1176 files changed, 79712 insertions, 64748 deletions

diff --git a/fs/9p/v9fs.c b/fs/9p/v9fs.c
index 61dbe52bb3..281a1ed03a 100644
--- a/fs/9p/v9fs.c
+++ b/fs/9p/v9fs.c
@@ -637,7 +637,7 @@ static int v9fs_init_inode_cache(void)
 	v9fs_inode_cache = kmem_cache_create("v9fs_inode_cache",
 					  sizeof(struct v9fs_inode),
 					  0, (SLAB_RECLAIM_ACCOUNT|
-					      SLAB_MEM_SPREAD|SLAB_ACCOUNT),
+					      SLAB_ACCOUNT),
 					  v9fs_inode_init_once);
 	if (!v9fs_inode_cache)
 		return -ENOMEM;
diff --git a/fs/9p/v9fs.h b/fs/9p/v9fs.h
index 698c43dd5d..1775fcc7f0 100644
--- a/fs/9p/v9fs.h
+++ b/fs/9p/v9fs.h
@@ -179,16 +179,14 @@ extern int v9fs_vfs_rename(struct mnt_idmap *idmap,
 			   struct inode *old_dir, struct dentry *old_dentry,
 			   struct inode *new_dir, struct dentry *new_dentry,
 			   unsigned int flags);
-extern struct inode *v9fs_inode_from_fid(struct v9fs_session_info *v9ses,
-					 struct p9_fid *fid,
-					 struct super_block *sb, int new);
+extern struct inode *v9fs_fid_iget(struct super_block *sb, struct p9_fid *fid,
+						bool new);
 extern const struct inode_operations v9fs_dir_inode_operations_dotl;
 extern const struct inode_operations v9fs_file_inode_operations_dotl;
 extern const struct inode_operations v9fs_symlink_inode_operations_dotl;
 extern const struct netfs_request_ops v9fs_req_ops;
-extern struct inode *v9fs_inode_from_fid_dotl(struct v9fs_session_info *v9ses,
-					      struct p9_fid *fid,
-					      struct super_block *sb, int new);
+extern struct inode *v9fs_fid_iget_dotl(struct super_block *sb,
+						struct p9_fid *fid, bool new);
 
 /* other default globals */
 #define V9FS_PORT	564
@@ -227,30 +225,12 @@ static inline int v9fs_proto_dotl(struct v9fs_session_info *v9ses)
  */
 static inline struct inode *
 v9fs_get_inode_from_fid(struct v9fs_session_info *v9ses, struct p9_fid *fid,
-			struct super_block *sb)
+			struct super_block *sb, bool new)
 {
 	if (v9fs_proto_dotl(v9ses))
-		return v9fs_inode_from_fid_dotl(v9ses, fid, sb, 0);
+		return v9fs_fid_iget_dotl(sb, fid, new);
 	else
-		return v9fs_inode_from_fid(v9ses, fid, sb, 0);
-}
-
-/**
- * v9fs_get_new_inode_from_fid - Helper routine to populate an inode by
- * issuing a attribute request
- * @v9ses: session information
- * @fid: fid to issue attribute request for
- * @sb: superblock on which to create inode
- *
- */
-static inline struct inode *
-v9fs_get_new_inode_from_fid(struct v9fs_session_info *v9ses, struct p9_fid *fid,
-			    struct super_block *sb)
-{
-	if (v9fs_proto_dotl(v9ses))
-		return v9fs_inode_from_fid_dotl(v9ses, fid, sb, 1);
-	else
-		return v9fs_inode_from_fid(v9ses, fid, sb, 1);
+		return v9fs_fid_iget(sb, fid, new);
 }
 
 #endif
diff --git a/fs/9p/v9fs_vfs.h b/fs/9p/v9fs_vfs.h
index 0e8418066a..7923c3c347 100644
--- a/fs/9p/v9fs_vfs.h
+++ b/fs/9p/v9fs_vfs.h
@@ -40,13 +40,16 @@ extern struct kmem_cache *v9fs_inode_cache;
 
 struct inode *v9fs_alloc_inode(struct super_block *sb);
 void v9fs_free_inode(struct inode *inode);
-struct inode *v9fs_get_inode(struct super_block *sb, umode_t mode,
-			     dev_t rdev);
 void v9fs_set_netfs_context(struct inode *inode);
 int v9fs_init_inode(struct v9fs_session_info *v9ses,
-		    struct inode *inode, umode_t mode, dev_t rdev);
+		    struct inode *inode, struct p9_qid *qid, umode_t mode, dev_t rdev);
 void v9fs_evict_inode(struct inode *inode);
-ino_t v9fs_qid2ino(struct p9_qid *qid);
+#if (BITS_PER_LONG == 32)
+#define QID2INO(q) ((ino_t) (((q)->path+2) ^ (((q)->path) >> 32)))
+#else
+#define QID2INO(q) ((ino_t) ((q)->path+2))
+#endif
+
 void v9fs_stat2inode(struct p9_wstat *stat, struct inode *inode,
 		      struct super_block *sb, unsigned int flags);
 void v9fs_stat2inode_dotl(struct p9_stat_dotl *stat, struct inode *inode,
diff --git a/fs/9p/vfs_addr.c b/fs/9p/vfs_addr.c
index 047855033d..a97ceb105c 100644
--- a/fs/9p/vfs_addr.c
+++ b/fs/9p/vfs_addr.c
@@ -26,36 +26,38 @@
 #include "cache.h"
 #include "fid.h"
 
-static void v9fs_upload_to_server(struct netfs_io_subrequest *subreq)
+/*
+ * Writeback calls this when it finds a folio that needs uploading.  This isn't
+ * called if writeback only has copy-to-cache to deal with.
+ */
+static void v9fs_begin_writeback(struct netfs_io_request *wreq)
 {
-	struct p9_fid *fid = subreq->rreq->netfs_priv;
-	int err, len;
-
-	trace_netfs_sreq(subreq, netfs_sreq_trace_submit);
-	len = p9_client_write(fid, subreq->start, &subreq->io_iter, &err);
-	netfs_write_subrequest_terminated(subreq, len ?: err, false);
-}
+	struct p9_fid *fid;
 
-static void v9fs_upload_to_server_worker(struct work_struct *work)
-{
-	struct netfs_io_subrequest *subreq =
-		container_of(work, struct netfs_io_subrequest, work);
+	fid = v9fs_fid_find_inode(wreq->inode, true, INVALID_UID, true);
+	if (!fid) {
+		WARN_ONCE(1, "folio expected an open fid inode->i_ino=%lx\n",
+			  wreq->inode->i_ino);
+		return;
+	}
 
-	v9fs_upload_to_server(subreq);
+	wreq->wsize = fid->clnt->msize - P9_IOHDRSZ;
+	if (fid->iounit)
+		wreq->wsize = min(wreq->wsize, fid->iounit);
+	wreq->netfs_priv = fid;
+	wreq->io_streams[0].avail = true;
 }
 
 /*
- * Set up write requests for a writeback slice.  We need to add a write request
- * for each write we want to make.
+ * Issue a subrequest to write to the server.
  */
-static void v9fs_create_write_requests(struct netfs_io_request *wreq, loff_t start, size_t len)
+static void v9fs_issue_write(struct netfs_io_subrequest *subreq)
 {
-	struct netfs_io_subrequest *subreq;
+	struct p9_fid *fid = subreq->rreq->netfs_priv;
+	int err, len;
 
-	subreq = netfs_create_write_request(wreq, NETFS_UPLOAD_TO_SERVER,
-					    start, len, v9fs_upload_to_server_worker);
-	if (subreq)
-		netfs_queue_write_request(subreq);
+	len = p9_client_write(fid, subreq->start, &subreq->io_iter, &err);
+	netfs_write_subrequest_terminated(subreq, len ?: err, false);
 }
 
 /**
@@ -87,12 +89,16 @@ static int v9fs_init_request(struct netfs_io_request *rreq, struct file *file)
 {
 	struct p9_fid *fid;
 	bool writing = (rreq->origin == NETFS_READ_FOR_WRITE ||
-			rreq->origin == NETFS_WRITEBACK ||
 			rreq->origin == NETFS_WRITETHROUGH ||
-			rreq->origin == NETFS_LAUNDER_WRITE ||
 			rreq->origin == NETFS_UNBUFFERED_WRITE ||
 			rreq->origin == NETFS_DIO_WRITE);
 
+	if (rreq->origin == NETFS_WRITEBACK)
+		return 0; /* We don't get the write handle until we find we
+			   * have actually dirty data and not just
+			   * copy-to-cache data.
+			   */
+
 	if (file) {
 		fid = file->private_data;
 		if (!fid)
@@ -104,6 +110,10 @@ static int v9fs_init_request(struct netfs_io_request *rreq, struct file *file)
 			goto no_fid;
 	}
 
+	rreq->wsize = fid->clnt->msize - P9_IOHDRSZ;
+	if (fid->iounit)
+		rreq->wsize = min(rreq->wsize, fid->iounit);
+
 	/* we might need to read from a fid that was opened write-only
 	 * for read-modify-write of page cache, use the writeback fid
 	 * for that */
@@ -132,7 +142,8 @@ const struct netfs_request_ops v9fs_req_ops = {
 	.init_request		= v9fs_init_request,
 	.free_request		= v9fs_free_request,
 	.issue_read		= v9fs_issue_read,
-	.create_write_requests	= v9fs_create_write_requests,
+	.begin_writeback	= v9fs_begin_writeback,
+	.issue_write		= v9fs_issue_write,
 };
 
 const struct address_space_operations v9fs_addr_operations = {
@@ -141,7 +152,6 @@ const struct address_space_operations v9fs_addr_operations = {
 	.dirty_folio		= netfs_dirty_folio,
 	.release_folio		= netfs_release_folio,
 	.invalidate_folio	= netfs_invalidate_folio,
-	.launder_folio		= netfs_launder_folio,
 	.direct_IO		= noop_direct_IO,
 	.writepages		= netfs_writepages,
 };
diff --git a/fs/9p/vfs_dentry.c b/fs/9p/vfs_dentry.c
index f16f735816..01338d4c2d 100644
--- a/fs/9p/vfs_dentry.c
+++ b/fs/9p/vfs_dentry.c
@@ -48,12 +48,17 @@ static int v9fs_cached_dentry_delete(const struct dentry *dentry)
 static void v9fs_dentry_release(struct dentry *dentry)
 {
 	struct hlist_node *p, *n;
+	struct hlist_head head;
 
 	p9_debug(P9_DEBUG_VFS, " dentry: %pd (%p)\n",
 		 dentry, dentry);
-	hlist_for_each_safe(p, n, (struct hlist_head *)&dentry->d_fsdata)
+
+	spin_lock(&dentry->d_lock);
+	hlist_move_list((struct hlist_head *)&dentry->d_fsdata, &head);
+	spin_unlock(&dentry->d_lock);
+
+	hlist_for_each_safe(p, n, &head)
 		p9_fid_put(hlist_entry(p, struct p9_fid, dlist));
-	dentry->d_fsdata = NULL;
 }
 
 static int v9fs_lookup_revalidate(struct dentry *dentry, unsigned int flags)
diff --git a/fs/9p/vfs_dir.c b/fs/9p/vfs_dir.c
index 4102759a5c..e0d34e4e90 100644
--- a/fs/9p/vfs_dir.c
+++ b/fs/9p/vfs_dir.c
@@ -127,7 +127,7 @@ static int v9fs_dir_readdir(struct file *file, struct dir_context *ctx)
 			}
 
 			over = !dir_emit(ctx, st.name, strlen(st.name),
-					 v9fs_qid2ino(&st.qid), dt_type(&st));
+					QID2INO(&st.qid), dt_type(&st));
 			p9stat_free(&st);
 			if (over)
 				return 0;
@@ -184,7 +184,7 @@ static int v9fs_dir_readdir_dotl(struct file *file, struct dir_context *ctx)
 
 			if (!dir_emit(ctx, curdirent.d_name,
 				      strlen(curdirent.d_name),
-				      v9fs_qid2ino(&curdirent.qid),
+				      QID2INO(&curdirent.qid),
 				      curdirent.d_type))
 				return 0;
 
diff --git a/fs/9p/vfs_file.c b/fs/9p/vfs_file.c
index a504240c81..348cc90bf9 100644
--- a/fs/9p/vfs_file.c
+++ b/fs/9p/vfs_file.c
@@ -107,7 +107,7 @@ static int v9fs_file_lock(struct file *filp, int cmd, struct file_lock *fl)
 
 	p9_debug(P9_DEBUG_VFS, "filp: %p lock: %p\n", filp, fl);
 
-	if ((IS_SETLK(cmd) || IS_SETLKW(cmd)) && fl->fl_type != F_UNLCK) {
+	if ((IS_SETLK(cmd) || IS_SETLKW(cmd)) && fl->c.flc_type != F_UNLCK) {
 		filemap_write_and_wait(inode->i_mapping);
 		invalidate_mapping_pages(&inode->i_data, 0, -1);
 	}
@@ -121,13 +121,12 @@ static int v9fs_file_do_lock(struct file *filp, int cmd, struct file_lock *fl)
 	struct p9_fid *fid;
 	uint8_t status = P9_LOCK_ERROR;
 	int res = 0;
-	unsigned char fl_type;
 	struct v9fs_session_info *v9ses;
 
 	fid = filp->private_data;
 	BUG_ON(fid == NULL);
 
-	BUG_ON((fl->fl_flags & FL_POSIX) != FL_POSIX);
+	BUG_ON((fl->c.flc_flags & FL_POSIX) != FL_POSIX);
 
 	res = locks_lock_file_wait(filp, fl);
 	if (res < 0)
@@ -136,7 +135,7 @@ static int v9fs_file_do_lock(struct file *filp, int cmd, struct file_lock *fl)
 	/* convert posix lock to p9 tlock args */
 	memset(&flock, 0, sizeof(flock));
 	/* map the lock type */
-	switch (fl->fl_type) {
+	switch (fl->c.flc_type) {
 	case F_RDLCK:
 		flock.type = P9_LOCK_TYPE_RDLCK;
 		break;
@@ -152,7 +151,7 @@ static int v9fs_file_do_lock(struct file *filp, int cmd, struct file_lock *fl)
 		flock.length = 0;
 	else
 		flock.length = fl->fl_end - fl->fl_start + 1;
-	flock.proc_id = fl->fl_pid;
+	flock.proc_id = fl->c.flc_pid;
 	flock.client_id = fid->clnt->name;
 	if (IS_SETLKW(cmd))
 		flock.flags = P9_LOCK_FLAGS_BLOCK;
@@ -207,12 +206,13 @@ out_unlock:
 	 * incase server returned error for lock request, revert
 	 * it locally
 	 */
-	if (res < 0 && fl->fl_type != F_UNLCK) {
-		fl_type = fl->fl_type;
-		fl->fl_type = F_UNLCK;
+	if (res < 0 && fl->c.flc_type != F_UNLCK) {
+		unsigned char type = fl->c.flc_type;
+
+		fl->c.flc_type = F_UNLCK;
 		/* Even if this fails we want to return the remote error */
 		locks_lock_file_wait(filp, fl);
-		fl->fl_type = fl_type;
+		fl->c.flc_type = type;
 	}
 	if (flock.client_id != fid->clnt->name)
 		kfree(flock.client_id);
@@ -234,7 +234,7 @@ static int v9fs_file_getlock(struct file *filp, struct file_lock *fl)
 	 * if we have a conflicting lock locally, no need to validate
 	 * with server
 	 */
-	if (fl->fl_type != F_UNLCK)
+	if (fl->c.flc_type != F_UNLCK)
 		return res;
 
 	/* convert posix lock to p9 tgetlock args */
@@ -245,7 +245,7 @@ static int v9fs_file_getlock(struct file *filp, struct file_lock *fl)
 		glock.length = 0;
 	else
 		glock.length = fl->fl_end - fl->fl_start + 1;
-	glock.proc_id = fl->fl_pid;
+	glock.proc_id = fl->c.flc_pid;
 	glock.client_id = fid->clnt->name;
 
 	res = p9_client_getlock_dotl(fid, &glock);
@@ -254,13 +254,13 @@ static int v9fs_file_getlock(struct file *filp, struct file_lock *fl)
 	/* map 9p lock type to os lock type */
 	switch (glock.type) {
 	case P9_LOCK_TYPE_RDLCK:
-		fl->fl_type = F_RDLCK;
+		fl->c.flc_type = F_RDLCK;
 		break;
 	case P9_LOCK_TYPE_WRLCK:
-		fl->fl_type = F_WRLCK;
+		fl->c.flc_type = F_WRLCK;
 		break;
 	case P9_LOCK_TYPE_UNLCK:
-		fl->fl_type = F_UNLCK;
+		fl->c.flc_type = F_UNLCK;
 		break;
 	}
 	if (glock.type != P9_LOCK_TYPE_UNLCK) {
@@ -269,7 +269,7 @@ static int v9fs_file_getlock(struct file *filp, struct file_lock *fl)
 			fl->fl_end = OFFSET_MAX;
 		else
 			fl->fl_end = glock.start + glock.length - 1;
-		fl->fl_pid = -glock.proc_id;
+		fl->c.flc_pid = -glock.proc_id;
 	}
 out:
 	if (glock.client_id != fid->clnt->name)
@@ -293,7 +293,7 @@ static int v9fs_file_lock_dotl(struct file *filp, int cmd, struct file_lock *fl)
 	p9_debug(P9_DEBUG_VFS, "filp: %p cmd:%d lock: %p name: %pD\n",
 		 filp, cmd, fl, filp);
 
-	if ((IS_SETLK(cmd) || IS_SETLKW(cmd)) && fl->fl_type != F_UNLCK) {
+	if ((IS_SETLK(cmd) || IS_SETLKW(cmd)) && fl->c.flc_type != F_UNLCK) {
 		filemap_write_and_wait(inode->i_mapping);
 		invalidate_mapping_pages(&inode->i_data, 0, -1);
 	}
@@ -324,16 +324,16 @@ static int v9fs_file_flock_dotl(struct file *filp, int cmd,
 	p9_debug(P9_DEBUG_VFS, "filp: %p cmd:%d lock: %p name: %pD\n",
 		 filp, cmd, fl, filp);
 
-	if (!(fl->fl_flags & FL_FLOCK))
+	if (!(fl->c.flc_flags & FL_FLOCK))
 		goto out_err;
 
-	if ((IS_SETLK(cmd) || IS_SETLKW(cmd)) && fl->fl_type != F_UNLCK) {
+	if ((IS_SETLK(cmd) || IS_SETLKW(cmd)) && fl->c.flc_type != F_UNLCK) {
 		filemap_write_and_wait(inode->i_mapping);
 		invalidate_mapping_pages(&inode->i_data, 0, -1);
 	}
 	/* Convert flock to posix lock */
-	fl->fl_flags |= FL_POSIX;
-	fl->fl_flags ^= FL_FLOCK;
+	fl->c.flc_flags |= FL_POSIX;
+	fl->c.flc_flags ^= FL_FLOCK;
 
 	if (IS_SETLK(cmd) | IS_SETLKW(cmd))
 		ret = v9fs_file_do_lock(filp, cmd, fl);
diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c
index 0fde0ab77e..fd72fc38c8 100644
--- a/fs/9p/vfs_inode.c
+++ b/fs/9p/vfs_inode.c
@@ -256,9 +256,12 @@ void v9fs_set_netfs_context(struct inode *inode)
 }
 
 int v9fs_init_inode(struct v9fs_session_info *v9ses,
-		    struct inode *inode, umode_t mode, dev_t rdev)
+		    struct inode *inode, struct p9_qid *qid, umode_t mode, dev_t rdev)
 {
 	int err = 0;
+	struct v9fs_inode *v9inode = V9FS_I(inode);
+
+	memcpy(&v9inode->qid, qid, sizeof(struct p9_qid));
 
 	inode_init_owner(&nop_mnt_idmap, inode, NULL, mode);
 	inode->i_blocks = 0;
@@ -335,36 +338,6 @@ error:
 }
 
 /**
- * v9fs_get_inode - helper function to setup an inode
- * @sb: superblock
- * @mode: mode to setup inode with
- * @rdev: The device numbers to set
- */
-
-struct inode *v9fs_get_inode(struct super_block *sb, umode_t mode, dev_t rdev)
-{
-	int err;
-	struct inode *inode;
-	struct v9fs_session_info *v9ses = sb->s_fs_info;
-
-	p9_debug(P9_DEBUG_VFS, "super block: %p mode: %ho\n", sb, mode);
-
-	inode = new_inode(sb);
-	if (!inode) {
-		pr_warn("%s (%d): Problem allocating inode\n",
-			__func__, task_pid_nr(current));
-		return ERR_PTR(-ENOMEM);
-	}
-	err = v9fs_init_inode(v9ses, inode, mode, rdev);
-	if (err) {
-		iput(inode);
-		return ERR_PTR(err);
-	}
-	v9fs_set_netfs_context(inode);
-	return inode;
-}
-
-/**
  * v9fs_evict_inode - Remove an inode from the inode cache
  * @inode: inode to release
  *
@@ -375,6 +348,7 @@ void v9fs_evict_inode(struct inode *inode)
 	__le32 __maybe_unused version;
 
 	if (!is_bad_inode(inode)) {
+		netfs_wait_for_outstanding_io(inode);
 		truncate_inode_pages_final(&inode->i_data);
 
 		version = cpu_to_le32(v9inode->qid.version);
@@ -391,107 +365,59 @@ void v9fs_evict_inode(struct inode *inode)
 		clear_inode(inode);
 }
 
-static int v9fs_test_inode(struct inode *inode, void *data)
-{
-	int umode;
-	dev_t rdev;
-	struct v9fs_inode *v9inode = V9FS_I(inode);
-	struct p9_wstat *st = (struct p9_wstat *)data;
-	struct v9fs_session_info *v9ses = v9fs_inode2v9ses(inode);
-
-	umode = p9mode2unixmode(v9ses, st, &rdev);
-	/* don't match inode of different type */
-	if (inode_wrong_type(inode, umode))
-		return 0;
-
-	/* compare qid details */
-	if (memcmp(&v9inode->qid.version,
-		   &st->qid.version, sizeof(v9inode->qid.version)))
-		return 0;
-
-	if (v9inode->qid.type != st->qid.type)
-		return 0;
-
-	if (v9inode->qid.path != st->qid.path)
-		return 0;
-	return 1;
-}
-
-static int v9fs_test_new_inode(struct inode *inode, void *data)
-{
-	return 0;
-}
-
-static int v9fs_set_inode(struct inode *inode,  void *data)
-{
-	struct v9fs_inode *v9inode = V9FS_I(inode);
-	struct p9_wstat *st = (struct p9_wstat *)data;
-
-	memcpy(&v9inode->qid, &st->qid, sizeof(st->qid));
-	return 0;
-}
-
-static struct inode *v9fs_qid_iget(struct super_block *sb,
-				   struct p9_qid *qid,
-				   struct p9_wstat *st,
-				   int new)
+struct inode *
+v9fs_fid_iget(struct super_block *sb, struct p9_fid *fid, bool new)
 {
 	dev_t rdev;
 	int retval;
 	umode_t umode;
-	unsigned long i_ino;
 	struct inode *inode;
+	struct p9_wstat *st;
 	struct v9fs_session_info *v9ses = sb->s_fs_info;
-	int (*test)(struct inode *inode, void *data);
-
-	if (new)
-		test = v9fs_test_new_inode;
-	else
-		test = v9fs_test_inode;
 
-	i_ino = v9fs_qid2ino(qid);
-	inode = iget5_locked(sb, i_ino, test, v9fs_set_inode, st);
-	if (!inode)
+	inode = iget_locked(sb, QID2INO(&fid->qid));
+	if (unlikely(!inode))
 		return ERR_PTR(-ENOMEM);
-	if (!(inode->i_state & I_NEW))
-		return inode;
+	if (!(inode->i_state & I_NEW)) {
+		if (!new) {
+			goto done;
+		} else {
+			p9_debug(P9_DEBUG_VFS, "WARNING: Inode collision %ld\n",
+						inode->i_ino);
+			iput(inode);
+			remove_inode_hash(inode);
+			inode = iget_locked(sb, QID2INO(&fid->qid));
+			WARN_ON(!(inode->i_state & I_NEW));
+		}
+	}
+
 	/*
 	 * initialize the inode with the stat info
 	 * FIXME!! we may need support for stale inodes
 	 * later.
 	 */
-	inode->i_ino = i_ino;
+	st = p9_client_stat(fid);
+	if (IS_ERR(st)) {
+		retval = PTR_ERR(st);
+		goto error;
+	}
+
 	umode = p9mode2unixmode(v9ses, st, &rdev);
-	retval = v9fs_init_inode(v9ses, inode, umode, rdev);
+	retval = v9fs_init_inode(v9ses, inode, &fid->qid, umode, rdev);
+	v9fs_stat2inode(st, inode, sb, 0);
+	p9stat_free(st);
+	kfree(st);
 	if (retval)
 		goto error;
 
-	v9fs_stat2inode(st, inode, sb, 0);
 	v9fs_set_netfs_context(inode);
 	v9fs_cache_inode_get_cookie(inode);
 	unlock_new_inode(inode);
+done:
 	return inode;
 error:
 	iget_failed(inode);
 	return ERR_PTR(retval);
-
-}
-
-struct inode *
-v9fs_inode_from_fid(struct v9fs_session_info *v9ses, struct p9_fid *fid,
-		    struct super_block *sb, int new)
-{
-	struct p9_wstat *st;
-	struct inode *inode = NULL;
-
-	st = p9_client_stat(fid);
-	if (IS_ERR(st))
-		return ERR_CAST(st);
-
-	inode = v9fs_qid_iget(sb, &st->qid, st, new);
-	p9stat_free(st);
-	kfree(st);
-	return inode;
 }
 
 /**
@@ -523,8 +449,15 @@ static int v9fs_at_to_dotl_flags(int flags)
  */
 static void v9fs_dec_count(struct inode *inode)
 {
-	if (!S_ISDIR(inode->i_mode) || inode->i_nlink > 2)
-		drop_nlink(inode);
+	if (!S_ISDIR(inode->i_mode) || inode->i_nlink > 2) {
+		if (inode->i_nlink) {
+			drop_nlink(inode);
+		} else {
+			p9_debug(P9_DEBUG_VFS,
+						"WARNING: unexpected i_nlink zero %d inode %ld\n",
+						inode->i_nlink, inode->i_ino);
+		}
+	}
 }
 
 /**
@@ -575,6 +508,9 @@ static int v9fs_remove(struct inode *dir, struct dentry *dentry, int flags)
 		} else
 			v9fs_dec_count(inode);
 
+		if (inode->i_nlink <= 0)	/* no more refs unhash it */
+			remove_inode_hash(inode);
+
 		v9fs_invalidate_inode_attr(inode);
 		v9fs_invalidate_inode_attr(dir);
 
@@ -640,7 +576,7 @@ v9fs_create(struct v9fs_session_info *v9ses, struct inode *dir,
 		/*
 		 * instantiate inode and assign the unopened fid to the dentry
 		 */
-		inode = v9fs_get_new_inode_from_fid(v9ses, fid, dir->i_sb);
+		inode = v9fs_get_inode_from_fid(v9ses, fid, dir->i_sb, true);
 		if (IS_ERR(inode)) {
 			err = PTR_ERR(inode);
 			p9_debug(P9_DEBUG_VFS,
@@ -768,10 +704,8 @@ struct dentry *v9fs_vfs_lookup(struct inode *dir, struct dentry *dentry,
 		inode = NULL;
 	else if (IS_ERR(fid))
 		inode = ERR_CAST(fid);
-	else if (v9ses->cache & (CACHE_META|CACHE_LOOSE))
-		inode = v9fs_get_inode_from_fid(v9ses, fid, dir->i_sb);
 	else
-		inode = v9fs_get_new_inode_from_fid(v9ses, fid, dir->i_sb);
+		inode = v9fs_get_inode_from_fid(v9ses, fid, dir->i_sb, false);
 	/*
 	 * If we had a rename on the server and a parallel lookup
 	 * for the new name, then make sure we instantiate with
@@ -1192,26 +1126,6 @@ v9fs_stat2inode(struct p9_wstat *stat, struct inode *inode,
 }
 
 /**
- * v9fs_qid2ino - convert qid into inode number
- * @qid: qid to hash
- *
- * BUG: potential for inode number collisions?
- */
-
-ino_t v9fs_qid2ino(struct p9_qid *qid)
-{
-	u64 path = qid->path + 2;
-	ino_t i = 0;
-
-	if (sizeof(ino_t) == sizeof(path))
-		memcpy(&i, &path, sizeof(ino_t));
-	else
-		i = (ino_t) (path ^ (path >> 32));
-
-	return i;
-}
-
-/**
  * v9fs_vfs_get_link - follow a symlink path
  * @dentry: dentry for symlink
  * @inode: inode for symlink
diff --git a/fs/9p/vfs_inode_dotl.c b/fs/9p/vfs_inode_dotl.c
index 3505227e17..c61b97bd13 100644
--- a/fs/9p/vfs_inode_dotl.c
+++ b/fs/9p/vfs_inode_dotl.c
@@ -52,82 +52,50 @@ static kgid_t v9fs_get_fsgid_for_create(struct inode *dir_inode)
 	return current_fsgid();
 }
 
-static int v9fs_test_inode_dotl(struct inode *inode, void *data)
-{
-	struct v9fs_inode *v9inode = V9FS_I(inode);
-	struct p9_stat_dotl *st = (struct p9_stat_dotl *)data;
-
-	/* don't match inode of different type */
-	if (inode_wrong_type(inode, st->st_mode))
-		return 0;
-
-	if (inode->i_generation != st->st_gen)
-		return 0;
-
-	/* compare qid details */
-	if (memcmp(&v9inode->qid.version,
-		   &st->qid.version, sizeof(v9inode->qid.version)))
-		return 0;
-
-	if (v9inode->qid.type != st->qid.type)
-		return 0;
 
-	if (v9inode->qid.path != st->qid.path)
-		return 0;
-	return 1;
-}
-
-/* Always get a new inode */
-static int v9fs_test_new_inode_dotl(struct inode *inode, void *data)
-{
-	return 0;
-}
 
-static int v9fs_set_inode_dotl(struct inode *inode,  void *data)
-{
-	struct v9fs_inode *v9inode = V9FS_I(inode);
-	struct p9_stat_dotl *st = (struct p9_stat_dotl *)data;
-
-	memcpy(&v9inode->qid, &st->qid, sizeof(st->qid));
-	inode->i_generation = st->st_gen;
-	return 0;
-}
-
-static struct inode *v9fs_qid_iget_dotl(struct super_block *sb,
-					struct p9_qid *qid,
-					struct p9_fid *fid,
-					struct p9_stat_dotl *st,
-					int new)
+struct inode *
+v9fs_fid_iget_dotl(struct super_block *sb, struct p9_fid *fid, bool new)
 {
 	int retval;
-	unsigned long i_ino;
 	struct inode *inode;
+	struct p9_stat_dotl *st;
 	struct v9fs_session_info *v9ses = sb->s_fs_info;
-	int (*test)(struct inode *inode, void *data);
 
-	if (new)
-		test = v9fs_test_new_inode_dotl;
-	else
-		test = v9fs_test_inode_dotl;
-
-	i_ino = v9fs_qid2ino(qid);
-	inode = iget5_locked(sb, i_ino, test, v9fs_set_inode_dotl, st);
-	if (!inode)
+	inode = iget_locked(sb, QID2INO(&fid->qid));
+	if (unlikely(!inode))
 		return ERR_PTR(-ENOMEM);
-	if (!(inode->i_state & I_NEW))
-		return inode;
+	if (!(inode->i_state & I_NEW)) {
+		if (!new) {
+			goto done;
+		} else { /* deal with race condition in inode number reuse */
+			p9_debug(P9_DEBUG_ERROR, "WARNING: Inode collision %lx\n",
+						inode->i_ino);
+			iput(inode);
+			remove_inode_hash(inode);
+			inode = iget_locked(sb, QID2INO(&fid->qid));
+			WARN_ON(!(inode->i_state & I_NEW));
+		}
+	}
+
 	/*
 	 * initialize the inode with the stat info
 	 * FIXME!! we may need support for stale inodes
 	 * later.
 	 */
-	inode->i_ino = i_ino;
-	retval = v9fs_init_inode(v9ses, inode,
+	st = p9_client_getattr_dotl(fid, P9_STATS_BASIC | P9_STATS_GEN);
+	if (IS_ERR(st)) {
+		retval = PTR_ERR(st);
+		goto error;
+	}
+
+	retval = v9fs_init_inode(v9ses, inode, &fid->qid,
 				 st->st_mode, new_decode_dev(st->st_rdev));
+	v9fs_stat2inode_dotl(st, inode, 0);
+	kfree(st);
 	if (retval)
 		goto error;
 
-	v9fs_stat2inode_dotl(st, inode, 0);
 	v9fs_set_netfs_context(inode);
 	v9fs_cache_inode_get_cookie(inode);
 	retval = v9fs_get_acl(inode, fid);
@@ -135,27 +103,11 @@ static struct inode *v9fs_qid_iget_dotl(struct super_block *sb,
 		goto error;
 
 	unlock_new_inode(inode);
+done:
 	return inode;
 error:
 	iget_failed(inode);
 	return ERR_PTR(retval);
-
-}
-
-struct inode *
-v9fs_inode_from_fid_dotl(struct v9fs_session_info *v9ses, struct p9_fid *fid,
-			 struct super_block *sb, int new)
-{
-	struct p9_stat_dotl *st;
-	struct inode *inode = NULL;
-
-	st = p9_client_getattr_dotl(fid, P9_STATS_BASIC | P9_STATS_GEN);
-	if (IS_ERR(st))
-		return ERR_CAST(st);
-
-	inode = v9fs_qid_iget_dotl(sb, &st->qid, fid, st, new);
-	kfree(st);
-	return inode;
 }
 
 struct dotl_openflag_map {
@@ -307,7 +259,7 @@ v9fs_vfs_atomic_open_dotl(struct inode *dir, struct dentry *dentry,
 		p9_debug(P9_DEBUG_VFS, "p9_client_walk failed %d\n", err);
 		goto out;
 	}
-	inode = v9fs_get_new_inode_from_fid(v9ses, fid, dir->i_sb);
+	inode = v9fs_fid_iget_dotl(dir->i_sb, fid, true);
 	if (IS_ERR(inode)) {
 		err = PTR_ERR(inode);
 		p9_debug(P9_DEBUG_VFS, "inode creation failed %d\n", err);
@@ -357,7 +309,6 @@ static int v9fs_vfs_mkdir_dotl(struct mnt_idmap *idmap,
 			       umode_t omode)
 {
 	int err;
-	struct v9fs_session_info *v9ses;
 	struct p9_fid *fid = NULL, *dfid = NULL;
 	kgid_t gid;
 	const unsigned char *name;
@@ -367,7 +318,6 @@ static int v9fs_vfs_mkdir_dotl(struct mnt_idmap *idmap,
 	struct posix_acl *dacl = NULL, *pacl = NULL;
 
 	p9_debug(P9_DEBUG_VFS, "name %pd\n", dentry);
-	v9ses = v9fs_inode2v9ses(dir);
 
 	omode |= S_IFDIR;
 	if (dir->i_mode & S_ISGID)
@@ -402,32 +352,17 @@ static int v9fs_vfs_mkdir_dotl(struct mnt_idmap *idmap,
 	}
 
 	/* instantiate inode and assign the unopened fid to the dentry */
-	if (v9ses->cache & (CACHE_META|CACHE_LOOSE)) {
-		inode = v9fs_get_new_inode_from_fid(v9ses, fid, dir->i_sb);
-		if (IS_ERR(inode)) {
-			err = PTR_ERR(inode);
-			p9_debug(P9_DEBUG_VFS, "inode creation failed %d\n",
-				 err);
-			goto error;
-		}
-		v9fs_fid_add(dentry, &fid);
-		v9fs_set_create_acl(inode, fid, dacl, pacl);
-		d_instantiate(dentry, inode);
-		err = 0;
-	} else {
-		/*
-		 * Not in cached mode. No need to populate
-		 * inode with stat. We need to get an inode
-		 * so that we can set the acl with dentry
-		 */
-		inode = v9fs_get_inode(dir->i_sb, mode, 0);
-		if (IS_ERR(inode)) {
-			err = PTR_ERR(inode);
-			goto error;
-		}
-		v9fs_set_create_acl(inode, fid, dacl, pacl);
-		d_instantiate(dentry, inode);
+	inode = v9fs_fid_iget_dotl(dir->i_sb, fid, true);
+	if (IS_ERR(inode)) {
+		err = PTR_ERR(inode);
+		p9_debug(P9_DEBUG_VFS, "inode creation failed %d\n",
+			 err);
+		goto error;
 	}
+	v9fs_fid_add(dentry, &fid);
+	v9fs_set_create_acl(inode, fid, dacl, pacl);
+	d_instantiate(dentry, inode);
+	err = 0;
 	inc_nlink(dir);
 	v9fs_invalidate_inode_attr(dir);
 error:
@@ -709,14 +644,11 @@ v9fs_vfs_symlink_dotl(struct mnt_idmap *idmap, struct inode *dir,
 	kgid_t gid;
 	const unsigned char *name;
 	struct p9_qid qid;
-	struct inode *inode;
 	struct p9_fid *dfid;
 	struct p9_fid *fid = NULL;
-	struct v9fs_session_info *v9ses;
 
 	name = dentry->d_name.name;
 	p9_debug(P9_DEBUG_VFS, "%lu,%s,%s\n", dir->i_ino, name, symname);
-	v9ses = v9fs_inode2v9ses(dir);
 
 	dfid = v9fs_parent_fid(dentry);
 	if (IS_ERR(dfid)) {
@@ -736,36 +668,6 @@ v9fs_vfs_symlink_dotl(struct mnt_idmap *idmap, struct inode *dir,
 	}
 
 	v9fs_invalidate_inode_attr(dir);
-	if (v9ses->cache & (CACHE_META|CACHE_LOOSE)) {
-		/* Now walk from the parent so we can get an unopened fid. */
-		fid = p9_client_walk(dfid, 1, &name, 1);
-		if (IS_ERR(fid)) {
-			err = PTR_ERR(fid);
-			p9_debug(P9_DEBUG_VFS, "p9_client_walk failed %d\n",
-				 err);
-			goto error;
-		}
-
-		/* instantiate inode and assign the unopened fid to dentry */
-		inode = v9fs_get_new_inode_from_fid(v9ses, fid, dir->i_sb);
-		if (IS_ERR(inode)) {
-			err = PTR_ERR(inode);
-			p9_debug(P9_DEBUG_VFS, "inode creation failed %d\n",
-				 err);
-			goto error;
-		}
-		v9fs_fid_add(dentry, &fid);
-		d_instantiate(dentry, inode);
-		err = 0;
-	} else {
-		/* Not in cached mode. No need to populate inode with stat */
-		inode = v9fs_get_inode(dir->i_sb, S_IFLNK, 0);
-		if (IS_ERR(inode)) {
-			err = PTR_ERR(inode);
-			goto error;
-		}
-		d_instantiate(dentry, inode);
-	}
 
 error:
 	p9_fid_put(fid);
@@ -847,7 +749,6 @@ v9fs_vfs_mknod_dotl(struct mnt_idmap *idmap, struct inode *dir,
 	kgid_t gid;
 	const unsigned char *name;
 	umode_t mode;
-	struct v9fs_session_info *v9ses;
 	struct p9_fid *fid = NULL, *dfid = NULL;
 	struct inode *inode;
 	struct p9_qid qid;
@@ -857,7 +758,6 @@ v9fs_vfs_mknod_dotl(struct mnt_idmap *idmap, struct inode *dir,
 		 dir->i_ino, dentry, omode,
 		 MAJOR(rdev), MINOR(rdev));
 
-	v9ses = v9fs_inode2v9ses(dir);
 	dfid = v9fs_parent_fid(dentry);
 	if (IS_ERR(dfid)) {
 		err = PTR_ERR(dfid);
@@ -888,33 +788,17 @@ v9fs_vfs_mknod_dotl(struct mnt_idmap *idmap, struct inode *dir,
 			 err);
 		goto error;
 	}
-
-	/* instantiate inode and assign the unopened fid to the dentry */
-	if (v9ses->cache & (CACHE_META|CACHE_LOOSE)) {
-		inode = v9fs_get_new_inode_from_fid(v9ses, fid, dir->i_sb);
-		if (IS_ERR(inode)) {
-			err = PTR_ERR(inode);
-			p9_debug(P9_DEBUG_VFS, "inode creation failed %d\n",
-				 err);
-			goto error;
-		}
-		v9fs_set_create_acl(inode, fid, dacl, pacl);
-		v9fs_fid_add(dentry, &fid);
-		d_instantiate(dentry, inode);
-		err = 0;
-	} else {
-		/*
-		 * Not in cached mode. No need to populate inode with stat.
-		 * socket syscall returns a fd, so we need instantiate
-		 */
-		inode = v9fs_get_inode(dir->i_sb, mode, rdev);
-		if (IS_ERR(inode)) {
-			err = PTR_ERR(inode);
-			goto error;
-		}
-		v9fs_set_create_acl(inode, fid, dacl, pacl);
-		d_instantiate(dentry, inode);
+	inode = v9fs_fid_iget_dotl(dir->i_sb, fid, true);
+	if (IS_ERR(inode)) {
+		err = PTR_ERR(inode);
+		p9_debug(P9_DEBUG_VFS, "inode creation failed %d\n",
+			 err);
+		goto error;
 	}
+	v9fs_set_create_acl(inode, fid, dacl, pacl);
+	v9fs_fid_add(dentry, &fid);
+	d_instantiate(dentry, inode);
+	err = 0;
 error:
 	p9_fid_put(fid);
 	v9fs_put_acl(dacl, pacl);
diff --git a/fs/9p/vfs_super.c b/fs/9p/vfs_super.c
index 23cc67f29a..f52fdf4294 100644
--- a/fs/9p/vfs_super.c
+++ b/fs/9p/vfs_super.c
@@ -110,7 +110,6 @@ static struct dentry *v9fs_mount(struct file_system_type *fs_type, int flags,
 	struct inode *inode = NULL;
 	struct dentry *root = NULL;
 	struct v9fs_session_info *v9ses = NULL;
-	umode_t mode = 0777 | S_ISVTX;
 	struct p9_fid *fid;
 	int retval = 0;
 
@@ -140,7 +139,7 @@ static struct dentry *v9fs_mount(struct file_system_type *fs_type, int flags,
 	else
 		sb->s_d_op = &v9fs_dentry_operations;
 
-	inode = v9fs_get_inode(sb, S_IFDIR | mode, 0);
+	inode = v9fs_get_inode_from_fid(v9ses, fid, sb, true);
 	if (IS_ERR(inode)) {
 		retval = PTR_ERR(inode);
 		goto release_sb;
@@ -152,32 +151,6 @@ static struct dentry *v9fs_mount(struct file_system_type *fs_type, int flags,
 		goto release_sb;
 	}
 	sb->s_root = root;
-	if (v9fs_proto_dotl(v9ses)) {
-		struct p9_stat_dotl *st = NULL;
-
-		st = p9_client_getattr_dotl(fid, P9_STATS_BASIC);
-		if (IS_ERR(st)) {
-			retval = PTR_ERR(st);
-			goto release_sb;
-		}
-		d_inode(root)->i_ino = v9fs_qid2ino(&st->qid);
-		v9fs_stat2inode_dotl(st, d_inode(root), 0);
-		kfree(st);
-	} else {
-		struct p9_wstat *st = NULL;
-
-		st = p9_client_stat(fid);
-		if (IS_ERR(st)) {
-			retval = PTR_ERR(st);
-			goto release_sb;
-		}
-
-		d_inode(root)->i_ino = v9fs_qid2ino(&st->qid);
-		v9fs_stat2inode(st, d_inode(root), sb, 0);
-
-		p9stat_free(st);
-		kfree(st);
-	}
 	retval = v9fs_get_acl(inode, fid);
 	if (retval)
 		goto release_sb;
diff --git a/fs/Kconfig b/fs/Kconfig
index 89fdbefd10..a46b0cbc4d 100644
--- a/fs/Kconfig
+++ b/fs/Kconfig
@@ -60,7 +60,6 @@ endif # BLOCK
 config FS_DAX
 	bool "File system based Direct Access (DAX) support"
 	depends on MMU
-	depends on !(ARM || MIPS || SPARC)
 	depends on ZONE_DEVICE || FS_DAX_LIMITED
 	select FS_IOMAP
 	select DAX
@@ -162,7 +161,6 @@ menu "DOS/FAT/EXFAT/NT Filesystems"
 
 source "fs/fat/Kconfig"
 source "fs/exfat/Kconfig"
-source "fs/ntfs/Kconfig"
 source "fs/ntfs3/Kconfig"
 
 endmenu
@@ -262,6 +260,7 @@ menuconfig HUGETLBFS
 	depends on X86 || SPARC64 || ARCH_SUPPORTS_HUGETLBFS || BROKEN
 	depends on (SYSFS || SYSCTL)
 	select MEMFD_CREATE
+	select PADATA if SMP
 	help
 	  hugetlbfs is a filesystem backing for HugeTLB pages, based on
 	  ramfs. For architectures that support it, say Y here and read
diff --git a/fs/Makefile b/fs/Makefile
index c09016257f..6ecc9b0a53 100644
--- a/fs/Makefile
+++ b/fs/Makefile
@@ -15,7 +15,7 @@ obj-y :=	open.o read_write.o file_table.o super.o \
 		pnode.o splice.o sync.o utimes.o d_path.o \
 		stack.o fs_struct.o statfs.o fs_pin.o nsfs.o \
 		fs_types.o fs_context.o fs_parser.o fsopen.o init.o \
-		kernel_read_file.o mnt_idmapping.o remap_range.o
+		kernel_read_file.o mnt_idmapping.o remap_range.o pidfs.o
 
 obj-$(CONFIG_BUFFER_HEAD)	+= buffer.o mpage.o
 obj-$(CONFIG_PROC_FS)		+= proc_namespace.o
@@ -91,7 +91,6 @@ obj-y				+= unicode/
 obj-$(CONFIG_SYSV_FS)		+= sysv/
 obj-$(CONFIG_SMBFS)		+= smb/
 obj-$(CONFIG_HPFS_FS)		+= hpfs/
-obj-$(CONFIG_NTFS_FS)		+= ntfs/
 obj-$(CONFIG_NTFS3_FS)		+= ntfs3/
 obj-$(CONFIG_UFS_FS)		+= ufs/
 obj-$(CONFIG_EFS_FS)		+= efs/
diff --git a/fs/adfs/super.c b/fs/adfs/super.c
index e8bfc38239..9354b14bbf 100644
--- a/fs/adfs/super.c
+++ b/fs/adfs/super.c
@@ -249,7 +249,7 @@ static int __init init_inodecache(void)
 	adfs_inode_cachep = kmem_cache_create("adfs_inode_cache",
 					     sizeof(struct adfs_inode_info),
 					     0, (SLAB_RECLAIM_ACCOUNT|
-						SLAB_MEM_SPREAD|SLAB_ACCOUNT),
+						SLAB_ACCOUNT),
 					     init_once);
 	if (adfs_inode_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/affs/super.c b/fs/affs/super.c
index b56a95cf41..3c58213396 100644
--- a/fs/affs/super.c
+++ b/fs/affs/super.c
@@ -130,8 +130,7 @@ static int __init init_inodecache(void)
 {
 	affs_inode_cachep = kmem_cache_create("affs_inode_cache",
 					     sizeof(struct affs_inode_info),
-					     0, (SLAB_RECLAIM_ACCOUNT|
-						SLAB_MEM_SPREAD|SLAB_ACCOUNT),
+					     0, (SLAB_RECLAIM_ACCOUNT | SLAB_ACCOUNT),
 					     init_once);
 	if (affs_inode_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/afs/file.c b/fs/afs/file.c
index ef2cc8f565..c3f0c45ae9 100644
--- a/fs/afs/file.c
+++ b/fs/afs/file.c
@@ -54,7 +54,6 @@ const struct address_space_operations afs_file_aops = {
 	.read_folio	= netfs_read_folio,
 	.readahead	= netfs_readahead,
 	.dirty_folio	= netfs_dirty_folio,
-	.launder_folio	= netfs_launder_folio,
 	.release_folio	= netfs_release_folio,
 	.invalidate_folio = netfs_invalidate_folio,
 	.migrate_folio	= filemap_migrate_folio,
@@ -354,7 +353,7 @@ static int afs_init_request(struct netfs_io_request *rreq, struct file *file)
 	if (file)
 		rreq->netfs_priv = key_get(afs_file_key(file));
 	rreq->rsize = 256 * 1024;
-	rreq->wsize = 256 * 1024;
+	rreq->wsize = 256 * 1024 * 1024;
 	return 0;
 }
 
@@ -369,6 +368,7 @@ static int afs_check_write_begin(struct file *file, loff_t pos, unsigned len,
 static void afs_free_request(struct netfs_io_request *rreq)
 {
 	key_put(rreq->netfs_priv);
+	afs_put_wb_key(rreq->netfs_priv2);
 }
 
 static void afs_update_i_size(struct inode *inode, loff_t new_i_size)
@@ -400,7 +400,9 @@ const struct netfs_request_ops afs_req_ops = {
 	.issue_read		= afs_issue_read,
 	.update_i_size		= afs_update_i_size,
 	.invalidate_cache	= afs_netfs_invalidate_cache,
-	.create_write_requests	= afs_create_write_requests,
+	.begin_writeback	= afs_begin_writeback,
+	.prepare_write		= afs_prepare_write,
+	.issue_write		= afs_issue_write,
 };
 
 static void afs_add_open_mmap(struct afs_vnode *vnode)
diff --git a/fs/afs/flock.c b/fs/afs/flock.c
index 9c6dea3139..f0e96a3509 100644
--- a/fs/afs/flock.c
+++ b/fs/afs/flock.c
@@ -93,13 +93,13 @@ static void afs_grant_locks(struct afs_vnode *vnode)
 	bool exclusive = (vnode->lock_type == AFS_LOCK_WRITE);
 
 	list_for_each_entry_safe(p, _p, &vnode->pending_locks, fl_u.afs.link) {
-		if (!exclusive && p->fl_type == F_WRLCK)
+		if (!exclusive && lock_is_write(p))
 			continue;
 
 		list_move_tail(&p->fl_u.afs.link, &vnode->granted_locks);
 		p->fl_u.afs.state = AFS_LOCK_GRANTED;
 		trace_afs_flock_op(vnode, p, afs_flock_op_grant);
-		wake_up(&p->fl_wait);
+		locks_wake_up(p);
 	}
 }
 
@@ -112,25 +112,24 @@ static void afs_next_locker(struct afs_vnode *vnode, int error)
 {
 	struct file_lock *p, *_p, *next = NULL;
 	struct key *key = vnode->lock_key;
-	unsigned int fl_type = F_RDLCK;
+	unsigned int type = F_RDLCK;
 
 	_enter("");
 
 	if (vnode->lock_type == AFS_LOCK_WRITE)
-		fl_type = F_WRLCK;
+		type = F_WRLCK;
 
 	list_for_each_entry_safe(p, _p, &vnode->pending_locks, fl_u.afs.link) {
 		if (error &&
-		    p->fl_type == fl_type &&
-		    afs_file_key(p->fl_file) == key) {
+		    p->c.flc_type == type &&
+		    afs_file_key(p->c.flc_file) == key) {
 			list_del_init(&p->fl_u.afs.link);
 			p->fl_u.afs.state = error;
-			wake_up(&p->fl_wait);
+			locks_wake_up(p);
 		}
 
 		/* Select the next locker to hand off to. */
-		if (next &&
-		    (next->fl_type == F_WRLCK || p->fl_type == F_RDLCK))
+		if (next && (lock_is_write(next) || lock_is_read(p)))
 			continue;
 		next = p;
 	}
@@ -142,7 +141,7 @@ static void afs_next_locker(struct afs_vnode *vnode, int error)
 		afs_set_lock_state(vnode, AFS_VNODE_LOCK_SETTING);
 		next->fl_u.afs.state = AFS_LOCK_YOUR_TRY;
 		trace_afs_flock_op(vnode, next, afs_flock_op_wake);
-		wake_up(&next->fl_wait);
+		locks_wake_up(next);
 	} else {
 		afs_set_lock_state(vnode, AFS_VNODE_LOCK_NONE);
 		trace_afs_flock_ev(vnode, NULL, afs_flock_no_lockers, 0);
@@ -166,7 +165,7 @@ static void afs_kill_lockers_enoent(struct afs_vnode *vnode)
 			       struct file_lock, fl_u.afs.link);
 		list_del_init(&p->fl_u.afs.link);
 		p->fl_u.afs.state = -ENOENT;
-		wake_up(&p->fl_wait);
+		locks_wake_up(p);
 	}
 
 	key_put(vnode->lock_key);
@@ -464,14 +463,14 @@ static int afs_do_setlk(struct file *file, struct file_lock *fl)
 
 	_enter("{%llx:%llu},%llu-%llu,%u,%u",
 	       vnode->fid.vid, vnode->fid.vnode,
-	       fl->fl_start, fl->fl_end, fl->fl_type, mode);
+	       fl->fl_start, fl->fl_end, fl->c.flc_type, mode);
 
 	fl->fl_ops = &afs_lock_ops;
 	INIT_LIST_HEAD(&fl->fl_u.afs.link);
 	fl->fl_u.afs.state = AFS_LOCK_PENDING;
 
 	partial = (fl->fl_start != 0 || fl->fl_end != OFFSET_MAX);
-	type = (fl->fl_type == F_RDLCK) ? AFS_LOCK_READ : AFS_LOCK_WRITE;
+	type = lock_is_read(fl) ? AFS_LOCK_READ : AFS_LOCK_WRITE;
 	if (mode == afs_flock_mode_write && partial)
 		type = AFS_LOCK_WRITE;
 
@@ -524,7 +523,7 @@ static int afs_do_setlk(struct file *file, struct file_lock *fl)
 	}
 
 	if (vnode->lock_state == AFS_VNODE_LOCK_NONE &&
-	    !(fl->fl_flags & FL_SLEEP)) {
+	    !(fl->c.flc_flags & FL_SLEEP)) {
 		ret = -EAGAIN;
 		if (type == AFS_LOCK_READ) {
 			if (vnode->status.lock_count == -1)
@@ -621,7 +620,7 @@ skip_server_lock:
 	return 0;
 
 lock_is_contended:
-	if (!(fl->fl_flags & FL_SLEEP)) {
+	if (!(fl->c.flc_flags & FL_SLEEP)) {
 		list_del_init(&fl->fl_u.afs.link);
 		afs_next_locker(vnode, 0);
 		ret = -EAGAIN;
@@ -641,7 +640,7 @@ need_to_wait:
 	spin_unlock(&vnode->lock);
 
 	trace_afs_flock_ev(vnode, fl, afs_flock_waiting, 0);
-	ret = wait_event_interruptible(fl->fl_wait,
+	ret = wait_event_interruptible(fl->c.flc_wait,
 				       fl->fl_u.afs.state != AFS_LOCK_PENDING);
 	trace_afs_flock_ev(vnode, fl, afs_flock_waited, ret);
 
@@ -704,7 +703,8 @@ static int afs_do_unlk(struct file *file, struct file_lock *fl)
 	struct afs_vnode *vnode = AFS_FS_I(file_inode(file));
 	int ret;
 
-	_enter("{%llx:%llu},%u", vnode->fid.vid, vnode->fid.vnode, fl->fl_type);
+	_enter("{%llx:%llu},%u", vnode->fid.vid, vnode->fid.vnode,
+	       fl->c.flc_type);
 
 	trace_afs_flock_op(vnode, fl, afs_flock_op_unlock);
 
@@ -730,11 +730,11 @@ static int afs_do_getlk(struct file *file, struct file_lock *fl)
 	if (vnode->lock_state == AFS_VNODE_LOCK_DELETED)
 		return -ENOENT;
 
-	fl->fl_type = F_UNLCK;
+	fl->c.flc_type = F_UNLCK;
 
 	/* check local lock records first */
 	posix_test_lock(file, fl);
-	if (fl->fl_type == F_UNLCK) {
+	if (lock_is_unlock(fl)) {
 		/* no local locks; consult the server */
 		ret = afs_fetch_status(vnode, key, false, NULL);
 		if (ret < 0)
@@ -743,18 +743,18 @@ static int afs_do_getlk(struct file *file, struct file_lock *fl)
 		lock_count = READ_ONCE(vnode->status.lock_count);
 		if (lock_count != 0) {
 			if (lock_count > 0)
-				fl->fl_type = F_RDLCK;
+				fl->c.flc_type = F_RDLCK;
 			else
-				fl->fl_type = F_WRLCK;
+				fl->c.flc_type = F_WRLCK;
 			fl->fl_start = 0;
 			fl->fl_end = OFFSET_MAX;
-			fl->fl_pid = 0;
+			fl->c.flc_pid = 0;
 		}
 	}
 
 	ret = 0;
 error:
-	_leave(" = %d [%hd]", ret, fl->fl_type);
+	_leave(" = %d [%hd]", ret, fl->c.flc_type);
 	return ret;
 }
 
@@ -769,7 +769,7 @@ int afs_lock(struct file *file, int cmd, struct file_lock *fl)
 
 	_enter("{%llx:%llu},%d,{t=%x,fl=%x,r=%Ld:%Ld}",
 	       vnode->fid.vid, vnode->fid.vnode, cmd,
-	       fl->fl_type, fl->fl_flags,
+	       fl->c.flc_type, fl->c.flc_flags,
 	       (long long) fl->fl_start, (long long) fl->fl_end);
 
 	if (IS_GETLK(cmd))
@@ -778,7 +778,7 @@ int afs_lock(struct file *file, int cmd, struct file_lock *fl)
 	fl->fl_u.afs.debug_id = atomic_inc_return(&afs_file_lock_debug_id);
 	trace_afs_flock_op(vnode, fl, afs_flock_op_lock);
 
-	if (fl->fl_type == F_UNLCK)
+	if (lock_is_unlock(fl))
 		ret = afs_do_unlk(file, fl);
 	else
 		ret = afs_do_setlk(file, fl);
@@ -804,7 +804,7 @@ int afs_flock(struct file *file, int cmd, struct file_lock *fl)
 
 	_enter("{%llx:%llu},%d,{t=%x,fl=%x}",
 	       vnode->fid.vid, vnode->fid.vnode, cmd,
-	       fl->fl_type, fl->fl_flags);
+	       fl->c.flc_type, fl->c.flc_flags);
 
 	/*
 	 * No BSD flocks over NFS allowed.
@@ -813,14 +813,14 @@ int afs_flock(struct file *file, int cmd, struct file_lock *fl)
 	 * Not sure whether that would be unique, though, or whether
 	 * that would break in other places.
 	 */
-	if (!(fl->fl_flags & FL_FLOCK))
+	if (!(fl->c.flc_flags & FL_FLOCK))
 		return -ENOLCK;
 
 	fl->fl_u.afs.debug_id = atomic_inc_return(&afs_file_lock_debug_id);
 	trace_afs_flock_op(vnode, fl, afs_flock_op_flock);
 
 	/* we're simulating flock() locks using posix locks on the server */
-	if (fl->fl_type == F_UNLCK)
+	if (lock_is_unlock(fl))
 		ret = afs_do_unlk(file, fl);
 	else
 		ret = afs_do_setlk(file, fl);
@@ -843,7 +843,7 @@ int afs_flock(struct file *file, int cmd, struct file_lock *fl)
  */
 static void afs_fl_copy_lock(struct file_lock *new, struct file_lock *fl)
 {
-	struct afs_vnode *vnode = AFS_FS_I(file_inode(fl->fl_file));
+	struct afs_vnode *vnode = AFS_FS_I(file_inode(fl->c.flc_file));
 
 	_enter("");
 
@@ -861,7 +861,7 @@ static void afs_fl_copy_lock(struct file_lock *new, struct file_lock *fl)
  */
 static void afs_fl_release_private(struct file_lock *fl)
 {
-	struct afs_vnode *vnode = AFS_FS_I(file_inode(fl->fl_file));
+	struct afs_vnode *vnode = AFS_FS_I(file_inode(fl->c.flc_file));
 
 	_enter("");
 
diff --git a/fs/afs/inode.c b/fs/afs/inode.c
index 94fc049aff..3acf5e0500 100644
--- a/fs/afs/inode.c
+++ b/fs/afs/inode.c
@@ -512,7 +512,7 @@ static int afs_iget5_set_root(struct inode *inode, void *opaque)
 	struct afs_vnode *vnode = AFS_FS_I(inode);
 
 	vnode->volume		= as->volume;
-	vnode->fid.vid		= as->volume->vid,
+	vnode->fid.vid		= as->volume->vid;
 	vnode->fid.vnode	= 1;
 	vnode->fid.unique	= 1;
 	inode->i_ino		= 1;
@@ -545,7 +545,7 @@ struct inode *afs_root_iget(struct super_block *sb, struct key *key)
 	BUG_ON(!(inode->i_state & I_NEW));
 
 	vnode = AFS_FS_I(inode);
-	vnode->cb_v_check = atomic_read(&as->volume->cb_v_break),
+	vnode->cb_v_check = atomic_read(&as->volume->cb_v_break);
 	afs_set_netfs_context(vnode);
 
 	op = afs_alloc_operation(key, as->volume);
@@ -648,6 +648,7 @@ void afs_evict_inode(struct inode *inode)
 
 	ASSERTCMP(inode->i_ino, ==, vnode->fid.vnode);
 
+	netfs_wait_for_outstanding_io(inode);
 	truncate_inode_pages_final(&inode->i_data);
 
 	afs_set_cache_aux(vnode, &aux);
diff --git a/fs/afs/internal.h b/fs/afs/internal.h
index 6ce5a61293..6e1d3c4daf 100644
--- a/fs/afs/internal.h
+++ b/fs/afs/internal.h
@@ -916,7 +916,6 @@ struct afs_operation {
 			loff_t	pos;
 			loff_t	size;
 			loff_t	i_size;
-			bool	laundering;	/* Laundering page, PG_writeback not set */
 		} store;
 		struct {
 			struct iattr	*attr;
@@ -1599,11 +1598,14 @@ extern int afs_check_volume_status(struct afs_volume *, struct afs_operation *);
 /*
  * write.c
  */
+void afs_prepare_write(struct netfs_io_subrequest *subreq);
+void afs_issue_write(struct netfs_io_subrequest *subreq);
+void afs_begin_writeback(struct netfs_io_request *wreq);
+void afs_retry_request(struct netfs_io_request *wreq, struct netfs_io_stream *stream);
 extern int afs_writepages(struct address_space *, struct writeback_control *);
 extern int afs_fsync(struct file *, loff_t, loff_t, int);
 extern vm_fault_t afs_page_mkwrite(struct vm_fault *vmf);
 extern void afs_prune_wb_keys(struct afs_vnode *);
-void afs_create_write_requests(struct netfs_io_request *wreq, loff_t start, size_t len);
 
 /*
  * xattr.c
diff --git a/fs/afs/mntpt.c b/fs/afs/mntpt.c
index 97f50e9fd9..297487ee83 100644
--- a/fs/afs/mntpt.c
+++ b/fs/afs/mntpt.c
@@ -140,6 +140,11 @@ static int afs_mntpt_set_params(struct fs_context *fc, struct dentry *mntpt)
 		put_page(page);
 		if (ret < 0)
 			return ret;
+
+		/* Don't cross a backup volume mountpoint from a backup volume */
+		if (src_as->volume && src_as->volume->type == AFSVL_BACKVOL &&
+		    ctx->type == AFSVL_BACKVOL)
+			return -ENODEV;
 	}
 
 	return 0;
diff --git a/fs/afs/rotate.c b/fs/afs/rotate.c
index ed04bd1eea..ed09d4d4c2 100644
--- a/fs/afs/rotate.c
+++ b/fs/afs/rotate.c
@@ -541,11 +541,13 @@ pick_server:
 		    test_bit(AFS_SE_EXCLUDED, &se->flags) ||
 		    !test_bit(AFS_SERVER_FL_RESPONDING, &s->flags))
 			continue;
-		es = op->server_states->endpoint_state;
+		es = op->server_states[i].endpoint_state;
 		sal = es->addresses;
 
 		afs_get_address_preferences_rcu(op->net, sal);
 		for (j = 0; j < sal->nr_addrs; j++) {
+			if (es->failed_set & (1 << j))
+				continue;
 			if (!sal->addrs[j].peer)
 				continue;
 			if (sal->addrs[j].prio > best_prio) {
@@ -605,6 +607,8 @@ iterate_address:
 	best_prio = -1;
 	addr_index = 0;
 	for (i = 0; i < alist->nr_addrs; i++) {
+		if (!(set & (1 << i)))
+			continue;
 		if (alist->addrs[i].prio > best_prio) {
 			addr_index = i;
 			best_prio = alist->addrs[i].prio;
@@ -674,7 +678,7 @@ no_more_servers:
 	for (i = 0; i < op->server_list->nr_servers; i++) {
 		struct afs_endpoint_state *estate;
 
-		estate = op->server_states->endpoint_state;
+		estate = op->server_states[i].endpoint_state;
 		error = READ_ONCE(estate->error);
 		if (error < 0)
 			afs_op_accumulate_error(op, error, estate->abort_code);
diff --git a/fs/afs/validation.c b/fs/afs/validation.c
index 32a53fc8df..bef8af12eb 100644
--- a/fs/afs/validation.c
+++ b/fs/afs/validation.c
@@ -365,9 +365,9 @@ static void afs_zap_data(struct afs_vnode *vnode)
 	 * written back in a regular file and completely discard the pages in a
 	 * directory or symlink */
 	if (S_ISREG(vnode->netfs.inode.i_mode))
-		invalidate_remote_inode(&vnode->netfs.inode);
+		filemap_invalidate_inode(&vnode->netfs.inode, true, 0, LLONG_MAX);
 	else
-		invalidate_inode_pages2(vnode->netfs.inode.i_mapping);
+		filemap_invalidate_inode(&vnode->netfs.inode, false, 0, LLONG_MAX);
 }
 
 /*
diff --git a/fs/afs/write.c b/fs/afs/write.c
index 74402d95a8..e959640694 100644
--- a/fs/afs/write.c
+++ b/fs/afs/write.c
@@ -29,43 +29,39 @@ static void afs_pages_written_back(struct afs_vnode *vnode, loff_t start, unsign
 
 /*
  * Find a key to use for the writeback.  We cached the keys used to author the
- * writes on the vnode.  *_wbk will contain the last writeback key used or NULL
- * and we need to start from there if it's set.
+ * writes on the vnode.  wreq->netfs_priv2 will contain the last writeback key
+ * record used or NULL and we need to start from there if it's set.
+ * wreq->netfs_priv will be set to the key itself or NULL.
  */
-static int afs_get_writeback_key(struct afs_vnode *vnode,
-				 struct afs_wb_key **_wbk)
+static void afs_get_writeback_key(struct netfs_io_request *wreq)
 {
-	struct afs_wb_key *wbk = NULL;
-	struct list_head *p;
-	int ret = -ENOKEY, ret2;
+	struct afs_wb_key *wbk, *old = wreq->netfs_priv2;
+	struct afs_vnode *vnode = AFS_FS_I(wreq->inode);
+
+	key_put(wreq->netfs_priv);
+	wreq->netfs_priv = NULL;
+	wreq->netfs_priv2 = NULL;
 
 	spin_lock(&vnode->wb_lock);
-	if (*_wbk)
-		p = (*_wbk)->vnode_link.next;
+	if (old)
+		wbk = list_next_entry(old, vnode_link);
 	else
-		p = vnode->wb_keys.next;
+		wbk = list_first_entry(&vnode->wb_keys, struct afs_wb_key, vnode_link);
 
-	while (p != &vnode->wb_keys) {
-		wbk = list_entry(p, struct afs_wb_key, vnode_link);
+	list_for_each_entry_from(wbk, &vnode->wb_keys, vnode_link) {
 		_debug("wbk %u", key_serial(wbk->key));
-		ret2 = key_validate(wbk->key);
-		if (ret2 == 0) {
+		if (key_validate(wbk->key) == 0) {
 			refcount_inc(&wbk->usage);
+			wreq->netfs_priv = key_get(wbk->key);
+			wreq->netfs_priv2 = wbk;
 			_debug("USE WB KEY %u", key_serial(wbk->key));
 			break;
 		}
-
-		wbk = NULL;
-		if (ret == -ENOKEY)
-			ret = ret2;
-		p = p->next;
 	}
 
 	spin_unlock(&vnode->wb_lock);
-	if (*_wbk)
-		afs_put_wb_key(*_wbk);
-	*_wbk = wbk;
-	return 0;
+
+	afs_put_wb_key(old);
 }
 
 static void afs_store_data_success(struct afs_operation *op)
@@ -75,8 +71,7 @@ static void afs_store_data_success(struct afs_operation *op)
 	op->ctime = op->file[0].scb.status.mtime_client;
 	afs_vnode_commit_status(op, &op->file[0]);
 	if (!afs_op_error(op)) {
-		if (!op->store.laundering)
-			afs_pages_written_back(vnode, op->store.pos, op->store.size);
+		afs_pages_written_back(vnode, op->store.pos, op->store.size);
 		afs_stat_v(vnode, n_stores);
 		atomic_long_add(op->store.size, &afs_v2net(vnode)->n_store_bytes);
 	}
@@ -89,113 +84,125 @@ static const struct afs_operation_ops afs_store_data_operation = {
 };
 
 /*
- * write to a file
+ * Prepare a subrequest to write to the server.  This sets the max_len
+ * parameter.
+ */
+void afs_prepare_write(struct netfs_io_subrequest *subreq)
+{
+	//if (test_bit(NETFS_SREQ_RETRYING, &subreq->flags))
+	//	subreq->max_len = 512 * 1024;
+	//else
+	subreq->max_len = 256 * 1024 * 1024;
+}
+
+/*
+ * Issue a subrequest to write to the server.
  */
-static int afs_store_data(struct afs_vnode *vnode, struct iov_iter *iter, loff_t pos,
-			  bool laundering)
+static void afs_issue_write_worker(struct work_struct *work)
 {
+	struct netfs_io_subrequest *subreq = container_of(work, struct netfs_io_subrequest, work);
+	struct netfs_io_request *wreq = subreq->rreq;
 	struct afs_operation *op;
-	struct afs_wb_key *wbk = NULL;
-	loff_t size = iov_iter_count(iter);
+	struct afs_vnode *vnode = AFS_FS_I(wreq->inode);
+	unsigned long long pos = subreq->start + subreq->transferred;
+	size_t len = subreq->len - subreq->transferred;
 	int ret = -ENOKEY;
 
-	_enter("%s{%llx:%llu.%u},%llx,%llx",
+	_enter("R=%x[%x],%s{%llx:%llu.%u},%llx,%zx",
+	       wreq->debug_id, subreq->debug_index,
 	       vnode->volume->name,
 	       vnode->fid.vid,
 	       vnode->fid.vnode,
 	       vnode->fid.unique,
-	       size, pos);
+	       pos, len);
 
-	ret = afs_get_writeback_key(vnode, &wbk);
-	if (ret) {
-		_leave(" = %d [no keys]", ret);
-		return ret;
-	}
+#if 0 // Error injection
+	if (subreq->debug_index == 3)
+		return netfs_write_subrequest_terminated(subreq, -ENOANO, false);
 
-	op = afs_alloc_operation(wbk->key, vnode->volume);
-	if (IS_ERR(op)) {
-		afs_put_wb_key(wbk);
-		return -ENOMEM;
+	if (!test_bit(NETFS_SREQ_RETRYING, &subreq->flags)) {
+		set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags);
+		return netfs_write_subrequest_terminated(subreq, -EAGAIN, false);
 	}
+#endif
+
+	op = afs_alloc_operation(wreq->netfs_priv, vnode->volume);
+	if (IS_ERR(op))
+		return netfs_write_subrequest_terminated(subreq, -EAGAIN, false);
 
 	afs_op_set_vnode(op, 0, vnode);
-	op->file[0].dv_delta = 1;
+	op->file[0].dv_delta	= 1;
 	op->file[0].modification = true;
-	op->store.pos = pos;
-	op->store.size = size;
-	op->store.laundering = laundering;
-	op->flags |= AFS_OPERATION_UNINTR;
-	op->ops = &afs_store_data_operation;
+	op->store.pos		= pos;
+	op->store.size		= len;
+	op->flags		|= AFS_OPERATION_UNINTR;
+	op->ops			= &afs_store_data_operation;
 
-try_next_key:
 	afs_begin_vnode_operation(op);
 
-	op->store.write_iter = iter;
-	op->store.i_size = max(pos + size, vnode->netfs.remote_i_size);
-	op->mtime = inode_get_mtime(&vnode->netfs.inode);
+	op->store.write_iter	= &subreq->io_iter;
+	op->store.i_size	= umax(pos + len, vnode->netfs.remote_i_size);
+	op->mtime		= inode_get_mtime(&vnode->netfs.inode);
 
 	afs_wait_for_operation(op);
-
-	switch (afs_op_error(op)) {
+	ret = afs_put_operation(op);
+	switch (ret) {
 	case -EACCES:
 	case -EPERM:
 	case -ENOKEY:
 	case -EKEYEXPIRED:
 	case -EKEYREJECTED:
 	case -EKEYREVOKED:
-		_debug("next");
-
-		ret = afs_get_writeback_key(vnode, &wbk);
-		if (ret == 0) {
-			key_put(op->key);
-			op->key = key_get(wbk->key);
-			goto try_next_key;
-		}
+		/* If there are more keys we can try, use the retry algorithm
+		 * to rotate the keys.
+		 */
+		if (wreq->netfs_priv2)
+			set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags);
 		break;
 	}
 
-	afs_put_wb_key(wbk);
-	_leave(" = %d", afs_op_error(op));
-	return afs_put_operation(op);
+	netfs_write_subrequest_terminated(subreq, ret < 0 ? ret : subreq->len, false);
 }
 
-static void afs_upload_to_server(struct netfs_io_subrequest *subreq)
+void afs_issue_write(struct netfs_io_subrequest *subreq)
 {
-	struct afs_vnode *vnode = AFS_FS_I(subreq->rreq->inode);
-	ssize_t ret;
-
-	_enter("%x[%x],%zx",
-	       subreq->rreq->debug_id, subreq->debug_index, subreq->io_iter.count);
-
-	trace_netfs_sreq(subreq, netfs_sreq_trace_submit);
-	ret = afs_store_data(vnode, &subreq->io_iter, subreq->start,
-			     subreq->rreq->origin == NETFS_LAUNDER_WRITE);
-	netfs_write_subrequest_terminated(subreq, ret < 0 ? ret : subreq->len,
-					  false);
+	subreq->work.func = afs_issue_write_worker;
+	if (!queue_work(system_unbound_wq, &subreq->work))
+		WARN_ON_ONCE(1);
 }
 
-static void afs_upload_to_server_worker(struct work_struct *work)
+/*
+ * Writeback calls this when it finds a folio that needs uploading.  This isn't
+ * called if writeback only has copy-to-cache to deal with.
+ */
+void afs_begin_writeback(struct netfs_io_request *wreq)
 {
-	struct netfs_io_subrequest *subreq =
-		container_of(work, struct netfs_io_subrequest, work);
-
-	afs_upload_to_server(subreq);
+	afs_get_writeback_key(wreq);
+	wreq->io_streams[0].avail = true;
 }
 
 /*
- * Set up write requests for a writeback slice.  We need to add a write request
- * for each write we want to make.
+ * Prepare to retry the writes in request.  Use this to try rotating the
+ * available writeback keys.
  */
-void afs_create_write_requests(struct netfs_io_request *wreq, loff_t start, size_t len)
+void afs_retry_request(struct netfs_io_request *wreq, struct netfs_io_stream *stream)
 {
-	struct netfs_io_subrequest *subreq;
-
-	_enter("%x,%llx-%llx", wreq->debug_id, start, start + len);
+	struct netfs_io_subrequest *subreq =
+		list_first_entry(&stream->subrequests,
+				 struct netfs_io_subrequest, rreq_link);
 
-	subreq = netfs_create_write_request(wreq, NETFS_UPLOAD_TO_SERVER,
-					    start, len, afs_upload_to_server_worker);
-	if (subreq)
-		netfs_queue_write_request(subreq);
+	switch (subreq->error) {
+	case -EACCES:
+	case -EPERM:
+	case -ENOKEY:
+	case -EKEYEXPIRED:
+	case -EKEYREJECTED:
+	case -EKEYREVOKED:
+		afs_get_writeback_key(wreq);
+		if (!wreq->netfs_priv)
+			stream->failed = true;
+		break;
+	}
 }
 
 /*
diff --git a/fs/aio.c b/fs/aio.c
index 0f4f531c97..57c9f7c077 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -122,7 +122,7 @@ struct kioctx {
 	unsigned long		mmap_base;
 	unsigned long		mmap_size;
 
-	struct page		**ring_pages;
+	struct folio		**ring_folios;
 	long			nr_pages;
 
 	struct rcu_work		free_rwork;	/* see free_ioctx() */
@@ -160,7 +160,7 @@ struct kioctx {
 		spinlock_t	completion_lock;
 	} ____cacheline_aligned_in_smp;
 
-	struct page		*internal_pages[AIO_RING_PAGES];
+	struct folio		*internal_folios[AIO_RING_PAGES];
 	struct file		*aio_ring_file;
 
 	unsigned		id;
@@ -334,19 +334,20 @@ static void aio_free_ring(struct kioctx *ctx)
 	put_aio_ring_file(ctx);
 
 	for (i = 0; i < ctx->nr_pages; i++) {
-		struct page *page;
-		pr_debug("pid(%d) [%d] page->count=%d\n", current->pid, i,
-				page_count(ctx->ring_pages[i]));
-		page = ctx->ring_pages[i];
-		if (!page)
+		struct folio *folio = ctx->ring_folios[i];
+
+		if (!folio)
 			continue;
-		ctx->ring_pages[i] = NULL;
-		put_page(page);
+
+		pr_debug("pid(%d) [%d] folio->count=%d\n", current->pid, i,
+			 folio_ref_count(folio));
+		ctx->ring_folios[i] = NULL;
+		folio_put(folio);
 	}
 
-	if (ctx->ring_pages && ctx->ring_pages != ctx->internal_pages) {
-		kfree(ctx->ring_pages);
-		ctx->ring_pages = NULL;
+	if (ctx->ring_folios && ctx->ring_folios != ctx->internal_folios) {
+		kfree(ctx->ring_folios);
+		ctx->ring_folios = NULL;
 	}
 }
 
@@ -441,7 +442,7 @@ static int aio_migrate_folio(struct address_space *mapping, struct folio *dst,
 	idx = src->index;
 	if (idx < (pgoff_t)ctx->nr_pages) {
 		/* Make sure the old folio hasn't already been changed */
-		if (ctx->ring_pages[idx] != &src->page)
+		if (ctx->ring_folios[idx] != src)
 			rc = -EAGAIN;
 	} else
 		rc = -EINVAL;
@@ -465,8 +466,8 @@ static int aio_migrate_folio(struct address_space *mapping, struct folio *dst,
 	 */
 	spin_lock_irqsave(&ctx->completion_lock, flags);
 	folio_migrate_copy(dst, src);
-	BUG_ON(ctx->ring_pages[idx] != &src->page);
-	ctx->ring_pages[idx] = &dst->page;
+	BUG_ON(ctx->ring_folios[idx] != src);
+	ctx->ring_folios[idx] = dst;
 	spin_unlock_irqrestore(&ctx->completion_lock, flags);
 
 	/* The old folio is no longer accessible. */
@@ -516,28 +517,30 @@ static int aio_setup_ring(struct kioctx *ctx, unsigned int nr_events)
 	nr_events = (PAGE_SIZE * nr_pages - sizeof(struct aio_ring))
 			/ sizeof(struct io_event);
 
-	ctx->ring_pages = ctx->internal_pages;
+	ctx->ring_folios = ctx->internal_folios;
 	if (nr_pages > AIO_RING_PAGES) {
-		ctx->ring_pages = kcalloc(nr_pages, sizeof(struct page *),
-					  GFP_KERNEL);
-		if (!ctx->ring_pages) {
+		ctx->ring_folios = kcalloc(nr_pages, sizeof(struct folio *),
+					   GFP_KERNEL);
+		if (!ctx->ring_folios) {
 			put_aio_ring_file(ctx);
 			return -ENOMEM;
 		}
 	}
 
 	for (i = 0; i < nr_pages; i++) {
-		struct page *page;
-		page = find_or_create_page(file->f_mapping,
-					   i, GFP_USER | __GFP_ZERO);
-		if (!page)
+		struct folio *folio;
+
+		folio = __filemap_get_folio(file->f_mapping, i,
+					    FGP_LOCK | FGP_ACCESSED | FGP_CREAT,
+					    GFP_USER | __GFP_ZERO);
+		if (IS_ERR(folio))
 			break;
-		pr_debug("pid(%d) page[%d]->count=%d\n",
-			 current->pid, i, page_count(page));
-		SetPageUptodate(page);
-		unlock_page(page);
 
-		ctx->ring_pages[i] = page;
+		pr_debug("pid(%d) [%d] folio->count=%d\n", current->pid, i,
+			 folio_ref_count(folio));
+		folio_end_read(folio, true);
+
+		ctx->ring_folios[i] = folio;
 	}
 	ctx->nr_pages = i;
 
@@ -570,7 +573,7 @@ static int aio_setup_ring(struct kioctx *ctx, unsigned int nr_events)
 	ctx->user_id = ctx->mmap_base;
 	ctx->nr_events = nr_events; /* trusted copy */
 
-	ring = page_address(ctx->ring_pages[0]);
+	ring = folio_address(ctx->ring_folios[0]);
 	ring->nr = nr_events;	/* user copy */
 	ring->id = ~0U;
 	ring->head = ring->tail = 0;
@@ -578,7 +581,7 @@ static int aio_setup_ring(struct kioctx *ctx, unsigned int nr_events)
 	ring->compat_features = AIO_RING_COMPAT_FEATURES;
 	ring->incompat_features = AIO_RING_INCOMPAT_FEATURES;
 	ring->header_length = sizeof(struct aio_ring);
-	flush_dcache_page(ctx->ring_pages[0]);
+	flush_dcache_folio(ctx->ring_folios[0]);
 
 	return 0;
 }
@@ -689,9 +692,9 @@ static int ioctx_add_table(struct kioctx *ctx, struct mm_struct *mm)
 
 					/* While kioctx setup is in progress,
 					 * we are protected from page migration
-					 * changes ring_pages by ->ring_lock.
+					 * changes ring_folios by ->ring_lock.
 					 */
-					ring = page_address(ctx->ring_pages[0]);
+					ring = folio_address(ctx->ring_folios[0]);
 					ring->id = ctx->id;
 					return 0;
 				}
@@ -1033,7 +1036,7 @@ static void user_refill_reqs_available(struct kioctx *ctx)
 		 * against ctx->completed_events below will make sure we do the
 		 * safe/right thing.
 		 */
-		ring = page_address(ctx->ring_pages[0]);
+		ring = folio_address(ctx->ring_folios[0]);
 		head = ring->head;
 
 		refill_reqs_available(ctx, head, ctx->tail);
@@ -1145,12 +1148,12 @@ static void aio_complete(struct aio_kiocb *iocb)
 	if (++tail >= ctx->nr_events)
 		tail = 0;
 
-	ev_page = page_address(ctx->ring_pages[pos / AIO_EVENTS_PER_PAGE]);
+	ev_page = folio_address(ctx->ring_folios[pos / AIO_EVENTS_PER_PAGE]);
 	event = ev_page + pos % AIO_EVENTS_PER_PAGE;
 
 	*event = iocb->ki_res;
 
-	flush_dcache_page(ctx->ring_pages[pos / AIO_EVENTS_PER_PAGE]);
+	flush_dcache_folio(ctx->ring_folios[pos / AIO_EVENTS_PER_PAGE]);
 
 	pr_debug("%p[%u]: %p: %p %Lx %Lx %Lx\n", ctx, tail, iocb,
 		 (void __user *)(unsigned long)iocb->ki_res.obj,
@@ -1163,10 +1166,10 @@ static void aio_complete(struct aio_kiocb *iocb)
 
 	ctx->tail = tail;
 
-	ring = page_address(ctx->ring_pages[0]);
+	ring = folio_address(ctx->ring_folios[0]);
 	head = ring->head;
 	ring->tail = tail;
-	flush_dcache_page(ctx->ring_pages[0]);
+	flush_dcache_folio(ctx->ring_folios[0]);
 
 	ctx->completed_events++;
 	if (ctx->completed_events > 1)
@@ -1238,8 +1241,8 @@ static long aio_read_events_ring(struct kioctx *ctx,
 	sched_annotate_sleep();
 	mutex_lock(&ctx->ring_lock);
 
-	/* Access to ->ring_pages here is protected by ctx->ring_lock. */
-	ring = page_address(ctx->ring_pages[0]);
+	/* Access to ->ring_folios here is protected by ctx->ring_lock. */
+	ring = folio_address(ctx->ring_folios[0]);
 	head = ring->head;
 	tail = ring->tail;
 
@@ -1260,20 +1263,20 @@ static long aio_read_events_ring(struct kioctx *ctx,
 	while (ret < nr) {
 		long avail;
 		struct io_event *ev;
-		struct page *page;
+		struct folio *folio;
 
 		avail = (head <= tail ?  tail : ctx->nr_events) - head;
 		if (head == tail)
 			break;
 
 		pos = head + AIO_EVENTS_OFFSET;
-		page = ctx->ring_pages[pos / AIO_EVENTS_PER_PAGE];
+		folio = ctx->ring_folios[pos / AIO_EVENTS_PER_PAGE];
 		pos %= AIO_EVENTS_PER_PAGE;
 
 		avail = min(avail, nr - ret);
 		avail = min_t(long, avail, AIO_EVENTS_PER_PAGE - pos);
 
-		ev = page_address(page);
+		ev = folio_address(folio);
 		copy_ret = copy_to_user(event + ret, ev + pos,
 					sizeof(*ev) * avail);
 
@@ -1287,9 +1290,9 @@ static long aio_read_events_ring(struct kioctx *ctx,
 		head %= ctx->nr_events;
 	}
 
-	ring = page_address(ctx->ring_pages[0]);
+	ring = folio_address(ctx->ring_folios[0]);
 	ring->head = head;
-	flush_dcache_page(ctx->ring_pages[0]);
+	flush_dcache_folio(ctx->ring_folios[0]);
 
 	pr_debug("%li  h%u t%u\n", ret, head, tail);
 out:
@@ -1605,7 +1608,7 @@ static int aio_read(struct kiocb *req, const struct iocb *iocb,
 		return ret;
 	ret = rw_verify_area(READ, file, &req->ki_pos, iov_iter_count(&iter));
 	if (!ret)
-		aio_rw_done(req, call_read_iter(file, req, &iter));
+		aio_rw_done(req, file->f_op->read_iter(req, &iter));
 	kfree(iovec);
 	return ret;
 }
@@ -1636,7 +1639,7 @@ static int aio_write(struct kiocb *req, const struct iocb *iocb,
 		if (S_ISREG(file_inode(file)->i_mode))
 			kiocb_start_write(req);
 		req->ki_flags |= IOCB_WRITE;
-		aio_rw_done(req, call_write_iter(file, req, &iter));
+		aio_rw_done(req, file->f_op->write_iter(req, &iter));
 	}
 	kfree(iovec);
 	return ret;
diff --git a/fs/anon_inodes.c b/fs/anon_inodes.c
index 0496cb5b6e..42bd1cb7c9 100644
--- a/fs/anon_inodes.c
+++ b/fs/anon_inodes.c
@@ -149,6 +149,38 @@ struct file *anon_inode_getfile(const char *name,
 EXPORT_SYMBOL_GPL(anon_inode_getfile);
 
 /**
+ * anon_inode_getfile_fmode - creates a new file instance by hooking it up to an
+ *                      anonymous inode, and a dentry that describe the "class"
+ *                      of the file
+ *
+ * @name:    [in]    name of the "class" of the new file
+ * @fops:    [in]    file operations for the new file
+ * @priv:    [in]    private data for the new file (will be file's private_data)
+ * @flags:   [in]    flags
+ * @f_mode:  [in]    fmode
+ *
+ * Creates a new file by hooking it on a single inode. This is useful for files
+ * that do not need to have a full-fledged inode in order to operate correctly.
+ * All the files created with anon_inode_getfile() will share a single inode,
+ * hence saving memory and avoiding code duplication for the file/inode/dentry
+ * setup. Allows setting the fmode. Returns the newly created file* or an error
+ * pointer.
+ */
+struct file *anon_inode_getfile_fmode(const char *name,
+				const struct file_operations *fops,
+				void *priv, int flags, fmode_t f_mode)
+{
+	struct file *file;
+
+	file = __anon_inode_getfile(name, fops, priv, flags, NULL, false);
+	if (!IS_ERR(file))
+		file->f_mode |= f_mode;
+
+	return file;
+}
+EXPORT_SYMBOL_GPL(anon_inode_getfile_fmode);
+
+/**
  * anon_inode_create_getfile - Like anon_inode_getfile(), but creates a new
  *                             !S_PRIVATE anon inode rather than reuse the
  *                             singleton anon inode and calls the
@@ -271,6 +303,7 @@ int anon_inode_create_getfd(const char *name, const struct file_operations *fops
 	return __anon_inode_getfd(name, fops, priv, flags, context_inode, true);
 }
 
+
 static int __init anon_inode_init(void)
 {
 	anon_inode_mnt = kern_mount(&anon_inode_fs_type);
diff --git a/fs/attr.c b/fs/attr.c
index 5a13f0c849..960a310581 100644
--- a/fs/attr.c
+++ b/fs/attr.c
@@ -16,8 +16,6 @@
 #include <linux/fcntl.h>
 #include <linux/filelock.h>
 #include <linux/security.h>
-#include <linux/evm.h>
-#include <linux/ima.h>
 
 #include "internal.h"
 
@@ -352,7 +350,7 @@ int may_setattr(struct mnt_idmap *idmap, struct inode *inode,
 EXPORT_SYMBOL(may_setattr);
 
 /**
- * notify_change - modify attributes of a filesytem object
+ * notify_change - modify attributes of a filesystem object
  * @idmap:	idmap of the mount the inode was found from
  * @dentry:	object affected
  * @attr:	new attributes
@@ -502,8 +500,7 @@ int notify_change(struct mnt_idmap *idmap, struct dentry *dentry,
 
 	if (!error) {
 		fsnotify_change(dentry, ia_valid);
-		ima_inode_post_setattr(idmap, dentry);
-		evm_inode_post_setattr(dentry, ia_valid);
+		security_inode_post_setattr(idmap, dentry, ia_valid);
 	}
 
 	return error;
diff --git a/fs/backing-file.c b/fs/backing-file.c
index a681f38d84..afb557446c 100644
--- a/fs/backing-file.c
+++ b/fs/backing-file.c
@@ -52,6 +52,29 @@ struct file *backing_file_open(const struct path *user_path, int flags,
 }
 EXPORT_SYMBOL_GPL(backing_file_open);
 
+struct file *backing_tmpfile_open(const struct path *user_path, int flags,
+				  const struct path *real_parentpath,
+				  umode_t mode, const struct cred *cred)
+{
+	struct mnt_idmap *real_idmap = mnt_idmap(real_parentpath->mnt);
+	struct file *f;
+	int error;
+
+	f = alloc_empty_backing_file(flags, cred);
+	if (IS_ERR(f))
+		return f;
+
+	path_get(user_path);
+	*backing_file_user_path(f) = *user_path;
+	error = vfs_tmpfile(real_idmap, real_parentpath, f, mode);
+	if (error) {
+		fput(f);
+		f = ERR_PTR(error);
+	}
+	return f;
+}
+EXPORT_SYMBOL(backing_tmpfile_open);
+
 struct backing_aio {
 	struct kiocb iocb;
 	refcount_t ref;
@@ -325,9 +348,7 @@ EXPORT_SYMBOL_GPL(backing_file_mmap);
 
 static int __init backing_aio_init(void)
 {
-	backing_aio_cachep = kmem_cache_create("backing_aio",
-					       sizeof(struct backing_aio),
-					       0, SLAB_HWCACHE_ALIGN, NULL);
+	backing_aio_cachep = KMEM_CACHE(backing_aio, SLAB_HWCACHE_ALIGN);
 	if (!backing_aio_cachep)
 		return -ENOMEM;
 
diff --git a/fs/bcachefs/Makefile b/fs/bcachefs/Makefile
index 1a05cecda7..66ca0bbee6 100644
--- a/fs/bcachefs/Makefile
+++ b/fs/bcachefs/Makefile
@@ -17,6 +17,7 @@ bcachefs-y		:=	\
 	btree_journal_iter.o	\
 	btree_key_cache.o	\
 	btree_locking.o		\
+	btree_node_scan.o	\
 	btree_trans_commit.o	\
 	btree_update.o		\
 	btree_update_interior.o	\
@@ -37,6 +38,7 @@ bcachefs-y		:=	\
 	error.o			\
 	extents.o		\
 	extent_update.o		\
+	eytzinger.o		\
 	fs.o			\
 	fs-common.o		\
 	fs-ioctl.o		\
@@ -67,6 +69,7 @@ bcachefs-y		:=	\
 	quota.o			\
 	rebalance.o		\
 	recovery.o		\
+	recovery_passes.o	\
 	reflink.o		\
 	replicas.o		\
 	sb-clean.o		\
@@ -82,6 +85,7 @@ bcachefs-y		:=	\
 	super-io.o		\
 	sysfs.o			\
 	tests.o			\
+	time_stats.o		\
 	thread_with_file.o	\
 	trace.o			\
 	two_state_shared_lock.o	\
@@ -90,3 +94,6 @@ bcachefs-y		:=	\
 	xattr.o
 
 obj-$(CONFIG_MEAN_AND_VARIANCE_UNIT_TEST)   += mean_and_variance_test.o
+
+# Silence "note: xyz changed in GCC X.X" messages
+subdir-ccflags-y += $(call cc-disable-warning, psabi)
diff --git a/fs/bcachefs/acl.c b/fs/bcachefs/acl.c
index 3640f417cc..250d6c6d3a 100644
--- a/fs/bcachefs/acl.c
+++ b/fs/bcachefs/acl.c
@@ -281,37 +281,29 @@ struct posix_acl *bch2_get_acl(struct mnt_idmap *idmap,
 	struct xattr_search_key search = X_SEARCH(acl_to_xattr_type(type), "", 0);
 	struct btree_trans *trans = bch2_trans_get(c);
 	struct btree_iter iter = { NULL };
-	struct bkey_s_c_xattr xattr;
 	struct posix_acl *acl = NULL;
-	struct bkey_s_c k;
-	int ret;
 retry:
 	bch2_trans_begin(trans);
 
-	ret = bch2_hash_lookup(trans, &iter, bch2_xattr_hash_desc,
-			&hash, inode_inum(inode), &search, 0);
-	if (ret) {
-		if (!bch2_err_matches(ret, ENOENT))
-			acl = ERR_PTR(ret);
-		goto out;
-	}
-
-	k = bch2_btree_iter_peek_slot(&iter);
-	ret = bkey_err(k);
-	if (ret) {
-		acl = ERR_PTR(ret);
-		goto out;
-	}
+	struct bkey_s_c k = bch2_hash_lookup(trans, &iter, bch2_xattr_hash_desc,
+					     &hash, inode_inum(inode), &search, 0);
+	int ret = bkey_err(k);
+	if (ret)
+		goto err;
 
-	xattr = bkey_s_c_to_xattr(k);
+	struct bkey_s_c_xattr xattr = bkey_s_c_to_xattr(k);
 	acl = bch2_acl_from_disk(trans, xattr_val(xattr.v),
-			le16_to_cpu(xattr.v->x_val_len));
+				 le16_to_cpu(xattr.v->x_val_len));
+	ret = PTR_ERR_OR_ZERO(acl);
+err:
+	if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
+		goto retry;
+
+	if (ret)
+		acl = !bch2_err_matches(ret, ENOENT) ? ERR_PTR(ret) : NULL;
 
-	if (!IS_ERR(acl))
+	if (!IS_ERR_OR_NULL(acl))
 		set_cached_acl(&inode->v, type, acl);
-out:
-	if (bch2_err_matches(PTR_ERR_OR_ZERO(acl), BCH_ERR_transaction_restart))
-		goto retry;
 
 	bch2_trans_iter_exit(trans, &iter);
 	bch2_trans_put(trans);
@@ -368,7 +360,7 @@ retry:
 
 	ret   = bch2_subvol_is_ro_trans(trans, inode->ei_subvol) ?:
 		bch2_inode_peek(trans, &inode_iter, &inode_u, inode_inum(inode),
-			      BTREE_ITER_INTENT);
+			      BTREE_ITER_intent);
 	if (ret)
 		goto btree_err;
 
@@ -416,39 +408,30 @@ int bch2_acl_chmod(struct btree_trans *trans, subvol_inum inum,
 	struct bch_hash_info hash_info = bch2_hash_info_init(trans->c, inode);
 	struct xattr_search_key search = X_SEARCH(KEY_TYPE_XATTR_INDEX_POSIX_ACL_ACCESS, "", 0);
 	struct btree_iter iter;
-	struct bkey_s_c_xattr xattr;
-	struct bkey_i_xattr *new;
 	struct posix_acl *acl = NULL;
-	struct bkey_s_c k;
-	int ret;
 
-	ret = bch2_hash_lookup(trans, &iter, bch2_xattr_hash_desc,
-			       &hash_info, inum, &search, BTREE_ITER_INTENT);
+	struct bkey_s_c k = bch2_hash_lookup(trans, &iter, bch2_xattr_hash_desc,
+			       &hash_info, inum, &search, BTREE_ITER_intent);
+	int ret = bkey_err(k);
 	if (ret)
 		return bch2_err_matches(ret, ENOENT) ? 0 : ret;
 
-	k = bch2_btree_iter_peek_slot(&iter);
-	ret = bkey_err(k);
-	if (ret)
-		goto err;
-	xattr = bkey_s_c_to_xattr(k);
+	struct bkey_s_c_xattr xattr = bkey_s_c_to_xattr(k);
 
 	acl = bch2_acl_from_disk(trans, xattr_val(xattr.v),
 			le16_to_cpu(xattr.v->x_val_len));
 	ret = PTR_ERR_OR_ZERO(acl);
-	if (IS_ERR_OR_NULL(acl))
+	if (ret)
 		goto err;
 
-	ret = allocate_dropping_locks_errcode(trans,
-				__posix_acl_chmod(&acl, _gfp, mode));
+	ret = allocate_dropping_locks_errcode(trans, __posix_acl_chmod(&acl, _gfp, mode));
 	if (ret)
 		goto err;
 
-	new = bch2_acl_to_xattr(trans, acl, ACL_TYPE_ACCESS);
-	if (IS_ERR(new)) {
-		ret = PTR_ERR(new);
+	struct bkey_i_xattr *new = bch2_acl_to_xattr(trans, acl, ACL_TYPE_ACCESS);
+	ret = PTR_ERR_OR_ZERO(new);
+	if (ret)
 		goto err;
-	}
 
 	new->k.p = iter.pos;
 	ret = bch2_trans_update(trans, &iter, &new->k_i, 0);
diff --git a/fs/bcachefs/alloc_background.c b/fs/bcachefs/alloc_background.c
index fd3e175d83..658f11aebd 100644
--- a/fs/bcachefs/alloc_background.c
+++ b/fs/bcachefs/alloc_background.c
@@ -3,6 +3,7 @@
 #include "alloc_background.h"
 #include "alloc_foreground.h"
 #include "backpointers.h"
+#include "bkey_buf.h"
 #include "btree_cache.h"
 #include "btree_io.h"
 #include "btree_key_cache.h"
@@ -29,6 +30,8 @@
 #include <linux/sched/task.h>
 #include <linux/sort.h>
 
+static void bch2_discard_one_bucket_fast(struct bch_dev *, u64);
+
 /* Persistent alloc info: */
 
 static const unsigned BCH_ALLOC_V1_FIELD_BYTES[] = {
@@ -193,7 +196,7 @@ static unsigned bch_alloc_v1_val_u64s(const struct bch_alloc *a)
 }
 
 int bch2_alloc_v1_invalid(struct bch_fs *c, struct bkey_s_c k,
-			  enum bkey_invalid_flags flags,
+			  enum bch_validate_flags flags,
 			  struct printbuf *err)
 {
 	struct bkey_s_c_alloc a = bkey_s_c_to_alloc(k);
@@ -209,7 +212,7 @@ fsck_err:
 }
 
 int bch2_alloc_v2_invalid(struct bch_fs *c, struct bkey_s_c k,
-			  enum bkey_invalid_flags flags,
+			  enum bch_validate_flags flags,
 			  struct printbuf *err)
 {
 	struct bkey_alloc_unpacked u;
@@ -223,7 +226,7 @@ fsck_err:
 }
 
 int bch2_alloc_v3_invalid(struct bch_fs *c, struct bkey_s_c k,
-			  enum bkey_invalid_flags flags,
+			  enum bch_validate_flags flags,
 			  struct printbuf *err)
 {
 	struct bkey_alloc_unpacked u;
@@ -237,15 +240,15 @@ fsck_err:
 }
 
 int bch2_alloc_v4_invalid(struct bch_fs *c, struct bkey_s_c k,
-			  enum bkey_invalid_flags flags, struct printbuf *err)
+			  enum bch_validate_flags flags, struct printbuf *err)
 {
 	struct bkey_s_c_alloc_v4 a = bkey_s_c_to_alloc_v4(k);
 	int ret = 0;
 
-	bkey_fsck_err_on(alloc_v4_u64s(a.v) > bkey_val_u64s(k.k), c, err,
+	bkey_fsck_err_on(alloc_v4_u64s_noerror(a.v) > bkey_val_u64s(k.k), c, err,
 			 alloc_v4_val_size_bad,
 			 "bad val size (%u > %zu)",
-			 alloc_v4_u64s(a.v), bkey_val_u64s(k.k));
+			 alloc_v4_u64s_noerror(a.v), bkey_val_u64s(k.k));
 
 	bkey_fsck_err_on(!BCH_ALLOC_V4_BACKPOINTERS_START(a.v) &&
 			 BCH_ALLOC_V4_NR_BACKPOINTERS(a.v), c, err,
@@ -257,11 +260,19 @@ int bch2_alloc_v4_invalid(struct bch_fs *c, struct bkey_s_c k,
 			 "invalid data type (got %u should be %u)",
 			 a.v->data_type, alloc_data_type(*a.v, a.v->data_type));
 
+	for (unsigned i = 0; i < 2; i++)
+		bkey_fsck_err_on(a.v->io_time[i] > LRU_TIME_MAX,
+				 c, err,
+				 alloc_key_io_time_bad,
+				 "invalid io_time[%s]: %llu, max %llu",
+				 i == READ ? "read" : "write",
+				 a.v->io_time[i], LRU_TIME_MAX);
+
 	switch (a.v->data_type) {
 	case BCH_DATA_free:
 	case BCH_DATA_need_gc_gens:
 	case BCH_DATA_need_discard:
-		bkey_fsck_err_on(bch2_bucket_sectors(*a.v) || a.v->stripe,
+		bkey_fsck_err_on(bch2_bucket_sectors_total(*a.v) || a.v->stripe,
 				 c, err, alloc_key_empty_but_have_data,
 				 "empty data type free but have data");
 		break;
@@ -328,27 +339,17 @@ void bch2_alloc_to_text(struct printbuf *out, struct bch_fs *c, struct bkey_s_c
 	prt_printf(out, "gen %u oldest_gen %u data_type ", a->gen, a->oldest_gen);
 	bch2_prt_data_type(out, a->data_type);
 	prt_newline(out);
-	prt_printf(out, "journal_seq       %llu",	a->journal_seq);
-	prt_newline(out);
-	prt_printf(out, "need_discard      %llu",	BCH_ALLOC_V4_NEED_DISCARD(a));
-	prt_newline(out);
-	prt_printf(out, "need_inc_gen      %llu",	BCH_ALLOC_V4_NEED_INC_GEN(a));
-	prt_newline(out);
-	prt_printf(out, "dirty_sectors     %u",	a->dirty_sectors);
-	prt_newline(out);
-	prt_printf(out, "cached_sectors    %u",	a->cached_sectors);
-	prt_newline(out);
-	prt_printf(out, "stripe            %u",	a->stripe);
-	prt_newline(out);
-	prt_printf(out, "stripe_redundancy %u",	a->stripe_redundancy);
-	prt_newline(out);
-	prt_printf(out, "io_time[READ]     %llu",	a->io_time[READ]);
-	prt_newline(out);
-	prt_printf(out, "io_time[WRITE]    %llu",	a->io_time[WRITE]);
-	prt_newline(out);
-	prt_printf(out, "fragmentation     %llu",	a->fragmentation_lru);
-	prt_newline(out);
-	prt_printf(out, "bp_start          %llu", BCH_ALLOC_V4_BACKPOINTERS_START(a));
+	prt_printf(out, "journal_seq       %llu\n",	a->journal_seq);
+	prt_printf(out, "need_discard      %llu\n",	BCH_ALLOC_V4_NEED_DISCARD(a));
+	prt_printf(out, "need_inc_gen      %llu\n",	BCH_ALLOC_V4_NEED_INC_GEN(a));
+	prt_printf(out, "dirty_sectors     %u\n",	a->dirty_sectors);
+	prt_printf(out, "cached_sectors    %u\n",	a->cached_sectors);
+	prt_printf(out, "stripe            %u\n",	a->stripe);
+	prt_printf(out, "stripe_redundancy %u\n",	a->stripe_redundancy);
+	prt_printf(out, "io_time[READ]     %llu\n",	a->io_time[READ]);
+	prt_printf(out, "io_time[WRITE]    %llu\n",	a->io_time[WRITE]);
+	prt_printf(out, "fragmentation     %llu\n",	a->fragmentation_lru);
+	prt_printf(out, "bp_start          %llu\n", BCH_ALLOC_V4_BACKPOINTERS_START(a));
 	printbuf_indent_sub(out, 2);
 }
 
@@ -437,22 +438,18 @@ struct bkey_i_alloc_v4 *bch2_alloc_to_v4_mut(struct btree_trans *trans, struct b
 }
 
 struct bkey_i_alloc_v4 *
-bch2_trans_start_alloc_update(struct btree_trans *trans, struct btree_iter *iter,
-			      struct bpos pos)
+bch2_trans_start_alloc_update_noupdate(struct btree_trans *trans, struct btree_iter *iter,
+				       struct bpos pos)
 {
-	struct bkey_s_c k;
-	struct bkey_i_alloc_v4 *a;
-	int ret;
-
-	k = bch2_bkey_get_iter(trans, iter, BTREE_ID_alloc, pos,
-			     BTREE_ITER_WITH_UPDATES|
-			     BTREE_ITER_CACHED|
-			     BTREE_ITER_INTENT);
-	ret = bkey_err(k);
+	struct bkey_s_c k = bch2_bkey_get_iter(trans, iter, BTREE_ID_alloc, pos,
+					       BTREE_ITER_with_updates|
+					       BTREE_ITER_cached|
+					       BTREE_ITER_intent);
+	int ret = bkey_err(k);
 	if (unlikely(ret))
 		return ERR_PTR(ret);
 
-	a = bch2_alloc_to_v4_mut_inlined(trans, k);
+	struct bkey_i_alloc_v4 *a = bch2_alloc_to_v4_mut_inlined(trans, k);
 	ret = PTR_ERR_OR_ZERO(a);
 	if (unlikely(ret))
 		goto err;
@@ -462,6 +459,20 @@ err:
 	return ERR_PTR(ret);
 }
 
+__flatten
+struct bkey_i_alloc_v4 *bch2_trans_start_alloc_update(struct btree_trans *trans, struct bpos pos)
+{
+	struct btree_iter iter;
+	struct bkey_i_alloc_v4 *a = bch2_trans_start_alloc_update_noupdate(trans, &iter, pos);
+	int ret = PTR_ERR_OR_ZERO(a);
+	if (ret)
+		return ERR_PTR(ret);
+
+	ret = bch2_trans_update(trans, &iter, &a->k_i, 0);
+	bch2_trans_iter_exit(trans, &iter);
+	return unlikely(ret) ? ERR_PTR(ret) : a;
+}
+
 static struct bpos alloc_gens_pos(struct bpos pos, unsigned *offset)
 {
 	*offset = pos.offset & KEY_TYPE_BUCKET_GENS_MASK;
@@ -485,7 +496,7 @@ static unsigned alloc_gen(struct bkey_s_c k, unsigned offset)
 }
 
 int bch2_bucket_gens_invalid(struct bch_fs *c, struct bkey_s_c k,
-			     enum bkey_invalid_flags flags,
+			     enum bch_validate_flags flags,
 			     struct printbuf *err)
 {
 	int ret = 0;
@@ -518,7 +529,7 @@ int bch2_bucket_gens_init(struct bch_fs *c)
 	int ret;
 
 	ret = for_each_btree_key(trans, iter, BTREE_ID_alloc, POS_MIN,
-				 BTREE_ITER_PREFETCH, k, ({
+				 BTREE_ITER_prefetch, k, ({
 		/*
 		 * Not a fsck error because this is checked/repaired by
 		 * bch2_check_alloc_key() which runs later:
@@ -530,13 +541,13 @@ int bch2_bucket_gens_init(struct bch_fs *c)
 		u8 gen = bch2_alloc_to_v4(k, &a)->gen;
 		unsigned offset;
 		struct bpos pos = alloc_gens_pos(iter.pos, &offset);
+		int ret2 = 0;
 
 		if (have_bucket_gens_key && bkey_cmp(iter.pos, pos)) {
-			ret = commit_do(trans, NULL, NULL,
-					BCH_TRANS_COMMIT_no_enospc,
-				bch2_btree_insert_trans(trans, BTREE_ID_bucket_gens, &g.k_i, 0));
-			if (ret)
-				break;
+			ret2 =  bch2_btree_insert_trans(trans, BTREE_ID_bucket_gens, &g.k_i, 0) ?:
+				bch2_trans_commit(trans, NULL, NULL, BCH_TRANS_COMMIT_no_enospc);
+			if (ret2)
+				goto iter_err;
 			have_bucket_gens_key = false;
 		}
 
@@ -547,7 +558,8 @@ int bch2_bucket_gens_init(struct bch_fs *c)
 		}
 
 		g.v.gens[offset] = gen;
-		0;
+iter_err:
+		ret2;
 	}));
 
 	if (have_bucket_gens_key && !ret)
@@ -564,29 +576,31 @@ int bch2_bucket_gens_init(struct bch_fs *c)
 int bch2_alloc_read(struct bch_fs *c)
 {
 	struct btree_trans *trans = bch2_trans_get(c);
+	struct bch_dev *ca = NULL;
 	int ret;
 
 	down_read(&c->gc_lock);
 
 	if (c->sb.version_upgrade_complete >= bcachefs_metadata_version_bucket_gens) {
 		ret = for_each_btree_key(trans, iter, BTREE_ID_bucket_gens, POS_MIN,
-					 BTREE_ITER_PREFETCH, k, ({
+					 BTREE_ITER_prefetch, k, ({
 			u64 start = bucket_gens_pos_to_alloc(k.k->p, 0).offset;
 			u64 end = bucket_gens_pos_to_alloc(bpos_nosnap_successor(k.k->p), 0).offset;
 
 			if (k.k->type != KEY_TYPE_bucket_gens)
 				continue;
 
-			const struct bch_bucket_gens *g = bkey_s_c_to_bucket_gens(k).v;
-
+			ca = bch2_dev_iterate(c, ca, k.k->p.inode);
 			/*
 			 * Not a fsck error because this is checked/repaired by
 			 * bch2_check_alloc_key() which runs later:
 			 */
-			if (!bch2_dev_exists2(c, k.k->p.inode))
+			if (!ca) {
+				bch2_btree_iter_set_pos(&iter, POS(k.k->p.inode + 1, 0));
 				continue;
+			}
 
-			struct bch_dev *ca = bch_dev_bkey_exists(c, k.k->p.inode);
+			const struct bch_bucket_gens *g = bkey_s_c_to_bucket_gens(k).v;
 
 			for (u64 b = max_t(u64, ca->mi.first_bucket, start);
 			     b < min_t(u64, ca->mi.nbuckets, end);
@@ -596,15 +610,16 @@ int bch2_alloc_read(struct bch_fs *c)
 		}));
 	} else {
 		ret = for_each_btree_key(trans, iter, BTREE_ID_alloc, POS_MIN,
-					 BTREE_ITER_PREFETCH, k, ({
+					 BTREE_ITER_prefetch, k, ({
+			ca = bch2_dev_iterate(c, ca, k.k->p.inode);
 			/*
 			 * Not a fsck error because this is checked/repaired by
 			 * bch2_check_alloc_key() which runs later:
 			 */
-			if (!bch2_dev_bucket_exists(c, k.k->p))
+			if (!ca) {
+				bch2_btree_iter_set_pos(&iter, POS(k.k->p.inode + 1, 0));
 				continue;
-
-			struct bch_dev *ca = bch_dev_bkey_exists(c, k.k->p.inode);
+			}
 
 			struct bch_alloc_v4 a;
 			*bucket_gen(ca, k.k->p.offset) = bch2_alloc_to_v4(k, &a)->gen;
@@ -612,6 +627,7 @@ int bch2_alloc_read(struct bch_fs *c)
 		}));
 	}
 
+	bch2_dev_put(ca);
 	bch2_trans_put(trans);
 	up_read(&c->gc_lock);
 
@@ -622,12 +638,12 @@ int bch2_alloc_read(struct bch_fs *c)
 /* Free space/discard btree: */
 
 static int bch2_bucket_do_index(struct btree_trans *trans,
+				struct bch_dev *ca,
 				struct bkey_s_c alloc_k,
 				const struct bch_alloc_v4 *a,
 				bool set)
 {
 	struct bch_fs *c = trans->c;
-	struct bch_dev *ca = bch_dev_bkey_exists(c, alloc_k.k->p.inode);
 	struct btree_iter iter;
 	struct bkey_s_c old;
 	struct bkey_i *k;
@@ -664,7 +680,7 @@ static int bch2_bucket_do_index(struct btree_trans *trans,
 
 	old = bch2_bkey_get_iter(trans, &iter, btree,
 			     bkey_start_pos(&k->k),
-			     BTREE_ITER_INTENT);
+			     BTREE_ITER_intent);
 	ret = bkey_err(old);
 	if (ret)
 		return ret;
@@ -708,8 +724,8 @@ static noinline int bch2_bucket_gen_update(struct btree_trans *trans,
 		return ret;
 
 	k = bch2_bkey_get_iter(trans, &iter, BTREE_ID_bucket_gens, pos,
-			       BTREE_ITER_INTENT|
-			       BTREE_ITER_WITH_UPDATES);
+			       BTREE_ITER_intent|
+			       BTREE_ITER_with_updates);
 	ret = bkey_err(k);
 	if (ret)
 		return ret;
@@ -731,28 +747,27 @@ static noinline int bch2_bucket_gen_update(struct btree_trans *trans,
 int bch2_trigger_alloc(struct btree_trans *trans,
 		       enum btree_id btree, unsigned level,
 		       struct bkey_s_c old, struct bkey_s new,
-		       unsigned flags)
+		       enum btree_iter_update_trigger_flags flags)
 {
 	struct bch_fs *c = trans->c;
+	struct printbuf buf = PRINTBUF;
 	int ret = 0;
 
-	if (bch2_trans_inconsistent_on(!bch2_dev_bucket_exists(c, new.k->p), trans,
-				       "alloc key for invalid device or bucket"))
+	struct bch_dev *ca = bch2_dev_bucket_tryget(c, new.k->p);
+	if (!ca)
 		return -EIO;
 
-	struct bch_dev *ca = bch_dev_bkey_exists(c, new.k->p.inode);
-
 	struct bch_alloc_v4 old_a_convert;
 	const struct bch_alloc_v4 *old_a = bch2_alloc_to_v4(old, &old_a_convert);
 
-	if (flags & BTREE_TRIGGER_TRANSACTIONAL) {
+	if (flags & BTREE_TRIGGER_transactional) {
 		struct bch_alloc_v4 *new_a = bkey_s_to_alloc_v4(new).v;
 
-		new_a->data_type = alloc_data_type(*new_a, new_a->data_type);
+		alloc_data_type_set(new_a, new_a->data_type);
 
-		if (bch2_bucket_sectors(*new_a) > bch2_bucket_sectors(*old_a)) {
-			new_a->io_time[READ] = max_t(u64, 1, atomic64_read(&c->io_clock[READ].now));
-			new_a->io_time[WRITE]= max_t(u64, 1, atomic64_read(&c->io_clock[WRITE].now));
+		if (bch2_bucket_sectors_total(*new_a) > bch2_bucket_sectors_total(*old_a)) {
+			new_a->io_time[READ] = bch2_current_io_time(c, READ);
+			new_a->io_time[WRITE]= bch2_current_io_time(c, WRITE);
 			SET_BCH_ALLOC_V4_NEED_INC_GEN(new_a, true);
 			SET_BCH_ALLOC_V4_NEED_DISCARD(new_a, true);
 		}
@@ -762,20 +777,21 @@ int bch2_trigger_alloc(struct btree_trans *trans,
 		    !bch2_bucket_is_open_safe(c, new.k->p.inode, new.k->p.offset)) {
 			new_a->gen++;
 			SET_BCH_ALLOC_V4_NEED_INC_GEN(new_a, false);
+			alloc_data_type_set(new_a, new_a->data_type);
 		}
 
 		if (old_a->data_type != new_a->data_type ||
 		    (new_a->data_type == BCH_DATA_free &&
 		     alloc_freespace_genbits(*old_a) != alloc_freespace_genbits(*new_a))) {
-			ret =   bch2_bucket_do_index(trans, old, old_a, false) ?:
-				bch2_bucket_do_index(trans, new.s_c, new_a, true);
+			ret =   bch2_bucket_do_index(trans, ca, old, old_a, false) ?:
+				bch2_bucket_do_index(trans, ca, new.s_c, new_a, true);
 			if (ret)
-				return ret;
+				goto err;
 		}
 
 		if (new_a->data_type == BCH_DATA_cached &&
 		    !new_a->io_time[READ])
-			new_a->io_time[READ] = max_t(u64, 1, atomic64_read(&c->io_clock[READ].now));
+			new_a->io_time[READ] = bch2_current_io_time(c, READ);
 
 		u64 old_lru = alloc_lru_idx_read(*old_a);
 		u64 new_lru = alloc_lru_idx_read(*new_a);
@@ -784,24 +800,23 @@ int bch2_trigger_alloc(struct btree_trans *trans,
 					      bucket_to_u64(new.k->p),
 					      old_lru, new_lru);
 			if (ret)
-				return ret;
+				goto err;
 		}
 
-		new_a->fragmentation_lru = alloc_lru_idx_fragmentation(*new_a,
-						bch_dev_bkey_exists(c, new.k->p.inode));
+		new_a->fragmentation_lru = alloc_lru_idx_fragmentation(*new_a, ca);
 		if (old_a->fragmentation_lru != new_a->fragmentation_lru) {
 			ret = bch2_lru_change(trans,
 					BCH_LRU_FRAGMENTATION_START,
 					bucket_to_u64(new.k->p),
 					old_a->fragmentation_lru, new_a->fragmentation_lru);
 			if (ret)
-				return ret;
+				goto err;
 		}
 
 		if (old_a->gen != new_a->gen) {
 			ret = bch2_bucket_gen_update(trans, new.k->p, new_a->gen);
 			if (ret)
-				return ret;
+				goto err;
 		}
 
 		/*
@@ -809,21 +824,21 @@ int bch2_trigger_alloc(struct btree_trans *trans,
 		 * not:
 		 */
 
-		if ((flags & BTREE_TRIGGER_BUCKET_INVALIDATE) &&
+		if ((flags & BTREE_TRIGGER_bucket_invalidate) &&
 		    old_a->cached_sectors) {
 			ret = bch2_update_cached_sectors_list(trans, new.k->p.inode,
 							      -((s64) old_a->cached_sectors));
 			if (ret)
-				return ret;
+				goto err;
 		}
 	}
 
-	if ((flags & BTREE_TRIGGER_ATOMIC) && (flags & BTREE_TRIGGER_INSERT)) {
+	if ((flags & BTREE_TRIGGER_atomic) && (flags & BTREE_TRIGGER_insert)) {
 		struct bch_alloc_v4 *new_a = bkey_s_to_alloc_v4(new).v;
 		u64 journal_seq = trans->journal_res.seq;
 		u64 bucket_journal_seq = new_a->journal_seq;
 
-		if ((flags & BTREE_TRIGGER_INSERT) &&
+		if ((flags & BTREE_TRIGGER_insert) &&
 		    data_type_is_empty(old_a->data_type) !=
 		    data_type_is_empty(new_a->data_type) &&
 		    new.k->type == KEY_TYPE_alloc_v4) {
@@ -850,42 +865,58 @@ int bch2_trigger_alloc(struct btree_trans *trans,
 					bucket_journal_seq);
 			if (ret) {
 				bch2_fs_fatal_error(c,
-					"error setting bucket_needs_journal_commit: %i", ret);
-				return ret;
+					"setting bucket_needs_journal_commit: %s", bch2_err_str(ret));
+				goto err;
 			}
 		}
 
 		percpu_down_read(&c->mark_lock);
-		if (new_a->gen != old_a->gen)
-			*bucket_gen(ca, new.k->p.offset) = new_a->gen;
+		if (new_a->gen != old_a->gen) {
+			u8 *gen = bucket_gen(ca, new.k->p.offset);
+			if (unlikely(!gen)) {
+				percpu_up_read(&c->mark_lock);
+				goto invalid_bucket;
+			}
+			*gen = new_a->gen;
+		}
 
 		bch2_dev_usage_update(c, ca, old_a, new_a, journal_seq, false);
+		percpu_up_read(&c->mark_lock);
 
-		if (new_a->data_type == BCH_DATA_free &&
-		    (!new_a->journal_seq || new_a->journal_seq < c->journal.flushed_seq_ondisk))
+#define eval_state(_a, expr)		({ const struct bch_alloc_v4 *a = _a; expr; })
+#define statechange(expr)		!eval_state(old_a, expr) && eval_state(new_a, expr)
+#define bucket_flushed(a)		(!a->journal_seq || a->journal_seq <= c->journal.flushed_seq_ondisk)
+
+		if (statechange(a->data_type == BCH_DATA_free) &&
+		    bucket_flushed(new_a))
 			closure_wake_up(&c->freelist_wait);
 
-		if (new_a->data_type == BCH_DATA_need_discard &&
-		    (!bucket_journal_seq || bucket_journal_seq < c->journal.flushed_seq_ondisk))
-			bch2_do_discards(c);
+		if (statechange(a->data_type == BCH_DATA_need_discard) &&
+		    !bch2_bucket_is_open_safe(c, new.k->p.inode, new.k->p.offset) &&
+		    bucket_flushed(new_a))
+			bch2_discard_one_bucket_fast(ca, new.k->p.offset);
 
-		if (old_a->data_type != BCH_DATA_cached &&
-		    new_a->data_type == BCH_DATA_cached &&
+		if (statechange(a->data_type == BCH_DATA_cached) &&
+		    !bch2_bucket_is_open(c, new.k->p.inode, new.k->p.offset) &&
 		    should_invalidate_buckets(ca, bch2_dev_usage_read(ca)))
-			bch2_do_invalidates(c);
+			bch2_dev_do_invalidates(ca);
 
-		if (new_a->data_type == BCH_DATA_need_gc_gens)
-			bch2_do_gc_gens(c);
-		percpu_up_read(&c->mark_lock);
+		if (statechange(a->data_type == BCH_DATA_need_gc_gens))
+			bch2_gc_gens_async(c);
 	}
 
-	if ((flags & BTREE_TRIGGER_GC) &&
-	    (flags & BTREE_TRIGGER_BUCKET_INVALIDATE)) {
+	if ((flags & BTREE_TRIGGER_gc) &&
+	    (flags & BTREE_TRIGGER_bucket_invalidate)) {
 		struct bch_alloc_v4 new_a_convert;
 		const struct bch_alloc_v4 *new_a = bch2_alloc_to_v4(new.s_c, &new_a_convert);
 
 		percpu_down_read(&c->mark_lock);
 		struct bucket *g = gc_bucket(ca, new.k->p.offset);
+		if (unlikely(!g)) {
+			percpu_up_read(&c->mark_lock);
+			goto invalid_bucket;
+		}
+		g->gen_valid	= 1;
 
 		bucket_lock(g);
 
@@ -900,12 +931,19 @@ int bch2_trigger_alloc(struct btree_trans *trans,
 		bucket_unlock(g);
 		percpu_up_read(&c->mark_lock);
 	}
-
-	return 0;
+err:
+	printbuf_exit(&buf);
+	bch2_dev_put(ca);
+	return ret;
+invalid_bucket:
+	bch2_fs_inconsistent(c, "reference to invalid bucket\n  %s",
+			     (bch2_bkey_val_to_text(&buf, c, new.s_c), buf.buf));
+	ret = -EIO;
+	goto err;
 }
 
 /*
- * This synthesizes deleted extents for holes, similar to BTREE_ITER_SLOTS for
+ * This synthesizes deleted extents for holes, similar to BTREE_ITER_slots for
  * extents style btrees, but works on non-extents btrees:
  */
 static struct bkey_s_c bch2_get_key_or_hole(struct btree_iter *iter, struct bpos end, struct bkey *hole)
@@ -950,35 +988,34 @@ static struct bkey_s_c bch2_get_key_or_hole(struct btree_iter *iter, struct bpos
 	}
 }
 
-static bool next_bucket(struct bch_fs *c, struct bpos *bucket)
+static bool next_bucket(struct bch_fs *c, struct bch_dev **ca, struct bpos *bucket)
 {
-	struct bch_dev *ca;
-
-	if (bch2_dev_bucket_exists(c, *bucket))
-		return true;
+	if (*ca) {
+		if (bucket->offset < (*ca)->mi.first_bucket)
+			bucket->offset = (*ca)->mi.first_bucket;
 
-	if (bch2_dev_exists2(c, bucket->inode)) {
-		ca = bch_dev_bkey_exists(c, bucket->inode);
-
-		if (bucket->offset < ca->mi.first_bucket) {
-			bucket->offset = ca->mi.first_bucket;
+		if (bucket->offset < (*ca)->mi.nbuckets)
 			return true;
-		}
 
+		bch2_dev_put(*ca);
+		*ca = NULL;
 		bucket->inode++;
 		bucket->offset = 0;
 	}
 
 	rcu_read_lock();
-	ca = __bch2_next_dev_idx(c, bucket->inode, NULL);
-	if (ca)
-		*bucket = POS(ca->dev_idx, ca->mi.first_bucket);
+	*ca = __bch2_next_dev_idx(c, bucket->inode, NULL);
+	if (*ca) {
+		*bucket = POS((*ca)->dev_idx, (*ca)->mi.first_bucket);
+		bch2_dev_get(*ca);
+	}
 	rcu_read_unlock();
 
-	return ca != NULL;
+	return *ca != NULL;
 }
 
-static struct bkey_s_c bch2_get_key_or_real_bucket_hole(struct btree_iter *iter, struct bkey *hole)
+static struct bkey_s_c bch2_get_key_or_real_bucket_hole(struct btree_iter *iter,
+					struct bch_dev **ca, struct bkey *hole)
 {
 	struct bch_fs *c = iter->trans->c;
 	struct bkey_s_c k;
@@ -987,22 +1024,21 @@ again:
 	if (bkey_err(k))
 		return k;
 
+	*ca = bch2_dev_iterate_noerror(c, *ca, k.k->p.inode);
+
 	if (!k.k->type) {
-		struct bpos bucket = bkey_start_pos(k.k);
+		struct bpos hole_start = bkey_start_pos(k.k);
 
-		if (!bch2_dev_bucket_exists(c, bucket)) {
-			if (!next_bucket(c, &bucket))
+		if (!*ca || !bucket_valid(*ca, hole_start.offset)) {
+			if (!next_bucket(c, ca, &hole_start))
 				return bkey_s_c_null;
 
-			bch2_btree_iter_set_pos(iter, bucket);
+			bch2_btree_iter_set_pos(iter, hole_start);
 			goto again;
 		}
 
-		if (!bch2_dev_bucket_exists(c, k.k->p)) {
-			struct bch_dev *ca = bch_dev_bkey_exists(c, bucket.inode);
-
-			bch2_key_resize(hole, ca->mi.nbuckets - bucket.offset);
-		}
+		if (k.k->p.offset > (*ca)->mi.nbuckets)
+			bch2_key_resize(hole, (*ca)->mi.nbuckets - hole_start.offset);
 	}
 
 	return k;
@@ -1017,24 +1053,25 @@ int bch2_check_alloc_key(struct btree_trans *trans,
 			 struct btree_iter *bucket_gens_iter)
 {
 	struct bch_fs *c = trans->c;
-	struct bch_dev *ca;
 	struct bch_alloc_v4 a_convert;
 	const struct bch_alloc_v4 *a;
 	unsigned discard_key_type, freespace_key_type;
 	unsigned gens_offset;
 	struct bkey_s_c k;
 	struct printbuf buf = PRINTBUF;
-	int ret;
+	int ret = 0;
 
-	if (fsck_err_on(!bch2_dev_bucket_exists(c, alloc_k.k->p), c,
-			alloc_key_to_missing_dev_bucket,
+	struct bch_dev *ca = bch2_dev_bucket_tryget_noerror(c, alloc_k.k->p);
+	if (fsck_err_on(!ca,
+			c, alloc_key_to_missing_dev_bucket,
 			"alloc key for invalid device:bucket %llu:%llu",
 			alloc_k.k->p.inode, alloc_k.k->p.offset))
-		return bch2_btree_delete_at(trans, alloc_iter, 0);
+		ret = bch2_btree_delete_at(trans, alloc_iter, 0);
+	if (!ca)
+		return ret;
 
-	ca = bch_dev_bkey_exists(c, alloc_k.k->p.inode);
 	if (!ca->mi.freespace_initialized)
-		return 0;
+		goto out;
 
 	a = bch2_alloc_to_v4(alloc_k, &a_convert);
 
@@ -1045,14 +1082,13 @@ int bch2_check_alloc_key(struct btree_trans *trans,
 	if (ret)
 		goto err;
 
-	if (k.k->type != discard_key_type &&
-	    (c->opts.reconstruct_alloc ||
-	     fsck_err(c, need_discard_key_wrong,
-		      "incorrect key in need_discard btree (got %s should be %s)\n"
-		      "  %s",
-		      bch2_bkey_types[k.k->type],
-		      bch2_bkey_types[discard_key_type],
-		      (bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf)))) {
+	if (fsck_err_on(k.k->type != discard_key_type,
+			c, need_discard_key_wrong,
+			"incorrect key in need_discard btree (got %s should be %s)\n"
+			"  %s",
+			bch2_bkey_types[k.k->type],
+			bch2_bkey_types[discard_key_type],
+			(bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf))) {
 		struct bkey_i *update =
 			bch2_trans_kmalloc(trans, sizeof(*update));
 
@@ -1076,15 +1112,14 @@ int bch2_check_alloc_key(struct btree_trans *trans,
 	if (ret)
 		goto err;
 
-	if (k.k->type != freespace_key_type &&
-	    (c->opts.reconstruct_alloc ||
-	     fsck_err(c, freespace_key_wrong,
-		      "incorrect key in freespace btree (got %s should be %s)\n"
-		      "  %s",
-		      bch2_bkey_types[k.k->type],
-		      bch2_bkey_types[freespace_key_type],
-		      (printbuf_reset(&buf),
-		       bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf)))) {
+	if (fsck_err_on(k.k->type != freespace_key_type,
+			c, freespace_key_wrong,
+			"incorrect key in freespace btree (got %s should be %s)\n"
+			"  %s",
+			bch2_bkey_types[k.k->type],
+			bch2_bkey_types[freespace_key_type],
+			(printbuf_reset(&buf),
+			 bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf))) {
 		struct bkey_i *update =
 			bch2_trans_kmalloc(trans, sizeof(*update));
 
@@ -1108,14 +1143,13 @@ int bch2_check_alloc_key(struct btree_trans *trans,
 	if (ret)
 		goto err;
 
-	if (a->gen != alloc_gen(k, gens_offset) &&
-	    (c->opts.reconstruct_alloc ||
-	     fsck_err(c, bucket_gens_key_wrong,
-		      "incorrect gen in bucket_gens btree (got %u should be %u)\n"
-		      "  %s",
-		      alloc_gen(k, gens_offset), a->gen,
-		      (printbuf_reset(&buf),
-		       bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf)))) {
+	if (fsck_err_on(a->gen != alloc_gen(k, gens_offset),
+			c, bucket_gens_key_wrong,
+			"incorrect gen in bucket_gens btree (got %u should be %u)\n"
+			"  %s",
+			alloc_gen(k, gens_offset), a->gen,
+			(printbuf_reset(&buf),
+			 bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf))) {
 		struct bkey_i_bucket_gens *g =
 			bch2_trans_kmalloc(trans, sizeof(*g));
 
@@ -1136,25 +1170,26 @@ int bch2_check_alloc_key(struct btree_trans *trans,
 		if (ret)
 			goto err;
 	}
+out:
 err:
 fsck_err:
+	bch2_dev_put(ca);
 	printbuf_exit(&buf);
 	return ret;
 }
 
 static noinline_for_stack
 int bch2_check_alloc_hole_freespace(struct btree_trans *trans,
+				    struct bch_dev *ca,
 				    struct bpos start,
 				    struct bpos *end,
 				    struct btree_iter *freespace_iter)
 {
 	struct bch_fs *c = trans->c;
-	struct bch_dev *ca;
 	struct bkey_s_c k;
 	struct printbuf buf = PRINTBUF;
 	int ret;
 
-	ca = bch_dev_bkey_exists(c, start.inode);
 	if (!ca->mi.freespace_initialized)
 		return 0;
 
@@ -1167,14 +1202,13 @@ int bch2_check_alloc_hole_freespace(struct btree_trans *trans,
 
 	*end = bkey_min(k.k->p, *end);
 
-	if (k.k->type != KEY_TYPE_set &&
-	    (c->opts.reconstruct_alloc ||
-	     fsck_err(c, freespace_hole_missing,
-		      "hole in alloc btree missing in freespace btree\n"
-		      "  device %llu buckets %llu-%llu",
-		      freespace_iter->pos.inode,
-		      freespace_iter->pos.offset,
-		      end->offset))) {
+	if (fsck_err_on(k.k->type != KEY_TYPE_set,
+			c, freespace_hole_missing,
+			"hole in alloc btree missing in freespace btree\n"
+			"  device %llu buckets %llu-%llu",
+			freespace_iter->pos.inode,
+			freespace_iter->pos.offset,
+			end->offset)) {
 		struct bkey_i *update =
 			bch2_trans_kmalloc(trans, sizeof(*update));
 
@@ -1309,7 +1343,7 @@ static noinline_for_stack int bch2_check_discard_freespace_key(struct btree_tran
 		goto delete;
 out:
 fsck_err:
-	set_btree_iter_dontneed(&alloc_iter);
+	bch2_set_btree_iter_dontneed(&alloc_iter);
 	bch2_trans_iter_exit(trans, &alloc_iter);
 	printbuf_exit(&buf);
 	return ret;
@@ -1333,30 +1367,25 @@ int bch2_check_bucket_gens_key(struct btree_trans *trans,
 {
 	struct bch_fs *c = trans->c;
 	struct bkey_i_bucket_gens g;
-	struct bch_dev *ca;
 	u64 start = bucket_gens_pos_to_alloc(k.k->p, 0).offset;
 	u64 end = bucket_gens_pos_to_alloc(bpos_nosnap_successor(k.k->p), 0).offset;
 	u64 b;
-	bool need_update = false, dev_exists;
+	bool need_update = false;
 	struct printbuf buf = PRINTBUF;
 	int ret = 0;
 
 	BUG_ON(k.k->type != KEY_TYPE_bucket_gens);
 	bkey_reassemble(&g.k_i, k);
 
-	/* if no bch_dev, skip out whether we repair or not */
-	dev_exists = bch2_dev_exists2(c, k.k->p.inode);
-	if (!dev_exists) {
-		if (fsck_err_on(!dev_exists, c,
-				bucket_gens_to_invalid_dev,
-				"bucket_gens key for invalid device:\n  %s",
-				(bch2_bkey_val_to_text(&buf, c, k), buf.buf))) {
+	struct bch_dev *ca = bch2_dev_tryget_noerror(c, k.k->p.inode);
+	if (!ca) {
+		if (fsck_err(c, bucket_gens_to_invalid_dev,
+			     "bucket_gens key for invalid device:\n  %s",
+			     (bch2_bkey_val_to_text(&buf, c, k), buf.buf)))
 			ret = bch2_btree_delete_at(trans, iter, 0);
-		}
 		goto out;
 	}
 
-	ca = bch_dev_bkey_exists(c, k.k->p.inode);
 	if (fsck_err_on(end <= ca->mi.first_bucket ||
 			start >= ca->mi.nbuckets, c,
 			bucket_gens_to_invalid_buckets,
@@ -1394,6 +1423,7 @@ int bch2_check_bucket_gens_key(struct btree_trans *trans,
 	}
 out:
 fsck_err:
+	bch2_dev_put(ca);
 	printbuf_exit(&buf);
 	return ret;
 }
@@ -1402,25 +1432,26 @@ int bch2_check_alloc_info(struct bch_fs *c)
 {
 	struct btree_trans *trans = bch2_trans_get(c);
 	struct btree_iter iter, discard_iter, freespace_iter, bucket_gens_iter;
+	struct bch_dev *ca = NULL;
 	struct bkey hole;
 	struct bkey_s_c k;
 	int ret = 0;
 
 	bch2_trans_iter_init(trans, &iter, BTREE_ID_alloc, POS_MIN,
-			     BTREE_ITER_PREFETCH);
+			     BTREE_ITER_prefetch);
 	bch2_trans_iter_init(trans, &discard_iter, BTREE_ID_need_discard, POS_MIN,
-			     BTREE_ITER_PREFETCH);
+			     BTREE_ITER_prefetch);
 	bch2_trans_iter_init(trans, &freespace_iter, BTREE_ID_freespace, POS_MIN,
-			     BTREE_ITER_PREFETCH);
+			     BTREE_ITER_prefetch);
 	bch2_trans_iter_init(trans, &bucket_gens_iter, BTREE_ID_bucket_gens, POS_MIN,
-			     BTREE_ITER_PREFETCH);
+			     BTREE_ITER_prefetch);
 
 	while (1) {
 		struct bpos next;
 
 		bch2_trans_begin(trans);
 
-		k = bch2_get_key_or_real_bucket_hole(&iter, &hole);
+		k = bch2_get_key_or_real_bucket_hole(&iter, &ca, &hole);
 		ret = bkey_err(k);
 		if (ret)
 			goto bkey_err;
@@ -1441,7 +1472,7 @@ int bch2_check_alloc_info(struct bch_fs *c)
 		} else {
 			next = k.k->p;
 
-			ret = bch2_check_alloc_hole_freespace(trans,
+			ret = bch2_check_alloc_hole_freespace(trans, ca,
 						    bkey_start_pos(k.k),
 						    &next,
 						    &freespace_iter) ?:
@@ -1469,19 +1500,21 @@ bkey_err:
 	bch2_trans_iter_exit(trans, &freespace_iter);
 	bch2_trans_iter_exit(trans, &discard_iter);
 	bch2_trans_iter_exit(trans, &iter);
+	bch2_dev_put(ca);
+	ca = NULL;
 
 	if (ret < 0)
 		goto err;
 
 	ret = for_each_btree_key(trans, iter,
 			BTREE_ID_need_discard, POS_MIN,
-			BTREE_ITER_PREFETCH, k,
+			BTREE_ITER_prefetch, k,
 		bch2_check_discard_freespace_key(trans, &iter));
 	if (ret)
 		goto err;
 
 	bch2_trans_iter_init(trans, &iter, BTREE_ID_freespace, POS_MIN,
-			     BTREE_ITER_PREFETCH);
+			     BTREE_ITER_prefetch);
 	while (1) {
 		bch2_trans_begin(trans);
 		k = bch2_btree_iter_peek(&iter);
@@ -1511,7 +1544,7 @@ bkey_err:
 
 	ret = for_each_btree_key_commit(trans, iter,
 			BTREE_ID_bucket_gens, POS_MIN,
-			BTREE_ITER_PREFETCH, k,
+			BTREE_ITER_prefetch, k,
 			NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
 		bch2_check_bucket_gens_key(trans, &iter, k));
 err:
@@ -1521,13 +1554,13 @@ err:
 }
 
 static int bch2_check_alloc_to_lru_ref(struct btree_trans *trans,
-				       struct btree_iter *alloc_iter)
+				       struct btree_iter *alloc_iter,
+				       struct bkey_buf *last_flushed)
 {
 	struct bch_fs *c = trans->c;
-	struct btree_iter lru_iter;
 	struct bch_alloc_v4 a_convert;
 	const struct bch_alloc_v4 *a;
-	struct bkey_s_c alloc_k, lru_k;
+	struct bkey_s_c alloc_k;
 	struct printbuf buf = PRINTBUF;
 	int ret;
 
@@ -1541,6 +1574,14 @@ static int bch2_check_alloc_to_lru_ref(struct btree_trans *trans,
 
 	a = bch2_alloc_to_v4(alloc_k, &a_convert);
 
+	if (a->fragmentation_lru) {
+		ret = bch2_lru_check_set(trans, BCH_LRU_FRAGMENTATION_START,
+					 a->fragmentation_lru,
+					 alloc_k, last_flushed);
+		if (ret)
+			return ret;
+	}
+
 	if (a->data_type != BCH_DATA_cached)
 		return 0;
 
@@ -1556,81 +1597,87 @@ static int bch2_check_alloc_to_lru_ref(struct btree_trans *trans,
 		if (ret)
 			goto err;
 
-		a_mut->v.io_time[READ] = atomic64_read(&c->io_clock[READ].now);
+		a_mut->v.io_time[READ] = bch2_current_io_time(c, READ);
 		ret = bch2_trans_update(trans, alloc_iter,
-					&a_mut->k_i, BTREE_TRIGGER_NORUN);
+					&a_mut->k_i, BTREE_TRIGGER_norun);
 		if (ret)
 			goto err;
 
 		a = &a_mut->v;
 	}
 
-	lru_k = bch2_bkey_get_iter(trans, &lru_iter, BTREE_ID_lru,
-			     lru_pos(alloc_k.k->p.inode,
-				     bucket_to_u64(alloc_k.k->p),
-				     a->io_time[READ]), 0);
-	ret = bkey_err(lru_k);
+	ret = bch2_lru_check_set(trans, alloc_k.k->p.inode, a->io_time[READ],
+				 alloc_k, last_flushed);
 	if (ret)
-		return ret;
-
-	if (fsck_err_on(lru_k.k->type != KEY_TYPE_set, c,
-			alloc_key_to_missing_lru_entry,
-			"missing lru entry\n"
-			"  %s",
-			(printbuf_reset(&buf),
-			 bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf))) {
-		ret = bch2_lru_set(trans,
-				   alloc_k.k->p.inode,
-				   bucket_to_u64(alloc_k.k->p),
-				   a->io_time[READ]);
-		if (ret)
-			goto err;
-	}
+		goto err;
 err:
 fsck_err:
-	bch2_trans_iter_exit(trans, &lru_iter);
 	printbuf_exit(&buf);
 	return ret;
 }
 
 int bch2_check_alloc_to_lru_refs(struct bch_fs *c)
 {
+	struct bkey_buf last_flushed;
+
+	bch2_bkey_buf_init(&last_flushed);
+	bkey_init(&last_flushed.k->k);
+
 	int ret = bch2_trans_run(c,
 		for_each_btree_key_commit(trans, iter, BTREE_ID_alloc,
-				POS_MIN, BTREE_ITER_PREFETCH, k,
+				POS_MIN, BTREE_ITER_prefetch, k,
 				NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
-			bch2_check_alloc_to_lru_ref(trans, &iter)));
+			bch2_check_alloc_to_lru_ref(trans, &iter, &last_flushed)));
+
+	bch2_bkey_buf_exit(&last_flushed, c);
 	bch_err_fn(c, ret);
 	return ret;
 }
 
+static int discard_in_flight_add(struct bch_dev *ca, u64 bucket, bool in_progress)
+{
+	int ret;
+
+	mutex_lock(&ca->discard_buckets_in_flight_lock);
+	darray_for_each(ca->discard_buckets_in_flight, i)
+		if (i->bucket == bucket) {
+			ret = -BCH_ERR_EEXIST_discard_in_flight_add;
+			goto out;
+		}
+
+	ret = darray_push(&ca->discard_buckets_in_flight, ((struct discard_in_flight) {
+			   .in_progress = in_progress,
+			   .bucket	= bucket,
+	}));
+out:
+	mutex_unlock(&ca->discard_buckets_in_flight_lock);
+	return ret;
+}
+
+static void discard_in_flight_remove(struct bch_dev *ca, u64 bucket)
+{
+	mutex_lock(&ca->discard_buckets_in_flight_lock);
+	darray_for_each(ca->discard_buckets_in_flight, i)
+		if (i->bucket == bucket) {
+			BUG_ON(!i->in_progress);
+			darray_remove_item(&ca->discard_buckets_in_flight, i);
+			goto found;
+		}
+	BUG();
+found:
+	mutex_unlock(&ca->discard_buckets_in_flight_lock);
+}
+
 struct discard_buckets_state {
 	u64		seen;
 	u64		open;
 	u64		need_journal_commit;
 	u64		discarded;
-	struct bch_dev	*ca;
 	u64		need_journal_commit_this_dev;
 };
 
-static void discard_buckets_next_dev(struct bch_fs *c, struct discard_buckets_state *s, struct bch_dev *ca)
-{
-	if (s->ca == ca)
-		return;
-
-	if (s->ca && s->need_journal_commit_this_dev >
-	    bch2_dev_usage_read(s->ca).d[BCH_DATA_free].buckets)
-		bch2_journal_flush_async(&c->journal, NULL);
-
-	if (s->ca)
-		percpu_ref_put(&s->ca->ref);
-	if (ca)
-		percpu_ref_get(&ca->ref);
-	s->ca = ca;
-	s->need_journal_commit_this_dev = 0;
-}
-
 static int bch2_discard_one_bucket(struct btree_trans *trans,
+				   struct bch_dev *ca,
 				   struct btree_iter *need_discard_iter,
 				   struct bpos *discard_pos_done,
 				   struct discard_buckets_state *s)
@@ -1639,20 +1686,11 @@ static int bch2_discard_one_bucket(struct btree_trans *trans,
 	struct bpos pos = need_discard_iter->pos;
 	struct btree_iter iter = { NULL };
 	struct bkey_s_c k;
-	struct bch_dev *ca;
 	struct bkey_i_alloc_v4 *a;
 	struct printbuf buf = PRINTBUF;
+	bool discard_locked = false;
 	int ret = 0;
 
-	ca = bch_dev_bkey_exists(c, pos.inode);
-
-	if (!percpu_ref_tryget(&ca->io_ref)) {
-		bch2_btree_iter_set_pos(need_discard_iter, POS(pos.inode + 1, 0));
-		return 0;
-	}
-
-	discard_buckets_next_dev(c, s, ca);
-
 	if (bch2_bucket_is_open_safe(c, pos.inode, pos.offset)) {
 		s->open++;
 		goto out;
@@ -1668,7 +1706,7 @@ static int bch2_discard_one_bucket(struct btree_trans *trans,
 
 	k = bch2_bkey_get_iter(trans, &iter, BTREE_ID_alloc,
 			       need_discard_iter->pos,
-			       BTREE_ITER_CACHED);
+			       BTREE_ITER_cached);
 	ret = bkey_err(k);
 	if (ret)
 		goto out;
@@ -1678,37 +1716,45 @@ static int bch2_discard_one_bucket(struct btree_trans *trans,
 	if (ret)
 		goto out;
 
-	if (BCH_ALLOC_V4_NEED_INC_GEN(&a->v)) {
-		a->v.gen++;
-		SET_BCH_ALLOC_V4_NEED_INC_GEN(&a->v, false);
-		goto write;
-	}
-
-	if (a->v.journal_seq > c->journal.flushed_seq_ondisk) {
-		if (c->curr_recovery_pass > BCH_RECOVERY_PASS_check_alloc_info) {
-			bch2_trans_inconsistent(trans,
-				"clearing need_discard but journal_seq %llu > flushed_seq %llu\n"
-				"%s",
-				a->v.journal_seq,
-				c->journal.flushed_seq_ondisk,
-				(bch2_bkey_val_to_text(&buf, c, k), buf.buf));
+	if (bch2_bucket_sectors_total(a->v)) {
+		if (bch2_trans_inconsistent_on(c->curr_recovery_pass > BCH_RECOVERY_PASS_check_alloc_info,
+					       trans, "attempting to discard bucket with dirty data\n%s",
+					       (bch2_bkey_val_to_text(&buf, c, k), buf.buf)))
 			ret = -EIO;
-		}
 		goto out;
 	}
 
 	if (a->v.data_type != BCH_DATA_need_discard) {
-		if (c->curr_recovery_pass > BCH_RECOVERY_PASS_check_alloc_info) {
-			bch2_trans_inconsistent(trans,
-				"bucket incorrectly set in need_discard btree\n"
-				"%s",
-				(bch2_bkey_val_to_text(&buf, c, k), buf.buf));
-			ret = -EIO;
+		if (data_type_is_empty(a->v.data_type) &&
+		    BCH_ALLOC_V4_NEED_INC_GEN(&a->v)) {
+			a->v.gen++;
+			SET_BCH_ALLOC_V4_NEED_INC_GEN(&a->v, false);
+			goto write;
 		}
 
+		if (bch2_trans_inconsistent_on(c->curr_recovery_pass > BCH_RECOVERY_PASS_check_alloc_info,
+					       trans, "bucket incorrectly set in need_discard btree\n"
+					       "%s",
+					       (bch2_bkey_val_to_text(&buf, c, k), buf.buf)))
+			ret = -EIO;
+		goto out;
+	}
+
+	if (a->v.journal_seq > c->journal.flushed_seq_ondisk) {
+		if (bch2_trans_inconsistent_on(c->curr_recovery_pass > BCH_RECOVERY_PASS_check_alloc_info,
+					       trans, "clearing need_discard but journal_seq %llu > flushed_seq %llu\n%s",
+					       a->v.journal_seq,
+					       c->journal.flushed_seq_ondisk,
+					       (bch2_bkey_val_to_text(&buf, c, k), buf.buf)))
+			ret = -EIO;
 		goto out;
 	}
 
+	if (discard_in_flight_add(ca, iter.pos.offset, true))
+		goto out;
+
+	discard_locked = true;
+
 	if (!bkey_eq(*discard_pos_done, iter.pos) &&
 	    ca->mi.discard && !c->opts.nochanges) {
 		/*
@@ -1728,8 +1774,9 @@ static int bch2_discard_one_bucket(struct btree_trans *trans,
 	}
 
 	SET_BCH_ALLOC_V4_NEED_DISCARD(&a->v, false);
-	a->v.data_type = alloc_data_type(a->v, a->v.data_type);
 write:
+	alloc_data_type_set(&a->v, a->v.data_type);
+
 	ret =   bch2_trans_update(trans, &iter, &a->k_i, 0) ?:
 		bch2_trans_commit(trans, NULL, NULL,
 				  BCH_WATERMARK_btree|
@@ -1740,16 +1787,18 @@ write:
 	count_event(c, bucket_discard);
 	s->discarded++;
 out:
+	if (discard_locked)
+		discard_in_flight_remove(ca, iter.pos.offset);
 	s->seen++;
 	bch2_trans_iter_exit(trans, &iter);
-	percpu_ref_put(&ca->io_ref);
 	printbuf_exit(&buf);
 	return ret;
 }
 
 static void bch2_do_discards_work(struct work_struct *work)
 {
-	struct bch_fs *c = container_of(work, struct bch_fs, discard_work);
+	struct bch_dev *ca = container_of(work, struct bch_dev, discard_work);
+	struct bch_fs *c = ca->fs;
 	struct discard_buckets_state s = {};
 	struct bpos discard_pos_done = POS_MAX;
 	int ret;
@@ -1760,23 +1809,132 @@ static void bch2_do_discards_work(struct work_struct *work)
 	 * successful commit:
 	 */
 	ret = bch2_trans_run(c,
-		for_each_btree_key(trans, iter,
-				   BTREE_ID_need_discard, POS_MIN, 0, k,
-			bch2_discard_one_bucket(trans, &iter, &discard_pos_done, &s)));
-
-	discard_buckets_next_dev(c, &s, NULL);
+		for_each_btree_key_upto(trans, iter,
+				   BTREE_ID_need_discard,
+				   POS(ca->dev_idx, 0),
+				   POS(ca->dev_idx, U64_MAX), 0, k,
+			bch2_discard_one_bucket(trans, ca, &iter, &discard_pos_done, &s)));
 
 	trace_discard_buckets(c, s.seen, s.open, s.need_journal_commit, s.discarded,
 			      bch2_err_str(ret));
 
 	bch2_write_ref_put(c, BCH_WRITE_REF_discard);
+	percpu_ref_put(&ca->io_ref);
+}
+
+void bch2_dev_do_discards(struct bch_dev *ca)
+{
+	struct bch_fs *c = ca->fs;
+
+	if (!bch2_dev_get_ioref(c, ca->dev_idx, WRITE))
+		return;
+
+	if (!bch2_write_ref_tryget(c, BCH_WRITE_REF_discard))
+		goto put_ioref;
+
+	if (queue_work(c->write_ref_wq, &ca->discard_work))
+		return;
+
+	bch2_write_ref_put(c, BCH_WRITE_REF_discard);
+put_ioref:
+	percpu_ref_put(&ca->io_ref);
 }
 
 void bch2_do_discards(struct bch_fs *c)
 {
-	if (bch2_write_ref_tryget(c, BCH_WRITE_REF_discard) &&
-	    !queue_work(c->write_ref_wq, &c->discard_work))
-		bch2_write_ref_put(c, BCH_WRITE_REF_discard);
+	for_each_member_device(c, ca)
+		bch2_dev_do_discards(ca);
+}
+
+static int bch2_clear_bucket_needs_discard(struct btree_trans *trans, struct bpos bucket)
+{
+	struct btree_iter iter;
+	bch2_trans_iter_init(trans, &iter, BTREE_ID_alloc, bucket, BTREE_ITER_intent);
+	struct bkey_s_c k = bch2_btree_iter_peek_slot(&iter);
+	int ret = bkey_err(k);
+	if (ret)
+		goto err;
+
+	struct bkey_i_alloc_v4 *a = bch2_alloc_to_v4_mut(trans, k);
+	ret = PTR_ERR_OR_ZERO(a);
+	if (ret)
+		goto err;
+
+	BUG_ON(a->v.dirty_sectors);
+	SET_BCH_ALLOC_V4_NEED_DISCARD(&a->v, false);
+	alloc_data_type_set(&a->v, a->v.data_type);
+
+	ret = bch2_trans_update(trans, &iter, &a->k_i, 0);
+err:
+	bch2_trans_iter_exit(trans, &iter);
+	return ret;
+}
+
+static void bch2_do_discards_fast_work(struct work_struct *work)
+{
+	struct bch_dev *ca = container_of(work, struct bch_dev, discard_fast_work);
+	struct bch_fs *c = ca->fs;
+
+	while (1) {
+		bool got_bucket = false;
+		u64 bucket;
+
+		mutex_lock(&ca->discard_buckets_in_flight_lock);
+		darray_for_each(ca->discard_buckets_in_flight, i) {
+			if (i->in_progress)
+				continue;
+
+			got_bucket = true;
+			bucket = i->bucket;
+			i->in_progress = true;
+			break;
+		}
+		mutex_unlock(&ca->discard_buckets_in_flight_lock);
+
+		if (!got_bucket)
+			break;
+
+		if (ca->mi.discard && !c->opts.nochanges)
+			blkdev_issue_discard(ca->disk_sb.bdev,
+					     bucket_to_sector(ca, bucket),
+					     ca->mi.bucket_size,
+					     GFP_KERNEL);
+
+		int ret = bch2_trans_do(c, NULL, NULL,
+			BCH_WATERMARK_btree|
+			BCH_TRANS_COMMIT_no_enospc,
+			bch2_clear_bucket_needs_discard(trans, POS(ca->dev_idx, bucket)));
+		bch_err_fn(c, ret);
+
+		discard_in_flight_remove(ca, bucket);
+
+		if (ret)
+			break;
+	}
+
+	bch2_write_ref_put(c, BCH_WRITE_REF_discard_fast);
+	percpu_ref_put(&ca->io_ref);
+}
+
+static void bch2_discard_one_bucket_fast(struct bch_dev *ca, u64 bucket)
+{
+	struct bch_fs *c = ca->fs;
+
+	if (discard_in_flight_add(ca, bucket, false))
+		return;
+
+	if (!bch2_dev_get_ioref(c, ca->dev_idx, WRITE))
+		return;
+
+	if (!bch2_write_ref_tryget(c, BCH_WRITE_REF_discard_fast))
+		goto put_ioref;
+
+	if (queue_work(c->write_ref_wq, &ca->discard_fast_work))
+		return;
+
+	bch2_write_ref_put(c, BCH_WRITE_REF_discard_fast);
+put_ioref:
+	percpu_ref_put(&ca->io_ref);
 }
 
 static int invalidate_one_bucket(struct btree_trans *trans,
@@ -1785,7 +1943,6 @@ static int invalidate_one_bucket(struct btree_trans *trans,
 				 s64 *nr_to_invalidate)
 {
 	struct bch_fs *c = trans->c;
-	struct btree_iter alloc_iter = { NULL };
 	struct bkey_i_alloc_v4 *a = NULL;
 	struct printbuf buf = PRINTBUF;
 	struct bpos bucket = u64_to_bucket(lru_k.k->p.offset);
@@ -1803,7 +1960,7 @@ static int invalidate_one_bucket(struct btree_trans *trans,
 	if (bch2_bucket_is_open_safe(c, bucket.inode, bucket.offset))
 		return 0;
 
-	a = bch2_trans_start_alloc_update(trans, &alloc_iter, bucket);
+	a = bch2_trans_start_alloc_update(trans, bucket);
 	ret = PTR_ERR_OR_ZERO(a);
 	if (ret)
 		goto out;
@@ -1813,6 +1970,7 @@ static int invalidate_one_bucket(struct btree_trans *trans,
 		goto out;
 
 	BUG_ON(a->v.data_type != BCH_DATA_cached);
+	BUG_ON(a->v.dirty_sectors);
 
 	if (!a->v.cached_sectors)
 		bch_err(c, "invalidating empty bucket, confused");
@@ -1824,21 +1982,18 @@ static int invalidate_one_bucket(struct btree_trans *trans,
 	a->v.data_type		= 0;
 	a->v.dirty_sectors	= 0;
 	a->v.cached_sectors	= 0;
-	a->v.io_time[READ]	= atomic64_read(&c->io_clock[READ].now);
-	a->v.io_time[WRITE]	= atomic64_read(&c->io_clock[WRITE].now);
+	a->v.io_time[READ]	= bch2_current_io_time(c, READ);
+	a->v.io_time[WRITE]	= bch2_current_io_time(c, WRITE);
 
-	ret =   bch2_trans_update(trans, &alloc_iter, &a->k_i,
-				BTREE_TRIGGER_BUCKET_INVALIDATE) ?:
-		bch2_trans_commit(trans, NULL, NULL,
-				  BCH_WATERMARK_btree|
-				  BCH_TRANS_COMMIT_no_enospc);
+	ret = bch2_trans_commit(trans, NULL, NULL,
+				BCH_WATERMARK_btree|
+				BCH_TRANS_COMMIT_no_enospc);
 	if (ret)
 		goto out;
 
 	trace_and_count(c, bucket_invalidate, c, bucket.inode, bucket.offset, cached_sectors);
 	--*nr_to_invalidate;
 out:
-	bch2_trans_iter_exit(trans, &alloc_iter);
 	printbuf_exit(&buf);
 	return ret;
 err:
@@ -1863,9 +2018,25 @@ err:
 	goto out;
 }
 
+static struct bkey_s_c next_lru_key(struct btree_trans *trans, struct btree_iter *iter,
+				    struct bch_dev *ca, bool *wrapped)
+{
+	struct bkey_s_c k;
+again:
+	k = bch2_btree_iter_peek_upto(iter, lru_pos(ca->dev_idx, U64_MAX, LRU_TIME_MAX));
+	if (!k.k && !*wrapped) {
+		bch2_btree_iter_set_pos(iter, lru_pos(ca->dev_idx, 0, 0));
+		*wrapped = true;
+		goto again;
+	}
+
+	return k;
+}
+
 static void bch2_do_invalidates_work(struct work_struct *work)
 {
-	struct bch_fs *c = container_of(work, struct bch_fs, invalidate_work);
+	struct bch_dev *ca = container_of(work, struct bch_dev, invalidate_work);
+	struct bch_fs *c = ca->fs;
 	struct btree_trans *trans = bch2_trans_get(c);
 	int ret = 0;
 
@@ -1873,31 +2044,63 @@ static void bch2_do_invalidates_work(struct work_struct *work)
 	if (ret)
 		goto err;
 
-	for_each_member_device(c, ca) {
-		s64 nr_to_invalidate =
-			should_invalidate_buckets(ca, bch2_dev_usage_read(ca));
+	s64 nr_to_invalidate =
+		should_invalidate_buckets(ca, bch2_dev_usage_read(ca));
+	struct btree_iter iter;
+	bool wrapped = false;
+
+	bch2_trans_iter_init(trans, &iter, BTREE_ID_lru,
+			     lru_pos(ca->dev_idx, 0,
+				     ((bch2_current_io_time(c, READ) + U32_MAX) &
+				      LRU_TIME_MAX)), 0);
 
-		ret = for_each_btree_key_upto(trans, iter, BTREE_ID_lru,
-				lru_pos(ca->dev_idx, 0, 0),
-				lru_pos(ca->dev_idx, U64_MAX, LRU_TIME_MAX),
-				BTREE_ITER_INTENT, k,
-			invalidate_one_bucket(trans, &iter, k, &nr_to_invalidate));
+	while (true) {
+		bch2_trans_begin(trans);
+
+		struct bkey_s_c k = next_lru_key(trans, &iter, ca, &wrapped);
+		ret = bkey_err(k);
+		if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
+			continue;
+		if (ret)
+			break;
+		if (!k.k)
+			break;
 
-		if (ret < 0) {
-			percpu_ref_put(&ca->ref);
+		ret = invalidate_one_bucket(trans, &iter, k, &nr_to_invalidate);
+		if (ret)
 			break;
-		}
+
+		bch2_btree_iter_advance(&iter);
 	}
+	bch2_trans_iter_exit(trans, &iter);
 err:
 	bch2_trans_put(trans);
 	bch2_write_ref_put(c, BCH_WRITE_REF_invalidate);
+	percpu_ref_put(&ca->io_ref);
+}
+
+void bch2_dev_do_invalidates(struct bch_dev *ca)
+{
+	struct bch_fs *c = ca->fs;
+
+	if (!bch2_dev_get_ioref(c, ca->dev_idx, WRITE))
+		return;
+
+	if (!bch2_write_ref_tryget(c, BCH_WRITE_REF_invalidate))
+		goto put_ioref;
+
+	if (queue_work(c->write_ref_wq, &ca->invalidate_work))
+		return;
+
+	bch2_write_ref_put(c, BCH_WRITE_REF_invalidate);
+put_ioref:
+	percpu_ref_put(&ca->io_ref);
 }
 
 void bch2_do_invalidates(struct bch_fs *c)
 {
-	if (bch2_write_ref_tryget(c, BCH_WRITE_REF_invalidate) &&
-	    !queue_work(c->write_ref_wq, &c->invalidate_work))
-		bch2_write_ref_put(c, BCH_WRITE_REF_invalidate);
+	for_each_member_device(c, ca)
+		bch2_dev_do_invalidates(ca);
 }
 
 int bch2_dev_freespace_init(struct bch_fs *c, struct bch_dev *ca,
@@ -1917,7 +2120,7 @@ int bch2_dev_freespace_init(struct bch_fs *c, struct bch_dev *ca,
 
 	bch2_trans_iter_init(trans, &iter, BTREE_ID_alloc,
 		POS(ca->dev_idx, max_t(u64, ca->mi.first_bucket, bucket_start)),
-		BTREE_ITER_PREFETCH);
+		BTREE_ITER_prefetch);
 	/*
 	 * Scan the alloc btree for every bucket on @ca, and add buckets to the
 	 * freespace/need_discard/need_gc_gens btrees as needed:
@@ -1949,7 +2152,7 @@ int bch2_dev_freespace_init(struct bch_fs *c, struct bch_dev *ca,
 			struct bch_alloc_v4 a_convert;
 			const struct bch_alloc_v4 *a = bch2_alloc_to_v4(k, &a_convert);
 
-			ret =   bch2_bucket_do_index(trans, k, a, true) ?:
+			ret =   bch2_bucket_do_index(trans, ca, k, a, true) ?:
 				bch2_trans_commit(trans, NULL, NULL,
 						  BCH_TRANS_COMMIT_no_enospc);
 			if (ret)
@@ -2021,7 +2224,7 @@ int bch2_fs_freespace_init(struct bch_fs *c)
 
 		ret = bch2_dev_freespace_init(c, ca, 0, ca->mi.nbuckets);
 		if (ret) {
-			percpu_ref_put(&ca->ref);
+			bch2_dev_put(ca);
 			bch_err_fn(c, ret);
 			return ret;
 		}
@@ -2048,12 +2251,15 @@ int bch2_bucket_io_time_reset(struct btree_trans *trans, unsigned dev,
 	u64 now;
 	int ret = 0;
 
-	a = bch2_trans_start_alloc_update(trans, &iter,  POS(dev, bucket_nr));
+	if (bch2_trans_relock(trans))
+		bch2_trans_begin(trans);
+
+	a = bch2_trans_start_alloc_update_noupdate(trans, &iter, POS(dev, bucket_nr));
 	ret = PTR_ERR_OR_ZERO(a);
 	if (ret)
 		return ret;
 
-	now = atomic64_read(&c->io_clock[rw].now);
+	now = bch2_current_io_time(c, rw);
 	if (a->v.io_time[rw] == now)
 		goto out;
 
@@ -2210,9 +2416,20 @@ void bch2_dev_allocator_add(struct bch_fs *c, struct bch_dev *ca)
 			set_bit(ca->dev_idx, c->rw_devs[i].d);
 }
 
+void bch2_dev_allocator_background_exit(struct bch_dev *ca)
+{
+	darray_exit(&ca->discard_buckets_in_flight);
+}
+
+void bch2_dev_allocator_background_init(struct bch_dev *ca)
+{
+	mutex_init(&ca->discard_buckets_in_flight_lock);
+	INIT_WORK(&ca->discard_work, bch2_do_discards_work);
+	INIT_WORK(&ca->discard_fast_work, bch2_do_discards_fast_work);
+	INIT_WORK(&ca->invalidate_work, bch2_do_invalidates_work);
+}
+
 void bch2_fs_allocator_background_init(struct bch_fs *c)
 {
 	spin_lock_init(&c->freelist_lock);
-	INIT_WORK(&c->discard_work, bch2_do_discards_work);
-	INIT_WORK(&c->invalidate_work, bch2_do_invalidates_work);
 }
diff --git a/fs/bcachefs/alloc_background.h b/fs/bcachefs/alloc_background.h
index e7f7e842ee..ba2c5557a3 100644
--- a/fs/bcachefs/alloc_background.h
+++ b/fs/bcachefs/alloc_background.h
@@ -8,21 +8,18 @@
 #include "debug.h"
 #include "super.h"
 
-enum bkey_invalid_flags;
+enum bch_validate_flags;
 
 /* How out of date a pointer gen is allowed to be: */
 #define BUCKET_GC_GEN_MAX	96U
 
 static inline bool bch2_dev_bucket_exists(struct bch_fs *c, struct bpos pos)
 {
-	struct bch_dev *ca;
-
-	if (!bch2_dev_exists2(c, pos.inode))
-		return false;
-
-	ca = bch_dev_bkey_exists(c, pos.inode);
-	return pos.offset >= ca->mi.first_bucket &&
-		pos.offset < ca->mi.nbuckets;
+	rcu_read_lock();
+	struct bch_dev *ca = bch2_dev_rcu(c, pos.inode);
+	bool ret = ca && bucket_valid(ca, pos.offset);
+	rcu_read_unlock();
+	return ret;
 }
 
 static inline u64 bucket_to_u64(struct bpos bucket)
@@ -40,38 +37,50 @@ static inline u8 alloc_gc_gen(struct bch_alloc_v4 a)
 	return a.gen - a.oldest_gen;
 }
 
-static inline enum bch_data_type __alloc_data_type(u32 dirty_sectors,
-						   u32 cached_sectors,
-						   u32 stripe,
-						   struct bch_alloc_v4 a,
-						   enum bch_data_type data_type)
+static inline void alloc_to_bucket(struct bucket *dst, struct bch_alloc_v4 src)
 {
-	if (stripe)
-		return data_type == BCH_DATA_parity ? data_type : BCH_DATA_stripe;
-	if (dirty_sectors)
-		return data_type;
-	if (cached_sectors)
-		return BCH_DATA_cached;
-	if (BCH_ALLOC_V4_NEED_DISCARD(&a))
-		return BCH_DATA_need_discard;
-	if (alloc_gc_gen(a) >= BUCKET_GC_GEN_MAX)
-		return BCH_DATA_need_gc_gens;
-	return BCH_DATA_free;
+	dst->gen		= src.gen;
+	dst->data_type		= src.data_type;
+	dst->dirty_sectors	= src.dirty_sectors;
+	dst->cached_sectors	= src.cached_sectors;
+	dst->stripe		= src.stripe;
 }
 
-static inline enum bch_data_type alloc_data_type(struct bch_alloc_v4 a,
-						 enum bch_data_type data_type)
+static inline void __bucket_m_to_alloc(struct bch_alloc_v4 *dst, struct bucket src)
+{
+	dst->gen		= src.gen;
+	dst->data_type		= src.data_type;
+	dst->dirty_sectors	= src.dirty_sectors;
+	dst->cached_sectors	= src.cached_sectors;
+	dst->stripe		= src.stripe;
+}
+
+static inline struct bch_alloc_v4 bucket_m_to_alloc(struct bucket b)
 {
-	return __alloc_data_type(a.dirty_sectors, a.cached_sectors,
-				 a.stripe, a, data_type);
+	struct bch_alloc_v4 ret = {};
+	__bucket_m_to_alloc(&ret, b);
+	return ret;
 }
 
 static inline enum bch_data_type bucket_data_type(enum bch_data_type data_type)
 {
-	return data_type == BCH_DATA_stripe ? BCH_DATA_user : data_type;
+	switch (data_type) {
+	case BCH_DATA_cached:
+	case BCH_DATA_stripe:
+		return BCH_DATA_user;
+	default:
+		return data_type;
+	}
 }
 
-static inline unsigned bch2_bucket_sectors(struct bch_alloc_v4 a)
+static inline bool bucket_data_type_mismatch(enum bch_data_type bucket,
+					     enum bch_data_type ptr)
+{
+	return !data_type_is_empty(bucket) &&
+		bucket_data_type(bucket) != bucket_data_type(ptr);
+}
+
+static inline unsigned bch2_bucket_sectors_total(struct bch_alloc_v4 a)
 {
 	return a.dirty_sectors + a.cached_sectors;
 }
@@ -89,6 +98,27 @@ static inline unsigned bch2_bucket_sectors_fragmented(struct bch_dev *ca,
 	return d ? max(0, ca->mi.bucket_size - d) : 0;
 }
 
+static inline enum bch_data_type alloc_data_type(struct bch_alloc_v4 a,
+						 enum bch_data_type data_type)
+{
+	if (a.stripe)
+		return data_type == BCH_DATA_parity ? data_type : BCH_DATA_stripe;
+	if (a.dirty_sectors)
+		return data_type;
+	if (a.cached_sectors)
+		return BCH_DATA_cached;
+	if (BCH_ALLOC_V4_NEED_DISCARD(&a))
+		return BCH_DATA_need_discard;
+	if (alloc_gc_gen(a) >= BUCKET_GC_GEN_MAX)
+		return BCH_DATA_need_gc_gens;
+	return BCH_DATA_free;
+}
+
+static inline void alloc_data_type_set(struct bch_alloc_v4 *a, enum bch_data_type data_type)
+{
+	a->data_type = alloc_data_type(*a, data_type);
+}
+
 static inline u64 alloc_lru_idx_read(struct bch_alloc_v4 a)
 {
 	return a.data_type == BCH_DATA_cached ? a.io_time[READ] : 0;
@@ -111,7 +141,13 @@ static inline u64 alloc_lru_idx_fragmentation(struct bch_alloc_v4 a,
 	    !bch2_bucket_sectors_fragmented(ca, a))
 		return 0;
 
-	u64 d = bch2_bucket_sectors_dirty(a);
+	/*
+	 * avoid overflowing LRU_TIME_BITS on a corrupted fs, when
+	 * bucket_sectors_dirty is (much) bigger than bucket_size
+	 */
+	u64 d = min(bch2_bucket_sectors_dirty(a),
+		    ca->mi.bucket_size);
+
 	return div_u64(d * (1ULL << 31), ca->mi.bucket_size);
 }
 
@@ -126,13 +162,17 @@ static inline struct bpos alloc_freespace_pos(struct bpos pos, struct bch_alloc_
 	return pos;
 }
 
-static inline unsigned alloc_v4_u64s(const struct bch_alloc_v4 *a)
+static inline unsigned alloc_v4_u64s_noerror(const struct bch_alloc_v4 *a)
 {
-	unsigned ret = (BCH_ALLOC_V4_BACKPOINTERS_START(a) ?:
+	return (BCH_ALLOC_V4_BACKPOINTERS_START(a) ?:
 			BCH_ALLOC_V4_U64s_V0) +
 		BCH_ALLOC_V4_NR_BACKPOINTERS(a) *
 		(sizeof(struct bch_backpointer) / sizeof(u64));
+}
 
+static inline unsigned alloc_v4_u64s(const struct bch_alloc_v4 *a)
+{
+	unsigned ret = alloc_v4_u64s_noerror(a);
 	BUG_ON(ret > U8_MAX - BKEY_U64s);
 	return ret;
 }
@@ -143,7 +183,9 @@ static inline void set_alloc_v4_u64s(struct bkey_i_alloc_v4 *a)
 }
 
 struct bkey_i_alloc_v4 *
-bch2_trans_start_alloc_update(struct btree_trans *, struct btree_iter *, struct bpos);
+bch2_trans_start_alloc_update_noupdate(struct btree_trans *, struct btree_iter *, struct bpos);
+struct bkey_i_alloc_v4 *
+bch2_trans_start_alloc_update(struct btree_trans *, struct bpos);
 
 void __bch2_alloc_to_v4(struct bkey_s_c, struct bch_alloc_v4 *);
 
@@ -169,13 +211,13 @@ struct bkey_i_alloc_v4 *bch2_alloc_to_v4_mut(struct btree_trans *, struct bkey_s
 int bch2_bucket_io_time_reset(struct btree_trans *, unsigned, size_t, int);
 
 int bch2_alloc_v1_invalid(struct bch_fs *, struct bkey_s_c,
-			  enum bkey_invalid_flags, struct printbuf *);
+			  enum bch_validate_flags, struct printbuf *);
 int bch2_alloc_v2_invalid(struct bch_fs *, struct bkey_s_c,
-			  enum bkey_invalid_flags, struct printbuf *);
+			  enum bch_validate_flags, struct printbuf *);
 int bch2_alloc_v3_invalid(struct bch_fs *, struct bkey_s_c,
-			  enum bkey_invalid_flags, struct printbuf *);
+			  enum bch_validate_flags, struct printbuf *);
 int bch2_alloc_v4_invalid(struct bch_fs *, struct bkey_s_c,
-			  enum bkey_invalid_flags, struct printbuf *);
+			  enum bch_validate_flags, struct printbuf *);
 void bch2_alloc_v4_swab(struct bkey_s);
 void bch2_alloc_to_text(struct printbuf *, struct bch_fs *, struct bkey_s_c);
 
@@ -209,7 +251,7 @@ void bch2_alloc_to_text(struct printbuf *, struct bch_fs *, struct bkey_s_c);
 })
 
 int bch2_bucket_gens_invalid(struct bch_fs *, struct bkey_s_c,
-			     enum bkey_invalid_flags, struct printbuf *);
+			     enum bch_validate_flags, struct printbuf *);
 void bch2_bucket_gens_to_text(struct printbuf *, struct bch_fs *, struct bkey_s_c);
 
 #define bch2_bkey_ops_bucket_gens ((struct bkey_ops) {	\
@@ -229,9 +271,11 @@ static inline bool bkey_is_alloc(const struct bkey *k)
 int bch2_alloc_read(struct bch_fs *);
 
 int bch2_trigger_alloc(struct btree_trans *, enum btree_id, unsigned,
-		       struct bkey_s_c, struct bkey_s, unsigned);
+		       struct bkey_s_c, struct bkey_s,
+		       enum btree_iter_update_trigger_flags);
 int bch2_check_alloc_info(struct bch_fs *);
 int bch2_check_alloc_to_lru_refs(struct bch_fs *);
+void bch2_dev_do_discards(struct bch_dev *);
 void bch2_do_discards(struct bch_fs *);
 
 static inline u64 should_invalidate_buckets(struct bch_dev *ca,
@@ -246,6 +290,7 @@ static inline u64 should_invalidate_buckets(struct bch_dev *ca,
 	return clamp_t(s64, want_free - free, 0, u.d[BCH_DATA_cached].buckets);
 }
 
+void bch2_dev_do_invalidates(struct bch_dev *);
 void bch2_do_invalidates(struct bch_fs *);
 
 static inline struct bch_backpointer *alloc_v4_backpointers(struct bch_alloc_v4 *a)
@@ -269,6 +314,9 @@ u64 bch2_min_rw_member_capacity(struct bch_fs *);
 void bch2_dev_allocator_remove(struct bch_fs *, struct bch_dev *);
 void bch2_dev_allocator_add(struct bch_fs *, struct bch_dev *);
 
+void bch2_dev_allocator_background_exit(struct bch_dev *);
+void bch2_dev_allocator_background_init(struct bch_dev *);
+
 void bch2_fs_allocator_background_init(struct bch_fs *);
 
 #endif /* _BCACHEFS_ALLOC_BACKGROUND_H */
diff --git a/fs/bcachefs/alloc_foreground.c b/fs/bcachefs/alloc_foreground.c
index 633d3223b3..27d97c22ae 100644
--- a/fs/bcachefs/alloc_foreground.c
+++ b/fs/bcachefs/alloc_foreground.c
@@ -71,7 +71,7 @@ void bch2_reset_alloc_cursors(struct bch_fs *c)
 {
 	rcu_read_lock();
 	for_each_member_device_rcu(c, ca, NULL)
-		ca->alloc_cursor = 0;
+		memset(ca->alloc_cursor, 0, sizeof(ca->alloc_cursor));
 	rcu_read_unlock();
 }
 
@@ -100,7 +100,7 @@ static void bch2_open_bucket_hash_remove(struct bch_fs *c, struct open_bucket *o
 
 void __bch2_open_bucket_put(struct bch_fs *c, struct open_bucket *ob)
 {
-	struct bch_dev *ca = bch_dev_bkey_exists(c, ob->dev);
+	struct bch_dev *ca = ob_dev(c, ob);
 
 	if (ob->ec) {
 		ec_stripe_new_put(c, ob->ec, STRIPE_REF_io);
@@ -188,8 +188,10 @@ long bch2_bucket_alloc_new_fs(struct bch_dev *ca)
 static inline unsigned open_buckets_reserved(enum bch_watermark watermark)
 {
 	switch (watermark) {
-	case BCH_WATERMARK_reclaim:
+	case BCH_WATERMARK_interior_updates:
 		return 0;
+	case BCH_WATERMARK_reclaim:
+		return OPEN_BUCKETS_COUNT / 6;
 	case BCH_WATERMARK_btree:
 	case BCH_WATERMARK_btree_copygc:
 		return OPEN_BUCKETS_COUNT / 4;
@@ -236,8 +238,7 @@ static struct open_bucket *__try_alloc_bucket(struct bch_fs *c, struct bch_dev *
 		if (cl)
 			closure_wait(&c->open_buckets_wait, cl);
 
-		track_event_change(&c->times[BCH_TIME_blocked_allocate_open_bucket],
-				   &c->blocked_allocate_open_bucket, true);
+		track_event_change(&c->times[BCH_TIME_blocked_allocate_open_bucket], true);
 		spin_unlock(&c->freelist_lock);
 		return ERR_PTR(-BCH_ERR_open_buckets_empty);
 	}
@@ -263,11 +264,8 @@ static struct open_bucket *__try_alloc_bucket(struct bch_fs *c, struct bch_dev *
 	ca->nr_open_buckets++;
 	bch2_open_bucket_hash_add(c, ob);
 
-	track_event_change(&c->times[BCH_TIME_blocked_allocate_open_bucket],
-			   &c->blocked_allocate_open_bucket, false);
-
-	track_event_change(&c->times[BCH_TIME_blocked_allocate],
-			   &c->blocked_allocate, false);
+	track_event_change(&c->times[BCH_TIME_blocked_allocate_open_bucket], false);
+	track_event_change(&c->times[BCH_TIME_blocked_allocate], false);
 
 	spin_unlock(&c->freelist_lock);
 	return ob;
@@ -302,7 +300,7 @@ static struct open_bucket *try_alloc_bucket(struct btree_trans *trans, struct bc
 
 	k = bch2_bkey_get_iter(trans, &iter,
 			       BTREE_ID_alloc, POS(ca->dev_idx, b),
-			       BTREE_ITER_CACHED);
+			       BTREE_ITER_cached);
 	ret = bkey_err(k);
 	if (ret) {
 		ob = ERR_PTR(ret);
@@ -344,9 +342,9 @@ static struct open_bucket *try_alloc_bucket(struct btree_trans *trans, struct bc
 		struct bch_backpointer bp;
 		struct bpos bp_pos = POS_MIN;
 
-		ret = bch2_get_next_backpointer(trans, POS(ca->dev_idx, b), -1,
+		ret = bch2_get_next_backpointer(trans, ca, POS(ca->dev_idx, b), -1,
 						&bp_pos, &bp,
-						BTREE_ITER_NOPRESERVE);
+						BTREE_ITER_nopreserve);
 		if (ret) {
 			ob = ERR_PTR(ret);
 			goto err;
@@ -365,10 +363,10 @@ static struct open_bucket *try_alloc_bucket(struct btree_trans *trans, struct bc
 
 	ob = __try_alloc_bucket(c, ca, b, watermark, a, s, cl);
 	if (!ob)
-		set_btree_iter_dontneed(&iter);
+		bch2_set_btree_iter_dontneed(&iter);
 err:
 	if (iter.path)
-		set_btree_iter_dontneed(&iter);
+		bch2_set_btree_iter_dontneed(&iter);
 	bch2_trans_iter_exit(trans, &iter);
 	printbuf_exit(&buf);
 	return ob;
@@ -391,7 +389,8 @@ bch2_bucket_alloc_early(struct btree_trans *trans,
 	struct bkey_s_c k, ck;
 	struct open_bucket *ob = NULL;
 	u64 first_bucket = max_t(u64, ca->mi.first_bucket, ca->new_fs_bucket_idx);
-	u64 alloc_start = max(first_bucket, READ_ONCE(ca->alloc_cursor));
+	u64 *dev_alloc_cursor = &ca->alloc_cursor[s->btree_bitmap];
+	u64 alloc_start = max(first_bucket, *dev_alloc_cursor);
 	u64 alloc_cursor = alloc_start;
 	int ret;
 
@@ -406,9 +405,8 @@ bch2_bucket_alloc_early(struct btree_trans *trans,
 	 */
 again:
 	for_each_btree_key_norestart(trans, iter, BTREE_ID_alloc, POS(ca->dev_idx, alloc_cursor),
-			   BTREE_ITER_SLOTS, k, ret) {
-		struct bch_alloc_v4 a_convert;
-		const struct bch_alloc_v4 *a;
+			   BTREE_ITER_slots, k, ret) {
+		u64 bucket = k.k->p.offset;
 
 		if (bkey_ge(k.k->p, POS(ca->dev_idx, ca->mi.nbuckets)))
 			break;
@@ -417,12 +415,29 @@ again:
 		    is_superblock_bucket(ca, k.k->p.offset))
 			continue;
 
-		a = bch2_alloc_to_v4(k, &a_convert);
+		if (s->btree_bitmap != BTREE_BITMAP_ANY &&
+		    s->btree_bitmap != bch2_dev_btree_bitmap_marked_sectors(ca,
+				bucket_to_sector(ca, bucket), ca->mi.bucket_size)) {
+			if (s->btree_bitmap == BTREE_BITMAP_YES &&
+			    bucket_to_sector(ca, bucket) > 64ULL << ca->mi.btree_bitmap_shift)
+				break;
+
+			bucket = sector_to_bucket(ca,
+					round_up(bucket_to_sector(ca, bucket) + 1,
+						 1ULL << ca->mi.btree_bitmap_shift));
+			bch2_btree_iter_set_pos(&iter, POS(ca->dev_idx, bucket));
+			s->buckets_seen++;
+			s->skipped_mi_btree_bitmap++;
+			continue;
+		}
+
+		struct bch_alloc_v4 a_convert;
+		const struct bch_alloc_v4 *a = bch2_alloc_to_v4(k, &a_convert);
 		if (a->data_type != BCH_DATA_free)
 			continue;
 
 		/* now check the cached key to serialize concurrent allocs of the bucket */
-		ck = bch2_bkey_get_iter(trans, &citer, BTREE_ID_alloc, k.k->p, BTREE_ITER_CACHED);
+		ck = bch2_bkey_get_iter(trans, &citer, BTREE_ID_alloc, k.k->p, BTREE_ITER_cached);
 		ret = bkey_err(ck);
 		if (ret)
 			break;
@@ -435,7 +450,7 @@ again:
 
 		ob = __try_alloc_bucket(trans->c, ca, k.k->p.offset, watermark, a, s, cl);
 next:
-		set_btree_iter_dontneed(&citer);
+		bch2_set_btree_iter_dontneed(&citer);
 		bch2_trans_iter_exit(trans, &citer);
 		if (ob)
 			break;
@@ -443,7 +458,6 @@ next:
 	bch2_trans_iter_exit(trans, &iter);
 
 	alloc_cursor = iter.pos.offset;
-	ca->alloc_cursor = alloc_cursor;
 
 	if (!ob && ret)
 		ob = ERR_PTR(ret);
@@ -453,6 +467,8 @@ next:
 		goto again;
 	}
 
+	*dev_alloc_cursor = alloc_cursor;
+
 	return ob;
 }
 
@@ -465,7 +481,8 @@ static struct open_bucket *bch2_bucket_alloc_freelist(struct btree_trans *trans,
 	struct btree_iter iter;
 	struct bkey_s_c k;
 	struct open_bucket *ob = NULL;
-	u64 alloc_start = max_t(u64, ca->mi.first_bucket, READ_ONCE(ca->alloc_cursor));
+	u64 *dev_alloc_cursor = &ca->alloc_cursor[s->btree_bitmap];
+	u64 alloc_start = max_t(u64, ca->mi.first_bucket, READ_ONCE(*dev_alloc_cursor));
 	u64 alloc_cursor = alloc_start;
 	int ret;
 
@@ -487,10 +504,30 @@ again:
 
 			s->buckets_seen++;
 
+			u64 bucket = alloc_cursor & ~(~0ULL << 56);
+			if (s->btree_bitmap != BTREE_BITMAP_ANY &&
+			    s->btree_bitmap != bch2_dev_btree_bitmap_marked_sectors(ca,
+					bucket_to_sector(ca, bucket), ca->mi.bucket_size)) {
+				if (s->btree_bitmap == BTREE_BITMAP_YES &&
+				    bucket_to_sector(ca, bucket) > 64ULL << ca->mi.btree_bitmap_shift)
+					goto fail;
+
+				bucket = sector_to_bucket(ca,
+						round_up(bucket_to_sector(ca, bucket) + 1,
+							 1ULL << ca->mi.btree_bitmap_shift));
+				u64 genbits = alloc_cursor >> 56;
+				alloc_cursor = bucket | (genbits << 56);
+
+				if (alloc_cursor > k.k->p.offset)
+					bch2_btree_iter_set_pos(&iter, POS(ca->dev_idx, alloc_cursor));
+				s->skipped_mi_btree_bitmap++;
+				continue;
+			}
+
 			ob = try_alloc_bucket(trans, ca, watermark,
 					      alloc_cursor, s, k, cl);
 			if (ob) {
-				set_btree_iter_dontneed(&iter);
+				bch2_set_btree_iter_dontneed(&iter);
 				break;
 			}
 		}
@@ -498,10 +535,9 @@ again:
 		if (ob || ret)
 			break;
 	}
+fail:
 	bch2_trans_iter_exit(trans, &iter);
 
-	ca->alloc_cursor = alloc_cursor;
-
 	if (!ob && ret)
 		ob = ERR_PTR(ret);
 
@@ -510,14 +546,56 @@ again:
 		goto again;
 	}
 
+	*dev_alloc_cursor = alloc_cursor;
+
 	return ob;
 }
 
+static noinline void trace_bucket_alloc2(struct bch_fs *c, struct bch_dev *ca,
+					 enum bch_watermark watermark,
+					 enum bch_data_type data_type,
+					 struct closure *cl,
+					 struct bch_dev_usage *usage,
+					 struct bucket_alloc_state *s,
+					 struct open_bucket *ob)
+{
+	struct printbuf buf = PRINTBUF;
+
+	printbuf_tabstop_push(&buf, 24);
+
+	prt_printf(&buf, "dev\t%s (%u)\n",	ca->name, ca->dev_idx);
+	prt_printf(&buf, "watermark\t%s\n",	bch2_watermarks[watermark]);
+	prt_printf(&buf, "data type\t%s\n",	__bch2_data_types[data_type]);
+	prt_printf(&buf, "blocking\t%u\n",	cl != NULL);
+	prt_printf(&buf, "free\t%llu\n",	usage->d[BCH_DATA_free].buckets);
+	prt_printf(&buf, "avail\t%llu\n",	dev_buckets_free(ca, *usage, watermark));
+	prt_printf(&buf, "copygc_wait\t%lu/%lli\n",
+		   bch2_copygc_wait_amount(c),
+		   c->copygc_wait - atomic64_read(&c->io_clock[WRITE].now));
+	prt_printf(&buf, "seen\t%llu\n",	s->buckets_seen);
+	prt_printf(&buf, "open\t%llu\n",	s->skipped_open);
+	prt_printf(&buf, "need journal commit\t%llu\n", s->skipped_need_journal_commit);
+	prt_printf(&buf, "nocow\t%llu\n",	s->skipped_nocow);
+	prt_printf(&buf, "nouse\t%llu\n",	s->skipped_nouse);
+	prt_printf(&buf, "mi_btree_bitmap\t%llu\n", s->skipped_mi_btree_bitmap);
+
+	if (!IS_ERR(ob)) {
+		prt_printf(&buf, "allocated\t%llu\n", ob->bucket);
+		trace_bucket_alloc(c, buf.buf);
+	} else {
+		prt_printf(&buf, "err\t%s\n", bch2_err_str(PTR_ERR(ob)));
+		trace_bucket_alloc_fail(c, buf.buf);
+	}
+
+	printbuf_exit(&buf);
+}
+
 /**
  * bch2_bucket_alloc_trans - allocate a single bucket from a specific device
  * @trans:	transaction object
  * @ca:		device to allocate from
  * @watermark:	how important is this allocation?
+ * @data_type:	BCH_DATA_journal, btree, user...
  * @cl:		if not NULL, closure to be used to wait if buckets not available
  * @usage:	for secondarily also returning the current device usage
  *
@@ -526,6 +604,7 @@ again:
 static struct open_bucket *bch2_bucket_alloc_trans(struct btree_trans *trans,
 				      struct bch_dev *ca,
 				      enum bch_watermark watermark,
+				      enum bch_data_type data_type,
 				      struct closure *cl,
 				      struct bch_dev_usage *usage)
 {
@@ -533,20 +612,22 @@ static struct open_bucket *bch2_bucket_alloc_trans(struct btree_trans *trans,
 	struct open_bucket *ob = NULL;
 	bool freespace = READ_ONCE(ca->mi.freespace_initialized);
 	u64 avail;
-	struct bucket_alloc_state s = { 0 };
+	struct bucket_alloc_state s = {
+		.btree_bitmap = data_type == BCH_DATA_btree,
+	};
 	bool waiting = false;
 again:
 	bch2_dev_usage_read_fast(ca, usage);
 	avail = dev_buckets_free(ca, *usage, watermark);
 
 	if (usage->d[BCH_DATA_need_discard].buckets > avail)
-		bch2_do_discards(c);
+		bch2_dev_do_discards(ca);
 
 	if (usage->d[BCH_DATA_need_gc_gens].buckets > avail)
-		bch2_do_gc_gens(c);
+		bch2_gc_gens_async(c);
 
 	if (should_invalidate_buckets(ca, *usage))
-		bch2_do_invalidates(c);
+		bch2_dev_do_invalidates(ca);
 
 	if (!avail) {
 		if (cl && !waiting) {
@@ -555,8 +636,7 @@ again:
 			goto again;
 		}
 
-		track_event_change(&c->times[BCH_TIME_blocked_allocate],
-				   &c->blocked_allocate, true);
+		track_event_change(&c->times[BCH_TIME_blocked_allocate], true);
 
 		ob = ERR_PTR(-BCH_ERR_freelist_empty);
 		goto err;
@@ -572,6 +652,11 @@ alloc:
 	if (s.skipped_need_journal_commit * 2 > avail)
 		bch2_journal_flush_async(&c->journal, NULL);
 
+	if (!ob && s.btree_bitmap != BTREE_BITMAP_ANY) {
+		s.btree_bitmap = BTREE_BITMAP_ANY;
+		goto alloc;
+	}
+
 	if (!ob && freespace && c->curr_recovery_pass <= BCH_RECOVERY_PASS_check_alloc_info) {
 		freespace = false;
 		goto alloc;
@@ -581,33 +666,24 @@ err:
 		ob = ERR_PTR(-BCH_ERR_no_buckets_found);
 
 	if (!IS_ERR(ob))
-		trace_and_count(c, bucket_alloc, ca,
-				bch2_watermarks[watermark],
-				ob->bucket,
-				usage->d[BCH_DATA_free].buckets,
-				avail,
-				bch2_copygc_wait_amount(c),
-				c->copygc_wait - atomic64_read(&c->io_clock[WRITE].now),
-				&s,
-				cl == NULL,
-				"");
+		ob->data_type = data_type;
+
+	if (!IS_ERR(ob))
+		count_event(c, bucket_alloc);
 	else if (!bch2_err_matches(PTR_ERR(ob), BCH_ERR_transaction_restart))
-		trace_and_count(c, bucket_alloc_fail, ca,
-				bch2_watermarks[watermark],
-				0,
-				usage->d[BCH_DATA_free].buckets,
-				avail,
-				bch2_copygc_wait_amount(c),
-				c->copygc_wait - atomic64_read(&c->io_clock[WRITE].now),
-				&s,
-				cl == NULL,
-				bch2_err_str(PTR_ERR(ob)));
+		count_event(c, bucket_alloc_fail);
+
+	if (!IS_ERR(ob)
+	    ? trace_bucket_alloc_enabled()
+	    : trace_bucket_alloc_fail_enabled())
+		trace_bucket_alloc2(c, ca, watermark, data_type, cl, usage, &s, ob);
 
 	return ob;
 }
 
 struct open_bucket *bch2_bucket_alloc(struct bch_fs *c, struct bch_dev *ca,
 				      enum bch_watermark watermark,
+				      enum bch_data_type data_type,
 				      struct closure *cl)
 {
 	struct bch_dev_usage usage;
@@ -615,7 +691,7 @@ struct open_bucket *bch2_bucket_alloc(struct bch_fs *c, struct bch_dev *ca,
 
 	bch2_trans_do(c, NULL, NULL, 0,
 		      PTR_ERR_OR_ZERO(ob = bch2_bucket_alloc_trans(trans, ca, watermark,
-							cl, &usage)));
+							data_type, cl, &usage)));
 	return ob;
 }
 
@@ -681,8 +757,7 @@ static int add_new_bucket(struct bch_fs *c,
 			   unsigned flags,
 			   struct open_bucket *ob)
 {
-	unsigned durability =
-		bch_dev_bkey_exists(c, ob->dev)->mi.durability;
+	unsigned durability = ob_dev(c, ob)->mi.durability;
 
 	BUG_ON(*nr_effective >= nr_replicas);
 
@@ -714,37 +789,28 @@ int bch2_bucket_alloc_set_trans(struct btree_trans *trans,
 	struct bch_fs *c = trans->c;
 	struct dev_alloc_list devs_sorted =
 		bch2_dev_alloc_list(c, stripe, devs_may_alloc);
-	unsigned dev;
-	struct bch_dev *ca;
 	int ret = -BCH_ERR_insufficient_devices;
-	unsigned i;
 
 	BUG_ON(*nr_effective >= nr_replicas);
 
-	for (i = 0; i < devs_sorted.nr; i++) {
+	for (unsigned i = 0; i < devs_sorted.nr; i++) {
 		struct bch_dev_usage usage;
 		struct open_bucket *ob;
 
-		dev = devs_sorted.devs[i];
-
-		rcu_read_lock();
-		ca = rcu_dereference(c->devs[dev]);
-		if (ca)
-			percpu_ref_get(&ca->ref);
-		rcu_read_unlock();
-
+		unsigned dev = devs_sorted.devs[i];
+		struct bch_dev *ca = bch2_dev_tryget_noerror(c, dev);
 		if (!ca)
 			continue;
 
 		if (!ca->mi.durability && *have_cache) {
-			percpu_ref_put(&ca->ref);
+			bch2_dev_put(ca);
 			continue;
 		}
 
-		ob = bch2_bucket_alloc_trans(trans, ca, watermark, cl, &usage);
+		ob = bch2_bucket_alloc_trans(trans, ca, watermark, data_type, cl, &usage);
 		if (!IS_ERR(ob))
 			bch2_dev_stripe_increment_inlined(ca, stripe, &usage);
-		percpu_ref_put(&ca->ref);
+		bch2_dev_put(ca);
 
 		if (IS_ERR(ob)) {
 			ret = PTR_ERR(ob);
@@ -753,8 +819,6 @@ int bch2_bucket_alloc_set_trans(struct btree_trans *trans,
 			continue;
 		}
 
-		ob->data_type = data_type;
-
 		if (add_new_bucket(c, ptrs, devs_may_alloc,
 				   nr_replicas, nr_effective,
 				   have_cache, flags, ob)) {
@@ -839,7 +903,7 @@ static bool want_bucket(struct bch_fs *c,
 			bool *have_cache, bool ec,
 			struct open_bucket *ob)
 {
-	struct bch_dev *ca = bch_dev_bkey_exists(c, ob->dev);
+	struct bch_dev *ca = ob_dev(c, ob);
 
 	if (!test_bit(ob->dev, devs_may_alloc->d))
 		return false;
@@ -909,7 +973,7 @@ static int bucket_alloc_set_partial(struct bch_fs *c,
 		struct open_bucket *ob = c->open_buckets + c->open_buckets_partial[i];
 
 		if (want_bucket(c, wp, devs_may_alloc, have_cache, ec, ob)) {
-			struct bch_dev *ca = bch_dev_bkey_exists(c, ob->dev);
+			struct bch_dev *ca = ob_dev(c, ob);
 			struct bch_dev_usage usage;
 			u64 avail;
 
@@ -1294,7 +1358,7 @@ deallocate_extra_replicas(struct bch_fs *c,
 	unsigned i;
 
 	open_bucket_for_each(c, ptrs, ob, i) {
-		unsigned d = bch_dev_bkey_exists(c, ob->dev)->mi.durability;
+		unsigned d = ob_dev(c, ob)->mi.durability;
 
 		if (d && d <= extra_replicas) {
 			extra_replicas -= d;
@@ -1345,6 +1409,10 @@ retry:
 
 	*wp_ret = wp = writepoint_find(trans, write_point.v);
 
+	ret = bch2_trans_relock(trans);
+	if (ret)
+		goto err;
+
 	/* metadata may not allocate on cache devices: */
 	if (wp->data_type != BCH_DATA_user)
 		have_cache = true;
@@ -1361,15 +1429,17 @@ retry:
 
 		/* Don't retry from all devices if we're out of open buckets: */
 		if (bch2_err_matches(ret, BCH_ERR_open_buckets_empty)) {
-			int ret = open_bucket_add_buckets(trans, &ptrs, wp, devs_have,
+			int ret2 = open_bucket_add_buckets(trans, &ptrs, wp, devs_have,
 					      target, erasure_code,
 					      nr_replicas, &nr_effective,
 					      &have_cache, watermark,
 					      flags, cl);
-			if (!ret ||
-			    bch2_err_matches(ret, BCH_ERR_transaction_restart) ||
-			    bch2_err_matches(ret, BCH_ERR_open_buckets_empty))
+			if (!ret2 ||
+			    bch2_err_matches(ret2, BCH_ERR_transaction_restart) ||
+			    bch2_err_matches(ret2, BCH_ERR_open_buckets_empty)) {
+				ret = ret2;
 				goto alloc_done;
+			}
 		}
 
 		/*
@@ -1445,7 +1515,7 @@ err:
 
 struct bch_extent_ptr bch2_ob_ptr(struct bch_fs *c, struct open_bucket *ob)
 {
-	struct bch_dev *ca = bch_dev_bkey_exists(c, ob->dev);
+	struct bch_dev *ca = ob_dev(c, ob);
 
 	return (struct bch_extent_ptr) {
 		.type	= 1 << BCH_EXTENT_ENTRY_ptr,
@@ -1521,7 +1591,7 @@ void bch2_fs_allocator_foreground_init(struct bch_fs *c)
 
 static void bch2_open_bucket_to_text(struct printbuf *out, struct bch_fs *c, struct open_bucket *ob)
 {
-	struct bch_dev *ca = bch_dev_bkey_exists(c, ob->dev);
+	struct bch_dev *ca = ob_dev(c, ob);
 	unsigned data_type = ob->data_type;
 	barrier(); /* READ_ONCE() doesn't work on bitfields */
 
@@ -1623,3 +1693,106 @@ void bch2_write_points_to_text(struct printbuf *out, struct bch_fs *c)
 	prt_str(out, "Btree write point\n");
 	bch2_write_point_to_text(out, c, &c->btree_write_point);
 }
+
+void bch2_fs_alloc_debug_to_text(struct printbuf *out, struct bch_fs *c)
+{
+	unsigned nr[BCH_DATA_NR];
+
+	memset(nr, 0, sizeof(nr));
+
+	for (unsigned i = 0; i < ARRAY_SIZE(c->open_buckets); i++)
+		nr[c->open_buckets[i].data_type]++;
+
+	printbuf_tabstops_reset(out);
+	printbuf_tabstop_push(out, 24);
+
+	percpu_down_read(&c->mark_lock);
+	prt_printf(out, "hidden\t%llu\n",			bch2_fs_usage_read_one(c, &c->usage_base->b.hidden));
+	prt_printf(out, "btree\t%llu\n",			bch2_fs_usage_read_one(c, &c->usage_base->b.btree));
+	prt_printf(out, "data\t%llu\n",				bch2_fs_usage_read_one(c, &c->usage_base->b.data));
+	prt_printf(out, "cached\t%llu\n",			bch2_fs_usage_read_one(c, &c->usage_base->b.cached));
+	prt_printf(out, "reserved\t%llu\n",			bch2_fs_usage_read_one(c, &c->usage_base->b.reserved));
+	prt_printf(out, "online_reserved\t%llu\n",		percpu_u64_get(c->online_reserved));
+	prt_printf(out, "nr_inodes\t%llu\n",			bch2_fs_usage_read_one(c, &c->usage_base->b.nr_inodes));
+	percpu_up_read(&c->mark_lock);
+
+	prt_newline(out);
+	prt_printf(out, "freelist_wait\t%s\n",			c->freelist_wait.list.first ? "waiting" : "empty");
+	prt_printf(out, "open buckets allocated\t%i\n",		OPEN_BUCKETS_COUNT - c->open_buckets_nr_free);
+	prt_printf(out, "open buckets total\t%u\n",		OPEN_BUCKETS_COUNT);
+	prt_printf(out, "open_buckets_wait\t%s\n",		c->open_buckets_wait.list.first ? "waiting" : "empty");
+	prt_printf(out, "open_buckets_btree\t%u\n",		nr[BCH_DATA_btree]);
+	prt_printf(out, "open_buckets_user\t%u\n",		nr[BCH_DATA_user]);
+	prt_printf(out, "btree reserve cache\t%u\n",		c->btree_reserve_cache_nr);
+}
+
+void bch2_dev_alloc_debug_to_text(struct printbuf *out, struct bch_dev *ca)
+{
+	struct bch_fs *c = ca->fs;
+	struct bch_dev_usage stats = bch2_dev_usage_read(ca);
+	unsigned nr[BCH_DATA_NR];
+
+	memset(nr, 0, sizeof(nr));
+
+	for (unsigned i = 0; i < ARRAY_SIZE(c->open_buckets); i++)
+		nr[c->open_buckets[i].data_type]++;
+
+	printbuf_tabstops_reset(out);
+	printbuf_tabstop_push(out, 12);
+	printbuf_tabstop_push(out, 16);
+	printbuf_tabstop_push(out, 16);
+	printbuf_tabstop_push(out, 16);
+	printbuf_tabstop_push(out, 16);
+
+	bch2_dev_usage_to_text(out, &stats);
+
+	prt_newline(out);
+
+	prt_printf(out, "reserves:\n");
+	for (unsigned i = 0; i < BCH_WATERMARK_NR; i++)
+		prt_printf(out, "%s\t%llu\r\n", bch2_watermarks[i], bch2_dev_buckets_reserved(ca, i));
+
+	prt_newline(out);
+
+	printbuf_tabstops_reset(out);
+	printbuf_tabstop_push(out, 12);
+	printbuf_tabstop_push(out, 16);
+
+	prt_printf(out, "open buckets\t%i\r\n",	ca->nr_open_buckets);
+	prt_printf(out, "buckets to invalidate\t%llu\r\n",	should_invalidate_buckets(ca, stats));
+}
+
+void bch2_print_allocator_stuck(struct bch_fs *c)
+{
+	struct printbuf buf = PRINTBUF;
+
+	prt_printf(&buf, "Allocator stuck? Waited for 10 seconds\n");
+
+	prt_printf(&buf, "Allocator debug:\n");
+	printbuf_indent_add(&buf, 2);
+	bch2_fs_alloc_debug_to_text(&buf, c);
+	printbuf_indent_sub(&buf, 2);
+	prt_newline(&buf);
+
+	for_each_online_member(c, ca) {
+		prt_printf(&buf, "Dev %u:\n", ca->dev_idx);
+		printbuf_indent_add(&buf, 2);
+		bch2_dev_alloc_debug_to_text(&buf, ca);
+		printbuf_indent_sub(&buf, 2);
+		prt_newline(&buf);
+	}
+
+	prt_printf(&buf, "Copygc debug:\n");
+	printbuf_indent_add(&buf, 2);
+	bch2_copygc_wait_to_text(&buf, c);
+	printbuf_indent_sub(&buf, 2);
+	prt_newline(&buf);
+
+	prt_printf(&buf, "Journal debug:\n");
+	printbuf_indent_add(&buf, 2);
+	bch2_journal_debug_to_text(&buf, &c->journal);
+	printbuf_indent_sub(&buf, 2);
+
+	bch2_print_string_as_lines(KERN_ERR, buf.buf);
+	printbuf_exit(&buf);
+}
diff --git a/fs/bcachefs/alloc_foreground.h b/fs/bcachefs/alloc_foreground.h
index 7aaeec44c7..a42c9730d3 100644
--- a/fs/bcachefs/alloc_foreground.h
+++ b/fs/bcachefs/alloc_foreground.h
@@ -30,8 +30,14 @@ void bch2_dev_stripe_increment(struct bch_dev *, struct dev_stripe_state *);
 
 long bch2_bucket_alloc_new_fs(struct bch_dev *);
 
+static inline struct bch_dev *ob_dev(struct bch_fs *c, struct open_bucket *ob)
+{
+	return bch2_dev_have_ref(c, ob->dev);
+}
+
 struct open_bucket *bch2_bucket_alloc(struct bch_fs *, struct bch_dev *,
-				      enum bch_watermark, struct closure *);
+				      enum bch_watermark, enum bch_data_type,
+				      struct closure *);
 
 static inline void ob_push(struct bch_fs *c, struct open_buckets *obs,
 			   struct open_bucket *ob)
@@ -184,7 +190,7 @@ bch2_alloc_sectors_append_ptrs_inlined(struct bch_fs *c, struct write_point *wp,
 	wp->sectors_allocated	+= sectors;
 
 	open_bucket_for_each(c, &wp->ptrs, ob, i) {
-		struct bch_dev *ca = bch_dev_bkey_exists(c, ob->dev);
+		struct bch_dev *ca = ob_dev(c, ob);
 		struct bch_extent_ptr ptr = bch2_ob_ptr(c, ob);
 
 		ptr.cached = cached ||
@@ -221,4 +227,9 @@ void bch2_open_buckets_partial_to_text(struct printbuf *, struct bch_fs *);
 
 void bch2_write_points_to_text(struct printbuf *, struct bch_fs *);
 
+void bch2_fs_alloc_debug_to_text(struct printbuf *, struct bch_fs *);
+void bch2_dev_alloc_debug_to_text(struct printbuf *, struct bch_dev *);
+
+void bch2_print_allocator_stuck(struct bch_fs *);
+
 #endif /* _BCACHEFS_ALLOC_FOREGROUND_H */
diff --git a/fs/bcachefs/alloc_types.h b/fs/bcachefs/alloc_types.h
index b91b7a4610..9bbb28e90b 100644
--- a/fs/bcachefs/alloc_types.h
+++ b/fs/bcachefs/alloc_types.h
@@ -9,11 +9,18 @@
 #include "fifo.h"
 
 struct bucket_alloc_state {
+	enum {
+		BTREE_BITMAP_NO,
+		BTREE_BITMAP_YES,
+		BTREE_BITMAP_ANY,
+	}	btree_bitmap;
+
 	u64	buckets_seen;
 	u64	skipped_open;
 	u64	skipped_need_journal_commit;
 	u64	skipped_nocow;
 	u64	skipped_nouse;
+	u64	skipped_mi_btree_bitmap;
 };
 
 #define BCH_WATERMARKS()		\
@@ -22,7 +29,8 @@ struct bucket_alloc_state {
 	x(copygc)			\
 	x(btree)			\
 	x(btree_copygc)			\
-	x(reclaim)
+	x(reclaim)			\
+	x(interior_updates)
 
 enum bch_watermark {
 #define x(name)	BCH_WATERMARK_##name,
diff --git a/fs/bcachefs/backpointers.c b/fs/bcachefs/backpointers.c
index 569b97904d..6d8b1bc90b 100644
--- a/fs/bcachefs/backpointers.c
+++ b/fs/bcachefs/backpointers.c
@@ -8,6 +8,7 @@
 #include "btree_update.h"
 #include "btree_update_interior.h"
 #include "btree_write_buffer.h"
+#include "checksum.h"
 #include "error.h"
 
 #include <linux/mm.h>
@@ -22,6 +23,7 @@ static bool extent_matches_bp(struct bch_fs *c,
 	const union bch_extent_entry *entry;
 	struct extent_ptr_decoded p;
 
+	rcu_read_lock();
 	bkey_for_each_ptr_decode(k.k, ptrs, p, entry) {
 		struct bpos bucket2;
 		struct bch_backpointer bp2;
@@ -29,28 +31,46 @@ static bool extent_matches_bp(struct bch_fs *c,
 		if (p.ptr.cached)
 			continue;
 
-		bch2_extent_ptr_to_bp(c, btree_id, level, k, p,
-				      &bucket2, &bp2);
+		struct bch_dev *ca = bch2_dev_rcu(c, p.ptr.dev);
+		if (!ca)
+			continue;
+
+		bch2_extent_ptr_to_bp(c, ca, btree_id, level, k, p, entry, &bucket2, &bp2);
 		if (bpos_eq(bucket, bucket2) &&
-		    !memcmp(&bp, &bp2, sizeof(bp)))
+		    !memcmp(&bp, &bp2, sizeof(bp))) {
+			rcu_read_unlock();
 			return true;
+		}
 	}
+	rcu_read_unlock();
 
 	return false;
 }
 
 int bch2_backpointer_invalid(struct bch_fs *c, struct bkey_s_c k,
-			     enum bkey_invalid_flags flags,
+			     enum bch_validate_flags flags,
 			     struct printbuf *err)
 {
 	struct bkey_s_c_backpointer bp = bkey_s_c_to_backpointer(k);
-	struct bpos bucket = bp_pos_to_bucket(c, bp.k->p);
+
+	rcu_read_lock();
+	struct bch_dev *ca = bch2_dev_rcu(c, bp.k->p.inode);
+	if (!ca) {
+		/* these will be caught by fsck */
+		rcu_read_unlock();
+		return 0;
+	}
+
+	struct bpos bucket = bp_pos_to_bucket(ca, bp.k->p);
+	struct bpos bp_pos = bucket_pos_to_bp_noerror(ca, bucket, bp.v->bucket_offset);
+	rcu_read_unlock();
 	int ret = 0;
 
-	bkey_fsck_err_on(!bpos_eq(bp.k->p, bucket_pos_to_bp(c, bucket, bp.v->bucket_offset)),
+	bkey_fsck_err_on((bp.v->bucket_offset >> MAX_EXTENT_COMPRESS_RATIO_SHIFT) >= ca->mi.bucket_size ||
+			 !bpos_eq(bp.k->p, bp_pos),
 			 c, err,
-			 backpointer_pos_wrong,
-			 "backpointer at wrong pos");
+			 backpointer_bucket_offset_wrong,
+			 "backpointer bucket_offset wrong");
 fsck_err:
 	return ret;
 }
@@ -68,10 +88,16 @@ void bch2_backpointer_to_text(struct printbuf *out, const struct bch_backpointer
 
 void bch2_backpointer_k_to_text(struct printbuf *out, struct bch_fs *c, struct bkey_s_c k)
 {
-	if (bch2_dev_exists2(c, k.k->p.inode)) {
+	rcu_read_lock();
+	struct bch_dev *ca = bch2_dev_rcu(c, k.k->p.inode);
+	if (ca) {
+		struct bpos bucket = bp_pos_to_bucket(ca, k.k->p);
+		rcu_read_unlock();
 		prt_str(out, "bucket=");
-		bch2_bpos_to_text(out, bp_pos_to_bucket(c, k.k->p));
+		bch2_bpos_to_text(out, bucket);
 		prt_str(out, " ");
+	} else {
+		rcu_read_unlock();
 	}
 
 	bch2_backpointer_to_text(out, bkey_s_c_to_backpointer(k).v);
@@ -110,8 +136,7 @@ static noinline int backpointer_mod_err(struct btree_trans *trans,
 
 		bch_err(c, "%s", buf.buf);
 	} else if (c->curr_recovery_pass > BCH_RECOVERY_PASS_check_extents_to_backpointers) {
-		prt_printf(&buf, "backpointer not found when deleting");
-		prt_newline(&buf);
+		prt_printf(&buf, "backpointer not found when deleting\n");
 		printbuf_indent_add(&buf, 2);
 
 		prt_printf(&buf, "searching for ");
@@ -131,14 +156,14 @@ static noinline int backpointer_mod_err(struct btree_trans *trans,
 	printbuf_exit(&buf);
 
 	if (c->curr_recovery_pass > BCH_RECOVERY_PASS_check_extents_to_backpointers) {
-		bch2_inconsistent_error(c);
-		return -EIO;
+		return bch2_inconsistent_error(c) ? BCH_ERR_erofs_unfixed_errors : 0;
 	} else {
 		return 0;
 	}
 }
 
 int bch2_bucket_backpointer_mod_nowritebuffer(struct btree_trans *trans,
+				struct bch_dev *ca,
 				struct bpos bucket,
 				struct bch_backpointer bp,
 				struct bkey_s_c orig_k,
@@ -155,7 +180,7 @@ int bch2_bucket_backpointer_mod_nowritebuffer(struct btree_trans *trans,
 		return ret;
 
 	bkey_backpointer_init(&bp_k->k_i);
-	bp_k->k.p = bucket_pos_to_bp(trans->c, bucket, bp.bucket_offset);
+	bp_k->k.p = bucket_pos_to_bp(ca, bucket, bp.bucket_offset);
 	bp_k->v = bp;
 
 	if (!insert) {
@@ -165,9 +190,9 @@ int bch2_bucket_backpointer_mod_nowritebuffer(struct btree_trans *trans,
 
 	k = bch2_bkey_get_iter(trans, &bp_iter, BTREE_ID_backpointers,
 			       bp_k->k.p,
-			       BTREE_ITER_INTENT|
-			       BTREE_ITER_SLOTS|
-			       BTREE_ITER_WITH_UPDATES);
+			       BTREE_ITER_intent|
+			       BTREE_ITER_slots|
+			       BTREE_ITER_with_updates);
 	ret = bkey_err(k);
 	if (ret)
 		goto err;
@@ -191,13 +216,13 @@ err:
  * Find the next backpointer >= *bp_offset:
  */
 int bch2_get_next_backpointer(struct btree_trans *trans,
+			      struct bch_dev *ca,
 			      struct bpos bucket, int gen,
 			      struct bpos *bp_pos,
 			      struct bch_backpointer *bp,
 			      unsigned iter_flags)
 {
-	struct bch_fs *c = trans->c;
-	struct bpos bp_end_pos = bucket_pos_to_bp(c, bpos_nosnap_successor(bucket), 0);
+	struct bpos bp_end_pos = bucket_pos_to_bp(ca, bpos_nosnap_successor(bucket), 0);
 	struct btree_iter alloc_iter = { NULL }, bp_iter = { NULL };
 	struct bkey_s_c k;
 	int ret = 0;
@@ -207,7 +232,7 @@ int bch2_get_next_backpointer(struct btree_trans *trans,
 
 	if (gen >= 0) {
 		k = bch2_bkey_get_iter(trans, &alloc_iter, BTREE_ID_alloc,
-				       bucket, BTREE_ITER_CACHED|iter_flags);
+				       bucket, BTREE_ITER_cached|iter_flags);
 		ret = bkey_err(k);
 		if (ret)
 			goto out;
@@ -217,7 +242,7 @@ int bch2_get_next_backpointer(struct btree_trans *trans,
 			goto done;
 	}
 
-	*bp_pos = bpos_max(*bp_pos, bucket_pos_to_bp(c, bucket, 0));
+	*bp_pos = bpos_max(*bp_pos, bucket_pos_to_bp(ca, bucket, 0));
 
 	for_each_btree_key_norestart(trans, bp_iter, BTREE_ID_backpointers,
 				     *bp_pos, iter_flags, k, ret) {
@@ -243,7 +268,6 @@ static void backpointer_not_found(struct btree_trans *trans,
 {
 	struct bch_fs *c = trans->c;
 	struct printbuf buf = PRINTBUF;
-	struct bpos bucket = bp_pos_to_bucket(c, bp_pos);
 
 	/*
 	 * If we're using the btree write buffer, the backpointer we were
@@ -253,6 +277,10 @@ static void backpointer_not_found(struct btree_trans *trans,
 	if (likely(!bch2_backpointers_no_use_write_buffer))
 		return;
 
+	struct bpos bucket;
+	if (!bp_pos_to_bucket_nodev(c, bp_pos, &bucket))
+		return;
+
 	prt_printf(&buf, "backpointer doesn't match %s it points to:\n  ",
 		   bp.level ? "btree node" : "extent");
 	prt_printf(&buf, "bucket: ");
@@ -282,15 +310,17 @@ struct bkey_s_c bch2_backpointer_get_key(struct btree_trans *trans,
 {
 	if (likely(!bp.level)) {
 		struct bch_fs *c = trans->c;
-		struct bpos bucket = bp_pos_to_bucket(c, bp_pos);
-		struct bkey_s_c k;
+
+		struct bpos bucket;
+		if (!bp_pos_to_bucket_nodev(c, bp_pos, &bucket))
+			return bkey_s_c_err(-EIO);
 
 		bch2_trans_node_iter_init(trans, iter,
 					  bp.btree_id,
 					  bp.pos,
 					  0, 0,
 					  iter_flags);
-		k = bch2_btree_iter_peek_slot(iter);
+		struct bkey_s_c k = bch2_btree_iter_peek_slot(iter);
 		if (bkey_err(k)) {
 			bch2_trans_iter_exit(trans, iter);
 			return k;
@@ -319,18 +349,20 @@ struct btree *bch2_backpointer_get_node(struct btree_trans *trans,
 					struct bch_backpointer bp)
 {
 	struct bch_fs *c = trans->c;
-	struct bpos bucket = bp_pos_to_bucket(c, bp_pos);
-	struct btree *b;
 
 	BUG_ON(!bp.level);
 
+	struct bpos bucket;
+	if (!bp_pos_to_bucket_nodev(c, bp_pos, &bucket))
+		return ERR_PTR(-EIO);
+
 	bch2_trans_node_iter_init(trans, iter,
 				  bp.btree_id,
 				  bp.pos,
 				  0,
 				  bp.level - 1,
 				  0);
-	b = bch2_btree_iter_peek_node(iter);
+	struct btree *b = bch2_btree_iter_peek_node(iter);
 	if (IS_ERR_OR_NULL(b))
 		goto err;
 
@@ -361,16 +393,16 @@ static int bch2_check_btree_backpointer(struct btree_trans *trans, struct btree_
 	struct printbuf buf = PRINTBUF;
 	int ret = 0;
 
-	if (fsck_err_on(!bch2_dev_exists2(c, k.k->p.inode), c,
-			backpointer_to_missing_device,
-			"backpointer for missing device:\n%s",
-			(bch2_bkey_val_to_text(&buf, c, k), buf.buf))) {
-		ret = bch2_btree_delete_at(trans, bp_iter, 0);
+	struct bpos bucket;
+	if (!bp_pos_to_bucket_nodev_noerror(c, k.k->p, &bucket)) {
+		if (fsck_err(c, backpointer_to_missing_device,
+			     "backpointer for missing device:\n%s",
+			     (bch2_bkey_val_to_text(&buf, c, k), buf.buf)))
+			ret = bch2_btree_delete_at(trans, bp_iter, 0);
 		goto out;
 	}
 
-	alloc_k = bch2_bkey_get_iter(trans, &alloc_iter, BTREE_ID_alloc,
-				     bp_pos_to_bucket(c, k.k->p), 0);
+	alloc_k = bch2_bkey_get_iter(trans, &alloc_iter, BTREE_ID_alloc, bucket, 0);
 	ret = bkey_err(alloc_k);
 	if (ret)
 		goto out;
@@ -379,7 +411,7 @@ static int bch2_check_btree_backpointer(struct btree_trans *trans, struct btree_
 			backpointer_to_missing_alloc,
 			"backpointer for nonexistent alloc key: %llu:%llu:0\n%s",
 			alloc_iter.pos.inode, alloc_iter.pos.offset,
-			(bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf))) {
+			(bch2_bkey_val_to_text(&buf, c, k), buf.buf))) {
 		ret = bch2_btree_delete_at(trans, bp_iter, 0);
 		goto out;
 	}
@@ -402,19 +434,90 @@ int bch2_check_btree_backpointers(struct bch_fs *c)
 	return ret;
 }
 
-static inline bool bkey_and_val_eq(struct bkey_s_c l, struct bkey_s_c r)
-{
-	return bpos_eq(l.k->p, r.k->p) &&
-		bkey_bytes(l.k) == bkey_bytes(r.k) &&
-		!memcmp(l.v, r.v, bkey_val_bytes(l.k));
-}
-
 struct extents_to_bp_state {
 	struct bpos	bucket_start;
 	struct bpos	bucket_end;
 	struct bkey_buf last_flushed;
 };
 
+static int drop_dev_and_update(struct btree_trans *trans, enum btree_id btree,
+			       struct bkey_s_c extent, unsigned dev)
+{
+	struct bkey_i *n = bch2_bkey_make_mut_noupdate(trans, extent);
+	int ret = PTR_ERR_OR_ZERO(n);
+	if (ret)
+		return ret;
+
+	bch2_bkey_drop_device(bkey_i_to_s(n), dev);
+	return bch2_btree_insert_trans(trans, btree, n, 0);
+}
+
+static int check_extent_checksum(struct btree_trans *trans,
+				 enum btree_id btree, struct bkey_s_c extent,
+				 enum btree_id o_btree, struct bkey_s_c extent2, unsigned dev)
+{
+	struct bch_fs *c = trans->c;
+	struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(extent);
+	const union bch_extent_entry *entry;
+	struct extent_ptr_decoded p;
+	struct printbuf buf = PRINTBUF;
+	void *data_buf = NULL;
+	struct bio *bio = NULL;
+	size_t bytes;
+	int ret = 0;
+
+	if (bkey_is_btree_ptr(extent.k))
+		return false;
+
+	bkey_for_each_ptr_decode(extent.k, ptrs, p, entry)
+		if (p.ptr.dev == dev)
+			goto found;
+	BUG();
+found:
+	if (!p.crc.csum_type)
+		return false;
+
+	bytes = p.crc.compressed_size << 9;
+
+	struct bch_dev *ca = bch2_dev_get_ioref(c, dev, READ);
+	if (!ca)
+		return false;
+
+	data_buf = kvmalloc(bytes, GFP_KERNEL);
+	if (!data_buf) {
+		ret = -ENOMEM;
+		goto err;
+	}
+
+	bio = bio_alloc(ca->disk_sb.bdev, buf_pages(data_buf, bytes), REQ_OP_READ, GFP_KERNEL);
+	bio->bi_iter.bi_sector = p.ptr.offset;
+	bch2_bio_map(bio, data_buf, bytes);
+	ret = submit_bio_wait(bio);
+	if (ret)
+		goto err;
+
+	prt_str(&buf, "extents pointing to same space, but first extent checksum bad:");
+	prt_printf(&buf, "\n  %s ", bch2_btree_id_str(btree));
+	bch2_bkey_val_to_text(&buf, c, extent);
+	prt_printf(&buf, "\n  %s ", bch2_btree_id_str(o_btree));
+	bch2_bkey_val_to_text(&buf, c, extent2);
+
+	struct nonce nonce = extent_nonce(extent.k->version, p.crc);
+	struct bch_csum csum = bch2_checksum(c, p.crc.csum_type, nonce, data_buf, bytes);
+	if (fsck_err_on(bch2_crc_cmp(csum, p.crc.csum),
+			c, dup_backpointer_to_bad_csum_extent,
+			"%s", buf.buf))
+		ret = drop_dev_and_update(trans, btree, extent, dev) ?: 1;
+fsck_err:
+err:
+	if (bio)
+		bio_put(bio);
+	kvfree(data_buf);
+	percpu_ref_put(&ca->io_ref);
+	printbuf_exit(&buf);
+	return ret;
+}
+
 static int check_bp_exists(struct btree_trans *trans,
 			   struct extents_to_bp_state *s,
 			   struct bpos bucket,
@@ -422,23 +525,29 @@ static int check_bp_exists(struct btree_trans *trans,
 			   struct bkey_s_c orig_k)
 {
 	struct bch_fs *c = trans->c;
-	struct btree_iter bp_iter = { NULL };
+	struct btree_iter bp_iter = {};
+	struct btree_iter other_extent_iter = {};
 	struct printbuf buf = PRINTBUF;
 	struct bkey_s_c bp_k;
-	struct bkey_buf tmp;
-	int ret;
+	int ret = 0;
 
-	bch2_bkey_buf_init(&tmp);
+	struct bch_dev *ca = bch2_dev_bucket_tryget(c, bucket);
+	if (!ca) {
+		prt_str(&buf, "extent for nonexistent device:bucket ");
+		bch2_bpos_to_text(&buf, bucket);
+		prt_str(&buf, "\n  ");
+		bch2_bkey_val_to_text(&buf, c, orig_k);
+		bch_err(c, "%s", buf.buf);
+		ret = -BCH_ERR_fsck_repair_unimplemented;
+		goto err;
+	}
 
 	if (bpos_lt(bucket, s->bucket_start) ||
 	    bpos_gt(bucket, s->bucket_end))
-		return 0;
-
-	if (!bch2_dev_bucket_exists(c, bucket))
-		goto missing;
+		goto out;
 
 	bp_k = bch2_bkey_get_iter(trans, &bp_iter, BTREE_ID_backpointers,
-				  bucket_pos_to_bp(c, bucket, bp.bucket_offset),
+				  bucket_pos_to_bp(ca, bucket, bp.bucket_offset),
 				  0);
 	ret = bkey_err(bp_k);
 	if (ret)
@@ -446,41 +555,100 @@ static int check_bp_exists(struct btree_trans *trans,
 
 	if (bp_k.k->type != KEY_TYPE_backpointer ||
 	    memcmp(bkey_s_c_to_backpointer(bp_k).v, &bp, sizeof(bp))) {
-		bch2_bkey_buf_reassemble(&tmp, c, orig_k);
-
-		if (!bkey_and_val_eq(orig_k, bkey_i_to_s_c(s->last_flushed.k))) {
-			if (bp.level) {
-				bch2_trans_unlock(trans);
-				bch2_btree_interior_updates_flush(c);
-			}
-
-			ret = bch2_btree_write_buffer_flush_sync(trans);
-			if (ret)
-				goto err;
+		ret = bch2_btree_write_buffer_maybe_flush(trans, orig_k, &s->last_flushed);
+		if (ret)
+			goto err;
 
-			bch2_bkey_buf_copy(&s->last_flushed, c, tmp.k);
-			ret = -BCH_ERR_transaction_restart_write_buffer_flush;
-			goto out;
-		}
-		goto missing;
+		goto check_existing_bp;
 	}
 out:
 err:
 fsck_err:
+	bch2_trans_iter_exit(trans, &other_extent_iter);
 	bch2_trans_iter_exit(trans, &bp_iter);
-	bch2_bkey_buf_exit(&tmp, c);
+	bch2_dev_put(ca);
 	printbuf_exit(&buf);
 	return ret;
+check_existing_bp:
+	/* Do we have a backpointer for a different extent? */
+	if (bp_k.k->type != KEY_TYPE_backpointer)
+		goto missing;
+
+	struct bch_backpointer other_bp = *bkey_s_c_to_backpointer(bp_k).v;
+
+	struct bkey_s_c other_extent =
+		bch2_backpointer_get_key(trans, &other_extent_iter, bp_k.k->p, other_bp, 0);
+	ret = bkey_err(other_extent);
+	if (ret == -BCH_ERR_backpointer_to_overwritten_btree_node)
+		ret = 0;
+	if (ret)
+		goto err;
+
+	if (!other_extent.k)
+		goto missing;
+
+	if (bch2_extents_match(orig_k, other_extent)) {
+		printbuf_reset(&buf);
+		prt_printf(&buf, "duplicate versions of same extent, deleting smaller\n  ");
+		bch2_bkey_val_to_text(&buf, c, orig_k);
+		prt_str(&buf, "\n  ");
+		bch2_bkey_val_to_text(&buf, c, other_extent);
+		bch_err(c, "%s", buf.buf);
+
+		if (other_extent.k->size <= orig_k.k->size) {
+			ret = drop_dev_and_update(trans, other_bp.btree_id, other_extent, bucket.inode);
+			if (ret)
+				goto err;
+			goto out;
+		} else {
+			ret = drop_dev_and_update(trans, bp.btree_id, orig_k, bucket.inode);
+			if (ret)
+				goto err;
+			goto missing;
+		}
+	}
+
+	ret = check_extent_checksum(trans, other_bp.btree_id, other_extent, bp.btree_id, orig_k, bucket.inode);
+	if (ret < 0)
+		goto err;
+	if (ret) {
+		ret = 0;
+		goto missing;
+	}
+
+	ret = check_extent_checksum(trans, bp.btree_id, orig_k, other_bp.btree_id, other_extent, bucket.inode);
+	if (ret < 0)
+		goto err;
+	if (ret) {
+		ret = 0;
+		goto out;
+	}
+
+	printbuf_reset(&buf);
+	prt_printf(&buf, "duplicate extents pointing to same space on dev %llu\n  ", bucket.inode);
+	bch2_bkey_val_to_text(&buf, c, orig_k);
+	prt_str(&buf, "\n  ");
+	bch2_bkey_val_to_text(&buf, c, other_extent);
+	bch_err(c, "%s", buf.buf);
+	ret = -BCH_ERR_fsck_repair_unimplemented;
+	goto err;
 missing:
+	printbuf_reset(&buf);
 	prt_printf(&buf, "missing backpointer for btree=%s l=%u ",
 	       bch2_btree_id_str(bp.btree_id), bp.level);
 	bch2_bkey_val_to_text(&buf, c, orig_k);
-	prt_printf(&buf, "\nbp pos ");
-	bch2_bpos_to_text(&buf, bp_iter.pos);
+	prt_printf(&buf, "\n  got:   ");
+	bch2_bkey_val_to_text(&buf, c, bp_k);
+
+	struct bkey_i_backpointer n_bp_k;
+	bkey_backpointer_init(&n_bp_k.k_i);
+	n_bp_k.k.p = bucket_pos_to_bp(ca, bucket, bp.bucket_offset);
+	n_bp_k.v = bp;
+	prt_printf(&buf, "\n  want:  ");
+	bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&n_bp_k.k_i));
 
-	if (c->opts.reconstruct_alloc ||
-	    fsck_err(c, ptr_to_missing_backpointer, "%s", buf.buf))
-		ret = bch2_bucket_backpointer_mod(trans, bucket, bp, orig_k, true);
+	if (fsck_err(c, ptr_to_missing_backpointer, "%s", buf.buf))
+		ret = bch2_bucket_backpointer_mod(trans, ca, bucket, bp, orig_k, true);
 
 	goto out;
 }
@@ -498,14 +666,20 @@ static int check_extent_to_backpointers(struct btree_trans *trans,
 
 	ptrs = bch2_bkey_ptrs_c(k);
 	bkey_for_each_ptr_decode(k.k, ptrs, p, entry) {
-		struct bpos bucket_pos;
+		struct bpos bucket_pos = POS_MIN;
 		struct bch_backpointer bp;
 
 		if (p.ptr.cached)
 			continue;
 
-		bch2_extent_ptr_to_bp(c, btree, level,
-				      k, p, &bucket_pos, &bp);
+		rcu_read_lock();
+		struct bch_dev *ca = bch2_dev_rcu(c, p.ptr.dev);
+		if (ca)
+			bch2_extent_ptr_to_bp(c, ca, btree, level, k, p, entry, &bucket_pos, &bp);
+		rcu_read_unlock();
+
+		if (!ca)
+			continue;
 
 		ret = check_bp_exists(trans, s, bucket_pos, bp, k);
 		if (ret)
@@ -555,60 +729,63 @@ static inline struct bbpos bp_to_bbpos(struct bch_backpointer bp)
 	};
 }
 
-static size_t btree_nodes_fit_in_ram(struct bch_fs *c)
+static u64 mem_may_pin_bytes(struct bch_fs *c)
 {
 	struct sysinfo i;
-	u64 mem_bytes;
-
 	si_meminfo(&i);
-	mem_bytes = i.totalram * i.mem_unit;
-	return div_u64(mem_bytes >> 1, c->opts.btree_node_size);
+
+	u64 mem_bytes = i.totalram * i.mem_unit;
+	return div_u64(mem_bytes * c->opts.fsck_memory_usage_percent, 100);
+}
+
+static size_t btree_nodes_fit_in_ram(struct bch_fs *c)
+{
+	return div_u64(mem_may_pin_bytes(c), c->opts.btree_node_size);
 }
 
 static int bch2_get_btree_in_memory_pos(struct btree_trans *trans,
-					unsigned btree_leaf_mask,
-					unsigned btree_interior_mask,
+					u64 btree_leaf_mask,
+					u64 btree_interior_mask,
 					struct bbpos start, struct bbpos *end)
 {
-	struct btree_iter iter;
-	struct bkey_s_c k;
-	size_t btree_nodes = btree_nodes_fit_in_ram(trans->c);
-	enum btree_id btree;
+	struct bch_fs *c = trans->c;
+	s64 mem_may_pin = mem_may_pin_bytes(c);
 	int ret = 0;
 
-	for (btree = start.btree; btree < BTREE_ID_NR && !ret; btree++) {
-		unsigned depth = ((1U << btree) & btree_leaf_mask) ? 1 : 2;
+	btree_interior_mask |= btree_leaf_mask;
+
+	c->btree_cache.pinned_nodes_leaf_mask		= btree_leaf_mask;
+	c->btree_cache.pinned_nodes_interior_mask	= btree_interior_mask;
+	c->btree_cache.pinned_nodes_start		= start;
+	c->btree_cache.pinned_nodes_end			= *end = BBPOS_MAX;
+
+	for (enum btree_id btree = start.btree;
+	     btree < BTREE_ID_NR && !ret;
+	     btree++) {
+		unsigned depth = ((1U << btree) & btree_leaf_mask) ? 0 : 1;
+		struct btree_iter iter;
+		struct btree *b;
 
 		if (!((1U << btree) & btree_leaf_mask) &&
 		    !((1U << btree) & btree_interior_mask))
 			continue;
 
-		bch2_trans_node_iter_init(trans, &iter, btree,
-					  btree == start.btree ? start.pos : POS_MIN,
-					  0, depth, 0);
-		/*
-		 * for_each_btree_key_contineu() doesn't check the return value
-		 * from bch2_btree_iter_advance(), which is needed when
-		 * iterating over interior nodes where we'll see keys at
-		 * SPOS_MAX:
-		 */
-		do {
-			k = __bch2_btree_iter_peek_and_restart(trans, &iter, 0);
-			ret = bkey_err(k);
-			if (!k.k || ret)
-				break;
-
-			--btree_nodes;
-			if (!btree_nodes) {
-				*end = BBPOS(btree, k.k->p);
+		bch2_trans_begin(trans);
+
+		__for_each_btree_node(trans, iter, btree,
+				      btree == start.btree ? start.pos : POS_MIN,
+				      0, depth, BTREE_ITER_prefetch, b, ret) {
+			mem_may_pin -= btree_buf_bytes(b);
+			if (mem_may_pin <= 0) {
+				c->btree_cache.pinned_nodes_end = *end =
+					BBPOS(btree, b->key.k.p);
 				bch2_trans_iter_exit(trans, &iter);
 				return 0;
 			}
-		} while (bch2_btree_iter_advance(&iter));
+		}
 		bch2_trans_iter_exit(trans, &iter);
 	}
 
-	*end = BBPOS_MAX;
 	return ret;
 }
 
@@ -631,31 +808,13 @@ static int bch2_check_extents_to_backpointers_pass(struct btree_trans *trans,
 
 		while (level >= depth) {
 			struct btree_iter iter;
-			bch2_trans_node_iter_init(trans, &iter, btree_id, POS_MIN, 0,
-						  level,
-						  BTREE_ITER_PREFETCH);
-			while (1) {
-				bch2_trans_begin(trans);
-
-				struct bkey_s_c k = bch2_btree_iter_peek(&iter);
-				if (!k.k)
-					break;
-				ret = bkey_err(k) ?:
-					check_extent_to_backpointers(trans, s, btree_id, level, k) ?:
-					bch2_trans_commit(trans, NULL, NULL,
-							  BCH_TRANS_COMMIT_no_enospc);
-				if (bch2_err_matches(ret, BCH_ERR_transaction_restart)) {
-					ret = 0;
-					continue;
-				}
-				if (ret)
-					break;
-				if (bpos_eq(iter.pos, SPOS_MAX))
-					break;
-				bch2_btree_iter_advance(&iter);
-			}
-			bch2_trans_iter_exit(trans, &iter);
+			bch2_trans_node_iter_init(trans, &iter, btree_id, POS_MIN, 0, level,
+						  BTREE_ITER_prefetch);
 
+			ret = for_each_btree_key_continue(trans, iter, 0, k, ({
+				check_extent_to_backpointers(trans, s, btree_id, level, k) ?:
+				bch2_trans_commit(trans, NULL, NULL, BCH_TRANS_COMMIT_no_enospc);
+			}));
 			if (ret)
 				return ret;
 
@@ -666,62 +825,6 @@ static int bch2_check_extents_to_backpointers_pass(struct btree_trans *trans,
 	return 0;
 }
 
-static struct bpos bucket_pos_to_bp_safe(const struct bch_fs *c,
-					 struct bpos bucket)
-{
-	return bch2_dev_exists2(c, bucket.inode)
-		? bucket_pos_to_bp(c, bucket, 0)
-		: bucket;
-}
-
-static int bch2_get_alloc_in_memory_pos(struct btree_trans *trans,
-					struct bpos start, struct bpos *end)
-{
-	struct btree_iter alloc_iter;
-	struct btree_iter bp_iter;
-	struct bkey_s_c alloc_k, bp_k;
-	size_t btree_nodes = btree_nodes_fit_in_ram(trans->c);
-	bool alloc_end = false, bp_end = false;
-	int ret = 0;
-
-	bch2_trans_node_iter_init(trans, &alloc_iter, BTREE_ID_alloc,
-				  start, 0, 1, 0);
-	bch2_trans_node_iter_init(trans, &bp_iter, BTREE_ID_backpointers,
-				  bucket_pos_to_bp_safe(trans->c, start), 0, 1, 0);
-	while (1) {
-		alloc_k = !alloc_end
-			? __bch2_btree_iter_peek_and_restart(trans, &alloc_iter, 0)
-			: bkey_s_c_null;
-		bp_k = !bp_end
-			? __bch2_btree_iter_peek_and_restart(trans, &bp_iter, 0)
-			: bkey_s_c_null;
-
-		ret = bkey_err(alloc_k) ?: bkey_err(bp_k);
-		if ((!alloc_k.k && !bp_k.k) || ret) {
-			*end = SPOS_MAX;
-			break;
-		}
-
-		--btree_nodes;
-		if (!btree_nodes) {
-			*end = alloc_k.k ? alloc_k.k->p : SPOS_MAX;
-			break;
-		}
-
-		if (bpos_lt(alloc_iter.pos, SPOS_MAX) &&
-		    bpos_lt(bucket_pos_to_bp_safe(trans->c, alloc_iter.pos), bp_iter.pos)) {
-			if (!bch2_btree_iter_advance(&alloc_iter))
-				alloc_end = true;
-		} else {
-			if (!bch2_btree_iter_advance(&bp_iter))
-				bp_end = true;
-		}
-	}
-	bch2_trans_iter_exit(trans, &bp_iter);
-	bch2_trans_iter_exit(trans, &alloc_iter);
-	return ret;
-}
-
 int bch2_check_extents_to_backpointers(struct bch_fs *c)
 {
 	struct btree_trans *trans = bch2_trans_get(c);
@@ -732,10 +835,16 @@ int bch2_check_extents_to_backpointers(struct bch_fs *c)
 	bkey_init(&s.last_flushed.k->k);
 
 	while (1) {
-		ret = bch2_get_alloc_in_memory_pos(trans, s.bucket_start, &s.bucket_end);
+		struct bbpos end;
+		ret = bch2_get_btree_in_memory_pos(trans,
+				BIT_ULL(BTREE_ID_backpointers),
+				BIT_ULL(BTREE_ID_backpointers),
+				BBPOS(BTREE_ID_backpointers, s.bucket_start), &end);
 		if (ret)
 			break;
 
+		s.bucket_end = end.pos;
+
 		if ( bpos_eq(s.bucket_start, POS_MIN) &&
 		    !bpos_eq(s.bucket_end, SPOS_MAX))
 			bch_verbose(c, "%s(): alloc info does not fit in ram, running in multiple passes with %zu nodes per pass",
@@ -763,6 +872,9 @@ int bch2_check_extents_to_backpointers(struct bch_fs *c)
 	bch2_trans_put(trans);
 	bch2_bkey_buf_exit(&s.last_flushed, c);
 
+	c->btree_cache.pinned_nodes_leaf_mask = 0;
+	c->btree_cache.pinned_nodes_interior_mask = 0;
+
 	bch_err_fn(c, ret);
 	return ret;
 }
@@ -771,7 +883,7 @@ static int check_one_backpointer(struct btree_trans *trans,
 				 struct bbpos start,
 				 struct bbpos end,
 				 struct bkey_s_c_backpointer bp,
-				 struct bpos *last_flushed_pos)
+				 struct bkey_buf *last_flushed)
 {
 	struct bch_fs *c = trans->c;
 	struct btree_iter iter;
@@ -791,20 +903,18 @@ static int check_one_backpointer(struct btree_trans *trans,
 	if (ret)
 		return ret;
 
-	if (!k.k && !bpos_eq(*last_flushed_pos, bp.k->p)) {
-		*last_flushed_pos = bp.k->p;
-		ret = bch2_btree_write_buffer_flush_sync(trans) ?:
-			-BCH_ERR_transaction_restart_write_buffer_flush;
-		goto out;
-	}
+	if (!k.k) {
+		ret = bch2_btree_write_buffer_maybe_flush(trans, bp.s_c, last_flushed);
+		if (ret)
+			goto out;
 
-	if (fsck_err_on(!k.k, c,
-			backpointer_to_missing_ptr,
-			"backpointer for missing %s\n  %s",
-			bp.v->level ? "btree node" : "extent",
-			(bch2_bkey_val_to_text(&buf, c, bp.s_c), buf.buf))) {
-		ret = bch2_btree_delete_at_buffered(trans, BTREE_ID_backpointers, bp.k->p);
-		goto out;
+		if (fsck_err(c, backpointer_to_missing_ptr,
+			     "backpointer for missing %s\n  %s",
+			     bp.v->level ? "btree node" : "extent",
+			     (bch2_bkey_val_to_text(&buf, c, bp.s_c), buf.buf))) {
+			ret = bch2_btree_delete_at_buffered(trans, BTREE_ID_backpointers, bp.k->p);
+			goto out;
+		}
 	}
 out:
 fsck_err:
@@ -817,14 +927,20 @@ static int bch2_check_backpointers_to_extents_pass(struct btree_trans *trans,
 						   struct bbpos start,
 						   struct bbpos end)
 {
-	struct bpos last_flushed_pos = SPOS_MAX;
+	struct bkey_buf last_flushed;
 
-	return for_each_btree_key_commit(trans, iter, BTREE_ID_backpointers,
-				  POS_MIN, BTREE_ITER_PREFETCH, k,
+	bch2_bkey_buf_init(&last_flushed);
+	bkey_init(&last_flushed.k->k);
+
+	int ret = for_each_btree_key_commit(trans, iter, BTREE_ID_backpointers,
+				  POS_MIN, BTREE_ITER_prefetch, k,
 				  NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
 		check_one_backpointer(trans, start, end,
 				      bkey_s_c_to_backpointer(k),
-				      &last_flushed_pos));
+				      &last_flushed));
+
+	bch2_bkey_buf_exit(&last_flushed, trans->c);
+	return ret;
 }
 
 int bch2_check_backpointers_to_extents(struct bch_fs *c)
@@ -868,6 +984,9 @@ int bch2_check_backpointers_to_extents(struct bch_fs *c)
 	}
 	bch2_trans_put(trans);
 
+	c->btree_cache.pinned_nodes_leaf_mask = 0;
+	c->btree_cache.pinned_nodes_interior_mask = 0;
+
 	bch_err_fn(c, ret);
 	return ret;
 }
diff --git a/fs/bcachefs/backpointers.h b/fs/bcachefs/backpointers.h
index 327365a9fe..6021de1c5e 100644
--- a/fs/bcachefs/backpointers.h
+++ b/fs/bcachefs/backpointers.h
@@ -6,6 +6,7 @@
 #include "btree_iter.h"
 #include "btree_update.h"
 #include "buckets.h"
+#include "error.h"
 #include "super.h"
 
 static inline u64 swab40(u64 x)
@@ -18,7 +19,7 @@ static inline u64 swab40(u64 x)
 }
 
 int bch2_backpointer_invalid(struct bch_fs *, struct bkey_s_c k,
-			     enum bkey_invalid_flags, struct printbuf *);
+			     enum bch_validate_flags, struct printbuf *);
 void bch2_backpointer_to_text(struct printbuf *, const struct bch_backpointer *);
 void bch2_backpointer_k_to_text(struct printbuf *, struct bch_fs *, struct bkey_s_c);
 void bch2_backpointer_swab(struct bkey_s);
@@ -36,50 +37,67 @@ void bch2_backpointer_swab(struct bkey_s);
  * Convert from pos in backpointer btree to pos of corresponding bucket in alloc
  * btree:
  */
-static inline struct bpos bp_pos_to_bucket(const struct bch_fs *c,
-					   struct bpos bp_pos)
+static inline struct bpos bp_pos_to_bucket(const struct bch_dev *ca, struct bpos bp_pos)
 {
-	struct bch_dev *ca = bch_dev_bkey_exists(c, bp_pos.inode);
 	u64 bucket_sector = bp_pos.offset >> MAX_EXTENT_COMPRESS_RATIO_SHIFT;
 
 	return POS(bp_pos.inode, sector_to_bucket(ca, bucket_sector));
 }
 
+static inline bool bp_pos_to_bucket_nodev_noerror(struct bch_fs *c, struct bpos bp_pos, struct bpos *bucket)
+{
+	rcu_read_lock();
+	struct bch_dev *ca = bch2_dev_rcu(c, bp_pos.inode);
+	if (ca)
+		*bucket = bp_pos_to_bucket(ca, bp_pos);
+	rcu_read_unlock();
+	return ca != NULL;
+}
+
+static inline bool bp_pos_to_bucket_nodev(struct bch_fs *c, struct bpos bp_pos, struct bpos *bucket)
+{
+	return !bch2_fs_inconsistent_on(!bp_pos_to_bucket_nodev_noerror(c, bp_pos, bucket),
+					c, "backpointer for missing device %llu", bp_pos.inode);
+}
+
+static inline struct bpos bucket_pos_to_bp_noerror(const struct bch_dev *ca,
+						   struct bpos bucket,
+						   u64 bucket_offset)
+{
+	return POS(bucket.inode,
+		   (bucket_to_sector(ca, bucket.offset) <<
+		    MAX_EXTENT_COMPRESS_RATIO_SHIFT) + bucket_offset);
+}
+
 /*
  * Convert from pos in alloc btree + bucket offset to pos in backpointer btree:
  */
-static inline struct bpos bucket_pos_to_bp(const struct bch_fs *c,
+static inline struct bpos bucket_pos_to_bp(const struct bch_dev *ca,
 					   struct bpos bucket,
 					   u64 bucket_offset)
 {
-	struct bch_dev *ca = bch_dev_bkey_exists(c, bucket.inode);
-	struct bpos ret;
-
-	ret = POS(bucket.inode,
-		  (bucket_to_sector(ca, bucket.offset) <<
-		   MAX_EXTENT_COMPRESS_RATIO_SHIFT) + bucket_offset);
-
-	EBUG_ON(!bkey_eq(bucket, bp_pos_to_bucket(c, ret)));
-
+	struct bpos ret = bucket_pos_to_bp_noerror(ca, bucket, bucket_offset);
+	EBUG_ON(!bkey_eq(bucket, bp_pos_to_bucket(ca, ret)));
 	return ret;
 }
 
-int bch2_bucket_backpointer_mod_nowritebuffer(struct btree_trans *, struct bpos bucket,
-				struct bch_backpointer, struct bkey_s_c, bool);
+int bch2_bucket_backpointer_mod_nowritebuffer(struct btree_trans *, struct bch_dev *,
+				struct bpos bucket, struct bch_backpointer, struct bkey_s_c, bool);
 
 static inline int bch2_bucket_backpointer_mod(struct btree_trans *trans,
+				struct bch_dev *ca,
 				struct bpos bucket,
 				struct bch_backpointer bp,
 				struct bkey_s_c orig_k,
 				bool insert)
 {
 	if (unlikely(bch2_backpointers_no_use_write_buffer))
-		return bch2_bucket_backpointer_mod_nowritebuffer(trans, bucket, bp, orig_k, insert);
+		return bch2_bucket_backpointer_mod_nowritebuffer(trans, ca, bucket, bp, orig_k, insert);
 
 	struct bkey_i_backpointer bp_k;
 
 	bkey_backpointer_init(&bp_k.k_i);
-	bp_k.k.p = bucket_pos_to_bp(trans->c, bucket, bp.bucket_offset);
+	bp_k.k.p = bucket_pos_to_bp(ca, bucket, bp.bucket_offset);
 	bp_k.v = bp;
 
 	if (!insert) {
@@ -90,24 +108,44 @@ static inline int bch2_bucket_backpointer_mod(struct btree_trans *trans,
 	return bch2_trans_update_buffered(trans, BTREE_ID_backpointers, &bp_k.k_i);
 }
 
-static inline enum bch_data_type bkey_ptr_data_type(enum btree_id btree_id, unsigned level,
-						    struct bkey_s_c k, struct extent_ptr_decoded p)
+static inline enum bch_data_type bch2_bkey_ptr_data_type(struct bkey_s_c k,
+							 struct extent_ptr_decoded p,
+							 const union bch_extent_entry *entry)
 {
-	return  level		? BCH_DATA_btree :
-		p.has_ec	? BCH_DATA_stripe :
-				  BCH_DATA_user;
+	switch (k.k->type) {
+	case KEY_TYPE_btree_ptr:
+	case KEY_TYPE_btree_ptr_v2:
+		return BCH_DATA_btree;
+	case KEY_TYPE_extent:
+	case KEY_TYPE_reflink_v:
+		return p.has_ec ? BCH_DATA_stripe : BCH_DATA_user;
+	case KEY_TYPE_stripe: {
+		const struct bch_extent_ptr *ptr = &entry->ptr;
+		struct bkey_s_c_stripe s = bkey_s_c_to_stripe(k);
+
+		BUG_ON(ptr < s.v->ptrs ||
+		       ptr >= s.v->ptrs + s.v->nr_blocks);
+
+		return ptr >= s.v->ptrs + s.v->nr_blocks - s.v->nr_redundant
+			? BCH_DATA_parity
+			: BCH_DATA_user;
+	}
+	default:
+		BUG();
+	}
 }
 
-static inline void bch2_extent_ptr_to_bp(struct bch_fs *c,
+static inline void bch2_extent_ptr_to_bp(struct bch_fs *c, struct bch_dev *ca,
 			   enum btree_id btree_id, unsigned level,
 			   struct bkey_s_c k, struct extent_ptr_decoded p,
+			   const union bch_extent_entry *entry,
 			   struct bpos *bucket_pos, struct bch_backpointer *bp)
 {
-	enum bch_data_type data_type = bkey_ptr_data_type(btree_id, level, k, p);
+	enum bch_data_type data_type = bch2_bkey_ptr_data_type(k, p, entry);
 	s64 sectors = level ? btree_sectors(c) : k.k->size;
 	u32 bucket_offset;
 
-	*bucket_pos = PTR_BUCKET_POS_OFFSET(c, &p.ptr, &bucket_offset);
+	*bucket_pos = PTR_BUCKET_POS_OFFSET(ca, &p.ptr, &bucket_offset);
 	*bp = (struct bch_backpointer) {
 		.btree_id	= btree_id,
 		.level		= level,
@@ -119,7 +157,7 @@ static inline void bch2_extent_ptr_to_bp(struct bch_fs *c,
 	};
 }
 
-int bch2_get_next_backpointer(struct btree_trans *, struct bpos, int,
+int bch2_get_next_backpointer(struct btree_trans *, struct bch_dev *ca, struct bpos, int,
 			      struct bpos *, struct bch_backpointer *, unsigned);
 struct bkey_s_c bch2_backpointer_get_key(struct btree_trans *, struct btree_iter *,
 					 struct bpos, struct bch_backpointer,
diff --git a/fs/bcachefs/bbpos_types.h b/fs/bcachefs/bbpos_types.h
index 5198e94cf3..f63893344f 100644
--- a/fs/bcachefs/bbpos_types.h
+++ b/fs/bcachefs/bbpos_types.h
@@ -13,6 +13,6 @@ static inline struct bbpos BBPOS(enum btree_id btree, struct bpos pos)
 }
 
 #define BBPOS_MIN	BBPOS(0, POS_MIN)
-#define BBPOS_MAX	BBPOS(BTREE_ID_NR - 1, POS_MAX)
+#define BBPOS_MAX	BBPOS(BTREE_ID_NR - 1, SPOS_MAX)
 
 #endif /* _BCACHEFS_BBPOS_TYPES_H */
diff --git a/fs/bcachefs/bcachefs.h b/fs/bcachefs/bcachefs.h
index 69d0d60d50..1106fec6e1 100644
--- a/fs/bcachefs/bcachefs.h
+++ b/fs/bcachefs/bcachefs.h
@@ -209,9 +209,10 @@
 #include "fifo.h"
 #include "nocow_locking_types.h"
 #include "opts.h"
-#include "recovery_types.h"
+#include "recovery_passes_types.h"
 #include "sb-errors_types.h"
 #include "seqmutex.h"
+#include "time_stats.h"
 #include "util.h"
 
 #ifdef CONFIG_BCACHEFS_DEBUG
@@ -266,6 +267,9 @@ do {									\
 #define bch2_fmt(_c, fmt)		bch2_log_msg(_c, fmt "\n")
 
 __printf(2, 3)
+void bch2_print_opts(struct bch_opts *, const char *, ...);
+
+__printf(2, 3)
 void __bch2_print(struct bch_fs *c, const char *fmt, ...);
 
 #define maybe_dev_to_fs(_c)	_Generic((_c),				\
@@ -355,6 +359,8 @@ do {									\
 #define BCH_DEBUG_PARAMS_ALWAYS()					\
 	BCH_DEBUG_PARAM(key_merging_disabled,				\
 		"Disables merging of extents")				\
+	BCH_DEBUG_PARAM(btree_node_merging_disabled,			\
+		"Disables merging of btree nodes")			\
 	BCH_DEBUG_PARAM(btree_gc_always_rewrite,			\
 		"Causes mark and sweep to compact and rewrite every "	\
 		"btree node it traverses")				\
@@ -451,7 +457,9 @@ enum bch_time_stats {
 };
 
 #include "alloc_types.h"
+#include "btree_gc_types.h"
 #include "btree_types.h"
+#include "btree_node_scan_types.h"
 #include "btree_write_buffer_types.h"
 #include "buckets_types.h"
 #include "buckets_waiting_for_journal_types.h"
@@ -463,6 +471,7 @@ enum bch_time_stats {
 #include "quota_types.h"
 #include "rebalance_types.h"
 #include "replicas_types.h"
+#include "sb-members_types.h"
 #include "subvolume_types.h"
 #include "super_types.h"
 #include "thread_with_file_types.h"
@@ -480,55 +489,24 @@ enum bch_time_stats {
 
 struct btree;
 
-enum gc_phase {
-	GC_PHASE_NOT_RUNNING,
-	GC_PHASE_START,
-	GC_PHASE_SB,
-
-	GC_PHASE_BTREE_stripes,
-	GC_PHASE_BTREE_extents,
-	GC_PHASE_BTREE_inodes,
-	GC_PHASE_BTREE_dirents,
-	GC_PHASE_BTREE_xattrs,
-	GC_PHASE_BTREE_alloc,
-	GC_PHASE_BTREE_quotas,
-	GC_PHASE_BTREE_reflink,
-	GC_PHASE_BTREE_subvolumes,
-	GC_PHASE_BTREE_snapshots,
-	GC_PHASE_BTREE_lru,
-	GC_PHASE_BTREE_freespace,
-	GC_PHASE_BTREE_need_discard,
-	GC_PHASE_BTREE_backpointers,
-	GC_PHASE_BTREE_bucket_gens,
-	GC_PHASE_BTREE_snapshot_trees,
-	GC_PHASE_BTREE_deleted_inodes,
-	GC_PHASE_BTREE_logged_ops,
-	GC_PHASE_BTREE_rebalance_work,
-
-	GC_PHASE_PENDING_DELETE,
-};
-
-struct gc_pos {
-	enum gc_phase		phase;
-	struct bpos		pos;
-	unsigned		level;
-};
-
-struct reflink_gc {
-	u64		offset;
-	u32		size;
-	u32		refcount;
-};
-
-typedef GENRADIX(struct reflink_gc) reflink_gc_table;
-
 struct io_count {
 	u64			sectors[2][BCH_DATA_NR];
 };
 
+struct discard_in_flight {
+	bool			in_progress:1;
+	u64			bucket:63;
+};
+
 struct bch_dev {
 	struct kobject		kobj;
+#ifdef CONFIG_BCACHEFS_DEBUG
+	atomic_long_t		ref;
+	bool			dying;
+	unsigned long		last_put;
+#else
 	struct percpu_ref	ref;
+#endif
 	struct completion	ref_completion;
 	struct percpu_ref	io_ref;
 	struct completion	io_ref_completion;
@@ -554,14 +532,11 @@ struct bch_dev {
 
 	struct bch_devs_mask	self;
 
-	/* biosets used in cloned bios for writing multiple replicas */
-	struct bio_set		replica_set;
-
 	/*
 	 * Buckets:
 	 * Per-bucket arrays are protected by c->mark_lock, bucket_lock and
 	 * gc_lock, for device resize - holding any is sufficient for access:
-	 * Or rcu_read_lock(), but only for ptr_stale():
+	 * Or rcu_read_lock(), but only for dev_ptr_stale():
 	 */
 	struct bucket_array __rcu *buckets_gc;
 	struct bucket_gens __rcu *bucket_gens;
@@ -575,7 +550,7 @@ struct bch_dev {
 
 	/* Allocator: */
 	u64			new_fs_bucket_idx;
-	u64			alloc_cursor;
+	u64			alloc_cursor[3];
 
 	unsigned		nr_open_buckets;
 	unsigned		nr_btree_reserve;
@@ -584,6 +559,12 @@ struct bch_dev {
 	size_t			inc_gen_really_needs_gc;
 	size_t			buckets_waiting_on_journal;
 
+	struct work_struct	invalidate_work;
+	struct work_struct	discard_work;
+	struct mutex		discard_buckets_in_flight_lock;
+	DARRAY(struct discard_in_flight)	discard_buckets_in_flight;
+	struct work_struct	discard_fast_work;
+
 	atomic64_t		rebalance_work;
 
 	struct journal_device	journal;
@@ -593,7 +574,7 @@ struct bch_dev {
 
 	/* The rest of this all shows up in sysfs */
 	atomic64_t		cur_latency[2];
-	struct bch2_time_stats	io_latency[2];
+	struct bch2_time_stats_quantiles io_latency[2];
 
 #define CONGESTED_MAX		1024
 	atomic_t		congested;
@@ -609,6 +590,7 @@ struct bch_dev {
  */
 
 #define BCH_FS_FLAGS()			\
+	x(new_fs)			\
 	x(started)			\
 	x(may_go_rw)			\
 	x(rw)				\
@@ -620,12 +602,12 @@ struct bch_dev {
 	x(clean_shutdown)		\
 	x(fsck_running)			\
 	x(initial_gc_unfixed)		\
-	x(need_another_gc)		\
 	x(need_delete_dead_snapshots)	\
 	x(error)			\
 	x(topology_error)		\
 	x(errors_fixed)			\
-	x(errors_not_fixed)
+	x(errors_not_fixed)		\
+	x(no_invalid_checks)
 
 enum bch_fs_flags {
 #define x(n)		BCH_FS_##n,
@@ -663,6 +645,8 @@ struct journal_seq_blacklist_table {
 };
 
 struct journal_keys {
+	/* must match layout in darray_types.h */
+	size_t			nr, size;
 	struct journal_key {
 		u64		journal_seq;
 		u32		journal_offset;
@@ -671,15 +655,13 @@ struct journal_keys {
 		bool		allocated;
 		bool		overwritten;
 		struct bkey_i	*k;
-	}			*d;
+	}			*data;
 	/*
 	 * Gap buffer: instead of all the empty space in the array being at the
 	 * end of the buffer - from @nr to @size - the empty space is at @gap.
 	 * This means that sequential insertions are O(n) instead of O(n^2).
 	 */
 	size_t			gap;
-	size_t			nr;
-	size_t			size;
 	atomic_t		ref;
 	bool			initial_ref_held;
 };
@@ -702,9 +684,13 @@ struct btree_trans_buf {
 	x(stripe_delete)						\
 	x(reflink)							\
 	x(fallocate)							\
+	x(fsync)							\
+	x(dio_write)							\
 	x(discard)							\
+	x(discard_fast)							\
 	x(invalidate)							\
 	x(delete_dead_snapshots)					\
+	x(gc_gens)							\
 	x(snapshot_delete_pagecache)					\
 	x(sysfs)							\
 	x(btree_write_buffer)
@@ -790,6 +776,7 @@ struct bch_fs {
 		u64		features;
 		u64		compat;
 		unsigned long	errors_silent[BITS_TO_LONGS(BCH_SB_ERR_MAX)];
+		u64		btrees_lost_data;
 	}			sb;
 
 
@@ -804,7 +791,6 @@ struct bch_fs {
 
 	/* snapshot.c: */
 	struct snapshot_table __rcu *snapshots;
-	size_t			snapshot_table_size;
 	struct mutex		snapshot_table_lock;
 	struct rw_semaphore	snapshot_create_lock;
 
@@ -815,7 +801,8 @@ struct bch_fs {
 
 	/* BTREE CACHE */
 	struct bio_set		btree_bio;
-	struct workqueue_struct	*io_complete_wq;
+	struct workqueue_struct	*btree_read_complete_wq;
+	struct workqueue_struct	*btree_write_submit_wq;
 
 	struct btree_root	btree_roots_known[BTREE_ID_NR];
 	DARRAY(struct btree_root) btree_roots_extra;
@@ -843,6 +830,8 @@ struct bch_fs {
 	struct workqueue_struct	*btree_interior_update_worker;
 	struct work_struct	btree_interior_update_work;
 
+	struct workqueue_struct	*btree_node_rewrite_worker;
+
 	struct list_head	pending_node_rewrites;
 	struct mutex		pending_node_rewrites_lock;
 
@@ -914,13 +903,10 @@ struct bch_fs {
 	/* JOURNAL SEQ BLACKLIST */
 	struct journal_seq_blacklist_table *
 				journal_seq_blacklist_table;
-	struct work_struct	journal_seq_blacklist_gc_work;
 
 	/* ALLOCATOR */
 	spinlock_t		freelist_lock;
 	struct closure_waitlist	freelist_wait;
-	u64			blocked_allocate;
-	u64			blocked_allocate_open_bucket;
 
 	open_bucket_idx_t	open_buckets_freelist;
 	open_bucket_idx_t	open_buckets_nr_free;
@@ -940,12 +926,9 @@ struct bch_fs {
 	unsigned		write_points_nr;
 
 	struct buckets_waiting_for_journal buckets_waiting_for_journal;
-	struct work_struct	discard_work;
-	struct work_struct	invalidate_work;
 
 	/* GARBAGE COLLECTION */
-	struct task_struct	*gc_thread;
-	atomic_t		kick_gc;
+	struct work_struct	gc_gens_work;
 	unsigned long		gc_count;
 
 	enum btree_id		gc_gens_btree;
@@ -975,6 +958,7 @@ struct bch_fs {
 	struct bio_set		bio_read;
 	struct bio_set		bio_read_split;
 	struct bio_set		bio_write;
+	struct bio_set		replica_set;
 	struct mutex		bio_bounce_pages_lock;
 	mempool_t		bio_bounce_pages;
 	struct bucket_nocow_lock_table
@@ -1095,12 +1079,13 @@ struct bch_fs {
 	struct journal_keys	journal_keys;
 	struct list_head	journal_iters;
 
+	struct find_btree_nodes	found_btree_nodes;
+
 	u64			last_bucket_seq_cleanup;
 
 	u64			counters_on_mount[BCH_COUNTER_NR];
 	u64 __percpu		*counters;
 
-	unsigned		btree_gc_periodic:1;
 	unsigned		copy_gc_enabled:1;
 	bool			promote_whole_extents;
 
@@ -1235,9 +1220,9 @@ static inline s64 bch2_current_time(const struct bch_fs *c)
 	return timespec_to_bch2_time(c, now);
 }
 
-static inline bool bch2_dev_exists2(const struct bch_fs *c, unsigned dev)
+static inline u64 bch2_current_io_time(const struct bch_fs *c, int rw)
 {
-	return dev < c->sb.nr_devices && c->devs[dev];
+	return max(1ULL, (u64) atomic64_read(&c->io_clock[rw].now) & LRU_TIME_MAX);
 }
 
 static inline struct stdio_redirect *bch2_fs_stdio_redirect(struct bch_fs *c)
diff --git a/fs/bcachefs/bcachefs_format.h b/fs/bcachefs/bcachefs_format.h
index 0668b682a2..e3b1bde489 100644
--- a/fs/bcachefs/bcachefs_format.h
+++ b/fs/bcachefs/bcachefs_format.h
@@ -76,6 +76,7 @@
 #include <asm/byteorder.h>
 #include <linux/kernel.h>
 #include <linux/uuid.h>
+#include <uapi/linux/magic.h>
 #include "vstructs.h"
 
 #ifdef __KERNEL__
@@ -189,7 +190,11 @@ struct bversion {
 	__u32		hi;
 	__u64		lo;
 #endif
-} __packed __aligned(4);
+} __packed
+#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
+__aligned(4)
+#endif
+;
 
 struct bkey {
 	/* Size of combined key and value, in u64s */
@@ -222,7 +227,36 @@ struct bkey {
 
 	__u8		pad[1];
 #endif
-} __packed __aligned(8);
+} __packed
+#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
+/*
+ * The big-endian version of bkey can't be compiled by rustc with the "aligned"
+ * attr since it doesn't allow types to have both "packed" and "aligned" attrs.
+ * So for Rust compatibility, don't include this. It can be included in the LE
+ * version because the "packed" attr is redundant in that case.
+ *
+ * History: (quoting Kent)
+ *
+ * Specifically, when i was designing bkey, I wanted the header to be no
+ * bigger than necessary so that bkey_packed could use the rest. That means that
+ * decently offten extent keys will fit into only 8 bytes, instead of spilling over
+ * to 16.
+ *
+ * But packed_bkey treats the part after the header - the packed section -
+ * as a single multi word, variable length integer. And bkey, the unpacked
+ * version, is just a special case version of a bkey_packed; all the packed
+ * bkey code will work on keys in any packed format, the in-memory
+ * representation of an unpacked key also is just one type of packed key...
+ *
+ * So that constrains the key part of a bkig endian bkey to start right
+ * after the header.
+ *
+ * If we ever do a bkey_v2 and need to expand the hedaer by another byte for
+ * some reason - that will clean up this wart.
+ */
+__aligned(8)
+#endif
+;
 
 struct bkey_packed {
 	__u64		_data[0];
@@ -442,6 +476,9 @@ struct bch_lru {
 
 #define LRU_ID_STRIPES		(1U << 16)
 
+#define LRU_TIME_BITS	48
+#define LRU_TIME_MAX	((1ULL << LRU_TIME_BITS) - 1)
+
 /* Optional/variable size superblock sections: */
 
 struct bch_sb_field {
@@ -469,16 +506,22 @@ struct bch_sb_field {
 
 #include "alloc_background_format.h"
 #include "extents_format.h"
-#include "reflink_format.h"
 #include "ec_format.h"
-#include "inode_format.h"
 #include "dirent_format.h"
-#include "xattr_format.h"
-#include "quota_format.h"
+#include "disk_groups_format.h"
+#include "inode_format.h"
+#include "journal_seq_blacklist_format.h"
 #include "logged_ops_format.h"
+#include "quota_format.h"
+#include "reflink_format.h"
+#include "replicas_format.h"
 #include "snapshot_format.h"
 #include "subvolume_format.h"
 #include "sb-counters_format.h"
+#include "sb-downgrade_format.h"
+#include "sb-errors_format.h"
+#include "sb-members_format.h"
+#include "xattr_format.h"
 
 enum bch_sb_field_type {
 #define x(f, nr)	BCH_SB_FIELD_##f = nr,
@@ -511,92 +554,6 @@ struct bch_sb_field_journal_v2 {
 	}			d[];
 };
 
-/* BCH_SB_FIELD_members_v1: */
-
-#define BCH_MIN_NR_NBUCKETS	(1 << 6)
-
-#define BCH_IOPS_MEASUREMENTS()			\
-	x(seqread,	0)			\
-	x(seqwrite,	1)			\
-	x(randread,	2)			\
-	x(randwrite,	3)
-
-enum bch_iops_measurement {
-#define x(t, n) BCH_IOPS_##t = n,
-	BCH_IOPS_MEASUREMENTS()
-#undef x
-	BCH_IOPS_NR
-};
-
-#define BCH_MEMBER_ERROR_TYPES()		\
-	x(read,		0)			\
-	x(write,	1)			\
-	x(checksum,	2)
-
-enum bch_member_error_type {
-#define x(t, n) BCH_MEMBER_ERROR_##t = n,
-	BCH_MEMBER_ERROR_TYPES()
-#undef x
-	BCH_MEMBER_ERROR_NR
-};
-
-struct bch_member {
-	__uuid_t		uuid;
-	__le64			nbuckets;	/* device size */
-	__le16			first_bucket;   /* index of first bucket used */
-	__le16			bucket_size;	/* sectors */
-	__le32			pad;
-	__le64			last_mount;	/* time_t */
-
-	__le64			flags;
-	__le32			iops[4];
-	__le64			errors[BCH_MEMBER_ERROR_NR];
-	__le64			errors_at_reset[BCH_MEMBER_ERROR_NR];
-	__le64			errors_reset_time;
-	__le64			seq;
-};
-
-#define BCH_MEMBER_V1_BYTES	56
-
-LE64_BITMASK(BCH_MEMBER_STATE,		struct bch_member, flags,  0,  4)
-/* 4-14 unused, was TIER, HAS_(META)DATA, REPLACEMENT */
-LE64_BITMASK(BCH_MEMBER_DISCARD,	struct bch_member, flags, 14, 15)
-LE64_BITMASK(BCH_MEMBER_DATA_ALLOWED,	struct bch_member, flags, 15, 20)
-LE64_BITMASK(BCH_MEMBER_GROUP,		struct bch_member, flags, 20, 28)
-LE64_BITMASK(BCH_MEMBER_DURABILITY,	struct bch_member, flags, 28, 30)
-LE64_BITMASK(BCH_MEMBER_FREESPACE_INITIALIZED,
-					struct bch_member, flags, 30, 31)
-
-#if 0
-LE64_BITMASK(BCH_MEMBER_NR_READ_ERRORS,	struct bch_member, flags[1], 0,  20);
-LE64_BITMASK(BCH_MEMBER_NR_WRITE_ERRORS,struct bch_member, flags[1], 20, 40);
-#endif
-
-#define BCH_MEMBER_STATES()			\
-	x(rw,		0)			\
-	x(ro,		1)			\
-	x(failed,	2)			\
-	x(spare,	3)
-
-enum bch_member_state {
-#define x(t, n) BCH_MEMBER_STATE_##t = n,
-	BCH_MEMBER_STATES()
-#undef x
-	BCH_MEMBER_STATE_NR
-};
-
-struct bch_sb_field_members_v1 {
-	struct bch_sb_field	field;
-	struct bch_member	_members[]; //Members are now variable size
-};
-
-struct bch_sb_field_members_v2 {
-	struct bch_sb_field	field;
-	__le16			member_bytes; //size of single member entry
-	u8			pad[6];
-	struct bch_member	_members[];
-};
-
 /* BCH_SB_FIELD_crypt: */
 
 struct nonce {
@@ -645,8 +602,6 @@ LE64_BITMASK(BCH_KDF_SCRYPT_N,	struct bch_sb_field_crypt, kdf_flags,  0, 16);
 LE64_BITMASK(BCH_KDF_SCRYPT_R,	struct bch_sb_field_crypt, kdf_flags, 16, 32);
 LE64_BITMASK(BCH_KDF_SCRYPT_P,	struct bch_sb_field_crypt, kdf_flags, 32, 48);
 
-/* BCH_SB_FIELD_replicas: */
-
 #define BCH_DATA_TYPES()		\
 	x(free,		0)		\
 	x(sb,		1)		\
@@ -689,50 +644,6 @@ static inline bool data_type_is_hidden(enum bch_data_type type)
 	}
 }
 
-struct bch_replicas_entry_v0 {
-	__u8			data_type;
-	__u8			nr_devs;
-	__u8			devs[];
-} __packed;
-
-struct bch_sb_field_replicas_v0 {
-	struct bch_sb_field	field;
-	struct bch_replicas_entry_v0 entries[];
-} __packed __aligned(8);
-
-struct bch_replicas_entry_v1 {
-	__u8			data_type;
-	__u8			nr_devs;
-	__u8			nr_required;
-	__u8			devs[];
-} __packed;
-
-#define replicas_entry_bytes(_i)					\
-	(offsetof(typeof(*(_i)), devs) + (_i)->nr_devs)
-
-struct bch_sb_field_replicas {
-	struct bch_sb_field	field;
-	struct bch_replicas_entry_v1 entries[];
-} __packed __aligned(8);
-
-/* BCH_SB_FIELD_disk_groups: */
-
-#define BCH_SB_LABEL_SIZE		32
-
-struct bch_disk_group {
-	__u8			label[BCH_SB_LABEL_SIZE];
-	__le64			flags[2];
-} __packed __aligned(8);
-
-LE64_BITMASK(BCH_GROUP_DELETED,		struct bch_disk_group, flags[0], 0,  1)
-LE64_BITMASK(BCH_GROUP_DATA_ALLOWED,	struct bch_disk_group, flags[0], 1,  6)
-LE64_BITMASK(BCH_GROUP_PARENT,		struct bch_disk_group, flags[0], 6, 24)
-
-struct bch_sb_field_disk_groups {
-	struct bch_sb_field	field;
-	struct bch_disk_group	entries[];
-} __packed __aligned(8);
-
 /*
  * On clean shutdown, store btree roots and current journal sequence number in
  * the superblock:
@@ -760,43 +671,11 @@ struct bch_sb_field_clean {
 	__u64			_data[];
 };
 
-struct journal_seq_blacklist_entry {
-	__le64			start;
-	__le64			end;
-};
-
-struct bch_sb_field_journal_seq_blacklist {
-	struct bch_sb_field	field;
-	struct journal_seq_blacklist_entry start[];
-};
-
-struct bch_sb_field_errors {
-	struct bch_sb_field	field;
-	struct bch_sb_field_error_entry {
-		__le64		v;
-		__le64		last_error_time;
-	}			entries[];
-};
-
-LE64_BITMASK(BCH_SB_ERROR_ENTRY_ID,	struct bch_sb_field_error_entry, v,  0, 16);
-LE64_BITMASK(BCH_SB_ERROR_ENTRY_NR,	struct bch_sb_field_error_entry, v, 16, 64);
-
 struct bch_sb_field_ext {
 	struct bch_sb_field	field;
 	__le64			recovery_passes_required[2];
 	__le64			errors_silent[8];
-};
-
-struct bch_sb_field_downgrade_entry {
-	__le16			version;
-	__le64			recovery_passes[2];
-	__le16			nr_errors;
-	__le16			errors[] __counted_by(nr_errors);
-} __packed __aligned(2);
-
-struct bch_sb_field_downgrade {
-	struct bch_sb_field	field;
-	struct bch_sb_field_downgrade_entry entries[];
+	__le64			btrees_lost_data;
 };
 
 /* Superblock: */
@@ -840,7 +719,10 @@ struct bch_sb_field_downgrade {
 	x(snapshot_skiplists,		BCH_VERSION(1,  1))		\
 	x(deleted_inodes,		BCH_VERSION(1,  2))		\
 	x(rebalance_work,		BCH_VERSION(1,  3))		\
-	x(member_seq,			BCH_VERSION(1,  4))
+	x(member_seq,			BCH_VERSION(1,  4))		\
+	x(subvolume_fs_parent,		BCH_VERSION(1,  5))		\
+	x(btree_subvolume_children,	BCH_VERSION(1,  6))		\
+	x(mi_btree_bitmap,		BCH_VERSION(1,  7))
 
 enum bcachefs_metadata_version {
 	bcachefs_metadata_version_min = 9,
@@ -856,7 +738,8 @@ unsigned bcachefs_metadata_required_upgrade_below = bcachefs_metadata_version_re
 #define bcachefs_metadata_version_current	(bcachefs_metadata_version_max - 1)
 
 #define BCH_SB_SECTOR			8
-#define BCH_SB_MEMBERS_MAX		64 /* XXX kill */
+
+#define BCH_SB_LAYOUT_SIZE_BITS_MAX	16 /* 32 MB */
 
 struct bch_sb_layout {
 	__uuid_t		magic;	/* bcachefs superblock UUID */
@@ -1107,8 +990,9 @@ enum bch_version_upgrade_opts {
 
 #define BCH_ERROR_ACTIONS()		\
 	x(continue,		0)	\
-	x(ro,			1)	\
-	x(panic,		2)
+	x(fix_safe,		1)	\
+	x(panic,		2)	\
+	x(ro,			3)
 
 enum bch_error_actions {
 #define x(t, n) BCH_ON_ERROR_##t = n,
@@ -1236,7 +1120,7 @@ enum bch_compression_opts {
 	UUID_INIT(0xc68573f6, 0x66ce, 0x90a9,				\
 		  0xd9, 0x6a, 0x60, 0xcf, 0x80, 0x3d, 0xf7, 0xef)
 
-#define BCACHEFS_STATFS_MAGIC		0xca451a4e
+#define BCACHEFS_STATFS_MAGIC		BCACHEFS_SUPER_MAGIC
 
 #define JSET_MAGIC		__cpu_to_le64(0x245235c1a3625032ULL)
 #define BSET_MAGIC		__cpu_to_le64(0x90135c78b99e07f5ULL)
@@ -1275,9 +1159,10 @@ static inline __u64 __bset_magic(struct bch_sb *sb)
 	x(dev_usage,		8)		\
 	x(log,			9)		\
 	x(overwrite,		10)		\
-	x(write_buffer_keys,	11)
+	x(write_buffer_keys,	11)		\
+	x(datetime,		12)
 
-enum {
+enum bch_jset_entry_type {
 #define x(f, nr)	BCH_JSET_ENTRY_##f	= nr,
 	BCH_JSET_ENTRY_TYPES()
 #undef x
@@ -1323,7 +1208,7 @@ struct jset_entry_blacklist_v2 {
 	x(inodes,		1)		\
 	x(key_version,		2)
 
-enum {
+enum bch_fs_usage_type {
 #define x(f, nr)	BCH_FS_USAGE_##f	= nr,
 	BCH_FS_USAGE_TYPES()
 #undef x
@@ -1376,6 +1261,11 @@ struct jset_entry_log {
 	u8			d[];
 } __packed __aligned(8);
 
+struct jset_entry_datetime {
+	struct jset_entry	entry;
+	__le64			seconds;
+} __packed __aligned(8);
+
 /*
  * On disk format for a journal entry:
  * seq is monotonically increasing; every journal entry has its own unique
@@ -1459,7 +1349,8 @@ enum btree_id_flags {
 	  BIT_ULL(KEY_TYPE_stripe))						\
 	x(reflink,		7,	BTREE_ID_EXTENTS|BTREE_ID_DATA,		\
 	  BIT_ULL(KEY_TYPE_reflink_v)|						\
-	  BIT_ULL(KEY_TYPE_indirect_inline_data))				\
+	  BIT_ULL(KEY_TYPE_indirect_inline_data)|				\
+	  BIT_ULL(KEY_TYPE_error))						\
 	x(subvolumes,		8,	0,					\
 	  BIT_ULL(KEY_TYPE_subvolume))						\
 	x(snapshots,		9,	0,					\
@@ -1482,7 +1373,9 @@ enum btree_id_flags {
 	  BIT_ULL(KEY_TYPE_logged_op_truncate)|					\
 	  BIT_ULL(KEY_TYPE_logged_op_finsert))					\
 	x(rebalance_work,	18,	BTREE_ID_SNAPSHOT_FIELD,		\
-	  BIT_ULL(KEY_TYPE_set)|BIT_ULL(KEY_TYPE_cookie))
+	  BIT_ULL(KEY_TYPE_set)|BIT_ULL(KEY_TYPE_cookie))			\
+	x(subvolume_children,	19,	0,					\
+	  BIT_ULL(KEY_TYPE_set))
 
 enum btree_id {
 #define x(name, nr, ...) BTREE_ID_##name = nr,
@@ -1491,6 +1384,27 @@ enum btree_id {
 	BTREE_ID_NR
 };
 
+/*
+ * Maximum number of btrees that we will _ever_ have under the current scheme,
+ * where we refer to them with 64 bit bitfields - and we also need a bit for
+ * the interior btree node type:
+ */
+#define BTREE_ID_NR_MAX		63
+
+static inline bool btree_id_is_alloc(enum btree_id id)
+{
+	switch (id) {
+	case BTREE_ID_alloc:
+	case BTREE_ID_backpointers:
+	case BTREE_ID_need_discard:
+	case BTREE_ID_freespace:
+	case BTREE_ID_bucket_gens:
+		return true;
+	default:
+		return false;
+	}
+}
+
 #define BTREE_MAX_DEPTH		4U
 
 /* Btree nodes */
diff --git a/fs/bcachefs/bkey.c b/fs/bcachefs/bkey.c
index 76e79a15ba..587d7318a2 100644
--- a/fs/bcachefs/bkey.c
+++ b/fs/bcachefs/bkey.c
@@ -640,7 +640,7 @@ struct bkey_format bch2_bkey_format_done(struct bkey_format_state *s)
 
 int bch2_bkey_format_invalid(struct bch_fs *c,
 			     struct bkey_format *f,
-			     enum bkey_invalid_flags flags,
+			     enum bch_validate_flags flags,
 			     struct printbuf *err)
 {
 	unsigned i, bits = KEY_PACKED_BITS_START;
@@ -656,20 +656,18 @@ int bch2_bkey_format_invalid(struct bch_fs *c,
 	 * unpacked format:
 	 */
 	for (i = 0; i < f->nr_fields; i++) {
-		if (!c || c->sb.version_min >= bcachefs_metadata_version_snapshot) {
+		if ((!c || c->sb.version_min >= bcachefs_metadata_version_snapshot) &&
+		    bch2_bkey_format_field_overflows(f, i)) {
 			unsigned unpacked_bits = bch2_bkey_format_current.bits_per_field[i];
 			u64 unpacked_max = ~((~0ULL << 1) << (unpacked_bits - 1));
-			u64 packed_max = f->bits_per_field[i]
-				? ~((~0ULL << 1) << (f->bits_per_field[i] - 1))
+			unsigned packed_bits = min(64, f->bits_per_field[i]);
+			u64 packed_max = packed_bits
+				? ~((~0ULL << 1) << (packed_bits - 1))
 				: 0;
-			u64 field_offset = le64_to_cpu(f->field_offset[i]);
 
-			if (packed_max + field_offset < packed_max ||
-			    packed_max + field_offset > unpacked_max) {
-				prt_printf(err, "field %u too large: %llu + %llu > %llu",
-					   i, packed_max, field_offset, unpacked_max);
-				return -BCH_ERR_invalid;
-			}
+			prt_printf(err, "field %u too large: %llu + %llu > %llu",
+				   i, packed_max, le64_to_cpu(f->field_offset[i]), unpacked_max);
+			return -BCH_ERR_invalid;
 		}
 
 		bits += f->bits_per_field[i];
@@ -1067,7 +1065,7 @@ void bch2_bkey_swab_key(const struct bkey_format *_f, struct bkey_packed *k)
 {
 	const struct bkey_format *f = bkey_packed(k) ? _f : &bch2_bkey_format_current;
 	u8 *l = k->key_start;
-	u8 *h = (u8 *) (k->_data + f->key_u64s) - 1;
+	u8 *h = (u8 *) ((u64 *) k->_data + f->key_u64s) - 1;
 
 	while (l < h) {
 		swap(*l, *h);
diff --git a/fs/bcachefs/bkey.h b/fs/bcachefs/bkey.h
index 831be01809..936357149c 100644
--- a/fs/bcachefs/bkey.h
+++ b/fs/bcachefs/bkey.h
@@ -4,15 +4,15 @@
 
 #include <linux/bug.h>
 #include "bcachefs_format.h"
-
+#include "bkey_types.h"
 #include "btree_types.h"
 #include "util.h"
 #include "vstructs.h"
 
-enum bkey_invalid_flags {
-	BKEY_INVALID_WRITE		= (1U << 0),
-	BKEY_INVALID_COMMIT		= (1U << 1),
-	BKEY_INVALID_JOURNAL		= (1U << 2),
+enum bch_validate_flags {
+	BCH_VALIDATE_write		= (1U << 0),
+	BCH_VALIDATE_commit		= (1U << 1),
+	BCH_VALIDATE_journal		= (1U << 2),
 };
 
 #if 0
@@ -31,57 +31,6 @@ void bch2_bkey_packed_to_binary_text(struct printbuf *,
 				     const struct bkey_format *,
 				     const struct bkey_packed *);
 
-/* bkey with split value, const */
-struct bkey_s_c {
-	const struct bkey	*k;
-	const struct bch_val	*v;
-};
-
-/* bkey with split value */
-struct bkey_s {
-	union {
-	struct {
-		struct bkey	*k;
-		struct bch_val	*v;
-	};
-	struct bkey_s_c		s_c;
-	};
-};
-
-#define bkey_p_next(_k)		vstruct_next(_k)
-
-static inline struct bkey_i *bkey_next(struct bkey_i *k)
-{
-	return (struct bkey_i *) ((u64 *) k->_data + k->k.u64s);
-}
-
-#define bkey_val_u64s(_k)	((_k)->u64s - BKEY_U64s)
-
-static inline size_t bkey_val_bytes(const struct bkey *k)
-{
-	return bkey_val_u64s(k) * sizeof(u64);
-}
-
-static inline void set_bkey_val_u64s(struct bkey *k, unsigned val_u64s)
-{
-	unsigned u64s = BKEY_U64s + val_u64s;
-
-	BUG_ON(u64s > U8_MAX);
-	k->u64s = u64s;
-}
-
-static inline void set_bkey_val_bytes(struct bkey *k, unsigned bytes)
-{
-	set_bkey_val_u64s(k, DIV_ROUND_UP(bytes, sizeof(u64)));
-}
-
-#define bkey_val_end(_k)	((void *) (((u64 *) (_k).v) + bkey_val_u64s((_k).k)))
-
-#define bkey_deleted(_k)	((_k)->type == KEY_TYPE_deleted)
-
-#define bkey_whiteout(_k)				\
-	((_k)->type == KEY_TYPE_deleted || (_k)->type == KEY_TYPE_whiteout)
-
 enum bkey_lr_packed {
 	BKEY_PACKED_BOTH,
 	BKEY_PACKED_RIGHT,
@@ -245,6 +194,13 @@ static inline struct bpos bkey_max(struct bpos l, struct bpos r)
 	return bkey_gt(l, r) ? l : r;
 }
 
+static inline bool bkey_and_val_eq(struct bkey_s_c l, struct bkey_s_c r)
+{
+	return bpos_eq(l.k->p, r.k->p) &&
+		bkey_bytes(l.k) == bkey_bytes(r.k) &&
+		!memcmp(l.v, r.v, bkey_val_bytes(l.k));
+}
+
 void bch2_bpos_swab(struct bpos *);
 void bch2_bkey_swab_key(const struct bkey_format *, struct bkey_packed *);
 
@@ -362,10 +318,13 @@ static inline struct bpos bkey_start_pos(const struct bkey *k)
 static inline unsigned bkeyp_key_u64s(const struct bkey_format *format,
 				      const struct bkey_packed *k)
 {
-	unsigned ret = bkey_packed(k) ? format->key_u64s : BKEY_U64s;
+	return bkey_packed(k) ? format->key_u64s : BKEY_U64s;
+}
 
-	EBUG_ON(k->u64s < ret);
-	return ret;
+static inline bool bkeyp_u64s_valid(const struct bkey_format *f,
+				    const struct bkey_packed *k)
+{
+	return ((unsigned) k->u64s - bkeyp_key_u64s(f, k) <= U8_MAX - BKEY_U64s);
 }
 
 static inline unsigned bkeyp_key_bytes(const struct bkey_format *format,
@@ -553,155 +512,6 @@ static inline void bkey_reassemble(struct bkey_i *dst,
 	memcpy_u64s_small(&dst->v, src.v, bkey_val_u64s(src.k));
 }
 
-#define bkey_s_null		((struct bkey_s)   { .k = NULL })
-#define bkey_s_c_null		((struct bkey_s_c) { .k = NULL })
-
-#define bkey_s_err(err)		((struct bkey_s)   { .k = ERR_PTR(err) })
-#define bkey_s_c_err(err)	((struct bkey_s_c) { .k = ERR_PTR(err) })
-
-static inline struct bkey_s bkey_to_s(struct bkey *k)
-{
-	return (struct bkey_s) { .k = k, .v = NULL };
-}
-
-static inline struct bkey_s_c bkey_to_s_c(const struct bkey *k)
-{
-	return (struct bkey_s_c) { .k = k, .v = NULL };
-}
-
-static inline struct bkey_s bkey_i_to_s(struct bkey_i *k)
-{
-	return (struct bkey_s) { .k = &k->k, .v = &k->v };
-}
-
-static inline struct bkey_s_c bkey_i_to_s_c(const struct bkey_i *k)
-{
-	return (struct bkey_s_c) { .k = &k->k, .v = &k->v };
-}
-
-/*
- * For a given type of value (e.g. struct bch_extent), generates the types for
- * bkey + bch_extent - inline, split, split const - and also all the conversion
- * functions, which also check that the value is of the correct type.
- *
- * We use anonymous unions for upcasting - e.g. converting from e.g. a
- * bkey_i_extent to a bkey_i - since that's always safe, instead of conversion
- * functions.
- */
-#define x(name, ...)					\
-struct bkey_i_##name {							\
-	union {								\
-		struct bkey		k;				\
-		struct bkey_i		k_i;				\
-	};								\
-	struct bch_##name		v;				\
-};									\
-									\
-struct bkey_s_c_##name {						\
-	union {								\
-	struct {							\
-		const struct bkey	*k;				\
-		const struct bch_##name	*v;				\
-	};								\
-	struct bkey_s_c			s_c;				\
-	};								\
-};									\
-									\
-struct bkey_s_##name {							\
-	union {								\
-	struct {							\
-		struct bkey		*k;				\
-		struct bch_##name	*v;				\
-	};								\
-	struct bkey_s_c_##name		c;				\
-	struct bkey_s			s;				\
-	struct bkey_s_c			s_c;				\
-	};								\
-};									\
-									\
-static inline struct bkey_i_##name *bkey_i_to_##name(struct bkey_i *k)	\
-{									\
-	EBUG_ON(!IS_ERR_OR_NULL(k) && k->k.type != KEY_TYPE_##name);	\
-	return container_of(&k->k, struct bkey_i_##name, k);		\
-}									\
-									\
-static inline const struct bkey_i_##name *				\
-bkey_i_to_##name##_c(const struct bkey_i *k)				\
-{									\
-	EBUG_ON(!IS_ERR_OR_NULL(k) && k->k.type != KEY_TYPE_##name);	\
-	return container_of(&k->k, struct bkey_i_##name, k);		\
-}									\
-									\
-static inline struct bkey_s_##name bkey_s_to_##name(struct bkey_s k)	\
-{									\
-	EBUG_ON(!IS_ERR_OR_NULL(k.k) && k.k->type != KEY_TYPE_##name);	\
-	return (struct bkey_s_##name) {					\
-		.k = k.k,						\
-		.v = container_of(k.v, struct bch_##name, v),		\
-	};								\
-}									\
-									\
-static inline struct bkey_s_c_##name bkey_s_c_to_##name(struct bkey_s_c k)\
-{									\
-	EBUG_ON(!IS_ERR_OR_NULL(k.k) && k.k->type != KEY_TYPE_##name);	\
-	return (struct bkey_s_c_##name) {				\
-		.k = k.k,						\
-		.v = container_of(k.v, struct bch_##name, v),		\
-	};								\
-}									\
-									\
-static inline struct bkey_s_##name name##_i_to_s(struct bkey_i_##name *k)\
-{									\
-	return (struct bkey_s_##name) {					\
-		.k = &k->k,						\
-		.v = &k->v,						\
-	};								\
-}									\
-									\
-static inline struct bkey_s_c_##name					\
-name##_i_to_s_c(const struct bkey_i_##name *k)				\
-{									\
-	return (struct bkey_s_c_##name) {				\
-		.k = &k->k,						\
-		.v = &k->v,						\
-	};								\
-}									\
-									\
-static inline struct bkey_s_##name bkey_i_to_s_##name(struct bkey_i *k)	\
-{									\
-	EBUG_ON(!IS_ERR_OR_NULL(k) && k->k.type != KEY_TYPE_##name);	\
-	return (struct bkey_s_##name) {					\
-		.k = &k->k,						\
-		.v = container_of(&k->v, struct bch_##name, v),		\
-	};								\
-}									\
-									\
-static inline struct bkey_s_c_##name					\
-bkey_i_to_s_c_##name(const struct bkey_i *k)				\
-{									\
-	EBUG_ON(!IS_ERR_OR_NULL(k) && k->k.type != KEY_TYPE_##name);	\
-	return (struct bkey_s_c_##name) {				\
-		.k = &k->k,						\
-		.v = container_of(&k->v, struct bch_##name, v),		\
-	};								\
-}									\
-									\
-static inline struct bkey_i_##name *bkey_##name##_init(struct bkey_i *_k)\
-{									\
-	struct bkey_i_##name *k =					\
-		container_of(&_k->k, struct bkey_i_##name, k);		\
-									\
-	bkey_init(&k->k);						\
-	memset(&k->v, 0, sizeof(k->v));					\
-	k->k.type = KEY_TYPE_##name;					\
-	set_bkey_val_bytes(&k->k, sizeof(k->v));			\
-									\
-	return k;							\
-}
-
-BCH_BKEY_TYPES();
-#undef x
-
 /* byte order helpers */
 
 #if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
@@ -771,8 +581,31 @@ static inline void bch2_bkey_format_add_key(struct bkey_format_state *s, const s
 
 void bch2_bkey_format_add_pos(struct bkey_format_state *, struct bpos);
 struct bkey_format bch2_bkey_format_done(struct bkey_format_state *);
+
+static inline bool bch2_bkey_format_field_overflows(struct bkey_format *f, unsigned i)
+{
+	unsigned f_bits = f->bits_per_field[i];
+	unsigned unpacked_bits = bch2_bkey_format_current.bits_per_field[i];
+	u64 unpacked_mask = ~((~0ULL << 1) << (unpacked_bits - 1));
+	u64 field_offset = le64_to_cpu(f->field_offset[i]);
+
+	if (f_bits > unpacked_bits)
+		return true;
+
+	if ((f_bits == unpacked_bits) && field_offset)
+		return true;
+
+	u64 f_mask = f_bits
+		? ~((~0ULL << (f_bits - 1)) << 1)
+		: 0;
+
+	if (((field_offset + f_mask) & unpacked_mask) < field_offset)
+		return true;
+	return false;
+}
+
 int bch2_bkey_format_invalid(struct bch_fs *, struct bkey_format *,
-			     enum bkey_invalid_flags, struct printbuf *);
+			     enum bch_validate_flags, struct printbuf *);
 void bch2_bkey_format_to_text(struct printbuf *, const struct bkey_format *);
 
 #endif /* _BCACHEFS_BKEY_H */
diff --git a/fs/bcachefs/bkey_methods.c b/fs/bcachefs/bkey_methods.c
index 5e52684764..bd32aac051 100644
--- a/fs/bcachefs/bkey_methods.c
+++ b/fs/bcachefs/bkey_methods.c
@@ -27,7 +27,7 @@ const char * const bch2_bkey_types[] = {
 };
 
 static int deleted_key_invalid(struct bch_fs *c, struct bkey_s_c k,
-			       enum bkey_invalid_flags flags, struct printbuf *err)
+			       enum bch_validate_flags flags, struct printbuf *err)
 {
 	return 0;
 }
@@ -41,7 +41,7 @@ static int deleted_key_invalid(struct bch_fs *c, struct bkey_s_c k,
 })
 
 static int empty_val_key_invalid(struct bch_fs *c, struct bkey_s_c k,
-				 enum bkey_invalid_flags flags, struct printbuf *err)
+				 enum bch_validate_flags flags, struct printbuf *err)
 {
 	int ret = 0;
 
@@ -58,7 +58,7 @@ fsck_err:
 })
 
 static int key_type_cookie_invalid(struct bch_fs *c, struct bkey_s_c k,
-				   enum bkey_invalid_flags flags, struct printbuf *err)
+				   enum bch_validate_flags flags, struct printbuf *err)
 {
 	return 0;
 }
@@ -82,7 +82,7 @@ static void key_type_cookie_to_text(struct printbuf *out, struct bch_fs *c,
 })
 
 static int key_type_inline_data_invalid(struct bch_fs *c, struct bkey_s_c k,
-					enum bkey_invalid_flags flags, struct printbuf *err)
+					enum bch_validate_flags flags, struct printbuf *err)
 {
 	return 0;
 }
@@ -123,9 +123,12 @@ const struct bkey_ops bch2_bkey_null_ops = {
 };
 
 int bch2_bkey_val_invalid(struct bch_fs *c, struct bkey_s_c k,
-			  enum bkey_invalid_flags flags,
+			  enum bch_validate_flags flags,
 			  struct printbuf *err)
 {
+	if (test_bit(BCH_FS_no_invalid_checks, &c->flags))
+		return 0;
+
 	const struct bkey_ops *ops = bch2_bkey_type_ops(k.k->type);
 	int ret = 0;
 
@@ -159,9 +162,12 @@ const char *bch2_btree_node_type_str(enum btree_node_type type)
 
 int __bch2_bkey_invalid(struct bch_fs *c, struct bkey_s_c k,
 			enum btree_node_type type,
-			enum bkey_invalid_flags flags,
+			enum bch_validate_flags flags,
 			struct printbuf *err)
 {
+	if (test_bit(BCH_FS_no_invalid_checks, &c->flags))
+		return 0;
+
 	int ret = 0;
 
 	bkey_fsck_err_on(k.k->u64s < BKEY_U64s, c, err,
@@ -171,11 +177,15 @@ int __bch2_bkey_invalid(struct bch_fs *c, struct bkey_s_c k,
 	if (type >= BKEY_TYPE_NR)
 		return 0;
 
-	bkey_fsck_err_on((flags & BKEY_INVALID_COMMIT) &&
+	bkey_fsck_err_on(k.k->type < KEY_TYPE_MAX &&
+			 (type == BKEY_TYPE_btree || (flags & BCH_VALIDATE_commit)) &&
 			 !(bch2_key_types_allowed[type] & BIT_ULL(k.k->type)), c, err,
 			 bkey_invalid_type_for_btree,
 			 "invalid key type for btree %s (%s)",
-			 bch2_btree_node_type_str(type), bch2_bkey_types[k.k->type]);
+			 bch2_btree_node_type_str(type),
+			 k.k->type < KEY_TYPE_MAX
+			 ? bch2_bkey_types[k.k->type]
+			 : "(unknown)");
 
 	if (btree_node_type_is_extents(type) && !bkey_whiteout(k.k)) {
 		bkey_fsck_err_on(k.k->size == 0, c, err,
@@ -220,7 +230,7 @@ fsck_err:
 
 int bch2_bkey_invalid(struct bch_fs *c, struct bkey_s_c k,
 		      enum btree_node_type type,
-		      enum bkey_invalid_flags flags,
+		      enum bch_validate_flags flags,
 		      struct printbuf *err)
 {
 	return __bch2_bkey_invalid(c, k, type, flags, err) ?:
@@ -388,8 +398,12 @@ void __bch2_bkey_compat(unsigned level, enum btree_id btree_id,
 	for (i = 0; i < nr_compat; i++)
 	switch (!write ? i : nr_compat - 1 - i) {
 	case 0:
-		if (big_endian != CPU_BIG_ENDIAN)
+		if (big_endian != CPU_BIG_ENDIAN) {
+			bch2_bkey_swab_key(f, k);
+		} else if (IS_ENABLED(CONFIG_BCACHEFS_DEBUG)) {
 			bch2_bkey_swab_key(f, k);
+			bch2_bkey_swab_key(f, k);
+		}
 		break;
 	case 1:
 		if (version < bcachefs_metadata_version_bkey_renumber)
diff --git a/fs/bcachefs/bkey_methods.h b/fs/bcachefs/bkey_methods.h
index 03efe8ee56..baef0722f5 100644
--- a/fs/bcachefs/bkey_methods.h
+++ b/fs/bcachefs/bkey_methods.h
@@ -22,14 +22,15 @@ extern const struct bkey_ops bch2_bkey_null_ops;
  */
 struct bkey_ops {
 	int		(*key_invalid)(struct bch_fs *c, struct bkey_s_c k,
-				       enum bkey_invalid_flags flags, struct printbuf *err);
+				       enum bch_validate_flags flags, struct printbuf *err);
 	void		(*val_to_text)(struct printbuf *, struct bch_fs *,
 				       struct bkey_s_c);
 	void		(*swab)(struct bkey_s);
 	bool		(*key_normalize)(struct bch_fs *, struct bkey_s);
 	bool		(*key_merge)(struct bch_fs *, struct bkey_s, struct bkey_s_c);
 	int		(*trigger)(struct btree_trans *, enum btree_id, unsigned,
-				   struct bkey_s_c, struct bkey_s, unsigned);
+				   struct bkey_s_c, struct bkey_s,
+				   enum btree_iter_update_trigger_flags);
 	void		(*compat)(enum btree_id id, unsigned version,
 				  unsigned big_endian, int write,
 				  struct bkey_s);
@@ -48,11 +49,11 @@ static inline const struct bkey_ops *bch2_bkey_type_ops(enum bch_bkey_type type)
 }
 
 int bch2_bkey_val_invalid(struct bch_fs *, struct bkey_s_c,
-			  enum bkey_invalid_flags, struct printbuf *);
+			  enum bch_validate_flags, struct printbuf *);
 int __bch2_bkey_invalid(struct bch_fs *, struct bkey_s_c, enum btree_node_type,
-			enum bkey_invalid_flags, struct printbuf *);
+			enum bch_validate_flags, struct printbuf *);
 int bch2_bkey_invalid(struct bch_fs *, struct bkey_s_c, enum btree_node_type,
-		      enum bkey_invalid_flags, struct printbuf *);
+		      enum bch_validate_flags, struct printbuf *);
 int bch2_bkey_in_btree_node(struct bch_fs *, struct btree *,
 			    struct bkey_s_c, struct printbuf *);
 
@@ -76,56 +77,10 @@ static inline bool bch2_bkey_maybe_mergable(const struct bkey *l, const struct b
 
 bool bch2_bkey_merge(struct bch_fs *, struct bkey_s, struct bkey_s_c);
 
-enum btree_update_flags {
-	__BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE = __BTREE_ITER_FLAGS_END,
-	__BTREE_UPDATE_NOJOURNAL,
-	__BTREE_UPDATE_KEY_CACHE_RECLAIM,
-
-	__BTREE_TRIGGER_NORUN,
-	__BTREE_TRIGGER_TRANSACTIONAL,
-	__BTREE_TRIGGER_ATOMIC,
-	__BTREE_TRIGGER_GC,
-	__BTREE_TRIGGER_INSERT,
-	__BTREE_TRIGGER_OVERWRITE,
-	__BTREE_TRIGGER_BUCKET_INVALIDATE,
-};
-
-#define BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE (1U << __BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE)
-#define BTREE_UPDATE_NOJOURNAL		(1U << __BTREE_UPDATE_NOJOURNAL)
-#define BTREE_UPDATE_KEY_CACHE_RECLAIM	(1U << __BTREE_UPDATE_KEY_CACHE_RECLAIM)
-
-/* Don't run triggers at all */
-#define BTREE_TRIGGER_NORUN		(1U << __BTREE_TRIGGER_NORUN)
-
-/*
- * If set, we're running transactional triggers as part of a transaction commit:
- * triggers may generate new updates
- *
- * If cleared, and either BTREE_TRIGGER_INSERT|BTREE_TRIGGER_OVERWRITE are set,
- * we're running atomic triggers during a transaction commit: we have our
- * journal reservation, we're holding btree node write locks, and we know the
- * transaction is going to commit (returning an error here is a fatal error,
- * causing us to go emergency read-only)
- */
-#define BTREE_TRIGGER_TRANSACTIONAL	(1U << __BTREE_TRIGGER_TRANSACTIONAL)
-#define BTREE_TRIGGER_ATOMIC		(1U << __BTREE_TRIGGER_ATOMIC)
-
-/* We're in gc/fsck: running triggers to recalculate e.g. disk usage */
-#define BTREE_TRIGGER_GC		(1U << __BTREE_TRIGGER_GC)
-
-/* @new is entering the btree */
-#define BTREE_TRIGGER_INSERT		(1U << __BTREE_TRIGGER_INSERT)
-
-/* @old is leaving the btree */
-#define BTREE_TRIGGER_OVERWRITE		(1U << __BTREE_TRIGGER_OVERWRITE)
-
-/* signal from bucket invalidate path to alloc trigger */
-#define BTREE_TRIGGER_BUCKET_INVALIDATE	(1U << __BTREE_TRIGGER_BUCKET_INVALIDATE)
-
 static inline int bch2_key_trigger(struct btree_trans *trans,
 		enum btree_id btree, unsigned level,
 		struct bkey_s_c old, struct bkey_s new,
-		unsigned flags)
+		enum btree_iter_update_trigger_flags flags)
 {
 	const struct bkey_ops *ops = bch2_bkey_type_ops(old.k->type ?: new.k->type);
 
@@ -135,8 +90,9 @@ static inline int bch2_key_trigger(struct btree_trans *trans,
 }
 
 static inline int bch2_key_trigger_old(struct btree_trans *trans,
-				       enum btree_id btree_id, unsigned level,
-				       struct bkey_s_c old, unsigned flags)
+			enum btree_id btree_id, unsigned level,
+			struct bkey_s_c old,
+			enum btree_iter_update_trigger_flags flags)
 {
 	struct bkey_i deleted;
 
@@ -144,12 +100,13 @@ static inline int bch2_key_trigger_old(struct btree_trans *trans,
 	deleted.k.p = old.k->p;
 
 	return bch2_key_trigger(trans, btree_id, level, old, bkey_i_to_s(&deleted),
-				BTREE_TRIGGER_OVERWRITE|flags);
+				BTREE_TRIGGER_overwrite|flags);
 }
 
 static inline int bch2_key_trigger_new(struct btree_trans *trans,
-				       enum btree_id btree_id, unsigned level,
-				       struct bkey_s new, unsigned flags)
+			enum btree_id btree_id, unsigned level,
+			struct bkey_s new,
+			enum btree_iter_update_trigger_flags flags)
 {
 	struct bkey_i deleted;
 
@@ -157,7 +114,7 @@ static inline int bch2_key_trigger_new(struct btree_trans *trans,
 	deleted.k.p = new.k->p;
 
 	return bch2_key_trigger(trans, btree_id, level, bkey_i_to_s_c(&deleted), new,
-				BTREE_TRIGGER_INSERT|flags);
+				BTREE_TRIGGER_insert|flags);
 }
 
 void bch2_bkey_renumber(enum btree_node_type, struct bkey_packed *, int);
@@ -172,7 +129,8 @@ static inline void bch2_bkey_compat(unsigned level, enum btree_id btree_id,
 			       struct bkey_packed *k)
 {
 	if (version < bcachefs_metadata_version_current ||
-	    big_endian != CPU_BIG_ENDIAN)
+	    big_endian != CPU_BIG_ENDIAN ||
+	    IS_ENABLED(CONFIG_BCACHEFS_DEBUG))
 		__bch2_bkey_compat(level, btree_id, version,
 				   big_endian, write, f, k);
 
diff --git a/fs/bcachefs/bkey_sort.c b/fs/bcachefs/bkey_sort.c
index bcca9e76a0..4536eb50fc 100644
--- a/fs/bcachefs/bkey_sort.c
+++ b/fs/bcachefs/bkey_sort.c
@@ -6,9 +6,9 @@
 #include "bset.h"
 #include "extents.h"
 
-typedef int (*sort_cmp_fn)(struct btree *,
-			   struct bkey_packed *,
-			   struct bkey_packed *);
+typedef int (*sort_cmp_fn)(const struct btree *,
+			   const struct bkey_packed *,
+			   const struct bkey_packed *);
 
 static inline bool sort_iter_end(struct sort_iter *iter)
 {
@@ -70,9 +70,9 @@ static inline struct bkey_packed *sort_iter_next(struct sort_iter *iter,
 /*
  * If keys compare equal, compare by pointer order:
  */
-static inline int key_sort_fix_overlapping_cmp(struct btree *b,
-					       struct bkey_packed *l,
-					       struct bkey_packed *r)
+static inline int key_sort_fix_overlapping_cmp(const struct btree *b,
+					       const struct bkey_packed *l,
+					       const struct bkey_packed *r)
 {
 	return bch2_bkey_cmp_packed(b, l, r) ?:
 		cmp_int((unsigned long) l, (unsigned long) r);
@@ -154,46 +154,59 @@ bch2_sort_repack(struct bset *dst, struct btree *src,
 	return nr;
 }
 
-static inline int sort_keys_cmp(struct btree *b,
-				struct bkey_packed *l,
-				struct bkey_packed *r)
+static inline int keep_unwritten_whiteouts_cmp(const struct btree *b,
+				const struct bkey_packed *l,
+				const struct bkey_packed *r)
 {
 	return bch2_bkey_cmp_packed_inlined(b, l, r) ?:
 		(int) bkey_deleted(r) - (int) bkey_deleted(l) ?:
-		(int) l->needs_whiteout - (int) r->needs_whiteout;
+		(long) l - (long) r;
 }
 
-unsigned bch2_sort_keys(struct bkey_packed *dst,
-			struct sort_iter *iter,
-			bool filter_whiteouts)
+#include "btree_update_interior.h"
+
+/*
+ * For sorting in the btree node write path: whiteouts not in the unwritten
+ * whiteouts area are dropped, whiteouts in the unwritten whiteouts area are
+ * dropped if overwritten by real keys:
+ */
+unsigned bch2_sort_keys_keep_unwritten_whiteouts(struct bkey_packed *dst, struct sort_iter *iter)
 {
-	const struct bkey_format *f = &iter->b->format;
 	struct bkey_packed *in, *next, *out = dst;
 
-	sort_iter_sort(iter, sort_keys_cmp);
+	sort_iter_sort(iter, keep_unwritten_whiteouts_cmp);
 
-	while ((in = sort_iter_next(iter, sort_keys_cmp))) {
-		bool needs_whiteout = false;
+	while ((in = sort_iter_next(iter, keep_unwritten_whiteouts_cmp))) {
+		if (bkey_deleted(in) && in < unwritten_whiteouts_start(iter->b))
+			continue;
 
-		if (bkey_deleted(in) &&
-		    (filter_whiteouts || !in->needs_whiteout))
+		if ((next = sort_iter_peek(iter)) &&
+		    !bch2_bkey_cmp_packed_inlined(iter->b, in, next))
 			continue;
 
-		while ((next = sort_iter_peek(iter)) &&
-		       !bch2_bkey_cmp_packed_inlined(iter->b, in, next)) {
-			BUG_ON(in->needs_whiteout &&
-			       next->needs_whiteout);
-			needs_whiteout |= in->needs_whiteout;
-			in = sort_iter_next(iter, sort_keys_cmp);
-		}
+		bkey_p_copy(out, in);
+		out = bkey_p_next(out);
+	}
 
-		if (bkey_deleted(in)) {
-			memcpy_u64s_small(out, in, bkeyp_key_u64s(f, in));
-			set_bkeyp_val_u64s(f, out, 0);
-		} else {
-			bkey_p_copy(out, in);
-		}
-		out->needs_whiteout |= needs_whiteout;
+	return (u64 *) out - (u64 *) dst;
+}
+
+/*
+ * Main sort routine for compacting a btree node in memory: we always drop
+ * whiteouts because any whiteouts that need to be written are in the unwritten
+ * whiteouts area:
+ */
+unsigned bch2_sort_keys(struct bkey_packed *dst, struct sort_iter *iter)
+{
+	struct bkey_packed *in, *out = dst;
+
+	sort_iter_sort(iter, bch2_bkey_cmp_packed_inlined);
+
+	while ((in = sort_iter_next(iter, bch2_bkey_cmp_packed_inlined))) {
+		if (bkey_deleted(in))
+			continue;
+
+		bkey_p_copy(out, in);
 		out = bkey_p_next(out);
 	}
 
diff --git a/fs/bcachefs/bkey_sort.h b/fs/bcachefs/bkey_sort.h
index 7c0f0b160f..9be969d468 100644
--- a/fs/bcachefs/bkey_sort.h
+++ b/fs/bcachefs/bkey_sort.h
@@ -48,7 +48,7 @@ bch2_sort_repack(struct bset *, struct btree *,
 		 struct btree_node_iter *,
 		 struct bkey_format *, bool);
 
-unsigned bch2_sort_keys(struct bkey_packed *,
-			struct sort_iter *, bool);
+unsigned bch2_sort_keys_keep_unwritten_whiteouts(struct bkey_packed *, struct sort_iter *);
+unsigned bch2_sort_keys(struct bkey_packed *, struct sort_iter *);
 
 #endif /* _BCACHEFS_BKEY_SORT_H */
diff --git a/fs/bcachefs/bkey_types.h b/fs/bcachefs/bkey_types.h
new file mode 100644
index 0000000000..c9ae9e42b3
--- /dev/null
+++ b/fs/bcachefs/bkey_types.h
@@ -0,0 +1,213 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _BCACHEFS_BKEY_TYPES_H
+#define _BCACHEFS_BKEY_TYPES_H
+
+#include "bcachefs_format.h"
+
+/*
+ * bkey_i	- bkey with inline value
+ * bkey_s	- bkey with split value
+ * bkey_s_c	- bkey with split value, const
+ */
+
+#define bkey_p_next(_k)		vstruct_next(_k)
+
+static inline struct bkey_i *bkey_next(struct bkey_i *k)
+{
+	return (struct bkey_i *) ((u64 *) k->_data + k->k.u64s);
+}
+
+#define bkey_val_u64s(_k)	((_k)->u64s - BKEY_U64s)
+
+static inline size_t bkey_val_bytes(const struct bkey *k)
+{
+	return bkey_val_u64s(k) * sizeof(u64);
+}
+
+static inline void set_bkey_val_u64s(struct bkey *k, unsigned val_u64s)
+{
+	unsigned u64s = BKEY_U64s + val_u64s;
+
+	BUG_ON(u64s > U8_MAX);
+	k->u64s = u64s;
+}
+
+static inline void set_bkey_val_bytes(struct bkey *k, unsigned bytes)
+{
+	set_bkey_val_u64s(k, DIV_ROUND_UP(bytes, sizeof(u64)));
+}
+
+#define bkey_val_end(_k)	((void *) (((u64 *) (_k).v) + bkey_val_u64s((_k).k)))
+
+#define bkey_deleted(_k)	((_k)->type == KEY_TYPE_deleted)
+
+#define bkey_whiteout(_k)				\
+	((_k)->type == KEY_TYPE_deleted || (_k)->type == KEY_TYPE_whiteout)
+
+/* bkey with split value, const */
+struct bkey_s_c {
+	const struct bkey	*k;
+	const struct bch_val	*v;
+};
+
+/* bkey with split value */
+struct bkey_s {
+	union {
+	struct {
+		struct bkey	*k;
+		struct bch_val	*v;
+	};
+	struct bkey_s_c		s_c;
+	};
+};
+
+#define bkey_s_null		((struct bkey_s)   { .k = NULL })
+#define bkey_s_c_null		((struct bkey_s_c) { .k = NULL })
+
+#define bkey_s_err(err)		((struct bkey_s)   { .k = ERR_PTR(err) })
+#define bkey_s_c_err(err)	((struct bkey_s_c) { .k = ERR_PTR(err) })
+
+static inline struct bkey_s bkey_to_s(struct bkey *k)
+{
+	return (struct bkey_s) { .k = k, .v = NULL };
+}
+
+static inline struct bkey_s_c bkey_to_s_c(const struct bkey *k)
+{
+	return (struct bkey_s_c) { .k = k, .v = NULL };
+}
+
+static inline struct bkey_s bkey_i_to_s(struct bkey_i *k)
+{
+	return (struct bkey_s) { .k = &k->k, .v = &k->v };
+}
+
+static inline struct bkey_s_c bkey_i_to_s_c(const struct bkey_i *k)
+{
+	return (struct bkey_s_c) { .k = &k->k, .v = &k->v };
+}
+
+/*
+ * For a given type of value (e.g. struct bch_extent), generates the types for
+ * bkey + bch_extent - inline, split, split const - and also all the conversion
+ * functions, which also check that the value is of the correct type.
+ *
+ * We use anonymous unions for upcasting - e.g. converting from e.g. a
+ * bkey_i_extent to a bkey_i - since that's always safe, instead of conversion
+ * functions.
+ */
+#define x(name, ...)					\
+struct bkey_i_##name {							\
+	union {								\
+		struct bkey		k;				\
+		struct bkey_i		k_i;				\
+	};								\
+	struct bch_##name		v;				\
+};									\
+									\
+struct bkey_s_c_##name {						\
+	union {								\
+	struct {							\
+		const struct bkey	*k;				\
+		const struct bch_##name	*v;				\
+	};								\
+	struct bkey_s_c			s_c;				\
+	};								\
+};									\
+									\
+struct bkey_s_##name {							\
+	union {								\
+	struct {							\
+		struct bkey		*k;				\
+		struct bch_##name	*v;				\
+	};								\
+	struct bkey_s_c_##name		c;				\
+	struct bkey_s			s;				\
+	struct bkey_s_c			s_c;				\
+	};								\
+};									\
+									\
+static inline struct bkey_i_##name *bkey_i_to_##name(struct bkey_i *k)	\
+{									\
+	EBUG_ON(!IS_ERR_OR_NULL(k) && k->k.type != KEY_TYPE_##name);	\
+	return container_of(&k->k, struct bkey_i_##name, k);		\
+}									\
+									\
+static inline const struct bkey_i_##name *				\
+bkey_i_to_##name##_c(const struct bkey_i *k)				\
+{									\
+	EBUG_ON(!IS_ERR_OR_NULL(k) && k->k.type != KEY_TYPE_##name);	\
+	return container_of(&k->k, struct bkey_i_##name, k);		\
+}									\
+									\
+static inline struct bkey_s_##name bkey_s_to_##name(struct bkey_s k)	\
+{									\
+	EBUG_ON(!IS_ERR_OR_NULL(k.k) && k.k->type != KEY_TYPE_##name);	\
+	return (struct bkey_s_##name) {					\
+		.k = k.k,						\
+		.v = container_of(k.v, struct bch_##name, v),		\
+	};								\
+}									\
+									\
+static inline struct bkey_s_c_##name bkey_s_c_to_##name(struct bkey_s_c k)\
+{									\
+	EBUG_ON(!IS_ERR_OR_NULL(k.k) && k.k->type != KEY_TYPE_##name);	\
+	return (struct bkey_s_c_##name) {				\
+		.k = k.k,						\
+		.v = container_of(k.v, struct bch_##name, v),		\
+	};								\
+}									\
+									\
+static inline struct bkey_s_##name name##_i_to_s(struct bkey_i_##name *k)\
+{									\
+	return (struct bkey_s_##name) {					\
+		.k = &k->k,						\
+		.v = &k->v,						\
+	};								\
+}									\
+									\
+static inline struct bkey_s_c_##name					\
+name##_i_to_s_c(const struct bkey_i_##name *k)				\
+{									\
+	return (struct bkey_s_c_##name) {				\
+		.k = &k->k,						\
+		.v = &k->v,						\
+	};								\
+}									\
+									\
+static inline struct bkey_s_##name bkey_i_to_s_##name(struct bkey_i *k)	\
+{									\
+	EBUG_ON(!IS_ERR_OR_NULL(k) && k->k.type != KEY_TYPE_##name);	\
+	return (struct bkey_s_##name) {					\
+		.k = &k->k,						\
+		.v = container_of(&k->v, struct bch_##name, v),		\
+	};								\
+}									\
+									\
+static inline struct bkey_s_c_##name					\
+bkey_i_to_s_c_##name(const struct bkey_i *k)				\
+{									\
+	EBUG_ON(!IS_ERR_OR_NULL(k) && k->k.type != KEY_TYPE_##name);	\
+	return (struct bkey_s_c_##name) {				\
+		.k = &k->k,						\
+		.v = container_of(&k->v, struct bch_##name, v),		\
+	};								\
+}									\
+									\
+static inline struct bkey_i_##name *bkey_##name##_init(struct bkey_i *_k)\
+{									\
+	struct bkey_i_##name *k =					\
+		container_of(&_k->k, struct bkey_i_##name, k);		\
+									\
+	bkey_init(&k->k);						\
+	memset(&k->v, 0, sizeof(k->v));					\
+	k->k.type = KEY_TYPE_##name;					\
+	set_bkey_val_bytes(&k->k, sizeof(k->v));			\
+									\
+	return k;							\
+}
+
+BCH_BKEY_TYPES();
+#undef x
+
+#endif /* _BCACHEFS_BKEY_TYPES_H */
diff --git a/fs/bcachefs/bset.c b/fs/bcachefs/bset.c
index 3fd1085b6c..575e1d0b6e 100644
--- a/fs/bcachefs/bset.c
+++ b/fs/bcachefs/bset.c
@@ -103,8 +103,6 @@ void bch2_dump_bset(struct bch_fs *c, struct btree *b,
 
 void bch2_dump_btree_node(struct bch_fs *c, struct btree *b)
 {
-	struct bset_tree *t;
-
 	console_lock();
 	for_each_bset(b, t)
 		bch2_dump_bset(c, b, bset(b, t), t - b->set);
@@ -134,18 +132,23 @@ void bch2_dump_btree_node_iter(struct btree *b,
 	printbuf_exit(&buf);
 }
 
-#ifdef CONFIG_BCACHEFS_DEBUG
-
-void __bch2_verify_btree_nr_keys(struct btree *b)
+struct btree_nr_keys bch2_btree_node_count_keys(struct btree *b)
 {
-	struct bset_tree *t;
 	struct bkey_packed *k;
-	struct btree_nr_keys nr = { 0 };
+	struct btree_nr_keys nr = {};
 
 	for_each_bset(b, t)
 		bset_tree_for_each_key(b, t, k)
 			if (!bkey_deleted(k))
 				btree_keys_account_key_add(&nr, t - b->set, k);
+	return nr;
+}
+
+#ifdef CONFIG_BCACHEFS_DEBUG
+
+void __bch2_verify_btree_nr_keys(struct btree *b)
+{
+	struct btree_nr_keys nr = bch2_btree_node_count_keys(b);
 
 	BUG_ON(memcmp(&nr, &b->nr, sizeof(nr)));
 }
@@ -192,7 +195,6 @@ void bch2_btree_node_iter_verify(struct btree_node_iter *iter,
 {
 	struct btree_node_iter_set *set, *s2;
 	struct bkey_packed *k, *p;
-	struct bset_tree *t;
 
 	if (bch2_btree_node_iter_end(iter))
 		return;
@@ -207,12 +209,14 @@ void bch2_btree_node_iter_verify(struct btree_node_iter *iter,
 	/* Verify that set->end is correct: */
 	btree_node_iter_for_each(iter, set) {
 		for_each_bset(b, t)
-			if (set->end == t->end_offset)
+			if (set->end == t->end_offset) {
+				BUG_ON(set->k < btree_bkey_first_offset(t) ||
+				       set->k >= t->end_offset);
 				goto found;
+			}
 		BUG();
 found:
-		BUG_ON(set->k < btree_bkey_first_offset(t) ||
-		       set->k >= t->end_offset);
+		do {} while (0);
 	}
 
 	/* Verify iterator is sorted: */
@@ -371,11 +375,9 @@ static struct bkey_float *bkey_float(const struct btree *b,
 	return ro_aux_tree_base(b, t)->f + idx;
 }
 
-static void bset_aux_tree_verify(const struct btree *b)
+static void bset_aux_tree_verify(struct btree *b)
 {
 #ifdef CONFIG_BCACHEFS_DEBUG
-	const struct bset_tree *t;
-
 	for_each_bset(b, t) {
 		if (t->aux_data_offset == U16_MAX)
 			continue;
@@ -679,20 +681,20 @@ static __always_inline void make_bfloat(struct btree *b, struct bset_tree *t,
 }
 
 /* bytes remaining - only valid for last bset: */
-static unsigned __bset_tree_capacity(const struct btree *b, const struct bset_tree *t)
+static unsigned __bset_tree_capacity(struct btree *b, const struct bset_tree *t)
 {
 	bset_aux_tree_verify(b);
 
 	return btree_aux_data_bytes(b) - t->aux_data_offset * sizeof(u64);
 }
 
-static unsigned bset_ro_tree_capacity(const struct btree *b, const struct bset_tree *t)
+static unsigned bset_ro_tree_capacity(struct btree *b, const struct bset_tree *t)
 {
 	return __bset_tree_capacity(b, t) /
 		(sizeof(struct bkey_float) + sizeof(u8));
 }
 
-static unsigned bset_rw_tree_capacity(const struct btree *b, const struct bset_tree *t)
+static unsigned bset_rw_tree_capacity(struct btree *b, const struct bset_tree *t)
 {
 	return __bset_tree_capacity(b, t) / sizeof(struct rw_aux_tree);
 }
@@ -1368,8 +1370,6 @@ void bch2_btree_node_iter_init(struct btree_node_iter *iter,
 void bch2_btree_node_iter_init_from_start(struct btree_node_iter *iter,
 					  struct btree *b)
 {
-	struct bset_tree *t;
-
 	memset(iter, 0, sizeof(*iter));
 
 	for_each_bset(b, t)
@@ -1475,7 +1475,6 @@ struct bkey_packed *bch2_btree_node_iter_prev_all(struct btree_node_iter *iter,
 {
 	struct bkey_packed *k, *prev = NULL;
 	struct btree_node_iter_set *set;
-	struct bset_tree *t;
 	unsigned end = 0;
 
 	if (bch2_expensive_debug_checks)
@@ -1544,9 +1543,7 @@ struct bkey_s_c bch2_btree_node_iter_peek_unpack(struct btree_node_iter *iter,
 
 void bch2_btree_keys_stats(const struct btree *b, struct bset_stats *stats)
 {
-	const struct bset_tree *t;
-
-	for_each_bset(b, t) {
+	for_each_bset_c(b, t) {
 		enum bset_aux_tree_type type = bset_aux_tree_type(t);
 		size_t j;
 
diff --git a/fs/bcachefs/bset.h b/fs/bcachefs/bset.h
index 79c77baaa3..5c6c7a14fa 100644
--- a/fs/bcachefs/bset.h
+++ b/fs/bcachefs/bset.h
@@ -206,7 +206,10 @@ static inline size_t btree_aux_data_u64s(const struct btree *b)
 }
 
 #define for_each_bset(_b, _t)						\
-	for (_t = (_b)->set; _t < (_b)->set + (_b)->nsets; _t++)
+	for (struct bset_tree *_t = (_b)->set; _t < (_b)->set + (_b)->nsets; _t++)
+
+#define for_each_bset_c(_b, _t)						\
+	for (const struct bset_tree *_t = (_b)->set; _t < (_b)->set + (_b)->nsets; _t++)
 
 #define bset_tree_for_each_key(_b, _t, _k)				\
 	for (_k = btree_bkey_first(_b, _t);				\
@@ -294,7 +297,6 @@ static inline struct bset_tree *
 bch2_bkey_to_bset_inlined(struct btree *b, struct bkey_packed *k)
 {
 	unsigned offset = __btree_node_key_to_offset(b, k);
-	struct bset_tree *t;
 
 	for_each_bset(b, t)
 		if (offset <= t->end_offset) {
@@ -458,6 +460,8 @@ struct bkey_s_c bch2_btree_node_iter_peek_unpack(struct btree_node_iter *,
 
 /* Accounting: */
 
+struct btree_nr_keys bch2_btree_node_count_keys(struct btree *);
+
 static inline void btree_keys_account_key(struct btree_nr_keys *n,
 					  unsigned bset,
 					  struct bkey_packed *k,
diff --git a/fs/bcachefs/btree_cache.c b/fs/bcachefs/btree_cache.c
index d7c81beac1..4f5e411771 100644
--- a/fs/bcachefs/btree_cache.c
+++ b/fs/bcachefs/btree_cache.c
@@ -1,6 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0
 
 #include "bcachefs.h"
+#include "bbpos.h"
 #include "bkey_buf.h"
 #include "btree_cache.h"
 #include "btree_io.h"
@@ -15,6 +16,12 @@
 #include <linux/prefetch.h>
 #include <linux/sched/mm.h>
 
+#define BTREE_CACHE_NOT_FREED_INCREMENT(counter) \
+do {						 \
+	if (shrinker_counter)			 \
+		bc->not_freed_##counter++;	 \
+} while (0)
+
 const char * const bch2_btree_node_flags[] = {
 #define x(f)	#f,
 	BTREE_FLAGS()
@@ -60,7 +67,7 @@ static void btree_node_data_free(struct bch_fs *c, struct btree *b)
 
 	clear_btree_node_just_written(b);
 
-	kvpfree(b->data, btree_buf_bytes(b));
+	kvfree(b->data);
 	b->data = NULL;
 #ifdef __KERNEL__
 	kvfree(b->aux_data);
@@ -84,17 +91,18 @@ static int bch2_btree_cache_cmp_fn(struct rhashtable_compare_arg *arg,
 }
 
 static const struct rhashtable_params bch_btree_cache_params = {
-	.head_offset	= offsetof(struct btree, hash),
-	.key_offset	= offsetof(struct btree, hash_val),
-	.key_len	= sizeof(u64),
-	.obj_cmpfn	= bch2_btree_cache_cmp_fn,
+	.head_offset		= offsetof(struct btree, hash),
+	.key_offset		= offsetof(struct btree, hash_val),
+	.key_len		= sizeof(u64),
+	.obj_cmpfn		= bch2_btree_cache_cmp_fn,
+	.automatic_shrinking	= true,
 };
 
 static int btree_node_data_alloc(struct bch_fs *c, struct btree *b, gfp_t gfp)
 {
 	BUG_ON(b->data || b->aux_data);
 
-	b->data = kvpmalloc(btree_buf_bytes(b), gfp);
+	b->data = kvmalloc(btree_buf_bytes(b), gfp);
 	if (!b->data)
 		return -BCH_ERR_ENOMEM_btree_node_mem_alloc;
 #ifdef __KERNEL__
@@ -107,7 +115,7 @@ static int btree_node_data_alloc(struct bch_fs *c, struct btree *b, gfp_t gfp)
 		b->aux_data = NULL;
 #endif
 	if (!b->aux_data) {
-		kvpfree(b->data, btree_buf_bytes(b));
+		kvfree(b->data);
 		b->data = NULL;
 		return -BCH_ERR_ENOMEM_btree_node_mem_alloc;
 	}
@@ -161,6 +169,9 @@ void bch2_btree_node_hash_remove(struct btree_cache *bc, struct btree *b)
 
 	/* Cause future lookups for this node to fail: */
 	b->hash_val = 0;
+
+	if (b->c.btree_id < BTREE_ID_NR)
+		--bc->used_by_btree[b->c.btree_id];
 }
 
 int __bch2_btree_node_hash_insert(struct btree_cache *bc, struct btree *b)
@@ -168,8 +179,11 @@ int __bch2_btree_node_hash_insert(struct btree_cache *bc, struct btree *b)
 	BUG_ON(b->hash_val);
 	b->hash_val = btree_ptr_hash_val(&b->key);
 
-	return rhashtable_lookup_insert_fast(&bc->table, &b->hash,
-					     bch_btree_cache_params);
+	int ret = rhashtable_lookup_insert_fast(&bc->table, &b->hash,
+						bch_btree_cache_params);
+	if (!ret && b->c.btree_id < BTREE_ID_NR)
+		bc->used_by_btree[b->c.btree_id]++;
+	return ret;
 }
 
 int bch2_btree_node_hash_insert(struct btree_cache *bc, struct btree *b,
@@ -189,6 +203,35 @@ int bch2_btree_node_hash_insert(struct btree_cache *bc, struct btree *b,
 	return ret;
 }
 
+void bch2_btree_node_update_key_early(struct btree_trans *trans,
+				      enum btree_id btree, unsigned level,
+				      struct bkey_s_c old, struct bkey_i *new)
+{
+	struct bch_fs *c = trans->c;
+	struct btree *b;
+	struct bkey_buf tmp;
+	int ret;
+
+	bch2_bkey_buf_init(&tmp);
+	bch2_bkey_buf_reassemble(&tmp, c, old);
+
+	b = bch2_btree_node_get_noiter(trans, tmp.k, btree, level, true);
+	if (!IS_ERR_OR_NULL(b)) {
+		mutex_lock(&c->btree_cache.lock);
+
+		bch2_btree_node_hash_remove(&c->btree_cache, b);
+
+		bkey_copy(&b->key, new);
+		ret = __bch2_btree_node_hash_insert(&c->btree_cache, b);
+		BUG_ON(ret);
+
+		mutex_unlock(&c->btree_cache.lock);
+		six_unlock_read(&b->c.lock);
+	}
+
+	bch2_bkey_buf_exit(&tmp, c);
+}
+
 __flatten
 static inline struct btree *btree_cache_find(struct btree_cache *bc,
 				     const struct bkey_i *k)
@@ -202,48 +245,86 @@ static inline struct btree *btree_cache_find(struct btree_cache *bc,
  * this version is for btree nodes that have already been freed (we're not
  * reaping a real btree node)
  */
-static int __btree_node_reclaim(struct bch_fs *c, struct btree *b, bool flush)
+static int __btree_node_reclaim(struct bch_fs *c, struct btree *b, bool flush, bool shrinker_counter)
 {
 	struct btree_cache *bc = &c->btree_cache;
 	int ret = 0;
 
 	lockdep_assert_held(&bc->lock);
+
+	struct bbpos pos = BBPOS(b->c.btree_id, b->key.k.p);
+
+	u64 mask = b->c.level
+		? bc->pinned_nodes_interior_mask
+		: bc->pinned_nodes_leaf_mask;
+
+	if ((mask & BIT_ULL(b->c.btree_id)) &&
+	    bbpos_cmp(bc->pinned_nodes_start, pos) < 0 &&
+	    bbpos_cmp(bc->pinned_nodes_end, pos) >= 0)
+		return -BCH_ERR_ENOMEM_btree_node_reclaim;
+
 wait_on_io:
 	if (b->flags & ((1U << BTREE_NODE_dirty)|
 			(1U << BTREE_NODE_read_in_flight)|
 			(1U << BTREE_NODE_write_in_flight))) {
-		if (!flush)
+		if (!flush) {
+			if (btree_node_dirty(b))
+				BTREE_CACHE_NOT_FREED_INCREMENT(dirty);
+			else if (btree_node_read_in_flight(b))
+				BTREE_CACHE_NOT_FREED_INCREMENT(read_in_flight);
+			else if (btree_node_write_in_flight(b))
+				BTREE_CACHE_NOT_FREED_INCREMENT(write_in_flight);
 			return -BCH_ERR_ENOMEM_btree_node_reclaim;
+		}
 
 		/* XXX: waiting on IO with btree cache lock held */
 		bch2_btree_node_wait_on_read(b);
 		bch2_btree_node_wait_on_write(b);
 	}
 
-	if (!six_trylock_intent(&b->c.lock))
+	if (!six_trylock_intent(&b->c.lock)) {
+		BTREE_CACHE_NOT_FREED_INCREMENT(lock_intent);
 		return -BCH_ERR_ENOMEM_btree_node_reclaim;
+	}
 
-	if (!six_trylock_write(&b->c.lock))
+	if (!six_trylock_write(&b->c.lock)) {
+		BTREE_CACHE_NOT_FREED_INCREMENT(lock_write);
 		goto out_unlock_intent;
+	}
 
 	/* recheck under lock */
 	if (b->flags & ((1U << BTREE_NODE_read_in_flight)|
 			(1U << BTREE_NODE_write_in_flight))) {
-		if (!flush)
+		if (!flush) {
+			if (btree_node_read_in_flight(b))
+				BTREE_CACHE_NOT_FREED_INCREMENT(read_in_flight);
+			else if (btree_node_write_in_flight(b))
+				BTREE_CACHE_NOT_FREED_INCREMENT(write_in_flight);
 			goto out_unlock;
+		}
 		six_unlock_write(&b->c.lock);
 		six_unlock_intent(&b->c.lock);
 		goto wait_on_io;
 	}
 
-	if (btree_node_noevict(b) ||
-	    btree_node_write_blocked(b) ||
-	    btree_node_will_make_reachable(b))
+	if (btree_node_noevict(b)) {
+		BTREE_CACHE_NOT_FREED_INCREMENT(noevict);
+		goto out_unlock;
+	}
+	if (btree_node_write_blocked(b)) {
+		BTREE_CACHE_NOT_FREED_INCREMENT(write_blocked);
+		goto out_unlock;
+	}
+	if (btree_node_will_make_reachable(b)) {
+		BTREE_CACHE_NOT_FREED_INCREMENT(will_make_reachable);
 		goto out_unlock;
+	}
 
 	if (btree_node_dirty(b)) {
-		if (!flush)
+		if (!flush) {
+			BTREE_CACHE_NOT_FREED_INCREMENT(dirty);
 			goto out_unlock;
+		}
 		/*
 		 * Using the underscore version because we don't want to compact
 		 * bsets after the write, since this node is about to be evicted
@@ -273,14 +354,14 @@ out_unlock_intent:
 	goto out;
 }
 
-static int btree_node_reclaim(struct bch_fs *c, struct btree *b)
+static int btree_node_reclaim(struct bch_fs *c, struct btree *b, bool shrinker_counter)
 {
-	return __btree_node_reclaim(c, b, false);
+	return __btree_node_reclaim(c, b, false, shrinker_counter);
 }
 
 static int btree_node_write_and_reclaim(struct bch_fs *c, struct btree *b)
 {
-	return __btree_node_reclaim(c, b, true);
+	return __btree_node_reclaim(c, b, true, false);
 }
 
 static unsigned long bch2_btree_cache_scan(struct shrinker *shrink,
@@ -328,11 +409,12 @@ static unsigned long bch2_btree_cache_scan(struct shrinker *shrink,
 		if (touched >= nr)
 			goto out;
 
-		if (!btree_node_reclaim(c, b)) {
+		if (!btree_node_reclaim(c, b, true)) {
 			btree_node_data_free(c, b);
 			six_unlock_write(&b->c.lock);
 			six_unlock_intent(&b->c.lock);
 			freed++;
+			bc->freed++;
 		}
 	}
 restart:
@@ -341,9 +423,11 @@ restart:
 
 		if (btree_node_accessed(b)) {
 			clear_btree_node_accessed(b);
-		} else if (!btree_node_reclaim(c, b)) {
+			bc->not_freed_access_bit++;
+		} else if (!btree_node_reclaim(c, b, true)) {
 			freed++;
 			btree_node_data_free(c, b);
+			bc->freed++;
 
 			bch2_btree_node_hash_remove(bc, b);
 			six_unlock_write(&b->c.lock);
@@ -408,7 +492,7 @@ void bch2_fs_btree_cache_exit(struct bch_fs *c)
 	if (c->verify_data)
 		list_move(&c->verify_data->list, &bc->live);
 
-	kvpfree(c->verify_ondisk, c->opts.btree_node_size);
+	kvfree(c->verify_ondisk);
 
 	for (i = 0; i < btree_id_nr_alive(c); i++) {
 		struct btree_root *r = bch2_btree_id_root(c, i);
@@ -551,7 +635,7 @@ static struct btree *btree_node_cannibalize(struct bch_fs *c)
 	struct btree *b;
 
 	list_for_each_entry_reverse(b, &bc->live, list)
-		if (!btree_node_reclaim(c, b))
+		if (!btree_node_reclaim(c, b, false))
 			return b;
 
 	while (1) {
@@ -587,7 +671,7 @@ struct btree *bch2_btree_node_mem_alloc(struct btree_trans *trans, bool pcpu_rea
 	 * disk node. Check the freed list before allocating a new one:
 	 */
 	list_for_each_entry(b, freed, list)
-		if (!btree_node_reclaim(c, b)) {
+		if (!btree_node_reclaim(c, b, false)) {
 			list_del_init(&b->list);
 			goto got_node;
 		}
@@ -613,7 +697,7 @@ got_node:
 	 * the list. Check if there's any freed nodes there:
 	 */
 	list_for_each_entry(b2, &bc->freeable, list)
-		if (!btree_node_reclaim(c, b2)) {
+		if (!btree_node_reclaim(c, b2, false)) {
 			swap(b->data, b2->data);
 			swap(b->aux_data, b2->aux_data);
 			btree_node_to_freedlist(bc, b2);
@@ -696,9 +780,31 @@ static noinline struct btree *bch2_btree_node_fill(struct btree_trans *trans,
 	struct bch_fs *c = trans->c;
 	struct btree_cache *bc = &c->btree_cache;
 	struct btree *b;
-	u32 seq;
 
-	BUG_ON(level + 1 >= BTREE_MAX_DEPTH);
+	if (unlikely(level >= BTREE_MAX_DEPTH)) {
+		int ret = bch2_fs_topology_error(c, "attempting to get btree node at level %u, >= max depth %u",
+						 level, BTREE_MAX_DEPTH);
+		return ERR_PTR(ret);
+	}
+
+	if (unlikely(!bkey_is_btree_ptr(&k->k))) {
+		struct printbuf buf = PRINTBUF;
+		bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(k));
+
+		int ret = bch2_fs_topology_error(c, "attempting to get btree node with non-btree key %s", buf.buf);
+		printbuf_exit(&buf);
+		return ERR_PTR(ret);
+	}
+
+	if (unlikely(k->k.u64s > BKEY_BTREE_PTR_U64s_MAX)) {
+		struct printbuf buf = PRINTBUF;
+		bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(k));
+
+		int ret = bch2_fs_topology_error(c, "attempting to get btree node with too big key %s", buf.buf);
+		printbuf_exit(&buf);
+		return ERR_PTR(ret);
+	}
+
 	/*
 	 * Parent node must be locked, else we could read in a btree node that's
 	 * been freed:
@@ -711,6 +817,9 @@ static noinline struct btree *bch2_btree_node_fill(struct btree_trans *trans,
 	b = bch2_btree_node_mem_alloc(trans, level != 0);
 
 	if (bch2_err_matches(PTR_ERR_OR_ZERO(b), ENOMEM)) {
+		if (!path)
+			return b;
+
 		trans->memory_allocation_failure = true;
 		trace_and_count(c, trans_restart_memory_allocation_failure, trans, _THIS_IP_, path);
 		return ERR_PTR(btree_trans_restart(trans, BCH_ERR_transaction_restart_fill_mem_alloc_fail));
@@ -736,33 +845,26 @@ static noinline struct btree *bch2_btree_node_fill(struct btree_trans *trans,
 	}
 
 	set_btree_node_read_in_flight(b);
-
 	six_unlock_write(&b->c.lock);
-	seq = six_lock_seq(&b->c.lock);
-	six_unlock_intent(&b->c.lock);
 
-	/* Unlock before doing IO: */
-	if (path && sync)
-		bch2_trans_unlock_noassert(trans);
+	if (path) {
+		u32 seq = six_lock_seq(&b->c.lock);
 
-	bch2_btree_node_read(trans, b, sync);
+		/* Unlock before doing IO: */
+		six_unlock_intent(&b->c.lock);
+		bch2_trans_unlock_noassert(trans);
 
-	if (!sync)
-		return NULL;
+		bch2_btree_node_read(trans, b, sync);
 
-	if (path) {
-		int ret = bch2_trans_relock(trans) ?:
-			bch2_btree_path_relock_intent(trans, path);
-		if (ret) {
-			BUG_ON(!trans->restarted);
-			return ERR_PTR(ret);
-		}
-	}
+		if (!sync)
+			return NULL;
 
-	if (!six_relock_type(&b->c.lock, lock_type, seq)) {
-		if (path)
-			trace_and_count(c, trans_restart_relock_after_fill, trans, _THIS_IP_, path);
-		return ERR_PTR(btree_trans_restart(trans, BCH_ERR_transaction_restart_relock_after_fill));
+		if (!six_relock_type(&b->c.lock, lock_type, seq))
+			b = NULL;
+	} else {
+		bch2_btree_node_read(trans, b, sync);
+		if (lock_type == SIX_LOCK_read)
+			six_lock_downgrade(&b->c.lock);
 	}
 
 	return b;
@@ -791,7 +893,8 @@ static noinline void btree_bad_header(struct bch_fs *c, struct btree *b)
 	prt_printf(&buf, "\nmax ");
 	bch2_bpos_to_text(&buf, b->data->max_key);
 
-	bch2_fs_inconsistent(c, "%s", buf.buf);
+	bch2_fs_topology_error(c, "%s", buf.buf);
+
 	printbuf_exit(&buf);
 }
 
@@ -814,7 +917,6 @@ static struct btree *__bch2_btree_node_get(struct btree_trans *trans, struct btr
 	struct bch_fs *c = trans->c;
 	struct btree_cache *bc = &c->btree_cache;
 	struct btree *b;
-	struct bset_tree *t;
 	bool need_relock = false;
 	int ret;
 
@@ -901,7 +1003,7 @@ retry:
 
 	if (unlikely(btree_node_read_error(b))) {
 		six_unlock_type(&b->c.lock, lock_type);
-		return ERR_PTR(-EIO);
+		return ERR_PTR(-BCH_ERR_btree_node_read_error);
 	}
 
 	EBUG_ON(b->c.btree_id != path->btree_id);
@@ -934,7 +1036,6 @@ struct btree *bch2_btree_node_get(struct btree_trans *trans, struct btree_path *
 {
 	struct bch_fs *c = trans->c;
 	struct btree *b;
-	struct bset_tree *t;
 	int ret;
 
 	EBUG_ON(level >= BTREE_MAX_DEPTH);
@@ -992,7 +1093,7 @@ struct btree *bch2_btree_node_get(struct btree_trans *trans, struct btree_path *
 
 	if (unlikely(btree_node_read_error(b))) {
 		six_unlock_type(&b->c.lock, lock_type);
-		return ERR_PTR(-EIO);
+		return ERR_PTR(-BCH_ERR_btree_node_read_error);
 	}
 
 	EBUG_ON(b->c.btree_id != path->btree_id);
@@ -1011,7 +1112,6 @@ struct btree *bch2_btree_node_get_noiter(struct btree_trans *trans,
 	struct bch_fs *c = trans->c;
 	struct btree_cache *bc = &c->btree_cache;
 	struct btree *b;
-	struct bset_tree *t;
 	int ret;
 
 	EBUG_ON(level >= BTREE_MAX_DEPTH);
@@ -1075,7 +1175,7 @@ lock_node:
 
 	if (unlikely(btree_node_read_error(b))) {
 		six_unlock_read(&b->c.lock);
-		b = ERR_PTR(-EIO);
+		b = ERR_PTR(-BCH_ERR_btree_node_read_error);
 		goto out;
 	}
 
@@ -1094,18 +1194,19 @@ int bch2_btree_node_prefetch(struct btree_trans *trans,
 {
 	struct bch_fs *c = trans->c;
 	struct btree_cache *bc = &c->btree_cache;
-	struct btree *b;
 
-	BUG_ON(trans && !btree_node_locked(path, level + 1));
+	BUG_ON(path && !btree_node_locked(path, level + 1));
 	BUG_ON(level >= BTREE_MAX_DEPTH);
 
-	b = btree_cache_find(bc, k);
+	struct btree *b = btree_cache_find(bc, k);
 	if (b)
 		return 0;
 
 	b = bch2_btree_node_fill(trans, path, k, btree_id,
 				 level, SIX_LOCK_read, false);
-	return PTR_ERR_OR_ZERO(b);
+	if (!IS_ERR_OR_NULL(b))
+		six_unlock_read(&b->c.lock);
+	return bch2_trans_relock(trans) ?: PTR_ERR_OR_ZERO(b);
 }
 
 void bch2_btree_node_evict(struct btree_trans *trans, const struct bkey_i *k)
@@ -1117,6 +1218,8 @@ void bch2_btree_node_evict(struct btree_trans *trans, const struct bkey_i *k)
 	b = btree_cache_find(bc, k);
 	if (!b)
 		return;
+
+	BUG_ON(b == btree_node_root(trans->c, b));
 wait_on_io:
 	/* not allowed to wait on io with btree locks held: */
 
@@ -1128,6 +1231,8 @@ wait_on_io:
 
 	btree_node_lock_nopath_nofail(trans, &b->c, SIX_LOCK_intent);
 	btree_node_lock_nopath_nofail(trans, &b->c, SIX_LOCK_write);
+	if (unlikely(b->hash_val != btree_ptr_hash_val(k)))
+		goto out;
 
 	if (btree_node_dirty(b)) {
 		__bch2_btree_node_write(c, b, BTREE_WRITE_cache_reclaim);
@@ -1142,7 +1247,7 @@ wait_on_io:
 	btree_node_data_free(c, b);
 	bch2_btree_node_hash_remove(bc, b);
 	mutex_unlock(&bc->lock);
-
+out:
 	six_unlock_write(&b->c.lock);
 	six_unlock_intent(&b->c.lock);
 }
@@ -1203,9 +1308,39 @@ void bch2_btree_node_to_text(struct printbuf *out, struct bch_fs *c, const struc
 	       stats.failed);
 }
 
-void bch2_btree_cache_to_text(struct printbuf *out, const struct bch_fs *c)
+static void prt_btree_cache_line(struct printbuf *out, const struct bch_fs *c,
+				 const char *label, unsigned nr)
 {
-	prt_printf(out, "nr nodes:\t\t%u\n", c->btree_cache.used);
-	prt_printf(out, "nr dirty:\t\t%u\n", atomic_read(&c->btree_cache.dirty));
-	prt_printf(out, "cannibalize lock:\t%p\n", c->btree_cache.alloc_lock);
+	prt_printf(out, "%s\t", label);
+	prt_human_readable_u64(out, nr * c->opts.btree_node_size);
+	prt_printf(out, " (%u)\n", nr);
+}
+
+void bch2_btree_cache_to_text(struct printbuf *out, const struct btree_cache *bc)
+{
+	struct bch_fs *c = container_of(bc, struct bch_fs, btree_cache);
+
+	if (!out->nr_tabstops)
+		printbuf_tabstop_push(out, 32);
+
+	prt_btree_cache_line(out, c, "total:",		bc->used);
+	prt_btree_cache_line(out, c, "nr dirty:",	atomic_read(&bc->dirty));
+	prt_printf(out, "cannibalize lock:\t%p\n",	bc->alloc_lock);
+	prt_newline(out);
+
+	for (unsigned i = 0; i < ARRAY_SIZE(bc->used_by_btree); i++)
+		prt_btree_cache_line(out, c, bch2_btree_id_str(i), bc->used_by_btree[i]);
+
+	prt_newline(out);
+	prt_printf(out, "freed:\t%u\n", bc->freed);
+	prt_printf(out, "not freed:\n");
+	prt_printf(out, "  dirty\t%u\n", bc->not_freed_dirty);
+	prt_printf(out, "  write in flight\t%u\n", bc->not_freed_write_in_flight);
+	prt_printf(out, "  read in flight\t%u\n", bc->not_freed_read_in_flight);
+	prt_printf(out, "  lock intent failed\t%u\n", bc->not_freed_lock_intent);
+	prt_printf(out, "  lock write failed\t%u\n", bc->not_freed_lock_write);
+	prt_printf(out, "  access bit\t%u\n", bc->not_freed_access_bit);
+	prt_printf(out, "  no evict failed\t%u\n", bc->not_freed_noevict);
+	prt_printf(out, "  write blocked\t%u\n", bc->not_freed_write_blocked);
+	prt_printf(out, "  will make reachable\t%u\n", bc->not_freed_will_make_reachable);
 }
diff --git a/fs/bcachefs/btree_cache.h b/fs/bcachefs/btree_cache.h
index 6d33885fdb..fed35de3e4 100644
--- a/fs/bcachefs/btree_cache.h
+++ b/fs/bcachefs/btree_cache.h
@@ -17,6 +17,9 @@ int __bch2_btree_node_hash_insert(struct btree_cache *, struct btree *);
 int bch2_btree_node_hash_insert(struct btree_cache *, struct btree *,
 				unsigned, enum btree_id);
 
+void bch2_btree_node_update_key_early(struct btree_trans *, enum btree_id, unsigned,
+				      struct bkey_s_c, struct bkey_i *);
+
 void bch2_btree_cache_cannibalize_unlock(struct btree_trans *);
 int bch2_btree_cache_cannibalize_lock(struct btree_trans *, struct closure *);
 
@@ -131,6 +134,6 @@ static inline struct btree *btree_node_root(struct bch_fs *c, struct btree *b)
 const char *bch2_btree_id_str(enum btree_id);
 void bch2_btree_pos_to_text(struct printbuf *, struct bch_fs *, const struct btree *);
 void bch2_btree_node_to_text(struct printbuf *, struct bch_fs *, const struct btree *);
-void bch2_btree_cache_to_text(struct printbuf *, const struct bch_fs *);
+void bch2_btree_cache_to_text(struct printbuf *, const struct btree_cache *);
 
 #endif /* _BCACHEFS_BTREE_CACHE_H */
diff --git a/fs/bcachefs/btree_gc.c b/fs/bcachefs/btree_gc.c
index 1102995643..a0deb82660 100644
--- a/fs/bcachefs/btree_gc.c
+++ b/fs/bcachefs/btree_gc.c
@@ -7,11 +7,13 @@
 #include "bcachefs.h"
 #include "alloc_background.h"
 #include "alloc_foreground.h"
+#include "backpointers.h"
 #include "bkey_methods.h"
 #include "bkey_buf.h"
 #include "btree_journal_iter.h"
 #include "btree_key_cache.h"
 #include "btree_locking.h"
+#include "btree_node_scan.h"
 #include "btree_update_interior.h"
 #include "btree_io.h"
 #include "btree_gc.h"
@@ -24,7 +26,7 @@
 #include "journal.h"
 #include "keylist.h"
 #include "move.h"
-#include "recovery.h"
+#include "recovery_passes.h"
 #include "reflink.h"
 #include "replicas.h"
 #include "super-io.h"
@@ -40,6 +42,7 @@
 
 #define DROP_THIS_NODE		10
 #define DROP_PREV_NODE		11
+#define DID_FILL_FROM_SCAN	12
 
 static struct bkey_s unsafe_bkey_s_c_to_s(struct bkey_s_c k)
 {
@@ -49,12 +52,6 @@ static struct bkey_s unsafe_bkey_s_c_to_s(struct bkey_s_c k)
 	}}};
 }
 
-static bool should_restart_for_topology_repair(struct bch_fs *c)
-{
-	return c->opts.fix_errors != FSCK_FIX_no &&
-		!(c->recovery_passes_complete & BIT_ULL(BCH_RECOVERY_PASS_check_topology));
-}
-
 static inline void __gc_pos_set(struct bch_fs *c, struct gc_pos new_pos)
 {
 	preempt_disable();
@@ -66,94 +63,10 @@ static inline void __gc_pos_set(struct bch_fs *c, struct gc_pos new_pos)
 
 static inline void gc_pos_set(struct bch_fs *c, struct gc_pos new_pos)
 {
-	BUG_ON(gc_pos_cmp(new_pos, c->gc_pos) <= 0);
+	BUG_ON(gc_pos_cmp(new_pos, c->gc_pos) < 0);
 	__gc_pos_set(c, new_pos);
 }
 
-/*
- * Missing: if an interior btree node is empty, we need to do something -
- * perhaps just kill it
- */
-static int bch2_gc_check_topology(struct bch_fs *c,
-				  struct btree *b,
-				  struct bkey_buf *prev,
-				  struct bkey_buf cur,
-				  bool is_last)
-{
-	struct bpos node_start	= b->data->min_key;
-	struct bpos node_end	= b->data->max_key;
-	struct bpos expected_start = bkey_deleted(&prev->k->k)
-		? node_start
-		: bpos_successor(prev->k->k.p);
-	struct printbuf buf1 = PRINTBUF, buf2 = PRINTBUF;
-	int ret = 0;
-
-	if (cur.k->k.type == KEY_TYPE_btree_ptr_v2) {
-		struct bkey_i_btree_ptr_v2 *bp = bkey_i_to_btree_ptr_v2(cur.k);
-
-		if (!bpos_eq(expected_start, bp->v.min_key)) {
-			bch2_topology_error(c);
-
-			if (bkey_deleted(&prev->k->k)) {
-				prt_printf(&buf1, "start of node: ");
-				bch2_bpos_to_text(&buf1, node_start);
-			} else {
-				bch2_bkey_val_to_text(&buf1, c, bkey_i_to_s_c(prev->k));
-			}
-			bch2_bkey_val_to_text(&buf2, c, bkey_i_to_s_c(cur.k));
-
-			if (__fsck_err(c,
-				       FSCK_CAN_FIX|
-				       FSCK_CAN_IGNORE|
-				       FSCK_NO_RATELIMIT,
-				       btree_node_topology_bad_min_key,
-				       "btree node with incorrect min_key at btree %s level %u:\n"
-				       "  prev %s\n"
-				       "  cur %s",
-				       bch2_btree_id_str(b->c.btree_id), b->c.level,
-				       buf1.buf, buf2.buf) && should_restart_for_topology_repair(c)) {
-				bch_info(c, "Halting mark and sweep to start topology repair pass");
-				ret = bch2_run_explicit_recovery_pass(c, BCH_RECOVERY_PASS_check_topology);
-				goto err;
-			} else {
-				set_bit(BCH_FS_initial_gc_unfixed, &c->flags);
-			}
-		}
-	}
-
-	if (is_last && !bpos_eq(cur.k->k.p, node_end)) {
-		bch2_topology_error(c);
-
-		printbuf_reset(&buf1);
-		printbuf_reset(&buf2);
-
-		bch2_bkey_val_to_text(&buf1, c, bkey_i_to_s_c(cur.k));
-		bch2_bpos_to_text(&buf2, node_end);
-
-		if (__fsck_err(c, FSCK_CAN_FIX|FSCK_CAN_IGNORE|FSCK_NO_RATELIMIT,
-			  btree_node_topology_bad_max_key,
-			  "btree node with incorrect max_key at btree %s level %u:\n"
-			  "  %s\n"
-			  "  expected %s",
-			  bch2_btree_id_str(b->c.btree_id), b->c.level,
-			  buf1.buf, buf2.buf) &&
-		    should_restart_for_topology_repair(c)) {
-			bch_info(c, "Halting mark and sweep to start topology repair pass");
-			ret = bch2_run_explicit_recovery_pass(c, BCH_RECOVERY_PASS_check_topology);
-			goto err;
-		} else {
-			set_bit(BCH_FS_initial_gc_unfixed, &c->flags);
-		}
-	}
-
-	bch2_bkey_buf_copy(prev, c, cur.k);
-err:
-fsck_err:
-	printbuf_exit(&buf2);
-	printbuf_exit(&buf1);
-	return ret;
-}
-
 static void btree_ptr_to_v2(struct btree *b, struct bkey_i_btree_ptr_v2 *dst)
 {
 	switch (b->key.k.type) {
@@ -178,40 +91,22 @@ static void btree_ptr_to_v2(struct btree *b, struct bkey_i_btree_ptr_v2 *dst)
 	}
 }
 
-static void bch2_btree_node_update_key_early(struct btree_trans *trans,
-					     enum btree_id btree, unsigned level,
-					     struct bkey_s_c old, struct bkey_i *new)
+static int set_node_min(struct bch_fs *c, struct btree *b, struct bpos new_min)
 {
-	struct bch_fs *c = trans->c;
-	struct btree *b;
-	struct bkey_buf tmp;
+	struct bkey_i_btree_ptr_v2 *new;
 	int ret;
 
-	bch2_bkey_buf_init(&tmp);
-	bch2_bkey_buf_reassemble(&tmp, c, old);
-
-	b = bch2_btree_node_get_noiter(trans, tmp.k, btree, level, true);
-	if (!IS_ERR_OR_NULL(b)) {
-		mutex_lock(&c->btree_cache.lock);
+	if (c->opts.verbose) {
+		struct printbuf buf = PRINTBUF;
 
-		bch2_btree_node_hash_remove(&c->btree_cache, b);
+		bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&b->key));
+		prt_str(&buf, " -> ");
+		bch2_bpos_to_text(&buf, new_min);
 
-		bkey_copy(&b->key, new);
-		ret = __bch2_btree_node_hash_insert(&c->btree_cache, b);
-		BUG_ON(ret);
-
-		mutex_unlock(&c->btree_cache.lock);
-		six_unlock_read(&b->c.lock);
+		bch_info(c, "%s(): %s", __func__, buf.buf);
+		printbuf_exit(&buf);
 	}
 
-	bch2_bkey_buf_exit(&tmp, c);
-}
-
-static int set_node_min(struct bch_fs *c, struct btree *b, struct bpos new_min)
-{
-	struct bkey_i_btree_ptr_v2 *new;
-	int ret;
-
 	new = kmalloc_array(BKEY_BTREE_PTR_U64s_MAX, sizeof(u64), GFP_KERNEL);
 	if (!new)
 		return -BCH_ERR_ENOMEM_gc_repair_key;
@@ -237,6 +132,17 @@ static int set_node_max(struct bch_fs *c, struct btree *b, struct bpos new_max)
 	struct bkey_i_btree_ptr_v2 *new;
 	int ret;
 
+	if (c->opts.verbose) {
+		struct printbuf buf = PRINTBUF;
+
+		bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&b->key));
+		prt_str(&buf, " -> ");
+		bch2_bpos_to_text(&buf, new_max);
+
+		bch_info(c, "%s(): %s", __func__, buf.buf);
+		printbuf_exit(&buf);
+	}
+
 	ret = bch2_journal_key_delete(c, b->c.btree_id, b->c.level + 1, b->key.k.p);
 	if (ret)
 		return ret;
@@ -268,128 +174,140 @@ static int set_node_max(struct bch_fs *c, struct btree *b, struct bpos new_max)
 	return 0;
 }
 
-static int btree_repair_node_boundaries(struct bch_fs *c, struct btree *b,
-					struct btree *prev, struct btree *cur)
+static int btree_check_node_boundaries(struct bch_fs *c, struct btree *b,
+				       struct btree *prev, struct btree *cur,
+				       struct bpos *pulled_from_scan)
 {
 	struct bpos expected_start = !prev
 		? b->data->min_key
 		: bpos_successor(prev->key.k.p);
-	struct printbuf buf1 = PRINTBUF, buf2 = PRINTBUF;
+	struct printbuf buf = PRINTBUF;
 	int ret = 0;
 
-	if (!prev) {
-		prt_printf(&buf1, "start of node: ");
-		bch2_bpos_to_text(&buf1, b->data->min_key);
-	} else {
-		bch2_bkey_val_to_text(&buf1, c, bkey_i_to_s_c(&prev->key));
+	BUG_ON(b->key.k.type == KEY_TYPE_btree_ptr_v2 &&
+	       !bpos_eq(bkey_i_to_btree_ptr_v2(&b->key)->v.min_key,
+			b->data->min_key));
+
+	if (bpos_eq(expected_start, cur->data->min_key))
+		return 0;
+
+	prt_printf(&buf, "  at btree %s level %u:\n  parent: ",
+		   bch2_btree_id_str(b->c.btree_id), b->c.level);
+	bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&b->key));
+
+	if (prev) {
+		prt_printf(&buf, "\n  prev: ");
+		bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&prev->key));
 	}
 
-	bch2_bkey_val_to_text(&buf2, c, bkey_i_to_s_c(&cur->key));
-
-	if (prev &&
-	    bpos_gt(expected_start, cur->data->min_key) &&
-	    BTREE_NODE_SEQ(cur->data) > BTREE_NODE_SEQ(prev->data)) {
-		/* cur overwrites prev: */
-
-		if (mustfix_fsck_err_on(bpos_ge(prev->data->min_key,
-						cur->data->min_key), c,
-				btree_node_topology_overwritten_by_next_node,
-				"btree node overwritten by next node at btree %s level %u:\n"
-				"  node %s\n"
-				"  next %s",
-				bch2_btree_id_str(b->c.btree_id), b->c.level,
-				buf1.buf, buf2.buf)) {
-			ret = DROP_PREV_NODE;
-			goto out;
-		}
+	prt_str(&buf, "\n  next: ");
+	bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&cur->key));
 
-		if (mustfix_fsck_err_on(!bpos_eq(prev->key.k.p,
-						 bpos_predecessor(cur->data->min_key)), c,
-				btree_node_topology_bad_max_key,
-				"btree node with incorrect max_key at btree %s level %u:\n"
-				"  node %s\n"
-				"  next %s",
-				bch2_btree_id_str(b->c.btree_id), b->c.level,
-				buf1.buf, buf2.buf))
-			ret = set_node_max(c, prev,
-					   bpos_predecessor(cur->data->min_key));
-	} else {
-		/* prev overwrites cur: */
-
-		if (mustfix_fsck_err_on(bpos_ge(expected_start,
-						cur->data->max_key), c,
-				btree_node_topology_overwritten_by_prev_node,
-				"btree node overwritten by prev node at btree %s level %u:\n"
-				"  prev %s\n"
-				"  node %s",
-				bch2_btree_id_str(b->c.btree_id), b->c.level,
-				buf1.buf, buf2.buf)) {
-			ret = DROP_THIS_NODE;
-			goto out;
-		}
+	if (bpos_lt(expected_start, cur->data->min_key)) {				/* gap */
+		if (b->c.level == 1 &&
+		    bpos_lt(*pulled_from_scan, cur->data->min_key)) {
+			ret = bch2_get_scanned_nodes(c, b->c.btree_id, 0,
+						     expected_start,
+						     bpos_predecessor(cur->data->min_key));
+			if (ret)
+				goto err;
 
-		if (mustfix_fsck_err_on(!bpos_eq(expected_start, cur->data->min_key), c,
-				btree_node_topology_bad_min_key,
-				"btree node with incorrect min_key at btree %s level %u:\n"
-				"  prev %s\n"
-				"  node %s",
-				bch2_btree_id_str(b->c.btree_id), b->c.level,
-				buf1.buf, buf2.buf))
-			ret = set_node_min(c, cur, expected_start);
+			*pulled_from_scan = cur->data->min_key;
+			ret = DID_FILL_FROM_SCAN;
+		} else {
+			if (mustfix_fsck_err(c, btree_node_topology_bad_min_key,
+					     "btree node with incorrect min_key%s", buf.buf))
+				ret = set_node_min(c, cur, expected_start);
+		}
+	} else {									/* overlap */
+		if (prev && BTREE_NODE_SEQ(cur->data) > BTREE_NODE_SEQ(prev->data)) {	/* cur overwrites prev */
+			if (bpos_ge(prev->data->min_key, cur->data->min_key)) {		/* fully? */
+				if (mustfix_fsck_err(c, btree_node_topology_overwritten_by_next_node,
+						     "btree node overwritten by next node%s", buf.buf))
+					ret = DROP_PREV_NODE;
+			} else {
+				if (mustfix_fsck_err(c, btree_node_topology_bad_max_key,
+						     "btree node with incorrect max_key%s", buf.buf))
+					ret = set_node_max(c, prev,
+							   bpos_predecessor(cur->data->min_key));
+			}
+		} else {
+			if (bpos_ge(expected_start, cur->data->max_key)) {		/* fully? */
+				if (mustfix_fsck_err(c, btree_node_topology_overwritten_by_prev_node,
+						     "btree node overwritten by prev node%s", buf.buf))
+					ret = DROP_THIS_NODE;
+			} else {
+				if (mustfix_fsck_err(c, btree_node_topology_bad_min_key,
+						     "btree node with incorrect min_key%s", buf.buf))
+					ret = set_node_min(c, cur, expected_start);
+			}
+		}
 	}
-out:
+err:
 fsck_err:
-	printbuf_exit(&buf2);
-	printbuf_exit(&buf1);
+	printbuf_exit(&buf);
 	return ret;
 }
 
 static int btree_repair_node_end(struct bch_fs *c, struct btree *b,
-				 struct btree *child)
+				 struct btree *child, struct bpos *pulled_from_scan)
 {
-	struct printbuf buf1 = PRINTBUF, buf2 = PRINTBUF;
+	struct printbuf buf = PRINTBUF;
 	int ret = 0;
 
-	bch2_bkey_val_to_text(&buf1, c, bkey_i_to_s_c(&child->key));
-	bch2_bpos_to_text(&buf2, b->key.k.p);
+	if (bpos_eq(child->key.k.p, b->key.k.p))
+		return 0;
 
-	if (mustfix_fsck_err_on(!bpos_eq(child->key.k.p, b->key.k.p), c,
-				btree_node_topology_bad_max_key,
-			"btree node with incorrect max_key at btree %s level %u:\n"
-			"  %s\n"
-			"  expected %s",
-			bch2_btree_id_str(b->c.btree_id), b->c.level,
-			buf1.buf, buf2.buf)) {
-		ret = set_node_max(c, child, b->key.k.p);
-		if (ret)
-			goto err;
+	prt_printf(&buf, "at btree %s level %u:\n  parent: ",
+		   bch2_btree_id_str(b->c.btree_id), b->c.level);
+	bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&b->key));
+
+	prt_str(&buf, "\n  child: ");
+	bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&child->key));
+
+	if (mustfix_fsck_err(c, btree_node_topology_bad_max_key,
+			     "btree node with incorrect max_key%s", buf.buf)) {
+		if (b->c.level == 1 &&
+		    bpos_lt(*pulled_from_scan, b->key.k.p)) {
+			ret = bch2_get_scanned_nodes(c, b->c.btree_id, 0,
+						bpos_successor(child->key.k.p), b->key.k.p);
+			if (ret)
+				goto err;
+
+			*pulled_from_scan = b->key.k.p;
+			ret = DID_FILL_FROM_SCAN;
+		} else {
+			ret = set_node_max(c, child, b->key.k.p);
+		}
 	}
 err:
 fsck_err:
-	printbuf_exit(&buf2);
-	printbuf_exit(&buf1);
+	printbuf_exit(&buf);
 	return ret;
 }
 
-static int bch2_btree_repair_topology_recurse(struct btree_trans *trans, struct btree *b)
+static int bch2_btree_repair_topology_recurse(struct btree_trans *trans, struct btree *b,
+					      struct bpos *pulled_from_scan)
 {
 	struct bch_fs *c = trans->c;
 	struct btree_and_journal_iter iter;
 	struct bkey_s_c k;
 	struct bkey_buf prev_k, cur_k;
 	struct btree *prev = NULL, *cur = NULL;
-	bool have_child, dropped_children = false;
+	bool have_child, new_pass = false;
 	struct printbuf buf = PRINTBUF;
 	int ret = 0;
 
 	if (!b->c.level)
 		return 0;
-again:
-	prev = NULL;
-	have_child = dropped_children = false;
+
 	bch2_bkey_buf_init(&prev_k);
 	bch2_bkey_buf_init(&cur_k);
-	bch2_btree_and_journal_iter_init_node_iter(&iter, c, b);
+again:
+	cur = prev = NULL;
+	have_child = new_pass = false;
+	bch2_btree_and_journal_iter_init_node_iter(trans, &iter, b);
+	iter.prefetch = true;
 
 	while ((k = bch2_btree_and_journal_iter_peek(&iter)).k) {
 		BUG_ON(bpos_lt(k.k->p, b->data->min_key));
@@ -406,7 +324,7 @@ again:
 		printbuf_reset(&buf);
 		bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(cur_k.k));
 
-		if (mustfix_fsck_err_on(ret == -EIO, c,
+		if (mustfix_fsck_err_on(bch2_err_matches(ret, EIO), c,
 				btree_node_unreadable,
 				"Topology repair: unreadable btree node at btree %s level %u:\n"
 				"  %s",
@@ -414,11 +332,17 @@ again:
 				b->c.level - 1,
 				buf.buf)) {
 			bch2_btree_node_evict(trans, cur_k.k);
+			cur = NULL;
 			ret = bch2_journal_key_delete(c, b->c.btree_id,
 						      b->c.level, cur_k.k->k.p);
-			cur = NULL;
 			if (ret)
 				break;
+
+			if (!btree_id_is_alloc(b->c.btree_id)) {
+				ret = bch2_run_explicit_recovery_pass(c, BCH_RECOVERY_PASS_scan_for_btree_nodes);
+				if (ret)
+					break;
+			}
 			continue;
 		}
 
@@ -426,7 +350,23 @@ again:
 		if (ret)
 			break;
 
-		ret = btree_repair_node_boundaries(c, b, prev, cur);
+		if (bch2_btree_node_is_stale(c, cur)) {
+			bch_info(c, "btree node %s older than nodes found by scanning", buf.buf);
+			six_unlock_read(&cur->c.lock);
+			bch2_btree_node_evict(trans, cur_k.k);
+			ret = bch2_journal_key_delete(c, b->c.btree_id,
+						      b->c.level, cur_k.k->k.p);
+			cur = NULL;
+			if (ret)
+				break;
+			continue;
+		}
+
+		ret = btree_check_node_boundaries(c, b, prev, cur, pulled_from_scan);
+		if (ret == DID_FILL_FROM_SCAN) {
+			new_pass = true;
+			ret = 0;
+		}
 
 		if (ret == DROP_THIS_NODE) {
 			six_unlock_read(&cur->c.lock);
@@ -444,6 +384,7 @@ again:
 		prev = NULL;
 
 		if (ret == DROP_PREV_NODE) {
+			bch_info(c, "dropped prev node");
 			bch2_btree_node_evict(trans, prev_k.k);
 			ret = bch2_journal_key_delete(c, b->c.btree_id,
 						      b->c.level, prev_k.k->k.p);
@@ -451,8 +392,6 @@ again:
 				break;
 
 			bch2_btree_and_journal_iter_exit(&iter);
-			bch2_bkey_buf_exit(&prev_k, c);
-			bch2_bkey_buf_exit(&cur_k, c);
 			goto again;
 		} else if (ret)
 			break;
@@ -464,7 +403,11 @@ again:
 
 	if (!ret && !IS_ERR_OR_NULL(prev)) {
 		BUG_ON(cur);
-		ret = btree_repair_node_end(c, b, prev);
+		ret = btree_repair_node_end(c, b, prev, pulled_from_scan);
+		if (ret == DID_FILL_FROM_SCAN) {
+			new_pass = true;
+			ret = 0;
+		}
 	}
 
 	if (!IS_ERR_OR_NULL(prev))
@@ -478,7 +421,12 @@ again:
 		goto err;
 
 	bch2_btree_and_journal_iter_exit(&iter);
-	bch2_btree_and_journal_iter_init_node_iter(&iter, c, b);
+
+	if (new_pass)
+		goto again;
+
+	bch2_btree_and_journal_iter_init_node_iter(trans, &iter, b);
+	iter.prefetch = true;
 
 	while ((k = bch2_btree_and_journal_iter_peek(&iter)).k) {
 		bch2_bkey_buf_reassemble(&cur_k, c, k);
@@ -493,7 +441,7 @@ again:
 		if (ret)
 			goto err;
 
-		ret = bch2_btree_repair_topology_recurse(trans, cur);
+		ret = bch2_btree_repair_topology_recurse(trans, cur, pulled_from_scan);
 		six_unlock_read(&cur->c.lock);
 		cur = NULL;
 
@@ -501,7 +449,7 @@ again:
 			bch2_btree_node_evict(trans, cur_k.k);
 			ret = bch2_journal_key_delete(c, b->c.btree_id,
 						      b->c.level, cur_k.k->k.p);
-			dropped_children = true;
+			new_pass = true;
 		}
 
 		if (ret)
@@ -528,12 +476,14 @@ fsck_err:
 		six_unlock_read(&cur->c.lock);
 
 	bch2_btree_and_journal_iter_exit(&iter);
-	bch2_bkey_buf_exit(&prev_k, c);
-	bch2_bkey_buf_exit(&cur_k, c);
 
-	if (!ret && dropped_children)
+	if (!ret && new_pass)
 		goto again;
 
+	BUG_ON(!ret && bch2_btree_node_check_topology(trans, b));
+
+	bch2_bkey_buf_exit(&prev_k, c);
+	bch2_bkey_buf_exit(&cur_k, c);
 	printbuf_exit(&buf);
 	return ret;
 }
@@ -541,547 +491,207 @@ fsck_err:
 int bch2_check_topology(struct bch_fs *c)
 {
 	struct btree_trans *trans = bch2_trans_get(c);
-	struct btree *b;
-	unsigned i;
+	struct bpos pulled_from_scan = POS_MIN;
 	int ret = 0;
 
-	for (i = 0; i < btree_id_nr_alive(c) && !ret; i++) {
+	for (unsigned i = 0; i < btree_id_nr_alive(c) && !ret; i++) {
 		struct btree_root *r = bch2_btree_id_root(c, i);
+		bool reconstructed_root = false;
 
-		if (!r->alive)
-			continue;
-
-		b = r->b;
-		if (btree_node_fake(b))
-			continue;
-
-		btree_node_lock_nopath_nofail(trans, &b->c, SIX_LOCK_read);
-		ret = bch2_btree_repair_topology_recurse(trans, b);
-		six_unlock_read(&b->c.lock);
-
-		if (ret == DROP_THIS_NODE) {
-			bch_err(c, "empty btree root - repair unimplemented");
-			ret = -BCH_ERR_fsck_repair_unimplemented;
-		}
-	}
-
-	bch2_trans_put(trans);
-
-	return ret;
-}
-
-static int bch2_check_fix_ptrs(struct btree_trans *trans, enum btree_id btree_id,
-			       unsigned level, bool is_root,
-			       struct bkey_s_c *k)
-{
-	struct bch_fs *c = trans->c;
-	struct bkey_ptrs_c ptrs_c = bch2_bkey_ptrs_c(*k);
-	const union bch_extent_entry *entry_c;
-	struct extent_ptr_decoded p = { 0 };
-	bool do_update = false;
-	struct printbuf buf = PRINTBUF;
-	int ret = 0;
-
-	/*
-	 * XXX
-	 * use check_bucket_ref here
-	 */
-	bkey_for_each_ptr_decode(k->k, ptrs_c, p, entry_c) {
-		struct bch_dev *ca = bch_dev_bkey_exists(c, p.ptr.dev);
-		struct bucket *g = PTR_GC_BUCKET(ca, &p.ptr);
-		enum bch_data_type data_type = bch2_bkey_ptr_data_type(*k, &entry_c->ptr);
-
-		if (!g->gen_valid &&
-		    (c->opts.reconstruct_alloc ||
-		     fsck_err(c, ptr_to_missing_alloc_key,
-			      "bucket %u:%zu data type %s ptr gen %u missing in alloc btree\n"
-			      "while marking %s",
-			      p.ptr.dev, PTR_BUCKET_NR(ca, &p.ptr),
-			      bch2_data_type_str(ptr_data_type(k->k, &p.ptr)),
-			      p.ptr.gen,
-			      (printbuf_reset(&buf),
-			       bch2_bkey_val_to_text(&buf, c, *k), buf.buf)))) {
-			if (!p.ptr.cached) {
-				g->gen_valid		= true;
-				g->gen			= p.ptr.gen;
-			} else {
-				do_update = true;
-			}
-		}
-
-		if (gen_cmp(p.ptr.gen, g->gen) > 0 &&
-		    (c->opts.reconstruct_alloc ||
-		     fsck_err(c, ptr_gen_newer_than_bucket_gen,
-			      "bucket %u:%zu data type %s ptr gen in the future: %u > %u\n"
-			      "while marking %s",
-			      p.ptr.dev, PTR_BUCKET_NR(ca, &p.ptr),
-			      bch2_data_type_str(ptr_data_type(k->k, &p.ptr)),
-			      p.ptr.gen, g->gen,
-			      (printbuf_reset(&buf),
-			       bch2_bkey_val_to_text(&buf, c, *k), buf.buf)))) {
-			if (!p.ptr.cached) {
-				g->gen_valid		= true;
-				g->gen			= p.ptr.gen;
-				g->data_type		= 0;
-				g->dirty_sectors	= 0;
-				g->cached_sectors	= 0;
-				set_bit(BCH_FS_need_another_gc, &c->flags);
-			} else {
-				do_update = true;
-			}
-		}
+		if (r->error) {
+			ret = bch2_run_explicit_recovery_pass(c, BCH_RECOVERY_PASS_scan_for_btree_nodes);
+			if (ret)
+				break;
+reconstruct_root:
+			bch_info(c, "btree root %s unreadable, must recover from scan", bch2_btree_id_str(i));
 
-		if (gen_cmp(g->gen, p.ptr.gen) > BUCKET_GC_GEN_MAX &&
-		    (c->opts.reconstruct_alloc ||
-		     fsck_err(c, ptr_gen_newer_than_bucket_gen,
-			      "bucket %u:%zu gen %u data type %s: ptr gen %u too stale\n"
-			      "while marking %s",
-			      p.ptr.dev, PTR_BUCKET_NR(ca, &p.ptr), g->gen,
-			      bch2_data_type_str(ptr_data_type(k->k, &p.ptr)),
-			      p.ptr.gen,
-			      (printbuf_reset(&buf),
-			       bch2_bkey_val_to_text(&buf, c, *k), buf.buf))))
-			do_update = true;
-
-		if (!p.ptr.cached && gen_cmp(p.ptr.gen, g->gen) < 0 &&
-		    (c->opts.reconstruct_alloc ||
-		     fsck_err(c, stale_dirty_ptr,
-			      "bucket %u:%zu data type %s stale dirty ptr: %u < %u\n"
-			      "while marking %s",
-			      p.ptr.dev, PTR_BUCKET_NR(ca, &p.ptr),
-			      bch2_data_type_str(ptr_data_type(k->k, &p.ptr)),
-			      p.ptr.gen, g->gen,
-			      (printbuf_reset(&buf),
-			       bch2_bkey_val_to_text(&buf, c, *k), buf.buf))))
-			do_update = true;
-
-		if (data_type != BCH_DATA_btree && p.ptr.gen != g->gen)
-			continue;
+			r->alive = false;
+			r->error = 0;
 
-		if (fsck_err_on(bucket_data_type(g->data_type) &&
-				bucket_data_type(g->data_type) != data_type, c,
-				ptr_bucket_data_type_mismatch,
-				"bucket %u:%zu different types of data in same bucket: %s, %s\n"
-				"while marking %s",
-				p.ptr.dev, PTR_BUCKET_NR(ca, &p.ptr),
-				bch2_data_type_str(g->data_type),
-				bch2_data_type_str(data_type),
-				(printbuf_reset(&buf),
-				 bch2_bkey_val_to_text(&buf, c, *k), buf.buf))) {
-			if (data_type == BCH_DATA_btree) {
-				g->data_type	= data_type;
-				set_bit(BCH_FS_need_another_gc, &c->flags);
+			if (!bch2_btree_has_scanned_nodes(c, i)) {
+				mustfix_fsck_err(c, btree_root_unreadable_and_scan_found_nothing,
+						 "no nodes found for btree %s, continue?", bch2_btree_id_str(i));
+				bch2_btree_root_alloc_fake_trans(trans, i, 0);
 			} else {
-				do_update = true;
+				bch2_btree_root_alloc_fake_trans(trans, i, 1);
+				bch2_shoot_down_journal_keys(c, i, 1, BTREE_MAX_DEPTH, POS_MIN, SPOS_MAX);
+				ret = bch2_get_scanned_nodes(c, i, 0, POS_MIN, SPOS_MAX);
+				if (ret)
+					break;
 			}
-		}
-
-		if (p.has_ec) {
-			struct gc_stripe *m = genradix_ptr(&c->gc_stripes, p.ec.idx);
-
-			if (fsck_err_on(!m || !m->alive, c,
-					ptr_to_missing_stripe,
-					"pointer to nonexistent stripe %llu\n"
-					"while marking %s",
-					(u64) p.ec.idx,
-					(printbuf_reset(&buf),
-					 bch2_bkey_val_to_text(&buf, c, *k), buf.buf)))
-				do_update = true;
-
-			if (fsck_err_on(m && m->alive && !bch2_ptr_matches_stripe_m(m, p), c,
-					ptr_to_incorrect_stripe,
-					"pointer does not match stripe %llu\n"
-					"while marking %s",
-					(u64) p.ec.idx,
-					(printbuf_reset(&buf),
-					 bch2_bkey_val_to_text(&buf, c, *k), buf.buf)))
-				do_update = true;
-		}
-	}
 
-	if (do_update) {
-		struct bkey_ptrs ptrs;
-		union bch_extent_entry *entry;
-		struct bch_extent_ptr *ptr;
-		struct bkey_i *new;
-
-		if (is_root) {
-			bch_err(c, "cannot update btree roots yet");
-			ret = -EINVAL;
-			goto err;
+			reconstructed_root = true;
 		}
 
-		new = kmalloc(bkey_bytes(k->k), GFP_KERNEL);
-		if (!new) {
-			ret = -BCH_ERR_ENOMEM_gc_repair_key;
-			bch_err_msg(c, ret, "allocating new key");
-			goto err;
-		}
+		struct btree *b = r->b;
 
-		bkey_reassemble(new, *k);
-
-		if (level) {
-			/*
-			 * We don't want to drop btree node pointers - if the
-			 * btree node isn't there anymore, the read path will
-			 * sort it out:
-			 */
-			ptrs = bch2_bkey_ptrs(bkey_i_to_s(new));
-			bkey_for_each_ptr(ptrs, ptr) {
-				struct bch_dev *ca = bch_dev_bkey_exists(c, ptr->dev);
-				struct bucket *g = PTR_GC_BUCKET(ca, ptr);
-
-				ptr->gen = g->gen;
-			}
-		} else {
-			bch2_bkey_drop_ptrs(bkey_i_to_s(new), ptr, ({
-				struct bch_dev *ca = bch_dev_bkey_exists(c, ptr->dev);
-				struct bucket *g = PTR_GC_BUCKET(ca, ptr);
-				enum bch_data_type data_type = bch2_bkey_ptr_data_type(*k, ptr);
-
-				(ptr->cached &&
-				 (!g->gen_valid || gen_cmp(ptr->gen, g->gen) > 0)) ||
-				(!ptr->cached &&
-				 gen_cmp(ptr->gen, g->gen) < 0) ||
-				gen_cmp(g->gen, ptr->gen) > BUCKET_GC_GEN_MAX ||
-				(g->data_type &&
-				 g->data_type != data_type);
-			}));
-again:
-			ptrs = bch2_bkey_ptrs(bkey_i_to_s(new));
-			bkey_extent_entry_for_each(ptrs, entry) {
-				if (extent_entry_type(entry) == BCH_EXTENT_ENTRY_stripe_ptr) {
-					struct gc_stripe *m = genradix_ptr(&c->gc_stripes,
-									entry->stripe_ptr.idx);
-					union bch_extent_entry *next_ptr;
-
-					bkey_extent_entry_for_each_from(ptrs, next_ptr, entry)
-						if (extent_entry_type(next_ptr) == BCH_EXTENT_ENTRY_ptr)
-							goto found;
-					next_ptr = NULL;
-found:
-					if (!next_ptr) {
-						bch_err(c, "aieee, found stripe ptr with no data ptr");
-						continue;
-					}
-
-					if (!m || !m->alive ||
-					    !__bch2_ptr_matches_stripe(&m->ptrs[entry->stripe_ptr.block],
-								       &next_ptr->ptr,
-								       m->sectors)) {
-						bch2_bkey_extent_entry_drop(new, entry);
-						goto again;
-					}
-				}
-			}
-		}
+		btree_node_lock_nopath_nofail(trans, &b->c, SIX_LOCK_read);
+		ret = bch2_btree_repair_topology_recurse(trans, b, &pulled_from_scan);
+		six_unlock_read(&b->c.lock);
 
-		ret = bch2_journal_key_insert_take(c, btree_id, level, new);
-		if (ret) {
-			kfree(new);
-			goto err;
-		}
+		if (ret == DROP_THIS_NODE) {
+			bch2_btree_node_hash_remove(&c->btree_cache, b);
+			mutex_lock(&c->btree_cache.lock);
+			list_move(&b->list, &c->btree_cache.freeable);
+			mutex_unlock(&c->btree_cache.lock);
 
-		if (level)
-			bch2_btree_node_update_key_early(trans, btree_id, level - 1, *k, new);
+			r->b = NULL;
 
-		if (0) {
-			printbuf_reset(&buf);
-			bch2_bkey_val_to_text(&buf, c, *k);
-			bch_info(c, "updated %s", buf.buf);
+			if (!reconstructed_root)
+				goto reconstruct_root;
 
-			printbuf_reset(&buf);
-			bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(new));
-			bch_info(c, "new key %s", buf.buf);
+			bch_err(c, "empty btree root %s", bch2_btree_id_str(i));
+			bch2_btree_root_alloc_fake_trans(trans, i, 0);
+			r->alive = false;
+			ret = 0;
 		}
-
-		*k = bkey_i_to_s_c(new);
 	}
-err:
 fsck_err:
-	printbuf_exit(&buf);
+	bch2_trans_put(trans);
 	return ret;
 }
 
 /* marking of btree keys/nodes: */
 
 static int bch2_gc_mark_key(struct btree_trans *trans, enum btree_id btree_id,
-			    unsigned level, bool is_root,
-			    struct bkey_s_c *k,
+			    unsigned level, struct btree **prev,
+			    struct btree_iter *iter, struct bkey_s_c k,
 			    bool initial)
 {
 	struct bch_fs *c = trans->c;
-	struct bkey deleted = KEY(0, 0, 0);
-	struct bkey_s_c old = (struct bkey_s_c) { &deleted, NULL };
-	int ret = 0;
-
-	deleted.p = k->k->p;
-
-	if (initial) {
-		BUG_ON(bch2_journal_seq_verify &&
-		       k->k->version.lo > atomic64_read(&c->journal.seq));
-
-		ret = bch2_check_fix_ptrs(trans, btree_id, level, is_root, k);
-		if (ret)
-			goto err;
-
-		if (fsck_err_on(k->k->version.lo > atomic64_read(&c->key_version), c,
-				bkey_version_in_future,
-				"key version number higher than recorded: %llu > %llu",
-				k->k->version.lo,
-				atomic64_read(&c->key_version)))
-			atomic64_set(&c->key_version, k->k->version.lo);
-	}
-
-	ret = commit_do(trans, NULL, NULL, 0,
-			bch2_key_trigger(trans, btree_id, level, old, unsafe_bkey_s_c_to_s(*k), BTREE_TRIGGER_GC));
-fsck_err:
-err:
-	bch_err_fn(c, ret);
-	return ret;
-}
-
-static int btree_gc_mark_node(struct btree_trans *trans, struct btree *b, bool initial)
-{
-	struct bch_fs *c = trans->c;
-	struct btree_node_iter iter;
-	struct bkey unpacked;
-	struct bkey_s_c k;
-	struct bkey_buf prev, cur;
-	int ret = 0;
-
-	if (!btree_node_type_needs_gc(btree_node_type(b)))
-		return 0;
 
-	bch2_btree_node_iter_init_from_start(&iter, b);
-	bch2_bkey_buf_init(&prev);
-	bch2_bkey_buf_init(&cur);
-	bkey_init(&prev.k->k);
-
-	while ((k = bch2_btree_node_iter_peek_unpack(&iter, b, &unpacked)).k) {
-		ret = bch2_gc_mark_key(trans, b->c.btree_id, b->c.level, false,
-				       &k, initial);
-		if (ret)
-			break;
+	if (iter) {
+		struct btree_path *path = btree_iter_path(trans, iter);
+		struct btree *b = path_l(path)->b;
 
-		bch2_btree_node_iter_advance(&iter, b);
-
-		if (b->c.level) {
-			bch2_bkey_buf_reassemble(&cur, c, k);
-
-			ret = bch2_gc_check_topology(c, b, &prev, cur,
-					bch2_btree_node_iter_end(&iter));
+		if (*prev != b) {
+			int ret = bch2_btree_node_check_topology(trans, b);
 			if (ret)
-				break;
+				return ret;
 		}
+		*prev = b;
 	}
 
-	bch2_bkey_buf_exit(&cur, c);
-	bch2_bkey_buf_exit(&prev, c);
-	return ret;
-}
-
-static int bch2_gc_btree(struct btree_trans *trans, enum btree_id btree_id,
-			 bool initial, bool metadata_only)
-{
-	struct bch_fs *c = trans->c;
-	struct btree_iter iter;
-	struct btree *b;
-	unsigned depth = metadata_only ? 1 : 0;
+	struct bkey deleted = KEY(0, 0, 0);
+	struct bkey_s_c old = (struct bkey_s_c) { &deleted, NULL };
+	struct printbuf buf = PRINTBUF;
 	int ret = 0;
 
-	gc_pos_set(c, gc_pos_btree(btree_id, POS_MIN, 0));
-
-	__for_each_btree_node(trans, iter, btree_id, POS_MIN,
-			      0, depth, BTREE_ITER_PREFETCH, b, ret) {
-		bch2_verify_btree_nr_keys(b);
+	deleted.p = k.k->p;
 
-		gc_pos_set(c, gc_pos_btree_node(b));
+	if (initial) {
+		BUG_ON(bch2_journal_seq_verify &&
+		       k.k->version.lo > atomic64_read(&c->journal.seq));
 
-		ret = btree_gc_mark_node(trans, b, initial);
-		if (ret)
-			break;
+		if (fsck_err_on(k.k->version.lo > atomic64_read(&c->key_version), c,
+				bkey_version_in_future,
+				"key version number higher than recorded %llu\n  %s",
+				atomic64_read(&c->key_version),
+				(bch2_bkey_val_to_text(&buf, c, k), buf.buf)))
+			atomic64_set(&c->key_version, k.k->version.lo);
 	}
-	bch2_trans_iter_exit(trans, &iter);
-
-	if (ret)
-		return ret;
 
-	mutex_lock(&c->btree_root_lock);
-	b = bch2_btree_id_root(c, btree_id)->b;
-	if (!btree_node_fake(b)) {
-		struct bkey_s_c k = bkey_i_to_s_c(&b->key);
-
-		ret = bch2_gc_mark_key(trans, b->c.btree_id, b->c.level + 1,
-				       true, &k, initial);
+	if (mustfix_fsck_err_on(level && !bch2_dev_btree_bitmap_marked(c, k),
+				c, btree_bitmap_not_marked,
+				"btree ptr not marked in member info btree allocated bitmap\n  %s",
+				(printbuf_reset(&buf),
+				 bch2_bkey_val_to_text(&buf, c, k),
+				 buf.buf))) {
+		mutex_lock(&c->sb_lock);
+		bch2_dev_btree_bitmap_mark(c, k);
+		bch2_write_super(c);
+		mutex_unlock(&c->sb_lock);
 	}
-	gc_pos_set(c, gc_pos_btree_root(b->c.btree_id));
-	mutex_unlock(&c->btree_root_lock);
-
-	return ret;
-}
 
-static int bch2_gc_btree_init_recurse(struct btree_trans *trans, struct btree *b,
-				      unsigned target_depth)
-{
-	struct bch_fs *c = trans->c;
-	struct btree_and_journal_iter iter;
-	struct bkey_s_c k;
-	struct bkey_buf cur, prev;
-	struct printbuf buf = PRINTBUF;
-	int ret = 0;
-
-	bch2_btree_and_journal_iter_init_node_iter(&iter, c, b);
-	bch2_bkey_buf_init(&prev);
-	bch2_bkey_buf_init(&cur);
-	bkey_init(&prev.k->k);
-
-	while ((k = bch2_btree_and_journal_iter_peek(&iter)).k) {
-		BUG_ON(bpos_lt(k.k->p, b->data->min_key));
-		BUG_ON(bpos_gt(k.k->p, b->data->max_key));
-
-		ret = bch2_gc_mark_key(trans, b->c.btree_id, b->c.level,
-				       false, &k, true);
-		if (ret)
-			goto fsck_err;
-
-		if (b->c.level) {
-			bch2_bkey_buf_reassemble(&cur, c, k);
-			k = bkey_i_to_s_c(cur.k);
+	/*
+	 * We require a commit before key_trigger() because
+	 * key_trigger(BTREE_TRIGGER_GC) is not idempotant; we'll calculate the
+	 * wrong result if we run it multiple times.
+	 */
+	unsigned flags = !iter ? BTREE_TRIGGER_is_root : 0;
 
-			bch2_btree_and_journal_iter_advance(&iter);
+	ret = bch2_key_trigger(trans, btree_id, level, old, unsafe_bkey_s_c_to_s(k),
+			       BTREE_TRIGGER_check_repair|flags);
+	if (ret)
+		goto out;
 
-			ret = bch2_gc_check_topology(c, b,
-					&prev, cur,
-					!bch2_btree_and_journal_iter_peek(&iter).k);
-			if (ret)
-				goto fsck_err;
-		} else {
-			bch2_btree_and_journal_iter_advance(&iter);
-		}
+	if (trans->nr_updates) {
+		ret = bch2_trans_commit(trans, NULL, NULL, 0) ?:
+			-BCH_ERR_transaction_restart_nested;
+		goto out;
 	}
 
-	if (b->c.level > target_depth) {
-		bch2_btree_and_journal_iter_exit(&iter);
-		bch2_btree_and_journal_iter_init_node_iter(&iter, c, b);
-
-		while ((k = bch2_btree_and_journal_iter_peek(&iter)).k) {
-			struct btree *child;
-
-			bch2_bkey_buf_reassemble(&cur, c, k);
-			bch2_btree_and_journal_iter_advance(&iter);
-
-			child = bch2_btree_node_get_noiter(trans, cur.k,
-						b->c.btree_id, b->c.level - 1,
-						false);
-			ret = PTR_ERR_OR_ZERO(child);
-
-			if (ret == -EIO) {
-				bch2_topology_error(c);
-
-				if (__fsck_err(c,
-					  FSCK_CAN_FIX|
-					  FSCK_CAN_IGNORE|
-					  FSCK_NO_RATELIMIT,
-					  btree_node_read_error,
-					  "Unreadable btree node at btree %s level %u:\n"
-					  "  %s",
-					  bch2_btree_id_str(b->c.btree_id),
-					  b->c.level - 1,
-					  (printbuf_reset(&buf),
-					   bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(cur.k)), buf.buf)) &&
-				    should_restart_for_topology_repair(c)) {
-					bch_info(c, "Halting mark and sweep to start topology repair pass");
-					ret = bch2_run_explicit_recovery_pass(c, BCH_RECOVERY_PASS_check_topology);
-					goto fsck_err;
-				} else {
-					/* Continue marking when opted to not
-					 * fix the error: */
-					ret = 0;
-					set_bit(BCH_FS_initial_gc_unfixed, &c->flags);
-					continue;
-				}
-			} else if (ret) {
-				bch_err_msg(c, ret, "getting btree node");
-				break;
-			}
-
-			ret = bch2_gc_btree_init_recurse(trans, child,
-							 target_depth);
-			six_unlock_read(&child->c.lock);
-
-			if (ret)
-				break;
-		}
-	}
+	ret = bch2_key_trigger(trans, btree_id, level, old, unsafe_bkey_s_c_to_s(k),
+			       BTREE_TRIGGER_gc|flags);
+out:
 fsck_err:
-	bch2_bkey_buf_exit(&cur, c);
-	bch2_bkey_buf_exit(&prev, c);
-	bch2_btree_and_journal_iter_exit(&iter);
 	printbuf_exit(&buf);
+	bch_err_fn(c, ret);
 	return ret;
 }
 
-static int bch2_gc_btree_init(struct btree_trans *trans,
-			      enum btree_id btree_id,
-			      bool metadata_only)
+static int bch2_gc_btree(struct btree_trans *trans, enum btree_id btree, bool initial)
 {
 	struct bch_fs *c = trans->c;
-	struct btree *b;
-	unsigned target_depth = metadata_only ? 1 : 0;
-	struct printbuf buf = PRINTBUF;
+	int level = 0, target_depth = btree_node_type_needs_gc(__btree_node_type(0, btree)) ? 0 : 1;
 	int ret = 0;
 
-	b = bch2_btree_id_root(c, btree_id)->b;
+	/* We need to make sure every leaf node is readable before going RW */
+	if (initial)
+		target_depth = 0;
 
-	if (btree_node_fake(b))
-		return 0;
+	/* root */
+	do {
+retry_root:
+		bch2_trans_begin(trans);
+
+		struct btree_iter iter;
+		bch2_trans_node_iter_init(trans, &iter, btree, POS_MIN,
+					  0, bch2_btree_id_root(c, btree)->b->c.level, 0);
+		struct btree *b = bch2_btree_iter_peek_node(&iter);
+		ret = PTR_ERR_OR_ZERO(b);
+		if (ret)
+			goto err_root;
 
-	six_lock_read(&b->c.lock, NULL, NULL);
-	printbuf_reset(&buf);
-	bch2_bpos_to_text(&buf, b->data->min_key);
-	if (mustfix_fsck_err_on(!bpos_eq(b->data->min_key, POS_MIN), c,
-				btree_root_bad_min_key,
-			"btree root with incorrect min_key: %s", buf.buf)) {
-		bch_err(c, "repair unimplemented");
-		ret = -BCH_ERR_fsck_repair_unimplemented;
-		goto fsck_err;
-	}
+		if (b != btree_node_root(c, b)) {
+			bch2_trans_iter_exit(trans, &iter);
+			goto retry_root;
+		}
 
-	printbuf_reset(&buf);
-	bch2_bpos_to_text(&buf, b->data->max_key);
-	if (mustfix_fsck_err_on(!bpos_eq(b->data->max_key, SPOS_MAX), c,
-				btree_root_bad_max_key,
-			"btree root with incorrect max_key: %s", buf.buf)) {
-		bch_err(c, "repair unimplemented");
-		ret = -BCH_ERR_fsck_repair_unimplemented;
-		goto fsck_err;
-	}
+		gc_pos_set(c, gc_pos_btree(btree, b->c.level + 1, SPOS_MAX));
+		struct bkey_s_c k = bkey_i_to_s_c(&b->key);
+		ret = bch2_gc_mark_key(trans, btree, b->c.level + 1, NULL, NULL, k, initial);
+		level = b->c.level;
+err_root:
+		bch2_trans_iter_exit(trans, &iter);
+	} while (bch2_err_matches(ret, BCH_ERR_transaction_restart));
 
-	if (b->c.level >= target_depth)
-		ret = bch2_gc_btree_init_recurse(trans, b, target_depth);
+	if (ret)
+		return ret;
 
-	if (!ret) {
-		struct bkey_s_c k = bkey_i_to_s_c(&b->key);
+	for (; level >= target_depth; --level) {
+		struct btree *prev = NULL;
+		struct btree_iter iter;
+		bch2_trans_node_iter_init(trans, &iter, btree, POS_MIN, 0, level,
+					  BTREE_ITER_prefetch);
 
-		ret = bch2_gc_mark_key(trans, b->c.btree_id, b->c.level + 1, true,
-				       &k, true);
+		ret = for_each_btree_key_continue(trans, iter, 0, k, ({
+			gc_pos_set(c, gc_pos_btree(btree, level, k.k->p));
+			bch2_gc_mark_key(trans, btree, level, &prev, &iter, k, initial);
+		}));
+		if (ret)
+			break;
 	}
-fsck_err:
-	six_unlock_read(&b->c.lock);
 
-	bch_err_fn(c, ret);
-	printbuf_exit(&buf);
 	return ret;
 }
 
 static inline int btree_id_gc_phase_cmp(enum btree_id l, enum btree_id r)
 {
-	return  (int) btree_id_to_gc_phase(l) -
-		(int) btree_id_to_gc_phase(r);
+	return cmp_int(gc_btree_order(l), gc_btree_order(r));
 }
 
-static int bch2_gc_btrees(struct bch_fs *c, bool initial, bool metadata_only)
+static int bch2_gc_btrees(struct bch_fs *c)
 {
 	struct btree_trans *trans = bch2_trans_get(c);
 	enum btree_id ids[BTREE_ID_NR];
@@ -1092,107 +702,43 @@ static int bch2_gc_btrees(struct bch_fs *c, bool initial, bool metadata_only)
 		ids[i] = i;
 	bubble_sort(ids, BTREE_ID_NR, btree_id_gc_phase_cmp);
 
-	for (i = 0; i < BTREE_ID_NR && !ret; i++)
-		ret = initial
-			? bch2_gc_btree_init(trans, ids[i], metadata_only)
-			: bch2_gc_btree(trans, ids[i], initial, metadata_only);
+	for (i = 0; i < btree_id_nr_alive(c) && !ret; i++) {
+		unsigned btree = i < BTREE_ID_NR ? ids[i] : i;
 
-	for (i = BTREE_ID_NR; i < btree_id_nr_alive(c) && !ret; i++) {
-		if (!bch2_btree_id_root(c, i)->alive)
+		if (IS_ERR_OR_NULL(bch2_btree_id_root(c, btree)->b))
 			continue;
 
-		ret = initial
-			? bch2_gc_btree_init(trans, i, metadata_only)
-			: bch2_gc_btree(trans, i, initial, metadata_only);
-	}
+		ret = bch2_gc_btree(trans, btree, true);
 
+		if (mustfix_fsck_err_on(bch2_err_matches(ret, EIO),
+					c, btree_node_read_error,
+			       "btree node read error for %s",
+			       bch2_btree_id_str(btree)))
+			ret = bch2_run_explicit_recovery_pass(c, BCH_RECOVERY_PASS_check_topology);
+	}
+fsck_err:
 	bch2_trans_put(trans);
 	bch_err_fn(c, ret);
 	return ret;
 }
 
-static void mark_metadata_sectors(struct bch_fs *c, struct bch_dev *ca,
-				  u64 start, u64 end,
-				  enum bch_data_type type,
-				  unsigned flags)
-{
-	u64 b = sector_to_bucket(ca, start);
-
-	do {
-		unsigned sectors =
-			min_t(u64, bucket_to_sector(ca, b + 1), end) - start;
-
-		bch2_mark_metadata_bucket(c, ca, b, type, sectors,
-					  gc_phase(GC_PHASE_SB), flags);
-		b++;
-		start += sectors;
-	} while (start < end);
-}
-
-static void bch2_mark_dev_superblock(struct bch_fs *c, struct bch_dev *ca,
-				     unsigned flags)
-{
-	struct bch_sb_layout *layout = &ca->disk_sb.sb->layout;
-	unsigned i;
-	u64 b;
-
-	for (i = 0; i < layout->nr_superblocks; i++) {
-		u64 offset = le64_to_cpu(layout->sb_offset[i]);
-
-		if (offset == BCH_SB_SECTOR)
-			mark_metadata_sectors(c, ca, 0, BCH_SB_SECTOR,
-					      BCH_DATA_sb, flags);
-
-		mark_metadata_sectors(c, ca, offset,
-				      offset + (1 << layout->sb_max_size_bits),
-				      BCH_DATA_sb, flags);
-	}
-
-	for (i = 0; i < ca->journal.nr; i++) {
-		b = ca->journal.buckets[i];
-		bch2_mark_metadata_bucket(c, ca, b, BCH_DATA_journal,
-					  ca->mi.bucket_size,
-					  gc_phase(GC_PHASE_SB), flags);
-	}
-}
-
-static void bch2_mark_superblocks(struct bch_fs *c)
+static int bch2_mark_superblocks(struct bch_fs *c)
 {
 	mutex_lock(&c->sb_lock);
-	gc_pos_set(c, gc_phase(GC_PHASE_SB));
+	gc_pos_set(c, gc_phase(GC_PHASE_sb));
 
-	for_each_online_member(c, ca)
-		bch2_mark_dev_superblock(c, ca, BTREE_TRIGGER_GC);
+	int ret = bch2_trans_mark_dev_sbs_flags(c, BTREE_TRIGGER_gc);
 	mutex_unlock(&c->sb_lock);
+	return ret;
 }
 
-#if 0
-/* Also see bch2_pending_btree_node_free_insert_done() */
-static void bch2_mark_pending_btree_node_frees(struct bch_fs *c)
-{
-	struct btree_update *as;
-	struct pending_btree_node_free *d;
-
-	mutex_lock(&c->btree_interior_update_lock);
-	gc_pos_set(c, gc_phase(GC_PHASE_PENDING_DELETE));
-
-	for_each_pending_btree_node_free(c, as, d)
-		if (d->index_update_done)
-			bch2_mark_key(c, bkey_i_to_s_c(&d->key), BTREE_TRIGGER_GC);
-
-	mutex_unlock(&c->btree_interior_update_lock);
-}
-#endif
-
 static void bch2_gc_free(struct bch_fs *c)
 {
 	genradix_free(&c->reflink_gc_table);
 	genradix_free(&c->gc_stripes);
 
 	for_each_member_device(c, ca) {
-		kvpfree(rcu_dereference_protected(ca->buckets_gc, 1),
-			sizeof(struct bucket_array) +
-			ca->mi.nbuckets * sizeof(struct bucket));
+		kvfree(rcu_dereference_protected(ca->buckets_gc, 1));
 		ca->buckets_gc = NULL;
 
 		free_percpu(ca->usage_gc);
@@ -1203,28 +749,23 @@ static void bch2_gc_free(struct bch_fs *c)
 	c->usage_gc = NULL;
 }
 
-static int bch2_gc_done(struct bch_fs *c,
-			bool initial, bool metadata_only)
+static int bch2_gc_done(struct bch_fs *c)
 {
 	struct bch_dev *ca = NULL;
 	struct printbuf buf = PRINTBUF;
-	bool verify = !metadata_only &&
-		!c->opts.reconstruct_alloc &&
-		(!initial || (c->sb.compat & (1ULL << BCH_COMPAT_alloc_info)));
 	unsigned i;
 	int ret = 0;
 
 	percpu_down_write(&c->mark_lock);
 
-#define copy_field(_err, _f, _msg, ...)					\
-	if (dst->_f != src->_f &&					\
-	    (!verify ||							\
-	     fsck_err(c, _err, _msg ": got %llu, should be %llu"	\
-		      , ##__VA_ARGS__, dst->_f, src->_f)))		\
+#define copy_field(_err, _f, _msg, ...)						\
+	if (fsck_err_on(dst->_f != src->_f, c, _err,				\
+			_msg ": got %llu, should be %llu" , ##__VA_ARGS__,	\
+			dst->_f, src->_f))					\
 		dst->_f = src->_f
-#define copy_dev_field(_err, _f, _msg, ...)				\
+#define copy_dev_field(_err, _f, _msg, ...)					\
 	copy_field(_err, _f, "dev %u has wrong " _msg, ca->dev_idx, ##__VA_ARGS__)
-#define copy_fs_field(_err, _f, _msg, ...)				\
+#define copy_fs_field(_err, _f, _msg, ...)					\
 	copy_field(_err, _f, "fs has wrong " _msg, ##__VA_ARGS__)
 
 	for (i = 0; i < ARRAY_SIZE(c->usage); i++)
@@ -1257,31 +798,24 @@ static int bch2_gc_done(struct bch_fs *c,
 		copy_fs_field(fs_usage_btree_wrong,
 			      b.btree,		"btree");
 
-		if (!metadata_only) {
-			copy_fs_field(fs_usage_data_wrong,
-				      b.data,	"data");
-			copy_fs_field(fs_usage_cached_wrong,
-				      b.cached,	"cached");
-			copy_fs_field(fs_usage_reserved_wrong,
-				      b.reserved,	"reserved");
-			copy_fs_field(fs_usage_nr_inodes_wrong,
-				      b.nr_inodes,"nr_inodes");
-
-			for (i = 0; i < BCH_REPLICAS_MAX; i++)
-				copy_fs_field(fs_usage_persistent_reserved_wrong,
-					      persistent_reserved[i],
-					      "persistent_reserved[%i]", i);
-		}
+		copy_fs_field(fs_usage_data_wrong,
+			      b.data,	"data");
+		copy_fs_field(fs_usage_cached_wrong,
+			      b.cached,	"cached");
+		copy_fs_field(fs_usage_reserved_wrong,
+			      b.reserved,	"reserved");
+		copy_fs_field(fs_usage_nr_inodes_wrong,
+			      b.nr_inodes,"nr_inodes");
+
+		for (i = 0; i < BCH_REPLICAS_MAX; i++)
+			copy_fs_field(fs_usage_persistent_reserved_wrong,
+				      persistent_reserved[i],
+				      "persistent_reserved[%i]", i);
 
 		for (i = 0; i < c->replicas.nr; i++) {
 			struct bch_replicas_entry_v1 *e =
 				cpu_replicas_entry(&c->replicas, i);
 
-			if (metadata_only &&
-			    (e->data_type == BCH_DATA_user ||
-			     e->data_type == BCH_DATA_cached))
-				continue;
-
 			printbuf_reset(&buf);
 			bch2_replicas_entry_to_text(&buf, e);
 
@@ -1295,10 +829,8 @@ static int bch2_gc_done(struct bch_fs *c,
 #undef copy_stripe_field
 #undef copy_field
 fsck_err:
-	if (ca)
-		percpu_ref_put(&ca->ref);
+	bch2_dev_put(ca);
 	bch_err_fn(c, ret);
-
 	percpu_up_write(&c->mark_lock);
 	printbuf_exit(&buf);
 	return ret;
@@ -1321,7 +853,7 @@ static int bch2_gc_start(struct bch_fs *c)
 		ca->usage_gc = alloc_percpu(struct bch_dev_usage);
 		if (!ca->usage_gc) {
 			bch_err(c, "error allocating ca->usage_gc");
-			percpu_ref_put(&ca->ref);
+			bch2_dev_put(ca);
 			return -BCH_ERR_ENOMEM_gc_start;
 		}
 
@@ -1332,19 +864,6 @@ static int bch2_gc_start(struct bch_fs *c)
 	return 0;
 }
 
-static int bch2_gc_reset(struct bch_fs *c)
-{
-	for_each_member_device(c, ca) {
-		free_percpu(ca->usage_gc);
-		ca->usage_gc = NULL;
-	}
-
-	free_percpu(c->usage_gc);
-	c->usage_gc = NULL;
-
-	return bch2_gc_start(c);
-}
-
 /* returns true if not equal */
 static inline bool bch2_alloc_v4_cmp(struct bch_alloc_v4 l,
 				     struct bch_alloc_v4 r)
@@ -1360,58 +879,47 @@ static inline bool bch2_alloc_v4_cmp(struct bch_alloc_v4 l,
 
 static int bch2_alloc_write_key(struct btree_trans *trans,
 				struct btree_iter *iter,
-				struct bkey_s_c k,
-				bool metadata_only)
+				struct bch_dev *ca,
+				struct bkey_s_c k)
 {
 	struct bch_fs *c = trans->c;
-	struct bch_dev *ca = bch_dev_bkey_exists(c, iter->pos.inode);
-	struct bucket gc, *b;
 	struct bkey_i_alloc_v4 *a;
-	struct bch_alloc_v4 old_convert, new;
+	struct bch_alloc_v4 old_gc, gc, old_convert, new;
 	const struct bch_alloc_v4 *old;
-	enum bch_data_type type;
 	int ret;
 
+	if (!bucket_valid(ca, k.k->p.offset))
+		return 0;
+
 	old = bch2_alloc_to_v4(k, &old_convert);
-	new = *old;
+	gc = new = *old;
 
 	percpu_down_read(&c->mark_lock);
-	b = gc_bucket(ca, iter->pos.offset);
+	__bucket_m_to_alloc(&gc, *gc_bucket(ca, iter->pos.offset));
+
+	old_gc = gc;
+
+	if ((old->data_type == BCH_DATA_sb ||
+	     old->data_type == BCH_DATA_journal) &&
+	    !bch2_dev_is_online(ca)) {
+		gc.data_type = old->data_type;
+		gc.dirty_sectors = old->dirty_sectors;
+	}
 
 	/*
-	 * b->data_type doesn't yet include need_discard & need_gc_gen states -
+	 * gc.data_type doesn't yet include need_discard & need_gc_gen states -
 	 * fix that here:
 	 */
-	type = __alloc_data_type(b->dirty_sectors,
-				 b->cached_sectors,
-				 b->stripe,
-				 *old,
-				 b->data_type);
-	if (b->data_type != type) {
-		struct bch_dev_usage *u;
-
-		preempt_disable();
-		u = this_cpu_ptr(ca->usage_gc);
-		u->d[b->data_type].buckets--;
-		b->data_type = type;
-		u->d[b->data_type].buckets++;
-		preempt_enable();
-	}
+	alloc_data_type_set(&gc, gc.data_type);
 
-	gc = *b;
+	if (gc.data_type != old_gc.data_type ||
+	    gc.dirty_sectors != old_gc.dirty_sectors)
+		bch2_dev_usage_update(c, ca, &old_gc, &gc, 0, true);
 	percpu_up_read(&c->mark_lock);
 
-	if (metadata_only &&
-	    gc.data_type != BCH_DATA_sb &&
-	    gc.data_type != BCH_DATA_journal &&
-	    gc.data_type != BCH_DATA_btree)
-		return 0;
-
-	if (gen_after(old->gen, gc.gen))
-		return 0;
+	gc.fragmentation_lru = alloc_lru_idx_fragmentation(gc, ca);
 
-	if (c->opts.reconstruct_alloc ||
-	    fsck_err_on(new.data_type != gc.data_type, c,
+	if (fsck_err_on(new.data_type != gc.data_type, c,
 			alloc_key_data_type_wrong,
 			"bucket %llu:%llu gen %u has wrong data_type"
 			": got %s, should be %s",
@@ -1422,26 +930,21 @@ static int bch2_alloc_write_key(struct btree_trans *trans,
 		new.data_type = gc.data_type;
 
 #define copy_bucket_field(_errtype, _f)					\
-	if (c->opts.reconstruct_alloc ||				\
-	    fsck_err_on(new._f != gc._f, c, _errtype,			\
+	if (fsck_err_on(new._f != gc._f, c, _errtype,			\
 			"bucket %llu:%llu gen %u data type %s has wrong " #_f	\
-			": got %u, should be %u",			\
+			": got %llu, should be %llu",			\
 			iter->pos.inode, iter->pos.offset,		\
 			gc.gen,						\
 			bch2_data_type_str(gc.data_type),		\
-			new._f, gc._f))					\
+			(u64) new._f, (u64) gc._f))				\
 		new._f = gc._f;						\
 
-	copy_bucket_field(alloc_key_gen_wrong,
-			  gen);
-	copy_bucket_field(alloc_key_dirty_sectors_wrong,
-			  dirty_sectors);
-	copy_bucket_field(alloc_key_cached_sectors_wrong,
-			  cached_sectors);
-	copy_bucket_field(alloc_key_stripe_wrong,
-			  stripe);
-	copy_bucket_field(alloc_key_stripe_redundancy_wrong,
-			  stripe_redundancy);
+	copy_bucket_field(alloc_key_gen_wrong,			gen);
+	copy_bucket_field(alloc_key_dirty_sectors_wrong,	dirty_sectors);
+	copy_bucket_field(alloc_key_cached_sectors_wrong,	cached_sectors);
+	copy_bucket_field(alloc_key_stripe_wrong,		stripe);
+	copy_bucket_field(alloc_key_stripe_redundancy_wrong,	stripe_redundancy);
+	copy_bucket_field(alloc_key_fragmentation_lru_wrong,	fragmentation_lru);
 #undef copy_bucket_field
 
 	if (!bch2_alloc_v4_cmp(*old, new))
@@ -1455,18 +958,18 @@ static int bch2_alloc_write_key(struct btree_trans *trans,
 	a->v = new;
 
 	/*
-	 * The trigger normally makes sure this is set, but we're not running
+	 * The trigger normally makes sure these are set, but we're not running
 	 * triggers:
 	 */
 	if (a->v.data_type == BCH_DATA_cached && !a->v.io_time[READ])
 		a->v.io_time[READ] = max_t(u64, 1, atomic64_read(&c->io_clock[READ].now));
 
-	ret = bch2_trans_update(trans, iter, &a->k_i, BTREE_TRIGGER_NORUN);
+	ret = bch2_trans_update(trans, iter, &a->k_i, BTREE_TRIGGER_norun);
 fsck_err:
 	return ret;
 }
 
-static int bch2_gc_alloc_done(struct bch_fs *c, bool metadata_only)
+static int bch2_gc_alloc_done(struct bch_fs *c)
 {
 	int ret = 0;
 
@@ -1475,11 +978,11 @@ static int bch2_gc_alloc_done(struct bch_fs *c, bool metadata_only)
 			for_each_btree_key_upto_commit(trans, iter, BTREE_ID_alloc,
 					POS(ca->dev_idx, ca->mi.first_bucket),
 					POS(ca->dev_idx, ca->mi.nbuckets - 1),
-					BTREE_ITER_SLOTS|BTREE_ITER_PREFETCH, k,
+					BTREE_ITER_slots|BTREE_ITER_prefetch, k,
 					NULL, NULL, BCH_TRANS_COMMIT_lazy_rw,
-				bch2_alloc_write_key(trans, &iter, k, metadata_only)));
+				bch2_alloc_write_key(trans, &iter, ca, k)));
 		if (ret) {
-			percpu_ref_put(&ca->ref);
+			bch2_dev_put(ca);
 			break;
 		}
 	}
@@ -1488,71 +991,50 @@ static int bch2_gc_alloc_done(struct bch_fs *c, bool metadata_only)
 	return ret;
 }
 
-static int bch2_gc_alloc_start(struct bch_fs *c, bool metadata_only)
+static int bch2_gc_alloc_start(struct bch_fs *c)
 {
 	for_each_member_device(c, ca) {
-		struct bucket_array *buckets = kvpmalloc(sizeof(struct bucket_array) +
+		struct bucket_array *buckets = kvmalloc(sizeof(struct bucket_array) +
 				ca->mi.nbuckets * sizeof(struct bucket),
 				GFP_KERNEL|__GFP_ZERO);
 		if (!buckets) {
-			percpu_ref_put(&ca->ref);
+			bch2_dev_put(ca);
 			bch_err(c, "error allocating ca->buckets[gc]");
 			return -BCH_ERR_ENOMEM_gc_alloc_start;
 		}
 
 		buckets->first_bucket	= ca->mi.first_bucket;
 		buckets->nbuckets	= ca->mi.nbuckets;
+		buckets->nbuckets_minus_first =
+			buckets->nbuckets - buckets->first_bucket;
 		rcu_assign_pointer(ca->buckets_gc, buckets);
 	}
 
+	struct bch_dev *ca = NULL;
 	int ret = bch2_trans_run(c,
 		for_each_btree_key(trans, iter, BTREE_ID_alloc, POS_MIN,
-					 BTREE_ITER_PREFETCH, k, ({
-			struct bch_dev *ca = bch_dev_bkey_exists(c, k.k->p.inode);
-			struct bucket *g = gc_bucket(ca, k.k->p.offset);
-
-			struct bch_alloc_v4 a_convert;
-			const struct bch_alloc_v4 *a = bch2_alloc_to_v4(k, &a_convert);
-
-			g->gen_valid	= 1;
-			g->gen		= a->gen;
-
-			if (metadata_only &&
-			    (a->data_type == BCH_DATA_user ||
-			     a->data_type == BCH_DATA_cached ||
-			     a->data_type == BCH_DATA_parity)) {
-				g->data_type		= a->data_type;
-				g->dirty_sectors	= a->dirty_sectors;
-				g->cached_sectors	= a->cached_sectors;
-				g->stripe		= a->stripe;
-				g->stripe_redundancy	= a->stripe_redundancy;
+					 BTREE_ITER_prefetch, k, ({
+			ca = bch2_dev_iterate(c, ca, k.k->p.inode);
+			if (!ca) {
+				bch2_btree_iter_set_pos(&iter, POS(k.k->p.inode + 1, 0));
+				continue;
 			}
 
+			if (bucket_valid(ca, k.k->p.offset)) {
+				struct bch_alloc_v4 a_convert;
+				const struct bch_alloc_v4 *a = bch2_alloc_to_v4(k, &a_convert);
+
+				struct bucket *g = gc_bucket(ca, k.k->p.offset);
+				g->gen_valid	= 1;
+				g->gen		= a->gen;
+			}
 			0;
 		})));
+	bch2_dev_put(ca);
 	bch_err_fn(c, ret);
 	return ret;
 }
 
-static void bch2_gc_alloc_reset(struct bch_fs *c, bool metadata_only)
-{
-	for_each_member_device(c, ca) {
-		struct bucket_array *buckets = gc_bucket_array(ca);
-		struct bucket *g;
-
-		for_each_bucket(g, buckets) {
-			if (metadata_only &&
-			    (g->data_type == BCH_DATA_user ||
-			     g->data_type == BCH_DATA_cached ||
-			     g->data_type == BCH_DATA_parity))
-				continue;
-			g->data_type = 0;
-			g->dirty_sectors = 0;
-			g->cached_sectors = 0;
-		}
-	}
-}
-
 static int bch2_gc_write_reflink_key(struct btree_trans *trans,
 				     struct btree_iter *iter,
 				     struct bkey_s_c k,
@@ -1585,51 +1067,44 @@ static int bch2_gc_write_reflink_key(struct btree_trans *trans,
 			"  should be %u",
 			(bch2_bkey_val_to_text(&buf, c, k), buf.buf),
 			r->refcount)) {
-		struct bkey_i *new = bch2_bkey_make_mut(trans, iter, &k, 0);
-
+		struct bkey_i *new = bch2_bkey_make_mut_noupdate(trans, k);
 		ret = PTR_ERR_OR_ZERO(new);
 		if (ret)
-			return ret;
+			goto out;
 
 		if (!r->refcount)
 			new->k.type = KEY_TYPE_deleted;
 		else
 			*bkey_refcount(bkey_i_to_s(new)) = cpu_to_le64(r->refcount);
+		ret = bch2_trans_update(trans, iter, new, 0);
 	}
+out:
 fsck_err:
 	printbuf_exit(&buf);
 	return ret;
 }
 
-static int bch2_gc_reflink_done(struct bch_fs *c, bool metadata_only)
+static int bch2_gc_reflink_done(struct bch_fs *c)
 {
 	size_t idx = 0;
 
-	if (metadata_only)
-		return 0;
-
 	int ret = bch2_trans_run(c,
 		for_each_btree_key_commit(trans, iter,
 				BTREE_ID_reflink, POS_MIN,
-				BTREE_ITER_PREFETCH, k,
+				BTREE_ITER_prefetch, k,
 				NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
 			bch2_gc_write_reflink_key(trans, &iter, k, &idx)));
 	c->reflink_gc_nr = 0;
 	return ret;
 }
 
-static int bch2_gc_reflink_start(struct bch_fs *c,
-				 bool metadata_only)
+static int bch2_gc_reflink_start(struct bch_fs *c)
 {
-
-	if (metadata_only)
-		return 0;
-
 	c->reflink_gc_nr = 0;
 
 	int ret = bch2_trans_run(c,
 		for_each_btree_key(trans, iter, BTREE_ID_reflink, POS_MIN,
-				   BTREE_ITER_PREFETCH, k, ({
+				   BTREE_ITER_prefetch, k, ({
 			const __le64 *refcount = bkey_refcount_c(k);
 
 			if (!refcount)
@@ -1652,15 +1127,6 @@ static int bch2_gc_reflink_start(struct bch_fs *c,
 	return ret;
 }
 
-static void bch2_gc_reflink_reset(struct bch_fs *c, bool metadata_only)
-{
-	struct genradix_iter iter;
-	struct reflink_gc *r;
-
-	genradix_for_each(&c->reflink_gc_table, iter, r)
-		r->refcount = 0;
-}
-
 static int bch2_gc_write_stripes_key(struct btree_trans *trans,
 				     struct btree_iter *iter,
 				     struct bkey_s_c k)
@@ -1714,30 +1180,20 @@ fsck_err:
 	return ret;
 }
 
-static int bch2_gc_stripes_done(struct bch_fs *c, bool metadata_only)
+static int bch2_gc_stripes_done(struct bch_fs *c)
 {
-	if (metadata_only)
-		return 0;
-
 	return bch2_trans_run(c,
 		for_each_btree_key_commit(trans, iter,
 				BTREE_ID_stripes, POS_MIN,
-				BTREE_ITER_PREFETCH, k,
+				BTREE_ITER_prefetch, k,
 				NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
 			bch2_gc_write_stripes_key(trans, &iter, k)));
 }
 
-static void bch2_gc_stripes_reset(struct bch_fs *c, bool metadata_only)
-{
-	genradix_free(&c->gc_stripes);
-}
-
 /**
- * bch2_gc - walk _all_ references to buckets, and recompute them:
+ * bch2_check_allocations - walk all references to buckets, and recompute them:
  *
  * @c:			filesystem object
- * @initial:		are we in recovery?
- * @metadata_only:	are we just checking metadata references, or everything?
  *
  * Returns: 0 on success, or standard errcode on failure
  *
@@ -1756,9 +1212,8 @@ static void bch2_gc_stripes_reset(struct bch_fs *c, bool metadata_only)
  *    move around - if references move backwards in the ordering GC
  *    uses, GC could skip past them
  */
-int bch2_gc(struct bch_fs *c, bool initial, bool metadata_only)
+int bch2_check_allocations(struct bch_fs *c)
 {
-	unsigned iter = 0;
 	int ret;
 
 	lockdep_assert_held(&c->state_lock);
@@ -1768,66 +1223,34 @@ int bch2_gc(struct bch_fs *c, bool initial, bool metadata_only)
 	bch2_btree_interior_updates_flush(c);
 
 	ret   = bch2_gc_start(c) ?:
-		bch2_gc_alloc_start(c, metadata_only) ?:
-		bch2_gc_reflink_start(c, metadata_only);
+		bch2_gc_alloc_start(c) ?:
+		bch2_gc_reflink_start(c);
 	if (ret)
 		goto out;
-again:
-	gc_pos_set(c, gc_phase(GC_PHASE_START));
 
-	bch2_mark_superblocks(c);
+	gc_pos_set(c, gc_phase(GC_PHASE_start));
 
-	ret = bch2_gc_btrees(c, initial, metadata_only);
+	ret = bch2_mark_superblocks(c);
+	BUG_ON(ret);
 
+	ret = bch2_gc_btrees(c);
 	if (ret)
 		goto out;
 
-#if 0
-	bch2_mark_pending_btree_node_frees(c);
-#endif
 	c->gc_count++;
 
-	if (test_bit(BCH_FS_need_another_gc, &c->flags) ||
-	    (!iter && bch2_test_restart_gc)) {
-		if (iter++ > 2) {
-			bch_info(c, "Unable to fix bucket gens, looping");
-			ret = -EINVAL;
-			goto out;
-		}
-
-		/*
-		 * XXX: make sure gens we fixed got saved
-		 */
-		bch_info(c, "Second GC pass needed, restarting:");
-		clear_bit(BCH_FS_need_another_gc, &c->flags);
-		__gc_pos_set(c, gc_phase(GC_PHASE_NOT_RUNNING));
-
-		bch2_gc_stripes_reset(c, metadata_only);
-		bch2_gc_alloc_reset(c, metadata_only);
-		bch2_gc_reflink_reset(c, metadata_only);
-		ret = bch2_gc_reset(c);
-		if (ret)
-			goto out;
-
-		/* flush fsck errors, reset counters */
-		bch2_flush_fsck_errs(c);
-		goto again;
-	}
+	bch2_journal_block(&c->journal);
 out:
-	if (!ret) {
-		bch2_journal_block(&c->journal);
-
-		ret   = bch2_gc_stripes_done(c, metadata_only) ?:
-			bch2_gc_reflink_done(c, metadata_only) ?:
-			bch2_gc_alloc_done(c, metadata_only) ?:
-			bch2_gc_done(c, initial, metadata_only);
+	ret   = bch2_gc_alloc_done(c) ?:
+		bch2_gc_done(c) ?:
+		bch2_gc_stripes_done(c) ?:
+		bch2_gc_reflink_done(c);
 
-		bch2_journal_unblock(&c->journal);
-	}
+	bch2_journal_unblock(&c->journal);
 
 	percpu_down_write(&c->mark_lock);
 	/* Indicates that gc is no longer in progress: */
-	__gc_pos_set(c, gc_phase(GC_PHASE_NOT_RUNNING));
+	__gc_pos_set(c, gc_phase(GC_PHASE_not_running));
 
 	bch2_gc_free(c);
 	percpu_up_write(&c->mark_lock);
@@ -1852,23 +1275,33 @@ static int gc_btree_gens_key(struct btree_trans *trans,
 	struct bkey_i *u;
 	int ret;
 
+	if (unlikely(test_bit(BCH_FS_going_ro, &c->flags)))
+		return -EROFS;
+
 	percpu_down_read(&c->mark_lock);
+	rcu_read_lock();
 	bkey_for_each_ptr(ptrs, ptr) {
-		struct bch_dev *ca = bch_dev_bkey_exists(c, ptr->dev);
+		struct bch_dev *ca = bch2_dev_rcu(c, ptr->dev);
+		if (!ca)
+			continue;
 
-		if (ptr_stale(ca, ptr) > 16) {
+		if (dev_ptr_stale(ca, ptr) > 16) {
+			rcu_read_unlock();
 			percpu_up_read(&c->mark_lock);
 			goto update;
 		}
 	}
 
 	bkey_for_each_ptr(ptrs, ptr) {
-		struct bch_dev *ca = bch_dev_bkey_exists(c, ptr->dev);
-		u8 *gen = &ca->oldest_gen[PTR_BUCKET_NR(ca, ptr)];
+		struct bch_dev *ca = bch2_dev_rcu(c, ptr->dev);
+		if (!ca)
+			continue;
 
+		u8 *gen = &ca->oldest_gen[PTR_BUCKET_NR(ca, ptr)];
 		if (gen_after(*gen, ptr->gen))
 			*gen = ptr->gen;
 	}
+	rcu_read_unlock();
 	percpu_up_read(&c->mark_lock);
 	return 0;
 update:
@@ -1881,10 +1314,9 @@ update:
 	return 0;
 }
 
-static int bch2_alloc_write_oldest_gen(struct btree_trans *trans, struct btree_iter *iter,
-				       struct bkey_s_c k)
+static int bch2_alloc_write_oldest_gen(struct btree_trans *trans, struct bch_dev *ca,
+				       struct btree_iter *iter, struct bkey_s_c k)
 {
-	struct bch_dev *ca = bch_dev_bkey_exists(trans->c, iter->pos.inode);
 	struct bch_alloc_v4 a_convert;
 	const struct bch_alloc_v4 *a = bch2_alloc_to_v4(k, &a_convert);
 	struct bkey_i_alloc_v4 *a_mut;
@@ -1899,7 +1331,7 @@ static int bch2_alloc_write_oldest_gen(struct btree_trans *trans, struct btree_i
 		return ret;
 
 	a_mut->v.oldest_gen = ca->oldest_gen[iter->pos.offset];
-	a_mut->v.data_type = alloc_data_type(a_mut->v, a_mut->v.data_type);
+	alloc_data_type_set(&a_mut->v, a_mut->v.data_type);
 
 	return bch2_trans_update(trans, iter, &a_mut->k_i, 0);
 }
@@ -1927,7 +1359,7 @@ int bch2_gc_gens(struct bch_fs *c)
 
 		ca->oldest_gen = kvmalloc(gens->nbuckets, GFP_KERNEL);
 		if (!ca->oldest_gen) {
-			percpu_ref_put(&ca->ref);
+			bch2_dev_put(ca);
 			ret = -BCH_ERR_ENOMEM_gc_gens;
 			goto err;
 		}
@@ -1945,7 +1377,7 @@ int bch2_gc_gens(struct bch_fs *c)
 			ret = bch2_trans_run(c,
 				for_each_btree_key_commit(trans, iter, i,
 						POS_MIN,
-						BTREE_ITER_PREFETCH|BTREE_ITER_ALL_SNAPSHOTS,
+						BTREE_ITER_prefetch|BTREE_ITER_all_snapshots,
 						k,
 						NULL, NULL,
 						BCH_TRANS_COMMIT_no_enospc,
@@ -1954,14 +1386,23 @@ int bch2_gc_gens(struct bch_fs *c)
 				goto err;
 		}
 
+	struct bch_dev *ca = NULL;
 	ret = bch2_trans_run(c,
 		for_each_btree_key_commit(trans, iter, BTREE_ID_alloc,
 				POS_MIN,
-				BTREE_ITER_PREFETCH,
+				BTREE_ITER_prefetch,
 				k,
 				NULL, NULL,
-				BCH_TRANS_COMMIT_no_enospc,
-			bch2_alloc_write_oldest_gen(trans, &iter, k)));
+				BCH_TRANS_COMMIT_no_enospc, ({
+			ca = bch2_dev_iterate(c, ca, k.k->p.inode);
+			if (!ca) {
+				bch2_btree_iter_set_pos(&iter, POS(k.k->p.inode + 1, 0));
+				continue;
+			}
+			bch2_alloc_write_oldest_gen(trans, ca, &iter, k);
+		})));
+	bch2_dev_put(ca);
+
 	if (ret)
 		goto err;
 
@@ -1985,87 +1426,23 @@ err:
 	return ret;
 }
 
-static int bch2_gc_thread(void *arg)
+static void bch2_gc_gens_work(struct work_struct *work)
 {
-	struct bch_fs *c = arg;
-	struct io_clock *clock = &c->io_clock[WRITE];
-	unsigned long last = atomic64_read(&clock->now);
-	unsigned last_kick = atomic_read(&c->kick_gc);
-
-	set_freezable();
-
-	while (1) {
-		while (1) {
-			set_current_state(TASK_INTERRUPTIBLE);
-
-			if (kthread_should_stop()) {
-				__set_current_state(TASK_RUNNING);
-				return 0;
-			}
-
-			if (atomic_read(&c->kick_gc) != last_kick)
-				break;
-
-			if (c->btree_gc_periodic) {
-				unsigned long next = last + c->capacity / 16;
-
-				if (atomic64_read(&clock->now) >= next)
-					break;
-
-				bch2_io_clock_schedule_timeout(clock, next);
-			} else {
-				schedule();
-			}
-
-			try_to_freeze();
-		}
-		__set_current_state(TASK_RUNNING);
-
-		last = atomic64_read(&clock->now);
-		last_kick = atomic_read(&c->kick_gc);
-
-		/*
-		 * Full gc is currently incompatible with btree key cache:
-		 */
-#if 0
-		ret = bch2_gc(c, false, false);
-#else
-		bch2_gc_gens(c);
-#endif
-		debug_check_no_locks_held();
-	}
-
-	return 0;
+	struct bch_fs *c = container_of(work, struct bch_fs, gc_gens_work);
+	bch2_gc_gens(c);
+	bch2_write_ref_put(c, BCH_WRITE_REF_gc_gens);
 }
 
-void bch2_gc_thread_stop(struct bch_fs *c)
+void bch2_gc_gens_async(struct bch_fs *c)
 {
-	struct task_struct *p;
-
-	p = c->gc_thread;
-	c->gc_thread = NULL;
-
-	if (p) {
-		kthread_stop(p);
-		put_task_struct(p);
-	}
+	if (bch2_write_ref_tryget(c, BCH_WRITE_REF_gc_gens) &&
+	    !queue_work(c->write_ref_wq, &c->gc_gens_work))
+		bch2_write_ref_put(c, BCH_WRITE_REF_gc_gens);
 }
 
-int bch2_gc_thread_start(struct bch_fs *c)
+void bch2_fs_gc_init(struct bch_fs *c)
 {
-	struct task_struct *p;
-
-	if (c->gc_thread)
-		return 0;
+	seqcount_init(&c->gc_pos_lock);
 
-	p = kthread_create(bch2_gc_thread, c, "bch-gc/%s", c->name);
-	if (IS_ERR(p)) {
-		bch_err_fn(c, PTR_ERR(p));
-		return PTR_ERR(p);
-	}
-
-	get_task_struct(p);
-	c->gc_thread = p;
-	wake_up_process(p);
-	return 0;
+	INIT_WORK(&c->gc_gens_work, bch2_gc_gens_work);
 }
diff --git a/fs/bcachefs/btree_gc.h b/fs/bcachefs/btree_gc.h
index 607575f83a..876d81e201 100644
--- a/fs/bcachefs/btree_gc.h
+++ b/fs/bcachefs/btree_gc.h
@@ -3,13 +3,11 @@
 #define _BCACHEFS_BTREE_GC_H
 
 #include "bkey.h"
+#include "btree_gc_types.h"
 #include "btree_types.h"
 
 int bch2_check_topology(struct bch_fs *);
-int bch2_gc(struct bch_fs *, bool, bool);
-int bch2_gc_gens(struct bch_fs *);
-void bch2_gc_thread_stop(struct bch_fs *);
-int bch2_gc_thread_start(struct bch_fs *);
+int bch2_check_allocations(struct bch_fs *);
 
 /*
  * For concurrent mark and sweep (with other index updates), we define a total
@@ -35,38 +33,17 @@ int bch2_gc_thread_start(struct bch_fs *);
 /* Position of (the start of) a gc phase: */
 static inline struct gc_pos gc_phase(enum gc_phase phase)
 {
-	return (struct gc_pos) {
-		.phase	= phase,
-		.pos	= POS_MIN,
-		.level	= 0,
-	};
+	return (struct gc_pos) { .phase	= phase, };
 }
 
-static inline int gc_pos_cmp(struct gc_pos l, struct gc_pos r)
-{
-	return  cmp_int(l.phase, r.phase) ?:
-		bpos_cmp(l.pos, r.pos) ?:
-		cmp_int(l.level, r.level);
-}
-
-static inline enum gc_phase btree_id_to_gc_phase(enum btree_id id)
-{
-	switch (id) {
-#define x(name, v, ...) case BTREE_ID_##name: return GC_PHASE_BTREE_##name;
-	BCH_BTREE_IDS()
-#undef x
-	default:
-		BUG();
-	}
-}
-
-static inline struct gc_pos gc_pos_btree(enum btree_id id,
-					 struct bpos pos, unsigned level)
+static inline struct gc_pos gc_pos_btree(enum btree_id btree, unsigned level,
+					 struct bpos pos)
 {
 	return (struct gc_pos) {
-		.phase	= btree_id_to_gc_phase(id),
-		.pos	= pos,
+		.phase	= GC_PHASE_btree,
+		.btree	= btree,
 		.level	= level,
+		.pos	= pos,
 	};
 }
 
@@ -76,19 +53,23 @@ static inline struct gc_pos gc_pos_btree(enum btree_id id,
  */
 static inline struct gc_pos gc_pos_btree_node(struct btree *b)
 {
-	return gc_pos_btree(b->c.btree_id, b->key.k.p, b->c.level);
+	return gc_pos_btree(b->c.btree_id, b->c.level, b->key.k.p);
 }
 
-/*
- * GC position of the pointer to a btree root: we don't use
- * gc_pos_pointer_to_btree_node() here to avoid a potential race with
- * btree_split() increasing the tree depth - the new root will have level > the
- * old root and thus have a greater gc position than the old root, but that
- * would be incorrect since once gc has marked the root it's not coming back.
- */
-static inline struct gc_pos gc_pos_btree_root(enum btree_id id)
+static inline int gc_btree_order(enum btree_id btree)
+{
+	if (btree == BTREE_ID_stripes)
+		return -1;
+	return btree;
+}
+
+static inline int gc_pos_cmp(struct gc_pos l, struct gc_pos r)
 {
-	return gc_pos_btree(id, SPOS_MAX, BTREE_MAX_DEPTH);
+	return   cmp_int(l.phase, r.phase) ?:
+		 cmp_int(gc_btree_order(l.btree),
+			 gc_btree_order(r.btree)) ?:
+		-cmp_int(l.level, r.level) ?:
+		 bpos_cmp(l.pos, r.pos);
 }
 
 static inline bool gc_visited(struct bch_fs *c, struct gc_pos pos)
@@ -104,11 +85,8 @@ static inline bool gc_visited(struct bch_fs *c, struct gc_pos pos)
 	return ret;
 }
 
-static inline void bch2_do_gc_gens(struct bch_fs *c)
-{
-	atomic_inc(&c->kick_gc);
-	if (c->gc_thread)
-		wake_up_process(c->gc_thread);
-}
+int bch2_gc_gens(struct bch_fs *);
+void bch2_gc_gens_async(struct bch_fs *);
+void bch2_fs_gc_init(struct bch_fs *);
 
 #endif /* _BCACHEFS_BTREE_GC_H */
diff --git a/fs/bcachefs/btree_gc_types.h b/fs/bcachefs/btree_gc_types.h
new file mode 100644
index 0000000000..b82c24bcc0
--- /dev/null
+++ b/fs/bcachefs/btree_gc_types.h
@@ -0,0 +1,29 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _BCACHEFS_BTREE_GC_TYPES_H
+#define _BCACHEFS_BTREE_GC_TYPES_H
+
+#include <linux/generic-radix-tree.h>
+
+enum gc_phase {
+	GC_PHASE_not_running,
+	GC_PHASE_start,
+	GC_PHASE_sb,
+	GC_PHASE_btree,
+};
+
+struct gc_pos {
+	enum gc_phase		phase:8;
+	enum btree_id		btree:8;
+	u16			level;
+	struct bpos		pos;
+};
+
+struct reflink_gc {
+	u64		offset;
+	u32		size;
+	u32		refcount;
+};
+
+typedef GENRADIX(struct reflink_gc) reflink_gc_table;
+
+#endif /* _BCACHEFS_BTREE_GC_TYPES_H */
diff --git a/fs/bcachefs/btree_io.c b/fs/bcachefs/btree_io.c
index aa9b6cbe32..7bca15c604 100644
--- a/fs/bcachefs/btree_io.c
+++ b/fs/bcachefs/btree_io.c
@@ -23,6 +23,18 @@
 
 #include <linux/sched/mm.h>
 
+static void bch2_btree_node_header_to_text(struct printbuf *out, struct btree_node *bn)
+{
+	prt_printf(out, "btree=%s l=%u seq %llux\n",
+		   bch2_btree_id_str(BTREE_NODE_ID(bn)),
+		   (unsigned) BTREE_NODE_LEVEL(bn), bn->keys.seq);
+	prt_str(out, "min: ");
+	bch2_bpos_to_text(out, bn->min_key);
+	prt_newline(out);
+	prt_str(out, "max: ");
+	bch2_bpos_to_text(out, bn->max_key);
+}
+
 void bch2_btree_node_io_unlock(struct btree *b)
 {
 	EBUG_ON(!btree_node_write_in_flight(b));
@@ -103,7 +115,7 @@ static void btree_bounce_free(struct bch_fs *c, size_t size,
 	if (used_mempool)
 		mempool_free(p, &c->btree_bounce_pool);
 	else
-		vpfree(p, size);
+		kvfree(p);
 }
 
 static void *btree_bounce_alloc(struct bch_fs *c, size_t size,
@@ -115,7 +127,7 @@ static void *btree_bounce_alloc(struct bch_fs *c, size_t size,
 	BUG_ON(size > c->opts.btree_node_size);
 
 	*used_mempool = false;
-	p = vpmalloc(size, __GFP_NOWARN|GFP_NOWAIT);
+	p = kvmalloc(size, __GFP_NOWARN|GFP_NOWAIT);
 	if (!p) {
 		*used_mempool = true;
 		p = mempool_alloc(&c->btree_bounce_pool, GFP_NOFS);
@@ -217,7 +229,6 @@ static bool should_compact_bset(struct btree *b, struct bset_tree *t,
 
 static bool bch2_drop_whiteouts(struct btree *b, enum compact_mode mode)
 {
-	struct bset_tree *t;
 	bool ret = false;
 
 	for_each_bset(b, t) {
@@ -288,8 +299,7 @@ bool bch2_compact_whiteouts(struct bch_fs *c, struct btree *b,
 
 static void btree_node_sort(struct bch_fs *c, struct btree *b,
 			    unsigned start_idx,
-			    unsigned end_idx,
-			    bool filter_whiteouts)
+			    unsigned end_idx)
 {
 	struct btree_node *out;
 	struct sort_iter_stack sort_iter;
@@ -320,7 +330,7 @@ static void btree_node_sort(struct bch_fs *c, struct btree *b,
 
 	start_time = local_clock();
 
-	u64s = bch2_sort_keys(out->keys.start, &sort_iter.iter, filter_whiteouts);
+	u64s = bch2_sort_keys(out->keys.start, &sort_iter.iter);
 
 	out->keys.u64s = cpu_to_le16(u64s);
 
@@ -426,13 +436,12 @@ static bool btree_node_compact(struct bch_fs *c, struct btree *b)
 			break;
 
 	if (b->nsets - unwritten_idx > 1) {
-		btree_node_sort(c, b, unwritten_idx,
-				b->nsets, false);
+		btree_node_sort(c, b, unwritten_idx, b->nsets);
 		ret = true;
 	}
 
 	if (unwritten_idx > 1) {
-		btree_node_sort(c, b, 0, unwritten_idx, false);
+		btree_node_sort(c, b, 0, unwritten_idx);
 		ret = true;
 	}
 
@@ -441,8 +450,6 @@ static bool btree_node_compact(struct bch_fs *c, struct btree *b)
 
 void bch2_btree_build_aux_trees(struct btree *b)
 {
-	struct bset_tree *t;
-
 	for_each_bset(b, t)
 		bch2_bset_build_aux_tree(b, t,
 				!bset_written(b, bset(b, t)) &&
@@ -512,7 +519,7 @@ void bch2_btree_init_next(struct btree_trans *trans, struct btree *b)
 
 static void btree_err_msg(struct printbuf *out, struct bch_fs *c,
 			  struct bch_dev *ca,
-			  struct btree *b, struct bset *i,
+			  struct btree *b, struct bset *i, struct bkey_packed *k,
 			  unsigned offset, int write)
 {
 	prt_printf(out, bch2_log_msg(c, "%s"),
@@ -524,28 +531,36 @@ static void btree_err_msg(struct printbuf *out, struct bch_fs *c,
 	prt_printf(out, "at btree ");
 	bch2_btree_pos_to_text(out, c, b);
 
-	prt_printf(out, "\n  node offset %u/%u",
+	printbuf_indent_add(out, 2);
+
+	prt_printf(out, "\nnode offset %u/%u",
 		   b->written, btree_ptr_sectors_written(&b->key));
 	if (i)
 		prt_printf(out, " bset u64s %u", le16_to_cpu(i->u64s));
+	if (k)
+		prt_printf(out, " bset byte offset %lu",
+			   (unsigned long)(void *)k -
+			   ((unsigned long)(void *)i & ~511UL));
 	prt_str(out, ": ");
 }
 
-__printf(9, 10)
+__printf(10, 11)
 static int __btree_err(int ret,
 		       struct bch_fs *c,
 		       struct bch_dev *ca,
 		       struct btree *b,
 		       struct bset *i,
+		       struct bkey_packed *k,
 		       int write,
 		       bool have_retry,
 		       enum bch_sb_error_id err_type,
 		       const char *fmt, ...)
 {
 	struct printbuf out = PRINTBUF;
+	bool silent = c->curr_recovery_pass == BCH_RECOVERY_PASS_scan_for_btree_nodes;
 	va_list args;
 
-	btree_err_msg(&out, c, ca, b, i, b->written, write);
+	btree_err_msg(&out, c, ca, b, i, k, b->written, write);
 
 	va_start(args, fmt);
 	prt_vprintf(&out, fmt, args);
@@ -564,12 +579,14 @@ static int __btree_err(int ret,
 	if (!have_retry && ret == -BCH_ERR_btree_node_read_err_must_retry)
 		ret = -BCH_ERR_btree_node_read_err_bad_node;
 
-	if (ret != -BCH_ERR_btree_node_read_err_fixable)
+	if (!silent && ret != -BCH_ERR_btree_node_read_err_fixable)
 		bch2_sb_error_count(c, err_type);
 
 	switch (ret) {
 	case -BCH_ERR_btree_node_read_err_fixable:
-		ret = bch2_fsck_err(c, FSCK_CAN_FIX, err_type, "%s", out.buf);
+		ret = !silent
+			? bch2_fsck_err(c, FSCK_CAN_FIX, err_type, "%s", out.buf)
+			: -BCH_ERR_fsck_fix;
 		if (ret != -BCH_ERR_fsck_fix &&
 		    ret != -BCH_ERR_fsck_ignore)
 			goto fsck_err;
@@ -577,15 +594,17 @@ static int __btree_err(int ret,
 		break;
 	case -BCH_ERR_btree_node_read_err_want_retry:
 	case -BCH_ERR_btree_node_read_err_must_retry:
-		bch2_print_string_as_lines(KERN_ERR, out.buf);
+		if (!silent)
+			bch2_print_string_as_lines(KERN_ERR, out.buf);
 		break;
 	case -BCH_ERR_btree_node_read_err_bad_node:
-		bch2_print_string_as_lines(KERN_ERR, out.buf);
-		bch2_topology_error(c);
-		ret = bch2_run_explicit_recovery_pass(c, BCH_RECOVERY_PASS_check_topology) ?: -EIO;
+		if (!silent)
+			bch2_print_string_as_lines(KERN_ERR, out.buf);
+		ret = bch2_topology_error(c);
 		break;
 	case -BCH_ERR_btree_node_read_err_incompatible:
-		bch2_print_string_as_lines(KERN_ERR, out.buf);
+		if (!silent)
+			bch2_print_string_as_lines(KERN_ERR, out.buf);
 		ret = -BCH_ERR_fsck_errors_not_fixed;
 		break;
 	default:
@@ -597,9 +616,9 @@ fsck_err:
 	return ret;
 }
 
-#define btree_err(type, c, ca, b, i, _err_type, msg, ...)		\
+#define btree_err(type, c, ca, b, i, k, _err_type, msg, ...)		\
 ({									\
-	int _ret = __btree_err(type, c, ca, b, i, write, have_retry,	\
+	int _ret = __btree_err(type, c, ca, b, i, k, write, have_retry,	\
 			       BCH_FSCK_ERR_##_err_type,		\
 			       msg, ##__VA_ARGS__);			\
 									\
@@ -620,8 +639,6 @@ fsck_err:
 __cold
 void bch2_btree_node_drop_keys_outside_node(struct btree *b)
 {
-	struct bset_tree *t;
-
 	for_each_bset(b, t) {
 		struct bset *i = bset(b, t);
 		struct bkey_packed *k;
@@ -655,6 +672,7 @@ void bch2_btree_node_drop_keys_outside_node(struct btree *b)
 	 */
 	bch2_bset_set_no_aux_tree(b, b->set);
 	bch2_btree_build_aux_trees(b);
+	b->nr = bch2_btree_node_count_keys(b);
 
 	struct bkey_s_c k;
 	struct bkey unpacked;
@@ -677,7 +695,7 @@ static int validate_bset(struct bch_fs *c, struct bch_dev *ca,
 
 	btree_err_on(!bch2_version_compatible(version),
 		     -BCH_ERR_btree_node_read_err_incompatible,
-		     c, ca, b, i,
+		     c, ca, b, i, NULL,
 		     btree_node_unsupported_version,
 		     "unsupported bset version %u.%u",
 		     BCH_VERSION_MAJOR(version),
@@ -685,7 +703,7 @@ static int validate_bset(struct bch_fs *c, struct bch_dev *ca,
 
 	if (btree_err_on(version < c->sb.version_min,
 			 -BCH_ERR_btree_node_read_err_fixable,
-			 c, NULL, b, i,
+			 c, NULL, b, i, NULL,
 			 btree_node_bset_older_than_sb_min,
 			 "bset version %u older than superblock version_min %u",
 			 version, c->sb.version_min)) {
@@ -698,7 +716,7 @@ static int validate_bset(struct bch_fs *c, struct bch_dev *ca,
 	if (btree_err_on(BCH_VERSION_MAJOR(version) >
 			 BCH_VERSION_MAJOR(c->sb.version),
 			 -BCH_ERR_btree_node_read_err_fixable,
-			 c, NULL, b, i,
+			 c, NULL, b, i, NULL,
 			 btree_node_bset_newer_than_sb,
 			 "bset version %u newer than superblock version %u",
 			 version, c->sb.version)) {
@@ -710,13 +728,13 @@ static int validate_bset(struct bch_fs *c, struct bch_dev *ca,
 
 	btree_err_on(BSET_SEPARATE_WHITEOUTS(i),
 		     -BCH_ERR_btree_node_read_err_incompatible,
-		     c, ca, b, i,
+		     c, ca, b, i, NULL,
 		     btree_node_unsupported_version,
 		     "BSET_SEPARATE_WHITEOUTS no longer supported");
 
 	if (btree_err_on(offset + sectors > btree_sectors(c),
 			 -BCH_ERR_btree_node_read_err_fixable,
-			 c, ca, b, i,
+			 c, ca, b, i, NULL,
 			 bset_past_end_of_btree_node,
 			 "bset past end of btree node")) {
 		i->u64s = 0;
@@ -726,13 +744,13 @@ static int validate_bset(struct bch_fs *c, struct bch_dev *ca,
 
 	btree_err_on(offset && !i->u64s,
 		     -BCH_ERR_btree_node_read_err_fixable,
-		     c, ca, b, i,
+		     c, ca, b, i, NULL,
 		     bset_empty,
 		     "empty bset");
 
 	btree_err_on(BSET_OFFSET(i) && BSET_OFFSET(i) != offset,
 		     -BCH_ERR_btree_node_read_err_want_retry,
-		     c, ca, b, i,
+		     c, ca, b, i, NULL,
 		     bset_wrong_sector_offset,
 		     "bset at wrong sector offset");
 
@@ -748,20 +766,20 @@ static int validate_bset(struct bch_fs *c, struct bch_dev *ca,
 			/* XXX endianness */
 			btree_err_on(bp->seq != bn->keys.seq,
 				     -BCH_ERR_btree_node_read_err_must_retry,
-				     c, ca, b, NULL,
+				     c, ca, b, NULL, NULL,
 				     bset_bad_seq,
 				     "incorrect sequence number (wrong btree node)");
 		}
 
 		btree_err_on(BTREE_NODE_ID(bn) != b->c.btree_id,
 			     -BCH_ERR_btree_node_read_err_must_retry,
-			     c, ca, b, i,
+			     c, ca, b, i, NULL,
 			     btree_node_bad_btree,
 			     "incorrect btree id");
 
 		btree_err_on(BTREE_NODE_LEVEL(bn) != b->c.level,
 			     -BCH_ERR_btree_node_read_err_must_retry,
-			     c, ca, b, i,
+			     c, ca, b, i, NULL,
 			     btree_node_bad_level,
 			     "incorrect level");
 
@@ -780,7 +798,7 @@ static int validate_bset(struct bch_fs *c, struct bch_dev *ca,
 
 			btree_err_on(!bpos_eq(b->data->min_key, bp->min_key),
 				     -BCH_ERR_btree_node_read_err_must_retry,
-				     c, ca, b, NULL,
+				     c, ca, b, NULL, NULL,
 				     btree_node_bad_min_key,
 				     "incorrect min_key: got %s should be %s",
 				     (printbuf_reset(&buf1),
@@ -791,7 +809,7 @@ static int validate_bset(struct bch_fs *c, struct bch_dev *ca,
 
 		btree_err_on(!bpos_eq(bn->max_key, b->key.k.p),
 			     -BCH_ERR_btree_node_read_err_must_retry,
-			     c, ca, b, i,
+			     c, ca, b, i, NULL,
 			     btree_node_bad_max_key,
 			     "incorrect max key %s",
 			     (printbuf_reset(&buf1),
@@ -803,7 +821,7 @@ static int validate_bset(struct bch_fs *c, struct bch_dev *ca,
 
 		btree_err_on(bch2_bkey_format_invalid(c, &bn->format, write, &buf1),
 			     -BCH_ERR_btree_node_read_err_bad_node,
-			     c, ca, b, i,
+			     c, ca, b, i, NULL,
 			     btree_node_bad_format,
 			     "invalid bkey format: %s\n  %s", buf1.buf,
 			     (printbuf_reset(&buf2),
@@ -831,7 +849,7 @@ static int bset_key_invalid(struct bch_fs *c, struct btree *b,
 		(rw == WRITE ? bch2_bkey_val_invalid(c, k, READ, err) : 0);
 }
 
-static bool __bkey_valid(struct bch_fs *c, struct btree *b,
+static bool bkey_packed_valid(struct bch_fs *c, struct btree *b,
 			 struct bset *i, struct bkey_packed *k)
 {
 	if (bkey_p_next(k) > vstruct_last(i))
@@ -840,6 +858,9 @@ static bool __bkey_valid(struct bch_fs *c, struct btree *b,
 	if (k->format > KEY_FORMAT_CURRENT)
 		return false;
 
+	if (!bkeyp_u64s_valid(&b->format, k))
+		return false;
+
 	struct printbuf buf = PRINTBUF;
 	struct bkey tmp;
 	struct bkey_s u = __bkey_disassemble(b, k, &tmp);
@@ -867,7 +888,7 @@ static int validate_bset_keys(struct bch_fs *c, struct btree *b,
 
 		if (btree_err_on(bkey_p_next(k) > vstruct_last(i),
 				 -BCH_ERR_btree_node_read_err_fixable,
-				 c, NULL, b, i,
+				 c, NULL, b, i, k,
 				 btree_node_bkey_past_bset_end,
 				 "key extends past end of bset")) {
 			i->u64s = cpu_to_le16((u64 *) k - i->_data);
@@ -876,12 +897,20 @@ static int validate_bset_keys(struct bch_fs *c, struct btree *b,
 
 		if (btree_err_on(k->format > KEY_FORMAT_CURRENT,
 				 -BCH_ERR_btree_node_read_err_fixable,
-				 c, NULL, b, i,
+				 c, NULL, b, i, k,
 				 btree_node_bkey_bad_format,
 				 "invalid bkey format %u", k->format))
 			goto drop_this_key;
 
-		/* XXX: validate k->u64s */
+		if (btree_err_on(!bkeyp_u64s_valid(&b->format, k),
+				 -BCH_ERR_btree_node_read_err_fixable,
+				 c, NULL, b, i, k,
+				 btree_node_bkey_bad_u64s,
+				 "bad k->u64s %u (min %u max %zu)", k->u64s,
+				 bkeyp_key_u64s(&b->format, k),
+				 U8_MAX - BKEY_U64s + bkeyp_key_u64s(&b->format, k)))
+			goto drop_this_key;
+
 		if (!write)
 			bch2_bkey_compat(b->c.level, b->c.btree_id, version,
 				    BSET_BIG_ENDIAN(i), write,
@@ -897,7 +926,7 @@ static int validate_bset_keys(struct bch_fs *c, struct btree *b,
 			bch2_bkey_val_to_text(&buf, c, u.s_c);
 
 			btree_err(-BCH_ERR_btree_node_read_err_fixable,
-				  c, NULL, b, i,
+				  c, NULL, b, i, k,
 				  btree_node_bad_bkey,
 				  "invalid bkey: %s", buf.buf);
 			goto drop_this_key;
@@ -918,7 +947,7 @@ static int validate_bset_keys(struct bch_fs *c, struct btree *b,
 			bch2_bkey_to_text(&buf, u.k);
 
 			if (btree_err(-BCH_ERR_btree_node_read_err_fixable,
-				      c, NULL, b, i,
+				      c, NULL, b, i, k,
 				      btree_node_bkey_out_of_order,
 				      "%s", buf.buf))
 				goto drop_this_key;
@@ -938,13 +967,12 @@ drop_this_key:
 			 * do
 			 */
 
-			if (!__bkey_valid(c, b, i, (void *) ((u64 *) k + next_good_key))) {
+			if (!bkey_packed_valid(c, b, i, (void *) ((u64 *) k + next_good_key))) {
 				for (next_good_key = 1;
 				     next_good_key < (u64 *) vstruct_last(i) - (u64 *) k;
 				     next_good_key++)
-					if (__bkey_valid(c, b, i, (void *) ((u64 *) k + next_good_key)))
+					if (bkey_packed_valid(c, b, i, (void *) ((u64 *) k + next_good_key)))
 						goto got_good_key;
-
 			}
 
 			/*
@@ -988,13 +1016,13 @@ int bch2_btree_node_read_done(struct bch_fs *c, struct bch_dev *ca,
 
 	if (bch2_meta_read_fault("btree"))
 		btree_err(-BCH_ERR_btree_node_read_err_must_retry,
-			  c, ca, b, NULL,
+			  c, ca, b, NULL, NULL,
 			  btree_node_fault_injected,
 			  "dynamic fault");
 
 	btree_err_on(le64_to_cpu(b->data->magic) != bset_magic(c),
 		     -BCH_ERR_btree_node_read_err_must_retry,
-		     c, ca, b, NULL,
+		     c, ca, b, NULL, NULL,
 		     btree_node_bad_magic,
 		     "bad magic: want %llx, got %llx",
 		     bset_magic(c), le64_to_cpu(b->data->magic));
@@ -1009,20 +1037,21 @@ int bch2_btree_node_read_done(struct bch_fs *c, struct bch_dev *ca,
 
 		btree_err_on(b->data->keys.seq != bp->seq,
 			     -BCH_ERR_btree_node_read_err_must_retry,
-			     c, ca, b, NULL,
+			     c, ca, b, NULL, NULL,
 			     btree_node_bad_seq,
-			     "got wrong btree node (want %llx got %llx)\n"
-			     "got btree %s level %llu pos %s",
-			     bp->seq, b->data->keys.seq,
-			     bch2_btree_id_str(BTREE_NODE_ID(b->data)),
-			     BTREE_NODE_LEVEL(b->data),
-			     buf.buf);
+			     "got wrong btree node: got\n%s",
+			     (printbuf_reset(&buf),
+			      bch2_btree_node_header_to_text(&buf, b->data),
+			      buf.buf));
 	} else {
 		btree_err_on(!b->data->keys.seq,
 			     -BCH_ERR_btree_node_read_err_must_retry,
-			     c, ca, b, NULL,
+			     c, ca, b, NULL, NULL,
 			     btree_node_bad_seq,
-			     "bad btree header: seq 0");
+			     "bad btree header: seq 0\n%s",
+			     (printbuf_reset(&buf),
+			      bch2_btree_node_header_to_text(&buf, b->data),
+			      buf.buf));
 	}
 
 	while (b->written < (ptr_written ?: btree_sectors(c))) {
@@ -1036,7 +1065,7 @@ int bch2_btree_node_read_done(struct bch_fs *c, struct bch_dev *ca,
 
 			btree_err_on(!bch2_checksum_type_valid(c, BSET_CSUM_TYPE(i)),
 				     -BCH_ERR_btree_node_read_err_want_retry,
-				     c, ca, b, i,
+				     c, ca, b, i, NULL,
 				     bset_unknown_csum,
 				     "unknown checksum type %llu", BSET_CSUM_TYPE(i));
 
@@ -1049,7 +1078,7 @@ int bch2_btree_node_read_done(struct bch_fs *c, struct bch_dev *ca,
 
 			btree_err_on(csum_bad,
 				     -BCH_ERR_btree_node_read_err_want_retry,
-				     c, ca, b, i,
+				     c, ca, b, i, NULL,
 				     bset_bad_csum,
 				     "%s",
 				     (printbuf_reset(&buf),
@@ -1058,13 +1087,13 @@ int bch2_btree_node_read_done(struct bch_fs *c, struct bch_dev *ca,
 
 			ret = bset_encrypt(c, i, b->written << 9);
 			if (bch2_fs_fatal_err_on(ret, c,
-					"error decrypting btree node: %i", ret))
+					"decrypting btree node: %s", bch2_err_str(ret)))
 				goto fsck_err;
 
 			btree_err_on(btree_node_type_is_extents(btree_node_type(b)) &&
 				     !BTREE_NODE_NEW_EXTENT_OVERWRITE(b->data),
 				     -BCH_ERR_btree_node_read_err_incompatible,
-				     c, NULL, b, NULL,
+				     c, NULL, b, NULL, NULL,
 				     btree_node_unsupported_version,
 				     "btree node does not have NEW_EXTENT_OVERWRITE set");
 
@@ -1078,19 +1107,19 @@ int bch2_btree_node_read_done(struct bch_fs *c, struct bch_dev *ca,
 
 			btree_err_on(!bch2_checksum_type_valid(c, BSET_CSUM_TYPE(i)),
 				     -BCH_ERR_btree_node_read_err_want_retry,
-				     c, ca, b, i,
+				     c, ca, b, i, NULL,
 				     bset_unknown_csum,
 				     "unknown checksum type %llu", BSET_CSUM_TYPE(i));
 
 			nonce = btree_nonce(i, b->written << 9);
 			struct bch_csum csum = csum_vstruct(c, BSET_CSUM_TYPE(i), nonce, bne);
 			csum_bad = bch2_crc_cmp(bne->csum, csum);
-			if (csum_bad)
+			if (ca && csum_bad)
 				bch2_io_error(ca, BCH_MEMBER_ERROR_checksum);
 
 			btree_err_on(csum_bad,
 				     -BCH_ERR_btree_node_read_err_want_retry,
-				     c, ca, b, i,
+				     c, ca, b, i, NULL,
 				     bset_bad_csum,
 				     "%s",
 				     (printbuf_reset(&buf),
@@ -1099,7 +1128,7 @@ int bch2_btree_node_read_done(struct bch_fs *c, struct bch_dev *ca,
 
 			ret = bset_encrypt(c, i, b->written << 9);
 			if (bch2_fs_fatal_err_on(ret, c,
-					"error decrypting btree node: %i\n", ret))
+					"decrypting btree node: %s", bch2_err_str(ret)))
 				goto fsck_err;
 
 			sectors = vstruct_sectors(bne, c->block_bits);
@@ -1128,14 +1157,14 @@ int bch2_btree_node_read_done(struct bch_fs *c, struct bch_dev *ca,
 
 		btree_err_on(blacklisted && first,
 			     -BCH_ERR_btree_node_read_err_fixable,
-			     c, ca, b, i,
+			     c, ca, b, i, NULL,
 			     bset_blacklisted_journal_seq,
 			     "first btree node bset has blacklisted journal seq (%llu)",
 			     le64_to_cpu(i->journal_seq));
 
 		btree_err_on(blacklisted && ptr_written,
 			     -BCH_ERR_btree_node_read_err_fixable,
-			     c, ca, b, i,
+			     c, ca, b, i, NULL,
 			     first_bset_blacklisted_journal_seq,
 			     "found blacklisted bset (journal seq %llu) in btree node at offset %u-%u/%u",
 			     le64_to_cpu(i->journal_seq),
@@ -1154,7 +1183,7 @@ int bch2_btree_node_read_done(struct bch_fs *c, struct bch_dev *ca,
 	if (ptr_written) {
 		btree_err_on(b->written < ptr_written,
 			     -BCH_ERR_btree_node_read_err_want_retry,
-			     c, ca, b, NULL,
+			     c, ca, b, NULL, NULL,
 			     btree_node_data_missing,
 			     "btree node data missing: expected %u sectors, found %u",
 			     ptr_written, b->written);
@@ -1167,7 +1196,7 @@ int bch2_btree_node_read_done(struct bch_fs *c, struct bch_dev *ca,
 								      le64_to_cpu(bne->keys.journal_seq),
 								      true),
 				     -BCH_ERR_btree_node_read_err_want_retry,
-				     c, ca, b, NULL,
+				     c, ca, b, NULL, NULL,
 				     btree_node_bset_after_end,
 				     "found bset signature after last bset");
 	}
@@ -1211,7 +1240,7 @@ int bch2_btree_node_read_done(struct bch_fs *c, struct bch_dev *ca,
 			bch2_bkey_val_to_text(&buf, c, u.s_c);
 
 			btree_err(-BCH_ERR_btree_node_read_err_fixable,
-				  c, NULL, b, i,
+				  c, NULL, b, i, k,
 				  btree_node_bad_bkey,
 				  "%s", buf.buf);
 
@@ -1239,12 +1268,14 @@ int bch2_btree_node_read_done(struct bch_fs *c, struct bch_dev *ca,
 
 	btree_node_reset_sib_u64s(b);
 
+	rcu_read_lock();
 	bkey_for_each_ptr(bch2_bkey_ptrs(bkey_i_to_s(&b->key)), ptr) {
-		struct bch_dev *ca2 = bch_dev_bkey_exists(c, ptr->dev);
+		struct bch_dev *ca2 = bch2_dev_rcu(c, ptr->dev);
 
-		if (ca2->mi.state != BCH_MEMBER_STATE_rw)
+		if (!ca2 || ca2->mi.state != BCH_MEMBER_STATE_rw)
 			set_btree_node_need_rewrite(b);
 	}
+	rcu_read_unlock();
 
 	if (!ptr_written)
 		set_btree_node_need_rewrite(b);
@@ -1255,10 +1286,12 @@ out:
 	return retry_read;
 fsck_err:
 	if (ret == -BCH_ERR_btree_node_read_err_want_retry ||
-	    ret == -BCH_ERR_btree_node_read_err_must_retry)
+	    ret == -BCH_ERR_btree_node_read_err_must_retry) {
 		retry_read = 1;
-	else
+	} else {
 		set_btree_node_read_error(b);
+		bch2_btree_lost_data(c, b->c.btree_id);
+	}
 	goto out;
 }
 
@@ -1267,8 +1300,8 @@ static void btree_node_read_work(struct work_struct *work)
 	struct btree_read_bio *rb =
 		container_of(work, struct btree_read_bio, work);
 	struct bch_fs *c	= rb->c;
+	struct bch_dev *ca	= rb->have_ioref ? bch2_dev_have_ref(c, rb->pick.ptr.dev) : NULL;
 	struct btree *b		= rb->b;
-	struct bch_dev *ca	= bch_dev_bkey_exists(c, rb->pick.ptr.dev);
 	struct bio *bio		= &rb->bio;
 	struct bch_io_failures failed = { .nr = 0 };
 	struct printbuf buf = PRINTBUF;
@@ -1280,8 +1313,8 @@ static void btree_node_read_work(struct work_struct *work)
 	while (1) {
 		retry = true;
 		bch_info(c, "retrying read");
-		ca = bch_dev_bkey_exists(c, rb->pick.ptr.dev);
-		rb->have_ioref		= bch2_dev_get_ioref(ca, READ);
+		ca = bch2_dev_get_ioref(c, rb->pick.ptr.dev, READ);
+		rb->have_ioref		= ca != NULL;
 		bio_reset(bio, NULL, REQ_OP_READ|REQ_SYNC|REQ_META);
 		bio->bi_iter.bi_sector	= rb->pick.ptr.offset;
 		bio->bi_iter.bi_size	= btree_buf_bytes(b);
@@ -1295,7 +1328,7 @@ static void btree_node_read_work(struct work_struct *work)
 start:
 		printbuf_reset(&buf);
 		bch2_btree_pos_to_text(&buf, c, b);
-		bch2_dev_io_err_on(bio->bi_status, ca, BCH_MEMBER_ERROR_read,
+		bch2_dev_io_err_on(ca && bio->bi_status, ca, BCH_MEMBER_ERROR_read,
 				   "btree read error %s for %s",
 				   bch2_blk_status_to_str(bio->bi_status), buf.buf);
 		if (rb->have_ioref)
@@ -1319,6 +1352,7 @@ start:
 
 		if (!can_retry) {
 			set_btree_node_read_error(b);
+			bch2_btree_lost_data(c, b->c.btree_id);
 			break;
 		}
 	}
@@ -1327,10 +1361,12 @@ start:
 			       rb->start_time);
 	bio_put(&rb->bio);
 
-	if (saw_error && !btree_node_read_error(b)) {
+	if (saw_error &&
+	    !btree_node_read_error(b) &&
+	    c->curr_recovery_pass != BCH_RECOVERY_PASS_scan_for_btree_nodes) {
 		printbuf_reset(&buf);
 		bch2_bpos_to_text(&buf, b->key.k.p);
-		bch_info(c, "%s: rewriting btree node at btree=%s level=%u %s due to error",
+		bch_err_ratelimited(c, "%s: rewriting btree node at btree=%s level=%u %s due to error",
 			 __func__, bch2_btree_id_str(b->c.btree_id), b->c.level, buf.buf);
 
 		bch2_btree_node_rewrite_async(c, b);
@@ -1348,12 +1384,12 @@ static void btree_node_read_endio(struct bio *bio)
 	struct bch_fs *c	= rb->c;
 
 	if (rb->have_ioref) {
-		struct bch_dev *ca = bch_dev_bkey_exists(c, rb->pick.ptr.dev);
+		struct bch_dev *ca = bch2_dev_have_ref(c, rb->pick.ptr.dev);
 
 		bch2_latency_acct(ca, rb->start_time, READ);
 	}
 
-	queue_work(c->io_complete_wq, &rb->work);
+	queue_work(c->btree_read_complete_wq, &rb->work);
 }
 
 struct btree_node_read_all {
@@ -1440,18 +1476,18 @@ static CLOSURE_CALLBACK(btree_node_read_all_replicas_done)
 
 		written2 = btree_node_sectors_written(c, ra->buf[i]);
 		if (btree_err_on(written2 != written, -BCH_ERR_btree_node_read_err_fixable,
-				 c, NULL, b, NULL,
+				 c, NULL, b, NULL, NULL,
 				 btree_node_replicas_sectors_written_mismatch,
 				 "btree node sectors written mismatch: %u != %u",
 				 written, written2) ||
 		    btree_err_on(btree_node_has_extra_bsets(c, written2, ra->buf[i]),
 				 -BCH_ERR_btree_node_read_err_fixable,
-				 c, NULL, b, NULL,
+				 c, NULL, b, NULL, NULL,
 				 btree_node_bset_after_end,
 				 "found bset signature after last bset") ||
 		    btree_err_on(memcmp(ra->buf[best], ra->buf[i], written << 9),
 				 -BCH_ERR_btree_node_read_err_fixable,
-				 c, NULL, b, NULL,
+				 c, NULL, b, NULL, NULL,
 				 btree_node_replicas_data_mismatch,
 				 "btree node replicas content mismatch"))
 			dump_bset_maps = true;
@@ -1518,9 +1554,10 @@ fsck_err:
 		ret = -1;
 	}
 
-	if (ret)
+	if (ret) {
 		set_btree_node_read_error(b);
-	else if (*saw_error)
+		bch2_btree_lost_data(c, b->c.btree_id);
+	} else if (*saw_error)
 		bch2_btree_node_rewrite_async(c, b);
 
 	for (i = 0; i < ra->nr; i++) {
@@ -1544,7 +1581,7 @@ static void btree_node_read_all_replicas_endio(struct bio *bio)
 	struct btree_node_read_all *ra = rb->ra;
 
 	if (rb->have_ioref) {
-		struct bch_dev *ca = bch_dev_bkey_exists(c, rb->pick.ptr.dev);
+		struct bch_dev *ca = bch2_dev_have_ref(c, rb->pick.ptr.dev);
 
 		bch2_latency_acct(ca, rb->start_time, READ);
 	}
@@ -1586,14 +1623,14 @@ static int btree_node_read_all_replicas(struct bch_fs *c, struct btree *b, bool
 
 	i = 0;
 	bkey_for_each_ptr_decode(k.k, ptrs, pick, entry) {
-		struct bch_dev *ca = bch_dev_bkey_exists(c, pick.ptr.dev);
+		struct bch_dev *ca = bch2_dev_get_ioref(c, pick.ptr.dev, READ);
 		struct btree_read_bio *rb =
 			container_of(ra->bio[i], struct btree_read_bio, bio);
 		rb->c			= c;
 		rb->b			= b;
 		rb->ra			= ra;
 		rb->start_time		= local_clock();
-		rb->have_ioref		= bch2_dev_get_ioref(ca, READ);
+		rb->have_ioref		= ca != NULL;
 		rb->idx			= i;
 		rb->pick		= pick;
 		rb->bio.bi_iter.bi_sector = pick.ptr.offset;
@@ -1619,7 +1656,7 @@ static int btree_node_read_all_replicas(struct bch_fs *c, struct btree *b, bool
 		btree_node_read_all_replicas_done(&ra->cl.work);
 	} else {
 		continue_at(&ra->cl, btree_node_read_all_replicas_done,
-			    c->io_complete_wq);
+			    c->btree_read_complete_wq);
 	}
 
 	return 0;
@@ -1649,20 +1686,21 @@ void bch2_btree_node_read(struct btree_trans *trans, struct btree *b,
 
 		prt_str(&buf, "btree node read error: no device to read from\n at ");
 		bch2_btree_pos_to_text(&buf, c, b);
-		bch_err(c, "%s", buf.buf);
+		bch_err_ratelimited(c, "%s", buf.buf);
 
 		if (c->recovery_passes_explicit & BIT_ULL(BCH_RECOVERY_PASS_check_topology) &&
 		    c->curr_recovery_pass > BCH_RECOVERY_PASS_check_topology)
 			bch2_fatal_error(c);
 
 		set_btree_node_read_error(b);
+		bch2_btree_lost_data(c, b->c.btree_id);
 		clear_btree_node_read_in_flight(b);
 		wake_up_bit(&b->flags, BTREE_NODE_read_in_flight);
 		printbuf_exit(&buf);
 		return;
 	}
 
-	ca = bch_dev_bkey_exists(c, pick.ptr.dev);
+	ca = bch2_dev_get_ioref(c, pick.ptr.dev, READ);
 
 	bio = bio_alloc_bioset(NULL,
 			       buf_pages(b->data, btree_buf_bytes(b)),
@@ -1674,7 +1712,7 @@ void bch2_btree_node_read(struct btree_trans *trans, struct btree *b,
 	rb->b			= b;
 	rb->ra			= NULL;
 	rb->start_time		= local_clock();
-	rb->have_ioref		= bch2_dev_get_ioref(ca, READ);
+	rb->have_ioref		= ca != NULL;
 	rb->pick		= pick;
 	INIT_WORK(&rb->work, btree_node_read_work);
 	bio->bi_iter.bi_sector	= pick.ptr.offset;
@@ -1699,7 +1737,7 @@ void bch2_btree_node_read(struct btree_trans *trans, struct btree *b,
 		if (sync)
 			btree_node_read_work(&rb->work);
 		else
-			queue_work(c->io_complete_wq, &rb->work);
+			queue_work(c->btree_read_complete_wq, &rb->work);
 	}
 }
 
@@ -1737,7 +1775,7 @@ static int __bch2_btree_root_read(struct btree_trans *trans, enum btree_id id,
 		list_move(&b->list, &c->btree_cache.freeable);
 		mutex_unlock(&c->btree_cache.lock);
 
-		ret = -EIO;
+		ret = -BCH_ERR_btree_node_read_error;
 		goto err;
 	}
 
@@ -1829,7 +1867,6 @@ static void btree_node_write_work(struct work_struct *work)
 		container_of(work, struct btree_write_bio, work);
 	struct bch_fs *c	= wbio->wbio.c;
 	struct btree *b		= wbio->wbio.bio.bi_private;
-	struct bch_extent_ptr *ptr;
 	int ret = 0;
 
 	btree_bounce_free(c,
@@ -1841,7 +1878,7 @@ static void btree_node_write_work(struct work_struct *work)
 		bch2_dev_list_has_dev(wbio->wbio.failed, ptr->dev));
 
 	if (!bch2_bkey_nr_ptrs(bkey_i_to_s_c(&wbio->key))) {
-		ret = -BCH_ERR_btree_write_all_failed;
+		ret = -BCH_ERR_btree_node_write_all_failed;
 		goto err;
 	}
 
@@ -1852,7 +1889,7 @@ static void btree_node_write_work(struct work_struct *work)
 	} else {
 		ret = bch2_trans_do(c, NULL, NULL, 0,
 			bch2_btree_node_update_key_get_iter(trans, b, &wbio->key,
-					BCH_WATERMARK_reclaim|
+					BCH_WATERMARK_interior_updates|
 					BCH_TRANS_COMMIT_journal_reclaim|
 					BCH_TRANS_COMMIT_no_enospc|
 					BCH_TRANS_COMMIT_no_check_rw,
@@ -1866,8 +1903,8 @@ out:
 	return;
 err:
 	set_btree_node_noevict(b);
-	if (!bch2_err_matches(ret, EROFS))
-		bch2_fs_fatal_error(c, "fatal error writing btree node: %s", bch2_err_str(ret));
+	bch2_fs_fatal_err_on(!bch2_err_matches(ret, EROFS), c,
+			     "writing btree node: %s", bch2_err_str(ret));
 	goto out;
 }
 
@@ -1879,13 +1916,14 @@ static void btree_node_write_endio(struct bio *bio)
 	struct btree_write_bio *wb	= container_of(orig, struct btree_write_bio, wbio);
 	struct bch_fs *c		= wbio->c;
 	struct btree *b			= wbio->bio.bi_private;
-	struct bch_dev *ca		= bch_dev_bkey_exists(c, wbio->dev);
+	struct bch_dev *ca		= wbio->have_ioref ? bch2_dev_have_ref(c, wbio->dev) : NULL;
 	unsigned long flags;
 
 	if (wbio->have_ioref)
 		bch2_latency_acct(ca, wbio->submit_time, WRITE);
 
-	if (bch2_dev_io_err_on(bio->bi_status, ca, BCH_MEMBER_ERROR_write,
+	if (!ca ||
+	    bch2_dev_io_err_on(bio->bi_status, ca, BCH_MEMBER_ERROR_write,
 			       "btree write error: %s",
 			       bch2_blk_status_to_str(bio->bi_status)) ||
 	    bch2_meta_write_fault("btree")) {
@@ -1952,7 +1990,6 @@ static void btree_write_submit(struct work_struct *work)
 void __bch2_btree_node_write(struct bch_fs *c, struct btree *b, unsigned flags)
 {
 	struct btree_write_bio *wbio;
-	struct bset_tree *t;
 	struct bset *i;
 	struct btree_node *bn = NULL;
 	struct btree_node_entry *bne = NULL;
@@ -2078,11 +2115,11 @@ do_write:
 		      unwritten_whiteouts_end(b));
 	SET_BSET_SEPARATE_WHITEOUTS(i, false);
 
-	b->whiteout_u64s = 0;
-
-	u64s = bch2_sort_keys(i->start, &sort_iter.iter, false);
+	u64s = bch2_sort_keys_keep_unwritten_whiteouts(i->start, &sort_iter.iter);
 	le16_add_cpu(&i->u64s, u64s);
 
+	b->whiteout_u64s = 0;
+
 	BUG_ON(!b->written && i->u64s != b->data->keys.u64s);
 
 	set_needs_whiteout(i, false);
@@ -2123,7 +2160,7 @@ do_write:
 
 	ret = bset_encrypt(c, i, b->written << 9);
 	if (bch2_fs_fatal_err_on(ret, c,
-			"error encrypting btree node: %i\n", ret))
+			"encrypting btree node: %s", bch2_err_str(ret)))
 		goto err;
 
 	nonce = btree_nonce(i, b->written << 9);
@@ -2192,7 +2229,7 @@ do_write:
 	atomic64_add(bytes_to_write, &c->btree_write_stats[type].bytes);
 
 	INIT_WORK(&wbio->work, btree_write_submit);
-	queue_work(c->io_complete_wq, &wbio->work);
+	queue_work(c->btree_write_submit_wq, &wbio->work);
 	return;
 err:
 	set_btree_node_noevict(b);
@@ -2209,7 +2246,6 @@ bool bch2_btree_post_write_cleanup(struct bch_fs *c, struct btree *b)
 {
 	bool invalidated_iter = false;
 	struct btree_node_entry *bne;
-	struct bset_tree *t;
 
 	if (!btree_node_just_written(b))
 		return false;
@@ -2232,7 +2268,7 @@ bool bch2_btree_post_write_cleanup(struct bch_fs *c, struct btree *b)
 	 * single bset:
 	 */
 	if (b->nsets > 1) {
-		btree_node_sort(c, b, 0, b->nsets, true);
+		btree_node_sort(c, b, 0, b->nsets);
 		invalidated_iter = true;
 	} else {
 		invalidated_iter = bch2_drop_whiteouts(b, COMPACT_ALL);
@@ -2329,20 +2365,13 @@ void bch2_btree_write_stats_to_text(struct printbuf *out, struct bch_fs *c)
 	printbuf_tabstop_push(out, 20);
 	printbuf_tabstop_push(out, 10);
 
-	prt_tab(out);
-	prt_str(out, "nr");
-	prt_tab(out);
-	prt_str(out, "size");
-	prt_newline(out);
+	prt_printf(out, "\tnr\tsize\n");
 
 	for (unsigned i = 0; i < BTREE_WRITE_TYPE_NR; i++) {
 		u64 nr		= atomic64_read(&c->btree_write_stats[i].nr);
 		u64 bytes	= atomic64_read(&c->btree_write_stats[i].bytes);
 
-		prt_printf(out, "%s:", bch2_btree_write_types[i]);
-		prt_tab(out);
-		prt_u64(out, nr);
-		prt_tab(out);
+		prt_printf(out, "%s:\t%llu\t", bch2_btree_write_types[i], nr);
 		prt_human_readable_u64(out, nr ? div64_u64(bytes, nr) : 0);
 		prt_newline(out);
 	}
diff --git a/fs/bcachefs/btree_io.h b/fs/bcachefs/btree_io.h
index e251cb6b96..2b8b564fc5 100644
--- a/fs/bcachefs/btree_io.h
+++ b/fs/bcachefs/btree_io.h
@@ -81,8 +81,6 @@ static inline bool should_compact_bset_lazy(struct btree *b,
 
 static inline bool bch2_maybe_compact_whiteouts(struct bch_fs *c, struct btree *b)
 {
-	struct bset_tree *t;
-
 	for_each_bset(b, t)
 		if (should_compact_bset_lazy(b, t))
 			return bch2_compact_whiteouts(c, b, COMPACT_LAZY);
diff --git a/fs/bcachefs/btree_iter.c b/fs/bcachefs/btree_iter.c
index 3ef338df82..19352a08ea 100644
--- a/fs/bcachefs/btree_iter.c
+++ b/fs/bcachefs/btree_iter.c
@@ -61,7 +61,7 @@ static inline int btree_path_cmp(const struct btree_path *l,
 static inline struct bpos bkey_successor(struct btree_iter *iter, struct bpos p)
 {
 	/* Are we iterating over keys in all snapshots? */
-	if (iter->flags & BTREE_ITER_ALL_SNAPSHOTS) {
+	if (iter->flags & BTREE_ITER_all_snapshots) {
 		p = bpos_successor(p);
 	} else {
 		p = bpos_nosnap_successor(p);
@@ -74,7 +74,7 @@ static inline struct bpos bkey_successor(struct btree_iter *iter, struct bpos p)
 static inline struct bpos bkey_predecessor(struct btree_iter *iter, struct bpos p)
 {
 	/* Are we iterating over keys in all snapshots? */
-	if (iter->flags & BTREE_ITER_ALL_SNAPSHOTS) {
+	if (iter->flags & BTREE_ITER_all_snapshots) {
 		p = bpos_predecessor(p);
 	} else {
 		p = bpos_nosnap_predecessor(p);
@@ -88,7 +88,7 @@ static inline struct bpos btree_iter_search_key(struct btree_iter *iter)
 {
 	struct bpos pos = iter->pos;
 
-	if ((iter->flags & BTREE_ITER_IS_EXTENTS) &&
+	if ((iter->flags & BTREE_ITER_is_extents) &&
 	    !bkey_eq(pos, POS_MAX))
 		pos = bkey_successor(iter, pos);
 	return pos;
@@ -221,11 +221,8 @@ static void bch2_btree_path_verify(struct btree_trans *trans,
 				   struct btree_path *path)
 {
 	struct bch_fs *c = trans->c;
-	unsigned i;
-
-	EBUG_ON(path->btree_id >= BTREE_ID_NR);
 
-	for (i = 0; i < (!path->cached ? BTREE_MAX_DEPTH : 1); i++) {
+	for (unsigned i = 0; i < (!path->cached ? BTREE_MAX_DEPTH : 1); i++) {
 		if (!path->l[i].b) {
 			BUG_ON(!path->cached &&
 			       bch2_btree_id_root(c, path->btree_id)->b->c.level > i);
@@ -251,15 +248,13 @@ static void bch2_btree_iter_verify(struct btree_iter *iter)
 {
 	struct btree_trans *trans = iter->trans;
 
-	BUG_ON(iter->btree_id >= BTREE_ID_NR);
+	BUG_ON(!!(iter->flags & BTREE_ITER_cached) != btree_iter_path(trans, iter)->cached);
 
-	BUG_ON(!!(iter->flags & BTREE_ITER_CACHED) != btree_iter_path(trans, iter)->cached);
+	BUG_ON((iter->flags & BTREE_ITER_is_extents) &&
+	       (iter->flags & BTREE_ITER_all_snapshots));
 
-	BUG_ON((iter->flags & BTREE_ITER_IS_EXTENTS) &&
-	       (iter->flags & BTREE_ITER_ALL_SNAPSHOTS));
-
-	BUG_ON(!(iter->flags & __BTREE_ITER_ALL_SNAPSHOTS) &&
-	       (iter->flags & BTREE_ITER_ALL_SNAPSHOTS) &&
+	BUG_ON(!(iter->flags & BTREE_ITER_snapshot_field) &&
+	       (iter->flags & BTREE_ITER_all_snapshots) &&
 	       !btree_type_has_snapshot_field(iter->btree_id));
 
 	if (iter->update_path)
@@ -269,10 +264,10 @@ static void bch2_btree_iter_verify(struct btree_iter *iter)
 
 static void bch2_btree_iter_verify_entry_exit(struct btree_iter *iter)
 {
-	BUG_ON((iter->flags & BTREE_ITER_FILTER_SNAPSHOTS) &&
+	BUG_ON((iter->flags & BTREE_ITER_filter_snapshots) &&
 	       !iter->pos.snapshot);
 
-	BUG_ON(!(iter->flags & BTREE_ITER_ALL_SNAPSHOTS) &&
+	BUG_ON(!(iter->flags & BTREE_ITER_all_snapshots) &&
 	       iter->pos.snapshot != iter->snapshot);
 
 	BUG_ON(bkey_lt(iter->pos, bkey_start_pos(&iter->k)) ||
@@ -289,7 +284,7 @@ static int bch2_btree_iter_verify_ret(struct btree_iter *iter, struct bkey_s_c k
 	if (!bch2_debug_check_iterators)
 		return 0;
 
-	if (!(iter->flags & BTREE_ITER_FILTER_SNAPSHOTS))
+	if (!(iter->flags & BTREE_ITER_filter_snapshots))
 		return 0;
 
 	if (bkey_err(k) || !k.k)
@@ -300,8 +295,8 @@ static int bch2_btree_iter_verify_ret(struct btree_iter *iter, struct bkey_s_c k
 					  k.k->p.snapshot));
 
 	bch2_trans_iter_init(trans, &copy, iter->btree_id, iter->pos,
-			     BTREE_ITER_NOPRESERVE|
-			     BTREE_ITER_ALL_SNAPSHOTS);
+			     BTREE_ITER_nopreserve|
+			     BTREE_ITER_all_snapshots);
 	prev = bch2_btree_iter_prev(&copy);
 	if (!prev.k)
 		goto out;
@@ -332,6 +327,8 @@ out:
 void bch2_assert_pos_locked(struct btree_trans *trans, enum btree_id id,
 			    struct bpos pos, bool key_cache)
 {
+	bch2_trans_verify_not_unlocked(trans);
+
 	struct btree_path *path;
 	struct trans_for_each_path_inorder_iter iter;
 	struct printbuf buf = PRINTBUF;
@@ -891,13 +888,13 @@ static noinline int btree_node_iter_and_journal_peek(struct btree_trans *trans,
 	struct bkey_s_c k;
 	int ret = 0;
 
-	__bch2_btree_and_journal_iter_init_node_iter(&jiter, c, l->b, l->iter, path->pos);
+	__bch2_btree_and_journal_iter_init_node_iter(trans, &jiter, l->b, l->iter, path->pos);
 
 	k = bch2_btree_and_journal_iter_peek(&jiter);
 
 	bch2_bkey_buf_reassemble(out, c, k);
 
-	if ((flags & BTREE_ITER_PREFETCH) &&
+	if ((flags & BTREE_ITER_prefetch) &&
 	    c->opts.btree_node_prefetch)
 		ret = btree_path_prefetch_j(trans, path, &jiter);
 
@@ -927,10 +924,24 @@ static __always_inline int btree_path_down(struct btree_trans *trans,
 		if (ret)
 			goto err;
 	} else {
-		bch2_bkey_buf_unpack(&tmp, c, l->b,
-				 bch2_btree_node_iter_peek(&l->iter, l->b));
+		struct bkey_packed *k = bch2_btree_node_iter_peek(&l->iter, l->b);
+		if (!k) {
+			struct printbuf buf = PRINTBUF;
+
+			prt_str(&buf, "node not found at pos ");
+			bch2_bpos_to_text(&buf, path->pos);
+			prt_str(&buf, " within parent node ");
+			bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&l->b->key));
+
+			bch2_fs_fatal_error(c, "%s", buf.buf);
+			printbuf_exit(&buf);
+			ret = -BCH_ERR_btree_need_topology_repair;
+			goto err;
+		}
+
+		bch2_bkey_buf_unpack(&tmp, c, l->b, k);
 
-		if ((flags & BTREE_ITER_PREFETCH) &&
+		if ((flags & BTREE_ITER_prefetch) &&
 		    c->opts.btree_node_prefetch) {
 			ret = btree_path_prefetch(trans, path);
 			if (ret)
@@ -962,7 +973,6 @@ err:
 	return ret;
 }
 
-
 static int bch2_btree_path_traverse_all(struct btree_trans *trans)
 {
 	struct bch_fs *c = trans->c;
@@ -986,6 +996,7 @@ retry_all:
 
 	bch2_trans_unlock(trans);
 	cond_resched();
+	trans_set_locked(trans);
 
 	if (unlikely(trans->memory_allocation_failure)) {
 		struct closure cl;
@@ -1146,9 +1157,10 @@ int bch2_btree_path_traverse_one(struct btree_trans *trans,
 	path = &trans->paths[path_idx];
 
 	if (unlikely(path->level >= BTREE_MAX_DEPTH))
-		goto out;
+		goto out_uptodate;
 
 	path->level = btree_path_up_until_good_node(trans, path, 0);
+	unsigned max_level = path->level;
 
 	EBUG_ON(btree_path_node(path, path->level) &&
 		!btree_node_locked(path, path->level));
@@ -1180,6 +1192,16 @@ int bch2_btree_path_traverse_one(struct btree_trans *trans,
 		}
 	}
 
+	if (unlikely(max_level > path->level)) {
+		struct btree_path *linked;
+		unsigned iter;
+
+		trans_for_each_path_with_node(trans, path_l(path)->b, linked, iter)
+			for (unsigned j = path->level + 1; j < max_level; j++)
+				linked->l[j] = path->l[j];
+	}
+
+out_uptodate:
 	path->uptodate = BTREE_ITER_UPTODATE;
 out:
 	if (bch2_err_matches(ret, BCH_ERR_transaction_restart) != !!trans->restarted)
@@ -1208,11 +1230,14 @@ static inline void btree_path_copy(struct btree_trans *trans, struct btree_path
 }
 
 static btree_path_idx_t btree_path_clone(struct btree_trans *trans, btree_path_idx_t src,
-					 bool intent)
+					 bool intent, unsigned long ip)
 {
 	btree_path_idx_t new = btree_path_alloc(trans, src);
 	btree_path_copy(trans, trans->paths + new, trans->paths + src);
 	__btree_path_get(trans->paths + new, intent);
+#ifdef TRACK_PATH_ALLOCATED
+	trans->paths[new].ip_allocated = ip;
+#endif
 	return new;
 }
 
@@ -1221,7 +1246,7 @@ btree_path_idx_t __bch2_btree_path_make_mut(struct btree_trans *trans,
 			btree_path_idx_t path, bool intent, unsigned long ip)
 {
 	__btree_path_put(trans->paths + path, intent);
-	path = btree_path_clone(trans, path, intent);
+	path = btree_path_clone(trans, path, intent, ip);
 	trans->paths[path].preserve = false;
 	return path;
 }
@@ -1321,6 +1346,26 @@ static inline void __bch2_path_free(struct btree_trans *trans, btree_path_idx_t
 	__clear_bit(path, trans->paths_allocated);
 }
 
+static bool bch2_btree_path_can_relock(struct btree_trans *trans, struct btree_path *path)
+{
+	unsigned l = path->level;
+
+	do {
+		if (!btree_path_node(path, l))
+			break;
+
+		if (!is_btree_node(path, l))
+			return false;
+
+		if (path->l[l].lock_seq != path->l[l].b->c.lock.seq)
+			return false;
+
+		l++;
+	} while (l < path->locks_want);
+
+	return true;
+}
+
 void bch2_path_put(struct btree_trans *trans, btree_path_idx_t path_idx, bool intent)
 {
 	struct btree_path *path = trans->paths + path_idx, *dup;
@@ -1335,10 +1380,15 @@ void bch2_path_put(struct btree_trans *trans, btree_path_idx_t path_idx, bool in
 	if (!dup && !(!path->preserve && !is_btree_node(path, path->level)))
 		return;
 
-	if (path->should_be_locked &&
-	    !trans->restarted &&
-	    (!dup || !bch2_btree_path_relock_norestart(trans, dup)))
-		return;
+	if (path->should_be_locked && !trans->restarted) {
+		if (!dup)
+			return;
+
+		if (!(trans->locked
+		      ? bch2_btree_path_relock_norestart(trans, dup)
+		      : bch2_btree_path_can_relock(trans, dup)))
+			return;
+	}
 
 	if (dup) {
 		dup->preserve		|= path->preserve;
@@ -1371,22 +1421,26 @@ void __noreturn bch2_trans_in_restart_error(struct btree_trans *trans)
 	      (void *) trans->last_restarted_ip);
 }
 
+void __noreturn bch2_trans_unlocked_error(struct btree_trans *trans)
+{
+	panic("trans should be locked, unlocked by %pS\n",
+	      (void *) trans->last_unlock_ip);
+}
+
 noinline __cold
 void bch2_trans_updates_to_text(struct printbuf *buf, struct btree_trans *trans)
 {
-	prt_printf(buf, "transaction updates for %s journal seq %llu",
+	prt_printf(buf, "transaction updates for %s journal seq %llu\n",
 	       trans->fn, trans->journal_res.seq);
-	prt_newline(buf);
 	printbuf_indent_add(buf, 2);
 
 	trans_for_each_update(trans, i) {
 		struct bkey_s_c old = { &i->old_k, i->old_v };
 
-		prt_printf(buf, "update: btree=%s cached=%u %pS",
+		prt_printf(buf, "update: btree=%s cached=%u %pS\n",
 		       bch2_btree_id_str(i->btree_id),
 		       i->cached,
 		       (void *) i->ip_allocated);
-		prt_newline(buf);
 
 		prt_printf(buf, "  old ");
 		bch2_bkey_val_to_text(buf, trans->c, old);
@@ -1415,23 +1469,63 @@ void bch2_dump_trans_updates(struct btree_trans *trans)
 	printbuf_exit(&buf);
 }
 
-static void bch2_btree_path_to_text(struct printbuf *out, struct btree_trans *trans, btree_path_idx_t path_idx)
+static void bch2_btree_path_to_text_short(struct printbuf *out, struct btree_trans *trans, btree_path_idx_t path_idx)
 {
 	struct btree_path *path = trans->paths + path_idx;
 
-	prt_printf(out, "path: idx %2u ref %u:%u %c %c btree=%s l=%u pos ",
+	prt_printf(out, "path: idx %2u ref %u:%u %c %c %c btree=%s l=%u pos ",
 		   path_idx, path->ref, path->intent_ref,
 		   path->preserve ? 'P' : ' ',
 		   path->should_be_locked ? 'S' : ' ',
+		   path->cached ? 'C' : 'B',
 		   bch2_btree_id_str(path->btree_id),
 		   path->level);
 	bch2_bpos_to_text(out, path->pos);
 
-	prt_printf(out, " locks %u", path->nodes_locked);
 #ifdef TRACK_PATH_ALLOCATED
 	prt_printf(out, " %pS", (void *) path->ip_allocated);
 #endif
+}
+
+static const char *btree_node_locked_str(enum btree_node_locked_type t)
+{
+	switch (t) {
+	case BTREE_NODE_UNLOCKED:
+		return "unlocked";
+	case BTREE_NODE_READ_LOCKED:
+		return "read";
+	case BTREE_NODE_INTENT_LOCKED:
+		return "intent";
+	case BTREE_NODE_WRITE_LOCKED:
+		return "write";
+	default:
+		return NULL;
+	}
+}
+
+void bch2_btree_path_to_text(struct printbuf *out, struct btree_trans *trans, btree_path_idx_t path_idx)
+{
+	bch2_btree_path_to_text_short(out, trans, path_idx);
+
+	struct btree_path *path = trans->paths + path_idx;
+
+	prt_printf(out, " uptodate %u locks_want %u", path->uptodate, path->locks_want);
 	prt_newline(out);
+
+	printbuf_indent_add(out, 2);
+	for (unsigned l = 0; l < BTREE_MAX_DEPTH; l++) {
+		prt_printf(out, "l=%u locks %s seq %u node ", l,
+			   btree_node_locked_str(btree_node_locked_type(path, l)),
+			   path->l[l].lock_seq);
+
+		int ret = PTR_ERR_OR_ZERO(path->l[l].b);
+		if (ret)
+			prt_str(out, bch2_err_str(ret));
+		else
+			prt_printf(out, "%px", path->l[l].b);
+		prt_newline(out);
+	}
+	printbuf_indent_sub(out, 2);
 }
 
 static noinline __cold
@@ -1443,8 +1537,10 @@ void __bch2_trans_paths_to_text(struct printbuf *out, struct btree_trans *trans,
 	if (!nosort)
 		btree_trans_sort_paths(trans);
 
-	trans_for_each_path_idx_inorder(trans, iter)
-		bch2_btree_path_to_text(out, trans, iter.path_idx);
+	trans_for_each_path_idx_inorder(trans, iter) {
+		bch2_btree_path_to_text_short(out, trans, iter.path_idx);
+		prt_newline(out);
+	}
 }
 
 noinline __cold
@@ -1520,7 +1616,7 @@ static noinline void btree_paths_realloc(struct btree_trans *trans)
 {
 	unsigned nr = trans->nr_paths * 2;
 
-	void *p = kzalloc(BITS_TO_LONGS(nr) * sizeof(unsigned long) +
+	void *p = kvzalloc(BITS_TO_LONGS(nr) * sizeof(unsigned long) +
 			  sizeof(struct btree_trans_paths) +
 			  nr * sizeof(struct btree_path) +
 			  nr * sizeof(btree_path_idx_t) + 8 +
@@ -1595,11 +1691,12 @@ btree_path_idx_t bch2_path_get(struct btree_trans *trans,
 			     unsigned flags, unsigned long ip)
 {
 	struct btree_path *path;
-	bool cached = flags & BTREE_ITER_CACHED;
-	bool intent = flags & BTREE_ITER_INTENT;
+	bool cached = flags & BTREE_ITER_cached;
+	bool intent = flags & BTREE_ITER_intent;
 	struct trans_for_each_path_inorder_iter iter;
 	btree_path_idx_t path_pos = 0, path_idx;
 
+	bch2_trans_verify_not_unlocked(trans);
 	bch2_trans_verify_not_in_restart(trans);
 	bch2_trans_verify_locks(trans);
 
@@ -1644,7 +1741,7 @@ btree_path_idx_t bch2_path_get(struct btree_trans *trans,
 		trans->paths_sorted		= false;
 	}
 
-	if (!(flags & BTREE_ITER_NOPRESERVE))
+	if (!(flags & BTREE_ITER_nopreserve))
 		path->preserve = true;
 
 	if (path->intent_ref)
@@ -1665,6 +1762,22 @@ btree_path_idx_t bch2_path_get(struct btree_trans *trans,
 	return path_idx;
 }
 
+btree_path_idx_t bch2_path_get_unlocked_mut(struct btree_trans *trans,
+					    enum btree_id btree_id,
+					    unsigned level,
+					    struct bpos pos)
+{
+	btree_path_idx_t path_idx = bch2_path_get(trans, btree_id, pos, level + 1, level,
+			     BTREE_ITER_nopreserve|
+			     BTREE_ITER_intent, _RET_IP_);
+	path_idx = bch2_btree_path_make_mut(trans, path_idx, true, _RET_IP_);
+
+	struct btree_path *path = trans->paths + path_idx;
+	bch2_btree_path_downgrade(trans, path);
+	__bch2_btree_path_unlock(trans, path);
+	return path_idx;
+}
+
 struct bkey_s_c bch2_btree_path_peek_slot(struct btree_path *path, struct bkey *u)
 {
 
@@ -1706,6 +1819,19 @@ hole:
 	return (struct bkey_s_c) { u, NULL };
 }
 
+
+void bch2_set_btree_iter_dontneed(struct btree_iter *iter)
+{
+	struct btree_trans *trans = iter->trans;
+
+	if (!iter->path || trans->restarted)
+		return;
+
+	struct btree_path *path = btree_iter_path(trans, iter);
+	path->preserve		= false;
+	if (path->ref == 1)
+		path->should_be_locked	= false;
+}
 /* Btree iterators: */
 
 int __must_check
@@ -1720,16 +1846,20 @@ bch2_btree_iter_traverse(struct btree_iter *iter)
 	struct btree_trans *trans = iter->trans;
 	int ret;
 
+	bch2_trans_verify_not_unlocked(trans);
+
 	iter->path = bch2_btree_path_set_pos(trans, iter->path,
 					btree_iter_search_key(iter),
-					iter->flags & BTREE_ITER_INTENT,
+					iter->flags & BTREE_ITER_intent,
 					btree_iter_ip_allocated(iter));
 
 	ret = bch2_btree_path_traverse(iter->trans, iter->path, iter->flags);
 	if (ret)
 		return ret;
 
-	btree_path_set_should_be_locked(trans->paths + iter->path);
+	struct btree_path *path = btree_iter_path(trans, iter);
+	if (btree_path_node(path, path->level))
+		btree_path_set_should_be_locked(path);
 	return 0;
 }
 
@@ -1759,7 +1889,7 @@ struct btree *bch2_btree_iter_peek_node(struct btree_iter *iter)
 	iter->k.p = iter->pos = b->key.k.p;
 
 	iter->path = bch2_btree_path_set_pos(trans, iter->path, b->key.k.p,
-					iter->flags & BTREE_ITER_INTENT,
+					iter->flags & BTREE_ITER_intent,
 					btree_iter_ip_allocated(iter));
 	btree_path_set_should_be_locked(btree_iter_path(trans, iter));
 out:
@@ -1820,13 +1950,16 @@ struct btree *bch2_btree_iter_next_node(struct btree_iter *iter)
 	if (bpos_eq(iter->pos, b->key.k.p)) {
 		__btree_path_set_level_up(trans, path, path->level++);
 	} else {
+		if (btree_lock_want(path, path->level + 1) == BTREE_NODE_UNLOCKED)
+			btree_node_unlock(trans, path, path->level + 1);
+
 		/*
 		 * Haven't gotten to the end of the parent node: go back down to
 		 * the next child node
 		 */
 		iter->path = bch2_btree_path_set_pos(trans, iter->path,
 					bpos_successor(iter->pos),
-					iter->flags & BTREE_ITER_INTENT,
+					iter->flags & BTREE_ITER_intent,
 					btree_iter_ip_allocated(iter));
 
 		path = btree_iter_path(trans, iter);
@@ -1844,7 +1977,7 @@ struct btree *bch2_btree_iter_next_node(struct btree_iter *iter)
 	iter->k.p = iter->pos = b->key.k.p;
 
 	iter->path = bch2_btree_path_set_pos(trans, iter->path, b->key.k.p,
-					iter->flags & BTREE_ITER_INTENT,
+					iter->flags & BTREE_ITER_intent,
 					btree_iter_ip_allocated(iter));
 	btree_path_set_should_be_locked(btree_iter_path(trans, iter));
 	EBUG_ON(btree_iter_path(trans, iter)->uptodate);
@@ -1863,11 +1996,11 @@ err:
 inline bool bch2_btree_iter_advance(struct btree_iter *iter)
 {
 	struct bpos pos = iter->k.p;
-	bool ret = !(iter->flags & BTREE_ITER_ALL_SNAPSHOTS
+	bool ret = !(iter->flags & BTREE_ITER_all_snapshots
 		     ? bpos_eq(pos, SPOS_MAX)
 		     : bkey_eq(pos, SPOS_MAX));
 
-	if (ret && !(iter->flags & BTREE_ITER_IS_EXTENTS))
+	if (ret && !(iter->flags & BTREE_ITER_is_extents))
 		pos = bkey_successor(iter, pos);
 	bch2_btree_iter_set_pos(iter, pos);
 	return ret;
@@ -1876,11 +2009,11 @@ inline bool bch2_btree_iter_advance(struct btree_iter *iter)
 inline bool bch2_btree_iter_rewind(struct btree_iter *iter)
 {
 	struct bpos pos = bkey_start_pos(&iter->k);
-	bool ret = !(iter->flags & BTREE_ITER_ALL_SNAPSHOTS
+	bool ret = !(iter->flags & BTREE_ITER_all_snapshots
 		     ? bpos_eq(pos, POS_MIN)
 		     : bkey_eq(pos, POS_MIN));
 
-	if (ret && !(iter->flags & BTREE_ITER_IS_EXTENTS))
+	if (ret && !(iter->flags & BTREE_ITER_is_extents))
 		pos = bkey_predecessor(iter, pos);
 	bch2_btree_iter_set_pos(iter, pos);
 	return ret;
@@ -1991,7 +2124,10 @@ struct bkey_s_c btree_trans_peek_key_cache(struct btree_iter *iter, struct bpos
 	struct bkey_s_c k;
 	int ret;
 
-	if ((iter->flags & BTREE_ITER_KEY_CACHE_FILL) &&
+	bch2_trans_verify_not_in_restart(trans);
+	bch2_trans_verify_not_unlocked(trans);
+
+	if ((iter->flags & BTREE_ITER_key_cache_fill) &&
 	    bpos_eq(iter->pos, pos))
 		return bkey_s_c_null;
 
@@ -2000,17 +2136,17 @@ struct bkey_s_c btree_trans_peek_key_cache(struct btree_iter *iter, struct bpos
 
 	if (!iter->key_cache_path)
 		iter->key_cache_path = bch2_path_get(trans, iter->btree_id, pos,
-						     iter->flags & BTREE_ITER_INTENT, 0,
-						     iter->flags|BTREE_ITER_CACHED|
-						     BTREE_ITER_CACHED_NOFILL,
+						     iter->flags & BTREE_ITER_intent, 0,
+						     iter->flags|BTREE_ITER_cached|
+						     BTREE_ITER_cached_nofill,
 						     _THIS_IP_);
 
 	iter->key_cache_path = bch2_btree_path_set_pos(trans, iter->key_cache_path, pos,
-					iter->flags & BTREE_ITER_INTENT,
+					iter->flags & BTREE_ITER_intent,
 					btree_iter_ip_allocated(iter));
 
 	ret =   bch2_btree_path_traverse(trans, iter->key_cache_path,
-					 iter->flags|BTREE_ITER_CACHED) ?:
+					 iter->flags|BTREE_ITER_cached) ?:
 		bch2_btree_path_relock(trans, btree_iter_path(trans, iter), _THIS_IP_);
 	if (unlikely(ret))
 		return bkey_s_c_err(ret);
@@ -2038,7 +2174,7 @@ static struct bkey_s_c __bch2_btree_iter_peek(struct btree_iter *iter, struct bp
 		struct btree_path_level *l;
 
 		iter->path = bch2_btree_path_set_pos(trans, iter->path, search_key,
-					iter->flags & BTREE_ITER_INTENT,
+					iter->flags & BTREE_ITER_intent,
 					btree_iter_ip_allocated(iter));
 
 		ret = bch2_btree_path_traverse(trans, iter->path, iter->flags);
@@ -2063,7 +2199,7 @@ static struct bkey_s_c __bch2_btree_iter_peek(struct btree_iter *iter, struct bp
 
 		k = btree_path_level_peek_all(trans->c, l, &iter->k);
 
-		if (unlikely(iter->flags & BTREE_ITER_WITH_KEY_CACHE) &&
+		if (unlikely(iter->flags & BTREE_ITER_with_key_cache) &&
 		    k.k &&
 		    (k2 = btree_trans_peek_key_cache(iter, k.k->p)).k) {
 			k = k2;
@@ -2074,10 +2210,10 @@ static struct bkey_s_c __bch2_btree_iter_peek(struct btree_iter *iter, struct bp
 			}
 		}
 
-		if (unlikely(iter->flags & BTREE_ITER_WITH_JOURNAL))
+		if (unlikely(iter->flags & BTREE_ITER_with_journal))
 			k = btree_trans_peek_journal(trans, iter, k);
 
-		if (unlikely((iter->flags & BTREE_ITER_WITH_UPDATES) &&
+		if (unlikely((iter->flags & BTREE_ITER_with_updates) &&
 			     trans->nr_updates))
 			bch2_btree_trans_peek_updates(trans, iter, &k);
 
@@ -2129,11 +2265,12 @@ struct bkey_s_c bch2_btree_iter_peek_upto(struct btree_iter *iter, struct bpos e
 	struct bpos iter_pos;
 	int ret;
 
-	EBUG_ON((iter->flags & BTREE_ITER_FILTER_SNAPSHOTS) && bkey_eq(end, POS_MAX));
+	bch2_trans_verify_not_unlocked(trans);
+	EBUG_ON((iter->flags & BTREE_ITER_filter_snapshots) && bkey_eq(end, POS_MAX));
 
 	if (iter->update_path) {
 		bch2_path_put_nokeep(trans, iter->update_path,
-				     iter->flags & BTREE_ITER_INTENT);
+				     iter->flags & BTREE_ITER_intent);
 		iter->update_path = 0;
 	}
 
@@ -2156,7 +2293,7 @@ struct bkey_s_c bch2_btree_iter_peek_upto(struct btree_iter *iter, struct bpos e
 		 * isn't monotonically increasing before FILTER_SNAPSHOTS, and
 		 * that's what we check against in extents mode:
 		 */
-		if (unlikely(!(iter->flags & BTREE_ITER_IS_EXTENTS)
+		if (unlikely(!(iter->flags & BTREE_ITER_is_extents)
 			     ? bkey_gt(k.k->p, end)
 			     : k.k->p.inode > end.inode))
 			goto end;
@@ -2164,13 +2301,13 @@ struct bkey_s_c bch2_btree_iter_peek_upto(struct btree_iter *iter, struct bpos e
 		if (iter->update_path &&
 		    !bkey_eq(trans->paths[iter->update_path].pos, k.k->p)) {
 			bch2_path_put_nokeep(trans, iter->update_path,
-					     iter->flags & BTREE_ITER_INTENT);
+					     iter->flags & BTREE_ITER_intent);
 			iter->update_path = 0;
 		}
 
-		if ((iter->flags & BTREE_ITER_FILTER_SNAPSHOTS) &&
-		    (iter->flags & BTREE_ITER_INTENT) &&
-		    !(iter->flags & BTREE_ITER_IS_EXTENTS) &&
+		if ((iter->flags & BTREE_ITER_filter_snapshots) &&
+		    (iter->flags & BTREE_ITER_intent) &&
+		    !(iter->flags & BTREE_ITER_is_extents) &&
 		    !iter->update_path) {
 			struct bpos pos = k.k->p;
 
@@ -2185,12 +2322,12 @@ struct bkey_s_c bch2_btree_iter_peek_upto(struct btree_iter *iter, struct bpos e
 			 * advance, same as on exit for iter->path, but only up
 			 * to snapshot
 			 */
-			__btree_path_get(trans->paths + iter->path, iter->flags & BTREE_ITER_INTENT);
+			__btree_path_get(trans->paths + iter->path, iter->flags & BTREE_ITER_intent);
 			iter->update_path = iter->path;
 
 			iter->update_path = bch2_btree_path_set_pos(trans,
 						iter->update_path, pos,
-						iter->flags & BTREE_ITER_INTENT,
+						iter->flags & BTREE_ITER_intent,
 						_THIS_IP_);
 			ret = bch2_btree_path_traverse(trans, iter->update_path, iter->flags);
 			if (unlikely(ret)) {
@@ -2203,7 +2340,7 @@ struct bkey_s_c bch2_btree_iter_peek_upto(struct btree_iter *iter, struct bpos e
 		 * We can never have a key in a leaf node at POS_MAX, so
 		 * we don't have to check these successor() calls:
 		 */
-		if ((iter->flags & BTREE_ITER_FILTER_SNAPSHOTS) &&
+		if ((iter->flags & BTREE_ITER_filter_snapshots) &&
 		    !bch2_snapshot_is_ancestor(trans->c,
 					       iter->snapshot,
 					       k.k->p.snapshot)) {
@@ -2212,7 +2349,7 @@ struct bkey_s_c bch2_btree_iter_peek_upto(struct btree_iter *iter, struct bpos e
 		}
 
 		if (bkey_whiteout(k.k) &&
-		    !(iter->flags & BTREE_ITER_ALL_SNAPSHOTS)) {
+		    !(iter->flags & BTREE_ITER_all_snapshots)) {
 			search_key = bkey_successor(iter, k.k->p);
 			continue;
 		}
@@ -2222,12 +2359,12 @@ struct bkey_s_c bch2_btree_iter_peek_upto(struct btree_iter *iter, struct bpos e
 		 * equal to the key we just returned - except extents can
 		 * straddle iter->pos:
 		 */
-		if (!(iter->flags & BTREE_ITER_IS_EXTENTS))
+		if (!(iter->flags & BTREE_ITER_is_extents))
 			iter_pos = k.k->p;
 		else
 			iter_pos = bkey_max(iter->pos, bkey_start_pos(k.k));
 
-		if (unlikely(!(iter->flags & BTREE_ITER_IS_EXTENTS)
+		if (unlikely(!(iter->flags & BTREE_ITER_is_extents)
 			     ? bkey_gt(iter_pos, end)
 			     : bkey_ge(iter_pos, end)))
 			goto end;
@@ -2238,7 +2375,7 @@ struct bkey_s_c bch2_btree_iter_peek_upto(struct btree_iter *iter, struct bpos e
 	iter->pos = iter_pos;
 
 	iter->path = bch2_btree_path_set_pos(trans, iter->path, k.k->p,
-				iter->flags & BTREE_ITER_INTENT,
+				iter->flags & BTREE_ITER_intent,
 				btree_iter_ip_allocated(iter));
 
 	btree_path_set_should_be_locked(btree_iter_path(trans, iter));
@@ -2251,7 +2388,7 @@ out_no_locked:
 			btree_path_set_should_be_locked(trans->paths + iter->update_path);
 	}
 
-	if (!(iter->flags & BTREE_ITER_ALL_SNAPSHOTS))
+	if (!(iter->flags & BTREE_ITER_all_snapshots))
 		iter->pos.snapshot = iter->snapshot;
 
 	ret = bch2_btree_iter_verify_ret(iter, k);
@@ -2301,21 +2438,22 @@ struct bkey_s_c bch2_btree_iter_peek_prev(struct btree_iter *iter)
 	btree_path_idx_t saved_path = 0;
 	int ret;
 
+	bch2_trans_verify_not_unlocked(trans);
 	EBUG_ON(btree_iter_path(trans, iter)->cached ||
 		btree_iter_path(trans, iter)->level);
 
-	if (iter->flags & BTREE_ITER_WITH_JOURNAL)
-		return bkey_s_c_err(-EIO);
+	if (iter->flags & BTREE_ITER_with_journal)
+		return bkey_s_c_err(-BCH_ERR_btree_iter_with_journal_not_supported);
 
 	bch2_btree_iter_verify(iter);
 	bch2_btree_iter_verify_entry_exit(iter);
 
-	if (iter->flags & BTREE_ITER_FILTER_SNAPSHOTS)
+	if (iter->flags & BTREE_ITER_filter_snapshots)
 		search_key.snapshot = U32_MAX;
 
 	while (1) {
 		iter->path = bch2_btree_path_set_pos(trans, iter->path, search_key,
-						iter->flags & BTREE_ITER_INTENT,
+						iter->flags & BTREE_ITER_intent,
 						btree_iter_ip_allocated(iter));
 
 		ret = bch2_btree_path_traverse(trans, iter->path, iter->flags);
@@ -2330,17 +2468,17 @@ struct bkey_s_c bch2_btree_iter_peek_prev(struct btree_iter *iter)
 
 		k = btree_path_level_peek(trans, path, &path->l[0], &iter->k);
 		if (!k.k ||
-		    ((iter->flags & BTREE_ITER_IS_EXTENTS)
+		    ((iter->flags & BTREE_ITER_is_extents)
 		     ? bpos_ge(bkey_start_pos(k.k), search_key)
 		     : bpos_gt(k.k->p, search_key)))
 			k = btree_path_level_prev(trans, path, &path->l[0], &iter->k);
 
-		if (unlikely((iter->flags & BTREE_ITER_WITH_UPDATES) &&
+		if (unlikely((iter->flags & BTREE_ITER_with_updates) &&
 			     trans->nr_updates))
 			bch2_btree_trans_peek_prev_updates(trans, iter, &k);
 
 		if (likely(k.k)) {
-			if (iter->flags & BTREE_ITER_FILTER_SNAPSHOTS) {
+			if (iter->flags & BTREE_ITER_filter_snapshots) {
 				if (k.k->p.snapshot == iter->snapshot)
 					goto got_key;
 
@@ -2351,7 +2489,7 @@ struct bkey_s_c bch2_btree_iter_peek_prev(struct btree_iter *iter)
 				 */
 				if (saved_path && !bkey_eq(k.k->p, saved_k.p)) {
 					bch2_path_put_nokeep(trans, iter->path,
-						      iter->flags & BTREE_ITER_INTENT);
+						      iter->flags & BTREE_ITER_intent);
 					iter->path = saved_path;
 					saved_path = 0;
 					iter->k	= saved_k;
@@ -2364,9 +2502,10 @@ struct bkey_s_c bch2_btree_iter_peek_prev(struct btree_iter *iter)
 							      k.k->p.snapshot)) {
 					if (saved_path)
 						bch2_path_put_nokeep(trans, saved_path,
-						      iter->flags & BTREE_ITER_INTENT);
+						      iter->flags & BTREE_ITER_intent);
 					saved_path = btree_path_clone(trans, iter->path,
-								iter->flags & BTREE_ITER_INTENT);
+								iter->flags & BTREE_ITER_intent,
+								_THIS_IP_);
 					path = btree_iter_path(trans, iter);
 					saved_k = *k.k;
 					saved_v = k.v;
@@ -2377,9 +2516,9 @@ struct bkey_s_c bch2_btree_iter_peek_prev(struct btree_iter *iter)
 			}
 got_key:
 			if (bkey_whiteout(k.k) &&
-			    !(iter->flags & BTREE_ITER_ALL_SNAPSHOTS)) {
+			    !(iter->flags & BTREE_ITER_all_snapshots)) {
 				search_key = bkey_predecessor(iter, k.k->p);
-				if (iter->flags & BTREE_ITER_FILTER_SNAPSHOTS)
+				if (iter->flags & BTREE_ITER_filter_snapshots)
 					search_key.snapshot = U32_MAX;
 				continue;
 			}
@@ -2403,11 +2542,11 @@ got_key:
 	if (bkey_lt(k.k->p, iter->pos))
 		iter->pos = k.k->p;
 
-	if (iter->flags & BTREE_ITER_FILTER_SNAPSHOTS)
+	if (iter->flags & BTREE_ITER_filter_snapshots)
 		iter->pos.snapshot = iter->snapshot;
 out_no_locked:
 	if (saved_path)
-		bch2_path_put_nokeep(trans, saved_path, iter->flags & BTREE_ITER_INTENT);
+		bch2_path_put_nokeep(trans, saved_path, iter->flags & BTREE_ITER_intent);
 
 	bch2_btree_iter_verify_entry_exit(iter);
 	bch2_btree_iter_verify(iter);
@@ -2437,12 +2576,13 @@ struct bkey_s_c bch2_btree_iter_peek_slot(struct btree_iter *iter)
 	struct bkey_s_c k;
 	int ret;
 
+	bch2_trans_verify_not_unlocked(trans);
 	bch2_btree_iter_verify(iter);
 	bch2_btree_iter_verify_entry_exit(iter);
-	EBUG_ON(btree_iter_path(trans, iter)->level && (iter->flags & BTREE_ITER_WITH_KEY_CACHE));
+	EBUG_ON(btree_iter_path(trans, iter)->level && (iter->flags & BTREE_ITER_with_key_cache));
 
 	/* extents can't span inode numbers: */
-	if ((iter->flags & BTREE_ITER_IS_EXTENTS) &&
+	if ((iter->flags & BTREE_ITER_is_extents) &&
 	    unlikely(iter->pos.offset == KEY_OFFSET_MAX)) {
 		if (iter->pos.inode == KEY_INODE_MAX)
 			return bkey_s_c_null;
@@ -2452,7 +2592,7 @@ struct bkey_s_c bch2_btree_iter_peek_slot(struct btree_iter *iter)
 
 	search_key = btree_iter_search_key(iter);
 	iter->path = bch2_btree_path_set_pos(trans, iter->path, search_key,
-					iter->flags & BTREE_ITER_INTENT,
+					iter->flags & BTREE_ITER_intent,
 					btree_iter_ip_allocated(iter));
 
 	ret = bch2_btree_path_traverse(trans, iter->path, iter->flags);
@@ -2461,22 +2601,22 @@ struct bkey_s_c bch2_btree_iter_peek_slot(struct btree_iter *iter)
 		goto out_no_locked;
 	}
 
-	if ((iter->flags & BTREE_ITER_CACHED) ||
-	    !(iter->flags & (BTREE_ITER_IS_EXTENTS|BTREE_ITER_FILTER_SNAPSHOTS))) {
+	if ((iter->flags & BTREE_ITER_cached) ||
+	    !(iter->flags & (BTREE_ITER_is_extents|BTREE_ITER_filter_snapshots))) {
 		k = bkey_s_c_null;
 
-		if (unlikely((iter->flags & BTREE_ITER_WITH_UPDATES) &&
+		if (unlikely((iter->flags & BTREE_ITER_with_updates) &&
 			     trans->nr_updates)) {
 			bch2_btree_trans_peek_slot_updates(trans, iter, &k);
 			if (k.k)
 				goto out;
 		}
 
-		if (unlikely(iter->flags & BTREE_ITER_WITH_JOURNAL) &&
+		if (unlikely(iter->flags & BTREE_ITER_with_journal) &&
 		    (k = btree_trans_peek_slot_journal(trans, iter)).k)
 			goto out;
 
-		if (unlikely(iter->flags & BTREE_ITER_WITH_KEY_CACHE) &&
+		if (unlikely(iter->flags & BTREE_ITER_with_key_cache) &&
 		    (k = btree_trans_peek_key_cache(iter, iter->pos)).k) {
 			if (!bkey_err(k))
 				iter->k = *k.k;
@@ -2491,18 +2631,19 @@ struct bkey_s_c bch2_btree_iter_peek_slot(struct btree_iter *iter)
 		struct bpos next;
 		struct bpos end = iter->pos;
 
-		if (iter->flags & BTREE_ITER_IS_EXTENTS)
+		if (iter->flags & BTREE_ITER_is_extents)
 			end.offset = U64_MAX;
 
 		EBUG_ON(btree_iter_path(trans, iter)->level);
 
-		if (iter->flags & BTREE_ITER_INTENT) {
+		if (iter->flags & BTREE_ITER_intent) {
 			struct btree_iter iter2;
 
 			bch2_trans_copy_iter(&iter2, iter);
 			k = bch2_btree_iter_peek_upto(&iter2, end);
 
 			if (k.k && !bkey_err(k)) {
+				swap(iter->key_cache_path, iter2.key_cache_path);
 				iter->k = iter2.k;
 				k.k = &iter->k;
 			}
@@ -2526,7 +2667,7 @@ struct bkey_s_c bch2_btree_iter_peek_slot(struct btree_iter *iter)
 			bkey_init(&iter->k);
 			iter->k.p = iter->pos;
 
-			if (iter->flags & BTREE_ITER_IS_EXTENTS) {
+			if (iter->flags & BTREE_ITER_is_extents) {
 				bch2_key_resize(&iter->k,
 						min_t(u64, KEY_SIZE_MAX,
 						      (next.inode == iter->pos.inode
@@ -2710,13 +2851,13 @@ void bch2_trans_iter_exit(struct btree_trans *trans, struct btree_iter *iter)
 {
 	if (iter->update_path)
 		bch2_path_put_nokeep(trans, iter->update_path,
-			      iter->flags & BTREE_ITER_INTENT);
+			      iter->flags & BTREE_ITER_intent);
 	if (iter->path)
 		bch2_path_put(trans, iter->path,
-			      iter->flags & BTREE_ITER_INTENT);
+			      iter->flags & BTREE_ITER_intent);
 	if (iter->key_cache_path)
 		bch2_path_put(trans, iter->key_cache_path,
-			      iter->flags & BTREE_ITER_INTENT);
+			      iter->flags & BTREE_ITER_intent);
 	iter->path		= 0;
 	iter->update_path	= 0;
 	iter->key_cache_path	= 0;
@@ -2741,9 +2882,9 @@ void bch2_trans_node_iter_init(struct btree_trans *trans,
 			       unsigned depth,
 			       unsigned flags)
 {
-	flags |= BTREE_ITER_NOT_EXTENTS;
-	flags |= __BTREE_ITER_ALL_SNAPSHOTS;
-	flags |= BTREE_ITER_ALL_SNAPSHOTS;
+	flags |= BTREE_ITER_not_extents;
+	flags |= BTREE_ITER_snapshot_field;
+	flags |= BTREE_ITER_all_snapshots;
 
 	bch2_trans_iter_init_common(trans, iter, btree_id, pos, locks_want, depth,
 			       __bch2_btree_iter_flags(trans, btree_id, flags),
@@ -2762,10 +2903,13 @@ void bch2_trans_copy_iter(struct btree_iter *dst, struct btree_iter *src)
 	struct btree_trans *trans = src->trans;
 
 	*dst = *src;
+#ifdef TRACK_PATH_ALLOCATED
+	dst->ip_allocated = _RET_IP_;
+#endif
 	if (src->path)
-		__btree_path_get(trans->paths + src->path, src->flags & BTREE_ITER_INTENT);
+		__btree_path_get(trans->paths + src->path, src->flags & BTREE_ITER_intent);
 	if (src->update_path)
-		__btree_path_get(trans->paths + src->update_path, src->flags & BTREE_ITER_INTENT);
+		__btree_path_get(trans->paths + src->update_path, src->flags & BTREE_ITER_intent);
 	dst->key_cache_path = 0;
 }
 
@@ -2784,6 +2928,31 @@ void *__bch2_trans_kmalloc(struct btree_trans *trans, size_t size)
 	struct btree_transaction_stats *s = btree_trans_stats(trans);
 	s->max_mem = max(s->max_mem, new_bytes);
 
+	if (trans->used_mempool) {
+		if (trans->mem_bytes >= new_bytes)
+			goto out_change_top;
+
+		/* No more space from mempool item, need malloc new one */
+		new_mem = kmalloc(new_bytes, GFP_NOWAIT|__GFP_NOWARN);
+		if (unlikely(!new_mem)) {
+			bch2_trans_unlock(trans);
+
+			new_mem = kmalloc(new_bytes, GFP_KERNEL);
+			if (!new_mem)
+				return ERR_PTR(-BCH_ERR_ENOMEM_trans_kmalloc);
+
+			ret = bch2_trans_relock(trans);
+			if (ret) {
+				kfree(new_mem);
+				return ERR_PTR(ret);
+			}
+		}
+		memcpy(new_mem, trans->mem, trans->mem_top);
+		trans->used_mempool = false;
+		mempool_free(trans->mem, &c->btree_trans_mem_pool);
+		goto out_new_mem;
+	}
+
 	new_mem = krealloc(trans->mem, new_bytes, GFP_NOWAIT|__GFP_NOWARN);
 	if (unlikely(!new_mem)) {
 		bch2_trans_unlock(trans);
@@ -2792,6 +2961,8 @@ void *__bch2_trans_kmalloc(struct btree_trans *trans, size_t size)
 		if (!new_mem && new_bytes <= BTREE_TRANS_MEM_MAX) {
 			new_mem = mempool_alloc(&c->btree_trans_mem_pool, GFP_KERNEL);
 			new_bytes = BTREE_TRANS_MEM_MAX;
+			memcpy(new_mem, trans->mem, trans->mem_top);
+			trans->used_mempool = true;
 			kfree(trans->mem);
 		}
 
@@ -2805,7 +2976,7 @@ void *__bch2_trans_kmalloc(struct btree_trans *trans, size_t size)
 		if (ret)
 			return ERR_PTR(ret);
 	}
-
+out_new_mem:
 	trans->mem = new_mem;
 	trans->mem_bytes = new_bytes;
 
@@ -2813,7 +2984,7 @@ void *__bch2_trans_kmalloc(struct btree_trans *trans, size_t size)
 		trace_and_count(c, trans_restart_mem_realloced, trans, _RET_IP_, new_bytes);
 		return ERR_PTR(btree_trans_restart(trans, BCH_ERR_transaction_restart_mem_realloced));
 	}
-
+out_change_top:
 	p = trans->mem + trans->mem_top;
 	trans->mem_top += size;
 	memset(p, 0, size);
@@ -2907,7 +3078,8 @@ u32 bch2_trans_begin(struct btree_trans *trans)
 	if (!trans->restarted &&
 	    (need_resched() ||
 	     time_after64(now, trans->last_begin_time + BTREE_TRANS_MAX_LOCK_HOLD_TIME_NS))) {
-		drop_locks_do(trans, (cond_resched(), 0));
+		bch2_trans_unlock(trans);
+		cond_resched();
 		now = local_clock();
 	}
 	trans->last_begin_time = now;
@@ -2917,11 +3089,15 @@ u32 bch2_trans_begin(struct btree_trans *trans)
 		bch2_trans_srcu_unlock(trans);
 
 	trans->last_begin_ip = _RET_IP_;
+
+	trans_set_locked(trans);
+
 	if (trans->restarted) {
 		bch2_btree_path_traverse_all(trans);
 		trans->notrace_relock_fail = false;
 	}
 
+	bch2_trans_verify_not_unlocked(trans);
 	return trans->restart_count;
 }
 
@@ -2955,7 +3131,6 @@ struct btree_trans *__bch2_trans_get(struct bch_fs *c, unsigned fn_idx)
 
 	trans = mempool_alloc(&c->btree_trans_pool, GFP_NOFS);
 	memset(trans, 0, sizeof(*trans));
-	closure_init_stack(&trans->ref);
 
 	seqmutex_lock(&c->btree_trans_lock);
 	if (IS_ENABLED(CONFIG_BCACHEFS_DEBUG)) {
@@ -2974,16 +3149,11 @@ struct btree_trans *__bch2_trans_get(struct bch_fs *c, unsigned fn_idx)
 			 */
 			BUG_ON(pos_task &&
 			       pid == pos_task->pid &&
-			       bch2_trans_locked(pos));
-
-			if (pos_task && pid < pos_task->pid) {
-				list_add_tail(&trans->list, &pos->list);
-				goto list_add_done;
-			}
+			       pos->locked);
 		}
 	}
-	list_add_tail(&trans->list, &c->btree_trans_list);
-list_add_done:
+
+	list_add(&trans->list, &c->btree_trans_list);
 	seqmutex_unlock(&c->btree_trans_lock);
 got_trans:
 	trans->c		= c;
@@ -2991,7 +3161,7 @@ got_trans:
 	trans->fn_idx		= fn_idx;
 	trans->locking_wait.task = current;
 	trans->journal_replay_not_finished =
-		unlikely(!test_bit(JOURNAL_REPLAY_DONE, &c->journal.flags)) &&
+		unlikely(!test_bit(JOURNAL_replay_done, &c->journal.flags)) &&
 		atomic_inc_not_zero(&c->journal_keys.ref);
 	trans->nr_paths		= ARRAY_SIZE(trans->_paths);
 	trans->paths_allocated	= trans->_paths_allocated;
@@ -3023,6 +3193,9 @@ got_trans:
 	trans->srcu_idx		= srcu_read_lock(&c->btree_trans_barrier);
 	trans->srcu_lock_time	= jiffies;
 	trans->srcu_held	= true;
+	trans_set_locked(trans);
+
+	closure_init_stack_release(&trans->ref);
 	return trans;
 }
 
@@ -3059,7 +3232,6 @@ void bch2_trans_put(struct btree_trans *trans)
 	trans_for_each_update(trans, i)
 		__btree_path_put(trans->paths + i->path, true);
 	trans->nr_updates	= 0;
-	trans->locking_wait.task = NULL;
 
 	check_btree_paths_leaked(trans);
 
@@ -3080,14 +3252,21 @@ void bch2_trans_put(struct btree_trans *trans)
 	if (unlikely(trans->journal_replay_not_finished))
 		bch2_journal_keys_put(c);
 
+	/*
+	 * trans->ref protects trans->locking_wait.task, btree_paths array; used
+	 * by cycle detector
+	 */
+	closure_return_sync(&trans->ref);
+	trans->locking_wait.task = NULL;
+
 	unsigned long *paths_allocated = trans->paths_allocated;
 	trans->paths_allocated	= NULL;
 	trans->paths		= NULL;
 
 	if (paths_allocated != trans->_paths_allocated)
-		kfree_rcu_mightsleep(paths_allocated);
+		kvfree_rcu_mightsleep(paths_allocated);
 
-	if (trans->mem_bytes == BTREE_TRANS_MEM_MAX)
+	if (trans->used_mempool)
 		mempool_free(trans->mem, &c->btree_trans_mem_pool);
 	else
 		kfree(trans->mem);
@@ -3097,8 +3276,6 @@ void bch2_trans_put(struct btree_trans *trans)
 		trans = this_cpu_xchg(c->btree_trans_bufs->trans, trans);
 
 	if (trans) {
-		closure_sync(&trans->ref);
-
 		seqmutex_lock(&c->btree_trans_lock);
 		list_del(&trans->list);
 		seqmutex_unlock(&c->btree_trans_lock);
@@ -3120,13 +3297,11 @@ bch2_btree_bkey_cached_common_to_text(struct printbuf *out,
 	pid = owner ? owner->pid : 0;
 	rcu_read_unlock();
 
-	prt_tab(out);
-	prt_printf(out, "%px %c l=%u %s:", b, b->cached ? 'c' : 'b',
+	prt_printf(out, "\t%px %c l=%u %s:", b, b->cached ? 'c' : 'b',
 		   b->level, bch2_btree_id_str(b->btree_id));
 	bch2_bpos_to_text(out, btree_node_pos(b));
 
-	prt_tab(out);
-	prt_printf(out, " locks %u:%u:%u held by pid %u",
+	prt_printf(out, "\t locks %u:%u:%u held by pid %u",
 		   c.n[0], c.n[1], c.n[2], pid);
 }
 
@@ -3183,10 +3358,8 @@ void bch2_btree_trans_to_text(struct printbuf *out, struct btree_trans *trans)
 
 	b = READ_ONCE(trans->locking);
 	if (b) {
-		prt_printf(out, "  blocked for %lluus on",
-			   div_u64(local_clock() - trans->locking_wait.start_time,
-				   1000));
-		prt_newline(out);
+		prt_printf(out, "  blocked for %lluus on\n",
+			   div_u64(local_clock() - trans->locking_wait.start_time, 1000));
 		prt_printf(out, "    %c", lock_types[trans->locking_wait.lock_want]);
 		bch2_btree_bkey_cached_common_to_text(out, b);
 		prt_newline(out);
@@ -3208,8 +3381,6 @@ void bch2_fs_btree_iter_exit(struct bch_fs *c)
 				per_cpu_ptr(c->btree_trans_bufs, cpu)->trans;
 
 			if (trans) {
-				closure_sync(&trans->ref);
-
 				seqmutex_lock(&c->btree_trans_lock);
 				list_del(&trans->list);
 				seqmutex_unlock(&c->btree_trans_lock);
@@ -3229,8 +3400,10 @@ void bch2_fs_btree_iter_exit(struct bch_fs *c)
 		bch2_time_stats_exit(&s->lock_hold_times);
 	}
 
-	if (c->btree_trans_barrier_initialized)
+	if (c->btree_trans_barrier_initialized) {
+		synchronize_srcu_expedited(&c->btree_trans_barrier);
 		cleanup_srcu_struct(&c->btree_trans_barrier);
+	}
 	mempool_exit(&c->btree_trans_mem_pool);
 	mempool_exit(&c->btree_trans_pool);
 }
diff --git a/fs/bcachefs/btree_iter.h b/fs/bcachefs/btree_iter.h
index 24772538e4..798eb1c479 100644
--- a/fs/bcachefs/btree_iter.h
+++ b/fs/bcachefs/btree_iter.h
@@ -216,9 +216,13 @@ int __must_check bch2_btree_path_traverse_one(struct btree_trans *,
 					      btree_path_idx_t,
 					      unsigned, unsigned long);
 
+static inline void bch2_trans_verify_not_unlocked(struct btree_trans *);
+
 static inline int __must_check bch2_btree_path_traverse(struct btree_trans *trans,
 					  btree_path_idx_t path, unsigned flags)
 {
+	bch2_trans_verify_not_unlocked(trans);
+
 	if (trans->paths[path].uptodate < BTREE_ITER_NEED_RELOCK)
 		return 0;
 
@@ -227,6 +231,9 @@ static inline int __must_check bch2_btree_path_traverse(struct btree_trans *tran
 
 btree_path_idx_t bch2_path_get(struct btree_trans *, enum btree_id, struct bpos,
 				 unsigned, unsigned, unsigned, unsigned long);
+btree_path_idx_t bch2_path_get_unlocked_mut(struct btree_trans *, enum btree_id,
+					    unsigned, struct bpos);
+
 struct bkey_s_c bch2_btree_path_peek_slot(struct btree_path *, struct bkey *);
 
 /*
@@ -283,7 +290,6 @@ int bch2_trans_relock(struct btree_trans *);
 int bch2_trans_relock_notrace(struct btree_trans *);
 void bch2_trans_unlock(struct btree_trans *);
 void bch2_trans_unlock_long(struct btree_trans *);
-bool bch2_trans_locked(struct btree_trans *);
 
 static inline int trans_was_restarted(struct btree_trans *trans, u32 restart_count)
 {
@@ -309,6 +315,14 @@ static inline void bch2_trans_verify_not_in_restart(struct btree_trans *trans)
 		bch2_trans_in_restart_error(trans);
 }
 
+void __noreturn bch2_trans_unlocked_error(struct btree_trans *);
+
+static inline void bch2_trans_verify_not_unlocked(struct btree_trans *trans)
+{
+	if (!trans->locked)
+		bch2_trans_unlocked_error(trans);
+}
+
 __always_inline
 static int btree_trans_restart_nounlock(struct btree_trans *trans, int err)
 {
@@ -386,10 +400,10 @@ static inline void bch2_btree_iter_set_pos(struct btree_iter *iter, struct bpos
 
 	if (unlikely(iter->update_path))
 		bch2_path_put(trans, iter->update_path,
-			      iter->flags & BTREE_ITER_INTENT);
+			      iter->flags & BTREE_ITER_intent);
 	iter->update_path = 0;
 
-	if (!(iter->flags & BTREE_ITER_ALL_SNAPSHOTS))
+	if (!(iter->flags & BTREE_ITER_all_snapshots))
 		new_pos.snapshot = iter->snapshot;
 
 	__bch2_btree_iter_set_pos(iter, new_pos);
@@ -397,7 +411,7 @@ static inline void bch2_btree_iter_set_pos(struct btree_iter *iter, struct bpos
 
 static inline void bch2_btree_iter_set_pos_to_extent_start(struct btree_iter *iter)
 {
-	BUG_ON(!(iter->flags & BTREE_ITER_IS_EXTENTS));
+	BUG_ON(!(iter->flags & BTREE_ITER_is_extents));
 	iter->pos = bkey_start_pos(&iter->k);
 }
 
@@ -416,20 +430,20 @@ static inline unsigned __bch2_btree_iter_flags(struct btree_trans *trans,
 					       unsigned btree_id,
 					       unsigned flags)
 {
-	if (!(flags & (BTREE_ITER_ALL_SNAPSHOTS|BTREE_ITER_NOT_EXTENTS)) &&
+	if (!(flags & (BTREE_ITER_all_snapshots|BTREE_ITER_not_extents)) &&
 	    btree_id_is_extents(btree_id))
-		flags |= BTREE_ITER_IS_EXTENTS;
+		flags |= BTREE_ITER_is_extents;
 
-	if (!(flags & __BTREE_ITER_ALL_SNAPSHOTS) &&
+	if (!(flags & BTREE_ITER_snapshot_field) &&
 	    !btree_type_has_snapshot_field(btree_id))
-		flags &= ~BTREE_ITER_ALL_SNAPSHOTS;
+		flags &= ~BTREE_ITER_all_snapshots;
 
-	if (!(flags & BTREE_ITER_ALL_SNAPSHOTS) &&
+	if (!(flags & BTREE_ITER_all_snapshots) &&
 	    btree_type_has_snapshots(btree_id))
-		flags |= BTREE_ITER_FILTER_SNAPSHOTS;
+		flags |= BTREE_ITER_filter_snapshots;
 
 	if (trans->journal_replay_not_finished)
-		flags |= BTREE_ITER_WITH_JOURNAL;
+		flags |= BTREE_ITER_with_journal;
 
 	return flags;
 }
@@ -439,10 +453,10 @@ static inline unsigned bch2_btree_iter_flags(struct btree_trans *trans,
 					     unsigned flags)
 {
 	if (!btree_id_cached(trans->c, btree_id)) {
-		flags &= ~BTREE_ITER_CACHED;
-		flags &= ~BTREE_ITER_WITH_KEY_CACHE;
-	} else if (!(flags & BTREE_ITER_CACHED))
-		flags |= BTREE_ITER_WITH_KEY_CACHE;
+		flags &= ~BTREE_ITER_cached;
+		flags &= ~BTREE_ITER_with_key_cache;
+	} else if (!(flags & BTREE_ITER_cached))
+		flags |= BTREE_ITER_with_key_cache;
 
 	return __bch2_btree_iter_flags(trans, btree_id, flags);
 }
@@ -494,13 +508,7 @@ void bch2_trans_node_iter_init(struct btree_trans *, struct btree_iter *,
 			       unsigned, unsigned, unsigned);
 void bch2_trans_copy_iter(struct btree_iter *, struct btree_iter *);
 
-static inline void set_btree_iter_dontneed(struct btree_iter *iter)
-{
-	struct btree_trans *trans = iter->trans;
-
-	if (!trans->restarted)
-		btree_iter_path(trans, iter)->preserve = false;
-}
+void bch2_set_btree_iter_dontneed(struct btree_iter *);
 
 void *__bch2_trans_kmalloc(struct btree_trans *, size_t);
 
@@ -614,14 +622,14 @@ u32 bch2_trans_begin(struct btree_trans *);
 static inline struct bkey_s_c bch2_btree_iter_peek_prev_type(struct btree_iter *iter,
 							     unsigned flags)
 {
-	return  flags & BTREE_ITER_SLOTS      ? bch2_btree_iter_peek_slot(iter) :
+	return  flags & BTREE_ITER_slots      ? bch2_btree_iter_peek_slot(iter) :
 						bch2_btree_iter_peek_prev(iter);
 }
 
 static inline struct bkey_s_c bch2_btree_iter_peek_type(struct btree_iter *iter,
 							unsigned flags)
 {
-	return  flags & BTREE_ITER_SLOTS      ? bch2_btree_iter_peek_slot(iter) :
+	return  flags & BTREE_ITER_slots      ? bch2_btree_iter_peek_slot(iter) :
 						bch2_btree_iter_peek(iter);
 }
 
@@ -629,7 +637,7 @@ static inline struct bkey_s_c bch2_btree_iter_peek_upto_type(struct btree_iter *
 							     struct bpos end,
 							     unsigned flags)
 {
-	if (!(flags & BTREE_ITER_SLOTS))
+	if (!(flags & BTREE_ITER_slots))
 		return bch2_btree_iter_peek_upto(iter, end);
 
 	if (bkey_gt(iter->pos, end))
@@ -642,7 +650,7 @@ int __bch2_btree_trans_too_many_iters(struct btree_trans *);
 
 static inline int btree_trans_too_many_iters(struct btree_trans *trans)
 {
-	if (bitmap_weight(trans->paths_allocated, trans->nr_paths) > BTREE_ITER_INITIAL - 8)
+	if (bitmap_weight(trans->paths_allocated, trans->nr_paths) > BTREE_ITER_NORMAL_LIMIT - 8)
 		return __bch2_btree_trans_too_many_iters(trans);
 
 	return 0;
@@ -694,16 +702,12 @@ transaction_restart:							\
 	_ret2 ?: trans_was_restarted(_trans, _restart_count);		\
 })
 
-#define for_each_btree_key_upto(_trans, _iter, _btree_id,		\
-				_start, _end, _flags, _k, _do)		\
+#define for_each_btree_key_upto_continue(_trans, _iter,			\
+					 _end, _flags, _k, _do)		\
 ({									\
-	struct btree_iter _iter;					\
 	struct bkey_s_c _k;						\
 	int _ret3 = 0;							\
 									\
-	bch2_trans_iter_init((_trans), &(_iter), (_btree_id),		\
-			     (_start), (_flags));			\
-									\
 	do {								\
 		_ret3 = lockrestart_do(_trans, ({			\
 			(_k) = bch2_btree_iter_peek_upto_type(&(_iter),	\
@@ -719,6 +723,21 @@ transaction_restart:							\
 	_ret3;								\
 })
 
+#define for_each_btree_key_continue(_trans, _iter, _flags, _k, _do)	\
+	for_each_btree_key_upto_continue(_trans, _iter, SPOS_MAX, _flags, _k, _do)
+
+#define for_each_btree_key_upto(_trans, _iter, _btree_id,		\
+				_start, _end, _flags, _k, _do)		\
+({									\
+	bch2_trans_begin(trans);					\
+									\
+	struct btree_iter _iter;					\
+	bch2_trans_iter_init((_trans), &(_iter), (_btree_id),		\
+			     (_start), (_flags));			\
+									\
+	for_each_btree_key_upto_continue(_trans, _iter, _end, _flags, _k, _do);\
+})
+
 #define for_each_btree_key(_trans, _iter, _btree_id,			\
 			   _start, _flags, _k, _do)			\
 	for_each_btree_key_upto(_trans, _iter, _btree_id, _start,	\
@@ -789,14 +808,6 @@ __bch2_btree_iter_peek_and_restart(struct btree_trans *trans,
 	return k;
 }
 
-#define for_each_btree_key_old(_trans, _iter, _btree_id,		\
-			   _start, _flags, _k, _ret)			\
-	for (bch2_trans_iter_init((_trans), &(_iter), (_btree_id),	\
-				  (_start), (_flags));			\
-	     (_k) = __bch2_btree_iter_peek_and_restart((_trans), &(_iter), _flags),\
-	     !((_ret) = bkey_err(_k)) && (_k).k;			\
-	     bch2_btree_iter_advance(&(_iter)))
-
 #define for_each_btree_key_upto_norestart(_trans, _iter, _btree_id,	\
 			   _start, _end, _flags, _k, _ret)		\
 	for (bch2_trans_iter_init((_trans), &(_iter), (_btree_id),	\
@@ -827,7 +838,7 @@ __bch2_btree_iter_peek_and_restart(struct btree_trans *trans,
 #define drop_locks_do(_trans, _do)					\
 ({									\
 	bch2_trans_unlock(_trans);					\
-	_do ?: bch2_trans_relock(_trans);				\
+	(_do) ?: bch2_trans_relock(_trans);				\
 })
 
 #define allocate_dropping_locks_errcode(_trans, _do)			\
@@ -856,6 +867,7 @@ __bch2_btree_iter_peek_and_restart(struct btree_trans *trans,
 })
 
 void bch2_trans_updates_to_text(struct printbuf *, struct btree_trans *);
+void bch2_btree_path_to_text(struct printbuf *, struct btree_trans *, btree_path_idx_t);
 void bch2_trans_paths_to_text(struct printbuf *, struct btree_trans *);
 void bch2_dump_trans_updates(struct btree_trans *);
 void bch2_dump_trans_paths_updates(struct btree_trans *);
diff --git a/fs/bcachefs/btree_journal_iter.c b/fs/bcachefs/btree_journal_iter.c
index 719a94a849..332dbf1649 100644
--- a/fs/bcachefs/btree_journal_iter.c
+++ b/fs/bcachefs/btree_journal_iter.c
@@ -1,7 +1,9 @@
 // SPDX-License-Identifier: GPL-2.0
 
 #include "bcachefs.h"
+#include "bkey_buf.h"
 #include "bset.h"
+#include "btree_cache.h"
 #include "btree_journal_iter.h"
 #include "journal_io.h"
 
@@ -40,7 +42,7 @@ static inline size_t idx_to_pos(struct journal_keys *keys, size_t idx)
 
 static inline struct journal_key *idx_to_key(struct journal_keys *keys, size_t idx)
 {
-	return keys->d + idx_to_pos(keys, idx);
+	return keys->data + idx_to_pos(keys, idx);
 }
 
 static size_t __bch2_journal_key_search(struct journal_keys *keys,
@@ -128,12 +130,30 @@ struct bkey_i *bch2_journal_keys_peek_slot(struct bch_fs *c, enum btree_id btree
 	return bch2_journal_keys_peek_upto(c, btree_id, level, pos, pos, &idx);
 }
 
+static void journal_iter_verify(struct journal_iter *iter)
+{
+	struct journal_keys *keys = iter->keys;
+	size_t gap_size = keys->size - keys->nr;
+
+	BUG_ON(iter->idx >= keys->gap &&
+	       iter->idx <  keys->gap + gap_size);
+
+	if (iter->idx < keys->size) {
+		struct journal_key *k = keys->data + iter->idx;
+
+		int cmp = cmp_int(k->btree_id,	iter->btree_id) ?:
+			  cmp_int(k->level,	iter->level);
+		BUG_ON(cmp < 0);
+	}
+}
+
 static void journal_iters_fix(struct bch_fs *c)
 {
 	struct journal_keys *keys = &c->journal_keys;
 	/* The key we just inserted is immediately before the gap: */
 	size_t gap_end = keys->gap + (keys->size - keys->nr);
-	struct btree_and_journal_iter *iter;
+	struct journal_key *new_key = &keys->data[keys->gap - 1];
+	struct journal_iter *iter;
 
 	/*
 	 * If an iterator points one after the key we just inserted, decrement
@@ -141,9 +161,14 @@ static void journal_iters_fix(struct bch_fs *c)
 	 * decrement was unnecessary, bch2_btree_and_journal_iter_peek() will
 	 * handle that:
 	 */
-	list_for_each_entry(iter, &c->journal_iters, journal.list)
-		if (iter->journal.idx == gap_end)
-			iter->journal.idx = keys->gap - 1;
+	list_for_each_entry(iter, &c->journal_iters, list) {
+		journal_iter_verify(iter);
+		if (iter->idx		== gap_end &&
+		    new_key->btree_id	== iter->btree_id &&
+		    new_key->level	== iter->level)
+			iter->idx = keys->gap - 1;
+		journal_iter_verify(iter);
+	}
 }
 
 static void journal_iters_move_gap(struct bch_fs *c, size_t old_gap, size_t new_gap)
@@ -180,33 +205,38 @@ int bch2_journal_key_insert_take(struct bch_fs *c, enum btree_id id,
 	BUG_ON(test_bit(BCH_FS_rw, &c->flags));
 
 	if (idx < keys->size &&
-	    journal_key_cmp(&n, &keys->d[idx]) == 0) {
-		if (keys->d[idx].allocated)
-			kfree(keys->d[idx].k);
-		keys->d[idx] = n;
+	    journal_key_cmp(&n, &keys->data[idx]) == 0) {
+		if (keys->data[idx].allocated)
+			kfree(keys->data[idx].k);
+		keys->data[idx] = n;
 		return 0;
 	}
 
 	if (idx > keys->gap)
 		idx -= keys->size - keys->nr;
 
+	size_t old_gap = keys->gap;
+
 	if (keys->nr == keys->size) {
+		journal_iters_move_gap(c, old_gap, keys->size);
+		old_gap = keys->size;
+
 		struct journal_keys new_keys = {
 			.nr			= keys->nr,
 			.size			= max_t(size_t, keys->size, 8) * 2,
 		};
 
-		new_keys.d = kvmalloc_array(new_keys.size, sizeof(new_keys.d[0]), GFP_KERNEL);
-		if (!new_keys.d) {
+		new_keys.data = kvmalloc_array(new_keys.size, sizeof(new_keys.data[0]), GFP_KERNEL);
+		if (!new_keys.data) {
 			bch_err(c, "%s: error allocating new key array (size %zu)",
 				__func__, new_keys.size);
 			return -BCH_ERR_ENOMEM_journal_key_insert;
 		}
 
 		/* Since @keys was full, there was no gap: */
-		memcpy(new_keys.d, keys->d, sizeof(keys->d[0]) * keys->nr);
-		kvfree(keys->d);
-		keys->d		= new_keys.d;
+		memcpy(new_keys.data, keys->data, sizeof(keys->data[0]) * keys->nr);
+		kvfree(keys->data);
+		keys->data	= new_keys.data;
 		keys->nr	= new_keys.nr;
 		keys->size	= new_keys.size;
 
@@ -214,13 +244,12 @@ int bch2_journal_key_insert_take(struct bch_fs *c, enum btree_id id,
 		keys->gap	= keys->nr;
 	}
 
-	journal_iters_move_gap(c, keys->gap, idx);
+	journal_iters_move_gap(c, old_gap, idx);
 
-	move_gap(keys->d, keys->nr, keys->size, keys->gap, idx);
-	keys->gap = idx;
+	move_gap(keys, idx);
 
 	keys->nr++;
-	keys->d[keys->gap++] = n;
+	keys->data[keys->gap++] = n;
 
 	journal_iters_fix(c);
 
@@ -260,6 +289,22 @@ int bch2_journal_key_delete(struct bch_fs *c, enum btree_id id,
 	return bch2_journal_key_insert(c, id, level, &whiteout);
 }
 
+bool bch2_key_deleted_in_journal(struct btree_trans *trans, enum btree_id btree,
+				 unsigned level, struct bpos pos)
+{
+	struct journal_keys *keys = &trans->c->journal_keys;
+	size_t idx = bch2_journal_key_search(keys, btree, level, pos);
+
+	if (!trans->journal_replay_not_finished)
+		return false;
+
+	return (idx < keys->size &&
+		keys->data[idx].btree_id	== btree &&
+		keys->data[idx].level		== level &&
+		bpos_eq(keys->data[idx].k->k.p, pos) &&
+		bkey_deleted(&keys->data[idx].k->k));
+}
+
 void bch2_journal_key_overwritten(struct bch_fs *c, enum btree_id btree,
 				  unsigned level, struct bpos pos)
 {
@@ -267,10 +312,10 @@ void bch2_journal_key_overwritten(struct bch_fs *c, enum btree_id btree,
 	size_t idx = bch2_journal_key_search(keys, btree, level, pos);
 
 	if (idx < keys->size &&
-	    keys->d[idx].btree_id	== btree &&
-	    keys->d[idx].level		== level &&
-	    bpos_eq(keys->d[idx].k->k.p, pos))
-		keys->d[idx].overwritten = true;
+	    keys->data[idx].btree_id	== btree &&
+	    keys->data[idx].level	== level &&
+	    bpos_eq(keys->data[idx].k->k.p, pos))
+		keys->data[idx].overwritten = true;
 }
 
 static void bch2_journal_iter_advance(struct journal_iter *iter)
@@ -284,16 +329,21 @@ static void bch2_journal_iter_advance(struct journal_iter *iter)
 
 static struct bkey_s_c bch2_journal_iter_peek(struct journal_iter *iter)
 {
-	struct journal_key *k = iter->keys->d + iter->idx;
+	journal_iter_verify(iter);
+
+	while (iter->idx < iter->keys->size) {
+		struct journal_key *k = iter->keys->data + iter->idx;
+
+		int cmp = cmp_int(k->btree_id,	iter->btree_id) ?:
+			  cmp_int(k->level,	iter->level);
+		if (cmp > 0)
+			break;
+		BUG_ON(cmp);
 
-	while (k < iter->keys->d + iter->keys->size &&
-	       k->btree_id	== iter->btree_id &&
-	       k->level		== iter->level) {
 		if (!k->overwritten)
 			return bkey_i_to_s_c(k->k);
 
 		bch2_journal_iter_advance(iter);
-		k = iter->keys->d + iter->idx;
 	}
 
 	return bkey_s_c_null;
@@ -313,6 +363,8 @@ static void bch2_journal_iter_init(struct bch_fs *c,
 	iter->level	= level;
 	iter->keys	= &c->journal_keys;
 	iter->idx	= bch2_journal_key_search(&c->journal_keys, id, level, pos);
+
+	journal_iter_verify(iter);
 }
 
 static struct bkey_s_c bch2_journal_iter_peek_btree(struct btree_and_journal_iter *iter)
@@ -334,9 +386,38 @@ void bch2_btree_and_journal_iter_advance(struct btree_and_journal_iter *iter)
 		iter->pos = bpos_successor(iter->pos);
 }
 
+static void btree_and_journal_iter_prefetch(struct btree_and_journal_iter *_iter)
+{
+	struct btree_and_journal_iter iter = *_iter;
+	struct bch_fs *c = iter.trans->c;
+	unsigned level = iter.journal.level;
+	struct bkey_buf tmp;
+	unsigned nr = test_bit(BCH_FS_started, &c->flags)
+		? (level > 1 ? 0 :  2)
+		: (level > 1 ? 1 : 16);
+
+	iter.prefetch = false;
+	bch2_bkey_buf_init(&tmp);
+
+	while (nr--) {
+		bch2_btree_and_journal_iter_advance(&iter);
+		struct bkey_s_c k = bch2_btree_and_journal_iter_peek(&iter);
+		if (!k.k)
+			break;
+
+		bch2_bkey_buf_reassemble(&tmp, c, k);
+		bch2_btree_node_prefetch(iter.trans, NULL, tmp.k, iter.journal.btree_id, level - 1);
+	}
+
+	bch2_bkey_buf_exit(&tmp, c);
+}
+
 struct bkey_s_c bch2_btree_and_journal_iter_peek(struct btree_and_journal_iter *iter)
 {
-	struct bkey_s_c btree_k, journal_k, ret;
+	struct bkey_s_c btree_k, journal_k = bkey_s_c_null, ret;
+
+	if (iter->prefetch && iter->journal.level)
+		btree_and_journal_iter_prefetch(iter);
 again:
 	if (iter->at_end)
 		return bkey_s_c_null;
@@ -345,9 +426,10 @@ again:
 	       bpos_lt(btree_k.k->p, iter->pos))
 		bch2_journal_iter_advance_btree(iter);
 
-	while ((journal_k = bch2_journal_iter_peek(&iter->journal)).k &&
-	       bpos_lt(journal_k.k->p, iter->pos))
-		bch2_journal_iter_advance(&iter->journal);
+	if (iter->trans->journal_replay_not_finished)
+		while ((journal_k = bch2_journal_iter_peek(&iter->journal)).k &&
+		       bpos_lt(journal_k.k->p, iter->pos))
+			bch2_journal_iter_advance(&iter->journal);
 
 	ret = journal_k.k &&
 		(!btree_k.k || bpos_le(journal_k.k->p, btree_k.k->p))
@@ -376,35 +458,40 @@ void bch2_btree_and_journal_iter_exit(struct btree_and_journal_iter *iter)
 	bch2_journal_iter_exit(&iter->journal);
 }
 
-void __bch2_btree_and_journal_iter_init_node_iter(struct btree_and_journal_iter *iter,
-						  struct bch_fs *c,
+void __bch2_btree_and_journal_iter_init_node_iter(struct btree_trans *trans,
+						  struct btree_and_journal_iter *iter,
 						  struct btree *b,
 						  struct btree_node_iter node_iter,
 						  struct bpos pos)
 {
 	memset(iter, 0, sizeof(*iter));
 
+	iter->trans = trans;
 	iter->b = b;
 	iter->node_iter = node_iter;
-	bch2_journal_iter_init(c, &iter->journal, b->c.btree_id, b->c.level, pos);
-	INIT_LIST_HEAD(&iter->journal.list);
 	iter->pos = b->data->min_key;
 	iter->at_end = false;
+	INIT_LIST_HEAD(&iter->journal.list);
+
+	if (trans->journal_replay_not_finished) {
+		bch2_journal_iter_init(trans->c, &iter->journal, b->c.btree_id, b->c.level, pos);
+		if (!test_bit(BCH_FS_may_go_rw, &trans->c->flags))
+			list_add(&iter->journal.list, &trans->c->journal_iters);
+	}
 }
 
 /*
  * this version is used by btree_gc before filesystem has gone RW and
  * multithreaded, so uses the journal_iters list:
  */
-void bch2_btree_and_journal_iter_init_node_iter(struct btree_and_journal_iter *iter,
-						struct bch_fs *c,
+void bch2_btree_and_journal_iter_init_node_iter(struct btree_trans *trans,
+						struct btree_and_journal_iter *iter,
 						struct btree *b)
 {
 	struct btree_node_iter node_iter;
 
 	bch2_btree_node_iter_init_from_start(&node_iter, b);
-	__bch2_btree_and_journal_iter_init_node_iter(iter, c, b, node_iter, b->data->min_key);
-	list_add(&iter->journal.list, &c->journal_iters);
+	__bch2_btree_and_journal_iter_init_node_iter(trans, iter, b, node_iter, b->data->min_key);
 }
 
 /* sort and dedup all keys in the journal: */
@@ -415,9 +502,7 @@ void bch2_journal_entries_free(struct bch_fs *c)
 	struct genradix_iter iter;
 
 	genradix_for_each(&c->journal_entries, iter, i)
-		if (*i)
-			kvpfree(*i, offsetof(struct journal_replay, j) +
-				vstruct_bytes(&(*i)->j));
+		kvfree(*i);
 	genradix_free(&c->journal_entries);
 }
 
@@ -437,22 +522,20 @@ static int journal_sort_key_cmp(const void *_l, const void *_r)
 void bch2_journal_keys_put(struct bch_fs *c)
 {
 	struct journal_keys *keys = &c->journal_keys;
-	struct journal_key *i;
 
 	BUG_ON(atomic_read(&keys->ref) <= 0);
 
 	if (!atomic_dec_and_test(&keys->ref))
 		return;
 
-	move_gap(keys->d, keys->nr, keys->size, keys->gap, keys->nr);
-	keys->gap = keys->nr;
+	move_gap(keys, keys->nr);
 
-	for (i = keys->d; i < keys->d + keys->nr; i++)
+	darray_for_each(*keys, i)
 		if (i->allocated)
 			kfree(i->k);
 
-	kvfree(keys->d);
-	keys->d = NULL;
+	kvfree(keys->data);
+	keys->data = NULL;
 	keys->nr = keys->gap = keys->size = 0;
 
 	bch2_journal_entries_free(c);
@@ -460,83 +543,38 @@ void bch2_journal_keys_put(struct bch_fs *c)
 
 static void __journal_keys_sort(struct journal_keys *keys)
 {
-	struct journal_key *src, *dst;
+	sort(keys->data, keys->nr, sizeof(keys->data[0]), journal_sort_key_cmp, NULL);
 
-	sort(keys->d, keys->nr, sizeof(keys->d[0]), journal_sort_key_cmp, NULL);
+	struct journal_key *dst = keys->data;
 
-	src = dst = keys->d;
-	while (src < keys->d + keys->nr) {
-		while (src + 1 < keys->d + keys->nr &&
-		       !journal_key_cmp(src, src + 1))
-			src++;
+	darray_for_each(*keys, src) {
+		if (src + 1 < &darray_top(*keys) &&
+		    !journal_key_cmp(src, src + 1))
+			continue;
 
-		*dst++ = *src++;
+		*dst++ = *src;
 	}
 
-	keys->nr = dst - keys->d;
+	keys->nr = dst - keys->data;
 }
 
 int bch2_journal_keys_sort(struct bch_fs *c)
 {
 	struct genradix_iter iter;
 	struct journal_replay *i, **_i;
-	struct jset_entry *entry;
-	struct bkey_i *k;
 	struct journal_keys *keys = &c->journal_keys;
-	size_t nr_keys = 0, nr_read = 0;
-
-	genradix_for_each(&c->journal_entries, iter, _i) {
-		i = *_i;
-
-		if (!i || i->ignore)
-			continue;
-
-		for_each_jset_key(k, entry, &i->j)
-			nr_keys++;
-	}
-
-	if (!nr_keys)
-		return 0;
-
-	keys->size = roundup_pow_of_two(nr_keys);
-
-	keys->d = kvmalloc_array(keys->size, sizeof(keys->d[0]), GFP_KERNEL);
-	if (!keys->d) {
-		bch_err(c, "Failed to allocate buffer for sorted journal keys (%zu keys); trying slowpath",
-			nr_keys);
-
-		do {
-			keys->size >>= 1;
-			keys->d = kvmalloc_array(keys->size, sizeof(keys->d[0]), GFP_KERNEL);
-		} while (!keys->d && keys->size > nr_keys / 8);
-
-		if (!keys->d) {
-			bch_err(c, "Failed to allocate %zu size buffer for sorted journal keys; exiting",
-				keys->size);
-			return -BCH_ERR_ENOMEM_journal_keys_sort;
-		}
-	}
+	size_t nr_read = 0;
 
 	genradix_for_each(&c->journal_entries, iter, _i) {
 		i = *_i;
 
-		if (!i || i->ignore)
+		if (journal_replay_ignore(i))
 			continue;
 
 		cond_resched();
 
 		for_each_jset_key(k, entry, &i->j) {
-			if (keys->nr == keys->size) {
-				__journal_keys_sort(keys);
-
-				if (keys->nr > keys->size * 7 / 8) {
-					bch_err(c, "Too many journal keys for slowpath; have %zu compacted, buf size %zu, processed %zu/%zu",
-						keys->nr, keys->size, nr_read, nr_keys);
-					return -BCH_ERR_ENOMEM_journal_keys_sort;
-				}
-			}
-
-			keys->d[keys->nr++] = (struct journal_key) {
+			struct journal_key n = (struct journal_key) {
 				.btree_id	= entry->btree_id,
 				.level		= entry->level,
 				.k		= k,
@@ -544,6 +582,18 @@ int bch2_journal_keys_sort(struct bch_fs *c)
 				.journal_offset	= k->_data - i->j._data,
 			};
 
+			if (darray_push(keys, n)) {
+				__journal_keys_sort(keys);
+
+				if (keys->nr * 8 > keys->size * 7) {
+					bch_err(c, "Too many journal keys for slowpath; have %zu compacted, buf size %zu, processed %zu keys at seq %llu",
+						keys->nr, keys->size, nr_read, le64_to_cpu(i->j.seq));
+					return -BCH_ERR_ENOMEM_journal_keys_sort;
+				}
+
+				BUG_ON(darray_push(keys, n));
+			}
+
 			nr_read++;
 		}
 	}
@@ -551,6 +601,42 @@ int bch2_journal_keys_sort(struct bch_fs *c)
 	__journal_keys_sort(keys);
 	keys->gap = keys->nr;
 
-	bch_verbose(c, "Journal keys: %zu read, %zu after sorting and compacting", nr_keys, keys->nr);
+	bch_verbose(c, "Journal keys: %zu read, %zu after sorting and compacting", nr_read, keys->nr);
 	return 0;
 }
+
+void bch2_shoot_down_journal_keys(struct bch_fs *c, enum btree_id btree,
+				  unsigned level_min, unsigned level_max,
+				  struct bpos start, struct bpos end)
+{
+	struct journal_keys *keys = &c->journal_keys;
+	size_t dst = 0;
+
+	move_gap(keys, keys->nr);
+
+	darray_for_each(*keys, i)
+		if (!(i->btree_id == btree &&
+		      i->level >= level_min &&
+		      i->level <= level_max &&
+		      bpos_ge(i->k->k.p, start) &&
+		      bpos_le(i->k->k.p, end)))
+			keys->data[dst++] = *i;
+	keys->nr = keys->gap = dst;
+}
+
+void bch2_journal_keys_dump(struct bch_fs *c)
+{
+	struct journal_keys *keys = &c->journal_keys;
+	struct printbuf buf = PRINTBUF;
+
+	pr_info("%zu keys:", keys->nr);
+
+	move_gap(keys, keys->nr);
+
+	darray_for_each(*keys, i) {
+		printbuf_reset(&buf);
+		bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(i->k));
+		pr_err("%s l=%u %s", bch2_btree_id_str(i->btree_id), i->level, buf.buf);
+	}
+	printbuf_exit(&buf);
+}
diff --git a/fs/bcachefs/btree_journal_iter.h b/fs/bcachefs/btree_journal_iter.h
index 8ca4c100b2..1ba4a79b0e 100644
--- a/fs/bcachefs/btree_journal_iter.h
+++ b/fs/bcachefs/btree_journal_iter.h
@@ -15,6 +15,7 @@ struct journal_iter {
  */
 
 struct btree_and_journal_iter {
+	struct btree_trans	*trans;
 	struct btree		*b;
 	struct btree_node_iter	node_iter;
 	struct bkey		unpacked;
@@ -22,6 +23,7 @@ struct btree_and_journal_iter {
 	struct journal_iter	journal;
 	struct bpos		pos;
 	bool			at_end;
+	bool			prefetch;
 };
 
 struct bkey_i *bch2_journal_keys_peek_upto(struct bch_fs *, enum btree_id,
@@ -29,25 +31,27 @@ struct bkey_i *bch2_journal_keys_peek_upto(struct bch_fs *, enum btree_id,
 struct bkey_i *bch2_journal_keys_peek_slot(struct bch_fs *, enum btree_id,
 					   unsigned, struct bpos);
 
+int bch2_btree_and_journal_iter_prefetch(struct btree_trans *, struct btree_path *,
+					 struct btree_and_journal_iter *);
+
 int bch2_journal_key_insert_take(struct bch_fs *, enum btree_id,
 				 unsigned, struct bkey_i *);
 int bch2_journal_key_insert(struct bch_fs *, enum btree_id,
 			    unsigned, struct bkey_i *);
 int bch2_journal_key_delete(struct bch_fs *, enum btree_id,
 			    unsigned, struct bpos);
-void bch2_journal_key_overwritten(struct bch_fs *, enum btree_id,
-				  unsigned, struct bpos);
+bool bch2_key_deleted_in_journal(struct btree_trans *, enum btree_id, unsigned, struct bpos);
+void bch2_journal_key_overwritten(struct bch_fs *, enum btree_id, unsigned, struct bpos);
 
 void bch2_btree_and_journal_iter_advance(struct btree_and_journal_iter *);
 struct bkey_s_c bch2_btree_and_journal_iter_peek(struct btree_and_journal_iter *);
 
 void bch2_btree_and_journal_iter_exit(struct btree_and_journal_iter *);
-void __bch2_btree_and_journal_iter_init_node_iter(struct btree_and_journal_iter *,
-				struct bch_fs *, struct btree *,
+void __bch2_btree_and_journal_iter_init_node_iter(struct btree_trans *,
+				struct btree_and_journal_iter *, struct btree *,
 				struct btree_node_iter, struct bpos);
-void bch2_btree_and_journal_iter_init_node_iter(struct btree_and_journal_iter *,
-						struct bch_fs *,
-						struct btree *);
+void bch2_btree_and_journal_iter_init_node_iter(struct btree_trans *,
+				struct btree_and_journal_iter *, struct btree *);
 
 void bch2_journal_keys_put(struct bch_fs *);
 
@@ -62,4 +66,10 @@ void bch2_journal_entries_free(struct bch_fs *);
 
 int bch2_journal_keys_sort(struct bch_fs *);
 
+void bch2_shoot_down_journal_keys(struct bch_fs *, enum btree_id,
+				  unsigned, unsigned,
+				  struct bpos, struct bpos);
+
+void bch2_journal_keys_dump(struct bch_fs *);
+
 #endif /* _BCACHEFS_BTREE_JOURNAL_ITER_H */
diff --git a/fs/bcachefs/btree_key_cache.c b/fs/bcachefs/btree_key_cache.c
index 74e52fd28a..2d3c0d45c3 100644
--- a/fs/bcachefs/btree_key_cache.c
+++ b/fs/bcachefs/btree_key_cache.c
@@ -32,10 +32,11 @@ static int bch2_btree_key_cache_cmp_fn(struct rhashtable_compare_arg *arg,
 }
 
 static const struct rhashtable_params bch2_btree_key_cache_params = {
-	.head_offset	= offsetof(struct bkey_cached, hash),
-	.key_offset	= offsetof(struct bkey_cached, key),
-	.key_len	= sizeof(struct bkey_cached_key),
-	.obj_cmpfn	= bch2_btree_key_cache_cmp_fn,
+	.head_offset		= offsetof(struct bkey_cached, hash),
+	.key_offset		= offsetof(struct bkey_cached, key),
+	.key_len		= sizeof(struct bkey_cached_key),
+	.obj_cmpfn		= bch2_btree_key_cache_cmp_fn,
+	.automatic_shrinking	= true,
 };
 
 __flatten
@@ -169,6 +170,7 @@ static void bkey_cached_move_to_freelist(struct btree_key_cache *bc,
 	} else {
 		mutex_lock(&bc->lock);
 		list_move_tail(&ck->list, &bc->freed_pcpu);
+		bc->nr_freed_pcpu++;
 		mutex_unlock(&bc->lock);
 	}
 }
@@ -245,6 +247,7 @@ bkey_cached_alloc(struct btree_trans *trans, struct btree_path *path,
 		if (!list_empty(&bc->freed_pcpu)) {
 			ck = list_last_entry(&bc->freed_pcpu, struct bkey_cached, list);
 			list_del_init(&ck->list);
+			bc->nr_freed_pcpu--;
 		}
 		mutex_unlock(&bc->lock);
 	}
@@ -380,9 +383,11 @@ static int btree_key_cache_fill(struct btree_trans *trans,
 	struct bkey_i *new_k = NULL;
 	int ret;
 
-	k = bch2_bkey_get_iter(trans, &iter, ck->key.btree_id, ck->key.pos,
-			       BTREE_ITER_KEY_CACHE_FILL|
-			       BTREE_ITER_CACHED_NOFILL);
+	bch2_trans_iter_init(trans, &iter, ck->key.btree_id, ck->key.pos,
+			     BTREE_ITER_key_cache_fill|
+			     BTREE_ITER_cached_nofill);
+	iter.flags &= ~BTREE_ITER_with_journal;
+	k = bch2_btree_iter_peek_slot(&iter);
 	ret = bkey_err(k);
 	if (ret)
 		goto err;
@@ -420,16 +425,16 @@ static int btree_key_cache_fill(struct btree_trans *trans,
 				goto err;
 			}
 
-			if (!bch2_btree_node_relock(trans, ck_path, 0)) {
+			ret = bch2_trans_relock(trans);
+			if (ret) {
 				kfree(new_k);
-				trace_and_count(trans->c, trans_restart_relock_key_cache_fill, trans, _THIS_IP_, ck_path);
-				ret = btree_trans_restart(trans, BCH_ERR_transaction_restart_key_cache_fill);
 				goto err;
 			}
 
-			ret = bch2_trans_relock(trans);
-			if (ret) {
+			if (!bch2_btree_node_relock(trans, ck_path, 0)) {
 				kfree(new_k);
+				trace_and_count(trans->c, trans_restart_relock_key_cache_fill, trans, _THIS_IP_, ck_path);
+				ret = btree_trans_restart(trans, BCH_ERR_transaction_restart_key_cache_fill);
 				goto err;
 			}
 		}
@@ -452,7 +457,7 @@ static int btree_key_cache_fill(struct btree_trans *trans,
 	bch2_btree_node_unlock_write(trans, ck_path, ck_path->l[0].b);
 
 	/* We're not likely to need this iterator again: */
-	set_btree_iter_dontneed(&iter);
+	bch2_set_btree_iter_dontneed(&iter);
 err:
 	bch2_trans_iter_exit(trans, &iter);
 	return ret;
@@ -511,23 +516,10 @@ retry:
 fill:
 	path->uptodate = BTREE_ITER_UPTODATE;
 
-	if (!ck->valid && !(flags & BTREE_ITER_CACHED_NOFILL)) {
-		/*
-		 * Using the underscore version because we haven't set
-		 * path->uptodate yet:
-		 */
-		if (!path->locks_want &&
-		    !__bch2_btree_path_upgrade(trans, path, 1, NULL)) {
-			trace_and_count(trans->c, trans_restart_key_cache_upgrade, trans, _THIS_IP_);
-			ret = btree_trans_restart(trans, BCH_ERR_transaction_restart_key_cache_upgrade);
-			goto err;
-		}
-
-		ret = btree_key_cache_fill(trans, path, ck);
-		if (ret)
-			goto err;
-
-		ret = bch2_btree_path_relock(trans, path, _THIS_IP_);
+	if (!ck->valid && !(flags & BTREE_ITER_cached_nofill)) {
+		ret =   bch2_btree_path_upgrade(trans, path, 1) ?:
+			btree_key_cache_fill(trans, path, ck) ?:
+			bch2_btree_path_relock(trans, path, _THIS_IP_);
 		if (ret)
 			goto err;
 
@@ -618,13 +610,13 @@ static int btree_key_cache_flush_pos(struct btree_trans *trans,
 	int ret;
 
 	bch2_trans_iter_init(trans, &b_iter, key.btree_id, key.pos,
-			     BTREE_ITER_SLOTS|
-			     BTREE_ITER_INTENT|
-			     BTREE_ITER_ALL_SNAPSHOTS);
+			     BTREE_ITER_slots|
+			     BTREE_ITER_intent|
+			     BTREE_ITER_all_snapshots);
 	bch2_trans_iter_init(trans, &c_iter, key.btree_id, key.pos,
-			     BTREE_ITER_CACHED|
-			     BTREE_ITER_INTENT);
-	b_iter.flags &= ~BTREE_ITER_WITH_KEY_CACHE;
+			     BTREE_ITER_cached|
+			     BTREE_ITER_intent);
+	b_iter.flags &= ~BTREE_ITER_with_key_cache;
 
 	ret = bch2_btree_iter_traverse(&c_iter);
 	if (ret)
@@ -657,14 +649,14 @@ static int btree_key_cache_flush_pos(struct btree_trans *trans,
 		commit_flags |= BCH_WATERMARK_reclaim;
 
 	if (ck->journal.seq != journal_last_seq(j) ||
-	    j->watermark == BCH_WATERMARK_stripe)
+	    !test_bit(JOURNAL_space_low, &c->journal.flags))
 		commit_flags |= BCH_TRANS_COMMIT_no_journal_res;
 
 	ret   = bch2_btree_iter_traverse(&b_iter) ?:
 		bch2_trans_update(trans, &b_iter, ck->k,
-				  BTREE_UPDATE_KEY_CACHE_RECLAIM|
-				  BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE|
-				  BTREE_TRIGGER_NORUN) ?:
+				  BTREE_UPDATE_key_cache_reclaim|
+				  BTREE_UPDATE_internal_snapshot_node|
+				  BTREE_TRIGGER_norun) ?:
 		bch2_trans_commit(trans, NULL, NULL,
 				  BCH_TRANS_COMMIT_no_check_rw|
 				  BCH_TRANS_COMMIT_no_enospc|
@@ -674,7 +666,7 @@ static int btree_key_cache_flush_pos(struct btree_trans *trans,
 			     !bch2_err_matches(ret, BCH_ERR_transaction_restart) &&
 			     !bch2_err_matches(ret, BCH_ERR_journal_reclaim_would_deadlock) &&
 			     !bch2_journal_error(j), c,
-			     "error flushing key cache: %s", bch2_err_str(ret));
+			     "flushing key cache: %s", bch2_err_str(ret));
 	if (ret)
 		goto out;
 
@@ -786,7 +778,7 @@ bool bch2_btree_insert_key_cached(struct btree_trans *trans,
 	 * flushing. The flush callback will not proceed unless ->seq matches
 	 * the latest pin, so make sure it starts with a consistent value.
 	 */
-	if (!(insert_entry->flags & BTREE_UPDATE_NOJOURNAL) ||
+	if (!(insert_entry->flags & BTREE_UPDATE_nojournal) ||
 	    !journal_pin_active(&ck->journal)) {
 		ck->seq = trans->journal_res.seq;
 	}
@@ -831,6 +823,8 @@ static unsigned long bch2_btree_key_cache_scan(struct shrinker *shrink,
 	int srcu_idx;
 
 	mutex_lock(&bc->lock);
+	bc->requested_to_free += sc->nr_to_scan;
+
 	srcu_idx = srcu_read_lock(&c->btree_trans_barrier);
 	flags = memalloc_nofs_save();
 
@@ -838,8 +832,6 @@ static unsigned long bch2_btree_key_cache_scan(struct shrinker *shrink,
 	 * Newest freed entries are at the end of the list - once we hit one
 	 * that's too new to be freed, we can bail out:
 	 */
-	scanned += bc->nr_freed_nonpcpu;
-
 	list_for_each_entry_safe(ck, t, &bc->freed_nonpcpu, list) {
 		if (!poll_state_synchronize_srcu(&c->btree_trans_barrier,
 						 ck->btree_trans_barrier_seq))
@@ -849,15 +841,10 @@ static unsigned long bch2_btree_key_cache_scan(struct shrinker *shrink,
 		six_lock_exit(&ck->c.lock);
 		kmem_cache_free(bch2_key_cache, ck);
 		atomic_long_dec(&bc->nr_freed);
-		freed++;
 		bc->nr_freed_nonpcpu--;
+		bc->freed++;
 	}
 
-	if (scanned >= nr)
-		goto out;
-
-	scanned += bc->nr_freed_pcpu;
-
 	list_for_each_entry_safe(ck, t, &bc->freed_pcpu, list) {
 		if (!poll_state_synchronize_srcu(&c->btree_trans_barrier,
 						 ck->btree_trans_barrier_seq))
@@ -867,13 +854,10 @@ static unsigned long bch2_btree_key_cache_scan(struct shrinker *shrink,
 		six_lock_exit(&ck->c.lock);
 		kmem_cache_free(bch2_key_cache, ck);
 		atomic_long_dec(&bc->nr_freed);
-		freed++;
 		bc->nr_freed_pcpu--;
+		bc->freed++;
 	}
 
-	if (scanned >= nr)
-		goto out;
-
 	rcu_read_lock();
 	tbl = rht_dereference_rcu(bc->table.tbl, &bc->table);
 	if (bc->shrink_iter >= tbl->size)
@@ -889,20 +873,24 @@ static unsigned long bch2_btree_key_cache_scan(struct shrinker *shrink,
 			next = rht_dereference_bucket_rcu(pos->next, tbl, bc->shrink_iter);
 			ck = container_of(pos, struct bkey_cached, hash);
 
-			if (test_bit(BKEY_CACHED_DIRTY, &ck->flags))
-				goto next;
-
-			if (test_bit(BKEY_CACHED_ACCESSED, &ck->flags))
+			if (test_bit(BKEY_CACHED_DIRTY, &ck->flags)) {
+				bc->skipped_dirty++;
+			} else if (test_bit(BKEY_CACHED_ACCESSED, &ck->flags)) {
 				clear_bit(BKEY_CACHED_ACCESSED, &ck->flags);
-			else if (bkey_cached_lock_for_evict(ck)) {
+				bc->skipped_accessed++;
+			} else if (!bkey_cached_lock_for_evict(ck)) {
+				bc->skipped_lock_fail++;
+			} else {
 				bkey_cached_evict(bc, ck);
 				bkey_cached_free(bc, ck);
+				bc->moved_to_freelist++;
+				freed++;
 			}
 
 			scanned++;
 			if (scanned >= nr)
 				break;
-next:
+
 			pos = next;
 		}
 
@@ -912,7 +900,6 @@ next:
 	} while (scanned < nr && bc->shrink_iter != start);
 
 	rcu_read_unlock();
-out:
 	memalloc_nofs_restore(flags);
 	srcu_read_unlock(&c->btree_trans_barrier, srcu_idx);
 	mutex_unlock(&bc->lock);
@@ -928,6 +915,14 @@ static unsigned long bch2_btree_key_cache_count(struct shrinker *shrink,
 	long nr = atomic_long_read(&bc->nr_keys) -
 		atomic_long_read(&bc->nr_dirty);
 
+	/*
+	 * Avoid hammering our shrinker too much if it's nearly empty - the
+	 * shrinker code doesn't take into account how big our cache is, if it's
+	 * mostly empty but the system is under memory pressure it causes nasty
+	 * lock contention:
+	 */
+	nr -= 128;
+
 	return max(0L, nr);
 }
 
@@ -963,13 +958,15 @@ void bch2_fs_btree_key_cache_exit(struct btree_key_cache *bc)
 	}
 
 #ifdef __KERNEL__
-	for_each_possible_cpu(cpu) {
-		struct btree_key_cache_freelist *f =
-			per_cpu_ptr(bc->pcpu_freed, cpu);
-
-		for (i = 0; i < f->nr; i++) {
-			ck = f->objs[i];
-			list_add(&ck->list, &items);
+	if (bc->pcpu_freed) {
+		for_each_possible_cpu(cpu) {
+			struct btree_key_cache_freelist *f =
+				per_cpu_ptr(bc->pcpu_freed, cpu);
+
+			for (i = 0; i < f->nr; i++) {
+				ck = f->objs[i];
+				list_add(&ck->list, &items);
+			}
 		}
 	}
 #endif
@@ -1034,22 +1031,56 @@ int bch2_fs_btree_key_cache_init(struct btree_key_cache *bc)
 	if (!shrink)
 		return -BCH_ERR_ENOMEM_fs_btree_cache_init;
 	bc->shrink = shrink;
-	shrink->seeks		= 0;
 	shrink->count_objects	= bch2_btree_key_cache_count;
 	shrink->scan_objects	= bch2_btree_key_cache_scan;
+	shrink->batch		= 1 << 14;
+	shrink->seeks		= 0;
 	shrink->private_data	= c;
 	shrinker_register(shrink);
 	return 0;
 }
 
-void bch2_btree_key_cache_to_text(struct printbuf *out, struct btree_key_cache *c)
+void bch2_btree_key_cache_to_text(struct printbuf *out, struct btree_key_cache *bc)
 {
-	prt_printf(out, "nr_freed:\t%lu",	atomic_long_read(&c->nr_freed));
-	prt_newline(out);
-	prt_printf(out, "nr_keys:\t%lu",	atomic_long_read(&c->nr_keys));
-	prt_newline(out);
-	prt_printf(out, "nr_dirty:\t%lu",	atomic_long_read(&c->nr_dirty));
-	prt_newline(out);
+	struct bch_fs *c = container_of(bc, struct bch_fs, btree_key_cache);
+
+	printbuf_tabstop_push(out, 24);
+	printbuf_tabstop_push(out, 12);
+
+	unsigned flags = memalloc_nofs_save();
+	mutex_lock(&bc->lock);
+	prt_printf(out, "keys:\t%lu\r\n",		atomic_long_read(&bc->nr_keys));
+	prt_printf(out, "dirty:\t%lu\r\n",		atomic_long_read(&bc->nr_dirty));
+	prt_printf(out, "freelist:\t%lu\r\n",		atomic_long_read(&bc->nr_freed));
+	prt_printf(out, "nonpcpu freelist:\t%zu\r\n",	bc->nr_freed_nonpcpu);
+	prt_printf(out, "pcpu freelist:\t%zu\r\n",	bc->nr_freed_pcpu);
+
+	prt_printf(out, "\nshrinker:\n");
+	prt_printf(out, "requested_to_free:\t%lu\r\n",	bc->requested_to_free);
+	prt_printf(out, "freed:\t%lu\r\n",		bc->freed);
+	prt_printf(out, "moved_to_freelist:\t%lu\r\n",	bc->moved_to_freelist);
+	prt_printf(out, "skipped_dirty:\t%lu\r\n",	bc->skipped_dirty);
+	prt_printf(out, "skipped_accessed:\t%lu\r\n",	bc->skipped_accessed);
+	prt_printf(out, "skipped_lock_fail:\t%lu\r\n",	bc->skipped_lock_fail);
+
+	prt_printf(out, "srcu seq:\t%lu\r\n",		get_state_synchronize_srcu(&c->btree_trans_barrier));
+
+	struct bkey_cached *ck;
+	unsigned iter = 0;
+	list_for_each_entry(ck, &bc->freed_nonpcpu, list) {
+		prt_printf(out, "freed_nonpcpu:\t%lu\r\n", ck->btree_trans_barrier_seq);
+		if (++iter > 10)
+			break;
+	}
+
+	iter = 0;
+	list_for_each_entry(ck, &bc->freed_pcpu, list) {
+		prt_printf(out, "freed_pcpu:\t%lu\r\n", ck->btree_trans_barrier_seq);
+		if (++iter > 10)
+			break;
+	}
+	mutex_unlock(&bc->lock);
+	memalloc_flags_restore(flags);
 }
 
 void bch2_btree_key_cache_exit(void)
diff --git a/fs/bcachefs/btree_key_cache_types.h b/fs/bcachefs/btree_key_cache_types.h
index 290e4e57df..237e8bb3ac 100644
--- a/fs/bcachefs/btree_key_cache_types.h
+++ b/fs/bcachefs/btree_key_cache_types.h
@@ -24,6 +24,14 @@ struct btree_key_cache {
 	atomic_long_t		nr_freed;
 	atomic_long_t		nr_keys;
 	atomic_long_t		nr_dirty;
+
+	/* shrinker stats */
+	unsigned long		requested_to_free;
+	unsigned long		freed;
+	unsigned long		moved_to_freelist;
+	unsigned long		skipped_dirty;
+	unsigned long		skipped_accessed;
+	unsigned long		skipped_lock_fail;
 };
 
 struct bkey_cached_key {
diff --git a/fs/bcachefs/btree_locking.c b/fs/bcachefs/btree_locking.c
index 6843974423..c51826fd55 100644
--- a/fs/bcachefs/btree_locking.c
+++ b/fs/bcachefs/btree_locking.c
@@ -83,8 +83,7 @@ static noinline void print_cycle(struct printbuf *out, struct lock_graph *g)
 {
 	struct trans_waiting_for_lock *i;
 
-	prt_printf(out, "Found lock cycle (%u entries):", g->nr);
-	prt_newline(out);
+	prt_printf(out, "Found lock cycle (%u entries):\n", g->nr);
 
 	for (i = g->g; i < g->g + g->nr; i++) {
 		struct task_struct *task = READ_ONCE(i->trans->locking_wait.task);
@@ -216,6 +215,7 @@ static noinline int break_cycle(struct lock_graph *g, struct printbuf *cycle)
 
 	if (unlikely(!best)) {
 		struct printbuf buf = PRINTBUF;
+		buf.atomic++;
 
 		prt_printf(&buf, bch2_fmt(g->g->trans->c, "cycle of nofail locks"));
 
@@ -224,15 +224,14 @@ static noinline int break_cycle(struct lock_graph *g, struct printbuf *cycle)
 
 			bch2_btree_trans_to_text(&buf, trans);
 
-			prt_printf(&buf, "backtrace:");
-			prt_newline(&buf);
+			prt_printf(&buf, "backtrace:\n");
 			printbuf_indent_add(&buf, 2);
 			bch2_prt_task_backtrace(&buf, trans->locking_wait.task, 2, GFP_NOWAIT);
 			printbuf_indent_sub(&buf, 2);
 			prt_newline(&buf);
 		}
 
-		bch2_print_string_as_lines(KERN_ERR, buf.buf);
+		bch2_print_string_as_lines_nonblocking(KERN_ERR, buf.buf);
 		printbuf_exit(&buf);
 		BUG();
 	}
@@ -440,33 +439,7 @@ void bch2_btree_node_lock_write_nofail(struct btree_trans *trans,
 				       struct btree_path *path,
 				       struct btree_bkey_cached_common *b)
 {
-	struct btree_path *linked;
-	unsigned i, iter;
-	int ret;
-
-	/*
-	 * XXX BIG FAT NOTICE
-	 *
-	 * Drop all read locks before taking a write lock:
-	 *
-	 * This is a hack, because bch2_btree_node_lock_write_nofail() is a
-	 * hack - but by dropping read locks first, this should never fail, and
-	 * we only use this in code paths where whatever read locks we've
-	 * already taken are no longer needed:
-	 */
-
-	trans_for_each_path(trans, linked, iter) {
-		if (!linked->nodes_locked)
-			continue;
-
-		for (i = 0; i < BTREE_MAX_DEPTH; i++)
-			if (btree_node_read_locked(linked, i)) {
-				btree_node_unlock(trans, linked, i);
-				btree_path_set_dirty(linked, BTREE_ITER_NEED_RELOCK);
-			}
-	}
-
-	ret = __btree_node_lock_write(trans, path, b, true);
+	int ret = __btree_node_lock_write(trans, path, b, true);
 	BUG_ON(ret);
 }
 
@@ -518,8 +491,6 @@ static inline bool btree_path_get_locks(struct btree_trans *trans,
 	if (path->uptodate == BTREE_ITER_NEED_RELOCK)
 		path->uptodate = BTREE_ITER_UPTODATE;
 
-	bch2_trans_verify_locks(trans);
-
 	return path->uptodate < BTREE_ITER_NEED_RELOCK;
 }
 
@@ -635,7 +606,9 @@ bool bch2_btree_path_relock_norestart(struct btree_trans *trans, struct btree_pa
 {
 	struct get_locks_fail f;
 
-	return btree_path_get_locks(trans, path, false, &f);
+	bool ret = btree_path_get_locks(trans, path, false, &f);
+	bch2_trans_verify_locks(trans);
+	return ret;
 }
 
 int __bch2_btree_path_relock(struct btree_trans *trans,
@@ -658,7 +631,9 @@ bool bch2_btree_path_upgrade_noupgrade_sibs(struct btree_trans *trans,
 
 	path->locks_want = new_locks_want;
 
-	return btree_path_get_locks(trans, path, true, f);
+	bool ret = btree_path_get_locks(trans, path, true, f);
+	bch2_trans_verify_locks(trans);
+	return ret;
 }
 
 bool __bch2_btree_path_upgrade(struct btree_trans *trans,
@@ -666,8 +641,9 @@ bool __bch2_btree_path_upgrade(struct btree_trans *trans,
 			       unsigned new_locks_want,
 			       struct get_locks_fail *f)
 {
-	if (bch2_btree_path_upgrade_noupgrade_sibs(trans, path, new_locks_want, f))
-		return true;
+	bool ret = bch2_btree_path_upgrade_noupgrade_sibs(trans, path, new_locks_want, f);
+	if (ret)
+		goto out;
 
 	/*
 	 * XXX: this is ugly - we'd prefer to not be mucking with other
@@ -701,8 +677,9 @@ bool __bch2_btree_path_upgrade(struct btree_trans *trans,
 				btree_path_get_locks(trans, linked, true, NULL);
 			}
 	}
-
-	return false;
+out:
+	bch2_trans_verify_locks(trans);
+	return ret;
 }
 
 void __bch2_btree_path_downgrade(struct btree_trans *trans,
@@ -747,85 +724,102 @@ void bch2_trans_downgrade(struct btree_trans *trans)
 		return;
 
 	trans_for_each_path(trans, path, i)
-		bch2_btree_path_downgrade(trans, path);
+		if (path->ref)
+			bch2_btree_path_downgrade(trans, path);
 }
 
-int bch2_trans_relock(struct btree_trans *trans)
+static inline void __bch2_trans_unlock(struct btree_trans *trans)
 {
 	struct btree_path *path;
 	unsigned i;
 
-	if (unlikely(trans->restarted))
-		return -((int) trans->restarted);
+	trans_for_each_path(trans, path, i)
+		__bch2_btree_path_unlock(trans, path);
+}
 
-	trans_for_each_path(trans, path, i) {
-		struct get_locks_fail f;
+static noinline __cold int bch2_trans_relock_fail(struct btree_trans *trans, struct btree_path *path,
+						  struct get_locks_fail *f, bool trace)
+{
+	if (!trace)
+		goto out;
 
-		if (path->should_be_locked &&
-		    !btree_path_get_locks(trans, path, false, &f)) {
-			if (trace_trans_restart_relock_enabled()) {
-				struct printbuf buf = PRINTBUF;
-
-				bch2_bpos_to_text(&buf, path->pos);
-				prt_printf(&buf, " l=%u seq=%u node seq=",
-					   f.l, path->l[f.l].lock_seq);
-				if (IS_ERR_OR_NULL(f.b)) {
-					prt_str(&buf, bch2_err_str(PTR_ERR(f.b)));
-				} else {
-					prt_printf(&buf, "%u", f.b->c.lock.seq);
-
-					struct six_lock_count c =
-						bch2_btree_node_lock_counts(trans, NULL, &f.b->c, f.l);
-					prt_printf(&buf, " self locked %u.%u.%u", c.n[0], c.n[1], c.n[2]);
-
-					c = six_lock_counts(&f.b->c.lock);
-					prt_printf(&buf, " total locked %u.%u.%u", c.n[0], c.n[1], c.n[2]);
-				}
+	if (trace_trans_restart_relock_enabled()) {
+		struct printbuf buf = PRINTBUF;
 
-				trace_trans_restart_relock(trans, _RET_IP_, buf.buf);
-				printbuf_exit(&buf);
-			}
+		bch2_bpos_to_text(&buf, path->pos);
+		prt_printf(&buf, " l=%u seq=%u node seq=", f->l, path->l[f->l].lock_seq);
+		if (IS_ERR_OR_NULL(f->b)) {
+			prt_str(&buf, bch2_err_str(PTR_ERR(f->b)));
+		} else {
+			prt_printf(&buf, "%u", f->b->c.lock.seq);
 
-			count_event(trans->c, trans_restart_relock);
-			return btree_trans_restart(trans, BCH_ERR_transaction_restart_relock);
+			struct six_lock_count c =
+				bch2_btree_node_lock_counts(trans, NULL, &f->b->c, f->l);
+			prt_printf(&buf, " self locked %u.%u.%u", c.n[0], c.n[1], c.n[2]);
+
+			c = six_lock_counts(&f->b->c.lock);
+			prt_printf(&buf, " total locked %u.%u.%u", c.n[0], c.n[1], c.n[2]);
 		}
+
+		trace_trans_restart_relock(trans, _RET_IP_, buf.buf);
+		printbuf_exit(&buf);
 	}
 
-	return 0;
+	count_event(trans->c, trans_restart_relock);
+out:
+	__bch2_trans_unlock(trans);
+	bch2_trans_verify_locks(trans);
+	return btree_trans_restart(trans, BCH_ERR_transaction_restart_relock);
 }
 
-int bch2_trans_relock_notrace(struct btree_trans *trans)
+static inline int __bch2_trans_relock(struct btree_trans *trans, bool trace)
 {
-	struct btree_path *path;
-	unsigned i;
+	bch2_trans_verify_locks(trans);
 
 	if (unlikely(trans->restarted))
 		return -((int) trans->restarted);
+	if (unlikely(trans->locked))
+		goto out;
+
+	struct btree_path *path;
+	unsigned i;
+
+	trans_for_each_path(trans, path, i) {
+		struct get_locks_fail f;
 
-	trans_for_each_path(trans, path, i)
 		if (path->should_be_locked &&
-		    !bch2_btree_path_relock_norestart(trans, path)) {
-			return btree_trans_restart(trans, BCH_ERR_transaction_restart_relock);
-		}
+		    !btree_path_get_locks(trans, path, false, &f))
+			return bch2_trans_relock_fail(trans, path, &f, trace);
+	}
+
+	trans_set_locked(trans);
+out:
+	bch2_trans_verify_locks(trans);
 	return 0;
 }
 
+int bch2_trans_relock(struct btree_trans *trans)
+{
+	return __bch2_trans_relock(trans, true);
+}
+
+int bch2_trans_relock_notrace(struct btree_trans *trans)
+{
+	return __bch2_trans_relock(trans, false);
+}
+
 void bch2_trans_unlock_noassert(struct btree_trans *trans)
 {
-	struct btree_path *path;
-	unsigned i;
+	__bch2_trans_unlock(trans);
 
-	trans_for_each_path(trans, path, i)
-		__bch2_btree_path_unlock(trans, path);
+	trans_set_unlocked(trans);
 }
 
 void bch2_trans_unlock(struct btree_trans *trans)
 {
-	struct btree_path *path;
-	unsigned i;
+	__bch2_trans_unlock(trans);
 
-	trans_for_each_path(trans, path, i)
-		__bch2_btree_path_unlock(trans, path);
+	trans_set_unlocked(trans);
 }
 
 void bch2_trans_unlock_long(struct btree_trans *trans)
@@ -834,17 +828,6 @@ void bch2_trans_unlock_long(struct btree_trans *trans)
 	bch2_trans_srcu_unlock(trans);
 }
 
-bool bch2_trans_locked(struct btree_trans *trans)
-{
-	struct btree_path *path;
-	unsigned i;
-
-	trans_for_each_path(trans, path, i)
-		if (path->nodes_locked)
-			return true;
-	return false;
-}
-
 int __bch2_trans_mutex_lock(struct btree_trans *trans,
 			    struct mutex *lock)
 {
@@ -861,15 +844,19 @@ int __bch2_trans_mutex_lock(struct btree_trans *trans,
 
 void bch2_btree_path_verify_locks(struct btree_path *path)
 {
-	unsigned l;
+	/*
+	 * A path may be uptodate and yet have nothing locked if and only if
+	 * there is no node at path->level, which generally means we were
+	 * iterating over all nodes and got to the end of the btree
+	 */
+	BUG_ON(path->uptodate == BTREE_ITER_UPTODATE &&
+	       btree_path_node(path, path->level) &&
+	       !path->nodes_locked);
 
-	if (!path->nodes_locked) {
-		BUG_ON(path->uptodate == BTREE_ITER_UPTODATE &&
-		       btree_path_node(path, path->level));
+	if (!path->nodes_locked)
 		return;
-	}
 
-	for (l = 0; l < BTREE_MAX_DEPTH; l++) {
+	for (unsigned l = 0; l < BTREE_MAX_DEPTH; l++) {
 		int want = btree_lock_want(path, l);
 		int have = btree_node_locked_type(path, l);
 
@@ -882,8 +869,24 @@ void bch2_btree_path_verify_locks(struct btree_path *path)
 	}
 }
 
+static bool bch2_trans_locked(struct btree_trans *trans)
+{
+	struct btree_path *path;
+	unsigned i;
+
+	trans_for_each_path(trans, path, i)
+		if (path->nodes_locked)
+			return true;
+	return false;
+}
+
 void bch2_trans_verify_locks(struct btree_trans *trans)
 {
+	if (!trans->locked) {
+		BUG_ON(bch2_trans_locked(trans));
+		return;
+	}
+
 	struct btree_path *path;
 	unsigned i;
 
diff --git a/fs/bcachefs/btree_locking.h b/fs/bcachefs/btree_locking.h
index 4bd72c855d..75a6274c7d 100644
--- a/fs/bcachefs/btree_locking.h
+++ b/fs/bcachefs/btree_locking.h
@@ -193,6 +193,28 @@ int bch2_six_check_for_deadlock(struct six_lock *lock, void *p);
 
 /* lock: */
 
+static inline void trans_set_locked(struct btree_trans *trans)
+{
+	if (!trans->locked) {
+		trans->locked = true;
+		trans->last_unlock_ip = 0;
+
+		trans->pf_memalloc_nofs = (current->flags & PF_MEMALLOC_NOFS) != 0;
+		current->flags |= PF_MEMALLOC_NOFS;
+	}
+}
+
+static inline void trans_set_unlocked(struct btree_trans *trans)
+{
+	if (trans->locked) {
+		trans->locked = false;
+		trans->last_unlock_ip = _RET_IP_;
+
+		if (!trans->pf_memalloc_nofs)
+			current->flags &= ~PF_MEMALLOC_NOFS;
+	}
+}
+
 static inline int __btree_node_lock_nopath(struct btree_trans *trans,
 					 struct btree_bkey_cached_common *b,
 					 enum six_lock_type type,
@@ -364,14 +386,14 @@ static inline int bch2_btree_path_upgrade(struct btree_trans *trans,
 					  struct btree_path *path,
 					  unsigned new_locks_want)
 {
-	struct get_locks_fail f;
+	struct get_locks_fail f = {};
 	unsigned old_locks_want = path->locks_want;
 
 	new_locks_want = min(new_locks_want, BTREE_MAX_DEPTH);
 
 	if (path->locks_want < new_locks_want
 	    ? __bch2_btree_path_upgrade(trans, path, new_locks_want, &f)
-	    : path->uptodate == BTREE_ITER_UPTODATE)
+	    : path->nodes_locked)
 		return 0;
 
 	trace_and_count(trans->c, trans_restart_upgrade, trans, _THIS_IP_, path,
diff --git a/fs/bcachefs/btree_node_scan.c b/fs/bcachefs/btree_node_scan.c
new file mode 100644
index 0000000000..2cb0442f6c
--- /dev/null
+++ b/fs/bcachefs/btree_node_scan.c
@@ -0,0 +1,525 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include "bcachefs.h"
+#include "btree_cache.h"
+#include "btree_io.h"
+#include "btree_journal_iter.h"
+#include "btree_node_scan.h"
+#include "btree_update_interior.h"
+#include "buckets.h"
+#include "error.h"
+#include "journal_io.h"
+#include "recovery_passes.h"
+
+#include <linux/kthread.h>
+#include <linux/sort.h>
+
+struct find_btree_nodes_worker {
+	struct closure		*cl;
+	struct find_btree_nodes	*f;
+	struct bch_dev		*ca;
+};
+
+static void found_btree_node_to_text(struct printbuf *out, struct bch_fs *c, const struct found_btree_node *n)
+{
+	prt_printf(out, "%s l=%u seq=%u cookie=%llx ", bch2_btree_id_str(n->btree_id), n->level, n->seq, n->cookie);
+	bch2_bpos_to_text(out, n->min_key);
+	prt_str(out, "-");
+	bch2_bpos_to_text(out, n->max_key);
+
+	if (n->range_updated)
+		prt_str(out, " range updated");
+	if (n->overwritten)
+		prt_str(out, " overwritten");
+
+	for (unsigned i = 0; i < n->nr_ptrs; i++) {
+		prt_char(out, ' ');
+		bch2_extent_ptr_to_text(out, c, n->ptrs + i);
+	}
+}
+
+static void found_btree_nodes_to_text(struct printbuf *out, struct bch_fs *c, found_btree_nodes nodes)
+{
+	printbuf_indent_add(out, 2);
+	darray_for_each(nodes, i) {
+		found_btree_node_to_text(out, c, i);
+		prt_newline(out);
+	}
+	printbuf_indent_sub(out, 2);
+}
+
+static void found_btree_node_to_key(struct bkey_i *k, const struct found_btree_node *f)
+{
+	struct bkey_i_btree_ptr_v2 *bp = bkey_btree_ptr_v2_init(k);
+
+	set_bkey_val_u64s(&bp->k, sizeof(struct bch_btree_ptr_v2) / sizeof(u64) + f->nr_ptrs);
+	bp->k.p			= f->max_key;
+	bp->v.seq		= cpu_to_le64(f->cookie);
+	bp->v.sectors_written	= 0;
+	bp->v.flags		= 0;
+	bp->v.sectors_written	= cpu_to_le16(f->sectors_written);
+	bp->v.min_key		= f->min_key;
+	SET_BTREE_PTR_RANGE_UPDATED(&bp->v, f->range_updated);
+	memcpy(bp->v.start, f->ptrs, sizeof(struct bch_extent_ptr) * f->nr_ptrs);
+}
+
+static bool found_btree_node_is_readable(struct btree_trans *trans,
+					 struct found_btree_node *f)
+{
+	struct { __BKEY_PADDED(k, BKEY_BTREE_PTR_VAL_U64s_MAX); } k;
+
+	found_btree_node_to_key(&k.k, f);
+
+	struct btree *b = bch2_btree_node_get_noiter(trans, &k.k, f->btree_id, f->level, false);
+	bool ret = !IS_ERR_OR_NULL(b);
+	if (!ret)
+		return ret;
+
+	f->sectors_written = b->written;
+	six_unlock_read(&b->c.lock);
+
+	/*
+	 * We might update this node's range; if that happens, we need the node
+	 * to be re-read so the read path can trim keys that are no longer in
+	 * this node
+	 */
+	if (b != btree_node_root(trans->c, b))
+		bch2_btree_node_evict(trans, &k.k);
+	return ret;
+}
+
+static int found_btree_node_cmp_cookie(const void *_l, const void *_r)
+{
+	const struct found_btree_node *l = _l;
+	const struct found_btree_node *r = _r;
+
+	return  cmp_int(l->btree_id,	r->btree_id) ?:
+		cmp_int(l->level,	r->level) ?:
+		cmp_int(l->cookie,	r->cookie);
+}
+
+/*
+ * Given two found btree nodes, if their sequence numbers are equal, take the
+ * one that's readable:
+ */
+static int found_btree_node_cmp_time(const struct found_btree_node *l,
+				     const struct found_btree_node *r)
+{
+	return cmp_int(l->seq, r->seq);
+}
+
+static int found_btree_node_cmp_pos(const void *_l, const void *_r)
+{
+	const struct found_btree_node *l = _l;
+	const struct found_btree_node *r = _r;
+
+	return  cmp_int(l->btree_id,	r->btree_id) ?:
+	       -cmp_int(l->level,	r->level) ?:
+		bpos_cmp(l->min_key,	r->min_key) ?:
+	       -found_btree_node_cmp_time(l, r);
+}
+
+static void try_read_btree_node(struct find_btree_nodes *f, struct bch_dev *ca,
+				struct bio *bio, struct btree_node *bn, u64 offset)
+{
+	struct bch_fs *c = container_of(f, struct bch_fs, found_btree_nodes);
+
+	bio_reset(bio, ca->disk_sb.bdev, REQ_OP_READ);
+	bio->bi_iter.bi_sector	= offset;
+	bch2_bio_map(bio, bn, PAGE_SIZE);
+
+	submit_bio_wait(bio);
+	if (bch2_dev_io_err_on(bio->bi_status, ca, BCH_MEMBER_ERROR_read,
+			       "IO error in try_read_btree_node() at %llu: %s",
+			       offset, bch2_blk_status_to_str(bio->bi_status)))
+		return;
+
+	if (le64_to_cpu(bn->magic) != bset_magic(c))
+		return;
+
+	if (bch2_csum_type_is_encryption(BSET_CSUM_TYPE(&bn->keys))) {
+		struct nonce nonce = btree_nonce(&bn->keys, 0);
+		unsigned bytes = (void *) &bn->keys - (void *) &bn->flags;
+
+		bch2_encrypt(c, BSET_CSUM_TYPE(&bn->keys), nonce, &bn->flags, bytes);
+	}
+
+	if (btree_id_is_alloc(BTREE_NODE_ID(bn)))
+		return;
+
+	if (BTREE_NODE_LEVEL(bn) >= BTREE_MAX_DEPTH)
+		return;
+
+	rcu_read_lock();
+	struct found_btree_node n = {
+		.btree_id	= BTREE_NODE_ID(bn),
+		.level		= BTREE_NODE_LEVEL(bn),
+		.seq		= BTREE_NODE_SEQ(bn),
+		.cookie		= le64_to_cpu(bn->keys.seq),
+		.min_key	= bn->min_key,
+		.max_key	= bn->max_key,
+		.nr_ptrs	= 1,
+		.ptrs[0].type	= 1 << BCH_EXTENT_ENTRY_ptr,
+		.ptrs[0].offset	= offset,
+		.ptrs[0].dev	= ca->dev_idx,
+		.ptrs[0].gen	= *bucket_gen(ca, sector_to_bucket(ca, offset)),
+	};
+	rcu_read_unlock();
+
+	if (bch2_trans_run(c, found_btree_node_is_readable(trans, &n))) {
+		mutex_lock(&f->lock);
+		if (BSET_BIG_ENDIAN(&bn->keys) != CPU_BIG_ENDIAN) {
+			bch_err(c, "try_read_btree_node() can't handle endian conversion");
+			f->ret = -EINVAL;
+			goto unlock;
+		}
+
+		if (darray_push(&f->nodes, n))
+			f->ret = -ENOMEM;
+unlock:
+		mutex_unlock(&f->lock);
+	}
+}
+
+static int read_btree_nodes_worker(void *p)
+{
+	struct find_btree_nodes_worker *w = p;
+	struct bch_fs *c = container_of(w->f, struct bch_fs, found_btree_nodes);
+	struct bch_dev *ca = w->ca;
+	void *buf = (void *) __get_free_page(GFP_KERNEL);
+	struct bio *bio = bio_alloc(NULL, 1, 0, GFP_KERNEL);
+	unsigned long last_print = jiffies;
+
+	if (!buf || !bio) {
+		bch_err(c, "read_btree_nodes_worker: error allocating bio/buf");
+		w->f->ret = -ENOMEM;
+		goto err;
+	}
+
+	for (u64 bucket = ca->mi.first_bucket; bucket < ca->mi.nbuckets; bucket++)
+		for (unsigned bucket_offset = 0;
+		     bucket_offset + btree_sectors(c) <= ca->mi.bucket_size;
+		     bucket_offset += btree_sectors(c)) {
+			if (time_after(jiffies, last_print + HZ * 30)) {
+				u64 cur_sector = bucket * ca->mi.bucket_size + bucket_offset;
+				u64 end_sector = ca->mi.nbuckets * ca->mi.bucket_size;
+
+				bch_info(ca, "%s: %2u%% done", __func__,
+					 (unsigned) div64_u64(cur_sector * 100, end_sector));
+				last_print = jiffies;
+			}
+
+			u64 sector = bucket * ca->mi.bucket_size + bucket_offset;
+
+			if (c->sb.version_upgrade_complete >= bcachefs_metadata_version_mi_btree_bitmap &&
+			    !bch2_dev_btree_bitmap_marked_sectors(ca, sector, btree_sectors(c)))
+				continue;
+
+			try_read_btree_node(w->f, ca, bio, buf, sector);
+		}
+err:
+	bio_put(bio);
+	free_page((unsigned long) buf);
+	percpu_ref_get(&ca->io_ref);
+	closure_put(w->cl);
+	kfree(w);
+	return 0;
+}
+
+static int read_btree_nodes(struct find_btree_nodes *f)
+{
+	struct bch_fs *c = container_of(f, struct bch_fs, found_btree_nodes);
+	struct closure cl;
+	int ret = 0;
+
+	closure_init_stack(&cl);
+
+	for_each_online_member(c, ca) {
+		if (!(ca->mi.data_allowed & BIT(BCH_DATA_btree)))
+			continue;
+
+		struct find_btree_nodes_worker *w = kmalloc(sizeof(*w), GFP_KERNEL);
+		struct task_struct *t;
+
+		if (!w) {
+			percpu_ref_put(&ca->io_ref);
+			ret = -ENOMEM;
+			goto err;
+		}
+
+		percpu_ref_get(&ca->io_ref);
+		closure_get(&cl);
+		w->cl		= &cl;
+		w->f		= f;
+		w->ca		= ca;
+
+		t = kthread_run(read_btree_nodes_worker, w, "read_btree_nodes/%s", ca->name);
+		ret = IS_ERR_OR_NULL(t);
+		if (ret) {
+			percpu_ref_put(&ca->io_ref);
+			closure_put(&cl);
+			f->ret = ret;
+			bch_err(c, "error starting kthread: %i", ret);
+			break;
+		}
+	}
+err:
+	closure_sync(&cl);
+	return f->ret ?: ret;
+}
+
+static void bubble_up(struct found_btree_node *n, struct found_btree_node *end)
+{
+	while (n + 1 < end &&
+	       found_btree_node_cmp_pos(n, n + 1) > 0) {
+		swap(n[0], n[1]);
+		n++;
+	}
+}
+
+static int handle_overwrites(struct bch_fs *c,
+			     struct found_btree_node *start,
+			     struct found_btree_node *end)
+{
+	struct found_btree_node *n;
+again:
+	for (n = start + 1;
+	     n < end &&
+	     n->btree_id	== start->btree_id &&
+	     n->level		== start->level &&
+	     bpos_lt(n->min_key, start->max_key);
+	     n++)  {
+		int cmp = found_btree_node_cmp_time(start, n);
+
+		if (cmp > 0) {
+			if (bpos_cmp(start->max_key, n->max_key) >= 0)
+				n->overwritten = true;
+			else {
+				n->range_updated = true;
+				n->min_key = bpos_successor(start->max_key);
+				n->range_updated = true;
+				bubble_up(n, end);
+				goto again;
+			}
+		} else if (cmp < 0) {
+			BUG_ON(bpos_cmp(n->min_key, start->min_key) <= 0);
+
+			start->max_key = bpos_predecessor(n->min_key);
+			start->range_updated = true;
+		} else if (n->level) {
+			n->overwritten = true;
+		} else {
+			struct printbuf buf = PRINTBUF;
+
+			prt_str(&buf, "overlapping btree nodes with same seq! halting\n  ");
+			found_btree_node_to_text(&buf, c, start);
+			prt_str(&buf, "\n  ");
+			found_btree_node_to_text(&buf, c, n);
+			bch_err(c, "%s", buf.buf);
+			printbuf_exit(&buf);
+			return -BCH_ERR_fsck_repair_unimplemented;
+		}
+	}
+
+	return 0;
+}
+
+int bch2_scan_for_btree_nodes(struct bch_fs *c)
+{
+	struct find_btree_nodes *f = &c->found_btree_nodes;
+	struct printbuf buf = PRINTBUF;
+	size_t dst;
+	int ret = 0;
+
+	if (f->nodes.nr)
+		return 0;
+
+	mutex_init(&f->lock);
+
+	ret = read_btree_nodes(f);
+	if (ret)
+		return ret;
+
+	if (!f->nodes.nr) {
+		bch_err(c, "%s: no btree nodes found", __func__);
+		ret = -EINVAL;
+		goto err;
+	}
+
+	if (0 && c->opts.verbose) {
+		printbuf_reset(&buf);
+		prt_printf(&buf, "%s: nodes found:\n", __func__);
+		found_btree_nodes_to_text(&buf, c, f->nodes);
+		bch2_print_string_as_lines(KERN_INFO, buf.buf);
+	}
+
+	sort(f->nodes.data, f->nodes.nr, sizeof(f->nodes.data[0]), found_btree_node_cmp_cookie, NULL);
+
+	dst = 0;
+	darray_for_each(f->nodes, i) {
+		struct found_btree_node *prev = dst ? f->nodes.data + dst - 1 : NULL;
+
+		if (prev &&
+		    prev->cookie == i->cookie) {
+			if (prev->nr_ptrs == ARRAY_SIZE(prev->ptrs)) {
+				bch_err(c, "%s: found too many replicas for btree node", __func__);
+				ret = -EINVAL;
+				goto err;
+			}
+			prev->ptrs[prev->nr_ptrs++] = i->ptrs[0];
+		} else {
+			f->nodes.data[dst++] = *i;
+		}
+	}
+	f->nodes.nr = dst;
+
+	sort(f->nodes.data, f->nodes.nr, sizeof(f->nodes.data[0]), found_btree_node_cmp_pos, NULL);
+
+	if (0 && c->opts.verbose) {
+		printbuf_reset(&buf);
+		prt_printf(&buf, "%s: nodes after merging replicas:\n", __func__);
+		found_btree_nodes_to_text(&buf, c, f->nodes);
+		bch2_print_string_as_lines(KERN_INFO, buf.buf);
+	}
+
+	dst = 0;
+	darray_for_each(f->nodes, i) {
+		if (i->overwritten)
+			continue;
+
+		ret = handle_overwrites(c, i, &darray_top(f->nodes));
+		if (ret)
+			goto err;
+
+		BUG_ON(i->overwritten);
+		f->nodes.data[dst++] = *i;
+	}
+	f->nodes.nr = dst;
+
+	if (c->opts.verbose) {
+		printbuf_reset(&buf);
+		prt_printf(&buf, "%s: nodes found after overwrites:\n", __func__);
+		found_btree_nodes_to_text(&buf, c, f->nodes);
+		bch2_print_string_as_lines(KERN_INFO, buf.buf);
+	}
+
+	eytzinger0_sort(f->nodes.data, f->nodes.nr, sizeof(f->nodes.data[0]), found_btree_node_cmp_pos, NULL);
+err:
+	printbuf_exit(&buf);
+	return ret;
+}
+
+static int found_btree_node_range_start_cmp(const void *_l, const void *_r)
+{
+	const struct found_btree_node *l = _l;
+	const struct found_btree_node *r = _r;
+
+	return  cmp_int(l->btree_id,	r->btree_id) ?:
+	       -cmp_int(l->level,	r->level) ?:
+		bpos_cmp(l->max_key,	r->min_key);
+}
+
+#define for_each_found_btree_node_in_range(_f, _search, _idx)				\
+	for (size_t _idx = eytzinger0_find_gt((_f)->nodes.data, (_f)->nodes.nr,		\
+					sizeof((_f)->nodes.data[0]),			\
+					found_btree_node_range_start_cmp, &search);	\
+	     _idx < (_f)->nodes.nr &&							\
+	     (_f)->nodes.data[_idx].btree_id == _search.btree_id &&			\
+	     (_f)->nodes.data[_idx].level == _search.level &&				\
+	     bpos_lt((_f)->nodes.data[_idx].min_key, _search.max_key);			\
+	     _idx = eytzinger0_next(_idx, (_f)->nodes.nr))
+
+bool bch2_btree_node_is_stale(struct bch_fs *c, struct btree *b)
+{
+	struct find_btree_nodes *f = &c->found_btree_nodes;
+
+	struct found_btree_node search = {
+		.btree_id	= b->c.btree_id,
+		.level		= b->c.level,
+		.min_key	= b->data->min_key,
+		.max_key	= b->key.k.p,
+	};
+
+	for_each_found_btree_node_in_range(f, search, idx)
+		if (f->nodes.data[idx].seq > BTREE_NODE_SEQ(b->data))
+			return true;
+	return false;
+}
+
+bool bch2_btree_has_scanned_nodes(struct bch_fs *c, enum btree_id btree)
+{
+	struct found_btree_node search = {
+		.btree_id	= btree,
+		.level		= 0,
+		.min_key	= POS_MIN,
+		.max_key	= SPOS_MAX,
+	};
+
+	for_each_found_btree_node_in_range(&c->found_btree_nodes, search, idx)
+		return true;
+	return false;
+}
+
+int bch2_get_scanned_nodes(struct bch_fs *c, enum btree_id btree,
+			   unsigned level, struct bpos node_min, struct bpos node_max)
+{
+	if (btree_id_is_alloc(btree))
+		return 0;
+
+	struct find_btree_nodes *f = &c->found_btree_nodes;
+
+	int ret = bch2_run_explicit_recovery_pass(c, BCH_RECOVERY_PASS_scan_for_btree_nodes);
+	if (ret)
+		return ret;
+
+	if (c->opts.verbose) {
+		struct printbuf buf = PRINTBUF;
+
+		prt_printf(&buf, "recovering %s l=%u ", bch2_btree_id_str(btree), level);
+		bch2_bpos_to_text(&buf, node_min);
+		prt_str(&buf, " - ");
+		bch2_bpos_to_text(&buf, node_max);
+
+		bch_info(c, "%s(): %s", __func__, buf.buf);
+		printbuf_exit(&buf);
+	}
+
+	struct found_btree_node search = {
+		.btree_id	= btree,
+		.level		= level,
+		.min_key	= node_min,
+		.max_key	= node_max,
+	};
+
+	for_each_found_btree_node_in_range(f, search, idx) {
+		struct found_btree_node n = f->nodes.data[idx];
+
+		n.range_updated |= bpos_lt(n.min_key, node_min);
+		n.min_key = bpos_max(n.min_key, node_min);
+
+		n.range_updated |= bpos_gt(n.max_key, node_max);
+		n.max_key = bpos_min(n.max_key, node_max);
+
+		struct { __BKEY_PADDED(k, BKEY_BTREE_PTR_VAL_U64s_MAX); } tmp;
+
+		found_btree_node_to_key(&tmp.k, &n);
+
+		struct printbuf buf = PRINTBUF;
+		bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&tmp.k));
+		bch_verbose(c, "%s(): recovering %s", __func__, buf.buf);
+		printbuf_exit(&buf);
+
+		BUG_ON(bch2_bkey_invalid(c, bkey_i_to_s_c(&tmp.k), BKEY_TYPE_btree, 0, NULL));
+
+		ret = bch2_journal_key_insert(c, btree, level + 1, &tmp.k);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+void bch2_find_btree_nodes_exit(struct find_btree_nodes *f)
+{
+	darray_exit(&f->nodes);
+}
diff --git a/fs/bcachefs/btree_node_scan.h b/fs/bcachefs/btree_node_scan.h
new file mode 100644
index 0000000000..08687b2097
--- /dev/null
+++ b/fs/bcachefs/btree_node_scan.h
@@ -0,0 +1,11 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _BCACHEFS_BTREE_NODE_SCAN_H
+#define _BCACHEFS_BTREE_NODE_SCAN_H
+
+int bch2_scan_for_btree_nodes(struct bch_fs *);
+bool bch2_btree_node_is_stale(struct bch_fs *, struct btree *);
+bool bch2_btree_has_scanned_nodes(struct bch_fs *, enum btree_id);
+int bch2_get_scanned_nodes(struct bch_fs *, enum btree_id, unsigned, struct bpos, struct bpos);
+void bch2_find_btree_nodes_exit(struct find_btree_nodes *);
+
+#endif /* _BCACHEFS_BTREE_NODE_SCAN_H */
diff --git a/fs/bcachefs/btree_node_scan_types.h b/fs/bcachefs/btree_node_scan_types.h
new file mode 100644
index 0000000000..5cfaeb5ac8
--- /dev/null
+++ b/fs/bcachefs/btree_node_scan_types.h
@@ -0,0 +1,31 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _BCACHEFS_BTREE_NODE_SCAN_TYPES_H
+#define _BCACHEFS_BTREE_NODE_SCAN_TYPES_H
+
+#include "darray.h"
+
+struct found_btree_node {
+	bool			range_updated:1;
+	bool			overwritten:1;
+	u8			btree_id;
+	u8			level;
+	unsigned		sectors_written;
+	u32			seq;
+	u64			cookie;
+
+	struct bpos		min_key;
+	struct bpos		max_key;
+
+	unsigned		nr_ptrs;
+	struct bch_extent_ptr	ptrs[BCH_REPLICAS_MAX];
+};
+
+typedef DARRAY(struct found_btree_node)	found_btree_nodes;
+
+struct find_btree_nodes {
+	int			ret;
+	struct mutex		lock;
+	found_btree_nodes	nodes;
+};
+
+#endif /* _BCACHEFS_BTREE_NODE_SCAN_TYPES_H */
diff --git a/fs/bcachefs/btree_trans_commit.c b/fs/bcachefs/btree_trans_commit.c
index 30d69a6d13..74e1ff2256 100644
--- a/fs/bcachefs/btree_trans_commit.c
+++ b/fs/bcachefs/btree_trans_commit.c
@@ -1,6 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0
 
 #include "bcachefs.h"
+#include "alloc_foreground.h"
 #include "btree_gc.h"
 #include "btree_io.h"
 #include "btree_iter.h"
@@ -19,6 +20,26 @@
 
 #include <linux/prefetch.h>
 
+static const char * const trans_commit_flags_strs[] = {
+#define x(n, ...) #n,
+	BCH_TRANS_COMMIT_FLAGS()
+#undef x
+	NULL
+};
+
+void bch2_trans_commit_flags_to_text(struct printbuf *out, enum bch_trans_commit_flags flags)
+{
+	enum bch_watermark watermark = flags & BCH_WATERMARK_MASK;
+
+	prt_printf(out, "watermark=%s", bch2_watermarks[watermark]);
+
+	flags >>= BCH_WATERMARK_BITS;
+	if (flags) {
+		prt_char(out, ' ');
+		bch2_prt_bitflags(out, trans_commit_flags_strs, flags);
+	}
+}
+
 static void verify_update_old_key(struct btree_trans *trans, struct btree_insert_entry *i)
 {
 #ifdef CONFIG_BCACHEFS_DEBUG
@@ -315,10 +336,10 @@ static inline void btree_insert_entry_checks(struct btree_trans *trans,
 	BUG_ON(i->btree_id	!= path->btree_id);
 	EBUG_ON(!i->level &&
 		btree_type_has_snapshots(i->btree_id) &&
-		!(i->flags & BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE) &&
-		test_bit(JOURNAL_REPLAY_DONE, &trans->c->journal.flags) &&
+		!(i->flags & BTREE_UPDATE_internal_snapshot_node) &&
+		test_bit(JOURNAL_replay_done, &trans->c->journal.flags) &&
 		i->k->k.p.snapshot &&
-		bch2_snapshot_is_internal_node(trans->c, i->k->k.p.snapshot));
+		bch2_snapshot_is_internal_node(trans->c, i->k->k.p.snapshot) > 0);
 }
 
 static __always_inline int bch2_trans_journal_res_get(struct btree_trans *trans,
@@ -397,12 +418,13 @@ static int btree_key_can_insert_cached(struct btree_trans *trans, unsigned flags
 	struct bkey_cached *ck = (void *) path->l[0].b;
 	unsigned new_u64s;
 	struct bkey_i *new_k;
+	unsigned watermark = flags & BCH_WATERMARK_MASK;
 
 	EBUG_ON(path->level);
 
-	if (!test_bit(BKEY_CACHED_DIRTY, &ck->flags) &&
-	    bch2_btree_key_cache_must_wait(c) &&
-	    !(flags & BCH_TRANS_COMMIT_journal_reclaim))
+	if (watermark < BCH_WATERMARK_reclaim &&
+	    !test_bit(BKEY_CACHED_DIRTY, &ck->flags) &&
+	    bch2_btree_key_cache_must_wait(c))
 		return -BCH_ERR_btree_insert_need_journal_reclaim;
 
 	/*
@@ -442,13 +464,13 @@ static int run_one_mem_trigger(struct btree_trans *trans,
 
 	verify_update_old_key(trans, i);
 
-	if (unlikely(flags & BTREE_TRIGGER_NORUN))
+	if (unlikely(flags & BTREE_TRIGGER_norun))
 		return 0;
 
 	if (old_ops->trigger == new_ops->trigger) {
 		ret   = bch2_key_trigger(trans, i->btree_id, i->level,
 				old, bkey_i_to_s(new),
-				BTREE_TRIGGER_INSERT|BTREE_TRIGGER_OVERWRITE|flags);
+				BTREE_TRIGGER_insert|BTREE_TRIGGER_overwrite|flags);
 	} else {
 		ret   = bch2_key_trigger_new(trans, i->btree_id, i->level,
 				bkey_i_to_s(new), flags) ?:
@@ -471,11 +493,11 @@ static int run_one_trans_trigger(struct btree_trans *trans, struct btree_insert_
 	struct bkey_s_c old = { &old_k, i->old_v };
 	const struct bkey_ops *old_ops = bch2_bkey_type_ops(old.k->type);
 	const struct bkey_ops *new_ops = bch2_bkey_type_ops(i->k->k.type);
-	unsigned flags = i->flags|BTREE_TRIGGER_TRANSACTIONAL;
+	unsigned flags = i->flags|BTREE_TRIGGER_transactional;
 
 	verify_update_old_key(trans, i);
 
-	if ((i->flags & BTREE_TRIGGER_NORUN) ||
+	if ((i->flags & BTREE_TRIGGER_norun) ||
 	    !(BTREE_NODE_TYPE_HAS_TRANS_TRIGGERS & (1U << i->bkey_type)))
 		return 0;
 
@@ -485,8 +507,8 @@ static int run_one_trans_trigger(struct btree_trans *trans, struct btree_insert_
 		i->overwrite_trigger_run = true;
 		i->insert_trigger_run = true;
 		return bch2_key_trigger(trans, i->btree_id, i->level, old, bkey_i_to_s(i->k),
-					BTREE_TRIGGER_INSERT|
-					BTREE_TRIGGER_OVERWRITE|flags) ?: 1;
+					BTREE_TRIGGER_insert|
+					BTREE_TRIGGER_overwrite|flags) ?: 1;
 	} else if (overwrite && !i->overwrite_trigger_run) {
 		i->overwrite_trigger_run = true;
 		return bch2_key_trigger_old(trans, i->btree_id, i->level, old, flags) ?: 1;
@@ -499,9 +521,8 @@ static int run_one_trans_trigger(struct btree_trans *trans, struct btree_insert_
 }
 
 static int run_btree_triggers(struct btree_trans *trans, enum btree_id btree_id,
-			      struct btree_insert_entry *btree_id_start)
+			      unsigned btree_id_start)
 {
-	struct btree_insert_entry *i;
 	bool trans_trigger_run;
 	int ret, overwrite;
 
@@ -514,13 +535,13 @@ static int run_btree_triggers(struct btree_trans *trans, enum btree_id btree_id,
 		do {
 			trans_trigger_run = false;
 
-			for (i = btree_id_start;
-			     i < trans->updates + trans->nr_updates && i->btree_id <= btree_id;
+			for (unsigned i = btree_id_start;
+			     i < trans->nr_updates && trans->updates[i].btree_id <= btree_id;
 			     i++) {
-				if (i->btree_id != btree_id)
+				if (trans->updates[i].btree_id != btree_id)
 					continue;
 
-				ret = run_one_trans_trigger(trans, i, overwrite);
+				ret = run_one_trans_trigger(trans, trans->updates + i, overwrite);
 				if (ret < 0)
 					return ret;
 				if (ret)
@@ -534,8 +555,7 @@ static int run_btree_triggers(struct btree_trans *trans, enum btree_id btree_id,
 
 static int bch2_trans_commit_run_triggers(struct btree_trans *trans)
 {
-	struct btree_insert_entry *btree_id_start = trans->updates;
-	unsigned btree_id = 0;
+	unsigned btree_id = 0, btree_id_start = 0;
 	int ret = 0;
 
 	/*
@@ -549,8 +569,8 @@ static int bch2_trans_commit_run_triggers(struct btree_trans *trans)
 		if (btree_id == BTREE_ID_alloc)
 			continue;
 
-		while (btree_id_start < trans->updates + trans->nr_updates &&
-		       btree_id_start->btree_id < btree_id)
+		while (btree_id_start < trans->nr_updates &&
+		       trans->updates[btree_id_start].btree_id < btree_id)
 			btree_id_start++;
 
 		ret = run_btree_triggers(trans, btree_id, btree_id_start);
@@ -558,11 +578,13 @@ static int bch2_trans_commit_run_triggers(struct btree_trans *trans)
 			return ret;
 	}
 
-	trans_for_each_update(trans, i) {
+	for (unsigned idx = 0; idx < trans->nr_updates; idx++) {
+		struct btree_insert_entry *i = trans->updates + idx;
+
 		if (i->btree_id > BTREE_ID_alloc)
 			break;
 		if (i->btree_id == BTREE_ID_alloc) {
-			ret = run_btree_triggers(trans, BTREE_ID_alloc, i);
+			ret = run_btree_triggers(trans, BTREE_ID_alloc, idx);
 			if (ret)
 				return ret;
 			break;
@@ -571,7 +593,7 @@ static int bch2_trans_commit_run_triggers(struct btree_trans *trans)
 
 #ifdef CONFIG_BCACHEFS_DEBUG
 	trans_for_each_update(trans, i)
-		BUG_ON(!(i->flags & BTREE_TRIGGER_NORUN) &&
+		BUG_ON(!(i->flags & BTREE_TRIGGER_norun) &&
 		       (BTREE_NODE_TYPE_HAS_TRANS_TRIGGERS & (1U << i->bkey_type)) &&
 		       (!i->insert_trigger_run || !i->overwrite_trigger_run));
 #endif
@@ -589,7 +611,7 @@ static noinline int bch2_trans_commit_run_gc_triggers(struct btree_trans *trans)
 
 		if (btree_node_type_needs_gc(__btree_node_type(i->level, i->btree_id)) &&
 		    gc_visited(trans->c, gc_pos_btree_node(insert_l(trans, i)->b))) {
-			int ret = run_one_mem_trigger(trans, i, i->flags|BTREE_TRIGGER_GC);
+			int ret = run_one_mem_trigger(trans, i, i->flags|BTREE_TRIGGER_gc);
 			if (ret)
 				return ret;
 		}
@@ -608,6 +630,9 @@ bch2_trans_commit_write_locked(struct btree_trans *trans, unsigned flags,
 	unsigned u64s = 0;
 	int ret;
 
+	bch2_trans_verify_not_unlocked(trans);
+	bch2_trans_verify_not_in_restart(trans);
+
 	if (race_fault()) {
 		trace_and_count(c, trans_restart_fault_inject, trans, trace_ip);
 		return btree_trans_restart_nounlock(trans, BCH_ERR_transaction_restart_fault_inject);
@@ -685,7 +710,7 @@ bch2_trans_commit_write_locked(struct btree_trans *trans, unsigned flags,
 
 	trans_for_each_update(trans, i)
 		if (BTREE_NODE_TYPE_HAS_ATOMIC_TRIGGERS & (1U << i->bkey_type)) {
-			ret = run_one_mem_trigger(trans, i, BTREE_TRIGGER_ATOMIC|i->flags);
+			ret = run_one_mem_trigger(trans, i, BTREE_TRIGGER_atomic|i->flags);
 			if (ret)
 				goto fatal_err;
 		}
@@ -704,7 +729,7 @@ bch2_trans_commit_write_locked(struct btree_trans *trans, unsigned flags,
 			if (i->key_cache_already_flushed)
 				continue;
 
-			if (i->flags & BTREE_UPDATE_NOJOURNAL)
+			if (i->flags & BTREE_UPDATE_nojournal)
 				continue;
 
 			verify_update_old_key(trans, i);
@@ -765,16 +790,15 @@ static noinline void bch2_drop_overwrites_from_journal(struct btree_trans *trans
 }
 
 static noinline int bch2_trans_commit_bkey_invalid(struct btree_trans *trans,
-						   enum bkey_invalid_flags flags,
+						   enum bch_validate_flags flags,
 						   struct btree_insert_entry *i,
 						   struct printbuf *err)
 {
 	struct bch_fs *c = trans->c;
 
 	printbuf_reset(err);
-	prt_printf(err, "invalid bkey on insert from %s -> %ps",
+	prt_printf(err, "invalid bkey on insert from %s -> %ps\n",
 		   trans->fn, (void *) i->ip_allocated);
-	prt_newline(err);
 	printbuf_indent_add(err, 2);
 
 	bch2_bkey_val_to_text(err, c, bkey_i_to_s_c(i->k));
@@ -795,8 +819,7 @@ static noinline int bch2_trans_commit_journal_entry_invalid(struct btree_trans *
 	struct bch_fs *c = trans->c;
 	struct printbuf buf = PRINTBUF;
 
-	prt_printf(&buf, "invalid bkey on insert from %s", trans->fn);
-	prt_newline(&buf);
+	prt_printf(&buf, "invalid bkey on insert from %s\n", trans->fn);
 	printbuf_indent_add(&buf, 2);
 
 	bch2_journal_entry_to_text(&buf, c, i);
@@ -826,7 +849,8 @@ static inline int do_bch2_trans_commit(struct btree_trans *trans, unsigned flags
 	struct bch_fs *c = trans->c;
 	int ret = 0, u64s_delta = 0;
 
-	trans_for_each_update(trans, i) {
+	for (unsigned idx = 0; idx < trans->nr_updates; idx++) {
+		struct btree_insert_entry *i = trans->updates + idx;
 		if (i->cached)
 			continue;
 
@@ -887,6 +911,7 @@ int bch2_trans_commit_error(struct btree_trans *trans, unsigned flags,
 			    int ret, unsigned long trace_ip)
 {
 	struct bch_fs *c = trans->c;
+	enum bch_watermark watermark = flags & BCH_WATERMARK_MASK;
 
 	switch (ret) {
 	case -BCH_ERR_btree_insert_btree_node_full:
@@ -905,7 +930,7 @@ int bch2_trans_commit_error(struct btree_trans *trans, unsigned flags,
 		 * flag
 		 */
 		if ((flags & BCH_TRANS_COMMIT_journal_reclaim) &&
-		    (flags & BCH_WATERMARK_MASK) != BCH_WATERMARK_reclaim) {
+		    watermark < BCH_WATERMARK_reclaim) {
 			ret = -BCH_ERR_journal_reclaim_would_deadlock;
 			break;
 		}
@@ -985,6 +1010,9 @@ int __bch2_trans_commit(struct btree_trans *trans, unsigned flags)
 	struct bch_fs *c = trans->c;
 	int ret = 0;
 
+	bch2_trans_verify_not_unlocked(trans);
+	bch2_trans_verify_not_in_restart(trans);
+
 	if (!trans->nr_updates &&
 	    !trans->journal_entries_u64s)
 		goto out_reset;
@@ -997,10 +1025,10 @@ int __bch2_trans_commit(struct btree_trans *trans, unsigned flags)
 
 	trans_for_each_update(trans, i) {
 		struct printbuf buf = PRINTBUF;
-		enum bkey_invalid_flags invalid_flags = 0;
+		enum bch_validate_flags invalid_flags = 0;
 
 		if (!(flags & BCH_TRANS_COMMIT_no_journal_res))
-			invalid_flags |= BKEY_INVALID_WRITE|BKEY_INVALID_COMMIT;
+			invalid_flags |= BCH_VALIDATE_write|BCH_VALIDATE_commit;
 
 		if (unlikely(bch2_bkey_invalid(c, bkey_i_to_s_c(i->k),
 					       i->bkey_type, invalid_flags, &buf)))
@@ -1015,10 +1043,10 @@ int __bch2_trans_commit(struct btree_trans *trans, unsigned flags)
 	for (struct jset_entry *i = trans->journal_entries;
 	     i != (void *) ((u64 *) trans->journal_entries + trans->journal_entries_u64s);
 	     i = vstruct_next(i)) {
-		enum bkey_invalid_flags invalid_flags = 0;
+		enum bch_validate_flags invalid_flags = 0;
 
 		if (!(flags & BCH_TRANS_COMMIT_no_journal_res))
-			invalid_flags |= BKEY_INVALID_WRITE|BKEY_INVALID_COMMIT;
+			invalid_flags |= BCH_VALIDATE_write|BCH_VALIDATE_commit;
 
 		if (unlikely(bch2_journal_entry_validate(c, NULL, i,
 					bcachefs_metadata_version_current,
@@ -1062,7 +1090,7 @@ int __bch2_trans_commit(struct btree_trans *trans, unsigned flags)
 		if (i->key_cache_already_flushed)
 			continue;
 
-		if (i->flags & BTREE_UPDATE_NOJOURNAL)
+		if (i->flags & BTREE_UPDATE_nojournal)
 			continue;
 
 		/* we're going to journal the key being updated: */
@@ -1083,6 +1111,7 @@ int __bch2_trans_commit(struct btree_trans *trans, unsigned flags)
 	}
 retry:
 	errored_at = NULL;
+	bch2_trans_verify_not_unlocked(trans);
 	bch2_trans_verify_not_in_restart(trans);
 	if (likely(!(flags & BCH_TRANS_COMMIT_no_journal_res)))
 		memset(&trans->journal_res, 0, sizeof(trans->journal_res));
diff --git a/fs/bcachefs/btree_types.h b/fs/bcachefs/btree_types.h
index 4a5a64499e..48cb1a7d31 100644
--- a/fs/bcachefs/btree_types.h
+++ b/fs/bcachefs/btree_types.h
@@ -5,6 +5,7 @@
 #include <linux/list.h>
 #include <linux/rhashtable.h>
 
+#include "bbpos_types.h"
 #include "btree_key_cache_types.h"
 #include "buckets_types.h"
 #include "darray.h"
@@ -162,9 +163,21 @@ struct btree_cache {
 	/* Number of elements in live + freeable lists */
 	unsigned		used;
 	unsigned		reserve;
+	unsigned		freed;
+	unsigned		not_freed_lock_intent;
+	unsigned		not_freed_lock_write;
+	unsigned		not_freed_dirty;
+	unsigned		not_freed_read_in_flight;
+	unsigned		not_freed_write_in_flight;
+	unsigned		not_freed_noevict;
+	unsigned		not_freed_write_blocked;
+	unsigned		not_freed_will_make_reachable;
+	unsigned		not_freed_access_bit;
 	atomic_t		dirty;
 	struct shrinker		*shrink;
 
+	unsigned		used_by_btree[BTREE_ID_NR];
+
 	/*
 	 * If we need to allocate memory for a new btree node and that
 	 * allocation fails, we can cannibalize another node in the btree cache
@@ -173,6 +186,11 @@ struct btree_cache {
 	 */
 	struct task_struct	*alloc_lock;
 	struct closure_waitlist	alloc_wait;
+
+	struct bbpos		pinned_nodes_start;
+	struct bbpos		pinned_nodes_end;
+	u64			pinned_nodes_leaf_mask;
+	u64			pinned_nodes_interior_mask;
 };
 
 struct btree_node_iter {
@@ -181,36 +199,89 @@ struct btree_node_iter {
 	} data[MAX_BSETS];
 };
 
+#define BTREE_ITER_FLAGS()			\
+	x(slots)				\
+	x(intent)				\
+	x(prefetch)				\
+	x(is_extents)				\
+	x(not_extents)				\
+	x(cached)				\
+	x(with_key_cache)			\
+	x(with_updates)				\
+	x(with_journal)				\
+	x(snapshot_field)			\
+	x(all_snapshots)			\
+	x(filter_snapshots)			\
+	x(nopreserve)				\
+	x(cached_nofill)			\
+	x(key_cache_fill)			\
+
+#define STR_HASH_FLAGS()			\
+	x(must_create)				\
+	x(must_replace)
+
+#define BTREE_UPDATE_FLAGS()			\
+	x(internal_snapshot_node)		\
+	x(nojournal)				\
+	x(key_cache_reclaim)
+
+
 /*
- * Iterate over all possible positions, synthesizing deleted keys for holes:
- */
-static const __maybe_unused u16 BTREE_ITER_SLOTS		= 1 << 0;
-/*
- * Indicates that intent locks should be taken on leaf nodes, because we expect
- * to be doing updates:
- */
-static const __maybe_unused u16 BTREE_ITER_INTENT		= 1 << 1;
-/*
- * Causes the btree iterator code to prefetch additional btree nodes from disk:
- */
-static const __maybe_unused u16 BTREE_ITER_PREFETCH		= 1 << 2;
-/*
- * Used in bch2_btree_iter_traverse(), to indicate whether we're searching for
- * @pos or the first key strictly greater than @pos
+ * BTREE_TRIGGER_norun - don't run triggers at all
+ *
+ * BTREE_TRIGGER_transactional - we're running transactional triggers as part of
+ * a transaction commit: triggers may generate new updates
+ *
+ * BTREE_TRIGGER_atomic - we're running atomic triggers during a transaction
+ * commit: we have our journal reservation, we're holding btree node write
+ * locks, and we know the transaction is going to commit (returning an error
+ * here is a fatal error, causing us to go emergency read-only)
+ *
+ * BTREE_TRIGGER_gc - we're in gc/fsck: running triggers to recalculate e.g. disk usage
+ *
+ * BTREE_TRIGGER_insert - @new is entering the btree
+ * BTREE_TRIGGER_overwrite - @old is leaving the btree
+ *
+ * BTREE_TRIGGER_bucket_invalidate - signal from bucket invalidate path to alloc
+ * trigger
  */
-static const __maybe_unused u16 BTREE_ITER_IS_EXTENTS		= 1 << 3;
-static const __maybe_unused u16 BTREE_ITER_NOT_EXTENTS		= 1 << 4;
-static const __maybe_unused u16 BTREE_ITER_CACHED		= 1 << 5;
-static const __maybe_unused u16 BTREE_ITER_WITH_KEY_CACHE	= 1 << 6;
-static const __maybe_unused u16 BTREE_ITER_WITH_UPDATES		= 1 << 7;
-static const __maybe_unused u16 BTREE_ITER_WITH_JOURNAL		= 1 << 8;
-static const __maybe_unused u16 __BTREE_ITER_ALL_SNAPSHOTS	= 1 << 9;
-static const __maybe_unused u16 BTREE_ITER_ALL_SNAPSHOTS	= 1 << 10;
-static const __maybe_unused u16 BTREE_ITER_FILTER_SNAPSHOTS	= 1 << 11;
-static const __maybe_unused u16 BTREE_ITER_NOPRESERVE		= 1 << 12;
-static const __maybe_unused u16 BTREE_ITER_CACHED_NOFILL	= 1 << 13;
-static const __maybe_unused u16 BTREE_ITER_KEY_CACHE_FILL	= 1 << 14;
-#define __BTREE_ITER_FLAGS_END					       15
+#define BTREE_TRIGGER_FLAGS()			\
+	x(norun)				\
+	x(transactional)			\
+	x(atomic)				\
+	x(check_repair)				\
+	x(gc)					\
+	x(insert)				\
+	x(overwrite)				\
+	x(is_root)				\
+	x(bucket_invalidate)
+
+enum {
+#define x(n) BTREE_ITER_FLAG_BIT_##n,
+	BTREE_ITER_FLAGS()
+	STR_HASH_FLAGS()
+	BTREE_UPDATE_FLAGS()
+	BTREE_TRIGGER_FLAGS()
+#undef x
+};
+
+/* iter flags must fit in a u16: */
+//BUILD_BUG_ON(BTREE_ITER_FLAG_BIT_key_cache_fill > 15);
+
+enum btree_iter_update_trigger_flags {
+#define x(n) BTREE_ITER_##n	= 1U << BTREE_ITER_FLAG_BIT_##n,
+	BTREE_ITER_FLAGS()
+#undef x
+#define x(n) STR_HASH_##n	= 1U << BTREE_ITER_FLAG_BIT_##n,
+	STR_HASH_FLAGS()
+#undef x
+#define x(n) BTREE_UPDATE_##n	= 1U << BTREE_ITER_FLAG_BIT_##n,
+	BTREE_UPDATE_FLAGS()
+#undef x
+#define x(n) BTREE_TRIGGER_##n	= 1U << BTREE_ITER_FLAG_BIT_##n,
+	BTREE_TRIGGER_FLAGS()
+#undef x
+};
 
 enum btree_path_uptodate {
 	BTREE_ITER_UPTODATE		= 0,
@@ -301,7 +372,7 @@ struct btree_iter {
 	 */
 	struct bkey		k;
 
-	/* BTREE_ITER_WITH_JOURNAL: */
+	/* BTREE_ITER_with_journal: */
 	size_t			journal_idx;
 #ifdef TRACK_PATH_ALLOCATED
 	unsigned long		ip_allocated;
@@ -315,9 +386,9 @@ struct bkey_cached {
 	struct btree_bkey_cached_common c;
 
 	unsigned long		flags;
+	unsigned long		btree_trans_barrier_seq;
 	u16			u64s;
 	bool			valid;
-	u32			btree_trans_barrier_seq;
 	struct bkey_cached_key	key;
 
 	struct rhash_head	hash;
@@ -358,7 +429,21 @@ struct btree_insert_entry {
 	unsigned long		ip_allocated;
 };
 
+/* Number of btree paths we preallocate, usually enough */
 #define BTREE_ITER_INITIAL		64
+/*
+ * Lmiit for btree_trans_too_many_iters(); this is enough that almost all code
+ * paths should run inside this limit, and if they don't it usually indicates a
+ * bug (leaking/duplicated btree paths).
+ *
+ * exception: some fsck paths
+ *
+ * bugs with excessive path usage seem to have possibly been eliminated now, so
+ * we might consider eliminating this (and btree_trans_too_many_iter()) at some
+ * point.
+ */
+#define BTREE_ITER_NORMAL_LIMIT		256
+/* never exceed limit */
 #define BTREE_ITER_MAX			(1U << 10)
 
 struct btree_trans_commit_hook;
@@ -398,6 +483,9 @@ struct btree_trans {
 	u8			lock_must_abort;
 	bool			lock_may_not_fail:1;
 	bool			srcu_held:1;
+	bool			locked:1;
+	bool			pf_memalloc_nofs:1;
+	bool			write_locked:1;
 	bool			used_mempool:1;
 	bool			in_traverse_all:1;
 	bool			paths_sorted:1;
@@ -405,13 +493,13 @@ struct btree_trans {
 	bool			journal_transaction_names:1;
 	bool			journal_replay_not_finished:1;
 	bool			notrace_relock_fail:1;
-	bool			write_locked:1;
 	enum bch_errcode	restarted:16;
 	u32			restart_count;
 
 	u64			last_begin_time;
 	unsigned long		last_begin_ip;
 	unsigned long		last_restarted_ip;
+	unsigned long		last_unlock_ip;
 	unsigned long		srcu_lock_time;
 
 	const char		*fn;
@@ -654,6 +742,7 @@ const char *bch2_btree_node_type_str(enum btree_node_type);
 	 BIT_ULL(BKEY_TYPE_inodes)|			\
 	 BIT_ULL(BKEY_TYPE_stripes)|			\
 	 BIT_ULL(BKEY_TYPE_reflink)|			\
+	 BIT_ULL(BKEY_TYPE_subvolumes)|			\
 	 BIT_ULL(BKEY_TYPE_btree))
 
 #define BTREE_NODE_TYPE_HAS_ATOMIC_TRIGGERS		\
@@ -673,13 +762,13 @@ static inline bool btree_node_type_needs_gc(enum btree_node_type type)
 
 static inline bool btree_node_type_is_extents(enum btree_node_type type)
 {
-	const unsigned mask = 0
+	const u64 mask = 0
 #define x(name, nr, flags, ...)	|((!!((flags) & BTREE_ID_EXTENTS)) << (nr + 1))
 	BCH_BTREE_IDS()
 #undef x
 	;
 
-	return (1U << type) & mask;
+	return BIT_ULL(type) & mask;
 }
 
 static inline bool btree_id_is_extents(enum btree_id btree)
@@ -689,35 +778,35 @@ static inline bool btree_id_is_extents(enum btree_id btree)
 
 static inline bool btree_type_has_snapshots(enum btree_id id)
 {
-	const unsigned mask = 0
+	const u64 mask = 0
 #define x(name, nr, flags, ...)	|((!!((flags) & BTREE_ID_SNAPSHOTS)) << nr)
 	BCH_BTREE_IDS()
 #undef x
 	;
 
-	return (1U << id) & mask;
+	return BIT_ULL(id) & mask;
 }
 
 static inline bool btree_type_has_snapshot_field(enum btree_id id)
 {
-	const unsigned mask = 0
+	const u64 mask = 0
 #define x(name, nr, flags, ...)	|((!!((flags) & (BTREE_ID_SNAPSHOT_FIELD|BTREE_ID_SNAPSHOTS))) << nr)
 	BCH_BTREE_IDS()
 #undef x
 	;
 
-	return (1U << id) & mask;
+	return BIT_ULL(id) & mask;
 }
 
 static inline bool btree_type_has_ptrs(enum btree_id id)
 {
-	const unsigned mask = 0
+	const u64 mask = 0
 #define x(name, nr, flags, ...)	|((!!((flags) & BTREE_ID_DATA)) << nr)
 	BCH_BTREE_IDS()
 #undef x
 	;
 
-	return (1U << id) & mask;
+	return BIT_ULL(id) & mask;
 }
 
 struct btree_root {
@@ -727,7 +816,7 @@ struct btree_root {
 	__BKEY_PADDED(key, BKEY_BTREE_PTR_VAL_U64s_MAX);
 	u8			level;
 	u8			alive;
-	s8			error;
+	s16			error;
 };
 
 enum btree_gc_coalesce_fail_reason {
diff --git a/fs/bcachefs/btree_update.c b/fs/bcachefs/btree_update.c
index c3ff365acc..f3c645a43d 100644
--- a/fs/bcachefs/btree_update.c
+++ b/fs/bcachefs/btree_update.c
@@ -25,19 +25,22 @@ static inline int btree_insert_entry_cmp(const struct btree_insert_entry *l,
 
 static int __must_check
 bch2_trans_update_by_path(struct btree_trans *, btree_path_idx_t,
-			  struct bkey_i *, enum btree_update_flags,
+			  struct bkey_i *, enum btree_iter_update_trigger_flags,
 			  unsigned long ip);
 
 static noinline int extent_front_merge(struct btree_trans *trans,
 				       struct btree_iter *iter,
 				       struct bkey_s_c k,
 				       struct bkey_i **insert,
-				       enum btree_update_flags flags)
+				       enum btree_iter_update_trigger_flags flags)
 {
 	struct bch_fs *c = trans->c;
 	struct bkey_i *update;
 	int ret;
 
+	if (unlikely(trans->journal_replay_not_finished))
+		return 0;
+
 	update = bch2_bkey_make_mut_noupdate(trans, k);
 	ret = PTR_ERR_OR_ZERO(update);
 	if (ret)
@@ -69,6 +72,9 @@ static noinline int extent_back_merge(struct btree_trans *trans,
 	struct bch_fs *c = trans->c;
 	int ret;
 
+	if (unlikely(trans->journal_replay_not_finished))
+		return 0;
+
 	ret =   bch2_key_has_snapshot_overwrites(trans, iter->btree_id, insert->k.p) ?:
 		bch2_key_has_snapshot_overwrites(trans, iter->btree_id, k.k->p);
 	if (ret < 0)
@@ -98,8 +104,8 @@ static int need_whiteout_for_snapshot(struct btree_trans *trans,
 	pos.snapshot++;
 
 	for_each_btree_key_norestart(trans, iter, btree_id, pos,
-			   BTREE_ITER_ALL_SNAPSHOTS|
-			   BTREE_ITER_NOPRESERVE, k, ret) {
+			   BTREE_ITER_all_snapshots|
+			   BTREE_ITER_nopreserve, k, ret) {
 		if (!bkey_eq(k.k->p, pos))
 			break;
 
@@ -132,8 +138,8 @@ int __bch2_insert_snapshot_whiteouts(struct btree_trans *trans,
 	darray_init(&s);
 
 	bch2_trans_iter_init(trans, &old_iter, id, old_pos,
-			     BTREE_ITER_NOT_EXTENTS|
-			     BTREE_ITER_ALL_SNAPSHOTS);
+			     BTREE_ITER_not_extents|
+			     BTREE_ITER_all_snapshots);
 	while ((old_k = bch2_btree_iter_prev(&old_iter)).k &&
 	       !(ret = bkey_err(old_k)) &&
 	       bkey_eq(old_pos, old_k.k->p)) {
@@ -145,8 +151,8 @@ int __bch2_insert_snapshot_whiteouts(struct btree_trans *trans,
 			continue;
 
 		new_k = bch2_bkey_get_iter(trans, &new_iter, id, whiteout_pos,
-					   BTREE_ITER_NOT_EXTENTS|
-					   BTREE_ITER_INTENT);
+					   BTREE_ITER_not_extents|
+					   BTREE_ITER_intent);
 		ret = bkey_err(new_k);
 		if (ret)
 			break;
@@ -162,7 +168,7 @@ int __bch2_insert_snapshot_whiteouts(struct btree_trans *trans,
 			update->k.type		= KEY_TYPE_whiteout;
 
 			ret = bch2_trans_update(trans, &new_iter, update,
-						BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
+						BTREE_UPDATE_internal_snapshot_node);
 		}
 		bch2_trans_iter_exit(trans, &new_iter);
 
@@ -179,7 +185,7 @@ int __bch2_insert_snapshot_whiteouts(struct btree_trans *trans,
 
 int bch2_trans_update_extent_overwrite(struct btree_trans *trans,
 				       struct btree_iter *iter,
-				       enum btree_update_flags flags,
+				       enum btree_iter_update_trigger_flags flags,
 				       struct bkey_s_c old,
 				       struct bkey_s_c new)
 {
@@ -212,7 +218,7 @@ int bch2_trans_update_extent_overwrite(struct btree_trans *trans,
 		ret =   bch2_insert_snapshot_whiteouts(trans, btree_id,
 					old.k->p, update->k.p) ?:
 			bch2_btree_insert_nonextent(trans, btree_id, update,
-					BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE|flags);
+					BTREE_UPDATE_internal_snapshot_node|flags);
 		if (ret)
 			return ret;
 	}
@@ -229,7 +235,7 @@ int bch2_trans_update_extent_overwrite(struct btree_trans *trans,
 		ret =   bch2_insert_snapshot_whiteouts(trans, btree_id,
 					old.k->p, update->k.p) ?:
 			bch2_btree_insert_nonextent(trans, btree_id, update,
-					  BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE|flags);
+					  BTREE_UPDATE_internal_snapshot_node|flags);
 		if (ret)
 			return ret;
 	}
@@ -254,7 +260,7 @@ int bch2_trans_update_extent_overwrite(struct btree_trans *trans,
 		}
 
 		ret = bch2_btree_insert_nonextent(trans, btree_id, update,
-					  BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE|flags);
+					  BTREE_UPDATE_internal_snapshot_node|flags);
 		if (ret)
 			return ret;
 	}
@@ -267,7 +273,7 @@ int bch2_trans_update_extent_overwrite(struct btree_trans *trans,
 		bch2_cut_front(new.k->p, update);
 
 		ret = bch2_trans_update_by_path(trans, iter->path, update,
-					  BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE|
+					  BTREE_UPDATE_internal_snapshot_node|
 					  flags, _RET_IP_);
 		if (ret)
 			return ret;
@@ -279,7 +285,7 @@ int bch2_trans_update_extent_overwrite(struct btree_trans *trans,
 static int bch2_trans_update_extent(struct btree_trans *trans,
 				    struct btree_iter *orig_iter,
 				    struct bkey_i *insert,
-				    enum btree_update_flags flags)
+				    enum btree_iter_update_trigger_flags flags)
 {
 	struct btree_iter iter;
 	struct bkey_s_c k;
@@ -287,9 +293,9 @@ static int bch2_trans_update_extent(struct btree_trans *trans,
 	int ret = 0;
 
 	bch2_trans_iter_init(trans, &iter, btree_id, bkey_start_pos(&insert->k),
-			     BTREE_ITER_INTENT|
-			     BTREE_ITER_WITH_UPDATES|
-			     BTREE_ITER_NOT_EXTENTS);
+			     BTREE_ITER_intent|
+			     BTREE_ITER_with_updates|
+			     BTREE_ITER_not_extents);
 	k = bch2_btree_iter_peek_upto(&iter, POS(insert->k.p.inode, U64_MAX));
 	if ((ret = bkey_err(k)))
 		goto err;
@@ -340,7 +346,7 @@ err:
 
 static noinline int flush_new_cached_update(struct btree_trans *trans,
 					    struct btree_insert_entry *i,
-					    enum btree_update_flags flags,
+					    enum btree_iter_update_trigger_flags flags,
 					    unsigned long ip)
 {
 	struct bkey k;
@@ -348,7 +354,7 @@ static noinline int flush_new_cached_update(struct btree_trans *trans,
 
 	btree_path_idx_t path_idx =
 		bch2_path_get(trans, i->btree_id, i->old_k.p, 1, 0,
-			      BTREE_ITER_INTENT, _THIS_IP_);
+			      BTREE_ITER_intent, _THIS_IP_);
 	ret = bch2_btree_path_traverse(trans, path_idx, 0);
 	if (ret)
 		goto out;
@@ -366,7 +372,7 @@ static noinline int flush_new_cached_update(struct btree_trans *trans,
 		goto out;
 
 	i->key_cache_already_flushed = true;
-	i->flags |= BTREE_TRIGGER_NORUN;
+	i->flags |= BTREE_TRIGGER_norun;
 
 	btree_path_set_should_be_locked(btree_path);
 	ret = bch2_trans_update_by_path(trans, path_idx, i->k, flags, ip);
@@ -377,7 +383,7 @@ out:
 
 static int __must_check
 bch2_trans_update_by_path(struct btree_trans *trans, btree_path_idx_t path_idx,
-			  struct bkey_i *k, enum btree_update_flags flags,
+			  struct bkey_i *k, enum btree_iter_update_trigger_flags flags,
 			  unsigned long ip)
 {
 	struct bch_fs *c = trans->c;
@@ -452,7 +458,7 @@ bch2_trans_update_by_path(struct btree_trans *trans, btree_path_idx_t path_idx,
 	 * the key cache - but the key has to exist in the btree for that to
 	 * work:
 	 */
-	if (path->cached && bkey_deleted(&i->old_k))
+	if (path->cached && !i->old_btree_u64s)
 		return flush_new_cached_update(trans, i, flags, ip);
 
 	return 0;
@@ -473,15 +479,15 @@ static noinline int bch2_trans_update_get_key_cache(struct btree_trans *trans,
 		if (!iter->key_cache_path)
 			iter->key_cache_path =
 				bch2_path_get(trans, path->btree_id, path->pos, 1, 0,
-					      BTREE_ITER_INTENT|
-					      BTREE_ITER_CACHED, _THIS_IP_);
+					      BTREE_ITER_intent|
+					      BTREE_ITER_cached, _THIS_IP_);
 
 		iter->key_cache_path =
 			bch2_btree_path_set_pos(trans, iter->key_cache_path, path->pos,
-						iter->flags & BTREE_ITER_INTENT,
+						iter->flags & BTREE_ITER_intent,
 						_THIS_IP_);
 
-		ret = bch2_btree_path_traverse(trans, iter->key_cache_path, BTREE_ITER_CACHED);
+		ret = bch2_btree_path_traverse(trans, iter->key_cache_path, BTREE_ITER_cached);
 		if (unlikely(ret))
 			return ret;
 
@@ -499,17 +505,17 @@ static noinline int bch2_trans_update_get_key_cache(struct btree_trans *trans,
 }
 
 int __must_check bch2_trans_update(struct btree_trans *trans, struct btree_iter *iter,
-				   struct bkey_i *k, enum btree_update_flags flags)
+				   struct bkey_i *k, enum btree_iter_update_trigger_flags flags)
 {
 	btree_path_idx_t path_idx = iter->update_path ?: iter->path;
 	int ret;
 
-	if (iter->flags & BTREE_ITER_IS_EXTENTS)
+	if (iter->flags & BTREE_ITER_is_extents)
 		return bch2_trans_update_extent(trans, iter, k, flags);
 
 	if (bkey_deleted(&k->k) &&
-	    !(flags & BTREE_UPDATE_KEY_CACHE_RECLAIM) &&
-	    (iter->flags & BTREE_ITER_FILTER_SNAPSHOTS)) {
+	    !(flags & BTREE_UPDATE_key_cache_reclaim) &&
+	    (iter->flags & BTREE_ITER_filter_snapshots)) {
 		ret = need_whiteout_for_snapshot(trans, iter->btree_id, k->k.p);
 		if (unlikely(ret < 0))
 			return ret;
@@ -522,7 +528,7 @@ int __must_check bch2_trans_update(struct btree_trans *trans, struct btree_iter
 	 * Ensure that updates to cached btrees go to the key cache:
 	 */
 	struct btree_path *path = trans->paths + path_idx;
-	if (!(flags & BTREE_UPDATE_KEY_CACHE_RECLAIM) &&
+	if (!(flags & BTREE_UPDATE_key_cache_reclaim) &&
 	    !path->cached &&
 	    !path->level &&
 	    btree_id_cached(trans->c, path->btree_id)) {
@@ -581,7 +587,7 @@ int bch2_bkey_get_empty_slot(struct btree_trans *trans, struct btree_iter *iter,
 	struct bkey_s_c k;
 	int ret = 0;
 
-	bch2_trans_iter_init(trans, iter, btree, POS_MAX, BTREE_ITER_INTENT);
+	bch2_trans_iter_init(trans, iter, btree, POS_MAX, BTREE_ITER_intent);
 	k = bch2_btree_iter_prev(iter);
 	ret = bkey_err(k);
 	if (ret)
@@ -615,15 +621,15 @@ void bch2_trans_commit_hook(struct btree_trans *trans,
 
 int bch2_btree_insert_nonextent(struct btree_trans *trans,
 				enum btree_id btree, struct bkey_i *k,
-				enum btree_update_flags flags)
+				enum btree_iter_update_trigger_flags flags)
 {
 	struct btree_iter iter;
 	int ret;
 
 	bch2_trans_iter_init(trans, &iter, btree, k->k.p,
-			     BTREE_ITER_CACHED|
-			     BTREE_ITER_NOT_EXTENTS|
-			     BTREE_ITER_INTENT);
+			     BTREE_ITER_cached|
+			     BTREE_ITER_not_extents|
+			     BTREE_ITER_intent);
 	ret   = bch2_btree_iter_traverse(&iter) ?:
 		bch2_trans_update(trans, &iter, k, flags);
 	bch2_trans_iter_exit(trans, &iter);
@@ -631,16 +637,13 @@ int bch2_btree_insert_nonextent(struct btree_trans *trans,
 }
 
 int bch2_btree_insert_trans(struct btree_trans *trans, enum btree_id id,
-			    struct bkey_i *k, enum btree_update_flags flags)
+			    struct bkey_i *k, enum btree_iter_update_trigger_flags flags)
 {
 	struct btree_iter iter;
-	int ret;
-
 	bch2_trans_iter_init(trans, &iter, id, bkey_start_pos(&k->k),
-			     BTREE_ITER_CACHED|
-			     BTREE_ITER_INTENT);
-	ret   = bch2_btree_iter_traverse(&iter) ?:
-		bch2_trans_update(trans, &iter, k, flags);
+			     BTREE_ITER_intent|flags);
+	int ret = bch2_btree_iter_traverse(&iter) ?:
+		  bch2_trans_update(trans, &iter, k, flags);
 	bch2_trans_iter_exit(trans, &iter);
 	return ret;
 }
@@ -692,8 +695,8 @@ int bch2_btree_delete(struct btree_trans *trans,
 	int ret;
 
 	bch2_trans_iter_init(trans, &iter, btree, pos,
-			     BTREE_ITER_CACHED|
-			     BTREE_ITER_INTENT);
+			     BTREE_ITER_cached|
+			     BTREE_ITER_intent);
 	ret   = bch2_btree_iter_traverse(&iter) ?:
 		bch2_btree_delete_at(trans, &iter, update_flags);
 	bch2_trans_iter_exit(trans, &iter);
@@ -711,7 +714,7 @@ int bch2_btree_delete_range_trans(struct btree_trans *trans, enum btree_id id,
 	struct bkey_s_c k;
 	int ret = 0;
 
-	bch2_trans_iter_init(trans, &iter, id, start, BTREE_ITER_INTENT);
+	bch2_trans_iter_init(trans, &iter, id, start, BTREE_ITER_intent);
 	while ((k = bch2_btree_iter_peek_upto(&iter, end)).k) {
 		struct disk_reservation disk_res =
 			bch2_disk_reservation_init(trans->c, 0);
@@ -739,7 +742,7 @@ int bch2_btree_delete_range_trans(struct btree_trans *trans, enum btree_id id,
 		 */
 		delete.k.p = iter.pos;
 
-		if (iter.flags & BTREE_ITER_IS_EXTENTS)
+		if (iter.flags & BTREE_ITER_is_extents)
 			bch2_key_resize(&delete.k,
 					bpos_min(end, k.k->p).offset -
 					iter.pos.offset);
@@ -788,6 +791,27 @@ int bch2_btree_delete_range(struct bch_fs *c, enum btree_id id,
 int bch2_btree_bit_mod(struct btree_trans *trans, enum btree_id btree,
 		       struct bpos pos, bool set)
 {
+	struct bkey_i *k = bch2_trans_kmalloc(trans, sizeof(*k));
+	int ret = PTR_ERR_OR_ZERO(k);
+	if (ret)
+		return ret;
+
+	bkey_init(&k->k);
+	k->k.type = set ? KEY_TYPE_set : KEY_TYPE_deleted;
+	k->k.p = pos;
+
+	struct btree_iter iter;
+	bch2_trans_iter_init(trans, &iter, btree, pos, BTREE_ITER_intent);
+
+	ret   = bch2_btree_iter_traverse(&iter) ?:
+		bch2_trans_update(trans, &iter, k, 0);
+	bch2_trans_iter_exit(trans, &iter);
+	return ret;
+}
+
+int bch2_btree_bit_mod_buffered(struct btree_trans *trans, enum btree_id btree,
+				struct bpos pos, bool set)
+{
 	struct bkey_i k;
 
 	bkey_init(&k.k);
@@ -825,7 +849,7 @@ __bch2_fs_log_msg(struct bch_fs *c, unsigned commit_flags, const char *fmt,
 	if (ret)
 		goto err;
 
-	if (!test_bit(JOURNAL_STARTED, &c->journal.flags)) {
+	if (!test_bit(JOURNAL_running, &c->journal.flags)) {
 		ret = darray_make_room(&c->journal.early_journal_entries, jset_u64s(u64s));
 		if (ret)
 			goto err;
diff --git a/fs/bcachefs/btree_update.h b/fs/bcachefs/btree_update.h
index b9382b7b28..b4894e4d54 100644
--- a/fs/bcachefs/btree_update.h
+++ b/fs/bcachefs/btree_update.h
@@ -44,16 +44,18 @@ enum bch_trans_commit_flags {
 #undef x
 };
 
+void bch2_trans_commit_flags_to_text(struct printbuf *, enum bch_trans_commit_flags);
+
 int bch2_btree_delete_extent_at(struct btree_trans *, struct btree_iter *,
 				unsigned, unsigned);
 int bch2_btree_delete_at(struct btree_trans *, struct btree_iter *, unsigned);
 int bch2_btree_delete(struct btree_trans *, enum btree_id, struct bpos, unsigned);
 
 int bch2_btree_insert_nonextent(struct btree_trans *, enum btree_id,
-				struct bkey_i *, enum btree_update_flags);
+				struct bkey_i *, enum btree_iter_update_trigger_flags);
 
 int bch2_btree_insert_trans(struct btree_trans *, enum btree_id, struct bkey_i *,
-			enum btree_update_flags);
+			enum btree_iter_update_trigger_flags);
 int bch2_btree_insert(struct bch_fs *, enum btree_id, struct bkey_i *,
 		     struct disk_reservation *, int flags);
 
@@ -63,11 +65,12 @@ int bch2_btree_delete_range(struct bch_fs *, enum btree_id,
 			    struct bpos, struct bpos, unsigned, u64 *);
 
 int bch2_btree_bit_mod(struct btree_trans *, enum btree_id, struct bpos, bool);
+int bch2_btree_bit_mod_buffered(struct btree_trans *, enum btree_id, struct bpos, bool);
 
 static inline int bch2_btree_delete_at_buffered(struct btree_trans *trans,
 						enum btree_id btree, struct bpos pos)
 {
-	return bch2_btree_bit_mod(trans, btree, pos, false);
+	return bch2_btree_bit_mod_buffered(trans, btree, pos, false);
 }
 
 int __bch2_insert_snapshot_whiteouts(struct btree_trans *, enum btree_id,
@@ -93,14 +96,14 @@ static inline int bch2_insert_snapshot_whiteouts(struct btree_trans *trans,
 }
 
 int bch2_trans_update_extent_overwrite(struct btree_trans *, struct btree_iter *,
-				       enum btree_update_flags,
+				       enum btree_iter_update_trigger_flags,
 				       struct bkey_s_c, struct bkey_s_c);
 
 int bch2_bkey_get_empty_slot(struct btree_trans *, struct btree_iter *,
 			     enum btree_id, struct bpos);
 
 int __must_check bch2_trans_update(struct btree_trans *, struct btree_iter *,
-				   struct bkey_i *, enum btree_update_flags);
+				   struct bkey_i *, enum btree_iter_update_trigger_flags);
 
 struct jset_entry *__bch2_trans_jset_entry_alloc(struct btree_trans *, unsigned);
 
@@ -275,7 +278,7 @@ static inline struct bkey_i *__bch2_bkey_get_mut_noupdate(struct btree_trans *tr
 					 unsigned flags, unsigned type, unsigned min_bytes)
 {
 	struct bkey_s_c k = __bch2_bkey_get_iter(trans, iter,
-				btree_id, pos, flags|BTREE_ITER_INTENT, type);
+				btree_id, pos, flags|BTREE_ITER_intent, type);
 	struct bkey_i *ret = IS_ERR(k.k)
 		? ERR_CAST(k.k)
 		: __bch2_bkey_make_mut_noupdate(trans, k, 0, min_bytes);
@@ -298,7 +301,7 @@ static inline struct bkey_i *__bch2_bkey_get_mut(struct btree_trans *trans,
 					 unsigned flags, unsigned type, unsigned min_bytes)
 {
 	struct bkey_i *mut = __bch2_bkey_get_mut_noupdate(trans, iter,
-				btree_id, pos, flags|BTREE_ITER_INTENT, type, min_bytes);
+				btree_id, pos, flags|BTREE_ITER_intent, type, min_bytes);
 	int ret;
 
 	if (IS_ERR(mut))
diff --git a/fs/bcachefs/btree_update_interior.c b/fs/bcachefs/btree_update_interior.c
index 4530b14ff2..60b8544cea 100644
--- a/fs/bcachefs/btree_update_interior.c
+++ b/fs/bcachefs/btree_update_interior.c
@@ -2,6 +2,7 @@
 
 #include "bcachefs.h"
 #include "alloc_foreground.h"
+#include "bkey_buf.h"
 #include "bkey_methods.h"
 #include "btree_cache.h"
 #include "btree_gc.h"
@@ -18,83 +19,140 @@
 #include "journal.h"
 #include "journal_reclaim.h"
 #include "keylist.h"
+#include "recovery_passes.h"
 #include "replicas.h"
+#include "sb-members.h"
 #include "super-io.h"
 #include "trace.h"
 
 #include <linux/random.h>
 
+static const char * const bch2_btree_update_modes[] = {
+#define x(t) #t,
+	BTREE_UPDATE_MODES()
+#undef x
+	NULL
+};
+
 static int bch2_btree_insert_node(struct btree_update *, struct btree_trans *,
-				  btree_path_idx_t, struct btree *,
-				  struct keylist *, unsigned);
+				  btree_path_idx_t, struct btree *, struct keylist *);
 static void bch2_btree_update_add_new_node(struct btree_update *, struct btree *);
 
-static btree_path_idx_t get_unlocked_mut_path(struct btree_trans *trans,
-					      enum btree_id btree_id,
-					      unsigned level,
-					      struct bpos pos)
-{
-	btree_path_idx_t path_idx = bch2_path_get(trans, btree_id, pos, level + 1, level,
-			     BTREE_ITER_NOPRESERVE|
-			     BTREE_ITER_INTENT, _RET_IP_);
-	path_idx = bch2_btree_path_make_mut(trans, path_idx, true, _RET_IP_);
-
-	struct btree_path *path = trans->paths + path_idx;
-	bch2_btree_path_downgrade(trans, path);
-	__bch2_btree_path_unlock(trans, path);
-	return path_idx;
-}
-
-/* Debug code: */
-
 /*
  * Verify that child nodes correctly span parent node's range:
  */
-static void btree_node_interior_verify(struct bch_fs *c, struct btree *b)
+int bch2_btree_node_check_topology(struct btree_trans *trans, struct btree *b)
 {
-#ifdef CONFIG_BCACHEFS_DEBUG
-	struct bpos next_node = b->data->min_key;
-	struct btree_node_iter iter;
+	struct bch_fs *c = trans->c;
+	struct bpos node_min = b->key.k.type == KEY_TYPE_btree_ptr_v2
+		? bkey_i_to_btree_ptr_v2(&b->key)->v.min_key
+		: b->data->min_key;
+	struct btree_and_journal_iter iter;
 	struct bkey_s_c k;
-	struct bkey_s_c_btree_ptr_v2 bp;
-	struct bkey unpacked;
-	struct printbuf buf1 = PRINTBUF, buf2 = PRINTBUF;
+	struct printbuf buf = PRINTBUF;
+	struct bkey_buf prev;
+	int ret = 0;
 
-	BUG_ON(!b->c.level);
+	BUG_ON(b->key.k.type == KEY_TYPE_btree_ptr_v2 &&
+	       !bpos_eq(bkey_i_to_btree_ptr_v2(&b->key)->v.min_key,
+			b->data->min_key));
+
+	if (b == btree_node_root(c, b)) {
+		if (!bpos_eq(b->data->min_key, POS_MIN)) {
+			printbuf_reset(&buf);
+			bch2_bpos_to_text(&buf, b->data->min_key);
+			need_fsck_err(c, btree_root_bad_min_key,
+				      "btree root with incorrect min_key: %s", buf.buf);
+			goto topology_repair;
+		}
 
-	if (!test_bit(JOURNAL_REPLAY_DONE, &c->journal.flags))
-		return;
+		if (!bpos_eq(b->data->max_key, SPOS_MAX)) {
+			printbuf_reset(&buf);
+			bch2_bpos_to_text(&buf, b->data->max_key);
+			need_fsck_err(c, btree_root_bad_max_key,
+				      "btree root with incorrect max_key: %s", buf.buf);
+			goto topology_repair;
+		}
+	}
 
-	bch2_btree_node_iter_init_from_start(&iter, b);
+	if (!b->c.level)
+		return 0;
 
-	while (1) {
-		k = bch2_btree_node_iter_peek_unpack(&iter, b, &unpacked);
+	bch2_bkey_buf_init(&prev);
+	bkey_init(&prev.k->k);
+	bch2_btree_and_journal_iter_init_node_iter(trans, &iter, b);
+
+	while ((k = bch2_btree_and_journal_iter_peek(&iter)).k) {
 		if (k.k->type != KEY_TYPE_btree_ptr_v2)
-			break;
-		bp = bkey_s_c_to_btree_ptr_v2(k);
+			goto out;
 
-		if (!bpos_eq(next_node, bp.v->min_key)) {
-			bch2_dump_btree_node(c, b);
-			bch2_bpos_to_text(&buf1, next_node);
-			bch2_bpos_to_text(&buf2, bp.v->min_key);
-			panic("expected next min_key %s got %s\n", buf1.buf, buf2.buf);
-		}
+		struct bkey_s_c_btree_ptr_v2 bp = bkey_s_c_to_btree_ptr_v2(k);
 
-		bch2_btree_node_iter_advance(&iter, b);
+		struct bpos expected_min = bkey_deleted(&prev.k->k)
+			? node_min
+			: bpos_successor(prev.k->k.p);
 
-		if (bch2_btree_node_iter_end(&iter)) {
-			if (!bpos_eq(k.k->p, b->key.k.p)) {
-				bch2_dump_btree_node(c, b);
-				bch2_bpos_to_text(&buf1, b->key.k.p);
-				bch2_bpos_to_text(&buf2, k.k->p);
-				panic("expected end %s got %s\n", buf1.buf, buf2.buf);
-			}
-			break;
+		if (!bpos_eq(expected_min, bp.v->min_key)) {
+			bch2_topology_error(c);
+
+			printbuf_reset(&buf);
+			prt_str(&buf, "end of prev node doesn't match start of next node\n"),
+			prt_printf(&buf, "  in btree %s level %u node ",
+				   bch2_btree_id_str(b->c.btree_id), b->c.level);
+			bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&b->key));
+			prt_str(&buf, "\n  prev ");
+			bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(prev.k));
+			prt_str(&buf, "\n  next ");
+			bch2_bkey_val_to_text(&buf, c, k);
+
+			need_fsck_err(c, btree_node_topology_bad_min_key, "%s", buf.buf);
+			goto topology_repair;
 		}
 
-		next_node = bpos_successor(k.k->p);
+		bch2_bkey_buf_reassemble(&prev, c, k);
+		bch2_btree_and_journal_iter_advance(&iter);
+	}
+
+	if (bkey_deleted(&prev.k->k)) {
+		bch2_topology_error(c);
+
+		printbuf_reset(&buf);
+		prt_str(&buf, "empty interior node\n");
+		prt_printf(&buf, "  in btree %s level %u node ",
+			   bch2_btree_id_str(b->c.btree_id), b->c.level);
+		bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&b->key));
+
+		need_fsck_err(c, btree_node_topology_empty_interior_node, "%s", buf.buf);
+		goto topology_repair;
+	} else if (!bpos_eq(prev.k->k.p, b->key.k.p)) {
+		bch2_topology_error(c);
+
+		printbuf_reset(&buf);
+		prt_str(&buf, "last child node doesn't end at end of parent node\n");
+		prt_printf(&buf, "  in btree %s level %u node ",
+			   bch2_btree_id_str(b->c.btree_id), b->c.level);
+		bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&b->key));
+		prt_str(&buf, "\n  last key ");
+		bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(prev.k));
+
+		need_fsck_err(c, btree_node_topology_bad_max_key, "%s", buf.buf);
+		goto topology_repair;
+	}
+out:
+fsck_err:
+	bch2_btree_and_journal_iter_exit(&iter);
+	bch2_bkey_buf_exit(&prev, c);
+	printbuf_exit(&buf);
+	return ret;
+topology_repair:
+	if ((c->recovery_passes_explicit & BIT_ULL(BCH_RECOVERY_PASS_check_topology)) &&
+	    c->curr_recovery_pass > BCH_RECOVERY_PASS_check_topology) {
+		bch2_inconsistent_error(c);
+		ret = -BCH_ERR_btree_need_topology_repair;
+	} else {
+		ret = bch2_run_explicit_recovery_pass(c, BCH_RECOVERY_PASS_check_topology);
 	}
-#endif
+	goto out;
 }
 
 /* Calculate ideal packed bkey format for new btree nodes: */
@@ -102,7 +160,6 @@ static void btree_node_interior_verify(struct bch_fs *c, struct btree *b)
 static void __bch2_btree_calc_format(struct bkey_format_state *s, struct btree *b)
 {
 	struct bkey_packed *k;
-	struct bset_tree *t;
 	struct bkey uk;
 
 	for_each_bset(b, t)
@@ -255,7 +312,7 @@ static struct btree *__bch2_btree_node_alloc(struct btree_trans *trans,
 	struct open_buckets obs = { .nr = 0 };
 	struct bch_devs_list devs_have = (struct bch_devs_list) { 0 };
 	enum bch_watermark watermark = flags & BCH_WATERMARK_MASK;
-	unsigned nr_reserve = watermark > BCH_WATERMARK_reclaim
+	unsigned nr_reserve = watermark < BCH_WATERMARK_reclaim
 		? BTREE_NODE_RESERVE
 		: 0;
 	int ret;
@@ -550,6 +607,26 @@ static void btree_update_add_key(struct btree_update *as,
 	bch2_keylist_push(keys);
 }
 
+static bool btree_update_new_nodes_marked_sb(struct btree_update *as)
+{
+	for_each_keylist_key(&as->new_keys, k)
+		if (!bch2_dev_btree_bitmap_marked(as->c, bkey_i_to_s_c(k)))
+			return false;
+	return true;
+}
+
+static void btree_update_new_nodes_mark_sb(struct btree_update *as)
+{
+	struct bch_fs *c = as->c;
+
+	mutex_lock(&c->sb_lock);
+	for_each_keylist_key(&as->new_keys, k)
+		bch2_dev_btree_bitmap_mark(c, bkey_i_to_s_c(k));
+
+	bch2_write_super(c);
+	mutex_unlock(&c->sb_lock);
+}
+
 /*
  * The transactional part of an interior btree node update, where we journal the
  * update we did to the interior node and update alloc info:
@@ -570,7 +647,7 @@ static int btree_update_nodes_written_trans(struct btree_trans *trans,
 		unsigned level = bkey_i_to_btree_ptr_v2(k)->v.mem_ptr;
 
 		ret = bch2_key_trigger_old(trans, as->btree_id, level, bkey_i_to_s_c(k),
-					   BTREE_TRIGGER_TRANSACTIONAL);
+					   BTREE_TRIGGER_transactional);
 		if (ret)
 			return ret;
 	}
@@ -579,7 +656,7 @@ static int btree_update_nodes_written_trans(struct btree_trans *trans,
 		unsigned level = bkey_i_to_btree_ptr_v2(k)->v.mem_ptr;
 
 		ret = bch2_key_trigger_new(trans, as->btree_id, level, bkey_i_to_s(k),
-					   BTREE_TRIGGER_TRANSACTIONAL);
+					   BTREE_TRIGGER_transactional);
 		if (ret)
 			return ret;
 	}
@@ -607,6 +684,9 @@ static void btree_update_nodes_written(struct btree_update *as)
 	if (ret)
 		goto err;
 
+	if (!btree_update_new_nodes_marked_sb(as))
+		btree_update_new_nodes_mark_sb(as);
+
 	/*
 	 * Wait for any in flight writes to finish before we free the old nodes
 	 * on disk:
@@ -639,7 +719,7 @@ static void btree_update_nodes_written(struct btree_update *as)
 	 * which may require allocations as well.
 	 */
 	ret = commit_do(trans, &as->disk_res, &journal_seq,
-			BCH_WATERMARK_reclaim|
+			BCH_WATERMARK_interior_updates|
 			BCH_TRANS_COMMIT_no_enospc|
 			BCH_TRANS_COMMIT_no_check_rw|
 			BCH_TRANS_COMMIT_journal_reclaim,
@@ -647,14 +727,15 @@ static void btree_update_nodes_written(struct btree_update *as)
 	bch2_trans_unlock(trans);
 
 	bch2_fs_fatal_err_on(ret && !bch2_journal_error(&c->journal), c,
-			     "%s(): error %s", __func__, bch2_err_str(ret));
+			     "%s", bch2_err_str(ret));
 err:
-	if (as->b) {
-
-		b = as->b;
-		btree_path_idx_t path_idx = get_unlocked_mut_path(trans,
-						as->btree_id, b->c.level, b->key.k.p);
-		struct btree_path *path = trans->paths + path_idx;
+	/*
+	 * We have to be careful because another thread might be getting ready
+	 * to free as->b and calling btree_update_reparent() on us - we'll
+	 * recheck under btree_update_lock below:
+	 */
+	b = READ_ONCE(as->b);
+	if (b) {
 		/*
 		 * @b is the node we did the final insert into:
 		 *
@@ -672,12 +753,16 @@ err:
 		 * btree_node_lock_nopath() (the use of which is always suspect,
 		 * we need to work on removing this in the future)
 		 *
-		 * It should be, but get_unlocked_mut_path() -> bch2_path_get()
+		 * It should be, but bch2_path_get_unlocked_mut() -> bch2_path_get()
 		 * calls bch2_path_upgrade(), before we call path_make_mut(), so
 		 * we may rarely end up with a locked path besides the one we
 		 * have here:
 		 */
 		bch2_trans_unlock(trans);
+		bch2_trans_begin(trans);
+		btree_path_idx_t path_idx = bch2_path_get_unlocked_mut(trans,
+						as->btree_id, b->c.level, b->key.k.p);
+		struct btree_path *path = trans->paths + path_idx;
 		btree_node_lock_nopath_nofail(trans, &b->c, SIX_LOCK_intent);
 		mark_btree_node_locked(trans, path, b->c.level, BTREE_NODE_INTENT_LOCKED);
 		path->l[b->c.level].lock_seq = six_lock_seq(&b->c.lock);
@@ -795,15 +880,17 @@ static void btree_update_updated_node(struct btree_update *as, struct btree *b)
 {
 	struct bch_fs *c = as->c;
 
-	mutex_lock(&c->btree_interior_update_lock);
-	list_add_tail(&as->unwritten_list, &c->btree_interior_updates_unwritten);
-
-	BUG_ON(as->mode != BTREE_INTERIOR_NO_UPDATE);
+	BUG_ON(as->mode != BTREE_UPDATE_none);
+	BUG_ON(as->update_level_end < b->c.level);
 	BUG_ON(!btree_node_dirty(b));
 	BUG_ON(!b->c.level);
 
-	as->mode	= BTREE_INTERIOR_UPDATING_NODE;
+	mutex_lock(&c->btree_interior_update_lock);
+	list_add_tail(&as->unwritten_list, &c->btree_interior_updates_unwritten);
+
+	as->mode	= BTREE_UPDATE_node;
 	as->b		= b;
+	as->update_level_end = b->c.level;
 
 	set_btree_node_write_blocked(b);
 	list_add(&as->write_blocked_list, &b->write_blocked);
@@ -825,7 +912,7 @@ static void btree_update_reparent(struct btree_update *as,
 	lockdep_assert_held(&c->btree_interior_update_lock);
 
 	child->b = NULL;
-	child->mode = BTREE_INTERIOR_UPDATING_AS;
+	child->mode = BTREE_UPDATE_update;
 
 	bch2_journal_pin_copy(&c->journal, &as->journal, &child->journal,
 			      bch2_update_reparent_journal_pin_flush);
@@ -836,7 +923,7 @@ static void btree_update_updated_root(struct btree_update *as, struct btree *b)
 	struct bkey_i *insert = &b->key;
 	struct bch_fs *c = as->c;
 
-	BUG_ON(as->mode != BTREE_INTERIOR_NO_UPDATE);
+	BUG_ON(as->mode != BTREE_UPDATE_none);
 
 	BUG_ON(as->journal_u64s + jset_u64s(insert->k.u64s) >
 	       ARRAY_SIZE(as->journal_entries));
@@ -850,7 +937,7 @@ static void btree_update_updated_root(struct btree_update *as, struct btree *b)
 	mutex_lock(&c->btree_interior_update_lock);
 	list_add_tail(&as->unwritten_list, &c->btree_interior_updates_unwritten);
 
-	as->mode	= BTREE_INTERIOR_UPDATING_ROOT;
+	as->mode	= BTREE_UPDATE_root;
 	mutex_unlock(&c->btree_interior_update_lock);
 }
 
@@ -1028,7 +1115,7 @@ static void bch2_btree_update_done(struct btree_update *as, struct btree_trans *
 	struct bch_fs *c = as->c;
 	u64 start_time = as->start_time;
 
-	BUG_ON(as->mode == BTREE_INTERIOR_NO_UPDATE);
+	BUG_ON(as->mode == BTREE_UPDATE_none);
 
 	if (as->took_gc_lock)
 		up_read(&as->c->gc_lock);
@@ -1045,7 +1132,7 @@ static void bch2_btree_update_done(struct btree_update *as, struct btree_trans *
 
 static struct btree_update *
 bch2_btree_update_start(struct btree_trans *trans, struct btree_path *path,
-			unsigned level, bool split, unsigned flags)
+			unsigned level_start, bool split, unsigned flags)
 {
 	struct bch_fs *c = trans->c;
 	struct btree_update *as;
@@ -1053,7 +1140,7 @@ bch2_btree_update_start(struct btree_trans *trans, struct btree_path *path,
 	int disk_res_flags = (flags & BCH_TRANS_COMMIT_no_enospc)
 		? BCH_DISK_RESERVATION_NOFAIL : 0;
 	unsigned nr_nodes[2] = { 0, 0 };
-	unsigned update_level = level;
+	unsigned level_end = level_start;
 	enum bch_watermark watermark = flags & BCH_WATERMARK_MASK;
 	int ret = 0;
 	u32 restart_count = trans->restart_count;
@@ -1068,29 +1155,29 @@ bch2_btree_update_start(struct btree_trans *trans, struct btree_path *path,
 	flags &= ~BCH_WATERMARK_MASK;
 	flags |= watermark;
 
-	if (!(flags & BCH_TRANS_COMMIT_journal_reclaim) &&
-	    watermark < c->journal.watermark) {
-		struct journal_res res = { 0 };
+	if (watermark < BCH_WATERMARK_reclaim &&
+	    test_bit(JOURNAL_space_low, &c->journal.flags)) {
+		if (flags & BCH_TRANS_COMMIT_journal_reclaim)
+			return ERR_PTR(-BCH_ERR_journal_reclaim_would_deadlock);
 
 		ret = drop_locks_do(trans,
-			bch2_journal_res_get(&c->journal, &res, 1,
-					     watermark|JOURNAL_RES_GET_CHECK));
+			({ wait_event(c->journal.wait, !test_bit(JOURNAL_space_low, &c->journal.flags)); 0; }));
 		if (ret)
 			return ERR_PTR(ret);
 	}
 
 	while (1) {
-		nr_nodes[!!update_level] += 1 + split;
-		update_level++;
+		nr_nodes[!!level_end] += 1 + split;
+		level_end++;
 
-		ret = bch2_btree_path_upgrade(trans, path, update_level + 1);
+		ret = bch2_btree_path_upgrade(trans, path, level_end + 1);
 		if (ret)
 			return ERR_PTR(ret);
 
-		if (!btree_path_node(path, update_level)) {
+		if (!btree_path_node(path, level_end)) {
 			/* Allocating new root? */
 			nr_nodes[1] += split;
-			update_level = BTREE_MAX_DEPTH;
+			level_end = BTREE_MAX_DEPTH;
 			break;
 		}
 
@@ -1098,11 +1185,11 @@ bch2_btree_update_start(struct btree_trans *trans, struct btree_path *path,
 		 * Always check for space for two keys, even if we won't have to
 		 * split at prior level - it might have been a merge instead:
 		 */
-		if (bch2_btree_node_insert_fits(path->l[update_level].b,
+		if (bch2_btree_node_insert_fits(path->l[level_end].b,
 						BKEY_BTREE_PTR_U64s_MAX * 2))
 			break;
 
-		split = path->l[update_level].b->nr.live_u64s > BTREE_SPLIT_THRESHOLD(c);
+		split = path->l[level_end].b->nr.live_u64s > BTREE_SPLIT_THRESHOLD(c);
 	}
 
 	if (!down_read_trylock(&c->gc_lock)) {
@@ -1116,12 +1203,15 @@ bch2_btree_update_start(struct btree_trans *trans, struct btree_path *path,
 	as = mempool_alloc(&c->btree_interior_update_pool, GFP_NOFS);
 	memset(as, 0, sizeof(*as));
 	closure_init(&as->cl, NULL);
-	as->c		= c;
-	as->start_time	= start_time;
-	as->mode	= BTREE_INTERIOR_NO_UPDATE;
-	as->took_gc_lock = true;
-	as->btree_id	= path->btree_id;
-	as->update_level = update_level;
+	as->c			= c;
+	as->start_time		= start_time;
+	as->ip_started		= _RET_IP_;
+	as->mode		= BTREE_UPDATE_none;
+	as->flags		= flags;
+	as->took_gc_lock	= true;
+	as->btree_id		= path->btree_id;
+	as->update_level_start	= level_start;
+	as->update_level_end	= level_end;
 	INIT_LIST_HEAD(&as->list);
 	INIT_LIST_HEAD(&as->unwritten_list);
 	INIT_LIST_HEAD(&as->write_blocked_list);
@@ -1163,7 +1253,7 @@ bch2_btree_update_start(struct btree_trans *trans, struct btree_path *path,
 		 */
 		if (bch2_err_matches(ret, ENOSPC) &&
 		    (flags & BCH_TRANS_COMMIT_journal_reclaim) &&
-		    watermark != BCH_WATERMARK_reclaim) {
+		    watermark < BCH_WATERMARK_reclaim) {
 			ret = -BCH_ERR_journal_reclaim_would_deadlock;
 			goto err;
 		}
@@ -1193,7 +1283,8 @@ bch2_btree_update_start(struct btree_trans *trans, struct btree_path *path,
 err:
 	bch2_btree_update_free(as, trans);
 	if (!bch2_err_matches(ret, ENOSPC) &&
-	    !bch2_err_matches(ret, EROFS))
+	    !bch2_err_matches(ret, EROFS) &&
+	    ret != -BCH_ERR_journal_reclaim_would_deadlock)
 		bch_err_fn_ratelimited(c, ret);
 	return ERR_PTR(ret);
 }
@@ -1208,33 +1299,35 @@ static void bch2_btree_set_root_inmem(struct bch_fs *c, struct btree *b)
 	mutex_unlock(&c->btree_cache.lock);
 
 	mutex_lock(&c->btree_root_lock);
-	BUG_ON(btree_node_root(c, b) &&
-	       (b->c.level < btree_node_root(c, b)->c.level ||
-		!btree_node_dying(btree_node_root(c, b))));
-
 	bch2_btree_id_root(c, b->c.btree_id)->b = b;
 	mutex_unlock(&c->btree_root_lock);
 
 	bch2_recalc_btree_reserve(c);
 }
 
-static void bch2_btree_set_root(struct btree_update *as,
-				struct btree_trans *trans,
-				struct btree_path *path,
-				struct btree *b)
+static int bch2_btree_set_root(struct btree_update *as,
+			       struct btree_trans *trans,
+			       struct btree_path *path,
+			       struct btree *b,
+			       bool nofail)
 {
 	struct bch_fs *c = as->c;
-	struct btree *old;
 
 	trace_and_count(c, btree_node_set_root, trans, b);
 
-	old = btree_node_root(c, b);
+	struct btree *old = btree_node_root(c, b);
 
 	/*
 	 * Ensure no one is using the old root while we switch to the
 	 * new root:
 	 */
-	bch2_btree_node_lock_write_nofail(trans, path, &old->c);
+	if (nofail) {
+		bch2_btree_node_lock_write_nofail(trans, path, &old->c);
+	} else {
+		int ret = bch2_btree_node_lock_write(trans, path, &old->c);
+		if (ret)
+			return ret;
+	}
 
 	bch2_btree_set_root_inmem(c, b);
 
@@ -1248,6 +1341,7 @@ static void bch2_btree_set_root(struct btree_update *as,
 	 * depend on the new root would have to update the new root.
 	 */
 	bch2_btree_node_unlock_write(trans, path, old);
+	return 0;
 }
 
 /* Interior node updates: */
@@ -1267,7 +1361,7 @@ static void bch2_insert_fixup_btree_ptr(struct btree_update *as,
 	BUG_ON(insert->k.type == KEY_TYPE_btree_ptr_v2 &&
 	       !btree_ptr_sectors_written(insert));
 
-	if (unlikely(!test_bit(JOURNAL_REPLAY_DONE, &c->journal.flags)))
+	if (unlikely(!test_bit(JOURNAL_replay_done, &c->journal.flags)))
 		bch2_journal_key_overwritten(c, b->c.btree_id, b->c.level, insert->k.p);
 
 	if (bch2_bkey_invalid(c, bkey_i_to_s_c(insert),
@@ -1314,12 +1408,12 @@ static void bch2_insert_fixup_btree_ptr(struct btree_update *as,
 }
 
 static void
-__bch2_btree_insert_keys_interior(struct btree_update *as,
-				  struct btree_trans *trans,
-				  struct btree_path *path,
-				  struct btree *b,
-				  struct btree_node_iter node_iter,
-				  struct keylist *keys)
+bch2_btree_insert_keys_interior(struct btree_update *as,
+				struct btree_trans *trans,
+				struct btree_path *path,
+				struct btree *b,
+				struct btree_node_iter node_iter,
+				struct keylist *keys)
 {
 	struct bkey_i *insert = bch2_keylist_front(keys);
 	struct bkey_packed *k;
@@ -1378,9 +1472,16 @@ static void __btree_split_node(struct btree_update *as,
 		if (bkey_deleted(k))
 			continue;
 
+		uk = bkey_unpack_key(b, k);
+
+		if (b->c.level &&
+		    u64s < n1_u64s &&
+		    u64s + k->u64s >= n1_u64s &&
+		    bch2_key_deleted_in_journal(trans, b->c.btree_id, b->c.level, uk.p))
+			n1_u64s += k->u64s;
+
 		i = u64s >= n1_u64s;
 		u64s += k->u64s;
-		uk = bkey_unpack_key(b, k);
 		if (!i)
 			n1_pos = uk.p;
 		bch2_bkey_format_add_key(&format[i], &uk);
@@ -1439,8 +1540,7 @@ static void __btree_split_node(struct btree_update *as,
 
 		bch2_verify_btree_nr_keys(n[i]);
 
-		if (b->c.level)
-			btree_node_interior_verify(as->c, n[i]);
+		BUG_ON(bch2_btree_node_check_topology(trans, n[i]));
 	}
 }
 
@@ -1469,15 +1569,15 @@ static void btree_split_insert_keys(struct btree_update *as,
 
 		bch2_btree_node_iter_init(&node_iter, b, &bch2_keylist_front(keys)->k.p);
 
-		__bch2_btree_insert_keys_interior(as, trans, path, b, node_iter, keys);
+		bch2_btree_insert_keys_interior(as, trans, path, b, node_iter, keys);
 
-		btree_node_interior_verify(as->c, b);
+		BUG_ON(bch2_btree_node_check_topology(trans, b));
 	}
 }
 
 static int btree_split(struct btree_update *as, struct btree_trans *trans,
 		       btree_path_idx_t path, struct btree *b,
-		       struct keylist *keys, unsigned flags)
+		       struct keylist *keys)
 {
 	struct bch_fs *c = as->c;
 	struct btree *parent = btree_node_parent(trans->paths + path, b);
@@ -1486,9 +1586,14 @@ static int btree_split(struct btree_update *as, struct btree_trans *trans,
 	u64 start_time = local_clock();
 	int ret = 0;
 
+	bch2_verify_btree_nr_keys(b);
 	BUG_ON(!parent && (b != btree_node_root(c, b)));
 	BUG_ON(parent && !btree_node_intent_locked(trans->paths + path, b->c.level + 1));
 
+	ret = bch2_btree_node_check_topology(trans, b);
+	if (ret)
+		return ret;
+
 	bch2_btree_interior_update_will_free_node(as, b);
 
 	if (b->nr.live_u64s > BTREE_SPLIT_THRESHOLD(c)) {
@@ -1515,12 +1620,12 @@ static int btree_split(struct btree_update *as, struct btree_trans *trans,
 		six_unlock_write(&n2->c.lock);
 		six_unlock_write(&n1->c.lock);
 
-		path1 = get_unlocked_mut_path(trans, as->btree_id, n1->c.level, n1->key.k.p);
+		path1 = bch2_path_get_unlocked_mut(trans, as->btree_id, n1->c.level, n1->key.k.p);
 		six_lock_increment(&n1->c.lock, SIX_LOCK_intent);
 		mark_btree_node_locked(trans, trans->paths + path1, n1->c.level, BTREE_NODE_INTENT_LOCKED);
 		bch2_btree_path_level_init(trans, trans->paths + path1, n1);
 
-		path2 = get_unlocked_mut_path(trans, as->btree_id, n2->c.level, n2->key.k.p);
+		path2 = bch2_path_get_unlocked_mut(trans, as->btree_id, n2->c.level, n2->key.k.p);
 		six_lock_increment(&n2->c.lock, SIX_LOCK_intent);
 		mark_btree_node_locked(trans, trans->paths + path2, n2->c.level, BTREE_NODE_INTENT_LOCKED);
 		bch2_btree_path_level_init(trans, trans->paths + path2, n2);
@@ -1565,7 +1670,7 @@ static int btree_split(struct btree_update *as, struct btree_trans *trans,
 		bch2_btree_update_add_new_node(as, n1);
 		six_unlock_write(&n1->c.lock);
 
-		path1 = get_unlocked_mut_path(trans, as->btree_id, n1->c.level, n1->key.k.p);
+		path1 = bch2_path_get_unlocked_mut(trans, as->btree_id, n1->c.level, n1->key.k.p);
 		six_lock_increment(&n1->c.lock, SIX_LOCK_intent);
 		mark_btree_node_locked(trans, trans->paths + path1, n1->c.level, BTREE_NODE_INTENT_LOCKED);
 		bch2_btree_path_level_init(trans, trans->paths + path1, n1);
@@ -1578,16 +1683,17 @@ static int btree_split(struct btree_update *as, struct btree_trans *trans,
 
 	if (parent) {
 		/* Split a non root node */
-		ret = bch2_btree_insert_node(as, trans, path, parent, &as->parent_keys, flags);
-		if (ret)
-			goto err;
+		ret = bch2_btree_insert_node(as, trans, path, parent, &as->parent_keys);
 	} else if (n3) {
-		bch2_btree_set_root(as, trans, trans->paths + path, n3);
+		ret = bch2_btree_set_root(as, trans, trans->paths + path, n3, false);
 	} else {
 		/* Root filled up but didn't need to be split */
-		bch2_btree_set_root(as, trans, trans->paths + path, n1);
+		ret = bch2_btree_set_root(as, trans, trans->paths + path, n1, false);
 	}
 
+	if (ret)
+		goto err;
+
 	if (n3) {
 		bch2_btree_update_get_open_buckets(as, n3);
 		bch2_btree_node_write(c, n3, SIX_LOCK_intent, 0);
@@ -1644,27 +1750,6 @@ err:
 	goto out;
 }
 
-static void
-bch2_btree_insert_keys_interior(struct btree_update *as,
-				struct btree_trans *trans,
-				struct btree_path *path,
-				struct btree *b,
-				struct keylist *keys)
-{
-	struct btree_path *linked;
-	unsigned i;
-
-	__bch2_btree_insert_keys_interior(as, trans, path, b,
-					  path->l[b->c.level].iter, keys);
-
-	btree_update_updated_node(as, b);
-
-	trans_for_each_path_with_node(trans, b, linked, i)
-		bch2_btree_node_iter_peek(&linked->l[b->c.level].iter, b);
-
-	bch2_trans_verify_paths(trans);
-}
-
 /**
  * bch2_btree_insert_node - insert bkeys into a given btree node
  *
@@ -1673,7 +1758,6 @@ bch2_btree_insert_keys_interior(struct btree_update *as,
  * @path_idx:		path that points to current node
  * @b:			node to insert keys into
  * @keys:		list of keys to insert
- * @flags:		transaction commit flags
  *
  * Returns: 0 on success, typically transaction restart error on failure
  *
@@ -1683,10 +1767,11 @@ bch2_btree_insert_keys_interior(struct btree_update *as,
  */
 static int bch2_btree_insert_node(struct btree_update *as, struct btree_trans *trans,
 				  btree_path_idx_t path_idx, struct btree *b,
-				  struct keylist *keys, unsigned flags)
+				  struct keylist *keys)
 {
 	struct bch_fs *c = as->c;
-	struct btree_path *path = trans->paths + path_idx;
+	struct btree_path *path = trans->paths + path_idx, *linked;
+	unsigned i;
 	int old_u64s = le16_to_cpu(btree_bset_last(b)->u64s);
 	int old_live_u64s = b->nr.live_u64s;
 	int live_u64s_added, u64s_added;
@@ -1709,9 +1794,19 @@ static int bch2_btree_insert_node(struct btree_update *as, struct btree_trans *t
 		goto split;
 	}
 
-	btree_node_interior_verify(c, b);
+	ret = bch2_btree_node_check_topology(trans, b);
+	if (ret) {
+		bch2_btree_node_unlock_write(trans, path, b);
+		return ret;
+	}
 
-	bch2_btree_insert_keys_interior(as, trans, path, b, keys);
+	bch2_btree_insert_keys_interior(as, trans, path, b,
+					path->l[b->c.level].iter, keys);
+
+	trans_for_each_path_with_node(trans, b, linked, i)
+		bch2_btree_node_iter_peek(&linked->l[b->c.level].iter, b);
+
+	bch2_trans_verify_paths(trans);
 
 	live_u64s_added = (int) b->nr.live_u64s - old_live_u64s;
 	u64s_added = (int) le16_to_cpu(btree_bset_last(b)->u64s) - old_u64s;
@@ -1725,21 +1820,22 @@ static int bch2_btree_insert_node(struct btree_update *as, struct btree_trans *t
 	    bch2_maybe_compact_whiteouts(c, b))
 		bch2_trans_node_reinit_iter(trans, b);
 
+	btree_update_updated_node(as, b);
 	bch2_btree_node_unlock_write(trans, path, b);
 
-	btree_node_interior_verify(c, b);
+	BUG_ON(bch2_btree_node_check_topology(trans, b));
 	return 0;
 split:
 	/*
 	 * We could attempt to avoid the transaction restart, by calling
 	 * bch2_btree_path_upgrade() and allocating more nodes:
 	 */
-	if (b->c.level >= as->update_level) {
+	if (b->c.level >= as->update_level_end) {
 		trace_and_count(c, trans_restart_split_race, trans, _THIS_IP_, b);
 		return btree_trans_restart(trans, BCH_ERR_transaction_restart_split_race);
 	}
 
-	return btree_split(as, trans, path_idx, b, keys, flags);
+	return btree_split(as, trans, path_idx, b, keys);
 }
 
 int bch2_btree_split_leaf(struct btree_trans *trans,
@@ -1747,7 +1843,6 @@ int bch2_btree_split_leaf(struct btree_trans *trans,
 			  unsigned flags)
 {
 	/* btree_split & merge may both cause paths array to be reallocated */
-
 	struct btree *b = path_l(trans->paths + path)->b;
 	struct btree_update *as;
 	unsigned l;
@@ -1759,7 +1854,7 @@ int bch2_btree_split_leaf(struct btree_trans *trans,
 	if (IS_ERR(as))
 		return PTR_ERR(as);
 
-	ret = btree_split(as, trans, path, b, NULL, flags);
+	ret = btree_split(as, trans, path, b, NULL);
 	if (ret) {
 		bch2_btree_update_free(as, trans);
 		return ret;
@@ -1775,6 +1870,65 @@ int bch2_btree_split_leaf(struct btree_trans *trans,
 	return ret;
 }
 
+static void __btree_increase_depth(struct btree_update *as, struct btree_trans *trans,
+				   btree_path_idx_t path_idx)
+{
+	struct bch_fs *c = as->c;
+	struct btree_path *path = trans->paths + path_idx;
+	struct btree *n, *b = bch2_btree_id_root(c, path->btree_id)->b;
+
+	BUG_ON(!btree_node_locked(path, b->c.level));
+
+	n = __btree_root_alloc(as, trans, b->c.level + 1);
+
+	bch2_btree_update_add_new_node(as, n);
+	six_unlock_write(&n->c.lock);
+
+	path->locks_want++;
+	BUG_ON(btree_node_locked(path, n->c.level));
+	six_lock_increment(&n->c.lock, SIX_LOCK_intent);
+	mark_btree_node_locked(trans, path, n->c.level, BTREE_NODE_INTENT_LOCKED);
+	bch2_btree_path_level_init(trans, path, n);
+
+	n->sib_u64s[0] = U16_MAX;
+	n->sib_u64s[1] = U16_MAX;
+
+	bch2_keylist_add(&as->parent_keys, &b->key);
+	btree_split_insert_keys(as, trans, path_idx, n, &as->parent_keys);
+
+	int ret = bch2_btree_set_root(as, trans, path, n, true);
+	BUG_ON(ret);
+
+	bch2_btree_update_get_open_buckets(as, n);
+	bch2_btree_node_write(c, n, SIX_LOCK_intent, 0);
+	bch2_trans_node_add(trans, path, n);
+	six_unlock_intent(&n->c.lock);
+
+	mutex_lock(&c->btree_cache.lock);
+	list_add_tail(&b->list, &c->btree_cache.live);
+	mutex_unlock(&c->btree_cache.lock);
+
+	bch2_trans_verify_locks(trans);
+}
+
+int bch2_btree_increase_depth(struct btree_trans *trans, btree_path_idx_t path, unsigned flags)
+{
+	struct bch_fs *c = trans->c;
+	struct btree *b = bch2_btree_id_root(c, trans->paths[path].btree_id)->b;
+
+	if (btree_node_fake(b))
+		return bch2_btree_split_leaf(trans, path, flags);
+
+	struct btree_update *as =
+		bch2_btree_update_start(trans, trans->paths + path, b->c.level, true, flags);
+	if (IS_ERR(as))
+		return PTR_ERR(as);
+
+	__btree_increase_depth(as, trans, path);
+	bch2_btree_update_done(as, trans);
+	return 0;
+}
+
 int __bch2_foreground_maybe_merge(struct btree_trans *trans,
 				  btree_path_idx_t path,
 				  unsigned level,
@@ -1794,9 +1948,27 @@ int __bch2_foreground_maybe_merge(struct btree_trans *trans,
 	u64 start_time = local_clock();
 	int ret = 0;
 
+	bch2_trans_verify_not_in_restart(trans);
+	bch2_trans_verify_not_unlocked(trans);
 	BUG_ON(!trans->paths[path].should_be_locked);
 	BUG_ON(!btree_node_locked(&trans->paths[path], level));
 
+	/*
+	 * Work around a deadlock caused by the btree write buffer not doing
+	 * merges and leaving tons of merges for us to do - we really don't need
+	 * to be doing merges at all from the interior update path, and if the
+	 * interior update path is generating too many new interior updates we
+	 * deadlock:
+	 */
+	if ((flags & BCH_WATERMARK_MASK) == BCH_WATERMARK_interior_updates)
+		return 0;
+
+	if ((flags & BCH_WATERMARK_MASK) <= BCH_WATERMARK_reclaim) {
+		flags &= ~BCH_WATERMARK_MASK;
+		flags |= BCH_WATERMARK_btree;
+		flags |= BCH_TRANS_COMMIT_journal_reclaim;
+	}
+
 	b = trans->paths[path].l[level].b;
 
 	if ((sib == btree_prev_sib && bpos_eq(b->data->min_key, POS_MIN)) ||
@@ -1810,7 +1982,7 @@ int __bch2_foreground_maybe_merge(struct btree_trans *trans,
 		: bpos_successor(b->data->max_key);
 
 	sib_path = bch2_path_get(trans, btree, sib_pos,
-				 U8_MAX, level, BTREE_ITER_INTENT, _THIS_IP_);
+				 U8_MAX, level, BTREE_ITER_intent, _THIS_IP_);
 	ret = bch2_btree_path_traverse(trans, sib_path, false);
 	if (ret)
 		goto err;
@@ -1845,8 +2017,7 @@ int __bch2_foreground_maybe_merge(struct btree_trans *trans,
 			__func__, buf1.buf, buf2.buf);
 		printbuf_exit(&buf1);
 		printbuf_exit(&buf2);
-		bch2_topology_error(c);
-		ret = -EIO;
+		ret = bch2_topology_error(c);
 		goto err;
 	}
 
@@ -1904,7 +2075,7 @@ int __bch2_foreground_maybe_merge(struct btree_trans *trans,
 	bch2_btree_update_add_new_node(as, n);
 	six_unlock_write(&n->c.lock);
 
-	new_path = get_unlocked_mut_path(trans, btree, n->c.level, n->key.k.p);
+	new_path = bch2_path_get_unlocked_mut(trans, btree, n->c.level, n->key.k.p);
 	six_lock_increment(&n->c.lock, SIX_LOCK_intent);
 	mark_btree_node_locked(trans, trans->paths + new_path, n->c.level, BTREE_NODE_INTENT_LOCKED);
 	bch2_btree_path_level_init(trans, trans->paths + new_path, n);
@@ -1916,7 +2087,7 @@ int __bch2_foreground_maybe_merge(struct btree_trans *trans,
 
 	bch2_trans_verify_paths(trans);
 
-	ret = bch2_btree_insert_node(as, trans, path, parent, &as->parent_keys, flags);
+	ret = bch2_btree_insert_node(as, trans, path, parent, &as->parent_keys);
 	if (ret)
 		goto err_free_update;
 
@@ -1943,6 +2114,10 @@ err:
 		bch2_path_put(trans, new_path, true);
 	bch2_path_put(trans, sib_path, true);
 	bch2_trans_verify_locks(trans);
+	if (ret == -BCH_ERR_journal_reclaim_would_deadlock)
+		ret = 0;
+	if (!ret)
+		ret = bch2_trans_relock(trans);
 	return ret;
 err_free_update:
 	bch2_btree_node_free_never_used(as, trans, n);
@@ -1978,7 +2153,7 @@ int bch2_btree_node_rewrite(struct btree_trans *trans,
 	bch2_btree_update_add_new_node(as, n);
 	six_unlock_write(&n->c.lock);
 
-	new_path = get_unlocked_mut_path(trans, iter->btree_id, n->c.level, n->key.k.p);
+	new_path = bch2_path_get_unlocked_mut(trans, iter->btree_id, n->c.level, n->key.k.p);
 	six_lock_increment(&n->c.lock, SIX_LOCK_intent);
 	mark_btree_node_locked(trans, trans->paths + new_path, n->c.level, BTREE_NODE_INTENT_LOCKED);
 	bch2_btree_path_level_init(trans, trans->paths + new_path, n);
@@ -1987,14 +2162,14 @@ int bch2_btree_node_rewrite(struct btree_trans *trans,
 
 	if (parent) {
 		bch2_keylist_add(&as->parent_keys, &n->key);
-		ret = bch2_btree_insert_node(as, trans, iter->path,
-					     parent, &as->parent_keys, flags);
-		if (ret)
-			goto err;
+		ret = bch2_btree_insert_node(as, trans, iter->path, parent, &as->parent_keys);
 	} else {
-		bch2_btree_set_root(as, trans, btree_iter_path(trans, iter), n);
+		ret = bch2_btree_set_root(as, trans, btree_iter_path(trans, iter), n, false);
 	}
 
+	if (ret)
+		goto err;
+
 	bch2_btree_update_get_open_buckets(as, n);
 	bch2_btree_node_write(c, n, SIX_LOCK_intent, 0);
 
@@ -2069,7 +2244,7 @@ static void async_btree_node_rewrite_work(struct work_struct *work)
 
 	ret = bch2_trans_do(c, NULL, NULL, 0,
 		      async_btree_node_rewrite_trans(trans, a));
-	bch_err_fn(c, ret);
+	bch_err_fn_ratelimited(c, ret);
 	bch2_write_ref_put(c, BCH_WRITE_REF_node_rewrite);
 	kfree(a);
 }
@@ -2116,7 +2291,7 @@ void bch2_btree_node_rewrite_async(struct bch_fs *c, struct btree *b)
 		bch2_write_ref_get(c, BCH_WRITE_REF_node_rewrite);
 	}
 
-	queue_work(c->btree_interior_update_worker, &a->work);
+	queue_work(c->btree_node_rewrite_worker, &a->work);
 }
 
 void bch2_do_pending_node_rewrites(struct bch_fs *c)
@@ -2128,7 +2303,7 @@ void bch2_do_pending_node_rewrites(struct bch_fs *c)
 		list_del(&a->list);
 
 		bch2_write_ref_get(c, BCH_WRITE_REF_node_rewrite);
-		queue_work(c->btree_interior_update_worker, &a->work);
+		queue_work(c->btree_node_rewrite_worker, &a->work);
 	}
 	mutex_unlock(&c->pending_node_rewrites_lock);
 }
@@ -2161,10 +2336,10 @@ static int __bch2_btree_node_update_key(struct btree_trans *trans,
 	if (!skip_triggers) {
 		ret   = bch2_key_trigger_old(trans, b->c.btree_id, b->c.level + 1,
 					     bkey_i_to_s_c(&b->key),
-					     BTREE_TRIGGER_TRANSACTIONAL) ?:
+					     BTREE_TRIGGER_transactional) ?:
 			bch2_key_trigger_new(trans, b->c.btree_id, b->c.level + 1,
 					     bkey_i_to_s(new_key),
-					     BTREE_TRIGGER_TRANSACTIONAL);
+					     BTREE_TRIGGER_transactional);
 		if (ret)
 			return ret;
 	}
@@ -2181,7 +2356,7 @@ static int __bch2_btree_node_update_key(struct btree_trans *trans,
 		bch2_trans_copy_iter(&iter2, iter);
 
 		iter2.path = bch2_btree_path_make_mut(trans, iter2.path,
-				iter2.flags & BTREE_ITER_INTENT,
+				iter2.flags & BTREE_ITER_intent,
 				_THIS_IP_);
 
 		struct btree_path *path2 = btree_iter_path(trans, &iter2);
@@ -2193,7 +2368,7 @@ static int __bch2_btree_node_update_key(struct btree_trans *trans,
 		trans->paths_sorted = false;
 
 		ret   = bch2_btree_iter_traverse(&iter2) ?:
-			bch2_trans_update(trans, &iter2, new_key, BTREE_TRIGGER_NORUN);
+			bch2_trans_update(trans, &iter2, new_key, BTREE_TRIGGER_norun);
 		if (ret)
 			goto err;
 	} else {
@@ -2301,7 +2476,7 @@ int bch2_btree_node_update_key_get_iter(struct btree_trans *trans,
 
 	bch2_trans_node_iter_init(trans, &iter, b->c.btree_id, b->key.k.p,
 				  BTREE_MAX_DEPTH, b->c.level,
-				  BTREE_ITER_INTENT);
+				  BTREE_ITER_intent);
 	ret = bch2_btree_iter_traverse(&iter);
 	if (ret)
 		goto out;
@@ -2315,7 +2490,6 @@ int bch2_btree_node_update_key_get_iter(struct btree_trans *trans,
 
 	BUG_ON(!btree_node_hashed(b));
 
-	struct bch_extent_ptr *ptr;
 	bch2_bkey_drop_ptrs(bkey_i_to_s(new_key), ptr,
 			    !bch2_bkey_has_device(bkey_i_to_s(&b->key), ptr->dev));
 
@@ -2339,7 +2513,7 @@ void bch2_btree_set_root_for_read(struct bch_fs *c, struct btree *b)
 	bch2_btree_set_root_inmem(c, b);
 }
 
-static int __bch2_btree_root_alloc(struct btree_trans *trans, enum btree_id id)
+int bch2_btree_root_alloc_fake_trans(struct btree_trans *trans, enum btree_id id, unsigned level)
 {
 	struct bch_fs *c = trans->c;
 	struct closure cl;
@@ -2358,7 +2532,7 @@ static int __bch2_btree_root_alloc(struct btree_trans *trans, enum btree_id id)
 
 	set_btree_node_fake(b);
 	set_btree_node_need_rewrite(b);
-	b->c.level	= 0;
+	b->c.level	= level;
 	b->c.btree_id	= id;
 
 	bkey_btree_ptr_init(&b->key);
@@ -2385,9 +2559,24 @@ static int __bch2_btree_root_alloc(struct btree_trans *trans, enum btree_id id)
 	return 0;
 }
 
-void bch2_btree_root_alloc(struct bch_fs *c, enum btree_id id)
+void bch2_btree_root_alloc_fake(struct bch_fs *c, enum btree_id id, unsigned level)
 {
-	bch2_trans_run(c, __bch2_btree_root_alloc(trans, id));
+	bch2_trans_run(c, bch2_btree_root_alloc_fake_trans(trans, id, level));
+}
+
+static void bch2_btree_update_to_text(struct printbuf *out, struct btree_update *as)
+{
+	prt_printf(out, "%ps: ", (void *) as->ip_started);
+	bch2_trans_commit_flags_to_text(out, as->flags);
+
+	prt_printf(out, " btree=%s l=%u-%u mode=%s nodes_written=%u cl.remaining=%u journal_seq=%llu\n",
+		   bch2_btree_id_str(as->btree_id),
+		   as->update_level_start,
+		   as->update_level_end,
+		   bch2_btree_update_modes[as->mode],
+		   as->nodes_written,
+		   closure_nr_remaining(&as->cl),
+		   as->journal.seq);
 }
 
 void bch2_btree_updates_to_text(struct printbuf *out, struct bch_fs *c)
@@ -2396,12 +2585,7 @@ void bch2_btree_updates_to_text(struct printbuf *out, struct bch_fs *c)
 
 	mutex_lock(&c->btree_interior_update_lock);
 	list_for_each_entry(as, &c->btree_interior_update_list, list)
-		prt_printf(out, "%p m %u w %u r %u j %llu\n",
-		       as,
-		       as->mode,
-		       as->nodes_written,
-		       closure_nr_remaining(&as->cl),
-		       as->journal.seq);
+		bch2_btree_update_to_text(out, as);
 	mutex_unlock(&c->btree_interior_update_lock);
 }
 
@@ -2465,6 +2649,8 @@ bch2_btree_roots_to_journal_entries(struct bch_fs *c,
 
 void bch2_fs_btree_interior_update_exit(struct bch_fs *c)
 {
+	if (c->btree_node_rewrite_worker)
+		destroy_workqueue(c->btree_node_rewrite_worker);
 	if (c->btree_interior_update_worker)
 		destroy_workqueue(c->btree_interior_update_worker);
 	mempool_exit(&c->btree_interior_update_pool);
@@ -2485,10 +2671,15 @@ void bch2_fs_btree_interior_update_init_early(struct bch_fs *c)
 int bch2_fs_btree_interior_update_init(struct bch_fs *c)
 {
 	c->btree_interior_update_worker =
-		alloc_workqueue("btree_update", WQ_UNBOUND|WQ_MEM_RECLAIM, 1);
+		alloc_workqueue("btree_update", WQ_UNBOUND|WQ_MEM_RECLAIM, 8);
 	if (!c->btree_interior_update_worker)
 		return -BCH_ERR_ENOMEM_btree_interior_update_worker_init;
 
+	c->btree_node_rewrite_worker =
+		alloc_ordered_workqueue("btree_node_rewrite", WQ_UNBOUND);
+	if (!c->btree_node_rewrite_worker)
+		return -BCH_ERR_ENOMEM_btree_interior_update_worker_init;
+
 	if (mempool_init_kmalloc_pool(&c->btree_interior_update_pool, 1,
 				      sizeof(struct btree_update)))
 		return -BCH_ERR_ENOMEM_btree_interior_update_pool_init;
diff --git a/fs/bcachefs/btree_update_interior.h b/fs/bcachefs/btree_update_interior.h
index c593c925d1..b5b76ce01c 100644
--- a/fs/bcachefs/btree_update_interior.h
+++ b/fs/bcachefs/btree_update_interior.h
@@ -10,6 +10,20 @@
 
 #define BTREE_UPDATE_JOURNAL_RES	(BTREE_UPDATE_NODES_MAX * (BKEY_BTREE_PTR_U64s_MAX + 1))
 
+int bch2_btree_node_check_topology(struct btree_trans *, struct btree *);
+
+#define BTREE_UPDATE_MODES()	\
+	x(none)			\
+	x(node)			\
+	x(root)			\
+	x(update)
+
+enum btree_update_mode {
+#define x(n)	BTREE_UPDATE_##n,
+	BTREE_UPDATE_MODES()
+#undef x
+};
+
 /*
  * Tracks an in progress split/rewrite of a btree node and the update to the
  * parent node:
@@ -32,28 +46,24 @@ struct btree_update {
 	struct closure			cl;
 	struct bch_fs			*c;
 	u64				start_time;
+	unsigned long			ip_started;
 
 	struct list_head		list;
 	struct list_head		unwritten_list;
 
-	/* What kind of update are we doing? */
-	enum {
-		BTREE_INTERIOR_NO_UPDATE,
-		BTREE_INTERIOR_UPDATING_NODE,
-		BTREE_INTERIOR_UPDATING_ROOT,
-		BTREE_INTERIOR_UPDATING_AS,
-	} mode;
-
+	enum btree_update_mode		mode;
+	enum bch_trans_commit_flags	flags;
 	unsigned			nodes_written:1;
 	unsigned			took_gc_lock:1;
 
 	enum btree_id			btree_id;
-	unsigned			update_level;
+	unsigned			update_level_start;
+	unsigned			update_level_end;
 
 	struct disk_reservation		disk_res;
 
 	/*
-	 * BTREE_INTERIOR_UPDATING_NODE:
+	 * BTREE_UPDATE_node:
 	 * The update that made the new nodes visible was a regular update to an
 	 * existing interior node - @b. We can't write out the update to @b
 	 * until the new nodes we created are finished writing, so we block @b
@@ -119,6 +129,8 @@ struct btree *__bch2_btree_node_alloc_replacement(struct btree_update *,
 
 int bch2_btree_split_leaf(struct btree_trans *, btree_path_idx_t, unsigned);
 
+int bch2_btree_increase_depth(struct btree_trans *, btree_path_idx_t, unsigned);
+
 int __bch2_foreground_maybe_merge(struct btree_trans *, btree_path_idx_t,
 				  unsigned, unsigned, enum btree_node_sibling);
 
@@ -132,6 +144,9 @@ static inline int bch2_foreground_maybe_merge_sibling(struct btree_trans *trans,
 
 	EBUG_ON(!btree_node_locked(path, level));
 
+	if (bch2_btree_node_merging_disabled)
+		return 0;
+
 	b = path->l[level].b;
 	if (b->sib_u64s[sib] > trans->c->btree_foreground_merge_threshold)
 		return 0;
@@ -160,7 +175,9 @@ int bch2_btree_node_update_key_get_iter(struct btree_trans *, struct btree *,
 					struct bkey_i *, unsigned, bool);
 
 void bch2_btree_set_root_for_read(struct bch_fs *, struct btree *);
-void bch2_btree_root_alloc(struct bch_fs *, enum btree_id);
+
+int bch2_btree_root_alloc_fake_trans(struct btree_trans *, enum btree_id, unsigned);
+void bch2_btree_root_alloc_fake(struct bch_fs *, enum btree_id, unsigned);
 
 static inline unsigned btree_update_reserve_required(struct bch_fs *c,
 						     struct btree *b)
diff --git a/fs/bcachefs/btree_write_buffer.c b/fs/bcachefs/btree_write_buffer.c
index ac78448619..d0e92d9480 100644
--- a/fs/bcachefs/btree_write_buffer.c
+++ b/fs/bcachefs/btree_write_buffer.c
@@ -1,16 +1,19 @@
 // SPDX-License-Identifier: GPL-2.0
 
 #include "bcachefs.h"
+#include "bkey_buf.h"
 #include "btree_locking.h"
 #include "btree_update.h"
 #include "btree_update_interior.h"
 #include "btree_write_buffer.h"
 #include "error.h"
+#include "extents.h"
 #include "journal.h"
 #include "journal_io.h"
 #include "journal_reclaim.h"
 
 #include <linux/prefetch.h>
+#include <linux/sort.h>
 
 static int bch2_btree_write_buffer_journal_flush(struct journal *,
 				struct journal_entry_pin *, u64);
@@ -46,6 +49,14 @@ static inline bool wb_key_ref_cmp(const struct wb_key_ref *l, const struct wb_ke
 #endif
 }
 
+static int wb_key_seq_cmp(const void *_l, const void *_r)
+{
+	const struct btree_write_buffered_key *l = _l;
+	const struct btree_write_buffered_key *r = _r;
+
+	return cmp_int(l->journal_seq, r->journal_seq);
+}
+
 /* Compare excluding idx, the low 24 bits: */
 static inline bool wb_key_eq(const void *_l, const void *_r)
 {
@@ -113,7 +124,7 @@ static noinline int wb_flush_one_slowpath(struct btree_trans *trans,
 	trans->journal_res.seq = wb->journal_seq;
 
 	return bch2_trans_update(trans, iter, &wb->k,
-				 BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE) ?:
+				 BTREE_UPDATE_internal_snapshot_node) ?:
 		bch2_trans_commit(trans, NULL, NULL,
 				  BCH_TRANS_COMMIT_no_enospc|
 				  BCH_TRANS_COMMIT_no_check_rw|
@@ -182,13 +193,13 @@ btree_write_buffered_insert(struct btree_trans *trans,
 	int ret;
 
 	bch2_trans_iter_init(trans, &iter, wb->btree, bkey_start_pos(&wb->k.k),
-			     BTREE_ITER_CACHED|BTREE_ITER_INTENT);
+			     BTREE_ITER_cached|BTREE_ITER_intent);
 
 	trans->journal_res.seq = wb->journal_seq;
 
 	ret   = bch2_btree_iter_traverse(&iter) ?:
 		bch2_trans_update(trans, &iter, &wb->k,
-				  BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
+				  BTREE_UPDATE_internal_snapshot_node);
 	bch2_trans_iter_exit(trans, &iter);
 	return ret;
 }
@@ -307,13 +318,23 @@ static int bch2_btree_write_buffer_flush_locked(struct btree_trans *trans)
 			    bpos_gt(k->k.k.p, path->l[0].b->key.k.p)) {
 				bch2_btree_node_unlock_write(trans, path, path->l[0].b);
 				write_locked = false;
+
+				ret = lockrestart_do(trans,
+					bch2_btree_iter_traverse(&iter) ?:
+					bch2_foreground_maybe_merge(trans, iter.path, 0,
+							BCH_WATERMARK_reclaim|
+							BCH_TRANS_COMMIT_journal_reclaim|
+							BCH_TRANS_COMMIT_no_check_rw|
+							BCH_TRANS_COMMIT_no_enospc));
+				if (ret)
+					goto err;
 			}
 		}
 
 		if (!iter.path || iter.btree_id != k->btree) {
 			bch2_trans_iter_exit(trans, &iter);
 			bch2_trans_iter_init(trans, &iter, k->btree, k->k.k.p,
-					     BTREE_ITER_INTENT|BTREE_ITER_ALL_SNAPSHOTS);
+					     BTREE_ITER_intent|BTREE_ITER_all_snapshots);
 		}
 
 		bch2_btree_iter_set_pos(&iter, k->k.k.p);
@@ -357,6 +378,11 @@ static int bch2_btree_write_buffer_flush_locked(struct btree_trans *trans)
 		 */
 		trace_and_count(c, write_buffer_flush_slowpath, trans, slowpath, wb->flushing.keys.nr);
 
+		sort(wb->flushing.keys.data,
+		     wb->flushing.keys.nr,
+		     sizeof(wb->flushing.keys.data[0]),
+		     wb_key_seq_cmp, NULL);
+
 		darray_for_each(wb->flushing.keys, i) {
 			if (!i->journal_seq)
 				continue;
@@ -368,17 +394,17 @@ static int bch2_btree_write_buffer_flush_locked(struct btree_trans *trans)
 
 			ret = commit_do(trans, NULL, NULL,
 					BCH_WATERMARK_reclaim|
+					BCH_TRANS_COMMIT_journal_reclaim|
 					BCH_TRANS_COMMIT_no_check_rw|
 					BCH_TRANS_COMMIT_no_enospc|
-					BCH_TRANS_COMMIT_no_journal_res|
-					BCH_TRANS_COMMIT_journal_reclaim,
+					BCH_TRANS_COMMIT_no_journal_res ,
 					btree_write_buffered_insert(trans, i));
 			if (ret)
 				goto err;
 		}
 	}
 err:
-	bch2_fs_fatal_err_on(ret, c, "%s: insert error %s", __func__, bch2_err_str(ret));
+	bch2_fs_fatal_err_on(ret, c, "%s", bch2_err_str(ret));
 	trace_write_buffer_flush(trans, wb->flushing.keys.nr, skipped, fast, 0);
 	bch2_journal_pin_drop(j, &wb->flushing.pin);
 	wb->flushing.keys.nr = 0;
@@ -468,6 +494,41 @@ int bch2_btree_write_buffer_tryflush(struct btree_trans *trans)
 	return ret;
 }
 
+/**
+ * In check and repair code, when checking references to write buffer btrees we
+ * need to issue a flush before we have a definitive error: this issues a flush
+ * if this is a key we haven't yet checked.
+ */
+int bch2_btree_write_buffer_maybe_flush(struct btree_trans *trans,
+					struct bkey_s_c referring_k,
+					struct bkey_buf *last_flushed)
+{
+	struct bch_fs *c = trans->c;
+	struct bkey_buf tmp;
+	int ret = 0;
+
+	bch2_bkey_buf_init(&tmp);
+
+	if (!bkey_and_val_eq(referring_k, bkey_i_to_s_c(last_flushed->k))) {
+		bch2_bkey_buf_reassemble(&tmp, c, referring_k);
+
+		if (bkey_is_btree_ptr(referring_k.k)) {
+			bch2_trans_unlock(trans);
+			bch2_btree_interior_updates_flush(c);
+		}
+
+		ret = bch2_btree_write_buffer_flush_sync(trans);
+		if (ret)
+			goto err;
+
+		bch2_bkey_buf_copy(last_flushed, c, tmp.k);
+		ret = -BCH_ERR_transaction_restart_write_buffer_flush;
+	}
+err:
+	bch2_bkey_buf_exit(&tmp, c);
+	return ret;
+}
+
 static void bch2_btree_write_buffer_flush_work(struct work_struct *work)
 {
 	struct bch_fs *c = container_of(work, struct bch_fs, btree_write_buffer.flush_work);
@@ -574,8 +635,6 @@ void bch2_journal_keys_to_write_buffer_end(struct bch_fs *c, struct journal_keys
 static int bch2_journal_keys_to_write_buffer(struct bch_fs *c, struct journal_buf *buf)
 {
 	struct journal_keys_to_wb dst;
-	struct jset_entry *entry;
-	struct bkey_i *k;
 	int ret = 0;
 
 	bch2_journal_keys_to_write_buffer_start(c, &dst, le64_to_cpu(buf->data->seq));
@@ -590,7 +649,9 @@ static int bch2_journal_keys_to_write_buffer(struct bch_fs *c, struct journal_bu
 		entry->type = BCH_JSET_ENTRY_btree_keys;
 	}
 
+	spin_lock(&c->journal.lock);
 	buf->need_flush_to_write_buffer = false;
+	spin_unlock(&c->journal.lock);
 out:
 	bch2_journal_keys_to_write_buffer_end(c, &dst);
 	return ret;
diff --git a/fs/bcachefs/btree_write_buffer.h b/fs/bcachefs/btree_write_buffer.h
index eebcd2b152..dd5e64218b 100644
--- a/fs/bcachefs/btree_write_buffer.h
+++ b/fs/bcachefs/btree_write_buffer.h
@@ -23,6 +23,9 @@ int bch2_btree_write_buffer_flush_sync(struct btree_trans *);
 int bch2_btree_write_buffer_flush_nocheck_rw(struct btree_trans *);
 int bch2_btree_write_buffer_tryflush(struct btree_trans *);
 
+struct bkey_buf;
+int bch2_btree_write_buffer_maybe_flush(struct btree_trans *, struct bkey_s_c, struct bkey_buf *);
+
 struct journal_keys_to_wb {
 	struct btree_write_buffer_keys	*wb;
 	size_t				room;
diff --git a/fs/bcachefs/buckets.c b/fs/bcachefs/buckets.c
index 54f7826ac4..314ee3e018 100644
--- a/fs/bcachefs/buckets.c
+++ b/fs/bcachefs/buckets.c
@@ -274,25 +274,14 @@ void bch2_dev_usage_init(struct bch_dev *ca)
 
 void bch2_dev_usage_to_text(struct printbuf *out, struct bch_dev_usage *usage)
 {
-	prt_tab(out);
-	prt_str(out, "buckets");
-	prt_tab_rjust(out);
-	prt_str(out, "sectors");
-	prt_tab_rjust(out);
-	prt_str(out, "fragmented");
-	prt_tab_rjust(out);
-	prt_newline(out);
+	prt_printf(out, "\tbuckets\rsectors\rfragmented\r\n");
 
 	for (unsigned i = 0; i < BCH_DATA_NR; i++) {
 		bch2_prt_data_type(out, i);
-		prt_tab(out);
-		prt_u64(out, usage->d[i].buckets);
-		prt_tab_rjust(out);
-		prt_u64(out, usage->d[i].sectors);
-		prt_tab_rjust(out);
-		prt_u64(out, usage->d[i].fragmented);
-		prt_tab_rjust(out);
-		prt_newline(out);
+		prt_printf(out, "\t%llu\r%llu\r%llu\r\n",
+			usage->d[i].buckets,
+			usage->d[i].sectors,
+			usage->d[i].fragmented);
 	}
 }
 
@@ -329,26 +318,6 @@ void bch2_dev_usage_update(struct bch_fs *c, struct bch_dev *ca,
 	preempt_enable();
 }
 
-static inline struct bch_alloc_v4 bucket_m_to_alloc(struct bucket b)
-{
-	return (struct bch_alloc_v4) {
-		.gen		= b.gen,
-		.data_type	= b.data_type,
-		.dirty_sectors	= b.dirty_sectors,
-		.cached_sectors	= b.cached_sectors,
-		.stripe		= b.stripe,
-	};
-}
-
-void bch2_dev_usage_update_m(struct bch_fs *c, struct bch_dev *ca,
-			     struct bucket *old, struct bucket *new)
-{
-	struct bch_alloc_v4 old_a = bucket_m_to_alloc(*old);
-	struct bch_alloc_v4 new_a = bucket_m_to_alloc(*new);
-
-	bch2_dev_usage_update(c, ca, &old_a, &new_a, 0, true);
-}
-
 static inline int __update_replicas(struct bch_fs *c,
 				    struct bch_fs_usage *fs_usage,
 				    struct bch_replicas_entry_v1 *r,
@@ -496,78 +465,305 @@ int bch2_update_cached_sectors_list(struct btree_trans *trans, unsigned dev, s64
 	return bch2_update_replicas_list(trans, &r.e, sectors);
 }
 
-int bch2_mark_metadata_bucket(struct bch_fs *c, struct bch_dev *ca,
-			      size_t b, enum bch_data_type data_type,
-			      unsigned sectors, struct gc_pos pos,
-			      unsigned flags)
+static int bch2_check_fix_ptr(struct btree_trans *trans,
+			      struct bkey_s_c k,
+			      struct extent_ptr_decoded p,
+			      const union bch_extent_entry *entry,
+			      bool *do_update)
 {
-	struct bucket old, new, *g;
+	struct bch_fs *c = trans->c;
+	struct printbuf buf = PRINTBUF;
 	int ret = 0;
 
-	BUG_ON(!(flags & BTREE_TRIGGER_GC));
-	BUG_ON(data_type != BCH_DATA_sb &&
-	       data_type != BCH_DATA_journal);
-
-	/*
-	 * Backup superblock might be past the end of our normal usable space:
-	 */
-	if (b >= ca->mi.nbuckets)
+	struct bch_dev *ca = bch2_dev_tryget(c, p.ptr.dev);
+	if (!ca) {
+		if (fsck_err(c, ptr_to_invalid_device,
+			     "pointer to missing device %u\n"
+			     "while marking %s",
+			     p.ptr.dev,
+			     (printbuf_reset(&buf),
+			      bch2_bkey_val_to_text(&buf, c, k), buf.buf)))
+			*do_update = true;
 		return 0;
+	}
 
-	percpu_down_read(&c->mark_lock);
-	g = gc_bucket(ca, b);
+	struct bucket *g = PTR_GC_BUCKET(ca, &p.ptr);
+	if (!g) {
+		if (fsck_err(c, ptr_to_invalid_device,
+			     "pointer to invalid bucket on device %u\n"
+			     "while marking %s",
+			     p.ptr.dev,
+			     (printbuf_reset(&buf),
+			      bch2_bkey_val_to_text(&buf, c, k), buf.buf)))
+			*do_update = true;
+		goto out;
+	}
 
-	bucket_lock(g);
-	old = *g;
+	enum bch_data_type data_type = bch2_bkey_ptr_data_type(k, p, entry);
 
-	if (bch2_fs_inconsistent_on(g->data_type &&
-			g->data_type != data_type, c,
-			"different types of data in same bucket: %s, %s",
+	if (fsck_err_on(!g->gen_valid,
+			c, ptr_to_missing_alloc_key,
+			"bucket %u:%zu data type %s ptr gen %u missing in alloc btree\n"
+			"while marking %s",
+			p.ptr.dev, PTR_BUCKET_NR(ca, &p.ptr),
+			bch2_data_type_str(ptr_data_type(k.k, &p.ptr)),
+			p.ptr.gen,
+			(printbuf_reset(&buf),
+			 bch2_bkey_val_to_text(&buf, c, k), buf.buf))) {
+		if (!p.ptr.cached) {
+			g->gen_valid		= true;
+			g->gen			= p.ptr.gen;
+		} else {
+			*do_update = true;
+		}
+	}
+
+	if (fsck_err_on(gen_cmp(p.ptr.gen, g->gen) > 0,
+			c, ptr_gen_newer_than_bucket_gen,
+			"bucket %u:%zu data type %s ptr gen in the future: %u > %u\n"
+			"while marking %s",
+			p.ptr.dev, PTR_BUCKET_NR(ca, &p.ptr),
+			bch2_data_type_str(ptr_data_type(k.k, &p.ptr)),
+			p.ptr.gen, g->gen,
+			(printbuf_reset(&buf),
+			 bch2_bkey_val_to_text(&buf, c, k), buf.buf))) {
+		if (!p.ptr.cached &&
+		    (g->data_type != BCH_DATA_btree ||
+		     data_type == BCH_DATA_btree)) {
+			g->gen_valid		= true;
+			g->gen			= p.ptr.gen;
+			g->data_type		= 0;
+			g->dirty_sectors	= 0;
+			g->cached_sectors	= 0;
+		} else {
+			*do_update = true;
+		}
+	}
+
+	if (fsck_err_on(gen_cmp(g->gen, p.ptr.gen) > BUCKET_GC_GEN_MAX,
+			c, ptr_gen_newer_than_bucket_gen,
+			"bucket %u:%zu gen %u data type %s: ptr gen %u too stale\n"
+			"while marking %s",
+			p.ptr.dev, PTR_BUCKET_NR(ca, &p.ptr), g->gen,
+			bch2_data_type_str(ptr_data_type(k.k, &p.ptr)),
+			p.ptr.gen,
+			(printbuf_reset(&buf),
+			 bch2_bkey_val_to_text(&buf, c, k), buf.buf)))
+		*do_update = true;
+
+	if (fsck_err_on(!p.ptr.cached && gen_cmp(p.ptr.gen, g->gen) < 0,
+			c, stale_dirty_ptr,
+			"bucket %u:%zu data type %s stale dirty ptr: %u < %u\n"
+			"while marking %s",
+			p.ptr.dev, PTR_BUCKET_NR(ca, &p.ptr),
+			bch2_data_type_str(ptr_data_type(k.k, &p.ptr)),
+			p.ptr.gen, g->gen,
+			(printbuf_reset(&buf),
+			 bch2_bkey_val_to_text(&buf, c, k), buf.buf)))
+		*do_update = true;
+
+	if (data_type != BCH_DATA_btree && p.ptr.gen != g->gen)
+		goto out;
+
+	if (fsck_err_on(bucket_data_type_mismatch(g->data_type, data_type),
+			c, ptr_bucket_data_type_mismatch,
+			"bucket %u:%zu gen %u different types of data in same bucket: %s, %s\n"
+			"while marking %s",
+			p.ptr.dev, PTR_BUCKET_NR(ca, &p.ptr), g->gen,
 			bch2_data_type_str(g->data_type),
-			bch2_data_type_str(data_type))) {
-		ret = -EIO;
-		goto err;
+			bch2_data_type_str(data_type),
+			(printbuf_reset(&buf),
+			 bch2_bkey_val_to_text(&buf, c, k), buf.buf))) {
+		if (data_type == BCH_DATA_btree) {
+			g->gen_valid		= true;
+			g->gen			= p.ptr.gen;
+			g->data_type		= data_type;
+			g->dirty_sectors	= 0;
+			g->cached_sectors	= 0;
+		} else {
+			*do_update = true;
+		}
 	}
 
-	if (bch2_fs_inconsistent_on((u64) g->dirty_sectors + sectors > ca->mi.bucket_size, c,
-			"bucket %u:%zu gen %u data type %s sector count overflow: %u + %u > bucket size",
-			ca->dev_idx, b, g->gen,
-			bch2_data_type_str(g->data_type ?: data_type),
-			g->dirty_sectors, sectors)) {
-		ret = -EIO;
-		goto err;
+	if (p.has_ec) {
+		struct gc_stripe *m = genradix_ptr(&c->gc_stripes, p.ec.idx);
+
+		if (fsck_err_on(!m || !m->alive,
+				c, ptr_to_missing_stripe,
+				"pointer to nonexistent stripe %llu\n"
+				"while marking %s",
+				(u64) p.ec.idx,
+				(printbuf_reset(&buf),
+				 bch2_bkey_val_to_text(&buf, c, k), buf.buf)))
+			*do_update = true;
+
+		if (fsck_err_on(m && m->alive && !bch2_ptr_matches_stripe_m(m, p),
+				c, ptr_to_incorrect_stripe,
+				"pointer does not match stripe %llu\n"
+				"while marking %s",
+				(u64) p.ec.idx,
+				(printbuf_reset(&buf),
+				 bch2_bkey_val_to_text(&buf, c, k), buf.buf)))
+			*do_update = true;
+	}
+out:
+fsck_err:
+	bch2_dev_put(ca);
+	printbuf_exit(&buf);
+	return ret;
+}
+
+int bch2_check_fix_ptrs(struct btree_trans *trans,
+			enum btree_id btree, unsigned level, struct bkey_s_c k,
+			enum btree_iter_update_trigger_flags flags)
+{
+	struct bch_fs *c = trans->c;
+	struct bkey_ptrs_c ptrs_c = bch2_bkey_ptrs_c(k);
+	const union bch_extent_entry *entry_c;
+	struct extent_ptr_decoded p = { 0 };
+	bool do_update = false;
+	struct printbuf buf = PRINTBUF;
+	int ret = 0;
+
+	percpu_down_read(&c->mark_lock);
+
+	bkey_for_each_ptr_decode(k.k, ptrs_c, p, entry_c) {
+		ret = bch2_check_fix_ptr(trans, k, p, entry_c, &do_update);
+		if (ret)
+			goto err;
 	}
 
-	g->data_type = data_type;
-	g->dirty_sectors += sectors;
-	new = *g;
+	if (do_update) {
+		if (flags & BTREE_TRIGGER_is_root) {
+			bch_err(c, "cannot update btree roots yet");
+			ret = -EINVAL;
+			goto err;
+		}
+
+		struct bkey_i *new = bch2_bkey_make_mut_noupdate(trans, k);
+		ret = PTR_ERR_OR_ZERO(new);
+		if (ret)
+			goto err;
+
+		rcu_read_lock();
+		bch2_bkey_drop_ptrs(bkey_i_to_s(new), ptr, !bch2_dev_rcu(c, ptr->dev));
+		rcu_read_unlock();
+
+		if (level) {
+			/*
+			 * We don't want to drop btree node pointers - if the
+			 * btree node isn't there anymore, the read path will
+			 * sort it out:
+			 */
+			struct bkey_ptrs ptrs = bch2_bkey_ptrs(bkey_i_to_s(new));
+			rcu_read_lock();
+			bkey_for_each_ptr(ptrs, ptr) {
+				struct bch_dev *ca = bch2_dev_rcu(c, ptr->dev);
+				struct bucket *g = PTR_GC_BUCKET(ca, ptr);
+
+				ptr->gen = g->gen;
+			}
+			rcu_read_unlock();
+		} else {
+			struct bkey_ptrs ptrs;
+			union bch_extent_entry *entry;
+
+			rcu_read_lock();
+restart_drop_ptrs:
+			ptrs = bch2_bkey_ptrs(bkey_i_to_s(new));
+			bkey_for_each_ptr_decode(bkey_i_to_s(new).k, ptrs, p, entry) {
+				struct bch_dev *ca = bch2_dev_rcu(c, p.ptr.dev);
+				struct bucket *g = PTR_GC_BUCKET(ca, &p.ptr);
+				enum bch_data_type data_type = bch2_bkey_ptr_data_type(bkey_i_to_s_c(new), p, entry);
+
+				if ((p.ptr.cached &&
+				     (!g->gen_valid || gen_cmp(p.ptr.gen, g->gen) > 0)) ||
+				    (!p.ptr.cached &&
+				     gen_cmp(p.ptr.gen, g->gen) < 0) ||
+				    gen_cmp(g->gen, p.ptr.gen) > BUCKET_GC_GEN_MAX ||
+				    (g->data_type &&
+				     g->data_type != data_type)) {
+					bch2_bkey_drop_ptr(bkey_i_to_s(new), &entry->ptr);
+					goto restart_drop_ptrs;
+				}
+			}
+			rcu_read_unlock();
+again:
+			ptrs = bch2_bkey_ptrs(bkey_i_to_s(new));
+			bkey_extent_entry_for_each(ptrs, entry) {
+				if (extent_entry_type(entry) == BCH_EXTENT_ENTRY_stripe_ptr) {
+					struct gc_stripe *m = genradix_ptr(&c->gc_stripes,
+									entry->stripe_ptr.idx);
+					union bch_extent_entry *next_ptr;
+
+					bkey_extent_entry_for_each_from(ptrs, next_ptr, entry)
+						if (extent_entry_type(next_ptr) == BCH_EXTENT_ENTRY_ptr)
+							goto found;
+					next_ptr = NULL;
+found:
+					if (!next_ptr) {
+						bch_err(c, "aieee, found stripe ptr with no data ptr");
+						continue;
+					}
+
+					if (!m || !m->alive ||
+					    !__bch2_ptr_matches_stripe(&m->ptrs[entry->stripe_ptr.block],
+								       &next_ptr->ptr,
+								       m->sectors)) {
+						bch2_bkey_extent_entry_drop(new, entry);
+						goto again;
+					}
+				}
+			}
+		}
+
+		if (0) {
+			printbuf_reset(&buf);
+			bch2_bkey_val_to_text(&buf, c, k);
+			bch_info(c, "updated %s", buf.buf);
+
+			printbuf_reset(&buf);
+			bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(new));
+			bch_info(c, "new key %s", buf.buf);
+		}
+
+		percpu_up_read(&c->mark_lock);
+		struct btree_iter iter;
+		bch2_trans_node_iter_init(trans, &iter, btree, new->k.p, 0, level,
+					  BTREE_ITER_intent|BTREE_ITER_all_snapshots);
+		ret =   bch2_btree_iter_traverse(&iter) ?:
+			bch2_trans_update(trans, &iter, new,
+					  BTREE_UPDATE_internal_snapshot_node|
+					  BTREE_TRIGGER_norun);
+		bch2_trans_iter_exit(trans, &iter);
+		percpu_down_read(&c->mark_lock);
+
+		if (ret)
+			goto err;
+
+		if (level)
+			bch2_btree_node_update_key_early(trans, btree, level - 1, k, new);
+	}
 err:
-	bucket_unlock(g);
-	if (!ret)
-		bch2_dev_usage_update_m(c, ca, &old, &new);
 	percpu_up_read(&c->mark_lock);
+	printbuf_exit(&buf);
 	return ret;
 }
 
-int bch2_check_bucket_ref(struct btree_trans *trans,
-			  struct bkey_s_c k,
-			  const struct bch_extent_ptr *ptr,
-			  s64 sectors, enum bch_data_type ptr_data_type,
-			  u8 b_gen, u8 bucket_data_type,
-			  u32 bucket_sectors)
+int bch2_bucket_ref_update(struct btree_trans *trans, struct bch_dev *ca,
+			   struct bkey_s_c k,
+			   const struct bch_extent_ptr *ptr,
+			   s64 sectors, enum bch_data_type ptr_data_type,
+			   u8 b_gen, u8 bucket_data_type,
+			   u32 *bucket_sectors)
 {
 	struct bch_fs *c = trans->c;
-	struct bch_dev *ca = bch_dev_bkey_exists(c, ptr->dev);
 	size_t bucket_nr = PTR_BUCKET_NR(ca, ptr);
 	struct printbuf buf = PRINTBUF;
+	bool inserting = sectors > 0;
 	int ret = 0;
 
-	if (bucket_data_type == BCH_DATA_cached)
-		bucket_data_type = BCH_DATA_user;
-
-	if ((bucket_data_type == BCH_DATA_stripe && ptr_data_type == BCH_DATA_user) ||
-	    (bucket_data_type == BCH_DATA_user   && ptr_data_type == BCH_DATA_stripe))
-		bucket_data_type = ptr_data_type = BCH_DATA_stripe;
+	BUG_ON(!sectors);
 
 	if (gen_after(ptr->gen, b_gen)) {
 		bch2_fsck_err(c, FSCK_CAN_IGNORE|FSCK_NEED_FSCK,
@@ -578,8 +774,9 @@ int bch2_check_bucket_ref(struct btree_trans *trans,
 			bch2_data_type_str(bucket_data_type ?: ptr_data_type),
 			ptr->gen,
 			(bch2_bkey_val_to_text(&buf, c, k), buf.buf));
-		ret = -EIO;
-		goto err;
+		if (inserting)
+			goto err;
+		goto out;
 	}
 
 	if (gen_cmp(b_gen, ptr->gen) > BUCKET_GC_GEN_MAX) {
@@ -592,33 +789,33 @@ int bch2_check_bucket_ref(struct btree_trans *trans,
 			ptr->gen,
 			(printbuf_reset(&buf),
 			 bch2_bkey_val_to_text(&buf, c, k), buf.buf));
-		ret = -EIO;
-		goto err;
+		if (inserting)
+			goto err;
+		goto out;
 	}
 
-	if (b_gen != ptr->gen && !ptr->cached) {
+	if (b_gen != ptr->gen && ptr->cached) {
+		ret = 1;
+		goto out;
+	}
+
+	if (b_gen != ptr->gen) {
 		bch2_fsck_err(c, FSCK_CAN_IGNORE|FSCK_NEED_FSCK,
 			      BCH_FSCK_ERR_stale_dirty_ptr,
 			"bucket %u:%zu gen %u (mem gen %u) data type %s: stale dirty ptr (gen %u)\n"
 			"while marking %s",
 			ptr->dev, bucket_nr, b_gen,
-			*bucket_gen(ca, bucket_nr),
+			bucket_gen_get(ca, bucket_nr),
 			bch2_data_type_str(bucket_data_type ?: ptr_data_type),
 			ptr->gen,
 			(printbuf_reset(&buf),
 			 bch2_bkey_val_to_text(&buf, c, k), buf.buf));
-		ret = -EIO;
-		goto err;
-	}
-
-	if (b_gen != ptr->gen) {
-		ret = 1;
+		if (inserting)
+			goto err;
 		goto out;
 	}
 
-	if (!data_type_is_empty(bucket_data_type) &&
-	    ptr_data_type &&
-	    bucket_data_type != ptr_data_type) {
+	if (bucket_data_type_mismatch(bucket_data_type, ptr_data_type)) {
 		bch2_fsck_err(c, FSCK_CAN_IGNORE|FSCK_NEED_FSCK,
 			      BCH_FSCK_ERR_ptr_bucket_data_type_mismatch,
 			"bucket %u:%zu gen %u different types of data in same bucket: %s, %s\n"
@@ -628,28 +825,33 @@ int bch2_check_bucket_ref(struct btree_trans *trans,
 			bch2_data_type_str(ptr_data_type),
 			(printbuf_reset(&buf),
 			 bch2_bkey_val_to_text(&buf, c, k), buf.buf));
-		ret = -EIO;
-		goto err;
+		if (inserting)
+			goto err;
+		goto out;
 	}
 
-	if ((u64) bucket_sectors + sectors > U32_MAX) {
+	if ((u64) *bucket_sectors + sectors > U32_MAX) {
 		bch2_fsck_err(c, FSCK_CAN_IGNORE|FSCK_NEED_FSCK,
 			      BCH_FSCK_ERR_bucket_sector_count_overflow,
 			"bucket %u:%zu gen %u data type %s sector count overflow: %u + %lli > U32_MAX\n"
 			"while marking %s",
 			ptr->dev, bucket_nr, b_gen,
 			bch2_data_type_str(bucket_data_type ?: ptr_data_type),
-			bucket_sectors, sectors,
+			*bucket_sectors, sectors,
 			(printbuf_reset(&buf),
 			 bch2_bkey_val_to_text(&buf, c, k), buf.buf));
-		ret = -EIO;
-		goto err;
+		if (inserting)
+			goto err;
+		sectors = -*bucket_sectors;
 	}
+
+	*bucket_sectors += sectors;
 out:
 	printbuf_exit(&buf);
 	return ret;
 err:
 	bch2_dump_trans_updates(trans);
+	ret = -EIO;
 	goto out;
 }
 
@@ -786,111 +988,102 @@ need_mark:
 
 /* KEY_TYPE_extent: */
 
-static int __mark_pointer(struct btree_trans *trans,
+static int __mark_pointer(struct btree_trans *trans, struct bch_dev *ca,
 			  struct bkey_s_c k,
 			  const struct bch_extent_ptr *ptr,
 			  s64 sectors, enum bch_data_type ptr_data_type,
-			  u8 bucket_gen, u8 *bucket_data_type,
-			  u32 *dirty_sectors, u32 *cached_sectors)
+			  struct bch_alloc_v4 *a)
 {
 	u32 *dst_sectors = !ptr->cached
-		? dirty_sectors
-		: cached_sectors;
-	int ret = bch2_check_bucket_ref(trans, k, ptr, sectors, ptr_data_type,
-				   bucket_gen, *bucket_data_type, *dst_sectors);
+		? &a->dirty_sectors
+		: &a->cached_sectors;
+	int ret = bch2_bucket_ref_update(trans, ca, k, ptr, sectors, ptr_data_type,
+					 a->gen, a->data_type, dst_sectors);
 
 	if (ret)
 		return ret;
 
-	*dst_sectors += sectors;
-
-	if (!*dirty_sectors && !*cached_sectors)
-		*bucket_data_type = 0;
-	else if (*bucket_data_type != BCH_DATA_stripe)
-		*bucket_data_type = ptr_data_type;
-
+	alloc_data_type_set(a, ptr_data_type);
 	return 0;
 }
 
 static int bch2_trigger_pointer(struct btree_trans *trans,
 			enum btree_id btree_id, unsigned level,
 			struct bkey_s_c k, struct extent_ptr_decoded p,
+			const union bch_extent_entry *entry,
 			s64 *sectors,
-			unsigned flags)
+			enum btree_iter_update_trigger_flags flags)
 {
-	bool insert = !(flags & BTREE_TRIGGER_OVERWRITE);
+	bool insert = !(flags & BTREE_TRIGGER_overwrite);
+	struct printbuf buf = PRINTBUF;
+	int ret = 0;
+
+	struct bch_fs *c = trans->c;
+	struct bch_dev *ca = bch2_dev_tryget(c, p.ptr.dev);
+	if (unlikely(!ca)) {
+		if (insert)
+			ret = -EIO;
+		goto err;
+	}
+
 	struct bpos bucket;
 	struct bch_backpointer bp;
-
-	bch2_extent_ptr_to_bp(trans->c, btree_id, level, k, p, &bucket, &bp);
+	bch2_extent_ptr_to_bp(trans->c, ca, btree_id, level, k, p, entry, &bucket, &bp);
 	*sectors = insert ? bp.bucket_len : -((s64) bp.bucket_len);
 
-	if (flags & BTREE_TRIGGER_TRANSACTIONAL) {
-		struct btree_iter iter;
-		struct bkey_i_alloc_v4 *a = bch2_trans_start_alloc_update(trans, &iter, bucket);
-		int ret = PTR_ERR_OR_ZERO(a);
+	if (flags & BTREE_TRIGGER_transactional) {
+		struct bkey_i_alloc_v4 *a = bch2_trans_start_alloc_update(trans, bucket);
+		ret = PTR_ERR_OR_ZERO(a) ?:
+			__mark_pointer(trans, ca, k, &p.ptr, *sectors, bp.data_type, &a->v);
 		if (ret)
-			return ret;
-
-		ret = __mark_pointer(trans, k, &p.ptr, *sectors, bp.data_type,
-				     a->v.gen, &a->v.data_type,
-				     &a->v.dirty_sectors, &a->v.cached_sectors) ?:
-			bch2_trans_update(trans, &iter, &a->k_i, 0);
-		bch2_trans_iter_exit(trans, &iter);
-
-		if (ret)
-			return ret;
+			goto err;
 
 		if (!p.ptr.cached) {
-			ret = bch2_bucket_backpointer_mod(trans, bucket, bp, k, insert);
+			ret = bch2_bucket_backpointer_mod(trans, ca, bucket, bp, k, insert);
 			if (ret)
-				return ret;
+				goto err;
 		}
 	}
 
-	if (flags & BTREE_TRIGGER_GC) {
-		struct bch_fs *c = trans->c;
-		struct bch_dev *ca = bch_dev_bkey_exists(c, p.ptr.dev);
-		enum bch_data_type data_type = bkey_ptr_data_type(btree_id, level, k, p);
-
+	if (flags & BTREE_TRIGGER_gc) {
 		percpu_down_read(&c->mark_lock);
-		struct bucket *g = PTR_GC_BUCKET(ca, &p.ptr);
-		bucket_lock(g);
-		struct bucket old = *g;
-
-		u8 bucket_data_type = g->data_type;
-		int ret = __mark_pointer(trans, k, &p.ptr, *sectors,
-				     data_type, g->gen,
-				     &bucket_data_type,
-				     &g->dirty_sectors,
-				     &g->cached_sectors);
-		if (ret) {
-			bucket_unlock(g);
-			percpu_up_read(&c->mark_lock);
-			return ret;
+		struct bucket *g = gc_bucket(ca, bucket.offset);
+		if (bch2_fs_inconsistent_on(!g, c, "reference to invalid bucket on device %u\n  %s",
+					    p.ptr.dev,
+					    (bch2_bkey_val_to_text(&buf, c, k), buf.buf))) {
+			ret = -EIO;
+			goto err_unlock;
 		}
 
-		g->data_type = bucket_data_type;
-		struct bucket new = *g;
+		bucket_lock(g);
+		struct bch_alloc_v4 old = bucket_m_to_alloc(*g), new = old;
+		ret = __mark_pointer(trans, ca, k, &p.ptr, *sectors, bp.data_type, &new);
+		if (!ret) {
+			alloc_to_bucket(g, new);
+			bch2_dev_usage_update(c, ca, &old, &new, 0, true);
+		}
 		bucket_unlock(g);
-		bch2_dev_usage_update_m(c, ca, &old, &new);
+err_unlock:
 		percpu_up_read(&c->mark_lock);
 	}
-
-	return 0;
+err:
+	bch2_dev_put(ca);
+	printbuf_exit(&buf);
+	return ret;
 }
 
 static int bch2_trigger_stripe_ptr(struct btree_trans *trans,
 				struct bkey_s_c k,
 				struct extent_ptr_decoded p,
 				enum bch_data_type data_type,
-				s64 sectors, unsigned flags)
+				s64 sectors,
+				enum btree_iter_update_trigger_flags flags)
 {
-	if (flags & BTREE_TRIGGER_TRANSACTIONAL) {
+	if (flags & BTREE_TRIGGER_transactional) {
 		struct btree_iter iter;
 		struct bkey_i_stripe *s = bch2_bkey_get_mut_typed(trans, &iter,
 				BTREE_ID_stripes, POS(0, p.ec.idx),
-				BTREE_ITER_WITH_UPDATES, stripe);
+				BTREE_ITER_with_updates, stripe);
 		int ret = PTR_ERR_OR_ZERO(s);
 		if (unlikely(ret)) {
 			bch2_trans_inconsistent_on(bch2_err_matches(ret, ENOENT), trans,
@@ -920,10 +1113,10 @@ err:
 		return ret;
 	}
 
-	if (flags & BTREE_TRIGGER_GC) {
+	if (flags & BTREE_TRIGGER_gc) {
 		struct bch_fs *c = trans->c;
 
-		BUG_ON(!(flags & BTREE_TRIGGER_GC));
+		BUG_ON(!(flags & BTREE_TRIGGER_gc));
 
 		struct gc_stripe *m = genradix_ptr_alloc(&c->gc_stripes, p.ec.idx, GFP_KERNEL);
 		if (!m) {
@@ -959,9 +1152,10 @@ err:
 
 static int __trigger_extent(struct btree_trans *trans,
 			    enum btree_id btree_id, unsigned level,
-			    struct bkey_s_c k, unsigned flags)
+			    struct bkey_s_c k,
+			    enum btree_iter_update_trigger_flags flags)
 {
-	bool gc = flags & BTREE_TRIGGER_GC;
+	bool gc = flags & BTREE_TRIGGER_gc;
 	struct bch_fs *c = trans->c;
 	struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k);
 	const union bch_extent_entry *entry;
@@ -970,7 +1164,7 @@ static int __trigger_extent(struct btree_trans *trans,
 	enum bch_data_type data_type = bkey_is_btree_ptr(k.k)
 		? BCH_DATA_btree
 		: BCH_DATA_user;
-	s64 dirty_sectors = 0;
+	s64 replicas_sectors = 0;
 	int ret = 0;
 
 	r.e.data_type	= data_type;
@@ -978,8 +1172,8 @@ static int __trigger_extent(struct btree_trans *trans,
 	r.e.nr_required	= 1;
 
 	bkey_for_each_ptr_decode(k.k, ptrs, p, entry) {
-		s64 disk_sectors;
-		ret = bch2_trigger_pointer(trans, btree_id, level, k, p, &disk_sectors, flags);
+		s64 disk_sectors = 0;
+		ret = bch2_trigger_pointer(trans, btree_id, level, k, p, entry, &disk_sectors, flags);
 		if (ret < 0)
 			return ret;
 
@@ -990,13 +1184,13 @@ static int __trigger_extent(struct btree_trans *trans,
 				ret = !gc
 					? bch2_update_cached_sectors_list(trans, p.ptr.dev, disk_sectors)
 					: update_cached_sectors(c, k, p.ptr.dev, disk_sectors, 0, true);
-				bch2_fs_fatal_err_on(ret && gc, c, "%s(): no replicas entry while updating cached sectors",
-						     __func__);
+				bch2_fs_fatal_err_on(ret && gc, c, "%s: no replicas entry while updating cached sectors",
+						     bch2_err_str(ret));
 				if (ret)
 					return ret;
 			}
 		} else if (!p.has_ec) {
-			dirty_sectors	       += disk_sectors;
+			replicas_sectors       += disk_sectors;
 			r.e.devs[r.e.nr_devs++]	= p.ptr.dev;
 		} else {
 			ret = bch2_trigger_stripe_ptr(trans, k, p, data_type, disk_sectors, flags);
@@ -1014,13 +1208,13 @@ static int __trigger_extent(struct btree_trans *trans,
 
 	if (r.e.nr_devs) {
 		ret = !gc
-			? bch2_update_replicas_list(trans, &r.e, dirty_sectors)
-			: bch2_update_replicas(c, k, &r.e, dirty_sectors, 0, true);
+			? bch2_update_replicas_list(trans, &r.e, replicas_sectors)
+			: bch2_update_replicas(c, k, &r.e, replicas_sectors, 0, true);
 		if (unlikely(ret && gc)) {
 			struct printbuf buf = PRINTBUF;
 
 			bch2_bkey_val_to_text(&buf, c, k);
-			bch2_fs_fatal_error(c, "%s(): no replicas entry for %s", __func__, buf.buf);
+			bch2_fs_fatal_error(c, ": no replicas entry for %s", buf.buf);
 			printbuf_exit(&buf);
 		}
 		if (ret)
@@ -1031,15 +1225,18 @@ static int __trigger_extent(struct btree_trans *trans,
 }
 
 int bch2_trigger_extent(struct btree_trans *trans,
-			enum btree_id btree_id, unsigned level,
+			enum btree_id btree, unsigned level,
 			struct bkey_s_c old, struct bkey_s new,
-			unsigned flags)
+			enum btree_iter_update_trigger_flags flags)
 {
 	struct bkey_ptrs_c new_ptrs = bch2_bkey_ptrs_c(new.s_c);
 	struct bkey_ptrs_c old_ptrs = bch2_bkey_ptrs_c(old);
 	unsigned new_ptrs_bytes = (void *) new_ptrs.end - (void *) new_ptrs.start;
 	unsigned old_ptrs_bytes = (void *) old_ptrs.end - (void *) old_ptrs.start;
 
+	if (unlikely(flags & BTREE_TRIGGER_check_repair))
+		return bch2_check_fix_ptrs(trans, btree, level, new.s_c, flags);
+
 	/* if pointers aren't changing - nothing to do: */
 	if (new_ptrs_bytes == old_ptrs_bytes &&
 	    !memcmp(new_ptrs.start,
@@ -1047,20 +1244,21 @@ int bch2_trigger_extent(struct btree_trans *trans,
 		    new_ptrs_bytes))
 		return 0;
 
-	if (flags & BTREE_TRIGGER_TRANSACTIONAL) {
+	if (flags & BTREE_TRIGGER_transactional) {
 		struct bch_fs *c = trans->c;
 		int mod = (int) bch2_bkey_needs_rebalance(c, new.s_c) -
 			  (int) bch2_bkey_needs_rebalance(c, old);
 
 		if (mod) {
-			int ret = bch2_btree_bit_mod(trans, BTREE_ID_rebalance_work, new.k->p, mod > 0);
+			int ret = bch2_btree_bit_mod_buffered(trans, BTREE_ID_rebalance_work,
+							      new.k->p, mod > 0);
 			if (ret)
 				return ret;
 		}
 	}
 
-	if (flags & (BTREE_TRIGGER_TRANSACTIONAL|BTREE_TRIGGER_GC))
-		return trigger_run_overwrite_then_insert(__trigger_extent, trans, btree_id, level, old, new, flags);
+	if (flags & (BTREE_TRIGGER_transactional|BTREE_TRIGGER_gc))
+		return trigger_run_overwrite_then_insert(__trigger_extent, trans, btree, level, old, new, flags);
 
 	return 0;
 }
@@ -1068,17 +1266,17 @@ int bch2_trigger_extent(struct btree_trans *trans,
 /* KEY_TYPE_reservation */
 
 static int __trigger_reservation(struct btree_trans *trans,
-				 enum btree_id btree_id, unsigned level,
-				 struct bkey_s_c k, unsigned flags)
+			enum btree_id btree_id, unsigned level, struct bkey_s_c k,
+			enum btree_iter_update_trigger_flags flags)
 {
 	struct bch_fs *c = trans->c;
 	unsigned replicas = bkey_s_c_to_reservation(k).v->nr_replicas;
 	s64 sectors = (s64) k.k->size * replicas;
 
-	if (flags & BTREE_TRIGGER_OVERWRITE)
+	if (flags & BTREE_TRIGGER_overwrite)
 		sectors = -sectors;
 
-	if (flags & BTREE_TRIGGER_TRANSACTIONAL) {
+	if (flags & BTREE_TRIGGER_transactional) {
 		int ret = bch2_replicas_deltas_realloc(trans, 0);
 		if (ret)
 			return ret;
@@ -1089,7 +1287,7 @@ static int __trigger_reservation(struct btree_trans *trans,
 		d->persistent_reserved[replicas - 1] += sectors;
 	}
 
-	if (flags & BTREE_TRIGGER_GC) {
+	if (flags & BTREE_TRIGGER_gc) {
 		percpu_down_read(&c->mark_lock);
 		preempt_disable();
 
@@ -1109,7 +1307,7 @@ static int __trigger_reservation(struct btree_trans *trans,
 int bch2_trigger_reservation(struct btree_trans *trans,
 			  enum btree_id btree_id, unsigned level,
 			  struct bkey_s_c old, struct bkey_s new,
-			  unsigned flags)
+			  enum btree_iter_update_trigger_flags flags)
 {
 	return trigger_run_overwrite_then_insert(__trigger_reservation, trans, btree_id, level, old, new, flags);
 }
@@ -1117,22 +1315,16 @@ int bch2_trigger_reservation(struct btree_trans *trans,
 /* Mark superblocks: */
 
 static int __bch2_trans_mark_metadata_bucket(struct btree_trans *trans,
-				    struct bch_dev *ca, size_t b,
+				    struct bch_dev *ca, u64 b,
 				    enum bch_data_type type,
 				    unsigned sectors)
 {
 	struct bch_fs *c = trans->c;
 	struct btree_iter iter;
-	struct bkey_i_alloc_v4 *a;
 	int ret = 0;
 
-	/*
-	 * Backup superblock might be past the end of our normal usable space:
-	 */
-	if (b >= ca->mi.nbuckets)
-		return 0;
-
-	a = bch2_trans_start_alloc_update(trans, &iter, POS(ca->dev_idx, b));
+	struct bkey_i_alloc_v4 *a =
+		bch2_trans_start_alloc_update_noupdate(trans, &iter, POS(ca->dev_idx, b));
 	if (IS_ERR(a))
 		return PTR_ERR(a);
 
@@ -1160,20 +1352,74 @@ err:
 	return ret;
 }
 
+static int bch2_mark_metadata_bucket(struct bch_fs *c, struct bch_dev *ca,
+			u64 b, enum bch_data_type data_type, unsigned sectors,
+			enum btree_iter_update_trigger_flags flags)
+{
+	percpu_down_read(&c->mark_lock);
+	struct bucket *g = gc_bucket(ca, b);
+	if (bch2_fs_inconsistent_on(!g, c, "reference to invalid bucket on device %u when marking metadata type %s",
+				    ca->dev_idx, bch2_data_type_str(data_type)))
+		goto err_unlock;
+
+	bucket_lock(g);
+	struct bch_alloc_v4 old = bucket_m_to_alloc(*g);
+
+	if (bch2_fs_inconsistent_on(g->data_type &&
+			g->data_type != data_type, c,
+			"different types of data in same bucket: %s, %s",
+			bch2_data_type_str(g->data_type),
+			bch2_data_type_str(data_type)))
+		goto err;
+
+	if (bch2_fs_inconsistent_on((u64) g->dirty_sectors + sectors > ca->mi.bucket_size, c,
+			"bucket %u:%llu gen %u data type %s sector count overflow: %u + %u > bucket size",
+			ca->dev_idx, b, g->gen,
+			bch2_data_type_str(g->data_type ?: data_type),
+			g->dirty_sectors, sectors))
+		goto err;
+
+	g->data_type = data_type;
+	g->dirty_sectors += sectors;
+	struct bch_alloc_v4 new = bucket_m_to_alloc(*g);
+	bch2_dev_usage_update(c, ca, &old, &new, 0, true);
+	percpu_up_read(&c->mark_lock);
+	return 0;
+err:
+	bucket_unlock(g);
+err_unlock:
+	percpu_up_read(&c->mark_lock);
+	return -EIO;
+}
+
 int bch2_trans_mark_metadata_bucket(struct btree_trans *trans,
-				    struct bch_dev *ca, size_t b,
-				    enum bch_data_type type,
-				    unsigned sectors)
+			struct bch_dev *ca, u64 b,
+			enum bch_data_type type, unsigned sectors,
+			enum btree_iter_update_trigger_flags flags)
 {
-	return commit_do(trans, NULL, NULL, 0,
-			__bch2_trans_mark_metadata_bucket(trans, ca, b, type, sectors));
+	BUG_ON(type != BCH_DATA_free &&
+	       type != BCH_DATA_sb &&
+	       type != BCH_DATA_journal);
+
+	/*
+	 * Backup superblock might be past the end of our normal usable space:
+	 */
+	if (b >= ca->mi.nbuckets)
+		return 0;
+
+	if (flags & BTREE_TRIGGER_gc)
+		return bch2_mark_metadata_bucket(trans->c, ca, b, type, sectors, flags);
+	else if (flags & BTREE_TRIGGER_transactional)
+		return commit_do(trans, NULL, NULL, 0,
+				 __bch2_trans_mark_metadata_bucket(trans, ca, b, type, sectors));
+	else
+		BUG();
 }
 
 static int bch2_trans_mark_metadata_sectors(struct btree_trans *trans,
-					    struct bch_dev *ca,
-					    u64 start, u64 end,
-					    enum bch_data_type type,
-					    u64 *bucket, unsigned *bucket_sectors)
+			struct bch_dev *ca, u64 start, u64 end,
+			enum bch_data_type type, u64 *bucket, unsigned *bucket_sectors,
+			enum btree_iter_update_trigger_flags flags)
 {
 	do {
 		u64 b = sector_to_bucket(ca, start);
@@ -1182,7 +1428,7 @@ static int bch2_trans_mark_metadata_sectors(struct btree_trans *trans,
 
 		if (b != *bucket && *bucket_sectors) {
 			int ret = bch2_trans_mark_metadata_bucket(trans, ca, *bucket,
-								  type, *bucket_sectors);
+							type, *bucket_sectors, flags);
 			if (ret)
 				return ret;
 
@@ -1197,8 +1443,8 @@ static int bch2_trans_mark_metadata_sectors(struct btree_trans *trans,
 	return 0;
 }
 
-static int __bch2_trans_mark_dev_sb(struct btree_trans *trans,
-				    struct bch_dev *ca)
+static int __bch2_trans_mark_dev_sb(struct btree_trans *trans, struct bch_dev *ca,
+			enum btree_iter_update_trigger_flags flags)
 {
 	struct bch_sb_layout *layout = &ca->disk_sb.sb->layout;
 	u64 bucket = 0;
@@ -1211,21 +1457,21 @@ static int __bch2_trans_mark_dev_sb(struct btree_trans *trans,
 		if (offset == BCH_SB_SECTOR) {
 			ret = bch2_trans_mark_metadata_sectors(trans, ca,
 						0, BCH_SB_SECTOR,
-						BCH_DATA_sb, &bucket, &bucket_sectors);
+						BCH_DATA_sb, &bucket, &bucket_sectors, flags);
 			if (ret)
 				return ret;
 		}
 
 		ret = bch2_trans_mark_metadata_sectors(trans, ca, offset,
 				      offset + (1 << layout->sb_max_size_bits),
-				      BCH_DATA_sb, &bucket, &bucket_sectors);
+				      BCH_DATA_sb, &bucket, &bucket_sectors, flags);
 		if (ret)
 			return ret;
 	}
 
 	if (bucket_sectors) {
 		ret = bch2_trans_mark_metadata_bucket(trans, ca,
-				bucket, BCH_DATA_sb, bucket_sectors);
+				bucket, BCH_DATA_sb, bucket_sectors, flags);
 		if (ret)
 			return ret;
 	}
@@ -1233,7 +1479,7 @@ static int __bch2_trans_mark_dev_sb(struct btree_trans *trans,
 	for (i = 0; i < ca->journal.nr; i++) {
 		ret = bch2_trans_mark_metadata_bucket(trans, ca,
 				ca->journal.buckets[i],
-				BCH_DATA_journal, ca->mi.bucket_size);
+				BCH_DATA_journal, ca->mi.bucket_size, flags);
 		if (ret)
 			return ret;
 	}
@@ -1241,20 +1487,22 @@ static int __bch2_trans_mark_dev_sb(struct btree_trans *trans,
 	return 0;
 }
 
-int bch2_trans_mark_dev_sb(struct bch_fs *c, struct bch_dev *ca)
+int bch2_trans_mark_dev_sb(struct bch_fs *c, struct bch_dev *ca,
+			enum btree_iter_update_trigger_flags flags)
 {
-	int ret = bch2_trans_run(c, __bch2_trans_mark_dev_sb(trans, ca));
-
+	int ret = bch2_trans_run(c,
+		__bch2_trans_mark_dev_sb(trans, ca, flags));
 	bch_err_fn(c, ret);
 	return ret;
 }
 
-int bch2_trans_mark_dev_sbs(struct bch_fs *c)
+int bch2_trans_mark_dev_sbs_flags(struct bch_fs *c,
+			enum btree_iter_update_trigger_flags flags)
 {
 	for_each_online_member(c, ca) {
-		int ret = bch2_trans_mark_dev_sb(c, ca);
+		int ret = bch2_trans_mark_dev_sb(c, ca, flags);
 		if (ret) {
-			percpu_ref_put(&ca->ref);
+			percpu_ref_put(&ca->io_ref);
 			return ret;
 		}
 	}
@@ -1262,6 +1510,11 @@ int bch2_trans_mark_dev_sbs(struct bch_fs *c)
 	return 0;
 }
 
+int bch2_trans_mark_dev_sbs(struct bch_fs *c)
+{
+	return bch2_trans_mark_dev_sbs_flags(c, BTREE_TRIGGER_transactional);
+}
+
 /* Disk reservations: */
 
 #define SECTORS_CACHE	1024
@@ -1330,37 +1583,57 @@ recalculate:
 
 /* Startup/shutdown: */
 
+void bch2_buckets_nouse_free(struct bch_fs *c)
+{
+	for_each_member_device(c, ca) {
+		kvfree_rcu_mightsleep(ca->buckets_nouse);
+		ca->buckets_nouse = NULL;
+	}
+}
+
+int bch2_buckets_nouse_alloc(struct bch_fs *c)
+{
+	for_each_member_device(c, ca) {
+		BUG_ON(ca->buckets_nouse);
+
+		ca->buckets_nouse = kvmalloc(BITS_TO_LONGS(ca->mi.nbuckets) *
+					    sizeof(unsigned long),
+					    GFP_KERNEL|__GFP_ZERO);
+		if (!ca->buckets_nouse) {
+			bch2_dev_put(ca);
+			return -BCH_ERR_ENOMEM_buckets_nouse;
+		}
+	}
+
+	return 0;
+}
+
 static void bucket_gens_free_rcu(struct rcu_head *rcu)
 {
 	struct bucket_gens *buckets =
 		container_of(rcu, struct bucket_gens, rcu);
 
-	kvpfree(buckets, sizeof(*buckets) + buckets->nbuckets);
+	kvfree(buckets);
 }
 
 int bch2_dev_buckets_resize(struct bch_fs *c, struct bch_dev *ca, u64 nbuckets)
 {
 	struct bucket_gens *bucket_gens = NULL, *old_bucket_gens = NULL;
-	unsigned long *buckets_nouse = NULL;
 	bool resize = ca->bucket_gens != NULL;
 	int ret;
 
-	if (!(bucket_gens	= kvpmalloc(sizeof(struct bucket_gens) + nbuckets,
-					    GFP_KERNEL|__GFP_ZERO))) {
-		ret = -BCH_ERR_ENOMEM_bucket_gens;
-		goto err;
-	}
+	BUG_ON(resize && ca->buckets_nouse);
 
-	if ((c->opts.buckets_nouse &&
-	     !(buckets_nouse	= kvpmalloc(BITS_TO_LONGS(nbuckets) *
-					    sizeof(unsigned long),
-					    GFP_KERNEL|__GFP_ZERO)))) {
-		ret = -BCH_ERR_ENOMEM_buckets_nouse;
+	if (!(bucket_gens	= kvmalloc(sizeof(struct bucket_gens) + nbuckets,
+					   GFP_KERNEL|__GFP_ZERO))) {
+		ret = -BCH_ERR_ENOMEM_bucket_gens;
 		goto err;
 	}
 
 	bucket_gens->first_bucket = ca->mi.first_bucket;
 	bucket_gens->nbuckets	= nbuckets;
+	bucket_gens->nbuckets_minus_first =
+		bucket_gens->nbuckets - bucket_gens->first_bucket;
 
 	if (resize) {
 		down_write(&c->gc_lock);
@@ -1376,17 +1649,11 @@ int bch2_dev_buckets_resize(struct bch_fs *c, struct bch_dev *ca, u64 nbuckets)
 		memcpy(bucket_gens->b,
 		       old_bucket_gens->b,
 		       n);
-		if (buckets_nouse)
-			memcpy(buckets_nouse,
-			       ca->buckets_nouse,
-			       BITS_TO_LONGS(n) * sizeof(unsigned long));
 	}
 
 	rcu_assign_pointer(ca->bucket_gens, bucket_gens);
 	bucket_gens	= old_bucket_gens;
 
-	swap(ca->buckets_nouse, buckets_nouse);
-
 	nbuckets = ca->mi.nbuckets;
 
 	if (resize) {
@@ -1397,8 +1664,6 @@ int bch2_dev_buckets_resize(struct bch_fs *c, struct bch_dev *ca, u64 nbuckets)
 
 	ret = 0;
 err:
-	kvpfree(buckets_nouse,
-		BITS_TO_LONGS(nbuckets) * sizeof(unsigned long));
 	if (bucket_gens)
 		call_rcu(&bucket_gens->rcu, bucket_gens_free_rcu);
 
@@ -1407,27 +1672,21 @@ err:
 
 void bch2_dev_buckets_free(struct bch_dev *ca)
 {
-	unsigned i;
+	kvfree(ca->buckets_nouse);
+	kvfree(rcu_dereference_protected(ca->bucket_gens, 1));
 
-	kvpfree(ca->buckets_nouse,
-		BITS_TO_LONGS(ca->mi.nbuckets) * sizeof(unsigned long));
-	kvpfree(rcu_dereference_protected(ca->bucket_gens, 1),
-		sizeof(struct bucket_gens) + ca->mi.nbuckets);
-
-	for (i = 0; i < ARRAY_SIZE(ca->usage); i++)
+	for (unsigned i = 0; i < ARRAY_SIZE(ca->usage); i++)
 		free_percpu(ca->usage[i]);
 	kfree(ca->usage_base);
 }
 
 int bch2_dev_buckets_alloc(struct bch_fs *c, struct bch_dev *ca)
 {
-	unsigned i;
-
 	ca->usage_base = kzalloc(sizeof(struct bch_dev_usage), GFP_KERNEL);
 	if (!ca->usage_base)
 		return -BCH_ERR_ENOMEM_usage_init;
 
-	for (i = 0; i < ARRAY_SIZE(ca->usage); i++) {
+	for (unsigned i = 0; i < ARRAY_SIZE(ca->usage); i++) {
 		ca->usage[i] = alloc_percpu(struct bch_dev_usage);
 		if (!ca->usage[i])
 			return -BCH_ERR_ENOMEM_usage_init;
diff --git a/fs/bcachefs/buckets.h b/fs/bcachefs/buckets.h
index 6387e039f7..8ad4be7386 100644
--- a/fs/bcachefs/buckets.h
+++ b/fs/bcachefs/buckets.h
@@ -12,7 +12,7 @@
 #include "extents.h"
 #include "sb-members.h"
 
-static inline size_t sector_to_bucket(const struct bch_dev *ca, sector_t s)
+static inline u64 sector_to_bucket(const struct bch_dev *ca, sector_t s)
 {
 	return div_u64(s, ca->mi.bucket_size);
 }
@@ -30,8 +30,7 @@ static inline sector_t bucket_remainder(const struct bch_dev *ca, sector_t s)
 	return remainder;
 }
 
-static inline size_t sector_to_bucket_and_offset(const struct bch_dev *ca, sector_t s,
-						 u32 *offset)
+static inline u64 sector_to_bucket_and_offset(const struct bch_dev *ca, sector_t s, u32 *offset)
 {
 	return div_u64_rem(s, ca->mi.bucket_size, offset);
 }
@@ -94,7 +93,8 @@ static inline struct bucket *gc_bucket(struct bch_dev *ca, size_t b)
 {
 	struct bucket_array *buckets = gc_bucket_array(ca);
 
-	BUG_ON(b < buckets->first_bucket || b >= buckets->nbuckets);
+	if (b - buckets->first_bucket >= buckets->nbuckets_minus_first)
+		return NULL;
 	return buckets->b + b;
 }
 
@@ -111,30 +111,35 @@ static inline u8 *bucket_gen(struct bch_dev *ca, size_t b)
 {
 	struct bucket_gens *gens = bucket_gens(ca);
 
-	BUG_ON(b < gens->first_bucket || b >= gens->nbuckets);
+	if (b - gens->first_bucket >= gens->nbuckets_minus_first)
+		return NULL;
 	return gens->b + b;
 }
 
+static inline u8 bucket_gen_get(struct bch_dev *ca, size_t b)
+{
+	rcu_read_lock();
+	u8 gen = *bucket_gen(ca, b);
+	rcu_read_unlock();
+	return gen;
+}
+
 static inline size_t PTR_BUCKET_NR(const struct bch_dev *ca,
 				   const struct bch_extent_ptr *ptr)
 {
 	return sector_to_bucket(ca, ptr->offset);
 }
 
-static inline struct bpos PTR_BUCKET_POS(const struct bch_fs *c,
-				   const struct bch_extent_ptr *ptr)
+static inline struct bpos PTR_BUCKET_POS(const struct bch_dev *ca,
+					 const struct bch_extent_ptr *ptr)
 {
-	struct bch_dev *ca = bch_dev_bkey_exists(c, ptr->dev);
-
 	return POS(ptr->dev, PTR_BUCKET_NR(ca, ptr));
 }
 
-static inline struct bpos PTR_BUCKET_POS_OFFSET(const struct bch_fs *c,
+static inline struct bpos PTR_BUCKET_POS_OFFSET(const struct bch_dev *ca,
 						const struct bch_extent_ptr *ptr,
 						u32 *bucket_offset)
 {
-	struct bch_dev *ca = bch_dev_bkey_exists(c, ptr->dev);
-
 	return POS(ptr->dev, sector_to_bucket_and_offset(ca, ptr->offset, bucket_offset));
 }
 
@@ -175,17 +180,22 @@ static inline int gen_after(u8 a, u8 b)
 	return r > 0 ? r : 0;
 }
 
+static inline int dev_ptr_stale_rcu(struct bch_dev *ca, const struct bch_extent_ptr *ptr)
+{
+	u8 *gen = bucket_gen(ca, PTR_BUCKET_NR(ca, ptr));
+	if (!gen)
+		return -1;
+	return gen_after(*gen, ptr->gen);
+}
+
 /**
- * ptr_stale() - check if a pointer points into a bucket that has been
+ * dev_ptr_stale() - check if a pointer points into a bucket that has been
  * invalidated.
  */
-static inline u8 ptr_stale(struct bch_dev *ca,
-			   const struct bch_extent_ptr *ptr)
+static inline int dev_ptr_stale(struct bch_dev *ca, const struct bch_extent_ptr *ptr)
 {
-	u8 ret;
-
 	rcu_read_lock();
-	ret = gen_after(*bucket_gen(ca, PTR_BUCKET_NR(ca, ptr)), ptr->gen);
+	int ret = dev_ptr_stale_rcu(ca, ptr);
 	rcu_read_unlock();
 
 	return ret;
@@ -226,6 +236,7 @@ static inline u64 bch2_dev_buckets_reserved(struct bch_dev *ca, enum bch_waterma
 		fallthrough;
 	case BCH_WATERMARK_btree_copygc:
 	case BCH_WATERMARK_reclaim:
+	case BCH_WATERMARK_interior_updates:
 		break;
 	}
 
@@ -305,8 +316,6 @@ bch2_fs_usage_read_short(struct bch_fs *);
 void bch2_dev_usage_update(struct bch_fs *, struct bch_dev *,
 			   const struct bch_alloc_v4 *,
 			   const struct bch_alloc_v4 *, u64, bool);
-void bch2_dev_usage_update_m(struct bch_fs *, struct bch_dev *,
-			     struct bucket *, struct bucket *);
 
 /* key/bucket marking: */
 
@@ -332,27 +341,29 @@ int bch2_replicas_deltas_realloc(struct btree_trans *, unsigned);
 
 void bch2_fs_usage_initialize(struct bch_fs *);
 
-int bch2_check_bucket_ref(struct btree_trans *, struct bkey_s_c,
-			  const struct bch_extent_ptr *,
-			  s64, enum bch_data_type, u8, u8, u32);
+int bch2_bucket_ref_update(struct btree_trans *, struct bch_dev *,
+			   struct bkey_s_c, const struct bch_extent_ptr *,
+			   s64, enum bch_data_type, u8, u8, u32 *);
 
-int bch2_mark_metadata_bucket(struct bch_fs *, struct bch_dev *,
-			      size_t, enum bch_data_type, unsigned,
-			      struct gc_pos, unsigned);
+int bch2_check_fix_ptrs(struct btree_trans *,
+			enum btree_id, unsigned, struct bkey_s_c,
+			enum btree_iter_update_trigger_flags);
 
 int bch2_trigger_extent(struct btree_trans *, enum btree_id, unsigned,
-			struct bkey_s_c, struct bkey_s, unsigned);
+			struct bkey_s_c, struct bkey_s,
+			enum btree_iter_update_trigger_flags);
 int bch2_trigger_reservation(struct btree_trans *, enum btree_id, unsigned,
-			  struct bkey_s_c, struct bkey_s, unsigned);
+			  struct bkey_s_c, struct bkey_s,
+			  enum btree_iter_update_trigger_flags);
 
 #define trigger_run_overwrite_then_insert(_fn, _trans, _btree_id, _level, _old, _new, _flags)\
 ({												\
 	int ret = 0;										\
 												\
 	if (_old.k->type)									\
-		ret = _fn(_trans, _btree_id, _level, _old, _flags & ~BTREE_TRIGGER_INSERT);	\
+		ret = _fn(_trans, _btree_id, _level, _old, _flags & ~BTREE_TRIGGER_insert);	\
 	if (!ret && _new.k->type)								\
-		ret = _fn(_trans, _btree_id, _level, _new.s_c, _flags & ~BTREE_TRIGGER_OVERWRITE);\
+		ret = _fn(_trans, _btree_id, _level, _new.s_c, _flags & ~BTREE_TRIGGER_overwrite);\
 	ret;											\
 })
 
@@ -361,9 +372,13 @@ void bch2_trans_account_disk_usage_change(struct btree_trans *);
 void bch2_trans_fs_usage_revert(struct btree_trans *, struct replicas_delta_list *);
 int bch2_trans_fs_usage_apply(struct btree_trans *, struct replicas_delta_list *);
 
-int bch2_trans_mark_metadata_bucket(struct btree_trans *, struct bch_dev *,
-				    size_t, enum bch_data_type, unsigned);
-int bch2_trans_mark_dev_sb(struct bch_fs *, struct bch_dev *);
+int bch2_trans_mark_metadata_bucket(struct btree_trans *, struct bch_dev *, u64,
+				    enum bch_data_type, unsigned,
+				    enum btree_iter_update_trigger_flags);
+int bch2_trans_mark_dev_sb(struct bch_fs *, struct bch_dev *,
+				    enum btree_iter_update_trigger_flags);
+int bch2_trans_mark_dev_sbs_flags(struct bch_fs *,
+				    enum btree_iter_update_trigger_flags);
 int bch2_trans_mark_dev_sbs(struct bch_fs *);
 
 static inline bool is_superblock_bucket(struct bch_dev *ca, u64 b)
@@ -394,14 +409,6 @@ static inline const char *bch2_data_type_str(enum bch_data_type type)
 		: "(invalid data type)";
 }
 
-static inline void bch2_prt_data_type(struct printbuf *out, enum bch_data_type type)
-{
-	if (type < BCH_DATA_NR)
-		prt_str(out, __bch2_data_types[type]);
-	else
-		prt_printf(out, "(invalid data type %u)", type);
-}
-
 /* disk reservations: */
 
 static inline void bch2_disk_reservation_put(struct bch_fs *c,
@@ -471,6 +478,9 @@ static inline u64 avail_factor(u64 r)
 	return div_u64(r << RESERVE_FACTOR, (1 << RESERVE_FACTOR) + 1);
 }
 
+void bch2_buckets_nouse_free(struct bch_fs *);
+int bch2_buckets_nouse_alloc(struct bch_fs *);
+
 int bch2_dev_buckets_resize(struct bch_fs *, struct bch_dev *, u64);
 void bch2_dev_buckets_free(struct bch_dev *);
 int bch2_dev_buckets_alloc(struct bch_fs *, struct bch_dev *);
diff --git a/fs/bcachefs/buckets_types.h b/fs/bcachefs/buckets_types.h
index 6a31740222..f636e17c4c 100644
--- a/fs/bcachefs/buckets_types.h
+++ b/fs/bcachefs/buckets_types.h
@@ -22,6 +22,7 @@ struct bucket_array {
 	struct rcu_head		rcu;
 	u16			first_bucket;
 	size_t			nbuckets;
+	size_t			nbuckets_minus_first;
 	struct bucket		b[];
 };
 
@@ -29,6 +30,7 @@ struct bucket_gens {
 	struct rcu_head		rcu;
 	u16			first_bucket;
 	size_t			nbuckets;
+	size_t			nbuckets_minus_first;
 	u8			b[];
 };
 
diff --git a/fs/bcachefs/chardev.c b/fs/bcachefs/chardev.c
index 226b39c176..6d82e1165a 100644
--- a/fs/bcachefs/chardev.c
+++ b/fs/bcachefs/chardev.c
@@ -7,7 +7,7 @@
 #include "chardev.h"
 #include "journal.h"
 #include "move.h"
-#include "recovery.h"
+#include "recovery_passes.h"
 #include "replicas.h"
 #include "super.h"
 #include "super-io.h"
@@ -22,12 +22,6 @@
 #include <linux/slab.h>
 #include <linux/uaccess.h>
 
-__must_check
-static int copy_to_user_errcode(void __user *to, const void *from, unsigned long n)
-{
-	return copy_to_user(to, from, n) ? -EFAULT : 0;
-}
-
 /* returns with ref on ca->ref */
 static struct bch_dev *bch2_device_lookup(struct bch_fs *c, u64 dev,
 					  unsigned flags)
@@ -38,12 +32,7 @@ static struct bch_dev *bch2_device_lookup(struct bch_fs *c, u64 dev,
 		if (dev >= c->sb.nr_devices)
 			return ERR_PTR(-EINVAL);
 
-		rcu_read_lock();
-		ca = rcu_dereference(c->devs[dev]);
-		if (ca)
-			percpu_ref_get(&ca->ref);
-		rcu_read_unlock();
-
+		ca = bch2_dev_tryget_noerror(c, dev);
 		if (!ca)
 			return ERR_PTR(-EINVAL);
 	} else {
@@ -140,39 +129,51 @@ static long bch2_ioctl_incremental(struct bch_ioctl_incremental __user *user_arg
 struct fsck_thread {
 	struct thread_with_stdio thr;
 	struct bch_fs		*c;
-	char			**devs;
-	size_t			nr_devs;
 	struct bch_opts		opts;
 };
 
 static void bch2_fsck_thread_exit(struct thread_with_stdio *_thr)
 {
 	struct fsck_thread *thr = container_of(_thr, struct fsck_thread, thr);
-	if (thr->devs)
-		for (size_t i = 0; i < thr->nr_devs; i++)
-			kfree(thr->devs[i]);
-	kfree(thr->devs);
 	kfree(thr);
 }
 
-static int bch2_fsck_offline_thread_fn(void *arg)
+static int bch2_fsck_offline_thread_fn(struct thread_with_stdio *stdio)
 {
-	struct fsck_thread *thr = container_of(arg, struct fsck_thread, thr);
-	struct bch_fs *c = bch2_fs_open(thr->devs, thr->nr_devs, thr->opts);
+	struct fsck_thread *thr = container_of(stdio, struct fsck_thread, thr);
+	struct bch_fs *c = thr->c;
 
-	thr->thr.thr.ret = PTR_ERR_OR_ZERO(c);
-	if (!thr->thr.thr.ret)
-		bch2_fs_stop(c);
+	int ret = PTR_ERR_OR_ZERO(c);
+	if (ret)
+		return ret;
 
-	thread_with_stdio_done(&thr->thr);
-	return 0;
+	ret = bch2_fs_start(thr->c);
+	if (ret)
+		goto err;
+
+	if (test_bit(BCH_FS_errors_fixed, &c->flags)) {
+		bch2_stdio_redirect_printf(&stdio->stdio, false, "%s: errors fixed\n", c->name);
+		ret |= 1;
+	}
+	if (test_bit(BCH_FS_error, &c->flags)) {
+		bch2_stdio_redirect_printf(&stdio->stdio, false, "%s: still has errors\n", c->name);
+		ret |= 4;
+	}
+err:
+	bch2_fs_stop(c);
+	return ret;
 }
 
+static const struct thread_with_stdio_ops bch2_offline_fsck_ops = {
+	.exit		= bch2_fsck_thread_exit,
+	.fn		= bch2_fsck_offline_thread_fn,
+};
+
 static long bch2_ioctl_fsck_offline(struct bch_ioctl_fsck_offline __user *user_arg)
 {
 	struct bch_ioctl_fsck_offline arg;
 	struct fsck_thread *thr = NULL;
-	u64 *devs = NULL;
+	darray_str(devs) = {};
 	long ret = 0;
 
 	if (copy_from_user(&arg, user_arg, sizeof(arg)))
@@ -184,35 +185,39 @@ static long bch2_ioctl_fsck_offline(struct bch_ioctl_fsck_offline __user *user_a
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
-	if (!(devs = kcalloc(arg.nr_devs, sizeof(*devs), GFP_KERNEL)) ||
-	    !(thr = kzalloc(sizeof(*thr), GFP_KERNEL)) ||
-	    !(thr->devs = kcalloc(arg.nr_devs, sizeof(*thr->devs), GFP_KERNEL))) {
-		ret = -ENOMEM;
-		goto err;
-	}
+	for (size_t i = 0; i < arg.nr_devs; i++) {
+		u64 dev_u64;
+		ret = copy_from_user_errcode(&dev_u64, &user_arg->devs[i], sizeof(u64));
+		if (ret)
+			goto err;
 
-	thr->opts = bch2_opts_empty();
-	thr->nr_devs = arg.nr_devs;
+		char *dev_str = strndup_user((char __user *)(unsigned long) dev_u64, PATH_MAX);
+		ret = PTR_ERR_OR_ZERO(dev_str);
+		if (ret)
+			goto err;
 
-	if (copy_from_user(devs, &user_arg->devs[0],
-			   array_size(sizeof(user_arg->devs[0]), arg.nr_devs))) {
-		ret = -EINVAL;
-		goto err;
+		ret = darray_push(&devs, dev_str);
+		if (ret) {
+			kfree(dev_str);
+			goto err;
+		}
 	}
 
-	for (size_t i = 0; i < arg.nr_devs; i++) {
-		thr->devs[i] = strndup_user((char __user *)(unsigned long) devs[i], PATH_MAX);
-		ret = PTR_ERR_OR_ZERO(thr->devs[i]);
-		if (ret)
-			goto err;
+	thr = kzalloc(sizeof(*thr), GFP_KERNEL);
+	if (!thr) {
+		ret = -ENOMEM;
+		goto err;
 	}
 
+	thr->opts = bch2_opts_empty();
+
 	if (arg.opts) {
 		char *optstr = strndup_user((char __user *)(unsigned long) arg.opts, 1 << 16);
 
 		ret =   PTR_ERR_OR_ZERO(optstr) ?:
 			bch2_parse_mount_opts(NULL, &thr->opts, optstr);
-		kfree(optstr);
+		if (!IS_ERR(optstr))
+			kfree(optstr);
 
 		if (ret)
 			goto err;
@@ -220,17 +225,28 @@ static long bch2_ioctl_fsck_offline(struct bch_ioctl_fsck_offline __user *user_a
 
 	opt_set(thr->opts, stdio, (u64)(unsigned long)&thr->thr.stdio);
 
-	ret = bch2_run_thread_with_stdio(&thr->thr,
-			bch2_fsck_thread_exit,
-			bch2_fsck_offline_thread_fn);
-err:
-	if (ret < 0) {
-		if (thr)
-			bch2_fsck_thread_exit(&thr->thr);
-		pr_err("ret %s", bch2_err_str(ret));
-	}
-	kfree(devs);
+	/* We need request_key() to be called before we punt to kthread: */
+	opt_set(thr->opts, nostart, true);
+
+	bch2_thread_with_stdio_init(&thr->thr, &bch2_offline_fsck_ops);
+
+	thr->c = bch2_fs_open(devs.data, arg.nr_devs, thr->opts);
+
+	if (!IS_ERR(thr->c) &&
+	    thr->c->opts.errors == BCH_ON_ERROR_panic)
+		thr->c->opts.errors = BCH_ON_ERROR_ro;
+
+	ret = __bch2_run_thread_with_stdio(&thr->thr);
+out:
+	darray_for_each(devs, i)
+		kfree(*i);
+	darray_exit(&devs);
 	return ret;
+err:
+	if (thr)
+		bch2_fsck_thread_exit(&thr->thr);
+	pr_err("ret %s", bch2_err_str(ret));
+	goto out;
 }
 
 static long bch2_global_ioctl(unsigned cmd, void __user *arg)
@@ -304,7 +320,8 @@ static long bch2_ioctl_disk_add(struct bch_fs *c, struct bch_ioctl_disk arg)
 		return ret;
 
 	ret = bch2_dev_add(c, path);
-	kfree(path);
+	if (!IS_ERR(path))
+		kfree(path);
 
 	return ret;
 }
@@ -371,7 +388,7 @@ static long bch2_ioctl_disk_offline(struct bch_fs *c, struct bch_ioctl_disk arg)
 		return PTR_ERR(ca);
 
 	ret = bch2_dev_offline(c, ca, arg.flags);
-	percpu_ref_put(&ca->ref);
+	bch2_dev_put(ca);
 	return ret;
 }
 
@@ -400,7 +417,7 @@ static long bch2_ioctl_disk_set_state(struct bch_fs *c,
 	if (ret)
 		bch_err(c, "Error setting device state: %s", bch2_err_str(ret));
 
-	percpu_ref_put(&ca->ref);
+	bch2_dev_put(ca);
 	return ret;
 }
 
@@ -595,7 +612,7 @@ static long bch2_ioctl_dev_usage(struct bch_fs *c,
 		arg.d[i].fragmented	= src.d[i].fragmented;
 	}
 
-	percpu_ref_put(&ca->ref);
+	bch2_dev_put(ca);
 
 	return copy_to_user_errcode(user_arg, &arg, sizeof(arg));
 }
@@ -647,7 +664,7 @@ static long bch2_ioctl_dev_usage_v2(struct bch_fs *c,
 			goto err;
 	}
 err:
-	percpu_ref_put(&ca->ref);
+	bch2_dev_put(ca);
 	return ret;
 }
 
@@ -669,11 +686,9 @@ static long bch2_ioctl_read_super(struct bch_fs *c,
 
 	if (arg.flags & BCH_READ_DEV) {
 		ca = bch2_device_lookup(c, arg.dev, arg.flags);
-
-		if (IS_ERR(ca)) {
-			ret = PTR_ERR(ca);
-			goto err;
-		}
+		ret = PTR_ERR_OR_ZERO(ca);
+		if (ret)
+			goto err_unlock;
 
 		sb = ca->disk_sb.sb;
 	} else {
@@ -688,8 +703,8 @@ static long bch2_ioctl_read_super(struct bch_fs *c,
 	ret = copy_to_user_errcode((void __user *)(unsigned long)arg.sb, sb,
 				   vstruct_bytes(sb));
 err:
-	if (!IS_ERR_OR_NULL(ca))
-		percpu_ref_put(&ca->ref);
+	bch2_dev_put(ca);
+err_unlock:
 	mutex_unlock(&c->sb_lock);
 	return ret;
 }
@@ -733,7 +748,7 @@ static long bch2_ioctl_disk_resize(struct bch_fs *c,
 
 	ret = bch2_dev_resize(c, ca, arg.nbuckets);
 
-	percpu_ref_put(&ca->ref);
+	bch2_dev_put(ca);
 	return ret;
 }
 
@@ -759,13 +774,13 @@ static long bch2_ioctl_disk_resize_journal(struct bch_fs *c,
 
 	ret = bch2_set_nr_journal_buckets(c, ca, arg.nbuckets);
 
-	percpu_ref_put(&ca->ref);
+	bch2_dev_put(ca);
 	return ret;
 }
 
-static int bch2_fsck_online_thread_fn(void *arg)
+static int bch2_fsck_online_thread_fn(struct thread_with_stdio *stdio)
 {
-	struct fsck_thread *thr = container_of(arg, struct fsck_thread, thr);
+	struct fsck_thread *thr = container_of(stdio, struct fsck_thread, thr);
 	struct bch_fs *c = thr->c;
 
 	c->stdio_filter = current;
@@ -793,13 +808,16 @@ static int bch2_fsck_online_thread_fn(void *arg)
 	c->stdio_filter = NULL;
 	c->opts.fix_errors = old_fix_errors;
 
-	thread_with_stdio_done(&thr->thr);
-
 	up(&c->online_fsck_mutex);
 	bch2_ro_ref_put(c);
-	return 0;
+	return ret;
 }
 
+static const struct thread_with_stdio_ops bch2_online_fsck_ops = {
+	.exit		= bch2_fsck_thread_exit,
+	.fn		= bch2_fsck_online_thread_fn,
+};
+
 static long bch2_ioctl_fsck_online(struct bch_fs *c,
 				   struct bch_ioctl_fsck_online arg)
 {
@@ -834,15 +852,14 @@ static long bch2_ioctl_fsck_online(struct bch_fs *c,
 
 		ret =   PTR_ERR_OR_ZERO(optstr) ?:
 			bch2_parse_mount_opts(c, &thr->opts, optstr);
-		kfree(optstr);
+		if (!IS_ERR(optstr))
+			kfree(optstr);
 
 		if (ret)
 			goto err;
 	}
 
-	ret = bch2_run_thread_with_stdio(&thr->thr,
-			bch2_fsck_thread_exit,
-			bch2_fsck_online_thread_fn);
+	ret = bch2_run_thread_with_stdio(&thr->thr, &bch2_online_fsck_ops);
 err:
 	if (ret < 0) {
 		bch_err_fn(c, ret);
@@ -940,7 +957,9 @@ static const struct file_operations bch_chardev_fops = {
 };
 
 static int bch_chardev_major;
-static struct class *bch_chardev_class;
+static const struct class bch_chardev_class = {
+	.name = "bcachefs",
+};
 static struct device *bch_chardev;
 
 void bch2_fs_chardev_exit(struct bch_fs *c)
@@ -957,7 +976,7 @@ int bch2_fs_chardev_init(struct bch_fs *c)
 	if (c->minor < 0)
 		return c->minor;
 
-	c->chardev = device_create(bch_chardev_class, NULL,
+	c->chardev = device_create(&bch_chardev_class, NULL,
 				   MKDEV(bch_chardev_major, c->minor), c,
 				   "bcachefs%u-ctl", c->minor);
 	if (IS_ERR(c->chardev))
@@ -968,32 +987,39 @@ int bch2_fs_chardev_init(struct bch_fs *c)
 
 void bch2_chardev_exit(void)
 {
-	if (!IS_ERR_OR_NULL(bch_chardev_class))
-		device_destroy(bch_chardev_class,
-			       MKDEV(bch_chardev_major, U8_MAX));
-	if (!IS_ERR_OR_NULL(bch_chardev_class))
-		class_destroy(bch_chardev_class);
+	device_destroy(&bch_chardev_class, MKDEV(bch_chardev_major, U8_MAX));
+	class_unregister(&bch_chardev_class);
 	if (bch_chardev_major > 0)
 		unregister_chrdev(bch_chardev_major, "bcachefs");
 }
 
 int __init bch2_chardev_init(void)
 {
+	int ret;
+
 	bch_chardev_major = register_chrdev(0, "bcachefs-ctl", &bch_chardev_fops);
 	if (bch_chardev_major < 0)
 		return bch_chardev_major;
 
-	bch_chardev_class = class_create("bcachefs");
-	if (IS_ERR(bch_chardev_class))
-		return PTR_ERR(bch_chardev_class);
+	ret = class_register(&bch_chardev_class);
+	if (ret)
+		goto major_out;
 
-	bch_chardev = device_create(bch_chardev_class, NULL,
+	bch_chardev = device_create(&bch_chardev_class, NULL,
 				    MKDEV(bch_chardev_major, U8_MAX),
 				    NULL, "bcachefs-ctl");
-	if (IS_ERR(bch_chardev))
-		return PTR_ERR(bch_chardev);
+	if (IS_ERR(bch_chardev)) {
+		ret = PTR_ERR(bch_chardev);
+		goto class_out;
+	}
 
 	return 0;
+
+class_out:
+	class_unregister(&bch_chardev_class);
+major_out:
+	unregister_chrdev(bch_chardev_major, "bcachefs-ctl");
+	return ret;
 }
 
 #endif /* NO_BCACHEFS_CHARDEV */
diff --git a/fs/bcachefs/checksum.c b/fs/bcachefs/checksum.c
index 3c761ad6b1..3bd3aba90d 100644
--- a/fs/bcachefs/checksum.c
+++ b/fs/bcachefs/checksum.c
@@ -102,6 +102,7 @@ static inline int do_encrypt_sg(struct crypto_sync_skcipher *tfm,
 	int ret;
 
 	skcipher_request_set_sync_tfm(req, tfm);
+	skcipher_request_set_callback(req, 0, NULL, NULL);
 	skcipher_request_set_crypt(req, sg, sg, len, nonce.d);
 
 	ret = crypto_skcipher_encrypt(req);
@@ -232,7 +233,7 @@ struct bch_csum bch2_checksum(struct bch_fs *c, unsigned type,
 		return ret;
 	}
 	default:
-		BUG();
+		return (struct bch_csum) {};
 	}
 }
 
@@ -306,7 +307,7 @@ static struct bch_csum __bch2_checksum_bio(struct bch_fs *c, unsigned type,
 		return ret;
 	}
 	default:
-		BUG();
+		return (struct bch_csum) {};
 	}
 }
 
@@ -351,8 +352,12 @@ int __bch2_encrypt_bio(struct bch_fs *c, unsigned type,
 		bytes += bv.bv_len;
 	}
 
-	sg_mark_end(sg - 1);
-	return do_encrypt_sg(c->chacha20, nonce, sgl, bytes);
+	if (sg != sgl) {
+		sg_mark_end(sg - 1);
+		return do_encrypt_sg(c->chacha20, nonce, sgl, bytes);
+	}
+
+	return ret;
 }
 
 struct bch_csum bch2_checksum_merge(unsigned type, struct bch_csum a,
@@ -429,15 +434,20 @@ int bch2_rechecksum_bio(struct bch_fs *c, struct bio *bio,
 				extent_nonce(version, crc_old), bio);
 
 	if (bch2_crc_cmp(merged, crc_old.csum) && !c->opts.no_data_io) {
-		bch_err(c, "checksum error in %s() (memory corruption or bug?)\n"
-			"expected %0llx:%0llx got %0llx:%0llx (old type %s new type %s)",
-			__func__,
-			crc_old.csum.hi,
-			crc_old.csum.lo,
-			merged.hi,
-			merged.lo,
-			bch2_csum_types[crc_old.csum_type],
-			bch2_csum_types[new_csum_type]);
+		struct printbuf buf = PRINTBUF;
+		prt_printf(&buf, "checksum error in %s() (memory corruption or bug?)\n"
+			   "expected %0llx:%0llx got %0llx:%0llx (old type ",
+			   __func__,
+			   crc_old.csum.hi,
+			   crc_old.csum.lo,
+			   merged.hi,
+			   merged.lo);
+		bch2_prt_csum_type(&buf, crc_old.csum_type);
+		prt_str(&buf, " new type ");
+		bch2_prt_csum_type(&buf, new_csum_type);
+		prt_str(&buf, ")");
+		bch_err(c, "%s", buf.buf);
+		printbuf_exit(&buf);
 		return -EIO;
 	}
 
@@ -463,9 +473,8 @@ int bch2_rechecksum_bio(struct bch_fs *c, struct bio *bio,
 
 /* BCH_SB_FIELD_crypt: */
 
-static int bch2_sb_crypt_validate(struct bch_sb *sb,
-				  struct bch_sb_field *f,
-				  struct printbuf *err)
+static int bch2_sb_crypt_validate(struct bch_sb *sb, struct bch_sb_field *f,
+				  enum bch_validate_flags flags, struct printbuf *err)
 {
 	struct bch_sb_field_crypt *crypt = field_to_type(f, crypt);
 
@@ -488,14 +497,10 @@ static void bch2_sb_crypt_to_text(struct printbuf *out, struct bch_sb *sb,
 {
 	struct bch_sb_field_crypt *crypt = field_to_type(f, crypt);
 
-	prt_printf(out, "KFD:               %llu", BCH_CRYPT_KDF_TYPE(crypt));
-	prt_newline(out);
-	prt_printf(out, "scrypt n:          %llu", BCH_KDF_SCRYPT_N(crypt));
-	prt_newline(out);
-	prt_printf(out, "scrypt r:          %llu", BCH_KDF_SCRYPT_R(crypt));
-	prt_newline(out);
-	prt_printf(out, "scrypt p:          %llu", BCH_KDF_SCRYPT_P(crypt));
-	prt_newline(out);
+	prt_printf(out, "KFD:               %llu\n", BCH_CRYPT_KDF_TYPE(crypt));
+	prt_printf(out, "scrypt n:          %llu\n", BCH_KDF_SCRYPT_N(crypt));
+	prt_printf(out, "scrypt r:          %llu\n", BCH_KDF_SCRYPT_R(crypt));
+	prt_printf(out, "scrypt p:          %llu\n", BCH_KDF_SCRYPT_P(crypt));
 }
 
 const struct bch_sb_field_ops bch_sb_field_ops_crypt = {
@@ -558,7 +563,7 @@ got_key:
 	return 0;
 }
 
-#include "../crypto.h"
+#include "crypto.h"
 #endif
 
 int bch2_request_key(struct bch_sb *sb, struct bch_key *key)
@@ -647,26 +652,26 @@ err:
 
 static int bch2_alloc_ciphers(struct bch_fs *c)
 {
-	int ret;
-
-	if (!c->chacha20)
-		c->chacha20 = crypto_alloc_sync_skcipher("chacha20", 0, 0);
-	ret = PTR_ERR_OR_ZERO(c->chacha20);
+	if (c->chacha20)
+		return 0;
 
+	struct crypto_sync_skcipher *chacha20 = crypto_alloc_sync_skcipher("chacha20", 0, 0);
+	int ret = PTR_ERR_OR_ZERO(chacha20);
 	if (ret) {
 		bch_err(c, "error requesting chacha20 module: %s", bch2_err_str(ret));
 		return ret;
 	}
 
-	if (!c->poly1305)
-		c->poly1305 = crypto_alloc_shash("poly1305", 0, 0);
-	ret = PTR_ERR_OR_ZERO(c->poly1305);
-
+	struct crypto_shash *poly1305 = crypto_alloc_shash("poly1305", 0, 0);
+	ret = PTR_ERR_OR_ZERO(poly1305);
 	if (ret) {
 		bch_err(c, "error requesting poly1305 module: %s", bch2_err_str(ret));
+		crypto_free_sync_skcipher(chacha20);
 		return ret;
 	}
 
+	c->chacha20	= chacha20;
+	c->poly1305	= poly1305;
 	return 0;
 }
 
@@ -761,11 +766,11 @@ err:
 
 void bch2_fs_encryption_exit(struct bch_fs *c)
 {
-	if (!IS_ERR_OR_NULL(c->poly1305))
+	if (c->poly1305)
 		crypto_free_shash(c->poly1305);
-	if (!IS_ERR_OR_NULL(c->chacha20))
+	if (c->chacha20)
 		crypto_free_sync_skcipher(c->chacha20);
-	if (!IS_ERR_OR_NULL(c->sha256))
+	if (c->sha256)
 		crypto_free_shash(c->sha256);
 }
 
@@ -778,6 +783,7 @@ int bch2_fs_encryption_init(struct bch_fs *c)
 	c->sha256 = crypto_alloc_shash("sha256", 0, 0);
 	ret = PTR_ERR_OR_ZERO(c->sha256);
 	if (ret) {
+		c->sha256 = NULL;
 		bch_err(c, "error requesting sha256 module: %s", bch2_err_str(ret));
 		goto out;
 	}
diff --git a/fs/bcachefs/checksum.h b/fs/bcachefs/checksum.h
index 1b8c2c1016..e40499fde9 100644
--- a/fs/bcachefs/checksum.h
+++ b/fs/bcachefs/checksum.h
@@ -61,11 +61,12 @@ static inline void bch2_csum_err_msg(struct printbuf *out,
 				     struct bch_csum expected,
 				     struct bch_csum got)
 {
-	prt_printf(out, "checksum error: got ");
+	prt_str(out, "checksum error, type ");
+	bch2_prt_csum_type(out, type);
+	prt_str(out, ": got ");
 	bch2_csum_to_text(out, type, got);
 	prt_str(out, " should be ");
 	bch2_csum_to_text(out, type, expected);
-	prt_printf(out, " type %s", bch2_csum_types[type]);
 }
 
 int bch2_chacha_encrypt_key(struct bch_key *, struct nonce, void *, size_t);
diff --git a/fs/bcachefs/clock.c b/fs/bcachefs/clock.c
index 3636444511..0f40b585ce 100644
--- a/fs/bcachefs/clock.c
+++ b/fs/bcachefs/clock.c
@@ -132,14 +132,9 @@ static struct io_timer *get_expired_timer(struct io_clock *clock,
 {
 	struct io_timer *ret = NULL;
 
-	spin_lock(&clock->timer_lock);
-
 	if (clock->timers.used &&
 	    time_after_eq(now, clock->timers.data[0]->expire))
 		heap_pop(&clock->timers, ret, io_timer_cmp, NULL);
-
-	spin_unlock(&clock->timer_lock);
-
 	return ret;
 }
 
@@ -148,8 +143,10 @@ void __bch2_increment_clock(struct io_clock *clock, unsigned sectors)
 	struct io_timer *timer;
 	unsigned long now = atomic64_add_return(sectors, &clock->now);
 
+	spin_lock(&clock->timer_lock);
 	while ((timer = get_expired_timer(clock, now)))
 		timer->fn(timer);
+	spin_unlock(&clock->timer_lock);
 }
 
 void bch2_io_timers_to_text(struct printbuf *out, struct io_clock *clock)
diff --git a/fs/bcachefs/compress.c b/fs/bcachefs/compress.c
index 33df8cf86b..1410365a88 100644
--- a/fs/bcachefs/compress.c
+++ b/fs/bcachefs/compress.c
@@ -601,13 +601,13 @@ static int __bch2_fs_compress_init(struct bch_fs *c, u64 features)
 		return 0;
 
 	if (!mempool_initialized(&c->compression_bounce[READ]) &&
-	    mempool_init_kvpmalloc_pool(&c->compression_bounce[READ],
-					1, c->opts.encoded_extent_max))
+	    mempool_init_kvmalloc_pool(&c->compression_bounce[READ],
+				       1, c->opts.encoded_extent_max))
 		return -BCH_ERR_ENOMEM_compression_bounce_read_init;
 
 	if (!mempool_initialized(&c->compression_bounce[WRITE]) &&
-	    mempool_init_kvpmalloc_pool(&c->compression_bounce[WRITE],
-					1, c->opts.encoded_extent_max))
+	    mempool_init_kvmalloc_pool(&c->compression_bounce[WRITE],
+				       1, c->opts.encoded_extent_max))
 		return -BCH_ERR_ENOMEM_compression_bounce_write_init;
 
 	for (i = compression_types;
@@ -622,15 +622,15 @@ static int __bch2_fs_compress_init(struct bch_fs *c, u64 features)
 		if (mempool_initialized(&c->compress_workspace[i->type]))
 			continue;
 
-		if (mempool_init_kvpmalloc_pool(
+		if (mempool_init_kvmalloc_pool(
 				&c->compress_workspace[i->type],
 				1, i->compress_workspace))
 			return -BCH_ERR_ENOMEM_compression_workspace_init;
 	}
 
 	if (!mempool_initialized(&c->decompress_workspace) &&
-	    mempool_init_kvpmalloc_pool(&c->decompress_workspace,
-					1, decompress_workspace_size))
+	    mempool_init_kvmalloc_pool(&c->decompress_workspace,
+				       1, decompress_workspace_size))
 		return -BCH_ERR_ENOMEM_decompression_workspace_init;
 
 	return 0;
diff --git a/fs/bcachefs/compress.h b/fs/bcachefs/compress.h
index 58c2eb4557..607fd5e232 100644
--- a/fs/bcachefs/compress.h
+++ b/fs/bcachefs/compress.h
@@ -47,14 +47,6 @@ static inline enum bch_compression_type bch2_compression_opt_to_type(unsigned v)
 	return __bch2_compression_opt_to_type[bch2_compression_decode(v).type];
 }
 
-static inline void bch2_prt_compression_type(struct printbuf *out, enum bch_compression_type type)
-{
-	if (type < BCH_COMPRESSION_TYPE_NR)
-		prt_str(out, __bch2_compression_types[type]);
-	else
-		prt_printf(out, "(invalid compression type %u)", type);
-}
-
 int bch2_bio_uncompress_inplace(struct bch_fs *, struct bio *,
 				struct bch_extent_crc_unpacked *);
 int bch2_bio_uncompress(struct bch_fs *, struct bio *, struct bio *,
diff --git a/fs/bcachefs/data_update.c b/fs/bcachefs/data_update.c
index 4150feca42..0087b8555e 100644
--- a/fs/bcachefs/data_update.c
+++ b/fs/bcachefs/data_update.c
@@ -5,7 +5,9 @@
 #include "bkey_buf.h"
 #include "btree_update.h"
 #include "buckets.h"
+#include "compress.h"
 #include "data_update.h"
+#include "disk_groups.h"
 #include "ec.h"
 #include "error.h"
 #include "extents.h"
@@ -14,6 +16,7 @@
 #include "move.h"
 #include "nocow_locking.h"
 #include "rebalance.h"
+#include "snapshot.h"
 #include "subvolume.h"
 #include "trace.h"
 
@@ -105,7 +108,7 @@ static int __bch2_data_update_index_update(struct btree_trans *trans,
 
 	bch2_trans_iter_init(trans, &iter, m->btree_id,
 			     bkey_start_pos(&bch2_keylist_front(keys)->k),
-			     BTREE_ITER_SLOTS|BTREE_ITER_INTENT);
+			     BTREE_ITER_slots|BTREE_ITER_intent);
 
 	while (1) {
 		struct bkey_s_c k;
@@ -201,6 +204,7 @@ restart_drop_conflicting_replicas:
 			bch2_bkey_durability(c, bkey_i_to_s_c(&new->k_i));
 
 		/* Now, drop excess replicas: */
+		rcu_read_lock();
 restart_drop_extra_replicas:
 		bkey_for_each_ptr_decode(old.k, bch2_bkey_ptrs(bkey_i_to_s(insert)), p, entry) {
 			unsigned ptr_durability = bch2_extent_ptr_durability(c, &p);
@@ -213,6 +217,7 @@ restart_drop_extra_replicas:
 				goto restart_drop_extra_replicas;
 			}
 		}
+		rcu_read_unlock();
 
 		/* Finally, add the pointers we just wrote: */
 		extent_for_each_ptr_decode(extent_i_to_s(new), p, entry)
@@ -287,7 +292,7 @@ restart_drop_extra_replicas:
 						k.k->p, insert->k.p) ?:
 			bch2_bkey_set_needs_rebalance(c, insert, &op->opts) ?:
 			bch2_trans_update(trans, &iter, insert,
-				BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE) ?:
+				BTREE_UPDATE_internal_snapshot_node) ?:
 			bch2_trans_commit(trans, &op->res,
 				NULL,
 				BCH_TRANS_COMMIT_no_check_rw|
@@ -356,10 +361,11 @@ void bch2_data_update_exit(struct data_update *update)
 		bch2_bkey_ptrs_c(bkey_i_to_s_c(update->k.k));
 
 	bkey_for_each_ptr(ptrs, ptr) {
+		struct bch_dev *ca = bch2_dev_have_ref(c, ptr->dev);
 		if (c->opts.nocow_enabled)
 			bch2_bucket_nocow_unlock(&c->nocow_locks,
-						 PTR_BUCKET_POS(c, ptr), 0);
-		percpu_ref_put(&bch_dev_bkey_exists(c, ptr->dev)->ref);
+						 PTR_BUCKET_POS(ca, ptr), 0);
+		bch2_dev_put(ca);
 	}
 
 	bch2_bkey_buf_exit(&update->k, c);
@@ -385,8 +391,10 @@ static void bch2_update_unwritten_extent(struct btree_trans *trans,
 	while (bio_sectors(bio)) {
 		unsigned sectors = bio_sectors(bio);
 
+		bch2_trans_begin(trans);
+
 		bch2_trans_iter_init(trans, &iter, update->btree_id, update->op.pos,
-				     BTREE_ITER_SLOTS);
+				     BTREE_ITER_slots);
 		ret = lockrestart_do(trans, ({
 			k = bch2_btree_iter_peek_slot(&iter);
 			bkey_err(k);
@@ -448,6 +456,38 @@ static void bch2_update_unwritten_extent(struct btree_trans *trans,
 	}
 }
 
+void bch2_data_update_opts_to_text(struct printbuf *out, struct bch_fs *c,
+				   struct bch_io_opts *io_opts,
+				   struct data_update_opts *data_opts)
+{
+	printbuf_tabstop_push(out, 20);
+	prt_str(out, "rewrite ptrs:\t");
+	bch2_prt_u64_base2(out, data_opts->rewrite_ptrs);
+	prt_newline(out);
+
+	prt_str(out, "kill ptrs:\t");
+	bch2_prt_u64_base2(out, data_opts->kill_ptrs);
+	prt_newline(out);
+
+	prt_str(out, "target:\t");
+	bch2_target_to_text(out, c, data_opts->target);
+	prt_newline(out);
+
+	prt_str(out, "compression:\t");
+	bch2_compression_opt_to_text(out, background_compression(*io_opts));
+	prt_newline(out);
+
+	prt_str(out, "extra replicas:\t");
+	prt_u64(out, data_opts->extra_replicas);
+}
+
+void bch2_data_update_to_text(struct printbuf *out, struct data_update *m)
+{
+	bch2_bkey_val_to_text(out, m->op.c, bkey_i_to_s_c(m->k.k));
+	prt_newline(out);
+	bch2_data_update_opts_to_text(out, m->op.c, &m->op.opts, &m->data_opts);
+}
+
 int bch2_extent_drop_ptrs(struct btree_trans *trans,
 			  struct btree_iter *iter,
 			  struct bkey_s_c k,
@@ -464,7 +504,6 @@ int bch2_extent_drop_ptrs(struct btree_trans *trans,
 
 	while (data_opts.kill_ptrs) {
 		unsigned i = 0, drop = __fls(data_opts.kill_ptrs);
-		struct bch_extent_ptr *ptr;
 
 		bch2_bkey_drop_ptrs(bkey_i_to_s(n), ptr, i++ == drop);
 		data_opts.kill_ptrs ^= 1U << drop;
@@ -479,15 +518,15 @@ int bch2_extent_drop_ptrs(struct btree_trans *trans,
 
 	/*
 	 * Since we're not inserting through an extent iterator
-	 * (BTREE_ITER_ALL_SNAPSHOTS iterators aren't extent iterators),
+	 * (BTREE_ITER_all_snapshots iterators aren't extent iterators),
 	 * we aren't using the extent overwrite path to delete, we're
 	 * just using the normal key deletion path:
 	 */
-	if (bkey_deleted(&n->k) && !(iter->flags & BTREE_ITER_IS_EXTENTS))
+	if (bkey_deleted(&n->k) && !(iter->flags & BTREE_ITER_is_extents))
 		n->k.size = 0;
 
 	return bch2_trans_relock(trans) ?:
-		bch2_trans_update(trans, iter, n, BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE) ?:
+		bch2_trans_update(trans, iter, n, BTREE_UPDATE_internal_snapshot_node) ?:
 		bch2_trans_commit(trans, NULL, NULL, BCH_TRANS_COMMIT_no_enospc);
 }
 
@@ -509,6 +548,14 @@ int bch2_data_update_init(struct btree_trans *trans,
 	unsigned ptrs_locked = 0;
 	int ret = 0;
 
+	/*
+	 * fs is corrupt  we have a key for a snapshot node that doesn't exist,
+	 * and we have to check for this because we go rw before repairing the
+	 * snapshots table - just skip it, we can move it later.
+	 */
+	if (unlikely(k.k->p.snapshot && !bch2_snapshot_equiv(c, k.k->p.snapshot)))
+		return -BCH_ERR_data_update_done;
+
 	bch2_bkey_buf_init(&m->k);
 	bch2_bkey_buf_reassemble(&m->k, c, k);
 	m->btree_id	= btree_id;
@@ -530,15 +577,26 @@ int bch2_data_update_init(struct btree_trans *trans,
 	m->op.compression_opt	= background_compression(io_opts);
 	m->op.watermark		= m->data_opts.btree_insert_flags & BCH_WATERMARK_MASK;
 
-	bkey_for_each_ptr(ptrs, ptr)
-		percpu_ref_get(&bch_dev_bkey_exists(c, ptr->dev)->ref);
+	bkey_for_each_ptr(ptrs, ptr) {
+		if (!bch2_dev_tryget(c, ptr->dev)) {
+			bkey_for_each_ptr(ptrs, ptr2) {
+				if (ptr2 == ptr)
+					break;
+				bch2_dev_put(bch2_dev_have_ref(c, ptr2->dev));
+			}
+			return -BCH_ERR_data_update_done;
+		}
+	}
 
 	unsigned durability_have = 0, durability_removing = 0;
 
 	i = 0;
 	bkey_for_each_ptr_decode(k.k, ptrs, p, entry) {
+		struct bch_dev *ca = bch2_dev_have_ref(c, p.ptr.dev);
+		struct bpos bucket = PTR_BUCKET_POS(ca, &p.ptr);
 		bool locked;
 
+		rcu_read_lock();
 		if (((1U << i) & m->data_opts.rewrite_ptrs)) {
 			BUG_ON(p.ptr.cached);
 
@@ -552,6 +610,7 @@ int bch2_data_update_init(struct btree_trans *trans,
 			bch2_dev_list_add_dev(&m->op.devs_have, p.ptr.dev);
 			durability_have += bch2_extent_ptr_durability(c, &p);
 		}
+		rcu_read_unlock();
 
 		/*
 		 * op->csum_type is normally initialized from the fs/file's
@@ -570,16 +629,13 @@ int bch2_data_update_init(struct btree_trans *trans,
 			if (ctxt) {
 				move_ctxt_wait_event(ctxt,
 						(locked = bch2_bucket_nocow_trylock(&c->nocow_locks,
-									  PTR_BUCKET_POS(c, &p.ptr), 0)) ||
-						(!atomic_read(&ctxt->read_sectors) &&
-						 !atomic_read(&ctxt->write_sectors)));
+									  bucket, 0)) ||
+						list_empty(&ctxt->ios));
 
 				if (!locked)
-					bch2_bucket_nocow_lock(&c->nocow_locks,
-							       PTR_BUCKET_POS(c, &p.ptr), 0);
+					bch2_bucket_nocow_lock(&c->nocow_locks, bucket, 0);
 			} else {
-				if (!bch2_bucket_nocow_trylock(&c->nocow_locks,
-							       PTR_BUCKET_POS(c, &p.ptr), 0)) {
+				if (!bch2_bucket_nocow_trylock(&c->nocow_locks, bucket, 0)) {
 					ret = -BCH_ERR_nocow_lock_blocked;
 					goto err;
 				}
@@ -590,6 +646,8 @@ int bch2_data_update_init(struct btree_trans *trans,
 		i++;
 	}
 
+	unsigned durability_required = max(0, (int) (io_opts.data_replicas - durability_have));
+
 	/*
 	 * If current extent durability is less than io_opts.data_replicas,
 	 * we're not trying to rereplicate the extent up to data_replicas here -
@@ -599,7 +657,7 @@ int bch2_data_update_init(struct btree_trans *trans,
 	 * rereplicate, currently, so that users don't get an unexpected -ENOSPC
 	 */
 	if (!(m->data_opts.write_flags & BCH_WRITE_CACHED) &&
-	    durability_have >= io_opts.data_replicas) {
+	    !durability_required) {
 		m->data_opts.kill_ptrs |= m->data_opts.rewrite_ptrs;
 		m->data_opts.rewrite_ptrs = 0;
 		/* if iter == NULL, it's just a promote */
@@ -608,11 +666,28 @@ int bch2_data_update_init(struct btree_trans *trans,
 		goto done;
 	}
 
-	m->op.nr_replicas = min(durability_removing, io_opts.data_replicas - durability_have) +
+	m->op.nr_replicas = min(durability_removing, durability_required) +
 		m->data_opts.extra_replicas;
-	m->op.nr_replicas_required = m->op.nr_replicas;
 
-	BUG_ON(!m->op.nr_replicas);
+	/*
+	 * If device(s) were set to durability=0 after data was written to them
+	 * we can end up with a duribilty=0 extent, and the normal algorithm
+	 * that tries not to increase durability doesn't work:
+	 */
+	if (!(durability_have + durability_removing))
+		m->op.nr_replicas = max((unsigned) m->op.nr_replicas, 1);
+
+	if (!m->op.nr_replicas) {
+		struct printbuf buf = PRINTBUF;
+
+		bch2_data_update_to_text(&buf, m);
+		WARN(1, "trying to move an extent, but nr_replicas=0\n%s", buf.buf);
+		printbuf_exit(&buf);
+		ret = -BCH_ERR_data_update_done;
+		goto done;
+	}
+
+	m->op.nr_replicas_required = m->op.nr_replicas;
 
 	if (reserve_sectors) {
 		ret = bch2_disk_reservation_add(c, &m->op.res, reserve_sectors,
@@ -632,10 +707,11 @@ int bch2_data_update_init(struct btree_trans *trans,
 err:
 	i = 0;
 	bkey_for_each_ptr_decode(k.k, ptrs, p, entry) {
+		struct bch_dev *ca = bch2_dev_have_ref(c, p.ptr.dev);
+		struct bpos bucket = PTR_BUCKET_POS(ca, &p.ptr);
 		if ((1U << i) & ptrs_locked)
-			bch2_bucket_nocow_unlock(&c->nocow_locks,
-						 PTR_BUCKET_POS(c, &p.ptr), 0);
-		percpu_ref_put(&bch_dev_bkey_exists(c, p.ptr.dev)->ref);
+			bch2_bucket_nocow_unlock(&c->nocow_locks, bucket, 0);
+		bch2_dev_put(ca);
 		i++;
 	}
 
diff --git a/fs/bcachefs/data_update.h b/fs/bcachefs/data_update.h
index 991095bbd4..8d36365bde 100644
--- a/fs/bcachefs/data_update.h
+++ b/fs/bcachefs/data_update.h
@@ -17,6 +17,9 @@ struct data_update_opts {
 	unsigned	write_flags;
 };
 
+void bch2_data_update_opts_to_text(struct printbuf *, struct bch_fs *,
+				   struct bch_io_opts *, struct data_update_opts *);
+
 struct data_update {
 	/* extent being updated: */
 	enum btree_id		btree_id;
@@ -27,6 +30,8 @@ struct data_update {
 	struct bch_write_op	op;
 };
 
+void bch2_data_update_to_text(struct printbuf *, struct data_update *);
+
 int bch2_data_update_index_update(struct bch_write_op *);
 
 void bch2_data_update_read_done(struct data_update *,
diff --git a/fs/bcachefs/debug.c b/fs/bcachefs/debug.c
index 7bdba8507f..ebabab171f 100644
--- a/fs/bcachefs/debug.c
+++ b/fs/bcachefs/debug.c
@@ -13,6 +13,7 @@
 #include "btree_iter.h"
 #include "btree_locking.h"
 #include "btree_update.h"
+#include "btree_update_interior.h"
 #include "buckets.h"
 #include "debug.h"
 #include "error.h"
@@ -36,11 +37,11 @@ static bool bch2_btree_verify_replica(struct bch_fs *c, struct btree *b,
 	struct btree_node *n_ondisk = c->verify_ondisk;
 	struct btree_node *n_sorted = c->verify_data->data;
 	struct bset *sorted, *inmemory = &b->data->keys;
-	struct bch_dev *ca = bch_dev_bkey_exists(c, pick.ptr.dev);
 	struct bio *bio;
 	bool failed = false, saw_error = false;
 
-	if (!bch2_dev_get_ioref(ca, READ))
+	struct bch_dev *ca = bch2_dev_get_ioref(c, pick.ptr.dev, READ);
+	if (!ca)
 		return false;
 
 	bio = bio_alloc_bioset(ca->disk_sb.bdev,
@@ -137,7 +138,7 @@ void __bch2_btree_verify(struct bch_fs *c, struct btree *b)
 	mutex_lock(&c->verify_lock);
 
 	if (!c->verify_ondisk) {
-		c->verify_ondisk = kvpmalloc(btree_buf_bytes(b), GFP_KERNEL);
+		c->verify_ondisk = kvmalloc(btree_buf_bytes(b), GFP_KERNEL);
 		if (!c->verify_ondisk)
 			goto out;
 	}
@@ -170,7 +171,7 @@ void __bch2_btree_verify(struct bch_fs *c, struct btree *b)
 		struct printbuf buf = PRINTBUF;
 
 		bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&b->key));
-		bch2_fs_fatal_error(c, "btree node verify failed for : %s\n", buf.buf);
+		bch2_fs_fatal_error(c, ": btree node verify failed for: %s\n", buf.buf);
 		printbuf_exit(&buf);
 	}
 out:
@@ -193,13 +194,13 @@ void bch2_btree_node_ondisk_to_text(struct printbuf *out, struct bch_fs *c,
 		return;
 	}
 
-	ca = bch_dev_bkey_exists(c, pick.ptr.dev);
-	if (!bch2_dev_get_ioref(ca, READ)) {
+	ca = bch2_dev_get_ioref(c, pick.ptr.dev, READ);
+	if (!ca) {
 		prt_printf(out, "error getting device to read from: not online\n");
 		return;
 	}
 
-	n_ondisk = kvpmalloc(btree_buf_bytes(b), GFP_KERNEL);
+	n_ondisk = kvmalloc(btree_buf_bytes(b), GFP_KERNEL);
 	if (!n_ondisk) {
 		prt_printf(out, "memory allocation failure\n");
 		goto out;
@@ -293,7 +294,7 @@ void bch2_btree_node_ondisk_to_text(struct printbuf *out, struct bch_fs *c,
 out:
 	if (bio)
 		bio_put(bio);
-	kvpfree(n_ondisk, btree_buf_bytes(b));
+	kvfree(n_ondisk);
 	percpu_ref_put(&ca->io_ref);
 }
 
@@ -374,8 +375,8 @@ static ssize_t bch2_read_btree(struct file *file, char __user *buf,
 	return flush_buf(i) ?:
 		bch2_trans_run(i->c,
 			for_each_btree_key(trans, iter, i->id, i->from,
-					   BTREE_ITER_PREFETCH|
-					   BTREE_ITER_ALL_SNAPSHOTS, k, ({
+					   BTREE_ITER_prefetch|
+					   BTREE_ITER_all_snapshots, k, ({
 				bch2_bkey_val_to_text(&i->buf, i->c, k);
 				prt_newline(&i->buf);
 				bch2_trans_unlock(trans);
@@ -458,8 +459,8 @@ static ssize_t bch2_read_bfloat_failed(struct file *file, char __user *buf,
 	return flush_buf(i) ?:
 		bch2_trans_run(i->c,
 			for_each_btree_key(trans, iter, i->id, i->from,
-					   BTREE_ITER_PREFETCH|
-					   BTREE_ITER_ALL_SNAPSHOTS, k, ({
+					   BTREE_ITER_prefetch|
+					   BTREE_ITER_all_snapshots, k, ({
 				struct btree_path_level *l =
 					&btree_iter_path(trans, &iter)->l[0];
 				struct bkey_packed *_k =
@@ -491,51 +492,26 @@ static void bch2_cached_btree_node_to_text(struct printbuf *out, struct bch_fs *
 	if (!out->nr_tabstops)
 		printbuf_tabstop_push(out, 32);
 
-	prt_printf(out, "%px btree=%s l=%u ",
-	       b,
-	       bch2_btree_id_str(b->c.btree_id),
-	       b->c.level);
-	prt_newline(out);
+	prt_printf(out, "%px btree=%s l=%u\n", b, bch2_btree_id_str(b->c.btree_id), b->c.level);
 
 	printbuf_indent_add(out, 2);
 
 	bch2_bkey_val_to_text(out, c, bkey_i_to_s_c(&b->key));
 	prt_newline(out);
 
-	prt_printf(out, "flags: ");
-	prt_tab(out);
+	prt_printf(out, "flags:\t");
 	prt_bitflags(out, bch2_btree_node_flags, b->flags);
 	prt_newline(out);
 
-	prt_printf(out, "pcpu read locks: ");
-	prt_tab(out);
-	prt_printf(out, "%u", b->c.lock.readers != NULL);
-	prt_newline(out);
-
-	prt_printf(out, "written:");
-	prt_tab(out);
-	prt_printf(out, "%u", b->written);
-	prt_newline(out);
-
-	prt_printf(out, "writes blocked:");
-	prt_tab(out);
-	prt_printf(out, "%u", !list_empty_careful(&b->write_blocked));
-	prt_newline(out);
+	prt_printf(out, "pcpu read locks:\t%u\n",	b->c.lock.readers != NULL);
+	prt_printf(out, "written:\t%u\n",		b->written);
+	prt_printf(out, "writes blocked:\t%u\n",	!list_empty_careful(&b->write_blocked));
+	prt_printf(out, "will make reachable:\t%lx\n",	b->will_make_reachable);
 
-	prt_printf(out, "will make reachable:");
-	prt_tab(out);
-	prt_printf(out, "%lx", b->will_make_reachable);
-	prt_newline(out);
-
-	prt_printf(out, "journal pin %px:", &b->writes[0].journal);
-	prt_tab(out);
-	prt_printf(out, "%llu", b->writes[0].journal.seq);
-	prt_newline(out);
-
-	prt_printf(out, "journal pin %px:", &b->writes[1].journal);
-	prt_tab(out);
-	prt_printf(out, "%llu", b->writes[1].journal.seq);
-	prt_newline(out);
+	prt_printf(out, "journal pin %px:\t%llu\n",
+		   &b->writes[0].journal, b->writes[0].journal.seq);
+	prt_printf(out, "journal pin %px:\t%llu\n",
+		   &b->writes[1].journal, b->writes[1].journal.seq);
 
 	printbuf_indent_sub(out, 2);
 }
@@ -592,6 +568,32 @@ static const struct file_operations cached_btree_nodes_ops = {
 	.read		= bch2_cached_btree_nodes_read,
 };
 
+typedef int (*list_cmp_fn)(const struct list_head *l, const struct list_head *r);
+
+static void list_sort(struct list_head *head, list_cmp_fn cmp)
+{
+	struct list_head *pos;
+
+	list_for_each(pos, head)
+		while (!list_is_last(pos, head) &&
+		       cmp(pos, pos->next) > 0) {
+			struct list_head *pos2, *next = pos->next;
+
+			list_del(next);
+			list_for_each(pos2, head)
+				if (cmp(next, pos2) < 0)
+					goto pos_found;
+			BUG();
+pos_found:
+			list_add_tail(next, pos2);
+		}
+}
+
+static int list_ptr_order_cmp(const struct list_head *l, const struct list_head *r)
+{
+	return cmp_int(l, r);
+}
+
 static ssize_t bch2_btree_transactions_read(struct file *file, char __user *buf,
 					    size_t size, loff_t *ppos)
 {
@@ -599,42 +601,39 @@ static ssize_t bch2_btree_transactions_read(struct file *file, char __user *buf,
 	struct bch_fs *c = i->c;
 	struct btree_trans *trans;
 	ssize_t ret = 0;
-	u32 seq;
 
 	i->ubuf = buf;
 	i->size	= size;
 	i->ret	= 0;
 restart:
 	seqmutex_lock(&c->btree_trans_lock);
-	list_for_each_entry(trans, &c->btree_trans_list, list) {
-		struct task_struct *task = READ_ONCE(trans->locking_wait.task);
+	list_sort(&c->btree_trans_list, list_ptr_order_cmp);
 
-		if (!task || task->pid <= i->iter)
+	list_for_each_entry(trans, &c->btree_trans_list, list) {
+		if ((ulong) trans <= i->iter)
 			continue;
 
-		closure_get(&trans->ref);
-		seq = seqmutex_seq(&c->btree_trans_lock);
-		seqmutex_unlock(&c->btree_trans_lock);
+		i->iter = (ulong) trans;
 
-		ret = flush_buf(i);
-		if (ret) {
-			closure_put(&trans->ref);
-			goto unlocked;
-		}
+		if (!closure_get_not_zero(&trans->ref))
+			continue;
+
+		u32 seq = seqmutex_unlock(&c->btree_trans_lock);
 
 		bch2_btree_trans_to_text(&i->buf, trans);
 
-		prt_printf(&i->buf, "backtrace:");
-		prt_newline(&i->buf);
+		prt_printf(&i->buf, "backtrace:\n");
 		printbuf_indent_add(&i->buf, 2);
-		bch2_prt_task_backtrace(&i->buf, task, 0, GFP_KERNEL);
+		bch2_prt_task_backtrace(&i->buf, trans->locking_wait.task, 0, GFP_KERNEL);
 		printbuf_indent_sub(&i->buf, 2);
 		prt_newline(&i->buf);
 
-		i->iter = task->pid;
-
 		closure_put(&trans->ref);
 
+		ret = flush_buf(i);
+		if (ret)
+			goto unlocked;
+
 		if (!seqmutex_relock(&c->btree_trans_lock, seq))
 			goto restart;
 	}
@@ -668,7 +667,7 @@ static ssize_t bch2_journal_pins_read(struct file *file, char __user *buf,
 	i->size	= size;
 	i->ret	= 0;
 
-	do {
+	while (1) {
 		err = flush_buf(i);
 		if (err)
 			return err;
@@ -676,9 +675,12 @@ static ssize_t bch2_journal_pins_read(struct file *file, char __user *buf,
 		if (!i->size)
 			break;
 
+		if (done)
+			break;
+
 		done = bch2_journal_seq_pins_to_text(&i->buf, &c->journal, &i->iter);
 		i->iter++;
-	} while (!done);
+	}
 
 	if (i->buf.allocation_failure)
 		return -ENOMEM;
@@ -693,13 +695,45 @@ static const struct file_operations journal_pins_ops = {
 	.read		= bch2_journal_pins_read,
 };
 
+static ssize_t bch2_btree_updates_read(struct file *file, char __user *buf,
+				       size_t size, loff_t *ppos)
+{
+	struct dump_iter *i = file->private_data;
+	struct bch_fs *c = i->c;
+	int err;
+
+	i->ubuf = buf;
+	i->size	= size;
+	i->ret	= 0;
+
+	if (!i->iter) {
+		bch2_btree_updates_to_text(&i->buf, c);
+		i->iter++;
+	}
+
+	err = flush_buf(i);
+	if (err)
+		return err;
+
+	if (i->buf.allocation_failure)
+		return -ENOMEM;
+
+	return i->ret;
+}
+
+static const struct file_operations btree_updates_ops = {
+	.owner		= THIS_MODULE,
+	.open		= bch2_dump_open,
+	.release	= bch2_dump_release,
+	.read		= bch2_btree_updates_read,
+};
+
 static int btree_transaction_stats_open(struct inode *inode, struct file *file)
 {
 	struct bch_fs *c = inode->i_private;
 	struct dump_iter *i;
 
 	i = kzalloc(sizeof(struct dump_iter), GFP_KERNEL);
-
 	if (!i)
 		return -ENOMEM;
 
@@ -746,25 +780,20 @@ static ssize_t btree_transaction_stats_read(struct file *file, char __user *buf,
 		    !bch2_btree_transaction_fns[i->iter])
 			break;
 
-		prt_printf(&i->buf, "%s: ", bch2_btree_transaction_fns[i->iter]);
-		prt_newline(&i->buf);
+		prt_printf(&i->buf, "%s:\n", bch2_btree_transaction_fns[i->iter]);
 		printbuf_indent_add(&i->buf, 2);
 
 		mutex_lock(&s->lock);
 
-		prt_printf(&i->buf, "Max mem used: %u", s->max_mem);
-		prt_newline(&i->buf);
-
-		prt_printf(&i->buf, "Transaction duration:");
-		prt_newline(&i->buf);
+		prt_printf(&i->buf, "Max mem used: %u\n", s->max_mem);
+		prt_printf(&i->buf, "Transaction duration:\n");
 
 		printbuf_indent_add(&i->buf, 2);
 		bch2_time_stats_to_text(&i->buf, &s->duration);
 		printbuf_indent_sub(&i->buf, 2);
 
 		if (IS_ENABLED(CONFIG_BCACHEFS_LOCK_TIME_STATS)) {
-			prt_printf(&i->buf, "Lock hold times:");
-			prt_newline(&i->buf);
+			prt_printf(&i->buf, "Lock hold times:\n");
 
 			printbuf_indent_add(&i->buf, 2);
 			bch2_time_stats_to_text(&i->buf, &s->lock_hold_times);
@@ -772,8 +801,7 @@ static ssize_t btree_transaction_stats_read(struct file *file, char __user *buf,
 		}
 
 		if (s->max_paths_text) {
-			prt_printf(&i->buf, "Maximum allocated btree paths (%u):", s->nr_max_paths);
-			prt_newline(&i->buf);
+			prt_printf(&i->buf, "Maximum allocated btree paths (%u):\n", s->nr_max_paths);
 
 			printbuf_indent_add(&i->buf, 2);
 			prt_str_indented(&i->buf, s->max_paths_text);
@@ -800,50 +828,55 @@ static const struct file_operations btree_transaction_stats_op = {
 	.read		= btree_transaction_stats_read,
 };
 
-static ssize_t bch2_btree_deadlock_read(struct file *file, char __user *buf,
-					    size_t size, loff_t *ppos)
+/* walk btree transactions until we find a deadlock and print it */
+static void btree_deadlock_to_text(struct printbuf *out, struct bch_fs *c)
 {
-	struct dump_iter *i = file->private_data;
-	struct bch_fs *c = i->c;
 	struct btree_trans *trans;
-	ssize_t ret = 0;
-	u32 seq;
-
-	i->ubuf = buf;
-	i->size	= size;
-	i->ret	= 0;
-
-	if (i->iter)
-		goto out;
+	ulong iter = 0;
 restart:
 	seqmutex_lock(&c->btree_trans_lock);
-	list_for_each_entry(trans, &c->btree_trans_list, list) {
-		struct task_struct *task = READ_ONCE(trans->locking_wait.task);
+	list_sort(&c->btree_trans_list, list_ptr_order_cmp);
 
-		if (!task || task->pid <= i->iter)
+	list_for_each_entry(trans, &c->btree_trans_list, list) {
+		if ((ulong) trans <= iter)
 			continue;
 
-		closure_get(&trans->ref);
-		seq = seqmutex_seq(&c->btree_trans_lock);
-		seqmutex_unlock(&c->btree_trans_lock);
+		iter = (ulong) trans;
 
-		ret = flush_buf(i);
-		if (ret) {
-			closure_put(&trans->ref);
-			goto out;
-		}
+		if (!closure_get_not_zero(&trans->ref))
+			continue;
 
-		bch2_check_for_deadlock(trans, &i->buf);
+		u32 seq = seqmutex_unlock(&c->btree_trans_lock);
 
-		i->iter = task->pid;
+		bool found = bch2_check_for_deadlock(trans, out) != 0;
 
 		closure_put(&trans->ref);
 
+		if (found)
+			return;
+
 		if (!seqmutex_relock(&c->btree_trans_lock, seq))
 			goto restart;
 	}
 	seqmutex_unlock(&c->btree_trans_lock);
-out:
+}
+
+static ssize_t bch2_btree_deadlock_read(struct file *file, char __user *buf,
+					    size_t size, loff_t *ppos)
+{
+	struct dump_iter *i = file->private_data;
+	struct bch_fs *c = i->c;
+	ssize_t ret = 0;
+
+	i->ubuf = buf;
+	i->size	= size;
+	i->ret	= 0;
+
+	if (!i->iter) {
+		btree_deadlock_to_text(&i->buf, c);
+		i->iter++;
+	}
+
 	if (i->buf.allocation_failure)
 		ret = -ENOMEM;
 
@@ -866,6 +899,20 @@ void bch2_fs_debug_exit(struct bch_fs *c)
 		debugfs_remove_recursive(c->fs_debug_dir);
 }
 
+static void bch2_fs_debug_btree_init(struct bch_fs *c, struct btree_debug *bd)
+{
+	struct dentry *d;
+
+	d = debugfs_create_dir(bch2_btree_id_str(bd->id), c->btree_debug_dir);
+
+	debugfs_create_file("keys", 0400, d, bd, &btree_debug_ops);
+
+	debugfs_create_file("formats", 0400, d, bd, &btree_format_debug_ops);
+
+	debugfs_create_file("bfloat-failed", 0400, d, bd,
+			    &bfloat_failed_debug_ops);
+}
+
 void bch2_fs_debug_init(struct bch_fs *c)
 {
 	struct btree_debug *bd;
@@ -888,6 +935,9 @@ void bch2_fs_debug_init(struct bch_fs *c)
 	debugfs_create_file("journal_pins", 0400, c->fs_debug_dir,
 			    c->btree_debug, &journal_pins_ops);
 
+	debugfs_create_file("btree_updates", 0400, c->fs_debug_dir,
+			    c->btree_debug, &btree_updates_ops);
+
 	debugfs_create_file("btree_transaction_stats", 0400, c->fs_debug_dir,
 			    c, &btree_transaction_stats_op);
 
@@ -902,21 +952,7 @@ void bch2_fs_debug_init(struct bch_fs *c)
 	     bd < c->btree_debug + ARRAY_SIZE(c->btree_debug);
 	     bd++) {
 		bd->id = bd - c->btree_debug;
-		debugfs_create_file(bch2_btree_id_str(bd->id),
-				    0400, c->btree_debug_dir, bd,
-				    &btree_debug_ops);
-
-		snprintf(name, sizeof(name), "%s-formats",
-			 bch2_btree_id_str(bd->id));
-
-		debugfs_create_file(name, 0400, c->btree_debug_dir, bd,
-				    &btree_format_debug_ops);
-
-		snprintf(name, sizeof(name), "%s-bfloat-failed",
-			 bch2_btree_id_str(bd->id));
-
-		debugfs_create_file(name, 0400, c->btree_debug_dir, bd,
-				    &bfloat_failed_debug_ops);
+		bch2_fs_debug_btree_init(c, bd);
 	}
 }
 
diff --git a/fs/bcachefs/dirent.c b/fs/bcachefs/dirent.c
index 4ae1e9f002..c67460d820 100644
--- a/fs/bcachefs/dirent.c
+++ b/fs/bcachefs/dirent.c
@@ -15,6 +15,9 @@
 
 static unsigned bch2_dirent_name_bytes(struct bkey_s_c_dirent d)
 {
+	if (bkey_val_bytes(d.k) < offsetof(struct bch_dirent, d_name))
+		return 0;
+
 	unsigned bkey_u64s = bkey_val_u64s(d.k);
 	unsigned bkey_bytes = bkey_u64s * sizeof(u64);
 	u64 last_u64 = ((u64*)d.v)[bkey_u64s - 1];
@@ -98,7 +101,7 @@ const struct bch_hash_desc bch2_dirent_hash_desc = {
 };
 
 int bch2_dirent_invalid(struct bch_fs *c, struct bkey_s_c k,
-			enum bkey_invalid_flags flags,
+			enum bch_validate_flags flags,
 			struct printbuf *err)
 {
 	struct bkey_s_c_dirent d = bkey_s_c_to_dirent(k);
@@ -118,7 +121,7 @@ int bch2_dirent_invalid(struct bch_fs *c, struct bkey_s_c k,
 	 * Check new keys don't exceed the max length
 	 * (older keys may be larger.)
 	 */
-	bkey_fsck_err_on((flags & BKEY_INVALID_COMMIT) && d_name.len > BCH_NAME_MAX, c, err,
+	bkey_fsck_err_on((flags & BCH_VALIDATE_commit) && d_name.len > BCH_NAME_MAX, c, err,
 			 dirent_name_too_long,
 			 "dirent name too big (%u > %u)",
 			 d_name.len, BCH_NAME_MAX);
@@ -144,19 +147,21 @@ fsck_err:
 	return ret;
 }
 
-void bch2_dirent_to_text(struct printbuf *out, struct bch_fs *c,
-			 struct bkey_s_c k)
+void bch2_dirent_to_text(struct printbuf *out, struct bch_fs *c, struct bkey_s_c k)
 {
 	struct bkey_s_c_dirent d = bkey_s_c_to_dirent(k);
 	struct qstr d_name = bch2_dirent_get_name(d);
 
-	prt_printf(out, "%.*s -> %llu type %s",
-	       d_name.len,
-	       d_name.name,
-	       d.v->d_type != DT_SUBVOL
-	       ? le64_to_cpu(d.v->d_inum)
-	       : le32_to_cpu(d.v->d_child_subvol),
-	       bch2_d_type_str(d.v->d_type));
+	prt_printf(out, "%.*s -> ", d_name.len, d_name.name);
+
+	if (d.v->d_type != DT_SUBVOL)
+		prt_printf(out, "%llu", le64_to_cpu(d.v->d_inum));
+	else
+		prt_printf(out, "%u -> %u",
+			   le32_to_cpu(d.v->d_parent_subvol),
+			   le32_to_cpu(d.v->d_child_subvol));
+
+	prt_printf(out, " type %s", bch2_d_type_str(d.v->d_type));
 }
 
 static struct bkey_i_dirent *dirent_create_key(struct btree_trans *trans,
@@ -199,17 +204,17 @@ static struct bkey_i_dirent *dirent_create_key(struct btree_trans *trans,
 }
 
 int bch2_dirent_create_snapshot(struct btree_trans *trans,
-			u64 dir, u32 snapshot,
+			u32 dir_subvol, u64 dir, u32 snapshot,
 			const struct bch_hash_info *hash_info,
 			u8 type, const struct qstr *name, u64 dst_inum,
 			u64 *dir_offset,
-			bch_str_hash_flags_t str_hash_flags)
+			enum btree_iter_update_trigger_flags flags)
 {
-	subvol_inum zero_inum = { 0 };
+	subvol_inum dir_inum = { .subvol = dir_subvol, .inum = dir };
 	struct bkey_i_dirent *dirent;
 	int ret;
 
-	dirent = dirent_create_key(trans, zero_inum, type, name, dst_inum);
+	dirent = dirent_create_key(trans, dir_inum, type, name, dst_inum);
 	ret = PTR_ERR_OR_ZERO(dirent);
 	if (ret)
 		return ret;
@@ -217,10 +222,9 @@ int bch2_dirent_create_snapshot(struct btree_trans *trans,
 	dirent->k.p.inode	= dir;
 	dirent->k.p.snapshot	= snapshot;
 
-	ret = bch2_hash_set_snapshot(trans, bch2_dirent_hash_desc, hash_info,
-				     zero_inum, snapshot,
-				     &dirent->k_i, str_hash_flags,
-				     BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
+	ret = bch2_hash_set_in_snapshot(trans, bch2_dirent_hash_desc, hash_info,
+					dir_inum, snapshot, &dirent->k_i,
+					flags|BTREE_UPDATE_internal_snapshot_node);
 	*dir_offset = dirent->k.p.offset;
 
 	return ret;
@@ -230,7 +234,7 @@ int bch2_dirent_create(struct btree_trans *trans, subvol_inum dir,
 		       const struct bch_hash_info *hash_info,
 		       u8 type, const struct qstr *name, u64 dst_inum,
 		       u64 *dir_offset,
-		       bch_str_hash_flags_t str_hash_flags)
+		       enum btree_iter_update_trigger_flags flags)
 {
 	struct bkey_i_dirent *dirent;
 	int ret;
@@ -241,7 +245,7 @@ int bch2_dirent_create(struct btree_trans *trans, subvol_inum dir,
 		return ret;
 
 	ret = bch2_hash_set(trans, bch2_dirent_hash_desc, hash_info,
-			    dir, &dirent->k_i, str_hash_flags);
+			    dir, &dirent->k_i, flags);
 	*dir_offset = dirent->k.p.offset;
 
 	return ret;
@@ -270,7 +274,7 @@ int bch2_dirent_read_target(struct btree_trans *trans, subvol_inum dir,
 	} else {
 		target->subvol	= le32_to_cpu(d.v->d_child_subvol);
 
-		ret = bch2_subvolume_get(trans, target->subvol, true, BTREE_ITER_CACHED, &s);
+		ret = bch2_subvolume_get(trans, target->subvol, true, BTREE_ITER_cached, &s);
 
 		target->inum	= le64_to_cpu(s.inode);
 	}
@@ -291,23 +295,17 @@ int bch2_dirent_rename(struct btree_trans *trans,
 	struct bkey_i_dirent *new_src = NULL, *new_dst = NULL;
 	struct bpos dst_pos =
 		POS(dst_dir.inum, bch2_dirent_hash(dst_hash, dst_name));
-	unsigned src_type = 0, dst_type = 0, src_update_flags = 0;
+	unsigned src_update_flags = 0;
+	bool delete_src, delete_dst;
 	int ret = 0;
 
-	if (src_dir.subvol != dst_dir.subvol)
-		return -EXDEV;
-
 	memset(src_inum, 0, sizeof(*src_inum));
 	memset(dst_inum, 0, sizeof(*dst_inum));
 
 	/* Lookup src: */
-	ret = bch2_hash_lookup(trans, &src_iter, bch2_dirent_hash_desc,
-			       src_hash, src_dir, src_name,
-			       BTREE_ITER_INTENT);
-	if (ret)
-		goto out;
-
-	old_src = bch2_btree_iter_peek_slot(&src_iter);
+	old_src = bch2_hash_lookup(trans, &src_iter, bch2_dirent_hash_desc,
+				   src_hash, src_dir, src_name,
+				   BTREE_ITER_intent);
 	ret = bkey_err(old_src);
 	if (ret)
 		goto out;
@@ -317,12 +315,6 @@ int bch2_dirent_rename(struct btree_trans *trans,
 	if (ret)
 		goto out;
 
-	src_type = bkey_s_c_to_dirent(old_src).v->d_type;
-
-	if (src_type == DT_SUBVOL && mode == BCH_RENAME_EXCHANGE)
-		return -EOPNOTSUPP;
-
-
 	/* Lookup dst: */
 	if (mode == BCH_RENAME) {
 		/*
@@ -335,13 +327,9 @@ int bch2_dirent_rename(struct btree_trans *trans,
 		if (ret)
 			goto out;
 	} else {
-		ret = bch2_hash_lookup(trans, &dst_iter, bch2_dirent_hash_desc,
-				       dst_hash, dst_dir, dst_name,
-				       BTREE_ITER_INTENT);
-		if (ret)
-			goto out;
-
-		old_dst = bch2_btree_iter_peek_slot(&dst_iter);
+		old_dst = bch2_hash_lookup(trans, &dst_iter, bch2_dirent_hash_desc,
+					    dst_hash, dst_dir, dst_name,
+					    BTREE_ITER_intent);
 		ret = bkey_err(old_dst);
 		if (ret)
 			goto out;
@@ -350,11 +338,6 @@ int bch2_dirent_rename(struct btree_trans *trans,
 				bkey_s_c_to_dirent(old_dst), dst_inum);
 		if (ret)
 			goto out;
-
-		dst_type = bkey_s_c_to_dirent(old_dst).v->d_type;
-
-		if (dst_type == DT_SUBVOL)
-			return -EOPNOTSUPP;
 	}
 
 	if (mode != BCH_RENAME_EXCHANGE)
@@ -424,28 +407,55 @@ int bch2_dirent_rename(struct btree_trans *trans,
 		}
 	}
 
+	if (new_dst->v.d_type == DT_SUBVOL)
+		new_dst->v.d_parent_subvol = cpu_to_le32(dst_dir.subvol);
+
+	if ((mode == BCH_RENAME_EXCHANGE) &&
+	    new_src->v.d_type == DT_SUBVOL)
+		new_src->v.d_parent_subvol = cpu_to_le32(src_dir.subvol);
+
 	ret = bch2_trans_update(trans, &dst_iter, &new_dst->k_i, 0);
 	if (ret)
 		goto out;
 out_set_src:
-
 	/*
-	 * If we're deleting a subvolume, we need to really delete the dirent,
-	 * not just emit a whiteout in the current snapshot:
+	 * If we're deleting a subvolume we need to really delete the dirent,
+	 * not just emit a whiteout in the current snapshot - there can only be
+	 * single dirent that points to a given subvolume.
+	 *
+	 * IOW, we don't maintain multiple versions in different snapshots of
+	 * dirents that point to subvolumes - dirents that point to subvolumes
+	 * are only visible in one particular subvolume so it's not necessary,
+	 * and it would be particularly confusing for fsck to have to deal with.
 	 */
-	if (src_type == DT_SUBVOL) {
-		bch2_btree_iter_set_snapshot(&src_iter, old_src.k->p.snapshot);
-		ret = bch2_btree_iter_traverse(&src_iter);
+	delete_src = bkey_s_c_to_dirent(old_src).v->d_type == DT_SUBVOL &&
+		new_src->k.p.snapshot != old_src.k->p.snapshot;
+
+	delete_dst = old_dst.k &&
+		bkey_s_c_to_dirent(old_dst).v->d_type == DT_SUBVOL &&
+		new_dst->k.p.snapshot != old_dst.k->p.snapshot;
+
+	if (!delete_src || !bkey_deleted(&new_src->k)) {
+		ret = bch2_trans_update(trans, &src_iter, &new_src->k_i, src_update_flags);
 		if (ret)
 			goto out;
+	}
 
-		new_src->k.p = src_iter.pos;
-		src_update_flags |= BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE;
+	if (delete_src) {
+		bch2_btree_iter_set_snapshot(&src_iter, old_src.k->p.snapshot);
+		ret =   bch2_btree_iter_traverse(&src_iter) ?:
+			bch2_btree_delete_at(trans, &src_iter, BTREE_UPDATE_internal_snapshot_node);
+		if (ret)
+			goto out;
 	}
 
-	ret = bch2_trans_update(trans, &src_iter, &new_src->k_i, src_update_flags);
-	if (ret)
-		goto out;
+	if (delete_dst) {
+		bch2_btree_iter_set_snapshot(&dst_iter, old_dst.k->p.snapshot);
+		ret =   bch2_btree_iter_traverse(&dst_iter) ?:
+			bch2_btree_delete_at(trans, &dst_iter, BTREE_UPDATE_internal_snapshot_node);
+		if (ret)
+			goto out;
+	}
 
 	if (mode == BCH_RENAME_EXCHANGE)
 		*src_offset = new_src->k.p.offset;
@@ -456,41 +466,25 @@ out:
 	return ret;
 }
 
-int __bch2_dirent_lookup_trans(struct btree_trans *trans,
-			       struct btree_iter *iter,
-			       subvol_inum dir,
-			       const struct bch_hash_info *hash_info,
-			       const struct qstr *name, subvol_inum *inum,
-			       unsigned flags)
+int bch2_dirent_lookup_trans(struct btree_trans *trans,
+			     struct btree_iter *iter,
+			     subvol_inum dir,
+			     const struct bch_hash_info *hash_info,
+			     const struct qstr *name, subvol_inum *inum,
+			     unsigned flags)
 {
-	struct bkey_s_c k;
-	struct bkey_s_c_dirent d;
-	u32 snapshot;
-	int ret;
-
-	ret = bch2_subvolume_get_snapshot(trans, dir.subvol, &snapshot);
-	if (ret)
-		return ret;
-
-	ret = bch2_hash_lookup(trans, iter, bch2_dirent_hash_desc,
-			       hash_info, dir, name, flags);
-	if (ret)
-		return ret;
-
-	k = bch2_btree_iter_peek_slot(iter);
-	ret = bkey_err(k);
+	struct bkey_s_c k = bch2_hash_lookup(trans, iter, bch2_dirent_hash_desc,
+					     hash_info, dir, name, flags);
+	int ret = bkey_err(k);
 	if (ret)
 		goto err;
 
-	d = bkey_s_c_to_dirent(k);
-
-	ret = bch2_dirent_read_target(trans, dir, d, inum);
+	ret = bch2_dirent_read_target(trans, dir, bkey_s_c_to_dirent(k), inum);
 	if (ret > 0)
 		ret = -ENOENT;
 err:
 	if (ret)
 		bch2_trans_iter_exit(trans, iter);
-
 	return ret;
 }
 
@@ -502,13 +496,13 @@ u64 bch2_dirent_lookup(struct bch_fs *c, subvol_inum dir,
 	struct btree_iter iter = { NULL };
 
 	int ret = lockrestart_do(trans,
-		__bch2_dirent_lookup_trans(trans, &iter, dir, hash_info, name, inum, 0));
+		bch2_dirent_lookup_trans(trans, &iter, dir, hash_info, name, inum, 0));
 	bch2_trans_iter_exit(trans, &iter);
 	bch2_trans_put(trans);
 	return ret;
 }
 
-int bch2_empty_dir_snapshot(struct btree_trans *trans, u64 dir, u32 snapshot)
+int bch2_empty_dir_snapshot(struct btree_trans *trans, u64 dir, u32 subvol, u32 snapshot)
 {
 	struct btree_iter iter;
 	struct bkey_s_c k;
@@ -518,7 +512,10 @@ int bch2_empty_dir_snapshot(struct btree_trans *trans, u64 dir, u32 snapshot)
 			   SPOS(dir, 0, snapshot),
 			   POS(dir, U64_MAX), 0, k, ret)
 		if (k.k->type == KEY_TYPE_dirent) {
-			ret = -ENOTEMPTY;
+			struct bkey_s_c_dirent d = bkey_s_c_to_dirent(k);
+			if (d.v->d_type == DT_SUBVOL && le32_to_cpu(d.v->d_parent_subvol) != subvol)
+				continue;
+			ret = -BCH_ERR_ENOTEMPTY_dir_not_empty;
 			break;
 		}
 	bch2_trans_iter_exit(trans, &iter);
@@ -531,7 +528,19 @@ int bch2_empty_dir_trans(struct btree_trans *trans, subvol_inum dir)
 	u32 snapshot;
 
 	return bch2_subvolume_get_snapshot(trans, dir.subvol, &snapshot) ?:
-		bch2_empty_dir_snapshot(trans, dir.inum, snapshot);
+		bch2_empty_dir_snapshot(trans, dir.inum, dir.subvol, snapshot);
+}
+
+static int bch2_dir_emit(struct dir_context *ctx, struct bkey_s_c_dirent d, subvol_inum target)
+{
+	struct qstr name = bch2_dirent_get_name(d);
+	bool ret = dir_emit(ctx, name.name,
+		      name.len,
+		      target.inum,
+		      vfs_d_type(d.v->d_type));
+	if (ret)
+		ctx->pos = d.k->p.offset + 1;
+	return ret;
 }
 
 int bch2_readdir(struct bch_fs *c, subvol_inum inum, struct dir_context *ctx)
@@ -539,11 +548,9 @@ int bch2_readdir(struct bch_fs *c, subvol_inum inum, struct dir_context *ctx)
 	struct btree_trans *trans = bch2_trans_get(c);
 	struct btree_iter iter;
 	struct bkey_s_c k;
-	struct bkey_s_c_dirent dirent;
 	subvol_inum target;
 	u32 snapshot;
 	struct bkey_buf sk;
-	struct qstr name;
 	int ret;
 
 	bch2_bkey_buf_init(&sk);
@@ -560,7 +567,9 @@ retry:
 		if (k.k->type != KEY_TYPE_dirent)
 			continue;
 
-		dirent = bkey_s_c_to_dirent(k);
+		/* dir_emit() can fault and block: */
+		bch2_bkey_buf_reassemble(&sk, c, k);
+		struct bkey_s_c_dirent dirent = bkey_i_to_s_c_dirent(sk.k);
 
 		ret = bch2_dirent_read_target(trans, inum, dirent, &target);
 		if (ret < 0)
@@ -568,28 +577,22 @@ retry:
 		if (ret)
 			continue;
 
-		/* dir_emit() can fault and block: */
-		bch2_bkey_buf_reassemble(&sk, c, k);
-		dirent = bkey_i_to_s_c_dirent(sk.k);
-		bch2_trans_unlock(trans);
-
-		name = bch2_dirent_get_name(dirent);
-
-		ctx->pos = dirent.k->p.offset;
-		if (!dir_emit(ctx, name.name,
-			      name.len,
-			      target.inum,
-			      vfs_d_type(dirent.v->d_type)))
-			break;
-		ctx->pos = dirent.k->p.offset + 1;
-
 		/*
 		 * read_target looks up subvolumes, we can overflow paths if the
 		 * directory has many subvolumes in it
+		 *
+		 * XXX: btree_trans_too_many_iters() is something we'd like to
+		 * get rid of, and there's no good reason to be using it here
+		 * except that we don't yet have a for_each_btree_key() helper
+		 * that does subvolume_get_snapshot().
 		 */
-		ret = btree_trans_too_many_iters(trans);
-		if (ret)
+		ret =   drop_locks_do(trans,
+				bch2_dir_emit(ctx, dirent, target)) ?:
+			btree_trans_too_many_iters(trans);
+		if (ret) {
+			ret = ret < 0 ? ret : 0;
 			break;
+		}
 	}
 	bch2_trans_iter_exit(trans, &iter);
 err:
diff --git a/fs/bcachefs/dirent.h b/fs/bcachefs/dirent.h
index 21ffeb78f0..24037e6e0a 100644
--- a/fs/bcachefs/dirent.h
+++ b/fs/bcachefs/dirent.h
@@ -4,11 +4,11 @@
 
 #include "str_hash.h"
 
-enum bkey_invalid_flags;
+enum bch_validate_flags;
 extern const struct bch_hash_desc bch2_dirent_hash_desc;
 
 int bch2_dirent_invalid(struct bch_fs *, struct bkey_s_c,
-			enum bkey_invalid_flags, struct printbuf *);
+			enum bch_validate_flags, struct printbuf *);
 void bch2_dirent_to_text(struct printbuf *, struct bch_fs *, struct bkey_s_c);
 
 #define bch2_bkey_ops_dirent ((struct bkey_ops) {	\
@@ -35,14 +35,14 @@ static inline unsigned dirent_val_u64s(unsigned len)
 int bch2_dirent_read_target(struct btree_trans *, subvol_inum,
 			    struct bkey_s_c_dirent, subvol_inum *);
 
-int bch2_dirent_create_snapshot(struct btree_trans *, u64, u32,
+int bch2_dirent_create_snapshot(struct btree_trans *, u32, u64, u32,
 			const struct bch_hash_info *, u8,
 			const struct qstr *, u64, u64 *,
-			bch_str_hash_flags_t);
+			enum btree_iter_update_trigger_flags);
 int bch2_dirent_create(struct btree_trans *, subvol_inum,
 		       const struct bch_hash_info *, u8,
 		       const struct qstr *, u64, u64 *,
-		       bch_str_hash_flags_t);
+		       enum btree_iter_update_trigger_flags);
 
 static inline unsigned vfs_d_type(unsigned type)
 {
@@ -62,14 +62,14 @@ int bch2_dirent_rename(struct btree_trans *,
 		       const struct qstr *, subvol_inum *, u64 *,
 		       enum bch_rename_mode);
 
-int __bch2_dirent_lookup_trans(struct btree_trans *, struct btree_iter *,
+int bch2_dirent_lookup_trans(struct btree_trans *, struct btree_iter *,
 			       subvol_inum, const struct bch_hash_info *,
 			       const struct qstr *, subvol_inum *, unsigned);
 u64 bch2_dirent_lookup(struct bch_fs *, subvol_inum,
 		       const struct bch_hash_info *,
 		       const struct qstr *, subvol_inum *);
 
-int bch2_empty_dir_snapshot(struct btree_trans *, u64, u32);
+int bch2_empty_dir_snapshot(struct btree_trans *, u64, u32, u32);
 int bch2_empty_dir_trans(struct btree_trans *, subvol_inum);
 int bch2_readdir(struct bch_fs *, subvol_inum, struct dir_context *);
 
diff --git a/fs/bcachefs/disk_groups.c b/fs/bcachefs/disk_groups.c
index 06a7df529b..521a86df5e 100644
--- a/fs/bcachefs/disk_groups.c
+++ b/fs/bcachefs/disk_groups.c
@@ -18,9 +18,8 @@ static int group_cmp(const void *_l, const void *_r)
 		strncmp(l->label, r->label, sizeof(l->label));
 }
 
-static int bch2_sb_disk_groups_validate(struct bch_sb *sb,
-					struct bch_sb_field *f,
-					struct printbuf *err)
+static int bch2_sb_disk_groups_validate(struct bch_sb *sb, struct bch_sb_field *f,
+				enum bch_validate_flags flags, struct printbuf *err)
 {
 	struct bch_sb_field_disk_groups *groups =
 		field_to_type(f, disk_groups);
@@ -177,7 +176,7 @@ int bch2_sb_disk_groups_to_cpu(struct bch_fs *c)
 		struct bch_member m = bch2_sb_member_get(c->disk_sb.sb, i);
 		struct bch_disk_group_cpu *dst;
 
-		if (!bch2_member_exists(&m))
+		if (!bch2_member_alive(&m))
 			continue;
 
 		g = BCH_MEMBER_GROUP(&m);
@@ -523,7 +522,7 @@ int bch2_opt_target_parse(struct bch_fs *c, const char *val, u64 *res,
 	ca = bch2_dev_lookup(c, val);
 	if (!IS_ERR(ca)) {
 		*res = dev_to_target(ca->dev_idx);
-		percpu_ref_put(&ca->ref);
+		bch2_dev_put(ca);
 		return 0;
 	}
 
@@ -588,7 +587,7 @@ static void bch2_target_to_text_sb(struct printbuf *out, struct bch_sb *sb, unsi
 	case TARGET_DEV: {
 		struct bch_member m = bch2_sb_member_get(sb, t.dev);
 
-		if (bch2_dev_exists(sb, t.dev)) {
+		if (bch2_member_exists(sb, t.dev)) {
 			prt_printf(out, "Device ");
 			pr_uuid(out, m.uuid.b);
 			prt_printf(out, " (%u)", t.dev);
diff --git a/fs/bcachefs/disk_groups_format.h b/fs/bcachefs/disk_groups_format.h
new file mode 100644
index 0000000000..698990bbf1
--- /dev/null
+++ b/fs/bcachefs/disk_groups_format.h
@@ -0,0 +1,21 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _BCACHEFS_DISK_GROUPS_FORMAT_H
+#define _BCACHEFS_DISK_GROUPS_FORMAT_H
+
+#define BCH_SB_LABEL_SIZE		32
+
+struct bch_disk_group {
+	__u8			label[BCH_SB_LABEL_SIZE];
+	__le64			flags[2];
+} __packed __aligned(8);
+
+LE64_BITMASK(BCH_GROUP_DELETED,		struct bch_disk_group, flags[0], 0,  1)
+LE64_BITMASK(BCH_GROUP_DATA_ALLOWED,	struct bch_disk_group, flags[0], 1,  6)
+LE64_BITMASK(BCH_GROUP_PARENT,		struct bch_disk_group, flags[0], 6, 24)
+
+struct bch_sb_field_disk_groups {
+	struct bch_sb_field	field;
+	struct bch_disk_group	entries[];
+} __packed __aligned(8);
+
+#endif /* _BCACHEFS_DISK_GROUPS_FORMAT_H */
diff --git a/fs/bcachefs/ec.c b/fs/bcachefs/ec.c
index d503af2700..83e279d418 100644
--- a/fs/bcachefs/ec.c
+++ b/fs/bcachefs/ec.c
@@ -107,7 +107,7 @@ struct ec_bio {
 /* Stripes btree keys: */
 
 int bch2_stripe_invalid(struct bch_fs *c, struct bkey_s_c k,
-			enum bkey_invalid_flags flags,
+			enum bch_validate_flags flags,
 			struct printbuf *err)
 {
 	const struct bch_stripe *s = bkey_s_c_to_stripe(k).v;
@@ -131,174 +131,231 @@ fsck_err:
 void bch2_stripe_to_text(struct printbuf *out, struct bch_fs *c,
 			 struct bkey_s_c k)
 {
-	const struct bch_stripe *s = bkey_s_c_to_stripe(k).v;
-	unsigned i, nr_data = s->nr_blocks - s->nr_redundant;
+	const struct bch_stripe *sp = bkey_s_c_to_stripe(k).v;
+	struct bch_stripe s = {};
+
+	memcpy(&s, sp, min(sizeof(s), bkey_val_bytes(k.k)));
+
+	unsigned nr_data = s.nr_blocks - s.nr_redundant;
 
-	prt_printf(out, "algo %u sectors %u blocks %u:%u csum %u gran %u",
-	       s->algorithm,
-	       le16_to_cpu(s->sectors),
-	       nr_data,
-	       s->nr_redundant,
-	       s->csum_type,
-	       1U << s->csum_granularity_bits);
+	prt_printf(out, "algo %u sectors %u blocks %u:%u csum ",
+		   s.algorithm,
+		   le16_to_cpu(s.sectors),
+		   nr_data,
+		   s.nr_redundant);
+	bch2_prt_csum_type(out, s.csum_type);
+	prt_printf(out, " gran %u", 1U << s.csum_granularity_bits);
+
+	for (unsigned i = 0; i < s.nr_blocks; i++) {
+		const struct bch_extent_ptr *ptr = sp->ptrs + i;
+
+		if ((void *) ptr >= bkey_val_end(k))
+			break;
 
-	for (i = 0; i < s->nr_blocks; i++) {
-		const struct bch_extent_ptr *ptr = s->ptrs + i;
-		struct bch_dev *ca = bch_dev_bkey_exists(c, ptr->dev);
-		u32 offset;
-		u64 b = sector_to_bucket_and_offset(ca, ptr->offset, &offset);
+		bch2_extent_ptr_to_text(out, c, ptr);
 
-		prt_printf(out, " %u:%llu:%u", ptr->dev, b, offset);
-		if (i < nr_data)
-			prt_printf(out, "#%u", stripe_blockcount_get(s, i));
-		prt_printf(out, " gen %u", ptr->gen);
-		if (ptr_stale(ca, ptr))
-			prt_printf(out, " stale");
+		if (s.csum_type < BCH_CSUM_NR &&
+		    i < nr_data &&
+		    stripe_blockcount_offset(&s, i) < bkey_val_bytes(k.k))
+			prt_printf(out,  "#%u", stripe_blockcount_get(sp, i));
 	}
 }
 
 /* Triggers: */
 
-static int bch2_trans_mark_stripe_bucket(struct btree_trans *trans,
-					 struct bkey_s_c_stripe s,
-					 unsigned idx, bool deleting)
+static int __mark_stripe_bucket(struct btree_trans *trans,
+				struct bch_dev *ca,
+				struct bkey_s_c_stripe s,
+				unsigned ptr_idx, bool deleting,
+				struct bpos bucket,
+				struct bch_alloc_v4 *a,
+				enum btree_iter_update_trigger_flags flags)
 {
-	struct bch_fs *c = trans->c;
-	const struct bch_extent_ptr *ptr = &s.v->ptrs[idx];
-	struct btree_iter iter;
-	struct bkey_i_alloc_v4 *a;
-	enum bch_data_type data_type = idx >= s.v->nr_blocks - s.v->nr_redundant
-		? BCH_DATA_parity : 0;
-	s64 sectors = data_type ? le16_to_cpu(s.v->sectors) : 0;
+	const struct bch_extent_ptr *ptr = s.v->ptrs + ptr_idx;
+	unsigned nr_data = s.v->nr_blocks - s.v->nr_redundant;
+	bool parity = ptr_idx >= nr_data;
+	enum bch_data_type data_type = parity ? BCH_DATA_parity : BCH_DATA_stripe;
+	s64 sectors = parity ? le16_to_cpu(s.v->sectors) : 0;
+	struct printbuf buf = PRINTBUF;
 	int ret = 0;
 
+	struct bch_fs *c = trans->c;
 	if (deleting)
 		sectors = -sectors;
 
-	a = bch2_trans_start_alloc_update(trans, &iter, PTR_BUCKET_POS(c, ptr));
-	if (IS_ERR(a))
-		return PTR_ERR(a);
-
-	ret = bch2_check_bucket_ref(trans, s.s_c, ptr, sectors, data_type,
-				    a->v.gen, a->v.data_type,
-				    a->v.dirty_sectors);
-	if (ret)
-		goto err;
-
 	if (!deleting) {
-		if (bch2_trans_inconsistent_on(a->v.stripe ||
-					       a->v.stripe_redundancy, trans,
-				"bucket %llu:%llu gen %u data type %s dirty_sectors %u: multiple stripes using same bucket (%u, %llu)",
-				iter.pos.inode, iter.pos.offset, a->v.gen,
-				bch2_data_type_str(a->v.data_type),
-				a->v.dirty_sectors,
-				a->v.stripe, s.k->p.offset)) {
+		if (bch2_trans_inconsistent_on(a->stripe ||
+					       a->stripe_redundancy, trans,
+				"bucket %llu:%llu gen %u data type %s dirty_sectors %u: multiple stripes using same bucket (%u, %llu)\n%s",
+				bucket.inode, bucket.offset, a->gen,
+				bch2_data_type_str(a->data_type),
+				a->dirty_sectors,
+				a->stripe, s.k->p.offset,
+				(bch2_bkey_val_to_text(&buf, c, s.s_c), buf.buf))) {
 			ret = -EIO;
 			goto err;
 		}
 
-		if (bch2_trans_inconsistent_on(data_type && a->v.dirty_sectors, trans,
-				"bucket %llu:%llu gen %u data type %s dirty_sectors %u: data already in stripe bucket %llu",
-				iter.pos.inode, iter.pos.offset, a->v.gen,
-				bch2_data_type_str(a->v.data_type),
-				a->v.dirty_sectors,
-				s.k->p.offset)) {
+		if (bch2_trans_inconsistent_on(parity && bch2_bucket_sectors_total(*a), trans,
+				"bucket %llu:%llu gen %u data type %s dirty_sectors %u cached_sectors %u: data already in parity bucket\n%s",
+				bucket.inode, bucket.offset, a->gen,
+				bch2_data_type_str(a->data_type),
+				a->dirty_sectors,
+				a->cached_sectors,
+				(bch2_bkey_val_to_text(&buf, c, s.s_c), buf.buf))) {
 			ret = -EIO;
 			goto err;
 		}
-
-		a->v.stripe		= s.k->p.offset;
-		a->v.stripe_redundancy	= s.v->nr_redundant;
-		a->v.data_type		= BCH_DATA_stripe;
 	} else {
-		if (bch2_trans_inconsistent_on(a->v.stripe != s.k->p.offset ||
-					       a->v.stripe_redundancy != s.v->nr_redundant, trans,
-				"bucket %llu:%llu gen %u: not marked as stripe when deleting stripe %llu (got %u)",
-				iter.pos.inode, iter.pos.offset, a->v.gen,
-				s.k->p.offset, a->v.stripe)) {
+		if (bch2_trans_inconsistent_on(a->stripe != s.k->p.offset ||
+					       a->stripe_redundancy != s.v->nr_redundant, trans,
+				"bucket %llu:%llu gen %u: not marked as stripe when deleting stripe (got %u)\n%s",
+				bucket.inode, bucket.offset, a->gen,
+				a->stripe,
+				(bch2_bkey_val_to_text(&buf, c, s.s_c), buf.buf))) {
 			ret = -EIO;
 			goto err;
 		}
 
-		a->v.stripe		= 0;
-		a->v.stripe_redundancy	= 0;
-		a->v.data_type		= alloc_data_type(a->v, BCH_DATA_user);
+		if (bch2_trans_inconsistent_on(a->data_type != data_type, trans,
+				"bucket %llu:%llu gen %u data type %s: wrong data type when stripe, should be %s\n%s",
+				bucket.inode, bucket.offset, a->gen,
+				bch2_data_type_str(a->data_type),
+				bch2_data_type_str(data_type),
+				(bch2_bkey_val_to_text(&buf, c, s.s_c), buf.buf))) {
+			ret = -EIO;
+			goto err;
+		}
+
+		if (bch2_trans_inconsistent_on(parity &&
+					       (a->dirty_sectors != -sectors ||
+						a->cached_sectors), trans,
+				"bucket %llu:%llu gen %u dirty_sectors %u cached_sectors %u: wrong sectors when deleting parity block of stripe\n%s",
+				bucket.inode, bucket.offset, a->gen,
+				a->dirty_sectors,
+				a->cached_sectors,
+				(bch2_bkey_val_to_text(&buf, c, s.s_c), buf.buf))) {
+			ret = -EIO;
+			goto err;
+		}
 	}
 
-	a->v.dirty_sectors += sectors;
-	if (data_type)
-		a->v.data_type = !deleting ? data_type : 0;
+	if (sectors) {
+		ret = bch2_bucket_ref_update(trans, ca, s.s_c, ptr, sectors, data_type,
+					     a->gen, a->data_type, &a->dirty_sectors);
+		if (ret)
+			goto err;
+	}
 
-	ret = bch2_trans_update(trans, &iter, &a->k_i, 0);
-	if (ret)
-		goto err;
+	if (!deleting) {
+		a->stripe		= s.k->p.offset;
+		a->stripe_redundancy	= s.v->nr_redundant;
+	} else {
+		a->stripe		= 0;
+		a->stripe_redundancy	= 0;
+	}
+
+	alloc_data_type_set(a, data_type);
 err:
-	bch2_trans_iter_exit(trans, &iter);
+	printbuf_exit(&buf);
 	return ret;
 }
 
 static int mark_stripe_bucket(struct btree_trans *trans,
-			      struct bkey_s_c k,
-			      unsigned ptr_idx,
-			      unsigned flags)
+			      struct bkey_s_c_stripe s,
+			      unsigned ptr_idx, bool deleting,
+			      enum btree_iter_update_trigger_flags flags)
 {
 	struct bch_fs *c = trans->c;
-	const struct bch_stripe *s = bkey_s_c_to_stripe(k).v;
-	unsigned nr_data = s->nr_blocks - s->nr_redundant;
-	bool parity = ptr_idx >= nr_data;
-	enum bch_data_type data_type = parity ? BCH_DATA_parity : BCH_DATA_stripe;
-	s64 sectors = parity ? le16_to_cpu(s->sectors) : 0;
-	const struct bch_extent_ptr *ptr = s->ptrs + ptr_idx;
-	struct bch_dev *ca = bch_dev_bkey_exists(c, ptr->dev);
-	struct bucket old, new, *g;
+	const struct bch_extent_ptr *ptr = s.v->ptrs + ptr_idx;
 	struct printbuf buf = PRINTBUF;
 	int ret = 0;
 
-	BUG_ON(!(flags & BTREE_TRIGGER_GC));
-
-	/* * XXX doesn't handle deletion */
-
-	percpu_down_read(&c->mark_lock);
-	g = PTR_GC_BUCKET(ca, ptr);
-
-	if (g->dirty_sectors ||
-	    (g->stripe && g->stripe != k.k->p.offset)) {
-		bch2_fs_inconsistent(c,
-			      "bucket %u:%zu gen %u: multiple stripes using same bucket\n%s",
-			      ptr->dev, PTR_BUCKET_NR(ca, ptr), g->gen,
-			      (bch2_bkey_val_to_text(&buf, c, k), buf.buf));
-		ret = -EINVAL;
+	struct bch_dev *ca = bch2_dev_tryget(c, ptr->dev);
+	if (unlikely(!ca)) {
+		if (!(flags & BTREE_TRIGGER_overwrite))
+			ret = -EIO;
 		goto err;
 	}
 
-	bucket_lock(g);
-	old = *g;
+	struct bpos bucket = PTR_BUCKET_POS(ca, ptr);
 
-	ret = bch2_check_bucket_ref(trans, k, ptr, sectors, data_type,
-				    g->gen, g->data_type,
-				    g->dirty_sectors);
-	if (ret)
-		goto err;
+	if (flags & BTREE_TRIGGER_transactional) {
+		struct bkey_i_alloc_v4 *a =
+			bch2_trans_start_alloc_update(trans, bucket);
+		ret = PTR_ERR_OR_ZERO(a) ?:
+			__mark_stripe_bucket(trans, ca, s, ptr_idx, deleting, bucket, &a->v, flags);
+	}
 
-	g->data_type = data_type;
-	g->dirty_sectors += sectors;
+	if (flags & BTREE_TRIGGER_gc) {
+		percpu_down_read(&c->mark_lock);
+		struct bucket *g = gc_bucket(ca, bucket.offset);
+		if (bch2_fs_inconsistent_on(!g, c, "reference to invalid bucket on device %u\n  %s",
+					    ptr->dev,
+					    (bch2_bkey_val_to_text(&buf, c, s.s_c), buf.buf))) {
+			ret = -EIO;
+			goto err_unlock;
+		}
 
-	g->stripe		= k.k->p.offset;
-	g->stripe_redundancy	= s->nr_redundant;
-	new = *g;
+		bucket_lock(g);
+		struct bch_alloc_v4 old = bucket_m_to_alloc(*g), new = old;
+		ret = __mark_stripe_bucket(trans, ca, s, ptr_idx, deleting, bucket, &new, flags);
+		if (!ret) {
+			alloc_to_bucket(g, new);
+			bch2_dev_usage_update(c, ca, &old, &new, 0, true);
+		}
+		bucket_unlock(g);
+err_unlock:
+		percpu_up_read(&c->mark_lock);
+	}
 err:
-	bucket_unlock(g);
-	if (!ret)
-		bch2_dev_usage_update_m(c, ca, &old, &new);
-	percpu_up_read(&c->mark_lock);
+	bch2_dev_put(ca);
 	printbuf_exit(&buf);
 	return ret;
 }
 
+static int mark_stripe_buckets(struct btree_trans *trans,
+			       struct bkey_s_c old, struct bkey_s_c new,
+			       enum btree_iter_update_trigger_flags flags)
+{
+	const struct bch_stripe *old_s = old.k->type == KEY_TYPE_stripe
+		? bkey_s_c_to_stripe(old).v : NULL;
+	const struct bch_stripe *new_s = new.k->type == KEY_TYPE_stripe
+		? bkey_s_c_to_stripe(new).v : NULL;
+
+	BUG_ON(old_s && new_s && old_s->nr_blocks != new_s->nr_blocks);
+
+	unsigned nr_blocks = new_s ? new_s->nr_blocks : old_s->nr_blocks;
+
+	for (unsigned i = 0; i < nr_blocks; i++) {
+		if (new_s && old_s &&
+		    !memcmp(&new_s->ptrs[i],
+			    &old_s->ptrs[i],
+			    sizeof(new_s->ptrs[i])))
+			continue;
+
+		if (new_s) {
+			int ret = mark_stripe_bucket(trans,
+					bkey_s_c_to_stripe(new), i, false, flags);
+			if (ret)
+				return ret;
+		}
+
+		if (old_s) {
+			int ret = mark_stripe_bucket(trans,
+					bkey_s_c_to_stripe(old), i, true, flags);
+			if (ret)
+				return ret;
+		}
+	}
+
+	return 0;
+}
+
 int bch2_trigger_stripe(struct btree_trans *trans,
-			enum btree_id btree_id, unsigned level,
+			enum btree_id btree, unsigned level,
 			struct bkey_s_c old, struct bkey_s _new,
-			unsigned flags)
+			enum btree_iter_update_trigger_flags flags)
 {
 	struct bkey_s_c new = _new.s_c;
 	struct bch_fs *c = trans->c;
@@ -308,7 +365,10 @@ int bch2_trigger_stripe(struct btree_trans *trans,
 	const struct bch_stripe *new_s = new.k->type == KEY_TYPE_stripe
 		? bkey_s_c_to_stripe(new).v : NULL;
 
-	if (flags & BTREE_TRIGGER_TRANSACTIONAL) {
+	if (unlikely(flags & BTREE_TRIGGER_check_repair))
+		return bch2_check_fix_ptrs(trans, btree, level, _new.s_c, flags);
+
+	if (flags & BTREE_TRIGGER_transactional) {
 		/*
 		 * If the pointers aren't changing, we don't need to do anything:
 		 */
@@ -343,31 +403,12 @@ int bch2_trigger_stripe(struct btree_trans *trans,
 				return ret;
 		}
 
-		unsigned nr_blocks = new_s ? new_s->nr_blocks : old_s->nr_blocks;
-		for (unsigned i = 0; i < nr_blocks; i++) {
-			if (new_s && old_s &&
-			    !memcmp(&new_s->ptrs[i],
-				    &old_s->ptrs[i],
-				    sizeof(new_s->ptrs[i])))
-				continue;
-
-			if (new_s) {
-				int ret = bch2_trans_mark_stripe_bucket(trans,
-						bkey_s_c_to_stripe(new), i, false);
-				if (ret)
-					return ret;
-			}
-
-			if (old_s) {
-				int ret = bch2_trans_mark_stripe_bucket(trans,
-						bkey_s_c_to_stripe(old), i, true);
-				if (ret)
-					return ret;
-			}
-		}
+		int ret = mark_stripe_buckets(trans, old, new, flags);
+		if (ret)
+			return ret;
 	}
 
-	if (flags & BTREE_TRIGGER_ATOMIC) {
+	if (flags & BTREE_TRIGGER_atomic) {
 		struct stripe *m = genradix_ptr(&c->stripes, idx);
 
 		if (!m) {
@@ -406,7 +447,7 @@ int bch2_trigger_stripe(struct btree_trans *trans,
 		}
 	}
 
-	if (flags & BTREE_TRIGGER_GC) {
+	if (flags & BTREE_TRIGGER_gc) {
 		struct gc_stripe *m =
 			genradix_ptr_alloc(&c->gc_stripes, idx, GFP_KERNEL);
 
@@ -435,20 +476,18 @@ int bch2_trigger_stripe(struct btree_trans *trans,
 		 */
 		memset(m->block_sectors, 0, sizeof(m->block_sectors));
 
-		for (unsigned i = 0; i < new_s->nr_blocks; i++) {
-			int ret = mark_stripe_bucket(trans, new, i, flags);
-			if (ret)
-				return ret;
-		}
+		int ret = mark_stripe_buckets(trans, old, new, flags);
+		if (ret)
+			return ret;
 
-		int ret = bch2_update_replicas(c, new, &m->r.e,
+		ret = bch2_update_replicas(c, new, &m->r.e,
 				      ((s64) m->sectors * m->nr_redundant),
 				      0, true);
 		if (ret) {
 			struct printbuf buf = PRINTBUF;
 
 			bch2_bkey_val_to_text(&buf, c, new);
-			bch2_fs_fatal_error(c, "no replicas entry for %s", buf.buf);
+			bch2_fs_fatal_error(c, ": no replicas entry for %s", buf.buf);
 			printbuf_exit(&buf);
 			return ret;
 		}
@@ -504,7 +543,7 @@ static void ec_stripe_buf_exit(struct ec_stripe_buf *buf)
 		unsigned i;
 
 		for (i = 0; i < s->v.nr_blocks; i++) {
-			kvpfree(buf->data[i], buf->size << 9);
+			kvfree(buf->data[i]);
 			buf->data[i] = NULL;
 		}
 	}
@@ -531,7 +570,7 @@ static int ec_stripe_buf_init(struct ec_stripe_buf *buf,
 	memset(buf->valid, 0xFF, sizeof(buf->valid));
 
 	for (i = 0; i < v->nr_blocks; i++) {
-		buf->data[i] = kvpmalloc(buf->size << 9, GFP_KERNEL);
+		buf->data[i] = kvmalloc(buf->size << 9, GFP_KERNEL);
 		if (!buf->data[i])
 			goto err;
 	}
@@ -604,21 +643,21 @@ static void ec_validate_checksums(struct bch_fs *c, struct ec_stripe_buf *buf)
 			struct bch_csum got = ec_block_checksum(buf, i, offset);
 
 			if (bch2_crc_cmp(want, got)) {
-				struct printbuf err = PRINTBUF;
-				struct bch_dev *ca = bch_dev_bkey_exists(c, v->ptrs[i].dev);
-
-				prt_printf(&err, "stripe checksum error: expected %0llx:%0llx got %0llx:%0llx (type %s)\n",
-					   want.hi, want.lo,
-					   got.hi, got.lo,
-					   bch2_csum_types[v->csum_type]);
-				prt_printf(&err, "  for %ps at %u of\n  ", (void *) _RET_IP_, i);
-				bch2_bkey_val_to_text(&err, c, bkey_i_to_s_c(&buf->key));
-				bch_err_ratelimited(ca, "%s", err.buf);
-				printbuf_exit(&err);
+				struct bch_dev *ca = bch2_dev_tryget(c, v->ptrs[i].dev);
+				if (ca) {
+					struct printbuf err = PRINTBUF;
 
-				clear_bit(i, buf->valid);
+					prt_str(&err, "stripe ");
+					bch2_csum_err_msg(&err, v->csum_type, want, got);
+					prt_printf(&err, "  for %ps at %u of\n  ", (void *) _RET_IP_, i);
+					bch2_bkey_val_to_text(&err, c, bkey_i_to_s_c(&buf->key));
+					bch_err_ratelimited(ca, "%s", err.buf);
+					printbuf_exit(&err);
 
-				bch2_io_error(ca, BCH_MEMBER_ERROR_checksum);
+					bch2_io_error(ca, BCH_MEMBER_ERROR_checksum);
+				}
+
+				clear_bit(i, buf->valid);
 				break;
 			}
 
@@ -685,10 +724,12 @@ static void ec_block_endio(struct bio *bio)
 			       bch2_blk_status_to_str(bio->bi_status)))
 		clear_bit(ec_bio->idx, ec_bio->buf->valid);
 
-	if (ptr_stale(ca, ptr)) {
+	int stale = dev_ptr_stale(ca, ptr);
+	if (stale) {
 		bch_err_ratelimited(ca->fs,
-				    "error %s stripe: stale pointer after io",
-				    bio_data_dir(bio) == READ ? "reading from" : "writing to");
+				    "error %s stripe: stale/invalid pointer (%i) after io",
+				    bio_data_dir(bio) == READ ? "reading from" : "writing to",
+				    stale);
 		clear_bit(ec_bio->idx, ec_bio->buf->valid);
 	}
 
@@ -703,25 +744,28 @@ static void ec_block_io(struct bch_fs *c, struct ec_stripe_buf *buf,
 	struct bch_stripe *v = &bkey_i_to_stripe(&buf->key)->v;
 	unsigned offset = 0, bytes = buf->size << 9;
 	struct bch_extent_ptr *ptr = &v->ptrs[idx];
-	struct bch_dev *ca = bch_dev_bkey_exists(c, ptr->dev);
 	enum bch_data_type data_type = idx < v->nr_blocks - v->nr_redundant
 		? BCH_DATA_user
 		: BCH_DATA_parity;
 	int rw = op_is_write(opf);
 
-	if (ptr_stale(ca, ptr)) {
-		bch_err_ratelimited(c,
-				    "error %s stripe: stale pointer",
-				    rw == READ ? "reading from" : "writing to");
+	struct bch_dev *ca = bch2_dev_get_ioref(c, ptr->dev, rw);
+	if (!ca) {
 		clear_bit(idx, buf->valid);
 		return;
 	}
 
-	if (!bch2_dev_get_ioref(ca, rw)) {
+	int stale = dev_ptr_stale(ca, ptr);
+	if (stale) {
+		bch_err_ratelimited(c,
+				    "error %s stripe: stale pointer (%i)",
+				    rw == READ ? "reading from" : "writing to",
+				    stale);
 		clear_bit(idx, buf->valid);
 		return;
 	}
 
+
 	this_cpu_add(ca->io_done->sectors[rw][data_type], buf->size);
 
 	while (offset < bytes) {
@@ -767,7 +811,7 @@ static int get_stripe_key_trans(struct btree_trans *trans, u64 idx,
 	int ret;
 
 	k = bch2_bkey_get_iter(trans, &iter, BTREE_ID_stripes,
-			       POS(0, idx), BTREE_ITER_SLOTS);
+			       POS(0, idx), BTREE_ITER_slots);
 	ret = bkey_err(k);
 	if (ret)
 		goto err;
@@ -878,7 +922,7 @@ static int __ec_stripe_mem_alloc(struct bch_fs *c, size_t idx, gfp_t gfp)
 	if (!genradix_ptr_alloc(&c->stripes, idx, gfp))
 		return -BCH_ERR_ENOMEM_ec_stripe_mem_alloc;
 
-	if (c->gc_pos.phase != GC_PHASE_NOT_RUNNING &&
+	if (c->gc_pos.phase != GC_PHASE_not_running &&
 	    !genradix_ptr_alloc(&c->gc_stripes, idx, gfp))
 		return -BCH_ERR_ENOMEM_ec_stripe_mem_alloc;
 
@@ -1058,7 +1102,7 @@ static int ec_stripe_delete(struct btree_trans *trans, u64 idx)
 	int ret;
 
 	k = bch2_bkey_get_iter(trans, &iter, BTREE_ID_stripes, POS(0, idx),
-			       BTREE_ITER_INTENT);
+			       BTREE_ITER_intent);
 	ret = bkey_err(k);
 	if (ret)
 		goto err;
@@ -1129,7 +1173,7 @@ static int ec_stripe_key_update(struct btree_trans *trans,
 	int ret;
 
 	k = bch2_bkey_get_iter(trans, &iter, BTREE_ID_stripes,
-			       new->k.p, BTREE_ITER_INTENT);
+			       new->k.p, BTREE_ITER_intent);
 	ret = bkey_err(k);
 	if (ret)
 		goto err;
@@ -1171,6 +1215,7 @@ err:
 }
 
 static int ec_stripe_update_extent(struct btree_trans *trans,
+				   struct bch_dev *ca,
 				   struct bpos bucket, u8 gen,
 				   struct ec_stripe_buf *s,
 				   struct bpos *bp_pos)
@@ -1181,13 +1226,13 @@ static int ec_stripe_update_extent(struct btree_trans *trans,
 	struct btree_iter iter;
 	struct bkey_s_c k;
 	const struct bch_extent_ptr *ptr_c;
-	struct bch_extent_ptr *ptr, *ec_ptr = NULL;
+	struct bch_extent_ptr *ec_ptr = NULL;
 	struct bch_extent_stripe_ptr stripe_ptr;
 	struct bkey_i *n;
 	int ret, dev, block;
 
-	ret = bch2_get_next_backpointer(trans, bucket, gen,
-				bp_pos, &bp, BTREE_ITER_CACHED);
+	ret = bch2_get_next_backpointer(trans, ca, bucket, gen,
+				bp_pos, &bp, BTREE_ITER_cached);
 	if (ret)
 		return ret;
 	if (bpos_eq(*bp_pos, SPOS_MAX))
@@ -1212,7 +1257,7 @@ static int ec_stripe_update_extent(struct btree_trans *trans,
 		return -EIO;
 	}
 
-	k = bch2_backpointer_get_key(trans, &iter, *bp_pos, bp, BTREE_ITER_INTENT);
+	k = bch2_backpointer_get_key(trans, &iter, *bp_pos, bp, BTREE_ITER_intent);
 	ret = bkey_err(k);
 	if (ret)
 		return ret;
@@ -1270,17 +1315,21 @@ static int ec_stripe_update_bucket(struct btree_trans *trans, struct ec_stripe_b
 {
 	struct bch_fs *c = trans->c;
 	struct bch_stripe *v = &bkey_i_to_stripe(&s->key)->v;
-	struct bch_extent_ptr bucket = v->ptrs[block];
-	struct bpos bucket_pos = PTR_BUCKET_POS(c, &bucket);
+	struct bch_extent_ptr ptr = v->ptrs[block];
 	struct bpos bp_pos = POS_MIN;
 	int ret = 0;
 
+	struct bch_dev *ca = bch2_dev_tryget(c, ptr.dev);
+	if (!ca)
+		return -EIO;
+
+	struct bpos bucket_pos = PTR_BUCKET_POS(ca, &ptr);
+
 	while (1) {
 		ret = commit_do(trans, NULL, NULL,
 				BCH_TRANS_COMMIT_no_check_rw|
 				BCH_TRANS_COMMIT_no_enospc,
-			ec_stripe_update_extent(trans, bucket_pos, bucket.gen,
-						s, &bp_pos));
+			ec_stripe_update_extent(trans, ca, bucket_pos, ptr.gen, s, &bp_pos));
 		if (ret)
 			break;
 		if (bkey_eq(bp_pos, POS_MAX))
@@ -1289,6 +1338,7 @@ static int ec_stripe_update_bucket(struct btree_trans *trans, struct ec_stripe_b
 		bp_pos = bpos_nosnap_successor(bp_pos);
 	}
 
+	bch2_dev_put(ca);
 	return ret;
 }
 
@@ -1319,20 +1369,18 @@ static void zero_out_rest_of_ec_bucket(struct bch_fs *c,
 				       unsigned block,
 				       struct open_bucket *ob)
 {
-	struct bch_dev *ca = bch_dev_bkey_exists(c, ob->dev);
-	unsigned offset = ca->mi.bucket_size - ob->sectors_free;
-	int ret;
-
-	if (!bch2_dev_get_ioref(ca, WRITE)) {
+	struct bch_dev *ca = bch2_dev_get_ioref(c, ob->dev, WRITE);
+	if (!ca) {
 		s->err = -BCH_ERR_erofs_no_writes;
 		return;
 	}
 
+	unsigned offset = ca->mi.bucket_size - ob->sectors_free;
 	memset(s->new_stripe.data[block] + (offset << 9),
 	       0,
 	       ob->sectors_free << 9);
 
-	ret = blkdev_issue_zeroout(ca->disk_sb.bdev,
+	int ret = blkdev_issue_zeroout(ca->disk_sb.bdev,
 			ob->bucket * ca->mi.bucket_size + offset,
 			ob->sectors_free,
 			GFP_KERNEL, 0);
@@ -1517,16 +1565,13 @@ void bch2_ec_bucket_cancel(struct bch_fs *c, struct open_bucket *ob)
 void *bch2_writepoint_ec_buf(struct bch_fs *c, struct write_point *wp)
 {
 	struct open_bucket *ob = ec_open_bucket(c, &wp->ptrs);
-	struct bch_dev *ca;
-	unsigned offset;
-
 	if (!ob)
 		return NULL;
 
 	BUG_ON(!ob->ec->new_stripe.data[ob->ec_idx]);
 
-	ca	= bch_dev_bkey_exists(c, ob->dev);
-	offset	= ca->mi.bucket_size - ob->sectors_free;
+	struct bch_dev *ca	= ob_dev(c, ob);
+	unsigned offset		= ca->mi.bucket_size - ob->sectors_free;
 
 	return ob->ec->new_stripe.data[ob->ec_idx] + (offset << 9);
 }
@@ -1868,10 +1913,10 @@ static int __bch2_ec_stripe_head_reuse(struct btree_trans *trans, struct ec_stri
 		return -BCH_ERR_stripe_alloc_blocked;
 
 	ret = get_stripe_key_trans(trans, idx, &h->s->existing_stripe);
+	bch2_fs_fatal_err_on(ret && !bch2_err_matches(ret, BCH_ERR_transaction_restart), c,
+			     "reading stripe key: %s", bch2_err_str(ret));
 	if (ret) {
 		bch2_stripe_close(c, h->s);
-		if (!bch2_err_matches(ret, BCH_ERR_transaction_restart))
-			bch2_fs_fatal_error(c, "error reading stripe key: %s", bch2_err_str(ret));
 		return ret;
 	}
 
@@ -1935,7 +1980,7 @@ static int __bch2_ec_stripe_head_reserve(struct btree_trans *trans, struct ec_st
 	}
 
 	for_each_btree_key_norestart(trans, iter, BTREE_ID_stripes, start_pos,
-			   BTREE_ITER_SLOTS|BTREE_ITER_INTENT, k, ret) {
+			   BTREE_ITER_slots|BTREE_ITER_intent, k, ret) {
 		if (bkey_gt(k.k->p, POS(0, U32_MAX))) {
 			if (start_pos.offset) {
 				start_pos = min_pos;
@@ -2125,7 +2170,7 @@ int bch2_stripes_read(struct bch_fs *c)
 {
 	int ret = bch2_trans_run(c,
 		for_each_btree_key(trans, iter, BTREE_ID_stripes, POS_MIN,
-				   BTREE_ITER_PREFETCH, k, ({
+				   BTREE_ITER_prefetch, k, ({
 			if (k.k->type != KEY_TYPE_stripe)
 				continue;
 
diff --git a/fs/bcachefs/ec.h b/fs/bcachefs/ec.h
index f4369b02e8..84a23eeb62 100644
--- a/fs/bcachefs/ec.h
+++ b/fs/bcachefs/ec.h
@@ -6,14 +6,15 @@
 #include "buckets_types.h"
 #include "extents_types.h"
 
-enum bkey_invalid_flags;
+enum bch_validate_flags;
 
 int bch2_stripe_invalid(struct bch_fs *, struct bkey_s_c,
-			enum bkey_invalid_flags, struct printbuf *);
+			enum bch_validate_flags, struct printbuf *);
 void bch2_stripe_to_text(struct printbuf *, struct bch_fs *,
 			 struct bkey_s_c);
 int bch2_trigger_stripe(struct btree_trans *, enum btree_id, unsigned,
-			struct bkey_s_c, struct bkey_s, unsigned);
+			struct bkey_s_c, struct bkey_s,
+			enum btree_iter_update_trigger_flags);
 
 #define bch2_bkey_ops_stripe ((struct bkey_ops) {	\
 	.key_invalid	= bch2_stripe_invalid,		\
@@ -32,6 +33,8 @@ static inline unsigned stripe_csums_per_device(const struct bch_stripe *s)
 static inline unsigned stripe_csum_offset(const struct bch_stripe *s,
 					  unsigned dev, unsigned csum_idx)
 {
+	EBUG_ON(s->csum_type >= BCH_CSUM_NR);
+
 	unsigned csum_bytes = bch_crc_bytes[s->csum_type];
 
 	return sizeof(struct bch_stripe) +
diff --git a/fs/bcachefs/errcode.c b/fs/bcachefs/errcode.c
index d260ff9bbf..43557bebd0 100644
--- a/fs/bcachefs/errcode.c
+++ b/fs/bcachefs/errcode.c
@@ -2,6 +2,7 @@
 
 #include "bcachefs.h"
 #include "errcode.h"
+#include "trace.h"
 
 #include <linux/errname.h>
 
@@ -49,15 +50,17 @@ bool __bch2_err_matches(int err, int class)
 	return err == class;
 }
 
-int __bch2_err_class(int err)
+int __bch2_err_class(int bch_err)
 {
-	err = -err;
-	BUG_ON((unsigned) err >= BCH_ERR_MAX);
+	int std_err = -bch_err;
+	BUG_ON((unsigned) std_err >= BCH_ERR_MAX);
 
-	while (err >= BCH_ERR_START && bch2_errcode_parents[err - BCH_ERR_START])
-		err = bch2_errcode_parents[err - BCH_ERR_START];
+	while (std_err >= BCH_ERR_START && bch2_errcode_parents[std_err - BCH_ERR_START])
+		std_err = bch2_errcode_parents[std_err - BCH_ERR_START];
+
+	trace_error_downcast(bch_err, std_err, _RET_IP_);
 
-	return -err;
+	return -std_err;
 }
 
 const char *bch2_blk_status_to_str(blk_status_t status)
diff --git a/fs/bcachefs/errcode.h b/fs/bcachefs/errcode.h
index 8c40c2067a..58612abf79 100644
--- a/fs/bcachefs/errcode.h
+++ b/fs/bcachefs/errcode.h
@@ -5,6 +5,10 @@
 #define BCH_ERRCODES()								\
 	x(ERANGE,			ERANGE_option_too_small)		\
 	x(ERANGE,			ERANGE_option_too_big)			\
+	x(EINVAL,			mount_option)				\
+	x(BCH_ERR_mount_option,		option_name)				\
+	x(BCH_ERR_mount_option,		option_value)				\
+	x(BCH_ERR_mount_option,         option_not_bool)                        \
 	x(ENOMEM,			ENOMEM_stripe_buf)			\
 	x(ENOMEM,			ENOMEM_replicas_table)			\
 	x(ENOMEM,			ENOMEM_cpu_replicas)			\
@@ -78,6 +82,7 @@
 	x(ENOMEM,			ENOMEM_fs_name_alloc)			\
 	x(ENOMEM,			ENOMEM_fs_other_alloc)			\
 	x(ENOMEM,			ENOMEM_dev_alloc)			\
+	x(ENOMEM,			ENOMEM_disk_accounting)			\
 	x(ENOSPC,			ENOSPC_disk_reservation)		\
 	x(ENOSPC,			ENOSPC_bucket_alloc)			\
 	x(ENOSPC,			ENOSPC_disk_label_add)			\
@@ -109,6 +114,11 @@
 	x(ENOENT,			ENOENT_dirent_doesnt_match_inode)	\
 	x(ENOENT,			ENOENT_dev_not_found)			\
 	x(ENOENT,			ENOENT_dev_idx_not_found)		\
+	x(ENOTEMPTY,			ENOTEMPTY_dir_not_empty)		\
+	x(ENOTEMPTY,			ENOTEMPTY_subvol_not_empty)		\
+	x(EEXIST,			EEXIST_str_hash_set)			\
+	x(EEXIST,			EEXIST_discard_in_flight_add)		\
+	x(EEXIST,			EEXIST_subvolume_create)		\
 	x(0,				open_buckets_empty)			\
 	x(0,				freelist_empty)				\
 	x(BCH_ERR_freelist_empty,	no_buckets_found)			\
@@ -168,6 +178,7 @@
 	x(EINVAL,			block_size_too_small)			\
 	x(EINVAL,			bucket_size_too_small)			\
 	x(EINVAL,			device_size_too_small)			\
+	x(EINVAL,			device_size_too_big)			\
 	x(EINVAL,			device_not_a_member_of_filesystem)	\
 	x(EINVAL,			device_has_been_removed)		\
 	x(EINVAL,			device_splitbrain)			\
@@ -176,6 +187,9 @@
 	x(EINVAL,			invalid)				\
 	x(EINVAL,			internal_fsck_err)			\
 	x(EINVAL,			opt_parse_error)			\
+	x(EINVAL,			remove_with_metadata_missing_unimplemented)\
+	x(EINVAL,			remove_would_lose_data)			\
+	x(EINVAL,			btree_iter_with_journal_not_supported)	\
 	x(EROFS,			erofs_trans_commit)			\
 	x(EROFS,			erofs_no_writes)			\
 	x(EROFS,			erofs_journal_err)			\
@@ -225,7 +239,10 @@
 	x(BCH_ERR_operation_blocked,    nocow_lock_blocked)			\
 	x(EIO,				btree_node_read_err)			\
 	x(EIO,				sb_not_downgraded)			\
-	x(EIO,				btree_write_all_failed)			\
+	x(EIO,				btree_node_write_all_failed)		\
+	x(EIO,				btree_node_read_error)			\
+	x(EIO,				btree_node_read_validate_error)		\
+	x(EIO,				btree_need_topology_repair)		\
 	x(BCH_ERR_btree_node_read_err,	btree_node_read_err_fixable)		\
 	x(BCH_ERR_btree_node_read_err,	btree_node_read_err_want_retry)		\
 	x(BCH_ERR_btree_node_read_err,	btree_node_read_err_must_retry)		\
@@ -238,7 +255,9 @@
 	x(BCH_ERR_nopromote,		nopromote_congested)			\
 	x(BCH_ERR_nopromote,		nopromote_in_flight)			\
 	x(BCH_ERR_nopromote,		nopromote_no_writes)			\
-	x(BCH_ERR_nopromote,		nopromote_enomem)
+	x(BCH_ERR_nopromote,		nopromote_enomem)			\
+	x(0,				need_inode_lock)			\
+	x(0,				invalid_snapshot_node)
 
 enum bch_errcode {
 	BCH_ERR_START		= 2048,
diff --git a/fs/bcachefs/error.c b/fs/bcachefs/error.c
index d32c8bebe4..d95c40f1b6 100644
--- a/fs/bcachefs/error.c
+++ b/fs/bcachefs/error.c
@@ -1,6 +1,8 @@
 // SPDX-License-Identifier: GPL-2.0
 #include "bcachefs.h"
 #include "error.h"
+#include "journal.h"
+#include "recovery_passes.h"
 #include "super.h"
 #include "thread_with_file.h"
 
@@ -13,9 +15,11 @@ bool bch2_inconsistent_error(struct bch_fs *c)
 	switch (c->opts.errors) {
 	case BCH_ON_ERROR_continue:
 		return false;
+	case BCH_ON_ERROR_fix_safe:
 	case BCH_ON_ERROR_ro:
 		if (bch2_fs_emergency_read_only(c))
-			bch_err(c, "inconsistency detected - emergency read only");
+			bch_err(c, "inconsistency detected - emergency read only at journal seq %llu",
+				journal_cur_seq(&c->journal));
 		return true;
 	case BCH_ON_ERROR_panic:
 		panic(bch2_fmt(c, "panic after error"));
@@ -25,11 +29,16 @@ bool bch2_inconsistent_error(struct bch_fs *c)
 	}
 }
 
-void bch2_topology_error(struct bch_fs *c)
+int bch2_topology_error(struct bch_fs *c)
 {
 	set_bit(BCH_FS_topology_error, &c->flags);
-	if (!test_bit(BCH_FS_fsck_running, &c->flags))
+	if (!test_bit(BCH_FS_fsck_running, &c->flags)) {
 		bch2_inconsistent_error(c);
+		return -BCH_ERR_btree_need_topology_repair;
+	} else {
+		return bch2_run_explicit_recovery_pass(c, BCH_RECOVERY_PASS_check_topology) ?:
+			-BCH_ERR_btree_node_read_validate_error;
+	}
 }
 
 void bch2_fatal_error(struct bch_fs *c)
@@ -168,6 +177,27 @@ static struct fsck_err_state *fsck_err_get(struct bch_fs *c, const char *fmt)
 	return s;
 }
 
+/* s/fix?/fixing/ s/recreate?/recreating/ */
+static void prt_actioning(struct printbuf *out, const char *action)
+{
+	unsigned len = strlen(action);
+
+	BUG_ON(action[len - 1] != '?');
+	--len;
+
+	if (action[len - 1] == 'e')
+		--len;
+
+	prt_bytes(out, action, len);
+	prt_str(out, "ing");
+}
+
+static const u8 fsck_flags_extra[] = {
+#define x(t, n, flags)		[BCH_FSCK_ERR_##t] = flags,
+	BCH_SB_ERRS()
+#undef x
+};
+
 int bch2_fsck_err(struct bch_fs *c,
 		  enum bch_fsck_flags flags,
 		  enum bch_sb_error_id err,
@@ -178,6 +208,10 @@ int bch2_fsck_err(struct bch_fs *c,
 	bool print = true, suppressing = false, inconsistent = false;
 	struct printbuf buf = PRINTBUF, *out = &buf;
 	int ret = -BCH_ERR_fsck_ignore;
+	const char *action_orig = "fix?", *action = action_orig;
+
+	if (!WARN_ON(err >= ARRAY_SIZE(fsck_flags_extra)))
+		flags |= fsck_flags_extra[err];
 
 	if ((flags & FSCK_CAN_FIX) &&
 	    test_bit(err, c->sb.errors_silent))
@@ -189,6 +223,19 @@ int bch2_fsck_err(struct bch_fs *c,
 	prt_vprintf(out, fmt, args);
 	va_end(args);
 
+	/* Custom fix/continue/recreate/etc.? */
+	if (out->buf[out->pos - 1] == '?') {
+		const char *p = strrchr(out->buf, ',');
+		if (p) {
+			out->pos = p - out->buf;
+			action = kstrdup(p + 2, GFP_KERNEL);
+			if (!action) {
+				ret = -ENOMEM;
+				goto err;
+			}
+		}
+	}
+
 	mutex_lock(&c->fsck_error_msgs_lock);
 	s = fsck_err_get(c, fmt);
 	if (s) {
@@ -200,12 +247,16 @@ int bch2_fsck_err(struct bch_fs *c,
 		if (s->last_msg && !strcmp(buf.buf, s->last_msg)) {
 			ret = s->ret;
 			mutex_unlock(&c->fsck_error_msgs_lock);
-			printbuf_exit(&buf);
-			return ret;
+			goto err;
 		}
 
 		kfree(s->last_msg);
 		s->last_msg = kstrdup(buf.buf, GFP_KERNEL);
+		if (!s->last_msg) {
+			mutex_unlock(&c->fsck_error_msgs_lock);
+			ret = -ENOMEM;
+			goto err;
+		}
 
 		if (c->opts.ratelimit_errors &&
 		    !(flags & FSCK_NO_RATELIMIT) &&
@@ -224,14 +275,22 @@ int bch2_fsck_err(struct bch_fs *c,
 		prt_printf(out, bch2_log_msg(c, ""));
 #endif
 
-	if (!test_bit(BCH_FS_fsck_running, &c->flags)) {
+	if ((flags & FSCK_CAN_FIX) &&
+	    (flags & FSCK_AUTOFIX) &&
+	    (c->opts.errors == BCH_ON_ERROR_continue ||
+	     c->opts.errors == BCH_ON_ERROR_fix_safe)) {
+		prt_str(out, ", ");
+		prt_actioning(out, action);
+		ret = -BCH_ERR_fsck_fix;
+	} else if (!test_bit(BCH_FS_fsck_running, &c->flags)) {
 		if (c->opts.errors != BCH_ON_ERROR_continue ||
 		    !(flags & (FSCK_CAN_FIX|FSCK_CAN_IGNORE))) {
 			prt_str(out, ", shutting down");
 			inconsistent = true;
 			ret = -BCH_ERR_fsck_errors_not_fixed;
 		} else if (flags & FSCK_CAN_FIX) {
-			prt_str(out, ", fixing");
+			prt_str(out, ", ");
+			prt_actioning(out, action);
 			ret = -BCH_ERR_fsck_fix;
 		} else {
 			prt_str(out, ", continuing");
@@ -246,16 +305,16 @@ int bch2_fsck_err(struct bch_fs *c,
 			: c->opts.fix_errors;
 
 		if (fix == FSCK_FIX_ask) {
-			int ask;
+			prt_str(out, ", ");
+			prt_str(out, action);
 
-			prt_str(out, ": fix?");
 			if (bch2_fs_stdio_redirect(c))
 				bch2_print(c, "%s", out->buf);
 			else
 				bch2_print_string_as_lines(KERN_ERR, out->buf);
 			print = false;
 
-			ask = bch2_fsck_ask_yn(c);
+			int ask = bch2_fsck_ask_yn(c);
 
 			if (ask >= YN_ALLNO && s)
 				s->fix = ask == YN_ALLNO
@@ -268,10 +327,12 @@ int bch2_fsck_err(struct bch_fs *c,
 		} else if (fix == FSCK_FIX_yes ||
 			   (c->opts.nochanges &&
 			    !(flags & FSCK_CAN_IGNORE))) {
-			prt_str(out, ", fixing");
+			prt_str(out, ", ");
+			prt_actioning(out, action);
 			ret = -BCH_ERR_fsck_fix;
 		} else {
-			prt_str(out, ", not fixing");
+			prt_str(out, ", not ");
+			prt_actioning(out, action);
 		}
 	} else if (flags & FSCK_NEED_FSCK) {
 		prt_str(out, " (run fsck to correct)");
@@ -303,8 +364,6 @@ int bch2_fsck_err(struct bch_fs *c,
 
 	mutex_unlock(&c->fsck_error_msgs_lock);
 
-	printbuf_exit(&buf);
-
 	if (inconsistent)
 		bch2_inconsistent_error(c);
 
@@ -314,7 +373,10 @@ int bch2_fsck_err(struct bch_fs *c,
 		set_bit(BCH_FS_errors_not_fixed, &c->flags);
 		set_bit(BCH_FS_error, &c->flags);
 	}
-
+err:
+	if (action != action_orig)
+		kfree(action);
+	printbuf_exit(&buf);
 	return ret;
 }
 
diff --git a/fs/bcachefs/error.h b/fs/bcachefs/error.h
index fec17d1353..777711504c 100644
--- a/fs/bcachefs/error.h
+++ b/fs/bcachefs/error.h
@@ -30,7 +30,13 @@ struct work_struct;
 
 bool bch2_inconsistent_error(struct bch_fs *);
 
-void bch2_topology_error(struct bch_fs *);
+int bch2_topology_error(struct bch_fs *);
+
+#define bch2_fs_topology_error(c, ...)					\
+({									\
+	bch_err(c, "btree topology error: " __VA_ARGS__);		\
+	bch2_topology_error(c);						\
+})
 
 #define bch2_fs_inconsistent(c, ...)					\
 ({									\
@@ -102,13 +108,6 @@ struct fsck_err_state {
 	char			*last_msg;
 };
 
-enum bch_fsck_flags {
-	FSCK_CAN_FIX		= 1 << 0,
-	FSCK_CAN_IGNORE		= 1 << 1,
-	FSCK_NEED_FSCK		= 1 << 2,
-	FSCK_NO_RATELIMIT	= 1 << 3,
-};
-
 #define fsck_err_count(_c, _err)	bch2_sb_err_count(_c, BCH_FSCK_ERR_##_err)
 
 __printf(4, 5) __cold
@@ -191,9 +190,9 @@ do {									\
 
 void bch2_fatal_error(struct bch_fs *);
 
-#define bch2_fs_fatal_error(c, ...)					\
+#define bch2_fs_fatal_error(c, _msg, ...)				\
 do {									\
-	bch_err(c, __VA_ARGS__);					\
+	bch_err(c, "%s(): fatal error " _msg, __func__, ##__VA_ARGS__);	\
 	bch2_fatal_error(c);						\
 } while (0)
 
diff --git a/fs/bcachefs/extent_update.c b/fs/bcachefs/extent_update.c
index b9033bb4f1..5f4fecb358 100644
--- a/fs/bcachefs/extent_update.c
+++ b/fs/bcachefs/extent_update.c
@@ -72,7 +72,7 @@ static int count_iters_for_insert(struct btree_trans *trans,
 
 		for_each_btree_key_norestart(trans, iter,
 				   BTREE_ID_reflink, POS(0, idx + offset),
-				   BTREE_ITER_SLOTS, r_k, ret2) {
+				   BTREE_ITER_slots, r_k, ret2) {
 			if (bkey_ge(bkey_start_pos(r_k.k), POS(0, idx + sectors)))
 				break;
 
diff --git a/fs/bcachefs/extents.c b/fs/bcachefs/extents.c
index 61395b113d..410b8bd81b 100644
--- a/fs/bcachefs/extents.c
+++ b/fs/bcachefs/extents.c
@@ -71,6 +71,12 @@ void bch2_mark_io_failure(struct bch_io_failures *failed,
 	}
 }
 
+static inline u64 dev_latency(struct bch_fs *c, unsigned dev)
+{
+	struct bch_dev *ca = bch2_dev_rcu(c, dev);
+	return ca ? atomic64_read(&ca->cur_latency[READ]) : S64_MAX;
+}
+
 /*
  * returns true if p1 is better than p2:
  */
@@ -79,11 +85,8 @@ static inline bool ptr_better(struct bch_fs *c,
 			      const struct extent_ptr_decoded p2)
 {
 	if (likely(!p1.idx && !p2.idx)) {
-		struct bch_dev *dev1 = bch_dev_bkey_exists(c, p1.ptr.dev);
-		struct bch_dev *dev2 = bch_dev_bkey_exists(c, p2.ptr.dev);
-
-		u64 l1 = atomic64_read(&dev1->cur_latency[READ]);
-		u64 l2 = atomic64_read(&dev2->cur_latency[READ]);
+		u64 l1 = dev_latency(c, p1.ptr.dev);
+		u64 l2 = dev_latency(c, p2.ptr.dev);
 
 		/* Pick at random, biased in favor of the faster device: */
 
@@ -109,21 +112,21 @@ int bch2_bkey_pick_read_device(struct bch_fs *c, struct bkey_s_c k,
 	const union bch_extent_entry *entry;
 	struct extent_ptr_decoded p;
 	struct bch_dev_io_failures *f;
-	struct bch_dev *ca;
 	int ret = 0;
 
 	if (k.k->type == KEY_TYPE_error)
 		return -EIO;
 
+	rcu_read_lock();
 	bkey_for_each_ptr_decode(k.k, ptrs, p, entry) {
 		/*
 		 * Unwritten extent: no need to actually read, treat it as a
 		 * hole and return 0s:
 		 */
-		if (p.ptr.unwritten)
-			return 0;
-
-		ca = bch_dev_bkey_exists(c, p.ptr.dev);
+		if (p.ptr.unwritten) {
+			ret = 0;
+			break;
+		}
 
 		/*
 		 * If there are any dirty pointers it's an error if we can't
@@ -132,7 +135,9 @@ int bch2_bkey_pick_read_device(struct bch_fs *c, struct bkey_s_c k,
 		if (!ret && !p.ptr.cached)
 			ret = -EIO;
 
-		if (p.ptr.cached && ptr_stale(ca, &p.ptr))
+		struct bch_dev *ca = bch2_dev_rcu(c, p.ptr.dev);
+
+		if (p.ptr.cached && (!ca || dev_ptr_stale_rcu(ca, &p.ptr)))
 			continue;
 
 		f = failed ? dev_io_failures(failed, p.ptr.dev) : NULL;
@@ -141,12 +146,13 @@ int bch2_bkey_pick_read_device(struct bch_fs *c, struct bkey_s_c k,
 				? f->idx
 				: f->idx + 1;
 
-		if (!p.idx &&
-		    !bch2_dev_is_readable(ca))
+		if (!p.idx && !ca)
 			p.idx++;
 
-		if (bch2_force_reconstruct_read &&
-		    !p.idx && p.has_ec)
+		if (!p.idx && p.has_ec && bch2_force_reconstruct_read)
+			p.idx++;
+
+		if (!p.idx && !bch2_dev_is_readable(ca))
 			p.idx++;
 
 		if (p.idx >= (unsigned) p.has_ec + 1)
@@ -158,6 +164,7 @@ int bch2_bkey_pick_read_device(struct bch_fs *c, struct bkey_s_c k,
 		*pick = p;
 		ret = 1;
 	}
+	rcu_read_unlock();
 
 	return ret;
 }
@@ -165,7 +172,7 @@ int bch2_bkey_pick_read_device(struct bch_fs *c, struct bkey_s_c k,
 /* KEY_TYPE_btree_ptr: */
 
 int bch2_btree_ptr_invalid(struct bch_fs *c, struct bkey_s_c k,
-			   enum bkey_invalid_flags flags,
+			   enum bch_validate_flags flags,
 			   struct printbuf *err)
 {
 	int ret = 0;
@@ -186,16 +193,26 @@ void bch2_btree_ptr_to_text(struct printbuf *out, struct bch_fs *c,
 }
 
 int bch2_btree_ptr_v2_invalid(struct bch_fs *c, struct bkey_s_c k,
-			      enum bkey_invalid_flags flags,
+			      enum bch_validate_flags flags,
 			      struct printbuf *err)
 {
+	struct bkey_s_c_btree_ptr_v2 bp = bkey_s_c_to_btree_ptr_v2(k);
 	int ret = 0;
 
-	bkey_fsck_err_on(bkey_val_u64s(k.k) > BKEY_BTREE_PTR_VAL_U64s_MAX, c, err,
-			 btree_ptr_v2_val_too_big,
+	bkey_fsck_err_on(bkey_val_u64s(k.k) > BKEY_BTREE_PTR_VAL_U64s_MAX,
+			 c, err, btree_ptr_v2_val_too_big,
 			 "value too big (%zu > %zu)",
 			 bkey_val_u64s(k.k), BKEY_BTREE_PTR_VAL_U64s_MAX);
 
+	bkey_fsck_err_on(bpos_ge(bp.v->min_key, bp.k->p),
+			 c, err, btree_ptr_v2_min_key_bad,
+			 "min_key > key");
+
+	if (flags & BCH_VALIDATE_write)
+		bkey_fsck_err_on(!bp.v->sectors_written,
+				 c, err, btree_ptr_v2_written_0,
+				 "sectors_written == 0");
+
 	ret = bch2_bkey_ptrs_invalid(c, k, flags, err);
 fsck_err:
 	return ret;
@@ -242,7 +259,6 @@ bool bch2_extent_merge(struct bch_fs *c, struct bkey_s l, struct bkey_s_c r)
 	const union bch_extent_entry *en_r;
 	struct extent_ptr_decoded lp, rp;
 	bool use_right_ptr;
-	struct bch_dev *ca;
 
 	en_l = l_ptrs.start;
 	en_r = r_ptrs.start;
@@ -273,8 +289,12 @@ bool bch2_extent_merge(struct bch_fs *c, struct bkey_s l, struct bkey_s_c r)
 			return false;
 
 		/* Extents may not straddle buckets: */
-		ca = bch_dev_bkey_exists(c, lp.ptr.dev);
-		if (PTR_BUCKET_NR(ca, &lp.ptr) != PTR_BUCKET_NR(ca, &rp.ptr))
+		rcu_read_lock();
+		struct bch_dev *ca = bch2_dev_rcu(c, lp.ptr.dev);
+		bool same_bucket = ca && PTR_BUCKET_NR(ca, &lp.ptr) == PTR_BUCKET_NR(ca, &rp.ptr);
+		rcu_read_unlock();
+
+		if (!same_bucket)
 			return false;
 
 		if (lp.has_ec			!= rp.has_ec ||
@@ -380,7 +400,7 @@ bool bch2_extent_merge(struct bch_fs *c, struct bkey_s l, struct bkey_s_c r)
 /* KEY_TYPE_reservation: */
 
 int bch2_reservation_invalid(struct bch_fs *c, struct bkey_s_c k,
-			     enum bkey_invalid_flags flags,
+			     enum bch_validate_flags flags,
 			     struct printbuf *err)
 {
 	struct bkey_s_c_reservation r = bkey_s_c_to_reservation(k);
@@ -662,16 +682,16 @@ static inline unsigned __extent_ptr_durability(struct bch_dev *ca, struct extent
 
 unsigned bch2_extent_ptr_desired_durability(struct bch_fs *c, struct extent_ptr_decoded *p)
 {
-	struct bch_dev *ca = bch_dev_bkey_exists(c, p->ptr.dev);
+	struct bch_dev *ca = bch2_dev_rcu(c, p->ptr.dev);
 
-	return __extent_ptr_durability(ca, p);
+	return ca ? __extent_ptr_durability(ca, p) : 0;
 }
 
 unsigned bch2_extent_ptr_durability(struct bch_fs *c, struct extent_ptr_decoded *p)
 {
-	struct bch_dev *ca = bch_dev_bkey_exists(c, p->ptr.dev);
+	struct bch_dev *ca = bch2_dev_rcu(c, p->ptr.dev);
 
-	if (ca->mi.state == BCH_MEMBER_STATE_failed)
+	if (!ca || ca->mi.state == BCH_MEMBER_STATE_failed)
 		return 0;
 
 	return __extent_ptr_durability(ca, p);
@@ -684,8 +704,10 @@ unsigned bch2_bkey_durability(struct bch_fs *c, struct bkey_s_c k)
 	struct extent_ptr_decoded p;
 	unsigned durability = 0;
 
+	rcu_read_lock();
 	bkey_for_each_ptr_decode(k.k, ptrs, p, entry)
 		durability += bch2_extent_ptr_durability(c, &p);
+	rcu_read_unlock();
 
 	return durability;
 }
@@ -697,9 +719,11 @@ static unsigned bch2_bkey_durability_safe(struct bch_fs *c, struct bkey_s_c k)
 	struct extent_ptr_decoded p;
 	unsigned durability = 0;
 
+	rcu_read_lock();
 	bkey_for_each_ptr_decode(k.k, ptrs, p, entry)
 		if (p.ptr.dev < c->sb.nr_devices && c->devs[p.ptr.dev])
 			durability += bch2_extent_ptr_durability(c, &p);
+	rcu_read_unlock();
 
 	return durability;
 }
@@ -828,8 +852,6 @@ union bch_extent_entry *bch2_bkey_drop_ptr(struct bkey_s k,
 
 void bch2_bkey_drop_device(struct bkey_s k, unsigned dev)
 {
-	struct bch_extent_ptr *ptr;
-
 	bch2_bkey_drop_ptrs(k, ptr, ptr->dev == dev);
 }
 
@@ -855,14 +877,21 @@ const struct bch_extent_ptr *bch2_bkey_has_device_c(struct bkey_s_c k, unsigned
 bool bch2_bkey_has_target(struct bch_fs *c, struct bkey_s_c k, unsigned target)
 {
 	struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k);
+	struct bch_dev *ca;
+	bool ret = false;
 
+	rcu_read_lock();
 	bkey_for_each_ptr(ptrs, ptr)
 		if (bch2_dev_in_target(c, ptr->dev, target) &&
+		    (ca = bch2_dev_rcu(c, ptr->dev)) &&
 		    (!ptr->cached ||
-		     !ptr_stale(bch_dev_bkey_exists(c, ptr->dev), ptr)))
-			return true;
+		     !dev_ptr_stale_rcu(ca, ptr))) {
+			ret = true;
+			break;
+		}
+	rcu_read_unlock();
 
-	return false;
+	return ret;
 }
 
 bool bch2_bkey_matches_ptr(struct bch_fs *c, struct bkey_s_c k,
@@ -964,15 +993,47 @@ void bch2_extent_ptr_set_cached(struct bkey_s k, struct bch_extent_ptr *ptr)
  */
 bool bch2_extent_normalize(struct bch_fs *c, struct bkey_s k)
 {
-	struct bch_extent_ptr *ptr;
+	struct bch_dev *ca;
 
+	rcu_read_lock();
 	bch2_bkey_drop_ptrs(k, ptr,
 		ptr->cached &&
-		ptr_stale(bch_dev_bkey_exists(c, ptr->dev), ptr));
+		(ca = bch2_dev_rcu(c, ptr->dev)) &&
+		dev_ptr_stale_rcu(ca, ptr) > 0);
+	rcu_read_unlock();
 
 	return bkey_deleted(k.k);
 }
 
+void bch2_extent_ptr_to_text(struct printbuf *out, struct bch_fs *c, const struct bch_extent_ptr *ptr)
+{
+	out->atomic++;
+	rcu_read_lock();
+	struct bch_dev *ca = bch2_dev_rcu(c, ptr->dev);
+	if (!ca) {
+		prt_printf(out, "ptr: %u:%llu gen %u%s", ptr->dev,
+			   (u64) ptr->offset, ptr->gen,
+			   ptr->cached ? " cached" : "");
+	} else {
+		u32 offset;
+		u64 b = sector_to_bucket_and_offset(ca, ptr->offset, &offset);
+
+		prt_printf(out, "ptr: %u:%llu:%u gen %u",
+			   ptr->dev, b, offset, ptr->gen);
+		if (ptr->cached)
+			prt_str(out, " cached");
+		if (ptr->unwritten)
+			prt_str(out, " unwritten");
+		int stale = dev_ptr_stale_rcu(ca, ptr);
+		if (stale > 0)
+			prt_printf(out, " stale");
+		else if (stale)
+			prt_printf(out, " invalid");
+	}
+	rcu_read_unlock();
+	--out->atomic;
+}
+
 void bch2_bkey_ptrs_to_text(struct printbuf *out, struct bch_fs *c,
 			    struct bkey_s_c k)
 {
@@ -988,42 +1049,22 @@ void bch2_bkey_ptrs_to_text(struct printbuf *out, struct bch_fs *c,
 			prt_printf(out, " ");
 
 		switch (__extent_entry_type(entry)) {
-		case BCH_EXTENT_ENTRY_ptr: {
-			const struct bch_extent_ptr *ptr = entry_to_ptr(entry);
-			struct bch_dev *ca = c && ptr->dev < c->sb.nr_devices && c->devs[ptr->dev]
-				? bch_dev_bkey_exists(c, ptr->dev)
-				: NULL;
-
-			if (!ca) {
-				prt_printf(out, "ptr: %u:%llu gen %u%s", ptr->dev,
-				       (u64) ptr->offset, ptr->gen,
-				       ptr->cached ? " cached" : "");
-			} else {
-				u32 offset;
-				u64 b = sector_to_bucket_and_offset(ca, ptr->offset, &offset);
-
-				prt_printf(out, "ptr: %u:%llu:%u gen %u",
-					   ptr->dev, b, offset, ptr->gen);
-				if (ptr->cached)
-					prt_str(out, " cached");
-				if (ptr->unwritten)
-					prt_str(out, " unwritten");
-				if (ca && ptr_stale(ca, ptr))
-					prt_printf(out, " stale");
-			}
+		case BCH_EXTENT_ENTRY_ptr:
+			bch2_extent_ptr_to_text(out, c, entry_to_ptr(entry));
 			break;
-		}
+
 		case BCH_EXTENT_ENTRY_crc32:
 		case BCH_EXTENT_ENTRY_crc64:
 		case BCH_EXTENT_ENTRY_crc128: {
 			struct bch_extent_crc_unpacked crc =
 				bch2_extent_crc_unpack(k.k, entry_to_crc(entry));
 
-			prt_printf(out, "crc: c_size %u size %u offset %u nonce %u csum %s compress ",
+			prt_printf(out, "crc: c_size %u size %u offset %u nonce %u csum ",
 			       crc.compressed_size,
 			       crc.uncompressed_size,
-			       crc.offset, crc.nonce,
-			       bch2_csum_types[crc.csum_type]);
+			       crc.offset, crc.nonce);
+			bch2_prt_csum_type(out, crc.csum_type);
+			prt_str(out, " compress ");
 			bch2_prt_compression_type(out, crc.compression_type);
 			break;
 		}
@@ -1057,55 +1098,50 @@ void bch2_bkey_ptrs_to_text(struct printbuf *out, struct bch_fs *c,
 
 static int extent_ptr_invalid(struct bch_fs *c,
 			      struct bkey_s_c k,
-			      enum bkey_invalid_flags flags,
+			      enum bch_validate_flags flags,
 			      const struct bch_extent_ptr *ptr,
 			      unsigned size_ondisk,
 			      bool metadata,
 			      struct printbuf *err)
 {
-	struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k);
-	u64 bucket;
-	u32 bucket_offset;
-	struct bch_dev *ca;
 	int ret = 0;
 
-	if (!bch2_dev_exists2(c, ptr->dev)) {
-		/*
-		 * If we're in the write path this key might have already been
-		 * overwritten, and we could be seeing a device that doesn't
-		 * exist anymore due to racing with device removal:
-		 */
-		if (flags & BKEY_INVALID_WRITE)
-			return 0;
-
-		bkey_fsck_err(c, err, ptr_to_invalid_device,
-			   "pointer to invalid device (%u)", ptr->dev);
+	rcu_read_lock();
+	struct bch_dev *ca = bch2_dev_rcu(c, ptr->dev);
+	if (!ca) {
+		rcu_read_unlock();
+		return 0;
 	}
+	u32 bucket_offset;
+	u64 bucket = sector_to_bucket_and_offset(ca, ptr->offset, &bucket_offset);
+	unsigned first_bucket	= ca->mi.first_bucket;
+	u64 nbuckets		= ca->mi.nbuckets;
+	unsigned bucket_size	= ca->mi.bucket_size;
+	rcu_read_unlock();
 
-	ca = bch_dev_bkey_exists(c, ptr->dev);
+	struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k);
 	bkey_for_each_ptr(ptrs, ptr2)
 		bkey_fsck_err_on(ptr != ptr2 && ptr->dev == ptr2->dev, c, err,
 				 ptr_to_duplicate_device,
 				 "multiple pointers to same device (%u)", ptr->dev);
 
-	bucket = sector_to_bucket_and_offset(ca, ptr->offset, &bucket_offset);
 
-	bkey_fsck_err_on(bucket >= ca->mi.nbuckets, c, err,
+	bkey_fsck_err_on(bucket >= nbuckets, c, err,
 			 ptr_after_last_bucket,
-			 "pointer past last bucket (%llu > %llu)", bucket, ca->mi.nbuckets);
-	bkey_fsck_err_on(ptr->offset < bucket_to_sector(ca, ca->mi.first_bucket), c, err,
+			 "pointer past last bucket (%llu > %llu)", bucket, nbuckets);
+	bkey_fsck_err_on(bucket < first_bucket, c, err,
 			 ptr_before_first_bucket,
-			 "pointer before first bucket (%llu < %u)", bucket, ca->mi.first_bucket);
-	bkey_fsck_err_on(bucket_offset + size_ondisk > ca->mi.bucket_size, c, err,
+			 "pointer before first bucket (%llu < %u)", bucket, first_bucket);
+	bkey_fsck_err_on(bucket_offset + size_ondisk > bucket_size, c, err,
 			 ptr_spans_multiple_buckets,
 			 "pointer spans multiple buckets (%u + %u > %u)",
-		       bucket_offset, size_ondisk, ca->mi.bucket_size);
+		       bucket_offset, size_ondisk, bucket_size);
 fsck_err:
 	return ret;
 }
 
 int bch2_bkey_ptrs_invalid(struct bch_fs *c, struct bkey_s_c k,
-			   enum bkey_invalid_flags flags,
+			   enum bch_validate_flags flags,
 			   struct printbuf *err)
 {
 	struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k);
@@ -1181,7 +1217,7 @@ int bch2_bkey_ptrs_invalid(struct bch_fs *c, struct bkey_s_c k,
 
 			bkey_fsck_err_on(crc_is_encoded(crc) &&
 					 (crc.uncompressed_size > c->opts.encoded_extent_max >> 9) &&
-					 (flags & (BKEY_INVALID_WRITE|BKEY_INVALID_COMMIT)), c, err,
+					 (flags & (BCH_VALIDATE_write|BCH_VALIDATE_commit)), c, err,
 					 ptr_crc_uncompressed_size_too_big,
 					 "too large encoded extent");
 
diff --git a/fs/bcachefs/extents.h b/fs/bcachefs/extents.h
index 6bf839d69e..1ade959652 100644
--- a/fs/bcachefs/extents.h
+++ b/fs/bcachefs/extents.h
@@ -8,7 +8,7 @@
 
 struct bch_fs;
 struct btree_trans;
-enum bkey_invalid_flags;
+enum bch_validate_flags;
 
 /* extent entries: */
 
@@ -43,6 +43,11 @@ enum bkey_invalid_flags;
 #define extent_entry_next(_entry)					\
 	((typeof(_entry)) ((void *) (_entry) + extent_entry_bytes(_entry)))
 
+#define extent_entry_next_safe(_entry, _end)				\
+	(likely(__extent_entry_type(_entry) < BCH_EXTENT_ENTRY_MAX)	\
+	 ? extent_entry_next(_entry)					\
+	 : _end)
+
 static inline unsigned
 __extent_entry_type(const union bch_extent_entry *e)
 {
@@ -103,17 +108,17 @@ static inline void extent_entry_drop(struct bkey_s k, union bch_extent_entry *en
 
 static inline bool extent_entry_is_ptr(const union bch_extent_entry *e)
 {
-	return extent_entry_type(e) == BCH_EXTENT_ENTRY_ptr;
+	return __extent_entry_type(e) == BCH_EXTENT_ENTRY_ptr;
 }
 
 static inline bool extent_entry_is_stripe_ptr(const union bch_extent_entry *e)
 {
-	return extent_entry_type(e) == BCH_EXTENT_ENTRY_stripe_ptr;
+	return __extent_entry_type(e) == BCH_EXTENT_ENTRY_stripe_ptr;
 }
 
 static inline bool extent_entry_is_crc(const union bch_extent_entry *e)
 {
-	switch (extent_entry_type(e)) {
+	switch (__extent_entry_type(e)) {
 	case BCH_EXTENT_ENTRY_crc32:
 	case BCH_EXTENT_ENTRY_crc64:
 	case BCH_EXTENT_ENTRY_crc128:
@@ -280,7 +285,7 @@ static inline struct bkey_ptrs bch2_bkey_ptrs(struct bkey_s k)
 #define __bkey_extent_entry_for_each_from(_start, _end, _entry)		\
 	for ((_entry) = (_start);					\
 	     (_entry) < (_end);						\
-	     (_entry) = extent_entry_next(_entry))
+	     (_entry) = extent_entry_next_safe(_entry, _end))
 
 #define __bkey_ptr_next(_ptr, _end)					\
 ({									\
@@ -318,7 +323,7 @@ static inline struct bkey_ptrs bch2_bkey_ptrs(struct bkey_s k)
 	(_ptr).has_ec	= false;					\
 									\
 	__bkey_extent_entry_for_each_from(_entry, _end, _entry)		\
-		switch (extent_entry_type(_entry)) {			\
+		switch (__extent_entry_type(_entry)) {			\
 		case BCH_EXTENT_ENTRY_ptr:				\
 			(_ptr).ptr		= _entry->ptr;		\
 			goto out;					\
@@ -344,7 +349,7 @@ out:									\
 	for ((_ptr).crc = bch2_extent_crc_unpack(_k, NULL),		\
 	     (_entry) = _start;						\
 	     __bkey_ptr_next_decode(_k, _end, _ptr, _entry);		\
-	     (_entry) = extent_entry_next(_entry))
+	     (_entry) = extent_entry_next_safe(_entry, _end))
 
 #define bkey_for_each_ptr_decode(_k, _p, _ptr, _entry)			\
 	__bkey_for_each_ptr_decode(_k, (_p).start, (_p).end,		\
@@ -401,12 +406,12 @@ int bch2_bkey_pick_read_device(struct bch_fs *, struct bkey_s_c,
 /* KEY_TYPE_btree_ptr: */
 
 int bch2_btree_ptr_invalid(struct bch_fs *, struct bkey_s_c,
-			   enum bkey_invalid_flags, struct printbuf *);
+			   enum bch_validate_flags, struct printbuf *);
 void bch2_btree_ptr_to_text(struct printbuf *, struct bch_fs *,
 			    struct bkey_s_c);
 
 int bch2_btree_ptr_v2_invalid(struct bch_fs *, struct bkey_s_c,
-			      enum bkey_invalid_flags, struct printbuf *);
+			      enum bch_validate_flags, struct printbuf *);
 void bch2_btree_ptr_v2_to_text(struct printbuf *, struct bch_fs *, struct bkey_s_c);
 void bch2_btree_ptr_v2_compat(enum btree_id, unsigned, unsigned,
 			      int, struct bkey_s);
@@ -443,7 +448,7 @@ bool bch2_extent_merge(struct bch_fs *, struct bkey_s, struct bkey_s_c);
 /* KEY_TYPE_reservation: */
 
 int bch2_reservation_invalid(struct bch_fs *, struct bkey_s_c,
-			     enum bkey_invalid_flags, struct printbuf *);
+			     enum bch_validate_flags, struct printbuf *);
 void bch2_reservation_to_text(struct printbuf *, struct bch_fs *, struct bkey_s_c);
 bool bch2_reservation_merge(struct bch_fs *, struct bkey_s, struct bkey_s_c);
 
@@ -591,30 +596,6 @@ static inline struct bch_devs_list bch2_bkey_cached_devs(struct bkey_s_c k)
 	return ret;
 }
 
-static inline unsigned bch2_bkey_ptr_data_type(struct bkey_s_c k, const struct bch_extent_ptr *ptr)
-{
-	switch (k.k->type) {
-	case KEY_TYPE_btree_ptr:
-	case KEY_TYPE_btree_ptr_v2:
-		return BCH_DATA_btree;
-	case KEY_TYPE_extent:
-	case KEY_TYPE_reflink_v:
-		return BCH_DATA_user;
-	case KEY_TYPE_stripe: {
-		struct bkey_s_c_stripe s = bkey_s_c_to_stripe(k);
-
-		BUG_ON(ptr < s.v->ptrs ||
-		       ptr >= s.v->ptrs + s.v->nr_blocks);
-
-		return ptr >= s.v->ptrs + s.v->nr_blocks - s.v->nr_redundant
-			? BCH_DATA_parity
-			: BCH_DATA_user;
-	}
-	default:
-		BUG();
-	}
-}
-
 unsigned bch2_bkey_nr_ptrs(struct bkey_s_c);
 unsigned bch2_bkey_nr_ptrs_allocated(struct bkey_s_c);
 unsigned bch2_bkey_nr_ptrs_fully_allocated(struct bkey_s_c);
@@ -673,7 +654,7 @@ union bch_extent_entry *bch2_bkey_drop_ptr(struct bkey_s,
 do {									\
 	struct bkey_ptrs _ptrs = bch2_bkey_ptrs(_k);			\
 									\
-	_ptr = &_ptrs.start->ptr;					\
+	struct bch_extent_ptr *_ptr = &_ptrs.start->ptr;		\
 									\
 	while ((_ptr = bkey_ptr_next(_ptrs, _ptr))) {			\
 		if (_cond) {						\
@@ -695,10 +676,11 @@ bch2_extent_has_ptr(struct bkey_s_c, struct extent_ptr_decoded, struct bkey_s);
 void bch2_extent_ptr_set_cached(struct bkey_s, struct bch_extent_ptr *);
 
 bool bch2_extent_normalize(struct bch_fs *, struct bkey_s);
+void bch2_extent_ptr_to_text(struct printbuf *out, struct bch_fs *, const struct bch_extent_ptr *);
 void bch2_bkey_ptrs_to_text(struct printbuf *, struct bch_fs *,
 			    struct bkey_s_c);
 int bch2_bkey_ptrs_invalid(struct bch_fs *, struct bkey_s_c,
-			   enum bkey_invalid_flags, struct printbuf *);
+			   enum bch_validate_flags, struct printbuf *);
 
 void bch2_ptr_swab(struct bkey_s);
 
diff --git a/fs/bcachefs/eytzinger.c b/fs/bcachefs/eytzinger.c
new file mode 100644
index 0000000000..2eaffe37b5
--- /dev/null
+++ b/fs/bcachefs/eytzinger.c
@@ -0,0 +1,305 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include "eytzinger.h"
+
+/**
+ * is_aligned - is this pointer & size okay for word-wide copying?
+ * @base: pointer to data
+ * @size: size of each element
+ * @align: required alignment (typically 4 or 8)
+ *
+ * Returns true if elements can be copied using word loads and stores.
+ * The size must be a multiple of the alignment, and the base address must
+ * be if we do not have CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS.
+ *
+ * For some reason, gcc doesn't know to optimize "if (a & mask || b & mask)"
+ * to "if ((a | b) & mask)", so we do that by hand.
+ */
+__attribute_const__ __always_inline
+static bool is_aligned(const void *base, size_t size, unsigned char align)
+{
+	unsigned char lsbits = (unsigned char)size;
+
+	(void)base;
+#ifndef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
+	lsbits |= (unsigned char)(uintptr_t)base;
+#endif
+	return (lsbits & (align - 1)) == 0;
+}
+
+/**
+ * swap_words_32 - swap two elements in 32-bit chunks
+ * @a: pointer to the first element to swap
+ * @b: pointer to the second element to swap
+ * @n: element size (must be a multiple of 4)
+ *
+ * Exchange the two objects in memory.  This exploits base+index addressing,
+ * which basically all CPUs have, to minimize loop overhead computations.
+ *
+ * For some reason, on x86 gcc 7.3.0 adds a redundant test of n at the
+ * bottom of the loop, even though the zero flag is still valid from the
+ * subtract (since the intervening mov instructions don't alter the flags).
+ * Gcc 8.1.0 doesn't have that problem.
+ */
+static void swap_words_32(void *a, void *b, size_t n)
+{
+	do {
+		u32 t = *(u32 *)(a + (n -= 4));
+		*(u32 *)(a + n) = *(u32 *)(b + n);
+		*(u32 *)(b + n) = t;
+	} while (n);
+}
+
+/**
+ * swap_words_64 - swap two elements in 64-bit chunks
+ * @a: pointer to the first element to swap
+ * @b: pointer to the second element to swap
+ * @n: element size (must be a multiple of 8)
+ *
+ * Exchange the two objects in memory.  This exploits base+index
+ * addressing, which basically all CPUs have, to minimize loop overhead
+ * computations.
+ *
+ * We'd like to use 64-bit loads if possible.  If they're not, emulating
+ * one requires base+index+4 addressing which x86 has but most other
+ * processors do not.  If CONFIG_64BIT, we definitely have 64-bit loads,
+ * but it's possible to have 64-bit loads without 64-bit pointers (e.g.
+ * x32 ABI).  Are there any cases the kernel needs to worry about?
+ */
+static void swap_words_64(void *a, void *b, size_t n)
+{
+	do {
+#ifdef CONFIG_64BIT
+		u64 t = *(u64 *)(a + (n -= 8));
+		*(u64 *)(a + n) = *(u64 *)(b + n);
+		*(u64 *)(b + n) = t;
+#else
+		/* Use two 32-bit transfers to avoid base+index+4 addressing */
+		u32 t = *(u32 *)(a + (n -= 4));
+		*(u32 *)(a + n) = *(u32 *)(b + n);
+		*(u32 *)(b + n) = t;
+
+		t = *(u32 *)(a + (n -= 4));
+		*(u32 *)(a + n) = *(u32 *)(b + n);
+		*(u32 *)(b + n) = t;
+#endif
+	} while (n);
+}
+
+/**
+ * swap_bytes - swap two elements a byte at a time
+ * @a: pointer to the first element to swap
+ * @b: pointer to the second element to swap
+ * @n: element size
+ *
+ * This is the fallback if alignment doesn't allow using larger chunks.
+ */
+static void swap_bytes(void *a, void *b, size_t n)
+{
+	do {
+		char t = ((char *)a)[--n];
+		((char *)a)[n] = ((char *)b)[n];
+		((char *)b)[n] = t;
+	} while (n);
+}
+
+/*
+ * The values are arbitrary as long as they can't be confused with
+ * a pointer, but small integers make for the smallest compare
+ * instructions.
+ */
+#define SWAP_WORDS_64 (swap_r_func_t)0
+#define SWAP_WORDS_32 (swap_r_func_t)1
+#define SWAP_BYTES    (swap_r_func_t)2
+#define SWAP_WRAPPER  (swap_r_func_t)3
+
+struct wrapper {
+	cmp_func_t cmp;
+	swap_func_t swap_func;
+};
+
+/*
+ * The function pointer is last to make tail calls most efficient if the
+ * compiler decides not to inline this function.
+ */
+static void do_swap(void *a, void *b, size_t size, swap_r_func_t swap_func, const void *priv)
+{
+	if (swap_func == SWAP_WRAPPER) {
+		((const struct wrapper *)priv)->swap_func(a, b, (int)size);
+		return;
+	}
+
+	if (swap_func == SWAP_WORDS_64)
+		swap_words_64(a, b, size);
+	else if (swap_func == SWAP_WORDS_32)
+		swap_words_32(a, b, size);
+	else if (swap_func == SWAP_BYTES)
+		swap_bytes(a, b, size);
+	else
+		swap_func(a, b, (int)size, priv);
+}
+
+#define _CMP_WRAPPER ((cmp_r_func_t)0L)
+
+static int do_cmp(const void *a, const void *b, cmp_r_func_t cmp, const void *priv)
+{
+	if (cmp == _CMP_WRAPPER)
+		return ((const struct wrapper *)priv)->cmp(a, b);
+	return cmp(a, b, priv);
+}
+
+static inline int eytzinger0_do_cmp(void *base, size_t n, size_t size,
+			 cmp_r_func_t cmp_func, const void *priv,
+			 size_t l, size_t r)
+{
+	return do_cmp(base + inorder_to_eytzinger0(l, n) * size,
+		      base + inorder_to_eytzinger0(r, n) * size,
+		      cmp_func, priv);
+}
+
+static inline void eytzinger0_do_swap(void *base, size_t n, size_t size,
+			   swap_r_func_t swap_func, const void *priv,
+			   size_t l, size_t r)
+{
+	do_swap(base + inorder_to_eytzinger0(l, n) * size,
+		base + inorder_to_eytzinger0(r, n) * size,
+		size, swap_func, priv);
+}
+
+void eytzinger0_sort_r(void *base, size_t n, size_t size,
+		       cmp_r_func_t cmp_func,
+		       swap_r_func_t swap_func,
+		       const void *priv)
+{
+	int i, j, k;
+
+	/* called from 'sort' without swap function, let's pick the default */
+	if (swap_func == SWAP_WRAPPER && !((struct wrapper *)priv)->swap_func)
+		swap_func = NULL;
+
+	if (!swap_func) {
+		if (is_aligned(base, size, 8))
+			swap_func = SWAP_WORDS_64;
+		else if (is_aligned(base, size, 4))
+			swap_func = SWAP_WORDS_32;
+		else
+			swap_func = SWAP_BYTES;
+	}
+
+	/* heapify */
+	for (i = n / 2 - 1; i >= 0; --i) {
+		/* Find the sift-down path all the way to the leaves. */
+		for (j = i; k = j * 2 + 1, k + 1 < n;)
+			j = eytzinger0_do_cmp(base, n, size, cmp_func, priv, k, k + 1) > 0 ? k : k + 1;
+
+		/* Special case for the last leaf with no sibling. */
+		if (j * 2 + 2 == n)
+			j = j * 2 + 1;
+
+		/* Backtrack to the correct location. */
+		while (j != i && eytzinger0_do_cmp(base, n, size, cmp_func, priv, i, j) >= 0)
+			j = (j - 1) / 2;
+
+		/* Shift the element into its correct place. */
+		for (k = j; j != i;) {
+			j = (j - 1) / 2;
+			eytzinger0_do_swap(base, n, size, swap_func, priv, j, k);
+		}
+	}
+
+	/* sort */
+	for (i = n - 1; i > 0; --i) {
+		eytzinger0_do_swap(base, n, size, swap_func, priv, 0, i);
+
+		/* Find the sift-down path all the way to the leaves. */
+		for (j = 0; k = j * 2 + 1, k + 1 < i;)
+			j = eytzinger0_do_cmp(base, n, size, cmp_func, priv, k, k + 1) > 0 ? k : k + 1;
+
+		/* Special case for the last leaf with no sibling. */
+		if (j * 2 + 2 == i)
+			j = j * 2 + 1;
+
+		/* Backtrack to the correct location. */
+		while (j && eytzinger0_do_cmp(base, n, size, cmp_func, priv, 0, j) >= 0)
+			j = (j - 1) / 2;
+
+		/* Shift the element into its correct place. */
+		for (k = j; j;) {
+			j = (j - 1) / 2;
+			eytzinger0_do_swap(base, n, size, swap_func, priv, j, k);
+		}
+	}
+}
+
+void eytzinger0_sort(void *base, size_t n, size_t size,
+		     cmp_func_t cmp_func,
+		     swap_func_t swap_func)
+{
+	struct wrapper w = {
+		.cmp  = cmp_func,
+		.swap_func = swap_func,
+	};
+
+	return eytzinger0_sort_r(base, n, size, _CMP_WRAPPER, SWAP_WRAPPER, &w);
+}
+
+#if 0
+#include <linux/slab.h>
+#include <linux/random.h>
+#include <linux/ktime.h>
+
+static u64 cmp_count;
+
+static int mycmp(const void *a, const void *b)
+{
+	u32 _a = *(u32 *)a;
+	u32 _b = *(u32 *)b;
+
+	cmp_count++;
+	if (_a < _b)
+		return -1;
+	else if (_a > _b)
+		return 1;
+	else
+		return 0;
+}
+
+static int test(void)
+{
+	size_t N, i;
+	ktime_t start, end;
+	s64 delta;
+	u32 *arr;
+
+	for (N = 10000; N <= 100000; N += 10000) {
+		arr = kmalloc_array(N, sizeof(u32), GFP_KERNEL);
+		cmp_count = 0;
+
+		for (i = 0; i < N; i++)
+			arr[i] = get_random_u32();
+
+		start = ktime_get();
+		eytzinger0_sort(arr, N, sizeof(u32), mycmp, NULL);
+		end = ktime_get();
+
+		delta = ktime_us_delta(end, start);
+		printk(KERN_INFO "time: %lld\n", delta);
+		printk(KERN_INFO "comparisons: %lld\n", cmp_count);
+
+		u32 prev = 0;
+
+		eytzinger0_for_each(i, N) {
+			if (prev > arr[i])
+				goto err;
+			prev = arr[i];
+		}
+
+		kfree(arr);
+	}
+	return 0;
+
+err:
+	kfree(arr);
+	return -1;
+}
+#endif
diff --git a/fs/bcachefs/eytzinger.h b/fs/bcachefs/eytzinger.h
index b04750dbf8..795f4fc0ba 100644
--- a/fs/bcachefs/eytzinger.h
+++ b/fs/bcachefs/eytzinger.h
@@ -5,23 +5,33 @@
 #include <linux/bitops.h>
 #include <linux/log2.h>
 
-#include "util.h"
+#ifdef EYTZINGER_DEBUG
+#define EYTZINGER_BUG_ON(cond)		BUG_ON(cond)
+#else
+#define EYTZINGER_BUG_ON(cond)
+#endif
 
 /*
  * Traversal for trees in eytzinger layout - a full binary tree layed out in an
- * array
- */
-
-/*
- * One based indexing version:
+ * array.
+ *
+ * Consider using an eytzinger tree any time you would otherwise be doing binary
+ * search over an array. Binary search is a worst case scenario for branch
+ * prediction and prefetching, but in an eytzinger tree every node's children
+ * are adjacent in memory, thus we can prefetch children before knowing the
+ * result of the comparison, assuming multiple nodes fit on a cacheline.
  *
- * With one based indexing each level of the tree starts at a power of two -
- * good for cacheline alignment:
+ * Two variants are provided, for one based indexing and zero based indexing.
+ *
+ * Zero based indexing is more convenient, but one based indexing has better
+ * alignment and thus better performance because each new level of the tree
+ * starts at a power of two, and thus if element 0 was cacheline aligned, each
+ * new level will be as well.
  */
 
 static inline unsigned eytzinger1_child(unsigned i, unsigned child)
 {
-	EBUG_ON(child > 1);
+	EYTZINGER_BUG_ON(child > 1);
 
 	return (i << 1) + child;
 }
@@ -38,7 +48,7 @@ static inline unsigned eytzinger1_right_child(unsigned i)
 
 static inline unsigned eytzinger1_first(unsigned size)
 {
-	return rounddown_pow_of_two(size);
+	return size ? rounddown_pow_of_two(size) : 0;
 }
 
 static inline unsigned eytzinger1_last(unsigned size)
@@ -58,7 +68,7 @@ static inline unsigned eytzinger1_last(unsigned size)
 
 static inline unsigned eytzinger1_next(unsigned i, unsigned size)
 {
-	EBUG_ON(i > size);
+	EYTZINGER_BUG_ON(i > size);
 
 	if (eytzinger1_right_child(i) <= size) {
 		i = eytzinger1_right_child(i);
@@ -74,7 +84,7 @@ static inline unsigned eytzinger1_next(unsigned i, unsigned size)
 
 static inline unsigned eytzinger1_prev(unsigned i, unsigned size)
 {
-	EBUG_ON(i > size);
+	EYTZINGER_BUG_ON(i > size);
 
 	if (eytzinger1_left_child(i) <= size) {
 		i = eytzinger1_left_child(i) + 1;
@@ -91,7 +101,9 @@ static inline unsigned eytzinger1_prev(unsigned i, unsigned size)
 
 static inline unsigned eytzinger1_extra(unsigned size)
 {
-	return (size + 1 - rounddown_pow_of_two(size)) << 1;
+	return size
+		? (size + 1 - rounddown_pow_of_two(size)) << 1
+		: 0;
 }
 
 static inline unsigned __eytzinger1_to_inorder(unsigned i, unsigned size,
@@ -101,7 +113,7 @@ static inline unsigned __eytzinger1_to_inorder(unsigned i, unsigned size,
 	unsigned shift = __fls(size) - b;
 	int s;
 
-	EBUG_ON(!i || i > size);
+	EYTZINGER_BUG_ON(!i || i > size);
 
 	i  ^= 1U << b;
 	i <<= 1;
@@ -126,7 +138,7 @@ static inline unsigned __inorder_to_eytzinger1(unsigned i, unsigned size,
 	unsigned shift;
 	int s;
 
-	EBUG_ON(!i || i > size);
+	EYTZINGER_BUG_ON(!i || i > size);
 
 	/*
 	 * sign bit trick:
@@ -164,7 +176,7 @@ static inline unsigned inorder_to_eytzinger1(unsigned i, unsigned size)
 
 static inline unsigned eytzinger0_child(unsigned i, unsigned child)
 {
-	EBUG_ON(child > 1);
+	EYTZINGER_BUG_ON(child > 1);
 
 	return (i << 1) + 1 + child;
 }
@@ -231,11 +243,9 @@ static inline unsigned inorder_to_eytzinger0(unsigned i, unsigned size)
 	     (_i) != -1;				\
 	     (_i) = eytzinger0_next((_i), (_size)))
 
-typedef int (*eytzinger_cmp_fn)(const void *l, const void *r, size_t size);
-
 /* return greatest node <= @search, or -1 if not found */
-static inline ssize_t eytzinger0_find_le(void *base, size_t nr, size_t size,
-					 eytzinger_cmp_fn cmp, const void *search)
+static inline int eytzinger0_find_le(void *base, size_t nr, size_t size,
+				     cmp_func_t cmp, const void *search)
 {
 	unsigned i, n = 0;
 
@@ -244,21 +254,38 @@ static inline ssize_t eytzinger0_find_le(void *base, size_t nr, size_t size,
 
 	do {
 		i = n;
-		n = eytzinger0_child(i, cmp(search, base + i * size, size) >= 0);
+		n = eytzinger0_child(i, cmp(base + i * size, search) <= 0);
 	} while (n < nr);
 
 	if (n & 1) {
-		/* @i was greater than @search, return previous node: */
-
-		if (i == eytzinger0_first(nr))
-			return -1;
-
+		/*
+		 * @i was greater than @search, return previous node:
+		 *
+		 * if @i was leftmost/smallest element,
+		 * eytzinger0_prev(eytzinger0_first())) returns -1, as expected
+		 */
 		return eytzinger0_prev(i, nr);
 	} else {
 		return i;
 	}
 }
 
+static inline int eytzinger0_find_gt(void *base, size_t nr, size_t size,
+				     cmp_func_t cmp, const void *search)
+{
+	ssize_t idx = eytzinger0_find_le(base, nr, size, cmp, search);
+
+	/*
+	 * if eytitzinger0_find_le() returned -1 - no element was <= search - we
+	 * want to return the first element; next/prev identities mean this work
+	 * as expected
+	 *
+	 * similarly if find_le() returns last element, we should return -1;
+	 * identities mean this all works out:
+	 */
+	return eytzinger0_next(idx, nr);
+}
+
 #define eytzinger0_find(base, nr, size, _cmp, search)			\
 ({									\
 	void *_base		= (base);				\
@@ -269,13 +296,13 @@ static inline ssize_t eytzinger0_find_le(void *base, size_t nr, size_t size,
 	int _res;							\
 									\
 	while (_i < _nr &&						\
-	       (_res = _cmp(_search, _base + _i * _size, _size)))	\
+	       (_res = _cmp(_search, _base + _i * _size)))		\
 		_i = eytzinger0_child(_i, _res > 0);			\
 	_i;								\
 })
 
-void eytzinger0_sort(void *, size_t, size_t,
-		    int (*cmp_func)(const void *, const void *, size_t),
-		    void (*swap_func)(void *, void *, size_t));
+void eytzinger0_sort_r(void *, size_t, size_t,
+		       cmp_r_func_t, swap_r_func_t, const void *);
+void eytzinger0_sort(void *, size_t, size_t, cmp_func_t, swap_func_t);
 
 #endif /* _EYTZINGER_H */
diff --git a/fs/bcachefs/fifo.h b/fs/bcachefs/fifo.h
index 66b945be10..d8153fe270 100644
--- a/fs/bcachefs/fifo.h
+++ b/fs/bcachefs/fifo.h
@@ -24,12 +24,12 @@ struct {								\
 	(fifo)->mask	= (fifo)->size					\
 		? roundup_pow_of_two((fifo)->size) - 1			\
 		: 0;							\
-	(fifo)->data	= kvpmalloc(fifo_buf_size(fifo), (_gfp));	\
+	(fifo)->data	= kvmalloc(fifo_buf_size(fifo), (_gfp));	\
 })
 
 #define free_fifo(fifo)							\
 do {									\
-	kvpfree((fifo)->data, fifo_buf_size(fifo));			\
+	kvfree((fifo)->data);						\
 	(fifo)->data = NULL;						\
 } while (0)
 
diff --git a/fs/bcachefs/fs-common.c b/fs/bcachefs/fs-common.c
index 1c1ea0f0c6..508d029ac5 100644
--- a/fs/bcachefs/fs-common.c
+++ b/fs/bcachefs/fs-common.c
@@ -42,7 +42,7 @@ int bch2_create_trans(struct btree_trans *trans,
 	if (ret)
 		goto err;
 
-	ret = bch2_inode_peek(trans, &dir_iter, dir_u, dir, BTREE_ITER_INTENT);
+	ret = bch2_inode_peek(trans, &dir_iter, dir_u, dir, BTREE_ITER_intent);
 	if (ret)
 		goto err;
 
@@ -70,7 +70,7 @@ int bch2_create_trans(struct btree_trans *trans,
 			struct bch_subvolume s;
 
 			ret = bch2_subvolume_get(trans, snapshot_src.subvol, true,
-						 BTREE_ITER_CACHED, &s);
+						 BTREE_ITER_cached, &s);
 			if (ret)
 				goto err;
 
@@ -78,7 +78,7 @@ int bch2_create_trans(struct btree_trans *trans,
 		}
 
 		ret = bch2_inode_peek(trans, &inode_iter, new_inode, snapshot_src,
-				      BTREE_ITER_INTENT);
+				      BTREE_ITER_intent);
 		if (ret)
 			goto err;
 
@@ -107,6 +107,7 @@ int bch2_create_trans(struct btree_trans *trans,
 		u32 new_subvol, dir_snapshot;
 
 		ret = bch2_subvolume_create(trans, new_inode->bi_inum,
+					    dir.subvol,
 					    snapshot_src.subvol,
 					    &new_subvol, &snapshot,
 					    (flags & BCH_CREATE_SNAPSHOT_RO) != 0);
@@ -162,7 +163,7 @@ int bch2_create_trans(struct btree_trans *trans,
 					 name,
 					 dir_target,
 					 &dir_offset,
-					 BCH_HASH_SET_MUST_CREATE);
+					 STR_HASH_must_create);
 		if (ret)
 			goto err;
 
@@ -170,7 +171,7 @@ int bch2_create_trans(struct btree_trans *trans,
 		new_inode->bi_dir_offset	= dir_offset;
 	}
 
-	inode_iter.flags &= ~BTREE_ITER_ALL_SNAPSHOTS;
+	inode_iter.flags &= ~BTREE_ITER_all_snapshots;
 	bch2_btree_iter_set_snapshot(&inode_iter, snapshot);
 
 	ret   = bch2_btree_iter_traverse(&inode_iter) ?:
@@ -197,16 +198,16 @@ int bch2_link_trans(struct btree_trans *trans,
 	if (dir.subvol != inum.subvol)
 		return -EXDEV;
 
-	ret = bch2_inode_peek(trans, &inode_iter, inode_u, inum, BTREE_ITER_INTENT);
+	ret = bch2_inode_peek(trans, &inode_iter, inode_u, inum, BTREE_ITER_intent);
 	if (ret)
-		goto err;
+		return ret;
 
 	inode_u->bi_ctime = now;
 	ret = bch2_inode_nlink_inc(inode_u);
 	if (ret)
-		return ret;
+		goto err;
 
-	ret = bch2_inode_peek(trans, &dir_iter, dir_u, dir, BTREE_ITER_INTENT);
+	ret = bch2_inode_peek(trans, &dir_iter, dir_u, dir, BTREE_ITER_intent);
 	if (ret)
 		goto err;
 
@@ -222,7 +223,7 @@ int bch2_link_trans(struct btree_trans *trans,
 	ret = bch2_dirent_create(trans, dir, &dir_hash,
 				 mode_to_type(inode_u->bi_mode),
 				 name, inum.inum, &dir_offset,
-				 BCH_HASH_SET_MUST_CREATE);
+				 STR_HASH_must_create);
 	if (ret)
 		goto err;
 
@@ -242,7 +243,7 @@ int bch2_unlink_trans(struct btree_trans *trans,
 		      struct bch_inode_unpacked *dir_u,
 		      struct bch_inode_unpacked *inode_u,
 		      const struct qstr *name,
-		      bool deleting_snapshot)
+		      bool deleting_subvol)
 {
 	struct bch_fs *c = trans->c;
 	struct btree_iter dir_iter = { NULL };
@@ -254,34 +255,41 @@ int bch2_unlink_trans(struct btree_trans *trans,
 	struct bkey_s_c k;
 	int ret;
 
-	ret = bch2_inode_peek(trans, &dir_iter, dir_u, dir, BTREE_ITER_INTENT);
+	ret = bch2_inode_peek(trans, &dir_iter, dir_u, dir, BTREE_ITER_intent);
 	if (ret)
 		goto err;
 
 	dir_hash = bch2_hash_info_init(c, dir_u);
 
-	ret = __bch2_dirent_lookup_trans(trans, &dirent_iter, dir, &dir_hash,
-					 name, &inum, BTREE_ITER_INTENT);
+	ret = bch2_dirent_lookup_trans(trans, &dirent_iter, dir, &dir_hash,
+				       name, &inum, BTREE_ITER_intent);
 	if (ret)
 		goto err;
 
 	ret = bch2_inode_peek(trans, &inode_iter, inode_u, inum,
-			      BTREE_ITER_INTENT);
+			      BTREE_ITER_intent);
 	if (ret)
 		goto err;
 
-	if (!deleting_snapshot && S_ISDIR(inode_u->bi_mode)) {
+	if (!deleting_subvol && S_ISDIR(inode_u->bi_mode)) {
 		ret = bch2_empty_dir_trans(trans, inum);
 		if (ret)
 			goto err;
 	}
 
-	if (deleting_snapshot && !inode_u->bi_subvol) {
+	if (deleting_subvol && !inode_u->bi_subvol) {
 		ret = -BCH_ERR_ENOENT_not_subvol;
 		goto err;
 	}
 
-	if (deleting_snapshot || inode_u->bi_subvol) {
+	if (inode_u->bi_subvol) {
+		/* Recursive subvolume destroy not allowed (yet?) */
+		ret = bch2_subvol_has_children(trans, inode_u->bi_subvol);
+		if (ret)
+			goto err;
+	}
+
+	if (deleting_subvol || inode_u->bi_subvol) {
 		ret = bch2_subvolume_unlink(trans, inode_u->bi_subvol);
 		if (ret)
 			goto err;
@@ -314,7 +322,7 @@ int bch2_unlink_trans(struct btree_trans *trans,
 
 	ret =   bch2_hash_delete_at(trans, bch2_dirent_hash_desc,
 				    &dir_hash, &dirent_iter,
-				    BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE) ?:
+				    BTREE_UPDATE_internal_snapshot_node) ?:
 		bch2_inode_write(trans, &dir_iter, dir_u) ?:
 		bch2_inode_write(trans, &inode_iter, inode_u);
 err:
@@ -349,6 +357,22 @@ bool bch2_reinherit_attrs(struct bch_inode_unpacked *dst_u,
 	return ret;
 }
 
+static int subvol_update_parent(struct btree_trans *trans, u32 subvol, u32 new_parent)
+{
+	struct btree_iter iter;
+	struct bkey_i_subvolume *s =
+		bch2_bkey_get_mut_typed(trans, &iter,
+			BTREE_ID_subvolumes, POS(0, subvol),
+			BTREE_ITER_cached, subvolume);
+	int ret = PTR_ERR_OR_ZERO(s);
+	if (ret)
+		return ret;
+
+	s->v.fs_path_parent = cpu_to_le32(new_parent);
+	bch2_trans_iter_exit(trans, &iter);
+	return 0;
+}
+
 int bch2_rename_trans(struct btree_trans *trans,
 		      subvol_inum src_dir, struct bch_inode_unpacked *src_dir_u,
 		      subvol_inum dst_dir, struct bch_inode_unpacked *dst_dir_u,
@@ -370,7 +394,7 @@ int bch2_rename_trans(struct btree_trans *trans,
 	int ret;
 
 	ret = bch2_inode_peek(trans, &src_dir_iter, src_dir_u, src_dir,
-			      BTREE_ITER_INTENT);
+			      BTREE_ITER_intent);
 	if (ret)
 		goto err;
 
@@ -379,7 +403,7 @@ int bch2_rename_trans(struct btree_trans *trans,
 	if (dst_dir.inum	!= src_dir.inum ||
 	    dst_dir.subvol	!= src_dir.subvol) {
 		ret = bch2_inode_peek(trans, &dst_dir_iter, dst_dir_u, dst_dir,
-				      BTREE_ITER_INTENT);
+				      BTREE_ITER_intent);
 		if (ret)
 			goto err;
 
@@ -399,17 +423,47 @@ int bch2_rename_trans(struct btree_trans *trans,
 		goto err;
 
 	ret = bch2_inode_peek(trans, &src_inode_iter, src_inode_u, src_inum,
-			      BTREE_ITER_INTENT);
+			      BTREE_ITER_intent);
 	if (ret)
 		goto err;
 
 	if (dst_inum.inum) {
 		ret = bch2_inode_peek(trans, &dst_inode_iter, dst_inode_u, dst_inum,
-				      BTREE_ITER_INTENT);
+				      BTREE_ITER_intent);
+		if (ret)
+			goto err;
+	}
+
+	if (src_inode_u->bi_subvol &&
+	    dst_dir.subvol != src_inode_u->bi_parent_subvol) {
+		ret = subvol_update_parent(trans, src_inode_u->bi_subvol, dst_dir.subvol);
+		if (ret)
+			goto err;
+	}
+
+	if (mode == BCH_RENAME_EXCHANGE &&
+	    dst_inode_u->bi_subvol &&
+	    src_dir.subvol != dst_inode_u->bi_parent_subvol) {
+		ret = subvol_update_parent(trans, dst_inode_u->bi_subvol, src_dir.subvol);
 		if (ret)
 			goto err;
 	}
 
+	/* Can't move across subvolumes, unless it's a subvolume root: */
+	if (src_dir.subvol != dst_dir.subvol &&
+	    (!src_inode_u->bi_subvol ||
+	     (dst_inum.inum && !dst_inode_u->bi_subvol))) {
+		ret = -EXDEV;
+		goto err;
+	}
+
+	if (src_inode_u->bi_parent_subvol)
+		src_inode_u->bi_parent_subvol = dst_dir.subvol;
+
+	if ((mode == BCH_RENAME_EXCHANGE) &&
+	    dst_inode_u->bi_parent_subvol)
+		dst_inode_u->bi_parent_subvol = src_dir.subvol;
+
 	src_inode_u->bi_dir		= dst_dir_u->bi_inum;
 	src_inode_u->bi_dir_offset	= dst_offset;
 
@@ -432,10 +486,10 @@ int bch2_rename_trans(struct btree_trans *trans,
 			goto err;
 		}
 
-		if (S_ISDIR(dst_inode_u->bi_mode) &&
-		    bch2_empty_dir_trans(trans, dst_inum)) {
-			ret = -ENOTEMPTY;
-			goto err;
+		if (S_ISDIR(dst_inode_u->bi_mode)) {
+			ret = bch2_empty_dir_trans(trans, dst_inum);
+			if (ret)
+				goto err;
 		}
 	}
 
diff --git a/fs/bcachefs/fs-io-buffered.c b/fs/bcachefs/fs-io-buffered.c
index 27710cdd57..54873ecc63 100644
--- a/fs/bcachefs/fs-io-buffered.c
+++ b/fs/bcachefs/fs-io-buffered.c
@@ -30,15 +30,8 @@ static void bch2_readpages_end_io(struct bio *bio)
 {
 	struct folio_iter fi;
 
-	bio_for_each_folio_all(fi, bio) {
-		if (!bio->bi_status) {
-			folio_mark_uptodate(fi.folio);
-		} else {
-			folio_clear_uptodate(fi.folio);
-			folio_set_error(fi.folio);
-		}
-		folio_unlock(fi.folio);
-	}
+	bio_for_each_folio_all(fi, bio)
+		folio_end_read(fi.folio, bio->bi_status == BLK_STS_OK);
 
 	bio_put(bio);
 }
@@ -176,7 +169,7 @@ retry:
 
 	bch2_trans_iter_init(trans, &iter, BTREE_ID_extents,
 			     SPOS(inum.inum, rbio->bio.bi_iter.bi_sector, snapshot),
-			     BTREE_ITER_SLOTS);
+			     BTREE_ITER_slots);
 	while (1) {
 		struct bkey_s_c k;
 		unsigned bytes, sectors, offset_into_extent;
@@ -264,7 +257,6 @@ void bch2_readahead(struct readahead_control *ractl)
 	struct bch_inode_info *inode = to_bch_ei(ractl->mapping->host);
 	struct bch_fs *c = inode->v.i_sb->s_fs_info;
 	struct bch_io_opts opts;
-	struct btree_trans *trans = bch2_trans_get(c);
 	struct folio *folio;
 	struct readpages_iter readpages_iter;
 
@@ -276,6 +268,7 @@ void bch2_readahead(struct readahead_control *ractl)
 
 	bch2_pagecache_add_get(inode);
 
+	struct btree_trans *trans = bch2_trans_get(c);
 	while ((folio = readpage_iter_peek(&readpages_iter))) {
 		unsigned n = min_t(unsigned,
 				   readpages_iter.folios.nr -
@@ -296,10 +289,10 @@ void bch2_readahead(struct readahead_control *ractl)
 			   &readpages_iter);
 		bch2_trans_unlock(trans);
 	}
+	bch2_trans_put(trans);
 
 	bch2_pagecache_add_put(inode);
 
-	bch2_trans_put(trans);
 	darray_exit(&readpages_iter.folios);
 }
 
@@ -408,7 +401,6 @@ static void bch2_writepage_io_done(struct bch_write_op *op)
 		bio_for_each_folio_all(fi, bio) {
 			struct bch_folio *s;
 
-			folio_set_error(fi.folio);
 			mapping_set_error(fi.folio->mapping, -EIO);
 
 			s = __bch2_folio(fi.folio);
@@ -445,8 +437,8 @@ static void bch2_writepage_io_done(struct bch_write_op *op)
 	 */
 
 	/*
-	 * PageWriteback is effectively our ref on the inode - fixup i_blocks
-	 * before calling end_page_writeback:
+	 * The writeback flag is effectively our ref on the inode -
+	 * fixup i_blocks before calling folio_end_writeback:
 	 */
 	bch2_i_sectors_acct(c, io->inode, NULL, io->op.i_sectors_delta);
 
@@ -810,7 +802,8 @@ static noinline void folios_trunc(folios *fs, struct folio **fi)
 static int __bch2_buffered_write(struct bch_inode_info *inode,
 				 struct address_space *mapping,
 				 struct iov_iter *iter,
-				 loff_t pos, unsigned len)
+				 loff_t pos, unsigned len,
+				 bool inode_locked)
 {
 	struct bch_fs *c = inode->v.i_sb->s_fs_info;
 	struct bch2_folio_reservation res;
@@ -835,6 +828,15 @@ static int __bch2_buffered_write(struct bch_inode_info *inode,
 
 	BUG_ON(!fs.nr);
 
+	/*
+	 * If we're not using the inode lock, we need to lock all the folios for
+	 * atomiticity of writes vs. other writes:
+	 */
+	if (!inode_locked && folio_end_pos(darray_last(fs)) < end) {
+		ret = -BCH_ERR_need_inode_lock;
+		goto out;
+	}
+
 	f = darray_first(fs);
 	if (pos != folio_pos(f) && !folio_test_uptodate(f)) {
 		ret = bch2_read_single_folio(f, mapping);
@@ -896,7 +898,7 @@ static int __bch2_buffered_write(struct bch_inode_info *inode,
 	darray_for_each(fs, fi) {
 		f = *fi;
 		f_len = min(end, folio_end_pos(f)) - f_pos;
-		f_copied = copy_page_from_iter_atomic(&f->page, f_offset, f_len, iter);
+		f_copied = copy_folio_from_iter_atomic(f, f_offset, f_len, iter);
 		if (!f_copied) {
 			folios_trunc(&fs, fi);
 			break;
@@ -929,8 +931,10 @@ static int __bch2_buffered_write(struct bch_inode_info *inode,
 	end = pos + copied;
 
 	spin_lock(&inode->v.i_lock);
-	if (end > inode->v.i_size)
+	if (end > inode->v.i_size) {
+		BUG_ON(!inode_locked);
 		i_size_write(&inode->v, end);
+	}
 	spin_unlock(&inode->v.i_lock);
 
 	f_pos = pos;
@@ -974,12 +978,68 @@ static ssize_t bch2_buffered_write(struct kiocb *iocb, struct iov_iter *iter)
 	struct file *file = iocb->ki_filp;
 	struct address_space *mapping = file->f_mapping;
 	struct bch_inode_info *inode = file_bch_inode(file);
-	loff_t pos = iocb->ki_pos;
-	ssize_t written = 0;
-	int ret = 0;
+	loff_t pos;
+	bool inode_locked = false;
+	ssize_t written = 0, written2 = 0, ret = 0;
+
+	/*
+	 * We don't take the inode lock unless i_size will be changing. Folio
+	 * locks provide exclusion with other writes, and the pagecache add lock
+	 * provides exclusion with truncate and hole punching.
+	 *
+	 * There is one nasty corner case where atomicity would be broken
+	 * without great care: when copying data from userspace to the page
+	 * cache, we do that with faults disable - a page fault would recurse
+	 * back into the filesystem, taking filesystem locks again, and
+	 * deadlock; so it's done with faults disabled, and we fault in the user
+	 * buffer when we aren't holding locks.
+	 *
+	 * If we do part of the write, but we then race and in the userspace
+	 * buffer have been evicted and are no longer resident, then we have to
+	 * drop our folio locks to re-fault them in, breaking write atomicity.
+	 *
+	 * To fix this, we restart the write from the start, if we weren't
+	 * holding the inode lock.
+	 *
+	 * There is another wrinkle after that; if we restart the write from the
+	 * start, and then get an unrecoverable error, we _cannot_ claim to
+	 * userspace that we did not write data we actually did - so we must
+	 * track (written2) the most we ever wrote.
+	 */
+
+	if ((iocb->ki_flags & IOCB_APPEND) ||
+	    (iocb->ki_pos + iov_iter_count(iter) > i_size_read(&inode->v))) {
+		inode_lock(&inode->v);
+		inode_locked = true;
+	}
+
+	ret = generic_write_checks(iocb, iter);
+	if (ret <= 0)
+		goto unlock;
+
+	ret = file_remove_privs_flags(file, !inode_locked ? IOCB_NOWAIT : 0);
+	if (ret) {
+		if (!inode_locked) {
+			inode_lock(&inode->v);
+			inode_locked = true;
+			ret = file_remove_privs_flags(file, 0);
+		}
+		if (ret)
+			goto unlock;
+	}
+
+	ret = file_update_time(file);
+	if (ret)
+		goto unlock;
+
+	pos = iocb->ki_pos;
 
 	bch2_pagecache_add_get(inode);
 
+	if (!inode_locked &&
+	    (iocb->ki_pos + iov_iter_count(iter) > i_size_read(&inode->v)))
+		goto get_inode_lock;
+
 	do {
 		unsigned offset = pos & (PAGE_SIZE - 1);
 		unsigned bytes = iov_iter_count(iter);
@@ -1004,12 +1064,17 @@ again:
 			}
 		}
 
+		if (unlikely(bytes != iov_iter_count(iter) && !inode_locked))
+			goto get_inode_lock;
+
 		if (unlikely(fatal_signal_pending(current))) {
 			ret = -EINTR;
 			break;
 		}
 
-		ret = __bch2_buffered_write(inode, mapping, iter, pos, bytes);
+		ret = __bch2_buffered_write(inode, mapping, iter, pos, bytes, inode_locked);
+		if (ret == -BCH_ERR_need_inode_lock)
+			goto get_inode_lock;
 		if (unlikely(ret < 0))
 			break;
 
@@ -1030,50 +1095,46 @@ again:
 		}
 		pos += ret;
 		written += ret;
+		written2 = max(written, written2);
+
+		if (ret != bytes && !inode_locked)
+			goto get_inode_lock;
 		ret = 0;
 
 		balance_dirty_pages_ratelimited(mapping);
-	} while (iov_iter_count(iter));
 
+		if (0) {
+get_inode_lock:
+			bch2_pagecache_add_put(inode);
+			inode_lock(&inode->v);
+			inode_locked = true;
+			bch2_pagecache_add_get(inode);
+
+			iov_iter_revert(iter, written);
+			pos -= written;
+			written = 0;
+			ret = 0;
+		}
+	} while (iov_iter_count(iter));
 	bch2_pagecache_add_put(inode);
+unlock:
+	if (inode_locked)
+		inode_unlock(&inode->v);
+
+	iocb->ki_pos += written;
 
-	return written ? written : ret;
+	ret = max(written, written2) ?: ret;
+	if (ret > 0)
+		ret = generic_write_sync(iocb, ret);
+	return ret;
 }
 
-ssize_t bch2_write_iter(struct kiocb *iocb, struct iov_iter *from)
+ssize_t bch2_write_iter(struct kiocb *iocb, struct iov_iter *iter)
 {
-	struct file *file = iocb->ki_filp;
-	struct bch_inode_info *inode = file_bch_inode(file);
-	ssize_t ret;
-
-	if (iocb->ki_flags & IOCB_DIRECT) {
-		ret = bch2_direct_write(iocb, from);
-		goto out;
-	}
+	ssize_t ret = iocb->ki_flags & IOCB_DIRECT
+		? bch2_direct_write(iocb, iter)
+		: bch2_buffered_write(iocb, iter);
 
-	inode_lock(&inode->v);
-
-	ret = generic_write_checks(iocb, from);
-	if (ret <= 0)
-		goto unlock;
-
-	ret = file_remove_privs(file);
-	if (ret)
-		goto unlock;
-
-	ret = file_update_time(file);
-	if (ret)
-		goto unlock;
-
-	ret = bch2_buffered_write(iocb, from);
-	if (likely(ret > 0))
-		iocb->ki_pos += ret;
-unlock:
-	inode_unlock(&inode->v);
-
-	if (ret > 0)
-		ret = generic_write_sync(iocb, ret);
-out:
 	return bch2_err_class(ret);
 }
 
diff --git a/fs/bcachefs/fs-io-direct.c b/fs/bcachefs/fs-io-direct.c
index 33cb6da3a5..049b61bc9a 100644
--- a/fs/bcachefs/fs-io-direct.c
+++ b/fs/bcachefs/fs-io-direct.c
@@ -254,7 +254,7 @@ retry:
 
 	for_each_btree_key_norestart(trans, iter, BTREE_ID_extents,
 			   SPOS(inum.inum, offset, snapshot),
-			   BTREE_ITER_SLOTS, k, err) {
+			   BTREE_ITER_slots, k, err) {
 		if (bkey_ge(bkey_start_pos(k.k), POS(inum.inum, end)))
 			break;
 
@@ -387,6 +387,8 @@ static __always_inline long bch2_dio_write_done(struct dio_write *dio)
 	ret = dio->op.error ?: ((long) dio->written << 9);
 	bio_put(&dio->op.wbio.bio);
 
+	bch2_write_ref_put(dio->op.c, BCH_WRITE_REF_dio_write);
+
 	/* inode->i_dio_count is our ref on inode and thus bch_fs */
 	inode_dio_end(&inode->v);
 
@@ -536,7 +538,7 @@ static __always_inline long bch2_dio_write_loop(struct dio_write *dio)
 		if (likely(!dio->iter.count) || dio->op.error)
 			break;
 
-		bio_reset(bio, NULL, REQ_OP_WRITE);
+		bio_reset(bio, NULL, REQ_OP_WRITE | REQ_SYNC | REQ_IDLE);
 	}
 out:
 	return bch2_dio_write_done(dio);
@@ -590,22 +592,27 @@ ssize_t bch2_direct_write(struct kiocb *req, struct iov_iter *iter)
 	prefetch(&inode->ei_inode);
 	prefetch((void *) &inode->ei_inode + 64);
 
+	if (!bch2_write_ref_tryget(c, BCH_WRITE_REF_dio_write))
+		return -EROFS;
+
 	inode_lock(&inode->v);
 
 	ret = generic_write_checks(req, iter);
 	if (unlikely(ret <= 0))
-		goto err;
+		goto err_put_write_ref;
 
 	ret = file_remove_privs(file);
 	if (unlikely(ret))
-		goto err;
+		goto err_put_write_ref;
 
 	ret = file_update_time(file);
 	if (unlikely(ret))
-		goto err;
+		goto err_put_write_ref;
 
-	if (unlikely((req->ki_pos|iter->count) & (block_bytes(c) - 1)))
-		goto err;
+	if (unlikely((req->ki_pos|iter->count) & (block_bytes(c) - 1))) {
+		ret = -EINVAL;
+		goto err_put_write_ref;
+	}
 
 	inode_dio_begin(&inode->v);
 	bch2_pagecache_block_get(inode);
@@ -618,7 +625,7 @@ ssize_t bch2_direct_write(struct kiocb *req, struct iov_iter *iter)
 
 	bio = bio_alloc_bioset(NULL,
 			       bio_iov_vecs_to_alloc(iter, BIO_MAX_VECS),
-			       REQ_OP_WRITE,
+			       REQ_OP_WRITE | REQ_SYNC | REQ_IDLE,
 			       GFP_KERNEL,
 			       &c->dio_write_bioset);
 	dio = container_of(bio, struct dio_write, op.wbio.bio);
@@ -645,7 +652,7 @@ ssize_t bch2_direct_write(struct kiocb *req, struct iov_iter *iter)
 	}
 
 	ret = bch2_dio_write_loop(dio);
-err:
+out:
 	if (locked)
 		inode_unlock(&inode->v);
 	return ret;
@@ -653,7 +660,9 @@ err_put_bio:
 	bch2_pagecache_block_put(inode);
 	bio_put(bio);
 	inode_dio_end(&inode->v);
-	goto err;
+err_put_write_ref:
+	bch2_write_ref_put(c, BCH_WRITE_REF_dio_write);
+	goto out;
 }
 
 void bch2_fs_fs_io_direct_exit(struct bch_fs *c)
diff --git a/fs/bcachefs/fs-io-pagecache.c b/fs/bcachefs/fs-io-pagecache.c
index d359aa9b33..872283e5bd 100644
--- a/fs/bcachefs/fs-io-pagecache.c
+++ b/fs/bcachefs/fs-io-pagecache.c
@@ -214,7 +214,7 @@ retry:
 
 	for_each_btree_key_norestart(trans, iter, BTREE_ID_extents,
 			   SPOS(inum.inum, offset, snapshot),
-			   BTREE_ITER_SLOTS, k, ret) {
+			   BTREE_ITER_slots, k, ret) {
 		unsigned nr_ptrs = bch2_bkey_nr_ptrs_fully_allocated(k);
 		unsigned state = bkey_to_sector_state(k);
 
diff --git a/fs/bcachefs/fs-io-pagecache.h b/fs/bcachefs/fs-io-pagecache.h
index 8cbaba6565..828c3d7c8f 100644
--- a/fs/bcachefs/fs-io-pagecache.h
+++ b/fs/bcachefs/fs-io-pagecache.h
@@ -51,13 +51,10 @@ enum bch_folio_sector_state {
 
 struct bch_folio_sector {
 	/* Uncompressed, fully allocated replicas (or on disk reservation): */
-	unsigned		nr_replicas:4;
-
+	u8			nr_replicas:4,
 	/* Owns PAGE_SECTORS * replicas_reserved sized in memory reservation: */
-	unsigned		replicas_reserved:4;
-
-	/* i_sectors: */
-	enum bch_folio_sector_state state:8;
+				replicas_reserved:4;
+	u8			state;
 };
 
 struct bch_folio {
diff --git a/fs/bcachefs/fs-io.c b/fs/bcachefs/fs-io.c
index 8c70123b6a..ef20b64033 100644
--- a/fs/bcachefs/fs-io.c
+++ b/fs/bcachefs/fs-io.c
@@ -174,18 +174,18 @@ void __bch2_i_sectors_acct(struct bch_fs *c, struct bch_inode_info *inode,
 static int bch2_flush_inode(struct bch_fs *c,
 			    struct bch_inode_info *inode)
 {
-	struct bch_inode_unpacked u;
-	int ret;
-
 	if (c->opts.journal_flush_disabled)
 		return 0;
 
-	ret = bch2_inode_find_by_inum(c, inode_inum(inode), &u);
-	if (ret)
-		return ret;
+	if (!bch2_write_ref_tryget(c, BCH_WRITE_REF_fsync))
+		return -EROFS;
 
-	return bch2_journal_flush_seq(&c->journal, u.bi_journal_seq) ?:
-		bch2_inode_flush_nocow_writes(c, inode);
+	struct bch_inode_unpacked u;
+	int ret = bch2_inode_find_by_inum(c, inode_inum(inode), &u) ?:
+		  bch2_journal_flush_seq(&c->journal, u.bi_journal_seq) ?:
+		  bch2_inode_flush_nocow_writes(c, inode);
+	bch2_write_ref_put(c, BCH_WRITE_REF_fsync);
+	return ret;
 }
 
 int bch2_fsync(struct file *file, loff_t start, loff_t end, int datasync)
@@ -202,7 +202,10 @@ int bch2_fsync(struct file *file, loff_t start, loff_t end, int datasync)
 		goto out;
 	ret = bch2_flush_inode(c, inode);
 out:
-	return bch2_err_class(ret);
+	ret = bch2_err_class(ret);
+	if (ret == -EROFS)
+		ret = -EIO;
+	return ret;
 }
 
 /* truncate: */
@@ -594,7 +597,7 @@ static int __bchfs_fallocate(struct bch_inode_info *inode, int mode,
 
 	bch2_trans_iter_init(trans, &iter, BTREE_ID_extents,
 			POS(inode->v.i_ino, start_sector),
-			BTREE_ITER_SLOTS|BTREE_ITER_INTENT);
+			BTREE_ITER_slots|BTREE_ITER_intent);
 
 	while (!ret && bkey_lt(iter.pos, end_pos)) {
 		s64 i_sectors_delta = 0;
@@ -1009,7 +1012,7 @@ retry:
 
 	for_each_btree_key_norestart(trans, iter, BTREE_ID_extents,
 			   SPOS(inode->v.i_ino, offset >> 9, snapshot),
-			   BTREE_ITER_SLOTS, k, ret) {
+			   BTREE_ITER_slots, k, ret) {
 		if (k.k->p.inode != inode->v.i_ino) {
 			next_hole = bch2_seek_pagecache_hole(&inode->v,
 					offset, MAX_LFS_FILESIZE, 0, false);
diff --git a/fs/bcachefs/fs-ioctl.c b/fs/bcachefs/fs-ioctl.c
index 3dc8630ff9..79a0c8732b 100644
--- a/fs/bcachefs/fs-ioctl.c
+++ b/fs/bcachefs/fs-ioctl.c
@@ -308,8 +308,8 @@ static int bch2_ioc_goingdown(struct bch_fs *c, u32 __user *arg)
 	return ret;
 }
 
-static long __bch2_ioctl_subvolume_create(struct bch_fs *c, struct file *filp,
-					  struct bch_ioctl_subvolume arg)
+static long bch2_ioctl_subvolume_create(struct bch_fs *c, struct file *filp,
+					struct bch_ioctl_subvolume arg)
 {
 	struct inode *dir;
 	struct bch_inode_info *inode;
@@ -373,7 +373,7 @@ retry:
 	}
 
 	if (dst_dentry->d_inode) {
-		error = -EEXIST;
+		error = -BCH_ERR_EEXIST_subvolume_create;
 		goto err3;
 	}
 
@@ -406,9 +406,12 @@ retry:
 	    !arg.src_ptr)
 		snapshot_src.subvol = inode_inum(to_bch_ei(dir)).subvol;
 
+	down_write(&c->snapshot_create_lock);
 	inode = __bch2_create(file_mnt_idmap(filp), to_bch_ei(dir),
 			      dst_dentry, arg.mode|S_IFDIR,
 			      0, snapshot_src, create_flags);
+	up_write(&c->snapshot_create_lock);
+
 	error = PTR_ERR_OR_ZERO(inode);
 	if (error)
 		goto err3;
@@ -429,16 +432,6 @@ err1:
 	return error;
 }
 
-static long bch2_ioctl_subvolume_create(struct bch_fs *c, struct file *filp,
-					struct bch_ioctl_subvolume arg)
-{
-	down_write(&c->snapshot_create_lock);
-	long ret = __bch2_ioctl_subvolume_create(c, filp, arg);
-	up_write(&c->snapshot_create_lock);
-
-	return ret;
-}
-
 static long bch2_ioctl_subvolume_destroy(struct bch_fs *c, struct file *filp,
 				struct bch_ioctl_subvolume arg)
 {
@@ -548,7 +541,7 @@ long bch2_compat_fs_ioctl(struct file *file, unsigned cmd, unsigned long arg)
 {
 	/* These are just misnamed, they actually get/put from/to user an int */
 	switch (cmd) {
-	case FS_IOC_GETFLAGS:
+	case FS_IOC32_GETFLAGS:
 		cmd = FS_IOC_GETFLAGS;
 		break;
 	case FS_IOC32_SETFLAGS:
diff --git a/fs/bcachefs/fs.c b/fs/bcachefs/fs.c
index 77ae65542d..fa1fee05cf 100644
--- a/fs/bcachefs/fs.c
+++ b/fs/bcachefs/fs.c
@@ -90,7 +90,7 @@ retry:
 	bch2_trans_begin(trans);
 
 	ret   = bch2_inode_peek(trans, &iter, &inode_u, inode_inum(inode),
-				BTREE_ITER_INTENT) ?:
+				BTREE_ITER_intent) ?:
 		(set ? set(trans, inode, &inode_u, p) : 0) ?:
 		bch2_inode_write(trans, &iter, &inode_u) ?:
 		bch2_trans_commit(trans, NULL, NULL, BCH_TRANS_COMMIT_no_enospc);
@@ -108,7 +108,8 @@ retry:
 		goto retry;
 
 	bch2_fs_fatal_err_on(bch2_err_matches(ret, ENOENT), c,
-			     "inode %u:%llu not found when updating",
+			     "%s: inode %u:%llu not found when updating",
+			     bch2_err_str(ret),
 			     inode_inum(inode).subvol,
 			     inode_inum(inode).inum);
 
@@ -176,45 +177,130 @@ static unsigned bch2_inode_hash(subvol_inum inum)
 	return jhash_3words(inum.subvol, inum.inum >> 32, inum.inum, JHASH_INITVAL);
 }
 
-struct inode *bch2_vfs_inode_get(struct bch_fs *c, subvol_inum inum)
+static struct bch_inode_info *bch2_inode_insert(struct bch_fs *c, struct bch_inode_info *inode)
 {
-	struct bch_inode_unpacked inode_u;
-	struct bch_inode_info *inode;
-	struct btree_trans *trans;
-	struct bch_subvolume subvol;
-	int ret;
+	subvol_inum inum = inode_inum(inode);
+	struct bch_inode_info *old = to_bch_ei(inode_insert5(&inode->v,
+				      bch2_inode_hash(inum),
+				      bch2_iget5_test,
+				      bch2_iget5_set,
+				      &inum));
+	BUG_ON(!old);
 
-	inode = to_bch_ei(iget5_locked(c->vfs_sb,
-				       bch2_inode_hash(inum),
-				       bch2_iget5_test,
-				       bch2_iget5_set,
-				       &inum));
-	if (unlikely(!inode))
-		return ERR_PTR(-ENOMEM);
-	if (!(inode->v.i_state & I_NEW))
-		return &inode->v;
+	if (unlikely(old != inode)) {
+		/*
+		 * bcachefs doesn't use I_NEW; we have no use for it since we
+		 * only insert fully created inodes in the inode hash table. But
+		 * discard_new_inode() expects it to be set...
+		 */
+		inode->v.i_flags |= I_NEW;
+		/*
+		 * We don't want bch2_evict_inode() to delete the inode on disk,
+		 * we just raced and had another inode in cache. Normally new
+		 * inodes don't have nlink == 0 - except tmpfiles do...
+		 */
+		set_nlink(&inode->v, 1);
+		discard_new_inode(&inode->v);
+		inode = old;
+	} else {
+		mutex_lock(&c->vfs_inodes_lock);
+		list_add(&inode->ei_vfs_inode_list, &c->vfs_inodes_list);
+		mutex_unlock(&c->vfs_inodes_lock);
+		/*
+		 * Again, I_NEW makes no sense for bcachefs. This is only needed
+		 * for clearing I_NEW, but since the inode was already fully
+		 * created and initialized we didn't actually want
+		 * inode_insert5() to set it for us.
+		 */
+		unlock_new_inode(&inode->v);
+	}
 
-	trans = bch2_trans_get(c);
-	ret = lockrestart_do(trans,
-		bch2_subvolume_get(trans, inum.subvol, true, 0, &subvol) ?:
-		bch2_inode_find_by_inum_trans(trans, inum, &inode_u));
+	return inode;
+}
 
-	if (!ret)
-		bch2_vfs_inode_init(trans, inum, inode, &inode_u, &subvol);
-	bch2_trans_put(trans);
+#define memalloc_flags_do(_flags, _do)						\
+({										\
+	unsigned _saved_flags = memalloc_flags_save(_flags);			\
+	typeof(_do) _ret = _do;							\
+	memalloc_noreclaim_restore(_saved_flags);				\
+	_ret;									\
+})
 
-	if (ret) {
-		iget_failed(&inode->v);
-		return ERR_PTR(bch2_err_class(ret));
+static struct inode *bch2_alloc_inode(struct super_block *sb)
+{
+	BUG();
+}
+
+static struct bch_inode_info *__bch2_new_inode(struct bch_fs *c)
+{
+	struct bch_inode_info *inode = kmem_cache_alloc(bch2_inode_cache, GFP_NOFS);
+	if (!inode)
+		return NULL;
+
+	inode_init_once(&inode->v);
+	mutex_init(&inode->ei_update_lock);
+	two_state_lock_init(&inode->ei_pagecache_lock);
+	INIT_LIST_HEAD(&inode->ei_vfs_inode_list);
+	inode->ei_flags = 0;
+	mutex_init(&inode->ei_quota_lock);
+	memset(&inode->ei_devs_need_flush, 0, sizeof(inode->ei_devs_need_flush));
+	inode->v.i_state = 0;
+
+	if (unlikely(inode_init_always(c->vfs_sb, &inode->v))) {
+		kmem_cache_free(bch2_inode_cache, inode);
+		return NULL;
 	}
 
-	mutex_lock(&c->vfs_inodes_lock);
-	list_add(&inode->ei_vfs_inode_list, &c->vfs_inodes_list);
-	mutex_unlock(&c->vfs_inodes_lock);
+	return inode;
+}
+
+/*
+ * Allocate a new inode, dropping/retaking btree locks if necessary:
+ */
+static struct bch_inode_info *bch2_new_inode(struct btree_trans *trans)
+{
+	struct bch_inode_info *inode =
+		memalloc_flags_do(PF_MEMALLOC_NORECLAIM|PF_MEMALLOC_NOWARN,
+				  __bch2_new_inode(trans->c));
+
+	if (unlikely(!inode)) {
+		int ret = drop_locks_do(trans, (inode = __bch2_new_inode(trans->c)) ? 0 : -ENOMEM);
+		if (ret && inode) {
+			__destroy_inode(&inode->v);
+			kmem_cache_free(bch2_inode_cache, inode);
+		}
+		if (ret)
+			return ERR_PTR(ret);
+	}
+
+	return inode;
+}
+
+struct inode *bch2_vfs_inode_get(struct bch_fs *c, subvol_inum inum)
+{
+	struct bch_inode_info *inode =
+		to_bch_ei(ilookup5_nowait(c->vfs_sb,
+					  bch2_inode_hash(inum),
+					  bch2_iget5_test,
+					  &inum));
+	if (inode)
+		return &inode->v;
+
+	struct btree_trans *trans = bch2_trans_get(c);
 
-	unlock_new_inode(&inode->v);
+	struct bch_inode_unpacked inode_u;
+	struct bch_subvolume subvol;
+	int ret = lockrestart_do(trans,
+		bch2_subvolume_get(trans, inum.subvol, true, 0, &subvol) ?:
+		bch2_inode_find_by_inum_trans(trans, inum, &inode_u)) ?:
+		PTR_ERR_OR_ZERO(inode = bch2_new_inode(trans));
+	if (!ret) {
+		bch2_vfs_inode_init(trans, inum, inode, &inode_u, &subvol);
+		inode = bch2_inode_insert(c, inode);
+	}
+	bch2_trans_put(trans);
 
-	return &inode->v;
+	return ret ? ERR_PTR(ret) : &inode->v;
 }
 
 struct bch_inode_info *
@@ -226,7 +312,7 @@ __bch2_create(struct mnt_idmap *idmap,
 	struct bch_fs *c = dir->v.i_sb->s_fs_info;
 	struct btree_trans *trans;
 	struct bch_inode_unpacked dir_u;
-	struct bch_inode_info *inode, *old;
+	struct bch_inode_info *inode;
 	struct bch_inode_unpacked inode_u;
 	struct posix_acl *default_acl = NULL, *acl = NULL;
 	subvol_inum inum;
@@ -243,7 +329,7 @@ __bch2_create(struct mnt_idmap *idmap,
 	if (ret)
 		return ERR_PTR(ret);
 #endif
-	inode = to_bch_ei(new_inode(c->vfs_sb));
+	inode = __bch2_new_inode(c);
 	if (unlikely(!inode)) {
 		inode = ERR_PTR(-ENOMEM);
 		goto err;
@@ -276,7 +362,7 @@ retry:
 	inum.inum = inode_u.bi_inum;
 
 	ret   = bch2_subvolume_get(trans, inum.subvol, true,
-				   BTREE_ITER_WITH_UPDATES, &subvol) ?:
+				   BTREE_ITER_with_updates, &subvol) ?:
 		bch2_trans_commit(trans, NULL, &journal_seq, 0);
 	if (unlikely(ret)) {
 		bch2_quota_acct(c, bch_qid(&inode_u), Q_INO, -1,
@@ -293,7 +379,6 @@ err_before_quota:
 		mutex_unlock(&dir->ei_update_lock);
 	}
 
-	bch2_iget5_set(&inode->v, &inum);
 	bch2_vfs_inode_init(trans, inum, inode, &inode_u, &subvol);
 
 	set_cached_acl(&inode->v, ACL_TYPE_ACCESS, acl);
@@ -304,36 +389,7 @@ err_before_quota:
 	 * bch2_trans_exit() and dropping locks, else we could race with another
 	 * thread pulling the inode in and modifying it:
 	 */
-
-	inode->v.i_state |= I_CREATING;
-
-	old = to_bch_ei(inode_insert5(&inode->v,
-				      bch2_inode_hash(inum),
-				      bch2_iget5_test,
-				      bch2_iget5_set,
-				      &inum));
-	BUG_ON(!old);
-
-	if (unlikely(old != inode)) {
-		/*
-		 * We raced, another process pulled the new inode into cache
-		 * before us:
-		 */
-		make_bad_inode(&inode->v);
-		iput(&inode->v);
-
-		inode = old;
-	} else {
-		mutex_lock(&c->vfs_inodes_lock);
-		list_add(&inode->ei_vfs_inode_list, &c->vfs_inodes_list);
-		mutex_unlock(&c->vfs_inodes_lock);
-		/*
-		 * we really don't want insert_inode_locked2() to be setting
-		 * I_NEW...
-		 */
-		unlock_new_inode(&inode->v);
-	}
-
+	inode = bch2_inode_insert(c, inode);
 	bch2_trans_put(trans);
 err:
 	posix_acl_release(default_acl);
@@ -352,23 +408,86 @@ err_trans:
 
 /* methods */
 
+static struct bch_inode_info *bch2_lookup_trans(struct btree_trans *trans,
+			subvol_inum dir, struct bch_hash_info *dir_hash_info,
+			const struct qstr *name)
+{
+	struct bch_fs *c = trans->c;
+	struct btree_iter dirent_iter = {};
+	subvol_inum inum = {};
+	struct printbuf buf = PRINTBUF;
+
+	struct bkey_s_c k = bch2_hash_lookup(trans, &dirent_iter, bch2_dirent_hash_desc,
+					     dir_hash_info, dir, name, 0);
+	int ret = bkey_err(k);
+	if (ret)
+		return ERR_PTR(ret);
+
+	ret = bch2_dirent_read_target(trans, dir, bkey_s_c_to_dirent(k), &inum);
+	if (ret > 0)
+		ret = -ENOENT;
+	if (ret)
+		goto err;
+
+	struct bch_inode_info *inode =
+		to_bch_ei(ilookup5_nowait(c->vfs_sb,
+					  bch2_inode_hash(inum),
+					  bch2_iget5_test,
+					  &inum));
+	if (inode)
+		goto out;
+
+	struct bch_subvolume subvol;
+	struct bch_inode_unpacked inode_u;
+	ret =   bch2_subvolume_get(trans, inum.subvol, true, 0, &subvol) ?:
+		bch2_inode_find_by_inum_nowarn_trans(trans, inum, &inode_u) ?:
+		PTR_ERR_OR_ZERO(inode = bch2_new_inode(trans));
+
+	bch2_fs_inconsistent_on(bch2_err_matches(ret, ENOENT),
+				c, "dirent to missing inode:\n  %s",
+				(bch2_bkey_val_to_text(&buf, c, k), buf.buf));
+	if (ret)
+		goto err;
+
+	/* regular files may have hardlinks: */
+	if (bch2_fs_inconsistent_on(bch2_inode_should_have_bp(&inode_u) &&
+				    !bkey_eq(k.k->p, POS(inode_u.bi_dir, inode_u.bi_dir_offset)),
+				    c,
+				    "dirent points to inode that does not point back:\n  %s",
+				    (bch2_bkey_val_to_text(&buf, c, k),
+				     prt_printf(&buf, "\n  "),
+				     bch2_inode_unpacked_to_text(&buf, &inode_u),
+				     buf.buf))) {
+		ret = -ENOENT;
+		goto err;
+	}
+
+	bch2_vfs_inode_init(trans, inum, inode, &inode_u, &subvol);
+	inode = bch2_inode_insert(c, inode);
+out:
+	bch2_trans_iter_exit(trans, &dirent_iter);
+	printbuf_exit(&buf);
+	return inode;
+err:
+	inode = ERR_PTR(ret);
+	goto out;
+}
+
 static struct dentry *bch2_lookup(struct inode *vdir, struct dentry *dentry,
 				  unsigned int flags)
 {
 	struct bch_fs *c = vdir->i_sb->s_fs_info;
 	struct bch_inode_info *dir = to_bch_ei(vdir);
 	struct bch_hash_info hash = bch2_hash_info_init(c, &dir->ei_inode);
-	struct inode *vinode = NULL;
-	subvol_inum inum = { .subvol = 1 };
-	int ret;
-
-	ret = bch2_dirent_lookup(c, inode_inum(dir), &hash,
-				 &dentry->d_name, &inum);
 
-	if (!ret)
-		vinode = bch2_vfs_inode_get(c, inum);
+	struct bch_inode_info *inode;
+	bch2_trans_do(c, NULL, NULL, 0,
+		PTR_ERR_OR_ZERO(inode = bch2_lookup_trans(trans, inode_inum(dir),
+							  &hash, &dentry->d_name)));
+	if (IS_ERR(inode))
+		inode = NULL;
 
-	return d_splice_alias(vinode, dentry);
+	return d_splice_alias(&inode->v, dentry);
 }
 
 static int bch2_mknod(struct mnt_idmap *idmap,
@@ -715,7 +834,7 @@ retry:
 	acl = NULL;
 
 	ret = bch2_inode_peek(trans, &inode_iter, &inode_u, inode_inum(inode),
-			      BTREE_ITER_INTENT);
+			      BTREE_ITER_intent);
 	if (ret)
 		goto btree_err;
 
@@ -772,6 +891,9 @@ static int bch2_getattr(struct mnt_idmap *idmap,
 	stat->blksize	= block_bytes(c);
 	stat->blocks	= inode->v.i_blocks;
 
+	stat->subvol	= inode->ei_subvol;
+	stat->result_mask |= STATX_SUBVOL;
+
 	if (request_mask & STATX_BTIME) {
 		stat->result_mask |= STATX_BTIME;
 		stat->btime = bch2_time_to_timespec(c, inode->ei_inode.bi_otime);
@@ -892,7 +1014,6 @@ static int bch2_fiemap(struct inode *vinode, struct fiemap_extent_info *info,
 	struct btree_iter iter;
 	struct bkey_s_c k;
 	struct bkey_buf cur, prev;
-	struct bpos end = POS(ei->v.i_ino, (start + len) >> 9);
 	unsigned offset_into_extent, sectors;
 	bool have_extent = false;
 	u32 snapshot;
@@ -902,6 +1023,7 @@ static int bch2_fiemap(struct inode *vinode, struct fiemap_extent_info *info,
 	if (ret)
 		return ret;
 
+	struct bpos end = POS(ei->v.i_ino, (start + len) >> 9);
 	if (start + len < start)
 		return -EINVAL;
 
@@ -968,6 +1090,10 @@ retry:
 
 		bch2_btree_iter_set_pos(&iter,
 			POS(iter.pos.inode, iter.pos.offset + sectors));
+
+		ret = bch2_trans_relock(trans);
+		if (ret)
+			break;
 	}
 	start = iter.pos.offset;
 	bch2_trans_iter_exit(trans, &iter);
@@ -1034,6 +1160,8 @@ static int bch2_open(struct inode *vinode, struct file *file)
 			return ret;
 	}
 
+	file->f_mode |= FMODE_CAN_ODIRECT;
+
 	return generic_file_open(vinode, file);
 }
 
@@ -1043,6 +1171,7 @@ static const struct file_operations bch_file_operations = {
 	.read_iter	= bch2_read_iter,
 	.write_iter	= bch2_write_iter,
 	.mmap		= bch2_mmap,
+	.get_unmapped_area = thp_get_unmapped_area,
 	.fsync		= bch2_fsync,
 	.splice_read	= filemap_splice_read,
 	.splice_write	= iter_file_splice_write,
@@ -1126,7 +1255,6 @@ static const struct address_space_operations bch_address_space_operations = {
 	.write_end	= bch2_write_end,
 	.invalidate_folio = bch2_invalidate_folio,
 	.release_folio	= bch2_release_folio,
-	.direct_IO	= noop_direct_IO,
 #ifdef CONFIG_MIGRATION
 	.migrate_folio	= filemap_migrate_folio,
 #endif
@@ -1372,13 +1500,9 @@ static void bch2_vfs_inode_init(struct btree_trans *trans, subvol_inum inum,
 				struct bch_inode_unpacked *bi,
 				struct bch_subvolume *subvol)
 {
+	bch2_iget5_set(&inode->v, &inum);
 	bch2_inode_update_after_write(trans, inode, bi, ~0);
 
-	if (BCH_SUBVOLUME_SNAP(subvol))
-		set_bit(EI_INODE_SNAPSHOT, &inode->ei_flags);
-	else
-		clear_bit(EI_INODE_SNAPSHOT, &inode->ei_flags);
-
 	inode->v.i_blocks	= bi->bi_sectors;
 	inode->v.i_ino		= bi->bi_inum;
 	inode->v.i_rdev		= bi->bi_dev;
@@ -1390,6 +1514,9 @@ static void bch2_vfs_inode_init(struct btree_trans *trans, subvol_inum inum,
 	inode->ei_qid		= bch_qid(bi);
 	inode->ei_subvol	= inum.subvol;
 
+	if (BCH_SUBVOLUME_SNAP(subvol))
+		set_bit(EI_INODE_SNAPSHOT, &inode->ei_flags);
+
 	inode->v.i_mapping->a_ops = &bch_address_space_operations;
 
 	switch (inode->v.i_mode & S_IFMT) {
@@ -1414,34 +1541,9 @@ static void bch2_vfs_inode_init(struct btree_trans *trans, subvol_inum inum,
 	mapping_set_large_folios(inode->v.i_mapping);
 }
 
-static struct inode *bch2_alloc_inode(struct super_block *sb)
-{
-	struct bch_inode_info *inode;
-
-	inode = kmem_cache_alloc(bch2_inode_cache, GFP_NOFS);
-	if (!inode)
-		return NULL;
-
-	inode_init_once(&inode->v);
-	mutex_init(&inode->ei_update_lock);
-	two_state_lock_init(&inode->ei_pagecache_lock);
-	INIT_LIST_HEAD(&inode->ei_vfs_inode_list);
-	mutex_init(&inode->ei_quota_lock);
-
-	return &inode->v;
-}
-
-static void bch2_i_callback(struct rcu_head *head)
+static void bch2_free_inode(struct inode *vinode)
 {
-	struct inode *vinode = container_of(head, struct inode, i_rcu);
-	struct bch_inode_info *inode = to_bch_ei(vinode);
-
-	kmem_cache_free(bch2_inode_cache, inode);
-}
-
-static void bch2_destroy_inode(struct inode *vinode)
-{
-	call_rcu(&vinode->i_rcu, bch2_i_callback);
+	kmem_cache_free(bch2_inode_cache, to_bch_ei(vinode));
 }
 
 static int inode_update_times_fn(struct btree_trans *trans,
@@ -1572,7 +1674,6 @@ static int bch2_statfs(struct dentry *dentry, struct kstatfs *buf)
 	 * number:
 	 */
 	u64 avail_inodes = ((usage.capacity - usage.used) << 3);
-	u64 fsid;
 
 	buf->f_type	= BCACHEFS_STATFS_MAGIC;
 	buf->f_bsize	= sb->s_blocksize;
@@ -1583,10 +1684,7 @@ static int bch2_statfs(struct dentry *dentry, struct kstatfs *buf)
 	buf->f_files	= usage.nr_inodes + avail_inodes;
 	buf->f_ffree	= avail_inodes;
 
-	fsid = le64_to_cpup((void *) c->sb.user_uuid.b) ^
-	       le64_to_cpup((void *) c->sb.user_uuid.b + sizeof(u64));
-	buf->f_fsid.val[0] = fsid & 0xFFFFFFFFUL;
-	buf->f_fsid.val[1] = (fsid >> 32) & 0xFFFFFFFFUL;
+	buf->f_fsid	= uuid_to_fsid(c->sb.user_uuid.b);
 	buf->f_namelen	= BCH_NAME_MAX;
 
 	return 0;
@@ -1753,7 +1851,7 @@ static int bch2_unfreeze(struct super_block *sb)
 
 static const struct super_operations bch_super_operations = {
 	.alloc_inode	= bch2_alloc_inode,
-	.destroy_inode	= bch2_destroy_inode,
+	.free_inode	= bch2_free_inode,
 	.write_inode	= bch2_vfs_write_inode,
 	.evict_inode	= bch2_evict_inode,
 	.sync_fs	= bch2_sync_fs,
@@ -1805,8 +1903,10 @@ static struct dentry *bch2_mount(struct file_system_type *fs_type,
 	opt_set(opts, read_only, (flags & SB_RDONLY) != 0);
 
 	ret = bch2_parse_mount_opts(NULL, &opts, data);
-	if (ret)
+	if (ret) {
+		ret = bch2_err_class(ret);
 		return ERR_PTR(ret);
+	}
 
 	if (!dev_name || strlen(dev_name) == 0)
 		return ERR_PTR(-EINVAL);
@@ -1854,8 +1954,7 @@ got_sb:
 
 	if (IS_ERR(sb)) {
 		ret = PTR_ERR(sb);
-		ret = bch2_err_class(ret);
-		return ERR_PTR(ret);
+		goto err;
 	}
 
 	c = sb->s_fs_info;
@@ -1882,6 +1981,8 @@ got_sb:
 	sb->s_time_gran		= c->sb.nsec_per_time_unit;
 	sb->s_time_min		= div_s64(S64_MIN, c->sb.time_units_per_sec) + 1;
 	sb->s_time_max		= div_s64(S64_MAX, c->sb.time_units_per_sec);
+	sb->s_uuid		= c->sb.user_uuid;
+	sb->s_shrink->seeks	= 0;
 	c->vfs_sb		= sb;
 	strscpy(sb->s_id, c->name, sizeof(sb->s_id));
 
@@ -1928,7 +2029,19 @@ out:
 	return dget(sb->s_root);
 
 err_put_super:
+	__bch2_fs_stop(c);
 	deactivate_locked_super(sb);
+err:
+	if (ret)
+		pr_err("error: %s", bch2_err_str(ret));
+	/*
+	 * On an inconsistency error in recovery we might see an -EROFS derived
+	 * errorcode (from the journal), but we don't want to return that to
+	 * userspace as that causes util-linux to retry the mount RO - which is
+	 * confusing:
+	 */
+	if (bch2_err_matches(ret, EROFS) && ret != -EROFS)
+		ret = -EIO;
 	return ERR_PTR(bch2_err_class(ret));
 }
 
diff --git a/fs/bcachefs/fsck.c b/fs/bcachefs/fsck.c
index 6a760777ba..921bcdb3e5 100644
--- a/fs/bcachefs/fsck.c
+++ b/fs/bcachefs/fsck.c
@@ -12,7 +12,7 @@
 #include "fsck.h"
 #include "inode.h"
 #include "keylist.h"
-#include "recovery.h"
+#include "recovery_passes.h"
 #include "snapshot.h"
 #include "super.h"
 #include "xattr.h"
@@ -63,9 +63,7 @@ static int subvol_lookup(struct btree_trans *trans, u32 subvol,
 			 u32 *snapshot, u64 *inum)
 {
 	struct bch_subvolume s;
-	int ret;
-
-	ret = bch2_subvolume_get(trans, subvol, false, 0, &s);
+	int ret = bch2_subvolume_get(trans, subvol, false, 0, &s);
 
 	*snapshot = le32_to_cpu(s.snapshot);
 	*inum = le64_to_cpu(s.inode);
@@ -79,29 +77,25 @@ static int lookup_first_inode(struct btree_trans *trans, u64 inode_nr,
 	struct bkey_s_c k;
 	int ret;
 
-	bch2_trans_iter_init(trans, &iter, BTREE_ID_inodes,
-			     POS(0, inode_nr),
-			     BTREE_ITER_ALL_SNAPSHOTS);
-	k = bch2_btree_iter_peek(&iter);
-	ret = bkey_err(k);
-	if (ret)
-		goto err;
-
-	if (!k.k || !bkey_eq(k.k->p, POS(0, inode_nr))) {
-		ret = -BCH_ERR_ENOENT_inode;
-		goto err;
+	for_each_btree_key_norestart(trans, iter, BTREE_ID_inodes, POS(0, inode_nr),
+				     BTREE_ITER_all_snapshots, k, ret) {
+		if (k.k->p.offset != inode_nr)
+			break;
+		if (!bkey_is_inode(k.k))
+			continue;
+		ret = bch2_inode_unpack(k, inode);
+		goto found;
 	}
-
-	ret = bch2_inode_unpack(k, inode);
-err:
+	ret = -BCH_ERR_ENOENT_inode;
+found:
 	bch_err_msg(trans->c, ret, "fetching inode %llu", inode_nr);
 	bch2_trans_iter_exit(trans, &iter);
 	return ret;
 }
 
 static int lookup_inode(struct btree_trans *trans, u64 inode_nr,
-			  struct bch_inode_unpacked *inode,
-			  u32 *snapshot)
+			struct bch_inode_unpacked *inode,
+			u32 *snapshot)
 {
 	struct btree_iter iter;
 	struct bkey_s_c k;
@@ -129,47 +123,19 @@ static int lookup_dirent_in_snapshot(struct btree_trans *trans,
 			   u64 *target, unsigned *type, u32 snapshot)
 {
 	struct btree_iter iter;
-	struct bkey_s_c_dirent d;
-	int ret = bch2_hash_lookup_in_snapshot(trans, &iter, bch2_dirent_hash_desc,
-			       &hash_info, dir, name, 0, snapshot);
+	struct bkey_s_c k = bch2_hash_lookup_in_snapshot(trans, &iter, bch2_dirent_hash_desc,
+							 &hash_info, dir, name, 0, snapshot);
+	int ret = bkey_err(k);
 	if (ret)
 		return ret;
 
-	d = bkey_s_c_to_dirent(bch2_btree_iter_peek_slot(&iter));
+	struct bkey_s_c_dirent d = bkey_s_c_to_dirent(bch2_btree_iter_peek_slot(&iter));
 	*target = le64_to_cpu(d.v->d_inum);
 	*type = d.v->d_type;
 	bch2_trans_iter_exit(trans, &iter);
 	return 0;
 }
 
-static int __write_inode(struct btree_trans *trans,
-			 struct bch_inode_unpacked *inode,
-			 u32 snapshot)
-{
-	struct bkey_inode_buf *inode_p =
-		bch2_trans_kmalloc(trans, sizeof(*inode_p));
-
-	if (IS_ERR(inode_p))
-		return PTR_ERR(inode_p);
-
-	bch2_inode_pack(inode_p, inode);
-	inode_p->inode.k.p.snapshot = snapshot;
-
-	return bch2_btree_insert_nonextent(trans, BTREE_ID_inodes,
-				&inode_p->inode.k_i,
-				BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
-}
-
-static int fsck_write_inode(struct btree_trans *trans,
-			    struct bch_inode_unpacked *inode,
-			    u32 snapshot)
-{
-	int ret = commit_do(trans, NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
-			    __write_inode(trans, inode, snapshot));
-	bch_err_fn(trans->c, ret);
-	return ret;
-}
-
 static int __remove_dirent(struct btree_trans *trans, struct bpos pos)
 {
 	struct bch_fs *c = trans->c;
@@ -184,11 +150,12 @@ static int __remove_dirent(struct btree_trans *trans, struct bpos pos)
 
 	dir_hash_info = bch2_hash_info_init(c, &dir_inode);
 
-	bch2_trans_iter_init(trans, &iter, BTREE_ID_dirents, pos, BTREE_ITER_INTENT);
+	bch2_trans_iter_init(trans, &iter, BTREE_ID_dirents, pos, BTREE_ITER_intent);
 
-	ret = bch2_hash_delete_at(trans, bch2_dirent_hash_desc,
-				  &dir_hash_info, &iter,
-				  BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
+	ret =   bch2_btree_iter_traverse(&iter) ?:
+		bch2_hash_delete_at(trans, bch2_dirent_hash_desc,
+				    &dir_hash_info, &iter,
+				    BTREE_UPDATE_internal_snapshot_node);
 	bch2_trans_iter_exit(trans, &iter);
 err:
 	bch_err_fn(c, ret);
@@ -197,7 +164,8 @@ err:
 
 /* Get lost+found, create if it doesn't exist: */
 static int lookup_lostfound(struct btree_trans *trans, u32 snapshot,
-			    struct bch_inode_unpacked *lostfound)
+			    struct bch_inode_unpacked *lostfound,
+			    u64 reattaching_inum)
 {
 	struct bch_fs *c = trans->c;
 	struct qstr lostfound_str = QSTR("lost+found");
@@ -212,19 +180,36 @@ static int lookup_lostfound(struct btree_trans *trans, u32 snapshot,
 		return ret;
 
 	subvol_inum root_inum = { .subvol = le32_to_cpu(st.master_subvol) };
-	u32 subvol_snapshot;
 
-	ret = subvol_lookup(trans, le32_to_cpu(st.master_subvol),
-			    &subvol_snapshot, &root_inum.inum);
-	bch_err_msg(c, ret, "looking up root subvol");
+	struct bch_subvolume subvol;
+	ret = bch2_subvolume_get(trans, le32_to_cpu(st.master_subvol),
+				 false, 0, &subvol);
+	bch_err_msg(c, ret, "looking up root subvol %u for snapshot %u",
+		    le32_to_cpu(st.master_subvol), snapshot);
 	if (ret)
 		return ret;
 
+	if (!subvol.inode) {
+		struct btree_iter iter;
+		struct bkey_i_subvolume *subvol = bch2_bkey_get_mut_typed(trans, &iter,
+				BTREE_ID_subvolumes, POS(0, le32_to_cpu(st.master_subvol)),
+				0, subvolume);
+		ret = PTR_ERR_OR_ZERO(subvol);
+		if (ret)
+			return ret;
+
+		subvol->v.inode = cpu_to_le64(reattaching_inum);
+		bch2_trans_iter_exit(trans, &iter);
+	}
+
+	root_inum.inum = le64_to_cpu(subvol.inode);
+
 	struct bch_inode_unpacked root_inode;
 	struct bch_hash_info root_hash_info;
 	u32 root_inode_snapshot = snapshot;
 	ret = lookup_inode(trans, root_inum.inum, &root_inode, &root_inode_snapshot);
-	bch_err_msg(c, ret, "looking up root inode");
+	bch_err_msg(c, ret, "looking up root inode %llu for subvol %u",
+		    root_inum.inum, le32_to_cpu(st.master_subvol));
 	if (ret)
 		return ret;
 
@@ -280,14 +265,14 @@ create_lostfound:
 		goto err;
 
 	ret =   bch2_dirent_create_snapshot(trans,
-				root_inode.bi_inum, snapshot, &root_hash_info,
+				0, root_inode.bi_inum, snapshot, &root_hash_info,
 				mode_to_type(lostfound->bi_mode),
 				&lostfound_str,
 				lostfound->bi_inum,
 				&lostfound->bi_dir_offset,
-				BCH_HASH_SET_MUST_CREATE) ?:
+				STR_HASH_must_create) ?:
 		bch2_inode_write_flags(trans, &lostfound_iter, lostfound,
-				       BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
+				       BTREE_UPDATE_internal_snapshot_node);
 err:
 	bch_err_msg(c, ret, "creating lost+found");
 	bch2_trans_iter_exit(trans, &lostfound_iter);
@@ -303,38 +288,55 @@ static int reattach_inode(struct btree_trans *trans,
 	char name_buf[20];
 	struct qstr name;
 	u64 dir_offset = 0;
+	u32 dirent_snapshot = inode_snapshot;
 	int ret;
 
-	ret = lookup_lostfound(trans, inode_snapshot, &lostfound);
+	if (inode->bi_subvol) {
+		inode->bi_parent_subvol = BCACHEFS_ROOT_SUBVOL;
+
+		u64 root_inum;
+		ret = subvol_lookup(trans, inode->bi_parent_subvol,
+				    &dirent_snapshot, &root_inum);
+		if (ret)
+			return ret;
+
+		snprintf(name_buf, sizeof(name_buf), "subvol-%u", inode->bi_subvol);
+	} else {
+		snprintf(name_buf, sizeof(name_buf), "%llu", inode->bi_inum);
+	}
+
+	ret = lookup_lostfound(trans, dirent_snapshot, &lostfound, inode->bi_inum);
 	if (ret)
 		return ret;
 
 	if (S_ISDIR(inode->bi_mode)) {
 		lostfound.bi_nlink++;
 
-		ret = __write_inode(trans, &lostfound, U32_MAX);
+		ret = __bch2_fsck_write_inode(trans, &lostfound, U32_MAX);
 		if (ret)
 			return ret;
 	}
 
 	dir_hash = bch2_hash_info_init(trans->c, &lostfound);
 
-	snprintf(name_buf, sizeof(name_buf), "%llu", inode->bi_inum);
 	name = (struct qstr) QSTR(name_buf);
 
 	ret = bch2_dirent_create_snapshot(trans,
-				lostfound.bi_inum, inode_snapshot,
+				inode->bi_parent_subvol, lostfound.bi_inum,
+				dirent_snapshot,
 				&dir_hash,
 				inode_d_type(inode),
-				&name, inode->bi_inum, &dir_offset,
-				BCH_HASH_SET_MUST_CREATE);
+				&name,
+				inode->bi_subvol ?: inode->bi_inum,
+				&dir_offset,
+				STR_HASH_must_create);
 	if (ret)
 		return ret;
 
 	inode->bi_dir		= lostfound.bi_inum;
 	inode->bi_dir_offset	= dir_offset;
 
-	return __write_inode(trans, inode, inode_snapshot);
+	return __bch2_fsck_write_inode(trans, inode, inode_snapshot);
 }
 
 static int remove_backpointer(struct btree_trans *trans,
@@ -353,14 +355,136 @@ static int remove_backpointer(struct btree_trans *trans,
 	return ret;
 }
 
-struct snapshots_seen_entry {
-	u32				id;
-	u32				equiv;
-};
+static int reattach_subvol(struct btree_trans *trans, struct bkey_s_c_subvolume s)
+{
+	struct bch_fs *c = trans->c;
+
+	struct bch_inode_unpacked inode;
+	int ret = bch2_inode_find_by_inum_trans(trans,
+				(subvol_inum) { s.k->p.offset, le64_to_cpu(s.v->inode) },
+				&inode);
+	if (ret)
+		return ret;
+
+	ret = remove_backpointer(trans, &inode);
+	bch_err_msg(c, ret, "removing dirent");
+	if (ret)
+		return ret;
+
+	ret = reattach_inode(trans, &inode, le32_to_cpu(s.v->snapshot));
+	bch_err_msg(c, ret, "reattaching inode %llu", inode.bi_inum);
+	return ret;
+}
+
+static int reconstruct_subvol(struct btree_trans *trans, u32 snapshotid, u32 subvolid, u64 inum)
+{
+	struct bch_fs *c = trans->c;
+
+	if (!bch2_snapshot_is_leaf(c, snapshotid)) {
+		bch_err(c, "need to reconstruct subvol, but have interior node snapshot");
+		return -BCH_ERR_fsck_repair_unimplemented;
+	}
+
+	/*
+	 * If inum isn't set, that means we're being called from check_dirents,
+	 * not check_inodes - the root of this subvolume doesn't exist or we
+	 * would have found it there:
+	 */
+	if (!inum) {
+		struct btree_iter inode_iter = {};
+		struct bch_inode_unpacked new_inode;
+		u64 cpu = raw_smp_processor_id();
+
+		bch2_inode_init_early(c, &new_inode);
+		bch2_inode_init_late(&new_inode, bch2_current_time(c), 0, 0, S_IFDIR|0755, 0, NULL);
+
+		new_inode.bi_subvol = subvolid;
+
+		int ret = bch2_inode_create(trans, &inode_iter, &new_inode, snapshotid, cpu) ?:
+			  bch2_btree_iter_traverse(&inode_iter) ?:
+			  bch2_inode_write(trans, &inode_iter, &new_inode);
+		bch2_trans_iter_exit(trans, &inode_iter);
+		if (ret)
+			return ret;
+
+		inum = new_inode.bi_inum;
+	}
+
+	bch_info(c, "reconstructing subvol %u with root inode %llu", subvolid, inum);
+
+	struct bkey_i_subvolume *new_subvol = bch2_trans_kmalloc(trans, sizeof(*new_subvol));
+	int ret = PTR_ERR_OR_ZERO(new_subvol);
+	if (ret)
+		return ret;
+
+	bkey_subvolume_init(&new_subvol->k_i);
+	new_subvol->k.p.offset	= subvolid;
+	new_subvol->v.snapshot	= cpu_to_le32(snapshotid);
+	new_subvol->v.inode	= cpu_to_le64(inum);
+	ret = bch2_btree_insert_trans(trans, BTREE_ID_subvolumes, &new_subvol->k_i, 0);
+	if (ret)
+		return ret;
+
+	struct btree_iter iter;
+	struct bkey_i_snapshot *s = bch2_bkey_get_mut_typed(trans, &iter,
+			BTREE_ID_snapshots, POS(0, snapshotid),
+			0, snapshot);
+	ret = PTR_ERR_OR_ZERO(s);
+	bch_err_msg(c, ret, "getting snapshot %u", snapshotid);
+	if (ret)
+		return ret;
+
+	u32 snapshot_tree = le32_to_cpu(s->v.tree);
+
+	s->v.subvol = cpu_to_le32(subvolid);
+	SET_BCH_SNAPSHOT_SUBVOL(&s->v, true);
+	bch2_trans_iter_exit(trans, &iter);
+
+	struct bkey_i_snapshot_tree *st = bch2_bkey_get_mut_typed(trans, &iter,
+			BTREE_ID_snapshot_trees, POS(0, snapshot_tree),
+			0, snapshot_tree);
+	ret = PTR_ERR_OR_ZERO(st);
+	bch_err_msg(c, ret, "getting snapshot tree %u", snapshot_tree);
+	if (ret)
+		return ret;
+
+	if (!st->v.master_subvol)
+		st->v.master_subvol = cpu_to_le32(subvolid);
+
+	bch2_trans_iter_exit(trans, &iter);
+	return 0;
+}
+
+static int reconstruct_inode(struct btree_trans *trans, u32 snapshot, u64 inum, u64 size, unsigned mode)
+{
+	struct bch_fs *c = trans->c;
+	struct bch_inode_unpacked new_inode;
+
+	bch2_inode_init_early(c, &new_inode);
+	bch2_inode_init_late(&new_inode, bch2_current_time(c), 0, 0, mode|0755, 0, NULL);
+	new_inode.bi_size = size;
+	new_inode.bi_inum = inum;
+
+	return __bch2_fsck_write_inode(trans, &new_inode, snapshot);
+}
+
+static int reconstruct_reg_inode(struct btree_trans *trans, u32 snapshot, u64 inum)
+{
+	struct btree_iter iter = {};
+
+	bch2_trans_iter_init(trans, &iter, BTREE_ID_extents, SPOS(inum, U64_MAX, snapshot), 0);
+	struct bkey_s_c k = bch2_btree_iter_peek_prev(&iter);
+	bch2_trans_iter_exit(trans, &iter);
+	int ret = bkey_err(k);
+	if (ret)
+		return ret;
+
+	return reconstruct_inode(trans, snapshot, inum, k.k->p.offset << 9, S_IFREG);
+}
 
 struct snapshots_seen {
 	struct bpos			pos;
-	DARRAY(struct snapshots_seen_entry) ids;
+	snapshot_id_list		ids;
 };
 
 static inline void snapshots_seen_exit(struct snapshots_seen *s)
@@ -375,20 +499,15 @@ static inline void snapshots_seen_init(struct snapshots_seen *s)
 
 static int snapshots_seen_add_inorder(struct bch_fs *c, struct snapshots_seen *s, u32 id)
 {
-	struct snapshots_seen_entry *i, n = {
-		.id	= id,
-		.equiv	= bch2_snapshot_equiv(c, id),
-	};
-	int ret = 0;
-
+	u32 *i;
 	__darray_for_each(s->ids, i) {
-		if (i->id == id)
+		if (*i == id)
 			return 0;
-		if (i->id > id)
+		if (*i > id)
 			break;
 	}
 
-	ret = darray_insert_item(&s->ids, i - s->ids.data, n);
+	int ret = darray_insert_item(&s->ids, i - s->ids.data, id);
 	if (ret)
 		bch_err(c, "error reallocating snapshots_seen table (size %zu)",
 			s->ids.size);
@@ -398,42 +517,11 @@ static int snapshots_seen_add_inorder(struct bch_fs *c, struct snapshots_seen *s
 static int snapshots_seen_update(struct bch_fs *c, struct snapshots_seen *s,
 				 enum btree_id btree_id, struct bpos pos)
 {
-	struct snapshots_seen_entry n = {
-		.id	= pos.snapshot,
-		.equiv	= bch2_snapshot_equiv(c, pos.snapshot),
-	};
-	int ret = 0;
-
 	if (!bkey_eq(s->pos, pos))
 		s->ids.nr = 0;
-
 	s->pos = pos;
-	s->pos.snapshot = n.equiv;
-
-	darray_for_each(s->ids, i) {
-		if (i->id == n.id)
-			return 0;
 
-		/*
-		 * We currently don't rigorously track for snapshot cleanup
-		 * needing to be run, so it shouldn't be a fsck error yet:
-		 */
-		if (i->equiv == n.equiv) {
-			bch_err(c, "snapshot deletion did not finish:\n"
-				"  duplicate keys in btree %s at %llu:%llu snapshots %u, %u (equiv %u)\n",
-				bch2_btree_id_str(btree_id),
-				pos.inode, pos.offset,
-				i->id, n.id, n.equiv);
-			set_bit(BCH_FS_need_delete_dead_snapshots, &c->flags);
-			return bch2_run_explicit_recovery_pass(c, BCH_RECOVERY_PASS_delete_dead_snapshots);
-		}
-	}
-
-	ret = darray_push(&s->ids, n);
-	if (ret)
-		bch_err(c, "error reallocating snapshots_seen table (size %zu)",
-			s->ids.size);
-	return ret;
+	return snapshot_list_add_nodup(c, &s->ids, pos.snapshot);
 }
 
 /**
@@ -453,12 +541,10 @@ static bool key_visible_in_snapshot(struct bch_fs *c, struct snapshots_seen *see
 	ssize_t i;
 
 	EBUG_ON(id > ancestor);
-	EBUG_ON(!bch2_snapshot_is_equiv(c, id));
-	EBUG_ON(!bch2_snapshot_is_equiv(c, ancestor));
 
 	/* @ancestor should be the snapshot most recently added to @seen */
 	EBUG_ON(ancestor != seen->pos.snapshot);
-	EBUG_ON(ancestor != seen->ids.data[seen->ids.nr - 1].equiv);
+	EBUG_ON(ancestor != darray_last(seen->ids));
 
 	if (id == ancestor)
 		return true;
@@ -477,9 +563,9 @@ static bool key_visible_in_snapshot(struct bch_fs *c, struct snapshots_seen *see
 	 */
 
 	for (i = seen->ids.nr - 2;
-	     i >= 0 && seen->ids.data[i].equiv >= id;
+	     i >= 0 && seen->ids.data[i] >= id;
 	     --i)
-		if (bch2_snapshot_is_ancestor(c, id, seen->ids.data[i].equiv))
+		if (bch2_snapshot_is_ancestor(c, id, seen->ids.data[i]))
 			return false;
 
 	return true;
@@ -510,9 +596,6 @@ static int ref_visible2(struct bch_fs *c,
 			u32 src, struct snapshots_seen *src_seen,
 			u32 dst, struct snapshots_seen *dst_seen)
 {
-	src = bch2_snapshot_equiv(c, src);
-	dst = bch2_snapshot_equiv(c, dst);
-
 	if (dst > src) {
 		swap(dst, src);
 		swap(dst_seen, src_seen);
@@ -559,7 +642,7 @@ static int add_inode(struct bch_fs *c, struct inode_walker *w,
 
 	return darray_push(&w->inodes, ((struct inode_walker_entry) {
 		.inode		= u,
-		.snapshot	= bch2_snapshot_equiv(c, inode.k->p.snapshot),
+		.snapshot	= inode.k->p.snapshot,
 	}));
 }
 
@@ -575,7 +658,7 @@ static int get_inodes_all_snapshots(struct btree_trans *trans,
 	w->inodes.nr = 0;
 
 	for_each_btree_key_norestart(trans, iter, BTREE_ID_inodes, POS(0, inum),
-				     BTREE_ITER_ALL_SNAPSHOTS, k, ret) {
+				     BTREE_ITER_all_snapshots, k, ret) {
 		if (k.k->p.offset != inum)
 			break;
 
@@ -592,37 +675,39 @@ static int get_inodes_all_snapshots(struct btree_trans *trans,
 }
 
 static struct inode_walker_entry *
-lookup_inode_for_snapshot(struct bch_fs *c, struct inode_walker *w,
-			  u32 snapshot, bool is_whiteout)
+lookup_inode_for_snapshot(struct bch_fs *c, struct inode_walker *w, struct bkey_s_c k)
 {
-	struct inode_walker_entry *i;
-
-	snapshot = bch2_snapshot_equiv(c, snapshot);
+	bool is_whiteout = k.k->type == KEY_TYPE_whiteout;
 
+	struct inode_walker_entry *i;
 	__darray_for_each(w->inodes, i)
-		if (bch2_snapshot_is_ancestor(c, snapshot, i->snapshot))
+		if (bch2_snapshot_is_ancestor(c, k.k->p.snapshot, i->snapshot))
 			goto found;
 
 	return NULL;
 found:
-	BUG_ON(snapshot > i->snapshot);
+	BUG_ON(k.k->p.snapshot > i->snapshot);
 
-	if (snapshot != i->snapshot && !is_whiteout) {
+	if (k.k->p.snapshot != i->snapshot && !is_whiteout) {
 		struct inode_walker_entry new = *i;
-		size_t pos;
-		int ret;
 
-		new.snapshot = snapshot;
+		new.snapshot = k.k->p.snapshot;
 		new.count = 0;
 
-		bch_info(c, "have key for inode %llu:%u but have inode in ancestor snapshot %u",
-			 w->last_pos.inode, snapshot, i->snapshot);
+		struct printbuf buf = PRINTBUF;
+		bch2_bkey_val_to_text(&buf, c, k);
 
-		while (i > w->inodes.data && i[-1].snapshot > snapshot)
+		bch_info(c, "have key for inode %llu:%u but have inode in ancestor snapshot %u\n"
+			 "unexpected because we should always update the inode when we update a key in that inode\n"
+			 "%s",
+			 w->last_pos.inode, k.k->p.snapshot, i->snapshot, buf.buf);
+		printbuf_exit(&buf);
+
+		while (i > w->inodes.data && i[-1].snapshot > k.k->p.snapshot)
 			--i;
 
-		pos = i - w->inodes.data;
-		ret = darray_insert_item(&w->inodes, pos, new);
+		size_t pos = i - w->inodes.data;
+		int ret = darray_insert_item(&w->inodes, pos, new);
 		if (ret)
 			return ERR_PTR(ret);
 
@@ -633,27 +718,27 @@ found:
 }
 
 static struct inode_walker_entry *walk_inode(struct btree_trans *trans,
-					     struct inode_walker *w, struct bpos pos,
-					     bool is_whiteout)
+					     struct inode_walker *w,
+					     struct bkey_s_c k)
 {
-	if (w->last_pos.inode != pos.inode) {
-		int ret = get_inodes_all_snapshots(trans, w, pos.inode);
+	if (w->last_pos.inode != k.k->p.inode) {
+		int ret = get_inodes_all_snapshots(trans, w, k.k->p.inode);
 		if (ret)
 			return ERR_PTR(ret);
-	} else if (bkey_cmp(w->last_pos, pos)) {
+	} else if (bkey_cmp(w->last_pos, k.k->p)) {
 		darray_for_each(w->inodes, i)
 			i->seen_this_pos = false;
 	}
 
-	w->last_pos = pos;
+	w->last_pos = k.k->p;
 
-	return lookup_inode_for_snapshot(trans->c, w, pos.snapshot, is_whiteout);
+	return lookup_inode_for_snapshot(trans->c, w, k);
 }
 
-static int __get_visible_inodes(struct btree_trans *trans,
-				struct inode_walker *w,
-				struct snapshots_seen *s,
-				u64 inum)
+static int get_visible_inodes(struct btree_trans *trans,
+			      struct inode_walker *w,
+			      struct snapshots_seen *s,
+			      u64 inum)
 {
 	struct bch_fs *c = trans->c;
 	struct btree_iter iter;
@@ -663,19 +748,17 @@ static int __get_visible_inodes(struct btree_trans *trans,
 	w->inodes.nr = 0;
 
 	for_each_btree_key_norestart(trans, iter, BTREE_ID_inodes, POS(0, inum),
-			   BTREE_ITER_ALL_SNAPSHOTS, k, ret) {
-		u32 equiv = bch2_snapshot_equiv(c, k.k->p.snapshot);
-
+			   BTREE_ITER_all_snapshots, k, ret) {
 		if (k.k->p.offset != inum)
 			break;
 
-		if (!ref_visible(c, s, s->pos.snapshot, equiv))
+		if (!ref_visible(c, s, s->pos.snapshot, k.k->p.snapshot))
 			continue;
 
 		if (bkey_is_inode(k.k))
 			add_inode(c, w, k);
 
-		if (equiv >= s->pos.snapshot)
+		if (k.k->p.snapshot >= s->pos.snapshot)
 			break;
 	}
 	bch2_trans_iter_exit(trans, &iter);
@@ -683,25 +766,6 @@ static int __get_visible_inodes(struct btree_trans *trans,
 	return ret;
 }
 
-static int check_key_has_snapshot(struct btree_trans *trans,
-				  struct btree_iter *iter,
-				  struct bkey_s_c k)
-{
-	struct bch_fs *c = trans->c;
-	struct printbuf buf = PRINTBUF;
-	int ret = 0;
-
-	if (mustfix_fsck_err_on(!bch2_snapshot_equiv(c, k.k->p.snapshot), c,
-				bkey_in_missing_snapshot,
-				"key in missing snapshot: %s",
-				(bch2_bkey_val_to_text(&buf, c, k), buf.buf)))
-		ret = bch2_btree_delete_at(trans, iter,
-					    BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE) ?: 1;
-fsck_err:
-	printbuf_exit(&buf);
-	return ret;
-}
-
 static int hash_redo_key(struct btree_trans *trans,
 			 const struct bch_hash_desc desc,
 			 struct bch_hash_info *hash_info,
@@ -722,11 +786,11 @@ static int hash_redo_key(struct btree_trans *trans,
 	delete->k.p = k_iter->pos;
 	return  bch2_btree_iter_traverse(k_iter) ?:
 		bch2_trans_update(trans, k_iter, delete, 0) ?:
-		bch2_hash_set_snapshot(trans, desc, hash_info,
+		bch2_hash_set_in_snapshot(trans, desc, hash_info,
 				       (subvol_inum) { 0, k.k->p.inode },
 				       k.k->p.snapshot, tmp,
-				       BCH_HASH_SET_MUST_CREATE,
-				       BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE) ?:
+				       STR_HASH_must_create|
+				       BTREE_UPDATE_internal_snapshot_node) ?:
 		bch2_trans_commit(trans, NULL, NULL, BCH_TRANS_COMMIT_no_enospc);
 }
 
@@ -755,7 +819,7 @@ static int hash_check_key(struct btree_trans *trans,
 
 	for_each_btree_key_norestart(trans, iter, desc.btree_id,
 				     SPOS(hash_k.k->p.inode, hash, hash_k.k->p.snapshot),
-				     BTREE_ITER_SLOTS, k, ret) {
+				     BTREE_ITER_slots, k, ret) {
 		if (bkey_eq(k.k->p, hash_k.k->p))
 			break;
 
@@ -795,16 +859,93 @@ fsck_err:
 	goto out;
 }
 
+static struct bkey_s_c_dirent dirent_get_by_pos(struct btree_trans *trans,
+						struct btree_iter *iter,
+						struct bpos pos)
+{
+	return bch2_bkey_get_iter_typed(trans, iter, BTREE_ID_dirents, pos, 0, dirent);
+}
+
+static struct bkey_s_c_dirent inode_get_dirent(struct btree_trans *trans,
+					       struct btree_iter *iter,
+					       struct bch_inode_unpacked *inode,
+					       u32 *snapshot)
+{
+	if (inode->bi_subvol) {
+		u64 inum;
+		int ret = subvol_lookup(trans, inode->bi_parent_subvol, snapshot, &inum);
+		if (ret)
+			return ((struct bkey_s_c_dirent) { .k = ERR_PTR(ret) });
+	}
+
+	return dirent_get_by_pos(trans, iter, SPOS(inode->bi_dir, inode->bi_dir_offset, *snapshot));
+}
+
+static bool inode_points_to_dirent(struct bch_inode_unpacked *inode,
+				   struct bkey_s_c_dirent d)
+{
+	return  inode->bi_dir		== d.k->p.inode &&
+		inode->bi_dir_offset	== d.k->p.offset;
+}
+
+static bool dirent_points_to_inode(struct bkey_s_c_dirent d,
+				   struct bch_inode_unpacked *inode)
+{
+	return d.v->d_type == DT_SUBVOL
+		? le32_to_cpu(d.v->d_child_subvol)	== inode->bi_subvol
+		: le64_to_cpu(d.v->d_inum)		== inode->bi_inum;
+}
+
 static int check_inode_deleted_list(struct btree_trans *trans, struct bpos p)
 {
 	struct btree_iter iter;
 	struct bkey_s_c k = bch2_bkey_get_iter(trans, &iter, BTREE_ID_deleted_inodes, p, 0);
-	int ret = bkey_err(k);
-	if (ret)
+	int ret = bkey_err(k) ?: k.k->type == KEY_TYPE_set;
+	bch2_trans_iter_exit(trans, &iter);
+	return ret;
+}
+
+static int check_inode_dirent_inode(struct btree_trans *trans, struct bkey_s_c inode_k,
+				    struct bch_inode_unpacked *inode,
+				    u32 inode_snapshot, bool *write_inode)
+{
+	struct bch_fs *c = trans->c;
+	struct printbuf buf = PRINTBUF;
+
+	struct btree_iter dirent_iter = {};
+	struct bkey_s_c_dirent d = inode_get_dirent(trans, &dirent_iter, inode, &inode_snapshot);
+	int ret = bkey_err(d);
+	if (ret && !bch2_err_matches(ret, ENOENT))
 		return ret;
 
-	bch2_trans_iter_exit(trans, &iter);
-	return k.k->type == KEY_TYPE_set;
+	if (fsck_err_on(ret,
+			c, inode_points_to_missing_dirent,
+			"inode points to missing dirent\n%s",
+			(bch2_bkey_val_to_text(&buf, c, inode_k), buf.buf)) ||
+	    fsck_err_on(!ret && !dirent_points_to_inode(d, inode),
+			c, inode_points_to_wrong_dirent,
+			"inode points to dirent that does not point back:\n%s",
+			(bch2_bkey_val_to_text(&buf, c, inode_k),
+			 prt_newline(&buf),
+			 bch2_bkey_val_to_text(&buf, c, d.s_c), buf.buf))) {
+		/*
+		 * We just clear the backpointer fields for now. If we find a
+		 * dirent that points to this inode in check_dirents(), we'll
+		 * update it then; then when we get to check_path() if the
+		 * backpointer is still 0 we'll reattach it.
+		 */
+		inode->bi_dir = 0;
+		inode->bi_dir_offset = 0;
+		inode->bi_flags &= ~BCH_INODE_backptr_untrusted;
+		*write_inode = true;
+	}
+
+	ret = 0;
+fsck_err:
+	bch2_trans_iter_exit(trans, &dirent_iter);
+	printbuf_exit(&buf);
+	bch_err_fn(c, ret);
+	return ret;
 }
 
 static int check_inode(struct btree_trans *trans,
@@ -819,7 +960,7 @@ static int check_inode(struct btree_trans *trans,
 	bool do_update = false;
 	int ret;
 
-	ret = check_key_has_snapshot(trans, iter, k);
+	ret = bch2_check_key_has_snapshot(trans, iter, k);
 	if (ret < 0)
 		goto err;
 	if (ret)
@@ -861,7 +1002,8 @@ static int check_inode(struct btree_trans *trans,
 
 		u.bi_flags &= ~BCH_INODE_i_size_dirty|BCH_INODE_unlinked;
 
-		ret = __write_inode(trans, &u, iter->pos.snapshot);
+		ret = __bch2_fsck_write_inode(trans, &u, iter->pos.snapshot);
+
 		bch_err_msg(c, ret, "in fsck updating inode");
 		if (ret)
 			return ret;
@@ -876,7 +1018,7 @@ static int check_inode(struct btree_trans *trans,
 		if (ret < 0)
 			return ret;
 
-		fsck_err_on(ret, c, unlinked_inode_not_on_deleted_list,
+		fsck_err_on(!ret, c, unlinked_inode_not_on_deleted_list,
 			    "inode %llu:%u unlinked, but not on deleted list",
 			    u.bi_inum, k.k->p.snapshot);
 		ret = 0;
@@ -950,8 +1092,54 @@ static int check_inode(struct btree_trans *trans,
 		do_update = true;
 	}
 
+	if (u.bi_dir || u.bi_dir_offset) {
+		ret = check_inode_dirent_inode(trans, k, &u, k.k->p.snapshot, &do_update);
+		if (ret)
+			goto err;
+	}
+
+	if (fsck_err_on(u.bi_parent_subvol &&
+			(u.bi_subvol == 0 ||
+			 u.bi_subvol == BCACHEFS_ROOT_SUBVOL),
+			c, inode_bi_parent_nonzero,
+			"inode %llu:%u has subvol %u but nonzero parent subvol %u",
+			u.bi_inum, k.k->p.snapshot, u.bi_subvol, u.bi_parent_subvol)) {
+		u.bi_parent_subvol = 0;
+		do_update = true;
+	}
+
+	if (u.bi_subvol) {
+		struct bch_subvolume s;
+
+		ret = bch2_subvolume_get(trans, u.bi_subvol, false, 0, &s);
+		if (ret && !bch2_err_matches(ret, ENOENT))
+			goto err;
+
+		if (ret && (c->sb.btrees_lost_data & BIT_ULL(BTREE_ID_subvolumes))) {
+			ret = reconstruct_subvol(trans, k.k->p.snapshot, u.bi_subvol, u.bi_inum);
+			goto do_update;
+		}
+
+		if (fsck_err_on(ret,
+				c, inode_bi_subvol_missing,
+				"inode %llu:%u bi_subvol points to missing subvolume %u",
+				u.bi_inum, k.k->p.snapshot, u.bi_subvol) ||
+		    fsck_err_on(le64_to_cpu(s.inode) != u.bi_inum ||
+				!bch2_snapshot_is_ancestor(c, le32_to_cpu(s.snapshot),
+							   k.k->p.snapshot),
+				c, inode_bi_subvol_wrong,
+				"inode %llu:%u points to subvol %u, but subvol points to %llu:%u",
+				u.bi_inum, k.k->p.snapshot, u.bi_subvol,
+				le64_to_cpu(s.inode),
+				le32_to_cpu(s.snapshot))) {
+			u.bi_subvol = 0;
+			u.bi_parent_subvol = 0;
+			do_update = true;
+		}
+	}
+do_update:
 	if (do_update) {
-		ret = __write_inode(trans, &u, iter->pos.snapshot);
+		ret = __bch2_fsck_write_inode(trans, &u, iter->pos.snapshot);
 		bch_err_msg(c, ret, "in fsck updating inode");
 		if (ret)
 			return ret;
@@ -973,7 +1161,7 @@ int bch2_check_inodes(struct bch_fs *c)
 	int ret = bch2_trans_run(c,
 		for_each_btree_key_commit(trans, iter, BTREE_ID_inodes,
 				POS_MIN,
-				BTREE_ITER_PREFETCH|BTREE_ITER_ALL_SNAPSHOTS, k,
+				BTREE_ITER_prefetch|BTREE_ITER_all_snapshots, k,
 				NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
 			check_inode(trans, &iter, k, &prev, &s, full)));
 
@@ -982,32 +1170,9 @@ int bch2_check_inodes(struct bch_fs *c)
 	return ret;
 }
 
-static struct bkey_s_c_dirent dirent_get_by_pos(struct btree_trans *trans,
-						struct btree_iter *iter,
-						struct bpos pos)
-{
-	return bch2_bkey_get_iter_typed(trans, iter, BTREE_ID_dirents, pos, 0, dirent);
-}
-
-static bool inode_points_to_dirent(struct bch_inode_unpacked *inode,
-				   struct bkey_s_c_dirent d)
-{
-	return  inode->bi_dir		== d.k->p.inode &&
-		inode->bi_dir_offset	== d.k->p.offset;
-}
-
-static bool dirent_points_to_inode(struct bkey_s_c_dirent d,
-				   struct bch_inode_unpacked *inode)
-{
-	return d.v->d_type == DT_SUBVOL
-		? le32_to_cpu(d.v->d_child_subvol)	== inode->bi_subvol
-		: le64_to_cpu(d.v->d_inum)		== inode->bi_inum;
-}
-
-static int check_i_sectors(struct btree_trans *trans, struct inode_walker *w)
+static int check_i_sectors_notnested(struct btree_trans *trans, struct inode_walker *w)
 {
 	struct bch_fs *c = trans->c;
-	u32 restart_count = trans->restart_count;
 	int ret = 0;
 	s64 count2;
 
@@ -1021,8 +1186,8 @@ static int check_i_sectors(struct btree_trans *trans, struct inode_walker *w)
 			i->count = count2;
 
 		if (i->count != count2) {
-			bch_err(c, "fsck counted i_sectors wrong for inode %llu:%u: got %llu should be %llu",
-				w->last_pos.inode, i->snapshot, i->count, count2);
+			bch_err_ratelimited(c, "fsck counted i_sectors wrong for inode %llu:%u: got %llu should be %llu",
+					    w->last_pos.inode, i->snapshot, i->count, count2);
 			return -BCH_ERR_internal_fsck_err;
 		}
 
@@ -1032,14 +1197,21 @@ static int check_i_sectors(struct btree_trans *trans, struct inode_walker *w)
 				w->last_pos.inode, i->snapshot,
 				i->inode.bi_sectors, i->count)) {
 			i->inode.bi_sectors = i->count;
-			ret = fsck_write_inode(trans, &i->inode, i->snapshot);
+			ret = bch2_fsck_write_inode(trans, &i->inode, i->snapshot);
 			if (ret)
 				break;
 		}
 	}
 fsck_err:
 	bch_err_fn(c, ret);
-	return ret ?: trans_was_restarted(trans, restart_count);
+	return ret;
+}
+
+static int check_i_sectors(struct btree_trans *trans, struct inode_walker *w)
+{
+	u32 restart_count = trans->restart_count;
+	return check_i_sectors_notnested(trans, w) ?:
+		trans_was_restarted(trans, restart_count);
 }
 
 struct extent_end {
@@ -1118,8 +1290,8 @@ static int overlapping_extents_found(struct btree_trans *trans,
 	BUG_ON(bkey_le(pos1, bkey_start_pos(&pos2)));
 
 	bch2_trans_iter_init(trans, &iter1, btree, pos1,
-			     BTREE_ITER_ALL_SNAPSHOTS|
-			     BTREE_ITER_NOT_EXTENTS);
+			     BTREE_ITER_all_snapshots|
+			     BTREE_ITER_not_extents);
 	k1 = bch2_btree_iter_peek_upto(&iter1, POS(pos1.inode, U64_MAX));
 	ret = bkey_err(k1);
 	if (ret)
@@ -1181,7 +1353,7 @@ static int overlapping_extents_found(struct btree_trans *trans,
 		trans->extra_disk_res += bch2_bkey_sectors_compressed(k2);
 
 		ret =   bch2_trans_update_extent_overwrite(trans, old_iter,
-				BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE,
+				BTREE_UPDATE_internal_snapshot_node,
 				k1, k2) ?:
 			bch2_trans_commit(trans, &res, NULL, BCH_TRANS_COMMIT_no_enospc);
 		bch2_disk_reservation_put(c, &res);
@@ -1222,7 +1394,6 @@ static int check_overlapping_extents(struct btree_trans *trans,
 			      struct snapshots_seen *seen,
 			      struct extent_ends *extent_ends,
 			      struct bkey_s_c k,
-			      u32 equiv,
 			      struct btree_iter *iter,
 			      bool *fixed)
 {
@@ -1255,10 +1426,6 @@ static int check_overlapping_extents(struct btree_trans *trans,
 			goto err;
 	}
 
-	ret = extent_ends_at(c, extent_ends, seen, k);
-	if (ret)
-		goto err;
-
 	extent_ends->last_pos = k.k->p;
 err:
 	return ret;
@@ -1295,12 +1462,9 @@ static int check_extent(struct btree_trans *trans, struct btree_iter *iter,
 	struct bch_fs *c = trans->c;
 	struct inode_walker_entry *i;
 	struct printbuf buf = PRINTBUF;
-	struct bpos equiv = k.k->p;
 	int ret = 0;
 
-	equiv.snapshot = bch2_snapshot_equiv(c, k.k->p.snapshot);
-
-	ret = check_key_has_snapshot(trans, iter, k);
+	ret = bch2_check_key_has_snapshot(trans, iter, k);
 	if (ret) {
 		ret = ret < 0 ? ret : 0;
 		goto out;
@@ -1312,7 +1476,7 @@ static int check_extent(struct btree_trans *trans, struct btree_iter *iter,
 			goto err;
 	}
 
-	i = walk_inode(trans, inode, equiv, k.k->type == KEY_TYPE_whiteout);
+	i = walk_inode(trans, inode, k);
 	ret = PTR_ERR_OR_ZERO(i);
 	if (ret)
 		goto err;
@@ -1322,6 +1486,17 @@ static int check_extent(struct btree_trans *trans, struct btree_iter *iter,
 		goto err;
 
 	if (k.k->type != KEY_TYPE_whiteout) {
+		if (!i && (c->sb.btrees_lost_data & BIT_ULL(BTREE_ID_inodes))) {
+			ret =   reconstruct_reg_inode(trans, k.k->p.snapshot, k.k->p.inode) ?:
+				bch2_trans_commit(trans, NULL, NULL, BCH_TRANS_COMMIT_no_enospc);
+			if (ret)
+				goto err;
+
+			inode->last_pos.inode--;
+			ret = -BCH_ERR_transaction_restart_nested;
+			goto err;
+		}
+
 		if (fsck_err_on(!i, c, extent_in_missing_inode,
 				"extent in missing inode:\n  %s",
 				(printbuf_reset(&buf),
@@ -1338,8 +1513,7 @@ static int check_extent(struct btree_trans *trans, struct btree_iter *iter,
 				 bch2_bkey_val_to_text(&buf, c, k), buf.buf)))
 			goto delete;
 
-		ret = check_overlapping_extents(trans, s, extent_ends, k,
-						equiv.snapshot, iter,
+		ret = check_overlapping_extents(trans, s, extent_ends, k, iter,
 						&inode->recalculate_sums);
 		if (ret)
 			goto err;
@@ -1356,8 +1530,8 @@ static int check_extent(struct btree_trans *trans, struct btree_iter *iter,
 	for (;
 	     inode->inodes.data && i >= inode->inodes.data;
 	     --i) {
-		if (i->snapshot > equiv.snapshot ||
-		    !key_visible_in_snapshot(c, s, i->snapshot, equiv.snapshot))
+		if (i->snapshot > k.k->p.snapshot ||
+		    !key_visible_in_snapshot(c, s, i->snapshot, k.k->p.snapshot))
 			continue;
 
 		if (k.k->type != KEY_TYPE_whiteout) {
@@ -1374,7 +1548,7 @@ static int check_extent(struct btree_trans *trans, struct btree_iter *iter,
 				bch2_btree_iter_set_snapshot(&iter2, i->snapshot);
 				ret =   bch2_btree_iter_traverse(&iter2) ?:
 					bch2_btree_delete_at(trans, &iter2,
-						BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
+						BTREE_UPDATE_internal_snapshot_node);
 				bch2_trans_iter_exit(trans, &iter2);
 				if (ret)
 					goto err;
@@ -1388,6 +1562,12 @@ static int check_extent(struct btree_trans *trans, struct btree_iter *iter,
 
 		i->seen_this_pos = true;
 	}
+
+	if (k.k->type != KEY_TYPE_whiteout) {
+		ret = extent_ends_at(c, extent_ends, s, k);
+		if (ret)
+			goto err;
+	}
 out:
 err:
 fsck_err:
@@ -1395,7 +1575,7 @@ fsck_err:
 	bch_err_fn(c, ret);
 	return ret;
 delete:
-	ret = bch2_btree_delete_at(trans, iter, BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
+	ret = bch2_btree_delete_at(trans, iter, BTREE_UPDATE_internal_snapshot_node);
 	goto out;
 }
 
@@ -1416,14 +1596,14 @@ int bch2_check_extents(struct bch_fs *c)
 	int ret = bch2_trans_run(c,
 		for_each_btree_key_commit(trans, iter, BTREE_ID_extents,
 				POS(BCACHEFS_ROOT_INO, 0),
-				BTREE_ITER_PREFETCH|BTREE_ITER_ALL_SNAPSHOTS, k,
+				BTREE_ITER_prefetch|BTREE_ITER_all_snapshots, k,
 				&res, NULL,
 				BCH_TRANS_COMMIT_no_enospc, ({
 			bch2_disk_reservation_put(c, &res);
 			check_extent(trans, &iter, k, &w, &s, &extent_ends) ?:
 			check_extent_overbig(trans, &iter, k);
 		})) ?:
-		check_i_sectors(trans, &w));
+		check_i_sectors_notnested(trans, &w));
 
 	bch2_disk_reservation_put(c, &res);
 	extent_ends_exit(&extent_ends);
@@ -1441,7 +1621,7 @@ int bch2_check_indirect_extents(struct bch_fs *c)
 	int ret = bch2_trans_run(c,
 		for_each_btree_key_commit(trans, iter, BTREE_ID_reflink,
 				POS_MIN,
-				BTREE_ITER_PREFETCH, k,
+				BTREE_ITER_prefetch, k,
 				&res, NULL,
 				BCH_TRANS_COMMIT_no_enospc, ({
 			bch2_disk_reservation_put(c, &res);
@@ -1453,10 +1633,9 @@ int bch2_check_indirect_extents(struct bch_fs *c)
 	return ret;
 }
 
-static int check_subdir_count(struct btree_trans *trans, struct inode_walker *w)
+static int check_subdir_count_notnested(struct btree_trans *trans, struct inode_walker *w)
 {
 	struct bch_fs *c = trans->c;
-	u32 restart_count = trans->restart_count;
 	int ret = 0;
 	s64 count2;
 
@@ -1469,8 +1648,8 @@ static int check_subdir_count(struct btree_trans *trans, struct inode_walker *w)
 			return count2;
 
 		if (i->count != count2) {
-			bch_err(c, "fsck counted subdirectories wrong: got %llu should be %llu",
-				i->count, count2);
+			bch_err_ratelimited(c, "fsck counted subdirectories wrong for inum %llu:%u: got %llu should be %llu",
+					    w->last_pos.inode, i->snapshot, i->count, count2);
 			i->count = count2;
 			if (i->inode.bi_nlink == i->count)
 				continue;
@@ -1481,96 +1660,135 @@ static int check_subdir_count(struct btree_trans *trans, struct inode_walker *w)
 				"directory %llu:%u with wrong i_nlink: got %u, should be %llu",
 				w->last_pos.inode, i->snapshot, i->inode.bi_nlink, i->count)) {
 			i->inode.bi_nlink = i->count;
-			ret = fsck_write_inode(trans, &i->inode, i->snapshot);
+			ret = bch2_fsck_write_inode(trans, &i->inode, i->snapshot);
 			if (ret)
 				break;
 		}
 	}
 fsck_err:
 	bch_err_fn(c, ret);
-	return ret ?: trans_was_restarted(trans, restart_count);
+	return ret;
 }
 
-static int check_dirent_target(struct btree_trans *trans,
-			       struct btree_iter *iter,
-			       struct bkey_s_c_dirent d,
-			       struct bch_inode_unpacked *target,
-			       u32 target_snapshot)
+static int check_subdir_count(struct btree_trans *trans, struct inode_walker *w)
+{
+	u32 restart_count = trans->restart_count;
+	return check_subdir_count_notnested(trans, w) ?:
+		trans_was_restarted(trans, restart_count);
+}
+
+noinline_for_stack
+static int check_dirent_inode_dirent(struct btree_trans *trans,
+				   struct btree_iter *iter,
+				   struct bkey_s_c_dirent d,
+				   struct bch_inode_unpacked *target,
+				   u32 target_snapshot)
 {
 	struct bch_fs *c = trans->c;
-	struct bkey_i_dirent *n;
 	struct printbuf buf = PRINTBUF;
-	struct btree_iter bp_iter = { NULL };
 	int ret = 0;
 
+	if (inode_points_to_dirent(target, d))
+		return 0;
+
+	if (bch2_inode_should_have_bp(target) &&
+	    !fsck_err(c, inode_wrong_backpointer,
+		      "dirent points to inode that does not point back:\n  %s",
+		      (bch2_bkey_val_to_text(&buf, c, d.s_c),
+		       prt_printf(&buf, "\n  "),
+		       bch2_inode_unpacked_to_text(&buf, target),
+		       buf.buf)))
+		goto out_noiter;
+
 	if (!target->bi_dir &&
 	    !target->bi_dir_offset) {
 		target->bi_dir		= d.k->p.inode;
 		target->bi_dir_offset	= d.k->p.offset;
-
-		ret = __write_inode(trans, target, target_snapshot);
-		if (ret)
-			goto err;
+		return __bch2_fsck_write_inode(trans, target, target_snapshot);
 	}
 
-	if (!inode_points_to_dirent(target, d)) {
-		struct bkey_s_c_dirent bp_dirent = dirent_get_by_pos(trans, &bp_iter,
-				      SPOS(target->bi_dir, target->bi_dir_offset, target_snapshot));
-		ret = bkey_err(bp_dirent);
-		if (ret && !bch2_err_matches(ret, ENOENT))
-			goto err;
+	struct btree_iter bp_iter = { NULL };
+	struct bkey_s_c_dirent bp_dirent = dirent_get_by_pos(trans, &bp_iter,
+			      SPOS(target->bi_dir, target->bi_dir_offset, target_snapshot));
+	ret = bkey_err(bp_dirent);
+	if (ret && !bch2_err_matches(ret, ENOENT))
+		goto err;
 
-		bool backpointer_exists = !ret;
-		ret = 0;
+	bool backpointer_exists = !ret;
+	ret = 0;
+
+	if (fsck_err_on(!backpointer_exists,
+			c, inode_wrong_backpointer,
+			"inode %llu:%u has wrong backpointer:\n"
+			"got       %llu:%llu\n"
+			"should be %llu:%llu",
+			target->bi_inum, target_snapshot,
+			target->bi_dir,
+			target->bi_dir_offset,
+			d.k->p.inode,
+			d.k->p.offset)) {
+		target->bi_dir		= d.k->p.inode;
+		target->bi_dir_offset	= d.k->p.offset;
+		ret = __bch2_fsck_write_inode(trans, target, target_snapshot);
+		goto out;
+	}
 
-		bch2_bkey_val_to_text(&buf, c, d.s_c);
-		prt_newline(&buf);
-		if (backpointer_exists)
-			bch2_bkey_val_to_text(&buf, c, bp_dirent.s_c);
+	bch2_bkey_val_to_text(&buf, c, d.s_c);
+	prt_newline(&buf);
+	if (backpointer_exists)
+		bch2_bkey_val_to_text(&buf, c, bp_dirent.s_c);
+
+	if (fsck_err_on(backpointer_exists &&
+			(S_ISDIR(target->bi_mode) ||
+			 target->bi_subvol),
+			c, inode_dir_multiple_links,
+			"%s %llu:%u with multiple links\n%s",
+			S_ISDIR(target->bi_mode) ? "directory" : "subvolume",
+			target->bi_inum, target_snapshot, buf.buf)) {
+		ret = __remove_dirent(trans, d.k->p);
+		goto out;
+	}
 
-		if (fsck_err_on(S_ISDIR(target->bi_mode) && backpointer_exists,
-				c, inode_dir_multiple_links,
-				"directory %llu:%u with multiple links\n%s",
-				target->bi_inum, target_snapshot, buf.buf)) {
-			ret = __remove_dirent(trans, d.k->p);
-			goto out;
-		}
+	/*
+	 * hardlinked file with nlink 0:
+	 * We're just adjusting nlink here so check_nlinks() will pick
+	 * it up, it ignores inodes with nlink 0
+	 */
+	if (fsck_err_on(backpointer_exists && !target->bi_nlink,
+			c, inode_multiple_links_but_nlink_0,
+			"inode %llu:%u type %s has multiple links but i_nlink 0\n%s",
+			target->bi_inum, target_snapshot, bch2_d_types[d.v->d_type], buf.buf)) {
+		target->bi_nlink++;
+		target->bi_flags &= ~BCH_INODE_unlinked;
+		ret = __bch2_fsck_write_inode(trans, target, target_snapshot);
+		if (ret)
+			goto err;
+	}
+out:
+err:
+fsck_err:
+	bch2_trans_iter_exit(trans, &bp_iter);
+out_noiter:
+	printbuf_exit(&buf);
+	bch_err_fn(c, ret);
+	return ret;
+}
 
-		/*
-		 * hardlinked file with nlink 0:
-		 * We're just adjusting nlink here so check_nlinks() will pick
-		 * it up, it ignores inodes with nlink 0
-		 */
-		if (fsck_err_on(backpointer_exists && !target->bi_nlink,
-				c, inode_multiple_links_but_nlink_0,
-				"inode %llu:%u type %s has multiple links but i_nlink 0\n%s",
-				target->bi_inum, target_snapshot, bch2_d_types[d.v->d_type], buf.buf)) {
-			target->bi_nlink++;
-			target->bi_flags &= ~BCH_INODE_unlinked;
-
-			ret = __write_inode(trans, target, target_snapshot);
-			if (ret)
-				goto err;
-		}
+noinline_for_stack
+static int check_dirent_target(struct btree_trans *trans,
+			       struct btree_iter *iter,
+			       struct bkey_s_c_dirent d,
+			       struct bch_inode_unpacked *target,
+			       u32 target_snapshot)
+{
+	struct bch_fs *c = trans->c;
+	struct bkey_i_dirent *n;
+	struct printbuf buf = PRINTBUF;
+	int ret = 0;
 
-		if (fsck_err_on(!backpointer_exists,
-				c, inode_wrong_backpointer,
-				"inode %llu:%u has wrong backpointer:\n"
-				"got       %llu:%llu\n"
-				"should be %llu:%llu",
-				target->bi_inum, target_snapshot,
-				target->bi_dir,
-				target->bi_dir_offset,
-				d.k->p.inode,
-				d.k->p.offset)) {
-			target->bi_dir		= d.k->p.inode;
-			target->bi_dir_offset	= d.k->p.offset;
-
-			ret = __write_inode(trans, target, target_snapshot);
-			if (ret)
-				goto err;
-		}
-	}
+	ret = check_dirent_inode_dirent(trans, iter, d, target, target_snapshot);
+	if (ret)
+		goto err;
 
 	if (fsck_err_on(d.v->d_type != inode_d_type(target),
 			c, dirent_d_type_wrong,
@@ -1586,6 +1804,12 @@ static int check_dirent_target(struct btree_trans *trans,
 
 		bkey_reassemble(&n->k_i, d.s_c);
 		n->v.d_type = inode_d_type(target);
+		if (n->v.d_type == DT_SUBVOL) {
+			n->v.d_parent_subvol = cpu_to_le32(target->bi_parent_subvol);
+			n->v.d_child_subvol = cpu_to_le32(target->bi_subvol);
+		} else {
+			n->v.d_inum = cpu_to_le64(target->bi_inum);
+		}
 
 		ret = bch2_trans_update(trans, iter, &n->k_i, 0);
 		if (ret)
@@ -1593,33 +1817,164 @@ static int check_dirent_target(struct btree_trans *trans,
 
 		d = dirent_i_to_s_c(n);
 	}
+err:
+fsck_err:
+	printbuf_exit(&buf);
+	bch_err_fn(c, ret);
+	return ret;
+}
+
+/* find a subvolume that's a descendent of @snapshot: */
+static int find_snapshot_subvol(struct btree_trans *trans, u32 snapshot, u32 *subvolid)
+{
+	struct btree_iter iter;
+	struct bkey_s_c k;
+	int ret;
+
+	for_each_btree_key_norestart(trans, iter, BTREE_ID_subvolumes, POS_MIN, 0, k, ret) {
+		if (k.k->type != KEY_TYPE_subvolume)
+			continue;
+
+		struct bkey_s_c_subvolume s = bkey_s_c_to_subvolume(k);
+		if (bch2_snapshot_is_ancestor(trans->c, le32_to_cpu(s.v->snapshot), snapshot)) {
+			bch2_trans_iter_exit(trans, &iter);
+			*subvolid = k.k->p.offset;
+			goto found;
+		}
+	}
+	if (!ret)
+		ret = -ENOENT;
+found:
+	bch2_trans_iter_exit(trans, &iter);
+	return ret;
+}
 
-	if (fsck_err_on(d.v->d_type == DT_SUBVOL &&
-			target->bi_parent_subvol != le32_to_cpu(d.v->d_parent_subvol),
-			c, dirent_d_parent_subvol_wrong,
-			"dirent has wrong d_parent_subvol field: got %u, should be %u",
-			le32_to_cpu(d.v->d_parent_subvol),
-			target->bi_parent_subvol)) {
-		n = bch2_trans_kmalloc(trans, bkey_bytes(d.k));
-		ret = PTR_ERR_OR_ZERO(n);
+noinline_for_stack
+static int check_dirent_to_subvol(struct btree_trans *trans, struct btree_iter *iter,
+				  struct bkey_s_c_dirent d)
+{
+	struct bch_fs *c = trans->c;
+	struct btree_iter subvol_iter = {};
+	struct bch_inode_unpacked subvol_root;
+	u32 parent_subvol = le32_to_cpu(d.v->d_parent_subvol);
+	u32 target_subvol = le32_to_cpu(d.v->d_child_subvol);
+	u32 parent_snapshot;
+	u32 new_parent_subvol = 0;
+	u64 parent_inum;
+	struct printbuf buf = PRINTBUF;
+	int ret = 0;
+
+	ret = subvol_lookup(trans, parent_subvol, &parent_snapshot, &parent_inum);
+	if (ret && !bch2_err_matches(ret, ENOENT))
+		return ret;
+
+	if (ret ||
+	    (!ret && !bch2_snapshot_is_ancestor(c, parent_snapshot, d.k->p.snapshot))) {
+		int ret2 = find_snapshot_subvol(trans, d.k->p.snapshot, &new_parent_subvol);
+		if (ret2 && !bch2_err_matches(ret, ENOENT))
+			return ret2;
+	}
+
+	if (ret &&
+	    !new_parent_subvol &&
+	    (c->sb.btrees_lost_data & BIT_ULL(BTREE_ID_subvolumes))) {
+		/*
+		 * Couldn't find a subvol for dirent's snapshot - but we lost
+		 * subvols, so we need to reconstruct:
+		 */
+		ret = reconstruct_subvol(trans, d.k->p.snapshot, parent_subvol, 0);
+		if (ret)
+			return ret;
+
+		parent_snapshot = d.k->p.snapshot;
+	}
+
+	if (fsck_err_on(ret, c, dirent_to_missing_parent_subvol,
+			"dirent parent_subvol points to missing subvolume\n%s",
+			(bch2_bkey_val_to_text(&buf, c, d.s_c), buf.buf)) ||
+	    fsck_err_on(!ret && !bch2_snapshot_is_ancestor(c, parent_snapshot, d.k->p.snapshot),
+			c, dirent_not_visible_in_parent_subvol,
+			"dirent not visible in parent_subvol (not an ancestor of subvol snap %u)\n%s",
+			parent_snapshot,
+			(bch2_bkey_val_to_text(&buf, c, d.s_c), buf.buf))) {
+		if (!new_parent_subvol) {
+			bch_err(c, "could not find a subvol for snapshot %u", d.k->p.snapshot);
+			return -BCH_ERR_fsck_repair_unimplemented;
+		}
+
+		struct bkey_i_dirent *new_dirent = bch2_bkey_make_mut_typed(trans, iter, &d.s_c, 0, dirent);
+		ret = PTR_ERR_OR_ZERO(new_dirent);
 		if (ret)
 			goto err;
 
-		bkey_reassemble(&n->k_i, d.s_c);
-		n->v.d_parent_subvol = cpu_to_le32(target->bi_parent_subvol);
+		new_dirent->v.d_parent_subvol = cpu_to_le32(new_parent_subvol);
+	}
 
-		ret = bch2_trans_update(trans, iter, &n->k_i, 0);
+	struct bkey_s_c_subvolume s =
+		bch2_bkey_get_iter_typed(trans, &subvol_iter,
+					 BTREE_ID_subvolumes, POS(0, target_subvol),
+					 0, subvolume);
+	ret = bkey_err(s.s_c);
+	if (ret && !bch2_err_matches(ret, ENOENT))
+		return ret;
+
+	if (ret) {
+		if (fsck_err(c, dirent_to_missing_subvol,
+			     "dirent points to missing subvolume\n%s",
+			     (bch2_bkey_val_to_text(&buf, c, d.s_c), buf.buf)))
+			return __remove_dirent(trans, d.k->p);
+		ret = 0;
+		goto out;
+	}
+
+	if (fsck_err_on(le32_to_cpu(s.v->fs_path_parent) != parent_subvol,
+			c, subvol_fs_path_parent_wrong,
+			"subvol with wrong fs_path_parent, should be be %u\n%s",
+			parent_subvol,
+			(bch2_bkey_val_to_text(&buf, c, s.s_c), buf.buf))) {
+		struct bkey_i_subvolume *n =
+			bch2_bkey_make_mut_typed(trans, &subvol_iter, &s.s_c, 0, subvolume);
+		ret = PTR_ERR_OR_ZERO(n);
 		if (ret)
 			goto err;
 
-		d = dirent_i_to_s_c(n);
+		n->v.fs_path_parent = cpu_to_le32(parent_subvol);
+	}
+
+	u64 target_inum = le64_to_cpu(s.v->inode);
+	u32 target_snapshot = le32_to_cpu(s.v->snapshot);
+
+	ret = lookup_inode(trans, target_inum, &subvol_root, &target_snapshot);
+	if (ret && !bch2_err_matches(ret, ENOENT))
+		goto err;
+
+	if (ret) {
+		bch_err(c, "subvol %u points to missing inode root %llu", target_subvol, target_inum);
+		ret = -BCH_ERR_fsck_repair_unimplemented;
+		ret = 0;
+		goto err;
+	}
+
+	if (fsck_err_on(!ret && parent_subvol != subvol_root.bi_parent_subvol,
+			c, inode_bi_parent_wrong,
+			"subvol root %llu has wrong bi_parent_subvol: got %u, should be %u",
+			target_inum,
+			subvol_root.bi_parent_subvol, parent_subvol)) {
+		subvol_root.bi_parent_subvol = parent_subvol;
+		ret = __bch2_fsck_write_inode(trans, &subvol_root, target_snapshot);
+		if (ret)
+			goto err;
 	}
+
+	ret = check_dirent_target(trans, iter, d, &subvol_root,
+				  target_snapshot);
+	if (ret)
+		goto err;
 out:
 err:
 fsck_err:
-	bch2_trans_iter_exit(trans, &bp_iter);
+	bch2_trans_iter_exit(trans, &subvol_iter);
 	printbuf_exit(&buf);
-	bch_err_fn(c, ret);
 	return ret;
 }
 
@@ -1631,21 +1986,16 @@ static int check_dirent(struct btree_trans *trans, struct btree_iter *iter,
 			struct snapshots_seen *s)
 {
 	struct bch_fs *c = trans->c;
-	struct bkey_s_c_dirent d;
 	struct inode_walker_entry *i;
 	struct printbuf buf = PRINTBUF;
-	struct bpos equiv;
 	int ret = 0;
 
-	ret = check_key_has_snapshot(trans, iter, k);
+	ret = bch2_check_key_has_snapshot(trans, iter, k);
 	if (ret) {
 		ret = ret < 0 ? ret : 0;
 		goto out;
 	}
 
-	equiv = k.k->p;
-	equiv.snapshot = bch2_snapshot_equiv(c, k.k->p.snapshot);
-
 	ret = snapshots_seen_update(c, s, iter->btree_id, k.k->p);
 	if (ret)
 		goto err;
@@ -1661,7 +2011,7 @@ static int check_dirent(struct btree_trans *trans, struct btree_iter *iter,
 
 	BUG_ON(!btree_iter_path(trans, iter)->should_be_locked);
 
-	i = walk_inode(trans, dir, equiv, k.k->type == KEY_TYPE_whiteout);
+	i = walk_inode(trans, dir, k);
 	ret = PTR_ERR_OR_ZERO(i);
 	if (ret < 0)
 		goto err;
@@ -1670,12 +2020,23 @@ static int check_dirent(struct btree_trans *trans, struct btree_iter *iter,
 		*hash_info = bch2_hash_info_init(c, &dir->inodes.data[0].inode);
 	dir->first_this_inode = false;
 
+	if (!i && (c->sb.btrees_lost_data & BIT_ULL(BTREE_ID_inodes))) {
+		ret =   reconstruct_inode(trans, k.k->p.snapshot, k.k->p.inode, 0, S_IFDIR) ?:
+			bch2_trans_commit(trans, NULL, NULL, BCH_TRANS_COMMIT_no_enospc);
+		if (ret)
+			goto err;
+
+		dir->last_pos.inode--;
+		ret = -BCH_ERR_transaction_restart_nested;
+		goto err;
+	}
+
 	if (fsck_err_on(!i, c, dirent_in_missing_dir_inode,
 			"dirent in nonexisting directory:\n%s",
 			(printbuf_reset(&buf),
 			 bch2_bkey_val_to_text(&buf, c, k), buf.buf))) {
 		ret = bch2_btree_delete_at(trans, iter,
-				BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
+				BTREE_UPDATE_internal_snapshot_node);
 		goto out;
 	}
 
@@ -1704,64 +2065,20 @@ static int check_dirent(struct btree_trans *trans, struct btree_iter *iter,
 	if (k.k->type != KEY_TYPE_dirent)
 		goto out;
 
-	d = bkey_s_c_to_dirent(k);
+	struct bkey_s_c_dirent d = bkey_s_c_to_dirent(k);
 
 	if (d.v->d_type == DT_SUBVOL) {
-		struct bch_inode_unpacked subvol_root;
-		u32 target_subvol = le32_to_cpu(d.v->d_child_subvol);
-		u32 target_snapshot;
-		u64 target_inum;
-
-		ret = subvol_lookup(trans, target_subvol,
-				      &target_snapshot, &target_inum);
-		if (ret && !bch2_err_matches(ret, ENOENT))
-			goto err;
-
-		if (fsck_err_on(ret, c, dirent_to_missing_subvol,
-				"dirent points to missing subvolume %u",
-				le32_to_cpu(d.v->d_child_subvol))) {
-			ret = __remove_dirent(trans, d.k->p);
-			goto err;
-		}
-
-		ret = lookup_inode(trans, target_inum,
-				   &subvol_root, &target_snapshot);
-		if (ret && !bch2_err_matches(ret, ENOENT))
-			goto err;
-
-		if (fsck_err_on(ret, c, subvol_to_missing_root,
-				"subvolume %u points to missing subvolume root %llu",
-				target_subvol,
-				target_inum)) {
-			bch_err(c, "repair not implemented yet");
-			ret = -EINVAL;
-			goto err;
-		}
-
-		if (fsck_err_on(subvol_root.bi_subvol != target_subvol,
-				c, subvol_root_wrong_bi_subvol,
-				"subvol root %llu has wrong bi_subvol field: got %u, should be %u",
-				target_inum,
-				subvol_root.bi_subvol, target_subvol)) {
-			subvol_root.bi_subvol = target_subvol;
-			ret = __write_inode(trans, &subvol_root, target_snapshot);
-			if (ret)
-				goto err;
-		}
-
-		ret = check_dirent_target(trans, iter, d, &subvol_root,
-					  target_snapshot);
+		ret = check_dirent_to_subvol(trans, iter, d);
 		if (ret)
 			goto err;
 	} else {
-		ret = __get_visible_inodes(trans, target, s, le64_to_cpu(d.v->d_inum));
+		ret = get_visible_inodes(trans, target, s, le64_to_cpu(d.v->d_inum));
 		if (ret)
 			goto err;
 
 		if (fsck_err_on(!target->inodes.nr,
 				c, dirent_to_missing_inode,
-				"dirent points to missing inode: (equiv %u)\n%s",
-				equiv.snapshot,
+				"dirent points to missing inode:\n%s",
 				(printbuf_reset(&buf),
 				 bch2_bkey_val_to_text(&buf, c, k),
 				 buf.buf))) {
@@ -1776,12 +2093,11 @@ static int check_dirent(struct btree_trans *trans, struct btree_iter *iter,
 			if (ret)
 				goto err;
 		}
-	}
-
-	if (d.v->d_type == DT_DIR)
-		for_each_visible_inode(c, s, dir, equiv.snapshot, i)
-			i->count++;
 
+		if (d.v->d_type == DT_DIR)
+			for_each_visible_inode(c, s, dir, d.k->p.snapshot, i)
+				i->count++;
+	}
 out:
 err:
 fsck_err:
@@ -1806,11 +2122,12 @@ int bch2_check_dirents(struct bch_fs *c)
 	int ret = bch2_trans_run(c,
 		for_each_btree_key_commit(trans, iter, BTREE_ID_dirents,
 				POS(BCACHEFS_ROOT_INO, 0),
-				BTREE_ITER_PREFETCH|BTREE_ITER_ALL_SNAPSHOTS,
+				BTREE_ITER_prefetch|BTREE_ITER_all_snapshots,
 				k,
 				NULL, NULL,
 				BCH_TRANS_COMMIT_no_enospc,
-			check_dirent(trans, &iter, k, &hash_info, &dir, &target, &s)));
+			check_dirent(trans, &iter, k, &hash_info, &dir, &target, &s)) ?:
+		check_subdir_count_notnested(trans, &dir));
 
 	snapshots_seen_exit(&s);
 	inode_walker_exit(&dir);
@@ -1828,11 +2145,13 @@ static int check_xattr(struct btree_trans *trans, struct btree_iter *iter,
 	struct inode_walker_entry *i;
 	int ret;
 
-	ret = check_key_has_snapshot(trans, iter, k);
-	if (ret)
+	ret = bch2_check_key_has_snapshot(trans, iter, k);
+	if (ret < 0)
 		return ret;
+	if (ret)
+		return 0;
 
-	i = walk_inode(trans, inode, k.k->p, k.k->type == KEY_TYPE_whiteout);
+	i = walk_inode(trans, inode, k);
 	ret = PTR_ERR_OR_ZERO(i);
 	if (ret)
 		return ret;
@@ -1867,7 +2186,7 @@ int bch2_check_xattrs(struct bch_fs *c)
 	ret = bch2_trans_run(c,
 		for_each_btree_key_commit(trans, iter, BTREE_ID_xattrs,
 			POS(BCACHEFS_ROOT_INO, 0),
-			BTREE_ITER_PREFETCH|BTREE_ITER_ALL_SNAPSHOTS,
+			BTREE_ITER_prefetch|BTREE_ITER_all_snapshots,
 			k,
 			NULL, NULL,
 			BCH_TRANS_COMMIT_no_enospc,
@@ -1890,17 +2209,21 @@ static int check_root_trans(struct btree_trans *trans)
 
 	if (mustfix_fsck_err_on(ret, c, root_subvol_missing,
 				"root subvol missing")) {
-		struct bkey_i_subvolume root_subvol;
+		struct bkey_i_subvolume *root_subvol =
+			bch2_trans_kmalloc(trans, sizeof(*root_subvol));
+		ret = PTR_ERR_OR_ZERO(root_subvol);
+		if (ret)
+			goto err;
 
 		snapshot	= U32_MAX;
 		inum		= BCACHEFS_ROOT_INO;
 
-		bkey_subvolume_init(&root_subvol.k_i);
-		root_subvol.k.p.offset = BCACHEFS_ROOT_SUBVOL;
-		root_subvol.v.flags	= 0;
-		root_subvol.v.snapshot	= cpu_to_le32(snapshot);
-		root_subvol.v.inode	= cpu_to_le64(inum);
-		ret = bch2_btree_insert_trans(trans, BTREE_ID_subvolumes, &root_subvol.k_i, 0);
+		bkey_subvolume_init(&root_subvol->k_i);
+		root_subvol->k.p.offset = BCACHEFS_ROOT_SUBVOL;
+		root_subvol->v.flags	= 0;
+		root_subvol->v.snapshot	= cpu_to_le32(snapshot);
+		root_subvol->v.inode	= cpu_to_le64(inum);
+		ret = bch2_btree_insert_trans(trans, BTREE_ID_subvolumes, &root_subvol->k_i, 0);
 		bch_err_msg(c, ret, "writing root subvol");
 		if (ret)
 			goto err;
@@ -1919,7 +2242,7 @@ static int check_root_trans(struct btree_trans *trans)
 				0, NULL);
 		root_inode.bi_inum = inum;
 
-		ret = __write_inode(trans, &root_inode, snapshot);
+		ret = __bch2_fsck_write_inode(trans, &root_inode, snapshot);
 		bch_err_msg(c, ret, "writing root inode");
 	}
 err:
@@ -1936,6 +2259,107 @@ int bch2_check_root(struct bch_fs *c)
 	return ret;
 }
 
+typedef DARRAY(u32) darray_u32;
+
+static bool darray_u32_has(darray_u32 *d, u32 v)
+{
+	darray_for_each(*d, i)
+		if (*i == v)
+			return true;
+	return false;
+}
+
+/*
+ * We've checked that inode backpointers point to valid dirents; here, it's
+ * sufficient to check that the subvolume root has a dirent:
+ */
+static int subvol_has_dirent(struct btree_trans *trans, struct bkey_s_c_subvolume s)
+{
+	struct bch_inode_unpacked inode;
+	int ret = bch2_inode_find_by_inum_trans(trans,
+				(subvol_inum) { s.k->p.offset, le64_to_cpu(s.v->inode) },
+				&inode);
+	if (ret)
+		return ret;
+
+	return inode.bi_dir != 0;
+}
+
+static int check_subvol_path(struct btree_trans *trans, struct btree_iter *iter, struct bkey_s_c k)
+{
+	struct bch_fs *c = trans->c;
+	struct btree_iter parent_iter = {};
+	darray_u32 subvol_path = {};
+	struct printbuf buf = PRINTBUF;
+	int ret = 0;
+
+	if (k.k->type != KEY_TYPE_subvolume)
+		return 0;
+
+	while (k.k->p.offset != BCACHEFS_ROOT_SUBVOL) {
+		ret = darray_push(&subvol_path, k.k->p.offset);
+		if (ret)
+			goto err;
+
+		struct bkey_s_c_subvolume s = bkey_s_c_to_subvolume(k);
+
+		ret = subvol_has_dirent(trans, s);
+		if (ret < 0)
+			break;
+
+		if (fsck_err_on(!ret,
+				c, subvol_unreachable,
+				"unreachable subvolume %s",
+				(bch2_bkey_val_to_text(&buf, c, s.s_c),
+				 buf.buf))) {
+			ret = reattach_subvol(trans, s);
+			break;
+		}
+
+		u32 parent = le32_to_cpu(s.v->fs_path_parent);
+
+		if (darray_u32_has(&subvol_path, parent)) {
+			if (fsck_err(c, subvol_loop, "subvolume loop"))
+				ret = reattach_subvol(trans, s);
+			break;
+		}
+
+		bch2_trans_iter_exit(trans, &parent_iter);
+		bch2_trans_iter_init(trans, &parent_iter,
+				     BTREE_ID_subvolumes, POS(0, parent), 0);
+		k = bch2_btree_iter_peek_slot(&parent_iter);
+		ret = bkey_err(k);
+		if (ret)
+			goto err;
+
+		if (fsck_err_on(k.k->type != KEY_TYPE_subvolume,
+				c, subvol_unreachable,
+				"unreachable subvolume %s",
+				(bch2_bkey_val_to_text(&buf, c, s.s_c),
+				 buf.buf))) {
+			ret = reattach_subvol(trans, s);
+			break;
+		}
+	}
+fsck_err:
+err:
+	printbuf_exit(&buf);
+	darray_exit(&subvol_path);
+	bch2_trans_iter_exit(trans, &parent_iter);
+	return ret;
+}
+
+int bch2_check_subvolume_structure(struct bch_fs *c)
+{
+	int ret = bch2_trans_run(c,
+		for_each_btree_key_commit(trans, iter,
+				BTREE_ID_subvolumes, POS_MIN, BTREE_ITER_prefetch, k,
+				NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
+			check_subvol_path(trans, &iter, k)));
+	bch_err_fn(c, ret);
+	return ret;
+}
+
 struct pathbuf_entry {
 	u64	inum;
 	u32	snapshot;
@@ -1952,89 +2376,71 @@ static bool path_is_dup(pathbuf *p, u64 inum, u32 snapshot)
 	return false;
 }
 
-static int path_down(struct bch_fs *c, pathbuf *p,
-		     u64 inum, u32 snapshot)
-{
-	int ret = darray_push(p, ((struct pathbuf_entry) {
-		.inum		= inum,
-		.snapshot	= snapshot,
-	}));
-
-	if (ret)
-		bch_err(c, "fsck: error allocating memory for pathbuf, size %zu",
-			p->size);
-	return ret;
-}
-
 /*
- * Check that a given inode is reachable from the root:
+ * Check that a given inode is reachable from its subvolume root - we already
+ * verified subvolume connectivity:
  *
  * XXX: we should also be verifying that inodes are in the right subvolumes
  */
-static int check_path(struct btree_trans *trans,
-		      pathbuf *p,
-		      struct bch_inode_unpacked *inode,
-		      u32 snapshot)
+static int check_path(struct btree_trans *trans, pathbuf *p, struct bkey_s_c inode_k)
 {
 	struct bch_fs *c = trans->c;
+	struct btree_iter inode_iter = {};
+	struct bch_inode_unpacked inode;
+	struct printbuf buf = PRINTBUF;
+	u32 snapshot = inode_k.k->p.snapshot;
 	int ret = 0;
 
-	snapshot = bch2_snapshot_equiv(c, snapshot);
 	p->nr = 0;
 
-	while (!(inode->bi_inum == BCACHEFS_ROOT_INO &&
-		 inode->bi_subvol == BCACHEFS_ROOT_SUBVOL)) {
+	BUG_ON(bch2_inode_unpack(inode_k, &inode));
+
+	while (!inode.bi_subvol) {
 		struct btree_iter dirent_iter;
 		struct bkey_s_c_dirent d;
 		u32 parent_snapshot = snapshot;
 
-		if (inode->bi_subvol) {
-			u64 inum;
-
-			ret = subvol_lookup(trans, inode->bi_parent_subvol,
-					    &parent_snapshot, &inum);
-			if (ret)
-				break;
-		}
-
-		d = dirent_get_by_pos(trans, &dirent_iter,
-				      SPOS(inode->bi_dir, inode->bi_dir_offset,
-					   parent_snapshot));
+		d = inode_get_dirent(trans, &dirent_iter, &inode, &parent_snapshot);
 		ret = bkey_err(d.s_c);
 		if (ret && !bch2_err_matches(ret, ENOENT))
 			break;
 
-		if (!ret && !dirent_points_to_inode(d, inode)) {
+		if (!ret && !dirent_points_to_inode(d, &inode)) {
 			bch2_trans_iter_exit(trans, &dirent_iter);
 			ret = -BCH_ERR_ENOENT_dirent_doesnt_match_inode;
 		}
 
 		if (bch2_err_matches(ret, ENOENT)) {
-			if (fsck_err(c,  inode_unreachable,
-				     "unreachable inode %llu:%u, type %s nlink %u backptr %llu:%llu",
-				     inode->bi_inum, snapshot,
-				     bch2_d_type_str(inode_d_type(inode)),
-				     inode->bi_nlink,
-				     inode->bi_dir,
-				     inode->bi_dir_offset))
-				ret = reattach_inode(trans, inode, snapshot);
-			break;
+			ret = 0;
+			if (fsck_err(c, inode_unreachable,
+				     "unreachable inode\n%s",
+				     (printbuf_reset(&buf),
+				      bch2_bkey_val_to_text(&buf, c, inode_k),
+				      buf.buf)))
+				ret = reattach_inode(trans, &inode, snapshot);
+			goto out;
 		}
 
 		bch2_trans_iter_exit(trans, &dirent_iter);
 
-		if (!S_ISDIR(inode->bi_mode))
+		if (!S_ISDIR(inode.bi_mode))
 			break;
 
-		ret = path_down(c, p, inode->bi_inum, snapshot);
-		if (ret) {
-			bch_err(c, "memory allocation failure");
+		ret = darray_push(p, ((struct pathbuf_entry) {
+			.inum		= inode.bi_inum,
+			.snapshot	= snapshot,
+		}));
+		if (ret)
 			return ret;
-		}
 
 		snapshot = parent_snapshot;
 
-		ret = lookup_inode(trans, inode->bi_dir, inode, &snapshot);
+		bch2_trans_iter_exit(trans, &inode_iter);
+		inode_k = bch2_bkey_get_iter(trans, &inode_iter, BTREE_ID_inodes,
+					     SPOS(0, inode.bi_dir, snapshot), 0);
+		ret = bkey_err(inode_k) ?:
+			!bkey_is_inode(inode_k.k) ? -BCH_ERR_ENOENT_inode
+			: bch2_inode_unpack(inode_k, &inode);
 		if (ret) {
 			/* Should have been caught in dirents pass */
 			if (!bch2_err_matches(ret, BCH_ERR_transaction_restart))
@@ -2042,30 +2448,32 @@ static int check_path(struct btree_trans *trans,
 			break;
 		}
 
-		if (path_is_dup(p, inode->bi_inum, snapshot)) {
+		snapshot = inode_k.k->p.snapshot;
+
+		if (path_is_dup(p, inode.bi_inum, snapshot)) {
 			/* XXX print path */
 			bch_err(c, "directory structure loop");
 
 			darray_for_each(*p, i)
 				pr_err("%llu:%u", i->inum, i->snapshot);
-			pr_err("%llu:%u", inode->bi_inum, snapshot);
+			pr_err("%llu:%u", inode.bi_inum, snapshot);
 
-			if (!fsck_err(c, dir_loop, "directory structure loop"))
-				return 0;
-
-			ret = remove_backpointer(trans, inode);
-			if (ret && !bch2_err_matches(ret, BCH_ERR_transaction_restart))
+			if (fsck_err(c, dir_loop, "directory structure loop")) {
+				ret = remove_backpointer(trans, &inode);
 				bch_err_msg(c, ret, "removing dirent");
-			if (ret)
-				break;
+				if (ret)
+					break;
 
-			ret = reattach_inode(trans, inode, snapshot);
-			if (ret && !bch2_err_matches(ret, BCH_ERR_transaction_restart))
-				bch_err_msg(c, ret, "reattaching inode %llu", inode->bi_inum);
+				ret = reattach_inode(trans, &inode, snapshot);
+				bch_err_msg(c, ret, "reattaching inode %llu", inode.bi_inum);
+			}
 			break;
 		}
 	}
+out:
 fsck_err:
+	bch2_trans_iter_exit(trans, &inode_iter);
+	printbuf_exit(&buf);
 	bch_err_fn(c, ret);
 	return ret;
 }
@@ -2077,25 +2485,22 @@ fsck_err:
  */
 int bch2_check_directory_structure(struct bch_fs *c)
 {
-	struct bch_inode_unpacked u;
 	pathbuf path = { 0, };
 	int ret;
 
 	ret = bch2_trans_run(c,
 		for_each_btree_key_commit(trans, iter, BTREE_ID_inodes, POS_MIN,
-					  BTREE_ITER_INTENT|
-					  BTREE_ITER_PREFETCH|
-					  BTREE_ITER_ALL_SNAPSHOTS, k,
+					  BTREE_ITER_intent|
+					  BTREE_ITER_prefetch|
+					  BTREE_ITER_all_snapshots, k,
 					  NULL, NULL, BCH_TRANS_COMMIT_no_enospc, ({
 			if (!bkey_is_inode(k.k))
 				continue;
 
-			BUG_ON(bch2_inode_unpack(k, &u));
-
-			if (u.bi_flags & BCH_INODE_unlinked)
+			if (bch2_inode_flags(k) & BCH_INODE_unlinked)
 				continue;
 
-			check_path(trans, &path, &u, iter.pos.snapshot);
+			check_path(trans, &path, k);
 		})));
 	darray_exit(&path);
 
@@ -2187,9 +2592,9 @@ static int check_nlinks_find_hardlinks(struct bch_fs *c,
 	int ret = bch2_trans_run(c,
 		for_each_btree_key(trans, iter, BTREE_ID_inodes,
 				   POS(0, start),
-				   BTREE_ITER_INTENT|
-				   BTREE_ITER_PREFETCH|
-				   BTREE_ITER_ALL_SNAPSHOTS, k, ({
+				   BTREE_ITER_intent|
+				   BTREE_ITER_prefetch|
+				   BTREE_ITER_all_snapshots, k, ({
 			if (!bkey_is_inode(k.k))
 				continue;
 
@@ -2230,9 +2635,9 @@ static int check_nlinks_walk_dirents(struct bch_fs *c, struct nlink_table *links
 
 	int ret = bch2_trans_run(c,
 		for_each_btree_key(trans, iter, BTREE_ID_dirents, POS_MIN,
-				   BTREE_ITER_INTENT|
-				   BTREE_ITER_PREFETCH|
-				   BTREE_ITER_ALL_SNAPSHOTS, k, ({
+				   BTREE_ITER_intent|
+				   BTREE_ITER_prefetch|
+				   BTREE_ITER_all_snapshots, k, ({
 			ret = snapshots_seen_update(c, &s, iter.btree_id, k.k->p);
 			if (ret)
 				break;
@@ -2243,8 +2648,7 @@ static int check_nlinks_walk_dirents(struct bch_fs *c, struct nlink_table *links
 				if (d.v->d_type != DT_DIR &&
 				    d.v->d_type != DT_SUBVOL)
 					inc_link(c, &s, links, range_start, range_end,
-						 le64_to_cpu(d.v->d_inum),
-						 bch2_snapshot_equiv(c, d.k->p.snapshot));
+						 le64_to_cpu(d.v->d_inum), d.k->p.snapshot);
 			}
 			0;
 		})));
@@ -2291,7 +2695,7 @@ static int check_nlinks_update_inode(struct btree_trans *trans, struct btree_ite
 			u.bi_inum, bch2_d_types[mode_to_type(u.bi_mode)],
 			bch2_inode_nlink_get(&u), link->count)) {
 		bch2_inode_nlink_set(&u, link->count);
-		ret = __write_inode(trans, &u, k.k->p.snapshot);
+		ret = __bch2_fsck_write_inode(trans, &u, k.k->p.snapshot);
 	}
 fsck_err:
 	return ret;
@@ -2307,7 +2711,7 @@ static int check_nlinks_update_hardlinks(struct bch_fs *c,
 	int ret = bch2_trans_run(c,
 		for_each_btree_key_commit(trans, iter, BTREE_ID_inodes,
 				POS(0, range_start),
-				BTREE_ITER_INTENT|BTREE_ITER_PREFETCH|BTREE_ITER_ALL_SNAPSHOTS, k,
+				BTREE_ITER_intent|BTREE_ITER_prefetch|BTREE_ITER_all_snapshots, k,
 				NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
 			check_nlinks_update_inode(trans, &iter, k, links, &idx, range_end)));
 	if (ret < 0) {
@@ -2375,7 +2779,7 @@ static int fix_reflink_p_key(struct btree_trans *trans, struct btree_iter *iter,
 	u->v.front_pad	= 0;
 	u->v.back_pad	= 0;
 
-	return bch2_trans_update(trans, iter, &u->k_i, BTREE_TRIGGER_NORUN);
+	return bch2_trans_update(trans, iter, &u->k_i, BTREE_TRIGGER_norun);
 }
 
 int bch2_fix_reflink_p(struct bch_fs *c)
@@ -2386,8 +2790,8 @@ int bch2_fix_reflink_p(struct bch_fs *c)
 	int ret = bch2_trans_run(c,
 		for_each_btree_key_commit(trans, iter,
 				BTREE_ID_extents, POS_MIN,
-				BTREE_ITER_INTENT|BTREE_ITER_PREFETCH|
-				BTREE_ITER_ALL_SNAPSHOTS, k,
+				BTREE_ITER_intent|BTREE_ITER_prefetch|
+				BTREE_ITER_all_snapshots, k,
 				NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
 			fix_reflink_p_key(trans, &iter, k)));
 	bch_err_fn(c, ret);
diff --git a/fs/bcachefs/fsck.h b/fs/bcachefs/fsck.h
index da991e8cf2..a4ef942717 100644
--- a/fs/bcachefs/fsck.h
+++ b/fs/bcachefs/fsck.h
@@ -8,6 +8,7 @@ int bch2_check_indirect_extents(struct bch_fs *);
 int bch2_check_dirents(struct bch_fs *);
 int bch2_check_xattrs(struct bch_fs *);
 int bch2_check_root(struct bch_fs *);
+int bch2_check_subvolume_structure(struct bch_fs *);
 int bch2_check_directory_structure(struct bch_fs *);
 int bch2_check_nlinks(struct bch_fs *);
 int bch2_fix_reflink_p(struct bch_fs *);
diff --git a/fs/bcachefs/inode.c b/fs/bcachefs/inode.c
index 086f0090b0..aafa79fa63 100644
--- a/fs/bcachefs/inode.c
+++ b/fs/bcachefs/inode.c
@@ -324,7 +324,7 @@ int bch2_inode_unpack(struct bkey_s_c k,
 	return bch2_inode_unpack_slowpath(k, unpacked);
 }
 
-static int bch2_inode_peek_nowarn(struct btree_trans *trans,
+int bch2_inode_peek_nowarn(struct btree_trans *trans,
 		    struct btree_iter *iter,
 		    struct bch_inode_unpacked *inode,
 		    subvol_inum inum, unsigned flags)
@@ -339,7 +339,7 @@ static int bch2_inode_peek_nowarn(struct btree_trans *trans,
 
 	k = bch2_bkey_get_iter(trans, iter, BTREE_ID_inodes,
 			       SPOS(0, inum.inum, snapshot),
-			       flags|BTREE_ITER_CACHED);
+			       flags|BTREE_ITER_cached);
 	ret = bkey_err(k);
 	if (ret)
 		return ret;
@@ -371,7 +371,7 @@ int bch2_inode_peek(struct btree_trans *trans,
 int bch2_inode_write_flags(struct btree_trans *trans,
 		     struct btree_iter *iter,
 		     struct bch_inode_unpacked *inode,
-		     enum btree_update_flags flags)
+		     enum btree_iter_update_trigger_flags flags)
 {
 	struct bkey_inode_buf *inode_p;
 
@@ -384,6 +384,34 @@ int bch2_inode_write_flags(struct btree_trans *trans,
 	return bch2_trans_update(trans, iter, &inode_p->inode.k_i, flags);
 }
 
+int __bch2_fsck_write_inode(struct btree_trans *trans,
+			 struct bch_inode_unpacked *inode,
+			 u32 snapshot)
+{
+	struct bkey_inode_buf *inode_p =
+		bch2_trans_kmalloc(trans, sizeof(*inode_p));
+
+	if (IS_ERR(inode_p))
+		return PTR_ERR(inode_p);
+
+	bch2_inode_pack(inode_p, inode);
+	inode_p->inode.k.p.snapshot = snapshot;
+
+	return bch2_btree_insert_nonextent(trans, BTREE_ID_inodes,
+				&inode_p->inode.k_i,
+				BTREE_UPDATE_internal_snapshot_node);
+}
+
+int bch2_fsck_write_inode(struct btree_trans *trans,
+			    struct bch_inode_unpacked *inode,
+			    u32 snapshot)
+{
+	int ret = commit_do(trans, NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
+			    __bch2_fsck_write_inode(trans, inode, snapshot));
+	bch_err_fn(trans->c, ret);
+	return ret;
+}
+
 struct bkey_i *bch2_inode_to_v3(struct btree_trans *trans, struct bkey_i *k)
 {
 	struct bch_inode_unpacked u;
@@ -445,7 +473,7 @@ fsck_err:
 }
 
 int bch2_inode_invalid(struct bch_fs *c, struct bkey_s_c k,
-		       enum bkey_invalid_flags flags,
+		       enum bch_validate_flags flags,
 		       struct printbuf *err)
 {
 	struct bkey_s_c_inode inode = bkey_s_c_to_inode(k);
@@ -462,7 +490,7 @@ fsck_err:
 }
 
 int bch2_inode_v2_invalid(struct bch_fs *c, struct bkey_s_c k,
-			  enum bkey_invalid_flags flags,
+			  enum bch_validate_flags flags,
 			  struct printbuf *err)
 {
 	struct bkey_s_c_inode_v2 inode = bkey_s_c_to_inode_v2(k);
@@ -479,7 +507,7 @@ fsck_err:
 }
 
 int bch2_inode_v3_invalid(struct bch_fs *c, struct bkey_s_c k,
-			  enum bkey_invalid_flags flags,
+			  enum bch_validate_flags flags,
 			  struct printbuf *err)
 {
 	struct bkey_s_c_inode_v3 inode = bkey_s_c_to_inode_v3(k);
@@ -507,29 +535,19 @@ static void __bch2_inode_unpacked_to_text(struct printbuf *out,
 					  struct bch_inode_unpacked *inode)
 {
 	printbuf_indent_add(out, 2);
-	prt_printf(out, "mode=%o", inode->bi_mode);
-	prt_newline(out);
+	prt_printf(out, "mode=%o\n", inode->bi_mode);
 
 	prt_str(out, "flags=");
 	prt_bitflags(out, bch2_inode_flag_strs, inode->bi_flags & ((1U << 20) - 1));
-	prt_printf(out, " (%x)", inode->bi_flags);
-	prt_newline(out);
+	prt_printf(out, " (%x)\n", inode->bi_flags);
 
-	prt_printf(out, "journal_seq=%llu", inode->bi_journal_seq);
-	prt_newline(out);
-
-	prt_printf(out, "bi_size=%llu", inode->bi_size);
-	prt_newline(out);
-
-	prt_printf(out, "bi_sectors=%llu", inode->bi_sectors);
-	prt_newline(out);
-
-	prt_newline(out);
-	prt_printf(out, "bi_version=%llu", inode->bi_version);
+	prt_printf(out, "journal_seq=%llu\n",	inode->bi_journal_seq);
+	prt_printf(out, "bi_size=%llu\n",	inode->bi_size);
+	prt_printf(out, "bi_sectors=%llu\n",	inode->bi_sectors);
+	prt_printf(out, "bi_version=%llu\n",	inode->bi_version);
 
 #define x(_name, _bits)						\
-	prt_printf(out, #_name "=%llu", (u64) inode->_name);	\
-	prt_newline(out);
+	prt_printf(out, #_name "=%llu\n", (u64) inode->_name);
 	BCH_INODE_FIELDS_v3()
 #undef  x
 	printbuf_indent_sub(out, 2);
@@ -576,11 +594,11 @@ int bch2_trigger_inode(struct btree_trans *trans,
 		       enum btree_id btree_id, unsigned level,
 		       struct bkey_s_c old,
 		       struct bkey_s new,
-		       unsigned flags)
+		       enum btree_iter_update_trigger_flags flags)
 {
-	s64 nr = bkey_is_inode(new.k) - bkey_is_inode(old.k);
+	s64 nr = (s64) bkey_is_inode(new.k) - (s64) bkey_is_inode(old.k);
 
-	if (flags & BTREE_TRIGGER_TRANSACTIONAL) {
+	if (flags & BTREE_TRIGGER_transactional) {
 		if (nr) {
 			int ret = bch2_replicas_deltas_realloc(trans, 0);
 			if (ret)
@@ -592,19 +610,20 @@ int bch2_trigger_inode(struct btree_trans *trans,
 		bool old_deleted = bkey_is_deleted_inode(old);
 		bool new_deleted = bkey_is_deleted_inode(new.s_c);
 		if (old_deleted != new_deleted) {
-			int ret = bch2_btree_bit_mod(trans, BTREE_ID_deleted_inodes, new.k->p, new_deleted);
+			int ret = bch2_btree_bit_mod_buffered(trans, BTREE_ID_deleted_inodes,
+							      new.k->p, new_deleted);
 			if (ret)
 				return ret;
 		}
 	}
 
-	if ((flags & BTREE_TRIGGER_ATOMIC) && (flags & BTREE_TRIGGER_INSERT)) {
+	if ((flags & BTREE_TRIGGER_atomic) && (flags & BTREE_TRIGGER_insert)) {
 		BUG_ON(!trans->journal_res.seq);
 
 		bkey_s_to_inode_v3(new).v->bi_journal_seq = cpu_to_le64(trans->journal_res.seq);
 	}
 
-	if (flags & BTREE_TRIGGER_GC) {
+	if (flags & BTREE_TRIGGER_gc) {
 		struct bch_fs *c = trans->c;
 
 		percpu_down_read(&c->mark_lock);
@@ -616,7 +635,7 @@ int bch2_trigger_inode(struct btree_trans *trans,
 }
 
 int bch2_inode_generation_invalid(struct bch_fs *c, struct bkey_s_c k,
-				  enum bkey_invalid_flags flags,
+				  enum bch_validate_flags flags,
 				  struct printbuf *err)
 {
 	int ret = 0;
@@ -733,8 +752,8 @@ int bch2_inode_create(struct btree_trans *trans,
 
 	pos = start;
 	bch2_trans_iter_init(trans, iter, BTREE_ID_inodes, POS(0, pos),
-			     BTREE_ITER_ALL_SNAPSHOTS|
-			     BTREE_ITER_INTENT);
+			     BTREE_ITER_all_snapshots|
+			     BTREE_ITER_intent);
 again:
 	while ((k = bch2_btree_iter_peek(iter)).k &&
 	       !(ret = bkey_err(k)) &&
@@ -795,7 +814,7 @@ static int bch2_inode_delete_keys(struct btree_trans *trans,
 	 * extent iterator:
 	 */
 	bch2_trans_iter_init(trans, &iter, id, POS(inum.inum, 0),
-			     BTREE_ITER_INTENT);
+			     BTREE_ITER_intent);
 
 	while (1) {
 		bch2_trans_begin(trans);
@@ -817,7 +836,7 @@ static int bch2_inode_delete_keys(struct btree_trans *trans,
 		bkey_init(&delete.k);
 		delete.k.p = iter.pos;
 
-		if (iter.flags & BTREE_ITER_IS_EXTENTS)
+		if (iter.flags & BTREE_ITER_is_extents)
 			bch2_key_resize(&delete.k,
 					bpos_min(end, k.k->p).offset -
 					iter.pos.offset);
@@ -866,7 +885,7 @@ retry:
 
 	k = bch2_bkey_get_iter(trans, &iter, BTREE_ID_inodes,
 			       SPOS(0, inum.inum, snapshot),
-			       BTREE_ITER_INTENT|BTREE_ITER_CACHED);
+			       BTREE_ITER_intent|BTREE_ITER_cached);
 	ret = bkey_err(k);
 	if (ret)
 		goto err;
@@ -1026,7 +1045,7 @@ retry:
 	bch2_trans_begin(trans);
 
 	k = bch2_bkey_get_iter(trans, &iter, BTREE_ID_inodes,
-			       SPOS(0, inum, snapshot), BTREE_ITER_INTENT);
+			       SPOS(0, inum, snapshot), BTREE_ITER_intent);
 	ret = bkey_err(k);
 	if (ret)
 		goto err;
@@ -1071,7 +1090,7 @@ static int may_delete_deleted_inode(struct btree_trans *trans,
 	struct bch_inode_unpacked inode;
 	int ret;
 
-	k = bch2_bkey_get_iter(trans, &inode_iter, BTREE_ID_inodes, pos, BTREE_ITER_CACHED);
+	k = bch2_bkey_get_iter(trans, &inode_iter, BTREE_ID_inodes, pos, BTREE_ITER_cached);
 	ret = bkey_err(k);
 	if (ret)
 		return ret;
@@ -1088,8 +1107,9 @@ static int may_delete_deleted_inode(struct btree_trans *trans,
 		goto out;
 
 	if (S_ISDIR(inode.bi_mode)) {
-		ret = bch2_empty_dir_snapshot(trans, pos.offset, pos.snapshot);
-		if (fsck_err_on(ret == -ENOTEMPTY, c, deleted_inode_is_dir,
+		ret = bch2_empty_dir_snapshot(trans, pos.offset, 0, pos.snapshot);
+		if (fsck_err_on(bch2_err_matches(ret, ENOTEMPTY),
+				c, deleted_inode_is_dir,
 				"non empty directory %llu:%u in deleted_inodes btree",
 				pos.offset, pos.snapshot))
 			goto delete;
@@ -1122,7 +1142,7 @@ static int may_delete_deleted_inode(struct btree_trans *trans,
 		inode.bi_flags &= ~BCH_INODE_unlinked;
 
 		ret = bch2_inode_write_flags(trans, &inode_iter, &inode,
-					     BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
+					     BTREE_UPDATE_internal_snapshot_node);
 		bch_err_msg(c, ret, "clearing inode unlinked flag");
 		if (ret)
 			goto out;
@@ -1141,7 +1161,7 @@ fsck_err:
 	bch2_trans_iter_exit(trans, &inode_iter);
 	return ret;
 delete:
-	ret = bch2_btree_bit_mod(trans, BTREE_ID_deleted_inodes, pos, false);
+	ret = bch2_btree_bit_mod_buffered(trans, BTREE_ID_deleted_inodes, pos, false);
 	goto out;
 }
 
@@ -1151,6 +1171,15 @@ int bch2_delete_dead_inodes(struct bch_fs *c)
 	bool need_another_pass;
 	int ret;
 again:
+	/*
+	 * if we ran check_inodes() unlinked inodes will have already been
+	 * cleaned up but the write buffer will be out of sync; therefore we
+	 * alway need a write buffer flush
+	 */
+	ret = bch2_btree_write_buffer_flush_sync(trans);
+	if (ret)
+		goto err;
+
 	need_another_pass = false;
 
 	/*
@@ -1160,7 +1189,7 @@ again:
 	 * flushed and we'd spin:
 	 */
 	ret = for_each_btree_key_commit(trans, iter, BTREE_ID_deleted_inodes, POS_MIN,
-					BTREE_ITER_PREFETCH|BTREE_ITER_ALL_SNAPSHOTS, k,
+					BTREE_ITER_prefetch|BTREE_ITER_all_snapshots, k,
 					NULL, NULL, BCH_TRANS_COMMIT_no_enospc, ({
 		ret = may_delete_deleted_inode(trans, &iter, k.k->p, &need_another_pass);
 		if (ret > 0) {
@@ -1183,12 +1212,8 @@ again:
 		ret;
 	}));
 
-	if (!ret && need_another_pass) {
-		ret = bch2_btree_write_buffer_flush_sync(trans);
-		if (ret)
-			goto err;
+	if (!ret && need_another_pass)
 		goto again;
-	}
 err:
 	bch2_trans_put(trans);
 	return ret;
diff --git a/fs/bcachefs/inode.h b/fs/bcachefs/inode.h
index b63f312581..679f5f5e5d 100644
--- a/fs/bcachefs/inode.h
+++ b/fs/bcachefs/inode.h
@@ -6,19 +6,20 @@
 #include "bkey_methods.h"
 #include "opts.h"
 
-enum bkey_invalid_flags;
+enum bch_validate_flags;
 extern const char * const bch2_inode_opts[];
 
 int bch2_inode_invalid(struct bch_fs *, struct bkey_s_c,
-		       enum bkey_invalid_flags, struct printbuf *);
+		       enum bch_validate_flags, struct printbuf *);
 int bch2_inode_v2_invalid(struct bch_fs *, struct bkey_s_c,
-			  enum bkey_invalid_flags, struct printbuf *);
+			  enum bch_validate_flags, struct printbuf *);
 int bch2_inode_v3_invalid(struct bch_fs *, struct bkey_s_c,
-			  enum bkey_invalid_flags, struct printbuf *);
+			  enum bch_validate_flags, struct printbuf *);
 void bch2_inode_to_text(struct printbuf *, struct bch_fs *, struct bkey_s_c);
 
 int bch2_trigger_inode(struct btree_trans *, enum btree_id, unsigned,
-			  struct bkey_s_c, struct bkey_s, unsigned);
+		       struct bkey_s_c, struct bkey_s,
+		       enum btree_iter_update_trigger_flags);
 
 #define bch2_bkey_ops_inode ((struct bkey_ops) {	\
 	.key_invalid	= bch2_inode_invalid,		\
@@ -49,7 +50,7 @@ static inline bool bkey_is_inode(const struct bkey *k)
 }
 
 int bch2_inode_generation_invalid(struct bch_fs *, struct bkey_s_c,
-				  enum bkey_invalid_flags, struct printbuf *);
+				  enum bch_validate_flags, struct printbuf *);
 void bch2_inode_generation_to_text(struct printbuf *, struct bch_fs *, struct bkey_s_c);
 
 #define bch2_bkey_ops_inode_generation ((struct bkey_ops) {	\
@@ -95,11 +96,13 @@ struct bkey_i *bch2_inode_to_v3(struct btree_trans *, struct bkey_i *);
 
 void bch2_inode_unpacked_to_text(struct printbuf *, struct bch_inode_unpacked *);
 
+int bch2_inode_peek_nowarn(struct btree_trans *, struct btree_iter *,
+		    struct bch_inode_unpacked *, subvol_inum, unsigned);
 int bch2_inode_peek(struct btree_trans *, struct btree_iter *,
 		    struct bch_inode_unpacked *, subvol_inum, unsigned);
 
 int bch2_inode_write_flags(struct btree_trans *, struct btree_iter *,
-		     struct bch_inode_unpacked *, enum btree_update_flags);
+		     struct bch_inode_unpacked *, enum btree_iter_update_trigger_flags);
 
 static inline int bch2_inode_write(struct btree_trans *trans,
 		     struct btree_iter *iter,
@@ -108,6 +111,9 @@ static inline int bch2_inode_write(struct btree_trans *trans,
 	return bch2_inode_write_flags(trans, iter, inode, 0);
 }
 
+int __bch2_fsck_write_inode(struct btree_trans *, struct bch_inode_unpacked *, u32);
+int bch2_fsck_write_inode(struct btree_trans *, struct bch_inode_unpacked *, u32);
+
 void bch2_inode_init_early(struct bch_fs *,
 			   struct bch_inode_unpacked *);
 void bch2_inode_init_late(struct bch_inode_unpacked *, u64,
@@ -172,6 +178,20 @@ static inline u8 inode_d_type(struct bch_inode_unpacked *inode)
 	return inode->bi_subvol ? DT_SUBVOL : mode_to_type(inode->bi_mode);
 }
 
+static inline u32 bch2_inode_flags(struct bkey_s_c k)
+{
+	switch (k.k->type) {
+	case KEY_TYPE_inode:
+		return le32_to_cpu(bkey_s_c_to_inode(k).v->bi_flags);
+	case KEY_TYPE_inode_v2:
+		return le64_to_cpu(bkey_s_c_to_inode_v2(k).v->bi_flags);
+	case KEY_TYPE_inode_v3:
+		return le64_to_cpu(bkey_s_c_to_inode_v3(k).v->bi_flags);
+	default:
+		return 0;
+	}
+}
+
 /* i_nlink: */
 
 static inline unsigned nlink_bias(umode_t mode)
@@ -201,6 +221,14 @@ static inline void bch2_inode_nlink_set(struct bch_inode_unpacked *bi,
 int bch2_inode_nlink_inc(struct bch_inode_unpacked *);
 void bch2_inode_nlink_dec(struct btree_trans *, struct bch_inode_unpacked *);
 
+static inline bool bch2_inode_should_have_bp(struct bch_inode_unpacked *inode)
+{
+	bool inode_has_bp = inode->bi_dir || inode->bi_dir_offset;
+
+	return S_ISDIR(inode->bi_mode) ||
+		(!inode->bi_nlink && inode_has_bp);
+}
+
 struct bch_opts bch2_inode_opts_to_opts(struct bch_inode_unpacked *);
 void bch2_inode_opts_get(struct bch_io_opts *, struct bch_fs *,
 			 struct bch_inode_unpacked *);
diff --git a/fs/bcachefs/io_misc.c b/fs/bcachefs/io_misc.c
index 1baf78594c..4583c9386e 100644
--- a/fs/bcachefs/io_misc.c
+++ b/fs/bcachefs/io_misc.c
@@ -125,7 +125,7 @@ err_noprint:
 	bch2_bkey_buf_exit(&old, c);
 
 	if (closure_nr_remaining(&cl) != 1) {
-		bch2_trans_unlock(trans);
+		bch2_trans_unlock_long(trans);
 		closure_sync(&cl);
 	}
 
@@ -198,7 +198,7 @@ int bch2_fpunch(struct bch_fs *c, subvol_inum inum, u64 start, u64 end,
 
 	bch2_trans_iter_init(trans, &iter, BTREE_ID_extents,
 			     POS(inum.inum, start),
-			     BTREE_ITER_INTENT);
+			     BTREE_ITER_intent);
 
 	ret = bch2_fpunch_at(trans, &iter, inum, end, i_sectors_delta);
 
@@ -230,7 +230,7 @@ static int truncate_set_isize(struct btree_trans *trans,
 	struct bch_inode_unpacked inode_u;
 	int ret;
 
-	ret   = bch2_inode_peek(trans, &iter, &inode_u, inum, BTREE_ITER_INTENT) ?:
+	ret   = bch2_inode_peek(trans, &iter, &inode_u, inum, BTREE_ITER_intent) ?:
 		(inode_u.bi_size = new_i_size, 0) ?:
 		bch2_inode_write(trans, &iter, &inode_u);
 
@@ -256,7 +256,7 @@ static int __bch2_resume_logged_op_truncate(struct btree_trans *trans,
 
 	bch2_trans_iter_init(trans, &fpunch_iter, BTREE_ID_extents,
 			     POS(inum.inum, round_up(new_i_size, block_bytes(c)) >> 9),
-			     BTREE_ITER_INTENT);
+			     BTREE_ITER_intent);
 	ret = bch2_fpunch_at(trans, &fpunch_iter, inum, U64_MAX, i_sectors_delta);
 	bch2_trans_iter_exit(trans, &fpunch_iter);
 
@@ -264,6 +264,7 @@ static int __bch2_resume_logged_op_truncate(struct btree_trans *trans,
 		ret = 0;
 err:
 	bch2_logged_op_finish(trans, op_k);
+	bch_err_fn(c, ret);
 	return ret;
 }
 
@@ -316,7 +317,7 @@ static int adjust_i_size(struct btree_trans *trans, subvol_inum inum, u64 offset
 	offset	<<= 9;
 	len	<<= 9;
 
-	ret = bch2_inode_peek(trans, &iter, &inode_u, inum, BTREE_ITER_INTENT);
+	ret = bch2_inode_peek(trans, &iter, &inode_u, inum, BTREE_ITER_intent);
 	if (ret)
 		return ret;
 
@@ -364,7 +365,7 @@ static int __bch2_resume_logged_op_finsert(struct btree_trans *trans,
 
 	bch2_trans_iter_init(trans, &iter, BTREE_ID_extents,
 			     POS(inum.inum, 0),
-			     BTREE_ITER_INTENT);
+			     BTREE_ITER_intent);
 
 	switch (op->v.state) {
 case LOGGED_OP_FINSERT_start:
@@ -476,6 +477,7 @@ case LOGGED_OP_FINSERT_finish:
 	break;
 	}
 err:
+	bch_err_fn(c, ret);
 	bch2_logged_op_finish(trans, op_k);
 	bch2_trans_iter_exit(trans, &iter);
 	return ret;
diff --git a/fs/bcachefs/io_read.c b/fs/bcachefs/io_read.c
index 3c574d8873..ebf39ef72f 100644
--- a/fs/bcachefs/io_read.c
+++ b/fs/bcachefs/io_read.c
@@ -84,9 +84,10 @@ struct promote_op {
 };
 
 static const struct rhashtable_params bch_promote_params = {
-	.head_offset	= offsetof(struct promote_op, hash),
-	.key_offset	= offsetof(struct promote_op, pos),
-	.key_len	= sizeof(struct bpos),
+	.head_offset		= offsetof(struct promote_op, hash),
+	.key_offset		= offsetof(struct promote_op, pos),
+	.key_len		= sizeof(struct bpos),
+	.automatic_shrinking	= true,
 };
 
 static inline int should_promote(struct bch_fs *c, struct bkey_s_c k,
@@ -174,7 +175,7 @@ static struct promote_op *__promote_alloc(struct btree_trans *trans,
 	if (!bch2_write_ref_tryget(c, BCH_WRITE_REF_promote))
 		return ERR_PTR(-BCH_ERR_nopromote_no_writes);
 
-	op = kzalloc(sizeof(*op) + sizeof(struct bio_vec) * pages, GFP_KERNEL);
+	op = kzalloc(struct_size(op, bi_inline_vecs, pages), GFP_KERNEL);
 	if (!op) {
 		ret = -BCH_ERR_nopromote_enomem;
 		goto err;
@@ -378,7 +379,7 @@ static void bch2_read_retry_nodecode(struct bch_fs *c, struct bch_read_bio *rbio
 	bch2_bkey_buf_init(&sk);
 
 	bch2_trans_iter_init(trans, &iter, rbio->data_btree,
-			     rbio->read_pos, BTREE_ITER_SLOTS);
+			     rbio->read_pos, BTREE_ITER_slots);
 retry:
 	rbio->bio.bi_status = 0;
 
@@ -388,7 +389,6 @@ retry:
 
 	bch2_bkey_buf_reassemble(&sk, c, k);
 	k = bkey_i_to_s_c(sk.k);
-	bch2_trans_unlock(trans);
 
 	if (!bch2_bkey_matches_ptr(c, k,
 				   rbio->pick.ptr,
@@ -487,7 +487,7 @@ static int __bch2_rbio_narrow_crcs(struct btree_trans *trans,
 		return 0;
 
 	k = bch2_bkey_get_iter(trans, &iter, rbio->data_btree, rbio->data_pos,
-			       BTREE_ITER_SLOTS|BTREE_ITER_INTENT);
+			       BTREE_ITER_slots|BTREE_ITER_intent);
 	if ((ret = bkey_err(k)))
 		goto out;
 
@@ -523,7 +523,7 @@ static int __bch2_rbio_narrow_crcs(struct btree_trans *trans,
 		goto out;
 
 	ret = bch2_trans_update(trans, &iter, new,
-				BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
+				BTREE_UPDATE_internal_snapshot_node);
 out:
 	bch2_trans_iter_exit(trans, &iter);
 	return ret;
@@ -541,7 +541,6 @@ static void __bch2_read_endio(struct work_struct *work)
 	struct bch_read_bio *rbio =
 		container_of(work, struct bch_read_bio, work);
 	struct bch_fs *c	= rbio->c;
-	struct bch_dev *ca	= bch_dev_bkey_exists(c, rbio->pick.ptr.dev);
 	struct bio *src		= &rbio->bio;
 	struct bio *dst		= &bch2_rbio_parent(rbio)->bio;
 	struct bvec_iter dst_iter = rbio->bvec_iter;
@@ -647,13 +646,15 @@ csum_err:
 	prt_str(&buf, "data ");
 	bch2_csum_err_msg(&buf, crc.csum_type, rbio->pick.crc.csum, csum);
 
-	bch_err_inum_offset_ratelimited(ca,
-		rbio->read_pos.inode,
-		rbio->read_pos.offset << 9,
-		"data %s", buf.buf);
+	struct bch_dev *ca = rbio->have_ioref ? bch2_dev_have_ref(c, rbio->pick.ptr.dev) : NULL;
+	if (ca) {
+		bch_err_inum_offset_ratelimited(ca,
+			rbio->read_pos.inode,
+			rbio->read_pos.offset << 9,
+			"data %s", buf.buf);
+		bch2_io_error(ca, BCH_MEMBER_ERROR_checksum);
+	}
 	printbuf_exit(&buf);
-
-	bch2_io_error(ca, BCH_MEMBER_ERROR_checksum);
 	bch2_rbio_error(rbio, READ_RETRY_AVOID, BLK_STS_IOERR);
 	goto out;
 decompression_err:
@@ -675,7 +676,7 @@ static void bch2_read_endio(struct bio *bio)
 	struct bch_read_bio *rbio =
 		container_of(bio, struct bch_read_bio, bio);
 	struct bch_fs *c	= rbio->c;
-	struct bch_dev *ca	= bch_dev_bkey_exists(c, rbio->pick.ptr.dev);
+	struct bch_dev *ca = rbio->have_ioref ? bch2_dev_have_ref(c, rbio->pick.ptr.dev) : NULL;
 	struct workqueue_struct *wq = NULL;
 	enum rbio_context context = RBIO_CONTEXT_NULL;
 
@@ -687,17 +688,21 @@ static void bch2_read_endio(struct bio *bio)
 	if (!rbio->split)
 		rbio->bio.bi_end_io = rbio->end_io;
 
-	if (bch2_dev_inum_io_err_on(bio->bi_status, ca, BCH_MEMBER_ERROR_read,
-				    rbio->read_pos.inode,
-				    rbio->read_pos.offset,
-				    "data read error: %s",
-			       bch2_blk_status_to_str(bio->bi_status))) {
+	if (bio->bi_status) {
+		if (ca) {
+			bch_err_inum_offset_ratelimited(ca,
+				rbio->read_pos.inode,
+				rbio->read_pos.offset,
+				"data read error: %s",
+				bch2_blk_status_to_str(bio->bi_status));
+			bch2_io_error(ca, BCH_MEMBER_ERROR_read);
+		}
 		bch2_rbio_error(rbio, READ_RETRY_AVOID, bio->bi_status);
 		return;
 	}
 
 	if (((rbio->flags & BCH_READ_RETRY_IF_STALE) && race_fault()) ||
-	    ptr_stale(ca, &rbio->pick.ptr)) {
+	    (ca && dev_ptr_stale(ca, &rbio->pick.ptr))) {
 		trace_and_count(c, read_reuse_race, &rbio->bio);
 
 		if (rbio->flags & BCH_READ_RETRY_IF_STALE)
@@ -758,32 +763,45 @@ err:
 }
 
 static noinline void read_from_stale_dirty_pointer(struct btree_trans *trans,
+						   struct bch_dev *ca,
 						   struct bkey_s_c k,
 						   struct bch_extent_ptr ptr)
 {
 	struct bch_fs *c = trans->c;
-	struct bch_dev *ca = bch_dev_bkey_exists(c, ptr.dev);
 	struct btree_iter iter;
 	struct printbuf buf = PRINTBUF;
 	int ret;
 
 	bch2_trans_iter_init(trans, &iter, BTREE_ID_alloc,
-			     PTR_BUCKET_POS(c, &ptr),
-			     BTREE_ITER_CACHED);
-
-	prt_printf(&buf, "Attempting to read from stale dirty pointer:");
-	printbuf_indent_add(&buf, 2);
-	prt_newline(&buf);
+			     PTR_BUCKET_POS(ca, &ptr),
+			     BTREE_ITER_cached);
 
-	bch2_bkey_val_to_text(&buf, c, k);
-	prt_newline(&buf);
+	u8 *gen = bucket_gen(ca, iter.pos.offset);
+	if (gen) {
 
-	prt_printf(&buf, "memory gen: %u", *bucket_gen(ca, iter.pos.offset));
+		prt_printf(&buf, "Attempting to read from stale dirty pointer:\n");
+		printbuf_indent_add(&buf, 2);
 
-	ret = lockrestart_do(trans, bkey_err(k = bch2_btree_iter_peek_slot(&iter)));
-	if (!ret) {
+		bch2_bkey_val_to_text(&buf, c, k);
 		prt_newline(&buf);
+
+		prt_printf(&buf, "memory gen: %u", *gen);
+
+		ret = lockrestart_do(trans, bkey_err(k = bch2_btree_iter_peek_slot(&iter)));
+		if (!ret) {
+			prt_newline(&buf);
+			bch2_bkey_val_to_text(&buf, c, k);
+		}
+	} else {
+		prt_printf(&buf, "Attempting to read from invalid bucket %llu:%llu:\n",
+			   iter.pos.inode, iter.pos.offset);
+		printbuf_indent_add(&buf, 2);
+
+		prt_printf(&buf, "first bucket %u nbuckets %llu\n",
+			   ca->mi.first_bucket, ca->mi.nbuckets);
+
 		bch2_bkey_val_to_text(&buf, c, k);
+		prt_newline(&buf);
 	}
 
 	bch2_fs_inconsistent(c, "%s", buf.buf);
@@ -801,7 +819,6 @@ int __bch2_read_extent(struct btree_trans *trans, struct bch_read_bio *orig,
 	struct bch_fs *c = trans->c;
 	struct extent_ptr_decoded pick;
 	struct bch_read_bio *rbio = NULL;
-	struct bch_dev *ca = NULL;
 	struct promote_op *promote = NULL;
 	bool bounce = false, read_full = false, narrow_crcs = false;
 	struct bpos data_pos = bkey_start_pos(k.k);
@@ -832,7 +849,7 @@ retry_pick:
 		goto err;
 	}
 
-	ca = bch_dev_bkey_exists(c, pick.ptr.dev);
+	struct bch_dev *ca = bch2_dev_get_ioref(c, pick.ptr.dev, READ);
 
 	/*
 	 * Stale dirty pointers are treated as IO errors, but @failed isn't
@@ -842,9 +859,11 @@ retry_pick:
 	 */
 	if ((flags & BCH_READ_IN_RETRY) &&
 	    !pick.ptr.cached &&
-	    unlikely(ptr_stale(ca, &pick.ptr))) {
-		read_from_stale_dirty_pointer(trans, k, pick.ptr);
+	    ca &&
+	    unlikely(dev_ptr_stale(ca, &pick.ptr))) {
+		read_from_stale_dirty_pointer(trans, ca, k, pick.ptr);
 		bch2_mark_io_failure(failed, &pick);
+		percpu_ref_put(&ca->io_ref);
 		goto retry_pick;
 	}
 
@@ -859,8 +878,11 @@ retry_pick:
 		 * can happen if we retry, and the extent we were going to read
 		 * has been merged in the meantime:
 		 */
-		if (pick.crc.compressed_size > orig->bio.bi_vcnt * PAGE_SECTORS)
+		if (pick.crc.compressed_size > orig->bio.bi_vcnt * PAGE_SECTORS) {
+			if (ca)
+				percpu_ref_put(&ca->io_ref);
 			goto hole;
+		}
 
 		iter.bi_size	= pick.crc.compressed_size << 9;
 		goto get_bio;
@@ -965,7 +987,7 @@ get_bio:
 	rbio->bvec_iter		= iter;
 	rbio->offset_into_extent= offset_into_extent;
 	rbio->flags		= flags;
-	rbio->have_ioref	= pick_ret > 0 && bch2_dev_get_ioref(ca, READ);
+	rbio->have_ioref	= ca != NULL;
 	rbio->narrow_crcs	= narrow_crcs;
 	rbio->hole		= 0;
 	rbio->retry		= 0;
@@ -981,6 +1003,9 @@ get_bio:
 	rbio->promote		= promote;
 	INIT_WORK(&rbio->work, NULL);
 
+	if (flags & BCH_READ_NODECODE)
+		orig->pick = pick;
+
 	rbio->bio.bi_opf	= orig->bio.bi_opf;
 	rbio->bio.bi_iter.bi_sector = pick.ptr.offset;
 	rbio->bio.bi_end_io	= bch2_read_endio;
@@ -995,7 +1020,7 @@ get_bio:
 	 * If it's being moved internally, we don't want to flag it as a cache
 	 * hit:
 	 */
-	if (pick.ptr.cached && !(flags & BCH_READ_NODECODE))
+	if (ca && pick.ptr.cached && !(flags & BCH_READ_NODECODE))
 		bch2_bucket_io_time_reset(trans, pick.ptr.dev,
 			PTR_BUCKET_NR(ca, &pick.ptr), READ);
 
@@ -1113,7 +1138,7 @@ retry:
 
 	bch2_trans_iter_init(trans, &iter, BTREE_ID_extents,
 			     SPOS(inum.inum, bvec_iter.bi_sector, snapshot),
-			     BTREE_ITER_SLOTS);
+			     BTREE_ITER_slots);
 	while (1) {
 		unsigned bytes, sectors, offset_into_extent;
 		enum btree_id data_btree = BTREE_ID_extents;
diff --git a/fs/bcachefs/io_write.c b/fs/bcachefs/io_write.c
index 2c098ac017..05e0cbef42 100644
--- a/fs/bcachefs/io_write.c
+++ b/fs/bcachefs/io_write.c
@@ -88,7 +88,7 @@ void bch2_latency_acct(struct bch_dev *ca, u64 submit_time, int rw)
 
 	bch2_congested_acct(ca, io_latency, now, rw);
 
-	__bch2_time_stats_update(&ca->io_latency[rw], submit_time, now);
+	__bch2_time_stats_update(&ca->io_latency[rw].stats, submit_time, now);
 }
 
 #endif
@@ -166,7 +166,7 @@ int bch2_sum_sector_overwrites(struct btree_trans *trans,
 	bch2_trans_copy_iter(&iter, extent_iter);
 
 	for_each_btree_key_upto_continue_norestart(iter,
-				new->k.p, BTREE_ITER_SLOTS, old, ret) {
+				new->k.p, BTREE_ITER_slots, old, ret) {
 		s64 sectors = min(new->k.p.offset, old.k->p.offset) -
 			max(bkey_start_offset(&new->k),
 			    bkey_start_offset(old.k));
@@ -199,9 +199,6 @@ static inline int bch2_extent_update_i_size_sectors(struct btree_trans *trans,
 						    u64 new_i_size,
 						    s64 i_sectors_delta)
 {
-	struct btree_iter iter;
-	struct bkey_i *k;
-	struct bkey_i_inode_v3 *inode;
 	/*
 	 * Crazy performance optimization:
 	 * Every extent update needs to also update the inode: the inode trigger
@@ -213,26 +210,37 @@ static inline int bch2_extent_update_i_size_sectors(struct btree_trans *trans,
 	 * to be journalled - if we crash, the bi_journal_seq update will be
 	 * lost, but that's fine.
 	 */
-	unsigned inode_update_flags = BTREE_UPDATE_NOJOURNAL;
-	int ret;
+	unsigned inode_update_flags = BTREE_UPDATE_nojournal;
 
-	k = bch2_bkey_get_mut_noupdate(trans, &iter, BTREE_ID_inodes,
+	struct btree_iter iter;
+	struct bkey_s_c k = bch2_bkey_get_iter(trans, &iter, BTREE_ID_inodes,
 			      SPOS(0,
 				   extent_iter->pos.inode,
 				   extent_iter->snapshot),
-			      BTREE_ITER_CACHED);
-	ret = PTR_ERR_OR_ZERO(k);
+			      BTREE_ITER_cached);
+	int ret = bkey_err(k);
 	if (unlikely(ret))
 		return ret;
 
-	if (unlikely(k->k.type != KEY_TYPE_inode_v3)) {
-		k = bch2_inode_to_v3(trans, k);
-		ret = PTR_ERR_OR_ZERO(k);
+	/*
+	 * varint_decode_fast(), in the inode .invalid method, reads up to 7
+	 * bytes past the end of the buffer:
+	 */
+	struct bkey_i *k_mut = bch2_trans_kmalloc_nomemzero(trans, bkey_bytes(k.k) + 8);
+	ret = PTR_ERR_OR_ZERO(k_mut);
+	if (unlikely(ret))
+		goto err;
+
+	bkey_reassemble(k_mut, k);
+
+	if (unlikely(k_mut->k.type != KEY_TYPE_inode_v3)) {
+		k_mut = bch2_inode_to_v3(trans, k_mut);
+		ret = PTR_ERR_OR_ZERO(k_mut);
 		if (unlikely(ret))
 			goto err;
 	}
 
-	inode = bkey_i_to_inode_v3(k);
+	struct bkey_i_inode_v3 *inode = bkey_i_to_inode_v3(k_mut);
 
 	if (!(le64_to_cpu(inode->v.bi_flags) & BCH_INODE_i_size_dirty) &&
 	    new_i_size > le64_to_cpu(inode->v.bi_size)) {
@@ -251,7 +259,7 @@ static inline int bch2_extent_update_i_size_sectors(struct btree_trans *trans,
 	}
 
 	ret = bch2_trans_update(trans, &iter, &inode->k_i,
-				BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE|
+				BTREE_UPDATE_internal_snapshot_node|
 				inode_update_flags);
 err:
 	bch2_trans_iter_exit(trans, &iter);
@@ -360,7 +368,7 @@ static int bch2_write_index_default(struct bch_write_op *op)
 
 		bch2_trans_iter_init(trans, &iter, BTREE_ID_extents,
 				     bkey_start_pos(&sk.k->k),
-				     BTREE_ITER_SLOTS|BTREE_ITER_INTENT);
+				     BTREE_ITER_slots|BTREE_ITER_intent);
 
 		ret =   bch2_bkey_set_needs_rebalance(c, sk.k, &op->opts) ?:
 			bch2_extent_update(trans, inum, &iter, sk.k,
@@ -399,13 +407,12 @@ void bch2_submit_wbio_replicas(struct bch_write_bio *wbio, struct bch_fs *c,
 	BUG_ON(c->opts.nochanges);
 
 	bkey_for_each_ptr(ptrs, ptr) {
-		BUG_ON(!bch2_dev_exists2(c, ptr->dev));
-
-		struct bch_dev *ca = bch_dev_bkey_exists(c, ptr->dev);
+		struct bch_dev *ca = nocow
+			? bch2_dev_have_ref(c, ptr->dev)
+			: bch2_dev_get_ioref(c, ptr->dev, type == BCH_DATA_btree ? READ : WRITE);
 
 		if (to_entry(ptr + 1) < ptrs.end) {
-			n = to_wbio(bio_alloc_clone(NULL, &wbio->bio,
-						GFP_NOFS, &ca->replica_set));
+			n = to_wbio(bio_alloc_clone(NULL, &wbio->bio, GFP_NOFS, &c->replica_set));
 
 			n->bio.bi_end_io	= wbio->bio.bi_end_io;
 			n->bio.bi_private	= wbio->bio.bi_private;
@@ -422,11 +429,12 @@ void bch2_submit_wbio_replicas(struct bch_write_bio *wbio, struct bch_fs *c,
 
 		n->c			= c;
 		n->dev			= ptr->dev;
-		n->have_ioref		= nocow || bch2_dev_get_ioref(ca,
-					type == BCH_DATA_btree ? READ : WRITE);
+		n->have_ioref		= ca != NULL;
 		n->nocow		= nocow;
 		n->submit_time		= local_clock();
 		n->inode_offset		= bkey_start_offset(&k->k);
+		if (nocow)
+			n->nocow_bucket	= PTR_BUCKET_NR(ca, ptr);
 		n->bio.bi_iter.bi_sector = ptr->offset;
 
 		if (likely(n->have_ioref)) {
@@ -473,7 +481,6 @@ static void bch2_write_done(struct closure *cl)
 static noinline int bch2_write_drop_io_error_ptrs(struct bch_write_op *op)
 {
 	struct keylist *keys = &op->insert_keys;
-	struct bch_extent_ptr *ptr;
 	struct bkey_i *src, *dst = keys->keys, *n;
 
 	for (src = keys->keys; src != keys->top; src = n) {
@@ -530,7 +537,8 @@ static void __bch2_write_index(struct bch_write_op *op)
 
 			bch_err_inum_offset_ratelimited(c,
 				insert->k.p.inode, insert->k.p.offset << 9,
-				"write error while doing btree update: %s",
+				"%s write error while doing btree update: %s",
+				op->flags & BCH_WRITE_MOVE ? "move" : "user",
 				bch2_err_str(ret));
 		}
 
@@ -641,7 +649,9 @@ static void bch2_write_endio(struct bio *bio)
 	struct bch_write_bio *wbio	= to_wbio(bio);
 	struct bch_write_bio *parent	= wbio->split ? wbio->parent : NULL;
 	struct bch_fs *c		= wbio->c;
-	struct bch_dev *ca		= bch_dev_bkey_exists(c, wbio->dev);
+	struct bch_dev *ca		= wbio->have_ioref
+		? bch2_dev_have_ref(c, wbio->dev)
+		: NULL;
 
 	if (bch2_dev_inum_io_err_on(bio->bi_status, ca, BCH_MEMBER_ERROR_write,
 				    op->pos.inode,
@@ -652,8 +662,12 @@ static void bch2_write_endio(struct bio *bio)
 		op->flags |= BCH_WRITE_IO_ERROR;
 	}
 
-	if (wbio->nocow)
+	if (wbio->nocow) {
+		bch2_bucket_nocow_unlock(&c->nocow_locks,
+					 POS(ca->dev_idx, wbio->nocow_bucket),
+					 BUCKET_NOCOW_LOCK_UPDATE);
 		set_bit(wbio->dev, op->devs_need_flush->d);
+	}
 
 	if (wbio->have_ioref) {
 		bch2_latency_acct(ca, wbio->submit_time, WRITE);
@@ -1067,7 +1081,8 @@ do_write:
 	*_dst = dst;
 	return more;
 csum_err:
-	bch_err(c, "error verifying existing checksum while rewriting existing data (memory corruption?)");
+	bch_err(c, "%s writ error: error verifying existing checksum while rewriting existing data (memory corruption?)",
+		op->flags & BCH_WRITE_MOVE ? "move" : "user");
 	ret = -EIO;
 err:
 	if (to_wbio(dst)->bounce)
@@ -1091,30 +1106,21 @@ static bool bch2_extent_is_writeable(struct bch_write_op *op,
 		return false;
 
 	e = bkey_s_c_to_extent(k);
+
+	rcu_read_lock();
 	extent_for_each_ptr_decode(e, p, entry) {
-		if (crc_is_encoded(p.crc) || p.has_ec)
+		if (crc_is_encoded(p.crc) || p.has_ec) {
+			rcu_read_unlock();
 			return false;
+		}
 
 		replicas += bch2_extent_ptr_durability(c, &p);
 	}
+	rcu_read_unlock();
 
 	return replicas >= op->opts.data_replicas;
 }
 
-static inline void bch2_nocow_write_unlock(struct bch_write_op *op)
-{
-	struct bch_fs *c = op->c;
-
-	for_each_keylist_key(&op->insert_keys, k) {
-		struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(bkey_i_to_s_c(k));
-
-		bkey_for_each_ptr(ptrs, ptr)
-			bch2_bucket_nocow_unlock(&c->nocow_locks,
-						 PTR_BUCKET_POS(c, ptr),
-						 BUCKET_NOCOW_LOCK_UPDATE);
-	}
-}
-
 static int bch2_nocow_write_convert_one_unwritten(struct btree_trans *trans,
 						  struct btree_iter *iter,
 						  struct bkey_i *orig,
@@ -1148,7 +1154,7 @@ static int bch2_nocow_write_convert_one_unwritten(struct btree_trans *trans,
 	return  bch2_extent_update_i_size_sectors(trans, iter,
 					min(new->k.p.offset << 9, new_i_size), 0) ?:
 		bch2_trans_update(trans, iter, new,
-				  BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
+				  BTREE_UPDATE_internal_snapshot_node);
 }
 
 static void bch2_nocow_write_convert_unwritten(struct bch_write_op *op)
@@ -1159,7 +1165,7 @@ static void bch2_nocow_write_convert_unwritten(struct bch_write_op *op)
 	for_each_keylist_key(&op->insert_keys, orig) {
 		int ret = for_each_btree_key_upto_commit(trans, iter, BTREE_ID_extents,
 				     bkey_start_pos(&orig->k), orig->k.p,
-				     BTREE_ITER_INTENT, k,
+				     BTREE_ITER_intent, k,
 				     NULL, NULL, BCH_TRANS_COMMIT_no_enospc, ({
 			bch2_nocow_write_convert_one_unwritten(trans, &iter, orig, k, op->new_i_size);
 		}));
@@ -1169,7 +1175,8 @@ static void bch2_nocow_write_convert_unwritten(struct bch_write_op *op)
 
 			bch_err_inum_offset_ratelimited(c,
 				insert->k.p.inode, insert->k.p.offset << 9,
-				"write error while doing btree update: %s",
+				"%s write error while doing btree update: %s",
+				op->flags & BCH_WRITE_MOVE ? "move" : "user",
 				bch2_err_str(ret));
 		}
 
@@ -1184,8 +1191,6 @@ static void bch2_nocow_write_convert_unwritten(struct bch_write_op *op)
 
 static void __bch2_nocow_write_done(struct bch_write_op *op)
 {
-	bch2_nocow_write_unlock(op);
-
 	if (unlikely(op->flags & BCH_WRITE_IO_ERROR)) {
 		op->error = -EIO;
 	} else if (unlikely(op->flags & BCH_WRITE_CONVERT_UNWRITTEN))
@@ -1215,7 +1220,7 @@ static void bch2_nocow_write(struct bch_write_op *op)
 	DARRAY_PREALLOCATED(struct bucket_to_lock, 3) buckets;
 	u32 snapshot;
 	struct bucket_to_lock *stale_at;
-	int ret;
+	int stale, ret;
 
 	if (op->flags & BCH_WRITE_MOVE)
 		return;
@@ -1231,12 +1236,16 @@ retry:
 
 	bch2_trans_iter_init(trans, &iter, BTREE_ID_extents,
 			     SPOS(op->pos.inode, op->pos.offset, snapshot),
-			     BTREE_ITER_SLOTS);
+			     BTREE_ITER_slots);
 	while (1) {
 		struct bio *bio = &op->wbio.bio;
 
 		buckets.nr = 0;
 
+		ret = bch2_trans_relock(trans);
+		if (ret)
+			break;
+
 		k = bch2_btree_iter_peek_slot(&iter);
 		ret = bkey_err(k);
 		if (ret)
@@ -1256,14 +1265,15 @@ retry:
 		/* Get iorefs before dropping btree locks: */
 		struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k);
 		bkey_for_each_ptr(ptrs, ptr) {
-			struct bpos b = PTR_BUCKET_POS(c, ptr);
+			struct bch_dev *ca = bch2_dev_get_ioref(c, ptr->dev, WRITE);
+			if (unlikely(!ca))
+				goto err_get_ioref;
+
+			struct bpos b = PTR_BUCKET_POS(ca, ptr);
 			struct nocow_lock_bucket *l =
 				bucket_nocow_lock(&c->nocow_locks, bucket_to_u64(b));
 			prefetch(l);
 
-			if (unlikely(!bch2_dev_get_ioref(bch_dev_bkey_exists(c, ptr->dev), WRITE)))
-				goto err_get_ioref;
-
 			/* XXX allocating memory with btree locks held - rare */
 			darray_push_gfp(&buckets, ((struct bucket_to_lock) {
 						   .b = b, .gen = ptr->gen, .l = l,
@@ -1282,14 +1292,15 @@ retry:
 			bch2_cut_back(POS(op->pos.inode, op->pos.offset + bio_sectors(bio)), op->insert_keys.top);
 
 		darray_for_each(buckets, i) {
-			struct bch_dev *ca = bch_dev_bkey_exists(c, i->b.inode);
+			struct bch_dev *ca = bch2_dev_have_ref(c, i->b.inode);
 
 			__bch2_bucket_nocow_lock(&c->nocow_locks, i->l,
 						 bucket_to_u64(i->b),
 						 BUCKET_NOCOW_LOCK_UPDATE);
 
 			rcu_read_lock();
-			bool stale = gen_after(*bucket_gen(ca, i->b.offset), i->gen);
+			u8 *gen = bucket_gen(ca, i->b.offset);
+			stale = !gen ? -1 : gen_after(*gen, i->gen);
 			rcu_read_unlock();
 
 			if (unlikely(stale)) {
@@ -1359,7 +1370,7 @@ err:
 	return;
 err_get_ioref:
 	darray_for_each(buckets, i)
-		percpu_ref_put(&bch_dev_bkey_exists(c, i->b.inode)->io_ref);
+		percpu_ref_put(&bch2_dev_have_ref(c, i->b.inode)->io_ref);
 
 	/* Fall back to COW path: */
 	goto out;
@@ -1370,8 +1381,18 @@ err_bucket_stale:
 			break;
 	}
 
-	/* We can retry this: */
-	ret = -BCH_ERR_transaction_restart;
+	struct printbuf buf = PRINTBUF;
+	if (bch2_fs_inconsistent_on(stale < 0, c,
+				    "pointer to invalid bucket in nocow path on device %llu\n  %s",
+				    stale_at->b.inode,
+				    (bch2_bkey_val_to_text(&buf, c, k), buf.buf))) {
+		ret = -EIO;
+	} else {
+		/* We can retry this: */
+		ret = -BCH_ERR_transaction_restart;
+	}
+	printbuf_exit(&buf);
+
 	goto err_get_ioref;
 }
 
@@ -1449,7 +1470,9 @@ err:
 					bch_err_inum_offset_ratelimited(c,
 						op->pos.inode,
 						op->pos.offset << 9,
-						"%s(): error: %s", __func__, bch2_err_str(ret));
+						"%s(): %s error: %s", __func__,
+						op->flags & BCH_WRITE_MOVE ? "move" : "user",
+						bch2_err_str(ret));
 				op->error = ret;
 				break;
 			}
@@ -1478,7 +1501,11 @@ err:
 	if ((op->flags & BCH_WRITE_SYNC) ||
 	    (!(op->flags & BCH_WRITE_DONE) &&
 	     !(op->flags & BCH_WRITE_IN_WORKER))) {
-		closure_sync(&op->cl);
+		if (closure_sync_timeout(&op->cl, HZ * 10)) {
+			bch2_print_allocator_stuck(c);
+			closure_sync(&op->cl);
+		}
+
 		__bch2_write_index(op);
 
 		if (!(op->flags & BCH_WRITE_DONE))
@@ -1500,6 +1527,8 @@ static void bch2_write_data_inline(struct bch_write_op *op, unsigned data_len)
 	unsigned sectors;
 	int ret;
 
+	memset(&op->failed, 0, sizeof(op->failed));
+
 	op->flags |= BCH_WRITE_WROTE_DATA_INLINE;
 	op->flags |= BCH_WRITE_DONE;
 
@@ -1573,7 +1602,8 @@ CLOSURE_CALLBACK(bch2_write)
 		bch_err_inum_offset_ratelimited(c,
 			op->pos.inode,
 			op->pos.offset << 9,
-			"misaligned write");
+			"%s write error: misaligned write",
+			op->flags & BCH_WRITE_MOVE ? "move" : "user");
 		op->error = -EIO;
 		goto err;
 	}
@@ -1633,8 +1663,7 @@ void bch2_write_op_to_text(struct printbuf *out, struct bch_write_op *op)
 	prt_bitflags(out, bch2_write_flags, op->flags);
 	prt_newline(out);
 
-	prt_printf(out, "ref: %u", closure_nr_remaining(&op->cl));
-	prt_newline(out);
+	prt_printf(out, "ref: %u\n", closure_nr_remaining(&op->cl));
 
 	printbuf_indent_sub(out, 2);
 }
@@ -1642,13 +1671,14 @@ void bch2_write_op_to_text(struct printbuf *out, struct bch_write_op *op)
 void bch2_fs_io_write_exit(struct bch_fs *c)
 {
 	mempool_exit(&c->bio_bounce_pages);
+	bioset_exit(&c->replica_set);
 	bioset_exit(&c->bio_write);
 }
 
 int bch2_fs_io_write_init(struct bch_fs *c)
 {
-	if (bioset_init(&c->bio_write, 1, offsetof(struct bch_write_bio, bio),
-			BIOSET_NEED_BVECS))
+	if (bioset_init(&c->bio_write,   1, offsetof(struct bch_write_bio, bio), BIOSET_NEED_BVECS) ||
+	    bioset_init(&c->replica_set, 4, offsetof(struct bch_write_bio, bio), 0))
 		return -BCH_ERR_ENOMEM_bio_write_init;
 
 	if (mempool_init_page_pool(&c->bio_bounce_pages,
diff --git a/fs/bcachefs/io_write_types.h b/fs/bcachefs/io_write_types.h
index c7f97c2c48..6e878a6f2f 100644
--- a/fs/bcachefs/io_write_types.h
+++ b/fs/bcachefs/io_write_types.h
@@ -20,6 +20,7 @@ struct bch_write_bio {
 
 	u64			submit_time;
 	u64			inode_offset;
+	u64			nocow_bucket;
 
 	struct bch_devs_list	failed;
 	u8			dev;
diff --git a/fs/bcachefs/journal.c b/fs/bcachefs/journal.c
index bc890776eb..10b19791ec 100644
--- a/fs/bcachefs/journal.c
+++ b/fs/bcachefs/journal.c
@@ -27,31 +27,59 @@ static const char * const bch2_journal_errors[] = {
 	NULL
 };
 
+static inline bool journal_seq_unwritten(struct journal *j, u64 seq)
+{
+	return seq > j->seq_ondisk;
+}
+
+static bool __journal_entry_is_open(union journal_res_state state)
+{
+	return state.cur_entry_offset < JOURNAL_ENTRY_CLOSED_VAL;
+}
+
+static inline unsigned nr_unwritten_journal_entries(struct journal *j)
+{
+	return atomic64_read(&j->seq) - j->seq_ondisk;
+}
+
+static bool journal_entry_is_open(struct journal *j)
+{
+	return __journal_entry_is_open(j->reservations);
+}
+
 static void bch2_journal_buf_to_text(struct printbuf *out, struct journal *j, u64 seq)
 {
 	union journal_res_state s = READ_ONCE(j->reservations);
 	unsigned i = seq & JOURNAL_BUF_MASK;
 	struct journal_buf *buf = j->buf + i;
 
-	prt_printf(out, "seq:");
-	prt_tab(out);
-	prt_printf(out, "%llu", seq);
-	prt_newline(out);
+	prt_printf(out, "seq:\t%llu\n", seq);
 	printbuf_indent_add(out, 2);
 
-	prt_printf(out, "refcount:");
-	prt_tab(out);
-	prt_printf(out, "%u", journal_state_count(s, i));
-	prt_newline(out);
+	prt_printf(out, "refcount:\t%u\n", journal_state_count(s, i));
 
-	prt_printf(out, "size:");
-	prt_tab(out);
+	prt_printf(out, "size:\t");
 	prt_human_readable_u64(out, vstruct_bytes(buf->data));
 	prt_newline(out);
 
-	prt_printf(out, "expires");
-	prt_tab(out);
-	prt_printf(out, "%li jiffies", buf->expires - jiffies);
+	prt_printf(out, "expires:\t");
+	prt_printf(out, "%li jiffies\n", buf->expires - jiffies);
+
+	prt_printf(out, "flags:\t");
+	if (buf->noflush)
+		prt_str(out, "noflush ");
+	if (buf->must_flush)
+		prt_str(out, "must_flush ");
+	if (buf->separate_flush)
+		prt_str(out, "separate_flush ");
+	if (buf->need_flush_to_write_buffer)
+		prt_str(out, "need_flush_to_write_buffer ");
+	if (buf->write_started)
+		prt_str(out, "write_started ");
+	if (buf->write_allocated)
+		prt_str(out, "write_allocated ");
+	if (buf->write_done)
+		prt_str(out, "write_done");
 	prt_newline(out);
 
 	printbuf_indent_sub(out, 2);
@@ -66,26 +94,7 @@ static void bch2_journal_bufs_to_text(struct printbuf *out, struct journal *j)
 	     seq <= journal_cur_seq(j);
 	     seq++)
 		bch2_journal_buf_to_text(out, j, seq);
-}
-
-static inline bool journal_seq_unwritten(struct journal *j, u64 seq)
-{
-	return seq > j->seq_ondisk;
-}
-
-static bool __journal_entry_is_open(union journal_res_state state)
-{
-	return state.cur_entry_offset < JOURNAL_ENTRY_CLOSED_VAL;
-}
-
-static inline unsigned nr_unwritten_journal_entries(struct journal *j)
-{
-	return atomic64_read(&j->seq) - j->seq_ondisk;
-}
-
-static bool journal_entry_is_open(struct journal *j)
-{
-	return __journal_entry_is_open(j->reservations);
+	prt_printf(out, "last buf %s\n", journal_entry_is_open(j) ? "open" : "closed");
 }
 
 static inline struct journal_buf *
@@ -174,21 +183,40 @@ journal_error_check_stuck(struct journal *j, int error, unsigned flags)
 	return stuck;
 }
 
+void bch2_journal_do_writes(struct journal *j)
+{
+	for (u64 seq = journal_last_unwritten_seq(j);
+	     seq <= journal_cur_seq(j);
+	     seq++) {
+		unsigned idx = seq & JOURNAL_BUF_MASK;
+		struct journal_buf *w = j->buf + idx;
+
+		if (w->write_started && !w->write_allocated)
+			break;
+		if (w->write_started)
+			continue;
+
+		if (!journal_state_count(j->reservations, idx)) {
+			w->write_started = true;
+			closure_call(&w->io, bch2_journal_write, j->wq, NULL);
+		}
+
+		break;
+	}
+}
+
 /*
  * Final processing when the last reference of a journal buffer has been
  * dropped. Drop the pin list reference acquired at journal entry open and write
  * the buffer, if requested.
  */
-void bch2_journal_buf_put_final(struct journal *j, u64 seq, bool write)
+void bch2_journal_buf_put_final(struct journal *j, u64 seq)
 {
-	struct bch_fs *c = container_of(j, struct bch_fs, journal);
-
 	lockdep_assert_held(&j->lock);
 
 	if (__bch2_journal_pin_put(j, seq))
 		bch2_journal_reclaim_fast(j);
-	if (write)
-		closure_call(&j->io, bch2_journal_write, c->io_complete_wq, NULL);
+	bch2_journal_do_writes(j);
 }
 
 /*
@@ -380,11 +408,14 @@ static int journal_entry_open(struct journal *j)
 	BUG_ON(j->buf + (journal_cur_seq(j) & JOURNAL_BUF_MASK) != buf);
 
 	bkey_extent_init(&buf->key);
-	buf->noflush	= false;
-	buf->must_flush	= false;
-	buf->separate_flush = false;
-	buf->flush_time	= 0;
+	buf->noflush		= false;
+	buf->must_flush		= false;
+	buf->separate_flush	= false;
+	buf->flush_time		= 0;
 	buf->need_flush_to_write_buffer = true;
+	buf->write_started	= false;
+	buf->write_allocated	= false;
+	buf->write_done		= false;
 
 	memset(buf->data, 0, sizeof(*buf->data));
 	buf->data->seq	= cpu_to_le64(journal_cur_seq(j));
@@ -418,9 +449,10 @@ static int journal_entry_open(struct journal *j)
 	} while ((v = atomic64_cmpxchg(&j->reservations.counter,
 				       old.v, new.v)) != old.v);
 
-	mod_delayed_work(c->io_complete_wq,
-			 &j->write_work,
-			 msecs_to_jiffies(c->opts.journal_flush_delay));
+	if (nr_unwritten_journal_entries(j) == 1)
+		mod_delayed_work(j->wq,
+				 &j->write_work,
+				 msecs_to_jiffies(c->opts.journal_flush_delay));
 	journal_wake(j);
 
 	if (j->early_journal_entries.nr)
@@ -445,20 +477,16 @@ static void journal_quiesce(struct journal *j)
 static void journal_write_work(struct work_struct *work)
 {
 	struct journal *j = container_of(work, struct journal, write_work.work);
-	struct bch_fs *c = container_of(j, struct bch_fs, journal);
-	long delta;
 
 	spin_lock(&j->lock);
-	if (!__journal_entry_is_open(j->reservations))
-		goto unlock;
-
-	delta = journal_cur_buf(j)->expires - jiffies;
+	if (__journal_entry_is_open(j->reservations)) {
+		long delta = journal_cur_buf(j)->expires - jiffies;
 
-	if (delta > 0)
-		mod_delayed_work(c->io_complete_wq, &j->write_work, delta);
-	else
-		__journal_entry_close(j, JOURNAL_ENTRY_CLOSED_VAL, true);
-unlock:
+		if (delta > 0)
+			mod_delayed_work(j->wq, &j->write_work, delta);
+		else
+			__journal_entry_close(j, JOURNAL_ENTRY_CLOSED_VAL, true);
+	}
 	spin_unlock(&j->lock);
 }
 
@@ -476,30 +504,29 @@ retry:
 	if (bch2_journal_error(j))
 		return -BCH_ERR_erofs_journal_err;
 
-	spin_lock(&j->lock);
+	if (j->blocked)
+		return -BCH_ERR_journal_res_get_blocked;
 
-	/* check once more in case somebody else shut things down... */
-	if (bch2_journal_error(j)) {
-		spin_unlock(&j->lock);
-		return -BCH_ERR_erofs_journal_err;
+	if ((flags & BCH_WATERMARK_MASK) < j->watermark) {
+		ret = JOURNAL_ERR_journal_full;
+		can_discard = j->can_discard;
+		goto out;
 	}
 
+	if (nr_unwritten_journal_entries(j) == ARRAY_SIZE(j->buf) && !journal_entry_is_open(j)) {
+		ret = JOURNAL_ERR_max_in_flight;
+		goto out;
+	}
+
+	spin_lock(&j->lock);
+
 	/*
 	 * Recheck after taking the lock, so we don't race with another thread
 	 * that just did journal_entry_open() and call bch2_journal_entry_close()
 	 * unnecessarily
 	 */
 	if (journal_res_get_fast(j, res, flags)) {
-		spin_unlock(&j->lock);
-		return 0;
-	}
-
-	if ((flags & BCH_WATERMARK_MASK) < j->watermark) {
-		/*
-		 * Don't want to close current journal entry, just need to
-		 * invoke reclaim:
-		 */
-		ret = JOURNAL_ERR_journal_full;
+		ret = 0;
 		goto unlock;
 	}
 
@@ -515,30 +542,30 @@ retry:
 		j->buf_size_want = max(j->buf_size_want, buf->buf_size << 1);
 
 	__journal_entry_close(j, JOURNAL_ENTRY_CLOSED_VAL, false);
-	ret = journal_entry_open(j);
-
-	if (ret == JOURNAL_ERR_max_in_flight) {
-		track_event_change(&c->times[BCH_TIME_blocked_journal_max_in_flight],
-				   &j->max_in_flight_start, true);
-		if (trace_journal_entry_full_enabled()) {
-			struct printbuf buf = PRINTBUF;
-			buf.atomic++;
-
-			bch2_journal_bufs_to_text(&buf, j);
-			trace_journal_entry_full(c, buf.buf);
-			printbuf_exit(&buf);
-		}
-		count_event(c, journal_entry_full);
-	}
+	ret = journal_entry_open(j) ?: JOURNAL_ERR_retry;
 unlock:
 	can_discard = j->can_discard;
 	spin_unlock(&j->lock);
-
-	if (!ret)
+out:
+	if (ret == JOURNAL_ERR_retry)
 		goto retry;
+	if (!ret)
+		return 0;
+
 	if (journal_error_check_stuck(j, ret, flags))
 		ret = -BCH_ERR_journal_res_get_blocked;
 
+	if (ret == JOURNAL_ERR_max_in_flight &&
+	    track_event_change(&c->times[BCH_TIME_blocked_journal_max_in_flight], true)) {
+
+		struct printbuf buf = PRINTBUF;
+		prt_printf(&buf, "seq %llu\n", journal_cur_seq(j));
+		bch2_journal_bufs_to_text(&buf, j);
+		trace_journal_entry_full(c, buf.buf);
+		printbuf_exit(&buf);
+		count_event(c, journal_entry_full);
+	}
+
 	/*
 	 * Journal is full - can't rely on reclaim from work item due to
 	 * freezing:
@@ -669,12 +696,18 @@ recheck_need_open:
 
 		spin_unlock(&j->lock);
 
+		/*
+		 * We're called from bch2_journal_flush_seq() -> wait_event();
+		 * but this might block. We won't usually block, so we won't
+		 * livelock:
+		 */
+		sched_annotate_sleep();
 		ret = bch2_journal_res_get(j, &res, jset_u64s(0), 0);
 		if (ret)
 			return ret;
 
 		seq = res.seq;
-		buf = j->buf + (seq & JOURNAL_BUF_MASK);
+		buf = journal_seq_to_buf(j, seq);
 		buf->must_flush = true;
 
 		if (!buf->flush_time) {
@@ -692,8 +725,8 @@ recheck_need_open:
 	}
 
 	/*
-	 * if write was kicked off without a flush, flush the next sequence
-	 * number instead
+	 * if write was kicked off without a flush, or if we promised it
+	 * wouldn't be a flush, flush the next sequence number instead
 	 */
 	buf = journal_seq_to_buf(j, seq);
 	if (buf->noflush) {
@@ -771,8 +804,8 @@ bool bch2_journal_noflush_seq(struct journal *j, u64 seq)
 	     unwritten_seq++) {
 		struct journal_buf *buf = journal_seq_to_buf(j, unwritten_seq);
 
-		/* journal write is already in flight, and was a flush write: */
-		if (unwritten_seq == journal_last_unwritten_seq(j) && !buf->noflush)
+		/* journal flush already in flight, or flush requseted */
+		if (buf->must_flush)
 			goto out;
 
 		buf->noflush = true;
@@ -833,6 +866,8 @@ static struct journal_buf *__bch2_next_write_buffer_flush_journal_buf(struct jou
 {
 	struct journal_buf *ret = NULL;
 
+	/* We're inside wait_event(), but using mutex_lock(: */
+	sched_annotate_sleep();
 	mutex_lock(&j->buf_lock);
 	spin_lock(&j->lock);
 	max_seq = min(max_seq, journal_cur_seq(j));
@@ -903,7 +938,8 @@ static int __bch2_set_nr_journal_buckets(struct bch_dev *ca, unsigned nr,
 				break;
 			}
 		} else {
-			ob[nr_got] = bch2_bucket_alloc(c, ca, BCH_WATERMARK_normal, cl);
+			ob[nr_got] = bch2_bucket_alloc(c, ca, BCH_WATERMARK_normal,
+						       BCH_DATA_journal, cl);
 			ret = PTR_ERR_OR_ZERO(ob[nr_got]);
 			if (ret)
 				break;
@@ -911,7 +947,7 @@ static int __bch2_set_nr_journal_buckets(struct bch_dev *ca, unsigned nr,
 			ret = bch2_trans_run(c,
 				bch2_trans_mark_metadata_bucket(trans, ca,
 						ob[nr_got]->bucket, BCH_DATA_journal,
-						ca->mi.bucket_size));
+						ca->mi.bucket_size, BTREE_TRIGGER_transactional));
 			if (ret) {
 				bch2_open_bucket_put(c, ob[nr_got]);
 				bch_err_msg(c, ret, "marking new journal buckets");
@@ -991,7 +1027,8 @@ err_unblock:
 		for (i = 0; i < nr_got; i++)
 			bch2_trans_run(c,
 				bch2_trans_mark_metadata_bucket(trans, ca,
-						bu[i], BCH_DATA_free, 0));
+						bu[i], BCH_DATA_free, 0,
+						BTREE_TRIGGER_transactional));
 err_free:
 	if (!new_fs)
 		for (i = 0; i < nr_got; i++)
@@ -1058,7 +1095,7 @@ unlock:
 	return ret;
 }
 
-int bch2_dev_journal_alloc(struct bch_dev *ca)
+int bch2_dev_journal_alloc(struct bch_dev *ca, bool new_fs)
 {
 	unsigned nr;
 	int ret;
@@ -1080,7 +1117,7 @@ int bch2_dev_journal_alloc(struct bch_dev *ca)
 		     min(1 << 13,
 			 (1 << 24) / ca->mi.bucket_size));
 
-	ret = __bch2_set_nr_journal_buckets(ca, nr, true, NULL);
+	ret = __bch2_set_nr_journal_buckets(ca, nr, new_fs, NULL);
 err:
 	bch_err_fn(ca, ret);
 	return ret;
@@ -1092,7 +1129,7 @@ int bch2_fs_journal_alloc(struct bch_fs *c)
 		if (ca->journal.nr)
 			continue;
 
-		int ret = bch2_dev_journal_alloc(ca);
+		int ret = bch2_dev_journal_alloc(ca, true);
 		if (ret) {
 			percpu_ref_put(&ca->io_ref);
 			return ret;
@@ -1130,6 +1167,9 @@ void bch2_dev_journal_stop(struct journal *j, struct bch_dev *ca)
 
 void bch2_fs_journal_stop(struct journal *j)
 {
+	if (!test_bit(JOURNAL_running, &j->flags))
+		return;
+
 	bch2_journal_reclaim_stop(j);
 	bch2_journal_flush_all_pins(j);
 
@@ -1142,12 +1182,16 @@ void bch2_fs_journal_stop(struct journal *j)
 	bch2_journal_meta(j);
 
 	journal_quiesce(j);
+	cancel_delayed_work_sync(&j->write_work);
 
-	BUG_ON(!bch2_journal_error(j) &&
-	       test_bit(JOURNAL_REPLAY_DONE, &j->flags) &&
-	       j->last_empty_seq != journal_cur_seq(j));
+	WARN(!bch2_journal_error(j) &&
+	     test_bit(JOURNAL_replay_done, &j->flags) &&
+	     j->last_empty_seq != journal_cur_seq(j),
+	     "journal shutdown error: cur seq %llu but last empty seq %llu",
+	     journal_cur_seq(j), j->last_empty_seq);
 
-	cancel_delayed_work_sync(&j->write_work);
+	if (!bch2_journal_error(j))
+		clear_bit(JOURNAL_running, &j->flags);
 }
 
 int bch2_fs_journal_start(struct journal *j, u64 cur_seq)
@@ -1157,13 +1201,12 @@ int bch2_fs_journal_start(struct journal *j, u64 cur_seq)
 	struct journal_replay *i, **_i;
 	struct genradix_iter iter;
 	bool had_entries = false;
-	unsigned ptr;
 	u64 last_seq = cur_seq, nr, seq;
 
 	genradix_for_each_reverse(&c->journal_entries, iter, _i) {
 		i = *_i;
 
-		if (!i || i->ignore)
+		if (journal_replay_ignore(i))
 			continue;
 
 		last_seq = le64_to_cpu(i->j.last_seq);
@@ -1196,7 +1239,7 @@ int bch2_fs_journal_start(struct journal *j, u64 cur_seq)
 	genradix_for_each(&c->journal_entries, iter, _i) {
 		i = *_i;
 
-		if (!i || i->ignore)
+		if (journal_replay_ignore(i))
 			continue;
 
 		seq = le64_to_cpu(i->j.seq);
@@ -1211,8 +1254,8 @@ int bch2_fs_journal_start(struct journal *j, u64 cur_seq)
 		p = journal_seq_pin(j, seq);
 
 		p->devs.nr = 0;
-		for (ptr = 0; ptr < i->nr_ptrs; ptr++)
-			bch2_dev_list_add_dev(&p->devs, i->ptrs[ptr].dev);
+		darray_for_each(i->ptrs, ptr)
+			bch2_dev_list_add_dev(&p->devs, ptr->dev);
 
 		had_entries = true;
 	}
@@ -1222,7 +1265,7 @@ int bch2_fs_journal_start(struct journal *j, u64 cur_seq)
 
 	spin_lock(&j->lock);
 
-	set_bit(JOURNAL_STARTED, &j->flags);
+	set_bit(JOURNAL_running, &j->flags);
 	j->last_flush_write = jiffies;
 
 	j->reservations.idx = j->reservations.unwritten_idx = journal_cur_seq(j);
@@ -1240,13 +1283,17 @@ int bch2_fs_journal_start(struct journal *j, u64 cur_seq)
 
 void bch2_dev_journal_exit(struct bch_dev *ca)
 {
-	kfree(ca->journal.bio);
-	kfree(ca->journal.buckets);
-	kfree(ca->journal.bucket_seq);
+	struct journal_device *ja = &ca->journal;
 
-	ca->journal.bio		= NULL;
-	ca->journal.buckets	= NULL;
-	ca->journal.bucket_seq	= NULL;
+	for (unsigned i = 0; i < ARRAY_SIZE(ja->bio); i++) {
+		kfree(ja->bio[i]);
+		ja->bio[i] = NULL;
+	}
+
+	kfree(ja->buckets);
+	kfree(ja->bucket_seq);
+	ja->buckets	= NULL;
+	ja->bucket_seq	= NULL;
 }
 
 int bch2_dev_journal_init(struct bch_dev *ca, struct bch_sb *sb)
@@ -1256,14 +1303,13 @@ int bch2_dev_journal_init(struct bch_dev *ca, struct bch_sb *sb)
 		bch2_sb_field_get(sb, journal);
 	struct bch_sb_field_journal_v2 *journal_buckets_v2 =
 		bch2_sb_field_get(sb, journal_v2);
-	unsigned i, nr_bvecs;
 
 	ja->nr = 0;
 
 	if (journal_buckets_v2) {
 		unsigned nr = bch2_sb_field_journal_v2_nr_entries(journal_buckets_v2);
 
-		for (i = 0; i < nr; i++)
+		for (unsigned i = 0; i < nr; i++)
 			ja->nr += le64_to_cpu(journal_buckets_v2->d[i].nr);
 	} else if (journal_buckets) {
 		ja->nr = bch2_nr_journal_buckets(journal_buckets);
@@ -1273,13 +1319,18 @@ int bch2_dev_journal_init(struct bch_dev *ca, struct bch_sb *sb)
 	if (!ja->bucket_seq)
 		return -BCH_ERR_ENOMEM_dev_journal_init;
 
-	nr_bvecs = DIV_ROUND_UP(JOURNAL_ENTRY_SIZE_MAX, PAGE_SIZE);
+	unsigned nr_bvecs = DIV_ROUND_UP(JOURNAL_ENTRY_SIZE_MAX, PAGE_SIZE);
 
-	ca->journal.bio = bio_kmalloc(nr_bvecs, GFP_KERNEL);
-	if (!ca->journal.bio)
-		return -BCH_ERR_ENOMEM_dev_journal_init;
+	for (unsigned i = 0; i < ARRAY_SIZE(ja->bio); i++) {
+		ja->bio[i] = kmalloc(struct_size(ja->bio[i], bio.bi_inline_vecs,
+				     nr_bvecs), GFP_KERNEL);
+		if (!ja->bio[i])
+			return -BCH_ERR_ENOMEM_dev_journal_init;
 
-	bio_init(ca->journal.bio, NULL, ca->journal.bio->bi_inline_vecs, nr_bvecs, 0);
+		ja->bio[i]->ca = ca;
+		ja->bio[i]->buf_idx = i;
+		bio_init(&ja->bio[i]->bio, NULL, ja->bio[i]->bio.bi_inline_vecs, nr_bvecs, 0);
+	}
 
 	ja->buckets = kcalloc(ja->nr, sizeof(u64), GFP_KERNEL);
 	if (!ja->buckets)
@@ -1287,14 +1338,14 @@ int bch2_dev_journal_init(struct bch_dev *ca, struct bch_sb *sb)
 
 	if (journal_buckets_v2) {
 		unsigned nr = bch2_sb_field_journal_v2_nr_entries(journal_buckets_v2);
-		unsigned j, dst = 0;
+		unsigned dst = 0;
 
-		for (i = 0; i < nr; i++)
-			for (j = 0; j < le64_to_cpu(journal_buckets_v2->d[i].nr); j++)
+		for (unsigned i = 0; i < nr; i++)
+			for (unsigned j = 0; j < le64_to_cpu(journal_buckets_v2->d[i].nr); j++)
 				ja->buckets[dst++] =
 					le64_to_cpu(journal_buckets_v2->d[i].start) + j;
 	} else if (journal_buckets) {
-		for (i = 0; i < ja->nr; i++)
+		for (unsigned i = 0; i < ja->nr; i++)
 			ja->buckets[i] = le64_to_cpu(journal_buckets->buckets[i]);
 	}
 
@@ -1303,19 +1354,19 @@ int bch2_dev_journal_init(struct bch_dev *ca, struct bch_sb *sb)
 
 void bch2_fs_journal_exit(struct journal *j)
 {
-	unsigned i;
+	if (j->wq)
+		destroy_workqueue(j->wq);
 
 	darray_exit(&j->early_journal_entries);
 
-	for (i = 0; i < ARRAY_SIZE(j->buf); i++)
-		kvpfree(j->buf[i].data, j->buf[i].buf_size);
+	for (unsigned i = 0; i < ARRAY_SIZE(j->buf); i++)
+		kvfree(j->buf[i].data);
 	free_fifo(&j->pin);
 }
 
 int bch2_fs_journal_init(struct journal *j)
 {
 	static struct lock_class_key res_key;
-	unsigned i;
 
 	mutex_init(&j->buf_lock);
 	spin_lock_init(&j->lock);
@@ -1336,19 +1387,32 @@ int bch2_fs_journal_init(struct journal *j)
 	if (!(init_fifo(&j->pin, JOURNAL_PIN, GFP_KERNEL)))
 		return -BCH_ERR_ENOMEM_journal_pin_fifo;
 
-	for (i = 0; i < ARRAY_SIZE(j->buf); i++) {
+	for (unsigned i = 0; i < ARRAY_SIZE(j->buf); i++) {
 		j->buf[i].buf_size = JOURNAL_ENTRY_SIZE_MIN;
-		j->buf[i].data = kvpmalloc(j->buf[i].buf_size, GFP_KERNEL);
+		j->buf[i].data = kvmalloc(j->buf[i].buf_size, GFP_KERNEL);
 		if (!j->buf[i].data)
 			return -BCH_ERR_ENOMEM_journal_buf;
+		j->buf[i].idx = i;
 	}
 
 	j->pin.front = j->pin.back = 1;
+
+	j->wq = alloc_workqueue("bcachefs_journal",
+				WQ_HIGHPRI|WQ_FREEZABLE|WQ_UNBOUND|WQ_MEM_RECLAIM, 512);
+	if (!j->wq)
+		return -BCH_ERR_ENOMEM_fs_other_alloc;
 	return 0;
 }
 
 /* debug: */
 
+static const char * const bch2_journal_flags_strs[] = {
+#define x(n)	#n,
+	JOURNAL_FLAGS()
+#undef x
+	NULL
+};
+
 void __bch2_journal_debug_to_text(struct printbuf *out, struct journal *j)
 {
 	struct bch_fs *c = container_of(j, struct bch_fs, journal);
@@ -1356,20 +1420,23 @@ void __bch2_journal_debug_to_text(struct printbuf *out, struct journal *j)
 	unsigned long now = jiffies;
 	u64 nr_writes = j->nr_flush_writes + j->nr_noflush_writes;
 
-	if (!out->nr_tabstops)
-		printbuf_tabstop_push(out, 24);
+	printbuf_tabstops_reset(out);
+	printbuf_tabstop_push(out, 28);
 	out->atomic++;
 
 	rcu_read_lock();
 	s = READ_ONCE(j->reservations);
 
+	prt_printf(out, "flags:\t");
+	prt_bitflags(out, bch2_journal_flags_strs, j->flags);
+	prt_newline(out);
 	prt_printf(out, "dirty journal entries:\t%llu/%llu\n",	fifo_used(&j->pin), j->pin.size);
-	prt_printf(out, "seq:\t\t\t%llu\n",			journal_cur_seq(j));
-	prt_printf(out, "seq_ondisk:\t\t%llu\n",		j->seq_ondisk);
-	prt_printf(out, "last_seq:\t\t%llu\n",			journal_last_seq(j));
+	prt_printf(out, "seq:\t%llu\n",				journal_cur_seq(j));
+	prt_printf(out, "seq_ondisk:\t%llu\n",			j->seq_ondisk);
+	prt_printf(out, "last_seq:\t%llu\n",			journal_last_seq(j));
 	prt_printf(out, "last_seq_ondisk:\t%llu\n",		j->last_seq_ondisk);
 	prt_printf(out, "flushed_seq_ondisk:\t%llu\n",		j->flushed_seq_ondisk);
-	prt_printf(out, "watermark:\t\t%s\n",			bch2_watermarks[j->watermark]);
+	prt_printf(out, "watermark:\t%s\n",			bch2_watermarks[j->watermark]);
 	prt_printf(out, "each entry reserved:\t%u\n",		j->entry_u64s_reserved);
 	prt_printf(out, "nr flush writes:\t%llu\n",		j->nr_flush_writes);
 	prt_printf(out, "nr noflush writes:\t%llu\n",		j->nr_noflush_writes);
@@ -1378,47 +1445,44 @@ void __bch2_journal_debug_to_text(struct printbuf *out, struct journal *j)
 	prt_newline(out);
 	prt_printf(out, "nr direct reclaim:\t%llu\n",		j->nr_direct_reclaim);
 	prt_printf(out, "nr background reclaim:\t%llu\n",	j->nr_background_reclaim);
-	prt_printf(out, "reclaim kicked:\t\t%u\n",		j->reclaim_kicked);
+	prt_printf(out, "reclaim kicked:\t%u\n",		j->reclaim_kicked);
 	prt_printf(out, "reclaim runs in:\t%u ms\n",		time_after(j->next_reclaim, now)
 	       ? jiffies_to_msecs(j->next_reclaim - jiffies) : 0);
+	prt_printf(out, "blocked:\t%u\n",			j->blocked);
 	prt_printf(out, "current entry sectors:\t%u\n",		j->cur_entry_sectors);
 	prt_printf(out, "current entry error:\t%s\n",		bch2_journal_errors[j->cur_entry_error]);
-	prt_printf(out, "current entry:\t\t");
+	prt_printf(out, "current entry:\t");
 
 	switch (s.cur_entry_offset) {
 	case JOURNAL_ENTRY_ERROR_VAL:
-		prt_printf(out, "error");
+		prt_printf(out, "error\n");
 		break;
 	case JOURNAL_ENTRY_CLOSED_VAL:
-		prt_printf(out, "closed");
+		prt_printf(out, "closed\n");
 		break;
 	default:
-		prt_printf(out, "%u/%u", s.cur_entry_offset, j->cur_entry_u64s);
+		prt_printf(out, "%u/%u\n", s.cur_entry_offset, j->cur_entry_u64s);
 		break;
 	}
 
-	prt_newline(out);
-	prt_printf(out, "unwritten entries:");
-	prt_newline(out);
+	prt_printf(out, "unwritten entries:\n");
 	bch2_journal_bufs_to_text(out, j);
 
-	prt_printf(out,
-	       "replay done:\t\t%i\n",
-	       test_bit(JOURNAL_REPLAY_DONE,	&j->flags));
-
 	prt_printf(out, "space:\n");
-	prt_printf(out, "\tdiscarded\t%u:%u\n",
+	printbuf_indent_add(out, 2);
+	prt_printf(out, "discarded\t%u:%u\n",
 	       j->space[journal_space_discarded].next_entry,
 	       j->space[journal_space_discarded].total);
-	prt_printf(out, "\tclean ondisk\t%u:%u\n",
+	prt_printf(out, "clean ondisk\t%u:%u\n",
 	       j->space[journal_space_clean_ondisk].next_entry,
 	       j->space[journal_space_clean_ondisk].total);
-	prt_printf(out, "\tclean\t\t%u:%u\n",
+	prt_printf(out, "clean\t%u:%u\n",
 	       j->space[journal_space_clean].next_entry,
 	       j->space[journal_space_clean].total);
-	prt_printf(out, "\ttotal\t\t%u:%u\n",
+	prt_printf(out, "total\t%u:%u\n",
 	       j->space[journal_space_total].next_entry,
 	       j->space[journal_space_total].total);
+	printbuf_indent_sub(out, 2);
 
 	for_each_member_device_rcu(c, ca, &c->rw_devs[BCH_DATA_journal]) {
 		struct journal_device *ja = &ca->journal;
@@ -1429,14 +1493,16 @@ void __bch2_journal_debug_to_text(struct printbuf *out, struct journal *j)
 		if (!ja->nr)
 			continue;
 
-		prt_printf(out, "dev %u:\n",		ca->dev_idx);
-		prt_printf(out, "\tnr\t\t%u\n",		ja->nr);
-		prt_printf(out, "\tbucket size\t%u\n",	ca->mi.bucket_size);
-		prt_printf(out, "\tavailable\t%u:%u\n",	bch2_journal_dev_buckets_available(j, ja, journal_space_discarded), ja->sectors_free);
-		prt_printf(out, "\tdiscard_idx\t%u\n",	ja->discard_idx);
-		prt_printf(out, "\tdirty_ondisk\t%u (seq %llu)\n", ja->dirty_idx_ondisk,	ja->bucket_seq[ja->dirty_idx_ondisk]);
-		prt_printf(out, "\tdirty_idx\t%u (seq %llu)\n", ja->dirty_idx,		ja->bucket_seq[ja->dirty_idx]);
-		prt_printf(out, "\tcur_idx\t\t%u (seq %llu)\n", ja->cur_idx,		ja->bucket_seq[ja->cur_idx]);
+		prt_printf(out, "dev %u:\n",			ca->dev_idx);
+		printbuf_indent_add(out, 2);
+		prt_printf(out, "nr\t%u\n",			ja->nr);
+		prt_printf(out, "bucket size\t%u\n",		ca->mi.bucket_size);
+		prt_printf(out, "available\t%u:%u\n",		bch2_journal_dev_buckets_available(j, ja, journal_space_discarded), ja->sectors_free);
+		prt_printf(out, "discard_idx\t%u\n",		ja->discard_idx);
+		prt_printf(out, "dirty_ondisk\t%u (seq %llu)\n",ja->dirty_idx_ondisk,	ja->bucket_seq[ja->dirty_idx_ondisk]);
+		prt_printf(out, "dirty_idx\t%u (seq %llu)\n",	ja->dirty_idx,		ja->bucket_seq[ja->dirty_idx]);
+		prt_printf(out, "cur_idx\t%u (seq %llu)\n",	ja->cur_idx,		ja->bucket_seq[ja->cur_idx]);
+		printbuf_indent_sub(out, 2);
 	}
 
 	rcu_read_unlock();
@@ -1455,9 +1521,13 @@ bool bch2_journal_seq_pins_to_text(struct printbuf *out, struct journal *j, u64
 {
 	struct journal_entry_pin_list *pin_list;
 	struct journal_entry_pin *pin;
-	unsigned i;
 
 	spin_lock(&j->lock);
+	if (!test_bit(JOURNAL_running, &j->flags)) {
+		spin_unlock(&j->lock);
+		return true;
+	}
+
 	*seq = max(*seq, j->pin.front);
 
 	if (*seq >= j->pin.back) {
@@ -1469,25 +1539,18 @@ bool bch2_journal_seq_pins_to_text(struct printbuf *out, struct journal *j, u64
 
 	pin_list = journal_seq_pin(j, *seq);
 
-	prt_printf(out, "%llu: count %u", *seq, atomic_read(&pin_list->count));
-	prt_newline(out);
+	prt_printf(out, "%llu: count %u\n", *seq, atomic_read(&pin_list->count));
 	printbuf_indent_add(out, 2);
 
-	for (i = 0; i < ARRAY_SIZE(pin_list->list); i++)
-		list_for_each_entry(pin, &pin_list->list[i], list) {
-			prt_printf(out, "\t%px %ps", pin, pin->flush);
-			prt_newline(out);
-		}
+	for (unsigned i = 0; i < ARRAY_SIZE(pin_list->list); i++)
+		list_for_each_entry(pin, &pin_list->list[i], list)
+			prt_printf(out, "\t%px %ps\n", pin, pin->flush);
 
-	if (!list_empty(&pin_list->flushed)) {
-		prt_printf(out, "flushed:");
-		prt_newline(out);
-	}
+	if (!list_empty(&pin_list->flushed))
+		prt_printf(out, "flushed:\n");
 
-	list_for_each_entry(pin, &pin_list->flushed, list) {
-		prt_printf(out, "\t%px %ps", pin, pin->flush);
-		prt_newline(out);
-	}
+	list_for_each_entry(pin, &pin_list->flushed, list)
+		prt_printf(out, "\t%px %ps\n", pin, pin->flush);
 
 	printbuf_indent_sub(out, 2);
 
diff --git a/fs/bcachefs/journal.h b/fs/bcachefs/journal.h
index 4544ce24bb..bc6b9c39dc 100644
--- a/fs/bcachefs/journal.h
+++ b/fs/bcachefs/journal.h
@@ -264,7 +264,8 @@ static inline union journal_res_state journal_state_buf_put(struct journal *j, u
 }
 
 bool bch2_journal_entry_close(struct journal *);
-void bch2_journal_buf_put_final(struct journal *, u64, bool);
+void bch2_journal_do_writes(struct journal *);
+void bch2_journal_buf_put_final(struct journal *, u64);
 
 static inline void __bch2_journal_buf_put(struct journal *j, unsigned idx, u64 seq)
 {
@@ -272,7 +273,7 @@ static inline void __bch2_journal_buf_put(struct journal *j, unsigned idx, u64 s
 
 	s = journal_state_buf_put(j, idx);
 	if (!journal_state_count(s, idx))
-		bch2_journal_buf_put_final(j, seq, idx == s.unwritten_idx);
+		bch2_journal_buf_put_final(j, seq);
 }
 
 static inline void bch2_journal_buf_put(struct journal *j, unsigned idx, u64 seq)
@@ -282,7 +283,7 @@ static inline void bch2_journal_buf_put(struct journal *j, unsigned idx, u64 seq
 	s = journal_state_buf_put(j, idx);
 	if (!journal_state_count(s, idx)) {
 		spin_lock(&j->lock);
-		bch2_journal_buf_put_final(j, seq, idx == s.unwritten_idx);
+		bch2_journal_buf_put_final(j, seq);
 		spin_unlock(&j->lock);
 	}
 }
@@ -371,7 +372,7 @@ static inline int bch2_journal_res_get(struct journal *j, struct journal_res *re
 	int ret;
 
 	EBUG_ON(res->ref);
-	EBUG_ON(!test_bit(JOURNAL_STARTED, &j->flags));
+	EBUG_ON(!test_bit(JOURNAL_running, &j->flags));
 
 	res->u64s = u64s;
 
@@ -417,8 +418,8 @@ struct bch_dev;
 
 static inline void bch2_journal_set_replay_done(struct journal *j)
 {
-	BUG_ON(!test_bit(JOURNAL_STARTED, &j->flags));
-	set_bit(JOURNAL_REPLAY_DONE, &j->flags);
+	BUG_ON(!test_bit(JOURNAL_running, &j->flags));
+	set_bit(JOURNAL_replay_done, &j->flags);
 }
 
 void bch2_journal_unblock(struct journal *);
@@ -432,7 +433,7 @@ bool bch2_journal_seq_pins_to_text(struct printbuf *, struct journal *, u64 *);
 
 int bch2_set_nr_journal_buckets(struct bch_fs *, struct bch_dev *,
 				unsigned nr);
-int bch2_dev_journal_alloc(struct bch_dev *);
+int bch2_dev_journal_alloc(struct bch_dev *, bool);
 int bch2_fs_journal_alloc(struct bch_fs *);
 
 void bch2_dev_journal_stop(struct journal *, struct bch_dev *);
diff --git a/fs/bcachefs/journal_io.c b/fs/bcachefs/journal_io.c
index 47805193f1..2326e2cb9c 100644
--- a/fs/bcachefs/journal_io.c
+++ b/fs/bcachefs/journal_io.c
@@ -17,6 +17,60 @@
 #include "sb-clean.h"
 #include "trace.h"
 
+void bch2_journal_pos_from_member_info_set(struct bch_fs *c)
+{
+	lockdep_assert_held(&c->sb_lock);
+
+	for_each_member_device(c, ca) {
+		struct bch_member *m = bch2_members_v2_get_mut(c->disk_sb.sb, ca->dev_idx);
+
+		m->last_journal_bucket = cpu_to_le32(ca->journal.cur_idx);
+		m->last_journal_bucket_offset = cpu_to_le32(ca->mi.bucket_size - ca->journal.sectors_free);
+	}
+}
+
+void bch2_journal_pos_from_member_info_resume(struct bch_fs *c)
+{
+	mutex_lock(&c->sb_lock);
+	for_each_member_device(c, ca) {
+		struct bch_member m = bch2_sb_member_get(c->disk_sb.sb, ca->dev_idx);
+
+		unsigned idx = le32_to_cpu(m.last_journal_bucket);
+		if (idx < ca->journal.nr)
+			ca->journal.cur_idx = idx;
+		unsigned offset = le32_to_cpu(m.last_journal_bucket_offset);
+		if (offset <= ca->mi.bucket_size)
+			ca->journal.sectors_free = ca->mi.bucket_size - offset;
+	}
+	mutex_unlock(&c->sb_lock);
+}
+
+void bch2_journal_ptrs_to_text(struct printbuf *out, struct bch_fs *c,
+			       struct journal_replay *j)
+{
+	darray_for_each(j->ptrs, i) {
+		if (i != j->ptrs.data)
+			prt_printf(out, " ");
+		prt_printf(out, "%u:%u:%u (sector %llu)",
+			   i->dev, i->bucket, i->bucket_offset, i->sector);
+	}
+}
+
+static void bch2_journal_replay_to_text(struct printbuf *out, struct bch_fs *c,
+					struct journal_replay *j)
+{
+	prt_printf(out, "seq %llu ", le64_to_cpu(j->j.seq));
+
+	bch2_journal_ptrs_to_text(out, c, j);
+
+	for_each_jset_entry_type(entry, &j->j, BCH_JSET_ENTRY_datetime) {
+		struct jset_entry_datetime *datetime =
+			container_of(entry, struct jset_entry_datetime, entry);
+		bch2_prt_datetime(out, le64_to_cpu(datetime->seconds));
+		break;
+	}
+}
+
 static struct nonce journal_nonce(const struct jset *jset)
 {
 	return (struct nonce) {{
@@ -52,13 +106,15 @@ static void __journal_replay_free(struct bch_fs *c,
 
 	BUG_ON(*p != i);
 	*p = NULL;
-	kvpfree(i, offsetof(struct journal_replay, j) +
-		vstruct_bytes(&i->j));
+	kvfree(i);
 }
 
-static void journal_replay_free(struct bch_fs *c, struct journal_replay *i)
+static void journal_replay_free(struct bch_fs *c, struct journal_replay *i, bool blacklisted)
 {
-	i->ignore = true;
+	if (blacklisted)
+		i->ignore_blacklisted = true;
+	else
+		i->ignore_not_dirty = true;
 
 	if (!c->opts.read_entire_journal)
 		__journal_replay_free(c, i);
@@ -84,11 +140,15 @@ static int journal_entry_add(struct bch_fs *c, struct bch_dev *ca,
 {
 	struct genradix_iter iter;
 	struct journal_replay **_i, *i, *dup;
-	struct journal_ptr *ptr;
 	size_t bytes = vstruct_bytes(j);
 	u64 last_seq = !JSET_NO_FLUSH(j) ? le64_to_cpu(j->last_seq) : 0;
+	struct printbuf buf = PRINTBUF;
 	int ret = JOURNAL_ENTRY_ADD_OK;
 
+	if (!c->journal.oldest_seq_found_ondisk ||
+	    le64_to_cpu(j->seq) < c->journal.oldest_seq_found_ondisk)
+		c->journal.oldest_seq_found_ondisk = le64_to_cpu(j->seq);
+
 	/* Is this entry older than the range we need? */
 	if (!c->opts.read_entire_journal &&
 	    le64_to_cpu(j->seq) < jlist->last_seq)
@@ -108,12 +168,13 @@ static int journal_entry_add(struct bch_fs *c, struct bch_dev *ca,
 				       journal_entry_radix_idx(c, jlist->last_seq)) {
 			i = *_i;
 
-			if (!i || i->ignore)
+			if (journal_replay_ignore(i))
 				continue;
 
 			if (le64_to_cpu(i->j.seq) >= last_seq)
 				break;
-			journal_replay_free(c, i);
+
+			journal_replay_free(c, i, false);
 		}
 	}
 
@@ -131,72 +192,62 @@ static int journal_entry_add(struct bch_fs *c, struct bch_dev *ca,
 	 */
 	dup = *_i;
 	if (dup) {
-		if (bytes == vstruct_bytes(&dup->j) &&
-		    !memcmp(j, &dup->j, bytes)) {
-			i = dup;
-			goto found;
-		}
+		bool identical = bytes == vstruct_bytes(&dup->j) &&
+			!memcmp(j, &dup->j, bytes);
+		bool not_identical = !identical &&
+			entry_ptr.csum_good &&
+			dup->csum_good;
+
+		bool same_device = false;
+		darray_for_each(dup->ptrs, ptr)
+			if (ptr->dev == ca->dev_idx)
+				same_device = true;
+
+		ret = darray_push(&dup->ptrs, entry_ptr);
+		if (ret)
+			goto out;
 
-		if (!entry_ptr.csum_good) {
-			i = dup;
-			goto found;
-		}
+		bch2_journal_replay_to_text(&buf, c, dup);
+
+		fsck_err_on(same_device,
+			    c, journal_entry_dup_same_device,
+			    "duplicate journal entry on same device\n  %s",
+			    buf.buf);
+
+		fsck_err_on(not_identical,
+			    c, journal_entry_replicas_data_mismatch,
+			    "found duplicate but non identical journal entries\n  %s",
+			    buf.buf);
 
-		if (!dup->csum_good)
+		if (entry_ptr.csum_good && !identical)
 			goto replace;
 
-		fsck_err(c, journal_entry_replicas_data_mismatch,
-			 "found duplicate but non identical journal entries (seq %llu)",
-			 le64_to_cpu(j->seq));
-		i = dup;
-		goto found;
+		goto out;
 	}
 replace:
-	i = kvpmalloc(offsetof(struct journal_replay, j) + bytes, GFP_KERNEL);
+	i = kvmalloc(offsetof(struct journal_replay, j) + bytes, GFP_KERNEL);
 	if (!i)
 		return -BCH_ERR_ENOMEM_journal_entry_add;
 
-	i->nr_ptrs	= 0;
-	i->csum_good	= entry_ptr.csum_good;
-	i->ignore	= false;
+	darray_init(&i->ptrs);
+	i->csum_good		= entry_ptr.csum_good;
+	i->ignore_blacklisted	= false;
+	i->ignore_not_dirty	= false;
 	unsafe_memcpy(&i->j, j, bytes, "embedded variable length struct");
-	i->ptrs[i->nr_ptrs++] = entry_ptr;
 
 	if (dup) {
-		if (dup->nr_ptrs >= ARRAY_SIZE(dup->ptrs)) {
-			bch_err(c, "found too many copies of journal entry %llu",
-				le64_to_cpu(i->j.seq));
-			dup->nr_ptrs = ARRAY_SIZE(dup->ptrs) - 1;
-		}
-
 		/* The first ptr should represent the jset we kept: */
-		memcpy(i->ptrs + i->nr_ptrs,
-		       dup->ptrs,
-		       sizeof(dup->ptrs[0]) * dup->nr_ptrs);
-		i->nr_ptrs += dup->nr_ptrs;
+		darray_for_each(dup->ptrs, ptr)
+			darray_push(&i->ptrs, *ptr);
 		__journal_replay_free(c, dup);
+	} else {
+		darray_push(&i->ptrs, entry_ptr);
 	}
 
 	*_i = i;
-	return 0;
-found:
-	for (ptr = i->ptrs; ptr < i->ptrs + i->nr_ptrs; ptr++) {
-		if (ptr->dev == ca->dev_idx) {
-			bch_err(c, "duplicate journal entry %llu on same device",
-				le64_to_cpu(i->j.seq));
-			goto out;
-		}
-	}
-
-	if (i->nr_ptrs >= ARRAY_SIZE(i->ptrs)) {
-		bch_err(c, "found too many copies of journal entry %llu",
-			le64_to_cpu(i->j.seq));
-		goto out;
-	}
-
-	i->ptrs[i->nr_ptrs++] = entry_ptr;
 out:
 fsck_err:
+	printbuf_exit(&buf);
 	return ret;
 }
 
@@ -223,7 +274,7 @@ static void journal_entry_err_msg(struct printbuf *out,
 
 	if (entry) {
 		prt_str(out, " type=");
-		prt_str(out, bch2_jset_entry_types[entry->type]);
+		bch2_prt_jset_entry_type(out, entry->type);
 	}
 
 	if (!jset) {
@@ -248,7 +299,7 @@ static void journal_entry_err_msg(struct printbuf *out,
 	journal_entry_err_msg(&_buf, version, jset, entry);		\
 	prt_printf(&_buf, msg, ##__VA_ARGS__);				\
 									\
-	switch (flags & BKEY_INVALID_WRITE) {				\
+	switch (flags & BCH_VALIDATE_write) {				\
 	case READ:							\
 		mustfix_fsck_err(c, _err, "%s", _buf.buf);		\
 		break;							\
@@ -277,9 +328,9 @@ static int journal_validate_key(struct bch_fs *c,
 				unsigned level, enum btree_id btree_id,
 				struct bkey_i *k,
 				unsigned version, int big_endian,
-				enum bkey_invalid_flags flags)
+				enum bch_validate_flags flags)
 {
-	int write = flags & BKEY_INVALID_WRITE;
+	int write = flags & BCH_VALIDATE_write;
 	void *next = vstruct_next(entry);
 	struct printbuf buf = PRINTBUF;
 	int ret = 0;
@@ -352,7 +403,7 @@ static int journal_entry_btree_keys_validate(struct bch_fs *c,
 				struct jset *jset,
 				struct jset_entry *entry,
 				unsigned version, int big_endian,
-				enum bkey_invalid_flags flags)
+				enum bch_validate_flags flags)
 {
 	struct bkey_i *k = entry->start;
 
@@ -361,9 +412,11 @@ static int journal_entry_btree_keys_validate(struct bch_fs *c,
 					       entry->level,
 					       entry->btree_id,
 					       k, version, big_endian,
-					       flags|BKEY_INVALID_JOURNAL);
+					       flags|BCH_VALIDATE_journal);
 		if (ret == FSCK_DELETED_KEY)
 			continue;
+		else if (ret)
+			return ret;
 
 		k = bkey_next(k);
 	}
@@ -374,13 +427,13 @@ static int journal_entry_btree_keys_validate(struct bch_fs *c,
 static void journal_entry_btree_keys_to_text(struct printbuf *out, struct bch_fs *c,
 					     struct jset_entry *entry)
 {
-	struct bkey_i *k;
 	bool first = true;
 
 	jset_entry_for_each_key(entry, k) {
 		if (!first) {
 			prt_newline(out);
-			prt_printf(out, "%s: ", bch2_jset_entry_types[entry->type]);
+			bch2_prt_jset_entry_type(out, entry->type);
+			prt_str(out, ": ");
 		}
 		prt_printf(out, "btree=%s l=%u ", bch2_btree_id_str(entry->btree_id), entry->level);
 		bch2_bkey_val_to_text(out, c, bkey_i_to_s_c(k));
@@ -392,7 +445,7 @@ static int journal_entry_btree_root_validate(struct bch_fs *c,
 				struct jset *jset,
 				struct jset_entry *entry,
 				unsigned version, int big_endian,
-				enum bkey_invalid_flags flags)
+				enum bch_validate_flags flags)
 {
 	struct bkey_i *k = entry->start;
 	int ret = 0;
@@ -431,7 +484,7 @@ static int journal_entry_prio_ptrs_validate(struct bch_fs *c,
 				struct jset *jset,
 				struct jset_entry *entry,
 				unsigned version, int big_endian,
-				enum bkey_invalid_flags flags)
+				enum bch_validate_flags flags)
 {
 	/* obsolete, don't care: */
 	return 0;
@@ -446,7 +499,7 @@ static int journal_entry_blacklist_validate(struct bch_fs *c,
 				struct jset *jset,
 				struct jset_entry *entry,
 				unsigned version, int big_endian,
-				enum bkey_invalid_flags flags)
+				enum bch_validate_flags flags)
 {
 	int ret = 0;
 
@@ -473,7 +526,7 @@ static int journal_entry_blacklist_v2_validate(struct bch_fs *c,
 				struct jset *jset,
 				struct jset_entry *entry,
 				unsigned version, int big_endian,
-				enum bkey_invalid_flags flags)
+				enum bch_validate_flags flags)
 {
 	struct jset_entry_blacklist_v2 *bl_entry;
 	int ret = 0;
@@ -515,7 +568,7 @@ static int journal_entry_usage_validate(struct bch_fs *c,
 				struct jset *jset,
 				struct jset_entry *entry,
 				unsigned version, int big_endian,
-				enum bkey_invalid_flags flags)
+				enum bch_validate_flags flags)
 {
 	struct jset_entry_usage *u =
 		container_of(entry, struct jset_entry_usage, entry);
@@ -540,16 +593,16 @@ static void journal_entry_usage_to_text(struct printbuf *out, struct bch_fs *c,
 	struct jset_entry_usage *u =
 		container_of(entry, struct jset_entry_usage, entry);
 
-	prt_printf(out, "type=%s v=%llu",
-	       bch2_fs_usage_types[u->entry.btree_id],
-	       le64_to_cpu(u->v));
+	prt_str(out, "type=");
+	bch2_prt_fs_usage_type(out, u->entry.btree_id);
+	prt_printf(out, " v=%llu", le64_to_cpu(u->v));
 }
 
 static int journal_entry_data_usage_validate(struct bch_fs *c,
 				struct jset *jset,
 				struct jset_entry *entry,
 				unsigned version, int big_endian,
-				enum bkey_invalid_flags flags)
+				enum bch_validate_flags flags)
 {
 	struct jset_entry_data_usage *u =
 		container_of(entry, struct jset_entry_data_usage, entry);
@@ -593,7 +646,7 @@ static int journal_entry_clock_validate(struct bch_fs *c,
 				struct jset *jset,
 				struct jset_entry *entry,
 				unsigned version, int big_endian,
-				enum bkey_invalid_flags flags)
+				enum bch_validate_flags flags)
 {
 	struct jset_entry_clock *clock =
 		container_of(entry, struct jset_entry_clock, entry);
@@ -633,13 +686,12 @@ static int journal_entry_dev_usage_validate(struct bch_fs *c,
 				struct jset *jset,
 				struct jset_entry *entry,
 				unsigned version, int big_endian,
-				enum bkey_invalid_flags flags)
+				enum bch_validate_flags flags)
 {
 	struct jset_entry_dev_usage *u =
 		container_of(entry, struct jset_entry_dev_usage, entry);
 	unsigned bytes = jset_u64s(le16_to_cpu(entry->u64s)) * sizeof(u64);
 	unsigned expected = sizeof(*u);
-	unsigned dev;
 	int ret = 0;
 
 	if (journal_entry_err_on(bytes < expected,
@@ -651,16 +703,6 @@ static int journal_entry_dev_usage_validate(struct bch_fs *c,
 		return ret;
 	}
 
-	dev = le32_to_cpu(u->dev);
-
-	if (journal_entry_err_on(!bch2_dev_exists2(c, dev),
-				 c, version, jset, entry,
-				 journal_entry_dev_usage_bad_dev,
-				 "bad dev")) {
-		journal_entry_null_range(entry, vstruct_next(entry));
-		return ret;
-	}
-
 	if (journal_entry_err_on(u->pad,
 				 c, version, jset, entry,
 				 journal_entry_dev_usage_bad_pad,
@@ -695,7 +737,7 @@ static int journal_entry_log_validate(struct bch_fs *c,
 				struct jset *jset,
 				struct jset_entry *entry,
 				unsigned version, int big_endian,
-				enum bkey_invalid_flags flags)
+				enum bch_validate_flags flags)
 {
 	return 0;
 }
@@ -713,7 +755,7 @@ static int journal_entry_overwrite_validate(struct bch_fs *c,
 				struct jset *jset,
 				struct jset_entry *entry,
 				unsigned version, int big_endian,
-				enum bkey_invalid_flags flags)
+				enum bch_validate_flags flags)
 {
 	return journal_entry_btree_keys_validate(c, jset, entry,
 				version, big_endian, READ);
@@ -729,7 +771,7 @@ static int journal_entry_write_buffer_keys_validate(struct bch_fs *c,
 				struct jset *jset,
 				struct jset_entry *entry,
 				unsigned version, int big_endian,
-				enum bkey_invalid_flags flags)
+				enum bch_validate_flags flags)
 {
 	return journal_entry_btree_keys_validate(c, jset, entry,
 				version, big_endian, READ);
@@ -741,10 +783,41 @@ static void journal_entry_write_buffer_keys_to_text(struct printbuf *out, struct
 	journal_entry_btree_keys_to_text(out, c, entry);
 }
 
+static int journal_entry_datetime_validate(struct bch_fs *c,
+				struct jset *jset,
+				struct jset_entry *entry,
+				unsigned version, int big_endian,
+				enum bch_validate_flags flags)
+{
+	unsigned bytes = vstruct_bytes(entry);
+	unsigned expected = 16;
+	int ret = 0;
+
+	if (journal_entry_err_on(vstruct_bytes(entry) < expected,
+				 c, version, jset, entry,
+				 journal_entry_dev_usage_bad_size,
+				 "bad size (%u < %u)",
+				 bytes, expected)) {
+		journal_entry_null_range(entry, vstruct_next(entry));
+		return ret;
+	}
+fsck_err:
+	return ret;
+}
+
+static void journal_entry_datetime_to_text(struct printbuf *out, struct bch_fs *c,
+					    struct jset_entry *entry)
+{
+	struct jset_entry_datetime *datetime =
+		container_of(entry, struct jset_entry_datetime, entry);
+
+	bch2_prt_datetime(out, le64_to_cpu(datetime->seconds));
+}
+
 struct jset_entry_ops {
 	int (*validate)(struct bch_fs *, struct jset *,
 			struct jset_entry *, unsigned, int,
-			enum bkey_invalid_flags);
+			enum bch_validate_flags);
 	void (*to_text)(struct printbuf *, struct bch_fs *, struct jset_entry *);
 };
 
@@ -762,7 +835,7 @@ int bch2_journal_entry_validate(struct bch_fs *c,
 				struct jset *jset,
 				struct jset_entry *entry,
 				unsigned version, int big_endian,
-				enum bkey_invalid_flags flags)
+				enum bch_validate_flags flags)
 {
 	return entry->type < BCH_JSET_ENTRY_NR
 		? bch2_jset_entry_ops[entry->type].validate(c, jset, entry,
@@ -773,16 +846,16 @@ int bch2_journal_entry_validate(struct bch_fs *c,
 void bch2_journal_entry_to_text(struct printbuf *out, struct bch_fs *c,
 				struct jset_entry *entry)
 {
+	bch2_prt_jset_entry_type(out, entry->type);
+
 	if (entry->type < BCH_JSET_ENTRY_NR) {
-		prt_printf(out, "%s: ", bch2_jset_entry_types[entry->type]);
+		prt_str(out, ": ");
 		bch2_jset_entry_ops[entry->type].to_text(out, c, entry);
-	} else {
-		prt_printf(out, "(unknown type %u)", entry->type);
 	}
 }
 
 static int jset_validate_entries(struct bch_fs *c, struct jset *jset,
-				 enum bkey_invalid_flags flags)
+				 enum bch_validate_flags flags)
 {
 	unsigned version = le32_to_cpu(jset->version);
 	int ret = 0;
@@ -808,7 +881,7 @@ fsck_err:
 static int jset_validate(struct bch_fs *c,
 			 struct bch_dev *ca,
 			 struct jset *jset, u64 sector,
-			 enum bkey_invalid_flags flags)
+			 enum bch_validate_flags flags)
 {
 	unsigned version;
 	int ret = 0;
@@ -863,7 +936,7 @@ static int jset_validate_early(struct bch_fs *c,
 {
 	size_t bytes = vstruct_bytes(jset);
 	unsigned version;
-	enum bkey_invalid_flags flags = BKEY_INVALID_JOURNAL;
+	enum bch_validate_flags flags = BCH_VALIDATE_journal;
 	int ret = 0;
 
 	if (le64_to_cpu(jset->magic) != jset_magic(c))
@@ -913,11 +986,11 @@ static int journal_read_buf_realloc(struct journal_read_buf *b,
 		return -BCH_ERR_ENOMEM_journal_read_buf_realloc;
 
 	new_size = roundup_pow_of_two(new_size);
-	n = kvpmalloc(new_size, GFP_KERNEL);
+	n = kvmalloc(new_size, GFP_KERNEL);
 	if (!n)
 		return -BCH_ERR_ENOMEM_journal_read_buf_realloc;
 
-	kvpfree(b->data, b->size);
+	kvfree(b->data);
 	b->data = n;
 	b->size = new_size;
 	return 0;
@@ -1002,6 +1075,13 @@ reread:
 			goto err;
 		}
 
+		if (le64_to_cpu(j->seq) > ja->highest_seq_found) {
+			ja->highest_seq_found = le64_to_cpu(j->seq);
+			ja->cur_idx = bucket;
+			ja->sectors_free = ca->mi.bucket_size -
+				bucket_remainder(ca, offset) - sectors;
+		}
+
 		/*
 		 * This happens sometimes if we don't have discards on -
 		 * when we've partially overwritten a bucket with new
@@ -1028,9 +1108,7 @@ reread:
 		ret = bch2_encrypt(c, JSET_CSUM_TYPE(j), journal_nonce(j),
 			     j->encrypted_start,
 			     vstruct_end(j) - (void *) j->encrypted_start);
-		bch2_fs_fatal_err_on(ret, c,
-				"error decrypting journal entry: %s",
-				bch2_err_str(ret));
+		bch2_fs_fatal_err_on(ret, c, "decrypting journal entry: %s", bch2_err_str(ret));
 
 		mutex_lock(&jlist->lock);
 		ret = journal_entry_add(c, ca, (struct journal_ptr) {
@@ -1072,8 +1150,6 @@ static CLOSURE_CALLBACK(bch2_journal_read_device)
 	struct bch_fs *c = ca->fs;
 	struct journal_list *jlist =
 		container_of(cl->parent, struct journal_list, cl);
-	struct journal_replay *r, **_r;
-	struct genradix_iter iter;
 	struct journal_read_buf buf = { NULL, 0 };
 	unsigned i;
 	int ret = 0;
@@ -1093,48 +1169,6 @@ static CLOSURE_CALLBACK(bch2_journal_read_device)
 			goto err;
 	}
 
-	ja->sectors_free = ca->mi.bucket_size;
-
-	mutex_lock(&jlist->lock);
-	genradix_for_each_reverse(&c->journal_entries, iter, _r) {
-		r = *_r;
-
-		if (!r)
-			continue;
-
-		for (i = 0; i < r->nr_ptrs; i++) {
-			if (r->ptrs[i].dev == ca->dev_idx) {
-				unsigned wrote = bucket_remainder(ca, r->ptrs[i].sector) +
-					vstruct_sectors(&r->j, c->block_bits);
-
-				ja->cur_idx = r->ptrs[i].bucket;
-				ja->sectors_free = ca->mi.bucket_size - wrote;
-				goto found;
-			}
-		}
-	}
-found:
-	mutex_unlock(&jlist->lock);
-
-	if (ja->bucket_seq[ja->cur_idx] &&
-	    ja->sectors_free == ca->mi.bucket_size) {
-#if 0
-		/*
-		 * Debug code for ZNS support, where we (probably) want to be
-		 * correlated where we stopped in the journal to the zone write
-		 * points:
-		 */
-		bch_err(c, "ja->sectors_free == ca->mi.bucket_size");
-		bch_err(c, "cur_idx %u/%u", ja->cur_idx, ja->nr);
-		for (i = 0; i < 3; i++) {
-			unsigned idx = (ja->cur_idx + ja->nr - 1 + i) % ja->nr;
-
-			bch_err(c, "bucket_seq[%u] = %llu", idx, ja->bucket_seq[idx]);
-		}
-#endif
-		ja->sectors_free = 0;
-	}
-
 	/*
 	 * Set dirty_idx to indicate the entire journal is full and needs to be
 	 * reclaimed - journal reclaim will immediately reclaim whatever isn't
@@ -1144,7 +1178,7 @@ found:
 		ja->dirty_idx = (ja->cur_idx + 1) % ja->nr;
 out:
 	bch_verbose(c, "journal read done on device %s, ret %i", ca->name, ret);
-	kvpfree(buf.data, buf.size);
+	kvfree(buf.data);
 	percpu_ref_put(&ca->io_ref);
 	closure_return(cl);
 	return;
@@ -1155,27 +1189,6 @@ err:
 	goto out;
 }
 
-void bch2_journal_ptrs_to_text(struct printbuf *out, struct bch_fs *c,
-			       struct journal_replay *j)
-{
-	unsigned i;
-
-	for (i = 0; i < j->nr_ptrs; i++) {
-		struct bch_dev *ca = bch_dev_bkey_exists(c, j->ptrs[i].dev);
-		u64 offset;
-
-		div64_u64_rem(j->ptrs[i].sector, ca->mi.bucket_size, &offset);
-
-		if (i)
-			prt_printf(out, " ");
-		prt_printf(out, "%u:%u:%u (sector %llu)",
-		       j->ptrs[i].dev,
-		       j->ptrs[i].bucket,
-		       j->ptrs[i].bucket_offset,
-		       j->ptrs[i].sector);
-	}
-}
-
 int bch2_journal_read(struct bch_fs *c,
 		      u64 *last_seq,
 		      u64 *blacklist_seq,
@@ -1224,24 +1237,24 @@ int bch2_journal_read(struct bch_fs *c,
 	 * those entries will be blacklisted:
 	 */
 	genradix_for_each_reverse(&c->journal_entries, radix_iter, _i) {
-		enum bkey_invalid_flags flags = BKEY_INVALID_JOURNAL;
+		enum bch_validate_flags flags = BCH_VALIDATE_journal;
 
 		i = *_i;
 
-		if (!i || i->ignore)
+		if (journal_replay_ignore(i))
 			continue;
 
 		if (!*start_seq)
 			*blacklist_seq = *start_seq = le64_to_cpu(i->j.seq) + 1;
 
 		if (JSET_NO_FLUSH(&i->j)) {
-			i->ignore = true;
+			i->ignore_blacklisted = true;
 			continue;
 		}
 
 		if (!last_write_torn && !i->csum_good) {
 			last_write_torn = true;
-			i->ignore = true;
+			i->ignore_blacklisted = true;
 			continue;
 		}
 
@@ -1280,12 +1293,12 @@ int bch2_journal_read(struct bch_fs *c,
 	genradix_for_each(&c->journal_entries, radix_iter, _i) {
 		i = *_i;
 
-		if (!i || i->ignore)
+		if (journal_replay_ignore(i))
 			continue;
 
 		seq = le64_to_cpu(i->j.seq);
 		if (seq < *last_seq) {
-			journal_replay_free(c, i);
+			journal_replay_free(c, i, false);
 			continue;
 		}
 
@@ -1293,7 +1306,7 @@ int bch2_journal_read(struct bch_fs *c,
 			fsck_err_on(!JSET_NO_FLUSH(&i->j), c,
 				    jset_seq_blacklisted,
 				    "found blacklisted journal entry %llu", seq);
-			i->ignore = true;
+			i->ignore_blacklisted = true;
 		}
 	}
 
@@ -1302,7 +1315,7 @@ int bch2_journal_read(struct bch_fs *c,
 	genradix_for_each(&c->journal_entries, radix_iter, _i) {
 		i = *_i;
 
-		if (!i || i->ignore)
+		if (journal_replay_ignore(i))
 			continue;
 
 		BUG_ON(seq > le64_to_cpu(i->j.seq));
@@ -1335,7 +1348,7 @@ int bch2_journal_read(struct bch_fs *c,
 			fsck_err(c, journal_entries_missing,
 				 "journal entries %llu-%llu missing! (replaying %llu-%llu)\n"
 				 "  prev at %s\n"
-				 "  next at %s",
+				 "  next at %s, continue?",
 				 missing_start, missing_end,
 				 *last_seq, *blacklist_seq - 1,
 				 buf1.buf, buf2.buf);
@@ -1353,32 +1366,31 @@ int bch2_journal_read(struct bch_fs *c,
 			.e.data_type = BCH_DATA_journal,
 			.e.nr_required = 1,
 		};
-		unsigned ptr;
 
 		i = *_i;
-		if (!i || i->ignore)
+		if (journal_replay_ignore(i))
 			continue;
 
-		for (ptr = 0; ptr < i->nr_ptrs; ptr++) {
-			struct bch_dev *ca = bch_dev_bkey_exists(c, i->ptrs[ptr].dev);
+		darray_for_each(i->ptrs, ptr) {
+			struct bch_dev *ca = bch2_dev_have_ref(c, ptr->dev);
 
-			if (!i->ptrs[ptr].csum_good)
-				bch_err_dev_offset(ca, i->ptrs[ptr].sector,
+			if (!ptr->csum_good)
+				bch_err_dev_offset(ca, ptr->sector,
 						   "invalid journal checksum, seq %llu%s",
 						   le64_to_cpu(i->j.seq),
 						   i->csum_good ? " (had good copy on another device)" : "");
 		}
 
 		ret = jset_validate(c,
-				    bch_dev_bkey_exists(c, i->ptrs[0].dev),
+				    bch2_dev_have_ref(c, i->ptrs.data[0].dev),
 				    &i->j,
-				    i->ptrs[0].sector,
+				    i->ptrs.data[0].sector,
 				    READ);
 		if (ret)
 			goto err;
 
-		for (ptr = 0; ptr < i->nr_ptrs; ptr++)
-			replicas.e.devs[replicas.e.nr_devs++] = i->ptrs[ptr].dev;
+		darray_for_each(i->ptrs, ptr)
+			replicas.e.devs[replicas.e.nr_devs++] = ptr->dev;
 
 		bch2_replicas_entry_sort(&replicas.e);
 
@@ -1547,7 +1559,7 @@ static void journal_buf_realloc(struct journal *j, struct journal_buf *buf)
 	if (bch2_btree_write_buffer_resize(c, btree_write_buffer_size))
 		return;
 
-	new_buf = kvpmalloc(new_size, GFP_NOFS|__GFP_NOWARN);
+	new_buf = kvmalloc(new_size, GFP_NOFS|__GFP_NOWARN);
 	if (!new_buf)
 		return;
 
@@ -1558,7 +1570,7 @@ static void journal_buf_realloc(struct journal *j, struct journal_buf *buf)
 	swap(buf->buf_size,	new_size);
 	spin_unlock(&j->lock);
 
-	kvpfree(new_buf, new_size);
+	kvfree(new_buf);
 }
 
 static inline struct journal_buf *journal_last_unwritten_buf(struct journal *j)
@@ -1568,12 +1580,12 @@ static inline struct journal_buf *journal_last_unwritten_buf(struct journal *j)
 
 static CLOSURE_CALLBACK(journal_write_done)
 {
-	closure_type(j, struct journal, io);
+	closure_type(w, struct journal_buf, io);
+	struct journal *j = container_of(w, struct journal, buf[w->idx]);
 	struct bch_fs *c = container_of(j, struct bch_fs, journal);
-	struct journal_buf *w = journal_last_unwritten_buf(j);
 	struct bch_replicas_padded replicas;
 	union journal_res_state old, new;
-	u64 v, seq;
+	u64 v, seq = le64_to_cpu(w->data->seq);
 	int err = 0;
 
 	bch2_time_stats_update(!JSET_NO_FLUSH(w->data)
@@ -1593,63 +1605,68 @@ static CLOSURE_CALLBACK(journal_write_done)
 	if (err)
 		bch2_fatal_error(c);
 
-	spin_lock(&j->lock);
-	seq = le64_to_cpu(w->data->seq);
+	closure_debug_destroy(cl);
 
+	spin_lock(&j->lock);
 	if (seq >= j->pin.front)
 		journal_seq_pin(j, seq)->devs = w->devs_written;
+	if (err && (!j->err_seq || seq < j->err_seq))
+		j->err_seq	= seq;
+	w->write_done = true;
+
+	bool completed = false;
+
+	for (seq = journal_last_unwritten_seq(j);
+	     seq <= journal_cur_seq(j);
+	     seq++) {
+		w = j->buf + (seq & JOURNAL_BUF_MASK);
+		if (!w->write_done)
+			break;
 
-	if (!err) {
-		if (!JSET_NO_FLUSH(w->data)) {
+		if (!j->err_seq && !JSET_NO_FLUSH(w->data)) {
 			j->flushed_seq_ondisk = seq;
 			j->last_seq_ondisk = w->last_seq;
 
 			bch2_do_discards(c);
 			closure_wake_up(&c->freelist_wait);
-
 			bch2_reset_alloc_cursors(c);
 		}
-	} else if (!j->err_seq || seq < j->err_seq)
-		j->err_seq	= seq;
 
-	j->seq_ondisk		= seq;
+		j->seq_ondisk = seq;
 
-	/*
-	 * Updating last_seq_ondisk may let bch2_journal_reclaim_work() discard
-	 * more buckets:
-	 *
-	 * Must come before signaling write completion, for
-	 * bch2_fs_journal_stop():
-	 */
-	if (j->watermark != BCH_WATERMARK_stripe)
-		journal_reclaim_kick(&c->journal);
+		/*
+		 * Updating last_seq_ondisk may let bch2_journal_reclaim_work() discard
+		 * more buckets:
+		 *
+		 * Must come before signaling write completion, for
+		 * bch2_fs_journal_stop():
+		 */
+		if (j->watermark != BCH_WATERMARK_stripe)
+			journal_reclaim_kick(&c->journal);
 
-	/* also must come before signalling write completion: */
-	closure_debug_destroy(cl);
+		v = atomic64_read(&j->reservations.counter);
+		do {
+			old.v = new.v = v;
+			BUG_ON(journal_state_count(new, new.unwritten_idx));
+			BUG_ON(new.unwritten_idx != (seq & JOURNAL_BUF_MASK));
 
-	v = atomic64_read(&j->reservations.counter);
-	do {
-		old.v = new.v = v;
-		BUG_ON(journal_state_count(new, new.unwritten_idx));
+			new.unwritten_idx++;
+		} while ((v = atomic64_cmpxchg(&j->reservations.counter, old.v, new.v)) != old.v);
 
-		new.unwritten_idx++;
-	} while ((v = atomic64_cmpxchg(&j->reservations.counter,
-				       old.v, new.v)) != old.v);
+		closure_wake_up(&w->wait);
+		completed = true;
+	}
 
-	bch2_journal_reclaim_fast(j);
-	bch2_journal_space_available(j);
+	if (completed) {
+		bch2_journal_reclaim_fast(j);
+		bch2_journal_space_available(j);
 
-	track_event_change(&c->times[BCH_TIME_blocked_journal_max_in_flight],
-			   &j->max_in_flight_start, false);
+		track_event_change(&c->times[BCH_TIME_blocked_journal_max_in_flight], false);
 
-	closure_wake_up(&w->wait);
-	journal_wake(j);
+		journal_wake(j);
+	}
 
-	if (!journal_state_count(new, new.unwritten_idx) &&
-	    journal_last_unwritten_seq(j) <= journal_cur_seq(j)) {
-		spin_unlock(&j->lock);
-		closure_call(&j->io, bch2_journal_write, c->io_complete_wq, NULL);
-	} else if (journal_last_unwritten_seq(j) == journal_cur_seq(j) &&
+	if (journal_last_unwritten_seq(j) == journal_cur_seq(j) &&
 		   new.cur_entry_offset < JOURNAL_ENTRY_CLOSED_VAL) {
 		struct journal_buf *buf = journal_cur_buf(j);
 		long delta = buf->expires - jiffies;
@@ -1659,47 +1676,52 @@ static CLOSURE_CALLBACK(journal_write_done)
 		 * previous entries still in flight - the current journal entry
 		 * might want to be written now:
 		 */
-
-		spin_unlock(&j->lock);
-		mod_delayed_work(c->io_complete_wq, &j->write_work, max(0L, delta));
-	} else {
-		spin_unlock(&j->lock);
+		mod_delayed_work(j->wq, &j->write_work, max(0L, delta));
 	}
+
+	/*
+	 * We don't typically trigger journal writes from her - the next journal
+	 * write will be triggered immediately after the previous one is
+	 * allocated, in bch2_journal_write() - but the journal write error path
+	 * is special:
+	 */
+	bch2_journal_do_writes(j);
+	spin_unlock(&j->lock);
 }
 
 static void journal_write_endio(struct bio *bio)
 {
-	struct bch_dev *ca = bio->bi_private;
+	struct journal_bio *jbio = container_of(bio, struct journal_bio, bio);
+	struct bch_dev *ca = jbio->ca;
 	struct journal *j = &ca->fs->journal;
-	struct journal_buf *w = journal_last_unwritten_buf(j);
-	unsigned long flags;
+	struct journal_buf *w = j->buf + jbio->buf_idx;
 
 	if (bch2_dev_io_err_on(bio->bi_status, ca, BCH_MEMBER_ERROR_write,
 			       "error writing journal entry %llu: %s",
 			       le64_to_cpu(w->data->seq),
 			       bch2_blk_status_to_str(bio->bi_status)) ||
 	    bch2_meta_write_fault("journal")) {
+		unsigned long flags;
+
 		spin_lock_irqsave(&j->err_lock, flags);
 		bch2_dev_list_drop_dev(&w->devs_written, ca->dev_idx);
 		spin_unlock_irqrestore(&j->err_lock, flags);
 	}
 
-	closure_put(&j->io);
+	closure_put(&w->io);
 	percpu_ref_put(&ca->io_ref);
 }
 
-static CLOSURE_CALLBACK(do_journal_write)
+static CLOSURE_CALLBACK(journal_write_submit)
 {
-	closure_type(j, struct journal, io);
+	closure_type(w, struct journal_buf, io);
+	struct journal *j = container_of(w, struct journal, buf[w->idx]);
 	struct bch_fs *c = container_of(j, struct bch_fs, journal);
-	struct bch_dev *ca;
-	struct journal_buf *w = journal_last_unwritten_buf(j);
-	struct bio *bio;
 	unsigned sectors = vstruct_sectors(w->data, c->block_bits);
 
 	extent_for_each_ptr(bkey_i_to_s_extent(&w->key), ptr) {
-		ca = bch_dev_bkey_exists(c, ptr->dev);
-		if (!percpu_ref_tryget(&ca->io_ref)) {
+		struct bch_dev *ca = bch2_dev_get_ioref(c, ptr->dev, WRITE);
+		if (!ca) {
 			/* XXX: fix this */
 			bch_err(c, "missing device for journal write\n");
 			continue;
@@ -1708,7 +1730,8 @@ static CLOSURE_CALLBACK(do_journal_write)
 		this_cpu_add(ca->io_done->sectors[WRITE][BCH_DATA_journal],
 			     sectors);
 
-		bio = ca->journal.bio;
+		struct journal_device *ja = &ca->journal;
+		struct bio *bio = &ja->bio[w->idx]->bio;
 		bio_reset(bio, ca->disk_sb.bdev, REQ_OP_WRITE|REQ_SYNC|REQ_META);
 		bio->bi_iter.bi_sector	= ptr->offset;
 		bio->bi_end_io		= journal_write_endio;
@@ -1727,11 +1750,50 @@ static CLOSURE_CALLBACK(do_journal_write)
 		trace_and_count(c, journal_write, bio);
 		closure_bio_submit(bio, cl);
 
-		ca->journal.bucket_seq[ca->journal.cur_idx] =
-			le64_to_cpu(w->data->seq);
+		ja->bucket_seq[ja->cur_idx] = le64_to_cpu(w->data->seq);
+	}
+
+	continue_at(cl, journal_write_done, j->wq);
+}
+
+static CLOSURE_CALLBACK(journal_write_preflush)
+{
+	closure_type(w, struct journal_buf, io);
+	struct journal *j = container_of(w, struct journal, buf[w->idx]);
+	struct bch_fs *c = container_of(j, struct bch_fs, journal);
+
+	if (j->seq_ondisk + 1 != le64_to_cpu(w->data->seq)) {
+		spin_lock(&j->lock);
+		if (j->seq_ondisk + 1 != le64_to_cpu(w->data->seq)) {
+			closure_wait(&j->async_wait, cl);
+			spin_unlock(&j->lock);
+			continue_at(cl, journal_write_preflush, j->wq);
+			return;
+		}
+		spin_unlock(&j->lock);
 	}
 
-	continue_at(cl, journal_write_done, c->io_complete_wq);
+	if (w->separate_flush) {
+		for_each_rw_member(c, ca) {
+			percpu_ref_get(&ca->io_ref);
+
+			struct journal_device *ja = &ca->journal;
+			struct bio *bio = &ja->bio[w->idx]->bio;
+			bio_reset(bio, ca->disk_sb.bdev,
+				  REQ_OP_WRITE|REQ_SYNC|REQ_META|REQ_PREFLUSH);
+			bio->bi_end_io		= journal_write_endio;
+			bio->bi_private		= ca;
+			closure_bio_submit(bio, cl);
+		}
+
+		continue_at(cl, journal_write_submit, j->wq);
+	} else {
+		/*
+		 * no need to punt to another work item if we're not waiting on
+		 * preflushes
+		 */
+		journal_write_submit(&cl->work);
+	}
 }
 
 static int bch2_journal_write_prep(struct journal *j, struct journal_buf *w)
@@ -1782,11 +1844,11 @@ static int bch2_journal_write_prep(struct journal *j, struct journal_buf *w)
 			if (!wb.wb)
 				bch2_journal_keys_to_write_buffer_start(c, &wb, seq);
 
-			struct bkey_i *k;
 			jset_entry_for_each_key(i, k) {
 				ret = bch2_journal_key_to_wb(c, &wb, i->btree_id, k);
 				if (ret) {
-					bch2_fs_fatal_error(c, "-ENOMEM flushing journal keys to btree write buffer");
+					bch2_fs_fatal_error(c, "flushing journal keys to btree write buffer: %s",
+							    bch2_err_str(ret));
 					bch2_journal_keys_to_write_buffer_end(c, &wb);
 					return ret;
 				}
@@ -1798,15 +1860,24 @@ static int bch2_journal_write_prep(struct journal *j, struct journal_buf *w)
 
 	if (wb.wb)
 		bch2_journal_keys_to_write_buffer_end(c, &wb);
+
+	spin_lock(&c->journal.lock);
 	w->need_flush_to_write_buffer = false;
+	spin_unlock(&c->journal.lock);
 
 	start = end = vstruct_last(jset);
 
 	end	= bch2_btree_roots_to_journal_entries(c, end, btree_roots_have);
 
+	struct jset_entry_datetime *d =
+		container_of(jset_entry_init(&end, sizeof(*d)), struct jset_entry_datetime, entry);
+	d->entry.type	= BCH_JSET_ENTRY_datetime;
+	d->seconds	= cpu_to_le64(ktime_get_real_seconds());
+
 	bch2_journal_super_entries_add_common(c, &end, seq);
 	u64s	= (u64 *) end - (u64 *) start;
-	BUG_ON(u64s > j->entry_u64s_reserved);
+
+	WARN_ON(u64s > j->entry_u64s_reserved);
 
 	le32_add_cpu(&jset->u64s, u64s);
 
@@ -1814,7 +1885,7 @@ static int bch2_journal_write_prep(struct journal *j, struct journal_buf *w)
 	bytes	= vstruct_bytes(jset);
 
 	if (sectors > w->sectors) {
-		bch2_fs_fatal_error(c, "aieeee! journal write overran available space, %zu > %u (extra %u reserved %u/%u)",
+		bch2_fs_fatal_error(c, ": journal write overran available space, %zu > %u (extra %u reserved %u/%u)",
 				    vstruct_bytes(jset), w->sectors << 9,
 				    u64s, w->u64s_reserved, j->entry_u64s_reserved);
 		return -EINVAL;
@@ -1842,8 +1913,7 @@ static int bch2_journal_write_prep(struct journal *j, struct journal_buf *w)
 	ret = bch2_encrypt(c, JSET_CSUM_TYPE(jset), journal_nonce(jset),
 		    jset->encrypted_start,
 		    vstruct_end(jset) - (void *) jset->encrypted_start);
-	if (bch2_fs_fatal_err_on(ret, c,
-			"error decrypting journal entry: %i", ret))
+	if (bch2_fs_fatal_err_on(ret, c, "decrypting journal entry: %s", bch2_err_str(ret)))
 		return ret;
 
 	jset->csum = csum_vstruct(c, JSET_CSUM_TYPE(jset),
@@ -1878,14 +1948,14 @@ static int bch2_journal_write_pick_flush(struct journal *j, struct journal_buf *
 	 * So if we're in an error state, and we're still starting up, we don't
 	 * write anything at all.
 	 */
-	if (error && test_bit(JOURNAL_NEED_FLUSH_WRITE, &j->flags))
+	if (error && test_bit(JOURNAL_need_flush_write, &j->flags))
 		return -EIO;
 
 	if (error ||
 	    w->noflush ||
 	    (!w->must_flush &&
 	     (jiffies - j->last_flush_write) < msecs_to_jiffies(c->opts.journal_flush_delay) &&
-	     test_bit(JOURNAL_MAY_SKIP_FLUSH, &j->flags))) {
+	     test_bit(JOURNAL_may_skip_flush, &j->flags))) {
 		w->noflush = true;
 		SET_JSET_NO_FLUSH(w->data, true);
 		w->data->last_seq	= 0;
@@ -1893,9 +1963,10 @@ static int bch2_journal_write_pick_flush(struct journal *j, struct journal_buf *
 
 		j->nr_noflush_writes++;
 	} else {
+		w->must_flush = true;
 		j->last_flush_write = jiffies;
 		j->nr_flush_writes++;
-		clear_bit(JOURNAL_NEED_FLUSH_WRITE, &j->flags);
+		clear_bit(JOURNAL_need_flush_write, &j->flags);
 	}
 
 	return 0;
@@ -1903,20 +1974,27 @@ static int bch2_journal_write_pick_flush(struct journal *j, struct journal_buf *
 
 CLOSURE_CALLBACK(bch2_journal_write)
 {
-	closure_type(j, struct journal, io);
+	closure_type(w, struct journal_buf, io);
+	struct journal *j = container_of(w, struct journal, buf[w->idx]);
 	struct bch_fs *c = container_of(j, struct bch_fs, journal);
-	struct journal_buf *w = journal_last_unwritten_buf(j);
 	struct bch_replicas_padded replicas;
-	struct bio *bio;
-	struct printbuf journal_debug_buf = PRINTBUF;
 	unsigned nr_rw_members = 0;
 	int ret;
 
+	for_each_rw_member(c, ca)
+		nr_rw_members++;
+
 	BUG_ON(BCH_SB_CLEAN(c->disk_sb.sb));
+	BUG_ON(!w->write_started);
+	BUG_ON(w->write_allocated);
+	BUG_ON(w->write_done);
 
 	j->write_start_time = local_clock();
 
 	spin_lock(&j->lock);
+	if (nr_rw_members > 1)
+		w->separate_flush = true;
+
 	ret = bch2_journal_write_pick_flush(j, w);
 	spin_unlock(&j->lock);
 	if (ret)
@@ -1943,11 +2021,15 @@ CLOSURE_CALLBACK(bch2_journal_write)
 	}
 
 	if (ret) {
-		__bch2_journal_debug_to_text(&journal_debug_buf, j);
+		struct printbuf buf = PRINTBUF;
+		buf.atomic++;
+
+		prt_printf(&buf, bch2_fmt(c, "Unable to allocate journal write: %s"),
+			   bch2_err_str(ret));
+		__bch2_journal_debug_to_text(&buf, j);
 		spin_unlock(&j->lock);
-		bch_err(c, "Unable to allocate journal write:\n%s",
-			journal_debug_buf.buf);
-		printbuf_exit(&journal_debug_buf);
+		bch2_print_string_as_lines(KERN_ERR, buf.buf);
+		printbuf_exit(&buf);
 		goto err;
 	}
 
@@ -1956,12 +2038,14 @@ CLOSURE_CALLBACK(bch2_journal_write)
 	 * bch2_journal_space_available():
 	 */
 	w->sectors = 0;
+	w->write_allocated = true;
 
 	/*
 	 * journal entry has been compacted and allocated, recalculate space
 	 * available:
 	 */
 	bch2_journal_space_available(j);
+	bch2_journal_do_writes(j);
 	spin_unlock(&j->lock);
 
 	w->devs_written = bch2_bkey_devs(bkey_i_to_s_c(&w->key));
@@ -1969,12 +2053,6 @@ CLOSURE_CALLBACK(bch2_journal_write)
 	if (c->opts.nochanges)
 		goto no_io;
 
-	for_each_rw_member(c, ca)
-		nr_rw_members++;
-
-	if (nr_rw_members > 1)
-		w->separate_flush = true;
-
 	/*
 	 * Mark journal replicas before we submit the write to guarantee
 	 * recovery will find the journal entries after a crash.
@@ -1985,25 +2063,15 @@ CLOSURE_CALLBACK(bch2_journal_write)
 	if (ret)
 		goto err;
 
-	if (!JSET_NO_FLUSH(w->data) && w->separate_flush) {
-		for_each_rw_member(c, ca) {
-			percpu_ref_get(&ca->io_ref);
-
-			bio = ca->journal.bio;
-			bio_reset(bio, ca->disk_sb.bdev,
-				  REQ_OP_WRITE|REQ_PREFLUSH);
-			bio->bi_end_io		= journal_write_endio;
-			bio->bi_private		= ca;
-			closure_bio_submit(bio, cl);
-		}
-	}
-
-	continue_at(cl, do_journal_write, c->io_complete_wq);
+	if (!JSET_NO_FLUSH(w->data))
+		continue_at(cl, journal_write_preflush, j->wq);
+	else
+		continue_at(cl, journal_write_submit, j->wq);
 	return;
 no_io:
-	continue_at(cl, journal_write_done, c->io_complete_wq);
+	continue_at(cl, journal_write_done, j->wq);
 	return;
 err:
 	bch2_fatal_error(c);
-	continue_at(cl, journal_write_done, c->io_complete_wq);
+	continue_at(cl, journal_write_done, j->wq);
 }
diff --git a/fs/bcachefs/journal_io.h b/fs/bcachefs/journal_io.h
index c035e7c108..2ca9cde30e 100644
--- a/fs/bcachefs/journal_io.h
+++ b/fs/bcachefs/journal_io.h
@@ -2,26 +2,38 @@
 #ifndef _BCACHEFS_JOURNAL_IO_H
 #define _BCACHEFS_JOURNAL_IO_H
 
+#include "darray.h"
+
+void bch2_journal_pos_from_member_info_set(struct bch_fs *);
+void bch2_journal_pos_from_member_info_resume(struct bch_fs *);
+
+struct journal_ptr {
+	bool		csum_good;
+	u8		dev;
+	u32		bucket;
+	u32		bucket_offset;
+	u64		sector;
+};
+
 /*
  * Only used for holding the journal entries we read in btree_journal_read()
  * during cache_registration
  */
 struct journal_replay {
-	struct journal_ptr {
-		bool		csum_good;
-		u8		dev;
-		u32		bucket;
-		u32		bucket_offset;
-		u64		sector;
-	}			ptrs[BCH_REPLICAS_MAX];
-	unsigned		nr_ptrs;
+	DARRAY_PREALLOCATED(struct journal_ptr, 8) ptrs;
 
 	bool			csum_good;
-	bool			ignore;
+	bool			ignore_blacklisted;
+	bool			ignore_not_dirty;
 	/* must be last: */
 	struct jset		j;
 };
 
+static inline bool journal_replay_ignore(struct journal_replay *i)
+{
+	return !i || i->ignore_blacklisted || i->ignore_not_dirty;
+}
+
 static inline struct jset_entry *__jset_entry_type_next(struct jset *jset,
 					struct jset_entry *entry, unsigned type)
 {
@@ -36,12 +48,12 @@ static inline struct jset_entry *__jset_entry_type_next(struct jset *jset,
 }
 
 #define for_each_jset_entry_type(entry, jset, type)			\
-	for (entry = (jset)->start;					\
+	for (struct jset_entry *entry = (jset)->start;			\
 	     (entry = __jset_entry_type_next(jset, entry, type));	\
 	     entry = vstruct_next(entry))
 
 #define jset_entry_for_each_key(_e, _k)					\
-	for (_k = (_e)->start;						\
+	for (struct bkey_i *_k = (_e)->start;				\
 	     _k < vstruct_last(_e);					\
 	     _k = bkey_next(_k))
 
@@ -51,7 +63,7 @@ static inline struct jset_entry *__jset_entry_type_next(struct jset *jset,
 
 int bch2_journal_entry_validate(struct bch_fs *, struct jset *,
 				struct jset_entry *, unsigned, int,
-				enum bkey_invalid_flags);
+				enum bch_validate_flags);
 void bch2_journal_entry_to_text(struct printbuf *, struct bch_fs *,
 				struct jset_entry *);
 
@@ -62,4 +74,20 @@ int bch2_journal_read(struct bch_fs *, u64 *, u64 *, u64 *);
 
 CLOSURE_CALLBACK(bch2_journal_write);
 
+static inline struct jset_entry *jset_entry_init(struct jset_entry **end, size_t size)
+{
+	struct jset_entry *entry = *end;
+	unsigned u64s = DIV_ROUND_UP(size, sizeof(u64));
+
+	memset(entry, 0, u64s * sizeof(u64));
+	/*
+	 * The u64s field counts from the start of data, ignoring the shared
+	 * fields.
+	 */
+	entry->u64s = cpu_to_le16(u64s - 1);
+
+	*end = vstruct_next(*end);
+	return entry;
+}
+
 #endif /* _BCACHEFS_JOURNAL_IO_H */
diff --git a/fs/bcachefs/journal_reclaim.c b/fs/bcachefs/journal_reclaim.c
index c33dca6415..79be0eaddf 100644
--- a/fs/bcachefs/journal_reclaim.c
+++ b/fs/bcachefs/journal_reclaim.c
@@ -62,14 +62,13 @@ void bch2_journal_set_watermark(struct journal *j)
 		? BCH_WATERMARK_reclaim
 		: BCH_WATERMARK_stripe;
 
-	if (track_event_change(&c->times[BCH_TIME_blocked_journal_low_on_space],
-			       &j->low_on_space_start, low_on_space) ||
-	    track_event_change(&c->times[BCH_TIME_blocked_journal_low_on_pin],
-			       &j->low_on_pin_start, low_on_pin) ||
-	    track_event_change(&c->times[BCH_TIME_blocked_write_buffer_full],
-			       &j->write_buffer_full_start, low_on_wb))
+	if (track_event_change(&c->times[BCH_TIME_blocked_journal_low_on_space], low_on_space) ||
+	    track_event_change(&c->times[BCH_TIME_blocked_journal_low_on_pin], low_on_pin) ||
+	    track_event_change(&c->times[BCH_TIME_blocked_write_buffer_full], low_on_wb))
 		trace_and_count(c, journal_full, c);
 
+	mod_bit(JOURNAL_space_low, &j->flags, low_on_space || low_on_pin);
+
 	swap(watermark, j->watermark);
 	if (watermark > j->watermark)
 		journal_wake(j);
@@ -226,9 +225,9 @@ void bch2_journal_space_available(struct journal *j)
 	     j->space[journal_space_clean_ondisk].total) &&
 	    (clean - clean_ondisk <= total / 8) &&
 	    (clean_ondisk * 2 > clean))
-		set_bit(JOURNAL_MAY_SKIP_FLUSH, &j->flags);
+		set_bit(JOURNAL_may_skip_flush, &j->flags);
 	else
-		clear_bit(JOURNAL_MAY_SKIP_FLUSH, &j->flags);
+		clear_bit(JOURNAL_may_skip_flush, &j->flags);
 
 	bch2_journal_set_watermark(j);
 out:
@@ -394,8 +393,6 @@ void bch2_journal_pin_copy(struct journal *j,
 			   struct journal_entry_pin *src,
 			   journal_pin_flush_fn flush_fn)
 {
-	bool reclaim;
-
 	spin_lock(&j->lock);
 
 	u64 seq = READ_ONCE(src->seq);
@@ -411,44 +408,44 @@ void bch2_journal_pin_copy(struct journal *j,
 		return;
 	}
 
-	reclaim = __journal_pin_drop(j, dst);
+	bool reclaim = __journal_pin_drop(j, dst);
 
 	bch2_journal_pin_set_locked(j, seq, dst, flush_fn, journal_pin_type(flush_fn));
 
 	if (reclaim)
 		bch2_journal_reclaim_fast(j);
-	spin_unlock(&j->lock);
 
 	/*
 	 * If the journal is currently full,  we might want to call flush_fn
 	 * immediately:
 	 */
-	journal_wake(j);
+	if (seq == journal_last_seq(j))
+		journal_wake(j);
+	spin_unlock(&j->lock);
 }
 
 void bch2_journal_pin_set(struct journal *j, u64 seq,
 			  struct journal_entry_pin *pin,
 			  journal_pin_flush_fn flush_fn)
 {
-	bool reclaim;
-
 	spin_lock(&j->lock);
 
 	BUG_ON(seq < journal_last_seq(j));
 
-	reclaim = __journal_pin_drop(j, pin);
+	bool reclaim = __journal_pin_drop(j, pin);
 
 	bch2_journal_pin_set_locked(j, seq, pin, flush_fn, journal_pin_type(flush_fn));
 
 	if (reclaim)
 		bch2_journal_reclaim_fast(j);
-	spin_unlock(&j->lock);
-
 	/*
 	 * If the journal is currently full,  we might want to call flush_fn
 	 * immediately:
 	 */
-	journal_wake(j);
+	if (seq == journal_last_seq(j))
+		journal_wake(j);
+
+	spin_unlock(&j->lock);
 }
 
 /**
@@ -821,7 +818,7 @@ static int journal_flush_done(struct journal *j, u64 seq_to_flush,
 	 * If journal replay hasn't completed, the unreplayed journal entries
 	 * hold refs on their corresponding sequence numbers
 	 */
-	ret = !test_bit(JOURNAL_REPLAY_DONE, &j->flags) ||
+	ret = !test_bit(JOURNAL_replay_done, &j->flags) ||
 		journal_last_seq(j) > seq_to_flush ||
 		!fifo_used(&j->pin);
 
@@ -836,7 +833,7 @@ bool bch2_journal_flush_pins(struct journal *j, u64 seq_to_flush)
 	/* time_stats this */
 	bool did_work = false;
 
-	if (!test_bit(JOURNAL_STARTED, &j->flags))
+	if (!test_bit(JOURNAL_running, &j->flags))
 		return false;
 
 	closure_wait_event(&j->async_wait,
diff --git a/fs/bcachefs/journal_sb.c b/fs/bcachefs/journal_sb.c
index ae4fb8c3a2..db80e506e3 100644
--- a/fs/bcachefs/journal_sb.c
+++ b/fs/bcachefs/journal_sb.c
@@ -16,9 +16,8 @@ static int u64_cmp(const void *_l, const void *_r)
 	return cmp_int(*l, *r);
 }
 
-static int bch2_sb_journal_validate(struct bch_sb *sb,
-				    struct bch_sb_field *f,
-				    struct printbuf *err)
+static int bch2_sb_journal_validate(struct bch_sb *sb, struct bch_sb_field *f,
+				enum bch_validate_flags flags, struct printbuf *err)
 {
 	struct bch_sb_field_journal *journal = field_to_type(f, journal);
 	struct bch_member m = bch2_sb_member_get(sb, sb->dev_idx);
@@ -99,9 +98,8 @@ static int u64_range_cmp(const void *_l, const void *_r)
 	return cmp_int(l->start, r->start);
 }
 
-static int bch2_sb_journal_v2_validate(struct bch_sb *sb,
-				    struct bch_sb_field *f,
-				    struct printbuf *err)
+static int bch2_sb_journal_v2_validate(struct bch_sb *sb, struct bch_sb_field *f,
+				enum bch_validate_flags flags, struct printbuf *err)
 {
 	struct bch_sb_field_journal_v2 *journal = field_to_type(f, journal_v2);
 	struct bch_member m = bch2_sb_member_get(sb, sb->dev_idx);
diff --git a/fs/bcachefs/journal_seq_blacklist.c b/fs/bcachefs/journal_seq_blacklist.c
index 0200e299cf..1f25c111c5 100644
--- a/fs/bcachefs/journal_seq_blacklist.c
+++ b/fs/bcachefs/journal_seq_blacklist.c
@@ -1,8 +1,8 @@
 // SPDX-License-Identifier: GPL-2.0
 
 #include "bcachefs.h"
-#include "btree_iter.h"
 #include "eytzinger.h"
+#include "journal.h"
 #include "journal_seq_blacklist.h"
 #include "super-io.h"
 
@@ -43,61 +43,36 @@ static unsigned sb_blacklist_u64s(unsigned nr)
 	return (sizeof(*bl) + sizeof(bl->start[0]) * nr) / sizeof(u64);
 }
 
-static struct bch_sb_field_journal_seq_blacklist *
-blacklist_entry_try_merge(struct bch_fs *c,
-			  struct bch_sb_field_journal_seq_blacklist *bl,
-			  unsigned i)
-{
-	unsigned nr = blacklist_nr_entries(bl);
-
-	if (le64_to_cpu(bl->start[i].end) >=
-	    le64_to_cpu(bl->start[i + 1].start)) {
-		bl->start[i].end = bl->start[i + 1].end;
-		--nr;
-		memmove(&bl->start[i],
-			&bl->start[i + 1],
-			sizeof(bl->start[0]) * (nr - i));
-
-		bl = bch2_sb_field_resize(&c->disk_sb, journal_seq_blacklist,
-					  sb_blacklist_u64s(nr));
-		BUG_ON(!bl);
-	}
-
-	return bl;
-}
-
-static bool bl_entry_contig_or_overlaps(struct journal_seq_blacklist_entry *e,
-					u64 start, u64 end)
-{
-	return !(end < le64_to_cpu(e->start) || le64_to_cpu(e->end) < start);
-}
-
 int bch2_journal_seq_blacklist_add(struct bch_fs *c, u64 start, u64 end)
 {
 	struct bch_sb_field_journal_seq_blacklist *bl;
-	unsigned i, nr;
+	unsigned i = 0, nr;
 	int ret = 0;
 
 	mutex_lock(&c->sb_lock);
 	bl = bch2_sb_field_get(c->disk_sb.sb, journal_seq_blacklist);
 	nr = blacklist_nr_entries(bl);
 
-	for (i = 0; i < nr; i++) {
+	while (i < nr) {
 		struct journal_seq_blacklist_entry *e =
 			bl->start + i;
 
-		if (bl_entry_contig_or_overlaps(e, start, end)) {
-			e->start = cpu_to_le64(min(start, le64_to_cpu(e->start)));
-			e->end	= cpu_to_le64(max(end, le64_to_cpu(e->end)));
-
-			if (i + 1 < nr)
-				bl = blacklist_entry_try_merge(c,
-							bl, i);
-			if (i)
-				bl = blacklist_entry_try_merge(c,
-							bl, i - 1);
-			goto out_write_sb;
+		if (end < le64_to_cpu(e->start))
+			break;
+
+		if (start > le64_to_cpu(e->end)) {
+			i++;
+			continue;
 		}
+
+		/*
+		 * Entry is contiguous or overlapping with new entry: merge it
+		 * with new entry, and delete:
+		 */
+
+		start	= min(start,	le64_to_cpu(e->start));
+		end	= max(end,	le64_to_cpu(e->end));
+		array_remove_item(bl->start, nr, i);
 	}
 
 	bl = bch2_sb_field_resize(&c->disk_sb, journal_seq_blacklist,
@@ -107,9 +82,10 @@ int bch2_journal_seq_blacklist_add(struct bch_fs *c, u64 start, u64 end)
 		goto out;
 	}
 
-	bl->start[nr].start	= cpu_to_le64(start);
-	bl->start[nr].end	= cpu_to_le64(end);
-out_write_sb:
+	array_insert_item(bl->start, nr, i, ((struct journal_seq_blacklist_entry) {
+		.start	= cpu_to_le64(start),
+		.end	= cpu_to_le64(end),
+	}));
 	c->disk_sb.sb->features[0] |= cpu_to_le64(1ULL << BCH_FEATURE_journal_seq_blacklist_v3);
 
 	ret = bch2_write_super(c);
@@ -119,8 +95,7 @@ out:
 	return ret ?: bch2_blacklist_table_initialize(c);
 }
 
-static int journal_seq_blacklist_table_cmp(const void *_l,
-					   const void *_r, size_t size)
+static int journal_seq_blacklist_table_cmp(const void *_l, const void *_r)
 {
 	const struct journal_seq_blacklist_table_entry *l = _l;
 	const struct journal_seq_blacklist_table_entry *r = _r;
@@ -165,8 +140,7 @@ int bch2_blacklist_table_initialize(struct bch_fs *c)
 	if (!bl)
 		return 0;
 
-	t = kzalloc(sizeof(*t) + sizeof(t->entries[0]) * nr,
-		    GFP_KERNEL);
+	t = kzalloc(struct_size(t, entries, nr), GFP_KERNEL);
 	if (!t)
 		return -BCH_ERR_ENOMEM_blacklist_table_init;
 
@@ -188,9 +162,8 @@ int bch2_blacklist_table_initialize(struct bch_fs *c)
 	return 0;
 }
 
-static int bch2_sb_journal_seq_blacklist_validate(struct bch_sb *sb,
-						  struct bch_sb_field *f,
-						  struct printbuf *err)
+static int bch2_sb_journal_seq_blacklist_validate(struct bch_sb *sb, struct bch_sb_field *f,
+				enum bch_validate_flags flags, struct printbuf *err)
 {
 	struct bch_sb_field_journal_seq_blacklist *bl =
 		field_to_type(f, journal_seq_blacklist);
@@ -243,78 +216,40 @@ const struct bch_sb_field_ops bch_sb_field_ops_journal_seq_blacklist = {
 	.to_text	= bch2_sb_journal_seq_blacklist_to_text
 };
 
-void bch2_blacklist_entries_gc(struct work_struct *work)
+bool bch2_blacklist_entries_gc(struct bch_fs *c)
 {
-	struct bch_fs *c = container_of(work, struct bch_fs,
-					journal_seq_blacklist_gc_work);
-	struct journal_seq_blacklist_table *t;
-	struct bch_sb_field_journal_seq_blacklist *bl;
 	struct journal_seq_blacklist_entry *src, *dst;
-	struct btree_trans *trans = bch2_trans_get(c);
-	unsigned i, nr, new_nr;
-	int ret;
-
-	for (i = 0; i < BTREE_ID_NR; i++) {
-		struct btree_iter iter;
-		struct btree *b;
-
-		bch2_trans_node_iter_init(trans, &iter, i, POS_MIN,
-					  0, 0, BTREE_ITER_PREFETCH);
-retry:
-		bch2_trans_begin(trans);
-
-		b = bch2_btree_iter_peek_node(&iter);
-
-		while (!(ret = PTR_ERR_OR_ZERO(b)) &&
-		       b &&
-		       !test_bit(BCH_FS_stopping, &c->flags))
-			b = bch2_btree_iter_next_node(&iter);
-
-		if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
-			goto retry;
 
-		bch2_trans_iter_exit(trans, &iter);
-	}
-
-	bch2_trans_put(trans);
-	if (ret)
-		return;
-
-	mutex_lock(&c->sb_lock);
-	bl = bch2_sb_field_get(c->disk_sb.sb, journal_seq_blacklist);
+	struct bch_sb_field_journal_seq_blacklist *bl =
+		bch2_sb_field_get(c->disk_sb.sb, journal_seq_blacklist);
 	if (!bl)
-		goto out;
+		return false;
 
-	nr = blacklist_nr_entries(bl);
+	unsigned nr = blacklist_nr_entries(bl);
 	dst = bl->start;
 
-	t = c->journal_seq_blacklist_table;
+	struct journal_seq_blacklist_table *t = c->journal_seq_blacklist_table;
 	BUG_ON(nr != t->nr);
 
-	for (src = bl->start, i = eytzinger0_first(t->nr);
+	unsigned i;
+	for (src = bl->start, i = t->nr == 0 ? 0 : eytzinger0_first(t->nr);
 	     src < bl->start + nr;
 	     src++, i = eytzinger0_next(i, nr)) {
 		BUG_ON(t->entries[i].start	!= le64_to_cpu(src->start));
 		BUG_ON(t->entries[i].end	!= le64_to_cpu(src->end));
 
-		if (t->entries[i].dirty)
+		if (t->entries[i].dirty || t->entries[i].end >= c->journal.oldest_seq_found_ondisk)
 			*dst++ = *src;
 	}
 
-	new_nr = dst - bl->start;
-
-	bch_info(c, "nr blacklist entries was %u, now %u", nr, new_nr);
-
-	if (new_nr != nr) {
-		bl = bch2_sb_field_resize(&c->disk_sb, journal_seq_blacklist,
-				new_nr ? sb_blacklist_u64s(new_nr) : 0);
-		BUG_ON(new_nr && !bl);
+	unsigned new_nr = dst - bl->start;
+	if (new_nr == nr)
+		return false;
 
-		if (!new_nr)
-			c->disk_sb.sb->features[0] &= cpu_to_le64(~(1ULL << BCH_FEATURE_journal_seq_blacklist_v3));
+	bch_verbose(c, "nr blacklist entries was %u, now %u", nr, new_nr);
 
-		bch2_write_super(c);
-	}
-out:
-	mutex_unlock(&c->sb_lock);
+	bl = bch2_sb_field_resize(&c->disk_sb, journal_seq_blacklist,
+				  new_nr ? sb_blacklist_u64s(new_nr) : 0);
+	BUG_ON(new_nr && !bl);
+	return true;
 }
diff --git a/fs/bcachefs/journal_seq_blacklist.h b/fs/bcachefs/journal_seq_blacklist.h
index afb886ec8e..d47636f96f 100644
--- a/fs/bcachefs/journal_seq_blacklist.h
+++ b/fs/bcachefs/journal_seq_blacklist.h
@@ -17,6 +17,6 @@ int bch2_blacklist_table_initialize(struct bch_fs *);
 
 extern const struct bch_sb_field_ops bch_sb_field_ops_journal_seq_blacklist;
 
-void bch2_blacklist_entries_gc(struct work_struct *);
+bool bch2_blacklist_entries_gc(struct bch_fs *);
 
 #endif /* _BCACHEFS_JOURNAL_SEQ_BLACKLIST_H */
diff --git a/fs/bcachefs/journal_seq_blacklist_format.h b/fs/bcachefs/journal_seq_blacklist_format.h
new file mode 100644
index 0000000000..2566b12dbc
--- /dev/null
+++ b/fs/bcachefs/journal_seq_blacklist_format.h
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _BCACHEFS_JOURNAL_SEQ_BLACKLIST_FORMAT_H
+#define _BCACHEFS_JOURNAL_SEQ_BLACKLIST_FORMAT_H
+
+struct journal_seq_blacklist_entry {
+	__le64			start;
+	__le64			end;
+};
+
+struct bch_sb_field_journal_seq_blacklist {
+	struct bch_sb_field	field;
+	struct journal_seq_blacklist_entry start[];
+};
+
+#endif /* _BCACHEFS_JOURNAL_SEQ_BLACKLIST_FORMAT_H */
diff --git a/fs/bcachefs/journal_types.h b/fs/bcachefs/journal_types.h
index 38817c7a08..19183fcf7a 100644
--- a/fs/bcachefs/journal_types.h
+++ b/fs/bcachefs/journal_types.h
@@ -18,6 +18,7 @@
  * the journal that are being staged or in flight.
  */
 struct journal_buf {
+	struct closure		io;
 	struct jset		*data;
 
 	__BKEY_PADDED(key, BCH_REPLICAS_MAX);
@@ -33,10 +34,14 @@ struct journal_buf {
 	unsigned		disk_sectors;	/* maximum size entry could have been, if
 						   buf_size was bigger */
 	unsigned		u64s_reserved;
-	bool			noflush;	/* write has already been kicked off, and was noflush */
-	bool			must_flush;	/* something wants a flush */
-	bool			separate_flush;
-	bool			need_flush_to_write_buffer;
+	bool			noflush:1;	/* write has already been kicked off, and was noflush */
+	bool			must_flush:1;	/* something wants a flush */
+	bool			separate_flush:1;
+	bool			need_flush_to_write_buffer:1;
+	bool			write_started:1;
+	bool			write_allocated:1;
+	bool			write_done:1;
+	u8			idx;
 };
 
 /*
@@ -124,16 +129,23 @@ enum journal_space_from {
 	journal_space_nr,
 };
 
+#define JOURNAL_FLAGS()			\
+	x(replay_done)			\
+	x(running)			\
+	x(may_skip_flush)		\
+	x(need_flush_write)		\
+	x(space_low)
+
 enum journal_flags {
-	JOURNAL_REPLAY_DONE,
-	JOURNAL_STARTED,
-	JOURNAL_MAY_SKIP_FLUSH,
-	JOURNAL_NEED_FLUSH_WRITE,
+#define x(n)	JOURNAL_##n,
+	JOURNAL_FLAGS()
+#undef x
 };
 
 /* Reasons we may fail to get a journal reservation: */
 #define JOURNAL_ERRORS()		\
 	x(ok)				\
+	x(retry)			\
 	x(blocked)			\
 	x(max_in_flight)		\
 	x(journal_full)			\
@@ -149,6 +161,13 @@ enum journal_errors {
 
 typedef DARRAY(u64)		darray_u64;
 
+struct journal_bio {
+	struct bch_dev		*ca;
+	unsigned		buf_idx;
+
+	struct bio		bio;
+};
+
 /* Embedded in struct bch_fs */
 struct journal {
 	/* Fastpath stuff up front: */
@@ -203,8 +222,8 @@ struct journal {
 	wait_queue_head_t	wait;
 	struct closure_waitlist	async_wait;
 
-	struct closure		io;
 	struct delayed_work	write_work;
+	struct workqueue_struct *wq;
 
 	/* Sequence number of most recent journal entry (last entry in @pin) */
 	atomic64_t		seq;
@@ -215,6 +234,7 @@ struct journal {
 	u64			last_seq_ondisk;
 	u64			err_seq;
 	u64			last_empty_seq;
+	u64			oldest_seq_found_ondisk;
 
 	/*
 	 * FIFO of journal entries whose btree updates have not yet been
@@ -274,11 +294,6 @@ struct journal {
 	u64			nr_noflush_writes;
 	u64			entry_bytes_written;
 
-	u64			low_on_space_start;
-	u64			low_on_pin_start;
-	u64			max_in_flight_start;
-	u64			write_buffer_full_start;
-
 	struct bch2_time_stats	*flush_write_time;
 	struct bch2_time_stats	*noflush_write_time;
 	struct bch2_time_stats	*flush_seq_time;
@@ -313,10 +328,11 @@ struct journal_device {
 	u64			*buckets;
 
 	/* Bio for journal reads/writes to this device */
-	struct bio		*bio;
+	struct journal_bio	*bio[JOURNAL_BUF_NR];
 
 	/* for bch_journal_read_device */
 	struct closure		read;
+	u64			highest_seq_found;
 };
 
 /*
diff --git a/fs/bcachefs/logged_ops.c b/fs/bcachefs/logged_ops.c
index ad598105c5..f49fdca1d0 100644
--- a/fs/bcachefs/logged_ops.c
+++ b/fs/bcachefs/logged_ops.c
@@ -37,7 +37,6 @@ static int resume_logged_op(struct btree_trans *trans, struct btree_iter *iter,
 	const struct bch_logged_op_fn *fn = logged_op_fn(k.k->type);
 	struct bkey_buf sk;
 	u32 restart_count = trans->restart_count;
-	int ret;
 
 	if (!fn)
 		return 0;
@@ -45,11 +44,11 @@ static int resume_logged_op(struct btree_trans *trans, struct btree_iter *iter,
 	bch2_bkey_buf_init(&sk);
 	bch2_bkey_buf_reassemble(&sk, c, k);
 
-	ret =   drop_locks_do(trans, (bch2_fs_lazy_rw(c), 0)) ?:
-		fn->resume(trans, sk.k) ?: trans_was_restarted(trans, restart_count);
+	fn->resume(trans, sk.k);
 
 	bch2_bkey_buf_exit(&sk, c);
-	return ret;
+
+	return trans_was_restarted(trans, restart_count);
 }
 
 int bch2_resume_logged_ops(struct bch_fs *c)
@@ -57,7 +56,7 @@ int bch2_resume_logged_ops(struct bch_fs *c)
 	int ret = bch2_trans_run(c,
 		for_each_btree_key(trans, iter,
 				   BTREE_ID_logged_ops, POS_MIN,
-				   BTREE_ITER_PREFETCH, k,
+				   BTREE_ITER_prefetch, k,
 			resume_logged_op(trans, &iter, k)));
 	bch_err_fn(c, ret);
 	return ret;
@@ -101,8 +100,8 @@ void bch2_logged_op_finish(struct btree_trans *trans, struct bkey_i *k)
 		struct printbuf buf = PRINTBUF;
 
 		bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(k));
-		bch2_fs_fatal_error(c, "%s: error deleting logged operation %s: %s",
-				     __func__, buf.buf, bch2_err_str(ret));
+		bch2_fs_fatal_error(c, "deleting logged operation %s: %s",
+				    buf.buf, bch2_err_str(ret));
 		printbuf_exit(&buf);
 	}
 }
diff --git a/fs/bcachefs/lru.c b/fs/bcachefs/lru.c
index 7a4ca5a28b..b12894ef44 100644
--- a/fs/bcachefs/lru.c
+++ b/fs/bcachefs/lru.c
@@ -11,7 +11,7 @@
 
 /* KEY_TYPE_lru is obsolete: */
 int bch2_lru_invalid(struct bch_fs *c, struct bkey_s_c k,
-		     enum bkey_invalid_flags flags,
+		     enum bch_validate_flags flags,
 		     struct printbuf *err)
 {
 	int ret = 0;
@@ -44,8 +44,8 @@ static int __bch2_lru_set(struct btree_trans *trans, u16 lru_id,
 			  u64 dev_bucket, u64 time, bool set)
 {
 	return time
-		? bch2_btree_bit_mod(trans, BTREE_ID_lru,
-				     lru_pos(lru_id, dev_bucket, time), set)
+		? bch2_btree_bit_mod_buffered(trans, BTREE_ID_lru,
+					      lru_pos(lru_id, dev_bucket, time), set)
 		: 0;
 }
 
@@ -77,6 +77,45 @@ static const char * const bch2_lru_types[] = {
 	NULL
 };
 
+int bch2_lru_check_set(struct btree_trans *trans,
+		       u16 lru_id, u64 time,
+		       struct bkey_s_c referring_k,
+		       struct bkey_buf *last_flushed)
+{
+	struct bch_fs *c = trans->c;
+	struct printbuf buf = PRINTBUF;
+	struct btree_iter lru_iter;
+	struct bkey_s_c lru_k =
+		bch2_bkey_get_iter(trans, &lru_iter, BTREE_ID_lru,
+				   lru_pos(lru_id,
+					   bucket_to_u64(referring_k.k->p),
+					   time), 0);
+	int ret = bkey_err(lru_k);
+	if (ret)
+		return ret;
+
+	if (lru_k.k->type != KEY_TYPE_set) {
+		ret = bch2_btree_write_buffer_maybe_flush(trans, referring_k, last_flushed);
+		if (ret)
+			goto err;
+
+		if (fsck_err(c, alloc_key_to_missing_lru_entry,
+			     "missing %s lru entry\n"
+			     "  %s",
+			     bch2_lru_types[lru_type(lru_k)],
+			     (bch2_bkey_val_to_text(&buf, c, referring_k), buf.buf))) {
+			ret = bch2_lru_set(trans, lru_id, bucket_to_u64(referring_k.k->p), time);
+			if (ret)
+				goto err;
+		}
+	}
+err:
+fsck_err:
+	bch2_trans_iter_exit(trans, &lru_iter);
+	printbuf_exit(&buf);
+	return ret;
+}
+
 static int bch2_check_lru_key(struct btree_trans *trans,
 			      struct btree_iter *lru_iter,
 			      struct bkey_s_c lru_k,
@@ -125,8 +164,7 @@ static int bch2_check_lru_key(struct btree_trans *trans,
 			goto out;
 		}
 
-		if (c->opts.reconstruct_alloc ||
-		    fsck_err(c, lru_entry_bad,
+		if (fsck_err(c, lru_entry_bad,
 			     "incorrect lru entry: lru %s time %llu\n"
 			     "  %s\n"
 			     "  for %s",
@@ -150,7 +188,7 @@ int bch2_check_lrus(struct bch_fs *c)
 	struct bpos last_flushed_pos = POS_MIN;
 	int ret = bch2_trans_run(c,
 		for_each_btree_key_commit(trans, iter,
-				BTREE_ID_lru, POS_MIN, BTREE_ITER_PREFETCH, k,
+				BTREE_ID_lru, POS_MIN, BTREE_ITER_prefetch, k,
 				NULL, NULL, BCH_TRANS_COMMIT_no_enospc|BCH_TRANS_COMMIT_lazy_rw,
 			bch2_check_lru_key(trans, &iter, k, &last_flushed_pos)));
 	bch_err_fn(c, ret);
diff --git a/fs/bcachefs/lru.h b/fs/bcachefs/lru.h
index 429dca816d..ed75bcf59d 100644
--- a/fs/bcachefs/lru.h
+++ b/fs/bcachefs/lru.h
@@ -2,9 +2,6 @@
 #ifndef _BCACHEFS_LRU_H
 #define _BCACHEFS_LRU_H
 
-#define LRU_TIME_BITS	48
-#define LRU_TIME_MAX	((1ULL << LRU_TIME_BITS) - 1)
-
 static inline u64 lru_pos_id(struct bpos pos)
 {
 	return pos.inode >> LRU_TIME_BITS;
@@ -49,7 +46,7 @@ static inline enum bch_lru_type lru_type(struct bkey_s_c l)
 }
 
 int bch2_lru_invalid(struct bch_fs *, struct bkey_s_c,
-		     enum bkey_invalid_flags, struct printbuf *);
+		     enum bch_validate_flags, struct printbuf *);
 void bch2_lru_to_text(struct printbuf *, struct bch_fs *, struct bkey_s_c);
 
 void bch2_lru_pos_to_text(struct printbuf *, struct bpos);
@@ -64,6 +61,9 @@ int bch2_lru_del(struct btree_trans *, u16, u64, u64);
 int bch2_lru_set(struct btree_trans *, u16, u64, u64);
 int bch2_lru_change(struct btree_trans *, u16, u64, u64, u64);
 
+struct bkey_buf;
+int bch2_lru_check_set(struct btree_trans *, u16, u64, struct bkey_s_c, struct bkey_buf *);
+
 int bch2_check_lrus(struct bch_fs *);
 
 #endif /* _BCACHEFS_LRU_H */
diff --git a/fs/bcachefs/mean_and_variance.c b/fs/bcachefs/mean_and_variance.c
index bf0ef668fd..0ea9f30803 100644
--- a/fs/bcachefs/mean_and_variance.c
+++ b/fs/bcachefs/mean_and_variance.c
@@ -103,14 +103,17 @@ EXPORT_SYMBOL_GPL(mean_and_variance_get_stddev);
  * mean_and_variance_weighted_update() - exponentially weighted variant of mean_and_variance_update()
  * @s: mean and variance number of samples and their sums
  * @x: new value to include in the &mean_and_variance_weighted
+ * @initted: caller must track whether this is the first use or not
+ * @weight: ewma weight
  *
  * see linked pdf: function derived from equations 140-143 where alpha = 2^w.
  * values are stored bitshifted for performance and added precision.
  */
-void mean_and_variance_weighted_update(struct mean_and_variance_weighted *s, s64 x)
+void mean_and_variance_weighted_update(struct mean_and_variance_weighted *s,
+		s64 x, bool initted, u8 weight)
 {
 	// previous weighted variance.
-	u8 w		= s->weight;
+	u8 w		= weight;
 	u64 var_w0	= s->variance;
 	// new value weighted.
 	s64 x_w		= x << w;
@@ -119,45 +122,50 @@ void mean_and_variance_weighted_update(struct mean_and_variance_weighted *s, s64
 	// new mean weighted.
 	s64 u_w1	= s->mean + diff;
 
-	if (!s->init) {
+	if (!initted) {
 		s->mean = x_w;
 		s->variance = 0;
 	} else {
 		s->mean = u_w1;
 		s->variance = ((var_w0 << w) - var_w0 + ((diff_w * (x_w - u_w1)) >> w)) >> w;
 	}
-	s->init = true;
 }
 EXPORT_SYMBOL_GPL(mean_and_variance_weighted_update);
 
 /**
  * mean_and_variance_weighted_get_mean() - get mean from @s
  * @s: mean and variance number of samples and their sums
+ * @weight: ewma weight
  */
-s64 mean_and_variance_weighted_get_mean(struct mean_and_variance_weighted s)
+s64 mean_and_variance_weighted_get_mean(struct mean_and_variance_weighted s,
+		u8 weight)
 {
-	return fast_divpow2(s.mean, s.weight);
+	return fast_divpow2(s.mean, weight);
 }
 EXPORT_SYMBOL_GPL(mean_and_variance_weighted_get_mean);
 
 /**
  * mean_and_variance_weighted_get_variance() -- get variance from @s
  * @s: mean and variance number of samples and their sums
+ * @weight: ewma weight
  */
-u64 mean_and_variance_weighted_get_variance(struct mean_and_variance_weighted s)
+u64 mean_and_variance_weighted_get_variance(struct mean_and_variance_weighted s,
+		u8 weight)
 {
 	// always positive don't need fast divpow2
-	return s.variance >> s.weight;
+	return s.variance >> weight;
 }
 EXPORT_SYMBOL_GPL(mean_and_variance_weighted_get_variance);
 
 /**
  * mean_and_variance_weighted_get_stddev() - get standard deviation from @s
  * @s: mean and variance number of samples and their sums
+ * @weight: ewma weight
  */
-u32 mean_and_variance_weighted_get_stddev(struct mean_and_variance_weighted s)
+u32 mean_and_variance_weighted_get_stddev(struct mean_and_variance_weighted s,
+		u8 weight)
 {
-	return int_sqrt64(mean_and_variance_weighted_get_variance(s));
+	return int_sqrt64(mean_and_variance_weighted_get_variance(s, weight));
 }
 EXPORT_SYMBOL_GPL(mean_and_variance_weighted_get_stddev);
 
diff --git a/fs/bcachefs/mean_and_variance.h b/fs/bcachefs/mean_and_variance.h
index 64df11ab42..4fcf062dd2 100644
--- a/fs/bcachefs/mean_and_variance.h
+++ b/fs/bcachefs/mean_and_variance.h
@@ -154,8 +154,6 @@ struct mean_and_variance {
 
 /* expontentially weighted variant */
 struct mean_and_variance_weighted {
-	bool	init;
-	u8	weight;	/* base 2 logarithim */
 	s64	mean;
 	u64	variance;
 };
@@ -192,10 +190,14 @@ s64 mean_and_variance_get_mean(struct mean_and_variance s);
 u64 mean_and_variance_get_variance(struct mean_and_variance s1);
 u32 mean_and_variance_get_stddev(struct mean_and_variance s);
 
-void mean_and_variance_weighted_update(struct mean_and_variance_weighted *s, s64 v);
+void mean_and_variance_weighted_update(struct mean_and_variance_weighted *s,
+		s64 v, bool initted, u8 weight);
 
-s64 mean_and_variance_weighted_get_mean(struct mean_and_variance_weighted s);
-u64 mean_and_variance_weighted_get_variance(struct mean_and_variance_weighted s);
-u32 mean_and_variance_weighted_get_stddev(struct mean_and_variance_weighted s);
+s64 mean_and_variance_weighted_get_mean(struct mean_and_variance_weighted s,
+		u8 weight);
+u64 mean_and_variance_weighted_get_variance(struct mean_and_variance_weighted s,
+		u8 weight);
+u32 mean_and_variance_weighted_get_stddev(struct mean_and_variance_weighted s,
+		u8 weight);
 
 #endif // MEAN_AND_VAIRANCE_H_
diff --git a/fs/bcachefs/mean_and_variance_test.c b/fs/bcachefs/mean_and_variance_test.c
index 51093fa848..e9d9c0212e 100644
--- a/fs/bcachefs/mean_and_variance_test.c
+++ b/fs/bcachefs/mean_and_variance_test.c
@@ -31,53 +31,59 @@ static void mean_and_variance_basic_test(struct kunit *test)
 
 static void mean_and_variance_weighted_test(struct kunit *test)
 {
-	struct mean_and_variance_weighted s = { .weight = 2 };
+	struct mean_and_variance_weighted s = { };
 
-	mean_and_variance_weighted_update(&s, 10);
-	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s), 10);
-	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s), 0);
+	mean_and_variance_weighted_update(&s, 10, false, 2);
+	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s, 2), 10);
+	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s, 2), 0);
 
-	mean_and_variance_weighted_update(&s, 20);
-	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s), 12);
-	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s), 18);
+	mean_and_variance_weighted_update(&s, 20, true, 2);
+	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s, 2), 12);
+	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s, 2), 18);
 
-	mean_and_variance_weighted_update(&s, 30);
-	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s), 16);
-	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s), 72);
+	mean_and_variance_weighted_update(&s, 30, true, 2);
+	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s, 2), 16);
+	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s, 2), 72);
 
-	s = (struct mean_and_variance_weighted) { .weight = 2 };
+	s = (struct mean_and_variance_weighted) { };
 
-	mean_and_variance_weighted_update(&s, -10);
-	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s), -10);
-	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s), 0);
+	mean_and_variance_weighted_update(&s, -10, false, 2);
+	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s, 2), -10);
+	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s, 2), 0);
 
-	mean_and_variance_weighted_update(&s, -20);
-	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s), -12);
-	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s), 18);
+	mean_and_variance_weighted_update(&s, -20, true, 2);
+	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s, 2), -12);
+	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s, 2), 18);
 
-	mean_and_variance_weighted_update(&s, -30);
-	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s), -16);
-	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s), 72);
+	mean_and_variance_weighted_update(&s, -30, true, 2);
+	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s, 2), -16);
+	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s, 2), 72);
 }
 
 static void mean_and_variance_weighted_advanced_test(struct kunit *test)
 {
-	struct mean_and_variance_weighted s = { .weight = 8 };
+	struct mean_and_variance_weighted s = { };
+	bool initted = false;
 	s64 i;
 
-	for (i = 10; i <= 100; i += 10)
-		mean_and_variance_weighted_update(&s, i);
+	for (i = 10; i <= 100; i += 10) {
+		mean_and_variance_weighted_update(&s, i, initted, 8);
+		initted = true;
+	}
 
-	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s), 11);
-	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s), 107);
+	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s, 8), 11);
+	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s, 8), 107);
 
-	s = (struct mean_and_variance_weighted) { .weight = 8 };
+	s = (struct mean_and_variance_weighted) { };
+	initted = false;
 
-	for (i = -10; i >= -100; i -= 10)
-		mean_and_variance_weighted_update(&s, i);
+	for (i = -10; i >= -100; i -= 10) {
+		mean_and_variance_weighted_update(&s, i, initted, 8);
+		initted = true;
+	}
 
-	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s), -11);
-	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s), 107);
+	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s, 8), -11);
+	KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s, 8), 107);
 }
 
 static void do_mean_and_variance_test(struct kunit *test,
@@ -92,26 +98,26 @@ static void do_mean_and_variance_test(struct kunit *test,
 				      s64 *weighted_stddev)
 {
 	struct mean_and_variance mv = {};
-	struct mean_and_variance_weighted vw = { .weight = weight };
+	struct mean_and_variance_weighted vw = { };
 
 	for (unsigned i = 0; i < initial_n; i++) {
 		mean_and_variance_update(&mv, initial_value);
-		mean_and_variance_weighted_update(&vw, initial_value);
+		mean_and_variance_weighted_update(&vw, initial_value, false, weight);
 
 		KUNIT_EXPECT_EQ(test, mean_and_variance_get_mean(mv),		initial_value);
 		KUNIT_EXPECT_EQ(test, mean_and_variance_get_stddev(mv),		0);
-		KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(vw),	initial_value);
-		KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_stddev(vw),0);
+		KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(vw, weight),	initial_value);
+		KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_stddev(vw, weight),0);
 	}
 
 	for (unsigned i = 0; i < n; i++) {
 		mean_and_variance_update(&mv, data[i]);
-		mean_and_variance_weighted_update(&vw, data[i]);
+		mean_and_variance_weighted_update(&vw, data[i], true, weight);
 
 		KUNIT_EXPECT_EQ(test, mean_and_variance_get_mean(mv),		mean[i]);
 		KUNIT_EXPECT_EQ(test, mean_and_variance_get_stddev(mv),		stddev[i]);
-		KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(vw),	weighted_mean[i]);
-		KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_stddev(vw),weighted_stddev[i]);
+		KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(vw, weight),	weighted_mean[i]);
+		KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_stddev(vw, weight),weighted_stddev[i]);
 	}
 
 	KUNIT_EXPECT_EQ(test, mv.n, initial_n + n);
@@ -211,4 +217,5 @@ static struct kunit_suite mean_and_variance_test_suite = {
 kunit_test_suite(mean_and_variance_test_suite);
 
 MODULE_AUTHOR("Daniel B. Hill");
+MODULE_DESCRIPTION("bcachefs filesystem mean and variance unit tests");
 MODULE_LICENSE("GPL");
diff --git a/fs/bcachefs/migrate.c b/fs/bcachefs/migrate.c
index 5623cee3ef..ddc187fb69 100644
--- a/fs/bcachefs/migrate.c
+++ b/fs/bcachefs/migrate.c
@@ -31,7 +31,7 @@ static int drop_dev_ptrs(struct bch_fs *c, struct bkey_s k,
 	nr_good = bch2_bkey_durability(c, k.s_c);
 	if ((!nr_good && !(flags & lost)) ||
 	    (nr_good < replicas && !(flags & degraded)))
-		return -EINVAL;
+		return -BCH_ERR_remove_would_lose_data;
 
 	return 0;
 }
@@ -49,7 +49,7 @@ static int bch2_dev_usrdata_drop_key(struct btree_trans *trans,
 	if (!bch2_bkey_has_device_c(k, dev_idx))
 		return 0;
 
-	n = bch2_bkey_make_mut(trans, iter, &k, BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
+	n = bch2_bkey_make_mut(trans, iter, &k, BTREE_UPDATE_internal_snapshot_node);
 	ret = PTR_ERR_OR_ZERO(n);
 	if (ret)
 		return ret;
@@ -67,7 +67,7 @@ static int bch2_dev_usrdata_drop_key(struct btree_trans *trans,
 
 	/*
 	 * Since we're not inserting through an extent iterator
-	 * (BTREE_ITER_ALL_SNAPSHOTS iterators aren't extent iterators),
+	 * (BTREE_ITER_all_snapshots iterators aren't extent iterators),
 	 * we aren't using the extent overwrite path to delete, we're
 	 * just using the normal key deletion path:
 	 */
@@ -87,7 +87,7 @@ static int bch2_dev_usrdata_drop(struct bch_fs *c, unsigned dev_idx, int flags)
 			continue;
 
 		ret = for_each_btree_key_commit(trans, iter, id, POS_MIN,
-				BTREE_ITER_PREFETCH|BTREE_ITER_ALL_SNAPSHOTS, k,
+				BTREE_ITER_prefetch|BTREE_ITER_all_snapshots, k,
 				NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
 			bch2_dev_usrdata_drop_key(trans, &iter, k, dev_idx, flags));
 		if (ret)
@@ -111,7 +111,7 @@ static int bch2_dev_metadata_drop(struct bch_fs *c, unsigned dev_idx, int flags)
 
 	/* don't handle this yet: */
 	if (flags & BCH_FORCE_IF_METADATA_LOST)
-		return -EINVAL;
+		return -BCH_ERR_remove_with_metadata_missing_unimplemented;
 
 	trans = bch2_trans_get(c);
 	bch2_bkey_buf_init(&k);
@@ -119,7 +119,7 @@ static int bch2_dev_metadata_drop(struct bch_fs *c, unsigned dev_idx, int flags)
 
 	for (id = 0; id < BTREE_ID_NR; id++) {
 		bch2_trans_node_iter_init(trans, &iter, id, POS_MIN, 0, 0,
-					  BTREE_ITER_PREFETCH);
+					  BTREE_ITER_prefetch);
 retry:
 		ret = 0;
 		while (bch2_trans_begin(trans),
@@ -132,10 +132,8 @@ retry:
 
 			ret = drop_dev_ptrs(c, bkey_i_to_s(k.k),
 					    dev_idx, flags, true);
-			if (ret) {
-				bch_err(c, "Cannot drop device without losing data");
+			if (ret)
 				break;
-			}
 
 			ret = bch2_btree_node_update_key(trans, &iter, b, k.k, 0, false);
 			if (bch2_err_matches(ret, BCH_ERR_transaction_restart)) {
diff --git a/fs/bcachefs/move.c b/fs/bcachefs/move.c
index bf68ea4944..e714e3bd5b 100644
--- a/fs/bcachefs/move.c
+++ b/fs/bcachefs/move.c
@@ -36,36 +36,6 @@ const char * const bch2_data_ops_strs[] = {
 	NULL
 };
 
-static void bch2_data_update_opts_to_text(struct printbuf *out, struct bch_fs *c,
-					  struct bch_io_opts *io_opts,
-					  struct data_update_opts *data_opts)
-{
-	printbuf_tabstop_push(out, 20);
-	prt_str(out, "rewrite ptrs:");
-	prt_tab(out);
-	bch2_prt_u64_base2(out, data_opts->rewrite_ptrs);
-	prt_newline(out);
-
-	prt_str(out, "kill ptrs: ");
-	prt_tab(out);
-	bch2_prt_u64_base2(out, data_opts->kill_ptrs);
-	prt_newline(out);
-
-	prt_str(out, "target: ");
-	prt_tab(out);
-	bch2_target_to_text(out, c, data_opts->target);
-	prt_newline(out);
-
-	prt_str(out, "compression: ");
-	prt_tab(out);
-	bch2_compression_opt_to_text(out, background_compression(*io_opts));
-	prt_newline(out);
-
-	prt_str(out, "extra replicas: ");
-	prt_tab(out);
-	prt_u64(out, data_opts->extra_replicas);
-}
-
 static void trace_move_extent2(struct bch_fs *c, struct bkey_s_c k,
 			       struct bch_io_opts *io_opts,
 			       struct data_update_opts *data_opts)
@@ -421,7 +391,7 @@ struct bch_io_opts *bch2_move_get_io_opts(struct btree_trans *trans,
 		io_opts->d.nr = 0;
 
 		ret = for_each_btree_key(trans, iter, BTREE_ID_inodes, POS(0, extent_k.k->p.inode),
-					 BTREE_ITER_ALL_SNAPSHOTS, k, ({
+					 BTREE_ITER_all_snapshots, k, ({
 			if (k.k->p.offset != extent_k.k->p.inode)
 				break;
 
@@ -467,7 +437,7 @@ int bch2_move_get_io_opts_one(struct btree_trans *trans,
 
 	k = bch2_bkey_get_iter(trans, &iter, BTREE_ID_inodes,
 			       SPOS(0, extent_k.k->p.inode, extent_k.k->p.snapshot),
-			       BTREE_ITER_CACHED);
+			       BTREE_ITER_cached);
 	ret = bkey_err(k);
 	if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
 		return ret;
@@ -552,9 +522,10 @@ static int bch2_move_data_btree(struct moving_context *ctxt,
 		ctxt->stats->pos	= BBPOS(btree_id, start);
 	}
 
+	bch2_trans_begin(trans);
 	bch2_trans_iter_init(trans, &iter, btree_id, start,
-			     BTREE_ITER_PREFETCH|
-			     BTREE_ITER_ALL_SNAPSHOTS);
+			     BTREE_ITER_prefetch|
+			     BTREE_ITER_all_snapshots);
 
 	if (ctxt->rate)
 		bch2_ratelimit_reset(ctxt->rate);
@@ -695,6 +666,10 @@ int bch2_evacuate_bucket(struct moving_context *ctxt,
 	struct bpos bp_pos = POS_MIN;
 	int ret = 0;
 
+	struct bch_dev *ca = bch2_dev_tryget(c, bucket.inode);
+	if (!ca)
+		return 0;
+
 	trace_bucket_evacuate(c, &bucket);
 
 	bch2_bkey_buf_init(&sk);
@@ -705,7 +680,7 @@ int bch2_evacuate_bucket(struct moving_context *ctxt,
 	bch2_trans_begin(trans);
 
 	bch2_trans_iter_init(trans, &iter, BTREE_ID_alloc,
-			     bucket, BTREE_ITER_CACHED);
+			     bucket, BTREE_ITER_cached);
 	ret = lockrestart_do(trans,
 			bkey_err(k = bch2_btree_iter_peek_slot(&iter)));
 	bch2_trans_iter_exit(trans, &iter);
@@ -716,7 +691,7 @@ int bch2_evacuate_bucket(struct moving_context *ctxt,
 
 	a = bch2_alloc_to_v4(k, &a_convert);
 	dirty_sectors = bch2_bucket_sectors_dirty(*a);
-	bucket_size = bch_dev_bkey_exists(c, bucket.inode)->mi.bucket_size;
+	bucket_size = ca->mi.bucket_size;
 	fragmentation = a->fragmentation_lru;
 
 	ret = bch2_btree_write_buffer_tryflush(trans);
@@ -730,9 +705,9 @@ int bch2_evacuate_bucket(struct moving_context *ctxt,
 
 		bch2_trans_begin(trans);
 
-		ret = bch2_get_next_backpointer(trans, bucket, gen,
+		ret = bch2_get_next_backpointer(trans, ca, bucket, gen,
 						&bp_pos, &bp,
-						BTREE_ITER_CACHED);
+						BTREE_ITER_cached);
 		if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
 			continue;
 		if (ret)
@@ -828,6 +803,7 @@ next:
 
 	trace_evacuate_bucket(c, &bucket, dirty_sectors, bucket_size, fragmentation, ret);
 err:
+	bch2_dev_put(ca);
 	bch2_bkey_buf_exit(&sk, c);
 	return ret;
 }
@@ -868,7 +844,7 @@ static int bch2_move_btree(struct bch_fs *c,
 			continue;
 
 		bch2_trans_node_iter_init(trans, &iter, btree, POS_MIN, 0, 0,
-					  BTREE_ITER_PREFETCH);
+					  BTREE_ITER_prefetch);
 retry:
 		ret = 0;
 		while (bch2_trans_begin(trans),
@@ -920,7 +896,20 @@ static bool rereplicate_pred(struct bch_fs *c, void *arg,
 		? c->opts.metadata_replicas
 		: io_opts->data_replicas;
 
-	if (!nr_good || nr_good >= replicas)
+	rcu_read_lock();
+	struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k);
+	unsigned i = 0;
+	bkey_for_each_ptr(ptrs, ptr) {
+		struct bch_dev *ca = bch2_dev_rcu(c, ptr->dev);
+		if (!ptr->cached &&
+		    (!ca || !ca->mi.durability))
+			data_opts->kill_ptrs |= BIT(i);
+		i++;
+	}
+	rcu_read_unlock();
+
+	if (!data_opts->kill_ptrs &&
+	    (!nr_good || nr_good >= replicas))
 		return false;
 
 	data_opts->target		= 0;
@@ -968,27 +957,17 @@ static bool migrate_btree_pred(struct bch_fs *c, void *arg,
 	return migrate_pred(c, arg, bkey_i_to_s_c(&b->key), io_opts, data_opts);
 }
 
+/*
+ * Ancient versions of bcachefs produced packed formats which could represent
+ * keys that the in memory format cannot represent; this checks for those
+ * formats so we can get rid of them.
+ */
 static bool bformat_needs_redo(struct bkey_format *f)
 {
-	unsigned i;
-
-	for (i = 0; i < f->nr_fields; i++) {
-		unsigned unpacked_bits = bch2_bkey_format_current.bits_per_field[i];
-		u64 unpacked_mask = ~((~0ULL << 1) << (unpacked_bits - 1));
-		u64 field_offset = le64_to_cpu(f->field_offset[i]);
-
-		if (f->bits_per_field[i] > unpacked_bits)
-			return true;
-
-		if ((f->bits_per_field[i] == unpacked_bits) && field_offset)
+	for (unsigned i = 0; i < f->nr_fields; i++)
+		if (bch2_bkey_format_field_overflows(f, i))
 			return true;
 
-		if (((field_offset + ((1ULL << f->bits_per_field[i]) - 1)) &
-		     unpacked_mask) <
-		    field_offset)
-			return true;
-	}
-
 	return false;
 }
 
@@ -1043,6 +1022,7 @@ static bool drop_extra_replicas_pred(struct bch_fs *c, void *arg,
 	struct extent_ptr_decoded p;
 	unsigned i = 0;
 
+	rcu_read_lock();
 	bkey_for_each_ptr_decode(k.k, bch2_bkey_ptrs_c(k), p, entry) {
 		unsigned d = bch2_extent_ptr_durability(c, &p);
 
@@ -1053,6 +1033,7 @@ static bool drop_extra_replicas_pred(struct bch_fs *c, void *arg,
 
 		i++;
 	}
+	rcu_read_unlock();
 
 	return data_opts->kill_ptrs != 0;
 }
@@ -1137,23 +1118,17 @@ void bch2_move_stats_to_text(struct printbuf *out, struct bch_move_stats *stats)
 	prt_newline(out);
 	printbuf_indent_add(out, 2);
 
-	prt_str(out, "keys moved:  ");
-	prt_u64(out, atomic64_read(&stats->keys_moved));
-	prt_newline(out);
-
-	prt_str(out, "keys raced:  ");
-	prt_u64(out, atomic64_read(&stats->keys_raced));
-	prt_newline(out);
-
-	prt_str(out, "bytes seen:  ");
+	prt_printf(out, "keys moved:  %llu\n",	atomic64_read(&stats->keys_moved));
+	prt_printf(out, "keys raced:  %llu\n",	atomic64_read(&stats->keys_raced));
+	prt_printf(out, "bytes seen:  ");
 	prt_human_readable_u64(out, atomic64_read(&stats->sectors_seen) << 9);
 	prt_newline(out);
 
-	prt_str(out, "bytes moved: ");
+	prt_printf(out, "bytes moved: ");
 	prt_human_readable_u64(out, atomic64_read(&stats->sectors_moved) << 9);
 	prt_newline(out);
 
-	prt_str(out, "bytes raced: ");
+	prt_printf(out, "bytes raced: ");
 	prt_human_readable_u64(out, atomic64_read(&stats->sectors_raced) << 9);
 	prt_newline(out);
 
@@ -1167,19 +1142,17 @@ static void bch2_moving_ctxt_to_text(struct printbuf *out, struct bch_fs *c, str
 	bch2_move_stats_to_text(out, ctxt->stats);
 	printbuf_indent_add(out, 2);
 
-	prt_printf(out, "reads: ios %u/%u sectors %u/%u",
+	prt_printf(out, "reads: ios %u/%u sectors %u/%u\n",
 		   atomic_read(&ctxt->read_ios),
 		   c->opts.move_ios_in_flight,
 		   atomic_read(&ctxt->read_sectors),
 		   c->opts.move_bytes_in_flight >> 9);
-	prt_newline(out);
 
-	prt_printf(out, "writes: ios %u/%u sectors %u/%u",
+	prt_printf(out, "writes: ios %u/%u sectors %u/%u\n",
 		   atomic_read(&ctxt->write_ios),
 		   c->opts.move_ios_in_flight,
 		   atomic_read(&ctxt->write_sectors),
 		   c->opts.move_bytes_in_flight >> 9);
-	prt_newline(out);
 
 	printbuf_indent_add(out, 2);
 
diff --git a/fs/bcachefs/movinggc.c b/fs/bcachefs/movinggc.c
index 69e06a84da..eb49dd045e 100644
--- a/fs/bcachefs/movinggc.c
+++ b/fs/bcachefs/movinggc.c
@@ -35,9 +35,10 @@ struct buckets_in_flight {
 };
 
 static const struct rhashtable_params bch_move_bucket_params = {
-	.head_offset	= offsetof(struct move_bucket_in_flight, hash),
-	.key_offset	= offsetof(struct move_bucket_in_flight, bucket.k),
-	.key_len	= sizeof(struct move_bucket_key),
+	.head_offset		= offsetof(struct move_bucket_in_flight, hash),
+	.key_offset		= offsetof(struct move_bucket_in_flight, bucket.k),
+	.key_len		= sizeof(struct move_bucket_key),
+	.automatic_shrinking	= true,
 };
 
 static struct move_bucket_in_flight *
@@ -84,7 +85,7 @@ static int bch2_bucket_is_movable(struct btree_trans *trans,
 		return 0;
 
 	k = bch2_bkey_get_iter(trans, &iter, BTREE_ID_alloc,
-			       b->k.bucket, BTREE_ITER_CACHED);
+			       b->k.bucket, BTREE_ITER_cached);
 	ret = bkey_err(k);
 	if (ret)
 		return ret;
@@ -155,10 +156,11 @@ static int bch2_copygc_get_buckets(struct moving_context *ctxt,
 	if (bch2_err_matches(ret, EROFS))
 		return ret;
 
-	if (bch2_fs_fatal_err_on(ret, c, "%s: error %s from bch2_btree_write_buffer_tryflush()",
-				 __func__, bch2_err_str(ret)))
+	if (bch2_fs_fatal_err_on(ret, c, "%s: from bch2_btree_write_buffer_tryflush()", bch2_err_str(ret)))
 		return ret;
 
+	bch2_trans_begin(trans);
+
 	ret = for_each_btree_key_upto(trans, iter, BTREE_ID_lru,
 				  lru_pos(BCH_LRU_FRAGMENTATION_START, 0, 0),
 				  lru_pos(BCH_LRU_FRAGMENTATION_START, U64_MAX, LRU_TIME_MAX),
diff --git a/fs/bcachefs/opts.c b/fs/bcachefs/opts.c
index b1ed0b9a20..bb068fd724 100644
--- a/fs/bcachefs/opts.c
+++ b/fs/bcachefs/opts.c
@@ -7,6 +7,7 @@
 #include "disk_groups.h"
 #include "error.h"
 #include "opts.h"
+#include "recovery_passes.h"
 #include "super-io.h"
 #include "util.h"
 
@@ -42,7 +43,7 @@ const char * const __bch2_btree_ids[] = {
 	NULL
 };
 
-const char * const bch2_csum_types[] = {
+static const char * const __bch2_csum_types[] = {
 	BCH_CSUM_TYPES()
 	NULL
 };
@@ -52,7 +53,7 @@ const char * const bch2_csum_opts[] = {
 	NULL
 };
 
-const char * const __bch2_compression_types[] = {
+static const char * const __bch2_compression_types[] = {
 	BCH_COMPRESSION_TYPES()
 	NULL
 };
@@ -82,18 +83,39 @@ const char * const bch2_member_states[] = {
 	NULL
 };
 
-const char * const bch2_jset_entry_types[] = {
+static const char * const __bch2_jset_entry_types[] = {
 	BCH_JSET_ENTRY_TYPES()
 	NULL
 };
 
-const char * const bch2_fs_usage_types[] = {
+static const char * const __bch2_fs_usage_types[] = {
 	BCH_FS_USAGE_TYPES()
 	NULL
 };
 
 #undef x
 
+static void prt_str_opt_boundscheck(struct printbuf *out, const char * const opts[],
+				    unsigned nr, const char *type, unsigned idx)
+{
+	if (idx < nr)
+		prt_str(out, opts[idx]);
+	else
+		prt_printf(out, "(unknown %s %u)", type, idx);
+}
+
+#define PRT_STR_OPT_BOUNDSCHECKED(name, type)					\
+void bch2_prt_##name(struct printbuf *out, type t)				\
+{										\
+	prt_str_opt_boundscheck(out, __bch2_##name##s, ARRAY_SIZE(__bch2_##name##s) - 1, #name, t);\
+}
+
+PRT_STR_OPT_BOUNDSCHECKED(jset_entry_type,	enum bch_jset_entry_type);
+PRT_STR_OPT_BOUNDSCHECKED(fs_usage_type,	enum bch_fs_usage_type);
+PRT_STR_OPT_BOUNDSCHECKED(data_type,		enum bch_data_type);
+PRT_STR_OPT_BOUNDSCHECKED(csum_type,		enum bch_csum_type);
+PRT_STR_OPT_BOUNDSCHECKED(compression_type,	enum bch_compression_type);
+
 static int bch2_opt_fix_errors_parse(struct bch_fs *c, const char *val, u64 *res,
 				     struct printbuf *err)
 {
@@ -205,6 +227,9 @@ const struct bch_option bch2_opt_table[] = {
 #define OPT_STR(_choices)	.type = BCH_OPT_STR,			\
 				.min = 0, .max = ARRAY_SIZE(_choices),	\
 				.choices = _choices
+#define OPT_STR_NOLIMIT(_choices)	.type = BCH_OPT_STR,		\
+				.min = 0, .max = U64_MAX,		\
+				.choices = _choices
 #define OPT_FN(_fn)		.type = BCH_OPT_FN, .fn	= _fn
 
 #define x(_name, _bits, _flags, _type, _sb_opt, _default, _hint, _help)	\
@@ -314,7 +339,7 @@ int bch2_opt_parse(struct bch_fs *c,
 		if (ret < 0 || (*res != 0 && *res != 1)) {
 			if (err)
 				prt_printf(err, "%s: must be bool", opt->attr.name);
-			return ret;
+			return ret < 0 ? ret : -BCH_ERR_option_not_bool;
 		}
 		break;
 	case BCH_OPT_UINT:
@@ -456,7 +481,7 @@ int bch2_parse_mount_opts(struct bch_fs *c, struct bch_opts *opts,
 
 	copied_opts = kstrdup(options, GFP_KERNEL);
 	if (!copied_opts)
-		return -1;
+		return -ENOMEM;
 	copied_opts_start = copied_opts;
 
 	while ((opt = strsep(&copied_opts, ",")) != NULL) {
@@ -501,11 +526,11 @@ int bch2_parse_mount_opts(struct bch_fs *c, struct bch_opts *opts,
 
 bad_opt:
 	pr_err("Bad mount option %s", name);
-	ret = -1;
+	ret = -BCH_ERR_option_name;
 	goto out;
 bad_val:
 	pr_err("Invalid mount option %s", err.buf);
-	ret = -1;
+	ret = -BCH_ERR_option_value;
 	goto out;
 out:
 	kfree(copied_opts_start);
diff --git a/fs/bcachefs/opts.h b/fs/bcachefs/opts.h
index 9a4b7faa37..b197ec90d4 100644
--- a/fs/bcachefs/opts.h
+++ b/fs/bcachefs/opts.h
@@ -16,18 +16,20 @@ extern const char * const bch2_version_upgrade_opts[];
 extern const char * const bch2_sb_features[];
 extern const char * const bch2_sb_compat[];
 extern const char * const __bch2_btree_ids[];
-extern const char * const bch2_csum_types[];
 extern const char * const bch2_csum_opts[];
-extern const char * const __bch2_compression_types[];
 extern const char * const bch2_compression_opts[];
 extern const char * const bch2_str_hash_types[];
 extern const char * const bch2_str_hash_opts[];
 extern const char * const __bch2_data_types[];
 extern const char * const bch2_member_states[];
-extern const char * const bch2_jset_entry_types[];
-extern const char * const bch2_fs_usage_types[];
 extern const char * const bch2_d_types[];
 
+void bch2_prt_jset_entry_type(struct printbuf *,	enum bch_jset_entry_type);
+void bch2_prt_fs_usage_type(struct printbuf *,		enum bch_fs_usage_type);
+void bch2_prt_data_type(struct printbuf *,		enum bch_data_type);
+void bch2_prt_csum_type(struct printbuf *,		enum bch_csum_type);
+void bch2_prt_compression_type(struct printbuf *,	enum bch_compression_type);
+
 static inline const char *bch2_d_type_str(unsigned d_type)
 {
 	return (d_type < BCH_DT_MAX ? bch2_d_types[d_type] : NULL) ?: "(bad d_type)";
@@ -135,7 +137,7 @@ enum fsck_err_opts {
 	x(errors,			u8,				\
 	  OPT_FS|OPT_FORMAT|OPT_MOUNT|OPT_RUNTIME,			\
 	  OPT_STR(bch2_error_actions),					\
-	  BCH_SB_ERROR_ACTION,		BCH_ON_ERROR_ro,		\
+	  BCH_SB_ERROR_ACTION,		BCH_ON_ERROR_fix_safe,		\
 	  NULL,		"Action to take on filesystem error")		\
 	x(metadata_replicas,		u8,				\
 	  OPT_FS|OPT_FORMAT|OPT_MOUNT|OPT_RUNTIME,			\
@@ -290,6 +292,11 @@ enum fsck_err_opts {
 	  OPT_BOOL(),							\
 	  BCH2_NO_SB_OPT,		false,				\
 	  NULL,		"Allow mounting in when data will be missing")	\
+	x(no_splitbrain_check,		u8,				\
+	  OPT_FS|OPT_MOUNT,						\
+	  OPT_BOOL(),							\
+	  BCH2_NO_SB_OPT,		false,				\
+	  NULL,		"Don't kick drives out when splitbrain detected")\
 	x(discard,			u8,				\
 	  OPT_FS|OPT_MOUNT|OPT_DEVICE,					\
 	  OPT_BOOL(),							\
@@ -332,6 +339,11 @@ enum fsck_err_opts {
 	  OPT_BOOL(),							\
 	  BCH2_NO_SB_OPT,		false,				\
 	  NULL,		"Run fsck on mount")				\
+	x(fsck_memory_usage_percent,	u8,				\
+	  OPT_FS|OPT_MOUNT,						\
+	  OPT_UINT(20, 70),						\
+	  BCH2_NO_SB_OPT,		50,				\
+	  NULL,		"Maximum percentage of system ram fsck is allowed to pin")\
 	x(fix_errors,			u8,				\
 	  OPT_FS|OPT_MOUNT,						\
 	  OPT_FN(bch2_opt_fix_errors),					\
@@ -352,12 +364,17 @@ enum fsck_err_opts {
 	  OPT_FS|OPT_MOUNT,						\
 	  OPT_BOOL(),							\
 	  BCH2_NO_SB_OPT,		false,				\
-	  NULL,		"Don't replay the journal")			\
-	x(keep_journal,			u8,				\
+	  NULL,		"Exit recovery immediately prior to journal replay")\
+	x(recovery_pass_last,		u8,				\
+	  OPT_FS|OPT_MOUNT,						\
+	  OPT_STR_NOLIMIT(bch2_recovery_passes),			\
+	  BCH2_NO_SB_OPT,		0,				\
+	  NULL,		"Exit recovery after specified pass")		\
+	x(retain_recovery_info,		u8,				\
 	  0,								\
 	  OPT_BOOL(),							\
 	  BCH2_NO_SB_OPT,		false,				\
-	  NULL,		"Don't free journal entries/keys after startup")\
+	  NULL,		"Don't free journal entries/keys, scanned btree nodes after startup")\
 	x(read_entire_journal,		u8,				\
 	  0,								\
 	  OPT_BOOL(),							\
@@ -409,11 +426,6 @@ enum fsck_err_opts {
 	  BCH_SB_VERSION_UPGRADE,	BCH_VERSION_UPGRADE_compatible,	\
 	  NULL,		"Set superblock to latest version,\n"		\
 			"allowing any new features to be used")		\
-	x(buckets_nouse,		u8,				\
-	  0,								\
-	  OPT_BOOL(),							\
-	  BCH2_NO_SB_OPT,		false,				\
-	  NULL,		"Allocate the buckets_nouse bitmap")		\
 	x(stdio,			u64,				\
 	  0,								\
 	  OPT_UINT(0, S64_MAX),						\
@@ -463,7 +475,7 @@ enum fsck_err_opts {
 	  OPT_FS|OPT_MOUNT|OPT_RUNTIME,					\
 	  OPT_BOOL(),							\
 	  BCH2_NO_SB_OPT,		true,				\
-	  NULL,		"BTREE_ITER_PREFETCH casuse btree nodes to be\n"\
+	  NULL,		"BTREE_ITER_prefetch casuse btree nodes to be\n"\
 	  " prefetched sequentially")
 
 struct bch_opts {
diff --git a/fs/bcachefs/printbuf.c b/fs/bcachefs/printbuf.c
index b27d229259..9f529e4c1b 100644
--- a/fs/bcachefs/printbuf.c
+++ b/fs/bcachefs/printbuf.c
@@ -10,35 +10,57 @@
 
 #include "printbuf.h"
 
-static inline unsigned printbuf_linelen(struct printbuf *buf)
+static inline unsigned __printbuf_linelen(struct printbuf *buf, unsigned pos)
 {
-	return buf->pos - buf->last_newline;
+	return pos - buf->last_newline;
 }
 
-int bch2_printbuf_make_room(struct printbuf *out, unsigned extra)
+static inline unsigned printbuf_linelen(struct printbuf *buf)
 {
-	unsigned new_size;
-	char *buf;
+	return __printbuf_linelen(buf, buf->pos);
+}
 
-	if (!out->heap_allocated)
-		return 0;
+/*
+ * Returns spaces from start of line, if set, or 0 if unset:
+ */
+static inline unsigned cur_tabstop(struct printbuf *buf)
+{
+	return buf->cur_tabstop < buf->nr_tabstops
+		? buf->_tabstops[buf->cur_tabstop]
+		: 0;
+}
 
+int bch2_printbuf_make_room(struct printbuf *out, unsigned extra)
+{
 	/* Reserved space for terminating nul: */
 	extra += 1;
 
-	if (out->pos + extra < out->size)
+	if (out->pos + extra <= out->size)
 		return 0;
 
-	new_size = roundup_pow_of_two(out->size + extra);
+	if (!out->heap_allocated) {
+		out->overflow = true;
+		return 0;
+	}
+
+	unsigned new_size = roundup_pow_of_two(out->size + extra);
+
+	/* Sanity check... */
+	if (new_size > PAGE_SIZE << MAX_PAGE_ORDER) {
+		out->allocation_failure = true;
+		out->overflow = true;
+		return -ENOMEM;
+	}
 
 	/*
 	 * Note: output buffer must be freeable with kfree(), it's not required
 	 * that the user use printbuf_exit().
 	 */
-	buf = krealloc(out->buf, new_size, !out->atomic ? GFP_KERNEL : GFP_NOWAIT);
+	char *buf = krealloc(out->buf, new_size, !out->atomic ? GFP_KERNEL : GFP_NOWAIT);
 
 	if (!buf) {
 		out->allocation_failure = true;
+		out->overflow = true;
 		return -ENOMEM;
 	}
 
@@ -47,6 +69,92 @@ int bch2_printbuf_make_room(struct printbuf *out, unsigned extra)
 	return 0;
 }
 
+static void printbuf_advance_pos(struct printbuf *out, unsigned len)
+{
+	out->pos += min(len, printbuf_remaining(out));
+}
+
+static void printbuf_insert_spaces(struct printbuf *out, unsigned pos, unsigned nr)
+{
+	unsigned move = out->pos - pos;
+
+	bch2_printbuf_make_room(out, nr);
+
+	if (pos + nr < out->size)
+		memmove(out->buf + pos + nr,
+			out->buf + pos,
+			min(move, out->size - 1 - pos - nr));
+
+	if (pos < out->size)
+		memset(out->buf + pos, ' ', min(nr, out->size - pos));
+
+	printbuf_advance_pos(out, nr);
+	printbuf_nul_terminate_reserved(out);
+}
+
+static void __printbuf_do_indent(struct printbuf *out, unsigned pos)
+{
+	while (true) {
+		int pad;
+		unsigned len = out->pos - pos;
+		char *p = out->buf + pos;
+		char *n = memscan(p, '\n', len);
+		if (cur_tabstop(out)) {
+			n = min(n, (char *) memscan(p, '\r', len));
+			n = min(n, (char *) memscan(p, '\t', len));
+		}
+
+		pos = n - out->buf;
+		if (pos == out->pos)
+			break;
+
+		switch (*n) {
+		case '\n':
+			pos++;
+			out->last_newline = pos;
+
+			printbuf_insert_spaces(out, pos, out->indent);
+
+			pos = min(pos + out->indent, out->pos);
+			out->last_field = pos;
+			out->cur_tabstop = 0;
+			break;
+		case '\r':
+			memmove(n, n + 1, out->pos - pos);
+			--out->pos;
+			pad = (int) cur_tabstop(out) - (int) __printbuf_linelen(out, pos);
+			if (pad > 0) {
+				printbuf_insert_spaces(out, out->last_field, pad);
+				pos += pad;
+			}
+
+			out->last_field = pos;
+			out->cur_tabstop++;
+			break;
+		case '\t':
+			pad = (int) cur_tabstop(out) - (int) __printbuf_linelen(out, pos) - 1;
+			if (pad > 0) {
+				*n = ' ';
+				printbuf_insert_spaces(out, pos, pad - 1);
+				pos += pad;
+			} else {
+				memmove(n, n + 1, out->pos - pos);
+				--out->pos;
+			}
+
+			out->last_field = pos;
+			out->cur_tabstop++;
+			break;
+		}
+	}
+}
+
+static inline void printbuf_do_indent(struct printbuf *out, unsigned pos)
+{
+	if (out->has_indent_or_tabstops && !out->suppress_indent_tabstop_handling)
+		__printbuf_do_indent(out, pos);
+}
+
 void bch2_prt_vprintf(struct printbuf *out, const char *fmt, va_list args)
 {
 	int len;
@@ -55,14 +163,14 @@ void bch2_prt_vprintf(struct printbuf *out, const char *fmt, va_list args)
 		va_list args2;
 
 		va_copy(args2, args);
-		len = vsnprintf(out->buf + out->pos, printbuf_remaining(out), fmt, args2);
+		len = vsnprintf(out->buf + out->pos, printbuf_remaining_size(out), fmt, args2);
 		va_end(args2);
-	} while (len + 1 >= printbuf_remaining(out) &&
-		 !bch2_printbuf_make_room(out, len + 1));
+	} while (len > printbuf_remaining(out) &&
+		 !bch2_printbuf_make_room(out, len));
 
-	len = min_t(size_t, len,
-		  printbuf_remaining(out) ? printbuf_remaining(out) - 1 : 0);
-	out->pos += len;
+	unsigned indent_pos = out->pos;
+	printbuf_advance_pos(out, len);
+	printbuf_do_indent(out, indent_pos);
 }
 
 void bch2_prt_printf(struct printbuf *out, const char *fmt, ...)
@@ -72,14 +180,14 @@ void bch2_prt_printf(struct printbuf *out, const char *fmt, ...)
 
 	do {
 		va_start(args, fmt);
-		len = vsnprintf(out->buf + out->pos, printbuf_remaining(out), fmt, args);
+		len = vsnprintf(out->buf + out->pos, printbuf_remaining_size(out), fmt, args);
 		va_end(args);
-	} while (len + 1 >= printbuf_remaining(out) &&
-		 !bch2_printbuf_make_room(out, len + 1));
+	} while (len > printbuf_remaining(out) &&
+		 !bch2_printbuf_make_room(out, len));
 
-	len = min_t(size_t, len,
-		  printbuf_remaining(out) ? printbuf_remaining(out) - 1 : 0);
-	out->pos += len;
+	unsigned indent_pos = out->pos;
+	printbuf_advance_pos(out, len);
+	printbuf_do_indent(out, indent_pos);
 }
 
 /**
@@ -194,33 +302,20 @@ void bch2_printbuf_indent_sub(struct printbuf *buf, unsigned spaces)
 
 void bch2_prt_newline(struct printbuf *buf)
 {
-	unsigned i;
-
 	bch2_printbuf_make_room(buf, 1 + buf->indent);
 
-	__prt_char(buf, '\n');
+	__prt_char_reserved(buf, '\n');
 
 	buf->last_newline	= buf->pos;
 
-	for (i = 0; i < buf->indent; i++)
-		__prt_char(buf, ' ');
+	__prt_chars_reserved(buf, ' ', buf->indent);
 
-	printbuf_nul_terminate(buf);
+	printbuf_nul_terminate_reserved(buf);
 
 	buf->last_field		= buf->pos;
 	buf->cur_tabstop	= 0;
 }
 
-/*
- * Returns spaces from start of line, if set, or 0 if unset:
- */
-static inline unsigned cur_tabstop(struct printbuf *buf)
-{
-	return buf->cur_tabstop < buf->nr_tabstops
-		? buf->_tabstops[buf->cur_tabstop]
-		: 0;
-}
-
 static void __prt_tab(struct printbuf *out)
 {
 	int spaces = max_t(int, 0, cur_tabstop(out) - printbuf_linelen(out));
@@ -247,24 +342,9 @@ void bch2_prt_tab(struct printbuf *out)
 
 static void __prt_tab_rjust(struct printbuf *buf)
 {
-	unsigned move = buf->pos - buf->last_field;
 	int pad = (int) cur_tabstop(buf) - (int) printbuf_linelen(buf);
-
-	if (pad > 0) {
-		bch2_printbuf_make_room(buf, pad);
-
-		if (buf->last_field + pad < buf->size)
-			memmove(buf->buf + buf->last_field + pad,
-				buf->buf + buf->last_field,
-				min(move, buf->size - 1 - buf->last_field - pad));
-
-		if (buf->last_field < buf->size)
-			memset(buf->buf + buf->last_field, ' ',
-			       min((unsigned) pad, buf->size - buf->last_field));
-
-		buf->pos += pad;
-		printbuf_nul_terminate(buf);
-	}
+	if (pad > 0)
+		printbuf_insert_spaces(buf, buf->last_field, pad);
 
 	buf->last_field = buf->pos;
 	buf->cur_tabstop++;
@@ -301,41 +381,9 @@ void bch2_prt_tab_rjust(struct printbuf *buf)
  */
 void bch2_prt_bytes_indented(struct printbuf *out, const char *str, unsigned count)
 {
-	const char *unprinted_start = str;
-	const char *end = str + count;
-
-	if (!out->has_indent_or_tabstops || out->suppress_indent_tabstop_handling) {
-		prt_bytes(out, str, count);
-		return;
-	}
-
-	while (str != end) {
-		switch (*str) {
-		case '\n':
-			prt_bytes(out, unprinted_start, str - unprinted_start);
-			unprinted_start = str + 1;
-			bch2_prt_newline(out);
-			break;
-		case '\t':
-			if (likely(cur_tabstop(out))) {
-				prt_bytes(out, unprinted_start, str - unprinted_start);
-				unprinted_start = str + 1;
-				__prt_tab(out);
-			}
-			break;
-		case '\r':
-			if (likely(cur_tabstop(out))) {
-				prt_bytes(out, unprinted_start, str - unprinted_start);
-				unprinted_start = str + 1;
-				__prt_tab_rjust(out);
-			}
-			break;
-		}
-
-		str++;
-	}
-
-	prt_bytes(out, unprinted_start, str - unprinted_start);
+	unsigned indent_pos = out->pos;
+	prt_bytes(out, str, count);
+	printbuf_do_indent(out, indent_pos);
 }
 
 /**
@@ -348,9 +396,10 @@ void bch2_prt_bytes_indented(struct printbuf *out, const char *str, unsigned cou
 void bch2_prt_human_readable_u64(struct printbuf *out, u64 v)
 {
 	bch2_printbuf_make_room(out, 10);
-	out->pos += string_get_size(v, 1, !out->si_units,
-				    out->buf + out->pos,
-				    printbuf_remaining_size(out));
+	unsigned len = string_get_size(v, 1, !out->si_units,
+				       out->buf + out->pos,
+				       printbuf_remaining_size(out));
+	printbuf_advance_pos(out, len);
 }
 
 /**
@@ -402,9 +451,7 @@ void bch2_prt_string_option(struct printbuf *out,
 			    const char * const list[],
 			    size_t selected)
 {
-	size_t i;
-
-	for (i = 0; list[i]; i++)
+	for (size_t i = 0; list[i]; i++)
 		bch2_prt_printf(out, i == selected ? "[%s] " : "%s ", list[i]);
 }
 
diff --git a/fs/bcachefs/printbuf.h b/fs/bcachefs/printbuf.h
index 9a4a56c409..9ecc56bc96 100644
--- a/fs/bcachefs/printbuf.h
+++ b/fs/bcachefs/printbuf.h
@@ -86,6 +86,7 @@ struct printbuf {
 	u8			atomic;
 	bool			allocation_failure:1;
 	bool			heap_allocated:1;
+	bool			overflow:1;
 	enum printbuf_si	si_units:1;
 	bool			human_readable_units:1;
 	bool			has_indent_or_tabstops:1;
@@ -142,7 +143,9 @@ void bch2_prt_bitflags_vector(struct printbuf *, const char * const[],
  */
 static inline unsigned printbuf_remaining_size(struct printbuf *out)
 {
-	return out->pos < out->size ? out->size - out->pos : 0;
+	if (WARN_ON(out->size && out->pos >= out->size))
+		out->pos = out->size - 1;
+	return out->size - out->pos;
 }
 
 /*
@@ -151,7 +154,7 @@ static inline unsigned printbuf_remaining_size(struct printbuf *out)
  */
 static inline unsigned printbuf_remaining(struct printbuf *out)
 {
-	return out->pos < out->size ? out->size - out->pos - 1 : 0;
+	return out->size ? printbuf_remaining_size(out) - 1 : 0;
 }
 
 static inline unsigned printbuf_written(struct printbuf *out)
@@ -159,30 +162,25 @@ static inline unsigned printbuf_written(struct printbuf *out)
 	return out->size ? min(out->pos, out->size - 1) : 0;
 }
 
-/*
- * Returns true if output was truncated:
- */
-static inline bool printbuf_overflowed(struct printbuf *out)
+static inline void printbuf_nul_terminate_reserved(struct printbuf *out)
 {
-	return out->pos >= out->size;
+	if (WARN_ON(out->size && out->pos >= out->size))
+		out->pos = out->size - 1;
+	if (out->size)
+		out->buf[out->pos] = 0;
 }
 
 static inline void printbuf_nul_terminate(struct printbuf *out)
 {
 	bch2_printbuf_make_room(out, 1);
-
-	if (out->pos < out->size)
-		out->buf[out->pos] = 0;
-	else if (out->size)
-		out->buf[out->size - 1] = 0;
+	printbuf_nul_terminate_reserved(out);
 }
 
 /* Doesn't call bch2_printbuf_make_room(), doesn't nul terminate: */
 static inline void __prt_char_reserved(struct printbuf *out, char c)
 {
 	if (printbuf_remaining(out))
-		out->buf[out->pos] = c;
-	out->pos++;
+		out->buf[out->pos++] = c;
 }
 
 /* Doesn't nul terminate: */
@@ -194,37 +192,34 @@ static inline void __prt_char(struct printbuf *out, char c)
 
 static inline void prt_char(struct printbuf *out, char c)
 {
-	__prt_char(out, c);
-	printbuf_nul_terminate(out);
+	bch2_printbuf_make_room(out, 2);
+	__prt_char_reserved(out, c);
+	printbuf_nul_terminate_reserved(out);
 }
 
 static inline void __prt_chars_reserved(struct printbuf *out, char c, unsigned n)
 {
-	unsigned i, can_print = min(n, printbuf_remaining(out));
+	unsigned can_print = min(n, printbuf_remaining(out));
 
-	for (i = 0; i < can_print; i++)
+	for (unsigned i = 0; i < can_print; i++)
 		out->buf[out->pos++] = c;
-	out->pos += n - can_print;
 }
 
 static inline void prt_chars(struct printbuf *out, char c, unsigned n)
 {
 	bch2_printbuf_make_room(out, n);
 	__prt_chars_reserved(out, c, n);
-	printbuf_nul_terminate(out);
+	printbuf_nul_terminate_reserved(out);
 }
 
 static inline void prt_bytes(struct printbuf *out, const void *b, unsigned n)
 {
-	unsigned i, can_print;
-
 	bch2_printbuf_make_room(out, n);
 
-	can_print = min(n, printbuf_remaining(out));
+	unsigned can_print = min(n, printbuf_remaining(out));
 
-	for (i = 0; i < can_print; i++)
+	for (unsigned i = 0; i < can_print; i++)
 		out->buf[out->pos++] = ((char *) b)[i];
-	out->pos += n - can_print;
 
 	printbuf_nul_terminate(out);
 }
@@ -241,18 +236,18 @@ static inline void prt_str_indented(struct printbuf *out, const char *str)
 
 static inline void prt_hex_byte(struct printbuf *out, u8 byte)
 {
-	bch2_printbuf_make_room(out, 2);
+	bch2_printbuf_make_room(out, 3);
 	__prt_char_reserved(out, hex_asc_hi(byte));
 	__prt_char_reserved(out, hex_asc_lo(byte));
-	printbuf_nul_terminate(out);
+	printbuf_nul_terminate_reserved(out);
 }
 
 static inline void prt_hex_byte_upper(struct printbuf *out, u8 byte)
 {
-	bch2_printbuf_make_room(out, 2);
+	bch2_printbuf_make_room(out, 3);
 	__prt_char_reserved(out, hex_asc_upper_hi(byte));
 	__prt_char_reserved(out, hex_asc_upper_lo(byte));
-	printbuf_nul_terminate(out);
+	printbuf_nul_terminate_reserved(out);
 }
 
 /**
diff --git a/fs/bcachefs/quota.c b/fs/bcachefs/quota.c
index e68b34eab9..a0cca8b70e 100644
--- a/fs/bcachefs/quota.c
+++ b/fs/bcachefs/quota.c
@@ -20,7 +20,7 @@ static const char * const bch2_quota_counters[] = {
 };
 
 static int bch2_sb_quota_validate(struct bch_sb *sb, struct bch_sb_field *f,
-				  struct printbuf *err)
+				  enum bch_validate_flags flags, struct printbuf *err)
 {
 	struct bch_sb_field_quota *q = field_to_type(f, quota);
 
@@ -60,8 +60,7 @@ const struct bch_sb_field_ops bch_sb_field_ops_quota = {
 };
 
 int bch2_quota_invalid(struct bch_fs *c, struct bkey_s_c k,
-		       enum bkey_invalid_flags flags,
-		       struct printbuf *err)
+		       enum bch_validate_flags flags, struct printbuf *err)
 {
 	int ret = 0;
 
@@ -97,45 +96,14 @@ static void qc_info_to_text(struct printbuf *out, struct qc_info *i)
 	printbuf_tabstops_reset(out);
 	printbuf_tabstop_push(out, 20);
 
-	prt_str(out, "i_fieldmask");
-	prt_tab(out);
-	prt_printf(out, "%x", i->i_fieldmask);
-	prt_newline(out);
-
-	prt_str(out, "i_flags");
-	prt_tab(out);
-	prt_printf(out, "%u", i->i_flags);
-	prt_newline(out);
-
-	prt_str(out, "i_spc_timelimit");
-	prt_tab(out);
-	prt_printf(out, "%u", i->i_spc_timelimit);
-	prt_newline(out);
-
-	prt_str(out, "i_ino_timelimit");
-	prt_tab(out);
-	prt_printf(out, "%u", i->i_ino_timelimit);
-	prt_newline(out);
-
-	prt_str(out, "i_rt_spc_timelimit");
-	prt_tab(out);
-	prt_printf(out, "%u", i->i_rt_spc_timelimit);
-	prt_newline(out);
-
-	prt_str(out, "i_spc_warnlimit");
-	prt_tab(out);
-	prt_printf(out, "%u", i->i_spc_warnlimit);
-	prt_newline(out);
-
-	prt_str(out, "i_ino_warnlimit");
-	prt_tab(out);
-	prt_printf(out, "%u", i->i_ino_warnlimit);
-	prt_newline(out);
-
-	prt_str(out, "i_rt_spc_warnlimit");
-	prt_tab(out);
-	prt_printf(out, "%u", i->i_rt_spc_warnlimit);
-	prt_newline(out);
+	prt_printf(out, "i_fieldmask\t%x\n",		i->i_fieldmask);
+	prt_printf(out, "i_flags\t%u\n",		i->i_flags);
+	prt_printf(out, "i_spc_timelimit\t%u\n",	i->i_spc_timelimit);
+	prt_printf(out, "i_ino_timelimit\t%u\n",	i->i_ino_timelimit);
+	prt_printf(out, "i_rt_spc_timelimit\t%u\n",	i->i_rt_spc_timelimit);
+	prt_printf(out, "i_spc_warnlimit\t%u\n",	i->i_spc_warnlimit);
+	prt_printf(out, "i_ino_warnlimit\t%u\n",	i->i_ino_warnlimit);
+	prt_printf(out, "i_rt_spc_warnlimit\t%u\n",	i->i_rt_spc_warnlimit);
 }
 
 static void qc_dqblk_to_text(struct printbuf *out, struct qc_dqblk *q)
@@ -143,60 +111,17 @@ static void qc_dqblk_to_text(struct printbuf *out, struct qc_dqblk *q)
 	printbuf_tabstops_reset(out);
 	printbuf_tabstop_push(out, 20);
 
-	prt_str(out, "d_fieldmask");
-	prt_tab(out);
-	prt_printf(out, "%x", q->d_fieldmask);
-	prt_newline(out);
-
-	prt_str(out, "d_spc_hardlimit");
-	prt_tab(out);
-	prt_printf(out, "%llu", q->d_spc_hardlimit);
-	prt_newline(out);
-
-	prt_str(out, "d_spc_softlimit");
-	prt_tab(out);
-	prt_printf(out, "%llu", q->d_spc_softlimit);
-	prt_newline(out);
-
-	prt_str(out, "d_ino_hardlimit");
-	prt_tab(out);
-	prt_printf(out, "%llu", q->d_ino_hardlimit);
-	prt_newline(out);
-
-	prt_str(out, "d_ino_softlimit");
-	prt_tab(out);
-	prt_printf(out, "%llu", q->d_ino_softlimit);
-	prt_newline(out);
-
-	prt_str(out, "d_space");
-	prt_tab(out);
-	prt_printf(out, "%llu", q->d_space);
-	prt_newline(out);
-
-	prt_str(out, "d_ino_count");
-	prt_tab(out);
-	prt_printf(out, "%llu", q->d_ino_count);
-	prt_newline(out);
-
-	prt_str(out, "d_ino_timer");
-	prt_tab(out);
-	prt_printf(out, "%llu", q->d_ino_timer);
-	prt_newline(out);
-
-	prt_str(out, "d_spc_timer");
-	prt_tab(out);
-	prt_printf(out, "%llu", q->d_spc_timer);
-	prt_newline(out);
-
-	prt_str(out, "d_ino_warns");
-	prt_tab(out);
-	prt_printf(out, "%i", q->d_ino_warns);
-	prt_newline(out);
-
-	prt_str(out, "d_spc_warns");
-	prt_tab(out);
-	prt_printf(out, "%i", q->d_spc_warns);
-	prt_newline(out);
+	prt_printf(out, "d_fieldmask\t%x\n",		q->d_fieldmask);
+	prt_printf(out, "d_spc_hardlimit\t%llu\n",	q->d_spc_hardlimit);
+	prt_printf(out, "d_spc_softlimit\t%llu\n",	q->d_spc_softlimit);
+	prt_printf(out, "d_ino_hardlimit\%llu\n",	q->d_ino_hardlimit);
+	prt_printf(out, "d_ino_softlimit\t%llu\n",	q->d_ino_softlimit);
+	prt_printf(out, "d_space\t%llu\n",		q->d_space);
+	prt_printf(out, "d_ino_count\t%llu\n",		q->d_ino_count);
+	prt_printf(out, "d_ino_timer\t%llu\n",		q->d_ino_timer);
+	prt_printf(out, "d_spc_timer\t%llu\n",		q->d_spc_timer);
+	prt_printf(out, "d_ino_warns\t%i\n",		q->d_ino_warns);
+	prt_printf(out, "d_spc_warns\t%i\n",		q->d_spc_warns);
 }
 
 static inline unsigned __next_qtype(unsigned i, unsigned qtypes)
@@ -560,13 +485,11 @@ static int bch2_fs_quota_read_inode(struct btree_trans *trans,
 	struct bch_fs *c = trans->c;
 	struct bch_inode_unpacked u;
 	struct bch_snapshot_tree s_t;
-	int ret;
+	u32 tree = bch2_snapshot_tree(c, k.k->p.snapshot);
 
-	ret = bch2_snapshot_tree_lookup(trans,
-			bch2_snapshot_tree(c, k.k->p.snapshot), &s_t);
+	int ret = bch2_snapshot_tree_lookup(trans, tree, &s_t);
 	bch2_fs_inconsistent_on(bch2_err_matches(ret, ENOENT), c,
-			"%s: snapshot tree %u not found", __func__,
-			snapshot_t(c, k.k->p.snapshot)->tree);
+			"%s: snapshot tree %u not found", __func__, tree);
 	if (ret)
 		return ret;
 
@@ -612,10 +535,10 @@ int bch2_fs_quota_read(struct bch_fs *c)
 
 	int ret = bch2_trans_run(c,
 		for_each_btree_key(trans, iter, BTREE_ID_quotas, POS_MIN,
-				   BTREE_ITER_PREFETCH, k,
+				   BTREE_ITER_prefetch, k,
 			__bch2_quota_set(c, k, NULL)) ?:
 		for_each_btree_key(trans, iter, BTREE_ID_inodes, POS_MIN,
-				   BTREE_ITER_PREFETCH|BTREE_ITER_ALL_SNAPSHOTS, k,
+				   BTREE_ITER_prefetch|BTREE_ITER_all_snapshots, k,
 			bch2_fs_quota_read_inode(trans, &iter, k)));
 	bch_err_fn(c, ret);
 	return ret;
@@ -902,7 +825,7 @@ static int bch2_set_quota_trans(struct btree_trans *trans,
 	int ret;
 
 	k = bch2_bkey_get_iter(trans, &iter, BTREE_ID_quotas, new_quota->k.p,
-			       BTREE_ITER_SLOTS|BTREE_ITER_INTENT);
+			       BTREE_ITER_slots|BTREE_ITER_intent);
 	ret = bkey_err(k);
 	if (unlikely(ret))
 		return ret;
diff --git a/fs/bcachefs/quota.h b/fs/bcachefs/quota.h
index 884f601f41..02d37a3322 100644
--- a/fs/bcachefs/quota.h
+++ b/fs/bcachefs/quota.h
@@ -5,11 +5,11 @@
 #include "inode.h"
 #include "quota_types.h"
 
-enum bkey_invalid_flags;
+enum bch_validate_flags;
 extern const struct bch_sb_field_ops bch_sb_field_ops_quota;
 
 int bch2_quota_invalid(struct bch_fs *, struct bkey_s_c,
-		       enum bkey_invalid_flags, struct printbuf *);
+		       enum bch_validate_flags, struct printbuf *);
 void bch2_quota_to_text(struct printbuf *, struct bch_fs *, struct bkey_s_c);
 
 #define bch2_bkey_ops_quota ((struct bkey_ops) {	\
diff --git a/fs/bcachefs/rebalance.c b/fs/bcachefs/rebalance.c
index 22d1017aa4..cf81e5128c 100644
--- a/fs/bcachefs/rebalance.c
+++ b/fs/bcachefs/rebalance.c
@@ -42,7 +42,7 @@ static int __bch2_set_rebalance_needs_scan(struct btree_trans *trans, u64 inum)
 
 	bch2_trans_iter_init(trans, &iter, BTREE_ID_rebalance_work,
 			     SPOS(inum, REBALANCE_WORK_SCAN_OFFSET, U32_MAX),
-			     BTREE_ITER_INTENT);
+			     BTREE_ITER_intent);
 	k = bch2_btree_iter_peek_slot(&iter);
 	ret = bkey_err(k);
 	if (ret)
@@ -89,7 +89,7 @@ static int bch2_clear_rebalance_needs_scan(struct btree_trans *trans, u64 inum,
 
 	bch2_trans_iter_init(trans, &iter, BTREE_ID_rebalance_work,
 			     SPOS(inum, REBALANCE_WORK_SCAN_OFFSET, U32_MAX),
-			     BTREE_ITER_INTENT);
+			     BTREE_ITER_intent);
 	k = bch2_btree_iter_peek_slot(&iter);
 	ret = bkey_err(k);
 	if (ret)
@@ -140,7 +140,7 @@ static struct bkey_s_c next_rebalance_extent(struct btree_trans *trans,
 	bch2_trans_iter_init(trans, extent_iter,
 			     work_pos.inode ? BTREE_ID_extents : BTREE_ID_reflink,
 			     work_pos,
-			     BTREE_ITER_ALL_SNAPSHOTS);
+			     BTREE_ITER_all_snapshots);
 	k = bch2_btree_iter_peek_slot(extent_iter);
 	if (bkey_err(k))
 		return k;
@@ -323,12 +323,14 @@ static int do_rebalance(struct moving_context *ctxt)
 	struct bkey_s_c k;
 	int ret = 0;
 
+	bch2_trans_begin(trans);
+
 	bch2_move_stats_init(&r->work_stats, "rebalance_work");
 	bch2_move_stats_init(&r->scan_stats, "rebalance_scan");
 
 	bch2_trans_iter_init(trans, &rebalance_work_iter,
 			     BTREE_ID_rebalance_work, POS_MIN,
-			     BTREE_ITER_ALL_SNAPSHOTS);
+			     BTREE_ITER_all_snapshots);
 
 	while (!bch2_move_ratelimit(ctxt)) {
 		if (!r->enabled) {
@@ -412,11 +414,11 @@ void bch2_rebalance_status_to_text(struct printbuf *out, struct bch_fs *c)
 		u64 now = atomic64_read(&c->io_clock[WRITE].now);
 
 		prt_str(out, "io wait duration:  ");
-		bch2_prt_human_readable_s64(out, r->wait_iotime_end - r->wait_iotime_start);
+		bch2_prt_human_readable_s64(out, (r->wait_iotime_end - r->wait_iotime_start) << 9);
 		prt_newline(out);
 
 		prt_str(out, "io wait remaining: ");
-		bch2_prt_human_readable_s64(out, r->wait_iotime_end - now);
+		bch2_prt_human_readable_s64(out, (r->wait_iotime_end - now) << 9);
 		prt_newline(out);
 
 		prt_str(out, "duration waited:   ");
diff --git a/fs/bcachefs/recovery.c b/fs/bcachefs/recovery.c
index 21e13bb433..1f9d044ed9 100644
--- a/fs/bcachefs/recovery.c
+++ b/fs/bcachefs/recovery.c
@@ -1,35 +1,31 @@
 // SPDX-License-Identifier: GPL-2.0
 
 #include "bcachefs.h"
-#include "backpointers.h"
-#include "bkey_buf.h"
 #include "alloc_background.h"
-#include "btree_gc.h"
+#include "bkey_buf.h"
 #include "btree_journal_iter.h"
+#include "btree_node_scan.h"
 #include "btree_update.h"
 #include "btree_update_interior.h"
 #include "btree_io.h"
 #include "buckets.h"
 #include "dirent.h"
-#include "ec.h"
 #include "errcode.h"
 #include "error.h"
 #include "fs-common.h"
-#include "fsck.h"
 #include "journal_io.h"
 #include "journal_reclaim.h"
 #include "journal_seq_blacklist.h"
-#include "lru.h"
 #include "logged_ops.h"
 #include "move.h"
 #include "quota.h"
 #include "rebalance.h"
 #include "recovery.h"
+#include "recovery_passes.h"
 #include "replicas.h"
 #include "sb-clean.h"
 #include "sb-downgrade.h"
 #include "snapshot.h"
-#include "subvolume.h"
 #include "super-io.h"
 
 #include <linux/sort.h>
@@ -37,30 +33,81 @@
 
 #define QSTR(n) { { { .len = strlen(n) } }, .name = n }
 
-static bool btree_id_is_alloc(enum btree_id id)
+void bch2_btree_lost_data(struct bch_fs *c, enum btree_id btree)
 {
-	switch (id) {
-	case BTREE_ID_alloc:
-	case BTREE_ID_backpointers:
-	case BTREE_ID_need_discard:
-	case BTREE_ID_freespace:
-	case BTREE_ID_bucket_gens:
-		return true;
-	default:
-		return false;
+	if (btree >= BTREE_ID_NR_MAX)
+		return;
+
+	u64 b = BIT_ULL(btree);
+
+	if (!(c->sb.btrees_lost_data & b)) {
+		bch_err(c, "flagging btree %s lost data", bch2_btree_id_str(btree));
+
+		mutex_lock(&c->sb_lock);
+		bch2_sb_field_get(c->disk_sb.sb, ext)->btrees_lost_data |= cpu_to_le64(b);
+		bch2_write_super(c);
+		mutex_unlock(&c->sb_lock);
 	}
 }
 
 /* for -o reconstruct_alloc: */
-static void drop_alloc_keys(struct journal_keys *keys)
+static void bch2_reconstruct_alloc(struct bch_fs *c)
 {
-	size_t src, dst;
+	bch2_journal_log_msg(c, "dropping alloc info");
+	bch_info(c, "dropping and reconstructing all alloc info");
+
+	mutex_lock(&c->sb_lock);
+	struct bch_sb_field_ext *ext = bch2_sb_field_get(c->disk_sb.sb, ext);
+
+	__set_bit_le64(BCH_RECOVERY_PASS_STABLE_check_allocations, ext->recovery_passes_required);
+	__set_bit_le64(BCH_RECOVERY_PASS_STABLE_check_alloc_info, ext->recovery_passes_required);
+	__set_bit_le64(BCH_RECOVERY_PASS_STABLE_check_lrus, ext->recovery_passes_required);
+	__set_bit_le64(BCH_RECOVERY_PASS_STABLE_check_extents_to_backpointers, ext->recovery_passes_required);
+	__set_bit_le64(BCH_RECOVERY_PASS_STABLE_check_alloc_to_lru_refs, ext->recovery_passes_required);
+
+	__set_bit_le64(BCH_FSCK_ERR_ptr_to_missing_alloc_key, ext->errors_silent);
+	__set_bit_le64(BCH_FSCK_ERR_ptr_gen_newer_than_bucket_gen, ext->errors_silent);
+	__set_bit_le64(BCH_FSCK_ERR_stale_dirty_ptr, ext->errors_silent);
+
+	__set_bit_le64(BCH_FSCK_ERR_dev_usage_buckets_wrong, ext->errors_silent);
+	__set_bit_le64(BCH_FSCK_ERR_dev_usage_sectors_wrong, ext->errors_silent);
+	__set_bit_le64(BCH_FSCK_ERR_dev_usage_fragmented_wrong, ext->errors_silent);
+
+	__set_bit_le64(BCH_FSCK_ERR_fs_usage_btree_wrong, ext->errors_silent);
+	__set_bit_le64(BCH_FSCK_ERR_fs_usage_cached_wrong, ext->errors_silent);
+	__set_bit_le64(BCH_FSCK_ERR_fs_usage_persistent_reserved_wrong, ext->errors_silent);
+	__set_bit_le64(BCH_FSCK_ERR_fs_usage_replicas_wrong, ext->errors_silent);
+
+	__set_bit_le64(BCH_FSCK_ERR_alloc_key_data_type_wrong, ext->errors_silent);
+	__set_bit_le64(BCH_FSCK_ERR_alloc_key_gen_wrong, ext->errors_silent);
+	__set_bit_le64(BCH_FSCK_ERR_alloc_key_dirty_sectors_wrong, ext->errors_silent);
+	__set_bit_le64(BCH_FSCK_ERR_alloc_key_cached_sectors_wrong, ext->errors_silent);
+	__set_bit_le64(BCH_FSCK_ERR_alloc_key_stripe_wrong, ext->errors_silent);
+	__set_bit_le64(BCH_FSCK_ERR_alloc_key_stripe_redundancy_wrong, ext->errors_silent);
+	__set_bit_le64(BCH_FSCK_ERR_need_discard_key_wrong, ext->errors_silent);
+	__set_bit_le64(BCH_FSCK_ERR_freespace_key_wrong, ext->errors_silent);
+	__set_bit_le64(BCH_FSCK_ERR_bucket_gens_key_wrong, ext->errors_silent);
+	__set_bit_le64(BCH_FSCK_ERR_freespace_hole_missing, ext->errors_silent);
+	__set_bit_le64(BCH_FSCK_ERR_ptr_to_missing_backpointer, ext->errors_silent);
+	__set_bit_le64(BCH_FSCK_ERR_lru_entry_bad, ext->errors_silent);
+	c->sb.compat &= ~(1ULL << BCH_COMPAT_alloc_info);
 
-	for (src = 0, dst = 0; src < keys->nr; src++)
-		if (!btree_id_is_alloc(keys->d[src].btree_id))
-			keys->d[dst++] = keys->d[src];
+	bch2_write_super(c);
+	mutex_unlock(&c->sb_lock);
 
-	keys->nr = dst;
+	c->recovery_passes_explicit |= bch2_recovery_passes_from_stable(le64_to_cpu(ext->recovery_passes_required[0]));
+
+
+	bch2_shoot_down_journal_keys(c, BTREE_ID_alloc,
+				     0, BTREE_MAX_DEPTH, POS_MIN, SPOS_MAX);
+	bch2_shoot_down_journal_keys(c, BTREE_ID_backpointers,
+				     0, BTREE_MAX_DEPTH, POS_MIN, SPOS_MAX);
+	bch2_shoot_down_journal_keys(c, BTREE_ID_need_discard,
+				     0, BTREE_MAX_DEPTH, POS_MIN, SPOS_MAX);
+	bch2_shoot_down_journal_keys(c, BTREE_ID_freespace,
+				     0, BTREE_MAX_DEPTH, POS_MIN, SPOS_MAX);
+	bch2_shoot_down_journal_keys(c, BTREE_ID_bucket_gens,
+				     0, BTREE_MAX_DEPTH, POS_MIN, SPOS_MAX);
 }
 
 /*
@@ -70,9 +117,7 @@ static void drop_alloc_keys(struct journal_keys *keys)
  */
 static void zero_out_btree_mem_ptr(struct journal_keys *keys)
 {
-	struct journal_key *i;
-
-	for (i = keys->d; i < keys->d + keys->nr; i++)
+	darray_for_each(*keys, i)
 		if (i->k->k.type == KEY_TYPE_btree_ptr_v2)
 			bkey_i_to_btree_ptr_v2(i->k)->v.mem_ptr = 0;
 }
@@ -94,9 +139,9 @@ static int bch2_journal_replay_key(struct btree_trans *trans,
 {
 	struct btree_iter iter;
 	unsigned iter_flags =
-		BTREE_ITER_INTENT|
-		BTREE_ITER_NOT_EXTENTS;
-	unsigned update_flags = BTREE_TRIGGER_NORUN;
+		BTREE_ITER_intent|
+		BTREE_ITER_not_extents;
+	unsigned update_flags = BTREE_TRIGGER_norun;
 	int ret;
 
 	if (k->overwritten)
@@ -105,17 +150,17 @@ static int bch2_journal_replay_key(struct btree_trans *trans,
 	trans->journal_res.seq = k->journal_seq;
 
 	/*
-	 * BTREE_UPDATE_KEY_CACHE_RECLAIM disables key cache lookup/update to
+	 * BTREE_UPDATE_key_cache_reclaim disables key cache lookup/update to
 	 * keep the key cache coherent with the underlying btree. Nothing
 	 * besides the allocator is doing updates yet so we don't need key cache
 	 * coherency for non-alloc btrees, and key cache fills for snapshots
-	 * btrees use BTREE_ITER_FILTER_SNAPSHOTS, which isn't available until
+	 * btrees use BTREE_ITER_filter_snapshots, which isn't available until
 	 * the snapshots recovery pass runs.
 	 */
 	if (!k->level && k->btree_id == BTREE_ID_alloc)
-		iter_flags |= BTREE_ITER_CACHED;
+		iter_flags |= BTREE_ITER_cached;
 	else
-		update_flags |= BTREE_UPDATE_KEY_CACHE_RECLAIM;
+		update_flags |= BTREE_UPDATE_key_cache_reclaim;
 
 	bch2_trans_node_iter_init(trans, &iter, k->btree_id, k->k->k.p,
 				  BTREE_MAX_DEPTH, k->level,
@@ -124,6 +169,17 @@ static int bch2_journal_replay_key(struct btree_trans *trans,
 	if (ret)
 		goto out;
 
+	struct btree_path *path = btree_iter_path(trans, &iter);
+	if (unlikely(!btree_path_node(path, k->level))) {
+		bch2_trans_iter_exit(trans, &iter);
+		bch2_trans_node_iter_init(trans, &iter, k->btree_id, k->k->k.p,
+					  BTREE_MAX_DEPTH, 0, iter_flags);
+		ret =   bch2_btree_iter_traverse(&iter) ?:
+			bch2_btree_increase_depth(trans, iter.path, 0) ?:
+			-BCH_ERR_transaction_restart_nested;
+		goto out;
+	}
+
 	/* Must be checked with btree locked: */
 	if (k->overwritten)
 		goto out;
@@ -142,14 +198,15 @@ static int journal_sort_seq_cmp(const void *_l, const void *_r)
 	return cmp_int(l->journal_seq, r->journal_seq);
 }
 
-static int bch2_journal_replay(struct bch_fs *c)
+int bch2_journal_replay(struct bch_fs *c)
 {
 	struct journal_keys *keys = &c->journal_keys;
 	DARRAY(struct journal_key *) keys_sorted = { 0 };
 	struct journal *j = &c->journal;
 	u64 start_seq	= c->journal_replay_seq_start;
 	u64 end_seq	= c->journal_replay_seq_start;
-	struct btree_trans *trans = bch2_trans_get(c);
+	struct btree_trans *trans = NULL;
+	bool immediate_flush = false;
 	int ret = 0;
 
 	if (keys->nr) {
@@ -161,15 +218,23 @@ static int bch2_journal_replay(struct bch_fs *c)
 
 	BUG_ON(!atomic_read(&keys->ref));
 
+	move_gap(keys, keys->nr);
+	trans = bch2_trans_get(c);
+
 	/*
 	 * First, attempt to replay keys in sorted order. This is more
 	 * efficient - better locality of btree access -  but some might fail if
 	 * that would cause a journal deadlock.
 	 */
-	for (size_t i = 0; i < keys->nr; i++) {
+	darray_for_each(*keys, k) {
 		cond_resched();
 
-		struct journal_key *k = keys->d + i;
+		/*
+		 * k->allocated means the key wasn't read in from the journal,
+		 * rather it was from early repair code
+		 */
+		if (k->allocated)
+			immediate_flush = true;
 
 		/* Skip fastpath if we're low on space in the journal */
 		ret = c->journal.watermark ? -1 :
@@ -199,7 +264,10 @@ static int bch2_journal_replay(struct bch_fs *c)
 
 		struct journal_key *k = *kp;
 
-		replay_now_at(j, k->journal_seq);
+		if (k->journal_seq)
+			replay_now_at(j, k->journal_seq);
+		else
+			replay_now_at(j, j->replay_journal_seq_end);
 
 		ret = commit_do(trans, NULL, NULL,
 				BCH_TRANS_COMMIT_no_enospc|
@@ -222,7 +290,8 @@ static int bch2_journal_replay(struct bch_fs *c)
 	bch2_trans_put(trans);
 	trans = NULL;
 
-	if (!c->opts.keep_journal)
+	if (!c->opts.retain_recovery_info &&
+	    c->recovery_pass_done >= BCH_RECOVERY_PASS_journal_replay)
 		bch2_journal_keys_put_initial(c);
 
 	replay_now_at(j, j->replay_journal_seq_end);
@@ -230,6 +299,12 @@ static int bch2_journal_replay(struct bch_fs *c)
 
 	bch2_journal_set_replay_done(j);
 
+	/* if we did any repair, flush it immediately */
+	if (immediate_flush) {
+		bch2_journal_flush_all_pins(&c->journal);
+		ret = bch2_journal_meta(&c->journal);
+	}
+
 	if (keys->nr)
 		bch2_journal_log_msg(c, "journal replay finished");
 err:
@@ -251,6 +326,12 @@ static int journal_replay_entry_early(struct bch_fs *c,
 	case BCH_JSET_ENTRY_btree_root: {
 		struct btree_root *r;
 
+		if (fsck_err_on(entry->btree_id >= BTREE_ID_NR_MAX,
+				c, invalid_btree_id,
+				"invalid btree id %u (max %u)",
+				entry->btree_id, BTREE_ID_NR_MAX))
+			return 0;
+
 		while (entry->btree_id >= c->btree_roots_extra.nr + BTREE_ID_NR) {
 			ret = darray_push(&c->btree_roots_extra, (struct btree_root) { NULL });
 			if (ret)
@@ -264,7 +345,7 @@ static int journal_replay_entry_early(struct bch_fs *c,
 			bkey_copy(&r->key, (struct bkey_i *) entry->start);
 			r->error = 0;
 		} else {
-			r->error = -EIO;
+			r->error = -BCH_ERR_btree_node_read_error;
 		}
 		r->alive = true;
 		break;
@@ -301,14 +382,17 @@ static int journal_replay_entry_early(struct bch_fs *c,
 	case BCH_JSET_ENTRY_dev_usage: {
 		struct jset_entry_dev_usage *u =
 			container_of(entry, struct jset_entry_dev_usage, entry);
-		struct bch_dev *ca = bch_dev_bkey_exists(c, le32_to_cpu(u->dev));
-		unsigned i, nr_types = jset_entry_dev_usage_nr_types(u);
-
-		for (i = 0; i < min_t(unsigned, nr_types, BCH_DATA_NR); i++) {
-			ca->usage_base->d[i].buckets	= le64_to_cpu(u->d[i].buckets);
-			ca->usage_base->d[i].sectors	= le64_to_cpu(u->d[i].sectors);
-			ca->usage_base->d[i].fragmented	= le64_to_cpu(u->d[i].fragmented);
-		}
+		unsigned nr_types = jset_entry_dev_usage_nr_types(u);
+
+		rcu_read_lock();
+		struct bch_dev *ca = bch2_dev_rcu(c, le32_to_cpu(u->dev));
+		if (ca)
+			for (unsigned i = 0; i < min_t(unsigned, nr_types, BCH_DATA_NR); i++) {
+				ca->usage_base->d[i].buckets	= le64_to_cpu(u->d[i].buckets);
+				ca->usage_base->d[i].sectors	= le64_to_cpu(u->d[i].sectors);
+				ca->usage_base->d[i].fragmented	= le64_to_cpu(u->d[i].fragmented);
+			}
+		rcu_read_unlock();
 
 		break;
 	}
@@ -337,7 +421,7 @@ static int journal_replay_entry_early(struct bch_fs *c,
 		atomic64_set(&c->io_clock[clock->rw].now, le64_to_cpu(clock->time));
 	}
 	}
-
+fsck_err:
 	return ret;
 }
 
@@ -359,7 +443,7 @@ static int journal_replay_early(struct bch_fs *c,
 		genradix_for_each(&c->journal_entries, iter, _i) {
 			i = *_i;
 
-			if (!i || i->ignore)
+			if (journal_replay_ignore(i))
 				continue;
 
 			vstruct_for_each(&i->j, entry) {
@@ -379,202 +463,57 @@ static int journal_replay_early(struct bch_fs *c,
 
 static int read_btree_roots(struct bch_fs *c)
 {
-	unsigned i;
 	int ret = 0;
 
-	for (i = 0; i < btree_id_nr_alive(c); i++) {
+	for (unsigned i = 0; i < btree_id_nr_alive(c); i++) {
 		struct btree_root *r = bch2_btree_id_root(c, i);
 
 		if (!r->alive)
 			continue;
 
-		if (btree_id_is_alloc(i) &&
-		    c->opts.reconstruct_alloc) {
-			c->sb.compat &= ~(1ULL << BCH_COMPAT_alloc_info);
+		if (btree_id_is_alloc(i) && c->opts.reconstruct_alloc)
 			continue;
-		}
 
-		if (r->error) {
-			__fsck_err(c,
-				   btree_id_is_alloc(i)
-				   ? FSCK_CAN_IGNORE : 0,
-				   btree_root_bkey_invalid,
-				   "invalid btree root %s",
-				   bch2_btree_id_str(i));
-			if (i == BTREE_ID_alloc)
+		if (mustfix_fsck_err_on((ret = r->error),
+					c, btree_root_bkey_invalid,
+					"invalid btree root %s",
+					bch2_btree_id_str(i)) ||
+		    mustfix_fsck_err_on((ret = r->error = bch2_btree_root_read(c, i, &r->key, r->level)),
+					c, btree_root_read_error,
+					"error reading btree root %s l=%u: %s",
+					bch2_btree_id_str(i), r->level, bch2_err_str(ret))) {
+			if (btree_id_is_alloc(i)) {
+				c->recovery_passes_explicit |= BIT_ULL(BCH_RECOVERY_PASS_check_allocations);
+				c->recovery_passes_explicit |= BIT_ULL(BCH_RECOVERY_PASS_check_alloc_info);
+				c->recovery_passes_explicit |= BIT_ULL(BCH_RECOVERY_PASS_check_lrus);
+				c->recovery_passes_explicit |= BIT_ULL(BCH_RECOVERY_PASS_check_extents_to_backpointers);
+				c->recovery_passes_explicit |= BIT_ULL(BCH_RECOVERY_PASS_check_alloc_to_lru_refs);
 				c->sb.compat &= ~(1ULL << BCH_COMPAT_alloc_info);
-		}
+				r->error = 0;
+			} else if (!(c->recovery_passes_explicit & BIT_ULL(BCH_RECOVERY_PASS_scan_for_btree_nodes))) {
+				bch_info(c, "will run btree node scan");
+				c->recovery_passes_explicit |= BIT_ULL(BCH_RECOVERY_PASS_scan_for_btree_nodes);
+				c->recovery_passes_explicit |= BIT_ULL(BCH_RECOVERY_PASS_check_topology);
+			}
 
-		ret = bch2_btree_root_read(c, i, &r->key, r->level);
-		if (ret) {
-			fsck_err(c,
-				 btree_root_read_error,
-				 "error reading btree root %s",
-				 bch2_btree_id_str(i));
-			if (btree_id_is_alloc(i))
-				c->sb.compat &= ~(1ULL << BCH_COMPAT_alloc_info);
 			ret = 0;
+			bch2_btree_lost_data(c, i);
 		}
 	}
 
-	for (i = 0; i < BTREE_ID_NR; i++) {
+	for (unsigned i = 0; i < BTREE_ID_NR; i++) {
 		struct btree_root *r = bch2_btree_id_root(c, i);
 
-		if (!r->b) {
+		if (!r->b && !r->error) {
 			r->alive = false;
 			r->level = 0;
-			bch2_btree_root_alloc(c, i);
+			bch2_btree_root_alloc_fake(c, i, 0);
 		}
 	}
 fsck_err:
 	return ret;
 }
 
-static int bch2_initialize_subvolumes(struct bch_fs *c)
-{
-	struct bkey_i_snapshot_tree	root_tree;
-	struct bkey_i_snapshot		root_snapshot;
-	struct bkey_i_subvolume		root_volume;
-	int ret;
-
-	bkey_snapshot_tree_init(&root_tree.k_i);
-	root_tree.k.p.offset		= 1;
-	root_tree.v.master_subvol	= cpu_to_le32(1);
-	root_tree.v.root_snapshot	= cpu_to_le32(U32_MAX);
-
-	bkey_snapshot_init(&root_snapshot.k_i);
-	root_snapshot.k.p.offset = U32_MAX;
-	root_snapshot.v.flags	= 0;
-	root_snapshot.v.parent	= 0;
-	root_snapshot.v.subvol	= cpu_to_le32(BCACHEFS_ROOT_SUBVOL);
-	root_snapshot.v.tree	= cpu_to_le32(1);
-	SET_BCH_SNAPSHOT_SUBVOL(&root_snapshot.v, true);
-
-	bkey_subvolume_init(&root_volume.k_i);
-	root_volume.k.p.offset = BCACHEFS_ROOT_SUBVOL;
-	root_volume.v.flags	= 0;
-	root_volume.v.snapshot	= cpu_to_le32(U32_MAX);
-	root_volume.v.inode	= cpu_to_le64(BCACHEFS_ROOT_INO);
-
-	ret =   bch2_btree_insert(c, BTREE_ID_snapshot_trees,	&root_tree.k_i, NULL, 0) ?:
-		bch2_btree_insert(c, BTREE_ID_snapshots,	&root_snapshot.k_i, NULL, 0) ?:
-		bch2_btree_insert(c, BTREE_ID_subvolumes,	&root_volume.k_i, NULL, 0);
-	bch_err_fn(c, ret);
-	return ret;
-}
-
-static int __bch2_fs_upgrade_for_subvolumes(struct btree_trans *trans)
-{
-	struct btree_iter iter;
-	struct bkey_s_c k;
-	struct bch_inode_unpacked inode;
-	int ret;
-
-	k = bch2_bkey_get_iter(trans, &iter, BTREE_ID_inodes,
-			       SPOS(0, BCACHEFS_ROOT_INO, U32_MAX), 0);
-	ret = bkey_err(k);
-	if (ret)
-		return ret;
-
-	if (!bkey_is_inode(k.k)) {
-		bch_err(trans->c, "root inode not found");
-		ret = -BCH_ERR_ENOENT_inode;
-		goto err;
-	}
-
-	ret = bch2_inode_unpack(k, &inode);
-	BUG_ON(ret);
-
-	inode.bi_subvol = BCACHEFS_ROOT_SUBVOL;
-
-	ret = bch2_inode_write(trans, &iter, &inode);
-err:
-	bch2_trans_iter_exit(trans, &iter);
-	return ret;
-}
-
-/* set bi_subvol on root inode */
-noinline_for_stack
-static int bch2_fs_upgrade_for_subvolumes(struct bch_fs *c)
-{
-	int ret = bch2_trans_do(c, NULL, NULL, BCH_TRANS_COMMIT_lazy_rw,
-				__bch2_fs_upgrade_for_subvolumes(trans));
-	bch_err_fn(c, ret);
-	return ret;
-}
-
-const char * const bch2_recovery_passes[] = {
-#define x(_fn, ...)	#_fn,
-	BCH_RECOVERY_PASSES()
-#undef x
-	NULL
-};
-
-static int bch2_check_allocations(struct bch_fs *c)
-{
-	return bch2_gc(c, true, c->opts.norecovery);
-}
-
-static int bch2_set_may_go_rw(struct bch_fs *c)
-{
-	struct journal_keys *keys = &c->journal_keys;
-
-	/*
-	 * After we go RW, the journal keys buffer can't be modified (except for
-	 * setting journal_key->overwritten: it will be accessed by multiple
-	 * threads
-	 */
-	move_gap(keys->d, keys->nr, keys->size, keys->gap, keys->nr);
-	keys->gap = keys->nr;
-
-	set_bit(BCH_FS_may_go_rw, &c->flags);
-
-	if (keys->nr || c->opts.fsck || !c->sb.clean)
-		return bch2_fs_read_write_early(c);
-	return 0;
-}
-
-struct recovery_pass_fn {
-	int		(*fn)(struct bch_fs *);
-	unsigned	when;
-};
-
-static struct recovery_pass_fn recovery_pass_fns[] = {
-#define x(_fn, _id, _when)	{ .fn = bch2_##_fn, .when = _when },
-	BCH_RECOVERY_PASSES()
-#undef x
-};
-
-u64 bch2_recovery_passes_to_stable(u64 v)
-{
-	static const u8 map[] = {
-#define x(n, id, ...)	[BCH_RECOVERY_PASS_##n] = BCH_RECOVERY_PASS_STABLE_##n,
-	BCH_RECOVERY_PASSES()
-#undef x
-	};
-
-	u64 ret = 0;
-	for (unsigned i = 0; i < ARRAY_SIZE(map); i++)
-		if (v & BIT_ULL(i))
-			ret |= BIT_ULL(map[i]);
-	return ret;
-}
-
-u64 bch2_recovery_passes_from_stable(u64 v)
-{
-	static const u8 map[] = {
-#define x(n, id, ...)	[BCH_RECOVERY_PASS_STABLE_##n] = BCH_RECOVERY_PASS_##n,
-	BCH_RECOVERY_PASSES()
-#undef x
-	};
-
-	u64 ret = 0;
-	for (unsigned i = 0; i < ARRAY_SIZE(map); i++)
-		if (v & BIT_ULL(i))
-			ret |= BIT_ULL(map[i]);
-	return ret;
-}
-
 static bool check_version_upgrade(struct bch_fs *c)
 {
 	unsigned latest_version	= bcachefs_metadata_version_current;
@@ -647,96 +586,6 @@ static bool check_version_upgrade(struct bch_fs *c)
 	return false;
 }
 
-u64 bch2_fsck_recovery_passes(void)
-{
-	u64 ret = 0;
-
-	for (unsigned i = 0; i < ARRAY_SIZE(recovery_pass_fns); i++)
-		if (recovery_pass_fns[i].when & PASS_FSCK)
-			ret |= BIT_ULL(i);
-	return ret;
-}
-
-static bool should_run_recovery_pass(struct bch_fs *c, enum bch_recovery_pass pass)
-{
-	struct recovery_pass_fn *p = recovery_pass_fns + pass;
-
-	if (c->opts.norecovery && pass > BCH_RECOVERY_PASS_snapshots_read)
-		return false;
-	if (c->recovery_passes_explicit & BIT_ULL(pass))
-		return true;
-	if ((p->when & PASS_FSCK) && c->opts.fsck)
-		return true;
-	if ((p->when & PASS_UNCLEAN) && !c->sb.clean)
-		return true;
-	if (p->when & PASS_ALWAYS)
-		return true;
-	return false;
-}
-
-static int bch2_run_recovery_pass(struct bch_fs *c, enum bch_recovery_pass pass)
-{
-	struct recovery_pass_fn *p = recovery_pass_fns + pass;
-	int ret;
-
-	if (!(p->when & PASS_SILENT))
-		bch2_print(c, KERN_INFO bch2_log_msg(c, "%s..."),
-			   bch2_recovery_passes[pass]);
-	ret = p->fn(c);
-	if (ret)
-		return ret;
-	if (!(p->when & PASS_SILENT))
-		bch2_print(c, KERN_CONT " done\n");
-
-	return 0;
-}
-
-static int bch2_run_recovery_passes(struct bch_fs *c)
-{
-	int ret = 0;
-
-	while (c->curr_recovery_pass < ARRAY_SIZE(recovery_pass_fns)) {
-		if (should_run_recovery_pass(c, c->curr_recovery_pass)) {
-			unsigned pass = c->curr_recovery_pass;
-
-			ret = bch2_run_recovery_pass(c, c->curr_recovery_pass);
-			if (bch2_err_matches(ret, BCH_ERR_restart_recovery) ||
-			    (ret && c->curr_recovery_pass < pass))
-				continue;
-			if (ret)
-				break;
-
-			c->recovery_passes_complete |= BIT_ULL(c->curr_recovery_pass);
-		}
-		c->curr_recovery_pass++;
-		c->recovery_pass_done = max(c->recovery_pass_done, c->curr_recovery_pass);
-	}
-
-	return ret;
-}
-
-int bch2_run_online_recovery_passes(struct bch_fs *c)
-{
-	int ret = 0;
-
-	for (unsigned i = 0; i < ARRAY_SIZE(recovery_pass_fns); i++) {
-		struct recovery_pass_fn *p = recovery_pass_fns + i;
-
-		if (!(p->when & PASS_ONLINE))
-			continue;
-
-		ret = bch2_run_recovery_pass(c, i);
-		if (bch2_err_matches(ret, BCH_ERR_restart_recovery)) {
-			i = c->curr_recovery_pass;
-			continue;
-		}
-		if (ret)
-			break;
-	}
-
-	return ret;
-}
-
 int bch2_fs_recovery(struct bch_fs *c)
 {
 	struct bch_sb_field_clean *clean = NULL;
@@ -769,69 +618,57 @@ int bch2_fs_recovery(struct bch_fs *c)
 		goto err;
 	}
 
-	if (c->opts.fsck && c->opts.norecovery) {
-		bch_err(c, "cannot select both norecovery and fsck");
-		ret = -EINVAL;
-		goto err;
-	}
-
-	if (!c->opts.nochanges) {
-		mutex_lock(&c->sb_lock);
-		bool write_sb = false;
-
-		struct bch_sb_field_ext *ext =
-			bch2_sb_field_get_minsize(&c->disk_sb, ext, sizeof(*ext) / sizeof(u64));
-		if (!ext) {
-			ret = -BCH_ERR_ENOSPC_sb;
-			mutex_unlock(&c->sb_lock);
-			goto err;
-		}
+	if (c->opts.norecovery)
+		c->opts.recovery_pass_last = BCH_RECOVERY_PASS_journal_replay - 1;
 
-		if (BCH_SB_HAS_TOPOLOGY_ERRORS(c->disk_sb.sb)) {
-			ext->recovery_passes_required[0] |=
-				cpu_to_le64(bch2_recovery_passes_to_stable(BIT_ULL(BCH_RECOVERY_PASS_check_topology)));
-			write_sb = true;
-		}
+	mutex_lock(&c->sb_lock);
+	struct bch_sb_field_ext *ext = bch2_sb_field_get(c->disk_sb.sb, ext);
+	bool write_sb = false;
 
-		u64 sb_passes = bch2_recovery_passes_from_stable(le64_to_cpu(ext->recovery_passes_required[0]));
-		if (sb_passes) {
-			struct printbuf buf = PRINTBUF;
-			prt_str(&buf, "superblock requires following recovery passes to be run:\n  ");
-			prt_bitflags(&buf, bch2_recovery_passes, sb_passes);
-			bch_info(c, "%s", buf.buf);
-			printbuf_exit(&buf);
-		}
+	if (BCH_SB_HAS_TOPOLOGY_ERRORS(c->disk_sb.sb)) {
+		ext->recovery_passes_required[0] |=
+			cpu_to_le64(bch2_recovery_passes_to_stable(BIT_ULL(BCH_RECOVERY_PASS_check_topology)));
+		write_sb = true;
+	}
 
-		if (bch2_check_version_downgrade(c)) {
-			struct printbuf buf = PRINTBUF;
+	u64 sb_passes = bch2_recovery_passes_from_stable(le64_to_cpu(ext->recovery_passes_required[0]));
+	if (sb_passes) {
+		struct printbuf buf = PRINTBUF;
+		prt_str(&buf, "superblock requires following recovery passes to be run:\n  ");
+		prt_bitflags(&buf, bch2_recovery_passes, sb_passes);
+		bch_info(c, "%s", buf.buf);
+		printbuf_exit(&buf);
+	}
 
-			prt_str(&buf, "Version downgrade required:");
+	if (bch2_check_version_downgrade(c)) {
+		struct printbuf buf = PRINTBUF;
 
-			__le64 passes = ext->recovery_passes_required[0];
-			bch2_sb_set_downgrade(c,
-					BCH_VERSION_MINOR(bcachefs_metadata_version_current),
-					BCH_VERSION_MINOR(c->sb.version));
-			passes = ext->recovery_passes_required[0] & ~passes;
-			if (passes) {
-				prt_str(&buf, "\n  running recovery passes: ");
-				prt_bitflags(&buf, bch2_recovery_passes,
-					     bch2_recovery_passes_from_stable(le64_to_cpu(passes)));
-			}
+		prt_str(&buf, "Version downgrade required:");
 
-			bch_info(c, "%s", buf.buf);
-			printbuf_exit(&buf);
-			write_sb = true;
+		__le64 passes = ext->recovery_passes_required[0];
+		bch2_sb_set_downgrade(c,
+				      BCH_VERSION_MINOR(bcachefs_metadata_version_current),
+				      BCH_VERSION_MINOR(c->sb.version));
+		passes = ext->recovery_passes_required[0] & ~passes;
+		if (passes) {
+			prt_str(&buf, "\n  running recovery passes: ");
+			prt_bitflags(&buf, bch2_recovery_passes,
+				     bch2_recovery_passes_from_stable(le64_to_cpu(passes)));
 		}
 
-		if (check_version_upgrade(c))
-			write_sb = true;
+		bch_info(c, "%s", buf.buf);
+		printbuf_exit(&buf);
+		write_sb = true;
+	}
 
-		if (write_sb)
-			bch2_write_super(c);
+	if (check_version_upgrade(c))
+		write_sb = true;
 
-		c->recovery_passes_explicit |= bch2_recovery_passes_from_stable(le64_to_cpu(ext->recovery_passes_required[0]));
-		mutex_unlock(&c->sb_lock);
-	}
+	c->recovery_passes_explicit |= bch2_recovery_passes_from_stable(le64_to_cpu(ext->recovery_passes_required[0]));
+
+	if (write_sb)
+		bch2_write_super(c);
+	mutex_unlock(&c->sb_lock);
 
 	if (c->opts.fsck && IS_ENABLED(CONFIG_BCACHEFS_DEBUG))
 		c->recovery_passes_explicit |= BIT_ULL(BCH_RECOVERY_PASS_check_topology);
@@ -845,7 +682,9 @@ int bch2_fs_recovery(struct bch_fs *c)
 		goto err;
 	}
 
-	if (!c->sb.clean || c->opts.fsck || c->opts.keep_journal) {
+	bch2_journal_pos_from_member_info_resume(c);
+
+	if (!c->sb.clean || c->opts.retain_recovery_info) {
 		struct genradix_iter iter;
 		struct journal_replay **i;
 
@@ -862,7 +701,7 @@ int bch2_fs_recovery(struct bch_fs *c)
 			goto out;
 
 		genradix_for_each_reverse(&c->journal_entries, iter, i)
-			if (*i && !(*i)->ignore) {
+			if (!journal_replay_ignore(*i)) {
 				last_journal_entry = &(*i)->j;
 				break;
 			}
@@ -887,7 +726,8 @@ int bch2_fs_recovery(struct bch_fs *c)
 			genradix_for_each_reverse(&c->journal_entries, iter, i)
 				if (*i) {
 					last_journal_entry = &(*i)->j;
-					(*i)->ignore = false;
+					(*i)->ignore_blacklisted = false;
+					(*i)->ignore_not_dirty= false;
 					/*
 					 * This was probably a NO_FLUSH entry,
 					 * so last_seq was garbage - but we know
@@ -923,10 +763,8 @@ use_clean:
 	c->journal_replay_seq_start	= last_seq;
 	c->journal_replay_seq_end	= blacklist_seq - 1;
 
-	if (c->opts.reconstruct_alloc) {
-		c->sb.compat &= ~(1ULL << BCH_COMPAT_alloc_info);
-		drop_alloc_keys(&c->journal_keys);
-	}
+	if (c->opts.reconstruct_alloc)
+		bch2_reconstruct_alloc(c);
 
 	zero_out_btree_mem_ptr(&c->journal_keys);
 
@@ -950,7 +788,7 @@ use_clean:
 			bch2_journal_seq_blacklist_add(c,
 					blacklist_seq, journal_seq);
 		if (ret) {
-			bch_err(c, "error creating new journal seq blacklist entry");
+			bch_err_msg(c, ret, "error creating new journal seq blacklist entry");
 			goto err;
 		}
 	}
@@ -961,9 +799,6 @@ use_clean:
 	if (ret)
 		goto err;
 
-	if (c->opts.reconstruct_alloc)
-		bch2_journal_log_msg(c, "dropping alloc info");
-
 	/*
 	 * Skip past versions that might have possibly been used (as nonces),
 	 * but hadn't had their pointers written:
@@ -981,6 +816,14 @@ use_clean:
 
 	clear_bit(BCH_FS_fsck_running, &c->flags);
 
+	/* fsync if we fixed errors */
+	if (test_bit(BCH_FS_errors_fixed, &c->flags) &&
+	    bch2_write_ref_tryget(c, BCH_WRITE_REF_fsync)) {
+		bch2_journal_flush_all_pins(&c->journal);
+		bch2_journal_meta(&c->journal);
+		bch2_write_ref_put(c, BCH_WRITE_REF_fsync);
+	}
+
 	/* If we fixed errors, verify that fs is actually clean now: */
 	if (IS_ENABLED(CONFIG_BCACHEFS_DEBUG) &&
 	    test_bit(BCH_FS_errors_fixed, &c->flags) &&
@@ -1015,7 +858,8 @@ use_clean:
 	}
 
 	mutex_lock(&c->sb_lock);
-	bool write_sb = false;
+	ext = bch2_sb_field_get(c->disk_sb.sb, ext);
+	write_sb = false;
 
 	if (BCH_SB_VERSION_UPGRADE_COMPLETE(c->disk_sb.sb) != le16_to_cpu(c->disk_sb.sb->version)) {
 		SET_BCH_SB_VERSION_UPGRADE_COMPLETE(c->disk_sb.sb, le16_to_cpu(c->disk_sb.sb->version));
@@ -1028,15 +872,18 @@ use_clean:
 		write_sb = true;
 	}
 
-	if (!test_bit(BCH_FS_error, &c->flags)) {
-		struct bch_sb_field_ext *ext = bch2_sb_field_get(c->disk_sb.sb, ext);
-		if (ext &&
-		    (!bch2_is_zero(ext->recovery_passes_required, sizeof(ext->recovery_passes_required)) ||
-		     !bch2_is_zero(ext->errors_silent, sizeof(ext->errors_silent)))) {
-			memset(ext->recovery_passes_required, 0, sizeof(ext->recovery_passes_required));
-			memset(ext->errors_silent, 0, sizeof(ext->errors_silent));
-			write_sb = true;
-		}
+	if (!test_bit(BCH_FS_error, &c->flags) &&
+	    !bch2_is_zero(ext->errors_silent, sizeof(ext->errors_silent))) {
+		memset(ext->errors_silent, 0, sizeof(ext->errors_silent));
+		write_sb = true;
+	}
+
+	if (c->opts.fsck &&
+	    !test_bit(BCH_FS_error, &c->flags) &&
+	    c->recovery_pass_done == BCH_RECOVERY_PASS_NR - 1 &&
+	    ext->btrees_lost_data) {
+		ext->btrees_lost_data = 0;
+		write_sb = true;
 	}
 
 	if (c->opts.fsck &&
@@ -1047,6 +894,9 @@ use_clean:
 		write_sb = true;
 	}
 
+	if (bch2_blacklist_entries_gc(c))
+		write_sb = true;
+
 	if (write_sb)
 		bch2_write_super(c);
 	mutex_unlock(&c->sb_lock);
@@ -1069,18 +919,16 @@ use_clean:
 		bch_info(c, "scanning for old btree nodes done");
 	}
 
-	if (c->journal_seq_blacklist_table &&
-	    c->journal_seq_blacklist_table->nr > 128)
-		queue_work(system_long_wq, &c->journal_seq_blacklist_gc_work);
-
 	ret = 0;
 out:
 	bch2_flush_fsck_errs(c);
 
-	if (!c->opts.keep_journal &&
-	    test_bit(JOURNAL_REPLAY_DONE, &c->journal.flags))
+	if (!c->opts.retain_recovery_info) {
 		bch2_journal_keys_put_initial(c);
-	kfree(clean);
+		bch2_find_btree_nodes_exit(&c->found_btree_nodes);
+	}
+	if (!IS_ERR(clean))
+		kfree(clean);
 
 	if (!ret &&
 	    test_bit(BCH_FS_need_delete_dead_snapshots, &c->flags) &&
@@ -1105,6 +953,7 @@ int bch2_fs_initialize(struct bch_fs *c)
 	int ret;
 
 	bch_notice(c, "initializing new filesystem");
+	set_bit(BCH_FS_new_fs, &c->flags);
 
 	mutex_lock(&c->sb_lock);
 	c->disk_sb.sb->compat[0] |= cpu_to_le64(1ULL << BCH_COMPAT_extents_above_btree_updates_done);
@@ -1119,11 +968,11 @@ int bch2_fs_initialize(struct bch_fs *c)
 	}
 	mutex_unlock(&c->sb_lock);
 
-	c->curr_recovery_pass = ARRAY_SIZE(recovery_pass_fns);
+	c->curr_recovery_pass = BCH_RECOVERY_PASS_NR;
 	set_bit(BCH_FS_may_go_rw, &c->flags);
 
 	for (unsigned i = 0; i < BTREE_ID_NR; i++)
-		bch2_btree_root_alloc(c, i);
+		bch2_btree_root_alloc_fake(c, i, 0);
 
 	for_each_member_device(c, ca)
 		bch2_dev_usage_init(ca);
@@ -1194,7 +1043,7 @@ int bch2_fs_initialize(struct bch_fs *c)
 	if (ret)
 		goto err;
 
-	c->recovery_pass_done = ARRAY_SIZE(recovery_pass_fns) - 1;
+	c->recovery_pass_done = BCH_RECOVERY_PASS_NR - 1;
 
 	if (enabled_qtypes(c)) {
 		ret = bch2_fs_quota_read(c);
diff --git a/fs/bcachefs/recovery.h b/fs/bcachefs/recovery.h
index 4e9d24719b..4bf818de1f 100644
--- a/fs/bcachefs/recovery.h
+++ b/fs/bcachefs/recovery.h
@@ -2,37 +2,9 @@
 #ifndef _BCACHEFS_RECOVERY_H
 #define _BCACHEFS_RECOVERY_H
 
-extern const char * const bch2_recovery_passes[];
+void bch2_btree_lost_data(struct bch_fs *, enum btree_id);
 
-u64 bch2_recovery_passes_to_stable(u64 v);
-u64 bch2_recovery_passes_from_stable(u64 v);
-
-/*
- * For when we need to rewind recovery passes and run a pass we skipped:
- */
-static inline int bch2_run_explicit_recovery_pass(struct bch_fs *c,
-						  enum bch_recovery_pass pass)
-{
-	if (c->recovery_passes_explicit & BIT_ULL(pass))
-		return 0;
-
-	bch_info(c, "running explicit recovery pass %s (%u), currently at %s (%u)",
-		 bch2_recovery_passes[pass], pass,
-		 bch2_recovery_passes[c->curr_recovery_pass], c->curr_recovery_pass);
-
-	c->recovery_passes_explicit |= BIT_ULL(pass);
-
-	if (c->curr_recovery_pass >= pass) {
-		c->curr_recovery_pass = pass;
-		c->recovery_passes_complete &= (1ULL << pass) >> 1;
-		return -BCH_ERR_restart_recovery;
-	} else {
-		return 0;
-	}
-}
-
-int bch2_run_online_recovery_passes(struct bch_fs *);
-u64 bch2_fsck_recovery_passes(void);
+int bch2_journal_replay(struct bch_fs *);
 
 int bch2_fs_recovery(struct bch_fs *);
 int bch2_fs_initialize(struct bch_fs *);
diff --git a/fs/bcachefs/recovery_passes.c b/fs/bcachefs/recovery_passes.c
new file mode 100644
index 0000000000..4a9eb9582b
--- /dev/null
+++ b/fs/bcachefs/recovery_passes.c
@@ -0,0 +1,245 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include "bcachefs.h"
+#include "alloc_background.h"
+#include "backpointers.h"
+#include "btree_gc.h"
+#include "btree_node_scan.h"
+#include "ec.h"
+#include "fsck.h"
+#include "inode.h"
+#include "journal.h"
+#include "lru.h"
+#include "logged_ops.h"
+#include "rebalance.h"
+#include "recovery.h"
+#include "recovery_passes.h"
+#include "snapshot.h"
+#include "subvolume.h"
+#include "super.h"
+#include "super-io.h"
+
+const char * const bch2_recovery_passes[] = {
+#define x(_fn, ...)	#_fn,
+	BCH_RECOVERY_PASSES()
+#undef x
+	NULL
+};
+
+static int bch2_set_may_go_rw(struct bch_fs *c)
+{
+	struct journal_keys *keys = &c->journal_keys;
+
+	/*
+	 * After we go RW, the journal keys buffer can't be modified (except for
+	 * setting journal_key->overwritten: it will be accessed by multiple
+	 * threads
+	 */
+	move_gap(keys, keys->nr);
+
+	set_bit(BCH_FS_may_go_rw, &c->flags);
+
+	if (keys->nr || c->opts.fsck || !c->sb.clean || c->recovery_passes_explicit)
+		return bch2_fs_read_write_early(c);
+	return 0;
+}
+
+struct recovery_pass_fn {
+	int		(*fn)(struct bch_fs *);
+	unsigned	when;
+};
+
+static struct recovery_pass_fn recovery_pass_fns[] = {
+#define x(_fn, _id, _when)	{ .fn = bch2_##_fn, .when = _when },
+	BCH_RECOVERY_PASSES()
+#undef x
+};
+
+static const u8 passes_to_stable_map[] = {
+#define x(n, id, ...)	[BCH_RECOVERY_PASS_##n] = BCH_RECOVERY_PASS_STABLE_##n,
+	BCH_RECOVERY_PASSES()
+#undef x
+};
+
+static enum bch_recovery_pass_stable bch2_recovery_pass_to_stable(enum bch_recovery_pass pass)
+{
+	return passes_to_stable_map[pass];
+}
+
+u64 bch2_recovery_passes_to_stable(u64 v)
+{
+	u64 ret = 0;
+	for (unsigned i = 0; i < ARRAY_SIZE(passes_to_stable_map); i++)
+		if (v & BIT_ULL(i))
+			ret |= BIT_ULL(passes_to_stable_map[i]);
+	return ret;
+}
+
+u64 bch2_recovery_passes_from_stable(u64 v)
+{
+	static const u8 map[] = {
+#define x(n, id, ...)	[BCH_RECOVERY_PASS_STABLE_##n] = BCH_RECOVERY_PASS_##n,
+	BCH_RECOVERY_PASSES()
+#undef x
+	};
+
+	u64 ret = 0;
+	for (unsigned i = 0; i < ARRAY_SIZE(map); i++)
+		if (v & BIT_ULL(i))
+			ret |= BIT_ULL(map[i]);
+	return ret;
+}
+
+/*
+ * For when we need to rewind recovery passes and run a pass we skipped:
+ */
+int bch2_run_explicit_recovery_pass(struct bch_fs *c,
+				    enum bch_recovery_pass pass)
+{
+	if (c->recovery_passes_explicit & BIT_ULL(pass))
+		return 0;
+
+	bch_info(c, "running explicit recovery pass %s (%u), currently at %s (%u)",
+		 bch2_recovery_passes[pass], pass,
+		 bch2_recovery_passes[c->curr_recovery_pass], c->curr_recovery_pass);
+
+	c->recovery_passes_explicit |= BIT_ULL(pass);
+
+	if (c->curr_recovery_pass >= pass) {
+		c->curr_recovery_pass = pass;
+		c->recovery_passes_complete &= (1ULL << pass) >> 1;
+		return -BCH_ERR_restart_recovery;
+	} else {
+		return 0;
+	}
+}
+
+int bch2_run_explicit_recovery_pass_persistent(struct bch_fs *c,
+					       enum bch_recovery_pass pass)
+{
+	enum bch_recovery_pass_stable s = bch2_recovery_pass_to_stable(pass);
+
+	mutex_lock(&c->sb_lock);
+	struct bch_sb_field_ext *ext = bch2_sb_field_get(c->disk_sb.sb, ext);
+
+	if (!test_bit_le64(s, ext->recovery_passes_required)) {
+		__set_bit_le64(s, ext->recovery_passes_required);
+		bch2_write_super(c);
+	}
+	mutex_unlock(&c->sb_lock);
+
+	return bch2_run_explicit_recovery_pass(c, pass);
+}
+
+static void bch2_clear_recovery_pass_required(struct bch_fs *c,
+					      enum bch_recovery_pass pass)
+{
+	enum bch_recovery_pass_stable s = bch2_recovery_pass_to_stable(pass);
+
+	mutex_lock(&c->sb_lock);
+	struct bch_sb_field_ext *ext = bch2_sb_field_get(c->disk_sb.sb, ext);
+
+	if (test_bit_le64(s, ext->recovery_passes_required)) {
+		__clear_bit_le64(s, ext->recovery_passes_required);
+		bch2_write_super(c);
+	}
+	mutex_unlock(&c->sb_lock);
+}
+
+u64 bch2_fsck_recovery_passes(void)
+{
+	u64 ret = 0;
+
+	for (unsigned i = 0; i < ARRAY_SIZE(recovery_pass_fns); i++)
+		if (recovery_pass_fns[i].when & PASS_FSCK)
+			ret |= BIT_ULL(i);
+	return ret;
+}
+
+static bool should_run_recovery_pass(struct bch_fs *c, enum bch_recovery_pass pass)
+{
+	struct recovery_pass_fn *p = recovery_pass_fns + pass;
+
+	if (c->recovery_passes_explicit & BIT_ULL(pass))
+		return true;
+	if ((p->when & PASS_FSCK) && c->opts.fsck)
+		return true;
+	if ((p->when & PASS_UNCLEAN) && !c->sb.clean)
+		return true;
+	if (p->when & PASS_ALWAYS)
+		return true;
+	return false;
+}
+
+static int bch2_run_recovery_pass(struct bch_fs *c, enum bch_recovery_pass pass)
+{
+	struct recovery_pass_fn *p = recovery_pass_fns + pass;
+	int ret;
+
+	if (!(p->when & PASS_SILENT))
+		bch2_print(c, KERN_INFO bch2_log_msg(c, "%s..."),
+			   bch2_recovery_passes[pass]);
+	ret = p->fn(c);
+	if (ret)
+		return ret;
+	if (!(p->when & PASS_SILENT))
+		bch2_print(c, KERN_CONT " done\n");
+
+	return 0;
+}
+
+int bch2_run_online_recovery_passes(struct bch_fs *c)
+{
+	int ret = 0;
+
+	for (unsigned i = 0; i < ARRAY_SIZE(recovery_pass_fns); i++) {
+		struct recovery_pass_fn *p = recovery_pass_fns + i;
+
+		if (!(p->when & PASS_ONLINE))
+			continue;
+
+		ret = bch2_run_recovery_pass(c, i);
+		if (bch2_err_matches(ret, BCH_ERR_restart_recovery)) {
+			i = c->curr_recovery_pass;
+			continue;
+		}
+		if (ret)
+			break;
+	}
+
+	return ret;
+}
+
+int bch2_run_recovery_passes(struct bch_fs *c)
+{
+	int ret = 0;
+
+	while (c->curr_recovery_pass < ARRAY_SIZE(recovery_pass_fns)) {
+		if (c->opts.recovery_pass_last &&
+		    c->curr_recovery_pass > c->opts.recovery_pass_last)
+			break;
+
+		if (should_run_recovery_pass(c, c->curr_recovery_pass)) {
+			unsigned pass = c->curr_recovery_pass;
+
+			ret =   bch2_run_recovery_pass(c, c->curr_recovery_pass) ?:
+				bch2_journal_flush(&c->journal);
+			if (bch2_err_matches(ret, BCH_ERR_restart_recovery) ||
+			    (ret && c->curr_recovery_pass < pass))
+				continue;
+			if (ret)
+				break;
+
+			c->recovery_passes_complete |= BIT_ULL(c->curr_recovery_pass);
+		}
+
+		c->recovery_pass_done = max(c->recovery_pass_done, c->curr_recovery_pass);
+
+		if (!test_bit(BCH_FS_error, &c->flags))
+			bch2_clear_recovery_pass_required(c, c->curr_recovery_pass);
+
+		c->curr_recovery_pass++;
+	}
+
+	return ret;
+}
diff --git a/fs/bcachefs/recovery_passes.h b/fs/bcachefs/recovery_passes.h
new file mode 100644
index 0000000000..99b464e127
--- /dev/null
+++ b/fs/bcachefs/recovery_passes.h
@@ -0,0 +1,17 @@
+#ifndef _BCACHEFS_RECOVERY_PASSES_H
+#define _BCACHEFS_RECOVERY_PASSES_H
+
+extern const char * const bch2_recovery_passes[];
+
+u64 bch2_recovery_passes_to_stable(u64 v);
+u64 bch2_recovery_passes_from_stable(u64 v);
+
+u64 bch2_fsck_recovery_passes(void);
+
+int bch2_run_explicit_recovery_pass(struct bch_fs *, enum bch_recovery_pass);
+int bch2_run_explicit_recovery_pass_persistent(struct bch_fs *, enum bch_recovery_pass);
+
+int bch2_run_online_recovery_passes(struct bch_fs *);
+int bch2_run_recovery_passes(struct bch_fs *);
+
+#endif /* _BCACHEFS_RECOVERY_PASSES_H */
diff --git a/fs/bcachefs/recovery_types.h b/fs/bcachefs/recovery_passes_types.h
index fa0c8efd2a..773aea9a00 100644
--- a/fs/bcachefs/recovery_types.h
+++ b/fs/bcachefs/recovery_passes_types.h
@@ -1,6 +1,6 @@
 /* SPDX-License-Identifier: GPL-2.0 */
-#ifndef _BCACHEFS_RECOVERY_TYPES_H
-#define _BCACHEFS_RECOVERY_TYPES_H
+#ifndef _BCACHEFS_RECOVERY_PASSES_TYPES_H
+#define _BCACHEFS_RECOVERY_PASSES_TYPES_H
 
 #define PASS_SILENT		BIT(0)
 #define PASS_FSCK		BIT(1)
@@ -13,11 +13,12 @@
  * must never change:
  */
 #define BCH_RECOVERY_PASSES()							\
+	x(scan_for_btree_nodes,			37, 0)				\
+	x(check_topology,			 4, 0)				\
 	x(alloc_read,				 0, PASS_ALWAYS)		\
 	x(stripes_read,				 1, PASS_ALWAYS)		\
 	x(initialize_subvolumes,		 2, 0)				\
 	x(snapshots_read,			 3, PASS_ALWAYS)		\
-	x(check_topology,			 4, 0)				\
 	x(check_allocations,			 5, PASS_FSCK)			\
 	x(trans_mark_dev_sbs,			 6, PASS_ALWAYS|PASS_SILENT)	\
 	x(fs_journal_alloc,			 7, PASS_ALWAYS|PASS_SILENT)	\
@@ -31,20 +32,23 @@
 	x(check_alloc_to_lru_refs,		15, PASS_ONLINE|PASS_FSCK)	\
 	x(fs_freespace_init,			16, PASS_ALWAYS|PASS_SILENT)	\
 	x(bucket_gens_init,			17, 0)				\
+	x(reconstruct_snapshots,		38, 0)				\
 	x(check_snapshot_trees,			18, PASS_ONLINE|PASS_FSCK)	\
 	x(check_snapshots,			19, PASS_ONLINE|PASS_FSCK)	\
 	x(check_subvols,			20, PASS_ONLINE|PASS_FSCK)	\
+	x(check_subvol_children,		35, PASS_ONLINE|PASS_FSCK)	\
 	x(delete_dead_snapshots,		21, PASS_ONLINE|PASS_FSCK)	\
 	x(fs_upgrade_for_subvolumes,		22, 0)				\
-	x(resume_logged_ops,			23, PASS_ALWAYS)		\
 	x(check_inodes,				24, PASS_FSCK)			\
 	x(check_extents,			25, PASS_FSCK)			\
 	x(check_indirect_extents,		26, PASS_FSCK)			\
 	x(check_dirents,			27, PASS_FSCK)			\
 	x(check_xattrs,				28, PASS_FSCK)			\
 	x(check_root,				29, PASS_ONLINE|PASS_FSCK)	\
+	x(check_subvolume_structure,		36, PASS_ONLINE|PASS_FSCK)	\
 	x(check_directory_structure,		30, PASS_ONLINE|PASS_FSCK)	\
 	x(check_nlinks,				31, PASS_FSCK)			\
+	x(resume_logged_ops,			23, PASS_ALWAYS)		\
 	x(delete_dead_inodes,			32, PASS_FSCK|PASS_UNCLEAN)	\
 	x(fix_reflink_p,			33, 0)				\
 	x(set_fs_needs_rebalance,		34, 0)				\
@@ -54,6 +58,7 @@ enum bch_recovery_pass {
 #define x(n, id, when)	BCH_RECOVERY_PASS_##n,
 	BCH_RECOVERY_PASSES()
 #undef x
+	BCH_RECOVERY_PASS_NR
 };
 
 /* But we also need stable identifiers that can be used in the superblock */
@@ -63,4 +68,4 @@ enum bch_recovery_pass_stable {
 #undef x
 };
 
-#endif /* _BCACHEFS_RECOVERY_TYPES_H */
+#endif /* _BCACHEFS_RECOVERY_PASSES_TYPES_H */
diff --git a/fs/bcachefs/reflink.c b/fs/bcachefs/reflink.c
index c47c66c2b3..9ac6cf21cf 100644
--- a/fs/bcachefs/reflink.c
+++ b/fs/bcachefs/reflink.c
@@ -30,7 +30,7 @@ static inline unsigned bkey_type_to_indirect(const struct bkey *k)
 /* reflink pointers */
 
 int bch2_reflink_p_invalid(struct bch_fs *c, struct bkey_s_c k,
-			   enum bkey_invalid_flags flags,
+			   enum bch_validate_flags flags,
 			   struct printbuf *err)
 {
 	struct bkey_s_c_reflink_p p = bkey_s_c_to_reflink_p(k);
@@ -74,20 +74,20 @@ bool bch2_reflink_p_merge(struct bch_fs *c, struct bkey_s _l, struct bkey_s_c _r
 }
 
 static int trans_trigger_reflink_p_segment(struct btree_trans *trans,
-			struct bkey_s_c_reflink_p p,
-			u64 *idx, unsigned flags)
+			struct bkey_s_c_reflink_p p, u64 *idx,
+			enum btree_iter_update_trigger_flags flags)
 {
 	struct bch_fs *c = trans->c;
 	struct btree_iter iter;
 	struct bkey_i *k;
 	__le64 *refcount;
-	int add = !(flags & BTREE_TRIGGER_OVERWRITE) ? 1 : -1;
+	int add = !(flags & BTREE_TRIGGER_overwrite) ? 1 : -1;
 	struct printbuf buf = PRINTBUF;
 	int ret;
 
 	k = bch2_bkey_get_mut_noupdate(trans, &iter,
 			BTREE_ID_reflink, POS(0, *idx),
-			BTREE_ITER_WITH_UPDATES);
+			BTREE_ITER_with_updates);
 	ret = PTR_ERR_OR_ZERO(k);
 	if (ret)
 		goto err;
@@ -102,7 +102,7 @@ static int trans_trigger_reflink_p_segment(struct btree_trans *trans,
 		goto err;
 	}
 
-	if (!*refcount && (flags & BTREE_TRIGGER_OVERWRITE)) {
+	if (!*refcount && (flags & BTREE_TRIGGER_overwrite)) {
 		bch2_bkey_val_to_text(&buf, c, p.s_c);
 		bch2_trans_inconsistent(trans,
 			"indirect extent refcount underflow at %llu while marking\n  %s",
@@ -111,7 +111,7 @@ static int trans_trigger_reflink_p_segment(struct btree_trans *trans,
 		goto err;
 	}
 
-	if (flags & BTREE_TRIGGER_INSERT) {
+	if (flags & BTREE_TRIGGER_insert) {
 		struct bch_reflink_p *v = (struct bch_reflink_p *) p.v;
 		u64 pad;
 
@@ -141,12 +141,13 @@ err:
 }
 
 static s64 gc_trigger_reflink_p_segment(struct btree_trans *trans,
-				struct bkey_s_c_reflink_p p,
-				u64 *idx, unsigned flags, size_t r_idx)
+				struct bkey_s_c_reflink_p p, u64 *idx,
+				enum btree_iter_update_trigger_flags flags,
+				size_t r_idx)
 {
 	struct bch_fs *c = trans->c;
 	struct reflink_gc *r;
-	int add = !(flags & BTREE_TRIGGER_OVERWRITE) ? 1 : -1;
+	int add = !(flags & BTREE_TRIGGER_overwrite) ? 1 : -1;
 	u64 start = le64_to_cpu(p.v->idx);
 	u64 end = le64_to_cpu(p.v->idx) + p.k->size;
 	u64 next_idx = end + le32_to_cpu(p.v->back_pad);
@@ -163,10 +164,13 @@ static s64 gc_trigger_reflink_p_segment(struct btree_trans *trans,
 
 	BUG_ON((s64) r->refcount + add < 0);
 
-	r->refcount += add;
+	if (flags & BTREE_TRIGGER_gc)
+		r->refcount += add;
 	*idx = r->offset;
 	return 0;
 not_found:
+	BUG_ON(!(flags & BTREE_TRIGGER_check_repair));
+
 	if (fsck_err(c, reflink_p_to_missing_reflink_v,
 		     "pointer to missing indirect extent\n"
 		     "  %s\n"
@@ -185,12 +189,11 @@ not_found:
 		} else {
 			bkey_error_init(update);
 			update->k.p		= p.k->p;
-			update->k.p.offset	= next_idx;
-			update->k.size		= next_idx - *idx;
+			update->k.size		= p.k->size;
 			set_bkey_val_u64s(&update->k, 0);
 		}
 
-		ret = bch2_btree_insert_trans(trans, BTREE_ID_extents, update, BTREE_TRIGGER_NORUN);
+		ret = bch2_btree_insert_trans(trans, BTREE_ID_extents, update, BTREE_TRIGGER_norun);
 	}
 
 	*idx = next_idx;
@@ -201,8 +204,8 @@ fsck_err:
 }
 
 static int __trigger_reflink_p(struct btree_trans *trans,
-			    enum btree_id btree_id, unsigned level,
-			    struct bkey_s_c k, unsigned flags)
+		enum btree_id btree_id, unsigned level, struct bkey_s_c k,
+		enum btree_iter_update_trigger_flags flags)
 {
 	struct bch_fs *c = trans->c;
 	struct bkey_s_c_reflink_p p = bkey_s_c_to_reflink_p(k);
@@ -211,12 +214,12 @@ static int __trigger_reflink_p(struct btree_trans *trans,
 	u64 idx = le64_to_cpu(p.v->idx) - le32_to_cpu(p.v->front_pad);
 	u64 end = le64_to_cpu(p.v->idx) + p.k->size + le32_to_cpu(p.v->back_pad);
 
-	if (flags & BTREE_TRIGGER_TRANSACTIONAL) {
+	if (flags & BTREE_TRIGGER_transactional) {
 		while (idx < end && !ret)
 			ret = trans_trigger_reflink_p_segment(trans, p, &idx, flags);
 	}
 
-	if (flags & BTREE_TRIGGER_GC) {
+	if (flags & (BTREE_TRIGGER_check_repair|BTREE_TRIGGER_gc)) {
 		size_t l = 0, r = c->reflink_gc_nr;
 
 		while (l < r) {
@@ -239,10 +242,10 @@ int bch2_trigger_reflink_p(struct btree_trans *trans,
 			   enum btree_id btree_id, unsigned level,
 			   struct bkey_s_c old,
 			   struct bkey_s new,
-			   unsigned flags)
+			   enum btree_iter_update_trigger_flags flags)
 {
-	if ((flags & BTREE_TRIGGER_TRANSACTIONAL) &&
-	    (flags & BTREE_TRIGGER_INSERT)) {
+	if ((flags & BTREE_TRIGGER_transactional) &&
+	    (flags & BTREE_TRIGGER_insert)) {
 		struct bch_reflink_p *v = bkey_s_to_reflink_p(new).v;
 
 		v->front_pad = v->back_pad = 0;
@@ -254,7 +257,7 @@ int bch2_trigger_reflink_p(struct btree_trans *trans,
 /* indirect extents */
 
 int bch2_reflink_v_invalid(struct bch_fs *c, struct bkey_s_c k,
-			   enum bkey_invalid_flags flags,
+			   enum bch_validate_flags flags,
 			   struct printbuf *err)
 {
 	return bch2_bkey_ptrs_invalid(c, k, flags, err);
@@ -282,23 +285,25 @@ bool bch2_reflink_v_merge(struct bch_fs *c, struct bkey_s _l, struct bkey_s_c _r
 }
 #endif
 
-static inline void check_indirect_extent_deleting(struct bkey_s new, unsigned *flags)
+static inline void
+check_indirect_extent_deleting(struct bkey_s new,
+			       enum btree_iter_update_trigger_flags *flags)
 {
-	if ((*flags & BTREE_TRIGGER_INSERT) && !*bkey_refcount(new)) {
+	if ((*flags & BTREE_TRIGGER_insert) && !*bkey_refcount(new)) {
 		new.k->type = KEY_TYPE_deleted;
 		new.k->size = 0;
 		set_bkey_val_u64s(new.k, 0);
-		*flags &= ~BTREE_TRIGGER_INSERT;
+		*flags &= ~BTREE_TRIGGER_insert;
 	}
 }
 
 int bch2_trigger_reflink_v(struct btree_trans *trans,
 			   enum btree_id btree_id, unsigned level,
 			   struct bkey_s_c old, struct bkey_s new,
-			   unsigned flags)
+			   enum btree_iter_update_trigger_flags flags)
 {
-	if ((flags & BTREE_TRIGGER_TRANSACTIONAL) &&
-	    (flags & BTREE_TRIGGER_INSERT))
+	if ((flags & BTREE_TRIGGER_transactional) &&
+	    (flags & BTREE_TRIGGER_insert))
 		check_indirect_extent_deleting(new, &flags);
 
 	return bch2_trigger_extent(trans, btree_id, level, old, new, flags);
@@ -307,7 +312,7 @@ int bch2_trigger_reflink_v(struct btree_trans *trans,
 /* indirect inline data */
 
 int bch2_indirect_inline_data_invalid(struct bch_fs *c, struct bkey_s_c k,
-				      enum bkey_invalid_flags flags,
+				      enum bch_validate_flags flags,
 				      struct printbuf *err)
 {
 	return 0;
@@ -327,7 +332,7 @@ void bch2_indirect_inline_data_to_text(struct printbuf *out,
 int bch2_trigger_indirect_inline_data(struct btree_trans *trans,
 			      enum btree_id btree_id, unsigned level,
 			      struct bkey_s_c old, struct bkey_s new,
-			      unsigned flags)
+			      enum btree_iter_update_trigger_flags flags)
 {
 	check_indirect_extent_deleting(new, &flags);
 
@@ -350,7 +355,7 @@ static int bch2_make_extent_indirect(struct btree_trans *trans,
 		bch2_check_set_feature(c, BCH_FEATURE_reflink_inline_data);
 
 	bch2_trans_iter_init(trans, &reflink_iter, BTREE_ID_reflink, POS_MAX,
-			     BTREE_ITER_INTENT);
+			     BTREE_ITER_intent);
 	k = bch2_btree_iter_peek_prev(&reflink_iter);
 	ret = bkey_err(k);
 	if (ret)
@@ -395,7 +400,7 @@ static int bch2_make_extent_indirect(struct btree_trans *trans,
 	r_p->v.idx = cpu_to_le64(bkey_start_offset(&r_v->k));
 
 	ret = bch2_trans_update(trans, extent_iter, &r_p->k_i,
-				BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
+				BTREE_UPDATE_internal_snapshot_node);
 err:
 	bch2_trans_iter_exit(trans, &reflink_iter);
 
@@ -456,9 +461,9 @@ s64 bch2_remap_range(struct bch_fs *c,
 		goto err;
 
 	bch2_trans_iter_init(trans, &src_iter, BTREE_ID_extents, src_start,
-			     BTREE_ITER_INTENT);
+			     BTREE_ITER_intent);
 	bch2_trans_iter_init(trans, &dst_iter, BTREE_ID_extents, dst_start,
-			     BTREE_ITER_INTENT);
+			     BTREE_ITER_intent);
 
 	while ((ret == 0 ||
 		bch2_err_matches(ret, BCH_ERR_transaction_restart)) &&
@@ -568,7 +573,7 @@ s64 bch2_remap_range(struct bch_fs *c,
 		bch2_trans_begin(trans);
 
 		ret2 = bch2_inode_peek(trans, &inode_iter, &inode_u,
-				       dst_inum, BTREE_ITER_INTENT);
+				       dst_inum, BTREE_ITER_intent);
 
 		if (!ret2 &&
 		    inode_u.bi_size < new_i_size) {
diff --git a/fs/bcachefs/reflink.h b/fs/bcachefs/reflink.h
index 4d88672897..e894f3a2c6 100644
--- a/fs/bcachefs/reflink.h
+++ b/fs/bcachefs/reflink.h
@@ -2,15 +2,16 @@
 #ifndef _BCACHEFS_REFLINK_H
 #define _BCACHEFS_REFLINK_H
 
-enum bkey_invalid_flags;
+enum bch_validate_flags;
 
 int bch2_reflink_p_invalid(struct bch_fs *, struct bkey_s_c,
-			   enum bkey_invalid_flags, struct printbuf *);
+			   enum bch_validate_flags, struct printbuf *);
 void bch2_reflink_p_to_text(struct printbuf *, struct bch_fs *,
 			    struct bkey_s_c);
 bool bch2_reflink_p_merge(struct bch_fs *, struct bkey_s, struct bkey_s_c);
 int bch2_trigger_reflink_p(struct btree_trans *, enum btree_id, unsigned,
-			   struct bkey_s_c, struct bkey_s, unsigned);
+			   struct bkey_s_c, struct bkey_s,
+			   enum btree_iter_update_trigger_flags);
 
 #define bch2_bkey_ops_reflink_p ((struct bkey_ops) {		\
 	.key_invalid	= bch2_reflink_p_invalid,		\
@@ -21,11 +22,12 @@ int bch2_trigger_reflink_p(struct btree_trans *, enum btree_id, unsigned,
 })
 
 int bch2_reflink_v_invalid(struct bch_fs *, struct bkey_s_c,
-			   enum bkey_invalid_flags, struct printbuf *);
+			   enum bch_validate_flags, struct printbuf *);
 void bch2_reflink_v_to_text(struct printbuf *, struct bch_fs *,
 			    struct bkey_s_c);
 int bch2_trigger_reflink_v(struct btree_trans *, enum btree_id, unsigned,
-			      struct bkey_s_c, struct bkey_s, unsigned);
+			   struct bkey_s_c, struct bkey_s,
+			   enum btree_iter_update_trigger_flags);
 
 #define bch2_bkey_ops_reflink_v ((struct bkey_ops) {		\
 	.key_invalid	= bch2_reflink_v_invalid,		\
@@ -36,13 +38,13 @@ int bch2_trigger_reflink_v(struct btree_trans *, enum btree_id, unsigned,
 })
 
 int bch2_indirect_inline_data_invalid(struct bch_fs *, struct bkey_s_c,
-				      enum bkey_invalid_flags, struct printbuf *);
+				      enum bch_validate_flags, struct printbuf *);
 void bch2_indirect_inline_data_to_text(struct printbuf *,
 				struct bch_fs *, struct bkey_s_c);
 int bch2_trigger_indirect_inline_data(struct btree_trans *,
 					 enum btree_id, unsigned,
 			      struct bkey_s_c, struct bkey_s,
-			      unsigned);
+			      enum btree_iter_update_trigger_flags);
 
 #define bch2_bkey_ops_indirect_inline_data ((struct bkey_ops) {	\
 	.key_invalid	= bch2_indirect_inline_data_invalid,	\
diff --git a/fs/bcachefs/replicas.c b/fs/bcachefs/replicas.c
index cc2672c120..57a1f09cca 100644
--- a/fs/bcachefs/replicas.c
+++ b/fs/bcachefs/replicas.c
@@ -6,12 +6,15 @@
 #include "replicas.h"
 #include "super-io.h"
 
+#include <linux/sort.h>
+
 static int bch2_cpu_replicas_to_sb_replicas(struct bch_fs *,
 					    struct bch_replicas_cpu *);
 
 /* Some (buggy!) compilers don't allow memcmp to be passed as a pointer */
-static int bch2_memcmp(const void *l, const void *r, size_t size)
+static int bch2_memcmp(const void *l, const void *r,  const void *priv)
 {
+	size_t size = (size_t) priv;
 	return memcmp(l, r, size);
 }
 
@@ -20,14 +23,12 @@ static int bch2_memcmp(const void *l, const void *r, size_t size)
 static void verify_replicas_entry(struct bch_replicas_entry_v1 *e)
 {
 #ifdef CONFIG_BCACHEFS_DEBUG
-	unsigned i;
-
 	BUG_ON(e->data_type >= BCH_DATA_NR);
 	BUG_ON(!e->nr_devs);
 	BUG_ON(e->nr_required > 1 &&
 	       e->nr_required >= e->nr_devs);
 
-	for (i = 0; i + 1 < e->nr_devs; i++)
+	for (unsigned i = 0; i + 1 < e->nr_devs; i++)
 		BUG_ON(e->devs[i] >= e->devs[i + 1]);
 #endif
 }
@@ -39,7 +40,8 @@ void bch2_replicas_entry_sort(struct bch_replicas_entry_v1 *e)
 
 static void bch2_cpu_replicas_sort(struct bch_replicas_cpu *r)
 {
-	eytzinger0_sort(r->entries, r->nr, r->entry_size, bch2_memcmp, NULL);
+	eytzinger0_sort_r(r->entries, r->nr, r->entry_size,
+			  bch2_memcmp, NULL, (void *)(size_t)r->entry_size);
 }
 
 static void bch2_replicas_entry_v0_to_text(struct printbuf *out,
@@ -80,7 +82,7 @@ int bch2_replicas_entry_validate(struct bch_replicas_entry_v1 *r,
 	}
 
 	for (unsigned i = 0; i < r->nr_devs; i++)
-		if (!bch2_dev_exists(sb, r->devs[i])) {
+		if (!bch2_member_exists(sb, r->devs[i])) {
 			prt_printf(err, "invalid device %u in entry ", r->devs[i]);
 			goto bad;
 		}
@@ -188,24 +190,17 @@ cpu_replicas_add_entry(struct bch_fs *c,
 		       struct bch_replicas_cpu *old,
 		       struct bch_replicas_entry_v1 *new_entry)
 {
-	unsigned i;
 	struct bch_replicas_cpu new = {
 		.nr		= old->nr + 1,
 		.entry_size	= max_t(unsigned, old->entry_size,
 					replicas_entry_bytes(new_entry)),
 	};
 
-	for (i = 0; i < new_entry->nr_devs; i++)
-		BUG_ON(!bch2_dev_exists2(c, new_entry->devs[i]));
-
-	BUG_ON(!new_entry->data_type);
-	verify_replicas_entry(new_entry);
-
 	new.entries = kcalloc(new.nr, new.entry_size, GFP_KERNEL);
 	if (!new.entries)
 		return new;
 
-	for (i = 0; i < old->nr; i++)
+	for (unsigned i = 0; i < old->nr; i++)
 		memcpy(cpu_replicas_entry(&new, i),
 		       cpu_replicas_entry(old, i),
 		       old->entry_size);
@@ -226,9 +221,7 @@ static inline int __replicas_entry_idx(struct bch_replicas_cpu *r,
 	if (unlikely(entry_size > r->entry_size))
 		return -1;
 
-	verify_replicas_entry(search);
-
-#define entry_cmp(_l, _r, size)	memcmp(_l, _r, entry_size)
+#define entry_cmp(_l, _r)	memcmp(_l, _r, entry_size)
 	idx = eytzinger0_find(r->entries, r->nr, r->entry_size,
 			      entry_cmp, search);
 #undef entry_cmp
@@ -520,13 +513,16 @@ int bch2_replicas_gc_start(struct bch_fs *c, unsigned typemask)
 	c->replicas_gc.nr		= 0;
 	c->replicas_gc.entry_size	= 0;
 
-	for_each_cpu_replicas_entry(&c->replicas, e)
-		if (!((1 << e->data_type) & typemask)) {
+	for_each_cpu_replicas_entry(&c->replicas, e) {
+		/* Preserve unknown data types */
+		if (e->data_type >= BCH_DATA_NR ||
+		    !((1 << e->data_type) & typemask)) {
 			c->replicas_gc.nr++;
 			c->replicas_gc.entry_size =
 				max_t(unsigned, c->replicas_gc.entry_size,
 				      replicas_entry_bytes(e));
 		}
+	}
 
 	c->replicas_gc.entries = kcalloc(c->replicas_gc.nr,
 					 c->replicas_gc.entry_size,
@@ -538,7 +534,8 @@ int bch2_replicas_gc_start(struct bch_fs *c, unsigned typemask)
 	}
 
 	for_each_cpu_replicas_entry(&c->replicas, e)
-		if (!((1 << e->data_type) & typemask))
+		if (e->data_type >= BCH_DATA_NR ||
+		    !((1 << e->data_type) & typemask))
 			memcpy(cpu_replicas_entry(&c->replicas_gc, i++),
 			       e, c->replicas_gc.entry_size);
 
@@ -824,10 +821,11 @@ static int bch2_cpu_replicas_validate(struct bch_replicas_cpu *cpu_r,
 {
 	unsigned i;
 
-	sort_cmp_size(cpu_r->entries,
-		      cpu_r->nr,
-		      cpu_r->entry_size,
-		      bch2_memcmp, NULL);
+	sort_r(cpu_r->entries,
+	       cpu_r->nr,
+	       cpu_r->entry_size,
+	       bch2_memcmp, NULL,
+	       (void *)(size_t)cpu_r->entry_size);
 
 	for (i = 0; i < cpu_r->nr; i++) {
 		struct bch_replicas_entry_v1 *e =
@@ -855,7 +853,7 @@ static int bch2_cpu_replicas_validate(struct bch_replicas_cpu *cpu_r,
 }
 
 static int bch2_sb_replicas_validate(struct bch_sb *sb, struct bch_sb_field *f,
-				     struct printbuf *err)
+				     enum bch_validate_flags flags, struct printbuf *err)
 {
 	struct bch_sb_field_replicas *sb_r = field_to_type(f, replicas);
 	struct bch_replicas_cpu cpu_r;
@@ -894,7 +892,7 @@ const struct bch_sb_field_ops bch_sb_field_ops_replicas = {
 };
 
 static int bch2_sb_replicas_v0_validate(struct bch_sb *sb, struct bch_sb_field *f,
-					struct printbuf *err)
+					enum bch_validate_flags flags, struct printbuf *err)
 {
 	struct bch_sb_field_replicas_v0 *sb_r = field_to_type(f, replicas_v0);
 	struct bch_replicas_cpu cpu_r;
@@ -942,18 +940,20 @@ bool bch2_have_enough_devs(struct bch_fs *c, struct bch_devs_mask devs,
 
 	percpu_down_read(&c->mark_lock);
 	for_each_cpu_replicas_entry(&c->replicas, e) {
-		unsigned i, nr_online = 0, nr_failed = 0, dflags = 0;
+		unsigned nr_online = 0, nr_failed = 0, dflags = 0;
 		bool metadata = e->data_type < BCH_DATA_user;
 
 		if (e->data_type == BCH_DATA_cached)
 			continue;
 
-		for (i = 0; i < e->nr_devs; i++) {
-			struct bch_dev *ca = bch_dev_bkey_exists(c, e->devs[i]);
-
+		rcu_read_lock();
+		for (unsigned i = 0; i < e->nr_devs; i++) {
 			nr_online += test_bit(e->devs[i], devs.d);
-			nr_failed += ca->mi.state == BCH_MEMBER_STATE_failed;
+
+			struct bch_dev *ca = bch2_dev_rcu(c, e->devs[i]);
+			nr_failed += ca && ca->mi.state == BCH_MEMBER_STATE_failed;
 		}
+		rcu_read_unlock();
 
 		if (nr_failed == e->nr_devs)
 			continue;
@@ -991,7 +991,7 @@ unsigned bch2_sb_dev_has_data(struct bch_sb *sb, unsigned dev)
 {
 	struct bch_sb_field_replicas *replicas;
 	struct bch_sb_field_replicas_v0 *replicas_v0;
-	unsigned i, data_has = 0;
+	unsigned data_has = 0;
 
 	replicas = bch2_sb_field_get(sb, replicas);
 	replicas_v0 = bch2_sb_field_get(sb, replicas_v0);
@@ -999,17 +999,26 @@ unsigned bch2_sb_dev_has_data(struct bch_sb *sb, unsigned dev)
 	if (replicas) {
 		struct bch_replicas_entry_v1 *r;
 
-		for_each_replicas_entry(replicas, r)
-			for (i = 0; i < r->nr_devs; i++)
+		for_each_replicas_entry(replicas, r) {
+			if (r->data_type >= sizeof(data_has) * 8)
+				continue;
+
+			for (unsigned i = 0; i < r->nr_devs; i++)
 				if (r->devs[i] == dev)
 					data_has |= 1 << r->data_type;
+		}
+
 	} else if (replicas_v0) {
 		struct bch_replicas_entry_v0 *r;
 
-		for_each_replicas_entry_v0(replicas_v0, r)
-			for (i = 0; i < r->nr_devs; i++)
+		for_each_replicas_entry_v0(replicas_v0, r) {
+			if (r->data_type >= sizeof(data_has) * 8)
+				continue;
+
+			for (unsigned i = 0; i < r->nr_devs; i++)
 				if (r->devs[i] == dev)
 					data_has |= 1 << r->data_type;
+		}
 	}
 
 
diff --git a/fs/bcachefs/replicas_format.h b/fs/bcachefs/replicas_format.h
new file mode 100644
index 0000000000..b97208195d
--- /dev/null
+++ b/fs/bcachefs/replicas_format.h
@@ -0,0 +1,31 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _BCACHEFS_REPLICAS_FORMAT_H
+#define _BCACHEFS_REPLICAS_FORMAT_H
+
+struct bch_replicas_entry_v0 {
+	__u8			data_type;
+	__u8			nr_devs;
+	__u8			devs[];
+} __packed;
+
+struct bch_sb_field_replicas_v0 {
+	struct bch_sb_field	field;
+	struct bch_replicas_entry_v0 entries[];
+} __packed __aligned(8);
+
+struct bch_replicas_entry_v1 {
+	__u8			data_type;
+	__u8			nr_devs;
+	__u8			nr_required;
+	__u8			devs[];
+} __packed;
+
+struct bch_sb_field_replicas {
+	struct bch_sb_field	field;
+	struct bch_replicas_entry_v1 entries[];
+} __packed __aligned(8);
+
+#define replicas_entry_bytes(_i)					\
+	(offsetof(typeof(*(_i)), devs) + (_i)->nr_devs)
+
+#endif /* _BCACHEFS_REPLICAS_FORMAT_H */
diff --git a/fs/bcachefs/sb-clean.c b/fs/bcachefs/sb-clean.c
index b6bf0ebe7e..47f10ab57f 100644
--- a/fs/bcachefs/sb-clean.c
+++ b/fs/bcachefs/sb-clean.c
@@ -29,6 +29,14 @@ int bch2_sb_clean_validate_late(struct bch_fs *c, struct bch_sb_field_clean *cle
 	for (entry = clean->start;
 	     entry < (struct jset_entry *) vstruct_end(&clean->field);
 	     entry = vstruct_next(entry)) {
+		if (vstruct_end(entry) > vstruct_end(&clean->field)) {
+			bch_err(c, "journal entry (u64s %u) overran end of superblock clean section (u64s %u) by %zu",
+				le16_to_cpu(entry->u64s), le32_to_cpu(clean->field.u64s),
+				(u64 *) vstruct_end(entry) - (u64 *) vstruct_end(&clean->field));
+			bch2_sb_error_count(c, BCH_FSCK_ERR_sb_clean_entry_overrun);
+			return -BCH_ERR_fsck_repair_unimplemented;
+		}
+
 		ret = bch2_journal_entry_validate(c, NULL, entry,
 						  le16_to_cpu(c->disk_sb.sb->version),
 						  BCH_SB_BIG_ENDIAN(c->disk_sb.sb),
@@ -171,22 +179,6 @@ fsck_err:
 	return ERR_PTR(ret);
 }
 
-static struct jset_entry *jset_entry_init(struct jset_entry **end, size_t size)
-{
-	struct jset_entry *entry = *end;
-	unsigned u64s = DIV_ROUND_UP(size, sizeof(u64));
-
-	memset(entry, 0, u64s * sizeof(u64));
-	/*
-	 * The u64s field counts from the start of data, ignoring the shared
-	 * fields.
-	 */
-	entry->u64s = cpu_to_le16(u64s - 1);
-
-	*end = vstruct_next(*end);
-	return entry;
-}
-
 void bch2_journal_super_entries_add_common(struct bch_fs *c,
 					   struct jset_entry **end,
 					   u64 journal_seq)
@@ -274,9 +266,8 @@ void bch2_journal_super_entries_add_common(struct bch_fs *c,
 	}
 }
 
-static int bch2_sb_clean_validate(struct bch_sb *sb,
-				  struct bch_sb_field *f,
-				  struct printbuf *err)
+static int bch2_sb_clean_validate(struct bch_sb *sb, struct bch_sb_field *f,
+				  enum bch_validate_flags flags, struct printbuf *err)
 {
 	struct bch_sb_field_clean *clean = field_to_type(f, clean);
 
@@ -286,6 +277,17 @@ static int bch2_sb_clean_validate(struct bch_sb *sb,
 		return -BCH_ERR_invalid_sb_clean;
 	}
 
+	for (struct jset_entry *entry = clean->start;
+	     entry != vstruct_end(&clean->field);
+	     entry = vstruct_next(entry)) {
+		if ((void *) vstruct_next(entry) > vstruct_end(&clean->field)) {
+			prt_str(err, "entry type ");
+			bch2_prt_jset_entry_type(err, entry->type);
+			prt_str(err, " overruns end of section");
+			return -BCH_ERR_invalid_sb_clean;
+		}
+	}
+
 	return 0;
 }
 
@@ -295,14 +297,15 @@ static void bch2_sb_clean_to_text(struct printbuf *out, struct bch_sb *sb,
 	struct bch_sb_field_clean *clean = field_to_type(f, clean);
 	struct jset_entry *entry;
 
-	prt_printf(out, "flags:          %x",	le32_to_cpu(clean->flags));
-	prt_newline(out);
-	prt_printf(out, "journal_seq:    %llu",	le64_to_cpu(clean->journal_seq));
-	prt_newline(out);
+	prt_printf(out, "flags:          %x\n",		le32_to_cpu(clean->flags));
+	prt_printf(out, "journal_seq:    %llu\n",	le64_to_cpu(clean->journal_seq));
 
 	for (entry = clean->start;
 	     entry != vstruct_end(&clean->field);
 	     entry = vstruct_next(entry)) {
+		if ((void *) vstruct_next(entry) > vstruct_end(&clean->field))
+			break;
+
 		if (entry->type == BCH_JSET_ENTRY_btree_keys &&
 		    !entry->u64s)
 			continue;
@@ -386,6 +389,8 @@ void bch2_fs_mark_clean(struct bch_fs *c)
 		goto out;
 	}
 
+	bch2_journal_pos_from_member_info_set(c);
+
 	bch2_write_super(c);
 out:
 	mutex_unlock(&c->sb_lock);
diff --git a/fs/bcachefs/sb-counters.c b/fs/bcachefs/sb-counters.c
index 7dc898761b..6992e74691 100644
--- a/fs/bcachefs/sb-counters.c
+++ b/fs/bcachefs/sb-counters.c
@@ -20,9 +20,8 @@ static size_t bch2_sb_counter_nr_entries(struct bch_sb_field_counters *ctrs)
 	return (__le64 *) vstruct_end(&ctrs->field) - &ctrs->d[0];
 };
 
-static int bch2_sb_counters_validate(struct bch_sb *sb,
-				     struct bch_sb_field *f,
-				     struct printbuf *err)
+static int bch2_sb_counters_validate(struct bch_sb *sb, struct bch_sb_field *f,
+				enum bch_validate_flags flags, struct printbuf *err)
 {
 	return 0;
 };
@@ -31,19 +30,12 @@ static void bch2_sb_counters_to_text(struct printbuf *out, struct bch_sb *sb,
 			      struct bch_sb_field *f)
 {
 	struct bch_sb_field_counters *ctrs = field_to_type(f, counters);
-	unsigned int i;
 	unsigned int nr = bch2_sb_counter_nr_entries(ctrs);
 
-	for (i = 0; i < nr; i++) {
-		if (i < BCH_COUNTER_NR)
-			prt_printf(out, "%s ", bch2_counter_names[i]);
-		else
-			prt_printf(out, "(unknown)");
-
-		prt_tab(out);
-		prt_printf(out, "%llu", le64_to_cpu(ctrs->d[i]));
-		prt_newline(out);
-	}
+	for (unsigned i = 0; i < nr; i++)
+		prt_printf(out, "%s \t%llu\n",
+			   i < BCH_COUNTER_NR ? bch2_counter_names[i] : "(unknown)",
+			   le64_to_cpu(ctrs->d[i]));
 };
 
 int bch2_sb_counters_to_cpu(struct bch_fs *c)
diff --git a/fs/bcachefs/sb-downgrade.c b/fs/bcachefs/sb-downgrade.c
index 441dcb1bf1..4710b61631 100644
--- a/fs/bcachefs/sb-downgrade.c
+++ b/fs/bcachefs/sb-downgrade.c
@@ -7,7 +7,7 @@
 
 #include "bcachefs.h"
 #include "darray.h"
-#include "recovery.h"
+#include "recovery_passes.h"
 #include "sb-downgrade.h"
 #include "sb-errors.h"
 #include "super-io.h"
@@ -45,7 +45,16 @@
 	  BIT_ULL(BCH_RECOVERY_PASS_check_inodes),		\
 	  BCH_FSCK_ERR_unlinked_inode_not_on_deleted_list)	\
 	x(rebalance_work,					\
-	  BIT_ULL(BCH_RECOVERY_PASS_set_fs_needs_rebalance))
+	  BIT_ULL(BCH_RECOVERY_PASS_set_fs_needs_rebalance))	\
+	x(subvolume_fs_parent,					\
+	  BIT_ULL(BCH_RECOVERY_PASS_check_dirents),		\
+	  BCH_FSCK_ERR_subvol_fs_path_parent_wrong)		\
+	x(btree_subvolume_children,				\
+	  BIT_ULL(BCH_RECOVERY_PASS_check_subvols),		\
+	  BCH_FSCK_ERR_subvol_children_not_set)			\
+	x(mi_btree_bitmap,					\
+	  BIT_ULL(BCH_RECOVERY_PASS_check_allocations),		\
+	  BCH_FSCK_ERR_btree_bitmap_not_marked)
 
 #define DOWNGRADE_TABLE()
 
@@ -125,15 +134,32 @@ downgrade_entry_next_c(const struct bch_sb_field_downgrade_entry *e)
 #define for_each_downgrade_entry(_d, _i)						\
 	for (const struct bch_sb_field_downgrade_entry *_i = (_d)->entries;		\
 	     (void *) _i	< vstruct_end(&(_d)->field) &&				\
-	     (void *) &_i->errors[0] < vstruct_end(&(_d)->field);			\
+	     (void *) &_i->errors[0] <= vstruct_end(&(_d)->field) &&			\
+	     (void *) downgrade_entry_next_c(_i) <= vstruct_end(&(_d)->field);		\
 	     _i = downgrade_entry_next_c(_i))
 
 static int bch2_sb_downgrade_validate(struct bch_sb *sb, struct bch_sb_field *f,
-				      struct printbuf *err)
+				      enum bch_validate_flags flags, struct printbuf *err)
 {
 	struct bch_sb_field_downgrade *e = field_to_type(f, downgrade);
 
-	for_each_downgrade_entry(e, i) {
+	for (const struct bch_sb_field_downgrade_entry *i = e->entries;
+	     (void *) i	< vstruct_end(&e->field);
+	     i = downgrade_entry_next_c(i)) {
+		/*
+		 * Careful: sb_field_downgrade_entry is only 2 byte aligned, but
+		 * section sizes are 8 byte aligned - an empty entry spanning
+		 * the end of the section is allowed (and ignored):
+		 */
+		if ((void *) &i->errors[0] > vstruct_end(&e->field))
+			break;
+
+		if (flags & BCH_VALIDATE_write &&
+		    (void *) downgrade_entry_next_c(i) > vstruct_end(&e->field)) {
+			prt_printf(err, "downgrade entry overruns end of superblock section");
+			return -BCH_ERR_invalid_sb_downgrade;
+		}
+
 		if (BCH_VERSION_MAJOR(le16_to_cpu(i->version)) !=
 		    BCH_VERSION_MAJOR(le16_to_cpu(sb->version))) {
 			prt_printf(err, "downgrade entry with mismatched major version (%u != %u)",
@@ -155,19 +181,16 @@ static void bch2_sb_downgrade_to_text(struct printbuf *out, struct bch_sb *sb,
 		printbuf_tabstop_push(out, 16);
 
 	for_each_downgrade_entry(e, i) {
-		prt_str(out, "version:");
-		prt_tab(out);
+		prt_str(out, "version:\t");
 		bch2_version_to_text(out, le16_to_cpu(i->version));
 		prt_newline(out);
 
-		prt_str(out, "recovery passes:");
-		prt_tab(out);
+		prt_str(out, "recovery passes:\t");
 		prt_bitflags(out, bch2_recovery_passes,
 			     bch2_recovery_passes_from_stable(le64_to_cpu(i->recovery_passes[0])));
 		prt_newline(out);
 
-		prt_str(out, "errors:");
-		prt_tab(out);
+		prt_str(out, "errors:\t");
 		bool first = true;
 		for (unsigned j = 0; j < le16_to_cpu(i->nr_errors); j++) {
 			if (!first)
@@ -205,7 +228,7 @@ int bch2_sb_downgrade_update(struct bch_fs *c)
 
 		dst = (void *) &darray_top(table);
 		dst->version = cpu_to_le16(src->version);
-		dst->recovery_passes[0]	= cpu_to_le64(src->recovery_passes);
+		dst->recovery_passes[0]	= cpu_to_le64(bch2_recovery_passes_to_stable(src->recovery_passes));
 		dst->recovery_passes[1]	= 0;
 		dst->nr_errors		= cpu_to_le16(src->nr_errors);
 		for (unsigned i = 0; i < src->nr_errors; i++)
@@ -253,7 +276,7 @@ void bch2_sb_set_downgrade(struct bch_fs *c, unsigned new_minor, unsigned old_mi
 				if (e < BCH_SB_ERR_MAX)
 					__set_bit(e, c->sb.errors_silent);
 				if (e < sizeof(ext->errors_silent) * 8)
-					ext->errors_silent[e / 64] |= cpu_to_le64(BIT_ULL(e % 64));
+					__set_bit_le64(e, ext->errors_silent);
 			}
 		}
 	}
diff --git a/fs/bcachefs/sb-downgrade_format.h b/fs/bcachefs/sb-downgrade_format.h
new file mode 100644
index 0000000000..cffd932be3
--- /dev/null
+++ b/fs/bcachefs/sb-downgrade_format.h
@@ -0,0 +1,17 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _BCACHEFS_SB_DOWNGRADE_FORMAT_H
+#define _BCACHEFS_SB_DOWNGRADE_FORMAT_H
+
+struct bch_sb_field_downgrade_entry {
+	__le16			version;
+	__le64			recovery_passes[2];
+	__le16			nr_errors;
+	__le16			errors[] __counted_by(nr_errors);
+} __packed __aligned(2);
+
+struct bch_sb_field_downgrade {
+	struct bch_sb_field	field;
+	struct bch_sb_field_downgrade_entry entries[];
+};
+
+#endif /* _BCACHEFS_SB_DOWNGRADE_FORMAT_H */
diff --git a/fs/bcachefs/sb-errors.c b/fs/bcachefs/sb-errors.c
index 5f5bcae391..c1270d790e 100644
--- a/fs/bcachefs/sb-errors.c
+++ b/fs/bcachefs/sb-errors.c
@@ -30,7 +30,7 @@ static inline unsigned bch2_sb_field_errors_u64s(unsigned nr)
 }
 
 static int bch2_sb_errors_validate(struct bch_sb *sb, struct bch_sb_field *f,
-				   struct printbuf *err)
+				   enum bch_validate_flags flags, struct printbuf *err)
 {
 	struct bch_sb_field_errors *e = field_to_type(f, errors);
 	unsigned i, nr = bch2_sb_field_errors_nr_entries(e);
@@ -110,19 +110,25 @@ out:
 void bch2_sb_errors_from_cpu(struct bch_fs *c)
 {
 	bch_sb_errors_cpu *src = &c->fsck_error_counts;
-	struct bch_sb_field_errors *dst =
-		bch2_sb_field_resize(&c->disk_sb, errors,
-				     bch2_sb_field_errors_u64s(src->nr));
+	struct bch_sb_field_errors *dst;
 	unsigned i;
 
+	mutex_lock(&c->fsck_error_counts_lock);
+
+	dst = bch2_sb_field_resize(&c->disk_sb, errors,
+				   bch2_sb_field_errors_u64s(src->nr));
+
 	if (!dst)
-		return;
+		goto err;
 
 	for (i = 0; i < src->nr; i++) {
 		SET_BCH_SB_ERROR_ENTRY_ID(&dst->entries[i], src->data[i].id);
 		SET_BCH_SB_ERROR_ENTRY_NR(&dst->entries[i], src->data[i].nr);
 		dst->entries[i].last_error_time = cpu_to_le64(src->data[i].last_error_time);
 	}
+
+err:
+	mutex_unlock(&c->fsck_error_counts_lock);
 }
 
 static int bch2_sb_errors_to_cpu(struct bch_fs *c)
diff --git a/fs/bcachefs/sb-errors_format.h b/fs/bcachefs/sb-errors_format.h
new file mode 100644
index 0000000000..d54121ec09
--- /dev/null
+++ b/fs/bcachefs/sb-errors_format.h
@@ -0,0 +1,310 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _BCACHEFS_SB_ERRORS_FORMAT_H
+#define _BCACHEFS_SB_ERRORS_FORMAT_H
+
+enum bch_fsck_flags {
+	FSCK_CAN_FIX		= 1 << 0,
+	FSCK_CAN_IGNORE		= 1 << 1,
+	FSCK_NEED_FSCK		= 1 << 2,
+	FSCK_NO_RATELIMIT	= 1 << 3,
+	FSCK_AUTOFIX		= 1 << 4,
+};
+
+#define BCH_SB_ERRS()									\
+	x(clean_but_journal_not_empty,				  0,	0)		\
+	x(dirty_but_no_journal_entries,				  1,	0)		\
+	x(dirty_but_no_journal_entries_post_drop_nonflushes,	  2,	0)		\
+	x(sb_clean_journal_seq_mismatch,			  3,	0)		\
+	x(sb_clean_btree_root_mismatch,				  4,	0)		\
+	x(sb_clean_missing,					  5,	0)		\
+	x(jset_unsupported_version,				  6,	0)		\
+	x(jset_unknown_csum,					  7,	0)		\
+	x(jset_last_seq_newer_than_seq,				  8,	0)		\
+	x(jset_past_bucket_end,					  9,	0)		\
+	x(jset_seq_blacklisted,					 10,	0)		\
+	x(journal_entries_missing,				 11,	0)		\
+	x(journal_entry_replicas_not_marked,			 12,	0)		\
+	x(journal_entry_past_jset_end,				 13,	0)		\
+	x(journal_entry_replicas_data_mismatch,			 14,	0)		\
+	x(journal_entry_bkey_u64s_0,				 15,	0)		\
+	x(journal_entry_bkey_past_end,				 16,	0)		\
+	x(journal_entry_bkey_bad_format,			 17,	0)		\
+	x(journal_entry_bkey_invalid,				 18,	0)		\
+	x(journal_entry_btree_root_bad_size,			 19,	0)		\
+	x(journal_entry_blacklist_bad_size,			 20,	0)		\
+	x(journal_entry_blacklist_v2_bad_size,			 21,	0)		\
+	x(journal_entry_blacklist_v2_start_past_end,		 22,	0)		\
+	x(journal_entry_usage_bad_size,				 23,	0)		\
+	x(journal_entry_data_usage_bad_size,			 24,	0)		\
+	x(journal_entry_clock_bad_size,				 25,	0)		\
+	x(journal_entry_clock_bad_rw,				 26,	0)		\
+	x(journal_entry_dev_usage_bad_size,			 27,	0)		\
+	x(journal_entry_dev_usage_bad_dev,			 28,	0)		\
+	x(journal_entry_dev_usage_bad_pad,			 29,	0)		\
+	x(btree_node_unreadable,				 30,	0)		\
+	x(btree_node_fault_injected,				 31,	0)		\
+	x(btree_node_bad_magic,					 32,	0)		\
+	x(btree_node_bad_seq,					 33,	0)		\
+	x(btree_node_unsupported_version,			 34,	0)		\
+	x(btree_node_bset_older_than_sb_min,			 35,	0)		\
+	x(btree_node_bset_newer_than_sb,			 36,	0)		\
+	x(btree_node_data_missing,				 37,	0)		\
+	x(btree_node_bset_after_end,				 38,	0)		\
+	x(btree_node_replicas_sectors_written_mismatch,		 39,	0)		\
+	x(btree_node_replicas_data_mismatch,			 40,	0)		\
+	x(bset_unknown_csum,					 41,	0)		\
+	x(bset_bad_csum,					 42,	0)		\
+	x(bset_past_end_of_btree_node,				 43,	0)		\
+	x(bset_wrong_sector_offset,				 44,	0)		\
+	x(bset_empty,						 45,	0)		\
+	x(bset_bad_seq,						 46,	0)		\
+	x(bset_blacklisted_journal_seq,				 47,	0)		\
+	x(first_bset_blacklisted_journal_seq,			 48,	0)		\
+	x(btree_node_bad_btree,					 49,	0)		\
+	x(btree_node_bad_level,					 50,	0)		\
+	x(btree_node_bad_min_key,				 51,	0)		\
+	x(btree_node_bad_max_key,				 52,	0)		\
+	x(btree_node_bad_format,				 53,	0)		\
+	x(btree_node_bkey_past_bset_end,			 54,	0)		\
+	x(btree_node_bkey_bad_format,				 55,	0)		\
+	x(btree_node_bad_bkey,					 56,	0)		\
+	x(btree_node_bkey_out_of_order,				 57,	0)		\
+	x(btree_root_bkey_invalid,				 58,	0)		\
+	x(btree_root_read_error,				 59,	0)		\
+	x(btree_root_bad_min_key,				 60,	0)		\
+	x(btree_root_bad_max_key,				 61,	0)		\
+	x(btree_node_read_error,				 62,	0)		\
+	x(btree_node_topology_bad_min_key,			 63,	0)		\
+	x(btree_node_topology_bad_max_key,			 64,	0)		\
+	x(btree_node_topology_overwritten_by_prev_node,		 65,	0)		\
+	x(btree_node_topology_overwritten_by_next_node,		 66,	0)		\
+	x(btree_node_topology_interior_node_empty,		 67,	0)		\
+	x(fs_usage_hidden_wrong,				 68,	FSCK_AUTOFIX)	\
+	x(fs_usage_btree_wrong,					 69,	FSCK_AUTOFIX)	\
+	x(fs_usage_data_wrong,					 70,	FSCK_AUTOFIX)	\
+	x(fs_usage_cached_wrong,				 71,	FSCK_AUTOFIX)	\
+	x(fs_usage_reserved_wrong,				 72,	FSCK_AUTOFIX)	\
+	x(fs_usage_persistent_reserved_wrong,			 73,	FSCK_AUTOFIX)	\
+	x(fs_usage_nr_inodes_wrong,				 74,	FSCK_AUTOFIX)	\
+	x(fs_usage_replicas_wrong,				 75,	FSCK_AUTOFIX)	\
+	x(dev_usage_buckets_wrong,				 76,	FSCK_AUTOFIX)	\
+	x(dev_usage_sectors_wrong,				 77,	FSCK_AUTOFIX)	\
+	x(dev_usage_fragmented_wrong,				 78,	FSCK_AUTOFIX)	\
+	x(dev_usage_buckets_ec_wrong,				 79,	FSCK_AUTOFIX)	\
+	x(bkey_version_in_future,				 80,	0)		\
+	x(bkey_u64s_too_small,					 81,	0)		\
+	x(bkey_invalid_type_for_btree,				 82,	0)		\
+	x(bkey_extent_size_zero,				 83,	0)		\
+	x(bkey_extent_size_greater_than_offset,			 84,	0)		\
+	x(bkey_size_nonzero,					 85,	0)		\
+	x(bkey_snapshot_nonzero,				 86,	0)		\
+	x(bkey_snapshot_zero,					 87,	0)		\
+	x(bkey_at_pos_max,					 88,	0)		\
+	x(bkey_before_start_of_btree_node,			 89,	0)		\
+	x(bkey_after_end_of_btree_node,				 90,	0)		\
+	x(bkey_val_size_nonzero,				 91,	0)		\
+	x(bkey_val_size_too_small,				 92,	0)		\
+	x(alloc_v1_val_size_bad,				 93,	0)		\
+	x(alloc_v2_unpack_error,				 94,	0)		\
+	x(alloc_v3_unpack_error,				 95,	0)		\
+	x(alloc_v4_val_size_bad,				 96,	0)		\
+	x(alloc_v4_backpointers_start_bad,			 97,	0)		\
+	x(alloc_key_data_type_bad,				 98,	0)		\
+	x(alloc_key_empty_but_have_data,			 99,	0)		\
+	x(alloc_key_dirty_sectors_0,				100,	0)		\
+	x(alloc_key_data_type_inconsistency,			101,	0)		\
+	x(alloc_key_to_missing_dev_bucket,			102,	0)		\
+	x(alloc_key_cached_inconsistency,			103,	0)		\
+	x(alloc_key_cached_but_read_time_zero,			104,	0)		\
+	x(alloc_key_to_missing_lru_entry,			105,	0)		\
+	x(alloc_key_data_type_wrong,				106,	FSCK_AUTOFIX)	\
+	x(alloc_key_gen_wrong,					107,	FSCK_AUTOFIX)	\
+	x(alloc_key_dirty_sectors_wrong,			108,	FSCK_AUTOFIX)	\
+	x(alloc_key_cached_sectors_wrong,			109,	FSCK_AUTOFIX)	\
+	x(alloc_key_stripe_wrong,				110,	FSCK_AUTOFIX)	\
+	x(alloc_key_stripe_redundancy_wrong,			111,	FSCK_AUTOFIX)	\
+	x(bucket_sector_count_overflow,				112,	0)		\
+	x(bucket_metadata_type_mismatch,			113,	0)		\
+	x(need_discard_key_wrong,				114,	0)		\
+	x(freespace_key_wrong,					115,	0)		\
+	x(freespace_hole_missing,				116,	0)		\
+	x(bucket_gens_val_size_bad,				117,	0)		\
+	x(bucket_gens_key_wrong,				118,	0)		\
+	x(bucket_gens_hole_wrong,				119,	0)		\
+	x(bucket_gens_to_invalid_dev,				120,	0)		\
+	x(bucket_gens_to_invalid_buckets,			121,	0)		\
+	x(bucket_gens_nonzero_for_invalid_buckets,		122,	0)		\
+	x(need_discard_freespace_key_to_invalid_dev_bucket,	123,	0)		\
+	x(need_discard_freespace_key_bad,			124,	0)		\
+	x(backpointer_bucket_offset_wrong,			125,	0)		\
+	x(backpointer_to_missing_device,			126,	0)		\
+	x(backpointer_to_missing_alloc,				127,	0)		\
+	x(backpointer_to_missing_ptr,				128,	0)		\
+	x(lru_entry_at_time_0,					129,	0)		\
+	x(lru_entry_to_invalid_bucket,				130,	0)		\
+	x(lru_entry_bad,					131,	0)		\
+	x(btree_ptr_val_too_big,				132,	0)		\
+	x(btree_ptr_v2_val_too_big,				133,	0)		\
+	x(btree_ptr_has_non_ptr,				134,	0)		\
+	x(extent_ptrs_invalid_entry,				135,	0)		\
+	x(extent_ptrs_no_ptrs,					136,	0)		\
+	x(extent_ptrs_too_many_ptrs,				137,	0)		\
+	x(extent_ptrs_redundant_crc,				138,	0)		\
+	x(extent_ptrs_redundant_stripe,				139,	0)		\
+	x(extent_ptrs_unwritten,				140,	0)		\
+	x(extent_ptrs_written_and_unwritten,			141,	0)		\
+	x(ptr_to_invalid_device,				142,	0)		\
+	x(ptr_to_duplicate_device,				143,	0)		\
+	x(ptr_after_last_bucket,				144,	0)		\
+	x(ptr_before_first_bucket,				145,	0)		\
+	x(ptr_spans_multiple_buckets,				146,	0)		\
+	x(ptr_to_missing_backpointer,				147,	0)		\
+	x(ptr_to_missing_alloc_key,				148,	0)		\
+	x(ptr_to_missing_replicas_entry,			149,	0)		\
+	x(ptr_to_missing_stripe,				150,	0)		\
+	x(ptr_to_incorrect_stripe,				151,	0)		\
+	x(ptr_gen_newer_than_bucket_gen,			152,	0)		\
+	x(ptr_too_stale,					153,	0)		\
+	x(stale_dirty_ptr,					154,	0)		\
+	x(ptr_bucket_data_type_mismatch,			155,	0)		\
+	x(ptr_cached_and_erasure_coded,				156,	0)		\
+	x(ptr_crc_uncompressed_size_too_small,			157,	0)		\
+	x(ptr_crc_csum_type_unknown,				158,	0)		\
+	x(ptr_crc_compression_type_unknown,			159,	0)		\
+	x(ptr_crc_redundant,					160,	0)		\
+	x(ptr_crc_uncompressed_size_too_big,			161,	0)		\
+	x(ptr_crc_nonce_mismatch,				162,	0)		\
+	x(ptr_stripe_redundant,					163,	0)		\
+	x(reservation_key_nr_replicas_invalid,			164,	0)		\
+	x(reflink_v_refcount_wrong,				165,	0)		\
+	x(reflink_p_to_missing_reflink_v,			166,	0)		\
+	x(stripe_pos_bad,					167,	0)		\
+	x(stripe_val_size_bad,					168,	0)		\
+	x(stripe_sector_count_wrong,				169,	0)		\
+	x(snapshot_tree_pos_bad,				170,	0)		\
+	x(snapshot_tree_to_missing_snapshot,			171,	0)		\
+	x(snapshot_tree_to_missing_subvol,			172,	0)		\
+	x(snapshot_tree_to_wrong_subvol,			173,	0)		\
+	x(snapshot_tree_to_snapshot_subvol,			174,	0)		\
+	x(snapshot_pos_bad,					175,	0)		\
+	x(snapshot_parent_bad,					176,	0)		\
+	x(snapshot_children_not_normalized,			177,	0)		\
+	x(snapshot_child_duplicate,				178,	0)		\
+	x(snapshot_child_bad,					179,	0)		\
+	x(snapshot_skiplist_not_normalized,			180,	0)		\
+	x(snapshot_skiplist_bad,				181,	0)		\
+	x(snapshot_should_not_have_subvol,			182,	0)		\
+	x(snapshot_to_bad_snapshot_tree,			183,	0)		\
+	x(snapshot_bad_depth,					184,	0)		\
+	x(snapshot_bad_skiplist,				185,	0)		\
+	x(subvol_pos_bad,					186,	0)		\
+	x(subvol_not_master_and_not_snapshot,			187,	0)		\
+	x(subvol_to_missing_root,				188,	0)		\
+	x(subvol_root_wrong_bi_subvol,				189,	0)		\
+	x(bkey_in_missing_snapshot,				190,	0)		\
+	x(inode_pos_inode_nonzero,				191,	0)		\
+	x(inode_pos_blockdev_range,				192,	0)		\
+	x(inode_unpack_error,					193,	0)		\
+	x(inode_str_hash_invalid,				194,	0)		\
+	x(inode_v3_fields_start_bad,				195,	0)		\
+	x(inode_snapshot_mismatch,				196,	0)		\
+	x(inode_unlinked_but_clean,				197,	0)		\
+	x(inode_unlinked_but_nlink_nonzero,			198,	0)		\
+	x(inode_checksum_type_invalid,				199,	0)		\
+	x(inode_compression_type_invalid,			200,	0)		\
+	x(inode_subvol_root_but_not_dir,			201,	0)		\
+	x(inode_i_size_dirty_but_clean,				202,	0)		\
+	x(inode_i_sectors_dirty_but_clean,			203,	0)		\
+	x(inode_i_sectors_wrong,				204,	0)		\
+	x(inode_dir_wrong_nlink,				205,	0)		\
+	x(inode_dir_multiple_links,				206,	0)		\
+	x(inode_multiple_links_but_nlink_0,			207,	0)		\
+	x(inode_wrong_backpointer,				208,	0)		\
+	x(inode_wrong_nlink,					209,	0)		\
+	x(inode_unreachable,					210,	0)		\
+	x(deleted_inode_but_clean,				211,	0)		\
+	x(deleted_inode_missing,				212,	0)		\
+	x(deleted_inode_is_dir,					213,	0)		\
+	x(deleted_inode_not_unlinked,				214,	0)		\
+	x(extent_overlapping,					215,	0)		\
+	x(extent_in_missing_inode,				216,	0)		\
+	x(extent_in_non_reg_inode,				217,	0)		\
+	x(extent_past_end_of_inode,				218,	0)		\
+	x(dirent_empty_name,					219,	0)		\
+	x(dirent_val_too_big,					220,	0)		\
+	x(dirent_name_too_long,					221,	0)		\
+	x(dirent_name_embedded_nul,				222,	0)		\
+	x(dirent_name_dot_or_dotdot,				223,	0)		\
+	x(dirent_name_has_slash,				224,	0)		\
+	x(dirent_d_type_wrong,					225,	0)		\
+	x(inode_bi_parent_wrong,				226,	0)		\
+	x(dirent_in_missing_dir_inode,				227,	0)		\
+	x(dirent_in_non_dir_inode,				228,	0)		\
+	x(dirent_to_missing_inode,				229,	0)		\
+	x(dirent_to_missing_subvol,				230,	0)		\
+	x(dirent_to_itself,					231,	0)		\
+	x(quota_type_invalid,					232,	0)		\
+	x(xattr_val_size_too_small,				233,	0)		\
+	x(xattr_val_size_too_big,				234,	0)		\
+	x(xattr_invalid_type,					235,	0)		\
+	x(xattr_name_invalid_chars,				236,	0)		\
+	x(xattr_in_missing_inode,				237,	0)		\
+	x(root_subvol_missing,					238,	0)		\
+	x(root_dir_missing,					239,	0)		\
+	x(root_inode_not_dir,					240,	0)		\
+	x(dir_loop,						241,	0)		\
+	x(hash_table_key_duplicate,				242,	0)		\
+	x(hash_table_key_wrong_offset,				243,	0)		\
+	x(unlinked_inode_not_on_deleted_list,			244,	0)		\
+	x(reflink_p_front_pad_bad,				245,	0)		\
+	x(journal_entry_dup_same_device,			246,	0)		\
+	x(inode_bi_subvol_missing,				247,	0)		\
+	x(inode_bi_subvol_wrong,				248,	0)		\
+	x(inode_points_to_missing_dirent,			249,	0)		\
+	x(inode_points_to_wrong_dirent,				250,	0)		\
+	x(inode_bi_parent_nonzero,				251,	0)		\
+	x(dirent_to_missing_parent_subvol,			252,	0)		\
+	x(dirent_not_visible_in_parent_subvol,			253,	0)		\
+	x(subvol_fs_path_parent_wrong,				254,	0)		\
+	x(subvol_root_fs_path_parent_nonzero,			255,	0)		\
+	x(subvol_children_not_set,				256,	0)		\
+	x(subvol_children_bad,					257,	0)		\
+	x(subvol_loop,						258,	0)		\
+	x(subvol_unreachable,					259,	0)		\
+	x(btree_node_bkey_bad_u64s,				260,	0)		\
+	x(btree_node_topology_empty_interior_node,		261,	0)		\
+	x(btree_ptr_v2_min_key_bad,				262,	0)		\
+	x(btree_root_unreadable_and_scan_found_nothing,		263,	0)		\
+	x(snapshot_node_missing,				264,	0)		\
+	x(dup_backpointer_to_bad_csum_extent,			265,	0)		\
+	x(btree_bitmap_not_marked,				266,	0)		\
+	x(sb_clean_entry_overrun,				267,	0)		\
+	x(btree_ptr_v2_written_0,				268,	0)		\
+	x(subvol_snapshot_bad,					269,	0)		\
+	x(subvol_inode_bad,					270,	0)		\
+	x(alloc_key_stripe_sectors_wrong,			271,	0)		\
+	x(accounting_mismatch,					272,	0)		\
+	x(accounting_replicas_not_marked,			273,	0)		\
+	x(invalid_btree_id,					274,	0)		\
+	x(alloc_key_io_time_bad,				275,	0)		\
+	x(alloc_key_fragmentation_lru_wrong,			276,	FSCK_AUTOFIX)
+
+enum bch_sb_error_id {
+#define x(t, n, ...) BCH_FSCK_ERR_##t = n,
+	BCH_SB_ERRS()
+#undef x
+	BCH_SB_ERR_MAX
+};
+
+struct bch_sb_field_errors {
+	struct bch_sb_field	field;
+	struct bch_sb_field_error_entry {
+		__le64		v;
+		__le64		last_error_time;
+	}			entries[];
+};
+
+LE64_BITMASK(BCH_SB_ERROR_ENTRY_ID,	struct bch_sb_field_error_entry, v,  0, 16);
+LE64_BITMASK(BCH_SB_ERROR_ENTRY_NR,	struct bch_sb_field_error_entry, v, 16, 64);
+
+#endif /* _BCACHEFS_SB_ERRORS_FORMAT_H */
diff --git a/fs/bcachefs/sb-errors_types.h b/fs/bcachefs/sb-errors_types.h
index c08aacdfd0..40325239c3 100644
--- a/fs/bcachefs/sb-errors_types.h
+++ b/fs/bcachefs/sb-errors_types.h
@@ -4,261 +4,6 @@
 
 #include "darray.h"
 
-#define BCH_SB_ERRS()							\
-	x(clean_but_journal_not_empty,				0)	\
-	x(dirty_but_no_journal_entries,				1)	\
-	x(dirty_but_no_journal_entries_post_drop_nonflushes,	2)	\
-	x(sb_clean_journal_seq_mismatch,			3)	\
-	x(sb_clean_btree_root_mismatch,				4)	\
-	x(sb_clean_missing,					5)	\
-	x(jset_unsupported_version,				6)	\
-	x(jset_unknown_csum,					7)	\
-	x(jset_last_seq_newer_than_seq,				8)	\
-	x(jset_past_bucket_end,					9)	\
-	x(jset_seq_blacklisted,					10)	\
-	x(journal_entries_missing,				11)	\
-	x(journal_entry_replicas_not_marked,			12)	\
-	x(journal_entry_past_jset_end,				13)	\
-	x(journal_entry_replicas_data_mismatch,			14)	\
-	x(journal_entry_bkey_u64s_0,				15)	\
-	x(journal_entry_bkey_past_end,				16)	\
-	x(journal_entry_bkey_bad_format,			17)	\
-	x(journal_entry_bkey_invalid,				18)	\
-	x(journal_entry_btree_root_bad_size,			19)	\
-	x(journal_entry_blacklist_bad_size,			20)	\
-	x(journal_entry_blacklist_v2_bad_size,			21)	\
-	x(journal_entry_blacklist_v2_start_past_end,		22)	\
-	x(journal_entry_usage_bad_size,				23)	\
-	x(journal_entry_data_usage_bad_size,			24)	\
-	x(journal_entry_clock_bad_size,				25)	\
-	x(journal_entry_clock_bad_rw,				26)	\
-	x(journal_entry_dev_usage_bad_size,			27)	\
-	x(journal_entry_dev_usage_bad_dev,			28)	\
-	x(journal_entry_dev_usage_bad_pad,			29)	\
-	x(btree_node_unreadable,				30)	\
-	x(btree_node_fault_injected,				31)	\
-	x(btree_node_bad_magic,					32)	\
-	x(btree_node_bad_seq,					33)	\
-	x(btree_node_unsupported_version,			34)	\
-	x(btree_node_bset_older_than_sb_min,			35)	\
-	x(btree_node_bset_newer_than_sb,			36)	\
-	x(btree_node_data_missing,				37)	\
-	x(btree_node_bset_after_end,				38)	\
-	x(btree_node_replicas_sectors_written_mismatch,		39)	\
-	x(btree_node_replicas_data_mismatch,			40)	\
-	x(bset_unknown_csum,					41)	\
-	x(bset_bad_csum,					42)	\
-	x(bset_past_end_of_btree_node,				43)	\
-	x(bset_wrong_sector_offset,				44)	\
-	x(bset_empty,						45)	\
-	x(bset_bad_seq,						46)	\
-	x(bset_blacklisted_journal_seq,				47)	\
-	x(first_bset_blacklisted_journal_seq,			48)	\
-	x(btree_node_bad_btree,					49)	\
-	x(btree_node_bad_level,					50)	\
-	x(btree_node_bad_min_key,				51)	\
-	x(btree_node_bad_max_key,				52)	\
-	x(btree_node_bad_format,				53)	\
-	x(btree_node_bkey_past_bset_end,			54)	\
-	x(btree_node_bkey_bad_format,				55)	\
-	x(btree_node_bad_bkey,					56)	\
-	x(btree_node_bkey_out_of_order,				57)	\
-	x(btree_root_bkey_invalid,				58)	\
-	x(btree_root_read_error,				59)	\
-	x(btree_root_bad_min_key,				60)	\
-	x(btree_root_bad_max_key,				61)	\
-	x(btree_node_read_error,				62)	\
-	x(btree_node_topology_bad_min_key,			63)	\
-	x(btree_node_topology_bad_max_key,			64)	\
-	x(btree_node_topology_overwritten_by_prev_node,		65)	\
-	x(btree_node_topology_overwritten_by_next_node,		66)	\
-	x(btree_node_topology_interior_node_empty,		67)	\
-	x(fs_usage_hidden_wrong,				68)	\
-	x(fs_usage_btree_wrong,					69)	\
-	x(fs_usage_data_wrong,					70)	\
-	x(fs_usage_cached_wrong,				71)	\
-	x(fs_usage_reserved_wrong,				72)	\
-	x(fs_usage_persistent_reserved_wrong,			73)	\
-	x(fs_usage_nr_inodes_wrong,				74)	\
-	x(fs_usage_replicas_wrong,				75)	\
-	x(dev_usage_buckets_wrong,				76)	\
-	x(dev_usage_sectors_wrong,				77)	\
-	x(dev_usage_fragmented_wrong,				78)	\
-	x(dev_usage_buckets_ec_wrong,				79)	\
-	x(bkey_version_in_future,				80)	\
-	x(bkey_u64s_too_small,					81)	\
-	x(bkey_invalid_type_for_btree,				82)	\
-	x(bkey_extent_size_zero,				83)	\
-	x(bkey_extent_size_greater_than_offset,			84)	\
-	x(bkey_size_nonzero,					85)	\
-	x(bkey_snapshot_nonzero,				86)	\
-	x(bkey_snapshot_zero,					87)	\
-	x(bkey_at_pos_max,					88)	\
-	x(bkey_before_start_of_btree_node,			89)	\
-	x(bkey_after_end_of_btree_node,				90)	\
-	x(bkey_val_size_nonzero,				91)	\
-	x(bkey_val_size_too_small,				92)	\
-	x(alloc_v1_val_size_bad,				93)	\
-	x(alloc_v2_unpack_error,				94)	\
-	x(alloc_v3_unpack_error,				95)	\
-	x(alloc_v4_val_size_bad,				96)	\
-	x(alloc_v4_backpointers_start_bad,			97)	\
-	x(alloc_key_data_type_bad,				98)	\
-	x(alloc_key_empty_but_have_data,			99)	\
-	x(alloc_key_dirty_sectors_0,				100)	\
-	x(alloc_key_data_type_inconsistency,			101)	\
-	x(alloc_key_to_missing_dev_bucket,			102)	\
-	x(alloc_key_cached_inconsistency,			103)	\
-	x(alloc_key_cached_but_read_time_zero,			104)	\
-	x(alloc_key_to_missing_lru_entry,			105)	\
-	x(alloc_key_data_type_wrong,				106)	\
-	x(alloc_key_gen_wrong,					107)	\
-	x(alloc_key_dirty_sectors_wrong,			108)	\
-	x(alloc_key_cached_sectors_wrong,			109)	\
-	x(alloc_key_stripe_wrong,				110)	\
-	x(alloc_key_stripe_redundancy_wrong,			111)	\
-	x(bucket_sector_count_overflow,				112)	\
-	x(bucket_metadata_type_mismatch,			113)	\
-	x(need_discard_key_wrong,				114)	\
-	x(freespace_key_wrong,					115)	\
-	x(freespace_hole_missing,				116)	\
-	x(bucket_gens_val_size_bad,				117)	\
-	x(bucket_gens_key_wrong,				118)	\
-	x(bucket_gens_hole_wrong,				119)	\
-	x(bucket_gens_to_invalid_dev,				120)	\
-	x(bucket_gens_to_invalid_buckets,			121)	\
-	x(bucket_gens_nonzero_for_invalid_buckets,		122)	\
-	x(need_discard_freespace_key_to_invalid_dev_bucket,	123)	\
-	x(need_discard_freespace_key_bad,			124)	\
-	x(backpointer_pos_wrong,				125)	\
-	x(backpointer_to_missing_device,			126)	\
-	x(backpointer_to_missing_alloc,				127)	\
-	x(backpointer_to_missing_ptr,				128)	\
-	x(lru_entry_at_time_0,					129)	\
-	x(lru_entry_to_invalid_bucket,				130)	\
-	x(lru_entry_bad,					131)	\
-	x(btree_ptr_val_too_big,				132)	\
-	x(btree_ptr_v2_val_too_big,				133)	\
-	x(btree_ptr_has_non_ptr,				134)	\
-	x(extent_ptrs_invalid_entry,				135)	\
-	x(extent_ptrs_no_ptrs,					136)	\
-	x(extent_ptrs_too_many_ptrs,				137)	\
-	x(extent_ptrs_redundant_crc,				138)	\
-	x(extent_ptrs_redundant_stripe,				139)	\
-	x(extent_ptrs_unwritten,				140)	\
-	x(extent_ptrs_written_and_unwritten,			141)	\
-	x(ptr_to_invalid_device,				142)	\
-	x(ptr_to_duplicate_device,				143)	\
-	x(ptr_after_last_bucket,				144)	\
-	x(ptr_before_first_bucket,				145)	\
-	x(ptr_spans_multiple_buckets,				146)	\
-	x(ptr_to_missing_backpointer,				147)	\
-	x(ptr_to_missing_alloc_key,				148)	\
-	x(ptr_to_missing_replicas_entry,			149)	\
-	x(ptr_to_missing_stripe,				150)	\
-	x(ptr_to_incorrect_stripe,				151)	\
-	x(ptr_gen_newer_than_bucket_gen,			152)	\
-	x(ptr_too_stale,					153)	\
-	x(stale_dirty_ptr,					154)	\
-	x(ptr_bucket_data_type_mismatch,			155)	\
-	x(ptr_cached_and_erasure_coded,				156)	\
-	x(ptr_crc_uncompressed_size_too_small,			157)	\
-	x(ptr_crc_csum_type_unknown,				158)	\
-	x(ptr_crc_compression_type_unknown,			159)	\
-	x(ptr_crc_redundant,					160)	\
-	x(ptr_crc_uncompressed_size_too_big,			161)	\
-	x(ptr_crc_nonce_mismatch,				162)	\
-	x(ptr_stripe_redundant,					163)	\
-	x(reservation_key_nr_replicas_invalid,			164)	\
-	x(reflink_v_refcount_wrong,				165)	\
-	x(reflink_p_to_missing_reflink_v,			166)	\
-	x(stripe_pos_bad,					167)	\
-	x(stripe_val_size_bad,					168)	\
-	x(stripe_sector_count_wrong,				169)	\
-	x(snapshot_tree_pos_bad,				170)	\
-	x(snapshot_tree_to_missing_snapshot,			171)	\
-	x(snapshot_tree_to_missing_subvol,			172)	\
-	x(snapshot_tree_to_wrong_subvol,			173)	\
-	x(snapshot_tree_to_snapshot_subvol,			174)	\
-	x(snapshot_pos_bad,					175)	\
-	x(snapshot_parent_bad,					176)	\
-	x(snapshot_children_not_normalized,			177)	\
-	x(snapshot_child_duplicate,				178)	\
-	x(snapshot_child_bad,					179)	\
-	x(snapshot_skiplist_not_normalized,			180)	\
-	x(snapshot_skiplist_bad,				181)	\
-	x(snapshot_should_not_have_subvol,			182)	\
-	x(snapshot_to_bad_snapshot_tree,			183)	\
-	x(snapshot_bad_depth,					184)	\
-	x(snapshot_bad_skiplist,				185)	\
-	x(subvol_pos_bad,					186)	\
-	x(subvol_not_master_and_not_snapshot,			187)	\
-	x(subvol_to_missing_root,				188)	\
-	x(subvol_root_wrong_bi_subvol,				189)	\
-	x(bkey_in_missing_snapshot,				190)	\
-	x(inode_pos_inode_nonzero,				191)	\
-	x(inode_pos_blockdev_range,				192)	\
-	x(inode_unpack_error,					193)	\
-	x(inode_str_hash_invalid,				194)	\
-	x(inode_v3_fields_start_bad,				195)	\
-	x(inode_snapshot_mismatch,				196)	\
-	x(inode_unlinked_but_clean,				197)	\
-	x(inode_unlinked_but_nlink_nonzero,			198)	\
-	x(inode_checksum_type_invalid,				199)	\
-	x(inode_compression_type_invalid,			200)	\
-	x(inode_subvol_root_but_not_dir,			201)	\
-	x(inode_i_size_dirty_but_clean,				202)	\
-	x(inode_i_sectors_dirty_but_clean,			203)	\
-	x(inode_i_sectors_wrong,				204)	\
-	x(inode_dir_wrong_nlink,				205)	\
-	x(inode_dir_multiple_links,				206)	\
-	x(inode_multiple_links_but_nlink_0,			207)	\
-	x(inode_wrong_backpointer,				208)	\
-	x(inode_wrong_nlink,					209)	\
-	x(inode_unreachable,					210)	\
-	x(deleted_inode_but_clean,				211)	\
-	x(deleted_inode_missing,				212)	\
-	x(deleted_inode_is_dir,					213)	\
-	x(deleted_inode_not_unlinked,				214)	\
-	x(extent_overlapping,					215)	\
-	x(extent_in_missing_inode,				216)	\
-	x(extent_in_non_reg_inode,				217)	\
-	x(extent_past_end_of_inode,				218)	\
-	x(dirent_empty_name,					219)	\
-	x(dirent_val_too_big,					220)	\
-	x(dirent_name_too_long,					221)	\
-	x(dirent_name_embedded_nul,				222)	\
-	x(dirent_name_dot_or_dotdot,				223)	\
-	x(dirent_name_has_slash,				224)	\
-	x(dirent_d_type_wrong,					225)	\
-	x(dirent_d_parent_subvol_wrong,				226)	\
-	x(dirent_in_missing_dir_inode,				227)	\
-	x(dirent_in_non_dir_inode,				228)	\
-	x(dirent_to_missing_inode,				229)	\
-	x(dirent_to_missing_subvol,				230)	\
-	x(dirent_to_itself,					231)	\
-	x(quota_type_invalid,					232)	\
-	x(xattr_val_size_too_small,				233)	\
-	x(xattr_val_size_too_big,				234)	\
-	x(xattr_invalid_type,					235)	\
-	x(xattr_name_invalid_chars,				236)	\
-	x(xattr_in_missing_inode,				237)	\
-	x(root_subvol_missing,					238)	\
-	x(root_dir_missing,					239)	\
-	x(root_inode_not_dir,					240)	\
-	x(dir_loop,						241)	\
-	x(hash_table_key_duplicate,				242)	\
-	x(hash_table_key_wrong_offset,				243)	\
-	x(unlinked_inode_not_on_deleted_list,			244)	\
-	x(reflink_p_front_pad_bad,				245)
-
-enum bch_sb_error_id {
-#define x(t, n) BCH_FSCK_ERR_##t = n,
-	BCH_SB_ERRS()
-#undef x
-	BCH_SB_ERR_MAX
-};
-
 struct bch_sb_error_entry_cpu {
 	u64			id:16,
 				nr:48;
@@ -268,4 +13,3 @@ struct bch_sb_error_entry_cpu {
 typedef DARRAY(struct bch_sb_error_entry_cpu) bch_sb_errors_cpu;
 
 #endif /* _BCACHEFS_SB_ERRORS_TYPES_H */
-
diff --git a/fs/bcachefs/sb-members.c b/fs/bcachefs/sb-members.c
index eff5ce18c6..39196f2a41 100644
--- a/fs/bcachefs/sb-members.c
+++ b/fs/bcachefs/sb-members.c
@@ -1,12 +1,24 @@
 // SPDX-License-Identifier: GPL-2.0
 
 #include "bcachefs.h"
+#include "btree_cache.h"
 #include "disk_groups.h"
+#include "error.h"
 #include "opts.h"
 #include "replicas.h"
 #include "sb-members.h"
 #include "super-io.h"
 
+void bch2_dev_missing(struct bch_fs *c, unsigned dev)
+{
+	bch2_fs_inconsistent(c, "pointer to nonexistent device %u", dev);
+}
+
+void bch2_dev_bucket_missing(struct bch_fs *c, struct bpos bucket)
+{
+	bch2_fs_inconsistent(c, "pointer to nonexistent bucket %llu:%llu", bucket.inode, bucket.offset);
+}
+
 #define x(t, n, ...) [n] = #t,
 static const char * const bch2_iops_measurements[] = {
 	BCH_IOPS_MEASUREMENTS()
@@ -123,9 +135,9 @@ static int validate_member(struct printbuf *err,
 			   struct bch_sb *sb,
 			   int i)
 {
-	if (le64_to_cpu(m.nbuckets) > LONG_MAX) {
-		prt_printf(err, "device %u: too many buckets (got %llu, max %lu)",
-			   i, le64_to_cpu(m.nbuckets), LONG_MAX);
+	if (le64_to_cpu(m.nbuckets) > BCH_MEMBER_NBUCKETS_MAX) {
+		prt_printf(err, "device %u: too many buckets (got %llu, max %u)",
+			   i, le64_to_cpu(m.nbuckets), BCH_MEMBER_NBUCKETS_MAX);
 		return -BCH_ERR_invalid_sb_members;
 	}
 
@@ -163,18 +175,14 @@ static void member_to_text(struct printbuf *out,
 	u64 bucket_size = le16_to_cpu(m.bucket_size);
 	u64 device_size = le64_to_cpu(m.nbuckets) * bucket_size;
 
-	if (!bch2_member_exists(&m))
+	if (!bch2_member_alive(&m))
 		return;
 
-	prt_printf(out, "Device:");
-	prt_tab(out);
-	prt_printf(out, "%u", i);
-	prt_newline(out);
+	prt_printf(out, "Device:\t%u\n", i);
 
 	printbuf_indent_add(out, 2);
 
-	prt_printf(out, "Label:");
-	prt_tab(out);
+	prt_printf(out, "Label:\t");
 	if (BCH_MEMBER_GROUP(&m)) {
 		unsigned idx = BCH_MEMBER_GROUP(&m) - 1;
 
@@ -188,103 +196,73 @@ static void member_to_text(struct printbuf *out,
 	}
 	prt_newline(out);
 
-	prt_printf(out, "UUID:");
-	prt_tab(out);
+	prt_printf(out, "UUID:\t");
 	pr_uuid(out, m.uuid.b);
 	prt_newline(out);
 
-	prt_printf(out, "Size:");
-	prt_tab(out);
+	prt_printf(out, "Size:\t");
 	prt_units_u64(out, device_size << 9);
 	prt_newline(out);
 
-	for (unsigned i = 0; i < BCH_MEMBER_ERROR_NR; i++) {
-		prt_printf(out, "%s errors:", bch2_member_error_strs[i]);
-		prt_tab(out);
-		prt_u64(out, le64_to_cpu(m.errors[i]));
-		prt_newline(out);
-	}
+	for (unsigned i = 0; i < BCH_MEMBER_ERROR_NR; i++)
+		prt_printf(out, "%s errors:\t%llu\n", bch2_member_error_strs[i], le64_to_cpu(m.errors[i]));
 
-	for (unsigned i = 0; i < BCH_IOPS_NR; i++) {
-		prt_printf(out, "%s iops:", bch2_iops_measurements[i]);
-		prt_tab(out);
-		prt_printf(out, "%u", le32_to_cpu(m.iops[i]));
-		prt_newline(out);
-	}
+	for (unsigned i = 0; i < BCH_IOPS_NR; i++)
+		prt_printf(out, "%s iops:\t%u\n", bch2_iops_measurements[i], le32_to_cpu(m.iops[i]));
 
-	prt_printf(out, "Bucket size:");
-	prt_tab(out);
+	prt_printf(out, "Bucket size:\t");
 	prt_units_u64(out, bucket_size << 9);
 	prt_newline(out);
 
-	prt_printf(out, "First bucket:");
-	prt_tab(out);
-	prt_printf(out, "%u", le16_to_cpu(m.first_bucket));
-	prt_newline(out);
-
-	prt_printf(out, "Buckets:");
-	prt_tab(out);
-	prt_printf(out, "%llu", le64_to_cpu(m.nbuckets));
-	prt_newline(out);
+	prt_printf(out, "First bucket:\t%u\n", le16_to_cpu(m.first_bucket));
+	prt_printf(out, "Buckets:\t%llu\n", le64_to_cpu(m.nbuckets));
 
-	prt_printf(out, "Last mount:");
-	prt_tab(out);
+	prt_printf(out, "Last mount:\t");
 	if (m.last_mount)
 		bch2_prt_datetime(out, le64_to_cpu(m.last_mount));
 	else
 		prt_printf(out, "(never)");
 	prt_newline(out);
 
-	prt_printf(out, "Last superblock write:");
-	prt_tab(out);
-	prt_u64(out, le64_to_cpu(m.seq));
-	prt_newline(out);
+	prt_printf(out, "Last superblock write:\t%llu\n", le64_to_cpu(m.seq));
 
-	prt_printf(out, "State:");
-	prt_tab(out);
-	prt_printf(out, "%s",
+	prt_printf(out, "State:\t%s\n",
 		   BCH_MEMBER_STATE(&m) < BCH_MEMBER_STATE_NR
 		   ? bch2_member_states[BCH_MEMBER_STATE(&m)]
 		   : "unknown");
-	prt_newline(out);
 
-	prt_printf(out, "Data allowed:");
-	prt_tab(out);
+	prt_printf(out, "Data allowed:\t");
 	if (BCH_MEMBER_DATA_ALLOWED(&m))
 		prt_bitflags(out, __bch2_data_types, BCH_MEMBER_DATA_ALLOWED(&m));
 	else
 		prt_printf(out, "(none)");
 	prt_newline(out);
 
-	prt_printf(out, "Has data:");
-	prt_tab(out);
+	prt_printf(out, "Has data:\t");
 	if (data_have)
 		prt_bitflags(out, __bch2_data_types, data_have);
 	else
 		prt_printf(out, "(none)");
 	prt_newline(out);
 
-	prt_str(out, "Durability:");
-	prt_tab(out);
-	prt_printf(out, "%llu", BCH_MEMBER_DURABILITY(&m) ? BCH_MEMBER_DURABILITY(&m) - 1 : 1);
+	prt_printf(out, "Btree allocated bitmap blocksize:\t");
+	prt_units_u64(out, 1ULL << m.btree_bitmap_shift);
 	prt_newline(out);
 
-	prt_printf(out, "Discard:");
-	prt_tab(out);
-	prt_printf(out, "%llu", BCH_MEMBER_DISCARD(&m));
+	prt_printf(out, "Btree allocated bitmap:\t");
+	bch2_prt_u64_base2_nbits(out, le64_to_cpu(m.btree_allocated_bitmap), 64);
 	prt_newline(out);
 
-	prt_printf(out, "Freespace initialized:");
-	prt_tab(out);
-	prt_printf(out, "%llu", BCH_MEMBER_FREESPACE_INITIALIZED(&m));
-	prt_newline(out);
+	prt_printf(out, "Durability:\t%llu\n", BCH_MEMBER_DURABILITY(&m) ? BCH_MEMBER_DURABILITY(&m) - 1 : 1);
+
+	prt_printf(out, "Discard:\t%llu\n", BCH_MEMBER_DISCARD(&m));
+	prt_printf(out, "Freespace initialized:\t%llu\n", BCH_MEMBER_FREESPACE_INITIALIZED(&m));
 
 	printbuf_indent_sub(out, 2);
 }
 
-static int bch2_sb_members_v1_validate(struct bch_sb *sb,
-				    struct bch_sb_field *f,
-				    struct printbuf *err)
+static int bch2_sb_members_v1_validate(struct bch_sb *sb, struct bch_sb_field *f,
+				enum bch_validate_flags flags, struct printbuf *err)
 {
 	struct bch_sb_field_members_v1 *mi = field_to_type(f, members_v1);
 	unsigned i;
@@ -332,9 +310,8 @@ static void bch2_sb_members_v2_to_text(struct printbuf *out, struct bch_sb *sb,
 		member_to_text(out, members_v2_get(mi, i), gi, sb, i);
 }
 
-static int bch2_sb_members_v2_validate(struct bch_sb *sb,
-				       struct bch_sb_field *f,
-				       struct printbuf *err)
+static int bch2_sb_members_v2_validate(struct bch_sb *sb, struct bch_sb_field *f,
+				enum bch_validate_flags flags, struct printbuf *err)
 {
 	struct bch_sb_field_members_v2 *mi = field_to_type(f, members_v2);
 	size_t mi_bytes = (void *) __bch2_members_v2_get_mut(mi, sb->nr_devices) -
@@ -389,12 +366,8 @@ void bch2_dev_io_errors_to_text(struct printbuf *out, struct bch_dev *ca)
 	prt_newline(out);
 
 	printbuf_indent_add(out, 2);
-	for (unsigned i = 0; i < BCH_MEMBER_ERROR_NR; i++) {
-		prt_printf(out, "%s:", bch2_member_error_strs[i]);
-		prt_tab(out);
-		prt_u64(out, atomic64_read(&ca->errors[i]));
-		prt_newline(out);
-	}
+	for (unsigned i = 0; i < BCH_MEMBER_ERROR_NR; i++)
+		prt_printf(out, "%s:\t%llu\n", bch2_member_error_strs[i], atomic64_read(&ca->errors[i]));
 	printbuf_indent_sub(out, 2);
 
 	prt_str(out, "IO errors since ");
@@ -403,12 +376,9 @@ void bch2_dev_io_errors_to_text(struct printbuf *out, struct bch_dev *ca)
 	prt_newline(out);
 
 	printbuf_indent_add(out, 2);
-	for (unsigned i = 0; i < BCH_MEMBER_ERROR_NR; i++) {
-		prt_printf(out, "%s:", bch2_member_error_strs[i]);
-		prt_tab(out);
-		prt_u64(out, atomic64_read(&ca->errors[i]) - le64_to_cpu(m.errors_at_reset[i]));
-		prt_newline(out);
-	}
+	for (unsigned i = 0; i < BCH_MEMBER_ERROR_NR; i++)
+		prt_printf(out, "%s:\t%llu\n", bch2_member_error_strs[i],
+			   atomic64_read(&ca->errors[i]) - le64_to_cpu(m.errors_at_reset[i]));
 	printbuf_indent_sub(out, 2);
 }
 
@@ -426,3 +396,71 @@ void bch2_dev_errors_reset(struct bch_dev *ca)
 	bch2_write_super(c);
 	mutex_unlock(&c->sb_lock);
 }
+
+/*
+ * Per member "range has btree nodes" bitmap:
+ *
+ * This is so that if we ever have to run the btree node scan to repair we don't
+ * have to scan full devices:
+ */
+
+bool bch2_dev_btree_bitmap_marked(struct bch_fs *c, struct bkey_s_c k)
+{
+	bool ret = true;
+	rcu_read_lock();
+	bkey_for_each_ptr(bch2_bkey_ptrs_c(k), ptr) {
+		struct bch_dev *ca = bch2_dev_rcu(c, ptr->dev);
+		if (!ca)
+			continue;
+
+		if (!bch2_dev_btree_bitmap_marked_sectors(ca, ptr->offset, btree_sectors(c))) {
+			ret = false;
+			break;
+		}
+	}
+	rcu_read_unlock();
+	return ret;
+}
+
+static void __bch2_dev_btree_bitmap_mark(struct bch_sb_field_members_v2 *mi, unsigned dev,
+				u64 start, unsigned sectors)
+{
+	struct bch_member *m = __bch2_members_v2_get_mut(mi, dev);
+	u64 bitmap = le64_to_cpu(m->btree_allocated_bitmap);
+
+	u64 end = start + sectors;
+
+	int resize = ilog2(roundup_pow_of_two(end)) - (m->btree_bitmap_shift + 6);
+	if (resize > 0) {
+		u64 new_bitmap = 0;
+
+		for (unsigned i = 0; i < 64; i++)
+			if (bitmap & BIT_ULL(i))
+				new_bitmap |= BIT_ULL(i >> resize);
+		bitmap = new_bitmap;
+		m->btree_bitmap_shift += resize;
+	}
+
+	BUG_ON(m->btree_bitmap_shift > 57);
+	BUG_ON(end > 64ULL << m->btree_bitmap_shift);
+
+	for (unsigned bit = start >> m->btree_bitmap_shift;
+	     (u64) bit << m->btree_bitmap_shift < end;
+	     bit++)
+		bitmap |= BIT_ULL(bit);
+
+	m->btree_allocated_bitmap = cpu_to_le64(bitmap);
+}
+
+void bch2_dev_btree_bitmap_mark(struct bch_fs *c, struct bkey_s_c k)
+{
+	lockdep_assert_held(&c->sb_lock);
+
+	struct bch_sb_field_members_v2 *mi = bch2_sb_field_get(c->disk_sb.sb, members_v2);
+	bkey_for_each_ptr(bch2_bkey_ptrs_c(k), ptr) {
+		if (!bch2_member_exists(c->disk_sb.sb, ptr->dev))
+			continue;
+
+		__bch2_dev_btree_bitmap_mark(mi, ptr->dev, ptr->offset, btree_sectors(c));
+	}
+}
diff --git a/fs/bcachefs/sb-members.h b/fs/bcachefs/sb-members.h
index be0a941832..dd93192ec0 100644
--- a/fs/bcachefs/sb-members.h
+++ b/fs/bcachefs/sb-members.h
@@ -3,6 +3,7 @@
 #define _BCACHEFS_SB_MEMBERS_H
 
 #include "darray.h"
+#include "bkey_types.h"
 
 extern char * const bch2_member_error_strs[];
 
@@ -28,19 +29,6 @@ static inline bool bch2_dev_is_readable(struct bch_dev *ca)
 		ca->mi.state != BCH_MEMBER_STATE_failed;
 }
 
-static inline bool bch2_dev_get_ioref(struct bch_dev *ca, int rw)
-{
-	if (!percpu_ref_tryget(&ca->io_ref))
-		return false;
-
-	if (ca->mi.state == BCH_MEMBER_STATE_rw ||
-	    (ca->mi.state == BCH_MEMBER_STATE_ro && rw == READ))
-		return true;
-
-	percpu_ref_put(&ca->io_ref);
-	return false;
-}
-
 static inline unsigned dev_mask_nr(const struct bch_devs_mask *devs)
 {
 	return bitmap_weight(devs->d, BCH_SB_MEMBERS_MAX);
@@ -104,14 +92,41 @@ static inline struct bch_dev *__bch2_next_dev(struct bch_fs *c, struct bch_dev *
 	for (struct bch_dev *_ca = NULL;				\
 	     (_ca = __bch2_next_dev((_c), _ca, (_mask)));)
 
-static inline struct bch_dev *bch2_get_next_dev(struct bch_fs *c, struct bch_dev *ca)
+static inline void bch2_dev_get(struct bch_dev *ca)
+{
+#ifdef CONFIG_BCACHEFS_DEBUG
+	BUG_ON(atomic_long_inc_return(&ca->ref) <= 1L);
+#else
+	percpu_ref_get(&ca->ref);
+#endif
+}
+
+static inline void __bch2_dev_put(struct bch_dev *ca)
+{
+#ifdef CONFIG_BCACHEFS_DEBUG
+	long r = atomic_long_dec_return(&ca->ref);
+	if (r < (long) !ca->dying)
+		panic("bch_dev->ref underflow, last put: %pS\n", (void *) ca->last_put);
+	ca->last_put = _THIS_IP_;
+	if (!r)
+		complete(&ca->ref_completion);
+#else
+	percpu_ref_put(&ca->ref);
+#endif
+}
+
+static inline void bch2_dev_put(struct bch_dev *ca)
 {
 	if (ca)
-		percpu_ref_put(&ca->ref);
+		__bch2_dev_put(ca);
+}
 
+static inline struct bch_dev *bch2_get_next_dev(struct bch_fs *c, struct bch_dev *ca)
+{
 	rcu_read_lock();
+	bch2_dev_put(ca);
 	if ((ca = __bch2_next_dev(c, ca, NULL)))
-		percpu_ref_get(&ca->ref);
+		bch2_dev_get(ca);
 	rcu_read_unlock();
 
 	return ca;
@@ -131,10 +146,10 @@ static inline struct bch_dev *bch2_get_next_online_dev(struct bch_fs *c,
 						       struct bch_dev *ca,
 						       unsigned state_mask)
 {
+	rcu_read_lock();
 	if (ca)
 		percpu_ref_put(&ca->io_ref);
 
-	rcu_read_lock();
 	while ((ca = __bch2_next_dev(c, ca, NULL)) &&
 	       (!((1 << ca->mi.state) & state_mask) ||
 		!percpu_ref_tryget(&ca->io_ref)))
@@ -157,26 +172,113 @@ static inline struct bch_dev *bch2_get_next_online_dev(struct bch_fs *c,
 #define for_each_readable_member(c, ca)				\
 	__for_each_online_member(c, ca,	BIT( BCH_MEMBER_STATE_rw)|BIT(BCH_MEMBER_STATE_ro))
 
-/*
- * If a key exists that references a device, the device won't be going away and
- * we can omit rcu_read_lock():
- */
-static inline struct bch_dev *bch_dev_bkey_exists(const struct bch_fs *c, unsigned idx)
+static inline bool bch2_dev_exists(const struct bch_fs *c, unsigned dev)
 {
-	EBUG_ON(idx >= c->sb.nr_devices || !c->devs[idx]);
+	return dev < c->sb.nr_devices && c->devs[dev];
+}
 
-	return rcu_dereference_check(c->devs[idx], 1);
+static inline bool bucket_valid(const struct bch_dev *ca, u64 b)
+{
+	return b - ca->mi.first_bucket < ca->mi.nbuckets_minus_first;
 }
 
-static inline struct bch_dev *bch_dev_locked(struct bch_fs *c, unsigned idx)
+static inline struct bch_dev *bch2_dev_have_ref(const struct bch_fs *c, unsigned dev)
 {
-	EBUG_ON(idx >= c->sb.nr_devices || !c->devs[idx]);
+	EBUG_ON(!bch2_dev_exists(c, dev));
 
-	return rcu_dereference_protected(c->devs[idx],
+	return rcu_dereference_check(c->devs[dev], 1);
+}
+
+static inline struct bch_dev *bch2_dev_locked(struct bch_fs *c, unsigned dev)
+{
+	EBUG_ON(!bch2_dev_exists(c, dev));
+
+	return rcu_dereference_protected(c->devs[dev],
 					 lockdep_is_held(&c->sb_lock) ||
 					 lockdep_is_held(&c->state_lock));
 }
 
+static inline struct bch_dev *bch2_dev_rcu(struct bch_fs *c, unsigned dev)
+{
+	return c && dev < c->sb.nr_devices
+		? rcu_dereference(c->devs[dev])
+		: NULL;
+}
+
+static inline struct bch_dev *bch2_dev_tryget_noerror(struct bch_fs *c, unsigned dev)
+{
+	rcu_read_lock();
+	struct bch_dev *ca = bch2_dev_rcu(c, dev);
+	if (ca)
+		bch2_dev_get(ca);
+	rcu_read_unlock();
+	return ca;
+}
+
+void bch2_dev_missing(struct bch_fs *, unsigned);
+
+static inline struct bch_dev *bch2_dev_tryget(struct bch_fs *c, unsigned dev)
+{
+	struct bch_dev *ca = bch2_dev_tryget_noerror(c, dev);
+	if (!ca)
+		bch2_dev_missing(c, dev);
+	return ca;
+}
+
+static inline struct bch_dev *bch2_dev_bucket_tryget_noerror(struct bch_fs *c, struct bpos bucket)
+{
+	struct bch_dev *ca = bch2_dev_tryget_noerror(c, bucket.inode);
+	if (ca && !bucket_valid(ca, bucket.offset)) {
+		bch2_dev_put(ca);
+		ca = NULL;
+	}
+	return ca;
+}
+
+void bch2_dev_bucket_missing(struct bch_fs *, struct bpos);
+
+static inline struct bch_dev *bch2_dev_bucket_tryget(struct bch_fs *c, struct bpos bucket)
+{
+	struct bch_dev *ca = bch2_dev_bucket_tryget_noerror(c, bucket);
+	if (!ca)
+		bch2_dev_bucket_missing(c, bucket);
+	return ca;
+}
+
+static inline struct bch_dev *bch2_dev_iterate_noerror(struct bch_fs *c, struct bch_dev *ca, unsigned dev_idx)
+{
+	if (ca && ca->dev_idx == dev_idx)
+		return ca;
+	bch2_dev_put(ca);
+	return bch2_dev_tryget_noerror(c, dev_idx);
+}
+
+static inline struct bch_dev *bch2_dev_iterate(struct bch_fs *c, struct bch_dev *ca, unsigned dev_idx)
+{
+	if (ca && ca->dev_idx == dev_idx)
+		return ca;
+	bch2_dev_put(ca);
+	return bch2_dev_tryget(c, dev_idx);
+}
+
+static inline struct bch_dev *bch2_dev_get_ioref(struct bch_fs *c, unsigned dev, int rw)
+{
+	rcu_read_lock();
+	struct bch_dev *ca = bch2_dev_rcu(c, dev);
+	if (ca && !percpu_ref_tryget(&ca->io_ref))
+		ca = NULL;
+	rcu_read_unlock();
+
+	if (ca &&
+	    (ca->mi.state == BCH_MEMBER_STATE_rw ||
+	    (ca->mi.state == BCH_MEMBER_STATE_ro && rw == READ)))
+		return ca;
+
+	if (ca)
+		percpu_ref_put(&ca->io_ref);
+	return NULL;
+}
+
 /* XXX kill, move to struct bch_fs */
 static inline struct bch_devs_mask bch2_online_devs(struct bch_fs *c)
 {
@@ -191,16 +293,16 @@ static inline struct bch_devs_mask bch2_online_devs(struct bch_fs *c)
 extern const struct bch_sb_field_ops bch_sb_field_ops_members_v1;
 extern const struct bch_sb_field_ops bch_sb_field_ops_members_v2;
 
-static inline bool bch2_member_exists(struct bch_member *m)
+static inline bool bch2_member_alive(struct bch_member *m)
 {
 	return !bch2_is_zero(&m->uuid, sizeof(m->uuid));
 }
 
-static inline bool bch2_dev_exists(struct bch_sb *sb, unsigned dev)
+static inline bool bch2_member_exists(struct bch_sb *sb, unsigned dev)
 {
 	if (dev < sb->nr_devices) {
 		struct bch_member m = bch2_sb_member_get(sb, dev);
-		return bch2_member_exists(&m);
+		return bch2_member_alive(&m);
 	}
 	return false;
 }
@@ -209,6 +311,8 @@ static inline struct bch_member_cpu bch2_mi_to_cpu(struct bch_member *mi)
 {
 	return (struct bch_member_cpu) {
 		.nbuckets	= le64_to_cpu(mi->nbuckets),
+		.nbuckets_minus_first = le64_to_cpu(mi->nbuckets) -
+			le16_to_cpu(mi->first_bucket),
 		.first_bucket	= le16_to_cpu(mi->first_bucket),
 		.bucket_size	= le16_to_cpu(mi->bucket_size),
 		.group		= BCH_MEMBER_GROUP(mi),
@@ -219,7 +323,9 @@ static inline struct bch_member_cpu bch2_mi_to_cpu(struct bch_member *mi)
 			? BCH_MEMBER_DURABILITY(mi) - 1
 			: 1,
 		.freespace_initialized = BCH_MEMBER_FREESPACE_INITIALIZED(mi),
-		.valid		= bch2_member_exists(mi),
+		.valid		= bch2_member_alive(mi),
+		.btree_bitmap_shift	= mi->btree_bitmap_shift,
+		.btree_allocated_bitmap = le64_to_cpu(mi->btree_allocated_bitmap),
 	};
 }
 
@@ -228,4 +334,22 @@ void bch2_sb_members_from_cpu(struct bch_fs *);
 void bch2_dev_io_errors_to_text(struct printbuf *, struct bch_dev *);
 void bch2_dev_errors_reset(struct bch_dev *);
 
+static inline bool bch2_dev_btree_bitmap_marked_sectors(struct bch_dev *ca, u64 start, unsigned sectors)
+{
+	u64 end = start + sectors;
+
+	if (end > 64ULL << ca->mi.btree_bitmap_shift)
+		return false;
+
+	for (unsigned bit = start >> ca->mi.btree_bitmap_shift;
+	     (u64) bit << ca->mi.btree_bitmap_shift < end;
+	     bit++)
+		if (!(ca->mi.btree_allocated_bitmap & BIT_ULL(bit)))
+			return false;
+	return true;
+}
+
+bool bch2_dev_btree_bitmap_marked(struct bch_fs *, struct bkey_s_c);
+void bch2_dev_btree_bitmap_mark(struct bch_fs *, struct bkey_s_c);
+
 #endif /* _BCACHEFS_SB_MEMBERS_H */
diff --git a/fs/bcachefs/sb-members_format.h b/fs/bcachefs/sb-members_format.h
new file mode 100644
index 0000000000..e2630548c0
--- /dev/null
+++ b/fs/bcachefs/sb-members_format.h
@@ -0,0 +1,110 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _BCACHEFS_SB_MEMBERS_FORMAT_H
+#define _BCACHEFS_SB_MEMBERS_FORMAT_H
+
+/*
+ * We refer to members with bitmasks in various places - but we need to get rid
+ * of this limit:
+ */
+#define BCH_SB_MEMBERS_MAX		64
+
+#define BCH_MIN_NR_NBUCKETS	(1 << 6)
+
+#define BCH_IOPS_MEASUREMENTS()			\
+	x(seqread,	0)			\
+	x(seqwrite,	1)			\
+	x(randread,	2)			\
+	x(randwrite,	3)
+
+enum bch_iops_measurement {
+#define x(t, n) BCH_IOPS_##t = n,
+	BCH_IOPS_MEASUREMENTS()
+#undef x
+	BCH_IOPS_NR
+};
+
+#define BCH_MEMBER_ERROR_TYPES()		\
+	x(read,		0)			\
+	x(write,	1)			\
+	x(checksum,	2)
+
+enum bch_member_error_type {
+#define x(t, n) BCH_MEMBER_ERROR_##t = n,
+	BCH_MEMBER_ERROR_TYPES()
+#undef x
+	BCH_MEMBER_ERROR_NR
+};
+
+struct bch_member {
+	__uuid_t		uuid;
+	__le64			nbuckets;	/* device size */
+	__le16			first_bucket;   /* index of first bucket used */
+	__le16			bucket_size;	/* sectors */
+	__u8			btree_bitmap_shift;
+	__u8			pad[3];
+	__le64			last_mount;	/* time_t */
+
+	__le64			flags;
+	__le32			iops[4];
+	__le64			errors[BCH_MEMBER_ERROR_NR];
+	__le64			errors_at_reset[BCH_MEMBER_ERROR_NR];
+	__le64			errors_reset_time;
+	__le64			seq;
+	__le64			btree_allocated_bitmap;
+	/*
+	 * On recovery from a clean shutdown we don't normally read the journal,
+	 * but we still want to resume writing from where we left off so we
+	 * don't overwrite more than is necessary, for list journal debugging:
+	 */
+	__le32			last_journal_bucket;
+	__le32			last_journal_bucket_offset;
+};
+
+/*
+ * This limit comes from the bucket_gens array - it's a single allocation, and
+ * kernel allocation are limited to INT_MAX
+ */
+#define BCH_MEMBER_NBUCKETS_MAX	(INT_MAX - 64)
+
+#define BCH_MEMBER_V1_BYTES	56
+
+LE64_BITMASK(BCH_MEMBER_STATE,		struct bch_member, flags,  0,  4)
+/* 4-14 unused, was TIER, HAS_(META)DATA, REPLACEMENT */
+LE64_BITMASK(BCH_MEMBER_DISCARD,	struct bch_member, flags, 14, 15)
+LE64_BITMASK(BCH_MEMBER_DATA_ALLOWED,	struct bch_member, flags, 15, 20)
+LE64_BITMASK(BCH_MEMBER_GROUP,		struct bch_member, flags, 20, 28)
+LE64_BITMASK(BCH_MEMBER_DURABILITY,	struct bch_member, flags, 28, 30)
+LE64_BITMASK(BCH_MEMBER_FREESPACE_INITIALIZED,
+					struct bch_member, flags, 30, 31)
+
+#if 0
+LE64_BITMASK(BCH_MEMBER_NR_READ_ERRORS,	struct bch_member, flags[1], 0,  20);
+LE64_BITMASK(BCH_MEMBER_NR_WRITE_ERRORS,struct bch_member, flags[1], 20, 40);
+#endif
+
+#define BCH_MEMBER_STATES()			\
+	x(rw,		0)			\
+	x(ro,		1)			\
+	x(failed,	2)			\
+	x(spare,	3)
+
+enum bch_member_state {
+#define x(t, n) BCH_MEMBER_STATE_##t = n,
+	BCH_MEMBER_STATES()
+#undef x
+	BCH_MEMBER_STATE_NR
+};
+
+struct bch_sb_field_members_v1 {
+	struct bch_sb_field	field;
+	struct bch_member	_members[]; //Members are now variable size
+};
+
+struct bch_sb_field_members_v2 {
+	struct bch_sb_field	field;
+	__le16			member_bytes; //size of single member entry
+	u8			pad[6];
+	struct bch_member	_members[];
+};
+
+#endif /* _BCACHEFS_SB_MEMBERS_FORMAT_H */
diff --git a/fs/bcachefs/sb-members_types.h b/fs/bcachefs/sb-members_types.h
new file mode 100644
index 0000000000..c0eda888fe
--- /dev/null
+++ b/fs/bcachefs/sb-members_types.h
@@ -0,0 +1,21 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _BCACHEFS_SB_MEMBERS_TYPES_H
+#define _BCACHEFS_SB_MEMBERS_TYPES_H
+
+struct bch_member_cpu {
+	u64			nbuckets;	/* device size */
+	u64			nbuckets_minus_first;
+	u16			first_bucket;   /* index of first bucket used */
+	u16			bucket_size;	/* sectors */
+	u16			group;
+	u8			state;
+	u8			discard;
+	u8			data_allowed;
+	u8			durability;
+	u8			freespace_initialized;
+	u8			valid;
+	u8			btree_bitmap_shift;
+	u64			btree_allocated_bitmap;
+};
+
+#endif /* _BCACHEFS_SB_MEMBERS_H */
diff --git a/fs/bcachefs/seqmutex.h b/fs/bcachefs/seqmutex.h
index c1860d8163..c4b3d8d3f4 100644
--- a/fs/bcachefs/seqmutex.h
+++ b/fs/bcachefs/seqmutex.h
@@ -19,17 +19,14 @@ static inline bool seqmutex_trylock(struct seqmutex *lock)
 static inline void seqmutex_lock(struct seqmutex *lock)
 {
 	mutex_lock(&lock->lock);
-}
-
-static inline void seqmutex_unlock(struct seqmutex *lock)
-{
 	lock->seq++;
-	mutex_unlock(&lock->lock);
 }
 
-static inline u32 seqmutex_seq(struct seqmutex *lock)
+static inline u32 seqmutex_unlock(struct seqmutex *lock)
 {
-	return lock->seq;
+	u32 seq = lock->seq;
+	mutex_unlock(&lock->lock);
+	return seq;
 }
 
 static inline bool seqmutex_relock(struct seqmutex *lock, u32 seq)
diff --git a/fs/bcachefs/snapshot.c b/fs/bcachefs/snapshot.c
index ac6ba04d55..24023d6a96 100644
--- a/fs/bcachefs/snapshot.c
+++ b/fs/bcachefs/snapshot.c
@@ -8,6 +8,7 @@
 #include "errcode.h"
 #include "error.h"
 #include "fs.h"
+#include "recovery_passes.h"
 #include "snapshot.h"
 
 #include <linux/random.h>
@@ -31,7 +32,7 @@ void bch2_snapshot_tree_to_text(struct printbuf *out, struct bch_fs *c,
 }
 
 int bch2_snapshot_tree_invalid(struct bch_fs *c, struct bkey_s_c k,
-			       enum bkey_invalid_flags flags,
+			       enum bch_validate_flags flags,
 			       struct printbuf *err)
 {
 	int ret = 0;
@@ -48,7 +49,7 @@ int bch2_snapshot_tree_lookup(struct btree_trans *trans, u32 id,
 			      struct bch_snapshot_tree *s)
 {
 	int ret = bch2_bkey_get_val_typed(trans, BTREE_ID_snapshot_trees, POS(0, id),
-					  BTREE_ITER_WITH_UPDATES, snapshot_tree, s);
+					  BTREE_ITER_with_updates, snapshot_tree, s);
 
 	if (bch2_err_matches(ret, ENOENT))
 		ret = -BCH_ERR_ENOENT_snapshot_tree;
@@ -91,23 +92,29 @@ static int bch2_snapshot_tree_create(struct btree_trans *trans,
 
 /* Snapshot nodes: */
 
-static bool bch2_snapshot_is_ancestor_early(struct bch_fs *c, u32 id, u32 ancestor)
+static bool __bch2_snapshot_is_ancestor_early(struct snapshot_table *t, u32 id, u32 ancestor)
 {
-	struct snapshot_table *t;
+	while (id && id < ancestor) {
+		const struct snapshot_t *s = __snapshot_t(t, id);
+		id = s ? s->parent : 0;
+	}
+	return id == ancestor;
+}
 
+static bool bch2_snapshot_is_ancestor_early(struct bch_fs *c, u32 id, u32 ancestor)
+{
 	rcu_read_lock();
-	t = rcu_dereference(c->snapshots);
-
-	while (id && id < ancestor)
-		id = __snapshot_t(t, id)->parent;
+	bool ret = __bch2_snapshot_is_ancestor_early(rcu_dereference(c->snapshots), id, ancestor);
 	rcu_read_unlock();
 
-	return id == ancestor;
+	return ret;
 }
 
 static inline u32 get_ancestor_below(struct snapshot_table *t, u32 id, u32 ancestor)
 {
 	const struct snapshot_t *s = __snapshot_t(t, id);
+	if (!s)
+		return 0;
 
 	if (s->skip[2] <= ancestor)
 		return s->skip[2];
@@ -118,27 +125,36 @@ static inline u32 get_ancestor_below(struct snapshot_table *t, u32 id, u32 ances
 	return s->parent;
 }
 
+static bool test_ancestor_bitmap(struct snapshot_table *t, u32 id, u32 ancestor)
+{
+	const struct snapshot_t *s = __snapshot_t(t, id);
+	if (!s)
+		return false;
+
+	return test_bit(ancestor - id - 1, s->is_ancestor);
+}
+
 bool __bch2_snapshot_is_ancestor(struct bch_fs *c, u32 id, u32 ancestor)
 {
-	struct snapshot_table *t;
 	bool ret;
 
-	EBUG_ON(c->recovery_pass_done <= BCH_RECOVERY_PASS_check_snapshots);
-
 	rcu_read_lock();
-	t = rcu_dereference(c->snapshots);
+	struct snapshot_table *t = rcu_dereference(c->snapshots);
+
+	if (unlikely(c->recovery_pass_done < BCH_RECOVERY_PASS_check_snapshots)) {
+		ret = __bch2_snapshot_is_ancestor_early(t, id, ancestor);
+		goto out;
+	}
 
 	while (id && id < ancestor - IS_ANCESTOR_BITMAP)
 		id = get_ancestor_below(t, id, ancestor);
 
-	if (id && id < ancestor) {
-		ret = test_bit(ancestor - id - 1, __snapshot_t(t, id)->is_ancestor);
-
-		EBUG_ON(ret != bch2_snapshot_is_ancestor_early(c, id, ancestor));
-	} else {
-		ret = id == ancestor;
-	}
+	ret = id && id < ancestor
+		? test_ancestor_bitmap(t, id, ancestor)
+		: id == ancestor;
 
+	EBUG_ON(ret != __bch2_snapshot_is_ancestor_early(t, id, ancestor));
+out:
 	rcu_read_unlock();
 
 	return ret;
@@ -147,36 +163,42 @@ bool __bch2_snapshot_is_ancestor(struct bch_fs *c, u32 id, u32 ancestor)
 static noinline struct snapshot_t *__snapshot_t_mut(struct bch_fs *c, u32 id)
 {
 	size_t idx = U32_MAX - id;
-	size_t new_size;
 	struct snapshot_table *new, *old;
 
-	new_size = max(16UL, roundup_pow_of_two(idx + 1));
+	size_t new_bytes = kmalloc_size_roundup(struct_size(new, s, idx + 1));
+	size_t new_size = (new_bytes - sizeof(*new)) / sizeof(new->s[0]);
 
-	new = kvzalloc(struct_size(new, s, new_size), GFP_KERNEL);
+	if (unlikely(new_bytes > INT_MAX))
+		return NULL;
+
+	new = kvzalloc(new_bytes, GFP_KERNEL);
 	if (!new)
 		return NULL;
 
+	new->nr = new_size;
+
 	old = rcu_dereference_protected(c->snapshots, true);
 	if (old)
-		memcpy(new->s,
-		       rcu_dereference_protected(c->snapshots, true)->s,
-		       sizeof(new->s[0]) * c->snapshot_table_size);
+		memcpy(new->s, old->s, sizeof(old->s[0]) * old->nr);
 
 	rcu_assign_pointer(c->snapshots, new);
-	c->snapshot_table_size = new_size;
-	kvfree_rcu_mightsleep(old);
+	kvfree_rcu(old, rcu);
 
-	return &rcu_dereference_protected(c->snapshots, true)->s[idx];
+	return &rcu_dereference_protected(c->snapshots,
+				lockdep_is_held(&c->snapshot_table_lock))->s[idx];
 }
 
 static inline struct snapshot_t *snapshot_t_mut(struct bch_fs *c, u32 id)
 {
 	size_t idx = U32_MAX - id;
+	struct snapshot_table *table =
+		rcu_dereference_protected(c->snapshots,
+				lockdep_is_held(&c->snapshot_table_lock));
 
 	lockdep_assert_held(&c->snapshot_table_lock);
 
-	if (likely(idx < c->snapshot_table_size))
-		return &rcu_dereference_protected(c->snapshots, true)->s[idx];
+	if (likely(table && idx < table->nr))
+		return &table->s[idx];
 
 	return __snapshot_t_mut(c, id);
 }
@@ -204,7 +226,7 @@ void bch2_snapshot_to_text(struct printbuf *out, struct bch_fs *c,
 }
 
 int bch2_snapshot_invalid(struct bch_fs *c, struct bkey_s_c k,
-			  enum bkey_invalid_flags flags,
+			  enum bch_validate_flags flags,
 			  struct printbuf *err)
 {
 	struct bkey_s_c_snapshot s;
@@ -279,7 +301,7 @@ static void set_is_ancestor_bitmap(struct bch_fs *c, u32 id)
 static int __bch2_mark_snapshot(struct btree_trans *trans,
 		       enum btree_id btree, unsigned level,
 		       struct bkey_s_c old, struct bkey_s_c new,
-		       unsigned flags)
+		       enum btree_iter_update_trigger_flags flags)
 {
 	struct bch_fs *c = trans->c;
 	struct snapshot_t *t;
@@ -333,7 +355,7 @@ err:
 int bch2_mark_snapshot(struct btree_trans *trans,
 		       enum btree_id btree, unsigned level,
 		       struct bkey_s_c old, struct bkey_s new,
-		       unsigned flags)
+		       enum btree_iter_update_trigger_flags flags)
 {
 	return __bch2_mark_snapshot(trans, btree, level, old, new.s_c, flags);
 }
@@ -342,7 +364,7 @@ int bch2_snapshot_lookup(struct btree_trans *trans, u32 id,
 			 struct bch_snapshot *s)
 {
 	return bch2_bkey_get_val_typed(trans, BTREE_ID_snapshots, POS(0, id),
-				       BTREE_ITER_WITH_UPDATES, snapshot, s);
+				       BTREE_ITER_with_updates, snapshot, s);
 }
 
 static int bch2_snapshot_live(struct btree_trans *trans, u32 id)
@@ -547,7 +569,7 @@ static int check_snapshot_tree(struct btree_trans *trans,
 			"snapshot tree points to missing subvolume:\n  %s",
 			(printbuf_reset(&buf),
 			 bch2_bkey_val_to_text(&buf, c, st.s_c), buf.buf)) ||
-	    fsck_err_on(!bch2_snapshot_is_ancestor_early(c,
+	    fsck_err_on(!bch2_snapshot_is_ancestor(c,
 						le32_to_cpu(subvol.snapshot),
 						root_id),
 			c, snapshot_tree_to_wrong_subvol,
@@ -563,6 +585,13 @@ static int check_snapshot_tree(struct btree_trans *trans,
 		u32 subvol_id;
 
 		ret = bch2_snapshot_tree_master_subvol(trans, root_id, &subvol_id);
+		bch_err_fn(c, ret);
+
+		if (bch2_err_matches(ret, ENOENT)) { /* nothing to be done here */
+			ret = 0;
+			goto err;
+		}
+
 		if (ret)
 			goto err;
 
@@ -592,7 +621,7 @@ int bch2_check_snapshot_trees(struct bch_fs *c)
 	int ret = bch2_trans_run(c,
 		for_each_btree_key_commit(trans, iter,
 			BTREE_ID_snapshot_trees, POS_MIN,
-			BTREE_ITER_PREFETCH, k,
+			BTREE_ITER_prefetch, k,
 			NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
 		check_snapshot_tree(trans, &iter, k)));
 	bch_err_fn(c, ret);
@@ -669,7 +698,7 @@ static int snapshot_tree_ptr_repair(struct btree_trans *trans,
 
 	root = bch2_bkey_get_iter_typed(trans, &root_iter,
 			       BTREE_ID_snapshots, POS(0, root_id),
-			       BTREE_ITER_WITH_UPDATES, snapshot);
+			       BTREE_ITER_with_updates, snapshot);
 	ret = bkey_err(root);
 	if (ret)
 		goto err;
@@ -720,7 +749,6 @@ static int check_snapshot(struct btree_trans *trans,
 	u32 parent_id = bch2_snapshot_parent_early(c, k.k->p.offset);
 	u32 real_depth;
 	struct printbuf buf = PRINTBUF;
-	bool should_have_subvol;
 	u32 i, id;
 	int ret = 0;
 
@@ -766,7 +794,7 @@ static int check_snapshot(struct btree_trans *trans,
 		}
 	}
 
-	should_have_subvol = BCH_SNAPSHOT_SUBVOL(&s) &&
+	bool should_have_subvol = BCH_SNAPSHOT_SUBVOL(&s) &&
 		!BCH_SNAPSHOT_DELETED(&s);
 
 	if (should_have_subvol) {
@@ -861,13 +889,181 @@ int bch2_check_snapshots(struct bch_fs *c)
 	int ret = bch2_trans_run(c,
 		for_each_btree_key_reverse_commit(trans, iter,
 				BTREE_ID_snapshots, POS_MAX,
-				BTREE_ITER_PREFETCH, k,
+				BTREE_ITER_prefetch, k,
 				NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
 			check_snapshot(trans, &iter, k)));
 	bch_err_fn(c, ret);
 	return ret;
 }
 
+static int check_snapshot_exists(struct btree_trans *trans, u32 id)
+{
+	struct bch_fs *c = trans->c;
+
+	if (bch2_snapshot_equiv(c, id))
+		return 0;
+
+	/* 0 is an invalid tree ID */
+	u32 tree_id = 0;
+	int ret = bch2_snapshot_tree_create(trans, id, 0, &tree_id);
+	if (ret)
+		return ret;
+
+	struct bkey_i_snapshot *snapshot = bch2_trans_kmalloc(trans, sizeof(*snapshot));
+	ret = PTR_ERR_OR_ZERO(snapshot);
+	if (ret)
+		return ret;
+
+	bkey_snapshot_init(&snapshot->k_i);
+	snapshot->k.p		= POS(0, id);
+	snapshot->v.tree	= cpu_to_le32(tree_id);
+	snapshot->v.btime.lo	= cpu_to_le64(bch2_current_time(c));
+
+	return  bch2_btree_insert_trans(trans, BTREE_ID_snapshots, &snapshot->k_i, 0) ?:
+		bch2_mark_snapshot(trans, BTREE_ID_snapshots, 0,
+				   bkey_s_c_null, bkey_i_to_s(&snapshot->k_i), 0) ?:
+		bch2_snapshot_set_equiv(trans, bkey_i_to_s_c(&snapshot->k_i));
+}
+
+/* Figure out which snapshot nodes belong in the same tree: */
+struct snapshot_tree_reconstruct {
+	enum btree_id			btree;
+	struct bpos			cur_pos;
+	snapshot_id_list		cur_ids;
+	DARRAY(snapshot_id_list)	trees;
+};
+
+static void snapshot_tree_reconstruct_exit(struct snapshot_tree_reconstruct *r)
+{
+	darray_for_each(r->trees, i)
+		darray_exit(i);
+	darray_exit(&r->trees);
+	darray_exit(&r->cur_ids);
+}
+
+static inline bool same_snapshot(struct snapshot_tree_reconstruct *r, struct bpos pos)
+{
+	return r->btree == BTREE_ID_inodes
+		? r->cur_pos.offset == pos.offset
+		: r->cur_pos.inode == pos.inode;
+}
+
+static inline bool snapshot_id_lists_have_common(snapshot_id_list *l, snapshot_id_list *r)
+{
+	darray_for_each(*l, i)
+		if (snapshot_list_has_id(r, *i))
+			return true;
+	return false;
+}
+
+static void snapshot_id_list_to_text(struct printbuf *out, snapshot_id_list *s)
+{
+	bool first = true;
+	darray_for_each(*s, i) {
+		if (!first)
+			prt_char(out, ' ');
+		first = false;
+		prt_printf(out, "%u", *i);
+	}
+}
+
+static int snapshot_tree_reconstruct_next(struct bch_fs *c, struct snapshot_tree_reconstruct *r)
+{
+	if (r->cur_ids.nr) {
+		darray_for_each(r->trees, i)
+			if (snapshot_id_lists_have_common(i, &r->cur_ids)) {
+				int ret = snapshot_list_merge(c, i, &r->cur_ids);
+				if (ret)
+					return ret;
+				goto out;
+			}
+		darray_push(&r->trees, r->cur_ids);
+		darray_init(&r->cur_ids);
+	}
+out:
+	r->cur_ids.nr = 0;
+	return 0;
+}
+
+static int get_snapshot_trees(struct bch_fs *c, struct snapshot_tree_reconstruct *r, struct bpos pos)
+{
+	if (!same_snapshot(r, pos))
+		snapshot_tree_reconstruct_next(c, r);
+	r->cur_pos = pos;
+	return snapshot_list_add_nodup(c, &r->cur_ids, pos.snapshot);
+}
+
+int bch2_reconstruct_snapshots(struct bch_fs *c)
+{
+	struct btree_trans *trans = bch2_trans_get(c);
+	struct printbuf buf = PRINTBUF;
+	struct snapshot_tree_reconstruct r = {};
+	int ret = 0;
+
+	for (unsigned btree = 0; btree < BTREE_ID_NR; btree++) {
+		if (btree_type_has_snapshots(btree)) {
+			r.btree = btree;
+
+			ret = for_each_btree_key(trans, iter, btree, POS_MIN,
+					BTREE_ITER_all_snapshots|BTREE_ITER_prefetch, k, ({
+				get_snapshot_trees(c, &r, k.k->p);
+			}));
+			if (ret)
+				goto err;
+
+			snapshot_tree_reconstruct_next(c, &r);
+		}
+	}
+
+	darray_for_each(r.trees, t) {
+		printbuf_reset(&buf);
+		snapshot_id_list_to_text(&buf, t);
+
+		darray_for_each(*t, id) {
+			if (fsck_err_on(!bch2_snapshot_equiv(c, *id),
+					c, snapshot_node_missing,
+					"snapshot node %u from tree %s missing, recreate?", *id, buf.buf)) {
+				if (t->nr > 1) {
+					bch_err(c, "cannot reconstruct snapshot trees with multiple nodes");
+					ret = -BCH_ERR_fsck_repair_unimplemented;
+					goto err;
+				}
+
+				ret = commit_do(trans, NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
+						check_snapshot_exists(trans, *id));
+				if (ret)
+					goto err;
+			}
+		}
+	}
+fsck_err:
+err:
+	bch2_trans_put(trans);
+	snapshot_tree_reconstruct_exit(&r);
+	printbuf_exit(&buf);
+	bch_err_fn(c, ret);
+	return ret;
+}
+
+int bch2_check_key_has_snapshot(struct btree_trans *trans,
+				struct btree_iter *iter,
+				struct bkey_s_c k)
+{
+	struct bch_fs *c = trans->c;
+	struct printbuf buf = PRINTBUF;
+	int ret = 0;
+
+	if (fsck_err_on(!bch2_snapshot_equiv(c, k.k->p.snapshot), c,
+			bkey_in_missing_snapshot,
+			"key in missing snapshot %s, delete?",
+			(bch2_bkey_val_to_text(&buf, c, k), buf.buf)))
+		ret = bch2_btree_delete_at(trans, iter,
+					    BTREE_UPDATE_internal_snapshot_node) ?: 1;
+fsck_err:
+	printbuf_exit(&buf);
+	return ret;
+}
+
 /*
  * Mark a snapshot as deleted, for future cleanup:
  */
@@ -917,7 +1113,7 @@ static int bch2_snapshot_node_delete(struct btree_trans *trans, u32 id)
 	int ret = 0;
 
 	s = bch2_bkey_get_iter_typed(trans, &iter, BTREE_ID_snapshots, POS(0, id),
-				     BTREE_ITER_INTENT, snapshot);
+				     BTREE_ITER_intent, snapshot);
 	ret = bkey_err(s);
 	bch2_fs_inconsistent_on(bch2_err_matches(ret, ENOENT), c,
 				"missing snapshot %u", id);
@@ -1026,7 +1222,7 @@ static int create_snapids(struct btree_trans *trans, u32 parent, u32 tree,
 	int ret;
 
 	bch2_trans_iter_init(trans, &iter, BTREE_ID_snapshots,
-			     POS_MIN, BTREE_ITER_INTENT);
+			     POS_MIN, BTREE_ITER_intent);
 	k = bch2_btree_iter_peek(&iter);
 	ret = bkey_err(k);
 	if (ret)
@@ -1177,35 +1373,39 @@ int bch2_snapshot_node_create(struct btree_trans *trans, u32 parent,
  * that key to snapshot leaf nodes, where we can mutate it
  */
 
-static int snapshot_delete_key(struct btree_trans *trans,
+static int delete_dead_snapshots_process_key(struct btree_trans *trans,
 			       struct btree_iter *iter,
 			       struct bkey_s_c k,
 			       snapshot_id_list *deleted,
 			       snapshot_id_list *equiv_seen,
 			       struct bpos *last_pos)
 {
+	int ret = bch2_check_key_has_snapshot(trans, iter, k);
+	if (ret)
+		return ret < 0 ? ret : 0;
+
 	struct bch_fs *c = trans->c;
 	u32 equiv = bch2_snapshot_equiv(c, k.k->p.snapshot);
+	if (!equiv) /* key for invalid snapshot node, but we chose not to delete */
+		return 0;
 
 	if (!bkey_eq(k.k->p, *last_pos))
 		equiv_seen->nr = 0;
-	*last_pos = k.k->p;
 
-	if (snapshot_list_has_id(deleted, k.k->p.snapshot) ||
-	    snapshot_list_has_id(equiv_seen, equiv)) {
+	if (snapshot_list_has_id(deleted, k.k->p.snapshot))
 		return bch2_btree_delete_at(trans, iter,
-					    BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
-	} else {
-		return snapshot_list_add(c, equiv_seen, equiv);
-	}
-}
+					    BTREE_UPDATE_internal_snapshot_node);
 
-static int move_key_to_correct_snapshot(struct btree_trans *trans,
-			       struct btree_iter *iter,
-			       struct bkey_s_c k)
-{
-	struct bch_fs *c = trans->c;
-	u32 equiv = bch2_snapshot_equiv(c, k.k->p.snapshot);
+	if (!bpos_eq(*last_pos, k.k->p) &&
+	    snapshot_list_has_id(equiv_seen, equiv))
+		return bch2_btree_delete_at(trans, iter,
+					    BTREE_UPDATE_internal_snapshot_node);
+
+	*last_pos = k.k->p;
+
+	ret = snapshot_list_add_nodup(c, equiv_seen, equiv);
+	if (ret)
+		return ret;
 
 	/*
 	 * When we have a linear chain of snapshot nodes, we consider
@@ -1215,31 +1415,30 @@ static int move_key_to_correct_snapshot(struct btree_trans *trans,
 	 *
 	 * If there are multiple keys in different snapshots at the same
 	 * position, we're only going to keep the one in the newest
-	 * snapshot - the rest have been overwritten and are redundant,
-	 * and for the key we're going to keep we need to move it to the
-	 * equivalance class ID if it's not there already.
+	 * snapshot (we delete the others above) - the rest have been
+	 * overwritten and are redundant, and for the key we're going to keep we
+	 * need to move it to the equivalance class ID if it's not there
+	 * already.
 	 */
 	if (equiv != k.k->p.snapshot) {
 		struct bkey_i *new = bch2_bkey_make_mut_noupdate(trans, k);
-		struct btree_iter new_iter;
-		int ret;
-
-		ret = PTR_ERR_OR_ZERO(new);
+		int ret = PTR_ERR_OR_ZERO(new);
 		if (ret)
 			return ret;
 
 		new->k.p.snapshot = equiv;
 
+		struct btree_iter new_iter;
 		bch2_trans_iter_init(trans, &new_iter, iter->btree_id, new->k.p,
-				     BTREE_ITER_ALL_SNAPSHOTS|
-				     BTREE_ITER_CACHED|
-				     BTREE_ITER_INTENT);
+				     BTREE_ITER_all_snapshots|
+				     BTREE_ITER_cached|
+				     BTREE_ITER_intent);
 
 		ret =   bch2_btree_iter_traverse(&new_iter) ?:
 			bch2_trans_update(trans, &new_iter, new,
-					BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE) ?:
+					BTREE_UPDATE_internal_snapshot_node) ?:
 			bch2_btree_delete_at(trans, iter,
-					BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
+					BTREE_UPDATE_internal_snapshot_node);
 		bch2_trans_iter_exit(trans, &new_iter);
 		if (ret)
 			return ret;
@@ -1364,19 +1563,11 @@ int bch2_delete_dead_snapshots(struct bch_fs *c)
 	struct btree_trans *trans;
 	snapshot_id_list deleted = { 0 };
 	snapshot_id_list deleted_interior = { 0 };
-	u32 id;
 	int ret = 0;
 
 	if (!test_and_clear_bit(BCH_FS_need_delete_dead_snapshots, &c->flags))
 		return 0;
 
-	if (!test_bit(BCH_FS_started, &c->flags)) {
-		ret = bch2_fs_read_write_early(c);
-		bch_err_msg(c, ret, "deleting dead snapshots: error going rw");
-		if (ret)
-			return ret;
-	}
-
 	trans = bch2_trans_get(c);
 
 	/*
@@ -1411,33 +1602,20 @@ int bch2_delete_dead_snapshots(struct bch_fs *c)
 	if (ret)
 		goto err;
 
-	for (id = 0; id < BTREE_ID_NR; id++) {
+	for (unsigned btree = 0; btree < BTREE_ID_NR; btree++) {
 		struct bpos last_pos = POS_MIN;
 		snapshot_id_list equiv_seen = { 0 };
 		struct disk_reservation res = { 0 };
 
-		if (!btree_type_has_snapshots(id))
-			continue;
-
-		/*
-		 * deleted inodes btree is maintained by a trigger on the inodes
-		 * btree - no work for us to do here, and it's not safe to scan
-		 * it because we'll see out of date keys due to the btree write
-		 * buffer:
-		 */
-		if (id == BTREE_ID_deleted_inodes)
+		if (!btree_type_has_snapshots(btree))
 			continue;
 
 		ret = for_each_btree_key_commit(trans, iter,
-				id, POS_MIN,
-				BTREE_ITER_PREFETCH|BTREE_ITER_ALL_SNAPSHOTS, k,
-				&res, NULL, BCH_TRANS_COMMIT_no_enospc,
-			snapshot_delete_key(trans, &iter, k, &deleted, &equiv_seen, &last_pos)) ?:
-		      for_each_btree_key_commit(trans, iter,
-				id, POS_MIN,
-				BTREE_ITER_PREFETCH|BTREE_ITER_ALL_SNAPSHOTS, k,
+				btree, POS_MIN,
+				BTREE_ITER_prefetch|BTREE_ITER_all_snapshots, k,
 				&res, NULL, BCH_TRANS_COMMIT_no_enospc,
-			move_key_to_correct_snapshot(trans, &iter, k));
+			delete_dead_snapshots_process_key(trans, &iter, k, &deleted,
+							  &equiv_seen, &last_pos));
 
 		bch2_disk_reservation_put(c, &res);
 		darray_exit(&equiv_seen);
@@ -1470,7 +1648,7 @@ int bch2_delete_dead_snapshots(struct bch_fs *c)
 	 * nodes some depth fields will be off:
 	 */
 	ret = for_each_btree_key_commit(trans, iter, BTREE_ID_snapshots, POS_MIN,
-				  BTREE_ITER_INTENT, k,
+				  BTREE_ITER_intent, k,
 				  NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
 		bch2_fix_child_of_deleted_snapshot(trans, &iter, k, &deleted_interior));
 	if (ret)
@@ -1505,6 +1683,8 @@ void bch2_delete_dead_snapshots_work(struct work_struct *work)
 {
 	struct bch_fs *c = container_of(work, struct bch_fs, snapshot_delete_work);
 
+	set_worker_desc("bcachefs-delete-dead-snapshots/%s", c->name);
+
 	bch2_delete_dead_snapshots(c);
 	bch2_write_ref_put(c, BCH_WRITE_REF_delete_dead_snapshots);
 }
@@ -1526,8 +1706,8 @@ int __bch2_key_has_snapshot_overwrites(struct btree_trans *trans,
 	int ret;
 
 	bch2_trans_iter_init(trans, &iter, id, pos,
-			     BTREE_ITER_NOT_EXTENTS|
-			     BTREE_ITER_ALL_SNAPSHOTS);
+			     BTREE_ITER_not_extents|
+			     BTREE_ITER_all_snapshots);
 	while (1) {
 		k = bch2_btree_iter_prev(&iter);
 		ret = bkey_err(k);
@@ -1579,7 +1759,7 @@ static int bch2_propagate_key_to_snapshot_leaf(struct btree_trans *trans,
 
 	pos.snapshot = leaf_id;
 
-	bch2_trans_iter_init(trans, &iter, btree, pos, BTREE_ITER_INTENT);
+	bch2_trans_iter_init(trans, &iter, btree, pos, BTREE_ITER_intent);
 	k = bch2_btree_iter_peek_slot(&iter);
 	ret = bkey_err(k);
 	if (ret)
@@ -1678,6 +1858,20 @@ int bch2_snapshots_read(struct bch_fs *c)
 				   POS_MIN, 0, k,
 			   (set_is_ancestor_bitmap(c, k.k->p.offset), 0)));
 	bch_err_fn(c, ret);
+
+	/*
+	 * It's important that we check if we need to reconstruct snapshots
+	 * before going RW, so we mark that pass as required in the superblock -
+	 * otherwise, we could end up deleting keys with missing snapshot nodes
+	 * instead
+	 */
+	BUG_ON(!test_bit(BCH_FS_new_fs, &c->flags) &&
+	       test_bit(BCH_FS_may_go_rw, &c->flags));
+
+	if (bch2_err_matches(ret, EIO) ||
+	    (c->sb.btrees_lost_data & BIT_ULL(BTREE_ID_snapshots)))
+		ret = bch2_run_explicit_recovery_pass_persistent(c, BCH_RECOVERY_PASS_reconstruct_snapshots);
+
 	return ret;
 }
 
diff --git a/fs/bcachefs/snapshot.h b/fs/bcachefs/snapshot.h
index 7c66ffc063..31b0ee03e9 100644
--- a/fs/bcachefs/snapshot.h
+++ b/fs/bcachefs/snapshot.h
@@ -2,11 +2,11 @@
 #ifndef _BCACHEFS_SNAPSHOT_H
 #define _BCACHEFS_SNAPSHOT_H
 
-enum bkey_invalid_flags;
+enum bch_validate_flags;
 
 void bch2_snapshot_tree_to_text(struct printbuf *, struct bch_fs *, struct bkey_s_c);
 int bch2_snapshot_tree_invalid(struct bch_fs *, struct bkey_s_c,
-			       enum bkey_invalid_flags, struct printbuf *);
+			       enum bch_validate_flags, struct printbuf *);
 
 #define bch2_bkey_ops_snapshot_tree ((struct bkey_ops) {	\
 	.key_invalid	= bch2_snapshot_tree_invalid,		\
@@ -20,9 +20,10 @@ int bch2_snapshot_tree_lookup(struct btree_trans *, u32, struct bch_snapshot_tre
 
 void bch2_snapshot_to_text(struct printbuf *, struct bch_fs *, struct bkey_s_c);
 int bch2_snapshot_invalid(struct bch_fs *, struct bkey_s_c,
-			  enum bkey_invalid_flags, struct printbuf *);
+			  enum bch_validate_flags, struct printbuf *);
 int bch2_mark_snapshot(struct btree_trans *, enum btree_id, unsigned,
-		       struct bkey_s_c, struct bkey_s, unsigned);
+		       struct bkey_s_c, struct bkey_s,
+		       enum btree_iter_update_trigger_flags);
 
 #define bch2_bkey_ops_snapshot ((struct bkey_ops) {		\
 	.key_invalid	= bch2_snapshot_invalid,		\
@@ -33,7 +34,11 @@ int bch2_mark_snapshot(struct btree_trans *, enum btree_id, unsigned,
 
 static inline struct snapshot_t *__snapshot_t(struct snapshot_table *t, u32 id)
 {
-	return &t->s[U32_MAX - id];
+	u32 idx = U32_MAX - id;
+
+	return likely(t && idx < t->nr)
+		? &t->s[idx]
+		: NULL;
 }
 
 static inline const struct snapshot_t *snapshot_t(struct bch_fs *c, u32 id)
@@ -44,7 +49,8 @@ static inline const struct snapshot_t *snapshot_t(struct bch_fs *c, u32 id)
 static inline u32 bch2_snapshot_tree(struct bch_fs *c, u32 id)
 {
 	rcu_read_lock();
-	id = snapshot_t(c, id)->tree;
+	const struct snapshot_t *s = snapshot_t(c, id);
+	id = s ? s->tree : 0;
 	rcu_read_unlock();
 
 	return id;
@@ -52,7 +58,8 @@ static inline u32 bch2_snapshot_tree(struct bch_fs *c, u32 id)
 
 static inline u32 __bch2_snapshot_parent_early(struct bch_fs *c, u32 id)
 {
-	return snapshot_t(c, id)->parent;
+	const struct snapshot_t *s = snapshot_t(c, id);
+	return s ? s->parent : 0;
 }
 
 static inline u32 bch2_snapshot_parent_early(struct bch_fs *c, u32 id)
@@ -66,19 +73,19 @@ static inline u32 bch2_snapshot_parent_early(struct bch_fs *c, u32 id)
 
 static inline u32 __bch2_snapshot_parent(struct bch_fs *c, u32 id)
 {
-#ifdef CONFIG_BCACHEFS_DEBUG
-	u32 parent = snapshot_t(c, id)->parent;
+	const struct snapshot_t *s = snapshot_t(c, id);
+	if (!s)
+		return 0;
 
-	if (parent &&
-	    snapshot_t(c, id)->depth != snapshot_t(c, parent)->depth + 1)
+	u32 parent = s->parent;
+	if (IS_ENABLED(CONFIG_BCACHEFS_DEBUG) &&
+	    parent &&
+	    s->depth != snapshot_t(c, parent)->depth + 1)
 		panic("id %u depth=%u parent %u depth=%u\n",
 		      id, snapshot_t(c, id)->depth,
 		      parent, snapshot_t(c, parent)->depth);
 
 	return parent;
-#else
-	return snapshot_t(c, id)->parent;
-#endif
 }
 
 static inline u32 bch2_snapshot_parent(struct bch_fs *c, u32 id)
@@ -116,7 +123,8 @@ static inline u32 bch2_snapshot_root(struct bch_fs *c, u32 id)
 
 static inline u32 __bch2_snapshot_equiv(struct bch_fs *c, u32 id)
 {
-	return snapshot_t(c, id)->equiv;
+	const struct snapshot_t *s = snapshot_t(c, id);
+	return s ? s->equiv : 0;
 }
 
 static inline u32 bch2_snapshot_equiv(struct bch_fs *c, u32 id)
@@ -128,43 +136,22 @@ static inline u32 bch2_snapshot_equiv(struct bch_fs *c, u32 id)
 	return id;
 }
 
-static inline bool bch2_snapshot_is_equiv(struct bch_fs *c, u32 id)
+static inline int bch2_snapshot_is_internal_node(struct bch_fs *c, u32 id)
 {
-	return id == bch2_snapshot_equiv(c, id);
-}
-
-static inline bool bch2_snapshot_is_internal_node(struct bch_fs *c, u32 id)
-{
-	const struct snapshot_t *s;
-	bool ret;
-
 	rcu_read_lock();
-	s = snapshot_t(c, id);
-	ret = s->children[0];
+	const struct snapshot_t *s = snapshot_t(c, id);
+	int ret = s ? s->children[0] : -BCH_ERR_invalid_snapshot_node;
 	rcu_read_unlock();
 
 	return ret;
 }
 
-static inline u32 bch2_snapshot_is_leaf(struct bch_fs *c, u32 id)
+static inline int bch2_snapshot_is_leaf(struct bch_fs *c, u32 id)
 {
-	return !bch2_snapshot_is_internal_node(c, id);
-}
-
-static inline u32 bch2_snapshot_sibling(struct bch_fs *c, u32 id)
-{
-	const struct snapshot_t *s;
-	u32 parent = __bch2_snapshot_parent(c, id);
-
-	if (!parent)
-		return 0;
-
-	s = snapshot_t(c, __bch2_snapshot_parent(c, id));
-	if (id == s->children[0])
-		return s->children[1];
-	if (id == s->children[1])
-		return s->children[0];
-	return 0;
+	int ret = bch2_snapshot_is_internal_node(c, id);
+	if (ret < 0)
+		return ret;
+	return !ret;
 }
 
 static inline u32 bch2_snapshot_depth(struct bch_fs *c, u32 parent)
@@ -189,12 +176,9 @@ static inline bool bch2_snapshot_is_ancestor(struct bch_fs *c, u32 id, u32 ances
 
 static inline bool bch2_snapshot_has_children(struct bch_fs *c, u32 id)
 {
-	const struct snapshot_t *t;
-	bool ret;
-
 	rcu_read_lock();
-	t = snapshot_t(c, id);
-	ret = (t->children[0]|t->children[1]) != 0;
+	const struct snapshot_t *t = snapshot_t(c, id);
+	bool ret = t && (t->children[0]|t->children[1]) != 0;
 	rcu_read_unlock();
 
 	return ret;
@@ -218,15 +202,34 @@ static inline bool snapshot_list_has_ancestor(struct bch_fs *c, snapshot_id_list
 
 static inline int snapshot_list_add(struct bch_fs *c, snapshot_id_list *s, u32 id)
 {
-	int ret;
-
 	BUG_ON(snapshot_list_has_id(s, id));
-	ret = darray_push(s, id);
+	int ret = darray_push(s, id);
 	if (ret)
 		bch_err(c, "error reallocating snapshot_id_list (size %zu)", s->size);
 	return ret;
 }
 
+static inline int snapshot_list_add_nodup(struct bch_fs *c, snapshot_id_list *s, u32 id)
+{
+	int ret = snapshot_list_has_id(s, id)
+		? 0
+		: darray_push(s, id);
+	if (ret)
+		bch_err(c, "error reallocating snapshot_id_list (size %zu)", s->size);
+	return ret;
+}
+
+static inline int snapshot_list_merge(struct bch_fs *c, snapshot_id_list *dst, snapshot_id_list *src)
+{
+	darray_for_each(*src, i) {
+		int ret = snapshot_list_add_nodup(c, dst, *i);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
 int bch2_snapshot_lookup(struct btree_trans *trans, u32 id,
 			 struct bch_snapshot *s);
 int bch2_snapshot_get_subvol(struct btree_trans *, u32,
@@ -238,6 +241,8 @@ int bch2_snapshot_node_create(struct btree_trans *, u32,
 
 int bch2_check_snapshot_trees(struct bch_fs *);
 int bch2_check_snapshots(struct bch_fs *);
+int bch2_reconstruct_snapshots(struct bch_fs *);
+int bch2_check_key_has_snapshot(struct btree_trans *, struct btree_iter *, struct bkey_s_c);
 
 int bch2_snapshot_node_set_deleted(struct btree_trans *, u32);
 void bch2_delete_dead_snapshots_work(struct work_struct *);
@@ -249,7 +254,7 @@ static inline int bch2_key_has_snapshot_overwrites(struct btree_trans *trans,
 					  struct bpos pos)
 {
 	if (!btree_type_has_snapshots(id) ||
-	    bch2_snapshot_is_leaf(trans->c, pos.snapshot))
+	    bch2_snapshot_is_leaf(trans->c, pos.snapshot) > 0)
 		return 0;
 
 	return __bch2_key_has_snapshot_overwrites(trans, id, pos);
diff --git a/fs/bcachefs/str_hash.h b/fs/bcachefs/str_hash.h
index fcaa5a8887..c8c266cb57 100644
--- a/fs/bcachefs/str_hash.h
+++ b/fs/bcachefs/str_hash.h
@@ -15,16 +15,6 @@
 #include <crypto/hash.h>
 #include <crypto/sha2.h>
 
-typedef unsigned __bitwise bch_str_hash_flags_t;
-
-enum bch_str_hash_flags {
-	__BCH_HASH_SET_MUST_CREATE,
-	__BCH_HASH_SET_MUST_REPLACE,
-};
-
-#define BCH_HASH_SET_MUST_CREATE	(__force bch_str_hash_flags_t) BIT(__BCH_HASH_SET_MUST_CREATE)
-#define BCH_HASH_SET_MUST_REPLACE	(__force bch_str_hash_flags_t) BIT(__BCH_HASH_SET_MUST_REPLACE)
-
 static inline enum bch_str_hash_type
 bch2_str_hash_opt_to_type(struct bch_fs *c, enum bch_str_hash_opts opt)
 {
@@ -159,13 +149,14 @@ static inline bool is_visible_key(struct bch_hash_desc desc, subvol_inum inum, s
 		 desc.is_visible(inum, k));
 }
 
-static __always_inline int
+static __always_inline struct bkey_s_c
 bch2_hash_lookup_in_snapshot(struct btree_trans *trans,
 		 struct btree_iter *iter,
 		 const struct bch_hash_desc desc,
 		 const struct bch_hash_info *info,
 		 subvol_inum inum, const void *key,
-		 unsigned flags, u32 snapshot)
+		 enum btree_iter_update_trigger_flags flags,
+		 u32 snapshot)
 {
 	struct bkey_s_c k;
 	int ret;
@@ -173,10 +164,10 @@ bch2_hash_lookup_in_snapshot(struct btree_trans *trans,
 	for_each_btree_key_upto_norestart(trans, *iter, desc.btree_id,
 			   SPOS(inum.inum, desc.hash_key(info, key), snapshot),
 			   POS(inum.inum, U64_MAX),
-			   BTREE_ITER_SLOTS|flags, k, ret) {
+			   BTREE_ITER_slots|flags, k, ret) {
 		if (is_visible_key(desc, inum, k)) {
 			if (!desc.cmp_key(k, key))
-				return 0;
+				return k;
 		} else if (k.k->type == KEY_TYPE_hash_whiteout) {
 			;
 		} else {
@@ -186,20 +177,23 @@ bch2_hash_lookup_in_snapshot(struct btree_trans *trans,
 	}
 	bch2_trans_iter_exit(trans, iter);
 
-	return ret ?: -BCH_ERR_ENOENT_str_hash_lookup;
+	return bkey_s_c_err(ret ?: -BCH_ERR_ENOENT_str_hash_lookup);
 }
 
-static __always_inline int
+static __always_inline struct bkey_s_c
 bch2_hash_lookup(struct btree_trans *trans,
 		 struct btree_iter *iter,
 		 const struct bch_hash_desc desc,
 		 const struct bch_hash_info *info,
 		 subvol_inum inum, const void *key,
-		 unsigned flags)
+		 enum btree_iter_update_trigger_flags flags)
 {
 	u32 snapshot;
-	return  bch2_subvolume_get_snapshot(trans, inum.subvol, &snapshot) ?:
-		bch2_hash_lookup_in_snapshot(trans, iter, desc, info, inum, key, flags, snapshot);
+	int ret = bch2_subvolume_get_snapshot(trans, inum.subvol, &snapshot);
+	if (ret)
+		return bkey_s_c_err(ret);
+
+	return bch2_hash_lookup_in_snapshot(trans, iter, desc, info, inum, key, flags, snapshot);
 }
 
 static __always_inline int
@@ -220,7 +214,7 @@ bch2_hash_hole(struct btree_trans *trans,
 	for_each_btree_key_upto_norestart(trans, *iter, desc.btree_id,
 			   SPOS(inum.inum, desc.hash_key(info, key), snapshot),
 			   POS(inum.inum, U64_MAX),
-			   BTREE_ITER_SLOTS|BTREE_ITER_INTENT, k, ret)
+			   BTREE_ITER_slots|BTREE_ITER_intent, k, ret)
 		if (!is_visible_key(desc, inum, k))
 			return 0;
 	bch2_trans_iter_exit(trans, iter);
@@ -242,7 +236,7 @@ int bch2_hash_needs_whiteout(struct btree_trans *trans,
 
 	bch2_btree_iter_advance(&iter);
 
-	for_each_btree_key_continue_norestart(iter, BTREE_ITER_SLOTS, k, ret) {
+	for_each_btree_key_continue_norestart(iter, BTREE_ITER_slots, k, ret) {
 		if (k.k->type != desc.key_type &&
 		    k.k->type != KEY_TYPE_hash_whiteout)
 			break;
@@ -259,13 +253,12 @@ int bch2_hash_needs_whiteout(struct btree_trans *trans,
 }
 
 static __always_inline
-int bch2_hash_set_snapshot(struct btree_trans *trans,
+int bch2_hash_set_in_snapshot(struct btree_trans *trans,
 			   const struct bch_hash_desc desc,
 			   const struct bch_hash_info *info,
 			   subvol_inum inum, u32 snapshot,
 			   struct bkey_i *insert,
-			   bch_str_hash_flags_t str_hash_flags,
-			   int update_flags)
+			   enum btree_iter_update_trigger_flags flags)
 {
 	struct btree_iter iter, slot = { NULL };
 	struct bkey_s_c k;
@@ -277,7 +270,7 @@ int bch2_hash_set_snapshot(struct btree_trans *trans,
 				desc.hash_bkey(info, bkey_i_to_s_c(insert)),
 				snapshot),
 			   POS(insert->k.p.inode, U64_MAX),
-			   BTREE_ITER_SLOTS|BTREE_ITER_INTENT, k, ret) {
+			   BTREE_ITER_slots|BTREE_ITER_intent, k, ret) {
 		if (is_visible_key(desc, inum, k)) {
 			if (!desc.cmp_bkey(k, bkey_i_to_s_c(insert)))
 				goto found;
@@ -286,8 +279,7 @@ int bch2_hash_set_snapshot(struct btree_trans *trans,
 			continue;
 		}
 
-		if (!slot.path &&
-		    !(str_hash_flags & BCH_HASH_SET_MUST_REPLACE))
+		if (!slot.path && !(flags & STR_HASH_must_replace))
 			bch2_trans_copy_iter(&slot, &iter);
 
 		if (k.k->type != KEY_TYPE_hash_whiteout)
@@ -305,16 +297,16 @@ found:
 	found = true;
 not_found:
 
-	if (!found && (str_hash_flags & BCH_HASH_SET_MUST_REPLACE)) {
+	if (!found && (flags & STR_HASH_must_replace)) {
 		ret = -BCH_ERR_ENOENT_str_hash_set_must_replace;
-	} else if (found && (str_hash_flags & BCH_HASH_SET_MUST_CREATE)) {
-		ret = -EEXIST;
+	} else if (found && (flags & STR_HASH_must_create)) {
+		ret = -BCH_ERR_EEXIST_str_hash_set;
 	} else {
 		if (!found && slot.path)
 			swap(iter, slot);
 
 		insert->k.p = iter.pos;
-		ret = bch2_trans_update(trans, &iter, insert, update_flags);
+		ret = bch2_trans_update(trans, &iter, insert, flags);
 	}
 
 	goto out;
@@ -326,19 +318,14 @@ int bch2_hash_set(struct btree_trans *trans,
 		  const struct bch_hash_info *info,
 		  subvol_inum inum,
 		  struct bkey_i *insert,
-		  bch_str_hash_flags_t str_hash_flags)
+		  enum btree_iter_update_trigger_flags flags)
 {
-	u32 snapshot;
-	int ret;
-
-	ret = bch2_subvolume_get_snapshot(trans, inum.subvol, &snapshot);
-	if (ret)
-		return ret;
-
 	insert->k.p.inode = inum.inum;
 
-	return bch2_hash_set_snapshot(trans, desc, info, inum,
-				      snapshot, insert, str_hash_flags, 0);
+	u32 snapshot;
+	return  bch2_subvolume_get_snapshot(trans, inum.subvol, &snapshot) ?:
+		bch2_hash_set_in_snapshot(trans, desc, info, inum,
+					  snapshot, insert, flags);
 }
 
 static __always_inline
@@ -346,7 +333,7 @@ int bch2_hash_delete_at(struct btree_trans *trans,
 			const struct bch_hash_desc desc,
 			const struct bch_hash_info *info,
 			struct btree_iter *iter,
-			unsigned update_flags)
+			enum btree_iter_update_trigger_flags flags)
 {
 	struct bkey_i *delete;
 	int ret;
@@ -364,7 +351,7 @@ int bch2_hash_delete_at(struct btree_trans *trans,
 	delete->k.p = iter->pos;
 	delete->k.type = ret ? KEY_TYPE_hash_whiteout : KEY_TYPE_deleted;
 
-	return bch2_trans_update(trans, iter, delete, update_flags);
+	return bch2_trans_update(trans, iter, delete, flags);
 }
 
 static __always_inline
@@ -374,14 +361,10 @@ int bch2_hash_delete(struct btree_trans *trans,
 		     subvol_inum inum, const void *key)
 {
 	struct btree_iter iter;
-	int ret;
-
-	ret = bch2_hash_lookup(trans, &iter, desc, info, inum, key,
-				BTREE_ITER_INTENT);
-	if (ret)
-		return ret;
-
-	ret = bch2_hash_delete_at(trans, desc, info, &iter, 0);
+	struct bkey_s_c k = bch2_hash_lookup(trans, &iter, desc, info, inum, key,
+					     BTREE_ITER_intent);
+	int ret = bkey_err(k) ?:
+		  bch2_hash_delete_at(trans, desc, info, &iter, 0);
 	bch2_trans_iter_exit(trans, &iter);
 	return ret;
 }
diff --git a/fs/bcachefs/subvolume.c b/fs/bcachefs/subvolume.c
index 7c67c28d3e..dfc9cf3057 100644
--- a/fs/bcachefs/subvolume.c
+++ b/fs/bcachefs/subvolume.c
@@ -13,13 +13,26 @@
 
 static int bch2_subvolume_delete(struct btree_trans *, u32);
 
+static struct bpos subvolume_children_pos(struct bkey_s_c k)
+{
+	if (k.k->type != KEY_TYPE_subvolume)
+		return POS_MIN;
+
+	struct bkey_s_c_subvolume s = bkey_s_c_to_subvolume(k);
+	if (!s.v->fs_path_parent)
+		return POS_MIN;
+	return POS(le32_to_cpu(s.v->fs_path_parent), s.k->p.offset);
+}
+
 static int check_subvol(struct btree_trans *trans,
 			struct btree_iter *iter,
 			struct bkey_s_c k)
 {
 	struct bch_fs *c = trans->c;
 	struct bkey_s_c_subvolume subvol;
+	struct btree_iter subvol_children_iter = {};
 	struct bch_snapshot snapshot;
+	struct printbuf buf = PRINTBUF;
 	unsigned snapid;
 	int ret = 0;
 
@@ -42,6 +55,72 @@ static int check_subvol(struct btree_trans *trans,
 		return ret ?: -BCH_ERR_transaction_restart_nested;
 	}
 
+	if (fsck_err_on(subvol.k->p.offset == BCACHEFS_ROOT_SUBVOL &&
+			subvol.v->fs_path_parent,
+			c, subvol_root_fs_path_parent_nonzero,
+			"root subvolume has nonzero fs_path_parent\n%s",
+			(bch2_bkey_val_to_text(&buf, c, k), buf.buf))) {
+		struct bkey_i_subvolume *n =
+			bch2_bkey_make_mut_typed(trans, iter, &subvol.s_c, 0, subvolume);
+		ret = PTR_ERR_OR_ZERO(n);
+		if (ret)
+			goto err;
+
+		n->v.fs_path_parent = 0;
+	}
+
+	if (subvol.v->fs_path_parent) {
+		struct bpos pos = subvolume_children_pos(k);
+
+		struct bkey_s_c subvol_children_k =
+			bch2_bkey_get_iter(trans, &subvol_children_iter,
+					   BTREE_ID_subvolume_children, pos, 0);
+		ret = bkey_err(subvol_children_k);
+		if (ret)
+			goto err;
+
+		if (fsck_err_on(subvol_children_k.k->type != KEY_TYPE_set,
+				c, subvol_children_not_set,
+				"subvolume not set in subvolume_children btree at %llu:%llu\n%s",
+				pos.inode, pos.offset,
+				(printbuf_reset(&buf),
+				 bch2_bkey_val_to_text(&buf, c, k), buf.buf))) {
+			ret = bch2_btree_bit_mod(trans, BTREE_ID_subvolume_children, pos, true);
+			if (ret)
+				goto err;
+		}
+	}
+
+	struct bch_inode_unpacked inode;
+	struct btree_iter inode_iter = {};
+	ret = bch2_inode_peek_nowarn(trans, &inode_iter, &inode,
+				    (subvol_inum) { k.k->p.offset, le64_to_cpu(subvol.v->inode) },
+				    0);
+	bch2_trans_iter_exit(trans, &inode_iter);
+
+	if (ret && !bch2_err_matches(ret, ENOENT))
+		return ret;
+
+	if (fsck_err_on(ret, c, subvol_to_missing_root,
+			"subvolume %llu points to missing subvolume root %llu:%u",
+			k.k->p.offset, le64_to_cpu(subvol.v->inode),
+			le32_to_cpu(subvol.v->snapshot))) {
+		ret = bch2_subvolume_delete(trans, iter->pos.offset);
+		bch_err_msg(c, ret, "deleting subvolume %llu", iter->pos.offset);
+		return ret ?: -BCH_ERR_transaction_restart_nested;
+	}
+
+	if (fsck_err_on(inode.bi_subvol != subvol.k->p.offset,
+			c, subvol_root_wrong_bi_subvol,
+			"subvol root %llu:%u has wrong bi_subvol field: got %u, should be %llu",
+			inode.bi_inum, inode_iter.k.p.snapshot,
+			inode.bi_subvol, subvol.k->p.offset)) {
+		inode.bi_subvol = subvol.k->p.offset;
+		ret = __bch2_fsck_write_inode(trans, &inode, le32_to_cpu(subvol.v->snapshot));
+		if (ret)
+			goto err;
+	}
+
 	if (!BCH_SUBVOLUME_SNAP(subvol.v)) {
 		u32 snapshot_root = bch2_snapshot_root(c, le32_to_cpu(subvol.v->snapshot));
 		u32 snapshot_tree;
@@ -72,8 +151,10 @@ static int check_subvol(struct btree_trans *trans,
 			SET_BCH_SUBVOLUME_SNAP(&s->v, true);
 		}
 	}
-
+err:
 fsck_err:
+	bch2_trans_iter_exit(trans, &subvol_children_iter);
+	printbuf_exit(&buf);
 	return ret;
 }
 
@@ -81,24 +162,69 @@ int bch2_check_subvols(struct bch_fs *c)
 {
 	int ret = bch2_trans_run(c,
 		for_each_btree_key_commit(trans, iter,
-				BTREE_ID_subvolumes, POS_MIN, BTREE_ITER_PREFETCH, k,
+				BTREE_ID_subvolumes, POS_MIN, BTREE_ITER_prefetch, k,
 				NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
 			check_subvol(trans, &iter, k)));
 	bch_err_fn(c, ret);
 	return ret;
 }
 
+static int check_subvol_child(struct btree_trans *trans,
+			      struct btree_iter *child_iter,
+			      struct bkey_s_c child_k)
+{
+	struct bch_fs *c = trans->c;
+	struct bch_subvolume s;
+	int ret = bch2_bkey_get_val_typed(trans, BTREE_ID_subvolumes, POS(0, child_k.k->p.offset),
+					  0, subvolume, &s);
+	if (ret && !bch2_err_matches(ret, ENOENT))
+		return ret;
+
+	if (fsck_err_on(ret ||
+			le32_to_cpu(s.fs_path_parent) != child_k.k->p.inode,
+			c, subvol_children_bad,
+			"incorrect entry in subvolume_children btree %llu:%llu",
+			child_k.k->p.inode, child_k.k->p.offset)) {
+		ret = bch2_btree_delete_at(trans, child_iter, 0);
+		if (ret)
+			goto err;
+	}
+err:
+fsck_err:
+	return ret;
+}
+
+int bch2_check_subvol_children(struct bch_fs *c)
+{
+	int ret = bch2_trans_run(c,
+		for_each_btree_key_commit(trans, iter,
+				BTREE_ID_subvolume_children, POS_MIN, BTREE_ITER_prefetch, k,
+				NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
+			check_subvol_child(trans, &iter, k)));
+	bch_err_fn(c, ret);
+	return 0;
+}
+
 /* Subvolumes: */
 
 int bch2_subvolume_invalid(struct bch_fs *c, struct bkey_s_c k,
-			   enum bkey_invalid_flags flags, struct printbuf *err)
+			   enum bch_validate_flags flags, struct printbuf *err)
 {
+	struct bkey_s_c_subvolume subvol = bkey_s_c_to_subvolume(k);
 	int ret = 0;
 
 	bkey_fsck_err_on(bkey_lt(k.k->p, SUBVOL_POS_MIN) ||
 			 bkey_gt(k.k->p, SUBVOL_POS_MAX), c, err,
 			 subvol_pos_bad,
 			 "invalid pos");
+
+	bkey_fsck_err_on(!subvol.v->snapshot, c, err,
+			 subvol_snapshot_bad,
+			 "invalid snapshot");
+
+	bkey_fsck_err_on(!subvol.v->inode, c, err,
+			 subvol_inode_bad,
+			 "invalid inode");
 fsck_err:
 	return ret;
 }
@@ -112,8 +238,50 @@ void bch2_subvolume_to_text(struct printbuf *out, struct bch_fs *c,
 		   le64_to_cpu(s.v->inode),
 		   le32_to_cpu(s.v->snapshot));
 
-	if (bkey_val_bytes(s.k) > offsetof(struct bch_subvolume, parent))
-		prt_printf(out, " parent %u", le32_to_cpu(s.v->parent));
+	if (bkey_val_bytes(s.k) > offsetof(struct bch_subvolume, creation_parent)) {
+		prt_printf(out, " creation_parent %u", le32_to_cpu(s.v->creation_parent));
+		prt_printf(out, " fs_parent %u", le32_to_cpu(s.v->fs_path_parent));
+	}
+}
+
+static int subvolume_children_mod(struct btree_trans *trans, struct bpos pos, bool set)
+{
+	return !bpos_eq(pos, POS_MIN)
+		? bch2_btree_bit_mod(trans, BTREE_ID_subvolume_children, pos, set)
+		: 0;
+}
+
+int bch2_subvolume_trigger(struct btree_trans *trans,
+			   enum btree_id btree_id, unsigned level,
+			   struct bkey_s_c old, struct bkey_s new,
+			   enum btree_iter_update_trigger_flags flags)
+{
+	if (flags & BTREE_TRIGGER_transactional) {
+		struct bpos children_pos_old = subvolume_children_pos(old);
+		struct bpos children_pos_new = subvolume_children_pos(new.s_c);
+
+		if (!bpos_eq(children_pos_old, children_pos_new)) {
+			int ret = subvolume_children_mod(trans, children_pos_old, false) ?:
+				  subvolume_children_mod(trans, children_pos_new, true);
+			if (ret)
+				return ret;
+		}
+	}
+
+	return 0;
+}
+
+int bch2_subvol_has_children(struct btree_trans *trans, u32 subvol)
+{
+	struct btree_iter iter;
+
+	bch2_trans_iter_init(trans, &iter, BTREE_ID_subvolume_children, POS(subvol, 0), 0);
+	struct bkey_s_c k = bch2_btree_iter_peek(&iter);
+	bch2_trans_iter_exit(trans, &iter);
+
+	return bkey_err(k) ?: k.k && k.k->p.inode == subvol
+		? -BCH_ERR_ENOTEMPTY_subvol_not_empty
+		: 0;
 }
 
 static __always_inline int
@@ -174,7 +342,7 @@ int bch2_subvolume_get_snapshot(struct btree_trans *trans, u32 subvolid,
 
 	subvol = bch2_bkey_get_iter_typed(trans, &iter,
 					  BTREE_ID_subvolumes, POS(0, subvolid),
-					  BTREE_ITER_CACHED|BTREE_ITER_WITH_UPDATES,
+					  BTREE_ITER_cached|BTREE_ITER_with_updates,
 					  subvolume);
 	ret = bkey_err(subvol);
 	bch2_fs_inconsistent_on(bch2_err_matches(ret, ENOENT), trans->c,
@@ -197,8 +365,8 @@ static int bch2_subvolume_reparent(struct btree_trans *trans,
 	if (k.k->type != KEY_TYPE_subvolume)
 		return 0;
 
-	if (bkey_val_bytes(k.k) > offsetof(struct bch_subvolume, parent) &&
-	    le32_to_cpu(bkey_s_c_to_subvolume(k).v->parent) != old_parent)
+	if (bkey_val_bytes(k.k) > offsetof(struct bch_subvolume, creation_parent) &&
+	    le32_to_cpu(bkey_s_c_to_subvolume(k).v->creation_parent) != old_parent)
 		return 0;
 
 	s = bch2_bkey_make_mut_typed(trans, iter, &k, 0, subvolume);
@@ -206,7 +374,7 @@ static int bch2_subvolume_reparent(struct btree_trans *trans,
 	if (ret)
 		return ret;
 
-	s->v.parent = cpu_to_le32(new_parent);
+	s->v.creation_parent = cpu_to_le32(new_parent);
 	return 0;
 }
 
@@ -224,12 +392,12 @@ static int bch2_subvolumes_reparent(struct btree_trans *trans, u32 subvolid_to_d
 
 	return lockrestart_do(trans,
 			bch2_subvolume_get(trans, subvolid_to_delete, true,
-				   BTREE_ITER_CACHED, &s)) ?:
+				   BTREE_ITER_cached, &s)) ?:
 		for_each_btree_key_commit(trans, iter,
-				BTREE_ID_subvolumes, POS_MIN, BTREE_ITER_PREFETCH, k,
+				BTREE_ID_subvolumes, POS_MIN, BTREE_ITER_prefetch, k,
 				NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
 			bch2_subvolume_reparent(trans, &iter, k,
-					subvolid_to_delete, le32_to_cpu(s.parent)));
+					subvolid_to_delete, le32_to_cpu(s.creation_parent)));
 }
 
 /*
@@ -245,7 +413,7 @@ static int __bch2_subvolume_delete(struct btree_trans *trans, u32 subvolid)
 
 	subvol = bch2_bkey_get_iter_typed(trans, &iter,
 				BTREE_ID_subvolumes, POS(0, subvolid),
-				BTREE_ITER_CACHED|BTREE_ITER_INTENT,
+				BTREE_ITER_cached|BTREE_ITER_intent,
 				subvolume);
 	ret = bkey_err(subvol);
 	bch2_fs_inconsistent_on(bch2_err_matches(ret, ENOENT), trans->c,
@@ -346,7 +514,7 @@ int bch2_subvolume_unlink(struct btree_trans *trans, u32 subvolid)
 
 	n = bch2_bkey_get_mut_typed(trans, &iter,
 			BTREE_ID_subvolumes, POS(0, subvolid),
-			BTREE_ITER_CACHED, subvolume);
+			BTREE_ITER_cached, subvolume);
 	ret = PTR_ERR_OR_ZERO(n);
 	if (unlikely(ret)) {
 		bch2_fs_inconsistent_on(bch2_err_matches(ret, ENOENT), trans->c,
@@ -360,6 +528,7 @@ int bch2_subvolume_unlink(struct btree_trans *trans, u32 subvolid)
 }
 
 int bch2_subvolume_create(struct btree_trans *trans, u64 inode,
+			  u32 parent_subvolid,
 			  u32 src_subvolid,
 			  u32 *new_subvolid,
 			  u32 *new_snapshotid,
@@ -387,7 +556,7 @@ int bch2_subvolume_create(struct btree_trans *trans, u64 inode,
 
 		src_subvol = bch2_bkey_get_mut_typed(trans, &src_iter,
 				BTREE_ID_subvolumes, POS(0, src_subvolid),
-				BTREE_ITER_CACHED, subvolume);
+				BTREE_ITER_cached, subvolume);
 		ret = PTR_ERR_OR_ZERO(src_subvol);
 		if (unlikely(ret)) {
 			bch2_fs_inconsistent_on(bch2_err_matches(ret, ENOENT), c,
@@ -416,12 +585,13 @@ int bch2_subvolume_create(struct btree_trans *trans, u64 inode,
 	if (ret)
 		goto err;
 
-	new_subvol->v.flags	= 0;
-	new_subvol->v.snapshot	= cpu_to_le32(new_nodes[0]);
-	new_subvol->v.inode	= cpu_to_le64(inode);
-	new_subvol->v.parent	= cpu_to_le32(src_subvolid);
-	new_subvol->v.otime.lo	= cpu_to_le64(bch2_current_time(c));
-	new_subvol->v.otime.hi	= 0;
+	new_subvol->v.flags		= 0;
+	new_subvol->v.snapshot		= cpu_to_le32(new_nodes[0]);
+	new_subvol->v.inode		= cpu_to_le64(inode);
+	new_subvol->v.creation_parent	= cpu_to_le32(src_subvolid);
+	new_subvol->v.fs_path_parent	= cpu_to_le32(parent_subvolid);
+	new_subvol->v.otime.lo		= cpu_to_le64(bch2_current_time(c));
+	new_subvol->v.otime.hi		= 0;
 
 	SET_BCH_SUBVOLUME_RO(&new_subvol->v, ro);
 	SET_BCH_SUBVOLUME_SNAP(&new_subvol->v, src_subvolid != 0);
@@ -434,6 +604,78 @@ err:
 	return ret;
 }
 
+int bch2_initialize_subvolumes(struct bch_fs *c)
+{
+	struct bkey_i_snapshot_tree	root_tree;
+	struct bkey_i_snapshot		root_snapshot;
+	struct bkey_i_subvolume		root_volume;
+	int ret;
+
+	bkey_snapshot_tree_init(&root_tree.k_i);
+	root_tree.k.p.offset		= 1;
+	root_tree.v.master_subvol	= cpu_to_le32(1);
+	root_tree.v.root_snapshot	= cpu_to_le32(U32_MAX);
+
+	bkey_snapshot_init(&root_snapshot.k_i);
+	root_snapshot.k.p.offset = U32_MAX;
+	root_snapshot.v.flags	= 0;
+	root_snapshot.v.parent	= 0;
+	root_snapshot.v.subvol	= cpu_to_le32(BCACHEFS_ROOT_SUBVOL);
+	root_snapshot.v.tree	= cpu_to_le32(1);
+	SET_BCH_SNAPSHOT_SUBVOL(&root_snapshot.v, true);
+
+	bkey_subvolume_init(&root_volume.k_i);
+	root_volume.k.p.offset = BCACHEFS_ROOT_SUBVOL;
+	root_volume.v.flags	= 0;
+	root_volume.v.snapshot	= cpu_to_le32(U32_MAX);
+	root_volume.v.inode	= cpu_to_le64(BCACHEFS_ROOT_INO);
+
+	ret =   bch2_btree_insert(c, BTREE_ID_snapshot_trees,	&root_tree.k_i, NULL, 0) ?:
+		bch2_btree_insert(c, BTREE_ID_snapshots,	&root_snapshot.k_i, NULL, 0) ?:
+		bch2_btree_insert(c, BTREE_ID_subvolumes,	&root_volume.k_i, NULL, 0);
+	bch_err_fn(c, ret);
+	return ret;
+}
+
+static int __bch2_fs_upgrade_for_subvolumes(struct btree_trans *trans)
+{
+	struct btree_iter iter;
+	struct bkey_s_c k;
+	struct bch_inode_unpacked inode;
+	int ret;
+
+	k = bch2_bkey_get_iter(trans, &iter, BTREE_ID_inodes,
+			       SPOS(0, BCACHEFS_ROOT_INO, U32_MAX), 0);
+	ret = bkey_err(k);
+	if (ret)
+		return ret;
+
+	if (!bkey_is_inode(k.k)) {
+		bch_err(trans->c, "root inode not found");
+		ret = -BCH_ERR_ENOENT_inode;
+		goto err;
+	}
+
+	ret = bch2_inode_unpack(k, &inode);
+	BUG_ON(ret);
+
+	inode.bi_subvol = BCACHEFS_ROOT_SUBVOL;
+
+	ret = bch2_inode_write(trans, &iter, &inode);
+err:
+	bch2_trans_iter_exit(trans, &iter);
+	return ret;
+}
+
+/* set bi_subvol on root inode */
+int bch2_fs_upgrade_for_subvolumes(struct bch_fs *c)
+{
+	int ret = bch2_trans_do(c, NULL, NULL, BCH_TRANS_COMMIT_lazy_rw,
+				__bch2_fs_upgrade_for_subvolumes(trans));
+	bch_err_fn(c, ret);
+	return ret;
+}
+
 int bch2_fs_subvolumes_init(struct bch_fs *c)
 {
 	INIT_WORK(&c->snapshot_delete_work, bch2_delete_dead_snapshots_work);
diff --git a/fs/bcachefs/subvolume.h b/fs/bcachefs/subvolume.h
index a6f56f66e2..afa5e871ef 100644
--- a/fs/bcachefs/subvolume.h
+++ b/fs/bcachefs/subvolume.h
@@ -5,20 +5,26 @@
 #include "darray.h"
 #include "subvolume_types.h"
 
-enum bkey_invalid_flags;
+enum bch_validate_flags;
 
 int bch2_check_subvols(struct bch_fs *);
+int bch2_check_subvol_children(struct bch_fs *);
 
 int bch2_subvolume_invalid(struct bch_fs *, struct bkey_s_c,
-			   enum bkey_invalid_flags, struct printbuf *);
+			   enum bch_validate_flags, struct printbuf *);
 void bch2_subvolume_to_text(struct printbuf *, struct bch_fs *, struct bkey_s_c);
+int bch2_subvolume_trigger(struct btree_trans *, enum btree_id, unsigned,
+			   struct bkey_s_c, struct bkey_s,
+			   enum btree_iter_update_trigger_flags);
 
 #define bch2_bkey_ops_subvolume ((struct bkey_ops) {		\
 	.key_invalid	= bch2_subvolume_invalid,		\
 	.val_to_text	= bch2_subvolume_to_text,		\
+	.trigger	= bch2_subvolume_trigger,		\
 	.min_val_size	= 16,					\
 })
 
+int bch2_subvol_has_children(struct btree_trans *, u32);
 int bch2_subvolume_get(struct btree_trans *, unsigned,
 		       bool, int, struct bch_subvolume *);
 int bch2_subvolume_get_snapshot(struct btree_trans *, u32, u32 *);
@@ -30,8 +36,10 @@ int bch2_delete_dead_snapshots(struct bch_fs *);
 void bch2_delete_dead_snapshots_async(struct bch_fs *);
 
 int bch2_subvolume_unlink(struct btree_trans *, u32);
-int bch2_subvolume_create(struct btree_trans *, u64, u32,
-			  u32 *, u32 *, bool);
+int bch2_subvolume_create(struct btree_trans *, u64, u32, u32, u32 *, u32 *, bool);
+
+int bch2_initialize_subvolumes(struct bch_fs *);
+int bch2_fs_upgrade_for_subvolumes(struct bch_fs *);
 
 int bch2_fs_subvolumes_init(struct bch_fs *);
 
diff --git a/fs/bcachefs/subvolume_format.h b/fs/bcachefs/subvolume_format.h
index af79134b07..e029df7ba8 100644
--- a/fs/bcachefs/subvolume_format.h
+++ b/fs/bcachefs/subvolume_format.h
@@ -19,8 +19,8 @@ struct bch_subvolume {
 	 * This is _not_ necessarily the subvolume of the directory containing
 	 * this subvolume:
 	 */
-	__le32			parent;
-	__le32			pad;
+	__le32			creation_parent;
+	__le32			fs_path_parent;
 	bch_le128		otime;
 };
 
diff --git a/fs/bcachefs/subvolume_types.h b/fs/bcachefs/subvolume_types.h
index ae644adfc3..9b10c89478 100644
--- a/fs/bcachefs/subvolume_types.h
+++ b/fs/bcachefs/subvolume_types.h
@@ -20,6 +20,8 @@ struct snapshot_t {
 };
 
 struct snapshot_table {
+	struct rcu_head		rcu;
+	size_t			nr;
 #ifndef RUST_BINDGEN
 	DECLARE_FLEX_ARRAY(struct snapshot_t, s);
 #else
diff --git a/fs/bcachefs/super-io.c b/fs/bcachefs/super-io.c
index 36988add58..b156fc85b8 100644
--- a/fs/bcachefs/super-io.c
+++ b/fs/bcachefs/super-io.c
@@ -8,7 +8,7 @@
 #include "journal.h"
 #include "journal_sb.h"
 #include "journal_seq_blacklist.h"
-#include "recovery.h"
+#include "recovery_passes.h"
 #include "replicas.h"
 #include "quota.h"
 #include "sb-clean.h"
@@ -76,7 +76,7 @@ const char * const bch2_sb_fields[] = {
 };
 
 static int bch2_sb_field_validate(struct bch_sb *, struct bch_sb_field *,
-				  struct printbuf *);
+				  enum bch_validate_flags, struct printbuf *);
 
 struct bch_sb_field *bch2_sb_field_get_id(struct bch_sb *sb,
 				      enum bch_sb_field_type type)
@@ -142,8 +142,8 @@ void bch2_sb_field_delete(struct bch_sb_handle *sb,
 void bch2_free_super(struct bch_sb_handle *sb)
 {
 	kfree(sb->bio);
-	if (!IS_ERR_OR_NULL(sb->bdev_handle))
-		bdev_release(sb->bdev_handle);
+	if (!IS_ERR_OR_NULL(sb->s_bdev_file))
+		bdev_fput(sb->s_bdev_file);
 	kfree(sb->holder);
 	kfree(sb->sb_name);
 
@@ -232,7 +232,7 @@ struct bch_sb_field *bch2_sb_field_resize_id(struct bch_sb_handle *sb,
 			struct bch_sb_handle *dev_sb = &ca->disk_sb;
 
 			if (bch2_sb_realloc(dev_sb, le32_to_cpu(dev_sb->sb->u64s) + d)) {
-				percpu_ref_put(&ca->ref);
+				percpu_ref_put(&ca->io_ref);
 				return NULL;
 			}
 		}
@@ -344,8 +344,8 @@ static int bch2_sb_compatible(struct bch_sb *sb, struct printbuf *out)
 	return 0;
 }
 
-static int bch2_sb_validate(struct bch_sb_handle *disk_sb, struct printbuf *out,
-			    int rw)
+static int bch2_sb_validate(struct bch_sb_handle *disk_sb,
+			    enum bch_validate_flags flags, struct printbuf *out)
 {
 	struct bch_sb *sb = disk_sb->sb;
 	struct bch_sb_field_members_v1 *mi;
@@ -401,7 +401,7 @@ static int bch2_sb_validate(struct bch_sb_handle *disk_sb, struct printbuf *out,
 		return -BCH_ERR_invalid_sb_time_precision;
 	}
 
-	if (rw == READ) {
+	if (!flags) {
 		/*
 		 * Been seeing a bug where these are getting inexplicably
 		 * zeroed, so we're now validating them, but we have to be
@@ -457,7 +457,7 @@ static int bch2_sb_validate(struct bch_sb_handle *disk_sb, struct printbuf *out,
 		return -BCH_ERR_invalid_sb_members_missing;
 	}
 
-	ret = bch2_sb_field_validate(sb, &mi->field, out);
+	ret = bch2_sb_field_validate(sb, &mi->field, flags, out);
 	if (ret)
 		return ret;
 
@@ -465,11 +465,19 @@ static int bch2_sb_validate(struct bch_sb_handle *disk_sb, struct printbuf *out,
 		if (le32_to_cpu(f->type) == BCH_SB_FIELD_members_v1)
 			continue;
 
-		ret = bch2_sb_field_validate(sb, f, out);
+		ret = bch2_sb_field_validate(sb, f, flags, out);
 		if (ret)
 			return ret;
 	}
 
+	if ((flags & BCH_VALIDATE_write) &&
+	    bch2_sb_member_get(sb, sb->dev_idx).seq != sb->seq) {
+		prt_printf(out, "Invalid superblock: member seq %llu != sb seq %llu",
+			   le64_to_cpu(bch2_sb_member_get(sb, sb->dev_idx).seq),
+			   le64_to_cpu(sb->seq));
+		return -BCH_ERR_invalid_sb_members_missing;
+	}
+
 	return 0;
 }
 
@@ -519,9 +527,11 @@ static void bch2_sb_update(struct bch_fs *c)
 	memset(c->sb.errors_silent, 0, sizeof(c->sb.errors_silent));
 
 	struct bch_sb_field_ext *ext = bch2_sb_field_get(src, ext);
-	if (ext)
+	if (ext) {
 		le_bitvector_to_cpu(c->sb.errors_silent, (void *) ext->errors_silent,
 				    sizeof(c->sb.errors_silent) * 8);
+		c->sb.btrees_lost_data = le64_to_cpu(ext->btrees_lost_data);
+	}
 
 	for_each_member_device(c, ca) {
 		struct bch_member m = bch2_sb_member_get(src, ca->dev_idx);
@@ -639,9 +649,10 @@ reread:
 
 	bytes = vstruct_bytes(sb->sb);
 
-	if (bytes > 512 << sb->sb->layout.sb_max_size_bits) {
-		prt_printf(err, "Invalid superblock: too big (got %zu bytes, layout max %lu)",
-		       bytes, 512UL << sb->sb->layout.sb_max_size_bits);
+	u64 sb_size = 512ULL << min(BCH_SB_LAYOUT_SIZE_BITS_MAX, sb->sb->layout.sb_max_size_bits);
+	if (bytes > sb_size) {
+		prt_printf(err, "Invalid superblock: too big (got %zu bytes, layout max %llu)",
+			   bytes, sb_size);
 		return -BCH_ERR_invalid_sb_too_big;
 	}
 
@@ -690,8 +701,11 @@ retry:
 		return -ENOMEM;
 
 	sb->sb_name = kstrdup(path, GFP_KERNEL);
-	if (!sb->sb_name)
-		return -ENOMEM;
+	if (!sb->sb_name) {
+		ret = -ENOMEM;
+		prt_printf(&err, "error allocating memory for sb_name");
+		goto err;
+	}
 
 #ifndef __KERNEL__
 	if (opt_get(*opts, direct_io) == false)
@@ -704,22 +718,23 @@ retry:
 	if (!opt_get(*opts, nochanges))
 		sb->mode |= BLK_OPEN_WRITE;
 
-	sb->bdev_handle = bdev_open_by_path(path, sb->mode, sb->holder, &bch2_sb_handle_bdev_ops);
-	if (IS_ERR(sb->bdev_handle) &&
-	    PTR_ERR(sb->bdev_handle) == -EACCES &&
+	sb->s_bdev_file = bdev_file_open_by_path(path, sb->mode, sb->holder, &bch2_sb_handle_bdev_ops);
+	if (IS_ERR(sb->s_bdev_file) &&
+	    PTR_ERR(sb->s_bdev_file) == -EACCES &&
 	    opt_get(*opts, read_only)) {
 		sb->mode &= ~BLK_OPEN_WRITE;
 
-		sb->bdev_handle = bdev_open_by_path(path, sb->mode, sb->holder, &bch2_sb_handle_bdev_ops);
-		if (!IS_ERR(sb->bdev_handle))
+		sb->s_bdev_file = bdev_file_open_by_path(path, sb->mode, sb->holder, &bch2_sb_handle_bdev_ops);
+		if (!IS_ERR(sb->s_bdev_file))
 			opt_set(*opts, nochanges, true);
 	}
 
-	if (IS_ERR(sb->bdev_handle)) {
-		ret = PTR_ERR(sb->bdev_handle);
+	if (IS_ERR(sb->s_bdev_file)) {
+		ret = PTR_ERR(sb->s_bdev_file);
+		prt_printf(&err, "error opening %s: %s", path, bch2_err_str(ret));
 		goto err;
 	}
-	sb->bdev = sb->bdev_handle->bdev;
+	sb->bdev = file_bdev(sb->s_bdev_file);
 
 	ret = bch2_sb_realloc(sb, 0);
 	if (ret) {
@@ -743,9 +758,9 @@ retry:
 	prt_printf(&err2, "bcachefs (%s): error reading default superblock: %s\n",
 	       path, err.buf);
 	if (ret == -BCH_ERR_invalid_sb_magic && ignore_notbchfs_msg)
-		printk(KERN_INFO "%s", err2.buf);
+		bch2_print_opts(opts, KERN_INFO "%s", err2.buf);
 	else
-		printk(KERN_ERR "%s", err2.buf);
+		bch2_print_opts(opts, KERN_ERR "%s", err2.buf);
 
 	printbuf_exit(&err2);
 	printbuf_reset(&err);
@@ -803,21 +818,20 @@ got_super:
 		goto err;
 	}
 
-	ret = 0;
 	sb->have_layout = true;
 
-	ret = bch2_sb_validate(sb, &err, READ);
+	ret = bch2_sb_validate(sb, 0, &err);
 	if (ret) {
-		printk(KERN_ERR "bcachefs (%s): error validating superblock: %s\n",
-		       path, err.buf);
+		bch2_print_opts(opts, KERN_ERR "bcachefs (%s): error validating superblock: %s\n",
+				path, err.buf);
 		goto err_no_print;
 	}
 out:
 	printbuf_exit(&err);
 	return ret;
 err:
-	printk(KERN_ERR "bcachefs (%s): error reading superblock: %s\n",
-	       path, err.buf);
+	bch2_print_opts(opts, KERN_ERR "bcachefs (%s): error reading superblock: %s\n",
+			path, err.buf);
 err_no_print:
 	bch2_free_super(sb);
 	goto out;
@@ -910,6 +924,7 @@ int bch2_write_super(struct bch_fs *c)
 	struct bch_devs_mask sb_written;
 	bool wrote, can_mount_without_written, can_mount_with_written;
 	unsigned degraded_flags = BCH_FORCE_IF_DEGRADED;
+	DARRAY(struct bch_dev *) online_devices = {};
 	int ret = 0;
 
 	trace_and_count(c, write_super, c, _RET_IP_);
@@ -922,6 +937,15 @@ int bch2_write_super(struct bch_fs *c)
 	closure_init_stack(cl);
 	memset(&sb_written, 0, sizeof(sb_written));
 
+	for_each_online_member(c, ca) {
+		ret = darray_push(&online_devices, ca);
+		if (bch2_fs_fatal_err_on(ret, c, "%s: error allocating online devices", __func__)) {
+			percpu_ref_put(&ca->io_ref);
+			goto out;
+		}
+		percpu_ref_get(&ca->io_ref);
+	}
+
 	/* Make sure we're using the new magic numbers: */
 	c->disk_sb.sb->magic = BCHFS_MAGIC;
 	c->disk_sb.sb->layout.magic = BCHFS_MAGIC;
@@ -929,8 +953,8 @@ int bch2_write_super(struct bch_fs *c)
 	le64_add_cpu(&c->disk_sb.sb->seq, 1);
 
 	struct bch_sb_field_members_v2 *mi = bch2_sb_field_get(c->disk_sb.sb, members_v2);
-	for_each_online_member(c, ca)
-		__bch2_members_v2_get_mut(mi, ca->dev_idx)->seq = c->disk_sb.sb->seq;
+	darray_for_each(online_devices, ca)
+		__bch2_members_v2_get_mut(mi, (*ca)->dev_idx)->seq = c->disk_sb.sb->seq;
 	c->disk_sb.sb->write_time = cpu_to_le64(ktime_get_real_seconds());
 
 	if (test_bit(BCH_FS_error, &c->flags))
@@ -946,16 +970,15 @@ int bch2_write_super(struct bch_fs *c)
 	bch2_sb_errors_from_cpu(c);
 	bch2_sb_downgrade_update(c);
 
-	for_each_online_member(c, ca)
-		bch2_sb_from_fs(c, ca);
+	darray_for_each(online_devices, ca)
+		bch2_sb_from_fs(c, (*ca));
 
-	for_each_online_member(c, ca) {
+	darray_for_each(online_devices, ca) {
 		printbuf_reset(&err);
 
-		ret = bch2_sb_validate(&ca->disk_sb, &err, WRITE);
+		ret = bch2_sb_validate(&(*ca)->disk_sb, BCH_VALIDATE_write, &err);
 		if (ret) {
 			bch2_fs_inconsistent(c, "sb invalid before write: %s", err.buf);
-			percpu_ref_put(&ca->io_ref);
 			goto out;
 		}
 	}
@@ -977,58 +1000,72 @@ int bch2_write_super(struct bch_fs *c)
 		prt_str(&buf, " > ");
 		bch2_version_to_text(&buf, bcachefs_metadata_version_current);
 		prt_str(&buf, ")");
-		bch2_fs_fatal_error(c, "%s", buf.buf);
+		bch2_fs_fatal_error(c, ": %s", buf.buf);
 		printbuf_exit(&buf);
 		return -BCH_ERR_sb_not_downgraded;
 	}
 
-	for_each_online_member(c, ca) {
-		__set_bit(ca->dev_idx, sb_written.d);
-		ca->sb_write_error = 0;
+	darray_for_each(online_devices, ca) {
+		__set_bit((*ca)->dev_idx, sb_written.d);
+		(*ca)->sb_write_error = 0;
 	}
 
-	for_each_online_member(c, ca)
-		read_back_super(c, ca);
+	darray_for_each(online_devices, ca)
+		read_back_super(c, *ca);
 	closure_sync(cl);
 
-	for_each_online_member(c, ca) {
+	darray_for_each(online_devices, cap) {
+		struct bch_dev *ca = *cap;
+
 		if (ca->sb_write_error)
 			continue;
 
 		if (le64_to_cpu(ca->sb_read_scratch->seq) < ca->disk_sb.seq) {
-			bch2_fs_fatal_error(c,
-				"Superblock write was silently dropped! (seq %llu expected %llu)",
+			struct printbuf buf = PRINTBUF;
+			prt_char(&buf, ' ');
+			prt_bdevname(&buf, ca->disk_sb.bdev);
+			prt_printf(&buf,
+				": Superblock write was silently dropped! (seq %llu expected %llu)",
 				le64_to_cpu(ca->sb_read_scratch->seq),
 				ca->disk_sb.seq);
-			percpu_ref_put(&ca->io_ref);
+			bch2_fs_fatal_error(c, "%s", buf.buf);
+			printbuf_exit(&buf);
 			ret = -BCH_ERR_erofs_sb_err;
-			goto out;
 		}
 
 		if (le64_to_cpu(ca->sb_read_scratch->seq) > ca->disk_sb.seq) {
-			bch2_fs_fatal_error(c,
-				"Superblock modified by another process (seq %llu expected %llu)",
+			struct printbuf buf = PRINTBUF;
+			prt_char(&buf, ' ');
+			prt_bdevname(&buf, ca->disk_sb.bdev);
+			prt_printf(&buf,
+				": Superblock modified by another process (seq %llu expected %llu)",
 				le64_to_cpu(ca->sb_read_scratch->seq),
 				ca->disk_sb.seq);
-			percpu_ref_put(&ca->io_ref);
+			bch2_fs_fatal_error(c, "%s", buf.buf);
+			printbuf_exit(&buf);
 			ret = -BCH_ERR_erofs_sb_err;
-			goto out;
 		}
 	}
 
+	if (ret)
+		goto out;
+
 	do {
 		wrote = false;
-		for_each_online_member(c, ca)
+		darray_for_each(online_devices, cap) {
+			struct bch_dev *ca = *cap;
 			if (!ca->sb_write_error &&
 			    sb < ca->disk_sb.sb->layout.nr_superblocks) {
 				write_one_super(c, ca, sb);
 				wrote = true;
 			}
+		}
 		closure_sync(cl);
 		sb++;
 	} while (wrote);
 
-	for_each_online_member(c, ca) {
+	darray_for_each(online_devices, cap) {
+		struct bch_dev *ca = *cap;
 		if (ca->sb_write_error)
 			__clear_bit(ca->dev_idx, sb_written.d);
 		else
@@ -1058,12 +1095,15 @@ int bch2_write_super(struct bch_fs *c)
 				 !can_mount_with_written ||
 				 (can_mount_without_written &&
 				  !can_mount_with_written), c,
-		"Unable to write superblock to sufficient devices (from %ps)",
+		": Unable to write superblock to sufficient devices (from %ps)",
 		(void *) _RET_IP_))
 		ret = -1;
 out:
 	/* Make new options visible after they're persistent: */
 	bch2_sb_update(c);
+	darray_for_each(online_devices, ca)
+		percpu_ref_put(&(*ca)->io_ref);
+	darray_exit(&online_devices);
 	printbuf_exit(&err);
 	return ret;
 }
@@ -1093,18 +1133,12 @@ bool bch2_check_version_downgrade(struct bch_fs *c)
 	 * c->sb will be checked before we write the superblock, so update it as
 	 * well:
 	 */
-	if (BCH_SB_VERSION_UPGRADE_COMPLETE(c->disk_sb.sb) > bcachefs_metadata_version_current) {
+	if (BCH_SB_VERSION_UPGRADE_COMPLETE(c->disk_sb.sb) > bcachefs_metadata_version_current)
 		SET_BCH_SB_VERSION_UPGRADE_COMPLETE(c->disk_sb.sb, bcachefs_metadata_version_current);
-		c->sb.version_upgrade_complete = bcachefs_metadata_version_current;
-	}
-	if (c->sb.version > bcachefs_metadata_version_current) {
+	if (c->sb.version > bcachefs_metadata_version_current)
 		c->disk_sb.sb->version = cpu_to_le16(bcachefs_metadata_version_current);
-		c->sb.version = bcachefs_metadata_version_current;
-	}
-	if (c->sb.version_min > bcachefs_metadata_version_current) {
+	if (c->sb.version_min > bcachefs_metadata_version_current)
 		c->disk_sb.sb->version_min = cpu_to_le16(bcachefs_metadata_version_current);
-		c->sb.version_min = bcachefs_metadata_version_current;
-	}
 	c->disk_sb.sb->compat[0] &= cpu_to_le64((1ULL << BCH_COMPAT_NR) - 1);
 	return ret;
 }
@@ -1122,7 +1156,7 @@ void bch2_sb_upgrade(struct bch_fs *c, unsigned new_version)
 }
 
 static int bch2_sb_ext_validate(struct bch_sb *sb, struct bch_sb_field *f,
-				struct printbuf *err)
+				enum bch_validate_flags flags, struct printbuf *err)
 {
 	if (vstruct_bytes(f) < 88) {
 		prt_printf(err, "field too small (%zu < %u)", vstruct_bytes(f), 88);
@@ -1137,8 +1171,7 @@ static void bch2_sb_ext_to_text(struct printbuf *out, struct bch_sb *sb,
 {
 	struct bch_sb_field_ext *e = field_to_type(f, ext);
 
-	prt_printf(out, "Recovery passes required:");
-	prt_tab(out);
+	prt_printf(out, "Recovery passes required:\t");
 	prt_bitflags(out, bch2_recovery_passes,
 		     bch2_recovery_passes_from_stable(le64_to_cpu(e->recovery_passes_required[0])));
 	prt_newline(out);
@@ -1147,13 +1180,16 @@ static void bch2_sb_ext_to_text(struct printbuf *out, struct bch_sb *sb,
 	if (errors_silent) {
 		le_bitvector_to_cpu(errors_silent, (void *) e->errors_silent, sizeof(e->errors_silent) * 8);
 
-		prt_printf(out, "Errors to silently fix:");
-		prt_tab(out);
+		prt_printf(out, "Errors to silently fix:\t");
 		prt_bitflags_vector(out, bch2_sb_error_strs, errors_silent, sizeof(e->errors_silent) * 8);
 		prt_newline(out);
 
 		kfree(errors_silent);
 	}
+
+	prt_printf(out, "Btrees with missing data:\t");
+	prt_bitflags(out, __bch2_btree_ids, le64_to_cpu(e->btrees_lost_data));
+	prt_newline(out);
 }
 
 static const struct bch_sb_field_ops bch_sb_field_ops_ext = {
@@ -1178,14 +1214,14 @@ static const struct bch_sb_field_ops *bch2_sb_field_type_ops(unsigned type)
 }
 
 static int bch2_sb_field_validate(struct bch_sb *sb, struct bch_sb_field *f,
-				  struct printbuf *err)
+				  enum bch_validate_flags flags, struct printbuf *err)
 {
 	unsigned type = le32_to_cpu(f->type);
 	struct printbuf field_err = PRINTBUF;
 	const struct bch_sb_field_ops *ops = bch2_sb_field_type_ops(type);
 	int ret;
 
-	ret = ops->validate ? ops->validate(sb, f, &field_err) : 0;
+	ret = ops->validate ? ops->validate(sb, f, flags, &field_err) : 0;
 	if (ret) {
 		prt_printf(err, "Invalid superblock section %s: %s",
 			   bch2_sb_fields[type], field_err.buf);
@@ -1259,97 +1295,73 @@ void bch2_sb_to_text(struct printbuf *out, struct bch_sb *sb,
 		printbuf_tabstop_push(out, 44);
 
 	for (int i = 0; i < sb->nr_devices; i++)
-		nr_devices += bch2_dev_exists(sb, i);
+		nr_devices += bch2_member_exists(sb, i);
 
-	prt_printf(out, "External UUID:");
-	prt_tab(out);
+	prt_printf(out, "External UUID:\t");
 	pr_uuid(out, sb->user_uuid.b);
 	prt_newline(out);
 
-	prt_printf(out, "Internal UUID:");
-	prt_tab(out);
+	prt_printf(out, "Internal UUID:\t");
 	pr_uuid(out, sb->uuid.b);
 	prt_newline(out);
 
-	prt_printf(out, "Magic number:");
-	prt_tab(out);
+	prt_printf(out, "Magic number:\t");
 	pr_uuid(out, sb->magic.b);
 	prt_newline(out);
 
-	prt_str(out, "Device index:");
-	prt_tab(out);
-	prt_printf(out, "%u", sb->dev_idx);
-	prt_newline(out);
+	prt_printf(out, "Device index:\t%u\n", sb->dev_idx);
 
-	prt_str(out, "Label:");
-	prt_tab(out);
+	prt_printf(out, "Label:\t");
 	prt_printf(out, "%.*s", (int) sizeof(sb->label), sb->label);
 	prt_newline(out);
 
-	prt_str(out, "Version:");
-	prt_tab(out);
+	prt_printf(out, "Version:\t");
 	bch2_version_to_text(out, le16_to_cpu(sb->version));
 	prt_newline(out);
 
-	prt_str(out, "Version upgrade complete:");
-	prt_tab(out);
+	prt_printf(out, "Version upgrade complete:\t");
 	bch2_version_to_text(out, BCH_SB_VERSION_UPGRADE_COMPLETE(sb));
 	prt_newline(out);
 
-	prt_printf(out, "Oldest version on disk:");
-	prt_tab(out);
+	prt_printf(out, "Oldest version on disk:\t");
 	bch2_version_to_text(out, le16_to_cpu(sb->version_min));
 	prt_newline(out);
 
-	prt_printf(out, "Created:");
-	prt_tab(out);
+	prt_printf(out, "Created:\t");
 	if (sb->time_base_lo)
 		bch2_prt_datetime(out, div_u64(le64_to_cpu(sb->time_base_lo), NSEC_PER_SEC));
 	else
 		prt_printf(out, "(not set)");
 	prt_newline(out);
 
-	prt_printf(out, "Sequence number:");
-	prt_tab(out);
+	prt_printf(out, "Sequence number:\t");
 	prt_printf(out, "%llu", le64_to_cpu(sb->seq));
 	prt_newline(out);
 
-	prt_printf(out, "Time of last write:");
-	prt_tab(out);
+	prt_printf(out, "Time of last write:\t");
 	bch2_prt_datetime(out, le64_to_cpu(sb->write_time));
 	prt_newline(out);
 
-	prt_printf(out, "Superblock size:");
-	prt_tab(out);
+	prt_printf(out, "Superblock size:\t");
 	prt_units_u64(out, vstruct_bytes(sb));
 	prt_str(out, "/");
 	prt_units_u64(out, 512ULL << sb->layout.sb_max_size_bits);
 	prt_newline(out);
 
-	prt_printf(out, "Clean:");
-	prt_tab(out);
-	prt_printf(out, "%llu", BCH_SB_CLEAN(sb));
-	prt_newline(out);
-
-	prt_printf(out, "Devices:");
-	prt_tab(out);
-	prt_printf(out, "%u", nr_devices);
-	prt_newline(out);
+	prt_printf(out, "Clean:\t%llu\n", BCH_SB_CLEAN(sb));
+	prt_printf(out, "Devices:\t%u\n", nr_devices);
 
-	prt_printf(out, "Sections:");
+	prt_printf(out, "Sections:\t");
 	vstruct_for_each(sb, f)
 		fields_have |= 1 << le32_to_cpu(f->type);
-	prt_tab(out);
 	prt_bitflags(out, bch2_sb_fields, fields_have);
 	prt_newline(out);
 
-	prt_printf(out, "Features:");
-	prt_tab(out);
+	prt_printf(out, "Features:\t");
 	prt_bitflags(out, bch2_sb_features, le64_to_cpu(sb->features[0]));
 	prt_newline(out);
 
-	prt_printf(out, "Compat features:");
-	prt_tab(out);
+	prt_printf(out, "Compat features:\t");
 	prt_bitflags(out, bch2_sb_compat, le64_to_cpu(sb->compat[0]));
 	prt_newline(out);
 
@@ -1366,8 +1378,7 @@ void bch2_sb_to_text(struct printbuf *out, struct bch_sb *sb,
 			if (opt->get_sb != BCH2_NO_SB_OPT) {
 				u64 v = bch2_opt_from_sb(sb, id);
 
-				prt_printf(out, "%s:", opt->attr.name);
-				prt_tab(out);
+				prt_printf(out, "%s:\t", opt->attr.name);
 				bch2_opt_to_text(out, NULL, sb, opt, v,
 						 OPT_HUMAN_READABLE|OPT_SHOW_FULL_LIST);
 				prt_newline(out);
diff --git a/fs/bcachefs/super-io.h b/fs/bcachefs/super-io.h
index 95e80e0631..fadd364e28 100644
--- a/fs/bcachefs/super-io.h
+++ b/fs/bcachefs/super-io.h
@@ -51,7 +51,8 @@ void bch2_sb_field_delete(struct bch_sb_handle *, enum bch_sb_field_type);
 extern const char * const bch2_sb_fields[];
 
 struct bch_sb_field_ops {
-	int	(*validate)(struct bch_sb *, struct bch_sb_field *, struct printbuf *);
+	int	(*validate)(struct bch_sb *, struct bch_sb_field *,
+			    enum bch_validate_flags, struct printbuf *);
 	void	(*to_text)(struct printbuf *, struct bch_sb *, struct bch_sb_field *);
 };
 
diff --git a/fs/bcachefs/super.c b/fs/bcachefs/super.c
index 6b23e11825..da735608d4 100644
--- a/fs/bcachefs/super.c
+++ b/fs/bcachefs/super.c
@@ -15,6 +15,7 @@
 #include "btree_gc.h"
 #include "btree_journal_iter.h"
 #include "btree_key_cache.h"
+#include "btree_node_scan.h"
 #include "btree_update_interior.h"
 #include "btree_io.h"
 #include "btree_write_buffer.h"
@@ -56,6 +57,7 @@
 #include "super.h"
 #include "super-io.h"
 #include "sysfs.h"
+#include "thread_with_file.h"
 #include "trace.h"
 
 #include <linux/backing-dev.h>
@@ -86,26 +88,38 @@ const char * const bch2_fs_flag_strs[] = {
 	NULL
 };
 
-void __bch2_print(struct bch_fs *c, const char *fmt, ...)
+__printf(2, 0)
+static void bch2_print_maybe_redirect(struct stdio_redirect *stdio, const char *fmt, va_list args)
 {
-	struct stdio_redirect *stdio = bch2_fs_stdio_redirect(c);
+#ifdef __KERNEL__
+	if (unlikely(stdio)) {
+		if (fmt[0] == KERN_SOH[0])
+			fmt += 2;
+
+		bch2_stdio_redirect_vprintf(stdio, true, fmt, args);
+		return;
+	}
+#endif
+	vprintk(fmt, args);
+}
+
+void bch2_print_opts(struct bch_opts *opts, const char *fmt, ...)
+{
+	struct stdio_redirect *stdio = (void *)(unsigned long)opts->stdio;
 
 	va_list args;
 	va_start(args, fmt);
-	if (likely(!stdio)) {
-		vprintk(fmt, args);
-	} else {
-		unsigned long flags;
-
-		if (fmt[0] == KERN_SOH[0])
-			fmt += 2;
+	bch2_print_maybe_redirect(stdio, fmt, args);
+	va_end(args);
+}
 
-		spin_lock_irqsave(&stdio->output_lock, flags);
-		prt_vprintf(&stdio->output_buf, fmt, args);
-		spin_unlock_irqrestore(&stdio->output_lock, flags);
+void __bch2_print(struct bch_fs *c, const char *fmt, ...)
+{
+	struct stdio_redirect *stdio = bch2_fs_stdio_redirect(c);
 
-		wake_up(&stdio->output_wait);
-	}
+	va_list args;
+	va_start(args, fmt);
+	bch2_print_maybe_redirect(stdio, fmt, args);
 	va_end(args);
 }
 
@@ -250,7 +264,6 @@ static void __bch2_fs_read_only(struct bch_fs *c)
 	bch2_open_buckets_stop(c, NULL, true);
 	bch2_rebalance_stop(c);
 	bch2_copygc_stop(c);
-	bch2_gc_thread_stop(c);
 	bch2_fs_ec_flush(c);
 
 	bch_verbose(c, "flushing journal and stopping allocators, journal seq %llu",
@@ -271,11 +284,16 @@ static void __bch2_fs_read_only(struct bch_fs *c)
 	bch_verbose(c, "flushing journal and stopping allocators complete, journal seq %llu",
 		    journal_cur_seq(&c->journal));
 
-	if (test_bit(JOURNAL_REPLAY_DONE, &c->journal.flags) &&
+	if (test_bit(JOURNAL_replay_done, &c->journal.flags) &&
 	    !test_bit(BCH_FS_emergency_ro, &c->flags))
 		set_bit(BCH_FS_clean_shutdown, &c->flags);
+
 	bch2_fs_journal_stop(&c->journal);
 
+	bch_info(c, "%sshutdown complete, journal seq %llu",
+		 test_bit(BCH_FS_clean_shutdown, &c->flags) ? "" : "un",
+		 c->journal.seq_ondisk);
+
 	/*
 	 * After stopping journal:
 	 */
@@ -352,7 +370,7 @@ void bch2_fs_read_only(struct bch_fs *c)
 	    !test_bit(BCH_FS_emergency_ro, &c->flags) &&
 	    test_bit(BCH_FS_started, &c->flags) &&
 	    test_bit(BCH_FS_clean_shutdown, &c->flags) &&
-	    !c->opts.norecovery) {
+	    c->recovery_pass_done >= BCH_RECOVERY_PASS_journal_replay) {
 		BUG_ON(c->journal.last_empty_seq != journal_cur_seq(&c->journal));
 		BUG_ON(atomic_read(&c->btree_cache.dirty));
 		BUG_ON(atomic_long_read(&c->btree_key_cache.nr_dirty));
@@ -448,7 +466,8 @@ static int __bch2_fs_read_write(struct bch_fs *c, bool early)
 	 * overwriting whatever was there previously, and there must always be
 	 * at least one non-flush write in the journal or recovery will fail:
 	 */
-	set_bit(JOURNAL_NEED_FLUSH_WRITE, &c->journal.flags);
+	set_bit(JOURNAL_need_flush_write, &c->journal.flags);
+	set_bit(JOURNAL_running, &c->journal.flags);
 
 	for_each_rw_member(c, ca)
 		bch2_dev_allocator_add(c, ca);
@@ -466,12 +485,6 @@ static int __bch2_fs_read_write(struct bch_fs *c, bool early)
 	}
 #endif
 
-	ret = bch2_gc_thread_start(c);
-	if (ret) {
-		bch_err(c, "error starting gc thread");
-		return ret;
-	}
-
 	ret = bch2_journal_reclaim_start(&c->journal);
 	if (ret)
 		goto err;
@@ -497,7 +510,8 @@ err:
 
 int bch2_fs_read_write(struct bch_fs *c)
 {
-	if (c->opts.norecovery)
+	if (c->opts.recovery_pass_last &&
+	    c->opts.recovery_pass_last < BCH_RECOVERY_PASS_journal_replay)
 		return -BCH_ERR_erofs_norecovery;
 
 	if (c->opts.nochanges)
@@ -517,11 +531,10 @@ int bch2_fs_read_write_early(struct bch_fs *c)
 
 static void __bch2_fs_free(struct bch_fs *c)
 {
-	unsigned i;
-
-	for (i = 0; i < BCH_TIME_STAT_NR; i++)
+	for (unsigned i = 0; i < BCH_TIME_STAT_NR; i++)
 		bch2_time_stats_exit(&c->times[i]);
 
+	bch2_find_btree_nodes_exit(&c->found_btree_nodes);
 	bch2_free_pending_node_rewrites(c);
 	bch2_fs_sb_errors_exit(c);
 	bch2_fs_counters_exit(c);
@@ -537,19 +550,24 @@ static void __bch2_fs_free(struct bch_fs *c)
 	bch2_fs_io_read_exit(c);
 	bch2_fs_buckets_waiting_for_journal_exit(c);
 	bch2_fs_btree_interior_update_exit(c);
-	bch2_fs_btree_iter_exit(c);
 	bch2_fs_btree_key_cache_exit(&c->btree_key_cache);
 	bch2_fs_btree_cache_exit(c);
+	bch2_fs_btree_iter_exit(c);
 	bch2_fs_replicas_exit(c);
 	bch2_fs_journal_exit(&c->journal);
 	bch2_io_clock_exit(&c->io_clock[WRITE]);
 	bch2_io_clock_exit(&c->io_clock[READ]);
 	bch2_fs_compress_exit(c);
 	bch2_journal_keys_put_initial(c);
+	bch2_find_btree_nodes_exit(&c->found_btree_nodes);
 	BUG_ON(atomic_read(&c->journal_keys.ref));
 	bch2_fs_btree_write_buffer_exit(c);
 	percpu_free_rwsem(&c->mark_lock);
-	free_percpu(c->online_reserved);
+	if (c->online_reserved) {
+		u64 v = percpu_u64_get(c->online_reserved);
+		WARN(v, "online_reserved not 0 at shutdown: %lli", v);
+		free_percpu(c->online_reserved);
+	}
 
 	darray_exit(&c->btree_roots_extra);
 	free_percpu(c->pcpu);
@@ -566,8 +584,10 @@ static void __bch2_fs_free(struct bch_fs *c)
 
 	if (c->write_ref_wq)
 		destroy_workqueue(c->write_ref_wq);
-	if (c->io_complete_wq)
-		destroy_workqueue(c->io_complete_wq);
+	if (c->btree_write_submit_wq)
+		destroy_workqueue(c->btree_write_submit_wq);
+	if (c->btree_read_complete_wq)
+		destroy_workqueue(c->btree_read_complete_wq);
 	if (c->copygc_wq)
 		destroy_workqueue(c->copygc_wq);
 	if (c->btree_io_complete_wq)
@@ -576,7 +596,7 @@ static void __bch2_fs_free(struct bch_fs *c)
 		destroy_workqueue(c->btree_update_wq);
 
 	bch2_free_super(&c->disk_sb);
-	kvpfree(c, sizeof(*c));
+	kvfree(c);
 	module_put(THIS_MODULE);
 }
 
@@ -593,8 +613,6 @@ void __bch2_fs_stop(struct bch_fs *c)
 
 	set_bit(BCH_FS_stopping, &c->flags);
 
-	cancel_work_sync(&c->journal_seq_blacklist_gc_work);
-
 	down_write(&c->state_lock);
 	bch2_fs_read_only(c);
 	up_write(&c->state_lock);
@@ -642,6 +660,7 @@ void bch2_fs_free(struct bch_fs *c)
 		struct bch_dev *ca = rcu_dereference_protected(c->devs[i], true);
 
 		if (ca) {
+			EBUG_ON(atomic_long_read(&ca->ref) != 1);
 			bch2_free_super(&ca->disk_sb);
 			bch2_dev_free(ca);
 		}
@@ -696,7 +715,7 @@ static int bch2_fs_online(struct bch_fs *c)
 		ret = bch2_dev_sysfs_online(c, ca);
 		if (ret) {
 			bch_err(c, "error creating sysfs objects");
-			percpu_ref_put(&ca->ref);
+			bch2_dev_put(ca);
 			goto err;
 		}
 	}
@@ -715,7 +734,7 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts)
 	unsigned i, iter_size;
 	int ret = 0;
 
-	c = kvpmalloc(sizeof(struct bch_fs), GFP_KERNEL|__GFP_ZERO);
+	c = kvmalloc(sizeof(struct bch_fs), GFP_KERNEL|__GFP_ZERO);
 	if (!c) {
 		c = ERR_PTR(-BCH_ERR_ENOMEM_fs_alloc);
 		goto out;
@@ -755,6 +774,7 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts)
 	for (i = 0; i < BCH_TIME_STAT_NR; i++)
 		bch2_time_stats_init(&c->times[i]);
 
+	bch2_fs_gc_init(c);
 	bch2_fs_copygc_init(c);
 	bch2_fs_btree_key_cache_init_early(&c->btree_key_cache);
 	bch2_fs_btree_iter_init_early(c);
@@ -777,16 +797,11 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts)
 
 	spin_lock_init(&c->btree_write_error_lock);
 
-	INIT_WORK(&c->journal_seq_blacklist_gc_work,
-		  bch2_blacklist_entries_gc);
-
 	INIT_LIST_HEAD(&c->journal_iters);
 
 	INIT_LIST_HEAD(&c->fsck_error_msgs);
 	mutex_init(&c->fsck_error_msgs_lock);
 
-	seqcount_init(&c->gc_pos_lock);
-
 	seqcount_init(&c->usage_lock);
 
 	sema_init(&c->io_in_flight, 128);
@@ -818,13 +833,13 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts)
 		goto err;
 
 	pr_uuid(&name, c->sb.user_uuid.b);
-	strscpy(c->name, name.buf, sizeof(c->name));
-	printbuf_exit(&name);
-
 	ret = name.allocation_failure ? -BCH_ERR_ENOMEM_fs_name_alloc : 0;
 	if (ret)
 		goto err;
 
+	strscpy(c->name, name.buf, sizeof(c->name));
+	printbuf_exit(&name);
+
 	/* Compat: */
 	if (le16_to_cpu(sb->version) <= bcachefs_metadata_version_inode_v2 &&
 	    !BCH_SB_JOURNAL_FLUSH_DELAY(sb))
@@ -862,13 +877,15 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts)
 	c->inode_shard_bits = ilog2(roundup_pow_of_two(num_possible_cpus()));
 
 	if (!(c->btree_update_wq = alloc_workqueue("bcachefs",
-				WQ_FREEZABLE|WQ_UNBOUND|WQ_MEM_RECLAIM, 512)) ||
+				WQ_HIGHPRI|WQ_FREEZABLE|WQ_MEM_RECLAIM|WQ_UNBOUND, 512)) ||
 	    !(c->btree_io_complete_wq = alloc_workqueue("bcachefs_btree_io",
-				WQ_FREEZABLE|WQ_MEM_RECLAIM, 1)) ||
+				WQ_HIGHPRI|WQ_FREEZABLE|WQ_MEM_RECLAIM, 1)) ||
 	    !(c->copygc_wq = alloc_workqueue("bcachefs_copygc",
-				WQ_FREEZABLE|WQ_MEM_RECLAIM|WQ_CPU_INTENSIVE, 1)) ||
-	    !(c->io_complete_wq = alloc_workqueue("bcachefs_io",
-				WQ_FREEZABLE|WQ_HIGHPRI|WQ_MEM_RECLAIM, 512)) ||
+				WQ_HIGHPRI|WQ_FREEZABLE|WQ_MEM_RECLAIM|WQ_CPU_INTENSIVE, 1)) ||
+	    !(c->btree_read_complete_wq = alloc_workqueue("bcachefs_btree_read_complete",
+				WQ_HIGHPRI|WQ_FREEZABLE|WQ_MEM_RECLAIM, 512)) ||
+	    !(c->btree_write_submit_wq = alloc_workqueue("bcachefs_btree_write_sumit",
+				WQ_HIGHPRI|WQ_FREEZABLE|WQ_MEM_RECLAIM, 1)) ||
 	    !(c->write_ref_wq = alloc_workqueue("bcachefs_write_ref",
 				WQ_FREEZABLE, 0)) ||
 #ifndef BCH_WRITE_REF_DEBUG
@@ -882,8 +899,8 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts)
 			BIOSET_NEED_BVECS) ||
 	    !(c->pcpu = alloc_percpu(struct bch_fs_pcpu)) ||
 	    !(c->online_reserved = alloc_percpu(u64)) ||
-	    mempool_init_kvpmalloc_pool(&c->btree_bounce_pool, 1,
-					c->opts.btree_node_size) ||
+	    mempool_init_kvmalloc_pool(&c->btree_bounce_pool, 1,
+				       c->opts.btree_node_size) ||
 	    mempool_init_kmalloc_pool(&c->large_bkey_pool, 1, 2048) ||
 	    !(c->unused_inode_hints = kcalloc(1U << c->inode_shard_bits,
 					      sizeof(u64), GFP_KERNEL))) {
@@ -897,9 +914,9 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts)
 	    bch2_io_clock_init(&c->io_clock[WRITE]) ?:
 	    bch2_fs_journal_init(&c->journal) ?:
 	    bch2_fs_replicas_init(c) ?:
+	    bch2_fs_btree_iter_init(c) ?:
 	    bch2_fs_btree_cache_init(c) ?:
 	    bch2_fs_btree_key_cache_init(&c->btree_key_cache) ?:
-	    bch2_fs_btree_iter_init(c) ?:
 	    bch2_fs_btree_interior_update_init(c) ?:
 	    bch2_fs_buckets_waiting_for_journal_init(c) ?:
 	    bch2_fs_btree_write_buffer_init(c) ?:
@@ -916,12 +933,13 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts)
 	if (ret)
 		goto err;
 
-	for (i = 0; i < c->sb.nr_devices; i++)
-		if (bch2_dev_exists(c->disk_sb.sb, i) &&
-		    bch2_dev_alloc(c, i)) {
-			ret = -EEXIST;
+	for (i = 0; i < c->sb.nr_devices; i++) {
+		if (!bch2_member_exists(c->disk_sb.sb, i))
+			continue;
+		ret = bch2_dev_alloc(c, i);
+		if (ret)
 			goto err;
-		}
+	}
 
 	bch2_journal_entry_res_resize(&c->journal,
 			&c->btree_root_journal_res,
@@ -1002,8 +1020,16 @@ int bch2_fs_start(struct bch_fs *c)
 	for_each_online_member(c, ca)
 		bch2_members_v2_get_mut(c->disk_sb.sb, ca->dev_idx)->last_mount = cpu_to_le64(now);
 
+	struct bch_sb_field_ext *ext =
+		bch2_sb_field_get_minsize(&c->disk_sb, ext, sizeof(*ext) / sizeof(u64));
 	mutex_unlock(&c->sb_lock);
 
+	if (!ext) {
+		bch_err(c, "insufficient space in superblock for sb_field_ext");
+		ret = -BCH_ERR_ENOSPC_sb;
+		goto err;
+	}
+
 	for_each_rw_member(c, ca)
 		bch2_dev_allocator_add(c, ca);
 	bch2_recalc_capacity(c);
@@ -1061,7 +1087,8 @@ static int bch2_dev_may_add(struct bch_sb *sb, struct bch_fs *c)
 }
 
 static int bch2_dev_in_fs(struct bch_sb_handle *fs,
-			  struct bch_sb_handle *sb)
+			  struct bch_sb_handle *sb,
+			  struct bch_opts *opts)
 {
 	if (fs == sb)
 		return 0;
@@ -1069,7 +1096,7 @@ static int bch2_dev_in_fs(struct bch_sb_handle *fs,
 	if (!uuid_equal(&fs->sb->uuid, &sb->sb->uuid))
 		return -BCH_ERR_device_not_a_member_of_filesystem;
 
-	if (!bch2_dev_exists(fs->sb, sb->sb->dev_idx))
+	if (!bch2_member_exists(fs->sb, sb->sb->dev_idx))
 		return -BCH_ERR_device_has_been_removed;
 
 	if (fs->sb->block_size != sb->sb->block_size)
@@ -1102,11 +1129,14 @@ static int bch2_dev_in_fs(struct bch_sb_handle *fs,
 		bch2_prt_datetime(&buf, le64_to_cpu(sb->sb->write_time));;
 		prt_newline(&buf);
 
-		prt_printf(&buf, "Not using older sb");
+		if (!opts->no_splitbrain_check)
+			prt_printf(&buf, "Not using older sb");
 
 		pr_err("%s", buf.buf);
 		printbuf_exit(&buf);
-		return -BCH_ERR_device_splitbrain;
+
+		if (!opts->no_splitbrain_check)
+			return -BCH_ERR_device_splitbrain;
 	}
 
 	struct bch_member m = bch2_sb_member_get(fs->sb, sb->sb->dev_idx);
@@ -1124,17 +1154,22 @@ static int bch2_dev_in_fs(struct bch_sb_handle *fs,
 		prt_newline(&buf);
 
 		prt_bdevname(&buf, fs->bdev);
-		prt_str(&buf, "believes seq of ");
+		prt_str(&buf, " believes seq of ");
 		prt_bdevname(&buf, sb->bdev);
 		prt_printf(&buf, " to be %llu, but ", seq_from_fs);
 		prt_bdevname(&buf, sb->bdev);
 		prt_printf(&buf, " has %llu\n", seq_from_member);
-		prt_str(&buf, "Not using ");
-		prt_bdevname(&buf, sb->bdev);
+
+		if (!opts->no_splitbrain_check) {
+			prt_str(&buf, "Not using ");
+			prt_bdevname(&buf, sb->bdev);
+		}
 
 		pr_err("%s", buf.buf);
 		printbuf_exit(&buf);
-		return -BCH_ERR_device_splitbrain;
+
+		if (!opts->no_splitbrain_check)
+			return -BCH_ERR_device_splitbrain;
 	}
 
 	return 0;
@@ -1160,19 +1195,22 @@ static void bch2_dev_free(struct bch_dev *ca)
 	if (ca->kobj.state_in_sysfs)
 		kobject_del(&ca->kobj);
 
+	kfree(ca->buckets_nouse);
 	bch2_free_super(&ca->disk_sb);
+	bch2_dev_allocator_background_exit(ca);
 	bch2_dev_journal_exit(ca);
 
 	free_percpu(ca->io_done);
-	bioset_exit(&ca->replica_set);
 	bch2_dev_buckets_free(ca);
 	free_page((unsigned long) ca->sb_read_scratch);
 
-	bch2_time_stats_exit(&ca->io_latency[WRITE]);
-	bch2_time_stats_exit(&ca->io_latency[READ]);
+	bch2_time_stats_quantiles_exit(&ca->io_latency[WRITE]);
+	bch2_time_stats_quantiles_exit(&ca->io_latency[READ]);
 
 	percpu_ref_exit(&ca->io_ref);
+#ifndef CONFIG_BCACHEFS_DEBUG
 	percpu_ref_exit(&ca->ref);
+#endif
 	kobject_put(&ca->kobj);
 }
 
@@ -1199,12 +1237,14 @@ static void __bch2_dev_offline(struct bch_fs *c, struct bch_dev *ca)
 	bch2_dev_journal_exit(ca);
 }
 
+#ifndef CONFIG_BCACHEFS_DEBUG
 static void bch2_dev_ref_complete(struct percpu_ref *ref)
 {
 	struct bch_dev *ca = container_of(ref, struct bch_dev, ref);
 
 	complete(&ca->ref_completion);
 }
+#endif
 
 static void bch2_dev_io_ref_complete(struct percpu_ref *ref)
 {
@@ -1260,8 +1300,8 @@ static struct bch_dev *__bch2_dev_alloc(struct bch_fs *c,
 
 	INIT_WORK(&ca->io_error_work, bch2_io_error_work);
 
-	bch2_time_stats_init(&ca->io_latency[READ]);
-	bch2_time_stats_init(&ca->io_latency[WRITE]);
+	bch2_time_stats_quantiles_init(&ca->io_latency[READ]);
+	bch2_time_stats_quantiles_init(&ca->io_latency[WRITE]);
 
 	ca->mi = bch2_mi_to_cpu(member);
 
@@ -1273,14 +1313,19 @@ static struct bch_dev *__bch2_dev_alloc(struct bch_fs *c,
 	ca->nr_btree_reserve = DIV_ROUND_UP(BTREE_NODE_RESERVE,
 			     ca->mi.bucket_size / btree_sectors(c));
 
-	if (percpu_ref_init(&ca->ref, bch2_dev_ref_complete,
-			    0, GFP_KERNEL) ||
-	    percpu_ref_init(&ca->io_ref, bch2_dev_io_ref_complete,
+#ifndef CONFIG_BCACHEFS_DEBUG
+	if (percpu_ref_init(&ca->ref, bch2_dev_ref_complete, 0, GFP_KERNEL))
+		goto err;
+#else
+	atomic_long_set(&ca->ref, 1);
+#endif
+
+	bch2_dev_allocator_background_init(ca);
+
+	if (percpu_ref_init(&ca->io_ref, bch2_dev_io_ref_complete,
 			    PERCPU_REF_INIT_DEAD, GFP_KERNEL) ||
 	    !(ca->sb_read_scratch = (void *) __get_free_page(GFP_KERNEL)) ||
 	    bch2_dev_buckets_alloc(c, ca) ||
-	    bioset_init(&ca->replica_set, 4,
-			offsetof(struct bch_write_bio, bio), 0) ||
 	    !(ca->io_done	= alloc_percpu(*ca->io_done)))
 		goto err;
 
@@ -1371,10 +1416,9 @@ static int bch2_dev_attach_bdev(struct bch_fs *c, struct bch_sb_handle *sb)
 	    le64_to_cpu(c->disk_sb.sb->seq))
 		bch2_sb_to_fs(c, sb->sb);
 
-	BUG_ON(sb->sb->dev_idx >= c->sb.nr_devices ||
-	       !c->devs[sb->sb->dev_idx]);
+	BUG_ON(!bch2_dev_exists(c, sb->sb->dev_idx));
 
-	ca = bch_dev_locked(c, sb->sb->dev_idx);
+	ca = bch2_dev_locked(c, sb->sb->dev_idx);
 
 	ret = __bch2_dev_attach_bdev(ca, sb);
 	if (ret)
@@ -1466,10 +1510,10 @@ static bool bch2_fs_may_start(struct bch_fs *c)
 		mutex_lock(&c->sb_lock);
 
 		for (i = 0; i < c->disk_sb.sb->nr_devices; i++) {
-			if (!bch2_dev_exists(c->disk_sb.sb, i))
+			if (!bch2_member_exists(c->disk_sb.sb, i))
 				continue;
 
-			ca = bch_dev_locked(c, i);
+			ca = bch2_dev_locked(c, i);
 
 			if (!bch2_dev_is_online(ca) &&
 			    (ca->mi.state == BCH_MEMBER_STATE_rw ||
@@ -1490,6 +1534,7 @@ static void __bch2_dev_read_only(struct bch_fs *c, struct bch_dev *ca)
 	 * The allocator thread itself allocates btree nodes, so stop it first:
 	 */
 	bch2_dev_allocator_remove(c, ca);
+	bch2_recalc_capacity(c);
 	bch2_dev_journal_stop(&c->journal, ca);
 }
 
@@ -1501,6 +1546,7 @@ static void __bch2_dev_read_write(struct bch_fs *c, struct bch_dev *ca)
 
 	bch2_dev_allocator_add(c, ca);
 	bch2_recalc_capacity(c);
+	bch2_dev_do_discards(ca);
 }
 
 int __bch2_dev_set_state(struct bch_fs *c, struct bch_dev *ca,
@@ -1559,17 +1605,17 @@ static int bch2_dev_remove_alloc(struct bch_fs *c, struct bch_dev *ca)
 	 * with bch2_do_invalidates() and bch2_do_discards()
 	 */
 	ret =   bch2_btree_delete_range(c, BTREE_ID_lru, start, end,
-					BTREE_TRIGGER_NORUN, NULL) ?:
+					BTREE_TRIGGER_norun, NULL) ?:
 		bch2_btree_delete_range(c, BTREE_ID_need_discard, start, end,
-					BTREE_TRIGGER_NORUN, NULL) ?:
+					BTREE_TRIGGER_norun, NULL) ?:
 		bch2_btree_delete_range(c, BTREE_ID_freespace, start, end,
-					BTREE_TRIGGER_NORUN, NULL) ?:
+					BTREE_TRIGGER_norun, NULL) ?:
 		bch2_btree_delete_range(c, BTREE_ID_backpointers, start, end,
-					BTREE_TRIGGER_NORUN, NULL) ?:
+					BTREE_TRIGGER_norun, NULL) ?:
 		bch2_btree_delete_range(c, BTREE_ID_alloc, start, end,
-					BTREE_TRIGGER_NORUN, NULL) ?:
+					BTREE_TRIGGER_norun, NULL) ?:
 		bch2_btree_delete_range(c, BTREE_ID_bucket_gens, start, end,
-					BTREE_TRIGGER_NORUN, NULL);
+					BTREE_TRIGGER_norun, NULL);
 	bch_err_msg(c, ret, "removing dev alloc info");
 	return ret;
 }
@@ -1586,7 +1632,7 @@ int bch2_dev_remove(struct bch_fs *c, struct bch_dev *ca, int flags)
 	 * We consume a reference to ca->ref, regardless of whether we succeed
 	 * or fail:
 	 */
-	percpu_ref_put(&ca->ref);
+	bch2_dev_put(ca);
 
 	if (!bch2_dev_state_allowed(c, ca, BCH_MEMBER_STATE_failed, flags)) {
 		bch_err(ca, "Cannot remove without losing data");
@@ -1597,27 +1643,27 @@ int bch2_dev_remove(struct bch_fs *c, struct bch_dev *ca, int flags)
 	__bch2_dev_read_only(c, ca);
 
 	ret = bch2_dev_data_drop(c, ca->dev_idx, flags);
-	bch_err_msg(ca, ret, "dropping data");
+	bch_err_msg(ca, ret, "bch2_dev_data_drop()");
 	if (ret)
 		goto err;
 
 	ret = bch2_dev_remove_alloc(c, ca);
-	bch_err_msg(ca, ret, "deleting alloc info");
+	bch_err_msg(ca, ret, "bch2_dev_remove_alloc()");
 	if (ret)
 		goto err;
 
 	ret = bch2_journal_flush_device_pins(&c->journal, ca->dev_idx);
-	bch_err_msg(ca, ret, "flushing journal");
+	bch_err_msg(ca, ret, "bch2_journal_flush_device_pins()");
 	if (ret)
 		goto err;
 
 	ret = bch2_journal_flush(&c->journal);
-	bch_err(ca, "journal error");
+	bch_err_msg(ca, ret, "bch2_journal_flush()");
 	if (ret)
 		goto err;
 
 	ret = bch2_replicas_gc2(c);
-	bch_err_msg(ca, ret, "in replicas_gc2()");
+	bch_err_msg(ca, ret, "bch2_replicas_gc2()");
 	if (ret)
 		goto err;
 
@@ -1638,7 +1684,12 @@ int bch2_dev_remove(struct bch_fs *c, struct bch_dev *ca, int flags)
 	rcu_assign_pointer(c->devs[ca->dev_idx], NULL);
 	mutex_unlock(&c->sb_lock);
 
+#ifndef CONFIG_BCACHEFS_DEBUG
 	percpu_ref_kill(&ca->ref);
+#else
+	ca->dying = true;
+	bch2_dev_put(ca);
+#endif
 	wait_for_completion(&ca->ref_completion);
 
 	bch2_dev_free(ca);
@@ -1721,7 +1772,7 @@ int bch2_dev_add(struct bch_fs *c, const char *path)
 	if (ret)
 		goto err;
 
-	ret = bch2_dev_journal_alloc(ca);
+	ret = bch2_dev_journal_alloc(ca, true);
 	bch_err_msg(c, ret, "allocating journal");
 	if (ret)
 		goto err;
@@ -1737,9 +1788,28 @@ int bch2_dev_add(struct bch_fs *c, const char *path)
 	if (dynamic_fault("bcachefs:add:no_slot"))
 		goto no_slot;
 
-	for (dev_idx = 0; dev_idx < BCH_SB_MEMBERS_MAX; dev_idx++)
-		if (!bch2_dev_exists(c->disk_sb.sb, dev_idx))
-			goto have_slot;
+	if (c->sb.nr_devices < BCH_SB_MEMBERS_MAX) {
+		dev_idx = c->sb.nr_devices;
+		goto have_slot;
+	}
+
+	int best = -1;
+	u64 best_last_mount = 0;
+	for (dev_idx = 0; dev_idx < BCH_SB_MEMBERS_MAX; dev_idx++) {
+		struct bch_member m = bch2_sb_member_get(c->disk_sb.sb, dev_idx);
+		if (bch2_member_alive(&m))
+			continue;
+
+		u64 last_mount = le64_to_cpu(m.last_mount);
+		if (best < 0 || last_mount < best_last_mount) {
+			best = dev_idx;
+			best_last_mount = last_mount;
+		}
+	}
+	if (best >= 0) {
+		dev_idx = best;
+		goto have_slot;
+	}
 no_slot:
 	ret = -BCH_ERR_ENOSPC_sb_members;
 	bch_err_msg(c, ret, "setting up new superblock");
@@ -1781,7 +1851,7 @@ have_slot:
 
 	bch2_dev_usage_journal_reserve(c);
 
-	ret = bch2_trans_mark_dev_sb(c, ca);
+	ret = bch2_trans_mark_dev_sb(c, ca, BTREE_TRIGGER_transactional);
 	bch_err_msg(ca, ret, "marking new superblock");
 	if (ret)
 		goto err_late;
@@ -1835,7 +1905,7 @@ int bch2_dev_online(struct bch_fs *c, const char *path)
 
 	dev_idx = sb.sb->dev_idx;
 
-	ret = bch2_dev_in_fs(&c->disk_sb, &sb);
+	ret = bch2_dev_in_fs(&c->disk_sb, &sb, &c->opts);
 	bch_err_msg(c, ret, "bringing %s online", path);
 	if (ret)
 		goto err;
@@ -1844,9 +1914,9 @@ int bch2_dev_online(struct bch_fs *c, const char *path)
 	if (ret)
 		goto err;
 
-	ca = bch_dev_locked(c, dev_idx);
+	ca = bch2_dev_locked(c, dev_idx);
 
-	ret = bch2_trans_mark_dev_sb(c, ca);
+	ret = bch2_trans_mark_dev_sb(c, ca, BTREE_TRIGGER_transactional);
 	bch_err_msg(c, ret, "bringing %s online: error from bch2_trans_mark_dev_sb", path);
 	if (ret)
 		goto err;
@@ -1862,7 +1932,7 @@ int bch2_dev_online(struct bch_fs *c, const char *path)
 	}
 
 	if (!ca->journal.nr) {
-		ret = bch2_dev_journal_alloc(ca);
+		ret = bch2_dev_journal_alloc(ca, false);
 		bch_err_msg(ca, ret, "allocating journal");
 		if (ret)
 			goto err;
@@ -1919,6 +1989,13 @@ int bch2_dev_resize(struct bch_fs *c, struct bch_dev *ca, u64 nbuckets)
 		goto err;
 	}
 
+	if (nbuckets > BCH_MEMBER_NBUCKETS_MAX) {
+		bch_err(ca, "New device size too big (%llu greater than max %u)",
+			nbuckets, BCH_MEMBER_NBUCKETS_MAX);
+		ret = -BCH_ERR_device_size_too_big;
+		goto err;
+	}
+
 	if (bch2_dev_is_online(ca) &&
 	    get_capacity(ca->disk_sb.bdev->bd_disk) <
 	    ca->mi.bucket_size * nbuckets) {
@@ -1932,7 +2009,7 @@ int bch2_dev_resize(struct bch_fs *c, struct bch_dev *ca, u64 nbuckets)
 	if (ret)
 		goto err;
 
-	ret = bch2_trans_mark_dev_sb(c, ca);
+	ret = bch2_trans_mark_dev_sb(c, ca, BTREE_TRIGGER_transactional);
 	if (ret)
 		goto err;
 
@@ -1964,13 +2041,9 @@ err:
 /* return with ref on ca->ref: */
 struct bch_dev *bch2_dev_lookup(struct bch_fs *c, const char *name)
 {
-	rcu_read_lock();
-	for_each_member_device_rcu(c, ca, NULL)
-		if (!strcmp(name, ca->name)) {
-			rcu_read_unlock();
+	for_each_member_device(c, ca)
+		if (!strcmp(name, ca->name))
 			return ca;
-		}
-	rcu_read_unlock();
 	return ERR_PTR(-BCH_ERR_ENOENT_dev_not_found);
 }
 
@@ -2023,7 +2096,7 @@ struct bch_fs *bch2_fs_open(char * const *devices, unsigned nr_devices,
 			best = sb;
 
 	darray_for_each_reverse(sbs, sb) {
-		ret = bch2_dev_in_fs(best, sb);
+		ret = bch2_dev_in_fs(best, sb, &opts);
 
 		if (ret == -BCH_ERR_device_has_been_removed ||
 		    ret == -BCH_ERR_device_splitbrain) {
diff --git a/fs/bcachefs/super_types.h b/fs/bcachefs/super_types.h
index 0e5a14fc8e..368a63d938 100644
--- a/fs/bcachefs/super_types.h
+++ b/fs/bcachefs/super_types.h
@@ -4,7 +4,7 @@
 
 struct bch_sb_handle {
 	struct bch_sb		*sb;
-	struct bdev_handle	*bdev_handle;
+	struct file		*s_bdev_file;
 	struct block_device	*bdev;
 	char			*sb_name;
 	struct bio		*bio;
@@ -26,17 +26,4 @@ struct bch_devs_list {
 	u8			data[BCH_BKEY_PTRS_MAX];
 };
 
-struct bch_member_cpu {
-	u64			nbuckets;	/* device size */
-	u16			first_bucket;   /* index of first bucket used */
-	u16			bucket_size;	/* sectors */
-	u16			group;
-	u8			state;
-	u8			discard;
-	u8			data_allowed;
-	u8			durability;
-	u8			freespace_initialized;
-	u8			valid;
-};
-
 #endif /* _BCACHEFS_SUPER_TYPES_H */
diff --git a/fs/bcachefs/sysfs.c b/fs/bcachefs/sysfs.c
index cee80c47fe..93ca74d108 100644
--- a/fs/bcachefs/sysfs.c
+++ b/fs/bcachefs/sysfs.c
@@ -17,7 +17,6 @@
 #include "btree_iter.h"
 #include "btree_key_cache.h"
 #include "btree_update.h"
-#include "btree_update_interior.h"
 #include "btree_gc.h"
 #include "buckets.h"
 #include "clock.h"
@@ -26,6 +25,7 @@
 #include "ec.h"
 #include "inode.h"
 #include "journal.h"
+#include "journal_reclaim.h"
 #include "keylist.h"
 #include "move.h"
 #include "movinggc.h"
@@ -139,9 +139,9 @@ do {									\
 write_attribute(trigger_gc);
 write_attribute(trigger_discards);
 write_attribute(trigger_invalidates);
-write_attribute(prune_cache);
-write_attribute(btree_wakeup);
-rw_attribute(btree_gc_periodic);
+write_attribute(trigger_journal_flush);
+write_attribute(trigger_btree_cache_shrink);
+write_attribute(trigger_btree_key_cache_shrink);
 rw_attribute(gc_gens_pos);
 
 read_attribute(uuid);
@@ -166,7 +166,6 @@ read_attribute(btree_write_stats);
 read_attribute(btree_cache_size);
 read_attribute(compression_stats);
 read_attribute(journal_debug);
-read_attribute(btree_updates);
 read_attribute(btree_cache);
 read_attribute(btree_key_cache);
 read_attribute(stripes_heap);
@@ -189,12 +188,8 @@ static void bch2_write_refs_to_text(struct printbuf *out, struct bch_fs *c)
 {
 	bch2_printbuf_tabstop_push(out, 24);
 
-	for (unsigned i = 0; i < ARRAY_SIZE(c->writes); i++) {
-		prt_str(out, bch2_write_refs[i]);
-		prt_tab(out);
-		prt_printf(out, "%li", atomic_long_read(&c->writes[i]));
-		prt_newline(out);
-	}
+	for (unsigned i = 0; i < ARRAY_SIZE(c->writes); i++)
+		prt_printf(out, "%s\t%li\n", bch2_write_refs[i], atomic_long_read(&c->writes[i]));
 }
 #endif
 
@@ -278,7 +273,7 @@ static int bch2_compression_stats_to_text(struct printbuf *out, struct bch_fs *c
 			continue;
 
 		ret = for_each_btree_key(trans, iter, id, POS_MIN,
-					 BTREE_ITER_ALL_SNAPSHOTS, k, ({
+					 BTREE_ITER_all_snapshots, k, ({
 			struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k);
 			struct bch_extent_crc_unpacked crc;
 			const union bch_extent_entry *entry;
@@ -313,22 +308,11 @@ static int bch2_compression_stats_to_text(struct printbuf *out, struct bch_fs *c
 	if (ret)
 		return ret;
 
-	prt_str(out, "type");
 	printbuf_tabstop_push(out, 12);
-	prt_tab(out);
-
-	prt_str(out, "compressed");
 	printbuf_tabstop_push(out, 16);
-	prt_tab_rjust(out);
-
-	prt_str(out, "uncompressed");
 	printbuf_tabstop_push(out, 16);
-	prt_tab_rjust(out);
-
-	prt_str(out, "average extent size");
 	printbuf_tabstop_push(out, 24);
-	prt_tab_rjust(out);
-	prt_newline(out);
+	prt_printf(out, "type\tcompressed\runcompressed\raverage extent size\r\n");
 
 	for (unsigned i = 0; i < ARRAY_SIZE(s); i++) {
 		bch2_prt_compression_type(out, i);
@@ -362,21 +346,6 @@ static void bch2_gc_gens_pos_to_text(struct printbuf *out, struct bch_fs *c)
 	prt_printf(out, "\n");
 }
 
-static void bch2_btree_wakeup_all(struct bch_fs *c)
-{
-	struct btree_trans *trans;
-
-	seqmutex_lock(&c->btree_trans_lock);
-	list_for_each_entry(trans, &c->btree_trans_list, list) {
-		struct btree_bkey_cached_common *b = READ_ONCE(trans->locking);
-
-		if (b)
-			six_lock_wakeup_all(&b->lock);
-
-	}
-	seqmutex_unlock(&c->btree_trans_lock);
-}
-
 SHOW(bch2_fs)
 {
 	struct bch_fs *c = container_of(kobj, struct bch_fs, kobj);
@@ -392,8 +361,6 @@ SHOW(bch2_fs)
 	if (attr == &sysfs_btree_write_stats)
 		bch2_btree_write_stats_to_text(out, c);
 
-	sysfs_printf(btree_gc_periodic, "%u",	(int) c->btree_gc_periodic);
-
 	if (attr == &sysfs_gc_gens_pos)
 		bch2_gc_gens_pos_to_text(out, c);
 
@@ -415,11 +382,8 @@ SHOW(bch2_fs)
 	if (attr == &sysfs_journal_debug)
 		bch2_journal_debug_to_text(out, &c->journal);
 
-	if (attr == &sysfs_btree_updates)
-		bch2_btree_updates_to_text(out, c);
-
 	if (attr == &sysfs_btree_cache)
-		bch2_btree_cache_to_text(out, c);
+		bch2_btree_cache_to_text(out, &c->btree_cache);
 
 	if (attr == &sysfs_btree_key_cache)
 		bch2_btree_key_cache_to_text(out, &c->btree_key_cache);
@@ -462,6 +426,9 @@ SHOW(bch2_fs)
 	if (attr == &sysfs_disk_groups)
 		bch2_disk_groups_to_text(out, c);
 
+	if (attr == &sysfs_alloc_debug)
+		bch2_fs_alloc_debug_to_text(out, c);
+
 	return 0;
 }
 
@@ -469,14 +436,6 @@ STORE(bch2_fs)
 {
 	struct bch_fs *c = container_of(kobj, struct bch_fs, kobj);
 
-	if (attr == &sysfs_btree_gc_periodic) {
-		ssize_t ret = strtoul_safe(buf, c->btree_gc_periodic)
-			?: (ssize_t) size;
-
-		wake_up_process(c->gc_thread);
-		return ret;
-	}
-
 	if (attr == &sysfs_copy_gc_enabled) {
 		ssize_t ret = strtoul_safe(buf, c->copy_gc_enabled)
 			?: (ssize_t) size;
@@ -505,10 +464,10 @@ STORE(bch2_fs)
 
 	/* Debugging: */
 
-	if (!test_bit(BCH_FS_rw, &c->flags))
+	if (!bch2_write_ref_tryget(c, BCH_WRITE_REF_sysfs))
 		return -EROFS;
 
-	if (attr == &sysfs_prune_cache) {
+	if (attr == &sysfs_trigger_btree_cache_shrink) {
 		struct shrink_control sc;
 
 		sc.gfp_mask = GFP_KERNEL;
@@ -516,28 +475,28 @@ STORE(bch2_fs)
 		c->btree_cache.shrink->scan_objects(c->btree_cache.shrink, &sc);
 	}
 
-	if (attr == &sysfs_btree_wakeup)
-		bch2_btree_wakeup_all(c);
-
-	if (attr == &sysfs_trigger_gc) {
-		/*
-		 * Full gc is currently incompatible with btree key cache:
-		 */
-#if 0
-		down_read(&c->state_lock);
-		bch2_gc(c, false, false);
-		up_read(&c->state_lock);
-#else
-		bch2_gc_gens(c);
-#endif
+	if (attr == &sysfs_trigger_btree_key_cache_shrink) {
+		struct shrink_control sc;
+
+		sc.gfp_mask = GFP_KERNEL;
+		sc.nr_to_scan = strtoul_or_return(buf);
+		c->btree_key_cache.shrink->scan_objects(c->btree_cache.shrink, &sc);
 	}
 
+	if (attr == &sysfs_trigger_gc)
+		bch2_gc_gens(c);
+
 	if (attr == &sysfs_trigger_discards)
 		bch2_do_discards(c);
 
 	if (attr == &sysfs_trigger_invalidates)
 		bch2_do_invalidates(c);
 
+	if (attr == &sysfs_trigger_journal_flush) {
+		bch2_journal_flush_all_pins(&c->journal);
+		bch2_journal_meta(&c->journal);
+	}
+
 #ifdef CONFIG_BCACHEFS_TESTS
 	if (attr == &sysfs_perf_test) {
 		char *tmp = kstrdup(buf, GFP_KERNEL), *p = tmp;
@@ -558,6 +517,7 @@ STORE(bch2_fs)
 			size = ret;
 	}
 #endif
+	bch2_write_ref_put(c, BCH_WRITE_REF_sysfs);
 	return size;
 }
 SYSFS_OPS(bch2_fs);
@@ -591,13 +551,11 @@ SHOW(bch2_fs_counters)
 		if (attr == &sysfs_##t) {					\
 			counter             = percpu_u64_get(&c->counters[BCH_COUNTER_##t]);\
 			counter_since_mount = counter - c->counters_on_mount[BCH_COUNTER_##t];\
-			prt_printf(out, "since mount:");				\
-			prt_tab(out);						\
+			prt_printf(out, "since mount:\t");			\
 			prt_human_readable_u64(out, counter_since_mount);	\
 			prt_newline(out);					\
 										\
-			prt_printf(out, "since filesystem creation:");		\
-			prt_tab(out);						\
+			prt_printf(out, "since filesystem creation:\t");	\
 			prt_human_readable_u64(out, counter);			\
 			prt_newline(out);					\
 		}
@@ -639,7 +597,6 @@ SYSFS_OPS(bch2_fs_internal);
 struct attribute *bch2_fs_internal_files[] = {
 	&sysfs_flags,
 	&sysfs_journal_debug,
-	&sysfs_btree_updates,
 	&sysfs_btree_cache,
 	&sysfs_btree_key_cache,
 	&sysfs_new_stripes,
@@ -657,8 +614,9 @@ struct attribute *bch2_fs_internal_files[] = {
 	&sysfs_trigger_gc,
 	&sysfs_trigger_discards,
 	&sysfs_trigger_invalidates,
-	&sysfs_prune_cache,
-	&sysfs_btree_wakeup,
+	&sysfs_trigger_journal_flush,
+	&sysfs_trigger_btree_cache_shrink,
+	&sysfs_trigger_btree_key_cache_shrink,
 
 	&sysfs_gc_gens_pos,
 
@@ -674,6 +632,7 @@ struct attribute *bch2_fs_internal_files[] = {
 	&sysfs_internal_uuid,
 
 	&sysfs_disk_groups,
+	&sysfs_alloc_debug,
 	NULL
 };
 
@@ -789,88 +748,6 @@ struct attribute *bch2_fs_time_stats_files[] = {
 	NULL
 };
 
-static void dev_alloc_debug_to_text(struct printbuf *out, struct bch_dev *ca)
-{
-	struct bch_fs *c = ca->fs;
-	struct bch_dev_usage stats = bch2_dev_usage_read(ca);
-	unsigned i, nr[BCH_DATA_NR];
-
-	memset(nr, 0, sizeof(nr));
-
-	for (i = 0; i < ARRAY_SIZE(c->open_buckets); i++)
-		nr[c->open_buckets[i].data_type]++;
-
-	printbuf_tabstop_push(out, 8);
-	printbuf_tabstop_push(out, 16);
-	printbuf_tabstop_push(out, 16);
-	printbuf_tabstop_push(out, 16);
-	printbuf_tabstop_push(out, 16);
-
-	bch2_dev_usage_to_text(out, &stats);
-
-	prt_newline(out);
-
-	prt_printf(out, "reserves:");
-	prt_newline(out);
-	for (i = 0; i < BCH_WATERMARK_NR; i++) {
-		prt_str(out, bch2_watermarks[i]);
-		prt_tab(out);
-		prt_u64(out, bch2_dev_buckets_reserved(ca, i));
-		prt_tab_rjust(out);
-		prt_newline(out);
-	}
-
-	prt_newline(out);
-
-	printbuf_tabstops_reset(out);
-	printbuf_tabstop_push(out, 24);
-
-	prt_str(out, "freelist_wait");
-	prt_tab(out);
-	prt_str(out, c->freelist_wait.list.first ? "waiting" : "empty");
-	prt_newline(out);
-
-	prt_str(out, "open buckets allocated");
-	prt_tab(out);
-	prt_u64(out, OPEN_BUCKETS_COUNT - c->open_buckets_nr_free);
-	prt_newline(out);
-
-	prt_str(out, "open buckets this dev");
-	prt_tab(out);
-	prt_u64(out, ca->nr_open_buckets);
-	prt_newline(out);
-
-	prt_str(out, "open buckets total");
-	prt_tab(out);
-	prt_u64(out, OPEN_BUCKETS_COUNT);
-	prt_newline(out);
-
-	prt_str(out, "open_buckets_wait");
-	prt_tab(out);
-	prt_str(out, c->open_buckets_wait.list.first ? "waiting" : "empty");
-	prt_newline(out);
-
-	prt_str(out, "open_buckets_btree");
-	prt_tab(out);
-	prt_u64(out, nr[BCH_DATA_btree]);
-	prt_newline(out);
-
-	prt_str(out, "open_buckets_user");
-	prt_tab(out);
-	prt_u64(out, nr[BCH_DATA_user]);
-	prt_newline(out);
-
-	prt_str(out, "buckets_to_invalidate");
-	prt_tab(out);
-	prt_u64(out, should_invalidate_buckets(ca, stats));
-	prt_newline(out);
-
-	prt_str(out, "btree reserve cache");
-	prt_tab(out);
-	prt_u64(out, c->btree_reserve_cache_nr);
-	prt_newline(out);
-}
-
 static const char * const bch2_rw[] = {
 	"read",
 	"write",
@@ -930,17 +807,17 @@ SHOW(bch2_dev)
 	sysfs_print(io_latency_write,		atomic64_read(&ca->cur_latency[WRITE]));
 
 	if (attr == &sysfs_io_latency_stats_read)
-		bch2_time_stats_to_text(out, &ca->io_latency[READ]);
+		bch2_time_stats_to_text(out, &ca->io_latency[READ].stats);
 
 	if (attr == &sysfs_io_latency_stats_write)
-		bch2_time_stats_to_text(out, &ca->io_latency[WRITE]);
+		bch2_time_stats_to_text(out, &ca->io_latency[WRITE].stats);
 
 	sysfs_printf(congested,			"%u%%",
 		     clamp(atomic_read(&ca->congested), 0, CONGESTED_MAX)
 		     * 100 / CONGESTED_MAX);
 
 	if (attr == &sysfs_alloc_debug)
-		dev_alloc_debug_to_text(out, ca);
+		bch2_dev_alloc_debug_to_text(out, ca);
 
 	return 0;
 }
diff --git a/fs/bcachefs/tests.c b/fs/bcachefs/tests.c
index b3fe9fc577..68104b2056 100644
--- a/fs/bcachefs/tests.c
+++ b/fs/bcachefs/tests.c
@@ -40,7 +40,7 @@ static int test_delete(struct bch_fs *c, u64 nr)
 	k.k.p.snapshot = U32_MAX;
 
 	bch2_trans_iter_init(trans, &iter, BTREE_ID_xattrs, k.k.p,
-			     BTREE_ITER_INTENT);
+			     BTREE_ITER_intent);
 
 	ret = commit_do(trans, NULL, NULL, 0,
 		bch2_btree_iter_traverse(&iter) ?:
@@ -81,7 +81,7 @@ static int test_delete_written(struct bch_fs *c, u64 nr)
 	k.k.p.snapshot = U32_MAX;
 
 	bch2_trans_iter_init(trans, &iter, BTREE_ID_xattrs, k.k.p,
-			     BTREE_ITER_INTENT);
+			     BTREE_ITER_intent);
 
 	ret = commit_do(trans, NULL, NULL, 0,
 		bch2_btree_iter_traverse(&iter) ?:
@@ -261,7 +261,7 @@ static int test_iterate_slots(struct bch_fs *c, u64 nr)
 	ret = bch2_trans_run(c,
 		for_each_btree_key_upto(trans, iter, BTREE_ID_xattrs,
 					SPOS(0, 0, U32_MAX), POS(0, U64_MAX),
-					BTREE_ITER_SLOTS, k, ({
+					BTREE_ITER_slots, k, ({
 			if (i >= nr * 2)
 				break;
 
@@ -322,7 +322,7 @@ static int test_iterate_slots_extents(struct bch_fs *c, u64 nr)
 	ret = bch2_trans_run(c,
 		for_each_btree_key_upto(trans, iter, BTREE_ID_extents,
 					SPOS(0, 0, U32_MAX), POS(0, U64_MAX),
-					BTREE_ITER_SLOTS, k, ({
+					BTREE_ITER_slots, k, ({
 			if (i == nr)
 				break;
 			BUG_ON(bkey_deleted(k.k) != !(i % 16));
@@ -452,7 +452,7 @@ static int insert_test_overlapping_extent(struct bch_fs *c, u64 inum, u64 start,
 
 	ret = bch2_trans_do(c, NULL, NULL, 0,
 		bch2_btree_insert_nonextent(trans, BTREE_ID_extents, &k.k_i,
-					    BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE));
+					    BTREE_UPDATE_internal_snapshot_node));
 	bch_err_fn(c, ret);
 	return ret;
 }
@@ -671,8 +671,8 @@ static int __do_delete(struct btree_trans *trans, struct bpos pos)
 	int ret = 0;
 
 	bch2_trans_iter_init(trans, &iter, BTREE_ID_xattrs, pos,
-			     BTREE_ITER_INTENT);
-	k = bch2_btree_iter_peek(&iter);
+			     BTREE_ITER_intent);
+	k = bch2_btree_iter_peek_upto(&iter, POS(0, U64_MAX));
 	ret = bkey_err(k);
 	if (ret)
 		goto err;
@@ -714,7 +714,7 @@ static int seq_insert(struct bch_fs *c, u64 nr)
 	return bch2_trans_run(c,
 		for_each_btree_key_commit(trans, iter, BTREE_ID_xattrs,
 					SPOS(0, 0, U32_MAX),
-					BTREE_ITER_SLOTS|BTREE_ITER_INTENT, k,
+					BTREE_ITER_slots|BTREE_ITER_intent, k,
 					NULL, NULL, 0, ({
 			if (iter.pos.offset >= nr)
 				break;
@@ -737,7 +737,7 @@ static int seq_overwrite(struct bch_fs *c, u64 nr)
 	return bch2_trans_run(c,
 		for_each_btree_key_commit(trans, iter, BTREE_ID_xattrs,
 					SPOS(0, 0, U32_MAX),
-					BTREE_ITER_INTENT, k,
+					BTREE_ITER_intent, k,
 					NULL, NULL, 0, ({
 			struct bkey_i_cookie u;
 
diff --git a/fs/bcachefs/thread_with_file.c b/fs/bcachefs/thread_with_file.c
index 9220d7de10..b1af7ac430 100644
--- a/fs/bcachefs/thread_with_file.c
+++ b/fs/bcachefs/thread_with_file.c
@@ -2,7 +2,6 @@
 #ifndef NO_BCACHEFS_FS
 
 #include "bcachefs.h"
-#include "printbuf.h"
 #include "thread_with_file.h"
 
 #include <linux/anon_inodes.h>
@@ -10,6 +9,7 @@
 #include <linux/kthread.h>
 #include <linux/pagemap.h>
 #include <linux/poll.h>
+#include <linux/sched/sysctl.h>
 
 void bch2_thread_with_file_exit(struct thread_with_file *thr)
 {
@@ -65,68 +65,82 @@ err:
 	return ret;
 }
 
-static inline bool thread_with_stdio_has_output(struct thread_with_stdio *thr)
+/* stdio_redirect */
+
+static bool stdio_redirect_has_input(struct stdio_redirect *stdio)
 {
-	return thr->stdio.output_buf.pos ||
-		thr->output2.nr ||
-		thr->thr.done;
+	return stdio->input.buf.nr || stdio->done;
 }
 
-static ssize_t thread_with_stdio_read(struct file *file, char __user *buf,
-				      size_t len, loff_t *ppos)
+static bool stdio_redirect_has_output(struct stdio_redirect *stdio)
 {
-	struct thread_with_stdio *thr =
-		container_of(file->private_data, struct thread_with_stdio, thr);
-	size_t copied = 0, b;
-	int ret = 0;
+	return stdio->output.buf.nr || stdio->done;
+}
 
-	if ((file->f_flags & O_NONBLOCK) &&
-	    !thread_with_stdio_has_output(thr))
-		return -EAGAIN;
+#define STDIO_REDIRECT_BUFSIZE		4096
 
-	ret = wait_event_interruptible(thr->stdio.output_wait,
-		thread_with_stdio_has_output(thr));
-	if (ret)
-		return ret;
+static bool stdio_redirect_has_input_space(struct stdio_redirect *stdio)
+{
+	return stdio->input.buf.nr < STDIO_REDIRECT_BUFSIZE || stdio->done;
+}
 
-	if (thr->thr.done)
-		return 0;
+static bool stdio_redirect_has_output_space(struct stdio_redirect *stdio)
+{
+	return stdio->output.buf.nr < STDIO_REDIRECT_BUFSIZE || stdio->done;
+}
 
-	while (len) {
-		ret = darray_make_room(&thr->output2, thr->stdio.output_buf.pos);
-		if (ret)
-			break;
+static void stdio_buf_init(struct stdio_buf *buf)
+{
+	spin_lock_init(&buf->lock);
+	init_waitqueue_head(&buf->wait);
+	darray_init(&buf->buf);
+}
 
-		spin_lock_irq(&thr->stdio.output_lock);
-		b = min_t(size_t, darray_room(thr->output2), thr->stdio.output_buf.pos);
+/* thread_with_stdio */
 
-		memcpy(&darray_top(thr->output2), thr->stdio.output_buf.buf, b);
-		memmove(thr->stdio.output_buf.buf,
-			thr->stdio.output_buf.buf + b,
-			thr->stdio.output_buf.pos - b);
+static void thread_with_stdio_done(struct thread_with_stdio *thr)
+{
+	thr->thr.done = true;
+	thr->stdio.done = true;
+	wake_up(&thr->stdio.input.wait);
+	wake_up(&thr->stdio.output.wait);
+}
 
-		thr->output2.nr += b;
-		thr->stdio.output_buf.pos -= b;
-		spin_unlock_irq(&thr->stdio.output_lock);
+static ssize_t thread_with_stdio_read(struct file *file, char __user *ubuf,
+				      size_t len, loff_t *ppos)
+{
+	struct thread_with_stdio *thr =
+		container_of(file->private_data, struct thread_with_stdio, thr);
+	struct stdio_buf *buf = &thr->stdio.output;
+	size_t copied = 0, b;
+	int ret = 0;
 
-		b = min(len, thr->output2.nr);
-		if (!b)
-			break;
+	if (!(file->f_flags & O_NONBLOCK)) {
+		ret = wait_event_interruptible(buf->wait, stdio_redirect_has_output(&thr->stdio));
+		if (ret)
+			return ret;
+	} else if (!stdio_redirect_has_output(&thr->stdio))
+		return -EAGAIN;
 
-		b -= copy_to_user(buf, thr->output2.data, b);
-		if (!b) {
+	while (len && buf->buf.nr) {
+		if (fault_in_writeable(ubuf, len) == len) {
 			ret = -EFAULT;
 			break;
 		}
 
-		copied	+= b;
-		buf	+= b;
-		len	-= b;
-
-		memmove(thr->output2.data,
-			thr->output2.data + b,
-			thr->output2.nr - b);
-		thr->output2.nr -= b;
+		spin_lock_irq(&buf->lock);
+		b = min_t(size_t, len, buf->buf.nr);
+
+		if (b && !copy_to_user_nofault(ubuf, buf->buf.data, b)) {
+			ubuf	+= b;
+			len	-= b;
+			copied	+= b;
+			buf->buf.nr -= b;
+			memmove(buf->buf.data,
+				buf->buf.data + b,
+				buf->buf.nr);
+		}
+		spin_unlock_irq(&buf->lock);
 	}
 
 	return copied ?: ret;
@@ -137,27 +151,20 @@ static int thread_with_stdio_release(struct inode *inode, struct file *file)
 	struct thread_with_stdio *thr =
 		container_of(file->private_data, struct thread_with_stdio, thr);
 
+	thread_with_stdio_done(thr);
 	bch2_thread_with_file_exit(&thr->thr);
-	printbuf_exit(&thr->stdio.input_buf);
-	printbuf_exit(&thr->stdio.output_buf);
-	darray_exit(&thr->output2);
-	thr->exit(thr);
+	darray_exit(&thr->stdio.input.buf);
+	darray_exit(&thr->stdio.output.buf);
+	thr->ops->exit(thr);
 	return 0;
 }
 
-#define WRITE_BUFFER		4096
-
-static inline bool thread_with_stdio_has_input_space(struct thread_with_stdio *thr)
-{
-	return thr->stdio.input_buf.pos < WRITE_BUFFER || thr->thr.done;
-}
-
 static ssize_t thread_with_stdio_write(struct file *file, const char __user *ubuf,
 				       size_t len, loff_t *ppos)
 {
 	struct thread_with_stdio *thr =
 		container_of(file->private_data, struct thread_with_stdio, thr);
-	struct printbuf *buf = &thr->stdio.input_buf;
+	struct stdio_buf *buf = &thr->stdio.input;
 	size_t copied = 0;
 	ssize_t ret = 0;
 
@@ -173,29 +180,30 @@ static ssize_t thread_with_stdio_write(struct file *file, const char __user *ubu
 			break;
 		}
 
-		spin_lock(&thr->stdio.input_lock);
-		if (buf->pos < WRITE_BUFFER)
-			bch2_printbuf_make_room(buf, min(b, WRITE_BUFFER - buf->pos));
-		b = min(len, printbuf_remaining_size(buf));
-
-		if (b && !copy_from_user_nofault(&buf->buf[buf->pos], ubuf, b)) {
-			ubuf += b;
-			len -= b;
-			copied += b;
-			buf->pos += b;
+		spin_lock(&buf->lock);
+		if (buf->buf.nr < STDIO_REDIRECT_BUFSIZE)
+			darray_make_room_gfp(&buf->buf,
+				min(b, STDIO_REDIRECT_BUFSIZE - buf->buf.nr), GFP_NOWAIT);
+		b = min(len, darray_room(buf->buf));
+
+		if (b && !copy_from_user_nofault(&darray_top(buf->buf), ubuf, b)) {
+			buf->buf.nr += b;
+			ubuf	+= b;
+			len	-= b;
+			copied	+= b;
 		}
-		spin_unlock(&thr->stdio.input_lock);
+		spin_unlock(&buf->lock);
 
 		if (b) {
-			wake_up(&thr->stdio.input_wait);
+			wake_up(&buf->wait);
 		} else {
 			if ((file->f_flags & O_NONBLOCK)) {
 				ret = -EAGAIN;
 				break;
 			}
 
-			ret = wait_event_interruptible(thr->stdio.input_wait,
-					thread_with_stdio_has_input_space(thr));
+			ret = wait_event_interruptible(buf->wait,
+					stdio_redirect_has_input_space(&thr->stdio));
 			if (ret)
 				break;
 		}
@@ -209,90 +217,244 @@ static __poll_t thread_with_stdio_poll(struct file *file, struct poll_table_stru
 	struct thread_with_stdio *thr =
 		container_of(file->private_data, struct thread_with_stdio, thr);
 
-	poll_wait(file, &thr->stdio.output_wait, wait);
-	poll_wait(file, &thr->stdio.input_wait, wait);
+	poll_wait(file, &thr->stdio.output.wait, wait);
+	poll_wait(file, &thr->stdio.input.wait, wait);
 
 	__poll_t mask = 0;
 
-	if (thread_with_stdio_has_output(thr))
+	if (stdio_redirect_has_output(&thr->stdio))
 		mask |= EPOLLIN;
-	if (thread_with_stdio_has_input_space(thr))
+	if (stdio_redirect_has_input_space(&thr->stdio))
 		mask |= EPOLLOUT;
 	if (thr->thr.done)
 		mask |= EPOLLHUP|EPOLLERR;
 	return mask;
 }
 
+static __poll_t thread_with_stdout_poll(struct file *file, struct poll_table_struct *wait)
+{
+	struct thread_with_stdio *thr =
+		container_of(file->private_data, struct thread_with_stdio, thr);
+
+	poll_wait(file, &thr->stdio.output.wait, wait);
+
+	__poll_t mask = 0;
+
+	if (stdio_redirect_has_output(&thr->stdio))
+		mask |= EPOLLIN;
+	if (thr->thr.done)
+		mask |= EPOLLHUP|EPOLLERR;
+	return mask;
+}
+
+static int thread_with_stdio_flush(struct file *file, fl_owner_t id)
+{
+	struct thread_with_stdio *thr =
+		container_of(file->private_data, struct thread_with_stdio, thr);
+
+	return thr->thr.ret;
+}
+
+static long thread_with_stdio_ioctl(struct file *file, unsigned int cmd, unsigned long p)
+{
+	struct thread_with_stdio *thr =
+		container_of(file->private_data, struct thread_with_stdio, thr);
+
+	if (thr->ops->unlocked_ioctl)
+		return thr->ops->unlocked_ioctl(thr, cmd, p);
+	return -ENOTTY;
+}
+
 static const struct file_operations thread_with_stdio_fops = {
-	.release	= thread_with_stdio_release,
+	.llseek		= no_llseek,
 	.read		= thread_with_stdio_read,
 	.write		= thread_with_stdio_write,
 	.poll		= thread_with_stdio_poll,
+	.flush		= thread_with_stdio_flush,
+	.release	= thread_with_stdio_release,
+	.unlocked_ioctl	= thread_with_stdio_ioctl,
+};
+
+static const struct file_operations thread_with_stdout_fops = {
 	.llseek		= no_llseek,
+	.read		= thread_with_stdio_read,
+	.poll		= thread_with_stdout_poll,
+	.flush		= thread_with_stdio_flush,
+	.release	= thread_with_stdio_release,
+	.unlocked_ioctl	= thread_with_stdio_ioctl,
 };
 
+static int thread_with_stdio_fn(void *arg)
+{
+	struct thread_with_stdio *thr = arg;
+
+	thr->thr.ret = thr->ops->fn(thr);
+
+	thread_with_stdio_done(thr);
+	return 0;
+}
+
+void bch2_thread_with_stdio_init(struct thread_with_stdio *thr,
+				 const struct thread_with_stdio_ops *ops)
+{
+	stdio_buf_init(&thr->stdio.input);
+	stdio_buf_init(&thr->stdio.output);
+	thr->ops = ops;
+}
+
+int __bch2_run_thread_with_stdio(struct thread_with_stdio *thr)
+{
+	return bch2_run_thread_with_file(&thr->thr, &thread_with_stdio_fops, thread_with_stdio_fn);
+}
+
 int bch2_run_thread_with_stdio(struct thread_with_stdio *thr,
-			       void (*exit)(struct thread_with_stdio *),
-			       int (*fn)(void *))
+			       const struct thread_with_stdio_ops *ops)
 {
-	thr->stdio.input_buf = PRINTBUF;
-	thr->stdio.input_buf.atomic++;
-	spin_lock_init(&thr->stdio.input_lock);
-	init_waitqueue_head(&thr->stdio.input_wait);
+	bch2_thread_with_stdio_init(thr, ops);
 
-	thr->stdio.output_buf = PRINTBUF;
-	thr->stdio.output_buf.atomic++;
-	spin_lock_init(&thr->stdio.output_lock);
-	init_waitqueue_head(&thr->stdio.output_wait);
+	return __bch2_run_thread_with_stdio(thr);
+}
 
-	darray_init(&thr->output2);
-	thr->exit = exit;
+int bch2_run_thread_with_stdout(struct thread_with_stdio *thr,
+				const struct thread_with_stdio_ops *ops)
+{
+	stdio_buf_init(&thr->stdio.input);
+	stdio_buf_init(&thr->stdio.output);
+	thr->ops = ops;
 
-	return bch2_run_thread_with_file(&thr->thr, &thread_with_stdio_fops, fn);
+	return bch2_run_thread_with_file(&thr->thr, &thread_with_stdout_fops, thread_with_stdio_fn);
 }
+EXPORT_SYMBOL_GPL(bch2_run_thread_with_stdout);
 
-int bch2_stdio_redirect_read(struct stdio_redirect *stdio, char *buf, size_t len)
+int bch2_stdio_redirect_read(struct stdio_redirect *stdio, char *ubuf, size_t len)
 {
-	wait_event(stdio->input_wait,
-		   stdio->input_buf.pos || stdio->done);
+	struct stdio_buf *buf = &stdio->input;
+
+	/*
+	 * we're waiting on user input (or for the file descriptor to be
+	 * closed), don't want a hung task warning:
+	 */
+	do {
+		wait_event_timeout(buf->wait, stdio_redirect_has_input(stdio),
+				   sysctl_hung_task_timeout_secs * HZ / 2);
+	} while (!stdio_redirect_has_input(stdio));
 
 	if (stdio->done)
 		return -1;
 
-	spin_lock(&stdio->input_lock);
-	int ret = min(len, stdio->input_buf.pos);
-	stdio->input_buf.pos -= ret;
-	memcpy(buf, stdio->input_buf.buf, ret);
-	memmove(stdio->input_buf.buf,
-		stdio->input_buf.buf + ret,
-		stdio->input_buf.pos);
-	spin_unlock(&stdio->input_lock);
+	spin_lock(&buf->lock);
+	int ret = min(len, buf->buf.nr);
+	buf->buf.nr -= ret;
+	memcpy(ubuf, buf->buf.data, ret);
+	memmove(buf->buf.data,
+		buf->buf.data + ret,
+		buf->buf.nr);
+	spin_unlock(&buf->lock);
 
-	wake_up(&stdio->input_wait);
+	wake_up(&buf->wait);
 	return ret;
 }
 
-int bch2_stdio_redirect_readline(struct stdio_redirect *stdio, char *buf, size_t len)
+int bch2_stdio_redirect_readline(struct stdio_redirect *stdio, char *ubuf, size_t len)
 {
-	wait_event(stdio->input_wait,
-		   stdio->input_buf.pos || stdio->done);
-
-	if (stdio->done)
-		return -1;
+	struct stdio_buf *buf = &stdio->input;
+	size_t copied = 0;
+	ssize_t ret = 0;
+again:
+	do {
+		wait_event_timeout(buf->wait, stdio_redirect_has_input(stdio),
+				   sysctl_hung_task_timeout_secs * HZ / 2);
+	} while (!stdio_redirect_has_input(stdio));
+
+	if (stdio->done) {
+		ret = -1;
+		goto out;
+	}
 
-	spin_lock(&stdio->input_lock);
-	int ret = min(len, stdio->input_buf.pos);
-	char *n = memchr(stdio->input_buf.buf, '\n', ret);
+	spin_lock(&buf->lock);
+	size_t b = min(len, buf->buf.nr);
+	char *n = memchr(buf->buf.data, '\n', b);
 	if (n)
-		ret = min(ret, n + 1 - stdio->input_buf.buf);
-	stdio->input_buf.pos -= ret;
-	memcpy(buf, stdio->input_buf.buf, ret);
-	memmove(stdio->input_buf.buf,
-		stdio->input_buf.buf + ret,
-		stdio->input_buf.pos);
-	spin_unlock(&stdio->input_lock);
-
-	wake_up(&stdio->input_wait);
+		b = min_t(size_t, b, n + 1 - buf->buf.data);
+	buf->buf.nr -= b;
+	memcpy(ubuf, buf->buf.data, b);
+	memmove(buf->buf.data,
+		buf->buf.data + b,
+		buf->buf.nr);
+	ubuf += b;
+	len -= b;
+	copied += b;
+	spin_unlock(&buf->lock);
+
+	wake_up(&buf->wait);
+
+	if (!n && len)
+		goto again;
+out:
+	return copied ?: ret;
+}
+
+__printf(3, 0)
+static ssize_t bch2_darray_vprintf(darray_char *out, gfp_t gfp, const char *fmt, va_list args)
+{
+	ssize_t ret;
+
+	do {
+		va_list args2;
+		size_t len;
+
+		va_copy(args2, args);
+		len = vsnprintf(out->data + out->nr, darray_room(*out), fmt, args2);
+		va_end(args2);
+
+		if (len + 1 <= darray_room(*out)) {
+			out->nr += len;
+			return len;
+		}
+
+		ret = darray_make_room_gfp(out, len + 1, gfp);
+	} while (ret == 0);
+
+	return ret;
+}
+
+ssize_t bch2_stdio_redirect_vprintf(struct stdio_redirect *stdio, bool nonblocking,
+				    const char *fmt, va_list args)
+{
+	struct stdio_buf *buf = &stdio->output;
+	unsigned long flags;
+	ssize_t ret;
+
+again:
+	spin_lock_irqsave(&buf->lock, flags);
+	ret = bch2_darray_vprintf(&buf->buf, GFP_NOWAIT, fmt, args);
+	spin_unlock_irqrestore(&buf->lock, flags);
+
+	if (ret < 0) {
+		if (nonblocking)
+			return -EAGAIN;
+
+		ret = wait_event_interruptible(buf->wait,
+				stdio_redirect_has_output_space(stdio));
+		if (ret)
+			return ret;
+		goto again;
+	}
+
+	wake_up(&buf->wait);
+	return ret;
+}
+
+ssize_t bch2_stdio_redirect_printf(struct stdio_redirect *stdio, bool nonblocking,
+				const char *fmt, ...)
+{
+	va_list args;
+	ssize_t ret;
+
+	va_start(args, fmt);
+	ret = bch2_stdio_redirect_vprintf(stdio, nonblocking, fmt, args);
+	va_end(args);
+
 	return ret;
 }
 
diff --git a/fs/bcachefs/thread_with_file.h b/fs/bcachefs/thread_with_file.h
index 05879c5048..1d63d14d7d 100644
--- a/fs/bcachefs/thread_with_file.h
+++ b/fs/bcachefs/thread_with_file.h
@@ -4,6 +4,38 @@
 
 #include "thread_with_file_types.h"
 
+/*
+ * Thread with file: Run a kthread and connect it to a file descriptor, so that
+ * it can be interacted with via fd read/write methods and closing the file
+ * descriptor stops the kthread.
+ *
+ * We have two different APIs:
+ *
+ * thread_with_file, the low level version.
+ * You get to define the full file_operations, including your release function,
+ * which means that you must call bch2_thread_with_file_exit() from your
+ * .release method
+ *
+ * thread_with_stdio, the higher level version
+ * This implements full piping of input and output, including .poll.
+ *
+ * Notes on behaviour:
+ *  - kthread shutdown behaves like writing or reading from a pipe that has been
+ *    closed
+ *  - Input and output buffers are 4096 bytes, although buffers may in some
+ *    situations slightly exceed that limit so as to avoid chopping off a
+ *    message in the middle in nonblocking mode.
+ *  - Input/output buffers are lazily allocated, with GFP_NOWAIT allocations -
+ *    should be fine but might change in future revisions.
+ *  - Output buffer may grow past 4096 bytes to deal with messages that are
+ *    bigger than 4096 bytes
+ *  - Writing may be done blocking or nonblocking; in nonblocking mode, we only
+ *    drop entire messages.
+ *
+ * To write, use stdio_redirect_printf()
+ * To read, use stdio_redirect_read() or stdio_redirect_readline()
+ */
+
 struct task_struct;
 
 struct thread_with_file {
@@ -17,25 +49,31 @@ int bch2_run_thread_with_file(struct thread_with_file *,
 			      const struct file_operations *,
 			      int (*fn)(void *));
 
+struct thread_with_stdio;
+
+struct thread_with_stdio_ops {
+	void (*exit)(struct thread_with_stdio *);
+	int (*fn)(struct thread_with_stdio *);
+	long (*unlocked_ioctl)(struct thread_with_stdio *, unsigned int, unsigned long);
+};
+
 struct thread_with_stdio {
 	struct thread_with_file	thr;
 	struct stdio_redirect	stdio;
-	DARRAY(char)		output2;
-	void			(*exit)(struct thread_with_stdio *);
+	const struct thread_with_stdio_ops	*ops;
 };
 
-static inline void thread_with_stdio_done(struct thread_with_stdio *thr)
-{
-	thr->thr.done = true;
-	thr->stdio.done = true;
-	wake_up(&thr->stdio.input_wait);
-	wake_up(&thr->stdio.output_wait);
-}
-
+void bch2_thread_with_stdio_init(struct thread_with_stdio *,
+				 const struct thread_with_stdio_ops *);
+int __bch2_run_thread_with_stdio(struct thread_with_stdio *);
 int bch2_run_thread_with_stdio(struct thread_with_stdio *,
-			       void (*exit)(struct thread_with_stdio *),
-			       int (*fn)(void *));
+			       const struct thread_with_stdio_ops *);
+int bch2_run_thread_with_stdout(struct thread_with_stdio *,
+				const struct thread_with_stdio_ops *);
 int bch2_stdio_redirect_read(struct stdio_redirect *, char *, size_t);
 int bch2_stdio_redirect_readline(struct stdio_redirect *, char *, size_t);
 
+__printf(3, 0) ssize_t bch2_stdio_redirect_vprintf(struct stdio_redirect *, bool, const char *, va_list);
+__printf(3, 4) ssize_t bch2_stdio_redirect_printf(struct stdio_redirect *, bool, const char *, ...);
+
 #endif /* _BCACHEFS_THREAD_WITH_FILE_H */
diff --git a/fs/bcachefs/thread_with_file_types.h b/fs/bcachefs/thread_with_file_types.h
index 90b5e645e9..e0daf4eec3 100644
--- a/fs/bcachefs/thread_with_file_types.h
+++ b/fs/bcachefs/thread_with_file_types.h
@@ -2,14 +2,21 @@
 #ifndef _BCACHEFS_THREAD_WITH_FILE_TYPES_H
 #define _BCACHEFS_THREAD_WITH_FILE_TYPES_H
 
+#include "darray.h"
+
+struct stdio_buf {
+	spinlock_t		lock;
+	wait_queue_head_t	wait;
+	darray_char		buf;
+};
+
 struct stdio_redirect {
-	spinlock_t		output_lock;
-	wait_queue_head_t	output_wait;
-	struct printbuf		output_buf;
+	struct stdio_buf	input;
+	struct stdio_buf	output;
 
 	spinlock_t		input_lock;
 	wait_queue_head_t	input_wait;
-	struct printbuf		input_buf;
+	darray_char		input_buf;
 	bool			done;
 };
 
diff --git a/fs/bcachefs/time_stats.c b/fs/bcachefs/time_stats.c
new file mode 100644
index 0000000000..4508e9dcbe
--- /dev/null
+++ b/fs/bcachefs/time_stats.c
@@ -0,0 +1,165 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <linux/jiffies.h>
+#include <linux/module.h>
+#include <linux/percpu.h>
+#include <linux/preempt.h>
+#include <linux/time.h>
+#include <linux/spinlock.h>
+
+#include "eytzinger.h"
+#include "time_stats.h"
+
+static const struct time_unit time_units[] = {
+	{ "ns",		1		 },
+	{ "us",		NSEC_PER_USEC	 },
+	{ "ms",		NSEC_PER_MSEC	 },
+	{ "s",		NSEC_PER_SEC	 },
+	{ "m",          (u64) NSEC_PER_SEC * 60},
+	{ "h",          (u64) NSEC_PER_SEC * 3600},
+	{ "d",          (u64) NSEC_PER_SEC * 3600 * 24},
+	{ "w",          (u64) NSEC_PER_SEC * 3600 * 24 * 7},
+	{ "y",          (u64) NSEC_PER_SEC * ((3600 * 24 * 7 * 365) + (3600 * (24 / 4) * 7))}, /* 365.25d */
+	{ "eon",        U64_MAX          },
+};
+
+const struct time_unit *bch2_pick_time_units(u64 ns)
+{
+	const struct time_unit *u;
+
+	for (u = time_units;
+	     u + 1 < time_units + ARRAY_SIZE(time_units) &&
+	     ns >= u[1].nsecs << 1;
+	     u++)
+		;
+
+	return u;
+}
+
+static void quantiles_update(struct quantiles *q, u64 v)
+{
+	unsigned i = 0;
+
+	while (i < ARRAY_SIZE(q->entries)) {
+		struct quantile_entry *e = q->entries + i;
+
+		if (unlikely(!e->step)) {
+			e->m = v;
+			e->step = max_t(unsigned, v / 2, 1024);
+		} else if (e->m > v) {
+			e->m = e->m >= e->step
+				? e->m - e->step
+				: 0;
+		} else if (e->m < v) {
+			e->m = e->m + e->step > e->m
+				? e->m + e->step
+				: U32_MAX;
+		}
+
+		if ((e->m > v ? e->m - v : v - e->m) < e->step)
+			e->step = max_t(unsigned, e->step / 2, 1);
+
+		if (v >= e->m)
+			break;
+
+		i = eytzinger0_child(i, v > e->m);
+	}
+}
+
+static inline void time_stats_update_one(struct bch2_time_stats *stats,
+					      u64 start, u64 end)
+{
+	u64 duration, freq;
+	bool initted = stats->last_event != 0;
+
+	if (time_after64(end, start)) {
+		struct quantiles *quantiles = time_stats_to_quantiles(stats);
+
+		duration = end - start;
+		mean_and_variance_update(&stats->duration_stats, duration);
+		mean_and_variance_weighted_update(&stats->duration_stats_weighted,
+				duration, initted, TIME_STATS_MV_WEIGHT);
+		stats->max_duration = max(stats->max_duration, duration);
+		stats->min_duration = min(stats->min_duration, duration);
+		stats->total_duration += duration;
+
+		if (quantiles)
+			quantiles_update(quantiles, duration);
+	}
+
+	if (stats->last_event && time_after64(end, stats->last_event)) {
+		freq = end - stats->last_event;
+		mean_and_variance_update(&stats->freq_stats, freq);
+		mean_and_variance_weighted_update(&stats->freq_stats_weighted,
+				freq, initted, TIME_STATS_MV_WEIGHT);
+		stats->max_freq = max(stats->max_freq, freq);
+		stats->min_freq = min(stats->min_freq, freq);
+	}
+
+	stats->last_event = end;
+}
+
+void __bch2_time_stats_clear_buffer(struct bch2_time_stats *stats,
+				    struct time_stat_buffer *b)
+{
+	for (struct time_stat_buffer_entry *i = b->entries;
+	     i < b->entries + ARRAY_SIZE(b->entries);
+	     i++)
+		time_stats_update_one(stats, i->start, i->end);
+	b->nr = 0;
+}
+
+static noinline void time_stats_clear_buffer(struct bch2_time_stats *stats,
+					     struct time_stat_buffer *b)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&stats->lock, flags);
+	__bch2_time_stats_clear_buffer(stats, b);
+	spin_unlock_irqrestore(&stats->lock, flags);
+}
+
+void __bch2_time_stats_update(struct bch2_time_stats *stats, u64 start, u64 end)
+{
+	unsigned long flags;
+
+	if (!stats->buffer) {
+		spin_lock_irqsave(&stats->lock, flags);
+		time_stats_update_one(stats, start, end);
+
+		if (mean_and_variance_weighted_get_mean(stats->freq_stats_weighted, TIME_STATS_MV_WEIGHT) < 32 &&
+		    stats->duration_stats.n > 1024)
+			stats->buffer =
+				alloc_percpu_gfp(struct time_stat_buffer,
+						 GFP_ATOMIC);
+		spin_unlock_irqrestore(&stats->lock, flags);
+	} else {
+		struct time_stat_buffer *b;
+
+		preempt_disable();
+		b = this_cpu_ptr(stats->buffer);
+
+		BUG_ON(b->nr >= ARRAY_SIZE(b->entries));
+		b->entries[b->nr++] = (struct time_stat_buffer_entry) {
+			.start = start,
+			.end = end
+		};
+
+		if (unlikely(b->nr == ARRAY_SIZE(b->entries)))
+			time_stats_clear_buffer(stats, b);
+		preempt_enable();
+	}
+}
+
+void bch2_time_stats_exit(struct bch2_time_stats *stats)
+{
+	free_percpu(stats->buffer);
+}
+
+void bch2_time_stats_init(struct bch2_time_stats *stats)
+{
+	memset(stats, 0, sizeof(*stats));
+	stats->min_duration = U64_MAX;
+	stats->min_freq = U64_MAX;
+	spin_lock_init(&stats->lock);
+}
diff --git a/fs/bcachefs/time_stats.h b/fs/bcachefs/time_stats.h
new file mode 100644
index 0000000000..5df6140374
--- /dev/null
+++ b/fs/bcachefs/time_stats.h
@@ -0,0 +1,159 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * bch2_time_stats - collect statistics on events that have a duration, with nicely
+ * formatted textual output on demand
+ *
+ * - percpu buffering of event collection: cheap enough to shotgun
+ *   everywhere without worrying about overhead
+ *
+ * tracks:
+ *  - number of events
+ *  - maximum event duration ever seen
+ *  - sum of all event durations
+ *  - average event duration, standard and weighted
+ *  - standard deviation of event durations, standard and weighted
+ * and analagous statistics for the frequency of events
+ *
+ * We provide both mean and weighted mean (exponentially weighted), and standard
+ * deviation and weighted standard deviation, to give an efficient-to-compute
+ * view of current behaviour versus. average behaviour - "did this event source
+ * just become wonky, or is this typical?".
+ *
+ * Particularly useful for tracking down latency issues.
+ */
+#ifndef _BCACHEFS_TIME_STATS_H
+#define _BCACHEFS_TIME_STATS_H
+
+#include <linux/sched/clock.h>
+#include <linux/spinlock_types.h>
+#include <linux/string.h>
+
+#include "mean_and_variance.h"
+
+struct time_unit {
+	const char	*name;
+	u64		nsecs;
+};
+
+/*
+ * given a nanosecond value, pick the preferred time units for printing:
+ */
+const struct time_unit *bch2_pick_time_units(u64 ns);
+
+/*
+ * quantiles - do not use:
+ *
+ * Only enabled if bch2_time_stats->quantiles_enabled has been manually set - don't
+ * use in new code.
+ */
+
+#define NR_QUANTILES	15
+#define QUANTILE_IDX(i)	inorder_to_eytzinger0(i, NR_QUANTILES)
+#define QUANTILE_FIRST	eytzinger0_first(NR_QUANTILES)
+#define QUANTILE_LAST	eytzinger0_last(NR_QUANTILES)
+
+struct quantiles {
+	struct quantile_entry {
+		u64	m;
+		u64	step;
+	}		entries[NR_QUANTILES];
+};
+
+struct time_stat_buffer {
+	unsigned	nr;
+	struct time_stat_buffer_entry {
+		u64	start;
+		u64	end;
+	}		entries[31];
+};
+
+struct bch2_time_stats {
+	spinlock_t	lock;
+	bool		have_quantiles;
+	/* all fields are in nanoseconds */
+	u64             min_duration;
+	u64		max_duration;
+	u64		total_duration;
+	u64             max_freq;
+	u64             min_freq;
+	u64		last_event;
+	u64		last_event_start;
+
+	struct mean_and_variance	  duration_stats;
+	struct mean_and_variance	  freq_stats;
+
+/* default weight for weighted mean and variance calculations */
+#define TIME_STATS_MV_WEIGHT	8
+
+	struct mean_and_variance_weighted duration_stats_weighted;
+	struct mean_and_variance_weighted freq_stats_weighted;
+	struct time_stat_buffer __percpu *buffer;
+};
+
+struct bch2_time_stats_quantiles {
+	struct bch2_time_stats	stats;
+	struct quantiles	quantiles;
+};
+
+static inline struct quantiles *time_stats_to_quantiles(struct bch2_time_stats *stats)
+{
+	return stats->have_quantiles
+		? &container_of(stats, struct bch2_time_stats_quantiles, stats)->quantiles
+		: NULL;
+}
+
+void __bch2_time_stats_clear_buffer(struct bch2_time_stats *, struct time_stat_buffer *);
+void __bch2_time_stats_update(struct bch2_time_stats *stats, u64, u64);
+
+/**
+ * time_stats_update - collect a new event being tracked
+ *
+ * @stats	- bch2_time_stats to update
+ * @start	- start time of event, recorded with local_clock()
+ *
+ * The end duration of the event will be the current time
+ */
+static inline void bch2_time_stats_update(struct bch2_time_stats *stats, u64 start)
+{
+	__bch2_time_stats_update(stats, start, local_clock());
+}
+
+/**
+ * track_event_change - track state change events
+ *
+ * @stats	- bch2_time_stats to update
+ * @v		- new state, true or false
+ *
+ * Use this when tracking time stats for state changes, i.e. resource X becoming
+ * blocked/unblocked.
+ */
+static inline bool track_event_change(struct bch2_time_stats *stats, bool v)
+{
+	if (v != !!stats->last_event_start) {
+		if (!v) {
+			bch2_time_stats_update(stats, stats->last_event_start);
+			stats->last_event_start = 0;
+		} else {
+			stats->last_event_start = local_clock() ?: 1;
+			return true;
+		}
+	}
+
+	return false;
+}
+
+void bch2_time_stats_exit(struct bch2_time_stats *);
+void bch2_time_stats_init(struct bch2_time_stats *);
+
+static inline void bch2_time_stats_quantiles_exit(struct bch2_time_stats_quantiles *statq)
+{
+	bch2_time_stats_exit(&statq->stats);
+}
+static inline void bch2_time_stats_quantiles_init(struct bch2_time_stats_quantiles *statq)
+{
+	bch2_time_stats_init(&statq->stats);
+	statq->stats.have_quantiles = true;
+	memset(&statq->quantiles, 0, sizeof(statq->quantiles));
+}
+
+#endif /* _BCACHEFS_TIME_STATS_H */
diff --git a/fs/bcachefs/trace.h b/fs/bcachefs/trace.h
index 293b90d704..84fcf26e30 100644
--- a/fs/bcachefs/trace.h
+++ b/fs/bcachefs/trace.h
@@ -43,7 +43,7 @@ DECLARE_EVENT_CLASS(fs_str,
 
 	TP_fast_assign(
 		__entry->dev		= c->dev;
-		__assign_str(str, str);
+		__assign_str(str);
 	),
 
 	TP_printk("%d,%d\n%s", MAJOR(__entry->dev), MINOR(__entry->dev), __get_str(str))
@@ -64,7 +64,7 @@ DECLARE_EVENT_CLASS(trans_str,
 		__entry->dev		= trans->c->dev;
 		strscpy(__entry->trans_fn, trans->fn, sizeof(__entry->trans_fn));
 		__entry->caller_ip		= caller_ip;
-		__assign_str(str, str);
+		__assign_str(str);
 	),
 
 	TP_printk("%d,%d %s %pS %s",
@@ -85,7 +85,7 @@ DECLARE_EVENT_CLASS(trans_str_nocaller,
 	TP_fast_assign(
 		__entry->dev		= trans->c->dev;
 		strscpy(__entry->trans_fn, trans->fn, sizeof(__entry->trans_fn));
-		__assign_str(str, str);
+		__assign_str(str);
 	),
 
 	TP_printk("%d,%d %s %s",
@@ -638,99 +638,14 @@ DEFINE_EVENT(bch_fs, gc_gens_end,
 
 /* Allocator */
 
-DECLARE_EVENT_CLASS(bucket_alloc,
-	TP_PROTO(struct bch_dev *ca, const char *alloc_reserve,
-		 u64 bucket,
-		 u64 free,
-		 u64 avail,
-		 u64 copygc_wait_amount,
-		 s64 copygc_waiting_for,
-		 struct bucket_alloc_state *s,
-		 bool nonblocking,
-		 const char *err),
-	TP_ARGS(ca, alloc_reserve, bucket, free, avail,
-		copygc_wait_amount, copygc_waiting_for,
-		s, nonblocking, err),
-
-	TP_STRUCT__entry(
-		__field(u8,			dev			)
-		__array(char,	reserve,	16			)
-		__field(u64,			bucket	)
-		__field(u64,			free			)
-		__field(u64,			avail			)
-		__field(u64,			copygc_wait_amount	)
-		__field(s64,			copygc_waiting_for	)
-		__field(u64,			seen			)
-		__field(u64,			open			)
-		__field(u64,			need_journal_commit	)
-		__field(u64,			nouse			)
-		__field(bool,			nonblocking		)
-		__field(u64,			nocow			)
-		__array(char,			err,	32		)
-	),
-
-	TP_fast_assign(
-		__entry->dev		= ca->dev_idx;
-		strscpy(__entry->reserve, alloc_reserve, sizeof(__entry->reserve));
-		__entry->bucket		= bucket;
-		__entry->free		= free;
-		__entry->avail		= avail;
-		__entry->copygc_wait_amount	= copygc_wait_amount;
-		__entry->copygc_waiting_for	= copygc_waiting_for;
-		__entry->seen		= s->buckets_seen;
-		__entry->open		= s->skipped_open;
-		__entry->need_journal_commit = s->skipped_need_journal_commit;
-		__entry->nouse		= s->skipped_nouse;
-		__entry->nonblocking	= nonblocking;
-		__entry->nocow		= s->skipped_nocow;
-		strscpy(__entry->err, err, sizeof(__entry->err));
-	),
-
-	TP_printk("reserve %s bucket %u:%llu free %llu avail %llu copygc_wait %llu/%lli seen %llu open %llu need_journal_commit %llu nouse %llu nocow %llu nonblocking %u err %s",
-		  __entry->reserve,
-		  __entry->dev,
-		  __entry->bucket,
-		  __entry->free,
-		  __entry->avail,
-		  __entry->copygc_wait_amount,
-		  __entry->copygc_waiting_for,
-		  __entry->seen,
-		  __entry->open,
-		  __entry->need_journal_commit,
-		  __entry->nouse,
-		  __entry->nocow,
-		  __entry->nonblocking,
-		  __entry->err)
+DEFINE_EVENT(fs_str, bucket_alloc,
+	TP_PROTO(struct bch_fs *c, const char *str),
+	TP_ARGS(c, str)
 );
 
-DEFINE_EVENT(bucket_alloc, bucket_alloc,
-	TP_PROTO(struct bch_dev *ca, const char *alloc_reserve,
-		 u64 bucket,
-		 u64 free,
-		 u64 avail,
-		 u64 copygc_wait_amount,
-		 s64 copygc_waiting_for,
-		 struct bucket_alloc_state *s,
-		 bool nonblocking,
-		 const char *err),
-	TP_ARGS(ca, alloc_reserve, bucket, free, avail,
-		copygc_wait_amount, copygc_waiting_for,
-		s, nonblocking, err)
-);
-
-DEFINE_EVENT(bucket_alloc, bucket_alloc_fail,
-	TP_PROTO(struct bch_dev *ca, const char *alloc_reserve,
-		 u64 bucket,
-		 u64 free,
-		 u64 avail,
-		 u64 copygc_wait_amount,
-		 s64 copygc_waiting_for,
-		 struct bucket_alloc_state *s,
-		 bool nonblocking,
-		 const char *err),
-	TP_ARGS(ca, alloc_reserve, bucket, free, avail,
-		copygc_wait_amount, copygc_waiting_for,
-		s, nonblocking, err)
+DEFINE_EVENT(fs_str, bucket_alloc_fail,
+	TP_PROTO(struct bch_fs *c, const char *str),
+	TP_ARGS(c, str)
 );
 
 TRACE_EVENT(discard_buckets,
@@ -1431,6 +1346,25 @@ DEFINE_EVENT(fs_str, data_update,
 	TP_ARGS(c, str)
 );
 
+TRACE_EVENT(error_downcast,
+	TP_PROTO(int bch_err, int std_err, unsigned long ip),
+	TP_ARGS(bch_err, std_err, ip),
+
+	TP_STRUCT__entry(
+		__array(char,		bch_err, 32		)
+		__array(char,		std_err, 32		)
+		__array(char,		ip, 32			)
+	),
+
+	TP_fast_assign(
+		strscpy(__entry->bch_err, bch2_err_str(bch_err), sizeof(__entry->bch_err));
+		strscpy(__entry->std_err, bch2_err_str(std_err), sizeof(__entry->std_err));
+		snprintf(__entry->ip, sizeof(__entry->ip), "%ps", (void *) ip);
+	),
+
+	TP_printk("%s -> %s %s", __entry->bch_err, __entry->std_err, __entry->ip)
+);
+
 #endif /* _TRACE_BCACHEFS_H */
 
 /* This part must be outside protection */
diff --git a/fs/bcachefs/util.c b/fs/bcachefs/util.c
index 3a32faa86b..4ec7e44d6e 100644
--- a/fs/bcachefs/util.c
+++ b/fs/bcachefs/util.c
@@ -252,8 +252,10 @@ void bch2_prt_u64_base2(struct printbuf *out, u64 v)
 	bch2_prt_u64_base2_nbits(out, v, fls64(v) ?: 1);
 }
 
-void bch2_print_string_as_lines(const char *prefix, const char *lines)
+static void __bch2_print_string_as_lines(const char *prefix, const char *lines,
+					 bool nonblocking)
 {
+	bool locked = false;
 	const char *p;
 
 	if (!lines) {
@@ -261,7 +263,13 @@ void bch2_print_string_as_lines(const char *prefix, const char *lines)
 		return;
 	}
 
-	console_lock();
+	if (!nonblocking) {
+		console_lock();
+		locked = true;
+	} else {
+		locked = console_trylock();
+	}
+
 	while (1) {
 		p = strchrnul(lines, '\n');
 		printk("%s%.*s\n", prefix, (int) (p - lines), lines);
@@ -269,7 +277,18 @@ void bch2_print_string_as_lines(const char *prefix, const char *lines)
 			break;
 		lines = p + 1;
 	}
-	console_unlock();
+	if (locked)
+		console_unlock();
+}
+
+void bch2_print_string_as_lines(const char *prefix, const char *lines)
+{
+	return __bch2_print_string_as_lines(prefix, lines, false);
+}
+
+void bch2_print_string_as_lines_nonblocking(const char *prefix, const char *lines)
+{
+	return __bch2_print_string_as_lines(prefix, lines, true);
 }
 
 int bch2_save_backtrace(bch_stacktrace *stack, struct task_struct *task, unsigned skipnr,
@@ -337,167 +356,23 @@ void bch2_prt_datetime(struct printbuf *out, time64_t sec)
 }
 #endif
 
-static const struct time_unit {
-	const char	*name;
-	u64		nsecs;
-} time_units[] = {
-	{ "ns",		1		 },
-	{ "us",		NSEC_PER_USEC	 },
-	{ "ms",		NSEC_PER_MSEC	 },
-	{ "s",		NSEC_PER_SEC	 },
-	{ "m",          (u64) NSEC_PER_SEC * 60},
-	{ "h",          (u64) NSEC_PER_SEC * 3600},
-	{ "eon",        U64_MAX          },
-};
-
-static const struct time_unit *pick_time_units(u64 ns)
-{
-	const struct time_unit *u;
-
-	for (u = time_units;
-	     u + 1 < time_units + ARRAY_SIZE(time_units) &&
-	     ns >= u[1].nsecs << 1;
-	     u++)
-		;
-
-	return u;
-}
-
 void bch2_pr_time_units(struct printbuf *out, u64 ns)
 {
-	const struct time_unit *u = pick_time_units(ns);
+	const struct time_unit *u = bch2_pick_time_units(ns);
 
 	prt_printf(out, "%llu %s", div_u64(ns, u->nsecs), u->name);
 }
 
-/* time stats: */
-
-#ifndef CONFIG_BCACHEFS_NO_LATENCY_ACCT
-static void bch2_quantiles_update(struct bch2_quantiles *q, u64 v)
-{
-	unsigned i = 0;
-
-	while (i < ARRAY_SIZE(q->entries)) {
-		struct bch2_quantile_entry *e = q->entries + i;
-
-		if (unlikely(!e->step)) {
-			e->m = v;
-			e->step = max_t(unsigned, v / 2, 1024);
-		} else if (e->m > v) {
-			e->m = e->m >= e->step
-				? e->m - e->step
-				: 0;
-		} else if (e->m < v) {
-			e->m = e->m + e->step > e->m
-				? e->m + e->step
-				: U32_MAX;
-		}
-
-		if ((e->m > v ? e->m - v : v - e->m) < e->step)
-			e->step = max_t(unsigned, e->step / 2, 1);
-
-		if (v >= e->m)
-			break;
-
-		i = eytzinger0_child(i, v > e->m);
-	}
-}
-
-static inline void bch2_time_stats_update_one(struct bch2_time_stats *stats,
-					      u64 start, u64 end)
-{
-	u64 duration, freq;
-
-	if (time_after64(end, start)) {
-		duration = end - start;
-		mean_and_variance_update(&stats->duration_stats, duration);
-		mean_and_variance_weighted_update(&stats->duration_stats_weighted, duration);
-		stats->max_duration = max(stats->max_duration, duration);
-		stats->min_duration = min(stats->min_duration, duration);
-		stats->total_duration += duration;
-		bch2_quantiles_update(&stats->quantiles, duration);
-	}
-
-	if (stats->last_event && time_after64(end, stats->last_event)) {
-		freq = end - stats->last_event;
-		mean_and_variance_update(&stats->freq_stats, freq);
-		mean_and_variance_weighted_update(&stats->freq_stats_weighted, freq);
-		stats->max_freq = max(stats->max_freq, freq);
-		stats->min_freq = min(stats->min_freq, freq);
-	}
-
-	stats->last_event = end;
-}
-
-static void __bch2_time_stats_clear_buffer(struct bch2_time_stats *stats,
-					   struct bch2_time_stat_buffer *b)
-{
-	for (struct bch2_time_stat_buffer_entry *i = b->entries;
-	     i < b->entries + ARRAY_SIZE(b->entries);
-	     i++)
-		bch2_time_stats_update_one(stats, i->start, i->end);
-	b->nr = 0;
-}
-
-static noinline void bch2_time_stats_clear_buffer(struct bch2_time_stats *stats,
-						  struct bch2_time_stat_buffer *b)
-{
-	unsigned long flags;
-
-	spin_lock_irqsave(&stats->lock, flags);
-	__bch2_time_stats_clear_buffer(stats, b);
-	spin_unlock_irqrestore(&stats->lock, flags);
-}
-
-void __bch2_time_stats_update(struct bch2_time_stats *stats, u64 start, u64 end)
-{
-	unsigned long flags;
-
-	WARN_ONCE(!stats->duration_stats_weighted.weight ||
-		  !stats->freq_stats_weighted.weight,
-		  "uninitialized time_stats");
-
-	if (!stats->buffer) {
-		spin_lock_irqsave(&stats->lock, flags);
-		bch2_time_stats_update_one(stats, start, end);
-
-		if (mean_and_variance_weighted_get_mean(stats->freq_stats_weighted) < 32 &&
-		    stats->duration_stats.n > 1024)
-			stats->buffer =
-				alloc_percpu_gfp(struct bch2_time_stat_buffer,
-						 GFP_ATOMIC);
-		spin_unlock_irqrestore(&stats->lock, flags);
-	} else {
-		struct bch2_time_stat_buffer *b;
-
-		preempt_disable();
-		b = this_cpu_ptr(stats->buffer);
-
-		BUG_ON(b->nr >= ARRAY_SIZE(b->entries));
-		b->entries[b->nr++] = (struct bch2_time_stat_buffer_entry) {
-			.start = start,
-			.end = end
-		};
-
-		if (unlikely(b->nr == ARRAY_SIZE(b->entries)))
-			bch2_time_stats_clear_buffer(stats, b);
-		preempt_enable();
-	}
-}
-
 static void bch2_pr_time_units_aligned(struct printbuf *out, u64 ns)
 {
-	const struct time_unit *u = pick_time_units(ns);
+	const struct time_unit *u = bch2_pick_time_units(ns);
 
-	prt_printf(out, "%llu ", div64_u64(ns, u->nsecs));
-	prt_tab_rjust(out);
-	prt_printf(out, "%s", u->name);
+	prt_printf(out, "%llu \r%s", div64_u64(ns, u->nsecs), u->name);
 }
 
 static inline void pr_name_and_units(struct printbuf *out, const char *name, u64 ns)
 {
-	prt_str(out, name);
-	prt_tab(out);
+	prt_printf(out, "%s\t", name);
 	bch2_pr_time_units_aligned(out, ns);
 	prt_newline(out);
 }
@@ -506,10 +381,9 @@ static inline void pr_name_and_units(struct printbuf *out, const char *name, u64
 
 void bch2_time_stats_to_text(struct printbuf *out, struct bch2_time_stats *stats)
 {
-	const struct time_unit *u;
+	struct quantiles *quantiles = time_stats_to_quantiles(stats);
 	s64 f_mean = 0, d_mean = 0;
-	u64 q, last_q = 0, f_stddev = 0, d_stddev = 0;
-	int i;
+	u64 f_stddev = 0, d_stddev = 0;
 
 	if (stats->buffer) {
 		int cpu;
@@ -531,12 +405,8 @@ void bch2_time_stats_to_text(struct printbuf *out, struct bch2_time_stats *stats
 	}
 
 	printbuf_tabstop_push(out, out->indent + TABSTOP_SIZE);
-	prt_printf(out, "count:");
-	prt_tab(out);
-	prt_printf(out, "%llu ",
-			 stats->duration_stats.n);
+	prt_printf(out, "count:\t%llu\n", stats->duration_stats.n);
 	printbuf_tabstop_pop(out);
-	prt_newline(out);
 
 	printbuf_tabstops_reset(out);
 
@@ -545,13 +415,8 @@ void bch2_time_stats_to_text(struct printbuf *out, struct bch2_time_stats *stats
 	printbuf_tabstop_push(out, 0);
 	printbuf_tabstop_push(out, TABSTOP_SIZE + 2);
 
-	prt_tab(out);
-	prt_printf(out, "since mount");
-	prt_tab_rjust(out);
-	prt_tab(out);
+	prt_printf(out, "\tsince mount\r\trecent\r\n");
 	prt_printf(out, "recent");
-	prt_tab_rjust(out);
-	prt_newline(out);
 
 	printbuf_tabstops_reset(out);
 	printbuf_tabstop_push(out, out->indent + 20);
@@ -559,88 +424,67 @@ void bch2_time_stats_to_text(struct printbuf *out, struct bch2_time_stats *stats
 	printbuf_tabstop_push(out, 2);
 	printbuf_tabstop_push(out, TABSTOP_SIZE);
 
-	prt_printf(out, "duration of events");
-	prt_newline(out);
+	prt_printf(out, "duration of events\n");
 	printbuf_indent_add(out, 2);
 
 	pr_name_and_units(out, "min:", stats->min_duration);
 	pr_name_and_units(out, "max:", stats->max_duration);
 	pr_name_and_units(out, "total:", stats->total_duration);
 
-	prt_printf(out, "mean:");
-	prt_tab(out);
+	prt_printf(out, "mean:\t");
 	bch2_pr_time_units_aligned(out, d_mean);
 	prt_tab(out);
-	bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_mean(stats->duration_stats_weighted));
+	bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_mean(stats->duration_stats_weighted, TIME_STATS_MV_WEIGHT));
 	prt_newline(out);
 
-	prt_printf(out, "stddev:");
-	prt_tab(out);
+	prt_printf(out, "stddev:\t");
 	bch2_pr_time_units_aligned(out, d_stddev);
 	prt_tab(out);
-	bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_stddev(stats->duration_stats_weighted));
+	bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_stddev(stats->duration_stats_weighted, TIME_STATS_MV_WEIGHT));
 
 	printbuf_indent_sub(out, 2);
 	prt_newline(out);
 
-	prt_printf(out, "time between events");
-	prt_newline(out);
+	prt_printf(out, "time between events\n");
 	printbuf_indent_add(out, 2);
 
 	pr_name_and_units(out, "min:", stats->min_freq);
 	pr_name_and_units(out, "max:", stats->max_freq);
 
-	prt_printf(out, "mean:");
-	prt_tab(out);
+	prt_printf(out, "mean:\t");
 	bch2_pr_time_units_aligned(out, f_mean);
 	prt_tab(out);
-	bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_mean(stats->freq_stats_weighted));
+	bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_mean(stats->freq_stats_weighted, TIME_STATS_MV_WEIGHT));
 	prt_newline(out);
 
-	prt_printf(out, "stddev:");
-	prt_tab(out);
+	prt_printf(out, "stddev:\t");
 	bch2_pr_time_units_aligned(out, f_stddev);
 	prt_tab(out);
-	bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_stddev(stats->freq_stats_weighted));
+	bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_stddev(stats->freq_stats_weighted, TIME_STATS_MV_WEIGHT));
 
 	printbuf_indent_sub(out, 2);
 	prt_newline(out);
 
 	printbuf_tabstops_reset(out);
 
-	i = eytzinger0_first(NR_QUANTILES);
-	u = pick_time_units(stats->quantiles.entries[i].m);
-
-	prt_printf(out, "quantiles (%s):\t", u->name);
-	eytzinger0_for_each(i, NR_QUANTILES) {
-		bool is_last = eytzinger0_next(i, NR_QUANTILES) == -1;
-
-		q = max(stats->quantiles.entries[i].m, last_q);
-		prt_printf(out, "%llu ",
-		       div_u64(q, u->nsecs));
-		if (is_last)
-			prt_newline(out);
-		last_q = q;
+	if (quantiles) {
+		int i = eytzinger0_first(NR_QUANTILES);
+		const struct time_unit *u =
+			bch2_pick_time_units(quantiles->entries[i].m);
+		u64 last_q = 0;
+
+		prt_printf(out, "quantiles (%s):\t", u->name);
+		eytzinger0_for_each(i, NR_QUANTILES) {
+			bool is_last = eytzinger0_next(i, NR_QUANTILES) == -1;
+
+			u64 q = max(quantiles->entries[i].m, last_q);
+			prt_printf(out, "%llu ", div_u64(q, u->nsecs));
+			if (is_last)
+				prt_newline(out);
+			last_q = q;
+		}
 	}
 }
-#else
-void bch2_time_stats_to_text(struct printbuf *out, struct bch2_time_stats *stats) {}
-#endif
-
-void bch2_time_stats_exit(struct bch2_time_stats *stats)
-{
-	free_percpu(stats->buffer);
-}
-
-void bch2_time_stats_init(struct bch2_time_stats *stats)
-{
-	memset(stats, 0, sizeof(*stats));
-	stats->duration_stats_weighted.weight = 8;
-	stats->freq_stats_weighted.weight = 8;
-	stats->min_duration = U64_MAX;
-	stats->min_freq = U64_MAX;
-	spin_lock_init(&stats->lock);
-}
 
 /* ratelimit: */
 
@@ -746,40 +590,31 @@ void bch2_pd_controller_debug_to_text(struct printbuf *out, struct bch_pd_contro
 	if (!out->nr_tabstops)
 		printbuf_tabstop_push(out, 20);
 
-	prt_printf(out, "rate:");
-	prt_tab(out);
+	prt_printf(out, "rate:\t");
 	prt_human_readable_s64(out, pd->rate.rate);
 	prt_newline(out);
 
-	prt_printf(out, "target:");
-	prt_tab(out);
+	prt_printf(out, "target:\t");
 	prt_human_readable_u64(out, pd->last_target);
 	prt_newline(out);
 
-	prt_printf(out, "actual:");
-	prt_tab(out);
+	prt_printf(out, "actual:\t");
 	prt_human_readable_u64(out, pd->last_actual);
 	prt_newline(out);
 
-	prt_printf(out, "proportional:");
-	prt_tab(out);
+	prt_printf(out, "proportional:\t");
 	prt_human_readable_s64(out, pd->last_proportional);
 	prt_newline(out);
 
-	prt_printf(out, "derivative:");
-	prt_tab(out);
+	prt_printf(out, "derivative:\t");
 	prt_human_readable_s64(out, pd->last_derivative);
 	prt_newline(out);
 
-	prt_printf(out, "change:");
-	prt_tab(out);
+	prt_printf(out, "change:\t");
 	prt_human_readable_s64(out, pd->last_change);
 	prt_newline(out);
 
-	prt_printf(out, "next io:");
-	prt_tab(out);
-	prt_printf(out, "%llims", div64_s64(pd->rate.next - local_clock(), NSEC_PER_MSEC));
-	prt_newline(out);
+	prt_printf(out, "next io:\t%llims\n", div64_s64(pd->rate.next - local_clock(), NSEC_PER_MSEC));
 }
 
 /* misc: */
@@ -864,171 +699,6 @@ void memcpy_from_bio(void *dst, struct bio *src, struct bvec_iter src_iter)
 	}
 }
 
-static int alignment_ok(const void *base, size_t align)
-{
-	return IS_ENABLED(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS) ||
-		((unsigned long)base & (align - 1)) == 0;
-}
-
-static void u32_swap(void *a, void *b, size_t size)
-{
-	u32 t = *(u32 *)a;
-	*(u32 *)a = *(u32 *)b;
-	*(u32 *)b = t;
-}
-
-static void u64_swap(void *a, void *b, size_t size)
-{
-	u64 t = *(u64 *)a;
-	*(u64 *)a = *(u64 *)b;
-	*(u64 *)b = t;
-}
-
-static void generic_swap(void *a, void *b, size_t size)
-{
-	char t;
-
-	do {
-		t = *(char *)a;
-		*(char *)a++ = *(char *)b;
-		*(char *)b++ = t;
-	} while (--size > 0);
-}
-
-static inline int do_cmp(void *base, size_t n, size_t size,
-			 int (*cmp_func)(const void *, const void *, size_t),
-			 size_t l, size_t r)
-{
-	return cmp_func(base + inorder_to_eytzinger0(l, n) * size,
-			base + inorder_to_eytzinger0(r, n) * size,
-			size);
-}
-
-static inline void do_swap(void *base, size_t n, size_t size,
-			   void (*swap_func)(void *, void *, size_t),
-			   size_t l, size_t r)
-{
-	swap_func(base + inorder_to_eytzinger0(l, n) * size,
-		  base + inorder_to_eytzinger0(r, n) * size,
-		  size);
-}
-
-void eytzinger0_sort(void *base, size_t n, size_t size,
-		     int (*cmp_func)(const void *, const void *, size_t),
-		     void (*swap_func)(void *, void *, size_t))
-{
-	int i, c, r;
-
-	if (!swap_func) {
-		if (size == 4 && alignment_ok(base, 4))
-			swap_func = u32_swap;
-		else if (size == 8 && alignment_ok(base, 8))
-			swap_func = u64_swap;
-		else
-			swap_func = generic_swap;
-	}
-
-	/* heapify */
-	for (i = n / 2 - 1; i >= 0; --i) {
-		for (r = i; r * 2 + 1 < n; r = c) {
-			c = r * 2 + 1;
-
-			if (c + 1 < n &&
-			    do_cmp(base, n, size, cmp_func, c, c + 1) < 0)
-				c++;
-
-			if (do_cmp(base, n, size, cmp_func, r, c) >= 0)
-				break;
-
-			do_swap(base, n, size, swap_func, r, c);
-		}
-	}
-
-	/* sort */
-	for (i = n - 1; i > 0; --i) {
-		do_swap(base, n, size, swap_func, 0, i);
-
-		for (r = 0; r * 2 + 1 < i; r = c) {
-			c = r * 2 + 1;
-
-			if (c + 1 < i &&
-			    do_cmp(base, n, size, cmp_func, c, c + 1) < 0)
-				c++;
-
-			if (do_cmp(base, n, size, cmp_func, r, c) >= 0)
-				break;
-
-			do_swap(base, n, size, swap_func, r, c);
-		}
-	}
-}
-
-void sort_cmp_size(void *base, size_t num, size_t size,
-	  int (*cmp_func)(const void *, const void *, size_t),
-	  void (*swap_func)(void *, void *, size_t size))
-{
-	/* pre-scale counters for performance */
-	int i = (num/2 - 1) * size, n = num * size, c, r;
-
-	if (!swap_func) {
-		if (size == 4 && alignment_ok(base, 4))
-			swap_func = u32_swap;
-		else if (size == 8 && alignment_ok(base, 8))
-			swap_func = u64_swap;
-		else
-			swap_func = generic_swap;
-	}
-
-	/* heapify */
-	for ( ; i >= 0; i -= size) {
-		for (r = i; r * 2 + size < n; r  = c) {
-			c = r * 2 + size;
-			if (c < n - size &&
-			    cmp_func(base + c, base + c + size, size) < 0)
-				c += size;
-			if (cmp_func(base + r, base + c, size) >= 0)
-				break;
-			swap_func(base + r, base + c, size);
-		}
-	}
-
-	/* sort */
-	for (i = n - size; i > 0; i -= size) {
-		swap_func(base, base + i, size);
-		for (r = 0; r * 2 + size < i; r = c) {
-			c = r * 2 + size;
-			if (c < i - size &&
-			    cmp_func(base + c, base + c + size, size) < 0)
-				c += size;
-			if (cmp_func(base + r, base + c, size) >= 0)
-				break;
-			swap_func(base + r, base + c, size);
-		}
-	}
-}
-
-static void mempool_free_vp(void *element, void *pool_data)
-{
-	size_t size = (size_t) pool_data;
-
-	vpfree(element, size);
-}
-
-static void *mempool_alloc_vp(gfp_t gfp_mask, void *pool_data)
-{
-	size_t size = (size_t) pool_data;
-
-	return vpmalloc(size, gfp_mask);
-}
-
-int mempool_init_kvpmalloc_pool(mempool_t *pool, int min_nr, size_t size)
-{
-	return size < PAGE_SIZE
-		? mempool_init_kmalloc_pool(pool, min_nr, size)
-		: mempool_init(pool, min_nr, mempool_alloc_vp,
-			       mempool_free_vp, (void *) size);
-}
-
 #if 0
 void eytzinger1_test(void)
 {
diff --git a/fs/bcachefs/util.h b/fs/bcachefs/util.h
index b414736d59..5b0533ec4c 100644
--- a/fs/bcachefs/util.h
+++ b/fs/bcachefs/util.h
@@ -21,6 +21,7 @@
 #include "mean_and_variance.h"
 
 #include "darray.h"
+#include "time_stats.h"
 
 struct closure;
 
@@ -53,38 +54,6 @@ static inline size_t buf_pages(void *p, size_t len)
 			    PAGE_SIZE);
 }
 
-static inline void vpfree(void *p, size_t size)
-{
-	if (is_vmalloc_addr(p))
-		vfree(p);
-	else
-		free_pages((unsigned long) p, get_order(size));
-}
-
-static inline void *vpmalloc(size_t size, gfp_t gfp_mask)
-{
-	return (void *) __get_free_pages(gfp_mask|__GFP_NOWARN,
-					 get_order(size)) ?:
-		__vmalloc(size, gfp_mask);
-}
-
-static inline void kvpfree(void *p, size_t size)
-{
-	if (size < PAGE_SIZE)
-		kfree(p);
-	else
-		vpfree(p, size);
-}
-
-static inline void *kvpmalloc(size_t size, gfp_t gfp_mask)
-{
-	return size < PAGE_SIZE
-		? kmalloc(size, gfp_mask)
-		: vpmalloc(size, gfp_mask);
-}
-
-int mempool_init_kvpmalloc_pool(mempool_t *, int, size_t);
-
 #define HEAP(type)							\
 struct {								\
 	size_t size, used;						\
@@ -97,13 +66,13 @@ struct {								\
 ({									\
 	(heap)->used = 0;						\
 	(heap)->size = (_size);						\
-	(heap)->data = kvpmalloc((heap)->size * sizeof((heap)->data[0]),\
+	(heap)->data = kvmalloc((heap)->size * sizeof((heap)->data[0]),\
 				 (gfp));				\
 })
 
 #define free_heap(heap)							\
 do {									\
-	kvpfree((heap)->data, (heap)->size * sizeof((heap)->data[0]));	\
+	kvfree((heap)->data);						\
 	(heap)->data = NULL;						\
 } while (0)
 
@@ -346,6 +315,7 @@ void bch2_prt_u64_base2_nbits(struct printbuf *, u64, unsigned);
 void bch2_prt_u64_base2(struct printbuf *, u64);
 
 void bch2_print_string_as_lines(const char *prefix, const char *lines);
+void bch2_print_string_as_lines_nonblocking(const char *prefix, const char *lines);
 
 typedef DARRAY(unsigned long) bch_stacktrace;
 int bch2_save_backtrace(bch_stacktrace *stack, struct task_struct *, unsigned, gfp_t);
@@ -361,84 +331,8 @@ static inline void prt_bdevname(struct printbuf *out, struct block_device *bdev)
 #endif
 }
 
-#define NR_QUANTILES	15
-#define QUANTILE_IDX(i)	inorder_to_eytzinger0(i, NR_QUANTILES)
-#define QUANTILE_FIRST	eytzinger0_first(NR_QUANTILES)
-#define QUANTILE_LAST	eytzinger0_last(NR_QUANTILES)
-
-struct bch2_quantiles {
-	struct bch2_quantile_entry {
-		u64	m;
-		u64	step;
-	}		entries[NR_QUANTILES];
-};
-
-struct bch2_time_stat_buffer {
-	unsigned	nr;
-	struct bch2_time_stat_buffer_entry {
-		u64	start;
-		u64	end;
-	}		entries[32];
-};
-
-struct bch2_time_stats {
-	spinlock_t	lock;
-	/* all fields are in nanoseconds */
-	u64             min_duration;
-	u64		max_duration;
-	u64		total_duration;
-	u64             max_freq;
-	u64             min_freq;
-	u64		last_event;
-	struct bch2_quantiles quantiles;
-
-	struct mean_and_variance	  duration_stats;
-	struct mean_and_variance_weighted duration_stats_weighted;
-	struct mean_and_variance	  freq_stats;
-	struct mean_and_variance_weighted freq_stats_weighted;
-	struct bch2_time_stat_buffer __percpu *buffer;
-};
-
-#ifndef CONFIG_BCACHEFS_NO_LATENCY_ACCT
-void __bch2_time_stats_update(struct bch2_time_stats *stats, u64, u64);
-
-static inline void bch2_time_stats_update(struct bch2_time_stats *stats, u64 start)
-{
-	__bch2_time_stats_update(stats, start, local_clock());
-}
-
-static inline bool track_event_change(struct bch2_time_stats *stats,
-				      u64 *start, bool v)
-{
-	if (v != !!*start) {
-		if (!v) {
-			bch2_time_stats_update(stats, *start);
-			*start = 0;
-		} else {
-			*start = local_clock() ?: 1;
-			return true;
-		}
-	}
-
-	return false;
-}
-#else
-static inline void __bch2_time_stats_update(struct bch2_time_stats *stats, u64 start, u64 end) {}
-static inline void bch2_time_stats_update(struct bch2_time_stats *stats, u64 start) {}
-static inline bool track_event_change(struct bch2_time_stats *stats,
-				      u64 *start, bool v)
-{
-	bool ret = v && !*start;
-	*start = v;
-	return ret;
-}
-#endif
-
 void bch2_time_stats_to_text(struct printbuf *, struct bch2_time_stats *);
 
-void bch2_time_stats_exit(struct bch2_time_stats *);
-void bch2_time_stats_init(struct bch2_time_stats *);
-
 #define ewma_add(ewma, val, weight)					\
 ({									\
 	typeof(ewma) _ewma = (ewma);					\
@@ -552,11 +446,6 @@ static inline unsigned fract_exp_two(unsigned x, unsigned fract_bits)
 void bch2_bio_map(struct bio *bio, void *base, size_t);
 int bch2_bio_alloc_pages(struct bio *, size_t, gfp_t);
 
-static inline sector_t bdev_sectors(struct block_device *bdev)
-{
-	return bdev->bd_inode->i_size >> 9;
-}
-
 #define closure_bio_submit(bio, cl)					\
 do {									\
 	closure_get(cl);						\
@@ -738,10 +627,6 @@ static inline void memset_u64s_tail(void *s, int c, unsigned bytes)
 	memset(s + bytes, c, rem);
 }
 
-void sort_cmp_size(void *base, size_t num, size_t size,
-	  int (*cmp_func)(const void *, const void *, size_t),
-	  void (*swap_func)(void *, void *, size_t));
-
 /* just the memmove, doesn't update @_nr */
 #define __array_insert_item(_array, _nr, _pos)				\
 	memmove(&(_array)[(_pos) + 1],					\
@@ -788,8 +673,15 @@ static inline void __move_gap(void *array, size_t element_size,
 }
 
 /* Move the gap in a gap buffer: */
-#define move_gap(_array, _nr, _size, _old_gap, _new_gap)	\
-	__move_gap(_array, sizeof(_array[0]), _nr, _size, _old_gap, _new_gap)
+#define move_gap(_d, _new_gap)						\
+do {									\
+	BUG_ON(_new_gap > (_d)->nr);					\
+	BUG_ON((_d)->gap > (_d)->nr);					\
+									\
+	__move_gap((_d)->data, sizeof((_d)->data[0]),			\
+		   (_d)->nr, (_d)->size, (_d)->gap, _new_gap);		\
+	(_d)->gap = _new_gap;						\
+} while (0)
 
 #define bubble_sort(_base, _nr, _cmp)					\
 do {									\
@@ -876,4 +768,43 @@ static inline bool qstr_eq(const struct qstr l, const struct qstr r)
 void bch2_darray_str_exit(darray_str *);
 int bch2_split_devs(const char *, darray_str *);
 
+#ifdef __KERNEL__
+
+__must_check
+static inline int copy_to_user_errcode(void __user *to, const void *from, unsigned long n)
+{
+	return copy_to_user(to, from, n) ? -EFAULT : 0;
+}
+
+__must_check
+static inline int copy_from_user_errcode(void *to, const void __user *from, unsigned long n)
+{
+	return copy_from_user(to, from, n) ? -EFAULT : 0;
+}
+
+#endif
+
+static inline void mod_bit(long nr, volatile unsigned long *addr, bool v)
+{
+	if (v)
+		set_bit(nr, addr);
+	else
+		clear_bit(nr, addr);
+}
+
+static inline void __set_bit_le64(size_t bit, __le64 *addr)
+{
+	addr[bit / 64] |= cpu_to_le64(BIT_ULL(bit % 64));
+}
+
+static inline void __clear_bit_le64(size_t bit, __le64 *addr)
+{
+	addr[bit / 64] &= ~cpu_to_le64(BIT_ULL(bit % 64));
+}
+
+static inline bool test_bit_le64(size_t bit, __le64 *addr)
+{
+	return (addr[bit / 64] & cpu_to_le64(BIT_ULL(bit % 64))) != 0;
+}
+
 #endif /* _BCACHEFS_UTIL_H */
diff --git a/fs/bcachefs/xattr.c b/fs/bcachefs/xattr.c
index 9c0d231603..c11bf6dacc 100644
--- a/fs/bcachefs/xattr.c
+++ b/fs/bcachefs/xattr.c
@@ -71,7 +71,7 @@ const struct bch_hash_desc bch2_xattr_hash_desc = {
 };
 
 int bch2_xattr_invalid(struct bch_fs *c, struct bkey_s_c k,
-		       enum bkey_invalid_flags flags,
+		       enum bch_validate_flags flags,
 		       struct printbuf *err)
 {
 	struct bkey_s_c_xattr xattr = bkey_s_c_to_xattr(k);
@@ -118,11 +118,17 @@ void bch2_xattr_to_text(struct printbuf *out, struct bch_fs *c,
 	else
 		prt_printf(out, "(unknown type %u)", xattr.v->x_type);
 
+	unsigned name_len = xattr.v->x_name_len;
+	unsigned val_len  = le16_to_cpu(xattr.v->x_val_len);
+	unsigned max_name_val_bytes = bkey_val_bytes(xattr.k) -
+		offsetof(struct bch_xattr, x_name);
+
+	val_len  = min_t(int, val_len, max_name_val_bytes - name_len);
+	name_len = min(name_len, max_name_val_bytes);
+
 	prt_printf(out, "%.*s:%.*s",
-	       xattr.v->x_name_len,
-	       xattr.v->x_name,
-	       le16_to_cpu(xattr.v->x_val_len),
-	       (char *) xattr_val(xattr.v));
+		   name_len, xattr.v->x_name,
+		   val_len,  (char *) xattr_val(xattr.v));
 
 	if (xattr.v->x_type == KEY_TYPE_XATTR_INDEX_POSIX_ACL_ACCESS ||
 	    xattr.v->x_type == KEY_TYPE_XATTR_INDEX_POSIX_ACL_DEFAULT) {
@@ -138,21 +144,13 @@ static int bch2_xattr_get_trans(struct btree_trans *trans, struct bch_inode_info
 	struct bch_hash_info hash = bch2_hash_info_init(trans->c, &inode->ei_inode);
 	struct xattr_search_key search = X_SEARCH(type, name, strlen(name));
 	struct btree_iter iter;
-	struct bkey_s_c_xattr xattr;
-	struct bkey_s_c k;
-	int ret;
-
-	ret = bch2_hash_lookup(trans, &iter, bch2_xattr_hash_desc, &hash,
-			       inode_inum(inode), &search, 0);
+	struct bkey_s_c k = bch2_hash_lookup(trans, &iter, bch2_xattr_hash_desc, &hash,
+					     inode_inum(inode), &search, 0);
+	int ret = bkey_err(k);
 	if (ret)
-		goto err1;
-
-	k = bch2_btree_iter_peek_slot(&iter);
-	ret = bkey_err(k);
-	if (ret)
-		goto err2;
+		return ret;
 
-	xattr = bkey_s_c_to_xattr(k);
+	struct bkey_s_c_xattr xattr = bkey_s_c_to_xattr(k);
 	ret = le16_to_cpu(xattr.v->x_val_len);
 	if (buffer) {
 		if (ret > size)
@@ -160,10 +158,8 @@ static int bch2_xattr_get_trans(struct btree_trans *trans, struct bch_inode_info
 		else
 			memcpy(buffer, xattr_val(xattr.v), ret);
 	}
-err2:
 	bch2_trans_iter_exit(trans, &iter);
-err1:
-	return ret < 0 && bch2_err_matches(ret, ENOENT) ? -ENODATA : ret;
+	return ret;
 }
 
 int bch2_xattr_set(struct btree_trans *trans, subvol_inum inum,
@@ -177,7 +173,7 @@ int bch2_xattr_set(struct btree_trans *trans, subvol_inum inum,
 	int ret;
 
 	ret   = bch2_subvol_is_ro_trans(trans, inum.subvol) ?:
-		bch2_inode_peek(trans, &inode_iter, inode_u, inum, BTREE_ITER_INTENT);
+		bch2_inode_peek(trans, &inode_iter, inode_u, inum, BTREE_ITER_intent);
 	if (ret)
 		return ret;
 
@@ -212,8 +208,8 @@ int bch2_xattr_set(struct btree_trans *trans, subvol_inum inum,
 
 		ret = bch2_hash_set(trans, bch2_xattr_hash_desc, hash_info,
 			      inum, &xattr->k_i,
-			      (flags & XATTR_CREATE ? BCH_HASH_SET_MUST_CREATE : 0)|
-			      (flags & XATTR_REPLACE ? BCH_HASH_SET_MUST_REPLACE : 0));
+			      (flags & XATTR_CREATE ? STR_HASH_must_create : 0)|
+			      (flags & XATTR_REPLACE ? STR_HASH_must_replace : 0));
 	} else {
 		struct xattr_search_key search =
 			X_SEARCH(type, name, strlen(name));
@@ -359,6 +355,9 @@ static int bch2_xattr_get_handler(const struct xattr_handler *handler,
 	int ret = bch2_trans_do(c, NULL, NULL, 0,
 		bch2_xattr_get_trans(trans, inode, name, buffer, size, handler->flags));
 
+	if (ret < 0 && bch2_err_matches(ret, ENOENT))
+		ret = -ENODATA;
+
 	return bch2_err_class(ret);
 }
 
@@ -544,11 +543,11 @@ static int bch2_xattr_bcachefs_set(const struct xattr_handler *handler,
 		kfree(buf);
 
 		if (ret < 0)
-			return ret;
+			goto err_class_exit;
 
 		ret = bch2_opt_check_may_set(c, opt_id, v);
 		if (ret < 0)
-			return ret;
+			goto err_class_exit;
 
 		s.v = v + 1;
 		s.defined = true;
@@ -595,6 +594,7 @@ err:
 	     (opt_id == Opt_compression && !inode_opt_get(c, &inode->ei_inode, background_compression))))
 		bch2_set_rebalance_needs_scan(c, inode->ei_inode.bi_inum);
 
+err_class_exit:
 	return bch2_err_class(ret);
 }
 
diff --git a/fs/bcachefs/xattr.h b/fs/bcachefs/xattr.h
index 1337f31a5c..1574b9eb4c 100644
--- a/fs/bcachefs/xattr.h
+++ b/fs/bcachefs/xattr.h
@@ -7,7 +7,7 @@
 extern const struct bch_hash_desc bch2_xattr_hash_desc;
 
 int bch2_xattr_invalid(struct bch_fs *, struct bkey_s_c,
-		       enum bkey_invalid_flags, struct printbuf *);
+		       enum bch_validate_flags, struct printbuf *);
 void bch2_xattr_to_text(struct printbuf *, struct bch_fs *, struct bkey_s_c);
 
 #define bch2_bkey_ops_xattr ((struct bkey_ops) {	\
diff --git a/fs/befs/linuxvfs.c b/fs/befs/linuxvfs.c
index 2b4dda0474..d76f406d3b 100644
--- a/fs/befs/linuxvfs.c
+++ b/fs/befs/linuxvfs.c
@@ -435,8 +435,7 @@ befs_init_inodecache(void)
 {
 	befs_inode_cachep = kmem_cache_create_usercopy("befs_inode_cache",
 				sizeof(struct befs_inode_info), 0,
-				(SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD|
-					SLAB_ACCOUNT),
+				SLAB_RECLAIM_ACCOUNT | SLAB_ACCOUNT,
 				offsetof(struct befs_inode_info,
 					i_data.symlink),
 				sizeof_field(struct befs_inode_info,
diff --git a/fs/bfs/inode.c b/fs/bfs/inode.c
index 355957dbce..db81570c96 100644
--- a/fs/bfs/inode.c
+++ b/fs/bfs/inode.c
@@ -259,7 +259,7 @@ static int __init init_inodecache(void)
 	bfs_inode_cachep = kmem_cache_create("bfs_inode_cache",
 					     sizeof(struct bfs_inode_info),
 					     0, (SLAB_RECLAIM_ACCOUNT|
-						SLAB_MEM_SPREAD|SLAB_ACCOUNT),
+						SLAB_ACCOUNT),
 					     init_once);
 	if (bfs_inode_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c
index 5397b552fb..a43897b03c 100644
--- a/fs/binfmt_elf.c
+++ b/fs/binfmt_elf.c
@@ -1262,6 +1262,9 @@ out_free_interp:
 		if (IS_ENABLED(CONFIG_ARCH_HAS_ELF_RANDOMIZE) &&
 		    elf_ex->e_type == ET_DYN && !interpreter) {
 			mm->brk = mm->start_brk = ELF_ET_DYN_BASE;
+		} else {
+			/* Otherwise leave a gap between .bss and brk. */
+			mm->brk = mm->start_brk = mm->brk + PAGE_SIZE;
 		}
 
 		mm->brk = mm->start_brk = arch_randomize_brk(mm);
@@ -1564,7 +1567,6 @@ static void fill_siginfo_note(struct memelfnote *note, user_siginfo_t *csigdata,
 	fill_note(note, "CORE", NT_SIGINFO, sizeof(*csigdata), csigdata);
 }
 
-#define MAX_FILE_NOTE_SIZE (4*1024*1024)
 /*
  * Format of NT_FILE note:
  *
@@ -1592,8 +1594,12 @@ static int fill_files_note(struct memelfnote *note, struct coredump_params *cprm
 
 	names_ofs = (2 + 3 * count) * sizeof(data[0]);
  alloc:
-	if (size >= MAX_FILE_NOTE_SIZE) /* paranoia check */
+	/* paranoia check */
+	if (size >= core_file_note_size_limit) {
+		pr_warn_once("coredump Note size too large: %u (does kernel.core_file_note_size_limit sysctl need adjustment?\n",
+			      size);
 		return -EINVAL;
+	}
 	size = round_up(size, PAGE_SIZE);
 	/*
 	 * "size" can be 0 here legitimately.
@@ -1928,7 +1934,7 @@ static void free_note_info(struct elf_note_info *info)
 		threads = t->next;
 		WARN_ON(t->notes[0].data && t->notes[0].data != &t->prstatus);
 		for (i = 1; i < info->thread_notes; ++i)
-			kfree(t->notes[i].data);
+			kvfree(t->notes[i].data);
 		kfree(t);
 	}
 	kfree(info->psinfo.data);
diff --git a/fs/binfmt_elf_fdpic.c b/fs/binfmt_elf_fdpic.c
index fefc642541..b799701454 100644
--- a/fs/binfmt_elf_fdpic.c
+++ b/fs/binfmt_elf_fdpic.c
@@ -320,7 +320,7 @@ static int load_elf_fdpic_binary(struct linux_binprm *bprm)
 	else
 		executable_stack = EXSTACK_DEFAULT;
 
-	if (stack_size == 0) {
+	if (stack_size == 0 && interp_params.flags & ELF_FDPIC_FLAG_PRESENT) {
 		stack_size = interp_params.stack_size;
 		if (interp_params.flags & ELF_FDPIC_FLAG_EXEC_STACK)
 			executable_stack = EXSTACK_ENABLE_X;
@@ -505,8 +505,9 @@ static int create_elf_fdpic_tables(struct linux_binprm *bprm,
 	char *k_platform, *k_base_platform;
 	char __user *u_platform, *u_base_platform, *p;
 	int loop;
-	int nr;	/* reset for each csp adjustment */
 	unsigned long flags = 0;
+	int ei_index;
+	elf_addr_t *elf_info;
 
 #ifdef CONFIG_MMU
 	/* In some cases (e.g. Hyper-Threading), we want to avoid L1 evictions
@@ -601,44 +602,24 @@ static int create_elf_fdpic_tables(struct linux_binprm *bprm,
 	csp -= sp & 15UL;
 	sp -= sp & 15UL;
 
-	/* put the ELF interpreter info on the stack */
-#define NEW_AUX_ENT(id, val)						\
-	do {								\
-		struct { unsigned long _id, _val; } __user *ent, v;	\
-									\
-		ent = (void __user *) csp;				\
-		v._id = (id);						\
-		v._val = (val);						\
-		if (copy_to_user(ent + nr, &v, sizeof(v)))		\
-			return -EFAULT;					\
-		nr++;							\
+	/* Create the ELF interpreter info */
+	elf_info = (elf_addr_t *)mm->saved_auxv;
+	/* update AT_VECTOR_SIZE_BASE if the number of NEW_AUX_ENT() changes */
+#define NEW_AUX_ENT(id, val) \
+	do { \
+		*elf_info++ = id; \
+		*elf_info++ = val; \
 	} while (0)
 
-	nr = 0;
-	csp -= 2 * sizeof(unsigned long);
-	NEW_AUX_ENT(AT_NULL, 0);
-	if (k_platform) {
-		nr = 0;
-		csp -= 2 * sizeof(unsigned long);
-		NEW_AUX_ENT(AT_PLATFORM,
-			    (elf_addr_t) (unsigned long) u_platform);
-	}
-
-	if (k_base_platform) {
-		nr = 0;
-		csp -= 2 * sizeof(unsigned long);
-		NEW_AUX_ENT(AT_BASE_PLATFORM,
-			    (elf_addr_t) (unsigned long) u_base_platform);
-	}
-
-	if (bprm->have_execfd) {
-		nr = 0;
-		csp -= 2 * sizeof(unsigned long);
-		NEW_AUX_ENT(AT_EXECFD, bprm->execfd);
-	}
-
-	nr = 0;
-	csp -= DLINFO_ITEMS * 2 * sizeof(unsigned long);
+#ifdef ARCH_DLINFO
+	/*
+	 * ARCH_DLINFO must come first so PPC can do its special alignment of
+	 * AUXV.
+	 * update AT_VECTOR_SIZE_ARCH if the number of NEW_AUX_ENT() in
+	 * ARCH_DLINFO changes
+	 */
+	ARCH_DLINFO;
+#endif
 	NEW_AUX_ENT(AT_HWCAP,	ELF_HWCAP);
 #ifdef ELF_HWCAP2
 	NEW_AUX_ENT(AT_HWCAP2,	ELF_HWCAP2);
@@ -659,17 +640,29 @@ static int create_elf_fdpic_tables(struct linux_binprm *bprm,
 	NEW_AUX_ENT(AT_EGID,	(elf_addr_t) from_kgid_munged(cred->user_ns, cred->egid));
 	NEW_AUX_ENT(AT_SECURE,	bprm->secureexec);
 	NEW_AUX_ENT(AT_EXECFN,	bprm->exec);
+	if (k_platform)
+		NEW_AUX_ENT(AT_PLATFORM,
+			    (elf_addr_t)(unsigned long)u_platform);
+	if (k_base_platform)
+		NEW_AUX_ENT(AT_BASE_PLATFORM,
+			    (elf_addr_t)(unsigned long)u_base_platform);
+	if (bprm->have_execfd)
+		NEW_AUX_ENT(AT_EXECFD, bprm->execfd);
+#undef NEW_AUX_ENT
+	/* AT_NULL is zero; clear the rest too */
+	memset(elf_info, 0, (char *)mm->saved_auxv +
+	       sizeof(mm->saved_auxv) - (char *)elf_info);
 
-#ifdef ARCH_DLINFO
-	nr = 0;
-	csp -= AT_VECTOR_SIZE_ARCH * 2 * sizeof(unsigned long);
+	/* And advance past the AT_NULL entry.  */
+	elf_info += 2;
 
-	/* ARCH_DLINFO must come last so platform specific code can enforce
-	 * special alignment requirements on the AUXV if necessary (eg. PPC).
-	 */
-	ARCH_DLINFO;
-#endif
-#undef NEW_AUX_ENT
+	ei_index = elf_info - (elf_addr_t *)mm->saved_auxv;
+	csp -= ei_index * sizeof(elf_addr_t);
+
+	/* Put the elf_info on the stack in the right place.  */
+	if (copy_to_user((void __user *)csp, mm->saved_auxv,
+			 ei_index * sizeof(elf_addr_t)))
+		return -EFAULT;
 
 	/* allocate room for argv[] and envv[] */
 	csp -= (bprm->envc + 1) * sizeof(elf_caddr_t);
@@ -1359,7 +1352,7 @@ static int fill_psinfo(struct elf_prpsinfo *psinfo, struct task_struct *p,
 	SET_UID(psinfo->pr_uid, from_kuid_munged(cred->user_ns, cred->uid));
 	SET_GID(psinfo->pr_gid, from_kgid_munged(cred->user_ns, cred->gid));
 	rcu_read_unlock();
-	strncpy(psinfo->pr_fname, p->comm, sizeof(psinfo->pr_fname));
+	get_task_comm(psinfo->pr_fname, p);
 
 	return 0;
 }
diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c
index c26545d71d..cd6d5bbb4b 100644
--- a/fs/binfmt_flat.c
+++ b/fs/binfmt_flat.c
@@ -72,8 +72,10 @@
 
 #ifdef CONFIG_BINFMT_FLAT_NO_DATA_START_OFFSET
 #define DATA_START_OFFSET_WORDS		(0)
+#define MAX_SHARED_LIBS_UPDATE		(0)
 #else
 #define DATA_START_OFFSET_WORDS		(MAX_SHARED_LIBS)
+#define MAX_SHARED_LIBS_UPDATE		(MAX_SHARED_LIBS)
 #endif
 
 struct lib_info {
@@ -880,7 +882,7 @@ static int load_flat_binary(struct linux_binprm *bprm)
 		return res;
 
 	/* Update data segment pointers for all libraries */
-	for (i = 0; i < MAX_SHARED_LIBS; i++) {
+	for (i = 0; i < MAX_SHARED_LIBS_UPDATE; i++) {
 		if (!libinfo.lib_list[i].loaded)
 			continue;
 		for (j = 0; j < MAX_SHARED_LIBS; j++) {
diff --git a/fs/btrfs/accessors.c b/fs/btrfs/accessors.c
index 1925a0919c..79026917db 100644
--- a/fs/btrfs/accessors.c
+++ b/fs/btrfs/accessors.c
@@ -5,7 +5,8 @@
 
 #include <asm/unaligned.h>
 #include "messages.h"
-#include "ctree.h"
+#include "extent_io.h"
+#include "fs.h"
 #include "accessors.h"
 
 static bool check_setget_bounds(const struct extent_buffer *eb,
@@ -63,8 +64,8 @@ u##bits btrfs_get_token_##bits(struct btrfs_map_token *token,		\
 	const unsigned long idx = get_eb_folio_index(token->eb, member_offset); \
 	const unsigned long oil = get_eb_offset_in_folio(token->eb,	\
 							 member_offset);\
-	const int unit_size = folio_size(token->eb->folios[0]);		\
-	const int unit_shift = folio_shift(token->eb->folios[0]);	\
+	const int unit_size = token->eb->folio_size;			\
+	const int unit_shift = token->eb->folio_shift;			\
 	const int size = sizeof(u##bits);				\
 	u8 lebytes[sizeof(u##bits)];					\
 	const int part = unit_size - oil;				\
@@ -94,7 +95,7 @@ u##bits btrfs_get_##bits(const struct extent_buffer *eb,		\
 	const unsigned long idx = get_eb_folio_index(eb, member_offset);\
 	const unsigned long oil = get_eb_offset_in_folio(eb,		\
 							 member_offset);\
-	const int unit_size = folio_size(eb->folios[0]);		\
+	const int unit_size = eb->folio_size;				\
 	char *kaddr = folio_address(eb->folios[idx]);			\
 	const int size = sizeof(u##bits);				\
 	const int part = unit_size - oil;				\
@@ -117,8 +118,8 @@ void btrfs_set_token_##bits(struct btrfs_map_token *token,		\
 	const unsigned long idx = get_eb_folio_index(token->eb, member_offset); \
 	const unsigned long oil = get_eb_offset_in_folio(token->eb,	\
 							 member_offset);\
-	const int unit_size = folio_size(token->eb->folios[0]);		\
-	const int unit_shift = folio_shift(token->eb->folios[0]);	\
+	const int unit_size = token->eb->folio_size;			\
+	const int unit_shift = token->eb->folio_shift;			\
 	const int size = sizeof(u##bits);				\
 	u8 lebytes[sizeof(u##bits)];					\
 	const int part = unit_size - oil;				\
@@ -151,7 +152,7 @@ void btrfs_set_##bits(const struct extent_buffer *eb, void *ptr,	\
 	const unsigned long idx = get_eb_folio_index(eb, member_offset);\
 	const unsigned long oil = get_eb_offset_in_folio(eb,		\
 							 member_offset);\
-	const int unit_size = folio_size(eb->folios[0]);		\
+	const int unit_size = eb->folio_size;				\
 	char *kaddr = folio_address(eb->folios[idx]);			\
 	const int size = sizeof(u##bits);				\
 	const int part = unit_size - oil;				\
diff --git a/fs/btrfs/accessors.h b/fs/btrfs/accessors.h
index ed7aa32972..6fce3e8d3d 100644
--- a/fs/btrfs/accessors.h
+++ b/fs/btrfs/accessors.h
@@ -3,8 +3,17 @@
 #ifndef BTRFS_ACCESSORS_H
 #define BTRFS_ACCESSORS_H
 
-#include <linux/stddef.h>
 #include <asm/unaligned.h>
+#include <linux/stddef.h>
+#include <linux/types.h>
+#include <linux/align.h>
+#include <linux/build_bug.h>
+#include <linux/compiler.h>
+#include <linux/string.h>
+#include <linux/mm.h>
+#include <uapi/linux/btrfs_tree.h>
+
+struct extent_buffer;
 
 struct btrfs_map_token {
 	struct extent_buffer *eb;
@@ -844,45 +853,6 @@ static inline void btrfs_set_balance_sys(struct extent_buffer *eb,
 	write_eb_member(eb, bi, struct btrfs_balance_item, sys, ba);
 }
 
-static inline void btrfs_disk_balance_args_to_cpu(struct btrfs_balance_args *cpu,
-			       const struct btrfs_disk_balance_args *disk)
-{
-	memset(cpu, 0, sizeof(*cpu));
-
-	cpu->profiles = le64_to_cpu(disk->profiles);
-	cpu->usage = le64_to_cpu(disk->usage);
-	cpu->devid = le64_to_cpu(disk->devid);
-	cpu->pstart = le64_to_cpu(disk->pstart);
-	cpu->pend = le64_to_cpu(disk->pend);
-	cpu->vstart = le64_to_cpu(disk->vstart);
-	cpu->vend = le64_to_cpu(disk->vend);
-	cpu->target = le64_to_cpu(disk->target);
-	cpu->flags = le64_to_cpu(disk->flags);
-	cpu->limit = le64_to_cpu(disk->limit);
-	cpu->stripes_min = le32_to_cpu(disk->stripes_min);
-	cpu->stripes_max = le32_to_cpu(disk->stripes_max);
-}
-
-static inline void btrfs_cpu_balance_args_to_disk(
-				struct btrfs_disk_balance_args *disk,
-				const struct btrfs_balance_args *cpu)
-{
-	memset(disk, 0, sizeof(*disk));
-
-	disk->profiles = cpu_to_le64(cpu->profiles);
-	disk->usage = cpu_to_le64(cpu->usage);
-	disk->devid = cpu_to_le64(cpu->devid);
-	disk->pstart = cpu_to_le64(cpu->pstart);
-	disk->pend = cpu_to_le64(cpu->pend);
-	disk->vstart = cpu_to_le64(cpu->vstart);
-	disk->vend = cpu_to_le64(cpu->vend);
-	disk->target = cpu_to_le64(cpu->target);
-	disk->flags = cpu_to_le64(cpu->flags);
-	disk->limit = cpu_to_le64(cpu->limit);
-	disk->stripes_min = cpu_to_le32(cpu->stripes_min);
-	disk->stripes_max = cpu_to_le32(cpu->stripes_max);
-}
-
 /* struct btrfs_super_block */
 BTRFS_SETGET_STACK_FUNCS(super_bytenr, struct btrfs_super_block, bytenr, 64);
 BTRFS_SETGET_STACK_FUNCS(super_flags, struct btrfs_super_block, flags, 64);
diff --git a/fs/btrfs/acl.c b/fs/btrfs/acl.c
index 7427449a04..e0ba00d64e 100644
--- a/fs/btrfs/acl.c
+++ b/fs/btrfs/acl.c
@@ -12,7 +12,6 @@
 #include <linux/sched/mm.h>
 #include <linux/slab.h>
 #include "ctree.h"
-#include "btrfs_inode.h"
 #include "xattr.h"
 #include "acl.h"
 
diff --git a/fs/btrfs/acl.h b/fs/btrfs/acl.h
index a270e71ec0..48b9ddae4a 100644
--- a/fs/btrfs/acl.h
+++ b/fs/btrfs/acl.h
@@ -3,8 +3,15 @@
 #ifndef BTRFS_ACL_H
 #define BTRFS_ACL_H
 
+struct posix_acl;
+struct inode;
+struct btrfs_trans_handle;
+
 #ifdef CONFIG_BTRFS_FS_POSIX_ACL
 
+struct mnt_idmap;
+struct dentry;
+
 struct posix_acl *btrfs_get_acl(struct inode *inode, int type, bool rcu);
 int btrfs_set_acl(struct mnt_idmap *idmap, struct dentry *dentry,
 		  struct posix_acl *acl, int type);
@@ -13,6 +20,10 @@ int __btrfs_set_acl(struct btrfs_trans_handle *trans, struct inode *inode,
 
 #else
 
+#include <linux/errno.h>
+
+struct btrfs_trans_handle;
+
 #define btrfs_get_acl NULL
 #define btrfs_set_acl NULL
 static inline int __btrfs_set_acl(struct btrfs_trans_handle *trans,
diff --git a/fs/btrfs/async-thread.c b/fs/btrfs/async-thread.c
index 9e261aac67..361a866c19 100644
--- a/fs/btrfs/async-thread.c
+++ b/fs/btrfs/async-thread.c
@@ -11,7 +11,6 @@
 #include <linux/freezer.h>
 #include <trace/events/btrfs.h>
 #include "async-thread.h"
-#include "ctree.h"
 
 enum {
 	WORK_DONE_BIT,
diff --git a/fs/btrfs/async-thread.h b/fs/btrfs/async-thread.h
index 62b8a0d578..04c2f31758 100644
--- a/fs/btrfs/async-thread.h
+++ b/fs/btrfs/async-thread.h
@@ -7,11 +7,14 @@
 #ifndef BTRFS_ASYNC_THREAD_H
 #define BTRFS_ASYNC_THREAD_H
 
+#include <linux/compiler_types.h>
 #include <linux/workqueue.h>
+#include <linux/list.h>
 
 struct btrfs_fs_info;
 struct btrfs_workqueue;
 struct btrfs_work;
+
 typedef void (*btrfs_func_t)(struct btrfs_work *arg);
 typedef void (*btrfs_ordered_func_t)(struct btrfs_work *arg, bool);
 
diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
index fc0930dd92..a2de5c05f9 100644
--- a/fs/btrfs/backref.c
+++ b/fs/btrfs/backref.c
@@ -198,10 +198,7 @@ static struct kmem_cache *btrfs_prelim_ref_cache;
 int __init btrfs_prelim_ref_init(void)
 {
 	btrfs_prelim_ref_cache = kmem_cache_create("btrfs_prelim_ref",
-					sizeof(struct prelim_ref),
-					0,
-					SLAB_MEM_SPREAD,
-					NULL);
+					sizeof(struct prelim_ref), 0, 0, NULL);
 	if (!btrfs_prelim_ref_cache)
 		return -ENOMEM;
 	return 0;
@@ -264,7 +261,7 @@ static void update_share_count(struct share_check *sc, int oldcount,
 	else if (oldcount < 1 && newcount > 0)
 		sc->share_count++;
 
-	if (newref->root_id == sc->root->root_key.objectid &&
+	if (newref->root_id == btrfs_root_id(sc->root) &&
 	    newref->wanted_disk_byte == sc->data_bytenr &&
 	    newref->key_for_search.objectid == sc->inum)
 		sc->self_ref_count += newref->count;
@@ -772,7 +769,7 @@ static int resolve_indirect_refs(struct btrfs_backref_walk_ctx *ctx,
 			continue;
 		}
 
-		if (sc && ref->root_id != sc->root->root_key.objectid) {
+		if (sc && ref->root_id != btrfs_root_id(sc->root)) {
 			free_pref(ref);
 			ret = BACKREF_FOUND_SHARED;
 			goto out;
@@ -922,40 +919,38 @@ static int add_delayed_refs(const struct btrfs_fs_info *fs_info,
 		switch (node->type) {
 		case BTRFS_TREE_BLOCK_REF_KEY: {
 			/* NORMAL INDIRECT METADATA backref */
-			struct btrfs_delayed_tree_ref *ref;
 			struct btrfs_key *key_ptr = NULL;
+			/* The owner of a tree block ref is the level. */
+			int level = btrfs_delayed_ref_owner(node);
 
 			if (head->extent_op && head->extent_op->update_key) {
 				btrfs_disk_key_to_cpu(&key, &head->extent_op->key);
 				key_ptr = &key;
 			}
 
-			ref = btrfs_delayed_node_to_tree_ref(node);
-			ret = add_indirect_ref(fs_info, preftrees, ref->root,
-					       key_ptr, ref->level + 1,
-					       node->bytenr, count, sc,
-					       GFP_ATOMIC);
+			ret = add_indirect_ref(fs_info, preftrees, node->ref_root,
+					       key_ptr, level + 1, node->bytenr,
+					       count, sc, GFP_ATOMIC);
 			break;
 		}
 		case BTRFS_SHARED_BLOCK_REF_KEY: {
-			/* SHARED DIRECT METADATA backref */
-			struct btrfs_delayed_tree_ref *ref;
-
-			ref = btrfs_delayed_node_to_tree_ref(node);
+			/*
+			 * SHARED DIRECT METADATA backref
+			 *
+			 * The owner of a tree block ref is the level.
+			 */
+			int level = btrfs_delayed_ref_owner(node);
 
-			ret = add_direct_ref(fs_info, preftrees, ref->level + 1,
-					     ref->parent, node->bytenr, count,
+			ret = add_direct_ref(fs_info, preftrees, level + 1,
+					     node->parent, node->bytenr, count,
 					     sc, GFP_ATOMIC);
 			break;
 		}
 		case BTRFS_EXTENT_DATA_REF_KEY: {
 			/* NORMAL INDIRECT DATA backref */
-			struct btrfs_delayed_data_ref *ref;
-			ref = btrfs_delayed_node_to_data_ref(node);
-
-			key.objectid = ref->objectid;
+			key.objectid = btrfs_delayed_ref_owner(node);
 			key.type = BTRFS_EXTENT_DATA_KEY;
-			key.offset = ref->offset;
+			key.offset = btrfs_delayed_ref_offset(node);
 
 			/*
 			 * If we have a share check context and a reference for
@@ -975,18 +970,14 @@ static int add_delayed_refs(const struct btrfs_fs_info *fs_info,
 			if (sc && count < 0)
 				sc->have_delayed_delete_refs = true;
 
-			ret = add_indirect_ref(fs_info, preftrees, ref->root,
+			ret = add_indirect_ref(fs_info, preftrees, node->ref_root,
 					       &key, 0, node->bytenr, count, sc,
 					       GFP_ATOMIC);
 			break;
 		}
 		case BTRFS_SHARED_DATA_REF_KEY: {
 			/* SHARED DIRECT FULL backref */
-			struct btrfs_delayed_data_ref *ref;
-
-			ref = btrfs_delayed_node_to_data_ref(node);
-
-			ret = add_direct_ref(fs_info, preftrees, 0, ref->parent,
+			ret = add_direct_ref(fs_info, preftrees, 0, node->parent,
 					     node->bytenr, count, sc,
 					     GFP_ATOMIC);
 			break;
@@ -1036,8 +1027,6 @@ static int add_inline_refs(struct btrfs_backref_walk_ctx *ctx,
 	slot = path->slots[0];
 
 	item_size = btrfs_item_size(leaf, slot);
-	BUG_ON(item_size < sizeof(*ei));
-
 	ei = btrfs_item_ptr(leaf, slot, struct btrfs_extent_item);
 
 	if (ctx->check_extent_item) {
@@ -1435,8 +1424,10 @@ again:
 	if (ret < 0)
 		goto out;
 	if (ret == 0) {
-		/* This shouldn't happen, indicates a bug or fs corruption. */
-		ASSERT(ret != 0);
+		/*
+		 * Key with offset -1 found, there would have to exist an extent
+		 * item with such offset, but this is out of the valid range.
+		 */
 		ret = -EUCLEAN;
 		goto out;
 	}
@@ -2225,6 +2216,13 @@ int extent_from_logical(struct btrfs_fs_info *fs_info, u64 logical,
 	ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
 	if (ret < 0)
 		return ret;
+	if (ret == 0) {
+		/*
+		 * Key with offset -1 found, there would have to exist an extent
+		 * item with such offset, but this is out of the valid range.
+		 */
+		return -EUCLEAN;
+	}
 
 	ret = btrfs_previous_extent_item(extent_root, path, 0);
 	if (ret) {
@@ -2247,7 +2245,6 @@ int extent_from_logical(struct btrfs_fs_info *fs_info, u64 logical,
 
 	eb = path->nodes[0];
 	item_size = btrfs_item_size(eb, path->slots[0]);
-	BUG_ON(item_size < sizeof(*ei));
 
 	ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item);
 	flags = btrfs_extent_flags(eb, ei);
@@ -2626,7 +2623,7 @@ static int iterate_inode_refs(u64 inum, struct inode_fs_paths *ipath)
 			btrfs_debug(fs_root->fs_info,
 				"following ref at offset %u for inode %llu in tree %llu",
 				cur, found_key.objectid,
-				fs_root->root_key.objectid);
+				btrfs_root_id(fs_root));
 			ret = inode_to_path(parent, name_len,
 				      (unsigned long)(iref + 1), eb, ipath);
 			if (ret)
@@ -2844,6 +2841,16 @@ struct btrfs_backref_iter *btrfs_backref_iter_alloc(struct btrfs_fs_info *fs_inf
 	return ret;
 }
 
+static void btrfs_backref_iter_release(struct btrfs_backref_iter *iter)
+{
+	iter->bytenr = 0;
+	iter->item_ptr = 0;
+	iter->cur_ptr = 0;
+	iter->end_ptr = 0;
+	btrfs_release_path(iter->path);
+	memset(&iter->cur_key, 0, sizeof(iter->cur_key));
+}
+
 int btrfs_backref_iter_start(struct btrfs_backref_iter *iter, u64 bytenr)
 {
 	struct btrfs_fs_info *fs_info = iter->fs_info;
@@ -2862,6 +2869,10 @@ int btrfs_backref_iter_start(struct btrfs_backref_iter *iter, u64 bytenr)
 	if (ret < 0)
 		return ret;
 	if (ret == 0) {
+		/*
+		 * Key with offset -1 found, there would have to exist an extent
+		 * item with such offset, but this is out of the valid range.
+		 */
 		ret = -EUCLEAN;
 		goto release;
 	}
@@ -2932,6 +2943,14 @@ release:
 	return ret;
 }
 
+static bool btrfs_backref_iter_is_inline_ref(struct btrfs_backref_iter *iter)
+{
+	if (iter->cur_key.type == BTRFS_EXTENT_ITEM_KEY ||
+	    iter->cur_key.type == BTRFS_METADATA_ITEM_KEY)
+		return true;
+	return false;
+}
+
 /*
  * Go to the next backref item of current bytenr, can be either inlined or
  * keyed.
@@ -2944,7 +2963,7 @@ release:
  */
 int btrfs_backref_iter_next(struct btrfs_backref_iter *iter)
 {
-	struct extent_buffer *eb = btrfs_backref_get_eb(iter);
+	struct extent_buffer *eb = iter->path->nodes[0];
 	struct btrfs_root *extent_root;
 	struct btrfs_path *path = iter->path;
 	struct btrfs_extent_inline_ref *iref;
@@ -3032,6 +3051,19 @@ struct btrfs_backref_node *btrfs_backref_alloc_node(
 	return node;
 }
 
+void btrfs_backref_free_node(struct btrfs_backref_cache *cache,
+			     struct btrfs_backref_node *node)
+{
+	if (node) {
+		ASSERT(list_empty(&node->list));
+		ASSERT(list_empty(&node->lower));
+		ASSERT(node->eb == NULL);
+		cache->nr_nodes--;
+		btrfs_put_root(node->root);
+		kfree(node);
+	}
+}
+
 struct btrfs_backref_edge *btrfs_backref_alloc_edge(
 		struct btrfs_backref_cache *cache)
 {
@@ -3043,6 +3075,52 @@ struct btrfs_backref_edge *btrfs_backref_alloc_edge(
 	return edge;
 }
 
+void btrfs_backref_free_edge(struct btrfs_backref_cache *cache,
+			     struct btrfs_backref_edge *edge)
+{
+	if (edge) {
+		cache->nr_edges--;
+		kfree(edge);
+	}
+}
+
+void btrfs_backref_unlock_node_buffer(struct btrfs_backref_node *node)
+{
+	if (node->locked) {
+		btrfs_tree_unlock(node->eb);
+		node->locked = 0;
+	}
+}
+
+void btrfs_backref_drop_node_buffer(struct btrfs_backref_node *node)
+{
+	if (node->eb) {
+		btrfs_backref_unlock_node_buffer(node);
+		free_extent_buffer(node->eb);
+		node->eb = NULL;
+	}
+}
+
+/*
+ * Drop the backref node from cache without cleaning up its children
+ * edges.
+ *
+ * This can only be called on node without parent edges.
+ * The children edges are still kept as is.
+ */
+void btrfs_backref_drop_node(struct btrfs_backref_cache *tree,
+			     struct btrfs_backref_node *node)
+{
+	ASSERT(list_empty(&node->upper));
+
+	btrfs_backref_drop_node_buffer(node);
+	list_del_init(&node->list);
+	list_del_init(&node->lower);
+	if (!RB_EMPTY_NODE(&node->rb_node))
+		rb_erase(&node->rb_node, &tree->rb_root);
+	btrfs_backref_free_node(tree, node);
+}
+
 /*
  * Drop the backref node from cache, also cleaning up all its
  * upper edges and any uncached nodes in the path.
@@ -3114,6 +3192,19 @@ void btrfs_backref_release_cache(struct btrfs_backref_cache *cache)
 	ASSERT(!cache->nr_edges);
 }
 
+void btrfs_backref_link_edge(struct btrfs_backref_edge *edge,
+			     struct btrfs_backref_node *lower,
+			     struct btrfs_backref_node *upper,
+			     int link_which)
+{
+	ASSERT(upper && lower && upper->level == lower->level + 1);
+	edge->node[LOWER] = lower;
+	edge->node[UPPER] = upper;
+	if (link_which & LINK_LOWER)
+		list_add_tail(&edge->list[LOWER], &lower->upper);
+	if (link_which & LINK_UPPER)
+		list_add_tail(&edge->list[UPPER], &upper->lower);
+}
 /*
  * Handle direct tree backref
  *
@@ -3264,7 +3355,7 @@ static int handle_indirect_tree_backref(struct btrfs_trans_handle *trans,
 	if (btrfs_node_blockptr(eb, path->slots[level]) != cur->bytenr) {
 		btrfs_err(fs_info,
 "couldn't find block (%llu) (level %d) in tree (%llu) with key (%llu %u %llu)",
-			  cur->bytenr, level - 1, root->root_key.objectid,
+			  cur->bytenr, level - 1, btrfs_root_id(root),
 			  tree_key->objectid, tree_key->type, tree_key->offset);
 		btrfs_put_root(root);
 		ret = -ENOENT;
@@ -3422,7 +3513,7 @@ int btrfs_backref_add_tree_node(struct btrfs_trans_handle *trans,
 		int type;
 
 		cond_resched();
-		eb = btrfs_backref_get_eb(iter);
+		eb = iter->path->nodes[0];
 
 		key.objectid = iter->bytenr;
 		if (btrfs_backref_iter_is_inline_ref(iter)) {
diff --git a/fs/btrfs/backref.h b/fs/btrfs/backref.h
index ab4ca0eda6..e8c22cccb5 100644
--- a/fs/btrfs/backref.h
+++ b/fs/btrfs/backref.h
@@ -6,11 +6,23 @@
 #ifndef BTRFS_BACKREF_H
 #define BTRFS_BACKREF_H
 
-#include <linux/btrfs.h>
+#include <linux/types.h>
+#include <linux/rbtree.h>
+#include <linux/list.h>
+#include <linux/slab.h>
+#include <uapi/linux/btrfs.h>
+#include <uapi/linux/btrfs_tree.h>
 #include "messages.h"
-#include "ulist.h"
+#include "locking.h"
 #include "disk-io.h"
 #include "extent_io.h"
+#include "ctree.h"
+
+struct extent_inode_elem;
+struct ulist;
+struct btrfs_extent_item;
+struct btrfs_trans_handle;
+struct btrfs_fs_info;
 
 /*
  * Used by implementations of iterate_extent_inodes_t (see definition below) to
@@ -271,22 +283,6 @@ struct btrfs_backref_iter {
 
 struct btrfs_backref_iter *btrfs_backref_iter_alloc(struct btrfs_fs_info *fs_info);
 
-static inline void btrfs_backref_iter_free(struct btrfs_backref_iter *iter)
-{
-	if (!iter)
-		return;
-	btrfs_free_path(iter->path);
-	kfree(iter);
-}
-
-static inline struct extent_buffer *btrfs_backref_get_eb(
-		struct btrfs_backref_iter *iter)
-{
-	if (!iter)
-		return NULL;
-	return iter->path->nodes[0];
-}
-
 /*
  * For metadata with EXTENT_ITEM key (non-skinny) case, the first inline data
  * is btrfs_tree_block_info, without a btrfs_extent_inline_ref header.
@@ -306,25 +302,6 @@ int btrfs_backref_iter_start(struct btrfs_backref_iter *iter, u64 bytenr);
 
 int btrfs_backref_iter_next(struct btrfs_backref_iter *iter);
 
-static inline bool btrfs_backref_iter_is_inline_ref(
-		struct btrfs_backref_iter *iter)
-{
-	if (iter->cur_key.type == BTRFS_EXTENT_ITEM_KEY ||
-	    iter->cur_key.type == BTRFS_METADATA_ITEM_KEY)
-		return true;
-	return false;
-}
-
-static inline void btrfs_backref_iter_release(struct btrfs_backref_iter *iter)
-{
-	iter->bytenr = 0;
-	iter->item_ptr = 0;
-	iter->cur_ptr = 0;
-	iter->end_ptr = 0;
-	btrfs_release_path(iter->path);
-	memset(&iter->cur_key, 0, sizeof(iter->cur_key));
-}
-
 /*
  * Backref cache related structures
  *
@@ -452,83 +429,22 @@ struct btrfs_backref_edge *btrfs_backref_alloc_edge(
 
 #define		LINK_LOWER	(1 << 0)
 #define		LINK_UPPER	(1 << 1)
-static inline void btrfs_backref_link_edge(struct btrfs_backref_edge *edge,
-					   struct btrfs_backref_node *lower,
-					   struct btrfs_backref_node *upper,
-					   int link_which)
-{
-	ASSERT(upper && lower && upper->level == lower->level + 1);
-	edge->node[LOWER] = lower;
-	edge->node[UPPER] = upper;
-	if (link_which & LINK_LOWER)
-		list_add_tail(&edge->list[LOWER], &lower->upper);
-	if (link_which & LINK_UPPER)
-		list_add_tail(&edge->list[UPPER], &upper->lower);
-}
-
-static inline void btrfs_backref_free_node(struct btrfs_backref_cache *cache,
-					   struct btrfs_backref_node *node)
-{
-	if (node) {
-		ASSERT(list_empty(&node->list));
-		ASSERT(list_empty(&node->lower));
-		ASSERT(node->eb == NULL);
-		cache->nr_nodes--;
-		btrfs_put_root(node->root);
-		kfree(node);
-	}
-}
-
-static inline void btrfs_backref_free_edge(struct btrfs_backref_cache *cache,
-					   struct btrfs_backref_edge *edge)
-{
-	if (edge) {
-		cache->nr_edges--;
-		kfree(edge);
-	}
-}
-
-static inline void btrfs_backref_unlock_node_buffer(
-		struct btrfs_backref_node *node)
-{
-	if (node->locked) {
-		btrfs_tree_unlock(node->eb);
-		node->locked = 0;
-	}
-}
 
-static inline void btrfs_backref_drop_node_buffer(
-		struct btrfs_backref_node *node)
-{
-	if (node->eb) {
-		btrfs_backref_unlock_node_buffer(node);
-		free_extent_buffer(node->eb);
-		node->eb = NULL;
-	}
-}
-
-/*
- * Drop the backref node from cache without cleaning up its children
- * edges.
- *
- * This can only be called on node without parent edges.
- * The children edges are still kept as is.
- */
-static inline void btrfs_backref_drop_node(struct btrfs_backref_cache *tree,
-					   struct btrfs_backref_node *node)
-{
-	ASSERT(list_empty(&node->upper));
-
-	btrfs_backref_drop_node_buffer(node);
-	list_del_init(&node->list);
-	list_del_init(&node->lower);
-	if (!RB_EMPTY_NODE(&node->rb_node))
-		rb_erase(&node->rb_node, &tree->rb_root);
-	btrfs_backref_free_node(tree, node);
-}
+void btrfs_backref_link_edge(struct btrfs_backref_edge *edge,
+			     struct btrfs_backref_node *lower,
+			     struct btrfs_backref_node *upper,
+			     int link_which);
+void btrfs_backref_free_node(struct btrfs_backref_cache *cache,
+			     struct btrfs_backref_node *node);
+void btrfs_backref_free_edge(struct btrfs_backref_cache *cache,
+			     struct btrfs_backref_edge *edge);
+void btrfs_backref_unlock_node_buffer(struct btrfs_backref_node *node);
+void btrfs_backref_drop_node_buffer(struct btrfs_backref_node *node);
 
 void btrfs_backref_cleanup_node(struct btrfs_backref_cache *cache,
 				struct btrfs_backref_node *node);
+void btrfs_backref_drop_node(struct btrfs_backref_cache *tree,
+			     struct btrfs_backref_node *node);
 
 void btrfs_backref_release_cache(struct btrfs_backref_cache *cache);
 
diff --git a/fs/btrfs/bio.c b/fs/btrfs/bio.c
index 928f512cdb..e3a57196b0 100644
--- a/fs/btrfs/bio.c
+++ b/fs/btrfs/bio.c
@@ -11,7 +11,6 @@
 #include "raid56.h"
 #include "async-thread.h"
 #include "dev-replace.h"
-#include "rcu-string.h"
 #include "zoned.h"
 #include "file-item.h"
 #include "raid-stripe-tree.h"
@@ -509,8 +508,6 @@ static void __btrfs_submit_bio(struct bio *bio, struct btrfs_io_context *bioc,
 	if (!bioc) {
 		/* Single mirror read/write fast path. */
 		btrfs_bio(bio)->mirror_num = mirror_num;
-		if (bio_op(bio) != REQ_OP_READ)
-			btrfs_bio(bio)->orig_physical = smap->physical;
 		bio->bi_iter.bi_sector = smap->physical >> SECTOR_SHIFT;
 		if (bio_op(bio) != REQ_OP_READ)
 			btrfs_bio(bio)->orig_physical = smap->physical;
@@ -611,8 +608,20 @@ static void run_one_async_done(struct btrfs_work *work, bool do_free)
 
 static bool should_async_write(struct btrfs_bio *bbio)
 {
+	bool auto_csum_mode = true;
+
+#ifdef CONFIG_BTRFS_DEBUG
+	struct btrfs_fs_devices *fs_devices = bbio->fs_info->fs_devices;
+	enum btrfs_offload_csum_mode csum_mode = READ_ONCE(fs_devices->offload_csum_mode);
+
+	if (csum_mode == BTRFS_OFFLOAD_CSUM_FORCE_OFF)
+		return false;
+
+	auto_csum_mode = (csum_mode == BTRFS_OFFLOAD_CSUM_AUTO);
+#endif
+
 	/* Submit synchronously if the checksum implementation is fast. */
-	if (test_bit(BTRFS_FS_CSUM_IMPL_FAST, &bbio->fs_info->flags))
+	if (auto_csum_mode && test_bit(BTRFS_FS_CSUM_IMPL_FAST, &bbio->fs_info->flags))
 		return false;
 
 	/*
@@ -732,7 +741,9 @@ static bool btrfs_submit_chunk(struct btrfs_bio *bbio, int mirror_num)
 			ret = btrfs_bio_csum(bbio);
 			if (ret)
 				goto fail_put_bio;
-		} else if (use_append) {
+		} else if (use_append ||
+			   (btrfs_is_zoned(fs_info) && inode &&
+			    inode->flags & BTRFS_INODE_NODATASUM)) {
 			ret = btrfs_alloc_dummy_sum(bbio);
 			if (ret)
 				goto fail_put_bio;
diff --git a/fs/btrfs/bio.h b/fs/btrfs/bio.h
index bbaed31716..d9dd527609 100644
--- a/fs/btrfs/bio.h
+++ b/fs/btrfs/bio.h
@@ -7,12 +7,14 @@
 #ifndef BTRFS_BIO_H
 #define BTRFS_BIO_H
 
+#include <linux/types.h>
 #include <linux/bio.h>
 #include <linux/workqueue.h>
 #include "tree-checker.h"
 
 struct btrfs_bio;
 struct btrfs_fs_info;
+struct btrfs_inode;
 
 #define BTRFS_BIO_INLINE_CSUM_SIZE	64
 
diff --git a/fs/btrfs/block-group.c b/fs/btrfs/block-group.c
index be70acea87..350cd1201c 100644
--- a/fs/btrfs/block-group.c
+++ b/fs/btrfs/block-group.c
@@ -418,7 +418,7 @@ struct btrfs_caching_control *btrfs_get_caching_control(
 	return ctl;
 }
 
-void btrfs_put_caching_control(struct btrfs_caching_control *ctl)
+static void btrfs_put_caching_control(struct btrfs_caching_control *ctl)
 {
 	if (refcount_dec_and_test(&ctl->count))
 		kfree(ctl);
@@ -1063,7 +1063,9 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
 	bool remove_rsv = false;
 
 	block_group = btrfs_lookup_block_group(fs_info, map->start);
-	BUG_ON(!block_group);
+	if (!block_group)
+		return -ENOENT;
+
 	BUG_ON(!block_group->ro);
 
 	trace_btrfs_remove_block_group(block_group);
@@ -1214,8 +1216,8 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
 	block_group->space_info->total_bytes -= block_group->length;
 	block_group->space_info->bytes_readonly -=
 		(block_group->length - block_group->zone_unusable);
-	block_group->space_info->bytes_zone_unusable -=
-		block_group->zone_unusable;
+	btrfs_space_info_update_bytes_zone_unusable(fs_info, block_group->space_info,
+						    -block_group->zone_unusable);
 	block_group->space_info->disk_total -= block_group->length * factor;
 
 	spin_unlock(&block_group->space_info->lock);
@@ -1387,7 +1389,8 @@ static int inc_block_group_ro(struct btrfs_block_group *cache, int force)
 		if (btrfs_is_zoned(cache->fs_info)) {
 			/* Migrate zone_unusable bytes to readonly */
 			sinfo->bytes_readonly += cache->zone_unusable;
-			sinfo->bytes_zone_unusable -= cache->zone_unusable;
+			btrfs_space_info_update_bytes_zone_unusable(cache->fs_info, sinfo,
+								    -cache->zone_unusable);
 			cache->zone_unusable = 0;
 		}
 		cache->ro++;
@@ -1429,7 +1432,7 @@ static bool clean_pinned_extents(struct btrfs_trans_handle *trans,
 	 * group in pinned_extents before we were able to clear the whole block
 	 * group range from pinned_extents. This means that task can lookup for
 	 * the block group after we unpinned it from pinned_extents and removed
-	 * it, leading to a BUG_ON() at unpin_extent_range().
+	 * it, leading to an error at unpin_extent_range().
 	 */
 	mutex_lock(&fs_info->unused_bg_unpin_mutex);
 	if (prev_trans) {
@@ -1522,6 +1525,13 @@ void btrfs_delete_unused_bgs(struct btrfs_fs_info *fs_info)
 			 * outstanding allocations in this block group.  We do
 			 * the ro check in case balance is currently acting on
 			 * this block group.
+			 *
+			 * Also bail out if this is the only block group for its
+			 * type, because otherwise we would lose profile
+			 * information from fs_info->avail_*_alloc_bits and the
+			 * next block group of this type would be created with a
+			 * "single" profile (even if we're in a raid fs) because
+			 * fs_info->avail_*_alloc_bits would be 0.
 			 */
 			trace_btrfs_skip_unused_block_group(block_group);
 			spin_unlock(&block_group->lock);
@@ -1776,6 +1786,7 @@ void btrfs_reclaim_bgs_work(struct work_struct *work)
 		container_of(work, struct btrfs_fs_info, reclaim_bgs_work);
 	struct btrfs_block_group *bg;
 	struct btrfs_space_info *space_info;
+	LIST_HEAD(retry_list);
 
 	if (!test_bit(BTRFS_FS_OPEN, &fs_info->flags))
 		return;
@@ -1912,8 +1923,20 @@ void btrfs_reclaim_bgs_work(struct work_struct *work)
 		}
 
 next:
-		if (ret)
-			btrfs_mark_bg_to_reclaim(bg);
+		if (ret) {
+			/* Refcount held by the reclaim_bgs list after splice. */
+			spin_lock(&fs_info->unused_bgs_lock);
+			/*
+			 * This block group might be added to the unused list
+			 * during the above process. Move it back to the
+			 * reclaim list otherwise.
+			 */
+			if (list_empty(&bg->bg_list)) {
+				btrfs_get_block_group(bg);
+				list_add_tail(&bg->bg_list, &retry_list);
+			}
+			spin_unlock(&fs_info->unused_bgs_lock);
+		}
 		btrfs_put_block_group(bg);
 
 		mutex_unlock(&fs_info->reclaim_bgs_lock);
@@ -1933,6 +1956,9 @@ next:
 	spin_unlock(&fs_info->unused_bgs_lock);
 	mutex_unlock(&fs_info->reclaim_bgs_lock);
 end:
+	spin_lock(&fs_info->unused_bgs_lock);
+	list_splice_tail(&retry_list, &fs_info->reclaim_bgs);
+	spin_unlock(&fs_info->unused_bgs_lock);
 	btrfs_exclop_finish(fs_info);
 	sb_end_write(fs_info->sb);
 }
@@ -3009,9 +3035,11 @@ void btrfs_dec_block_group_ro(struct btrfs_block_group *cache)
 		if (btrfs_is_zoned(cache->fs_info)) {
 			/* Migrate zone_unusable bytes back */
 			cache->zone_unusable =
-				(cache->alloc_offset - cache->used) +
+				(cache->alloc_offset - cache->used - cache->pinned -
+				 cache->reserved) +
 				(cache->length - cache->zone_capacity);
-			sinfo->bytes_zone_unusable += cache->zone_unusable;
+			btrfs_space_info_update_bytes_zone_unusable(cache->fs_info, sinfo,
+								    cache->zone_unusable);
 			sinfo->bytes_readonly -= cache->zone_unusable;
 		}
 		num_bytes = cache->length - cache->reserved -
diff --git a/fs/btrfs/block-group.h b/fs/btrfs/block-group.h
index 962b119839..85e2d4cd12 100644
--- a/fs/btrfs/block-group.h
+++ b/fs/btrfs/block-group.h
@@ -3,9 +3,22 @@
 #ifndef BTRFS_BLOCK_GROUP_H
 #define BTRFS_BLOCK_GROUP_H
 
+#include <linux/atomic.h>
+#include <linux/mutex.h>
+#include <linux/list.h>
+#include <linux/spinlock.h>
+#include <linux/refcount.h>
+#include <linux/wait.h>
+#include <linux/sizes.h>
+#include <linux/rwsem.h>
+#include <linux/rbtree.h>
+#include <uapi/linux/btrfs_tree.h>
 #include "free-space-cache.h"
 
 struct btrfs_chunk_map;
+struct btrfs_fs_info;
+struct btrfs_inode;
+struct btrfs_trans_handle;
 
 enum btrfs_disk_cache_state {
 	BTRFS_DC_WRITTEN,
@@ -297,7 +310,6 @@ void btrfs_wait_nocow_writers(struct btrfs_block_group *bg);
 void btrfs_wait_block_group_cache_progress(struct btrfs_block_group *cache,
 				           u64 num_bytes);
 int btrfs_cache_block_group(struct btrfs_block_group *cache, bool wait);
-void btrfs_put_caching_control(struct btrfs_caching_control *ctl);
 struct btrfs_caching_control *btrfs_get_caching_control(
 		struct btrfs_block_group *cache);
 int btrfs_add_new_free_space(struct btrfs_block_group *block_group,
diff --git a/fs/btrfs/block-rsv.c b/fs/btrfs/block-rsv.c
index 1043a81423..b299b82d67 100644
--- a/fs/btrfs/block-rsv.c
+++ b/fs/btrfs/block-rsv.c
@@ -6,7 +6,6 @@
 #include "space-info.h"
 #include "transaction.h"
 #include "block-group.h"
-#include "disk-io.h"
 #include "fs.h"
 #include "accessors.h"
 
@@ -342,9 +341,9 @@ void btrfs_update_global_block_rsv(struct btrfs_fs_info *fs_info)
 	read_lock(&fs_info->global_root_lock);
 	rbtree_postorder_for_each_entry_safe(root, tmp, &fs_info->global_root_tree,
 					     rb_node) {
-		if (root->root_key.objectid == BTRFS_EXTENT_TREE_OBJECTID ||
-		    root->root_key.objectid == BTRFS_CSUM_TREE_OBJECTID ||
-		    root->root_key.objectid == BTRFS_FREE_SPACE_TREE_OBJECTID) {
+		if (btrfs_root_id(root) == BTRFS_EXTENT_TREE_OBJECTID ||
+		    btrfs_root_id(root) == BTRFS_CSUM_TREE_OBJECTID ||
+		    btrfs_root_id(root) == BTRFS_FREE_SPACE_TREE_OBJECTID) {
 			num_bytes += btrfs_root_used(&root->root_item);
 			min_items++;
 		}
@@ -407,7 +406,7 @@ void btrfs_init_root_block_rsv(struct btrfs_root *root)
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
 
-	switch (root->root_key.objectid) {
+	switch (btrfs_root_id(root)) {
 	case BTRFS_CSUM_TREE_OBJECTID:
 	case BTRFS_EXTENT_TREE_OBJECTID:
 	case BTRFS_FREE_SPACE_TREE_OBJECTID:
@@ -469,8 +468,7 @@ static struct btrfs_block_rsv *get_block_rsv(
 
 	if (test_bit(BTRFS_ROOT_SHAREABLE, &root->state) ||
 	    (root == fs_info->uuid_root) ||
-	    (trans->adding_csums &&
-	     root->root_key.objectid == BTRFS_CSUM_TREE_OBJECTID))
+	    (trans->adding_csums && btrfs_root_id(root) == BTRFS_CSUM_TREE_OBJECTID))
 		block_rsv = trans->block_rsv;
 
 	if (!block_rsv)
diff --git a/fs/btrfs/block-rsv.h b/fs/btrfs/block-rsv.h
index 43a9a6b5a7..1f53b967d0 100644
--- a/fs/btrfs/block-rsv.h
+++ b/fs/btrfs/block-rsv.h
@@ -3,8 +3,15 @@
 #ifndef BTRFS_BLOCK_RSV_H
 #define BTRFS_BLOCK_RSV_H
 
+#include <linux/types.h>
+#include <linux/compiler.h>
+#include <linux/spinlock.h>
+
 struct btrfs_trans_handle;
 struct btrfs_root;
+struct btrfs_space_info;
+struct btrfs_block_rsv;
+struct btrfs_fs_info;
 enum btrfs_reserve_flush_enum;
 
 /*
diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index 7f7c5a92d2..6ed495ca7a 100644
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -8,13 +8,32 @@
 
 #include <linux/hash.h>
 #include <linux/refcount.h>
+#include <linux/spinlock.h>
+#include <linux/mutex.h>
+#include <linux/rwsem.h>
+#include <linux/fs.h>
+#include <linux/mm.h>
+#include <linux/compiler.h>
 #include <linux/fscrypt.h>
+#include <linux/lockdep.h>
+#include <uapi/linux/btrfs_tree.h>
 #include <trace/events/btrfs.h>
+#include "block-rsv.h"
+#include "btrfs_inode.h"
 #include "extent_map.h"
 #include "extent_io.h"
+#include "extent-io-tree.h"
 #include "ordered-data.h"
 #include "delayed-inode.h"
 
+struct extent_state;
+struct posix_acl;
+struct iov_iter;
+struct writeback_control;
+struct btrfs_root;
+struct btrfs_fs_info;
+struct btrfs_trans_handle;
+
 /*
  * Since we search a directory based on f_pos (struct dir_context::pos) we have
  * to start at 2 since '.' and '..' have f_pos of 0 and 1 respectively, so
@@ -41,7 +60,6 @@ enum {
 	  */
 	BTRFS_INODE_NEEDS_FULL_SYNC,
 	BTRFS_INODE_COPY_EVERYTHING,
-	BTRFS_INODE_IN_DELALLOC_LIST,
 	BTRFS_INODE_HAS_PROPS,
 	BTRFS_INODE_SNAPSHOT_FLUSH,
 	/*
@@ -71,6 +89,16 @@ enum {
 	BTRFS_INODE_FREE_SPACE_INODE,
 	/* Set when there are no capabilities in XATTs for the inode. */
 	BTRFS_INODE_NO_CAP_XATTR,
+	/*
+	 * Set if an error happened when doing a COW write before submitting a
+	 * bio or during writeback. Used for both buffered writes and direct IO
+	 * writes. This is to signal a fast fsync that it has to wait for
+	 * ordered extents to complete and therefore not log extent maps that
+	 * point to unwritten extents (when an ordered extent completes and it
+	 * has the BTRFS_ORDERED_IOERR flag set, it drops extent maps in its
+	 * range).
+	 */
+	BTRFS_INODE_COW_WRITE_ERROR,
 };
 
 /* in memory btrfs inode */
@@ -363,9 +391,11 @@ static inline void btrfs_set_inode_last_sub_trans(struct btrfs_inode *inode)
 }
 
 /*
- * Should be called while holding the inode's VFS lock in exclusive mode or in a
- * context where no one else can access the inode concurrently (during inode
- * creation or when loading an inode from disk).
+ * Should be called while holding the inode's VFS lock in exclusive mode, or
+ * while holding the inode's mmap lock (struct btrfs_inode::i_mmap_lock) in
+ * either shared or exclusive mode, or in a context where no one else can access
+ * the inode concurrently (during inode creation or when loading an inode from
+ * disk).
  */
 static inline void btrfs_set_inode_full_sync(struct btrfs_inode *inode)
 {
@@ -428,7 +458,7 @@ noinline int can_nocow_extent(struct inode *inode, u64 offset, u64 *len,
 			      u64 *orig_start, u64 *orig_block_len,
 			      u64 *ram_bytes, bool nowait, bool strict);
 
-void __btrfs_del_delalloc_inode(struct btrfs_root *root, struct btrfs_inode *inode);
+void btrfs_del_delalloc_inode(struct btrfs_inode *inode);
 struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry);
 int btrfs_set_inode_index(struct btrfs_inode *dir, u64 *index);
 int btrfs_unlink_inode(struct btrfs_trans_handle *trans,
@@ -478,7 +508,6 @@ void btrfs_merge_delalloc_extent(struct btrfs_inode *inode, struct extent_state
 void btrfs_split_delalloc_extent(struct btrfs_inode *inode,
 				 struct extent_state *orig, u64 split);
 void btrfs_set_range_writeback(struct btrfs_inode *inode, u64 start, u64 end);
-vm_fault_t btrfs_page_mkwrite(struct vm_fault *vmf);
 void btrfs_evict_inode(struct inode *inode);
 struct inode *btrfs_alloc_inode(struct super_block *sb);
 void btrfs_destroy_inode(struct inode *inode);
@@ -490,8 +519,7 @@ struct inode *btrfs_iget_path(struct super_block *s, u64 ino,
 			      struct btrfs_root *root, struct btrfs_path *path);
 struct inode *btrfs_iget(struct super_block *s, u64 ino, struct btrfs_root *root);
 struct extent_map *btrfs_get_extent(struct btrfs_inode *inode,
-				    struct page *page, size_t pg_offset,
-				    u64 start, u64 len);
+				    struct page *page, u64 start, u64 len);
 int btrfs_update_inode(struct btrfs_trans_handle *trans,
 		       struct btrfs_inode *inode);
 int btrfs_update_inode_fallback(struct btrfs_trans_handle *trans,
@@ -527,6 +555,7 @@ ssize_t btrfs_dio_read(struct kiocb *iocb, struct iov_iter *iter,
 		       size_t done_before);
 struct iomap_dio *btrfs_dio_write(struct kiocb *iocb, struct iov_iter *iter,
 				  size_t done_before);
+struct btrfs_inode *btrfs_find_first_inode(struct btrfs_root *root, u64 min_ino);
 
 extern const struct dentry_operations btrfs_dentry_operations;
 
diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c
index 0041613a36..7c22f5b8c5 100644
--- a/fs/btrfs/compression.c
+++ b/fs/btrfs/compression.c
@@ -25,8 +25,6 @@
 #include "misc.h"
 #include "ctree.h"
 #include "fs.h"
-#include "disk-io.h"
-#include "transaction.h"
 #include "btrfs_inode.h"
 #include "bio.h"
 #include "ordered-data.h"
@@ -34,8 +32,7 @@
 #include "extent_io.h"
 #include "extent_map.h"
 #include "subpage.h"
-#include "zoned.h"
-#include "file-item.h"
+#include "messages.h"
 #include "super.h"
 
 static struct bio_set btrfs_compressed_bioset;
@@ -93,20 +90,20 @@ bool btrfs_compress_is_valid_type(const char *str, size_t len)
 }
 
 static int compression_compress_pages(int type, struct list_head *ws,
-               struct address_space *mapping, u64 start, struct page **pages,
-               unsigned long *out_pages, unsigned long *total_in,
-               unsigned long *total_out)
+				      struct address_space *mapping, u64 start,
+				      struct folio **folios, unsigned long *out_folios,
+				      unsigned long *total_in, unsigned long *total_out)
 {
 	switch (type) {
 	case BTRFS_COMPRESS_ZLIB:
-		return zlib_compress_pages(ws, mapping, start, pages,
-				out_pages, total_in, total_out);
+		return zlib_compress_folios(ws, mapping, start, folios,
+					    out_folios, total_in, total_out);
 	case BTRFS_COMPRESS_LZO:
-		return lzo_compress_pages(ws, mapping, start, pages,
-				out_pages, total_in, total_out);
+		return lzo_compress_folios(ws, mapping, start, folios,
+					   out_folios, total_in, total_out);
 	case BTRFS_COMPRESS_ZSTD:
-		return zstd_compress_pages(ws, mapping, start, pages,
-				out_pages, total_in, total_out);
+		return zstd_compress_folios(ws, mapping, start, folios,
+					    out_folios, total_in, total_out);
 	case BTRFS_COMPRESS_NONE:
 	default:
 		/*
@@ -118,7 +115,7 @@ static int compression_compress_pages(int type, struct list_head *ws,
 		 * Not a big deal, just need to inform caller that we
 		 * haven't allocated any pages yet.
 		 */
-		*out_pages = 0;
+		*out_folios = 0;
 		return -E2BIG;
 	}
 }
@@ -161,11 +158,11 @@ static int compression_decompress(int type, struct list_head *ws,
 	}
 }
 
-static void btrfs_free_compressed_pages(struct compressed_bio *cb)
+static void btrfs_free_compressed_folios(struct compressed_bio *cb)
 {
-	for (unsigned int i = 0; i < cb->nr_pages; i++)
-		btrfs_free_compr_page(cb->compressed_pages[i]);
-	kfree(cb->compressed_pages);
+	for (unsigned int i = 0; i < cb->nr_folios; i++)
+		btrfs_free_compr_folio(cb->compressed_folios[i]);
+	kfree(cb->compressed_folios);
 }
 
 static int btrfs_decompress_bio(struct compressed_bio *cb);
@@ -226,25 +223,25 @@ static unsigned long btrfs_compr_pool_scan(struct shrinker *sh, struct shrink_co
 /*
  * Common wrappers for page allocation from compression wrappers
  */
-struct page *btrfs_alloc_compr_page(void)
+struct folio *btrfs_alloc_compr_folio(void)
 {
-	struct page *page = NULL;
+	struct folio *folio = NULL;
 
 	spin_lock(&compr_pool.lock);
 	if (compr_pool.count > 0) {
-		page = list_first_entry(&compr_pool.list, struct page, lru);
-		list_del_init(&page->lru);
+		folio = list_first_entry(&compr_pool.list, struct folio, lru);
+		list_del_init(&folio->lru);
 		compr_pool.count--;
 	}
 	spin_unlock(&compr_pool.lock);
 
-	if (page)
-		return page;
+	if (folio)
+		return folio;
 
-	return alloc_page(GFP_NOFS);
+	return folio_alloc(GFP_NOFS, 0);
 }
 
-void btrfs_free_compr_page(struct page *page)
+void btrfs_free_compr_folio(struct folio *folio)
 {
 	bool do_free = false;
 
@@ -252,7 +249,7 @@ void btrfs_free_compr_page(struct page *page)
 	if (compr_pool.count > compr_pool.thresh) {
 		do_free = true;
 	} else {
-		list_add(&page->lru, &compr_pool.list);
+		list_add(&folio->lru, &compr_pool.list);
 		compr_pool.count++;
 	}
 	spin_unlock(&compr_pool.lock);
@@ -260,8 +257,8 @@ void btrfs_free_compr_page(struct page *page)
 	if (!do_free)
 		return;
 
-	ASSERT(page_ref_count(page) == 1);
-	put_page(page);
+	ASSERT(folio_ref_count(folio) == 1);
+	folio_put(folio);
 }
 
 static void end_bbio_comprssed_read(struct btrfs_bio *bbio)
@@ -272,7 +269,7 @@ static void end_bbio_comprssed_read(struct btrfs_bio *bbio)
 	if (!status)
 		status = errno_to_blk_status(btrfs_decompress_bio(cb));
 
-	btrfs_free_compressed_pages(cb);
+	btrfs_free_compressed_folios(cb);
 	btrfs_bio_end_io(cb->orig_bbio, status);
 	bio_put(&bbio->bio);
 }
@@ -326,7 +323,7 @@ static void btrfs_finish_compressed_write_work(struct work_struct *work)
 		end_compressed_writeback(cb);
 	/* Note, our inode could be gone now */
 
-	btrfs_free_compressed_pages(cb);
+	btrfs_free_compressed_folios(cb);
 	bio_put(&cb->bbio.bio);
 }
 
@@ -345,17 +342,19 @@ static void end_bbio_comprssed_write(struct btrfs_bio *bbio)
 	queue_work(fs_info->compressed_write_workers, &cb->write_end_work);
 }
 
-static void btrfs_add_compressed_bio_pages(struct compressed_bio *cb)
+static void btrfs_add_compressed_bio_folios(struct compressed_bio *cb)
 {
 	struct bio *bio = &cb->bbio.bio;
 	u32 offset = 0;
 
 	while (offset < cb->compressed_len) {
+		int ret;
 		u32 len = min_t(u32, cb->compressed_len - offset, PAGE_SIZE);
 
 		/* Maximum compressed extent is smaller than bio size limit. */
-		__bio_add_page(bio, cb->compressed_pages[offset >> PAGE_SHIFT],
-			       len, 0);
+		ret = bio_add_folio(bio, cb->compressed_folios[offset >> PAGE_SHIFT],
+				    len, 0);
+		ASSERT(ret);
 		offset += len;
 	}
 }
@@ -370,8 +369,8 @@ static void btrfs_add_compressed_bio_pages(struct compressed_bio *cb)
  * the end io hooks.
  */
 void btrfs_submit_compressed_write(struct btrfs_ordered_extent *ordered,
-				   struct page **compressed_pages,
-				   unsigned int nr_pages,
+				   struct folio **compressed_folios,
+				   unsigned int nr_folios,
 				   blk_opf_t write_flags,
 				   bool writeback)
 {
@@ -387,14 +386,14 @@ void btrfs_submit_compressed_write(struct btrfs_ordered_extent *ordered,
 				  end_bbio_comprssed_write);
 	cb->start = ordered->file_offset;
 	cb->len = ordered->num_bytes;
-	cb->compressed_pages = compressed_pages;
+	cb->compressed_folios = compressed_folios;
 	cb->compressed_len = ordered->disk_num_bytes;
 	cb->writeback = writeback;
 	INIT_WORK(&cb->write_end_work, btrfs_finish_compressed_write_work);
-	cb->nr_pages = nr_pages;
+	cb->nr_folios = nr_folios;
 	cb->bbio.bio.bi_iter.bi_sector = ordered->disk_bytenr >> SECTOR_SHIFT;
 	cb->bbio.ordered = ordered;
-	btrfs_add_compressed_bio_pages(cb);
+	btrfs_add_compressed_bio_folios(cb);
 
 	btrfs_submit_bio(&cb->bbio, 0);
 }
@@ -515,6 +514,7 @@ static noinline int add_ra_bio_pages(struct inode *inode,
 			put_page(page);
 			break;
 		}
+		add_size = min(em->start + em->len, page_end + 1) - cur;
 		free_extent_map(em);
 
 		if (page->index == end_index) {
@@ -527,7 +527,6 @@ static noinline int add_ra_bio_pages(struct inode *inode,
 			}
 		}
 
-		add_size = min(em->start + em->len, page_end + 1) - cur;
 		ret = bio_add_page(orig_bio, page, add_size, offset_in_page(cur));
 		if (ret != add_size) {
 			unlock_extent(tree, cur, page_end, NULL);
@@ -602,14 +601,14 @@ void btrfs_submit_compressed_read(struct btrfs_bio *bbio)
 
 	free_extent_map(em);
 
-	cb->nr_pages = DIV_ROUND_UP(compressed_len, PAGE_SIZE);
-	cb->compressed_pages = kcalloc(cb->nr_pages, sizeof(struct page *), GFP_NOFS);
-	if (!cb->compressed_pages) {
+	cb->nr_folios = DIV_ROUND_UP(compressed_len, PAGE_SIZE);
+	cb->compressed_folios = kcalloc(cb->nr_folios, sizeof(struct page *), GFP_NOFS);
+	if (!cb->compressed_folios) {
 		ret = BLK_STS_RESOURCE;
 		goto out_free_bio;
 	}
 
-	ret2 = btrfs_alloc_page_array(cb->nr_pages, cb->compressed_pages, 0);
+	ret2 = btrfs_alloc_folio_array(cb->nr_folios, cb->compressed_folios, 0);
 	if (ret2) {
 		ret = BLK_STS_RESOURCE;
 		goto out_free_compressed_pages;
@@ -621,7 +620,7 @@ void btrfs_submit_compressed_read(struct btrfs_bio *bbio)
 	/* include any pages we added in add_ra-bio_pages */
 	cb->len = bbio->bio.bi_iter.bi_size;
 	cb->bbio.bio.bi_iter.bi_sector = bbio->bio.bi_iter.bi_sector;
-	btrfs_add_compressed_bio_pages(cb);
+	btrfs_add_compressed_bio_folios(cb);
 
 	if (memstall)
 		psi_memstall_leave(&pflags);
@@ -630,7 +629,7 @@ void btrfs_submit_compressed_read(struct btrfs_bio *bbio)
 	return;
 
 out_free_compressed_pages:
-	kfree(cb->compressed_pages);
+	kfree(cb->compressed_folios);
 out_free_bio:
 	bio_put(&cb->bbio.bio);
 out:
@@ -977,6 +976,29 @@ static unsigned int btrfs_compress_set_level(int type, unsigned level)
 	return level;
 }
 
+/* Wrapper around find_get_page(), with extra error message. */
+int btrfs_compress_filemap_get_folio(struct address_space *mapping, u64 start,
+				     struct folio **in_folio_ret)
+{
+	struct folio *in_folio;
+
+	/*
+	 * The compressed write path should have the folio locked already, thus
+	 * we only need to grab one reference.
+	 */
+	in_folio = filemap_get_folio(mapping, start >> PAGE_SHIFT);
+	if (IS_ERR(in_folio)) {
+		struct btrfs_inode *inode = BTRFS_I(mapping->host);
+
+		btrfs_crit(inode->root->fs_info,
+		"failed to get page cache, root %lld ino %llu file offset %llu",
+			   btrfs_root_id(inode->root), btrfs_ino(inode), start);
+		return -ENOENT;
+	}
+	*in_folio_ret = in_folio;
+	return 0;
+}
+
 /*
  * Given an address space and start and length, compress the bytes into @pages
  * that are allocated on demand.
@@ -997,11 +1019,9 @@ static unsigned int btrfs_compress_set_level(int type, unsigned level)
  * @total_out is an in/out parameter, must be set to the input length and will
  * be also used to return the total number of compressed bytes
  */
-int btrfs_compress_pages(unsigned int type_level, struct address_space *mapping,
-			 u64 start, struct page **pages,
-			 unsigned long *out_pages,
-			 unsigned long *total_in,
-			 unsigned long *total_out)
+int btrfs_compress_folios(unsigned int type_level, struct address_space *mapping,
+			 u64 start, struct folio **folios, unsigned long *out_folios,
+			 unsigned long *total_in, unsigned long *total_out)
 {
 	int type = btrfs_compress_type(type_level);
 	int level = btrfs_compress_level(type_level);
@@ -1010,8 +1030,8 @@ int btrfs_compress_pages(unsigned int type_level, struct address_space *mapping,
 
 	level = btrfs_compress_set_level(type, level);
 	workspace = get_workspace(type, level);
-	ret = compression_compress_pages(type, workspace, mapping, start, pages,
-					 out_pages, total_in, total_out);
+	ret = compression_compress_pages(type, workspace, mapping, start, folios,
+					 out_folios, total_in, total_out);
 	put_workspace(type, workspace);
 	return ret;
 }
@@ -1479,11 +1499,6 @@ static void heuristic_collect_sample(struct inode *inode, u64 start, u64 end,
 /*
  * Compression heuristic.
  *
- * For now is's a naive and optimistic 'return true', we'll extend the logic to
- * quickly (compared to direct compression) detect data characteristics
- * (compressible/incompressible) to avoid wasting CPU time on incompressible
- * data.
- *
  * The following types of analysis can be performed:
  * - detect mostly zero data
  * - detect data with low "byte set" size (text, etc)
diff --git a/fs/btrfs/compression.h b/fs/btrfs/compression.h
index afd7e50d07..c20c1a1b09 100644
--- a/fs/btrfs/compression.h
+++ b/fs/btrfs/compression.h
@@ -7,10 +7,18 @@
 #define BTRFS_COMPRESSION_H
 
 #include <linux/sizes.h>
+#include <linux/mm.h>
+#include <linux/list.h>
+#include <linux/workqueue.h>
+#include <linux/wait.h>
 #include "bio.h"
 
+struct address_space;
+struct page;
+struct inode;
 struct btrfs_inode;
 struct btrfs_ordered_extent;
+struct btrfs_bio;
 
 /*
  * We want to make sure that amount of RAM required to uncompress an extent is
@@ -32,14 +40,12 @@ static_assert((BTRFS_MAX_COMPRESSED % PAGE_SIZE) == 0);
 
 #define	BTRFS_ZLIB_DEFAULT_LEVEL		3
 
-struct page;
-
 struct compressed_bio {
-	/* Number of compressed pages in the array */
-	unsigned int nr_pages;
+	/* Number of compressed folios in the array. */
+	unsigned int nr_folios;
 
-	/* the pages with the compressed data on them */
-	struct page **compressed_pages;
+	/* The folios with the compressed data on them. */
+	struct folio **compressed_folios;
 
 	/* starting offset in the inode for our pages */
 	u64 start;
@@ -79,27 +85,24 @@ static inline unsigned int btrfs_compress_level(unsigned int type_level)
 int __init btrfs_init_compress(void);
 void __cold btrfs_exit_compress(void);
 
-int btrfs_compress_pages(unsigned int type_level, struct address_space *mapping,
-			 u64 start, struct page **pages,
-			 unsigned long *out_pages,
-			 unsigned long *total_in,
-			 unsigned long *total_out);
+int btrfs_compress_folios(unsigned int type_level, struct address_space *mapping,
+			  u64 start, struct folio **folios, unsigned long *out_folios,
+			 unsigned long *total_in, unsigned long *total_out);
 int btrfs_decompress(int type, const u8 *data_in, struct page *dest_page,
 		     unsigned long start_byte, size_t srclen, size_t destlen);
 int btrfs_decompress_buf2page(const char *buf, u32 buf_len,
 			      struct compressed_bio *cb, u32 decompressed);
 
 void btrfs_submit_compressed_write(struct btrfs_ordered_extent *ordered,
-				  struct page **compressed_pages,
-				  unsigned int nr_pages,
-				  blk_opf_t write_flags,
-				  bool writeback);
+				   struct folio **compressed_folios,
+				   unsigned int nr_folios, blk_opf_t write_flags,
+				   bool writeback);
 void btrfs_submit_compressed_read(struct btrfs_bio *bbio);
 
 unsigned int btrfs_compress_str2level(unsigned int type, const char *str);
 
-struct page *btrfs_alloc_compr_page(void);
-void btrfs_free_compr_page(struct page *page);
+struct folio *btrfs_alloc_compr_folio(void);
+void btrfs_free_compr_folio(struct folio *folio);
 
 enum btrfs_compression_type {
 	BTRFS_COMPRESS_NONE  = 0,
@@ -143,8 +146,11 @@ bool btrfs_compress_is_valid_type(const char *str, size_t len);
 
 int btrfs_compress_heuristic(struct inode *inode, u64 start, u64 end);
 
-int zlib_compress_pages(struct list_head *ws, struct address_space *mapping,
-		u64 start, struct page **pages, unsigned long *out_pages,
+int btrfs_compress_filemap_get_folio(struct address_space *mapping, u64 start,
+				     struct folio **in_folio_ret);
+
+int zlib_compress_folios(struct list_head *ws, struct address_space *mapping,
+			 u64 start, struct folio **folios, unsigned long *out_folios,
 		unsigned long *total_in, unsigned long *total_out);
 int zlib_decompress_bio(struct list_head *ws, struct compressed_bio *cb);
 int zlib_decompress(struct list_head *ws, const u8 *data_in,
@@ -154,8 +160,8 @@ struct list_head *zlib_alloc_workspace(unsigned int level);
 void zlib_free_workspace(struct list_head *ws);
 struct list_head *zlib_get_workspace(unsigned int level);
 
-int lzo_compress_pages(struct list_head *ws, struct address_space *mapping,
-		u64 start, struct page **pages, unsigned long *out_pages,
+int lzo_compress_folios(struct list_head *ws, struct address_space *mapping,
+			u64 start, struct folio **folios, unsigned long *out_folios,
 		unsigned long *total_in, unsigned long *total_out);
 int lzo_decompress_bio(struct list_head *ws, struct compressed_bio *cb);
 int lzo_decompress(struct list_head *ws, const u8 *data_in,
@@ -164,12 +170,12 @@ int lzo_decompress(struct list_head *ws, const u8 *data_in,
 struct list_head *lzo_alloc_workspace(unsigned int level);
 void lzo_free_workspace(struct list_head *ws);
 
-int zstd_compress_pages(struct list_head *ws, struct address_space *mapping,
-		u64 start, struct page **pages, unsigned long *out_pages,
+int zstd_compress_folios(struct list_head *ws, struct address_space *mapping,
+			 u64 start, struct folio **folios, unsigned long *out_folios,
 		unsigned long *total_in, unsigned long *total_out);
 int zstd_decompress_bio(struct list_head *ws, struct compressed_bio *cb);
 int zstd_decompress(struct list_head *ws, const u8 *data_in,
-		struct page *dest_page, unsigned long start_byte, size_t srclen,
+		struct page *dest_page, unsigned long dest_pgoff, size_t srclen,
 		size_t destlen);
 void zstd_init_workspace_manager(void);
 void zstd_cleanup_workspace_manager(void);
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index e65e012bac..8a791b648a 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -291,7 +291,7 @@ static void add_root_to_dirty_list(struct btrfs_root *root)
 	spin_lock(&fs_info->trans_lock);
 	if (!test_and_set_bit(BTRFS_ROOT_DIRTY, &root->state)) {
 		/* Want the extent tree to be the last on the list */
-		if (root->root_key.objectid == BTRFS_EXTENT_TREE_OBJECTID)
+		if (btrfs_root_id(root) == BTRFS_EXTENT_TREE_OBJECTID)
 			list_move_tail(&root->dirty_list,
 				       &fs_info->dirty_cowonly_roots);
 		else
@@ -321,7 +321,7 @@ int btrfs_copy_root(struct btrfs_trans_handle *trans,
 	WARN_ON(test_bit(BTRFS_ROOT_SHAREABLE, &root->state) &&
 		trans->transid != fs_info->running_transaction->transid);
 	WARN_ON(test_bit(BTRFS_ROOT_SHAREABLE, &root->state) &&
-		trans->transid != root->last_trans);
+		trans->transid != btrfs_get_root_last_trans(root));
 
 	level = btrfs_header_level(buf);
 	if (level == 0)
@@ -454,7 +454,7 @@ static noinline int update_ref_for_cow(struct btrfs_trans_handle *trans,
 		}
 	} else {
 		refs = 1;
-		if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID ||
+		if (btrfs_root_id(root) == BTRFS_TREE_RELOC_OBJECTID ||
 		    btrfs_header_backref_rev(buf) < BTRFS_MIXED_BACKREF_REV)
 			flags = BTRFS_BLOCK_FLAG_FULL_BACKREF;
 		else
@@ -466,15 +466,14 @@ static noinline int update_ref_for_cow(struct btrfs_trans_handle *trans,
 	       !(flags & BTRFS_BLOCK_FLAG_FULL_BACKREF));
 
 	if (refs > 1) {
-		if ((owner == root->root_key.objectid ||
-		     root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) &&
+		if ((owner == btrfs_root_id(root) ||
+		     btrfs_root_id(root) == BTRFS_TREE_RELOC_OBJECTID) &&
 		    !(flags & BTRFS_BLOCK_FLAG_FULL_BACKREF)) {
 			ret = btrfs_inc_ref(trans, root, buf, 1);
 			if (ret)
 				return ret;
 
-			if (root->root_key.objectid ==
-			    BTRFS_TREE_RELOC_OBJECTID) {
+			if (btrfs_root_id(root) == BTRFS_TREE_RELOC_OBJECTID) {
 				ret = btrfs_dec_ref(trans, root, buf, 0);
 				if (ret)
 					return ret;
@@ -485,8 +484,7 @@ static noinline int update_ref_for_cow(struct btrfs_trans_handle *trans,
 			new_flags |= BTRFS_BLOCK_FLAG_FULL_BACKREF;
 		} else {
 
-			if (root->root_key.objectid ==
-			    BTRFS_TREE_RELOC_OBJECTID)
+			if (btrfs_root_id(root) == BTRFS_TREE_RELOC_OBJECTID)
 				ret = btrfs_inc_ref(trans, root, cow, 1);
 			else
 				ret = btrfs_inc_ref(trans, root, cow, 0);
@@ -500,8 +498,7 @@ static noinline int update_ref_for_cow(struct btrfs_trans_handle *trans,
 		}
 	} else {
 		if (flags & BTRFS_BLOCK_FLAG_FULL_BACKREF) {
-			if (root->root_key.objectid ==
-			    BTRFS_TREE_RELOC_OBJECTID)
+			if (btrfs_root_id(root) == BTRFS_TREE_RELOC_OBJECTID)
 				ret = btrfs_inc_ref(trans, root, cow, 1);
 			else
 				ret = btrfs_inc_ref(trans, root, cow, 0);
@@ -554,7 +551,7 @@ int btrfs_force_cow_block(struct btrfs_trans_handle *trans,
 	WARN_ON(test_bit(BTRFS_ROOT_SHAREABLE, &root->state) &&
 		trans->transid != fs_info->running_transaction->transid);
 	WARN_ON(test_bit(BTRFS_ROOT_SHAREABLE, &root->state) &&
-		trans->transid != root->last_trans);
+		trans->transid != btrfs_get_root_last_trans(root));
 
 	level = btrfs_header_level(buf);
 
@@ -563,13 +560,13 @@ int btrfs_force_cow_block(struct btrfs_trans_handle *trans,
 	else
 		btrfs_node_key(buf, &disk_key, 0);
 
-	if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) {
+	if (btrfs_root_id(root) == BTRFS_TREE_RELOC_OBJECTID) {
 		if (parent)
 			parent_start = parent->start;
 		reloc_src_root = btrfs_header_owner(buf);
 	}
 	cow = btrfs_alloc_tree_block(trans, root, parent_start,
-				     root->root_key.objectid, &disk_key, level,
+				     btrfs_root_id(root), &disk_key, level,
 				     search_start, empty_size, reloc_src_root, nest);
 	if (IS_ERR(cow))
 		return PTR_ERR(cow);
@@ -582,10 +579,10 @@ int btrfs_force_cow_block(struct btrfs_trans_handle *trans,
 	btrfs_set_header_backref_rev(cow, BTRFS_MIXED_BACKREF_REV);
 	btrfs_clear_header_flag(cow, BTRFS_HEADER_FLAG_WRITTEN |
 				     BTRFS_HEADER_FLAG_RELOC);
-	if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID)
+	if (btrfs_root_id(root) == BTRFS_TREE_RELOC_OBJECTID)
 		btrfs_set_header_flag(cow, BTRFS_HEADER_FLAG_RELOC);
 	else
-		btrfs_set_header_owner(cow, root->root_key.objectid);
+		btrfs_set_header_owner(cow, btrfs_root_id(root));
 
 	write_extent_buffer_fsid(cow, fs_info->fs_devices->metadata_uuid);
 
@@ -609,7 +606,7 @@ int btrfs_force_cow_block(struct btrfs_trans_handle *trans,
 
 	if (buf == root->node) {
 		WARN_ON(parent && parent != buf);
-		if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID ||
+		if (btrfs_root_id(root) == BTRFS_TREE_RELOC_OBJECTID ||
 		    btrfs_header_backref_rev(buf) < BTRFS_MIXED_BACKREF_REV)
 			parent_start = buf->start;
 
@@ -623,10 +620,16 @@ int btrfs_force_cow_block(struct btrfs_trans_handle *trans,
 		atomic_inc(&cow->refs);
 		rcu_assign_pointer(root->node, cow);
 
-		btrfs_free_tree_block(trans, btrfs_root_id(root), buf,
-				      parent_start, last_ref);
+		ret = btrfs_free_tree_block(trans, btrfs_root_id(root), buf,
+					    parent_start, last_ref);
 		free_extent_buffer(buf);
 		add_root_to_dirty_list(root);
+		if (ret < 0) {
+			btrfs_tree_unlock(cow);
+			free_extent_buffer(cow);
+			btrfs_abort_transaction(trans, ret);
+			return ret;
+		}
 	} else {
 		WARN_ON(trans->transid != btrfs_header_generation(parent));
 		ret = btrfs_tree_mod_log_insert_key(parent, parent_slot,
@@ -651,8 +654,14 @@ int btrfs_force_cow_block(struct btrfs_trans_handle *trans,
 				return ret;
 			}
 		}
-		btrfs_free_tree_block(trans, btrfs_root_id(root), buf,
-				      parent_start, last_ref);
+		ret = btrfs_free_tree_block(trans, btrfs_root_id(root), buf,
+					    parent_start, last_ref);
+		if (ret < 0) {
+			btrfs_tree_unlock(cow);
+			free_extent_buffer(cow);
+			btrfs_abort_transaction(trans, ret);
+			return ret;
+		}
 	}
 	if (unlock_orig)
 		btrfs_tree_unlock(buf);
@@ -685,7 +694,7 @@ static inline int should_cow_block(struct btrfs_trans_handle *trans,
 	 */
 	if (btrfs_header_generation(buf) == trans->transid &&
 	    !btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN) &&
-	    !(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID &&
+	    !(btrfs_root_id(root) != BTRFS_TREE_RELOC_OBJECTID &&
 	      btrfs_header_flag(buf, BTRFS_HEADER_FLAG_RELOC)) &&
 	    !test_bit(BTRFS_ROOT_FORCE_COW, &root->state))
 		return 0;
@@ -820,7 +829,7 @@ int btrfs_bin_search(struct extent_buffer *eb, int first_slot,
 	}
 
 	while (low < high) {
-		const int unit_size = folio_size(eb->folios[0]);
+		const int unit_size = eb->folio_size;
 		unsigned long oil;
 		unsigned long offset;
 		struct btrfs_disk_key *tmp;
@@ -986,9 +995,13 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 		free_extent_buffer(mid);
 
 		root_sub_used_bytes(root);
-		btrfs_free_tree_block(trans, btrfs_root_id(root), mid, 0, 1);
+		ret = btrfs_free_tree_block(trans, btrfs_root_id(root), mid, 0, 1);
 		/* once for the root ptr */
 		free_extent_buffer_stale(mid);
+		if (ret < 0) {
+			btrfs_abort_transaction(trans, ret);
+			goto out;
+		}
 		return 0;
 	}
 	if (btrfs_header_nritems(mid) >
@@ -1003,7 +1016,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 			goto out;
 		}
 
-		__btrfs_tree_lock(left, BTRFS_NESTING_LEFT);
+		btrfs_tree_lock_nested(left, BTRFS_NESTING_LEFT);
 		wret = btrfs_cow_block(trans, root, left,
 				       parent, pslot - 1, &left,
 				       BTRFS_NESTING_LEFT_COW);
@@ -1021,7 +1034,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 			goto out;
 		}
 
-		__btrfs_tree_lock(right, BTRFS_NESTING_RIGHT);
+		btrfs_tree_lock_nested(right, BTRFS_NESTING_RIGHT);
 		wret = btrfs_cow_block(trans, root, right,
 				       parent, pslot + 1, &right,
 				       BTRFS_NESTING_RIGHT_COW);
@@ -1056,10 +1069,14 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 				goto out;
 			}
 			root_sub_used_bytes(root);
-			btrfs_free_tree_block(trans, btrfs_root_id(root), right,
-					      0, 1);
+			ret = btrfs_free_tree_block(trans, btrfs_root_id(root),
+						    right, 0, 1);
 			free_extent_buffer_stale(right);
 			right = NULL;
+			if (ret < 0) {
+				btrfs_abort_transaction(trans, ret);
+				goto out;
+			}
 		} else {
 			struct btrfs_disk_key right_key;
 			btrfs_node_key(right, &right_key, 0);
@@ -1114,9 +1131,13 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 			goto out;
 		}
 		root_sub_used_bytes(root);
-		btrfs_free_tree_block(trans, btrfs_root_id(root), mid, 0, 1);
+		ret = btrfs_free_tree_block(trans, btrfs_root_id(root), mid, 0, 1);
 		free_extent_buffer_stale(mid);
 		mid = NULL;
+		if (ret < 0) {
+			btrfs_abort_transaction(trans, ret);
+			goto out;
+		}
 	} else {
 		/* update the parent key to reflect our changes */
 		struct btrfs_disk_key mid_key;
@@ -1205,7 +1226,7 @@ static noinline int push_nodes_for_insert(struct btrfs_trans_handle *trans,
 		if (IS_ERR(left))
 			return PTR_ERR(left);
 
-		__btrfs_tree_lock(left, BTRFS_NESTING_LEFT);
+		btrfs_tree_lock_nested(left, BTRFS_NESTING_LEFT);
 
 		left_nr = btrfs_header_nritems(left);
 		if (left_nr >= BTRFS_NODEPTRS_PER_BLOCK(fs_info) - 1) {
@@ -1265,7 +1286,7 @@ static noinline int push_nodes_for_insert(struct btrfs_trans_handle *trans,
 		if (IS_ERR(right))
 			return PTR_ERR(right);
 
-		__btrfs_tree_lock(right, BTRFS_NESTING_RIGHT);
+		btrfs_tree_lock_nested(right, BTRFS_NESTING_RIGHT);
 
 		right_nr = btrfs_header_nritems(right);
 		if (right_nr >= BTRFS_NODEPTRS_PER_BLOCK(fs_info) - 1) {
@@ -1511,7 +1532,7 @@ read_block_for_search(struct btrfs_root *root, struct btrfs_path *p,
 	check.has_first_key = true;
 	check.level = parent_level - 1;
 	check.transid = gen;
-	check.owner_root = root->root_key.objectid;
+	check.owner_root = btrfs_root_id(root);
 
 	/*
 	 * If we need to read an extent buffer from disk and we are holding locks
@@ -1556,7 +1577,7 @@ read_block_for_search(struct btrfs_root *root, struct btrfs_path *p,
 			btrfs_release_path(p);
 			return -EIO;
 		}
-		if (btrfs_check_eb_owner(tmp, root->root_key.objectid)) {
+		if (btrfs_check_eb_owner(tmp, btrfs_root_id(root))) {
 			free_extent_buffer(tmp);
 			btrfs_release_path(p);
 			return -EUCLEAN;
@@ -2865,7 +2886,7 @@ static noinline int insert_new_root(struct btrfs_trans_handle *trans,
 	else
 		btrfs_node_key(lower, &lower_key, 0);
 
-	c = btrfs_alloc_tree_block(trans, root, 0, root->root_key.objectid,
+	c = btrfs_alloc_tree_block(trans, root, 0, btrfs_root_id(root),
 				   &lower_key, level, root->node->start, 0,
 				   0, BTRFS_NESTING_NEW_ROOT);
 	if (IS_ERR(c))
@@ -2886,7 +2907,11 @@ static noinline int insert_new_root(struct btrfs_trans_handle *trans,
 	old = root->node;
 	ret = btrfs_tree_mod_log_insert_root(root->node, c, false);
 	if (ret < 0) {
-		btrfs_free_tree_block(trans, btrfs_root_id(root), c, 0, 1);
+		int ret2;
+
+		ret2 = btrfs_free_tree_block(trans, btrfs_root_id(root), c, 0, 1);
+		if (ret2 < 0)
+			btrfs_abort_transaction(trans, ret2);
 		btrfs_tree_unlock(c);
 		free_extent_buffer(c);
 		return ret;
@@ -3009,7 +3034,7 @@ static noinline int split_node(struct btrfs_trans_handle *trans,
 	mid = (c_nritems + 1) / 2;
 	btrfs_node_key(c, &disk_key, mid);
 
-	split = btrfs_alloc_tree_block(trans, root, 0, root->root_key.objectid,
+	split = btrfs_alloc_tree_block(trans, root, 0, btrfs_root_id(root),
 				       &disk_key, level, c->start, 0,
 				       0, BTRFS_NESTING_SPLIT);
 	if (IS_ERR(split))
@@ -3267,7 +3292,7 @@ static int push_leaf_right(struct btrfs_trans_handle *trans, struct btrfs_root
 	if (IS_ERR(right))
 		return PTR_ERR(right);
 
-	__btrfs_tree_lock(right, BTRFS_NESTING_RIGHT);
+	btrfs_tree_lock_nested(right, BTRFS_NESTING_RIGHT);
 
 	free_space = btrfs_leaf_free_space(right);
 	if (free_space < data_size)
@@ -3483,7 +3508,7 @@ static int push_leaf_left(struct btrfs_trans_handle *trans, struct btrfs_root
 	if (IS_ERR(left))
 		return PTR_ERR(left);
 
-	__btrfs_tree_lock(left, BTRFS_NESTING_LEFT);
+	btrfs_tree_lock_nested(left, BTRFS_NESTING_LEFT);
 
 	free_space = btrfs_leaf_free_space(left);
 	if (free_space < data_size) {
@@ -3761,7 +3786,7 @@ again:
 	 * BTRFS_NESTING_SPLIT_THE_SPLITTENING if we need to, but for now just
 	 * use BTRFS_NESTING_NEW_ROOT.
 	 */
-	right = btrfs_alloc_tree_block(trans, root, 0, root->root_key.objectid,
+	right = btrfs_alloc_tree_block(trans, root, 0, btrfs_root_id(root),
 				       &disk_key, 0, l->start, 0, 0,
 				       num_doubles ? BTRFS_NESTING_NEW_ROOT :
 				       BTRFS_NESTING_SPLIT);
@@ -4280,6 +4305,10 @@ void btrfs_setup_item_for_insert(struct btrfs_trans_handle *trans,
 /*
  * Given a key and some data, insert items into the tree.
  * This does all the path init required, making room in the tree if needed.
+ *
+ * Returns: 0        on success
+ *          -EEXIST  if the first key already exists
+ *          < 0      on other errors
  */
 int btrfs_insert_empty_items(struct btrfs_trans_handle *trans,
 			    struct btrfs_root *root,
@@ -4451,9 +4480,12 @@ static noinline int btrfs_del_leaf(struct btrfs_trans_handle *trans,
 	root_sub_used_bytes(root);
 
 	atomic_inc(&leaf->refs);
-	btrfs_free_tree_block(trans, btrfs_root_id(root), leaf, 0, 1);
+	ret = btrfs_free_tree_block(trans, btrfs_root_id(root), leaf, 0, 1);
 	free_extent_buffer_stale(leaf);
-	return 0;
+	if (ret < 0)
+		btrfs_abort_transaction(trans, ret);
+
+	return ret;
 }
 /*
  * delete the item at the leaf level in path.  If that empties
@@ -5082,9 +5114,7 @@ int btrfs_previous_extent_item(struct btrfs_root *root,
 
 int __init btrfs_ctree_init(void)
 {
-	btrfs_path_cachep = kmem_cache_create("btrfs_path",
-			sizeof(struct btrfs_path), 0,
-			SLAB_MEM_SPREAD, NULL);
+	btrfs_path_cachep = KMEM_CACHE(btrfs_path, 0);
 	if (!btrfs_path_cachep)
 		return -ENOMEM;
 	return 0;
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 70e828d331..a56209d275 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -7,25 +7,24 @@
 #define BTRFS_CTREE_H
 
 #include <linux/pagemap.h>
+#include <linux/spinlock.h>
+#include <linux/rbtree.h>
+#include <linux/mutex.h>
+#include <linux/wait.h>
+#include <linux/list.h>
+#include <linux/atomic.h>
+#include <linux/xarray.h>
+#include <linux/refcount.h>
+#include <uapi/linux/btrfs_tree.h>
 #include "locking.h"
 #include "fs.h"
 #include "accessors.h"
+#include "extent-io-tree.h"
 
+struct extent_buffer;
+struct btrfs_block_rsv;
 struct btrfs_trans_handle;
-struct btrfs_transaction;
-struct btrfs_pending_snapshot;
-struct btrfs_delayed_ref_root;
-struct btrfs_space_info;
 struct btrfs_block_group;
-struct btrfs_ordered_sum;
-struct btrfs_ref;
-struct btrfs_bio;
-struct btrfs_ioctl_encoded_io_args;
-struct btrfs_device;
-struct btrfs_fs_devices;
-struct btrfs_balance_control;
-struct btrfs_delayed_root;
-struct reloc_control;
 
 /* Read ahead values for struct btrfs_path.reada */
 enum {
@@ -355,6 +354,16 @@ static inline void btrfs_set_root_last_log_commit(struct btrfs_root *root, int c
 	WRITE_ONCE(root->last_log_commit, commit_id);
 }
 
+static inline u64 btrfs_get_root_last_trans(const struct btrfs_root *root)
+{
+	return READ_ONCE(root->last_trans);
+}
+
+static inline void btrfs_set_root_last_trans(struct btrfs_root *root, u64 transid)
+{
+	WRITE_ONCE(root->last_trans, transid);
+}
+
 /*
  * Structure that conveys information about an extent that is going to replace
  * all the extents in a file range.
@@ -448,6 +457,7 @@ struct btrfs_file_private {
 	void *filldir_buf;
 	u64 last_index;
 	struct extent_state *llseek_cached_state;
+	bool fsync_skip_inode_lock;
 };
 
 static inline u32 BTRFS_LEAF_DATA_SIZE(const struct btrfs_fs_info *info)
@@ -478,8 +488,7 @@ static inline gfp_t btrfs_alloc_write_mask(struct address_space *mapping)
 	return mapping_gfp_constraint(mapping, ~__GFP_FS);
 }
 
-int btrfs_error_unpin_extent_range(struct btrfs_fs_info *fs_info,
-				   u64 start, u64 end);
+void btrfs_error_unpin_extent_range(struct btrfs_fs_info *fs_info, u64 start, u64 end);
 int btrfs_discard_extent(struct btrfs_fs_info *fs_info, u64 bytenr,
 			 u64 num_bytes, u64 *actual_bytes);
 int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range);
diff --git a/fs/btrfs/defrag.c b/fs/btrfs/defrag.c
index a77be9896d..f664678c71 100644
--- a/fs/btrfs/defrag.c
+++ b/fs/btrfs/defrag.c
@@ -6,7 +6,6 @@
 #include <linux/sched.h>
 #include "ctree.h"
 #include "disk-io.h"
-#include "print-tree.h"
 #include "transaction.h"
 #include "locking.h"
 #include "accessors.h"
@@ -140,7 +139,7 @@ int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans,
 	if (trans)
 		transid = trans->transid;
 	else
-		transid = inode->root->last_trans;
+		transid = btrfs_get_root_last_trans(root);
 
 	defrag = kmem_cache_zalloc(btrfs_inode_defrag_cachep, GFP_NOFS);
 	if (!defrag)
@@ -148,7 +147,7 @@ int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans,
 
 	defrag->ino = btrfs_ino(inode);
 	defrag->transid = transid;
-	defrag->root = root->root_key.objectid;
+	defrag->root = btrfs_root_id(root);
 	defrag->extent_thresh = extent_thresh;
 
 	spin_lock(&fs_info->defrag_inodes_lock);
@@ -521,7 +520,7 @@ static int btrfs_defrag_leaves(struct btrfs_trans_handle *trans,
 	 * keep_locks set and lowest_level is 1, regardless of the value of
 	 * path->slots[1].
 	 */
-	BUG_ON(path->locks[1] == 0);
+	ASSERT(path->locks[1] != 0);
 	ret = btrfs_realloc_node(trans, root,
 				 path->nodes[1], 0,
 				 &last_ret,
@@ -861,20 +860,21 @@ out:
  * NOTE: Caller should also wait for page writeback after the cluster is
  * prepared, here we don't do writeback wait for each page.
  */
-static struct page *defrag_prepare_one_page(struct btrfs_inode *inode, pgoff_t index)
+static struct folio *defrag_prepare_one_folio(struct btrfs_inode *inode, pgoff_t index)
 {
 	struct address_space *mapping = inode->vfs_inode.i_mapping;
 	gfp_t mask = btrfs_alloc_write_mask(mapping);
 	u64 page_start = (u64)index << PAGE_SHIFT;
 	u64 page_end = page_start + PAGE_SIZE - 1;
 	struct extent_state *cached_state = NULL;
-	struct page *page;
+	struct folio *folio;
 	int ret;
 
 again:
-	page = find_or_create_page(mapping, index, mask);
-	if (!page)
-		return ERR_PTR(-ENOMEM);
+	folio = __filemap_get_folio(mapping, index,
+				    FGP_LOCK | FGP_ACCESSED | FGP_CREAT, mask);
+	if (IS_ERR(folio))
+		return folio;
 
 	/*
 	 * Since we can defragment files opened read-only, we can encounter
@@ -884,16 +884,16 @@ again:
 	 * executables that explicitly enable them, so this isn't very
 	 * restrictive.
 	 */
-	if (PageCompound(page)) {
-		unlock_page(page);
-		put_page(page);
+	if (folio_test_large(folio)) {
+		folio_unlock(folio);
+		folio_put(folio);
 		return ERR_PTR(-ETXTBSY);
 	}
 
-	ret = set_page_extent_mapped(page);
+	ret = set_folio_extent_mapped(folio);
 	if (ret < 0) {
-		unlock_page(page);
-		put_page(page);
+		folio_unlock(folio);
+		folio_put(folio);
 		return ERR_PTR(ret);
 	}
 
@@ -908,17 +908,17 @@ again:
 		if (!ordered)
 			break;
 
-		unlock_page(page);
+		folio_unlock(folio);
 		btrfs_start_ordered_extent(ordered);
 		btrfs_put_ordered_extent(ordered);
-		lock_page(page);
+		folio_lock(folio);
 		/*
-		 * We unlocked the page above, so we need check if it was
+		 * We unlocked the folio above, so we need check if it was
 		 * released or not.
 		 */
-		if (page->mapping != mapping || !PagePrivate(page)) {
-			unlock_page(page);
-			put_page(page);
+		if (folio->mapping != mapping || !folio->private) {
+			folio_unlock(folio);
+			folio_put(folio);
 			goto again;
 		}
 	}
@@ -927,21 +927,21 @@ again:
 	 * Now the page range has no ordered extent any more.  Read the page to
 	 * make it uptodate.
 	 */
-	if (!PageUptodate(page)) {
-		btrfs_read_folio(NULL, page_folio(page));
-		lock_page(page);
-		if (page->mapping != mapping || !PagePrivate(page)) {
-			unlock_page(page);
-			put_page(page);
+	if (!folio_test_uptodate(folio)) {
+		btrfs_read_folio(NULL, folio);
+		folio_lock(folio);
+		if (folio->mapping != mapping || !folio->private) {
+			folio_unlock(folio);
+			folio_put(folio);
 			goto again;
 		}
-		if (!PageUptodate(page)) {
-			unlock_page(page);
-			put_page(page);
+		if (!folio_test_uptodate(folio)) {
+			folio_unlock(folio);
+			folio_put(folio);
 			return ERR_PTR(-EIO);
 		}
 	}
-	return page;
+	return folio;
 }
 
 struct defrag_target_range {
@@ -1162,7 +1162,7 @@ static_assert(PAGE_ALIGNED(CLUSTER_SIZE));
  */
 static int defrag_one_locked_target(struct btrfs_inode *inode,
 				    struct defrag_target_range *target,
-				    struct page **pages, int nr_pages,
+				    struct folio **folios, int nr_pages,
 				    struct extent_state **cached_state)
 {
 	struct btrfs_fs_info *fs_info = inode->root->fs_info;
@@ -1171,7 +1171,7 @@ static int defrag_one_locked_target(struct btrfs_inode *inode,
 	const u64 len = target->len;
 	unsigned long last_index = (start + len - 1) >> PAGE_SHIFT;
 	unsigned long start_index = start >> PAGE_SHIFT;
-	unsigned long first_index = page_index(pages[0]);
+	unsigned long first_index = folios[0]->index;
 	int ret = 0;
 	int i;
 
@@ -1188,8 +1188,8 @@ static int defrag_one_locked_target(struct btrfs_inode *inode,
 
 	/* Update the page status */
 	for (i = start_index - first_index; i <= last_index - first_index; i++) {
-		ClearPageChecked(pages[i]);
-		btrfs_folio_clamp_set_dirty(fs_info, page_folio(pages[i]), start, len);
+		folio_clear_checked(folios[i]);
+		btrfs_folio_clamp_set_dirty(fs_info, folios[i], start, len);
 	}
 	btrfs_delalloc_release_extents(inode, len);
 	extent_changeset_free(data_reserved);
@@ -1205,7 +1205,7 @@ static int defrag_one_range(struct btrfs_inode *inode, u64 start, u32 len,
 	struct defrag_target_range *entry;
 	struct defrag_target_range *tmp;
 	LIST_HEAD(target_list);
-	struct page **pages;
+	struct folio **folios;
 	const u32 sectorsize = inode->root->fs_info->sectorsize;
 	u64 last_index = (start + len - 1) >> PAGE_SHIFT;
 	u64 start_index = start >> PAGE_SHIFT;
@@ -1216,21 +1216,21 @@ static int defrag_one_range(struct btrfs_inode *inode, u64 start, u32 len,
 	ASSERT(nr_pages <= CLUSTER_SIZE / PAGE_SIZE);
 	ASSERT(IS_ALIGNED(start, sectorsize) && IS_ALIGNED(len, sectorsize));
 
-	pages = kcalloc(nr_pages, sizeof(struct page *), GFP_NOFS);
-	if (!pages)
+	folios = kcalloc(nr_pages, sizeof(struct folio *), GFP_NOFS);
+	if (!folios)
 		return -ENOMEM;
 
 	/* Prepare all pages */
 	for (i = 0; i < nr_pages; i++) {
-		pages[i] = defrag_prepare_one_page(inode, start_index + i);
-		if (IS_ERR(pages[i])) {
-			ret = PTR_ERR(pages[i]);
-			pages[i] = NULL;
-			goto free_pages;
+		folios[i] = defrag_prepare_one_folio(inode, start_index + i);
+		if (IS_ERR(folios[i])) {
+			ret = PTR_ERR(folios[i]);
+			nr_pages = i;
+			goto free_folios;
 		}
 	}
 	for (i = 0; i < nr_pages; i++)
-		wait_on_page_writeback(pages[i]);
+		folio_wait_writeback(folios[i]);
 
 	/* Lock the pages range */
 	lock_extent(&inode->io_tree, start_index << PAGE_SHIFT,
@@ -1250,7 +1250,7 @@ static int defrag_one_range(struct btrfs_inode *inode, u64 start, u32 len,
 		goto unlock_extent;
 
 	list_for_each_entry(entry, &target_list, list) {
-		ret = defrag_one_locked_target(inode, entry, pages, nr_pages,
+		ret = defrag_one_locked_target(inode, entry, folios, nr_pages,
 					       &cached_state);
 		if (ret < 0)
 			break;
@@ -1264,14 +1264,12 @@ unlock_extent:
 	unlock_extent(&inode->io_tree, start_index << PAGE_SHIFT,
 		      (last_index << PAGE_SHIFT) + PAGE_SIZE - 1,
 		      &cached_state);
-free_pages:
+free_folios:
 	for (i = 0; i < nr_pages; i++) {
-		if (pages[i]) {
-			unlock_page(pages[i]);
-			put_page(pages[i]);
-		}
+		folio_unlock(folios[i]);
+		folio_put(folios[i]);
 	}
-	kfree(pages);
+	kfree(folios);
 	return ret;
 }
 
@@ -1512,9 +1510,7 @@ void __cold btrfs_auto_defrag_exit(void)
 int __init btrfs_auto_defrag_init(void)
 {
 	btrfs_inode_defrag_cachep = kmem_cache_create("btrfs_inode_defrag",
-					sizeof(struct inode_defrag), 0,
-					SLAB_MEM_SPREAD,
-					NULL);
+					sizeof(struct inode_defrag), 0, 0, NULL);
 	if (!btrfs_inode_defrag_cachep)
 		return -ENOMEM;
 
diff --git a/fs/btrfs/defrag.h b/fs/btrfs/defrag.h
index 5a62763528..878528e086 100644
--- a/fs/btrfs/defrag.h
+++ b/fs/btrfs/defrag.h
@@ -3,6 +3,16 @@
 #ifndef BTRFS_DEFRAG_H
 #define BTRFS_DEFRAG_H
 
+#include <linux/types.h>
+#include <linux/compiler_types.h>
+
+struct inode;
+struct file_ra_state;
+struct btrfs_fs_info;
+struct btrfs_root;
+struct btrfs_trans_handle;
+struct btrfs_ioctl_defrag_range_args;
+
 int btrfs_defrag_file(struct inode *inode, struct file_ra_state *ra,
 		      struct btrfs_ioctl_defrag_range_args *range,
 		      u64 newer_than, unsigned long max_to_defrag);
diff --git a/fs/btrfs/delalloc-space.c b/fs/btrfs/delalloc-space.c
index acf9f4b6c0..b3527efd0b 100644
--- a/fs/btrfs/delalloc-space.c
+++ b/fs/btrfs/delalloc-space.c
@@ -6,9 +6,7 @@
 #include "block-rsv.h"
 #include "btrfs_inode.h"
 #include "space-info.h"
-#include "transaction.h"
 #include "qgroup.h"
-#include "block-group.h"
 #include "fs.h"
 
 /*
diff --git a/fs/btrfs/delalloc-space.h b/fs/btrfs/delalloc-space.h
index c5d573f236..ce4f889e4f 100644
--- a/fs/btrfs/delalloc-space.h
+++ b/fs/btrfs/delalloc-space.h
@@ -3,7 +3,11 @@
 #ifndef BTRFS_DELALLOC_SPACE_H
 #define BTRFS_DELALLOC_SPACE_H
 
+#include <linux/types.h>
+
 struct extent_changeset;
+struct btrfs_inode;
+struct btrfs_fs_info;
 
 int btrfs_alloc_data_chunk_ondemand(struct btrfs_inode *inode, u64 bytes);
 int btrfs_check_data_free_space(struct btrfs_inode *inode,
diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c
index ab5a833e7d..95a0497fa8 100644
--- a/fs/btrfs/delayed-inode.c
+++ b/fs/btrfs/delayed-inode.c
@@ -28,11 +28,7 @@ static struct kmem_cache *delayed_node_cache;
 
 int __init btrfs_delayed_inode_init(void)
 {
-	delayed_node_cache = kmem_cache_create("btrfs_delayed_node",
-					sizeof(struct btrfs_delayed_node),
-					0,
-					SLAB_MEM_SPREAD,
-					NULL);
+	delayed_node_cache = KMEM_CACHE(btrfs_delayed_node, 0);
 	if (!delayed_node_cache)
 		return -ENOMEM;
 	return 0;
@@ -43,6 +39,17 @@ void __cold btrfs_delayed_inode_exit(void)
 	kmem_cache_destroy(delayed_node_cache);
 }
 
+void btrfs_init_delayed_root(struct btrfs_delayed_root *delayed_root)
+{
+	atomic_set(&delayed_root->items, 0);
+	atomic_set(&delayed_root->items_seq, 0);
+	delayed_root->nodes = 0;
+	spin_lock_init(&delayed_root->lock);
+	init_waitqueue_head(&delayed_root->wait);
+	INIT_LIST_HEAD(&delayed_root->node_list);
+	INIT_LIST_HEAD(&delayed_root->prepare_list);
+}
+
 static inline void btrfs_init_delayed_node(
 				struct btrfs_delayed_node *delayed_node,
 				struct btrfs_root *root, u64 inode_id)
@@ -430,8 +437,6 @@ static void __btrfs_remove_delayed_item(struct btrfs_delayed_item *delayed_item)
 
 	delayed_root = delayed_node->root->fs_info->delayed_root;
 
-	BUG_ON(!delayed_root);
-
 	if (delayed_item->type == BTRFS_DELAYED_INSERTION_ITEM)
 		root = &delayed_node->ins_root;
 	else
@@ -980,7 +985,7 @@ static void btrfs_release_delayed_inode(struct btrfs_delayed_node *delayed_node)
 
 	if (delayed_node &&
 	    test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) {
-		BUG_ON(!delayed_node->root);
+		ASSERT(delayed_node->root);
 		clear_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags);
 		delayed_node->count--;
 
@@ -1646,7 +1651,7 @@ int btrfs_delete_delayed_dir_index(struct btrfs_trans_handle *trans,
 	if (unlikely(ret)) {
 		btrfs_err(trans->fs_info,
 			  "err add delayed dir index item(index: %llu) into the deletion tree of the delayed node(root id: %llu, inode id: %llu, errno: %d)",
-			  index, node->root->root_key.objectid,
+			  index, btrfs_root_id(node->root),
 			  node->inode_id, ret);
 		btrfs_delayed_item_release_metadata(dir->root, item);
 		btrfs_release_delayed_item(item);
diff --git a/fs/btrfs/delayed-inode.h b/fs/btrfs/delayed-inode.h
index 5cceb31bbd..64e115d974 100644
--- a/fs/btrfs/delayed-inode.h
+++ b/fs/btrfs/delayed-inode.h
@@ -7,15 +7,23 @@
 #ifndef BTRFS_DELAYED_INODE_H
 #define BTRFS_DELAYED_INODE_H
 
+#include <linux/types.h>
 #include <linux/rbtree.h>
 #include <linux/spinlock.h>
 #include <linux/mutex.h>
 #include <linux/list.h>
 #include <linux/wait.h>
+#include <linux/fs.h>
 #include <linux/atomic.h>
 #include <linux/refcount.h>
 #include "ctree.h"
 
+struct btrfs_disk_key;
+struct btrfs_fs_info;
+struct btrfs_inode;
+struct btrfs_root;
+struct btrfs_trans_handle;
+
 enum btrfs_delayed_item_type {
 	BTRFS_DELAYED_INSERTION_ITEM,
 	BTRFS_DELAYED_DELETION_ITEM
@@ -98,18 +106,7 @@ struct btrfs_delayed_item {
 	char data[] __counted_by(data_len);
 };
 
-static inline void btrfs_init_delayed_root(
-				struct btrfs_delayed_root *delayed_root)
-{
-	atomic_set(&delayed_root->items, 0);
-	atomic_set(&delayed_root->items_seq, 0);
-	delayed_root->nodes = 0;
-	spin_lock_init(&delayed_root->lock);
-	init_waitqueue_head(&delayed_root->wait);
-	INIT_LIST_HEAD(&delayed_root->node_list);
-	INIT_LIST_HEAD(&delayed_root->prepare_list);
-}
-
+void btrfs_init_delayed_root(struct btrfs_delayed_root *delayed_root);
 int btrfs_insert_delayed_dir_index(struct btrfs_trans_handle *trans,
 				   const char *name, int name_len,
 				   struct btrfs_inode *dir,
diff --git a/fs/btrfs/delayed-ref.c b/fs/btrfs/delayed-ref.c
index 891ea2fa26..6cc80fb10d 100644
--- a/fs/btrfs/delayed-ref.c
+++ b/fs/btrfs/delayed-ref.c
@@ -16,8 +16,7 @@
 #include "fs.h"
 
 struct kmem_cache *btrfs_delayed_ref_head_cachep;
-struct kmem_cache *btrfs_delayed_tree_ref_cachep;
-struct kmem_cache *btrfs_delayed_data_ref_cachep;
+struct kmem_cache *btrfs_delayed_ref_node_cachep;
 struct kmem_cache *btrfs_delayed_extent_op_cachep;
 /*
  * delayed back reference update tracking.  For subvolume trees
@@ -305,50 +304,19 @@ int btrfs_delayed_refs_rsv_refill(struct btrfs_fs_info *fs_info,
 }
 
 /*
- * compare two delayed tree backrefs with same bytenr and type
- */
-static int comp_tree_refs(struct btrfs_delayed_tree_ref *ref1,
-			  struct btrfs_delayed_tree_ref *ref2)
-{
-	if (ref1->node.type == BTRFS_TREE_BLOCK_REF_KEY) {
-		if (ref1->root < ref2->root)
-			return -1;
-		if (ref1->root > ref2->root)
-			return 1;
-	} else {
-		if (ref1->parent < ref2->parent)
-			return -1;
-		if (ref1->parent > ref2->parent)
-			return 1;
-	}
-	return 0;
-}
-
-/*
  * compare two delayed data backrefs with same bytenr and type
  */
-static int comp_data_refs(struct btrfs_delayed_data_ref *ref1,
-			  struct btrfs_delayed_data_ref *ref2)
+static int comp_data_refs(struct btrfs_delayed_ref_node *ref1,
+			  struct btrfs_delayed_ref_node *ref2)
 {
-	if (ref1->node.type == BTRFS_EXTENT_DATA_REF_KEY) {
-		if (ref1->root < ref2->root)
-			return -1;
-		if (ref1->root > ref2->root)
-			return 1;
-		if (ref1->objectid < ref2->objectid)
-			return -1;
-		if (ref1->objectid > ref2->objectid)
-			return 1;
-		if (ref1->offset < ref2->offset)
-			return -1;
-		if (ref1->offset > ref2->offset)
-			return 1;
-	} else {
-		if (ref1->parent < ref2->parent)
-			return -1;
-		if (ref1->parent > ref2->parent)
-			return 1;
-	}
+	if (ref1->data_ref.objectid < ref2->data_ref.objectid)
+		return -1;
+	if (ref1->data_ref.objectid > ref2->data_ref.objectid)
+		return 1;
+	if (ref1->data_ref.offset < ref2->data_ref.offset)
+		return -1;
+	if (ref1->data_ref.offset > ref2->data_ref.offset)
+		return 1;
 	return 0;
 }
 
@@ -362,13 +330,20 @@ static int comp_refs(struct btrfs_delayed_ref_node *ref1,
 		return -1;
 	if (ref1->type > ref2->type)
 		return 1;
-	if (ref1->type == BTRFS_TREE_BLOCK_REF_KEY ||
-	    ref1->type == BTRFS_SHARED_BLOCK_REF_KEY)
-		ret = comp_tree_refs(btrfs_delayed_node_to_tree_ref(ref1),
-				     btrfs_delayed_node_to_tree_ref(ref2));
-	else
-		ret = comp_data_refs(btrfs_delayed_node_to_data_ref(ref1),
-				     btrfs_delayed_node_to_data_ref(ref2));
+	if (ref1->type == BTRFS_SHARED_BLOCK_REF_KEY ||
+	    ref1->type == BTRFS_SHARED_DATA_REF_KEY) {
+		if (ref1->parent < ref2->parent)
+			return -1;
+		if (ref1->parent > ref2->parent)
+			return 1;
+	} else {
+		if (ref1->ref_root < ref2->ref_root)
+			return -1;
+		if (ref1->ref_root > ref2->ref_root)
+			return -1;
+		if (ref1->type == BTRFS_EXTENT_DATA_REF_KEY)
+			ret = comp_data_refs(ref1, ref2);
+	}
 	if (ret)
 		return ret;
 	if (check_seq) {
@@ -828,18 +803,20 @@ static noinline void update_existing_head_ref(struct btrfs_trans_handle *trans,
 }
 
 static void init_delayed_ref_head(struct btrfs_delayed_ref_head *head_ref,
+				  struct btrfs_ref *generic_ref,
 				  struct btrfs_qgroup_extent_record *qrecord,
-				  u64 bytenr, u64 num_bytes, u64 ref_root,
-				  u64 reserved, int action, bool is_data,
-				  bool is_system, u64 owning_root)
+				  u64 reserved)
 {
 	int count_mod = 1;
 	bool must_insert_reserved = false;
 
 	/* If reserved is provided, it must be a data extent. */
-	BUG_ON(!is_data && reserved);
+	BUG_ON(generic_ref->type != BTRFS_REF_DATA && reserved);
 
-	switch (action) {
+	switch (generic_ref->action) {
+	case BTRFS_ADD_DELAYED_REF:
+		/* count_mod is already set to 1. */
+		break;
 	case BTRFS_UPDATE_DELAYED_HEAD:
 		count_mod = 0;
 		break;
@@ -868,14 +845,14 @@ static void init_delayed_ref_head(struct btrfs_delayed_ref_head *head_ref,
 	}
 
 	refcount_set(&head_ref->refs, 1);
-	head_ref->bytenr = bytenr;
-	head_ref->num_bytes = num_bytes;
+	head_ref->bytenr = generic_ref->bytenr;
+	head_ref->num_bytes = generic_ref->num_bytes;
 	head_ref->ref_mod = count_mod;
 	head_ref->reserved_bytes = reserved;
 	head_ref->must_insert_reserved = must_insert_reserved;
-	head_ref->owning_root = owning_root;
-	head_ref->is_data = is_data;
-	head_ref->is_system = is_system;
+	head_ref->owning_root = generic_ref->owning_root;
+	head_ref->is_data = (generic_ref->type == BTRFS_REF_DATA);
+	head_ref->is_system = (generic_ref->ref_root == BTRFS_CHUNK_TREE_OBJECTID);
 	head_ref->ref_tree = RB_ROOT_CACHED;
 	INIT_LIST_HEAD(&head_ref->ref_add_list);
 	RB_CLEAR_NODE(&head_ref->href_node);
@@ -885,12 +862,12 @@ static void init_delayed_ref_head(struct btrfs_delayed_ref_head *head_ref,
 	mutex_init(&head_ref->mutex);
 
 	if (qrecord) {
-		if (ref_root && reserved) {
+		if (generic_ref->ref_root && reserved) {
 			qrecord->data_rsv = reserved;
-			qrecord->data_rsv_refroot = ref_root;
+			qrecord->data_rsv_refroot = generic_ref->ref_root;
 		}
-		qrecord->bytenr = bytenr;
-		qrecord->num_bytes = num_bytes;
+		qrecord->bytenr = generic_ref->bytenr;
+		qrecord->num_bytes = generic_ref->num_bytes;
 		qrecord->old_roots = NULL;
 	}
 }
@@ -982,89 +959,104 @@ add_delayed_ref_head(struct btrfs_trans_handle *trans,
  */
 static void init_delayed_ref_common(struct btrfs_fs_info *fs_info,
 				    struct btrfs_delayed_ref_node *ref,
-				    u64 bytenr, u64 num_bytes, u64 ref_root,
-				    int action, u8 ref_type)
+				    struct btrfs_ref *generic_ref)
 {
+	int action = generic_ref->action;
 	u64 seq = 0;
 
 	if (action == BTRFS_ADD_DELAYED_EXTENT)
 		action = BTRFS_ADD_DELAYED_REF;
 
-	if (is_fstree(ref_root))
+	if (is_fstree(generic_ref->ref_root))
 		seq = atomic64_read(&fs_info->tree_mod_seq);
 
 	refcount_set(&ref->refs, 1);
-	ref->bytenr = bytenr;
-	ref->num_bytes = num_bytes;
+	ref->bytenr = generic_ref->bytenr;
+	ref->num_bytes = generic_ref->num_bytes;
 	ref->ref_mod = 1;
 	ref->action = action;
 	ref->seq = seq;
-	ref->type = ref_type;
+	ref->type = btrfs_ref_type(generic_ref);
+	ref->ref_root = generic_ref->ref_root;
+	ref->parent = generic_ref->parent;
 	RB_CLEAR_NODE(&ref->ref_node);
 	INIT_LIST_HEAD(&ref->add_list);
+
+	if (generic_ref->type == BTRFS_REF_DATA)
+		ref->data_ref = generic_ref->data_ref;
+	else
+		ref->tree_ref = generic_ref->tree_ref;
 }
 
-/*
- * add a delayed tree ref.  This does all of the accounting required
- * to make sure the delayed ref is eventually processed before this
- * transaction commits.
- */
-int btrfs_add_delayed_tree_ref(struct btrfs_trans_handle *trans,
-			       struct btrfs_ref *generic_ref,
-			       struct btrfs_delayed_extent_op *extent_op)
+void btrfs_init_tree_ref(struct btrfs_ref *generic_ref, int level, u64 mod_root,
+			 bool skip_qgroup)
+{
+#ifdef CONFIG_BTRFS_FS_REF_VERIFY
+	/* If @real_root not set, use @root as fallback */
+	generic_ref->real_root = mod_root ?: generic_ref->ref_root;
+#endif
+	generic_ref->tree_ref.level = level;
+	generic_ref->type = BTRFS_REF_METADATA;
+	if (skip_qgroup || !(is_fstree(generic_ref->ref_root) &&
+			     (!mod_root || is_fstree(mod_root))))
+		generic_ref->skip_qgroup = true;
+	else
+		generic_ref->skip_qgroup = false;
+
+}
+
+void btrfs_init_data_ref(struct btrfs_ref *generic_ref, u64 ino, u64 offset,
+			 u64 mod_root, bool skip_qgroup)
+{
+#ifdef CONFIG_BTRFS_FS_REF_VERIFY
+	/* If @real_root not set, use @root as fallback */
+	generic_ref->real_root = mod_root ?: generic_ref->ref_root;
+#endif
+	generic_ref->data_ref.objectid = ino;
+	generic_ref->data_ref.offset = offset;
+	generic_ref->type = BTRFS_REF_DATA;
+	if (skip_qgroup || !(is_fstree(generic_ref->ref_root) &&
+			     (!mod_root || is_fstree(mod_root))))
+		generic_ref->skip_qgroup = true;
+	else
+		generic_ref->skip_qgroup = false;
+}
+
+static int add_delayed_ref(struct btrfs_trans_handle *trans,
+			   struct btrfs_ref *generic_ref,
+			   struct btrfs_delayed_extent_op *extent_op,
+			   u64 reserved)
 {
 	struct btrfs_fs_info *fs_info = trans->fs_info;
-	struct btrfs_delayed_tree_ref *ref;
+	struct btrfs_delayed_ref_node *node;
 	struct btrfs_delayed_ref_head *head_ref;
 	struct btrfs_delayed_ref_root *delayed_refs;
 	struct btrfs_qgroup_extent_record *record = NULL;
 	bool qrecord_inserted;
-	bool is_system;
-	bool merged;
 	int action = generic_ref->action;
-	int level = generic_ref->tree_ref.level;
-	u64 bytenr = generic_ref->bytenr;
-	u64 num_bytes = generic_ref->len;
-	u64 parent = generic_ref->parent;
-	u8 ref_type;
-
-	is_system = (generic_ref->tree_ref.ref_root == BTRFS_CHUNK_TREE_OBJECTID);
+	bool merged;
 
-	ASSERT(generic_ref->type == BTRFS_REF_METADATA && generic_ref->action);
-	ref = kmem_cache_alloc(btrfs_delayed_tree_ref_cachep, GFP_NOFS);
-	if (!ref)
+	node = kmem_cache_alloc(btrfs_delayed_ref_node_cachep, GFP_NOFS);
+	if (!node)
 		return -ENOMEM;
 
 	head_ref = kmem_cache_alloc(btrfs_delayed_ref_head_cachep, GFP_NOFS);
 	if (!head_ref) {
-		kmem_cache_free(btrfs_delayed_tree_ref_cachep, ref);
+		kmem_cache_free(btrfs_delayed_ref_node_cachep, node);
 		return -ENOMEM;
 	}
 
 	if (btrfs_qgroup_full_accounting(fs_info) && !generic_ref->skip_qgroup) {
 		record = kzalloc(sizeof(*record), GFP_NOFS);
 		if (!record) {
-			kmem_cache_free(btrfs_delayed_tree_ref_cachep, ref);
+			kmem_cache_free(btrfs_delayed_ref_node_cachep, node);
 			kmem_cache_free(btrfs_delayed_ref_head_cachep, head_ref);
 			return -ENOMEM;
 		}
 	}
 
-	if (parent)
-		ref_type = BTRFS_SHARED_BLOCK_REF_KEY;
-	else
-		ref_type = BTRFS_TREE_BLOCK_REF_KEY;
-
-	init_delayed_ref_common(fs_info, &ref->node, bytenr, num_bytes,
-				generic_ref->tree_ref.ref_root, action,
-				ref_type);
-	ref->root = generic_ref->tree_ref.ref_root;
-	ref->parent = parent;
-	ref->level = level;
-
-	init_delayed_ref_head(head_ref, record, bytenr, num_bytes,
-			      generic_ref->tree_ref.ref_root, 0, action,
-			      false, is_system, generic_ref->owning_root);
+	init_delayed_ref_common(fs_info, node, generic_ref);
+	init_delayed_ref_head(head_ref, generic_ref, record, reserved);
 	head_ref->extent_op = extent_op;
 
 	delayed_refs = &trans->transaction->delayed_refs;
@@ -1077,7 +1069,7 @@ int btrfs_add_delayed_tree_ref(struct btrfs_trans_handle *trans,
 	head_ref = add_delayed_ref_head(trans, head_ref, record,
 					action, &qrecord_inserted);
 
-	merged = insert_delayed_ref(trans, head_ref, &ref->node);
+	merged = insert_delayed_ref(trans, head_ref, node);
 	spin_unlock(&delayed_refs->lock);
 
 	/*
@@ -1086,107 +1078,39 @@ int btrfs_add_delayed_tree_ref(struct btrfs_trans_handle *trans,
 	 */
 	btrfs_update_delayed_refs_rsv(trans);
 
-	trace_add_delayed_tree_ref(fs_info, &ref->node, ref,
-				   action == BTRFS_ADD_DELAYED_EXTENT ?
-				   BTRFS_ADD_DELAYED_REF : action);
+	if (generic_ref->type == BTRFS_REF_DATA)
+		trace_add_delayed_data_ref(trans->fs_info, node);
+	else
+		trace_add_delayed_tree_ref(trans->fs_info, node);
 	if (merged)
-		kmem_cache_free(btrfs_delayed_tree_ref_cachep, ref);
+		kmem_cache_free(btrfs_delayed_ref_node_cachep, node);
 
 	if (qrecord_inserted)
-		btrfs_qgroup_trace_extent_post(trans, record);
-
+		return btrfs_qgroup_trace_extent_post(trans, record);
 	return 0;
 }
 
 /*
+ * Add a delayed tree ref. This does all of the accounting required to make sure
+ * the delayed ref is eventually processed before this transaction commits.
+ */
+int btrfs_add_delayed_tree_ref(struct btrfs_trans_handle *trans,
+			       struct btrfs_ref *generic_ref,
+			       struct btrfs_delayed_extent_op *extent_op)
+{
+	ASSERT(generic_ref->type == BTRFS_REF_METADATA && generic_ref->action);
+	return add_delayed_ref(trans, generic_ref, extent_op, 0);
+}
+
+/*
  * add a delayed data ref. it's similar to btrfs_add_delayed_tree_ref.
  */
 int btrfs_add_delayed_data_ref(struct btrfs_trans_handle *trans,
 			       struct btrfs_ref *generic_ref,
 			       u64 reserved)
 {
-	struct btrfs_fs_info *fs_info = trans->fs_info;
-	struct btrfs_delayed_data_ref *ref;
-	struct btrfs_delayed_ref_head *head_ref;
-	struct btrfs_delayed_ref_root *delayed_refs;
-	struct btrfs_qgroup_extent_record *record = NULL;
-	bool qrecord_inserted;
-	int action = generic_ref->action;
-	bool merged;
-	u64 bytenr = generic_ref->bytenr;
-	u64 num_bytes = generic_ref->len;
-	u64 parent = generic_ref->parent;
-	u64 ref_root = generic_ref->data_ref.ref_root;
-	u64 owner = generic_ref->data_ref.ino;
-	u64 offset = generic_ref->data_ref.offset;
-	u8 ref_type;
-
-	ASSERT(generic_ref->type == BTRFS_REF_DATA && action);
-	ref = kmem_cache_alloc(btrfs_delayed_data_ref_cachep, GFP_NOFS);
-	if (!ref)
-		return -ENOMEM;
-
-	if (parent)
-	        ref_type = BTRFS_SHARED_DATA_REF_KEY;
-	else
-	        ref_type = BTRFS_EXTENT_DATA_REF_KEY;
-	init_delayed_ref_common(fs_info, &ref->node, bytenr, num_bytes,
-				ref_root, action, ref_type);
-	ref->root = ref_root;
-	ref->parent = parent;
-	ref->objectid = owner;
-	ref->offset = offset;
-
-
-	head_ref = kmem_cache_alloc(btrfs_delayed_ref_head_cachep, GFP_NOFS);
-	if (!head_ref) {
-		kmem_cache_free(btrfs_delayed_data_ref_cachep, ref);
-		return -ENOMEM;
-	}
-
-	if (btrfs_qgroup_full_accounting(fs_info) && !generic_ref->skip_qgroup) {
-		record = kzalloc(sizeof(*record), GFP_NOFS);
-		if (!record) {
-			kmem_cache_free(btrfs_delayed_data_ref_cachep, ref);
-			kmem_cache_free(btrfs_delayed_ref_head_cachep,
-					head_ref);
-			return -ENOMEM;
-		}
-	}
-
-	init_delayed_ref_head(head_ref, record, bytenr, num_bytes, ref_root,
-			      reserved, action, true, false, generic_ref->owning_root);
-	head_ref->extent_op = NULL;
-
-	delayed_refs = &trans->transaction->delayed_refs;
-	spin_lock(&delayed_refs->lock);
-
-	/*
-	 * insert both the head node and the new ref without dropping
-	 * the spin lock
-	 */
-	head_ref = add_delayed_ref_head(trans, head_ref, record,
-					action, &qrecord_inserted);
-
-	merged = insert_delayed_ref(trans, head_ref, &ref->node);
-	spin_unlock(&delayed_refs->lock);
-
-	/*
-	 * Need to update the delayed_refs_rsv with any changes we may have
-	 * made.
-	 */
-	btrfs_update_delayed_refs_rsv(trans);
-
-	trace_add_delayed_data_ref(trans->fs_info, &ref->node, ref,
-				   action == BTRFS_ADD_DELAYED_EXTENT ?
-				   BTRFS_ADD_DELAYED_REF : action);
-	if (merged)
-		kmem_cache_free(btrfs_delayed_data_ref_cachep, ref);
-
-
-	if (qrecord_inserted)
-		return btrfs_qgroup_trace_extent_post(trans, record);
-	return 0;
+	ASSERT(generic_ref->type == BTRFS_REF_DATA && generic_ref->action);
+	return add_delayed_ref(trans, generic_ref, NULL, reserved);
 }
 
 int btrfs_add_delayed_extent_op(struct btrfs_trans_handle *trans,
@@ -1195,13 +1119,18 @@ int btrfs_add_delayed_extent_op(struct btrfs_trans_handle *trans,
 {
 	struct btrfs_delayed_ref_head *head_ref;
 	struct btrfs_delayed_ref_root *delayed_refs;
+	struct btrfs_ref generic_ref = {
+		.type = BTRFS_REF_METADATA,
+		.action = BTRFS_UPDATE_DELAYED_HEAD,
+		.bytenr = bytenr,
+		.num_bytes = num_bytes,
+	};
 
 	head_ref = kmem_cache_alloc(btrfs_delayed_ref_head_cachep, GFP_NOFS);
 	if (!head_ref)
 		return -ENOMEM;
 
-	init_delayed_ref_head(head_ref, NULL, bytenr, num_bytes, 0, 0,
-			      BTRFS_UPDATE_DELAYED_HEAD, false, false, 0);
+	init_delayed_ref_head(head_ref, &generic_ref, NULL, 0);
 	head_ref->extent_op = extent_op;
 
 	delayed_refs = &trans->transaction->delayed_refs;
@@ -1220,6 +1149,14 @@ int btrfs_add_delayed_extent_op(struct btrfs_trans_handle *trans,
 	return 0;
 }
 
+void btrfs_put_delayed_ref(struct btrfs_delayed_ref_node *ref)
+{
+	if (refcount_dec_and_test(&ref->refs)) {
+		WARN_ON(!RB_EMPTY_NODE(&ref->ref_node));
+		kmem_cache_free(btrfs_delayed_ref_node_cachep, ref);
+	}
+}
+
 /*
  * This does a simple search for the head node for a given extent.  Returns the
  * head node if found, or NULL if not.
@@ -1235,38 +1172,21 @@ btrfs_find_delayed_ref_head(struct btrfs_delayed_ref_root *delayed_refs, u64 byt
 void __cold btrfs_delayed_ref_exit(void)
 {
 	kmem_cache_destroy(btrfs_delayed_ref_head_cachep);
-	kmem_cache_destroy(btrfs_delayed_tree_ref_cachep);
-	kmem_cache_destroy(btrfs_delayed_data_ref_cachep);
+	kmem_cache_destroy(btrfs_delayed_ref_node_cachep);
 	kmem_cache_destroy(btrfs_delayed_extent_op_cachep);
 }
 
 int __init btrfs_delayed_ref_init(void)
 {
-	btrfs_delayed_ref_head_cachep = kmem_cache_create(
-				"btrfs_delayed_ref_head",
-				sizeof(struct btrfs_delayed_ref_head), 0,
-				SLAB_MEM_SPREAD, NULL);
+	btrfs_delayed_ref_head_cachep = KMEM_CACHE(btrfs_delayed_ref_head, 0);
 	if (!btrfs_delayed_ref_head_cachep)
 		goto fail;
 
-	btrfs_delayed_tree_ref_cachep = kmem_cache_create(
-				"btrfs_delayed_tree_ref",
-				sizeof(struct btrfs_delayed_tree_ref), 0,
-				SLAB_MEM_SPREAD, NULL);
-	if (!btrfs_delayed_tree_ref_cachep)
-		goto fail;
-
-	btrfs_delayed_data_ref_cachep = kmem_cache_create(
-				"btrfs_delayed_data_ref",
-				sizeof(struct btrfs_delayed_data_ref), 0,
-				SLAB_MEM_SPREAD, NULL);
-	if (!btrfs_delayed_data_ref_cachep)
+	btrfs_delayed_ref_node_cachep = KMEM_CACHE(btrfs_delayed_ref_node, 0);
+	if (!btrfs_delayed_ref_node_cachep)
 		goto fail;
 
-	btrfs_delayed_extent_op_cachep = kmem_cache_create(
-				"btrfs_delayed_extent_op",
-				sizeof(struct btrfs_delayed_extent_op), 0,
-				SLAB_MEM_SPREAD, NULL);
+	btrfs_delayed_extent_op_cachep = KMEM_CACHE(btrfs_delayed_extent_op, 0);
 	if (!btrfs_delayed_extent_op_cachep)
 		goto fail;
 
diff --git a/fs/btrfs/delayed-ref.h b/fs/btrfs/delayed-ref.h
index 62d679d40f..04b180ebe1 100644
--- a/fs/btrfs/delayed-ref.h
+++ b/fs/btrfs/delayed-ref.h
@@ -6,7 +6,17 @@
 #ifndef BTRFS_DELAYED_REF_H
 #define BTRFS_DELAYED_REF_H
 
+#include <linux/types.h>
 #include <linux/refcount.h>
+#include <linux/list.h>
+#include <linux/rbtree.h>
+#include <linux/mutex.h>
+#include <linux/spinlock.h>
+#include <linux/slab.h>
+#include <uapi/linux/btrfs_tree.h>
+
+struct btrfs_trans_handle;
+struct btrfs_fs_info;
 
 /* these are the possible values of struct btrfs_delayed_ref_node->action */
 enum btrfs_delayed_ref_action {
@@ -20,6 +30,32 @@ enum btrfs_delayed_ref_action {
 	BTRFS_UPDATE_DELAYED_HEAD,
 } __packed;
 
+struct btrfs_data_ref {
+	/* For EXTENT_DATA_REF */
+
+	/* Inode which refers to this data extent */
+	u64 objectid;
+
+	/*
+	 * file_offset - extent_offset
+	 *
+	 * file_offset is the key.offset of the EXTENT_DATA key.
+	 * extent_offset is btrfs_file_extent_offset() of the EXTENT_DATA data.
+	 */
+	u64 offset;
+};
+
+struct btrfs_tree_ref {
+	/*
+	 * Level of this tree block.
+	 *
+	 * Shared for skinny (TREE_BLOCK_REF) and normal tree ref.
+	 */
+	int level;
+
+	/* For non-skinny metadata, no special member needed */
+};
+
 struct btrfs_delayed_ref_node {
 	struct rb_node ref_node;
 	/*
@@ -38,6 +74,15 @@ struct btrfs_delayed_ref_node {
 	/* seq number to keep track of insertion order */
 	u64 seq;
 
+	/* The ref_root for this ref */
+	u64 ref_root;
+
+	/*
+	 * The parent for this ref, if this isn't set the ref_root is the
+	 * reference owner.
+	 */
+	u64 parent;
+
 	/* ref count on this data structure */
 	refcount_t refs;
 
@@ -54,6 +99,11 @@ struct btrfs_delayed_ref_node {
 
 	unsigned int action:8;
 	unsigned int type:8;
+
+	union {
+		struct btrfs_tree_ref tree_ref;
+		struct btrfs_data_ref data_ref;
+	};
 };
 
 struct btrfs_delayed_extent_op {
@@ -141,21 +191,6 @@ struct btrfs_delayed_ref_head {
 	bool processing;
 };
 
-struct btrfs_delayed_tree_ref {
-	struct btrfs_delayed_ref_node node;
-	u64 root;
-	u64 parent;
-	int level;
-};
-
-struct btrfs_delayed_data_ref {
-	struct btrfs_delayed_ref_node node;
-	u64 root;
-	u64 parent;
-	u64 objectid;
-	u64 offset;
-};
-
 enum btrfs_delayed_ref_flags {
 	/* Indicate that we are flushing delayed refs for the commit */
 	BTRFS_DELAYED_REFS_FLUSHING,
@@ -204,42 +239,6 @@ enum btrfs_ref_type {
 	BTRFS_REF_LAST,
 } __packed;
 
-struct btrfs_data_ref {
-	/* For EXTENT_DATA_REF */
-
-	/* Root which owns this data reference. */
-	u64 ref_root;
-
-	/* Inode which refers to this data extent */
-	u64 ino;
-
-	/*
-	 * file_offset - extent_offset
-	 *
-	 * file_offset is the key.offset of the EXTENT_DATA key.
-	 * extent_offset is btrfs_file_extent_offset() of the EXTENT_DATA data.
-	 */
-	u64 offset;
-};
-
-struct btrfs_tree_ref {
-	/*
-	 * Level of this tree block
-	 *
-	 * Shared for skinny (TREE_BLOCK_REF) and normal tree ref.
-	 */
-	int level;
-
-	/*
-	 * Root which owns this tree block reference.
-	 *
-	 * For TREE_BLOCK_REF (skinny metadata, either inline or keyed)
-	 */
-	u64 ref_root;
-
-	/* For non-skinny metadata, no special member needed */
-};
-
 struct btrfs_ref {
 	enum btrfs_ref_type type;
 	enum btrfs_delayed_ref_action action;
@@ -257,9 +256,15 @@ struct btrfs_ref {
 	u64 real_root;
 #endif
 	u64 bytenr;
-	u64 len;
+	u64 num_bytes;
 	u64 owning_root;
 
+	/*
+	 * The root that owns the reference for this reference, this will be set
+	 * or ->parent will be set, depending on what type of reference this is.
+	 */
+	u64 ref_root;
+
 	/* Bytenr of the parent tree block */
 	u64 parent;
 	union {
@@ -269,8 +274,7 @@ struct btrfs_ref {
 };
 
 extern struct kmem_cache *btrfs_delayed_ref_head_cachep;
-extern struct kmem_cache *btrfs_delayed_tree_ref_cachep;
-extern struct kmem_cache *btrfs_delayed_data_ref_cachep;
+extern struct kmem_cache *btrfs_delayed_ref_node_cachep;
 extern struct kmem_cache *btrfs_delayed_extent_op_cachep;
 
 int __init btrfs_delayed_ref_init(void);
@@ -308,53 +312,10 @@ static inline u64 btrfs_calc_delayed_ref_csum_bytes(const struct btrfs_fs_info *
 	return btrfs_calc_metadata_size(fs_info, num_csum_items);
 }
 
-static inline void btrfs_init_generic_ref(struct btrfs_ref *generic_ref,
-					  int action, u64 bytenr, u64 len,
-					  u64 parent, u64 owning_root)
-{
-	generic_ref->action = action;
-	generic_ref->bytenr = bytenr;
-	generic_ref->len = len;
-	generic_ref->parent = parent;
-	generic_ref->owning_root = owning_root;
-}
-
-static inline void btrfs_init_tree_ref(struct btrfs_ref *generic_ref, int level,
-				       u64 root, u64 mod_root, bool skip_qgroup)
-{
-#ifdef CONFIG_BTRFS_FS_REF_VERIFY
-	/* If @real_root not set, use @root as fallback */
-	generic_ref->real_root = mod_root ?: root;
-#endif
-	generic_ref->tree_ref.level = level;
-	generic_ref->tree_ref.ref_root = root;
-	generic_ref->type = BTRFS_REF_METADATA;
-	if (skip_qgroup || !(is_fstree(root) &&
-			     (!mod_root || is_fstree(mod_root))))
-		generic_ref->skip_qgroup = true;
-	else
-		generic_ref->skip_qgroup = false;
-
-}
-
-static inline void btrfs_init_data_ref(struct btrfs_ref *generic_ref,
-				u64 ref_root, u64 ino, u64 offset, u64 mod_root,
-				bool skip_qgroup)
-{
-#ifdef CONFIG_BTRFS_FS_REF_VERIFY
-	/* If @real_root not set, use @root as fallback */
-	generic_ref->real_root = mod_root ?: ref_root;
-#endif
-	generic_ref->data_ref.ref_root = ref_root;
-	generic_ref->data_ref.ino = ino;
-	generic_ref->data_ref.offset = offset;
-	generic_ref->type = BTRFS_REF_DATA;
-	if (skip_qgroup || !(is_fstree(ref_root) &&
-			     (!mod_root || is_fstree(mod_root))))
-		generic_ref->skip_qgroup = true;
-	else
-		generic_ref->skip_qgroup = false;
-}
+void btrfs_init_tree_ref(struct btrfs_ref *generic_ref, int level, u64 mod_root,
+			 bool skip_qgroup);
+void btrfs_init_data_ref(struct btrfs_ref *generic_ref, u64 ino, u64 offset,
+			 u64 mod_root, bool skip_qgroup);
 
 static inline struct btrfs_delayed_extent_op *
 btrfs_alloc_delayed_extent_op(void)
@@ -369,24 +330,7 @@ btrfs_free_delayed_extent_op(struct btrfs_delayed_extent_op *op)
 		kmem_cache_free(btrfs_delayed_extent_op_cachep, op);
 }
 
-static inline void btrfs_put_delayed_ref(struct btrfs_delayed_ref_node *ref)
-{
-	if (refcount_dec_and_test(&ref->refs)) {
-		WARN_ON(!RB_EMPTY_NODE(&ref->ref_node));
-		switch (ref->type) {
-		case BTRFS_TREE_BLOCK_REF_KEY:
-		case BTRFS_SHARED_BLOCK_REF_KEY:
-			kmem_cache_free(btrfs_delayed_tree_ref_cachep, ref);
-			break;
-		case BTRFS_EXTENT_DATA_REF_KEY:
-		case BTRFS_SHARED_DATA_REF_KEY:
-			kmem_cache_free(btrfs_delayed_data_ref_cachep, ref);
-			break;
-		default:
-			BUG();
-		}
-	}
-}
+void btrfs_put_delayed_ref(struct btrfs_delayed_ref_node *ref);
 
 static inline u64 btrfs_ref_head_to_space_flags(
 				struct btrfs_delayed_ref_head *head_ref)
@@ -446,19 +390,39 @@ void btrfs_migrate_to_delayed_refs_rsv(struct btrfs_fs_info *fs_info,
 				       u64 num_bytes);
 bool btrfs_check_space_for_delayed_refs(struct btrfs_fs_info *fs_info);
 
-/*
- * helper functions to cast a node into its container
- */
-static inline struct btrfs_delayed_tree_ref *
-btrfs_delayed_node_to_tree_ref(struct btrfs_delayed_ref_node *node)
+static inline u64 btrfs_delayed_ref_owner(struct btrfs_delayed_ref_node *node)
+{
+	if (node->type == BTRFS_EXTENT_DATA_REF_KEY ||
+	    node->type == BTRFS_SHARED_DATA_REF_KEY)
+		return node->data_ref.objectid;
+	return node->tree_ref.level;
+}
+
+static inline u64 btrfs_delayed_ref_offset(struct btrfs_delayed_ref_node *node)
 {
-	return container_of(node, struct btrfs_delayed_tree_ref, node);
+	if (node->type == BTRFS_EXTENT_DATA_REF_KEY ||
+	    node->type == BTRFS_SHARED_DATA_REF_KEY)
+		return node->data_ref.offset;
+	return 0;
 }
 
-static inline struct btrfs_delayed_data_ref *
-btrfs_delayed_node_to_data_ref(struct btrfs_delayed_ref_node *node)
+static inline u8 btrfs_ref_type(struct btrfs_ref *ref)
 {
-	return container_of(node, struct btrfs_delayed_data_ref, node);
+	ASSERT(ref->type == BTRFS_REF_DATA || ref->type == BTRFS_REF_METADATA);
+
+	if (ref->type == BTRFS_REF_DATA) {
+		if (ref->parent)
+			return BTRFS_SHARED_DATA_REF_KEY;
+		else
+			return BTRFS_EXTENT_DATA_REF_KEY;
+	} else {
+		if (ref->parent)
+			return BTRFS_SHARED_BLOCK_REF_KEY;
+		else
+			return BTRFS_TREE_BLOCK_REF_KEY;
+	}
+
+	return 0;
 }
 
 #endif
diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c
index 79c4293ddf..7130040d92 100644
--- a/fs/btrfs/dev-replace.c
+++ b/fs/btrfs/dev-replace.c
@@ -11,10 +11,8 @@
 #include <linux/math64.h>
 #include "misc.h"
 #include "ctree.h"
-#include "extent_map.h"
 #include "disk-io.h"
 #include "transaction.h"
-#include "print-tree.h"
 #include "volumes.h"
 #include "async-thread.h"
 #include "dev-replace.h"
@@ -246,7 +244,7 @@ static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 {
 	struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
 	struct btrfs_device *device;
-	struct bdev_handle *bdev_handle;
+	struct file *bdev_file;
 	struct block_device *bdev;
 	u64 devid = BTRFS_DEV_REPLACE_DEVID;
 	int ret = 0;
@@ -257,13 +255,13 @@ static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 		return -EINVAL;
 	}
 
-	bdev_handle = bdev_open_by_path(device_path, BLK_OPEN_WRITE,
+	bdev_file = bdev_file_open_by_path(device_path, BLK_OPEN_WRITE,
 					fs_info->bdev_holder, NULL);
-	if (IS_ERR(bdev_handle)) {
+	if (IS_ERR(bdev_file)) {
 		btrfs_err(fs_info, "target device %s is invalid!", device_path);
-		return PTR_ERR(bdev_handle);
+		return PTR_ERR(bdev_file);
 	}
-	bdev = bdev_handle->bdev;
+	bdev = file_bdev(bdev_file);
 
 	if (!btrfs_check_device_zone_type(fs_info, bdev)) {
 		btrfs_err(fs_info,
@@ -314,11 +312,11 @@ static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 	device->commit_bytes_used = device->bytes_used;
 	device->fs_info = fs_info;
 	device->bdev = bdev;
-	device->bdev_handle = bdev_handle;
+	device->bdev_file = bdev_file;
 	set_bit(BTRFS_DEV_STATE_IN_FS_METADATA, &device->dev_state);
 	set_bit(BTRFS_DEV_STATE_REPLACE_TGT, &device->dev_state);
 	device->dev_stats_valid = 1;
-	set_blocksize(device->bdev, BTRFS_BDEV_BLOCKSIZE);
+	set_blocksize(bdev_file, BTRFS_BDEV_BLOCKSIZE);
 	device->fs_devices = fs_devices;
 
 	ret = btrfs_get_dev_zone_info(device, false);
@@ -335,7 +333,7 @@ static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 	return 0;
 
 error:
-	bdev_release(bdev_handle);
+	fput(bdev_file);
 	return ret;
 }
 
@@ -1000,8 +998,7 @@ error:
 	btrfs_sysfs_remove_device(src_device);
 	btrfs_sysfs_update_devid(tgt_device);
 	if (test_bit(BTRFS_DEV_STATE_WRITEABLE, &src_device->dev_state))
-		btrfs_scratch_superblocks(fs_info, src_device->bdev,
-					  src_device->name->str);
+		btrfs_scratch_superblocks(fs_info, src_device);
 
 	/* write back the superblocks */
 	trans = btrfs_start_transaction(root, 0);
diff --git a/fs/btrfs/dev-replace.h b/fs/btrfs/dev-replace.h
index 675082ccec..23e480efe5 100644
--- a/fs/btrfs/dev-replace.h
+++ b/fs/btrfs/dev-replace.h
@@ -6,11 +6,15 @@
 #ifndef BTRFS_DEV_REPLACE_H
 #define BTRFS_DEV_REPLACE_H
 
+#include <linux/types.h>
+#include <linux/compiler_types.h>
+
 struct btrfs_ioctl_dev_replace_args;
 struct btrfs_fs_info;
 struct btrfs_trans_handle;
 struct btrfs_dev_replace;
 struct btrfs_block_group;
+struct btrfs_device;
 
 int btrfs_init_dev_replace(struct btrfs_fs_info *fs_info);
 int btrfs_run_dev_replace(struct btrfs_trans_handle *trans);
diff --git a/fs/btrfs/dir-item.h b/fs/btrfs/dir-item.h
index e40a226373..00b3d83d75 100644
--- a/fs/btrfs/dir-item.h
+++ b/fs/btrfs/dir-item.h
@@ -3,9 +3,15 @@
 #ifndef BTRFS_DIR_ITEM_H
 #define BTRFS_DIR_ITEM_H
 
+#include <linux/types.h>
 #include <linux/crc32c.h>
 
 struct fscrypt_str;
+struct btrfs_fs_info;
+struct btrfs_key;
+struct btrfs_path;
+struct btrfs_root;
+struct btrfs_trans_handle;
 
 int btrfs_check_dir_item_collision(struct btrfs_root *root, u64 dir,
 			  const struct fscrypt_str *name);
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 2567821224..3791813dc7 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -29,7 +29,6 @@
 #include "tree-log.h"
 #include "free-space-cache.h"
 #include "free-space-tree.h"
-#include "rcu-string.h"
 #include "dev-replace.h"
 #include "raid56.h"
 #include "sysfs.h"
@@ -193,7 +192,7 @@ static int btrfs_repair_eb_io_failure(const struct extent_buffer *eb,
 		struct folio *folio = eb->folios[i];
 		u64 start = max_t(u64, eb->start, folio_pos(folio));
 		u64 end = min_t(u64, eb->start + eb->len,
-				folio_pos(folio) + folio_size(folio));
+				folio_pos(folio) + eb->folio_size);
 		u32 len = end - start;
 
 		ret = btrfs_repair_io_failure(fs_info, 0, start, len,
@@ -647,7 +646,7 @@ struct extent_buffer *read_tree_block(struct btrfs_fs_info *fs_info, u64 bytenr,
 static void __setup_root(struct btrfs_root *root, struct btrfs_fs_info *fs_info,
 			 u64 objectid)
 {
-	bool dummy = test_bit(BTRFS_FS_STATE_DUMMY_FS_INFO, &fs_info->fs_state);
+	bool dummy = btrfs_is_testing(fs_info);
 
 	memset(&root->root_key, 0, sizeof(root->root_key));
 	memset(&root->root_item, 0, sizeof(root->root_item));
@@ -659,13 +658,12 @@ static void __setup_root(struct btrfs_root *root, struct btrfs_fs_info *fs_info,
 	root->state = 0;
 	RB_CLEAR_NODE(&root->rb_node);
 
-	root->last_trans = 0;
+	btrfs_set_root_last_trans(root, 0);
 	root->free_objectid = 0;
 	root->nr_delalloc_inodes = 0;
 	root->nr_ordered_extents = 0;
 	root->inode_tree = RB_ROOT;
-	/* GFP flags are compatible with XA_FLAGS_*. */
-	xa_init_flags(&root->delayed_nodes, GFP_ATOMIC);
+	xa_init(&root->delayed_nodes);
 
 	btrfs_init_root_block_rsv(root);
 
@@ -777,7 +775,7 @@ int btrfs_global_root_insert(struct btrfs_root *root)
 	if (tmp) {
 		ret = -EEXIST;
 		btrfs_warn(fs_info, "global root %llu %llu already exists",
-				root->root_key.objectid, root->root_key.offset);
+			   btrfs_root_id(root), root->root_key.offset);
 	}
 	return ret;
 }
@@ -1012,8 +1010,8 @@ int btrfs_add_log_tree(struct btrfs_trans_handle *trans,
 		return ret;
 	}
 
-	log_root->last_trans = trans->transid;
-	log_root->root_key.offset = root->root_key.objectid;
+	btrfs_set_root_last_trans(log_root, trans->transid);
+	log_root->root_key.offset = btrfs_root_id(root);
 
 	inode_item = &log_root->root_item.inode;
 	btrfs_set_stack_inode_generation(inode_item, 1);
@@ -1077,15 +1075,15 @@ static struct btrfs_root *read_tree_root_path(struct btrfs_root *tree_root,
 	 * For real fs, and not log/reloc trees, root owner must
 	 * match its root node owner
 	 */
-	if (!test_bit(BTRFS_FS_STATE_DUMMY_FS_INFO, &fs_info->fs_state) &&
-	    root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID &&
-	    root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID &&
-	    root->root_key.objectid != btrfs_header_owner(root->node)) {
+	if (!btrfs_is_testing(fs_info) &&
+	    btrfs_root_id(root) != BTRFS_TREE_LOG_OBJECTID &&
+	    btrfs_root_id(root) != BTRFS_TREE_RELOC_OBJECTID &&
+	    btrfs_root_id(root) != btrfs_header_owner(root->node)) {
 		btrfs_crit(fs_info,
 "root=%llu block=%llu, tree root owner mismatch, have %llu expect %llu",
-			   root->root_key.objectid, root->node->start,
+			   btrfs_root_id(root), root->node->start,
 			   btrfs_header_owner(root->node),
-			   root->root_key.objectid);
+			   btrfs_root_id(root));
 		ret = -EUCLEAN;
 		goto fail;
 	}
@@ -1122,9 +1120,9 @@ static int btrfs_init_fs_root(struct btrfs_root *root, dev_t anon_dev)
 
 	btrfs_drew_lock_init(&root->snapshot_lock);
 
-	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID &&
+	if (btrfs_root_id(root) != BTRFS_TREE_LOG_OBJECTID &&
 	    !btrfs_is_data_reloc_root(root) &&
-	    is_fstree(root->root_key.objectid)) {
+	    is_fstree(btrfs_root_id(root))) {
 		set_bit(BTRFS_ROOT_SHAREABLE, &root->state);
 		btrfs_check_and_init_root_item(&root->root_item);
 	}
@@ -1133,7 +1131,7 @@ static int btrfs_init_fs_root(struct btrfs_root *root, dev_t anon_dev)
 	 * Don't assign anonymous block device to roots that are not exposed to
 	 * userspace, the id pool is limited to 1M
 	 */
-	if (is_fstree(root->root_key.objectid) &&
+	if (is_fstree(btrfs_root_id(root)) &&
 	    btrfs_root_refs(&root->root_item) > 0) {
 		if (!anon_dev) {
 			ret = get_anon_bdev(&root->anon_dev);
@@ -1220,7 +1218,7 @@ int btrfs_insert_fs_root(struct btrfs_fs_info *fs_info,
 
 	spin_lock(&fs_info->fs_roots_radix_lock);
 	ret = radix_tree_insert(&fs_info->fs_roots_radix,
-				(unsigned long)root->root_key.objectid,
+				(unsigned long)btrfs_root_id(root),
 				root);
 	if (ret == 0) {
 		btrfs_grab_root(root);
@@ -1245,6 +1243,7 @@ void btrfs_check_leaked_roots(struct btrfs_fs_info *fs_info)
 		btrfs_err(fs_info, "leaked root %s refcount %d",
 			  btrfs_root_name(&root->root_key, buf),
 			  refcount_read(&root->refs));
+		WARN_ON_ONCE(1);
 		while (refcount_read(&root->refs) > 1)
 			btrfs_put_root(root);
 		btrfs_put_root(root);
@@ -1266,9 +1265,14 @@ static void free_global_roots(struct btrfs_fs_info *fs_info)
 
 void btrfs_free_fs_info(struct btrfs_fs_info *fs_info)
 {
+	struct percpu_counter *em_counter = &fs_info->evictable_extent_maps;
+
 	percpu_counter_destroy(&fs_info->dirty_metadata_bytes);
 	percpu_counter_destroy(&fs_info->delalloc_bytes);
 	percpu_counter_destroy(&fs_info->ordered_bytes);
+	if (percpu_counter_initialized(em_counter))
+		ASSERT(percpu_counter_sum_positive(em_counter) == 0);
+	percpu_counter_destroy(em_counter);
 	percpu_counter_destroy(&fs_info->dev_replace.bio_counter);
 	btrfs_free_csum_hash(fs_info);
 	btrfs_free_stripe_hash_table(fs_info);
@@ -2240,7 +2244,7 @@ static int btrfs_read_roots(struct btrfs_fs_info *fs_info)
 	struct btrfs_key location;
 	int ret;
 
-	BUG_ON(!fs_info->tree_root);
+	ASSERT(fs_info->tree_root);
 
 	ret = load_global_roots(tree_root);
 	if (ret)
@@ -2584,7 +2588,7 @@ static int load_super_root(struct btrfs_root *root, u64 bytenr, u64 gen, int lev
 	struct btrfs_tree_parent_check check = {
 		.level = level,
 		.transid = gen,
-		.owner_root = root->root_key.objectid
+		.owner_root = btrfs_root_id(root)
 	};
 	int ret = 0;
 
@@ -2848,6 +2852,12 @@ static int init_mount_fs_info(struct btrfs_fs_info *fs_info, struct super_block
 	if (ret)
 		return ret;
 
+	ret = percpu_counter_init(&fs_info->evictable_extent_maps, 0, GFP_KERNEL);
+	if (ret)
+		return ret;
+
+	spin_lock_init(&fs_info->extent_map_shrinker_lock);
+
 	ret = percpu_counter_init(&fs_info->dirty_metadata_bytes, 0, GFP_KERNEL);
 	if (ret)
 		return ret;
@@ -2930,7 +2940,7 @@ static int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info)
 			spin_unlock(&fs_info->fs_roots_radix_lock);
 			break;
 		}
-		root_objectid = gang[ret - 1]->root_key.objectid + 1;
+		root_objectid = btrfs_root_id(gang[ret - 1]) + 1;
 
 		for (i = 0; i < ret; i++) {
 			/* Avoid to grab roots in dead_roots. */
@@ -2946,7 +2956,7 @@ static int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info)
 		for (i = 0; i < ret; i++) {
 			if (!gang[i])
 				continue;
-			root_objectid = gang[i]->root_key.objectid;
+			root_objectid = btrfs_root_id(gang[i]);
 			err = btrfs_orphan_cleanup(gang[i]);
 			if (err)
 				goto out;
@@ -3618,28 +3628,25 @@ ALLOW_ERROR_INJECTION(open_ctree, ERRNO);
 static void btrfs_end_super_write(struct bio *bio)
 {
 	struct btrfs_device *device = bio->bi_private;
-	struct bio_vec *bvec;
-	struct bvec_iter_all iter_all;
-	struct page *page;
-
-	bio_for_each_segment_all(bvec, bio, iter_all) {
-		page = bvec->bv_page;
+	struct folio_iter fi;
 
+	bio_for_each_folio_all(fi, bio) {
 		if (bio->bi_status) {
 			btrfs_warn_rl_in_rcu(device->fs_info,
-				"lost page write due to IO error on %s (%d)",
+				"lost super block write due to IO error on %s (%d)",
 				btrfs_dev_name(device),
 				blk_status_to_errno(bio->bi_status));
-			ClearPageUptodate(page);
-			SetPageError(page);
 			btrfs_dev_stat_inc_and_print(device,
 						     BTRFS_DEV_STAT_WRITE_ERRS);
-		} else {
-			SetPageUptodate(page);
+			/* Ensure failure if the primary sb fails. */
+			if (bio->bi_opf & REQ_FUA)
+				atomic_add(BTRFS_SUPER_PRIMARY_WRITE_ERROR,
+					   &device->sb_write_errors);
+			else
+				atomic_inc(&device->sb_write_errors);
 		}
-
-		put_page(page);
-		unlock_page(page);
+		folio_unlock(fi.folio);
+		folio_put(fi.folio);
 	}
 
 	bio_put(bio);
@@ -3651,7 +3658,7 @@ struct btrfs_super_block *btrfs_read_dev_one_super(struct block_device *bdev,
 	struct btrfs_super_block *super;
 	struct page *page;
 	u64 bytenr, bytenr_orig;
-	struct address_space *mapping = bdev->bd_inode->i_mapping;
+	struct address_space *mapping = bdev->bd_mapping;
 	int ret;
 
 	bytenr_orig = btrfs_sb_offset(copy_num);
@@ -3726,34 +3733,36 @@ struct btrfs_super_block *btrfs_read_dev_super(struct block_device *bdev)
 
 /*
  * Write superblock @sb to the @device. Do not wait for completion, all the
- * pages we use for writing are locked.
+ * folios we use for writing are locked.
  *
  * Write @max_mirrors copies of the superblock, where 0 means default that fit
  * the expected device size at commit time. Note that max_mirrors must be
  * same for write and wait phases.
  *
- * Return number of errors when page is not found or submission fails.
+ * Return number of errors when folio is not found or submission fails.
  */
 static int write_dev_supers(struct btrfs_device *device,
 			    struct btrfs_super_block *sb, int max_mirrors)
 {
 	struct btrfs_fs_info *fs_info = device->fs_info;
-	struct address_space *mapping = device->bdev->bd_inode->i_mapping;
+	struct address_space *mapping = device->bdev->bd_mapping;
 	SHASH_DESC_ON_STACK(shash, fs_info->csum_shash);
 	int i;
-	int errors = 0;
 	int ret;
 	u64 bytenr, bytenr_orig;
 
+	atomic_set(&device->sb_write_errors, 0);
+
 	if (max_mirrors == 0)
 		max_mirrors = BTRFS_SUPER_MIRROR_MAX;
 
 	shash->tfm = fs_info->csum_shash;
 
 	for (i = 0; i < max_mirrors; i++) {
-		struct page *page;
+		struct folio *folio;
 		struct bio *bio;
 		struct btrfs_super_block *disk_super;
+		size_t offset;
 
 		bytenr_orig = btrfs_sb_offset(i);
 		ret = btrfs_sb_log_location(device, i, WRITE, &bytenr);
@@ -3763,7 +3772,7 @@ static int write_dev_supers(struct btrfs_device *device,
 			btrfs_err(device->fs_info,
 				"couldn't get super block location for mirror %d",
 				i);
-			errors++;
+			atomic_inc(&device->sb_write_errors);
 			continue;
 		}
 		if (bytenr + BTRFS_SUPER_INFO_SIZE >=
@@ -3776,20 +3785,20 @@ static int write_dev_supers(struct btrfs_device *device,
 				    BTRFS_SUPER_INFO_SIZE - BTRFS_CSUM_SIZE,
 				    sb->csum);
 
-		page = find_or_create_page(mapping, bytenr >> PAGE_SHIFT,
-					   GFP_NOFS);
-		if (!page) {
+		folio = __filemap_get_folio(mapping, bytenr >> PAGE_SHIFT,
+					    FGP_LOCK | FGP_ACCESSED | FGP_CREAT,
+					    GFP_NOFS);
+		if (IS_ERR(folio)) {
 			btrfs_err(device->fs_info,
 			    "couldn't get super block page for bytenr %llu",
 			    bytenr);
-			errors++;
+			atomic_inc(&device->sb_write_errors);
 			continue;
 		}
+		ASSERT(folio_order(folio) == 0);
 
-		/* Bump the refcount for wait_dev_supers() */
-		get_page(page);
-
-		disk_super = page_address(page);
+		offset = offset_in_folio(folio, bytenr);
+		disk_super = folio_address(folio) + offset;
 		memcpy(disk_super, sb, BTRFS_SUPER_INFO_SIZE);
 
 		/*
@@ -3803,8 +3812,7 @@ static int write_dev_supers(struct btrfs_device *device,
 		bio->bi_iter.bi_sector = bytenr >> SECTOR_SHIFT;
 		bio->bi_private = device;
 		bio->bi_end_io = btrfs_end_super_write;
-		__bio_add_page(bio, page, BTRFS_SUPER_INFO_SIZE,
-			       offset_in_page(bytenr));
+		bio_add_folio_nofail(bio, folio, BTRFS_SUPER_INFO_SIZE, offset);
 
 		/*
 		 * We FUA only the first super block.  The others we allow to
@@ -3816,17 +3824,17 @@ static int write_dev_supers(struct btrfs_device *device,
 		submit_bio(bio);
 
 		if (btrfs_advance_sb_log(device, i))
-			errors++;
+			atomic_inc(&device->sb_write_errors);
 	}
-	return errors < i ? 0 : -1;
+	return atomic_read(&device->sb_write_errors) < i ? 0 : -1;
 }
 
 /*
  * Wait for write completion of superblocks done by write_dev_supers,
  * @max_mirrors same for write and wait phases.
  *
- * Return number of errors when page is not found or not marked up to
- * date.
+ * Return -1 if primary super block write failed or when there were no super block
+ * copies written. Otherwise 0.
  */
 static int wait_dev_supers(struct btrfs_device *device, int max_mirrors)
 {
@@ -3840,7 +3848,7 @@ static int wait_dev_supers(struct btrfs_device *device, int max_mirrors)
 		max_mirrors = BTRFS_SUPER_MIRROR_MAX;
 
 	for (i = 0; i < max_mirrors; i++) {
-		struct page *page;
+		struct folio *folio;
 
 		ret = btrfs_sb_log_location(device, i, READ, &bytenr);
 		if (ret == -ENOENT) {
@@ -3855,30 +3863,21 @@ static int wait_dev_supers(struct btrfs_device *device, int max_mirrors)
 		    device->commit_total_bytes)
 			break;
 
-		page = find_get_page(device->bdev->bd_inode->i_mapping,
-				     bytenr >> PAGE_SHIFT);
-		if (!page) {
-			errors++;
-			if (i == 0)
-				primary_failed = true;
+		folio = filemap_get_folio(device->bdev->bd_mapping,
+					  bytenr >> PAGE_SHIFT);
+		/* If the folio has been removed, then we know it completed. */
+		if (IS_ERR(folio))
 			continue;
-		}
-		/* Page is submitted locked and unlocked once the IO completes */
-		wait_on_page_locked(page);
-		if (PageError(page)) {
-			errors++;
-			if (i == 0)
-				primary_failed = true;
-		}
-
-		/* Drop our reference */
-		put_page(page);
+		ASSERT(folio_order(folio) == 0);
 
-		/* Drop the reference from the writing run */
-		put_page(page);
+		/* Folio will be unlocked once the write completes. */
+		folio_wait_locked(folio);
+		folio_put(folio);
 	}
 
-	/* log error, force error return */
+	errors += atomic_read(&device->sb_write_errors);
+	if (errors >= BTRFS_SUPER_PRIMARY_WRITE_ERROR)
+		primary_failed = true;
 	if (primary_failed) {
 		btrfs_err(device->fs_info, "error writing primary super block to device %llu",
 			  device->devid);
@@ -4139,7 +4138,7 @@ void btrfs_drop_and_free_fs_root(struct btrfs_fs_info *fs_info,
 
 	spin_lock(&fs_info->fs_roots_radix_lock);
 	radix_tree_delete(&fs_info->fs_roots_radix,
-			  (unsigned long)root->root_key.objectid);
+			  (unsigned long)btrfs_root_id(root));
 	if (test_and_clear_bit(BTRFS_ROOT_IN_RADIX, &root->state))
 		drop_ref = true;
 	spin_unlock(&fs_info->fs_roots_radix_lock);
@@ -4182,9 +4181,6 @@ static void warn_about_uncommitted_trans(struct btrfs_fs_info *fs_info)
 	struct btrfs_transaction *tmp;
 	bool found = false;
 
-	if (list_empty(&fs_info->trans_list))
-		return;
-
 	/*
 	 * This function is only called at the very end of close_ctree(),
 	 * thus no other running transaction, no need to take trans_lock.
@@ -4484,7 +4480,7 @@ static void btrfs_drop_all_logs(struct btrfs_fs_info *fs_info)
 		for (i = 0; i < ret; i++) {
 			if (!gang[i])
 				continue;
-			root_objectid = gang[i]->root_key.objectid;
+			root_objectid = btrfs_root_id(gang[i]);
 			btrfs_free_log(NULL, gang[i]);
 			btrfs_put_root(gang[i]);
 		}
@@ -4544,18 +4540,10 @@ static void btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
 				       struct btrfs_fs_info *fs_info)
 {
 	struct rb_node *node;
-	struct btrfs_delayed_ref_root *delayed_refs;
+	struct btrfs_delayed_ref_root *delayed_refs = &trans->delayed_refs;
 	struct btrfs_delayed_ref_node *ref;
 
-	delayed_refs = &trans->delayed_refs;
-
 	spin_lock(&delayed_refs->lock);
-	if (atomic_read(&delayed_refs->num_entries) == 0) {
-		spin_unlock(&delayed_refs->lock);
-		btrfs_debug(fs_info, "delayed_refs has NO entry");
-		return;
-	}
-
 	while ((node = rb_first_cached(&delayed_refs->href_root)) != NULL) {
 		struct btrfs_delayed_ref_head *head;
 		struct rb_node *n;
@@ -4629,7 +4617,7 @@ static void btrfs_destroy_delalloc_inodes(struct btrfs_root *root)
 		struct inode *inode = NULL;
 		btrfs_inode = list_first_entry(&splice, struct btrfs_inode,
 					       delalloc_inodes);
-		__btrfs_del_delalloc_inode(root, btrfs_inode);
+		btrfs_del_delalloc_inode(btrfs_inode);
 		spin_unlock(&root->delalloc_lock);
 
 		/*
@@ -4815,7 +4803,7 @@ static void btrfs_free_all_qgroup_pertrans(struct btrfs_fs_info *fs_info)
 
 			btrfs_qgroup_free_meta_all_pertrans(root);
 			radix_tree_tag_clear(&fs_info->fs_roots_radix,
-					(unsigned long)root->root_key.objectid,
+					(unsigned long)btrfs_root_id(root),
 					BTRFS_ROOT_TRANS_TAG);
 		}
 	}
@@ -4844,14 +4832,10 @@ void btrfs_cleanup_one_transaction(struct btrfs_transaction *cur_trans,
 	cur_trans->state = TRANS_STATE_UNBLOCKED;
 	wake_up(&fs_info->transaction_wait);
 
-	btrfs_destroy_delayed_inodes(fs_info);
-
 	btrfs_destroy_marked_extents(fs_info, &cur_trans->dirty_pages,
 				     EXTENT_DIRTY);
 	btrfs_destroy_pinned_extent(fs_info, &cur_trans->pinned_extents);
 
-	btrfs_free_all_qgroup_pertrans(fs_info);
-
 	cur_trans->state =TRANS_STATE_COMPLETED;
 	wake_up(&cur_trans->commit_wait);
 }
@@ -4904,6 +4888,7 @@ static int btrfs_cleanup_transaction(struct btrfs_fs_info *fs_info)
 	btrfs_assert_delayed_root_empty(fs_info);
 	btrfs_destroy_all_delalloc_inodes(fs_info);
 	btrfs_drop_all_logs(fs_info);
+	btrfs_free_all_qgroup_pertrans(fs_info);
 	mutex_unlock(&fs_info->transaction_kthread_mutex);
 
 	return 0;
@@ -4928,7 +4913,14 @@ int btrfs_init_root_free_objectid(struct btrfs_root *root)
 	ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
 	if (ret < 0)
 		goto error;
-	BUG_ON(ret == 0); /* Corruption */
+	if (ret == 0) {
+		/*
+		 * Key with offset -1 found, there would have to exist a root
+		 * with such id, but this is out of valid range.
+		 */
+		ret = -EUCLEAN;
+		goto error;
+	}
 	if (path->slots[0] > 0) {
 		slot = path->slots[0] - 1;
 		l = path->nodes[0];
@@ -4952,7 +4944,7 @@ int btrfs_get_free_objectid(struct btrfs_root *root, u64 *objectid)
 	if (unlikely(root->free_objectid >= BTRFS_LAST_FREE_OBJECTID)) {
 		btrfs_warn(root->fs_info,
 			   "the objectid of root %llu reaches its highest value",
-			   root->root_key.objectid);
+			   btrfs_root_id(root));
 		ret = -ENOSPC;
 		goto out;
 	}
diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h
index eb3473d1c1..76eb53fe7a 100644
--- a/fs/btrfs/disk-io.h
+++ b/fs/btrfs/disk-io.h
@@ -6,6 +6,22 @@
 #ifndef BTRFS_DISK_IO_H
 #define BTRFS_DISK_IO_H
 
+#include <linux/sizes.h>
+#include <linux/compiler_types.h>
+#include "ctree.h"
+#include "fs.h"
+
+struct block_device;
+struct super_block;
+struct extent_buffer;
+struct btrfs_device;
+struct btrfs_fs_devices;
+struct btrfs_fs_info;
+struct btrfs_super_block;
+struct btrfs_trans_handle;
+struct btrfs_tree_parent_check;
+struct btrfs_transaction;
+
 #define BTRFS_SUPER_MIRROR_MAX	 3
 #define BTRFS_SUPER_MIRROR_SHIFT 12
 
@@ -25,10 +41,6 @@ static inline u64 btrfs_sb_offset(int mirror)
 	return BTRFS_SUPER_INFO_OFFSET;
 }
 
-struct btrfs_device;
-struct btrfs_fs_devices;
-struct btrfs_tree_parent_check;
-
 void btrfs_check_leaked_roots(struct btrfs_fs_info *fs_info);
 void btrfs_init_fs_info(struct btrfs_fs_info *fs_info);
 struct extent_buffer *read_tree_block(struct btrfs_fs_info *fs_info, u64 bytenr,
diff --git a/fs/btrfs/export.c b/fs/btrfs/export.c
index 1ac00ee795..9e81f89e76 100644
--- a/fs/btrfs/export.c
+++ b/fs/btrfs/export.c
@@ -5,7 +5,6 @@
 #include "ctree.h"
 #include "disk-io.h"
 #include "btrfs_inode.h"
-#include "print-tree.h"
 #include "export.h"
 #include "accessors.h"
 #include "super.h"
@@ -35,7 +34,7 @@ static int btrfs_encode_fh(struct inode *inode, u32 *fh, int *max_len,
 	type = FILEID_BTRFS_WITHOUT_PARENT;
 
 	fid->objectid = btrfs_ino(BTRFS_I(inode));
-	fid->root_objectid = BTRFS_I(inode)->root->root_key.objectid;
+	fid->root_objectid = btrfs_root_id(BTRFS_I(inode)->root);
 	fid->gen = inode->i_generation;
 
 	if (parent) {
@@ -43,7 +42,7 @@ static int btrfs_encode_fh(struct inode *inode, u32 *fh, int *max_len,
 
 		fid->parent_objectid = BTRFS_I(parent)->location.objectid;
 		fid->parent_gen = parent->i_generation;
-		parent_root_id = BTRFS_I(parent)->root->root_key.objectid;
+		parent_root_id = btrfs_root_id(BTRFS_I(parent)->root);
 
 		if (parent_root_id != fid->root_objectid) {
 			fid->parent_root_objectid = parent_root_id;
@@ -161,7 +160,7 @@ struct dentry *btrfs_get_parent(struct dentry *child)
 		return ERR_PTR(-ENOMEM);
 
 	if (btrfs_ino(BTRFS_I(dir)) == BTRFS_FIRST_FREE_OBJECTID) {
-		key.objectid = root->root_key.objectid;
+		key.objectid = btrfs_root_id(root);
 		key.type = BTRFS_ROOT_BACKREF_KEY;
 		key.offset = (u64)-1;
 		root = fs_info->tree_root;
@@ -244,7 +243,7 @@ static int btrfs_get_name(struct dentry *parent, char *name,
 		return -ENOMEM;
 
 	if (ino == BTRFS_FIRST_FREE_OBJECTID) {
-		key.objectid = BTRFS_I(inode)->root->root_key.objectid;
+		key.objectid = btrfs_root_id(BTRFS_I(inode)->root);
 		key.type = BTRFS_ROOT_BACKREF_KEY;
 		key.offset = (u64)-1;
 		root = fs_info->tree_root;
diff --git a/fs/btrfs/export.h b/fs/btrfs/export.h
index eba6bc4f5a..464582273a 100644
--- a/fs/btrfs/export.h
+++ b/fs/btrfs/export.h
@@ -4,6 +4,10 @@
 #define BTRFS_EXPORT_H
 
 #include <linux/exportfs.h>
+#include <linux/types.h>
+
+struct dentry;
+struct super_block;
 
 extern const struct export_operations btrfs_export_ops;
 
diff --git a/fs/btrfs/extent-io-tree.c b/fs/btrfs/extent-io-tree.c
index e3ee5449cc..ed2cfc3d5d 100644
--- a/fs/btrfs/extent-io-tree.c
+++ b/fs/btrfs/extent-io-tree.c
@@ -6,7 +6,6 @@
 #include "ctree.h"
 #include "extent-io-tree.h"
 #include "btrfs_inode.h"
-#include "misc.h"
 
 static struct kmem_cache *extent_state_cache;
 
@@ -48,6 +47,7 @@ static inline void btrfs_extent_state_leak_debug_check(void)
 		       extent_state_in_tree(state),
 		       refcount_read(&state->refs));
 		list_del(&state->leak_list);
+		WARN_ON_ONCE(1);
 		kmem_cache_free(extent_state_cache, state);
 	}
 }
@@ -1059,7 +1059,7 @@ static int __set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 	struct extent_state *prealloc = NULL;
 	struct rb_node **p = NULL;
 	struct rb_node *parent = NULL;
-	int err = 0;
+	int ret = 0;
 	u64 last_start;
 	u64 last_end;
 	u32 exclusive_bits = (bits & EXTENT_LOCKED);
@@ -1122,7 +1122,7 @@ hit_next:
 		if (state->state & exclusive_bits) {
 			*failed_start = state->start;
 			cache_state(state, failed_state);
-			err = -EEXIST;
+			ret = -EEXIST;
 			goto out;
 		}
 
@@ -1158,7 +1158,7 @@ hit_next:
 		if (state->state & exclusive_bits) {
 			*failed_start = start;
 			cache_state(state, failed_state);
-			err = -EEXIST;
+			ret = -EEXIST;
 			goto out;
 		}
 
@@ -1175,12 +1175,12 @@ hit_next:
 		prealloc = alloc_extent_state_atomic(prealloc);
 		if (!prealloc)
 			goto search_again;
-		err = split_state(tree, state, prealloc, start);
-		if (err)
-			extent_io_tree_panic(tree, state, "split", err);
+		ret = split_state(tree, state, prealloc, start);
+		if (ret)
+			extent_io_tree_panic(tree, state, "split", ret);
 
 		prealloc = NULL;
-		if (err)
+		if (ret)
 			goto out;
 		if (state->end <= end) {
 			set_state_bits(tree, state, bits, changeset);
@@ -1224,8 +1224,8 @@ hit_next:
 		prealloc->end = this_end;
 		inserted_state = insert_state(tree, prealloc, bits, changeset);
 		if (IS_ERR(inserted_state)) {
-			err = PTR_ERR(inserted_state);
-			extent_io_tree_panic(tree, prealloc, "insert", err);
+			ret = PTR_ERR(inserted_state);
+			extent_io_tree_panic(tree, prealloc, "insert", ret);
 		}
 
 		cache_state(inserted_state, cached_state);
@@ -1244,16 +1244,16 @@ hit_next:
 		if (state->state & exclusive_bits) {
 			*failed_start = start;
 			cache_state(state, failed_state);
-			err = -EEXIST;
+			ret = -EEXIST;
 			goto out;
 		}
 
 		prealloc = alloc_extent_state_atomic(prealloc);
 		if (!prealloc)
 			goto search_again;
-		err = split_state(tree, state, prealloc, end + 1);
-		if (err)
-			extent_io_tree_panic(tree, state, "split", err);
+		ret = split_state(tree, state, prealloc, end + 1);
+		if (ret)
+			extent_io_tree_panic(tree, state, "split", ret);
 
 		set_state_bits(tree, prealloc, bits, changeset);
 		cache_state(prealloc, cached_state);
@@ -1275,7 +1275,7 @@ out:
 	if (prealloc)
 		free_extent_state(prealloc);
 
-	return err;
+	return ret;
 
 }
 
@@ -1312,7 +1312,7 @@ int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 	struct extent_state *prealloc = NULL;
 	struct rb_node **p = NULL;
 	struct rb_node *parent = NULL;
-	int err = 0;
+	int ret = 0;
 	u64 last_start;
 	u64 last_end;
 	bool first_iteration = true;
@@ -1351,7 +1351,7 @@ again:
 	if (!state) {
 		prealloc = alloc_extent_state_atomic(prealloc);
 		if (!prealloc) {
-			err = -ENOMEM;
+			ret = -ENOMEM;
 			goto out;
 		}
 		prealloc->start = start;
@@ -1402,14 +1402,14 @@ hit_next:
 	if (state->start < start) {
 		prealloc = alloc_extent_state_atomic(prealloc);
 		if (!prealloc) {
-			err = -ENOMEM;
+			ret = -ENOMEM;
 			goto out;
 		}
-		err = split_state(tree, state, prealloc, start);
-		if (err)
-			extent_io_tree_panic(tree, state, "split", err);
+		ret = split_state(tree, state, prealloc, start);
+		if (ret)
+			extent_io_tree_panic(tree, state, "split", ret);
 		prealloc = NULL;
-		if (err)
+		if (ret)
 			goto out;
 		if (state->end <= end) {
 			set_state_bits(tree, state, bits, NULL);
@@ -1442,7 +1442,7 @@ hit_next:
 
 		prealloc = alloc_extent_state_atomic(prealloc);
 		if (!prealloc) {
-			err = -ENOMEM;
+			ret = -ENOMEM;
 			goto out;
 		}
 
@@ -1454,8 +1454,8 @@ hit_next:
 		prealloc->end = this_end;
 		inserted_state = insert_state(tree, prealloc, bits, NULL);
 		if (IS_ERR(inserted_state)) {
-			err = PTR_ERR(inserted_state);
-			extent_io_tree_panic(tree, prealloc, "insert", err);
+			ret = PTR_ERR(inserted_state);
+			extent_io_tree_panic(tree, prealloc, "insert", ret);
 		}
 		cache_state(inserted_state, cached_state);
 		if (inserted_state == prealloc)
@@ -1472,13 +1472,13 @@ hit_next:
 	if (state->start <= end && state->end > end) {
 		prealloc = alloc_extent_state_atomic(prealloc);
 		if (!prealloc) {
-			err = -ENOMEM;
+			ret = -ENOMEM;
 			goto out;
 		}
 
-		err = split_state(tree, state, prealloc, end + 1);
-		if (err)
-			extent_io_tree_panic(tree, state, "split", err);
+		ret = split_state(tree, state, prealloc, end + 1);
+		if (ret)
+			extent_io_tree_panic(tree, state, "split", ret);
 
 		set_state_bits(tree, prealloc, bits, NULL);
 		cache_state(prealloc, cached_state);
@@ -1500,7 +1500,7 @@ out:
 	if (prealloc)
 		free_extent_state(prealloc);
 
-	return err;
+	return ret;
 }
 
 /*
@@ -1883,8 +1883,8 @@ void __cold extent_state_free_cachep(void)
 int __init extent_state_init_cachep(void)
 {
 	extent_state_cache = kmem_cache_create("btrfs_extent_state",
-			sizeof(struct extent_state), 0,
-			SLAB_MEM_SPREAD, NULL);
+					       sizeof(struct extent_state), 0, 0,
+					       NULL);
 	if (!extent_state_cache)
 		return -ENOMEM;
 
diff --git a/fs/btrfs/extent-io-tree.h b/fs/btrfs/extent-io-tree.h
index ebe6390d65..9d3a52d8f5 100644
--- a/fs/btrfs/extent-io-tree.h
+++ b/fs/btrfs/extent-io-tree.h
@@ -3,9 +3,16 @@
 #ifndef BTRFS_EXTENT_IO_TREE_H
 #define BTRFS_EXTENT_IO_TREE_H
 
+#include <linux/rbtree.h>
+#include <linux/spinlock.h>
+#include <linux/refcount.h>
+#include <linux/list.h>
+#include <linux/wait.h>
 #include "misc.h"
 
 struct extent_changeset;
+struct btrfs_fs_info;
+struct btrfs_inode;
 
 /* Bits for the extent state */
 enum {
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 8e8cc11112..844b677d05 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -18,7 +18,7 @@
 #include <linux/crc32c.h>
 #include "ctree.h"
 #include "extent-tree.h"
-#include "tree-log.h"
+#include "transaction.h"
 #include "disk-io.h"
 #include "print-tree.h"
 #include "volumes.h"
@@ -26,14 +26,11 @@
 #include "locking.h"
 #include "free-space-cache.h"
 #include "free-space-tree.h"
-#include "sysfs.h"
 #include "qgroup.h"
 #include "ref-verify.h"
 #include "space-info.h"
 #include "block-rsv.h"
-#include "delalloc-space.h"
 #include "discard.h"
-#include "rcu-string.h"
 #include "zoned.h"
 #include "dev-replace.h"
 #include "fs.h"
@@ -49,9 +46,7 @@
 
 static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 			       struct btrfs_delayed_ref_head *href,
-			       struct btrfs_delayed_ref_node *node, u64 parent,
-			       u64 root_objectid, u64 owner_objectid,
-			       u64 owner_offset,
+			       struct btrfs_delayed_ref_node *node,
 			       struct btrfs_delayed_extent_op *extra_op);
 static void __run_delayed_extent_op(struct btrfs_delayed_extent_op *extent_op,
 				    struct extent_buffer *leaf,
@@ -109,10 +104,7 @@ int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans,
 	struct btrfs_delayed_ref_head *head;
 	struct btrfs_delayed_ref_root *delayed_refs;
 	struct btrfs_path *path;
-	struct btrfs_extent_item *ei;
-	struct extent_buffer *leaf;
 	struct btrfs_key key;
-	u32 item_size;
 	u64 num_refs;
 	u64 extent_flags;
 	u64 owner = 0;
@@ -162,16 +154,11 @@ search_again:
 	}
 
 	if (ret == 0) {
-		leaf = path->nodes[0];
-		item_size = btrfs_item_size(leaf, path->slots[0]);
-		if (item_size >= sizeof(*ei)) {
-			ei = btrfs_item_ptr(leaf, path->slots[0],
-					    struct btrfs_extent_item);
-			num_refs = btrfs_extent_refs(leaf, ei);
-			extent_flags = btrfs_extent_flags(leaf, ei);
-			owner = btrfs_get_extent_owner_root(fs_info, leaf,
-							    path->slots[0]);
-		} else {
+		struct extent_buffer *leaf = path->nodes[0];
+		struct btrfs_extent_item *ei;
+		const u32 item_size = btrfs_item_size(leaf, path->slots[0]);
+
+		if (unlikely(item_size < sizeof(*ei))) {
 			ret = -EUCLEAN;
 			btrfs_err(fs_info,
 			"unexpected extent item size, has %u expect >= %zu",
@@ -184,6 +171,10 @@ search_again:
 			goto out_free;
 		}
 
+		ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
+		num_refs = btrfs_extent_refs(leaf, ei);
+		extent_flags = btrfs_extent_flags(leaf, ei);
+		owner = btrfs_get_extent_owner_root(fs_info, leaf, path->slots[0]);
 		BUG_ON(num_refs == 0);
 	} else {
 		num_refs = 0;
@@ -451,9 +442,8 @@ static noinline int lookup_extent_data_ref(struct btrfs_trans_handle *trans,
 	struct btrfs_extent_data_ref *ref;
 	struct extent_buffer *leaf;
 	u32 nritems;
-	int ret;
 	int recow;
-	int err = -ENOENT;
+	int ret;
 
 	key.objectid = bytenr;
 	if (parent) {
@@ -467,26 +457,26 @@ static noinline int lookup_extent_data_ref(struct btrfs_trans_handle *trans,
 again:
 	recow = 0;
 	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
-	if (ret < 0) {
-		err = ret;
-		goto fail;
-	}
+	if (ret < 0)
+		return ret;
 
 	if (parent) {
-		if (!ret)
-			return 0;
-		goto fail;
+		if (ret)
+			return -ENOENT;
+		return 0;
 	}
 
+	ret = -ENOENT;
 	leaf = path->nodes[0];
 	nritems = btrfs_header_nritems(leaf);
 	while (1) {
 		if (path->slots[0] >= nritems) {
 			ret = btrfs_next_leaf(root, path);
-			if (ret < 0)
-				err = ret;
-			if (ret)
-				goto fail;
+			if (ret) {
+				if (ret > 0)
+					return -ENOENT;
+				return ret;
+			}
 
 			leaf = path->nodes[0];
 			nritems = btrfs_header_nritems(leaf);
@@ -507,37 +497,37 @@ again:
 				btrfs_release_path(path);
 				goto again;
 			}
-			err = 0;
+			ret = 0;
 			break;
 		}
 		path->slots[0]++;
 	}
 fail:
-	return err;
+	return ret;
 }
 
 static noinline int insert_extent_data_ref(struct btrfs_trans_handle *trans,
 					   struct btrfs_path *path,
-					   u64 bytenr, u64 parent,
-					   u64 root_objectid, u64 owner,
-					   u64 offset, int refs_to_add)
+					   struct btrfs_delayed_ref_node *node,
+					   u64 bytenr)
 {
 	struct btrfs_root *root = btrfs_extent_root(trans->fs_info, bytenr);
 	struct btrfs_key key;
 	struct extent_buffer *leaf;
+	u64 owner = btrfs_delayed_ref_owner(node);
+	u64 offset = btrfs_delayed_ref_offset(node);
 	u32 size;
 	u32 num_refs;
 	int ret;
 
 	key.objectid = bytenr;
-	if (parent) {
+	if (node->parent) {
 		key.type = BTRFS_SHARED_DATA_REF_KEY;
-		key.offset = parent;
+		key.offset = node->parent;
 		size = sizeof(struct btrfs_shared_data_ref);
 	} else {
 		key.type = BTRFS_EXTENT_DATA_REF_KEY;
-		key.offset = hash_extent_data_ref(root_objectid,
-						  owner, offset);
+		key.offset = hash_extent_data_ref(node->ref_root, owner, offset);
 		size = sizeof(struct btrfs_extent_data_ref);
 	}
 
@@ -546,15 +536,15 @@ static noinline int insert_extent_data_ref(struct btrfs_trans_handle *trans,
 		goto fail;
 
 	leaf = path->nodes[0];
-	if (parent) {
+	if (node->parent) {
 		struct btrfs_shared_data_ref *ref;
 		ref = btrfs_item_ptr(leaf, path->slots[0],
 				     struct btrfs_shared_data_ref);
 		if (ret == 0) {
-			btrfs_set_shared_data_ref_count(leaf, ref, refs_to_add);
+			btrfs_set_shared_data_ref_count(leaf, ref, node->ref_mod);
 		} else {
 			num_refs = btrfs_shared_data_ref_count(leaf, ref);
-			num_refs += refs_to_add;
+			num_refs += node->ref_mod;
 			btrfs_set_shared_data_ref_count(leaf, ref, num_refs);
 		}
 	} else {
@@ -562,7 +552,7 @@ static noinline int insert_extent_data_ref(struct btrfs_trans_handle *trans,
 		while (ret == -EEXIST) {
 			ref = btrfs_item_ptr(leaf, path->slots[0],
 					     struct btrfs_extent_data_ref);
-			if (match_extent_data_ref(leaf, ref, root_objectid,
+			if (match_extent_data_ref(leaf, ref, node->ref_root,
 						  owner, offset))
 				break;
 			btrfs_release_path(path);
@@ -577,14 +567,13 @@ static noinline int insert_extent_data_ref(struct btrfs_trans_handle *trans,
 		ref = btrfs_item_ptr(leaf, path->slots[0],
 				     struct btrfs_extent_data_ref);
 		if (ret == 0) {
-			btrfs_set_extent_data_ref_root(leaf, ref,
-						       root_objectid);
+			btrfs_set_extent_data_ref_root(leaf, ref, node->ref_root);
 			btrfs_set_extent_data_ref_objectid(leaf, ref, owner);
 			btrfs_set_extent_data_ref_offset(leaf, ref, offset);
-			btrfs_set_extent_data_ref_count(leaf, ref, refs_to_add);
+			btrfs_set_extent_data_ref_count(leaf, ref, node->ref_mod);
 		} else {
 			num_refs = btrfs_extent_data_ref_count(leaf, ref);
-			num_refs += refs_to_add;
+			num_refs += node->ref_mod;
 			btrfs_set_extent_data_ref_count(leaf, ref, num_refs);
 		}
 	}
@@ -708,20 +697,20 @@ static noinline int lookup_tree_block_ref(struct btrfs_trans_handle *trans,
 
 static noinline int insert_tree_block_ref(struct btrfs_trans_handle *trans,
 					  struct btrfs_path *path,
-					  u64 bytenr, u64 parent,
-					  u64 root_objectid)
+					  struct btrfs_delayed_ref_node *node,
+					  u64 bytenr)
 {
 	struct btrfs_root *root = btrfs_extent_root(trans->fs_info, bytenr);
 	struct btrfs_key key;
 	int ret;
 
 	key.objectid = bytenr;
-	if (parent) {
+	if (node->parent) {
 		key.type = BTRFS_SHARED_BLOCK_REF_KEY;
-		key.offset = parent;
+		key.offset = node->parent;
 	} else {
 		key.type = BTRFS_TREE_BLOCK_REF_KEY;
-		key.offset = root_objectid;
+		key.offset = node->ref_root;
 	}
 
 	ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
@@ -1442,7 +1431,7 @@ int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
 	ASSERT(generic_ref->type != BTRFS_REF_NOT_SET &&
 	       generic_ref->action);
 	BUG_ON(generic_ref->type == BTRFS_REF_METADATA &&
-	       generic_ref->tree_ref.ref_root == BTRFS_TREE_LOG_OBJECTID);
+	       generic_ref->ref_root == BTRFS_TREE_LOG_OBJECTID);
 
 	if (generic_ref->type == BTRFS_REF_METADATA)
 		ret = btrfs_add_delayed_tree_ref(trans, generic_ref, NULL);
@@ -1465,34 +1454,12 @@ int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
  * @node:	    The delayed ref node used to get the bytenr/length for
  *		    extent whose references are incremented.
  *
- * @parent:	    If this is a shared extent (BTRFS_SHARED_DATA_REF_KEY/
- *		    BTRFS_SHARED_BLOCK_REF_KEY) then it holds the logical
- *		    bytenr of the parent block. Since new extents are always
- *		    created with indirect references, this will only be the case
- *		    when relocating a shared extent. In that case, root_objectid
- *		    will be BTRFS_TREE_RELOC_OBJECTID. Otherwise, parent must
- *		    be 0
- *
- * @root_objectid:  The id of the root where this modification has originated,
- *		    this can be either one of the well-known metadata trees or
- *		    the subvolume id which references this extent.
- *
- * @owner:	    For data extents it is the inode number of the owning file.
- *		    For metadata extents this parameter holds the level in the
- *		    tree of the extent.
- *
- * @offset:	    For metadata extents the offset is ignored and is currently
- *		    always passed as 0. For data extents it is the fileoffset
- *		    this extent belongs to.
- *
  * @extent_op       Pointer to a structure, holding information necessary when
  *                  updating a tree block's flags
  *
  */
 static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
 				  struct btrfs_delayed_ref_node *node,
-				  u64 parent, u64 root_objectid,
-				  u64 owner, u64 offset,
 				  struct btrfs_delayed_extent_op *extent_op)
 {
 	struct btrfs_path *path;
@@ -1501,6 +1468,8 @@ static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
 	struct btrfs_key key;
 	u64 bytenr = node->bytenr;
 	u64 num_bytes = node->num_bytes;
+	u64 owner = btrfs_delayed_ref_owner(node);
+	u64 offset = btrfs_delayed_ref_offset(node);
 	u64 refs;
 	int refs_to_add = node->ref_mod;
 	int ret;
@@ -1511,7 +1480,7 @@ static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
 
 	/* this will setup the path even if it fails to insert the back ref */
 	ret = insert_inline_extent_backref(trans, path, bytenr, num_bytes,
-					   parent, root_objectid, owner,
+					   node->parent, node->ref_root, owner,
 					   offset, refs_to_add, extent_op);
 	if ((ret < 0 && ret != -EAGAIN) || !ret)
 		goto out;
@@ -1534,12 +1503,9 @@ static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
 
 	/* now insert the actual backref */
 	if (owner < BTRFS_FIRST_FREE_OBJECTID)
-		ret = insert_tree_block_ref(trans, path, bytenr, parent,
-					    root_objectid);
+		ret = insert_tree_block_ref(trans, path, node, bytenr);
 	else
-		ret = insert_extent_data_ref(trans, path, bytenr, parent,
-					     root_objectid, owner, offset,
-					     refs_to_add);
+		ret = insert_extent_data_ref(trans, path, node, bytenr);
 
 	if (ret)
 		btrfs_abort_transaction(trans, ret);
@@ -1572,15 +1538,13 @@ static int run_delayed_data_ref(struct btrfs_trans_handle *trans,
 				bool insert_reserved)
 {
 	int ret = 0;
-	struct btrfs_delayed_data_ref *ref;
 	u64 parent = 0;
 	u64 flags = 0;
 
-	ref = btrfs_delayed_node_to_data_ref(node);
-	trace_run_delayed_data_ref(trans->fs_info, node, ref, node->action);
+	trace_run_delayed_data_ref(trans->fs_info, node);
 
 	if (node->type == BTRFS_SHARED_DATA_REF_KEY)
-		parent = ref->parent;
+		parent = node->parent;
 
 	if (node->action == BTRFS_ADD_DELAYED_REF && insert_reserved) {
 		struct btrfs_key key;
@@ -1591,6 +1555,8 @@ static int run_delayed_data_ref(struct btrfs_trans_handle *trans,
 			.is_inc	= true,
 			.generation = trans->transid,
 		};
+		u64 owner = btrfs_delayed_ref_owner(node);
+		u64 offset = btrfs_delayed_ref_offset(node);
 
 		if (extent_op)
 			flags |= extent_op->flags_to_set;
@@ -1599,21 +1565,17 @@ static int run_delayed_data_ref(struct btrfs_trans_handle *trans,
 		key.type = BTRFS_EXTENT_ITEM_KEY;
 		key.offset = node->num_bytes;
 
-		ret = alloc_reserved_file_extent(trans, parent, ref->root,
-						 flags, ref->objectid,
-						 ref->offset, &key,
-						 node->ref_mod, href->owning_root);
+		ret = alloc_reserved_file_extent(trans, parent, node->ref_root,
+						 flags, owner, offset, &key,
+						 node->ref_mod,
+						 href->owning_root);
 		free_head_ref_squota_rsv(trans->fs_info, href);
 		if (!ret)
 			ret = btrfs_record_squota_delta(trans->fs_info, &delta);
 	} else if (node->action == BTRFS_ADD_DELAYED_REF) {
-		ret = __btrfs_inc_extent_ref(trans, node, parent, ref->root,
-					     ref->objectid, ref->offset,
-					     extent_op);
+		ret = __btrfs_inc_extent_ref(trans, node, extent_op);
 	} else if (node->action == BTRFS_DROP_DELAYED_REF) {
-		ret = __btrfs_free_extent(trans, href, node, parent,
-					  ref->root, ref->objectid,
-					  ref->offset, extent_op);
+		ret = __btrfs_free_extent(trans, href, node, extent_op);
 	} else {
 		BUG();
 	}
@@ -1735,16 +1697,14 @@ static int run_delayed_tree_ref(struct btrfs_trans_handle *trans,
 {
 	int ret = 0;
 	struct btrfs_fs_info *fs_info = trans->fs_info;
-	struct btrfs_delayed_tree_ref *ref;
 	u64 parent = 0;
 	u64 ref_root = 0;
 
-	ref = btrfs_delayed_node_to_tree_ref(node);
-	trace_run_delayed_tree_ref(trans->fs_info, node, ref, node->action);
+	trace_run_delayed_tree_ref(trans->fs_info, node);
 
 	if (node->type == BTRFS_SHARED_BLOCK_REF_KEY)
-		parent = ref->parent;
-	ref_root = ref->root;
+		parent = node->parent;
+	ref_root = node->ref_root;
 
 	if (unlikely(node->ref_mod != 1)) {
 		btrfs_err(trans->fs_info,
@@ -1767,11 +1727,9 @@ static int run_delayed_tree_ref(struct btrfs_trans_handle *trans,
 		if (!ret)
 			btrfs_record_squota_delta(fs_info, &delta);
 	} else if (node->action == BTRFS_ADD_DELAYED_REF) {
-		ret = __btrfs_inc_extent_ref(trans, node, parent, ref_root,
-					     ref->level, 0, extent_op);
+		ret = __btrfs_inc_extent_ref(trans, node, extent_op);
 	} else if (node->action == BTRFS_DROP_DELAYED_REF) {
-		ret = __btrfs_free_extent(trans, href, node, parent, ref_root,
-					  ref->level, 0, extent_op);
+		ret = __btrfs_free_extent(trans, href, node, extent_op);
 	} else {
 		BUG();
 	}
@@ -2295,7 +2253,6 @@ static noinline int check_delayed_ref(struct btrfs_root *root,
 {
 	struct btrfs_delayed_ref_head *head;
 	struct btrfs_delayed_ref_node *ref;
-	struct btrfs_delayed_data_ref *data_ref;
 	struct btrfs_delayed_ref_root *delayed_refs;
 	struct btrfs_transaction *cur_trans;
 	struct rb_node *node;
@@ -2349,6 +2306,9 @@ static noinline int check_delayed_ref(struct btrfs_root *root,
 	 */
 	for (node = rb_first_cached(&head->ref_tree); node;
 	     node = rb_next(node)) {
+		u64 ref_owner;
+		u64 ref_offset;
+
 		ref = rb_entry(node, struct btrfs_delayed_ref_node, ref_node);
 		/* If it's a shared ref we know a cross reference exists */
 		if (ref->type != BTRFS_EXTENT_DATA_REF_KEY) {
@@ -2356,15 +2316,15 @@ static noinline int check_delayed_ref(struct btrfs_root *root,
 			break;
 		}
 
-		data_ref = btrfs_delayed_node_to_data_ref(ref);
+		ref_owner = btrfs_delayed_ref_owner(ref);
+		ref_offset = btrfs_delayed_ref_offset(ref);
 
 		/*
 		 * If our ref doesn't match the one we're currently looking at
 		 * then we have a cross reference.
 		 */
-		if (data_ref->root != root->root_key.objectid ||
-		    data_ref->objectid != objectid ||
-		    data_ref->offset != offset) {
+		if (ref->ref_root != btrfs_root_id(root) ||
+		    ref_owner != objectid || ref_offset != offset) {
 			ret = 1;
 			break;
 		}
@@ -2399,7 +2359,14 @@ static noinline int check_committed_ref(struct btrfs_root *root,
 	ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
 	if (ret < 0)
 		goto out;
-	BUG_ON(ret == 0); /* Corruption */
+	if (ret == 0) {
+		/*
+		 * Key with offset -1 found, there would have to exist an extent
+		 * item with such offset, but this is out of the valid range.
+		 */
+		ret = -EUCLEAN;
+		goto out;
+	}
 
 	ret = -ENOENT;
 	if (path->slots[0] == 0)
@@ -2450,8 +2417,7 @@ static noinline int check_committed_ref(struct btrfs_root *root,
 	ref = (struct btrfs_extent_data_ref *)(&iref->offset);
 	if (btrfs_extent_refs(leaf, ei) !=
 	    btrfs_extent_data_ref_count(leaf, ref) ||
-	    btrfs_extent_data_ref_root(leaf, ref) !=
-	    root->root_key.objectid ||
+	    btrfs_extent_data_ref_root(leaf, ref) != btrfs_root_id(root) ||
 	    btrfs_extent_data_ref_objectid(leaf, ref) != objectid ||
 	    btrfs_extent_data_ref_offset(leaf, ref) != offset)
 		goto out;
@@ -2488,14 +2454,11 @@ static int __btrfs_mod_ref(struct btrfs_trans_handle *trans,
 			   int full_backref, int inc)
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
-	u64 bytenr;
-	u64 num_bytes;
 	u64 parent;
 	u64 ref_root;
 	u32 nritems;
 	struct btrfs_key key;
 	struct btrfs_file_extent_item *fi;
-	struct btrfs_ref generic_ref = { 0 };
 	bool for_reloc = btrfs_header_flag(buf, BTRFS_HEADER_FLAG_RELOC);
 	int i;
 	int action;
@@ -2522,6 +2485,12 @@ static int __btrfs_mod_ref(struct btrfs_trans_handle *trans,
 		action = BTRFS_DROP_DELAYED_REF;
 
 	for (i = 0; i < nritems; i++) {
+		struct btrfs_ref ref = {
+			.action = action,
+			.parent = parent,
+			.ref_root = ref_root,
+		};
+
 		if (level == 0) {
 			btrfs_item_key_to_cpu(buf, &key, i);
 			if (key.type != BTRFS_EXTENT_DATA_KEY)
@@ -2531,35 +2500,33 @@ static int __btrfs_mod_ref(struct btrfs_trans_handle *trans,
 			if (btrfs_file_extent_type(buf, fi) ==
 			    BTRFS_FILE_EXTENT_INLINE)
 				continue;
-			bytenr = btrfs_file_extent_disk_bytenr(buf, fi);
-			if (bytenr == 0)
+			ref.bytenr = btrfs_file_extent_disk_bytenr(buf, fi);
+			if (ref.bytenr == 0)
 				continue;
 
-			num_bytes = btrfs_file_extent_disk_num_bytes(buf, fi);
+			ref.num_bytes = btrfs_file_extent_disk_num_bytes(buf, fi);
+			ref.owning_root = ref_root;
+
 			key.offset -= btrfs_file_extent_offset(buf, fi);
-			btrfs_init_generic_ref(&generic_ref, action, bytenr,
-					       num_bytes, parent, ref_root);
-			btrfs_init_data_ref(&generic_ref, ref_root, key.objectid,
-					    key.offset, root->root_key.objectid,
-					    for_reloc);
+			btrfs_init_data_ref(&ref, key.objectid, key.offset,
+					    btrfs_root_id(root), for_reloc);
 			if (inc)
-				ret = btrfs_inc_extent_ref(trans, &generic_ref);
+				ret = btrfs_inc_extent_ref(trans, &ref);
 			else
-				ret = btrfs_free_extent(trans, &generic_ref);
+				ret = btrfs_free_extent(trans, &ref);
 			if (ret)
 				goto fail;
 		} else {
-			bytenr = btrfs_node_blockptr(buf, i);
-			num_bytes = fs_info->nodesize;
-			/* We don't know the owning_root, use 0. */
-			btrfs_init_generic_ref(&generic_ref, action, bytenr,
-					       num_bytes, parent, 0);
-			btrfs_init_tree_ref(&generic_ref, level - 1, ref_root,
-					    root->root_key.objectid, for_reloc);
+			/* We don't know the owning_root, leave as 0. */
+			ref.bytenr = btrfs_node_blockptr(buf, i);
+			ref.num_bytes = fs_info->nodesize;
+
+			btrfs_init_tree_ref(&ref, level - 1,
+					    btrfs_root_id(root), for_reloc);
 			if (inc)
-				ret = btrfs_inc_extent_ref(trans, &generic_ref);
+				ret = btrfs_inc_extent_ref(trans, &ref);
 			else
-				ret = btrfs_free_extent(trans, &generic_ref);
+				ret = btrfs_free_extent(trans, &ref);
 			if (ret)
 				goto fail;
 		}
@@ -2780,6 +2747,7 @@ static int unpin_extent_range(struct btrfs_fs_info *fs_info,
 	u64 total_unpinned = 0;
 	u64 empty_cluster = 0;
 	bool readonly;
+	int ret = 0;
 
 	while (start <= end) {
 		readonly = false;
@@ -2789,7 +2757,11 @@ static int unpin_extent_range(struct btrfs_fs_info *fs_info,
 				btrfs_put_block_group(cache);
 			total_unpinned = 0;
 			cache = btrfs_lookup_block_group(fs_info, start);
-			BUG_ON(!cache); /* Logic error */
+			if (cache == NULL) {
+				/* Logic error, something removed the block group. */
+				ret = -EUCLEAN;
+				goto out;
+			}
 
 			cluster = fetch_cluster_info(fs_info,
 						     cache->space_info,
@@ -2830,7 +2802,8 @@ static int unpin_extent_range(struct btrfs_fs_info *fs_info,
 			readonly = true;
 		} else if (btrfs_is_zoned(fs_info)) {
 			/* Need reset before reusing in a zoned block group */
-			space_info->bytes_zone_unusable += len;
+			btrfs_space_info_update_bytes_zone_unusable(fs_info, space_info,
+								    len);
 			readonly = true;
 		}
 		spin_unlock(&cache->lock);
@@ -2858,7 +2831,8 @@ static int unpin_extent_range(struct btrfs_fs_info *fs_info,
 
 	if (cache)
 		btrfs_put_block_group(cache);
-	return 0;
+out:
+	return ret;
 }
 
 int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans)
@@ -2888,7 +2862,8 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans)
 						   end + 1 - start, NULL);
 
 		clear_extent_dirty(unpin, start, end, &cached_state);
-		unpin_extent_range(fs_info, start, end, true);
+		ret = unpin_extent_range(fs_info, start, end, true);
+		BUG_ON(ret);
 		mutex_unlock(&fs_info->unused_bg_unpin_mutex);
 		free_extent_state(cached_state);
 		cond_resched();
@@ -3088,9 +3063,7 @@ static int do_free_extent_accounting(struct btrfs_trans_handle *trans,
  */
 static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 			       struct btrfs_delayed_ref_head *href,
-			       struct btrfs_delayed_ref_node *node, u64 parent,
-			       u64 root_objectid, u64 owner_objectid,
-			       u64 owner_offset,
+			       struct btrfs_delayed_ref_node *node,
 			       struct btrfs_delayed_extent_op *extent_op)
 {
 	struct btrfs_fs_info *info = trans->fs_info;
@@ -3110,6 +3083,8 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 	u64 refs;
 	u64 bytenr = node->bytenr;
 	u64 num_bytes = node->num_bytes;
+	u64 owner_objectid = btrfs_delayed_ref_owner(node);
+	u64 owner_offset = btrfs_delayed_ref_offset(node);
 	bool skinny_metadata = btrfs_fs_incompat(info, SKINNY_METADATA);
 	u64 delayed_ref_root = href->owning_root;
 
@@ -3135,7 +3110,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 		skinny_metadata = false;
 
 	ret = lookup_extent_backref(trans, path, &iref, bytenr, num_bytes,
-				    parent, root_objectid, owner_objectid,
+				    node->parent, node->ref_root, owner_objectid,
 				    owner_offset);
 	if (ret == 0) {
 		/*
@@ -3237,7 +3212,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 	} else if (WARN_ON(ret == -ENOENT)) {
 		abort_and_dump(trans, path,
 "unable to find ref byte nr %llu parent %llu root %llu owner %llu offset %llu slot %d",
-			       bytenr, parent, root_objectid, owner_objectid,
+			       bytenr, node->parent, node->ref_root, owner_objectid,
 			       owner_offset, path->slots[0]);
 		goto out;
 	} else {
@@ -3441,29 +3416,42 @@ out_delayed_unlock:
 	return 0;
 }
 
-void btrfs_free_tree_block(struct btrfs_trans_handle *trans,
-			   u64 root_id,
-			   struct extent_buffer *buf,
-			   u64 parent, int last_ref)
+int btrfs_free_tree_block(struct btrfs_trans_handle *trans,
+			  u64 root_id,
+			  struct extent_buffer *buf,
+			  u64 parent, int last_ref)
 {
 	struct btrfs_fs_info *fs_info = trans->fs_info;
-	struct btrfs_ref generic_ref = { 0 };
 	struct btrfs_block_group *bg;
 	int ret;
 
-	btrfs_init_generic_ref(&generic_ref, BTRFS_DROP_DELAYED_REF,
-			       buf->start, buf->len, parent, btrfs_header_owner(buf));
-	btrfs_init_tree_ref(&generic_ref, btrfs_header_level(buf),
-			    root_id, 0, false);
-
 	if (root_id != BTRFS_TREE_LOG_OBJECTID) {
+		struct btrfs_ref generic_ref = {
+			.action = BTRFS_DROP_DELAYED_REF,
+			.bytenr = buf->start,
+			.num_bytes = buf->len,
+			.parent = parent,
+			.owning_root = btrfs_header_owner(buf),
+			.ref_root = root_id,
+		};
+
+		/*
+		 * Assert that the extent buffer is not cleared due to
+		 * EXTENT_BUFFER_ZONED_ZEROOUT. Please refer
+		 * btrfs_clear_buffer_dirty() and btree_csum_one_bio() for
+		 * detail.
+		 */
+		ASSERT(btrfs_header_bytenr(buf) != 0);
+
+		btrfs_init_tree_ref(&generic_ref, btrfs_header_level(buf), 0, false);
 		btrfs_ref_tree_mod(fs_info, &generic_ref);
 		ret = btrfs_add_delayed_tree_ref(trans, &generic_ref, NULL);
-		BUG_ON(ret); /* -ENOMEM */
+		if (ret < 0)
+			return ret;
 	}
 
 	if (!last_ref)
-		return;
+		return 0;
 
 	if (btrfs_header_generation(buf) != trans->transid)
 		goto out;
@@ -3520,6 +3508,7 @@ out:
 	 * matter anymore.
 	 */
 	clear_bit(EXTENT_BUFFER_CORRUPT, &buf->bflags);
+	return 0;
 }
 
 /* Can return -ENOMEM */
@@ -3535,11 +3524,8 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_ref *ref)
 	 * tree log blocks never actually go into the extent allocation
 	 * tree, just update pinning info and exit early.
 	 */
-	if ((ref->type == BTRFS_REF_METADATA &&
-	     ref->tree_ref.ref_root == BTRFS_TREE_LOG_OBJECTID) ||
-	    (ref->type == BTRFS_REF_DATA &&
-	     ref->data_ref.ref_root == BTRFS_TREE_LOG_OBJECTID)) {
-		btrfs_pin_extent(trans, ref->bytenr, ref->len, 1);
+	if (ref->ref_root == BTRFS_TREE_LOG_OBJECTID) {
+		btrfs_pin_extent(trans, ref->bytenr, ref->num_bytes, 1);
 		ret = 0;
 	} else if (ref->type == BTRFS_REF_METADATA) {
 		ret = btrfs_add_delayed_tree_ref(trans, ref, NULL);
@@ -3547,10 +3533,7 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_ref *ref)
 		ret = btrfs_add_delayed_data_ref(trans, ref, 0);
 	}
 
-	if (!((ref->type == BTRFS_REF_METADATA &&
-	       ref->tree_ref.ref_root == BTRFS_TREE_LOG_OBJECTID) ||
-	      (ref->type == BTRFS_REF_DATA &&
-	       ref->data_ref.ref_root == BTRFS_TREE_LOG_OBJECTID)))
+	if (ref->ref_root != BTRFS_TREE_LOG_OBJECTID)
 		btrfs_ref_tree_mod(fs_info, ref);
 
 	return ret;
@@ -4685,7 +4668,7 @@ int btrfs_reserve_extent(struct btrfs_root *root, u64 ram_bytes,
 	bool final_tried = num_bytes == min_alloc_size;
 	u64 flags;
 	int ret;
-	bool for_treelog = (root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID);
+	bool for_treelog = (btrfs_root_id(root) == BTRFS_TREE_LOG_OBJECTID);
 	bool for_data_reloc = (btrfs_is_data_reloc_root(root) && is_data);
 
 	flags = get_alloc_profile_by_root(root, is_data);
@@ -4879,16 +4862,16 @@ static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
 	struct btrfs_extent_inline_ref *iref;
 	struct btrfs_path *path;
 	struct extent_buffer *leaf;
-	struct btrfs_delayed_tree_ref *ref;
 	u32 size = sizeof(*extent_item) + sizeof(*iref);
 	u64 flags = extent_op->flags_to_set;
+	/* The owner of a tree block is the level. */
+	int level = btrfs_delayed_ref_owner(node);
 	bool skinny_metadata = btrfs_fs_incompat(fs_info, SKINNY_METADATA);
 
-	ref = btrfs_delayed_node_to_tree_ref(node);
-
 	extent_key.objectid = node->bytenr;
 	if (skinny_metadata) {
-		extent_key.offset = ref->level;
+		/* The owner of a tree block is the level. */
+		extent_key.offset = level;
 		extent_key.type = BTRFS_METADATA_ITEM_KEY;
 	} else {
 		extent_key.offset = node->num_bytes;
@@ -4921,18 +4904,18 @@ static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
 	} else {
 		block_info = (struct btrfs_tree_block_info *)(extent_item + 1);
 		btrfs_set_tree_block_key(leaf, block_info, &extent_op->key);
-		btrfs_set_tree_block_level(leaf, block_info, ref->level);
+		btrfs_set_tree_block_level(leaf, block_info, level);
 		iref = (struct btrfs_extent_inline_ref *)(block_info + 1);
 	}
 
 	if (node->type == BTRFS_SHARED_BLOCK_REF_KEY) {
 		btrfs_set_extent_inline_ref_type(leaf, iref,
 						 BTRFS_SHARED_BLOCK_REF_KEY);
-		btrfs_set_extent_inline_ref_offset(leaf, iref, ref->parent);
+		btrfs_set_extent_inline_ref_offset(leaf, iref, node->parent);
 	} else {
 		btrfs_set_extent_inline_ref_type(leaf, iref,
 						 BTRFS_TREE_BLOCK_REF_KEY);
-		btrfs_set_extent_inline_ref_offset(leaf, iref, ref->root);
+		btrfs_set_extent_inline_ref_offset(leaf, iref, node->ref_root);
 	}
 
 	btrfs_mark_buffer_dirty(trans, leaf);
@@ -4946,19 +4929,20 @@ int btrfs_alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
 				     u64 offset, u64 ram_bytes,
 				     struct btrfs_key *ins)
 {
-	struct btrfs_ref generic_ref = { 0 };
-	u64 root_objectid = root->root_key.objectid;
-	u64 owning_root = root_objectid;
+	struct btrfs_ref generic_ref = {
+		.action = BTRFS_ADD_DELAYED_EXTENT,
+		.bytenr = ins->objectid,
+		.num_bytes = ins->offset,
+		.owning_root = btrfs_root_id(root),
+		.ref_root = btrfs_root_id(root),
+	};
 
-	BUG_ON(root_objectid == BTRFS_TREE_LOG_OBJECTID);
+	ASSERT(generic_ref.ref_root != BTRFS_TREE_LOG_OBJECTID);
 
 	if (btrfs_is_data_reloc_root(root) && is_fstree(root->relocation_src_root))
-		owning_root = root->relocation_src_root;
+		generic_ref.owning_root = root->relocation_src_root;
 
-	btrfs_init_generic_ref(&generic_ref, BTRFS_ADD_DELAYED_EXTENT,
-			       ins->objectid, ins->offset, 0, owning_root);
-	btrfs_init_data_ref(&generic_ref, root_objectid, owner,
-			    offset, 0, false);
+	btrfs_init_data_ref(&generic_ref, owner, offset, 0, false);
 	btrfs_ref_tree_mod(root->fs_info, &generic_ref);
 
 	return btrfs_add_delayed_data_ref(trans, &generic_ref, ram_bytes);
@@ -5081,7 +5065,7 @@ btrfs_init_new_buffer(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 	 */
 	btrfs_set_buffer_lockdep_class(lockdep_owner, buf, level);
 
-	__btrfs_tree_lock(buf, nest);
+	btrfs_tree_lock_nested(buf, nest);
 	btrfs_clear_buffer_dirty(trans, buf);
 	clear_bit(EXTENT_BUFFER_STALE, &buf->bflags);
 	clear_bit(EXTENT_BUFFER_ZONED_ZEROOUT, &buf->bflags);
@@ -5096,7 +5080,7 @@ btrfs_init_new_buffer(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 	btrfs_set_header_owner(buf, owner);
 	write_extent_buffer_fsid(buf, fs_info->fs_devices->metadata_uuid);
 	write_extent_buffer_chunk_tree_uuid(buf, fs_info->chunk_tree_uuid);
-	if (root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID) {
+	if (btrfs_root_id(root) == BTRFS_TREE_LOG_OBJECTID) {
 		buf->log_index = root->log_transid % 2;
 		/*
 		 * we allow two log transactions at a time, use different
@@ -5137,7 +5121,6 @@ struct extent_buffer *btrfs_alloc_tree_block(struct btrfs_trans_handle *trans,
 	struct btrfs_block_rsv *block_rsv;
 	struct extent_buffer *buf;
 	struct btrfs_delayed_extent_op *extent_op;
-	struct btrfs_ref generic_ref = { 0 };
 	u64 flags = 0;
 	int ret;
 	u32 blocksize = fs_info->nodesize;
@@ -5180,6 +5163,14 @@ struct extent_buffer *btrfs_alloc_tree_block(struct btrfs_trans_handle *trans,
 		BUG_ON(parent > 0);
 
 	if (root_objectid != BTRFS_TREE_LOG_OBJECTID) {
+		struct btrfs_ref generic_ref = {
+			.action = BTRFS_ADD_DELAYED_EXTENT,
+			.bytenr = ins.objectid,
+			.num_bytes = ins.offset,
+			.parent = parent,
+			.owning_root = owning_root,
+			.ref_root = root_objectid,
+		};
 		extent_op = btrfs_alloc_delayed_extent_op();
 		if (!extent_op) {
 			ret = -ENOMEM;
@@ -5194,10 +5185,7 @@ struct extent_buffer *btrfs_alloc_tree_block(struct btrfs_trans_handle *trans,
 		extent_op->update_flags = true;
 		extent_op->level = level;
 
-		btrfs_init_generic_ref(&generic_ref, BTRFS_ADD_DELAYED_EXTENT,
-				       ins.objectid, ins.offset, parent, owning_root);
-		btrfs_init_tree_ref(&generic_ref, level, root_objectid,
-				    root->root_key.objectid, false);
+		btrfs_init_tree_ref(&generic_ref, level, btrfs_root_id(root), false);
 		btrfs_ref_tree_mod(fs_info, &generic_ref);
 		ret = btrfs_add_delayed_tree_ref(trans, &generic_ref, extent_op);
 		if (ret)
@@ -5335,8 +5323,7 @@ static noinline int walk_down_proc(struct btrfs_trans_handle *trans,
 	u64 flag = BTRFS_BLOCK_FLAG_FULL_BACKREF;
 	int ret;
 
-	if (wc->stage == UPDATE_BACKREF &&
-	    btrfs_header_owner(eb) != root->root_key.objectid)
+	if (wc->stage == UPDATE_BACKREF && btrfs_header_owner(eb) != btrfs_root_id(root))
 		return 1;
 
 	/*
@@ -5410,7 +5397,7 @@ static int check_ref_exists(struct btrfs_trans_handle *trans,
 
 	ret = lookup_extent_backref(trans, path, &iref, bytenr,
 				    root->fs_info->nodesize, parent,
-				    root->root_key.objectid, level, 0);
+				    btrfs_root_id(root), level, 0);
 	btrfs_free_path(path);
 	if (ret == -ENOENT)
 		return 0;
@@ -5440,11 +5427,9 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans,
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	u64 bytenr;
 	u64 generation;
-	u64 parent;
 	u64 owner_root = 0;
 	struct btrfs_tree_parent_check check = { 0 };
 	struct btrfs_key key;
-	struct btrfs_ref ref = { 0 };
 	struct extent_buffer *next;
 	int level = wc->level;
 	int reada = 0;
@@ -5468,7 +5453,7 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans,
 
 	check.level = level - 1;
 	check.transid = generation;
-	check.owner_root = root->root_key.objectid;
+	check.owner_root = btrfs_root_id(root);
 	check.has_first_key = true;
 	btrfs_node_key_to_cpu(path->nodes[level], &check.first_key,
 			      path->slots[level]);
@@ -5476,7 +5461,7 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans,
 	next = find_extent_buffer(fs_info, bytenr);
 	if (!next) {
 		next = btrfs_find_create_tree_block(fs_info, bytenr,
-				root->root_key.objectid, level - 1);
+				btrfs_root_id(root), level - 1);
 		if (IS_ERR(next))
 			return PTR_ERR(next);
 		reada = 1;
@@ -5561,19 +5546,25 @@ skip:
 	wc->refs[level - 1] = 0;
 	wc->flags[level - 1] = 0;
 	if (wc->stage == DROP_REFERENCE) {
+		struct btrfs_ref ref = {
+			.action = BTRFS_DROP_DELAYED_REF,
+			.bytenr = bytenr,
+			.num_bytes = fs_info->nodesize,
+			.owning_root = owner_root,
+			.ref_root = btrfs_root_id(root),
+		};
 		if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF) {
-			parent = path->nodes[level]->start;
+			ref.parent = path->nodes[level]->start;
 		} else {
-			ASSERT(root->root_key.objectid ==
+			ASSERT(btrfs_root_id(root) ==
 			       btrfs_header_owner(path->nodes[level]));
-			if (root->root_key.objectid !=
+			if (btrfs_root_id(root) !=
 			    btrfs_header_owner(path->nodes[level])) {
 				btrfs_err(root->fs_info,
 						"mismatched block owner");
 				ret = -EIO;
 				goto out_unlock;
 			}
-			parent = 0;
 		}
 
 		/*
@@ -5583,7 +5574,7 @@ skip:
 		 * ->restarted flag.
 		 */
 		if (wc->restarted) {
-			ret = check_ref_exists(trans, root, bytenr, parent,
+			ret = check_ref_exists(trans, root, bytenr, ref.parent,
 					       level - 1);
 			if (ret < 0)
 				goto out_unlock;
@@ -5598,8 +5589,7 @@ skip:
 		 * already accounted them at merge time (replace_path),
 		 * thus we could skip expensive subtree trace here.
 		 */
-		if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID &&
-		    need_account) {
+		if (btrfs_root_id(root) != BTRFS_TREE_RELOC_OBJECTID && need_account) {
 			ret = btrfs_qgroup_trace_subtree(trans, next,
 							 generation, level - 1);
 			if (ret) {
@@ -5618,10 +5608,7 @@ skip:
 		wc->drop_level = level;
 		find_next_key(path, level, &wc->drop_progress);
 
-		btrfs_init_generic_ref(&ref, BTRFS_DROP_DELAYED_REF, bytenr,
-				       fs_info->nodesize, parent, owner_root);
-		btrfs_init_tree_ref(&ref, level - 1, root->root_key.objectid,
-				    0, false);
+		btrfs_init_tree_ref(&ref, level - 1, 0, false);
 		ret = btrfs_free_extent(trans, &ref);
 		if (ret)
 			goto out_unlock;
@@ -5655,7 +5642,7 @@ static noinline int walk_up_proc(struct btrfs_trans_handle *trans,
 				 struct walk_control *wc)
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
-	int ret;
+	int ret = 0;
 	int level = wc->level;
 	struct extent_buffer *eb = path->nodes[level];
 	u64 parent = 0;
@@ -5712,7 +5699,7 @@ static noinline int walk_up_proc(struct btrfs_trans_handle *trans,
 			else
 				ret = btrfs_dec_ref(trans, root, eb, 0);
 			BUG_ON(ret); /* -ENOMEM */
-			if (is_fstree(root->root_key.objectid)) {
+			if (is_fstree(btrfs_root_id(root))) {
 				ret = btrfs_qgroup_trace_leaf_items(trans, eb);
 				if (ret) {
 					btrfs_err_rl(fs_info,
@@ -5732,26 +5719,28 @@ static noinline int walk_up_proc(struct btrfs_trans_handle *trans,
 	if (eb == root->node) {
 		if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF)
 			parent = eb->start;
-		else if (root->root_key.objectid != btrfs_header_owner(eb))
+		else if (btrfs_root_id(root) != btrfs_header_owner(eb))
 			goto owner_mismatch;
 	} else {
 		if (wc->flags[level + 1] & BTRFS_BLOCK_FLAG_FULL_BACKREF)
 			parent = path->nodes[level + 1]->start;
-		else if (root->root_key.objectid !=
+		else if (btrfs_root_id(root) !=
 			 btrfs_header_owner(path->nodes[level + 1]))
 			goto owner_mismatch;
 	}
 
-	btrfs_free_tree_block(trans, btrfs_root_id(root), eb, parent,
-			      wc->refs[level] == 1);
+	ret = btrfs_free_tree_block(trans, btrfs_root_id(root), eb, parent,
+				    wc->refs[level] == 1);
+	if (ret < 0)
+		btrfs_abort_transaction(trans, ret);
 out:
 	wc->refs[level] = 0;
 	wc->flags[level] = 0;
-	return 0;
+	return ret;
 
 owner_mismatch:
 	btrfs_err_rl(fs_info, "unexpected tree owner, have %llu expect %llu",
-		     btrfs_header_owner(eb), root->root_key.objectid);
+		     btrfs_header_owner(eb), btrfs_root_id(root));
 	return -EUCLEAN;
 }
 
@@ -5837,8 +5826,7 @@ static noinline int walk_up_tree(struct btrfs_trans_handle *trans,
  */
 int btrfs_drop_snapshot(struct btrfs_root *root, int update_ref, int for_reloc)
 {
-	const bool is_reloc_root = (root->root_key.objectid ==
-				    BTRFS_TREE_RELOC_OBJECTID);
+	const bool is_reloc_root = (btrfs_root_id(root) == BTRFS_TREE_RELOC_OBJECTID);
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct btrfs_path *path;
 	struct btrfs_trans_handle *trans;
@@ -5852,7 +5840,7 @@ int btrfs_drop_snapshot(struct btrfs_root *root, int update_ref, int for_reloc)
 	bool root_dropped = false;
 	bool unfinished_drop = false;
 
-	btrfs_debug(fs_info, "Drop subvolume %llu", root->root_key.objectid);
+	btrfs_debug(fs_info, "Drop subvolume %llu", btrfs_root_id(root));
 
 	path = btrfs_alloc_path();
 	if (!path) {
@@ -6050,8 +6038,7 @@ int btrfs_drop_snapshot(struct btrfs_root *root, int update_ref, int for_reloc)
 			 *
 			 * The most common failure here is just -ENOENT.
 			 */
-			btrfs_del_orphan_item(trans, tree_root,
-					      root->root_key.objectid);
+			btrfs_del_orphan_item(trans, tree_root, btrfs_root_id(root));
 		}
 	}
 
@@ -6113,9 +6100,8 @@ int btrfs_drop_subtree(struct btrfs_trans_handle *trans,
 	int level;
 	int parent_level;
 	int ret = 0;
-	int wret;
 
-	BUG_ON(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID);
+	BUG_ON(btrfs_root_id(root) != BTRFS_TREE_RELOC_OBJECTID);
 
 	path = btrfs_alloc_path();
 	if (!path)
@@ -6149,17 +6135,16 @@ int btrfs_drop_subtree(struct btrfs_trans_handle *trans,
 	wc->reada_count = BTRFS_NODEPTRS_PER_BLOCK(fs_info);
 
 	while (1) {
-		wret = walk_down_tree(trans, root, path, wc);
-		if (wret < 0) {
-			ret = wret;
+		ret = walk_down_tree(trans, root, path, wc);
+		if (ret < 0)
 			break;
-		}
 
-		wret = walk_up_tree(trans, root, path, wc, parent_level);
-		if (wret < 0)
-			ret = wret;
-		if (wret != 0)
+		ret = walk_up_tree(trans, root, path, wc, parent_level);
+		if (ret) {
+			if (ret > 0)
+				ret = 0;
 			break;
+		}
 	}
 
 	kfree(wc);
@@ -6167,10 +6152,13 @@ int btrfs_drop_subtree(struct btrfs_trans_handle *trans,
 	return ret;
 }
 
-int btrfs_error_unpin_extent_range(struct btrfs_fs_info *fs_info,
-				   u64 start, u64 end)
+/*
+ * Unpin the extent range in an error context and don't add the space back.
+ * Errors are not propagated further.
+ */
+void btrfs_error_unpin_extent_range(struct btrfs_fs_info *fs_info, u64 start, u64 end)
 {
-	return unpin_extent_range(fs_info, start, end, false);
+	unpin_extent_range(fs_info, start, end, false);
 }
 
 /*
diff --git a/fs/btrfs/extent-tree.h b/fs/btrfs/extent-tree.h
index 2e066035cc..2ad51130c0 100644
--- a/fs/btrfs/extent-tree.h
+++ b/fs/btrfs/extent-tree.h
@@ -3,11 +3,21 @@
 #ifndef BTRFS_EXTENT_TREE_H
 #define BTRFS_EXTENT_TREE_H
 
+#include <linux/types.h>
 #include "misc.h"
 #include "block-group.h"
+#include "locking.h"
 
+struct extent_buffer;
 struct btrfs_free_cluster;
+struct btrfs_fs_info;
+struct btrfs_root;
+struct btrfs_path;
+struct btrfs_ref;
+struct btrfs_disk_key;
 struct btrfs_delayed_ref_head;
+struct btrfs_delayed_ref_root;
+struct btrfs_extent_inline_ref;
 
 enum btrfs_extent_allocation_policy {
 	BTRFS_EXTENT_ALLOC_CLUSTERED,
@@ -117,10 +127,10 @@ struct extent_buffer *btrfs_alloc_tree_block(struct btrfs_trans_handle *trans,
 					     u64 empty_size,
 					     u64 reloc_src_root,
 					     enum btrfs_lock_nesting nest);
-void btrfs_free_tree_block(struct btrfs_trans_handle *trans,
-			   u64 root_id,
-			   struct extent_buffer *buf,
-			   u64 parent, int last_ref);
+int btrfs_free_tree_block(struct btrfs_trans_handle *trans,
+			  u64 root_id,
+			  struct extent_buffer *buf,
+			  u64 parent, int last_ref);
 int btrfs_alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
 				     struct btrfs_root *root, u64 owner,
 				     u64 offset, u64 ram_bytes,
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index b5175b9929..0486b1f911 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -14,7 +14,6 @@
 #include <linux/pagevec.h>
 #include <linux/prefetch.h>
 #include <linux/fsverity.h>
-#include "misc.h"
 #include "extent_io.h"
 #include "extent-io-tree.h"
 #include "extent_map.h"
@@ -22,7 +21,6 @@
 #include "btrfs_inode.h"
 #include "bio.h"
 #include "locking.h"
-#include "rcu-string.h"
 #include "backref.h"
 #include "disk-io.h"
 #include "subpage.h"
@@ -78,10 +76,11 @@ void btrfs_extent_buffer_leak_debug_check(struct btrfs_fs_info *fs_info)
 		eb = list_first_entry(&fs_info->allocated_ebs,
 				      struct extent_buffer, leak_list);
 		pr_err(
-	"BTRFS: buffer leak start %llu len %lu refs %d bflags %lu owner %llu\n",
+	"BTRFS: buffer leak start %llu len %u refs %d bflags %lu owner %llu\n",
 		       eb->start, eb->len, atomic_read(&eb->refs), eb->bflags,
 		       btrfs_header_owner(eb));
 		list_del(&eb->leak_list);
+		WARN_ON_ONCE(1);
 		kmem_cache_free(extent_buffer_cache, eb);
 	}
 	spin_unlock_irqrestore(&fs_info->eb_leak_lock, flags);
@@ -147,8 +146,8 @@ static void submit_write_bio(struct btrfs_bio_ctrl *bio_ctrl, int ret)
 int __init extent_buffer_init_cachep(void)
 {
 	extent_buffer_cache = kmem_cache_create("btrfs_extent_buffer",
-			sizeof(struct extent_buffer), 0,
-			SLAB_MEM_SPREAD, NULL);
+						sizeof(struct extent_buffer), 0, 0,
+						NULL);
 	if (!extent_buffer_cache)
 		return -ENOMEM;
 
@@ -397,15 +396,14 @@ again:
 	/* then test to make sure it is all still delalloc */
 	ret = test_range_bit(tree, delalloc_start, delalloc_end,
 			     EXTENT_DELALLOC, cached_state);
+
+	unlock_extent(tree, delalloc_start, delalloc_end, &cached_state);
 	if (!ret) {
-		unlock_extent(tree, delalloc_start, delalloc_end,
-			      &cached_state);
 		__unlock_for_delalloc(inode, locked_page,
 			      delalloc_start, delalloc_end);
 		cond_resched();
 		goto again;
 	}
-	free_extent_state(cached_state);
 	*start = delalloc_start;
 	*end = delalloc_end;
 out_failed:
@@ -414,9 +412,10 @@ out_failed:
 
 void extent_clear_unlock_delalloc(struct btrfs_inode *inode, u64 start, u64 end,
 				  struct page *locked_page,
+				  struct extent_state **cached,
 				  u32 clear_bits, unsigned long page_ops)
 {
-	clear_extent_bit(&inode->io_tree, start, end, clear_bits, NULL);
+	clear_extent_bit(&inode->io_tree, start, end, clear_bits, cached);
 
 	__process_pages_contig(inode->vfs_inode.i_mapping, locked_page,
 			       start, end, page_ops);
@@ -462,16 +461,15 @@ static void end_page_read(struct page *page, bool uptodate, u64 start, u32 len)
  */
 static void end_bbio_data_write(struct btrfs_bio *bbio)
 {
+	struct btrfs_fs_info *fs_info = bbio->fs_info;
 	struct bio *bio = &bbio->bio;
 	int error = blk_status_to_errno(bio->bi_status);
 	struct folio_iter fi;
+	const u32 sectorsize = fs_info->sectorsize;
 
 	ASSERT(!bio_flagged(bio, BIO_CLONED));
 	bio_for_each_folio_all(fi, bio) {
 		struct folio *folio = fi.folio;
-		struct inode *inode = folio->mapping->host;
-		struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
-		const u32 sectorsize = fs_info->sectorsize;
 		u64 start = folio_pos(folio) + fi.offset;
 		u32 len = fi.length;
 
@@ -593,22 +591,17 @@ static void begin_page_read(struct btrfs_fs_info *fs_info, struct page *page)
  */
 static void end_bbio_data_read(struct btrfs_bio *bbio)
 {
+	struct btrfs_fs_info *fs_info = bbio->fs_info;
 	struct bio *bio = &bbio->bio;
 	struct processed_extent processed = { 0 };
 	struct folio_iter fi;
-	/*
-	 * The offset to the beginning of a bio, since one bio can never be
-	 * larger than UINT_MAX, u32 here is enough.
-	 */
-	u32 bio_offset = 0;
+	const u32 sectorsize = fs_info->sectorsize;
 
 	ASSERT(!bio_flagged(bio, BIO_CLONED));
 	bio_for_each_folio_all(fi, &bbio->bio) {
 		bool uptodate = !bio->bi_status;
 		struct folio *folio = fi.folio;
 		struct inode *inode = folio->mapping->host;
-		struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
-		const u32 sectorsize = fs_info->sectorsize;
 		u64 start;
 		u64 end;
 		u32 len;
@@ -667,10 +660,6 @@ static void end_bbio_data_read(struct btrfs_bio *bbio)
 		end_page_read(folio_page(folio, 0), uptodate, start, len);
 		endio_readpage_release_extent(&processed, BTRFS_I(inode),
 					      start, end, uptodate);
-
-		ASSERT(bio_offset + len > bio_offset);
-		bio_offset += len;
-
 	}
 	/* Release the last extent */
 	endio_readpage_release_extent(&processed, NULL, 0, 0, false);
@@ -678,6 +667,37 @@ static void end_bbio_data_read(struct btrfs_bio *bbio)
 }
 
 /*
+ * Populate every free slot in a provided array with folios.
+ *
+ * @nr_folios:   number of folios to allocate
+ * @folio_array: the array to fill with folios; any existing non-NULL entries in
+ *		 the array will be skipped
+ * @extra_gfp:	 the extra GFP flags for the allocation
+ *
+ * Return: 0        if all folios were able to be allocated;
+ *         -ENOMEM  otherwise, the partially allocated folios would be freed and
+ *                  the array slots zeroed
+ */
+int btrfs_alloc_folio_array(unsigned int nr_folios, struct folio **folio_array,
+			    gfp_t extra_gfp)
+{
+	for (int i = 0; i < nr_folios; i++) {
+		if (folio_array[i])
+			continue;
+		folio_array[i] = folio_alloc(GFP_NOFS | extra_gfp, 0);
+		if (!folio_array[i])
+			goto error;
+	}
+	return 0;
+error:
+	for (int i = 0; i < nr_folios; i++) {
+		if (folio_array[i])
+			folio_put(folio_array[i]);
+	}
+	return -ENOMEM;
+}
+
+/*
  * Populate every free slot in a provided array with pages.
  *
  * @nr_pages:   number of pages to allocate
@@ -728,6 +748,8 @@ static int alloc_eb_folio_array(struct extent_buffer *eb, gfp_t extra_gfp)
 
 	for (int i = 0; i < num_pages; i++)
 		eb->folios[i] = page_folio(page_array[i]);
+	eb->folio_size = PAGE_SIZE;
+	eb->folio_shift = PAGE_SHIFT;
 	return 0;
 }
 
@@ -964,13 +986,14 @@ void clear_page_extent_mapped(struct page *page)
 	folio_detach_private(folio);
 }
 
-static struct extent_map *
-__get_extent_map(struct inode *inode, struct page *page, size_t pg_offset,
+static struct extent_map *__get_extent_map(struct inode *inode, struct page *page,
 		 u64 start, u64 len, struct extent_map **em_cached)
 {
 	struct extent_map *em;
 
-	if (em_cached && *em_cached) {
+	ASSERT(em_cached);
+
+	if (*em_cached) {
 		em = *em_cached;
 		if (extent_map_in_tree(em) && start >= em->start &&
 		    start < extent_map_end(em)) {
@@ -982,8 +1005,8 @@ __get_extent_map(struct inode *inode, struct page *page, size_t pg_offset,
 		*em_cached = NULL;
 	}
 
-	em = btrfs_get_extent(BTRFS_I(inode), page, pg_offset, start, len);
-	if (em_cached && !IS_ERR(em)) {
+	em = btrfs_get_extent(BTRFS_I(inode), page, start, len);
+	if (!IS_ERR(em)) {
 		BUG_ON(*em_cached);
 		refcount_inc(&em->refs);
 		*em_cached = em;
@@ -1045,8 +1068,7 @@ static int btrfs_do_readpage(struct page *page, struct extent_map **em_cached,
 			end_page_read(page, true, cur, iosize);
 			break;
 		}
-		em = __get_extent_map(inode, page, pg_offset, cur,
-				      end - cur + 1, em_cached);
+		em = __get_extent_map(inode, page, cur, end - cur + 1, em_cached);
 		if (IS_ERR(em)) {
 			unlock_extent(tree, cur, end, NULL);
 			end_page_read(page, false, cur, end + 1 - cur);
@@ -1155,11 +1177,14 @@ int btrfs_read_folio(struct file *file, struct folio *folio)
 	u64 start = page_offset(page);
 	u64 end = start + PAGE_SIZE - 1;
 	struct btrfs_bio_ctrl bio_ctrl = { .opf = REQ_OP_READ };
+	struct extent_map *em_cached = NULL;
 	int ret;
 
 	btrfs_lock_and_flush_ordered_range(inode, start, end, NULL);
 
-	ret = btrfs_do_readpage(page, NULL, &bio_ctrl, NULL);
+	ret = btrfs_do_readpage(page, &em_cached, &bio_ctrl, NULL);
+	free_extent_map(em_cached);
+
 	/*
 	 * If btrfs_do_readpage() failed we will want to submit the assembled
 	 * bio to do the cleanup.
@@ -1177,6 +1202,8 @@ static inline void contiguous_readpages(struct page *pages[], int nr_pages,
 	struct btrfs_inode *inode = page_to_inode(pages[0]);
 	int index;
 
+	ASSERT(em_cached);
+
 	btrfs_lock_and_flush_ordered_range(inode, start, end, NULL);
 
 	for (index = 0; index < nr_pages; index++) {
@@ -1365,7 +1392,7 @@ static noinline_for_stack int __extent_writepage_io(struct btrfs_inode *inode,
 			continue;
 		}
 
-		em = btrfs_get_extent(inode, NULL, 0, cur, len);
+		em = btrfs_get_extent(inode, NULL, cur, len);
 		if (IS_ERR(em)) {
 			ret = PTR_ERR_OR_ZERO(em);
 			goto out_error;
@@ -1575,7 +1602,7 @@ static void set_btree_ioerr(struct extent_buffer *eb)
 	 * can be no longer dirty nor marked anymore for writeback (if a
 	 * subsequent modification to the extent buffer didn't happen before the
 	 * transaction commit), which makes filemap_fdata[write|wait]_range not
-	 * able to find the pages tagged with SetPageError at transaction
+	 * able to find the pages which contain errors at transaction
 	 * commit time. So if this happens we must abort the transaction,
 	 * otherwise we commit a super block with btree roots that point to
 	 * btree nodes/leafs whose content on disk is invalid - either garbage
@@ -1733,10 +1760,10 @@ static noinline_for_stack void write_one_eb(struct extent_buffer *eb,
 			folio_lock(folio);
 			folio_clear_dirty_for_io(folio);
 			folio_start_writeback(folio);
-			ret = bio_add_folio(&bbio->bio, folio, folio_size(folio), 0);
+			ret = bio_add_folio(&bbio->bio, folio, eb->folio_size, 0);
 			ASSERT(ret);
 			wbc_account_cgroup_owner(wbc, folio_page(folio, 0),
-						 folio_size(folio));
+						 eb->folio_size);
 			wbc->nr_to_write -= folio_nr_pages(folio);
 			folio_unlock(folio);
 		}
@@ -2219,10 +2246,8 @@ void extent_write_locked_range(struct inode *inode, struct page *locked_page,
 
 		page = find_get_page(mapping, cur >> PAGE_SHIFT);
 		ASSERT(PageLocked(page));
-		if (pages_dirty && page != locked_page) {
+		if (pages_dirty && page != locked_page)
 			ASSERT(PageDirty(page));
-			clear_page_dirty_for_io(page);
-		}
 
 		ret = __extent_writepage_io(BTRFS_I(inode), page, &bio_ctrl,
 					    i_size, &nr);
@@ -2250,8 +2275,7 @@ next_page:
 	submit_write_bio(&bio_ctrl, found_error ? ret : 0);
 }
 
-int extent_writepages(struct address_space *mapping,
-		      struct writeback_control *wbc)
+int btrfs_writepages(struct address_space *mapping, struct writeback_control *wbc)
 {
 	struct inode *inode = mapping->host;
 	int ret = 0;
@@ -2271,7 +2295,7 @@ int extent_writepages(struct address_space *mapping,
 	return ret;
 }
 
-void extent_readahead(struct readahead_control *rac)
+void btrfs_readahead(struct readahead_control *rac)
 {
 	struct btrfs_bio_ctrl bio_ctrl = { .opf = REQ_OP_READ | REQ_RAHEAD };
 	struct page *pagepool[16];
@@ -2329,19 +2353,20 @@ int extent_invalidate_folio(struct extent_io_tree *tree,
  * are locked or under IO and drops the related state bits if it is safe
  * to drop the page.
  */
-static int try_release_extent_state(struct extent_io_tree *tree,
+static bool try_release_extent_state(struct extent_io_tree *tree,
 				    struct page *page, gfp_t mask)
 {
 	u64 start = page_offset(page);
 	u64 end = start + PAGE_SIZE - 1;
-	int ret = 1;
+	bool ret;
 
 	if (test_range_bit_exists(tree, start, end, EXTENT_LOCKED)) {
-		ret = 0;
+		ret = false;
 	} else {
 		u32 clear_bits = ~(EXTENT_LOCKED | EXTENT_NODATASUM |
 				   EXTENT_DELALLOC_NEW | EXTENT_CTLBITS |
 				   EXTENT_QGROUP_RESERVED);
+		int ret2;
 
 		/*
 		 * At this point we can safely clear everything except the
@@ -2349,15 +2374,15 @@ static int try_release_extent_state(struct extent_io_tree *tree,
 		 * The delalloc new bit will be cleared by ordered extent
 		 * completion.
 		 */
-		ret = __clear_extent_bit(tree, start, end, clear_bits, NULL, NULL);
+		ret2 = __clear_extent_bit(tree, start, end, clear_bits, NULL, NULL);
 
 		/* if clear_extent_bit failed for enomem reasons,
 		 * we can't allow the release to continue.
 		 */
-		if (ret < 0)
-			ret = 0;
+		if (ret2 < 0)
+			ret = false;
 		else
-			ret = 1;
+			ret = true;
 	}
 	return ret;
 }
@@ -2367,84 +2392,80 @@ static int try_release_extent_state(struct extent_io_tree *tree,
  * in the range corresponding to the page, both state records and extent
  * map records are removed
  */
-int try_release_extent_mapping(struct page *page, gfp_t mask)
+bool try_release_extent_mapping(struct page *page, gfp_t mask)
 {
-	struct extent_map *em;
 	u64 start = page_offset(page);
 	u64 end = start + PAGE_SIZE - 1;
-	struct btrfs_inode *btrfs_inode = page_to_inode(page);
-	struct extent_io_tree *tree = &btrfs_inode->io_tree;
-	struct extent_map_tree *map = &btrfs_inode->extent_tree;
-
-	if (gfpflags_allow_blocking(mask) &&
-	    page->mapping->host->i_size > SZ_16M) {
-		u64 len;
-		while (start <= end) {
-			struct btrfs_fs_info *fs_info;
-			u64 cur_gen;
-
-			len = end - start + 1;
-			write_lock(&map->lock);
-			em = lookup_extent_mapping(map, start, len);
-			if (!em) {
-				write_unlock(&map->lock);
-				break;
-			}
-			if ((em->flags & EXTENT_FLAG_PINNED) ||
-			    em->start != start) {
-				write_unlock(&map->lock);
-				free_extent_map(em);
-				break;
-			}
-			if (test_range_bit_exists(tree, em->start,
-						  extent_map_end(em) - 1,
-						  EXTENT_LOCKED))
-				goto next;
-			/*
-			 * If it's not in the list of modified extents, used
-			 * by a fast fsync, we can remove it. If it's being
-			 * logged we can safely remove it since fsync took an
-			 * extra reference on the em.
-			 */
-			if (list_empty(&em->list) ||
-			    (em->flags & EXTENT_FLAG_LOGGING))
-				goto remove_em;
-			/*
-			 * If it's in the list of modified extents, remove it
-			 * only if its generation is older then the current one,
-			 * in which case we don't need it for a fast fsync.
-			 * Otherwise don't remove it, we could be racing with an
-			 * ongoing fast fsync that could miss the new extent.
-			 */
-			fs_info = btrfs_inode->root->fs_info;
-			spin_lock(&fs_info->trans_lock);
-			cur_gen = fs_info->generation;
-			spin_unlock(&fs_info->trans_lock);
-			if (em->generation >= cur_gen)
-				goto next;
-remove_em:
-			/*
-			 * We only remove extent maps that are not in the list of
-			 * modified extents or that are in the list but with a
-			 * generation lower then the current generation, so there
-			 * is no need to set the full fsync flag on the inode (it
-			 * hurts the fsync performance for workloads with a data
-			 * size that exceeds or is close to the system's memory).
-			 */
-			remove_extent_mapping(map, em);
-			/* once for the rb tree */
+	struct btrfs_inode *inode = page_to_inode(page);
+	struct extent_io_tree *io_tree = &inode->io_tree;
+
+	while (start <= end) {
+		const u64 cur_gen = btrfs_get_fs_generation(inode->root->fs_info);
+		const u64 len = end - start + 1;
+		struct extent_map_tree *extent_tree = &inode->extent_tree;
+		struct extent_map *em;
+
+		write_lock(&extent_tree->lock);
+		em = lookup_extent_mapping(extent_tree, start, len);
+		if (!em) {
+			write_unlock(&extent_tree->lock);
+			break;
+		}
+		if ((em->flags & EXTENT_FLAG_PINNED) || em->start != start) {
+			write_unlock(&extent_tree->lock);
 			free_extent_map(em);
+			break;
+		}
+		if (test_range_bit_exists(io_tree, em->start,
+					  extent_map_end(em) - 1, EXTENT_LOCKED))
+			goto next;
+		/*
+		 * If it's not in the list of modified extents, used by a fast
+		 * fsync, we can remove it. If it's being logged we can safely
+		 * remove it since fsync took an extra reference on the em.
+		 */
+		if (list_empty(&em->list) || (em->flags & EXTENT_FLAG_LOGGING))
+			goto remove_em;
+		/*
+		 * If it's in the list of modified extents, remove it only if
+		 * its generation is older then the current one, in which case
+		 * we don't need it for a fast fsync. Otherwise don't remove it,
+		 * we could be racing with an ongoing fast fsync that could miss
+		 * the new extent.
+		 */
+		if (em->generation >= cur_gen)
+			goto next;
+remove_em:
+		/*
+		 * We only remove extent maps that are not in the list of
+		 * modified extents or that are in the list but with a
+		 * generation lower then the current generation, so there is no
+		 * need to set the full fsync flag on the inode (it hurts the
+		 * fsync performance for workloads with a data size that exceeds
+		 * or is close to the system's memory).
+		 */
+		remove_extent_mapping(inode, em);
+		/* Once for the inode's extent map tree. */
+		free_extent_map(em);
 next:
-			start = extent_map_end(em);
-			write_unlock(&map->lock);
+		start = extent_map_end(em);
+		write_unlock(&extent_tree->lock);
 
-			/* once for us */
-			free_extent_map(em);
+		/* Once for us, for the lookup_extent_mapping() reference. */
+		free_extent_map(em);
 
-			cond_resched(); /* Allow large-extent preemption. */
+		if (need_resched()) {
+			/*
+			 * If we need to resched but we can't block just exit
+			 * and leave any remaining extent maps.
+			 */
+			if (!gfpflags_allow_blocking(mask))
+				break;
+
+			cond_resched();
 		}
 	}
-	return try_release_extent_state(tree, page, mask);
+	return try_release_extent_state(io_tree, page, mask);
 }
 
 struct btrfs_fiemap_entry {
@@ -2746,7 +2767,7 @@ static int emit_last_fiemap_cache(struct fiemap_extent_info *fieinfo,
 
 static int fiemap_next_leaf_item(struct btrfs_inode *inode, struct btrfs_path *path)
 {
-	struct extent_buffer *clone;
+	struct extent_buffer *clone = path->nodes[0];
 	struct btrfs_key key;
 	int slot;
 	int ret;
@@ -2755,29 +2776,51 @@ static int fiemap_next_leaf_item(struct btrfs_inode *inode, struct btrfs_path *p
 	if (path->slots[0] < btrfs_header_nritems(path->nodes[0]))
 		return 0;
 
+	/*
+	 * Add a temporary extra ref to an already cloned extent buffer to
+	 * prevent btrfs_next_leaf() freeing it, we want to reuse it to avoid
+	 * the cost of allocating a new one.
+	 */
+	ASSERT(test_bit(EXTENT_BUFFER_UNMAPPED, &clone->bflags));
+	atomic_inc(&clone->refs);
+
 	ret = btrfs_next_leaf(inode->root, path);
 	if (ret != 0)
-		return ret;
+		goto out;
 
 	/*
 	 * Don't bother with cloning if there are no more file extent items for
 	 * our inode.
 	 */
 	btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
-	if (key.objectid != btrfs_ino(inode) || key.type != BTRFS_EXTENT_DATA_KEY)
-		return 1;
+	if (key.objectid != btrfs_ino(inode) || key.type != BTRFS_EXTENT_DATA_KEY) {
+		ret = 1;
+		goto out;
+	}
 
+	/*
+	 * Important to preserve the start field, for the optimizations when
+	 * checking if extents are shared (see extent_fiemap()).
+	 *
+	 * We must set ->start before calling copy_extent_buffer_full().  If we
+	 * are on sub-pagesize blocksize, we use ->start to determine the offset
+	 * into the folio where our eb exists, and if we update ->start after
+	 * the fact then any subsequent reads of the eb may read from a
+	 * different offset in the folio than where we originally copied into.
+	 */
+	clone->start = path->nodes[0]->start;
 	/* See the comment at fiemap_search_slot() about why we clone. */
-	clone = btrfs_clone_extent_buffer(path->nodes[0]);
-	if (!clone)
-		return -ENOMEM;
+	copy_extent_buffer_full(clone, path->nodes[0]);
 
 	slot = path->slots[0];
 	btrfs_release_path(path);
 	path->nodes[0] = clone;
 	path->slots[0] = slot;
+out:
+	if (ret)
+		free_extent_buffer(clone);
 
-	return 0;
+	return ret;
 }
 
 /*
@@ -3508,7 +3551,7 @@ err:
 	for (int i = 0; i < num_folios; i++) {
 		if (eb->folios[i]) {
 			detach_extent_buffer_folio(eb, eb->folios[i]);
-			__folio_put(eb->folios[i]);
+			folio_put(eb->folios[i]);
 		}
 	}
 	__free_extent_buffer(eb);
@@ -3644,6 +3687,8 @@ static struct extent_buffer *grab_extent_buffer(
 	struct folio *folio = page_folio(page);
 	struct extent_buffer *exists;
 
+	lockdep_assert_held(&page->mapping->i_private_lock);
+
 	/*
 	 * For subpage case, we completely rely on radix tree to ensure we
 	 * don't try to insert two ebs for the same bytenr.  So here we always
@@ -3711,13 +3756,14 @@ static int check_eb_alignment(struct btrfs_fs_info *fs_info, u64 start)
  * The caller needs to free the existing folios and retry using the same order.
  */
 static int attach_eb_folio_to_filemap(struct extent_buffer *eb, int i,
+				      struct btrfs_subpage *prealloc,
 				      struct extent_buffer **found_eb_ret)
 {
 
 	struct btrfs_fs_info *fs_info = eb->fs_info;
 	struct address_space *mapping = fs_info->btree_inode->i_mapping;
 	const unsigned long index = eb->start >> PAGE_SHIFT;
-	struct folio *existing_folio;
+	struct folio *existing_folio = NULL;
 	int ret;
 
 	ASSERT(found_eb_ret);
@@ -3729,30 +3775,31 @@ retry:
 	ret = filemap_add_folio(mapping, eb->folios[i], index + i,
 				GFP_NOFS | __GFP_NOFAIL);
 	if (!ret)
-		return 0;
+		goto finish;
 
 	existing_folio = filemap_lock_folio(mapping, index + i);
 	/* The page cache only exists for a very short time, just retry. */
-	if (IS_ERR(existing_folio))
+	if (IS_ERR(existing_folio)) {
+		existing_folio = NULL;
 		goto retry;
+	}
 
 	/* For now, we should only have single-page folios for btree inode. */
 	ASSERT(folio_nr_pages(existing_folio) == 1);
 
-	if (folio_size(existing_folio) != folio_size(eb->folios[0])) {
+	if (folio_size(existing_folio) != eb->folio_size) {
 		folio_unlock(existing_folio);
 		folio_put(existing_folio);
 		return -EAGAIN;
 	}
 
-	if (fs_info->nodesize < PAGE_SIZE) {
-		/*
-		 * We're going to reuse the existing page, can drop our page
-		 * and subpage structure now.
-		 */
+finish:
+	spin_lock(&mapping->i_private_lock);
+	if (existing_folio && fs_info->nodesize < PAGE_SIZE) {
+		/* We're going to reuse the existing page, can drop our folio now. */
 		__free_page(folio_page(eb->folios[i], 0));
 		eb->folios[i] = existing_folio;
-	} else {
+	} else if (existing_folio) {
 		struct extent_buffer *existing_eb;
 
 		existing_eb = grab_extent_buffer(fs_info,
@@ -3760,6 +3807,7 @@ retry:
 		if (existing_eb) {
 			/* The extent buffer still exists, we can use it directly. */
 			*found_eb_ret = existing_eb;
+			spin_unlock(&mapping->i_private_lock);
 			folio_unlock(existing_folio);
 			folio_put(existing_folio);
 			return 1;
@@ -3768,6 +3816,22 @@ retry:
 		__free_page(folio_page(eb->folios[i], 0));
 		eb->folios[i] = existing_folio;
 	}
+	eb->folio_size = folio_size(eb->folios[i]);
+	eb->folio_shift = folio_shift(eb->folios[i]);
+	/* Should not fail, as we have preallocated the memory. */
+	ret = attach_extent_buffer_folio(eb, eb->folios[i], prealloc);
+	ASSERT(!ret);
+	/*
+	 * To inform we have an extra eb under allocation, so that
+	 * detach_extent_buffer_page() won't release the folio private when the
+	 * eb hasn't been inserted into radix tree yet.
+	 *
+	 * The ref will be decreased when the eb releases the page, in
+	 * detach_extent_buffer_page().  Thus needs no special handling in the
+	 * error path.
+	 */
+	btrfs_folio_inc_eb_refs(fs_info, eb->folios[i]);
+	spin_unlock(&mapping->i_private_lock);
 	return 0;
 }
 
@@ -3779,7 +3843,6 @@ struct extent_buffer *alloc_extent_buffer(struct btrfs_fs_info *fs_info,
 	int attached = 0;
 	struct extent_buffer *eb;
 	struct extent_buffer *existing_eb = NULL;
-	struct address_space *mapping = fs_info->btree_inode->i_mapping;
 	struct btrfs_subpage *prealloc = NULL;
 	u64 lockdep_owner = owner_root;
 	bool page_contig = true;
@@ -3845,7 +3908,7 @@ reallocate:
 	for (int i = 0; i < num_folios; i++) {
 		struct folio *folio;
 
-		ret = attach_eb_folio_to_filemap(eb, i, &existing_eb);
+		ret = attach_eb_folio_to_filemap(eb, i, prealloc, &existing_eb);
 		if (ret > 0) {
 			ASSERT(existing_eb);
 			goto out;
@@ -3882,22 +3945,6 @@ reallocate:
 		 * and free the allocated page.
 		 */
 		folio = eb->folios[i];
-		spin_lock(&mapping->i_private_lock);
-		/* Should not fail, as we have preallocated the memory */
-		ret = attach_extent_buffer_folio(eb, folio, prealloc);
-		ASSERT(!ret);
-		/*
-		 * To inform we have extra eb under allocation, so that
-		 * detach_extent_buffer_page() won't release the folio private
-		 * when the eb hasn't yet been inserted into radix tree.
-		 *
-		 * The ref will be decreased when the eb released the page, in
-		 * detach_extent_buffer_page().
-		 * Thus needs no special handling in error path.
-		 */
-		btrfs_folio_inc_eb_refs(fs_info, folio);
-		spin_unlock(&mapping->i_private_lock);
-
 		WARN_ON(btrfs_folio_test_dirty(fs_info, folio, eb->start, eb->len));
 
 		/*
@@ -4169,6 +4216,7 @@ void set_extent_buffer_dirty(struct extent_buffer *eb)
 	num_folios = num_extent_folios(eb);
 	WARN_ON(atomic_read(&eb->refs) == 0);
 	WARN_ON(!test_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags));
+	WARN_ON(test_bit(EXTENT_BUFFER_ZONED_ZEROOUT, &eb->bflags));
 
 	if (!was_dirty) {
 		bool subpage = eb->fs_info->nodesize < PAGE_SIZE;
@@ -4246,6 +4294,13 @@ void set_extent_buffer_uptodate(struct extent_buffer *eb)
 	}
 }
 
+static void clear_extent_buffer_reading(struct extent_buffer *eb)
+{
+	clear_bit(EXTENT_BUFFER_READING, &eb->bflags);
+	smp_mb__after_atomic();
+	wake_up_bit(&eb->bflags, EXTENT_BUFFER_READING);
+}
+
 static void end_bbio_meta_read(struct btrfs_bio *bbio)
 {
 	struct extent_buffer *eb = bbio->private;
@@ -4254,6 +4309,13 @@ static void end_bbio_meta_read(struct btrfs_bio *bbio)
 	struct folio_iter fi;
 	u32 bio_offset = 0;
 
+	/*
+	 * If the extent buffer is marked UPTODATE before the read operation
+	 * completes, other calls to read_extent_buffer_pages() will return
+	 * early without waiting for the read to finish, causing data races.
+	 */
+	WARN_ON(test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags));
+
 	eb->read_mirror = bbio->mirror_num;
 
 	if (uptodate &&
@@ -4280,9 +4342,7 @@ static void end_bbio_meta_read(struct btrfs_bio *bbio)
 		bio_offset += len;
 	}
 
-	clear_bit(EXTENT_BUFFER_READING, &eb->bflags);
-	smp_mb__after_atomic();
-	wake_up_bit(&eb->bflags, EXTENT_BUFFER_READING);
+	clear_extent_buffer_reading(eb);
 	free_extent_buffer(eb);
 
 	bio_put(&bbio->bio);
@@ -4316,9 +4376,7 @@ int read_extent_buffer_pages(struct extent_buffer *eb, int wait, int mirror_num,
 	 * will now be set, and we shouldn't read it in again.
 	 */
 	if (unlikely(test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags))) {
-		clear_bit(EXTENT_BUFFER_READING, &eb->bflags);
-		smp_mb__after_atomic();
-		wake_up_bit(&eb->bflags, EXTENT_BUFFER_READING);
+		clear_extent_buffer_reading(eb);
 		return 0;
 	}
 
@@ -4344,7 +4402,7 @@ int read_extent_buffer_pages(struct extent_buffer *eb, int wait, int mirror_num,
 		for (int i = 0; i < num_folios; i++) {
 			struct folio *folio = eb->folios[i];
 
-			ret = bio_add_folio(&bbio->bio, folio, folio_size(folio), 0);
+			ret = bio_add_folio(&bbio->bio, folio, eb->folio_size, 0);
 			ASSERT(ret);
 		}
 	}
@@ -4364,7 +4422,7 @@ static bool report_eb_range(const struct extent_buffer *eb, unsigned long start,
 			    unsigned long len)
 {
 	btrfs_warn(eb->fs_info,
-		"access to eb bytenr %llu len %lu out of range start %lu len %lu",
+		"access to eb bytenr %llu len %u out of range start %lu len %lu",
 		eb->start, eb->len, start, len);
 	WARN_ON(IS_ENABLED(CONFIG_BTRFS_DEBUG));
 
@@ -4393,7 +4451,7 @@ static inline int check_eb_range(const struct extent_buffer *eb,
 void read_extent_buffer(const struct extent_buffer *eb, void *dstv,
 			unsigned long start, unsigned long len)
 {
-	const int unit_size = folio_size(eb->folios[0]);
+	const int unit_size = eb->folio_size;
 	size_t cur;
 	size_t offset;
 	char *dst = (char *)dstv;
@@ -4433,7 +4491,7 @@ int read_extent_buffer_to_user_nofault(const struct extent_buffer *eb,
 				       void __user *dstv,
 				       unsigned long start, unsigned long len)
 {
-	const int unit_size = folio_size(eb->folios[0]);
+	const int unit_size = eb->folio_size;
 	size_t cur;
 	size_t offset;
 	char __user *dst = (char __user *)dstv;
@@ -4473,7 +4531,7 @@ int read_extent_buffer_to_user_nofault(const struct extent_buffer *eb,
 int memcmp_extent_buffer(const struct extent_buffer *eb, const void *ptrv,
 			 unsigned long start, unsigned long len)
 {
-	const int unit_size = folio_size(eb->folios[0]);
+	const int unit_size = eb->folio_size;
 	size_t cur;
 	size_t offset;
 	char *kaddr;
@@ -4544,7 +4602,7 @@ static void __write_extent_buffer(const struct extent_buffer *eb,
 				  const void *srcv, unsigned long start,
 				  unsigned long len, bool use_memmove)
 {
-	const int unit_size = folio_size(eb->folios[0]);
+	const int unit_size = eb->folio_size;
 	size_t cur;
 	size_t offset;
 	char *kaddr;
@@ -4593,7 +4651,7 @@ void write_extent_buffer(const struct extent_buffer *eb, const void *srcv,
 static void memset_extent_buffer(const struct extent_buffer *eb, int c,
 				 unsigned long start, unsigned long len)
 {
-	const int unit_size = folio_size(eb->folios[0]);
+	const int unit_size = eb->folio_size;
 	unsigned long cur = start;
 
 	if (eb->addr) {
@@ -4624,7 +4682,7 @@ void memzero_extent_buffer(const struct extent_buffer *eb, unsigned long start,
 void copy_extent_buffer_full(const struct extent_buffer *dst,
 			     const struct extent_buffer *src)
 {
-	const int unit_size = folio_size(src->folios[0]);
+	const int unit_size = src->folio_size;
 	unsigned long cur = 0;
 
 	ASSERT(dst->len == src->len);
@@ -4646,7 +4704,7 @@ void copy_extent_buffer(const struct extent_buffer *dst,
 			unsigned long dst_offset, unsigned long src_offset,
 			unsigned long len)
 {
-	const int unit_size = folio_size(dst->folios[0]);
+	const int unit_size = dst->folio_size;
 	u64 dst_len = dst->len;
 	size_t cur;
 	size_t offset;
@@ -4702,10 +4760,10 @@ static inline void eb_bitmap_offset(const struct extent_buffer *eb,
 	 * the bitmap item in the extent buffer + the offset of the byte in the
 	 * bitmap item.
 	 */
-	offset = start + offset_in_folio(eb->folios[0], eb->start) + byte_offset;
+	offset = start + offset_in_eb_folio(eb, eb->start) + byte_offset;
 
-	*folio_index = offset >> folio_shift(eb->folios[0]);
-	*folio_offset = offset_in_folio(eb->folios[0], offset);
+	*folio_index = offset >> eb->folio_shift;
+	*folio_offset = offset_in_eb_folio(eb, offset);
 }
 
 /*
@@ -4819,7 +4877,7 @@ void memcpy_extent_buffer(const struct extent_buffer *dst,
 			  unsigned long dst_offset, unsigned long src_offset,
 			  unsigned long len)
 {
-	const int unit_size = folio_size(dst->folios[0]);
+	const int unit_size = dst->folio_size;
 	unsigned long cur_off = 0;
 
 	if (check_eb_range(dst, dst_offset, len) ||
diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h
index 2c9d6570b0..dca6b12769 100644
--- a/fs/btrfs/extent_io.h
+++ b/fs/btrfs/extent_io.h
@@ -7,11 +7,33 @@
 #include <linux/refcount.h>
 #include <linux/fiemap.h>
 #include <linux/btrfs_tree.h>
+#include <linux/spinlock.h>
+#include <linux/atomic.h>
+#include <linux/rwsem.h>
+#include <linux/list.h>
+#include <linux/slab.h>
 #include "compression.h"
+#include "messages.h"
 #include "ulist.h"
 #include "misc.h"
 
+struct page;
+struct file;
+struct folio;
+struct inode;
+struct fiemap_extent_info;
+struct readahead_control;
+struct address_space;
+struct writeback_control;
+struct extent_io_tree;
+struct extent_map_tree;
+struct extent_state;
+struct btrfs_block_group;
+struct btrfs_fs_info;
+struct btrfs_inode;
+struct btrfs_root;
 struct btrfs_trans_handle;
+struct btrfs_tree_parent_check;
 
 enum {
 	EXTENT_BUFFER_UPTODATE,
@@ -63,11 +85,6 @@ enum {
 #define BITMAP_LAST_BYTE_MASK(nbits) \
 	(BYTE_MASK >> (-(nbits) & (BITS_PER_BYTE - 1)))
 
-struct btrfs_root;
-struct btrfs_inode;
-struct btrfs_fs_info;
-struct extent_io_tree;
-struct btrfs_tree_parent_check;
 
 int __init extent_buffer_init_cachep(void);
 void __cold extent_buffer_free_cachep(void);
@@ -75,7 +92,8 @@ void __cold extent_buffer_free_cachep(void);
 #define INLINE_EXTENT_BUFFER_PAGES     (BTRFS_MAX_METADATA_BLOCKSIZE / PAGE_SIZE)
 struct extent_buffer {
 	u64 start;
-	unsigned long len;
+	u32 len;
+	u32 folio_size;
 	unsigned long bflags;
 	struct btrfs_fs_info *fs_info;
 
@@ -90,6 +108,7 @@ struct extent_buffer {
 	int read_mirror;
 	/* >= 0 if eb belongs to a log tree, -1 otherwise */
 	s8 log_index;
+	u8 folio_shift;
 	struct rcu_head rcu_head;
 
 	struct rw_semaphore lock;
@@ -113,6 +132,13 @@ struct btrfs_eb_write_context {
 	struct btrfs_block_group *zoned_bg;
 };
 
+static inline unsigned long offset_in_eb_folio(const struct extent_buffer *eb,
+					       u64 start)
+{
+	ASSERT(eb->folio_size);
+	return start & (eb->folio_size - 1);
+}
+
 /*
  * Get the correct offset inside the page of extent buffer.
  *
@@ -151,13 +177,13 @@ static inline unsigned long get_eb_folio_index(const struct extent_buffer *eb,
 	 *	   the folio_shift would be large enough to always make us
 	 *	   return 0 as index.
 	 *    1.2) Several page sized folios
-	 *         The folio_shift() would be PAGE_SHIFT, giving us the correct
+	 *         The folio_shift would be PAGE_SHIFT, giving us the correct
 	 *         index.
 	 *
 	 * 2) sectorsize < PAGE_SIZE and nodesize < PAGE_SIZE case
 	 *    The folio would only be page sized, and always give us 0 as index.
 	 */
-	return offset >> folio_shift(eb->folios[0]);
+	return offset >> eb->folio_shift;
 }
 
 /*
@@ -205,20 +231,17 @@ static inline void extent_changeset_free(struct extent_changeset *changeset)
 	kfree(changeset);
 }
 
-struct extent_map_tree;
-
-int try_release_extent_mapping(struct page *page, gfp_t mask);
+bool try_release_extent_mapping(struct page *page, gfp_t mask);
 int try_release_extent_buffer(struct page *page);
 
 int btrfs_read_folio(struct file *file, struct folio *folio);
 void extent_write_locked_range(struct inode *inode, struct page *locked_page,
 			       u64 start, u64 end, struct writeback_control *wbc,
 			       bool pages_dirty);
-int extent_writepages(struct address_space *mapping,
-		      struct writeback_control *wbc);
+int btrfs_writepages(struct address_space *mapping, struct writeback_control *wbc);
 int btree_write_cache_pages(struct address_space *mapping,
 			    struct writeback_control *wbc);
-void extent_readahead(struct readahead_control *rac);
+void btrfs_readahead(struct readahead_control *rac);
 int extent_fiemap(struct btrfs_inode *inode, struct fiemap_extent_info *fieinfo,
 		  u64 start, u64 len);
 int set_folio_extent_mapped(struct folio *folio);
@@ -330,6 +353,7 @@ void clear_extent_buffer_uptodate(struct extent_buffer *eb);
 void extent_range_clear_dirty_for_io(struct inode *inode, u64 start, u64 end);
 void extent_clear_unlock_delalloc(struct btrfs_inode *inode, u64 start, u64 end,
 				  struct page *locked_page,
+				  struct extent_state **cached,
 				  u32 bits_to_clear, unsigned long page_ops);
 int extent_invalidate_folio(struct extent_io_tree *tree,
 			    struct folio *folio, size_t offset);
@@ -338,6 +362,8 @@ void btrfs_clear_buffer_dirty(struct btrfs_trans_handle *trans,
 
 int btrfs_alloc_page_array(unsigned int nr_pages, struct page **page_array,
 			   gfp_t extra_gfp);
+int btrfs_alloc_folio_array(unsigned int nr_folios, struct folio **folio_array,
+			    gfp_t extra_gfp);
 
 #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
 bool find_lock_delalloc_range(struct inode *inode,
diff --git a/fs/btrfs/extent_map.c b/fs/btrfs/extent_map.c
index 54e111b5fa..b4c9a6aa11 100644
--- a/fs/btrfs/extent_map.c
+++ b/fs/btrfs/extent_map.c
@@ -5,10 +5,10 @@
 #include <linux/spinlock.h>
 #include "messages.h"
 #include "ctree.h"
-#include "volumes.h"
 #include "extent_map.h"
 #include "compression.h"
 #include "btrfs_inode.h"
+#include "disk-io.h"
 
 
 static struct kmem_cache *extent_map_cache;
@@ -16,8 +16,7 @@ static struct kmem_cache *extent_map_cache;
 int __init extent_map_init(void)
 {
 	extent_map_cache = kmem_cache_create("btrfs_extent_map",
-			sizeof(struct extent_map), 0,
-			SLAB_MEM_SPREAD, NULL);
+					     sizeof(struct extent_map), 0, 0, NULL);
 	if (!extent_map_cache)
 		return -ENOMEM;
 	return 0;
@@ -78,6 +77,14 @@ static u64 range_end(u64 start, u64 len)
 	return start + len;
 }
 
+static void dec_evictable_extent_maps(struct btrfs_inode *inode)
+{
+	struct btrfs_fs_info *fs_info = inode->root->fs_info;
+
+	if (!btrfs_is_testing(fs_info) && is_fstree(btrfs_root_id(inode->root)))
+		percpu_counter_dec(&fs_info->evictable_extent_maps);
+}
+
 static int tree_insert(struct rb_root_cached *root, struct extent_map *em)
 {
 	struct rb_node **p = &root->rb_root.rb_node;
@@ -225,8 +232,9 @@ static bool mergeable_maps(const struct extent_map *prev, const struct extent_ma
 	return next->block_start == prev->block_start;
 }
 
-static void try_merge_map(struct extent_map_tree *tree, struct extent_map *em)
+static void try_merge_map(struct btrfs_inode *inode, struct extent_map *em)
 {
+	struct extent_map_tree *tree = &inode->extent_tree;
 	struct extent_map *merge = NULL;
 	struct rb_node *rb;
 
@@ -254,14 +262,13 @@ static void try_merge_map(struct extent_map_tree *tree, struct extent_map *em)
 			em->len += merge->len;
 			em->block_len += merge->block_len;
 			em->block_start = merge->block_start;
-			em->mod_len = (em->mod_len + em->mod_start) - merge->mod_start;
-			em->mod_start = merge->mod_start;
 			em->generation = max(em->generation, merge->generation);
 			em->flags |= EXTENT_FLAG_MERGED;
 
 			rb_erase_cached(&merge->rb_node, &tree->map);
 			RB_CLEAR_NODE(&merge->rb_node);
 			free_extent_map(merge);
+			dec_evictable_extent_maps(inode);
 		}
 	}
 
@@ -273,10 +280,10 @@ static void try_merge_map(struct extent_map_tree *tree, struct extent_map *em)
 		em->block_len += merge->block_len;
 		rb_erase_cached(&merge->rb_node, &tree->map);
 		RB_CLEAR_NODE(&merge->rb_node);
-		em->mod_len = (merge->mod_start + merge->mod_len) - em->mod_start;
 		em->generation = max(em->generation, merge->generation);
 		em->flags |= EXTENT_FLAG_MERGED;
 		free_extent_map(merge);
+		dec_evictable_extent_maps(inode);
 	}
 }
 
@@ -302,7 +309,6 @@ int unpin_extent_cache(struct btrfs_inode *inode, u64 start, u64 len, u64 gen)
 	struct extent_map_tree *tree = &inode->extent_tree;
 	int ret = 0;
 	struct extent_map *em;
-	bool prealloc = false;
 
 	write_lock(&tree->lock);
 	em = lookup_extent_mapping(tree, start, len);
@@ -327,20 +333,8 @@ int unpin_extent_cache(struct btrfs_inode *inode, u64 start, u64 len, u64 gen)
 
 	em->generation = gen;
 	em->flags &= ~EXTENT_FLAG_PINNED;
-	em->mod_start = em->start;
-	em->mod_len = em->len;
 
-	if (em->flags & EXTENT_FLAG_FILLING) {
-		prealloc = true;
-		em->flags &= ~EXTENT_FLAG_FILLING;
-	}
-
-	try_merge_map(tree, em);
-
-	if (prealloc) {
-		em->mod_start = em->start;
-		em->mod_len = em->len;
-	}
+	try_merge_map(inode, em);
 
 out:
 	write_unlock(&tree->lock);
@@ -349,58 +343,62 @@ out:
 
 }
 
-void clear_em_logging(struct extent_map_tree *tree, struct extent_map *em)
+void clear_em_logging(struct btrfs_inode *inode, struct extent_map *em)
 {
-	lockdep_assert_held_write(&tree->lock);
+	lockdep_assert_held_write(&inode->extent_tree.lock);
 
 	em->flags &= ~EXTENT_FLAG_LOGGING;
 	if (extent_map_in_tree(em))
-		try_merge_map(tree, em);
+		try_merge_map(inode, em);
 }
 
-static inline void setup_extent_mapping(struct extent_map_tree *tree,
+static inline void setup_extent_mapping(struct btrfs_inode *inode,
 					struct extent_map *em,
 					int modified)
 {
 	refcount_inc(&em->refs);
-	em->mod_start = em->start;
-	em->mod_len = em->len;
 
 	ASSERT(list_empty(&em->list));
 
 	if (modified)
-		list_add(&em->list, &tree->modified_extents);
+		list_add(&em->list, &inode->extent_tree.modified_extents);
 	else
-		try_merge_map(tree, em);
+		try_merge_map(inode, em);
 }
 
 /*
- * Add new extent map to the extent tree
+ * Add a new extent map to an inode's extent map tree.
  *
- * @tree:	tree to insert new map in
+ * @inode:	the target inode
  * @em:		map to insert
  * @modified:	indicate whether the given @em should be added to the
  *	        modified list, which indicates the extent needs to be logged
  *
- * Insert @em into @tree or perform a simple forward/backward merge with
- * existing mappings.  The extent_map struct passed in will be inserted
- * into the tree directly, with an additional reference taken, or a
- * reference dropped if the merge attempt was successful.
+ * Insert @em into the @inode's extent map tree or perform a simple
+ * forward/backward merge with existing mappings.  The extent_map struct passed
+ * in will be inserted into the tree directly, with an additional reference
+ * taken, or a reference dropped if the merge attempt was successful.
  */
-static int add_extent_mapping(struct extent_map_tree *tree,
+static int add_extent_mapping(struct btrfs_inode *inode,
 			      struct extent_map *em, int modified)
 {
-	int ret = 0;
+	struct extent_map_tree *tree = &inode->extent_tree;
+	struct btrfs_root *root = inode->root;
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	int ret;
 
 	lockdep_assert_held_write(&tree->lock);
 
 	ret = tree_insert(&tree->map, em);
 	if (ret)
-		goto out;
+		return ret;
 
-	setup_extent_mapping(tree, em, modified);
-out:
-	return ret;
+	setup_extent_mapping(inode, em, modified);
+
+	if (!btrfs_is_testing(fs_info) && is_fstree(btrfs_root_id(root)))
+		percpu_counter_inc(&fs_info->evictable_extent_maps);
+
+	return 0;
 }
 
 static struct extent_map *
@@ -466,16 +464,18 @@ struct extent_map *search_extent_mapping(struct extent_map_tree *tree,
 }
 
 /*
- * Remove an extent_map from the extent tree.
+ * Remove an extent_map from its inode's extent tree.
  *
- * @tree:	extent tree to remove from
+ * @inode:	the inode the extent map belongs to
  * @em:		extent map being removed
  *
- * Remove @em from @tree.  No reference counts are dropped, and no checks
- * are done to see if the range is in use.
+ * Remove @em from the extent tree of @inode.  No reference counts are dropped,
+ * and no checks are done to see if the range is in use.
  */
-void remove_extent_mapping(struct extent_map_tree *tree, struct extent_map *em)
+void remove_extent_mapping(struct btrfs_inode *inode, struct extent_map *em)
 {
+	struct extent_map_tree *tree = &inode->extent_tree;
+
 	lockdep_assert_held_write(&tree->lock);
 
 	WARN_ON(em->flags & EXTENT_FLAG_PINNED);
@@ -483,13 +483,17 @@ void remove_extent_mapping(struct extent_map_tree *tree, struct extent_map *em)
 	if (!(em->flags & EXTENT_FLAG_LOGGING))
 		list_del_init(&em->list);
 	RB_CLEAR_NODE(&em->rb_node);
+
+	dec_evictable_extent_maps(inode);
 }
 
-static void replace_extent_mapping(struct extent_map_tree *tree,
+static void replace_extent_mapping(struct btrfs_inode *inode,
 				   struct extent_map *cur,
 				   struct extent_map *new,
 				   int modified)
 {
+	struct extent_map_tree *tree = &inode->extent_tree;
+
 	lockdep_assert_held_write(&tree->lock);
 
 	WARN_ON(cur->flags & EXTENT_FLAG_PINNED);
@@ -499,7 +503,7 @@ static void replace_extent_mapping(struct extent_map_tree *tree,
 	rb_replace_node_cached(&cur->rb_node, &new->rb_node, &tree->map);
 	RB_CLEAR_NODE(&cur->rb_node);
 
-	setup_extent_mapping(tree, new, modified);
+	setup_extent_mapping(inode, new, modified);
 }
 
 static struct extent_map *next_extent_map(const struct extent_map *em)
@@ -528,7 +532,7 @@ static struct extent_map *prev_extent_map(struct extent_map *em)
  * and an extent that you want to insert, deal with overlap and insert
  * the best fitted new extent into the tree.
  */
-static noinline int merge_extent_mapping(struct extent_map_tree *em_tree,
+static noinline int merge_extent_mapping(struct btrfs_inode *inode,
 					 struct extent_map *existing,
 					 struct extent_map *em,
 					 u64 map_start)
@@ -539,7 +543,8 @@ static noinline int merge_extent_mapping(struct extent_map_tree *em_tree,
 	u64 end;
 	u64 start_diff;
 
-	BUG_ON(map_start < em->start || map_start >= extent_map_end(em));
+	if (map_start < em->start || map_start >= extent_map_end(em))
+		return -EINVAL;
 
 	if (existing->start > map_start) {
 		next = existing;
@@ -561,14 +566,13 @@ static noinline int merge_extent_mapping(struct extent_map_tree *em_tree,
 		em->block_start += start_diff;
 		em->block_len = em->len;
 	}
-	return add_extent_mapping(em_tree, em, 0);
+	return add_extent_mapping(inode, em, 0);
 }
 
 /*
- * Add extent mapping into em_tree.
+ * Add extent mapping into an inode's extent map tree.
  *
- * @fs_info:  the filesystem
- * @em_tree:  extent tree into which we want to insert the extent mapping
+ * @inode:    target inode
  * @em_in:    extent we are inserting
  * @start:    start of the logical range btrfs_get_extent() is requesting
  * @len:      length of the logical range btrfs_get_extent() is requesting
@@ -576,8 +580,8 @@ static noinline int merge_extent_mapping(struct extent_map_tree *em_tree,
  * Note that @em_in's range may be different from [start, start+len),
  * but they must be overlapped.
  *
- * Insert @em_in into @em_tree. In case there is an overlapping range, handle
- * the -EEXIST by either:
+ * Insert @em_in into the inode's extent map tree. In case there is an
+ * overlapping range, handle the -EEXIST by either:
  * a) Returning the existing extent in @em_in if @start is within the
  *    existing em.
  * b) Merge the existing extent with @em_in passed in.
@@ -585,12 +589,12 @@ static noinline int merge_extent_mapping(struct extent_map_tree *em_tree,
  * Return 0 on success, otherwise -EEXIST.
  *
  */
-int btrfs_add_extent_mapping(struct btrfs_fs_info *fs_info,
-			     struct extent_map_tree *em_tree,
+int btrfs_add_extent_mapping(struct btrfs_inode *inode,
 			     struct extent_map **em_in, u64 start, u64 len)
 {
 	int ret;
 	struct extent_map *em = *em_in;
+	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 
 	/*
 	 * Tree-checker should have rejected any inline extent with non-zero
@@ -599,7 +603,7 @@ int btrfs_add_extent_mapping(struct btrfs_fs_info *fs_info,
 	if (em->block_start == EXTENT_MAP_INLINE)
 		ASSERT(em->start == 0);
 
-	ret = add_extent_mapping(em_tree, em, 0);
+	ret = add_extent_mapping(inode, em, 0);
 	/* it is possible that someone inserted the extent into the tree
 	 * while we had the lock dropped.  It is also possible that
 	 * an overlapping map exists in the tree
@@ -607,7 +611,7 @@ int btrfs_add_extent_mapping(struct btrfs_fs_info *fs_info,
 	if (ret == -EEXIST) {
 		struct extent_map *existing;
 
-		existing = search_extent_mapping(em_tree, start, len);
+		existing = search_extent_mapping(&inode->extent_tree, start, len);
 
 		trace_btrfs_handle_em_exist(fs_info, existing, em, start, len);
 
@@ -628,15 +632,14 @@ int btrfs_add_extent_mapping(struct btrfs_fs_info *fs_info,
 			 * The existing extent map is the one nearest to
 			 * the [start, start + len) range which overlaps
 			 */
-			ret = merge_extent_mapping(em_tree, existing,
-						   em, start);
-			if (ret) {
+			ret = merge_extent_mapping(inode, existing, em, start);
+			if (WARN_ON(ret)) {
 				free_extent_map(em);
 				*em_in = NULL;
-				WARN_ONCE(ret,
-"unexpected error %d: merge existing(start %llu len %llu) with em(start %llu len %llu)\n",
-					  ret, existing->start, existing->len,
-					  orig_start, orig_len);
+				btrfs_warn(fs_info,
+"extent map merge error existing [%llu, %llu) with em [%llu, %llu) start %llu",
+					   existing->start, extent_map_end(existing),
+					   orig_start, orig_start + orig_len, start);
 			}
 			free_extent_map(existing);
 		}
@@ -651,8 +654,10 @@ int btrfs_add_extent_mapping(struct btrfs_fs_info *fs_info,
  * if needed. This avoids searching the tree, from the root down to the first
  * extent map, before each deletion.
  */
-static void drop_all_extent_maps_fast(struct extent_map_tree *tree)
+static void drop_all_extent_maps_fast(struct btrfs_inode *inode)
 {
+	struct extent_map_tree *tree = &inode->extent_tree;
+
 	write_lock(&tree->lock);
 	while (!RB_EMPTY_ROOT(&tree->map.rb_root)) {
 		struct extent_map *em;
@@ -661,7 +666,7 @@ static void drop_all_extent_maps_fast(struct extent_map_tree *tree)
 		node = rb_first_cached(&tree->map);
 		em = rb_entry(node, struct extent_map, rb_node);
 		em->flags &= ~(EXTENT_FLAG_PINNED | EXTENT_FLAG_LOGGING);
-		remove_extent_mapping(tree, em);
+		remove_extent_mapping(inode, em);
 		free_extent_map(em);
 		cond_resched_rwlock_write(&tree->lock);
 	}
@@ -694,7 +699,7 @@ void btrfs_drop_extent_map_range(struct btrfs_inode *inode, u64 start, u64 end,
 	WARN_ON(end < start);
 	if (end == (u64)-1) {
 		if (start == 0 && !skip_pinned) {
-			drop_all_extent_maps_fast(em_tree);
+			drop_all_extent_maps_fast(inode);
 			return;
 		}
 		len = (u64)-1;
@@ -791,7 +796,7 @@ void btrfs_drop_extent_map_range(struct btrfs_inode *inode, u64 start, u64 end,
 
 			split->generation = gen;
 			split->flags = flags;
-			replace_extent_mapping(em_tree, em, split, modified);
+			replace_extent_mapping(inode, em, split, modified);
 			free_extent_map(split);
 			split = split2;
 			split2 = NULL;
@@ -832,13 +837,11 @@ void btrfs_drop_extent_map_range(struct btrfs_inode *inode, u64 start, u64 end,
 			}
 
 			if (extent_map_in_tree(em)) {
-				replace_extent_mapping(em_tree, em, split,
-						       modified);
+				replace_extent_mapping(inode, em, split, modified);
 			} else {
 				int ret;
 
-				ret = add_extent_mapping(em_tree, split,
-							 modified);
+				ret = add_extent_mapping(inode, split, modified);
 				/* Logic error, shouldn't happen. */
 				ASSERT(ret == 0);
 				if (WARN_ON(ret != 0) && modified)
@@ -873,7 +876,7 @@ remove_em:
 				ASSERT(!split);
 				btrfs_set_inode_full_sync(inode);
 			}
-			remove_extent_mapping(em_tree, em);
+			remove_extent_mapping(inode, em);
 		}
 
 		/*
@@ -928,7 +931,7 @@ int btrfs_replace_extent_map_range(struct btrfs_inode *inode,
 	do {
 		btrfs_drop_extent_map_range(inode, new_em->start, end, false);
 		write_lock(&tree->lock);
-		ret = add_extent_mapping(tree, new_em, modified);
+		ret = add_extent_mapping(inode, new_em, modified);
 		write_unlock(&tree->lock);
 	} while (ret == -EEXIST);
 
@@ -992,7 +995,7 @@ int split_extent_map(struct btrfs_inode *inode, u64 start, u64 len, u64 pre,
 	split_pre->flags = flags;
 	split_pre->generation = em->generation;
 
-	replace_extent_mapping(em_tree, em, split_pre, 1);
+	replace_extent_mapping(inode, em, split_pre, 1);
 
 	/*
 	 * Now we only have an extent_map at:
@@ -1009,7 +1012,7 @@ int split_extent_map(struct btrfs_inode *inode, u64 start, u64 len, u64 pre,
 	split_mid->ram_bytes = split_mid->len;
 	split_mid->flags = flags;
 	split_mid->generation = em->generation;
-	add_extent_mapping(em_tree, split_mid, 1);
+	add_extent_mapping(inode, split_mid, 1);
 
 	/* Once for us */
 	free_extent_map(em);
@@ -1024,3 +1027,240 @@ out_free_pre:
 	free_extent_map(split_pre);
 	return ret;
 }
+
+struct btrfs_em_shrink_ctx {
+	long nr_to_scan;
+	long scanned;
+	u64 last_ino;
+	u64 last_root;
+};
+
+static long btrfs_scan_inode(struct btrfs_inode *inode, struct btrfs_em_shrink_ctx *ctx)
+{
+	const u64 cur_fs_gen = btrfs_get_fs_generation(inode->root->fs_info);
+	struct extent_map_tree *tree = &inode->extent_tree;
+	long nr_dropped = 0;
+	struct rb_node *node;
+
+	/*
+	 * Take the mmap lock so that we serialize with the inode logging phase
+	 * of fsync because we may need to set the full sync flag on the inode,
+	 * in case we have to remove extent maps in the tree's list of modified
+	 * extents. If we set the full sync flag in the inode while an fsync is
+	 * in progress, we may risk missing new extents because before the flag
+	 * is set, fsync decides to only wait for writeback to complete and then
+	 * during inode logging it sees the flag set and uses the subvolume tree
+	 * to find new extents, which may not be there yet because ordered
+	 * extents haven't completed yet.
+	 *
+	 * We also do a try lock because otherwise we could deadlock. This is
+	 * because the shrinker for this filesystem may be invoked while we are
+	 * in a path that is holding the mmap lock in write mode. For example in
+	 * a reflink operation while COWing an extent buffer, when allocating
+	 * pages for a new extent buffer and under memory pressure, the shrinker
+	 * may be invoked, and therefore we would deadlock by attempting to read
+	 * lock the mmap lock while we are holding already a write lock on it.
+	 */
+	if (!down_read_trylock(&inode->i_mmap_lock))
+		return 0;
+
+	/*
+	 * We want to be fast because we can be called from any path trying to
+	 * allocate memory, so if the lock is busy we don't want to spend time
+	 * waiting for it - either some task is about to do IO for the inode or
+	 * we may have another task shrinking extent maps, here in this code, so
+	 * skip this inode.
+	 */
+	if (!write_trylock(&tree->lock)) {
+		up_read(&inode->i_mmap_lock);
+		return 0;
+	}
+
+	node = rb_first_cached(&tree->map);
+	while (node) {
+		struct extent_map *em;
+
+		em = rb_entry(node, struct extent_map, rb_node);
+		node = rb_next(node);
+		ctx->scanned++;
+
+		if (em->flags & EXTENT_FLAG_PINNED)
+			goto next;
+
+		/*
+		 * If the inode is in the list of modified extents (new) and its
+		 * generation is the same (or is greater than) the current fs
+		 * generation, it means it was not yet persisted so we have to
+		 * set the full sync flag so that the next fsync will not miss
+		 * it.
+		 */
+		if (!list_empty(&em->list) && em->generation >= cur_fs_gen)
+			btrfs_set_inode_full_sync(inode);
+
+		remove_extent_mapping(inode, em);
+		trace_btrfs_extent_map_shrinker_remove_em(inode, em);
+		/* Drop the reference for the tree. */
+		free_extent_map(em);
+		nr_dropped++;
+next:
+		if (ctx->scanned >= ctx->nr_to_scan)
+			break;
+
+		/*
+		 * Stop if we need to reschedule or there's contention on the
+		 * lock. This is to avoid slowing other tasks trying to take the
+		 * lock and because the shrinker might be called during a memory
+		 * allocation path and we want to avoid taking a very long time
+		 * and slowing down all sorts of tasks.
+		 */
+		if (need_resched() || rwlock_needbreak(&tree->lock))
+			break;
+	}
+	write_unlock(&tree->lock);
+	up_read(&inode->i_mmap_lock);
+
+	return nr_dropped;
+}
+
+static long btrfs_scan_root(struct btrfs_root *root, struct btrfs_em_shrink_ctx *ctx)
+{
+	struct btrfs_inode *inode;
+	long nr_dropped = 0;
+	u64 min_ino = ctx->last_ino + 1;
+
+	inode = btrfs_find_first_inode(root, min_ino);
+	while (inode) {
+		nr_dropped += btrfs_scan_inode(inode, ctx);
+
+		min_ino = btrfs_ino(inode) + 1;
+		ctx->last_ino = btrfs_ino(inode);
+		btrfs_add_delayed_iput(inode);
+
+		if (ctx->scanned >= ctx->nr_to_scan)
+			break;
+
+		/*
+		 * We may be called from memory allocation paths, so we don't
+		 * want to take too much time and slowdown tasks.
+		 */
+		if (need_resched())
+			break;
+
+		inode = btrfs_find_first_inode(root, min_ino);
+	}
+
+	if (inode) {
+		/*
+		 * There are still inodes in this root or we happened to process
+		 * the last one and reached the scan limit. In either case set
+		 * the current root to this one, so we'll resume from the next
+		 * inode if there is one or we will find out this was the last
+		 * one and move to the next root.
+		 */
+		ctx->last_root = btrfs_root_id(root);
+	} else {
+		/*
+		 * No more inodes in this root, set extent_map_shrinker_last_ino to 0 so
+		 * that when processing the next root we start from its first inode.
+		 */
+		ctx->last_ino = 0;
+		ctx->last_root = btrfs_root_id(root) + 1;
+	}
+
+	return nr_dropped;
+}
+
+long btrfs_free_extent_maps(struct btrfs_fs_info *fs_info, long nr_to_scan)
+{
+	struct btrfs_em_shrink_ctx ctx;
+	u64 start_root_id;
+	u64 next_root_id;
+	bool cycled = false;
+	long nr_dropped = 0;
+
+	ctx.scanned = 0;
+	ctx.nr_to_scan = nr_to_scan;
+
+	/*
+	 * In case we have multiple tasks running this shrinker, make the next
+	 * one start from the next inode in case it starts before we finish.
+	 */
+	spin_lock(&fs_info->extent_map_shrinker_lock);
+	ctx.last_ino = fs_info->extent_map_shrinker_last_ino;
+	fs_info->extent_map_shrinker_last_ino++;
+	ctx.last_root = fs_info->extent_map_shrinker_last_root;
+	spin_unlock(&fs_info->extent_map_shrinker_lock);
+
+	start_root_id = ctx.last_root;
+	next_root_id = ctx.last_root;
+
+	if (trace_btrfs_extent_map_shrinker_scan_enter_enabled()) {
+		s64 nr = percpu_counter_sum_positive(&fs_info->evictable_extent_maps);
+
+		trace_btrfs_extent_map_shrinker_scan_enter(fs_info, nr_to_scan,
+							   nr, ctx.last_root,
+							   ctx.last_ino);
+	}
+
+	/*
+	 * We may be called from memory allocation paths, so we don't want to
+	 * take too much time and slowdown tasks, so stop if we need reschedule.
+	 */
+	while (ctx.scanned < ctx.nr_to_scan && !need_resched()) {
+		struct btrfs_root *root;
+		unsigned long count;
+
+		spin_lock(&fs_info->fs_roots_radix_lock);
+		count = radix_tree_gang_lookup(&fs_info->fs_roots_radix,
+					       (void **)&root,
+					       (unsigned long)next_root_id, 1);
+		if (count == 0) {
+			spin_unlock(&fs_info->fs_roots_radix_lock);
+			if (start_root_id > 0 && !cycled) {
+				next_root_id = 0;
+				ctx.last_root = 0;
+				ctx.last_ino = 0;
+				cycled = true;
+				continue;
+			}
+			break;
+		}
+		next_root_id = btrfs_root_id(root) + 1;
+		root = btrfs_grab_root(root);
+		spin_unlock(&fs_info->fs_roots_radix_lock);
+
+		if (!root)
+			continue;
+
+		if (is_fstree(btrfs_root_id(root)))
+			nr_dropped += btrfs_scan_root(root, &ctx);
+
+		btrfs_put_root(root);
+	}
+
+	/*
+	 * In case of multiple tasks running this extent map shrinking code this
+	 * isn't perfect but it's simple and silences things like KCSAN. It's
+	 * not possible to know which task made more progress because we can
+	 * cycle back to the first root and first inode if it's not the first
+	 * time the shrinker ran, see the above logic. Also a task that started
+	 * later may finish ealier than another task and made less progress. So
+	 * make this simple and update to the progress of the last task that
+	 * finished, with the occasional possiblity of having two consecutive
+	 * runs of the shrinker process the same inodes.
+	 */
+	spin_lock(&fs_info->extent_map_shrinker_lock);
+	fs_info->extent_map_shrinker_last_ino = ctx.last_ino;
+	fs_info->extent_map_shrinker_last_root = ctx.last_root;
+	spin_unlock(&fs_info->extent_map_shrinker_lock);
+
+	if (trace_btrfs_extent_map_shrinker_scan_exit_enabled()) {
+		s64 nr = percpu_counter_sum_positive(&fs_info->evictable_extent_maps);
+
+		trace_btrfs_extent_map_shrinker_scan_exit(fs_info, nr_dropped,
+							  nr, ctx.last_root,
+							  ctx.last_ino);
+	}
+
+	return nr_dropped;
+}
diff --git a/fs/btrfs/extent_map.h b/fs/btrfs/extent_map.h
index e380fc08bb..6d587111f7 100644
--- a/fs/btrfs/extent_map.h
+++ b/fs/btrfs/extent_map.h
@@ -3,10 +3,18 @@
 #ifndef BTRFS_EXTENT_MAP_H
 #define BTRFS_EXTENT_MAP_H
 
+#include <linux/compiler_types.h>
+#include <linux/rwlock_types.h>
 #include <linux/rbtree.h>
+#include <linux/list.h>
 #include <linux/refcount.h>
+#include "misc.h"
+#include "extent_map.h"
 #include "compression.h"
 
+struct btrfs_inode;
+struct btrfs_fs_info;
+
 #define EXTENT_MAP_LAST_BYTE ((u64)-4)
 #define EXTENT_MAP_HOLE ((u64)-3)
 #define EXTENT_MAP_INLINE ((u64)-2)
@@ -22,28 +30,77 @@ enum {
 	ENUM_BIT(EXTENT_FLAG_PREALLOC),
 	/* Logging this extent */
 	ENUM_BIT(EXTENT_FLAG_LOGGING),
-	/* Filling in a preallocated extent */
-	ENUM_BIT(EXTENT_FLAG_FILLING),
 	/* This em is merged from two or more physically adjacent ems */
 	ENUM_BIT(EXTENT_FLAG_MERGED),
 };
 
 /*
+ * This structure represents file extents and holes.
+ *
+ * Unlike on-disk file extent items, extent maps can be merged to save memory.
+ * This means members only match file extent items before any merging.
+ *
  * Keep this structure as compact as possible, as we can have really large
  * amounts of allocated extent maps at any time.
  */
 struct extent_map {
 	struct rb_node rb_node;
 
-	/* all of these are in bytes */
+	/* All of these are in bytes. */
+
+	/* File offset matching the offset of a BTRFS_EXTENT_ITEM_KEY key. */
 	u64 start;
+
+	/*
+	 * Length of the file extent.
+	 *
+	 * For non-inlined file extents it's btrfs_file_extent_item::num_bytes.
+	 * For inline extents it's sectorsize, since inline data starts at
+	 * offsetof(struct btrfs_file_extent_item, disk_bytenr) thus
+	 * btrfs_file_extent_item::num_bytes is not valid.
+	 */
 	u64 len;
-	u64 mod_start;
-	u64 mod_len;
+
+	/*
+	 * The file offset of the original file extent before splitting.
+	 *
+	 * This is an in-memory only member, matching
+	 * extent_map::start - btrfs_file_extent_item::offset for
+	 * regular/preallocated extents. EXTENT_MAP_HOLE otherwise.
+	 */
 	u64 orig_start;
+
+	/*
+	 * The full on-disk extent length, matching
+	 * btrfs_file_extent_item::disk_num_bytes.
+	 */
 	u64 orig_block_len;
+
+	/*
+	 * The decompressed size of the whole on-disk extent, matching
+	 * btrfs_file_extent_item::ram_bytes.
+	 */
 	u64 ram_bytes;
+
+	/*
+	 * The on-disk logical bytenr for the file extent.
+	 *
+	 * For compressed extents it matches btrfs_file_extent_item::disk_bytenr.
+	 * For uncompressed extents it matches
+	 * btrfs_file_extent_item::disk_bytenr + btrfs_file_extent_item::offset
+	 *
+	 * For holes it is EXTENT_MAP_HOLE and for inline extents it is
+	 * EXTENT_MAP_INLINE.
+	 */
 	u64 block_start;
+
+	/*
+	 * The on-disk length for the file extent.
+	 *
+	 * For compressed extents it matches btrfs_file_extent_item::disk_num_bytes.
+	 * For uncompressed extents it matches extent_map::len.
+	 * For holes and inline extents it's -1 and shouldn't be used.
+	 */
 	u64 block_len;
 
 	/*
@@ -116,7 +173,7 @@ static inline u64 extent_map_end(const struct extent_map *em)
 void extent_map_tree_init(struct extent_map_tree *tree);
 struct extent_map *lookup_extent_mapping(struct extent_map_tree *tree,
 					 u64 start, u64 len);
-void remove_extent_mapping(struct extent_map_tree *tree, struct extent_map *em);
+void remove_extent_mapping(struct btrfs_inode *inode, struct extent_map *em);
 int split_extent_map(struct btrfs_inode *inode, u64 start, u64 len, u64 pre,
 		     u64 new_logical);
 
@@ -125,11 +182,10 @@ void free_extent_map(struct extent_map *em);
 int __init extent_map_init(void);
 void __cold extent_map_exit(void);
 int unpin_extent_cache(struct btrfs_inode *inode, u64 start, u64 len, u64 gen);
-void clear_em_logging(struct extent_map_tree *tree, struct extent_map *em);
+void clear_em_logging(struct btrfs_inode *inode, struct extent_map *em);
 struct extent_map *search_extent_mapping(struct extent_map_tree *tree,
 					 u64 start, u64 len);
-int btrfs_add_extent_mapping(struct btrfs_fs_info *fs_info,
-			     struct extent_map_tree *em_tree,
+int btrfs_add_extent_mapping(struct btrfs_inode *inode,
 			     struct extent_map **em_in, u64 start, u64 len);
 void btrfs_drop_extent_map_range(struct btrfs_inode *inode,
 				 u64 start, u64 end,
@@ -137,5 +193,6 @@ void btrfs_drop_extent_map_range(struct btrfs_inode *inode,
 int btrfs_replace_extent_map_range(struct btrfs_inode *inode,
 				   struct extent_map *new_em,
 				   bool modified);
+long btrfs_free_extent_maps(struct btrfs_fs_info *fs_info, long nr_to_scan);
 
 #endif
diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c
index 81ac1d474b..bce95f8717 100644
--- a/fs/btrfs/file-item.c
+++ b/fs/btrfs/file-item.c
@@ -10,17 +10,14 @@
 #include <linux/sched/mm.h>
 #include <crypto/hash.h>
 #include "messages.h"
-#include "misc.h"
 #include "ctree.h"
 #include "disk-io.h"
 #include "transaction.h"
 #include "bio.h"
-#include "print-tree.h"
 #include "compression.h"
 #include "fs.h"
 #include "accessors.h"
 #include "file-item.h"
-#include "super.h"
 
 #define __MAX_CSUM_ITEMS(r, size) ((unsigned long)(((BTRFS_LEAF_DATA_SIZE(r) - \
 				   sizeof(struct btrfs_item) * 2) / \
@@ -179,7 +176,6 @@ int btrfs_insert_hole_extent(struct btrfs_trans_handle *trans,
 				      sizeof(*item));
 	if (ret < 0)
 		goto out;
-	BUG_ON(ret); /* Can't happen */
 	leaf = path->nodes[0];
 	item = btrfs_item_ptr(leaf, path->slots[0],
 			      struct btrfs_file_extent_item);
@@ -434,8 +430,7 @@ blk_status_t btrfs_lookup_bio_sums(struct btrfs_bio *bbio)
 			memset(csum_dst, 0, csum_size);
 			count = 1;
 
-			if (inode->root->root_key.objectid ==
-			    BTRFS_DATA_RELOC_TREE_OBJECTID) {
+			if (btrfs_root_id(inode->root) == BTRFS_DATA_RELOC_TREE_OBJECTID) {
 				u64 file_offset = bbio->file_offset + bio_offset;
 
 				set_extent_bit(&inode->io_tree, file_offset,
@@ -454,9 +449,22 @@ blk_status_t btrfs_lookup_bio_sums(struct btrfs_bio *bbio)
 	return ret;
 }
 
+/*
+ * Search for checksums for a given logical range.
+ *
+ * @root:		The root where to look for checksums.
+ * @start:		Logical address of target checksum range.
+ * @end:		End offset (inclusive) of the target checksum range.
+ * @list:		List for adding each checksum that was found.
+ *			Can be NULL in case the caller only wants to check if
+ *			there any checksums for the range.
+ * @nowait:		Indicate if the search must be non-blocking or not.
+ *
+ * Return < 0 on error, 0 if no checksums were found, or 1 if checksums were
+ * found.
+ */
 int btrfs_lookup_csums_list(struct btrfs_root *root, u64 start, u64 end,
-			    struct list_head *list, int search_commit,
-			    bool nowait)
+			    struct list_head *list, bool nowait)
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct btrfs_key key;
@@ -464,8 +472,8 @@ int btrfs_lookup_csums_list(struct btrfs_root *root, u64 start, u64 end,
 	struct extent_buffer *leaf;
 	struct btrfs_ordered_sum *sums;
 	struct btrfs_csum_item *item;
-	LIST_HEAD(tmplist);
 	int ret;
+	bool found_csums = false;
 
 	ASSERT(IS_ALIGNED(start, fs_info->sectorsize) &&
 	       IS_ALIGNED(end + 1, fs_info->sectorsize));
@@ -475,11 +483,6 @@ int btrfs_lookup_csums_list(struct btrfs_root *root, u64 start, u64 end,
 		return -ENOMEM;
 
 	path->nowait = nowait;
-	if (search_commit) {
-		path->skip_locking = 1;
-		path->reada = READA_FORWARD;
-		path->search_commit_root = 1;
-	}
 
 	key.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
 	key.offset = start;
@@ -487,7 +490,7 @@ int btrfs_lookup_csums_list(struct btrfs_root *root, u64 start, u64 end,
 
 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 	if (ret < 0)
-		goto fail;
+		goto out;
 	if (ret > 0 && path->slots[0] > 0) {
 		leaf = path->nodes[0];
 		btrfs_item_key_to_cpu(leaf, &key, path->slots[0] - 1);
@@ -522,7 +525,7 @@ int btrfs_lookup_csums_list(struct btrfs_root *root, u64 start, u64 end,
 		if (path->slots[0] >= btrfs_header_nritems(leaf)) {
 			ret = btrfs_next_leaf(root, path);
 			if (ret < 0)
-				goto fail;
+				goto out;
 			if (ret > 0)
 				break;
 			leaf = path->nodes[0];
@@ -544,6 +547,10 @@ int btrfs_lookup_csums_list(struct btrfs_root *root, u64 start, u64 end,
 			continue;
 		}
 
+		found_csums = true;
+		if (!list)
+			goto out;
+
 		csum_end = min(csum_end, end + 1);
 		item = btrfs_item_ptr(path->nodes[0], path->slots[0],
 				      struct btrfs_csum_item);
@@ -557,7 +564,7 @@ int btrfs_lookup_csums_list(struct btrfs_root *root, u64 start, u64 end,
 				       GFP_NOFS);
 			if (!sums) {
 				ret = -ENOMEM;
-				goto fail;
+				goto out;
 			}
 
 			sums->logical = start;
@@ -571,21 +578,24 @@ int btrfs_lookup_csums_list(struct btrfs_root *root, u64 start, u64 end,
 					   bytes_to_csum_size(fs_info, size));
 
 			start += size;
-			list_add_tail(&sums->list, &tmplist);
+			list_add_tail(&sums->list, list);
 		}
 		path->slots[0]++;
 	}
-	ret = 0;
-fail:
-	while (ret < 0 && !list_empty(&tmplist)) {
-		sums = list_entry(tmplist.next, struct btrfs_ordered_sum, list);
-		list_del(&sums->list);
-		kfree(sums);
+out:
+	btrfs_free_path(path);
+	if (ret < 0) {
+		if (list) {
+			struct btrfs_ordered_sum *tmp_sums;
+
+			list_for_each_entry_safe(sums, tmp_sums, list, list)
+				kfree(sums);
+		}
+
+		return ret;
 	}
-	list_splice_tail(&tmplist, list);
 
-	btrfs_free_path(path);
-	return ret;
+	return found_csums ? 1 : 0;
 }
 
 /*
@@ -874,8 +884,8 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans,
 	const u32 csum_size = fs_info->csum_size;
 	u32 blocksize_bits = fs_info->sectorsize_bits;
 
-	ASSERT(root->root_key.objectid == BTRFS_CSUM_TREE_OBJECTID ||
-	       root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID);
+	ASSERT(btrfs_root_id(root) == BTRFS_CSUM_TREE_OBJECTID ||
+	       btrfs_root_id(root) == BTRFS_TREE_LOG_OBJECTID);
 
 	path = btrfs_alloc_path();
 	if (!path)
@@ -1175,7 +1185,7 @@ extend_csum:
 		 * search, etc, because log trees are temporary anyway and it
 		 * would only save a few bytes of leaf space.
 		 */
-		if (root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID) {
+		if (btrfs_root_id(root) == BTRFS_TREE_LOG_OBJECTID) {
 			if (path->slots[0] + 1 >=
 			    btrfs_header_nritems(path->nodes[0])) {
 				ret = find_next_csum_offset(root, path, &next_offset);
@@ -1229,8 +1239,6 @@ insert:
 				      ins_size);
 	if (ret < 0)
 		goto out;
-	if (WARN_ON(ret != 0))
-		goto out;
 	leaf = path->nodes[0];
 csum:
 	item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_csum_item);
@@ -1271,20 +1279,19 @@ void btrfs_extent_item_to_extent_map(struct btrfs_inode *inode,
 	struct extent_buffer *leaf = path->nodes[0];
 	const int slot = path->slots[0];
 	struct btrfs_key key;
-	u64 extent_start, extent_end;
+	u64 extent_start;
 	u64 bytenr;
 	u8 type = btrfs_file_extent_type(leaf, fi);
 	int compress_type = btrfs_file_extent_compression(leaf, fi);
 
 	btrfs_item_key_to_cpu(leaf, &key, slot);
 	extent_start = key.offset;
-	extent_end = btrfs_file_extent_end(path);
 	em->ram_bytes = btrfs_file_extent_ram_bytes(leaf, fi);
 	em->generation = btrfs_file_extent_generation(leaf, fi);
 	if (type == BTRFS_FILE_EXTENT_REG ||
 	    type == BTRFS_FILE_EXTENT_PREALLOC) {
 		em->start = extent_start;
-		em->len = extent_end - extent_start;
+		em->len = btrfs_file_extent_end(path) - extent_start;
 		em->orig_start = extent_start -
 			btrfs_file_extent_offset(leaf, fi);
 		em->orig_block_len = btrfs_file_extent_disk_num_bytes(leaf, fi);
@@ -1305,9 +1312,12 @@ void btrfs_extent_item_to_extent_map(struct btrfs_inode *inode,
 				em->flags |= EXTENT_FLAG_PREALLOC;
 		}
 	} else if (type == BTRFS_FILE_EXTENT_INLINE) {
+		/* Tree-checker has ensured this. */
+		ASSERT(extent_start == 0);
+
 		em->block_start = EXTENT_MAP_INLINE;
-		em->start = extent_start;
-		em->len = extent_end - extent_start;
+		em->start = 0;
+		em->len = fs_info->sectorsize;
 		/*
 		 * Initialize orig_start and block_len with the same values
 		 * as in inode.c:btrfs_get_extent().
@@ -1319,7 +1329,7 @@ void btrfs_extent_item_to_extent_map(struct btrfs_inode *inode,
 		btrfs_err(fs_info,
 			  "unknown file extent item type %d, inode %llu, offset %llu, "
 			  "root %llu", type, btrfs_ino(inode), extent_start,
-			  root->root_key.objectid);
+			  btrfs_root_id(root));
 	}
 }
 
@@ -1340,12 +1350,10 @@ u64 btrfs_file_extent_end(const struct btrfs_path *path)
 	ASSERT(key.type == BTRFS_EXTENT_DATA_KEY);
 	fi = btrfs_item_ptr(leaf, slot, struct btrfs_file_extent_item);
 
-	if (btrfs_file_extent_type(leaf, fi) == BTRFS_FILE_EXTENT_INLINE) {
-		end = btrfs_file_extent_ram_bytes(leaf, fi);
-		end = ALIGN(key.offset + end, leaf->fs_info->sectorsize);
-	} else {
+	if (btrfs_file_extent_type(leaf, fi) == BTRFS_FILE_EXTENT_INLINE)
+		end = leaf->fs_info->sectorsize;
+	else
 		end = key.offset + btrfs_file_extent_num_bytes(leaf, fi);
-	}
 
 	return end;
 }
diff --git a/fs/btrfs/file-item.h b/fs/btrfs/file-item.h
index 04bd2d34ef..557dc43d71 100644
--- a/fs/btrfs/file-item.h
+++ b/fs/btrfs/file-item.h
@@ -3,8 +3,21 @@
 #ifndef BTRFS_FILE_ITEM_H
 #define BTRFS_FILE_ITEM_H
 
+#include <linux/list.h>
+#include <uapi/linux/btrfs_tree.h>
 #include "accessors.h"
 
+struct extent_map;
+struct btrfs_file_extent_item;
+struct btrfs_fs_info;
+struct btrfs_path;
+struct btrfs_bio;
+struct btrfs_trans_handle;
+struct btrfs_root;
+struct btrfs_ordered_sum;
+struct btrfs_path;
+struct btrfs_inode;
+
 #define BTRFS_FILE_EXTENT_INLINE_DATA_START		\
 		(offsetof(struct btrfs_file_extent_item, disk_bytenr))
 
@@ -55,8 +68,7 @@ int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end,
 			     struct list_head *list, int search_commit,
 			     bool nowait);
 int btrfs_lookup_csums_list(struct btrfs_root *root, u64 start, u64 end,
-			    struct list_head *list, int search_commit,
-			    bool nowait);
+			    struct list_head *list, bool nowait);
 int btrfs_lookup_csums_bitmap(struct btrfs_root *root, struct btrfs_path *path,
 			      u64 start, u64 end, u8 *csum_buf,
 			      unsigned long *csum_bitmap);
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index 616790d6e5..ca434f0cd2 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -22,10 +22,8 @@
 #include "disk-io.h"
 #include "transaction.h"
 #include "btrfs_inode.h"
-#include "print-tree.h"
 #include "tree-log.h"
 #include "locking.h"
-#include "volumes.h"
 #include "qgroup.h"
 #include "compression.h"
 #include "delalloc-space.h"
@@ -130,7 +128,7 @@ int btrfs_dirty_pages(struct btrfs_inode *inode, struct page **pages,
 		      struct extent_state **cached, bool noreserve)
 {
 	struct btrfs_fs_info *fs_info = inode->root->fs_info;
-	int err = 0;
+	int ret = 0;
 	int i;
 	u64 num_bytes;
 	u64 start_pos;
@@ -160,10 +158,10 @@ int btrfs_dirty_pages(struct btrfs_inode *inode, struct page **pages,
 			 EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING | EXTENT_DEFRAG,
 			 cached);
 
-	err = btrfs_set_extent_delalloc(inode, start_pos, end_of_last_block,
+	ret = btrfs_set_extent_delalloc(inode, start_pos, end_of_last_block,
 					extra_bits, cached);
-	if (err)
-		return err;
+	if (ret)
+		return ret;
 
 	for (i = 0; i < num_pages; i++) {
 		struct page *p = pages[i];
@@ -208,7 +206,6 @@ int btrfs_drop_extents(struct btrfs_trans_handle *trans,
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct extent_buffer *leaf;
 	struct btrfs_file_extent_item *fi;
-	struct btrfs_ref ref = { 0 };
 	struct btrfs_key key;
 	struct btrfs_key new_key;
 	u64 ino = btrfs_ino(inode);
@@ -248,7 +245,7 @@ int btrfs_drop_extents(struct btrfs_trans_handle *trans,
 	if (args->start >= inode->disk_i_size && !args->replace_extent)
 		modify_tree = 0;
 
-	update_refs = (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID);
+	update_refs = (btrfs_root_id(root) != BTRFS_TREE_LOG_OBJECTID);
 	while (1) {
 		recow = 0;
 		ret = btrfs_lookup_file_extent(trans, root, path, ino,
@@ -375,15 +372,17 @@ next_slot:
 			btrfs_mark_buffer_dirty(trans, leaf);
 
 			if (update_refs && disk_bytenr > 0) {
-				btrfs_init_generic_ref(&ref,
-						BTRFS_ADD_DELAYED_REF,
-						disk_bytenr, num_bytes, 0,
-						root->root_key.objectid);
-				btrfs_init_data_ref(&ref,
-						root->root_key.objectid,
-						new_key.objectid,
-						args->start - extent_offset,
-						0, false);
+				struct btrfs_ref ref = {
+					.action = BTRFS_ADD_DELAYED_REF,
+					.bytenr = disk_bytenr,
+					.num_bytes = num_bytes,
+					.parent = 0,
+					.owning_root = btrfs_root_id(root),
+					.ref_root = btrfs_root_id(root),
+				};
+				btrfs_init_data_ref(&ref, new_key.objectid,
+						    args->start - extent_offset,
+						    0, false);
 				ret = btrfs_inc_extent_ref(trans, &ref);
 				if (ret) {
 					btrfs_abort_transaction(trans, ret);
@@ -466,15 +465,17 @@ delete_extent_item:
 				extent_end = ALIGN(extent_end,
 						   fs_info->sectorsize);
 			} else if (update_refs && disk_bytenr > 0) {
-				btrfs_init_generic_ref(&ref,
-						BTRFS_DROP_DELAYED_REF,
-						disk_bytenr, num_bytes, 0,
-						root->root_key.objectid);
-				btrfs_init_data_ref(&ref,
-						root->root_key.objectid,
-						key.objectid,
-						key.offset - extent_offset, 0,
-						false);
+				struct btrfs_ref ref = {
+					.action = BTRFS_DROP_DELAYED_REF,
+					.bytenr = disk_bytenr,
+					.num_bytes = num_bytes,
+					.parent = 0,
+					.owning_root = btrfs_root_id(root),
+					.ref_root = btrfs_root_id(root),
+				};
+				btrfs_init_data_ref(&ref, key.objectid,
+						    key.offset - extent_offset,
+						    0, false);
 				ret = btrfs_free_extent(trans, &ref);
 				if (ret) {
 					btrfs_abort_transaction(trans, ret);
@@ -750,10 +751,13 @@ again:
 						extent_end - split);
 		btrfs_mark_buffer_dirty(trans, leaf);
 
-		btrfs_init_generic_ref(&ref, BTRFS_ADD_DELAYED_REF, bytenr,
-				       num_bytes, 0, root->root_key.objectid);
-		btrfs_init_data_ref(&ref, root->root_key.objectid, ino,
-				    orig_offset, 0, false);
+		ref.action = BTRFS_ADD_DELAYED_REF;
+		ref.bytenr = bytenr;
+		ref.num_bytes = num_bytes;
+		ref.parent = 0;
+		ref.owning_root = btrfs_root_id(root);
+		ref.ref_root = btrfs_root_id(root);
+		btrfs_init_data_ref(&ref, ino, orig_offset, 0, false);
 		ret = btrfs_inc_extent_ref(trans, &ref);
 		if (ret) {
 			btrfs_abort_transaction(trans, ret);
@@ -776,10 +780,14 @@ again:
 
 	other_start = end;
 	other_end = 0;
-	btrfs_init_generic_ref(&ref, BTRFS_DROP_DELAYED_REF, bytenr,
-			       num_bytes, 0, root->root_key.objectid);
-	btrfs_init_data_ref(&ref, root->root_key.objectid, ino, orig_offset,
-			    0, false);
+
+	ref.action = BTRFS_DROP_DELAYED_REF;
+	ref.bytenr = bytenr;
+	ref.num_bytes = num_bytes;
+	ref.parent = 0;
+	ref.owning_root = btrfs_root_id(root);
+	ref.ref_root = btrfs_root_id(root);
+	btrfs_init_data_ref(&ref, ino, orig_offset, 0, false);
 	if (extent_mergeable(leaf, path->slots[0] + 1,
 			     ino, bytenr, orig_offset,
 			     &other_start, &other_end)) {
@@ -917,7 +925,7 @@ static noinline int prepare_pages(struct inode *inode, struct page **pages,
 	unsigned long index = pos >> PAGE_SHIFT;
 	gfp_t mask = get_prepare_gfp_flags(inode, nowait);
 	fgf_t fgp_flags = get_prepare_fgp_flags(nowait);
-	int err = 0;
+	int ret = 0;
 	int faili;
 
 	for (i = 0; i < num_pages; i++) {
@@ -927,28 +935,28 @@ again:
 		if (!pages[i]) {
 			faili = i - 1;
 			if (nowait)
-				err = -EAGAIN;
+				ret = -EAGAIN;
 			else
-				err = -ENOMEM;
+				ret = -ENOMEM;
 			goto fail;
 		}
 
-		err = set_page_extent_mapped(pages[i]);
-		if (err < 0) {
+		ret = set_page_extent_mapped(pages[i]);
+		if (ret < 0) {
 			faili = i;
 			goto fail;
 		}
 
 		if (i == 0)
-			err = prepare_uptodate_page(inode, pages[i], pos,
+			ret = prepare_uptodate_page(inode, pages[i], pos,
 						    force_uptodate);
-		if (!err && i == num_pages - 1)
-			err = prepare_uptodate_page(inode, pages[i],
+		if (!ret && i == num_pages - 1)
+			ret = prepare_uptodate_page(inode, pages[i],
 						    pos + write_bytes, false);
-		if (err) {
+		if (ret) {
 			put_page(pages[i]);
-			if (!nowait && err == -EAGAIN) {
-				err = 0;
+			if (!nowait && ret == -EAGAIN) {
+				ret = 0;
 				goto again;
 			}
 			faili = i - 1;
@@ -964,7 +972,7 @@ fail:
 		put_page(pages[faili]);
 		faili--;
 	}
-	return err;
+	return ret;
 
 }
 
@@ -1467,7 +1475,7 @@ static ssize_t btrfs_direct_write(struct kiocb *iocb, struct iov_iter *from)
 	ssize_t written_buffered;
 	size_t prev_left = 0;
 	loff_t endbyte;
-	ssize_t err;
+	ssize_t ret;
 	unsigned int ilock_flags = 0;
 	struct iomap_dio *dio;
 
@@ -1484,9 +1492,9 @@ static ssize_t btrfs_direct_write(struct kiocb *iocb, struct iov_iter *from)
 		ilock_flags |= BTRFS_ILOCK_SHARED;
 
 relock:
-	err = btrfs_inode_lock(BTRFS_I(inode), ilock_flags);
-	if (err < 0)
-		return err;
+	ret = btrfs_inode_lock(BTRFS_I(inode), ilock_flags);
+	if (ret < 0)
+		return ret;
 
 	/* Shared lock cannot be used with security bits set. */
 	if ((ilock_flags & BTRFS_ILOCK_SHARED) && !IS_NOSEC(inode)) {
@@ -1495,14 +1503,14 @@ relock:
 		goto relock;
 	}
 
-	err = generic_write_checks(iocb, from);
-	if (err <= 0) {
+	ret = generic_write_checks(iocb, from);
+	if (ret <= 0) {
 		btrfs_inode_unlock(BTRFS_I(inode), ilock_flags);
-		return err;
+		return ret;
 	}
 
-	err = btrfs_write_check(iocb, from, err);
-	if (err < 0) {
+	ret = btrfs_write_check(iocb, from, ret);
+	if (ret < 0) {
 		btrfs_inode_unlock(BTRFS_I(inode), ilock_flags);
 		goto out;
 	}
@@ -1542,27 +1550,43 @@ relock:
 	 * So here we disable page faults in the iov_iter and then retry if we
 	 * got -EFAULT, faulting in the pages before the retry.
 	 */
+again:
 	from->nofault = true;
 	dio = btrfs_dio_write(iocb, from, written);
 	from->nofault = false;
 
-	/*
-	 * iomap_dio_complete() will call btrfs_sync_file() if we have a dsync
-	 * iocb, and that needs to lock the inode. So unlock it before calling
-	 * iomap_dio_complete() to avoid a deadlock.
-	 */
-	btrfs_inode_unlock(BTRFS_I(inode), ilock_flags);
+	if (IS_ERR_OR_NULL(dio)) {
+		ret = PTR_ERR_OR_ZERO(dio);
+	} else {
+		struct btrfs_file_private stack_private = { 0 };
+		struct btrfs_file_private *private;
+		const bool have_private = (file->private_data != NULL);
 
-	if (IS_ERR_OR_NULL(dio))
-		err = PTR_ERR_OR_ZERO(dio);
-	else
-		err = iomap_dio_complete(dio);
+		if (!have_private)
+			file->private_data = &stack_private;
+
+		/*
+		 * If we have a synchoronous write, we must make sure the fsync
+		 * triggered by the iomap_dio_complete() call below doesn't
+		 * deadlock on the inode lock - we are already holding it and we
+		 * can't call it after unlocking because we may need to complete
+		 * partial writes due to the input buffer (or parts of it) not
+		 * being already faulted in.
+		 */
+		private = file->private_data;
+		private->fsync_skip_inode_lock = true;
+		ret = iomap_dio_complete(dio);
+		private->fsync_skip_inode_lock = false;
+
+		if (!have_private)
+			file->private_data = NULL;
+	}
 
 	/* No increment (+=) because iomap returns a cumulative value. */
-	if (err > 0)
-		written = err;
+	if (ret > 0)
+		written = ret;
 
-	if (iov_iter_count(from) > 0 && (err == -EFAULT || err > 0)) {
+	if (iov_iter_count(from) > 0 && (ret == -EFAULT || ret > 0)) {
 		const size_t left = iov_iter_count(from);
 		/*
 		 * We have more data left to write. Try to fault in as many as
@@ -1579,19 +1603,21 @@ relock:
 		 * to buffered IO in case we haven't made any progress.
 		 */
 		if (left == prev_left) {
-			err = -ENOTBLK;
+			ret = -ENOTBLK;
 		} else {
 			fault_in_iov_iter_readable(from, left);
 			prev_left = left;
-			goto relock;
+			goto again;
 		}
 	}
 
+	btrfs_inode_unlock(BTRFS_I(inode), ilock_flags);
+
 	/*
-	 * If 'err' is -ENOTBLK or we have not written all data, then it means
+	 * If 'ret' is -ENOTBLK or we have not written all data, then it means
 	 * we must fallback to buffered IO.
 	 */
-	if ((err < 0 && err != -ENOTBLK) || !iov_iter_count(from))
+	if ((ret < 0 && ret != -ENOTBLK) || !iov_iter_count(from))
 		goto out;
 
 buffered:
@@ -1602,14 +1628,14 @@ buffered:
 	 * below, we will block when flushing and waiting for the IO.
 	 */
 	if (iocb->ki_flags & IOCB_NOWAIT) {
-		err = -EAGAIN;
+		ret = -EAGAIN;
 		goto out;
 	}
 
 	pos = iocb->ki_pos;
 	written_buffered = btrfs_buffered_write(iocb, from);
 	if (written_buffered < 0) {
-		err = written_buffered;
+		ret = written_buffered;
 		goto out;
 	}
 	/*
@@ -1617,18 +1643,18 @@ buffered:
 	 * able to read what was just written.
 	 */
 	endbyte = pos + written_buffered - 1;
-	err = btrfs_fdatawrite_range(inode, pos, endbyte);
-	if (err)
+	ret = btrfs_fdatawrite_range(inode, pos, endbyte);
+	if (ret)
 		goto out;
-	err = filemap_fdatawait_range(inode->i_mapping, pos, endbyte);
-	if (err)
+	ret = filemap_fdatawait_range(inode->i_mapping, pos, endbyte);
+	if (ret)
 		goto out;
 	written += written_buffered;
 	iocb->ki_pos = pos + written_buffered;
 	invalidate_mapping_pages(file->f_mapping, pos >> PAGE_SHIFT,
 				 endbyte >> PAGE_SHIFT);
 out:
-	return err < 0 ? err : written;
+	return ret < 0 ? ret : written;
 }
 
 static ssize_t btrfs_encoded_write(struct kiocb *iocb, struct iov_iter *from,
@@ -1785,6 +1811,7 @@ static inline bool skip_inode_logging(const struct btrfs_log_ctx *ctx)
  */
 int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 {
+	struct btrfs_file_private *private = file->private_data;
 	struct dentry *dentry = file_dentry(file);
 	struct inode *inode = d_inode(dentry);
 	struct btrfs_fs_info *fs_info = inode_to_fs_info(inode);
@@ -1794,6 +1821,7 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	int ret = 0, err;
 	u64 len;
 	bool full_sync;
+	const bool skip_ilock = (private ? private->fsync_skip_inode_lock : false);
 
 	trace_btrfs_sync_file(file, datasync);
 
@@ -1821,7 +1849,10 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	if (ret)
 		goto out;
 
-	btrfs_inode_lock(BTRFS_I(inode), BTRFS_ILOCK_MMAP);
+	if (skip_ilock)
+		down_write(&BTRFS_I(inode)->i_mmap_lock);
+	else
+		btrfs_inode_lock(BTRFS_I(inode), BTRFS_ILOCK_MMAP);
 
 	atomic_inc(&root->log_batch);
 
@@ -1845,7 +1876,10 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	 */
 	ret = start_ordered_ops(inode, start, end);
 	if (ret) {
-		btrfs_inode_unlock(BTRFS_I(inode), BTRFS_ILOCK_MMAP);
+		if (skip_ilock)
+			up_write(&BTRFS_I(inode)->i_mmap_lock);
+		else
+			btrfs_inode_unlock(BTRFS_I(inode), BTRFS_ILOCK_MMAP);
 		goto out;
 	}
 
@@ -1877,6 +1911,7 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	 */
 	if (full_sync || btrfs_is_zoned(fs_info)) {
 		ret = btrfs_wait_ordered_range(inode, start, len);
+		clear_bit(BTRFS_INODE_COW_WRITE_ERROR, &BTRFS_I(inode)->runtime_flags);
 	} else {
 		/*
 		 * Get our ordered extents as soon as possible to avoid doing
@@ -1886,6 +1921,21 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 		btrfs_get_ordered_extents_for_logging(BTRFS_I(inode),
 						      &ctx.ordered_extents);
 		ret = filemap_fdatawait_range(inode->i_mapping, start, end);
+		if (ret)
+			goto out_release_extents;
+
+		/*
+		 * Check and clear the BTRFS_INODE_COW_WRITE_ERROR now after
+		 * starting and waiting for writeback, because for buffered IO
+		 * it may have been set during the end IO callback
+		 * (end_bbio_data_write() -> btrfs_finish_ordered_extent()) in
+		 * case an error happened and we need to wait for ordered
+		 * extents to complete so that any extent maps that point to
+		 * unwritten locations are dropped and we don't log them.
+		 */
+		if (test_and_clear_bit(BTRFS_INODE_COW_WRITE_ERROR,
+				       &BTRFS_I(inode)->runtime_flags))
+			ret = btrfs_wait_ordered_range(inode, start, len);
 	}
 
 	if (ret)
@@ -1912,6 +1962,8 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 		goto out_release_extents;
 	}
 
+	btrfs_init_log_ctx_scratch_eb(&ctx);
+
 	/*
 	 * We use start here because we will need to wait on the IO to complete
 	 * in btrfs_sync_log, which could require joining a transaction (for
@@ -1931,6 +1983,15 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	trans->in_fsync = true;
 
 	ret = btrfs_log_dentry_safe(trans, dentry, &ctx);
+	/*
+	 * Scratch eb no longer needed, release before syncing log or commit
+	 * transaction, to avoid holding unnecessary memory during such long
+	 * operations.
+	 */
+	if (ctx.scratch_eb) {
+		free_extent_buffer(ctx.scratch_eb);
+		ctx.scratch_eb = NULL;
+	}
 	btrfs_release_log_ctx_extents(&ctx);
 	if (ret < 0) {
 		/* Fallthrough and commit/free transaction. */
@@ -1947,7 +2008,10 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	 * file again, but that will end up using the synchronization
 	 * inside btrfs_sync_log to keep things safe.
 	 */
-	btrfs_inode_unlock(BTRFS_I(inode), BTRFS_ILOCK_MMAP);
+	if (skip_ilock)
+		up_write(&BTRFS_I(inode)->i_mmap_lock);
+	else
+		btrfs_inode_unlock(BTRFS_I(inode), BTRFS_ILOCK_MMAP);
 
 	if (ret == BTRFS_NO_LOG_SYNC) {
 		ret = btrfs_end_transaction(trans);
@@ -2006,6 +2070,7 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 
 	ret = btrfs_commit_transaction(trans);
 out:
+	free_extent_buffer(ctx.scratch_eb);
 	ASSERT(list_empty(&ctx.list));
 	ASSERT(list_empty(&ctx.conflict_inodes));
 	err = file_check_and_advance_wb_err(file);
@@ -2015,10 +2080,179 @@ out:
 
 out_release_extents:
 	btrfs_release_log_ctx_extents(&ctx);
-	btrfs_inode_unlock(BTRFS_I(inode), BTRFS_ILOCK_MMAP);
+	if (skip_ilock)
+		up_write(&BTRFS_I(inode)->i_mmap_lock);
+	else
+		btrfs_inode_unlock(BTRFS_I(inode), BTRFS_ILOCK_MMAP);
 	goto out;
 }
 
+/*
+ * btrfs_page_mkwrite() is not allowed to change the file size as it gets
+ * called from a page fault handler when a page is first dirtied. Hence we must
+ * be careful to check for EOF conditions here. We set the page up correctly
+ * for a written page which means we get ENOSPC checking when writing into
+ * holes and correct delalloc and unwritten extent mapping on filesystems that
+ * support these features.
+ *
+ * We are not allowed to take the i_mutex here so we have to play games to
+ * protect against truncate races as the page could now be beyond EOF.  Because
+ * truncate_setsize() writes the inode size before removing pages, once we have
+ * the page lock we can determine safely if the page is beyond EOF. If it is not
+ * beyond EOF, then the page is guaranteed safe against truncation until we
+ * unlock the page.
+ */
+static vm_fault_t btrfs_page_mkwrite(struct vm_fault *vmf)
+{
+	struct page *page = vmf->page;
+	struct folio *folio = page_folio(page);
+	struct inode *inode = file_inode(vmf->vma->vm_file);
+	struct btrfs_fs_info *fs_info = inode_to_fs_info(inode);
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	struct btrfs_ordered_extent *ordered;
+	struct extent_state *cached_state = NULL;
+	struct extent_changeset *data_reserved = NULL;
+	unsigned long zero_start;
+	loff_t size;
+	vm_fault_t ret;
+	int ret2;
+	int reserved = 0;
+	u64 reserved_space;
+	u64 page_start;
+	u64 page_end;
+	u64 end;
+
+	ASSERT(folio_order(folio) == 0);
+
+	reserved_space = PAGE_SIZE;
+
+	sb_start_pagefault(inode->i_sb);
+	page_start = page_offset(page);
+	page_end = page_start + PAGE_SIZE - 1;
+	end = page_end;
+
+	/*
+	 * Reserving delalloc space after obtaining the page lock can lead to
+	 * deadlock. For example, if a dirty page is locked by this function
+	 * and the call to btrfs_delalloc_reserve_space() ends up triggering
+	 * dirty page write out, then the btrfs_writepages() function could
+	 * end up waiting indefinitely to get a lock on the page currently
+	 * being processed by btrfs_page_mkwrite() function.
+	 */
+	ret2 = btrfs_delalloc_reserve_space(BTRFS_I(inode), &data_reserved,
+					    page_start, reserved_space);
+	if (!ret2) {
+		ret2 = file_update_time(vmf->vma->vm_file);
+		reserved = 1;
+	}
+	if (ret2) {
+		ret = vmf_error(ret2);
+		if (reserved)
+			goto out;
+		goto out_noreserve;
+	}
+
+	/* Make the VM retry the fault. */
+	ret = VM_FAULT_NOPAGE;
+again:
+	down_read(&BTRFS_I(inode)->i_mmap_lock);
+	lock_page(page);
+	size = i_size_read(inode);
+
+	if ((page->mapping != inode->i_mapping) ||
+	    (page_start >= size)) {
+		/* Page got truncated out from underneath us. */
+		goto out_unlock;
+	}
+	wait_on_page_writeback(page);
+
+	lock_extent(io_tree, page_start, page_end, &cached_state);
+	ret2 = set_page_extent_mapped(page);
+	if (ret2 < 0) {
+		ret = vmf_error(ret2);
+		unlock_extent(io_tree, page_start, page_end, &cached_state);
+		goto out_unlock;
+	}
+
+	/*
+	 * We can't set the delalloc bits if there are pending ordered
+	 * extents.  Drop our locks and wait for them to finish.
+	 */
+	ordered = btrfs_lookup_ordered_range(BTRFS_I(inode), page_start, PAGE_SIZE);
+	if (ordered) {
+		unlock_extent(io_tree, page_start, page_end, &cached_state);
+		unlock_page(page);
+		up_read(&BTRFS_I(inode)->i_mmap_lock);
+		btrfs_start_ordered_extent(ordered);
+		btrfs_put_ordered_extent(ordered);
+		goto again;
+	}
+
+	if (page->index == ((size - 1) >> PAGE_SHIFT)) {
+		reserved_space = round_up(size - page_start, fs_info->sectorsize);
+		if (reserved_space < PAGE_SIZE) {
+			end = page_start + reserved_space - 1;
+			btrfs_delalloc_release_space(BTRFS_I(inode),
+					data_reserved, page_start,
+					PAGE_SIZE - reserved_space, true);
+		}
+	}
+
+	/*
+	 * page_mkwrite gets called when the page is firstly dirtied after it's
+	 * faulted in, but write(2) could also dirty a page and set delalloc
+	 * bits, thus in this case for space account reason, we still need to
+	 * clear any delalloc bits within this page range since we have to
+	 * reserve data&meta space before lock_page() (see above comments).
+	 */
+	clear_extent_bit(&BTRFS_I(inode)->io_tree, page_start, end,
+			  EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING |
+			  EXTENT_DEFRAG, &cached_state);
+
+	ret2 = btrfs_set_extent_delalloc(BTRFS_I(inode), page_start, end, 0,
+					&cached_state);
+	if (ret2) {
+		unlock_extent(io_tree, page_start, page_end, &cached_state);
+		ret = VM_FAULT_SIGBUS;
+		goto out_unlock;
+	}
+
+	/* Page is wholly or partially inside EOF. */
+	if (page_start + PAGE_SIZE > size)
+		zero_start = offset_in_page(size);
+	else
+		zero_start = PAGE_SIZE;
+
+	if (zero_start != PAGE_SIZE)
+		memzero_page(page, zero_start, PAGE_SIZE - zero_start);
+
+	btrfs_folio_clear_checked(fs_info, folio, page_start, PAGE_SIZE);
+	btrfs_folio_set_dirty(fs_info, folio, page_start, end + 1 - page_start);
+	btrfs_folio_set_uptodate(fs_info, folio, page_start, end + 1 - page_start);
+
+	btrfs_set_inode_last_sub_trans(BTRFS_I(inode));
+
+	unlock_extent(io_tree, page_start, page_end, &cached_state);
+	up_read(&BTRFS_I(inode)->i_mmap_lock);
+
+	btrfs_delalloc_release_extents(BTRFS_I(inode), PAGE_SIZE);
+	sb_end_pagefault(inode->i_sb);
+	extent_changeset_free(data_reserved);
+	return VM_FAULT_LOCKED;
+
+out_unlock:
+	unlock_page(page);
+	up_read(&BTRFS_I(inode)->i_mmap_lock);
+out:
+	btrfs_delalloc_release_extents(BTRFS_I(inode), PAGE_SIZE);
+	btrfs_delalloc_release_space(BTRFS_I(inode), data_reserved, page_start,
+				     reserved_space, (ret != 0));
+out_noreserve:
+	sb_end_pagefault(inode->i_sb);
+	extent_changeset_free(data_reserved);
+	return ret;
+}
+
 static const struct vm_operations_struct btrfs_file_vm_ops = {
 	.fault		= filemap_fault,
 	.map_pages	= filemap_map_pages,
@@ -2176,7 +2410,7 @@ static int find_first_non_hole(struct btrfs_inode *inode, u64 *start, u64 *len)
 	struct extent_map *em;
 	int ret = 0;
 
-	em = btrfs_get_extent(inode, NULL, 0,
+	em = btrfs_get_extent(inode, NULL,
 			      round_down(*start, fs_info->sectorsize),
 			      round_up(*len, fs_info->sectorsize));
 	if (IS_ERR(em))
@@ -2248,7 +2482,6 @@ static int btrfs_insert_replace_extent(struct btrfs_trans_handle *trans,
 	struct extent_buffer *leaf;
 	struct btrfs_key key;
 	int slot;
-	struct btrfs_ref ref = { 0 };
 	int ret;
 
 	if (replace_len == 0)
@@ -2304,15 +2537,17 @@ static int btrfs_insert_replace_extent(struct btrfs_trans_handle *trans,
 						       extent_info->qgroup_reserved,
 						       &key);
 	} else {
+		struct btrfs_ref ref = {
+			.action = BTRFS_ADD_DELAYED_REF,
+			.bytenr = extent_info->disk_offset,
+			.num_bytes = extent_info->disk_len,
+			.owning_root = btrfs_root_id(root),
+			.ref_root = btrfs_root_id(root),
+		};
 		u64 ref_offset;
 
-		btrfs_init_generic_ref(&ref, BTRFS_ADD_DELAYED_REF,
-				       extent_info->disk_offset,
-				       extent_info->disk_len, 0,
-				       root->root_key.objectid);
 		ref_offset = extent_info->file_offset - extent_info->data_offset;
-		btrfs_init_data_ref(&ref, root->root_key.objectid,
-				    btrfs_ino(inode), ref_offset, 0, false);
+		btrfs_init_data_ref(&ref, btrfs_ino(inode), ref_offset, 0, false);
 		ret = btrfs_inc_extent_ref(trans, &ref);
 	}
 
@@ -2835,7 +3070,7 @@ static int btrfs_zero_range_check_range_boundary(struct btrfs_inode *inode,
 	int ret;
 
 	offset = round_down(offset, sectorsize);
-	em = btrfs_get_extent(inode, NULL, 0, offset, sectorsize);
+	em = btrfs_get_extent(inode, NULL, offset, sectorsize);
 	if (IS_ERR(em))
 		return PTR_ERR(em);
 
@@ -2866,7 +3101,7 @@ static int btrfs_zero_range(struct inode *inode,
 	u64 bytes_to_reserve = 0;
 	bool space_reserved = false;
 
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, alloc_start,
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, alloc_start,
 			      alloc_end - alloc_start);
 	if (IS_ERR(em)) {
 		ret = PTR_ERR(em);
@@ -2909,8 +3144,7 @@ static int btrfs_zero_range(struct inode *inode,
 
 	if (BTRFS_BYTES_TO_BLKS(fs_info, offset) ==
 	    BTRFS_BYTES_TO_BLKS(fs_info, offset + len - 1)) {
-		em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, alloc_start,
-				      sectorsize);
+		em = btrfs_get_extent(BTRFS_I(inode), NULL, alloc_start, sectorsize);
 		if (IS_ERR(em)) {
 			ret = PTR_ERR(em);
 			goto out;
@@ -3005,7 +3239,7 @@ reserve_space:
 		}
 		ret = btrfs_prealloc_file_range(inode, mode, alloc_start,
 						alloc_end - alloc_start,
-						i_blocksize(inode),
+						fs_info->sectorsize,
 						offset + len, &alloc_hint);
 		unlock_extent(&BTRFS_I(inode)->io_tree, lockstart, lockend,
 			      &cached_state);
@@ -3126,7 +3360,7 @@ static long btrfs_fallocate(struct file *file, int mode,
 
 	/* First, check if we exceed the qgroup limit */
 	while (cur_offset < alloc_end) {
-		em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, cur_offset,
+		em = btrfs_get_extent(BTRFS_I(inode), NULL, cur_offset,
 				      alloc_end - cur_offset);
 		if (IS_ERR(em)) {
 			ret = PTR_ERR(em);
@@ -3177,7 +3411,7 @@ static long btrfs_fallocate(struct file *file, int mode,
 		if (!ret) {
 			ret = btrfs_prealloc_file_range(inode, mode,
 					range->start,
-					range->len, i_blocksize(inode),
+					range->len, blocksize,
 					offset + len, &alloc_hint);
 			/*
 			 * btrfs_prealloc_file_range() releases space even
@@ -3710,8 +3944,7 @@ static int btrfs_file_open(struct inode *inode, struct file *filp)
 {
 	int ret;
 
-	filp->f_mode |= FMODE_NOWAIT | FMODE_BUF_RASYNC | FMODE_BUF_WASYNC |
-		        FMODE_CAN_ODIRECT;
+	filp->f_mode |= FMODE_NOWAIT | FMODE_CAN_ODIRECT;
 
 	ret = fsverity_file_open(inode, filp);
 	if (ret)
@@ -3841,6 +4074,7 @@ const struct file_operations btrfs_file_operations = {
 	.compat_ioctl	= btrfs_compat_ioctl,
 #endif
 	.remap_file_range = btrfs_remap_file_range,
+	.fop_flags	= FOP_BUFFER_RASYNC | FOP_BUFFER_WASYNC,
 };
 
 int btrfs_fdatawrite_range(struct inode *inode, loff_t start, loff_t end)
diff --git a/fs/btrfs/file.h b/fs/btrfs/file.h
index 82b34fbb29..77aaca208c 100644
--- a/fs/btrfs/file.h
+++ b/fs/btrfs/file.h
@@ -3,6 +3,21 @@
 #ifndef BTRFS_FILE_H
 #define BTRFS_FILE_H
 
+#include <linux/types.h>
+
+struct file;
+struct extent_state;
+struct kiocb;
+struct iov_iter;
+struct page;
+struct btrfs_ioctl_encoded_io_args;
+struct btrfs_drop_extents_args;
+struct btrfs_inode;
+struct btrfs_root;
+struct btrfs_path;
+struct btrfs_replace_extent_info;
+struct btrfs_trans_handle;
+
 extern const struct file_operations btrfs_file_operations;
 
 int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync);
diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c
index 815bb146b1..62c3dea957 100644
--- a/fs/btrfs/free-space-cache.c
+++ b/fs/btrfs/free-space-cache.c
@@ -19,9 +19,7 @@
 #include "transaction.h"
 #include "disk-io.h"
 #include "extent_io.h"
-#include "volumes.h"
 #include "space-info.h"
-#include "delalloc-space.h"
 #include "block-group.h"
 #include "discard.h"
 #include "subpage.h"
@@ -860,6 +858,7 @@ static int __load_free_space_cache(struct btrfs_root *root, struct inode *inode,
 				spin_unlock(&ctl->tree_lock);
 				btrfs_err(fs_info,
 					"Duplicate entries in free space cache, dumping");
+				kmem_cache_free(btrfs_free_space_bitmap_cachep, e->bitmap);
 				kmem_cache_free(btrfs_free_space_cachep, e);
 				goto free_cache;
 			}
@@ -1913,9 +1912,9 @@ static inline void bitmap_clear_bits(struct btrfs_free_space_ctl *ctl,
 		ctl->free_space -= bytes;
 }
 
-static void bitmap_set_bits(struct btrfs_free_space_ctl *ctl,
-			    struct btrfs_free_space *info, u64 offset,
-			    u64 bytes)
+static void btrfs_bitmap_set_bits(struct btrfs_free_space_ctl *ctl,
+				  struct btrfs_free_space *info, u64 offset,
+				  u64 bytes)
 {
 	unsigned long start, count, end;
 	int extent_delta = 1;
@@ -2251,7 +2250,7 @@ static u64 add_bytes_to_bitmap(struct btrfs_free_space_ctl *ctl,
 
 	bytes_to_set = min(end - offset, bytes);
 
-	bitmap_set_bits(ctl, info, offset, bytes_to_set);
+	btrfs_bitmap_set_bits(ctl, info, offset, bytes_to_set);
 
 	return bytes_to_set;
 
@@ -2621,7 +2620,7 @@ static void steal_from_bitmap(struct btrfs_free_space_ctl *ctl,
 	}
 }
 
-int __btrfs_add_free_space(struct btrfs_block_group *block_group,
+static int __btrfs_add_free_space(struct btrfs_block_group *block_group,
 			   u64 offset, u64 bytes,
 			   enum btrfs_trim_state trim_state)
 {
@@ -2699,7 +2698,7 @@ static int __btrfs_add_free_space_zoned(struct btrfs_block_group *block_group,
 	u64 offset = bytenr - block_group->start;
 	u64 to_free, to_unusable;
 	int bg_reclaim_threshold = 0;
-	bool initial = (size == block_group->length);
+	bool initial = ((size == block_group->length) && (block_group->alloc_offset == 0));
 	u64 reclaimable_unusable;
 
 	WARN_ON(!initial && offset + size > block_group->zone_capacity);
@@ -2725,8 +2724,10 @@ static int __btrfs_add_free_space_zoned(struct btrfs_block_group *block_group,
 	 * If the block group is read-only, we should account freed space into
 	 * bytes_readonly.
 	 */
-	if (!block_group->ro)
+	if (!block_group->ro) {
 		block_group->zone_unusable += to_unusable;
+		WARN_ON(block_group->zone_unusable > block_group->length);
+	}
 	spin_unlock(&ctl->tree_lock);
 	if (!used) {
 		spin_lock(&block_group->lock);
@@ -4156,15 +4157,13 @@ out:
 
 int __init btrfs_free_space_init(void)
 {
-	btrfs_free_space_cachep = kmem_cache_create("btrfs_free_space",
-			sizeof(struct btrfs_free_space), 0,
-			SLAB_MEM_SPREAD, NULL);
+	btrfs_free_space_cachep = KMEM_CACHE(btrfs_free_space, 0);
 	if (!btrfs_free_space_cachep)
 		return -ENOMEM;
 
 	btrfs_free_space_bitmap_cachep = kmem_cache_create("btrfs_free_space_bitmap",
 							PAGE_SIZE, PAGE_SIZE,
-							SLAB_MEM_SPREAD, NULL);
+							0, NULL);
 	if (!btrfs_free_space_bitmap_cachep) {
 		kmem_cache_destroy(btrfs_free_space_cachep);
 		return -ENOMEM;
diff --git a/fs/btrfs/free-space-cache.h b/fs/btrfs/free-space-cache.h
index 33b4da3271..83774bfd7b 100644
--- a/fs/btrfs/free-space-cache.h
+++ b/fs/btrfs/free-space-cache.h
@@ -6,6 +6,19 @@
 #ifndef BTRFS_FREE_SPACE_CACHE_H
 #define BTRFS_FREE_SPACE_CACHE_H
 
+#include <linux/rbtree.h>
+#include <linux/list.h>
+#include <linux/spinlock.h>
+#include <linux/mutex.h>
+#include "fs.h"
+
+struct inode;
+struct page;
+struct btrfs_fs_info;
+struct btrfs_path;
+struct btrfs_trans_handle;
+struct btrfs_trim_block_group;
+
 /*
  * This is the trim state of an extent or bitmap.
  *
@@ -114,8 +127,6 @@ int btrfs_write_out_cache(struct btrfs_trans_handle *trans,
 
 void btrfs_init_free_space_ctl(struct btrfs_block_group *block_group,
 			       struct btrfs_free_space_ctl *ctl);
-int __btrfs_add_free_space(struct btrfs_block_group *block_group, u64 bytenr,
-			   u64 size, enum btrfs_trim_state trim_state);
 int btrfs_add_free_space(struct btrfs_block_group *block_group,
 			 u64 bytenr, u64 size);
 int btrfs_add_free_space_unused(struct btrfs_block_group *block_group,
diff --git a/fs/btrfs/free-space-tree.c b/fs/btrfs/free-space-tree.c
index 7b598b0707..7ba50e1339 100644
--- a/fs/btrfs/free-space-tree.c
+++ b/fs/btrfs/free-space-tree.c
@@ -1176,12 +1176,16 @@ int btrfs_create_free_space_tree(struct btrfs_fs_info *fs_info)
 					    BTRFS_FREE_SPACE_TREE_OBJECTID);
 	if (IS_ERR(free_space_root)) {
 		ret = PTR_ERR(free_space_root);
-		goto abort;
+		btrfs_abort_transaction(trans, ret);
+		btrfs_end_transaction(trans);
+		goto out_clear;
 	}
 	ret = btrfs_global_root_insert(free_space_root);
 	if (ret) {
 		btrfs_put_root(free_space_root);
-		goto abort;
+		btrfs_abort_transaction(trans, ret);
+		btrfs_end_transaction(trans);
+		goto out_clear;
 	}
 
 	node = rb_first_cached(&fs_info->block_group_cache_tree);
@@ -1189,8 +1193,11 @@ int btrfs_create_free_space_tree(struct btrfs_fs_info *fs_info)
 		block_group = rb_entry(node, struct btrfs_block_group,
 				       cache_node);
 		ret = populate_free_space_tree(trans, block_group);
-		if (ret)
-			goto abort;
+		if (ret) {
+			btrfs_abort_transaction(trans, ret);
+			btrfs_end_transaction(trans);
+			goto out_clear;
+		}
 		node = rb_next(node);
 	}
 
@@ -1206,11 +1213,9 @@ int btrfs_create_free_space_tree(struct btrfs_fs_info *fs_info)
 	clear_bit(BTRFS_FS_FREE_SPACE_TREE_UNTRUSTED, &fs_info->flags);
 	return ret;
 
-abort:
+out_clear:
 	clear_bit(BTRFS_FS_CREATING_FREE_SPACE_TREE, &fs_info->flags);
 	clear_bit(BTRFS_FS_FREE_SPACE_TREE_UNTRUSTED, &fs_info->flags);
-	btrfs_abort_transaction(trans, ret);
-	btrfs_end_transaction(trans);
 	return ret;
 }
 
@@ -1273,12 +1278,18 @@ int btrfs_delete_free_space_tree(struct btrfs_fs_info *fs_info)
 	btrfs_clear_fs_compat_ro(fs_info, FREE_SPACE_TREE_VALID);
 
 	ret = clear_free_space_tree(trans, free_space_root);
-	if (ret)
-		goto abort;
+	if (ret) {
+		btrfs_abort_transaction(trans, ret);
+		btrfs_end_transaction(trans);
+		return ret;
+	}
 
 	ret = btrfs_del_root(trans, &free_space_root->root_key);
-	if (ret)
-		goto abort;
+	if (ret) {
+		btrfs_abort_transaction(trans, ret);
+		btrfs_end_transaction(trans);
+		return ret;
+	}
 
 	btrfs_global_root_delete(free_space_root);
 
@@ -1289,17 +1300,16 @@ int btrfs_delete_free_space_tree(struct btrfs_fs_info *fs_info)
 	btrfs_tree_lock(free_space_root->node);
 	btrfs_clear_buffer_dirty(trans, free_space_root->node);
 	btrfs_tree_unlock(free_space_root->node);
-	btrfs_free_tree_block(trans, btrfs_root_id(free_space_root),
-			      free_space_root->node, 0, 1);
-
+	ret = btrfs_free_tree_block(trans, btrfs_root_id(free_space_root),
+				    free_space_root->node, 0, 1);
 	btrfs_put_root(free_space_root);
+	if (ret < 0) {
+		btrfs_abort_transaction(trans, ret);
+		btrfs_end_transaction(trans);
+		return ret;
+	}
 
 	return btrfs_commit_transaction(trans);
-
-abort:
-	btrfs_abort_transaction(trans, ret);
-	btrfs_end_transaction(trans);
-	return ret;
 }
 
 int btrfs_rebuild_free_space_tree(struct btrfs_fs_info *fs_info)
@@ -1322,8 +1332,11 @@ int btrfs_rebuild_free_space_tree(struct btrfs_fs_info *fs_info)
 	set_bit(BTRFS_FS_FREE_SPACE_TREE_UNTRUSTED, &fs_info->flags);
 
 	ret = clear_free_space_tree(trans, free_space_root);
-	if (ret)
-		goto abort;
+	if (ret) {
+		btrfs_abort_transaction(trans, ret);
+		btrfs_end_transaction(trans);
+		return ret;
+	}
 
 	node = rb_first_cached(&fs_info->block_group_cache_tree);
 	while (node) {
@@ -1332,8 +1345,11 @@ int btrfs_rebuild_free_space_tree(struct btrfs_fs_info *fs_info)
 		block_group = rb_entry(node, struct btrfs_block_group,
 				       cache_node);
 		ret = populate_free_space_tree(trans, block_group);
-		if (ret)
-			goto abort;
+		if (ret) {
+			btrfs_abort_transaction(trans, ret);
+			btrfs_end_transaction(trans);
+			return ret;
+		}
 		node = rb_next(node);
 	}
 
@@ -1344,10 +1360,6 @@ int btrfs_rebuild_free_space_tree(struct btrfs_fs_info *fs_info)
 	ret = btrfs_commit_transaction(trans);
 	clear_bit(BTRFS_FS_FREE_SPACE_TREE_UNTRUSTED, &fs_info->flags);
 	return ret;
-abort:
-	btrfs_abort_transaction(trans, ret);
-	btrfs_end_transaction(trans);
-	return ret;
 }
 
 static int __add_block_group_free_space(struct btrfs_trans_handle *trans,
diff --git a/fs/btrfs/free-space-tree.h b/fs/btrfs/free-space-tree.h
index 6d5551d0ce..e6c6d6f4f2 100644
--- a/fs/btrfs/free-space-tree.h
+++ b/fs/btrfs/free-space-tree.h
@@ -6,7 +6,13 @@
 #ifndef BTRFS_FREE_SPACE_TREE_H
 #define BTRFS_FREE_SPACE_TREE_H
 
+#include <linux/bits.h>
+
 struct btrfs_caching_control;
+struct btrfs_fs_info;
+struct btrfs_path;
+struct btrfs_block_group;
+struct btrfs_trans_handle;
 
 /*
  * The default size for new free space bitmap items. The last bitmap in a block
diff --git a/fs/btrfs/fs.h b/fs/btrfs/fs.h
index bd59cf0aae..833dc3fe0a 100644
--- a/fs/btrfs/fs.h
+++ b/fs/btrfs/fs.h
@@ -4,13 +4,49 @@
 #define BTRFS_FS_H
 
 #include <linux/blkdev.h>
-#include <linux/fs.h>
-#include <linux/btrfs_tree.h>
 #include <linux/sizes.h>
+#include <linux/time64.h>
+#include <linux/compiler.h>
+#include <linux/math.h>
+#include <linux/atomic.h>
+#include <linux/percpu_counter.h>
+#include <linux/completion.h>
+#include <linux/lockdep.h>
+#include <linux/spinlock.h>
+#include <linux/mutex.h>
+#include <linux/rwlock_types.h>
+#include <linux/rwsem.h>
+#include <linux/semaphore.h>
+#include <linux/list.h>
+#include <linux/radix-tree.h>
+#include <linux/workqueue.h>
+#include <linux/wait.h>
+#include <linux/wait_bit.h>
+#include <linux/sched.h>
+#include <linux/rbtree.h>
+#include <uapi/linux/btrfs.h>
+#include <uapi/linux/btrfs_tree.h>
 #include "extent-io-tree.h"
-#include "extent_map.h"
 #include "async-thread.h"
 #include "block-rsv.h"
+#include "fs.h"
+
+struct inode;
+struct super_block;
+struct kobject;
+struct reloc_control;
+struct crypto_shash;
+struct ulist;
+struct btrfs_device;
+struct btrfs_block_group;
+struct btrfs_root;
+struct btrfs_fs_devices;
+struct btrfs_transaction;
+struct btrfs_delayed_root;
+struct btrfs_balance_control;
+struct btrfs_subpage_info;
+struct btrfs_stripe_hash_table;
+struct btrfs_space_info;
 
 #define BTRFS_MAX_EXTENT_SIZE SZ_128M
 
@@ -593,6 +629,11 @@ struct btrfs_fs_info {
 	s32 dirty_metadata_batch;
 	s32 delalloc_batch;
 
+	struct percpu_counter evictable_extent_maps;
+	spinlock_t extent_map_shrinker_lock;
+	u64 extent_map_shrinker_last_root;
+	u64 extent_map_shrinker_last_ino;
+
 	/* Protected by 'trans_lock'. */
 	struct list_head dirty_cowonly_roots;
 
@@ -732,10 +773,13 @@ struct btrfs_fs_info {
 
 	/* Reclaim partially filled block groups in the background */
 	struct work_struct reclaim_bgs_work;
+	/* Protected by unused_bgs_lock. */
 	struct list_head reclaim_bgs;
 	int bg_reclaim_threshold;
 
+	/* Protects the lists unused_bgs and reclaim_bgs. */
 	spinlock_t unused_bgs_lock;
+	/* Protected by unused_bgs_lock. */
 	struct list_head unused_bgs;
 	struct mutex unused_bg_unpin_mutex;
 	/* Protect block groups that are going to be deleted */
@@ -933,6 +977,8 @@ void btrfs_exclop_finish(struct btrfs_fs_info *fs_info);
 void btrfs_exclop_balance(struct btrfs_fs_info *fs_info,
 			  enum btrfs_exclusive_operation op);
 
+int btrfs_check_ioctl_vol_args_path(const struct btrfs_ioctl_vol_args *vol_args);
+
 /* Compatibility and incompatibility defines */
 void __btrfs_set_fs_incompat(struct btrfs_fs_info *fs_info, u64 flag,
 			     const char *name);
diff --git a/fs/btrfs/inode-item.c b/fs/btrfs/inode-item.c
index 7d734830e5..84a94d19b2 100644
--- a/fs/btrfs/inode-item.c
+++ b/fs/btrfs/inode-item.c
@@ -9,7 +9,6 @@
 #include "inode-item.h"
 #include "disk-io.h"
 #include "transaction.h"
-#include "print-tree.h"
 #include "space-info.h"
 #include "accessors.h"
 #include "extent-tree.h"
@@ -671,16 +670,18 @@ delete:
 		}
 
 		if (del_item && extent_start != 0 && !control->skip_ref_updates) {
-			struct btrfs_ref ref = { 0 };
+			struct btrfs_ref ref = {
+				.action = BTRFS_DROP_DELAYED_REF,
+				.bytenr = extent_start,
+				.num_bytes = extent_num_bytes,
+				.owning_root = btrfs_root_id(root),
+				.ref_root = btrfs_header_owner(leaf),
+			};
 
 			bytes_deleted += extent_num_bytes;
 
-			btrfs_init_generic_ref(&ref, BTRFS_DROP_DELAYED_REF,
-					extent_start, extent_num_bytes, 0,
-					root->root_key.objectid);
-			btrfs_init_data_ref(&ref, btrfs_header_owner(leaf),
-					control->ino, extent_offset,
-					root->root_key.objectid, false);
+			btrfs_init_data_ref(&ref, control->ino, extent_offset,
+					    btrfs_root_id(root), false);
 			ret = btrfs_free_extent(trans, &ref);
 			if (ret) {
 				btrfs_abort_transaction(trans, ret);
diff --git a/fs/btrfs/inode-item.h b/fs/btrfs/inode-item.h
index 4337bb26f4..c4aded8270 100644
--- a/fs/btrfs/inode-item.h
+++ b/fs/btrfs/inode-item.h
@@ -6,14 +6,15 @@
 #include <linux/types.h>
 #include <linux/crc32c.h>
 
+struct fscrypt_str;
+struct extent_buffer;
 struct btrfs_trans_handle;
 struct btrfs_root;
 struct btrfs_path;
 struct btrfs_key;
 struct btrfs_inode_extref;
 struct btrfs_inode;
-struct extent_buffer;
-struct fscrypt_str;
+struct btrfs_truncate_control;
 
 /*
  * Return this if we need to call truncate_block for the last bit of the
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index d3c534c1bf..39d22693e4 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -39,14 +39,12 @@
 #include "disk-io.h"
 #include "transaction.h"
 #include "btrfs_inode.h"
-#include "print-tree.h"
 #include "ordered-data.h"
 #include "xattr.h"
 #include "tree-log.h"
 #include "bio.h"
 #include "compression.h"
 #include "locking.h"
-#include "free-space-cache.h"
 #include "props.h"
 #include "qgroup.h"
 #include "delalloc-space.h"
@@ -256,7 +254,7 @@ static void print_data_reloc_error(const struct btrfs_inode *inode, u64 file_off
 		btrfs_warn_rl(fs_info, "has data reloc tree but no running relocation");
 		btrfs_warn_rl(fs_info,
 "csum failed root %lld ino %llu off %llu csum " CSUM_FMT " expected csum " CSUM_FMT " mirror %d",
-			inode->root->root_key.objectid, btrfs_ino(inode), file_off,
+			btrfs_root_id(inode->root), btrfs_ino(inode), file_off,
 			CSUM_FMT_VALUE(csum_size, csum),
 			CSUM_FMT_VALUE(csum_size, csum_expected),
 			mirror_num);
@@ -266,7 +264,7 @@ static void print_data_reloc_error(const struct btrfs_inode *inode, u64 file_off
 	logical += file_off;
 	btrfs_warn_rl(fs_info,
 "csum failed root %lld ino %llu off %llu logical %llu csum " CSUM_FMT " expected csum " CSUM_FMT " mirror %d",
-			inode->root->root_key.objectid,
+			btrfs_root_id(inode->root),
 			btrfs_ino(inode), file_off, logical,
 			CSUM_FMT_VALUE(csum_size, csum),
 			CSUM_FMT_VALUE(csum_size, csum_expected),
@@ -333,15 +331,15 @@ static void __cold btrfs_print_data_csum_error(struct btrfs_inode *inode,
 	const u32 csum_size = root->fs_info->csum_size;
 
 	/* For data reloc tree, it's better to do a backref lookup instead. */
-	if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID)
+	if (btrfs_root_id(root) == BTRFS_DATA_RELOC_TREE_OBJECTID)
 		return print_data_reloc_error(inode, logical_start, csum,
 					      csum_expected, mirror_num);
 
 	/* Output without objectid, which is more meaningful */
-	if (root->root_key.objectid >= BTRFS_LAST_FREE_OBJECTID) {
+	if (btrfs_root_id(root) >= BTRFS_LAST_FREE_OBJECTID) {
 		btrfs_warn_rl(root->fs_info,
 "csum failed root %lld ino %lld off %llu csum " CSUM_FMT " expected csum " CSUM_FMT " mirror %d",
-			root->root_key.objectid, btrfs_ino(inode),
+			btrfs_root_id(root), btrfs_ino(inode),
 			logical_start,
 			CSUM_FMT_VALUE(csum_size, csum),
 			CSUM_FMT_VALUE(csum_size, csum_expected),
@@ -349,7 +347,7 @@ static void __cold btrfs_print_data_csum_error(struct btrfs_inode *inode,
 	} else {
 		btrfs_warn_rl(root->fs_info,
 "csum failed root %llu ino %llu off %llu csum " CSUM_FMT " expected csum " CSUM_FMT " mirror %d",
-			root->root_key.objectid, btrfs_ino(inode),
+			btrfs_root_id(root), btrfs_ino(inode),
 			logical_start,
 			CSUM_FMT_VALUE(csum_size, csum),
 			CSUM_FMT_VALUE(csum_size, csum_expected),
@@ -514,12 +512,13 @@ static int insert_inline_extent(struct btrfs_trans_handle *trans,
 				struct btrfs_inode *inode, bool extent_inserted,
 				size_t size, size_t compressed_size,
 				int compress_type,
-				struct page **compressed_pages,
+				struct folio *compressed_folio,
 				bool update_i_size)
 {
 	struct btrfs_root *root = inode->root;
 	struct extent_buffer *leaf;
 	struct page *page = NULL;
+	const u32 sectorsize = trans->fs_info->sectorsize;
 	char *kaddr;
 	unsigned long ptr;
 	struct btrfs_file_extent_item *ei;
@@ -527,10 +526,23 @@ static int insert_inline_extent(struct btrfs_trans_handle *trans,
 	size_t cur_size = size;
 	u64 i_size;
 
-	ASSERT((compressed_size > 0 && compressed_pages) ||
-	       (compressed_size == 0 && !compressed_pages));
+	/*
+	 * The decompressed size must still be no larger than a sector.  Under
+	 * heavy race, we can have size == 0 passed in, but that shouldn't be a
+	 * big deal and we can continue the insertion.
+	 */
+	ASSERT(size <= sectorsize);
 
-	if (compressed_size && compressed_pages)
+	/*
+	 * The compressed size also needs to be no larger than a sector.
+	 * That's also why we only need one page as the parameter.
+	 */
+	if (compressed_folio)
+		ASSERT(compressed_size <= sectorsize);
+	else
+		ASSERT(compressed_size == 0);
+
+	if (compressed_size && compressed_folio)
 		cur_size = compressed_size;
 
 	if (!extent_inserted) {
@@ -558,21 +570,10 @@ static int insert_inline_extent(struct btrfs_trans_handle *trans,
 	ptr = btrfs_file_extent_inline_start(ei);
 
 	if (compress_type != BTRFS_COMPRESS_NONE) {
-		struct page *cpage;
-		int i = 0;
-		while (compressed_size > 0) {
-			cpage = compressed_pages[i];
-			cur_size = min_t(unsigned long, compressed_size,
-				       PAGE_SIZE);
-
-			kaddr = kmap_local_page(cpage);
-			write_extent_buffer(leaf, kaddr, ptr, cur_size);
-			kunmap_local(kaddr);
+		kaddr = kmap_local_folio(compressed_folio, 0);
+		write_extent_buffer(leaf, kaddr, ptr, compressed_size);
+		kunmap_local(kaddr);
 
-			i++;
-			ptr += cur_size;
-			compressed_size -= cur_size;
-		}
 		btrfs_set_file_extent_compression(leaf, ei,
 						  compress_type);
 	} else {
@@ -613,17 +614,62 @@ fail:
 	return ret;
 }
 
+static bool can_cow_file_range_inline(struct btrfs_inode *inode,
+				      u64 offset, u64 size,
+				      size_t compressed_size)
+{
+	struct btrfs_fs_info *fs_info = inode->root->fs_info;
+	u64 data_len = (compressed_size ?: size);
+
+	/* Inline extents must start at offset 0. */
+	if (offset != 0)
+		return false;
+
+	/*
+	 * Due to the page size limit, for subpage we can only trigger the
+	 * writeback for the dirty sectors of page, that means data writeback
+	 * is doing more writeback than what we want.
+	 *
+	 * This is especially unexpected for some call sites like fallocate,
+	 * where we only increase i_size after everything is done.
+	 * This means we can trigger inline extent even if we didn't want to.
+	 * So here we skip inline extent creation completely.
+	 */
+	if (fs_info->sectorsize != PAGE_SIZE)
+		return false;
+
+	/* Inline extents are limited to sectorsize. */
+	if (size > fs_info->sectorsize)
+		return false;
+
+	/* We cannot exceed the maximum inline data size. */
+	if (data_len > BTRFS_MAX_INLINE_DATA_SIZE(fs_info))
+		return false;
+
+	/* We cannot exceed the user specified max_inline size. */
+	if (data_len > fs_info->max_inline)
+		return false;
+
+	/* Inline extents must be the entirety of the file. */
+	if (size < i_size_read(&inode->vfs_inode))
+		return false;
+
+	return true;
+}
 
 /*
  * conditionally insert an inline extent into the file.  This
  * does the checks required to make sure the data is small enough
  * to fit as an inline extent.
+ *
+ * If being used directly, you must have already checked we're allowed to cow
+ * the range by getting true from can_cow_file_range_inline().
  */
-static noinline int cow_file_range_inline(struct btrfs_inode *inode, u64 size,
-					  size_t compressed_size,
-					  int compress_type,
-					  struct page **compressed_pages,
-					  bool update_i_size)
+static noinline int __cow_file_range_inline(struct btrfs_inode *inode, u64 offset,
+					    u64 size, size_t compressed_size,
+					    int compress_type,
+					    struct folio *compressed_folio,
+					    bool update_i_size)
 {
 	struct btrfs_drop_extents_args drop_args = { 0 };
 	struct btrfs_root *root = inode->root;
@@ -633,18 +679,6 @@ static noinline int cow_file_range_inline(struct btrfs_inode *inode, u64 size,
 	int ret;
 	struct btrfs_path *path;
 
-	/*
-	 * We can create an inline extent if it ends at or beyond the current
-	 * i_size, is no larger than a sector (decompressed), and the (possibly
-	 * compressed) data fits in a leaf and the configured maximum inline
-	 * size.
-	 */
-	if (size < i_size_read(&inode->vfs_inode) ||
-	    size > fs_info->sectorsize ||
-	    data_len > BTRFS_MAX_INLINE_DATA_SIZE(fs_info) ||
-	    data_len > fs_info->max_inline)
-		return 1;
-
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
@@ -670,7 +704,7 @@ static noinline int cow_file_range_inline(struct btrfs_inode *inode, u64 size,
 
 	ret = insert_inline_extent(trans, path, inode, drop_args.extent_inserted,
 				   size, compressed_size, compress_type,
-				   compressed_pages, update_i_size);
+				   compressed_folio, update_i_size);
 	if (ret && ret != -ENOSPC) {
 		btrfs_abort_transaction(trans, ret);
 		goto out;
@@ -703,12 +737,48 @@ out:
 	return ret;
 }
 
+static noinline int cow_file_range_inline(struct btrfs_inode *inode,
+					  struct page *locked_page,
+					  u64 offset, u64 end,
+					  size_t compressed_size,
+					  int compress_type,
+					  struct folio *compressed_folio,
+					  bool update_i_size)
+{
+	struct extent_state *cached = NULL;
+	unsigned long clear_flags = EXTENT_DELALLOC | EXTENT_DELALLOC_NEW |
+		EXTENT_DEFRAG | EXTENT_DO_ACCOUNTING | EXTENT_LOCKED;
+	u64 size = min_t(u64, i_size_read(&inode->vfs_inode), end + 1);
+	int ret;
+
+	if (!can_cow_file_range_inline(inode, offset, size, compressed_size))
+		return 1;
+
+	lock_extent(&inode->io_tree, offset, end, &cached);
+	ret = __cow_file_range_inline(inode, offset, size, compressed_size,
+				      compress_type, compressed_folio,
+				      update_i_size);
+	if (ret > 0) {
+		unlock_extent(&inode->io_tree, offset, end, &cached);
+		return ret;
+	}
+
+	if (ret == 0)
+		locked_page = NULL;
+
+	extent_clear_unlock_delalloc(inode, offset, end, locked_page, &cached,
+				     clear_flags,
+				     PAGE_UNLOCK | PAGE_START_WRITEBACK |
+				     PAGE_END_WRITEBACK);
+	return ret;
+}
+
 struct async_extent {
 	u64 start;
 	u64 ram_size;
 	u64 compressed_size;
-	struct page **pages;
-	unsigned long nr_pages;
+	struct folio **folios;
+	unsigned long nr_folios;
 	int compress_type;
 	struct list_head list;
 };
@@ -733,19 +803,20 @@ struct async_cow {
 static noinline int add_async_extent(struct async_chunk *cow,
 				     u64 start, u64 ram_size,
 				     u64 compressed_size,
-				     struct page **pages,
-				     unsigned long nr_pages,
+				     struct folio **folios,
+				     unsigned long nr_folios,
 				     int compress_type)
 {
 	struct async_extent *async_extent;
 
 	async_extent = kmalloc(sizeof(*async_extent), GFP_NOFS);
-	BUG_ON(!async_extent); /* -ENOMEM */
+	if (!async_extent)
+		return -ENOMEM;
 	async_extent->start = start;
 	async_extent->ram_size = ram_size;
 	async_extent->compressed_size = compressed_size;
-	async_extent->pages = pages;
-	async_extent->nr_pages = nr_pages;
+	async_extent->folios = folios;
+	async_extent->nr_folios = nr_folios;
 	async_extent->compress_type = compress_type;
 	list_add_tail(&async_extent->list, &cow->extents);
 	return 0;
@@ -849,8 +920,8 @@ static void compress_file_range(struct btrfs_work *work)
 	u64 actual_end;
 	u64 i_size;
 	int ret = 0;
-	struct page **pages;
-	unsigned long nr_pages;
+	struct folio **folios;
+	unsigned long nr_folios;
 	unsigned long total_compressed = 0;
 	unsigned long total_in = 0;
 	unsigned int poff;
@@ -880,9 +951,9 @@ static void compress_file_range(struct btrfs_work *work)
 	barrier();
 	actual_end = min_t(u64, i_size, end + 1);
 again:
-	pages = NULL;
-	nr_pages = (end >> PAGE_SHIFT) - (start >> PAGE_SHIFT) + 1;
-	nr_pages = min_t(unsigned long, nr_pages, BTRFS_MAX_COMPRESSED_PAGES);
+	folios = NULL;
+	nr_folios = (end >> PAGE_SHIFT) - (start >> PAGE_SHIFT) + 1;
+	nr_folios = min_t(unsigned long, nr_folios, BTRFS_MAX_COMPRESSED_PAGES);
 
 	/*
 	 * we don't want to send crud past the end of i_size through
@@ -931,8 +1002,8 @@ again:
 	if (!inode_need_compress(inode, start, end))
 		goto cleanup_and_bail_uncompressed;
 
-	pages = kcalloc(nr_pages, sizeof(struct page *), GFP_NOFS);
-	if (!pages) {
+	folios = kcalloc(nr_folios, sizeof(struct folio *), GFP_NOFS);
+	if (!folios) {
 		/*
 		 * Memory allocation failure is not a fatal error, we can fall
 		 * back to uncompressed code.
@@ -946,9 +1017,9 @@ again:
 		compress_type = inode->prop_compress;
 
 	/* Compression level is applied here. */
-	ret = btrfs_compress_pages(compress_type | (fs_info->compress_level << 4),
-				   mapping, start, pages, &nr_pages, &total_in,
-				   &total_compressed);
+	ret = btrfs_compress_folios(compress_type | (fs_info->compress_level << 4),
+				    mapping, start, folios, &nr_folios, &total_in,
+				    &total_compressed);
 	if (ret)
 		goto mark_incompressible;
 
@@ -958,7 +1029,7 @@ again:
 	 */
 	poff = offset_in_page(total_compressed);
 	if (poff)
-		memzero_page(pages[nr_pages - 1], poff, PAGE_SIZE - poff);
+		folio_zero_range(folios[nr_folios - 1], poff, PAGE_SIZE - poff);
 
 	/*
 	 * Try to create an inline extent.
@@ -969,43 +1040,16 @@ again:
 	 * Check cow_file_range() for why we don't even try to create inline
 	 * extent for the subpage case.
 	 */
-	if (start == 0 && fs_info->sectorsize == PAGE_SIZE) {
-		if (total_in < actual_end) {
-			ret = cow_file_range_inline(inode, actual_end, 0,
-						    BTRFS_COMPRESS_NONE, NULL,
-						    false);
-		} else {
-			ret = cow_file_range_inline(inode, actual_end,
-						    total_compressed,
-						    compress_type, pages,
-						    false);
-		}
-		if (ret <= 0) {
-			unsigned long clear_flags = EXTENT_DELALLOC |
-				EXTENT_DELALLOC_NEW | EXTENT_DEFRAG |
-				EXTENT_DO_ACCOUNTING;
-
-			if (ret < 0)
-				mapping_set_error(mapping, -EIO);
-
-			/*
-			 * inline extent creation worked or returned error,
-			 * we don't need to create any more async work items.
-			 * Unlock and free up our temp pages.
-			 *
-			 * We use DO_ACCOUNTING here because we need the
-			 * delalloc_release_metadata to be done _after_ we drop
-			 * our outstanding extent for clearing delalloc for this
-			 * range.
-			 */
-			extent_clear_unlock_delalloc(inode, start, end,
-						     NULL,
-						     clear_flags,
-						     PAGE_UNLOCK |
-						     PAGE_START_WRITEBACK |
-						     PAGE_END_WRITEBACK);
-			goto free_pages;
-		}
+	if (total_in < actual_end)
+		ret = cow_file_range_inline(inode, NULL, start, end, 0,
+					    BTRFS_COMPRESS_NONE, NULL, false);
+	else
+		ret = cow_file_range_inline(inode, NULL, start, end, total_compressed,
+					    compress_type, folios[0], false);
+	if (ret <= 0) {
+		if (ret < 0)
+			mapping_set_error(mapping, -EIO);
+		goto free_pages;
 	}
 
 	/*
@@ -1027,8 +1071,9 @@ again:
 	 * The async work queues will take care of doing actual allocation on
 	 * disk for these compressed pages, and will submit the bios.
 	 */
-	add_async_extent(async_chunk, start, total_in, total_compressed, pages,
-			 nr_pages, compress_type);
+	ret = add_async_extent(async_chunk, start, total_in, total_compressed, folios,
+			       nr_folios, compress_type);
+	BUG_ON(ret);
 	if (start + total_in < end) {
 		start += total_in;
 		cond_resched();
@@ -1040,15 +1085,16 @@ mark_incompressible:
 	if (!btrfs_test_opt(fs_info, FORCE_COMPRESS) && !inode->prop_compress)
 		inode->flags |= BTRFS_INODE_NOCOMPRESS;
 cleanup_and_bail_uncompressed:
-	add_async_extent(async_chunk, start, end - start + 1, 0, NULL, 0,
-			 BTRFS_COMPRESS_NONE);
+	ret = add_async_extent(async_chunk, start, end - start + 1, 0, NULL, 0,
+			       BTRFS_COMPRESS_NONE);
+	BUG_ON(ret);
 free_pages:
-	if (pages) {
-		for (i = 0; i < nr_pages; i++) {
-			WARN_ON(pages[i]->mapping);
-			btrfs_free_compr_page(pages[i]);
+	if (folios) {
+		for (i = 0; i < nr_folios; i++) {
+			WARN_ON(folios[i]->mapping);
+			btrfs_free_compr_folio(folios[i]);
 		}
-		kfree(pages);
+		kfree(folios);
 	}
 }
 
@@ -1056,16 +1102,16 @@ static void free_async_extent_pages(struct async_extent *async_extent)
 {
 	int i;
 
-	if (!async_extent->pages)
+	if (!async_extent->folios)
 		return;
 
-	for (i = 0; i < async_extent->nr_pages; i++) {
-		WARN_ON(async_extent->pages[i]->mapping);
-		btrfs_free_compr_page(async_extent->pages[i]);
+	for (i = 0; i < async_extent->nr_folios; i++) {
+		WARN_ON(async_extent->folios[i]->mapping);
+		btrfs_free_compr_folio(async_extent->folios[i]);
 	}
-	kfree(async_extent->pages);
-	async_extent->nr_pages = 0;
-	async_extent->pages = NULL;
+	kfree(async_extent->folios);
+	async_extent->nr_folios = 0;
+	async_extent->folios = NULL;
 }
 
 static void submit_uncompressed_range(struct btrfs_inode *inode,
@@ -1112,6 +1158,7 @@ static void submit_one_async_extent(struct async_chunk *async_chunk,
 	struct btrfs_ordered_extent *ordered;
 	struct btrfs_key ins;
 	struct page *locked_page = NULL;
+	struct extent_state *cached = NULL;
 	struct extent_map *em;
 	int ret = 0;
 	u64 start = async_extent->start;
@@ -1131,7 +1178,6 @@ static void submit_one_async_extent(struct async_chunk *async_chunk,
 		if (!(start >= locked_page_end || end <= locked_page_start))
 			locked_page = async_chunk->locked_page;
 	}
-	lock_extent(io_tree, start, end, NULL);
 
 	if (async_extent->compress_type == BTRFS_COMPRESS_NONE) {
 		submit_uncompressed_range(inode, async_extent, locked_page);
@@ -1153,6 +1199,8 @@ static void submit_one_async_extent(struct async_chunk *async_chunk,
 		goto done;
 	}
 
+	lock_extent(io_tree, start, end, &cached);
+
 	/* Here we're doing allocation and writeback of the compressed pages */
 	em = create_io_em(inode, start,
 			  async_extent->ram_size,	/* len */
@@ -1186,11 +1234,11 @@ static void submit_one_async_extent(struct async_chunk *async_chunk,
 
 	/* Clear dirty, set writeback and unlock the pages. */
 	extent_clear_unlock_delalloc(inode, start, end,
-			NULL, EXTENT_LOCKED | EXTENT_DELALLOC,
+			NULL, &cached, EXTENT_LOCKED | EXTENT_DELALLOC,
 			PAGE_UNLOCK | PAGE_START_WRITEBACK);
 	btrfs_submit_compressed_write(ordered,
-			    async_extent->pages,	/* compressed_pages */
-			    async_extent->nr_pages,
+			    async_extent->folios,	/* compressed_folios */
+			    async_extent->nr_folios,
 			    async_chunk->write_flags, true);
 	*alloc_hint = ins.objectid + ins.offset;
 done:
@@ -1204,7 +1252,8 @@ out_free_reserve:
 	btrfs_free_reserved_extent(fs_info, ins.objectid, ins.offset, 1);
 	mapping_set_error(inode->vfs_inode.i_mapping, -EIO);
 	extent_clear_unlock_delalloc(inode, start, end,
-				     NULL, EXTENT_LOCKED | EXTENT_DELALLOC |
+				     NULL, &cached,
+				     EXTENT_LOCKED | EXTENT_DELALLOC |
 				     EXTENT_DELALLOC_NEW |
 				     EXTENT_DEFRAG | EXTENT_DO_ACCOUNTING,
 				     PAGE_UNLOCK | PAGE_START_WRITEBACK |
@@ -1214,7 +1263,7 @@ out_free_reserve:
 		kthread_associate_blkcg(NULL);
 	btrfs_debug(fs_info,
 "async extent submission failed root=%lld inode=%llu start=%llu len=%llu ret=%d",
-		    root->root_key.objectid, btrfs_ino(inode), start,
+		    btrfs_root_id(root), btrfs_ino(inode), start,
 		    async_extent->ram_size, ret);
 	kfree(async_extent);
 }
@@ -1286,6 +1335,7 @@ static noinline int cow_file_range(struct btrfs_inode *inode,
 {
 	struct btrfs_root *root = inode->root;
 	struct btrfs_fs_info *fs_info = root->fs_info;
+	struct extent_state *cached = NULL;
 	u64 alloc_hint = 0;
 	u64 orig_start = start;
 	u64 num_bytes;
@@ -1311,53 +1361,21 @@ static noinline int cow_file_range(struct btrfs_inode *inode,
 
 	inode_should_defrag(inode, start, end, num_bytes, SZ_64K);
 
-	/*
-	 * Due to the page size limit, for subpage we can only trigger the
-	 * writeback for the dirty sectors of page, that means data writeback
-	 * is doing more writeback than what we want.
-	 *
-	 * This is especially unexpected for some call sites like fallocate,
-	 * where we only increase i_size after everything is done.
-	 * This means we can trigger inline extent even if we didn't want to.
-	 * So here we skip inline extent creation completely.
-	 */
-	if (start == 0 && fs_info->sectorsize == PAGE_SIZE && !no_inline) {
-		u64 actual_end = min_t(u64, i_size_read(&inode->vfs_inode),
-				       end + 1);
-
+	if (!no_inline) {
 		/* lets try to make an inline extent */
-		ret = cow_file_range_inline(inode, actual_end, 0,
+		ret = cow_file_range_inline(inode, locked_page, start, end, 0,
 					    BTRFS_COMPRESS_NONE, NULL, false);
-		if (ret == 0) {
-			/*
-			 * We use DO_ACCOUNTING here because we need the
-			 * delalloc_release_metadata to be run _after_ we drop
-			 * our outstanding extent for clearing delalloc for this
-			 * range.
-			 */
-			extent_clear_unlock_delalloc(inode, start, end,
-				     locked_page,
-				     EXTENT_LOCKED | EXTENT_DELALLOC |
-				     EXTENT_DELALLOC_NEW | EXTENT_DEFRAG |
-				     EXTENT_DO_ACCOUNTING, PAGE_UNLOCK |
-				     PAGE_START_WRITEBACK | PAGE_END_WRITEBACK);
+		if (ret <= 0) {
 			/*
-			 * locked_page is locked by the caller of
-			 * writepage_delalloc(), not locked by
-			 * __process_pages_contig().
-			 *
-			 * We can't let __process_pages_contig() to unlock it,
-			 * as it doesn't have any subpage::writers recorded.
+			 * We succeeded, return 1 so the caller knows we're done
+			 * with this page and already handled the IO.
 			 *
-			 * Here we manually unlock the page, since the caller
-			 * can't determine if it's an inline extent or a
-			 * compressed extent.
+			 * If there was an error then cow_file_range_inline() has
+			 * already done the cleanup.
 			 */
-			unlock_page(locked_page);
-			ret = 1;
+			if (ret == 0)
+				ret = 1;
 			goto done;
-		} else if (ret < 0) {
-			goto out_unlock;
 		}
 	}
 
@@ -1417,6 +1435,10 @@ static noinline int cow_file_range(struct btrfs_inode *inode,
 		extent_reserved = true;
 
 		ram_size = ins.offset;
+
+		lock_extent(&inode->io_tree, start, start + ram_size - 1,
+			    &cached);
+
 		em = create_io_em(inode, start, ins.offset, /* len */
 				  start, /* orig_start */
 				  ins.objectid, /* block_start */
@@ -1426,6 +1448,8 @@ static noinline int cow_file_range(struct btrfs_inode *inode,
 				  BTRFS_COMPRESS_NONE, /* compress_type */
 				  BTRFS_ORDERED_REGULAR /* type */);
 		if (IS_ERR(em)) {
+			unlock_extent(&inode->io_tree, start,
+				      start + ram_size - 1, &cached);
 			ret = PTR_ERR(em);
 			goto out_reserve;
 		}
@@ -1436,6 +1460,8 @@ static noinline int cow_file_range(struct btrfs_inode *inode,
 					0, 1 << BTRFS_ORDERED_REGULAR,
 					BTRFS_COMPRESS_NONE);
 		if (IS_ERR(ordered)) {
+			unlock_extent(&inode->io_tree, start,
+				      start + ram_size - 1, &cached);
 			ret = PTR_ERR(ordered);
 			goto out_drop_extent_cache;
 		}
@@ -1475,7 +1501,7 @@ static noinline int cow_file_range(struct btrfs_inode *inode,
 		page_ops |= PAGE_SET_ORDERED;
 
 		extent_clear_unlock_delalloc(inode, start, start + ram_size - 1,
-					     locked_page,
+					     locked_page, &cached,
 					     EXTENT_LOCKED | EXTENT_DELALLOC,
 					     page_ops);
 		if (num_bytes < cur_alloc_size)
@@ -1534,10 +1560,17 @@ out_unlock:
 		if (!locked_page)
 			mapping_set_error(inode->vfs_inode.i_mapping, ret);
 		extent_clear_unlock_delalloc(inode, orig_start, start - 1,
-					     locked_page, 0, page_ops);
+					     locked_page, NULL, 0, page_ops);
 	}
 
 	/*
+	 * At this point we're unlocked, we want to make sure we're only
+	 * clearing these flags under the extent lock, so lock the rest of the
+	 * range and clear everything up.
+	 */
+	lock_extent(&inode->io_tree, start, end, NULL);
+
+	/*
 	 * For the range (2). If we reserved an extent for our delalloc range
 	 * (or a subrange) and failed to create the respective ordered extent,
 	 * then it means that when we reserved the extent we decremented the
@@ -1550,7 +1583,7 @@ out_unlock:
 	if (extent_reserved) {
 		extent_clear_unlock_delalloc(inode, start,
 					     start + cur_alloc_size - 1,
-					     locked_page,
+					     locked_page, &cached,
 					     clear_bits,
 					     page_ops);
 		start += cur_alloc_size;
@@ -1565,7 +1598,7 @@ out_unlock:
 	if (start < end) {
 		clear_bits |= EXTENT_CLEAR_DATA_RESV;
 		extent_clear_unlock_delalloc(inode, start, end, locked_page,
-					     clear_bits, page_ops);
+					     &cached, clear_bits, page_ops);
 	}
 	return ret;
 }
@@ -1638,7 +1671,6 @@ static bool run_delalloc_compressed(struct btrfs_inode *inode,
 	if (!ctx)
 		return false;
 
-	unlock_extent(&inode->io_tree, start, end, NULL);
 	set_bit(BTRFS_INODE_HAS_ASYNC_EXTENT, &inode->runtime_flags);
 
 	async_chunk = ctx->chunks;
@@ -1732,29 +1764,6 @@ static noinline int run_delalloc_cow(struct btrfs_inode *inode,
 	return 1;
 }
 
-static noinline int csum_exist_in_range(struct btrfs_fs_info *fs_info,
-					u64 bytenr, u64 num_bytes, bool nowait)
-{
-	struct btrfs_root *csum_root = btrfs_csum_root(fs_info, bytenr);
-	struct btrfs_ordered_sum *sums;
-	int ret;
-	LIST_HEAD(list);
-
-	ret = btrfs_lookup_csums_list(csum_root, bytenr, bytenr + num_bytes - 1,
-				      &list, 0, nowait);
-	if (ret == 0 && list_empty(&list))
-		return 0;
-
-	while (!list_empty(&list)) {
-		sums = list_entry(list.next, struct btrfs_ordered_sum, list);
-		list_del(&sums->list);
-		kfree(sums);
-	}
-	if (ret < 0)
-		return ret;
-	return 1;
-}
-
 static int fallback_to_cow(struct btrfs_inode *inode, struct page *locked_page,
 			   const u64 start, const u64 end)
 {
@@ -1762,6 +1771,7 @@ static int fallback_to_cow(struct btrfs_inode *inode, struct page *locked_page,
 	const bool is_reloc_ino = btrfs_is_data_reloc_root(inode->root);
 	const u64 range_bytes = end + 1 - start;
 	struct extent_io_tree *io_tree = &inode->io_tree;
+	struct extent_state *cached_state = NULL;
 	u64 range_start = start;
 	u64 count;
 	int ret;
@@ -1798,6 +1808,7 @@ static int fallback_to_cow(struct btrfs_inode *inode, struct page *locked_page,
 	 * group that contains that extent to RO mode and therefore force COW
 	 * when starting writeback.
 	 */
+	lock_extent(io_tree, start, end, &cached_state);
 	count = count_range_bits(io_tree, &range_start, end, range_bytes,
 				 EXTENT_NORESERVE, 0, NULL);
 	if (count > 0 || is_space_ino || is_reloc_ino) {
@@ -1816,6 +1827,7 @@ static int fallback_to_cow(struct btrfs_inode *inode, struct page *locked_page,
 			clear_extent_bit(io_tree, start, end, EXTENT_NORESERVE,
 					 NULL);
 	}
+	unlock_extent(io_tree, start, end, &cached_state);
 
 	/*
 	 * Don't try to create inline extents, as a mix of inline extent that
@@ -1869,6 +1881,7 @@ static int can_nocow_file_extent(struct btrfs_path *path,
 	struct extent_buffer *leaf = path->nodes[0];
 	struct btrfs_root *root = inode->root;
 	struct btrfs_file_extent_item *fi;
+	struct btrfs_root *csum_root;
 	u64 extent_end;
 	u8 extent_type;
 	int can_nocow = 0;
@@ -1929,7 +1942,7 @@ static int can_nocow_file_extent(struct btrfs_path *path,
 	if (args->free_path) {
 		/*
 		 * We don't need the path anymore, plus through the
-		 * csum_exist_in_range() call below we will end up allocating
+		 * btrfs_lookup_csums_list() call below we will end up allocating
 		 * another path. So free the path to avoid unnecessary extra
 		 * memory usage.
 		 */
@@ -1950,8 +1963,11 @@ static int can_nocow_file_extent(struct btrfs_path *path,
 	 * Force COW if csums exist in the range. This ensures that csums for a
 	 * given extent are either valid or do not exist.
 	 */
-	ret = csum_exist_in_range(root->fs_info, args->disk_bytenr, args->num_bytes,
-				  nowait);
+
+	csum_root = btrfs_csum_root(root->fs_info, args->disk_bytenr);
+	ret = btrfs_lookup_csums_list(csum_root, args->disk_bytenr,
+				      args->disk_bytenr + args->num_bytes - 1,
+				      NULL, nowait);
 	WARN_ON_ONCE(ret > 0 && is_freespace_inode);
 	if (ret != 0)
 		goto out;
@@ -2001,12 +2017,13 @@ static noinline int run_delalloc_nocow(struct btrfs_inode *inode,
 	nocow_args.end = end;
 	nocow_args.writeback_path = true;
 
-	while (1) {
+	while (cur_offset <= end) {
 		struct btrfs_block_group *nocow_bg = NULL;
 		struct btrfs_ordered_extent *ordered;
 		struct btrfs_key found_key;
 		struct btrfs_file_extent_item *fi;
 		struct extent_buffer *leaf;
+		struct extent_state *cached_state = NULL;
 		u64 extent_end;
 		u64 ram_bytes;
 		u64 nocow_end;
@@ -2144,6 +2161,8 @@ must_cow:
 		}
 
 		nocow_end = cur_offset + nocow_args.num_bytes - 1;
+		lock_extent(&inode->io_tree, cur_offset, nocow_end, &cached_state);
+
 		is_prealloc = extent_type == BTRFS_FILE_EXTENT_PREALLOC;
 		if (is_prealloc) {
 			u64 orig_start = found_key.offset - nocow_args.extent_offset;
@@ -2157,6 +2176,8 @@ must_cow:
 					  ram_bytes, BTRFS_COMPRESS_NONE,
 					  BTRFS_ORDERED_PREALLOC);
 			if (IS_ERR(em)) {
+				unlock_extent(&inode->io_tree, cur_offset,
+					      nocow_end, &cached_state);
 				btrfs_dec_nocow_writers(nocow_bg);
 				ret = PTR_ERR(em);
 				goto error;
@@ -2177,6 +2198,8 @@ must_cow:
 				btrfs_drop_extent_map_range(inode, cur_offset,
 							    nocow_end, false);
 			}
+			unlock_extent(&inode->io_tree, cur_offset,
+				      nocow_end, &cached_state);
 			ret = PTR_ERR(ordered);
 			goto error;
 		}
@@ -2191,8 +2214,8 @@ must_cow:
 		btrfs_put_ordered_extent(ordered);
 
 		extent_clear_unlock_delalloc(inode, cur_offset, nocow_end,
-					     locked_page, EXTENT_LOCKED |
-					     EXTENT_DELALLOC |
+					     locked_page, &cached_state,
+					     EXTENT_LOCKED | EXTENT_DELALLOC |
 					     EXTENT_CLEAR_DATA_RESV,
 					     PAGE_UNLOCK | PAGE_SET_ORDERED);
 
@@ -2205,8 +2228,6 @@ must_cow:
 		 */
 		if (ret)
 			goto error;
-		if (cur_offset > end)
-			break;
 	}
 	btrfs_release_path(path);
 
@@ -2232,13 +2253,23 @@ error:
 	 */
 	if (cow_start != (u64)-1)
 		cur_offset = cow_start;
-	if (cur_offset < end)
+
+	/*
+	 * We need to lock the extent here because we're clearing DELALLOC and
+	 * we're not locked at this point.
+	 */
+	if (cur_offset < end) {
+		struct extent_state *cached = NULL;
+
+		lock_extent(&inode->io_tree, cur_offset, end, &cached);
 		extent_clear_unlock_delalloc(inode, cur_offset, end,
-					     locked_page, EXTENT_LOCKED |
-					     EXTENT_DELALLOC | EXTENT_DEFRAG |
+					     locked_page, &cached,
+					     EXTENT_LOCKED | EXTENT_DELALLOC |
+					     EXTENT_DEFRAG |
 					     EXTENT_DO_ACCOUNTING, PAGE_UNLOCK |
 					     PAGE_START_WRITEBACK |
 					     PAGE_END_WRITEBACK);
+	}
 	btrfs_free_path(path);
 	return ret;
 }
@@ -2301,6 +2332,8 @@ void btrfs_split_delalloc_extent(struct btrfs_inode *inode,
 	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 	u64 size;
 
+	lockdep_assert_held(&inode->io_tree.lock);
+
 	/* not delalloc, ignore it */
 	if (!(orig->state & EXTENT_DELALLOC))
 		return;
@@ -2339,6 +2372,8 @@ void btrfs_merge_delalloc_extent(struct btrfs_inode *inode, struct extent_state
 	u64 new_size, old_size;
 	u32 num_extents;
 
+	lockdep_assert_held(&inode->io_tree.lock);
+
 	/* not delalloc, ignore it */
 	if (!(other->state & EXTENT_DELALLOC))
 		return;
@@ -2386,55 +2421,50 @@ void btrfs_merge_delalloc_extent(struct btrfs_inode *inode, struct extent_state
 	spin_unlock(&inode->lock);
 }
 
-static void btrfs_add_delalloc_inodes(struct btrfs_root *root,
-				      struct btrfs_inode *inode)
+static void btrfs_add_delalloc_inode(struct btrfs_inode *inode)
 {
-	struct btrfs_fs_info *fs_info = inode->root->fs_info;
+	struct btrfs_root *root = inode->root;
+	struct btrfs_fs_info *fs_info = root->fs_info;
 
 	spin_lock(&root->delalloc_lock);
-	if (list_empty(&inode->delalloc_inodes)) {
-		list_add_tail(&inode->delalloc_inodes, &root->delalloc_inodes);
-		set_bit(BTRFS_INODE_IN_DELALLOC_LIST, &inode->runtime_flags);
-		root->nr_delalloc_inodes++;
-		if (root->nr_delalloc_inodes == 1) {
-			spin_lock(&fs_info->delalloc_root_lock);
-			BUG_ON(!list_empty(&root->delalloc_root));
-			list_add_tail(&root->delalloc_root,
-				      &fs_info->delalloc_roots);
-			spin_unlock(&fs_info->delalloc_root_lock);
-		}
+	ASSERT(list_empty(&inode->delalloc_inodes));
+	list_add_tail(&inode->delalloc_inodes, &root->delalloc_inodes);
+	root->nr_delalloc_inodes++;
+	if (root->nr_delalloc_inodes == 1) {
+		spin_lock(&fs_info->delalloc_root_lock);
+		ASSERT(list_empty(&root->delalloc_root));
+		list_add_tail(&root->delalloc_root, &fs_info->delalloc_roots);
+		spin_unlock(&fs_info->delalloc_root_lock);
 	}
 	spin_unlock(&root->delalloc_lock);
 }
 
-void __btrfs_del_delalloc_inode(struct btrfs_root *root,
-				struct btrfs_inode *inode)
+void btrfs_del_delalloc_inode(struct btrfs_inode *inode)
 {
+	struct btrfs_root *root = inode->root;
 	struct btrfs_fs_info *fs_info = root->fs_info;
 
+	lockdep_assert_held(&root->delalloc_lock);
+
+	/*
+	 * We may be called after the inode was already deleted from the list,
+	 * namely in the transaction abort path btrfs_destroy_delalloc_inodes(),
+	 * and then later through btrfs_clear_delalloc_extent() while the inode
+	 * still has ->delalloc_bytes > 0.
+	 */
 	if (!list_empty(&inode->delalloc_inodes)) {
 		list_del_init(&inode->delalloc_inodes);
-		clear_bit(BTRFS_INODE_IN_DELALLOC_LIST,
-			  &inode->runtime_flags);
 		root->nr_delalloc_inodes--;
 		if (!root->nr_delalloc_inodes) {
 			ASSERT(list_empty(&root->delalloc_inodes));
 			spin_lock(&fs_info->delalloc_root_lock);
-			BUG_ON(list_empty(&root->delalloc_root));
+			ASSERT(!list_empty(&root->delalloc_root));
 			list_del_init(&root->delalloc_root);
 			spin_unlock(&fs_info->delalloc_root_lock);
 		}
 	}
 }
 
-static void btrfs_del_delalloc_inode(struct btrfs_root *root,
-				     struct btrfs_inode *inode)
-{
-	spin_lock(&root->delalloc_lock);
-	__btrfs_del_delalloc_inode(root, inode);
-	spin_unlock(&root->delalloc_lock);
-}
-
 /*
  * Properly track delayed allocation bytes in the inode and to maintain the
  * list of inodes that have pending delalloc work to be done.
@@ -2444,6 +2474,8 @@ void btrfs_set_delalloc_extent(struct btrfs_inode *inode, struct extent_state *s
 {
 	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 
+	lockdep_assert_held(&inode->io_tree.lock);
+
 	if ((bits & EXTENT_DEFRAG) && !(bits & EXTENT_DELALLOC))
 		WARN_ON(1);
 	/*
@@ -2452,10 +2484,9 @@ void btrfs_set_delalloc_extent(struct btrfs_inode *inode, struct extent_state *s
 	 * bit, which is only set or cleared with irqs on
 	 */
 	if (!(state->state & EXTENT_DELALLOC) && (bits & EXTENT_DELALLOC)) {
-		struct btrfs_root *root = inode->root;
 		u64 len = state->end + 1 - state->start;
+		u64 prev_delalloc_bytes;
 		u32 num_extents = count_max_extents(fs_info, len);
-		bool do_list = !btrfs_is_free_space_inode(inode);
 
 		spin_lock(&inode->lock);
 		btrfs_mod_outstanding_extents(inode, num_extents);
@@ -2468,13 +2499,20 @@ void btrfs_set_delalloc_extent(struct btrfs_inode *inode, struct extent_state *s
 		percpu_counter_add_batch(&fs_info->delalloc_bytes, len,
 					 fs_info->delalloc_batch);
 		spin_lock(&inode->lock);
+		prev_delalloc_bytes = inode->delalloc_bytes;
 		inode->delalloc_bytes += len;
 		if (bits & EXTENT_DEFRAG)
 			inode->defrag_bytes += len;
-		if (do_list && !test_bit(BTRFS_INODE_IN_DELALLOC_LIST,
-					 &inode->runtime_flags))
-			btrfs_add_delalloc_inodes(root, inode);
 		spin_unlock(&inode->lock);
+
+		/*
+		 * We don't need to be under the protection of the inode's lock,
+		 * because we are called while holding the inode's io_tree lock
+		 * and are therefore protected against concurrent calls of this
+		 * function and btrfs_clear_delalloc_extent().
+		 */
+		if (!btrfs_is_free_space_inode(inode) && prev_delalloc_bytes == 0)
+			btrfs_add_delalloc_inode(inode);
 	}
 
 	if (!(state->state & EXTENT_DELALLOC_NEW) &&
@@ -2496,6 +2534,8 @@ void btrfs_clear_delalloc_extent(struct btrfs_inode *inode,
 	u64 len = state->end + 1 - state->start;
 	u32 num_extents = count_max_extents(fs_info, len);
 
+	lockdep_assert_held(&inode->io_tree.lock);
+
 	if ((state->state & EXTENT_DEFRAG) && (bits & EXTENT_DEFRAG)) {
 		spin_lock(&inode->lock);
 		inode->defrag_bytes -= len;
@@ -2509,7 +2549,7 @@ void btrfs_clear_delalloc_extent(struct btrfs_inode *inode,
 	 */
 	if ((state->state & EXTENT_DELALLOC) && (bits & EXTENT_DELALLOC)) {
 		struct btrfs_root *root = inode->root;
-		bool do_list = !btrfs_is_free_space_inode(inode);
+		u64 new_delalloc_bytes;
 
 		spin_lock(&inode->lock);
 		btrfs_mod_outstanding_extents(inode, -num_extents);
@@ -2529,7 +2569,8 @@ void btrfs_clear_delalloc_extent(struct btrfs_inode *inode,
 			return;
 
 		if (!btrfs_is_data_reloc_root(root) &&
-		    do_list && !(state->state & EXTENT_NORESERVE) &&
+		    !btrfs_is_free_space_inode(inode) &&
+		    !(state->state & EXTENT_NORESERVE) &&
 		    (bits & EXTENT_CLEAR_DATA_RESV))
 			btrfs_free_reserved_data_space_noquota(fs_info, len);
 
@@ -2537,11 +2578,20 @@ void btrfs_clear_delalloc_extent(struct btrfs_inode *inode,
 					 fs_info->delalloc_batch);
 		spin_lock(&inode->lock);
 		inode->delalloc_bytes -= len;
-		if (do_list && inode->delalloc_bytes == 0 &&
-		    test_bit(BTRFS_INODE_IN_DELALLOC_LIST,
-					&inode->runtime_flags))
-			btrfs_del_delalloc_inode(root, inode);
+		new_delalloc_bytes = inode->delalloc_bytes;
 		spin_unlock(&inode->lock);
+
+		/*
+		 * We don't need to be under the protection of the inode's lock,
+		 * because we are called while holding the inode's io_tree lock
+		 * and are therefore protected against concurrent calls of this
+		 * function and btrfs_set_delalloc_extent().
+		 */
+		if (!btrfs_is_free_space_inode(inode) && new_delalloc_bytes == 0) {
+			spin_lock(&root->delalloc_lock);
+			btrfs_del_delalloc_inode(inode);
+			spin_unlock(&root->delalloc_lock);
+		}
 	}
 
 	if ((state->state & EXTENT_DELALLOC_NEW) &&
@@ -2631,7 +2681,7 @@ static int btrfs_find_new_delalloc_bytes(struct btrfs_inode *inode,
 		u64 em_len;
 		int ret = 0;
 
-		em = btrfs_get_extent(inode, NULL, 0, search_start, search_len);
+		em = btrfs_get_extent(inode, NULL, search_start, search_len);
 		if (IS_ERR(em))
 			return PTR_ERR(em);
 
@@ -3161,7 +3211,6 @@ int btrfs_finish_one_ordered(struct btrfs_ordered_extent *ordered_extent)
 		btrfs_abort_transaction(trans, ret);
 		goto out;
 	}
-	ret = 0;
 out:
 	clear_extent_bit(&inode->io_tree, start, end, clear_bits,
 			 &cached_state);
@@ -3180,9 +3229,8 @@ out:
 		 * set the mapping error, so we need to set it if we're the ones
 		 * marking this ordered extent as failed.
 		 */
-		if (ret && !test_and_set_bit(BTRFS_ORDERED_IOERR,
-					     &ordered_extent->flags))
-			mapping_set_error(ordered_extent->inode->i_mapping, -EIO);
+		if (ret)
+			btrfs_mark_ordered_extent_error(ordered_extent);
 
 		if (truncated)
 			unwritten_start += logical_len;
@@ -3236,7 +3284,7 @@ out:
 			 * Actually free the qgroup rsv which was released when
 			 * the ordered extent was created.
 			 */
-			btrfs_qgroup_free_refroot(fs_info, inode->root->root_key.objectid,
+			btrfs_qgroup_free_refroot(fs_info, btrfs_root_id(inode->root),
 						  ordered_extent->qgroup_rsv,
 						  BTRFS_QGROUP_RSV_DATA);
 		}
@@ -3903,7 +3951,7 @@ cache_acl:
 			btrfs_err(fs_info,
 				  "error loading props for ino %llu (root %llu): %d",
 				  btrfs_ino(BTRFS_I(inode)),
-				  root->root_key.objectid, ret);
+				  btrfs_root_id(root), ret);
 	}
 	if (path != in_path)
 		btrfs_free_path(path);
@@ -4262,7 +4310,7 @@ static int btrfs_unlink_subvol(struct btrfs_trans_handle *trans,
 	/* This needs to handle no-key deletions later on */
 
 	if (btrfs_ino(inode) == BTRFS_FIRST_FREE_OBJECTID) {
-		objectid = inode->root->root_key.objectid;
+		objectid = btrfs_root_id(inode->root);
 	} else if (btrfs_ino(inode) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID) {
 		objectid = inode->location.objectid;
 	} else {
@@ -4320,7 +4368,7 @@ static int btrfs_unlink_subvol(struct btrfs_trans_handle *trans,
 		btrfs_release_path(path);
 	} else {
 		ret = btrfs_del_root_ref(trans, objectid,
-					 root->root_key.objectid, dir_ino,
+					 btrfs_root_id(root), dir_ino,
 					 &index, &fname.disk_name);
 		if (ret) {
 			btrfs_abort_transaction(trans, ret);
@@ -4370,7 +4418,7 @@ static noinline int may_destroy_subvol(struct btrfs_root *root)
 				   dir_id, &name, 0);
 	if (di && !IS_ERR(di)) {
 		btrfs_dir_item_key_to_cpu(path->nodes[0], di, &key);
-		if (key.objectid == root->root_key.objectid) {
+		if (key.objectid == btrfs_root_id(root)) {
 			ret = -EPERM;
 			btrfs_err(fs_info,
 				  "deleting default subvolume %llu is not allowed",
@@ -4380,21 +4428,27 @@ static noinline int may_destroy_subvol(struct btrfs_root *root)
 		btrfs_release_path(path);
 	}
 
-	key.objectid = root->root_key.objectid;
+	key.objectid = btrfs_root_id(root);
 	key.type = BTRFS_ROOT_REF_KEY;
 	key.offset = (u64)-1;
 
 	ret = btrfs_search_slot(NULL, fs_info->tree_root, &key, path, 0, 0);
 	if (ret < 0)
 		goto out;
-	BUG_ON(ret == 0);
+	if (ret == 0) {
+		/*
+		 * Key with offset -1 found, there would have to exist a root
+		 * with such id, but this is out of valid range.
+		 */
+		ret = -EUCLEAN;
+		goto out;
+	}
 
 	ret = 0;
 	if (path->slots[0] > 0) {
 		path->slots[0]--;
 		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
-		if (key.objectid == root->root_key.objectid &&
-		    key.type == BTRFS_ROOT_REF_KEY)
+		if (key.objectid == btrfs_root_id(root) && key.type == BTRFS_ROOT_REF_KEY)
 			ret = -ENOTEMPTY;
 	}
 out:
@@ -4406,64 +4460,26 @@ out:
 static void btrfs_prune_dentries(struct btrfs_root *root)
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
-	struct rb_node *node;
-	struct rb_node *prev;
-	struct btrfs_inode *entry;
-	struct inode *inode;
-	u64 objectid = 0;
+	struct btrfs_inode *inode;
+	u64 min_ino = 0;
 
 	if (!BTRFS_FS_ERROR(fs_info))
 		WARN_ON(btrfs_root_refs(&root->root_item) != 0);
 
-	spin_lock(&root->inode_lock);
-again:
-	node = root->inode_tree.rb_node;
-	prev = NULL;
-	while (node) {
-		prev = node;
-		entry = rb_entry(node, struct btrfs_inode, rb_node);
-
-		if (objectid < btrfs_ino(entry))
-			node = node->rb_left;
-		else if (objectid > btrfs_ino(entry))
-			node = node->rb_right;
-		else
-			break;
-	}
-	if (!node) {
-		while (prev) {
-			entry = rb_entry(prev, struct btrfs_inode, rb_node);
-			if (objectid <= btrfs_ino(entry)) {
-				node = prev;
-				break;
-			}
-			prev = rb_next(prev);
-		}
-	}
-	while (node) {
-		entry = rb_entry(node, struct btrfs_inode, rb_node);
-		objectid = btrfs_ino(entry) + 1;
-		inode = igrab(&entry->vfs_inode);
-		if (inode) {
-			spin_unlock(&root->inode_lock);
-			if (atomic_read(&inode->i_count) > 1)
-				d_prune_aliases(inode);
-			/*
-			 * btrfs_drop_inode will have it removed from the inode
-			 * cache when its usage count hits zero.
-			 */
-			iput(inode);
-			cond_resched();
-			spin_lock(&root->inode_lock);
-			goto again;
-		}
-
-		if (cond_resched_lock(&root->inode_lock))
-			goto again;
+	inode = btrfs_find_first_inode(root, min_ino);
+	while (inode) {
+		if (atomic_read(&inode->vfs_inode.i_count) > 1)
+			d_prune_aliases(&inode->vfs_inode);
 
-		node = rb_next(node);
+		min_ino = btrfs_ino(inode) + 1;
+		/*
+		 * btrfs_drop_inode() will have it removed from the inode
+		 * cache when its usage count hits zero.
+		 */
+		iput(&inode->vfs_inode);
+		cond_resched();
+		inode = btrfs_find_first_inode(root, min_ino);
 	}
-	spin_unlock(&root->inode_lock);
 }
 
 int btrfs_delete_subvolume(struct btrfs_inode *dir, struct dentry *dentry)
@@ -4490,7 +4506,7 @@ int btrfs_delete_subvolume(struct btrfs_inode *dir, struct dentry *dentry)
 		spin_unlock(&dest->root_item_lock);
 		btrfs_warn(fs_info,
 			   "attempt to delete subvolume %llu during send",
-			   dest->root_key.objectid);
+			   btrfs_root_id(dest));
 		ret = -EPERM;
 		goto out_up_write;
 	}
@@ -4498,7 +4514,7 @@ int btrfs_delete_subvolume(struct btrfs_inode *dir, struct dentry *dentry)
 		spin_unlock(&dest->root_item_lock);
 		btrfs_warn(fs_info,
 			   "attempt to delete subvolume %llu with active swapfile",
-			   root->root_key.objectid);
+			   btrfs_root_id(root));
 		ret = -EPERM;
 		goto out_up_write;
 	}
@@ -4559,7 +4575,7 @@ int btrfs_delete_subvolume(struct btrfs_inode *dir, struct dentry *dentry)
 	if (!test_and_set_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, &dest->state)) {
 		ret = btrfs_insert_orphan_item(trans,
 					fs_info->tree_root,
-					dest->root_key.objectid);
+					btrfs_root_id(dest));
 		if (ret) {
 			btrfs_abort_transaction(trans, ret);
 			goto out_end_trans;
@@ -4567,8 +4583,7 @@ int btrfs_delete_subvolume(struct btrfs_inode *dir, struct dentry *dentry)
 	}
 
 	ret = btrfs_uuid_tree_remove(trans, dest->root_item.uuid,
-				  BTRFS_UUID_KEY_SUBVOL,
-				  dest->root_key.objectid);
+				     BTRFS_UUID_KEY_SUBVOL, btrfs_root_id(dest));
 	if (ret && ret != -ENOENT) {
 		btrfs_abort_transaction(trans, ret);
 		goto out_end_trans;
@@ -4577,7 +4592,7 @@ int btrfs_delete_subvolume(struct btrfs_inode *dir, struct dentry *dentry)
 		ret = btrfs_uuid_tree_remove(trans,
 					  dest->root_item.received_uuid,
 					  BTRFS_UUID_KEY_RECEIVED_SUBVOL,
-					  dest->root_key.objectid);
+					  btrfs_root_id(dest));
 		if (ret && ret != -ENOENT) {
 			btrfs_abort_transaction(trans, ret);
 			goto out_end_trans;
@@ -4618,7 +4633,7 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
 {
 	struct inode *inode = d_inode(dentry);
 	struct btrfs_fs_info *fs_info = BTRFS_I(inode)->root->fs_info;
-	int err = 0;
+	int ret = 0;
 	struct btrfs_trans_handle *trans;
 	u64 last_unlink_trans;
 	struct fscrypt_name fname;
@@ -4634,33 +4649,33 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
 		return btrfs_delete_subvolume(BTRFS_I(dir), dentry);
 	}
 
-	err = fscrypt_setup_filename(dir, &dentry->d_name, 1, &fname);
-	if (err)
-		return err;
+	ret = fscrypt_setup_filename(dir, &dentry->d_name, 1, &fname);
+	if (ret)
+		return ret;
 
 	/* This needs to handle no-key deletions later on */
 
 	trans = __unlink_start_trans(BTRFS_I(dir));
 	if (IS_ERR(trans)) {
-		err = PTR_ERR(trans);
+		ret = PTR_ERR(trans);
 		goto out_notrans;
 	}
 
 	if (unlikely(btrfs_ino(BTRFS_I(inode)) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) {
-		err = btrfs_unlink_subvol(trans, BTRFS_I(dir), dentry);
+		ret = btrfs_unlink_subvol(trans, BTRFS_I(dir), dentry);
 		goto out;
 	}
 
-	err = btrfs_orphan_add(trans, BTRFS_I(inode));
-	if (err)
+	ret = btrfs_orphan_add(trans, BTRFS_I(inode));
+	if (ret)
 		goto out;
 
 	last_unlink_trans = BTRFS_I(inode)->last_unlink_trans;
 
 	/* now the directory is empty */
-	err = btrfs_unlink_inode(trans, BTRFS_I(dir), BTRFS_I(d_inode(dentry)),
+	ret = btrfs_unlink_inode(trans, BTRFS_I(dir), BTRFS_I(d_inode(dentry)),
 				 &fname.disk_name);
-	if (!err) {
+	if (!ret) {
 		btrfs_i_size_write(BTRFS_I(inode), 0);
 		/*
 		 * Propagate the last_unlink_trans value of the deleted dir to
@@ -4682,7 +4697,7 @@ out_notrans:
 	btrfs_btree_balance_dirty(fs_info);
 	fscrypt_free_filename(&fname);
 
-	return err;
+	return ret;
 }
 
 /*
@@ -4710,7 +4725,7 @@ int btrfs_truncate_block(struct btrfs_inode *inode, loff_t from, loff_t len,
 	u32 blocksize = fs_info->sectorsize;
 	pgoff_t index = from >> PAGE_SHIFT;
 	unsigned offset = from & (blocksize - 1);
-	struct page *page;
+	struct folio *folio;
 	gfp_t mask = btrfs_alloc_write_mask(mapping);
 	size_t write_bytes = blocksize;
 	int ret = 0;
@@ -4742,8 +4757,9 @@ int btrfs_truncate_block(struct btrfs_inode *inode, loff_t from, loff_t len,
 		goto out;
 	}
 again:
-	page = find_or_create_page(mapping, index, mask);
-	if (!page) {
+	folio = __filemap_get_folio(mapping, index,
+				    FGP_LOCK | FGP_ACCESSED | FGP_CREAT, mask);
+	if (IS_ERR(folio)) {
 		btrfs_delalloc_release_space(inode, data_reserved, block_start,
 					     blocksize, true);
 		btrfs_delalloc_release_extents(inode, blocksize);
@@ -4751,15 +4767,15 @@ again:
 		goto out;
 	}
 
-	if (!PageUptodate(page)) {
-		ret = btrfs_read_folio(NULL, page_folio(page));
-		lock_page(page);
-		if (page->mapping != mapping) {
-			unlock_page(page);
-			put_page(page);
+	if (!folio_test_uptodate(folio)) {
+		ret = btrfs_read_folio(NULL, folio);
+		folio_lock(folio);
+		if (folio->mapping != mapping) {
+			folio_unlock(folio);
+			folio_put(folio);
 			goto again;
 		}
-		if (!PageUptodate(page)) {
+		if (!folio_test_uptodate(folio)) {
 			ret = -EIO;
 			goto out_unlock;
 		}
@@ -4771,19 +4787,19 @@ again:
 	 * folio private, but left the page in the mapping.  Set the page mapped
 	 * here to make sure it's properly set for the subpage stuff.
 	 */
-	ret = set_page_extent_mapped(page);
+	ret = set_folio_extent_mapped(folio);
 	if (ret < 0)
 		goto out_unlock;
 
-	wait_on_page_writeback(page);
+	folio_wait_writeback(folio);
 
 	lock_extent(io_tree, block_start, block_end, &cached_state);
 
 	ordered = btrfs_lookup_ordered_extent(inode, block_start);
 	if (ordered) {
 		unlock_extent(io_tree, block_start, block_end, &cached_state);
-		unlock_page(page);
-		put_page(page);
+		folio_unlock(folio);
+		folio_put(folio);
 		btrfs_start_ordered_extent(ordered);
 		btrfs_put_ordered_extent(ordered);
 		goto again;
@@ -4804,15 +4820,16 @@ again:
 		if (!len)
 			len = blocksize - offset;
 		if (front)
-			memzero_page(page, (block_start - page_offset(page)),
-				     offset);
+			folio_zero_range(folio, block_start - folio_pos(folio),
+					 offset);
 		else
-			memzero_page(page, (block_start - page_offset(page)) + offset,
-				     len);
+			folio_zero_range(folio,
+					 (block_start - folio_pos(folio)) + offset,
+					 len);
 	}
-	btrfs_folio_clear_checked(fs_info, page_folio(page), block_start,
+	btrfs_folio_clear_checked(fs_info, folio, block_start,
 				  block_end + 1 - block_start);
-	btrfs_folio_set_dirty(fs_info, page_folio(page), block_start,
+	btrfs_folio_set_dirty(fs_info, folio, block_start,
 			      block_end + 1 - block_start);
 	unlock_extent(io_tree, block_start, block_end, &cached_state);
 
@@ -4829,8 +4846,8 @@ out_unlock:
 					block_start, blocksize, true);
 	}
 	btrfs_delalloc_release_extents(inode, blocksize);
-	unlock_page(page);
-	put_page(page);
+	folio_unlock(folio);
+	folio_put(folio);
 out:
 	if (only_release_metadata)
 		btrfs_check_nocow_unlock(inode);
@@ -4904,16 +4921,16 @@ int btrfs_cont_expand(struct btrfs_inode *inode, loff_t oldsize, loff_t size)
 	u64 last_byte;
 	u64 cur_offset;
 	u64 hole_size;
-	int err = 0;
+	int ret = 0;
 
 	/*
 	 * If our size started in the middle of a block we need to zero out the
 	 * rest of the block before we expand the i_size, otherwise we could
 	 * expose stale data.
 	 */
-	err = btrfs_truncate_block(inode, oldsize, 0, 0);
-	if (err)
-		return err;
+	ret = btrfs_truncate_block(inode, oldsize, 0, 0);
+	if (ret)
+		return ret;
 
 	if (size <= hole_start)
 		return 0;
@@ -4922,10 +4939,9 @@ int btrfs_cont_expand(struct btrfs_inode *inode, loff_t oldsize, loff_t size)
 					   &cached_state);
 	cur_offset = hole_start;
 	while (1) {
-		em = btrfs_get_extent(inode, NULL, 0, cur_offset,
-				      block_end - cur_offset);
+		em = btrfs_get_extent(inode, NULL, cur_offset, block_end - cur_offset);
 		if (IS_ERR(em)) {
-			err = PTR_ERR(em);
+			ret = PTR_ERR(em);
 			em = NULL;
 			break;
 		}
@@ -4936,13 +4952,13 @@ int btrfs_cont_expand(struct btrfs_inode *inode, loff_t oldsize, loff_t size)
 		if (!(em->flags & EXTENT_FLAG_PREALLOC)) {
 			struct extent_map *hole_em;
 
-			err = maybe_insert_hole(inode, cur_offset, hole_size);
-			if (err)
+			ret = maybe_insert_hole(inode, cur_offset, hole_size);
+			if (ret)
 				break;
 
-			err = btrfs_inode_set_file_extent_range(inode,
+			ret = btrfs_inode_set_file_extent_range(inode,
 							cur_offset, hole_size);
-			if (err)
+			if (ret)
 				break;
 
 			hole_em = alloc_extent_map();
@@ -4963,12 +4979,12 @@ int btrfs_cont_expand(struct btrfs_inode *inode, loff_t oldsize, loff_t size)
 			hole_em->ram_bytes = hole_size;
 			hole_em->generation = btrfs_get_fs_generation(fs_info);
 
-			err = btrfs_replace_extent_map_range(inode, hole_em, true);
+			ret = btrfs_replace_extent_map_range(inode, hole_em, true);
 			free_extent_map(hole_em);
 		} else {
-			err = btrfs_inode_set_file_extent_range(inode,
+			ret = btrfs_inode_set_file_extent_range(inode,
 							cur_offset, hole_size);
-			if (err)
+			if (ret)
 				break;
 		}
 next:
@@ -4980,7 +4996,7 @@ next:
 	}
 	free_extent_map(em);
 	unlock_extent(io_tree, hole_start, block_end - 1, &cached_state);
-	return err;
+	return ret;
 }
 
 static int btrfs_setsize(struct inode *inode, struct iattr *attr)
@@ -5256,7 +5272,7 @@ void btrfs_evict_inode(struct inode *inode)
 
 	if (inode->i_nlink &&
 	    ((btrfs_root_refs(&root->root_item) != 0 &&
-	      root->root_key.objectid != BTRFS_ROOT_TREE_OBJECTID) ||
+	      btrfs_root_id(root) != BTRFS_ROOT_TREE_OBJECTID) ||
 	     btrfs_is_free_space_inode(BTRFS_I(inode))))
 		goto out;
 
@@ -5268,7 +5284,7 @@ void btrfs_evict_inode(struct inode *inode)
 
 	if (inode->i_nlink > 0) {
 		BUG_ON(btrfs_root_refs(&root->root_item) != 0 &&
-		       root->root_key.objectid != BTRFS_ROOT_TREE_OBJECTID);
+		       btrfs_root_id(root) != BTRFS_ROOT_TREE_OBJECTID);
 		goto out;
 	}
 
@@ -5440,7 +5456,7 @@ static int fixup_tree_root_location(struct btrfs_fs_info *fs_info,
 	}
 
 	err = -ENOENT;
-	key.objectid = dir->root->root_key.objectid;
+	key.objectid = btrfs_root_id(dir->root);
 	key.type = BTRFS_ROOT_REF_KEY;
 	key.offset = location->objectid;
 
@@ -5549,7 +5565,6 @@ static int btrfs_init_locked_inode(struct inode *inode, void *p)
 	BTRFS_I(inode)->location.type = BTRFS_INODE_ITEM_KEY;
 	BTRFS_I(inode)->location.offset = 0;
 	BTRFS_I(inode)->root = btrfs_grab_root(args->root);
-	BUG_ON(args->root && !BTRFS_I(inode)->root);
 
 	if (args->root && args->root == args->root->fs_info->tree_root &&
 	    args->ino != BTRFS_BTREE_INODE_OBJECTID)
@@ -6400,8 +6415,7 @@ int btrfs_create_new_inode(struct btrfs_trans_handle *trans,
 	if (ret) {
 		btrfs_err(fs_info,
 			  "error inheriting props for ino %llu (root %llu): %d",
-			  btrfs_ino(BTRFS_I(inode)), root->root_key.objectid,
-			  ret);
+			  btrfs_ino(BTRFS_I(inode)), btrfs_root_id(root), ret);
 	}
 
 	/*
@@ -6474,7 +6488,7 @@ int btrfs_add_link(struct btrfs_trans_handle *trans,
 
 	if (unlikely(ino == BTRFS_FIRST_FREE_OBJECTID)) {
 		ret = btrfs_add_root_ref(trans, key.objectid,
-					 root->root_key.objectid, parent_ino,
+					 btrfs_root_id(root), parent_ino,
 					 index, name);
 	} else if (add_backref) {
 		ret = btrfs_insert_inode_ref(trans, root, name,
@@ -6517,7 +6531,7 @@ fail_dir_item:
 		u64 local_index;
 		int err;
 		err = btrfs_del_root_ref(trans, key.objectid,
-					 root->root_key.objectid, parent_ino,
+					 btrfs_root_id(root), parent_ino,
 					 &local_index, name);
 		if (err)
 			btrfs_abort_transaction(trans, err);
@@ -6615,7 +6629,7 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir,
 	int drop_inode = 0;
 
 	/* do not allow sys_link's with other subvols of the same device */
-	if (root->root_key.objectid != BTRFS_I(inode)->root->root_key.objectid)
+	if (btrfs_root_id(root) != btrfs_root_id(BTRFS_I(inode)->root))
 		return -EXDEV;
 
 	if (inode->i_nlink >= BTRFS_LINK_MAX)
@@ -6772,7 +6786,6 @@ static int read_inline_extent(struct btrfs_inode *inode, struct btrfs_path *path
  *
  * @inode:	file to search in
  * @page:	page to read extent data into if the extent is inline
- * @pg_offset:	offset into @page to copy to
  * @start:	file offset
  * @len:	length of range starting at @start
  *
@@ -6786,8 +6799,7 @@ static int read_inline_extent(struct btrfs_inode *inode, struct btrfs_path *path
  * Return: ERR_PTR on error, non-NULL extent_map on success.
  */
 struct extent_map *btrfs_get_extent(struct btrfs_inode *inode,
-				    struct page *page, size_t pg_offset,
-				    u64 start, u64 len)
+				    struct page *page, u64 start, u64 len)
 {
 	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 	int ret = 0;
@@ -6930,7 +6942,6 @@ next:
 		 * ensured by tree-checker and inline extent creation path.
 		 * Thus all members representing file offsets should be zero.
 		 */
-		ASSERT(pg_offset == 0);
 		ASSERT(extent_start == 0);
 		ASSERT(em->start == 0);
 
@@ -6965,7 +6976,7 @@ insert:
 	}
 
 	write_lock(&em_tree->lock);
-	ret = btrfs_add_extent_mapping(fs_info, em_tree, &em, start, len);
+	ret = btrfs_add_extent_mapping(inode, &em, start, len);
 	write_unlock(&em_tree->lock);
 out:
 	btrfs_free_path(path);
@@ -7292,11 +7303,49 @@ static struct extent_map *create_io_em(struct btrfs_inode *inode, u64 start,
 	struct extent_map *em;
 	int ret;
 
+	/*
+	 * Note the missing NOCOW type.
+	 *
+	 * For pure NOCOW writes, we should not create an io extent map, but
+	 * just reusing the existing one.
+	 * Only PREALLOC writes (NOCOW write into preallocated range) can
+	 * create an io extent map.
+	 */
 	ASSERT(type == BTRFS_ORDERED_PREALLOC ||
 	       type == BTRFS_ORDERED_COMPRESSED ||
-	       type == BTRFS_ORDERED_NOCOW ||
 	       type == BTRFS_ORDERED_REGULAR);
 
+	switch (type) {
+	case BTRFS_ORDERED_PREALLOC:
+		/* Uncompressed extents. */
+		ASSERT(block_len == len);
+
+		/* We're only referring part of a larger preallocated extent. */
+		ASSERT(block_len <= ram_bytes);
+		break;
+	case BTRFS_ORDERED_REGULAR:
+		/* Uncompressed extents. */
+		ASSERT(block_len == len);
+
+		/* COW results a new extent matching our file extent size. */
+		ASSERT(orig_block_len == len);
+		ASSERT(ram_bytes == len);
+
+		/* Since it's a new extent, we should not have any offset. */
+		ASSERT(orig_start == start);
+		break;
+	case BTRFS_ORDERED_COMPRESSED:
+		/* Must be compressed. */
+		ASSERT(compress_type != BTRFS_COMPRESS_NONE);
+
+		/*
+		 * Encoded write can make us to refer to part of the
+		 * uncompressed extent.
+		 */
+		ASSERT(len <= ram_bytes);
+		break;
+	}
+
 	em = alloc_extent_map();
 	if (!em)
 		return ERR_PTR(-ENOMEM);
@@ -7310,9 +7359,7 @@ static struct extent_map *create_io_em(struct btrfs_inode *inode, u64 start,
 	em->ram_bytes = ram_bytes;
 	em->generation = -1;
 	em->flags |= EXTENT_FLAG_PINNED;
-	if (type == BTRFS_ORDERED_PREALLOC)
-		em->flags |= EXTENT_FLAG_FILLING;
-	else if (type == BTRFS_ORDERED_COMPRESSED)
+	if (type == BTRFS_ORDERED_COMPRESSED)
 		extent_map_set_compression(em, compress_type);
 
 	ret = btrfs_replace_extent_map_range(inode, em, true);
@@ -7571,7 +7618,7 @@ static int btrfs_dio_iomap_begin(struct inode *inode, loff_t start,
 	if (ret < 0)
 		goto err;
 
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, start, len);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, start, len);
 	if (IS_ERR(em)) {
 		ret = PTR_ERR(em);
 		goto unlock_err;
@@ -7899,17 +7946,6 @@ static int btrfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 	return ret;
 }
 
-static int btrfs_writepages(struct address_space *mapping,
-			    struct writeback_control *wbc)
-{
-	return extent_writepages(mapping, wbc);
-}
-
-static void btrfs_readahead(struct readahead_control *rac)
-{
-	extent_readahead(rac);
-}
-
 /*
  * For release_folio() and invalidate_folio() we have a race window where
  * folio_end_writeback() is called but the subpage spinlock is not yet released.
@@ -7946,13 +7982,12 @@ static void wait_subpage_spinlock(struct page *page)
 
 static bool __btrfs_release_folio(struct folio *folio, gfp_t gfp_flags)
 {
-	int ret = try_release_extent_mapping(&folio->page, gfp_flags);
-
-	if (ret == 1) {
+	if (try_release_extent_mapping(&folio->page, gfp_flags)) {
 		wait_subpage_spinlock(&folio->page);
 		clear_page_extent_mapped(&folio->page);
+		return true;
 	}
-	return ret;
+	return false;
 }
 
 static bool btrfs_release_folio(struct folio *folio, gfp_t gfp_flags)
@@ -8150,173 +8185,6 @@ next:
 	clear_page_extent_mapped(&folio->page);
 }
 
-/*
- * btrfs_page_mkwrite() is not allowed to change the file size as it gets
- * called from a page fault handler when a page is first dirtied. Hence we must
- * be careful to check for EOF conditions here. We set the page up correctly
- * for a written page which means we get ENOSPC checking when writing into
- * holes and correct delalloc and unwritten extent mapping on filesystems that
- * support these features.
- *
- * We are not allowed to take the i_mutex here so we have to play games to
- * protect against truncate races as the page could now be beyond EOF.  Because
- * truncate_setsize() writes the inode size before removing pages, once we have
- * the page lock we can determine safely if the page is beyond EOF. If it is not
- * beyond EOF, then the page is guaranteed safe against truncation until we
- * unlock the page.
- */
-vm_fault_t btrfs_page_mkwrite(struct vm_fault *vmf)
-{
-	struct page *page = vmf->page;
-	struct folio *folio = page_folio(page);
-	struct inode *inode = file_inode(vmf->vma->vm_file);
-	struct btrfs_fs_info *fs_info = inode_to_fs_info(inode);
-	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
-	struct btrfs_ordered_extent *ordered;
-	struct extent_state *cached_state = NULL;
-	struct extent_changeset *data_reserved = NULL;
-	unsigned long zero_start;
-	loff_t size;
-	vm_fault_t ret;
-	int ret2;
-	int reserved = 0;
-	u64 reserved_space;
-	u64 page_start;
-	u64 page_end;
-	u64 end;
-
-	ASSERT(folio_order(folio) == 0);
-
-	reserved_space = PAGE_SIZE;
-
-	sb_start_pagefault(inode->i_sb);
-	page_start = page_offset(page);
-	page_end = page_start + PAGE_SIZE - 1;
-	end = page_end;
-
-	/*
-	 * Reserving delalloc space after obtaining the page lock can lead to
-	 * deadlock. For example, if a dirty page is locked by this function
-	 * and the call to btrfs_delalloc_reserve_space() ends up triggering
-	 * dirty page write out, then the btrfs_writepages() function could
-	 * end up waiting indefinitely to get a lock on the page currently
-	 * being processed by btrfs_page_mkwrite() function.
-	 */
-	ret2 = btrfs_delalloc_reserve_space(BTRFS_I(inode), &data_reserved,
-					    page_start, reserved_space);
-	if (!ret2) {
-		ret2 = file_update_time(vmf->vma->vm_file);
-		reserved = 1;
-	}
-	if (ret2) {
-		ret = vmf_error(ret2);
-		if (reserved)
-			goto out;
-		goto out_noreserve;
-	}
-
-	ret = VM_FAULT_NOPAGE; /* make the VM retry the fault */
-again:
-	down_read(&BTRFS_I(inode)->i_mmap_lock);
-	lock_page(page);
-	size = i_size_read(inode);
-
-	if ((page->mapping != inode->i_mapping) ||
-	    (page_start >= size)) {
-		/* page got truncated out from underneath us */
-		goto out_unlock;
-	}
-	wait_on_page_writeback(page);
-
-	lock_extent(io_tree, page_start, page_end, &cached_state);
-	ret2 = set_page_extent_mapped(page);
-	if (ret2 < 0) {
-		ret = vmf_error(ret2);
-		unlock_extent(io_tree, page_start, page_end, &cached_state);
-		goto out_unlock;
-	}
-
-	/*
-	 * we can't set the delalloc bits if there are pending ordered
-	 * extents.  Drop our locks and wait for them to finish
-	 */
-	ordered = btrfs_lookup_ordered_range(BTRFS_I(inode), page_start,
-			PAGE_SIZE);
-	if (ordered) {
-		unlock_extent(io_tree, page_start, page_end, &cached_state);
-		unlock_page(page);
-		up_read(&BTRFS_I(inode)->i_mmap_lock);
-		btrfs_start_ordered_extent(ordered);
-		btrfs_put_ordered_extent(ordered);
-		goto again;
-	}
-
-	if (page->index == ((size - 1) >> PAGE_SHIFT)) {
-		reserved_space = round_up(size - page_start,
-					  fs_info->sectorsize);
-		if (reserved_space < PAGE_SIZE) {
-			end = page_start + reserved_space - 1;
-			btrfs_delalloc_release_space(BTRFS_I(inode),
-					data_reserved, page_start,
-					PAGE_SIZE - reserved_space, true);
-		}
-	}
-
-	/*
-	 * page_mkwrite gets called when the page is firstly dirtied after it's
-	 * faulted in, but write(2) could also dirty a page and set delalloc
-	 * bits, thus in this case for space account reason, we still need to
-	 * clear any delalloc bits within this page range since we have to
-	 * reserve data&meta space before lock_page() (see above comments).
-	 */
-	clear_extent_bit(&BTRFS_I(inode)->io_tree, page_start, end,
-			  EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING |
-			  EXTENT_DEFRAG, &cached_state);
-
-	ret2 = btrfs_set_extent_delalloc(BTRFS_I(inode), page_start, end, 0,
-					&cached_state);
-	if (ret2) {
-		unlock_extent(io_tree, page_start, page_end, &cached_state);
-		ret = VM_FAULT_SIGBUS;
-		goto out_unlock;
-	}
-
-	/* page is wholly or partially inside EOF */
-	if (page_start + PAGE_SIZE > size)
-		zero_start = offset_in_page(size);
-	else
-		zero_start = PAGE_SIZE;
-
-	if (zero_start != PAGE_SIZE)
-		memzero_page(page, zero_start, PAGE_SIZE - zero_start);
-
-	btrfs_folio_clear_checked(fs_info, folio, page_start, PAGE_SIZE);
-	btrfs_folio_set_dirty(fs_info, folio, page_start, end + 1 - page_start);
-	btrfs_folio_set_uptodate(fs_info, folio, page_start, end + 1 - page_start);
-
-	btrfs_set_inode_last_sub_trans(BTRFS_I(inode));
-
-	unlock_extent(io_tree, page_start, page_end, &cached_state);
-	up_read(&BTRFS_I(inode)->i_mmap_lock);
-
-	btrfs_delalloc_release_extents(BTRFS_I(inode), PAGE_SIZE);
-	sb_end_pagefault(inode->i_sb);
-	extent_changeset_free(data_reserved);
-	return VM_FAULT_LOCKED;
-
-out_unlock:
-	unlock_page(page);
-	up_read(&BTRFS_I(inode)->i_mmap_lock);
-out:
-	btrfs_delalloc_release_extents(BTRFS_I(inode), PAGE_SIZE);
-	btrfs_delalloc_release_space(BTRFS_I(inode), data_reserved, page_start,
-				     reserved_space, (ret != 0));
-out_noreserve:
-	sb_end_pagefault(inode->i_sb);
-	extent_changeset_free(data_reserved);
-	return ret;
-}
-
 static int btrfs_truncate(struct btrfs_inode *inode, bool skip_writeback)
 {
 	struct btrfs_truncate_control control = {
@@ -8716,7 +8584,7 @@ int __init btrfs_init_cachep(void)
 {
 	btrfs_inode_cachep = kmem_cache_create("btrfs_inode",
 			sizeof(struct btrfs_inode), 0,
-			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD | SLAB_ACCOUNT,
+			SLAB_RECLAIM_ACCOUNT | SLAB_ACCOUNT,
 			init_once);
 	if (!btrfs_inode_cachep)
 		goto fail;
@@ -8765,6 +8633,9 @@ static int btrfs_getattr(struct mnt_idmap *idmap,
 	generic_fillattr(idmap, request_mask, inode, stat);
 	stat->dev = BTRFS_I(inode)->root->anon_dev;
 
+	stat->subvol = BTRFS_I(inode)->root->root_key.objectid;
+	stat->result_mask |= STATX_SUBVOL;
+
 	spin_lock(&BTRFS_I(inode)->lock);
 	delalloc_bytes = BTRFS_I(inode)->new_delalloc_bytes;
 	inode_bytes = inode_get_bytes(inode);
@@ -9644,7 +9515,7 @@ free_qgroup:
 	 * or we leak qgroup data reservation.
 	 */
 	btrfs_qgroup_free_refroot(inode->root->fs_info,
-			inode->root->root_key.objectid, qgroup_released,
+			btrfs_root_id(inode->root), qgroup_released,
 			BTRFS_QGROUP_RSV_DATA);
 	return ERR_PTR(ret);
 }
@@ -10180,7 +10051,7 @@ ssize_t btrfs_encoded_read(struct kiocb *iocb, struct iov_iter *iter,
 		cond_resched();
 	}
 
-	em = btrfs_get_extent(inode, NULL, 0, start, lockend - start + 1);
+	em = btrfs_get_extent(inode, NULL, start, lockend - start + 1);
 	if (IS_ERR(em)) {
 		ret = PTR_ERR(em);
 		goto out_unlock_extent;
@@ -10292,8 +10163,8 @@ ssize_t btrfs_do_encoded_write(struct kiocb *iocb, struct iov_iter *from,
 	size_t orig_count;
 	u64 start, end;
 	u64 num_bytes, ram_bytes, disk_num_bytes;
-	unsigned long nr_pages, i;
-	struct page **pages;
+	unsigned long nr_folios, i;
+	struct folio **folios;
 	struct btrfs_key ins;
 	bool extent_reserved = false;
 	struct extent_map *em;
@@ -10382,24 +10253,24 @@ ssize_t btrfs_do_encoded_write(struct kiocb *iocb, struct iov_iter *from,
 	 * isn't.
 	 */
 	disk_num_bytes = ALIGN(orig_count, fs_info->sectorsize);
-	nr_pages = DIV_ROUND_UP(disk_num_bytes, PAGE_SIZE);
-	pages = kvcalloc(nr_pages, sizeof(struct page *), GFP_KERNEL_ACCOUNT);
-	if (!pages)
+	nr_folios = DIV_ROUND_UP(disk_num_bytes, PAGE_SIZE);
+	folios = kvcalloc(nr_folios, sizeof(struct page *), GFP_KERNEL_ACCOUNT);
+	if (!folios)
 		return -ENOMEM;
-	for (i = 0; i < nr_pages; i++) {
+	for (i = 0; i < nr_folios; i++) {
 		size_t bytes = min_t(size_t, PAGE_SIZE, iov_iter_count(from));
 		char *kaddr;
 
-		pages[i] = alloc_page(GFP_KERNEL_ACCOUNT);
-		if (!pages[i]) {
+		folios[i] = folio_alloc(GFP_KERNEL_ACCOUNT, 0);
+		if (!folios[i]) {
 			ret = -ENOMEM;
-			goto out_pages;
+			goto out_folios;
 		}
-		kaddr = kmap_local_page(pages[i]);
+		kaddr = kmap_local_folio(folios[i], 0);
 		if (copy_from_iter(kaddr, bytes, from) != bytes) {
 			kunmap_local(kaddr);
 			ret = -EFAULT;
-			goto out_pages;
+			goto out_folios;
 		}
 		if (bytes < PAGE_SIZE)
 			memset(kaddr + bytes, 0, PAGE_SIZE - bytes);
@@ -10411,12 +10282,12 @@ ssize_t btrfs_do_encoded_write(struct kiocb *iocb, struct iov_iter *from,
 
 		ret = btrfs_wait_ordered_range(&inode->vfs_inode, start, num_bytes);
 		if (ret)
-			goto out_pages;
+			goto out_folios;
 		ret = invalidate_inode_pages2_range(inode->vfs_inode.i_mapping,
 						    start >> PAGE_SHIFT,
 						    end >> PAGE_SHIFT);
 		if (ret)
-			goto out_pages;
+			goto out_folios;
 		lock_extent(io_tree, start, end, &cached_state);
 		ordered = btrfs_lookup_ordered_range(inode, start, num_bytes);
 		if (!ordered &&
@@ -10444,10 +10315,12 @@ ssize_t btrfs_do_encoded_write(struct kiocb *iocb, struct iov_iter *from,
 		goto out_qgroup_free_data;
 
 	/* Try an inline extent first. */
-	if (start == 0 && encoded->unencoded_len == encoded->len &&
-	    encoded->unencoded_offset == 0) {
-		ret = cow_file_range_inline(inode, encoded->len, orig_count,
-					    compression, pages, true);
+	if (encoded->unencoded_len == encoded->len &&
+	    encoded->unencoded_offset == 0 &&
+	    can_cow_file_range_inline(inode, start, encoded->len, orig_count)) {
+		ret = __cow_file_range_inline(inode, start, encoded->len,
+					      orig_count, compression, folios[0],
+					      true);
 		if (ret <= 0) {
 			if (ret == 0)
 				ret = orig_count;
@@ -10491,7 +10364,7 @@ ssize_t btrfs_do_encoded_write(struct kiocb *iocb, struct iov_iter *from,
 
 	btrfs_delalloc_release_extents(inode, num_bytes);
 
-	btrfs_submit_compressed_write(ordered, pages, nr_pages, 0, false);
+	btrfs_submit_compressed_write(ordered, folios, nr_folios, 0, false);
 	ret = orig_count;
 	goto out;
 
@@ -10513,12 +10386,12 @@ out_free_data_space:
 		btrfs_free_reserved_data_space_noquota(fs_info, disk_num_bytes);
 out_unlock:
 	unlock_extent(io_tree, start, end, &cached_state);
-out_pages:
-	for (i = 0; i < nr_pages; i++) {
-		if (pages[i])
-			__free_page(pages[i]);
+out_folios:
+	for (i = 0; i < nr_folios; i++) {
+		if (folios[i])
+			folio_put(folios[i]);
 	}
-	kvfree(pages);
+	kvfree(folios);
 out:
 	if (ret >= 0)
 		iocb->ki_pos += encoded->len;
@@ -10745,7 +10618,7 @@ static int btrfs_swap_activate(struct swap_info_struct *sis, struct file *file,
 		btrfs_exclop_finish(fs_info);
 		btrfs_warn(fs_info,
 		"cannot activate swapfile because subvolume %llu is being deleted",
-			root->root_key.objectid);
+			btrfs_root_id(root));
 		return -EPERM;
 	}
 	atomic_inc(&root->nr_swapfiles);
@@ -10760,7 +10633,7 @@ static int btrfs_swap_activate(struct swap_info_struct *sis, struct file *file,
 		struct btrfs_block_group *bg;
 		u64 len = isize - start;
 
-		em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, start, len);
+		em = btrfs_get_extent(BTRFS_I(inode), NULL, start, len);
 		if (IS_ERR(em)) {
 			ret = PTR_ERR(em);
 			goto out;
@@ -10971,7 +10844,7 @@ void btrfs_assert_inode_range_clean(struct btrfs_inode *inode, u64 start, u64 en
 	if (ordered) {
 		btrfs_err(root->fs_info,
 "found unexpected ordered extent in file range [%llu, %llu] for inode %llu root %llu (ordered range [%llu, %llu])",
-			  start, end, btrfs_ino(inode), root->root_key.objectid,
+			  start, end, btrfs_ino(inode), btrfs_root_id(root),
 			  ordered->file_offset,
 			  ordered->file_offset + ordered->num_bytes - 1);
 		btrfs_put_ordered_extent(ordered);
@@ -10980,6 +10853,65 @@ void btrfs_assert_inode_range_clean(struct btrfs_inode *inode, u64 start, u64 en
 	ASSERT(ordered == NULL);
 }
 
+/*
+ * Find the first inode with a minimum number.
+ *
+ * @root:	The root to search for.
+ * @min_ino:	The minimum inode number.
+ *
+ * Find the first inode in the @root with a number >= @min_ino and return it.
+ * Returns NULL if no such inode found.
+ */
+struct btrfs_inode *btrfs_find_first_inode(struct btrfs_root *root, u64 min_ino)
+{
+	struct rb_node *node;
+	struct rb_node *prev;
+	struct btrfs_inode *inode;
+
+	spin_lock(&root->inode_lock);
+again:
+	node = root->inode_tree.rb_node;
+	prev = NULL;
+	while (node) {
+		prev = node;
+		inode = rb_entry(node, struct btrfs_inode, rb_node);
+		if (min_ino < btrfs_ino(inode))
+			node = node->rb_left;
+		else if (min_ino > btrfs_ino(inode))
+			node = node->rb_right;
+		else
+			break;
+	}
+
+	if (!node) {
+		while (prev) {
+			inode = rb_entry(prev, struct btrfs_inode, rb_node);
+			if (min_ino <= btrfs_ino(inode)) {
+				node = prev;
+				break;
+			}
+			prev = rb_next(prev);
+		}
+	}
+
+	while (node) {
+		inode = rb_entry(prev, struct btrfs_inode, rb_node);
+		if (igrab(&inode->vfs_inode)) {
+			spin_unlock(&root->inode_lock);
+			return inode;
+		}
+
+		min_ino = btrfs_ino(inode) + 1;
+		if (cond_resched_lock(&root->inode_lock))
+			goto again;
+
+		node = rb_next(node);
+	}
+	spin_unlock(&root->inode_lock);
+
+	return NULL;
+}
+
 static const struct inode_operations btrfs_dir_inode_operations = {
 	.getattr	= btrfs_getattr,
 	.lookup		= btrfs_lookup,
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 8851ba7a1e..c1b0556e40 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -34,11 +34,9 @@
 #include "export.h"
 #include "transaction.h"
 #include "btrfs_inode.h"
-#include "print-tree.h"
 #include "volumes.h"
 #include "locking.h"
 #include "backref.h"
-#include "rcu-string.h"
 #include "send.h"
 #include "dev-replace.h"
 #include "props.h"
@@ -47,9 +45,7 @@
 #include "tree-log.h"
 #include "compression.h"
 #include "space-info.h"
-#include "delalloc-space.h"
 #include "block-group.h"
-#include "subpage.h"
 #include "fs.h"
 #include "accessors.h"
 #include "extent-tree.h"
@@ -231,6 +227,20 @@ static int check_fsflags_compatible(struct btrfs_fs_info *fs_info,
 	return 0;
 }
 
+int btrfs_check_ioctl_vol_args_path(const struct btrfs_ioctl_vol_args *vol_args)
+{
+	if (memchr(vol_args->name, 0, sizeof(vol_args->name)) == NULL)
+		return -ENAMETOOLONG;
+	return 0;
+}
+
+static int btrfs_check_ioctl_vol_args2_subvol_name(const struct btrfs_ioctl_vol_args_v2 *vol_args2)
+{
+	if (memchr(vol_args2->name, 0, sizeof(vol_args2->name)) == NULL)
+		return -ENAMETOOLONG;
+	return 0;
+}
+
 /*
  * Set flags/xflags from the internal inode flags. The remaining items of
  * fsxattr are zeroed.
@@ -658,7 +668,7 @@ static noinline int create_subvol(struct mnt_idmap *idmap,
 	/* Tree log can't currently deal with an inode which is a new root. */
 	btrfs_set_log_full_commit(trans);
 
-	ret = btrfs_qgroup_inherit(trans, 0, objectid, root->root_key.objectid, inherit);
+	ret = btrfs_qgroup_inherit(trans, 0, objectid, btrfs_root_id(root), inherit);
 	if (ret)
 		goto out;
 
@@ -709,6 +719,8 @@ static noinline int create_subvol(struct mnt_idmap *idmap,
 	ret = btrfs_insert_root(trans, fs_info->tree_root, &key,
 				root_item);
 	if (ret) {
+		int ret2;
+
 		/*
 		 * Since we don't abort the transaction in this case, free the
 		 * tree block so that we don't leak space and leave the
@@ -719,7 +731,9 @@ static noinline int create_subvol(struct mnt_idmap *idmap,
 		btrfs_tree_lock(leaf);
 		btrfs_clear_buffer_dirty(trans, leaf);
 		btrfs_tree_unlock(leaf);
-		btrfs_free_tree_block(trans, objectid, leaf, 0, 1);
+		ret2 = btrfs_free_tree_block(trans, objectid, leaf, 0, 1);
+		if (ret2 < 0)
+			btrfs_abort_transaction(trans, ret2);
 		free_extent_buffer(leaf);
 		goto out;
 	}
@@ -929,7 +943,9 @@ static int btrfs_may_delete(struct mnt_idmap *idmap,
 	if (d_really_is_negative(victim))
 		return -ENOENT;
 
-	BUG_ON(d_inode(victim->d_parent) != dir);
+	/* The @victim is not inside @dir. */
+	if (d_inode(victim->d_parent) != dir)
+		return -EINVAL;
 	audit_inode_child(dir, victim, AUDIT_TYPE_CHILD_DELETE);
 
 	error = inode_permission(idmap, dir, MAY_WRITE | MAY_EXEC);
@@ -1147,7 +1163,10 @@ static noinline int btrfs_ioctl_resize(struct file *file,
 		ret = PTR_ERR(vol_args);
 		goto out_drop;
 	}
-	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
+	ret = btrfs_check_ioctl_vol_args_path(vol_args);
+	if (ret < 0)
+		goto out_free;
+
 	sizestr = vol_args->name;
 	cancel = (strcmp("cancel", sizestr) == 0);
 	ret = exclop_start_or_cancel_reloc(fs_info, BTRFS_EXCLOP_RESIZE, cancel);
@@ -1347,12 +1366,15 @@ static noinline int btrfs_ioctl_snap_create(struct file *file,
 	vol_args = memdup_user(arg, sizeof(*vol_args));
 	if (IS_ERR(vol_args))
 		return PTR_ERR(vol_args);
-	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
+	ret = btrfs_check_ioctl_vol_args_path(vol_args);
+	if (ret < 0)
+		goto out;
 
 	ret = __btrfs_ioctl_snap_create(file, file_mnt_idmap(file),
 					vol_args->name, vol_args->fd, subvol,
 					false, NULL);
 
+out:
 	kfree(vol_args);
 	return ret;
 }
@@ -1371,7 +1393,9 @@ static noinline int btrfs_ioctl_snap_create_v2(struct file *file,
 	vol_args = memdup_user(arg, sizeof(*vol_args));
 	if (IS_ERR(vol_args))
 		return PTR_ERR(vol_args);
-	vol_args->name[BTRFS_SUBVOL_NAME_MAX] = '\0';
+	ret = btrfs_check_ioctl_vol_args2_subvol_name(vol_args);
+	if (ret < 0)
+		goto free_args;
 
 	if (vol_args->flags & ~BTRFS_SUBVOL_CREATE_ARGS_MASK) {
 		ret = -EOPNOTSUPP;
@@ -1490,7 +1514,7 @@ static noinline int btrfs_ioctl_subvol_setflags(struct file *file,
 			spin_unlock(&root->root_item_lock);
 			btrfs_warn(fs_info,
 				   "Attempt to set subvolume %llu read-write during send",
-				   root->root_key.objectid);
+				   btrfs_root_id(root));
 			ret = -EPERM;
 			goto out_drop_sem;
 		}
@@ -1899,7 +1923,7 @@ static int btrfs_search_path_in_tree_user(struct mnt_idmap *idmap,
 	struct btrfs_fs_info *fs_info = BTRFS_I(inode)->root->fs_info;
 	struct super_block *sb = inode->i_sb;
 	struct btrfs_key upper_limit = BTRFS_I(inode)->location;
-	u64 treeid = BTRFS_I(inode)->root->root_key.objectid;
+	u64 treeid = btrfs_root_id(BTRFS_I(inode)->root);
 	u64 dirid = args->dirid;
 	unsigned long item_off;
 	unsigned long item_len;
@@ -2071,7 +2095,7 @@ static noinline int btrfs_ioctl_ino_lookup(struct btrfs_root *root,
 	 * path is reset so it's consistent with btrfs_search_path_in_tree.
 	 */
 	if (args->treeid == 0)
-		args->treeid = root->root_key.objectid;
+		args->treeid = btrfs_root_id(root);
 
 	if (args->objectid == BTRFS_FIRST_FREE_OBJECTID) {
 		args->name[0] = 0;
@@ -2167,7 +2191,7 @@ static int btrfs_ioctl_get_subvol_info(struct inode *inode, void __user *argp)
 	fs_info = BTRFS_I(inode)->root->fs_info;
 
 	/* Get root_item of inode's subvolume */
-	key.objectid = BTRFS_I(inode)->root->root_key.objectid;
+	key.objectid = btrfs_root_id(BTRFS_I(inode)->root);
 	root = btrfs_get_fs_root(fs_info, key.objectid, true);
 	if (IS_ERR(root)) {
 		ret = PTR_ERR(root);
@@ -2282,7 +2306,7 @@ static int btrfs_ioctl_get_subvol_rootref(struct btrfs_root *root,
 		return PTR_ERR(rootrefs);
 	}
 
-	objectid = root->root_key.objectid;
+	objectid = btrfs_root_id(root);
 	key.objectid = objectid;
 	key.type = BTRFS_ROOT_REF_KEY;
 	key.offset = rootrefs->min_treeid;
@@ -2366,7 +2390,7 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 	struct mnt_idmap *idmap = file_mnt_idmap(file);
 	char *subvol_name, *subvol_name_ptr = NULL;
 	int subvol_namelen;
-	int err = 0;
+	int ret = 0;
 	bool destroy_parent = false;
 
 	/* We don't support snapshots with extent tree v2 yet. */
@@ -2382,7 +2406,7 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 			return PTR_ERR(vol_args2);
 
 		if (vol_args2->flags & ~BTRFS_SUBVOL_DELETE_ARGS_MASK) {
-			err = -EOPNOTSUPP;
+			ret = -EOPNOTSUPP;
 			goto out;
 		}
 
@@ -2391,29 +2415,31 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 		 * name, same as v1 currently does.
 		 */
 		if (!(vol_args2->flags & BTRFS_SUBVOL_SPEC_BY_ID)) {
-			vol_args2->name[BTRFS_SUBVOL_NAME_MAX] = 0;
+			ret = btrfs_check_ioctl_vol_args2_subvol_name(vol_args2);
+			if (ret < 0)
+				goto out;
 			subvol_name = vol_args2->name;
 
-			err = mnt_want_write_file(file);
-			if (err)
+			ret = mnt_want_write_file(file);
+			if (ret)
 				goto out;
 		} else {
 			struct inode *old_dir;
 
 			if (vol_args2->subvolid < BTRFS_FIRST_FREE_OBJECTID) {
-				err = -EINVAL;
+				ret = -EINVAL;
 				goto out;
 			}
 
-			err = mnt_want_write_file(file);
-			if (err)
+			ret = mnt_want_write_file(file);
+			if (ret)
 				goto out;
 
 			dentry = btrfs_get_dentry(fs_info->sb,
 					BTRFS_FIRST_FREE_OBJECTID,
 					vol_args2->subvolid, 0);
 			if (IS_ERR(dentry)) {
-				err = PTR_ERR(dentry);
+				ret = PTR_ERR(dentry);
 				goto out_drop_write;
 			}
 
@@ -2433,7 +2459,7 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 			 */
 			dput(dentry);
 			if (IS_ERR(parent)) {
-				err = PTR_ERR(parent);
+				ret = PTR_ERR(parent);
 				goto out_drop_write;
 			}
 			old_dir = dir;
@@ -2457,14 +2483,14 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 			 * to delete without an idmapped mount.
 			 */
 			if (old_dir != dir && idmap != &nop_mnt_idmap) {
-				err = -EOPNOTSUPP;
+				ret = -EOPNOTSUPP;
 				goto free_parent;
 			}
 
 			subvol_name_ptr = btrfs_get_subvol_name_from_objectid(
 						fs_info, vol_args2->subvolid);
 			if (IS_ERR(subvol_name_ptr)) {
-				err = PTR_ERR(subvol_name_ptr);
+				ret = PTR_ERR(subvol_name_ptr);
 				goto free_parent;
 			}
 			/* subvol_name_ptr is already nul terminated */
@@ -2475,11 +2501,14 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 		if (IS_ERR(vol_args))
 			return PTR_ERR(vol_args);
 
-		vol_args->name[BTRFS_PATH_NAME_MAX] = 0;
+		ret = btrfs_check_ioctl_vol_args_path(vol_args);
+		if (ret < 0)
+			goto out;
+
 		subvol_name = vol_args->name;
 
-		err = mnt_want_write_file(file);
-		if (err)
+		ret = mnt_want_write_file(file);
+		if (ret)
 			goto out;
 	}
 
@@ -2487,26 +2516,26 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 
 	if (strchr(subvol_name, '/') ||
 	    strncmp(subvol_name, "..", subvol_namelen) == 0) {
-		err = -EINVAL;
+		ret = -EINVAL;
 		goto free_subvol_name;
 	}
 
 	if (!S_ISDIR(dir->i_mode)) {
-		err = -ENOTDIR;
+		ret = -ENOTDIR;
 		goto free_subvol_name;
 	}
 
-	err = down_write_killable_nested(&dir->i_rwsem, I_MUTEX_PARENT);
-	if (err == -EINTR)
+	ret = down_write_killable_nested(&dir->i_rwsem, I_MUTEX_PARENT);
+	if (ret == -EINTR)
 		goto free_subvol_name;
 	dentry = lookup_one(idmap, subvol_name, parent, subvol_namelen);
 	if (IS_ERR(dentry)) {
-		err = PTR_ERR(dentry);
+		ret = PTR_ERR(dentry);
 		goto out_unlock_dir;
 	}
 
 	if (d_really_is_negative(dentry)) {
-		err = -ENOENT;
+		ret = -ENOENT;
 		goto out_dput;
 	}
 
@@ -2526,7 +2555,7 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 		 * Users who want to delete empty subvols should try
 		 * rmdir(2).
 		 */
-		err = -EPERM;
+		ret = -EPERM;
 		if (!btrfs_test_opt(fs_info, USER_SUBVOL_RM_ALLOWED))
 			goto out_dput;
 
@@ -2537,29 +2566,29 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 		 * of the subvol, not a random directory contained
 		 * within it.
 		 */
-		err = -EINVAL;
+		ret = -EINVAL;
 		if (root == dest)
 			goto out_dput;
 
-		err = inode_permission(idmap, inode, MAY_WRITE | MAY_EXEC);
-		if (err)
+		ret = inode_permission(idmap, inode, MAY_WRITE | MAY_EXEC);
+		if (ret)
 			goto out_dput;
 	}
 
 	/* check if subvolume may be deleted by a user */
-	err = btrfs_may_delete(idmap, dir, dentry, 1);
-	if (err)
+	ret = btrfs_may_delete(idmap, dir, dentry, 1);
+	if (ret)
 		goto out_dput;
 
 	if (btrfs_ino(BTRFS_I(inode)) != BTRFS_FIRST_FREE_OBJECTID) {
-		err = -EINVAL;
+		ret = -EINVAL;
 		goto out_dput;
 	}
 
 	btrfs_inode_lock(BTRFS_I(inode), 0);
-	err = btrfs_delete_subvolume(BTRFS_I(dir), dentry);
+	ret = btrfs_delete_subvolume(BTRFS_I(dir), dentry);
 	btrfs_inode_unlock(BTRFS_I(inode), 0);
-	if (!err)
+	if (!ret)
 		d_delete_notify(dir, dentry);
 
 out_dput:
@@ -2576,7 +2605,7 @@ out_drop_write:
 out:
 	kfree(vol_args2);
 	kfree(vol_args);
-	return err;
+	return ret;
 }
 
 static int btrfs_ioctl_defrag(struct file *file, void __user *argp)
@@ -2686,12 +2715,16 @@ static long btrfs_ioctl_add_dev(struct btrfs_fs_info *fs_info, void __user *arg)
 		goto out;
 	}
 
-	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
+	ret = btrfs_check_ioctl_vol_args_path(vol_args);
+	if (ret < 0)
+		goto out_free;
+
 	ret = btrfs_init_new_device(fs_info, vol_args->name);
 
 	if (!ret)
 		btrfs_info(fs_info, "disk added %s", vol_args->name);
 
+out_free:
 	kfree(vol_args);
 out:
 	if (restore_op)
@@ -2707,7 +2740,7 @@ static long btrfs_ioctl_rm_dev_v2(struct file *file, void __user *arg)
 	struct inode *inode = file_inode(file);
 	struct btrfs_fs_info *fs_info = inode_to_fs_info(inode);
 	struct btrfs_ioctl_vol_args_v2 *vol_args;
-	struct bdev_handle *bdev_handle = NULL;
+	struct file *bdev_file = NULL;
 	int ret;
 	bool cancel = false;
 
@@ -2723,7 +2756,10 @@ static long btrfs_ioctl_rm_dev_v2(struct file *file, void __user *arg)
 		goto out;
 	}
 
-	vol_args->name[BTRFS_SUBVOL_NAME_MAX] = '\0';
+	ret = btrfs_check_ioctl_vol_args2_subvol_name(vol_args);
+	if (ret < 0)
+		goto out;
+
 	if (vol_args->flags & BTRFS_DEVICE_SPEC_BY_ID) {
 		args.devid = vol_args->devid;
 	} else if (!strcmp("cancel", vol_args->name)) {
@@ -2744,7 +2780,7 @@ static long btrfs_ioctl_rm_dev_v2(struct file *file, void __user *arg)
 		goto err_drop;
 
 	/* Exclusive operation is now claimed */
-	ret = btrfs_rm_device(fs_info, &args, &bdev_handle);
+	ret = btrfs_rm_device(fs_info, &args, &bdev_file);
 
 	btrfs_exclop_finish(fs_info);
 
@@ -2758,8 +2794,8 @@ static long btrfs_ioctl_rm_dev_v2(struct file *file, void __user *arg)
 	}
 err_drop:
 	mnt_drop_write_file(file);
-	if (bdev_handle)
-		bdev_release(bdev_handle);
+	if (bdev_file)
+		fput(bdev_file);
 out:
 	btrfs_put_dev_args_from_path(&args);
 	kfree(vol_args);
@@ -2772,7 +2808,7 @@ static long btrfs_ioctl_rm_dev(struct file *file, void __user *arg)
 	struct inode *inode = file_inode(file);
 	struct btrfs_fs_info *fs_info = inode_to_fs_info(inode);
 	struct btrfs_ioctl_vol_args *vol_args;
-	struct bdev_handle *bdev_handle = NULL;
+	struct file *bdev_file = NULL;
 	int ret;
 	bool cancel = false;
 
@@ -2783,7 +2819,10 @@ static long btrfs_ioctl_rm_dev(struct file *file, void __user *arg)
 	if (IS_ERR(vol_args))
 		return PTR_ERR(vol_args);
 
-	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
+	ret = btrfs_check_ioctl_vol_args_path(vol_args);
+	if (ret < 0)
+		goto out_free;
+
 	if (!strcmp("cancel", vol_args->name)) {
 		cancel = true;
 	} else {
@@ -2799,17 +2838,18 @@ static long btrfs_ioctl_rm_dev(struct file *file, void __user *arg)
 	ret = exclop_start_or_cancel_reloc(fs_info, BTRFS_EXCLOP_DEV_REMOVE,
 					   cancel);
 	if (ret == 0) {
-		ret = btrfs_rm_device(fs_info, &args, &bdev_handle);
+		ret = btrfs_rm_device(fs_info, &args, &bdev_file);
 		if (!ret)
 			btrfs_info(fs_info, "disk deleted %s", vol_args->name);
 		btrfs_exclop_finish(fs_info);
 	}
 
 	mnt_drop_write_file(file);
-	if (bdev_handle)
-		bdev_release(bdev_handle);
+	if (bdev_file)
+		fput(bdev_file);
 out:
 	btrfs_put_dev_args_from_path(&args);
+out_free:
 	kfree(vol_args);
 	return ret;
 }
@@ -2945,7 +2985,7 @@ static long btrfs_ioctl_default_subvol(struct file *file, void __user *argp)
 		ret = PTR_ERR(new_root);
 		goto out;
 	}
-	if (!is_fstree(new_root->root_key.objectid)) {
+	if (!is_fstree(btrfs_root_id(new_root))) {
 		ret = -ENOENT;
 		goto out_free;
 	}
@@ -3911,7 +3951,7 @@ static long btrfs_ioctl_qgroup_limit(struct file *file, void __user *arg)
 	qgroupid = sa->qgroupid;
 	if (!qgroupid) {
 		/* take the current subvol as qgroup */
-		qgroupid = root->root_key.objectid;
+		qgroupid = btrfs_root_id(root);
 	}
 
 	ret = btrfs_limit_qgroup(trans, qgroupid, &sa->lim);
@@ -4042,7 +4082,7 @@ static long _btrfs_ioctl_set_received_subvol(struct file *file,
 	    !btrfs_is_empty_uuid(root_item->received_uuid)) {
 		ret = btrfs_uuid_tree_remove(trans, root_item->received_uuid,
 					  BTRFS_UUID_KEY_RECEIVED_SUBVOL,
-					  root->root_key.objectid);
+					  btrfs_root_id(root));
 		if (ret && ret != -ENOENT) {
 		        btrfs_abort_transaction(trans, ret);
 		        btrfs_end_transaction(trans);
@@ -4066,7 +4106,7 @@ static long _btrfs_ioctl_set_received_subvol(struct file *file,
 	if (received_uuid_changed && !btrfs_is_empty_uuid(sa->uuid)) {
 		ret = btrfs_uuid_tree_add(trans, sa->uuid,
 					  BTRFS_UUID_KEY_RECEIVED_SUBVOL,
-					  root->root_key.objectid);
+					  btrfs_root_id(root));
 		if (ret < 0 && ret != -EEXIST) {
 			btrfs_abort_transaction(trans, ret);
 			btrfs_end_transaction(trans);
diff --git a/fs/btrfs/ioctl.h b/fs/btrfs/ioctl.h
index d51b9a2f2f..2c5dc25ec6 100644
--- a/fs/btrfs/ioctl.h
+++ b/fs/btrfs/ioctl.h
@@ -3,6 +3,15 @@
 #ifndef BTRFS_IOCTL_H
 #define BTRFS_IOCTL_H
 
+#include <linux/types.h>
+
+struct file;
+struct dentry;
+struct mnt_idmap;
+struct fileattr;
+struct btrfs_fs_info;
+struct btrfs_ioctl_balance_args;
+
 long btrfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
 long btrfs_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
 int btrfs_fileattr_get(struct dentry *dentry, struct fileattr *fa);
diff --git a/fs/btrfs/locking.c b/fs/btrfs/locking.c
index 74d8e2003f..6a0b7abb5b 100644
--- a/fs/btrfs/locking.c
+++ b/fs/btrfs/locking.c
@@ -13,7 +13,6 @@
 #include "ctree.h"
 #include "extent_io.h"
 #include "locking.h"
-#include "accessors.h"
 
 /*
  * Lockdep class keys for extent_buffer->lock's in this root.  For a given
@@ -85,7 +84,7 @@ void btrfs_set_buffer_lockdep_class(u64 objectid, struct extent_buffer *eb, int
 {
 	struct btrfs_lockdep_keyset *ks;
 
-	BUG_ON(level >= ARRAY_SIZE(ks->keys));
+	ASSERT(level < ARRAY_SIZE(ks->keys));
 
 	/* Find the matching keyset, id 0 is the default entry */
 	for (ks = btrfs_lockdep_keysets; ks->id; ks++)
@@ -98,7 +97,7 @@ void btrfs_set_buffer_lockdep_class(u64 objectid, struct extent_buffer *eb, int
 void btrfs_maybe_reset_lockdep_class(struct btrfs_root *root, struct extent_buffer *eb)
 {
 	if (test_bit(BTRFS_ROOT_RESET_LOCKDEP_CLASS, &root->state))
-		btrfs_set_buffer_lockdep_class(root->root_key.objectid,
+		btrfs_set_buffer_lockdep_class(btrfs_root_id(root),
 					       eb, btrfs_header_level(eb));
 }
 
@@ -130,14 +129,14 @@ static void btrfs_set_eb_lock_owner(struct extent_buffer *eb, pid_t owner) { }
  */
 
 /*
- * __btrfs_tree_read_lock - lock extent buffer for read
+ * btrfs_tree_read_lock_nested - lock extent buffer for read
  * @eb:		the eb to be locked
  * @nest:	the nesting level to be used for lockdep
  *
  * This takes the read lock on the extent buffer, using the specified nesting
  * level for lockdep purposes.
  */
-void __btrfs_tree_read_lock(struct extent_buffer *eb, enum btrfs_lock_nesting nest)
+void btrfs_tree_read_lock_nested(struct extent_buffer *eb, enum btrfs_lock_nesting nest)
 {
 	u64 start_ns = 0;
 
@@ -148,11 +147,6 @@ void __btrfs_tree_read_lock(struct extent_buffer *eb, enum btrfs_lock_nesting ne
 	trace_btrfs_tree_read_lock(eb, start_ns);
 }
 
-void btrfs_tree_read_lock(struct extent_buffer *eb)
-{
-	__btrfs_tree_read_lock(eb, BTRFS_NESTING_NORMAL);
-}
-
 /*
  * Try-lock for read.
  *
@@ -199,7 +193,7 @@ void btrfs_tree_read_unlock(struct extent_buffer *eb)
  *
  * Returns with the eb->lock write locked.
  */
-void __btrfs_tree_lock(struct extent_buffer *eb, enum btrfs_lock_nesting nest)
+void btrfs_tree_lock_nested(struct extent_buffer *eb, enum btrfs_lock_nesting nest)
 	__acquires(&eb->lock)
 {
 	u64 start_ns = 0;
@@ -212,11 +206,6 @@ void __btrfs_tree_lock(struct extent_buffer *eb, enum btrfs_lock_nesting nest)
 	trace_btrfs_tree_lock(eb, start_ns);
 }
 
-void btrfs_tree_lock(struct extent_buffer *eb)
-{
-	__btrfs_tree_lock(eb, BTRFS_NESTING_NORMAL);
-}
-
 /*
  * Release the write lock.
  */
@@ -375,8 +364,12 @@ void btrfs_drew_write_lock(struct btrfs_drew_lock *lock)
 
 void btrfs_drew_write_unlock(struct btrfs_drew_lock *lock)
 {
-	atomic_dec(&lock->writers);
-	cond_wake_up(&lock->pending_readers);
+	/*
+	 * atomic_dec_and_test() implies a full barrier, so woken up readers are
+	 * guaranteed to see the decrement.
+	 */
+	if (atomic_dec_and_test(&lock->writers))
+		wake_up(&lock->pending_readers);
 }
 
 void btrfs_drew_read_lock(struct btrfs_drew_lock *lock)
diff --git a/fs/btrfs/locking.h b/fs/btrfs/locking.h
index 7d6ee1e609..1bc8e67388 100644
--- a/fs/btrfs/locking.h
+++ b/fs/btrfs/locking.h
@@ -8,8 +8,14 @@
 
 #include <linux/atomic.h>
 #include <linux/wait.h>
+#include <linux/lockdep.h>
 #include <linux/percpu_counter.h>
 #include "extent_io.h"
+#include "locking.h"
+
+struct extent_buffer;
+struct btrfs_path;
+struct btrfs_root;
 
 #define BTRFS_WRITE_LOCK 1
 #define BTRFS_READ_LOCK 2
@@ -157,14 +163,22 @@ enum btrfs_lockdep_trans_states {
 static_assert(BTRFS_NESTING_MAX <= MAX_LOCKDEP_SUBCLASSES,
 	      "too many lock subclasses defined");
 
-struct btrfs_path;
+void btrfs_tree_lock_nested(struct extent_buffer *eb, enum btrfs_lock_nesting nest);
+
+static inline void btrfs_tree_lock(struct extent_buffer *eb)
+{
+	btrfs_tree_lock_nested(eb, BTRFS_NESTING_NORMAL);
+}
 
-void __btrfs_tree_lock(struct extent_buffer *eb, enum btrfs_lock_nesting nest);
-void btrfs_tree_lock(struct extent_buffer *eb);
 void btrfs_tree_unlock(struct extent_buffer *eb);
 
-void __btrfs_tree_read_lock(struct extent_buffer *eb, enum btrfs_lock_nesting nest);
-void btrfs_tree_read_lock(struct extent_buffer *eb);
+void btrfs_tree_read_lock_nested(struct extent_buffer *eb, enum btrfs_lock_nesting nest);
+
+static inline void btrfs_tree_read_lock(struct extent_buffer *eb)
+{
+	btrfs_tree_read_lock_nested(eb, BTRFS_NESTING_NORMAL);
+}
+
 void btrfs_tree_read_unlock(struct extent_buffer *eb);
 int btrfs_try_tree_read_lock(struct extent_buffer *eb);
 int btrfs_try_tree_write_lock(struct extent_buffer *eb);
diff --git a/fs/btrfs/lru_cache.h b/fs/btrfs/lru_cache.h
index 00328c856b..e32906ab6f 100644
--- a/fs/btrfs/lru_cache.h
+++ b/fs/btrfs/lru_cache.h
@@ -3,8 +3,10 @@
 #ifndef BTRFS_LRU_CACHE_H
 #define BTRFS_LRU_CACHE_H
 
+#include <linux/types.h>
 #include <linux/maple_tree.h>
 #include <linux/list.h>
+#include "lru_cache.h"
 
 /*
  * A cache entry. This is meant to be embedded in a structure of a user of
@@ -50,11 +52,6 @@ struct btrfs_lru_cache {
 #define btrfs_lru_cache_for_each_entry_safe(cache, entry, tmp)		\
 	list_for_each_entry_safe_reverse((entry), (tmp), &(cache)->lru_list, lru_list)
 
-static inline unsigned int btrfs_lru_cache_size(const struct btrfs_lru_cache *cache)
-{
-	return cache->size;
-}
-
 static inline struct btrfs_lru_cache_entry *btrfs_lru_cache_lru_entry(
 					      struct btrfs_lru_cache *cache)
 {
diff --git a/fs/btrfs/lzo.c b/fs/btrfs/lzo.c
index 3e5d3b7028..1c396ac167 100644
--- a/fs/btrfs/lzo.c
+++ b/fs/btrfs/lzo.c
@@ -130,17 +130,17 @@ static inline size_t read_compress_length(const char *buf)
  */
 static int copy_compressed_data_to_page(char *compressed_data,
 					size_t compressed_size,
-					struct page **out_pages,
-					unsigned long max_nr_page,
+					struct folio **out_folios,
+					unsigned long max_nr_folio,
 					u32 *cur_out,
 					const u32 sectorsize)
 {
 	u32 sector_bytes_left;
 	u32 orig_out;
-	struct page *cur_page;
+	struct folio *cur_folio;
 	char *kaddr;
 
-	if ((*cur_out / PAGE_SIZE) >= max_nr_page)
+	if ((*cur_out / PAGE_SIZE) >= max_nr_folio)
 		return -E2BIG;
 
 	/*
@@ -149,16 +149,16 @@ static int copy_compressed_data_to_page(char *compressed_data,
 	 */
 	ASSERT((*cur_out / sectorsize) == (*cur_out + LZO_LEN - 1) / sectorsize);
 
-	cur_page = out_pages[*cur_out / PAGE_SIZE];
+	cur_folio = out_folios[*cur_out / PAGE_SIZE];
 	/* Allocate a new page */
-	if (!cur_page) {
-		cur_page = btrfs_alloc_compr_page();
-		if (!cur_page)
+	if (!cur_folio) {
+		cur_folio = btrfs_alloc_compr_folio();
+		if (!cur_folio)
 			return -ENOMEM;
-		out_pages[*cur_out / PAGE_SIZE] = cur_page;
+		out_folios[*cur_out / PAGE_SIZE] = cur_folio;
 	}
 
-	kaddr = kmap_local_page(cur_page);
+	kaddr = kmap_local_folio(cur_folio, 0);
 	write_compress_length(kaddr + offset_in_page(*cur_out),
 			      compressed_size);
 	*cur_out += LZO_LEN;
@@ -172,18 +172,18 @@ static int copy_compressed_data_to_page(char *compressed_data,
 
 		kunmap_local(kaddr);
 
-		if ((*cur_out / PAGE_SIZE) >= max_nr_page)
+		if ((*cur_out / PAGE_SIZE) >= max_nr_folio)
 			return -E2BIG;
 
-		cur_page = out_pages[*cur_out / PAGE_SIZE];
+		cur_folio = out_folios[*cur_out / PAGE_SIZE];
 		/* Allocate a new page */
-		if (!cur_page) {
-			cur_page = btrfs_alloc_compr_page();
-			if (!cur_page)
+		if (!cur_folio) {
+			cur_folio = btrfs_alloc_compr_folio();
+			if (!cur_folio)
 				return -ENOMEM;
-			out_pages[*cur_out / PAGE_SIZE] = cur_page;
+			out_folios[*cur_out / PAGE_SIZE] = cur_folio;
 		}
-		kaddr = kmap_local_page(cur_page);
+		kaddr = kmap_local_folio(cur_folio, 0);
 
 		memcpy(kaddr + offset_in_page(*cur_out),
 		       compressed_data + *cur_out - orig_out, copy_len);
@@ -209,15 +209,15 @@ out:
 	return 0;
 }
 
-int lzo_compress_pages(struct list_head *ws, struct address_space *mapping,
-		u64 start, struct page **pages, unsigned long *out_pages,
-		unsigned long *total_in, unsigned long *total_out)
+int lzo_compress_folios(struct list_head *ws, struct address_space *mapping,
+			u64 start, struct folio **folios, unsigned long *out_folios,
+			unsigned long *total_in, unsigned long *total_out)
 {
 	struct workspace *workspace = list_entry(ws, struct workspace, list);
 	const u32 sectorsize = inode_to_fs_info(mapping->host)->sectorsize;
-	struct page *page_in = NULL;
+	struct folio *folio_in = NULL;
 	char *sizes_ptr;
-	const unsigned long max_nr_page = *out_pages;
+	const unsigned long max_nr_folio = *out_folios;
 	int ret = 0;
 	/* Points to the file offset of input data */
 	u64 cur_in = start;
@@ -225,8 +225,8 @@ int lzo_compress_pages(struct list_head *ws, struct address_space *mapping,
 	u32 cur_out = 0;
 	u32 len = *total_out;
 
-	ASSERT(max_nr_page > 0);
-	*out_pages = 0;
+	ASSERT(max_nr_folio > 0);
+	*out_folios = 0;
 	*total_out = 0;
 	*total_in = 0;
 
@@ -243,15 +243,16 @@ int lzo_compress_pages(struct list_head *ws, struct address_space *mapping,
 		size_t out_len;
 
 		/* Get the input page first */
-		if (!page_in) {
-			page_in = find_get_page(mapping, cur_in >> PAGE_SHIFT);
-			ASSERT(page_in);
+		if (!folio_in) {
+			ret = btrfs_compress_filemap_get_folio(mapping, cur_in, &folio_in);
+			if (ret < 0)
+				goto out;
 		}
 
 		/* Compress at most one sector of data each time */
 		in_len = min_t(u32, start + len - cur_in, sectorsize - sector_off);
 		ASSERT(in_len);
-		data_in = kmap_local_page(page_in);
+		data_in = kmap_local_folio(folio_in, 0);
 		ret = lzo1x_1_compress(data_in +
 				       offset_in_page(cur_in), in_len,
 				       workspace->cbuf, &out_len,
@@ -264,7 +265,7 @@ int lzo_compress_pages(struct list_head *ws, struct address_space *mapping,
 		}
 
 		ret = copy_compressed_data_to_page(workspace->cbuf, out_len,
-						   pages, max_nr_page,
+						   folios, max_nr_folio,
 						   &cur_out, sectorsize);
 		if (ret < 0)
 			goto out;
@@ -282,13 +283,13 @@ int lzo_compress_pages(struct list_head *ws, struct address_space *mapping,
 
 		/* Check if we have reached page boundary */
 		if (PAGE_ALIGNED(cur_in)) {
-			put_page(page_in);
-			page_in = NULL;
+			folio_put(folio_in);
+			folio_in = NULL;
 		}
 	}
 
 	/* Store the size of all chunks of compressed data */
-	sizes_ptr = kmap_local_page(pages[0]);
+	sizes_ptr = kmap_local_folio(folios[0], 0);
 	write_compress_length(sizes_ptr, cur_out);
 	kunmap_local(sizes_ptr);
 
@@ -296,9 +297,9 @@ int lzo_compress_pages(struct list_head *ws, struct address_space *mapping,
 	*total_out = cur_out;
 	*total_in = cur_in - start;
 out:
-	if (page_in)
-		put_page(page_in);
-	*out_pages = DIV_ROUND_UP(cur_out, PAGE_SIZE);
+	if (folio_in)
+		folio_put(folio_in);
+	*out_folios = DIV_ROUND_UP(cur_out, PAGE_SIZE);
 	return ret;
 }
 
@@ -313,15 +314,15 @@ static void copy_compressed_segment(struct compressed_bio *cb,
 	u32 orig_in = *cur_in;
 
 	while (*cur_in < orig_in + len) {
-		struct page *cur_page;
+		struct folio *cur_folio;
 		u32 copy_len = min_t(u32, PAGE_SIZE - offset_in_page(*cur_in),
 					  orig_in + len - *cur_in);
 
 		ASSERT(copy_len);
-		cur_page = cb->compressed_pages[*cur_in / PAGE_SIZE];
+		cur_folio = cb->compressed_folios[*cur_in / PAGE_SIZE];
 
-		memcpy_from_page(dest + *cur_in - orig_in, cur_page,
-				 offset_in_page(*cur_in), copy_len);
+		memcpy_from_folio(dest + *cur_in - orig_in, cur_folio,
+				  offset_in_folio(cur_folio, *cur_in), copy_len);
 
 		*cur_in += copy_len;
 	}
@@ -341,7 +342,7 @@ int lzo_decompress_bio(struct list_head *ws, struct compressed_bio *cb)
 	/* Bytes decompressed so far */
 	u32 cur_out = 0;
 
-	kaddr = kmap_local_page(cb->compressed_pages[0]);
+	kaddr = kmap_local_folio(cb->compressed_folios[0], 0);
 	len_in = read_compress_length(kaddr);
 	kunmap_local(kaddr);
 	cur_in += LZO_LEN;
@@ -363,7 +364,7 @@ int lzo_decompress_bio(struct list_head *ws, struct compressed_bio *cb)
 
 	/* Go through each lzo segment */
 	while (cur_in < len_in) {
-		struct page *cur_page;
+		struct folio *cur_folio;
 		/* Length of the compressed segment */
 		u32 seg_len;
 		u32 sector_bytes_left;
@@ -375,9 +376,9 @@ int lzo_decompress_bio(struct list_head *ws, struct compressed_bio *cb)
 		 */
 		ASSERT(cur_in / sectorsize ==
 		       (cur_in + LZO_LEN - 1) / sectorsize);
-		cur_page = cb->compressed_pages[cur_in / PAGE_SIZE];
-		ASSERT(cur_page);
-		kaddr = kmap_local_page(cur_page);
+		cur_folio = cb->compressed_folios[cur_in / PAGE_SIZE];
+		ASSERT(cur_folio);
+		kaddr = kmap_local_folio(cur_folio, 0);
 		seg_len = read_compress_length(kaddr + offset_in_page(cur_in));
 		kunmap_local(kaddr);
 		cur_in += LZO_LEN;
diff --git a/fs/btrfs/messages.c b/fs/btrfs/messages.c
index cdada48658..210d9c82e2 100644
--- a/fs/btrfs/messages.c
+++ b/fs/btrfs/messages.c
@@ -3,13 +3,11 @@
 #include "fs.h"
 #include "messages.h"
 #include "discard.h"
-#include "transaction.h"
-#include "space-info.h"
 #include "super.h"
 
 #ifdef CONFIG_PRINTK
 
-#define STATE_STRING_PREFACE	": state "
+#define STATE_STRING_PREFACE	" state "
 #define STATE_STRING_BUF_LEN	(sizeof(STATE_STRING_PREFACE) + BTRFS_FS_STATE_COUNT + 1)
 
 /*
diff --git a/fs/btrfs/misc.h b/fs/btrfs/misc.h
index 40f2d9f1a1..dde4904aea 100644
--- a/fs/btrfs/misc.h
+++ b/fs/btrfs/misc.h
@@ -3,6 +3,8 @@
 #ifndef BTRFS_MISC_H
 #define BTRFS_MISC_H
 
+#include <linux/types.h>
+#include <linux/bitmap.h>
 #include <linux/sched.h>
 #include <linux/wait.h>
 #include <linux/math64.h>
diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c
index 81f67ebf74..35a413ce93 100644
--- a/fs/btrfs/ordered-data.c
+++ b/fs/btrfs/ordered-data.c
@@ -19,7 +19,6 @@
 #include "qgroup.h"
 #include "subpage.h"
 #include "file.h"
-#include "super.h"
 
 static struct kmem_cache *btrfs_ordered_extent_cache;
 
@@ -295,6 +294,12 @@ void btrfs_add_ordered_sum(struct btrfs_ordered_extent *entry,
 	spin_unlock_irq(&inode->ordered_tree_lock);
 }
 
+void btrfs_mark_ordered_extent_error(struct btrfs_ordered_extent *ordered)
+{
+	if (!test_and_set_bit(BTRFS_ORDERED_IOERR, &ordered->flags))
+		mapping_set_error(ordered->inode->i_mapping, -EIO);
+}
+
 static void finish_ordered_fn(struct btrfs_work *work)
 {
 	struct btrfs_ordered_extent *ordered_extent;
@@ -333,7 +338,7 @@ static bool can_finish_ordered_extent(struct btrfs_ordered_extent *ordered,
 	if (WARN_ON_ONCE(len > ordered->bytes_left)) {
 		btrfs_crit(fs_info,
 "bad ordered extent accounting, root=%llu ino=%llu OE offset=%llu OE len=%llu to_dec=%llu left=%llu",
-			   inode->root->root_key.objectid, btrfs_ino(inode),
+			   btrfs_root_id(inode->root), btrfs_ino(inode),
 			   ordered->file_offset, ordered->num_bytes,
 			   len, ordered->bytes_left);
 		ordered->bytes_left = 0;
@@ -383,6 +388,37 @@ bool btrfs_finish_ordered_extent(struct btrfs_ordered_extent *ordered,
 	ret = can_finish_ordered_extent(ordered, page, file_offset, len, uptodate);
 	spin_unlock_irqrestore(&inode->ordered_tree_lock, flags);
 
+	/*
+	 * If this is a COW write it means we created new extent maps for the
+	 * range and they point to unwritten locations if we got an error either
+	 * before submitting a bio or during IO.
+	 *
+	 * We have marked the ordered extent with BTRFS_ORDERED_IOERR, and we
+	 * are queuing its completion below. During completion, at
+	 * btrfs_finish_one_ordered(), we will drop the extent maps for the
+	 * unwritten extents.
+	 *
+	 * However because completion runs in a work queue we can end up having
+	 * a fast fsync running before that. In the case of direct IO, once we
+	 * unlock the inode the fsync might start, and we queue the completion
+	 * before unlocking the inode. In the case of buffered IO when writeback
+	 * finishes (end_bbio_data_write()) we queue the completion, so if the
+	 * writeback was triggered by a fast fsync, the fsync might start
+	 * logging before ordered extent completion runs in the work queue.
+	 *
+	 * The fast fsync will log file extent items based on the extent maps it
+	 * finds, so if by the time it collects extent maps the ordered extent
+	 * completion didn't happen yet, it will log file extent items that
+	 * point to unwritten extents, resulting in a corruption if a crash
+	 * happens and the log tree is replayed. Note that a fast fsync does not
+	 * wait for completion of ordered extents in order to reduce latency.
+	 *
+	 * Set a flag in the inode so that the next fast fsync will wait for
+	 * ordered extents to complete before starting to log.
+	 */
+	if (!uptodate && !test_bit(BTRFS_ORDERED_NOCOW, &ordered->flags))
+		set_bit(BTRFS_INODE_COW_WRITE_ERROR, &inode->runtime_flags);
+
 	if (ret)
 		btrfs_queue_ordered_fn(ordered);
 	return ret;
@@ -1237,10 +1273,7 @@ struct btrfs_ordered_extent *btrfs_split_ordered_extent(
 
 int __init ordered_data_init(void)
 {
-	btrfs_ordered_extent_cache = kmem_cache_create("btrfs_ordered_extent",
-				     sizeof(struct btrfs_ordered_extent), 0,
-				     SLAB_MEM_SPREAD,
-				     NULL);
+	btrfs_ordered_extent_cache = KMEM_CACHE(btrfs_ordered_extent, 0);
 	if (!btrfs_ordered_extent_cache)
 		return -ENOMEM;
 
diff --git a/fs/btrfs/ordered-data.h b/fs/btrfs/ordered-data.h
index 127ef8bf0f..b6f6c6b917 100644
--- a/fs/btrfs/ordered-data.h
+++ b/fs/btrfs/ordered-data.h
@@ -6,6 +6,21 @@
 #ifndef BTRFS_ORDERED_DATA_H
 #define BTRFS_ORDERED_DATA_H
 
+#include <linux/types.h>
+#include <linux/list.h>
+#include <linux/refcount.h>
+#include <linux/completion.h>
+#include <linux/rbtree.h>
+#include <linux/wait.h>
+#include "async-thread.h"
+
+struct inode;
+struct page;
+struct extent_state;
+struct btrfs_inode;
+struct btrfs_root;
+struct btrfs_fs_info;
+
 struct btrfs_ordered_sum {
 	/*
 	 * Logical start address and length for of the blocks covered by
@@ -188,6 +203,7 @@ bool btrfs_try_lock_ordered_range(struct btrfs_inode *inode, u64 start, u64 end,
 				  struct extent_state **cached_state);
 struct btrfs_ordered_extent *btrfs_split_ordered_extent(
 			struct btrfs_ordered_extent *ordered, u64 len);
+void btrfs_mark_ordered_extent_error(struct btrfs_ordered_extent *ordered);
 int __init ordered_data_init(void);
 void __cold ordered_data_exit(void);
 
diff --git a/fs/btrfs/orphan.c b/fs/btrfs/orphan.c
index 7a1b021b56..6195a2215b 100644
--- a/fs/btrfs/orphan.c
+++ b/fs/btrfs/orphan.c
@@ -4,7 +4,6 @@
  */
 
 #include "ctree.h"
-#include "disk-io.h"
 #include "orphan.h"
 
 int btrfs_insert_orphan_item(struct btrfs_trans_handle *trans,
diff --git a/fs/btrfs/orphan.h b/fs/btrfs/orphan.h
index 3faab5cbb5..aa54a88a60 100644
--- a/fs/btrfs/orphan.h
+++ b/fs/btrfs/orphan.h
@@ -3,6 +3,11 @@
 #ifndef BTRFS_ORPHAN_H
 #define BTRFS_ORPHAN_H
 
+#include <linux/types.h>
+
+struct btrfs_trans_handle;
+struct btrfs_root;
+
 int btrfs_insert_orphan_item(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *root, u64 offset);
 int btrfs_del_orphan_item(struct btrfs_trans_handle *trans,
diff --git a/fs/btrfs/print-tree.c b/fs/btrfs/print-tree.c
index 7e46aa8a04..b876156fcd 100644
--- a/fs/btrfs/print-tree.c
+++ b/fs/btrfs/print-tree.c
@@ -14,7 +14,7 @@
 
 struct root_name_map {
 	u64 id;
-	char name[16];
+	const char *name;
 };
 
 static const struct root_name_map root_map[] = {
diff --git a/fs/btrfs/print-tree.h b/fs/btrfs/print-tree.h
index c42bc666d5..8504bf1702 100644
--- a/fs/btrfs/print-tree.h
+++ b/fs/btrfs/print-tree.h
@@ -9,6 +9,9 @@
 /* Buffer size to contain tree name and possibly additional data (offset) */
 #define BTRFS_ROOT_NAME_BUF_LEN				48
 
+struct extent_buffer;
+struct btrfs_key;
+
 void btrfs_print_leaf(const struct extent_buffer *l);
 void btrfs_print_tree(const struct extent_buffer *c, bool follow);
 const char *btrfs_root_name(const struct btrfs_key *key, char *buf);
diff --git a/fs/btrfs/props.c b/fs/btrfs/props.c
index ac4a0af2b5..155570e20f 100644
--- a/fs/btrfs/props.c
+++ b/fs/btrfs/props.c
@@ -4,6 +4,7 @@
  */
 
 #include <linux/hashtable.h>
+#include <linux/xattr.h>
 #include "messages.h"
 #include "props.h"
 #include "btrfs_inode.h"
@@ -267,7 +268,7 @@ static void inode_prop_iterator(void *ctx,
 		btrfs_warn(root->fs_info,
 			   "error applying prop %s to ino %llu (root %llu): %d",
 			   handler->xattr_name, btrfs_ino(BTRFS_I(inode)),
-			   root->root_key.objectid, ret);
+			   btrfs_root_id(root), ret);
 	else
 		set_bit(BTRFS_INODE_HAS_PROPS, &BTRFS_I(inode)->runtime_flags);
 }
diff --git a/fs/btrfs/props.h b/fs/btrfs/props.h
index 6e283196e3..f60cd89feb 100644
--- a/fs/btrfs/props.h
+++ b/fs/btrfs/props.h
@@ -6,7 +6,12 @@
 #ifndef BTRFS_PROPS_H
 #define BTRFS_PROPS_H
 
-#include "ctree.h"
+#include <linux/compiler_types.h>
+
+struct inode;
+struct btrfs_inode;
+struct btrfs_path;
+struct btrfs_trans_handle;
 
 int __init btrfs_props_init(void);
 
diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c
index cacc12d0ff..29d6ca3b87 100644
--- a/fs/btrfs/qgroup.c
+++ b/fs/btrfs/qgroup.c
@@ -468,6 +468,7 @@ int btrfs_read_qgroup_config(struct btrfs_fs_info *fs_info)
 		}
 		if (!qgroup) {
 			struct btrfs_qgroup *prealloc;
+			struct btrfs_root *tree_root = fs_info->tree_root;
 
 			prealloc = kzalloc(sizeof(*prealloc), GFP_KERNEL);
 			if (!prealloc) {
@@ -475,6 +476,25 @@ int btrfs_read_qgroup_config(struct btrfs_fs_info *fs_info)
 				goto out;
 			}
 			qgroup = add_qgroup_rb(fs_info, prealloc, found_key.offset);
+			/*
+			 * If a qgroup exists for a subvolume ID, it is possible
+			 * that subvolume has been deleted, in which case
+			 * re-using that ID would lead to incorrect accounting.
+			 *
+			 * Ensure that we skip any such subvol ids.
+			 *
+			 * We don't need to lock because this is only called
+			 * during mount before we start doing things like creating
+			 * subvolumes.
+			 */
+			if (is_fstree(qgroup->qgroupid) &&
+			    qgroup->qgroupid > tree_root->free_objectid)
+				/*
+				 * Don't need to check against BTRFS_LAST_FREE_OBJECTID,
+				 * as it will get checked on the next call to
+				 * btrfs_get_free_objectid.
+				 */
+				tree_root->free_objectid = qgroup->qgroupid + 1;
 		}
 		ret = btrfs_sysfs_add_one_qgroup(fs_info, qgroup);
 		if (ret < 0)
@@ -1324,14 +1344,14 @@ static int flush_reservations(struct btrfs_fs_info *fs_info)
 	trans = btrfs_join_transaction(fs_info->tree_root);
 	if (IS_ERR(trans))
 		return PTR_ERR(trans);
-	btrfs_commit_transaction(trans);
+	ret = btrfs_commit_transaction(trans);
 
 	return ret;
 }
 
 int btrfs_quota_disable(struct btrfs_fs_info *fs_info)
 {
-	struct btrfs_root *quota_root;
+	struct btrfs_root *quota_root = NULL;
 	struct btrfs_trans_handle *trans = NULL;
 	int ret = 0;
 
@@ -1426,12 +1446,14 @@ int btrfs_quota_disable(struct btrfs_fs_info *fs_info)
 	btrfs_tree_lock(quota_root->node);
 	btrfs_clear_buffer_dirty(trans, quota_root->node);
 	btrfs_tree_unlock(quota_root->node);
-	btrfs_free_tree_block(trans, btrfs_root_id(quota_root),
-			      quota_root->node, 0, 1);
+	ret = btrfs_free_tree_block(trans, btrfs_root_id(quota_root),
+				    quota_root->node, 0, 1);
 
-	btrfs_put_root(quota_root);
+	if (ret < 0)
+		btrfs_abort_transaction(trans, ret);
 
 out:
+	btrfs_put_root(quota_root);
 	mutex_unlock(&fs_info->qgroup_ioctl_lock);
 	if (ret && trans)
 		btrfs_end_transaction(trans);
@@ -1536,18 +1558,15 @@ static int quick_update_accounting(struct btrfs_fs_info *fs_info,
 {
 	struct btrfs_qgroup *qgroup;
 	int ret = 1;
-	int err = 0;
 
 	qgroup = find_qgroup_rb(fs_info, src);
 	if (!qgroup)
 		goto out;
 	if (qgroup->excl == qgroup->rfer) {
-		ret = 0;
-		err = __qgroup_excl_accounting(fs_info, dst, qgroup, sign);
-		if (err < 0) {
-			ret = err;
+		ret = __qgroup_excl_accounting(fs_info, dst, qgroup, sign);
+		if (ret < 0)
 			goto out;
-		}
+		ret = 0;
 	}
 out:
 	if (ret)
@@ -2500,8 +2519,8 @@ int btrfs_qgroup_trace_subtree(struct btrfs_trans_handle *trans,
 	struct extent_buffer *eb = root_eb;
 	struct btrfs_path *path = NULL;
 
-	BUG_ON(root_level < 0 || root_level >= BTRFS_MAX_LEVEL);
-	BUG_ON(root_eb == NULL);
+	ASSERT(0 <= root_level && root_level < BTRFS_MAX_LEVEL);
+	ASSERT(root_eb != NULL);
 
 	if (!btrfs_qgroup_full_accounting(fs_info))
 		return 0;
@@ -2856,8 +2875,6 @@ int btrfs_qgroup_account_extent(struct btrfs_trans_handle *trans, u64 bytenr,
 	if (nr_old_roots == 0 && nr_new_roots == 0)
 		goto out_free;
 
-	BUG_ON(!fs_info->quota_root);
-
 	trace_btrfs_qgroup_account_extent(fs_info, trans->transid, bytenr,
 					num_bytes, nr_old_roots, nr_new_roots);
 
@@ -3047,8 +3064,6 @@ int btrfs_qgroup_check_inherit(struct btrfs_fs_info *fs_info,
 			       struct btrfs_qgroup_inherit *inherit,
 			       size_t size)
 {
-	if (!btrfs_qgroup_enabled(fs_info))
-		return 0;
 	if (inherit->flags & ~BTRFS_QGROUP_INHERIT_FLAGS_SUPP)
 		return -EOPNOTSUPP;
 	if (size < sizeof(*inherit) || size > PAGE_SIZE)
@@ -3066,13 +3081,18 @@ int btrfs_qgroup_check_inherit(struct btrfs_fs_info *fs_info,
 	if (inherit->num_ref_copies > 0 || inherit->num_excl_copies > 0)
 		return -EINVAL;
 
-	if (inherit->num_qgroups > PAGE_SIZE)
-		return -EINVAL;
-
 	if (size != struct_size(inherit, qgroups, inherit->num_qgroups))
 		return -EINVAL;
 
 	/*
+	 * Skip the inherit source qgroups check if qgroup is not enabled.
+	 * Qgroup can still be later enabled causing problems, but in that case
+	 * btrfs_qgroup_inherit() would just ignore those invalid ones.
+	 */
+	if (!btrfs_qgroup_enabled(fs_info))
+		return 0;
+
+	/*
 	 * Now check all the remaining qgroups, they should all:
 	 *
 	 * - Exist
@@ -3131,13 +3151,69 @@ static int qgroup_auto_inherit(struct btrfs_fs_info *fs_info,
 	qgids = res->qgroups;
 
 	list_for_each_entry(qg_list, &inode_qg->groups, next_group)
-		qgids[i] = qg_list->group->qgroupid;
+		qgids[i++] = qg_list->group->qgroupid;
 
 	*inherit = res;
 	return 0;
 }
 
 /*
+ * Check if we can skip rescan when inheriting qgroups.  If @src has a single
+ * @parent, and that @parent is owning all its bytes exclusively, we can skip
+ * the full rescan, by just adding nodesize to the @parent's excl/rfer.
+ *
+ * Return <0 for fatal errors (like srcid/parentid has no qgroup).
+ * Return 0 if a quick inherit is done.
+ * Return >0 if a quick inherit is not possible, and a full rescan is needed.
+ */
+static int qgroup_snapshot_quick_inherit(struct btrfs_fs_info *fs_info,
+					 u64 srcid, u64 parentid)
+{
+	struct btrfs_qgroup *src;
+	struct btrfs_qgroup *parent;
+	struct btrfs_qgroup_list *list;
+	int nr_parents = 0;
+
+	src = find_qgroup_rb(fs_info, srcid);
+	if (!src)
+		return -ENOENT;
+	parent = find_qgroup_rb(fs_info, parentid);
+	if (!parent)
+		return -ENOENT;
+
+	/*
+	 * Source has no parent qgroup, but our new qgroup would have one.
+	 * Qgroup numbers would become inconsistent.
+	 */
+	if (list_empty(&src->groups))
+		return 1;
+
+	list_for_each_entry(list, &src->groups, next_group) {
+		/* The parent is not the same, quick update is not possible. */
+		if (list->group->qgroupid != parentid)
+			return 1;
+		nr_parents++;
+		/*
+		 * More than one parent qgroup, we can't be sure about accounting
+		 * consistency.
+		 */
+		if (nr_parents > 1)
+			return 1;
+	}
+
+	/*
+	 * The parent is not exclusively owning all its bytes.  We're not sure
+	 * if the source has any bytes not fully owned by the parent.
+	 */
+	if (parent->excl != parent->rfer)
+		return 1;
+
+	parent->excl += fs_info->nodesize;
+	parent->rfer += fs_info->nodesize;
+	return 0;
+}
+
+/*
  * Copy the accounting information between qgroups. This is necessary
  * when a snapshot or a subvolume is created. Throwing an error will
  * cause a transaction abort so we take extra care here to only error
@@ -3305,6 +3381,13 @@ int btrfs_qgroup_inherit(struct btrfs_trans_handle *trans, u64 srcid,
 
 		qgroup_dirty(fs_info, dstgroup);
 		qgroup_dirty(fs_info, srcgroup);
+
+		/*
+		 * If the source qgroup has parent but the new one doesn't,
+		 * we need a full rescan.
+		 */
+		if (!inherit && !list_empty(&srcgroup->groups))
+			need_rescan = true;
 	}
 
 	if (!inherit)
@@ -3319,14 +3402,16 @@ int btrfs_qgroup_inherit(struct btrfs_trans_handle *trans, u64 srcid,
 			if (ret)
 				goto unlock;
 		}
+		if (srcid) {
+			/* Check if we can do a quick inherit. */
+			ret = qgroup_snapshot_quick_inherit(fs_info, srcid, *i_qgroups);
+			if (ret < 0)
+				goto unlock;
+			if (ret > 0)
+				need_rescan = true;
+			ret = 0;
+		}
 		++i_qgroups;
-
-		/*
-		 * If we're doing a snapshot, and adding the snapshot to a new
-		 * qgroup, the numbers are guaranteed to be incorrect.
-		 */
-		if (srcid)
-			need_rescan = true;
 	}
 
 	for (i = 0; i <  inherit->num_ref_copies; ++i, i_qgroups += 2) {
@@ -3408,7 +3493,7 @@ static int qgroup_reserve(struct btrfs_root *root, u64 num_bytes, bool enforce,
 {
 	struct btrfs_qgroup *qgroup;
 	struct btrfs_fs_info *fs_info = root->fs_info;
-	u64 ref_root = root->root_key.objectid;
+	u64 ref_root = btrfs_root_id(root);
 	int ret = 0;
 	LIST_HEAD(qgroup_list);
 
@@ -3643,7 +3728,6 @@ static void btrfs_qgroup_rescan_worker(struct btrfs_work *work)
 						     qgroup_rescan_work);
 	struct btrfs_path *path;
 	struct btrfs_trans_handle *trans = NULL;
-	int err = -ENOMEM;
 	int ret = 0;
 	bool stopped = false;
 	bool did_leaf_rescans = false;
@@ -3652,8 +3736,10 @@ static void btrfs_qgroup_rescan_worker(struct btrfs_work *work)
 		return;
 
 	path = btrfs_alloc_path();
-	if (!path)
+	if (!path) {
+		ret = -ENOMEM;
 		goto out;
+	}
 	/*
 	 * Rescan should only search for commit root, and any later difference
 	 * should be recorded by qgroup
@@ -3661,18 +3747,17 @@ static void btrfs_qgroup_rescan_worker(struct btrfs_work *work)
 	path->search_commit_root = 1;
 	path->skip_locking = 1;
 
-	err = 0;
-	while (!err && !(stopped = rescan_should_stop(fs_info))) {
+	while (!ret && !(stopped = rescan_should_stop(fs_info))) {
 		trans = btrfs_start_transaction(fs_info->fs_root, 0);
 		if (IS_ERR(trans)) {
-			err = PTR_ERR(trans);
+			ret = PTR_ERR(trans);
 			break;
 		}
 
-		err = qgroup_rescan_leaf(trans, path);
+		ret = qgroup_rescan_leaf(trans, path);
 		did_leaf_rescans = true;
 
-		if (err > 0)
+		if (ret > 0)
 			btrfs_commit_transaction(trans);
 		else
 			btrfs_end_transaction(trans);
@@ -3682,10 +3767,10 @@ out:
 	btrfs_free_path(path);
 
 	mutex_lock(&fs_info->qgroup_rescan_lock);
-	if (err > 0 &&
+	if (ret > 0 &&
 	    fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT) {
 		fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT;
-	} else if (err < 0 || stopped) {
+	} else if (ret < 0 || stopped) {
 		fs_info->qgroup_flags |= BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT;
 	}
 	mutex_unlock(&fs_info->qgroup_rescan_lock);
@@ -3700,11 +3785,11 @@ out:
 	if (did_leaf_rescans) {
 		trans = btrfs_start_transaction(fs_info->quota_root, 1);
 		if (IS_ERR(trans)) {
-			err = PTR_ERR(trans);
+			ret = PTR_ERR(trans);
 			trans = NULL;
 			btrfs_err(fs_info,
 				  "fail to start transaction for status update: %d",
-				  err);
+				  ret);
 		}
 	} else {
 		trans = NULL;
@@ -3715,11 +3800,11 @@ out:
 	    fs_info->qgroup_flags & BTRFS_QGROUP_RUNTIME_FLAG_CANCEL_RESCAN)
 		fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_RESCAN;
 	if (trans) {
-		ret = update_qgroup_status_item(trans);
-		if (ret < 0) {
-			err = ret;
-			btrfs_err(fs_info, "fail to update qgroup status: %d",
-				  err);
+		int ret2 = update_qgroup_status_item(trans);
+
+		if (ret2 < 0) {
+			ret = ret2;
+			btrfs_err(fs_info, "fail to update qgroup status: %d", ret);
 		}
 	}
 	fs_info->qgroup_rescan_running = false;
@@ -3736,11 +3821,11 @@ out:
 		btrfs_info(fs_info, "qgroup scan paused");
 	} else if (fs_info->qgroup_flags & BTRFS_QGROUP_RUNTIME_FLAG_CANCEL_RESCAN) {
 		btrfs_info(fs_info, "qgroup scan cancelled");
-	} else if (err >= 0) {
+	} else if (ret >= 0) {
 		btrfs_info(fs_info, "qgroup scan completed%s",
-			err > 0 ? " (inconsistency flag cleared)" : "");
+			ret > 0 ? " (inconsistency flag cleared)" : "");
 	} else {
-		btrfs_err(fs_info, "qgroup scan failed with %d", err);
+		btrfs_err(fs_info, "qgroup scan failed with %d", ret);
 	}
 }
 
@@ -3763,14 +3848,14 @@ qgroup_rescan_init(struct btrfs_fs_info *fs_info, u64 progress_objectid,
 		/* we're resuming qgroup rescan at mount time */
 		if (!(fs_info->qgroup_flags &
 		      BTRFS_QGROUP_STATUS_FLAG_RESCAN)) {
-			btrfs_warn(fs_info,
+			btrfs_debug(fs_info,
 			"qgroup rescan init failed, qgroup rescan is not queued");
 			ret = -EINVAL;
 		} else if (!(fs_info->qgroup_flags &
 			     BTRFS_QGROUP_STATUS_FLAG_ON)) {
-			btrfs_warn(fs_info,
+			btrfs_debug(fs_info,
 			"qgroup rescan init failed, qgroup is not enabled");
-			ret = -EINVAL;
+			ret = -ENOTCONN;
 		}
 
 		if (ret)
@@ -3781,14 +3866,12 @@ qgroup_rescan_init(struct btrfs_fs_info *fs_info, u64 progress_objectid,
 
 	if (init_flags) {
 		if (fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN) {
-			btrfs_warn(fs_info,
-				   "qgroup rescan is already in progress");
 			ret = -EINPROGRESS;
 		} else if (!(fs_info->qgroup_flags &
 			     BTRFS_QGROUP_STATUS_FLAG_ON)) {
-			btrfs_warn(fs_info,
+			btrfs_debug(fs_info,
 			"qgroup rescan init failed, qgroup is not enabled");
-			ret = -EINVAL;
+			ret = -ENOTCONN;
 		} else if (btrfs_qgroup_mode(fs_info) == BTRFS_QGROUP_MODE_DISABLED) {
 			/* Quota disable is in progress */
 			ret = -EBUSY;
@@ -4049,7 +4132,7 @@ static int qgroup_reserve_data(struct btrfs_inode *inode,
 	int ret;
 
 	if (btrfs_qgroup_mode(root->fs_info) == BTRFS_QGROUP_MODE_DISABLED ||
-	    !is_fstree(root->root_key.objectid) || len == 0)
+	    !is_fstree(btrfs_root_id(root)) || len == 0)
 		return 0;
 
 	/* @reserved parameter is mandatory for qgroup */
@@ -4165,7 +4248,7 @@ static int qgroup_free_reserved_data(struct btrfs_inode *inode,
 			goto out;
 		freed += changeset.bytes_changed;
 	}
-	btrfs_qgroup_free_refroot(root->fs_info, root->root_key.objectid, freed,
+	btrfs_qgroup_free_refroot(root->fs_info, btrfs_root_id(root), freed,
 				  BTRFS_QGROUP_RSV_DATA);
 	if (freed_ret)
 		*freed_ret = freed;
@@ -4206,7 +4289,7 @@ static int __btrfs_qgroup_release_data(struct btrfs_inode *inode,
 					changeset.bytes_changed, trace_op);
 	if (free)
 		btrfs_qgroup_free_refroot(inode->root->fs_info,
-				inode->root->root_key.objectid,
+				btrfs_root_id(inode->root),
 				changeset.bytes_changed, BTRFS_QGROUP_RSV_DATA);
 	if (released)
 		*released = changeset.bytes_changed;
@@ -4301,7 +4384,7 @@ int btrfs_qgroup_reserve_meta(struct btrfs_root *root, int num_bytes,
 	int ret;
 
 	if (btrfs_qgroup_mode(fs_info) == BTRFS_QGROUP_MODE_DISABLED ||
-	    !is_fstree(root->root_key.objectid) || num_bytes == 0)
+	    !is_fstree(btrfs_root_id(root)) || num_bytes == 0)
 		return 0;
 
 	BUG_ON(num_bytes != round_down(num_bytes, fs_info->nodesize));
@@ -4346,13 +4429,13 @@ void btrfs_qgroup_free_meta_all_pertrans(struct btrfs_root *root)
 	struct btrfs_fs_info *fs_info = root->fs_info;
 
 	if (btrfs_qgroup_mode(fs_info) == BTRFS_QGROUP_MODE_DISABLED ||
-	    !is_fstree(root->root_key.objectid))
+	    !is_fstree(btrfs_root_id(root)))
 		return;
 
 	/* TODO: Update trace point to handle such free */
 	trace_qgroup_meta_free_all_pertrans(root);
 	/* Special value -1 means to free all reserved space */
-	btrfs_qgroup_free_refroot(fs_info, root->root_key.objectid, (u64)-1,
+	btrfs_qgroup_free_refroot(fs_info, btrfs_root_id(root), (u64)-1,
 				  BTRFS_QGROUP_RSV_META_PERTRANS);
 }
 
@@ -4362,7 +4445,7 @@ void __btrfs_qgroup_free_meta(struct btrfs_root *root, int num_bytes,
 	struct btrfs_fs_info *fs_info = root->fs_info;
 
 	if (btrfs_qgroup_mode(fs_info) == BTRFS_QGROUP_MODE_DISABLED ||
-	    !is_fstree(root->root_key.objectid))
+	    !is_fstree(btrfs_root_id(root)))
 		return;
 
 	/*
@@ -4373,8 +4456,7 @@ void __btrfs_qgroup_free_meta(struct btrfs_root *root, int num_bytes,
 	num_bytes = sub_root_meta_rsv(root, num_bytes, type);
 	BUG_ON(num_bytes != round_down(num_bytes, fs_info->nodesize));
 	trace_qgroup_meta_reserve(root, -(s64)num_bytes, type);
-	btrfs_qgroup_free_refroot(fs_info, root->root_key.objectid,
-				  num_bytes, type);
+	btrfs_qgroup_free_refroot(fs_info, btrfs_root_id(root), num_bytes, type);
 }
 
 static void qgroup_convert_meta(struct btrfs_fs_info *fs_info, u64 ref_root,
@@ -4422,13 +4504,13 @@ void btrfs_qgroup_convert_reserved_meta(struct btrfs_root *root, int num_bytes)
 	struct btrfs_fs_info *fs_info = root->fs_info;
 
 	if (btrfs_qgroup_mode(fs_info) == BTRFS_QGROUP_MODE_DISABLED ||
-	    !is_fstree(root->root_key.objectid))
+	    !is_fstree(btrfs_root_id(root)))
 		return;
 	/* Same as btrfs_qgroup_free_meta_prealloc() */
 	num_bytes = sub_root_meta_rsv(root, num_bytes,
 				      BTRFS_QGROUP_RSV_META_PREALLOC);
 	trace_qgroup_meta_convert(root, num_bytes);
-	qgroup_convert_meta(fs_info, root->root_key.objectid, num_bytes);
+	qgroup_convert_meta(fs_info, btrfs_root_id(root), num_bytes);
 	if (!sb_rdonly(fs_info->sb))
 		add_root_meta_rsv(root, num_bytes, BTRFS_QGROUP_RSV_META_PERTRANS);
 }
@@ -4457,7 +4539,7 @@ void btrfs_qgroup_check_reserved_leak(struct btrfs_inode *inode)
 				btrfs_ino(inode), unode->val, unode->aux);
 		}
 		btrfs_qgroup_free_refroot(inode->root->fs_info,
-				inode->root->root_key.objectid,
+				btrfs_root_id(inode->root),
 				changeset.bytes_changed, BTRFS_QGROUP_RSV_DATA);
 
 	}
@@ -4643,7 +4725,7 @@ int btrfs_qgroup_trace_subtree_after_cow(struct btrfs_trans_handle *trans,
 
 	if (!btrfs_qgroup_full_accounting(fs_info))
 		return 0;
-	if (!is_fstree(root->root_key.objectid) || !root->reloc_root)
+	if (!is_fstree(btrfs_root_id(root)) || !root->reloc_root)
 		return 0;
 
 	spin_lock(&blocks->lock);
diff --git a/fs/btrfs/qgroup.h b/fs/btrfs/qgroup.h
index 45a7d8920d..706640be0e 100644
--- a/fs/btrfs/qgroup.h
+++ b/fs/btrfs/qgroup.h
@@ -6,12 +6,22 @@
 #ifndef BTRFS_QGROUP_H
 #define BTRFS_QGROUP_H
 
+#include <linux/types.h>
 #include <linux/spinlock.h>
 #include <linux/rbtree.h>
 #include <linux/kobject.h>
-#include "ulist.h"
-#include "delayed-ref.h"
-#include "misc.h"
+#include <linux/list.h>
+#include <uapi/linux/btrfs_tree.h>
+
+struct extent_buffer;
+struct extent_changeset;
+struct btrfs_delayed_extent_op;
+struct btrfs_fs_info;
+struct btrfs_root;
+struct btrfs_ioctl_quota_ctl_args;
+struct btrfs_trans_handle;
+struct btrfs_delayed_ref_root;
+struct btrfs_inode;
 
 /*
  * Btrfs qgroup overview
@@ -321,7 +331,6 @@ int btrfs_limit_qgroup(struct btrfs_trans_handle *trans, u64 qgroupid,
 		       struct btrfs_qgroup_limit *limit);
 int btrfs_read_qgroup_config(struct btrfs_fs_info *fs_info);
 void btrfs_free_qgroup_config(struct btrfs_fs_info *fs_info);
-struct btrfs_delayed_extent_op;
 
 int btrfs_qgroup_trace_extent_nolock(
 		struct btrfs_fs_info *fs_info,
diff --git a/fs/btrfs/raid-stripe-tree.c b/fs/btrfs/raid-stripe-tree.c
index 9589362acf..6af6b4b9a3 100644
--- a/fs/btrfs/raid-stripe-tree.c
+++ b/fs/btrfs/raid-stripe-tree.c
@@ -11,7 +11,6 @@
 #include "disk-io.h"
 #include "raid-stripe-tree.h"
 #include "volumes.h"
-#include "misc.h"
 #include "print-tree.h"
 
 int btrfs_delete_raid_extent(struct btrfs_trans_handle *trans, u64 start, u64 length)
diff --git a/fs/btrfs/raid-stripe-tree.h b/fs/btrfs/raid-stripe-tree.h
index cdb58b38fc..c9c258f849 100644
--- a/fs/btrfs/raid-stripe-tree.h
+++ b/fs/btrfs/raid-stripe-tree.h
@@ -6,6 +6,10 @@
 #ifndef BTRFS_RAID_STRIPE_TREE_H
 #define BTRFS_RAID_STRIPE_TREE_H
 
+#include <linux/types.h>
+#include <uapi/linux/btrfs_tree.h>
+#include "fs.h"
+
 #define BTRFS_RST_SUPP_BLOCK_GROUP_MASK    (BTRFS_BLOCK_GROUP_DUP |		\
 					    BTRFS_BLOCK_GROUP_RAID1_MASK |	\
 					    BTRFS_BLOCK_GROUP_RAID0 |		\
@@ -13,6 +17,7 @@
 
 struct btrfs_io_context;
 struct btrfs_io_stripe;
+struct btrfs_fs_info;
 struct btrfs_ordered_extent;
 struct btrfs_trans_handle;
 
diff --git a/fs/btrfs/raid56.c b/fs/btrfs/raid56.c
index 792c8e17c3..831fac45e7 100644
--- a/fs/btrfs/raid56.c
+++ b/fs/btrfs/raid56.c
@@ -14,7 +14,6 @@
 #include <linux/raid/xor.h>
 #include <linux/mm.h>
 #include "messages.h"
-#include "misc.h"
 #include "ctree.h"
 #include "disk-io.h"
 #include "volumes.h"
@@ -332,12 +331,11 @@ static void steal_rbio(struct btrfs_raid_bio *src, struct btrfs_raid_bio *dest)
 static void merge_rbio(struct btrfs_raid_bio *dest,
 		       struct btrfs_raid_bio *victim)
 {
-	bio_list_merge(&dest->bio_list, &victim->bio_list);
+	bio_list_merge_init(&dest->bio_list, &victim->bio_list);
 	dest->bio_list_bytes += victim->bio_list_bytes;
 	/* Also inherit the bitmaps from @victim. */
 	bitmap_or(&dest->dbitmap, &victim->dbitmap, &dest->dbitmap,
 		  dest->stripe_nsectors);
-	bio_list_init(&victim->bio_list);
 }
 
 /*
@@ -918,6 +916,13 @@ static struct btrfs_raid_bio *alloc_rbio(struct btrfs_fs_info *fs_info,
 	 */
 	ASSERT(stripe_nsectors <= BITS_PER_LONG);
 
+	/*
+	 * Real stripes must be between 2 (2 disks RAID5, aka RAID1) and 256
+	 * (limited by u8).
+	 */
+	ASSERT(real_stripes >= 2);
+	ASSERT(real_stripes <= U8_MAX);
+
 	rbio = kzalloc(sizeof(*rbio), GFP_NOFS);
 	if (!rbio)
 		return ERR_PTR(-ENOMEM);
@@ -955,6 +960,7 @@ static struct btrfs_raid_bio *alloc_rbio(struct btrfs_fs_info *fs_info,
 
 	ASSERT(btrfs_nr_parity_stripes(bioc->map_type));
 	rbio->nr_data = real_stripes - btrfs_nr_parity_stripes(bioc->map_type);
+	ASSERT(rbio->nr_data > 0);
 
 	return rbio;
 }
@@ -1181,6 +1187,26 @@ static inline void bio_list_put(struct bio_list *bio_list)
 		bio_put(bio);
 }
 
+static void assert_rbio(struct btrfs_raid_bio *rbio)
+{
+	if (!IS_ENABLED(CONFIG_BTRFS_DEBUG) ||
+	    !IS_ENABLED(CONFIG_BTRFS_ASSERT))
+		return;
+
+	/*
+	 * At least two stripes (2 disks RAID5), and since real_stripes is U8,
+	 * we won't go beyond 256 disks anyway.
+	 */
+	ASSERT(rbio->real_stripes >= 2);
+	ASSERT(rbio->nr_data > 0);
+
+	/*
+	 * This is another check to make sure nr data stripes is smaller
+	 * than total stripes.
+	 */
+	ASSERT(rbio->nr_data < rbio->real_stripes);
+}
+
 /* Generate PQ for one vertical stripe. */
 static void generate_pq_vertical(struct btrfs_raid_bio *rbio, int sectornr)
 {
@@ -1212,6 +1238,7 @@ static void generate_pq_vertical(struct btrfs_raid_bio *rbio, int sectornr)
 		pointers[stripe++] = kmap_local_page(sector->page) +
 				     sector->pgoff;
 
+		assert_rbio(rbio);
 		raid6_call.gen_syndrome(rbio->real_stripes, sectorsize,
 					pointers);
 	} else {
@@ -2473,6 +2500,7 @@ static int finish_parity_scrub(struct btrfs_raid_bio *rbio)
 		}
 
 		if (has_qstripe) {
+			assert_rbio(rbio);
 			/* RAID6, call the library function to fill in our P/Q */
 			raid6_call.gen_syndrome(rbio->real_stripes, sectorsize,
 						pointers);
diff --git a/fs/btrfs/raid56.h b/fs/btrfs/raid56.h
index 4702136888..0d7b4c2fb6 100644
--- a/fs/btrfs/raid56.h
+++ b/fs/btrfs/raid56.h
@@ -7,9 +7,18 @@
 #ifndef BTRFS_RAID56_H
 #define BTRFS_RAID56_H
 
+#include <linux/types.h>
+#include <linux/list.h>
+#include <linux/spinlock.h>
+#include <linux/bio.h>
+#include <linux/refcount.h>
 #include <linux/workqueue.h>
 #include "volumes.h"
 
+struct page;
+struct sector_ptr;
+struct btrfs_fs_info;
+
 enum btrfs_rbio_ops {
 	BTRFS_RBIO_WRITE,
 	BTRFS_RBIO_READ_REBUILD,
diff --git a/fs/btrfs/rcu-string.h b/fs/btrfs/rcu-string.h
index 5c2b66d155..1c2d7cb1fe 100644
--- a/fs/btrfs/rcu-string.h
+++ b/fs/btrfs/rcu-string.h
@@ -6,6 +6,12 @@
 #ifndef BTRFS_RCU_STRING_H
 #define BTRFS_RCU_STRING_H
 
+#include <linux/types.h>
+#include <linux/string.h>
+#include <linux/slab.h>
+#include <linux/rcupdate.h>
+#include <linux/printk.h>
+
 struct rcu_string {
 	struct rcu_head rcu;
 	char str[];
diff --git a/fs/btrfs/ref-verify.c b/fs/btrfs/ref-verify.c
index 8c4fc98ca9..9522a8b79d 100644
--- a/fs/btrfs/ref-verify.c
+++ b/fs/btrfs/ref-verify.c
@@ -441,7 +441,8 @@ static int process_extent_item(struct btrfs_fs_info *fs_info,
 	u32 item_size = btrfs_item_size(leaf, slot);
 	unsigned long end, ptr;
 	u64 offset, flags, count;
-	int type, ret;
+	int type;
+	int ret = 0;
 
 	ei = btrfs_item_ptr(leaf, slot, struct btrfs_extent_item);
 	flags = btrfs_extent_flags(leaf, ei);
@@ -486,7 +487,11 @@ static int process_extent_item(struct btrfs_fs_info *fs_info,
 						  key->objectid, key->offset);
 			break;
 		case BTRFS_EXTENT_OWNER_REF_KEY:
-			WARN_ON(!btrfs_fs_incompat(fs_info, SIMPLE_QUOTA));
+			if (!btrfs_fs_incompat(fs_info, SIMPLE_QUOTA)) {
+				btrfs_err(fs_info,
+			  "found extent owner ref without simple quotas enabled");
+				ret = -EINVAL;
+			}
 			break;
 		default:
 			btrfs_err(fs_info, "invalid key type in iref");
@@ -673,7 +678,7 @@ int btrfs_ref_tree_mod(struct btrfs_fs_info *fs_info,
 	int ret = 0;
 	bool metadata;
 	u64 bytenr = generic_ref->bytenr;
-	u64 num_bytes = generic_ref->len;
+	u64 num_bytes = generic_ref->num_bytes;
 	u64 parent = generic_ref->parent;
 	u64 ref_root = 0;
 	u64 owner = 0;
@@ -684,11 +689,11 @@ int btrfs_ref_tree_mod(struct btrfs_fs_info *fs_info,
 
 	if (generic_ref->type == BTRFS_REF_METADATA) {
 		if (!parent)
-			ref_root = generic_ref->tree_ref.ref_root;
+			ref_root = generic_ref->ref_root;
 		owner = generic_ref->tree_ref.level;
 	} else if (!parent) {
-		ref_root = generic_ref->data_ref.ref_root;
-		owner = generic_ref->data_ref.ino;
+		ref_root = generic_ref->ref_root;
+		owner = generic_ref->data_ref.objectid;
 		offset = generic_ref->data_ref.offset;
 	}
 	metadata = owner < BTRFS_FIRST_FREE_OBJECTID;
diff --git a/fs/btrfs/ref-verify.h b/fs/btrfs/ref-verify.h
index 855de37719..3511e1a5c9 100644
--- a/fs/btrfs/ref-verify.h
+++ b/fs/btrfs/ref-verify.h
@@ -6,7 +6,16 @@
 #ifndef BTRFS_REF_VERIFY_H
 #define BTRFS_REF_VERIFY_H
 
+#include <linux/types.h>
+#include <linux/rbtree_types.h>
+
+struct btrfs_fs_info;
+struct btrfs_ref;
+
 #ifdef CONFIG_BTRFS_FS_REF_VERIFY
+
+#include <linux/spinlock.h>
+
 int btrfs_build_ref_tree(struct btrfs_fs_info *fs_info);
 void btrfs_free_ref_cache(struct btrfs_fs_info *fs_info);
 int btrfs_ref_tree_mod(struct btrfs_fs_info *fs_info,
diff --git a/fs/btrfs/reflink.c b/fs/btrfs/reflink.c
index 08d0fb46ce..d0a3fcecc4 100644
--- a/fs/btrfs/reflink.c
+++ b/fs/btrfs/reflink.c
@@ -616,35 +616,6 @@ out:
 	return ret;
 }
 
-static void btrfs_double_extent_unlock(struct inode *inode1, u64 loff1,
-				       struct inode *inode2, u64 loff2, u64 len)
-{
-	unlock_extent(&BTRFS_I(inode1)->io_tree, loff1, loff1 + len - 1, NULL);
-	unlock_extent(&BTRFS_I(inode2)->io_tree, loff2, loff2 + len - 1, NULL);
-}
-
-static void btrfs_double_extent_lock(struct inode *inode1, u64 loff1,
-				     struct inode *inode2, u64 loff2, u64 len)
-{
-	u64 range1_end = loff1 + len - 1;
-	u64 range2_end = loff2 + len - 1;
-
-	if (inode1 < inode2) {
-		swap(inode1, inode2);
-		swap(loff1, loff2);
-		swap(range1_end, range2_end);
-	} else if (inode1 == inode2 && loff2 < loff1) {
-		swap(loff1, loff2);
-		swap(range1_end, range2_end);
-	}
-
-	lock_extent(&BTRFS_I(inode1)->io_tree, loff1, range1_end, NULL);
-	lock_extent(&BTRFS_I(inode2)->io_tree, loff2, range2_end, NULL);
-
-	btrfs_assert_inode_range_clean(BTRFS_I(inode1), loff1, range1_end);
-	btrfs_assert_inode_range_clean(BTRFS_I(inode2), loff2, range2_end);
-}
-
 static void btrfs_double_mmap_lock(struct inode *inode1, struct inode *inode2)
 {
 	if (inode1 < inode2)
@@ -662,17 +633,21 @@ static void btrfs_double_mmap_unlock(struct inode *inode1, struct inode *inode2)
 static int btrfs_extent_same_range(struct inode *src, u64 loff, u64 len,
 				   struct inode *dst, u64 dst_loff)
 {
+	const u64 end = dst_loff + len - 1;
+	struct extent_state *cached_state = NULL;
 	struct btrfs_fs_info *fs_info = BTRFS_I(src)->root->fs_info;
 	const u64 bs = fs_info->sectorsize;
 	int ret;
 
 	/*
-	 * Lock destination range to serialize with concurrent readahead() and
-	 * source range to serialize with relocation.
+	 * Lock destination range to serialize with concurrent readahead(), and
+	 * we are safe from concurrency with relocation of source extents
+	 * because we have already locked the inode's i_mmap_lock in exclusive
+	 * mode.
 	 */
-	btrfs_double_extent_lock(src, loff, dst, dst_loff, len);
+	lock_extent(&BTRFS_I(dst)->io_tree, dst_loff, end, &cached_state);
 	ret = btrfs_clone(src, dst, loff, len, ALIGN(len, bs), dst_loff, 1);
-	btrfs_double_extent_unlock(src, loff, dst, dst_loff, len);
+	unlock_extent(&BTRFS_I(dst)->io_tree, dst_loff, end, &cached_state);
 
 	btrfs_btree_balance_dirty(fs_info);
 
@@ -690,7 +665,7 @@ static int btrfs_extent_same(struct inode *src, u64 loff, u64 olen,
 	if (root_dst->send_in_progress) {
 		btrfs_warn_rl(root_dst->fs_info,
 "cannot deduplicate to root %llu while send operations are using it (%d in progress)",
-			      root_dst->root_key.objectid,
+			      btrfs_root_id(root_dst),
 			      root_dst->send_in_progress);
 		spin_unlock(&root_dst->root_item_lock);
 		return -EAGAIN;
@@ -724,6 +699,7 @@ out:
 static noinline int btrfs_clone_files(struct file *file, struct file *file_src,
 					u64 off, u64 olen, u64 destoff)
 {
+	struct extent_state *cached_state = NULL;
 	struct inode *inode = file_inode(file);
 	struct inode *src = file_inode(file_src);
 	struct btrfs_fs_info *fs_info = inode_to_fs_info(inode);
@@ -731,6 +707,7 @@ static noinline int btrfs_clone_files(struct file *file, struct file *file_src,
 	int wb_ret;
 	u64 len = olen;
 	u64 bs = fs_info->sectorsize;
+	u64 end;
 
 	/*
 	 * VFS's generic_remap_file_range_prep() protects us from cloning the
@@ -763,12 +740,15 @@ static noinline int btrfs_clone_files(struct file *file, struct file *file_src,
 	}
 
 	/*
-	 * Lock destination range to serialize with concurrent readahead() and
-	 * source range to serialize with relocation.
+	 * Lock destination range to serialize with concurrent readahead(), and
+	 * we are safe from concurrency with relocation of source extents
+	 * because we have already locked the inode's i_mmap_lock in exclusive
+	 * mode.
 	 */
-	btrfs_double_extent_lock(src, off, inode, destoff, len);
+	end = destoff + len - 1;
+	lock_extent(&BTRFS_I(inode)->io_tree, destoff, end, &cached_state);
 	ret = btrfs_clone(src, inode, off, olen, len, destoff, 0);
-	btrfs_double_extent_unlock(src, off, inode, destoff, len);
+	unlock_extent(&BTRFS_I(inode)->io_tree, destoff, end, &cached_state);
 
 	/*
 	 * We may have copied an inline extent into a page of the destination
diff --git a/fs/btrfs/reflink.h b/fs/btrfs/reflink.h
index ecb309b4da..1e291f7d85 100644
--- a/fs/btrfs/reflink.h
+++ b/fs/btrfs/reflink.h
@@ -3,7 +3,9 @@
 #ifndef BTRFS_REFLINK_H
 #define BTRFS_REFLINK_H
 
-#include <linux/fs.h>
+#include <linux/types.h>
+
+struct file;
 
 loff_t btrfs_remap_file_range(struct file *file_in, loff_t pos_in,
 			      struct file *file_out, loff_t pos_out,
diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 2fca67f2b3..f2935252b9 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -473,20 +473,19 @@ static noinline_for_stack struct btrfs_backref_node *build_backref_tree(
 	struct btrfs_backref_node *node = NULL;
 	struct btrfs_backref_edge *edge;
 	int ret;
-	int err = 0;
 
 	iter = btrfs_backref_iter_alloc(rc->extent_root->fs_info);
 	if (!iter)
 		return ERR_PTR(-ENOMEM);
 	path = btrfs_alloc_path();
 	if (!path) {
-		err = -ENOMEM;
+		ret = -ENOMEM;
 		goto out;
 	}
 
 	node = btrfs_backref_alloc_node(cache, bytenr, level);
 	if (!node) {
-		err = -ENOMEM;
+		ret = -ENOMEM;
 		goto out;
 	}
 
@@ -497,10 +496,9 @@ static noinline_for_stack struct btrfs_backref_node *build_backref_tree(
 	do {
 		ret = btrfs_backref_add_tree_node(trans, cache, path, iter,
 						  node_key, cur);
-		if (ret < 0) {
-			err = ret;
+		if (ret < 0)
 			goto out;
-		}
+
 		edge = list_first_entry_or_null(&cache->pending_edge,
 				struct btrfs_backref_edge, list[UPPER]);
 		/*
@@ -515,19 +513,18 @@ static noinline_for_stack struct btrfs_backref_node *build_backref_tree(
 
 	/* Finish the upper linkage of newly added edges/nodes */
 	ret = btrfs_backref_finish_upper_links(cache, node);
-	if (ret < 0) {
-		err = ret;
+	if (ret < 0)
 		goto out;
-	}
 
 	if (handle_useless_nodes(rc, node))
 		node = NULL;
 out:
-	btrfs_backref_iter_free(iter);
+	btrfs_free_path(iter->path);
+	kfree(iter);
 	btrfs_free_path(path);
-	if (err) {
+	if (ret) {
 		btrfs_backref_error_cleanup(cache, node);
-		return ERR_PTR(err);
+		return ERR_PTR(ret);
 	}
 	ASSERT(!node || !node->detached);
 	ASSERT(list_empty(&cache->useless_node) &&
@@ -753,7 +750,7 @@ static struct btrfs_root *create_reloc_root(struct btrfs_trans_handle *trans,
 	root_key.type = BTRFS_ROOT_ITEM_KEY;
 	root_key.offset = objectid;
 
-	if (root->root_key.objectid == objectid) {
+	if (btrfs_root_id(root) == objectid) {
 		u64 commit_root_gen;
 
 		/* called by btrfs_init_reloc_root */
@@ -797,7 +794,7 @@ static struct btrfs_root *create_reloc_root(struct btrfs_trans_handle *trans,
 	btrfs_set_root_level(root_item, btrfs_header_level(eb));
 	btrfs_set_root_generation(root_item, trans->transid);
 
-	if (root->root_key.objectid == objectid) {
+	if (btrfs_root_id(root) == objectid) {
 		btrfs_set_root_refs(root_item, 0);
 		memset(&root_item->drop_progress, 0,
 		       sizeof(struct btrfs_disk_key));
@@ -820,7 +817,7 @@ static struct btrfs_root *create_reloc_root(struct btrfs_trans_handle *trans,
 		goto abort;
 	}
 	set_bit(BTRFS_ROOT_SHAREABLE, &reloc_root->state);
-	reloc_root->last_trans = trans->transid;
+	btrfs_set_root_last_trans(reloc_root, trans->transid);
 	return reloc_root;
 fail:
 	kfree(root_item);
@@ -867,7 +864,7 @@ int btrfs_init_reloc_root(struct btrfs_trans_handle *trans,
 	 */
 	if (root->reloc_root) {
 		reloc_root = root->reloc_root;
-		reloc_root->last_trans = trans->transid;
+		btrfs_set_root_last_trans(reloc_root, trans->transid);
 		return 0;
 	}
 
@@ -875,8 +872,7 @@ int btrfs_init_reloc_root(struct btrfs_trans_handle *trans,
 	 * We are merging reloc roots, we do not need new reloc trees.  Also
 	 * reloc trees never need their own reloc tree.
 	 */
-	if (!rc->create_reloc_tree ||
-	    root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID)
+	if (!rc->create_reloc_tree || btrfs_root_id(root) == BTRFS_TREE_RELOC_OBJECTID)
 		return 0;
 
 	if (!trans->reloc_reserved) {
@@ -884,7 +880,7 @@ int btrfs_init_reloc_root(struct btrfs_trans_handle *trans,
 		trans->block_rsv = rc->block_rsv;
 		clear_rsv = 1;
 	}
-	reloc_root = create_reloc_root(trans, root, root->root_key.objectid);
+	reloc_root = create_reloc_root(trans, root, btrfs_root_id(root));
 	if (clear_rsv)
 		trans->block_rsv = rsv;
 	if (IS_ERR(reloc_root))
@@ -951,60 +947,6 @@ int btrfs_update_reloc_root(struct btrfs_trans_handle *trans,
 }
 
 /*
- * helper to find first cached inode with inode number >= objectid
- * in a subvolume
- */
-static struct inode *find_next_inode(struct btrfs_root *root, u64 objectid)
-{
-	struct rb_node *node;
-	struct rb_node *prev;
-	struct btrfs_inode *entry;
-	struct inode *inode;
-
-	spin_lock(&root->inode_lock);
-again:
-	node = root->inode_tree.rb_node;
-	prev = NULL;
-	while (node) {
-		prev = node;
-		entry = rb_entry(node, struct btrfs_inode, rb_node);
-
-		if (objectid < btrfs_ino(entry))
-			node = node->rb_left;
-		else if (objectid > btrfs_ino(entry))
-			node = node->rb_right;
-		else
-			break;
-	}
-	if (!node) {
-		while (prev) {
-			entry = rb_entry(prev, struct btrfs_inode, rb_node);
-			if (objectid <= btrfs_ino(entry)) {
-				node = prev;
-				break;
-			}
-			prev = rb_next(prev);
-		}
-	}
-	while (node) {
-		entry = rb_entry(node, struct btrfs_inode, rb_node);
-		inode = igrab(&entry->vfs_inode);
-		if (inode) {
-			spin_unlock(&root->inode_lock);
-			return inode;
-		}
-
-		objectid = btrfs_ino(entry) + 1;
-		if (cond_resched_lock(&root->inode_lock))
-			goto again;
-
-		node = rb_next(node);
-	}
-	spin_unlock(&root->inode_lock);
-	return NULL;
-}
-
-/*
  * get new location of data
  */
 static int get_new_location(struct inode *reloc_inode, u64 *new_bytenr,
@@ -1064,7 +1006,7 @@ int replace_file_extents(struct btrfs_trans_handle *trans,
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct btrfs_key key;
 	struct btrfs_file_extent_item *fi;
-	struct inode *inode = NULL;
+	struct btrfs_inode *inode = NULL;
 	u64 parent;
 	u64 bytenr;
 	u64 new_bytenr = 0;
@@ -1080,7 +1022,7 @@ int replace_file_extents(struct btrfs_trans_handle *trans,
 		return 0;
 
 	/* reloc trees always use full backref */
-	if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID)
+	if (btrfs_root_id(root) == BTRFS_TREE_RELOC_OBJECTID)
 		parent = leaf->start;
 	else
 		parent = 0;
@@ -1109,15 +1051,15 @@ int replace_file_extents(struct btrfs_trans_handle *trans,
 		 * if we are modifying block in fs tree, wait for read_folio
 		 * to complete and drop the extent cache
 		 */
-		if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID) {
+		if (btrfs_root_id(root) != BTRFS_TREE_RELOC_OBJECTID) {
 			if (first) {
-				inode = find_next_inode(root, key.objectid);
+				inode = btrfs_find_first_inode(root, key.objectid);
 				first = 0;
-			} else if (inode && btrfs_ino(BTRFS_I(inode)) < key.objectid) {
-				btrfs_add_delayed_iput(BTRFS_I(inode));
-				inode = find_next_inode(root, key.objectid);
+			} else if (inode && btrfs_ino(inode) < key.objectid) {
+				btrfs_add_delayed_iput(inode);
+				inode = btrfs_find_first_inode(root, key.objectid);
 			}
-			if (inode && btrfs_ino(BTRFS_I(inode)) == key.objectid) {
+			if (inode && btrfs_ino(inode) == key.objectid) {
 				struct extent_state *cached_state = NULL;
 
 				end = key.offset +
@@ -1126,16 +1068,20 @@ int replace_file_extents(struct btrfs_trans_handle *trans,
 						    fs_info->sectorsize));
 				WARN_ON(!IS_ALIGNED(end, fs_info->sectorsize));
 				end--;
-				ret = try_lock_extent(&BTRFS_I(inode)->io_tree,
-						      key.offset, end,
-						      &cached_state);
-				if (!ret)
+				/* Take mmap lock to serialize with reflinks. */
+				if (!down_read_trylock(&inode->i_mmap_lock))
+					continue;
+				ret = try_lock_extent(&inode->io_tree, key.offset,
+						      end, &cached_state);
+				if (!ret) {
+					up_read(&inode->i_mmap_lock);
 					continue;
+				}
 
-				btrfs_drop_extent_map_range(BTRFS_I(inode),
-							    key.offset, end, true);
-				unlock_extent(&BTRFS_I(inode)->io_tree,
-					      key.offset, end, &cached_state);
+				btrfs_drop_extent_map_range(inode, key.offset, end, true);
+				unlock_extent(&inode->io_tree, key.offset, end,
+					      &cached_state);
+				up_read(&inode->i_mmap_lock);
 			}
 		}
 
@@ -1153,22 +1099,28 @@ int replace_file_extents(struct btrfs_trans_handle *trans,
 		dirty = 1;
 
 		key.offset -= btrfs_file_extent_offset(leaf, fi);
-		btrfs_init_generic_ref(&ref, BTRFS_ADD_DELAYED_REF, new_bytenr,
-				       num_bytes, parent, root->root_key.objectid);
-		btrfs_init_data_ref(&ref, btrfs_header_owner(leaf),
-				    key.objectid, key.offset,
-				    root->root_key.objectid, false);
+		ref.action = BTRFS_ADD_DELAYED_REF;
+		ref.bytenr = new_bytenr;
+		ref.num_bytes = num_bytes;
+		ref.parent = parent;
+		ref.owning_root = btrfs_root_id(root);
+		ref.ref_root = btrfs_header_owner(leaf);
+		btrfs_init_data_ref(&ref, key.objectid, key.offset,
+				    btrfs_root_id(root), false);
 		ret = btrfs_inc_extent_ref(trans, &ref);
 		if (ret) {
 			btrfs_abort_transaction(trans, ret);
 			break;
 		}
 
-		btrfs_init_generic_ref(&ref, BTRFS_DROP_DELAYED_REF, bytenr,
-				       num_bytes, parent, root->root_key.objectid);
-		btrfs_init_data_ref(&ref, btrfs_header_owner(leaf),
-				    key.objectid, key.offset,
-				    root->root_key.objectid, false);
+		ref.action = BTRFS_DROP_DELAYED_REF;
+		ref.bytenr = bytenr;
+		ref.num_bytes = num_bytes;
+		ref.parent = parent;
+		ref.owning_root = btrfs_root_id(root);
+		ref.ref_root = btrfs_header_owner(leaf);
+		btrfs_init_data_ref(&ref, key.objectid, key.offset,
+				    btrfs_root_id(root), false);
 		ret = btrfs_free_extent(trans, &ref);
 		if (ret) {
 			btrfs_abort_transaction(trans, ret);
@@ -1178,7 +1130,7 @@ int replace_file_extents(struct btrfs_trans_handle *trans,
 	if (dirty)
 		btrfs_mark_buffer_dirty(trans, leaf);
 	if (inode)
-		btrfs_add_delayed_iput(BTRFS_I(inode));
+		btrfs_add_delayed_iput(inode);
 	return ret;
 }
 
@@ -1224,8 +1176,8 @@ int replace_path(struct btrfs_trans_handle *trans, struct reloc_control *rc,
 	int ret;
 	int slot;
 
-	ASSERT(src->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID);
-	ASSERT(dest->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID);
+	ASSERT(btrfs_root_id(src) == BTRFS_TREE_RELOC_OBJECTID);
+	ASSERT(btrfs_root_id(dest) != BTRFS_TREE_RELOC_OBJECTID);
 
 	last_snapshot = btrfs_root_last_snapshot(&src->root_item);
 again:
@@ -1377,20 +1329,26 @@ again:
 					      path->slots[level], old_ptr_gen);
 		btrfs_mark_buffer_dirty(trans, path->nodes[level]);
 
-		btrfs_init_generic_ref(&ref, BTRFS_ADD_DELAYED_REF, old_bytenr,
-				       blocksize, path->nodes[level]->start,
-				       src->root_key.objectid);
-		btrfs_init_tree_ref(&ref, level - 1, src->root_key.objectid,
-				    0, true);
+		ref.action = BTRFS_ADD_DELAYED_REF;
+		ref.bytenr = old_bytenr;
+		ref.num_bytes = blocksize;
+		ref.parent = path->nodes[level]->start;
+		ref.owning_root = btrfs_root_id(src);
+		ref.ref_root = btrfs_root_id(src);
+		btrfs_init_tree_ref(&ref, level - 1, 0, true);
 		ret = btrfs_inc_extent_ref(trans, &ref);
 		if (ret) {
 			btrfs_abort_transaction(trans, ret);
 			break;
 		}
-		btrfs_init_generic_ref(&ref, BTRFS_ADD_DELAYED_REF, new_bytenr,
-				       blocksize, 0, dest->root_key.objectid);
-		btrfs_init_tree_ref(&ref, level - 1, dest->root_key.objectid, 0,
-				    true);
+
+		ref.action = BTRFS_ADD_DELAYED_REF;
+		ref.bytenr = new_bytenr;
+		ref.num_bytes = blocksize;
+		ref.parent = 0;
+		ref.owning_root = btrfs_root_id(dest);
+		ref.ref_root = btrfs_root_id(dest);
+		btrfs_init_tree_ref(&ref, level - 1, 0, true);
 		ret = btrfs_inc_extent_ref(trans, &ref);
 		if (ret) {
 			btrfs_abort_transaction(trans, ret);
@@ -1398,10 +1356,13 @@ again:
 		}
 
 		/* We don't know the real owning_root, use 0. */
-		btrfs_init_generic_ref(&ref, BTRFS_DROP_DELAYED_REF, new_bytenr,
-				       blocksize, path->nodes[level]->start, 0);
-		btrfs_init_tree_ref(&ref, level - 1, src->root_key.objectid,
-				    0, true);
+		ref.action = BTRFS_DROP_DELAYED_REF;
+		ref.bytenr = new_bytenr;
+		ref.num_bytes = blocksize;
+		ref.parent = path->nodes[level]->start;
+		ref.owning_root = 0;
+		ref.ref_root = btrfs_root_id(src);
+		btrfs_init_tree_ref(&ref, level - 1, 0, true);
 		ret = btrfs_free_extent(trans, &ref);
 		if (ret) {
 			btrfs_abort_transaction(trans, ret);
@@ -1409,10 +1370,13 @@ again:
 		}
 
 		/* We don't know the real owning_root, use 0. */
-		btrfs_init_generic_ref(&ref, BTRFS_DROP_DELAYED_REF, old_bytenr,
-				       blocksize, 0, 0);
-		btrfs_init_tree_ref(&ref, level - 1, dest->root_key.objectid,
-				    0, true);
+		ref.action = BTRFS_DROP_DELAYED_REF;
+		ref.bytenr = old_bytenr;
+		ref.num_bytes = blocksize;
+		ref.parent = 0;
+		ref.owning_root = 0;
+		ref.ref_root = btrfs_root_id(dest);
+		btrfs_init_tree_ref(&ref, level - 1, 0, true);
 		ret = btrfs_free_extent(trans, &ref);
 		if (ret) {
 			btrfs_abort_transaction(trans, ret);
@@ -1520,7 +1484,7 @@ static int invalidate_extent_cache(struct btrfs_root *root,
 				   const struct btrfs_key *max_key)
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
-	struct inode *inode = NULL;
+	struct btrfs_inode *inode = NULL;
 	u64 objectid;
 	u64 start, end;
 	u64 ino;
@@ -1530,23 +1494,24 @@ static int invalidate_extent_cache(struct btrfs_root *root,
 		struct extent_state *cached_state = NULL;
 
 		cond_resched();
-		iput(inode);
+		if (inode)
+			iput(&inode->vfs_inode);
 
 		if (objectid > max_key->objectid)
 			break;
 
-		inode = find_next_inode(root, objectid);
+		inode = btrfs_find_first_inode(root, objectid);
 		if (!inode)
 			break;
-		ino = btrfs_ino(BTRFS_I(inode));
+		ino = btrfs_ino(inode);
 
 		if (ino > max_key->objectid) {
-			iput(inode);
+			iput(&inode->vfs_inode);
 			break;
 		}
 
 		objectid = ino + 1;
-		if (!S_ISREG(inode->i_mode))
+		if (!S_ISREG(inode->vfs_inode.i_mode))
 			continue;
 
 		if (unlikely(min_key->objectid == ino)) {
@@ -1579,9 +1544,9 @@ static int invalidate_extent_cache(struct btrfs_root *root,
 		}
 
 		/* the lock_extent waits for read_folio to complete */
-		lock_extent(&BTRFS_I(inode)->io_tree, start, end, &cached_state);
-		btrfs_drop_extent_map_range(BTRFS_I(inode), start, end, true);
-		unlock_extent(&BTRFS_I(inode)->io_tree, start, end, &cached_state);
+		lock_extent(&inode->io_tree, start, end, &cached_state);
+		btrfs_drop_extent_map_range(inode, start, end, true);
+		unlock_extent(&inode->io_tree, start, end, &cached_state);
 	}
 	return 0;
 }
@@ -1616,7 +1581,7 @@ static int insert_dirty_subvol(struct btrfs_trans_handle *trans,
 	int ret;
 
 	/* @root must be a subvolume tree root with a valid reloc tree */
-	ASSERT(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID);
+	ASSERT(btrfs_root_id(root) != BTRFS_TREE_RELOC_OBJECTID);
 	ASSERT(reloc_root);
 
 	reloc_root_item = &reloc_root->root_item;
@@ -1645,7 +1610,7 @@ static int clean_dirty_subvols(struct reloc_control *rc)
 
 	list_for_each_entry_safe(root, next, &rc->dirty_subvol_roots,
 				 reloc_dirty_list) {
-		if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID) {
+		if (btrfs_root_id(root) != BTRFS_TREE_RELOC_OBJECTID) {
 			/* Merged subvolume, cleanup its reloc root */
 			struct btrfs_root *reloc_root = root->reloc_root;
 
@@ -1774,7 +1739,7 @@ static noinline_for_stack int merge_reloc_root(struct reloc_control *rc,
 		 * btrfs_update_reloc_root() and update our root item
 		 * appropriately.
 		 */
-		reloc_root->last_trans = trans->transid;
+		btrfs_set_root_last_trans(reloc_root, trans->transid);
 		trans->block_rsv = rc->block_rsv;
 
 		replaced = 0;
@@ -1920,13 +1885,13 @@ again:
 			if (root->reloc_root) {
 				btrfs_err(fs_info,
 "reloc tree mismatch, root %lld has reloc root key (%lld %u %llu) gen %llu, expect reloc root key (%lld %u %llu) gen %llu",
-					  root->root_key.objectid,
-					  root->reloc_root->root_key.objectid,
+					  btrfs_root_id(root),
+					  btrfs_root_id(root->reloc_root),
 					  root->reloc_root->root_key.type,
 					  root->reloc_root->root_key.offset,
 					  btrfs_root_generation(
 						  &root->reloc_root->root_item),
-					  reloc_root->root_key.objectid,
+					  btrfs_root_id(reloc_root),
 					  reloc_root->root_key.type,
 					  reloc_root->root_key.offset,
 					  btrfs_root_generation(
@@ -1934,8 +1899,8 @@ again:
 			} else {
 				btrfs_err(fs_info,
 "reloc tree mismatch, root %lld has no reloc root, expect reloc root key (%lld %u %llu) gen %llu",
-					  root->root_key.objectid,
-					  reloc_root->root_key.objectid,
+					  btrfs_root_id(root),
+					  btrfs_root_id(reloc_root),
 					  reloc_root->root_key.type,
 					  reloc_root->root_key.offset,
 					  btrfs_root_generation(
@@ -2117,7 +2082,7 @@ static int record_reloc_root_in_trans(struct btrfs_trans_handle *trans,
 	struct btrfs_root *root;
 	int ret;
 
-	if (reloc_root->last_trans == trans->transid)
+	if (btrfs_get_root_last_trans(reloc_root) == trans->transid)
 		return 0;
 
 	root = btrfs_get_fs_root(fs_info, reloc_root->root_key.offset, false);
@@ -2192,7 +2157,7 @@ struct btrfs_root *select_reloc_root(struct btrfs_trans_handle *trans,
 			return ERR_PTR(-EUCLEAN);
 		}
 
-		if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) {
+		if (btrfs_root_id(root) == BTRFS_TREE_RELOC_OBJECTID) {
 			ret = record_reloc_root_in_trans(trans, root);
 			if (ret)
 				return ERR_PTR(ret);
@@ -2299,7 +2264,7 @@ struct btrfs_root *select_one_root(struct btrfs_backref_node *node)
 		if (!test_bit(BTRFS_ROOT_SHAREABLE, &root->state))
 			return root;
 
-		if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID)
+		if (btrfs_root_id(root) != BTRFS_TREE_RELOC_OBJECTID)
 			fs_root = root;
 
 		if (next != node)
@@ -2315,9 +2280,8 @@ struct btrfs_root *select_one_root(struct btrfs_backref_node *node)
 	return fs_root;
 }
 
-static noinline_for_stack
-u64 calcu_metadata_size(struct reloc_control *rc,
-			struct btrfs_backref_node *node, int reserve)
+static noinline_for_stack u64 calcu_metadata_size(struct reloc_control *rc,
+						  struct btrfs_backref_node *node)
 {
 	struct btrfs_fs_info *fs_info = rc->extent_root->fs_info;
 	struct btrfs_backref_node *next = node;
@@ -2326,12 +2290,12 @@ u64 calcu_metadata_size(struct reloc_control *rc,
 	u64 num_bytes = 0;
 	int index = 0;
 
-	BUG_ON(reserve && node->processed);
+	BUG_ON(node->processed);
 
 	while (next) {
 		cond_resched();
 		while (1) {
-			if (next->processed && (reserve || next != node))
+			if (next->processed)
 				break;
 
 			num_bytes += fs_info->nodesize;
@@ -2359,7 +2323,7 @@ static int reserve_metadata_space(struct btrfs_trans_handle *trans,
 	int ret;
 	u64 tmp;
 
-	num_bytes = calcu_metadata_size(rc, node, 1) * 2;
+	num_bytes = calcu_metadata_size(rc, node) * 2;
 
 	trans->block_rsv = rc->block_rsv;
 	rc->reserved_bytes += num_bytes;
@@ -2422,8 +2386,6 @@ static int do_relocation(struct btrfs_trans_handle *trans,
 	path->lowest_level = node->level + 1;
 	rc->backref_cache.path[node->level] = node;
 	list_for_each_entry(edge, &node->upper, list[LOWER]) {
-		struct btrfs_ref ref = { 0 };
-
 		cond_resched();
 
 		upper = edge->node[UPPER];
@@ -2511,19 +2473,23 @@ static int do_relocation(struct btrfs_trans_handle *trans,
 			 */
 			ASSERT(node->eb == eb);
 		} else {
+			struct btrfs_ref ref = {
+				.action = BTRFS_ADD_DELAYED_REF,
+				.bytenr = node->eb->start,
+				.num_bytes = blocksize,
+				.parent = upper->eb->start,
+				.owning_root = btrfs_header_owner(upper->eb),
+				.ref_root = btrfs_header_owner(upper->eb),
+			};
+
 			btrfs_set_node_blockptr(upper->eb, slot,
 						node->eb->start);
 			btrfs_set_node_ptr_generation(upper->eb, slot,
 						      trans->transid);
 			btrfs_mark_buffer_dirty(trans, upper->eb);
 
-			btrfs_init_generic_ref(&ref, BTRFS_ADD_DELAYED_REF,
-					       node->eb->start, blocksize,
-					       upper->eb->start,
-					       btrfs_header_owner(upper->eb));
 			btrfs_init_tree_ref(&ref, node->level,
-					    btrfs_header_owner(upper->eb),
-					    root->root_key.objectid, false);
+					    btrfs_root_id(root), false);
 			ret = btrfs_inc_extent_ref(trans, &ref);
 			if (!ret)
 				ret = btrfs_drop_subtree(trans, root, eb,
@@ -2775,12 +2741,11 @@ int relocate_tree_blocks(struct btrfs_trans_handle *trans,
 	struct btrfs_path *path;
 	struct tree_block *block;
 	struct tree_block *next;
-	int ret;
-	int err = 0;
+	int ret = 0;
 
 	path = btrfs_alloc_path();
 	if (!path) {
-		err = -ENOMEM;
+		ret = -ENOMEM;
 		goto out_free_blocks;
 	}
 
@@ -2795,8 +2760,8 @@ int relocate_tree_blocks(struct btrfs_trans_handle *trans,
 	/* Get first keys */
 	rbtree_postorder_for_each_entry_safe(block, next, blocks, rb_node) {
 		if (!block->key_ready) {
-			err = get_tree_block_key(fs_info, block);
-			if (err)
+			ret = get_tree_block_key(fs_info, block);
+			if (ret)
 				goto out_free_path;
 		}
 	}
@@ -2806,25 +2771,23 @@ int relocate_tree_blocks(struct btrfs_trans_handle *trans,
 		node = build_backref_tree(trans, rc, &block->key,
 					  block->level, block->bytenr);
 		if (IS_ERR(node)) {
-			err = PTR_ERR(node);
+			ret = PTR_ERR(node);
 			goto out;
 		}
 
 		ret = relocate_tree_block(trans, rc, node, &block->key,
 					  path);
-		if (ret < 0) {
-			err = ret;
+		if (ret < 0)
 			break;
-		}
 	}
 out:
-	err = finish_pending_nodes(trans, rc, path, err);
+	ret = finish_pending_nodes(trans, rc, path, ret);
 
 out_free_path:
 	btrfs_free_path(path);
 out_free_blocks:
 	free_block_list(blocks);
-	return err;
+	return ret;
 }
 
 static noinline_for_stack int prealloc_file_extent_cluster(
@@ -2849,7 +2812,7 @@ static noinline_for_stack int prealloc_file_extent_cluster(
 	 * btrfs_do_readpage() call of previously relocated file cluster.
 	 *
 	 * If the current cluster starts in the above range, btrfs_do_readpage()
-	 * will skip the read, and relocate_one_page() will later writeback
+	 * will skip the read, and relocate_one_folio() will later writeback
 	 * the padding zeros as new data, causing data corruption.
 	 *
 	 * Here we have to manually invalidate the range (i_size, PAGE_END + 1).
@@ -2858,7 +2821,7 @@ static noinline_for_stack int prealloc_file_extent_cluster(
 		struct address_space *mapping = inode->vfs_inode.i_mapping;
 		struct btrfs_fs_info *fs_info = inode->root->fs_info;
 		const u32 sectorsize = fs_info->sectorsize;
-		struct page *page;
+		struct folio *folio;
 
 		ASSERT(sectorsize < PAGE_SIZE);
 		ASSERT(IS_ALIGNED(i_size, sectorsize));
@@ -2889,16 +2852,16 @@ static noinline_for_stack int prealloc_file_extent_cluster(
 		clear_extent_bits(&inode->io_tree, i_size,
 				  round_up(i_size, PAGE_SIZE) - 1,
 				  EXTENT_UPTODATE);
-		page = find_lock_page(mapping, i_size >> PAGE_SHIFT);
+		folio = filemap_lock_folio(mapping, i_size >> PAGE_SHIFT);
 		/*
 		 * If page is freed we don't need to do anything then, as we
 		 * will re-read the whole page anyway.
 		 */
-		if (page) {
-			btrfs_subpage_clear_uptodate(fs_info, page_folio(page), i_size,
+		if (!IS_ERR(folio)) {
+			btrfs_subpage_clear_uptodate(fs_info, folio, i_size,
 					round_up(i_size, PAGE_SIZE) - i_size);
-			unlock_page(page);
-			put_page(page);
+			folio_unlock(folio);
+			folio_put(folio);
 		}
 	}
 
@@ -2983,68 +2946,71 @@ static u64 get_cluster_boundary_end(const struct file_extent_cluster *cluster,
 	return cluster->boundary[cluster_nr + 1] - 1;
 }
 
-static int relocate_one_page(struct inode *inode, struct file_ra_state *ra,
-			     const struct file_extent_cluster *cluster,
-			     int *cluster_nr, unsigned long page_index)
+static int relocate_one_folio(struct inode *inode, struct file_ra_state *ra,
+			      const struct file_extent_cluster *cluster,
+			      int *cluster_nr, unsigned long index)
 {
 	struct btrfs_fs_info *fs_info = inode_to_fs_info(inode);
 	u64 offset = BTRFS_I(inode)->index_cnt;
 	const unsigned long last_index = (cluster->end - offset) >> PAGE_SHIFT;
 	gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping);
-	struct page *page;
-	u64 page_start;
-	u64 page_end;
+	struct folio *folio;
+	u64 folio_start;
+	u64 folio_end;
 	u64 cur;
 	int ret;
 
-	ASSERT(page_index <= last_index);
-	page = find_lock_page(inode->i_mapping, page_index);
-	if (!page) {
+	ASSERT(index <= last_index);
+	folio = filemap_lock_folio(inode->i_mapping, index);
+	if (IS_ERR(folio)) {
 		page_cache_sync_readahead(inode->i_mapping, ra, NULL,
-				page_index, last_index + 1 - page_index);
-		page = find_or_create_page(inode->i_mapping, page_index, mask);
-		if (!page)
-			return -ENOMEM;
+					  index, last_index + 1 - index);
+		folio = __filemap_get_folio(inode->i_mapping, index,
+					    FGP_LOCK | FGP_ACCESSED | FGP_CREAT, mask);
+		if (IS_ERR(folio))
+			return PTR_ERR(folio);
 	}
 
-	if (PageReadahead(page))
+	WARN_ON(folio_order(folio));
+
+	if (folio_test_readahead(folio))
 		page_cache_async_readahead(inode->i_mapping, ra, NULL,
-				page_folio(page), page_index,
-				last_index + 1 - page_index);
+					   folio, index,
+					   last_index + 1 - index);
 
-	if (!PageUptodate(page)) {
-		btrfs_read_folio(NULL, page_folio(page));
-		lock_page(page);
-		if (!PageUptodate(page)) {
+	if (!folio_test_uptodate(folio)) {
+		btrfs_read_folio(NULL, folio);
+		folio_lock(folio);
+		if (!folio_test_uptodate(folio)) {
 			ret = -EIO;
-			goto release_page;
+			goto release_folio;
 		}
 	}
 
 	/*
-	 * We could have lost page private when we dropped the lock to read the
-	 * page above, make sure we set_page_extent_mapped here so we have any
+	 * We could have lost folio private when we dropped the lock to read the
+	 * folio above, make sure we set_page_extent_mapped here so we have any
 	 * of the subpage blocksize stuff we need in place.
 	 */
-	ret = set_page_extent_mapped(page);
+	ret = set_folio_extent_mapped(folio);
 	if (ret < 0)
-		goto release_page;
+		goto release_folio;
 
-	page_start = page_offset(page);
-	page_end = page_start + PAGE_SIZE - 1;
+	folio_start = folio_pos(folio);
+	folio_end = folio_start + PAGE_SIZE - 1;
 
 	/*
 	 * Start from the cluster, as for subpage case, the cluster can start
-	 * inside the page.
+	 * inside the folio.
 	 */
-	cur = max(page_start, cluster->boundary[*cluster_nr] - offset);
-	while (cur <= page_end) {
+	cur = max(folio_start, cluster->boundary[*cluster_nr] - offset);
+	while (cur <= folio_end) {
 		struct extent_state *cached_state = NULL;
 		u64 extent_start = cluster->boundary[*cluster_nr] - offset;
 		u64 extent_end = get_cluster_boundary_end(cluster,
 						*cluster_nr) - offset;
-		u64 clamped_start = max(page_start, extent_start);
-		u64 clamped_end = min(page_end, extent_end);
+		u64 clamped_start = max(folio_start, extent_start);
+		u64 clamped_end = min(folio_end, extent_end);
 		u32 clamped_len = clamped_end + 1 - clamped_start;
 
 		/* Reserve metadata for this range */
@@ -3052,7 +3018,7 @@ static int relocate_one_page(struct inode *inode, struct file_ra_state *ra,
 						      clamped_len, clamped_len,
 						      false);
 		if (ret)
-			goto release_page;
+			goto release_folio;
 
 		/* Mark the range delalloc and dirty for later writeback */
 		lock_extent(&BTRFS_I(inode)->io_tree, clamped_start, clamped_end,
@@ -3068,20 +3034,18 @@ static int relocate_one_page(struct inode *inode, struct file_ra_state *ra,
 							clamped_len, true);
 			btrfs_delalloc_release_extents(BTRFS_I(inode),
 						       clamped_len);
-			goto release_page;
+			goto release_folio;
 		}
-		btrfs_folio_set_dirty(fs_info, page_folio(page),
-				      clamped_start, clamped_len);
+		btrfs_folio_set_dirty(fs_info, folio, clamped_start, clamped_len);
 
 		/*
-		 * Set the boundary if it's inside the page.
+		 * Set the boundary if it's inside the folio.
 		 * Data relocation requires the destination extents to have the
 		 * same size as the source.
 		 * EXTENT_BOUNDARY bit prevents current extent from being merged
 		 * with previous extent.
 		 */
-		if (in_range(cluster->boundary[*cluster_nr] - offset,
-			     page_start, PAGE_SIZE)) {
+		if (in_range(cluster->boundary[*cluster_nr] - offset, folio_start, PAGE_SIZE)) {
 			u64 boundary_start = cluster->boundary[*cluster_nr] -
 						offset;
 			u64 boundary_end = boundary_start +
@@ -3104,8 +3068,8 @@ static int relocate_one_page(struct inode *inode, struct file_ra_state *ra,
 				break;
 		}
 	}
-	unlock_page(page);
-	put_page(page);
+	folio_unlock(folio);
+	folio_put(folio);
 
 	balance_dirty_pages_ratelimited(inode->i_mapping);
 	btrfs_throttle(fs_info);
@@ -3113,9 +3077,9 @@ static int relocate_one_page(struct inode *inode, struct file_ra_state *ra,
 		ret = -ECANCELED;
 	return ret;
 
-release_page:
-	unlock_page(page);
-	put_page(page);
+release_folio:
+	folio_unlock(folio);
+	folio_put(folio);
 	return ret;
 }
 
@@ -3150,7 +3114,7 @@ static int relocate_file_extent_cluster(struct inode *inode,
 	last_index = (cluster->end - offset) >> PAGE_SHIFT;
 	for (index = (cluster->start - offset) >> PAGE_SHIFT;
 	     index <= last_index && !ret; index++)
-		ret = relocate_one_page(inode, ra, cluster, &cluster_nr, index);
+		ret = relocate_one_folio(inode, ra, cluster, &cluster_nr, index);
 	if (ret == 0)
 		WARN_ON(cluster_nr != cluster->nr);
 out:
@@ -3927,7 +3891,7 @@ static noinline_for_stack struct inode *create_reloc_inode(
 	struct btrfs_trans_handle *trans;
 	struct btrfs_root *root;
 	u64 objectid;
-	int err = 0;
+	int ret = 0;
 
 	root = btrfs_grab_root(fs_info->data_reloc_root);
 	trans = btrfs_start_transaction(root, 6);
@@ -3936,31 +3900,31 @@ static noinline_for_stack struct inode *create_reloc_inode(
 		return ERR_CAST(trans);
 	}
 
-	err = btrfs_get_free_objectid(root, &objectid);
-	if (err)
+	ret = btrfs_get_free_objectid(root, &objectid);
+	if (ret)
 		goto out;
 
-	err = __insert_orphan_inode(trans, root, objectid);
-	if (err)
+	ret = __insert_orphan_inode(trans, root, objectid);
+	if (ret)
 		goto out;
 
 	inode = btrfs_iget(fs_info->sb, objectid, root);
 	if (IS_ERR(inode)) {
 		delete_orphan_inode(trans, root, objectid);
-		err = PTR_ERR(inode);
+		ret = PTR_ERR(inode);
 		inode = NULL;
 		goto out;
 	}
 	BTRFS_I(inode)->index_cnt = group->start;
 
-	err = btrfs_orphan_add(trans, BTRFS_I(inode));
+	ret = btrfs_orphan_add(trans, BTRFS_I(inode));
 out:
 	btrfs_put_root(root);
 	btrfs_end_transaction(trans);
 	btrfs_btree_balance_dirty(fs_info);
-	if (err) {
+	if (ret) {
 		iput(inode);
-		inode = ERR_PTR(err);
+		inode = ERR_PTR(ret);
 	}
 	return inode;
 }
@@ -4438,9 +4402,11 @@ int btrfs_reloc_clone_csums(struct btrfs_ordered_extent *ordered)
 
 	ret = btrfs_lookup_csums_list(csum_root, disk_bytenr,
 				      disk_bytenr + ordered->num_bytes - 1,
-				      &list, 0, false);
-	if (ret)
+				      &list, false);
+	if (ret < 0) {
+		btrfs_mark_ordered_extent_error(ordered);
 		return ret;
+	}
 
 	while (!list_empty(&list)) {
 		struct btrfs_ordered_sum *sums =
@@ -4490,8 +4456,7 @@ int btrfs_reloc_cow_block(struct btrfs_trans_handle *trans,
 	    btrfs_root_last_snapshot(&root->root_item))
 		first_cow = 1;
 
-	if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID &&
-	    rc->create_reloc_tree) {
+	if (btrfs_root_id(root) == BTRFS_TREE_RELOC_OBJECTID && rc->create_reloc_tree) {
 		WARN_ON(!first_cow && level == 0);
 
 		node = rc->backref_cache.path[level];
@@ -4584,8 +4549,7 @@ int btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans,
 	}
 
 	new_root = pending->snap;
-	reloc_root = create_reloc_root(trans, root->reloc_root,
-				       new_root->root_key.objectid);
+	reloc_root = create_reloc_root(trans, root->reloc_root, btrfs_root_id(new_root));
 	if (IS_ERR(reloc_root))
 		return PTR_ERR(reloc_root);
 
diff --git a/fs/btrfs/relocation.h b/fs/btrfs/relocation.h
index 5fb60f2deb..788c86d863 100644
--- a/fs/btrfs/relocation.h
+++ b/fs/btrfs/relocation.h
@@ -3,6 +3,15 @@
 #ifndef BTRFS_RELOCATION_H
 #define BTRFS_RELOCATION_H
 
+#include <linux/types.h>
+
+struct extent_buffer;
+struct btrfs_fs_info;
+struct btrfs_root;
+struct btrfs_trans_handle;
+struct btrfs_ordered_extent;
+struct btrfs_pending_snapshot;
+
 int btrfs_relocate_block_group(struct btrfs_fs_info *fs_info, u64 group_start);
 int btrfs_init_reloc_root(struct btrfs_trans_handle *trans, struct btrfs_root *root);
 int btrfs_update_reloc_root(struct btrfs_trans_handle *trans,
diff --git a/fs/btrfs/root-tree.c b/fs/btrfs/root-tree.c
index 5260677ad5..33962671a9 100644
--- a/fs/btrfs/root-tree.c
+++ b/fs/btrfs/root-tree.c
@@ -10,7 +10,6 @@
 #include "messages.h"
 #include "transaction.h"
 #include "disk-io.h"
-#include "print-tree.h"
 #include "qgroup.h"
 #include "space-info.h"
 #include "accessors.h"
@@ -82,7 +81,14 @@ int btrfs_find_root(struct btrfs_root *root, const struct btrfs_key *search_key,
 		if (ret > 0)
 			goto out;
 	} else {
-		BUG_ON(ret == 0);		/* Logical error */
+		/*
+		 * Key with offset -1 found, there would have to exist a root
+		 * with such id, but this is out of the valid range.
+		 */
+		if (ret == 0) {
+			ret = -EUCLEAN;
+			goto out;
+		}
 		if (path->slots[0] == 0)
 			goto out;
 		path->slots[0]--;
@@ -142,8 +148,7 @@ int btrfs_update_root(struct btrfs_trans_handle *trans, struct btrfs_root
 	if (ret > 0) {
 		btrfs_crit(fs_info,
 			"unable to find root key (%llu %u %llu) in tree %llu",
-			key->objectid, key->type, key->offset,
-			root->root_key.objectid);
+			key->objectid, key->type, key->offset, btrfs_root_id(root));
 		ret = -EUCLEAN;
 		btrfs_abort_transaction(trans, ret);
 		goto out;
@@ -323,8 +328,11 @@ int btrfs_del_root(struct btrfs_trans_handle *trans,
 	ret = btrfs_search_slot(trans, root, key, path, -1, 1);
 	if (ret < 0)
 		goto out;
-
-	BUG_ON(ret != 0);
+	if (ret != 0) {
+		/* The root must exist but we did not find it by the key. */
+		ret = -EUCLEAN;
+		goto out;
+	}
 
 	ret = btrfs_del_item(trans, root, path);
 out:
diff --git a/fs/btrfs/root-tree.h b/fs/btrfs/root-tree.h
index de0175cbdb..8f5739e732 100644
--- a/fs/btrfs/root-tree.h
+++ b/fs/btrfs/root-tree.h
@@ -3,7 +3,17 @@
 #ifndef BTRFS_ROOT_TREE_H
 #define BTRFS_ROOT_TREE_H
 
+#include <linux/types.h>
+
 struct fscrypt_str;
+struct extent_buffer;
+struct btrfs_key;
+struct btrfs_root;
+struct btrfs_root_item;
+struct btrfs_path;
+struct btrfs_fs_info;
+struct btrfs_block_rsv;
+struct btrfs_trans_handle;
 
 int btrfs_subvolume_reserve_metadata(struct btrfs_root *root,
 				     struct btrfs_block_rsv *rsv,
diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c
index d89d5c0a8c..d7caa3732f 100644
--- a/fs/btrfs/scrub.c
+++ b/fs/btrfs/scrub.c
@@ -1390,8 +1390,15 @@ static int find_first_extent_item(struct btrfs_root *extent_root,
 	ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
 	if (ret < 0)
 		return ret;
+	if (ret == 0) {
+		/*
+		 * Key with offset -1 found, there would have to exist an extent
+		 * item with such offset, but this is out of the valid range.
+		 */
+		btrfs_release_path(path);
+		return -EUCLEAN;
+	}
 
-	ASSERT(ret > 0);
 	/*
 	 * Here we intentionally pass 0 as @min_objectid, as there could be
 	 * an extent item starting before @search_start.
@@ -1681,20 +1688,24 @@ static void scrub_submit_extent_sector_read(struct scrub_ctx *sctx,
 					    (i << fs_info->sectorsize_bits);
 			int err;
 
-			bbio = btrfs_bio_alloc(stripe->nr_sectors, REQ_OP_READ,
-					       fs_info, scrub_read_endio, stripe);
-			bbio->bio.bi_iter.bi_sector = logical >> SECTOR_SHIFT;
-
 			io_stripe.is_scrub = true;
+			stripe_len = (nr_sectors - i) << fs_info->sectorsize_bits;
+			/*
+			 * For RST cases, we need to manually split the bbio to
+			 * follow the RST boundary.
+			 */
 			err = btrfs_map_block(fs_info, BTRFS_MAP_READ, logical,
-					      &stripe_len, &bioc, &io_stripe,
-					      &mirror);
+					      &stripe_len, &bioc, &io_stripe, &mirror);
 			btrfs_put_bioc(bioc);
-			if (err) {
-				btrfs_bio_end_io(bbio,
-						 errno_to_blk_status(err));
-				return;
+			if (err < 0) {
+				set_bit(i, &stripe->io_error_bitmap);
+				set_bit(i, &stripe->error_bitmap);
+				continue;
 			}
+
+			bbio = btrfs_bio_alloc(stripe->nr_sectors, REQ_OP_READ,
+					       fs_info, scrub_read_endio, stripe);
+			bbio->bio.bi_iter.bi_sector = logical >> SECTOR_SHIFT;
 		}
 
 		__bio_add_page(&bbio->bio, page, fs_info->sectorsize, pgoff);
@@ -2093,7 +2104,7 @@ static int scrub_simple_mirror(struct scrub_ctx *sctx,
 	struct btrfs_fs_info *fs_info = sctx->fs_info;
 	const u64 logical_end = logical_start + logical_length;
 	u64 cur_logical = logical_start;
-	int ret;
+	int ret = 0;
 
 	/* The range must be inside the bg */
 	ASSERT(logical_start >= bg->start && logical_end <= bg->start + bg->length);
diff --git a/fs/btrfs/scrub.h b/fs/btrfs/scrub.h
index 7639103ebf..f0df597b75 100644
--- a/fs/btrfs/scrub.h
+++ b/fs/btrfs/scrub.h
@@ -3,6 +3,12 @@
 #ifndef BTRFS_SCRUB_H
 #define BTRFS_SCRUB_H
 
+#include <linux/types.h>
+
+struct btrfs_fs_info;
+struct btrfs_device;
+struct btrfs_scrub_progress;
+
 int btrfs_scrub_dev(struct btrfs_fs_info *fs_info, u64 devid, u64 start,
 		    u64 end, struct btrfs_scrub_progress *progress,
 		    int readonly, int is_dev_replace);
diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c
index e818766915..3dd4a48479 100644
--- a/fs/btrfs/send.c
+++ b/fs/btrfs/send.c
@@ -25,7 +25,6 @@
 #include "btrfs_inode.h"
 #include "transaction.h"
 #include "compression.h"
-#include "xattr.h"
 #include "print-tree.h"
 #include "accessors.h"
 #include "dir-item.h"
@@ -393,9 +392,8 @@ static void inconsistent_snapshot_error(struct send_ctx *sctx,
 	btrfs_err(sctx->send_root->fs_info,
 		  "Send: inconsistent snapshot, found %s %s for inode %llu without updated inode item, send root is %llu, parent root is %llu",
 		  result_string, what, sctx->cmp_key->objectid,
-		  sctx->send_root->root_key.objectid,
-		  (sctx->parent_root ?
-		   sctx->parent_root->root_key.objectid : 0));
+		  btrfs_root_id(sctx->send_root),
+		  (sctx->parent_root ?  btrfs_root_id(sctx->parent_root) : 0));
 }
 
 __maybe_unused
@@ -777,7 +775,12 @@ static int begin_cmd(struct send_ctx *sctx, int cmd)
 	if (WARN_ON(!sctx->send_buf))
 		return -EINVAL;
 
-	BUG_ON(sctx->send_size);
+	if (unlikely(sctx->send_size != 0)) {
+		btrfs_err(sctx->send_root->fs_info,
+			  "send: command header buffer not empty cmd %d offset %llu",
+			  cmd, sctx->send_off);
+		return -EINVAL;
+	}
 
 	sctx->send_size += sizeof(*hdr);
 	hdr = (struct btrfs_cmd_header *)sctx->send_buf;
@@ -1312,9 +1315,9 @@ static int __clone_root_cmp_bsearch(const void *key, const void *elt)
 	u64 root = (u64)(uintptr_t)key;
 	const struct clone_root *cr = elt;
 
-	if (root < cr->root->root_key.objectid)
+	if (root < btrfs_root_id(cr->root))
 		return -1;
-	if (root > cr->root->root_key.objectid)
+	if (root > btrfs_root_id(cr->root))
 		return 1;
 	return 0;
 }
@@ -1324,9 +1327,9 @@ static int __clone_root_cmp_sort(const void *e1, const void *e2)
 	const struct clone_root *cr1 = e1;
 	const struct clone_root *cr2 = e2;
 
-	if (cr1->root->root_key.objectid < cr2->root->root_key.objectid)
+	if (btrfs_root_id(cr1->root) < btrfs_root_id(cr2->root))
 		return -1;
-	if (cr1->root->root_key.objectid > cr2->root->root_key.objectid)
+	if (btrfs_root_id(cr1->root) > btrfs_root_id(cr2->root))
 		return 1;
 	return 0;
 }
@@ -1414,7 +1417,7 @@ static bool lookup_backref_cache(u64 leaf_bytenr, void *ctx,
 	struct btrfs_lru_cache_entry *raw_entry;
 	struct backref_cache_entry *entry;
 
-	if (btrfs_lru_cache_size(&sctx->backref_cache) == 0)
+	if (sctx->backref_cache.size == 0)
 		return false;
 
 	/*
@@ -1512,7 +1515,7 @@ static void store_backref_cache(u64 leaf_bytenr, const struct ulist *root_ids,
 	 * transaction handle or holding fs_info->commit_root_sem, so no need
 	 * to take any lock here.
 	 */
-	if (btrfs_lru_cache_size(&sctx->backref_cache) == 1)
+	if (sctx->backref_cache.size == 1)
 		sctx->backref_cache_last_reloc_trans = fs_info->last_reloc_trans;
 }
 
@@ -1774,7 +1777,7 @@ static int read_symlink(struct btrfs_root *root,
 		 */
 		btrfs_err(root->fs_info,
 			  "Found empty symlink inode %llu at root %llu",
-			  ino, root->root_key.objectid);
+			  ino, btrfs_root_id(root));
 		ret = -EIO;
 		goto out;
 	}
@@ -2528,7 +2531,7 @@ static int send_subvol_begin(struct send_ctx *sctx)
 		return -ENOMEM;
 	}
 
-	key.objectid = send_root->root_key.objectid;
+	key.objectid = btrfs_root_id(send_root);
 	key.type = BTRFS_ROOT_BACKREF_KEY;
 	key.offset = 0;
 
@@ -2544,7 +2547,7 @@ static int send_subvol_begin(struct send_ctx *sctx)
 	leaf = path->nodes[0];
 	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
 	if (key.type != BTRFS_ROOT_BACKREF_KEY ||
-	    key.objectid != send_root->root_key.objectid) {
+	    key.objectid != btrfs_root_id(send_root)) {
 		ret = -ENOENT;
 		goto out;
 	}
@@ -2817,8 +2820,7 @@ static int cache_dir_utimes(struct send_ctx *sctx, u64 dir, u64 gen)
 
 static int trim_dir_utimes_cache(struct send_ctx *sctx)
 {
-	while (btrfs_lru_cache_size(&sctx->dir_utimes_cache) >
-	       SEND_MAX_DIR_UTIMES_CACHE_SIZE) {
+	while (sctx->dir_utimes_cache.size > SEND_MAX_DIR_UTIMES_CACHE_SIZE) {
 		struct btrfs_lru_cache_entry *lru;
 		int ret;
 
@@ -4190,7 +4192,13 @@ static int process_recorded_refs(struct send_ctx *sctx, int *pending_move)
 	 * This should never happen as the root dir always has the same ref
 	 * which is always '..'
 	 */
-	BUG_ON(sctx->cur_ino <= BTRFS_FIRST_FREE_OBJECTID);
+	if (unlikely(sctx->cur_ino <= BTRFS_FIRST_FREE_OBJECTID)) {
+		btrfs_err(fs_info,
+			  "send: unexpected inode %llu in process_recorded_refs()",
+			  sctx->cur_ino);
+		ret = -EINVAL;
+		goto out;
+	}
 
 	valid_path = fs_path_alloc();
 	if (!valid_path) {
@@ -5265,10 +5273,11 @@ static int put_file_data(struct send_ctx *sctx, u64 offset, u32 len)
 {
 	struct btrfs_root *root = sctx->send_root;
 	struct btrfs_fs_info *fs_info = root->fs_info;
-	struct page *page;
+	struct folio *folio;
 	pgoff_t index = offset >> PAGE_SHIFT;
 	pgoff_t last_index;
 	unsigned pg_offset = offset_in_page(offset);
+	struct address_space *mapping = sctx->cur_inode->i_mapping;
 	int ret;
 
 	ret = put_data_header(sctx, len);
@@ -5281,44 +5290,44 @@ static int put_file_data(struct send_ctx *sctx, u64 offset, u32 len)
 		unsigned cur_len = min_t(unsigned, len,
 					 PAGE_SIZE - pg_offset);
 
-		page = find_lock_page(sctx->cur_inode->i_mapping, index);
-		if (!page) {
-			page_cache_sync_readahead(sctx->cur_inode->i_mapping,
+		folio = filemap_lock_folio(mapping, index);
+		if (IS_ERR(folio)) {
+			page_cache_sync_readahead(mapping,
 						  &sctx->ra, NULL, index,
 						  last_index + 1 - index);
 
-			page = find_or_create_page(sctx->cur_inode->i_mapping,
-						   index, GFP_KERNEL);
-			if (!page) {
-				ret = -ENOMEM;
+	                folio = filemap_grab_folio(mapping, index);
+			if (IS_ERR(folio)) {
+				ret = PTR_ERR(folio);
 				break;
 			}
 		}
 
-		if (PageReadahead(page))
-			page_cache_async_readahead(sctx->cur_inode->i_mapping,
-						   &sctx->ra, NULL, page_folio(page),
+		WARN_ON(folio_order(folio));
+
+		if (folio_test_readahead(folio))
+			page_cache_async_readahead(mapping, &sctx->ra, NULL, folio,
 						   index, last_index + 1 - index);
 
-		if (!PageUptodate(page)) {
-			btrfs_read_folio(NULL, page_folio(page));
-			lock_page(page);
-			if (!PageUptodate(page)) {
-				unlock_page(page);
+		if (!folio_test_uptodate(folio)) {
+			btrfs_read_folio(NULL, folio);
+			folio_lock(folio);
+			if (!folio_test_uptodate(folio)) {
+				folio_unlock(folio);
 				btrfs_err(fs_info,
 			"send: IO error at offset %llu for inode %llu root %llu",
-					page_offset(page), sctx->cur_ino,
-					sctx->send_root->root_key.objectid);
-				put_page(page);
+					folio_pos(folio), sctx->cur_ino,
+					btrfs_root_id(sctx->send_root));
+				folio_put(folio);
 				ret = -EIO;
 				break;
 			}
 		}
 
-		memcpy_from_page(sctx->send_buf + sctx->send_size, page,
-				 pg_offset, cur_len);
-		unlock_page(page);
-		put_page(page);
+		memcpy_from_folio(sctx->send_buf + sctx->send_size, folio,
+				  pg_offset, cur_len);
+		folio_unlock(folio);
+		folio_put(folio);
 		index++;
 		pg_offset = 0;
 		len -= cur_len;
@@ -5379,7 +5388,7 @@ static int send_clone(struct send_ctx *sctx,
 
 	btrfs_debug(sctx->send_root->fs_info,
 		    "send_clone offset=%llu, len=%d, clone_root=%llu, clone_inode=%llu, clone_offset=%llu",
-		    offset, len, clone_root->root->root_key.objectid,
+		    offset, len, btrfs_root_id(clone_root->root),
 		    clone_root->ino, clone_root->offset);
 
 	p = fs_path_alloc();
@@ -6466,21 +6475,18 @@ static int maybe_send_hole(struct send_ctx *sctx, struct btrfs_path *path,
 	if (sctx->cur_ino != key->objectid || !need_send_hole(sctx))
 		return 0;
 
-	if (sctx->cur_inode_last_extent == (u64)-1) {
-		ret = get_last_extent(sctx, key->offset - 1);
-		if (ret)
-			return ret;
-	}
-
-	if (path->slots[0] == 0 &&
-	    sctx->cur_inode_last_extent < key->offset) {
-		/*
-		 * We might have skipped entire leafs that contained only
-		 * file extent items for our current inode. These leafs have
-		 * a generation number smaller (older) than the one in the
-		 * current leaf and the leaf our last extent came from, and
-		 * are located between these 2 leafs.
-		 */
+	/*
+	 * Get last extent's end offset (exclusive) if we haven't determined it
+	 * yet (we're processing the first file extent item that is new), or if
+	 * we're at the first slot of a leaf and the last extent's end is less
+	 * than the current extent's offset, because we might have skipped
+	 * entire leaves that contained only file extent items for our current
+	 * inode. These leaves have a generation number smaller (older) than the
+	 * one in the current leaf and the leaf our last extent came from, and
+	 * are located between these 2 leaves.
+	 */
+	if ((sctx->cur_inode_last_extent == (u64)-1) ||
+	    (path->slots[0] == 0 && sctx->cur_inode_last_extent < key->offset)) {
 		ret = get_last_extent(sctx, key->offset - 1);
 		if (ret)
 			return ret;
@@ -7331,7 +7337,7 @@ static int search_key_again(const struct send_ctx *sctx,
 "send: key (%llu %u %llu) not found in %s root %llu, lowest_level %d, slot %d",
 			  key->objectid, key->type, key->offset,
 			  (root == sctx->parent_root ? "parent" : "send"),
-			  root->root_key.objectid, path->lowest_level,
+			  btrfs_root_id(root), path->lowest_level,
 			  path->slots[path->lowest_level]);
 		return -EUCLEAN;
 	}
@@ -7437,8 +7443,8 @@ static int tree_move_down(struct btrfs_path *path, int *level, u64 reada_min_gen
 	u64 reada_done = 0;
 
 	lockdep_assert_held_read(&parent->fs_info->commit_root_sem);
+	ASSERT(*level != 0);
 
-	BUG_ON(*level == 0);
 	eb = btrfs_read_node_slot(parent, slot);
 	if (IS_ERR(eb))
 		return PTR_ERR(eb);
@@ -8065,7 +8071,7 @@ static void btrfs_root_dec_send_in_progress(struct btrfs_root* root)
 	if (root->send_in_progress < 0)
 		btrfs_err(root->fs_info,
 			  "send_in_progress unbalanced %d root %llu",
-			  root->send_in_progress, root->root_key.objectid);
+			  root->send_in_progress, btrfs_root_id(root));
 	spin_unlock(&root->root_item_lock);
 }
 
@@ -8073,7 +8079,7 @@ static void dedupe_in_progress_warn(const struct btrfs_root *root)
 {
 	btrfs_warn_rl(root->fs_info,
 "cannot use root %llu for send while deduplications on it are in progress (%d in progress)",
-		      root->root_key.objectid, root->dedupe_in_progress);
+		      btrfs_root_id(root), root->dedupe_in_progress);
 }
 
 long btrfs_ioctl_send(struct inode *inode, struct btrfs_ioctl_send_args *arg)
diff --git a/fs/btrfs/send.h b/fs/btrfs/send.h
index 4f5509cb18..dd1c9f02b0 100644
--- a/fs/btrfs/send.h
+++ b/fs/btrfs/send.h
@@ -8,6 +8,11 @@
 #define BTRFS_SEND_H
 
 #include <linux/types.h>
+#include <linux/sizes.h>
+#include <linux/align.h>
+
+struct inode;
+struct btrfs_ioctl_send_args;
 
 #define BTRFS_SEND_STREAM_MAGIC "btrfs-stream"
 /* Conditional support for the upcoming protocol version. */
@@ -25,9 +30,6 @@
 #define BTRFS_SEND_BUF_SIZE_V1				SZ_64K
 #define BTRFS_SEND_BUF_SIZE_V2	ALIGN(SZ_16K + BTRFS_MAX_COMPRESSED, PAGE_SIZE)
 
-struct inode;
-struct btrfs_ioctl_send_args;
-
 enum btrfs_tlv_type {
 	BTRFS_TLV_U8,
 	BTRFS_TLV_U16,
diff --git a/fs/btrfs/space-info.c b/fs/btrfs/space-info.c
index 3b54eb5834..8f194eefd3 100644
--- a/fs/btrfs/space-info.c
+++ b/fs/btrfs/space-info.c
@@ -9,7 +9,6 @@
 #include "ordered-data.h"
 #include "transaction.h"
 #include "block-group.h"
-#include "zoned.h"
 #include "fs.h"
 #include "accessors.h"
 #include "extent-tree.h"
@@ -312,7 +311,7 @@ void btrfs_add_bg_to_space_info(struct btrfs_fs_info *info,
 	found->bytes_used += block_group->used;
 	found->disk_used += block_group->used * factor;
 	found->bytes_readonly += block_group->bytes_super;
-	found->bytes_zone_unusable += block_group->zone_unusable;
+	btrfs_space_info_update_bytes_zone_unusable(info, found, block_group->zone_unusable);
 	if (block_group->length > 0)
 		found->full = 0;
 	btrfs_try_granting_tickets(info, found);
@@ -374,11 +373,18 @@ static u64 calc_available_free_space(struct btrfs_fs_info *fs_info,
 	 * "optimal" chunk size based on the fs size.  However when we actually
 	 * allocate the chunk we will strip this down further, making it no more
 	 * than 10% of the disk or 1G, whichever is smaller.
+	 *
+	 * On the zoned mode, we need to use zone_size (=
+	 * data_sinfo->chunk_size) as it is.
 	 */
 	data_sinfo = btrfs_find_space_info(fs_info, BTRFS_BLOCK_GROUP_DATA);
-	data_chunk_size = min(data_sinfo->chunk_size,
-			      mult_perc(fs_info->fs_devices->total_rw_bytes, 10));
-	data_chunk_size = min_t(u64, data_chunk_size, SZ_1G);
+	if (!btrfs_is_zoned(fs_info)) {
+		data_chunk_size = min(data_sinfo->chunk_size,
+				      mult_perc(fs_info->fs_devices->total_rw_bytes, 10));
+		data_chunk_size = min_t(u64, data_chunk_size, SZ_1G);
+	} else {
+		data_chunk_size = data_sinfo->chunk_size;
+	}
 
 	/*
 	 * Since data allocations immediately use block groups as part of the
@@ -406,6 +412,17 @@ static u64 calc_available_free_space(struct btrfs_fs_info *fs_info,
 		avail >>= 3;
 	else
 		avail >>= 1;
+
+	/*
+	 * On the zoned mode, we always allocate one zone as one chunk.
+	 * Returning non-zone size alingned bytes here will result in
+	 * less pressure for the async metadata reclaim process, and it
+	 * will over-commit too much leading to ENOSPC. Align down to the
+	 * zone size to avoid that.
+	 */
+	if (btrfs_is_zoned(fs_info))
+		avail = ALIGN_DOWN(avail, fs_info->zone_size);
+
 	return avail;
 }
 
@@ -556,8 +573,7 @@ again:
 
 		spin_lock(&cache->lock);
 		avail = cache->length - cache->used - cache->pinned -
-			cache->reserved - cache->delalloc_bytes -
-			cache->bytes_super - cache->zone_unusable;
+			cache->reserved - cache->bytes_super - cache->zone_unusable;
 		btrfs_info(fs_info,
 "block group %llu has %llu bytes, %llu used %llu pinned %llu reserved %llu delalloc %llu super %llu zone_unusable (%llu bytes available) %s",
 			   cache->start, cache->length, cache->used, cache->pinned,
diff --git a/fs/btrfs/space-info.h b/fs/btrfs/space-info.h
index 92c595fed1..3e304300fc 100644
--- a/fs/btrfs/space-info.h
+++ b/fs/btrfs/space-info.h
@@ -4,8 +4,17 @@
 #define BTRFS_SPACE_INFO_H
 
 #include <trace/events/btrfs.h>
+#include <linux/spinlock.h>
+#include <linux/list.h>
+#include <linux/kobject.h>
+#include <linux/lockdep.h>
+#include <linux/wait.h>
+#include <linux/rwsem.h>
 #include "volumes.h"
 
+struct btrfs_fs_info;
+struct btrfs_block_group;
+
 /*
  * Different levels for to flush space when doing space reservations.
  *
@@ -198,6 +207,7 @@ btrfs_space_info_update_##name(struct btrfs_fs_info *fs_info,		\
 
 DECLARE_SPACE_INFO_UPDATE(bytes_may_use, "space_info");
 DECLARE_SPACE_INFO_UPDATE(bytes_pinned, "pinned");
+DECLARE_SPACE_INFO_UPDATE(bytes_zone_unusable, "zone_unusable");
 
 int btrfs_init_space_info(struct btrfs_fs_info *fs_info);
 void btrfs_add_bg_to_space_info(struct btrfs_fs_info *info,
diff --git a/fs/btrfs/subpage.c b/fs/btrfs/subpage.c
index 0e49dab8da..54736f6238 100644
--- a/fs/btrfs/subpage.c
+++ b/fs/btrfs/subpage.c
@@ -111,6 +111,9 @@ void btrfs_init_subpage_info(struct btrfs_subpage_info *subpage_info, u32 sector
 	subpage_info->checked_offset = cur;
 	cur += nr_bits;
 
+	subpage_info->locked_offset = cur;
+	cur += nr_bits;
+
 	subpage_info->total_nr_bits = cur;
 }
 
@@ -237,28 +240,58 @@ static void btrfs_subpage_assert(const struct btrfs_fs_info *fs_info,
 		       start + len <= folio_pos(folio) + PAGE_SIZE);
 }
 
+#define subpage_calc_start_bit(fs_info, folio, name, start, len)	\
+({									\
+	unsigned int start_bit;						\
+									\
+	btrfs_subpage_assert(fs_info, folio, start, len);		\
+	start_bit = offset_in_page(start) >> fs_info->sectorsize_bits;	\
+	start_bit += fs_info->subpage_info->name##_offset;		\
+	start_bit;							\
+})
+
 void btrfs_subpage_start_reader(const struct btrfs_fs_info *fs_info,
 				struct folio *folio, u64 start, u32 len)
 {
 	struct btrfs_subpage *subpage = folio_get_private(folio);
+	const int start_bit = subpage_calc_start_bit(fs_info, folio, locked, start, len);
 	const int nbits = len >> fs_info->sectorsize_bits;
+	unsigned long flags;
+
 
 	btrfs_subpage_assert(fs_info, folio, start, len);
 
+	spin_lock_irqsave(&subpage->lock, flags);
+	/*
+	 * Even though it's just for reading the page, no one should have
+	 * locked the subpage range.
+	 */
+	ASSERT(bitmap_test_range_all_zero(subpage->bitmaps, start_bit, nbits));
+	bitmap_set(subpage->bitmaps, start_bit, nbits);
 	atomic_add(nbits, &subpage->readers);
+	spin_unlock_irqrestore(&subpage->lock, flags);
 }
 
 void btrfs_subpage_end_reader(const struct btrfs_fs_info *fs_info,
 			      struct folio *folio, u64 start, u32 len)
 {
 	struct btrfs_subpage *subpage = folio_get_private(folio);
+	const int start_bit = subpage_calc_start_bit(fs_info, folio, locked, start, len);
 	const int nbits = len >> fs_info->sectorsize_bits;
+	unsigned long flags;
 	bool is_data;
 	bool last;
 
 	btrfs_subpage_assert(fs_info, folio, start, len);
 	is_data = is_data_inode(folio->mapping->host);
+
+	spin_lock_irqsave(&subpage->lock, flags);
+
+	/* The range should have already been locked. */
+	ASSERT(bitmap_test_range_all_set(subpage->bitmaps, start_bit, nbits));
 	ASSERT(atomic_read(&subpage->readers) >= nbits);
+
+	bitmap_clear(subpage->bitmaps, start_bit, nbits);
 	last = atomic_sub_and_test(nbits, &subpage->readers);
 
 	/*
@@ -270,6 +303,7 @@ void btrfs_subpage_end_reader(const struct btrfs_fs_info *fs_info,
 	 */
 	if (is_data && last)
 		folio_unlock(folio);
+	spin_unlock_irqrestore(&subpage->lock, flags);
 }
 
 static void btrfs_subpage_clamp_range(struct folio *folio, u64 *start, u32 *len)
@@ -290,28 +324,38 @@ static void btrfs_subpage_clamp_range(struct folio *folio, u64 *start, u32 *len)
 			     orig_start + orig_len) - *start;
 }
 
-void btrfs_subpage_start_writer(const struct btrfs_fs_info *fs_info,
-				struct folio *folio, u64 start, u32 len)
+static void btrfs_subpage_start_writer(const struct btrfs_fs_info *fs_info,
+				       struct folio *folio, u64 start, u32 len)
 {
 	struct btrfs_subpage *subpage = folio_get_private(folio);
+	const int start_bit = subpage_calc_start_bit(fs_info, folio, locked, start, len);
 	const int nbits = (len >> fs_info->sectorsize_bits);
+	unsigned long flags;
 	int ret;
 
 	btrfs_subpage_assert(fs_info, folio, start, len);
 
+	spin_lock_irqsave(&subpage->lock, flags);
 	ASSERT(atomic_read(&subpage->readers) == 0);
+	ASSERT(bitmap_test_range_all_zero(subpage->bitmaps, start_bit, nbits));
+	bitmap_set(subpage->bitmaps, start_bit, nbits);
 	ret = atomic_add_return(nbits, &subpage->writers);
 	ASSERT(ret == nbits);
+	spin_unlock_irqrestore(&subpage->lock, flags);
 }
 
-bool btrfs_subpage_end_and_test_writer(const struct btrfs_fs_info *fs_info,
-				       struct folio *folio, u64 start, u32 len)
+static bool btrfs_subpage_end_and_test_writer(const struct btrfs_fs_info *fs_info,
+					      struct folio *folio, u64 start, u32 len)
 {
 	struct btrfs_subpage *subpage = folio_get_private(folio);
+	const int start_bit = subpage_calc_start_bit(fs_info, folio, locked, start, len);
 	const int nbits = (len >> fs_info->sectorsize_bits);
+	unsigned long flags;
+	bool last;
 
 	btrfs_subpage_assert(fs_info, folio, start, len);
 
+	spin_lock_irqsave(&subpage->lock, flags);
 	/*
 	 * We have call sites passing @lock_page into
 	 * extent_clear_unlock_delalloc() for compression path.
@@ -319,11 +363,18 @@ bool btrfs_subpage_end_and_test_writer(const struct btrfs_fs_info *fs_info,
 	 * This @locked_page is locked by plain lock_page(), thus its
 	 * subpage::writers is 0.  Handle them in a special way.
 	 */
-	if (atomic_read(&subpage->writers) == 0)
+	if (atomic_read(&subpage->writers) == 0) {
+		spin_unlock_irqrestore(&subpage->lock, flags);
 		return true;
+	}
 
 	ASSERT(atomic_read(&subpage->writers) >= nbits);
-	return atomic_sub_and_test(nbits, &subpage->writers);
+	/* The target range should have been locked. */
+	ASSERT(bitmap_test_range_all_set(subpage->bitmaps, start_bit, nbits));
+	bitmap_clear(subpage->bitmaps, start_bit, nbits);
+	last = atomic_sub_and_test(nbits, &subpage->writers);
+	spin_unlock_irqrestore(&subpage->lock, flags);
+	return last;
 }
 
 /*
@@ -365,16 +416,6 @@ void btrfs_folio_end_writer_lock(const struct btrfs_fs_info *fs_info,
 		folio_unlock(folio);
 }
 
-#define subpage_calc_start_bit(fs_info, folio, name, start, len)	\
-({									\
-	unsigned int start_bit;						\
-									\
-	btrfs_subpage_assert(fs_info, folio, start, len);		\
-	start_bit = offset_in_page(start) >> fs_info->sectorsize_bits;	\
-	start_bit += fs_info->subpage_info->name##_offset;		\
-	start_bit;							\
-})
-
 #define subpage_test_bitmap_all_set(fs_info, subpage, name)		\
 	bitmap_test_range_all_set(subpage->bitmaps,			\
 			fs_info->subpage_info->name##_offset,		\
@@ -751,6 +792,7 @@ void __cold btrfs_subpage_dump_bitmap(const struct btrfs_fs_info *fs_info,
 	GET_SUBPAGE_BITMAP(subpage, subpage_info, writeback, &writeback_bitmap);
 	GET_SUBPAGE_BITMAP(subpage, subpage_info, ordered, &ordered_bitmap);
 	GET_SUBPAGE_BITMAP(subpage, subpage_info, checked, &checked_bitmap);
+	GET_SUBPAGE_BITMAP(subpage, subpage_info, locked, &checked_bitmap);
 	spin_unlock_irqrestore(&subpage->lock, flags);
 
 	dump_page(folio_page(folio, 0), "btrfs subpage dump");
diff --git a/fs/btrfs/subpage.h b/fs/btrfs/subpage.h
index 793c2b314a..b6dc013b0f 100644
--- a/fs/btrfs/subpage.h
+++ b/fs/btrfs/subpage.h
@@ -4,6 +4,11 @@
 #define BTRFS_SUBPAGE_H
 
 #include <linux/spinlock.h>
+#include <linux/atomic.h>
+
+struct address_space;
+struct folio;
+struct btrfs_fs_info;
 
 /*
  * Extra info for subpapge bitmap.
@@ -28,7 +33,7 @@ struct btrfs_subpage_info {
 	unsigned int total_nr_bits;
 
 	/*
-	 * *_start indicates where the bitmap starts, the length is always
+	 * *_offset indicates where the bitmap starts, the length is always
 	 * @bitmap_size, which is calculated from PAGE_SIZE / sectorsize.
 	 */
 	unsigned int uptodate_offset;
@@ -36,6 +41,16 @@ struct btrfs_subpage_info {
 	unsigned int writeback_offset;
 	unsigned int ordered_offset;
 	unsigned int checked_offset;
+
+	/*
+	 * For locked bitmaps, normally it's subpage representation for folio
+	 * Locked flag, but metadata is different:
+	 *
+	 * - Metadata doesn't really lock the folio
+	 *   It's just to prevent page::private get cleared before the last
+	 *   end_page_read().
+	 */
+	unsigned int locked_offset;
 };
 
 /*
@@ -93,10 +108,6 @@ void btrfs_subpage_start_reader(const struct btrfs_fs_info *fs_info,
 void btrfs_subpage_end_reader(const struct btrfs_fs_info *fs_info,
 			      struct folio *folio, u64 start, u32 len);
 
-void btrfs_subpage_start_writer(const struct btrfs_fs_info *fs_info,
-				struct folio *folio, u64 start, u32 len);
-bool btrfs_subpage_end_and_test_writer(const struct btrfs_fs_info *fs_info,
-				       struct folio *folio, u64 start, u32 len);
 int btrfs_folio_start_writer_lock(const struct btrfs_fs_info *fs_info,
 				  struct folio *folio, u64 start, u32 len);
 void btrfs_folio_end_writer_lock(const struct btrfs_fs_info *fs_info,
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index c45fdaf24c..f05cce7c8b 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -34,13 +34,11 @@
 #include "disk-io.h"
 #include "transaction.h"
 #include "btrfs_inode.h"
-#include "print-tree.h"
 #include "props.h"
 #include "xattr.h"
 #include "bio.h"
 #include "export.h"
 #include "compression.h"
-#include "rcu-string.h"
 #include "dev-replace.h"
 #include "free-space-cache.h"
 #include "backref.h"
@@ -121,6 +119,7 @@ enum {
 	Opt_thread_pool,
 	Opt_treelog,
 	Opt_user_subvol_rm_allowed,
+	Opt_norecovery,
 
 	/* Rescue options */
 	Opt_rescue,
@@ -247,6 +246,8 @@ static const struct fs_parameter_spec btrfs_fs_parameters[] = {
 	__fsparam(NULL, "nologreplay", Opt_nologreplay, fs_param_deprecated, NULL),
 	/* Deprecated, with alias rescue=usebackuproot */
 	__fsparam(NULL, "usebackuproot", Opt_usebackuproot, fs_param_deprecated, NULL),
+	/* For compatibility only, alias for "rescue=nologreplay". */
+	fsparam_flag("norecovery", Opt_norecovery),
 
 	/* Debugging options. */
 	fsparam_flag_no("enospc_debug", Opt_enospc_debug),
@@ -440,6 +441,11 @@ static int btrfs_parse_param(struct fs_context *fc, struct fs_parameter *param)
 		"'nologreplay' is deprecated, use 'rescue=nologreplay' instead");
 		btrfs_set_opt(ctx->mount_opt, NOLOGREPLAY);
 		break;
+	case Opt_norecovery:
+		btrfs_info(NULL,
+"'norecovery' is for compatibility only, recommended to use 'rescue=nologreplay'");
+		btrfs_set_opt(ctx->mount_opt, NOLOGREPLAY);
+		break;
 	case Opt_flushoncommit:
 		if (result.negated)
 			btrfs_clear_opt(ctx->mount_opt, FLUSHONCOMMIT);
@@ -1099,10 +1105,9 @@ static int btrfs_show_options(struct seq_file *seq, struct dentry *dentry)
 #endif
 	if (btrfs_test_opt(info, REF_VERIFY))
 		seq_puts(seq, ",ref_verify");
-	seq_printf(seq, ",subvolid=%llu",
-		  BTRFS_I(d_inode(dentry))->root->root_key.objectid);
+	seq_printf(seq, ",subvolid=%llu", btrfs_root_id(BTRFS_I(d_inode(dentry))->root));
 	subvol_name = btrfs_get_subvol_name_from_objectid(info,
-			BTRFS_I(d_inode(dentry))->root->root_key.objectid);
+			btrfs_root_id(BTRFS_I(d_inode(dentry))->root));
 	if (!IS_ERR(subvol_name)) {
 		seq_puts(seq, ",subvol=");
 		seq_escape(seq, subvol_name, " \t\n\\");
@@ -1154,7 +1159,7 @@ static struct dentry *mount_subvol(const char *subvol_name, u64 subvol_objectid,
 		struct super_block *s = root->d_sb;
 		struct btrfs_fs_info *fs_info = btrfs_sb(s);
 		struct inode *root_inode = d_inode(root);
-		u64 root_objectid = BTRFS_I(root_inode)->root->root_key.objectid;
+		u64 root_objectid = btrfs_root_id(BTRFS_I(root_inode)->root);
 
 		ret = 0;
 		if (!is_subvolume_inode(root_inode)) {
@@ -1776,10 +1781,8 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf)
 	buf->f_fsid.val[0] = be32_to_cpu(fsid[0]) ^ be32_to_cpu(fsid[2]);
 	buf->f_fsid.val[1] = be32_to_cpu(fsid[1]) ^ be32_to_cpu(fsid[3]);
 	/* Mask in the root object ID too, to disambiguate subvols */
-	buf->f_fsid.val[0] ^=
-		BTRFS_I(d_inode(dentry))->root->root_key.objectid >> 32;
-	buf->f_fsid.val[1] ^=
-		BTRFS_I(d_inode(dentry))->root->root_key.objectid;
+	buf->f_fsid.val[0] ^= btrfs_root_id(BTRFS_I(d_inode(dentry))->root) >> 32;
+	buf->f_fsid.val[1] ^= btrfs_root_id(BTRFS_I(d_inode(dentry))->root);
 
 	return 0;
 }
@@ -2203,7 +2206,9 @@ static long btrfs_control_ioctl(struct file *file, unsigned int cmd,
 	vol = memdup_user((void __user *)arg, sizeof(*vol));
 	if (IS_ERR(vol))
 		return PTR_ERR(vol);
-	vol->name[BTRFS_PATH_NAME_MAX] = '\0';
+	ret = btrfs_check_ioctl_vol_args_path(vol);
+	if (ret < 0)
+		goto out;
 
 	switch (cmd) {
 	case BTRFS_IOC_SCAN_DEV:
@@ -2245,6 +2250,7 @@ static long btrfs_control_ioctl(struct file *file, unsigned int cmd,
 		break;
 	}
 
+out:
 	kfree(vol);
 	return ret;
 }
@@ -2373,6 +2379,24 @@ static int btrfs_show_devname(struct seq_file *m, struct dentry *root)
 	return 0;
 }
 
+static long btrfs_nr_cached_objects(struct super_block *sb, struct shrink_control *sc)
+{
+	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
+	const s64 nr = percpu_counter_sum_positive(&fs_info->evictable_extent_maps);
+
+	trace_btrfs_extent_map_shrinker_count(fs_info, nr);
+
+	return nr;
+}
+
+static long btrfs_free_cached_objects(struct super_block *sb, struct shrink_control *sc)
+{
+	const long nr_to_scan = min_t(unsigned long, LONG_MAX, sc->nr_to_scan);
+	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
+
+	return btrfs_free_extent_maps(fs_info, nr_to_scan);
+}
+
 static const struct super_operations btrfs_super_ops = {
 	.drop_inode	= btrfs_drop_inode,
 	.evict_inode	= btrfs_evict_inode,
@@ -2386,6 +2410,8 @@ static const struct super_operations btrfs_super_ops = {
 	.statfs		= btrfs_statfs,
 	.freeze_fs	= btrfs_freeze,
 	.unfreeze_fs	= btrfs_unfreeze,
+	.nr_cached_objects = btrfs_nr_cached_objects,
+	.free_cached_objects = btrfs_free_cached_objects,
 };
 
 static const struct file_operations btrfs_ctl_fops = {
diff --git a/fs/btrfs/super.h b/fs/btrfs/super.h
index f18253ca28..cbcab434b5 100644
--- a/fs/btrfs/super.h
+++ b/fs/btrfs/super.h
@@ -3,6 +3,13 @@
 #ifndef BTRFS_SUPER_H
 #define BTRFS_SUPER_H
 
+#include <linux/types.h>
+#include <linux/fs.h>
+#include "fs.h"
+
+struct super_block;
+struct btrfs_fs_info;
+
 bool btrfs_check_options(struct btrfs_fs_info *info, unsigned long *mount_opt,
 			 unsigned long flags);
 int btrfs_sync_fs(struct super_block *sb, int wait);
diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c
index 84c05246ff..af545b6b11 100644
--- a/fs/btrfs/sysfs.c
+++ b/fs/btrfs/sysfs.c
@@ -421,7 +421,7 @@ BTRFS_ATTR(static_feature, supported_sectorsizes,
 
 static ssize_t acl_show(struct kobject *kobj, struct kobj_attribute *a, char *buf)
 {
-	return sysfs_emit(buf, "%d\n", !!IS_ENABLED(CONFIG_BTRFS_FS_POSIX_ACL));
+	return sysfs_emit(buf, "%d\n", IS_ENABLED(CONFIG_BTRFS_FS_POSIX_ACL));
 }
 BTRFS_ATTR(static_feature, acl, acl_show);
 
@@ -1228,11 +1228,12 @@ static ssize_t btrfs_read_policy_show(struct kobject *kobj,
 				      struct kobj_attribute *a, char *buf)
 {
 	struct btrfs_fs_devices *fs_devices = to_fs_devs(kobj);
+	const enum btrfs_read_policy policy = READ_ONCE(fs_devices->read_policy);
 	ssize_t ret = 0;
 	int i;
 
 	for (i = 0; i < BTRFS_NR_READ_POLICY; i++) {
-		if (fs_devices->read_policy == i)
+		if (policy == i)
 			ret += sysfs_emit_at(buf, ret, "%s[%s]",
 					 (ret == 0 ? "" : " "),
 					 btrfs_read_policy_name[i]);
@@ -1256,8 +1257,8 @@ static ssize_t btrfs_read_policy_store(struct kobject *kobj,
 
 	for (i = 0; i < BTRFS_NR_READ_POLICY; i++) {
 		if (sysfs_streq(buf, btrfs_read_policy_name[i])) {
-			if (i != fs_devices->read_policy) {
-				fs_devices->read_policy = i;
+			if (i != READ_ONCE(fs_devices->read_policy)) {
+				WRITE_ONCE(fs_devices->read_policy, i);
 				btrfs_info(fs_devices->fs_info,
 					   "read policy set to '%s'",
 					   btrfs_read_policy_name[i]);
@@ -1306,6 +1307,47 @@ static ssize_t btrfs_bg_reclaim_threshold_store(struct kobject *kobj,
 BTRFS_ATTR_RW(, bg_reclaim_threshold, btrfs_bg_reclaim_threshold_show,
 	      btrfs_bg_reclaim_threshold_store);
 
+#ifdef CONFIG_BTRFS_DEBUG
+static ssize_t btrfs_offload_csum_show(struct kobject *kobj,
+				       struct kobj_attribute *a, char *buf)
+{
+	struct btrfs_fs_devices *fs_devices = to_fs_devs(kobj);
+
+	switch (READ_ONCE(fs_devices->offload_csum_mode)) {
+	case BTRFS_OFFLOAD_CSUM_AUTO:
+		return sysfs_emit(buf, "auto\n");
+	case BTRFS_OFFLOAD_CSUM_FORCE_ON:
+		return sysfs_emit(buf, "1\n");
+	case BTRFS_OFFLOAD_CSUM_FORCE_OFF:
+		return sysfs_emit(buf, "0\n");
+	default:
+		WARN_ON(1);
+		return -EINVAL;
+	}
+}
+
+static ssize_t btrfs_offload_csum_store(struct kobject *kobj,
+					struct kobj_attribute *a, const char *buf,
+					size_t len)
+{
+	struct btrfs_fs_devices *fs_devices = to_fs_devs(kobj);
+	int ret;
+	bool val;
+
+	ret = kstrtobool(buf, &val);
+	if (ret == 0)
+		WRITE_ONCE(fs_devices->offload_csum_mode,
+			   val ? BTRFS_OFFLOAD_CSUM_FORCE_ON : BTRFS_OFFLOAD_CSUM_FORCE_OFF);
+	else if (ret == -EINVAL && sysfs_streq(buf, "auto"))
+		WRITE_ONCE(fs_devices->offload_csum_mode, BTRFS_OFFLOAD_CSUM_AUTO);
+	else
+		return -EINVAL;
+
+	return len;
+}
+BTRFS_ATTR_RW(, offload_csum, btrfs_offload_csum_show, btrfs_offload_csum_store);
+#endif
+
 /*
  * Per-filesystem information and stats.
  *
@@ -1325,6 +1367,9 @@ static const struct attribute *btrfs_attrs[] = {
 	BTRFS_ATTR_PTR(, bg_reclaim_threshold),
 	BTRFS_ATTR_PTR(, commit_stats),
 	BTRFS_ATTR_PTR(, temp_fsid),
+#ifdef CONFIG_BTRFS_DEBUG
+	BTRFS_ATTR_PTR(, offload_csum),
+#endif
 	NULL,
 };
 
@@ -2294,7 +2339,7 @@ int btrfs_sysfs_add_one_qgroup(struct btrfs_fs_info *fs_info,
 	struct kobject *qgroups_kobj = fs_info->qgroups_kobj;
 	int ret;
 
-	if (test_bit(BTRFS_FS_STATE_DUMMY_FS_INFO, &fs_info->fs_state))
+	if (btrfs_is_testing(fs_info))
 		return 0;
 	if (qgroup->kobj.state_initialized)
 		return 0;
@@ -2315,7 +2360,7 @@ void btrfs_sysfs_del_qgroups(struct btrfs_fs_info *fs_info)
 	struct btrfs_qgroup *qgroup;
 	struct btrfs_qgroup *next;
 
-	if (test_bit(BTRFS_FS_STATE_DUMMY_FS_INFO, &fs_info->fs_state))
+	if (btrfs_is_testing(fs_info))
 		return;
 
 	rbtree_postorder_for_each_entry_safe(qgroup, next,
@@ -2336,7 +2381,7 @@ int btrfs_sysfs_add_qgroups(struct btrfs_fs_info *fs_info)
 	struct btrfs_qgroup *next;
 	int ret = 0;
 
-	if (test_bit(BTRFS_FS_STATE_DUMMY_FS_INFO, &fs_info->fs_state))
+	if (btrfs_is_testing(fs_info))
 		return 0;
 
 	ASSERT(fsid_kobj);
@@ -2368,7 +2413,7 @@ out:
 void btrfs_sysfs_del_one_qgroup(struct btrfs_fs_info *fs_info,
 				struct btrfs_qgroup *qgroup)
 {
-	if (test_bit(BTRFS_FS_STATE_DUMMY_FS_INFO, &fs_info->fs_state))
+	if (btrfs_is_testing(fs_info))
 		return;
 
 	if (qgroup->kobj.state_initialized) {
diff --git a/fs/btrfs/sysfs.h b/fs/btrfs/sysfs.h
index 86c7eef128..e6a284c598 100644
--- a/fs/btrfs/sysfs.h
+++ b/fs/btrfs/sysfs.h
@@ -3,8 +3,17 @@
 #ifndef BTRFS_SYSFS_H
 #define BTRFS_SYSFS_H
 
+#include <linux/types.h>
+#include <linux/compiler_types.h>
 #include <linux/kobject.h>
 
+struct btrfs_fs_info;
+struct btrfs_device;
+struct btrfs_fs_devices;
+struct btrfs_block_group;
+struct btrfs_space_info;
+struct btrfs_qgroup;
+
 enum btrfs_feature_set {
 	FEAT_COMPAT,
 	FEAT_COMPAT_RO,
diff --git a/fs/btrfs/tests/btrfs-tests.c b/fs/btrfs/tests/btrfs-tests.c
index 709c6cc970..dce0387ef1 100644
--- a/fs/btrfs/tests/btrfs-tests.c
+++ b/fs/btrfs/tests/btrfs-tests.c
@@ -160,8 +160,7 @@ void btrfs_free_dummy_fs_info(struct btrfs_fs_info *fs_info)
 	if (!fs_info)
 		return;
 
-	if (WARN_ON(!test_bit(BTRFS_FS_STATE_DUMMY_FS_INFO,
-			      &fs_info->fs_state)))
+	if (WARN_ON(!btrfs_is_testing(fs_info)))
 		return;
 
 	test_mnt->mnt_sb->s_fs_info = NULL;
diff --git a/fs/btrfs/tests/extent-map-tests.c b/fs/btrfs/tests/extent-map-tests.c
index 47b5d30103..ba36794ba2 100644
--- a/fs/btrfs/tests/extent-map-tests.c
+++ b/fs/btrfs/tests/extent-map-tests.c
@@ -11,19 +11,22 @@
 #include "../disk-io.h"
 #include "../block-group.h"
 
-static void free_extent_map_tree(struct extent_map_tree *em_tree)
+static int free_extent_map_tree(struct btrfs_inode *inode)
 {
+	struct extent_map_tree *em_tree = &inode->extent_tree;
 	struct extent_map *em;
 	struct rb_node *node;
+	int ret = 0;
 
 	write_lock(&em_tree->lock);
 	while (!RB_EMPTY_ROOT(&em_tree->map.rb_root)) {
 		node = rb_first_cached(&em_tree->map);
 		em = rb_entry(node, struct extent_map, rb_node);
-		remove_extent_mapping(em_tree, em);
+		remove_extent_mapping(inode, em);
 
 #ifdef CONFIG_BTRFS_DEBUG
 		if (refcount_read(&em->refs) != 1) {
+			ret = -EINVAL;
 			test_err(
 "em leak: em (start %llu len %llu block_start %llu block_len %llu) refs %d",
 				 em->start, em->len, em->block_start,
@@ -35,6 +38,8 @@ static void free_extent_map_tree(struct extent_map_tree *em_tree)
 		free_extent_map(em);
 	}
 	write_unlock(&em_tree->lock);
+
+	return ret;
 }
 
 /*
@@ -53,13 +58,14 @@ static void free_extent_map_tree(struct extent_map_tree *em_tree)
  *                                    ->add_extent_mapping(0, 16K)
  *                                    -> #handle -EEXIST
  */
-static int test_case_1(struct btrfs_fs_info *fs_info,
-		struct extent_map_tree *em_tree)
+static int test_case_1(struct btrfs_fs_info *fs_info, struct btrfs_inode *inode)
 {
+	struct extent_map_tree *em_tree = &inode->extent_tree;
 	struct extent_map *em;
 	u64 start = 0;
 	u64 len = SZ_8K;
 	int ret;
+	int ret2;
 
 	em = alloc_extent_map();
 	if (!em) {
@@ -73,7 +79,7 @@ static int test_case_1(struct btrfs_fs_info *fs_info,
 	em->block_start = 0;
 	em->block_len = SZ_16K;
 	write_lock(&em_tree->lock);
-	ret = btrfs_add_extent_mapping(fs_info, em_tree, &em, em->start, em->len);
+	ret = btrfs_add_extent_mapping(inode, &em, em->start, em->len);
 	write_unlock(&em_tree->lock);
 	if (ret < 0) {
 		test_err("cannot add extent range [0, 16K)");
@@ -94,7 +100,7 @@ static int test_case_1(struct btrfs_fs_info *fs_info,
 	em->block_start = SZ_32K; /* avoid merging */
 	em->block_len = SZ_4K;
 	write_lock(&em_tree->lock);
-	ret = btrfs_add_extent_mapping(fs_info, em_tree, &em, em->start, em->len);
+	ret = btrfs_add_extent_mapping(inode, &em, em->start, em->len);
 	write_unlock(&em_tree->lock);
 	if (ret < 0) {
 		test_err("cannot add extent range [16K, 20K)");
@@ -115,7 +121,7 @@ static int test_case_1(struct btrfs_fs_info *fs_info,
 	em->block_start = start;
 	em->block_len = len;
 	write_lock(&em_tree->lock);
-	ret = btrfs_add_extent_mapping(fs_info, em_tree, &em, em->start, em->len);
+	ret = btrfs_add_extent_mapping(inode, &em, em->start, em->len);
 	write_unlock(&em_tree->lock);
 	if (ret) {
 		test_err("case1 [%llu %llu]: ret %d", start, start + len, ret);
@@ -137,7 +143,9 @@ static int test_case_1(struct btrfs_fs_info *fs_info,
 	}
 	free_extent_map(em);
 out:
-	free_extent_map_tree(em_tree);
+	ret2 = free_extent_map_tree(inode);
+	if (ret == 0)
+		ret = ret2;
 
 	return ret;
 }
@@ -148,11 +156,12 @@ out:
  * Reading the inline ending up with EEXIST, ie. read an inline
  * extent and discard page cache and read it again.
  */
-static int test_case_2(struct btrfs_fs_info *fs_info,
-		struct extent_map_tree *em_tree)
+static int test_case_2(struct btrfs_fs_info *fs_info, struct btrfs_inode *inode)
 {
+	struct extent_map_tree *em_tree = &inode->extent_tree;
 	struct extent_map *em;
 	int ret;
+	int ret2;
 
 	em = alloc_extent_map();
 	if (!em) {
@@ -166,7 +175,7 @@ static int test_case_2(struct btrfs_fs_info *fs_info,
 	em->block_start = EXTENT_MAP_INLINE;
 	em->block_len = (u64)-1;
 	write_lock(&em_tree->lock);
-	ret = btrfs_add_extent_mapping(fs_info, em_tree, &em, em->start, em->len);
+	ret = btrfs_add_extent_mapping(inode, &em, em->start, em->len);
 	write_unlock(&em_tree->lock);
 	if (ret < 0) {
 		test_err("cannot add extent range [0, 1K)");
@@ -187,7 +196,7 @@ static int test_case_2(struct btrfs_fs_info *fs_info,
 	em->block_start = SZ_4K;
 	em->block_len = SZ_4K;
 	write_lock(&em_tree->lock);
-	ret = btrfs_add_extent_mapping(fs_info, em_tree, &em, em->start, em->len);
+	ret = btrfs_add_extent_mapping(inode, &em, em->start, em->len);
 	write_unlock(&em_tree->lock);
 	if (ret < 0) {
 		test_err("cannot add extent range [4K, 8K)");
@@ -208,7 +217,7 @@ static int test_case_2(struct btrfs_fs_info *fs_info,
 	em->block_start = EXTENT_MAP_INLINE;
 	em->block_len = (u64)-1;
 	write_lock(&em_tree->lock);
-	ret = btrfs_add_extent_mapping(fs_info, em_tree, &em, em->start, em->len);
+	ret = btrfs_add_extent_mapping(inode, &em, em->start, em->len);
 	write_unlock(&em_tree->lock);
 	if (ret) {
 		test_err("case2 [0 1K]: ret %d", ret);
@@ -229,17 +238,21 @@ static int test_case_2(struct btrfs_fs_info *fs_info,
 	}
 	free_extent_map(em);
 out:
-	free_extent_map_tree(em_tree);
+	ret2 = free_extent_map_tree(inode);
+	if (ret == 0)
+		ret = ret2;
 
 	return ret;
 }
 
 static int __test_case_3(struct btrfs_fs_info *fs_info,
-		struct extent_map_tree *em_tree, u64 start)
+			 struct btrfs_inode *inode, u64 start)
 {
+	struct extent_map_tree *em_tree = &inode->extent_tree;
 	struct extent_map *em;
 	u64 len = SZ_4K;
 	int ret;
+	int ret2;
 
 	em = alloc_extent_map();
 	if (!em) {
@@ -253,7 +266,7 @@ static int __test_case_3(struct btrfs_fs_info *fs_info,
 	em->block_start = SZ_4K;
 	em->block_len = SZ_4K;
 	write_lock(&em_tree->lock);
-	ret = btrfs_add_extent_mapping(fs_info, em_tree, &em, em->start, em->len);
+	ret = btrfs_add_extent_mapping(inode, &em, em->start, em->len);
 	write_unlock(&em_tree->lock);
 	if (ret < 0) {
 		test_err("cannot add extent range [4K, 8K)");
@@ -274,7 +287,7 @@ static int __test_case_3(struct btrfs_fs_info *fs_info,
 	em->block_start = 0;
 	em->block_len = SZ_16K;
 	write_lock(&em_tree->lock);
-	ret = btrfs_add_extent_mapping(fs_info, em_tree, &em, start, len);
+	ret = btrfs_add_extent_mapping(inode, &em, start, len);
 	write_unlock(&em_tree->lock);
 	if (ret) {
 		test_err("case3 [%llu %llu): ret %d",
@@ -301,7 +314,9 @@ static int __test_case_3(struct btrfs_fs_info *fs_info,
 	}
 	free_extent_map(em);
 out:
-	free_extent_map_tree(em_tree);
+	ret2 = free_extent_map_tree(inode);
+	if (ret == 0)
+		ret = ret2;
 
 	return ret;
 }
@@ -322,28 +337,29 @@ out:
  *   -> add_extent_mapping()
  *                            -> add_extent_mapping()
  */
-static int test_case_3(struct btrfs_fs_info *fs_info,
-		struct extent_map_tree *em_tree)
+static int test_case_3(struct btrfs_fs_info *fs_info, struct btrfs_inode *inode)
 {
 	int ret;
 
-	ret = __test_case_3(fs_info, em_tree, 0);
+	ret = __test_case_3(fs_info, inode, 0);
 	if (ret)
 		return ret;
-	ret = __test_case_3(fs_info, em_tree, SZ_8K);
+	ret = __test_case_3(fs_info, inode, SZ_8K);
 	if (ret)
 		return ret;
-	ret = __test_case_3(fs_info, em_tree, (12 * SZ_1K));
+	ret = __test_case_3(fs_info, inode, (12 * SZ_1K));
 
 	return ret;
 }
 
 static int __test_case_4(struct btrfs_fs_info *fs_info,
-		struct extent_map_tree *em_tree, u64 start)
+			 struct btrfs_inode *inode, u64 start)
 {
+	struct extent_map_tree *em_tree = &inode->extent_tree;
 	struct extent_map *em;
 	u64 len = SZ_4K;
 	int ret;
+	int ret2;
 
 	em = alloc_extent_map();
 	if (!em) {
@@ -357,7 +373,7 @@ static int __test_case_4(struct btrfs_fs_info *fs_info,
 	em->block_start = 0;
 	em->block_len = SZ_8K;
 	write_lock(&em_tree->lock);
-	ret = btrfs_add_extent_mapping(fs_info, em_tree, &em, em->start, em->len);
+	ret = btrfs_add_extent_mapping(inode, &em, em->start, em->len);
 	write_unlock(&em_tree->lock);
 	if (ret < 0) {
 		test_err("cannot add extent range [0, 8K)");
@@ -378,7 +394,7 @@ static int __test_case_4(struct btrfs_fs_info *fs_info,
 	em->block_start = SZ_16K; /* avoid merging */
 	em->block_len = 24 * SZ_1K;
 	write_lock(&em_tree->lock);
-	ret = btrfs_add_extent_mapping(fs_info, em_tree, &em, em->start, em->len);
+	ret = btrfs_add_extent_mapping(inode, &em, em->start, em->len);
 	write_unlock(&em_tree->lock);
 	if (ret < 0) {
 		test_err("cannot add extent range [8K, 32K)");
@@ -398,7 +414,7 @@ static int __test_case_4(struct btrfs_fs_info *fs_info,
 	em->block_start = 0;
 	em->block_len = SZ_32K;
 	write_lock(&em_tree->lock);
-	ret = btrfs_add_extent_mapping(fs_info, em_tree, &em, start, len);
+	ret = btrfs_add_extent_mapping(inode, &em, start, len);
 	write_unlock(&em_tree->lock);
 	if (ret) {
 		test_err("case4 [%llu %llu): ret %d",
@@ -420,7 +436,9 @@ static int __test_case_4(struct btrfs_fs_info *fs_info,
 	}
 	free_extent_map(em);
 out:
-	free_extent_map_tree(em_tree);
+	ret2 = free_extent_map_tree(inode);
+	if (ret == 0)
+		ret = ret2;
 
 	return ret;
 }
@@ -450,23 +468,22 @@ out:
  *                                             # handle -EEXIST when adding
  *                                             # [0, 32K)
  */
-static int test_case_4(struct btrfs_fs_info *fs_info,
-		struct extent_map_tree *em_tree)
+static int test_case_4(struct btrfs_fs_info *fs_info, struct btrfs_inode *inode)
 {
 	int ret;
 
-	ret = __test_case_4(fs_info, em_tree, 0);
+	ret = __test_case_4(fs_info, inode, 0);
 	if (ret)
 		return ret;
-	ret = __test_case_4(fs_info, em_tree, SZ_4K);
+	ret = __test_case_4(fs_info, inode, SZ_4K);
 
 	return ret;
 }
 
-static int add_compressed_extent(struct btrfs_fs_info *fs_info,
-				 struct extent_map_tree *em_tree,
+static int add_compressed_extent(struct btrfs_inode *inode,
 				 u64 start, u64 len, u64 block_start)
 {
+	struct extent_map_tree *em_tree = &inode->extent_tree;
 	struct extent_map *em;
 	int ret;
 
@@ -482,7 +499,7 @@ static int add_compressed_extent(struct btrfs_fs_info *fs_info,
 	em->block_len = SZ_4K;
 	em->flags |= EXTENT_FLAG_COMPRESS_ZLIB;
 	write_lock(&em_tree->lock);
-	ret = btrfs_add_extent_mapping(fs_info, em_tree, &em, em->start, em->len);
+	ret = btrfs_add_extent_mapping(inode, &em, em->start, em->len);
 	write_unlock(&em_tree->lock);
 	free_extent_map(em);
 	if (ret < 0) {
@@ -588,53 +605,44 @@ static int validate_range(struct extent_map_tree *em_tree, int index)
  * They'll have the EXTENT_FLAG_COMPRESSED flag set to keep the em tree from
  * merging the em's.
  */
-static int test_case_5(struct btrfs_fs_info *fs_info)
+static int test_case_5(struct btrfs_fs_info *fs_info, struct btrfs_inode *inode)
 {
-	struct extent_map_tree *em_tree;
-	struct inode *inode;
 	u64 start, end;
 	int ret;
+	int ret2;
 
 	test_msg("Running btrfs_drop_extent_map_range tests");
 
-	inode = btrfs_new_test_inode();
-	if (!inode) {
-		test_std_err(TEST_ALLOC_INODE);
-		return -ENOMEM;
-	}
-
-	em_tree = &BTRFS_I(inode)->extent_tree;
-
 	/* [0, 12k) */
-	ret = add_compressed_extent(fs_info, em_tree, 0, SZ_4K * 3, 0);
+	ret = add_compressed_extent(inode, 0, SZ_4K * 3, 0);
 	if (ret) {
 		test_err("cannot add extent range [0, 12K)");
 		goto out;
 	}
 
 	/* [12k, 24k) */
-	ret = add_compressed_extent(fs_info, em_tree, SZ_4K * 3, SZ_4K * 3, SZ_4K);
+	ret = add_compressed_extent(inode, SZ_4K * 3, SZ_4K * 3, SZ_4K);
 	if (ret) {
 		test_err("cannot add extent range [12k, 24k)");
 		goto out;
 	}
 
 	/* [24k, 36k) */
-	ret = add_compressed_extent(fs_info, em_tree, SZ_4K * 6, SZ_4K * 3, SZ_8K);
+	ret = add_compressed_extent(inode, SZ_4K * 6, SZ_4K * 3, SZ_8K);
 	if (ret) {
 		test_err("cannot add extent range [12k, 24k)");
 		goto out;
 	}
 
 	/* [36k, 40k) */
-	ret = add_compressed_extent(fs_info, em_tree, SZ_32K + SZ_4K, SZ_4K, SZ_4K * 3);
+	ret = add_compressed_extent(inode, SZ_32K + SZ_4K, SZ_4K, SZ_4K * 3);
 	if (ret) {
 		test_err("cannot add extent range [12k, 24k)");
 		goto out;
 	}
 
 	/* [40k, 64k) */
-	ret = add_compressed_extent(fs_info, em_tree, SZ_4K * 10, SZ_4K * 6, SZ_16K);
+	ret = add_compressed_extent(inode, SZ_4K * 10, SZ_4K * 6, SZ_16K);
 	if (ret) {
 		test_err("cannot add extent range [12k, 24k)");
 		goto out;
@@ -643,36 +651,39 @@ static int test_case_5(struct btrfs_fs_info *fs_info)
 	/* Drop [8k, 12k) */
 	start = SZ_8K;
 	end = (3 * SZ_4K) - 1;
-	btrfs_drop_extent_map_range(BTRFS_I(inode), start, end, false);
-	ret = validate_range(&BTRFS_I(inode)->extent_tree, 0);
+	btrfs_drop_extent_map_range(inode, start, end, false);
+	ret = validate_range(&inode->extent_tree, 0);
 	if (ret)
 		goto out;
 
 	/* Drop [12k, 20k) */
 	start = SZ_4K * 3;
 	end = SZ_16K + SZ_4K - 1;
-	btrfs_drop_extent_map_range(BTRFS_I(inode), start, end, false);
-	ret = validate_range(&BTRFS_I(inode)->extent_tree, 1);
+	btrfs_drop_extent_map_range(inode, start, end, false);
+	ret = validate_range(&inode->extent_tree, 1);
 	if (ret)
 		goto out;
 
 	/* Drop [28k, 32k) */
 	start = SZ_32K - SZ_4K;
 	end = SZ_32K - 1;
-	btrfs_drop_extent_map_range(BTRFS_I(inode), start, end, false);
-	ret = validate_range(&BTRFS_I(inode)->extent_tree, 2);
+	btrfs_drop_extent_map_range(inode, start, end, false);
+	ret = validate_range(&inode->extent_tree, 2);
 	if (ret)
 		goto out;
 
 	/* Drop [32k, 64k) */
 	start = SZ_32K;
 	end = SZ_64K - 1;
-	btrfs_drop_extent_map_range(BTRFS_I(inode), start, end, false);
-	ret = validate_range(&BTRFS_I(inode)->extent_tree, 3);
+	btrfs_drop_extent_map_range(inode, start, end, false);
+	ret = validate_range(&inode->extent_tree, 3);
 	if (ret)
 		goto out;
 out:
-	iput(inode);
+	ret2 = free_extent_map_tree(inode);
+	if (ret == 0)
+		ret = ret2;
+
 	return ret;
 }
 
@@ -681,23 +692,26 @@ out:
  * for areas between two existing ems.  Validate it doesn't do this when there
  * are two unmerged em's side by side.
  */
-static int test_case_6(struct btrfs_fs_info *fs_info, struct extent_map_tree *em_tree)
+static int test_case_6(struct btrfs_fs_info *fs_info, struct btrfs_inode *inode)
 {
+	struct extent_map_tree *em_tree = &inode->extent_tree;
 	struct extent_map *em = NULL;
 	int ret;
+	int ret2;
 
-	ret = add_compressed_extent(fs_info, em_tree, 0, SZ_4K, 0);
+	ret = add_compressed_extent(inode, 0, SZ_4K, 0);
 	if (ret)
 		goto out;
 
-	ret = add_compressed_extent(fs_info, em_tree, SZ_4K, SZ_4K, 0);
+	ret = add_compressed_extent(inode, SZ_4K, SZ_4K, 0);
 	if (ret)
 		goto out;
 
 	em = alloc_extent_map();
 	if (!em) {
 		test_std_err(TEST_ALLOC_EXTENT_MAP);
-		return -ENOMEM;
+		ret = -ENOMEM;
+		goto out;
 	}
 
 	em->start = SZ_4K;
@@ -705,7 +719,7 @@ static int test_case_6(struct btrfs_fs_info *fs_info, struct extent_map_tree *em
 	em->block_start = SZ_16K;
 	em->block_len = SZ_16K;
 	write_lock(&em_tree->lock);
-	ret = btrfs_add_extent_mapping(fs_info, em_tree, &em, 0, SZ_8K);
+	ret = btrfs_add_extent_mapping(inode, &em, 0, SZ_8K);
 	write_unlock(&em_tree->lock);
 
 	if (ret != 0) {
@@ -725,7 +739,10 @@ static int test_case_6(struct btrfs_fs_info *fs_info, struct extent_map_tree *em
 	ret = 0;
 out:
 	free_extent_map(em);
-	free_extent_map_tree(em_tree);
+	ret2 = free_extent_map_tree(inode);
+	if (ret == 0)
+		ret = ret2;
+
 	return ret;
 }
 
@@ -734,28 +751,19 @@ out:
  * true would mess up the start/end calculations and subsequent splits would be
  * incorrect.
  */
-static int test_case_7(struct btrfs_fs_info *fs_info)
+static int test_case_7(struct btrfs_fs_info *fs_info, struct btrfs_inode *inode)
 {
-	struct extent_map_tree *em_tree;
+	struct extent_map_tree *em_tree = &inode->extent_tree;
 	struct extent_map *em;
-	struct inode *inode;
 	int ret;
+	int ret2;
 
 	test_msg("Running btrfs_drop_extent_cache with pinned");
 
-	inode = btrfs_new_test_inode();
-	if (!inode) {
-		test_std_err(TEST_ALLOC_INODE);
-		return -ENOMEM;
-	}
-
-	em_tree = &BTRFS_I(inode)->extent_tree;
-
 	em = alloc_extent_map();
 	if (!em) {
 		test_std_err(TEST_ALLOC_EXTENT_MAP);
-		ret = -ENOMEM;
-		goto out;
+		return -ENOMEM;
 	}
 
 	/* [0, 16K), pinned */
@@ -765,7 +773,7 @@ static int test_case_7(struct btrfs_fs_info *fs_info)
 	em->block_len = SZ_4K;
 	em->flags |= EXTENT_FLAG_PINNED;
 	write_lock(&em_tree->lock);
-	ret = btrfs_add_extent_mapping(fs_info, em_tree, &em, em->start, em->len);
+	ret = btrfs_add_extent_mapping(inode, &em, em->start, em->len);
 	write_unlock(&em_tree->lock);
 	if (ret < 0) {
 		test_err("couldn't add extent map");
@@ -786,7 +794,7 @@ static int test_case_7(struct btrfs_fs_info *fs_info)
 	em->block_start = SZ_32K;
 	em->block_len = SZ_16K;
 	write_lock(&em_tree->lock);
-	ret = btrfs_add_extent_mapping(fs_info, em_tree, &em, em->start, em->len);
+	ret = btrfs_add_extent_mapping(inode, &em, em->start, em->len);
 	write_unlock(&em_tree->lock);
 	if (ret < 0) {
 		test_err("couldn't add extent map");
@@ -798,7 +806,7 @@ static int test_case_7(struct btrfs_fs_info *fs_info)
 	 * Drop [0, 36K) This should skip the [0, 4K) extent and then split the
 	 * [32K, 48K) extent.
 	 */
-	btrfs_drop_extent_map_range(BTRFS_I(inode), 0, (36 * SZ_1K) - 1, true);
+	btrfs_drop_extent_map_range(inode, 0, (36 * SZ_1K) - 1, true);
 
 	/* Make sure our extent maps look sane. */
 	ret = -EINVAL;
@@ -865,7 +873,14 @@ static int test_case_7(struct btrfs_fs_info *fs_info)
 	ret = 0;
 out:
 	free_extent_map(em);
-	iput(inode);
+	/* Unpin our extent to prevent warning when removing it below. */
+	ret2 = unpin_extent_cache(inode, 0, SZ_16K, 0);
+	if (ret == 0)
+		ret = ret2;
+	ret2 = free_extent_map_tree(inode);
+	if (ret == 0)
+		ret = ret2;
+
 	return ret;
 }
 
@@ -959,7 +974,8 @@ out_free:
 int btrfs_test_extent_map(void)
 {
 	struct btrfs_fs_info *fs_info = NULL;
-	struct extent_map_tree *em_tree;
+	struct inode *inode;
+	struct btrfs_root *root = NULL;
 	int ret = 0, i;
 	struct rmap_test_vector rmap_tests[] = {
 		{
@@ -1008,33 +1024,42 @@ int btrfs_test_extent_map(void)
 		return -ENOMEM;
 	}
 
-	em_tree = kzalloc(sizeof(*em_tree), GFP_KERNEL);
-	if (!em_tree) {
+	inode = btrfs_new_test_inode();
+	if (!inode) {
+		test_std_err(TEST_ALLOC_INODE);
 		ret = -ENOMEM;
 		goto out;
 	}
 
-	extent_map_tree_init(em_tree);
+	root = btrfs_alloc_dummy_root(fs_info);
+	if (IS_ERR(root)) {
+		test_std_err(TEST_ALLOC_ROOT);
+		ret = PTR_ERR(root);
+		root = NULL;
+		goto out;
+	}
 
-	ret = test_case_1(fs_info, em_tree);
+	BTRFS_I(inode)->root = root;
+
+	ret = test_case_1(fs_info, BTRFS_I(inode));
 	if (ret)
 		goto out;
-	ret = test_case_2(fs_info, em_tree);
+	ret = test_case_2(fs_info, BTRFS_I(inode));
 	if (ret)
 		goto out;
-	ret = test_case_3(fs_info, em_tree);
+	ret = test_case_3(fs_info, BTRFS_I(inode));
 	if (ret)
 		goto out;
-	ret = test_case_4(fs_info, em_tree);
+	ret = test_case_4(fs_info, BTRFS_I(inode));
 	if (ret)
 		goto out;
-	ret = test_case_5(fs_info);
+	ret = test_case_5(fs_info, BTRFS_I(inode));
 	if (ret)
 		goto out;
-	ret = test_case_6(fs_info, em_tree);
+	ret = test_case_6(fs_info, BTRFS_I(inode));
 	if (ret)
 		goto out;
-	ret = test_case_7(fs_info);
+	ret = test_case_7(fs_info, BTRFS_I(inode));
 	if (ret)
 		goto out;
 
@@ -1046,7 +1071,8 @@ int btrfs_test_extent_map(void)
 	}
 
 out:
-	kfree(em_tree);
+	iput(inode);
+	btrfs_free_dummy_root(root);
 	btrfs_free_dummy_fs_info(fs_info);
 
 	return ret;
diff --git a/fs/btrfs/tests/inode-tests.c b/fs/btrfs/tests/inode-tests.c
index 9957de9f78..99da9d34b7 100644
--- a/fs/btrfs/tests/inode-tests.c
+++ b/fs/btrfs/tests/inode-tests.c
@@ -258,7 +258,7 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 
 	/* First with no extents */
 	BTRFS_I(inode)->root = root;
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, 0, sectorsize);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, sectorsize);
 	if (IS_ERR(em)) {
 		em = NULL;
 		test_err("got an error when we shouldn't have");
@@ -278,7 +278,7 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 	 */
 	setup_file_extents(root, sectorsize);
 
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, 0, (u64)-1);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, (u64)-1);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
@@ -316,7 +316,7 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 	offset = em->start + em->len;
 	free_extent_map(em);
 
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, offset, sectorsize);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, offset, sectorsize);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
@@ -339,7 +339,7 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 	free_extent_map(em);
 
 	/* Regular extent */
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, offset, sectorsize);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, offset, sectorsize);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
@@ -367,7 +367,7 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 	free_extent_map(em);
 
 	/* The next 3 are split extents */
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, offset, sectorsize);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, offset, sectorsize);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
@@ -396,7 +396,7 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 	offset = em->start + em->len;
 	free_extent_map(em);
 
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, offset, sectorsize);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, offset, sectorsize);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
@@ -418,7 +418,7 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 	offset = em->start + em->len;
 	free_extent_map(em);
 
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, offset, sectorsize);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, offset, sectorsize);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
@@ -452,7 +452,7 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 	free_extent_map(em);
 
 	/* Prealloc extent */
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, offset, sectorsize);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, offset, sectorsize);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
@@ -481,7 +481,7 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 	free_extent_map(em);
 
 	/* The next 3 are a half written prealloc extent */
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, offset, sectorsize);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, offset, sectorsize);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
@@ -511,7 +511,7 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 	offset = em->start + em->len;
 	free_extent_map(em);
 
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, offset, sectorsize);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, offset, sectorsize);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
@@ -544,7 +544,7 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 	offset = em->start + em->len;
 	free_extent_map(em);
 
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, offset, sectorsize);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, offset, sectorsize);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
@@ -579,7 +579,7 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 	free_extent_map(em);
 
 	/* Now for the compressed extent */
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, offset, sectorsize);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, offset, sectorsize);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
@@ -613,7 +613,7 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 	free_extent_map(em);
 
 	/* Split compressed extent */
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, offset, sectorsize);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, offset, sectorsize);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
@@ -648,7 +648,7 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 	offset = em->start + em->len;
 	free_extent_map(em);
 
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, offset, sectorsize);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, offset, sectorsize);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
@@ -675,7 +675,7 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 	offset = em->start + em->len;
 	free_extent_map(em);
 
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, offset, sectorsize);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, offset, sectorsize);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
@@ -710,7 +710,7 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 	free_extent_map(em);
 
 	/* A hole between regular extents but no hole extent */
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, offset + 6, sectorsize);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, offset + 6, sectorsize);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
@@ -737,7 +737,7 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 	offset = em->start + em->len;
 	free_extent_map(em);
 
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, offset, SZ_4M);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, offset, SZ_4M);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
@@ -770,7 +770,7 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize)
 	offset = em->start + em->len;
 	free_extent_map(em);
 
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, offset, sectorsize);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, offset, sectorsize);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
@@ -850,7 +850,7 @@ static int test_hole_first(u32 sectorsize, u32 nodesize)
 	insert_inode_item_key(root);
 	insert_extent(root, sectorsize, sectorsize, sectorsize, 0, sectorsize,
 		      sectorsize, BTRFS_FILE_EXTENT_REG, 0, 1);
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, 0, 2 * sectorsize);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, 2 * sectorsize);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
@@ -872,7 +872,7 @@ static int test_hole_first(u32 sectorsize, u32 nodesize)
 	}
 	free_extent_map(em);
 
-	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, sectorsize, 2 * sectorsize);
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, sectorsize, 2 * sectorsize);
 	if (IS_ERR(em)) {
 		test_err("got an error when we shouldn't have");
 		goto out;
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index b93cdf5f17..76117bb2c7 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -23,12 +23,10 @@
 #include "qgroup.h"
 #include "block-group.h"
 #include "space-info.h"
-#include "zoned.h"
 #include "fs.h"
 #include "accessors.h"
 #include "extent-tree.h"
 #include "root-tree.h"
-#include "defrag.h"
 #include "dir-item.h"
 #include "uuid-tree.h"
 #include "ioctl.h"
@@ -407,7 +405,7 @@ static int record_root_in_trans(struct btrfs_trans_handle *trans,
 	int ret = 0;
 
 	if ((test_bit(BTRFS_ROOT_SHAREABLE, &root->state) &&
-	    root->last_trans < trans->transid) || force) {
+	    btrfs_get_root_last_trans(root) < trans->transid) || force) {
 		WARN_ON(!force && root->commit_root != root->node);
 
 		/*
@@ -423,15 +421,15 @@ static int record_root_in_trans(struct btrfs_trans_handle *trans,
 		smp_wmb();
 
 		spin_lock(&fs_info->fs_roots_radix_lock);
-		if (root->last_trans == trans->transid && !force) {
+		if (btrfs_get_root_last_trans(root) == trans->transid && !force) {
 			spin_unlock(&fs_info->fs_roots_radix_lock);
 			return 0;
 		}
 		radix_tree_tag_set(&fs_info->fs_roots_radix,
-				   (unsigned long)root->root_key.objectid,
+				   (unsigned long)btrfs_root_id(root),
 				   BTRFS_ROOT_TRANS_TAG);
 		spin_unlock(&fs_info->fs_roots_radix_lock);
-		root->last_trans = trans->transid;
+		btrfs_set_root_last_trans(root, trans->transid);
 
 		/* this is pretty tricky.  We don't want to
 		 * take the relocation lock in btrfs_record_root_in_trans
@@ -474,7 +472,7 @@ void btrfs_add_dropped_root(struct btrfs_trans_handle *trans,
 	/* Make sure we don't try to update the root at commit time */
 	spin_lock(&fs_info->fs_roots_radix_lock);
 	radix_tree_tag_clear(&fs_info->fs_roots_radix,
-			     (unsigned long)root->root_key.objectid,
+			     (unsigned long)btrfs_root_id(root),
 			     BTRFS_ROOT_TRANS_TAG);
 	spin_unlock(&fs_info->fs_roots_radix_lock);
 }
@@ -493,7 +491,7 @@ int btrfs_record_root_in_trans(struct btrfs_trans_handle *trans,
 	 * and barriers
 	 */
 	smp_rmb();
-	if (root->last_trans == trans->transid &&
+	if (btrfs_get_root_last_trans(root) == trans->transid &&
 	    !test_bit(BTRFS_ROOT_IN_TRANS_SETUP, &root->state))
 		return 0;
 
@@ -552,7 +550,7 @@ static inline bool need_reserve_reloc_root(struct btrfs_root *root)
 
 	if (!fs_info->reloc_ctl ||
 	    !test_bit(BTRFS_ROOT_SHAREABLE, &root->state) ||
-	    root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID ||
+	    btrfs_root_id(root) == BTRFS_TREE_RELOC_OBJECTID ||
 	    root->reloc_root)
 		return false;
 
@@ -1054,7 +1052,7 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
 {
 	struct btrfs_fs_info *info = trans->fs_info;
 	struct btrfs_transaction *cur_trans = trans->transaction;
-	int err = 0;
+	int ret = 0;
 
 	if (refcount_read(&trans->use_count) > 1) {
 		refcount_dec(&trans->use_count);
@@ -1093,13 +1091,13 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
 	if (TRANS_ABORTED(trans) || BTRFS_FS_ERROR(info)) {
 		wake_up_process(info->transaction_kthread);
 		if (TRANS_ABORTED(trans))
-			err = trans->aborted;
+			ret = trans->aborted;
 		else
-			err = -EROFS;
+			ret = -EROFS;
 	}
 
 	kmem_cache_free(btrfs_trans_handle_cachep, trans);
-	return err;
+	return ret;
 }
 
 int btrfs_end_transaction(struct btrfs_trans_handle *trans)
@@ -1120,8 +1118,7 @@ int btrfs_end_transaction_throttle(struct btrfs_trans_handle *trans)
 int btrfs_write_marked_extents(struct btrfs_fs_info *fs_info,
 			       struct extent_io_tree *dirty_pages, int mark)
 {
-	int err = 0;
-	int werr = 0;
+	int ret = 0;
 	struct address_space *mapping = fs_info->btree_inode->i_mapping;
 	struct extent_state *cached_state = NULL;
 	u64 start = 0;
@@ -1131,7 +1128,7 @@ int btrfs_write_marked_extents(struct btrfs_fs_info *fs_info,
 				     mark, &cached_state)) {
 		bool wait_writeback = false;
 
-		err = convert_extent_bit(dirty_pages, start, end,
+		ret = convert_extent_bit(dirty_pages, start, end,
 					 EXTENT_NEED_WAIT,
 					 mark, &cached_state);
 		/*
@@ -1147,22 +1144,22 @@ int btrfs_write_marked_extents(struct btrfs_fs_info *fs_info,
 		 * We cleanup any entries left in the io tree when committing
 		 * the transaction (through extent_io_tree_release()).
 		 */
-		if (err == -ENOMEM) {
-			err = 0;
+		if (ret == -ENOMEM) {
+			ret = 0;
 			wait_writeback = true;
 		}
-		if (!err)
-			err = filemap_fdatawrite_range(mapping, start, end);
-		if (err)
-			werr = err;
-		else if (wait_writeback)
-			werr = filemap_fdatawait_range(mapping, start, end);
+		if (!ret)
+			ret = filemap_fdatawrite_range(mapping, start, end);
+		if (!ret && wait_writeback)
+			ret = filemap_fdatawait_range(mapping, start, end);
 		free_extent_state(cached_state);
+		if (ret)
+			break;
 		cached_state = NULL;
 		cond_resched();
 		start = end + 1;
 	}
-	return werr;
+	return ret;
 }
 
 /*
@@ -1174,12 +1171,11 @@ int btrfs_write_marked_extents(struct btrfs_fs_info *fs_info,
 static int __btrfs_wait_marked_extents(struct btrfs_fs_info *fs_info,
 				       struct extent_io_tree *dirty_pages)
 {
-	int err = 0;
-	int werr = 0;
 	struct address_space *mapping = fs_info->btree_inode->i_mapping;
 	struct extent_state *cached_state = NULL;
 	u64 start = 0;
 	u64 end;
+	int ret = 0;
 
 	while (find_first_extent_bit(dirty_pages, start, &start, &end,
 				     EXTENT_NEED_WAIT, &cached_state)) {
@@ -1191,22 +1187,20 @@ static int __btrfs_wait_marked_extents(struct btrfs_fs_info *fs_info,
 		 * concurrently - we do it only at transaction commit time when
 		 * it's safe to do it (through extent_io_tree_release()).
 		 */
-		err = clear_extent_bit(dirty_pages, start, end,
+		ret = clear_extent_bit(dirty_pages, start, end,
 				       EXTENT_NEED_WAIT, &cached_state);
-		if (err == -ENOMEM)
-			err = 0;
-		if (!err)
-			err = filemap_fdatawait_range(mapping, start, end);
-		if (err)
-			werr = err;
+		if (ret == -ENOMEM)
+			ret = 0;
+		if (!ret)
+			ret = filemap_fdatawait_range(mapping, start, end);
 		free_extent_state(cached_state);
+		if (ret)
+			break;
 		cached_state = NULL;
 		cond_resched();
 		start = end + 1;
 	}
-	if (err)
-		werr = err;
-	return werr;
+	return ret;
 }
 
 static int btrfs_wait_extents(struct btrfs_fs_info *fs_info,
@@ -1231,7 +1225,7 @@ int btrfs_wait_tree_log_extents(struct btrfs_root *log_root, int mark)
 	bool errors = false;
 	int err;
 
-	ASSERT(log_root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID);
+	ASSERT(btrfs_root_id(log_root) == BTRFS_TREE_LOG_OBJECTID);
 
 	err = __btrfs_wait_marked_extents(fs_info, dirty_pages);
 	if ((mark & EXTENT_DIRTY) &&
@@ -1494,7 +1488,7 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans)
 			ASSERT(atomic_read(&root->log_commit[1]) == 0);
 
 			radix_tree_tag_clear(&fs_info->fs_roots_radix,
-					(unsigned long)root->root_key.objectid,
+					(unsigned long)btrfs_root_id(root),
 					BTRFS_ROOT_TRANS_TAG);
 			btrfs_qgroup_free_meta_all_pertrans(root);
 			spin_unlock(&fs_info->fs_roots_radix_lock);
@@ -1585,8 +1579,8 @@ static int qgroup_account_snapshot(struct btrfs_trans_handle *trans,
 		goto out;
 
 	/* Now qgroup are all updated, we can inherit it to new qgroups */
-	ret = btrfs_qgroup_inherit(trans, src->root_key.objectid, dst_objectid,
-				   parent->root_key.objectid, inherit);
+	ret = btrfs_qgroup_inherit(trans, btrfs_root_id(src), dst_objectid,
+				   btrfs_root_id(parent), inherit);
 	if (ret < 0)
 		goto out;
 
@@ -1824,7 +1818,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
 	 * insert root back/forward references
 	 */
 	ret = btrfs_add_root_ref(trans, objectid,
-				 parent_root->root_key.objectid,
+				 btrfs_root_id(parent_root),
 				 btrfs_ino(BTRFS_I(parent_inode)), index,
 				 &fname.disk_name);
 	if (ret) {
@@ -1857,16 +1851,14 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
 		ret = qgroup_account_snapshot(trans, root, parent_root,
 					      pending->inherit, objectid);
 	else if (btrfs_qgroup_mode(fs_info) == BTRFS_QGROUP_MODE_SIMPLE)
-		ret = btrfs_qgroup_inherit(trans, root->root_key.objectid, objectid,
-					   parent_root->root_key.objectid, pending->inherit);
+		ret = btrfs_qgroup_inherit(trans, btrfs_root_id(root), objectid,
+					   btrfs_root_id(parent_root), pending->inherit);
 	if (ret < 0)
 		goto fail;
 
 	ret = btrfs_insert_dir_item(trans, &fname.disk_name,
 				    BTRFS_I(parent_inode), &key, BTRFS_FT_DIR,
 				    index);
-	/* We have check then name at the beginning, so it is impossible. */
-	BUG_ON(ret == -EEXIST || ret == -EOVERFLOW);
 	if (ret) {
 		btrfs_abort_transaction(trans, ret);
 		goto fail;
@@ -1958,19 +1950,6 @@ static void update_super_roots(struct btrfs_fs_info *fs_info)
 		super->uuid_tree_generation = root_item->generation;
 }
 
-int btrfs_transaction_in_commit(struct btrfs_fs_info *info)
-{
-	struct btrfs_transaction *trans;
-	int ret = 0;
-
-	spin_lock(&info->trans_lock);
-	trans = info->running_transaction;
-	if (trans)
-		ret = (trans->state >= TRANS_STATE_COMMIT_START);
-	spin_unlock(&info->trans_lock);
-	return ret;
-}
-
 int btrfs_transaction_blocked(struct btrfs_fs_info *info)
 {
 	struct btrfs_transaction *trans;
@@ -2640,7 +2619,7 @@ int btrfs_clean_one_deleted_snapshot(struct btrfs_fs_info *fs_info)
 	list_del_init(&root->root_list);
 	spin_unlock(&fs_info->trans_lock);
 
-	btrfs_debug(fs_info, "cleaner removing %llu", root->root_key.objectid);
+	btrfs_debug(fs_info, "cleaner removing %llu", btrfs_root_id(root));
 
 	btrfs_kill_all_delayed_nodes(root);
 
@@ -2685,9 +2664,7 @@ void __cold __btrfs_abort_transaction(struct btrfs_trans_handle *trans,
 
 int __init btrfs_transaction_init(void)
 {
-	btrfs_trans_handle_cachep = kmem_cache_create("btrfs_trans_handle",
-			sizeof(struct btrfs_trans_handle), 0,
-			SLAB_TEMPORARY | SLAB_MEM_SPREAD, NULL);
+	btrfs_trans_handle_cachep = KMEM_CACHE(btrfs_trans_handle, SLAB_TEMPORARY);
 	if (!btrfs_trans_handle_cachep)
 		return -ENOMEM;
 	return 0;
diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h
index 2bf8bbdfd0..4e451ab173 100644
--- a/fs/btrfs/transaction.h
+++ b/fs/btrfs/transaction.h
@@ -6,12 +6,27 @@
 #ifndef BTRFS_TRANSACTION_H
 #define BTRFS_TRANSACTION_H
 
+#include <linux/atomic.h>
 #include <linux/refcount.h>
+#include <linux/list.h>
+#include <linux/time64.h>
+#include <linux/mutex.h>
+#include <linux/wait.h>
 #include "btrfs_inode.h"
 #include "delayed-ref.h"
-#include "ctree.h"
+#include "extent-io-tree.h"
+#include "block-rsv.h"
+#include "messages.h"
 #include "misc.h"
 
+struct dentry;
+struct inode;
+struct btrfs_pending_snapshot;
+struct btrfs_fs_info;
+struct btrfs_root_item;
+struct btrfs_root;
+struct btrfs_path;
+
 /* Radix-tree tag for roots that are part of the trasaction. */
 #define BTRFS_ROOT_TRANS_TAG			0
 
@@ -262,7 +277,6 @@ int btrfs_write_marked_extents(struct btrfs_fs_info *fs_info,
 				struct extent_io_tree *dirty_pages, int mark);
 int btrfs_wait_tree_log_extents(struct btrfs_root *root, int mark);
 int btrfs_transaction_blocked(struct btrfs_fs_info *info);
-int btrfs_transaction_in_commit(struct btrfs_fs_info *info);
 void btrfs_put_transaction(struct btrfs_transaction *transaction);
 void btrfs_add_dropped_root(struct btrfs_trans_handle *trans,
 			    struct btrfs_root *root);
diff --git a/fs/btrfs/tree-checker.c b/fs/btrfs/tree-checker.c
index 75c78a5556..a2c3651a3d 100644
--- a/fs/btrfs/tree-checker.c
+++ b/fs/btrfs/tree-checker.c
@@ -21,7 +21,6 @@
 #include "messages.h"
 #include "ctree.h"
 #include "tree-checker.h"
-#include "disk-io.h"
 #include "compression.h"
 #include "volumes.h"
 #include "misc.h"
@@ -30,7 +29,6 @@
 #include "file-item.h"
 #include "inode-item.h"
 #include "dir-item.h"
-#include "raid-stripe-tree.h"
 #include "extent-tree.h"
 
 /*
@@ -67,6 +65,7 @@ static void generic_err(const struct extent_buffer *eb, int slot,
 	vaf.fmt = fmt;
 	vaf.va = &args;
 
+	dump_page(folio_page(eb->folios[0], 0), "eb page dump");
 	btrfs_crit(fs_info,
 		"corrupt %s: root=%llu block=%llu slot=%d, %pV",
 		btrfs_header_level(eb) == 0 ? "leaf" : "node",
@@ -94,6 +93,7 @@ static void file_extent_err(const struct extent_buffer *eb, int slot,
 	vaf.fmt = fmt;
 	vaf.va = &args;
 
+	dump_page(folio_page(eb->folios[0], 0), "eb page dump");
 	btrfs_crit(fs_info,
 	"corrupt %s: root=%llu block=%llu slot=%d ino=%llu file_offset=%llu, %pV",
 		btrfs_header_level(eb) == 0 ? "leaf" : "node",
@@ -154,6 +154,7 @@ static void dir_item_err(const struct extent_buffer *eb, int slot,
 	vaf.fmt = fmt;
 	vaf.va = &args;
 
+	dump_page(folio_page(eb->folios[0], 0), "eb page dump");
 	btrfs_crit(fs_info,
 		"corrupt %s: root=%llu block=%llu slot=%d ino=%llu, %pV",
 		btrfs_header_level(eb) == 0 ? "leaf" : "node",
@@ -649,6 +650,7 @@ static void block_group_err(const struct extent_buffer *eb, int slot,
 	vaf.fmt = fmt;
 	vaf.va = &args;
 
+	dump_page(folio_page(eb->folios[0], 0), "eb page dump");
 	btrfs_crit(fs_info,
 	"corrupt %s: root=%llu block=%llu slot=%d bg_start=%llu bg_len=%llu, %pV",
 		btrfs_header_level(eb) == 0 ? "leaf" : "node",
@@ -1005,6 +1007,7 @@ static void dev_item_err(const struct extent_buffer *eb, int slot,
 	vaf.fmt = fmt;
 	vaf.va = &args;
 
+	dump_page(folio_page(eb->folios[0], 0), "eb page dump");
 	btrfs_crit(eb->fs_info,
 	"corrupt %s: root=%llu block=%llu slot=%d devid=%llu %pV",
 		btrfs_header_level(eb) == 0 ? "leaf" : "node",
@@ -1260,6 +1263,7 @@ static void extent_err(const struct extent_buffer *eb, int slot,
 	vaf.fmt = fmt;
 	vaf.va = &args;
 
+	dump_page(folio_page(eb->folios[0], 0), "eb page dump");
 	btrfs_crit(eb->fs_info,
 	"corrupt %s: block=%llu slot=%d extent bytenr=%llu len=%llu %pV",
 		btrfs_header_level(eb) == 0 ? "leaf" : "node",
@@ -2017,7 +2021,7 @@ int btrfs_check_eb_owner(const struct extent_buffer *eb, u64 root_owner)
 	 * Skip dummy fs, as selftests don't create unique ebs for each dummy
 	 * root.
 	 */
-	if (test_bit(BTRFS_FS_STATE_DUMMY_FS_INFO, &eb->fs_info->fs_state))
+	if (btrfs_is_testing(eb->fs_info))
 		return 0;
 	/*
 	 * There are several call sites (backref walking, qgroup, and data
diff --git a/fs/btrfs/tree-checker.h b/fs/btrfs/tree-checker.h
index 580ec4fde0..01669cfa65 100644
--- a/fs/btrfs/tree-checker.h
+++ b/fs/btrfs/tree-checker.h
@@ -6,10 +6,12 @@
 #ifndef BTRFS_TREE_CHECKER_H
 #define BTRFS_TREE_CHECKER_H
 
+#include <linux/types.h>
 #include <uapi/linux/btrfs_tree.h>
 
 struct extent_buffer;
 struct btrfs_chunk;
+struct btrfs_key;
 
 /* All the extra info needed to verify the parentness of a tree block. */
 struct btrfs_tree_parent_check {
diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
index 331fc74299..0bce1d45e2 100644
--- a/fs/btrfs/tree-log.c
+++ b/fs/btrfs/tree-log.c
@@ -13,13 +13,11 @@
 #include "tree-log.h"
 #include "disk-io.h"
 #include "locking.h"
-#include "print-tree.h"
 #include "backref.h"
 #include "compression.h"
 #include "qgroup.h"
 #include "block-group.h"
 #include "space-info.h"
-#include "zoned.h"
 #include "inode-item.h"
 #include "fs.h"
 #include "accessors.h"
@@ -140,6 +138,25 @@ static void wait_log_commit(struct btrfs_root *root, int transid);
  * and once to do all the other items.
  */
 
+static struct inode *btrfs_iget_logging(u64 objectid, struct btrfs_root *root)
+{
+	unsigned int nofs_flag;
+	struct inode *inode;
+
+	/*
+	 * We're holding a transaction handle whether we are logging or
+	 * replaying a log tree, so we must make sure NOFS semantics apply
+	 * because btrfs_alloc_inode() may be triggered and it uses GFP_KERNEL
+	 * to allocate an inode, which can recurse back into the filesystem and
+	 * attempt a transaction commit, resulting in a deadlock.
+	 */
+	nofs_flag = memalloc_nofs_save();
+	inode = btrfs_iget(root->fs_info->sb, objectid, root);
+	memalloc_nofs_restore(nofs_flag);
+
+	return inode;
+}
+
 /*
  * start a sub transaction and setup the log tree
  * this increments the log tree writer count to make the people
@@ -393,7 +410,7 @@ static int overwrite_item(struct btrfs_trans_handle *trans,
 	 * the leaf before writing into the log tree. See the comments at
 	 * copy_items() for more details.
 	 */
-	ASSERT(root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID);
+	ASSERT(btrfs_root_id(root) != BTRFS_TREE_LOG_OBJECTID);
 
 	item_size = btrfs_item_size(eb, slot);
 	src_ptr = btrfs_item_ptr_offset(eb, slot);
@@ -602,7 +619,7 @@ static noinline struct inode *read_one_inode(struct btrfs_root *root,
 {
 	struct inode *inode;
 
-	inode = btrfs_iget(root->fs_info->sb, objectid, root);
+	inode = btrfs_iget_logging(objectid, root);
 	if (IS_ERR(inode))
 		inode = NULL;
 	return inode;
@@ -750,7 +767,6 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
 			goto out;
 
 		if (ins.objectid > 0) {
-			struct btrfs_ref ref = { 0 };
 			u64 csum_start;
 			u64 csum_end;
 			LIST_HEAD(ordered_sums);
@@ -764,13 +780,15 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
 			if (ret < 0) {
 				goto out;
 			} else if (ret == 0) {
-				btrfs_init_generic_ref(&ref,
-						BTRFS_ADD_DELAYED_REF,
-						ins.objectid, ins.offset, 0,
-						root->root_key.objectid);
-				btrfs_init_data_ref(&ref,
-						root->root_key.objectid,
-						key->objectid, offset, 0, false);
+				struct btrfs_ref ref = {
+					.action = BTRFS_ADD_DELAYED_REF,
+					.bytenr = ins.objectid,
+					.num_bytes = ins.offset,
+					.owning_root = btrfs_root_id(root),
+					.ref_root = btrfs_root_id(root),
+				};
+				btrfs_init_data_ref(&ref, key->objectid, offset,
+						    0, false);
 				ret = btrfs_inc_extent_ref(trans, &ref);
 				if (ret)
 					goto out;
@@ -780,7 +798,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
 				 * allocation tree
 				 */
 				ret = btrfs_alloc_logged_file_extent(trans,
-						root->root_key.objectid,
+						btrfs_root_id(root),
 						key->objectid, offset, &ins);
 				if (ret)
 					goto out;
@@ -799,9 +817,10 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
 
 			ret = btrfs_lookup_csums_list(root->log_root,
 						csum_start, csum_end - 1,
-						&ordered_sums, 0, false);
-			if (ret)
+						&ordered_sums, false);
+			if (ret < 0)
 				goto out;
+			ret = 0;
 			/*
 			 * Now delete all existing cums in the csum root that
 			 * cover our range. We do this because we can have an
@@ -2820,6 +2839,52 @@ static void wait_for_writer(struct btrfs_root *root)
 	finish_wait(&root->log_writer_wait, &wait);
 }
 
+void btrfs_init_log_ctx(struct btrfs_log_ctx *ctx, struct inode *inode)
+{
+	ctx->log_ret = 0;
+	ctx->log_transid = 0;
+	ctx->log_new_dentries = false;
+	ctx->logging_new_name = false;
+	ctx->logging_new_delayed_dentries = false;
+	ctx->logged_before = false;
+	ctx->inode = inode;
+	INIT_LIST_HEAD(&ctx->list);
+	INIT_LIST_HEAD(&ctx->ordered_extents);
+	INIT_LIST_HEAD(&ctx->conflict_inodes);
+	ctx->num_conflict_inodes = 0;
+	ctx->logging_conflict_inodes = false;
+	ctx->scratch_eb = NULL;
+}
+
+void btrfs_init_log_ctx_scratch_eb(struct btrfs_log_ctx *ctx)
+{
+	struct btrfs_inode *inode = BTRFS_I(ctx->inode);
+
+	if (!test_bit(BTRFS_INODE_NEEDS_FULL_SYNC, &inode->runtime_flags) &&
+	    !test_bit(BTRFS_INODE_COPY_EVERYTHING, &inode->runtime_flags))
+		return;
+
+	/*
+	 * Don't care about allocation failure. This is just for optimization,
+	 * if we fail to allocate here, we will try again later if needed.
+	 */
+	ctx->scratch_eb = alloc_dummy_extent_buffer(inode->root->fs_info, 0);
+}
+
+void btrfs_release_log_ctx_extents(struct btrfs_log_ctx *ctx)
+{
+	struct btrfs_ordered_extent *ordered;
+	struct btrfs_ordered_extent *tmp;
+
+	ASSERT(inode_is_locked(ctx->inode));
+
+	list_for_each_entry_safe(ordered, tmp, &ctx->ordered_extents, log_list) {
+		list_del_init(&ordered->log_list);
+		btrfs_put_ordered_extent(ordered);
+	}
+}
+
+
 static inline void btrfs_remove_log_ctx(struct btrfs_root *root,
 					struct btrfs_log_ctx *ctx)
 {
@@ -3001,7 +3066,7 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
 		if (ret != -ENOSPC)
 			btrfs_err(fs_info,
 				  "failed to update log for root %llu ret %d",
-				  root->root_key.objectid, ret);
+				  btrfs_root_id(root), ret);
 		btrfs_wait_tree_log_extents(log, mark);
 		mutex_unlock(&log_root_tree->log_mutex);
 		goto out;
@@ -3619,6 +3684,30 @@ out:
 	return ret;
 }
 
+static int clone_leaf(struct btrfs_path *path, struct btrfs_log_ctx *ctx)
+{
+	const int slot = path->slots[0];
+
+	if (ctx->scratch_eb) {
+		copy_extent_buffer_full(ctx->scratch_eb, path->nodes[0]);
+	} else {
+		ctx->scratch_eb = btrfs_clone_extent_buffer(path->nodes[0]);
+		if (!ctx->scratch_eb)
+			return -ENOMEM;
+	}
+
+	btrfs_release_path(path);
+	path->nodes[0] = ctx->scratch_eb;
+	path->slots[0] = slot;
+	/*
+	 * Add extra ref to scratch eb so that it is not freed when callers
+	 * release the path, so we can reuse it later if needed.
+	 */
+	atomic_inc(&ctx->scratch_eb->refs);
+
+	return 0;
+}
+
 static int process_dir_items_leaf(struct btrfs_trans_handle *trans,
 				  struct btrfs_inode *inode,
 				  struct btrfs_path *path,
@@ -3633,23 +3722,20 @@ static int process_dir_items_leaf(struct btrfs_trans_handle *trans,
 	bool last_found = false;
 	int batch_start = 0;
 	int batch_size = 0;
-	int i;
+	int ret;
 
 	/*
 	 * We need to clone the leaf, release the read lock on it, and use the
 	 * clone before modifying the log tree. See the comment at copy_items()
 	 * about why we need to do this.
 	 */
-	src = btrfs_clone_extent_buffer(path->nodes[0]);
-	if (!src)
-		return -ENOMEM;
+	ret = clone_leaf(path, ctx);
+	if (ret < 0)
+		return ret;
 
-	i = path->slots[0];
-	btrfs_release_path(path);
-	path->nodes[0] = src;
-	path->slots[0] = i;
+	src = path->nodes[0];
 
-	for (; i < nritems; i++) {
+	for (int i = path->slots[0]; i < nritems; i++) {
 		struct btrfs_dir_item *di;
 		struct btrfs_key key;
 		int ret;
@@ -4259,17 +4345,16 @@ static noinline int copy_items(struct btrfs_trans_handle *trans,
 			       struct btrfs_path *dst_path,
 			       struct btrfs_path *src_path,
 			       int start_slot, int nr, int inode_only,
-			       u64 logged_isize)
+			       u64 logged_isize, struct btrfs_log_ctx *ctx)
 {
 	struct btrfs_root *log = inode->root->log_root;
 	struct btrfs_file_extent_item *extent;
 	struct extent_buffer *src;
-	int ret = 0;
+	int ret;
 	struct btrfs_key *ins_keys;
 	u32 *ins_sizes;
 	struct btrfs_item_batch batch;
 	char *ins_data;
-	int i;
 	int dst_index;
 	const bool skip_csum = (inode->flags & BTRFS_INODE_NODATASUM);
 	const u64 i_size = i_size_read(&inode->vfs_inode);
@@ -4302,14 +4387,11 @@ static noinline int copy_items(struct btrfs_trans_handle *trans,
 	 * while the other is holding the delayed node's mutex and wants to
 	 * write lock the same subvolume leaf for flushing delayed items.
 	 */
-	src = btrfs_clone_extent_buffer(src_path->nodes[0]);
-	if (!src)
-		return -ENOMEM;
+	ret = clone_leaf(src_path, ctx);
+	if (ret < 0)
+		return ret;
 
-	i = src_path->slots[0];
-	btrfs_release_path(src_path);
-	src_path->nodes[0] = src;
-	src_path->slots[0] = i;
+	src = src_path->nodes[0];
 
 	ins_data = kmalloc(nr * sizeof(struct btrfs_key) +
 			   nr * sizeof(u32), GFP_NOFS);
@@ -4324,7 +4406,7 @@ static noinline int copy_items(struct btrfs_trans_handle *trans,
 	batch.nr = 0;
 
 	dst_index = 0;
-	for (i = 0; i < nr; i++) {
+	for (int i = 0; i < nr; i++) {
 		const int src_slot = start_slot + i;
 		struct btrfs_root *csum_root;
 		struct btrfs_ordered_sum *sums;
@@ -4399,9 +4481,10 @@ static noinline int copy_items(struct btrfs_trans_handle *trans,
 		disk_bytenr += extent_offset;
 		ret = btrfs_lookup_csums_list(csum_root, disk_bytenr,
 					      disk_bytenr + extent_num_bytes - 1,
-					      &ordered_sums, 0, false);
-		if (ret)
+					      &ordered_sums, false);
+		if (ret < 0)
 			goto out;
+		ret = 0;
 
 		list_for_each_entry_safe(sums, sums_next, &ordered_sums, list) {
 			if (!ret)
@@ -4431,7 +4514,7 @@ add_to_batch:
 		goto out;
 
 	dst_index = 0;
-	for (i = 0; i < nr; i++) {
+	for (int i = 0; i < nr; i++) {
 		const int src_slot = start_slot + i;
 		const int dst_slot = dst_path->slots[0] + dst_index;
 		struct btrfs_key key;
@@ -4513,8 +4596,8 @@ static int log_extent_csums(struct btrfs_trans_handle *trans,
 	struct btrfs_root *csum_root;
 	u64 csum_offset;
 	u64 csum_len;
-	u64 mod_start = em->mod_start;
-	u64 mod_len = em->mod_len;
+	u64 mod_start = em->start;
+	u64 mod_len = em->len;
 	LIST_HEAD(ordered_sums);
 	int ret = 0;
 
@@ -4594,9 +4677,10 @@ static int log_extent_csums(struct btrfs_trans_handle *trans,
 	csum_root = btrfs_csum_root(trans->fs_info, em->block_start);
 	ret = btrfs_lookup_csums_list(csum_root, em->block_start + csum_offset,
 				      em->block_start + csum_offset +
-				      csum_len - 1, &ordered_sums, 0, false);
-	if (ret)
+				      csum_len - 1, &ordered_sums, false);
+	if (ret < 0)
 		return ret;
+	ret = 0;
 
 	while (!list_empty(&ordered_sums)) {
 		struct btrfs_ordered_sum *sums = list_entry(ordered_sums.next,
@@ -4704,7 +4788,8 @@ static int log_one_extent(struct btrfs_trans_handle *trans,
  */
 static int btrfs_log_prealloc_extents(struct btrfs_trans_handle *trans,
 				      struct btrfs_inode *inode,
-				      struct btrfs_path *path)
+				      struct btrfs_path *path,
+				      struct btrfs_log_ctx *ctx)
 {
 	struct btrfs_root *root = inode->root;
 	struct btrfs_key key;
@@ -4770,7 +4855,7 @@ static int btrfs_log_prealloc_extents(struct btrfs_trans_handle *trans,
 		if (slot >= btrfs_header_nritems(leaf)) {
 			if (ins_nr > 0) {
 				ret = copy_items(trans, inode, dst_path, path,
-						 start_slot, ins_nr, 1, 0);
+						 start_slot, ins_nr, 1, 0, ctx);
 				if (ret < 0)
 					goto out;
 				ins_nr = 0;
@@ -4794,18 +4879,23 @@ static int btrfs_log_prealloc_extents(struct btrfs_trans_handle *trans,
 			path->slots[0]++;
 			continue;
 		}
-		if (!dropped_extents) {
-			/*
-			 * Avoid logging extent items logged in past fsync calls
-			 * and leading to duplicate keys in the log tree.
-			 */
+		/*
+		 * Avoid overlapping items in the log tree. The first time we
+		 * get here, get rid of everything from a past fsync. After
+		 * that, if the current extent starts before the end of the last
+		 * extent we copied, truncate the last one. This can happen if
+		 * an ordered extent completion modifies the subvolume tree
+		 * while btrfs_next_leaf() has the tree unlocked.
+		 */
+		if (!dropped_extents || key.offset < truncate_offset) {
 			ret = truncate_inode_items(trans, root->log_root, inode,
-						   truncate_offset,
+						   min(key.offset, truncate_offset),
 						   BTRFS_EXTENT_DATA_KEY);
 			if (ret)
 				goto out;
 			dropped_extents = true;
 		}
+		truncate_offset = btrfs_file_extent_end(path);
 		if (ins_nr == 0)
 			start_slot = slot;
 		ins_nr++;
@@ -4820,7 +4910,7 @@ static int btrfs_log_prealloc_extents(struct btrfs_trans_handle *trans,
 	}
 	if (ins_nr > 0)
 		ret = copy_items(trans, inode, dst_path, path,
-				 start_slot, ins_nr, 1, 0);
+				 start_slot, ins_nr, 1, 0, ctx);
 out:
 	btrfs_release_path(path);
 	btrfs_free_path(dst_path);
@@ -4883,7 +4973,7 @@ process:
 		 * private list.
 		 */
 		if (ret) {
-			clear_em_logging(tree, em);
+			clear_em_logging(inode, em);
 			free_extent_map(em);
 			continue;
 		}
@@ -4892,14 +4982,14 @@ process:
 
 		ret = log_one_extent(trans, inode, em, path, ctx);
 		write_lock(&tree->lock);
-		clear_em_logging(tree, em);
+		clear_em_logging(inode, em);
 		free_extent_map(em);
 	}
 	WARN_ON(!list_empty(&extents));
 	write_unlock(&tree->lock);
 
 	if (!ret)
-		ret = btrfs_log_prealloc_extents(trans, inode, path);
+		ret = btrfs_log_prealloc_extents(trans, inode, path, ctx);
 	if (ret)
 		return ret;
 
@@ -4980,7 +5070,8 @@ static int logged_inode_size(struct btrfs_root *log, struct btrfs_inode *inode,
 static int btrfs_log_all_xattrs(struct btrfs_trans_handle *trans,
 				struct btrfs_inode *inode,
 				struct btrfs_path *path,
-				struct btrfs_path *dst_path)
+				struct btrfs_path *dst_path,
+				struct btrfs_log_ctx *ctx)
 {
 	struct btrfs_root *root = inode->root;
 	int ret;
@@ -5009,7 +5100,7 @@ static int btrfs_log_all_xattrs(struct btrfs_trans_handle *trans,
 		if (slot >= nritems) {
 			if (ins_nr > 0) {
 				ret = copy_items(trans, inode, dst_path, path,
-						 start_slot, ins_nr, 1, 0);
+						 start_slot, ins_nr, 1, 0, ctx);
 				if (ret < 0)
 					return ret;
 				ins_nr = 0;
@@ -5035,7 +5126,7 @@ static int btrfs_log_all_xattrs(struct btrfs_trans_handle *trans,
 	}
 	if (ins_nr > 0) {
 		ret = copy_items(trans, inode, dst_path, path,
-				 start_slot, ins_nr, 1, 0);
+				 start_slot, ins_nr, 1, 0, ctx);
 		if (ret < 0)
 			return ret;
 	}
@@ -5366,7 +5457,6 @@ static int log_new_dir_dentries(struct btrfs_trans_handle *trans,
 				struct btrfs_log_ctx *ctx)
 {
 	struct btrfs_root *root = start_inode->root;
-	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct btrfs_path *path;
 	LIST_HEAD(dir_list);
 	struct btrfs_dir_list *dir_elem;
@@ -5427,7 +5517,7 @@ again:
 				continue;
 
 			btrfs_release_path(path);
-			di_inode = btrfs_iget(fs_info->sb, di_key.objectid, root);
+			di_inode = btrfs_iget_logging(di_key.objectid, root);
 			if (IS_ERR(di_inode)) {
 				ret = PTR_ERR(di_inode);
 				goto out;
@@ -5487,7 +5577,7 @@ again:
 		btrfs_add_delayed_iput(curr_inode);
 		curr_inode = NULL;
 
-		vfs_inode = btrfs_iget(fs_info->sb, ino, root);
+		vfs_inode = btrfs_iget_logging(ino, root);
 		if (IS_ERR(vfs_inode)) {
 			ret = PTR_ERR(vfs_inode);
 			break;
@@ -5582,7 +5672,7 @@ static int add_conflicting_inode(struct btrfs_trans_handle *trans,
 	if (ctx->num_conflict_inodes >= MAX_CONFLICT_INODES)
 		return BTRFS_LOG_FORCE_COMMIT;
 
-	inode = btrfs_iget(root->fs_info->sb, ino, root);
+	inode = btrfs_iget_logging(ino, root);
 	/*
 	 * If the other inode that had a conflicting dir entry was deleted in
 	 * the current transaction then we either:
@@ -5683,7 +5773,6 @@ static int log_conflicting_inodes(struct btrfs_trans_handle *trans,
 				  struct btrfs_root *root,
 				  struct btrfs_log_ctx *ctx)
 {
-	struct btrfs_fs_info *fs_info = root->fs_info;
 	int ret = 0;
 
 	/*
@@ -5714,7 +5803,7 @@ static int log_conflicting_inodes(struct btrfs_trans_handle *trans,
 		list_del(&curr->list);
 		kfree(curr);
 
-		inode = btrfs_iget(fs_info->sb, ino, root);
+		inode = btrfs_iget_logging(ino, root);
 		/*
 		 * If the other inode that had a conflicting dir entry was
 		 * deleted in the current transaction, we need to log its parent
@@ -5725,7 +5814,7 @@ static int log_conflicting_inodes(struct btrfs_trans_handle *trans,
 			if (ret != -ENOENT)
 				break;
 
-			inode = btrfs_iget(fs_info->sb, parent, root);
+			inode = btrfs_iget_logging(parent, root);
 			if (IS_ERR(inode)) {
 				ret = PTR_ERR(inode);
 				break;
@@ -5847,7 +5936,7 @@ again:
 				}
 				ret = copy_items(trans, inode, dst_path, path,
 						 ins_start_slot, ins_nr,
-						 inode_only, logged_isize);
+						 inode_only, logged_isize, ctx);
 				if (ret < 0)
 					return ret;
 				ins_nr = 0;
@@ -5866,7 +5955,7 @@ again:
 				goto next_slot;
 			ret = copy_items(trans, inode, dst_path, path,
 					 ins_start_slot,
-					 ins_nr, inode_only, logged_isize);
+					 ins_nr, inode_only, logged_isize, ctx);
 			if (ret < 0)
 				return ret;
 			ins_nr = 0;
@@ -5883,7 +5972,7 @@ again:
 		}
 
 		ret = copy_items(trans, inode, dst_path, path, ins_start_slot,
-				 ins_nr, inode_only, logged_isize);
+				 ins_nr, inode_only, logged_isize, ctx);
 		if (ret < 0)
 			return ret;
 		ins_nr = 1;
@@ -5898,7 +5987,7 @@ next_slot:
 		if (ins_nr) {
 			ret = copy_items(trans, inode, dst_path, path,
 					 ins_start_slot, ins_nr, inode_only,
-					 logged_isize);
+					 logged_isize, ctx);
 			if (ret < 0)
 				return ret;
 			ins_nr = 0;
@@ -5923,7 +6012,7 @@ next_key:
 	}
 	if (ins_nr) {
 		ret = copy_items(trans, inode, dst_path, path, ins_start_slot,
-				 ins_nr, inode_only, logged_isize);
+				 ins_nr, inode_only, logged_isize, ctx);
 		if (ret)
 			return ret;
 	}
@@ -5934,7 +6023,7 @@ next_key:
 		 * lock the same leaf with btrfs_log_prealloc_extents() below.
 		 */
 		btrfs_release_path(path);
-		ret = btrfs_log_prealloc_extents(trans, inode, dst_path);
+		ret = btrfs_log_prealloc_extents(trans, inode, dst_path, ctx);
 	}
 
 	return ret;
@@ -6247,7 +6336,6 @@ static int log_new_delayed_dentries(struct btrfs_trans_handle *trans,
 				    struct btrfs_log_ctx *ctx)
 {
 	const bool orig_log_new_dentries = ctx->log_new_dentries;
-	struct btrfs_fs_info *fs_info = trans->fs_info;
 	struct btrfs_delayed_item *item;
 	int ret = 0;
 
@@ -6273,7 +6361,7 @@ static int log_new_delayed_dentries(struct btrfs_trans_handle *trans,
 		if (key.type == BTRFS_ROOT_ITEM_KEY)
 			continue;
 
-		di_inode = btrfs_iget(fs_info->sb, key.objectid, inode->root);
+		di_inode = btrfs_iget_logging(key.objectid, inode->root);
 		if (IS_ERR(di_inode)) {
 			ret = PTR_ERR(di_inode);
 			break;
@@ -6526,7 +6614,7 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans,
 
 	btrfs_release_path(path);
 	btrfs_release_path(dst_path);
-	ret = btrfs_log_all_xattrs(trans, inode, path, dst_path);
+	ret = btrfs_log_all_xattrs(trans, inode, path, dst_path, ctx);
 	if (ret)
 		goto out_unlock;
 	xattrs_logged = true;
@@ -6553,7 +6641,7 @@ log_extents:
 		 * BTRFS_INODE_COPY_EVERYTHING set.
 		 */
 		if (!xattrs_logged && inode->logged_trans < trans->transid) {
-			ret = btrfs_log_all_xattrs(trans, inode, path, dst_path);
+			ret = btrfs_log_all_xattrs(trans, inode, path, dst_path, ctx);
 			if (ret)
 				goto out_unlock;
 			btrfs_release_path(path);
@@ -6657,7 +6745,6 @@ static int btrfs_log_all_parents(struct btrfs_trans_handle *trans,
 				 struct btrfs_inode *inode,
 				 struct btrfs_log_ctx *ctx)
 {
-	struct btrfs_fs_info *fs_info = trans->fs_info;
 	int ret;
 	struct btrfs_path *path;
 	struct btrfs_key key;
@@ -6722,8 +6809,7 @@ static int btrfs_log_all_parents(struct btrfs_trans_handle *trans,
 				cur_offset = item_size;
 			}
 
-			dir_inode = btrfs_iget(fs_info->sb, inode_key.objectid,
-					       root);
+			dir_inode = btrfs_iget_logging(inode_key.objectid, root);
 			/*
 			 * If the parent inode was deleted, return an error to
 			 * fallback to a transaction commit. This is to prevent
@@ -6785,7 +6871,6 @@ static int log_new_ancestors(struct btrfs_trans_handle *trans,
 	btrfs_item_key_to_cpu(path->nodes[0], &found_key, path->slots[0]);
 
 	while (true) {
-		struct btrfs_fs_info *fs_info = root->fs_info;
 		struct extent_buffer *leaf;
 		int slot;
 		struct btrfs_key search_key;
@@ -6800,7 +6885,7 @@ static int log_new_ancestors(struct btrfs_trans_handle *trans,
 		search_key.objectid = found_key.offset;
 		search_key.type = BTRFS_INODE_ITEM_KEY;
 		search_key.offset = 0;
-		inode = btrfs_iget(fs_info->sb, ino, root);
+		inode = btrfs_iget_logging(ino, root);
 		if (IS_ERR(inode))
 			return PTR_ERR(inode);
 
@@ -7502,6 +7587,7 @@ void btrfs_log_new_name(struct btrfs_trans_handle *trans,
 
 	btrfs_init_log_ctx(&ctx, &inode->vfs_inode);
 	ctx.logging_new_name = true;
+	btrfs_init_log_ctx_scratch_eb(&ctx);
 	/*
 	 * We don't care about the return value. If we fail to log the new name
 	 * then we know the next attempt to sync the log will fallback to a full
@@ -7510,6 +7596,7 @@ void btrfs_log_new_name(struct btrfs_trans_handle *trans,
 	 * inconsistent state after a rename operation.
 	 */
 	btrfs_log_inode_parent(trans, inode, parent, LOG_INODE_EXISTS, &ctx);
+	free_extent_buffer(ctx.scratch_eb);
 	ASSERT(list_empty(&ctx.conflict_inodes));
 out:
 	/*
diff --git a/fs/btrfs/tree-log.h b/fs/btrfs/tree-log.h
index a550a8a375..22e9cbc815 100644
--- a/fs/btrfs/tree-log.h
+++ b/fs/btrfs/tree-log.h
@@ -6,10 +6,18 @@
 #ifndef BTRFS_TREE_LOG_H
 #define BTRFS_TREE_LOG_H
 
+#include <linux/list.h>
+#include <linux/fs.h>
 #include "messages.h"
 #include "ctree.h"
 #include "transaction.h"
 
+struct inode;
+struct dentry;
+struct btrfs_ordered_extent;
+struct btrfs_root;
+struct btrfs_trans_handle;
+
 /* return value for btrfs_log_dentry_safe that means we don't need to log it at all */
 #define BTRFS_NO_LOG_SYNC 256
 
@@ -36,37 +44,20 @@ struct btrfs_log_ctx {
 	struct list_head conflict_inodes;
 	int num_conflict_inodes;
 	bool logging_conflict_inodes;
+	/*
+	 * Used for fsyncs that need to copy items from the subvolume tree to
+	 * the log tree (full sync flag set or copy everything flag set) to
+	 * avoid allocating a temporary extent buffer while holding a lock on
+	 * an extent buffer of the subvolume tree and under the log transaction.
+	 * Also helps to avoid allocating and freeing a temporary extent buffer
+	 * in case we need to process multiple leaves from the subvolume tree.
+	 */
+	struct extent_buffer *scratch_eb;
 };
 
-static inline void btrfs_init_log_ctx(struct btrfs_log_ctx *ctx,
-				      struct inode *inode)
-{
-	ctx->log_ret = 0;
-	ctx->log_transid = 0;
-	ctx->log_new_dentries = false;
-	ctx->logging_new_name = false;
-	ctx->logging_new_delayed_dentries = false;
-	ctx->logged_before = false;
-	ctx->inode = inode;
-	INIT_LIST_HEAD(&ctx->list);
-	INIT_LIST_HEAD(&ctx->ordered_extents);
-	INIT_LIST_HEAD(&ctx->conflict_inodes);
-	ctx->num_conflict_inodes = 0;
-	ctx->logging_conflict_inodes = false;
-}
-
-static inline void btrfs_release_log_ctx_extents(struct btrfs_log_ctx *ctx)
-{
-	struct btrfs_ordered_extent *ordered;
-	struct btrfs_ordered_extent *tmp;
-
-	ASSERT(inode_is_locked(ctx->inode));
-
-	list_for_each_entry_safe(ordered, tmp, &ctx->ordered_extents, log_list) {
-		list_del_init(&ordered->log_list);
-		btrfs_put_ordered_extent(ordered);
-	}
-}
+void btrfs_init_log_ctx(struct btrfs_log_ctx *ctx, struct inode *inode);
+void btrfs_init_log_ctx_scratch_eb(struct btrfs_log_ctx *ctx);
+void btrfs_release_log_ctx_extents(struct btrfs_log_ctx *ctx);
 
 static inline void btrfs_set_log_full_commit(struct btrfs_trans_handle *trans)
 {
diff --git a/fs/btrfs/tree-mod-log.c b/fs/btrfs/tree-mod-log.c
index 3df6153d5d..fa45b5fb96 100644
--- a/fs/btrfs/tree-mod-log.c
+++ b/fs/btrfs/tree-mod-log.c
@@ -44,7 +44,7 @@ struct tree_mod_elem {
 /*
  * Pull a new tree mod seq number for our operation.
  */
-static inline u64 btrfs_inc_tree_mod_seq(struct btrfs_fs_info *fs_info)
+static u64 btrfs_inc_tree_mod_seq(struct btrfs_fs_info *fs_info)
 {
 	return atomic64_inc_return(&fs_info->tree_mod_seq);
 }
@@ -170,8 +170,7 @@ static noinline int tree_mod_log_insert(struct btrfs_fs_info *fs_info,
  * this until all tree mod log insertions are recorded in the rb tree and then
  * write unlock fs_info::tree_mod_log_lock.
  */
-static inline bool tree_mod_dont_log(struct btrfs_fs_info *fs_info,
-				    struct extent_buffer *eb)
+static bool tree_mod_dont_log(struct btrfs_fs_info *fs_info, struct extent_buffer *eb)
 {
 	if (!test_bit(BTRFS_FS_TREE_MOD_LOG_USERS, &fs_info->flags))
 		return true;
@@ -188,7 +187,7 @@ static inline bool tree_mod_dont_log(struct btrfs_fs_info *fs_info,
 }
 
 /* Similar to tree_mod_dont_log, but doesn't acquire any locks. */
-static inline bool tree_mod_need_log(const struct btrfs_fs_info *fs_info,
+static bool tree_mod_need_log(const struct btrfs_fs_info *fs_info,
 				    struct extent_buffer *eb)
 {
 	if (!test_bit(BTRFS_FS_TREE_MOD_LOG_USERS, &fs_info->flags))
@@ -367,9 +366,9 @@ free_tms:
 	return ret;
 }
 
-static inline int tree_mod_log_free_eb(struct btrfs_fs_info *fs_info,
-				       struct tree_mod_elem **tm_list,
-				       int nritems)
+static int tree_mod_log_free_eb(struct btrfs_fs_info *fs_info,
+				struct tree_mod_elem **tm_list,
+				int nritems)
 {
 	int i, j;
 	int ret;
@@ -1005,7 +1004,7 @@ struct extent_buffer *btrfs_get_old_root(struct btrfs_root *root, u64 time_seq)
 		free_extent_buffer(eb_root);
 
 		check.level = level;
-		check.owner_root = root->root_key.objectid;
+		check.owner_root = btrfs_root_id(root);
 
 		old = read_tree_block(fs_info, logical, &check);
 		if (WARN_ON(IS_ERR(old) || !extent_buffer_uptodate(old))) {
diff --git a/fs/btrfs/tree-mod-log.h b/fs/btrfs/tree-mod-log.h
index 94f10afeee..ff00c8e8a3 100644
--- a/fs/btrfs/tree-mod-log.h
+++ b/fs/btrfs/tree-mod-log.h
@@ -3,7 +3,13 @@
 #ifndef BTRFS_TREE_MOD_LOG_H
 #define BTRFS_TREE_MOD_LOG_H
 
-#include "ctree.h"
+#include <linux/list.h>
+
+struct extent_buffer;
+struct btrfs_fs_info;
+struct btrfs_path;
+struct btrfs_root;
+struct btrfs_seq_list;
 
 /* Represents a tree mod log user. */
 struct btrfs_seq_list {
diff --git a/fs/btrfs/ulist.c b/fs/btrfs/ulist.c
index b4ac2b0cd2..183863f4bf 100644
--- a/fs/btrfs/ulist.c
+++ b/fs/btrfs/ulist.c
@@ -7,7 +7,6 @@
 #include <linux/slab.h>
 #include "messages.h"
 #include "ulist.h"
-#include "ctree.h"
 
 /*
  * ulist is a generic data structure to hold a collection of unique u64
diff --git a/fs/btrfs/ulist.h b/fs/btrfs/ulist.h
index b2cef187ea..8e200fe1a2 100644
--- a/fs/btrfs/ulist.h
+++ b/fs/btrfs/ulist.h
@@ -7,6 +7,7 @@
 #ifndef BTRFS_ULIST_H
 #define BTRFS_ULIST_H
 
+#include <linux/types.h>
 #include <linux/list.h>
 #include <linux/rbtree.h>
 
diff --git a/fs/btrfs/uuid-tree.c b/fs/btrfs/uuid-tree.c
index 5be74f9e47..b0aff297d6 100644
--- a/fs/btrfs/uuid-tree.c
+++ b/fs/btrfs/uuid-tree.c
@@ -9,7 +9,6 @@
 #include "ctree.h"
 #include "transaction.h"
 #include "disk-io.h"
-#include "print-tree.h"
 #include "fs.h"
 #include "accessors.h"
 #include "uuid-tree.h"
@@ -114,7 +113,7 @@ int btrfs_uuid_tree_add(struct btrfs_trans_handle *trans, u8 *uuid, u8 type,
 
 	ret = btrfs_insert_empty_item(trans, uuid_root, path, &key,
 				      sizeof(subid_le));
-	if (ret >= 0) {
+	if (ret == 0) {
 		/* Add an item for the type for the first time */
 		eb = path->nodes[0];
 		slot = path->slots[0];
diff --git a/fs/btrfs/uuid-tree.h b/fs/btrfs/uuid-tree.h
index 5350c87fe2..080ede0227 100644
--- a/fs/btrfs/uuid-tree.h
+++ b/fs/btrfs/uuid-tree.h
@@ -3,6 +3,11 @@
 #ifndef BTRFS_UUID_TREE_H
 #define BTRFS_UUID_TREE_H
 
+#include <linux/types.h>
+
+struct btrfs_trans_handle;
+struct btrfs_fs_info;
+
 int btrfs_uuid_tree_add(struct btrfs_trans_handle *trans, u8 *uuid, u8 type,
 			u64 subid);
 int btrfs_uuid_tree_remove(struct btrfs_trans_handle *trans, u8 *uuid, u8 type,
diff --git a/fs/btrfs/verity.c b/fs/btrfs/verity.c
index 66e2270b0d..4042dd6437 100644
--- a/fs/btrfs/verity.c
+++ b/fs/btrfs/verity.c
@@ -14,7 +14,6 @@
 #include "ctree.h"
 #include "btrfs_inode.h"
 #include "transaction.h"
-#include "disk-io.h"
 #include "locking.h"
 #include "fs.h"
 #include "accessors.h"
diff --git a/fs/btrfs/verity.h b/fs/btrfs/verity.h
index 91c10f7d0a..d696659e43 100644
--- a/fs/btrfs/verity.h
+++ b/fs/btrfs/verity.h
@@ -3,8 +3,13 @@
 #ifndef BTRFS_VERITY_H
 #define BTRFS_VERITY_H
 
+struct inode;
+struct btrfs_inode;
+
 #ifdef CONFIG_FS_VERITY
 
+#include <linux/fsverity.h>
+
 extern const struct fsverity_operations btrfs_verityops;
 
 int btrfs_drop_verity_items(struct btrfs_inode *inode);
@@ -12,6 +17,8 @@ int btrfs_get_verity_descriptor(struct inode *inode, void *buf, size_t buf_size)
 
 #else
 
+#include <linux/errno.h>
+
 static inline int btrfs_drop_verity_items(struct btrfs_inode *inode)
 {
 	return 0;
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index 3ebadc5ec8..c39145e8c4 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -14,10 +14,8 @@
 #include <linux/namei.h>
 #include "misc.h"
 #include "ctree.h"
-#include "extent_map.h"
 #include "disk-io.h"
 #include "transaction.h"
-#include "print-tree.h"
 #include "volumes.h"
 #include "raid56.h"
 #include "rcu-string.h"
@@ -468,39 +466,42 @@ static noinline struct btrfs_fs_devices *find_fsid(
 
 static int
 btrfs_get_bdev_and_sb(const char *device_path, blk_mode_t flags, void *holder,
-		      int flush, struct bdev_handle **bdev_handle,
+		      int flush, struct file **bdev_file,
 		      struct btrfs_super_block **disk_super)
 {
 	struct block_device *bdev;
 	int ret;
 
-	*bdev_handle = bdev_open_by_path(device_path, flags, holder, NULL);
+	*bdev_file = bdev_file_open_by_path(device_path, flags, holder, NULL);
 
-	if (IS_ERR(*bdev_handle)) {
-		ret = PTR_ERR(*bdev_handle);
+	if (IS_ERR(*bdev_file)) {
+		ret = PTR_ERR(*bdev_file);
 		goto error;
 	}
-	bdev = (*bdev_handle)->bdev;
+	bdev = file_bdev(*bdev_file);
 
 	if (flush)
 		sync_blockdev(bdev);
-	ret = set_blocksize(bdev, BTRFS_BDEV_BLOCKSIZE);
-	if (ret) {
-		bdev_release(*bdev_handle);
-		goto error;
+	if (holder) {
+		ret = set_blocksize(*bdev_file, BTRFS_BDEV_BLOCKSIZE);
+		if (ret) {
+			fput(*bdev_file);
+			goto error;
+		}
 	}
 	invalidate_bdev(bdev);
 	*disk_super = btrfs_read_dev_super(bdev);
 	if (IS_ERR(*disk_super)) {
 		ret = PTR_ERR(*disk_super);
-		bdev_release(*bdev_handle);
+		fput(*bdev_file);
 		goto error;
 	}
 
 	return 0;
 
 error:
-	*bdev_handle = NULL;
+	*disk_super = NULL;
+	*bdev_file = NULL;
 	return ret;
 }
 
@@ -643,7 +644,7 @@ static int btrfs_open_one_device(struct btrfs_fs_devices *fs_devices,
 			struct btrfs_device *device, blk_mode_t flags,
 			void *holder)
 {
-	struct bdev_handle *bdev_handle;
+	struct file *bdev_file;
 	struct btrfs_super_block *disk_super;
 	u64 devid;
 	int ret;
@@ -654,7 +655,7 @@ static int btrfs_open_one_device(struct btrfs_fs_devices *fs_devices,
 		return -EINVAL;
 
 	ret = btrfs_get_bdev_and_sb(device->name->str, flags, holder, 1,
-				    &bdev_handle, &disk_super);
+				    &bdev_file, &disk_super);
 	if (ret)
 		return ret;
 
@@ -678,20 +679,20 @@ static int btrfs_open_one_device(struct btrfs_fs_devices *fs_devices,
 		clear_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
 		fs_devices->seeding = true;
 	} else {
-		if (bdev_read_only(bdev_handle->bdev))
+		if (bdev_read_only(file_bdev(bdev_file)))
 			clear_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
 		else
 			set_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
 	}
 
-	if (!bdev_nonrot(bdev_handle->bdev))
+	if (!bdev_nonrot(file_bdev(bdev_file)))
 		fs_devices->rotating = true;
 
-	if (bdev_max_discard_sectors(bdev_handle->bdev))
+	if (bdev_max_discard_sectors(file_bdev(bdev_file)))
 		fs_devices->discardable = true;
 
-	device->bdev_handle = bdev_handle;
-	device->bdev = bdev_handle->bdev;
+	device->bdev_file = bdev_file;
+	device->bdev = file_bdev(bdev_file);
 	clear_bit(BTRFS_DEV_STATE_IN_FS_METADATA, &device->dev_state);
 
 	if (device->devt != device->bdev->bd_dev) {
@@ -716,7 +717,7 @@ static int btrfs_open_one_device(struct btrfs_fs_devices *fs_devices,
 
 error_free_page:
 	btrfs_release_disk_super(disk_super);
-	bdev_release(bdev_handle);
+	fput(bdev_file);
 
 	return -EINVAL;
 }
@@ -779,8 +780,9 @@ static noinline struct btrfs_device *device_list_add(const char *path,
 		if (same_fsid_diff_dev) {
 			generate_random_uuid(fs_devices->fsid);
 			fs_devices->temp_fsid = true;
-			pr_info("BTRFS: device %s using temp-fsid %pU\n",
-				path, fs_devices->fsid);
+		pr_info("BTRFS: device %s (%d:%d) using temp-fsid %pU\n",
+				path, MAJOR(path_devt), MINOR(path_devt),
+				fs_devices->fsid);
 		}
 
 		mutex_lock(&fs_devices->device_list_mutex);
@@ -809,8 +811,9 @@ static noinline struct btrfs_device *device_list_add(const char *path,
 
 		if (fs_devices->opened) {
 			btrfs_err(NULL,
-"device %s belongs to fsid %pU, and the fs is already mounted, scanned by %s (%d)",
-				  path, fs_devices->fsid, current->comm,
+"device %s (%d:%d) belongs to fsid %pU, and the fs is already mounted, scanned by %s (%d)",
+				  path, MAJOR(path_devt), MINOR(path_devt),
+				  fs_devices->fsid, current->comm,
 				  task_pid_nr(current));
 			mutex_unlock(&fs_devices->device_list_mutex);
 			return ERR_PTR(-EBUSY);
@@ -836,13 +839,15 @@ static noinline struct btrfs_device *device_list_add(const char *path,
 
 		if (disk_super->label[0])
 			pr_info(
-	"BTRFS: device label %s devid %llu transid %llu %s scanned by %s (%d)\n",
+"BTRFS: device label %s devid %llu transid %llu %s (%d:%d) scanned by %s (%d)\n",
 				disk_super->label, devid, found_transid, path,
+				MAJOR(path_devt), MINOR(path_devt),
 				current->comm, task_pid_nr(current));
 		else
 			pr_info(
-	"BTRFS: device fsid %pU devid %llu transid %llu %s scanned by %s (%d)\n",
+"BTRFS: device fsid %pU devid %llu transid %llu %s (%d:%d) scanned by %s (%d)\n",
 				disk_super->fsid, devid, found_transid, path,
+				MAJOR(path_devt), MINOR(path_devt),
 				current->comm, task_pid_nr(current));
 
 	} else if (!device->name || strcmp(device->name->str, path)) {
@@ -1025,10 +1030,10 @@ static void __btrfs_free_extra_devids(struct btrfs_fs_devices *fs_devices,
 		if (device->devid == BTRFS_DEV_REPLACE_DEVID)
 			continue;
 
-		if (device->bdev_handle) {
-			bdev_release(device->bdev_handle);
+		if (device->bdev_file) {
+			fput(device->bdev_file);
 			device->bdev = NULL;
-			device->bdev_handle = NULL;
+			device->bdev_file = NULL;
 			fs_devices->open_devices--;
 		}
 		if (test_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state)) {
@@ -1073,7 +1078,7 @@ static void btrfs_close_bdev(struct btrfs_device *device)
 		invalidate_bdev(device->bdev);
 	}
 
-	bdev_release(device->bdev_handle);
+	fput(device->bdev_file);
 }
 
 static void btrfs_close_one_device(struct btrfs_device *device)
@@ -1285,7 +1290,7 @@ static struct btrfs_super_block *btrfs_read_disk_super(struct block_device *bdev
 		return ERR_PTR(-EINVAL);
 
 	/* pull in the page with our super */
-	page = read_cache_page_gfp(bdev->bd_inode->i_mapping, index, GFP_KERNEL);
+	page = read_cache_page_gfp(bdev->bd_mapping, index, GFP_KERNEL);
 
 	if (IS_ERR(page))
 		return ERR_CAST(page);
@@ -1374,8 +1379,9 @@ struct btrfs_device *btrfs_scan_one_device(const char *path, blk_mode_t flags,
 	struct btrfs_super_block *disk_super;
 	bool new_device_added = false;
 	struct btrfs_device *device = NULL;
-	struct bdev_handle *bdev_handle;
+	struct file *bdev_file;
 	u64 bytenr, bytenr_orig;
+	dev_t devt;
 	int ret;
 
 	lockdep_assert_held(&uuid_mutex);
@@ -1397,31 +1403,30 @@ struct btrfs_device *btrfs_scan_one_device(const char *path, blk_mode_t flags,
 	 * values temporarily, as the device paths of the fsid are the only
 	 * required information for assembling the volume.
 	 */
-	bdev_handle = bdev_open_by_path(path, flags, NULL, NULL);
-	if (IS_ERR(bdev_handle))
-		return ERR_CAST(bdev_handle);
+	bdev_file = bdev_file_open_by_path(path, flags, NULL, NULL);
+	if (IS_ERR(bdev_file))
+		return ERR_CAST(bdev_file);
 
 	bytenr_orig = btrfs_sb_offset(0);
-	ret = btrfs_sb_log_location_bdev(bdev_handle->bdev, 0, READ, &bytenr);
+	ret = btrfs_sb_log_location_bdev(file_bdev(bdev_file), 0, READ, &bytenr);
 	if (ret) {
 		device = ERR_PTR(ret);
 		goto error_bdev_put;
 	}
 
-	disk_super = btrfs_read_disk_super(bdev_handle->bdev, bytenr,
+	disk_super = btrfs_read_disk_super(file_bdev(bdev_file), bytenr,
 					   bytenr_orig);
 	if (IS_ERR(disk_super)) {
 		device = ERR_CAST(disk_super);
 		goto error_bdev_put;
 	}
 
-	if (btrfs_skip_registration(disk_super, path, bdev_handle->bdev->bd_dev,
-				    mount_arg_dev)) {
+	devt = file_bdev(bdev_file)->bd_dev;
+	if (btrfs_skip_registration(disk_super, path, devt, mount_arg_dev)) {
 		pr_debug("BTRFS: skip registering single non-seed device %s (%d:%d)\n",
-			  path, MAJOR(bdev_handle->bdev->bd_dev),
-			  MINOR(bdev_handle->bdev->bd_dev));
+			  path, MAJOR(devt), MINOR(devt));
 
-		btrfs_free_stale_devices(bdev_handle->bdev->bd_dev, NULL);
+		btrfs_free_stale_devices(devt, NULL);
 
 		device = NULL;
 		goto free_disk_super;
@@ -1435,7 +1440,7 @@ free_disk_super:
 	btrfs_release_disk_super(disk_super);
 
 error_bdev_put:
-	bdev_release(bdev_handle);
+	fput(bdev_file);
 
 	return device;
 }
@@ -2086,11 +2091,10 @@ static void btrfs_scratch_superblock(struct btrfs_fs_info *fs_info,
 			copy_num, ret);
 }
 
-void btrfs_scratch_superblocks(struct btrfs_fs_info *fs_info,
-			       struct block_device *bdev,
-			       const char *device_path)
+void btrfs_scratch_superblocks(struct btrfs_fs_info *fs_info, struct btrfs_device *device)
 {
 	int copy_num;
+	struct block_device *bdev = device->bdev;
 
 	if (!bdev)
 		return;
@@ -2106,12 +2110,12 @@ void btrfs_scratch_superblocks(struct btrfs_fs_info *fs_info,
 	btrfs_kobject_uevent(bdev, KOBJ_CHANGE);
 
 	/* Update ctime/mtime for device path for libblkid */
-	update_dev_time(device_path);
+	update_dev_time(device->name->str);
 }
 
 int btrfs_rm_device(struct btrfs_fs_info *fs_info,
 		    struct btrfs_dev_lookup_args *args,
-		    struct bdev_handle **bdev_handle)
+		    struct file **bdev_file)
 {
 	struct btrfs_trans_handle *trans;
 	struct btrfs_device *device;
@@ -2220,7 +2224,7 @@ int btrfs_rm_device(struct btrfs_fs_info *fs_info,
 
 	btrfs_assign_next_active_device(device, NULL);
 
-	if (device->bdev_handle) {
+	if (device->bdev_file) {
 		cur_devices->open_devices--;
 		/* remove sysfs entry */
 		btrfs_sysfs_remove_device(device);
@@ -2236,20 +2240,19 @@ int btrfs_rm_device(struct btrfs_fs_info *fs_info,
 	 * free the device.
 	 *
 	 * We cannot call btrfs_close_bdev() here because we're holding the sb
-	 * write lock, and bdev_release() will pull in the ->open_mutex on
-	 * the block device and it's dependencies.  Instead just flush the
-	 * device and let the caller do the final bdev_release.
+	 * write lock, and fput() on the block device will pull in the
+	 * ->open_mutex on the block device and it's dependencies.  Instead
+	 *  just flush the device and let the caller do the final bdev_release.
 	 */
 	if (test_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state)) {
-		btrfs_scratch_superblocks(fs_info, device->bdev,
-					  device->name->str);
+		btrfs_scratch_superblocks(fs_info, device);
 		if (device->bdev) {
 			sync_blockdev(device->bdev);
 			invalidate_bdev(device->bdev);
 		}
 	}
 
-	*bdev_handle = device->bdev_handle;
+	*bdev_file = device->bdev_file;
 	synchronize_rcu();
 	btrfs_free_device(device);
 
@@ -2355,8 +2358,7 @@ void btrfs_destroy_dev_replace_tgtdev(struct btrfs_device *tgtdev)
 
 	mutex_unlock(&fs_devices->device_list_mutex);
 
-	btrfs_scratch_superblocks(tgtdev->fs_info, tgtdev->bdev,
-				  tgtdev->name->str);
+	btrfs_scratch_superblocks(tgtdev->fs_info, tgtdev);
 
 	btrfs_close_bdev(tgtdev);
 	synchronize_rcu();
@@ -2386,7 +2388,7 @@ int btrfs_get_dev_args_from_path(struct btrfs_fs_info *fs_info,
 				 const char *path)
 {
 	struct btrfs_super_block *disk_super;
-	struct bdev_handle *bdev_handle;
+	struct file *bdev_file;
 	int ret;
 
 	if (!path || !path[0])
@@ -2404,7 +2406,7 @@ int btrfs_get_dev_args_from_path(struct btrfs_fs_info *fs_info,
 	}
 
 	ret = btrfs_get_bdev_and_sb(path, BLK_OPEN_READ, NULL, 0,
-				    &bdev_handle, &disk_super);
+				    &bdev_file, &disk_super);
 	if (ret) {
 		btrfs_put_dev_args_from_path(args);
 		return ret;
@@ -2417,7 +2419,7 @@ int btrfs_get_dev_args_from_path(struct btrfs_fs_info *fs_info,
 	else
 		memcpy(args->fsid, disk_super->fsid, BTRFS_FSID_SIZE);
 	btrfs_release_disk_super(disk_super);
-	bdev_release(bdev_handle);
+	fput(bdev_file);
 	return 0;
 }
 
@@ -2637,7 +2639,7 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 	struct btrfs_root *root = fs_info->dev_root;
 	struct btrfs_trans_handle *trans;
 	struct btrfs_device *device;
-	struct bdev_handle *bdev_handle;
+	struct file *bdev_file;
 	struct super_block *sb = fs_info->sb;
 	struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
 	struct btrfs_fs_devices *seed_devices = NULL;
@@ -2650,12 +2652,12 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 	if (sb_rdonly(sb) && !fs_devices->seeding)
 		return -EROFS;
 
-	bdev_handle = bdev_open_by_path(device_path, BLK_OPEN_WRITE,
+	bdev_file = bdev_file_open_by_path(device_path, BLK_OPEN_WRITE,
 					fs_info->bdev_holder, NULL);
-	if (IS_ERR(bdev_handle))
-		return PTR_ERR(bdev_handle);
+	if (IS_ERR(bdev_file))
+		return PTR_ERR(bdev_file);
 
-	if (!btrfs_check_device_zone_type(fs_info, bdev_handle->bdev)) {
+	if (!btrfs_check_device_zone_type(fs_info, file_bdev(bdev_file))) {
 		ret = -EINVAL;
 		goto error;
 	}
@@ -2667,11 +2669,11 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 		locked = true;
 	}
 
-	sync_blockdev(bdev_handle->bdev);
+	sync_blockdev(file_bdev(bdev_file));
 
 	rcu_read_lock();
 	list_for_each_entry_rcu(device, &fs_devices->devices, dev_list) {
-		if (device->bdev == bdev_handle->bdev) {
+		if (device->bdev == file_bdev(bdev_file)) {
 			ret = -EEXIST;
 			rcu_read_unlock();
 			goto error;
@@ -2687,8 +2689,8 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 	}
 
 	device->fs_info = fs_info;
-	device->bdev_handle = bdev_handle;
-	device->bdev = bdev_handle->bdev;
+	device->bdev_file = bdev_file;
+	device->bdev = file_bdev(bdev_file);
 	ret = lookup_bdev(device_path, &device->devt);
 	if (ret)
 		goto error_free_device;
@@ -2715,7 +2717,7 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 	set_bit(BTRFS_DEV_STATE_IN_FS_METADATA, &device->dev_state);
 	clear_bit(BTRFS_DEV_STATE_REPLACE_TGT, &device->dev_state);
 	device->dev_stats_valid = 1;
-	set_blocksize(device->bdev, BTRFS_BDEV_BLOCKSIZE);
+	set_blocksize(device->bdev_file, BTRFS_BDEV_BLOCKSIZE);
 
 	if (seeding_dev) {
 		btrfs_clear_sb_rdonly(sb);
@@ -2871,7 +2873,7 @@ error_free_zone:
 error_free_device:
 	btrfs_free_device(device);
 error:
-	bdev_release(bdev_handle);
+	fput(bdev_file);
 	if (locked) {
 		mutex_unlock(&uuid_mutex);
 		up_write(&sb->s_umount);
@@ -3545,6 +3547,44 @@ static int btrfs_may_alloc_data_chunk(struct btrfs_fs_info *fs_info,
 	return 0;
 }
 
+static void btrfs_disk_balance_args_to_cpu(struct btrfs_balance_args *cpu,
+					   const struct btrfs_disk_balance_args *disk)
+{
+	memset(cpu, 0, sizeof(*cpu));
+
+	cpu->profiles = le64_to_cpu(disk->profiles);
+	cpu->usage = le64_to_cpu(disk->usage);
+	cpu->devid = le64_to_cpu(disk->devid);
+	cpu->pstart = le64_to_cpu(disk->pstart);
+	cpu->pend = le64_to_cpu(disk->pend);
+	cpu->vstart = le64_to_cpu(disk->vstart);
+	cpu->vend = le64_to_cpu(disk->vend);
+	cpu->target = le64_to_cpu(disk->target);
+	cpu->flags = le64_to_cpu(disk->flags);
+	cpu->limit = le64_to_cpu(disk->limit);
+	cpu->stripes_min = le32_to_cpu(disk->stripes_min);
+	cpu->stripes_max = le32_to_cpu(disk->stripes_max);
+}
+
+static void btrfs_cpu_balance_args_to_disk(struct btrfs_disk_balance_args *disk,
+					   const struct btrfs_balance_args *cpu)
+{
+	memset(disk, 0, sizeof(*disk));
+
+	disk->profiles = cpu_to_le64(cpu->profiles);
+	disk->usage = cpu_to_le64(cpu->usage);
+	disk->devid = cpu_to_le64(cpu->devid);
+	disk->pstart = cpu_to_le64(cpu->pstart);
+	disk->pend = cpu_to_le64(cpu->pend);
+	disk->vstart = cpu_to_le64(cpu->vstart);
+	disk->vend = cpu_to_le64(cpu->vend);
+	disk->target = cpu_to_le64(cpu->target);
+	disk->flags = cpu_to_le64(cpu->flags);
+	disk->limit = cpu_to_le64(cpu->limit);
+	disk->stripes_min = cpu_to_le32(cpu->stripes_min);
+	disk->stripes_max = cpu_to_le32(cpu->stripes_max);
+}
+
 static int insert_balance_item(struct btrfs_fs_info *fs_info,
 			       struct btrfs_balance_control *bctl)
 {
@@ -3689,7 +3729,7 @@ static void reset_balance_state(struct btrfs_fs_info *fs_info)
 	struct btrfs_balance_control *bctl = fs_info->balance_ctl;
 	int ret;
 
-	BUG_ON(!fs_info->balance_ctl);
+	ASSERT(fs_info->balance_ctl);
 
 	spin_lock(&fs_info->balance_lock);
 	fs_info->balance_ctl = NULL;
@@ -5578,21 +5618,6 @@ struct btrfs_chunk_map *btrfs_alloc_chunk_map(int num_stripes, gfp_t gfp)
 	return map;
 }
 
-struct btrfs_chunk_map *btrfs_clone_chunk_map(struct btrfs_chunk_map *map, gfp_t gfp)
-{
-	const int size = btrfs_chunk_map_size(map->num_stripes);
-	struct btrfs_chunk_map *clone;
-
-	clone = kmemdup(map, size, gfp);
-	if (!clone)
-		return NULL;
-
-	refcount_set(&clone->refs, 1);
-	RB_CLEAR_NODE(&clone->rb_node);
-
-	return clone;
-}
-
 static struct btrfs_block_group *create_chunk(struct btrfs_trans_handle *trans,
 			struct alloc_chunk_ctl *ctl,
 			struct btrfs_device_info *devices_info)
@@ -6009,6 +6034,7 @@ static int find_live_mirror(struct btrfs_fs_info *fs_info,
 			    struct btrfs_chunk_map *map, int first,
 			    int dev_replace_is_ongoing)
 {
+	const enum btrfs_read_policy policy = READ_ONCE(fs_info->fs_devices->read_policy);
 	int i;
 	int num_stripes;
 	int preferred_mirror;
@@ -6023,13 +6049,12 @@ static int find_live_mirror(struct btrfs_fs_info *fs_info,
 	else
 		num_stripes = map->num_stripes;
 
-	switch (fs_info->fs_devices->read_policy) {
+	switch (policy) {
 	default:
 		/* Shouldn't happen, just warn and use pid instead of failing */
-		btrfs_warn_rl(fs_info,
-			      "unknown read_policy type %u, reset to pid",
-			      fs_info->fs_devices->read_policy);
-		fs_info->fs_devices->read_policy = BTRFS_READ_POLICY_PID;
+		btrfs_warn_rl(fs_info, "unknown read_policy type %u, reset to pid",
+			      policy);
+		WRITE_ONCE(fs_info->fs_devices->read_policy, BTRFS_READ_POLICY_PID);
 		fallthrough;
 	case BTRFS_READ_POLICY_PID:
 		preferred_mirror = first + (current->pid % num_stripes);
diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h
index 53f87f398d..66e6fc481e 100644
--- a/fs/btrfs/volumes.h
+++ b/fs/btrfs/volumes.h
@@ -6,13 +6,28 @@
 #ifndef BTRFS_VOLUMES_H
 #define BTRFS_VOLUMES_H
 
+#include <linux/blk_types.h>
+#include <linux/sizes.h>
+#include <linux/atomic.h>
 #include <linux/sort.h>
-#include <linux/btrfs.h>
-#include "async-thread.h"
+#include <linux/list.h>
+#include <linux/mutex.h>
+#include <linux/log2.h>
+#include <linux/kobject.h>
+#include <linux/refcount.h>
+#include <linux/completion.h>
+#include <linux/rbtree.h>
+#include <uapi/linux/btrfs.h>
 #include "messages.h"
-#include "tree-checker.h"
 #include "rcu-string.h"
 
+struct block_device;
+struct bdev_handle;
+struct btrfs_fs_info;
+struct btrfs_block_group;
+struct btrfs_trans_handle;
+struct btrfs_zoned_device_info;
+
 #define BTRFS_MAX_DATA_CHUNK_SIZE	(10ULL * SZ_1G)
 
 extern struct mutex uuid_mutex;
@@ -77,7 +92,10 @@ enum btrfs_raid_types {
 #define BTRFS_DEV_STATE_FLUSH_SENT	(4)
 #define BTRFS_DEV_STATE_NO_READA	(5)
 
-struct btrfs_zoned_device_info;
+/* Special value encoding failure to write primary super block. */
+#define BTRFS_SUPER_PRIMARY_WRITE_ERROR		(INT_MAX / 2)
+
+struct btrfs_fs_devices;
 
 struct btrfs_device {
 	struct list_head dev_list; /* device_list_mutex */
@@ -90,7 +108,7 @@ struct btrfs_device {
 
 	u64 generation;
 
-	struct bdev_handle *bdev_handle;
+	struct file *bdev_file;
 	struct block_device *bdev;
 
 	struct btrfs_zoned_device_info *zone_info;
@@ -127,6 +145,12 @@ struct btrfs_device {
 	/* type and info about this device */
 	u64 type;
 
+	/*
+	 * Counter of super block write errors, values larger than
+	 * BTRFS_SUPER_PRIMARY_WRITE_ERROR encode primary super block write failure.
+	 */
+	atomic_t sb_write_errors;
+
 	/* minimal io size for this device */
 	u32 sector_size;
 
@@ -276,6 +300,25 @@ enum btrfs_read_policy {
 	BTRFS_NR_READ_POLICY,
 };
 
+#ifdef CONFIG_BTRFS_DEBUG
+/*
+ * Checksum mode - offload it to workqueues or do it synchronously in
+ * btrfs_submit_chunk().
+ */
+enum btrfs_offload_csum_mode {
+	/*
+	 * Choose offloading checksum or do it synchronously automatically.
+	 * Do it synchronously if the checksum is fast, or offload to workqueues
+	 * otherwise.
+	 */
+	BTRFS_OFFLOAD_CSUM_AUTO,
+	/* Always offload checksum to workqueues. */
+	BTRFS_OFFLOAD_CSUM_FORCE_ON,
+	/* Never offload checksum to workqueues. */
+	BTRFS_OFFLOAD_CSUM_FORCE_OFF,
+};
+#endif
+
 struct btrfs_fs_devices {
 	u8 fsid[BTRFS_FSID_SIZE]; /* FS specific uuid */
 
@@ -380,6 +423,11 @@ struct btrfs_fs_devices {
 
 	/* Policy used to read the mirrored stripes. */
 	enum btrfs_read_policy read_policy;
+
+#ifdef CONFIG_BTRFS_DEBUG
+	/* Checksum mode - offload it or do it synchronously. */
+	enum btrfs_offload_csum_mode offload_csum_mode;
+#endif
 };
 
 #define BTRFS_MAX_DEVS(info) ((BTRFS_MAX_ITEM_SIZE(info)	\
@@ -557,8 +605,6 @@ static inline void btrfs_free_chunk_map(struct btrfs_chunk_map *map)
 	}
 }
 
-struct btrfs_balance_args;
-struct btrfs_balance_progress;
 struct btrfs_balance_control {
 	struct btrfs_balance_args data;
 	struct btrfs_balance_args meta;
@@ -661,7 +707,7 @@ struct btrfs_device *btrfs_alloc_device(struct btrfs_fs_info *fs_info,
 void btrfs_put_dev_args_from_path(struct btrfs_dev_lookup_args *args);
 int btrfs_rm_device(struct btrfs_fs_info *fs_info,
 		    struct btrfs_dev_lookup_args *args,
-		    struct bdev_handle **bdev_handle);
+		    struct file **bdev_file);
 void __exit btrfs_cleanup_fs_uuids(void);
 int btrfs_num_copies(struct btrfs_fs_info *fs_info, u64 logical, u64 len);
 int btrfs_grow_device(struct btrfs_trans_handle *trans,
@@ -706,7 +752,6 @@ struct btrfs_chunk_map *btrfs_alloc_chunk_map(int num_stripes, gfp_t gfp);
 int btrfs_add_chunk_map(struct btrfs_fs_info *fs_info, struct btrfs_chunk_map *map);
 #endif
 
-struct btrfs_chunk_map *btrfs_clone_chunk_map(struct btrfs_chunk_map *map, gfp_t gfp);
 struct btrfs_chunk_map *btrfs_find_chunk_map(struct btrfs_fs_info *fs_info,
 					     u64 logical, u64 length);
 struct btrfs_chunk_map *btrfs_find_chunk_map_nolock(struct btrfs_fs_info *fs_info,
@@ -780,9 +825,7 @@ void btrfs_commit_device_sizes(struct btrfs_transaction *trans);
 struct list_head * __attribute_const__ btrfs_get_fs_uuids(void);
 bool btrfs_check_rw_degradable(struct btrfs_fs_info *fs_info,
 					struct btrfs_device *failing_dev);
-void btrfs_scratch_superblocks(struct btrfs_fs_info *fs_info,
-			       struct block_device *bdev,
-			       const char *device_path);
+void btrfs_scratch_superblocks(struct btrfs_fs_info *fs_info, struct btrfs_device *device);
 
 enum btrfs_raid_types __attribute_const__ btrfs_bg_flags_to_raid_index(u64 flags);
 int btrfs_bg_type_to_factor(u64 flags);
diff --git a/fs/btrfs/xattr.c b/fs/btrfs/xattr.c
index 6287763fdc..15d0999e34 100644
--- a/fs/btrfs/xattr.c
+++ b/fs/btrfs/xattr.c
@@ -504,7 +504,7 @@ static int btrfs_initxattrs(struct inode *inode,
 	const struct xattr *xattr;
 	unsigned int nofs_flag;
 	char *name;
-	int err = 0;
+	int ret = 0;
 
 	/*
 	 * We're holding a transaction handle, so use a NOFS memory allocation
@@ -515,7 +515,7 @@ static int btrfs_initxattrs(struct inode *inode,
 		name = kmalloc(XATTR_SECURITY_PREFIX_LEN +
 			       strlen(xattr->name) + 1, GFP_KERNEL);
 		if (!name) {
-			err = -ENOMEM;
+			ret = -ENOMEM;
 			break;
 		}
 		strcpy(name, XATTR_SECURITY_PREFIX);
@@ -524,14 +524,14 @@ static int btrfs_initxattrs(struct inode *inode,
 		if (strcmp(name, XATTR_NAME_CAPS) == 0)
 			clear_bit(BTRFS_INODE_NO_CAP_XATTR, &BTRFS_I(inode)->runtime_flags);
 
-		err = btrfs_setxattr(trans, inode, name, xattr->value,
+		ret = btrfs_setxattr(trans, inode, name, xattr->value,
 				     xattr->value_len, 0);
 		kfree(name);
-		if (err < 0)
+		if (ret < 0)
 			break;
 	}
 	memalloc_nofs_restore(nofs_flag);
-	return err;
+	return ret;
 }
 
 int btrfs_xattr_security_init(struct btrfs_trans_handle *trans,
diff --git a/fs/btrfs/xattr.h b/fs/btrfs/xattr.h
index 118118ca3e..b9376ea258 100644
--- a/fs/btrfs/xattr.h
+++ b/fs/btrfs/xattr.h
@@ -6,7 +6,11 @@
 #ifndef BTRFS_XATTR_H
 #define BTRFS_XATTR_H
 
-#include <linux/xattr.h>
+struct dentry;
+struct inode;
+struct qstr;
+struct xattr_handler;
+struct btrfs_trans_handle;
 
 extern const struct xattr_handler * const btrfs_xattr_handlers[];
 
diff --git a/fs/btrfs/zlib.c b/fs/btrfs/zlib.c
index 8da66ea699..d9e5c88a0f 100644
--- a/fs/btrfs/zlib.c
+++ b/fs/btrfs/zlib.c
@@ -91,24 +91,24 @@ fail:
 	return ERR_PTR(-ENOMEM);
 }
 
-int zlib_compress_pages(struct list_head *ws, struct address_space *mapping,
-		u64 start, struct page **pages, unsigned long *out_pages,
-		unsigned long *total_in, unsigned long *total_out)
+int zlib_compress_folios(struct list_head *ws, struct address_space *mapping,
+			 u64 start, struct folio **folios, unsigned long *out_folios,
+			 unsigned long *total_in, unsigned long *total_out)
 {
 	struct workspace *workspace = list_entry(ws, struct workspace, list);
 	int ret;
 	char *data_in = NULL;
-	char *cpage_out;
-	int nr_pages = 0;
-	struct page *in_page = NULL;
-	struct page *out_page = NULL;
+	char *cfolio_out;
+	int nr_folios = 0;
+	struct folio *in_folio = NULL;
+	struct folio *out_folio = NULL;
 	unsigned long bytes_left;
-	unsigned int in_buf_pages;
+	unsigned int in_buf_folios;
 	unsigned long len = *total_out;
-	unsigned long nr_dest_pages = *out_pages;
-	const unsigned long max_out = nr_dest_pages * PAGE_SIZE;
+	unsigned long nr_dest_folios = *out_folios;
+	const unsigned long max_out = nr_dest_folios * PAGE_SIZE;
 
-	*out_pages = 0;
+	*out_folios = 0;
 	*total_out = 0;
 	*total_in = 0;
 
@@ -121,18 +121,18 @@ int zlib_compress_pages(struct list_head *ws, struct address_space *mapping,
 	workspace->strm.total_in = 0;
 	workspace->strm.total_out = 0;
 
-	out_page = btrfs_alloc_compr_page();
-	if (out_page == NULL) {
+	out_folio = btrfs_alloc_compr_folio();
+	if (out_folio == NULL) {
 		ret = -ENOMEM;
 		goto out;
 	}
-	cpage_out = page_address(out_page);
-	pages[0] = out_page;
-	nr_pages = 1;
+	cfolio_out = folio_address(out_folio);
+	folios[0] = out_folio;
+	nr_folios = 1;
 
 	workspace->strm.next_in = workspace->buf;
 	workspace->strm.avail_in = 0;
-	workspace->strm.next_out = cpage_out;
+	workspace->strm.next_out = cfolio_out;
 	workspace->strm.avail_out = PAGE_SIZE;
 
 	while (workspace->strm.total_in < len) {
@@ -142,19 +142,22 @@ int zlib_compress_pages(struct list_head *ws, struct address_space *mapping,
 		 */
 		if (workspace->strm.avail_in == 0) {
 			bytes_left = len - workspace->strm.total_in;
-			in_buf_pages = min(DIV_ROUND_UP(bytes_left, PAGE_SIZE),
-					   workspace->buf_size / PAGE_SIZE);
-			if (in_buf_pages > 1) {
+			in_buf_folios = min(DIV_ROUND_UP(bytes_left, PAGE_SIZE),
+					    workspace->buf_size / PAGE_SIZE);
+			if (in_buf_folios > 1) {
 				int i;
 
-				for (i = 0; i < in_buf_pages; i++) {
+				for (i = 0; i < in_buf_folios; i++) {
 					if (data_in) {
 						kunmap_local(data_in);
-						put_page(in_page);
+						folio_put(in_folio);
+						data_in = NULL;
 					}
-					in_page = find_get_page(mapping,
-								start >> PAGE_SHIFT);
-					data_in = kmap_local_page(in_page);
+					ret = btrfs_compress_filemap_get_folio(mapping,
+							start, &in_folio);
+					if (ret < 0)
+						goto out;
+					data_in = kmap_local_folio(in_folio, 0);
 					copy_page(workspace->buf + i * PAGE_SIZE,
 						  data_in);
 					start += PAGE_SIZE;
@@ -163,11 +166,14 @@ int zlib_compress_pages(struct list_head *ws, struct address_space *mapping,
 			} else {
 				if (data_in) {
 					kunmap_local(data_in);
-					put_page(in_page);
+					folio_put(in_folio);
+					data_in = NULL;
 				}
-				in_page = find_get_page(mapping,
-							start >> PAGE_SHIFT);
-				data_in = kmap_local_page(in_page);
+				ret = btrfs_compress_filemap_get_folio(mapping,
+						start, &in_folio);
+				if (ret < 0)
+					goto out;
+				data_in = kmap_local_folio(in_folio, 0);
 				start += PAGE_SIZE;
 				workspace->strm.next_in = data_in;
 			}
@@ -196,20 +202,20 @@ int zlib_compress_pages(struct list_head *ws, struct address_space *mapping,
 		 * the stream end if required
 		 */
 		if (workspace->strm.avail_out == 0) {
-			if (nr_pages == nr_dest_pages) {
+			if (nr_folios == nr_dest_folios) {
 				ret = -E2BIG;
 				goto out;
 			}
-			out_page = btrfs_alloc_compr_page();
-			if (out_page == NULL) {
+			out_folio = btrfs_alloc_compr_folio();
+			if (out_folio == NULL) {
 				ret = -ENOMEM;
 				goto out;
 			}
-			cpage_out = page_address(out_page);
-			pages[nr_pages] = out_page;
-			nr_pages++;
+			cfolio_out = folio_address(out_folio);
+			folios[nr_folios] = out_folio;
+			nr_folios++;
 			workspace->strm.avail_out = PAGE_SIZE;
-			workspace->strm.next_out = cpage_out;
+			workspace->strm.next_out = cfolio_out;
 		}
 		/* we're all done */
 		if (workspace->strm.total_in >= len)
@@ -231,21 +237,21 @@ int zlib_compress_pages(struct list_head *ws, struct address_space *mapping,
 			ret = -EIO;
 			goto out;
 		} else if (workspace->strm.avail_out == 0) {
-			/* get another page for the stream end */
-			if (nr_pages == nr_dest_pages) {
+			/* Get another folio for the stream end. */
+			if (nr_folios == nr_dest_folios) {
 				ret = -E2BIG;
 				goto out;
 			}
-			out_page = btrfs_alloc_compr_page();
-			if (out_page == NULL) {
+			out_folio = btrfs_alloc_compr_folio();
+			if (out_folio == NULL) {
 				ret = -ENOMEM;
 				goto out;
 			}
-			cpage_out = page_address(out_page);
-			pages[nr_pages] = out_page;
-			nr_pages++;
+			cfolio_out = folio_address(out_folio);
+			folios[nr_folios] = out_folio;
+			nr_folios++;
 			workspace->strm.avail_out = PAGE_SIZE;
-			workspace->strm.next_out = cpage_out;
+			workspace->strm.next_out = cfolio_out;
 		}
 	}
 	zlib_deflateEnd(&workspace->strm);
@@ -259,10 +265,10 @@ int zlib_compress_pages(struct list_head *ws, struct address_space *mapping,
 	*total_out = workspace->strm.total_out;
 	*total_in = workspace->strm.total_in;
 out:
-	*out_pages = nr_pages;
+	*out_folios = nr_folios;
 	if (data_in) {
 		kunmap_local(data_in);
-		put_page(in_page);
+		folio_put(in_folio);
 	}
 
 	return ret;
@@ -275,13 +281,13 @@ int zlib_decompress_bio(struct list_head *ws, struct compressed_bio *cb)
 	int wbits = MAX_WBITS;
 	char *data_in;
 	size_t total_out = 0;
-	unsigned long page_in_index = 0;
+	unsigned long folio_in_index = 0;
 	size_t srclen = cb->compressed_len;
-	unsigned long total_pages_in = DIV_ROUND_UP(srclen, PAGE_SIZE);
+	unsigned long total_folios_in = DIV_ROUND_UP(srclen, PAGE_SIZE);
 	unsigned long buf_start;
-	struct page **pages_in = cb->compressed_pages;
+	struct folio **folios_in = cb->compressed_folios;
 
-	data_in = kmap_local_page(pages_in[page_in_index]);
+	data_in = kmap_local_folio(folios_in[folio_in_index], 0);
 	workspace->strm.next_in = data_in;
 	workspace->strm.avail_in = min_t(size_t, srclen, PAGE_SIZE);
 	workspace->strm.total_in = 0;
@@ -331,12 +337,12 @@ int zlib_decompress_bio(struct list_head *ws, struct compressed_bio *cb)
 		if (workspace->strm.avail_in == 0) {
 			unsigned long tmp;
 			kunmap_local(data_in);
-			page_in_index++;
-			if (page_in_index >= total_pages_in) {
+			folio_in_index++;
+			if (folio_in_index >= total_folios_in) {
 				data_in = NULL;
 				break;
 			}
-			data_in = kmap_local_page(pages_in[page_in_index]);
+			data_in = kmap_local_folio(folios_in[folio_in_index], 0);
 			workspace->strm.next_in = data_in;
 			tmp = srclen - workspace->strm.total_in;
 			workspace->strm.avail_in = min(tmp, PAGE_SIZE);
@@ -398,7 +404,7 @@ int zlib_decompress(struct list_head *ws, const u8 *data_in,
 
 out:
 	if (unlikely(to_copy != destlen)) {
-		pr_warn_ratelimited("BTRFS: infalte failed, decompressed=%lu expected=%zu\n",
+		pr_warn_ratelimited("BTRFS: inflate failed, decompressed=%lu expected=%zu\n",
 					to_copy, destlen);
 		ret = -EIO;
 	} else {
diff --git a/fs/btrfs/zoned.c b/fs/btrfs/zoned.c
index 9729fa29c7..947a87576f 100644
--- a/fs/btrfs/zoned.c
+++ b/fs/btrfs/zoned.c
@@ -12,10 +12,8 @@
 #include "rcu-string.h"
 #include "disk-io.h"
 #include "block-group.h"
-#include "transaction.h"
 #include "dev-replace.h"
 #include "space-info.h"
-#include "super.h"
 #include "fs.h"
 #include "accessors.h"
 #include "bio.h"
@@ -120,7 +118,7 @@ static int sb_write_pointer(struct block_device *bdev, struct blk_zone *zones,
 		return -ENOENT;
 	} else if (full[0] && full[1]) {
 		/* Compare two super blocks */
-		struct address_space *mapping = bdev->bd_inode->i_mapping;
+		struct address_space *mapping = bdev->bd_mapping;
 		struct page *page[BTRFS_NR_SB_LOG_ZONES];
 		struct btrfs_super_block *super[BTRFS_NR_SB_LOG_ZONES];
 		int i;
@@ -824,11 +822,14 @@ static int sb_log_location(struct block_device *bdev, struct blk_zone *zones,
 			reset = &zones[1];
 
 		if (reset && reset->cond != BLK_ZONE_COND_EMPTY) {
+			unsigned int nofs_flags;
+
 			ASSERT(sb_zone_is_full(reset));
 
+			nofs_flags = memalloc_nofs_save();
 			ret = blkdev_zone_mgmt(bdev, REQ_OP_ZONE_RESET,
-					       reset->start, reset->len,
-					       GFP_NOFS);
+					       reset->start, reset->len);
+			memalloc_nofs_restore(nofs_flags);
 			if (ret)
 				return ret;
 
@@ -974,11 +975,14 @@ int btrfs_advance_sb_log(struct btrfs_device *device, int mirror)
 			 * explicit ZONE_FINISH is not necessary.
 			 */
 			if (zone->wp != zone->start + zone->capacity) {
+				unsigned int nofs_flags;
 				int ret;
 
+				nofs_flags = memalloc_nofs_save();
 				ret = blkdev_zone_mgmt(device->bdev,
 						REQ_OP_ZONE_FINISH, zone->start,
-						zone->len, GFP_NOFS);
+						zone->len);
+				memalloc_nofs_restore(nofs_flags);
 				if (ret)
 					return ret;
 			}
@@ -996,11 +1000,13 @@ int btrfs_advance_sb_log(struct btrfs_device *device, int mirror)
 
 int btrfs_reset_sb_log_zones(struct block_device *bdev, int mirror)
 {
+	unsigned int nofs_flags;
 	sector_t zone_sectors;
 	sector_t nr_sectors;
 	u8 zone_sectors_shift;
 	u32 sb_zone;
 	u32 nr_zones;
+	int ret;
 
 	zone_sectors = bdev_zone_sectors(bdev);
 	zone_sectors_shift = ilog2(zone_sectors);
@@ -1011,9 +1017,12 @@ int btrfs_reset_sb_log_zones(struct block_device *bdev, int mirror)
 	if (sb_zone + 1 >= nr_zones)
 		return -ENOENT;
 
-	return blkdev_zone_mgmt(bdev, REQ_OP_ZONE_RESET,
-				zone_start_sector(sb_zone, bdev),
-				zone_sectors * BTRFS_NR_SB_LOG_ZONES, GFP_NOFS);
+	nofs_flags = memalloc_nofs_save();
+	ret = blkdev_zone_mgmt(bdev, REQ_OP_ZONE_RESET,
+			       zone_start_sector(sb_zone, bdev),
+			       zone_sectors * BTRFS_NR_SB_LOG_ZONES);
+	memalloc_nofs_restore(nofs_flags);
+	return ret;
 }
 
 /*
@@ -1124,12 +1133,14 @@ static void btrfs_dev_clear_active_zone(struct btrfs_device *device, u64 pos)
 int btrfs_reset_device_zone(struct btrfs_device *device, u64 physical,
 			    u64 length, u64 *bytes)
 {
+	unsigned int nofs_flags;
 	int ret;
 
 	*bytes = 0;
+	nofs_flags = memalloc_nofs_save();
 	ret = blkdev_zone_mgmt(device->bdev, REQ_OP_ZONE_RESET,
-			       physical >> SECTOR_SHIFT, length >> SECTOR_SHIFT,
-			       GFP_NOFS);
+			       physical >> SECTOR_SHIFT, length >> SECTOR_SHIFT);
+	memalloc_nofs_restore(nofs_flags);
 	if (ret)
 		return ret;
 
@@ -1279,7 +1290,7 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx,
 				struct btrfs_chunk_map *map)
 {
 	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
-	struct btrfs_device *device = map->stripes[zone_idx].dev;
+	struct btrfs_device *device;
 	int dev_replace_is_ongoing = 0;
 	unsigned int nofs_flag;
 	struct blk_zone zone;
@@ -1287,7 +1298,11 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx,
 
 	info->physical = map->stripes[zone_idx].physical;
 
+	down_read(&dev_replace->rwsem);
+	device = map->stripes[zone_idx].dev;
+
 	if (!device->bdev) {
+		up_read(&dev_replace->rwsem);
 		info->alloc_offset = WP_MISSING_DEV;
 		return 0;
 	}
@@ -1297,6 +1312,7 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx,
 		__set_bit(zone_idx, active);
 
 	if (!btrfs_dev_is_sequential(device, info->physical)) {
+		up_read(&dev_replace->rwsem);
 		info->alloc_offset = WP_CONVENTIONAL;
 		return 0;
 	}
@@ -1304,11 +1320,9 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx,
 	/* This zone will be used for allocation, so mark this zone non-empty. */
 	btrfs_dev_clear_zone_empty(device, info->physical);
 
-	down_read(&dev_replace->rwsem);
 	dev_replace_is_ongoing = btrfs_dev_replace_is_ongoing(dev_replace);
 	if (dev_replace_is_ongoing && dev_replace->tgtdev != NULL)
 		btrfs_dev_clear_zone_empty(dev_replace->tgtdev, info->physical);
-	up_read(&dev_replace->rwsem);
 
 	/*
 	 * The group is mapped to a sequential zone. Get the zone write pointer
@@ -1319,6 +1333,7 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx,
 	ret = btrfs_get_dev_zone(device, info->physical, &zone);
 	memalloc_nofs_restore(nofs_flag);
 	if (ret) {
+		up_read(&dev_replace->rwsem);
 		if (ret != -EIO && ret != -EOPNOTSUPP)
 			return ret;
 		info->alloc_offset = WP_MISSING_DEV;
@@ -1330,6 +1345,7 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx,
 		"zoned: unexpected conventional zone %llu on device %s (devid %llu)",
 			zone.start << SECTOR_SHIFT, rcu_str_deref(device->name),
 			device->devid);
+		up_read(&dev_replace->rwsem);
 		return -EIO;
 	}
 
@@ -1357,6 +1373,8 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx,
 		break;
 	}
 
+	up_read(&dev_replace->rwsem);
+
 	return 0;
 }
 
@@ -2241,14 +2259,16 @@ static int do_zone_finish(struct btrfs_block_group *block_group, bool fully_writ
 		struct btrfs_device *device = map->stripes[i].dev;
 		const u64 physical = map->stripes[i].physical;
 		struct btrfs_zoned_device_info *zinfo = device->zone_info;
+		unsigned int nofs_flags;
 
 		if (zinfo->max_active_zones == 0)
 			continue;
 
+		nofs_flags = memalloc_nofs_save();
 		ret = blkdev_zone_mgmt(device->bdev, REQ_OP_ZONE_FINISH,
 				       physical >> SECTOR_SHIFT,
-				       zinfo->zone_size >> SECTOR_SHIFT,
-				       GFP_NOFS);
+				       zinfo->zone_size >> SECTOR_SHIFT);
+		memalloc_nofs_restore(nofs_flags);
 
 		if (ret) {
 			up_read(&dev_replace->rwsem);
diff --git a/fs/btrfs/zoned.h b/fs/btrfs/zoned.h
index f573bda496..77c4321e33 100644
--- a/fs/btrfs/zoned.h
+++ b/fs/btrfs/zoned.h
@@ -4,12 +4,27 @@
 #define BTRFS_ZONED_H
 
 #include <linux/types.h>
+#include <linux/atomic.h>
 #include <linux/blkdev.h>
+#include <linux/blkzoned.h>
+#include <linux/errno.h>
+#include <linux/spinlock.h>
+#include <linux/mutex.h>
 #include "messages.h"
 #include "volumes.h"
 #include "disk-io.h"
 #include "block-group.h"
 #include "btrfs_inode.h"
+#include "fs.h"
+
+struct block_device;
+struct extent_buffer;
+struct btrfs_bio;
+struct btrfs_ordered_extent;
+struct btrfs_fs_info;
+struct btrfs_space_info;
+struct btrfs_eb_write_context;
+struct btrfs_fs_devices;
 
 #define BTRFS_DEFAULT_RECLAIM_THRESH           			(75)
 
diff --git a/fs/btrfs/zstd.c b/fs/btrfs/zstd.c
index 0d66db8bc1..2b232b82c3 100644
--- a/fs/btrfs/zstd.c
+++ b/fs/btrfs/zstd.c
@@ -18,8 +18,9 @@
 #include <linux/slab.h>
 #include <linux/zstd.h>
 #include "misc.h"
+#include "fs.h"
 #include "compression.h"
-#include "ctree.h"
+#include "super.h"
 
 #define ZSTD_BTRFS_MAX_WINDOWLOG 17
 #define ZSTD_BTRFS_MAX_INPUT (1 << ZSTD_BTRFS_MAX_WINDOWLOG)
@@ -373,25 +374,25 @@ fail:
 	return ERR_PTR(-ENOMEM);
 }
 
-int zstd_compress_pages(struct list_head *ws, struct address_space *mapping,
-		u64 start, struct page **pages, unsigned long *out_pages,
-		unsigned long *total_in, unsigned long *total_out)
+int zstd_compress_folios(struct list_head *ws, struct address_space *mapping,
+			 u64 start, struct folio **folios, unsigned long *out_folios,
+			 unsigned long *total_in, unsigned long *total_out)
 {
 	struct workspace *workspace = list_entry(ws, struct workspace, list);
 	zstd_cstream *stream;
 	int ret = 0;
-	int nr_pages = 0;
-	struct page *in_page = NULL;  /* The current page to read */
-	struct page *out_page = NULL; /* The current page to write to */
+	int nr_folios = 0;
+	struct folio *in_folio = NULL;  /* The current folio to read. */
+	struct folio *out_folio = NULL; /* The current folio to write to. */
 	unsigned long tot_in = 0;
 	unsigned long tot_out = 0;
 	unsigned long len = *total_out;
-	const unsigned long nr_dest_pages = *out_pages;
-	unsigned long max_out = nr_dest_pages * PAGE_SIZE;
+	const unsigned long nr_dest_folios = *out_folios;
+	unsigned long max_out = nr_dest_folios * PAGE_SIZE;
 	zstd_parameters params = zstd_get_btrfs_parameters(workspace->req_level,
 							   len);
 
-	*out_pages = 0;
+	*out_folios = 0;
 	*total_out = 0;
 	*total_in = 0;
 
@@ -405,19 +406,21 @@ int zstd_compress_pages(struct list_head *ws, struct address_space *mapping,
 	}
 
 	/* map in the first page of input data */
-	in_page = find_get_page(mapping, start >> PAGE_SHIFT);
-	workspace->in_buf.src = kmap_local_page(in_page);
+	ret = btrfs_compress_filemap_get_folio(mapping, start, &in_folio);
+	if (ret < 0)
+		goto out;
+	workspace->in_buf.src = kmap_local_folio(in_folio, 0);
 	workspace->in_buf.pos = 0;
 	workspace->in_buf.size = min_t(size_t, len, PAGE_SIZE);
 
 	/* Allocate and map in the output buffer */
-	out_page = btrfs_alloc_compr_page();
-	if (out_page == NULL) {
+	out_folio = btrfs_alloc_compr_folio();
+	if (out_folio == NULL) {
 		ret = -ENOMEM;
 		goto out;
 	}
-	pages[nr_pages++] = out_page;
-	workspace->out_buf.dst = page_address(out_page);
+	folios[nr_folios++] = out_folio;
+	workspace->out_buf.dst = folio_address(out_folio);
 	workspace->out_buf.pos = 0;
 	workspace->out_buf.size = min_t(size_t, max_out, PAGE_SIZE);
 
@@ -452,17 +455,17 @@ int zstd_compress_pages(struct list_head *ws, struct address_space *mapping,
 		if (workspace->out_buf.pos == workspace->out_buf.size) {
 			tot_out += PAGE_SIZE;
 			max_out -= PAGE_SIZE;
-			if (nr_pages == nr_dest_pages) {
+			if (nr_folios == nr_dest_folios) {
 				ret = -E2BIG;
 				goto out;
 			}
-			out_page = btrfs_alloc_compr_page();
-			if (out_page == NULL) {
+			out_folio = btrfs_alloc_compr_folio();
+			if (out_folio == NULL) {
 				ret = -ENOMEM;
 				goto out;
 			}
-			pages[nr_pages++] = out_page;
-			workspace->out_buf.dst = page_address(out_page);
+			folios[nr_folios++] = out_folio;
+			workspace->out_buf.dst = folio_address(out_folio);
 			workspace->out_buf.pos = 0;
 			workspace->out_buf.size = min_t(size_t, max_out,
 							PAGE_SIZE);
@@ -478,11 +481,14 @@ int zstd_compress_pages(struct list_head *ws, struct address_space *mapping,
 		if (workspace->in_buf.pos == workspace->in_buf.size) {
 			tot_in += PAGE_SIZE;
 			kunmap_local(workspace->in_buf.src);
-			put_page(in_page);
+			workspace->in_buf.src = NULL;
+			folio_put(in_folio);
 			start += PAGE_SIZE;
 			len -= PAGE_SIZE;
-			in_page = find_get_page(mapping, start >> PAGE_SHIFT);
-			workspace->in_buf.src = kmap_local_page(in_page);
+			ret = btrfs_compress_filemap_get_folio(mapping, start, &in_folio);
+			if (ret < 0)
+				goto out;
+			workspace->in_buf.src = kmap_local_folio(in_folio, 0);
 			workspace->in_buf.pos = 0;
 			workspace->in_buf.size = min_t(size_t, len, PAGE_SIZE);
 		}
@@ -509,17 +515,17 @@ int zstd_compress_pages(struct list_head *ws, struct address_space *mapping,
 
 		tot_out += PAGE_SIZE;
 		max_out -= PAGE_SIZE;
-		if (nr_pages == nr_dest_pages) {
+		if (nr_folios == nr_dest_folios) {
 			ret = -E2BIG;
 			goto out;
 		}
-		out_page = btrfs_alloc_compr_page();
-		if (out_page == NULL) {
+		out_folio = btrfs_alloc_compr_folio();
+		if (out_folio == NULL) {
 			ret = -ENOMEM;
 			goto out;
 		}
-		pages[nr_pages++] = out_page;
-		workspace->out_buf.dst = page_address(out_page);
+		folios[nr_folios++] = out_folio;
+		workspace->out_buf.dst = folio_address(out_folio);
 		workspace->out_buf.pos = 0;
 		workspace->out_buf.size = min_t(size_t, max_out, PAGE_SIZE);
 	}
@@ -533,10 +539,10 @@ int zstd_compress_pages(struct list_head *ws, struct address_space *mapping,
 	*total_in = tot_in;
 	*total_out = tot_out;
 out:
-	*out_pages = nr_pages;
+	*out_folios = nr_folios;
 	if (workspace->in_buf.src) {
 		kunmap_local(workspace->in_buf.src);
-		put_page(in_page);
+		folio_put(in_folio);
 	}
 	return ret;
 }
@@ -544,12 +550,12 @@ out:
 int zstd_decompress_bio(struct list_head *ws, struct compressed_bio *cb)
 {
 	struct workspace *workspace = list_entry(ws, struct workspace, list);
-	struct page **pages_in = cb->compressed_pages;
+	struct folio **folios_in = cb->compressed_folios;
 	size_t srclen = cb->compressed_len;
 	zstd_dstream *stream;
 	int ret = 0;
-	unsigned long page_in_index = 0;
-	unsigned long total_pages_in = DIV_ROUND_UP(srclen, PAGE_SIZE);
+	unsigned long folio_in_index = 0;
+	unsigned long total_folios_in = DIV_ROUND_UP(srclen, PAGE_SIZE);
 	unsigned long buf_start;
 	unsigned long total_out = 0;
 
@@ -561,7 +567,7 @@ int zstd_decompress_bio(struct list_head *ws, struct compressed_bio *cb)
 		goto done;
 	}
 
-	workspace->in_buf.src = kmap_local_page(pages_in[page_in_index]);
+	workspace->in_buf.src = kmap_local_folio(folios_in[folio_in_index], 0);
 	workspace->in_buf.pos = 0;
 	workspace->in_buf.size = min_t(size_t, srclen, PAGE_SIZE);
 
@@ -598,14 +604,15 @@ int zstd_decompress_bio(struct list_head *ws, struct compressed_bio *cb)
 
 		if (workspace->in_buf.pos == workspace->in_buf.size) {
 			kunmap_local(workspace->in_buf.src);
-			page_in_index++;
-			if (page_in_index >= total_pages_in) {
+			folio_in_index++;
+			if (folio_in_index >= total_folios_in) {
 				workspace->in_buf.src = NULL;
 				ret = -EIO;
 				goto done;
 			}
 			srclen -= PAGE_SIZE;
-			workspace->in_buf.src = kmap_local_page(pages_in[page_in_index]);
+			workspace->in_buf.src =
+				kmap_local_folio(folios_in[folio_in_index], 0);
 			workspace->in_buf.pos = 0;
 			workspace->in_buf.size = min_t(size_t, srclen, PAGE_SIZE);
 		}
@@ -618,80 +625,48 @@ done:
 }
 
 int zstd_decompress(struct list_head *ws, const u8 *data_in,
-		struct page *dest_page, unsigned long start_byte, size_t srclen,
+		struct page *dest_page, unsigned long dest_pgoff, size_t srclen,
 		size_t destlen)
 {
 	struct workspace *workspace = list_entry(ws, struct workspace, list);
+	struct btrfs_fs_info *fs_info = btrfs_sb(dest_page->mapping->host->i_sb);
+	const u32 sectorsize = fs_info->sectorsize;
 	zstd_dstream *stream;
 	int ret = 0;
-	size_t ret2;
-	unsigned long total_out = 0;
-	unsigned long pg_offset = 0;
+	unsigned long to_copy = 0;
 
 	stream = zstd_init_dstream(
 			ZSTD_BTRFS_MAX_INPUT, workspace->mem, workspace->size);
 	if (!stream) {
 		pr_warn("BTRFS: zstd_init_dstream failed\n");
-		ret = -EIO;
 		goto finish;
 	}
 
-	destlen = min_t(size_t, destlen, PAGE_SIZE);
-
 	workspace->in_buf.src = data_in;
 	workspace->in_buf.pos = 0;
 	workspace->in_buf.size = srclen;
 
 	workspace->out_buf.dst = workspace->buf;
 	workspace->out_buf.pos = 0;
-	workspace->out_buf.size = PAGE_SIZE;
-
-	ret2 = 1;
-	while (pg_offset < destlen
-	       && workspace->in_buf.pos < workspace->in_buf.size) {
-		unsigned long buf_start;
-		unsigned long buf_offset;
-		unsigned long bytes;
-
-		/* Check if the frame is over and we still need more input */
-		if (ret2 == 0) {
-			pr_debug("BTRFS: zstd_decompress_stream ended early\n");
-			ret = -EIO;
-			goto finish;
-		}
-		ret2 = zstd_decompress_stream(stream, &workspace->out_buf,
-				&workspace->in_buf);
-		if (zstd_is_error(ret2)) {
-			pr_debug("BTRFS: zstd_decompress_stream returned %d\n",
-					zstd_get_error_code(ret2));
-			ret = -EIO;
-			goto finish;
-		}
-
-		buf_start = total_out;
-		total_out += workspace->out_buf.pos;
-		workspace->out_buf.pos = 0;
-
-		if (total_out <= start_byte)
-			continue;
-
-		if (total_out > start_byte && buf_start < start_byte)
-			buf_offset = start_byte - buf_start;
-		else
-			buf_offset = 0;
-
-		bytes = min_t(unsigned long, destlen - pg_offset,
-				workspace->out_buf.size - buf_offset);
-
-		memcpy_to_page(dest_page, pg_offset,
-			       workspace->out_buf.dst + buf_offset, bytes);
-
-		pg_offset += bytes;
+	workspace->out_buf.size = sectorsize;
+
+	/*
+	 * Since both input and output buffers should not exceed one sector,
+	 * one call should end the decompression.
+	 */
+	ret = zstd_decompress_stream(stream, &workspace->out_buf, &workspace->in_buf);
+	if (zstd_is_error(ret)) {
+		pr_warn_ratelimited("BTRFS: zstd_decompress_stream return %d\n",
+				    zstd_get_error_code(ret));
+		goto finish;
 	}
-	ret = 0;
+	to_copy = workspace->out_buf.pos;
+	memcpy_to_page(dest_page, dest_pgoff, workspace->out_buf.dst, to_copy);
 finish:
-	if (pg_offset < destlen) {
-		memzero_page(dest_page, pg_offset, destlen - pg_offset);
+	/* Error or early end. */
+	if (unlikely(to_copy < destlen)) {
+		ret = -EIO;
+		memzero_page(dest_page, dest_pgoff + to_copy, destlen - to_copy);
 	}
 	return ret;
 }
diff --git a/fs/buffer.c b/fs/buffer.c
index d3bcf601d3..645f0387df 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -55,7 +55,7 @@
 
 static int fsync_buffers_list(spinlock_t *lock, struct list_head *list);
 static void submit_bh_wbc(blk_opf_t opf, struct buffer_head *bh,
-			  struct writeback_control *wbc);
+			  enum rw_hint hint, struct writeback_control *wbc);
 
 #define BH_ENTRY(list) list_entry((list), struct buffer_head, b_assoc_buffers)
 
@@ -189,8 +189,8 @@ EXPORT_SYMBOL(end_buffer_write_sync);
 static struct buffer_head *
 __find_get_block_slow(struct block_device *bdev, sector_t block)
 {
-	struct inode *bd_inode = bdev->bd_inode;
-	struct address_space *bd_mapping = bd_inode->i_mapping;
+	struct address_space *bd_mapping = bdev->bd_mapping;
+	const int blkbits = bd_mapping->host->i_blkbits;
 	struct buffer_head *ret = NULL;
 	pgoff_t index;
 	struct buffer_head *bh;
@@ -199,7 +199,7 @@ __find_get_block_slow(struct block_device *bdev, sector_t block)
 	int all_mapped = 1;
 	static DEFINE_RATELIMIT_STATE(last_warned, HZ, 1);
 
-	index = ((loff_t)block << bd_inode->i_blkbits) / PAGE_SIZE;
+	index = ((loff_t)block << blkbits) / PAGE_SIZE;
 	folio = __filemap_get_folio(bd_mapping, index, FGP_ACCESSED, 0);
 	if (IS_ERR(folio))
 		goto out;
@@ -233,7 +233,7 @@ __find_get_block_slow(struct block_device *bdev, sector_t block)
 		       (unsigned long long)block,
 		       (unsigned long long)bh->b_blocknr,
 		       bh->b_state, bh->b_size, bdev,
-		       1 << bd_inode->i_blkbits);
+		       1 << blkbits);
 	}
 out_unlock:
 	spin_unlock(&bd_mapping->i_private_lock);
@@ -464,7 +464,7 @@ EXPORT_SYMBOL(mark_buffer_async_write);
  * a successful fsync().  For example, ext2 indirect blocks need to be
  * written back and waited upon before fsync() returns.
  *
- * The functions mark_buffer_inode_dirty(), fsync_inode_buffers(),
+ * The functions mark_buffer_dirty_inode(), fsync_inode_buffers(),
  * inode_has_buffers() and invalidate_inode_buffers() are provided for the
  * management of a list of dependent buffers at ->i_mapping->i_private_list.
  *
@@ -687,30 +687,37 @@ void mark_buffer_dirty_inode(struct buffer_head *bh, struct inode *inode)
 }
 EXPORT_SYMBOL(mark_buffer_dirty_inode);
 
-/*
- * Add a page to the dirty page list.
- *
- * It is a sad fact of life that this function is called from several places
- * deeply under spinlocking.  It may not sleep.
- *
- * If the page has buffers, the uptodate buffers are set dirty, to preserve
- * dirty-state coherency between the page and the buffers.  It the page does
- * not have buffers then when they are later attached they will all be set
- * dirty.
- *
- * The buffers are dirtied before the page is dirtied.  There's a small race
- * window in which a writepage caller may see the page cleanness but not the
- * buffer dirtiness.  That's fine.  If this code were to set the page dirty
- * before the buffers, a concurrent writepage caller could clear the page dirty
- * bit, see a bunch of clean buffers and we'd end up with dirty buffers/clean
- * page on the dirty page list.
- *
- * We use i_private_lock to lock against try_to_free_buffers while using the
- * page's buffer list.  Also use this to protect against clean buffers being
- * added to the page after it was set dirty.
- *
- * FIXME: may need to call ->reservepage here as well.  That's rather up to the
- * address_space though.
+/**
+ * block_dirty_folio - Mark a folio as dirty.
+ * @mapping: The address space containing this folio.
+ * @folio: The folio to mark dirty.
+ *
+ * Filesystems which use buffer_heads can use this function as their
+ * ->dirty_folio implementation.  Some filesystems need to do a little
+ * work before calling this function.  Filesystems which do not use
+ * buffer_heads should call filemap_dirty_folio() instead.
+ *
+ * If the folio has buffers, the uptodate buffers are set dirty, to
+ * preserve dirty-state coherency between the folio and the buffers.
+ * Buffers added to a dirty folio are created dirty.
+ *
+ * The buffers are dirtied before the folio is dirtied.  There's a small
+ * race window in which writeback may see the folio cleanness but not the
+ * buffer dirtiness.  That's fine.  If this code were to set the folio
+ * dirty before the buffers, writeback could clear the folio dirty flag,
+ * see a bunch of clean buffers and we'd end up with dirty buffers/clean
+ * folio on the dirty folio list.
+ *
+ * We use i_private_lock to lock against try_to_free_buffers() while
+ * using the folio's buffer list.  This also prevents clean buffers
+ * being added to the folio after it was set dirty.
+ *
+ * Context: May only be called from process context.  Does not sleep.
+ * Caller must ensure that @folio cannot be truncated during this call,
+ * typically by holding the folio lock or having a page in the folio
+ * mapped and holding the page table lock.
+ *
+ * Return: True if the folio was dirtied; false if it was already dirtied.
  */
 bool block_dirty_folio(struct address_space *mapping, struct folio *folio)
 {
@@ -1034,12 +1041,12 @@ static sector_t folio_init_buffers(struct folio *folio,
 static bool grow_dev_folio(struct block_device *bdev, sector_t block,
 		pgoff_t index, unsigned size, gfp_t gfp)
 {
-	struct inode *inode = bdev->bd_inode;
+	struct address_space *mapping = bdev->bd_mapping;
 	struct folio *folio;
 	struct buffer_head *bh;
 	sector_t end_block = 0;
 
-	folio = __filemap_get_folio(inode->i_mapping, index,
+	folio = __filemap_get_folio(mapping, index,
 			FGP_LOCK | FGP_ACCESSED | FGP_CREAT, gfp);
 	if (IS_ERR(folio))
 		return false;
@@ -1073,10 +1080,10 @@ static bool grow_dev_folio(struct block_device *bdev, sector_t block,
 	 * lock to be atomic wrt __find_get_block(), which does not
 	 * run under the folio lock.
 	 */
-	spin_lock(&inode->i_mapping->i_private_lock);
+	spin_lock(&mapping->i_private_lock);
 	link_dev_buffers(folio, bh);
 	end_block = folio_init_buffers(folio, bdev, size);
-	spin_unlock(&inode->i_mapping->i_private_lock);
+	spin_unlock(&mapping->i_private_lock);
 unlock:
 	folio_unlock(folio);
 	folio_put(folio);
@@ -1219,26 +1226,28 @@ void mark_buffer_write_io_error(struct buffer_head *bh)
 }
 EXPORT_SYMBOL(mark_buffer_write_io_error);
 
-/*
- * Decrement a buffer_head's reference count.  If all buffers against a page
- * have zero reference count, are clean and unlocked, and if the page is clean
- * and unlocked then try_to_free_buffers() may strip the buffers from the page
- * in preparation for freeing it (sometimes, rarely, buffers are removed from
- * a page but it ends up not being freed, and buffers may later be reattached).
+/**
+ * __brelse - Release a buffer.
+ * @bh: The buffer to release.
+ *
+ * This variant of brelse() can be called if @bh is guaranteed to not be NULL.
  */
-void __brelse(struct buffer_head * buf)
+void __brelse(struct buffer_head *bh)
 {
-	if (atomic_read(&buf->b_count)) {
-		put_bh(buf);
+	if (atomic_read(&bh->b_count)) {
+		put_bh(bh);
 		return;
 	}
 	WARN(1, KERN_ERR "VFS: brelse: Trying to free free buffer\n");
 }
 EXPORT_SYMBOL(__brelse);
 
-/*
- * bforget() is like brelse(), except it discards any
- * potentially dirty data.
+/**
+ * __bforget - Discard any dirty data in a buffer.
+ * @bh: The buffer to forget.
+ *
+ * This variant of bforget() can be called if @bh is guaranteed to not
+ * be NULL.
  */
 void __bforget(struct buffer_head *bh)
 {
@@ -1415,6 +1424,11 @@ EXPORT_SYMBOL(__find_get_block);
  * @size: The size of buffer_heads for this @bdev.
  * @gfp: The memory allocation flags to use.
  *
+ * The returned buffer head has its reference count incremented, but is
+ * not locked.  The caller should call brelse() when it has finished
+ * with the buffer.  The buffer may not be uptodate.  If needed, the
+ * caller can bring it uptodate either by reading it or overwriting it.
+ *
  * Return: The buffer head, or NULL if memory could not be allocated.
  */
 struct buffer_head *bdev_getblk(struct block_device *bdev, sector_t block,
@@ -1446,24 +1460,33 @@ void __breadahead(struct block_device *bdev, sector_t block, unsigned size)
 EXPORT_SYMBOL(__breadahead);
 
 /**
- *  __bread_gfp() - reads a specified block and returns the bh
- *  @bdev: the block_device to read from
- *  @block: number of block
- *  @size: size (in bytes) to read
- *  @gfp: page allocation flag
- *
- *  Reads a specified block, and returns buffer head that contains it.
- *  The page cache can be allocated from non-movable area
- *  not to prevent page migration if you set gfp to zero.
- *  It returns NULL if the block was unreadable.
+ * __bread_gfp() - Read a block.
+ * @bdev: The block device to read from.
+ * @block: Block number in units of block size.
+ * @size: The block size of this device in bytes.
+ * @gfp: Not page allocation flags; see below.
+ *
+ * You are not expected to call this function.  You should use one of
+ * sb_bread(), sb_bread_unmovable() or __bread().
+ *
+ * Read a specified block, and return the buffer head that refers to it.
+ * If @gfp is 0, the memory will be allocated using the block device's
+ * default GFP flags.  If @gfp is __GFP_MOVABLE, the memory may be
+ * allocated from a movable area.  Do not pass in a complete set of
+ * GFP flags.
+ *
+ * The returned buffer head has its refcount increased.  The caller should
+ * call brelse() when it has finished with the buffer.
+ *
+ * Context: May sleep waiting for I/O.
+ * Return: NULL if the block was unreadable.
  */
-struct buffer_head *
-__bread_gfp(struct block_device *bdev, sector_t block,
-		   unsigned size, gfp_t gfp)
+struct buffer_head *__bread_gfp(struct block_device *bdev, sector_t block,
+		unsigned size, gfp_t gfp)
 {
 	struct buffer_head *bh;
 
-	gfp |= mapping_gfp_constraint(bdev->bd_inode->i_mapping, ~__GFP_FS);
+	gfp |= mapping_gfp_constraint(bdev->bd_mapping, ~__GFP_FS);
 
 	/*
 	 * Prefer looping in the allocator rather than here, at least that
@@ -1696,16 +1719,16 @@ EXPORT_SYMBOL(create_empty_buffers);
  */
 void clean_bdev_aliases(struct block_device *bdev, sector_t block, sector_t len)
 {
-	struct inode *bd_inode = bdev->bd_inode;
-	struct address_space *bd_mapping = bd_inode->i_mapping;
+	struct address_space *bd_mapping = bdev->bd_mapping;
+	const int blkbits = bd_mapping->host->i_blkbits;
 	struct folio_batch fbatch;
-	pgoff_t index = ((loff_t)block << bd_inode->i_blkbits) / PAGE_SIZE;
+	pgoff_t index = ((loff_t)block << blkbits) / PAGE_SIZE;
 	pgoff_t end;
 	int i, count;
 	struct buffer_head *bh;
 	struct buffer_head *head;
 
-	end = ((loff_t)(block + len - 1) << bd_inode->i_blkbits) / PAGE_SIZE;
+	end = ((loff_t)(block + len - 1) << blkbits) / PAGE_SIZE;
 	folio_batch_init(&fbatch);
 	while (filemap_get_folios(bd_mapping, &index, end, &fbatch)) {
 		count = folio_batch_count(&fbatch);
@@ -1889,7 +1912,8 @@ int __block_write_full_folio(struct inode *inode, struct folio *folio,
 	do {
 		struct buffer_head *next = bh->b_this_page;
 		if (buffer_async_write(bh)) {
-			submit_bh_wbc(REQ_OP_WRITE | write_flags, bh, wbc);
+			submit_bh_wbc(REQ_OP_WRITE | write_flags, bh,
+				      inode->i_write_hint, wbc);
 			nr_underway++;
 		}
 		bh = next;
@@ -1944,7 +1968,8 @@ recover:
 		struct buffer_head *next = bh->b_this_page;
 		if (buffer_async_write(bh)) {
 			clear_buffer_dirty(bh);
-			submit_bh_wbc(REQ_OP_WRITE | write_flags, bh, wbc);
+			submit_bh_wbc(REQ_OP_WRITE | write_flags, bh,
+				      inode->i_write_hint, wbc);
 			nr_underway++;
 		}
 		bh = next;
@@ -2162,6 +2187,8 @@ static void __block_commit_write(struct folio *folio, size_t from, size_t to)
 	struct buffer_head *bh, *head;
 
 	bh = head = folio_buffers(folio);
+	if (!bh)
+		return;
 	blocksize = bh->b_size;
 
 	block_start = 0;
@@ -2756,6 +2783,7 @@ static void end_bio_bh_io_sync(struct bio *bio)
 }
 
 static void submit_bh_wbc(blk_opf_t opf, struct buffer_head *bh,
+			  enum rw_hint write_hint,
 			  struct writeback_control *wbc)
 {
 	const enum req_op op = opf & REQ_OP_MASK;
@@ -2783,6 +2811,7 @@ static void submit_bh_wbc(blk_opf_t opf, struct buffer_head *bh,
 	fscrypt_set_bio_crypt_ctx_bh(bio, bh, GFP_NOIO);
 
 	bio->bi_iter.bi_sector = bh->b_blocknr * (bh->b_size >> 9);
+	bio->bi_write_hint = write_hint;
 
 	__bio_add_page(bio, bh->b_page, bh->b_size, bh_offset(bh));
 
@@ -2802,7 +2831,7 @@ static void submit_bh_wbc(blk_opf_t opf, struct buffer_head *bh,
 
 void submit_bh(blk_opf_t opf, struct buffer_head *bh)
 {
-	submit_bh_wbc(opf, bh, NULL);
+	submit_bh_wbc(opf, bh, WRITE_LIFE_NOT_SET, NULL);
 }
 EXPORT_SYMBOL(submit_bh);
 
@@ -2857,26 +2886,6 @@ int sync_dirty_buffer(struct buffer_head *bh)
 }
 EXPORT_SYMBOL(sync_dirty_buffer);
 
-/*
- * try_to_free_buffers() checks if all the buffers on this particular folio
- * are unused, and releases them if so.
- *
- * Exclusion against try_to_free_buffers may be obtained by either
- * locking the folio or by holding its mapping's i_private_lock.
- *
- * If the folio is dirty but all the buffers are clean then we need to
- * be sure to mark the folio clean as well.  This is because the folio
- * may be against a block device, and a later reattachment of buffers
- * to a dirty folio will set *all* buffers dirty.  Which would corrupt
- * filesystem data on the same device.
- *
- * The same applies to regular filesystem folios: if all the buffers are
- * clean then we set the folio clean and proceed.  To do that, we require
- * total exclusion from block_dirty_folio().  That is obtained with
- * i_private_lock.
- *
- * try_to_free_buffers() is non-blocking.
- */
 static inline int buffer_busy(struct buffer_head *bh)
 {
 	return atomic_read(&bh->b_count) |
@@ -2910,6 +2919,30 @@ failed:
 	return false;
 }
 
+/**
+ * try_to_free_buffers - Release buffers attached to this folio.
+ * @folio: The folio.
+ *
+ * If any buffers are in use (dirty, under writeback, elevated refcount),
+ * no buffers will be freed.
+ *
+ * If the folio is dirty but all the buffers are clean then we need to
+ * be sure to mark the folio clean as well.  This is because the folio
+ * may be against a block device, and a later reattachment of buffers
+ * to a dirty folio will set *all* buffers dirty.  Which would corrupt
+ * filesystem data on the same device.
+ *
+ * The same applies to regular filesystem folios: if all the buffers are
+ * clean then we set the folio clean and proceed.  To do that, we require
+ * total exclusion from block_dirty_folio().  That is obtained with
+ * i_private_lock.
+ *
+ * Exclusion against try_to_free_buffers may be obtained by either
+ * locking the folio or by holding its mapping's i_private_lock.
+ *
+ * Context: Process context.  @folio must be locked.  Will not sleep.
+ * Return: true if all buffers attached to this folio were freed.
+ */
 bool try_to_free_buffers(struct folio *folio)
 {
 	struct address_space * const mapping = folio->mapping;
@@ -3121,12 +3154,8 @@ void __init buffer_init(void)
 	unsigned long nrpages;
 	int ret;
 
-	bh_cachep = kmem_cache_create("buffer_head",
-			sizeof(struct buffer_head), 0,
-				(SLAB_RECLAIM_ACCOUNT|SLAB_PANIC|
-				SLAB_MEM_SPREAD),
-				NULL);
-
+	bh_cachep = KMEM_CACHE(buffer_head,
+				SLAB_RECLAIM_ACCOUNT|SLAB_PANIC);
 	/*
 	 * Limit the bh occupancy to 10% of ZONE_NORMAL
 	 */
diff --git a/fs/cachefiles/cache.c b/fs/cachefiles/cache.c
index f449f7340a..9fb06dc165 100644
--- a/fs/cachefiles/cache.c
+++ b/fs/cachefiles/cache.c
@@ -8,6 +8,7 @@
 #include <linux/slab.h>
 #include <linux/statfs.h>
 #include <linux/namei.h>
+#include <trace/events/fscache.h>
 #include "internal.h"
 
 /*
@@ -312,19 +313,59 @@ static void cachefiles_withdraw_objects(struct cachefiles_cache *cache)
 }
 
 /*
- * Withdraw volumes.
+ * Withdraw fscache volumes.
+ */
+static void cachefiles_withdraw_fscache_volumes(struct cachefiles_cache *cache)
+{
+	struct list_head *cur;
+	struct cachefiles_volume *volume;
+	struct fscache_volume *vcookie;
+
+	_enter("");
+retry:
+	spin_lock(&cache->object_list_lock);
+	list_for_each(cur, &cache->volumes) {
+		volume = list_entry(cur, struct cachefiles_volume, cache_link);
+
+		if (atomic_read(&volume->vcookie->n_accesses) == 0)
+			continue;
+
+		vcookie = fscache_try_get_volume(volume->vcookie,
+						 fscache_volume_get_withdraw);
+		if (vcookie) {
+			spin_unlock(&cache->object_list_lock);
+			fscache_withdraw_volume(vcookie);
+			fscache_put_volume(vcookie, fscache_volume_put_withdraw);
+			goto retry;
+		}
+	}
+	spin_unlock(&cache->object_list_lock);
+
+	_leave("");
+}
+
+/*
+ * Withdraw cachefiles volumes.
  */
 static void cachefiles_withdraw_volumes(struct cachefiles_cache *cache)
 {
 	_enter("");
 
 	for (;;) {
+		struct fscache_volume *vcookie = NULL;
 		struct cachefiles_volume *volume = NULL;
 
 		spin_lock(&cache->object_list_lock);
 		if (!list_empty(&cache->volumes)) {
 			volume = list_first_entry(&cache->volumes,
 						  struct cachefiles_volume, cache_link);
+			vcookie = fscache_try_get_volume(volume->vcookie,
+							 fscache_volume_get_withdraw);
+			if (!vcookie) {
+				spin_unlock(&cache->object_list_lock);
+				cpu_relax();
+				continue;
+			}
 			list_del_init(&volume->cache_link);
 		}
 		spin_unlock(&cache->object_list_lock);
@@ -332,6 +373,7 @@ static void cachefiles_withdraw_volumes(struct cachefiles_cache *cache)
 			break;
 
 		cachefiles_withdraw_volume(volume);
+		fscache_put_volume(vcookie, fscache_volume_put_withdraw);
 	}
 
 	_leave("");
@@ -371,6 +413,7 @@ void cachefiles_withdraw_cache(struct cachefiles_cache *cache)
 	pr_info("File cache on %s unregistering\n", fscache->name);
 
 	fscache_withdraw_cache(fscache);
+	cachefiles_withdraw_fscache_volumes(cache);
 
 	/* we now have to destroy all the active objects pertaining to this
 	 * cache - which we do by passing them off to thread pool to be
diff --git a/fs/cachefiles/daemon.c b/fs/cachefiles/daemon.c
index 6465e25742..89b11336a8 100644
--- a/fs/cachefiles/daemon.c
+++ b/fs/cachefiles/daemon.c
@@ -133,7 +133,7 @@ static int cachefiles_daemon_open(struct inode *inode, struct file *file)
 	return 0;
 }
 
-static void cachefiles_flush_reqs(struct cachefiles_cache *cache)
+void cachefiles_flush_reqs(struct cachefiles_cache *cache)
 {
 	struct xarray *xa = &cache->reqs;
 	struct cachefiles_req *req;
@@ -159,6 +159,7 @@ static void cachefiles_flush_reqs(struct cachefiles_cache *cache)
 	xa_for_each(xa, index, req) {
 		req->error = -EIO;
 		complete(&req->done);
+		__xa_erase(xa, index);
 	}
 	xa_unlock(xa);
 
@@ -365,14 +366,14 @@ static __poll_t cachefiles_daemon_poll(struct file *file,
 
 	if (cachefiles_in_ondemand_mode(cache)) {
 		if (!xa_empty(&cache->reqs)) {
-			rcu_read_lock();
+			xas_lock(&xas);
 			xas_for_each_marked(&xas, req, ULONG_MAX, CACHEFILES_REQ_NEW) {
 				if (!cachefiles_ondemand_is_reopening_read(req)) {
 					mask |= EPOLLIN;
 					break;
 				}
 			}
-			rcu_read_unlock();
+			xas_unlock(&xas);
 		}
 	} else {
 		if (test_bit(CACHEFILES_STATE_CHANGED, &cache->flags))
diff --git a/fs/cachefiles/internal.h b/fs/cachefiles/internal.h
index d33169f001..7b99bd98de 100644
--- a/fs/cachefiles/internal.h
+++ b/fs/cachefiles/internal.h
@@ -48,6 +48,7 @@ enum cachefiles_object_state {
 	CACHEFILES_ONDEMAND_OBJSTATE_CLOSE, /* Anonymous fd closed by daemon or initial state */
 	CACHEFILES_ONDEMAND_OBJSTATE_OPEN, /* Anonymous fd associated with object is available */
 	CACHEFILES_ONDEMAND_OBJSTATE_REOPENING, /* Object that was closed and is being reopened. */
+	CACHEFILES_ONDEMAND_OBJSTATE_DROPPING, /* Object is being dropped. */
 };
 
 struct cachefiles_ondemand_info {
@@ -55,6 +56,7 @@ struct cachefiles_ondemand_info {
 	int				ondemand_id;
 	enum cachefiles_object_state	state;
 	struct cachefiles_object	*object;
+	spinlock_t			lock;
 };
 
 /*
@@ -127,6 +129,7 @@ struct cachefiles_cache {
 	unsigned long			req_id_next;
 	struct xarray			ondemand_ids;	/* xarray for ondemand_id allocation */
 	u32				ondemand_id_next;
+	u32				msg_id_next;
 };
 
 static inline bool cachefiles_in_ondemand_mode(struct cachefiles_cache *cache)
@@ -138,6 +141,7 @@ static inline bool cachefiles_in_ondemand_mode(struct cachefiles_cache *cache)
 struct cachefiles_req {
 	struct cachefiles_object *object;
 	struct completion done;
+	refcount_t ref;
 	int error;
 	struct cachefiles_msg msg;
 };
@@ -186,6 +190,7 @@ extern int cachefiles_has_space(struct cachefiles_cache *cache,
  * daemon.c
  */
 extern const struct file_operations cachefiles_daemon_fops;
+extern void cachefiles_flush_reqs(struct cachefiles_cache *cache);
 extern void cachefiles_get_unbind_pincount(struct cachefiles_cache *cache);
 extern void cachefiles_put_unbind_pincount(struct cachefiles_cache *cache);
 
@@ -332,6 +337,7 @@ cachefiles_ondemand_set_object_##_state(struct cachefiles_object *object) \
 CACHEFILES_OBJECT_STATE_FUNCS(open, OPEN);
 CACHEFILES_OBJECT_STATE_FUNCS(close, CLOSE);
 CACHEFILES_OBJECT_STATE_FUNCS(reopening, REOPENING);
+CACHEFILES_OBJECT_STATE_FUNCS(dropping, DROPPING);
 
 static inline bool cachefiles_ondemand_is_reopening_read(struct cachefiles_req *req)
 {
@@ -424,6 +430,8 @@ do {							\
 	pr_err("I/O Error: " FMT"\n", ##__VA_ARGS__);	\
 	fscache_io_error((___cache)->cache);		\
 	set_bit(CACHEFILES_DEAD, &(___cache)->flags);	\
+	if (cachefiles_in_ondemand_mode(___cache))	\
+		cachefiles_flush_reqs(___cache);	\
 } while (0)
 
 #define cachefiles_io_error_obj(object, FMT, ...)			\
diff --git a/fs/cachefiles/io.c b/fs/cachefiles/io.c
index 1d685357e6..e667dbcd20 100644
--- a/fs/cachefiles/io.c
+++ b/fs/cachefiles/io.c
@@ -9,6 +9,7 @@
 #include <linux/slab.h>
 #include <linux/file.h>
 #include <linux/uio.h>
+#include <linux/bio.h>
 #include <linux/falloc.h>
 #include <linux/sched/mm.h>
 #include <trace/events/fscache.h>
@@ -493,7 +494,7 @@ out_no_object:
  * boundary as appropriate.
  */
 static enum netfs_io_source cachefiles_prepare_read(struct netfs_io_subrequest *subreq,
-						    loff_t i_size)
+						    unsigned long long i_size)
 {
 	return cachefiles_do_prepare_read(&subreq->rreq->cache_resources,
 					  subreq->start, &subreq->len, i_size,
@@ -622,6 +623,77 @@ static int cachefiles_prepare_write(struct netfs_cache_resources *cres,
 	return ret;
 }
 
+static void cachefiles_prepare_write_subreq(struct netfs_io_subrequest *subreq)
+{
+	struct netfs_io_request *wreq = subreq->rreq;
+	struct netfs_cache_resources *cres = &wreq->cache_resources;
+
+	_enter("W=%x[%x] %llx", wreq->debug_id, subreq->debug_index, subreq->start);
+
+	subreq->max_len = ULONG_MAX;
+	subreq->max_nr_segs = BIO_MAX_VECS;
+
+	if (!cachefiles_cres_file(cres)) {
+		if (!fscache_wait_for_operation(cres, FSCACHE_WANT_WRITE))
+			return netfs_prepare_write_failed(subreq);
+		if (!cachefiles_cres_file(cres))
+			return netfs_prepare_write_failed(subreq);
+	}
+}
+
+static void cachefiles_issue_write(struct netfs_io_subrequest *subreq)
+{
+	struct netfs_io_request *wreq = subreq->rreq;
+	struct netfs_cache_resources *cres = &wreq->cache_resources;
+	struct cachefiles_object *object = cachefiles_cres_object(cres);
+	struct cachefiles_cache *cache = object->volume->cache;
+	const struct cred *saved_cred;
+	size_t off, pre, post, len = subreq->len;
+	loff_t start = subreq->start;
+	int ret;
+
+	_enter("W=%x[%x] %llx-%llx",
+	       wreq->debug_id, subreq->debug_index, start, start + len - 1);
+
+	/* We need to start on the cache granularity boundary */
+	off = start & (CACHEFILES_DIO_BLOCK_SIZE - 1);
+	if (off) {
+		pre = CACHEFILES_DIO_BLOCK_SIZE - off;
+		if (pre >= len) {
+			netfs_write_subrequest_terminated(subreq, len, false);
+			return;
+		}
+		subreq->transferred += pre;
+		start += pre;
+		len -= pre;
+		iov_iter_advance(&subreq->io_iter, pre);
+	}
+
+	/* We also need to end on the cache granularity boundary */
+	post = len & (CACHEFILES_DIO_BLOCK_SIZE - 1);
+	if (post) {
+		len -= post;
+		if (len == 0) {
+			netfs_write_subrequest_terminated(subreq, post, false);
+			return;
+		}
+		iov_iter_truncate(&subreq->io_iter, len);
+	}
+
+	cachefiles_begin_secure(cache, &saved_cred);
+	ret = __cachefiles_prepare_write(object, cachefiles_cres_file(cres),
+					 &start, &len, len, true);
+	cachefiles_end_secure(cache, saved_cred);
+	if (ret < 0) {
+		netfs_write_subrequest_terminated(subreq, ret, false);
+		return;
+	}
+
+	cachefiles_write(&subreq->rreq->cache_resources,
+			 subreq->start, &subreq->io_iter,
+			 netfs_write_subrequest_terminated, subreq);
+}
+
 /*
  * Clean up an operation.
  */
@@ -638,8 +710,10 @@ static const struct netfs_cache_ops cachefiles_netfs_cache_ops = {
 	.end_operation		= cachefiles_end_operation,
 	.read			= cachefiles_read,
 	.write			= cachefiles_write,
+	.issue_write		= cachefiles_issue_write,
 	.prepare_read		= cachefiles_prepare_read,
 	.prepare_write		= cachefiles_prepare_write,
+	.prepare_write_subreq	= cachefiles_prepare_write_subreq,
 	.prepare_ondemand_read	= cachefiles_prepare_ondemand_read,
 	.query_occupancy	= cachefiles_query_occupancy,
 };
diff --git a/fs/cachefiles/namei.c b/fs/cachefiles/namei.c
index 7ade836beb..f53977169d 100644
--- a/fs/cachefiles/namei.c
+++ b/fs/cachefiles/namei.c
@@ -563,8 +563,7 @@ static bool cachefiles_open_file(struct cachefiles_object *object,
 	 */
 	path.mnt = cache->mnt;
 	path.dentry = dentry;
-	file = kernel_file_open(&path, O_RDWR | O_LARGEFILE | O_DIRECT,
-				d_backing_inode(dentry), cache->cache_cred);
+	file = kernel_file_open(&path, O_RDWR | O_LARGEFILE | O_DIRECT, cache->cache_cred);
 	if (IS_ERR(file)) {
 		trace_cachefiles_vfs_error(object, d_backing_inode(dentry),
 					   PTR_ERR(file),
diff --git a/fs/cachefiles/ondemand.c b/fs/cachefiles/ondemand.c
index 4ba42f1fa3..470c966583 100644
--- a/fs/cachefiles/ondemand.c
+++ b/fs/cachefiles/ondemand.c
@@ -1,22 +1,42 @@
 // SPDX-License-Identifier: GPL-2.0-or-later
-#include <linux/fdtable.h>
 #include <linux/anon_inodes.h>
 #include <linux/uio.h>
 #include "internal.h"
 
+struct ondemand_anon_file {
+	struct file *file;
+	int fd;
+};
+
+static inline void cachefiles_req_put(struct cachefiles_req *req)
+{
+	if (refcount_dec_and_test(&req->ref))
+		kfree(req);
+}
+
 static int cachefiles_ondemand_fd_release(struct inode *inode,
 					  struct file *file)
 {
 	struct cachefiles_object *object = file->private_data;
-	struct cachefiles_cache *cache = object->volume->cache;
-	struct cachefiles_ondemand_info *info = object->ondemand;
-	int object_id = info->ondemand_id;
+	struct cachefiles_cache *cache;
+	struct cachefiles_ondemand_info *info;
+	int object_id;
 	struct cachefiles_req *req;
-	XA_STATE(xas, &cache->reqs, 0);
+	XA_STATE(xas, NULL, 0);
+
+	if (!object)
+		return 0;
+
+	info = object->ondemand;
+	cache = object->volume->cache;
+	xas.xa = &cache->reqs;
 
 	xa_lock(&cache->reqs);
+	spin_lock(&info->lock);
+	object_id = info->ondemand_id;
 	info->ondemand_id = CACHEFILES_ONDEMAND_ID_CLOSED;
 	cachefiles_ondemand_set_object_close(object);
+	spin_unlock(&info->lock);
 
 	/* Only flush CACHEFILES_REQ_NEW marked req to avoid race with daemon_read */
 	xas_for_each_marked(&xas, req, ULONG_MAX, CACHEFILES_REQ_NEW) {
@@ -76,12 +96,12 @@ static loff_t cachefiles_ondemand_fd_llseek(struct file *filp, loff_t pos,
 }
 
 static long cachefiles_ondemand_fd_ioctl(struct file *filp, unsigned int ioctl,
-					 unsigned long arg)
+					 unsigned long id)
 {
 	struct cachefiles_object *object = filp->private_data;
 	struct cachefiles_cache *cache = object->volume->cache;
 	struct cachefiles_req *req;
-	unsigned long id;
+	XA_STATE(xas, &cache->reqs, id);
 
 	if (ioctl != CACHEFILES_IOC_READ_COMPLETE)
 		return -EINVAL;
@@ -89,10 +109,15 @@ static long cachefiles_ondemand_fd_ioctl(struct file *filp, unsigned int ioctl,
 	if (!test_bit(CACHEFILES_ONDEMAND_MODE, &cache->flags))
 		return -EOPNOTSUPP;
 
-	id = arg;
-	req = xa_erase(&cache->reqs, id);
-	if (!req)
+	xa_lock(&cache->reqs);
+	req = xas_load(&xas);
+	if (!req || req->msg.opcode != CACHEFILES_OP_READ ||
+	    req->object != object) {
+		xa_unlock(&cache->reqs);
 		return -EINVAL;
+	}
+	xas_store(&xas, NULL);
+	xa_unlock(&cache->reqs);
 
 	trace_cachefiles_ondemand_cread(object, id);
 	complete(&req->done);
@@ -116,10 +141,12 @@ int cachefiles_ondemand_copen(struct cachefiles_cache *cache, char *args)
 {
 	struct cachefiles_req *req;
 	struct fscache_cookie *cookie;
+	struct cachefiles_ondemand_info *info;
 	char *pid, *psize;
 	unsigned long id;
 	long size;
 	int ret;
+	XA_STATE(xas, &cache->reqs, 0);
 
 	if (!test_bit(CACHEFILES_ONDEMAND_MODE, &cache->flags))
 		return -EOPNOTSUPP;
@@ -143,10 +170,18 @@ int cachefiles_ondemand_copen(struct cachefiles_cache *cache, char *args)
 	if (ret)
 		return ret;
 
-	req = xa_erase(&cache->reqs, id);
-	if (!req)
+	xa_lock(&cache->reqs);
+	xas.xa_index = id;
+	req = xas_load(&xas);
+	if (!req || req->msg.opcode != CACHEFILES_OP_OPEN ||
+	    !req->object->ondemand->ondemand_id) {
+		xa_unlock(&cache->reqs);
 		return -EINVAL;
+	}
+	xas_store(&xas, NULL);
+	xa_unlock(&cache->reqs);
 
+	info = req->object->ondemand;
 	/* fail OPEN request if copen format is invalid */
 	ret = kstrtol(psize, 0, &size);
 	if (ret) {
@@ -166,6 +201,32 @@ int cachefiles_ondemand_copen(struct cachefiles_cache *cache, char *args)
 		goto out;
 	}
 
+	spin_lock(&info->lock);
+	/*
+	 * The anonymous fd was closed before copen ? Fail the request.
+	 *
+	 *             t1             |             t2
+	 * ---------------------------------------------------------
+	 *                             cachefiles_ondemand_copen
+	 *                             req = xa_erase(&cache->reqs, id)
+	 * // Anon fd is maliciously closed.
+	 * cachefiles_ondemand_fd_release
+	 * xa_lock(&cache->reqs)
+	 * cachefiles_ondemand_set_object_close(object)
+	 * xa_unlock(&cache->reqs)
+	 *                             cachefiles_ondemand_set_object_open
+	 *                             // No one will ever close it again.
+	 * cachefiles_ondemand_daemon_read
+	 * cachefiles_ondemand_select_req
+	 *
+	 * Get a read req but its fd is already closed. The daemon can't
+	 * issue a cread ioctl with an closed fd, then hung.
+	 */
+	if (info->ondemand_id == CACHEFILES_ONDEMAND_ID_CLOSED) {
+		spin_unlock(&info->lock);
+		req->error = -EBADFD;
+		goto out;
+	}
 	cookie = req->object->cookie;
 	cookie->object_size = size;
 	if (size)
@@ -175,9 +236,15 @@ int cachefiles_ondemand_copen(struct cachefiles_cache *cache, char *args)
 	trace_cachefiles_ondemand_copen(req->object, id, size);
 
 	cachefiles_ondemand_set_object_open(req->object);
+	spin_unlock(&info->lock);
 	wake_up_all(&cache->daemon_pollwq);
 
 out:
+	spin_lock(&info->lock);
+	/* Need to set object close to avoid reopen status continuing */
+	if (info->ondemand_id == CACHEFILES_ONDEMAND_ID_CLOSED)
+		cachefiles_ondemand_set_object_close(req->object);
+	spin_unlock(&info->lock);
 	complete(&req->done);
 	return ret;
 }
@@ -205,14 +272,14 @@ int cachefiles_ondemand_restore(struct cachefiles_cache *cache, char *args)
 	return 0;
 }
 
-static int cachefiles_ondemand_get_fd(struct cachefiles_req *req)
+static int cachefiles_ondemand_get_fd(struct cachefiles_req *req,
+				      struct ondemand_anon_file *anon_file)
 {
 	struct cachefiles_object *object;
 	struct cachefiles_cache *cache;
 	struct cachefiles_open *load;
-	struct file *file;
 	u32 object_id;
-	int ret, fd;
+	int ret;
 
 	object = cachefiles_grab_object(req->object,
 			cachefiles_obj_get_ondemand_fd);
@@ -224,35 +291,53 @@ static int cachefiles_ondemand_get_fd(struct cachefiles_req *req)
 	if (ret < 0)
 		goto err;
 
-	fd = get_unused_fd_flags(O_WRONLY);
-	if (fd < 0) {
-		ret = fd;
+	anon_file->fd = get_unused_fd_flags(O_WRONLY);
+	if (anon_file->fd < 0) {
+		ret = anon_file->fd;
 		goto err_free_id;
 	}
 
-	file = anon_inode_getfile("[cachefiles]", &cachefiles_ondemand_fd_fops,
-				  object, O_WRONLY);
-	if (IS_ERR(file)) {
-		ret = PTR_ERR(file);
+	anon_file->file = anon_inode_getfile("[cachefiles]",
+				&cachefiles_ondemand_fd_fops, object, O_WRONLY);
+	if (IS_ERR(anon_file->file)) {
+		ret = PTR_ERR(anon_file->file);
 		goto err_put_fd;
 	}
 
-	file->f_mode |= FMODE_PWRITE | FMODE_LSEEK;
-	fd_install(fd, file);
+	spin_lock(&object->ondemand->lock);
+	if (object->ondemand->ondemand_id > 0) {
+		spin_unlock(&object->ondemand->lock);
+		/* Pair with check in cachefiles_ondemand_fd_release(). */
+		anon_file->file->private_data = NULL;
+		ret = -EEXIST;
+		goto err_put_file;
+	}
+
+	anon_file->file->f_mode |= FMODE_PWRITE | FMODE_LSEEK;
 
 	load = (void *)req->msg.data;
-	load->fd = fd;
+	load->fd = anon_file->fd;
 	object->ondemand->ondemand_id = object_id;
+	spin_unlock(&object->ondemand->lock);
 
 	cachefiles_get_unbind_pincount(cache);
 	trace_cachefiles_ondemand_open(object, &req->msg, load);
 	return 0;
 
+err_put_file:
+	fput(anon_file->file);
+	anon_file->file = NULL;
 err_put_fd:
-	put_unused_fd(fd);
+	put_unused_fd(anon_file->fd);
+	anon_file->fd = ret;
 err_free_id:
 	xa_erase(&cache->ondemand_ids, object_id);
 err:
+	spin_lock(&object->ondemand->lock);
+	/* Avoid marking an opened object as closed. */
+	if (object->ondemand->ondemand_id <= 0)
+		cachefiles_ondemand_set_object_close(object);
+	spin_unlock(&object->ondemand->lock);
 	cachefiles_put_object(object, cachefiles_obj_put_ondemand_fd);
 	return ret;
 }
@@ -294,14 +379,28 @@ static struct cachefiles_req *cachefiles_ondemand_select_req(struct xa_state *xa
 	return NULL;
 }
 
+static inline bool cachefiles_ondemand_finish_req(struct cachefiles_req *req,
+						  struct xa_state *xas, int err)
+{
+	if (unlikely(!xas || !req))
+		return false;
+
+	if (xa_cmpxchg(xas->xa, xas->xa_index, req, NULL, 0) != req)
+		return false;
+
+	req->error = err;
+	complete(&req->done);
+	return true;
+}
+
 ssize_t cachefiles_ondemand_daemon_read(struct cachefiles_cache *cache,
 					char __user *_buffer, size_t buflen)
 {
 	struct cachefiles_req *req;
 	struct cachefiles_msg *msg;
-	unsigned long id = 0;
 	size_t n;
 	int ret = 0;
+	struct ondemand_anon_file anon_file;
 	XA_STATE(xas, &cache->reqs, cache->req_id_next);
 
 	xa_lock(&cache->reqs);
@@ -330,42 +429,37 @@ ssize_t cachefiles_ondemand_daemon_read(struct cachefiles_cache *cache,
 
 	xas_clear_mark(&xas, CACHEFILES_REQ_NEW);
 	cache->req_id_next = xas.xa_index + 1;
+	refcount_inc(&req->ref);
+	cachefiles_grab_object(req->object, cachefiles_obj_get_read_req);
 	xa_unlock(&cache->reqs);
 
-	id = xas.xa_index;
-
 	if (msg->opcode == CACHEFILES_OP_OPEN) {
-		ret = cachefiles_ondemand_get_fd(req);
-		if (ret) {
-			cachefiles_ondemand_set_object_close(req->object);
-			goto error;
-		}
+		ret = cachefiles_ondemand_get_fd(req, &anon_file);
+		if (ret)
+			goto out;
 	}
 
-	msg->msg_id = id;
+	msg->msg_id = xas.xa_index;
 	msg->object_id = req->object->ondemand->ondemand_id;
 
-	if (copy_to_user(_buffer, msg, n) != 0) {
+	if (copy_to_user(_buffer, msg, n) != 0)
 		ret = -EFAULT;
-		goto err_put_fd;
-	}
 
-	/* CLOSE request has no reply */
-	if (msg->opcode == CACHEFILES_OP_CLOSE) {
-		xa_erase(&cache->reqs, id);
-		complete(&req->done);
+	if (msg->opcode == CACHEFILES_OP_OPEN) {
+		if (ret < 0) {
+			fput(anon_file.file);
+			put_unused_fd(anon_file.fd);
+			goto out;
+		}
+		fd_install(anon_file.fd, anon_file.file);
 	}
-
-	return n;
-
-err_put_fd:
-	if (msg->opcode == CACHEFILES_OP_OPEN)
-		close_fd(((struct cachefiles_open *)msg->data)->fd);
-error:
-	xa_erase(&cache->reqs, id);
-	req->error = ret;
-	complete(&req->done);
-	return ret;
+out:
+	cachefiles_put_object(req->object, cachefiles_obj_put_read_req);
+	/* Remove error request and CLOSE request has no reply */
+	if (ret || msg->opcode == CACHEFILES_OP_CLOSE)
+		cachefiles_ondemand_finish_req(req, &xas, ret);
+	cachefiles_req_put(req);
+	return ret ? ret : n;
 }
 
 typedef int (*init_req_fn)(struct cachefiles_req *req, void *private);
@@ -395,6 +489,7 @@ static int cachefiles_ondemand_send_req(struct cachefiles_object *object,
 		goto out;
 	}
 
+	refcount_set(&req->ref, 1);
 	req->object = object;
 	init_completion(&req->done);
 	req->msg.opcode = opcode;
@@ -422,7 +517,8 @@ static int cachefiles_ondemand_send_req(struct cachefiles_object *object,
 		 */
 		xas_lock(&xas);
 
-		if (test_bit(CACHEFILES_DEAD, &cache->flags)) {
+		if (test_bit(CACHEFILES_DEAD, &cache->flags) ||
+		    cachefiles_ondemand_object_is_dropping(object)) {
 			xas_unlock(&xas);
 			ret = -EIO;
 			goto out;
@@ -432,20 +528,32 @@ static int cachefiles_ondemand_send_req(struct cachefiles_object *object,
 		smp_mb();
 
 		if (opcode == CACHEFILES_OP_CLOSE &&
-			!cachefiles_ondemand_object_is_open(object)) {
+		    !cachefiles_ondemand_object_is_open(object)) {
 			WARN_ON_ONCE(object->ondemand->ondemand_id == 0);
 			xas_unlock(&xas);
 			ret = -EIO;
 			goto out;
 		}
 
-		xas.xa_index = 0;
+		/*
+		 * Cyclically find a free xas to avoid msg_id reuse that would
+		 * cause the daemon to successfully copen a stale msg_id.
+		 */
+		xas.xa_index = cache->msg_id_next;
 		xas_find_marked(&xas, UINT_MAX, XA_FREE_MARK);
+		if (xas.xa_node == XAS_RESTART) {
+			xas.xa_index = 0;
+			xas_find_marked(&xas, cache->msg_id_next - 1, XA_FREE_MARK);
+		}
 		if (xas.xa_node == XAS_RESTART)
 			xas_set_err(&xas, -EBUSY);
+
 		xas_store(&xas, req);
-		xas_clear_mark(&xas, XA_FREE_MARK);
-		xas_set_mark(&xas, CACHEFILES_REQ_NEW);
+		if (xas_valid(&xas)) {
+			cache->msg_id_next = xas.xa_index + 1;
+			xas_clear_mark(&xas, XA_FREE_MARK);
+			xas_set_mark(&xas, CACHEFILES_REQ_NEW);
+		}
 		xas_unlock(&xas);
 	} while (xas_nomem(&xas, GFP_KERNEL));
 
@@ -454,16 +562,27 @@ static int cachefiles_ondemand_send_req(struct cachefiles_object *object,
 		goto out;
 
 	wake_up_all(&cache->daemon_pollwq);
-	wait_for_completion(&req->done);
-	ret = req->error;
-	kfree(req);
+wait:
+	ret = wait_for_completion_killable(&req->done);
+	if (!ret) {
+		ret = req->error;
+	} else {
+		ret = -EINTR;
+		if (!cachefiles_ondemand_finish_req(req, &xas, ret)) {
+			/* Someone will complete it soon. */
+			cpu_relax();
+			goto wait;
+		}
+	}
+	cachefiles_req_put(req);
 	return ret;
 out:
 	/* Reset the object to close state in error handling path.
 	 * If error occurs after creating the anonymous fd,
 	 * cachefiles_ondemand_fd_release() will set object to close.
 	 */
-	if (opcode == CACHEFILES_OP_OPEN)
+	if (opcode == CACHEFILES_OP_OPEN &&
+	    !cachefiles_ondemand_object_is_dropping(object))
 		cachefiles_ondemand_set_object_close(object);
 	kfree(req);
 	return ret;
@@ -562,8 +681,34 @@ int cachefiles_ondemand_init_object(struct cachefiles_object *object)
 
 void cachefiles_ondemand_clean_object(struct cachefiles_object *object)
 {
+	unsigned long index;
+	struct cachefiles_req *req;
+	struct cachefiles_cache *cache;
+
+	if (!object->ondemand)
+		return;
+
 	cachefiles_ondemand_send_req(object, CACHEFILES_OP_CLOSE, 0,
 			cachefiles_ondemand_init_close_req, NULL);
+
+	if (!object->ondemand->ondemand_id)
+		return;
+
+	/* Cancel all requests for the object that is being dropped. */
+	cache = object->volume->cache;
+	xa_lock(&cache->reqs);
+	cachefiles_ondemand_set_object_dropping(object);
+	xa_for_each(&cache->reqs, index, req) {
+		if (req->object == object) {
+			req->error = -EIO;
+			complete(&req->done);
+			__xa_erase(&cache->reqs, index);
+		}
+	}
+	xa_unlock(&cache->reqs);
+
+	/* Wait for ondemand_object_worker() to finish to avoid UAF. */
+	cancel_work_sync(&object->ondemand->ondemand_work);
 }
 
 int cachefiles_ondemand_init_obj_info(struct cachefiles_object *object,
@@ -578,6 +723,7 @@ int cachefiles_ondemand_init_obj_info(struct cachefiles_object *object,
 		return -ENOMEM;
 
 	object->ondemand->object = object;
+	spin_lock_init(&object->ondemand->lock);
 	INIT_WORK(&object->ondemand->ondemand_work, ondemand_object_worker);
 	return 0;
 }
diff --git a/fs/cachefiles/volume.c b/fs/cachefiles/volume.c
index 89df0ba8ba..781aac4ef2 100644
--- a/fs/cachefiles/volume.c
+++ b/fs/cachefiles/volume.c
@@ -133,7 +133,6 @@ void cachefiles_free_volume(struct fscache_volume *vcookie)
 
 void cachefiles_withdraw_volume(struct cachefiles_volume *volume)
 {
-	fscache_withdraw_volume(volume->vcookie);
 	cachefiles_set_volume_xattr(volume);
 	__cachefiles_free_volume(volume);
 }
diff --git a/fs/cachefiles/xattr.c b/fs/cachefiles/xattr.c
index bcb6173943..4dd8a993c6 100644
--- a/fs/cachefiles/xattr.c
+++ b/fs/cachefiles/xattr.c
@@ -110,9 +110,11 @@ int cachefiles_check_auxdata(struct cachefiles_object *object, struct file *file
 	if (xlen == 0)
 		xlen = vfs_getxattr(&nop_mnt_idmap, dentry, cachefiles_xattr_cache, buf, tlen);
 	if (xlen != tlen) {
-		if (xlen < 0)
+		if (xlen < 0) {
+			ret = xlen;
 			trace_cachefiles_vfs_error(object, file_inode(file), xlen,
 						   cachefiles_trace_getxattr_error);
+		}
 		if (xlen == -EIO)
 			cachefiles_io_error_obj(
 				object,
@@ -252,6 +254,7 @@ int cachefiles_check_volume_xattr(struct cachefiles_volume *volume)
 		xlen = vfs_getxattr(&nop_mnt_idmap, dentry, cachefiles_xattr_cache, buf, len);
 	if (xlen != len) {
 		if (xlen < 0) {
+			ret = xlen;
 			trace_cachefiles_vfs_error(NULL, d_inode(dentry), xlen,
 						   cachefiles_trace_getxattr_error);
 			if (xlen == -EIO)
diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c
index ee9caf7916..8c16bc5250 100644
--- a/fs/ceph/addr.c
+++ b/fs/ceph/addr.c
@@ -193,7 +193,7 @@ static void ceph_netfs_expand_readahead(struct netfs_io_request *rreq)
 	 * block, but do not exceed the file size, unless the original
 	 * request already exceeds it.
 	 */
-	new_end = min(round_up(end, lo->stripe_unit), rreq->i_size);
+	new_end = umin(round_up(end, lo->stripe_unit), rreq->i_size);
 	if (new_end > end && new_end <= rreq->start + max_len)
 		rreq->len = new_end - rreq->start;
 
@@ -498,11 +498,6 @@ const struct netfs_request_ops ceph_netfs_ops = {
 };
 
 #ifdef CONFIG_CEPH_FSCACHE
-static void ceph_set_page_fscache(struct page *page)
-{
-	set_page_fscache(page);
-}
-
 static void ceph_fscache_write_terminated(void *priv, ssize_t error, bool was_async)
 {
 	struct inode *inode = priv;
@@ -517,13 +512,9 @@ static void ceph_fscache_write_to_cache(struct inode *inode, u64 off, u64 len, b
 	struct fscache_cookie *cookie = ceph_fscache_cookie(ci);
 
 	fscache_write_to_cache(cookie, inode->i_mapping, off, len, i_size_read(inode),
-			       ceph_fscache_write_terminated, inode, caching);
+			       ceph_fscache_write_terminated, inode, true, caching);
 }
 #else
-static inline void ceph_set_page_fscache(struct page *page)
-{
-}
-
 static inline void ceph_fscache_write_to_cache(struct inode *inode, u64 off, u64 len, bool caching)
 {
 }
@@ -715,8 +706,6 @@ static int writepage_nounlock(struct page *page, struct writeback_control *wbc)
 		len = wlen;
 
 	set_page_writeback(page);
-	if (caching)
-		ceph_set_page_fscache(page);
 	ceph_fscache_write_to_cache(inode, page_off, len, caching);
 
 	if (IS_ENCRYPTED(inode)) {
@@ -800,8 +789,6 @@ static int ceph_writepage(struct page *page, struct writeback_control *wbc)
 		return AOP_WRITEPAGE_ACTIVATE;
 	}
 
-	wait_on_page_fscache(page);
-
 	err = writepage_nounlock(page, wbc);
 	if (err == -ERESTARTSYS) {
 		/* direct memory reclaimer was killed by SIGKILL. return 0
@@ -1075,7 +1062,7 @@ get_more_pages:
 				unlock_page(page);
 				break;
 			}
-			if (PageWriteback(page) || PageFsCache(page)) {
+			if (PageWriteback(page)) {
 				if (wbc->sync_mode == WB_SYNC_NONE) {
 					doutc(cl, "%p under writeback\n", page);
 					unlock_page(page);
@@ -1083,7 +1070,6 @@ get_more_pages:
 				}
 				doutc(cl, "waiting on writeback %p\n", page);
 				wait_on_page_writeback(page);
-				wait_on_page_fscache(page);
 			}
 
 			if (!clear_page_dirty_for_io(page)) {
@@ -1268,8 +1254,6 @@ new_request:
 			}
 
 			set_page_writeback(page);
-			if (caching)
-				ceph_set_page_fscache(page);
 			len += thp_size(page);
 		}
 		ceph_fscache_write_to_cache(inode, offset, len, caching);
@@ -1513,7 +1497,7 @@ static int ceph_write_begin(struct file *file, struct address_space *mapping,
 	if (r < 0)
 		return r;
 
-	folio_wait_fscache(folio);
+	folio_wait_private_2(folio); /* [DEPRECATED] */
 	WARN_ON_ONCE(!folio_test_locked(folio));
 	*pagep = &folio->page;
 	return 0;
diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index e8d8e3e633..68df96800e 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -2016,6 +2016,8 @@ bool __ceph_should_report_size(struct ceph_inode_info *ci)
  *  CHECK_CAPS_AUTHONLY - we should only check the auth cap
  *  CHECK_CAPS_FLUSH - we should flush any dirty caps immediately, without
  *    further delay.
+ *  CHECK_CAPS_FLUSH_FORCE - we should flush any caps immediately, without
+ *    further delay.
  */
 void ceph_check_caps(struct ceph_inode_info *ci, int flags)
 {
@@ -2097,7 +2099,7 @@ retry:
 	}
 
 	doutc(cl, "%p %llx.%llx file_want %s used %s dirty %s "
-	      "flushing %s issued %s revoking %s retain %s %s%s%s\n",
+	      "flushing %s issued %s revoking %s retain %s %s%s%s%s\n",
 	     inode, ceph_vinop(inode), ceph_cap_string(file_wanted),
 	     ceph_cap_string(used), ceph_cap_string(ci->i_dirty_caps),
 	     ceph_cap_string(ci->i_flushing_caps),
@@ -2105,7 +2107,8 @@ retry:
 	     ceph_cap_string(retain),
 	     (flags & CHECK_CAPS_AUTHONLY) ? " AUTHONLY" : "",
 	     (flags & CHECK_CAPS_FLUSH) ? " FLUSH" : "",
-	     (flags & CHECK_CAPS_NOINVAL) ? " NOINVAL" : "");
+	     (flags & CHECK_CAPS_NOINVAL) ? " NOINVAL" : "",
+	     (flags & CHECK_CAPS_FLUSH_FORCE) ? " FLUSH_FORCE" : "");
 
 	/*
 	 * If we no longer need to hold onto old our caps, and we may
@@ -2180,6 +2183,11 @@ retry:
 				queue_writeback = true;
 		}
 
+		if (flags & CHECK_CAPS_FLUSH_FORCE) {
+			doutc(cl, "force to flush caps\n");
+			goto ack;
+		}
+
 		if (cap == ci->i_auth_cap &&
 		    (cap->issued & CEPH_CAP_FILE_WR)) {
 			/* request larger max_size from MDS? */
@@ -3504,6 +3512,8 @@ static void handle_cap_grant(struct inode *inode,
 	bool queue_invalidate = false;
 	bool deleted_inode = false;
 	bool fill_inline = false;
+	bool revoke_wait = false;
+	int flags = 0;
 
 	/*
 	 * If there is at least one crypto block then we'll trust
@@ -3699,16 +3709,18 @@ static void handle_cap_grant(struct inode *inode,
 		      ceph_cap_string(cap->issued), ceph_cap_string(newcaps),
 		      ceph_cap_string(revoking));
 		if (S_ISREG(inode->i_mode) &&
-		    (revoking & used & CEPH_CAP_FILE_BUFFER))
+		    (revoking & used & CEPH_CAP_FILE_BUFFER)) {
 			writeback = true;  /* initiate writeback; will delay ack */
-		else if (queue_invalidate &&
+			revoke_wait = true;
+		} else if (queue_invalidate &&
 			 revoking == CEPH_CAP_FILE_CACHE &&
-			 (newcaps & CEPH_CAP_FILE_LAZYIO) == 0)
-			; /* do nothing yet, invalidation will be queued */
-		else if (cap == ci->i_auth_cap)
+			 (newcaps & CEPH_CAP_FILE_LAZYIO) == 0) {
+			revoke_wait = true; /* do nothing yet, invalidation will be queued */
+		} else if (cap == ci->i_auth_cap) {
 			check_caps = 1; /* check auth cap only */
-		else
+		} else {
 			check_caps = 2; /* check all caps */
+		}
 		/* If there is new caps, try to wake up the waiters */
 		if (~cap->issued & newcaps)
 			wake = true;
@@ -3735,8 +3747,9 @@ static void handle_cap_grant(struct inode *inode,
 	BUG_ON(cap->issued & ~cap->implemented);
 
 	/* don't let check_caps skip sending a response to MDS for revoke msgs */
-	if (le32_to_cpu(grant->op) == CEPH_CAP_OP_REVOKE) {
+	if (!revoke_wait && le32_to_cpu(grant->op) == CEPH_CAP_OP_REVOKE) {
 		cap->mds_wanted = 0;
+		flags |= CHECK_CAPS_FLUSH_FORCE;
 		if (cap == ci->i_auth_cap)
 			check_caps = 1; /* check auth cap only */
 		else
@@ -3792,9 +3805,9 @@ static void handle_cap_grant(struct inode *inode,
 
 	mutex_unlock(&session->s_mutex);
 	if (check_caps == 1)
-		ceph_check_caps(ci, CHECK_CAPS_AUTHONLY | CHECK_CAPS_NOINVAL);
+		ceph_check_caps(ci, flags | CHECK_CAPS_AUTHONLY | CHECK_CAPS_NOINVAL);
 	else if (check_caps == 2)
-		ceph_check_caps(ci, CHECK_CAPS_NOINVAL);
+		ceph_check_caps(ci, flags | CHECK_CAPS_NOINVAL);
 }
 
 /*
@@ -4634,6 +4647,14 @@ unsigned long ceph_check_delayed_caps(struct ceph_mds_client *mdsc)
 			iput(inode);
 			spin_lock(&mdsc->cap_delay_lock);
 		}
+
+		/*
+		 * Make sure too many dirty caps or general
+		 * slowness doesn't block mdsc delayed work,
+		 * preventing send_renew_caps() from running.
+		 */
+		if (jiffies - loop_start >= 5 * HZ)
+			break;
 	}
 	spin_unlock(&mdsc->cap_delay_lock);
 	doutc(cl, "done\n");
diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c
index 0e9f56eaba..82a2e2a06a 100644
--- a/fs/ceph/dir.c
+++ b/fs/ceph/dir.c
@@ -1336,8 +1336,12 @@ static int ceph_unlink(struct inode *dir, struct dentry *dentry)
 	struct inode *inode = d_inode(dentry);
 	struct ceph_mds_request *req;
 	bool try_async = ceph_test_mount_opt(fsc, ASYNC_DIROPS);
+	struct dentry *dn;
 	int err = -EROFS;
 	int op;
+	char *path;
+	int pathlen;
+	u64 pathbase;
 
 	if (ceph_snap(dir) == CEPH_SNAPDIR) {
 		/* rmdir .snap/foo is RMSNAP */
@@ -1351,6 +1355,30 @@ static int ceph_unlink(struct inode *dir, struct dentry *dentry)
 			CEPH_MDS_OP_RMDIR : CEPH_MDS_OP_UNLINK;
 	} else
 		goto out;
+
+	dn = d_find_alias(dir);
+	if (!dn) {
+		try_async = false;
+	} else {
+		path = ceph_mdsc_build_path(mdsc, dn, &pathlen, &pathbase, 0);
+		if (IS_ERR(path)) {
+			try_async = false;
+			err = 0;
+		} else {
+			err = ceph_mds_check_access(mdsc, path, MAY_WRITE);
+		}
+		ceph_mdsc_free_path(path, pathlen);
+		dput(dn);
+
+		/* For none EACCES cases will let the MDS do the mds auth check */
+		if (err == -EACCES) {
+			return err;
+		} else if (err < 0) {
+			try_async = false;
+			err = 0;
+		}
+	}
+
 retry:
 	req = ceph_mdsc_create_request(mdsc, op, USE_AUTH_MDS);
 	if (IS_ERR(req)) {
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index 3d1cd079db..4b8d59ebda 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -366,6 +366,12 @@ int ceph_open(struct inode *inode, struct file *file)
 	struct ceph_file_info *fi = file->private_data;
 	int err;
 	int flags, fmode, wanted;
+	struct dentry *dentry;
+	char *path;
+	int pathlen;
+	u64 pathbase;
+	bool do_sync = false;
+	int mask = MAY_READ;
 
 	if (fi) {
 		doutc(cl, "file %p is already opened\n", file);
@@ -387,6 +393,31 @@ int ceph_open(struct inode *inode, struct file *file)
 	fmode = ceph_flags_to_mode(flags);
 	wanted = ceph_caps_for_mode(fmode);
 
+	if (fmode & CEPH_FILE_MODE_WR)
+		mask |= MAY_WRITE;
+	dentry = d_find_alias(inode);
+	if (!dentry) {
+		do_sync = true;
+	} else {
+		path = ceph_mdsc_build_path(mdsc, dentry, &pathlen, &pathbase, 0);
+		if (IS_ERR(path)) {
+			do_sync = true;
+			err = 0;
+		} else {
+			err = ceph_mds_check_access(mdsc, path, mask);
+		}
+		ceph_mdsc_free_path(path, pathlen);
+		dput(dentry);
+
+		/* For none EACCES cases will let the MDS do the mds auth check */
+		if (err == -EACCES) {
+			return err;
+		} else if (err < 0) {
+			do_sync = true;
+			err = 0;
+		}
+	}
+
 	/* snapped files are read-only */
 	if (ceph_snap(inode) != CEPH_NOSNAP && (file->f_mode & FMODE_WRITE))
 		return -EROFS;
@@ -402,7 +433,7 @@ int ceph_open(struct inode *inode, struct file *file)
 	 * asynchronously.
 	 */
 	spin_lock(&ci->i_ceph_lock);
-	if (__ceph_is_any_real_caps(ci) &&
+	if (!do_sync && __ceph_is_any_real_caps(ci) &&
 	    (((fmode & CEPH_FILE_MODE_WR) == 0) || ci->i_auth_cap)) {
 		int mds_wanted = __ceph_caps_mds_wanted(ci, true);
 		int issued = __ceph_caps_issued(ci, NULL);
@@ -420,7 +451,7 @@ int ceph_open(struct inode *inode, struct file *file)
 			ceph_check_caps(ci, 0);
 
 		return ceph_init_file(inode, file, fmode);
-	} else if (ceph_snap(inode) != CEPH_NOSNAP &&
+	} else if (!do_sync && ceph_snap(inode) != CEPH_NOSNAP &&
 		   (ci->i_snap_caps & wanted) == wanted) {
 		__ceph_touch_fmode(ci, mdsc, fmode);
 		spin_unlock(&ci->i_ceph_lock);
@@ -759,6 +790,9 @@ int ceph_atomic_open(struct inode *dir, struct dentry *dentry,
 	bool try_async = ceph_test_mount_opt(fsc, ASYNC_DIROPS);
 	int mask;
 	int err;
+	char *path;
+	int pathlen;
+	u64 pathbase;
 
 	doutc(cl, "%p %llx.%llx dentry %p '%pd' %s flags %d mode 0%o\n",
 	      dir, ceph_vinop(dir), dentry, dentry,
@@ -776,6 +810,34 @@ int ceph_atomic_open(struct inode *dir, struct dentry *dentry,
 	 */
 	flags &= ~O_TRUNC;
 
+	dn = d_find_alias(dir);
+	if (!dn) {
+		try_async = false;
+	} else {
+		path = ceph_mdsc_build_path(mdsc, dn, &pathlen, &pathbase, 0);
+		if (IS_ERR(path)) {
+			try_async = false;
+			err = 0;
+		} else {
+			int fmode = ceph_flags_to_mode(flags);
+
+			mask = MAY_READ;
+			if (fmode & CEPH_FILE_MODE_WR)
+				mask |= MAY_WRITE;
+			err = ceph_mds_check_access(mdsc, path, mask);
+		}
+		ceph_mdsc_free_path(path, pathlen);
+		dput(dn);
+
+		/* For none EACCES cases will let the MDS do the mds auth check */
+		if (err == -EACCES) {
+			return err;
+		} else if (err < 0) {
+			try_async = false;
+			err = 0;
+		}
+	}
+
 retry:
 	if (flags & O_CREAT) {
 		if (ceph_quota_is_max_files_exceeded(dir))
@@ -2129,14 +2191,16 @@ again:
 		int statret;
 		struct page *page = NULL;
 		loff_t i_size;
+		int mask = CEPH_STAT_CAP_SIZE;
 		if (retry_op == READ_INLINE) {
 			page = __page_cache_alloc(GFP_KERNEL);
 			if (!page)
 				return -ENOMEM;
+
+			mask = CEPH_STAT_CAP_INLINE_DATA;
 		}
 
-		statret = __ceph_do_getattr(inode, page,
-					    CEPH_STAT_CAP_INLINE_DATA, !!page);
+		statret = __ceph_do_getattr(inode, page, mask, !!page);
 		if (statret < 0) {
 			if (page)
 				__free_page(page);
@@ -2177,7 +2241,7 @@ again:
 		/* hit EOF or hole? */
 		if (retry_op == CHECK_EOF && iocb->ki_pos < i_size &&
 		    ret < len) {
-			doutc(cl, "hit hole, ppos %lld < size %lld, reading more\n",
+			doutc(cl, "may hit hole, ppos %lld < size %lld, reading more\n",
 			      iocb->ki_pos, i_size);
 
 			read += ret;
diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index 7b2e77517f..249ddfbb1b 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -577,6 +577,8 @@ struct inode *ceph_alloc_inode(struct super_block *sb)
 
 	/* Set parameters for the netfs library */
 	netfs_inode_init(&ci->netfs, &ceph_netfs_ops, false);
+	/* [DEPRECATED] Use PG_private_2 to mark folio being written to the cache. */
+	__set_bit(NETFS_ICTX_USE_PGPRIV2, &ci->netfs.flags);
 
 	spin_lock_init(&ci->i_ceph_lock);
 
@@ -2480,6 +2482,34 @@ int __ceph_setattr(struct mnt_idmap *idmap, struct inode *inode,
 	bool lock_snap_rwsem = false;
 	bool fill_fscrypt;
 	int truncate_retry = 20; /* The RMW will take around 50ms */
+	struct dentry *dentry;
+	char *path;
+	int pathlen;
+	u64 pathbase;
+	bool do_sync = false;
+
+	dentry = d_find_alias(inode);
+	if (!dentry) {
+		do_sync = true;
+	} else {
+		path = ceph_mdsc_build_path(mdsc, dentry, &pathlen, &pathbase, 0);
+		if (IS_ERR(path)) {
+			do_sync = true;
+			err = 0;
+		} else {
+			err = ceph_mds_check_access(mdsc, path, MAY_WRITE);
+		}
+		ceph_mdsc_free_path(path, pathlen);
+		dput(dentry);
+
+		/* For none EACCES cases will let the MDS do the mds auth check */
+		if (err == -EACCES) {
+			return err;
+		} else if (err < 0) {
+			do_sync = true;
+			err = 0;
+		}
+	}
 
 retry:
 	prealloc_cf = ceph_alloc_cap_flush();
@@ -2526,7 +2556,7 @@ retry:
 		/* It should never be re-set once set */
 		WARN_ON_ONCE(ci->fscrypt_auth);
 
-		if (issued & CEPH_CAP_AUTH_EXCL) {
+		if (!do_sync && (issued & CEPH_CAP_AUTH_EXCL)) {
 			dirtied |= CEPH_CAP_AUTH_EXCL;
 			kfree(ci->fscrypt_auth);
 			ci->fscrypt_auth = (u8 *)cia->fscrypt_auth;
@@ -2555,7 +2585,7 @@ retry:
 		      ceph_vinop(inode),
 		      from_kuid(&init_user_ns, inode->i_uid),
 		      from_kuid(&init_user_ns, attr->ia_uid));
-		if (issued & CEPH_CAP_AUTH_EXCL) {
+		if (!do_sync && (issued & CEPH_CAP_AUTH_EXCL)) {
 			inode->i_uid = fsuid;
 			dirtied |= CEPH_CAP_AUTH_EXCL;
 		} else if ((issued & CEPH_CAP_AUTH_SHARED) == 0 ||
@@ -2573,7 +2603,7 @@ retry:
 		      ceph_vinop(inode),
 		      from_kgid(&init_user_ns, inode->i_gid),
 		      from_kgid(&init_user_ns, attr->ia_gid));
-		if (issued & CEPH_CAP_AUTH_EXCL) {
+		if (!do_sync && (issued & CEPH_CAP_AUTH_EXCL)) {
 			inode->i_gid = fsgid;
 			dirtied |= CEPH_CAP_AUTH_EXCL;
 		} else if ((issued & CEPH_CAP_AUTH_SHARED) == 0 ||
@@ -2587,7 +2617,7 @@ retry:
 	if (ia_valid & ATTR_MODE) {
 		doutc(cl, "%p %llx.%llx mode 0%o -> 0%o\n", inode,
 		      ceph_vinop(inode), inode->i_mode, attr->ia_mode);
-		if (issued & CEPH_CAP_AUTH_EXCL) {
+		if (!do_sync && (issued & CEPH_CAP_AUTH_EXCL)) {
 			inode->i_mode = attr->ia_mode;
 			dirtied |= CEPH_CAP_AUTH_EXCL;
 		} else if ((issued & CEPH_CAP_AUTH_SHARED) == 0 ||
@@ -2606,11 +2636,11 @@ retry:
 		      inode, ceph_vinop(inode),
 		      atime.tv_sec, atime.tv_nsec,
 		      attr->ia_atime.tv_sec, attr->ia_atime.tv_nsec);
-		if (issued & CEPH_CAP_FILE_EXCL) {
+		if (!do_sync && (issued & CEPH_CAP_FILE_EXCL)) {
 			ci->i_time_warp_seq++;
 			inode_set_atime_to_ts(inode, attr->ia_atime);
 			dirtied |= CEPH_CAP_FILE_EXCL;
-		} else if ((issued & CEPH_CAP_FILE_WR) &&
+		} else if (!do_sync && (issued & CEPH_CAP_FILE_WR) &&
 			   timespec64_compare(&atime,
 					      &attr->ia_atime) < 0) {
 			inode_set_atime_to_ts(inode, attr->ia_atime);
@@ -2646,7 +2676,7 @@ retry:
 						     CEPH_FSCRYPT_BLOCK_SIZE));
 			req->r_fscrypt_file = attr->ia_size;
 			fill_fscrypt = true;
-		} else if ((issued & CEPH_CAP_FILE_EXCL) && attr->ia_size >= isize) {
+		} else if (!do_sync && (issued & CEPH_CAP_FILE_EXCL) && attr->ia_size >= isize) {
 			if (attr->ia_size > isize) {
 				i_size_write(inode, attr->ia_size);
 				inode->i_blocks = calc_inode_blocks(attr->ia_size);
@@ -2683,11 +2713,11 @@ retry:
 		      inode, ceph_vinop(inode),
 		      mtime.tv_sec, mtime.tv_nsec,
 		      attr->ia_mtime.tv_sec, attr->ia_mtime.tv_nsec);
-		if (issued & CEPH_CAP_FILE_EXCL) {
+		if (!do_sync && (issued & CEPH_CAP_FILE_EXCL)) {
 			ci->i_time_warp_seq++;
 			inode_set_mtime_to_ts(inode, attr->ia_mtime);
 			dirtied |= CEPH_CAP_FILE_EXCL;
-		} else if ((issued & CEPH_CAP_FILE_WR) &&
+		} else if (!do_sync && (issued & CEPH_CAP_FILE_WR) &&
 			   timespec64_compare(&mtime, &attr->ia_mtime) < 0) {
 			inode_set_mtime_to_ts(inode, attr->ia_mtime);
 			dirtied |= CEPH_CAP_FILE_WR;
diff --git a/fs/ceph/locks.c b/fs/ceph/locks.c
index e07ad29ff8..ebf4ac0055 100644
--- a/fs/ceph/locks.c
+++ b/fs/ceph/locks.c
@@ -33,7 +33,7 @@ void __init ceph_flock_init(void)
 
 static void ceph_fl_copy_lock(struct file_lock *dst, struct file_lock *src)
 {
-	struct inode *inode = file_inode(dst->fl_file);
+	struct inode *inode = file_inode(dst->c.flc_file);
 	atomic_inc(&ceph_inode(inode)->i_filelock_ref);
 	dst->fl_u.ceph.inode = igrab(inode);
 }
@@ -110,17 +110,18 @@ static int ceph_lock_message(u8 lock_type, u16 operation, struct inode *inode,
 	else
 		length = fl->fl_end - fl->fl_start + 1;
 
-	owner = secure_addr(fl->fl_owner);
+	owner = secure_addr(fl->c.flc_owner);
 
 	doutc(cl, "rule: %d, op: %d, owner: %llx, pid: %llu, "
 		    "start: %llu, length: %llu, wait: %d, type: %d\n",
-		    (int)lock_type, (int)operation, owner, (u64)fl->fl_pid,
-		    fl->fl_start, length, wait, fl->fl_type);
+		    (int)lock_type, (int)operation, owner,
+		    (u64) fl->c.flc_pid,
+		    fl->fl_start, length, wait, fl->c.flc_type);
 
 	req->r_args.filelock_change.rule = lock_type;
 	req->r_args.filelock_change.type = cmd;
 	req->r_args.filelock_change.owner = cpu_to_le64(owner);
-	req->r_args.filelock_change.pid = cpu_to_le64((u64)fl->fl_pid);
+	req->r_args.filelock_change.pid = cpu_to_le64((u64) fl->c.flc_pid);
 	req->r_args.filelock_change.start = cpu_to_le64(fl->fl_start);
 	req->r_args.filelock_change.length = cpu_to_le64(length);
 	req->r_args.filelock_change.wait = wait;
@@ -130,13 +131,13 @@ static int ceph_lock_message(u8 lock_type, u16 operation, struct inode *inode,
 		err = ceph_mdsc_wait_request(mdsc, req, wait ?
 					ceph_lock_wait_for_completion : NULL);
 	if (!err && operation == CEPH_MDS_OP_GETFILELOCK) {
-		fl->fl_pid = -le64_to_cpu(req->r_reply_info.filelock_reply->pid);
+		fl->c.flc_pid = -le64_to_cpu(req->r_reply_info.filelock_reply->pid);
 		if (CEPH_LOCK_SHARED == req->r_reply_info.filelock_reply->type)
-			fl->fl_type = F_RDLCK;
+			fl->c.flc_type = F_RDLCK;
 		else if (CEPH_LOCK_EXCL == req->r_reply_info.filelock_reply->type)
-			fl->fl_type = F_WRLCK;
+			fl->c.flc_type = F_WRLCK;
 		else
-			fl->fl_type = F_UNLCK;
+			fl->c.flc_type = F_UNLCK;
 
 		fl->fl_start = le64_to_cpu(req->r_reply_info.filelock_reply->start);
 		length = le64_to_cpu(req->r_reply_info.filelock_reply->start) +
@@ -150,8 +151,8 @@ static int ceph_lock_message(u8 lock_type, u16 operation, struct inode *inode,
 	ceph_mdsc_put_request(req);
 	doutc(cl, "rule: %d, op: %d, pid: %llu, start: %llu, "
 	      "length: %llu, wait: %d, type: %d, err code %d\n",
-	      (int)lock_type, (int)operation, (u64)fl->fl_pid,
-	      fl->fl_start, length, wait, fl->fl_type, err);
+	      (int)lock_type, (int)operation, (u64) fl->c.flc_pid,
+	      fl->fl_start, length, wait, fl->c.flc_type, err);
 	return err;
 }
 
@@ -227,10 +228,10 @@ static int ceph_lock_wait_for_completion(struct ceph_mds_client *mdsc,
 static int try_unlock_file(struct file *file, struct file_lock *fl)
 {
 	int err;
-	unsigned int orig_flags = fl->fl_flags;
-	fl->fl_flags |= FL_EXISTS;
+	unsigned int orig_flags = fl->c.flc_flags;
+	fl->c.flc_flags |= FL_EXISTS;
 	err = locks_lock_file_wait(file, fl);
-	fl->fl_flags = orig_flags;
+	fl->c.flc_flags = orig_flags;
 	if (err == -ENOENT) {
 		if (!(orig_flags & FL_EXISTS))
 			err = 0;
@@ -253,13 +254,13 @@ int ceph_lock(struct file *file, int cmd, struct file_lock *fl)
 	u8 wait = 0;
 	u8 lock_cmd;
 
-	if (!(fl->fl_flags & FL_POSIX))
+	if (!(fl->c.flc_flags & FL_POSIX))
 		return -ENOLCK;
 
 	if (ceph_inode_is_shutdown(inode))
 		return -ESTALE;
 
-	doutc(cl, "fl_owner: %p\n", fl->fl_owner);
+	doutc(cl, "fl_owner: %p\n", fl->c.flc_owner);
 
 	/* set wait bit as appropriate, then make command as Ceph expects it*/
 	if (IS_GETLK(cmd))
@@ -273,19 +274,19 @@ int ceph_lock(struct file *file, int cmd, struct file_lock *fl)
 	}
 	spin_unlock(&ci->i_ceph_lock);
 	if (err < 0) {
-		if (op == CEPH_MDS_OP_SETFILELOCK && F_UNLCK == fl->fl_type)
+		if (op == CEPH_MDS_OP_SETFILELOCK && lock_is_unlock(fl))
 			posix_lock_file(file, fl, NULL);
 		return err;
 	}
 
-	if (F_RDLCK == fl->fl_type)
+	if (lock_is_read(fl))
 		lock_cmd = CEPH_LOCK_SHARED;
-	else if (F_WRLCK == fl->fl_type)
+	else if (lock_is_write(fl))
 		lock_cmd = CEPH_LOCK_EXCL;
 	else
 		lock_cmd = CEPH_LOCK_UNLOCK;
 
-	if (op == CEPH_MDS_OP_SETFILELOCK && F_UNLCK == fl->fl_type) {
+	if (op == CEPH_MDS_OP_SETFILELOCK && lock_is_unlock(fl)) {
 		err = try_unlock_file(file, fl);
 		if (err <= 0)
 			return err;
@@ -293,7 +294,7 @@ int ceph_lock(struct file *file, int cmd, struct file_lock *fl)
 
 	err = ceph_lock_message(CEPH_LOCK_FCNTL, op, inode, lock_cmd, wait, fl);
 	if (!err) {
-		if (op == CEPH_MDS_OP_SETFILELOCK && F_UNLCK != fl->fl_type) {
+		if (op == CEPH_MDS_OP_SETFILELOCK && F_UNLCK != fl->c.flc_type) {
 			doutc(cl, "locking locally\n");
 			err = posix_lock_file(file, fl, NULL);
 			if (err) {
@@ -319,13 +320,13 @@ int ceph_flock(struct file *file, int cmd, struct file_lock *fl)
 	u8 wait = 0;
 	u8 lock_cmd;
 
-	if (!(fl->fl_flags & FL_FLOCK))
+	if (!(fl->c.flc_flags & FL_FLOCK))
 		return -ENOLCK;
 
 	if (ceph_inode_is_shutdown(inode))
 		return -ESTALE;
 
-	doutc(cl, "fl_file: %p\n", fl->fl_file);
+	doutc(cl, "fl_file: %p\n", fl->c.flc_file);
 
 	spin_lock(&ci->i_ceph_lock);
 	if (ci->i_ceph_flags & CEPH_I_ERROR_FILELOCK) {
@@ -333,7 +334,7 @@ int ceph_flock(struct file *file, int cmd, struct file_lock *fl)
 	}
 	spin_unlock(&ci->i_ceph_lock);
 	if (err < 0) {
-		if (F_UNLCK == fl->fl_type)
+		if (lock_is_unlock(fl))
 			locks_lock_file_wait(file, fl);
 		return err;
 	}
@@ -341,14 +342,14 @@ int ceph_flock(struct file *file, int cmd, struct file_lock *fl)
 	if (IS_SETLKW(cmd))
 		wait = 1;
 
-	if (F_RDLCK == fl->fl_type)
+	if (lock_is_read(fl))
 		lock_cmd = CEPH_LOCK_SHARED;
-	else if (F_WRLCK == fl->fl_type)
+	else if (lock_is_write(fl))
 		lock_cmd = CEPH_LOCK_EXCL;
 	else
 		lock_cmd = CEPH_LOCK_UNLOCK;
 
-	if (F_UNLCK == fl->fl_type) {
+	if (lock_is_unlock(fl)) {
 		err = try_unlock_file(file, fl);
 		if (err <= 0)
 			return err;
@@ -356,7 +357,7 @@ int ceph_flock(struct file *file, int cmd, struct file_lock *fl)
 
 	err = ceph_lock_message(CEPH_LOCK_FLOCK, CEPH_MDS_OP_SETFILELOCK,
 				inode, lock_cmd, wait, fl);
-	if (!err && F_UNLCK != fl->fl_type) {
+	if (!err && F_UNLCK != fl->c.flc_type) {
 		err = locks_lock_file_wait(file, fl);
 		if (err) {
 			ceph_lock_message(CEPH_LOCK_FLOCK,
@@ -385,9 +386,9 @@ void ceph_count_locks(struct inode *inode, int *fcntl_count, int *flock_count)
 	ctx = locks_inode_context(inode);
 	if (ctx) {
 		spin_lock(&ctx->flc_lock);
-		list_for_each_entry(lock, &ctx->flc_posix, fl_list)
+		for_each_file_lock(lock, &ctx->flc_posix)
 			++(*fcntl_count);
-		list_for_each_entry(lock, &ctx->flc_flock, fl_list)
+		for_each_file_lock(lock, &ctx->flc_flock)
 			++(*flock_count);
 		spin_unlock(&ctx->flc_lock);
 	}
@@ -408,10 +409,10 @@ static int lock_to_ceph_filelock(struct inode *inode,
 	cephlock->start = cpu_to_le64(lock->fl_start);
 	cephlock->length = cpu_to_le64(lock->fl_end - lock->fl_start + 1);
 	cephlock->client = cpu_to_le64(0);
-	cephlock->pid = cpu_to_le64((u64)lock->fl_pid);
-	cephlock->owner = cpu_to_le64(secure_addr(lock->fl_owner));
+	cephlock->pid = cpu_to_le64((u64) lock->c.flc_pid);
+	cephlock->owner = cpu_to_le64(secure_addr(lock->c.flc_owner));
 
-	switch (lock->fl_type) {
+	switch (lock->c.flc_type) {
 	case F_RDLCK:
 		cephlock->type = CEPH_LOCK_SHARED;
 		break;
@@ -422,7 +423,8 @@ static int lock_to_ceph_filelock(struct inode *inode,
 		cephlock->type = CEPH_LOCK_UNLOCK;
 		break;
 	default:
-		doutc(cl, "Have unknown lock type %d\n", lock->fl_type);
+		doutc(cl, "Have unknown lock type %d\n",
+		      lock->c.flc_type);
 		err = -EINVAL;
 	}
 
@@ -453,7 +455,7 @@ int ceph_encode_locks_to_buffer(struct inode *inode,
 		return 0;
 
 	spin_lock(&ctx->flc_lock);
-	list_for_each_entry(lock, &ctx->flc_posix, fl_list) {
+	for_each_file_lock(lock, &ctx->flc_posix) {
 		++seen_fcntl;
 		if (seen_fcntl > num_fcntl_locks) {
 			err = -ENOSPC;
@@ -464,7 +466,7 @@ int ceph_encode_locks_to_buffer(struct inode *inode,
 			goto fail;
 		++l;
 	}
-	list_for_each_entry(lock, &ctx->flc_flock, fl_list) {
+	for_each_file_lock(lock, &ctx->flc_flock) {
 		++seen_flock;
 		if (seen_flock > num_flock_locks) {
 			err = -ENOSPC;
diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
index 360b686c3c..c2157f6e0c 100644
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@ -4112,10 +4112,13 @@ static void handle_session(struct ceph_mds_session *session,
 	void *p = msg->front.iov_base;
 	void *end = p + msg->front.iov_len;
 	struct ceph_mds_session_head *h;
-	u32 op;
+	struct ceph_mds_cap_auth *cap_auths = NULL;
+	u32 op, cap_auths_num = 0;
 	u64 seq, features = 0;
 	int wake = 0;
 	bool blocklisted = false;
+	u32 i;
+
 
 	/* decode */
 	ceph_decode_need(&p, end, sizeof(*h), bad);
@@ -4160,7 +4163,101 @@ static void handle_session(struct ceph_mds_session *session,
 		}
 	}
 
+	if (msg_version >= 6) {
+		ceph_decode_32_safe(&p, end, cap_auths_num, bad);
+		doutc(cl, "cap_auths_num %d\n", cap_auths_num);
+
+		if (cap_auths_num && op != CEPH_SESSION_OPEN) {
+			WARN_ON_ONCE(op != CEPH_SESSION_OPEN);
+			goto skip_cap_auths;
+		}
+
+		cap_auths = kcalloc(cap_auths_num,
+				    sizeof(struct ceph_mds_cap_auth),
+				    GFP_KERNEL);
+		if (!cap_auths) {
+			pr_err_client(cl, "No memory for cap_auths\n");
+			return;
+		}
+
+		for (i = 0; i < cap_auths_num; i++) {
+			u32 _len, j;
+
+			/* struct_v, struct_compat, and struct_len in MDSCapAuth */
+			ceph_decode_skip_n(&p, end, 2 + sizeof(u32), bad);
+
+			/* struct_v, struct_compat, and struct_len in MDSCapMatch */
+			ceph_decode_skip_n(&p, end, 2 + sizeof(u32), bad);
+			ceph_decode_64_safe(&p, end, cap_auths[i].match.uid, bad);
+			ceph_decode_32_safe(&p, end, _len, bad);
+			if (_len) {
+				cap_auths[i].match.gids = kcalloc(_len, sizeof(u32),
+								  GFP_KERNEL);
+				if (!cap_auths[i].match.gids) {
+					pr_err_client(cl, "No memory for gids\n");
+					goto fail;
+				}
+
+				cap_auths[i].match.num_gids = _len;
+				for (j = 0; j < _len; j++)
+					ceph_decode_32_safe(&p, end,
+							    cap_auths[i].match.gids[j],
+							    bad);
+			}
+
+			ceph_decode_32_safe(&p, end, _len, bad);
+			if (_len) {
+				cap_auths[i].match.path = kcalloc(_len + 1, sizeof(char),
+								  GFP_KERNEL);
+				if (!cap_auths[i].match.path) {
+					pr_err_client(cl, "No memory for path\n");
+					goto fail;
+				}
+				ceph_decode_copy(&p, cap_auths[i].match.path, _len);
+
+				/* Remove the tailing '/' */
+				while (_len && cap_auths[i].match.path[_len - 1] == '/') {
+					cap_auths[i].match.path[_len - 1] = '\0';
+					_len -= 1;
+				}
+			}
+
+			ceph_decode_32_safe(&p, end, _len, bad);
+			if (_len) {
+				cap_auths[i].match.fs_name = kcalloc(_len + 1, sizeof(char),
+								     GFP_KERNEL);
+				if (!cap_auths[i].match.fs_name) {
+					pr_err_client(cl, "No memory for fs_name\n");
+					goto fail;
+				}
+				ceph_decode_copy(&p, cap_auths[i].match.fs_name, _len);
+			}
+
+			ceph_decode_8_safe(&p, end, cap_auths[i].match.root_squash, bad);
+			ceph_decode_8_safe(&p, end, cap_auths[i].readable, bad);
+			ceph_decode_8_safe(&p, end, cap_auths[i].writeable, bad);
+			doutc(cl, "uid %lld, num_gids %u, path %s, fs_name %s, root_squash %d, readable %d, writeable %d\n",
+			      cap_auths[i].match.uid, cap_auths[i].match.num_gids,
+			      cap_auths[i].match.path, cap_auths[i].match.fs_name,
+			      cap_auths[i].match.root_squash,
+			      cap_auths[i].readable, cap_auths[i].writeable);
+		}
+	}
+
+skip_cap_auths:
 	mutex_lock(&mdsc->mutex);
+	if (op == CEPH_SESSION_OPEN) {
+		if (mdsc->s_cap_auths) {
+			for (i = 0; i < mdsc->s_cap_auths_num; i++) {
+				kfree(mdsc->s_cap_auths[i].match.gids);
+				kfree(mdsc->s_cap_auths[i].match.path);
+				kfree(mdsc->s_cap_auths[i].match.fs_name);
+			}
+			kfree(mdsc->s_cap_auths);
+		}
+		mdsc->s_cap_auths_num = cap_auths_num;
+		mdsc->s_cap_auths = cap_auths;
+	}
 	if (op == CEPH_SESSION_CLOSE) {
 		ceph_get_mds_session(session);
 		__unregister_session(mdsc, session);
@@ -4290,6 +4387,13 @@ bad:
 	pr_err_client(cl, "corrupt message mds%d len %d\n", mds,
 		      (int)msg->front.iov_len);
 	ceph_msg_dump(msg);
+fail:
+	for (i = 0; i < cap_auths_num; i++) {
+		kfree(cap_auths[i].match.gids);
+		kfree(cap_auths[i].match.path);
+		kfree(cap_auths[i].match.fs_name);
+	}
+	kfree(cap_auths);
 	return;
 }
 
@@ -5499,6 +5603,170 @@ void send_flush_mdlog(struct ceph_mds_session *s)
 	mutex_unlock(&s->s_mutex);
 }
 
+static int ceph_mds_auth_match(struct ceph_mds_client *mdsc,
+			       struct ceph_mds_cap_auth *auth,
+			       char *tpath)
+{
+	const struct cred *cred = get_current_cred();
+	u32 caller_uid = from_kuid(&init_user_ns, cred->fsuid);
+	u32 caller_gid = from_kgid(&init_user_ns, cred->fsgid);
+	struct ceph_client *cl = mdsc->fsc->client;
+	const char *spath = mdsc->fsc->mount_options->server_path;
+	bool gid_matched = false;
+	u32 gid, tlen, len;
+	int i, j;
+
+	doutc(cl, "match.uid %lld\n", auth->match.uid);
+	if (auth->match.uid != MDS_AUTH_UID_ANY) {
+		if (auth->match.uid != caller_uid)
+			return 0;
+		if (auth->match.num_gids) {
+			for (i = 0; i < auth->match.num_gids; i++) {
+				if (caller_gid == auth->match.gids[i])
+					gid_matched = true;
+			}
+			if (!gid_matched && cred->group_info->ngroups) {
+				for (i = 0; i < cred->group_info->ngroups; i++) {
+					gid = from_kgid(&init_user_ns,
+							cred->group_info->gid[i]);
+					for (j = 0; j < auth->match.num_gids; j++) {
+						if (gid == auth->match.gids[j]) {
+							gid_matched = true;
+							break;
+						}
+					}
+					if (gid_matched)
+						break;
+				}
+			}
+			if (!gid_matched)
+				return 0;
+		}
+	}
+
+	/* path match */
+	if (auth->match.path) {
+		if (!tpath)
+			return 0;
+
+		tlen = strlen(tpath);
+		len = strlen(auth->match.path);
+		if (len) {
+			char *_tpath = tpath;
+			bool free_tpath = false;
+			int m, n;
+
+			doutc(cl, "server path %s, tpath %s, match.path %s\n",
+			      spath, tpath, auth->match.path);
+			if (spath && (m = strlen(spath)) != 1) {
+				/* mount path + '/' + tpath + an extra space */
+				n = m + 1 + tlen + 1;
+				_tpath = kmalloc(n, GFP_NOFS);
+				if (!_tpath)
+					return -ENOMEM;
+				/* remove the leading '/' */
+				snprintf(_tpath, n, "%s/%s", spath + 1, tpath);
+				free_tpath = true;
+				tlen = strlen(_tpath);
+			}
+
+			/*
+			 * Please note the tailing '/' for match.path has already
+			 * been removed when parsing.
+			 *
+			 * Remove the tailing '/' for the target path.
+			 */
+			while (tlen && _tpath[tlen - 1] == '/') {
+				_tpath[tlen - 1] = '\0';
+				tlen -= 1;
+			}
+			doutc(cl, "_tpath %s\n", _tpath);
+
+			/*
+			 * In case first == _tpath && tlen == len:
+			 *  match.path=/foo  --> /foo _path=/foo     --> match
+			 *  match.path=/foo/ --> /foo _path=/foo     --> match
+			 *
+			 * In case first == _tmatch.path && tlen > len:
+			 *  match.path=/foo/ --> /foo _path=/foo/    --> match
+			 *  match.path=/foo  --> /foo _path=/foo/    --> match
+			 *  match.path=/foo/ --> /foo _path=/foo/d   --> match
+			 *  match.path=/foo  --> /foo _path=/food    --> mismatch
+			 *
+			 * All the other cases                       --> mismatch
+			 */
+			char *first = strstr(_tpath, auth->match.path);
+			if (first != _tpath) {
+				if (free_tpath)
+					kfree(_tpath);
+				return 0;
+			}
+
+			if (tlen > len && _tpath[len] != '/') {
+				if (free_tpath)
+					kfree(_tpath);
+				return 0;
+			}
+		}
+	}
+
+	doutc(cl, "matched\n");
+	return 1;
+}
+
+int ceph_mds_check_access(struct ceph_mds_client *mdsc, char *tpath, int mask)
+{
+	const struct cred *cred = get_current_cred();
+	u32 caller_uid = from_kuid(&init_user_ns, cred->fsuid);
+	u32 caller_gid = from_kgid(&init_user_ns, cred->fsgid);
+	struct ceph_mds_cap_auth *rw_perms_s = NULL;
+	struct ceph_client *cl = mdsc->fsc->client;
+	bool root_squash_perms = true;
+	int i, err;
+
+	doutc(cl, "tpath '%s', mask %d, caller_uid %d, caller_gid %d\n",
+	      tpath, mask, caller_uid, caller_gid);
+
+	for (i = 0; i < mdsc->s_cap_auths_num; i++) {
+		struct ceph_mds_cap_auth *s = &mdsc->s_cap_auths[i];
+
+		err = ceph_mds_auth_match(mdsc, s, tpath);
+		if (err < 0) {
+			return err;
+		} else if (err > 0) {
+			/* always follow the last auth caps' permision */
+			root_squash_perms = true;
+			rw_perms_s = NULL;
+			if ((mask & MAY_WRITE) && s->writeable &&
+			    s->match.root_squash && (!caller_uid || !caller_gid))
+				root_squash_perms = false;
+
+			if (((mask & MAY_WRITE) && !s->writeable) ||
+			    ((mask & MAY_READ) && !s->readable))
+				rw_perms_s = s;
+		}
+	}
+
+	doutc(cl, "root_squash_perms %d, rw_perms_s %p\n", root_squash_perms,
+	      rw_perms_s);
+	if (root_squash_perms && rw_perms_s == NULL) {
+		doutc(cl, "access allowed\n");
+		return 0;
+	}
+
+	if (!root_squash_perms) {
+		doutc(cl, "root_squash is enabled and user(%d %d) isn't allowed to write",
+		      caller_uid, caller_gid);
+	}
+	if (rw_perms_s) {
+		doutc(cl, "mds auth caps readable/writeable %d/%d while request r/w %d/%d",
+		      rw_perms_s->readable, rw_perms_s->writeable,
+		      !!(mask & MAY_READ), !!(mask & MAY_WRITE));
+	}
+	doutc(cl, "access denied\n");
+	return -EACCES;
+}
+
 /*
  * called before mount is ro, and before dentries are torn down.
  * (hmm, does this still race with new lookups?)
diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h
index b88e804152..cfa18cf915 100644
--- a/fs/ceph/mds_client.h
+++ b/fs/ceph/mds_client.h
@@ -35,8 +35,9 @@ enum ceph_feature_type {
 	CEPHFS_FEATURE_32BITS_RETRY_FWD,
 	CEPHFS_FEATURE_NEW_SNAPREALM_INFO,
 	CEPHFS_FEATURE_HAS_OWNER_UIDGID,
+	CEPHFS_FEATURE_MDS_AUTH_CAPS_CHECK,
 
-	CEPHFS_FEATURE_MAX = CEPHFS_FEATURE_HAS_OWNER_UIDGID,
+	CEPHFS_FEATURE_MAX = CEPHFS_FEATURE_MDS_AUTH_CAPS_CHECK,
 };
 
 #define CEPHFS_FEATURES_CLIENT_SUPPORTED {	\
@@ -52,6 +53,7 @@ enum ceph_feature_type {
 	CEPHFS_FEATURE_OP_GETVXATTR,		\
 	CEPHFS_FEATURE_32BITS_RETRY_FWD,	\
 	CEPHFS_FEATURE_HAS_OWNER_UIDGID,	\
+	CEPHFS_FEATURE_MDS_AUTH_CAPS_CHECK,	\
 }
 
 /*
@@ -71,6 +73,24 @@ enum ceph_feature_type {
 struct ceph_fs_client;
 struct ceph_cap;
 
+#define MDS_AUTH_UID_ANY -1
+
+struct ceph_mds_cap_match {
+	s64 uid;  /* default to MDS_AUTH_UID_ANY */
+	u32 num_gids;
+	u32 *gids;  /* use these GIDs */
+	char *path;  /* require path to be child of this
+			(may be "" or "/" for any) */
+	char *fs_name;
+	bool root_squash;  /* default to false */
+};
+
+struct ceph_mds_cap_auth {
+	struct ceph_mds_cap_match match;
+	bool readable;
+	bool writeable;
+};
+
 /*
  * parsed info about a single inode.  pointers are into the encoded
  * on-wire structures within the mds reply message payload.
@@ -513,6 +533,9 @@ struct ceph_mds_client {
 	struct rw_semaphore     pool_perm_rwsem;
 	struct rb_root		pool_perm_tree;
 
+	u32			 s_cap_auths_num;
+	struct ceph_mds_cap_auth *s_cap_auths;
+
 	char nodename[__NEW_UTS_LEN + 1];
 };
 
@@ -581,6 +604,9 @@ extern void ceph_queue_cap_unlink_work(struct ceph_mds_client *mdsc);
 extern int ceph_iterate_session_caps(struct ceph_mds_session *session,
 				     int (*cb)(struct inode *, int mds, void *),
 				     void *arg);
+extern int ceph_mds_check_access(struct ceph_mds_client *mdsc, char *tpath,
+				 int mask);
+
 extern void ceph_mdsc_pre_umount(struct ceph_mds_client *mdsc);
 
 static inline void ceph_mdsc_free_path(char *path, int len)
diff --git a/fs/ceph/super.c b/fs/ceph/super.c
index 5ec102f6b1..0cdf84cd17 100644
--- a/fs/ceph/super.c
+++ b/fs/ceph/super.c
@@ -928,40 +928,41 @@ static int __init init_caches(void)
 	ceph_inode_cachep = kmem_cache_create("ceph_inode_info",
 				      sizeof(struct ceph_inode_info),
 				      __alignof__(struct ceph_inode_info),
-				      SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD|
-				      SLAB_ACCOUNT, ceph_inode_init_once);
+				      SLAB_RECLAIM_ACCOUNT | SLAB_ACCOUNT,
+				      ceph_inode_init_once);
 	if (!ceph_inode_cachep)
 		return -ENOMEM;
 
-	ceph_cap_cachep = KMEM_CACHE(ceph_cap, SLAB_MEM_SPREAD);
+	ceph_cap_cachep = KMEM_CACHE(ceph_cap, 0);
 	if (!ceph_cap_cachep)
 		goto bad_cap;
-	ceph_cap_snap_cachep = KMEM_CACHE(ceph_cap_snap, SLAB_MEM_SPREAD);
+	ceph_cap_snap_cachep = KMEM_CACHE(ceph_cap_snap, 0);
 	if (!ceph_cap_snap_cachep)
 		goto bad_cap_snap;
 	ceph_cap_flush_cachep = KMEM_CACHE(ceph_cap_flush,
-					   SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD);
+					   SLAB_RECLAIM_ACCOUNT);
 	if (!ceph_cap_flush_cachep)
 		goto bad_cap_flush;
 
 	ceph_dentry_cachep = KMEM_CACHE(ceph_dentry_info,
-					SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD);
+					SLAB_RECLAIM_ACCOUNT);
 	if (!ceph_dentry_cachep)
 		goto bad_dentry;
 
-	ceph_file_cachep = KMEM_CACHE(ceph_file_info, SLAB_MEM_SPREAD);
+	ceph_file_cachep = KMEM_CACHE(ceph_file_info, 0);
 	if (!ceph_file_cachep)
 		goto bad_file;
 
-	ceph_dir_file_cachep = KMEM_CACHE(ceph_dir_file_info, SLAB_MEM_SPREAD);
+	ceph_dir_file_cachep = KMEM_CACHE(ceph_dir_file_info, 0);
 	if (!ceph_dir_file_cachep)
 		goto bad_dir_file;
 
-	ceph_mds_request_cachep = KMEM_CACHE(ceph_mds_request, SLAB_MEM_SPREAD);
+	ceph_mds_request_cachep = KMEM_CACHE(ceph_mds_request, 0);
 	if (!ceph_mds_request_cachep)
 		goto bad_mds_req;
 
-	ceph_wb_pagevec_pool = mempool_create_kmalloc_pool(10, CEPH_MAX_WRITE_SIZE >> PAGE_SHIFT);
+	ceph_wb_pagevec_pool = mempool_create_kmalloc_pool(10,
+	    (CEPH_MAX_WRITE_SIZE >> PAGE_SHIFT) * sizeof(struct page *));
 	if (!ceph_wb_pagevec_pool)
 		goto bad_pagevec_pool;
 
diff --git a/fs/ceph/super.h b/fs/ceph/super.h
index b63b4cd9b5..6e817bf133 100644
--- a/fs/ceph/super.h
+++ b/fs/ceph/super.h
@@ -200,9 +200,10 @@ struct ceph_cap {
 	struct list_head caps_item;
 };
 
-#define CHECK_CAPS_AUTHONLY   1  /* only check auth cap */
-#define CHECK_CAPS_FLUSH      2  /* flush any dirty caps */
-#define CHECK_CAPS_NOINVAL    4  /* don't invalidate pagecache */
+#define CHECK_CAPS_AUTHONLY     1  /* only check auth cap */
+#define CHECK_CAPS_FLUSH        2  /* flush any dirty caps */
+#define CHECK_CAPS_NOINVAL      4  /* don't invalidate pagecache */
+#define CHECK_CAPS_FLUSH_FORCE  8  /* force flush any caps */
 
 struct ceph_cap_flush {
 	u64 tid;
diff --git a/fs/coda/inode.c b/fs/coda/inode.c
index 0c7c252879..6898dc6210 100644
--- a/fs/coda/inode.c
+++ b/fs/coda/inode.c
@@ -24,6 +24,8 @@
 #include <linux/pid_namespace.h>
 #include <linux/uaccess.h>
 #include <linux/fs.h>
+#include <linux/fs_context.h>
+#include <linux/fs_parser.h>
 #include <linux/vmalloc.h>
 
 #include <linux/coda.h>
@@ -70,8 +72,8 @@ int __init coda_init_inodecache(void)
 {
 	coda_inode_cachep = kmem_cache_create("coda_inode_cache",
 				sizeof(struct coda_inode_info), 0,
-				SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD|
-				SLAB_ACCOUNT, init_once);
+				SLAB_RECLAIM_ACCOUNT | SLAB_ACCOUNT,
+				init_once);
 	if (coda_inode_cachep == NULL)
 		return -ENOMEM;
 	return 0;
@@ -87,10 +89,10 @@ void coda_destroy_inodecache(void)
 	kmem_cache_destroy(coda_inode_cachep);
 }
 
-static int coda_remount(struct super_block *sb, int *flags, char *data)
+static int coda_reconfigure(struct fs_context *fc)
 {
-	sync_filesystem(sb);
-	*flags |= SB_NOATIME;
+	sync_filesystem(fc->root->d_sb);
+	fc->sb_flags |= SB_NOATIME;
 	return 0;
 }
 
@@ -102,78 +104,102 @@ static const struct super_operations coda_super_operations =
 	.evict_inode	= coda_evict_inode,
 	.put_super	= coda_put_super,
 	.statfs		= coda_statfs,
-	.remount_fs	= coda_remount,
 };
 
-static int get_device_index(struct coda_mount_data *data)
+struct coda_fs_context {
+	int	idx;
+};
+
+enum {
+	Opt_fd,
+};
+
+static const struct fs_parameter_spec coda_param_specs[] = {
+	fsparam_fd	("fd",	Opt_fd),
+	{}
+};
+
+static int coda_parse_fd(struct fs_context *fc, int fd)
 {
+	struct coda_fs_context *ctx = fc->fs_private;
 	struct fd f;
 	struct inode *inode;
 	int idx;
 
-	if (data == NULL) {
-		pr_warn("%s: Bad mount data\n", __func__);
-		return -1;
-	}
-
-	if (data->version != CODA_MOUNT_VERSION) {
-		pr_warn("%s: Bad mount version\n", __func__);
-		return -1;
-	}
-
-	f = fdget(data->fd);
+	f = fdget(fd);
 	if (!f.file)
-		goto Ebadf;
+		return -EBADF;
 	inode = file_inode(f.file);
 	if (!S_ISCHR(inode->i_mode) || imajor(inode) != CODA_PSDEV_MAJOR) {
 		fdput(f);
-		goto Ebadf;
+		return invalf(fc, "code: Not coda psdev");
 	}
 
 	idx = iminor(inode);
 	fdput(f);
 
-	if (idx < 0 || idx >= MAX_CODADEVS) {
-		pr_warn("%s: Bad minor number\n", __func__);
-		return -1;
+	if (idx < 0 || idx >= MAX_CODADEVS)
+		return invalf(fc, "coda: Bad minor number");
+	ctx->idx = idx;
+	return 0;
+}
+
+static int coda_parse_param(struct fs_context *fc, struct fs_parameter *param)
+{
+	struct fs_parse_result result;
+	int opt;
+
+	opt = fs_parse(fc, coda_param_specs, param, &result);
+	if (opt < 0)
+		return opt;
+
+	switch (opt) {
+	case Opt_fd:
+		return coda_parse_fd(fc, result.uint_32);
 	}
 
-	return idx;
-Ebadf:
-	pr_warn("%s: Bad file\n", __func__);
-	return -1;
+	return 0;
+}
+
+/*
+ * Parse coda's binary mount data form.  We ignore any errors and go with index
+ * 0 if we get one for backward compatibility.
+ */
+static int coda_parse_monolithic(struct fs_context *fc, void *_data)
+{
+	struct coda_mount_data *data = _data;
+
+	if (!data)
+		return invalf(fc, "coda: Bad mount data");
+
+	if (data->version != CODA_MOUNT_VERSION)
+		return invalf(fc, "coda: Bad mount version");
+
+	coda_parse_fd(fc, data->fd);
+	return 0;
 }
 
-static int coda_fill_super(struct super_block *sb, void *data, int silent)
+static int coda_fill_super(struct super_block *sb, struct fs_context *fc)
 {
+	struct coda_fs_context *ctx = fc->fs_private;
 	struct inode *root = NULL;
 	struct venus_comm *vc;
 	struct CodaFid fid;
 	int error;
-	int idx;
-
-	if (task_active_pid_ns(current) != &init_pid_ns)
-		return -EINVAL;
-
-	idx = get_device_index((struct coda_mount_data *) data);
 
-	/* Ignore errors in data, for backward compatibility */
-	if(idx == -1)
-		idx = 0;
-	
-	pr_info("%s: device index: %i\n", __func__,  idx);
+	infof(fc, "coda: device index: %i\n", ctx->idx);
 
-	vc = &coda_comms[idx];
+	vc = &coda_comms[ctx->idx];
 	mutex_lock(&vc->vc_mutex);
 
 	if (!vc->vc_inuse) {
-		pr_warn("%s: No pseudo device\n", __func__);
+		errorf(fc, "coda: No pseudo device");
 		error = -EINVAL;
 		goto unlock_out;
 	}
 
 	if (vc->vc_sb) {
-		pr_warn("%s: Device already mounted\n", __func__);
+		errorf(fc, "coda: Device already mounted");
 		error = -EBUSY;
 		goto unlock_out;
 	}
@@ -313,18 +339,45 @@ static int coda_statfs(struct dentry *dentry, struct kstatfs *buf)
 	return 0; 
 }
 
-/* init_coda: used by filesystems.c to register coda */
+static int coda_get_tree(struct fs_context *fc)
+{
+	if (task_active_pid_ns(current) != &init_pid_ns)
+		return -EINVAL;
 
-static struct dentry *coda_mount(struct file_system_type *fs_type,
-	int flags, const char *dev_name, void *data)
+	return get_tree_nodev(fc, coda_fill_super);
+}
+
+static void coda_free_fc(struct fs_context *fc)
 {
-	return mount_nodev(fs_type, flags, data, coda_fill_super);
+	kfree(fc->fs_private);
+}
+
+static const struct fs_context_operations coda_context_ops = {
+	.free		= coda_free_fc,
+	.parse_param	= coda_parse_param,
+	.parse_monolithic = coda_parse_monolithic,
+	.get_tree	= coda_get_tree,
+	.reconfigure	= coda_reconfigure,
+};
+
+static int coda_init_fs_context(struct fs_context *fc)
+{
+	struct coda_fs_context *ctx;
+
+	ctx = kzalloc(sizeof(struct coda_fs_context), GFP_KERNEL);
+	if (!ctx)
+		return -ENOMEM;
+
+	fc->fs_private = ctx;
+	fc->ops = &coda_context_ops;
+	return 0;
 }
 
 struct file_system_type coda_fs_type = {
 	.owner		= THIS_MODULE,
 	.name		= "coda",
-	.mount		= coda_mount,
+	.init_fs_context = coda_init_fs_context,
+	.parameters	= coda_param_specs,
 	.kill_sb	= kill_anon_super,
 	.fs_flags	= FS_BINARY_MOUNTDATA,
 };
diff --git a/fs/coredump.c b/fs/coredump.c
index be6403b4b1..a57a06b80f 100644
--- a/fs/coredump.c
+++ b/fs/coredump.c
@@ -56,10 +56,15 @@
 static bool dump_vma_snapshot(struct coredump_params *cprm);
 static void free_vma_snapshot(struct coredump_params *cprm);
 
+#define CORE_FILE_NOTE_SIZE_DEFAULT (4*1024*1024)
+/* Define a reasonable max cap */
+#define CORE_FILE_NOTE_SIZE_MAX (16*1024*1024)
+
 static int core_uses_pid;
 static unsigned int core_pipe_limit;
 static char core_pattern[CORENAME_MAX_SIZE] = "core";
 static int core_name_size = CORENAME_MAX_SIZE;
+unsigned int core_file_note_size_limit = CORE_FILE_NOTE_SIZE_DEFAULT;
 
 struct core_name {
 	char *corename;
@@ -371,9 +376,7 @@ static int zap_process(struct task_struct *start, int exit_code)
 		if (t != current && !(t->flags & PF_POSTCOREDUMP)) {
 			sigaddset(&t->pending.signal, SIGKILL);
 			signal_wake_up(t, 1);
-			/* The vhost_worker does not particpate in coredumps */
-			if ((t->flags & (PF_USER_WORKER | PF_IO_WORKER)) != PF_USER_WORKER)
-				nr++;
+			nr++;
 		}
 	}
 
@@ -998,6 +1001,9 @@ static int proc_dostring_coredump(struct ctl_table *table, int write,
 	return error;
 }
 
+static const unsigned int core_file_note_size_min = CORE_FILE_NOTE_SIZE_DEFAULT;
+static const unsigned int core_file_note_size_max = CORE_FILE_NOTE_SIZE_MAX;
+
 static struct ctl_table coredump_sysctls[] = {
 	{
 		.procname	= "core_uses_pid",
@@ -1020,6 +1026,15 @@ static struct ctl_table coredump_sysctls[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec,
 	},
+	{
+		.procname       = "core_file_note_size_limit",
+		.data           = &core_file_note_size_limit,
+		.maxlen         = sizeof(unsigned int),
+		.mode           = 0644,
+		.proc_handler	= proc_douintvec_minmax,
+		.extra1		= (unsigned int *)&core_file_note_size_min,
+		.extra2		= (unsigned int *)&core_file_note_size_max,
+	},
 };
 
 static int __init init_fs_coredump_sysctls(void)
diff --git a/fs/cramfs/inode.c b/fs/cramfs/inode.c
index 60dbfa0f88..460690ca01 100644
--- a/fs/cramfs/inode.c
+++ b/fs/cramfs/inode.c
@@ -183,7 +183,7 @@ static int next_buffer;
 static void *cramfs_blkdev_read(struct super_block *sb, unsigned int offset,
 				unsigned int len)
 {
-	struct address_space *mapping = sb->s_bdev->bd_inode->i_mapping;
+	struct address_space *mapping = sb->s_bdev->bd_mapping;
 	struct file_ra_state ra = {};
 	struct page *pages[BLKS_PER_BUF];
 	unsigned i, blocknr, buffer;
@@ -495,7 +495,7 @@ static void cramfs_kill_sb(struct super_block *sb)
 		sb->s_mtd = NULL;
 	} else if (IS_ENABLED(CONFIG_CRAMFS_BLOCKDEV) && sb->s_bdev) {
 		sync_blockdev(sb->s_bdev);
-		bdev_release(sb->s_bdev_handle);
+		bdev_fput(sb->s_bdev_file);
 	}
 	kfree(sbi);
 }
diff --git a/fs/crypto/fname.c b/fs/crypto/fname.c
index 7b3fc18959..0ad52fbe51 100644
--- a/fs/crypto/fname.c
+++ b/fs/crypto/fname.c
@@ -74,13 +74,7 @@ struct fscrypt_nokey_name {
 
 static inline bool fscrypt_is_dot_dotdot(const struct qstr *str)
 {
-	if (str->len == 1 && str->name[0] == '.')
-		return true;
-
-	if (str->len == 2 && str->name[0] == '.' && str->name[1] == '.')
-		return true;
-
-	return false;
+	return is_dot_dotdot(str->name, str->len);
 }
 
 /**
diff --git a/fs/crypto/fscrypt_private.h b/fs/crypto/fscrypt_private.h
index 1892356cf9..8371e4e1f5 100644
--- a/fs/crypto/fscrypt_private.h
+++ b/fs/crypto/fscrypt_private.h
@@ -222,16 +222,19 @@ struct fscrypt_inode_info {
 	struct fscrypt_prepared_key ci_enc_key;
 
 	/* True if ci_enc_key should be freed when this struct is freed */
-	bool ci_owns_key;
+	u8 ci_owns_key : 1;
 
 #ifdef CONFIG_FS_ENCRYPTION_INLINE_CRYPT
 	/*
 	 * True if this inode will use inline encryption (blk-crypto) instead of
 	 * the traditional filesystem-layer encryption.
 	 */
-	bool ci_inlinecrypt;
+	u8 ci_inlinecrypt : 1;
 #endif
 
+	/* True if ci_dirhash_key is initialized */
+	u8 ci_dirhash_key_initialized : 1;
+
 	/*
 	 * log2 of the data unit size (granularity of contents encryption) of
 	 * this file.  This is computable from ci_policy and ci_inode but is
@@ -242,6 +245,9 @@ struct fscrypt_inode_info {
 	/* Cached value: log2 of number of data units per FS block */
 	u8 ci_data_units_per_block_bits;
 
+	/* Hashed inode number.  Only set for IV_INO_LBLK_32 */
+	u32 ci_hashed_ino;
+
 	/*
 	 * Encryption mode used for this inode.  It corresponds to either the
 	 * contents or filenames encryption mode, depending on the inode type.
@@ -276,16 +282,12 @@ struct fscrypt_inode_info {
 	 * the plaintext filenames -- currently just casefolded directories.
 	 */
 	siphash_key_t ci_dirhash_key;
-	bool ci_dirhash_key_initialized;
 
 	/* The encryption policy used by this inode */
 	union fscrypt_policy ci_policy;
 
 	/* This inode's nonce, copied from the fscrypt_context */
 	u8 ci_nonce[FSCRYPT_FILE_NONCE_SIZE];
-
-	/* Hashed inode number.  Only set for IV_INO_LBLK_32 */
-	u32 ci_hashed_ino;
 };
 
 typedef enum {
diff --git a/fs/crypto/hooks.c b/fs/crypto/hooks.c
index 52504dd478..d8d5049b8f 100644
--- a/fs/crypto/hooks.c
+++ b/fs/crypto/hooks.c
@@ -30,21 +30,41 @@
 int fscrypt_file_open(struct inode *inode, struct file *filp)
 {
 	int err;
-	struct dentry *dir;
+	struct dentry *dentry, *dentry_parent;
+	struct inode *inode_parent;
 
 	err = fscrypt_require_key(inode);
 	if (err)
 		return err;
 
-	dir = dget_parent(file_dentry(filp));
-	if (IS_ENCRYPTED(d_inode(dir)) &&
-	    !fscrypt_has_permitted_context(d_inode(dir), inode)) {
+	dentry = file_dentry(filp);
+
+	/*
+	 * Getting a reference to the parent dentry is needed for the actual
+	 * encryption policy comparison, but it's expensive on multi-core
+	 * systems.  Since this function runs on unencrypted files too, start
+	 * with a lightweight RCU-mode check for the parent directory being
+	 * unencrypted (in which case it's fine for the child to be either
+	 * unencrypted, or encrypted with any policy).  Only continue on to the
+	 * full policy check if the parent directory is actually encrypted.
+	 */
+	rcu_read_lock();
+	dentry_parent = READ_ONCE(dentry->d_parent);
+	inode_parent = d_inode_rcu(dentry_parent);
+	if (inode_parent != NULL && !IS_ENCRYPTED(inode_parent)) {
+		rcu_read_unlock();
+		return 0;
+	}
+	rcu_read_unlock();
+
+	dentry_parent = dget_parent(dentry);
+	if (!fscrypt_has_permitted_context(d_inode(dentry_parent), inode)) {
 		fscrypt_warn(inode,
 			     "Inconsistent encryption context (parent directory: %lu)",
-			     d_inode(dir)->i_ino);
+			     d_inode(dentry_parent)->i_ino);
 		err = -EPERM;
 	}
-	dput(dir);
+	dput(dentry_parent);
 	return err;
 }
 EXPORT_SYMBOL_GPL(fscrypt_file_open);
@@ -102,11 +122,8 @@ int __fscrypt_prepare_lookup(struct inode *dir, struct dentry *dentry,
 	if (err && err != -ENOENT)
 		return err;
 
-	if (fname->is_nokey_name) {
-		spin_lock(&dentry->d_lock);
-		dentry->d_flags |= DCACHE_NOKEY_NAME;
-		spin_unlock(&dentry->d_lock);
-	}
+	fscrypt_prepare_dentry(dentry, fname->is_nokey_name);
+
 	return err;
 }
 EXPORT_SYMBOL_GPL(__fscrypt_prepare_lookup);
@@ -131,12 +148,10 @@ EXPORT_SYMBOL_GPL(__fscrypt_prepare_lookup);
 int fscrypt_prepare_lookup_partial(struct inode *dir, struct dentry *dentry)
 {
 	int err = fscrypt_get_encryption_info(dir, true);
+	bool is_nokey_name = (!err && !fscrypt_has_encryption_key(dir));
+
+	fscrypt_prepare_dentry(dentry, is_nokey_name);
 
-	if (!err && !fscrypt_has_encryption_key(dir)) {
-		spin_lock(&dentry->d_lock);
-		dentry->d_flags |= DCACHE_NOKEY_NAME;
-		spin_unlock(&dentry->d_lock);
-	}
 	return err;
 }
 EXPORT_SYMBOL_GPL(fscrypt_prepare_lookup_partial);
diff --git a/fs/crypto/inline_crypt.c b/fs/crypto/inline_crypt.c
index b4002aea7c..40de69860d 100644
--- a/fs/crypto/inline_crypt.c
+++ b/fs/crypto/inline_crypt.c
@@ -284,7 +284,7 @@ static bool bh_get_inode_and_lblk_num(const struct buffer_head *bh,
 				      const struct inode **inode_ret,
 				      u64 *lblk_num_ret)
 {
-	struct page *page = bh->b_page;
+	struct folio *folio = bh->b_folio;
 	const struct address_space *mapping;
 	const struct inode *inode;
 
@@ -292,13 +292,13 @@ static bool bh_get_inode_and_lblk_num(const struct buffer_head *bh,
 	 * The ext4 journal (jbd2) can submit a buffer_head it directly created
 	 * for a non-pagecache page.  fscrypt doesn't care about these.
 	 */
-	mapping = page_mapping(page);
+	mapping = folio_mapping(folio);
 	if (!mapping)
 		return false;
 	inode = mapping->host;
 
 	*inode_ret = inode;
-	*lblk_num_ret = ((u64)page->index << (PAGE_SHIFT - inode->i_blkbits)) +
+	*lblk_num_ret = ((u64)folio->index << (PAGE_SHIFT - inode->i_blkbits)) +
 			(bh_offset(bh) >> inode->i_blkbits);
 	return true;
 }
diff --git a/fs/crypto/keyring.c b/fs/crypto/keyring.c
index 0edf0b58da..6681a71625 100644
--- a/fs/crypto/keyring.c
+++ b/fs/crypto/keyring.c
@@ -74,8 +74,12 @@ void fscrypt_put_master_key(struct fscrypt_master_key *mk)
 	 * that concurrent keyring lookups can no longer find it.
 	 */
 	WARN_ON_ONCE(refcount_read(&mk->mk_active_refs) != 0);
-	key_put(mk->mk_users);
-	mk->mk_users = NULL;
+	if (mk->mk_users) {
+		/* Clear the keyring so the quota gets released right away. */
+		keyring_clear(mk->mk_users);
+		key_put(mk->mk_users);
+		mk->mk_users = NULL;
+	}
 	call_rcu(&mk->mk_rcu_head, fscrypt_free_master_key);
 }
 
diff --git a/fs/crypto/keysetup.c b/fs/crypto/keysetup.c
index d71f7c799e..b4fe01ea4b 100644
--- a/fs/crypto/keysetup.c
+++ b/fs/crypto/keysetup.c
@@ -23,7 +23,7 @@ struct fscrypt_mode fscrypt_modes[] = {
 		.blk_crypto_mode = BLK_ENCRYPTION_MODE_AES_256_XTS,
 	},
 	[FSCRYPT_MODE_AES_256_CTS] = {
-		.friendly_name = "AES-256-CTS-CBC",
+		.friendly_name = "AES-256-CBC-CTS",
 		.cipher_str = "cts(cbc(aes))",
 		.keysize = 32,
 		.security_strength = 32,
@@ -38,7 +38,7 @@ struct fscrypt_mode fscrypt_modes[] = {
 		.blk_crypto_mode = BLK_ENCRYPTION_MODE_AES_128_CBC_ESSIV,
 	},
 	[FSCRYPT_MODE_AES_128_CTS] = {
-		.friendly_name = "AES-128-CTS-CBC",
+		.friendly_name = "AES-128-CBC-CTS",
 		.cipher_str = "cts(cbc(aes))",
 		.keysize = 16,
 		.security_strength = 16,
@@ -53,7 +53,7 @@ struct fscrypt_mode fscrypt_modes[] = {
 		.blk_crypto_mode = BLK_ENCRYPTION_MODE_SM4_XTS,
 	},
 	[FSCRYPT_MODE_SM4_CTS] = {
-		.friendly_name = "SM4-CTS-CBC",
+		.friendly_name = "SM4-CBC-CTS",
 		.cipher_str = "cts(cbc(sm4))",
 		.keysize = 16,
 		.security_strength = 16,
@@ -687,7 +687,7 @@ int fscrypt_get_encryption_info(struct inode *inode, bool allow_unsupported)
 /**
  * fscrypt_prepare_new_inode() - prepare to create a new inode in a directory
  * @dir: a possibly-encrypted directory
- * @inode: the new inode.  ->i_mode must be set already.
+ * @inode: the new inode.  ->i_mode and ->i_blkbits must be set already.
  *	   ->i_ino doesn't need to be set yet.
  * @encrypt_ret: (output) set to %true if the new inode will be encrypted
  *
@@ -717,6 +717,9 @@ int fscrypt_prepare_new_inode(struct inode *dir, struct inode *inode,
 	if (IS_ERR(policy))
 		return PTR_ERR(policy);
 
+	if (WARN_ON_ONCE(inode->i_blkbits == 0))
+		return -EINVAL;
+
 	if (WARN_ON_ONCE(inode->i_mode == 0))
 		return -EINVAL;
 
diff --git a/fs/dax.c b/fs/dax.c
index 423fc1607d..becb4a6920 100644
--- a/fs/dax.c
+++ b/fs/dax.c
@@ -1207,17 +1207,17 @@ static vm_fault_t dax_pmd_load_hole(struct xa_state *xas, struct vm_fault *vmf,
 	struct vm_area_struct *vma = vmf->vma;
 	struct inode *inode = mapping->host;
 	pgtable_t pgtable = NULL;
-	struct page *zero_page;
+	struct folio *zero_folio;
 	spinlock_t *ptl;
 	pmd_t pmd_entry;
 	pfn_t pfn;
 
-	zero_page = mm_get_huge_zero_page(vmf->vma->vm_mm);
+	zero_folio = mm_get_huge_zero_folio(vmf->vma->vm_mm);
 
-	if (unlikely(!zero_page))
+	if (unlikely(!zero_folio))
 		goto fallback;
 
-	pfn = page_to_pfn_t(zero_page);
+	pfn = page_to_pfn_t(&zero_folio->page);
 	*entry = dax_insert_entry(xas, vmf, iter, *entry, pfn,
 				  DAX_PMD | DAX_ZERO_PAGE);
 
@@ -1237,17 +1237,17 @@ static vm_fault_t dax_pmd_load_hole(struct xa_state *xas, struct vm_fault *vmf,
 		pgtable_trans_huge_deposit(vma->vm_mm, vmf->pmd, pgtable);
 		mm_inc_nr_ptes(vma->vm_mm);
 	}
-	pmd_entry = mk_pmd(zero_page, vmf->vma->vm_page_prot);
+	pmd_entry = mk_pmd(&zero_folio->page, vmf->vma->vm_page_prot);
 	pmd_entry = pmd_mkhuge(pmd_entry);
 	set_pmd_at(vmf->vma->vm_mm, pmd_addr, vmf->pmd, pmd_entry);
 	spin_unlock(ptl);
-	trace_dax_pmd_load_hole(inode, vmf, zero_page, *entry);
+	trace_dax_pmd_load_hole(inode, vmf, zero_folio, *entry);
 	return VM_FAULT_NOPAGE;
 
 fallback:
 	if (pgtable)
 		pte_free(vma->vm_mm, pgtable);
-	trace_dax_pmd_load_hole_fallback(inode, vmf, zero_page, *entry);
+	trace_dax_pmd_load_hole_fallback(inode, vmf, zero_folio, *entry);
 	return VM_FAULT_FALLBACK;
 }
 #else
diff --git a/fs/dcache.c b/fs/dcache.c
index 6ebccba333..4c144519aa 100644
--- a/fs/dcache.c
+++ b/fs/dcache.c
@@ -355,7 +355,11 @@ static inline void __d_clear_type_and_inode(struct dentry *dentry)
 	flags &= ~DCACHE_ENTRY_TYPE;
 	WRITE_ONCE(dentry->d_flags, flags);
 	dentry->d_inode = NULL;
-	if (dentry->d_flags & DCACHE_LRU_LIST)
+	/*
+	 * The negative counter only tracks dentries on the LRU. Don't inc if
+	 * d_lru is on another list.
+	 */
+	if ((flags & (DCACHE_LRU_LIST|DCACHE_SHRINK_LIST)) == DCACHE_LRU_LIST)
 		this_cpu_inc(nr_dentry_negative);
 }
 
@@ -1844,9 +1848,11 @@ static void __d_instantiate(struct dentry *dentry, struct inode *inode)
 
 	spin_lock(&dentry->d_lock);
 	/*
-	 * Decrement negative dentry count if it was in the LRU list.
+	 * The negative counter only tracks dentries on the LRU. Don't dec if
+	 * d_lru is on another list.
 	 */
-	if (dentry->d_flags & DCACHE_LRU_LIST)
+	if ((dentry->d_flags &
+	     (DCACHE_LRU_LIST|DCACHE_SHRINK_LIST)) == DCACHE_LRU_LIST)
 		this_cpu_dec(nr_dentry_negative);
 	hlist_add_head(&dentry->d_u.d_alias, &inode->i_dentry);
 	raw_write_seqcount_begin(&dentry->d_seq);
@@ -3029,28 +3035,25 @@ EXPORT_SYMBOL(d_splice_alias);
   
 bool is_subdir(struct dentry *new_dentry, struct dentry *old_dentry)
 {
-	bool result;
+	bool subdir;
 	unsigned seq;
 
 	if (new_dentry == old_dentry)
 		return true;
 
-	do {
-		/* for restarting inner loop in case of seq retry */
-		seq = read_seqbegin(&rename_lock);
-		/*
-		 * Need rcu_readlock to protect against the d_parent trashing
-		 * due to d_move
-		 */
-		rcu_read_lock();
-		if (d_ancestor(old_dentry, new_dentry))
-			result = true;
-		else
-			result = false;
-		rcu_read_unlock();
-	} while (read_seqretry(&rename_lock, seq));
-
-	return result;
+	/* Access d_parent under rcu as d_move() may change it. */
+	rcu_read_lock();
+	seq = read_seqbegin(&rename_lock);
+	subdir = d_ancestor(old_dentry, new_dentry);
+	 /* Try lockless once... */
+	if (read_seqretry(&rename_lock, seq)) {
+		/* ...else acquire lock for progress even on deep chains. */
+		read_seqlock_excl(&rename_lock);
+		subdir = d_ancestor(old_dentry, new_dentry);
+		read_sequnlock_excl(&rename_lock);
+	}
+	rcu_read_unlock();
+	return subdir;
 }
 EXPORT_SYMBOL(is_subdir);
 
@@ -3139,7 +3142,7 @@ static void __init dcache_init(void)
 	 * of the dcache.
 	 */
 	dentry_cache = KMEM_CACHE_USERCOPY(dentry,
-		SLAB_RECLAIM_ACCOUNT|SLAB_PANIC|SLAB_MEM_SPREAD|SLAB_ACCOUNT,
+		SLAB_RECLAIM_ACCOUNT|SLAB_PANIC|SLAB_ACCOUNT,
 		d_iname);
 
 	/* Hash may have been set up in dcache_init_early */
diff --git a/fs/debugfs/inode.c b/fs/debugfs/inode.c
index a40da00654..8fd928899a 100644
--- a/fs/debugfs/inode.c
+++ b/fs/debugfs/inode.c
@@ -14,7 +14,8 @@
 
 #include <linux/module.h>
 #include <linux/fs.h>
-#include <linux/mount.h>
+#include <linux/fs_context.h>
+#include <linux/fs_parser.h>
 #include <linux/pagemap.h>
 #include <linux/init.h>
 #include <linux/kobject.h>
@@ -23,7 +24,6 @@
 #include <linux/fsnotify.h>
 #include <linux/string.h>
 #include <linux/seq_file.h>
-#include <linux/parser.h>
 #include <linux/magic.h>
 #include <linux/slab.h>
 #include <linux/security.h>
@@ -77,7 +77,7 @@ static struct inode *debugfs_get_inode(struct super_block *sb)
 	return inode;
 }
 
-struct debugfs_mount_opts {
+struct debugfs_fs_info {
 	kuid_t uid;
 	kgid_t gid;
 	umode_t mode;
@@ -89,68 +89,59 @@ enum {
 	Opt_uid,
 	Opt_gid,
 	Opt_mode,
-	Opt_err
 };
 
-static const match_table_t tokens = {
-	{Opt_uid, "uid=%u"},
-	{Opt_gid, "gid=%u"},
-	{Opt_mode, "mode=%o"},
-	{Opt_err, NULL}
+static const struct fs_parameter_spec debugfs_param_specs[] = {
+	fsparam_u32	("gid",		Opt_gid),
+	fsparam_u32oct	("mode",	Opt_mode),
+	fsparam_u32	("uid",		Opt_uid),
+	{}
 };
 
-struct debugfs_fs_info {
-	struct debugfs_mount_opts mount_opts;
-};
-
-static int debugfs_parse_options(char *data, struct debugfs_mount_opts *opts)
+static int debugfs_parse_param(struct fs_context *fc, struct fs_parameter *param)
 {
-	substring_t args[MAX_OPT_ARGS];
-	int option;
-	int token;
+	struct debugfs_fs_info *opts = fc->s_fs_info;
+	struct fs_parse_result result;
 	kuid_t uid;
 	kgid_t gid;
-	char *p;
-
-	opts->opts = 0;
-	opts->mode = DEBUGFS_DEFAULT_MODE;
-
-	while ((p = strsep(&data, ",")) != NULL) {
-		if (!*p)
-			continue;
-
-		token = match_token(p, tokens, args);
-		switch (token) {
-		case Opt_uid:
-			if (match_int(&args[0], &option))
-				return -EINVAL;
-			uid = make_kuid(current_user_ns(), option);
-			if (!uid_valid(uid))
-				return -EINVAL;
-			opts->uid = uid;
-			break;
-		case Opt_gid:
-			if (match_int(&args[0], &option))
-				return -EINVAL;
-			gid = make_kgid(current_user_ns(), option);
-			if (!gid_valid(gid))
-				return -EINVAL;
-			opts->gid = gid;
-			break;
-		case Opt_mode:
-			if (match_octal(&args[0], &option))
-				return -EINVAL;
-			opts->mode = option & S_IALLUGO;
-			break;
+	int opt;
+
+	opt = fs_parse(fc, debugfs_param_specs, param, &result);
+	if (opt < 0) {
 		/*
-		 * We might like to report bad mount options here;
-		 * but traditionally debugfs has ignored all mount options
-		 */
-		}
+                * We might like to report bad mount options here; but
+                * traditionally debugfs has ignored all mount options
+                */
+		if (opt == -ENOPARAM)
+			return 0;
 
-		opts->opts |= BIT(token);
+		return opt;
+	}
+
+	switch (opt) {
+	case Opt_uid:
+		uid = make_kuid(current_user_ns(), result.uint_32);
+		if (!uid_valid(uid))
+			return invalf(fc, "Unknown uid");
+		opts->uid = uid;
+		break;
+	case Opt_gid:
+		gid = make_kgid(current_user_ns(), result.uint_32);
+		if (!gid_valid(gid))
+			return invalf(fc, "Unknown gid");
+		opts->gid = gid;
+		break;
+	case Opt_mode:
+		opts->mode = result.uint_32 & S_IALLUGO;
+		break;
+	/*
+	 * We might like to report bad mount options here;
+	 * but traditionally debugfs has ignored all mount options
+	 */
 	}
 
+	opts->opts |= BIT(opt);
+
 	return 0;
 }
 
@@ -158,23 +149,22 @@ static void _debugfs_apply_options(struct super_block *sb, bool remount)
 {
 	struct debugfs_fs_info *fsi = sb->s_fs_info;
 	struct inode *inode = d_inode(sb->s_root);
-	struct debugfs_mount_opts *opts = &fsi->mount_opts;
 
 	/*
 	 * On remount, only reset mode/uid/gid if they were provided as mount
 	 * options.
 	 */
 
-	if (!remount || opts->opts & BIT(Opt_mode)) {
+	if (!remount || fsi->opts & BIT(Opt_mode)) {
 		inode->i_mode &= ~S_IALLUGO;
-		inode->i_mode |= opts->mode;
+		inode->i_mode |= fsi->mode;
 	}
 
-	if (!remount || opts->opts & BIT(Opt_uid))
-		inode->i_uid = opts->uid;
+	if (!remount || fsi->opts & BIT(Opt_uid))
+		inode->i_uid = fsi->uid;
 
-	if (!remount || opts->opts & BIT(Opt_gid))
-		inode->i_gid = opts->gid;
+	if (!remount || fsi->opts & BIT(Opt_gid))
+		inode->i_gid = fsi->gid;
 }
 
 static void debugfs_apply_options(struct super_block *sb)
@@ -187,35 +177,33 @@ static void debugfs_apply_options_remount(struct super_block *sb)
 	_debugfs_apply_options(sb, true);
 }
 
-static int debugfs_remount(struct super_block *sb, int *flags, char *data)
+static int debugfs_reconfigure(struct fs_context *fc)
 {
-	int err;
-	struct debugfs_fs_info *fsi = sb->s_fs_info;
+	struct super_block *sb = fc->root->d_sb;
+	struct debugfs_fs_info *sb_opts = sb->s_fs_info;
+	struct debugfs_fs_info *new_opts = fc->s_fs_info;
 
 	sync_filesystem(sb);
-	err = debugfs_parse_options(data, &fsi->mount_opts);
-	if (err)
-		goto fail;
 
+	/* structure copy of new mount options to sb */
+	*sb_opts = *new_opts;
 	debugfs_apply_options_remount(sb);
 
-fail:
-	return err;
+	return 0;
 }
 
 static int debugfs_show_options(struct seq_file *m, struct dentry *root)
 {
 	struct debugfs_fs_info *fsi = root->d_sb->s_fs_info;
-	struct debugfs_mount_opts *opts = &fsi->mount_opts;
 
-	if (!uid_eq(opts->uid, GLOBAL_ROOT_UID))
+	if (!uid_eq(fsi->uid, GLOBAL_ROOT_UID))
 		seq_printf(m, ",uid=%u",
-			   from_kuid_munged(&init_user_ns, opts->uid));
-	if (!gid_eq(opts->gid, GLOBAL_ROOT_GID))
+			   from_kuid_munged(&init_user_ns, fsi->uid));
+	if (!gid_eq(fsi->gid, GLOBAL_ROOT_GID))
 		seq_printf(m, ",gid=%u",
-			   from_kgid_munged(&init_user_ns, opts->gid));
-	if (opts->mode != DEBUGFS_DEFAULT_MODE)
-		seq_printf(m, ",mode=%o", opts->mode);
+			   from_kgid_munged(&init_user_ns, fsi->gid));
+	if (fsi->mode != DEBUGFS_DEFAULT_MODE)
+		seq_printf(m, ",mode=%o", fsi->mode);
 
 	return 0;
 }
@@ -229,7 +217,6 @@ static void debugfs_free_inode(struct inode *inode)
 
 static const struct super_operations debugfs_super_operations = {
 	.statfs		= simple_statfs,
-	.remount_fs	= debugfs_remount,
 	.show_options	= debugfs_show_options,
 	.free_inode	= debugfs_free_inode,
 };
@@ -263,26 +250,14 @@ static const struct dentry_operations debugfs_dops = {
 	.d_automount = debugfs_automount,
 };
 
-static int debug_fill_super(struct super_block *sb, void *data, int silent)
+static int debugfs_fill_super(struct super_block *sb, struct fs_context *fc)
 {
 	static const struct tree_descr debug_files[] = {{""}};
-	struct debugfs_fs_info *fsi;
 	int err;
 
-	fsi = kzalloc(sizeof(struct debugfs_fs_info), GFP_KERNEL);
-	sb->s_fs_info = fsi;
-	if (!fsi) {
-		err = -ENOMEM;
-		goto fail;
-	}
-
-	err = debugfs_parse_options(data, &fsi->mount_opts);
+	err = simple_fill_super(sb, DEBUGFS_MAGIC, debug_files);
 	if (err)
-		goto fail;
-
-	err  =  simple_fill_super(sb, DEBUGFS_MAGIC, debug_files);
-	if (err)
-		goto fail;
+		return err;
 
 	sb->s_op = &debugfs_super_operations;
 	sb->s_d_op = &debugfs_dops;
@@ -290,27 +265,48 @@ static int debug_fill_super(struct super_block *sb, void *data, int silent)
 	debugfs_apply_options(sb);
 
 	return 0;
-
-fail:
-	kfree(fsi);
-	sb->s_fs_info = NULL;
-	return err;
 }
 
-static struct dentry *debug_mount(struct file_system_type *fs_type,
-			int flags, const char *dev_name,
-			void *data)
+static int debugfs_get_tree(struct fs_context *fc)
 {
 	if (!(debugfs_allow & DEBUGFS_ALLOW_API))
-		return ERR_PTR(-EPERM);
+		return -EPERM;
+
+	return get_tree_single(fc, debugfs_fill_super);
+}
+
+static void debugfs_free_fc(struct fs_context *fc)
+{
+	kfree(fc->s_fs_info);
+}
 
-	return mount_single(fs_type, flags, data, debug_fill_super);
+static const struct fs_context_operations debugfs_context_ops = {
+	.free		= debugfs_free_fc,
+	.parse_param	= debugfs_parse_param,
+	.get_tree	= debugfs_get_tree,
+	.reconfigure	= debugfs_reconfigure,
+};
+
+static int debugfs_init_fs_context(struct fs_context *fc)
+{
+	struct debugfs_fs_info *fsi;
+
+	fsi = kzalloc(sizeof(struct debugfs_fs_info), GFP_KERNEL);
+	if (!fsi)
+		return -ENOMEM;
+
+	fsi->mode = DEBUGFS_DEFAULT_MODE;
+
+	fc->s_fs_info = fsi;
+	fc->ops = &debugfs_context_ops;
+	return 0;
 }
 
 static struct file_system_type debug_fs_type = {
 	.owner =	THIS_MODULE,
 	.name =		"debugfs",
-	.mount =	debug_mount,
+	.init_fs_context = debugfs_init_fs_context,
+	.parameters =	debugfs_param_specs,
 	.kill_sb =	kill_litter_super,
 };
 MODULE_ALIAS_FS("debugfs");
diff --git a/fs/direct-io.c b/fs/direct-io.c
index 60456263a3..b0aafe640f 100644
--- a/fs/direct-io.c
+++ b/fs/direct-io.c
@@ -410,6 +410,8 @@ dio_bio_alloc(struct dio *dio, struct dio_submit *sdio,
 		bio->bi_end_io = dio_bio_end_io;
 	if (dio->is_pinned)
 		bio_set_flag(bio, BIO_PAGE_PINNED);
+	bio->bi_write_hint = file_inode(dio->iocb->ki_filp)->i_write_hint;
+
 	sdio->bio = bio;
 	sdio->logical_offset_in_bio = sdio->cur_page_fs_offset;
 }
@@ -1215,7 +1217,6 @@ ssize_t __blockdev_direct_IO(struct kiocb *iocb, struct inode *inode,
 	 */
 	inode_dio_begin(inode);
 
-	retval = 0;
 	sdio.blkbits = blkbits;
 	sdio.blkfactor = i_blkbits - blkbits;
 	sdio.block_in_file = offset >> blkbits;
diff --git a/fs/dlm/ast.c b/fs/dlm/ast.c
index decedc4ee1..59711486d8 100644
--- a/fs/dlm/ast.c
+++ b/fs/dlm/ast.c
@@ -18,31 +18,32 @@
 #include "user.h"
 #include "ast.h"
 
-void dlm_release_callback(struct kref *ref)
+static void dlm_callback_work(struct work_struct *work)
 {
-	struct dlm_callback *cb = container_of(ref, struct dlm_callback, ref);
+	struct dlm_callback *cb = container_of(work, struct dlm_callback, work);
+
+	if (cb->flags & DLM_CB_BAST) {
+		trace_dlm_bast(cb->ls_id, cb->lkb_id, cb->mode, cb->res_name,
+			       cb->res_length);
+		cb->bastfn(cb->astparam, cb->mode);
+	} else if (cb->flags & DLM_CB_CAST) {
+		trace_dlm_ast(cb->ls_id, cb->lkb_id, cb->sb_status,
+			      cb->sb_flags, cb->res_name, cb->res_length);
+		cb->lkb_lksb->sb_status = cb->sb_status;
+		cb->lkb_lksb->sb_flags = cb->sb_flags;
+		cb->astfn(cb->astparam);
+	}
 
 	dlm_free_cb(cb);
 }
 
-void dlm_callback_set_last_ptr(struct dlm_callback **from,
-			       struct dlm_callback *to)
-{
-	if (*from)
-		kref_put(&(*from)->ref, dlm_release_callback);
-
-	if (to)
-		kref_get(&to->ref);
-
-	*from = to;
-}
-
-int dlm_enqueue_lkb_callback(struct dlm_lkb *lkb, uint32_t flags, int mode,
-			     int status, uint32_t sbflags)
+int dlm_queue_lkb_callback(struct dlm_lkb *lkb, uint32_t flags, int mode,
+			   int status, uint32_t sbflags,
+			   struct dlm_callback **cb)
 {
-	struct dlm_ls *ls = lkb->lkb_resource->res_ls;
+	struct dlm_rsb *rsb = lkb->lkb_resource;
 	int rv = DLM_ENQUEUE_CALLBACK_SUCCESS;
-	struct dlm_callback *cb;
+	struct dlm_ls *ls = rsb->res_ls;
 	int copy_lvb = 0;
 	int prev_mode;
 
@@ -50,11 +51,11 @@ int dlm_enqueue_lkb_callback(struct dlm_lkb *lkb, uint32_t flags, int mode,
 		/* if cb is a bast, it should be skipped if the blocking mode is
 		 * compatible with the last granted mode
 		 */
-		if (lkb->lkb_last_cast) {
-			if (dlm_modes_compat(mode, lkb->lkb_last_cast->mode)) {
+		if (lkb->lkb_last_cast_cb_mode != -1) {
+			if (dlm_modes_compat(mode, lkb->lkb_last_cast_cb_mode)) {
 				log_debug(ls, "skip %x bast mode %d for cast mode %d",
 					  lkb->lkb_id, mode,
-					  lkb->lkb_last_cast->mode);
+					  lkb->lkb_last_cast_cb_mode);
 				goto out;
 			}
 		}
@@ -65,8 +66,9 @@ int dlm_enqueue_lkb_callback(struct dlm_lkb *lkb, uint32_t flags, int mode,
 		 * is a bast for the same mode or a more restrictive mode.
 		 * (the addional > PR check is needed for PR/CW inversion)
 		 */
-		if (lkb->lkb_last_cb && lkb->lkb_last_cb->flags & DLM_CB_BAST) {
-			prev_mode = lkb->lkb_last_cb->mode;
+		if (lkb->lkb_last_cb_mode != -1 &&
+		    lkb->lkb_last_cb_flags & DLM_CB_BAST) {
+			prev_mode = lkb->lkb_last_cb_mode;
 
 			if ((prev_mode == mode) ||
 			    (prev_mode > mode && prev_mode > DLM_LOCK_PR)) {
@@ -75,65 +77,55 @@ int dlm_enqueue_lkb_callback(struct dlm_lkb *lkb, uint32_t flags, int mode,
 				goto out;
 			}
 		}
+
+		lkb->lkb_last_bast_time = ktime_get();
+		lkb->lkb_last_bast_cb_mode = mode;
 	} else if (flags & DLM_CB_CAST) {
 		if (test_bit(DLM_DFL_USER_BIT, &lkb->lkb_dflags)) {
-			if (lkb->lkb_last_cast)
-				prev_mode = lkb->lkb_last_cb->mode;
-			else
-				prev_mode = -1;
+			prev_mode = lkb->lkb_last_cast_cb_mode;
 
 			if (!status && lkb->lkb_lksb->sb_lvbptr &&
 			    dlm_lvb_operations[prev_mode + 1][mode + 1])
 				copy_lvb = 1;
 		}
+
+		lkb->lkb_last_cast_cb_mode = mode;
+		lkb->lkb_last_cast_time = ktime_get();
 	}
 
-	cb = dlm_allocate_cb();
-	if (!cb) {
+	lkb->lkb_last_cb_mode = mode;
+	lkb->lkb_last_cb_flags = flags;
+
+	*cb = dlm_allocate_cb();
+	if (!*cb) {
 		rv = DLM_ENQUEUE_CALLBACK_FAILURE;
 		goto out;
 	}
 
-	cb->flags = flags;
-	cb->mode = mode;
-	cb->sb_status = status;
-	cb->sb_flags = (sbflags & 0x000000FF);
-	cb->copy_lvb = copy_lvb;
-	kref_init(&cb->ref);
-	if (!test_and_set_bit(DLM_IFL_CB_PENDING_BIT, &lkb->lkb_iflags))
-		rv = DLM_ENQUEUE_CALLBACK_NEED_SCHED;
-
-	list_add_tail(&cb->list, &lkb->lkb_callbacks);
+	/* for tracing */
+	(*cb)->lkb_id = lkb->lkb_id;
+	(*cb)->ls_id = ls->ls_global_id;
+	memcpy((*cb)->res_name, rsb->res_name, rsb->res_length);
+	(*cb)->res_length = rsb->res_length;
 
-	if (flags & DLM_CB_CAST)
-		dlm_callback_set_last_ptr(&lkb->lkb_last_cast, cb);
+	(*cb)->flags = flags;
+	(*cb)->mode = mode;
+	(*cb)->sb_status = status;
+	(*cb)->sb_flags = (sbflags & 0x000000FF);
+	(*cb)->copy_lvb = copy_lvb;
+	(*cb)->lkb_lksb = lkb->lkb_lksb;
 
-	dlm_callback_set_last_ptr(&lkb->lkb_last_cb, cb);
+	rv = DLM_ENQUEUE_CALLBACK_NEED_SCHED;
 
- out:
+out:
 	return rv;
 }
 
-int dlm_dequeue_lkb_callback(struct dlm_lkb *lkb, struct dlm_callback **cb)
-{
-	/* oldest undelivered cb is callbacks first entry */
-	*cb = list_first_entry_or_null(&lkb->lkb_callbacks,
-				       struct dlm_callback, list);
-	if (!*cb)
-		return DLM_DEQUEUE_CALLBACK_EMPTY;
-
-	/* remove it from callbacks so shift others down */
-	list_del(&(*cb)->list);
-	if (list_empty(&lkb->lkb_callbacks))
-		return DLM_DEQUEUE_CALLBACK_LAST;
-
-	return DLM_DEQUEUE_CALLBACK_SUCCESS;
-}
-
 void dlm_add_cb(struct dlm_lkb *lkb, uint32_t flags, int mode, int status,
-		uint32_t sbflags)
+		  uint32_t sbflags)
 {
 	struct dlm_ls *ls = lkb->lkb_resource->res_ls;
+	struct dlm_callback *cb;
 	int rv;
 
 	if (test_bit(DLM_DFL_USER_BIT, &lkb->lkb_dflags)) {
@@ -141,88 +133,36 @@ void dlm_add_cb(struct dlm_lkb *lkb, uint32_t flags, int mode, int status,
 		return;
 	}
 
-	spin_lock(&lkb->lkb_cb_lock);
-	rv = dlm_enqueue_lkb_callback(lkb, flags, mode, status, sbflags);
+	rv = dlm_queue_lkb_callback(lkb, flags, mode, status, sbflags,
+				    &cb);
 	switch (rv) {
 	case DLM_ENQUEUE_CALLBACK_NEED_SCHED:
-		kref_get(&lkb->lkb_ref);
-
-		spin_lock(&ls->ls_cb_lock);
-		if (test_bit(LSFL_CB_DELAY, &ls->ls_flags)) {
-			list_add(&lkb->lkb_cb_list, &ls->ls_cb_delay);
-		} else {
-			queue_work(ls->ls_callback_wq, &lkb->lkb_cb_work);
-		}
-		spin_unlock(&ls->ls_cb_lock);
-		break;
-	case DLM_ENQUEUE_CALLBACK_FAILURE:
-		WARN_ON_ONCE(1);
+		cb->astfn = lkb->lkb_astfn;
+		cb->bastfn = lkb->lkb_bastfn;
+		cb->astparam = lkb->lkb_astparam;
+		INIT_WORK(&cb->work, dlm_callback_work);
+
+		spin_lock_bh(&ls->ls_cb_lock);
+		if (test_bit(LSFL_CB_DELAY, &ls->ls_flags))
+			list_add(&cb->list, &ls->ls_cb_delay);
+		else
+			queue_work(ls->ls_callback_wq, &cb->work);
+		spin_unlock_bh(&ls->ls_cb_lock);
 		break;
 	case DLM_ENQUEUE_CALLBACK_SUCCESS:
 		break;
+	case DLM_ENQUEUE_CALLBACK_FAILURE:
+		fallthrough;
 	default:
 		WARN_ON_ONCE(1);
 		break;
 	}
-	spin_unlock(&lkb->lkb_cb_lock);
-}
-
-void dlm_callback_work(struct work_struct *work)
-{
-	struct dlm_lkb *lkb = container_of(work, struct dlm_lkb, lkb_cb_work);
-	struct dlm_ls *ls = lkb->lkb_resource->res_ls;
-	void (*castfn) (void *astparam);
-	void (*bastfn) (void *astparam, int mode);
-	struct dlm_callback *cb;
-	int rv;
-
-	spin_lock(&lkb->lkb_cb_lock);
-	rv = dlm_dequeue_lkb_callback(lkb, &cb);
-	if (WARN_ON_ONCE(rv == DLM_DEQUEUE_CALLBACK_EMPTY)) {
-		clear_bit(DLM_IFL_CB_PENDING_BIT, &lkb->lkb_iflags);
-		spin_unlock(&lkb->lkb_cb_lock);
-		goto out;
-	}
-	spin_unlock(&lkb->lkb_cb_lock);
-
-	for (;;) {
-		castfn = lkb->lkb_astfn;
-		bastfn = lkb->lkb_bastfn;
-
-		if (cb->flags & DLM_CB_BAST) {
-			trace_dlm_bast(ls, lkb, cb->mode);
-			lkb->lkb_last_bast_time = ktime_get();
-			lkb->lkb_last_bast_mode = cb->mode;
-			bastfn(lkb->lkb_astparam, cb->mode);
-		} else if (cb->flags & DLM_CB_CAST) {
-			lkb->lkb_lksb->sb_status = cb->sb_status;
-			lkb->lkb_lksb->sb_flags = cb->sb_flags;
-			trace_dlm_ast(ls, lkb);
-			lkb->lkb_last_cast_time = ktime_get();
-			castfn(lkb->lkb_astparam);
-		}
-
-		kref_put(&cb->ref, dlm_release_callback);
-
-		spin_lock(&lkb->lkb_cb_lock);
-		rv = dlm_dequeue_lkb_callback(lkb, &cb);
-		if (rv == DLM_DEQUEUE_CALLBACK_EMPTY) {
-			clear_bit(DLM_IFL_CB_PENDING_BIT, &lkb->lkb_iflags);
-			spin_unlock(&lkb->lkb_cb_lock);
-			break;
-		}
-		spin_unlock(&lkb->lkb_cb_lock);
-	}
-
-out:
-	/* undo kref_get from dlm_add_callback, may cause lkb to be freed */
-	dlm_put_lkb(lkb);
 }
 
 int dlm_callback_start(struct dlm_ls *ls)
 {
-	ls->ls_callback_wq = alloc_workqueue("dlm_callback",
-					     WQ_HIGHPRI | WQ_MEM_RECLAIM, 0);
+	ls->ls_callback_wq = alloc_ordered_workqueue("dlm_callback",
+						     WQ_HIGHPRI | WQ_MEM_RECLAIM);
 	if (!ls->ls_callback_wq) {
 		log_print("can't start dlm_callback workqueue");
 		return -ENOMEM;
@@ -239,9 +179,9 @@ void dlm_callback_stop(struct dlm_ls *ls)
 void dlm_callback_suspend(struct dlm_ls *ls)
 {
 	if (ls->ls_callback_wq) {
-		spin_lock(&ls->ls_cb_lock);
+		spin_lock_bh(&ls->ls_cb_lock);
 		set_bit(LSFL_CB_DELAY, &ls->ls_flags);
-		spin_unlock(&ls->ls_cb_lock);
+		spin_unlock_bh(&ls->ls_cb_lock);
 
 		flush_workqueue(ls->ls_callback_wq);
 	}
@@ -251,7 +191,7 @@ void dlm_callback_suspend(struct dlm_ls *ls)
 
 void dlm_callback_resume(struct dlm_ls *ls)
 {
-	struct dlm_lkb *lkb, *safe;
+	struct dlm_callback *cb, *safe;
 	int count = 0, sum = 0;
 	bool empty;
 
@@ -259,10 +199,10 @@ void dlm_callback_resume(struct dlm_ls *ls)
 		return;
 
 more:
-	spin_lock(&ls->ls_cb_lock);
-	list_for_each_entry_safe(lkb, safe, &ls->ls_cb_delay, lkb_cb_list) {
-		list_del_init(&lkb->lkb_cb_list);
-		queue_work(ls->ls_callback_wq, &lkb->lkb_cb_work);
+	spin_lock_bh(&ls->ls_cb_lock);
+	list_for_each_entry_safe(cb, safe, &ls->ls_cb_delay, list) {
+		list_del(&cb->list);
+		queue_work(ls->ls_callback_wq, &cb->work);
 		count++;
 		if (count == MAX_CB_QUEUE)
 			break;
@@ -270,7 +210,7 @@ more:
 	empty = list_empty(&ls->ls_cb_delay);
 	if (empty)
 		clear_bit(LSFL_CB_DELAY, &ls->ls_flags);
-	spin_unlock(&ls->ls_cb_lock);
+	spin_unlock_bh(&ls->ls_cb_lock);
 
 	sum += count;
 	if (!empty) {
diff --git a/fs/dlm/ast.h b/fs/dlm/ast.h
index ce007892dc..9093ff043b 100644
--- a/fs/dlm/ast.h
+++ b/fs/dlm/ast.h
@@ -14,19 +14,12 @@
 #define DLM_ENQUEUE_CALLBACK_NEED_SCHED	1
 #define DLM_ENQUEUE_CALLBACK_SUCCESS	0
 #define DLM_ENQUEUE_CALLBACK_FAILURE	-1
-int dlm_enqueue_lkb_callback(struct dlm_lkb *lkb, uint32_t flags, int mode,
-			     int status, uint32_t sbflags);
-#define DLM_DEQUEUE_CALLBACK_EMPTY	2
-#define DLM_DEQUEUE_CALLBACK_LAST	1
-#define DLM_DEQUEUE_CALLBACK_SUCCESS	0
-int dlm_dequeue_lkb_callback(struct dlm_lkb *lkb, struct dlm_callback **cb);
+int dlm_queue_lkb_callback(struct dlm_lkb *lkb, uint32_t flags, int mode,
+			   int status, uint32_t sbflags,
+			   struct dlm_callback **cb);
 void dlm_add_cb(struct dlm_lkb *lkb, uint32_t flags, int mode, int status,
                 uint32_t sbflags);
-void dlm_callback_set_last_ptr(struct dlm_callback **from,
-			       struct dlm_callback *to);
 
-void dlm_release_callback(struct kref *ref);
-void dlm_callback_work(struct work_struct *work);
 int dlm_callback_start(struct dlm_ls *ls);
 void dlm_callback_stop(struct dlm_ls *ls);
 void dlm_callback_suspend(struct dlm_ls *ls);
diff --git a/fs/dlm/config.c b/fs/dlm/config.c
index e55e0a2cd2..517fa975dc 100644
--- a/fs/dlm/config.c
+++ b/fs/dlm/config.c
@@ -63,6 +63,14 @@ static void release_node(struct config_item *);
 static struct configfs_attribute *comm_attrs[];
 static struct configfs_attribute *node_attrs[];
 
+const struct rhashtable_params dlm_rhash_rsb_params = {
+	.nelem_hint = 3, /* start small */
+	.key_len = DLM_RESNAME_MAXLEN,
+	.key_offset = offsetof(struct dlm_rsb, res_name),
+	.head_offset = offsetof(struct dlm_rsb, res_node),
+	.automatic_shrinking = true,
+};
+
 struct dlm_cluster {
 	struct config_group group;
 	unsigned int cl_tcp_port;
diff --git a/fs/dlm/config.h b/fs/dlm/config.h
index 4c91fcca0f..ed237d9102 100644
--- a/fs/dlm/config.h
+++ b/fs/dlm/config.h
@@ -21,6 +21,8 @@ struct dlm_config_node {
 	uint32_t comm_seq;
 };
 
+extern const struct rhashtable_params dlm_rhash_rsb_params;
+
 #define DLM_MAX_ADDR_COUNT 3
 
 #define DLM_PROTO_TCP	0
diff --git a/fs/dlm/debug_fs.c b/fs/dlm/debug_fs.c
index 4fa11d9ddb..6ab3ed4074 100644
--- a/fs/dlm/debug_fs.c
+++ b/fs/dlm/debug_fs.c
@@ -247,7 +247,7 @@ static void print_format3_lock(struct seq_file *s, struct dlm_lkb *lkb,
 		   lkb->lkb_status,
 		   lkb->lkb_grmode,
 		   lkb->lkb_rqmode,
-		   lkb->lkb_last_bast_mode,
+		   lkb->lkb_last_bast_cb_mode,
 		   rsb_lookup,
 		   lkb->lkb_wait_type,
 		   lkb->lkb_lvbseq,
@@ -366,58 +366,10 @@ static void print_format4(struct dlm_rsb *r, struct seq_file *s)
 	unlock_rsb(r);
 }
 
-static void print_format5_lock(struct seq_file *s, struct dlm_lkb *lkb)
-{
-	struct dlm_callback *cb;
-
-	/* lkb_id lkb_flags mode flags sb_status sb_flags */
-
-	spin_lock(&lkb->lkb_cb_lock);
-	list_for_each_entry(cb, &lkb->lkb_callbacks, list) {
-		seq_printf(s, "%x %x %d %x %d %x\n",
-			   lkb->lkb_id,
-			   dlm_iflags_val(lkb),
-			   cb->mode,
-			   cb->flags,
-			   cb->sb_status,
-			   cb->sb_flags);
-	}
-	spin_unlock(&lkb->lkb_cb_lock);
-}
-
-static void print_format5(struct dlm_rsb *r, struct seq_file *s)
-{
-	struct dlm_lkb *lkb;
-
-	lock_rsb(r);
-
-	list_for_each_entry(lkb, &r->res_grantqueue, lkb_statequeue) {
-		print_format5_lock(s, lkb);
-		if (seq_has_overflowed(s))
-			goto out;
-	}
-
-	list_for_each_entry(lkb, &r->res_convertqueue, lkb_statequeue) {
-		print_format5_lock(s, lkb);
-		if (seq_has_overflowed(s))
-			goto out;
-	}
-
-	list_for_each_entry(lkb, &r->res_waitqueue, lkb_statequeue) {
-		print_format5_lock(s, lkb);
-		if (seq_has_overflowed(s))
-			goto out;
-	}
- out:
-	unlock_rsb(r);
-}
-
-struct rsbtbl_iter {
-	struct dlm_rsb *rsb;
-	unsigned bucket;
-	int format;
-	int header;
-};
+static const struct seq_operations format1_seq_ops;
+static const struct seq_operations format2_seq_ops;
+static const struct seq_operations format3_seq_ops;
+static const struct seq_operations format4_seq_ops;
 
 /*
  * If the buffer is full, seq_printf can be called again, but it
@@ -428,207 +380,61 @@ struct rsbtbl_iter {
 
 static int table_seq_show(struct seq_file *seq, void *iter_ptr)
 {
-	struct rsbtbl_iter *ri = iter_ptr;
-
-	switch (ri->format) {
-	case 1:
-		print_format1(ri->rsb, seq);
-		break;
-	case 2:
-		if (ri->header) {
-			seq_puts(seq, "id nodeid remid pid xid exflags flags sts grmode rqmode time_ms r_nodeid r_len r_name\n");
-			ri->header = 0;
-		}
-		print_format2(ri->rsb, seq);
-		break;
-	case 3:
-		if (ri->header) {
-			seq_puts(seq, "rsb ptr nodeid first_lkid flags !root_list_empty !recover_list_empty recover_locks_count len\n");
-			ri->header = 0;
-		}
-		print_format3(ri->rsb, seq);
-		break;
-	case 4:
-		if (ri->header) {
-			seq_puts(seq, "rsb ptr nodeid master_nodeid dir_nodeid our_nodeid toss_time flags len str|hex name\n");
-			ri->header = 0;
-		}
-		print_format4(ri->rsb, seq);
-		break;
-	case 5:
-		if (ri->header) {
-			seq_puts(seq, "lkb_id lkb_flags mode flags sb_status sb_flags\n");
-			ri->header = 0;
-		}
-		print_format5(ri->rsb, seq);
-		break;
-	}
+	struct dlm_rsb *rsb = list_entry(iter_ptr, struct dlm_rsb, res_rsbs_list);
+
+	if (seq->op == &format1_seq_ops)
+		print_format1(rsb, seq);
+	else if (seq->op == &format2_seq_ops)
+		print_format2(rsb, seq);
+	else if (seq->op == &format3_seq_ops)
+		print_format3(rsb, seq);
+	else if (seq->op == &format4_seq_ops)
+		print_format4(rsb, seq);
 
 	return 0;
 }
 
-static const struct seq_operations format1_seq_ops;
-static const struct seq_operations format2_seq_ops;
-static const struct seq_operations format3_seq_ops;
-static const struct seq_operations format4_seq_ops;
-static const struct seq_operations format5_seq_ops;
-
 static void *table_seq_start(struct seq_file *seq, loff_t *pos)
 {
-	struct rb_root *tree;
-	struct rb_node *node;
 	struct dlm_ls *ls = seq->private;
-	struct rsbtbl_iter *ri;
-	struct dlm_rsb *r;
-	loff_t n = *pos;
-	unsigned bucket, entry;
-	int toss = (seq->op == &format4_seq_ops);
-
-	bucket = n >> 32;
-	entry = n & ((1LL << 32) - 1);
-
-	if (bucket >= ls->ls_rsbtbl_size)
-		return NULL;
-
-	ri = kzalloc(sizeof(*ri), GFP_NOFS);
-	if (!ri)
-		return NULL;
-	if (n == 0)
-		ri->header = 1;
-	if (seq->op == &format1_seq_ops)
-		ri->format = 1;
-	if (seq->op == &format2_seq_ops)
-		ri->format = 2;
-	if (seq->op == &format3_seq_ops)
-		ri->format = 3;
-	if (seq->op == &format4_seq_ops)
-		ri->format = 4;
-	if (seq->op == &format5_seq_ops)
-		ri->format = 5;
-
-	tree = toss ? &ls->ls_rsbtbl[bucket].toss : &ls->ls_rsbtbl[bucket].keep;
-
-	spin_lock(&ls->ls_rsbtbl[bucket].lock);
-	if (!RB_EMPTY_ROOT(tree)) {
-		for (node = rb_first(tree); node; node = rb_next(node)) {
-			r = rb_entry(node, struct dlm_rsb, res_hashnode);
-			if (!entry--) {
-				dlm_hold_rsb(r);
-				ri->rsb = r;
-				ri->bucket = bucket;
-				spin_unlock(&ls->ls_rsbtbl[bucket].lock);
-				return ri;
-			}
-		}
-	}
-	spin_unlock(&ls->ls_rsbtbl[bucket].lock);
-
-	/*
-	 * move to the first rsb in the next non-empty bucket
-	 */
+	struct list_head *list;
 
-	/* zero the entry */
-	n &= ~((1LL << 32) - 1);
+	if (!*pos) {
+		if (seq->op == &format2_seq_ops)
+			seq_puts(seq, "id nodeid remid pid xid exflags flags sts grmode rqmode time_ms r_nodeid r_len r_name\n");
+		else if (seq->op == &format3_seq_ops)
+			seq_puts(seq, "rsb ptr nodeid first_lkid flags !root_list_empty !recover_list_empty recover_locks_count len\n");
+		else if (seq->op == &format4_seq_ops)
+			seq_puts(seq, "rsb ptr nodeid master_nodeid dir_nodeid our_nodeid toss_time flags len str|hex name\n");
+	}
 
-	while (1) {
-		bucket++;
-		n += 1LL << 32;
+	if (seq->op == &format4_seq_ops)
+		list = &ls->ls_toss;
+	else
+		list = &ls->ls_keep;
 
-		if (bucket >= ls->ls_rsbtbl_size) {
-			kfree(ri);
-			return NULL;
-		}
-		tree = toss ? &ls->ls_rsbtbl[bucket].toss : &ls->ls_rsbtbl[bucket].keep;
-
-		spin_lock(&ls->ls_rsbtbl[bucket].lock);
-		if (!RB_EMPTY_ROOT(tree)) {
-			node = rb_first(tree);
-			r = rb_entry(node, struct dlm_rsb, res_hashnode);
-			dlm_hold_rsb(r);
-			ri->rsb = r;
-			ri->bucket = bucket;
-			spin_unlock(&ls->ls_rsbtbl[bucket].lock);
-			*pos = n;
-			return ri;
-		}
-		spin_unlock(&ls->ls_rsbtbl[bucket].lock);
-	}
+	read_lock_bh(&ls->ls_rsbtbl_lock);
+	return seq_list_start(list, *pos);
 }
 
 static void *table_seq_next(struct seq_file *seq, void *iter_ptr, loff_t *pos)
 {
 	struct dlm_ls *ls = seq->private;
-	struct rsbtbl_iter *ri = iter_ptr;
-	struct rb_root *tree;
-	struct rb_node *next;
-	struct dlm_rsb *r, *rp;
-	loff_t n = *pos;
-	unsigned bucket;
-	int toss = (seq->op == &format4_seq_ops);
-
-	bucket = n >> 32;
-
-	/*
-	 * move to the next rsb in the same bucket
-	 */
-
-	spin_lock(&ls->ls_rsbtbl[bucket].lock);
-	rp = ri->rsb;
-	next = rb_next(&rp->res_hashnode);
-
-	if (next) {
-		r = rb_entry(next, struct dlm_rsb, res_hashnode);
-		dlm_hold_rsb(r);
-		ri->rsb = r;
-		spin_unlock(&ls->ls_rsbtbl[bucket].lock);
-		dlm_put_rsb(rp);
-		++*pos;
-		return ri;
-	}
-	spin_unlock(&ls->ls_rsbtbl[bucket].lock);
-	dlm_put_rsb(rp);
+	struct list_head *list;
 
-	/*
-	 * move to the first rsb in the next non-empty bucket
-	 */
-
-	/* zero the entry */
-	n &= ~((1LL << 32) - 1);
-
-	while (1) {
-		bucket++;
-		n += 1LL << 32;
+	if (seq->op == &format4_seq_ops)
+		list = &ls->ls_toss;
+	else
+		list = &ls->ls_keep;
 
-		if (bucket >= ls->ls_rsbtbl_size) {
-			kfree(ri);
-			++*pos;
-			return NULL;
-		}
-		tree = toss ? &ls->ls_rsbtbl[bucket].toss : &ls->ls_rsbtbl[bucket].keep;
-
-		spin_lock(&ls->ls_rsbtbl[bucket].lock);
-		if (!RB_EMPTY_ROOT(tree)) {
-			next = rb_first(tree);
-			r = rb_entry(next, struct dlm_rsb, res_hashnode);
-			dlm_hold_rsb(r);
-			ri->rsb = r;
-			ri->bucket = bucket;
-			spin_unlock(&ls->ls_rsbtbl[bucket].lock);
-			*pos = n;
-			return ri;
-		}
-		spin_unlock(&ls->ls_rsbtbl[bucket].lock);
-	}
+	return seq_list_next(iter_ptr, list, pos);
 }
 
 static void table_seq_stop(struct seq_file *seq, void *iter_ptr)
 {
-	struct rsbtbl_iter *ri = iter_ptr;
+	struct dlm_ls *ls = seq->private;
 
-	if (ri) {
-		dlm_put_rsb(ri->rsb);
-		kfree(ri);
-	}
+	read_unlock_bh(&ls->ls_rsbtbl_lock);
 }
 
 static const struct seq_operations format1_seq_ops = {
@@ -659,18 +465,10 @@ static const struct seq_operations format4_seq_ops = {
 	.show  = table_seq_show,
 };
 
-static const struct seq_operations format5_seq_ops = {
-	.start = table_seq_start,
-	.next  = table_seq_next,
-	.stop  = table_seq_stop,
-	.show  = table_seq_show,
-};
-
 static const struct file_operations format1_fops;
 static const struct file_operations format2_fops;
 static const struct file_operations format3_fops;
 static const struct file_operations format4_fops;
-static const struct file_operations format5_fops;
 
 static int table_open1(struct inode *inode, struct file *file)
 {
@@ -757,20 +555,6 @@ static int table_open4(struct inode *inode, struct file *file)
 	return 0;
 }
 
-static int table_open5(struct inode *inode, struct file *file)
-{
-	struct seq_file *seq;
-	int ret;
-
-	ret = seq_open(file, &format5_seq_ops);
-	if (ret)
-		return ret;
-
-	seq = file->private_data;
-	seq->private = inode->i_private; /* the dlm_ls */
-	return 0;
-}
-
 static const struct file_operations format1_fops = {
 	.owner   = THIS_MODULE,
 	.open    = table_open1,
@@ -804,14 +588,6 @@ static const struct file_operations format4_fops = {
 	.release = seq_release
 };
 
-static const struct file_operations format5_fops = {
-	.owner   = THIS_MODULE,
-	.open    = table_open5,
-	.read    = seq_read,
-	.llseek  = seq_lseek,
-	.release = seq_release
-};
-
 /*
  * dump lkb's on the ls_waiters list
  */
@@ -823,7 +599,13 @@ static ssize_t waiters_read(struct file *file, char __user *userbuf,
 	size_t len = DLM_DEBUG_BUF_LEN, pos = 0, ret, rv;
 
 	mutex_lock(&debug_buf_lock);
-	mutex_lock(&ls->ls_waiters_mutex);
+	ret = dlm_lock_recovery_try(ls);
+	if (!ret) {
+		rv = -EAGAIN;
+		goto out;
+	}
+
+	spin_lock_bh(&ls->ls_waiters_lock);
 	memset(debug_buf, 0, sizeof(debug_buf));
 
 	list_for_each_entry(lkb, &ls->ls_waiters, lkb_wait_reply) {
@@ -834,9 +616,11 @@ static ssize_t waiters_read(struct file *file, char __user *userbuf,
 			break;
 		pos += ret;
 	}
-	mutex_unlock(&ls->ls_waiters_mutex);
+	spin_unlock_bh(&ls->ls_waiters_lock);
+	dlm_unlock_recovery(ls);
 
 	rv = simple_read_from_buffer(userbuf, count, ppos, debug_buf, pos);
+out:
 	mutex_unlock(&debug_buf_lock);
 	return rv;
 }
@@ -858,7 +642,12 @@ static ssize_t waiters_write(struct file *file, const char __user *user_buf,
 	if (n != 3)
 		return -EINVAL;
 
+	error = dlm_lock_recovery_try(ls);
+	if (!error)
+		return -EAGAIN;
+
 	error = dlm_debug_add_lkb_to_waiters(ls, lkb_id, mstype, to_nodeid);
+	dlm_unlock_recovery(ls);
 	if (error)
 		return error;
 
@@ -1021,16 +810,6 @@ void dlm_create_debug_file(struct dlm_ls *ls)
 							  dlm_root,
 							  ls,
 							  &waiters_fops);
-
-	/* format 5 */
-
-	snprintf(name, sizeof(name), "%s_queued_asts", ls->ls_name);
-
-	ls->ls_debug_queued_asts_dentry = debugfs_create_file(name,
-							      0644,
-							      dlm_root,
-							      ls,
-							      &format5_fops);
 }
 
 void __init dlm_register_debugfs(void)
diff --git a/fs/dlm/dir.c b/fs/dlm/dir.c
index f6acba4310..b1ab0adbd9 100644
--- a/fs/dlm/dir.c
+++ b/fs/dlm/dir.c
@@ -47,15 +47,13 @@ int dlm_dir_nodeid(struct dlm_rsb *r)
 	return r->res_dir_nodeid;
 }
 
-void dlm_recover_dir_nodeid(struct dlm_ls *ls)
+void dlm_recover_dir_nodeid(struct dlm_ls *ls, const struct list_head *root_list)
 {
 	struct dlm_rsb *r;
 
-	down_read(&ls->ls_root_sem);
-	list_for_each_entry(r, &ls->ls_root_list, res_root_list) {
+	list_for_each_entry(r, root_list, res_root_list) {
 		r->res_dir_nodeid = dlm_hash2nodeid(ls, r->res_hash);
 	}
-	up_read(&ls->ls_root_sem);
 }
 
 int dlm_recover_directory(struct dlm_ls *ls, uint64_t seq)
@@ -200,35 +198,98 @@ static struct dlm_rsb *find_rsb_root(struct dlm_ls *ls, const char *name,
 				     int len)
 {
 	struct dlm_rsb *r;
-	uint32_t hash, bucket;
 	int rv;
 
-	hash = jhash(name, len, 0);
-	bucket = hash & (ls->ls_rsbtbl_size - 1);
-
-	spin_lock(&ls->ls_rsbtbl[bucket].lock);
-	rv = dlm_search_rsb_tree(&ls->ls_rsbtbl[bucket].keep, name, len, &r);
-	if (rv)
-		rv = dlm_search_rsb_tree(&ls->ls_rsbtbl[bucket].toss,
-					 name, len, &r);
-	spin_unlock(&ls->ls_rsbtbl[bucket].lock);
-
+	read_lock_bh(&ls->ls_rsbtbl_lock);
+	rv = dlm_search_rsb_tree(&ls->ls_rsbtbl, name, len, &r);
+	read_unlock_bh(&ls->ls_rsbtbl_lock);
 	if (!rv)
 		return r;
 
-	down_read(&ls->ls_root_sem);
-	list_for_each_entry(r, &ls->ls_root_list, res_root_list) {
+	list_for_each_entry(r, &ls->ls_masters_list, res_masters_list) {
 		if (len == r->res_length && !memcmp(name, r->res_name, len)) {
-			up_read(&ls->ls_root_sem);
 			log_debug(ls, "find_rsb_root revert to root_list %s",
 				  r->res_name);
 			return r;
 		}
 	}
-	up_read(&ls->ls_root_sem);
 	return NULL;
 }
 
+struct dlm_dir_dump {
+	/* init values to match if whole
+	 * dump fits to one seq. Sanity check only.
+	 */
+	uint64_t seq_init;
+	uint64_t nodeid_init;
+	/* compare local pointer with last lookup,
+	 * just a sanity check.
+	 */
+	struct list_head *last;
+
+	unsigned int sent_res; /* for log info */
+	unsigned int sent_msg; /* for log info */
+
+	struct list_head list;
+};
+
+static void drop_dir_ctx(struct dlm_ls *ls, int nodeid)
+{
+	struct dlm_dir_dump *dd, *safe;
+
+	write_lock_bh(&ls->ls_dir_dump_lock);
+	list_for_each_entry_safe(dd, safe, &ls->ls_dir_dump_list, list) {
+		if (dd->nodeid_init == nodeid) {
+			log_error(ls, "drop dump seq %llu",
+				 (unsigned long long)dd->seq_init);
+			list_del(&dd->list);
+			kfree(dd);
+		}
+	}
+	write_unlock_bh(&ls->ls_dir_dump_lock);
+}
+
+static struct dlm_dir_dump *lookup_dir_dump(struct dlm_ls *ls, int nodeid)
+{
+	struct dlm_dir_dump *iter, *dd = NULL;
+
+	read_lock_bh(&ls->ls_dir_dump_lock);
+	list_for_each_entry(iter, &ls->ls_dir_dump_list, list) {
+		if (iter->nodeid_init == nodeid) {
+			dd = iter;
+			break;
+		}
+	}
+	read_unlock_bh(&ls->ls_dir_dump_lock);
+
+	return dd;
+}
+
+static struct dlm_dir_dump *init_dir_dump(struct dlm_ls *ls, int nodeid)
+{
+	struct dlm_dir_dump *dd;
+
+	dd = lookup_dir_dump(ls, nodeid);
+	if (dd) {
+		log_error(ls, "found ongoing dir dump for node %d, will drop it",
+			  nodeid);
+		drop_dir_ctx(ls, nodeid);
+	}
+
+	dd = kzalloc(sizeof(*dd), GFP_ATOMIC);
+	if (!dd)
+		return NULL;
+
+	dd->seq_init = ls->ls_recover_seq;
+	dd->nodeid_init = nodeid;
+
+	write_lock_bh(&ls->ls_dir_dump_lock);
+	list_add(&dd->list, &ls->ls_dir_dump_list);
+	write_unlock_bh(&ls->ls_dir_dump_lock);
+
+	return dd;
+}
+
 /* Find the rsb where we left off (or start again), then send rsb names
    for rsb's we're master of and whose directory node matches the requesting
    node.  inbuf is the rsb name last sent, inlen is the name's length */
@@ -239,27 +300,50 @@ void dlm_copy_master_names(struct dlm_ls *ls, const char *inbuf, int inlen,
 	struct list_head *list;
 	struct dlm_rsb *r;
 	int offset = 0, dir_nodeid;
+	struct dlm_dir_dump *dd;
 	__be16 be_namelen;
 
-	down_read(&ls->ls_root_sem);
+	read_lock_bh(&ls->ls_masters_lock);
 
 	if (inlen > 1) {
+		dd = lookup_dir_dump(ls, nodeid);
+		if (!dd) {
+			log_error(ls, "failed to lookup dir dump context nodeid: %d",
+				  nodeid);
+			goto out;
+		}
+
+		/* next chunk in dump */
 		r = find_rsb_root(ls, inbuf, inlen);
 		if (!r) {
 			log_error(ls, "copy_master_names from %d start %d %.*s",
 				  nodeid, inlen, inlen, inbuf);
 			goto out;
 		}
-		list = r->res_root_list.next;
+		list = r->res_masters_list.next;
+
+		/* sanity checks */
+		if (dd->last != &r->res_masters_list ||
+		    dd->seq_init != ls->ls_recover_seq) {
+			log_error(ls, "failed dir dump sanity check seq_init: %llu seq: %llu",
+				  (unsigned long long)dd->seq_init,
+				  (unsigned long long)ls->ls_recover_seq);
+			goto out;
+		}
 	} else {
-		list = ls->ls_root_list.next;
-	}
+		dd = init_dir_dump(ls, nodeid);
+		if (!dd) {
+			log_error(ls, "failed to allocate dir dump context");
+			goto out;
+		}
 
-	for (offset = 0; list != &ls->ls_root_list; list = list->next) {
-		r = list_entry(list, struct dlm_rsb, res_root_list);
-		if (r->res_nodeid)
-			continue;
+		/* start dump */
+		list = ls->ls_masters_list.next;
+		dd->last = list;
+	}
 
+	for (offset = 0; list != &ls->ls_masters_list; list = list->next) {
+		r = list_entry(list, struct dlm_rsb, res_masters_list);
 		dir_nodeid = dlm_dir_nodeid(r);
 		if (dir_nodeid != nodeid)
 			continue;
@@ -277,7 +361,7 @@ void dlm_copy_master_names(struct dlm_ls *ls, const char *inbuf, int inlen,
 			be_namelen = cpu_to_be16(0);
 			memcpy(outbuf + offset, &be_namelen, sizeof(__be16));
 			offset += sizeof(__be16);
-			ls->ls_recover_dir_sent_msg++;
+			dd->sent_msg++;
 			goto out;
 		}
 
@@ -286,7 +370,8 @@ void dlm_copy_master_names(struct dlm_ls *ls, const char *inbuf, int inlen,
 		offset += sizeof(__be16);
 		memcpy(outbuf + offset, r->res_name, r->res_length);
 		offset += r->res_length;
-		ls->ls_recover_dir_sent_res++;
+		dd->sent_res++;
+		dd->last = list;
 	}
 
 	/*
@@ -294,14 +379,22 @@ void dlm_copy_master_names(struct dlm_ls *ls, const char *inbuf, int inlen,
 	 * terminating record.
 	 */
 
-	if ((list == &ls->ls_root_list) &&
+	if ((list == &ls->ls_masters_list) &&
 	    (offset + sizeof(uint16_t) <= outlen)) {
+		/* end dump */
 		be_namelen = cpu_to_be16(0xFFFF);
 		memcpy(outbuf + offset, &be_namelen, sizeof(__be16));
 		offset += sizeof(__be16);
-		ls->ls_recover_dir_sent_msg++;
+		dd->sent_msg++;
+		log_rinfo(ls, "dlm_recover_directory nodeid %d sent %u res out %u messages",
+			  nodeid, dd->sent_res, dd->sent_msg);
+
+		write_lock_bh(&ls->ls_dir_dump_lock);
+		list_del_init(&dd->list);
+		write_unlock_bh(&ls->ls_dir_dump_lock);
+		kfree(dd);
 	}
  out:
-	up_read(&ls->ls_root_sem);
+	read_unlock_bh(&ls->ls_masters_lock);
 }
 
diff --git a/fs/dlm/dir.h b/fs/dlm/dir.h
index 39ecb69d7e..5b2a7ee376 100644
--- a/fs/dlm/dir.h
+++ b/fs/dlm/dir.h
@@ -14,7 +14,8 @@
 
 int dlm_dir_nodeid(struct dlm_rsb *rsb);
 int dlm_hash2nodeid(struct dlm_ls *ls, uint32_t hash);
-void dlm_recover_dir_nodeid(struct dlm_ls *ls);
+void dlm_recover_dir_nodeid(struct dlm_ls *ls,
+			    const struct list_head *root_list);
 int dlm_recover_directory(struct dlm_ls *ls, uint64_t seq);
 void dlm_copy_master_names(struct dlm_ls *ls, const char *inbuf, int inlen,
 			   char *outbuf, int outlen, int nodeid);
diff --git a/fs/dlm/dlm_internal.h b/fs/dlm/dlm_internal.h
index 511d0b984f..9085ba3b2f 100644
--- a/fs/dlm/dlm_internal.h
+++ b/fs/dlm/dlm_internal.h
@@ -16,6 +16,7 @@
  * This is the main header file to be included in each DLM source file.
  */
 
+#include <uapi/linux/dlm_device.h>
 #include <linux/slab.h>
 #include <linux/sched.h>
 #include <linux/types.h>
@@ -33,6 +34,7 @@
 #include <linux/kernel.h>
 #include <linux/jhash.h>
 #include <linux/miscdevice.h>
+#include <linux/rhashtable.h>
 #include <linux/mutex.h>
 #include <linux/idr.h>
 #include <linux/ratelimit.h>
@@ -98,17 +100,6 @@ do { \
   } \
 }
 
-
-#define DLM_RTF_SHRINK_BIT	0
-
-struct dlm_rsbtable {
-	struct rb_root		keep;
-	struct rb_root		toss;
-	spinlock_t		lock;
-	unsigned long		flags;
-};
-
-
 /*
  * Lockspace member (per node in a ls)
  */
@@ -204,8 +195,7 @@ struct dlm_args {
 #define DLM_IFL_OVERLAP_CANCEL_BIT 20
 #define DLM_IFL_ENDOFLIFE_BIT	21
 #define DLM_IFL_DEADLOCK_CANCEL_BIT 24
-#define DLM_IFL_CB_PENDING_BIT	25
-#define __DLM_IFL_MAX_BIT	DLM_IFL_CB_PENDING_BIT
+#define __DLM_IFL_MAX_BIT	DLM_IFL_DEADLOCK_CANCEL_BIT
 
 /* lkb_dflags */
 
@@ -217,15 +207,47 @@ struct dlm_args {
 #define DLM_CB_CAST		0x00000001
 #define DLM_CB_BAST		0x00000002
 
+/* much of this is just saving user space pointers associated with the
+ * lock that we pass back to the user lib with an ast
+ */
+
+struct dlm_user_args {
+	struct dlm_user_proc	*proc; /* each process that opens the lockspace
+					* device has private data
+					* (dlm_user_proc) on the struct file,
+					* the process's locks point back to it
+					*/
+	struct dlm_lksb		lksb;
+	struct dlm_lksb __user	*user_lksb;
+	void __user		*castparam;
+	void __user		*castaddr;
+	void __user		*bastparam;
+	void __user		*bastaddr;
+	uint64_t		xid;
+};
+
 struct dlm_callback {
 	uint32_t		flags;		/* DLM_CBF_ */
 	int			sb_status;	/* copy to lksb status */
 	uint8_t			sb_flags;	/* copy to lksb flags */
 	int8_t			mode; /* rq mode of bast, gr mode of cast */
-	int			copy_lvb;
+	bool			copy_lvb;
+	struct dlm_lksb		*lkb_lksb;
+	unsigned char		lvbptr[DLM_USER_LVB_LEN];
+
+	union {
+		void			*astparam;	/* caller's ast arg */
+		struct dlm_user_args	ua;
+	};
+	struct work_struct	work;
+	void			(*bastfn)(void *astparam, int mode);
+	void			(*astfn)(void *astparam);
+	char			res_name[DLM_RESNAME_MAXLEN];
+	size_t			res_length;
+	uint32_t		ls_id;
+	uint32_t		lkb_id;
 
 	struct list_head	list;
-	struct kref		ref;
 };
 
 struct dlm_lkb {
@@ -247,7 +269,7 @@ struct dlm_lkb {
 	int8_t			lkb_highbast;	/* highest mode bast sent for */
 
 	int8_t			lkb_wait_type;	/* type of reply waiting for */
-	atomic_t		lkb_wait_count;
+	int8_t			lkb_wait_count;
 	int			lkb_wait_nodeid; /* for debugging */
 
 	struct list_head	lkb_statequeue;	/* rsb g/c/w list */
@@ -256,13 +278,10 @@ struct dlm_lkb {
 	struct list_head	lkb_ownqueue;	/* list of locks for a process */
 	ktime_t			lkb_timestamp;
 
-	spinlock_t		lkb_cb_lock;
-	struct work_struct	lkb_cb_work;
-	struct list_head	lkb_cb_list; /* for ls_cb_delay or proc->asts */
-	struct list_head	lkb_callbacks;
-	struct dlm_callback	*lkb_last_cast;
-	struct dlm_callback	*lkb_last_cb;
-	int			lkb_last_bast_mode;
+	int8_t			lkb_last_cast_cb_mode;
+	int8_t			lkb_last_bast_cb_mode;
+	int8_t			lkb_last_cb_mode;
+	uint8_t			lkb_last_cb_flags;
 	ktime_t			lkb_last_cast_time;	/* for debugging */
 	ktime_t			lkb_last_bast_time;	/* for debugging */
 
@@ -291,7 +310,7 @@ struct dlm_lkb {
 struct dlm_rsb {
 	struct dlm_ls		*res_ls;	/* the lockspace */
 	struct kref		res_ref;
-	struct mutex		res_mutex;
+	spinlock_t		res_lock;
 	unsigned long		res_flags;
 	int			res_length;	/* length of rsb name */
 	int			res_nodeid;
@@ -300,20 +319,22 @@ struct dlm_rsb {
 	int			res_id;		/* for ls_recover_idr */
 	uint32_t                res_lvbseq;
 	uint32_t		res_hash;
-	uint32_t		res_bucket;	/* rsbtbl */
 	unsigned long		res_toss_time;
 	uint32_t		res_first_lkid;
 	struct list_head	res_lookup;	/* lkbs waiting on first */
 	union {
 		struct list_head	res_hashchain;
-		struct rb_node		res_hashnode;	/* rsbtbl */
+		struct rhash_head	res_node; /* rsbtbl */
 	};
 	struct list_head	res_grantqueue;
 	struct list_head	res_convertqueue;
 	struct list_head	res_waitqueue;
 
+	struct list_head	res_rsbs_list;
 	struct list_head	res_root_list;	    /* used for recovery */
+	struct list_head	res_masters_list;   /* used for recovery */
 	struct list_head	res_recover_list;   /* used for recovery */
+	struct list_head	res_toss_q_list;
 	int			res_recover_locks_count;
 
 	char			*res_lvbptr;
@@ -347,6 +368,7 @@ enum rsb_flags {
 	RSB_RECOVER_CONVERT,
 	RSB_RECOVER_GRANT,
 	RSB_RECOVER_LVB_INVAL,
+	RSB_TOSS,
 };
 
 static inline void rsb_set_flag(struct dlm_rsb *r, enum rsb_flags flag)
@@ -560,24 +582,33 @@ struct dlm_ls {
 	struct kobject		ls_kobj;
 
 	struct idr		ls_lkbidr;
-	spinlock_t		ls_lkbidr_spin;
+	rwlock_t		ls_lkbidr_lock;
+
+	struct rhashtable	ls_rsbtbl;
+	rwlock_t		ls_rsbtbl_lock;
 
-	struct dlm_rsbtable	*ls_rsbtbl;
-	uint32_t		ls_rsbtbl_size;
+	struct list_head	ls_toss;
+	struct list_head	ls_keep;
 
-	struct mutex		ls_waiters_mutex;
+	struct timer_list	ls_timer;
+	/* this queue is ordered according the
+	 * absolute res_toss_time jiffies time
+	 * to mod_timer() with the first element
+	 * if necessary.
+	 */
+	struct list_head	ls_toss_q;
+	spinlock_t		ls_toss_q_lock;
+
+	spinlock_t		ls_waiters_lock;
 	struct list_head	ls_waiters;	/* lkbs needing a reply */
 
-	struct mutex		ls_orphans_mutex;
+	spinlock_t		ls_orphans_lock;
 	struct list_head	ls_orphans;
 
 	spinlock_t		ls_new_rsb_spin;
 	int			ls_new_rsb_count;
 	struct list_head	ls_new_rsb;	/* new rsb structs */
 
-	char			*ls_remove_names[DLM_REMOVE_NAMES_MAX];
-	int			ls_remove_lens[DLM_REMOVE_NAMES_MAX];
-
 	struct list_head	ls_nodes;	/* current nodes in ls */
 	struct list_head	ls_nodes_gone;	/* dead node list, recovery */
 	int			ls_num_nodes;	/* number of nodes in ls */
@@ -614,7 +645,6 @@ struct dlm_ls {
 
 	spinlock_t		ls_cb_lock;
 	struct list_head	ls_cb_delay; /* save for queue_work later */
-	struct timer_list	ls_timer;
 	struct task_struct	*ls_recoverd_task;
 	struct mutex		ls_recoverd_active;
 	spinlock_t		ls_recover_lock;
@@ -623,15 +653,11 @@ struct dlm_ls {
 	uint64_t		ls_recover_seq;
 	struct dlm_recover	*ls_recover_args;
 	struct rw_semaphore	ls_in_recovery;	/* block local requests */
-	struct rw_semaphore	ls_recv_active;	/* block dlm_recv */
+	rwlock_t		ls_recv_active;	/* block dlm_recv */
 	struct list_head	ls_requestqueue;/* queue remote requests */
-	atomic_t		ls_requestqueue_cnt;
-	wait_queue_head_t	ls_requestqueue_wait;
-	struct mutex		ls_requestqueue_mutex;
+	rwlock_t		ls_requestqueue_lock;
 	struct dlm_rcom		*ls_recover_buf;
 	int			ls_recover_nodeid; /* for debugging */
-	unsigned int		ls_recover_dir_sent_res; /* for log info */
-	unsigned int		ls_recover_dir_sent_msg; /* for log info */
 	unsigned int		ls_recover_locks_in; /* for log info */
 	uint64_t		ls_rcom_seq;
 	spinlock_t		ls_rcom_spin;
@@ -644,8 +670,10 @@ struct dlm_ls {
 	wait_queue_head_t	ls_recover_lock_wait;
 	spinlock_t		ls_clear_proc_locks;
 
-	struct list_head	ls_root_list;	/* root resources */
-	struct rw_semaphore	ls_root_sem;	/* protect root_list */
+	struct list_head	ls_masters_list; /* root resources */
+	rwlock_t		ls_masters_lock; /* protect root_list */
+	struct list_head	ls_dir_dump_list; /* root resources */
+	rwlock_t		ls_dir_dump_lock; /* protect root_list */
 
 	const struct dlm_lockspace_ops *ls_ops;
 	void			*ls_ops_arg;
@@ -687,23 +715,7 @@ struct dlm_ls {
 #define LSFL_UEVENT_WAIT	7
 #define LSFL_CB_DELAY		9
 #define LSFL_NODIR		10
-
-/* much of this is just saving user space pointers associated with the
-   lock that we pass back to the user lib with an ast */
-
-struct dlm_user_args {
-	struct dlm_user_proc	*proc; /* each process that opens the lockspace
-					  device has private data
-					  (dlm_user_proc) on the struct file,
-					  the process's locks point back to it*/
-	struct dlm_lksb		lksb;
-	struct dlm_lksb __user	*user_lksb;
-	void __user		*castparam;
-	void __user		*castaddr;
-	void __user		*bastparam;
-	void __user		*bastaddr;
-	uint64_t		xid;
-};
+#define LSFL_RECV_MSG_BLOCKED	11
 
 #define DLM_PROC_FLAGS_CLOSING 1
 #define DLM_PROC_FLAGS_COMPAT  2
diff --git a/fs/dlm/lock.c b/fs/dlm/lock.c
index 652c51fbbf..f103b8c305 100644
--- a/fs/dlm/lock.c
+++ b/fs/dlm/lock.c
@@ -201,7 +201,7 @@ void dlm_dump_rsb(struct dlm_rsb *r)
 
 /* Threads cannot use the lockspace while it's being recovered */
 
-static inline void dlm_lock_recovery(struct dlm_ls *ls)
+void dlm_lock_recovery(struct dlm_ls *ls)
 {
 	down_read(&ls->ls_in_recovery);
 }
@@ -320,11 +320,18 @@ static void queue_bast(struct dlm_rsb *r, struct dlm_lkb *lkb, int rqmode)
  * Basic operations on rsb's and lkb's
  */
 
+static inline unsigned long rsb_toss_jiffies(void)
+{
+	return jiffies + (READ_ONCE(dlm_config.ci_toss_secs) * HZ);
+}
+
 /* This is only called to add a reference when the code already holds
    a valid reference to the rsb, so there's no need for locking. */
 
 static inline void hold_rsb(struct dlm_rsb *r)
 {
+	/* rsbs in toss state never get referenced */
+	WARN_ON(rsb_flag(r, RSB_TOSS));
 	kref_get(&r->res_ref);
 }
 
@@ -333,19 +340,48 @@ void dlm_hold_rsb(struct dlm_rsb *r)
 	hold_rsb(r);
 }
 
+/* TODO move this to lib/refcount.c */
+static __must_check bool
+dlm_refcount_dec_and_write_lock_bh(refcount_t *r, rwlock_t *lock)
+__cond_acquires(lock)
+{
+	if (refcount_dec_not_one(r))
+		return false;
+
+	write_lock_bh(lock);
+	if (!refcount_dec_and_test(r)) {
+		write_unlock_bh(lock);
+		return false;
+	}
+
+	return true;
+}
+
+/* TODO move this to include/linux/kref.h */
+static inline int dlm_kref_put_write_lock_bh(struct kref *kref,
+					     void (*release)(struct kref *kref),
+					     rwlock_t *lock)
+{
+	if (dlm_refcount_dec_and_write_lock_bh(&kref->refcount, lock)) {
+		release(kref);
+		return 1;
+	}
+
+	return 0;
+}
+
 /* When all references to the rsb are gone it's transferred to
    the tossed list for later disposal. */
 
 static void put_rsb(struct dlm_rsb *r)
 {
 	struct dlm_ls *ls = r->res_ls;
-	uint32_t bucket = r->res_bucket;
 	int rv;
 
-	rv = kref_put_lock(&r->res_ref, toss_rsb,
-			   &ls->ls_rsbtbl[bucket].lock);
+	rv = dlm_kref_put_write_lock_bh(&r->res_ref, toss_rsb,
+					&ls->ls_rsbtbl_lock);
 	if (rv)
-		spin_unlock(&ls->ls_rsbtbl[bucket].lock);
+		write_unlock_bh(&ls->ls_rsbtbl_lock);
 }
 
 void dlm_put_rsb(struct dlm_rsb *r)
@@ -358,17 +394,17 @@ static int pre_rsb_struct(struct dlm_ls *ls)
 	struct dlm_rsb *r1, *r2;
 	int count = 0;
 
-	spin_lock(&ls->ls_new_rsb_spin);
+	spin_lock_bh(&ls->ls_new_rsb_spin);
 	if (ls->ls_new_rsb_count > dlm_config.ci_new_rsb_count / 2) {
-		spin_unlock(&ls->ls_new_rsb_spin);
+		spin_unlock_bh(&ls->ls_new_rsb_spin);
 		return 0;
 	}
-	spin_unlock(&ls->ls_new_rsb_spin);
+	spin_unlock_bh(&ls->ls_new_rsb_spin);
 
 	r1 = dlm_allocate_rsb(ls);
 	r2 = dlm_allocate_rsb(ls);
 
-	spin_lock(&ls->ls_new_rsb_spin);
+	spin_lock_bh(&ls->ls_new_rsb_spin);
 	if (r1) {
 		list_add(&r1->res_hashchain, &ls->ls_new_rsb);
 		ls->ls_new_rsb_count++;
@@ -378,13 +414,236 @@ static int pre_rsb_struct(struct dlm_ls *ls)
 		ls->ls_new_rsb_count++;
 	}
 	count = ls->ls_new_rsb_count;
-	spin_unlock(&ls->ls_new_rsb_spin);
+	spin_unlock_bh(&ls->ls_new_rsb_spin);
 
 	if (!count)
 		return -ENOMEM;
 	return 0;
 }
 
+/* connected with timer_delete_sync() in dlm_ls_stop() to stop
+ * new timers when recovery is triggered and don't run them
+ * again until a dlm_timer_resume() tries it again.
+ */
+static void __rsb_mod_timer(struct dlm_ls *ls, unsigned long jiffies)
+{
+	if (!dlm_locking_stopped(ls))
+		mod_timer(&ls->ls_timer, jiffies);
+}
+
+/* This function tries to resume the timer callback if a rsb
+ * is on the toss list and no timer is pending. It might that
+ * the first entry is on currently executed as timer callback
+ * but we don't care if a timer queued up again and does
+ * nothing. Should be a rare case.
+ */
+void dlm_timer_resume(struct dlm_ls *ls)
+{
+	struct dlm_rsb *r;
+
+	spin_lock_bh(&ls->ls_toss_q_lock);
+	r = list_first_entry_or_null(&ls->ls_toss_q, struct dlm_rsb,
+				     res_toss_q_list);
+	if (r && !timer_pending(&ls->ls_timer))
+		__rsb_mod_timer(ls, r->res_toss_time);
+	spin_unlock_bh(&ls->ls_toss_q_lock);
+}
+
+/* ls_rsbtbl_lock must be held and being sure the rsb is in toss state */
+static void rsb_delete_toss_timer(struct dlm_ls *ls, struct dlm_rsb *r)
+{
+	struct dlm_rsb *first;
+
+	spin_lock_bh(&ls->ls_toss_q_lock);
+	r->res_toss_time = 0;
+
+	/* if the rsb is not queued do nothing */
+	if (list_empty(&r->res_toss_q_list))
+		goto out;
+
+	/* get the first element before delete */
+	first = list_first_entry(&ls->ls_toss_q, struct dlm_rsb,
+				 res_toss_q_list);
+	list_del_init(&r->res_toss_q_list);
+	/* check if the first element was the rsb we deleted */
+	if (first == r) {
+		/* try to get the new first element, if the list
+		 * is empty now try to delete the timer, if we are
+		 * too late we don't care.
+		 *
+		 * if the list isn't empty and a new first element got
+		 * in place, set the new timer expire time.
+		 */
+		first = list_first_entry_or_null(&ls->ls_toss_q, struct dlm_rsb,
+						 res_toss_q_list);
+		if (!first)
+			timer_delete(&ls->ls_timer);
+		else
+			__rsb_mod_timer(ls, first->res_toss_time);
+	}
+
+out:
+	spin_unlock_bh(&ls->ls_toss_q_lock);
+}
+
+/* Caller must held ls_rsbtbl_lock and need to be called every time
+ * when either the rsb enters toss state or the toss state changes
+ * the dir/master nodeid.
+ */
+static void rsb_mod_timer(struct dlm_ls *ls, struct dlm_rsb *r)
+{
+	int our_nodeid = dlm_our_nodeid();
+	struct dlm_rsb *first;
+
+	/* If we're the directory record for this rsb, and
+	 * we're not the master of it, then we need to wait
+	 * for the master node to send us a dir remove for
+	 * before removing the dir record.
+	 */
+	if (!dlm_no_directory(ls) &&
+	    (r->res_master_nodeid != our_nodeid) &&
+	    (dlm_dir_nodeid(r) == our_nodeid)) {
+		rsb_delete_toss_timer(ls, r);
+		return;
+	}
+
+	spin_lock_bh(&ls->ls_toss_q_lock);
+	/* set the new rsb absolute expire time in the rsb */
+	r->res_toss_time = rsb_toss_jiffies();
+	if (list_empty(&ls->ls_toss_q)) {
+		/* if the queue is empty add the element and it's
+		 * our new expire time
+		 */
+		list_add_tail(&r->res_toss_q_list, &ls->ls_toss_q);
+		__rsb_mod_timer(ls, r->res_toss_time);
+	} else {
+		/* check if the rsb was already queued, if so delete
+		 * it from the toss queue
+		 */
+		if (!list_empty(&r->res_toss_q_list))
+			list_del(&r->res_toss_q_list);
+
+		/* try to get the maybe new first element and then add
+		 * to this rsb with the oldest expire time to the end
+		 * of the queue. If the list was empty before this
+		 * rsb expire time is our next expiration if it wasn't
+		 * the now new first elemet is our new expiration time
+		 */
+		first = list_first_entry_or_null(&ls->ls_toss_q, struct dlm_rsb,
+						 res_toss_q_list);
+		list_add_tail(&r->res_toss_q_list, &ls->ls_toss_q);
+		if (!first)
+			__rsb_mod_timer(ls, r->res_toss_time);
+		else
+			__rsb_mod_timer(ls, first->res_toss_time);
+	}
+	spin_unlock_bh(&ls->ls_toss_q_lock);
+}
+
+/* if we hit contention we do in 250 ms a retry to trylock.
+ * if there is any other mod_timer in between we don't care
+ * about that it expires earlier again this is only for the
+ * unlikely case nothing happened in this time.
+ */
+#define DLM_TOSS_TIMER_RETRY	(jiffies + msecs_to_jiffies(250))
+
+void dlm_rsb_toss_timer(struct timer_list *timer)
+{
+	struct dlm_ls *ls = from_timer(ls, timer, ls_timer);
+	int our_nodeid = dlm_our_nodeid();
+	struct dlm_rsb *r;
+	int rv;
+
+	while (1) {
+		/* interrupting point to leave iteration when
+		 * recovery waits for timer_delete_sync(), recovery
+		 * will take care to delete everything in toss queue.
+		 */
+		if (dlm_locking_stopped(ls))
+			break;
+
+		rv = spin_trylock(&ls->ls_toss_q_lock);
+		if (!rv) {
+			/* rearm again try timer */
+			__rsb_mod_timer(ls, DLM_TOSS_TIMER_RETRY);
+			break;
+		}
+
+		r = list_first_entry_or_null(&ls->ls_toss_q, struct dlm_rsb,
+					     res_toss_q_list);
+		if (!r) {
+			/* nothing to do anymore next rsb queue will
+			 * set next mod_timer() expire.
+			 */
+			spin_unlock(&ls->ls_toss_q_lock);
+			break;
+		}
+
+		/* test if the first rsb isn't expired yet, if
+		 * so we stop freeing rsb from toss queue as
+		 * the order in queue is ascending to the
+		 * absolute res_toss_time jiffies
+		 */
+		if (time_before(jiffies, r->res_toss_time)) {
+			/* rearm with the next rsb to expire in the future */
+			__rsb_mod_timer(ls, r->res_toss_time);
+			spin_unlock(&ls->ls_toss_q_lock);
+			break;
+		}
+
+		/* in find_rsb_dir/nodir there is a reverse order of this
+		 * lock, however this is only a trylock if we hit some
+		 * possible contention we try it again.
+		 *
+		 * This lock synchronized while holding ls_toss_q_lock
+		 * synchronize everything that rsb_delete_toss_timer()
+		 * or rsb_mod_timer() can't run after this timer callback
+		 * deletes the rsb from the ls_toss_q. Whereas the other
+		 * holders have always a priority to run as this is only
+		 * a caching handling and the other holders might to put
+		 * this rsb out of the toss state.
+		 */
+		rv = write_trylock(&ls->ls_rsbtbl_lock);
+		if (!rv) {
+			spin_unlock(&ls->ls_toss_q_lock);
+			/* rearm again try timer */
+			__rsb_mod_timer(ls, DLM_TOSS_TIMER_RETRY);
+			break;
+		}
+
+		list_del(&r->res_rsbs_list);
+		rhashtable_remove_fast(&ls->ls_rsbtbl, &r->res_node,
+				       dlm_rhash_rsb_params);
+
+		/* not necessary to held the ls_rsbtbl_lock when
+		 * calling send_remove()
+		 */
+		write_unlock(&ls->ls_rsbtbl_lock);
+
+		/* remove the rsb out of the toss queue its gone
+		 * drom DLM now
+		 */
+		list_del_init(&r->res_toss_q_list);
+		spin_unlock(&ls->ls_toss_q_lock);
+
+		/* no rsb in this state should ever run a timer */
+		WARN_ON(!dlm_no_directory(ls) &&
+			(r->res_master_nodeid != our_nodeid) &&
+			(dlm_dir_nodeid(r) == our_nodeid));
+
+		/* We're the master of this rsb but we're not
+		 * the directory record, so we need to tell the
+		 * dir node to remove the dir record
+		 */
+		if (!dlm_no_directory(ls) &&
+		    (r->res_master_nodeid == our_nodeid) &&
+		    (dlm_dir_nodeid(r) != our_nodeid))
+			send_remove(r);
+
+		free_toss_rsb(r);
+	}
+}
+
 /* If ls->ls_new_rsb is empty, return -EAGAIN, so the caller can
    unlock any spinlocks, go back and call pre_rsb_struct again.
    Otherwise, take an rsb off the list and return it. */
@@ -395,10 +654,10 @@ static int get_rsb_struct(struct dlm_ls *ls, const void *name, int len,
 	struct dlm_rsb *r;
 	int count;
 
-	spin_lock(&ls->ls_new_rsb_spin);
+	spin_lock_bh(&ls->ls_new_rsb_spin);
 	if (list_empty(&ls->ls_new_rsb)) {
 		count = ls->ls_new_rsb_count;
-		spin_unlock(&ls->ls_new_rsb_spin);
+		spin_unlock_bh(&ls->ls_new_rsb_spin);
 		log_debug(ls, "find_rsb retry %d %d %s",
 			  count, dlm_config.ci_new_rsb_count,
 			  (const char *)name);
@@ -407,88 +666,44 @@ static int get_rsb_struct(struct dlm_ls *ls, const void *name, int len,
 
 	r = list_first_entry(&ls->ls_new_rsb, struct dlm_rsb, res_hashchain);
 	list_del(&r->res_hashchain);
-	/* Convert the empty list_head to a NULL rb_node for tree usage: */
-	memset(&r->res_hashnode, 0, sizeof(struct rb_node));
 	ls->ls_new_rsb_count--;
-	spin_unlock(&ls->ls_new_rsb_spin);
+	spin_unlock_bh(&ls->ls_new_rsb_spin);
 
 	r->res_ls = ls;
 	r->res_length = len;
 	memcpy(r->res_name, name, len);
-	mutex_init(&r->res_mutex);
+	spin_lock_init(&r->res_lock);
 
 	INIT_LIST_HEAD(&r->res_lookup);
 	INIT_LIST_HEAD(&r->res_grantqueue);
 	INIT_LIST_HEAD(&r->res_convertqueue);
 	INIT_LIST_HEAD(&r->res_waitqueue);
 	INIT_LIST_HEAD(&r->res_root_list);
+	INIT_LIST_HEAD(&r->res_toss_q_list);
 	INIT_LIST_HEAD(&r->res_recover_list);
+	INIT_LIST_HEAD(&r->res_masters_list);
 
 	*r_ret = r;
 	return 0;
 }
 
-static int rsb_cmp(struct dlm_rsb *r, const char *name, int nlen)
+int dlm_search_rsb_tree(struct rhashtable *rhash, const void *name, int len,
+			struct dlm_rsb **r_ret)
 {
-	char maxname[DLM_RESNAME_MAXLEN];
+	char key[DLM_RESNAME_MAXLEN] = {};
 
-	memset(maxname, 0, DLM_RESNAME_MAXLEN);
-	memcpy(maxname, name, nlen);
-	return memcmp(r->res_name, maxname, DLM_RESNAME_MAXLEN);
-}
+	memcpy(key, name, len);
+	*r_ret = rhashtable_lookup_fast(rhash, &key, dlm_rhash_rsb_params);
+	if (*r_ret)
+		return 0;
 
-int dlm_search_rsb_tree(struct rb_root *tree, const void *name, int len,
-			struct dlm_rsb **r_ret)
-{
-	struct rb_node *node = tree->rb_node;
-	struct dlm_rsb *r;
-	int rc;
-
-	while (node) {
-		r = rb_entry(node, struct dlm_rsb, res_hashnode);
-		rc = rsb_cmp(r, name, len);
-		if (rc < 0)
-			node = node->rb_left;
-		else if (rc > 0)
-			node = node->rb_right;
-		else
-			goto found;
-	}
-	*r_ret = NULL;
 	return -EBADR;
-
- found:
-	*r_ret = r;
-	return 0;
 }
 
-static int rsb_insert(struct dlm_rsb *rsb, struct rb_root *tree)
+static int rsb_insert(struct dlm_rsb *rsb, struct rhashtable *rhash)
 {
-	struct rb_node **newn = &tree->rb_node;
-	struct rb_node *parent = NULL;
-	int rc;
-
-	while (*newn) {
-		struct dlm_rsb *cur = rb_entry(*newn, struct dlm_rsb,
-					       res_hashnode);
-
-		parent = *newn;
-		rc = rsb_cmp(cur, rsb->res_name, rsb->res_length);
-		if (rc < 0)
-			newn = &parent->rb_left;
-		else if (rc > 0)
-			newn = &parent->rb_right;
-		else {
-			log_print("rsb_insert match");
-			dlm_dump_rsb(rsb);
-			dlm_dump_rsb(cur);
-			return -EEXIST;
-		}
-	}
-
-	rb_link_node(&rsb->res_hashnode, parent, newn);
-	rb_insert_color(&rsb->res_hashnode, tree);
-	return 0;
+	return rhashtable_insert_fast(rhash, &rsb->res_node,
+				      dlm_rhash_rsb_params);
 }
 
 /*
@@ -536,8 +751,7 @@ static int rsb_insert(struct dlm_rsb *rsb, struct rb_root *tree)
  */
 
 static int find_rsb_dir(struct dlm_ls *ls, const void *name, int len,
-			uint32_t hash, uint32_t b,
-			int dir_nodeid, int from_nodeid,
+			uint32_t hash, int dir_nodeid, int from_nodeid,
 			unsigned int flags, struct dlm_rsb **r_ret)
 {
 	struct dlm_rsb *r = NULL;
@@ -584,24 +798,46 @@ static int find_rsb_dir(struct dlm_ls *ls, const void *name, int len,
 			goto out;
 	}
 
-	spin_lock(&ls->ls_rsbtbl[b].lock);
+ retry_lookup:
 
-	error = dlm_search_rsb_tree(&ls->ls_rsbtbl[b].keep, name, len, &r);
-	if (error)
-		goto do_toss;
+	/* check if the rsb is in keep state under read lock - likely path */
+	read_lock_bh(&ls->ls_rsbtbl_lock);
+	error = dlm_search_rsb_tree(&ls->ls_rsbtbl, name, len, &r);
+	if (error) {
+		read_unlock_bh(&ls->ls_rsbtbl_lock);
+		goto do_new;
+	}
 	
 	/*
 	 * rsb is active, so we can't check master_nodeid without lock_rsb.
 	 */
 
+	if (rsb_flag(r, RSB_TOSS)) {
+		read_unlock_bh(&ls->ls_rsbtbl_lock);
+		goto do_toss;
+	}
+
 	kref_get(&r->res_ref);
-	goto out_unlock;
+	read_unlock_bh(&ls->ls_rsbtbl_lock);
+	goto out;
 
 
  do_toss:
-	error = dlm_search_rsb_tree(&ls->ls_rsbtbl[b].toss, name, len, &r);
-	if (error)
+	write_lock_bh(&ls->ls_rsbtbl_lock);
+
+	/* retry lookup under write lock to see if its still in toss state
+	 * if not it's in keep state and we relookup - unlikely path.
+	 */
+	error = dlm_search_rsb_tree(&ls->ls_rsbtbl, name, len, &r);
+	if (!error) {
+		if (!rsb_flag(r, RSB_TOSS)) {
+			write_unlock_bh(&ls->ls_rsbtbl_lock);
+			goto retry_lookup;
+		}
+	} else {
+		write_unlock_bh(&ls->ls_rsbtbl_lock);
 		goto do_new;
+	}
 
 	/*
 	 * rsb found inactive (master_nodeid may be out of date unless
@@ -616,8 +852,9 @@ static int find_rsb_dir(struct dlm_ls *ls, const void *name, int len,
 		log_debug(ls, "find_rsb toss from_other %d master %d dir %d %s",
 			  from_nodeid, r->res_master_nodeid, dir_nodeid,
 			  r->res_name);
+		write_unlock_bh(&ls->ls_rsbtbl_lock);
 		error = -ENOTBLK;
-		goto out_unlock;
+		goto out;
 	}
 
 	if ((r->res_master_nodeid != our_nodeid) && from_dir) {
@@ -639,9 +876,17 @@ static int find_rsb_dir(struct dlm_ls *ls, const void *name, int len,
 		r->res_first_lkid = 0;
 	}
 
-	rb_erase(&r->res_hashnode, &ls->ls_rsbtbl[b].toss);
-	error = rsb_insert(r, &ls->ls_rsbtbl[b].keep);
-	goto out_unlock;
+	list_move(&r->res_rsbs_list, &ls->ls_keep);
+	rsb_clear_flag(r, RSB_TOSS);
+	/* rsb got out of toss state, it becomes alive again
+	 * and we reinit the reference counter that is only
+	 * valid for keep state rsbs
+	 */
+	kref_init(&r->res_ref);
+	rsb_delete_toss_timer(ls, r);
+	write_unlock_bh(&ls->ls_rsbtbl_lock);
+
+	goto out;
 
 
  do_new:
@@ -650,18 +895,15 @@ static int find_rsb_dir(struct dlm_ls *ls, const void *name, int len,
 	 */
 
 	if (error == -EBADR && !create)
-		goto out_unlock;
+		goto out;
 
 	error = get_rsb_struct(ls, name, len, &r);
-	if (error == -EAGAIN) {
-		spin_unlock(&ls->ls_rsbtbl[b].lock);
+	if (error == -EAGAIN)
 		goto retry;
-	}
 	if (error)
-		goto out_unlock;
+		goto out;
 
 	r->res_hash = hash;
-	r->res_bucket = b;
 	r->res_dir_nodeid = dir_nodeid;
 	kref_init(&r->res_ref);
 
@@ -681,7 +923,7 @@ static int find_rsb_dir(struct dlm_ls *ls, const void *name, int len,
 		dlm_free_rsb(r);
 		r = NULL;
 		error = -ENOTBLK;
-		goto out_unlock;
+		goto out;
 	}
 
 	if (from_other) {
@@ -701,9 +943,20 @@ static int find_rsb_dir(struct dlm_ls *ls, const void *name, int len,
 	}
 
  out_add:
-	error = rsb_insert(r, &ls->ls_rsbtbl[b].keep);
- out_unlock:
-	spin_unlock(&ls->ls_rsbtbl[b].lock);
+
+	write_lock_bh(&ls->ls_rsbtbl_lock);
+	error = rsb_insert(r, &ls->ls_rsbtbl);
+	if (error == -EEXIST) {
+		/* somebody else was faster and it seems the
+		 * rsb exists now, we do a whole relookup
+		 */
+		write_unlock_bh(&ls->ls_rsbtbl_lock);
+		dlm_free_rsb(r);
+		goto retry_lookup;
+	} else if (!error) {
+		list_add(&r->res_rsbs_list, &ls->ls_keep);
+	}
+	write_unlock_bh(&ls->ls_rsbtbl_lock);
  out:
 	*r_ret = r;
 	return error;
@@ -714,8 +967,7 @@ static int find_rsb_dir(struct dlm_ls *ls, const void *name, int len,
    dlm_recover_masters). */
 
 static int find_rsb_nodir(struct dlm_ls *ls, const void *name, int len,
-			  uint32_t hash, uint32_t b,
-			  int dir_nodeid, int from_nodeid,
+			  uint32_t hash, int dir_nodeid, int from_nodeid,
 			  unsigned int flags, struct dlm_rsb **r_ret)
 {
 	struct dlm_rsb *r = NULL;
@@ -728,24 +980,48 @@ static int find_rsb_nodir(struct dlm_ls *ls, const void *name, int len,
 	if (error < 0)
 		goto out;
 
-	spin_lock(&ls->ls_rsbtbl[b].lock);
+ retry_lookup:
 
-	error = dlm_search_rsb_tree(&ls->ls_rsbtbl[b].keep, name, len, &r);
-	if (error)
+	/* check if the rsb is in keep state under read lock - likely path */
+	read_lock_bh(&ls->ls_rsbtbl_lock);
+	error = dlm_search_rsb_tree(&ls->ls_rsbtbl, name, len, &r);
+	if (error) {
+		read_unlock_bh(&ls->ls_rsbtbl_lock);
+		goto do_new;
+	}
+
+	if (rsb_flag(r, RSB_TOSS)) {
+		read_unlock_bh(&ls->ls_rsbtbl_lock);
 		goto do_toss;
+	}
 
 	/*
 	 * rsb is active, so we can't check master_nodeid without lock_rsb.
 	 */
 
 	kref_get(&r->res_ref);
-	goto out_unlock;
+	read_unlock_bh(&ls->ls_rsbtbl_lock);
+
+	goto out;
 
 
  do_toss:
-	error = dlm_search_rsb_tree(&ls->ls_rsbtbl[b].toss, name, len, &r);
-	if (error)
+	write_lock_bh(&ls->ls_rsbtbl_lock);
+
+	/* retry lookup under write lock to see if its still in toss state
+	 * if not it's in keep state and we relookup - unlikely path.
+	 */
+	error = dlm_search_rsb_tree(&ls->ls_rsbtbl, name, len, &r);
+	if (!error) {
+		if (!rsb_flag(r, RSB_TOSS)) {
+			write_unlock_bh(&ls->ls_rsbtbl_lock);
+			goto retry_lookup;
+		}
+	} else {
+		write_unlock_bh(&ls->ls_rsbtbl_lock);
 		goto do_new;
+	}
+
 
 	/*
 	 * rsb found inactive. No other thread is using this rsb because
@@ -759,8 +1035,9 @@ static int find_rsb_nodir(struct dlm_ls *ls, const void *name, int len,
 		log_error(ls, "find_rsb toss from_nodeid %d master %d dir %d",
 			  from_nodeid, r->res_master_nodeid, dir_nodeid);
 		dlm_print_rsb(r);
+		write_unlock_bh(&ls->ls_rsbtbl_lock);
 		error = -ENOTBLK;
-		goto out_unlock;
+		goto out;
 	}
 
 	if (!recover && (r->res_master_nodeid != our_nodeid) &&
@@ -774,9 +1051,17 @@ static int find_rsb_nodir(struct dlm_ls *ls, const void *name, int len,
 		r->res_nodeid = 0;
 	}
 
-	rb_erase(&r->res_hashnode, &ls->ls_rsbtbl[b].toss);
-	error = rsb_insert(r, &ls->ls_rsbtbl[b].keep);
-	goto out_unlock;
+	list_move(&r->res_rsbs_list, &ls->ls_keep);
+	rsb_clear_flag(r, RSB_TOSS);
+	/* rsb got out of toss state, it becomes alive again
+	 * and we reinit the reference counter that is only
+	 * valid for keep state rsbs
+	 */
+	kref_init(&r->res_ref);
+	rsb_delete_toss_timer(ls, r);
+	write_unlock_bh(&ls->ls_rsbtbl_lock);
+
+	goto out;
 
 
  do_new:
@@ -786,22 +1071,31 @@ static int find_rsb_nodir(struct dlm_ls *ls, const void *name, int len,
 
 	error = get_rsb_struct(ls, name, len, &r);
 	if (error == -EAGAIN) {
-		spin_unlock(&ls->ls_rsbtbl[b].lock);
 		goto retry;
 	}
 	if (error)
-		goto out_unlock;
+		goto out;
 
 	r->res_hash = hash;
-	r->res_bucket = b;
 	r->res_dir_nodeid = dir_nodeid;
 	r->res_master_nodeid = dir_nodeid;
 	r->res_nodeid = (dir_nodeid == our_nodeid) ? 0 : dir_nodeid;
 	kref_init(&r->res_ref);
 
-	error = rsb_insert(r, &ls->ls_rsbtbl[b].keep);
- out_unlock:
-	spin_unlock(&ls->ls_rsbtbl[b].lock);
+	write_lock_bh(&ls->ls_rsbtbl_lock);
+	error = rsb_insert(r, &ls->ls_rsbtbl);
+	if (error == -EEXIST) {
+		/* somebody else was faster and it seems the
+		 * rsb exists now, we do a whole relookup
+		 */
+		write_unlock_bh(&ls->ls_rsbtbl_lock);
+		dlm_free_rsb(r);
+		goto retry_lookup;
+	} else if (!error) {
+		list_add(&r->res_rsbs_list, &ls->ls_keep);
+	}
+	write_unlock_bh(&ls->ls_rsbtbl_lock);
+
  out:
 	*r_ret = r;
 	return error;
@@ -811,23 +1105,21 @@ static int find_rsb(struct dlm_ls *ls, const void *name, int len,
 		    int from_nodeid, unsigned int flags,
 		    struct dlm_rsb **r_ret)
 {
-	uint32_t hash, b;
 	int dir_nodeid;
+	uint32_t hash;
 
 	if (len > DLM_RESNAME_MAXLEN)
 		return -EINVAL;
 
 	hash = jhash(name, len, 0);
-	b = hash & (ls->ls_rsbtbl_size - 1);
-
 	dir_nodeid = dlm_hash2nodeid(ls, hash);
 
 	if (dlm_no_directory(ls))
-		return find_rsb_nodir(ls, name, len, hash, b, dir_nodeid,
+		return find_rsb_nodir(ls, name, len, hash, dir_nodeid,
 				      from_nodeid, flags, r_ret);
 	else
-		return find_rsb_dir(ls, name, len, hash, b, dir_nodeid,
-				      from_nodeid, flags, r_ret);
+		return find_rsb_dir(ls, name, len, hash, dir_nodeid,
+				    from_nodeid, flags, r_ret);
 }
 
 /* we have received a request and found that res_master_nodeid != our_nodeid,
@@ -988,7 +1280,7 @@ int dlm_master_lookup(struct dlm_ls *ls, int from_nodeid, const char *name,
 		      int len, unsigned int flags, int *r_nodeid, int *result)
 {
 	struct dlm_rsb *r = NULL;
-	uint32_t hash, b;
+	uint32_t hash;
 	int our_nodeid = dlm_our_nodeid();
 	int dir_nodeid, error;
 
@@ -1002,8 +1294,6 @@ int dlm_master_lookup(struct dlm_ls *ls, int from_nodeid, const char *name,
 	}
 
 	hash = jhash(name, len, 0);
-	b = hash & (ls->ls_rsbtbl_size - 1);
-
 	dir_nodeid = dlm_hash2nodeid(ls, hash);
 	if (dir_nodeid != our_nodeid) {
 		log_error(ls, "dlm_master_lookup from %d dir %d our %d h %x %d",
@@ -1018,15 +1308,23 @@ int dlm_master_lookup(struct dlm_ls *ls, int from_nodeid, const char *name,
 	if (error < 0)
 		return error;
 
-	spin_lock(&ls->ls_rsbtbl[b].lock);
-	error = dlm_search_rsb_tree(&ls->ls_rsbtbl[b].keep, name, len, &r);
+ retry_lookup:
+
+	/* check if the rsb is in keep state under read lock - likely path */
+	read_lock_bh(&ls->ls_rsbtbl_lock);
+	error = dlm_search_rsb_tree(&ls->ls_rsbtbl, name, len, &r);
 	if (!error) {
+		if (rsb_flag(r, RSB_TOSS)) {
+			read_unlock_bh(&ls->ls_rsbtbl_lock);
+			goto do_toss;
+		}
+
 		/* because the rsb is active, we need to lock_rsb before
 		 * checking/changing re_master_nodeid
 		 */
 
 		hold_rsb(r);
-		spin_unlock(&ls->ls_rsbtbl[b].lock);
+		read_unlock_bh(&ls->ls_rsbtbl_lock);
 		lock_rsb(r);
 
 		__dlm_master_lookup(ls, r, our_nodeid, from_nodeid, false,
@@ -1037,11 +1335,31 @@ int dlm_master_lookup(struct dlm_ls *ls, int from_nodeid, const char *name,
 		put_rsb(r);
 
 		return 0;
+	} else {
+		read_unlock_bh(&ls->ls_rsbtbl_lock);
+		goto not_found;
 	}
 
-	error = dlm_search_rsb_tree(&ls->ls_rsbtbl[b].toss, name, len, &r);
-	if (error)
+ do_toss:
+	/* unlikely path - relookup under write */
+	write_lock_bh(&ls->ls_rsbtbl_lock);
+
+	/* rsb_mod_timer() requires to held ls_rsbtbl_lock in write lock
+	 * check if the rsb is still in toss state, if not relookup
+	 */
+	error = dlm_search_rsb_tree(&ls->ls_rsbtbl, name, len, &r);
+	if (!error) {
+		if (!rsb_flag(r, RSB_TOSS)) {
+			write_unlock_bh(&ls->ls_rsbtbl_lock);
+			/* something as changed, very unlikely but
+			 * try again
+			 */
+			goto retry_lookup;
+		}
+	} else {
+		write_unlock_bh(&ls->ls_rsbtbl_lock);
 		goto not_found;
+	}
 
 	/* because the rsb is inactive (on toss list), it's not refcounted
 	 * and lock_rsb is not used, but is protected by the rsbtbl lock
@@ -1050,83 +1368,78 @@ int dlm_master_lookup(struct dlm_ls *ls, int from_nodeid, const char *name,
 	__dlm_master_lookup(ls, r, our_nodeid, from_nodeid, true, flags,
 			    r_nodeid, result);
 
-	r->res_toss_time = jiffies;
+	rsb_mod_timer(ls, r);
 	/* the rsb was inactive (on toss list) */
-	spin_unlock(&ls->ls_rsbtbl[b].lock);
+	write_unlock_bh(&ls->ls_rsbtbl_lock);
 
 	return 0;
 
  not_found:
 	error = get_rsb_struct(ls, name, len, &r);
-	if (error == -EAGAIN) {
-		spin_unlock(&ls->ls_rsbtbl[b].lock);
+	if (error == -EAGAIN)
 		goto retry;
-	}
 	if (error)
-		goto out_unlock;
+		goto out;
 
 	r->res_hash = hash;
-	r->res_bucket = b;
 	r->res_dir_nodeid = our_nodeid;
 	r->res_master_nodeid = from_nodeid;
 	r->res_nodeid = from_nodeid;
 	kref_init(&r->res_ref);
-	r->res_toss_time = jiffies;
+	rsb_set_flag(r, RSB_TOSS);
 
-	error = rsb_insert(r, &ls->ls_rsbtbl[b].toss);
-	if (error) {
+	write_lock_bh(&ls->ls_rsbtbl_lock);
+	error = rsb_insert(r, &ls->ls_rsbtbl);
+	if (error == -EEXIST) {
+		/* somebody else was faster and it seems the
+		 * rsb exists now, we do a whole relookup
+		 */
+		write_unlock_bh(&ls->ls_rsbtbl_lock);
+		dlm_free_rsb(r);
+		goto retry_lookup;
+	} else if (error) {
+		write_unlock_bh(&ls->ls_rsbtbl_lock);
 		/* should never happen */
 		dlm_free_rsb(r);
-		spin_unlock(&ls->ls_rsbtbl[b].lock);
 		goto retry;
 	}
 
+	list_add(&r->res_rsbs_list, &ls->ls_toss);
+	rsb_mod_timer(ls, r);
+	write_unlock_bh(&ls->ls_rsbtbl_lock);
+
 	if (result)
 		*result = DLM_LU_ADD;
 	*r_nodeid = from_nodeid;
- out_unlock:
-	spin_unlock(&ls->ls_rsbtbl[b].lock);
+ out:
 	return error;
 }
 
 static void dlm_dump_rsb_hash(struct dlm_ls *ls, uint32_t hash)
 {
-	struct rb_node *n;
 	struct dlm_rsb *r;
-	int i;
 
-	for (i = 0; i < ls->ls_rsbtbl_size; i++) {
-		spin_lock(&ls->ls_rsbtbl[i].lock);
-		for (n = rb_first(&ls->ls_rsbtbl[i].keep); n; n = rb_next(n)) {
-			r = rb_entry(n, struct dlm_rsb, res_hashnode);
-			if (r->res_hash == hash)
-				dlm_dump_rsb(r);
-		}
-		spin_unlock(&ls->ls_rsbtbl[i].lock);
+	read_lock_bh(&ls->ls_rsbtbl_lock);
+	list_for_each_entry(r, &ls->ls_keep, res_rsbs_list) {
+		if (r->res_hash == hash)
+			dlm_dump_rsb(r);
 	}
+	read_unlock_bh(&ls->ls_rsbtbl_lock);
 }
 
 void dlm_dump_rsb_name(struct dlm_ls *ls, const char *name, int len)
 {
 	struct dlm_rsb *r = NULL;
-	uint32_t hash, b;
 	int error;
 
-	hash = jhash(name, len, 0);
-	b = hash & (ls->ls_rsbtbl_size - 1);
-
-	spin_lock(&ls->ls_rsbtbl[b].lock);
-	error = dlm_search_rsb_tree(&ls->ls_rsbtbl[b].keep, name, len, &r);
+	read_lock_bh(&ls->ls_rsbtbl_lock);
+	error = dlm_search_rsb_tree(&ls->ls_rsbtbl, name, len, &r);
 	if (!error)
-		goto out_dump;
-
-	error = dlm_search_rsb_tree(&ls->ls_rsbtbl[b].toss, name, len, &r);
-	if (error)
 		goto out;
- out_dump:
+
 	dlm_dump_rsb(r);
  out:
-	spin_unlock(&ls->ls_rsbtbl[b].lock);
+	read_unlock_bh(&ls->ls_rsbtbl_lock);
 }
 
 static void toss_rsb(struct kref *kref)
@@ -1135,11 +1448,10 @@ static void toss_rsb(struct kref *kref)
 	struct dlm_ls *ls = r->res_ls;
 
 	DLM_ASSERT(list_empty(&r->res_root_list), dlm_print_rsb(r););
-	kref_init(&r->res_ref);
-	rb_erase(&r->res_hashnode, &ls->ls_rsbtbl[r->res_bucket].keep);
-	rsb_insert(r, &ls->ls_rsbtbl[r->res_bucket].toss);
-	r->res_toss_time = jiffies;
-	set_bit(DLM_RTF_SHRINK_BIT, &ls->ls_rsbtbl[r->res_bucket].flags);
+	rsb_set_flag(r, RSB_TOSS);
+	list_move(&r->res_rsbs_list, &ls->ls_toss);
+	rsb_mod_timer(ls, r);
+
 	if (r->res_lvbptr) {
 		dlm_free_lvb(r->res_lvbptr);
 		r->res_lvbptr = NULL;
@@ -1151,23 +1463,27 @@ static void toss_rsb(struct kref *kref)
 static void unhold_rsb(struct dlm_rsb *r)
 {
 	int rv;
+
+	/* rsbs in toss state never get referenced */
+	WARN_ON(rsb_flag(r, RSB_TOSS));
 	rv = kref_put(&r->res_ref, toss_rsb);
 	DLM_ASSERT(!rv, dlm_dump_rsb(r););
 }
 
-static void kill_rsb(struct kref *kref)
+void free_toss_rsb(struct dlm_rsb *r)
 {
-	struct dlm_rsb *r = container_of(kref, struct dlm_rsb, res_ref);
-
-	/* All work is done after the return from kref_put() so we
-	   can release the write_lock before the remove and free. */
+	WARN_ON_ONCE(!rsb_flag(r, RSB_TOSS));
 
 	DLM_ASSERT(list_empty(&r->res_lookup), dlm_dump_rsb(r););
 	DLM_ASSERT(list_empty(&r->res_grantqueue), dlm_dump_rsb(r););
 	DLM_ASSERT(list_empty(&r->res_convertqueue), dlm_dump_rsb(r););
 	DLM_ASSERT(list_empty(&r->res_waitqueue), dlm_dump_rsb(r););
 	DLM_ASSERT(list_empty(&r->res_root_list), dlm_dump_rsb(r););
+	DLM_ASSERT(list_empty(&r->res_toss_q_list), dlm_dump_rsb(r););
 	DLM_ASSERT(list_empty(&r->res_recover_list), dlm_dump_rsb(r););
+	DLM_ASSERT(list_empty(&r->res_masters_list), dlm_dump_rsb(r););
+
+	dlm_free_rsb(r);
 }
 
 /* Attaching/detaching lkb's from rsb's is for rsb reference counting.
@@ -1197,24 +1513,20 @@ static int _create_lkb(struct dlm_ls *ls, struct dlm_lkb **lkb_ret,
 	if (!lkb)
 		return -ENOMEM;
 
-	lkb->lkb_last_bast_mode = -1;
+	lkb->lkb_last_bast_cb_mode = DLM_LOCK_IV;
+	lkb->lkb_last_cast_cb_mode = DLM_LOCK_IV;
+	lkb->lkb_last_cb_mode = DLM_LOCK_IV;
 	lkb->lkb_nodeid = -1;
 	lkb->lkb_grmode = DLM_LOCK_IV;
 	kref_init(&lkb->lkb_ref);
 	INIT_LIST_HEAD(&lkb->lkb_ownqueue);
 	INIT_LIST_HEAD(&lkb->lkb_rsb_lookup);
-	INIT_LIST_HEAD(&lkb->lkb_cb_list);
-	INIT_LIST_HEAD(&lkb->lkb_callbacks);
-	spin_lock_init(&lkb->lkb_cb_lock);
-	INIT_WORK(&lkb->lkb_cb_work, dlm_callback_work);
 
-	idr_preload(GFP_NOFS);
-	spin_lock(&ls->ls_lkbidr_spin);
+	write_lock_bh(&ls->ls_lkbidr_lock);
 	rv = idr_alloc(&ls->ls_lkbidr, lkb, start, end, GFP_NOWAIT);
 	if (rv >= 0)
 		lkb->lkb_id = rv;
-	spin_unlock(&ls->ls_lkbidr_spin);
-	idr_preload_end();
+	write_unlock_bh(&ls->ls_lkbidr_lock);
 
 	if (rv < 0) {
 		log_error(ls, "create_lkb idr error %d", rv);
@@ -1235,11 +1547,11 @@ static int find_lkb(struct dlm_ls *ls, uint32_t lkid, struct dlm_lkb **lkb_ret)
 {
 	struct dlm_lkb *lkb;
 
-	spin_lock(&ls->ls_lkbidr_spin);
+	read_lock_bh(&ls->ls_lkbidr_lock);
 	lkb = idr_find(&ls->ls_lkbidr, lkid);
 	if (lkb)
 		kref_get(&lkb->lkb_ref);
-	spin_unlock(&ls->ls_lkbidr_spin);
+	read_unlock_bh(&ls->ls_lkbidr_lock);
 
 	*lkb_ret = lkb;
 	return lkb ? 0 : -ENOENT;
@@ -1263,11 +1575,11 @@ static int __put_lkb(struct dlm_ls *ls, struct dlm_lkb *lkb)
 	uint32_t lkid = lkb->lkb_id;
 	int rv;
 
-	rv = kref_put_lock(&lkb->lkb_ref, kill_lkb,
-			   &ls->ls_lkbidr_spin);
+	rv = dlm_kref_put_write_lock_bh(&lkb->lkb_ref, kill_lkb,
+					&ls->ls_lkbidr_lock);
 	if (rv) {
 		idr_remove(&ls->ls_lkbidr, lkid);
-		spin_unlock(&ls->ls_lkbidr_spin);
+		write_unlock_bh(&ls->ls_lkbidr_lock);
 
 		detach_lkb(lkb);
 
@@ -1407,9 +1719,8 @@ static int add_to_waiters(struct dlm_lkb *lkb, int mstype, int to_nodeid)
 {
 	struct dlm_ls *ls = lkb->lkb_resource->res_ls;
 	int error = 0;
-	int wc;
 
-	mutex_lock(&ls->ls_waiters_mutex);
+	spin_lock_bh(&ls->ls_waiters_lock);
 
 	if (is_overlap_unlock(lkb) ||
 	    (is_overlap_cancel(lkb) && (mstype == DLM_MSG_CANCEL))) {
@@ -1429,17 +1740,20 @@ static int add_to_waiters(struct dlm_lkb *lkb, int mstype, int to_nodeid)
 			error = -EBUSY;
 			goto out;
 		}
-		wc = atomic_inc_return(&lkb->lkb_wait_count);
+		lkb->lkb_wait_count++;
 		hold_lkb(lkb);
 
 		log_debug(ls, "addwait %x cur %d overlap %d count %d f %x",
-			  lkb->lkb_id, lkb->lkb_wait_type, mstype, wc,
-			  dlm_iflags_val(lkb));
+			  lkb->lkb_id, lkb->lkb_wait_type, mstype,
+			  lkb->lkb_wait_count, dlm_iflags_val(lkb));
 		goto out;
 	}
 
-	wc = atomic_fetch_inc(&lkb->lkb_wait_count);
-	DLM_ASSERT(!wc, dlm_print_lkb(lkb); printk("wait_count %d\n", wc););
+	DLM_ASSERT(!lkb->lkb_wait_count,
+		   dlm_print_lkb(lkb);
+		   printk("wait_count %d\n", lkb->lkb_wait_count););
+
+	lkb->lkb_wait_count++;
 	lkb->lkb_wait_type = mstype;
 	lkb->lkb_wait_nodeid = to_nodeid; /* for debugging */
 	hold_lkb(lkb);
@@ -1449,7 +1763,7 @@ static int add_to_waiters(struct dlm_lkb *lkb, int mstype, int to_nodeid)
 		log_error(ls, "addwait error %x %d flags %x %d %d %s",
 			  lkb->lkb_id, error, dlm_iflags_val(lkb), mstype,
 			  lkb->lkb_wait_type, lkb->lkb_resource->res_name);
-	mutex_unlock(&ls->ls_waiters_mutex);
+	spin_unlock_bh(&ls->ls_waiters_lock);
 	return error;
 }
 
@@ -1502,7 +1816,7 @@ static int _remove_from_waiters(struct dlm_lkb *lkb, int mstype,
 		log_debug(ls, "remwait %x convert_reply zap overlap_cancel",
 			  lkb->lkb_id);
 		lkb->lkb_wait_type = 0;
-		atomic_dec(&lkb->lkb_wait_count);
+		lkb->lkb_wait_count--;
 		unhold_lkb(lkb);
 		goto out_del;
 	}
@@ -1529,15 +1843,16 @@ static int _remove_from_waiters(struct dlm_lkb *lkb, int mstype,
 	if (overlap_done && lkb->lkb_wait_type) {
 		log_error(ls, "remwait error %x reply %d wait_type %d overlap",
 			  lkb->lkb_id, mstype, lkb->lkb_wait_type);
-		atomic_dec(&lkb->lkb_wait_count);
+		lkb->lkb_wait_count--;
 		unhold_lkb(lkb);
 		lkb->lkb_wait_type = 0;
 	}
 
-	DLM_ASSERT(atomic_read(&lkb->lkb_wait_count), dlm_print_lkb(lkb););
+	DLM_ASSERT(lkb->lkb_wait_count, dlm_print_lkb(lkb););
 
 	clear_bit(DLM_IFL_RESEND_BIT, &lkb->lkb_iflags);
-	if (atomic_dec_and_test(&lkb->lkb_wait_count))
+	lkb->lkb_wait_count--;
+	if (!lkb->lkb_wait_count)
 		list_del_init(&lkb->lkb_wait_reply);
 	unhold_lkb(lkb);
 	return 0;
@@ -1548,14 +1863,18 @@ static int remove_from_waiters(struct dlm_lkb *lkb, int mstype)
 	struct dlm_ls *ls = lkb->lkb_resource->res_ls;
 	int error;
 
-	mutex_lock(&ls->ls_waiters_mutex);
+	spin_lock_bh(&ls->ls_waiters_lock);
 	error = _remove_from_waiters(lkb, mstype, NULL);
-	mutex_unlock(&ls->ls_waiters_mutex);
+	spin_unlock_bh(&ls->ls_waiters_lock);
 	return error;
 }
 
 /* Handles situations where we might be processing a "fake" or "local" reply in
-   which we can't try to take waiters_mutex again. */
+ * the recovery context which stops any locking activity. Only debugfs might
+ * change the lockspace waiters but they will held the recovery lock to ensure
+ * remove_from_waiters_ms() in local case will be the only user manipulating the
+ * lockspace waiters in recovery context.
+ */
 
 static int remove_from_waiters_ms(struct dlm_lkb *lkb,
 				  const struct dlm_message *ms, bool local)
@@ -1564,159 +1883,16 @@ static int remove_from_waiters_ms(struct dlm_lkb *lkb,
 	int error;
 
 	if (!local)
-		mutex_lock(&ls->ls_waiters_mutex);
+		spin_lock_bh(&ls->ls_waiters_lock);
+	else
+		WARN_ON_ONCE(!rwsem_is_locked(&ls->ls_in_recovery) ||
+			     !dlm_locking_stopped(ls));
 	error = _remove_from_waiters(lkb, le32_to_cpu(ms->m_type), ms);
 	if (!local)
-		mutex_unlock(&ls->ls_waiters_mutex);
+		spin_unlock_bh(&ls->ls_waiters_lock);
 	return error;
 }
 
-static void shrink_bucket(struct dlm_ls *ls, int b)
-{
-	struct rb_node *n, *next;
-	struct dlm_rsb *r;
-	char *name;
-	int our_nodeid = dlm_our_nodeid();
-	int remote_count = 0;
-	int need_shrink = 0;
-	int i, len, rv;
-
-	memset(&ls->ls_remove_lens, 0, sizeof(int) * DLM_REMOVE_NAMES_MAX);
-
-	spin_lock(&ls->ls_rsbtbl[b].lock);
-
-	if (!test_bit(DLM_RTF_SHRINK_BIT, &ls->ls_rsbtbl[b].flags)) {
-		spin_unlock(&ls->ls_rsbtbl[b].lock);
-		return;
-	}
-
-	for (n = rb_first(&ls->ls_rsbtbl[b].toss); n; n = next) {
-		next = rb_next(n);
-		r = rb_entry(n, struct dlm_rsb, res_hashnode);
-
-		/* If we're the directory record for this rsb, and
-		   we're not the master of it, then we need to wait
-		   for the master node to send us a dir remove for
-		   before removing the dir record. */
-
-		if (!dlm_no_directory(ls) &&
-		    (r->res_master_nodeid != our_nodeid) &&
-		    (dlm_dir_nodeid(r) == our_nodeid)) {
-			continue;
-		}
-
-		need_shrink = 1;
-
-		if (!time_after_eq(jiffies, r->res_toss_time +
-				   dlm_config.ci_toss_secs * HZ)) {
-			continue;
-		}
-
-		if (!dlm_no_directory(ls) &&
-		    (r->res_master_nodeid == our_nodeid) &&
-		    (dlm_dir_nodeid(r) != our_nodeid)) {
-
-			/* We're the master of this rsb but we're not
-			   the directory record, so we need to tell the
-			   dir node to remove the dir record. */
-
-			ls->ls_remove_lens[remote_count] = r->res_length;
-			memcpy(ls->ls_remove_names[remote_count], r->res_name,
-			       DLM_RESNAME_MAXLEN);
-			remote_count++;
-
-			if (remote_count >= DLM_REMOVE_NAMES_MAX)
-				break;
-			continue;
-		}
-
-		if (!kref_put(&r->res_ref, kill_rsb)) {
-			log_error(ls, "tossed rsb in use %s", r->res_name);
-			continue;
-		}
-
-		rb_erase(&r->res_hashnode, &ls->ls_rsbtbl[b].toss);
-		dlm_free_rsb(r);
-	}
-
-	if (need_shrink)
-		set_bit(DLM_RTF_SHRINK_BIT, &ls->ls_rsbtbl[b].flags);
-	else
-		clear_bit(DLM_RTF_SHRINK_BIT, &ls->ls_rsbtbl[b].flags);
-	spin_unlock(&ls->ls_rsbtbl[b].lock);
-
-	/*
-	 * While searching for rsb's to free, we found some that require
-	 * remote removal.  We leave them in place and find them again here
-	 * so there is a very small gap between removing them from the toss
-	 * list and sending the removal.  Keeping this gap small is
-	 * important to keep us (the master node) from being out of sync
-	 * with the remote dir node for very long.
-	 */
-
-	for (i = 0; i < remote_count; i++) {
-		name = ls->ls_remove_names[i];
-		len = ls->ls_remove_lens[i];
-
-		spin_lock(&ls->ls_rsbtbl[b].lock);
-		rv = dlm_search_rsb_tree(&ls->ls_rsbtbl[b].toss, name, len, &r);
-		if (rv) {
-			spin_unlock(&ls->ls_rsbtbl[b].lock);
-			log_debug(ls, "remove_name not toss %s", name);
-			continue;
-		}
-
-		if (r->res_master_nodeid != our_nodeid) {
-			spin_unlock(&ls->ls_rsbtbl[b].lock);
-			log_debug(ls, "remove_name master %d dir %d our %d %s",
-				  r->res_master_nodeid, r->res_dir_nodeid,
-				  our_nodeid, name);
-			continue;
-		}
-
-		if (r->res_dir_nodeid == our_nodeid) {
-			/* should never happen */
-			spin_unlock(&ls->ls_rsbtbl[b].lock);
-			log_error(ls, "remove_name dir %d master %d our %d %s",
-				  r->res_dir_nodeid, r->res_master_nodeid,
-				  our_nodeid, name);
-			continue;
-		}
-
-		if (!time_after_eq(jiffies, r->res_toss_time +
-				   dlm_config.ci_toss_secs * HZ)) {
-			spin_unlock(&ls->ls_rsbtbl[b].lock);
-			log_debug(ls, "remove_name toss_time %lu now %lu %s",
-				  r->res_toss_time, jiffies, name);
-			continue;
-		}
-
-		if (!kref_put(&r->res_ref, kill_rsb)) {
-			spin_unlock(&ls->ls_rsbtbl[b].lock);
-			log_error(ls, "remove_name in use %s", name);
-			continue;
-		}
-
-		rb_erase(&r->res_hashnode, &ls->ls_rsbtbl[b].toss);
-		send_remove(r);
-		spin_unlock(&ls->ls_rsbtbl[b].lock);
-
-		dlm_free_rsb(r);
-	}
-}
-
-void dlm_scan_rsbs(struct dlm_ls *ls)
-{
-	int i;
-
-	for (i = 0; i < ls->ls_rsbtbl_size; i++) {
-		shrink_bucket(ls, i);
-		if (dlm_locking_stopped(ls))
-			break;
-		cond_resched();
-	}
-}
-
 /* lkb is master or local copy */
 
 static void set_lvb_lock(struct dlm_rsb *r, struct dlm_lkb *lkb)
@@ -2535,7 +2711,6 @@ static void process_lookup_list(struct dlm_rsb *r)
 	list_for_each_entry_safe(lkb, safe, &r->res_lookup, lkb_rsb_lookup) {
 		list_del_init(&lkb->lkb_rsb_lookup);
 		_request_lock(r, lkb);
-		schedule();
 	}
 }
 
@@ -2666,7 +2841,7 @@ static int validate_lock_args(struct dlm_ls *ls, struct dlm_lkb *lkb,
 			goto out;
 
 		/* lock not allowed if there's any op in progress */
-		if (lkb->lkb_wait_type || atomic_read(&lkb->lkb_wait_count))
+		if (lkb->lkb_wait_type || lkb->lkb_wait_count)
 			goto out;
 
 		if (is_overlap(lkb))
@@ -2728,7 +2903,7 @@ static int validate_unlock_args(struct dlm_lkb *lkb, struct dlm_args *args)
 
 	/* normal unlock not allowed if there's any op in progress */
 	if (!(args->flags & (DLM_LKF_CANCEL | DLM_LKF_FORCEUNLOCK)) &&
-	    (lkb->lkb_wait_type || atomic_read(&lkb->lkb_wait_count)))
+	    (lkb->lkb_wait_type || lkb->lkb_wait_count))
 		goto out;
 
 	/* an lkb may be waiting for an rsb lookup to complete where the
@@ -3329,8 +3504,7 @@ int dlm_unlock(dlm_lockspace_t *lockspace,
 static int _create_message(struct dlm_ls *ls, int mb_len,
 			   int to_nodeid, int mstype,
 			   struct dlm_message **ms_ret,
-			   struct dlm_mhandle **mh_ret,
-			   gfp_t allocation)
+			   struct dlm_mhandle **mh_ret)
 {
 	struct dlm_message *ms;
 	struct dlm_mhandle *mh;
@@ -3340,7 +3514,7 @@ static int _create_message(struct dlm_ls *ls, int mb_len,
 	   pass into midcomms_commit and a message buffer (mb) that we
 	   write our data into */
 
-	mh = dlm_midcomms_get_mhandle(to_nodeid, mb_len, allocation, &mb);
+	mh = dlm_midcomms_get_mhandle(to_nodeid, mb_len, &mb);
 	if (!mh)
 		return -ENOBUFS;
 
@@ -3362,8 +3536,7 @@ static int _create_message(struct dlm_ls *ls, int mb_len,
 static int create_message(struct dlm_rsb *r, struct dlm_lkb *lkb,
 			  int to_nodeid, int mstype,
 			  struct dlm_message **ms_ret,
-			  struct dlm_mhandle **mh_ret,
-			  gfp_t allocation)
+			  struct dlm_mhandle **mh_ret)
 {
 	int mb_len = sizeof(struct dlm_message);
 
@@ -3384,7 +3557,7 @@ static int create_message(struct dlm_rsb *r, struct dlm_lkb *lkb,
 	}
 
 	return _create_message(r->res_ls, mb_len, to_nodeid, mstype,
-			       ms_ret, mh_ret, allocation);
+			       ms_ret, mh_ret);
 }
 
 /* further lowcomms enhancements or alternate implementations may make
@@ -3453,7 +3626,7 @@ static int send_common(struct dlm_rsb *r, struct dlm_lkb *lkb, int mstype)
 	if (error)
 		return error;
 
-	error = create_message(r, lkb, to_nodeid, mstype, &ms, &mh, GFP_NOFS);
+	error = create_message(r, lkb, to_nodeid, mstype, &ms, &mh);
 	if (error)
 		goto fail;
 
@@ -3513,8 +3686,7 @@ static int send_grant(struct dlm_rsb *r, struct dlm_lkb *lkb)
 
 	to_nodeid = lkb->lkb_nodeid;
 
-	error = create_message(r, lkb, to_nodeid, DLM_MSG_GRANT, &ms, &mh,
-			       GFP_NOFS);
+	error = create_message(r, lkb, to_nodeid, DLM_MSG_GRANT, &ms, &mh);
 	if (error)
 		goto out;
 
@@ -3535,8 +3707,7 @@ static int send_bast(struct dlm_rsb *r, struct dlm_lkb *lkb, int mode)
 
 	to_nodeid = lkb->lkb_nodeid;
 
-	error = create_message(r, NULL, to_nodeid, DLM_MSG_BAST, &ms, &mh,
-			       GFP_NOFS);
+	error = create_message(r, NULL, to_nodeid, DLM_MSG_BAST, &ms, &mh);
 	if (error)
 		goto out;
 
@@ -3561,8 +3732,7 @@ static int send_lookup(struct dlm_rsb *r, struct dlm_lkb *lkb)
 	if (error)
 		return error;
 
-	error = create_message(r, NULL, to_nodeid, DLM_MSG_LOOKUP, &ms, &mh,
-			       GFP_NOFS);
+	error = create_message(r, NULL, to_nodeid, DLM_MSG_LOOKUP, &ms, &mh);
 	if (error)
 		goto fail;
 
@@ -3586,8 +3756,7 @@ static int send_remove(struct dlm_rsb *r)
 
 	to_nodeid = dlm_dir_nodeid(r);
 
-	error = create_message(r, NULL, to_nodeid, DLM_MSG_REMOVE, &ms, &mh,
-			       GFP_ATOMIC);
+	error = create_message(r, NULL, to_nodeid, DLM_MSG_REMOVE, &ms, &mh);
 	if (error)
 		goto out;
 
@@ -3608,7 +3777,7 @@ static int send_common_reply(struct dlm_rsb *r, struct dlm_lkb *lkb,
 
 	to_nodeid = lkb->lkb_nodeid;
 
-	error = create_message(r, lkb, to_nodeid, mstype, &ms, &mh, GFP_NOFS);
+	error = create_message(r, lkb, to_nodeid, mstype, &ms, &mh);
 	if (error)
 		goto out;
 
@@ -3650,8 +3819,7 @@ static int send_lookup_reply(struct dlm_ls *ls,
 	struct dlm_mhandle *mh;
 	int error, nodeid = le32_to_cpu(ms_in->m_header.h_nodeid);
 
-	error = create_message(r, NULL, nodeid, DLM_MSG_LOOKUP_REPLY, &ms, &mh,
-			       GFP_NOFS);
+	error = create_message(r, NULL, nodeid, DLM_MSG_LOOKUP_REPLY, &ms, &mh);
 	if (error)
 		goto out;
 
@@ -4136,7 +4304,6 @@ static void receive_remove(struct dlm_ls *ls, const struct dlm_message *ms)
 {
 	char name[DLM_RESNAME_MAXLEN+1];
 	struct dlm_rsb *r;
-	uint32_t hash, b;
 	int rv, len, dir_nodeid, from_nodeid;
 
 	from_nodeid = le32_to_cpu(ms->m_header.h_nodeid);
@@ -4156,47 +4323,44 @@ static void receive_remove(struct dlm_ls *ls, const struct dlm_message *ms)
 		return;
 	}
 
-	/* Look for name on rsbtbl.toss, if it's there, kill it.
-	   If it's on rsbtbl.keep, it's being used, and we should ignore this
-	   message.  This is an expected race between the dir node sending a
-	   request to the master node at the same time as the master node sends
-	   a remove to the dir node.  The resolution to that race is for the
-	   dir node to ignore the remove message, and the master node to
-	   recreate the master rsb when it gets a request from the dir node for
-	   an rsb it doesn't have. */
+	/* Look for name in rsb toss state, if it's there, kill it.
+	 * If it's in non toss state, it's being used, and we should ignore this
+	 * message.  This is an expected race between the dir node sending a
+	 * request to the master node at the same time as the master node sends
+	 * a remove to the dir node.  The resolution to that race is for the
+	 * dir node to ignore the remove message, and the master node to
+	 * recreate the master rsb when it gets a request from the dir node for
+	 * an rsb it doesn't have.
+	 */
 
 	memset(name, 0, sizeof(name));
 	memcpy(name, ms->m_extra, len);
 
-	hash = jhash(name, len, 0);
-	b = hash & (ls->ls_rsbtbl_size - 1);
-
-	spin_lock(&ls->ls_rsbtbl[b].lock);
+	write_lock_bh(&ls->ls_rsbtbl_lock);
 
-	rv = dlm_search_rsb_tree(&ls->ls_rsbtbl[b].toss, name, len, &r);
+	rv = dlm_search_rsb_tree(&ls->ls_rsbtbl, name, len, &r);
 	if (rv) {
-		/* verify the rsb is on keep list per comment above */
-		rv = dlm_search_rsb_tree(&ls->ls_rsbtbl[b].keep, name, len, &r);
-		if (rv) {
-			/* should not happen */
-			log_error(ls, "receive_remove from %d not found %s",
-				  from_nodeid, name);
-			spin_unlock(&ls->ls_rsbtbl[b].lock);
-			return;
-		}
+		/* should not happen */
+		log_error(ls, "%s from %d not found %s", __func__,
+			  from_nodeid, name);
+		write_unlock_bh(&ls->ls_rsbtbl_lock);
+		return;
+	}
+
+	if (!rsb_flag(r, RSB_TOSS)) {
 		if (r->res_master_nodeid != from_nodeid) {
 			/* should not happen */
 			log_error(ls, "receive_remove keep from %d master %d",
 				  from_nodeid, r->res_master_nodeid);
 			dlm_print_rsb(r);
-			spin_unlock(&ls->ls_rsbtbl[b].lock);
+			write_unlock_bh(&ls->ls_rsbtbl_lock);
 			return;
 		}
 
 		log_debug(ls, "receive_remove from %d master %d first %x %s",
 			  from_nodeid, r->res_master_nodeid, r->res_first_lkid,
 			  name);
-		spin_unlock(&ls->ls_rsbtbl[b].lock);
+		write_unlock_bh(&ls->ls_rsbtbl_lock);
 		return;
 	}
 
@@ -4204,20 +4368,16 @@ static void receive_remove(struct dlm_ls *ls, const struct dlm_message *ms)
 		log_error(ls, "receive_remove toss from %d master %d",
 			  from_nodeid, r->res_master_nodeid);
 		dlm_print_rsb(r);
-		spin_unlock(&ls->ls_rsbtbl[b].lock);
+		write_unlock_bh(&ls->ls_rsbtbl_lock);
 		return;
 	}
 
-	if (kref_put(&r->res_ref, kill_rsb)) {
-		rb_erase(&r->res_hashnode, &ls->ls_rsbtbl[b].toss);
-		spin_unlock(&ls->ls_rsbtbl[b].lock);
-		dlm_free_rsb(r);
-	} else {
-		log_error(ls, "receive_remove from %d rsb ref error",
-			  from_nodeid);
-		dlm_print_rsb(r);
-		spin_unlock(&ls->ls_rsbtbl[b].lock);
-	}
+	list_del(&r->res_rsbs_list);
+	rhashtable_remove_fast(&ls->ls_rsbtbl, &r->res_node,
+			       dlm_rhash_rsb_params);
+	write_unlock_bh(&ls->ls_rsbtbl_lock);
+
+	free_toss_rsb(r);
 }
 
 static void receive_purge(struct dlm_ls *ls, const struct dlm_message *ms)
@@ -4404,7 +4564,6 @@ static void _receive_convert_reply(struct dlm_lkb *lkb,
 	if (error)
 		goto out;
 
-	/* local reply can happen with waiters_mutex held */
 	error = remove_from_waiters_ms(lkb, ms, local);
 	if (error)
 		goto out;
@@ -4443,7 +4602,6 @@ static void _receive_unlock_reply(struct dlm_lkb *lkb,
 	if (error)
 		goto out;
 
-	/* local reply can happen with waiters_mutex held */
 	error = remove_from_waiters_ms(lkb, ms, local);
 	if (error)
 		goto out;
@@ -4495,7 +4653,6 @@ static void _receive_cancel_reply(struct dlm_lkb *lkb,
 	if (error)
 		goto out;
 
-	/* local reply can happen with waiters_mutex held */
 	error = remove_from_waiters_ms(lkb, ms, local);
 	if (error)
 		goto out;
@@ -4754,20 +4911,32 @@ static void _receive_message(struct dlm_ls *ls, const struct dlm_message *ms,
 static void dlm_receive_message(struct dlm_ls *ls, const struct dlm_message *ms,
 				int nodeid)
 {
-	if (dlm_locking_stopped(ls)) {
+try_again:
+	read_lock_bh(&ls->ls_requestqueue_lock);
+	if (test_bit(LSFL_RECV_MSG_BLOCKED, &ls->ls_flags)) {
 		/* If we were a member of this lockspace, left, and rejoined,
 		   other nodes may still be sending us messages from the
 		   lockspace generation before we left. */
 		if (WARN_ON_ONCE(!ls->ls_generation)) {
+			read_unlock_bh(&ls->ls_requestqueue_lock);
 			log_limit(ls, "receive %d from %d ignore old gen",
 				  le32_to_cpu(ms->m_type), nodeid);
 			return;
 		}
 
+		read_unlock_bh(&ls->ls_requestqueue_lock);
+		write_lock_bh(&ls->ls_requestqueue_lock);
+		/* recheck because we hold writelock now */
+		if (!test_bit(LSFL_RECV_MSG_BLOCKED, &ls->ls_flags)) {
+			write_unlock_bh(&ls->ls_requestqueue_lock);
+			goto try_again;
+		}
+
 		dlm_add_requestqueue(ls, nodeid, ms);
+		write_unlock_bh(&ls->ls_requestqueue_lock);
 	} else {
-		dlm_wait_requestqueue(ls);
 		_receive_message(ls, ms, 0);
+		read_unlock_bh(&ls->ls_requestqueue_lock);
 	}
 }
 
@@ -4827,7 +4996,7 @@ void dlm_receive_buffer(const union dlm_packet *p, int nodeid)
 	/* this rwsem allows dlm_ls_stop() to wait for all dlm_recv threads to
 	   be inactive (in this ls) before transitioning to recovery mode */
 
-	down_read(&ls->ls_recv_active);
+	read_lock_bh(&ls->ls_recv_active);
 	if (hd->h_cmd == DLM_MSG)
 		dlm_receive_message(ls, &p->message, nodeid);
 	else if (hd->h_cmd == DLM_RCOM)
@@ -4835,7 +5004,7 @@ void dlm_receive_buffer(const union dlm_packet *p, int nodeid)
 	else
 		log_error(ls, "invalid h_cmd %d from %d lockspace %x",
 			  hd->h_cmd, nodeid, le32_to_cpu(hd->u.h_lockspace));
-	up_read(&ls->ls_recv_active);
+	read_unlock_bh(&ls->ls_recv_active);
 
 	dlm_put_lockspace(ls);
 }
@@ -4896,8 +5065,6 @@ void dlm_recover_waiters_pre(struct dlm_ls *ls)
 	if (!ms_local)
 		return;
 
-	mutex_lock(&ls->ls_waiters_mutex);
-
 	list_for_each_entry_safe(lkb, safe, &ls->ls_waiters, lkb_wait_reply) {
 
 		dir_nodeid = dlm_dir_nodeid(lkb->lkb_resource);
@@ -4990,7 +5157,6 @@ void dlm_recover_waiters_pre(struct dlm_ls *ls)
 		}
 		schedule();
 	}
-	mutex_unlock(&ls->ls_waiters_mutex);
 	kfree(ms_local);
 }
 
@@ -4998,7 +5164,7 @@ static struct dlm_lkb *find_resend_waiter(struct dlm_ls *ls)
 {
 	struct dlm_lkb *lkb = NULL, *iter;
 
-	mutex_lock(&ls->ls_waiters_mutex);
+	spin_lock_bh(&ls->ls_waiters_lock);
 	list_for_each_entry(iter, &ls->ls_waiters, lkb_wait_reply) {
 		if (test_bit(DLM_IFL_RESEND_BIT, &iter->lkb_iflags)) {
 			hold_lkb(iter);
@@ -5006,26 +5172,37 @@ static struct dlm_lkb *find_resend_waiter(struct dlm_ls *ls)
 			break;
 		}
 	}
-	mutex_unlock(&ls->ls_waiters_mutex);
+	spin_unlock_bh(&ls->ls_waiters_lock);
 
 	return lkb;
 }
 
-/* Deal with lookups and lkb's marked RESEND from _pre.  We may now be the
-   master or dir-node for r.  Processing the lkb may result in it being placed
-   back on waiters. */
-
-/* We do this after normal locking has been enabled and any saved messages
-   (in requestqueue) have been processed.  We should be confident that at
-   this point we won't get or process a reply to any of these waiting
-   operations.  But, new ops may be coming in on the rsbs/locks here from
-   userspace or remotely. */
-
-/* there may have been an overlap unlock/cancel prior to recovery or after
-   recovery.  if before, the lkb may still have a pos wait_count; if after, the
-   overlap flag would just have been set and nothing new sent.  we can be
-   confident here than any replies to either the initial op or overlap ops
-   prior to recovery have been received. */
+/*
+ * Forced state reset for locks that were in the middle of remote operations
+ * when recovery happened (i.e. lkbs that were on the waiters list, waiting
+ * for a reply from a remote operation.)  The lkbs remaining on the waiters
+ * list need to be reevaluated; some may need resending to a different node
+ * than previously, and some may now need local handling rather than remote.
+ *
+ * First, the lkb state for the voided remote operation is forcibly reset,
+ * equivalent to what remove_from_waiters() would normally do:
+ * . lkb removed from ls_waiters list
+ * . lkb wait_type cleared
+ * . lkb waiters_count cleared
+ * . lkb ref count decremented for each waiters_count (almost always 1,
+ *   but possibly 2 in case of cancel/unlock overlapping, which means
+ *   two remote replies were being expected for the lkb.)
+ *
+ * Second, the lkb is reprocessed like an original operation would be,
+ * by passing it to _request_lock or _convert_lock, which will either
+ * process the lkb operation locally, or send it to a remote node again
+ * and put the lkb back onto the waiters list.
+ *
+ * When reprocessing the lkb, we may find that it's flagged for an overlapping
+ * force-unlock or cancel, either from before recovery began, or after recovery
+ * finished.  If this is the case, the unlock/cancel is done directly, and the
+ * original operation is not initiated again (no _request_lock/_convert_lock.)
+ */
 
 int dlm_recover_waiters_post(struct dlm_ls *ls)
 {
@@ -5040,6 +5217,11 @@ int dlm_recover_waiters_post(struct dlm_ls *ls)
 			break;
 		}
 
+		/* 
+		 * Find an lkb from the waiters list that's been affected by
+		 * recovery node changes, and needs to be reprocessed.  Does
+		 * hold_lkb(), adding a refcount.
+		 */
 		lkb = find_resend_waiter(ls);
 		if (!lkb)
 			break;
@@ -5048,6 +5230,11 @@ int dlm_recover_waiters_post(struct dlm_ls *ls)
 		hold_rsb(r);
 		lock_rsb(r);
 
+		/*
+		 * If the lkb has been flagged for a force unlock or cancel,
+		 * then the reprocessing below will be replaced by just doing
+		 * the unlock/cancel directly.
+		 */
 		mstype = lkb->lkb_wait_type;
 		oc = test_and_clear_bit(DLM_IFL_OVERLAP_CANCEL_BIT,
 					&lkb->lkb_iflags);
@@ -5061,21 +5248,39 @@ int dlm_recover_waiters_post(struct dlm_ls *ls)
 			  r->res_nodeid, lkb->lkb_nodeid, lkb->lkb_wait_nodeid,
 			  dlm_dir_nodeid(r), oc, ou);
 
-		/* At this point we assume that we won't get a reply to any
-		   previous op or overlap op on this lock.  First, do a big
-		   remove_from_waiters() for all previous ops. */
+		/*
+		 * No reply to the pre-recovery operation will now be received,
+		 * so a forced equivalent of remove_from_waiters() is needed to
+		 * reset the waiters state that was in place before recovery.
+		 */
 
 		clear_bit(DLM_IFL_RESEND_BIT, &lkb->lkb_iflags);
+
+		/* Forcibly clear wait_type */
 		lkb->lkb_wait_type = 0;
-		/* drop all wait_count references we still
-		 * hold a reference for this iteration.
+
+		/*
+		 * Forcibly reset wait_count and associated refcount.  The
+		 * wait_count will almost always be 1, but in case of an
+		 * overlapping unlock/cancel it could be 2: see where
+		 * add_to_waiters() finds the lkb is already on the waiters
+		 * list and does lkb_wait_count++; hold_lkb().
 		 */
-		while (!atomic_dec_and_test(&lkb->lkb_wait_count))
+		while (lkb->lkb_wait_count) {
+			lkb->lkb_wait_count--;
 			unhold_lkb(lkb);
+		}
 
-		mutex_lock(&ls->ls_waiters_mutex);
+		/* Forcibly remove from waiters list */
+		spin_lock_bh(&ls->ls_waiters_lock);
 		list_del_init(&lkb->lkb_wait_reply);
-		mutex_unlock(&ls->ls_waiters_mutex);
+		spin_unlock_bh(&ls->ls_waiters_lock);
+
+		/*
+		 * The lkb is now clear of all prior waiters state and can be
+		 * processed locally, or sent to remote node again, or directly
+		 * cancelled/unlocked.
+		 */
 
 		if (oc || ou) {
 			/* do an unlock or cancel instead of resending */
@@ -5194,7 +5399,7 @@ static void purge_dead_list(struct dlm_ls *ls, struct dlm_rsb *r,
 
 /* Get rid of locks held by nodes that are gone. */
 
-void dlm_recover_purge(struct dlm_ls *ls)
+void dlm_recover_purge(struct dlm_ls *ls, const struct list_head *root_list)
 {
 	struct dlm_rsb *r;
 	struct dlm_member *memb;
@@ -5213,8 +5418,7 @@ void dlm_recover_purge(struct dlm_ls *ls)
 	if (!nodes_count)
 		return;
 
-	down_write(&ls->ls_root_sem);
-	list_for_each_entry(r, &ls->ls_root_list, res_root_list) {
+	list_for_each_entry(r, root_list, res_root_list) {
 		hold_rsb(r);
 		lock_rsb(r);
 		if (is_master(r)) {
@@ -5229,22 +5433,18 @@ void dlm_recover_purge(struct dlm_ls *ls)
 		unhold_rsb(r);
 		cond_resched();
 	}
-	up_write(&ls->ls_root_sem);
 
 	if (lkb_count)
 		log_rinfo(ls, "dlm_recover_purge %u locks for %u nodes",
 			  lkb_count, nodes_count);
 }
 
-static struct dlm_rsb *find_grant_rsb(struct dlm_ls *ls, int bucket)
+static struct dlm_rsb *find_grant_rsb(struct dlm_ls *ls)
 {
-	struct rb_node *n;
 	struct dlm_rsb *r;
 
-	spin_lock(&ls->ls_rsbtbl[bucket].lock);
-	for (n = rb_first(&ls->ls_rsbtbl[bucket].keep); n; n = rb_next(n)) {
-		r = rb_entry(n, struct dlm_rsb, res_hashnode);
-
+	read_lock_bh(&ls->ls_rsbtbl_lock);
+	list_for_each_entry(r, &ls->ls_keep, res_rsbs_list) {
 		if (!rsb_flag(r, RSB_RECOVER_GRANT))
 			continue;
 		if (!is_master(r)) {
@@ -5252,10 +5452,10 @@ static struct dlm_rsb *find_grant_rsb(struct dlm_ls *ls, int bucket)
 			continue;
 		}
 		hold_rsb(r);
-		spin_unlock(&ls->ls_rsbtbl[bucket].lock);
+		read_unlock_bh(&ls->ls_rsbtbl_lock);
 		return r;
 	}
-	spin_unlock(&ls->ls_rsbtbl[bucket].lock);
+	read_unlock_bh(&ls->ls_rsbtbl_lock);
 	return NULL;
 }
 
@@ -5279,19 +5479,15 @@ static struct dlm_rsb *find_grant_rsb(struct dlm_ls *ls, int bucket)
 void dlm_recover_grant(struct dlm_ls *ls)
 {
 	struct dlm_rsb *r;
-	int bucket = 0;
 	unsigned int count = 0;
 	unsigned int rsb_count = 0;
 	unsigned int lkb_count = 0;
 
 	while (1) {
-		r = find_grant_rsb(ls, bucket);
-		if (!r) {
-			if (bucket == ls->ls_rsbtbl_size - 1)
-				break;
-			bucket++;
-			continue;
-		}
+		r = find_grant_rsb(ls);
+		if (!r)
+			break;
+
 		rsb_count++;
 		count = 0;
 		lock_rsb(r);
@@ -5599,10 +5795,10 @@ int dlm_user_request(struct dlm_ls *ls, struct dlm_user_args *ua,
 	}
 
 	/* add this new lkb to the per-process list of locks */
-	spin_lock(&ua->proc->locks_spin);
+	spin_lock_bh(&ua->proc->locks_spin);
 	hold_lkb(lkb);
 	list_add_tail(&lkb->lkb_ownqueue, &ua->proc->locks);
-	spin_unlock(&ua->proc->locks_spin);
+	spin_unlock_bh(&ua->proc->locks_spin);
 	do_put = false;
  out_put:
 	trace_dlm_lock_end(ls, lkb, name, namelen, mode, flags, error, false);
@@ -5684,7 +5880,7 @@ int dlm_user_adopt_orphan(struct dlm_ls *ls, struct dlm_user_args *ua_tmp,
 	int found_other_mode = 0;
 	int rv = 0;
 
-	mutex_lock(&ls->ls_orphans_mutex);
+	spin_lock_bh(&ls->ls_orphans_lock);
 	list_for_each_entry(iter, &ls->ls_orphans, lkb_ownqueue) {
 		if (iter->lkb_resource->res_length != namelen)
 			continue;
@@ -5701,7 +5897,7 @@ int dlm_user_adopt_orphan(struct dlm_ls *ls, struct dlm_user_args *ua_tmp,
 		*lkid = iter->lkb_id;
 		break;
 	}
-	mutex_unlock(&ls->ls_orphans_mutex);
+	spin_unlock_bh(&ls->ls_orphans_lock);
 
 	if (!lkb && found_other_mode) {
 		rv = -EAGAIN;
@@ -5732,9 +5928,9 @@ int dlm_user_adopt_orphan(struct dlm_ls *ls, struct dlm_user_args *ua_tmp,
 	 * for the proc locks list.
 	 */
 
-	spin_lock(&ua->proc->locks_spin);
+	spin_lock_bh(&ua->proc->locks_spin);
 	list_add_tail(&lkb->lkb_ownqueue, &ua->proc->locks);
-	spin_unlock(&ua->proc->locks_spin);
+	spin_unlock_bh(&ua->proc->locks_spin);
  out:
 	kfree(ua_tmp);
 	return rv;
@@ -5778,11 +5974,11 @@ int dlm_user_unlock(struct dlm_ls *ls, struct dlm_user_args *ua_tmp,
 	if (error)
 		goto out_put;
 
-	spin_lock(&ua->proc->locks_spin);
+	spin_lock_bh(&ua->proc->locks_spin);
 	/* dlm_user_add_cb() may have already taken lkb off the proc list */
 	if (!list_empty(&lkb->lkb_ownqueue))
 		list_move(&lkb->lkb_ownqueue, &ua->proc->unlocking);
-	spin_unlock(&ua->proc->locks_spin);
+	spin_unlock_bh(&ua->proc->locks_spin);
  out_put:
 	trace_dlm_unlock_end(ls, lkb, flags, error);
 	dlm_put_lkb(lkb);
@@ -5893,9 +6089,9 @@ static int orphan_proc_lock(struct dlm_ls *ls, struct dlm_lkb *lkb)
 	int error;
 
 	hold_lkb(lkb); /* reference for the ls_orphans list */
-	mutex_lock(&ls->ls_orphans_mutex);
+	spin_lock_bh(&ls->ls_orphans_lock);
 	list_add_tail(&lkb->lkb_ownqueue, &ls->ls_orphans);
-	mutex_unlock(&ls->ls_orphans_mutex);
+	spin_unlock_bh(&ls->ls_orphans_lock);
 
 	set_unlock_args(0, lkb->lkb_ua, &args);
 
@@ -5933,7 +6129,7 @@ static struct dlm_lkb *del_proc_lock(struct dlm_ls *ls,
 {
 	struct dlm_lkb *lkb = NULL;
 
-	spin_lock(&ls->ls_clear_proc_locks);
+	spin_lock_bh(&ls->ls_clear_proc_locks);
 	if (list_empty(&proc->locks))
 		goto out;
 
@@ -5945,7 +6141,7 @@ static struct dlm_lkb *del_proc_lock(struct dlm_ls *ls,
 	else
 		set_bit(DLM_IFL_DEAD_BIT, &lkb->lkb_iflags);
  out:
-	spin_unlock(&ls->ls_clear_proc_locks);
+	spin_unlock_bh(&ls->ls_clear_proc_locks);
 	return lkb;
 }
 
@@ -5961,6 +6157,7 @@ static struct dlm_lkb *del_proc_lock(struct dlm_ls *ls,
 
 void dlm_clear_proc_locks(struct dlm_ls *ls, struct dlm_user_proc *proc)
 {
+	struct dlm_callback *cb, *cb_safe;
 	struct dlm_lkb *lkb, *safe;
 
 	dlm_lock_recovery(ls);
@@ -5981,7 +6178,7 @@ void dlm_clear_proc_locks(struct dlm_ls *ls, struct dlm_user_proc *proc)
 		dlm_put_lkb(lkb);
 	}
 
-	spin_lock(&ls->ls_clear_proc_locks);
+	spin_lock_bh(&ls->ls_clear_proc_locks);
 
 	/* in-progress unlocks */
 	list_for_each_entry_safe(lkb, safe, &proc->unlocking, lkb_ownqueue) {
@@ -5990,29 +6187,29 @@ void dlm_clear_proc_locks(struct dlm_ls *ls, struct dlm_user_proc *proc)
 		dlm_put_lkb(lkb);
 	}
 
-	list_for_each_entry_safe(lkb, safe, &proc->asts, lkb_cb_list) {
-		dlm_purge_lkb_callbacks(lkb);
-		list_del_init(&lkb->lkb_cb_list);
-		dlm_put_lkb(lkb);
+	list_for_each_entry_safe(cb, cb_safe, &proc->asts, list) {
+		list_del(&cb->list);
+		dlm_free_cb(cb);
 	}
 
-	spin_unlock(&ls->ls_clear_proc_locks);
+	spin_unlock_bh(&ls->ls_clear_proc_locks);
 	dlm_unlock_recovery(ls);
 }
 
 static void purge_proc_locks(struct dlm_ls *ls, struct dlm_user_proc *proc)
 {
+	struct dlm_callback *cb, *cb_safe;
 	struct dlm_lkb *lkb, *safe;
 
 	while (1) {
 		lkb = NULL;
-		spin_lock(&proc->locks_spin);
+		spin_lock_bh(&proc->locks_spin);
 		if (!list_empty(&proc->locks)) {
 			lkb = list_entry(proc->locks.next, struct dlm_lkb,
 					 lkb_ownqueue);
 			list_del_init(&lkb->lkb_ownqueue);
 		}
-		spin_unlock(&proc->locks_spin);
+		spin_unlock_bh(&proc->locks_spin);
 
 		if (!lkb)
 			break;
@@ -6022,21 +6219,20 @@ static void purge_proc_locks(struct dlm_ls *ls, struct dlm_user_proc *proc)
 		dlm_put_lkb(lkb); /* ref from proc->locks list */
 	}
 
-	spin_lock(&proc->locks_spin);
+	spin_lock_bh(&proc->locks_spin);
 	list_for_each_entry_safe(lkb, safe, &proc->unlocking, lkb_ownqueue) {
 		list_del_init(&lkb->lkb_ownqueue);
 		set_bit(DLM_IFL_DEAD_BIT, &lkb->lkb_iflags);
 		dlm_put_lkb(lkb);
 	}
-	spin_unlock(&proc->locks_spin);
+	spin_unlock_bh(&proc->locks_spin);
 
-	spin_lock(&proc->asts_spin);
-	list_for_each_entry_safe(lkb, safe, &proc->asts, lkb_cb_list) {
-		dlm_purge_lkb_callbacks(lkb);
-		list_del_init(&lkb->lkb_cb_list);
-		dlm_put_lkb(lkb);
+	spin_lock_bh(&proc->asts_spin);
+	list_for_each_entry_safe(cb, cb_safe, &proc->asts, list) {
+		list_del(&cb->list);
+		dlm_free_cb(cb);
 	}
-	spin_unlock(&proc->asts_spin);
+	spin_unlock_bh(&proc->asts_spin);
 }
 
 /* pid of 0 means purge all orphans */
@@ -6045,7 +6241,7 @@ static void do_purge(struct dlm_ls *ls, int nodeid, int pid)
 {
 	struct dlm_lkb *lkb, *safe;
 
-	mutex_lock(&ls->ls_orphans_mutex);
+	spin_lock_bh(&ls->ls_orphans_lock);
 	list_for_each_entry_safe(lkb, safe, &ls->ls_orphans, lkb_ownqueue) {
 		if (pid && lkb->lkb_ownpid != pid)
 			continue;
@@ -6053,7 +6249,7 @@ static void do_purge(struct dlm_ls *ls, int nodeid, int pid)
 		list_del_init(&lkb->lkb_ownqueue);
 		dlm_put_lkb(lkb);
 	}
-	mutex_unlock(&ls->ls_orphans_mutex);
+	spin_unlock_bh(&ls->ls_orphans_lock);
 }
 
 static int send_purge(struct dlm_ls *ls, int nodeid, int pid)
@@ -6063,7 +6259,7 @@ static int send_purge(struct dlm_ls *ls, int nodeid, int pid)
 	int error;
 
 	error = _create_message(ls, sizeof(struct dlm_message), nodeid,
-				DLM_MSG_PURGE, &ms, &mh, GFP_NOFS);
+				DLM_MSG_PURGE, &ms, &mh);
 	if (error)
 		return error;
 	ms->m_nodeid = cpu_to_le32(nodeid);
diff --git a/fs/dlm/lock.h b/fs/dlm/lock.h
index b54e2cbbe6..8de9dee4c0 100644
--- a/fs/dlm/lock.h
+++ b/fs/dlm/lock.h
@@ -11,6 +11,7 @@
 #ifndef __LOCK_DOT_H__
 #define __LOCK_DOT_H__
 
+void dlm_rsb_toss_timer(struct timer_list *timer);
 void dlm_dump_rsb(struct dlm_rsb *r);
 void dlm_dump_rsb_name(struct dlm_ls *ls, const char *name, int len);
 void dlm_print_lkb(struct dlm_lkb *lkb);
@@ -18,20 +19,23 @@ void dlm_receive_message_saved(struct dlm_ls *ls, const struct dlm_message *ms,
 			       uint32_t saved_seq);
 void dlm_receive_buffer(const union dlm_packet *p, int nodeid);
 int dlm_modes_compat(int mode1, int mode2);
+void free_toss_rsb(struct dlm_rsb *r);
 void dlm_put_rsb(struct dlm_rsb *r);
 void dlm_hold_rsb(struct dlm_rsb *r);
 int dlm_put_lkb(struct dlm_lkb *lkb);
 void dlm_scan_rsbs(struct dlm_ls *ls);
 int dlm_lock_recovery_try(struct dlm_ls *ls);
+void dlm_lock_recovery(struct dlm_ls *ls);
 void dlm_unlock_recovery(struct dlm_ls *ls);
+void dlm_timer_resume(struct dlm_ls *ls);
 
 int dlm_master_lookup(struct dlm_ls *ls, int from_nodeid, const char *name,
 		      int len, unsigned int flags, int *r_nodeid, int *result);
 
-int dlm_search_rsb_tree(struct rb_root *tree, const void *name, int len,
+int dlm_search_rsb_tree(struct rhashtable *rhash, const void *name, int len,
 			struct dlm_rsb **r_ret);
 
-void dlm_recover_purge(struct dlm_ls *ls);
+void dlm_recover_purge(struct dlm_ls *ls, const struct list_head *root_list);
 void dlm_purge_mstcpy_locks(struct dlm_rsb *r);
 void dlm_recover_grant(struct dlm_ls *ls);
 int dlm_recover_waiters_post(struct dlm_ls *ls);
@@ -68,12 +72,12 @@ static inline int is_master(struct dlm_rsb *r)
 
 static inline void lock_rsb(struct dlm_rsb *r)
 {
-	mutex_lock(&r->res_mutex);
+	spin_lock_bh(&r->res_lock);
 }
 
 static inline void unlock_rsb(struct dlm_rsb *r)
 {
-	mutex_unlock(&r->res_mutex);
+	spin_unlock_bh(&r->res_lock);
 }
 
 #endif
diff --git a/fs/dlm/lockspace.c b/fs/dlm/lockspace.c
index 0455dddb07..475ab4370d 100644
--- a/fs/dlm/lockspace.c
+++ b/fs/dlm/lockspace.c
@@ -29,8 +29,6 @@ static int			ls_count;
 static struct mutex		ls_lock;
 static struct list_head		lslist;
 static spinlock_t		lslist_lock;
-static struct task_struct *	scand_task;
-
 
 static ssize_t dlm_control_store(struct dlm_ls *ls, const char *buf, size_t len)
 {
@@ -247,66 +245,11 @@ void dlm_lockspace_exit(void)
 	kset_unregister(dlm_kset);
 }
 
-static struct dlm_ls *find_ls_to_scan(void)
-{
-	struct dlm_ls *ls;
-
-	spin_lock(&lslist_lock);
-	list_for_each_entry(ls, &lslist, ls_list) {
-		if (time_after_eq(jiffies, ls->ls_scan_time +
-					    dlm_config.ci_scan_secs * HZ)) {
-			spin_unlock(&lslist_lock);
-			return ls;
-		}
-	}
-	spin_unlock(&lslist_lock);
-	return NULL;
-}
-
-static int dlm_scand(void *data)
-{
-	struct dlm_ls *ls;
-
-	while (!kthread_should_stop()) {
-		ls = find_ls_to_scan();
-		if (ls) {
-			if (dlm_lock_recovery_try(ls)) {
-				ls->ls_scan_time = jiffies;
-				dlm_scan_rsbs(ls);
-				dlm_unlock_recovery(ls);
-			} else {
-				ls->ls_scan_time += HZ;
-			}
-			continue;
-		}
-		schedule_timeout_interruptible(dlm_config.ci_scan_secs * HZ);
-	}
-	return 0;
-}
-
-static int dlm_scand_start(void)
-{
-	struct task_struct *p;
-	int error = 0;
-
-	p = kthread_run(dlm_scand, NULL, "dlm_scand");
-	if (IS_ERR(p))
-		error = PTR_ERR(p);
-	else
-		scand_task = p;
-	return error;
-}
-
-static void dlm_scand_stop(void)
-{
-	kthread_stop(scand_task);
-}
-
 struct dlm_ls *dlm_find_lockspace_global(uint32_t id)
 {
 	struct dlm_ls *ls;
 
-	spin_lock(&lslist_lock);
+	spin_lock_bh(&lslist_lock);
 
 	list_for_each_entry(ls, &lslist, ls_list) {
 		if (ls->ls_global_id == id) {
@@ -316,7 +259,7 @@ struct dlm_ls *dlm_find_lockspace_global(uint32_t id)
 	}
 	ls = NULL;
  out:
-	spin_unlock(&lslist_lock);
+	spin_unlock_bh(&lslist_lock);
 	return ls;
 }
 
@@ -324,7 +267,7 @@ struct dlm_ls *dlm_find_lockspace_local(dlm_lockspace_t *lockspace)
 {
 	struct dlm_ls *ls;
 
-	spin_lock(&lslist_lock);
+	spin_lock_bh(&lslist_lock);
 	list_for_each_entry(ls, &lslist, ls_list) {
 		if (ls->ls_local_handle == lockspace) {
 			atomic_inc(&ls->ls_count);
@@ -333,7 +276,7 @@ struct dlm_ls *dlm_find_lockspace_local(dlm_lockspace_t *lockspace)
 	}
 	ls = NULL;
  out:
-	spin_unlock(&lslist_lock);
+	spin_unlock_bh(&lslist_lock);
 	return ls;
 }
 
@@ -341,7 +284,7 @@ struct dlm_ls *dlm_find_lockspace_device(int minor)
 {
 	struct dlm_ls *ls;
 
-	spin_lock(&lslist_lock);
+	spin_lock_bh(&lslist_lock);
 	list_for_each_entry(ls, &lslist, ls_list) {
 		if (ls->ls_device.minor == minor) {
 			atomic_inc(&ls->ls_count);
@@ -350,7 +293,7 @@ struct dlm_ls *dlm_find_lockspace_device(int minor)
 	}
 	ls = NULL;
  out:
-	spin_unlock(&lslist_lock);
+	spin_unlock_bh(&lslist_lock);
 	return ls;
 }
 
@@ -365,15 +308,15 @@ static void remove_lockspace(struct dlm_ls *ls)
 retry:
 	wait_event(ls->ls_count_wait, atomic_read(&ls->ls_count) == 0);
 
-	spin_lock(&lslist_lock);
+	spin_lock_bh(&lslist_lock);
 	if (atomic_read(&ls->ls_count) != 0) {
-		spin_unlock(&lslist_lock);
+		spin_unlock_bh(&lslist_lock);
 		goto retry;
 	}
 
 	WARN_ON(ls->ls_create_count != 0);
 	list_del(&ls->ls_list);
-	spin_unlock(&lslist_lock);
+	spin_unlock_bh(&lslist_lock);
 }
 
 static int threads_start(void)
@@ -382,22 +325,9 @@ static int threads_start(void)
 
 	/* Thread for sending/receiving messages for all lockspace's */
 	error = dlm_midcomms_start();
-	if (error) {
+	if (error)
 		log_print("cannot start dlm midcomms %d", error);
-		goto fail;
-	}
-
-	error = dlm_scand_start();
-	if (error) {
-		log_print("cannot start dlm_scand thread %d", error);
-		goto midcomms_fail;
-	}
-
-	return 0;
 
- midcomms_fail:
-	dlm_midcomms_stop();
- fail:
 	return error;
 }
 
@@ -407,9 +337,9 @@ static int new_lockspace(const char *name, const char *cluster,
 			 int *ops_result, dlm_lockspace_t **lockspace)
 {
 	struct dlm_ls *ls;
-	int i, size, error;
 	int do_unreg = 0;
 	int namelen = strlen(name);
+	int error;
 
 	if (namelen > DLM_LOCKSPACE_LEN || namelen == 0)
 		return -EINVAL;
@@ -448,7 +378,7 @@ static int new_lockspace(const char *name, const char *cluster,
 
 	error = 0;
 
-	spin_lock(&lslist_lock);
+	spin_lock_bh(&lslist_lock);
 	list_for_each_entry(ls, &lslist, ls_list) {
 		WARN_ON(ls->ls_create_count <= 0);
 		if (ls->ls_namelen != namelen)
@@ -464,7 +394,7 @@ static int new_lockspace(const char *name, const char *cluster,
 		error = 1;
 		break;
 	}
-	spin_unlock(&lslist_lock);
+	spin_unlock_bh(&lslist_lock);
 
 	if (error)
 		goto out;
@@ -492,32 +422,21 @@ static int new_lockspace(const char *name, const char *cluster,
 	 */
 	ls->ls_exflags = (flags & ~(DLM_LSFL_FS | DLM_LSFL_NEWEXCL));
 
-	size = READ_ONCE(dlm_config.ci_rsbtbl_size);
-	ls->ls_rsbtbl_size = size;
+	INIT_LIST_HEAD(&ls->ls_toss);
+	INIT_LIST_HEAD(&ls->ls_keep);
+	rwlock_init(&ls->ls_rsbtbl_lock);
 
-	ls->ls_rsbtbl = vmalloc(array_size(size, sizeof(struct dlm_rsbtable)));
-	if (!ls->ls_rsbtbl)
+	error = rhashtable_init(&ls->ls_rsbtbl, &dlm_rhash_rsb_params);
+	if (error)
 		goto out_lsfree;
-	for (i = 0; i < size; i++) {
-		ls->ls_rsbtbl[i].keep.rb_node = NULL;
-		ls->ls_rsbtbl[i].toss.rb_node = NULL;
-		spin_lock_init(&ls->ls_rsbtbl[i].lock);
-	}
-
-	for (i = 0; i < DLM_REMOVE_NAMES_MAX; i++) {
-		ls->ls_remove_names[i] = kzalloc(DLM_RESNAME_MAXLEN+1,
-						 GFP_KERNEL);
-		if (!ls->ls_remove_names[i])
-			goto out_rsbtbl;
-	}
 
 	idr_init(&ls->ls_lkbidr);
-	spin_lock_init(&ls->ls_lkbidr_spin);
+	rwlock_init(&ls->ls_lkbidr_lock);
 
 	INIT_LIST_HEAD(&ls->ls_waiters);
-	mutex_init(&ls->ls_waiters_mutex);
+	spin_lock_init(&ls->ls_waiters_lock);
 	INIT_LIST_HEAD(&ls->ls_orphans);
-	mutex_init(&ls->ls_orphans_mutex);
+	spin_lock_init(&ls->ls_orphans_lock);
 
 	INIT_LIST_HEAD(&ls->ls_new_rsb);
 	spin_lock_init(&ls->ls_new_rsb_spin);
@@ -552,11 +471,9 @@ static int new_lockspace(const char *name, const char *cluster,
 	ls->ls_recover_seq = get_random_u64();
 	ls->ls_recover_args = NULL;
 	init_rwsem(&ls->ls_in_recovery);
-	init_rwsem(&ls->ls_recv_active);
+	rwlock_init(&ls->ls_recv_active);
 	INIT_LIST_HEAD(&ls->ls_requestqueue);
-	atomic_set(&ls->ls_requestqueue_cnt, 0);
-	init_waitqueue_head(&ls->ls_requestqueue_wait);
-	mutex_init(&ls->ls_requestqueue_mutex);
+	rwlock_init(&ls->ls_requestqueue_lock);
 	spin_lock_init(&ls->ls_clear_proc_locks);
 
 	/* Due backwards compatibility with 3.1 we need to use maximum
@@ -565,8 +482,10 @@ static int new_lockspace(const char *name, const char *cluster,
 	 * might send less.
 	 */
 	ls->ls_recover_buf = kmalloc(DLM_MAX_SOCKET_BUFSIZE, GFP_NOFS);
-	if (!ls->ls_recover_buf)
+	if (!ls->ls_recover_buf) {
+		error = -ENOMEM;
 		goto out_lkbidr;
+	}
 
 	ls->ls_slot = 0;
 	ls->ls_num_slots = 0;
@@ -580,13 +499,20 @@ static int new_lockspace(const char *name, const char *cluster,
 	ls->ls_recover_list_count = 0;
 	ls->ls_local_handle = ls;
 	init_waitqueue_head(&ls->ls_wait_general);
-	INIT_LIST_HEAD(&ls->ls_root_list);
-	init_rwsem(&ls->ls_root_sem);
+	INIT_LIST_HEAD(&ls->ls_masters_list);
+	rwlock_init(&ls->ls_masters_lock);
+	INIT_LIST_HEAD(&ls->ls_dir_dump_list);
+	rwlock_init(&ls->ls_dir_dump_lock);
+
+	INIT_LIST_HEAD(&ls->ls_toss_q);
+	spin_lock_init(&ls->ls_toss_q_lock);
+	timer_setup(&ls->ls_timer, dlm_rsb_toss_timer,
+		    TIMER_DEFERRABLE);
 
-	spin_lock(&lslist_lock);
+	spin_lock_bh(&lslist_lock);
 	ls->ls_create_count = 1;
 	list_add(&ls->ls_list, &lslist);
-	spin_unlock(&lslist_lock);
+	spin_unlock_bh(&lslist_lock);
 
 	if (flags & DLM_LSFL_FS) {
 		error = dlm_callback_start(ls);
@@ -655,17 +581,14 @@ static int new_lockspace(const char *name, const char *cluster,
  out_callback:
 	dlm_callback_stop(ls);
  out_delist:
-	spin_lock(&lslist_lock);
+	spin_lock_bh(&lslist_lock);
 	list_del(&ls->ls_list);
-	spin_unlock(&lslist_lock);
+	spin_unlock_bh(&lslist_lock);
 	idr_destroy(&ls->ls_recover_idr);
 	kfree(ls->ls_recover_buf);
  out_lkbidr:
 	idr_destroy(&ls->ls_lkbidr);
- out_rsbtbl:
-	for (i = 0; i < DLM_REMOVE_NAMES_MAX; i++)
-		kfree(ls->ls_remove_names[i]);
-	vfree(ls->ls_rsbtbl);
+	rhashtable_destroy(&ls->ls_rsbtbl);
  out_lsfree:
 	if (do_unreg)
 		kobject_put(&ls->ls_kobj);
@@ -697,7 +620,6 @@ static int __dlm_new_lockspace(const char *name, const char *cluster,
 	if (error > 0)
 		error = 0;
 	if (!ls_count) {
-		dlm_scand_stop();
 		dlm_midcomms_shutdown();
 		dlm_midcomms_stop();
 	}
@@ -756,7 +678,7 @@ static int lockspace_busy(struct dlm_ls *ls, int force)
 {
 	int rv;
 
-	spin_lock(&ls->ls_lkbidr_spin);
+	read_lock_bh(&ls->ls_lkbidr_lock);
 	if (force == 0) {
 		rv = idr_for_each(&ls->ls_lkbidr, lkb_idr_is_any, ls);
 	} else if (force == 1) {
@@ -764,19 +686,25 @@ static int lockspace_busy(struct dlm_ls *ls, int force)
 	} else {
 		rv = 0;
 	}
-	spin_unlock(&ls->ls_lkbidr_spin);
+	read_unlock_bh(&ls->ls_lkbidr_lock);
 	return rv;
 }
 
+static void rhash_free_rsb(void *ptr, void *arg)
+{
+	struct dlm_rsb *rsb = ptr;
+
+	dlm_free_rsb(rsb);
+}
+
 static int release_lockspace(struct dlm_ls *ls, int force)
 {
 	struct dlm_rsb *rsb;
-	struct rb_node *n;
-	int i, busy, rv;
+	int busy, rv;
 
 	busy = lockspace_busy(ls, force);
 
-	spin_lock(&lslist_lock);
+	spin_lock_bh(&lslist_lock);
 	if (ls->ls_create_count == 1) {
 		if (busy) {
 			rv = -EBUSY;
@@ -790,7 +718,7 @@ static int release_lockspace(struct dlm_ls *ls, int force)
 	} else {
 		rv = -EINVAL;
 	}
-	spin_unlock(&lslist_lock);
+	spin_unlock_bh(&lslist_lock);
 
 	if (rv) {
 		log_debug(ls, "release_lockspace no remove %d", rv);
@@ -807,8 +735,13 @@ static int release_lockspace(struct dlm_ls *ls, int force)
 
 	dlm_recoverd_stop(ls);
 
+	/* clear the LSFL_RUNNING flag to fast up
+	 * time_shutdown_sync(), we don't care anymore
+	 */
+	clear_bit(LSFL_RUNNING, &ls->ls_flags);
+	timer_shutdown_sync(&ls->ls_timer);
+
 	if (ls_count == 1) {
-		dlm_scand_stop();
 		dlm_clear_members(ls);
 		dlm_midcomms_shutdown();
 	}
@@ -830,27 +763,9 @@ static int release_lockspace(struct dlm_ls *ls, int force)
 	idr_destroy(&ls->ls_lkbidr);
 
 	/*
-	 * Free all rsb's on rsbtbl[] lists
+	 * Free all rsb's on rsbtbl
 	 */
-
-	for (i = 0; i < ls->ls_rsbtbl_size; i++) {
-		while ((n = rb_first(&ls->ls_rsbtbl[i].keep))) {
-			rsb = rb_entry(n, struct dlm_rsb, res_hashnode);
-			rb_erase(n, &ls->ls_rsbtbl[i].keep);
-			dlm_free_rsb(rsb);
-		}
-
-		while ((n = rb_first(&ls->ls_rsbtbl[i].toss))) {
-			rsb = rb_entry(n, struct dlm_rsb, res_hashnode);
-			rb_erase(n, &ls->ls_rsbtbl[i].toss);
-			dlm_free_rsb(rsb);
-		}
-	}
-
-	vfree(ls->ls_rsbtbl);
-
-	for (i = 0; i < DLM_REMOVE_NAMES_MAX; i++)
-		kfree(ls->ls_remove_names[i]);
+	rhashtable_free_and_destroy(&ls->ls_rsbtbl, rhash_free_rsb, NULL);
 
 	while (!list_empty(&ls->ls_new_rsb)) {
 		rsb = list_first_entry(&ls->ls_new_rsb, struct dlm_rsb,
@@ -918,20 +833,19 @@ void dlm_stop_lockspaces(void)
 
  restart:
 	count = 0;
-	spin_lock(&lslist_lock);
+	spin_lock_bh(&lslist_lock);
 	list_for_each_entry(ls, &lslist, ls_list) {
 		if (!test_bit(LSFL_RUNNING, &ls->ls_flags)) {
 			count++;
 			continue;
 		}
-		spin_unlock(&lslist_lock);
+		spin_unlock_bh(&lslist_lock);
 		log_error(ls, "no userland control daemon, stopping lockspace");
 		dlm_ls_stop(ls);
 		goto restart;
 	}
-	spin_unlock(&lslist_lock);
+	spin_unlock_bh(&lslist_lock);
 
 	if (count)
 		log_print("dlm user daemon left %d lockspaces", count);
 }
-
diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcomms.c
index 6296c62c10..6b8078085e 100644
--- a/fs/dlm/lowcomms.c
+++ b/fs/dlm/lowcomms.c
@@ -204,6 +204,7 @@ static void process_dlm_messages(struct work_struct *work);
 static DECLARE_WORK(process_work, process_dlm_messages);
 static DEFINE_SPINLOCK(processqueue_lock);
 static bool process_dlm_messages_pending;
+static DECLARE_WAIT_QUEUE_HEAD(processqueue_wq);
 static atomic_t processqueue_count;
 static LIST_HEAD(processqueue);
 
@@ -248,7 +249,7 @@ struct kmem_cache *dlm_lowcomms_writequeue_cache_create(void)
 
 struct kmem_cache *dlm_lowcomms_msg_cache_create(void)
 {
-	return kmem_cache_create("dlm_msg", sizeof(struct dlm_msg), 0, 0, NULL);
+	return KMEM_CACHE(dlm_msg, 0);
 }
 
 /* need to held writequeue_lock */
@@ -867,36 +868,38 @@ static void process_dlm_messages(struct work_struct *work)
 {
 	struct processqueue_entry *pentry;
 
-	spin_lock(&processqueue_lock);
+	spin_lock_bh(&processqueue_lock);
 	pentry = list_first_entry_or_null(&processqueue,
 					  struct processqueue_entry, list);
 	if (WARN_ON_ONCE(!pentry)) {
 		process_dlm_messages_pending = false;
-		spin_unlock(&processqueue_lock);
+		spin_unlock_bh(&processqueue_lock);
 		return;
 	}
 
 	list_del(&pentry->list);
-	atomic_dec(&processqueue_count);
-	spin_unlock(&processqueue_lock);
+	if (atomic_dec_and_test(&processqueue_count))
+		wake_up(&processqueue_wq);
+	spin_unlock_bh(&processqueue_lock);
 
 	for (;;) {
 		dlm_process_incoming_buffer(pentry->nodeid, pentry->buf,
 					    pentry->buflen);
 		free_processqueue_entry(pentry);
 
-		spin_lock(&processqueue_lock);
+		spin_lock_bh(&processqueue_lock);
 		pentry = list_first_entry_or_null(&processqueue,
 						  struct processqueue_entry, list);
 		if (!pentry) {
 			process_dlm_messages_pending = false;
-			spin_unlock(&processqueue_lock);
+			spin_unlock_bh(&processqueue_lock);
 			break;
 		}
 
 		list_del(&pentry->list);
-		atomic_dec(&processqueue_count);
-		spin_unlock(&processqueue_lock);
+		if (atomic_dec_and_test(&processqueue_count))
+			wake_up(&processqueue_wq);
+		spin_unlock_bh(&processqueue_lock);
 	}
 }
 
@@ -966,14 +969,14 @@ again:
 	memmove(con->rx_leftover_buf, pentry->buf + ret,
 		con->rx_leftover);
 
-	spin_lock(&processqueue_lock);
+	spin_lock_bh(&processqueue_lock);
 	ret = atomic_inc_return(&processqueue_count);
 	list_add_tail(&pentry->list, &processqueue);
 	if (!process_dlm_messages_pending) {
 		process_dlm_messages_pending = true;
 		queue_work(process_workqueue, &process_work);
 	}
-	spin_unlock(&processqueue_lock);
+	spin_unlock_bh(&processqueue_lock);
 
 	if (ret > DLM_MAX_PROCESS_BUFFERS)
 		return DLM_IO_FLUSH;
@@ -1229,14 +1232,13 @@ out:
 };
 
 static struct dlm_msg *dlm_lowcomms_new_msg_con(struct connection *con, int len,
-						gfp_t allocation, char **ppc,
-						void (*cb)(void *data),
+						char **ppc, void (*cb)(void *data),
 						void *data)
 {
 	struct writequeue_entry *e;
 	struct dlm_msg *msg;
 
-	msg = dlm_allocate_msg(allocation);
+	msg = dlm_allocate_msg();
 	if (!msg)
 		return NULL;
 
@@ -1261,9 +1263,8 @@ static struct dlm_msg *dlm_lowcomms_new_msg_con(struct connection *con, int len,
  * dlm_lowcomms_commit_msg which is a must call if success
  */
 #ifndef __CHECKER__
-struct dlm_msg *dlm_lowcomms_new_msg(int nodeid, int len, gfp_t allocation,
-				     char **ppc, void (*cb)(void *data),
-				     void *data)
+struct dlm_msg *dlm_lowcomms_new_msg(int nodeid, int len, char **ppc,
+				     void (*cb)(void *data), void *data)
 {
 	struct connection *con;
 	struct dlm_msg *msg;
@@ -1284,7 +1285,7 @@ struct dlm_msg *dlm_lowcomms_new_msg(int nodeid, int len, gfp_t allocation,
 		return NULL;
 	}
 
-	msg = dlm_lowcomms_new_msg_con(con, len, allocation, ppc, cb, data);
+	msg = dlm_lowcomms_new_msg_con(con, len, ppc, cb, data);
 	if (!msg) {
 		srcu_read_unlock(&connections_srcu, idx);
 		return NULL;
@@ -1348,8 +1349,8 @@ int dlm_lowcomms_resend_msg(struct dlm_msg *msg)
 	if (msg->retransmit)
 		return 1;
 
-	msg_resend = dlm_lowcomms_new_msg_con(msg->entry->con, msg->len,
-					      GFP_ATOMIC, &ppc, NULL, NULL);
+	msg_resend = dlm_lowcomms_new_msg_con(msg->entry->con, msg->len, &ppc,
+					      NULL, NULL);
 	if (!msg_resend)
 		return -ENOMEM;
 
@@ -1513,7 +1514,20 @@ static void process_recv_sockets(struct work_struct *work)
 		/* CF_RECV_PENDING cleared */
 		break;
 	case DLM_IO_FLUSH:
-		flush_workqueue(process_workqueue);
+		/* we can't flush the process_workqueue here because a
+		 * WQ_MEM_RECLAIM workequeue can occurr a deadlock for a non
+		 * WQ_MEM_RECLAIM workqueue such as process_workqueue. Instead
+		 * we have a waitqueue to wait until all messages are
+		 * processed.
+		 *
+		 * This handling is only necessary to backoff the sender and
+		 * not queue all messages from the socket layer into DLM
+		 * processqueue. When DLM is capable to parse multiple messages
+		 * on an e.g. per socket basis this handling can might be
+		 * removed. Especially in a message burst we are too slow to
+		 * process messages and the queue will fill up memory.
+		 */
+		wait_event(processqueue_wq, !atomic_read(&processqueue_count));
 		fallthrough;
 	case DLM_IO_RESCHED:
 		cond_resched();
@@ -1703,11 +1717,7 @@ static int work_start(void)
 		return -ENOMEM;
 	}
 
-	/* ordered dlm message process queue,
-	 * should be converted to a tasklet
-	 */
-	process_workqueue = alloc_ordered_workqueue("dlm_process",
-						    WQ_HIGHPRI | WQ_MEM_RECLAIM);
+	process_workqueue = alloc_workqueue("dlm_process", WQ_HIGHPRI | WQ_BH, 0);
 	if (!process_workqueue) {
 		log_print("can't start dlm_process");
 		destroy_workqueue(io_workqueue);
diff --git a/fs/dlm/lowcomms.h b/fs/dlm/lowcomms.h
index 3e8dca6618..8deb16f8f6 100644
--- a/fs/dlm/lowcomms.h
+++ b/fs/dlm/lowcomms.h
@@ -39,9 +39,8 @@ void dlm_lowcomms_stop(void);
 void dlm_lowcomms_init(void);
 void dlm_lowcomms_exit(void);
 int dlm_lowcomms_close(int nodeid);
-struct dlm_msg *dlm_lowcomms_new_msg(int nodeid, int len, gfp_t allocation,
-				     char **ppc, void (*cb)(void *data),
-				     void *data);
+struct dlm_msg *dlm_lowcomms_new_msg(int nodeid, int len, char **ppc,
+				     void (*cb)(void *data), void *data);
 void dlm_lowcomms_commit_msg(struct dlm_msg *msg);
 void dlm_lowcomms_put_msg(struct dlm_msg *msg);
 int dlm_lowcomms_resend_msg(struct dlm_msg *msg);
diff --git a/fs/dlm/member.c b/fs/dlm/member.c
index be7909ead7..c46e306f2e 100644
--- a/fs/dlm/member.c
+++ b/fs/dlm/member.c
@@ -630,7 +630,7 @@ int dlm_ls_stop(struct dlm_ls *ls)
 	 * message to the requestqueue without races.
 	 */
 
-	down_write(&ls->ls_recv_active);
+	write_lock_bh(&ls->ls_recv_active);
 
 	/*
 	 * Abort any recovery that's in progress (see RECOVER_STOP,
@@ -638,18 +638,25 @@ int dlm_ls_stop(struct dlm_ls *ls)
 	 * dlm to quit any processing (see RUNNING, dlm_locking_stopped()).
 	 */
 
-	spin_lock(&ls->ls_recover_lock);
+	spin_lock_bh(&ls->ls_recover_lock);
 	set_bit(LSFL_RECOVER_STOP, &ls->ls_flags);
 	new = test_and_clear_bit(LSFL_RUNNING, &ls->ls_flags);
+	if (new)
+		timer_delete_sync(&ls->ls_timer);
 	ls->ls_recover_seq++;
-	spin_unlock(&ls->ls_recover_lock);
+
+	/* activate requestqueue and stop processing */
+	write_lock_bh(&ls->ls_requestqueue_lock);
+	set_bit(LSFL_RECV_MSG_BLOCKED, &ls->ls_flags);
+	write_unlock_bh(&ls->ls_requestqueue_lock);
+	spin_unlock_bh(&ls->ls_recover_lock);
 
 	/*
 	 * Let dlm_recv run again, now any normal messages will be saved on the
 	 * requestqueue for later.
 	 */
 
-	up_write(&ls->ls_recv_active);
+	write_unlock_bh(&ls->ls_recv_active);
 
 	/*
 	 * This in_recovery lock does two things:
@@ -674,13 +681,13 @@ int dlm_ls_stop(struct dlm_ls *ls)
 
 	dlm_recoverd_suspend(ls);
 
-	spin_lock(&ls->ls_recover_lock);
+	spin_lock_bh(&ls->ls_recover_lock);
 	kfree(ls->ls_slots);
 	ls->ls_slots = NULL;
 	ls->ls_num_slots = 0;
 	ls->ls_slots_size = 0;
 	ls->ls_recover_status = 0;
-	spin_unlock(&ls->ls_recover_lock);
+	spin_unlock_bh(&ls->ls_recover_lock);
 
 	dlm_recoverd_resume(ls);
 
@@ -714,12 +721,12 @@ int dlm_ls_start(struct dlm_ls *ls)
 	if (error < 0)
 		goto fail_rv;
 
-	spin_lock(&ls->ls_recover_lock);
+	spin_lock_bh(&ls->ls_recover_lock);
 
 	/* the lockspace needs to be stopped before it can be started */
 
 	if (!dlm_locking_stopped(ls)) {
-		spin_unlock(&ls->ls_recover_lock);
+		spin_unlock_bh(&ls->ls_recover_lock);
 		log_error(ls, "start ignored: lockspace running");
 		error = -EINVAL;
 		goto fail;
@@ -730,7 +737,7 @@ int dlm_ls_start(struct dlm_ls *ls)
 	rv->seq = ++ls->ls_recover_seq;
 	rv_old = ls->ls_recover_args;
 	ls->ls_recover_args = rv;
-	spin_unlock(&ls->ls_recover_lock);
+	spin_unlock_bh(&ls->ls_recover_lock);
 
 	if (rv_old) {
 		log_error(ls, "unused recovery %llx %d",
diff --git a/fs/dlm/memory.c b/fs/dlm/memory.c
index 64f212a066..15a8b1cee4 100644
--- a/fs/dlm/memory.c
+++ b/fs/dlm/memory.c
@@ -84,7 +84,7 @@ char *dlm_allocate_lvb(struct dlm_ls *ls)
 {
 	char *p;
 
-	p = kzalloc(ls->ls_lvblen, GFP_NOFS);
+	p = kzalloc(ls->ls_lvblen, GFP_ATOMIC);
 	return p;
 }
 
@@ -97,7 +97,7 @@ struct dlm_rsb *dlm_allocate_rsb(struct dlm_ls *ls)
 {
 	struct dlm_rsb *r;
 
-	r = kmem_cache_zalloc(rsb_cache, GFP_NOFS);
+	r = kmem_cache_zalloc(rsb_cache, GFP_ATOMIC);
 	return r;
 }
 
@@ -112,7 +112,7 @@ struct dlm_lkb *dlm_allocate_lkb(struct dlm_ls *ls)
 {
 	struct dlm_lkb *lkb;
 
-	lkb = kmem_cache_zalloc(lkb_cache, GFP_NOFS);
+	lkb = kmem_cache_zalloc(lkb_cache, GFP_ATOMIC);
 	return lkb;
 }
 
@@ -127,16 +127,12 @@ void dlm_free_lkb(struct dlm_lkb *lkb)
 		}
 	}
 
-	/* drop references if they are set */
-	dlm_callback_set_last_ptr(&lkb->lkb_last_cast, NULL);
-	dlm_callback_set_last_ptr(&lkb->lkb_last_cb, NULL);
-
 	kmem_cache_free(lkb_cache, lkb);
 }
 
-struct dlm_mhandle *dlm_allocate_mhandle(gfp_t allocation)
+struct dlm_mhandle *dlm_allocate_mhandle(void)
 {
-	return kmem_cache_alloc(mhandle_cache, allocation);
+	return kmem_cache_alloc(mhandle_cache, GFP_ATOMIC);
 }
 
 void dlm_free_mhandle(struct dlm_mhandle *mhandle)
@@ -154,9 +150,9 @@ void dlm_free_writequeue(struct writequeue_entry *writequeue)
 	kmem_cache_free(writequeue_cache, writequeue);
 }
 
-struct dlm_msg *dlm_allocate_msg(gfp_t allocation)
+struct dlm_msg *dlm_allocate_msg(void)
 {
-	return kmem_cache_alloc(msg_cache, allocation);
+	return kmem_cache_alloc(msg_cache, GFP_ATOMIC);
 }
 
 void dlm_free_msg(struct dlm_msg *msg)
diff --git a/fs/dlm/memory.h b/fs/dlm/memory.h
index 6b29563d24..15198d46b4 100644
--- a/fs/dlm/memory.h
+++ b/fs/dlm/memory.h
@@ -20,11 +20,11 @@ struct dlm_lkb *dlm_allocate_lkb(struct dlm_ls *ls);
 void dlm_free_lkb(struct dlm_lkb *l);
 char *dlm_allocate_lvb(struct dlm_ls *ls);
 void dlm_free_lvb(char *l);
-struct dlm_mhandle *dlm_allocate_mhandle(gfp_t allocation);
+struct dlm_mhandle *dlm_allocate_mhandle(void);
 void dlm_free_mhandle(struct dlm_mhandle *mhandle);
 struct writequeue_entry *dlm_allocate_writequeue(void);
 void dlm_free_writequeue(struct writequeue_entry *writequeue);
-struct dlm_msg *dlm_allocate_msg(gfp_t allocation);
+struct dlm_msg *dlm_allocate_msg(void);
 void dlm_free_msg(struct dlm_msg *msg);
 struct dlm_callback *dlm_allocate_cb(void);
 void dlm_free_cb(struct dlm_callback *cb);
diff --git a/fs/dlm/midcomms.c b/fs/dlm/midcomms.c
index 2247ebb61b..c34f38e9ee 100644
--- a/fs/dlm/midcomms.c
+++ b/fs/dlm/midcomms.c
@@ -226,8 +226,7 @@ static DEFINE_MUTEX(close_lock);
 
 struct kmem_cache *dlm_midcomms_cache_create(void)
 {
-	return kmem_cache_create("dlm_mhandle", sizeof(struct dlm_mhandle),
-				 0, 0, NULL);
+	return KMEM_CACHE(dlm_mhandle, 0);
 }
 
 static inline const char *dlm_state_str(int state)
@@ -365,9 +364,9 @@ int dlm_midcomms_addr(int nodeid, struct sockaddr_storage *addr, int len)
 	node->users = 0;
 	midcomms_node_reset(node);
 
-	spin_lock(&nodes_lock);
+	spin_lock_bh(&nodes_lock);
 	hlist_add_head_rcu(&node->hlist, &node_hash[r]);
-	spin_unlock(&nodes_lock);
+	spin_unlock_bh(&nodes_lock);
 
 	node->debugfs = dlm_create_debug_comms_file(nodeid, node);
 	return 0;
@@ -380,8 +379,7 @@ static int dlm_send_ack(int nodeid, uint32_t seq)
 	struct dlm_msg *msg;
 	char *ppc;
 
-	msg = dlm_lowcomms_new_msg(nodeid, mb_len, GFP_ATOMIC, &ppc,
-				   NULL, NULL);
+	msg = dlm_lowcomms_new_msg(nodeid, mb_len, &ppc, NULL, NULL);
 	if (!msg)
 		return -ENOMEM;
 
@@ -429,7 +427,7 @@ static int dlm_send_fin(struct midcomms_node *node,
 	struct dlm_mhandle *mh;
 	char *ppc;
 
-	mh = dlm_midcomms_get_mhandle(node->nodeid, mb_len, GFP_ATOMIC, &ppc);
+	mh = dlm_midcomms_get_mhandle(node->nodeid, mb_len, &ppc);
 	if (!mh)
 		return -ENOMEM;
 
@@ -479,7 +477,7 @@ static void dlm_receive_ack(struct midcomms_node *node, uint32_t seq)
 
 static void dlm_pas_fin_ack_rcv(struct midcomms_node *node)
 {
-	spin_lock(&node->state_lock);
+	spin_lock_bh(&node->state_lock);
 	pr_debug("receive passive fin ack from node %d with state %s\n",
 		 node->nodeid, dlm_state_str(node->state));
 
@@ -493,13 +491,13 @@ static void dlm_pas_fin_ack_rcv(struct midcomms_node *node)
 		wake_up(&node->shutdown_wait);
 		break;
 	default:
-		spin_unlock(&node->state_lock);
+		spin_unlock_bh(&node->state_lock);
 		log_print("%s: unexpected state: %d",
 			  __func__, node->state);
 		WARN_ON_ONCE(1);
 		return;
 	}
-	spin_unlock(&node->state_lock);
+	spin_unlock_bh(&node->state_lock);
 }
 
 static void dlm_receive_buffer_3_2_trace(uint32_t seq,
@@ -536,7 +534,7 @@ static void dlm_midcomms_receive_buffer(const union dlm_packet *p,
 	if (is_expected_seq) {
 		switch (p->header.h_cmd) {
 		case DLM_FIN:
-			spin_lock(&node->state_lock);
+			spin_lock_bh(&node->state_lock);
 			pr_debug("receive fin msg from node %d with state %s\n",
 				 node->nodeid, dlm_state_str(node->state));
 
@@ -577,13 +575,13 @@ static void dlm_midcomms_receive_buffer(const union dlm_packet *p,
 				/* probably remove_member caught it, do nothing */
 				break;
 			default:
-				spin_unlock(&node->state_lock);
+				spin_unlock_bh(&node->state_lock);
 				log_print("%s: unexpected state: %d",
 					  __func__, node->state);
 				WARN_ON_ONCE(1);
 				return;
 			}
-			spin_unlock(&node->state_lock);
+			spin_unlock_bh(&node->state_lock);
 			break;
 		default:
 			WARN_ON_ONCE(test_bit(DLM_NODE_FLAG_STOP_RX, &node->flags));
@@ -977,13 +975,13 @@ static void midcomms_new_msg_cb(void *data)
 }
 
 static struct dlm_msg *dlm_midcomms_get_msg_3_2(struct dlm_mhandle *mh, int nodeid,
-						int len, gfp_t allocation, char **ppc)
+						int len, char **ppc)
 {
 	struct dlm_opts *opts;
 	struct dlm_msg *msg;
 
 	msg = dlm_lowcomms_new_msg(nodeid, len + DLM_MIDCOMMS_OPT_LEN,
-				   allocation, ppc, midcomms_new_msg_cb, mh);
+				   ppc, midcomms_new_msg_cb, mh);
 	if (!msg)
 		return NULL;
 
@@ -1002,8 +1000,7 @@ static struct dlm_msg *dlm_midcomms_get_msg_3_2(struct dlm_mhandle *mh, int node
  * dlm_midcomms_commit_mhandle which is a must call if success
  */
 #ifndef __CHECKER__
-struct dlm_mhandle *dlm_midcomms_get_mhandle(int nodeid, int len,
-					     gfp_t allocation, char **ppc)
+struct dlm_mhandle *dlm_midcomms_get_mhandle(int nodeid, int len, char **ppc)
 {
 	struct midcomms_node *node;
 	struct dlm_mhandle *mh;
@@ -1018,7 +1015,7 @@ struct dlm_mhandle *dlm_midcomms_get_mhandle(int nodeid, int len,
 	/* this is a bug, however we going on and hope it will be resolved */
 	WARN_ON_ONCE(test_bit(DLM_NODE_FLAG_STOP_TX, &node->flags));
 
-	mh = dlm_allocate_mhandle(allocation);
+	mh = dlm_allocate_mhandle();
 	if (!mh)
 		goto err;
 
@@ -1029,8 +1026,7 @@ struct dlm_mhandle *dlm_midcomms_get_mhandle(int nodeid, int len,
 
 	switch (node->version) {
 	case DLM_VERSION_3_1:
-		msg = dlm_lowcomms_new_msg(nodeid, len, allocation, ppc,
-					   NULL, NULL);
+		msg = dlm_lowcomms_new_msg(nodeid, len, ppc, NULL, NULL);
 		if (!msg) {
 			dlm_free_mhandle(mh);
 			goto err;
@@ -1041,8 +1037,7 @@ struct dlm_mhandle *dlm_midcomms_get_mhandle(int nodeid, int len,
 		/* send ack back if necessary */
 		dlm_send_ack_threshold(node, DLM_SEND_ACK_BACK_MSG_THRESHOLD);
 
-		msg = dlm_midcomms_get_msg_3_2(mh, nodeid, len, allocation,
-					       ppc);
+		msg = dlm_midcomms_get_msg_3_2(mh, nodeid, len, ppc);
 		if (!msg) {
 			dlm_free_mhandle(mh);
 			goto err;
@@ -1187,7 +1182,7 @@ void dlm_midcomms_exit(void)
 
 static void dlm_act_fin_ack_rcv(struct midcomms_node *node)
 {
-	spin_lock(&node->state_lock);
+	spin_lock_bh(&node->state_lock);
 	pr_debug("receive active fin ack from node %d with state %s\n",
 		 node->nodeid, dlm_state_str(node->state));
 
@@ -1207,13 +1202,13 @@ static void dlm_act_fin_ack_rcv(struct midcomms_node *node)
 		wake_up(&node->shutdown_wait);
 		break;
 	default:
-		spin_unlock(&node->state_lock);
+		spin_unlock_bh(&node->state_lock);
 		log_print("%s: unexpected state: %d",
 			  __func__, node->state);
 		WARN_ON_ONCE(1);
 		return;
 	}
-	spin_unlock(&node->state_lock);
+	spin_unlock_bh(&node->state_lock);
 }
 
 void dlm_midcomms_add_member(int nodeid)
@@ -1228,7 +1223,7 @@ void dlm_midcomms_add_member(int nodeid)
 		return;
 	}
 
-	spin_lock(&node->state_lock);
+	spin_lock_bh(&node->state_lock);
 	if (!node->users) {
 		pr_debug("receive add member from node %d with state %s\n",
 			 node->nodeid, dlm_state_str(node->state));
@@ -1256,7 +1251,7 @@ void dlm_midcomms_add_member(int nodeid)
 
 	node->users++;
 	pr_debug("node %d users inc count %d\n", nodeid, node->users);
-	spin_unlock(&node->state_lock);
+	spin_unlock_bh(&node->state_lock);
 
 	srcu_read_unlock(&nodes_srcu, idx);
 }
@@ -1274,13 +1269,13 @@ void dlm_midcomms_remove_member(int nodeid)
 		return;
 	}
 
-	spin_lock(&node->state_lock);
+	spin_lock_bh(&node->state_lock);
 	/* case of dlm_midcomms_addr() created node but
 	 * was not added before because dlm_midcomms_close()
 	 * removed the node
 	 */
 	if (!node->users) {
-		spin_unlock(&node->state_lock);
+		spin_unlock_bh(&node->state_lock);
 		srcu_read_unlock(&nodes_srcu, idx);
 		return;
 	}
@@ -1318,7 +1313,7 @@ void dlm_midcomms_remove_member(int nodeid)
 			break;
 		}
 	}
-	spin_unlock(&node->state_lock);
+	spin_unlock_bh(&node->state_lock);
 
 	srcu_read_unlock(&nodes_srcu, idx);
 }
@@ -1356,7 +1351,7 @@ static void midcomms_shutdown(struct midcomms_node *node)
 		return;
 	}
 
-	spin_lock(&node->state_lock);
+	spin_lock_bh(&node->state_lock);
 	pr_debug("receive active shutdown for node %d with state %s\n",
 		 node->nodeid, dlm_state_str(node->state));
 	switch (node->state) {
@@ -1375,7 +1370,7 @@ static void midcomms_shutdown(struct midcomms_node *node)
 		 */
 		break;
 	}
-	spin_unlock(&node->state_lock);
+	spin_unlock_bh(&node->state_lock);
 
 	if (DLM_DEBUG_FENCE_TERMINATION)
 		msleep(5000);
@@ -1446,9 +1441,9 @@ int dlm_midcomms_close(int nodeid)
 	ret = dlm_lowcomms_close(nodeid);
 	dlm_delete_debug_comms_file(node->debugfs);
 
-	spin_lock(&nodes_lock);
+	spin_lock_bh(&nodes_lock);
 	hlist_del_rcu(&node->hlist);
-	spin_unlock(&nodes_lock);
+	spin_unlock_bh(&nodes_lock);
 	srcu_read_unlock(&nodes_srcu, idx);
 
 	/* wait that all readers left until flush send queue */
@@ -1502,8 +1497,8 @@ int dlm_midcomms_rawmsg_send(struct midcomms_node *node, void *buf,
 	rd.node = node;
 	rd.buf = buf;
 
-	msg = dlm_lowcomms_new_msg(node->nodeid, buflen, GFP_NOFS,
-				   &msgbuf, midcomms_new_rawmsg_cb, &rd);
+	msg = dlm_lowcomms_new_msg(node->nodeid, buflen, &msgbuf,
+				   midcomms_new_rawmsg_cb, &rd);
 	if (!msg)
 		return -ENOMEM;
 
diff --git a/fs/dlm/midcomms.h b/fs/dlm/midcomms.h
index e7246fb3ef..278d26fdeb 100644
--- a/fs/dlm/midcomms.h
+++ b/fs/dlm/midcomms.h
@@ -16,8 +16,7 @@ struct midcomms_node;
 
 int dlm_validate_incoming_buffer(int nodeid, unsigned char *buf, int len);
 int dlm_process_incoming_buffer(int nodeid, unsigned char *buf, int buflen);
-struct dlm_mhandle *dlm_midcomms_get_mhandle(int nodeid, int len,
-					     gfp_t allocation, char **ppc);
+struct dlm_mhandle *dlm_midcomms_get_mhandle(int nodeid, int len, char **ppc);
 void dlm_midcomms_commit_mhandle(struct dlm_mhandle *mh, const void *name,
 				 int namelen);
 int dlm_midcomms_addr(int nodeid, struct sockaddr_storage *addr, int len);
diff --git a/fs/dlm/plock.c b/fs/dlm/plock.c
index d814c51213..9ca83ef70e 100644
--- a/fs/dlm/plock.c
+++ b/fs/dlm/plock.c
@@ -138,14 +138,14 @@ int dlm_posix_lock(dlm_lockspace_t *lockspace, u64 number, struct file *file,
 	}
 
 	op->info.optype		= DLM_PLOCK_OP_LOCK;
-	op->info.pid		= fl->fl_pid;
-	op->info.ex		= (fl->fl_type == F_WRLCK);
-	op->info.wait		= !!(fl->fl_flags & FL_SLEEP);
+	op->info.pid		= fl->c.flc_pid;
+	op->info.ex		= lock_is_write(fl);
+	op->info.wait		= !!(fl->c.flc_flags & FL_SLEEP);
 	op->info.fsid		= ls->ls_global_id;
 	op->info.number		= number;
 	op->info.start		= fl->fl_start;
 	op->info.end		= fl->fl_end;
-	op->info.owner = (__u64)(long)fl->fl_owner;
+	op->info.owner = (__u64)(long) fl->c.flc_owner;
 	/* async handling */
 	if (fl->fl_lmops && fl->fl_lmops->lm_grant) {
 		op_data = kzalloc(sizeof(*op_data), GFP_NOFS);
@@ -258,7 +258,7 @@ static int dlm_plock_callback(struct plock_op *op)
 	}
 
 	/* got fs lock; bookkeep locally as well: */
-	flc->fl_flags &= ~FL_SLEEP;
+	flc->c.flc_flags &= ~FL_SLEEP;
 	if (posix_lock_file(file, flc, NULL)) {
 		/*
 		 * This can only happen in the case of kmalloc() failure.
@@ -291,7 +291,7 @@ int dlm_posix_unlock(dlm_lockspace_t *lockspace, u64 number, struct file *file,
 	struct dlm_ls *ls;
 	struct plock_op *op;
 	int rv;
-	unsigned char fl_flags = fl->fl_flags;
+	unsigned char saved_flags = fl->c.flc_flags;
 
 	ls = dlm_find_lockspace_local(lockspace);
 	if (!ls)
@@ -304,7 +304,7 @@ int dlm_posix_unlock(dlm_lockspace_t *lockspace, u64 number, struct file *file,
 	}
 
 	/* cause the vfs unlock to return ENOENT if lock is not found */
-	fl->fl_flags |= FL_EXISTS;
+	fl->c.flc_flags |= FL_EXISTS;
 
 	rv = locks_lock_file_wait(file, fl);
 	if (rv == -ENOENT) {
@@ -317,14 +317,14 @@ int dlm_posix_unlock(dlm_lockspace_t *lockspace, u64 number, struct file *file,
 	}
 
 	op->info.optype		= DLM_PLOCK_OP_UNLOCK;
-	op->info.pid		= fl->fl_pid;
+	op->info.pid		= fl->c.flc_pid;
 	op->info.fsid		= ls->ls_global_id;
 	op->info.number		= number;
 	op->info.start		= fl->fl_start;
 	op->info.end		= fl->fl_end;
-	op->info.owner = (__u64)(long)fl->fl_owner;
+	op->info.owner = (__u64)(long) fl->c.flc_owner;
 
-	if (fl->fl_flags & FL_CLOSE) {
+	if (fl->c.flc_flags & FL_CLOSE) {
 		op->info.flags |= DLM_PLOCK_FL_CLOSE;
 		send_op(op);
 		rv = 0;
@@ -345,7 +345,7 @@ out_free:
 	dlm_release_plock_op(op);
 out:
 	dlm_put_lockspace(ls);
-	fl->fl_flags = fl_flags;
+	fl->c.flc_flags = saved_flags;
 	return rv;
 }
 EXPORT_SYMBOL_GPL(dlm_posix_unlock);
@@ -375,14 +375,14 @@ int dlm_posix_cancel(dlm_lockspace_t *lockspace, u64 number, struct file *file,
 		return -EINVAL;
 
 	memset(&info, 0, sizeof(info));
-	info.pid = fl->fl_pid;
-	info.ex = (fl->fl_type == F_WRLCK);
+	info.pid = fl->c.flc_pid;
+	info.ex = lock_is_write(fl);
 	info.fsid = ls->ls_global_id;
 	dlm_put_lockspace(ls);
 	info.number = number;
 	info.start = fl->fl_start;
 	info.end = fl->fl_end;
-	info.owner = (__u64)(long)fl->fl_owner;
+	info.owner = (__u64)(long) fl->c.flc_owner;
 
 	rv = do_lock_cancel(&info);
 	switch (rv) {
@@ -437,13 +437,13 @@ int dlm_posix_get(dlm_lockspace_t *lockspace, u64 number, struct file *file,
 	}
 
 	op->info.optype		= DLM_PLOCK_OP_GET;
-	op->info.pid		= fl->fl_pid;
-	op->info.ex		= (fl->fl_type == F_WRLCK);
+	op->info.pid		= fl->c.flc_pid;
+	op->info.ex		= lock_is_write(fl);
 	op->info.fsid		= ls->ls_global_id;
 	op->info.number		= number;
 	op->info.start		= fl->fl_start;
 	op->info.end		= fl->fl_end;
-	op->info.owner = (__u64)(long)fl->fl_owner;
+	op->info.owner = (__u64)(long) fl->c.flc_owner;
 
 	send_op(op);
 	wait_event(recv_wq, (op->done != 0));
@@ -455,16 +455,16 @@ int dlm_posix_get(dlm_lockspace_t *lockspace, u64 number, struct file *file,
 
 	rv = op->info.rv;
 
-	fl->fl_type = F_UNLCK;
+	fl->c.flc_type = F_UNLCK;
 	if (rv == -ENOENT)
 		rv = 0;
 	else if (rv > 0) {
 		locks_init_lock(fl);
-		fl->fl_type = (op->info.ex) ? F_WRLCK : F_RDLCK;
-		fl->fl_flags = FL_POSIX;
-		fl->fl_pid = op->info.pid;
+		fl->c.flc_type = (op->info.ex) ? F_WRLCK : F_RDLCK;
+		fl->c.flc_flags = FL_POSIX;
+		fl->c.flc_pid = op->info.pid;
 		if (op->info.nodeid != dlm_our_nodeid())
-			fl->fl_pid = -fl->fl_pid;
+			fl->c.flc_pid = -fl->c.flc_pid;
 		fl->fl_start = op->info.start;
 		fl->fl_end = op->info.end;
 		rv = 0;
diff --git a/fs/dlm/rcom.c b/fs/dlm/rcom.c
index 3b734aed26..be1a71a630 100644
--- a/fs/dlm/rcom.c
+++ b/fs/dlm/rcom.c
@@ -55,7 +55,7 @@ static int create_rcom(struct dlm_ls *ls, int to_nodeid, int type, int len,
 	struct dlm_mhandle *mh;
 	char *mb;
 
-	mh = dlm_midcomms_get_mhandle(to_nodeid, mb_len, GFP_NOFS, &mb);
+	mh = dlm_midcomms_get_mhandle(to_nodeid, mb_len, &mb);
 	if (!mh) {
 		log_print("%s to %d type %d len %d ENOBUFS",
 			  __func__, to_nodeid, type, len);
@@ -75,8 +75,7 @@ static int create_rcom_stateless(struct dlm_ls *ls, int to_nodeid, int type,
 	struct dlm_msg *msg;
 	char *mb;
 
-	msg = dlm_lowcomms_new_msg(to_nodeid, mb_len, GFP_NOFS, &mb,
-				   NULL, NULL);
+	msg = dlm_lowcomms_new_msg(to_nodeid, mb_len, &mb, NULL, NULL);
 	if (!msg) {
 		log_print("create_rcom to %d type %d len %d ENOBUFS",
 			  to_nodeid, type, len);
@@ -144,18 +143,18 @@ static int check_rcom_config(struct dlm_ls *ls, struct dlm_rcom *rc, int nodeid)
 
 static void allow_sync_reply(struct dlm_ls *ls, __le64 *new_seq)
 {
-	spin_lock(&ls->ls_rcom_spin);
+	spin_lock_bh(&ls->ls_rcom_spin);
 	*new_seq = cpu_to_le64(++ls->ls_rcom_seq);
 	set_bit(LSFL_RCOM_WAIT, &ls->ls_flags);
-	spin_unlock(&ls->ls_rcom_spin);
+	spin_unlock_bh(&ls->ls_rcom_spin);
 }
 
 static void disallow_sync_reply(struct dlm_ls *ls)
 {
-	spin_lock(&ls->ls_rcom_spin);
+	spin_lock_bh(&ls->ls_rcom_spin);
 	clear_bit(LSFL_RCOM_WAIT, &ls->ls_flags);
 	clear_bit(LSFL_RCOM_READY, &ls->ls_flags);
-	spin_unlock(&ls->ls_rcom_spin);
+	spin_unlock_bh(&ls->ls_rcom_spin);
 }
 
 /*
@@ -246,10 +245,10 @@ static void receive_rcom_status(struct dlm_ls *ls,
 		goto do_create;
 	}
 
-	spin_lock(&ls->ls_recover_lock);
+	spin_lock_bh(&ls->ls_recover_lock);
 	status = ls->ls_recover_status;
 	num_slots = ls->ls_num_slots;
-	spin_unlock(&ls->ls_recover_lock);
+	spin_unlock_bh(&ls->ls_recover_lock);
 	len += num_slots * sizeof(struct rcom_slot);
 
  do_create:
@@ -267,9 +266,9 @@ static void receive_rcom_status(struct dlm_ls *ls,
 	if (!num_slots)
 		goto do_send;
 
-	spin_lock(&ls->ls_recover_lock);
+	spin_lock_bh(&ls->ls_recover_lock);
 	if (ls->ls_num_slots != num_slots) {
-		spin_unlock(&ls->ls_recover_lock);
+		spin_unlock_bh(&ls->ls_recover_lock);
 		log_debug(ls, "receive_rcom_status num_slots %d to %d",
 			  num_slots, ls->ls_num_slots);
 		rc->rc_result = 0;
@@ -278,7 +277,7 @@ static void receive_rcom_status(struct dlm_ls *ls,
 	}
 
 	dlm_slots_copy_out(ls, rc);
-	spin_unlock(&ls->ls_recover_lock);
+	spin_unlock_bh(&ls->ls_recover_lock);
 
  do_send:
 	send_rcom_stateless(msg, rc);
@@ -286,7 +285,7 @@ static void receive_rcom_status(struct dlm_ls *ls,
 
 static void receive_sync_reply(struct dlm_ls *ls, const struct dlm_rcom *rc_in)
 {
-	spin_lock(&ls->ls_rcom_spin);
+	spin_lock_bh(&ls->ls_rcom_spin);
 	if (!test_bit(LSFL_RCOM_WAIT, &ls->ls_flags) ||
 	    le64_to_cpu(rc_in->rc_id) != ls->ls_rcom_seq) {
 		log_debug(ls, "reject reply %d from %d seq %llx expect %llx",
@@ -302,7 +301,7 @@ static void receive_sync_reply(struct dlm_ls *ls, const struct dlm_rcom *rc_in)
 	clear_bit(LSFL_RCOM_WAIT, &ls->ls_flags);
 	wake_up(&ls->ls_wait_general);
  out:
-	spin_unlock(&ls->ls_rcom_spin);
+	spin_unlock_bh(&ls->ls_rcom_spin);
 }
 
 int dlm_rcom_names(struct dlm_ls *ls, int nodeid, char *last_name,
@@ -510,7 +509,7 @@ int dlm_send_ls_not_ready(int nodeid, const struct dlm_rcom *rc_in)
 	char *mb;
 	int mb_len = sizeof(struct dlm_rcom) + sizeof(struct rcom_config);
 
-	mh = dlm_midcomms_get_mhandle(nodeid, mb_len, GFP_NOFS, &mb);
+	mh = dlm_midcomms_get_mhandle(nodeid, mb_len, &mb);
 	if (!mh)
 		return -ENOBUFS;
 
@@ -614,11 +613,11 @@ void dlm_receive_rcom(struct dlm_ls *ls, const struct dlm_rcom *rc, int nodeid)
 		break;
 	}
 
-	spin_lock(&ls->ls_recover_lock);
+	spin_lock_bh(&ls->ls_recover_lock);
 	status = ls->ls_recover_status;
 	stop = dlm_recovery_stopped(ls);
 	seq = ls->ls_recover_seq;
-	spin_unlock(&ls->ls_recover_lock);
+	spin_unlock_bh(&ls->ls_recover_lock);
 
 	if (stop && (rc->rc_type != cpu_to_le32(DLM_RCOM_STATUS)))
 		goto ignore;
diff --git a/fs/dlm/recover.c b/fs/dlm/recover.c
index 53917c0aa3..f493d5f30c 100644
--- a/fs/dlm/recover.c
+++ b/fs/dlm/recover.c
@@ -74,9 +74,9 @@ int dlm_wait_function(struct dlm_ls *ls, int (*testfn) (struct dlm_ls *ls))
 uint32_t dlm_recover_status(struct dlm_ls *ls)
 {
 	uint32_t status;
-	spin_lock(&ls->ls_recover_lock);
+	spin_lock_bh(&ls->ls_recover_lock);
 	status = ls->ls_recover_status;
-	spin_unlock(&ls->ls_recover_lock);
+	spin_unlock_bh(&ls->ls_recover_lock);
 	return status;
 }
 
@@ -87,9 +87,9 @@ static void _set_recover_status(struct dlm_ls *ls, uint32_t status)
 
 void dlm_set_recover_status(struct dlm_ls *ls, uint32_t status)
 {
-	spin_lock(&ls->ls_recover_lock);
+	spin_lock_bh(&ls->ls_recover_lock);
 	_set_recover_status(ls, status);
-	spin_unlock(&ls->ls_recover_lock);
+	spin_unlock_bh(&ls->ls_recover_lock);
 }
 
 static int wait_status_all(struct dlm_ls *ls, uint32_t wait_status,
@@ -188,13 +188,13 @@ int dlm_recover_members_wait(struct dlm_ls *ls, uint64_t seq)
 
 		rv = dlm_slots_assign(ls, &num_slots, &slots_size, &slots, &gen);
 		if (!rv) {
-			spin_lock(&ls->ls_recover_lock);
+			spin_lock_bh(&ls->ls_recover_lock);
 			_set_recover_status(ls, DLM_RS_NODES_ALL);
 			ls->ls_num_slots = num_slots;
 			ls->ls_slots_size = slots_size;
 			ls->ls_slots = slots;
 			ls->ls_generation = gen;
-			spin_unlock(&ls->ls_recover_lock);
+			spin_unlock_bh(&ls->ls_recover_lock);
 		} else {
 			dlm_set_recover_status(ls, DLM_RS_NODES_ALL);
 		}
@@ -241,9 +241,9 @@ static int recover_list_empty(struct dlm_ls *ls)
 {
 	int empty;
 
-	spin_lock(&ls->ls_recover_list_lock);
+	spin_lock_bh(&ls->ls_recover_list_lock);
 	empty = list_empty(&ls->ls_recover_list);
-	spin_unlock(&ls->ls_recover_list_lock);
+	spin_unlock_bh(&ls->ls_recover_list_lock);
 
 	return empty;
 }
@@ -252,23 +252,23 @@ static void recover_list_add(struct dlm_rsb *r)
 {
 	struct dlm_ls *ls = r->res_ls;
 
-	spin_lock(&ls->ls_recover_list_lock);
+	spin_lock_bh(&ls->ls_recover_list_lock);
 	if (list_empty(&r->res_recover_list)) {
 		list_add_tail(&r->res_recover_list, &ls->ls_recover_list);
 		ls->ls_recover_list_count++;
 		dlm_hold_rsb(r);
 	}
-	spin_unlock(&ls->ls_recover_list_lock);
+	spin_unlock_bh(&ls->ls_recover_list_lock);
 }
 
 static void recover_list_del(struct dlm_rsb *r)
 {
 	struct dlm_ls *ls = r->res_ls;
 
-	spin_lock(&ls->ls_recover_list_lock);
+	spin_lock_bh(&ls->ls_recover_list_lock);
 	list_del_init(&r->res_recover_list);
 	ls->ls_recover_list_count--;
-	spin_unlock(&ls->ls_recover_list_lock);
+	spin_unlock_bh(&ls->ls_recover_list_lock);
 
 	dlm_put_rsb(r);
 }
@@ -277,7 +277,7 @@ static void recover_list_clear(struct dlm_ls *ls)
 {
 	struct dlm_rsb *r, *s;
 
-	spin_lock(&ls->ls_recover_list_lock);
+	spin_lock_bh(&ls->ls_recover_list_lock);
 	list_for_each_entry_safe(r, s, &ls->ls_recover_list, res_recover_list) {
 		list_del_init(&r->res_recover_list);
 		r->res_recover_locks_count = 0;
@@ -290,17 +290,17 @@ static void recover_list_clear(struct dlm_ls *ls)
 			  ls->ls_recover_list_count);
 		ls->ls_recover_list_count = 0;
 	}
-	spin_unlock(&ls->ls_recover_list_lock);
+	spin_unlock_bh(&ls->ls_recover_list_lock);
 }
 
 static int recover_idr_empty(struct dlm_ls *ls)
 {
 	int empty = 1;
 
-	spin_lock(&ls->ls_recover_idr_lock);
+	spin_lock_bh(&ls->ls_recover_idr_lock);
 	if (ls->ls_recover_list_count)
 		empty = 0;
-	spin_unlock(&ls->ls_recover_idr_lock);
+	spin_unlock_bh(&ls->ls_recover_idr_lock);
 
 	return empty;
 }
@@ -310,8 +310,7 @@ static int recover_idr_add(struct dlm_rsb *r)
 	struct dlm_ls *ls = r->res_ls;
 	int rv;
 
-	idr_preload(GFP_NOFS);
-	spin_lock(&ls->ls_recover_idr_lock);
+	spin_lock_bh(&ls->ls_recover_idr_lock);
 	if (r->res_id) {
 		rv = -1;
 		goto out_unlock;
@@ -325,8 +324,7 @@ static int recover_idr_add(struct dlm_rsb *r)
 	dlm_hold_rsb(r);
 	rv = 0;
 out_unlock:
-	spin_unlock(&ls->ls_recover_idr_lock);
-	idr_preload_end();
+	spin_unlock_bh(&ls->ls_recover_idr_lock);
 	return rv;
 }
 
@@ -334,11 +332,11 @@ static void recover_idr_del(struct dlm_rsb *r)
 {
 	struct dlm_ls *ls = r->res_ls;
 
-	spin_lock(&ls->ls_recover_idr_lock);
+	spin_lock_bh(&ls->ls_recover_idr_lock);
 	idr_remove(&ls->ls_recover_idr, r->res_id);
 	r->res_id = 0;
 	ls->ls_recover_list_count--;
-	spin_unlock(&ls->ls_recover_idr_lock);
+	spin_unlock_bh(&ls->ls_recover_idr_lock);
 
 	dlm_put_rsb(r);
 }
@@ -347,9 +345,9 @@ static struct dlm_rsb *recover_idr_find(struct dlm_ls *ls, uint64_t id)
 {
 	struct dlm_rsb *r;
 
-	spin_lock(&ls->ls_recover_idr_lock);
+	spin_lock_bh(&ls->ls_recover_idr_lock);
 	r = idr_find(&ls->ls_recover_idr, (int)id);
-	spin_unlock(&ls->ls_recover_idr_lock);
+	spin_unlock_bh(&ls->ls_recover_idr_lock);
 	return r;
 }
 
@@ -358,7 +356,7 @@ static void recover_idr_clear(struct dlm_ls *ls)
 	struct dlm_rsb *r;
 	int id;
 
-	spin_lock(&ls->ls_recover_idr_lock);
+	spin_lock_bh(&ls->ls_recover_idr_lock);
 
 	idr_for_each_entry(&ls->ls_recover_idr, r, id) {
 		idr_remove(&ls->ls_recover_idr, id);
@@ -374,7 +372,7 @@ static void recover_idr_clear(struct dlm_ls *ls)
 			  ls->ls_recover_list_count);
 		ls->ls_recover_list_count = 0;
 	}
-	spin_unlock(&ls->ls_recover_idr_lock);
+	spin_unlock_bh(&ls->ls_recover_idr_lock);
 }
 
 
@@ -521,7 +519,8 @@ static int recover_master_static(struct dlm_rsb *r, unsigned int *count)
  * the correct dir node.
  */
 
-int dlm_recover_masters(struct dlm_ls *ls, uint64_t seq)
+int dlm_recover_masters(struct dlm_ls *ls, uint64_t seq,
+			const struct list_head *root_list)
 {
 	struct dlm_rsb *r;
 	unsigned int total = 0;
@@ -531,10 +530,8 @@ int dlm_recover_masters(struct dlm_ls *ls, uint64_t seq)
 
 	log_rinfo(ls, "dlm_recover_masters");
 
-	down_read(&ls->ls_root_sem);
-	list_for_each_entry(r, &ls->ls_root_list, res_root_list) {
+	list_for_each_entry(r, root_list, res_root_list) {
 		if (dlm_recovery_stopped(ls)) {
-			up_read(&ls->ls_root_sem);
 			error = -EINTR;
 			goto out;
 		}
@@ -548,12 +545,9 @@ int dlm_recover_masters(struct dlm_ls *ls, uint64_t seq)
 		cond_resched();
 		total++;
 
-		if (error) {
-			up_read(&ls->ls_root_sem);
+		if (error)
 			goto out;
-		}
 	}
-	up_read(&ls->ls_root_sem);
 
 	log_rinfo(ls, "dlm_recover_masters %u of %u", count, total);
 
@@ -658,13 +652,13 @@ static int recover_locks(struct dlm_rsb *r, uint64_t seq)
 	return error;
 }
 
-int dlm_recover_locks(struct dlm_ls *ls, uint64_t seq)
+int dlm_recover_locks(struct dlm_ls *ls, uint64_t seq,
+		      const struct list_head *root_list)
 {
 	struct dlm_rsb *r;
 	int error, count = 0;
 
-	down_read(&ls->ls_root_sem);
-	list_for_each_entry(r, &ls->ls_root_list, res_root_list) {
+	list_for_each_entry(r, root_list, res_root_list) {
 		if (is_master(r)) {
 			rsb_clear_flag(r, RSB_NEW_MASTER);
 			continue;
@@ -675,19 +669,15 @@ int dlm_recover_locks(struct dlm_ls *ls, uint64_t seq)
 
 		if (dlm_recovery_stopped(ls)) {
 			error = -EINTR;
-			up_read(&ls->ls_root_sem);
 			goto out;
 		}
 
 		error = recover_locks(r, seq);
-		if (error) {
-			up_read(&ls->ls_root_sem);
+		if (error)
 			goto out;
-		}
 
 		count += r->res_recover_locks_count;
 	}
-	up_read(&ls->ls_root_sem);
 
 	log_rinfo(ls, "dlm_recover_locks %d out", count);
 
@@ -856,13 +846,12 @@ static void recover_grant(struct dlm_rsb *r)
 		rsb_set_flag(r, RSB_RECOVER_GRANT);
 }
 
-void dlm_recover_rsbs(struct dlm_ls *ls)
+void dlm_recover_rsbs(struct dlm_ls *ls, const struct list_head *root_list)
 {
 	struct dlm_rsb *r;
 	unsigned int count = 0;
 
-	down_read(&ls->ls_root_sem);
-	list_for_each_entry(r, &ls->ls_root_list, res_root_list) {
+	list_for_each_entry(r, root_list, res_root_list) {
 		lock_rsb(r);
 		if (is_master(r)) {
 			if (rsb_flag(r, RSB_RECOVER_CONVERT))
@@ -883,7 +872,6 @@ void dlm_recover_rsbs(struct dlm_ls *ls)
 		rsb_clear_flag(r, RSB_NEW_MASTER2);
 		unlock_rsb(r);
 	}
-	up_read(&ls->ls_root_sem);
 
 	if (count)
 		log_rinfo(ls, "dlm_recover_rsbs %d done", count);
@@ -891,66 +879,25 @@ void dlm_recover_rsbs(struct dlm_ls *ls)
 
 /* Create a single list of all root rsb's to be used during recovery */
 
-int dlm_create_root_list(struct dlm_ls *ls)
-{
-	struct rb_node *n;
-	struct dlm_rsb *r;
-	int i, error = 0;
-
-	down_write(&ls->ls_root_sem);
-	if (!list_empty(&ls->ls_root_list)) {
-		log_error(ls, "root list not empty");
-		error = -EINVAL;
-		goto out;
-	}
-
-	for (i = 0; i < ls->ls_rsbtbl_size; i++) {
-		spin_lock(&ls->ls_rsbtbl[i].lock);
-		for (n = rb_first(&ls->ls_rsbtbl[i].keep); n; n = rb_next(n)) {
-			r = rb_entry(n, struct dlm_rsb, res_hashnode);
-			list_add(&r->res_root_list, &ls->ls_root_list);
-			dlm_hold_rsb(r);
-		}
-
-		if (!RB_EMPTY_ROOT(&ls->ls_rsbtbl[i].toss))
-			log_error(ls, "dlm_create_root_list toss not empty");
-		spin_unlock(&ls->ls_rsbtbl[i].lock);
-	}
- out:
-	up_write(&ls->ls_root_sem);
-	return error;
-}
-
-void dlm_release_root_list(struct dlm_ls *ls)
+void dlm_clear_toss(struct dlm_ls *ls)
 {
 	struct dlm_rsb *r, *safe;
+	unsigned int count = 0;
 
-	down_write(&ls->ls_root_sem);
-	list_for_each_entry_safe(r, safe, &ls->ls_root_list, res_root_list) {
-		list_del_init(&r->res_root_list);
-		dlm_put_rsb(r);
-	}
-	up_write(&ls->ls_root_sem);
-}
+	write_lock_bh(&ls->ls_rsbtbl_lock);
+	list_for_each_entry_safe(r, safe, &ls->ls_toss, res_rsbs_list) {
+		list_del(&r->res_rsbs_list);
+		rhashtable_remove_fast(&ls->ls_rsbtbl, &r->res_node,
+				       dlm_rhash_rsb_params);
 
-void dlm_clear_toss(struct dlm_ls *ls)
-{
-	struct rb_node *n, *next;
-	struct dlm_rsb *r;
-	unsigned int count = 0;
-	int i;
-
-	for (i = 0; i < ls->ls_rsbtbl_size; i++) {
-		spin_lock(&ls->ls_rsbtbl[i].lock);
-		for (n = rb_first(&ls->ls_rsbtbl[i].toss); n; n = next) {
-			next = rb_next(n);
-			r = rb_entry(n, struct dlm_rsb, res_hashnode);
-			rb_erase(n, &ls->ls_rsbtbl[i].toss);
-			dlm_free_rsb(r);
-			count++;
-		}
-		spin_unlock(&ls->ls_rsbtbl[i].lock);
+		/* remove it from the toss queue if its part of it */
+		if (!list_empty(&r->res_toss_q_list))
+			list_del_init(&r->res_toss_q_list);
+
+		free_toss_rsb(r);
+		count++;
 	}
+	write_unlock_bh(&ls->ls_rsbtbl_lock);
 
 	if (count)
 		log_rinfo(ls, "dlm_clear_toss %u done", count);
diff --git a/fs/dlm/recover.h b/fs/dlm/recover.h
index dbc51013ec..efc79a6e57 100644
--- a/fs/dlm/recover.h
+++ b/fs/dlm/recover.h
@@ -19,14 +19,14 @@ int dlm_recover_members_wait(struct dlm_ls *ls, uint64_t seq);
 int dlm_recover_directory_wait(struct dlm_ls *ls, uint64_t seq);
 int dlm_recover_locks_wait(struct dlm_ls *ls, uint64_t seq);
 int dlm_recover_done_wait(struct dlm_ls *ls, uint64_t seq);
-int dlm_recover_masters(struct dlm_ls *ls, uint64_t seq);
+int dlm_recover_masters(struct dlm_ls *ls, uint64_t seq,
+			const struct list_head *root_list);
 int dlm_recover_master_reply(struct dlm_ls *ls, const struct dlm_rcom *rc);
-int dlm_recover_locks(struct dlm_ls *ls, uint64_t seq);
+int dlm_recover_locks(struct dlm_ls *ls, uint64_t seq,
+		      const struct list_head *root_list);
 void dlm_recovered_lock(struct dlm_rsb *r);
-int dlm_create_root_list(struct dlm_ls *ls);
-void dlm_release_root_list(struct dlm_ls *ls);
 void dlm_clear_toss(struct dlm_ls *ls);
-void dlm_recover_rsbs(struct dlm_ls *ls);
+void dlm_recover_rsbs(struct dlm_ls *ls, const struct list_head *root_list);
 
 #endif				/* __RECOVER_DOT_H__ */
 
diff --git a/fs/dlm/recoverd.c b/fs/dlm/recoverd.c
index 4d17491dea..17a40d1e60 100644
--- a/fs/dlm/recoverd.c
+++ b/fs/dlm/recoverd.c
@@ -20,6 +20,67 @@
 #include "requestqueue.h"
 #include "recoverd.h"
 
+static int dlm_create_masters_list(struct dlm_ls *ls)
+{
+	struct dlm_rsb *r;
+	int error = 0;
+
+	write_lock_bh(&ls->ls_masters_lock);
+	if (!list_empty(&ls->ls_masters_list)) {
+		log_error(ls, "root list not empty");
+		error = -EINVAL;
+		goto out;
+	}
+
+	read_lock_bh(&ls->ls_rsbtbl_lock);
+	list_for_each_entry(r, &ls->ls_keep, res_rsbs_list) {
+		if (r->res_nodeid)
+			continue;
+
+		list_add(&r->res_masters_list, &ls->ls_masters_list);
+		dlm_hold_rsb(r);
+	}
+	read_unlock_bh(&ls->ls_rsbtbl_lock);
+ out:
+	write_unlock_bh(&ls->ls_masters_lock);
+	return error;
+}
+
+static void dlm_release_masters_list(struct dlm_ls *ls)
+{
+	struct dlm_rsb *r, *safe;
+
+	write_lock_bh(&ls->ls_masters_lock);
+	list_for_each_entry_safe(r, safe, &ls->ls_masters_list, res_masters_list) {
+		list_del_init(&r->res_masters_list);
+		dlm_put_rsb(r);
+	}
+	write_unlock_bh(&ls->ls_masters_lock);
+}
+
+static void dlm_create_root_list(struct dlm_ls *ls, struct list_head *root_list)
+{
+	struct dlm_rsb *r;
+
+	read_lock_bh(&ls->ls_rsbtbl_lock);
+	list_for_each_entry(r, &ls->ls_keep, res_rsbs_list) {
+		list_add(&r->res_root_list, root_list);
+		dlm_hold_rsb(r);
+	}
+
+	WARN_ON_ONCE(!list_empty(&ls->ls_toss));
+	read_unlock_bh(&ls->ls_rsbtbl_lock);
+}
+
+static void dlm_release_root_list(struct list_head *root_list)
+{
+	struct dlm_rsb *r, *safe;
+
+	list_for_each_entry_safe(r, safe, root_list, res_root_list) {
+		list_del_init(&r->res_root_list);
+		dlm_put_rsb(r);
+	}
+}
 
 /* If the start for which we're re-enabling locking (seq) has been superseded
    by a newer stop (ls_recover_seq), we need to leave locking disabled.
@@ -32,24 +93,35 @@ static int enable_locking(struct dlm_ls *ls, uint64_t seq)
 {
 	int error = -EINTR;
 
-	down_write(&ls->ls_recv_active);
+	write_lock_bh(&ls->ls_recv_active);
 
-	spin_lock(&ls->ls_recover_lock);
+	spin_lock_bh(&ls->ls_recover_lock);
 	if (ls->ls_recover_seq == seq) {
 		set_bit(LSFL_RUNNING, &ls->ls_flags);
+		/* Schedule next timer if recovery put something on toss.
+		 *
+		 * The rsbs that was queued while recovery on toss hasn't
+		 * started yet because LSFL_RUNNING was set everything
+		 * else recovery hasn't started as well because ls_in_recovery
+		 * is still hold. So we should not run into the case that
+		 * dlm_timer_resume() queues a timer that can occur in
+		 * a no op.
+		 */
+		dlm_timer_resume(ls);
 		/* unblocks processes waiting to enter the dlm */
 		up_write(&ls->ls_in_recovery);
 		clear_bit(LSFL_RECOVER_LOCK, &ls->ls_flags);
 		error = 0;
 	}
-	spin_unlock(&ls->ls_recover_lock);
+	spin_unlock_bh(&ls->ls_recover_lock);
 
-	up_write(&ls->ls_recv_active);
+	write_unlock_bh(&ls->ls_recv_active);
 	return error;
 }
 
 static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 {
+	LIST_HEAD(root_list);
 	unsigned long start;
 	int error, neg = 0;
 
@@ -66,7 +138,7 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 	 * routines.
 	 */
 
-	dlm_create_root_list(ls);
+	dlm_create_root_list(ls, &root_list);
 
 	/*
 	 * Add or remove nodes from the lockspace's ls_nodes list.
@@ -82,10 +154,25 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 		goto fail;
 	}
 
-	dlm_recover_dir_nodeid(ls);
+	dlm_recover_dir_nodeid(ls, &root_list);
+
+	/* Create a snapshot of all active rsbs were we are the master of.
+	 * During the barrier between dlm_recover_members_wait() and
+	 * dlm_recover_directory() other nodes can dump their necessary
+	 * directory dlm_rsb (r->res_dir_nodeid == nodeid) in rcom
+	 * communication dlm_copy_master_names() handling.
+	 *
+	 * TODO We should create a per lockspace list that contains rsbs
+	 * that we are the master of. Instead of creating this list while
+	 * recovery we keep track of those rsbs while locking handling and
+	 * recovery can use it when necessary.
+	 */
+	error = dlm_create_masters_list(ls);
+	if (error) {
+		log_rinfo(ls, "dlm_create_masters_list error %d", error);
+		goto fail_root_list;
+	}
 
-	ls->ls_recover_dir_sent_res = 0;
-	ls->ls_recover_dir_sent_msg = 0;
 	ls->ls_recover_locks_in = 0;
 
 	dlm_set_recover_status(ls, DLM_RS_NODES);
@@ -93,7 +180,8 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 	error = dlm_recover_members_wait(ls, rv->seq);
 	if (error) {
 		log_rinfo(ls, "dlm_recover_members_wait error %d", error);
-		goto fail;
+		dlm_release_masters_list(ls);
+		goto fail_root_list;
 	}
 
 	start = jiffies;
@@ -106,7 +194,8 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 	error = dlm_recover_directory(ls, rv->seq);
 	if (error) {
 		log_rinfo(ls, "dlm_recover_directory error %d", error);
-		goto fail;
+		dlm_release_masters_list(ls);
+		goto fail_root_list;
 	}
 
 	dlm_set_recover_status(ls, DLM_RS_DIR);
@@ -114,11 +203,11 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 	error = dlm_recover_directory_wait(ls, rv->seq);
 	if (error) {
 		log_rinfo(ls, "dlm_recover_directory_wait error %d", error);
-		goto fail;
+		dlm_release_masters_list(ls);
+		goto fail_root_list;
 	}
 
-	log_rinfo(ls, "dlm_recover_directory %u out %u messages",
-		  ls->ls_recover_dir_sent_res, ls->ls_recover_dir_sent_msg);
+	dlm_release_masters_list(ls);
 
 	/*
 	 * We may have outstanding operations that are waiting for a reply from
@@ -130,7 +219,7 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 
 	if (dlm_recovery_stopped(ls)) {
 		error = -EINTR;
-		goto fail;
+		goto fail_root_list;
 	}
 
 	if (neg || dlm_no_directory(ls)) {
@@ -138,27 +227,27 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 		 * Clear lkb's for departed nodes.
 		 */
 
-		dlm_recover_purge(ls);
+		dlm_recover_purge(ls, &root_list);
 
 		/*
 		 * Get new master nodeid's for rsb's that were mastered on
 		 * departed nodes.
 		 */
 
-		error = dlm_recover_masters(ls, rv->seq);
+		error = dlm_recover_masters(ls, rv->seq, &root_list);
 		if (error) {
 			log_rinfo(ls, "dlm_recover_masters error %d", error);
-			goto fail;
+			goto fail_root_list;
 		}
 
 		/*
 		 * Send our locks on remastered rsb's to the new masters.
 		 */
 
-		error = dlm_recover_locks(ls, rv->seq);
+		error = dlm_recover_locks(ls, rv->seq, &root_list);
 		if (error) {
 			log_rinfo(ls, "dlm_recover_locks error %d", error);
-			goto fail;
+			goto fail_root_list;
 		}
 
 		dlm_set_recover_status(ls, DLM_RS_LOCKS);
@@ -166,7 +255,7 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 		error = dlm_recover_locks_wait(ls, rv->seq);
 		if (error) {
 			log_rinfo(ls, "dlm_recover_locks_wait error %d", error);
-			goto fail;
+			goto fail_root_list;
 		}
 
 		log_rinfo(ls, "dlm_recover_locks %u in",
@@ -178,7 +267,7 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 		 * settings.
 		 */
 
-		dlm_recover_rsbs(ls);
+		dlm_recover_rsbs(ls, &root_list);
 	} else {
 		/*
 		 * Other lockspace members may be going through the "neg" steps
@@ -190,11 +279,11 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 		error = dlm_recover_locks_wait(ls, rv->seq);
 		if (error) {
 			log_rinfo(ls, "dlm_recover_locks_wait error %d", error);
-			goto fail;
+			goto fail_root_list;
 		}
 	}
 
-	dlm_release_root_list(ls);
+	dlm_release_root_list(&root_list);
 
 	/*
 	 * Purge directory-related requests that are saved in requestqueue.
@@ -243,8 +332,9 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 
 	return 0;
 
+ fail_root_list:
+	dlm_release_root_list(&root_list);
  fail:
-	dlm_release_root_list(ls);
 	mutex_unlock(&ls->ls_recoverd_active);
 
 	return error;
@@ -259,12 +349,12 @@ static void do_ls_recovery(struct dlm_ls *ls)
 	struct dlm_recover *rv = NULL;
 	int error;
 
-	spin_lock(&ls->ls_recover_lock);
+	spin_lock_bh(&ls->ls_recover_lock);
 	rv = ls->ls_recover_args;
 	ls->ls_recover_args = NULL;
 	if (rv && ls->ls_recover_seq == rv->seq)
 		clear_bit(LSFL_RECOVER_STOP, &ls->ls_flags);
-	spin_unlock(&ls->ls_recover_lock);
+	spin_unlock_bh(&ls->ls_recover_lock);
 
 	if (rv) {
 		error = ls_recover(ls, rv);
diff --git a/fs/dlm/requestqueue.c b/fs/dlm/requestqueue.c
index 892d6ca21e..719a5243a0 100644
--- a/fs/dlm/requestqueue.c
+++ b/fs/dlm/requestqueue.c
@@ -37,7 +37,7 @@ void dlm_add_requestqueue(struct dlm_ls *ls, int nodeid,
 	int length = le16_to_cpu(ms->m_header.h_length) -
 		sizeof(struct dlm_message);
 
-	e = kmalloc(sizeof(struct rq_entry) + length, GFP_NOFS);
+	e = kmalloc(sizeof(struct rq_entry) + length, GFP_ATOMIC);
 	if (!e) {
 		log_print("dlm_add_requestqueue: out of memory len %d", length);
 		return;
@@ -48,10 +48,7 @@ void dlm_add_requestqueue(struct dlm_ls *ls, int nodeid,
 	memcpy(&e->request, ms, sizeof(*ms));
 	memcpy(&e->request.m_extra, ms->m_extra, length);
 
-	atomic_inc(&ls->ls_requestqueue_cnt);
-	mutex_lock(&ls->ls_requestqueue_mutex);
 	list_add_tail(&e->list, &ls->ls_requestqueue);
-	mutex_unlock(&ls->ls_requestqueue_mutex);
 }
 
 /*
@@ -71,16 +68,14 @@ int dlm_process_requestqueue(struct dlm_ls *ls)
 	struct dlm_message *ms;
 	int error = 0;
 
-	mutex_lock(&ls->ls_requestqueue_mutex);
-
+	write_lock_bh(&ls->ls_requestqueue_lock);
 	for (;;) {
 		if (list_empty(&ls->ls_requestqueue)) {
-			mutex_unlock(&ls->ls_requestqueue_mutex);
+			clear_bit(LSFL_RECV_MSG_BLOCKED, &ls->ls_flags);
 			error = 0;
 			break;
 		}
-		e = list_entry(ls->ls_requestqueue.next, struct rq_entry, list);
-		mutex_unlock(&ls->ls_requestqueue_mutex);
+		e = list_first_entry(&ls->ls_requestqueue, struct rq_entry, list);
 
 		ms = &e->request;
 
@@ -93,41 +88,23 @@ int dlm_process_requestqueue(struct dlm_ls *ls)
 			  e->recover_seq);
 
 		dlm_receive_message_saved(ls, &e->request, e->recover_seq);
-
-		mutex_lock(&ls->ls_requestqueue_mutex);
 		list_del(&e->list);
-		if (atomic_dec_and_test(&ls->ls_requestqueue_cnt))
-			wake_up(&ls->ls_requestqueue_wait);
 		kfree(e);
 
 		if (dlm_locking_stopped(ls)) {
 			log_debug(ls, "process_requestqueue abort running");
-			mutex_unlock(&ls->ls_requestqueue_mutex);
 			error = -EINTR;
 			break;
 		}
+		write_unlock_bh(&ls->ls_requestqueue_lock);
 		schedule();
+		write_lock_bh(&ls->ls_requestqueue_lock);
 	}
+	write_unlock_bh(&ls->ls_requestqueue_lock);
 
 	return error;
 }
 
-/*
- * After recovery is done, locking is resumed and dlm_recoverd takes all the
- * saved requests and processes them as they would have been by dlm_recv.  At
- * the same time, dlm_recv will start receiving new requests from remote nodes.
- * We want to delay dlm_recv processing new requests until dlm_recoverd has
- * finished processing the old saved requests.  We don't check for locking
- * stopped here because dlm_ls_stop won't stop locking until it's suspended us
- * (dlm_recv).
- */
-
-void dlm_wait_requestqueue(struct dlm_ls *ls)
-{
-	wait_event(ls->ls_requestqueue_wait,
-		   atomic_read(&ls->ls_requestqueue_cnt) == 0);
-}
-
 static int purge_request(struct dlm_ls *ls, struct dlm_message *ms, int nodeid)
 {
 	__le32 type = ms->m_type;
@@ -158,17 +135,15 @@ void dlm_purge_requestqueue(struct dlm_ls *ls)
 	struct dlm_message *ms;
 	struct rq_entry *e, *safe;
 
-	mutex_lock(&ls->ls_requestqueue_mutex);
+	write_lock_bh(&ls->ls_requestqueue_lock);
 	list_for_each_entry_safe(e, safe, &ls->ls_requestqueue, list) {
 		ms =  &e->request;
 
 		if (purge_request(ls, ms, e->nodeid)) {
 			list_del(&e->list);
-			if (atomic_dec_and_test(&ls->ls_requestqueue_cnt))
-				wake_up(&ls->ls_requestqueue_wait);
 			kfree(e);
 		}
 	}
-	mutex_unlock(&ls->ls_requestqueue_mutex);
+	write_unlock_bh(&ls->ls_requestqueue_lock);
 }
 
diff --git a/fs/dlm/user.c b/fs/dlm/user.c
index 12a483deee..3173b974e8 100644
--- a/fs/dlm/user.c
+++ b/fs/dlm/user.c
@@ -21,6 +21,7 @@
 #include "dlm_internal.h"
 #include "lockspace.h"
 #include "lock.h"
+#include "lvb_table.h"
 #include "user.h"
 #include "ast.h"
 #include "config.h"
@@ -144,24 +145,6 @@ static void compat_output(struct dlm_lock_result *res,
 }
 #endif
 
-/* should held proc->asts_spin lock */
-void dlm_purge_lkb_callbacks(struct dlm_lkb *lkb)
-{
-	struct dlm_callback *cb, *safe;
-
-	list_for_each_entry_safe(cb, safe, &lkb->lkb_callbacks, list) {
-		list_del(&cb->list);
-		kref_put(&cb->ref, dlm_release_callback);
-	}
-
-	clear_bit(DLM_IFL_CB_PENDING_BIT, &lkb->lkb_iflags);
-
-	/* invalidate */
-	dlm_callback_set_last_ptr(&lkb->lkb_last_cast, NULL);
-	dlm_callback_set_last_ptr(&lkb->lkb_last_cb, NULL);
-	lkb->lkb_last_bast_mode = -1;
-}
-
 /* Figure out if this lock is at the end of its life and no longer
    available for the application to use.  The lkb still exists until
    the final ast is read.  A lock becomes EOL in three situations:
@@ -198,6 +181,7 @@ void dlm_user_add_ast(struct dlm_lkb *lkb, uint32_t flags, int mode,
 	struct dlm_ls *ls;
 	struct dlm_user_args *ua;
 	struct dlm_user_proc *proc;
+	struct dlm_callback *cb;
 	int rv;
 
 	if (test_bit(DLM_DFL_ORPHAN_BIT, &lkb->lkb_dflags) ||
@@ -205,7 +189,7 @@ void dlm_user_add_ast(struct dlm_lkb *lkb, uint32_t flags, int mode,
 		return;
 
 	ls = lkb->lkb_resource->res_ls;
-	spin_lock(&ls->ls_clear_proc_locks);
+	spin_lock_bh(&ls->ls_clear_proc_locks);
 
 	/* If ORPHAN/DEAD flag is set, it means the process is dead so an ast
 	   can't be delivered.  For ORPHAN's, dlm_clear_proc_locks() freed
@@ -227,38 +211,44 @@ void dlm_user_add_ast(struct dlm_lkb *lkb, uint32_t flags, int mode,
 	if ((flags & DLM_CB_CAST) && lkb_is_endoflife(mode, status))
 		set_bit(DLM_IFL_ENDOFLIFE_BIT, &lkb->lkb_iflags);
 
-	spin_lock(&proc->asts_spin);
+	spin_lock_bh(&proc->asts_spin);
 
-	rv = dlm_enqueue_lkb_callback(lkb, flags, mode, status, sbflags);
+	rv = dlm_queue_lkb_callback(lkb, flags, mode, status, sbflags, &cb);
 	switch (rv) {
-	case DLM_ENQUEUE_CALLBACK_FAILURE:
-		spin_unlock(&proc->asts_spin);
-		WARN_ON_ONCE(1);
-		goto out;
 	case DLM_ENQUEUE_CALLBACK_NEED_SCHED:
-		kref_get(&lkb->lkb_ref);
-		list_add_tail(&lkb->lkb_cb_list, &proc->asts);
+		cb->ua = *ua;
+		cb->lkb_lksb = &cb->ua.lksb;
+		if (cb->copy_lvb) {
+			memcpy(cb->lvbptr, ua->lksb.sb_lvbptr,
+			       DLM_USER_LVB_LEN);
+			cb->lkb_lksb->sb_lvbptr = cb->lvbptr;
+		}
+
+		list_add_tail(&cb->list, &proc->asts);
 		wake_up_interruptible(&proc->wait);
 		break;
 	case DLM_ENQUEUE_CALLBACK_SUCCESS:
 		break;
+	case DLM_ENQUEUE_CALLBACK_FAILURE:
+		fallthrough;
 	default:
+		spin_unlock_bh(&proc->asts_spin);
 		WARN_ON_ONCE(1);
-		break;
+		goto out;
 	}
-	spin_unlock(&proc->asts_spin);
+	spin_unlock_bh(&proc->asts_spin);
 
 	if (test_bit(DLM_IFL_ENDOFLIFE_BIT, &lkb->lkb_iflags)) {
 		/* N.B. spin_lock locks_spin, not asts_spin */
-		spin_lock(&proc->locks_spin);
+		spin_lock_bh(&proc->locks_spin);
 		if (!list_empty(&lkb->lkb_ownqueue)) {
 			list_del_init(&lkb->lkb_ownqueue);
 			dlm_put_lkb(lkb);
 		}
-		spin_unlock(&proc->locks_spin);
+		spin_unlock_bh(&proc->locks_spin);
 	}
  out:
-	spin_unlock(&ls->ls_clear_proc_locks);
+	spin_unlock_bh(&ls->ls_clear_proc_locks);
 }
 
 static int device_user_lock(struct dlm_user_proc *proc,
@@ -802,7 +792,6 @@ static ssize_t device_read(struct file *file, char __user *buf, size_t count,
 			   loff_t *ppos)
 {
 	struct dlm_user_proc *proc = file->private_data;
-	struct dlm_lkb *lkb;
 	DECLARE_WAITQUEUE(wait, current);
 	struct dlm_callback *cb;
 	int rv, ret;
@@ -824,16 +813,14 @@ static ssize_t device_read(struct file *file, char __user *buf, size_t count,
 #endif
 		return -EINVAL;
 
- try_another:
-
 	/* do we really need this? can a read happen after a close? */
 	if (test_bit(DLM_PROC_FLAGS_CLOSING, &proc->flags))
 		return -EINVAL;
 
-	spin_lock(&proc->asts_spin);
+	spin_lock_bh(&proc->asts_spin);
 	if (list_empty(&proc->asts)) {
 		if (file->f_flags & O_NONBLOCK) {
-			spin_unlock(&proc->asts_spin);
+			spin_unlock_bh(&proc->asts_spin);
 			return -EAGAIN;
 		}
 
@@ -842,16 +829,16 @@ static ssize_t device_read(struct file *file, char __user *buf, size_t count,
 	repeat:
 		set_current_state(TASK_INTERRUPTIBLE);
 		if (list_empty(&proc->asts) && !signal_pending(current)) {
-			spin_unlock(&proc->asts_spin);
+			spin_unlock_bh(&proc->asts_spin);
 			schedule();
-			spin_lock(&proc->asts_spin);
+			spin_lock_bh(&proc->asts_spin);
 			goto repeat;
 		}
 		set_current_state(TASK_RUNNING);
 		remove_wait_queue(&proc->wait, &wait);
 
 		if (signal_pending(current)) {
-			spin_unlock(&proc->asts_spin);
+			spin_unlock_bh(&proc->asts_spin);
 			return -ERESTARTSYS;
 		}
 	}
@@ -860,51 +847,24 @@ static ssize_t device_read(struct file *file, char __user *buf, size_t count,
 	   without removing lkb_cb_list; so empty lkb_cb_list is always
 	   consistent with empty lkb_callbacks */
 
-	lkb = list_first_entry(&proc->asts, struct dlm_lkb, lkb_cb_list);
-
-	rv = dlm_dequeue_lkb_callback(lkb, &cb);
-	switch (rv) {
-	case DLM_DEQUEUE_CALLBACK_EMPTY:
-		/* this shouldn't happen; lkb should have been removed from
-		 * list when last item was dequeued
-		 */
-		log_print("dlm_rem_lkb_callback empty %x", lkb->lkb_id);
-		list_del_init(&lkb->lkb_cb_list);
-		spin_unlock(&proc->asts_spin);
-		/* removes ref for proc->asts, may cause lkb to be freed */
-		dlm_put_lkb(lkb);
-		WARN_ON_ONCE(1);
-		goto try_another;
-	case DLM_DEQUEUE_CALLBACK_LAST:
-		list_del_init(&lkb->lkb_cb_list);
-		clear_bit(DLM_IFL_CB_PENDING_BIT, &lkb->lkb_iflags);
-		break;
-	case DLM_DEQUEUE_CALLBACK_SUCCESS:
-		break;
-	default:
-		WARN_ON_ONCE(1);
-		break;
-	}
-	spin_unlock(&proc->asts_spin);
+	cb = list_first_entry(&proc->asts, struct dlm_callback, list);
+	list_del(&cb->list);
+	spin_unlock_bh(&proc->asts_spin);
 
 	if (cb->flags & DLM_CB_BAST) {
-		trace_dlm_bast(lkb->lkb_resource->res_ls, lkb, cb->mode);
+		trace_dlm_bast(cb->ls_id, cb->lkb_id, cb->mode, cb->res_name,
+			       cb->res_length);
 	} else if (cb->flags & DLM_CB_CAST) {
-		lkb->lkb_lksb->sb_status = cb->sb_status;
-		lkb->lkb_lksb->sb_flags = cb->sb_flags;
-		trace_dlm_ast(lkb->lkb_resource->res_ls, lkb);
+		cb->lkb_lksb->sb_status = cb->sb_status;
+		cb->lkb_lksb->sb_flags = cb->sb_flags;
+		trace_dlm_ast(cb->ls_id, cb->lkb_id, cb->sb_status,
+			      cb->sb_flags, cb->res_name, cb->res_length);
 	}
 
-	ret = copy_result_to_user(lkb->lkb_ua,
+	ret = copy_result_to_user(&cb->ua,
 				  test_bit(DLM_PROC_FLAGS_COMPAT, &proc->flags),
 				  cb->flags, cb->mode, cb->copy_lvb, buf, count);
-
-	kref_put(&cb->ref, dlm_release_callback);
-
-	/* removes ref for proc->asts, may cause lkb to be freed */
-	if (rv == DLM_DEQUEUE_CALLBACK_LAST)
-		dlm_put_lkb(lkb);
-
+	dlm_free_cb(cb);
 	return ret;
 }
 
@@ -914,12 +874,12 @@ static __poll_t device_poll(struct file *file, poll_table *wait)
 
 	poll_wait(file, &proc->wait, wait);
 
-	spin_lock(&proc->asts_spin);
+	spin_lock_bh(&proc->asts_spin);
 	if (!list_empty(&proc->asts)) {
-		spin_unlock(&proc->asts_spin);
+		spin_unlock_bh(&proc->asts_spin);
 		return EPOLLIN | EPOLLRDNORM;
 	}
-	spin_unlock(&proc->asts_spin);
+	spin_unlock_bh(&proc->asts_spin);
 	return 0;
 }
 
diff --git a/fs/ecryptfs/crypto.c b/fs/ecryptfs/crypto.c
index 03bd55069d..d39a1a69fe 100644
--- a/fs/ecryptfs/crypto.c
+++ b/fs/ecryptfs/crypto.c
@@ -1606,9 +1606,7 @@ ecryptfs_add_new_key_tfm(struct ecryptfs_key_tfm **key_tfm, char *cipher_name,
 		goto out;
 	}
 	mutex_init(&tmp_tfm->key_tfm_mutex);
-	strncpy(tmp_tfm->cipher_name, cipher_name,
-		ECRYPTFS_MAX_CIPHER_NAME_SIZE);
-	tmp_tfm->cipher_name[ECRYPTFS_MAX_CIPHER_NAME_SIZE] = '\0';
+	strscpy(tmp_tfm->cipher_name, cipher_name);
 	tmp_tfm->key_size = key_size;
 	rc = ecryptfs_process_key_cipher(&tmp_tfm->key_tfm,
 					 tmp_tfm->cipher_name,
@@ -1949,16 +1947,6 @@ out:
 	return rc;
 }
 
-static bool is_dot_dotdot(const char *name, size_t name_size)
-{
-	if (name_size == 1 && name[0] == '.')
-		return true;
-	else if (name_size == 2 && name[0] == '.' && name[1] == '.')
-		return true;
-
-	return false;
-}
-
 /**
  * ecryptfs_decode_and_decrypt_filename - converts the encoded cipher text name to decoded plaintext
  * @plaintext_name: The plaintext name
diff --git a/fs/ecryptfs/main.c b/fs/ecryptfs/main.c
index 2dc927ba06..577c563023 100644
--- a/fs/ecryptfs/main.c
+++ b/fs/ecryptfs/main.c
@@ -256,11 +256,8 @@ static int ecryptfs_parse_options(struct ecryptfs_sb_info *sbi, char *options,
 	substring_t args[MAX_OPT_ARGS];
 	int token;
 	char *sig_src;
-	char *cipher_name_dst;
 	char *cipher_name_src;
-	char *fn_cipher_name_dst;
 	char *fn_cipher_name_src;
-	char *fnek_dst;
 	char *fnek_src;
 	char *cipher_key_bytes_src;
 	char *fn_cipher_key_bytes_src;
@@ -293,12 +290,8 @@ static int ecryptfs_parse_options(struct ecryptfs_sb_info *sbi, char *options,
 		case ecryptfs_opt_cipher:
 		case ecryptfs_opt_ecryptfs_cipher:
 			cipher_name_src = args[0].from;
-			cipher_name_dst =
-				mount_crypt_stat->
-				global_default_cipher_name;
-			strncpy(cipher_name_dst, cipher_name_src,
-				ECRYPTFS_MAX_CIPHER_NAME_SIZE);
-			cipher_name_dst[ECRYPTFS_MAX_CIPHER_NAME_SIZE] = '\0';
+			strscpy(mount_crypt_stat->global_default_cipher_name,
+				cipher_name_src);
 			cipher_name_set = 1;
 			break;
 		case ecryptfs_opt_ecryptfs_key_bytes:
@@ -326,11 +319,8 @@ static int ecryptfs_parse_options(struct ecryptfs_sb_info *sbi, char *options,
 			break;
 		case ecryptfs_opt_fnek_sig:
 			fnek_src = args[0].from;
-			fnek_dst =
-				mount_crypt_stat->global_default_fnek_sig;
-			strncpy(fnek_dst, fnek_src, ECRYPTFS_SIG_SIZE_HEX);
-			mount_crypt_stat->global_default_fnek_sig[
-				ECRYPTFS_SIG_SIZE_HEX] = '\0';
+			strscpy(mount_crypt_stat->global_default_fnek_sig,
+				fnek_src);
 			rc = ecryptfs_add_global_auth_tok(
 				mount_crypt_stat,
 				mount_crypt_stat->global_default_fnek_sig,
@@ -348,12 +338,8 @@ static int ecryptfs_parse_options(struct ecryptfs_sb_info *sbi, char *options,
 			break;
 		case ecryptfs_opt_fn_cipher:
 			fn_cipher_name_src = args[0].from;
-			fn_cipher_name_dst =
-				mount_crypt_stat->global_default_fn_cipher_name;
-			strncpy(fn_cipher_name_dst, fn_cipher_name_src,
-				ECRYPTFS_MAX_CIPHER_NAME_SIZE);
-			mount_crypt_stat->global_default_fn_cipher_name[
-				ECRYPTFS_MAX_CIPHER_NAME_SIZE] = '\0';
+			strscpy(mount_crypt_stat->global_default_fn_cipher_name,
+				fn_cipher_name_src);
 			fn_cipher_name_set = 1;
 			break;
 		case ecryptfs_opt_fn_cipher_key_bytes:
diff --git a/fs/efivarfs/internal.h b/fs/efivarfs/internal.h
index f7206158ee..d71d2e0842 100644
--- a/fs/efivarfs/internal.h
+++ b/fs/efivarfs/internal.h
@@ -24,11 +24,8 @@ struct efivarfs_fs_info {
 struct efi_variable {
 	efi_char16_t  VariableName[EFI_VAR_NAME_LEN/sizeof(efi_char16_t)];
 	efi_guid_t    VendorGuid;
-	unsigned long DataSize;
-	__u8          Data[1024];
-	efi_status_t  Status;
 	__u32         Attributes;
-} __attribute__((packed));
+};
 
 struct efivar_entry {
 	struct efi_variable var;
diff --git a/fs/efivarfs/vars.c b/fs/efivarfs/vars.c
index 4d722af101..3cc89bb624 100644
--- a/fs/efivarfs/vars.c
+++ b/fs/efivarfs/vars.c
@@ -295,9 +295,9 @@ static bool variable_is_present(efi_char16_t *variable_name, efi_guid_t *vendor,
 	unsigned long strsize1, strsize2;
 	bool found = false;
 
-	strsize1 = ucs2_strsize(variable_name, 1024);
+	strsize1 = ucs2_strsize(variable_name, EFI_VAR_NAME_LEN);
 	list_for_each_entry_safe(entry, n, head, list) {
-		strsize2 = ucs2_strsize(entry->var.VariableName, 1024);
+		strsize2 = ucs2_strsize(entry->var.VariableName, EFI_VAR_NAME_LEN);
 		if (strsize1 == strsize2 &&
 			!memcmp(variable_name, &(entry->var.VariableName),
 				strsize2) &&
@@ -396,6 +396,7 @@ int efivar_init(int (*func)(efi_char16_t *, efi_guid_t, unsigned long, void *,
 
 	do {
 		variable_name_size = 512;
+		BUILD_BUG_ON(EFI_VAR_NAME_LEN < 512);
 
 		status = efivar_get_next_variable(&variable_name_size,
 						  variable_name,
diff --git a/fs/efs/super.c b/fs/efs/super.c
index f17fdac76b..e4421c10ca 100644
--- a/fs/efs/super.c
+++ b/fs/efs/super.c
@@ -14,19 +14,14 @@
 #include <linux/buffer_head.h>
 #include <linux/vfs.h>
 #include <linux/blkdev.h>
-
+#include <linux/fs_context.h>
+#include <linux/fs_parser.h>
 #include "efs.h"
 #include <linux/efs_vh.h>
 #include <linux/efs_fs_sb.h>
 
 static int efs_statfs(struct dentry *dentry, struct kstatfs *buf);
-static int efs_fill_super(struct super_block *s, void *d, int silent);
-
-static struct dentry *efs_mount(struct file_system_type *fs_type,
-	int flags, const char *dev_name, void *data)
-{
-	return mount_bdev(fs_type, flags, dev_name, data, efs_fill_super);
-}
+static int efs_init_fs_context(struct fs_context *fc);
 
 static void efs_kill_sb(struct super_block *s)
 {
@@ -35,15 +30,6 @@ static void efs_kill_sb(struct super_block *s)
 	kfree(sbi);
 }
 
-static struct file_system_type efs_fs_type = {
-	.owner		= THIS_MODULE,
-	.name		= "efs",
-	.mount		= efs_mount,
-	.kill_sb	= efs_kill_sb,
-	.fs_flags	= FS_REQUIRES_DEV,
-};
-MODULE_ALIAS_FS("efs");
-
 static struct pt_types sgi_pt_types[] = {
 	{0x00,		"SGI vh"},
 	{0x01,		"SGI trkrepl"},
@@ -63,6 +49,27 @@ static struct pt_types sgi_pt_types[] = {
 	{0,		NULL}
 };
 
+enum {
+	Opt_explicit_open,
+};
+
+static const struct fs_parameter_spec efs_param_spec[] = {
+	fsparam_flag    ("explicit-open",       Opt_explicit_open),
+	{}
+};
+
+/*
+ * File system definition and registration.
+ */
+static struct file_system_type efs_fs_type = {
+	.owner			= THIS_MODULE,
+	.name			= "efs",
+	.kill_sb		= efs_kill_sb,
+	.fs_flags		= FS_REQUIRES_DEV,
+	.init_fs_context	= efs_init_fs_context,
+	.parameters		= efs_param_spec,
+};
+MODULE_ALIAS_FS("efs");
 
 static struct kmem_cache * efs_inode_cachep;
 
@@ -91,8 +98,8 @@ static int __init init_inodecache(void)
 {
 	efs_inode_cachep = kmem_cache_create("efs_inode_cache",
 				sizeof(struct efs_inode_info), 0,
-				SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD|
-				SLAB_ACCOUNT, init_once);
+				SLAB_RECLAIM_ACCOUNT|SLAB_ACCOUNT,
+				init_once);
 	if (efs_inode_cachep == NULL)
 		return -ENOMEM;
 	return 0;
@@ -108,18 +115,10 @@ static void destroy_inodecache(void)
 	kmem_cache_destroy(efs_inode_cachep);
 }
 
-static int efs_remount(struct super_block *sb, int *flags, char *data)
-{
-	sync_filesystem(sb);
-	*flags |= SB_RDONLY;
-	return 0;
-}
-
 static const struct super_operations efs_superblock_operations = {
 	.alloc_inode	= efs_alloc_inode,
 	.free_inode	= efs_free_inode,
 	.statfs		= efs_statfs,
-	.remount_fs	= efs_remount,
 };
 
 static const struct export_operations efs_export_ops = {
@@ -249,26 +248,26 @@ static int efs_validate_super(struct efs_sb_info *sb, struct efs_super *super) {
 	return 0;    
 }
 
-static int efs_fill_super(struct super_block *s, void *d, int silent)
+static int efs_fill_super(struct super_block *s, struct fs_context *fc)
 {
 	struct efs_sb_info *sb;
 	struct buffer_head *bh;
 	struct inode *root;
 
- 	sb = kzalloc(sizeof(struct efs_sb_info), GFP_KERNEL);
+	sb = kzalloc(sizeof(struct efs_sb_info), GFP_KERNEL);
 	if (!sb)
 		return -ENOMEM;
 	s->s_fs_info = sb;
 	s->s_time_min = 0;
 	s->s_time_max = U32_MAX;
- 
+
 	s->s_magic		= EFS_SUPER_MAGIC;
 	if (!sb_set_blocksize(s, EFS_BLOCKSIZE)) {
 		pr_err("device does not support %d byte blocks\n",
 			EFS_BLOCKSIZE);
 		return -EINVAL;
 	}
-  
+
 	/* read the vh (volume header) block */
 	bh = sb_bread(s, 0);
 
@@ -294,7 +293,7 @@ static int efs_fill_super(struct super_block *s, void *d, int silent)
 		pr_err("cannot read superblock\n");
 		return -EIO;
 	}
-		
+
 	if (efs_validate_super(sb, (struct efs_super *) bh->b_data)) {
 #ifdef DEBUG
 		pr_warn("invalid superblock at block %u\n",
@@ -328,6 +327,61 @@ static int efs_fill_super(struct super_block *s, void *d, int silent)
 	return 0;
 }
 
+static void efs_free_fc(struct fs_context *fc)
+{
+	kfree(fc->fs_private);
+}
+
+static int efs_get_tree(struct fs_context *fc)
+{
+	return get_tree_bdev(fc, efs_fill_super);
+}
+
+static int efs_parse_param(struct fs_context *fc, struct fs_parameter *param)
+{
+	int token;
+	struct fs_parse_result result;
+
+	token = fs_parse(fc, efs_param_spec, param, &result);
+	if (token < 0)
+		return token;
+	return 0;
+}
+
+static int efs_reconfigure(struct fs_context *fc)
+{
+	sync_filesystem(fc->root->d_sb);
+
+	return 0;
+}
+
+struct efs_context {
+	unsigned long s_mount_opts;
+};
+
+static const struct fs_context_operations efs_context_opts = {
+	.parse_param	= efs_parse_param,
+	.get_tree	= efs_get_tree,
+	.reconfigure	= efs_reconfigure,
+	.free		= efs_free_fc,
+};
+
+/*
+ * Set up the filesystem mount context.
+ */
+static int efs_init_fs_context(struct fs_context *fc)
+{
+	struct efs_context *ctx;
+
+	ctx = kzalloc(sizeof(struct efs_context), GFP_KERNEL);
+	if (!ctx)
+		return -ENOMEM;
+	fc->fs_private = ctx;
+	fc->ops = &efs_context_opts;
+
+	return 0;
+}
+
 static int efs_statfs(struct dentry *dentry, struct kstatfs *buf) {
 	struct super_block *sb = dentry->d_sb;
 	struct efs_sb_info *sbi = SUPER_INFO(sb);
diff --git a/fs/erofs/Kconfig b/fs/erofs/Kconfig
index fffd391934..7dcdce660c 100644
--- a/fs/erofs/Kconfig
+++ b/fs/erofs/Kconfig
@@ -112,6 +112,21 @@ config EROFS_FS_ZIP_DEFLATE
 
 	  If unsure, say N.
 
+config EROFS_FS_ZIP_ZSTD
+	bool "EROFS Zstandard compressed data support"
+	depends on EROFS_FS_ZIP
+	select ZSTD_DECOMPRESS
+	help
+	  Saying Y here includes support for reading EROFS file systems
+	  containing Zstandard compressed data.  It gives better compression
+	  ratios than the default LZ4 format, while it costs more CPU
+	  overhead.
+
+	  Zstandard support is an experimental feature for now and so most
+	  file systems will be readable without selecting this option.
+
+	  If unsure, say N.
+
 config EROFS_FS_ONDEMAND
 	bool "EROFS fscache-based on-demand read support"
 	depends on EROFS_FS
diff --git a/fs/erofs/Makefile b/fs/erofs/Makefile
index 994d0b9ded..097d672e6b 100644
--- a/fs/erofs/Makefile
+++ b/fs/erofs/Makefile
@@ -1,9 +1,10 @@
 # SPDX-License-Identifier: GPL-2.0-only
 
 obj-$(CONFIG_EROFS_FS) += erofs.o
-erofs-objs := super.o inode.o data.o namei.o dir.o utils.o sysfs.o
+erofs-objs := super.o inode.o data.o namei.o dir.o sysfs.o
 erofs-$(CONFIG_EROFS_FS_XATTR) += xattr.o
-erofs-$(CONFIG_EROFS_FS_ZIP) += decompressor.o zmap.o zdata.o pcpubuf.o
+erofs-$(CONFIG_EROFS_FS_ZIP) += decompressor.o zmap.o zdata.o zutil.o
 erofs-$(CONFIG_EROFS_FS_ZIP_LZMA) += decompressor_lzma.o
 erofs-$(CONFIG_EROFS_FS_ZIP_DEFLATE) += decompressor_deflate.o
+erofs-$(CONFIG_EROFS_FS_ZIP_ZSTD) += decompressor_zstd.o
 erofs-$(CONFIG_EROFS_FS_ONDEMAND) += fscache.o
diff --git a/fs/erofs/compress.h b/fs/erofs/compress.h
index 7cc5841577..19d53c30c8 100644
--- a/fs/erofs/compress.h
+++ b/fs/erofs/compress.h
@@ -81,13 +81,6 @@ static inline bool z_erofs_put_shortlivedpage(struct page **pagepool,
 	return true;
 }
 
-#define MNGD_MAPPING(sbi)	((sbi)->managed_cache->i_mapping)
-static inline bool erofs_page_is_managed(const struct erofs_sb_info *sbi,
-					 struct page *page)
-{
-	return page->mapping == MNGD_MAPPING(sbi);
-}
-
 int z_erofs_fixup_insize(struct z_erofs_decompress_req *rq, const char *padbuf,
 			 unsigned int padbufsize);
 extern const struct z_erofs_decompressor erofs_decompressors[];
@@ -97,8 +90,12 @@ int z_erofs_load_lzma_config(struct super_block *sb,
 			struct erofs_super_block *dsb, void *data, int size);
 int z_erofs_load_deflate_config(struct super_block *sb,
 			struct erofs_super_block *dsb, void *data, int size);
+int z_erofs_load_zstd_config(struct super_block *sb,
+			struct erofs_super_block *dsb, void *data, int size);
 int z_erofs_lzma_decompress(struct z_erofs_decompress_req *rq,
 			    struct page **pagepool);
 int z_erofs_deflate_decompress(struct z_erofs_decompress_req *rq,
 			       struct page **pagepool);
+int z_erofs_zstd_decompress(struct z_erofs_decompress_req *rq,
+			    struct page **pgpl);
 #endif
diff --git a/fs/erofs/data.c b/fs/erofs/data.c
index 3d9721b3fa..8be60797ea 100644
--- a/fs/erofs/data.c
+++ b/fs/erofs/data.c
@@ -29,11 +29,9 @@ void erofs_put_metabuf(struct erofs_buf *buf)
  * Derive the block size from inode->i_blkbits to make compatible with
  * anonymous inode in fscache mode.
  */
-void *erofs_bread(struct erofs_buf *buf, erofs_blk_t blkaddr,
+void *erofs_bread(struct erofs_buf *buf, erofs_off_t offset,
 		  enum erofs_kmap_type type)
 {
-	struct inode *inode = buf->inode;
-	erofs_off_t offset = (erofs_off_t)blkaddr << inode->i_blkbits;
 	pgoff_t index = offset >> PAGE_SHIFT;
 	struct page *page = buf->page;
 	struct folio *folio;
@@ -43,7 +41,7 @@ void *erofs_bread(struct erofs_buf *buf, erofs_blk_t blkaddr,
 		erofs_put_metabuf(buf);
 
 		nofs_flag = memalloc_nofs_save();
-		folio = read_cache_folio(inode->i_mapping, index, NULL, NULL);
+		folio = read_cache_folio(buf->mapping, index, NULL, NULL);
 		memalloc_nofs_restore(nofs_flag);
 		if (IS_ERR(folio))
 			return folio;
@@ -68,16 +66,16 @@ void *erofs_bread(struct erofs_buf *buf, erofs_blk_t blkaddr,
 void erofs_init_metabuf(struct erofs_buf *buf, struct super_block *sb)
 {
 	if (erofs_is_fscache_mode(sb))
-		buf->inode = EROFS_SB(sb)->s_fscache->inode;
+		buf->mapping = EROFS_SB(sb)->s_fscache->inode->i_mapping;
 	else
-		buf->inode = sb->s_bdev->bd_inode;
+		buf->mapping = sb->s_bdev->bd_mapping;
 }
 
 void *erofs_read_metabuf(struct erofs_buf *buf, struct super_block *sb,
-			 erofs_blk_t blkaddr, enum erofs_kmap_type type)
+			 erofs_off_t offset, enum erofs_kmap_type type)
 {
 	erofs_init_metabuf(buf, sb);
-	return erofs_bread(buf, blkaddr, type);
+	return erofs_bread(buf, offset, type);
 }
 
 static int erofs_map_blocks_flatmode(struct inode *inode,
@@ -154,7 +152,7 @@ int erofs_map_blocks(struct inode *inode, struct erofs_map_blocks *map)
 	pos = ALIGN(erofs_iloc(inode) + vi->inode_isize +
 		    vi->xattr_isize, unit) + unit * chunknr;
 
-	kaddr = erofs_read_metabuf(&buf, sb, erofs_blknr(sb, pos), EROFS_KMAP);
+	kaddr = erofs_read_metabuf(&buf, sb, pos, EROFS_KMAP);
 	if (IS_ERR(kaddr)) {
 		err = PTR_ERR(kaddr);
 		goto out;
@@ -165,7 +163,7 @@ int erofs_map_blocks(struct inode *inode, struct erofs_map_blocks *map)
 
 	/* handle block map */
 	if (!(vi->chunkformat & EROFS_CHUNK_FORMAT_INDEXES)) {
-		__le32 *blkaddr = kaddr + erofs_blkoff(sb, pos);
+		__le32 *blkaddr = kaddr;
 
 		if (le32_to_cpu(*blkaddr) == EROFS_NULL_ADDR) {
 			map->m_flags = 0;
@@ -176,7 +174,7 @@ int erofs_map_blocks(struct inode *inode, struct erofs_map_blocks *map)
 		goto out_unlock;
 	}
 	/* parse chunk indexes */
-	idx = kaddr + erofs_blkoff(sb, pos);
+	idx = kaddr;
 	switch (le32_to_cpu(idx->blkaddr)) {
 	case EROFS_NULL_ADDR:
 		map->m_flags = 0;
@@ -220,7 +218,7 @@ int erofs_map_dev(struct super_block *sb, struct erofs_map_dev *map)
 			up_read(&devs->rwsem);
 			return 0;
 		}
-		map->m_bdev = dif->bdev_handle ? dif->bdev_handle->bdev : NULL;
+		map->m_bdev = dif->bdev_file ? file_bdev(dif->bdev_file) : NULL;
 		map->m_daxdev = dif->dax_dev;
 		map->m_dax_part_off = dif->dax_part_off;
 		map->m_fscache = dif->fscache;
@@ -238,8 +236,8 @@ int erofs_map_dev(struct super_block *sb, struct erofs_map_dev *map)
 			if (map->m_pa >= startoff &&
 			    map->m_pa < startoff + length) {
 				map->m_pa -= startoff;
-				map->m_bdev = dif->bdev_handle ?
-					      dif->bdev_handle->bdev : NULL;
+				map->m_bdev = dif->bdev_file ?
+					      file_bdev(dif->bdev_file) : NULL;
 				map->m_daxdev = dif->dax_dev;
 				map->m_dax_part_off = dif->dax_part_off;
 				map->m_fscache = dif->fscache;
@@ -296,11 +294,10 @@ static int erofs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
 		struct erofs_buf buf = __EROFS_BUF_INITIALIZER;
 
 		iomap->type = IOMAP_INLINE;
-		ptr = erofs_read_metabuf(&buf, sb,
-				erofs_blknr(sb, mdev.m_pa), EROFS_KMAP);
+		ptr = erofs_read_metabuf(&buf, sb, mdev.m_pa, EROFS_KMAP);
 		if (IS_ERR(ptr))
 			return PTR_ERR(ptr);
-		iomap->inline_data = ptr + erofs_blkoff(sb, mdev.m_pa);
+		iomap->inline_data = ptr;
 		iomap->private = buf.base;
 	} else {
 		iomap->type = IOMAP_MAPPED;
diff --git a/fs/erofs/decompressor.c b/fs/erofs/decompressor.c
index 2ec9b2bb62..9d85b6c11c 100644
--- a/fs/erofs/decompressor.c
+++ b/fs/erofs/decompressor.c
@@ -54,7 +54,7 @@ static int z_erofs_load_lz4_config(struct super_block *sb,
 	sbi->lz4.max_distance_pages = distance ?
 					DIV_ROUND_UP(distance, PAGE_SIZE) + 1 :
 					LZ4_MAX_DISTANCE_PAGES;
-	return erofs_pcpubuf_growsize(sbi->lz4.max_pclusterblks);
+	return z_erofs_gbuf_growsize(sbi->lz4.max_pclusterblks);
 }
 
 /*
@@ -111,7 +111,7 @@ static int z_erofs_lz4_prepare_dstpages(struct z_erofs_lz4_decompress_ctx *ctx,
 			victim = availables[--top];
 			get_page(victim);
 		} else {
-			victim = erofs_allocpage(pagepool, rq->gfp);
+			victim = __erofs_allocpage(pagepool, rq->gfp, true);
 			if (!victim)
 				return -ENOMEM;
 			set_page_private(victim, Z_EROFS_SHORTLIVED_PAGE);
@@ -159,7 +159,7 @@ static void *z_erofs_lz4_handle_overlap(struct z_erofs_lz4_decompress_ctx *ctx,
 docopy:
 	/* Or copy compressed data which can be overlapped to per-CPU buffer */
 	in = rq->in;
-	src = erofs_get_pcpubuf(ctx->inpages);
+	src = z_erofs_get_gbuf(ctx->inpages);
 	if (!src) {
 		DBG_BUGON(1);
 		kunmap_local(inpage);
@@ -260,7 +260,7 @@ static int z_erofs_lz4_decompress_mem(struct z_erofs_lz4_decompress_ctx *ctx,
 	} else if (maptype == 1) {
 		vm_unmap_ram(src, ctx->inpages);
 	} else if (maptype == 2) {
-		erofs_put_pcpubuf(src);
+		z_erofs_put_gbuf(src);
 	} else if (maptype != 3) {
 		DBG_BUGON(1);
 		return -EFAULT;
@@ -399,6 +399,13 @@ const struct z_erofs_decompressor erofs_decompressors[] = {
 		.name = "deflate"
 	},
 #endif
+#ifdef CONFIG_EROFS_FS_ZIP_ZSTD
+	[Z_EROFS_COMPRESSION_ZSTD] = {
+		.config = z_erofs_load_zstd_config,
+		.decompress = z_erofs_zstd_decompress,
+		.name = "zstd"
+	},
+#endif
 };
 
 int z_erofs_parse_cfgs(struct super_block *sb, struct erofs_super_block *dsb)
diff --git a/fs/erofs/decompressor_deflate.c b/fs/erofs/decompressor_deflate.c
index b98872058a..3a3461561a 100644
--- a/fs/erofs/decompressor_deflate.c
+++ b/fs/erofs/decompressor_deflate.c
@@ -46,39 +46,15 @@ int __init z_erofs_deflate_init(void)
 	/* by default, use # of possible CPUs instead */
 	if (!z_erofs_deflate_nstrms)
 		z_erofs_deflate_nstrms = num_possible_cpus();
-
-	for (; z_erofs_deflate_avail_strms < z_erofs_deflate_nstrms;
-	     ++z_erofs_deflate_avail_strms) {
-		struct z_erofs_deflate *strm;
-
-		strm = kzalloc(sizeof(*strm), GFP_KERNEL);
-		if (!strm)
-			goto out_failed;
-
-		/* XXX: in-kernel zlib cannot shrink windowbits currently */
-		strm->z.workspace = vmalloc(zlib_inflate_workspacesize());
-		if (!strm->z.workspace) {
-			kfree(strm);
-			goto out_failed;
-		}
-
-		spin_lock(&z_erofs_deflate_lock);
-		strm->next = z_erofs_deflate_head;
-		z_erofs_deflate_head = strm;
-		spin_unlock(&z_erofs_deflate_lock);
-	}
 	return 0;
-
-out_failed:
-	erofs_err(NULL, "failed to allocate zlib workspace");
-	z_erofs_deflate_exit();
-	return -ENOMEM;
 }
 
 int z_erofs_load_deflate_config(struct super_block *sb,
 			struct erofs_super_block *dsb, void *data, int size)
 {
 	struct z_erofs_deflate_cfgs *dfl = data;
+	static DEFINE_MUTEX(deflate_resize_mutex);
+	static bool inited;
 
 	if (!dfl || size < sizeof(struct z_erofs_deflate_cfgs)) {
 		erofs_err(sb, "invalid deflate cfgs, size=%u", size);
@@ -89,9 +65,36 @@ int z_erofs_load_deflate_config(struct super_block *sb,
 		erofs_err(sb, "unsupported windowbits %u", dfl->windowbits);
 		return -EOPNOTSUPP;
 	}
+	mutex_lock(&deflate_resize_mutex);
+	if (!inited) {
+		for (; z_erofs_deflate_avail_strms < z_erofs_deflate_nstrms;
+		     ++z_erofs_deflate_avail_strms) {
+			struct z_erofs_deflate *strm;
+
+			strm = kzalloc(sizeof(*strm), GFP_KERNEL);
+			if (!strm)
+				goto failed;
+			/* XXX: in-kernel zlib cannot customize windowbits */
+			strm->z.workspace = vmalloc(zlib_inflate_workspacesize());
+			if (!strm->z.workspace) {
+				kfree(strm);
+				goto failed;
+			}
 
+			spin_lock(&z_erofs_deflate_lock);
+			strm->next = z_erofs_deflate_head;
+			z_erofs_deflate_head = strm;
+			spin_unlock(&z_erofs_deflate_lock);
+		}
+		inited = true;
+	}
+	mutex_unlock(&deflate_resize_mutex);
 	erofs_info(sb, "EXPERIMENTAL DEFLATE feature in use. Use at your own risk!");
 	return 0;
+failed:
+	mutex_unlock(&deflate_resize_mutex);
+	z_erofs_deflate_exit();
+	return -ENOMEM;
 }
 
 int z_erofs_deflate_decompress(struct z_erofs_decompress_req *rq,
@@ -212,9 +215,6 @@ again:
 
 			if (rq->out[no] != rq->in[j])
 				continue;
-
-			DBG_BUGON(erofs_page_is_managed(EROFS_SB(sb),
-							rq->in[j]));
 			tmppage = erofs_allocpage(pgpl, rq->gfp);
 			if (!tmppage) {
 				err = -ENOMEM;
diff --git a/fs/erofs/decompressor_lzma.c b/fs/erofs/decompressor_lzma.c
index 6ca357d83c..4b28dc130c 100644
--- a/fs/erofs/decompressor_lzma.c
+++ b/fs/erofs/decompressor_lzma.c
@@ -258,9 +258,6 @@ again:
 
 			if (rq->out[no] != rq->in[j])
 				continue;
-
-			DBG_BUGON(erofs_page_is_managed(EROFS_SB(rq->sb),
-							rq->in[j]));
 			tmppage = erofs_allocpage(pgpl, rq->gfp);
 			if (!tmppage) {
 				err = -ENOMEM;
diff --git a/fs/erofs/decompressor_zstd.c b/fs/erofs/decompressor_zstd.c
new file mode 100644
index 0000000000..63a23cac3a
--- /dev/null
+++ b/fs/erofs/decompressor_zstd.c
@@ -0,0 +1,279 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+#include <linux/zstd.h>
+#include "compress.h"
+
+struct z_erofs_zstd {
+	struct z_erofs_zstd *next;
+	u8 bounce[PAGE_SIZE];
+	void *wksp;
+	unsigned int wkspsz;
+};
+
+static DEFINE_SPINLOCK(z_erofs_zstd_lock);
+static unsigned int z_erofs_zstd_max_dictsize;
+static unsigned int z_erofs_zstd_nstrms, z_erofs_zstd_avail_strms;
+static struct z_erofs_zstd *z_erofs_zstd_head;
+static DECLARE_WAIT_QUEUE_HEAD(z_erofs_zstd_wq);
+
+module_param_named(zstd_streams, z_erofs_zstd_nstrms, uint, 0444);
+
+static struct z_erofs_zstd *z_erofs_isolate_strms(bool all)
+{
+	struct z_erofs_zstd *strm;
+
+again:
+	spin_lock(&z_erofs_zstd_lock);
+	strm = z_erofs_zstd_head;
+	if (!strm) {
+		spin_unlock(&z_erofs_zstd_lock);
+		wait_event(z_erofs_zstd_wq, READ_ONCE(z_erofs_zstd_head));
+		goto again;
+	}
+	z_erofs_zstd_head = all ? NULL : strm->next;
+	spin_unlock(&z_erofs_zstd_lock);
+	return strm;
+}
+
+void z_erofs_zstd_exit(void)
+{
+	while (z_erofs_zstd_avail_strms) {
+		struct z_erofs_zstd *strm, *n;
+
+		for (strm = z_erofs_isolate_strms(true); strm; strm = n) {
+			n = strm->next;
+
+			kvfree(strm->wksp);
+			kfree(strm);
+			--z_erofs_zstd_avail_strms;
+		}
+	}
+}
+
+int __init z_erofs_zstd_init(void)
+{
+	/* by default, use # of possible CPUs instead */
+	if (!z_erofs_zstd_nstrms)
+		z_erofs_zstd_nstrms = num_possible_cpus();
+
+	for (; z_erofs_zstd_avail_strms < z_erofs_zstd_nstrms;
+	     ++z_erofs_zstd_avail_strms) {
+		struct z_erofs_zstd *strm;
+
+		strm = kzalloc(sizeof(*strm), GFP_KERNEL);
+		if (!strm) {
+			z_erofs_zstd_exit();
+			return -ENOMEM;
+		}
+		spin_lock(&z_erofs_zstd_lock);
+		strm->next = z_erofs_zstd_head;
+		z_erofs_zstd_head = strm;
+		spin_unlock(&z_erofs_zstd_lock);
+	}
+	return 0;
+}
+
+int z_erofs_load_zstd_config(struct super_block *sb,
+			struct erofs_super_block *dsb, void *data, int size)
+{
+	static DEFINE_MUTEX(zstd_resize_mutex);
+	struct z_erofs_zstd_cfgs *zstd = data;
+	unsigned int dict_size, wkspsz;
+	struct z_erofs_zstd *strm, *head = NULL;
+	void *wksp;
+
+	if (!zstd || size < sizeof(struct z_erofs_zstd_cfgs) || zstd->format) {
+		erofs_err(sb, "unsupported zstd format, size=%u", size);
+		return -EINVAL;
+	}
+
+	if (zstd->windowlog > ilog2(Z_EROFS_ZSTD_MAX_DICT_SIZE) - 10) {
+		erofs_err(sb, "unsupported zstd window log %u", zstd->windowlog);
+		return -EINVAL;
+	}
+	dict_size = 1U << (zstd->windowlog + 10);
+
+	/* in case 2 z_erofs_load_zstd_config() race to avoid deadlock */
+	mutex_lock(&zstd_resize_mutex);
+	if (z_erofs_zstd_max_dictsize >= dict_size) {
+		mutex_unlock(&zstd_resize_mutex);
+		return 0;
+	}
+
+	/* 1. collect/isolate all streams for the following check */
+	while (z_erofs_zstd_avail_strms) {
+		struct z_erofs_zstd *n;
+
+		for (strm = z_erofs_isolate_strms(true); strm; strm = n) {
+			n = strm->next;
+			strm->next = head;
+			head = strm;
+			--z_erofs_zstd_avail_strms;
+		}
+	}
+
+	/* 2. walk each isolated stream and grow max dict_size if needed */
+	wkspsz = zstd_dstream_workspace_bound(dict_size);
+	for (strm = head; strm; strm = strm->next) {
+		wksp = kvmalloc(wkspsz, GFP_KERNEL);
+		if (!wksp)
+			break;
+		kvfree(strm->wksp);
+		strm->wksp = wksp;
+		strm->wkspsz = wkspsz;
+	}
+
+	/* 3. push back all to the global list and update max dict_size */
+	spin_lock(&z_erofs_zstd_lock);
+	DBG_BUGON(z_erofs_zstd_head);
+	z_erofs_zstd_head = head;
+	spin_unlock(&z_erofs_zstd_lock);
+	z_erofs_zstd_avail_strms = z_erofs_zstd_nstrms;
+	wake_up_all(&z_erofs_zstd_wq);
+	if (!strm)
+		z_erofs_zstd_max_dictsize = dict_size;
+	mutex_unlock(&zstd_resize_mutex);
+	return strm ? -ENOMEM : 0;
+}
+
+int z_erofs_zstd_decompress(struct z_erofs_decompress_req *rq,
+			    struct page **pgpl)
+{
+	const unsigned int nrpages_out =
+		PAGE_ALIGN(rq->pageofs_out + rq->outputsize) >> PAGE_SHIFT;
+	const unsigned int nrpages_in =
+		PAGE_ALIGN(rq->inputsize) >> PAGE_SHIFT;
+	zstd_dstream *stream;
+	struct super_block *sb = rq->sb;
+	unsigned int insz, outsz, pofs;
+	struct z_erofs_zstd *strm;
+	zstd_in_buffer in_buf = { NULL, 0, 0 };
+	zstd_out_buffer out_buf = { NULL, 0, 0 };
+	u8 *kin, *kout = NULL;
+	bool bounced = false;
+	int no = -1, ni = 0, j = 0, zerr, err;
+
+	/* 1. get the exact compressed size */
+	kin = kmap_local_page(*rq->in);
+	err = z_erofs_fixup_insize(rq, kin + rq->pageofs_in,
+			min_t(unsigned int, rq->inputsize,
+			      sb->s_blocksize - rq->pageofs_in));
+	if (err) {
+		kunmap_local(kin);
+		return err;
+	}
+
+	/* 2. get an available ZSTD context */
+	strm = z_erofs_isolate_strms(false);
+
+	/* 3. multi-call decompress */
+	insz = rq->inputsize;
+	outsz = rq->outputsize;
+	stream = zstd_init_dstream(z_erofs_zstd_max_dictsize, strm->wksp, strm->wkspsz);
+	if (!stream) {
+		err = -EIO;
+		goto failed_zinit;
+	}
+
+	pofs = rq->pageofs_out;
+	in_buf.size = min_t(u32, insz, PAGE_SIZE - rq->pageofs_in);
+	insz -= in_buf.size;
+	in_buf.src = kin + rq->pageofs_in;
+	do {
+		if (out_buf.size == out_buf.pos) {
+			if (++no >= nrpages_out || !outsz) {
+				erofs_err(sb, "insufficient space for decompressed data");
+				err = -EFSCORRUPTED;
+				break;
+			}
+
+			if (kout)
+				kunmap_local(kout);
+			out_buf.size = min_t(u32, outsz, PAGE_SIZE - pofs);
+			outsz -= out_buf.size;
+			if (!rq->out[no]) {
+				rq->out[no] = erofs_allocpage(pgpl, rq->gfp);
+				if (!rq->out[no]) {
+					kout = NULL;
+					err = -ENOMEM;
+					break;
+				}
+				set_page_private(rq->out[no],
+						 Z_EROFS_SHORTLIVED_PAGE);
+			}
+			kout = kmap_local_page(rq->out[no]);
+			out_buf.dst = kout + pofs;
+			out_buf.pos = 0;
+			pofs = 0;
+		}
+
+		if (in_buf.size == in_buf.pos && insz) {
+			if (++ni >= nrpages_in) {
+				erofs_err(sb, "invalid compressed data");
+				err = -EFSCORRUPTED;
+				break;
+			}
+
+			if (kout) /* unlike kmap(), take care of the orders */
+				kunmap_local(kout);
+			kunmap_local(kin);
+			in_buf.size = min_t(u32, insz, PAGE_SIZE);
+			insz -= in_buf.size;
+			kin = kmap_local_page(rq->in[ni]);
+			in_buf.src = kin;
+			in_buf.pos = 0;
+			bounced = false;
+			if (kout) {
+				j = (u8 *)out_buf.dst - kout;
+				kout = kmap_local_page(rq->out[no]);
+				out_buf.dst = kout + j;
+			}
+		}
+
+		/*
+		 * Handle overlapping: Use bounced buffer if the compressed
+		 * data is under processing; Or use short-lived pages from the
+		 * on-stack pagepool where pages share among the same request
+		 * and not _all_ inplace I/O pages are needed to be doubled.
+		 */
+		if (!bounced && rq->out[no] == rq->in[ni]) {
+			memcpy(strm->bounce, in_buf.src, in_buf.size);
+			in_buf.src = strm->bounce;
+			bounced = true;
+		}
+
+		for (j = ni + 1; j < nrpages_in; ++j) {
+			struct page *tmppage;
+
+			if (rq->out[no] != rq->in[j])
+				continue;
+			tmppage = erofs_allocpage(pgpl, rq->gfp);
+			if (!tmppage) {
+				err = -ENOMEM;
+				goto failed;
+			}
+			set_page_private(tmppage, Z_EROFS_SHORTLIVED_PAGE);
+			copy_highpage(tmppage, rq->in[j]);
+			rq->in[j] = tmppage;
+		}
+		zerr = zstd_decompress_stream(stream, &out_buf, &in_buf);
+		if (zstd_is_error(zerr) || (!zerr && outsz)) {
+			erofs_err(sb, "failed to decompress in[%u] out[%u]: %s",
+				  rq->inputsize, rq->outputsize,
+				  zerr ? zstd_get_error_name(zerr) : "unexpected end of stream");
+			err = -EFSCORRUPTED;
+			break;
+		}
+	} while (outsz || out_buf.pos < out_buf.size);
+failed:
+	if (kout)
+		kunmap_local(kout);
+failed_zinit:
+	kunmap_local(kin);
+	/* 4. push back ZSTD stream context to the global list */
+	spin_lock(&z_erofs_zstd_lock);
+	strm->next = z_erofs_zstd_head;
+	z_erofs_zstd_head = strm;
+	spin_unlock(&z_erofs_zstd_lock);
+	wake_up(&z_erofs_zstd_wq);
+	return err;
+}
diff --git a/fs/erofs/dir.c b/fs/erofs/dir.c
index b80abec053..2193a6710c 100644
--- a/fs/erofs/dir.c
+++ b/fs/erofs/dir.c
@@ -58,12 +58,12 @@ static int erofs_readdir(struct file *f, struct dir_context *ctx)
 	int err = 0;
 	bool initial = true;
 
-	buf.inode = dir;
+	buf.mapping = dir->i_mapping;
 	while (ctx->pos < dirsize) {
 		struct erofs_dirent *de;
 		unsigned int nameoff, maxsize;
 
-		de = erofs_bread(&buf, i, EROFS_KMAP);
+		de = erofs_bread(&buf, erofs_pos(sb, i), EROFS_KMAP);
 		if (IS_ERR(de)) {
 			erofs_err(sb, "fail to readdir of logical block %u of nid %llu",
 				  i, EROFS_I(dir)->nid);
diff --git a/fs/erofs/erofs_fs.h b/fs/erofs/erofs_fs.h
index a03ec70ba6..6c0c270c42 100644
--- a/fs/erofs/erofs_fs.h
+++ b/fs/erofs/erofs_fs.h
@@ -296,6 +296,7 @@ enum {
 	Z_EROFS_COMPRESSION_LZ4		= 0,
 	Z_EROFS_COMPRESSION_LZMA	= 1,
 	Z_EROFS_COMPRESSION_DEFLATE	= 2,
+	Z_EROFS_COMPRESSION_ZSTD	= 3,
 	Z_EROFS_COMPRESSION_MAX
 };
 #define Z_EROFS_ALL_COMPR_ALGS		((1 << Z_EROFS_COMPRESSION_MAX) - 1)
@@ -322,6 +323,15 @@ struct z_erofs_deflate_cfgs {
 	u8 reserved[5];
 } __packed;
 
+/* 6 bytes (+ length field = 8 bytes) */
+struct z_erofs_zstd_cfgs {
+	u8 format;
+	u8 windowlog;           /* windowLog - ZSTD_WINDOWLOG_ABSOLUTEMIN(10) */
+	u8 reserved[4];
+} __packed;
+
+#define Z_EROFS_ZSTD_MAX_DICT_SIZE      Z_EROFS_PCLUSTER_MAX_SIZE
+
 /*
  * bit 0 : COMPACTED_2B indexes (0 - off; 1 - on)
  *  e.g. for 4k logical cluster size,      4B        if compacted 2B is off;
@@ -396,8 +406,7 @@ enum {
 	Z_EROFS_LCLUSTER_TYPE_MAX
 };
 
-#define Z_EROFS_LI_LCLUSTER_TYPE_BITS        2
-#define Z_EROFS_LI_LCLUSTER_TYPE_BIT         0
+#define Z_EROFS_LI_LCLUSTER_TYPE_MASK	(Z_EROFS_LCLUSTER_TYPE_MAX - 1)
 
 /* (noncompact only, HEAD) This pcluster refers to partial decompressed data */
 #define Z_EROFS_LI_PARTIAL_REF		(1 << 15)
@@ -451,8 +460,6 @@ static inline void erofs_check_ondisk_layout_definitions(void)
 		     sizeof(struct z_erofs_lcluster_index));
 	BUILD_BUG_ON(sizeof(struct erofs_deviceslot) != 128);
 
-	BUILD_BUG_ON(BIT(Z_EROFS_LI_LCLUSTER_TYPE_BITS) <
-		     Z_EROFS_LCLUSTER_TYPE_MAX - 1);
 	/* exclude old compiler versions like gcc 7.5.0 */
 	BUILD_BUG_ON(__builtin_constant_p(fmh) ?
 		     fmh != cpu_to_le64(1ULL << 63) : 0);
diff --git a/fs/erofs/fscache.c b/fs/erofs/fscache.c
index 122a4753ec..fda16eedaf 100644
--- a/fs/erofs/fscache.c
+++ b/fs/erofs/fscache.c
@@ -25,9 +25,15 @@ static struct file_system_type erofs_anon_fs_type = {
 	.kill_sb        = kill_anon_super,
 };
 
-struct erofs_fscache_request {
-	struct erofs_fscache_request *primary;
-	struct netfs_cache_resources cache_resources;
+struct erofs_fscache_io {
+	struct netfs_cache_resources cres;
+	struct iov_iter		iter;
+	netfs_io_terminated_t	end_io;
+	void			*private;
+	refcount_t		ref;
+};
+
+struct erofs_fscache_rq {
 	struct address_space	*mapping;	/* The mapping being accessed */
 	loff_t			start;		/* Start position */
 	size_t			len;		/* Length of the request */
@@ -36,44 +42,17 @@ struct erofs_fscache_request {
 	refcount_t		ref;
 };
 
-static struct erofs_fscache_request *erofs_fscache_req_alloc(struct address_space *mapping,
-					     loff_t start, size_t len)
+static bool erofs_fscache_io_put(struct erofs_fscache_io *io)
 {
-	struct erofs_fscache_request *req;
-
-	req = kzalloc(sizeof(struct erofs_fscache_request), GFP_KERNEL);
-	if (!req)
-		return ERR_PTR(-ENOMEM);
-
-	req->mapping = mapping;
-	req->start   = start;
-	req->len     = len;
-	refcount_set(&req->ref, 1);
-
-	return req;
-}
-
-static struct erofs_fscache_request *erofs_fscache_req_chain(struct erofs_fscache_request *primary,
-					     size_t len)
-{
-	struct erofs_fscache_request *req;
-
-	/* use primary request for the first submission */
-	if (!primary->submitted) {
-		refcount_inc(&primary->ref);
-		return primary;
-	}
-
-	req = erofs_fscache_req_alloc(primary->mapping,
-			primary->start + primary->submitted, len);
-	if (!IS_ERR(req)) {
-		req->primary = primary;
-		refcount_inc(&primary->ref);
-	}
-	return req;
+	if (!refcount_dec_and_test(&io->ref))
+		return false;
+	if (io->cres.ops)
+		io->cres.ops->end_operation(&io->cres);
+	kfree(io);
+	return true;
 }
 
-static void erofs_fscache_req_complete(struct erofs_fscache_request *req)
+static void erofs_fscache_req_complete(struct erofs_fscache_rq *req)
 {
 	struct folio *folio;
 	bool failed = req->error;
@@ -93,120 +72,196 @@ static void erofs_fscache_req_complete(struct erofs_fscache_request *req)
 	rcu_read_unlock();
 }
 
-static void erofs_fscache_req_put(struct erofs_fscache_request *req)
+static void erofs_fscache_req_put(struct erofs_fscache_rq *req)
 {
-	if (refcount_dec_and_test(&req->ref)) {
-		if (req->cache_resources.ops)
-			req->cache_resources.ops->end_operation(&req->cache_resources);
-		if (!req->primary)
-			erofs_fscache_req_complete(req);
-		else
-			erofs_fscache_req_put(req->primary);
-		kfree(req);
-	}
+	if (!refcount_dec_and_test(&req->ref))
+		return;
+	erofs_fscache_req_complete(req);
+	kfree(req);
+}
+
+static struct erofs_fscache_rq *erofs_fscache_req_alloc(struct address_space *mapping,
+						loff_t start, size_t len)
+{
+	struct erofs_fscache_rq *req = kzalloc(sizeof(*req), GFP_KERNEL);
+
+	if (!req)
+		return NULL;
+	req->mapping = mapping;
+	req->start = start;
+	req->len = len;
+	refcount_set(&req->ref, 1);
+	return req;
+}
+
+static void erofs_fscache_req_io_put(struct erofs_fscache_io *io)
+{
+	struct erofs_fscache_rq *req = io->private;
+
+	if (erofs_fscache_io_put(io))
+		erofs_fscache_req_put(req);
 }
 
-static void erofs_fscache_subreq_complete(void *priv,
+static void erofs_fscache_req_end_io(void *priv,
 		ssize_t transferred_or_error, bool was_async)
 {
-	struct erofs_fscache_request *req = priv;
+	struct erofs_fscache_io *io = priv;
+	struct erofs_fscache_rq *req = io->private;
 
-	if (IS_ERR_VALUE(transferred_or_error)) {
-		if (req->primary)
-			req->primary->error = transferred_or_error;
-		else
-			req->error = transferred_or_error;
-	}
-	erofs_fscache_req_put(req);
+	if (IS_ERR_VALUE(transferred_or_error))
+		req->error = transferred_or_error;
+	erofs_fscache_req_io_put(io);
+}
+
+static struct erofs_fscache_io *erofs_fscache_req_io_alloc(struct erofs_fscache_rq *req)
+{
+	struct erofs_fscache_io *io = kzalloc(sizeof(*io), GFP_KERNEL);
+
+	if (!io)
+		return NULL;
+	io->end_io = erofs_fscache_req_end_io;
+	io->private = req;
+	refcount_inc(&req->ref);
+	refcount_set(&io->ref, 1);
+	return io;
 }
 
 /*
- * Read data from fscache (cookie, pstart, len), and fill the read data into
- * page cache described by (req->mapping, lstart, len). @pstart describeis the
- * start physical address in the cache file.
+ * Read data from fscache described by cookie at pstart physical address
+ * offset, and fill the read data into buffer described by io->iter.
  */
-static int erofs_fscache_read_folios_async(struct fscache_cookie *cookie,
-		struct erofs_fscache_request *req, loff_t pstart, size_t len)
+static int erofs_fscache_read_io_async(struct fscache_cookie *cookie,
+		loff_t pstart, struct erofs_fscache_io *io)
 {
 	enum netfs_io_source source;
-	struct super_block *sb = req->mapping->host->i_sb;
-	struct netfs_cache_resources *cres = &req->cache_resources;
-	struct iov_iter iter;
-	loff_t lstart = req->start + req->submitted;
-	size_t done = 0;
+	struct netfs_cache_resources *cres = &io->cres;
+	struct iov_iter *iter = &io->iter;
 	int ret;
 
-	DBG_BUGON(len > req->len - req->submitted);
-
 	ret = fscache_begin_read_operation(cres, cookie);
 	if (ret)
 		return ret;
 
-	while (done < len) {
-		loff_t sstart = pstart + done;
-		size_t slen = len - done;
+	while (iov_iter_count(iter)) {
+		size_t orig_count = iov_iter_count(iter), len = orig_count;
 		unsigned long flags = 1 << NETFS_SREQ_ONDEMAND;
 
 		source = cres->ops->prepare_ondemand_read(cres,
-				sstart, &slen, LLONG_MAX, &flags, 0);
-		if (WARN_ON(slen == 0))
+				pstart, &len, LLONG_MAX, &flags, 0);
+		if (WARN_ON(len == 0))
 			source = NETFS_INVALID_READ;
 		if (source != NETFS_READ_FROM_CACHE) {
-			erofs_err(sb, "failed to fscache prepare_read (source %d)", source);
+			erofs_err(NULL, "prepare_ondemand_read failed (source %d)", source);
 			return -EIO;
 		}
 
-		refcount_inc(&req->ref);
-		iov_iter_xarray(&iter, ITER_DEST, &req->mapping->i_pages,
-				lstart + done, slen);
-
-		ret = fscache_read(cres, sstart, &iter, NETFS_READ_HOLE_FAIL,
-				   erofs_fscache_subreq_complete, req);
+		iov_iter_truncate(iter, len);
+		refcount_inc(&io->ref);
+		ret = fscache_read(cres, pstart, iter, NETFS_READ_HOLE_FAIL,
+				   io->end_io, io);
 		if (ret == -EIOCBQUEUED)
 			ret = 0;
 		if (ret) {
-			erofs_err(sb, "failed to fscache_read (ret %d)", ret);
+			erofs_err(NULL, "fscache_read failed (ret %d)", ret);
 			return ret;
 		}
+		if (WARN_ON(iov_iter_count(iter)))
+			return -EIO;
 
-		done += slen;
+		iov_iter_reexpand(iter, orig_count - len);
+		pstart += len;
 	}
-	DBG_BUGON(done != len);
 	return 0;
 }
 
-static int erofs_fscache_meta_read_folio(struct file *data, struct folio *folio)
+struct erofs_fscache_bio {
+	struct erofs_fscache_io io;
+	struct bio bio;		/* w/o bdev to share bio_add_page/endio() */
+	struct bio_vec bvecs[BIO_MAX_VECS];
+};
+
+static void erofs_fscache_bio_endio(void *priv,
+		ssize_t transferred_or_error, bool was_async)
+{
+	struct erofs_fscache_bio *io = priv;
+
+	if (IS_ERR_VALUE(transferred_or_error))
+		io->bio.bi_status = errno_to_blk_status(transferred_or_error);
+	io->bio.bi_end_io(&io->bio);
+	BUILD_BUG_ON(offsetof(struct erofs_fscache_bio, io) != 0);
+	erofs_fscache_io_put(&io->io);
+}
+
+struct bio *erofs_fscache_bio_alloc(struct erofs_map_dev *mdev)
 {
+	struct erofs_fscache_bio *io;
+
+	io = kmalloc(sizeof(*io), GFP_KERNEL | __GFP_NOFAIL);
+	bio_init(&io->bio, NULL, io->bvecs, BIO_MAX_VECS, REQ_OP_READ);
+	io->io.private = mdev->m_fscache->cookie;
+	io->io.end_io = erofs_fscache_bio_endio;
+	refcount_set(&io->io.ref, 1);
+	return &io->bio;
+}
+
+void erofs_fscache_submit_bio(struct bio *bio)
+{
+	struct erofs_fscache_bio *io = container_of(bio,
+			struct erofs_fscache_bio, bio);
 	int ret;
+
+	iov_iter_bvec(&io->io.iter, ITER_DEST, io->bvecs, bio->bi_vcnt,
+		      bio->bi_iter.bi_size);
+	ret = erofs_fscache_read_io_async(io->io.private,
+				bio->bi_iter.bi_sector << 9, &io->io);
+	erofs_fscache_io_put(&io->io);
+	if (!ret)
+		return;
+	bio->bi_status = errno_to_blk_status(ret);
+	bio->bi_end_io(bio);
+}
+
+static int erofs_fscache_meta_read_folio(struct file *data, struct folio *folio)
+{
 	struct erofs_fscache *ctx = folio->mapping->host->i_private;
-	struct erofs_fscache_request *req;
+	int ret = -ENOMEM;
+	struct erofs_fscache_rq *req;
+	struct erofs_fscache_io *io;
 
 	req = erofs_fscache_req_alloc(folio->mapping,
 				folio_pos(folio), folio_size(folio));
-	if (IS_ERR(req)) {
+	if (!req) {
 		folio_unlock(folio);
-		return PTR_ERR(req);
+		return ret;
 	}
 
-	ret = erofs_fscache_read_folios_async(ctx->cookie, req,
-				folio_pos(folio), folio_size(folio));
+	io = erofs_fscache_req_io_alloc(req);
+	if (!io) {
+		req->error = ret;
+		goto out;
+	}
+	iov_iter_xarray(&io->iter, ITER_DEST, &folio->mapping->i_pages,
+			folio_pos(folio), folio_size(folio));
+
+	ret = erofs_fscache_read_io_async(ctx->cookie, folio_pos(folio), io);
 	if (ret)
 		req->error = ret;
 
+	erofs_fscache_req_io_put(io);
+out:
 	erofs_fscache_req_put(req);
 	return ret;
 }
 
-static int erofs_fscache_data_read_slice(struct erofs_fscache_request *primary)
+static int erofs_fscache_data_read_slice(struct erofs_fscache_rq *req)
 {
-	struct address_space *mapping = primary->mapping;
+	struct address_space *mapping = req->mapping;
 	struct inode *inode = mapping->host;
 	struct super_block *sb = inode->i_sb;
-	struct erofs_fscache_request *req;
+	struct erofs_fscache_io *io;
 	struct erofs_map_blocks map;
 	struct erofs_map_dev mdev;
-	struct iov_iter iter;
-	loff_t pos = primary->start + primary->submitted;
+	loff_t pos = req->start + req->submitted;
 	size_t count;
 	int ret;
 
@@ -217,35 +272,32 @@ static int erofs_fscache_data_read_slice(struct erofs_fscache_request *primary)
 
 	if (map.m_flags & EROFS_MAP_META) {
 		struct erofs_buf buf = __EROFS_BUF_INITIALIZER;
-		erofs_blk_t blknr;
-		size_t offset, size;
+		struct iov_iter iter;
+		size_t size = map.m_llen;
 		void *src;
 
-		/* For tail packing layout, the offset may be non-zero. */
-		offset = erofs_blkoff(sb, map.m_pa);
-		blknr = erofs_blknr(sb, map.m_pa);
-		size = map.m_llen;
-
-		src = erofs_read_metabuf(&buf, sb, blknr, EROFS_KMAP);
+		src = erofs_read_metabuf(&buf, sb, map.m_pa, EROFS_KMAP);
 		if (IS_ERR(src))
 			return PTR_ERR(src);
 
 		iov_iter_xarray(&iter, ITER_DEST, &mapping->i_pages, pos, PAGE_SIZE);
-		if (copy_to_iter(src + offset, size, &iter) != size) {
+		if (copy_to_iter(src, size, &iter) != size) {
 			erofs_put_metabuf(&buf);
 			return -EFAULT;
 		}
 		iov_iter_zero(PAGE_SIZE - size, &iter);
 		erofs_put_metabuf(&buf);
-		primary->submitted += PAGE_SIZE;
+		req->submitted += PAGE_SIZE;
 		return 0;
 	}
 
-	count = primary->len - primary->submitted;
+	count = req->len - req->submitted;
 	if (!(map.m_flags & EROFS_MAP_MAPPED)) {
+		struct iov_iter iter;
+
 		iov_iter_xarray(&iter, ITER_DEST, &mapping->i_pages, pos, count);
 		iov_iter_zero(count, &iter);
-		primary->submitted += count;
+		req->submitted += count;
 		return 0;
 	}
 
@@ -260,18 +312,19 @@ static int erofs_fscache_data_read_slice(struct erofs_fscache_request *primary)
 	if (ret)
 		return ret;
 
-	req = erofs_fscache_req_chain(primary, count);
-	if (IS_ERR(req))
-		return PTR_ERR(req);
+	io = erofs_fscache_req_io_alloc(req);
+	if (!io)
+		return -ENOMEM;
+	iov_iter_xarray(&io->iter, ITER_DEST, &mapping->i_pages, pos, count);
+	ret = erofs_fscache_read_io_async(mdev.m_fscache->cookie,
+			mdev.m_pa + (pos - map.m_la), io);
+	erofs_fscache_req_io_put(io);
 
-	ret = erofs_fscache_read_folios_async(mdev.m_fscache->cookie,
-			req, mdev.m_pa + (pos - map.m_la), count);
-	erofs_fscache_req_put(req);
-	primary->submitted += count;
+	req->submitted += count;
 	return ret;
 }
 
-static int erofs_fscache_data_read(struct erofs_fscache_request *req)
+static int erofs_fscache_data_read(struct erofs_fscache_rq *req)
 {
 	int ret;
 
@@ -280,20 +333,19 @@ static int erofs_fscache_data_read(struct erofs_fscache_request *req)
 		if (ret)
 			req->error = ret;
 	} while (!ret && req->submitted < req->len);
-
 	return ret;
 }
 
 static int erofs_fscache_read_folio(struct file *file, struct folio *folio)
 {
-	struct erofs_fscache_request *req;
+	struct erofs_fscache_rq *req;
 	int ret;
 
 	req = erofs_fscache_req_alloc(folio->mapping,
 			folio_pos(folio), folio_size(folio));
-	if (IS_ERR(req)) {
+	if (!req) {
 		folio_unlock(folio);
-		return PTR_ERR(req);
+		return -ENOMEM;
 	}
 
 	ret = erofs_fscache_data_read(req);
@@ -303,14 +355,14 @@ static int erofs_fscache_read_folio(struct file *file, struct folio *folio)
 
 static void erofs_fscache_readahead(struct readahead_control *rac)
 {
-	struct erofs_fscache_request *req;
+	struct erofs_fscache_rq *req;
 
 	if (!readahead_count(rac))
 		return;
 
 	req = erofs_fscache_req_alloc(rac->mapping,
 			readahead_pos(rac), readahead_length(rac));
-	if (IS_ERR(req))
+	if (!req)
 		return;
 
 	/* The request completion will drop refs on the folios. */
diff --git a/fs/erofs/inode.c b/fs/erofs/inode.c
index 36e638e8b5..5f6439a63a 100644
--- a/fs/erofs/inode.c
+++ b/fs/erofs/inode.c
@@ -26,7 +26,7 @@ static void *erofs_read_inode(struct erofs_buf *buf,
 	blkaddr = erofs_blknr(sb, inode_loc);
 	*ofs = erofs_blkoff(sb, inode_loc);
 
-	kaddr = erofs_read_metabuf(buf, sb, blkaddr, EROFS_KMAP);
+	kaddr = erofs_read_metabuf(buf, sb, erofs_pos(sb, blkaddr), EROFS_KMAP);
 	if (IS_ERR(kaddr)) {
 		erofs_err(sb, "failed to get inode (nid: %llu) page, err %ld",
 			  vi->nid, PTR_ERR(kaddr));
@@ -66,7 +66,7 @@ static void *erofs_read_inode(struct erofs_buf *buf,
 				goto err_out;
 			}
 			memcpy(copied, dic, gotten);
-			kaddr = erofs_read_metabuf(buf, sb, blkaddr + 1,
+			kaddr = erofs_read_metabuf(buf, sb, erofs_pos(sb, blkaddr + 1),
 						   EROFS_KMAP);
 			if (IS_ERR(kaddr)) {
 				erofs_err(sb, "failed to get inode payload block (nid: %llu), err %ld",
@@ -259,14 +259,12 @@ static int erofs_fill_inode(struct inode *inode)
 
 	if (erofs_inode_is_data_compressed(vi->datalayout)) {
 #ifdef CONFIG_EROFS_FS_ZIP
-		if (!erofs_is_fscache_mode(inode->i_sb)) {
-			DO_ONCE_LITE_IF(inode->i_sb->s_blocksize != PAGE_SIZE,
-				  erofs_info, inode->i_sb,
-				  "EXPERIMENTAL EROFS subpage compressed block support in use. Use at your own risk!");
-			inode->i_mapping->a_ops = &z_erofs_aops;
-			err = 0;
-			goto out_unlock;
-		}
+		DO_ONCE_LITE_IF(inode->i_blkbits != PAGE_SHIFT,
+			  erofs_info, inode->i_sb,
+			  "EXPERIMENTAL EROFS subpage compressed block support in use. Use at your own risk!");
+		inode->i_mapping->a_ops = &z_erofs_aops;
+		err = 0;
+		goto out_unlock;
 #endif
 		err = -EOPNOTSUPP;
 		goto out_unlock;
diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
index c69174675c..0c1b44ac95 100644
--- a/fs/erofs/internal.h
+++ b/fs/erofs/internal.h
@@ -49,7 +49,7 @@ typedef u32 erofs_blk_t;
 struct erofs_device_info {
 	char *path;
 	struct erofs_fscache *fscache;
-	struct bdev_handle *bdev_handle;
+	struct file *bdev_file;
 	struct dax_device *dax_dev;
 	u64 dax_part_off;
 
@@ -64,15 +64,12 @@ enum {
 };
 
 struct erofs_mount_opts {
-#ifdef CONFIG_EROFS_FS_ZIP
 	/* current strategy of how to use managed cache */
 	unsigned char cache_strategy;
 	/* strategy of sync decompression (0 - auto, 1 - force on, 2 - force off) */
 	unsigned int sync_decompress;
-
 	/* threshold for decompression synchronously */
 	unsigned int max_sync_decompress_pages;
-#endif
 	unsigned int mount_opt;
 };
 
@@ -216,7 +213,7 @@ enum erofs_kmap_type {
 };
 
 struct erofs_buf {
-	struct inode *inode;
+	struct address_space *mapping;
 	struct page *page;
 	void *base;
 	enum erofs_kmap_type kmap_type;
@@ -402,11 +399,11 @@ void *erofs_read_metadata(struct super_block *sb, struct erofs_buf *buf,
 			  erofs_off_t *offset, int *lengthp);
 void erofs_unmap_metabuf(struct erofs_buf *buf);
 void erofs_put_metabuf(struct erofs_buf *buf);
-void *erofs_bread(struct erofs_buf *buf, erofs_blk_t blkaddr,
+void *erofs_bread(struct erofs_buf *buf, erofs_off_t offset,
 		  enum erofs_kmap_type type);
 void erofs_init_metabuf(struct erofs_buf *buf, struct super_block *sb);
 void *erofs_read_metabuf(struct erofs_buf *buf, struct super_block *sb,
-			 erofs_blk_t blkaddr, enum erofs_kmap_type type);
+			 erofs_off_t offset, enum erofs_kmap_type type);
 int erofs_map_dev(struct super_block *sb, struct erofs_map_dev *dev);
 int erofs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 		 u64 start, u64 len);
@@ -438,7 +435,11 @@ void erofs_unregister_sysfs(struct super_block *sb);
 int __init erofs_init_sysfs(void);
 void erofs_exit_sysfs(void);
 
-struct page *erofs_allocpage(struct page **pagepool, gfp_t gfp);
+struct page *__erofs_allocpage(struct page **pagepool, gfp_t gfp, bool tryrsv);
+static inline struct page *erofs_allocpage(struct page **pagepool, gfp_t gfp)
+{
+	return __erofs_allocpage(pagepool, gfp, false);
+}
 static inline void erofs_pagepool_add(struct page **pagepool, struct page *page)
 {
 	set_page_private(page, (unsigned long)*pagepool);
@@ -459,15 +460,15 @@ int __init erofs_init_shrinker(void);
 void erofs_exit_shrinker(void);
 int __init z_erofs_init_zip_subsystem(void);
 void z_erofs_exit_zip_subsystem(void);
-int erofs_try_to_free_all_cached_pages(struct erofs_sb_info *sbi,
-				       struct erofs_workgroup *egrp);
+int erofs_try_to_free_all_cached_folios(struct erofs_sb_info *sbi,
+					struct erofs_workgroup *egrp);
 int z_erofs_map_blocks_iter(struct inode *inode, struct erofs_map_blocks *map,
 			    int flags);
-void *erofs_get_pcpubuf(unsigned int requiredpages);
-void erofs_put_pcpubuf(void *ptr);
-int erofs_pcpubuf_growsize(unsigned int nrpages);
-void __init erofs_pcpubuf_init(void);
-void erofs_pcpubuf_exit(void);
+void *z_erofs_get_gbuf(unsigned int requiredpages);
+void z_erofs_put_gbuf(void *ptr);
+int z_erofs_gbuf_growsize(unsigned int nrpages);
+int __init z_erofs_gbuf_init(void);
+void z_erofs_gbuf_exit(void);
 int erofs_init_managed_cache(struct super_block *sb);
 int z_erofs_parse_cfgs(struct super_block *sb, struct erofs_super_block *dsb);
 #else
@@ -477,8 +478,8 @@ static inline int erofs_init_shrinker(void) { return 0; }
 static inline void erofs_exit_shrinker(void) {}
 static inline int z_erofs_init_zip_subsystem(void) { return 0; }
 static inline void z_erofs_exit_zip_subsystem(void) {}
-static inline void erofs_pcpubuf_init(void) {}
-static inline void erofs_pcpubuf_exit(void) {}
+static inline int z_erofs_gbuf_init(void) { return 0; }
+static inline void z_erofs_gbuf_exit(void) {}
 static inline int erofs_init_managed_cache(struct super_block *sb) { return 0; }
 #endif	/* !CONFIG_EROFS_FS_ZIP */
 
@@ -498,6 +499,14 @@ static inline int z_erofs_deflate_init(void) { return 0; }
 static inline int z_erofs_deflate_exit(void) { return 0; }
 #endif	/* !CONFIG_EROFS_FS_ZIP_DEFLATE */
 
+#ifdef CONFIG_EROFS_FS_ZIP_ZSTD
+int __init z_erofs_zstd_init(void);
+void z_erofs_zstd_exit(void);
+#else
+static inline int z_erofs_zstd_init(void) { return 0; }
+static inline int z_erofs_zstd_exit(void) { return 0; }
+#endif	/* !CONFIG_EROFS_FS_ZIP_ZSTD */
+
 #ifdef CONFIG_EROFS_FS_ONDEMAND
 int erofs_fscache_register_fs(struct super_block *sb);
 void erofs_fscache_unregister_fs(struct super_block *sb);
@@ -505,6 +514,8 @@ void erofs_fscache_unregister_fs(struct super_block *sb);
 struct erofs_fscache *erofs_fscache_register_cookie(struct super_block *sb,
 					char *name, unsigned int flags);
 void erofs_fscache_unregister_cookie(struct erofs_fscache *fscache);
+struct bio *erofs_fscache_bio_alloc(struct erofs_map_dev *mdev);
+void erofs_fscache_submit_bio(struct bio *bio);
 #else
 static inline int erofs_fscache_register_fs(struct super_block *sb)
 {
@@ -522,6 +533,8 @@ struct erofs_fscache *erofs_fscache_register_cookie(struct super_block *sb,
 static inline void erofs_fscache_unregister_cookie(struct erofs_fscache *fscache)
 {
 }
+static inline struct bio *erofs_fscache_bio_alloc(struct erofs_map_dev *mdev) { return NULL; }
+static inline void erofs_fscache_submit_bio(struct bio *bio) {}
 #endif
 
 #define EFSCORRUPTED    EUCLEAN         /* Filesystem is corrupted */
diff --git a/fs/erofs/namei.c b/fs/erofs/namei.c
index f0110a78ac..c94d0c1608 100644
--- a/fs/erofs/namei.c
+++ b/fs/erofs/namei.c
@@ -99,8 +99,8 @@ static void *erofs_find_target_block(struct erofs_buf *target,
 		struct erofs_buf buf = __EROFS_BUF_INITIALIZER;
 		struct erofs_dirent *de;
 
-		buf.inode = dir;
-		de = erofs_bread(&buf, mid, EROFS_KMAP);
+		buf.mapping = dir->i_mapping;
+		de = erofs_bread(&buf, erofs_pos(dir->i_sb, mid), EROFS_KMAP);
 		if (!IS_ERR(de)) {
 			const int nameoff = nameoff_from_disk(de->nameoff, bsz);
 			const int ndirents = nameoff / sizeof(*de);
@@ -171,7 +171,7 @@ int erofs_namei(struct inode *dir, const struct qstr *name, erofs_nid_t *nid,
 
 	qn.name = name->name;
 	qn.end = name->name + name->len;
-	buf.inode = dir;
+	buf.mapping = dir->i_mapping;
 
 	ndirents = 0;
 	de = erofs_find_target_block(&buf, dir, &qn, &ndirents);
diff --git a/fs/erofs/pcpubuf.c b/fs/erofs/pcpubuf.c
deleted file mode 100644
index c7a4b1d770..0000000000
--- a/fs/erofs/pcpubuf.c
+++ /dev/null
@@ -1,148 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-only
-/*
- * Copyright (C) Gao Xiang <xiang@kernel.org>
- *
- * For low-latency decompression algorithms (e.g. lz4), reserve consecutive
- * per-CPU virtual memory (in pages) in advance to store such inplace I/O
- * data if inplace decompression is failed (due to unmet inplace margin for
- * example).
- */
-#include "internal.h"
-
-struct erofs_pcpubuf {
-	raw_spinlock_t lock;
-	void *ptr;
-	struct page **pages;
-	unsigned int nrpages;
-};
-
-static DEFINE_PER_CPU(struct erofs_pcpubuf, erofs_pcb);
-
-void *erofs_get_pcpubuf(unsigned int requiredpages)
-	__acquires(pcb->lock)
-{
-	struct erofs_pcpubuf *pcb = &get_cpu_var(erofs_pcb);
-
-	raw_spin_lock(&pcb->lock);
-	/* check if the per-CPU buffer is too small */
-	if (requiredpages > pcb->nrpages) {
-		raw_spin_unlock(&pcb->lock);
-		put_cpu_var(erofs_pcb);
-		/* (for sparse checker) pretend pcb->lock is still taken */
-		__acquire(pcb->lock);
-		return NULL;
-	}
-	return pcb->ptr;
-}
-
-void erofs_put_pcpubuf(void *ptr) __releases(pcb->lock)
-{
-	struct erofs_pcpubuf *pcb = &per_cpu(erofs_pcb, smp_processor_id());
-
-	DBG_BUGON(pcb->ptr != ptr);
-	raw_spin_unlock(&pcb->lock);
-	put_cpu_var(erofs_pcb);
-}
-
-/* the next step: support per-CPU page buffers hotplug */
-int erofs_pcpubuf_growsize(unsigned int nrpages)
-{
-	static DEFINE_MUTEX(pcb_resize_mutex);
-	static unsigned int pcb_nrpages;
-	struct page *pagepool = NULL;
-	int delta, cpu, ret, i;
-
-	mutex_lock(&pcb_resize_mutex);
-	delta = nrpages - pcb_nrpages;
-	ret = 0;
-	/* avoid shrinking pcpubuf, since no idea how many fses rely on */
-	if (delta <= 0)
-		goto out;
-
-	for_each_possible_cpu(cpu) {
-		struct erofs_pcpubuf *pcb = &per_cpu(erofs_pcb, cpu);
-		struct page **pages, **oldpages;
-		void *ptr, *old_ptr;
-
-		pages = kmalloc_array(nrpages, sizeof(*pages), GFP_KERNEL);
-		if (!pages) {
-			ret = -ENOMEM;
-			break;
-		}
-
-		for (i = 0; i < nrpages; ++i) {
-			pages[i] = erofs_allocpage(&pagepool, GFP_KERNEL);
-			if (!pages[i]) {
-				ret = -ENOMEM;
-				oldpages = pages;
-				goto free_pagearray;
-			}
-		}
-		ptr = vmap(pages, nrpages, VM_MAP, PAGE_KERNEL);
-		if (!ptr) {
-			ret = -ENOMEM;
-			oldpages = pages;
-			goto free_pagearray;
-		}
-		raw_spin_lock(&pcb->lock);
-		old_ptr = pcb->ptr;
-		pcb->ptr = ptr;
-		oldpages = pcb->pages;
-		pcb->pages = pages;
-		i = pcb->nrpages;
-		pcb->nrpages = nrpages;
-		raw_spin_unlock(&pcb->lock);
-
-		if (!oldpages) {
-			DBG_BUGON(old_ptr);
-			continue;
-		}
-
-		if (old_ptr)
-			vunmap(old_ptr);
-free_pagearray:
-		while (i)
-			erofs_pagepool_add(&pagepool, oldpages[--i]);
-		kfree(oldpages);
-		if (ret)
-			break;
-	}
-	pcb_nrpages = nrpages;
-	erofs_release_pages(&pagepool);
-out:
-	mutex_unlock(&pcb_resize_mutex);
-	return ret;
-}
-
-void __init erofs_pcpubuf_init(void)
-{
-	int cpu;
-
-	for_each_possible_cpu(cpu) {
-		struct erofs_pcpubuf *pcb = &per_cpu(erofs_pcb, cpu);
-
-		raw_spin_lock_init(&pcb->lock);
-	}
-}
-
-void erofs_pcpubuf_exit(void)
-{
-	int cpu, i;
-
-	for_each_possible_cpu(cpu) {
-		struct erofs_pcpubuf *pcb = &per_cpu(erofs_pcb, cpu);
-
-		if (pcb->ptr) {
-			vunmap(pcb->ptr);
-			pcb->ptr = NULL;
-		}
-		if (!pcb->pages)
-			continue;
-
-		for (i = 0; i < pcb->nrpages; ++i)
-			if (pcb->pages[i])
-				put_page(pcb->pages[i]);
-		kfree(pcb->pages);
-		pcb->pages = NULL;
-	}
-}
diff --git a/fs/erofs/super.c b/fs/erofs/super.c
index a2fa745585..1b91d95130 100644
--- a/fs/erofs/super.c
+++ b/fs/erofs/super.c
@@ -132,11 +132,11 @@ void *erofs_read_metadata(struct super_block *sb, struct erofs_buf *buf,
 	int len, i, cnt;
 
 	*offset = round_up(*offset, 4);
-	ptr = erofs_bread(buf, erofs_blknr(sb, *offset), EROFS_KMAP);
+	ptr = erofs_bread(buf, *offset, EROFS_KMAP);
 	if (IS_ERR(ptr))
 		return ptr;
 
-	len = le16_to_cpu(*(__le16 *)&ptr[erofs_blkoff(sb, *offset)]);
+	len = le16_to_cpu(*(__le16 *)ptr);
 	if (!len)
 		len = U16_MAX + 1;
 	buffer = kmalloc(len, GFP_KERNEL);
@@ -148,12 +148,12 @@ void *erofs_read_metadata(struct super_block *sb, struct erofs_buf *buf,
 	for (i = 0; i < len; i += cnt) {
 		cnt = min_t(int, sb->s_blocksize - erofs_blkoff(sb, *offset),
 			    len - i);
-		ptr = erofs_bread(buf, erofs_blknr(sb, *offset), EROFS_KMAP);
+		ptr = erofs_bread(buf, *offset, EROFS_KMAP);
 		if (IS_ERR(ptr)) {
 			kfree(buffer);
 			return ptr;
 		}
-		memcpy(buffer + i, ptr + erofs_blkoff(sb, *offset), cnt);
+		memcpy(buffer + i, ptr, cnt);
 		*offset += cnt;
 	}
 	return buffer;
@@ -177,13 +177,11 @@ static int erofs_init_device(struct erofs_buf *buf, struct super_block *sb,
 	struct erofs_sb_info *sbi = EROFS_SB(sb);
 	struct erofs_fscache *fscache;
 	struct erofs_deviceslot *dis;
-	struct bdev_handle *bdev_handle;
-	void *ptr;
+	struct file *bdev_file;
 
-	ptr = erofs_read_metabuf(buf, sb, erofs_blknr(sb, *pos), EROFS_KMAP);
-	if (IS_ERR(ptr))
-		return PTR_ERR(ptr);
-	dis = ptr + erofs_blkoff(sb, *pos);
+	dis = erofs_read_metabuf(buf, sb, *pos, EROFS_KMAP);
+	if (IS_ERR(dis))
+		return PTR_ERR(dis);
 
 	if (!sbi->devs->flatdev && !dif->path) {
 		if (!dis->tag[0]) {
@@ -201,12 +199,12 @@ static int erofs_init_device(struct erofs_buf *buf, struct super_block *sb,
 			return PTR_ERR(fscache);
 		dif->fscache = fscache;
 	} else if (!sbi->devs->flatdev) {
-		bdev_handle = bdev_open_by_path(dif->path, BLK_OPEN_READ,
+		bdev_file = bdev_file_open_by_path(dif->path, BLK_OPEN_READ,
 						sb->s_type, NULL);
-		if (IS_ERR(bdev_handle))
-			return PTR_ERR(bdev_handle);
-		dif->bdev_handle = bdev_handle;
-		dif->dax_dev = fs_dax_get_by_bdev(bdev_handle->bdev,
+		if (IS_ERR(bdev_file))
+			return PTR_ERR(bdev_file);
+		dif->bdev_file = bdev_file;
+		dif->dax_dev = fs_dax_get_by_bdev(file_bdev(bdev_file),
 				&dif->dax_part_off, NULL, NULL);
 	}
 
@@ -345,7 +343,7 @@ static int erofs_read_superblock(struct super_block *sb)
 	sbi->build_time = le64_to_cpu(dsb->build_time);
 	sbi->build_time_nsec = le32_to_cpu(dsb->build_time_nsec);
 
-	memcpy(&sb->s_uuid, dsb->uuid, sizeof(dsb->uuid));
+	super_set_uuid(sb, (void *)dsb->uuid, sizeof(dsb->uuid));
 
 	ret = strscpy(sbi->volume_name, dsb->volume_name,
 		      sizeof(dsb->volume_name));
@@ -430,7 +428,6 @@ static bool erofs_fc_set_dax_mode(struct fs_context *fc, unsigned int mode)
 
 	switch (mode) {
 	case EROFS_MOUNT_DAX_ALWAYS:
-		warnfc(fc, "DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
 		set_opt(&sbi->opt, DAX_ALWAYS);
 		clear_opt(&sbi->opt, DAX_NEVER);
 		return true;
@@ -728,8 +725,8 @@ static int erofs_release_device_info(int id, void *ptr, void *data)
 	struct erofs_device_info *dif = ptr;
 
 	fs_put_dax(dif->dax_dev, NULL);
-	if (dif->bdev_handle)
-		bdev_release(dif->bdev_handle);
+	if (dif->bdev_file)
+		fput(dif->bdev_file);
 	erofs_fscache_unregister_cookie(dif->fscache);
 	dif->fscache = NULL;
 	kfree(dif->path);
@@ -843,7 +840,7 @@ static int __init erofs_module_init(void)
 
 	erofs_inode_cachep = kmem_cache_create("erofs_inode",
 			sizeof(struct erofs_inode), 0,
-			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD | SLAB_ACCOUNT,
+			SLAB_RECLAIM_ACCOUNT | SLAB_ACCOUNT,
 			erofs_inode_init_once);
 	if (!erofs_inode_cachep)
 		return -ENOMEM;
@@ -860,7 +857,14 @@ static int __init erofs_module_init(void)
 	if (err)
 		goto deflate_err;
 
-	erofs_pcpubuf_init();
+	err = z_erofs_zstd_init();
+	if (err)
+		goto zstd_err;
+
+	err = z_erofs_gbuf_init();
+	if (err)
+		goto gbuf_err;
+
 	err = z_erofs_init_zip_subsystem();
 	if (err)
 		goto zip_err;
@@ -880,6 +884,10 @@ fs_err:
 sysfs_err:
 	z_erofs_exit_zip_subsystem();
 zip_err:
+	z_erofs_gbuf_exit();
+gbuf_err:
+	z_erofs_zstd_exit();
+zstd_err:
 	z_erofs_deflate_exit();
 deflate_err:
 	z_erofs_lzma_exit();
@@ -899,33 +907,32 @@ static void __exit erofs_module_exit(void)
 
 	erofs_exit_sysfs();
 	z_erofs_exit_zip_subsystem();
+	z_erofs_zstd_exit();
 	z_erofs_deflate_exit();
 	z_erofs_lzma_exit();
 	erofs_exit_shrinker();
 	kmem_cache_destroy(erofs_inode_cachep);
-	erofs_pcpubuf_exit();
+	z_erofs_gbuf_exit();
 }
 
 static int erofs_statfs(struct dentry *dentry, struct kstatfs *buf)
 {
 	struct super_block *sb = dentry->d_sb;
 	struct erofs_sb_info *sbi = EROFS_SB(sb);
-	u64 id = 0;
-
-	if (!erofs_is_fscache_mode(sb))
-		id = huge_encode_dev(sb->s_bdev->bd_dev);
 
 	buf->f_type = sb->s_magic;
 	buf->f_bsize = sb->s_blocksize;
 	buf->f_blocks = sbi->total_blocks;
 	buf->f_bfree = buf->f_bavail = 0;
-
 	buf->f_files = ULLONG_MAX;
 	buf->f_ffree = ULLONG_MAX - sbi->inos;
-
 	buf->f_namelen = EROFS_NAME_LEN;
 
-	buf->f_fsid    = u64_to_fsid(id);
+	if (uuid_is_null(&sb->s_uuid))
+		buf->f_fsid = u64_to_fsid(erofs_is_fscache_mode(sb) ? 0 :
+				huge_encode_dev(sb->s_bdev->bd_dev));
+	else
+		buf->f_fsid = uuid_to_fsid(sb->s_uuid.b);
 	return 0;
 }
 
@@ -934,26 +941,14 @@ static int erofs_show_options(struct seq_file *seq, struct dentry *root)
 	struct erofs_sb_info *sbi = EROFS_SB(root->d_sb);
 	struct erofs_mount_opts *opt = &sbi->opt;
 
-#ifdef CONFIG_EROFS_FS_XATTR
-	if (test_opt(opt, XATTR_USER))
-		seq_puts(seq, ",user_xattr");
-	else
-		seq_puts(seq, ",nouser_xattr");
-#endif
-#ifdef CONFIG_EROFS_FS_POSIX_ACL
-	if (test_opt(opt, POSIX_ACL))
-		seq_puts(seq, ",acl");
-	else
-		seq_puts(seq, ",noacl");
-#endif
-#ifdef CONFIG_EROFS_FS_ZIP
-	if (opt->cache_strategy == EROFS_ZIP_CACHE_DISABLED)
-		seq_puts(seq, ",cache_strategy=disabled");
-	else if (opt->cache_strategy == EROFS_ZIP_CACHE_READAHEAD)
-		seq_puts(seq, ",cache_strategy=readahead");
-	else if (opt->cache_strategy == EROFS_ZIP_CACHE_READAROUND)
-		seq_puts(seq, ",cache_strategy=readaround");
-#endif
+	if (IS_ENABLED(CONFIG_EROFS_FS_XATTR))
+		seq_puts(seq, test_opt(opt, XATTR_USER) ?
+				",user_xattr" : ",nouser_xattr");
+	if (IS_ENABLED(CONFIG_EROFS_FS_POSIX_ACL))
+		seq_puts(seq, test_opt(opt, POSIX_ACL) ? ",acl" : ",noacl");
+	if (IS_ENABLED(CONFIG_EROFS_FS_ZIP))
+		seq_printf(seq, ",cache_strategy=%s",
+			  erofs_param_cache_strategy[opt->cache_strategy].name);
 	if (test_opt(opt, DAX_ALWAYS))
 		seq_puts(seq, ",dax=always");
 	if (test_opt(opt, DAX_NEVER))
diff --git a/fs/erofs/xattr.c b/fs/erofs/xattr.c
index b58316b49a..a90d7d6497 100644
--- a/fs/erofs/xattr.c
+++ b/fs/erofs/xattr.c
@@ -81,13 +81,13 @@ static int erofs_init_inode_xattrs(struct inode *inode)
 	it.pos = erofs_iloc(inode) + vi->inode_isize;
 
 	/* read in shared xattr array (non-atomic, see kmalloc below) */
-	it.kaddr = erofs_bread(&it.buf, erofs_blknr(sb, it.pos), EROFS_KMAP);
+	it.kaddr = erofs_bread(&it.buf, it.pos, EROFS_KMAP);
 	if (IS_ERR(it.kaddr)) {
 		ret = PTR_ERR(it.kaddr);
 		goto out_unlock;
 	}
 
-	ih = it.kaddr + erofs_blkoff(sb, it.pos);
+	ih = it.kaddr;
 	vi->xattr_name_filter = le32_to_cpu(ih->h_name_filter);
 	vi->xattr_shared_count = ih->h_shared_count;
 	vi->xattr_shared_xattrs = kmalloc_array(vi->xattr_shared_count,
@@ -102,16 +102,14 @@ static int erofs_init_inode_xattrs(struct inode *inode)
 	it.pos += sizeof(struct erofs_xattr_ibody_header);
 
 	for (i = 0; i < vi->xattr_shared_count; ++i) {
-		it.kaddr = erofs_bread(&it.buf, erofs_blknr(sb, it.pos),
-				       EROFS_KMAP);
+		it.kaddr = erofs_bread(&it.buf, it.pos, EROFS_KMAP);
 		if (IS_ERR(it.kaddr)) {
 			kfree(vi->xattr_shared_xattrs);
 			vi->xattr_shared_xattrs = NULL;
 			ret = PTR_ERR(it.kaddr);
 			goto out_unlock;
 		}
-		vi->xattr_shared_xattrs[i] = le32_to_cpu(*(__le32 *)
-				(it.kaddr + erofs_blkoff(sb, it.pos)));
+		vi->xattr_shared_xattrs[i] = le32_to_cpu(*(__le32 *)it.kaddr);
 		it.pos += sizeof(__le32);
 	}
 	erofs_put_metabuf(&it.buf);
@@ -185,12 +183,11 @@ static int erofs_xattr_copy_to_buffer(struct erofs_xattr_iter *it,
 	void *src;
 
 	for (processed = 0; processed < len; processed += slice) {
-		it->kaddr = erofs_bread(&it->buf, erofs_blknr(sb, it->pos),
-					EROFS_KMAP);
+		it->kaddr = erofs_bread(&it->buf, it->pos, EROFS_KMAP);
 		if (IS_ERR(it->kaddr))
 			return PTR_ERR(it->kaddr);
 
-		src = it->kaddr + erofs_blkoff(sb, it->pos);
+		src = it->kaddr;
 		slice = min_t(unsigned int, sb->s_blocksize -
 				erofs_blkoff(sb, it->pos), len - processed);
 		memcpy(it->buffer + it->buffer_ofs, src, slice);
@@ -208,8 +205,7 @@ static int erofs_listxattr_foreach(struct erofs_xattr_iter *it)
 	int err;
 
 	/* 1. handle xattr entry */
-	entry = *(struct erofs_xattr_entry *)
-			(it->kaddr + erofs_blkoff(it->sb, it->pos));
+	entry = *(struct erofs_xattr_entry *)it->kaddr;
 	it->pos += sizeof(struct erofs_xattr_entry);
 
 	base_index = entry.e_name_index;
@@ -259,8 +255,7 @@ static int erofs_getxattr_foreach(struct erofs_xattr_iter *it)
 	unsigned int slice, processed, value_sz;
 
 	/* 1. handle xattr entry */
-	entry = *(struct erofs_xattr_entry *)
-			(it->kaddr + erofs_blkoff(sb, it->pos));
+	entry = *(struct erofs_xattr_entry *)it->kaddr;
 	it->pos += sizeof(struct erofs_xattr_entry);
 	value_sz = le16_to_cpu(entry.e_value_size);
 
@@ -291,8 +286,7 @@ static int erofs_getxattr_foreach(struct erofs_xattr_iter *it)
 
 	/* 2. handle xattr name */
 	for (processed = 0; processed < entry.e_name_len; processed += slice) {
-		it->kaddr = erofs_bread(&it->buf, erofs_blknr(sb, it->pos),
-					EROFS_KMAP);
+		it->kaddr = erofs_bread(&it->buf, it->pos, EROFS_KMAP);
 		if (IS_ERR(it->kaddr))
 			return PTR_ERR(it->kaddr);
 
@@ -300,7 +294,7 @@ static int erofs_getxattr_foreach(struct erofs_xattr_iter *it)
 				sb->s_blocksize - erofs_blkoff(sb, it->pos),
 				entry.e_name_len - processed);
 		if (memcmp(it->name.name + it->infix_len + processed,
-			   it->kaddr + erofs_blkoff(sb, it->pos), slice))
+			   it->kaddr, slice))
 			return -ENOATTR;
 		it->pos += slice;
 	}
@@ -336,13 +330,11 @@ static int erofs_xattr_iter_inline(struct erofs_xattr_iter *it,
 	it->pos = erofs_iloc(inode) + vi->inode_isize + xattr_header_sz;
 
 	while (remaining) {
-		it->kaddr = erofs_bread(&it->buf, erofs_blknr(it->sb, it->pos),
-					EROFS_KMAP);
+		it->kaddr = erofs_bread(&it->buf, it->pos, EROFS_KMAP);
 		if (IS_ERR(it->kaddr))
 			return PTR_ERR(it->kaddr);
 
-		entry_sz = erofs_xattr_entry_size(it->kaddr +
-				erofs_blkoff(it->sb, it->pos));
+		entry_sz = erofs_xattr_entry_size(it->kaddr);
 		/* xattr on-disk corruption: xattr entry beyond xattr_isize */
 		if (remaining < entry_sz) {
 			DBG_BUGON(1);
@@ -375,8 +367,7 @@ static int erofs_xattr_iter_shared(struct erofs_xattr_iter *it,
 	for (i = 0; i < vi->xattr_shared_count; ++i) {
 		it->pos = erofs_pos(sb, sbi->xattr_blkaddr) +
 				vi->xattr_shared_xattrs[i] * sizeof(__le32);
-		it->kaddr = erofs_bread(&it->buf, erofs_blknr(sb, it->pos),
-					EROFS_KMAP);
+		it->kaddr = erofs_bread(&it->buf, it->pos, EROFS_KMAP);
 		if (IS_ERR(it->kaddr))
 			return PTR_ERR(it->kaddr);
 
@@ -492,7 +483,7 @@ int erofs_xattr_prefixes_init(struct super_block *sb)
 		return -ENOMEM;
 
 	if (sbi->packed_inode)
-		buf.inode = sbi->packed_inode;
+		buf.mapping = sbi->packed_inode->i_mapping;
 	else
 		erofs_init_metabuf(&buf, sb);
 
diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c
index ff0aa72b0d..d6fe002a4a 100644
--- a/fs/erofs/zdata.c
+++ b/fs/erofs/zdata.c
@@ -19,7 +19,10 @@
 typedef void *z_erofs_next_pcluster_t;
 
 struct z_erofs_bvec {
-	struct page *page;
+	union {
+		struct page *page;
+		struct folio *folio;
+	};
 	int offset;
 	unsigned int end;
 };
@@ -116,47 +119,46 @@ static inline unsigned int z_erofs_pclusterpages(struct z_erofs_pcluster *pcl)
 	return PAGE_ALIGN(pcl->pclustersize) >> PAGE_SHIFT;
 }
 
+#define MNGD_MAPPING(sbi)	((sbi)->managed_cache->i_mapping)
+static bool erofs_folio_is_managed(struct erofs_sb_info *sbi, struct folio *fo)
+{
+	return fo->mapping == MNGD_MAPPING(sbi);
+}
+
 /*
- * bit 30: I/O error occurred on this page
- * bit 0 - 29: remaining parts to complete this page
+ * bit 30: I/O error occurred on this folio
+ * bit 0 - 29: remaining parts to complete this folio
  */
-#define Z_EROFS_PAGE_EIO			(1 << 30)
+#define Z_EROFS_FOLIO_EIO			(1 << 30)
 
-static inline void z_erofs_onlinepage_init(struct page *page)
+static void z_erofs_onlinefolio_init(struct folio *folio)
 {
 	union {
 		atomic_t o;
-		unsigned long v;
+		void *v;
 	} u = { .o = ATOMIC_INIT(1) };
 
-	set_page_private(page, u.v);
-	smp_wmb();
-	SetPagePrivate(page);
+	folio->private = u.v;	/* valid only if file-backed folio is locked */
 }
 
-static inline void z_erofs_onlinepage_split(struct page *page)
+static void z_erofs_onlinefolio_split(struct folio *folio)
 {
-	atomic_inc((atomic_t *)&page->private);
+	atomic_inc((atomic_t *)&folio->private);
 }
 
-static void z_erofs_onlinepage_endio(struct page *page, int err)
+static void z_erofs_onlinefolio_end(struct folio *folio, int err)
 {
 	int orig, v;
 
-	DBG_BUGON(!PagePrivate(page));
-
 	do {
-		orig = atomic_read((atomic_t *)&page->private);
-		v = (orig - 1) | (err ? Z_EROFS_PAGE_EIO : 0);
-	} while (atomic_cmpxchg((atomic_t *)&page->private, orig, v) != orig);
-
-	if (!(v & ~Z_EROFS_PAGE_EIO)) {
-		set_page_private(page, 0);
-		ClearPagePrivate(page);
-		if (!(v & Z_EROFS_PAGE_EIO))
-			SetPageUptodate(page);
-		unlock_page(page);
-	}
+		orig = atomic_read((atomic_t *)&folio->private);
+		v = (orig - 1) | (err ? Z_EROFS_FOLIO_EIO : 0);
+	} while (atomic_cmpxchg((atomic_t *)&folio->private, orig, v) != orig);
+
+	if (v & ~Z_EROFS_FOLIO_EIO)
+		return;
+	folio->private = 0;
+	folio_end_read(folio, !(v & Z_EROFS_FOLIO_EIO));
 }
 
 #define Z_EROFS_ONSTACK_PAGES		32
@@ -572,17 +574,13 @@ static void z_erofs_bind_cache(struct z_erofs_decompress_frontend *fe)
 
 	for (i = 0; i < pclusterpages; ++i) {
 		struct page *page, *newpage;
-		void *t;	/* mark pages just found for debugging */
 
 		/* Inaccurate check w/o locking to avoid unneeded lookups */
 		if (READ_ONCE(pcl->compressed_bvecs[i].page))
 			continue;
 
 		page = find_get_page(mc, pcl->obj.index + i);
-		if (page) {
-			t = (void *)((unsigned long)page | 1);
-			newpage = NULL;
-		} else {
+		if (!page) {
 			/* I/O is needed, no possible to decompress directly */
 			standalone = false;
 			if (!shouldalloc)
@@ -596,11 +594,10 @@ static void z_erofs_bind_cache(struct z_erofs_decompress_frontend *fe)
 			if (!newpage)
 				continue;
 			set_page_private(newpage, Z_EROFS_PREALLOCATED_PAGE);
-			t = (void *)((unsigned long)newpage | 1);
 		}
 		spin_lock(&pcl->obj.lockref.lock);
 		if (!pcl->compressed_bvecs[i].page) {
-			pcl->compressed_bvecs[i].page = t;
+			pcl->compressed_bvecs[i].page = page ? page : newpage;
 			spin_unlock(&pcl->obj.lockref.lock);
 			continue;
 		}
@@ -620,9 +617,9 @@ static void z_erofs_bind_cache(struct z_erofs_decompress_frontend *fe)
 		fe->mode = Z_EROFS_PCLUSTER_FOLLOWED_NOINPLACE;
 }
 
-/* called by erofs_shrinker to get rid of all compressed_pages */
-int erofs_try_to_free_all_cached_pages(struct erofs_sb_info *sbi,
-				       struct erofs_workgroup *grp)
+/* called by erofs_shrinker to get rid of all cached compressed bvecs */
+int erofs_try_to_free_all_cached_folios(struct erofs_sb_info *sbi,
+					struct erofs_workgroup *grp)
 {
 	struct z_erofs_pcluster *const pcl =
 		container_of(grp, struct z_erofs_pcluster, obj);
@@ -630,27 +627,22 @@ int erofs_try_to_free_all_cached_pages(struct erofs_sb_info *sbi,
 	int i;
 
 	DBG_BUGON(z_erofs_is_inline_pcluster(pcl));
-	/*
-	 * refcount of workgroup is now freezed as 0,
-	 * therefore no need to worry about available decompression users.
-	 */
+	/* There is no actice user since the pcluster is now freezed */
 	for (i = 0; i < pclusterpages; ++i) {
-		struct page *page = pcl->compressed_bvecs[i].page;
+		struct folio *folio = pcl->compressed_bvecs[i].folio;
 
-		if (!page)
+		if (!folio)
 			continue;
 
-		/* block other users from reclaiming or migrating the page */
-		if (!trylock_page(page))
+		/* Avoid reclaiming or migrating this folio */
+		if (!folio_trylock(folio))
 			return -EBUSY;
 
-		if (!erofs_page_is_managed(sbi, page))
+		if (!erofs_folio_is_managed(sbi, folio))
 			continue;
-
-		/* barrier is implied in the following 'unlock_page' */
-		WRITE_ONCE(pcl->compressed_bvecs[i].page, NULL);
-		detach_page_private(page);
-		unlock_page(page);
+		pcl->compressed_bvecs[i].folio = NULL;
+		folio_detach_private(folio);
+		folio_unlock(folio);
 	}
 	return 0;
 }
@@ -667,20 +659,17 @@ static bool z_erofs_cache_release_folio(struct folio *folio, gfp_t gfp)
 
 	ret = false;
 	spin_lock(&pcl->obj.lockref.lock);
-	if (pcl->obj.lockref.count > 0)
-		goto out;
-
-	DBG_BUGON(z_erofs_is_inline_pcluster(pcl));
-	for (i = 0; i < pclusterpages; ++i) {
-		if (pcl->compressed_bvecs[i].page == &folio->page) {
-			WRITE_ONCE(pcl->compressed_bvecs[i].page, NULL);
-			ret = true;
-			break;
+	if (pcl->obj.lockref.count <= 0) {
+		DBG_BUGON(z_erofs_is_inline_pcluster(pcl));
+		for (i = 0; i < pclusterpages; ++i) {
+			if (pcl->compressed_bvecs[i].folio == folio) {
+				pcl->compressed_bvecs[i].folio = NULL;
+				folio_detach_private(folio);
+				ret = true;
+				break;
+			}
 		}
 	}
-	if (ret)
-		folio_detach_private(folio);
-out:
 	spin_unlock(&pcl->obj.lockref.lock);
 	return ret;
 }
@@ -879,7 +868,7 @@ static int z_erofs_pcluster_begin(struct z_erofs_decompress_frontend *fe)
 	} else {
 		void *mptr;
 
-		mptr = erofs_read_metabuf(&map->buf, sb, blknr, EROFS_NO_KMAP);
+		mptr = erofs_read_metabuf(&map->buf, sb, map->m_pa, EROFS_NO_KMAP);
 		if (IS_ERR(mptr)) {
 			ret = PTR_ERR(mptr);
 			erofs_err(sb, "failed to get inline data %d", ret);
@@ -947,35 +936,35 @@ static int z_erofs_read_fragment(struct super_block *sb, struct page *page,
 	if (!packed_inode)
 		return -EFSCORRUPTED;
 
-	buf.inode = packed_inode;
+	buf.mapping = packed_inode->i_mapping;
 	for (; cur < end; cur += cnt, pos += cnt) {
 		cnt = min_t(unsigned int, end - cur,
 			    sb->s_blocksize - erofs_blkoff(sb, pos));
-		src = erofs_bread(&buf, erofs_blknr(sb, pos), EROFS_KMAP);
+		src = erofs_bread(&buf, pos, EROFS_KMAP);
 		if (IS_ERR(src)) {
 			erofs_put_metabuf(&buf);
 			return PTR_ERR(src);
 		}
-		memcpy_to_page(page, cur, src + erofs_blkoff(sb, pos), cnt);
+		memcpy_to_page(page, cur, src, cnt);
 	}
 	erofs_put_metabuf(&buf);
 	return 0;
 }
 
-static int z_erofs_do_read_page(struct z_erofs_decompress_frontend *fe,
-				struct page *page, bool ra)
+static int z_erofs_scan_folio(struct z_erofs_decompress_frontend *fe,
+			      struct folio *folio, bool ra)
 {
 	struct inode *const inode = fe->inode;
 	struct erofs_map_blocks *const map = &fe->map;
-	const loff_t offset = page_offset(page);
-	const unsigned int bs = i_blocksize(inode);
+	const loff_t offset = folio_pos(folio);
+	const unsigned int bs = i_blocksize(inode), fs = folio_size(folio);
 	bool tight = true, exclusive;
 	unsigned int cur, end, len, split;
 	int err = 0;
 
-	z_erofs_onlinepage_init(page);
+	z_erofs_onlinefolio_init(folio);
 	split = 0;
-	end = PAGE_SIZE;
+	end = fs;
 repeat:
 	if (offset + end - 1 < map->m_la ||
 	    offset + end - 1 >= map->m_la + map->m_llen) {
@@ -992,7 +981,7 @@ repeat:
 	++split;
 
 	if (!(map->m_flags & EROFS_MAP_MAPPED)) {
-		zero_user_segment(page, cur, end);
+		folio_zero_segment(folio, cur, end);
 		tight = false;
 		goto next_part;
 	}
@@ -1001,8 +990,8 @@ repeat:
 		erofs_off_t fpos = offset + cur - map->m_la;
 
 		len = min_t(unsigned int, map->m_llen - fpos, end - cur);
-		err = z_erofs_read_fragment(inode->i_sb, page, cur, cur + len,
-				EROFS_I(inode)->z_fragmentoff + fpos);
+		err = z_erofs_read_fragment(inode->i_sb, &folio->page, cur,
+			cur + len, EROFS_I(inode)->z_fragmentoff + fpos);
 		if (err)
 			goto out;
 		tight = false;
@@ -1017,25 +1006,25 @@ repeat:
 	}
 
 	/*
-	 * Ensure the current partial page belongs to this submit chain rather
+	 * Ensure the current partial folio belongs to this submit chain rather
 	 * than other concurrent submit chains or the noio(bypass) chain since
-	 * those chains are handled asynchronously thus the page cannot be used
+	 * those chains are handled asynchronously thus the folio cannot be used
 	 * for inplace I/O or bvpage (should be processed in a strict order.)
 	 */
 	tight &= (fe->mode > Z_EROFS_PCLUSTER_FOLLOWED_NOINPLACE);
-	exclusive = (!cur && ((split <= 1) || (tight && bs == PAGE_SIZE)));
+	exclusive = (!cur && ((split <= 1) || (tight && bs == fs)));
 	if (cur)
 		tight &= (fe->mode >= Z_EROFS_PCLUSTER_FOLLOWED);
 
 	err = z_erofs_attach_page(fe, &((struct z_erofs_bvec) {
-					.page = page,
+					.page = &folio->page,
 					.offset = offset - map->m_la,
 					.end = end,
 				  }), exclusive);
 	if (err)
 		goto out;
 
-	z_erofs_onlinepage_split(page);
+	z_erofs_onlinefolio_split(folio);
 	if (fe->pcl->pageofs_out != (map->m_la & ~PAGE_MASK))
 		fe->pcl->multibases = true;
 	if (fe->pcl->length < offset + end - map->m_la) {
@@ -1056,7 +1045,7 @@ next_part:
 		goto repeat;
 
 out:
-	z_erofs_onlinepage_endio(page, err);
+	z_erofs_onlinefolio_end(folio, err);
 	return err;
 }
 
@@ -1159,7 +1148,7 @@ static void z_erofs_fill_other_copies(struct z_erofs_decompress_backend *be,
 			cur += len;
 		}
 		kunmap_local(dst);
-		z_erofs_onlinepage_endio(bvi->bvec.page, err);
+		z_erofs_onlinefolio_end(page_folio(bvi->bvec.page), err);
 		list_del(p);
 		kfree(bvi);
 	}
@@ -1210,7 +1199,7 @@ static int z_erofs_parse_in_bvecs(struct z_erofs_decompress_backend *be,
 		be->compressed_pages[i] = page;
 
 		if (z_erofs_is_inline_pcluster(pcl) ||
-		    erofs_page_is_managed(EROFS_SB(be->sb), page)) {
+		    erofs_folio_is_managed(EROFS_SB(be->sb), page_folio(page))) {
 			if (!PageUptodate(page))
 				err = -EIO;
 			continue;
@@ -1295,7 +1284,8 @@ static int z_erofs_decompress_pcluster(struct z_erofs_decompress_backend *be,
 			/* consider shortlived pages added when decompressing */
 			page = be->compressed_pages[i];
 
-			if (!page || erofs_page_is_managed(sbi, page))
+			if (!page ||
+			    erofs_folio_is_managed(sbi, page_folio(page)))
 				continue;
 			(void)z_erofs_put_shortlivedpage(be->pagepool, page);
 			WRITE_ONCE(pcl->compressed_bvecs[i].page, NULL);
@@ -1316,7 +1306,7 @@ static int z_erofs_decompress_pcluster(struct z_erofs_decompress_backend *be,
 		/* recycle all individual short-lived pages */
 		if (z_erofs_put_shortlivedpage(be->pagepool, page))
 			continue;
-		z_erofs_onlinepage_endio(page, err);
+		z_erofs_onlinefolio_end(page_folio(page), err);
 	}
 
 	if (be->decompressed_pages != be->onstack_pages)
@@ -1430,38 +1420,34 @@ static void z_erofs_fill_bio_vec(struct bio_vec *bvec,
 	struct z_erofs_bvec zbv;
 	struct address_space *mapping;
 	struct page *page;
-	int justfound, bs = i_blocksize(f->inode);
+	int bs = i_blocksize(f->inode);
 
-	/* Except for inplace pages, the entire page can be used for I/Os */
+	/* Except for inplace folios, the entire folio can be used for I/Os */
 	bvec->bv_offset = 0;
 	bvec->bv_len = PAGE_SIZE;
 repeat:
 	spin_lock(&pcl->obj.lockref.lock);
 	zbv = pcl->compressed_bvecs[nr];
-	page = zbv.page;
-	justfound = (unsigned long)page & 1UL;
-	page = (struct page *)((unsigned long)page & ~1UL);
-	pcl->compressed_bvecs[nr].page = page;
 	spin_unlock(&pcl->obj.lockref.lock);
-	if (!page)
-		goto out_allocpage;
+	if (!zbv.folio)
+		goto out_allocfolio;
 
-	bvec->bv_page = page;
-	DBG_BUGON(z_erofs_is_shortlived_page(page));
+	bvec->bv_page = &zbv.folio->page;
+	DBG_BUGON(z_erofs_is_shortlived_page(bvec->bv_page));
 	/*
-	 * Handle preallocated cached pages.  We tried to allocate such pages
+	 * Handle preallocated cached folios.  We tried to allocate such folios
 	 * without triggering direct reclaim.  If allocation failed, inplace
-	 * file-backed pages will be used instead.
+	 * file-backed folios will be used instead.
 	 */
-	if (page->private == Z_EROFS_PREALLOCATED_PAGE) {
-		set_page_private(page, 0);
+	if (zbv.folio->private == (void *)Z_EROFS_PREALLOCATED_PAGE) {
+		zbv.folio->private = 0;
 		tocache = true;
 		goto out_tocache;
 	}
 
-	mapping = READ_ONCE(page->mapping);
+	mapping = READ_ONCE(zbv.folio->mapping);
 	/*
-	 * File-backed pages for inplace I/Os are all locked steady,
+	 * File-backed folios for inplace I/Os are all locked steady,
 	 * therefore it is impossible for `mapping` to be NULL.
 	 */
 	if (mapping && mapping != mc) {
@@ -1471,26 +1457,21 @@ repeat:
 		return;
 	}
 
-	lock_page(page);
-	/* only true if page reclaim goes wrong, should never happen */
-	DBG_BUGON(justfound && PagePrivate(page));
-
-	/* the cached page is still in managed cache */
-	if (page->mapping == mc) {
+	folio_lock(zbv.folio);
+	if (zbv.folio->mapping == mc) {
 		/*
-		 * The cached page is still available but without a valid
-		 * `->private` pcluster hint.  Let's reconnect them.
+		 * The cached folio is still in managed cache but without
+		 * a valid `->private` pcluster hint.  Let's reconnect them.
 		 */
-		if (!PagePrivate(page)) {
-			DBG_BUGON(!justfound);
-			/* compressed_bvecs[] already takes a ref */
-			attach_page_private(page, pcl);
-			put_page(page);
+		if (!folio_test_private(zbv.folio)) {
+			folio_attach_private(zbv.folio, pcl);
+			/* compressed_bvecs[] already takes a ref before */
+			folio_put(zbv.folio);
 		}
 
 		/* no need to submit if it is already up-to-date */
-		if (PageUptodate(page)) {
-			unlock_page(page);
+		if (folio_test_uptodate(zbv.folio)) {
+			folio_unlock(zbv.folio);
 			bvec->bv_page = NULL;
 		}
 		return;
@@ -1500,34 +1481,32 @@ repeat:
 	 * It has been truncated, so it's unsafe to reuse this one. Let's
 	 * allocate a new page for compressed data.
 	 */
-	DBG_BUGON(page->mapping);
-	DBG_BUGON(!justfound);
-
+	DBG_BUGON(zbv.folio->mapping);
 	tocache = true;
-	unlock_page(page);
-	put_page(page);
-out_allocpage:
+	folio_unlock(zbv.folio);
+	folio_put(zbv.folio);
+out_allocfolio:
 	page = erofs_allocpage(&f->pagepool, gfp | __GFP_NOFAIL);
 	spin_lock(&pcl->obj.lockref.lock);
-	if (pcl->compressed_bvecs[nr].page) {
+	if (pcl->compressed_bvecs[nr].folio) {
 		erofs_pagepool_add(&f->pagepool, page);
 		spin_unlock(&pcl->obj.lockref.lock);
 		cond_resched();
 		goto repeat;
 	}
-	pcl->compressed_bvecs[nr].page = page;
+	pcl->compressed_bvecs[nr].folio = zbv.folio = page_folio(page);
 	spin_unlock(&pcl->obj.lockref.lock);
 	bvec->bv_page = page;
 out_tocache:
 	if (!tocache || bs != PAGE_SIZE ||
-	    add_to_page_cache_lru(page, mc, pcl->obj.index + nr, gfp)) {
-		/* turn into a temporary shortlived page (1 ref) */
-		set_page_private(page, Z_EROFS_SHORTLIVED_PAGE);
+	    filemap_add_folio(mc, zbv.folio, pcl->obj.index + nr, gfp)) {
+		/* turn into a temporary shortlived folio (1 ref) */
+		zbv.folio->private = (void *)Z_EROFS_SHORTLIVED_PAGE;
 		return;
 	}
-	attach_page_private(page, pcl);
+	folio_attach_private(zbv.folio, pcl);
 	/* drop a refcount added by allocpage (then 2 refs in total here) */
-	put_page(page);
+	folio_put(zbv.folio);
 }
 
 static struct z_erofs_decompressqueue *jobqueue_init(struct super_block *sb,
@@ -1582,28 +1561,29 @@ static void move_to_bypass_jobqueue(struct z_erofs_pcluster *pcl,
 	qtail[JQ_BYPASS] = &pcl->next;
 }
 
-static void z_erofs_submissionqueue_endio(struct bio *bio)
+static void z_erofs_endio(struct bio *bio)
 {
 	struct z_erofs_decompressqueue *q = bio->bi_private;
 	blk_status_t err = bio->bi_status;
-	struct bio_vec *bvec;
-	struct bvec_iter_all iter_all;
+	struct folio_iter fi;
 
-	bio_for_each_segment_all(bvec, bio, iter_all) {
-		struct page *page = bvec->bv_page;
+	bio_for_each_folio_all(fi, bio) {
+		struct folio *folio = fi.folio;
 
-		DBG_BUGON(PageUptodate(page));
-		DBG_BUGON(z_erofs_page_is_invalidated(page));
-		if (erofs_page_is_managed(EROFS_SB(q->sb), page)) {
-			if (!err)
-				SetPageUptodate(page);
-			unlock_page(page);
-		}
+		DBG_BUGON(folio_test_uptodate(folio));
+		DBG_BUGON(z_erofs_page_is_invalidated(&folio->page));
+		if (!erofs_folio_is_managed(EROFS_SB(q->sb), folio))
+			continue;
+
+		if (!err)
+			folio_mark_uptodate(folio);
+		folio_unlock(folio);
 	}
 	if (err)
 		q->eio = true;
 	z_erofs_decompress_kickoff(q, -1);
-	bio_put(bio);
+	if (bio->bi_bdev)
+		bio_put(bio);
 }
 
 static void z_erofs_submit_queue(struct z_erofs_decompress_frontend *f,
@@ -1617,7 +1597,6 @@ static void z_erofs_submit_queue(struct z_erofs_decompress_frontend *f,
 	z_erofs_next_pcluster_t owned_head = f->owned_head;
 	/* bio is NULL initially, so no need to initialize last_{index,bdev} */
 	erofs_off_t last_pa;
-	struct block_device *last_bdev;
 	unsigned int nr_bios = 0;
 	struct bio *bio = NULL;
 	unsigned long pflags;
@@ -1664,9 +1643,13 @@ static void z_erofs_submit_queue(struct z_erofs_decompress_frontend *f,
 				continue;
 
 			if (bio && (cur != last_pa ||
-				    last_bdev != mdev.m_bdev)) {
-submit_bio_retry:
-				submit_bio(bio);
+				    bio->bi_bdev != mdev.m_bdev)) {
+io_retry:
+				if (!erofs_is_fscache_mode(sb))
+					submit_bio(bio);
+				else
+					erofs_fscache_submit_bio(bio);
+
 				if (memstall) {
 					psi_memstall_leave(&pflags);
 					memstall = 0;
@@ -1681,15 +1664,16 @@ submit_bio_retry:
 			}
 
 			if (!bio) {
-				bio = bio_alloc(mdev.m_bdev, BIO_MAX_VECS,
-						REQ_OP_READ, GFP_NOIO);
-				bio->bi_end_io = z_erofs_submissionqueue_endio;
+				bio = erofs_is_fscache_mode(sb) ?
+					erofs_fscache_bio_alloc(&mdev) :
+					bio_alloc(mdev.m_bdev, BIO_MAX_VECS,
+						  REQ_OP_READ, GFP_NOIO);
+				bio->bi_end_io = z_erofs_endio;
 				bio->bi_iter.bi_sector = cur >> 9;
 				bio->bi_private = q[JQ_SUBMIT];
 				if (readahead)
 					bio->bi_opf |= REQ_RAHEAD;
 				++nr_bios;
-				last_bdev = mdev.m_bdev;
 			}
 
 			if (cur + bvec.bv_len > end)
@@ -1697,7 +1681,7 @@ submit_bio_retry:
 			DBG_BUGON(bvec.bv_len < sb->s_blocksize);
 			if (!bio_add_page(bio, bvec.bv_page, bvec.bv_len,
 					  bvec.bv_offset))
-				goto submit_bio_retry;
+				goto io_retry;
 
 			last_pa = cur + bvec.bv_len;
 			bypass = false;
@@ -1710,7 +1694,10 @@ submit_bio_retry:
 	} while (owned_head != Z_EROFS_PCLUSTER_TAIL);
 
 	if (bio) {
-		submit_bio(bio);
+		if (!erofs_is_fscache_mode(sb))
+			submit_bio(bio);
+		else
+			erofs_fscache_submit_bio(bio);
 		if (memstall)
 			psi_memstall_leave(&pflags);
 	}
@@ -1795,7 +1782,7 @@ static void z_erofs_pcluster_readmore(struct z_erofs_decompress_frontend *f,
 			if (PageUptodate(page))
 				unlock_page(page);
 			else
-				(void)z_erofs_do_read_page(f, page, !!rac);
+				z_erofs_scan_folio(f, page_folio(page), !!rac);
 			put_page(page);
 		}
 
@@ -1816,7 +1803,7 @@ static int z_erofs_read_folio(struct file *file, struct folio *folio)
 	f.headoffset = (erofs_off_t)folio->index << PAGE_SHIFT;
 
 	z_erofs_pcluster_readmore(&f, NULL, true);
-	err = z_erofs_do_read_page(&f, &folio->page, false);
+	err = z_erofs_scan_folio(&f, folio, false);
 	z_erofs_pcluster_readmore(&f, NULL, false);
 	z_erofs_pcluster_end(&f);
 
@@ -1857,7 +1844,7 @@ static void z_erofs_readahead(struct readahead_control *rac)
 		folio = head;
 		head = folio_get_private(folio);
 
-		err = z_erofs_do_read_page(&f, &folio->page, true);
+		err = z_erofs_scan_folio(&f, folio, true);
 		if (err && err != -EINTR)
 			erofs_err(inode->i_sb, "readahead error at folio %lu @ nid %llu",
 				  folio->index, EROFS_I(inode)->nid);
diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c
index e313c93635..74d3d7bffc 100644
--- a/fs/erofs/zmap.c
+++ b/fs/erofs/zmap.c
@@ -31,22 +31,20 @@ static int z_erofs_load_full_lcluster(struct z_erofs_maprecorder *m,
 			vi->inode_isize + vi->xattr_isize) +
 			lcn * sizeof(struct z_erofs_lcluster_index);
 	struct z_erofs_lcluster_index *di;
-	unsigned int advise, type;
+	unsigned int advise;
 
 	m->kaddr = erofs_read_metabuf(&m->map->buf, inode->i_sb,
-				      erofs_blknr(inode->i_sb, pos), EROFS_KMAP);
+				      pos, EROFS_KMAP);
 	if (IS_ERR(m->kaddr))
 		return PTR_ERR(m->kaddr);
 
 	m->nextpackoff = pos + sizeof(struct z_erofs_lcluster_index);
 	m->lcn = lcn;
-	di = m->kaddr + erofs_blkoff(inode->i_sb, pos);
+	di = m->kaddr;
 
 	advise = le16_to_cpu(di->di_advise);
-	type = (advise >> Z_EROFS_LI_LCLUSTER_TYPE_BIT) &
-		((1 << Z_EROFS_LI_LCLUSTER_TYPE_BITS) - 1);
-	switch (type) {
-	case Z_EROFS_LCLUSTER_TYPE_NONHEAD:
+	m->type = advise & Z_EROFS_LI_LCLUSTER_TYPE_MASK;
+	if (m->type == Z_EROFS_LCLUSTER_TYPE_NONHEAD) {
 		m->clusterofs = 1 << vi->z_logical_clusterbits;
 		m->delta[0] = le16_to_cpu(di->di_u.delta[0]);
 		if (m->delta[0] & Z_EROFS_LI_D0_CBLKCNT) {
@@ -60,24 +58,15 @@ static int z_erofs_load_full_lcluster(struct z_erofs_maprecorder *m,
 			m->delta[0] = 1;
 		}
 		m->delta[1] = le16_to_cpu(di->di_u.delta[1]);
-		break;
-	case Z_EROFS_LCLUSTER_TYPE_PLAIN:
-	case Z_EROFS_LCLUSTER_TYPE_HEAD1:
-	case Z_EROFS_LCLUSTER_TYPE_HEAD2:
-		if (advise & Z_EROFS_LI_PARTIAL_REF)
-			m->partialref = true;
+	} else {
+		m->partialref = !!(advise & Z_EROFS_LI_PARTIAL_REF);
 		m->clusterofs = le16_to_cpu(di->di_clusterofs);
 		if (m->clusterofs >= 1 << vi->z_logical_clusterbits) {
 			DBG_BUGON(1);
 			return -EFSCORRUPTED;
 		}
 		m->pblk = le32_to_cpu(di->di_u.blkaddr);
-		break;
-	default:
-		DBG_BUGON(1);
-		return -EOPNOTSUPP;
 	}
-	m->type = type;
 	return 0;
 }
 
@@ -120,7 +109,7 @@ static int unpack_compacted_index(struct z_erofs_maprecorder *m,
 {
 	struct erofs_inode *const vi = EROFS_I(m->inode);
 	const unsigned int lclusterbits = vi->z_logical_clusterbits;
-	unsigned int vcnt, base, lo, lobits, encodebits, nblk, eofs;
+	unsigned int vcnt, lo, lobits, encodebits, nblk, bytes;
 	int i;
 	u8 *in, type;
 	bool big_pcluster;
@@ -138,11 +127,11 @@ static int unpack_compacted_index(struct z_erofs_maprecorder *m,
 	big_pcluster = vi->z_advise & Z_EROFS_ADVISE_BIG_PCLUSTER_1;
 	lobits = max(lclusterbits, ilog2(Z_EROFS_LI_D0_CBLKCNT) + 1U);
 	encodebits = ((vcnt << amortizedshift) - sizeof(__le32)) * 8 / vcnt;
-	eofs = erofs_blkoff(m->inode->i_sb, pos);
-	base = round_down(eofs, vcnt << amortizedshift);
-	in = m->kaddr + base;
+	bytes = pos & ((vcnt << amortizedshift) - 1);
 
-	i = (eofs - base) >> amortizedshift;
+	in = m->kaddr - bytes;
+
+	i = bytes >> amortizedshift;
 
 	lo = decode_compactedbits(lobits, in, encodebits * i, &type);
 	m->type = type;
@@ -267,7 +256,7 @@ static int z_erofs_load_compact_lcluster(struct z_erofs_maprecorder *m,
 out:
 	pos += lcn * (1 << amortizedshift);
 	m->kaddr = erofs_read_metabuf(&m->map->buf, inode->i_sb,
-				      erofs_blknr(inode->i_sb, pos), EROFS_KMAP);
+				      pos, EROFS_KMAP);
 	if (IS_ERR(m->kaddr))
 		return PTR_ERR(m->kaddr);
 	return unpack_compacted_index(m, amortizedshift, pos, lookahead);
@@ -561,7 +550,8 @@ static int z_erofs_do_map_blocks(struct inode *inode,
 	if ((flags & EROFS_GET_BLOCKS_FIEMAP) ||
 	    ((flags & EROFS_GET_BLOCKS_READMORE) &&
 	     (map->m_algorithmformat == Z_EROFS_COMPRESSION_LZMA ||
-	      map->m_algorithmformat == Z_EROFS_COMPRESSION_DEFLATE) &&
+	      map->m_algorithmformat == Z_EROFS_COMPRESSION_DEFLATE ||
+	      map->m_algorithmformat == Z_EROFS_COMPRESSION_ZSTD) &&
 	      map->m_llen >= i_blocksize(inode))) {
 		err = z_erofs_get_extent_decompressedlen(&m);
 		if (!err)
@@ -580,7 +570,6 @@ static int z_erofs_fill_inode_lazy(struct inode *inode)
 	int err, headnr;
 	erofs_off_t pos;
 	struct erofs_buf buf = __EROFS_BUF_INITIALIZER;
-	void *kaddr;
 	struct z_erofs_map_header *h;
 
 	if (test_bit(EROFS_I_Z_INITED_BIT, &vi->flags)) {
@@ -600,13 +589,12 @@ static int z_erofs_fill_inode_lazy(struct inode *inode)
 		goto out_unlock;
 
 	pos = ALIGN(erofs_iloc(inode) + vi->inode_isize + vi->xattr_isize, 8);
-	kaddr = erofs_read_metabuf(&buf, sb, erofs_blknr(sb, pos), EROFS_KMAP);
-	if (IS_ERR(kaddr)) {
-		err = PTR_ERR(kaddr);
+	h = erofs_read_metabuf(&buf, sb, pos, EROFS_KMAP);
+	if (IS_ERR(h)) {
+		err = PTR_ERR(h);
 		goto out_unlock;
 	}
 
-	h = kaddr + erofs_blkoff(sb, pos);
 	/*
 	 * if the highest bit of the 8-byte map header is set, the whole file
 	 * is stored in the packed inode. The rest bits keeps z_fragmentoff.
@@ -723,6 +711,8 @@ int z_erofs_map_blocks_iter(struct inode *inode, struct erofs_map_blocks *map,
 
 	err = z_erofs_do_map_blocks(inode, map, flags);
 out:
+	if (err)
+		map->m_llen = 0;
 	trace_z_erofs_map_blocks_iter_exit(inode, map, flags, err);
 	return err;
 }
diff --git a/fs/erofs/utils.c b/fs/erofs/zutil.c
index e146d09151..9b53883e5c 100644
--- a/fs/erofs/utils.c
+++ b/fs/erofs/zutil.c
@@ -5,16 +5,189 @@
  */
 #include "internal.h"
 
-struct page *erofs_allocpage(struct page **pagepool, gfp_t gfp)
+struct z_erofs_gbuf {
+	spinlock_t lock;
+	void *ptr;
+	struct page **pages;
+	unsigned int nrpages;
+};
+
+static struct z_erofs_gbuf *z_erofs_gbufpool, *z_erofs_rsvbuf;
+static unsigned int z_erofs_gbuf_count, z_erofs_gbuf_nrpages,
+		z_erofs_rsv_nrpages;
+
+module_param_named(global_buffers, z_erofs_gbuf_count, uint, 0444);
+module_param_named(reserved_pages, z_erofs_rsv_nrpages, uint, 0444);
+
+static atomic_long_t erofs_global_shrink_cnt;	/* for all mounted instances */
+/* protected by 'erofs_sb_list_lock' */
+static unsigned int shrinker_run_no;
+
+/* protects the mounted 'erofs_sb_list' */
+static DEFINE_SPINLOCK(erofs_sb_list_lock);
+static LIST_HEAD(erofs_sb_list);
+static struct shrinker *erofs_shrinker_info;
+
+static unsigned int z_erofs_gbuf_id(void)
+{
+	return raw_smp_processor_id() % z_erofs_gbuf_count;
+}
+
+void *z_erofs_get_gbuf(unsigned int requiredpages)
+	__acquires(gbuf->lock)
+{
+	struct z_erofs_gbuf *gbuf;
+
+	migrate_disable();
+	gbuf = &z_erofs_gbufpool[z_erofs_gbuf_id()];
+	spin_lock(&gbuf->lock);
+	/* check if the buffer is too small */
+	if (requiredpages > gbuf->nrpages) {
+		spin_unlock(&gbuf->lock);
+		migrate_enable();
+		/* (for sparse checker) pretend gbuf->lock is still taken */
+		__acquire(gbuf->lock);
+		return NULL;
+	}
+	return gbuf->ptr;
+}
+
+void z_erofs_put_gbuf(void *ptr) __releases(gbuf->lock)
+{
+	struct z_erofs_gbuf *gbuf;
+
+	gbuf = &z_erofs_gbufpool[z_erofs_gbuf_id()];
+	DBG_BUGON(gbuf->ptr != ptr);
+	spin_unlock(&gbuf->lock);
+	migrate_enable();
+}
+
+int z_erofs_gbuf_growsize(unsigned int nrpages)
+{
+	static DEFINE_MUTEX(gbuf_resize_mutex);
+	struct page **tmp_pages = NULL;
+	struct z_erofs_gbuf *gbuf;
+	void *ptr, *old_ptr;
+	int last, i, j;
+
+	mutex_lock(&gbuf_resize_mutex);
+	/* avoid shrinking gbufs, since no idea how many fses rely on */
+	if (nrpages <= z_erofs_gbuf_nrpages) {
+		mutex_unlock(&gbuf_resize_mutex);
+		return 0;
+	}
+
+	for (i = 0; i < z_erofs_gbuf_count; ++i) {
+		gbuf = &z_erofs_gbufpool[i];
+		tmp_pages = kcalloc(nrpages, sizeof(*tmp_pages), GFP_KERNEL);
+		if (!tmp_pages)
+			goto out;
+
+		for (j = 0; j < gbuf->nrpages; ++j)
+			tmp_pages[j] = gbuf->pages[j];
+		do {
+			last = j;
+			j = alloc_pages_bulk_array(GFP_KERNEL, nrpages,
+						   tmp_pages);
+			if (last == j)
+				goto out;
+		} while (j != nrpages);
+
+		ptr = vmap(tmp_pages, nrpages, VM_MAP, PAGE_KERNEL);
+		if (!ptr)
+			goto out;
+
+		spin_lock(&gbuf->lock);
+		kfree(gbuf->pages);
+		gbuf->pages = tmp_pages;
+		old_ptr = gbuf->ptr;
+		gbuf->ptr = ptr;
+		gbuf->nrpages = nrpages;
+		spin_unlock(&gbuf->lock);
+		if (old_ptr)
+			vunmap(old_ptr);
+	}
+	z_erofs_gbuf_nrpages = nrpages;
+out:
+	if (i < z_erofs_gbuf_count && tmp_pages) {
+		for (j = 0; j < nrpages; ++j)
+			if (tmp_pages[j] && tmp_pages[j] != gbuf->pages[j])
+				__free_page(tmp_pages[j]);
+		kfree(tmp_pages);
+	}
+	mutex_unlock(&gbuf_resize_mutex);
+	return i < z_erofs_gbuf_count ? -ENOMEM : 0;
+}
+
+int __init z_erofs_gbuf_init(void)
+{
+	unsigned int i, total = num_possible_cpus();
+
+	if (z_erofs_gbuf_count)
+		total = min(z_erofs_gbuf_count, total);
+	z_erofs_gbuf_count = total;
+
+	/* The last (special) global buffer is the reserved buffer */
+	total += !!z_erofs_rsv_nrpages;
+
+	z_erofs_gbufpool = kcalloc(total, sizeof(*z_erofs_gbufpool),
+				   GFP_KERNEL);
+	if (!z_erofs_gbufpool)
+		return -ENOMEM;
+
+	if (z_erofs_rsv_nrpages) {
+		z_erofs_rsvbuf = &z_erofs_gbufpool[total - 1];
+		z_erofs_rsvbuf->pages = kcalloc(z_erofs_rsv_nrpages,
+				sizeof(*z_erofs_rsvbuf->pages), GFP_KERNEL);
+		if (!z_erofs_rsvbuf->pages) {
+			z_erofs_rsvbuf = NULL;
+			z_erofs_rsv_nrpages = 0;
+		}
+	}
+	for (i = 0; i < total; ++i)
+		spin_lock_init(&z_erofs_gbufpool[i].lock);
+	return 0;
+}
+
+void z_erofs_gbuf_exit(void)
+{
+	int i, j;
+
+	for (i = 0; i < z_erofs_gbuf_count + (!!z_erofs_rsvbuf); ++i) {
+		struct z_erofs_gbuf *gbuf = &z_erofs_gbufpool[i];
+
+		if (gbuf->ptr) {
+			vunmap(gbuf->ptr);
+			gbuf->ptr = NULL;
+		}
+
+		if (!gbuf->pages)
+			continue;
+
+		for (j = 0; j < gbuf->nrpages; ++j)
+			if (gbuf->pages[j])
+				put_page(gbuf->pages[j]);
+		kfree(gbuf->pages);
+		gbuf->pages = NULL;
+	}
+	kfree(z_erofs_gbufpool);
+}
+
+struct page *__erofs_allocpage(struct page **pagepool, gfp_t gfp, bool tryrsv)
 {
 	struct page *page = *pagepool;
 
 	if (page) {
-		DBG_BUGON(page_ref_count(page) != 1);
 		*pagepool = (struct page *)page_private(page);
-	} else {
-		page = alloc_page(gfp);
+	} else if (tryrsv && z_erofs_rsvbuf && z_erofs_rsvbuf->nrpages) {
+		spin_lock(&z_erofs_rsvbuf->lock);
+		if (z_erofs_rsvbuf->nrpages)
+			page = z_erofs_rsvbuf->pages[--z_erofs_rsvbuf->nrpages];
+		spin_unlock(&z_erofs_rsvbuf->lock);
 	}
+	if (!page)
+		page = alloc_page(gfp);
+	DBG_BUGON(page && page_ref_count(page) != 1);
 	return page;
 }
 
@@ -24,14 +197,22 @@ void erofs_release_pages(struct page **pagepool)
 		struct page *page = *pagepool;
 
 		*pagepool = (struct page *)page_private(page);
+		/* try to fill reserved global pool first */
+		if (z_erofs_rsvbuf && z_erofs_rsvbuf->nrpages <
+				z_erofs_rsv_nrpages) {
+			spin_lock(&z_erofs_rsvbuf->lock);
+			if (z_erofs_rsvbuf->nrpages < z_erofs_rsv_nrpages) {
+				z_erofs_rsvbuf->pages[z_erofs_rsvbuf->nrpages++]
+						= page;
+				spin_unlock(&z_erofs_rsvbuf->lock);
+				continue;
+			}
+			spin_unlock(&z_erofs_rsvbuf->lock);
+		}
 		put_page(page);
 	}
 }
 
-#ifdef CONFIG_EROFS_FS_ZIP
-/* global shrink count (for all mounted EROFS instances) */
-static atomic_long_t erofs_global_shrink_cnt;
-
 static bool erofs_workgroup_get(struct erofs_workgroup *grp)
 {
 	if (lockref_get_not_zero(&grp->lockref))
@@ -129,7 +310,7 @@ static bool erofs_try_to_release_workgroup(struct erofs_sb_info *sbi,
 	 * the XArray. Otherwise some cached pages could be still attached to
 	 * the orphan old workgroup when the new one is available in the tree.
 	 */
-	if (erofs_try_to_free_all_cached_pages(sbi, grp))
+	if (erofs_try_to_free_all_cached_folios(sbi, grp))
 		goto out;
 
 	/*
@@ -171,13 +352,6 @@ static unsigned long erofs_shrink_workstation(struct erofs_sb_info *sbi,
 	return freed;
 }
 
-/* protected by 'erofs_sb_list_lock' */
-static unsigned int shrinker_run_no;
-
-/* protects the mounted 'erofs_sb_list' */
-static DEFINE_SPINLOCK(erofs_sb_list_lock);
-static LIST_HEAD(erofs_sb_list);
-
 void erofs_shrinker_register(struct super_block *sb)
 {
 	struct erofs_sb_info *sbi = EROFS_SB(sb);
@@ -264,8 +438,6 @@ static unsigned long erofs_shrink_scan(struct shrinker *shrink,
 	return freed;
 }
 
-static struct shrinker *erofs_shrinker_info;
-
 int __init erofs_init_shrinker(void)
 {
 	erofs_shrinker_info = shrinker_alloc(0, "erofs-shrinker");
@@ -274,9 +446,7 @@ int __init erofs_init_shrinker(void)
 
 	erofs_shrinker_info->count_objects = erofs_shrink_count;
 	erofs_shrinker_info->scan_objects = erofs_shrink_scan;
-
 	shrinker_register(erofs_shrinker_info);
-
 	return 0;
 }
 
@@ -284,4 +454,3 @@ void erofs_exit_shrinker(void)
 {
 	shrinker_free(erofs_shrinker_info);
 }
-#endif	/* !CONFIG_EROFS_FS_ZIP */
diff --git a/fs/eventfd.c b/fs/eventfd.c
index ad8186d47b..9afdb722fa 100644
--- a/fs/eventfd.c
+++ b/fs/eventfd.c
@@ -251,7 +251,7 @@ static ssize_t eventfd_write(struct file *file, const char __user *buf, size_t c
 	ssize_t res;
 	__u64 ucnt;
 
-	if (count < sizeof(ucnt))
+	if (count != sizeof(ucnt))
 		return -EINVAL;
 	if (copy_from_user(&ucnt, buf, sizeof(ucnt)))
 		return -EFAULT;
@@ -283,13 +283,18 @@ static ssize_t eventfd_write(struct file *file, const char __user *buf, size_t c
 static void eventfd_show_fdinfo(struct seq_file *m, struct file *f)
 {
 	struct eventfd_ctx *ctx = f->private_data;
+	__u64 cnt;
 
 	spin_lock_irq(&ctx->wqh.lock);
-	seq_printf(m, "eventfd-count: %16llx\n",
-		   (unsigned long long)ctx->count);
+	cnt = ctx->count;
 	spin_unlock_irq(&ctx->wqh.lock);
-	seq_printf(m, "eventfd-id: %d\n", ctx->id);
-	seq_printf(m, "eventfd-semaphore: %d\n",
+
+	seq_printf(m,
+		   "eventfd-count: %16llx\n"
+		   "eventfd-id: %d\n"
+		   "eventfd-semaphore: %d\n",
+		   cnt,
+		   ctx->id,
 		   !!(ctx->flags & EFD_SEMAPHORE));
 }
 #endif
@@ -383,6 +388,7 @@ static int do_eventfd(unsigned int count, int flags)
 	/* Check the EFD_* constants for consistency.  */
 	BUILD_BUG_ON(EFD_CLOEXEC != O_CLOEXEC);
 	BUILD_BUG_ON(EFD_NONBLOCK != O_NONBLOCK);
+	BUILD_BUG_ON(EFD_SEMAPHORE != (1 << 0));
 
 	if (flags & ~EFD_FLAGS_SET)
 		return -EINVAL;
diff --git a/fs/eventpoll.c b/fs/eventpoll.c
index c5a9a483fb..f53ca4f7fc 100644
--- a/fs/eventpoll.c
+++ b/fs/eventpoll.c
@@ -37,6 +37,7 @@
 #include <linux/seq_file.h>
 #include <linux/compat.h>
 #include <linux/rculist.h>
+#include <linux/capability.h>
 #include <net/busy_poll.h>
 
 /*
@@ -206,7 +207,7 @@ struct eventpoll {
 	 */
 	struct epitem *ovflist;
 
-	/* wakeup_source used when ep_scan_ready_list is running */
+	/* wakeup_source used when ep_send_events or __ep_eventpoll_poll is running */
 	struct wakeup_source *ws;
 
 	/* The user that created the eventpoll descriptor */
@@ -227,6 +228,11 @@ struct eventpoll {
 #ifdef CONFIG_NET_RX_BUSY_POLL
 	/* used to track busy poll napi_id */
 	unsigned int napi_id;
+	/* busy poll timeout */
+	u32 busy_poll_usecs;
+	/* busy poll packet budget */
+	u16 busy_poll_budget;
+	bool prefer_busy_poll;
 #endif
 
 #ifdef CONFIG_DEBUG_LOCK_ALLOC
@@ -387,11 +393,41 @@ static inline int ep_events_available(struct eventpoll *ep)
 }
 
 #ifdef CONFIG_NET_RX_BUSY_POLL
+/**
+ * busy_loop_ep_timeout - check if busy poll has timed out. The timeout value
+ * from the epoll instance ep is preferred, but if it is not set fallback to
+ * the system-wide global via busy_loop_timeout.
+ *
+ * @start_time: The start time used to compute the remaining time until timeout.
+ * @ep: Pointer to the eventpoll context.
+ *
+ * Return: true if the timeout has expired, false otherwise.
+ */
+static bool busy_loop_ep_timeout(unsigned long start_time,
+				 struct eventpoll *ep)
+{
+	unsigned long bp_usec = READ_ONCE(ep->busy_poll_usecs);
+
+	if (bp_usec) {
+		unsigned long end_time = start_time + bp_usec;
+		unsigned long now = busy_loop_current_time();
+
+		return time_after(now, end_time);
+	} else {
+		return busy_loop_timeout(start_time);
+	}
+}
+
+static bool ep_busy_loop_on(struct eventpoll *ep)
+{
+	return !!ep->busy_poll_usecs || net_busy_loop_on();
+}
+
 static bool ep_busy_loop_end(void *p, unsigned long start_time)
 {
 	struct eventpoll *ep = p;
 
-	return ep_events_available(ep) || busy_loop_timeout(start_time);
+	return ep_events_available(ep) || busy_loop_ep_timeout(start_time, ep);
 }
 
 /*
@@ -403,10 +439,15 @@ static bool ep_busy_loop_end(void *p, unsigned long start_time)
 static bool ep_busy_loop(struct eventpoll *ep, int nonblock)
 {
 	unsigned int napi_id = READ_ONCE(ep->napi_id);
+	u16 budget = READ_ONCE(ep->busy_poll_budget);
+	bool prefer_busy_poll = READ_ONCE(ep->prefer_busy_poll);
 
-	if ((napi_id >= MIN_NAPI_ID) && net_busy_loop_on()) {
-		napi_busy_loop(napi_id, nonblock ? NULL : ep_busy_loop_end, ep, false,
-			       BUSY_POLL_BUDGET);
+	if (!budget)
+		budget = BUSY_POLL_BUDGET;
+
+	if (napi_id >= MIN_NAPI_ID && ep_busy_loop_on(ep)) {
+		napi_busy_loop(napi_id, nonblock ? NULL : ep_busy_loop_end,
+			       ep, prefer_busy_poll, budget);
 		if (ep_events_available(ep))
 			return true;
 		/*
@@ -425,12 +466,12 @@ static bool ep_busy_loop(struct eventpoll *ep, int nonblock)
  */
 static inline void ep_set_busy_poll_napi_id(struct epitem *epi)
 {
-	struct eventpoll *ep;
+	struct eventpoll *ep = epi->ep;
 	unsigned int napi_id;
 	struct socket *sock;
 	struct sock *sk;
 
-	if (!net_busy_loop_on())
+	if (!ep_busy_loop_on(ep))
 		return;
 
 	sock = sock_from_file(epi->ffd.file);
@@ -442,7 +483,6 @@ static inline void ep_set_busy_poll_napi_id(struct epitem *epi)
 		return;
 
 	napi_id = READ_ONCE(sk->sk_napi_id);
-	ep = epi->ep;
 
 	/* Non-NAPI IDs can be rejected
 	 *	or
@@ -455,6 +495,49 @@ static inline void ep_set_busy_poll_napi_id(struct epitem *epi)
 	ep->napi_id = napi_id;
 }
 
+static long ep_eventpoll_bp_ioctl(struct file *file, unsigned int cmd,
+				  unsigned long arg)
+{
+	struct eventpoll *ep = file->private_data;
+	void __user *uarg = (void __user *)arg;
+	struct epoll_params epoll_params;
+
+	switch (cmd) {
+	case EPIOCSPARAMS:
+		if (copy_from_user(&epoll_params, uarg, sizeof(epoll_params)))
+			return -EFAULT;
+
+		/* pad byte must be zero */
+		if (epoll_params.__pad)
+			return -EINVAL;
+
+		if (epoll_params.busy_poll_usecs > S32_MAX)
+			return -EINVAL;
+
+		if (epoll_params.prefer_busy_poll > 1)
+			return -EINVAL;
+
+		if (epoll_params.busy_poll_budget > NAPI_POLL_WEIGHT &&
+		    !capable(CAP_NET_ADMIN))
+			return -EPERM;
+
+		WRITE_ONCE(ep->busy_poll_usecs, epoll_params.busy_poll_usecs);
+		WRITE_ONCE(ep->busy_poll_budget, epoll_params.busy_poll_budget);
+		WRITE_ONCE(ep->prefer_busy_poll, epoll_params.prefer_busy_poll);
+		return 0;
+	case EPIOCGPARAMS:
+		memset(&epoll_params, 0, sizeof(epoll_params));
+		epoll_params.busy_poll_usecs = READ_ONCE(ep->busy_poll_usecs);
+		epoll_params.busy_poll_budget = READ_ONCE(ep->busy_poll_budget);
+		epoll_params.prefer_busy_poll = READ_ONCE(ep->prefer_busy_poll);
+		if (copy_to_user(uarg, &epoll_params, sizeof(epoll_params)))
+			return -EFAULT;
+		return 0;
+	default:
+		return -ENOIOCTLCMD;
+	}
+}
+
 #else
 
 static inline bool ep_busy_loop(struct eventpoll *ep, int nonblock)
@@ -466,6 +549,12 @@ static inline void ep_set_busy_poll_napi_id(struct epitem *epi)
 {
 }
 
+static long ep_eventpoll_bp_ioctl(struct file *file, unsigned int cmd,
+				  unsigned long arg)
+{
+	return -EOPNOTSUPP;
+}
+
 #endif /* CONFIG_NET_RX_BUSY_POLL */
 
 /*
@@ -678,12 +767,6 @@ static void ep_done_scan(struct eventpoll *ep,
 	write_unlock_irq(&ep->lock);
 }
 
-static void epi_rcu_free(struct rcu_head *head)
-{
-	struct epitem *epi = container_of(head, struct epitem, rcu);
-	kmem_cache_free(epi_cache, epi);
-}
-
 static void ep_get(struct eventpoll *ep)
 {
 	refcount_inc(&ep->refcount);
@@ -767,7 +850,7 @@ static bool __ep_remove(struct eventpoll *ep, struct epitem *epi, bool force)
 	 * ep->mtx. The rcu read side, reverse_path_check_proc(), does not make
 	 * use of the rbn field.
 	 */
-	call_rcu(&epi->rcu, epi_rcu_free);
+	kfree_rcu(epi, rcu);
 
 	percpu_counter_dec(&ep->user->epoll_watches);
 	return ep_refcount_dec_and_test(ep);
@@ -825,6 +908,27 @@ static void ep_clear_and_put(struct eventpoll *ep)
 		ep_free(ep);
 }
 
+static long ep_eventpoll_ioctl(struct file *file, unsigned int cmd,
+			       unsigned long arg)
+{
+	int ret;
+
+	if (!is_file_epoll(file))
+		return -EINVAL;
+
+	switch (cmd) {
+	case EPIOCSPARAMS:
+	case EPIOCGPARAMS:
+		ret = ep_eventpoll_bp_ioctl(file, cmd, arg);
+		break;
+	default:
+		ret = -EINVAL;
+		break;
+	}
+
+	return ret;
+}
+
 static int ep_eventpoll_release(struct inode *inode, struct file *file)
 {
 	struct eventpoll *ep = file->private_data;
@@ -967,6 +1071,8 @@ static const struct file_operations eventpoll_fops = {
 	.release	= ep_eventpoll_release,
 	.poll		= ep_eventpoll_poll,
 	.llseek		= noop_llseek,
+	.unlocked_ioctl	= ep_eventpoll_ioctl,
+	.compat_ioctl   = compat_ptr_ioctl,
 };
 
 /*
@@ -1189,7 +1295,7 @@ static inline bool chain_epi_lockless(struct epitem *epi)
  * This callback takes a read lock in order not to contend with concurrent
  * events from another file descriptor, thus all modifications to ->rdllist
  * or ->ovflist are lockless.  Read lock is paired with the write lock from
- * ep_scan_ready_list(), which stops all list modifications and guarantees
+ * ep_start/done_scan(), which stops all list modifications and guarantees
  * that lists state is seen correctly.
  *
  * Another thing worth to mention is that ep_poll_callback() can be called
@@ -1787,7 +1893,7 @@ static int ep_send_events(struct eventpoll *ep,
 			 * availability. At this point, no one can insert
 			 * into ep->rdllist besides us. The epoll_ctl()
 			 * callers are locked out by
-			 * ep_scan_ready_list() holding "mtx" and the
+			 * ep_send_events() holding "mtx" and the
 			 * poll callback will queue them in ep->ovflist.
 			 */
 			list_add_tail(&epi->rdllink, &ep->rdllist);
@@ -1940,7 +2046,7 @@ static int ep_poll(struct eventpoll *ep, struct epoll_event __user *events,
 		__set_current_state(TASK_INTERRUPTIBLE);
 
 		/*
-		 * Do the final check under the lock. ep_scan_ready_list()
+		 * Do the final check under the lock. ep_start/done_scan()
 		 * plays with two lists (->rdllist and ->ovflist) and there
 		 * is always a race when both lists are empty for short
 		 * period of time although events are pending, so lock is
@@ -2094,6 +2200,11 @@ static int do_epoll_create(int flags)
 		error = PTR_ERR(file);
 		goto out_free_fd;
 	}
+#ifdef CONFIG_NET_RX_BUSY_POLL
+	ep->busy_poll_usecs = 0;
+	ep->busy_poll_budget = 0;
+	ep->prefer_busy_poll = false;
+#endif
 	ep->file = file;
 	fd_install(fd, file);
 	return fd;
diff --git a/fs/exec.c b/fs/exec.c
index f11cfd7bce..0c17e59e37 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -1170,7 +1170,6 @@ static int de_thread(struct task_struct *tsk)
 
 		BUG_ON(leader->exit_state != EXIT_ZOMBIE);
 		leader->exit_state = EXIT_DEAD;
-
 		/*
 		 * We are going to release_task()->ptrace_unlink() silently,
 		 * the tracer can sleep in do_wait(). EXIT_DEAD guarantees
@@ -1280,6 +1279,14 @@ int begin_new_exec(struct linux_binprm * bprm)
 		return retval;
 
 	/*
+	 * This tracepoint marks the point before flushing the old exec where
+	 * the current task is still unchanged, but errors are fatal (point of
+	 * no return). The later "sched_process_exec" tracepoint is called after
+	 * the current task has successfully switched to the new exec.
+	 */
+	trace_sched_prepare_exec(current, bprm);
+
+	/*
 	 * Ensure all future errors are fatal.
 	 */
 	bprm->point_of_no_return = true;
@@ -1661,6 +1668,7 @@ static void bprm_fill_uid(struct linux_binprm *bprm, struct file *file)
 	unsigned int mode;
 	vfsuid_t vfsuid;
 	vfsgid_t vfsgid;
+	int err;
 
 	if (!mnt_may_suid(file->f_path.mnt))
 		return;
@@ -1677,12 +1685,17 @@ static void bprm_fill_uid(struct linux_binprm *bprm, struct file *file)
 	/* Be careful if suid/sgid is set */
 	inode_lock(inode);
 
-	/* reload atomically mode/uid/gid now that lock held */
+	/* Atomically reload and check mode/uid/gid now that lock held. */
 	mode = inode->i_mode;
 	vfsuid = i_uid_into_vfsuid(idmap, inode);
 	vfsgid = i_gid_into_vfsgid(idmap, inode);
+	err = inode_permission(idmap, inode, MAY_EXEC);
 	inode_unlock(inode);
 
+	/* Did the exec bit vanish out from under us? Give up. */
+	if (err)
+		return;
+
 	/* We ignore suid/sgid if there are no mappings for them in the ns */
 	if (!vfsuid_has_mapping(bprm->cred->user_ns, vfsuid) ||
 	    !vfsgid_has_mapping(bprm->cred->user_ns, vfsgid))
@@ -1732,7 +1745,6 @@ static int prepare_binprm(struct linux_binprm *bprm)
  */
 int remove_arg_zero(struct linux_binprm *bprm)
 {
-	int ret = 0;
 	unsigned long offset;
 	char *kaddr;
 	struct page *page;
@@ -1743,10 +1755,8 @@ int remove_arg_zero(struct linux_binprm *bprm)
 	do {
 		offset = bprm->p & ~PAGE_MASK;
 		page = get_arg_page(bprm, bprm->p, 0);
-		if (!page) {
-			ret = -EFAULT;
-			goto out;
-		}
+		if (!page)
+			return -EFAULT;
 		kaddr = kmap_local_page(page);
 
 		for (; offset < PAGE_SIZE && kaddr[offset];
@@ -1759,10 +1769,8 @@ int remove_arg_zero(struct linux_binprm *bprm)
 
 	bprm->p++;
 	bprm->argc--;
-	ret = 0;
 
-out:
-	return ret;
+	return 0;
 }
 EXPORT_SYMBOL(remove_arg_zero);
 
diff --git a/fs/exfat/cache.c b/fs/exfat/cache.c
index 5a2f119b7e..7cc200d898 100644
--- a/fs/exfat/cache.c
+++ b/fs/exfat/cache.c
@@ -46,7 +46,7 @@ int exfat_cache_init(void)
 {
 	exfat_cachep = kmem_cache_create("exfat_cache",
 				sizeof(struct exfat_cache),
-				0, SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD,
+				0, SLAB_RECLAIM_ACCOUNT,
 				exfat_cache_init_once);
 	if (!exfat_cachep)
 		return -ENOMEM;
diff --git a/fs/exfat/dir.c b/fs/exfat/dir.c
index 9f9295847a..7446bf09a0 100644
--- a/fs/exfat/dir.c
+++ b/fs/exfat/dir.c
@@ -420,6 +420,7 @@ static void exfat_set_entry_type(struct exfat_dentry *ep, unsigned int type)
 static void exfat_init_stream_entry(struct exfat_dentry *ep,
 		unsigned int start_clu, unsigned long long size)
 {
+	memset(ep, 0, sizeof(*ep));
 	exfat_set_entry_type(ep, TYPE_STREAM);
 	if (size == 0)
 		ep->dentry.stream.flags = ALLOC_FAT_CHAIN;
@@ -448,88 +449,35 @@ static void exfat_init_name_entry(struct exfat_dentry *ep,
 	}
 }
 
-int exfat_init_dir_entry(struct inode *inode, struct exfat_chain *p_dir,
-		int entry, unsigned int type, unsigned int start_clu,
-		unsigned long long size)
+void exfat_init_dir_entry(struct exfat_entry_set_cache *es,
+		unsigned int type, unsigned int start_clu,
+		unsigned long long size, struct timespec64 *ts)
 {
-	struct super_block *sb = inode->i_sb;
+	struct super_block *sb = es->sb;
 	struct exfat_sb_info *sbi = EXFAT_SB(sb);
-	struct timespec64 ts = current_time(inode);
 	struct exfat_dentry *ep;
-	struct buffer_head *bh;
-
-	/*
-	 * We cannot use exfat_get_dentry_set here because file ep is not
-	 * initialized yet.
-	 */
-	ep = exfat_get_dentry(sb, p_dir, entry, &bh);
-	if (!ep)
-		return -EIO;
 
+	ep = exfat_get_dentry_cached(es, ES_IDX_FILE);
+	memset(ep, 0, sizeof(*ep));
 	exfat_set_entry_type(ep, type);
-	exfat_set_entry_time(sbi, &ts,
+	exfat_set_entry_time(sbi, ts,
 			&ep->dentry.file.create_tz,
 			&ep->dentry.file.create_time,
 			&ep->dentry.file.create_date,
 			&ep->dentry.file.create_time_cs);
-	exfat_set_entry_time(sbi, &ts,
+	exfat_set_entry_time(sbi, ts,
 			&ep->dentry.file.modify_tz,
 			&ep->dentry.file.modify_time,
 			&ep->dentry.file.modify_date,
 			&ep->dentry.file.modify_time_cs);
-	exfat_set_entry_time(sbi, &ts,
+	exfat_set_entry_time(sbi, ts,
 			&ep->dentry.file.access_tz,
 			&ep->dentry.file.access_time,
 			&ep->dentry.file.access_date,
 			NULL);
 
-	exfat_update_bh(bh, IS_DIRSYNC(inode));
-	brelse(bh);
-
-	ep = exfat_get_dentry(sb, p_dir, entry + 1, &bh);
-	if (!ep)
-		return -EIO;
-
+	ep = exfat_get_dentry_cached(es, ES_IDX_STREAM);
 	exfat_init_stream_entry(ep, start_clu, size);
-	exfat_update_bh(bh, IS_DIRSYNC(inode));
-	brelse(bh);
-
-	return 0;
-}
-
-int exfat_update_dir_chksum(struct inode *inode, struct exfat_chain *p_dir,
-		int entry)
-{
-	struct super_block *sb = inode->i_sb;
-	int ret = 0;
-	int i, num_entries;
-	u16 chksum;
-	struct exfat_dentry *ep, *fep;
-	struct buffer_head *fbh, *bh;
-
-	fep = exfat_get_dentry(sb, p_dir, entry, &fbh);
-	if (!fep)
-		return -EIO;
-
-	num_entries = fep->dentry.file.num_ext + 1;
-	chksum = exfat_calc_chksum16(fep, DENTRY_SIZE, 0, CS_DIR_ENTRY);
-
-	for (i = 1; i < num_entries; i++) {
-		ep = exfat_get_dentry(sb, p_dir, entry + i, &bh);
-		if (!ep) {
-			ret = -EIO;
-			goto release_fbh;
-		}
-		chksum = exfat_calc_chksum16(ep, DENTRY_SIZE, chksum,
-				CS_DEFAULT);
-		brelse(bh);
-	}
-
-	fep->dentry.file.checksum = cpu_to_le16(chksum);
-	exfat_update_bh(fbh, IS_DIRSYNC(inode));
-release_fbh:
-	brelse(fbh);
-	return ret;
 }
 
 static void exfat_free_benign_secondary_clusters(struct inode *inode,
@@ -551,76 +499,49 @@ static void exfat_free_benign_secondary_clusters(struct inode *inode,
 	exfat_free_cluster(inode, &dir);
 }
 
-int exfat_init_ext_entry(struct inode *inode, struct exfat_chain *p_dir,
-		int entry, int num_entries, struct exfat_uni_name *p_uniname)
+void exfat_init_ext_entry(struct exfat_entry_set_cache *es, int num_entries,
+		struct exfat_uni_name *p_uniname)
 {
-	struct super_block *sb = inode->i_sb;
 	int i;
 	unsigned short *uniname = p_uniname->name;
 	struct exfat_dentry *ep;
-	struct buffer_head *bh;
-	int sync = IS_DIRSYNC(inode);
-
-	ep = exfat_get_dentry(sb, p_dir, entry, &bh);
-	if (!ep)
-		return -EIO;
 
+	ep = exfat_get_dentry_cached(es, ES_IDX_FILE);
 	ep->dentry.file.num_ext = (unsigned char)(num_entries - 1);
-	exfat_update_bh(bh, sync);
-	brelse(bh);
-
-	ep = exfat_get_dentry(sb, p_dir, entry + 1, &bh);
-	if (!ep)
-		return -EIO;
 
+	ep = exfat_get_dentry_cached(es, ES_IDX_STREAM);
 	ep->dentry.stream.name_len = p_uniname->name_len;
 	ep->dentry.stream.name_hash = cpu_to_le16(p_uniname->name_hash);
-	exfat_update_bh(bh, sync);
-	brelse(bh);
-
-	for (i = EXFAT_FIRST_CLUSTER; i < num_entries; i++) {
-		ep = exfat_get_dentry(sb, p_dir, entry + i, &bh);
-		if (!ep)
-			return -EIO;
-
-		if (exfat_get_entry_type(ep) & TYPE_BENIGN_SEC)
-			exfat_free_benign_secondary_clusters(inode, ep);
 
+	for (i = ES_IDX_FIRST_FILENAME; i < num_entries; i++) {
+		ep = exfat_get_dentry_cached(es, i);
 		exfat_init_name_entry(ep, uniname);
-		exfat_update_bh(bh, sync);
-		brelse(bh);
 		uniname += EXFAT_FILE_NAME_LEN;
 	}
 
-	exfat_update_dir_chksum(inode, p_dir, entry);
-	return 0;
+	exfat_update_dir_chksum(es);
 }
 
-int exfat_remove_entries(struct inode *inode, struct exfat_chain *p_dir,
-		int entry, int order, int num_entries)
+void exfat_remove_entries(struct inode *inode, struct exfat_entry_set_cache *es,
+		int order)
 {
-	struct super_block *sb = inode->i_sb;
 	int i;
 	struct exfat_dentry *ep;
-	struct buffer_head *bh;
 
-	for (i = order; i < num_entries; i++) {
-		ep = exfat_get_dentry(sb, p_dir, entry + i, &bh);
-		if (!ep)
-			return -EIO;
+	for (i = order; i < es->num_entries; i++) {
+		ep = exfat_get_dentry_cached(es, i);
 
 		if (exfat_get_entry_type(ep) & TYPE_BENIGN_SEC)
 			exfat_free_benign_secondary_clusters(inode, ep);
 
 		exfat_set_entry_type(ep, TYPE_DELETED);
-		exfat_update_bh(bh, IS_DIRSYNC(inode));
-		brelse(bh);
 	}
 
-	return 0;
+	if (order < es->num_entries)
+		es->modified = true;
 }
 
-void exfat_update_dir_chksum_with_entry_set(struct exfat_entry_set_cache *es)
+void exfat_update_dir_chksum(struct exfat_entry_set_cache *es)
 {
 	int chksum_type = CS_DIR_ENTRY, i;
 	unsigned short chksum = 0;
@@ -775,7 +696,6 @@ struct exfat_dentry *exfat_get_dentry(struct super_block *sb,
 }
 
 enum exfat_validate_dentry_mode {
-	ES_MODE_STARTED,
 	ES_MODE_GET_FILE_ENTRY,
 	ES_MODE_GET_STRM_ENTRY,
 	ES_MODE_GET_NAME_ENTRY,
@@ -790,11 +710,6 @@ static bool exfat_validate_entry(unsigned int type,
 		return false;
 
 	switch (*mode) {
-	case ES_MODE_STARTED:
-		if  (type != TYPE_FILE && type != TYPE_DIR)
-			return false;
-		*mode = ES_MODE_GET_FILE_ENTRY;
-		break;
 	case ES_MODE_GET_FILE_ENTRY:
 		if (type != TYPE_STREAM)
 			return false;
@@ -834,7 +749,7 @@ struct exfat_dentry *exfat_get_dentry_cached(
 }
 
 /*
- * Returns a set of dentries for a file or dir.
+ * Returns a set of dentries.
  *
  * Note It provides a direct pointer to bh->data via exfat_get_dentry_cached().
  * User should call exfat_get_dentry_set() after setting 'modified' to apply
@@ -842,22 +757,24 @@ struct exfat_dentry *exfat_get_dentry_cached(
  *
  * in:
  *   sb+p_dir+entry: indicates a file/dir
- *   type:  specifies how many dentries should be included.
+ *   num_entries: specifies how many dentries should be included.
+ *                It will be set to es->num_entries if it is not 0.
+ *                If num_entries is 0, es->num_entries will be obtained
+ *                from the first dentry.
+ * out:
+ *   es: pointer of entry set on success.
  * return:
- *   pointer of entry set on success,
- *   NULL on failure.
+ *   0 on success
+ *   -error code on failure
  */
-int exfat_get_dentry_set(struct exfat_entry_set_cache *es,
+static int __exfat_get_dentry_set(struct exfat_entry_set_cache *es,
 		struct super_block *sb, struct exfat_chain *p_dir, int entry,
-		unsigned int type)
+		unsigned int num_entries)
 {
 	int ret, i, num_bh;
 	unsigned int off;
 	sector_t sec;
 	struct exfat_sb_info *sbi = EXFAT_SB(sb);
-	struct exfat_dentry *ep;
-	int num_entries;
-	enum exfat_validate_dentry_mode mode = ES_MODE_STARTED;
 	struct buffer_head *bh;
 
 	if (p_dir->dir == DIR_DELETED) {
@@ -880,17 +797,23 @@ int exfat_get_dentry_set(struct exfat_entry_set_cache *es,
 		return -EIO;
 	es->bh[es->num_bh++] = bh;
 
-	ep = exfat_get_dentry_cached(es, ES_IDX_FILE);
-	if (!exfat_validate_entry(exfat_get_entry_type(ep), &mode))
-		goto put_es;
+	if (num_entries == ES_ALL_ENTRIES) {
+		struct exfat_dentry *ep;
+
+		ep = exfat_get_dentry_cached(es, ES_IDX_FILE);
+		if (ep->type != EXFAT_FILE) {
+			brelse(bh);
+			return -EIO;
+		}
+
+		num_entries = ep->dentry.file.num_ext + 1;
+	}
 
-	num_entries = type == ES_ALL_ENTRIES ?
-		ep->dentry.file.num_ext + 1 : type;
 	es->num_entries = num_entries;
 
 	num_bh = EXFAT_B_TO_BLK_ROUND_UP(off + num_entries * DENTRY_SIZE, sb);
 	if (num_bh > ARRAY_SIZE(es->__bh)) {
-		es->bh = kmalloc_array(num_bh, sizeof(*es->bh), GFP_KERNEL);
+		es->bh = kmalloc_array(num_bh, sizeof(*es->bh), GFP_NOFS);
 		if (!es->bh) {
 			brelse(bh);
 			return -ENOMEM;
@@ -918,8 +841,27 @@ int exfat_get_dentry_set(struct exfat_entry_set_cache *es,
 		es->bh[es->num_bh++] = bh;
 	}
 
+	return 0;
+
+put_es:
+	exfat_put_dentry_set(es, false);
+	return -EIO;
+}
+
+int exfat_get_dentry_set(struct exfat_entry_set_cache *es,
+		struct super_block *sb, struct exfat_chain *p_dir,
+		int entry, unsigned int num_entries)
+{
+	int ret, i;
+	struct exfat_dentry *ep;
+	enum exfat_validate_dentry_mode mode = ES_MODE_GET_FILE_ENTRY;
+
+	ret = __exfat_get_dentry_set(es, sb, p_dir, entry, num_entries);
+	if (ret < 0)
+		return ret;
+
 	/* validate cached dentries */
-	for (i = ES_IDX_STREAM; i < num_entries; i++) {
+	for (i = ES_IDX_STREAM; i < es->num_entries; i++) {
 		ep = exfat_get_dentry_cached(es, i);
 		if (!exfat_validate_entry(exfat_get_entry_type(ep), &mode))
 			goto put_es;
@@ -931,6 +873,85 @@ put_es:
 	return -EIO;
 }
 
+static int exfat_validate_empty_dentry_set(struct exfat_entry_set_cache *es)
+{
+	struct exfat_dentry *ep;
+	struct buffer_head *bh;
+	int i, off;
+	bool unused_hit = false;
+
+	/*
+	 * ONLY UNUSED OR DELETED DENTRIES ARE ALLOWED:
+	 * Although it violates the specification for a deleted entry to
+	 * follow an unused entry, some exFAT implementations could work
+	 * like this. Therefore, to improve compatibility, let's allow it.
+	 */
+	for (i = 0; i < es->num_entries; i++) {
+		ep = exfat_get_dentry_cached(es, i);
+		if (ep->type == EXFAT_UNUSED) {
+			unused_hit = true;
+		} else if (!IS_EXFAT_DELETED(ep->type)) {
+			if (unused_hit)
+				goto err_used_follow_unused;
+			i++;
+			goto count_skip_entries;
+		}
+	}
+
+	return 0;
+
+err_used_follow_unused:
+	off = es->start_off + (i << DENTRY_SIZE_BITS);
+	bh = es->bh[EXFAT_B_TO_BLK(off, es->sb)];
+
+	exfat_fs_error(es->sb,
+		"in sector %lld, dentry %d should be unused, but 0x%x",
+		bh->b_blocknr, off >> DENTRY_SIZE_BITS, ep->type);
+
+	return -EIO;
+
+count_skip_entries:
+	es->num_entries = EXFAT_B_TO_DEN(EXFAT_BLK_TO_B(es->num_bh, es->sb) - es->start_off);
+	for (; i < es->num_entries; i++) {
+		ep = exfat_get_dentry_cached(es, i);
+		if (IS_EXFAT_DELETED(ep->type))
+			break;
+	}
+
+	return i;
+}
+
+/*
+ * Get an empty dentry set.
+ *
+ * in:
+ *   sb+p_dir+entry: indicates the empty dentry location
+ *   num_entries: specifies how many empty dentries should be included.
+ * out:
+ *   es: pointer of empty dentry set on success.
+ * return:
+ *   0  : on success
+ *   >0 : the dentries are not empty, the return value is the number of
+ *        dentries to be skipped for the next lookup.
+ *   <0 : on failure
+ */
+int exfat_get_empty_dentry_set(struct exfat_entry_set_cache *es,
+		struct super_block *sb, struct exfat_chain *p_dir,
+		int entry, unsigned int num_entries)
+{
+	int ret;
+
+	ret = __exfat_get_dentry_set(es, sb, p_dir, entry, num_entries);
+	if (ret < 0)
+		return ret;
+
+	ret = exfat_validate_empty_dentry_set(es);
+	if (ret)
+		exfat_put_dentry_set(es, false);
+
+	return ret;
+}
+
 static inline void exfat_reset_empty_hint(struct exfat_hint_femp *hint_femp)
 {
 	hint_femp->eidx = EXFAT_HINT_NONE;
@@ -1187,27 +1208,6 @@ found:
 	return dentry - num_ext;
 }
 
-int exfat_count_ext_entries(struct super_block *sb, struct exfat_chain *p_dir,
-		int entry, struct exfat_dentry *ep)
-{
-	int i, count = 0;
-	unsigned int type;
-	struct exfat_dentry *ext_ep;
-	struct buffer_head *bh;
-
-	for (i = 0, entry++; i < ep->dentry.file.num_ext; i++, entry++) {
-		ext_ep = exfat_get_dentry(sb, p_dir, entry, &bh);
-		if (!ext_ep)
-			return -EIO;
-
-		type = exfat_get_entry_type(ext_ep);
-		brelse(bh);
-		if (type & TYPE_CRITICAL_SEC || type & TYPE_BENIGN_SEC)
-			count++;
-	}
-	return count;
-}
-
 int exfat_count_dir_entries(struct super_block *sb, struct exfat_chain *p_dir)
 {
 	int i, count = 0;
diff --git a/fs/exfat/exfat_fs.h b/fs/exfat/exfat_fs.h
index 3615954334..ecc5db952d 100644
--- a/fs/exfat/exfat_fs.h
+++ b/fs/exfat/exfat_fs.h
@@ -431,8 +431,6 @@ int exfat_ent_get(struct super_block *sb, unsigned int loc,
 		unsigned int *content);
 int exfat_ent_set(struct super_block *sb, unsigned int loc,
 		unsigned int content);
-int exfat_count_ext_entries(struct super_block *sb, struct exfat_chain *p_dir,
-		int entry, struct exfat_dentry *p_entry);
 int exfat_chain_cont_cluster(struct super_block *sb, unsigned int chain,
 		unsigned int len);
 int exfat_zeroed_cluster(struct inode *dir, unsigned int clu);
@@ -480,16 +478,14 @@ int exfat_get_cluster(struct inode *inode, unsigned int cluster,
 extern const struct inode_operations exfat_dir_inode_operations;
 extern const struct file_operations exfat_dir_operations;
 unsigned int exfat_get_entry_type(struct exfat_dentry *p_entry);
-int exfat_init_dir_entry(struct inode *inode, struct exfat_chain *p_dir,
-		int entry, unsigned int type, unsigned int start_clu,
-		unsigned long long size);
-int exfat_init_ext_entry(struct inode *inode, struct exfat_chain *p_dir,
-		int entry, int num_entries, struct exfat_uni_name *p_uniname);
-int exfat_remove_entries(struct inode *inode, struct exfat_chain *p_dir,
-		int entry, int order, int num_entries);
-int exfat_update_dir_chksum(struct inode *inode, struct exfat_chain *p_dir,
-		int entry);
-void exfat_update_dir_chksum_with_entry_set(struct exfat_entry_set_cache *es);
+void exfat_init_dir_entry(struct exfat_entry_set_cache *es,
+		unsigned int type, unsigned int start_clu,
+		unsigned long long size, struct timespec64 *ts);
+void exfat_init_ext_entry(struct exfat_entry_set_cache *es, int num_entries,
+		struct exfat_uni_name *p_uniname);
+void exfat_remove_entries(struct inode *inode, struct exfat_entry_set_cache *es,
+		int order);
+void exfat_update_dir_chksum(struct exfat_entry_set_cache *es);
 int exfat_calc_num_entries(struct exfat_uni_name *p_uniname);
 int exfat_find_dir_entry(struct super_block *sb, struct exfat_inode_info *ei,
 		struct exfat_chain *p_dir, struct exfat_uni_name *p_uniname,
@@ -501,7 +497,10 @@ struct exfat_dentry *exfat_get_dentry_cached(struct exfat_entry_set_cache *es,
 		int num);
 int exfat_get_dentry_set(struct exfat_entry_set_cache *es,
 		struct super_block *sb, struct exfat_chain *p_dir, int entry,
-		unsigned int type);
+		unsigned int num_entries);
+int exfat_get_empty_dentry_set(struct exfat_entry_set_cache *es,
+		struct super_block *sb, struct exfat_chain *p_dir, int entry,
+		unsigned int num_entries);
 int exfat_put_dentry_set(struct exfat_entry_set_cache *es, int sync);
 int exfat_count_dir_entries(struct super_block *sb, struct exfat_chain *p_dir);
 
diff --git a/fs/exfat/inode.c b/fs/exfat/inode.c
index 0687f95295..dd894e558c 100644
--- a/fs/exfat/inode.c
+++ b/fs/exfat/inode.c
@@ -94,7 +94,7 @@ int __exfat_write_inode(struct inode *inode, int sync)
 		ep2->dentry.stream.start_clu = EXFAT_FREE_CLUSTER;
 	}
 
-	exfat_update_dir_chksum_with_entry_set(&es);
+	exfat_update_dir_chksum(&es);
 	return exfat_put_dentry_set(&es, sync);
 }
 
diff --git a/fs/exfat/namei.c b/fs/exfat/namei.c
index 9c549fd11f..631ad9e8e3 100644
--- a/fs/exfat/namei.c
+++ b/fs/exfat/namei.c
@@ -204,21 +204,16 @@ const struct dentry_operations exfat_utf8_dentry_ops = {
 	.d_compare	= exfat_utf8_d_cmp,
 };
 
-/* used only in search empty_slot() */
-#define CNT_UNUSED_NOHIT        (-1)
-#define CNT_UNUSED_HIT          (-2)
 /* search EMPTY CONTINUOUS "num_entries" entries */
 static int exfat_search_empty_slot(struct super_block *sb,
 		struct exfat_hint_femp *hint_femp, struct exfat_chain *p_dir,
-		int num_entries)
+		int num_entries, struct exfat_entry_set_cache *es)
 {
-	int i, dentry, num_empty = 0;
+	int i, dentry, ret;
 	int dentries_per_clu;
-	unsigned int type;
 	struct exfat_chain clu;
-	struct exfat_dentry *ep;
 	struct exfat_sb_info *sbi = EXFAT_SB(sb);
-	struct buffer_head *bh;
+	int total_entries = EXFAT_CLU_TO_DEN(p_dir->size, sbi);
 
 	dentries_per_clu = sbi->dentries_per_clu;
 
@@ -231,7 +226,7 @@ static int exfat_search_empty_slot(struct super_block *sb,
 		 * Otherwise, and if "dentry + hint_famp->count" is also equal
 		 * to "p_dir->size * dentries_per_clu", it means ENOSPC.
 		 */
-		if (dentry + hint_femp->count == p_dir->size * dentries_per_clu &&
+		if (dentry + hint_femp->count == total_entries &&
 		    num_entries > hint_femp->count)
 			return -ENOSPC;
 
@@ -242,69 +237,41 @@ static int exfat_search_empty_slot(struct super_block *sb,
 		dentry = 0;
 	}
 
-	while (clu.dir != EXFAT_EOF_CLUSTER) {
+	while (dentry + num_entries < total_entries &&
+	       clu.dir != EXFAT_EOF_CLUSTER) {
 		i = dentry & (dentries_per_clu - 1);
 
-		for (; i < dentries_per_clu; i++, dentry++) {
-			ep = exfat_get_dentry(sb, &clu, i, &bh);
-			if (!ep)
-				return -EIO;
-			type = exfat_get_entry_type(ep);
-			brelse(bh);
-
-			if (type == TYPE_UNUSED || type == TYPE_DELETED) {
-				num_empty++;
-				if (hint_femp->eidx == EXFAT_HINT_NONE) {
-					hint_femp->eidx = dentry;
-					hint_femp->count = CNT_UNUSED_NOHIT;
-					exfat_chain_set(&hint_femp->cur,
-						clu.dir, clu.size, clu.flags);
-				}
-
-				if (type == TYPE_UNUSED &&
-				    hint_femp->count != CNT_UNUSED_HIT)
-					hint_femp->count = CNT_UNUSED_HIT;
+		ret = exfat_get_empty_dentry_set(es, sb, &clu, i, num_entries);
+		if (ret < 0)
+			return ret;
+		else if (ret == 0)
+			return dentry;
+
+		dentry += ret;
+		i += ret;
+
+		while (i >= dentries_per_clu) {
+			if (clu.flags == ALLOC_NO_FAT_CHAIN) {
+				if (--clu.size > 0)
+					clu.dir++;
+				else
+					clu.dir = EXFAT_EOF_CLUSTER;
 			} else {
-				if (hint_femp->eidx != EXFAT_HINT_NONE &&
-				    hint_femp->count == CNT_UNUSED_HIT) {
-					/* unused empty group means
-					 * an empty group which includes
-					 * unused dentry
-					 */
-					exfat_fs_error(sb,
-						"found bogus dentry(%d) beyond unused empty group(%d) (start_clu : %u, cur_clu : %u)",
-						dentry, hint_femp->eidx,
-						p_dir->dir, clu.dir);
+				if (exfat_get_next_cluster(sb, &clu.dir))
 					return -EIO;
-				}
-
-				num_empty = 0;
-				hint_femp->eidx = EXFAT_HINT_NONE;
 			}
 
-			if (num_empty >= num_entries) {
-				/* found and invalidate hint_femp */
-				hint_femp->eidx = EXFAT_HINT_NONE;
-				return (dentry - (num_entries - 1));
-			}
-		}
-
-		if (clu.flags == ALLOC_NO_FAT_CHAIN) {
-			if (--clu.size > 0)
-				clu.dir++;
-			else
-				clu.dir = EXFAT_EOF_CLUSTER;
-		} else {
-			if (exfat_get_next_cluster(sb, &clu.dir))
-				return -EIO;
+			i -= dentries_per_clu;
 		}
 	}
 
-	hint_femp->eidx = p_dir->size * dentries_per_clu - num_empty;
-	hint_femp->count = num_empty;
-	if (num_empty == 0)
+	hint_femp->eidx = dentry;
+	hint_femp->count = 0;
+	if (dentry == total_entries || clu.dir == EXFAT_EOF_CLUSTER)
 		exfat_chain_set(&hint_femp->cur, EXFAT_EOF_CLUSTER, 0,
 				clu.flags);
+	else
+		hint_femp->cur = clu;
 
 	return -ENOSPC;
 }
@@ -325,7 +292,8 @@ static int exfat_check_max_dentries(struct inode *inode)
  * if there isn't any empty slot, expand cluster chain.
  */
 static int exfat_find_empty_entry(struct inode *inode,
-		struct exfat_chain *p_dir, int num_entries)
+		struct exfat_chain *p_dir, int num_entries,
+		struct exfat_entry_set_cache *es)
 {
 	int dentry;
 	unsigned int ret, last_clu;
@@ -344,7 +312,7 @@ static int exfat_find_empty_entry(struct inode *inode,
 	}
 
 	while ((dentry = exfat_search_empty_slot(sb, &hint_femp, p_dir,
-					num_entries)) < 0) {
+					num_entries, es)) < 0) {
 		if (dentry == -EIO)
 			break;
 
@@ -499,6 +467,8 @@ static int exfat_add_entry(struct inode *inode, const char *path,
 	struct exfat_sb_info *sbi = EXFAT_SB(sb);
 	struct exfat_uni_name uniname;
 	struct exfat_chain clu;
+	struct timespec64 ts = current_time(inode);
+	struct exfat_entry_set_cache es;
 	int clu_size = 0;
 	unsigned int start_clu = EXFAT_FREE_CLUSTER;
 
@@ -513,7 +483,7 @@ static int exfat_add_entry(struct inode *inode, const char *path,
 	}
 
 	/* exfat_find_empty_entry must be called before alloc_cluster() */
-	dentry = exfat_find_empty_entry(inode, p_dir, num_entries);
+	dentry = exfat_find_empty_entry(inode, p_dir, num_entries, &es);
 	if (dentry < 0) {
 		ret = dentry; /* -EIO or -ENOSPC */
 		goto out;
@@ -521,8 +491,10 @@ static int exfat_add_entry(struct inode *inode, const char *path,
 
 	if (type == TYPE_DIR && !sbi->options.zero_size_dir) {
 		ret = exfat_alloc_new_dir(inode, &clu);
-		if (ret)
+		if (ret) {
+			exfat_put_dentry_set(&es, false);
 			goto out;
+		}
 		start_clu = clu.dir;
 		clu_size = sbi->cluster_size;
 	}
@@ -531,12 +503,10 @@ static int exfat_add_entry(struct inode *inode, const char *path,
 	/* fill the dos name directory entry information of the created file.
 	 * the first cluster is not determined yet. (0)
 	 */
-	ret = exfat_init_dir_entry(inode, p_dir, dentry, type,
-		start_clu, clu_size);
-	if (ret)
-		goto out;
+	exfat_init_dir_entry(&es, type, start_clu, clu_size, &ts);
+	exfat_init_ext_entry(&es, num_entries, &uniname);
 
-	ret = exfat_init_ext_entry(inode, p_dir, dentry, num_entries, &uniname);
+	ret = exfat_put_dentry_set(&es, IS_DIRSYNC(inode));
 	if (ret)
 		goto out;
 
@@ -577,6 +547,7 @@ static int exfat_create(struct mnt_idmap *idmap, struct inode *dir,
 	struct exfat_dir_entry info;
 	loff_t i_pos;
 	int err;
+	loff_t size = i_size_read(dir);
 
 	mutex_lock(&EXFAT_SB(sb)->s_lock);
 	exfat_set_volume_dirty(sb);
@@ -587,7 +558,7 @@ static int exfat_create(struct mnt_idmap *idmap, struct inode *dir,
 
 	inode_inc_iversion(dir);
 	inode_set_mtime_to_ts(dir, inode_set_ctime_current(dir));
-	if (IS_DIRSYNC(dir))
+	if (IS_DIRSYNC(dir) && size != i_size_read(dir))
 		exfat_sync_inode(dir);
 	else
 		mark_inode_dirty(dir);
@@ -795,12 +766,11 @@ unlock:
 static int exfat_unlink(struct inode *dir, struct dentry *dentry)
 {
 	struct exfat_chain cdir;
-	struct exfat_dentry *ep;
 	struct super_block *sb = dir->i_sb;
 	struct inode *inode = dentry->d_inode;
 	struct exfat_inode_info *ei = EXFAT_I(inode);
-	struct buffer_head *bh;
-	int num_entries, entry, err = 0;
+	struct exfat_entry_set_cache es;
+	int entry, err = 0;
 
 	mutex_lock(&EXFAT_SB(sb)->s_lock);
 	exfat_chain_dup(&cdir, &ei->dir);
@@ -811,26 +781,20 @@ static int exfat_unlink(struct inode *dir, struct dentry *dentry)
 		goto unlock;
 	}
 
-	ep = exfat_get_dentry(sb, &cdir, entry, &bh);
-	if (!ep) {
-		err = -EIO;
-		goto unlock;
-	}
-	num_entries = exfat_count_ext_entries(sb, &cdir, entry, ep);
-	if (num_entries < 0) {
+	err = exfat_get_dentry_set(&es, sb, &cdir, entry, ES_ALL_ENTRIES);
+	if (err) {
 		err = -EIO;
-		brelse(bh);
 		goto unlock;
 	}
-	num_entries++;
-	brelse(bh);
 
 	exfat_set_volume_dirty(sb);
+
 	/* update the directory entry */
-	if (exfat_remove_entries(dir, &cdir, entry, 0, num_entries)) {
-		err = -EIO;
+	exfat_remove_entries(inode, &es, ES_IDX_FILE);
+
+	err = exfat_put_dentry_set(&es, IS_DIRSYNC(inode));
+	if (err)
 		goto unlock;
-	}
 
 	/* This doesn't modify ei */
 	ei->dir.dir = DIR_DELETED;
@@ -838,10 +802,7 @@ static int exfat_unlink(struct inode *dir, struct dentry *dentry)
 	inode_inc_iversion(dir);
 	simple_inode_init_ts(dir);
 	exfat_truncate_inode_atime(dir);
-	if (IS_DIRSYNC(dir))
-		exfat_sync_inode(dir);
-	else
-		mark_inode_dirty(dir);
+	mark_inode_dirty(dir);
 
 	clear_nlink(inode);
 	simple_inode_init_ts(inode);
@@ -862,6 +823,7 @@ static int exfat_mkdir(struct mnt_idmap *idmap, struct inode *dir,
 	struct exfat_chain cdir;
 	loff_t i_pos;
 	int err;
+	loff_t size = i_size_read(dir);
 
 	mutex_lock(&EXFAT_SB(sb)->s_lock);
 	exfat_set_volume_dirty(sb);
@@ -872,7 +834,7 @@ static int exfat_mkdir(struct mnt_idmap *idmap, struct inode *dir,
 
 	inode_inc_iversion(dir);
 	inode_set_mtime_to_ts(dir, inode_set_ctime_current(dir));
-	if (IS_DIRSYNC(dir))
+	if (IS_DIRSYNC(dir) && size != i_size_read(dir))
 		exfat_sync_inode(dir);
 	else
 		mark_inode_dirty(dir);
@@ -946,13 +908,12 @@ static int exfat_check_dir_empty(struct super_block *sb,
 static int exfat_rmdir(struct inode *dir, struct dentry *dentry)
 {
 	struct inode *inode = dentry->d_inode;
-	struct exfat_dentry *ep;
 	struct exfat_chain cdir, clu_to_free;
 	struct super_block *sb = inode->i_sb;
 	struct exfat_sb_info *sbi = EXFAT_SB(sb);
 	struct exfat_inode_info *ei = EXFAT_I(inode);
-	struct buffer_head *bh;
-	int num_entries, entry, err;
+	struct exfat_entry_set_cache es;
+	int entry, err;
 
 	mutex_lock(&EXFAT_SB(inode->i_sb)->s_lock);
 
@@ -976,27 +937,20 @@ static int exfat_rmdir(struct inode *dir, struct dentry *dentry)
 		goto unlock;
 	}
 
-	ep = exfat_get_dentry(sb, &cdir, entry, &bh);
-	if (!ep) {
-		err = -EIO;
-		goto unlock;
-	}
-
-	num_entries = exfat_count_ext_entries(sb, &cdir, entry, ep);
-	if (num_entries < 0) {
+	err = exfat_get_dentry_set(&es, sb, &cdir, entry, ES_ALL_ENTRIES);
+	if (err) {
 		err = -EIO;
-		brelse(bh);
 		goto unlock;
 	}
-	num_entries++;
-	brelse(bh);
 
 	exfat_set_volume_dirty(sb);
-	err = exfat_remove_entries(dir, &cdir, entry, 0, num_entries);
-	if (err) {
-		exfat_err(sb, "failed to exfat_remove_entries : err(%d)", err);
+
+	exfat_remove_entries(inode, &es, ES_IDX_FILE);
+
+	err = exfat_put_dentry_set(&es, IS_DIRSYNC(dir));
+	if (err)
 		goto unlock;
-	}
+
 	ei->dir.dir = DIR_DELETED;
 
 	inode_inc_iversion(dir);
@@ -1022,67 +976,52 @@ static int exfat_rename_file(struct inode *inode, struct exfat_chain *p_dir,
 		int oldentry, struct exfat_uni_name *p_uniname,
 		struct exfat_inode_info *ei)
 {
-	int ret, num_old_entries, num_new_entries;
+	int ret, num_new_entries;
 	struct exfat_dentry *epold, *epnew;
 	struct super_block *sb = inode->i_sb;
-	struct buffer_head *new_bh, *old_bh;
+	struct exfat_entry_set_cache old_es, new_es;
 	int sync = IS_DIRSYNC(inode);
 
-	epold = exfat_get_dentry(sb, p_dir, oldentry, &old_bh);
-	if (!epold)
-		return -EIO;
-
-	num_old_entries = exfat_count_ext_entries(sb, p_dir, oldentry, epold);
-	if (num_old_entries < 0)
-		return -EIO;
-	num_old_entries++;
-
 	num_new_entries = exfat_calc_num_entries(p_uniname);
 	if (num_new_entries < 0)
 		return num_new_entries;
 
-	if (num_old_entries < num_new_entries) {
-		int newentry;
+	ret = exfat_get_dentry_set(&old_es, sb, p_dir, oldentry, ES_ALL_ENTRIES);
+	if (ret) {
+		ret = -EIO;
+		return ret;
+	}
 
-		newentry =
-			exfat_find_empty_entry(inode, p_dir, num_new_entries);
-		if (newentry < 0)
-			return newentry; /* -EIO or -ENOSPC */
+	epold = exfat_get_dentry_cached(&old_es, ES_IDX_FILE);
 
-		epnew = exfat_get_dentry(sb, p_dir, newentry, &new_bh);
-		if (!epnew)
-			return -EIO;
+	if (old_es.num_entries < num_new_entries) {
+		int newentry;
 
+		newentry = exfat_find_empty_entry(inode, p_dir, num_new_entries,
+				&new_es);
+		if (newentry < 0) {
+			ret = newentry; /* -EIO or -ENOSPC */
+			goto put_old_es;
+		}
+
+		epnew = exfat_get_dentry_cached(&new_es, ES_IDX_FILE);
 		*epnew = *epold;
 		if (exfat_get_entry_type(epnew) == TYPE_FILE) {
 			epnew->dentry.file.attr |= cpu_to_le16(EXFAT_ATTR_ARCHIVE);
 			ei->attr |= EXFAT_ATTR_ARCHIVE;
 		}
-		exfat_update_bh(new_bh, sync);
-		brelse(old_bh);
-		brelse(new_bh);
-
-		epold = exfat_get_dentry(sb, p_dir, oldentry + 1, &old_bh);
-		if (!epold)
-			return -EIO;
-		epnew = exfat_get_dentry(sb, p_dir, newentry + 1, &new_bh);
-		if (!epnew) {
-			brelse(old_bh);
-			return -EIO;
-		}
 
+		epold = exfat_get_dentry_cached(&old_es, ES_IDX_STREAM);
+		epnew = exfat_get_dentry_cached(&new_es, ES_IDX_STREAM);
 		*epnew = *epold;
-		exfat_update_bh(new_bh, sync);
-		brelse(old_bh);
-		brelse(new_bh);
 
-		ret = exfat_init_ext_entry(inode, p_dir, newentry,
-			num_new_entries, p_uniname);
+		exfat_init_ext_entry(&new_es, num_new_entries, p_uniname);
+
+		ret = exfat_put_dentry_set(&new_es, sync);
 		if (ret)
-			return ret;
+			goto put_old_es;
 
-		exfat_remove_entries(inode, p_dir, oldentry, 0,
-			num_old_entries);
+		exfat_remove_entries(inode, &old_es, ES_IDX_FILE);
 		ei->dir = *p_dir;
 		ei->entry = newentry;
 	} else {
@@ -1090,85 +1029,72 @@ static int exfat_rename_file(struct inode *inode, struct exfat_chain *p_dir,
 			epold->dentry.file.attr |= cpu_to_le16(EXFAT_ATTR_ARCHIVE);
 			ei->attr |= EXFAT_ATTR_ARCHIVE;
 		}
-		exfat_update_bh(old_bh, sync);
-		brelse(old_bh);
-		ret = exfat_init_ext_entry(inode, p_dir, oldentry,
-			num_new_entries, p_uniname);
-		if (ret)
-			return ret;
 
-		exfat_remove_entries(inode, p_dir, oldentry, num_new_entries,
-			num_old_entries);
+		exfat_remove_entries(inode, &old_es, ES_IDX_FIRST_FILENAME + 1);
+		exfat_init_ext_entry(&old_es, num_new_entries, p_uniname);
 	}
-	return 0;
+	return exfat_put_dentry_set(&old_es, sync);
+
+put_old_es:
+	exfat_put_dentry_set(&old_es, false);
+	return ret;
 }
 
 static int exfat_move_file(struct inode *inode, struct exfat_chain *p_olddir,
 		int oldentry, struct exfat_chain *p_newdir,
 		struct exfat_uni_name *p_uniname, struct exfat_inode_info *ei)
 {
-	int ret, newentry, num_new_entries, num_old_entries;
+	int ret, newentry, num_new_entries;
 	struct exfat_dentry *epmov, *epnew;
 	struct super_block *sb = inode->i_sb;
-	struct buffer_head *mov_bh, *new_bh;
-
-	epmov = exfat_get_dentry(sb, p_olddir, oldentry, &mov_bh);
-	if (!epmov)
-		return -EIO;
-
-	num_old_entries = exfat_count_ext_entries(sb, p_olddir, oldentry,
-		epmov);
-	if (num_old_entries < 0)
-		return -EIO;
-	num_old_entries++;
+	struct exfat_entry_set_cache mov_es, new_es;
 
 	num_new_entries = exfat_calc_num_entries(p_uniname);
 	if (num_new_entries < 0)
 		return num_new_entries;
 
-	newentry = exfat_find_empty_entry(inode, p_newdir, num_new_entries);
-	if (newentry < 0)
-		return newentry; /* -EIO or -ENOSPC */
-
-	epnew = exfat_get_dentry(sb, p_newdir, newentry, &new_bh);
-	if (!epnew)
+	ret = exfat_get_dentry_set(&mov_es, sb, p_olddir, oldentry,
+			ES_ALL_ENTRIES);
+	if (ret)
 		return -EIO;
 
+	newentry = exfat_find_empty_entry(inode, p_newdir, num_new_entries,
+			&new_es);
+	if (newentry < 0) {
+		ret = newentry; /* -EIO or -ENOSPC */
+		goto put_mov_es;
+	}
+
+	epmov = exfat_get_dentry_cached(&mov_es, ES_IDX_FILE);
+	epnew = exfat_get_dentry_cached(&new_es, ES_IDX_FILE);
 	*epnew = *epmov;
 	if (exfat_get_entry_type(epnew) == TYPE_FILE) {
 		epnew->dentry.file.attr |= cpu_to_le16(EXFAT_ATTR_ARCHIVE);
 		ei->attr |= EXFAT_ATTR_ARCHIVE;
 	}
-	exfat_update_bh(new_bh, IS_DIRSYNC(inode));
-	brelse(mov_bh);
-	brelse(new_bh);
-
-	epmov = exfat_get_dentry(sb, p_olddir, oldentry + 1, &mov_bh);
-	if (!epmov)
-		return -EIO;
-	epnew = exfat_get_dentry(sb, p_newdir, newentry + 1, &new_bh);
-	if (!epnew) {
-		brelse(mov_bh);
-		return -EIO;
-	}
 
+	epmov = exfat_get_dentry_cached(&mov_es, ES_IDX_STREAM);
+	epnew = exfat_get_dentry_cached(&new_es, ES_IDX_STREAM);
 	*epnew = *epmov;
-	exfat_update_bh(new_bh, IS_DIRSYNC(inode));
-	brelse(mov_bh);
-	brelse(new_bh);
-
-	ret = exfat_init_ext_entry(inode, p_newdir, newentry, num_new_entries,
-		p_uniname);
-	if (ret)
-		return ret;
 
-	exfat_remove_entries(inode, p_olddir, oldentry, 0, num_old_entries);
+	exfat_init_ext_entry(&new_es, num_new_entries, p_uniname);
+	exfat_remove_entries(inode, &mov_es, ES_IDX_FILE);
 
 	exfat_chain_set(&ei->dir, p_newdir->dir, p_newdir->size,
 		p_newdir->flags);
 
 	ei->entry = newentry;
-	return 0;
+
+	ret = exfat_put_dentry_set(&new_es, IS_DIRSYNC(inode));
+	if (ret)
+		goto put_mov_es;
+
+	return exfat_put_dentry_set(&mov_es, IS_DIRSYNC(inode));
+
+put_mov_es:
+	exfat_put_dentry_set(&mov_es, false);
+
+	return ret;
 }
 
 /* rename or move a old file into a new file */
@@ -1186,7 +1112,6 @@ static int __exfat_rename(struct inode *old_parent_inode,
 	struct exfat_sb_info *sbi = EXFAT_SB(sb);
 	const unsigned char *new_path = new_dentry->d_name.name;
 	struct inode *new_inode = new_dentry->d_inode;
-	int num_entries;
 	struct exfat_inode_info *new_ei = NULL;
 	unsigned int new_entry_type = TYPE_UNUSED;
 	int new_entry = 0;
@@ -1257,25 +1182,21 @@ static int __exfat_rename(struct inode *old_parent_inode,
 				&newdir, &uni_name, ei);
 
 	if (!ret && new_inode) {
+		struct exfat_entry_set_cache es;
+
 		/* delete entries of new_dir */
-		ep = exfat_get_dentry(sb, p_dir, new_entry, &new_bh);
-		if (!ep) {
+		ret = exfat_get_dentry_set(&es, sb, p_dir, new_entry,
+				ES_ALL_ENTRIES);
+		if (ret) {
 			ret = -EIO;
 			goto del_out;
 		}
 
-		num_entries = exfat_count_ext_entries(sb, p_dir, new_entry, ep);
-		if (num_entries < 0) {
-			ret = -EIO;
-			goto del_out;
-		}
-		brelse(new_bh);
+		exfat_remove_entries(new_inode, &es, ES_IDX_FILE);
 
-		if (exfat_remove_entries(new_inode, p_dir, new_entry, 0,
-				num_entries + 1)) {
-			ret = -EIO;
+		ret = exfat_put_dentry_set(&es, IS_DIRSYNC(new_inode));
+		if (ret)
 			goto del_out;
-		}
 
 		/* Free the clusters if new_inode is a dir(as if exfat_rmdir) */
 		if (new_entry_type == TYPE_DIR &&
@@ -1317,6 +1238,7 @@ static int exfat_rename(struct mnt_idmap *idmap,
 	struct super_block *sb = old_dir->i_sb;
 	loff_t i_pos;
 	int err;
+	loff_t size = i_size_read(new_dir);
 
 	/*
 	 * The VFS already checks for existence, so for local filesystems
@@ -1338,7 +1260,7 @@ static int exfat_rename(struct mnt_idmap *idmap,
 	simple_rename_timestamp(old_dir, old_dentry, new_dir, new_dentry);
 	EXFAT_I(new_dir)->i_crtime = current_time(new_dir);
 	exfat_truncate_inode_atime(new_dir);
-	if (IS_DIRSYNC(new_dir))
+	if (IS_DIRSYNC(new_dir) && size != i_size_read(new_dir))
 		exfat_sync_inode(new_dir);
 	else
 		mark_inode_dirty(new_dir);
@@ -1359,9 +1281,7 @@ static int exfat_rename(struct mnt_idmap *idmap,
 	}
 
 	inode_inc_iversion(old_dir);
-	if (IS_DIRSYNC(old_dir))
-		exfat_sync_inode(old_dir);
-	else
+	if (new_dir != old_dir)
 		mark_inode_dirty(old_dir);
 
 	if (new_inode) {
diff --git a/fs/exfat/super.c b/fs/exfat/super.c
index fcb6582677..3d5ea2cfad 100644
--- a/fs/exfat/super.c
+++ b/fs/exfat/super.c
@@ -813,7 +813,7 @@ static int __init init_exfat_fs(void)
 
 	exfat_inode_cachep = kmem_cache_create("exfat_inode_cache",
 			sizeof(struct exfat_inode_info),
-			0, SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD,
+			0, SLAB_RECLAIM_ACCOUNT,
 			exfat_inode_init_once);
 	if (!exfat_inode_cachep) {
 		err = -ENOMEM;
diff --git a/fs/exportfs/expfs.c b/fs/exportfs/expfs.c
index 3ae0154c56..07ea3d62b2 100644
--- a/fs/exportfs/expfs.c
+++ b/fs/exportfs/expfs.c
@@ -255,7 +255,7 @@ static bool filldir_one(struct dir_context *ctx, const char *name, int len,
 		container_of(ctx, struct getdents_callback, ctx);
 
 	buf->sequence++;
-	if (buf->ino == ino && len <= NAME_MAX) {
+	if (buf->ino == ino && len <= NAME_MAX && !is_dot_dotdot(name, len)) {
 		memcpy(buf->name, name, len);
 		buf->name[len] = '\0';
 		buf->found = 1;
diff --git a/fs/ext2/Kconfig b/fs/ext2/Kconfig
index 74d9896590..d5bce83ad9 100644
--- a/fs/ext2/Kconfig
+++ b/fs/ext2/Kconfig
@@ -1,16 +1,22 @@
 # SPDX-License-Identifier: GPL-2.0-only
 config EXT2_FS
-	tristate "Second extended fs support"
+	tristate "Second extended fs support (DEPRECATED)"
 	select BUFFER_HEAD
 	select FS_IOMAP
-	select LEGACY_DIRECT_IO
 	help
 	  Ext2 is a standard Linux file system for hard disks.
 
-	  To compile this file system support as a module, choose M here: the
-	  module will be called ext2.
+	  This filesystem driver is deprecated because it does not properly
+	  support inode time stamps beyond 03:14:07 UTC on 19 January 2038.
 
-	  If unsure, say Y.
+	  Ext2 users are advised to use ext4 driver to access their filesystem.
+	  The driver is fully compatible, supports filesystems without journal
+          or extents, and also supports larger time stamps if the filesystem
+          is created with at least 256 byte inodes.
+
+	  This code is kept as a simple reference for filesystem developers.
+
+	  If unsure, say N.
 
 config EXT2_FS_XATTR
 	bool "Ext2 extended attributes"
diff --git a/fs/ext2/balloc.c b/fs/ext2/balloc.c
index e124f3d709..b8cfab8f98 100644
--- a/fs/ext2/balloc.c
+++ b/fs/ext2/balloc.c
@@ -77,26 +77,33 @@ static int ext2_valid_block_bitmap(struct super_block *sb,
 	ext2_grpblk_t next_zero_bit;
 	ext2_fsblk_t bitmap_blk;
 	ext2_fsblk_t group_first_block;
+	ext2_grpblk_t max_bit;
 
 	group_first_block = ext2_group_first_block_no(sb, block_group);
+	max_bit = ext2_group_last_block_no(sb, block_group) - group_first_block;
 
 	/* check whether block bitmap block number is set */
 	bitmap_blk = le32_to_cpu(desc->bg_block_bitmap);
 	offset = bitmap_blk - group_first_block;
-	if (!ext2_test_bit(offset, bh->b_data))
+	if (offset < 0 || offset > max_bit ||
+	    !ext2_test_bit(offset, bh->b_data))
 		/* bad block bitmap */
 		goto err_out;
 
 	/* check whether the inode bitmap block number is set */
 	bitmap_blk = le32_to_cpu(desc->bg_inode_bitmap);
 	offset = bitmap_blk - group_first_block;
-	if (!ext2_test_bit(offset, bh->b_data))
+	if (offset < 0 || offset > max_bit ||
+	    !ext2_test_bit(offset, bh->b_data))
 		/* bad block bitmap */
 		goto err_out;
 
 	/* check whether the inode table block number is set */
 	bitmap_blk = le32_to_cpu(desc->bg_inode_table);
 	offset = bitmap_blk - group_first_block;
+	if (offset < 0 || offset > max_bit ||
+	    offset + EXT2_SB(sb)->s_itb_per_group - 1 > max_bit)
+		goto err_out;
 	next_zero_bit = ext2_find_next_zero_bit(bh->b_data,
 				offset + EXT2_SB(sb)->s_itb_per_group,
 				offset);
@@ -412,7 +419,7 @@ void ext2_init_block_alloc_info(struct inode *inode)
 	struct ext2_block_alloc_info *block_i;
 	struct super_block *sb = inode->i_sb;
 
-	block_i = kmalloc(sizeof(*block_i), GFP_NOFS);
+	block_i = kmalloc(sizeof(*block_i), GFP_KERNEL);
 	if (block_i) {
 		struct ext2_reserve_window_node *rsv = &block_i->rsv_window_node;
 
diff --git a/fs/ext2/dir.c b/fs/ext2/dir.c
index 4fb155b5a9..087457061c 100644
--- a/fs/ext2/dir.c
+++ b/fs/ext2/dir.c
@@ -175,7 +175,6 @@ Eend:
 			(unsigned long) le32_to_cpu(p->inode));
 	}
 fail:
-	folio_set_error(folio);
 	return false;
 }
 
diff --git a/fs/ext2/file.c b/fs/ext2/file.c
index 4ddc36f4db..10b061ac5b 100644
--- a/fs/ext2/file.c
+++ b/fs/ext2/file.c
@@ -302,6 +302,12 @@ static ssize_t ext2_file_write_iter(struct kiocb *iocb, struct iov_iter *from)
 	return generic_file_write_iter(iocb, from);
 }
 
+static int ext2_file_open(struct inode *inode, struct file *filp)
+{
+	filp->f_mode |= FMODE_CAN_ODIRECT;
+	return dquot_file_open(inode, filp);
+}
+
 const struct file_operations ext2_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read_iter	= ext2_file_read_iter,
@@ -311,7 +317,7 @@ const struct file_operations ext2_file_operations = {
 	.compat_ioctl	= ext2_compat_ioctl,
 #endif
 	.mmap		= ext2_file_mmap,
-	.open		= dquot_file_open,
+	.open		= ext2_file_open,
 	.release	= ext2_release_file,
 	.fsync		= ext2_fsync,
 	.get_unmapped_area = thp_get_unmapped_area,
diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c
index 5a4272b2c6..0caa1650ce 100644
--- a/fs/ext2/inode.c
+++ b/fs/ext2/inode.c
@@ -754,7 +754,7 @@ static int ext2_get_blocks(struct inode *inode,
 		 */
 		err = sb_issue_zeroout(inode->i_sb,
 				le32_to_cpu(chain[depth-1].key), count,
-				GFP_NOFS);
+				GFP_KERNEL);
 		if (err) {
 			mutex_unlock(&ei->truncate_mutex);
 			goto cleanup;
@@ -965,7 +965,6 @@ const struct address_space_operations ext2_aops = {
 	.write_begin		= ext2_write_begin,
 	.write_end		= ext2_write_end,
 	.bmap			= ext2_bmap,
-	.direct_IO		= noop_direct_IO,
 	.writepages		= ext2_writepages,
 	.migrate_folio		= buffer_migrate_folio,
 	.is_partially_uptodate	= block_is_partially_uptodate,
@@ -974,7 +973,6 @@ const struct address_space_operations ext2_aops = {
 
 static const struct address_space_operations ext2_dax_aops = {
 	.writepages		= ext2_dax_writepages,
-	.direct_IO		= noop_direct_IO,
 	.dirty_folio		= noop_dirty_folio,
 };
 
diff --git a/fs/ext2/super.c b/fs/ext2/super.c
index 6d8587505c..37f7ce56ad 100644
--- a/fs/ext2/super.c
+++ b/fs/ext2/super.c
@@ -213,8 +213,7 @@ static int __init init_inodecache(void)
 {
 	ext2_inode_cachep = kmem_cache_create_usercopy("ext2_inode_cache",
 				sizeof(struct ext2_inode_info), 0,
-				(SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD|
-					SLAB_ACCOUNT),
+				SLAB_RECLAIM_ACCOUNT | SLAB_ACCOUNT,
 				offsetof(struct ext2_inode_info, i_data),
 				sizeof_field(struct ext2_inode_info, i_data),
 				init_once);
diff --git a/fs/ext2/xattr.c b/fs/ext2/xattr.c
index e849241ebb..c885dcc3bd 100644
--- a/fs/ext2/xattr.c
+++ b/fs/ext2/xattr.c
@@ -874,7 +874,7 @@ ext2_xattr_cache_insert(struct mb_cache *cache, struct buffer_head *bh)
 	__u32 hash = le32_to_cpu(HDR(bh)->h_hash);
 	int error;
 
-	error = mb_cache_entry_create(cache, GFP_NOFS, hash, bh->b_blocknr,
+	error = mb_cache_entry_create(cache, GFP_KERNEL, hash, bh->b_blocknr,
 				      true);
 	if (error) {
 		if (error == -EBUSY) {
diff --git a/fs/ext4/acl.h b/fs/ext4/acl.h
index ef4c19e5f5..0c5a79c3b5 100644
--- a/fs/ext4/acl.h
+++ b/fs/ext4/acl.h
@@ -68,11 +68,6 @@ extern int ext4_init_acl(handle_t *, struct inode *, struct inode *);
 static inline int
 ext4_init_acl(handle_t *handle, struct inode *inode, struct inode *dir)
 {
-	/* usually, the umask is applied by posix_acl_create(), but if
-	   ext4 ACL support is disabled at compile time, we need to do
-	   it here, because posix_acl_create() will never be called */
-	inode->i_mode &= ~current_umask();
-
 	return 0;
 }
 #endif  /* CONFIG_EXT4_FS_POSIX_ACL */
diff --git a/fs/ext4/dir.c b/fs/ext4/dir.c
index 3985f8c33f..ff4514e462 100644
--- a/fs/ext4/dir.c
+++ b/fs/ext4/dir.c
@@ -192,7 +192,7 @@ static int ext4_readdir(struct file *file, struct dir_context *ctx)
 					(PAGE_SHIFT - inode->i_blkbits);
 			if (!ra_has_index(&file->f_ra, index))
 				page_cache_sync_readahead(
-					sb->s_bdev->bd_inode->i_mapping,
+					sb->s_bdev->bd_mapping,
 					&file->f_ra, file,
 					index, 1);
 			file->f_ra.prev_pos = (loff_t)index << PAGE_SHIFT;
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 4fb938ba3f..efed7f0987 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -213,11 +213,14 @@ enum criteria {
 #define EXT4_MB_USE_RESERVED		0x2000
 /* Do strict check for free blocks while retrying block allocation */
 #define EXT4_MB_STRICT_CHECK		0x4000
-/* Large fragment size list lookup succeeded at least once for cr = 0 */
+/* Large fragment size list lookup succeeded at least once for
+ * CR_POWER2_ALIGNED */
 #define EXT4_MB_CR_POWER2_ALIGNED_OPTIMIZED		0x8000
-/* Avg fragment size rb tree lookup succeeded at least once for cr = 1 */
+/* Avg fragment size rb tree lookup succeeded at least once for
+ * CR_GOAL_LEN_FAST */
 #define EXT4_MB_CR_GOAL_LEN_FAST_OPTIMIZED		0x00010000
-/* Avg fragment size rb tree lookup succeeded at least once for cr = 1.5 */
+/* Avg fragment size rb tree lookup succeeded at least once for
+ * CR_BEST_AVAIL_LEN */
 #define EXT4_MB_CR_BEST_AVAIL_LEN_OPTIMIZED		0x00020000
 
 struct ext4_allocation_request {
@@ -1344,7 +1347,7 @@ struct ext4_super_block {
 /*60*/	__le32	s_feature_incompat;	/* incompatible feature set */
 	__le32	s_feature_ro_compat;	/* readonly-compatible feature set */
 /*68*/	__u8	s_uuid[16];		/* 128-bit uuid for volume */
-/*78*/	char	s_volume_name[EXT4_LABEL_MAX];	/* volume name */
+/*78*/	char	s_volume_name[EXT4_LABEL_MAX] __nonstring; /* volume name */
 /*88*/	char	s_last_mounted[64] __nonstring;	/* directory where last mounted */
 /*C8*/	__le32	s_algorithm_usage_bitmap; /* For compression */
 	/*
@@ -1550,7 +1553,7 @@ struct ext4_sb_info {
 	unsigned long s_commit_interval;
 	u32 s_max_batch_time;
 	u32 s_min_batch_time;
-	struct bdev_handle *s_journal_bdev_handle;
+	struct file *s_journal_bdev_file;
 #ifdef CONFIG_QUOTA
 	/* Names of quota files with journalled quota */
 	char __rcu *s_qf_names[EXT4_MAXQUOTAS];
diff --git a/fs/ext4/ext4_jbd2.c b/fs/ext4/ext4_jbd2.c
index 5d8055161a..da4a824563 100644
--- a/fs/ext4/ext4_jbd2.c
+++ b/fs/ext4/ext4_jbd2.c
@@ -206,7 +206,7 @@ static void ext4_journal_abort_handle(const char *caller, unsigned int line,
 
 static void ext4_check_bdev_write_error(struct super_block *sb)
 {
-	struct address_space *mapping = sb->s_bdev->bd_inode->i_mapping;
+	struct address_space *mapping = sb->s_bdev->bd_mapping;
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
 	int err;
 
diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index 7669d154c0..e067f2dd03 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -3402,9 +3402,10 @@ static int ext4_ext_convert_to_initialized(handle_t *handle,
 	struct ext4_extent *ex, *abut_ex;
 	ext4_lblk_t ee_block, eof_block;
 	unsigned int ee_len, depth, map_len = map->m_len;
-	int allocated = 0, max_zeroout = 0;
 	int err = 0;
 	int split_flag = EXT4_EXT_DATA_VALID2;
+	int allocated = 0;
+	unsigned int max_zeroout = 0;
 
 	ext_debug(inode, "logical block %llu, max_blocks %u\n",
 		  (unsigned long long)map->m_lblk, map_len);
@@ -4111,10 +4112,10 @@ insert_hole:
  *
  * Need to be called with
  * down_read(&EXT4_I(inode)->i_data_sem) if not allocating file system block
- * (ie, create is zero). Otherwise down_write(&EXT4_I(inode)->i_data_sem)
+ * (ie, flags is zero). Otherwise down_write(&EXT4_I(inode)->i_data_sem)
  *
  * return > 0, number of blocks already mapped/allocated
- *          if create == 0 and these are pre-allocated blocks
+ *          if flags doesn't contain EXT4_GET_BLOCKS_CREATE and these are pre-allocated blocks
  *          	buffer head is unmapped
  *          otherwise blocks are mapped
  *
@@ -4218,7 +4219,7 @@ int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,
 
 	/*
 	 * requested block isn't allocated yet;
-	 * we couldn't try to create block if create flag is zero
+	 * we couldn't try to create block if flags doesn't contain EXT4_GET_BLOCKS_CREATE
 	 */
 	if ((flags & EXT4_GET_BLOCKS_CREATE) == 0) {
 		ext4_lblk_t len;
diff --git a/fs/ext4/extents_status.c b/fs/ext4/extents_status.c
index 4a00e2f019..3a53dbb85e 100644
--- a/fs/ext4/extents_status.c
+++ b/fs/ext4/extents_status.c
@@ -310,6 +310,8 @@ void ext4_es_find_extent_range(struct inode *inode,
 			       ext4_lblk_t lblk, ext4_lblk_t end,
 			       struct extent_status *es)
 {
+	es->es_lblk = es->es_len = es->es_pblk = 0;
+
 	if (EXT4_SB(inode->i_sb)->s_mount_state & EXT4_FC_REPLAY)
 		return;
 
diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c
index 87c009e0c5..d3a67bc06d 100644
--- a/fs/ext4/fast_commit.c
+++ b/fs/ext4/fast_commit.c
@@ -649,6 +649,12 @@ void ext4_fc_track_range(handle_t *handle, struct inode *inode, ext4_lblk_t star
 	if (ext4_test_mount_flag(inode->i_sb, EXT4_MF_FC_INELIGIBLE))
 		return;
 
+	if (ext4_has_inline_data(inode)) {
+		ext4_fc_mark_ineligible(inode->i_sb, EXT4_FC_REASON_XATTR,
+					handle);
+		return;
+	}
+
 	args.start = start;
 	args.end = end;
 
diff --git a/fs/ext4/file.c b/fs/ext4/file.c
index 54d6ff2258..c89e434db6 100644
--- a/fs/ext4/file.c
+++ b/fs/ext4/file.c
@@ -844,8 +844,7 @@ static int ext4_sample_last_mounted(struct super_block *sb,
 	if (err)
 		goto out_journal;
 	lock_buffer(sbi->s_sbh);
-	strncpy(sbi->s_es->s_last_mounted, cp,
-		sizeof(sbi->s_es->s_last_mounted));
+	strtomem_pad(sbi->s_es->s_last_mounted, cp, 0);
 	ext4_superblock_csum_set(sb);
 	unlock_buffer(sbi->s_sbh);
 	ext4_handle_dirty_metadata(handle, NULL, sbi->s_sbh);
@@ -885,8 +884,7 @@ static int ext4_file_open(struct inode *inode, struct file *filp)
 			return ret;
 	}
 
-	filp->f_mode |= FMODE_NOWAIT | FMODE_BUF_RASYNC |
-			FMODE_DIO_PARALLEL_WRITE;
+	filp->f_mode |= FMODE_NOWAIT | FMODE_CAN_ODIRECT;
 	return dquot_file_open(inode, filp);
 }
 
@@ -938,7 +936,6 @@ const struct file_operations ext4_file_operations = {
 	.compat_ioctl	= ext4_compat_ioctl,
 #endif
 	.mmap		= ext4_file_mmap,
-	.mmap_supported_flags = MAP_SYNC,
 	.open		= ext4_file_open,
 	.release	= ext4_release_file,
 	.fsync		= ext4_sync_file,
@@ -946,6 +943,8 @@ const struct file_operations ext4_file_operations = {
 	.splice_read	= ext4_file_splice_read,
 	.splice_write	= iter_file_splice_write,
 	.fallocate	= ext4_fallocate,
+	.fop_flags	= FOP_MMAP_SYNC | FOP_BUFFER_RASYNC |
+			  FOP_DIO_PARALLEL_WRITE,
 };
 
 const struct inode_operations ext4_file_inode_operations = {
diff --git a/fs/ext4/fsmap.c b/fs/ext4/fsmap.c
index 11e6f33677..df853c4d3a 100644
--- a/fs/ext4/fsmap.c
+++ b/fs/ext4/fsmap.c
@@ -576,9 +576,9 @@ static bool ext4_getfsmap_is_valid_device(struct super_block *sb,
 	if (fm->fmr_device == 0 || fm->fmr_device == UINT_MAX ||
 	    fm->fmr_device == new_encode_dev(sb->s_bdev->bd_dev))
 		return true;
-	if (EXT4_SB(sb)->s_journal_bdev_handle &&
+	if (EXT4_SB(sb)->s_journal_bdev_file &&
 	    fm->fmr_device ==
-	    new_encode_dev(EXT4_SB(sb)->s_journal_bdev_handle->bdev->bd_dev))
+	    new_encode_dev(file_bdev(EXT4_SB(sb)->s_journal_bdev_file)->bd_dev))
 		return true;
 	return false;
 }
@@ -648,9 +648,9 @@ int ext4_getfsmap(struct super_block *sb, struct ext4_fsmap_head *head,
 	memset(handlers, 0, sizeof(handlers));
 	handlers[0].gfd_dev = new_encode_dev(sb->s_bdev->bd_dev);
 	handlers[0].gfd_fn = ext4_getfsmap_datadev;
-	if (EXT4_SB(sb)->s_journal_bdev_handle) {
+	if (EXT4_SB(sb)->s_journal_bdev_file) {
 		handlers[1].gfd_dev = new_encode_dev(
-			EXT4_SB(sb)->s_journal_bdev_handle->bdev->bd_dev);
+			file_bdev(EXT4_SB(sb)->s_journal_bdev_file)->bd_dev);
 		handlers[1].gfd_fn = ext4_getfsmap_logdev;
 	}
 
diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c
index d5bd1e3a5d..e7a09a9983 100644
--- a/fs/ext4/inline.c
+++ b/fs/ext4/inline.c
@@ -1410,7 +1410,11 @@ int ext4_inlinedir_to_tree(struct file *dir_file,
 			hinfo->hash = EXT4_DIRENT_HASH(de);
 			hinfo->minor_hash = EXT4_DIRENT_MINOR_HASH(de);
 		} else {
-			ext4fs_dirhash(dir, de->name, de->name_len, hinfo);
+			err = ext4fs_dirhash(dir, de->name, de->name_len, hinfo);
+			if (err) {
+				ret = err;
+				goto out;
+			}
 		}
 		if ((hinfo->hash < start_hash) ||
 		    ((hinfo->hash == start_hash) &&
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 31604907af..238e196338 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -453,6 +453,35 @@ static void ext4_map_blocks_es_recheck(handle_t *handle,
 }
 #endif /* ES_AGGRESSIVE_TEST */
 
+static int ext4_map_query_blocks(handle_t *handle, struct inode *inode,
+				 struct ext4_map_blocks *map)
+{
+	unsigned int status;
+	int retval;
+
+	if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
+		retval = ext4_ext_map_blocks(handle, inode, map, 0);
+	else
+		retval = ext4_ind_map_blocks(handle, inode, map, 0);
+
+	if (retval <= 0)
+		return retval;
+
+	if (unlikely(retval != map->m_len)) {
+		ext4_warning(inode->i_sb,
+			     "ES len assertion failed for inode "
+			     "%lu: retval %d != map->m_len %d",
+			     inode->i_ino, retval, map->m_len);
+		WARN_ON(1);
+	}
+
+	status = map->m_flags & EXT4_MAP_UNWRITTEN ?
+			EXTENT_STATUS_UNWRITTEN : EXTENT_STATUS_WRITTEN;
+	ext4_es_insert_extent(inode, map->m_lblk, map->m_len,
+			      map->m_pblk, status);
+	return retval;
+}
+
 /*
  * The ext4_map_blocks() function tries to look up the requested blocks,
  * and returns if the blocks are already mapped.
@@ -465,9 +494,10 @@ static void ext4_map_blocks_es_recheck(handle_t *handle,
  * Otherwise, call with ext4_ind_map_blocks() to handle indirect mapping
  * based files
  *
- * On success, it returns the number of blocks being mapped or allocated.  if
- * create==0 and the blocks are pre-allocated and unwritten, the resulting @map
- * is marked as unwritten. If the create == 1, it will mark @map as mapped.
+ * On success, it returns the number of blocks being mapped or allocated.
+ * If flags doesn't contain EXT4_GET_BLOCKS_CREATE the blocks are
+ * pre-allocated and unwritten, the resulting @map is marked as unwritten.
+ * If the flags contain EXT4_GET_BLOCKS_CREATE, it will mark @map as mapped.
  *
  * It returns 0 if plain look up failed (blocks have not been allocated), in
  * that case, @map is returned as unmapped but we still do fill map->m_len to
@@ -589,8 +619,7 @@ found:
 	 * Returns if the blocks have already allocated
 	 *
 	 * Note that if blocks have been preallocated
-	 * ext4_ext_get_block() returns the create = 0
-	 * with buffer head unmapped.
+	 * ext4_ext_map_blocks() returns with buffer head unmapped
 	 */
 	if (retval > 0 && map->m_flags & EXT4_MAP_MAPPED)
 		/*
@@ -1708,6 +1737,7 @@ static int ext4_da_map_blocks(struct inode *inode, sector_t iblock,
 		if (ext4_es_is_hole(&es))
 			goto add_delayed;
 
+found:
 		/*
 		 * Delayed extent could be allocated by fallocate.
 		 * So we need to check it.
@@ -1744,36 +1774,34 @@ static int ext4_da_map_blocks(struct inode *inode, sector_t iblock,
 	down_read(&EXT4_I(inode)->i_data_sem);
 	if (ext4_has_inline_data(inode))
 		retval = 0;
-	else if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
-		retval = ext4_ext_map_blocks(NULL, inode, map, 0);
 	else
-		retval = ext4_ind_map_blocks(NULL, inode, map, 0);
-	if (retval < 0) {
-		up_read(&EXT4_I(inode)->i_data_sem);
+		retval = ext4_map_query_blocks(NULL, inode, map);
+	up_read(&EXT4_I(inode)->i_data_sem);
+	if (retval)
 		return retval;
-	}
-	if (retval > 0) {
-		unsigned int status;
 
-		if (unlikely(retval != map->m_len)) {
-			ext4_warning(inode->i_sb,
-				     "ES len assertion failed for inode "
-				     "%lu: retval %d != map->m_len %d",
-				     inode->i_ino, retval, map->m_len);
-			WARN_ON(1);
+add_delayed:
+	down_write(&EXT4_I(inode)->i_data_sem);
+	/*
+	 * Page fault path (ext4_page_mkwrite does not take i_rwsem)
+	 * and fallocate path (no folio lock) can race. Make sure we
+	 * lookup the extent status tree here again while i_data_sem
+	 * is held in write mode, before inserting a new da entry in
+	 * the extent status tree.
+	 */
+	if (ext4_es_lookup_extent(inode, iblock, NULL, &es)) {
+		if (!ext4_es_is_hole(&es)) {
+			up_write(&EXT4_I(inode)->i_data_sem);
+			goto found;
+		}
+	} else if (!ext4_has_inline_data(inode)) {
+		retval = ext4_map_query_blocks(NULL, inode, map);
+		if (retval) {
+			up_write(&EXT4_I(inode)->i_data_sem);
+			return retval;
 		}
-
-		status = map->m_flags & EXT4_MAP_UNWRITTEN ?
-				EXTENT_STATUS_UNWRITTEN : EXTENT_STATUS_WRITTEN;
-		ext4_es_insert_extent(inode, map->m_lblk, map->m_len,
-				      map->m_pblk, status);
-		up_read(&EXT4_I(inode)->i_data_sem);
-		return retval;
 	}
-	up_read(&EXT4_I(inode)->i_data_sem);
 
-add_delayed:
-	down_write(&EXT4_I(inode)->i_data_sem);
 	retval = ext4_insert_delayed_block(inode, map->m_lblk);
 	up_write(&EXT4_I(inode)->i_data_sem);
 	if (retval)
@@ -1865,7 +1893,7 @@ static int mpage_submit_folio(struct mpage_da_data *mpd, struct folio *folio)
 	len = folio_size(folio);
 	if (folio_pos(folio) + len > size &&
 	    !ext4_verity_in_progress(mpd->inode))
-		len = size & ~PAGE_MASK;
+		len = size & (len - 1);
 	err = ext4_bio_write_folio(&mpd->io_submit, folio, len);
 	if (!err)
 		mpd->wbc->nr_to_write--;
@@ -2334,7 +2362,7 @@ static int mpage_journal_page_buffers(handle_t *handle,
 
 	if (folio_pos(folio) + len > size &&
 	    !ext4_verity_in_progress(inode))
-		len = size - folio_pos(folio);
+		len = size & (len - 1);
 
 	return ext4_journal_folio_buffers(handle, folio, len);
 }
@@ -2945,6 +2973,11 @@ static int ext4_da_do_write_end(struct address_space *mapping,
 	bool disksize_changed = false;
 	loff_t new_i_size;
 
+	if (unlikely(!folio_buffers(folio))) {
+		folio_unlock(folio);
+		folio_put(folio);
+		return -EIO;
+	}
 	/*
 	 * block_write_end() will mark the inode as dirty with I_DIRTY_PAGES
 	 * flag, which all that's needed to trigger page writeback.
@@ -3527,7 +3560,6 @@ static const struct address_space_operations ext4_aops = {
 	.bmap			= ext4_bmap,
 	.invalidate_folio	= ext4_invalidate_folio,
 	.release_folio		= ext4_release_folio,
-	.direct_IO		= noop_direct_IO,
 	.migrate_folio		= buffer_migrate_folio,
 	.is_partially_uptodate  = block_is_partially_uptodate,
 	.error_remove_folio	= generic_error_remove_folio,
@@ -3544,7 +3576,6 @@ static const struct address_space_operations ext4_journalled_aops = {
 	.bmap			= ext4_bmap,
 	.invalidate_folio	= ext4_journalled_invalidate_folio,
 	.release_folio		= ext4_release_folio,
-	.direct_IO		= noop_direct_IO,
 	.migrate_folio		= buffer_migrate_folio_norefs,
 	.is_partially_uptodate  = block_is_partially_uptodate,
 	.error_remove_folio	= generic_error_remove_folio,
@@ -3561,7 +3592,6 @@ static const struct address_space_operations ext4_da_aops = {
 	.bmap			= ext4_bmap,
 	.invalidate_folio	= ext4_invalidate_folio,
 	.release_folio		= ext4_release_folio,
-	.direct_IO		= noop_direct_IO,
 	.migrate_folio		= buffer_migrate_folio,
 	.is_partially_uptodate  = block_is_partially_uptodate,
 	.error_remove_folio	= generic_error_remove_folio,
@@ -3570,7 +3600,6 @@ static const struct address_space_operations ext4_da_aops = {
 
 static const struct address_space_operations ext4_dax_aops = {
 	.writepages		= ext4_dax_writepages,
-	.direct_IO		= noop_direct_IO,
 	.dirty_folio		= noop_dirty_folio,
 	.bmap			= ext4_bmap,
 	.swap_activate		= ext4_iomap_swap_activate,
diff --git a/fs/ext4/ioctl.c b/fs/ext4/ioctl.c
index 7160a71044..e8bf5972dd 100644
--- a/fs/ext4/ioctl.c
+++ b/fs/ext4/ioctl.c
@@ -1150,9 +1150,8 @@ static int ext4_ioctl_getlabel(struct ext4_sb_info *sbi, char __user *user_label
 	 */
 	BUILD_BUG_ON(EXT4_LABEL_MAX >= FSLABEL_MAX);
 
-	memset(label, 0, sizeof(label));
 	lock_buffer(sbi->s_sbh);
-	strncpy(label, sbi->s_es->s_volume_name, EXT4_LABEL_MAX);
+	memtostr_pad(label, sbi->s_es->s_volume_name);
 	unlock_buffer(sbi->s_sbh);
 
 	if (copy_to_user(user_label, label, sizeof(label)))
diff --git a/fs/ext4/mballoc-test.c b/fs/ext4/mballoc-test.c
index f94901fd38..bb2a223b20 100644
--- a/fs/ext4/mballoc-test.c
+++ b/fs/ext4/mballoc-test.c
@@ -5,6 +5,7 @@
 
 #include <kunit/test.h>
 #include <kunit/static_stub.h>
+#include <linux/random.h>
 
 #include "ext4.h"
 
@@ -20,41 +21,159 @@ struct mbt_ctx {
 };
 
 struct mbt_ext4_super_block {
-	struct super_block sb;
+	struct ext4_super_block es;
+	struct ext4_sb_info sbi;
 	struct mbt_ctx mbt_ctx;
 };
 
-#define MBT_CTX(_sb) (&(container_of((_sb), struct mbt_ext4_super_block, sb)->mbt_ctx))
+#define MBT_SB(_sb) (container_of((_sb)->s_fs_info, struct mbt_ext4_super_block, sbi))
+#define MBT_CTX(_sb) (&MBT_SB(_sb)->mbt_ctx)
 #define MBT_GRP_CTX(_sb, _group) (&MBT_CTX(_sb)->grp_ctx[_group])
 
+static struct inode *mbt_alloc_inode(struct super_block *sb)
+{
+	struct ext4_inode_info *ei;
+
+	ei = kmalloc(sizeof(struct ext4_inode_info), GFP_KERNEL);
+	if (!ei)
+		return NULL;
+
+	INIT_LIST_HEAD(&ei->i_orphan);
+	init_rwsem(&ei->xattr_sem);
+	init_rwsem(&ei->i_data_sem);
+	inode_init_once(&ei->vfs_inode);
+	ext4_fc_init_inode(&ei->vfs_inode);
+
+	return &ei->vfs_inode;
+}
+
+static void mbt_free_inode(struct inode *inode)
+{
+	kfree(EXT4_I(inode));
+}
+
+static const struct super_operations mbt_sops = {
+	.alloc_inode	= mbt_alloc_inode,
+	.free_inode	= mbt_free_inode,
+};
+
+static void mbt_kill_sb(struct super_block *sb)
+{
+	generic_shutdown_super(sb);
+}
+
+static struct file_system_type mbt_fs_type = {
+	.name			= "mballoc test",
+	.kill_sb		= mbt_kill_sb,
+};
+
+static int mbt_mb_init(struct super_block *sb)
+{
+	ext4_fsblk_t block;
+	int ret;
+
+	/* needed by ext4_mb_init->bdev_nonrot(sb->s_bdev) */
+	sb->s_bdev = kzalloc(sizeof(*sb->s_bdev), GFP_KERNEL);
+	if (sb->s_bdev == NULL)
+		return -ENOMEM;
+
+	sb->s_bdev->bd_queue = kzalloc(sizeof(struct request_queue), GFP_KERNEL);
+	if (sb->s_bdev->bd_queue == NULL) {
+		kfree(sb->s_bdev);
+		return -ENOMEM;
+	}
+
+	/*
+	 * needed by ext4_mb_init->ext4_mb_init_backend-> sbi->s_buddy_cache =
+	 * new_inode(sb);
+	 */
+	INIT_LIST_HEAD(&sb->s_inodes);
+	sb->s_op = &mbt_sops;
+
+	ret = ext4_mb_init(sb);
+	if (ret != 0)
+		goto err_out;
+
+	block = ext4_count_free_clusters(sb);
+	ret = percpu_counter_init(&EXT4_SB(sb)->s_freeclusters_counter, block,
+				  GFP_KERNEL);
+	if (ret != 0)
+		goto err_mb_release;
+
+	ret = percpu_counter_init(&EXT4_SB(sb)->s_dirtyclusters_counter, 0,
+				  GFP_KERNEL);
+	if (ret != 0)
+		goto err_freeclusters;
+
+	return 0;
+
+err_freeclusters:
+	percpu_counter_destroy(&EXT4_SB(sb)->s_freeclusters_counter);
+err_mb_release:
+	ext4_mb_release(sb);
+err_out:
+	kfree(sb->s_bdev->bd_queue);
+	kfree(sb->s_bdev);
+	return ret;
+}
+
+static void mbt_mb_release(struct super_block *sb)
+{
+	percpu_counter_destroy(&EXT4_SB(sb)->s_dirtyclusters_counter);
+	percpu_counter_destroy(&EXT4_SB(sb)->s_freeclusters_counter);
+	ext4_mb_release(sb);
+	kfree(sb->s_bdev->bd_queue);
+	kfree(sb->s_bdev);
+}
+
+static int mbt_set(struct super_block *sb, void *data)
+{
+	return 0;
+}
+
 static struct super_block *mbt_ext4_alloc_super_block(void)
 {
-	struct ext4_super_block *es = kzalloc(sizeof(*es), GFP_KERNEL);
-	struct ext4_sb_info *sbi = kzalloc(sizeof(*sbi), GFP_KERNEL);
-	struct mbt_ext4_super_block *fsb = kzalloc(sizeof(*fsb), GFP_KERNEL);
+	struct mbt_ext4_super_block *fsb;
+	struct super_block *sb;
+	struct ext4_sb_info *sbi;
 
-	if (fsb == NULL || sbi == NULL || es == NULL)
+	fsb = kzalloc(sizeof(*fsb), GFP_KERNEL);
+	if (fsb == NULL)
+		return NULL;
+
+	sb = sget(&mbt_fs_type, NULL, mbt_set, 0, NULL);
+	if (IS_ERR(sb))
 		goto out;
 
-	sbi->s_es = es;
-	fsb->sb.s_fs_info = sbi;
-	return &fsb->sb;
+	sbi = &fsb->sbi;
+
+	sbi->s_blockgroup_lock =
+		kzalloc(sizeof(struct blockgroup_lock), GFP_KERNEL);
+	if (!sbi->s_blockgroup_lock)
+		goto out_deactivate;
+
+	bgl_lock_init(sbi->s_blockgroup_lock);
 
+	sbi->s_es = &fsb->es;
+	sb->s_fs_info = sbi;
+
+	up_write(&sb->s_umount);
+	return sb;
+
+out_deactivate:
+	deactivate_locked_super(sb);
 out:
 	kfree(fsb);
-	kfree(sbi);
-	kfree(es);
 	return NULL;
 }
 
 static void mbt_ext4_free_super_block(struct super_block *sb)
 {
-	struct mbt_ext4_super_block *fsb =
-		container_of(sb, struct mbt_ext4_super_block, sb);
+	struct mbt_ext4_super_block *fsb = MBT_SB(sb);
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
 
-	kfree(sbi->s_es);
-	kfree(sbi);
+	kfree(sbi->s_blockgroup_lock);
+	deactivate_super(sb);
 	kfree(fsb);
 }
 
@@ -82,6 +201,9 @@ static void mbt_init_sb_layout(struct super_block *sb,
 	sbi->s_clusters_per_group = layout->blocks_per_group >>
 				    layout->cluster_bits;
 	sbi->s_desc_size = layout->desc_size;
+	sbi->s_desc_per_block_bits =
+		sb->s_blocksize_bits - (fls(layout->desc_size) - 1);
+	sbi->s_desc_per_block = 1 << sbi->s_desc_per_block_bits;
 
 	es->s_first_data_block = cpu_to_le32(0);
 	es->s_blocks_count_lo = cpu_to_le32(layout->blocks_per_group *
@@ -91,9 +213,13 @@ static void mbt_init_sb_layout(struct super_block *sb,
 static int mbt_grp_ctx_init(struct super_block *sb,
 			    struct mbt_grp_ctx *grp_ctx)
 {
+	ext4_grpblk_t max = EXT4_CLUSTERS_PER_GROUP(sb);
+
 	grp_ctx->bitmap_bh.b_data = kzalloc(EXT4_BLOCK_SIZE(sb), GFP_KERNEL);
 	if (grp_ctx->bitmap_bh.b_data == NULL)
 		return -ENOMEM;
+	mb_set_bits(grp_ctx->bitmap_bh.b_data, max, sb->s_blocksize * 8 - max);
+	ext4_free_group_clusters_set(sb, &grp_ctx->desc, max);
 
 	return 0;
 }
@@ -112,6 +238,13 @@ static void mbt_ctx_mark_used(struct super_block *sb, ext4_group_t group,
 	mb_set_bits(grp_ctx->bitmap_bh.b_data, start, len);
 }
 
+static void *mbt_ctx_bitmap(struct super_block *sb, ext4_group_t group)
+{
+	struct mbt_grp_ctx *grp_ctx = MBT_GRP_CTX(sb, group);
+
+	return grp_ctx->bitmap_bh.b_data;
+}
+
 /* called after mbt_init_sb_layout */
 static int mbt_ctx_init(struct super_block *sb)
 {
@@ -133,6 +266,8 @@ static int mbt_ctx_init(struct super_block *sb)
 	 * block which will fail ext4_sb_block_valid check.
 	 */
 	mb_set_bits(ctx->grp_ctx[0].bitmap_bh.b_data, 0, 1);
+	ext4_free_group_clusters_set(sb, &ctx->grp_ctx[0].desc,
+				     EXT4_CLUSTERS_PER_GROUP(sb) - 1);
 
 	return 0;
 out:
@@ -167,6 +302,13 @@ static int ext4_wait_block_bitmap_stub(struct super_block *sb,
 				       ext4_group_t block_group,
 				       struct buffer_head *bh)
 {
+	/*
+	 * real ext4_wait_block_bitmap will set these flags and
+	 * functions like ext4_mb_init_cache will verify the flags.
+	 */
+	set_buffer_uptodate(bh);
+	set_bitmap_uptodate(bh);
+	set_buffer_verified(bh);
 	return 0;
 }
 
@@ -232,6 +374,14 @@ static int mbt_kunit_init(struct kunit *test)
 	kunit_activate_static_stub(test,
 				   ext4_mb_mark_context,
 				   ext4_mb_mark_context_stub);
+
+	/* stub function will be called in mbt_mb_init->ext4_mb_init */
+	if (mbt_mb_init(sb) != 0) {
+		mbt_ctx_release(sb);
+		mbt_ext4_free_super_block(sb);
+		return -ENOMEM;
+	}
+
 	return 0;
 }
 
@@ -239,6 +389,7 @@ static void mbt_kunit_exit(struct kunit *test)
 {
 	struct super_block *sb = (struct super_block *)test->priv;
 
+	mbt_mb_release(sb);
 	mbt_ctx_release(sb);
 	mbt_ext4_free_super_block(sb);
 }
@@ -246,14 +397,19 @@ static void mbt_kunit_exit(struct kunit *test)
 static void test_new_blocks_simple(struct kunit *test)
 {
 	struct super_block *sb = (struct super_block *)test->priv;
-	struct inode inode = { .i_sb = sb, };
+	struct inode *inode;
 	struct ext4_allocation_request ar;
 	ext4_group_t i, goal_group = TEST_GOAL_GROUP;
 	int err = 0;
 	ext4_fsblk_t found;
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
 
-	ar.inode = &inode;
+	inode = kunit_kzalloc(test, sizeof(*inode), GFP_KERNEL);
+	if (!inode)
+		return;
+
+	inode->i_sb = sb;
+	ar.inode = inode;
 
 	/* get block at goal */
 	ar.goal = ext4_group_first_block_no(sb, goal_group);
@@ -297,6 +453,486 @@ static void test_new_blocks_simple(struct kunit *test)
 		"unexpectedly get block when no block is available");
 }
 
+#define TEST_RANGE_COUNT 8
+
+struct test_range {
+	ext4_grpblk_t start;
+	ext4_grpblk_t len;
+};
+
+static void
+mbt_generate_test_ranges(struct super_block *sb, struct test_range *ranges,
+			 int count)
+{
+	ext4_grpblk_t start, len, max;
+	int i;
+
+	max = EXT4_CLUSTERS_PER_GROUP(sb) / count;
+	for (i = 0; i < count; i++) {
+		start = get_random_u32() % max;
+		len = get_random_u32() % max;
+		len = min(len, max - start);
+
+		ranges[i].start = start + i * max;
+		ranges[i].len = len;
+	}
+}
+
+static void
+validate_free_blocks_simple(struct kunit *test, struct super_block *sb,
+			    ext4_group_t goal_group, ext4_grpblk_t start,
+			    ext4_grpblk_t len)
+{
+	void *bitmap;
+	ext4_grpblk_t bit, max = EXT4_CLUSTERS_PER_GROUP(sb);
+	ext4_group_t i;
+
+	for (i = 0; i < ext4_get_groups_count(sb); i++) {
+		if (i == goal_group)
+			continue;
+
+		bitmap = mbt_ctx_bitmap(sb, i);
+		bit = mb_find_next_zero_bit(bitmap, max, 0);
+		KUNIT_ASSERT_EQ_MSG(test, bit, max,
+				    "free block on unexpected group %d", i);
+	}
+
+	bitmap = mbt_ctx_bitmap(sb, goal_group);
+	bit = mb_find_next_zero_bit(bitmap, max, 0);
+	KUNIT_ASSERT_EQ(test, bit, start);
+
+	bit = mb_find_next_bit(bitmap, max, bit + 1);
+	KUNIT_ASSERT_EQ(test, bit, start + len);
+}
+
+static void
+test_free_blocks_simple_range(struct kunit *test, ext4_group_t goal_group,
+			      ext4_grpblk_t start, ext4_grpblk_t len)
+{
+	struct super_block *sb = (struct super_block *)test->priv;
+	struct ext4_sb_info *sbi = EXT4_SB(sb);
+	struct inode *inode;
+	ext4_fsblk_t block;
+
+	inode = kunit_kzalloc(test, sizeof(*inode), GFP_KERNEL);
+	if (!inode)
+		return;
+	inode->i_sb = sb;
+
+	if (len == 0)
+		return;
+
+	block = ext4_group_first_block_no(sb, goal_group) +
+		EXT4_C2B(sbi, start);
+	ext4_free_blocks_simple(inode, block, len);
+	validate_free_blocks_simple(test, sb, goal_group, start, len);
+	mbt_ctx_mark_used(sb, goal_group, 0, EXT4_CLUSTERS_PER_GROUP(sb));
+}
+
+static void test_free_blocks_simple(struct kunit *test)
+{
+	struct super_block *sb = (struct super_block *)test->priv;
+	ext4_grpblk_t max = EXT4_CLUSTERS_PER_GROUP(sb);
+	ext4_group_t i;
+	struct test_range ranges[TEST_RANGE_COUNT];
+
+	for (i = 0; i < ext4_get_groups_count(sb); i++)
+		mbt_ctx_mark_used(sb, i, 0, max);
+
+	mbt_generate_test_ranges(sb, ranges, TEST_RANGE_COUNT);
+	for (i = 0; i < TEST_RANGE_COUNT; i++)
+		test_free_blocks_simple_range(test, TEST_GOAL_GROUP,
+			ranges[i].start, ranges[i].len);
+}
+
+static void
+test_mark_diskspace_used_range(struct kunit *test,
+			       struct ext4_allocation_context *ac,
+			       ext4_grpblk_t start,
+			       ext4_grpblk_t len)
+{
+	struct super_block *sb = (struct super_block *)test->priv;
+	int ret;
+	void *bitmap;
+	ext4_grpblk_t i, max;
+
+	/* ext4_mb_mark_diskspace_used will BUG if len is 0 */
+	if (len == 0)
+		return;
+
+	ac->ac_b_ex.fe_group = TEST_GOAL_GROUP;
+	ac->ac_b_ex.fe_start = start;
+	ac->ac_b_ex.fe_len = len;
+
+	bitmap = mbt_ctx_bitmap(sb, TEST_GOAL_GROUP);
+	memset(bitmap, 0, sb->s_blocksize);
+	ret = ext4_mb_mark_diskspace_used(ac, NULL, 0);
+	KUNIT_ASSERT_EQ(test, ret, 0);
+
+	max = EXT4_CLUSTERS_PER_GROUP(sb);
+	i = mb_find_next_bit(bitmap, max, 0);
+	KUNIT_ASSERT_EQ(test, i, start);
+	i = mb_find_next_zero_bit(bitmap, max, i + 1);
+	KUNIT_ASSERT_EQ(test, i, start + len);
+	i = mb_find_next_bit(bitmap, max, i + 1);
+	KUNIT_ASSERT_EQ(test, max, i);
+}
+
+static void test_mark_diskspace_used(struct kunit *test)
+{
+	struct super_block *sb = (struct super_block *)test->priv;
+	struct inode *inode;
+	struct ext4_allocation_context ac;
+	struct test_range ranges[TEST_RANGE_COUNT];
+	int i;
+
+	mbt_generate_test_ranges(sb, ranges, TEST_RANGE_COUNT);
+
+	inode = kunit_kzalloc(test, sizeof(*inode), GFP_KERNEL);
+	if (!inode)
+		return;
+	inode->i_sb = sb;
+
+	ac.ac_status = AC_STATUS_FOUND;
+	ac.ac_sb = sb;
+	ac.ac_inode = inode;
+	for (i = 0; i < TEST_RANGE_COUNT; i++)
+		test_mark_diskspace_used_range(test, &ac, ranges[i].start,
+					       ranges[i].len);
+}
+
+static void mbt_generate_buddy(struct super_block *sb, void *buddy,
+			       void *bitmap, struct ext4_group_info *grp)
+{
+	struct ext4_sb_info *sbi = EXT4_SB(sb);
+	uint32_t order, off;
+	void *bb, *bb_h;
+	int max;
+
+	memset(buddy, 0xff, sb->s_blocksize);
+	memset(grp, 0, offsetof(struct ext4_group_info,
+				 bb_counters[MB_NUM_ORDERS(sb)]));
+
+	bb = bitmap;
+	max = EXT4_CLUSTERS_PER_GROUP(sb);
+	bb_h = buddy + sbi->s_mb_offsets[1];
+
+	off = mb_find_next_zero_bit(bb, max, 0);
+	grp->bb_first_free = off;
+	while (off < max) {
+		grp->bb_counters[0]++;
+		grp->bb_free++;
+
+		if (!(off & 1) && !mb_test_bit(off + 1, bb)) {
+			grp->bb_free++;
+			grp->bb_counters[0]--;
+			mb_clear_bit(off >> 1, bb_h);
+			grp->bb_counters[1]++;
+			grp->bb_largest_free_order = 1;
+			off++;
+		}
+
+		off = mb_find_next_zero_bit(bb, max, off + 1);
+	}
+
+	for (order = 1; order < MB_NUM_ORDERS(sb) - 1; order++) {
+		bb = buddy + sbi->s_mb_offsets[order];
+		bb_h = buddy + sbi->s_mb_offsets[order + 1];
+		max = max >> 1;
+		off = mb_find_next_zero_bit(bb, max, 0);
+
+		while (off < max) {
+			if (!(off & 1) && !mb_test_bit(off + 1, bb)) {
+				mb_set_bits(bb, off, 2);
+				grp->bb_counters[order] -= 2;
+				mb_clear_bit(off >> 1, bb_h);
+				grp->bb_counters[order + 1]++;
+				grp->bb_largest_free_order = order + 1;
+				off++;
+			}
+
+			off = mb_find_next_zero_bit(bb, max, off + 1);
+		}
+	}
+
+	max = EXT4_CLUSTERS_PER_GROUP(sb);
+	off = mb_find_next_zero_bit(bitmap, max, 0);
+	while (off < max) {
+		grp->bb_fragments++;
+
+		off = mb_find_next_bit(bitmap, max, off + 1);
+		if (off + 1 >= max)
+			break;
+
+		off = mb_find_next_zero_bit(bitmap, max, off + 1);
+	}
+}
+
+static void
+mbt_validate_group_info(struct kunit *test, struct ext4_group_info *grp1,
+			struct ext4_group_info *grp2)
+{
+	struct super_block *sb = (struct super_block *)test->priv;
+	int i;
+
+	KUNIT_ASSERT_EQ(test, grp1->bb_first_free,
+			grp2->bb_first_free);
+	KUNIT_ASSERT_EQ(test, grp1->bb_fragments,
+			grp2->bb_fragments);
+	KUNIT_ASSERT_EQ(test, grp1->bb_free, grp2->bb_free);
+	KUNIT_ASSERT_EQ(test, grp1->bb_largest_free_order,
+			grp2->bb_largest_free_order);
+
+	for (i = 1; i < MB_NUM_ORDERS(sb); i++) {
+		KUNIT_ASSERT_EQ_MSG(test, grp1->bb_counters[i],
+				    grp2->bb_counters[i],
+				    "bb_counters[%d] diffs, expected %d, generated %d",
+				    i, grp1->bb_counters[i],
+				    grp2->bb_counters[i]);
+	}
+}
+
+static void
+do_test_generate_buddy(struct kunit *test, struct super_block *sb, void *bitmap,
+			   void *mbt_buddy, struct ext4_group_info *mbt_grp,
+			   void *ext4_buddy, struct ext4_group_info *ext4_grp)
+{
+	int i;
+
+	mbt_generate_buddy(sb, mbt_buddy, bitmap, mbt_grp);
+
+	for (i = 0; i < MB_NUM_ORDERS(sb); i++)
+		ext4_grp->bb_counters[i] = 0;
+	/* needed by validation in ext4_mb_generate_buddy */
+	ext4_grp->bb_free = mbt_grp->bb_free;
+	memset(ext4_buddy, 0xff, sb->s_blocksize);
+	ext4_mb_generate_buddy(sb, ext4_buddy, bitmap, TEST_GOAL_GROUP,
+			       ext4_grp);
+
+	KUNIT_ASSERT_EQ(test, memcmp(mbt_buddy, ext4_buddy, sb->s_blocksize),
+			0);
+	mbt_validate_group_info(test, mbt_grp, ext4_grp);
+}
+
+static void test_mb_generate_buddy(struct kunit *test)
+{
+	struct super_block *sb = (struct super_block *)test->priv;
+	void *bitmap, *expected_bb, *generate_bb;
+	struct ext4_group_info *expected_grp, *generate_grp;
+	struct test_range ranges[TEST_RANGE_COUNT];
+	int i;
+
+	bitmap = kunit_kzalloc(test, sb->s_blocksize, GFP_KERNEL);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, bitmap);
+	expected_bb = kunit_kzalloc(test, sb->s_blocksize, GFP_KERNEL);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, expected_bb);
+	generate_bb = kunit_kzalloc(test, sb->s_blocksize, GFP_KERNEL);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, generate_bb);
+	expected_grp = kunit_kzalloc(test, offsetof(struct ext4_group_info,
+				bb_counters[MB_NUM_ORDERS(sb)]), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, expected_grp);
+	generate_grp = ext4_get_group_info(sb, TEST_GOAL_GROUP);
+	KUNIT_ASSERT_NOT_NULL(test, generate_grp);
+
+	mbt_generate_test_ranges(sb, ranges, TEST_RANGE_COUNT);
+	for (i = 0; i < TEST_RANGE_COUNT; i++) {
+		mb_set_bits(bitmap, ranges[i].start, ranges[i].len);
+		do_test_generate_buddy(test, sb, bitmap, expected_bb,
+				       expected_grp, generate_bb, generate_grp);
+	}
+}
+
+static void
+test_mb_mark_used_range(struct kunit *test, struct ext4_buddy *e4b,
+			ext4_grpblk_t start, ext4_grpblk_t len, void *bitmap,
+			void *buddy, struct ext4_group_info *grp)
+{
+	struct super_block *sb = (struct super_block *)test->priv;
+	struct ext4_free_extent ex;
+	int i;
+
+	/* mb_mark_used only accepts non-zero len */
+	if (len == 0)
+		return;
+
+	ex.fe_start = start;
+	ex.fe_len = len;
+	ex.fe_group = TEST_GOAL_GROUP;
+
+	ext4_lock_group(sb, TEST_GOAL_GROUP);
+	mb_mark_used(e4b, &ex);
+	ext4_unlock_group(sb, TEST_GOAL_GROUP);
+
+	mb_set_bits(bitmap, start, len);
+	/* bypass bb_free validatoin in ext4_mb_generate_buddy */
+	grp->bb_free -= len;
+	memset(buddy, 0xff, sb->s_blocksize);
+	for (i = 0; i < MB_NUM_ORDERS(sb); i++)
+		grp->bb_counters[i] = 0;
+	ext4_mb_generate_buddy(sb, buddy, bitmap, 0, grp);
+
+	KUNIT_ASSERT_EQ(test, memcmp(buddy, e4b->bd_buddy, sb->s_blocksize),
+			0);
+	mbt_validate_group_info(test, grp, e4b->bd_info);
+}
+
+static void test_mb_mark_used(struct kunit *test)
+{
+	struct ext4_buddy e4b;
+	struct super_block *sb = (struct super_block *)test->priv;
+	void *bitmap, *buddy;
+	struct ext4_group_info *grp;
+	int ret;
+	struct test_range ranges[TEST_RANGE_COUNT];
+	int i;
+
+	/* buddy cache assumes that each page contains at least one block */
+	if (sb->s_blocksize > PAGE_SIZE)
+		kunit_skip(test, "blocksize exceeds pagesize");
+
+	bitmap = kunit_kzalloc(test, sb->s_blocksize, GFP_KERNEL);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, bitmap);
+	buddy = kunit_kzalloc(test, sb->s_blocksize, GFP_KERNEL);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, buddy);
+	grp = kunit_kzalloc(test, offsetof(struct ext4_group_info,
+				bb_counters[MB_NUM_ORDERS(sb)]), GFP_KERNEL);
+
+	ret = ext4_mb_load_buddy(sb, TEST_GOAL_GROUP, &e4b);
+	KUNIT_ASSERT_EQ(test, ret, 0);
+
+	grp->bb_free = EXT4_CLUSTERS_PER_GROUP(sb);
+	mbt_generate_test_ranges(sb, ranges, TEST_RANGE_COUNT);
+	for (i = 0; i < TEST_RANGE_COUNT; i++)
+		test_mb_mark_used_range(test, &e4b, ranges[i].start,
+					ranges[i].len, bitmap, buddy, grp);
+
+	ext4_mb_unload_buddy(&e4b);
+}
+
+static void
+test_mb_free_blocks_range(struct kunit *test, struct ext4_buddy *e4b,
+			  ext4_grpblk_t start, ext4_grpblk_t len, void *bitmap,
+			  void *buddy, struct ext4_group_info *grp)
+{
+	struct super_block *sb = (struct super_block *)test->priv;
+	int i;
+
+	/* mb_free_blocks will WARN if len is 0 */
+	if (len == 0)
+		return;
+
+	ext4_lock_group(sb, e4b->bd_group);
+	mb_free_blocks(NULL, e4b, start, len);
+	ext4_unlock_group(sb, e4b->bd_group);
+
+	mb_clear_bits(bitmap, start, len);
+	/* bypass bb_free validatoin in ext4_mb_generate_buddy */
+	grp->bb_free += len;
+	memset(buddy, 0xff, sb->s_blocksize);
+	for (i = 0; i < MB_NUM_ORDERS(sb); i++)
+		grp->bb_counters[i] = 0;
+	ext4_mb_generate_buddy(sb, buddy, bitmap, 0, grp);
+
+	KUNIT_ASSERT_EQ(test, memcmp(buddy, e4b->bd_buddy, sb->s_blocksize),
+			0);
+	mbt_validate_group_info(test, grp, e4b->bd_info);
+
+}
+
+static void test_mb_free_blocks(struct kunit *test)
+{
+	struct ext4_buddy e4b;
+	struct super_block *sb = (struct super_block *)test->priv;
+	void *bitmap, *buddy;
+	struct ext4_group_info *grp;
+	struct ext4_free_extent ex;
+	int ret;
+	int i;
+	struct test_range ranges[TEST_RANGE_COUNT];
+
+	/* buddy cache assumes that each page contains at least one block */
+	if (sb->s_blocksize > PAGE_SIZE)
+		kunit_skip(test, "blocksize exceeds pagesize");
+
+	bitmap = kunit_kzalloc(test, sb->s_blocksize, GFP_KERNEL);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, bitmap);
+	buddy = kunit_kzalloc(test, sb->s_blocksize, GFP_KERNEL);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, buddy);
+	grp = kunit_kzalloc(test, offsetof(struct ext4_group_info,
+				bb_counters[MB_NUM_ORDERS(sb)]), GFP_KERNEL);
+
+	ret = ext4_mb_load_buddy(sb, TEST_GOAL_GROUP, &e4b);
+	KUNIT_ASSERT_EQ(test, ret, 0);
+
+	ex.fe_start = 0;
+	ex.fe_len = EXT4_CLUSTERS_PER_GROUP(sb);
+	ex.fe_group = TEST_GOAL_GROUP;
+
+	ext4_lock_group(sb, TEST_GOAL_GROUP);
+	mb_mark_used(&e4b, &ex);
+	ext4_unlock_group(sb, TEST_GOAL_GROUP);
+
+	grp->bb_free = 0;
+	memset(bitmap, 0xff, sb->s_blocksize);
+
+	mbt_generate_test_ranges(sb, ranges, TEST_RANGE_COUNT);
+	for (i = 0; i < TEST_RANGE_COUNT; i++)
+		test_mb_free_blocks_range(test, &e4b, ranges[i].start,
+					  ranges[i].len, bitmap, buddy, grp);
+
+	ext4_mb_unload_buddy(&e4b);
+}
+
+#define COUNT_FOR_ESTIMATE 100000
+static void test_mb_mark_used_cost(struct kunit *test)
+{
+	struct ext4_buddy e4b;
+	struct super_block *sb = (struct super_block *)test->priv;
+	struct ext4_free_extent ex;
+	int ret;
+	struct test_range ranges[TEST_RANGE_COUNT];
+	int i, j;
+	unsigned long start, end, all = 0;
+
+	/* buddy cache assumes that each page contains at least one block */
+	if (sb->s_blocksize > PAGE_SIZE)
+		kunit_skip(test, "blocksize exceeds pagesize");
+
+	ret = ext4_mb_load_buddy(sb, TEST_GOAL_GROUP, &e4b);
+	KUNIT_ASSERT_EQ(test, ret, 0);
+
+	ex.fe_group = TEST_GOAL_GROUP;
+	for (j = 0; j < COUNT_FOR_ESTIMATE; j++) {
+		mbt_generate_test_ranges(sb, ranges, TEST_RANGE_COUNT);
+		start = jiffies;
+		for (i = 0; i < TEST_RANGE_COUNT; i++) {
+			if (ranges[i].len == 0)
+				continue;
+
+			ex.fe_start = ranges[i].start;
+			ex.fe_len = ranges[i].len;
+			ext4_lock_group(sb, TEST_GOAL_GROUP);
+			mb_mark_used(&e4b, &ex);
+			ext4_unlock_group(sb, TEST_GOAL_GROUP);
+		}
+		end = jiffies;
+		all += (end - start);
+
+		for (i = 0; i < TEST_RANGE_COUNT; i++) {
+			if (ranges[i].len == 0)
+				continue;
+
+			ext4_lock_group(sb, TEST_GOAL_GROUP);
+			mb_free_blocks(NULL, &e4b, ranges[i].start,
+				       ranges[i].len);
+			ext4_unlock_group(sb, TEST_GOAL_GROUP);
+		}
+	}
+
+	kunit_info(test, "costed jiffies %lu\n", all);
+	ext4_mb_unload_buddy(&e4b);
+}
+
 static const struct mbt_ext4_block_layout mbt_test_layouts[] = {
 	{
 		.blocksize_bits = 10,
@@ -334,6 +970,13 @@ KUNIT_ARRAY_PARAM(mbt_layouts, mbt_test_layouts, mbt_show_layout);
 
 static struct kunit_case mbt_test_cases[] = {
 	KUNIT_CASE_PARAM(test_new_blocks_simple, mbt_layouts_gen_params),
+	KUNIT_CASE_PARAM(test_free_blocks_simple, mbt_layouts_gen_params),
+	KUNIT_CASE_PARAM(test_mb_generate_buddy, mbt_layouts_gen_params),
+	KUNIT_CASE_PARAM(test_mb_mark_used, mbt_layouts_gen_params),
+	KUNIT_CASE_PARAM(test_mb_free_blocks, mbt_layouts_gen_params),
+	KUNIT_CASE_PARAM(test_mark_diskspace_used, mbt_layouts_gen_params),
+	KUNIT_CASE_PARAM_ATTR(test_mb_mark_used_cost, mbt_layouts_gen_params,
+			      { .speed = KUNIT_SPEED_SLOW }),
 	{}
 };
 
diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c
index 00b0839b53..9dda9cd68a 100644
--- a/fs/ext4/mballoc.c
+++ b/fs/ext4/mballoc.c
@@ -831,6 +831,8 @@ static int mb_avg_fragment_size_order(struct super_block *sb, ext4_grpblk_t len)
 		return 0;
 	if (order == MB_NUM_ORDERS(sb))
 		order--;
+	if (WARN_ON_ONCE(order > MB_NUM_ORDERS(sb)))
+		order = MB_NUM_ORDERS(sb) - 1;
 	return order;
 }
 
@@ -1008,6 +1010,8 @@ static void ext4_mb_choose_next_group_best_avail(struct ext4_allocation_context
 	 * goal length.
 	 */
 	order = fls(ac->ac_g_ex.fe_len) - 1;
+	if (WARN_ON_ONCE(order - 1 > MB_NUM_ORDERS(ac->ac_sb)))
+		order = MB_NUM_ORDERS(ac->ac_sb);
 	min_order = order - sbi->s_mb_best_avail_max_trim_order;
 	if (min_order < 0)
 		min_order = 0;
@@ -1076,23 +1080,11 @@ static inline int should_optimize_scan(struct ext4_allocation_context *ac)
 }
 
 /*
- * Return next linear group for allocation. If linear traversal should not be
- * performed, this function just returns the same group
+ * Return next linear group for allocation.
  */
 static ext4_group_t
-next_linear_group(struct ext4_allocation_context *ac, ext4_group_t group,
-		  ext4_group_t ngroups)
+next_linear_group(ext4_group_t group, ext4_group_t ngroups)
 {
-	if (!should_optimize_scan(ac))
-		goto inc_and_return;
-
-	if (ac->ac_groups_linear_remaining) {
-		ac->ac_groups_linear_remaining--;
-		goto inc_and_return;
-	}
-
-	return group;
-inc_and_return:
 	/*
 	 * Artificially restricted ngroups for non-extent
 	 * files makes group > ngroups possible on first loop.
@@ -1118,8 +1110,19 @@ static void ext4_mb_choose_next_group(struct ext4_allocation_context *ac,
 {
 	*new_cr = ac->ac_criteria;
 
-	if (!should_optimize_scan(ac) || ac->ac_groups_linear_remaining) {
-		*group = next_linear_group(ac, *group, ngroups);
+	if (!should_optimize_scan(ac)) {
+		*group = next_linear_group(*group, ngroups);
+		return;
+	}
+
+	/*
+	 * Optimized scanning can return non adjacent groups which can cause
+	 * seek overhead for rotational disks. So try few linear groups before
+	 * trying optimized scan.
+	 */
+	if (ac->ac_groups_linear_remaining) {
+		*group = next_linear_group(*group, ngroups);
+		ac->ac_groups_linear_remaining--;
 		return;
 	}
 
@@ -1131,8 +1134,9 @@ static void ext4_mb_choose_next_group(struct ext4_allocation_context *ac,
 		ext4_mb_choose_next_group_best_avail(ac, new_cr, group);
 	} else {
 		/*
-		 * TODO: For CR=2, we can arrange groups in an rb tree sorted by
-		 * bb_free. But until that happens, we should never come here.
+		 * TODO: For CR_GOAL_LEN_SLOW, we can arrange groups in an
+		 * rb tree sorted by bb_free. But until that happens, we should
+		 * never come here.
 		 */
 		WARN_ON(1);
 	}
@@ -1270,7 +1274,7 @@ static void mb_regenerate_buddy(struct ext4_buddy *e4b)
  * for this page; do not hold this lock when calling this routine!
  */
 
-static int ext4_mb_init_cache(struct page *page, char *incore, gfp_t gfp)
+static int ext4_mb_init_cache(struct folio *folio, char *incore, gfp_t gfp)
 {
 	ext4_group_t ngroups;
 	unsigned int blocksize;
@@ -1288,13 +1292,13 @@ static int ext4_mb_init_cache(struct page *page, char *incore, gfp_t gfp)
 	char *bitmap;
 	struct ext4_group_info *grinfo;
 
-	inode = page->mapping->host;
+	inode = folio->mapping->host;
 	sb = inode->i_sb;
 	ngroups = ext4_get_groups_count(sb);
 	blocksize = i_blocksize(inode);
 	blocks_per_page = PAGE_SIZE / blocksize;
 
-	mb_debug(sb, "init page %lu\n", page->index);
+	mb_debug(sb, "init folio %lu\n", folio->index);
 
 	groups_per_page = blocks_per_page >> 1;
 	if (groups_per_page == 0)
@@ -1309,9 +1313,9 @@ static int ext4_mb_init_cache(struct page *page, char *incore, gfp_t gfp)
 	} else
 		bh = &bhs;
 
-	first_group = page->index * blocks_per_page / 2;
+	first_group = folio->index * blocks_per_page / 2;
 
-	/* read all groups the page covers into the cache */
+	/* read all groups the folio covers into the cache */
 	for (i = 0, group = first_group; i < groups_per_page; i++, group++) {
 		if (group >= ngroups)
 			break;
@@ -1322,10 +1326,11 @@ static int ext4_mb_init_cache(struct page *page, char *incore, gfp_t gfp)
 		/*
 		 * If page is uptodate then we came here after online resize
 		 * which added some new uninitialized group info structs, so
-		 * we must skip all initialized uptodate buddies on the page,
+		 * we must skip all initialized uptodate buddies on the folio,
 		 * which may be currently in use by an allocating task.
 		 */
-		if (PageUptodate(page) && !EXT4_MB_GRP_NEED_INIT(grinfo)) {
+		if (folio_test_uptodate(folio) &&
+				!EXT4_MB_GRP_NEED_INIT(grinfo)) {
 			bh[i] = NULL;
 			continue;
 		}
@@ -1349,7 +1354,7 @@ static int ext4_mb_init_cache(struct page *page, char *incore, gfp_t gfp)
 			err = err2;
 	}
 
-	first_block = page->index * blocks_per_page;
+	first_block = folio->index * blocks_per_page;
 	for (i = 0; i < blocks_per_page; i++) {
 		group = (first_block + i) >> 1;
 		if (group >= ngroups)
@@ -1370,7 +1375,7 @@ static int ext4_mb_init_cache(struct page *page, char *incore, gfp_t gfp)
 		 * above
 		 *
 		 */
-		data = page_address(page) + (i * blocksize);
+		data = folio_address(folio) + (i * blocksize);
 		bitmap = bh[group - first_group]->b_data;
 
 		/*
@@ -1385,8 +1390,8 @@ static int ext4_mb_init_cache(struct page *page, char *incore, gfp_t gfp)
 		if ((first_block + i) & 1) {
 			/* this is block of buddy */
 			BUG_ON(incore == NULL);
-			mb_debug(sb, "put buddy for group %u in page %lu/%x\n",
-				group, page->index, i * blocksize);
+			mb_debug(sb, "put buddy for group %u in folio %lu/%x\n",
+				group, folio->index, i * blocksize);
 			trace_ext4_mb_buddy_bitmap_load(sb, group);
 			grinfo->bb_fragments = 0;
 			memset(grinfo->bb_counters, 0,
@@ -1404,8 +1409,8 @@ static int ext4_mb_init_cache(struct page *page, char *incore, gfp_t gfp)
 		} else {
 			/* this is block of bitmap */
 			BUG_ON(incore != NULL);
-			mb_debug(sb, "put bitmap for group %u in page %lu/%x\n",
-				group, page->index, i * blocksize);
+			mb_debug(sb, "put bitmap for group %u in folio %lu/%x\n",
+				group, folio->index, i * blocksize);
 			trace_ext4_mb_bitmap_load(sb, group);
 
 			/* see comments in ext4_mb_put_pa() */
@@ -1423,7 +1428,7 @@ static int ext4_mb_init_cache(struct page *page, char *incore, gfp_t gfp)
 			incore = data;
 		}
 	}
-	SetPageUptodate(page);
+	folio_mark_uptodate(folio);
 
 out:
 	if (bh) {
@@ -1439,7 +1444,7 @@ out:
  * Lock the buddy and bitmap pages. This make sure other parallel init_group
  * on the same buddy page doesn't happen whild holding the buddy page lock.
  * Return locked buddy and bitmap pages on e4b struct. If buddy and bitmap
- * are on the same page e4b->bd_buddy_page is NULL and return value is 0.
+ * are on the same page e4b->bd_buddy_folio is NULL and return value is 0.
  */
 static int ext4_mb_get_buddy_page_lock(struct super_block *sb,
 		ext4_group_t group, struct ext4_buddy *e4b, gfp_t gfp)
@@ -1447,10 +1452,10 @@ static int ext4_mb_get_buddy_page_lock(struct super_block *sb,
 	struct inode *inode = EXT4_SB(sb)->s_buddy_cache;
 	int block, pnum, poff;
 	int blocks_per_page;
-	struct page *page;
+	struct folio *folio;
 
-	e4b->bd_buddy_page = NULL;
-	e4b->bd_bitmap_page = NULL;
+	e4b->bd_buddy_folio = NULL;
+	e4b->bd_bitmap_folio = NULL;
 
 	blocks_per_page = PAGE_SIZE / sb->s_blocksize;
 	/*
@@ -1461,12 +1466,13 @@ static int ext4_mb_get_buddy_page_lock(struct super_block *sb,
 	block = group * 2;
 	pnum = block / blocks_per_page;
 	poff = block % blocks_per_page;
-	page = find_or_create_page(inode->i_mapping, pnum, gfp);
-	if (!page)
-		return -ENOMEM;
-	BUG_ON(page->mapping != inode->i_mapping);
-	e4b->bd_bitmap_page = page;
-	e4b->bd_bitmap = page_address(page) + (poff * sb->s_blocksize);
+	folio = __filemap_get_folio(inode->i_mapping, pnum,
+			FGP_LOCK | FGP_ACCESSED | FGP_CREAT, gfp);
+	if (IS_ERR(folio))
+		return PTR_ERR(folio);
+	BUG_ON(folio->mapping != inode->i_mapping);
+	e4b->bd_bitmap_folio = folio;
+	e4b->bd_bitmap = folio_address(folio) + (poff * sb->s_blocksize);
 
 	if (blocks_per_page >= 2) {
 		/* buddy and bitmap are on the same page */
@@ -1474,23 +1480,24 @@ static int ext4_mb_get_buddy_page_lock(struct super_block *sb,
 	}
 
 	/* blocks_per_page == 1, hence we need another page for the buddy */
-	page = find_or_create_page(inode->i_mapping, block + 1, gfp);
-	if (!page)
-		return -ENOMEM;
-	BUG_ON(page->mapping != inode->i_mapping);
-	e4b->bd_buddy_page = page;
+	folio = __filemap_get_folio(inode->i_mapping, block + 1,
+			FGP_LOCK | FGP_ACCESSED | FGP_CREAT, gfp);
+	if (IS_ERR(folio))
+		return PTR_ERR(folio);
+	BUG_ON(folio->mapping != inode->i_mapping);
+	e4b->bd_buddy_folio = folio;
 	return 0;
 }
 
 static void ext4_mb_put_buddy_page_lock(struct ext4_buddy *e4b)
 {
-	if (e4b->bd_bitmap_page) {
-		unlock_page(e4b->bd_bitmap_page);
-		put_page(e4b->bd_bitmap_page);
+	if (e4b->bd_bitmap_folio) {
+		folio_unlock(e4b->bd_bitmap_folio);
+		folio_put(e4b->bd_bitmap_folio);
 	}
-	if (e4b->bd_buddy_page) {
-		unlock_page(e4b->bd_buddy_page);
-		put_page(e4b->bd_buddy_page);
+	if (e4b->bd_buddy_folio) {
+		folio_unlock(e4b->bd_buddy_folio);
+		folio_put(e4b->bd_buddy_folio);
 	}
 }
 
@@ -1505,7 +1512,7 @@ int ext4_mb_init_group(struct super_block *sb, ext4_group_t group, gfp_t gfp)
 
 	struct ext4_group_info *this_grp;
 	struct ext4_buddy e4b;
-	struct page *page;
+	struct folio *folio;
 	int ret = 0;
 
 	might_sleep();
@@ -1532,16 +1539,16 @@ int ext4_mb_init_group(struct super_block *sb, ext4_group_t group, gfp_t gfp)
 		goto err;
 	}
 
-	page = e4b.bd_bitmap_page;
-	ret = ext4_mb_init_cache(page, NULL, gfp);
+	folio = e4b.bd_bitmap_folio;
+	ret = ext4_mb_init_cache(folio, NULL, gfp);
 	if (ret)
 		goto err;
-	if (!PageUptodate(page)) {
+	if (!folio_test_uptodate(folio)) {
 		ret = -EIO;
 		goto err;
 	}
 
-	if (e4b.bd_buddy_page == NULL) {
+	if (e4b.bd_buddy_folio == NULL) {
 		/*
 		 * If both the bitmap and buddy are in
 		 * the same page we don't need to force
@@ -1551,11 +1558,11 @@ int ext4_mb_init_group(struct super_block *sb, ext4_group_t group, gfp_t gfp)
 		goto err;
 	}
 	/* init buddy cache */
-	page = e4b.bd_buddy_page;
-	ret = ext4_mb_init_cache(page, e4b.bd_bitmap, gfp);
+	folio = e4b.bd_buddy_folio;
+	ret = ext4_mb_init_cache(folio, e4b.bd_bitmap, gfp);
 	if (ret)
 		goto err;
-	if (!PageUptodate(page)) {
+	if (!folio_test_uptodate(folio)) {
 		ret = -EIO;
 		goto err;
 	}
@@ -1577,7 +1584,7 @@ ext4_mb_load_buddy_gfp(struct super_block *sb, ext4_group_t group,
 	int block;
 	int pnum;
 	int poff;
-	struct page *page;
+	struct folio *folio;
 	int ret;
 	struct ext4_group_info *grp;
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
@@ -1595,8 +1602,8 @@ ext4_mb_load_buddy_gfp(struct super_block *sb, ext4_group_t group,
 	e4b->bd_info = grp;
 	e4b->bd_sb = sb;
 	e4b->bd_group = group;
-	e4b->bd_buddy_page = NULL;
-	e4b->bd_bitmap_page = NULL;
+	e4b->bd_buddy_folio = NULL;
+	e4b->bd_bitmap_folio = NULL;
 
 	if (unlikely(EXT4_MB_GRP_NEED_INIT(grp))) {
 		/*
@@ -1617,102 +1624,103 @@ ext4_mb_load_buddy_gfp(struct super_block *sb, ext4_group_t group,
 	pnum = block / blocks_per_page;
 	poff = block % blocks_per_page;
 
-	/* we could use find_or_create_page(), but it locks page
-	 * what we'd like to avoid in fast path ... */
-	page = find_get_page_flags(inode->i_mapping, pnum, FGP_ACCESSED);
-	if (page == NULL || !PageUptodate(page)) {
-		if (page)
+	/* Avoid locking the folio in the fast path ... */
+	folio = __filemap_get_folio(inode->i_mapping, pnum, FGP_ACCESSED, 0);
+	if (IS_ERR(folio) || !folio_test_uptodate(folio)) {
+		if (!IS_ERR(folio))
 			/*
-			 * drop the page reference and try
-			 * to get the page with lock. If we
+			 * drop the folio reference and try
+			 * to get the folio with lock. If we
 			 * are not uptodate that implies
-			 * somebody just created the page but
-			 * is yet to initialize the same. So
+			 * somebody just created the folio but
+			 * is yet to initialize it. So
 			 * wait for it to initialize.
 			 */
-			put_page(page);
-		page = find_or_create_page(inode->i_mapping, pnum, gfp);
-		if (page) {
-			if (WARN_RATELIMIT(page->mapping != inode->i_mapping,
-	"ext4: bitmap's paging->mapping != inode->i_mapping\n")) {
+			folio_put(folio);
+		folio = __filemap_get_folio(inode->i_mapping, pnum,
+				FGP_LOCK | FGP_ACCESSED | FGP_CREAT, gfp);
+		if (!IS_ERR(folio)) {
+			if (WARN_RATELIMIT(folio->mapping != inode->i_mapping,
+	"ext4: bitmap's mapping != inode->i_mapping\n")) {
 				/* should never happen */
-				unlock_page(page);
+				folio_unlock(folio);
 				ret = -EINVAL;
 				goto err;
 			}
-			if (!PageUptodate(page)) {
-				ret = ext4_mb_init_cache(page, NULL, gfp);
+			if (!folio_test_uptodate(folio)) {
+				ret = ext4_mb_init_cache(folio, NULL, gfp);
 				if (ret) {
-					unlock_page(page);
+					folio_unlock(folio);
 					goto err;
 				}
-				mb_cmp_bitmaps(e4b, page_address(page) +
+				mb_cmp_bitmaps(e4b, folio_address(folio) +
 					       (poff * sb->s_blocksize));
 			}
-			unlock_page(page);
+			folio_unlock(folio);
 		}
 	}
-	if (page == NULL) {
-		ret = -ENOMEM;
+	if (IS_ERR(folio)) {
+		ret = PTR_ERR(folio);
 		goto err;
 	}
-	if (!PageUptodate(page)) {
+	if (!folio_test_uptodate(folio)) {
 		ret = -EIO;
 		goto err;
 	}
 
-	/* Pages marked accessed already */
-	e4b->bd_bitmap_page = page;
-	e4b->bd_bitmap = page_address(page) + (poff * sb->s_blocksize);
+	/* Folios marked accessed already */
+	e4b->bd_bitmap_folio = folio;
+	e4b->bd_bitmap = folio_address(folio) + (poff * sb->s_blocksize);
 
 	block++;
 	pnum = block / blocks_per_page;
 	poff = block % blocks_per_page;
 
-	page = find_get_page_flags(inode->i_mapping, pnum, FGP_ACCESSED);
-	if (page == NULL || !PageUptodate(page)) {
-		if (page)
-			put_page(page);
-		page = find_or_create_page(inode->i_mapping, pnum, gfp);
-		if (page) {
-			if (WARN_RATELIMIT(page->mapping != inode->i_mapping,
-	"ext4: buddy bitmap's page->mapping != inode->i_mapping\n")) {
+	folio = __filemap_get_folio(inode->i_mapping, pnum, FGP_ACCESSED, 0);
+	if (IS_ERR(folio) || !folio_test_uptodate(folio)) {
+		if (!IS_ERR(folio))
+			folio_put(folio);
+		folio = __filemap_get_folio(inode->i_mapping, pnum,
+				FGP_LOCK | FGP_ACCESSED | FGP_CREAT, gfp);
+		if (!IS_ERR(folio)) {
+			if (WARN_RATELIMIT(folio->mapping != inode->i_mapping,
+	"ext4: buddy bitmap's mapping != inode->i_mapping\n")) {
 				/* should never happen */
-				unlock_page(page);
+				folio_unlock(folio);
 				ret = -EINVAL;
 				goto err;
 			}
-			if (!PageUptodate(page)) {
-				ret = ext4_mb_init_cache(page, e4b->bd_bitmap,
+			if (!folio_test_uptodate(folio)) {
+				ret = ext4_mb_init_cache(folio, e4b->bd_bitmap,
 							 gfp);
 				if (ret) {
-					unlock_page(page);
+					folio_unlock(folio);
 					goto err;
 				}
 			}
-			unlock_page(page);
+			folio_unlock(folio);
 		}
 	}
-	if (page == NULL) {
-		ret = -ENOMEM;
+	if (IS_ERR(folio)) {
+		ret = PTR_ERR(folio);
 		goto err;
 	}
-	if (!PageUptodate(page)) {
+	if (!folio_test_uptodate(folio)) {
 		ret = -EIO;
 		goto err;
 	}
 
-	/* Pages marked accessed already */
-	e4b->bd_buddy_page = page;
-	e4b->bd_buddy = page_address(page) + (poff * sb->s_blocksize);
+	/* Folios marked accessed already */
+	e4b->bd_buddy_folio = folio;
+	e4b->bd_buddy = folio_address(folio) + (poff * sb->s_blocksize);
 
 	return 0;
 
 err:
-	if (page)
-		put_page(page);
-	if (e4b->bd_bitmap_page)
-		put_page(e4b->bd_bitmap_page);
+	if (!IS_ERR_OR_NULL(folio))
+		folio_put(folio);
+	if (e4b->bd_bitmap_folio)
+		folio_put(e4b->bd_bitmap_folio);
 
 	e4b->bd_buddy = NULL;
 	e4b->bd_bitmap = NULL;
@@ -1727,10 +1735,10 @@ static int ext4_mb_load_buddy(struct super_block *sb, ext4_group_t group,
 
 static void ext4_mb_unload_buddy(struct ext4_buddy *e4b)
 {
-	if (e4b->bd_bitmap_page)
-		put_page(e4b->bd_bitmap_page);
-	if (e4b->bd_buddy_page)
-		put_page(e4b->bd_buddy_page);
+	if (e4b->bd_bitmap_folio)
+		folio_put(e4b->bd_bitmap_folio);
+	if (e4b->bd_buddy_folio)
+		folio_put(e4b->bd_buddy_folio);
 }
 
 
@@ -2040,13 +2048,12 @@ static int mb_mark_used(struct ext4_buddy *e4b, struct ext4_free_extent *ex)
 	int ord;
 	int mlen = 0;
 	int max = 0;
-	int cur;
 	int start = ex->fe_start;
 	int len = ex->fe_len;
 	unsigned ret = 0;
 	int len0 = len;
 	void *buddy;
-	bool split = false;
+	int ord_start, ord_end;
 
 	BUG_ON(start + len > (e4b->bd_sb->s_blocksize << 3));
 	BUG_ON(e4b->bd_group != ex->fe_group);
@@ -2071,16 +2078,12 @@ static int mb_mark_used(struct ext4_buddy *e4b, struct ext4_free_extent *ex)
 
 	/* let's maintain buddy itself */
 	while (len) {
-		if (!split)
-			ord = mb_find_order_for_block(e4b, start);
+		ord = mb_find_order_for_block(e4b, start);
 
 		if (((start >> ord) << ord) == start && len >= (1 << ord)) {
 			/* the whole chunk may be allocated at once! */
 			mlen = 1 << ord;
-			if (!split)
-				buddy = mb_find_buddy(e4b, ord, &max);
-			else
-				split = false;
+			buddy = mb_find_buddy(e4b, ord, &max);
 			BUG_ON((start >> ord) >= max);
 			mb_set_bit(start >> ord, buddy);
 			e4b->bd_info->bb_counters[ord]--;
@@ -2094,20 +2097,29 @@ static int mb_mark_used(struct ext4_buddy *e4b, struct ext4_free_extent *ex)
 		if (ret == 0)
 			ret = len | (ord << 16);
 
-		/* we have to split large buddy */
 		BUG_ON(ord <= 0);
 		buddy = mb_find_buddy(e4b, ord, &max);
 		mb_set_bit(start >> ord, buddy);
 		e4b->bd_info->bb_counters[ord]--;
 
-		ord--;
-		cur = (start >> ord) & ~1U;
-		buddy = mb_find_buddy(e4b, ord, &max);
-		mb_clear_bit(cur, buddy);
-		mb_clear_bit(cur + 1, buddy);
-		e4b->bd_info->bb_counters[ord]++;
-		e4b->bd_info->bb_counters[ord]++;
-		split = true;
+		ord_start = (start >> ord) << ord;
+		ord_end = ord_start + (1 << ord);
+		/* first chunk */
+		if (start > ord_start)
+			ext4_mb_mark_free_simple(e4b->bd_sb, e4b->bd_buddy,
+						 ord_start, start - ord_start,
+						 e4b->bd_info);
+
+		/* last chunk */
+		if (start + len < ord_end) {
+			ext4_mb_mark_free_simple(e4b->bd_sb, e4b->bd_buddy,
+						 start + len,
+						 ord_end - (start + len),
+						 e4b->bd_info);
+			break;
+		}
+		len = start + len - ord_end;
+		start = ord_end;
 	}
 	mb_set_largest_free_order(e4b->bd_sb, e4b->bd_info);
 
@@ -2149,10 +2161,10 @@ static void ext4_mb_use_best_found(struct ext4_allocation_context *ac,
 	 * double allocate blocks. The reference is dropped
 	 * in ext4_mb_release_context
 	 */
-	ac->ac_bitmap_page = e4b->bd_bitmap_page;
-	get_page(ac->ac_bitmap_page);
-	ac->ac_buddy_page = e4b->bd_buddy_page;
-	get_page(ac->ac_buddy_page);
+	ac->ac_bitmap_folio = e4b->bd_bitmap_folio;
+	folio_get(ac->ac_bitmap_folio);
+	ac->ac_buddy_folio = e4b->bd_buddy_folio;
+	folio_get(ac->ac_buddy_folio);
 	/* store last allocated for subsequent stream allocation */
 	if (ac->ac_flags & EXT4_MB_STREAM_ALLOC) {
 		spin_lock(&sbi->s_md_lock);
@@ -2675,7 +2687,7 @@ static int ext4_mb_good_group_nolock(struct ext4_allocation_context *ac,
 		int ret;
 
 		/*
-		 * cr=CR_POWER2_ALIGNED/CR_GOAL_LEN_FAST is a very optimistic
+		 * CR_POWER2_ALIGNED/CR_GOAL_LEN_FAST is a very optimistic
 		 * search to find large good chunks almost for free. If buddy
 		 * data is not ready, then this optimization makes no sense. But
 		 * we never skip the first block group in a flex_bg, since this
@@ -2856,6 +2868,7 @@ repeat:
 		group = ac->ac_g_ex.fe_group;
 		ac->ac_groups_linear_remaining = sbi->s_mb_max_linear_groups;
 		prefetch_grp = group;
+		nr = 0;
 
 		for (i = 0, new_cr = cr; i < ngroups; i++,
 		     ext4_mb_choose_next_group(ac, &new_cr, &group, ngroups)) {
@@ -3015,8 +3028,8 @@ static int ext4_mb_seq_groups_show(struct seq_file *seq, void *v)
 {
 	struct super_block *sb = pde_data(file_inode(seq->file));
 	ext4_group_t group = (ext4_group_t) ((unsigned long) v);
-	int i;
-	int err, buddy_loaded = 0;
+	int i, err;
+	char nbuf[16];
 	struct ext4_buddy e4b;
 	struct ext4_group_info *grinfo;
 	unsigned char blocksize_bits = min_t(unsigned char,
@@ -3043,17 +3056,17 @@ static int ext4_mb_seq_groups_show(struct seq_file *seq, void *v)
 	if (unlikely(EXT4_MB_GRP_NEED_INIT(grinfo))) {
 		err = ext4_mb_load_buddy(sb, group, &e4b);
 		if (err) {
-			seq_printf(seq, "#%-5u: I/O error\n", group);
+			seq_printf(seq, "#%-5u: %s\n", group, ext4_decode_error(NULL, err, nbuf));
 			return 0;
 		}
-		buddy_loaded = 1;
+		ext4_mb_unload_buddy(&e4b);
 	}
 
+	/*
+	 * We care only about free space counters in the group info and
+	 * these are safe to access even after the buddy has been unloaded
+	 */
 	memcpy(&sg, grinfo, i);
-
-	if (buddy_loaded)
-		ext4_mb_unload_buddy(&e4b);
-
 	seq_printf(seq, "#%-5u: %-5u %-5u %-5u [", group, sg.info.bb_free,
 			sg.info.bb_fragments, sg.info.bb_first_free);
 	for (i = 0; i <= 13; i++)
@@ -3186,7 +3199,6 @@ int ext4_seq_mb_stats_show(struct seq_file *seq, void *offset)
 }
 
 static void *ext4_mb_seq_structs_summary_start(struct seq_file *seq, loff_t *pos)
-__acquires(&EXT4_SB(sb)->s_mb_rb_lock)
 {
 	struct super_block *sb = pde_data(file_inode(seq->file));
 	unsigned long position;
@@ -3440,10 +3452,11 @@ static int ext4_mb_init_backend(struct super_block *sb)
 	}
 	if (sbi->s_mb_prefetch > ext4_get_groups_count(sb))
 		sbi->s_mb_prefetch = ext4_get_groups_count(sb);
-	/* now many real IOs to prefetch within a single allocation at cr=0
-	 * given cr=0 is an CPU-related optimization we shouldn't try to
-	 * load too many groups, at some point we should start to use what
-	 * we've got in memory.
+	/*
+	 * now many real IOs to prefetch within a single allocation at
+	 * CR_POWER2_ALIGNED. Given CR_POWER2_ALIGNED is an CPU-related
+	 * optimization we shouldn't try to load too many groups, at some point
+	 * we should start to use what we've got in memory.
 	 * with an average random access time 5ms, it'd take a second to get
 	 * 200 groups (* N with flex_bg), so let's make this limit 4
 	 */
@@ -3832,8 +3845,7 @@ void ext4_mb_release(struct super_block *sb)
 }
 
 static inline int ext4_issue_discard(struct super_block *sb,
-		ext4_group_t block_group, ext4_grpblk_t cluster, int count,
-		struct bio **biop)
+		ext4_group_t block_group, ext4_grpblk_t cluster, int count)
 {
 	ext4_fsblk_t discard_block;
 
@@ -3842,13 +3854,8 @@ static inline int ext4_issue_discard(struct super_block *sb,
 	count = EXT4_C2B(EXT4_SB(sb), count);
 	trace_ext4_discard_blocks(sb,
 			(unsigned long long) discard_block, count);
-	if (biop) {
-		return __blkdev_issue_discard(sb->s_bdev,
-			(sector_t)discard_block << (sb->s_blocksize_bits - 9),
-			(sector_t)count << (sb->s_blocksize_bits - 9),
-			GFP_NOFS, biop);
-	} else
-		return sb_issue_discard(sb, discard_block, count, GFP_NOFS, 0);
+
+	return sb_issue_discard(sb, discard_block, count, GFP_NOFS, 0);
 }
 
 static void ext4_free_data_in_buddy(struct super_block *sb,
@@ -3890,8 +3897,8 @@ static void ext4_free_data_in_buddy(struct super_block *sb,
 		/* No more items in the per group rb tree
 		 * balance refcounts from ext4_mb_free_metadata()
 		 */
-		put_page(e4b.bd_buddy_page);
-		put_page(e4b.bd_bitmap_page);
+		folio_put(e4b.bd_buddy_folio);
+		folio_put(e4b.bd_bitmap_folio);
 	}
 	ext4_unlock_group(sb, entry->efd_group);
 	ext4_mb_unload_buddy(&e4b);
@@ -5995,10 +6002,10 @@ static void ext4_mb_release_context(struct ext4_allocation_context *ac)
 
 		ext4_mb_put_pa(ac, ac->ac_sb, pa);
 	}
-	if (ac->ac_bitmap_page)
-		put_page(ac->ac_bitmap_page);
-	if (ac->ac_buddy_page)
-		put_page(ac->ac_buddy_page);
+	if (ac->ac_bitmap_folio)
+		folio_put(ac->ac_bitmap_folio);
+	if (ac->ac_buddy_folio)
+		folio_put(ac->ac_buddy_folio);
 	if (ac->ac_flags & EXT4_MB_HINT_GROUP_ALLOC)
 		mutex_unlock(&ac->ac_lg->lg_mutex);
 	ext4_mb_collect_stats(ac);
@@ -6314,8 +6321,8 @@ ext4_mb_free_metadata(handle_t *handle, struct ext4_buddy *e4b,
 	struct rb_node *parent = NULL, *new_node;
 
 	BUG_ON(!ext4_handle_valid(handle));
-	BUG_ON(e4b->bd_bitmap_page == NULL);
-	BUG_ON(e4b->bd_buddy_page == NULL);
+	BUG_ON(e4b->bd_bitmap_folio == NULL);
+	BUG_ON(e4b->bd_buddy_folio == NULL);
 
 	new_node = &new_entry->efd_node;
 	cluster = new_entry->efd_start_cluster;
@@ -6326,8 +6333,8 @@ ext4_mb_free_metadata(handle_t *handle, struct ext4_buddy *e4b,
 		 * otherwise we'll refresh it from
 		 * on-disk bitmap and lose not-yet-available
 		 * blocks */
-		get_page(e4b->bd_buddy_page);
-		get_page(e4b->bd_bitmap_page);
+		folio_get(e4b->bd_buddy_folio);
+		folio_get(e4b->bd_bitmap_folio);
 	}
 	while (*n) {
 		parent = *n;
@@ -6496,8 +6503,14 @@ do_more:
 	} else {
 		if (test_opt(sb, DISCARD)) {
 			err = ext4_issue_discard(sb, block_group, bit,
-						 count_clusters, NULL);
-			if (err && err != -EOPNOTSUPP)
+						 count_clusters);
+			/*
+			 * Ignore EOPNOTSUPP error. This is consistent with
+			 * what happens when using journal.
+			 */
+			if (err == -EOPNOTSUPP)
+				err = 0;
+			if (err)
 				ext4_msg(sb, KERN_WARNING, "discard request in"
 					 " group:%u block:%d count:%lu failed"
 					 " with %d", block_group, bit, count,
@@ -6747,7 +6760,7 @@ __acquires(bitlock)
 	 */
 	mb_mark_used(e4b, &ex);
 	ext4_unlock_group(sb, group);
-	ret = ext4_issue_discard(sb, group, start, count, NULL);
+	ret = ext4_issue_discard(sb, group, start, count);
 	ext4_lock_group(sb, group);
 	mb_free_blocks(NULL, e4b, start, ex.fe_len);
 	return ret;
diff --git a/fs/ext4/mballoc.h b/fs/ext4/mballoc.h
index 56938532b4..d8553f1498 100644
--- a/fs/ext4/mballoc.h
+++ b/fs/ext4/mballoc.h
@@ -187,14 +187,14 @@ struct ext4_allocation_context {
 	struct ext4_free_extent ac_f_ex;
 
 	/*
-	 * goal len can change in CR1.5, so save the original len. This is
-	 * used while adjusting the PA window and for accounting.
+	 * goal len can change in CR_BEST_AVAIL_LEN, so save the original len.
+	 * This is used while adjusting the PA window and for accounting.
 	 */
 	ext4_grpblk_t	ac_orig_goal_len;
 
 	__u32 ac_flags;		/* allocation hints */
+	__u32 ac_groups_linear_remaining;
 	__u16 ac_groups_scanned;
-	__u16 ac_groups_linear_remaining;
 	__u16 ac_found;
 	__u16 ac_cX_found[EXT4_MB_NUM_CRS];
 	__u16 ac_tail;
@@ -204,8 +204,8 @@ struct ext4_allocation_context {
 	__u8 ac_2order;		/* if request is to allocate 2^N blocks and
 				 * N > 0, the field stores N, otherwise 0 */
 	__u8 ac_op;		/* operation, for history only */
-	struct page *ac_bitmap_page;
-	struct page *ac_buddy_page;
+	struct folio *ac_bitmap_folio;
+	struct folio *ac_buddy_folio;
 	struct ext4_prealloc_space *ac_pa;
 	struct ext4_locality_group *ac_lg;
 };
@@ -215,9 +215,9 @@ struct ext4_allocation_context {
 #define AC_STATUS_BREAK		3
 
 struct ext4_buddy {
-	struct page *bd_buddy_page;
+	struct folio *bd_buddy_folio;
 	void *bd_buddy;
-	struct page *bd_bitmap_page;
+	struct folio *bd_bitmap_folio;
 	void *bd_bitmap;
 	struct ext4_group_info *bd_info;
 	struct super_block *bd_sb;
diff --git a/fs/ext4/move_extent.c b/fs/ext4/move_extent.c
index 7cd4afa4de..204f53b236 100644
--- a/fs/ext4/move_extent.c
+++ b/fs/ext4/move_extent.c
@@ -199,10 +199,8 @@ mext_page_mkuptodate(struct folio *folio, unsigned from, unsigned to)
 			continue;
 		if (!buffer_mapped(bh)) {
 			err = ext4_get_block(inode, block, bh, 0);
-			if (err) {
-				folio_set_error(folio);
+			if (err)
 				return err;
-			}
 			if (!buffer_mapped(bh)) {
 				folio_zero_range(folio, block_start, blocksize);
 				set_buffer_uptodate(bh);
diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index 58fee3c6fe..1311ad0464 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -151,10 +151,11 @@ static struct buffer_head *__ext4_read_dirblock(struct inode *inode,
 
 		return bh;
 	}
-	if (!bh && (type == INDEX || type == DIRENT_HTREE)) {
+	/* The first directory block must not be a hole. */
+	if (!bh && (type == INDEX || type == DIRENT_HTREE || block == 0)) {
 		ext4_error_inode(inode, func, line, block,
-				 "Directory hole found for htree %s block",
-				 (type == INDEX) ? "index" : "leaf");
+				 "Directory hole found for htree %s block %u",
+				 (type == INDEX) ? "index" : "leaf", block);
 		return ERR_PTR(-EFSCORRUPTED);
 	}
 	if (!bh)
@@ -1762,7 +1763,6 @@ static struct buffer_head *ext4_lookup_entry(struct inode *dir,
 	struct buffer_head *bh;
 
 	err = ext4_fname_prepare_lookup(dir, dentry, &fname);
-	generic_set_encrypted_ci_d_ops(dentry);
 	if (err == -ENOENT)
 		return NULL;
 	if (err)
@@ -2218,6 +2218,52 @@ static int add_dirent_to_buf(handle_t *handle, struct ext4_filename *fname,
 	return err ? err : err2;
 }
 
+static bool ext4_check_dx_root(struct inode *dir, struct dx_root *root)
+{
+	struct fake_dirent *fde;
+	const char *error_msg;
+	unsigned int rlen;
+	unsigned int blocksize = dir->i_sb->s_blocksize;
+	char *blockend = (char *)root + dir->i_sb->s_blocksize;
+
+	fde = &root->dot;
+	if (unlikely(fde->name_len != 1)) {
+		error_msg = "invalid name_len for '.'";
+		goto corrupted;
+	}
+	if (unlikely(strncmp(root->dot_name, ".", fde->name_len))) {
+		error_msg = "invalid name for '.'";
+		goto corrupted;
+	}
+	rlen = ext4_rec_len_from_disk(fde->rec_len, blocksize);
+	if (unlikely((char *)fde + rlen >= blockend)) {
+		error_msg = "invalid rec_len for '.'";
+		goto corrupted;
+	}
+
+	fde = &root->dotdot;
+	if (unlikely(fde->name_len != 2)) {
+		error_msg = "invalid name_len for '..'";
+		goto corrupted;
+	}
+	if (unlikely(strncmp(root->dotdot_name, "..", fde->name_len))) {
+		error_msg = "invalid name for '..'";
+		goto corrupted;
+	}
+	rlen = ext4_rec_len_from_disk(fde->rec_len, blocksize);
+	if (unlikely((char *)fde + rlen >= blockend)) {
+		error_msg = "invalid rec_len for '..'";
+		goto corrupted;
+	}
+
+	return true;
+
+corrupted:
+	EXT4_ERROR_INODE(dir, "Corrupt dir, %s, running e2fsck is recommended",
+			 error_msg);
+	return false;
+}
+
 /*
  * This converts a one block unindexed directory to a 3 block indexed
  * directory, and adds the dentry to the indexed directory.
@@ -2252,17 +2298,17 @@ static int make_indexed_dir(handle_t *handle, struct ext4_filename *fname,
 		brelse(bh);
 		return retval;
 	}
+
 	root = (struct dx_root *) bh->b_data;
+	if (!ext4_check_dx_root(dir, root)) {
+		brelse(bh);
+		return -EFSCORRUPTED;
+	}
 
 	/* The 0th block becomes the root, move the dirents out */
 	fde = &root->dotdot;
 	de = (struct ext4_dir_entry_2 *)((char *)fde +
 		ext4_rec_len_from_disk(fde->rec_len, blocksize));
-	if ((char *) de >= (((char *) root) + blocksize)) {
-		EXT4_ERROR_INODE(dir, "invalid rec_len for '..'");
-		brelse(bh);
-		return -EFSCORRUPTED;
-	}
 	len = ((char *) root) + (blocksize - csum_size) - (char *) de;
 
 	/* Allocate new block for the 0th block's dirents */
@@ -3084,10 +3130,7 @@ bool ext4_empty_dir(struct inode *inode)
 		EXT4_ERROR_INODE(inode, "invalid size");
 		return false;
 	}
-	/* The first directory block must not be a hole,
-	 * so treat it as DIRENT_HTREE
-	 */
-	bh = ext4_read_dirblock(inode, 0, DIRENT_HTREE);
+	bh = ext4_read_dirblock(inode, 0, EITHER);
 	if (IS_ERR(bh))
 		return false;
 
@@ -3532,10 +3575,7 @@ static struct buffer_head *ext4_get_first_dir_block(handle_t *handle,
 		struct ext4_dir_entry_2 *de;
 		unsigned int offset;
 
-		/* The first directory block must not be a hole, so
-		 * treat it as DIRENT_HTREE
-		 */
-		bh = ext4_read_dirblock(inode, 0, DIRENT_HTREE);
+		bh = ext4_read_dirblock(inode, 0, EITHER);
 		if (IS_ERR(bh)) {
 			*retval = PTR_ERR(bh);
 			return NULL;
diff --git a/fs/ext4/page-io.c b/fs/ext4/page-io.c
index 312bc68133..ad5543866d 100644
--- a/fs/ext4/page-io.c
+++ b/fs/ext4/page-io.c
@@ -117,7 +117,6 @@ static void ext4_finish_bio(struct bio *bio)
 
 		if (bio->bi_status) {
 			int err = blk_status_to_errno(bio->bi_status);
-			folio_set_error(folio);
 			mapping_set_error(folio->mapping, err);
 		}
 		bh = head = folio_buffers(folio);
@@ -441,8 +440,6 @@ int ext4_bio_write_folio(struct ext4_io_submit *io, struct folio *folio,
 	BUG_ON(!folio_test_locked(folio));
 	BUG_ON(folio_test_writeback(folio));
 
-	folio_clear_error(folio);
-
 	/*
 	 * Comments copied from block_write_full_folio:
 	 *
diff --git a/fs/ext4/readpage.c b/fs/ext4/readpage.c
index 21e8f0aebb..8494492582 100644
--- a/fs/ext4/readpage.c
+++ b/fs/ext4/readpage.c
@@ -289,7 +289,6 @@ int ext4_mpage_readpages(struct inode *inode,
 
 				if (ext4_map_blocks(NULL, inode, &map, 0) < 0) {
 				set_error_page:
-					folio_set_error(folio);
 					folio_zero_segment(folio, 0,
 							  folio_size(folio));
 					folio_unlock(folio);
diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c
index 3c0d12382e..0ba9837d65 100644
--- a/fs/ext4/resize.c
+++ b/fs/ext4/resize.c
@@ -2085,7 +2085,7 @@ retry:
 		}
 	}
 
-	if ((!resize_inode && !meta_bg) || n_blocks_count == o_blocks_count) {
+	if ((!resize_inode && !meta_bg && n_desc_blocks > o_desc_blocks) || n_blocks_count == o_blocks_count) {
 		err = ext4_convert_meta_bg(sb, resize_inode);
 		if (err)
 			goto out;
diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index 4866657c6d..c682fb927b 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -244,7 +244,7 @@ static struct buffer_head *__ext4_sb_bread_gfp(struct super_block *sb,
 struct buffer_head *ext4_sb_bread(struct super_block *sb, sector_t block,
 				   blk_opf_t op_flags)
 {
-	gfp_t gfp = mapping_gfp_constraint(sb->s_bdev->bd_inode->i_mapping,
+	gfp_t gfp = mapping_gfp_constraint(sb->s_bdev->bd_mapping,
 			~__GFP_FS) | __GFP_MOVABLE;
 
 	return __ext4_sb_bread_gfp(sb, block, op_flags, gfp);
@@ -253,7 +253,7 @@ struct buffer_head *ext4_sb_bread(struct super_block *sb, sector_t block,
 struct buffer_head *ext4_sb_bread_unmovable(struct super_block *sb,
 					    sector_t block)
 {
-	gfp_t gfp = mapping_gfp_constraint(sb->s_bdev->bd_inode->i_mapping,
+	gfp_t gfp = mapping_gfp_constraint(sb->s_bdev->bd_mapping,
 			~__GFP_FS);
 
 	return __ext4_sb_bread_gfp(sb, block, 0, gfp);
@@ -492,22 +492,6 @@ static void ext4_maybe_update_superblock(struct super_block *sb)
 		schedule_work(&EXT4_SB(sb)->s_sb_upd_work);
 }
 
-/*
- * The del_gendisk() function uninitializes the disk-specific data
- * structures, including the bdi structure, without telling anyone
- * else.  Once this happens, any attempt to call mark_buffer_dirty()
- * (for example, by ext4_commit_super), will cause a kernel OOPS.
- * This is a kludge to prevent these oops until we can put in a proper
- * hook in del_gendisk() to inform the VFS and file system layers.
- */
-static int block_device_ejected(struct super_block *sb)
-{
-	struct inode *bd_inode = sb->s_bdev->bd_inode;
-	struct backing_dev_info *bdi = inode_to_bdi(bd_inode);
-
-	return bdi->dev == NULL;
-}
-
 static void ext4_journal_commit_callback(journal_t *journal, transaction_t *txn)
 {
 	struct super_block		*sb = journal->j_private;
@@ -1359,14 +1343,14 @@ static void ext4_put_super(struct super_block *sb)
 
 	sync_blockdev(sb->s_bdev);
 	invalidate_bdev(sb->s_bdev);
-	if (sbi->s_journal_bdev_handle) {
+	if (sbi->s_journal_bdev_file) {
 		/*
 		 * Invalidate the journal device's buffers.  We don't want them
 		 * floating about in memory - the physical journal device may
 		 * hotswapped, and it breaks the `ro-after' testing code.
 		 */
-		sync_blockdev(sbi->s_journal_bdev_handle->bdev);
-		invalidate_bdev(sbi->s_journal_bdev_handle->bdev);
+		sync_blockdev(file_bdev(sbi->s_journal_bdev_file));
+		invalidate_bdev(file_bdev(sbi->s_journal_bdev_file));
 	}
 
 	ext4_xattr_destroy_cache(sbi->s_ea_inode_cache);
@@ -1500,8 +1484,7 @@ static int __init init_inodecache(void)
 {
 	ext4_inode_cachep = kmem_cache_create_usercopy("ext4_inode_cache",
 				sizeof(struct ext4_inode_info), 0,
-				(SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD|
-					SLAB_ACCOUNT),
+				SLAB_RECLAIM_ACCOUNT | SLAB_ACCOUNT,
 				offsetof(struct ext4_inode_info, i_data),
 				sizeof_field(struct ext4_inode_info, i_data),
 				init_once);
@@ -1724,10 +1707,6 @@ static const struct constant_table ext4_param_dax[] = {
 	{}
 };
 
-/* String parameter that allows empty argument */
-#define fsparam_string_empty(NAME, OPT) \
-	__fsparam(fs_param_is_string, NAME, OPT, fs_param_can_be_empty, NULL)
-
 /*
  * Mount option specification
  * We don't use fsparam_flag_no because of the way we set the
@@ -2079,8 +2058,7 @@ static int unnote_qf_name(struct fs_context *fc, int qtype)
 {
 	struct ext4_fs_context *ctx = fc->fs_private;
 
-	if (ctx->s_qf_names[qtype])
-		kfree(ctx->s_qf_names[qtype]);
+	kfree(ctx->s_qf_names[qtype]);
 
 	ctx->s_qf_names[qtype] = NULL;
 	ctx->qname_spec |= 1 << qtype;
@@ -2485,8 +2463,7 @@ static int parse_options(struct fs_context *fc, char *options)
 			param.size = v_len;
 
 			ret = ext4_parse_param(fc, &param);
-			if (param.string)
-				kfree(param.string);
+			kfree(param.string);
 			if (ret < 0)
 				return ret;
 		}
@@ -4233,7 +4210,7 @@ int ext4_calculate_overhead(struct super_block *sb)
 	 * Add the internal journal blocks whether the journal has been
 	 * loaded or not
 	 */
-	if (sbi->s_journal && !sbi->s_journal_bdev_handle)
+	if (sbi->s_journal && !sbi->s_journal_bdev_file)
 		overhead += EXT4_NUM_B2C(sbi, sbi->s_journal->j_total_len);
 	else if (ext4_has_feature_journal(sb) && !sbi->s_journal && j_inum) {
 		/* j_inum for internal journal is non-zero */
@@ -4422,22 +4399,6 @@ static int ext4_handle_clustersize(struct super_block *sb)
 		}
 		sbi->s_cluster_bits = le32_to_cpu(es->s_log_cluster_size) -
 			le32_to_cpu(es->s_log_block_size);
-		sbi->s_clusters_per_group =
-			le32_to_cpu(es->s_clusters_per_group);
-		if (sbi->s_clusters_per_group > sb->s_blocksize * 8) {
-			ext4_msg(sb, KERN_ERR,
-				 "#clusters per group too big: %lu",
-				 sbi->s_clusters_per_group);
-			return -EINVAL;
-		}
-		if (sbi->s_blocks_per_group !=
-		    (sbi->s_clusters_per_group * (clustersize / sb->s_blocksize))) {
-			ext4_msg(sb, KERN_ERR, "blocks per group (%lu) and "
-				 "clusters per group (%lu) inconsistent",
-				 sbi->s_blocks_per_group,
-				 sbi->s_clusters_per_group);
-			return -EINVAL;
-		}
 	} else {
 		if (clustersize != sb->s_blocksize) {
 			ext4_msg(sb, KERN_ERR,
@@ -4451,9 +4412,21 @@ static int ext4_handle_clustersize(struct super_block *sb)
 				 sbi->s_blocks_per_group);
 			return -EINVAL;
 		}
-		sbi->s_clusters_per_group = sbi->s_blocks_per_group;
 		sbi->s_cluster_bits = 0;
 	}
+	sbi->s_clusters_per_group = le32_to_cpu(es->s_clusters_per_group);
+	if (sbi->s_clusters_per_group > sb->s_blocksize * 8) {
+		ext4_msg(sb, KERN_ERR, "#clusters per group too big: %lu",
+			 sbi->s_clusters_per_group);
+		return -EINVAL;
+	}
+	if (sbi->s_blocks_per_group !=
+	    (sbi->s_clusters_per_group * (clustersize / sb->s_blocksize))) {
+		ext4_msg(sb, KERN_ERR,
+			 "blocks per group (%lu) and clusters per group (%lu) inconsistent",
+			 sbi->s_blocks_per_group, sbi->s_clusters_per_group);
+		return -EINVAL;
+	}
 	sbi->s_cluster_ratio = clustersize / sb->s_blocksize;
 
 	/* Do we have standard group size of clustersize * 8 blocks ? */
@@ -5346,7 +5319,8 @@ static int __ext4_fill_super(struct fs_context *fc, struct super_block *sb)
 		sb->s_qcop = &ext4_qctl_operations;
 	sb->s_quota_types = QTYPE_MASK_USR | QTYPE_MASK_GRP | QTYPE_MASK_PRJ;
 #endif
-	memcpy(&sb->s_uuid, es->s_uuid, sizeof(es->s_uuid));
+	super_set_uuid(sb, es->s_uuid, sizeof(es->s_uuid));
+	super_set_sysfs_name_bdev(sb);
 
 	INIT_LIST_HEAD(&sbi->s_orphan); /* unlinked but open files */
 	mutex_init(&sbi->s_orphan_lock);
@@ -5484,6 +5458,7 @@ static int __ext4_fill_super(struct fs_context *fc, struct super_block *sb)
 		goto failed_mount4;
 	}
 
+	generic_set_sb_d_ops(sb);
 	sb->s_root = d_make_root(root);
 	if (!sb->s_root) {
 		ext4_msg(sb, KERN_ERR, "get root dentry failed");
@@ -5555,19 +5530,15 @@ static int __ext4_fill_super(struct fs_context *fc, struct super_block *sb)
 	if (err)
 		goto failed_mount6;
 
-	err = ext4_register_sysfs(sb);
-	if (err)
-		goto failed_mount7;
-
 	err = ext4_init_orphan_info(sb);
 	if (err)
-		goto failed_mount8;
+		goto failed_mount7;
 #ifdef CONFIG_QUOTA
 	/* Enable quota usage during mount. */
 	if (ext4_has_feature_quota(sb) && !sb_rdonly(sb)) {
 		err = ext4_enable_quotas(sb);
 		if (err)
-			goto failed_mount9;
+			goto failed_mount8;
 	}
 #endif  /* CONFIG_QUOTA */
 
@@ -5576,7 +5547,7 @@ static int __ext4_fill_super(struct fs_context *fc, struct super_block *sb)
 	 * used to detect the metadata async write error.
 	 */
 	spin_lock_init(&sbi->s_bdev_wb_lock);
-	errseq_check_and_advance(&sb->s_bdev->bd_inode->i_mapping->wb_err,
+	errseq_check_and_advance(&sb->s_bdev->bd_mapping->wb_err,
 				 &sbi->s_bdev_wb_err);
 	EXT4_SB(sb)->s_mount_state |= EXT4_ORPHAN_FS;
 	ext4_orphan_cleanup(sb, es);
@@ -5593,7 +5564,7 @@ static int __ext4_fill_super(struct fs_context *fc, struct super_block *sb)
 		ext4_msg(sb, KERN_INFO, "recovery complete");
 		err = ext4_mark_recovery_complete(sb, es);
 		if (err)
-			goto failed_mount10;
+			goto failed_mount9;
 	}
 
 	if (test_opt(sb, DISCARD) && !bdev_max_discard_sectors(sb->s_bdev))
@@ -5610,15 +5581,17 @@ static int __ext4_fill_super(struct fs_context *fc, struct super_block *sb)
 	atomic_set(&sbi->s_warning_count, 0);
 	atomic_set(&sbi->s_msg_count, 0);
 
+	/* Register sysfs after all initializations are complete. */
+	err = ext4_register_sysfs(sb);
+	if (err)
+		goto failed_mount9;
+
 	return 0;
 
-failed_mount10:
+failed_mount9:
 	ext4_quotas_off(sb, EXT4_MAXQUOTAS);
-failed_mount9: __maybe_unused
+failed_mount8: __maybe_unused
 	ext4_release_orphan_info(sb);
-failed_mount8:
-	ext4_unregister_sysfs(sb);
-	kobject_put(&sbi->s_kobj);
 failed_mount7:
 	ext4_unregister_li_request(sb);
 failed_mount6:
@@ -5670,9 +5643,9 @@ failed_mount:
 #endif
 	fscrypt_free_dummy_policy(&sbi->s_dummy_enc_policy);
 	brelse(sbi->s_sbh);
-	if (sbi->s_journal_bdev_handle) {
-		invalidate_bdev(sbi->s_journal_bdev_handle->bdev);
-		bdev_release(sbi->s_journal_bdev_handle);
+	if (sbi->s_journal_bdev_file) {
+		invalidate_bdev(file_bdev(sbi->s_journal_bdev_file));
+		bdev_fput(sbi->s_journal_bdev_file);
 	}
 out_fail:
 	invalidate_bdev(sb->s_bdev);
@@ -5842,30 +5815,30 @@ static journal_t *ext4_open_inode_journal(struct super_block *sb,
 	return journal;
 }
 
-static struct bdev_handle *ext4_get_journal_blkdev(struct super_block *sb,
+static struct file *ext4_get_journal_blkdev(struct super_block *sb,
 					dev_t j_dev, ext4_fsblk_t *j_start,
 					ext4_fsblk_t *j_len)
 {
 	struct buffer_head *bh;
 	struct block_device *bdev;
-	struct bdev_handle *bdev_handle;
+	struct file *bdev_file;
 	int hblock, blocksize;
 	ext4_fsblk_t sb_block;
 	unsigned long offset;
 	struct ext4_super_block *es;
 	int errno;
 
-	bdev_handle = bdev_open_by_dev(j_dev,
+	bdev_file = bdev_file_open_by_dev(j_dev,
 		BLK_OPEN_READ | BLK_OPEN_WRITE | BLK_OPEN_RESTRICT_WRITES,
 		sb, &fs_holder_ops);
-	if (IS_ERR(bdev_handle)) {
+	if (IS_ERR(bdev_file)) {
 		ext4_msg(sb, KERN_ERR,
 			 "failed to open journal device unknown-block(%u,%u) %ld",
-			 MAJOR(j_dev), MINOR(j_dev), PTR_ERR(bdev_handle));
-		return bdev_handle;
+			 MAJOR(j_dev), MINOR(j_dev), PTR_ERR(bdev_file));
+		return bdev_file;
 	}
 
-	bdev = bdev_handle->bdev;
+	bdev = file_bdev(bdev_file);
 	blocksize = sb->s_blocksize;
 	hblock = bdev_logical_block_size(bdev);
 	if (blocksize < hblock) {
@@ -5877,7 +5850,7 @@ static struct bdev_handle *ext4_get_journal_blkdev(struct super_block *sb,
 
 	sb_block = EXT4_MIN_BLOCK_SIZE / blocksize;
 	offset = EXT4_MIN_BLOCK_SIZE % blocksize;
-	set_blocksize(bdev, blocksize);
+	set_blocksize(bdev_file, blocksize);
 	bh = __bread(bdev, sb_block, blocksize);
 	if (!bh) {
 		ext4_msg(sb, KERN_ERR, "couldn't read superblock of "
@@ -5912,12 +5885,12 @@ static struct bdev_handle *ext4_get_journal_blkdev(struct super_block *sb,
 	*j_start = sb_block + 1;
 	*j_len = ext4_blocks_count(es);
 	brelse(bh);
-	return bdev_handle;
+	return bdev_file;
 
 out_bh:
 	brelse(bh);
 out_bdev:
-	bdev_release(bdev_handle);
+	bdev_fput(bdev_file);
 	return ERR_PTR(errno);
 }
 
@@ -5927,14 +5900,14 @@ static journal_t *ext4_open_dev_journal(struct super_block *sb,
 	journal_t *journal;
 	ext4_fsblk_t j_start;
 	ext4_fsblk_t j_len;
-	struct bdev_handle *bdev_handle;
+	struct file *bdev_file;
 	int errno = 0;
 
-	bdev_handle = ext4_get_journal_blkdev(sb, j_dev, &j_start, &j_len);
-	if (IS_ERR(bdev_handle))
-		return ERR_CAST(bdev_handle);
+	bdev_file = ext4_get_journal_blkdev(sb, j_dev, &j_start, &j_len);
+	if (IS_ERR(bdev_file))
+		return ERR_CAST(bdev_file);
 
-	journal = jbd2_journal_init_dev(bdev_handle->bdev, sb->s_bdev, j_start,
+	journal = jbd2_journal_init_dev(file_bdev(bdev_file), sb->s_bdev, j_start,
 					j_len, sb->s_blocksize);
 	if (IS_ERR(journal)) {
 		ext4_msg(sb, KERN_ERR, "failed to create device journal");
@@ -5949,14 +5922,14 @@ static journal_t *ext4_open_dev_journal(struct super_block *sb,
 		goto out_journal;
 	}
 	journal->j_private = sb;
-	EXT4_SB(sb)->s_journal_bdev_handle = bdev_handle;
+	EXT4_SB(sb)->s_journal_bdev_file = bdev_file;
 	ext4_init_journal_params(sb, journal);
 	return journal;
 
 out_journal:
 	jbd2_journal_destroy(journal);
 out_bdev:
-	bdev_release(bdev_handle);
+	bdev_fput(bdev_file);
 	return ERR_PTR(errno);
 }
 
@@ -6134,8 +6107,8 @@ static void ext4_update_super(struct super_block *sb)
 			__ext4_update_tstamp(&es->s_first_error_time,
 					     &es->s_first_error_time_hi,
 					     sbi->s_first_error_time);
-			strncpy(es->s_first_error_func, sbi->s_first_error_func,
-				sizeof(es->s_first_error_func));
+			strtomem_pad(es->s_first_error_func,
+				     sbi->s_first_error_func, 0);
 			es->s_first_error_line =
 				cpu_to_le32(sbi->s_first_error_line);
 			es->s_first_error_ino =
@@ -6148,8 +6121,7 @@ static void ext4_update_super(struct super_block *sb)
 		__ext4_update_tstamp(&es->s_last_error_time,
 				     &es->s_last_error_time_hi,
 				     sbi->s_last_error_time);
-		strncpy(es->s_last_error_func, sbi->s_last_error_func,
-			sizeof(es->s_last_error_func));
+		strtomem_pad(es->s_last_error_func, sbi->s_last_error_func, 0);
 		es->s_last_error_line = cpu_to_le32(sbi->s_last_error_line);
 		es->s_last_error_ino = cpu_to_le32(sbi->s_last_error_ino);
 		es->s_last_error_block = cpu_to_le64(sbi->s_last_error_block);
@@ -6176,8 +6148,6 @@ static int ext4_commit_super(struct super_block *sb)
 
 	if (!sbh)
 		return -EINVAL;
-	if (block_device_ejected(sb))
-		return -ENODEV;
 
 	ext4_update_super(sb);
 
@@ -7326,12 +7296,12 @@ static inline int ext3_feature_set_ok(struct super_block *sb)
 static void ext4_kill_sb(struct super_block *sb)
 {
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
-	struct bdev_handle *handle = sbi ? sbi->s_journal_bdev_handle : NULL;
+	struct file *bdev_file = sbi ? sbi->s_journal_bdev_file : NULL;
 
 	kill_block_super(sb);
 
-	if (handle)
-		bdev_release(handle);
+	if (bdev_file)
+		bdev_fput(bdev_file);
 }
 
 static struct file_system_type ext4_fs_type = {
diff --git a/fs/ext4/sysfs.c b/fs/ext4/sysfs.c
index 6d332dff79..ddb54608ca 100644
--- a/fs/ext4/sysfs.c
+++ b/fs/ext4/sysfs.c
@@ -29,7 +29,10 @@ typedef enum {
 	attr_trigger_test_error,
 	attr_first_error_time,
 	attr_last_error_time,
+	attr_clusters_in_group,
+	attr_mb_order,
 	attr_feature,
+	attr_pointer_pi,
 	attr_pointer_ui,
 	attr_pointer_ul,
 	attr_pointer_u64,
@@ -104,7 +107,7 @@ static ssize_t reserved_clusters_store(struct ext4_sb_info *sbi,
 	int ret;
 
 	ret = kstrtoull(skip_spaces(buf), 0, &val);
-	if (ret || val >= clusters)
+	if (ret || val >= clusters || (s64)val < 0)
 		return -EINVAL;
 
 	atomic64_set(&sbi->s_resv_clusters, val);
@@ -178,6 +181,9 @@ static struct ext4_attr ext4_attr_##_name = {			\
 #define EXT4_RO_ATTR_ES_STRING(_name,_elname,_size)			\
 	EXT4_ATTR_STRING(_name, 0444, _size, ext4_super_block, _elname)
 
+#define EXT4_RW_ATTR_SBI_PI(_name,_elname)      \
+	EXT4_ATTR_OFFSET(_name, 0644, pointer_pi, ext4_sb_info, _elname)
+
 #define EXT4_RW_ATTR_SBI_UI(_name,_elname)	\
 	EXT4_ATTR_OFFSET(_name, 0644, pointer_ui, ext4_sb_info, _elname)
 
@@ -207,23 +213,25 @@ EXT4_ATTR_FUNC(sra_exceeded_retry_limit, 0444);
 
 EXT4_ATTR_OFFSET(inode_readahead_blks, 0644, inode_readahead,
 		 ext4_sb_info, s_inode_readahead_blks);
+EXT4_ATTR_OFFSET(mb_group_prealloc, 0644, clusters_in_group,
+		 ext4_sb_info, s_mb_group_prealloc);
+EXT4_ATTR_OFFSET(mb_best_avail_max_trim_order, 0644, mb_order,
+		 ext4_sb_info, s_mb_best_avail_max_trim_order);
 EXT4_RW_ATTR_SBI_UI(inode_goal, s_inode_goal);
 EXT4_RW_ATTR_SBI_UI(mb_stats, s_mb_stats);
 EXT4_RW_ATTR_SBI_UI(mb_max_to_scan, s_mb_max_to_scan);
 EXT4_RW_ATTR_SBI_UI(mb_min_to_scan, s_mb_min_to_scan);
 EXT4_RW_ATTR_SBI_UI(mb_order2_req, s_mb_order2_reqs);
 EXT4_RW_ATTR_SBI_UI(mb_stream_req, s_mb_stream_request);
-EXT4_RW_ATTR_SBI_UI(mb_group_prealloc, s_mb_group_prealloc);
 EXT4_RW_ATTR_SBI_UI(mb_max_linear_groups, s_mb_max_linear_groups);
 EXT4_RW_ATTR_SBI_UI(extent_max_zeroout_kb, s_extent_max_zeroout_kb);
 EXT4_ATTR(trigger_fs_error, 0200, trigger_test_error);
-EXT4_RW_ATTR_SBI_UI(err_ratelimit_interval_ms, s_err_ratelimit_state.interval);
-EXT4_RW_ATTR_SBI_UI(err_ratelimit_burst, s_err_ratelimit_state.burst);
-EXT4_RW_ATTR_SBI_UI(warning_ratelimit_interval_ms, s_warning_ratelimit_state.interval);
-EXT4_RW_ATTR_SBI_UI(warning_ratelimit_burst, s_warning_ratelimit_state.burst);
-EXT4_RW_ATTR_SBI_UI(msg_ratelimit_interval_ms, s_msg_ratelimit_state.interval);
-EXT4_RW_ATTR_SBI_UI(msg_ratelimit_burst, s_msg_ratelimit_state.burst);
-EXT4_RW_ATTR_SBI_UI(mb_best_avail_max_trim_order, s_mb_best_avail_max_trim_order);
+EXT4_RW_ATTR_SBI_PI(err_ratelimit_interval_ms, s_err_ratelimit_state.interval);
+EXT4_RW_ATTR_SBI_PI(err_ratelimit_burst, s_err_ratelimit_state.burst);
+EXT4_RW_ATTR_SBI_PI(warning_ratelimit_interval_ms, s_warning_ratelimit_state.interval);
+EXT4_RW_ATTR_SBI_PI(warning_ratelimit_burst, s_warning_ratelimit_state.burst);
+EXT4_RW_ATTR_SBI_PI(msg_ratelimit_interval_ms, s_msg_ratelimit_state.interval);
+EXT4_RW_ATTR_SBI_PI(msg_ratelimit_burst, s_msg_ratelimit_state.burst);
 #ifdef CONFIG_EXT4_DEBUG
 EXT4_RW_ATTR_SBI_UL(simulate_fail, s_simulate_fail);
 #endif
@@ -366,13 +374,45 @@ static ssize_t __print_tstamp(char *buf, __le32 lo, __u8 hi)
 #define print_tstamp(buf, es, tstamp) \
 	__print_tstamp(buf, (es)->tstamp, (es)->tstamp ## _hi)
 
+static ssize_t ext4_generic_attr_show(struct ext4_attr *a,
+				      struct ext4_sb_info *sbi, char *buf)
+{
+	void *ptr = calc_ptr(a, sbi);
+
+	if (!ptr)
+		return 0;
+
+	switch (a->attr_id) {
+	case attr_inode_readahead:
+	case attr_clusters_in_group:
+	case attr_mb_order:
+	case attr_pointer_pi:
+	case attr_pointer_ui:
+		if (a->attr_ptr == ptr_ext4_super_block_offset)
+			return sysfs_emit(buf, "%u\n", le32_to_cpup(ptr));
+		return sysfs_emit(buf, "%u\n", *((unsigned int *) ptr));
+	case attr_pointer_ul:
+		return sysfs_emit(buf, "%lu\n", *((unsigned long *) ptr));
+	case attr_pointer_u8:
+		return sysfs_emit(buf, "%u\n", *((unsigned char *) ptr));
+	case attr_pointer_u64:
+		if (a->attr_ptr == ptr_ext4_super_block_offset)
+			return sysfs_emit(buf, "%llu\n", le64_to_cpup(ptr));
+		return sysfs_emit(buf, "%llu\n", *((unsigned long long *) ptr));
+	case attr_pointer_string:
+		return sysfs_emit(buf, "%.*s\n", a->attr_size, (char *) ptr);
+	case attr_pointer_atomic:
+		return sysfs_emit(buf, "%d\n", atomic_read((atomic_t *) ptr));
+	}
+	return 0;
+}
+
 static ssize_t ext4_attr_show(struct kobject *kobj,
 			      struct attribute *attr, char *buf)
 {
 	struct ext4_sb_info *sbi = container_of(kobj, struct ext4_sb_info,
 						s_kobj);
 	struct ext4_attr *a = container_of(attr, struct ext4_attr, attr);
-	void *ptr = calc_ptr(a, sbi);
 
 	switch (a->attr_id) {
 	case attr_delayed_allocation_blocks:
@@ -391,45 +431,6 @@ static ssize_t ext4_attr_show(struct kobject *kobj,
 		return sysfs_emit(buf, "%llu\n",
 				(unsigned long long)
 			percpu_counter_sum(&sbi->s_sra_exceeded_retry_limit));
-	case attr_inode_readahead:
-	case attr_pointer_ui:
-		if (!ptr)
-			return 0;
-		if (a->attr_ptr == ptr_ext4_super_block_offset)
-			return sysfs_emit(buf, "%u\n",
-					le32_to_cpup(ptr));
-		else
-			return sysfs_emit(buf, "%u\n",
-					*((unsigned int *) ptr));
-	case attr_pointer_ul:
-		if (!ptr)
-			return 0;
-		return sysfs_emit(buf, "%lu\n",
-				*((unsigned long *) ptr));
-	case attr_pointer_u8:
-		if (!ptr)
-			return 0;
-		return sysfs_emit(buf, "%u\n",
-				*((unsigned char *) ptr));
-	case attr_pointer_u64:
-		if (!ptr)
-			return 0;
-		if (a->attr_ptr == ptr_ext4_super_block_offset)
-			return sysfs_emit(buf, "%llu\n",
-					le64_to_cpup(ptr));
-		else
-			return sysfs_emit(buf, "%llu\n",
-					*((unsigned long long *) ptr));
-	case attr_pointer_string:
-		if (!ptr)
-			return 0;
-		return sysfs_emit(buf, "%.*s\n", a->attr_size,
-				(char *) ptr);
-	case attr_pointer_atomic:
-		if (!ptr)
-			return 0;
-		return sysfs_emit(buf, "%d\n",
-				atomic_read((atomic_t *) ptr));
 	case attr_feature:
 		return sysfs_emit(buf, "supported\n");
 	case attr_first_error_time:
@@ -438,29 +439,34 @@ static ssize_t ext4_attr_show(struct kobject *kobj,
 		return print_tstamp(buf, sbi->s_es, s_last_error_time);
 	case attr_journal_task:
 		return journal_task_show(sbi, buf);
+	default:
+		return ext4_generic_attr_show(a, sbi, buf);
 	}
-
-	return 0;
 }
 
-static ssize_t ext4_attr_store(struct kobject *kobj,
-			       struct attribute *attr,
-			       const char *buf, size_t len)
+static ssize_t ext4_generic_attr_store(struct ext4_attr *a,
+				       struct ext4_sb_info *sbi,
+				       const char *buf, size_t len)
 {
-	struct ext4_sb_info *sbi = container_of(kobj, struct ext4_sb_info,
-						s_kobj);
-	struct ext4_attr *a = container_of(attr, struct ext4_attr, attr);
-	void *ptr = calc_ptr(a, sbi);
-	unsigned long t;
 	int ret;
+	unsigned int t;
+	unsigned long lt;
+	void *ptr = calc_ptr(a, sbi);
+
+	if (!ptr)
+		return 0;
 
 	switch (a->attr_id) {
-	case attr_reserved_clusters:
-		return reserved_clusters_store(sbi, buf, len);
+	case attr_pointer_pi:
+		ret = kstrtouint(skip_spaces(buf), 0, &t);
+		if (ret)
+			return ret;
+		if ((int)t < 0)
+			return -EINVAL;
+		*((unsigned int *) ptr) = t;
+		return len;
 	case attr_pointer_ui:
-		if (!ptr)
-			return 0;
-		ret = kstrtoul(skip_spaces(buf), 0, &t);
+		ret = kstrtouint(skip_spaces(buf), 0, &t);
 		if (ret)
 			return ret;
 		if (a->attr_ptr == ptr_ext4_super_block_offset)
@@ -468,20 +474,50 @@ static ssize_t ext4_attr_store(struct kobject *kobj,
 		else
 			*((unsigned int *) ptr) = t;
 		return len;
+	case attr_mb_order:
+		ret = kstrtouint(skip_spaces(buf), 0, &t);
+		if (ret)
+			return ret;
+		if (t > 64)
+			return -EINVAL;
+		*((unsigned int *) ptr) = t;
+		return len;
+	case attr_clusters_in_group:
+		ret = kstrtouint(skip_spaces(buf), 0, &t);
+		if (ret)
+			return ret;
+		if (t > sbi->s_clusters_per_group)
+			return -EINVAL;
+		*((unsigned int *) ptr) = t;
+		return len;
 	case attr_pointer_ul:
-		if (!ptr)
-			return 0;
-		ret = kstrtoul(skip_spaces(buf), 0, &t);
+		ret = kstrtoul(skip_spaces(buf), 0, &lt);
 		if (ret)
 			return ret;
-		*((unsigned long *) ptr) = t;
+		*((unsigned long *) ptr) = lt;
 		return len;
+	}
+	return 0;
+}
+
+static ssize_t ext4_attr_store(struct kobject *kobj,
+			       struct attribute *attr,
+			       const char *buf, size_t len)
+{
+	struct ext4_sb_info *sbi = container_of(kobj, struct ext4_sb_info,
+						s_kobj);
+	struct ext4_attr *a = container_of(attr, struct ext4_attr, attr);
+
+	switch (a->attr_id) {
+	case attr_reserved_clusters:
+		return reserved_clusters_store(sbi, buf, len);
 	case attr_inode_readahead:
 		return inode_readahead_blks_store(sbi, buf, len);
 	case attr_trigger_test_error:
 		return trigger_test_error(sbi, buf, len);
+	default:
+		return ext4_generic_attr_store(a, sbi, buf, len);
 	}
-	return 0;
 }
 
 static void ext4_sb_release(struct kobject *kobj)
diff --git a/fs/ext4/xattr.c b/fs/ext4/xattr.c
index 82dc5e673d..46ce2f21fe 100644
--- a/fs/ext4/xattr.c
+++ b/fs/ext4/xattr.c
@@ -1433,6 +1433,12 @@ retry:
 			goto out;
 
 		memcpy(bh->b_data, buf, csize);
+		/*
+		 * Zero out block tail to avoid writing uninitialized memory
+		 * to disk.
+		 */
+		if (csize < blocksize)
+			memset(bh->b_data + csize, 0, blocksize - csize);
 		set_buffer_uptodate(bh);
 		ext4_handle_dirty_metadata(handle, ea_inode, bh);
 
@@ -1565,46 +1571,49 @@ ext4_xattr_inode_cache_find(struct inode *inode, const void *value,
 /*
  * Add value of the EA in an inode.
  */
-static int ext4_xattr_inode_lookup_create(handle_t *handle, struct inode *inode,
-					  const void *value, size_t value_len,
-					  struct inode **ret_inode)
+static struct inode *ext4_xattr_inode_lookup_create(handle_t *handle,
+		struct inode *inode, const void *value, size_t value_len)
 {
 	struct inode *ea_inode;
 	u32 hash;
 	int err;
 
+	/* Account inode & space to quota even if sharing... */
+	err = ext4_xattr_inode_alloc_quota(inode, value_len);
+	if (err)
+		return ERR_PTR(err);
+
 	hash = ext4_xattr_inode_hash(EXT4_SB(inode->i_sb), value, value_len);
 	ea_inode = ext4_xattr_inode_cache_find(inode, value, value_len, hash);
 	if (ea_inode) {
 		err = ext4_xattr_inode_inc_ref(handle, ea_inode);
-		if (err) {
-			iput(ea_inode);
-			return err;
-		}
-
-		*ret_inode = ea_inode;
-		return 0;
+		if (err)
+			goto out_err;
+		return ea_inode;
 	}
 
 	/* Create an inode for the EA value */
 	ea_inode = ext4_xattr_inode_create(handle, inode, hash);
-	if (IS_ERR(ea_inode))
-		return PTR_ERR(ea_inode);
+	if (IS_ERR(ea_inode)) {
+		ext4_xattr_inode_free_quota(inode, NULL, value_len);
+		return ea_inode;
+	}
 
 	err = ext4_xattr_inode_write(handle, ea_inode, value, value_len);
 	if (err) {
 		if (ext4_xattr_inode_dec_ref(handle, ea_inode))
 			ext4_warning_inode(ea_inode, "cleanup dec ref error %d", err);
-		iput(ea_inode);
-		return err;
+		goto out_err;
 	}
 
 	if (EA_INODE_CACHE(inode))
 		mb_cache_entry_create(EA_INODE_CACHE(inode), GFP_NOFS, hash,
 				      ea_inode->i_ino, true /* reusable */);
-
-	*ret_inode = ea_inode;
-	return 0;
+	return ea_inode;
+out_err:
+	iput(ea_inode);
+	ext4_xattr_inode_free_quota(inode, NULL, value_len);
+	return ERR_PTR(err);
 }
 
 /*
@@ -1616,6 +1625,7 @@ static int ext4_xattr_inode_lookup_create(handle_t *handle, struct inode *inode,
 static int ext4_xattr_set_entry(struct ext4_xattr_info *i,
 				struct ext4_xattr_search *s,
 				handle_t *handle, struct inode *inode,
+				struct inode *new_ea_inode,
 				bool is_block)
 {
 	struct ext4_xattr_entry *last, *next;
@@ -1623,7 +1633,6 @@ static int ext4_xattr_set_entry(struct ext4_xattr_info *i,
 	size_t min_offs = s->end - s->base, name_len = strlen(i->name);
 	int in_inode = i->in_inode;
 	struct inode *old_ea_inode = NULL;
-	struct inode *new_ea_inode = NULL;
 	size_t old_size, new_size;
 	int ret;
 
@@ -1708,43 +1717,11 @@ static int ext4_xattr_set_entry(struct ext4_xattr_info *i,
 			old_ea_inode = NULL;
 			goto out;
 		}
-	}
-	if (i->value && in_inode) {
-		WARN_ON_ONCE(!i->value_len);
 
-		ret = ext4_xattr_inode_alloc_quota(inode, i->value_len);
-		if (ret)
-			goto out;
-
-		ret = ext4_xattr_inode_lookup_create(handle, inode, i->value,
-						     i->value_len,
-						     &new_ea_inode);
-		if (ret) {
-			new_ea_inode = NULL;
-			ext4_xattr_inode_free_quota(inode, NULL, i->value_len);
-			goto out;
-		}
-	}
-
-	if (old_ea_inode) {
 		/* We are ready to release ref count on the old_ea_inode. */
 		ret = ext4_xattr_inode_dec_ref(handle, old_ea_inode);
-		if (ret) {
-			/* Release newly required ref count on new_ea_inode. */
-			if (new_ea_inode) {
-				int err;
-
-				err = ext4_xattr_inode_dec_ref(handle,
-							       new_ea_inode);
-				if (err)
-					ext4_warning_inode(new_ea_inode,
-						  "dec ref new_ea_inode err=%d",
-						  err);
-				ext4_xattr_inode_free_quota(inode, new_ea_inode,
-							    i->value_len);
-			}
+		if (ret)
 			goto out;
-		}
 
 		ext4_xattr_inode_free_quota(inode, old_ea_inode,
 					    le32_to_cpu(here->e_value_size));
@@ -1868,7 +1845,6 @@ update_hash:
 	ret = 0;
 out:
 	iput(old_ea_inode);
-	iput(new_ea_inode);
 	return ret;
 }
 
@@ -1931,9 +1907,21 @@ ext4_xattr_block_set(handle_t *handle, struct inode *inode,
 	size_t old_ea_inode_quota = 0;
 	unsigned int ea_ino;
 
-
 #define header(x) ((struct ext4_xattr_header *)(x))
 
+	/* If we need EA inode, prepare it before locking the buffer */
+	if (i->value && i->in_inode) {
+		WARN_ON_ONCE(!i->value_len);
+
+		ea_inode = ext4_xattr_inode_lookup_create(handle, inode,
+					i->value, i->value_len);
+		if (IS_ERR(ea_inode)) {
+			error = PTR_ERR(ea_inode);
+			ea_inode = NULL;
+			goto cleanup;
+		}
+	}
+
 	if (s->base) {
 		int offset = (char *)s->here - bs->bh->b_data;
 
@@ -1942,6 +1930,7 @@ ext4_xattr_block_set(handle_t *handle, struct inode *inode,
 						      EXT4_JTR_NONE);
 		if (error)
 			goto cleanup;
+
 		lock_buffer(bs->bh);
 
 		if (header(s->base)->h_refcount == cpu_to_le32(1)) {
@@ -1968,7 +1957,7 @@ ext4_xattr_block_set(handle_t *handle, struct inode *inode,
 			}
 			ea_bdebug(bs->bh, "modifying in-place");
 			error = ext4_xattr_set_entry(i, s, handle, inode,
-						     true /* is_block */);
+					     ea_inode, true /* is_block */);
 			ext4_xattr_block_csum_set(inode, bs->bh);
 			unlock_buffer(bs->bh);
 			if (error == -EFSCORRUPTED)
@@ -2036,33 +2025,22 @@ clone_block:
 		s->end = s->base + sb->s_blocksize;
 	}
 
-	error = ext4_xattr_set_entry(i, s, handle, inode, true /* is_block */);
+	error = ext4_xattr_set_entry(i, s, handle, inode, ea_inode,
+				     true /* is_block */);
 	if (error == -EFSCORRUPTED)
 		goto bad_block;
 	if (error)
 		goto cleanup;
 
-	if (i->value && s->here->e_value_inum) {
-		/*
-		 * A ref count on ea_inode has been taken as part of the call to
-		 * ext4_xattr_set_entry() above. We would like to drop this
-		 * extra ref but we have to wait until the xattr block is
-		 * initialized and has its own ref count on the ea_inode.
-		 */
-		ea_ino = le32_to_cpu(s->here->e_value_inum);
-		error = ext4_xattr_inode_iget(inode, ea_ino,
-					      le32_to_cpu(s->here->e_hash),
-					      &ea_inode);
-		if (error) {
-			ea_inode = NULL;
+inserted:
+	if (!IS_LAST_ENTRY(s->first)) {
+		new_bh = ext4_xattr_block_cache_find(inode, header(s->base), &ce);
+		if (IS_ERR(new_bh)) {
+			error = PTR_ERR(new_bh);
+			new_bh = NULL;
 			goto cleanup;
 		}
-	}
 
-inserted:
-	if (!IS_LAST_ENTRY(s->first)) {
-		new_bh = ext4_xattr_block_cache_find(inode, header(s->base),
-						     &ce);
 		if (new_bh) {
 			/* We found an identical block in the cache. */
 			if (new_bh == bs->bh)
@@ -2211,17 +2189,16 @@ getblk_failed:
 
 cleanup:
 	if (ea_inode) {
-		int error2;
-
-		error2 = ext4_xattr_inode_dec_ref(handle, ea_inode);
-		if (error2)
-			ext4_warning_inode(ea_inode, "dec ref error=%d",
-					   error2);
+		if (error) {
+			int error2;
 
-		/* If there was an error, revert the quota charge. */
-		if (error)
+			error2 = ext4_xattr_inode_dec_ref(handle, ea_inode);
+			if (error2)
+				ext4_warning_inode(ea_inode, "dec ref error=%d",
+						   error2);
 			ext4_xattr_inode_free_quota(inode, ea_inode,
 						    i_size_read(ea_inode));
+		}
 		iput(ea_inode);
 	}
 	if (ce)
@@ -2279,14 +2256,38 @@ int ext4_xattr_ibody_set(handle_t *handle, struct inode *inode,
 {
 	struct ext4_xattr_ibody_header *header;
 	struct ext4_xattr_search *s = &is->s;
+	struct inode *ea_inode = NULL;
 	int error;
 
 	if (!EXT4_INODE_HAS_XATTR_SPACE(inode))
 		return -ENOSPC;
 
-	error = ext4_xattr_set_entry(i, s, handle, inode, false /* is_block */);
-	if (error)
+	/* If we need EA inode, prepare it before locking the buffer */
+	if (i->value && i->in_inode) {
+		WARN_ON_ONCE(!i->value_len);
+
+		ea_inode = ext4_xattr_inode_lookup_create(handle, inode,
+					i->value, i->value_len);
+		if (IS_ERR(ea_inode))
+			return PTR_ERR(ea_inode);
+	}
+	error = ext4_xattr_set_entry(i, s, handle, inode, ea_inode,
+				     false /* is_block */);
+	if (error) {
+		if (ea_inode) {
+			int error2;
+
+			error2 = ext4_xattr_inode_dec_ref(handle, ea_inode);
+			if (error2)
+				ext4_warning_inode(ea_inode, "dec ref error=%d",
+						   error2);
+
+			ext4_xattr_inode_free_quota(inode, ea_inode,
+						    i_size_read(ea_inode));
+			iput(ea_inode);
+		}
 		return error;
+	}
 	header = IHDR(inode, ext4_raw_inode(&is->iloc));
 	if (!IS_LAST_ENTRY(s->first)) {
 		header->h_magic = cpu_to_le32(EXT4_XATTR_MAGIC);
@@ -2295,6 +2296,7 @@ int ext4_xattr_ibody_set(handle_t *handle, struct inode *inode,
 		header->h_magic = cpu_to_le32(0);
 		ext4_clear_inode_state(inode, EXT4_STATE_XATTR);
 	}
+	iput(ea_inode);
 	return 0;
 }
 
@@ -3103,8 +3105,8 @@ ext4_xattr_cmp(struct ext4_xattr_header *header1,
  *
  * Find an identical extended attribute block.
  *
- * Returns a pointer to the block found, or NULL if such a block was
- * not found or an error occurred.
+ * Returns a pointer to the block found, or NULL if such a block was not
+ * found, or an error pointer if an error occurred while reading ea block.
  */
 static struct buffer_head *
 ext4_xattr_block_cache_find(struct inode *inode,
@@ -3126,11 +3128,11 @@ ext4_xattr_block_cache_find(struct inode *inode,
 
 		bh = ext4_sb_bread(inode->i_sb, ce->e_value, REQ_PRIO);
 		if (IS_ERR(bh)) {
-			if (PTR_ERR(bh) == -ENOMEM)
-				return NULL;
-			bh = NULL;
-			EXT4_ERROR_INODE(inode, "block %lu read error",
-					 (unsigned long)ce->e_value);
+			if (PTR_ERR(bh) != -ENOMEM)
+				EXT4_ERROR_INODE(inode, "block %lu read error",
+						 (unsigned long)ce->e_value);
+			mb_cache_entry_put(ea_block_cache, ce);
+			return bh;
 		} else if (ext4_xattr_cmp(header, BHDR(bh)) == 0) {
 			*pce = ce;
 			return bh;
diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c
index 9afc8d24dc..b52d10b457 100644
--- a/fs/f2fs/checkpoint.c
+++ b/fs/f2fs/checkpoint.c
@@ -154,49 +154,47 @@ static bool __is_bitmap_valid(struct f2fs_sb_info *sbi, block_t blkaddr,
 	if (unlikely(f2fs_cp_error(sbi)))
 		return exist;
 
-	if (exist && type == DATA_GENERIC_ENHANCE_UPDATE) {
-		f2fs_err(sbi, "Inconsistent error blkaddr:%u, sit bitmap:%d",
-			 blkaddr, exist);
-		set_sbi_flag(sbi, SBI_NEED_FSCK);
-		return exist;
-	}
+	if ((exist && type == DATA_GENERIC_ENHANCE_UPDATE) ||
+		(!exist && type == DATA_GENERIC_ENHANCE))
+		goto out_err;
+	if (!exist && type != DATA_GENERIC_ENHANCE_UPDATE)
+		goto out_handle;
+	return exist;
 
-	if (!exist && type == DATA_GENERIC_ENHANCE) {
-		f2fs_err(sbi, "Inconsistent error blkaddr:%u, sit bitmap:%d",
-			 blkaddr, exist);
-		set_sbi_flag(sbi, SBI_NEED_FSCK);
-		dump_stack();
-	}
+out_err:
+	f2fs_err(sbi, "Inconsistent error blkaddr:%u, sit bitmap:%d",
+		 blkaddr, exist);
+	set_sbi_flag(sbi, SBI_NEED_FSCK);
+	dump_stack();
+out_handle:
+	f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR);
 	return exist;
 }
 
-bool f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi,
+static bool __f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi,
 					block_t blkaddr, int type)
 {
-	if (time_to_inject(sbi, FAULT_BLKADDR))
-		return false;
-
 	switch (type) {
 	case META_NAT:
 		break;
 	case META_SIT:
 		if (unlikely(blkaddr >= SIT_BLK_CNT(sbi)))
-			return false;
+			goto check_only;
 		break;
 	case META_SSA:
 		if (unlikely(blkaddr >= MAIN_BLKADDR(sbi) ||
 			blkaddr < SM_I(sbi)->ssa_blkaddr))
-			return false;
+			goto check_only;
 		break;
 	case META_CP:
 		if (unlikely(blkaddr >= SIT_I(sbi)->sit_base_addr ||
 			blkaddr < __start_cp_addr(sbi)))
-			return false;
+			goto check_only;
 		break;
 	case META_POR:
 		if (unlikely(blkaddr >= MAX_BLKADDR(sbi) ||
 			blkaddr < MAIN_BLKADDR(sbi)))
-			return false;
+			goto check_only;
 		break;
 	case DATA_GENERIC:
 	case DATA_GENERIC_ENHANCE:
@@ -213,7 +211,7 @@ bool f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi,
 				  blkaddr);
 			set_sbi_flag(sbi, SBI_NEED_FSCK);
 			dump_stack();
-			return false;
+			goto err;
 		} else {
 			return __is_bitmap_valid(sbi, blkaddr, type);
 		}
@@ -221,13 +219,31 @@ bool f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi,
 	case META_GENERIC:
 		if (unlikely(blkaddr < SEG0_BLKADDR(sbi) ||
 			blkaddr >= MAIN_BLKADDR(sbi)))
-			return false;
+			goto err;
 		break;
 	default:
 		BUG();
 	}
 
 	return true;
+err:
+	f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR);
+check_only:
+	return false;
+}
+
+bool f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi,
+					block_t blkaddr, int type)
+{
+	if (time_to_inject(sbi, FAULT_BLKADDR_VALIDITY))
+		return false;
+	return __f2fs_is_valid_blkaddr(sbi, blkaddr, type);
+}
+
+bool f2fs_is_valid_blkaddr_raw(struct f2fs_sb_info *sbi,
+					block_t blkaddr, int type)
+{
+	return __f2fs_is_valid_blkaddr(sbi, blkaddr, type);
 }
 
 /*
@@ -330,7 +346,7 @@ static int __f2fs_write_meta_page(struct page *page,
 {
 	struct f2fs_sb_info *sbi = F2FS_P_SB(page);
 
-	trace_f2fs_writepage(page, META);
+	trace_f2fs_writepage(page_folio(page), META);
 
 	if (unlikely(f2fs_cp_error(sbi))) {
 		if (is_sbi_flag_set(sbi, SBI_IS_CLOSE)) {
@@ -477,7 +493,7 @@ stop:
 static bool f2fs_dirty_meta_folio(struct address_space *mapping,
 		struct folio *folio)
 {
-	trace_f2fs_set_page_dirty(&folio->page, META);
+	trace_f2fs_set_page_dirty(folio, META);
 
 	if (!folio_test_uptodate(folio))
 		folio_mark_uptodate(folio);
@@ -889,7 +905,7 @@ static struct page *validate_checkpoint(struct f2fs_sb_info *sbi,
 
 	cp_blocks = le32_to_cpu(cp_block->cp_pack_total_block_count);
 
-	if (cp_blocks > sbi->blocks_per_seg || cp_blocks <= F2FS_CP_PACKS) {
+	if (cp_blocks > BLKS_PER_SEG(sbi) || cp_blocks <= F2FS_CP_PACKS) {
 		f2fs_warn(sbi, "invalid cp_pack_total_block_count:%u",
 			  le32_to_cpu(cp_block->cp_pack_total_block_count));
 		goto invalid_cp;
@@ -1170,6 +1186,11 @@ static void __prepare_cp_block(struct f2fs_sb_info *sbi)
 	ckpt->valid_node_count = cpu_to_le32(valid_node_count(sbi));
 	ckpt->valid_inode_count = cpu_to_le32(valid_inode_count(sbi));
 	ckpt->next_free_nid = cpu_to_le32(last_nid);
+
+	/* update user_block_counts */
+	sbi->last_valid_block_count = sbi->total_valid_block_count;
+	percpu_counter_set(&sbi->alloc_valid_block_count, 0);
+	percpu_counter_set(&sbi->rf_node_block_count, 0);
 }
 
 static bool __need_flush_quota(struct f2fs_sb_info *sbi)
@@ -1324,7 +1345,7 @@ static void update_ckpt_flags(struct f2fs_sb_info *sbi, struct cp_control *cpc)
 
 	if (cpc->reason & CP_UMOUNT) {
 		if (le32_to_cpu(ckpt->cp_pack_total_block_count) +
-			NM_I(sbi)->nat_bits_blocks > sbi->blocks_per_seg) {
+			NM_I(sbi)->nat_bits_blocks > BLKS_PER_SEG(sbi)) {
 			clear_ckpt_flags(sbi, CP_NAT_BITS_FLAG);
 			f2fs_notice(sbi, "Disable nat_bits due to no space");
 		} else if (!is_set_ckpt_flags(sbi, CP_NAT_BITS_FLAG) &&
@@ -1527,7 +1548,7 @@ static int do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc)
 		cp_ver |= ((__u64)crc32 << 32);
 		*(__le64 *)nm_i->nat_bits = cpu_to_le64(cp_ver);
 
-		blk = start_blk + sbi->blocks_per_seg - nm_i->nat_bits_blocks;
+		blk = start_blk + BLKS_PER_SEG(sbi) - nm_i->nat_bits_blocks;
 		for (i = 0; i < nm_i->nat_bits_blocks; i++)
 			f2fs_update_meta_page(sbi, nm_i->nat_bits +
 					(i << F2FS_BLKSIZE_BITS), blk + i);
@@ -1559,11 +1580,6 @@ static int do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc)
 		start_blk += NR_CURSEG_NODE_TYPE;
 	}
 
-	/* update user_block_counts */
-	sbi->last_valid_block_count = sbi->total_valid_block_count;
-	percpu_counter_set(&sbi->alloc_valid_block_count, 0);
-	percpu_counter_set(&sbi->rf_node_block_count, 0);
-
 	/* Here, we have one bio having CP pack except cp pack 2 page */
 	f2fs_sync_meta_pages(sbi, META, LONG_MAX, FS_CP_META_IO);
 	/* Wait for all dirty meta pages to be submitted for IO */
@@ -1731,9 +1747,9 @@ void f2fs_init_ino_entry_info(struct f2fs_sb_info *sbi)
 		im->ino_num = 0;
 	}
 
-	sbi->max_orphans = (sbi->blocks_per_seg - F2FS_CP_PACKS -
+	sbi->max_orphans = (BLKS_PER_SEG(sbi) - F2FS_CP_PACKS -
 			NR_CURSEG_PERSIST_TYPE - __cp_payload(sbi)) *
-				F2FS_ORPHANS_PER_BLOCK;
+			F2FS_ORPHANS_PER_BLOCK;
 }
 
 int __init f2fs_create_checkpoint_caches(void)
diff --git a/fs/f2fs/compress.c b/fs/f2fs/compress.c
index 15f330d8fd..1ef82a5463 100644
--- a/fs/f2fs/compress.c
+++ b/fs/f2fs/compress.c
@@ -198,8 +198,8 @@ static int lzo_compress_pages(struct compress_ctx *cc)
 	ret = lzo1x_1_compress(cc->rbuf, cc->rlen, cc->cbuf->cdata,
 					&cc->clen, cc->private);
 	if (ret != LZO_E_OK) {
-		printk_ratelimited("%sF2FS-fs (%s): lzo compress failed, ret:%d\n",
-				KERN_ERR, F2FS_I_SB(cc->inode)->sb->s_id, ret);
+		f2fs_err_ratelimited(F2FS_I_SB(cc->inode),
+				"lzo compress failed, ret:%d", ret);
 		return -EIO;
 	}
 	return 0;
@@ -212,17 +212,15 @@ static int lzo_decompress_pages(struct decompress_io_ctx *dic)
 	ret = lzo1x_decompress_safe(dic->cbuf->cdata, dic->clen,
 						dic->rbuf, &dic->rlen);
 	if (ret != LZO_E_OK) {
-		printk_ratelimited("%sF2FS-fs (%s): lzo decompress failed, ret:%d\n",
-				KERN_ERR, F2FS_I_SB(dic->inode)->sb->s_id, ret);
+		f2fs_err_ratelimited(F2FS_I_SB(dic->inode),
+				"lzo decompress failed, ret:%d", ret);
 		return -EIO;
 	}
 
 	if (dic->rlen != PAGE_SIZE << dic->log_cluster_size) {
-		printk_ratelimited("%sF2FS-fs (%s): lzo invalid rlen:%zu, "
-					"expected:%lu\n", KERN_ERR,
-					F2FS_I_SB(dic->inode)->sb->s_id,
-					dic->rlen,
-					PAGE_SIZE << dic->log_cluster_size);
+		f2fs_err_ratelimited(F2FS_I_SB(dic->inode),
+				"lzo invalid rlen:%zu, expected:%lu",
+				dic->rlen, PAGE_SIZE << dic->log_cluster_size);
 		return -EIO;
 	}
 	return 0;
@@ -294,16 +292,15 @@ static int lz4_decompress_pages(struct decompress_io_ctx *dic)
 	ret = LZ4_decompress_safe(dic->cbuf->cdata, dic->rbuf,
 						dic->clen, dic->rlen);
 	if (ret < 0) {
-		printk_ratelimited("%sF2FS-fs (%s): lz4 decompress failed, ret:%d\n",
-				KERN_ERR, F2FS_I_SB(dic->inode)->sb->s_id, ret);
+		f2fs_err_ratelimited(F2FS_I_SB(dic->inode),
+				"lz4 decompress failed, ret:%d", ret);
 		return -EIO;
 	}
 
 	if (ret != PAGE_SIZE << dic->log_cluster_size) {
-		printk_ratelimited("%sF2FS-fs (%s): lz4 invalid ret:%d, "
-					"expected:%lu\n", KERN_ERR,
-					F2FS_I_SB(dic->inode)->sb->s_id, ret,
-					PAGE_SIZE << dic->log_cluster_size);
+		f2fs_err_ratelimited(F2FS_I_SB(dic->inode),
+				"lz4 invalid ret:%d, expected:%lu",
+				ret, PAGE_SIZE << dic->log_cluster_size);
 		return -EIO;
 	}
 	return 0;
@@ -350,9 +347,8 @@ static int zstd_init_compress_ctx(struct compress_ctx *cc)
 
 	stream = zstd_init_cstream(&params, 0, workspace, workspace_size);
 	if (!stream) {
-		printk_ratelimited("%sF2FS-fs (%s): %s zstd_init_cstream failed\n",
-				KERN_ERR, F2FS_I_SB(cc->inode)->sb->s_id,
-				__func__);
+		f2fs_err_ratelimited(F2FS_I_SB(cc->inode),
+				"%s zstd_init_cstream failed", __func__);
 		kvfree(workspace);
 		return -EIO;
 	}
@@ -390,16 +386,16 @@ static int zstd_compress_pages(struct compress_ctx *cc)
 
 	ret = zstd_compress_stream(stream, &outbuf, &inbuf);
 	if (zstd_is_error(ret)) {
-		printk_ratelimited("%sF2FS-fs (%s): %s zstd_compress_stream failed, ret: %d\n",
-				KERN_ERR, F2FS_I_SB(cc->inode)->sb->s_id,
+		f2fs_err_ratelimited(F2FS_I_SB(cc->inode),
+				"%s zstd_compress_stream failed, ret: %d",
 				__func__, zstd_get_error_code(ret));
 		return -EIO;
 	}
 
 	ret = zstd_end_stream(stream, &outbuf);
 	if (zstd_is_error(ret)) {
-		printk_ratelimited("%sF2FS-fs (%s): %s zstd_end_stream returned %d\n",
-				KERN_ERR, F2FS_I_SB(cc->inode)->sb->s_id,
+		f2fs_err_ratelimited(F2FS_I_SB(cc->inode),
+				"%s zstd_end_stream returned %d",
 				__func__, zstd_get_error_code(ret));
 		return -EIO;
 	}
@@ -432,9 +428,8 @@ static int zstd_init_decompress_ctx(struct decompress_io_ctx *dic)
 
 	stream = zstd_init_dstream(max_window_size, workspace, workspace_size);
 	if (!stream) {
-		printk_ratelimited("%sF2FS-fs (%s): %s zstd_init_dstream failed\n",
-				KERN_ERR, F2FS_I_SB(dic->inode)->sb->s_id,
-				__func__);
+		f2fs_err_ratelimited(F2FS_I_SB(dic->inode),
+				"%s zstd_init_dstream failed", __func__);
 		kvfree(workspace);
 		return -EIO;
 	}
@@ -469,16 +464,15 @@ static int zstd_decompress_pages(struct decompress_io_ctx *dic)
 
 	ret = zstd_decompress_stream(stream, &outbuf, &inbuf);
 	if (zstd_is_error(ret)) {
-		printk_ratelimited("%sF2FS-fs (%s): %s zstd_decompress_stream failed, ret: %d\n",
-				KERN_ERR, F2FS_I_SB(dic->inode)->sb->s_id,
+		f2fs_err_ratelimited(F2FS_I_SB(dic->inode),
+				"%s zstd_decompress_stream failed, ret: %d",
 				__func__, zstd_get_error_code(ret));
 		return -EIO;
 	}
 
 	if (dic->rlen != outbuf.pos) {
-		printk_ratelimited("%sF2FS-fs (%s): %s ZSTD invalid rlen:%zu, "
-				"expected:%lu\n", KERN_ERR,
-				F2FS_I_SB(dic->inode)->sb->s_id,
+		f2fs_err_ratelimited(F2FS_I_SB(dic->inode),
+				"%s ZSTD invalid rlen:%zu, expected:%lu",
 				__func__, dic->rlen,
 				PAGE_SIZE << dic->log_cluster_size);
 		return -EIO;
@@ -512,8 +506,8 @@ static int lzorle_compress_pages(struct compress_ctx *cc)
 	ret = lzorle1x_1_compress(cc->rbuf, cc->rlen, cc->cbuf->cdata,
 					&cc->clen, cc->private);
 	if (ret != LZO_E_OK) {
-		printk_ratelimited("%sF2FS-fs (%s): lzo-rle compress failed, ret:%d\n",
-				KERN_ERR, F2FS_I_SB(cc->inode)->sb->s_id, ret);
+		f2fs_err_ratelimited(F2FS_I_SB(cc->inode),
+				"lzo-rle compress failed, ret:%d", ret);
 		return -EIO;
 	}
 	return 0;
@@ -780,9 +774,9 @@ void f2fs_decompress_cluster(struct decompress_io_ctx *dic, bool in_task)
 		if (provided != calculated) {
 			if (!is_inode_flag_set(dic->inode, FI_COMPRESS_CORRUPT)) {
 				set_inode_flag(dic->inode, FI_COMPRESS_CORRUPT);
-				printk_ratelimited(
-					"%sF2FS-fs (%s): checksum invalid, nid = %lu, %x vs %x",
-					KERN_INFO, sbi->sb->s_id, dic->inode->i_ino,
+				f2fs_info_ratelimited(sbi,
+					"checksum invalid, nid = %lu, %x vs %x",
+					dic->inode->i_ino,
 					provided, calculated);
 			}
 			set_sbi_flag(sbi, SBI_NEED_FSCK);
@@ -1031,6 +1025,31 @@ static void set_cluster_writeback(struct compress_ctx *cc)
 	}
 }
 
+static void cancel_cluster_writeback(struct compress_ctx *cc,
+			struct compress_io_ctx *cic, int submitted)
+{
+	int i;
+
+	/* Wait for submitted IOs. */
+	if (submitted > 1) {
+		f2fs_submit_merged_write(F2FS_I_SB(cc->inode), DATA);
+		while (atomic_read(&cic->pending_pages) !=
+					(cc->valid_nr_cpages - submitted + 1))
+			f2fs_io_schedule_timeout(DEFAULT_IO_TIMEOUT);
+	}
+
+	/* Cancel writeback and stay locked. */
+	for (i = 0; i < cc->cluster_size; i++) {
+		if (i < submitted) {
+			inode_inc_dirty_pages(cc->inode);
+			lock_page(cc->rpages[i]);
+		}
+		clear_page_private_gcing(cc->rpages[i]);
+		if (folio_test_writeback(page_folio(cc->rpages[i])))
+			end_page_writeback(cc->rpages[i]);
+	}
+}
+
 static void set_cluster_dirty(struct compress_ctx *cc)
 {
 	int i;
@@ -1232,7 +1251,6 @@ static int f2fs_write_compressed_pages(struct compress_ctx *cc,
 		.page = NULL,
 		.encrypted_page = NULL,
 		.compressed_page = NULL,
-		.submitted = 0,
 		.io_type = io_type,
 		.io_wbc = wbc,
 		.encrypted = fscrypt_inode_uses_fs_layer_crypto(cc->inode) ?
@@ -1358,7 +1376,16 @@ static int f2fs_write_compressed_pages(struct compress_ctx *cc,
 			fio.compressed_page = cc->cpages[i - 1];
 
 		cc->cpages[i - 1] = NULL;
+		fio.submitted = 0;
 		f2fs_outplace_write_data(&dn, &fio);
+		if (unlikely(!fio.submitted)) {
+			cancel_cluster_writeback(cc, cic, i);
+
+			/* To call fscrypt_finalize_bounce_page */
+			i = cc->valid_nr_cpages;
+			*submitted = 0;
+			goto out_destroy_crypt;
+		}
 		(*submitted)++;
 unlock_continue:
 		inode_dec_dirty_pages(cc->inode);
@@ -1392,8 +1419,11 @@ unlock_continue:
 out_destroy_crypt:
 	page_array_free(cc->inode, cic->rpages, cc->cluster_size);
 
-	for (--i; i >= 0; i--)
+	for (--i; i >= 0; i--) {
+		if (!cc->cpages[i])
+			continue;
 		fscrypt_finalize_bounce_page(&cc->cpages[i]);
+	}
 out_put_cic:
 	kmem_cache_free(cic_entry_slab, cic);
 out_put_dnode:
@@ -1484,7 +1514,7 @@ continue_unlock:
 		if (!PageDirty(cc->rpages[i]))
 			goto continue_unlock;
 
-		if (PageWriteback(cc->rpages[i])) {
+		if (folio_test_writeback(page_folio(cc->rpages[i]))) {
 			if (wbc->sync_mode == WB_SYNC_NONE)
 				goto continue_unlock;
 			f2fs_wait_on_page_writeback(cc->rpages[i], DATA, true, true);
@@ -1891,12 +1921,8 @@ void f2fs_cache_compressed_page(struct f2fs_sb_info *sbi, struct page *page,
 
 	set_page_private_data(cpage, ino);
 
-	if (!f2fs_is_valid_blkaddr(sbi, blkaddr, DATA_GENERIC_ENHANCE_READ))
-		goto out;
-
 	memcpy(page_address(cpage), page_address(page), PAGE_SIZE);
 	SetPageUptodate(cpage);
-out:
 	f2fs_put_page(cpage, 1);
 }
 
diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
index 05158f89ef..467f67cf2b 100644
--- a/fs/f2fs/data.c
+++ b/fs/f2fs/data.c
@@ -338,17 +338,6 @@ static void f2fs_write_end_io(struct bio *bio)
 		struct page *page = bvec->bv_page;
 		enum count_type type = WB_DATA_TYPE(page, false);
 
-		if (page_private_dummy(page)) {
-			clear_page_private_dummy(page);
-			unlock_page(page);
-			mempool_free(page, sbi->write_io_dummy);
-
-			if (unlikely(bio->bi_status))
-				f2fs_stop_checkpoint(sbi, true,
-						STOP_CP_REASON_WRITE_FAIL);
-			continue;
-		}
-
 		fscrypt_finalize_bounce_page(&page);
 
 #ifdef CONFIG_F2FS_FS_COMPRESSION
@@ -476,6 +465,8 @@ static struct bio *__bio_alloc(struct f2fs_io_info *fio, int npages)
 	} else {
 		bio->bi_end_io = f2fs_write_end_io;
 		bio->bi_private = sbi;
+		bio->bi_write_hint = f2fs_io_type_to_rw_hint(sbi,
+						fio->type, fio->temp);
 	}
 	iostat_alloc_and_bind_ctx(sbi, bio, NULL);
 
@@ -522,50 +513,13 @@ void f2fs_submit_read_bio(struct f2fs_sb_info *sbi, struct bio *bio,
 	submit_bio(bio);
 }
 
-static void f2fs_align_write_bio(struct f2fs_sb_info *sbi, struct bio *bio)
-{
-	unsigned int start =
-		(bio->bi_iter.bi_size >> F2FS_BLKSIZE_BITS) % F2FS_IO_SIZE(sbi);
-
-	if (start == 0)
-		return;
-
-	/* fill dummy pages */
-	for (; start < F2FS_IO_SIZE(sbi); start++) {
-		struct page *page =
-			mempool_alloc(sbi->write_io_dummy,
-				      GFP_NOIO | __GFP_NOFAIL);
-		f2fs_bug_on(sbi, !page);
-
-		lock_page(page);
-
-		zero_user_segment(page, 0, PAGE_SIZE);
-		set_page_private_dummy(page);
-
-		if (bio_add_page(bio, page, PAGE_SIZE, 0) < PAGE_SIZE)
-			f2fs_bug_on(sbi, 1);
-	}
-}
-
 static void f2fs_submit_write_bio(struct f2fs_sb_info *sbi, struct bio *bio,
 				  enum page_type type)
 {
 	WARN_ON_ONCE(is_read_io(bio_op(bio)));
 
-	if (type == DATA || type == NODE) {
-		if (f2fs_lfs_mode(sbi) && current->plug)
-			blk_finish_plug(current->plug);
-
-		if (F2FS_IO_ALIGNED(sbi)) {
-			f2fs_align_write_bio(sbi, bio);
-			/*
-			 * In the NODE case, we lose next block address chain.
-			 * So, we need to do checkpoint in f2fs_sync_file.
-			 */
-			if (type == NODE)
-				set_sbi_flag(sbi, SBI_NEED_CP);
-		}
-	}
+	if (f2fs_lfs_mode(sbi) && current->plug && PAGE_TYPE_ON_MAIN(type))
+		blk_finish_plug(current->plug);
 
 	trace_f2fs_submit_write_bio(sbi->sb, type, bio);
 	iostat_update_submit_ctx(bio, type);
@@ -641,17 +595,20 @@ int f2fs_init_write_merge_io(struct f2fs_sb_info *sbi)
 			return -ENOMEM;
 
 		for (j = HOT; j < n; j++) {
-			init_f2fs_rwsem(&sbi->write_io[i][j].io_rwsem);
-			sbi->write_io[i][j].sbi = sbi;
-			sbi->write_io[i][j].bio = NULL;
-			spin_lock_init(&sbi->write_io[i][j].io_lock);
-			INIT_LIST_HEAD(&sbi->write_io[i][j].io_list);
-			INIT_LIST_HEAD(&sbi->write_io[i][j].bio_list);
-			init_f2fs_rwsem(&sbi->write_io[i][j].bio_list_lock);
+			struct f2fs_bio_info *io = &sbi->write_io[i][j];
+
+			init_f2fs_rwsem(&io->io_rwsem);
+			io->sbi = sbi;
+			io->bio = NULL;
+			io->last_block_in_bio = 0;
+			spin_lock_init(&io->io_lock);
+			INIT_LIST_HEAD(&io->io_list);
+			INIT_LIST_HEAD(&io->bio_list);
+			init_f2fs_rwsem(&io->bio_list_lock);
 #ifdef CONFIG_BLK_DEV_ZONED
-			init_completion(&sbi->write_io[i][j].zone_wait);
-			sbi->write_io[i][j].zone_pending_bio = NULL;
-			sbi->write_io[i][j].bi_private = NULL;
+			init_completion(&io->zone_wait);
+			io->zone_pending_bio = NULL;
+			io->bi_private = NULL;
 #endif
 		}
 	}
@@ -738,10 +695,8 @@ int f2fs_submit_page_bio(struct f2fs_io_info *fio)
 
 	if (!f2fs_is_valid_blkaddr(fio->sbi, fio->new_blkaddr,
 			fio->is_por ? META_POR : (__is_meta_io(fio) ?
-			META_GENERIC : DATA_GENERIC_ENHANCE))) {
-		f2fs_handle_error(fio->sbi, ERROR_INVALID_BLKADDR);
+			META_GENERIC : DATA_GENERIC_ENHANCE)))
 		return -EFSCORRUPTED;
-	}
 
 	trace_f2fs_submit_page_bio(page, fio);
 
@@ -794,16 +749,6 @@ static bool io_is_mergeable(struct f2fs_sb_info *sbi, struct bio *bio,
 					block_t last_blkaddr,
 					block_t cur_blkaddr)
 {
-	if (F2FS_IO_ALIGNED(sbi) && (fio->type == DATA || fio->type == NODE)) {
-		unsigned int filled_blocks =
-				F2FS_BYTES_TO_BLK(bio->bi_iter.bi_size);
-		unsigned int io_size = F2FS_IO_SIZE(sbi);
-		unsigned int left_vecs = bio->bi_max_vecs - bio->bi_vcnt;
-
-		/* IOs in bio is aligned and left space of vectors is not enough */
-		if (!(filled_blocks % io_size) && left_vecs < io_size)
-			return false;
-	}
 	if (!page_is_mergeable(sbi, bio, last_blkaddr, cur_blkaddr))
 		return false;
 	return io_type_is_mergeable(io, fio);
@@ -946,10 +891,8 @@ int f2fs_merge_page_bio(struct f2fs_io_info *fio)
 			fio->encrypted_page : fio->page;
 
 	if (!f2fs_is_valid_blkaddr(fio->sbi, fio->new_blkaddr,
-			__is_meta_io(fio) ? META_GENERIC : DATA_GENERIC)) {
-		f2fs_handle_error(fio->sbi, ERROR_INVALID_BLKADDR);
+			__is_meta_io(fio) ? META_GENERIC : DATA_GENERIC))
 		return -EFSCORRUPTED;
-	}
 
 	trace_f2fs_submit_page_bio(page, fio);
 
@@ -982,6 +925,7 @@ alloc_new:
 #ifdef CONFIG_BLK_DEV_ZONED
 static bool is_end_zone_blkaddr(struct f2fs_sb_info *sbi, block_t blkaddr)
 {
+	struct block_device *bdev = sbi->sb->s_bdev;
 	int devi = 0;
 
 	if (f2fs_is_multi_device(sbi)) {
@@ -992,8 +936,9 @@ static bool is_end_zone_blkaddr(struct f2fs_sb_info *sbi, block_t blkaddr)
 			return false;
 		}
 		blkaddr -= FDEV(devi).start_blk;
+		bdev = FDEV(devi).bdev;
 	}
-	return bdev_is_zoned(FDEV(devi).bdev) &&
+	return bdev_is_zoned(bdev) &&
 		f2fs_blkz_is_seq(sbi, devi, blkaddr) &&
 		(blkaddr % sbi->blocks_per_blkz == sbi->blocks_per_blkz - 1);
 }
@@ -1055,14 +1000,6 @@ next:
 		__submit_merged_bio(io);
 alloc_new:
 	if (io->bio == NULL) {
-		if (F2FS_IO_ALIGNED(sbi) &&
-				(fio->type == DATA || fio->type == NODE) &&
-				fio->new_blkaddr & F2FS_IO_SIZE_MASK(sbi)) {
-			dec_page_count(sbi, WB_DATA_TYPE(bio_page,
-						fio->compressed_page));
-			fio->retry = 1;
-			goto skip;
-		}
 		io->bio = __bio_alloc(fio, BIO_MAX_VECS);
 		f2fs_set_bio_crypt_ctx(io->bio, fio->page->mapping->host,
 				       bio_page->index, fio, GFP_NOIO);
@@ -1092,7 +1029,6 @@ alloc_new:
 		__submit_merged_bio(io);
 	}
 #endif
-skip:
 	if (fio->in_list)
 		goto next;
 out:
@@ -1286,8 +1222,6 @@ struct page *f2fs_get_read_data_page(struct inode *inode, pgoff_t index,
 		if (!f2fs_is_valid_blkaddr(F2FS_I_SB(inode), dn.data_blkaddr,
 						DATA_GENERIC_ENHANCE_READ)) {
 			err = -EFSCORRUPTED;
-			f2fs_handle_error(F2FS_I_SB(inode),
-						ERROR_INVALID_BLKADDR);
 			goto put_err;
 		}
 		goto got_it;
@@ -1313,8 +1247,6 @@ struct page *f2fs_get_read_data_page(struct inode *inode, pgoff_t index,
 						dn.data_blkaddr,
 						DATA_GENERIC_ENHANCE)) {
 		err = -EFSCORRUPTED;
-		f2fs_handle_error(F2FS_I_SB(inode),
-					ERROR_INVALID_BLKADDR);
 		goto put_err;
 	}
 got_it:
@@ -1483,8 +1415,11 @@ static int __allocate_data_block(struct dnode_of_data *dn, int seg_type)
 
 	set_summary(&sum, dn->nid, dn->ofs_in_node, ni.version);
 	old_blkaddr = dn->data_blkaddr;
-	f2fs_allocate_data_block(sbi, NULL, old_blkaddr, &dn->data_blkaddr,
-				&sum, seg_type, NULL);
+	err = f2fs_allocate_data_block(sbi, NULL, old_blkaddr,
+				&dn->data_blkaddr, &sum, seg_type, NULL);
+	if (err)
+		return err;
+
 	if (GET_SEGNO(sbi, old_blkaddr) != NULL_SEGNO)
 		f2fs_invalidate_internal_cache(sbi, old_blkaddr);
 
@@ -1579,6 +1514,25 @@ static bool f2fs_map_blocks_cached(struct inode *inode,
 	return true;
 }
 
+static bool map_is_mergeable(struct f2fs_sb_info *sbi,
+				struct f2fs_map_blocks *map,
+				block_t blkaddr, int flag, int bidx,
+				int ofs)
+{
+	if (map->m_multidev_dio && map->m_bdev != FDEV(bidx).bdev)
+		return false;
+	if (map->m_pblk != NEW_ADDR && blkaddr == (map->m_pblk + ofs))
+		return true;
+	if (map->m_pblk == NEW_ADDR && blkaddr == NEW_ADDR)
+		return true;
+	if (flag == F2FS_GET_BLOCK_PRE_DIO)
+		return true;
+	if (flag == F2FS_GET_BLOCK_DIO &&
+		map->m_pblk == NULL_ADDR && blkaddr == NULL_ADDR)
+		return true;
+	return false;
+}
+
 /*
  * f2fs_map_blocks() tries to find or build mapping relationship which
  * maps continuous logical blocks to physical blocks, and return such
@@ -1642,13 +1596,13 @@ next_block:
 	if (!is_hole &&
 	    !f2fs_is_valid_blkaddr(sbi, blkaddr, DATA_GENERIC_ENHANCE)) {
 		err = -EFSCORRUPTED;
-		f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR);
 		goto sync_out;
 	}
 
 	/* use out-place-update for direct IO under LFS mode */
-	if (map->m_may_create &&
-	    (is_hole || (f2fs_lfs_mode(sbi) && flag == F2FS_GET_BLOCK_DIO))) {
+	if (map->m_may_create && (is_hole ||
+		(flag == F2FS_GET_BLOCK_DIO && f2fs_lfs_mode(sbi) &&
+		!f2fs_is_pinned_file(inode)))) {
 		if (unlikely(f2fs_cp_error(sbi))) {
 			err = -EIO;
 			goto sync_out;
@@ -1701,6 +1655,10 @@ next_block:
 				goto sync_out;
 			}
 			break;
+		case F2FS_GET_BLOCK_DIO:
+			if (map->m_next_pgofs)
+				*map->m_next_pgofs = pgofs + 1;
+			break;
 		default:
 			/* for defragment case */
 			if (map->m_next_pgofs)
@@ -1719,19 +1677,15 @@ next_block:
 		/* reserved delalloc block should be mapped for fiemap. */
 		if (blkaddr == NEW_ADDR)
 			map->m_flags |= F2FS_MAP_DELALLOC;
-		map->m_flags |= F2FS_MAP_MAPPED;
+		if (flag != F2FS_GET_BLOCK_DIO || !is_hole)
+			map->m_flags |= F2FS_MAP_MAPPED;
 
 		map->m_pblk = blkaddr;
 		map->m_len = 1;
 
 		if (map->m_multidev_dio)
 			map->m_bdev = FDEV(bidx).bdev;
-	} else if ((map->m_pblk != NEW_ADDR &&
-			blkaddr == (map->m_pblk + ofs)) ||
-			(map->m_pblk == NEW_ADDR && blkaddr == NEW_ADDR) ||
-			flag == F2FS_GET_BLOCK_PRE_DIO) {
-		if (map->m_multidev_dio && map->m_bdev != FDEV(bidx).bdev)
-			goto sync_out;
+	} else if (map_is_mergeable(sbi, map, blkaddr, flag, bidx, ofs)) {
 		ofs++;
 		map->m_len++;
 	} else {
@@ -2115,7 +2069,7 @@ static inline loff_t f2fs_readpage_limit(struct inode *inode)
 	return i_size_read(inode);
 }
 
-static int f2fs_read_single_page(struct inode *inode, struct page *page,
+static int f2fs_read_single_page(struct inode *inode, struct folio *folio,
 					unsigned nr_pages,
 					struct f2fs_map_blocks *map,
 					struct bio **bio_ret,
@@ -2128,9 +2082,10 @@ static int f2fs_read_single_page(struct inode *inode, struct page *page,
 	sector_t last_block;
 	sector_t last_block_in_file;
 	sector_t block_nr;
+	pgoff_t index = folio_index(folio);
 	int ret = 0;
 
-	block_in_file = (sector_t)page_index(page);
+	block_in_file = (sector_t)index;
 	last_block = block_in_file + nr_pages;
 	last_block_in_file = bytes_to_blks(inode,
 			f2fs_readpage_limit(inode) + blocksize - 1);
@@ -2161,26 +2116,24 @@ static int f2fs_read_single_page(struct inode *inode, struct page *page,
 got_it:
 	if ((map->m_flags & F2FS_MAP_MAPPED)) {
 		block_nr = map->m_pblk + block_in_file - map->m_lblk;
-		SetPageMappedToDisk(page);
+		folio_set_mappedtodisk(folio);
 
 		if (!f2fs_is_valid_blkaddr(F2FS_I_SB(inode), block_nr,
 						DATA_GENERIC_ENHANCE_READ)) {
 			ret = -EFSCORRUPTED;
-			f2fs_handle_error(F2FS_I_SB(inode),
-						ERROR_INVALID_BLKADDR);
 			goto out;
 		}
 	} else {
 zero_out:
-		zero_user_segment(page, 0, PAGE_SIZE);
-		if (f2fs_need_verity(inode, page->index) &&
-		    !fsverity_verify_page(page)) {
+		folio_zero_segment(folio, 0, folio_size(folio));
+		if (f2fs_need_verity(inode, index) &&
+		    !fsverity_verify_folio(folio)) {
 			ret = -EIO;
 			goto out;
 		}
-		if (!PageUptodate(page))
-			SetPageUptodate(page);
-		unlock_page(page);
+		if (!folio_test_uptodate(folio))
+			folio_mark_uptodate(folio);
+		folio_unlock(folio);
 		goto out;
 	}
 
@@ -2190,14 +2143,14 @@ zero_out:
 	 */
 	if (bio && (!page_is_mergeable(F2FS_I_SB(inode), bio,
 				       *last_block_in_bio, block_nr) ||
-		    !f2fs_crypt_mergeable_bio(bio, inode, page->index, NULL))) {
+		    !f2fs_crypt_mergeable_bio(bio, inode, index, NULL))) {
 submit_and_realloc:
 		f2fs_submit_read_bio(F2FS_I_SB(inode), bio, DATA);
 		bio = NULL;
 	}
 	if (bio == NULL) {
 		bio = f2fs_grab_read_bio(inode, block_nr, nr_pages,
-				is_readahead ? REQ_RAHEAD : 0, page->index,
+				is_readahead ? REQ_RAHEAD : 0, index,
 				false);
 		if (IS_ERR(bio)) {
 			ret = PTR_ERR(bio);
@@ -2212,7 +2165,7 @@ submit_and_realloc:
 	 */
 	f2fs_wait_on_block_writeback(inode, block_nr);
 
-	if (bio_add_page(bio, page, blocksize, 0) < blocksize)
+	if (!bio_add_folio(bio, folio, blocksize, 0))
 		goto submit_and_realloc;
 
 	inc_page_count(F2FS_I_SB(inode), F2FS_RD_DATA);
@@ -2399,7 +2352,7 @@ out:
  * Major change was from block_size == page_size in f2fs by default.
  */
 static int f2fs_mpage_readpages(struct inode *inode,
-		struct readahead_control *rac, struct page *page)
+		struct readahead_control *rac, struct folio *folio)
 {
 	struct bio *bio = NULL;
 	sector_t last_block_in_bio = 0;
@@ -2419,6 +2372,7 @@ static int f2fs_mpage_readpages(struct inode *inode,
 #endif
 	unsigned nr_pages = rac ? readahead_count(rac) : 1;
 	unsigned max_nr_pages = nr_pages;
+	pgoff_t index;
 	int ret = 0;
 
 	map.m_pblk = 0;
@@ -2432,64 +2386,63 @@ static int f2fs_mpage_readpages(struct inode *inode,
 
 	for (; nr_pages; nr_pages--) {
 		if (rac) {
-			page = readahead_page(rac);
-			prefetchw(&page->flags);
+			folio = readahead_folio(rac);
+			prefetchw(&folio->flags);
 		}
 
-#ifdef CONFIG_F2FS_FS_COMPRESSION
-		if (f2fs_compressed_file(inode)) {
-			/* there are remained compressed pages, submit them */
-			if (!f2fs_cluster_can_merge_page(&cc, page->index)) {
-				ret = f2fs_read_multi_pages(&cc, &bio,
-							max_nr_pages,
-							&last_block_in_bio,
-							rac != NULL, false);
-				f2fs_destroy_compress_ctx(&cc, false);
-				if (ret)
-					goto set_error_page;
-			}
-			if (cc.cluster_idx == NULL_CLUSTER) {
-				if (nc_cluster_idx ==
-					page->index >> cc.log_cluster_size) {
-					goto read_single_page;
-				}
+		index = folio_index(folio);
 
-				ret = f2fs_is_compressed_cluster(inode, page->index);
-				if (ret < 0)
-					goto set_error_page;
-				else if (!ret) {
-					nc_cluster_idx =
-						page->index >> cc.log_cluster_size;
-					goto read_single_page;
-				}
-
-				nc_cluster_idx = NULL_CLUSTER;
-			}
-			ret = f2fs_init_compress_ctx(&cc);
+#ifdef CONFIG_F2FS_FS_COMPRESSION
+		if (!f2fs_compressed_file(inode))
+			goto read_single_page;
+
+		/* there are remained compressed pages, submit them */
+		if (!f2fs_cluster_can_merge_page(&cc, index)) {
+			ret = f2fs_read_multi_pages(&cc, &bio,
+						max_nr_pages,
+						&last_block_in_bio,
+						rac != NULL, false);
+			f2fs_destroy_compress_ctx(&cc, false);
 			if (ret)
 				goto set_error_page;
+		}
+		if (cc.cluster_idx == NULL_CLUSTER) {
+			if (nc_cluster_idx == index >> cc.log_cluster_size)
+				goto read_single_page;
 
-			f2fs_compress_ctx_add_page(&cc, page);
+			ret = f2fs_is_compressed_cluster(inode, index);
+			if (ret < 0)
+				goto set_error_page;
+			else if (!ret) {
+				nc_cluster_idx =
+					index >> cc.log_cluster_size;
+				goto read_single_page;
+			}
 
-			goto next_page;
+			nc_cluster_idx = NULL_CLUSTER;
 		}
+		ret = f2fs_init_compress_ctx(&cc);
+		if (ret)
+			goto set_error_page;
+
+		f2fs_compress_ctx_add_page(&cc, &folio->page);
+
+		goto next_page;
 read_single_page:
 #endif
 
-		ret = f2fs_read_single_page(inode, page, max_nr_pages, &map,
+		ret = f2fs_read_single_page(inode, folio, max_nr_pages, &map,
 					&bio, &last_block_in_bio, rac);
 		if (ret) {
 #ifdef CONFIG_F2FS_FS_COMPRESSION
 set_error_page:
 #endif
-			zero_user_segment(page, 0, PAGE_SIZE);
-			unlock_page(page);
+			folio_zero_segment(folio, 0, folio_size(folio));
+			folio_unlock(folio);
 		}
 #ifdef CONFIG_F2FS_FS_COMPRESSION
 next_page:
 #endif
-		if (rac)
-			put_page(page);
 
 #ifdef CONFIG_F2FS_FS_COMPRESSION
 		if (f2fs_compressed_file(inode)) {
@@ -2511,22 +2464,21 @@ next_page:
 
 static int f2fs_read_data_folio(struct file *file, struct folio *folio)
 {
-	struct page *page = &folio->page;
-	struct inode *inode = page_file_mapping(page)->host;
+	struct inode *inode = folio_file_mapping(folio)->host;
 	int ret = -EAGAIN;
 
-	trace_f2fs_readpage(page, DATA);
+	trace_f2fs_readpage(folio, DATA);
 
 	if (!f2fs_is_compress_backend_ready(inode)) {
-		unlock_page(page);
+		folio_unlock(folio);
 		return -EOPNOTSUPP;
 	}
 
 	/* If the file has inline data, try to read it directly */
 	if (f2fs_has_inline_data(inode))
-		ret = f2fs_read_inline_data(inode, page);
+		ret = f2fs_read_inline_data(inode, folio);
 	if (ret == -EAGAIN)
-		ret = f2fs_mpage_readpages(inode, NULL, page);
+		ret = f2fs_mpage_readpages(inode, NULL, folio);
 	return ret;
 }
 
@@ -2651,7 +2603,7 @@ bool f2fs_should_update_outplace(struct inode *inode, struct f2fs_io_info *fio)
 		return true;
 	if (IS_NOQUOTA(inode))
 		return true;
-	if (f2fs_is_atomic_file(inode))
+	if (f2fs_used_in_atomic_write(inode))
 		return true;
 	/* rewrite low ratio compress data w/ OPU mode to avoid fragmentation */
 	if (f2fs_compressed_file(inode) &&
@@ -2669,8 +2621,6 @@ bool f2fs_should_update_outplace(struct inode *inode, struct f2fs_io_info *fio)
 	if (fio) {
 		if (page_private_gcing(fio->page))
 			return true;
-		if (page_private_dummy(fio->page))
-			return true;
 		if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED) &&
 			f2fs_is_checkpointed_data(sbi, fio->old_blkaddr)))
 			return true;
@@ -2707,11 +2657,8 @@ int f2fs_do_write_data_page(struct f2fs_io_info *fio)
 	    f2fs_lookup_read_extent_cache_block(inode, page->index,
 						&fio->old_blkaddr)) {
 		if (!f2fs_is_valid_blkaddr(fio->sbi, fio->old_blkaddr,
-						DATA_GENERIC_ENHANCE)) {
-			f2fs_handle_error(fio->sbi,
-						ERROR_INVALID_BLKADDR);
+						DATA_GENERIC_ENHANCE))
 			return -EFSCORRUPTED;
-		}
 
 		ipu_force = true;
 		fio->need_lock = LOCK_DONE;
@@ -2739,12 +2686,11 @@ got_it:
 		!f2fs_is_valid_blkaddr(fio->sbi, fio->old_blkaddr,
 						DATA_GENERIC_ENHANCE)) {
 		err = -EFSCORRUPTED;
-		f2fs_handle_error(fio->sbi, ERROR_INVALID_BLKADDR);
 		goto out_writepage;
 	}
 
 	/* wait for GCed page writeback via META_MAPPING */
-	if (fio->post_read)
+	if (fio->meta_gc)
 		f2fs_wait_on_block_writeback(inode, fio->old_blkaddr);
 
 	/*
@@ -2766,12 +2712,11 @@ got_it:
 		if (err) {
 			if (fscrypt_inode_uses_fs_layer_crypto(inode))
 				fscrypt_finalize_bounce_page(&fio->encrypted_page);
-			if (PageWriteback(page))
-				end_page_writeback(page);
+			end_page_writeback(page);
 		} else {
 			set_inode_flag(inode, FI_UPDATE_WRITE);
 		}
-		trace_f2fs_do_write_data_page(fio->page, IPU);
+		trace_f2fs_do_write_data_page(page_folio(page), IPU);
 		return err;
 	}
 
@@ -2800,7 +2745,7 @@ got_it:
 
 	/* LFS mode write path */
 	f2fs_outplace_write_data(&dn, fio);
-	trace_f2fs_do_write_data_page(page, OPU);
+	trace_f2fs_do_write_data_page(page_folio(page), OPU);
 	set_inode_flag(inode, FI_APPEND_WRITE);
 out_writepage:
 	f2fs_put_dnode(&dn);
@@ -2840,14 +2785,14 @@ int f2fs_write_single_data_page(struct page *page, int *submitted,
 		.submitted = 0,
 		.compr_blocks = compr_blocks,
 		.need_lock = compr_blocks ? LOCK_DONE : LOCK_RETRY,
-		.post_read = f2fs_post_read_required(inode) ? 1 : 0,
+		.meta_gc = f2fs_meta_inode_gc_required(inode) ? 1 : 0,
 		.io_type = io_type,
 		.io_wbc = wbc,
 		.bio = bio,
 		.last_block = last_block,
 	};
 
-	trace_f2fs_writepage(page, DATA);
+	trace_f2fs_writepage(page_folio(page), DATA);
 
 	/* we should bypass data pages to proceed the kworker jobs */
 	if (unlikely(f2fs_cp_error(sbi))) {
@@ -3460,7 +3405,7 @@ restart:
 
 	if (f2fs_has_inline_data(inode)) {
 		if (pos + len <= MAX_INLINE_DATA(inode)) {
-			f2fs_do_read_inline_data(page, ipage);
+			f2fs_do_read_inline_data(page_folio(page), ipage);
 			set_inode_flag(inode, FI_DATA_EXIST);
 			if (inode->i_nlink)
 				set_page_private_inline(ipage);
@@ -3706,7 +3651,6 @@ repeat:
 		if (!f2fs_is_valid_blkaddr(sbi, blkaddr,
 				DATA_GENERIC_ENHANCE_READ)) {
 			err = -EFSCORRUPTED;
-			f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR);
 			goto fail;
 		}
 		err = f2fs_submit_page_read(use_cow ?
@@ -3822,7 +3766,7 @@ static bool f2fs_dirty_data_folio(struct address_space *mapping,
 {
 	struct inode *inode = mapping->host;
 
-	trace_f2fs_set_page_dirty(&folio->page, DATA);
+	trace_f2fs_set_page_dirty(folio, DATA);
 
 	if (!folio_test_uptodate(folio))
 		folio_mark_uptodate(folio);
@@ -3907,26 +3851,36 @@ static int f2fs_migrate_blocks(struct inode *inode, block_t start_blk,
 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 	unsigned int blkofs;
 	unsigned int blk_per_sec = BLKS_PER_SEC(sbi);
+	unsigned int end_blk = start_blk + blkcnt - 1;
 	unsigned int secidx = start_blk / blk_per_sec;
-	unsigned int end_sec = secidx + blkcnt / blk_per_sec;
+	unsigned int end_sec;
 	int ret = 0;
 
+	if (!blkcnt)
+		return 0;
+	end_sec = end_blk / blk_per_sec;
+
 	f2fs_down_write(&F2FS_I(inode)->i_gc_rwsem[WRITE]);
 	filemap_invalidate_lock(inode->i_mapping);
 
 	set_inode_flag(inode, FI_ALIGNED_WRITE);
 	set_inode_flag(inode, FI_OPU_WRITE);
 
-	for (; secidx < end_sec; secidx++) {
+	for (; secidx <= end_sec; secidx++) {
+		unsigned int blkofs_end = secidx == end_sec ?
+				end_blk % blk_per_sec : blk_per_sec - 1;
+
 		f2fs_down_write(&sbi->pin_sem);
 
-		f2fs_lock_op(sbi);
-		f2fs_allocate_new_section(sbi, CURSEG_COLD_DATA_PINNED, false);
-		f2fs_unlock_op(sbi);
+		ret = f2fs_allocate_pinning_section(sbi);
+		if (ret) {
+			f2fs_up_write(&sbi->pin_sem);
+			break;
+		}
 
 		set_inode_flag(inode, FI_SKIP_WRITES);
 
-		for (blkofs = 0; blkofs < blk_per_sec; blkofs++) {
+		for (blkofs = 0; blkofs <= blkofs_end; blkofs++) {
 			struct page *page;
 			unsigned int blkidx = secidx * blk_per_sec + blkofs;
 
@@ -3968,15 +3922,14 @@ static int check_swap_activate(struct swap_info_struct *sis,
 	struct address_space *mapping = swap_file->f_mapping;
 	struct inode *inode = mapping->host;
 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
-	sector_t cur_lblock;
-	sector_t last_lblock;
-	sector_t pblock;
-	sector_t lowest_pblock = -1;
-	sector_t highest_pblock = 0;
+	block_t cur_lblock;
+	block_t last_lblock;
+	block_t pblock;
+	block_t lowest_pblock = -1;
+	block_t highest_pblock = 0;
 	int nr_extents = 0;
-	unsigned long nr_pblocks;
+	unsigned int nr_pblocks;
 	unsigned int blks_per_sec = BLKS_PER_SEC(sbi);
-	unsigned int sec_blks_mask = BLKS_PER_SEC(sbi) - 1;
 	unsigned int not_aligned = 0;
 	int ret = 0;
 
@@ -4014,28 +3967,35 @@ retry:
 		pblock = map.m_pblk;
 		nr_pblocks = map.m_len;
 
-		if ((pblock - SM_I(sbi)->main_blkaddr) & sec_blks_mask ||
-				nr_pblocks & sec_blks_mask) {
+		if ((pblock - SM_I(sbi)->main_blkaddr) % blks_per_sec ||
+				nr_pblocks % blks_per_sec ||
+				!f2fs_valid_pinned_area(sbi, pblock)) {
+			bool last_extent = false;
+
 			not_aligned++;
 
 			nr_pblocks = roundup(nr_pblocks, blks_per_sec);
 			if (cur_lblock + nr_pblocks > sis->max)
 				nr_pblocks -= blks_per_sec;
 
+			/* this extent is last one */
 			if (!nr_pblocks) {
-				/* this extent is last one */
-				nr_pblocks = map.m_len;
-				f2fs_warn(sbi, "Swapfile: last extent is not aligned to section");
-				goto next;
+				nr_pblocks = last_lblock - cur_lblock;
+				last_extent = true;
 			}
 
 			ret = f2fs_migrate_blocks(inode, cur_lblock,
 							nr_pblocks);
-			if (ret)
+			if (ret) {
+				if (ret == -ENOENT)
+					ret = -EINVAL;
 				goto out;
-			goto retry;
+			}
+
+			if (!last_extent)
+				goto retry;
 		}
-next:
+
 		if (cur_lblock + nr_pblocks >= sis->max)
 			nr_pblocks = sis->max - cur_lblock;
 
@@ -4073,17 +4033,17 @@ static int f2fs_swap_activate(struct swap_info_struct *sis, struct file *file,
 				sector_t *span)
 {
 	struct inode *inode = file_inode(file);
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 	int ret;
 
 	if (!S_ISREG(inode->i_mode))
 		return -EINVAL;
 
-	if (f2fs_readonly(F2FS_I_SB(inode)->sb))
+	if (f2fs_readonly(sbi->sb))
 		return -EROFS;
 
-	if (f2fs_lfs_mode(F2FS_I_SB(inode))) {
-		f2fs_err(F2FS_I_SB(inode),
-			"Swapfile not supported in LFS mode");
+	if (f2fs_lfs_mode(sbi) && !f2fs_sb_has_blkzoned(sbi)) {
+		f2fs_err(sbi, "Swapfile not supported in LFS mode");
 		return -EINVAL;
 	}
 
@@ -4094,6 +4054,10 @@ static int f2fs_swap_activate(struct swap_info_struct *sis, struct file *file,
 	if (!f2fs_disable_compressed_file(inode))
 		return -EINVAL;
 
+	ret = filemap_fdatawrite(inode->i_mapping);
+	if (ret < 0)
+		return ret;
+
 	f2fs_precache_extents(inode);
 
 	ret = check_swap_activate(sis, file, span);
@@ -4102,7 +4066,7 @@ static int f2fs_swap_activate(struct swap_info_struct *sis, struct file *file,
 
 	stat_inc_swapfile_inode(inode);
 	set_inode_flag(inode, FI_PIN_FILE);
-	f2fs_update_time(F2FS_I_SB(inode), REQ_TIME);
+	f2fs_update_time(sbi, REQ_TIME);
 	return ret;
 }
 
@@ -4143,11 +4107,12 @@ const struct address_space_operations f2fs_dblock_aops = {
 
 void f2fs_clear_page_cache_dirty_tag(struct page *page)
 {
-	struct address_space *mapping = page_mapping(page);
+	struct folio *folio = page_folio(page);
+	struct address_space *mapping = folio->mapping;
 	unsigned long flags;
 
 	xa_lock_irqsave(&mapping->i_pages, flags);
-	__xa_clear_mark(&mapping->i_pages, page_index(page),
+	__xa_clear_mark(&mapping->i_pages, folio->index,
 						PAGECACHE_TAG_DIRTY);
 	xa_unlock_irqrestore(&mapping->i_pages, flags);
 }
@@ -4220,7 +4185,8 @@ static int f2fs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
 	map.m_lblk = bytes_to_blks(inode, offset);
 	map.m_len = bytes_to_blks(inode, offset + length - 1) - map.m_lblk + 1;
 	map.m_next_pgofs = &next_pgofs;
-	map.m_seg_type = f2fs_rw_hint_to_seg_type(inode->i_write_hint);
+	map.m_seg_type = f2fs_rw_hint_to_seg_type(F2FS_I_SB(inode),
+						inode->i_write_hint);
 	if (flags & IOMAP_WRITE)
 		map.m_may_create = true;
 
@@ -4241,12 +4207,13 @@ static int f2fs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
 	 * We should never see delalloc or compressed extents here based on
 	 * prior flushing and checks.
 	 */
-	if (WARN_ON_ONCE(map.m_pblk == NEW_ADDR))
-		return -EINVAL;
 	if (WARN_ON_ONCE(map.m_pblk == COMPRESS_ADDR))
 		return -EINVAL;
 
-	if (map.m_pblk != NULL_ADDR) {
+	if (map.m_flags & F2FS_MAP_MAPPED) {
+		if (WARN_ON_ONCE(map.m_pblk == NEW_ADDR))
+			return -EINVAL;
+
 		iomap->length = blks_to_bytes(inode, map.m_len);
 		iomap->type = IOMAP_MAPPED;
 		iomap->flags |= IOMAP_F_MERGED;
@@ -4255,9 +4222,17 @@ static int f2fs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
 	} else {
 		if (flags & IOMAP_WRITE)
 			return -ENOTBLK;
-		iomap->length = blks_to_bytes(inode, next_pgofs) -
-				iomap->offset;
-		iomap->type = IOMAP_HOLE;
+
+		if (map.m_pblk == NULL_ADDR) {
+			iomap->length = blks_to_bytes(inode, next_pgofs) -
+								iomap->offset;
+			iomap->type = IOMAP_HOLE;
+		} else if (map.m_pblk == NEW_ADDR) {
+			iomap->length = blks_to_bytes(inode, map.m_len);
+			iomap->type = IOMAP_UNWRITTEN;
+		} else {
+			f2fs_bug_on(F2FS_I_SB(inode), 1);
+		}
 		iomap->addr = IOMAP_NULL_ADDR;
 	}
 
diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c
index fdbf994f12..8b0e1e71b6 100644
--- a/fs/f2fs/debug.c
+++ b/fs/f2fs/debug.c
@@ -41,7 +41,7 @@ void f2fs_update_sit_info(struct f2fs_sb_info *sbi)
 	total_vblocks = 0;
 	blks_per_sec = CAP_BLKS_PER_SEC(sbi);
 	hblks_per_sec = blks_per_sec / 2;
-	for (segno = 0; segno < MAIN_SEGS(sbi); segno += sbi->segs_per_sec) {
+	for (segno = 0; segno < MAIN_SEGS(sbi); segno += SEGS_PER_SEC(sbi)) {
 		vblocks = get_valid_blocks(sbi, segno, true);
 		dist = abs(vblocks - hblks_per_sec);
 		bimodal += dist * dist;
@@ -135,7 +135,7 @@ static void update_general_status(struct f2fs_sb_info *sbi)
 	si->cur_ckpt_time = sbi->cprc_info.cur_time;
 	si->peak_ckpt_time = sbi->cprc_info.peak_time;
 	spin_unlock(&sbi->cprc_info.stat_lock);
-	si->total_count = (int)sbi->user_block_count / sbi->blocks_per_seg;
+	si->total_count = BLKS_TO_SEGS(sbi, (int)sbi->user_block_count);
 	si->rsvd_segs = reserved_segments(sbi);
 	si->overp_segs = overprovision_segments(sbi);
 	si->valid_count = valid_user_blocks(sbi);
@@ -176,11 +176,10 @@ static void update_general_status(struct f2fs_sb_info *sbi)
 	si->alloc_nids = NM_I(sbi)->nid_cnt[PREALLOC_NID];
 	si->io_skip_bggc = sbi->io_skip_bggc;
 	si->other_skip_bggc = sbi->other_skip_bggc;
-	si->util_free = (int)(free_user_blocks(sbi) >> sbi->log_blocks_per_seg)
+	si->util_free = (int)(BLKS_TO_SEGS(sbi, free_user_blocks(sbi)))
 		* 100 / (int)(sbi->user_block_count >> sbi->log_blocks_per_seg)
 		/ 2;
-	si->util_valid = (int)(written_block_count(sbi) >>
-						sbi->log_blocks_per_seg)
+	si->util_valid = (int)(BLKS_TO_SEGS(sbi, written_block_count(sbi)))
 		* 100 / (int)(sbi->user_block_count >> sbi->log_blocks_per_seg)
 		/ 2;
 	si->util_invalid = 50 - si->util_free - si->util_valid;
@@ -208,7 +207,7 @@ static void update_general_status(struct f2fs_sb_info *sbi)
 		if (!blks)
 			continue;
 
-		if (blks == sbi->blocks_per_seg)
+		if (blks == BLKS_PER_SEG(sbi))
 			si->full_seg[type]++;
 		else
 			si->dirty_seg[type]++;
diff --git a/fs/f2fs/dir.c b/fs/f2fs/dir.c
index 1b937f7d04..02c9355176 100644
--- a/fs/f2fs/dir.c
+++ b/fs/f2fs/dir.c
@@ -996,9 +996,8 @@ int f2fs_fill_dentries(struct dir_context *ctx, struct f2fs_dentry_ptr *d,
 		de = &d->dentry[bit_pos];
 		if (de->name_len == 0) {
 			if (found_valid_dirent || !bit_pos) {
-				printk_ratelimited(
-					"%sF2FS-fs (%s): invalid namelen(0), ino:%u, run fsck to fix.",
-					KERN_WARNING, sbi->sb->s_id,
+				f2fs_warn_ratelimited(sbi,
+					"invalid namelen(0), ino:%u, run fsck to fix.",
 					le32_to_cpu(de->ino));
 				set_sbi_flag(sbi, SBI_NEED_FSCK);
 			}
diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c
index ad8dfac73b..fd1fc06359 100644
--- a/fs/f2fs/extent_cache.c
+++ b/fs/f2fs/extent_cache.c
@@ -19,34 +19,23 @@
 #include "node.h"
 #include <trace/events/f2fs.h>
 
-bool sanity_check_extent_cache(struct inode *inode)
+bool sanity_check_extent_cache(struct inode *inode, struct page *ipage)
 {
 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
-	struct f2fs_inode_info *fi = F2FS_I(inode);
-	struct extent_tree *et = fi->extent_tree[EX_READ];
-	struct extent_info *ei;
-
-	if (!et)
-		return true;
+	struct f2fs_extent *i_ext = &F2FS_INODE(ipage)->i_ext;
+	struct extent_info ei;
 
-	ei = &et->largest;
-	if (!ei->len)
-		return true;
+	get_read_extent_info(&ei, i_ext);
 
-	/* Let's drop, if checkpoint got corrupted. */
-	if (is_set_ckpt_flags(sbi, CP_ERROR_FLAG)) {
-		ei->len = 0;
-		et->largest_updated = true;
+	if (!ei.len)
 		return true;
-	}
 
-	if (!f2fs_is_valid_blkaddr(sbi, ei->blk, DATA_GENERIC_ENHANCE) ||
-	    !f2fs_is_valid_blkaddr(sbi, ei->blk + ei->len - 1,
+	if (!f2fs_is_valid_blkaddr(sbi, ei.blk, DATA_GENERIC_ENHANCE) ||
+	    !f2fs_is_valid_blkaddr(sbi, ei.blk + ei.len - 1,
 					DATA_GENERIC_ENHANCE)) {
-		set_sbi_flag(sbi, SBI_NEED_FSCK);
 		f2fs_warn(sbi, "%s: inode (ino=%lx) extent info [%u, %u, %u] is incorrect, run fsck to fix",
 			  __func__, inode->i_ino,
-			  ei->blk, ei->fofs, ei->len);
+			  ei.blk, ei.fofs, ei.len);
 		return false;
 	}
 	return true;
@@ -395,24 +384,22 @@ void f2fs_init_read_extent_tree(struct inode *inode, struct page *ipage)
 
 	if (!__may_extent_tree(inode, EX_READ)) {
 		/* drop largest read extent */
-		if (i_ext && i_ext->len) {
+		if (i_ext->len) {
 			f2fs_wait_on_page_writeback(ipage, NODE, true, true);
 			i_ext->len = 0;
 			set_page_dirty(ipage);
 		}
-		goto out;
+		set_inode_flag(inode, FI_NO_EXTENT);
+		return;
 	}
 
 	et = __grab_extent_tree(inode, EX_READ);
 
-	if (!i_ext || !i_ext->len)
-		goto out;
-
 	get_read_extent_info(&ei, i_ext);
 
 	write_lock(&et->lock);
-	if (atomic_read(&et->node_cnt))
-		goto unlock_out;
+	if (atomic_read(&et->node_cnt) || !ei.len)
+		goto skip;
 
 	en = __attach_extent_node(sbi, et, &ei, NULL,
 				&et->root.rb_root.rb_node, true);
@@ -424,11 +411,13 @@ void f2fs_init_read_extent_tree(struct inode *inode, struct page *ipage)
 		list_add_tail(&en->list, &eti->extent_list);
 		spin_unlock(&eti->extent_lock);
 	}
-unlock_out:
+skip:
+	/* Let's drop, if checkpoint got corrupted. */
+	if (f2fs_cp_error(sbi)) {
+		et->largest.len = 0;
+		et->largest_updated = true;
+	}
 	write_unlock(&et->lock);
-out:
-	if (!F2FS_I(inode)->extent_tree[EX_READ])
-		set_inode_flag(inode, FI_NO_EXTENT);
 }
 
 void f2fs_init_age_extent_tree(struct inode *inode)
@@ -856,10 +845,8 @@ static int __get_new_block_age(struct inode *inode, struct extent_info *ei,
 		goto out;
 
 	if (__is_valid_data_blkaddr(blkaddr) &&
-	    !f2fs_is_valid_blkaddr(sbi, blkaddr, DATA_GENERIC_ENHANCE)) {
-		f2fs_bug_on(sbi, 1);
+	    !f2fs_is_valid_blkaddr(sbi, blkaddr, DATA_GENERIC_ENHANCE))
 		return -EINVAL;
-	}
 out:
 	/*
 	 * init block age with zero, this can happen when the block age extent
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index 6610ff6d7e..5556ab4913 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -24,6 +24,7 @@
 #include <linux/blkdev.h>
 #include <linux/quotaops.h>
 #include <linux/part_stat.h>
+#include <linux/rw_hint.h>
 #include <crypto/hash.h>
 
 #include <linux/fscrypt.h>
@@ -60,7 +61,9 @@ enum {
 	FAULT_SLAB_ALLOC,
 	FAULT_DQUOT_INIT,
 	FAULT_LOCK_OP,
-	FAULT_BLKADDR,
+	FAULT_BLKADDR_VALIDITY,
+	FAULT_BLKADDR_CONSISTENCE,
+	FAULT_NO_SEGMENT,
 	FAULT_MAX,
 };
 
@@ -69,7 +72,7 @@ enum {
 
 struct f2fs_fault_info {
 	atomic_t inject_ops;
-	unsigned int inject_rate;
+	int inject_rate;
 	unsigned int inject_type;
 };
 
@@ -147,7 +150,6 @@ struct f2fs_rwsem {
 
 struct f2fs_mount_info {
 	unsigned int opt;
-	int write_io_size_bits;		/* Write IO size bits */
 	block_t root_reserved_blocks;	/* root reserved blocks */
 	kuid_t s_resuid;		/* reserved blocks for uid */
 	kgid_t s_resgid;		/* reserved blocks for gid */
@@ -763,11 +765,6 @@ enum {
 
 #define DEF_DIR_LEVEL		0
 
-enum {
-	GC_FAILURE_PIN,
-	MAX_GC_FAILURE
-};
-
 /* used for f2fs_inode_info->flags */
 enum {
 	FI_NEW_INODE,		/* indicate newly allocated inode */
@@ -806,6 +803,7 @@ enum {
 	FI_COW_FILE,		/* indicate COW file */
 	FI_ATOMIC_COMMITTED,	/* indicate atomic commit completed except disk sync */
 	FI_ATOMIC_REPLACE,	/* indicate atomic replace */
+	FI_OPENED_FILE,		/* indicate file has been opened */
 	FI_MAX,			/* max flag, never be used */
 };
 
@@ -814,9 +812,10 @@ struct f2fs_inode_info {
 	unsigned long i_flags;		/* keep an inode flags for ioctl */
 	unsigned char i_advise;		/* use to give file attribute hints */
 	unsigned char i_dir_level;	/* use for dentry level for large dir */
-	unsigned int i_current_depth;	/* only for directory depth */
-	/* for gc failure statistic */
-	unsigned int i_gc_failures[MAX_GC_FAILURE];
+	union {
+		unsigned int i_current_depth;	/* only for directory depth */
+		unsigned short i_gc_failures;	/* for gc failure statistic */
+	};
 	unsigned int i_pino;		/* parent inode number */
 	umode_t i_acl_mode;		/* keep file acl mode temporarily */
 
@@ -844,7 +843,11 @@ struct f2fs_inode_info {
 	struct task_struct *atomic_write_task;	/* store atomic write task */
 	struct extent_tree *extent_tree[NR_EXTENT_CACHES];
 					/* cached extent_tree entry */
-	struct inode *cow_inode;	/* copy-on-write inode for atomic write */
+	union {
+		struct inode *cow_inode;	/* copy-on-write inode for atomic write */
+		struct inode *atomic_inode;
+					/* point to atomic_inode, available only for cow_inode */
+	};
 
 	/* avoid racing between foreground op and gc */
 	struct f2fs_rwsem i_gc_rwsem[2];
@@ -1116,6 +1119,7 @@ enum count_type {
  * ...			Only can be used with META.
  */
 #define PAGE_TYPE_OF_BIO(type)	((type) > META ? META : (type))
+#define PAGE_TYPE_ON_MAIN(type)	((type) == DATA || (type) == NODE)
 enum page_type {
 	DATA = 0,
 	NODE = 1,	/* should not change this */
@@ -1210,9 +1214,8 @@ struct f2fs_io_info {
 	unsigned int submitted:1;	/* indicate IO submission */
 	unsigned int in_list:1;		/* indicate fio is in io_list */
 	unsigned int is_por:1;		/* indicate IO is from recovery or not */
-	unsigned int retry:1;		/* need to reallocate block address */
 	unsigned int encrypted:1;	/* indicate file is encrypted */
-	unsigned int post_read:1;	/* require post read */
+	unsigned int meta_gc:1;		/* require meta inode GC */
 	enum iostat_type io_type;	/* io type */
 	struct writeback_control *io_wbc; /* writeback control */
 	struct bio **bio;		/* bio for ipu */
@@ -1245,7 +1248,7 @@ struct f2fs_bio_info {
 #define FDEV(i)				(sbi->devs[i])
 #define RDEV(i)				(raw_super->devs[i])
 struct f2fs_dev_info {
-	struct bdev_handle *bdev_handle;
+	struct file *bdev_file;
 	struct block_device *bdev;
 	char path[MAX_PATH_LEN];
 	unsigned int total_segments;
@@ -1412,18 +1415,16 @@ static inline void f2fs_clear_bit(unsigned int nr, char *addr);
  * Layout A: lowest bit should be 1
  * | bit0 = 1 | bit1 | bit2 | ... | bit MAX | private data .... |
  * bit 0	PAGE_PRIVATE_NOT_POINTER
- * bit 1	PAGE_PRIVATE_DUMMY_WRITE
- * bit 2	PAGE_PRIVATE_ONGOING_MIGRATION
- * bit 3	PAGE_PRIVATE_INLINE_INODE
- * bit 4	PAGE_PRIVATE_REF_RESOURCE
- * bit 5-	f2fs private data
+ * bit 1	PAGE_PRIVATE_ONGOING_MIGRATION
+ * bit 2	PAGE_PRIVATE_INLINE_INODE
+ * bit 3	PAGE_PRIVATE_REF_RESOURCE
+ * bit 4-	f2fs private data
  *
  * Layout B: lowest bit should be 0
  * page.private is a wrapped pointer.
  */
 enum {
 	PAGE_PRIVATE_NOT_POINTER,		/* private contains non-pointer data */
-	PAGE_PRIVATE_DUMMY_WRITE,		/* data page for padding aligned IO */
 	PAGE_PRIVATE_ONGOING_MIGRATION,		/* data page which is on-going migrating */
 	PAGE_PRIVATE_INLINE_INODE,		/* inode page contains inline data */
 	PAGE_PRIVATE_REF_RESOURCE,		/* dirty page has referenced resources */
@@ -1557,6 +1558,7 @@ struct f2fs_sb_info {
 
 #ifdef CONFIG_BLK_DEV_ZONED
 	unsigned int blocks_per_blkz;		/* F2FS blocks per zone */
+	unsigned int max_open_zones;		/* max open zone resources of the zoned device */
 #endif
 
 	/* for node-related operations */
@@ -1570,7 +1572,6 @@ struct f2fs_sb_info {
 	struct f2fs_bio_info *write_io[NR_PAGE_TYPE];	/* for write bios */
 	/* keep migration IO order for LFS mode */
 	struct f2fs_rwsem io_order_lock;
-	mempool_t *write_io_dummy;		/* Dummy pages */
 	pgoff_t page_eio_ofs[NR_PAGE_TYPE];	/* EIO page offset */
 	int page_eio_cnt[NR_PAGE_TYPE];		/* EIO count */
 
@@ -1677,7 +1678,7 @@ struct f2fs_sb_info {
 	unsigned long long skipped_gc_rwsem;		/* FG_GC only */
 
 	/* threshold for gc trials on pinned files */
-	u64 gc_pin_file_threshold;
+	unsigned short gc_pin_file_threshold;
 	struct f2fs_rwsem pin_sem;
 
 	/* maximum # of trials to find a victim segment for SSR and GC */
@@ -1816,6 +1817,37 @@ struct f2fs_sb_info {
 #endif
 };
 
+/* Definitions to access f2fs_sb_info */
+#define SEGS_TO_BLKS(sbi, segs)					\
+		((segs) << (sbi)->log_blocks_per_seg)
+#define BLKS_TO_SEGS(sbi, blks)					\
+		((blks) >> (sbi)->log_blocks_per_seg)
+
+#define BLKS_PER_SEG(sbi)	((sbi)->blocks_per_seg)
+#define BLKS_PER_SEC(sbi)	(SEGS_TO_BLKS(sbi, (sbi)->segs_per_sec))
+#define SEGS_PER_SEC(sbi)	((sbi)->segs_per_sec)
+
+__printf(3, 4)
+void f2fs_printk(struct f2fs_sb_info *sbi, bool limit_rate, const char *fmt, ...);
+
+#define f2fs_err(sbi, fmt, ...)						\
+	f2fs_printk(sbi, false, KERN_ERR fmt, ##__VA_ARGS__)
+#define f2fs_warn(sbi, fmt, ...)					\
+	f2fs_printk(sbi, false, KERN_WARNING fmt, ##__VA_ARGS__)
+#define f2fs_notice(sbi, fmt, ...)					\
+	f2fs_printk(sbi, false, KERN_NOTICE fmt, ##__VA_ARGS__)
+#define f2fs_info(sbi, fmt, ...)					\
+	f2fs_printk(sbi, false, KERN_INFO fmt, ##__VA_ARGS__)
+#define f2fs_debug(sbi, fmt, ...)					\
+	f2fs_printk(sbi, false, KERN_DEBUG fmt, ##__VA_ARGS__)
+
+#define f2fs_err_ratelimited(sbi, fmt, ...)				\
+	f2fs_printk(sbi, true, KERN_ERR fmt, ##__VA_ARGS__)
+#define f2fs_warn_ratelimited(sbi, fmt, ...)				\
+	f2fs_printk(sbi, true, KERN_WARNING fmt, ##__VA_ARGS__)
+#define f2fs_info_ratelimited(sbi, fmt, ...)				\
+	f2fs_printk(sbi, true, KERN_INFO fmt, ##__VA_ARGS__)
+
 #ifdef CONFIG_F2FS_FAULT_INJECTION
 #define time_to_inject(sbi, type) __time_to_inject(sbi, type, __func__,	\
 									__builtin_return_address(0))
@@ -1833,9 +1865,8 @@ static inline bool __time_to_inject(struct f2fs_sb_info *sbi, int type,
 	atomic_inc(&ffi->inject_ops);
 	if (atomic_read(&ffi->inject_ops) >= ffi->inject_rate) {
 		atomic_set(&ffi->inject_ops, 0);
-		printk_ratelimited("%sF2FS-fs (%s) : inject %s in %s of %pS\n",
-			KERN_INFO, sbi->sb->s_id, f2fs_fault_name[type],
-			func, parent_func);
+		f2fs_info_ratelimited(sbi, "inject %s in %s of %pS",
+				f2fs_fault_name[type], func, parent_func);
 		return true;
 	}
 	return false;
@@ -2255,11 +2286,32 @@ static inline bool __allow_reserved_blocks(struct f2fs_sb_info *sbi,
 	return false;
 }
 
+static inline unsigned int get_available_block_count(struct f2fs_sb_info *sbi,
+						struct inode *inode, bool cap)
+{
+	block_t avail_user_block_count;
+
+	avail_user_block_count = sbi->user_block_count -
+					sbi->current_reserved_blocks;
+
+	if (!__allow_reserved_blocks(sbi, inode, cap))
+		avail_user_block_count -= F2FS_OPTION(sbi).root_reserved_blocks;
+
+	if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED))) {
+		if (avail_user_block_count > sbi->unusable_block_count)
+			avail_user_block_count -= sbi->unusable_block_count;
+		else
+			avail_user_block_count = 0;
+	}
+
+	return avail_user_block_count;
+}
+
 static inline void f2fs_i_blocks_write(struct inode *, block_t, bool, bool);
 static inline int inc_valid_block_count(struct f2fs_sb_info *sbi,
 				 struct inode *inode, blkcnt_t *count, bool partial)
 {
-	blkcnt_t diff = 0, release = 0;
+	long long diff = 0, release = 0;
 	block_t avail_user_block_count;
 	int ret;
 
@@ -2279,40 +2331,27 @@ static inline int inc_valid_block_count(struct f2fs_sb_info *sbi,
 	percpu_counter_add(&sbi->alloc_valid_block_count, (*count));
 
 	spin_lock(&sbi->stat_lock);
-	sbi->total_valid_block_count += (block_t)(*count);
-	avail_user_block_count = sbi->user_block_count -
-					sbi->current_reserved_blocks;
 
-	if (!__allow_reserved_blocks(sbi, inode, true))
-		avail_user_block_count -= F2FS_OPTION(sbi).root_reserved_blocks;
-
-	if (F2FS_IO_ALIGNED(sbi))
-		avail_user_block_count -= sbi->blocks_per_seg *
-				SM_I(sbi)->additional_reserved_segments;
-
-	if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED))) {
-		if (avail_user_block_count > sbi->unusable_block_count)
-			avail_user_block_count -= sbi->unusable_block_count;
-		else
-			avail_user_block_count = 0;
-	}
-	if (unlikely(sbi->total_valid_block_count > avail_user_block_count)) {
+	avail_user_block_count = get_available_block_count(sbi, inode, true);
+	diff = (long long)sbi->total_valid_block_count + *count -
+						avail_user_block_count;
+	if (unlikely(diff > 0)) {
 		if (!partial) {
 			spin_unlock(&sbi->stat_lock);
+			release = *count;
 			goto enospc;
 		}
-
-		diff = sbi->total_valid_block_count - avail_user_block_count;
 		if (diff > *count)
 			diff = *count;
 		*count -= diff;
 		release = diff;
-		sbi->total_valid_block_count -= diff;
 		if (!*count) {
 			spin_unlock(&sbi->stat_lock);
 			goto enospc;
 		}
 	}
+	sbi->total_valid_block_count += (block_t)(*count);
+
 	spin_unlock(&sbi->stat_lock);
 
 	if (unlikely(release)) {
@@ -2329,20 +2368,6 @@ release_quota:
 	return -ENOSPC;
 }
 
-__printf(2, 3)
-void f2fs_printk(struct f2fs_sb_info *sbi, const char *fmt, ...);
-
-#define f2fs_err(sbi, fmt, ...)						\
-	f2fs_printk(sbi, KERN_ERR fmt, ##__VA_ARGS__)
-#define f2fs_warn(sbi, fmt, ...)					\
-	f2fs_printk(sbi, KERN_WARNING fmt, ##__VA_ARGS__)
-#define f2fs_notice(sbi, fmt, ...)					\
-	f2fs_printk(sbi, KERN_NOTICE fmt, ##__VA_ARGS__)
-#define f2fs_info(sbi, fmt, ...)					\
-	f2fs_printk(sbi, KERN_INFO fmt, ##__VA_ARGS__)
-#define f2fs_debug(sbi, fmt, ...)					\
-	f2fs_printk(sbi, KERN_DEBUG fmt, ##__VA_ARGS__)
-
 #define PAGE_PRIVATE_GET_FUNC(name, flagname) \
 static inline bool page_private_##name(struct page *page) \
 { \
@@ -2371,17 +2396,14 @@ static inline void clear_page_private_##name(struct page *page) \
 PAGE_PRIVATE_GET_FUNC(nonpointer, NOT_POINTER);
 PAGE_PRIVATE_GET_FUNC(inline, INLINE_INODE);
 PAGE_PRIVATE_GET_FUNC(gcing, ONGOING_MIGRATION);
-PAGE_PRIVATE_GET_FUNC(dummy, DUMMY_WRITE);
 
 PAGE_PRIVATE_SET_FUNC(reference, REF_RESOURCE);
 PAGE_PRIVATE_SET_FUNC(inline, INLINE_INODE);
 PAGE_PRIVATE_SET_FUNC(gcing, ONGOING_MIGRATION);
-PAGE_PRIVATE_SET_FUNC(dummy, DUMMY_WRITE);
 
 PAGE_PRIVATE_CLEAR_FUNC(reference, REF_RESOURCE);
 PAGE_PRIVATE_CLEAR_FUNC(inline, INLINE_INODE);
 PAGE_PRIVATE_CLEAR_FUNC(gcing, ONGOING_MIGRATION);
-PAGE_PRIVATE_CLEAR_FUNC(dummy, DUMMY_WRITE);
 
 static inline unsigned long get_page_private_data(struct page *page)
 {
@@ -2515,11 +2537,8 @@ static inline int get_dirty_pages(struct inode *inode)
 
 static inline int get_blocktype_secs(struct f2fs_sb_info *sbi, int block_type)
 {
-	unsigned int pages_per_sec = sbi->segs_per_sec * sbi->blocks_per_seg;
-	unsigned int segs = (get_pages(sbi, block_type) + pages_per_sec - 1) >>
-						sbi->log_blocks_per_seg;
-
-	return segs / sbi->segs_per_sec;
+	return div_u64(get_pages(sbi, block_type) + BLKS_PER_SEC(sbi) - 1,
+							BLKS_PER_SEC(sbi));
 }
 
 static inline block_t valid_user_blocks(struct f2fs_sb_info *sbi)
@@ -2583,7 +2602,7 @@ static inline block_t __start_cp_addr(struct f2fs_sb_info *sbi)
 	block_t start_addr = le32_to_cpu(F2FS_RAW_SUPER(sbi)->cp_blkaddr);
 
 	if (sbi->cur_cp_pack == 2)
-		start_addr += sbi->blocks_per_seg;
+		start_addr += BLKS_PER_SEG(sbi);
 	return start_addr;
 }
 
@@ -2592,7 +2611,7 @@ static inline block_t __start_cp_next_addr(struct f2fs_sb_info *sbi)
 	block_t start_addr = le32_to_cpu(F2FS_RAW_SUPER(sbi)->cp_blkaddr);
 
 	if (sbi->cur_cp_pack == 1)
-		start_addr += sbi->blocks_per_seg;
+		start_addr += BLKS_PER_SEG(sbi);
 	return start_addr;
 }
 
@@ -2611,7 +2630,8 @@ static inline int inc_valid_node_count(struct f2fs_sb_info *sbi,
 					struct inode *inode, bool is_inode)
 {
 	block_t	valid_block_count;
-	unsigned int valid_node_count, user_block_count;
+	unsigned int valid_node_count;
+	unsigned int avail_user_block_count;
 	int err;
 
 	if (is_inode) {
@@ -2631,21 +2651,10 @@ static inline int inc_valid_node_count(struct f2fs_sb_info *sbi,
 
 	spin_lock(&sbi->stat_lock);
 
-	valid_block_count = sbi->total_valid_block_count +
-					sbi->current_reserved_blocks + 1;
-
-	if (!__allow_reserved_blocks(sbi, inode, false))
-		valid_block_count += F2FS_OPTION(sbi).root_reserved_blocks;
+	valid_block_count = sbi->total_valid_block_count + 1;
+	avail_user_block_count = get_available_block_count(sbi, inode, false);
 
-	if (F2FS_IO_ALIGNED(sbi))
-		valid_block_count += sbi->blocks_per_seg *
-				SM_I(sbi)->additional_reserved_segments;
-
-	user_block_count = sbi->user_block_count;
-	if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED)))
-		user_block_count -= sbi->unusable_block_count;
-
-	if (unlikely(valid_block_count > user_block_count)) {
+	if (unlikely(valid_block_count > avail_user_block_count)) {
 		spin_unlock(&sbi->stat_lock);
 		goto enospc;
 	}
@@ -3126,7 +3135,7 @@ static inline void f2fs_i_depth_write(struct inode *inode, unsigned int depth)
 static inline void f2fs_i_gc_failures_write(struct inode *inode,
 					unsigned int count)
 {
-	F2FS_I(inode)->i_gc_failures[GC_FAILURE_PIN] = count;
+	F2FS_I(inode)->i_gc_failures = count;
 	f2fs_mark_inode_dirty_sync(inode, true);
 }
 
@@ -3376,17 +3385,6 @@ static inline bool f2fs_cp_error(struct f2fs_sb_info *sbi)
 	return is_set_ckpt_flags(sbi, CP_ERROR_FLAG);
 }
 
-static inline bool is_dot_dotdot(const u8 *name, size_t len)
-{
-	if (len == 1 && name[0] == '.')
-		return true;
-
-	if (len == 2 && name[0] == '.' && name[1] == '.')
-		return true;
-
-	return false;
-}
-
 static inline void *f2fs_kmalloc(struct f2fs_sb_info *sbi,
 					size_t size, gfp_t flags)
 {
@@ -3467,7 +3465,7 @@ static inline __le32 *get_dnode_addr(struct inode *inode,
 		sizeof((f2fs_inode)->field))			\
 		<= (F2FS_OLD_ATTRIBUTE_SIZE + (extra_isize)))	\
 
-#define __is_large_section(sbi)		((sbi)->segs_per_sec > 1)
+#define __is_large_section(sbi)		(SEGS_PER_SEC(sbi) > 1)
 
 #define __is_meta_io(fio) (PAGE_TYPE_OF_BIO((fio)->type) == META)
 
@@ -3502,6 +3500,8 @@ int f2fs_setattr(struct mnt_idmap *idmap, struct dentry *dentry,
 		 struct iattr *attr);
 int f2fs_truncate_hole(struct inode *inode, pgoff_t pg_start, pgoff_t pg_end);
 void f2fs_truncate_data_blocks_range(struct dnode_of_data *dn, int count);
+int f2fs_do_shutdown(struct f2fs_sb_info *sbi, unsigned int flag,
+							bool readonly);
 int f2fs_precache_extents(struct inode *inode);
 int f2fs_fileattr_get(struct dentry *dentry, struct fileattr *fa);
 int f2fs_fileattr_set(struct mnt_idmap *idmap,
@@ -3696,15 +3696,14 @@ int f2fs_disable_cp_again(struct f2fs_sb_info *sbi, block_t unusable);
 void f2fs_release_discard_addrs(struct f2fs_sb_info *sbi);
 int f2fs_npages_for_summary_flush(struct f2fs_sb_info *sbi, bool for_ra);
 bool f2fs_segment_has_free_slot(struct f2fs_sb_info *sbi, int segno);
-void f2fs_init_inmem_curseg(struct f2fs_sb_info *sbi);
+int f2fs_init_inmem_curseg(struct f2fs_sb_info *sbi);
 void f2fs_save_inmem_curseg(struct f2fs_sb_info *sbi);
 void f2fs_restore_inmem_curseg(struct f2fs_sb_info *sbi);
-void f2fs_get_new_segment(struct f2fs_sb_info *sbi,
-			unsigned int *newseg, bool new_sec, int dir);
-void f2fs_allocate_segment_for_resize(struct f2fs_sb_info *sbi, int type,
+int f2fs_allocate_segment_for_resize(struct f2fs_sb_info *sbi, int type,
 					unsigned int start, unsigned int end);
-void f2fs_allocate_new_section(struct f2fs_sb_info *sbi, int type, bool force);
-void f2fs_allocate_new_segments(struct f2fs_sb_info *sbi);
+int f2fs_allocate_new_section(struct f2fs_sb_info *sbi, int type, bool force);
+int f2fs_allocate_pinning_section(struct f2fs_sb_info *sbi);
+int f2fs_allocate_new_segments(struct f2fs_sb_info *sbi);
 int f2fs_trim_fs(struct f2fs_sb_info *sbi, struct fstrim_range *range);
 bool f2fs_exist_trim_candidates(struct f2fs_sb_info *sbi,
 					struct cp_control *cpc);
@@ -3725,7 +3724,8 @@ void f2fs_replace_block(struct f2fs_sb_info *sbi, struct dnode_of_data *dn,
 			block_t old_addr, block_t new_addr,
 			unsigned char version, bool recover_curseg,
 			bool recover_newaddr);
-void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
+int f2fs_get_segment_temp(int seg_type);
+int f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
 			block_t old_blkaddr, block_t *new_blkaddr,
 			struct f2fs_summary *sum, int type,
 			struct f2fs_io_info *fio);
@@ -3747,7 +3747,9 @@ int f2fs_build_segment_manager(struct f2fs_sb_info *sbi);
 void f2fs_destroy_segment_manager(struct f2fs_sb_info *sbi);
 int __init f2fs_create_segment_manager_caches(void);
 void f2fs_destroy_segment_manager_caches(void);
-int f2fs_rw_hint_to_seg_type(enum rw_hint hint);
+int f2fs_rw_hint_to_seg_type(struct f2fs_sb_info *sbi, enum rw_hint hint);
+enum rw_hint f2fs_io_type_to_rw_hint(struct f2fs_sb_info *sbi,
+			enum page_type type, enum temp_type temp);
 unsigned int f2fs_usable_segs_in_sec(struct f2fs_sb_info *sbi,
 			unsigned int segno);
 unsigned int f2fs_usable_blks_in_seg(struct f2fs_sb_info *sbi,
@@ -3775,6 +3777,8 @@ struct page *f2fs_get_meta_page_retry(struct f2fs_sb_info *sbi, pgoff_t index);
 struct page *f2fs_get_tmp_page(struct f2fs_sb_info *sbi, pgoff_t index);
 bool f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi,
 					block_t blkaddr, int type);
+bool f2fs_is_valid_blkaddr_raw(struct f2fs_sb_info *sbi,
+					block_t blkaddr, int type);
 int f2fs_ra_meta_pages(struct f2fs_sb_info *sbi, block_t start, int nrpages,
 			int type, bool sync);
 void f2fs_ra_meta_pages_cond(struct f2fs_sb_info *sbi, pgoff_t index,
@@ -3879,6 +3883,9 @@ void f2fs_stop_gc_thread(struct f2fs_sb_info *sbi);
 block_t f2fs_start_bidx_of_node(unsigned int node_ofs, struct inode *inode);
 int f2fs_gc(struct f2fs_sb_info *sbi, struct f2fs_gc_control *gc_control);
 void f2fs_build_gc_manager(struct f2fs_sb_info *sbi);
+int f2fs_gc_range(struct f2fs_sb_info *sbi,
+		unsigned int start_seg, unsigned int end_seg,
+		bool dry_run, unsigned int dry_run_sections);
 int f2fs_resize_fs(struct file *filp, __u64 block_count);
 int __init f2fs_create_garbage_collection_cache(void);
 void f2fs_destroy_garbage_collection_cache(void);
@@ -4149,10 +4156,10 @@ extern struct kmem_cache *f2fs_inode_entry_slab;
 bool f2fs_may_inline_data(struct inode *inode);
 bool f2fs_sanity_check_inline_data(struct inode *inode);
 bool f2fs_may_inline_dentry(struct inode *inode);
-void f2fs_do_read_inline_data(struct page *page, struct page *ipage);
+void f2fs_do_read_inline_data(struct folio *folio, struct page *ipage);
 void f2fs_truncate_inline_inode(struct inode *inode,
 						struct page *ipage, u64 from);
-int f2fs_read_inline_data(struct inode *inode, struct page *page);
+int f2fs_read_inline_data(struct inode *inode, struct folio *folio);
 int f2fs_convert_inline_page(struct dnode_of_data *dn, struct page *page);
 int f2fs_convert_inline_inode(struct inode *inode);
 int f2fs_try_convert_inline_dir(struct inode *dir, struct dentry *dentry);
@@ -4188,7 +4195,7 @@ void f2fs_leave_shrinker(struct f2fs_sb_info *sbi);
 /*
  * extent_cache.c
  */
-bool sanity_check_extent_cache(struct inode *inode);
+bool sanity_check_extent_cache(struct inode *inode, struct page *ipage);
 void f2fs_init_extent_tree(struct inode *inode);
 void f2fs_drop_extent_tree(struct inode *inode);
 void f2fs_destroy_extent_node(struct inode *inode);
@@ -4259,6 +4266,16 @@ static inline bool f2fs_post_read_required(struct inode *inode)
 		f2fs_compressed_file(inode);
 }
 
+static inline bool f2fs_used_in_atomic_write(struct inode *inode)
+{
+	return f2fs_is_atomic_file(inode) || f2fs_is_cow_file(inode);
+}
+
+static inline bool f2fs_meta_inode_gc_required(struct inode *inode)
+{
+	return f2fs_post_read_required(inode) || f2fs_used_in_atomic_write(inode);
+}
+
 /*
  * compress.c
  */
@@ -4535,6 +4552,17 @@ static inline bool f2fs_lfs_mode(struct f2fs_sb_info *sbi)
 	return F2FS_OPTION(sbi).fs_mode == FS_MODE_LFS;
 }
 
+static inline bool f2fs_valid_pinned_area(struct f2fs_sb_info *sbi,
+					  block_t blkaddr)
+{
+	if (f2fs_sb_has_blkzoned(sbi)) {
+		int devi = f2fs_target_device_index(sbi, blkaddr);
+
+		return !bdev_is_zoned(FDEV(devi).bdev);
+	}
+	return true;
+}
+
 static inline bool f2fs_low_mem_mode(struct f2fs_sb_info *sbi)
 {
 	return F2FS_OPTION(sbi).memory_mode == MEMORY_MODE_LOW;
@@ -4586,10 +4614,14 @@ static inline bool f2fs_need_verity(const struct inode *inode, pgoff_t idx)
 }
 
 #ifdef CONFIG_F2FS_FAULT_INJECTION
-extern void f2fs_build_fault_attr(struct f2fs_sb_info *sbi, unsigned int rate,
-							unsigned int type);
+extern int f2fs_build_fault_attr(struct f2fs_sb_info *sbi, unsigned long rate,
+							unsigned long type);
 #else
-#define f2fs_build_fault_attr(sbi, rate, type)		do { } while (0)
+static inline int f2fs_build_fault_attr(struct f2fs_sb_info *sbi,
+					unsigned long rate, unsigned long type)
+{
+	return 0;
+}
 #endif
 
 static inline bool is_journalled_quota(struct f2fs_sb_info *sbi)
@@ -4647,7 +4679,7 @@ static inline void f2fs_truncate_meta_inode_pages(struct f2fs_sb_info *sbi,
 
 		page = find_get_page(META_MAPPING(sbi), blkaddr + i);
 		if (page) {
-			if (PageWriteback(page))
+			if (folio_test_writeback(page_folio(page)))
 				need_submit = true;
 			f2fs_put_page(page, 0);
 		}
diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index 51970b345e..387ce167dd 100644
--- a/fs/f2fs/file.c
+++ b/fs/f2fs/file.c
@@ -58,7 +58,7 @@ static vm_fault_t f2fs_vm_page_mkwrite(struct vm_fault *vmf)
 	struct inode *inode = file_inode(vmf->vma->vm_file);
 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 	struct dnode_of_data dn;
-	bool need_alloc = true;
+	bool need_alloc = !f2fs_is_pinned_file(inode);
 	int err = 0;
 	vm_fault_t ret;
 
@@ -115,19 +115,18 @@ static vm_fault_t f2fs_vm_page_mkwrite(struct vm_fault *vmf)
 		goto out_sem;
 	}
 
+	set_new_dnode(&dn, inode, NULL, NULL, 0);
 	if (need_alloc) {
 		/* block allocation */
-		set_new_dnode(&dn, inode, NULL, NULL, 0);
 		err = f2fs_get_block_locked(&dn, page->index);
-	}
-
-#ifdef CONFIG_F2FS_FS_COMPRESSION
-	if (!need_alloc) {
-		set_new_dnode(&dn, inode, NULL, NULL, 0);
+	} else {
 		err = f2fs_get_dnode_of_data(&dn, page->index, LOOKUP_NODE);
 		f2fs_put_dnode(&dn);
+		if (f2fs_is_pinned_file(inode) &&
+		    !__is_valid_data_blkaddr(dn.data_blkaddr))
+			err = -EIO;
 	}
-#endif
+
 	if (err) {
 		unlock_page(page);
 		goto out_sem;
@@ -395,9 +394,20 @@ int f2fs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	return f2fs_do_sync_file(file, start, end, datasync, false);
 }
 
-static bool __found_offset(struct address_space *mapping, block_t blkaddr,
-				pgoff_t index, int whence)
+static bool __found_offset(struct address_space *mapping,
+		struct dnode_of_data *dn, pgoff_t index, int whence)
 {
+	block_t blkaddr = f2fs_data_blkaddr(dn);
+	struct inode *inode = mapping->host;
+	bool compressed_cluster = false;
+
+	if (f2fs_compressed_file(inode)) {
+		block_t first_blkaddr = data_blkaddr(dn->inode, dn->node_page,
+		    ALIGN_DOWN(dn->ofs_in_node, F2FS_I(inode)->i_cluster_size));
+
+		compressed_cluster = first_blkaddr == COMPRESS_ADDR;
+	}
+
 	switch (whence) {
 	case SEEK_DATA:
 		if (__is_valid_data_blkaddr(blkaddr))
@@ -405,8 +415,12 @@ static bool __found_offset(struct address_space *mapping, block_t blkaddr,
 		if (blkaddr == NEW_ADDR &&
 		    xa_get_mark(&mapping->i_pages, index, PAGECACHE_TAG_DIRTY))
 			return true;
+		if (compressed_cluster)
+			return true;
 		break;
 	case SEEK_HOLE:
+		if (compressed_cluster)
+			return false;
 		if (blkaddr == NULL_ADDR)
 			return true;
 		break;
@@ -475,7 +489,7 @@ static loff_t f2fs_seek_block(struct file *file, loff_t offset, int whence)
 				goto fail;
 			}
 
-			if (__found_offset(file->f_mapping, blkaddr,
+			if (__found_offset(file->f_mapping, &dn,
 							pgofs, whence)) {
 				f2fs_put_dnode(&dn);
 				goto found;
@@ -540,6 +554,42 @@ static int f2fs_file_mmap(struct file *file, struct vm_area_struct *vma)
 	return 0;
 }
 
+static int finish_preallocate_blocks(struct inode *inode)
+{
+	int ret;
+
+	inode_lock(inode);
+	if (is_inode_flag_set(inode, FI_OPENED_FILE)) {
+		inode_unlock(inode);
+		return 0;
+	}
+
+	if (!file_should_truncate(inode)) {
+		set_inode_flag(inode, FI_OPENED_FILE);
+		inode_unlock(inode);
+		return 0;
+	}
+
+	f2fs_down_write(&F2FS_I(inode)->i_gc_rwsem[WRITE]);
+	filemap_invalidate_lock(inode->i_mapping);
+
+	truncate_setsize(inode, i_size_read(inode));
+	ret = f2fs_truncate(inode);
+
+	filemap_invalidate_unlock(inode->i_mapping);
+	f2fs_up_write(&F2FS_I(inode)->i_gc_rwsem[WRITE]);
+
+	if (!ret)
+		set_inode_flag(inode, FI_OPENED_FILE);
+
+	inode_unlock(inode);
+	if (ret)
+		return ret;
+
+	file_dont_truncate(inode);
+	return 0;
+}
+
 static int f2fs_file_open(struct inode *inode, struct file *filp)
 {
 	int err = fscrypt_file_open(inode, filp);
@@ -554,10 +604,14 @@ static int f2fs_file_open(struct inode *inode, struct file *filp)
 	if (err)
 		return err;
 
-	filp->f_mode |= FMODE_NOWAIT | FMODE_BUF_RASYNC;
+	filp->f_mode |= FMODE_NOWAIT;
 	filp->f_mode |= FMODE_CAN_ODIRECT;
 
-	return dquot_file_open(inode, filp);
+	err = dquot_file_open(inode, filp);
+	if (err)
+		return err;
+
+	return finish_preallocate_blocks(inode);
 }
 
 void f2fs_truncate_data_blocks_range(struct dnode_of_data *dn, int count)
@@ -591,8 +645,10 @@ void f2fs_truncate_data_blocks_range(struct dnode_of_data *dn, int count)
 		f2fs_set_data_blkaddr(dn, NULL_ADDR);
 
 		if (__is_valid_data_blkaddr(blkaddr)) {
-			if (!f2fs_is_valid_blkaddr(sbi, blkaddr,
-					DATA_GENERIC_ENHANCE))
+			if (time_to_inject(sbi, FAULT_BLKADDR_CONSISTENCE))
+				continue;
+			if (!f2fs_is_valid_blkaddr_raw(sbi, blkaddr,
+						DATA_GENERIC_ENHANCE))
 				continue;
 			if (compressed_cluster)
 				valid_blocks++;
@@ -809,6 +865,8 @@ static bool f2fs_force_buffered_io(struct inode *inode, int rw)
 		return true;
 	if (f2fs_compressed_file(inode))
 		return true;
+	if (f2fs_has_inline_data(inode))
+		return true;
 
 	/* disallow direct IO if any of devices has unaligned blksize */
 	if (f2fs_is_multi_device(sbi) && !sbi->aligned_blksize)
@@ -817,9 +875,8 @@ static bool f2fs_force_buffered_io(struct inode *inode, int rw)
 	 * for blkzoned device, fallback direct IO to buffered IO, so
 	 * all IOs can be serialized by log-structured write.
 	 */
-	if (f2fs_sb_has_blkzoned(sbi) && (rw == WRITE))
-		return true;
-	if (f2fs_lfs_mode(sbi) && rw == WRITE && F2FS_IO_ALIGNED(sbi))
+	if (f2fs_sb_has_blkzoned(sbi) && (rw == WRITE) &&
+	    !f2fs_is_pinned_file(inode))
 		return true;
 	if (is_sbi_flag_set(sbi, SBI_CP_DISABLED))
 		return true;
@@ -937,9 +994,14 @@ int f2fs_setattr(struct mnt_idmap *idmap, struct dentry *dentry,
 				  ATTR_GID | ATTR_TIMES_SET))))
 		return -EPERM;
 
-	if ((attr->ia_valid & ATTR_SIZE) &&
-		!f2fs_is_compress_backend_ready(inode))
-		return -EOPNOTSUPP;
+	if ((attr->ia_valid & ATTR_SIZE)) {
+		if (!f2fs_is_compress_backend_ready(inode))
+			return -EOPNOTSUPP;
+		if (is_inode_flag_set(inode, FI_COMPRESS_RELEASED) &&
+			!IS_ALIGNED(attr->ia_size,
+			F2FS_BLK_TO_BYTES(F2FS_I(inode)->i_cluster_size)))
+			return -EINVAL;
+	}
 
 	err = setattr_prepare(idmap, dentry, attr);
 	if (err)
@@ -1193,7 +1255,6 @@ next_dnode:
 			!f2fs_is_valid_blkaddr(sbi, *blkaddr,
 					DATA_GENERIC_ENHANCE)) {
 			f2fs_put_dnode(&dn);
-			f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR);
 			return -EFSCORRUPTED;
 		}
 
@@ -1311,6 +1372,9 @@ static int __clone_blkaddrs(struct inode *src_inode, struct inode *dst_inode,
 				f2fs_put_page(psrc, 1);
 				return PTR_ERR(pdst);
 			}
+
+			f2fs_wait_on_page_writeback(pdst, DATA, true, true);
+
 			memcpy_page(pdst, 0, psrc, 0, PAGE_SIZE);
 			set_page_dirty(pdst);
 			set_page_private_gcing(pdst);
@@ -1479,7 +1543,6 @@ static int f2fs_do_zero_range(struct dnode_of_data *dn, pgoff_t start,
 		if (!f2fs_is_valid_blkaddr(sbi, dn->data_blkaddr,
 					DATA_GENERIC_ENHANCE)) {
 			ret = -EFSCORRUPTED;
-			f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR);
 			break;
 		}
 
@@ -1663,10 +1726,12 @@ static int f2fs_insert_range(struct inode *inode, loff_t offset, loff_t len)
 	}
 	filemap_invalidate_unlock(mapping);
 	f2fs_up_write(&F2FS_I(inode)->i_gc_rwsem[WRITE]);
+	if (ret)
+		return ret;
 
 	/* write out all moved pages, if possible */
 	filemap_invalidate_lock(mapping);
-	filemap_write_and_wait_range(mapping, offset, LLONG_MAX);
+	ret = filemap_write_and_wait_range(mapping, offset, LLONG_MAX);
 	truncate_pagecache(inode, offset);
 	filemap_invalidate_unlock(mapping);
 
@@ -1732,9 +1797,11 @@ next_alloc:
 
 		f2fs_down_write(&sbi->pin_sem);
 
-		f2fs_lock_op(sbi);
-		f2fs_allocate_new_section(sbi, CURSEG_COLD_DATA_PINNED, false);
-		f2fs_unlock_op(sbi);
+		err = f2fs_allocate_pinning_section(sbi);
+		if (err) {
+			f2fs_up_write(&sbi->pin_sem);
+			goto out_err;
+		}
 
 		map.m_seg_type = CURSEG_COLD_DATA_PINNED;
 		err = f2fs_map_blocks(inode, &map, F2FS_GET_BLOCK_PRE_DIO);
@@ -1800,15 +1867,6 @@ static long f2fs_fallocate(struct file *file, int mode,
 		(mode & (FALLOC_FL_COLLAPSE_RANGE | FALLOC_FL_INSERT_RANGE)))
 		return -EOPNOTSUPP;
 
-	/*
-	 * Pinned file should not support partial truncation since the block
-	 * can be used by applications.
-	 */
-	if ((f2fs_compressed_file(inode) || f2fs_is_pinned_file(inode)) &&
-		(mode & (FALLOC_FL_PUNCH_HOLE | FALLOC_FL_COLLAPSE_RANGE |
-			FALLOC_FL_ZERO_RANGE | FALLOC_FL_INSERT_RANGE)))
-		return -EOPNOTSUPP;
-
 	if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE |
 			FALLOC_FL_COLLAPSE_RANGE | FALLOC_FL_ZERO_RANGE |
 			FALLOC_FL_INSERT_RANGE))
@@ -1816,6 +1874,17 @@ static long f2fs_fallocate(struct file *file, int mode,
 
 	inode_lock(inode);
 
+	/*
+	 * Pinned file should not support partial truncation since the block
+	 * can be used by applications.
+	 */
+	if ((f2fs_compressed_file(inode) || f2fs_is_pinned_file(inode)) &&
+		(mode & (FALLOC_FL_PUNCH_HOLE | FALLOC_FL_COLLAPSE_RANGE |
+			FALLOC_FL_ZERO_RANGE | FALLOC_FL_INSERT_RANGE))) {
+		ret = -EOPNOTSUPP;
+		goto out;
+	}
+
 	ret = file_modified(file);
 	if (ret)
 		goto out;
@@ -2067,7 +2136,8 @@ static int f2fs_ioc_start_atomic_write(struct file *filp, bool truncate)
 
 	inode_lock(inode);
 
-	if (!f2fs_disable_compressed_file(inode)) {
+	if (!f2fs_disable_compressed_file(inode) ||
+			f2fs_is_pinned_file(inode)) {
 		ret = -EINVAL;
 		goto out;
 	}
@@ -2113,6 +2183,9 @@ static int f2fs_ioc_start_atomic_write(struct file *filp, bool truncate)
 
 		set_inode_flag(fi->cow_inode, FI_COW_FILE);
 		clear_inode_flag(fi->cow_inode, FI_INLINE_DATA);
+
+		/* Set the COW inode's atomic_inode to the atomic inode */
+		F2FS_I(fi->cow_inode)->atomic_inode = inode;
 	} else {
 		/* Reuse the already created COW inode */
 		ret = f2fs_do_truncate_blocks(fi->cow_inode, 0, true);
@@ -2206,34 +2279,13 @@ static int f2fs_ioc_abort_atomic_write(struct file *filp)
 	return ret;
 }
 
-static int f2fs_ioc_shutdown(struct file *filp, unsigned long arg)
+int f2fs_do_shutdown(struct f2fs_sb_info *sbi, unsigned int flag,
+							bool readonly)
 {
-	struct inode *inode = file_inode(filp);
-	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 	struct super_block *sb = sbi->sb;
-	__u32 in;
 	int ret = 0;
 
-	if (!capable(CAP_SYS_ADMIN))
-		return -EPERM;
-
-	if (get_user(in, (__u32 __user *)arg))
-		return -EFAULT;
-
-	if (in != F2FS_GOING_DOWN_FULLSYNC) {
-		ret = mnt_want_write_file(filp);
-		if (ret) {
-			if (ret == -EROFS) {
-				ret = 0;
-				f2fs_stop_checkpoint(sbi, false,
-						STOP_CP_REASON_SHUTDOWN);
-				trace_f2fs_shutdown(sbi, in, ret);
-			}
-			return ret;
-		}
-	}
-
-	switch (in) {
+	switch (flag) {
 	case F2FS_GOING_DOWN_FULLSYNC:
 		ret = bdev_freeze(sb->s_bdev);
 		if (ret)
@@ -2244,8 +2296,11 @@ static int f2fs_ioc_shutdown(struct file *filp, unsigned long arg)
 	case F2FS_GOING_DOWN_METASYNC:
 		/* do checkpoint only */
 		ret = f2fs_sync_fs(sb, 1);
-		if (ret)
+		if (ret) {
+			if (ret == -EIO)
+				ret = 0;
 			goto out;
+		}
 		f2fs_stop_checkpoint(sbi, false, STOP_CP_REASON_SHUTDOWN);
 		break;
 	case F2FS_GOING_DOWN_NOSYNC:
@@ -2261,12 +2316,17 @@ static int f2fs_ioc_shutdown(struct file *filp, unsigned long arg)
 		set_sbi_flag(sbi, SBI_IS_DIRTY);
 		/* do checkpoint only */
 		ret = f2fs_sync_fs(sb, 1);
+		if (ret == -EIO)
+			ret = 0;
 		goto out;
 	default:
 		ret = -EINVAL;
 		goto out;
 	}
 
+	if (readonly)
+		goto out;
+
 	f2fs_stop_gc_thread(sbi);
 	f2fs_stop_discard_thread(sbi);
 
@@ -2275,10 +2335,44 @@ static int f2fs_ioc_shutdown(struct file *filp, unsigned long arg)
 
 	f2fs_update_time(sbi, REQ_TIME);
 out:
-	if (in != F2FS_GOING_DOWN_FULLSYNC)
-		mnt_drop_write_file(filp);
 
-	trace_f2fs_shutdown(sbi, in, ret);
+	trace_f2fs_shutdown(sbi, flag, ret);
+
+	return ret;
+}
+
+static int f2fs_ioc_shutdown(struct file *filp, unsigned long arg)
+{
+	struct inode *inode = file_inode(filp);
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+	__u32 in;
+	int ret;
+	bool need_drop = false, readonly = false;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	if (get_user(in, (__u32 __user *)arg))
+		return -EFAULT;
+
+	if (in != F2FS_GOING_DOWN_FULLSYNC) {
+		ret = mnt_want_write_file(filp);
+		if (ret) {
+			if (ret != -EROFS)
+				return ret;
+
+			/* fallback to nosync shutdown for readonly fs */
+			in = F2FS_GOING_DOWN_NOSYNC;
+			readonly = true;
+		} else {
+			need_drop = true;
+		}
+	}
+
+	ret = f2fs_do_shutdown(sbi, in, readonly);
+
+	if (need_drop)
+		mnt_drop_write_file(filp);
 
 	return ret;
 }
@@ -2331,13 +2425,14 @@ static bool uuid_is_nonzero(__u8 u[16])
 static int f2fs_ioc_set_encryption_policy(struct file *filp, unsigned long arg)
 {
 	struct inode *inode = file_inode(filp);
+	int ret;
 
 	if (!f2fs_sb_has_encrypt(F2FS_I_SB(inode)))
 		return -EOPNOTSUPP;
 
+	ret = fscrypt_ioctl_set_policy(filp, (const void __user *)arg);
 	f2fs_update_time(F2FS_I_SB(inode), REQ_TIME);
-
-	return fscrypt_ioctl_set_policy(filp, (const void __user *)arg);
+	return ret;
 }
 
 static int f2fs_ioc_get_encryption_policy(struct file *filp, unsigned long arg)
@@ -2579,18 +2674,18 @@ static int f2fs_defragment_range(struct f2fs_sb_info *sbi,
 					.m_may_create = false };
 	struct extent_info ei = {};
 	pgoff_t pg_start, pg_end, next_pgofs;
-	unsigned int blk_per_seg = sbi->blocks_per_seg;
 	unsigned int total = 0, sec_num;
 	block_t blk_end = 0;
 	bool fragmented = false;
 	int err;
 
-	pg_start = range->start >> PAGE_SHIFT;
-	pg_end = (range->start + range->len) >> PAGE_SHIFT;
-
 	f2fs_balance_fs(sbi, true);
 
 	inode_lock(inode);
+	pg_start = range->start >> PAGE_SHIFT;
+	pg_end = min_t(pgoff_t,
+				(range->start + range->len) >> PAGE_SHIFT,
+				DIV_ROUND_UP(i_size_read(inode), PAGE_SIZE));
 
 	if (is_inode_flag_set(inode, FI_COMPRESS_RELEASED)) {
 		err = -EINVAL;
@@ -2605,8 +2700,9 @@ static int f2fs_defragment_range(struct f2fs_sb_info *sbi,
 	}
 
 	/* writeback all dirty pages in the range */
-	err = filemap_write_and_wait_range(inode->i_mapping, range->start,
-						range->start + range->len - 1);
+	err = filemap_write_and_wait_range(inode->i_mapping,
+						pg_start << PAGE_SHIFT,
+						(pg_end << PAGE_SHIFT) - 1);
 	if (err)
 		goto out;
 
@@ -2688,7 +2784,8 @@ do_map:
 		set_inode_flag(inode, FI_SKIP_WRITES);
 
 		idx = map.m_lblk;
-		while (idx < map.m_lblk + map.m_len && cnt < blk_per_seg) {
+		while (idx < map.m_lblk + map.m_len &&
+						cnt < BLKS_PER_SEG(sbi)) {
 			struct page *page;
 
 			page = f2fs_get_lock_data_page(inode, idx, true);
@@ -2708,7 +2805,7 @@ do_map:
 
 		map.m_lblk = idx;
 check:
-		if (map.m_lblk < pg_end && cnt < blk_per_seg)
+		if (map.m_lblk < pg_end && cnt < BLKS_PER_SEG(sbi))
 			goto do_map;
 
 		clear_inode_flag(inode, FI_SKIP_WRITES);
@@ -2763,7 +2860,8 @@ static int f2fs_ioc_defragment(struct file *filp, unsigned long arg)
 	err = f2fs_defragment_range(sbi, filp, &range);
 	mnt_drop_write_file(filp);
 
-	f2fs_update_time(sbi, REQ_TIME);
+	if (range.len)
+		f2fs_update_time(sbi, REQ_TIME);
 	if (err < 0)
 		return err;
 
@@ -2814,7 +2912,8 @@ static int f2fs_move_file_range(struct file *file_in, loff_t pos_in,
 			goto out;
 	}
 
-	if (f2fs_compressed_file(src) || f2fs_compressed_file(dst)) {
+	if (f2fs_compressed_file(src) || f2fs_compressed_file(dst) ||
+		f2fs_is_pinned_file(src) || f2fs_is_pinned_file(dst)) {
 		ret = -EOPNOTSUPP;
 		goto out_unlock;
 	}
@@ -2977,8 +3076,8 @@ static int f2fs_ioc_flush_device(struct file *filp, unsigned long arg)
 
 	if (!f2fs_is_multi_device(sbi) || sbi->s_ndevs - 1 <= range.dev_num ||
 			__is_large_section(sbi)) {
-		f2fs_warn(sbi, "Can't flush %u in %d for segs_per_sec %u != 1",
-			  range.dev_num, sbi->s_ndevs, sbi->segs_per_sec);
+		f2fs_warn(sbi, "Can't flush %u in %d for SEGS_PER_SEC %u != 1",
+			  range.dev_num, sbi->s_ndevs, SEGS_PER_SEC(sbi));
 		return -EINVAL;
 	}
 
@@ -3166,24 +3265,24 @@ int f2fs_pin_file_control(struct inode *inode, bool inc)
 	struct f2fs_inode_info *fi = F2FS_I(inode);
 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 
-	/* Use i_gc_failures for normal file as a risk signal. */
-	if (inc)
-		f2fs_i_gc_failures_write(inode,
-				fi->i_gc_failures[GC_FAILURE_PIN] + 1);
-
-	if (fi->i_gc_failures[GC_FAILURE_PIN] > sbi->gc_pin_file_threshold) {
+	if (fi->i_gc_failures >= sbi->gc_pin_file_threshold) {
 		f2fs_warn(sbi, "%s: Enable GC = ino %lx after %x GC trials",
-			  __func__, inode->i_ino,
-			  fi->i_gc_failures[GC_FAILURE_PIN]);
+			  __func__, inode->i_ino, fi->i_gc_failures);
 		clear_inode_flag(inode, FI_PIN_FILE);
 		return -EAGAIN;
 	}
+
+	/* Use i_gc_failures for normal file as a risk signal. */
+	if (inc)
+		f2fs_i_gc_failures_write(inode, fi->i_gc_failures + 1);
+
 	return 0;
 }
 
 static int f2fs_ioc_set_pin_file(struct file *filp, unsigned long arg)
 {
 	struct inode *inode = file_inode(filp);
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 	__u32 pin;
 	int ret = 0;
 
@@ -3193,7 +3292,7 @@ static int f2fs_ioc_set_pin_file(struct file *filp, unsigned long arg)
 	if (!S_ISREG(inode->i_mode))
 		return -EINVAL;
 
-	if (f2fs_readonly(F2FS_I_SB(inode)->sb))
+	if (f2fs_readonly(sbi->sb))
 		return -EROFS;
 
 	ret = mnt_want_write_file(filp);
@@ -3206,9 +3305,18 @@ static int f2fs_ioc_set_pin_file(struct file *filp, unsigned long arg)
 		clear_inode_flag(inode, FI_PIN_FILE);
 		f2fs_i_gc_failures_write(inode, 0);
 		goto done;
+	} else if (f2fs_is_pinned_file(inode)) {
+		goto done;
 	}
 
-	if (f2fs_should_update_outplace(inode, NULL)) {
+	if (F2FS_HAS_BLOCKS(inode)) {
+		ret = -EFBIG;
+		goto out;
+	}
+
+	/* Let's allow file pinning on zoned device. */
+	if (!f2fs_sb_has_blkzoned(sbi) &&
+	    f2fs_should_update_outplace(inode, NULL)) {
 		ret = -EINVAL;
 		goto out;
 	}
@@ -3228,9 +3336,9 @@ static int f2fs_ioc_set_pin_file(struct file *filp, unsigned long arg)
 	}
 
 	set_inode_flag(inode, FI_PIN_FILE);
-	ret = F2FS_I(inode)->i_gc_failures[GC_FAILURE_PIN];
+	ret = F2FS_I(inode)->i_gc_failures;
 done:
-	f2fs_update_time(F2FS_I_SB(inode), REQ_TIME);
+	f2fs_update_time(sbi, REQ_TIME);
 out:
 	inode_unlock(inode);
 	mnt_drop_write_file(filp);
@@ -3243,7 +3351,7 @@ static int f2fs_ioc_get_pin_file(struct file *filp, unsigned long arg)
 	__u32 pin = 0;
 
 	if (is_inode_flag_set(inode, FI_PIN_FILE))
-		pin = F2FS_I(inode)->i_gc_failures[GC_FAILURE_PIN];
+		pin = F2FS_I(inode)->i_gc_failures;
 	return put_user(pin, (u32 __user *)arg);
 }
 
@@ -3439,10 +3547,8 @@ static int release_compress_blocks(struct dnode_of_data *dn, pgoff_t count)
 		if (!__is_valid_data_blkaddr(blkaddr))
 			continue;
 		if (unlikely(!f2fs_is_valid_blkaddr(sbi, blkaddr,
-					DATA_GENERIC_ENHANCE))) {
-			f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR);
+					DATA_GENERIC_ENHANCE)))
 			return -EFSCORRUPTED;
-		}
 	}
 
 	while (count) {
@@ -3491,9 +3597,6 @@ static int f2fs_release_compress_blocks(struct file *filp, unsigned long arg)
 	if (!f2fs_sb_has_compression(sbi))
 		return -EOPNOTSUPP;
 
-	if (!f2fs_compressed_file(inode))
-		return -EINVAL;
-
 	if (f2fs_readonly(sbi->sb))
 		return -EROFS;
 
@@ -3512,7 +3615,8 @@ static int f2fs_release_compress_blocks(struct file *filp, unsigned long arg)
 		goto out;
 	}
 
-	if (is_inode_flag_set(inode, FI_COMPRESS_RELEASED)) {
+	if (!f2fs_compressed_file(inode) ||
+		is_inode_flag_set(inode, FI_COMPRESS_RELEASED)) {
 		ret = -EINVAL;
 		goto out;
 	}
@@ -3539,9 +3643,12 @@ static int f2fs_release_compress_blocks(struct file *filp, unsigned long arg)
 		struct dnode_of_data dn;
 		pgoff_t end_offset, count;
 
+		f2fs_lock_op(sbi);
+
 		set_new_dnode(&dn, inode, NULL, NULL, 0);
 		ret = f2fs_get_dnode_of_data(&dn, page_idx, LOOKUP_NODE);
 		if (ret) {
+			f2fs_unlock_op(sbi);
 			if (ret == -ENOENT) {
 				page_idx = f2fs_get_next_page_offset(&dn,
 								page_idx);
@@ -3559,6 +3666,8 @@ static int f2fs_release_compress_blocks(struct file *filp, unsigned long arg)
 
 		f2fs_put_dnode(&dn);
 
+		f2fs_unlock_op(sbi);
+
 		if (ret < 0)
 			break;
 
@@ -3569,6 +3678,8 @@ static int f2fs_release_compress_blocks(struct file *filp, unsigned long arg)
 	filemap_invalidate_unlock(inode->i_mapping);
 	f2fs_up_write(&F2FS_I(inode)->i_gc_rwsem[WRITE]);
 out:
+	if (released_blocks)
+		f2fs_update_time(sbi, REQ_TIME);
 	inode_unlock(inode);
 
 	mnt_drop_write_file(filp);
@@ -3604,15 +3715,14 @@ static int reserve_compress_blocks(struct dnode_of_data *dn, pgoff_t count,
 		if (!__is_valid_data_blkaddr(blkaddr))
 			continue;
 		if (unlikely(!f2fs_is_valid_blkaddr(sbi, blkaddr,
-					DATA_GENERIC_ENHANCE))) {
-			f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR);
+					DATA_GENERIC_ENHANCE)))
 			return -EFSCORRUPTED;
-		}
 	}
 
 	while (count) {
 		int compr_blocks = 0;
-		blkcnt_t reserved;
+		blkcnt_t reserved = 0;
+		blkcnt_t to_reserved;
 		int ret;
 
 		for (i = 0; i < cluster_size; i++) {
@@ -3632,20 +3742,26 @@ static int reserve_compress_blocks(struct dnode_of_data *dn, pgoff_t count,
 			 * fails in release_compress_blocks(), so NEW_ADDR
 			 * is a possible case.
 			 */
-			if (blkaddr == NEW_ADDR ||
-				__is_valid_data_blkaddr(blkaddr)) {
+			if (blkaddr == NEW_ADDR) {
+				reserved++;
+				continue;
+			}
+			if (__is_valid_data_blkaddr(blkaddr)) {
 				compr_blocks++;
 				continue;
 			}
 		}
 
-		reserved = cluster_size - compr_blocks;
+		to_reserved = cluster_size - compr_blocks - reserved;
 
 		/* for the case all blocks in cluster were reserved */
-		if (reserved == 1)
+		if (to_reserved == 1) {
+			dn->ofs_in_node += cluster_size;
 			goto next;
+		}
 
-		ret = inc_valid_block_count(sbi, dn->inode, &reserved, false);
+		ret = inc_valid_block_count(sbi, dn->inode,
+						&to_reserved, false);
 		if (unlikely(ret))
 			return ret;
 
@@ -3656,7 +3772,7 @@ static int reserve_compress_blocks(struct dnode_of_data *dn, pgoff_t count,
 
 		f2fs_i_compr_blocks_update(dn->inode, compr_blocks, true);
 
-		*reserved_blocks += reserved;
+		*reserved_blocks += to_reserved;
 next:
 		count -= cluster_size;
 	}
@@ -3675,9 +3791,6 @@ static int f2fs_reserve_compress_blocks(struct file *filp, unsigned long arg)
 	if (!f2fs_sb_has_compression(sbi))
 		return -EOPNOTSUPP;
 
-	if (!f2fs_compressed_file(inode))
-		return -EINVAL;
-
 	if (f2fs_readonly(sbi->sb))
 		return -EROFS;
 
@@ -3689,7 +3802,8 @@ static int f2fs_reserve_compress_blocks(struct file *filp, unsigned long arg)
 
 	inode_lock(inode);
 
-	if (!is_inode_flag_set(inode, FI_COMPRESS_RELEASED)) {
+	if (!f2fs_compressed_file(inode) ||
+		!is_inode_flag_set(inode, FI_COMPRESS_RELEASED)) {
 		ret = -EINVAL;
 		goto unlock_inode;
 	}
@@ -3706,9 +3820,12 @@ static int f2fs_reserve_compress_blocks(struct file *filp, unsigned long arg)
 		struct dnode_of_data dn;
 		pgoff_t end_offset, count;
 
+		f2fs_lock_op(sbi);
+
 		set_new_dnode(&dn, inode, NULL, NULL, 0);
 		ret = f2fs_get_dnode_of_data(&dn, page_idx, LOOKUP_NODE);
 		if (ret) {
+			f2fs_unlock_op(sbi);
 			if (ret == -ENOENT) {
 				page_idx = f2fs_get_next_page_offset(&dn,
 								page_idx);
@@ -3726,6 +3843,8 @@ static int f2fs_reserve_compress_blocks(struct file *filp, unsigned long arg)
 
 		f2fs_put_dnode(&dn);
 
+		f2fs_unlock_op(sbi);
+
 		if (ret < 0)
 			break;
 
@@ -3741,6 +3860,8 @@ static int f2fs_reserve_compress_blocks(struct file *filp, unsigned long arg)
 		f2fs_mark_inode_dirty_sync(inode, true);
 	}
 unlock_inode:
+	if (reserved_blocks)
+		f2fs_update_time(sbi, REQ_TIME);
 	inode_unlock(inode);
 	mnt_drop_write_file(filp);
 
@@ -3749,7 +3870,7 @@ unlock_inode:
 	} else if (reserved_blocks &&
 			atomic_read(&F2FS_I(inode)->i_compr_blocks)) {
 		set_sbi_flag(sbi, SBI_NEED_FSCK);
-		f2fs_warn(sbi, "%s: partial blocks were released i_ino=%lx "
+		f2fs_warn(sbi, "%s: partial blocks were reserved i_ino=%lx "
 			"iblocks=%llu, reserved=%u, compr_blocks=%u, "
 			"run fsck to fix.",
 			__func__, inode->i_ino, inode->i_blocks,
@@ -3889,8 +4010,6 @@ static int f2fs_sec_trim_file(struct file *filp, unsigned long arg)
 						DATA_GENERIC_ENHANCE)) {
 				ret = -EFSCORRUPTED;
 				f2fs_put_dnode(&dn);
-				f2fs_handle_error(sbi,
-						ERROR_INVALID_BLKADDR);
 				goto out;
 			}
 
@@ -3939,6 +4058,7 @@ static int f2fs_sec_trim_file(struct file *filp, unsigned long arg)
 	if (len)
 		ret = f2fs_secure_erase(prev_bdev, inode, prev_index,
 				prev_block, len, range.flags);
+	f2fs_update_time(sbi, REQ_TIME);
 out:
 	filemap_invalidate_unlock(mapping);
 	f2fs_up_write(&F2FS_I(inode)->i_gc_rwsem[WRITE]);
@@ -4082,7 +4202,6 @@ static int f2fs_ioc_decompress_file(struct file *filp)
 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 	struct f2fs_inode_info *fi = F2FS_I(inode);
 	pgoff_t page_idx = 0, last_idx;
-	unsigned int blk_per_seg = sbi->blocks_per_seg;
 	int cluster_size = fi->i_cluster_size;
 	int count, ret;
 
@@ -4093,9 +4212,6 @@ static int f2fs_ioc_decompress_file(struct file *filp)
 	if (!(filp->f_mode & FMODE_WRITE))
 		return -EBADF;
 
-	if (!f2fs_compressed_file(inode))
-		return -EINVAL;
-
 	f2fs_balance_fs(sbi, true);
 
 	file_start_write(filp);
@@ -4106,7 +4222,8 @@ static int f2fs_ioc_decompress_file(struct file *filp)
 		goto out;
 	}
 
-	if (is_inode_flag_set(inode, FI_COMPRESS_RELEASED)) {
+	if (!f2fs_compressed_file(inode) ||
+		is_inode_flag_set(inode, FI_COMPRESS_RELEASED)) {
 		ret = -EINVAL;
 		goto out;
 	}
@@ -4126,7 +4243,7 @@ static int f2fs_ioc_decompress_file(struct file *filp)
 		if (ret < 0)
 			break;
 
-		if (get_dirty_pages(inode) >= blk_per_seg) {
+		if (get_dirty_pages(inode) >= BLKS_PER_SEG(sbi)) {
 			ret = filemap_fdatawrite(inode->i_mapping);
 			if (ret < 0)
 				break;
@@ -4149,6 +4266,7 @@ static int f2fs_ioc_decompress_file(struct file *filp)
 	if (ret)
 		f2fs_warn(sbi, "%s: The file might be partially decompressed (errno=%d). Please delete the file.",
 			  __func__, ret);
+	f2fs_update_time(sbi, REQ_TIME);
 out:
 	inode_unlock(inode);
 	file_end_write(filp);
@@ -4161,7 +4279,6 @@ static int f2fs_ioc_compress_file(struct file *filp)
 	struct inode *inode = file_inode(filp);
 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 	pgoff_t page_idx = 0, last_idx;
-	unsigned int blk_per_seg = sbi->blocks_per_seg;
 	int cluster_size = F2FS_I(inode)->i_cluster_size;
 	int count, ret;
 
@@ -4172,9 +4289,6 @@ static int f2fs_ioc_compress_file(struct file *filp)
 	if (!(filp->f_mode & FMODE_WRITE))
 		return -EBADF;
 
-	if (!f2fs_compressed_file(inode))
-		return -EINVAL;
-
 	f2fs_balance_fs(sbi, true);
 
 	file_start_write(filp);
@@ -4185,7 +4299,8 @@ static int f2fs_ioc_compress_file(struct file *filp)
 		goto out;
 	}
 
-	if (is_inode_flag_set(inode, FI_COMPRESS_RELEASED)) {
+	if (!f2fs_compressed_file(inode) ||
+		is_inode_flag_set(inode, FI_COMPRESS_RELEASED)) {
 		ret = -EINVAL;
 		goto out;
 	}
@@ -4204,7 +4319,7 @@ static int f2fs_ioc_compress_file(struct file *filp)
 		if (ret < 0)
 			break;
 
-		if (get_dirty_pages(inode) >= blk_per_seg) {
+		if (get_dirty_pages(inode) >= BLKS_PER_SEG(sbi)) {
 			ret = filemap_fdatawrite(inode->i_mapping);
 			if (ret < 0)
 				break;
@@ -4229,6 +4344,7 @@ static int f2fs_ioc_compress_file(struct file *filp)
 	if (ret)
 		f2fs_warn(sbi, "%s: The file might be partially compressed (errno=%d). Please delete the file.",
 			  __func__, ret);
+	f2fs_update_time(sbi, REQ_TIME);
 out:
 	inode_unlock(inode);
 	file_end_write(filp);
@@ -4587,7 +4703,8 @@ static int f2fs_preallocate_blocks(struct kiocb *iocb, struct iov_iter *iter,
 
 	map.m_may_create = true;
 	if (dio) {
-		map.m_seg_type = f2fs_rw_hint_to_seg_type(inode->i_write_hint);
+		map.m_seg_type = f2fs_rw_hint_to_seg_type(sbi,
+						inode->i_write_hint);
 		flag = F2FS_GET_BLOCK_PRE_DIO;
 	} else {
 		map.m_seg_type = NO_CHECK_TYPE;
@@ -4635,8 +4752,21 @@ static int f2fs_dio_write_end_io(struct kiocb *iocb, ssize_t size, int error,
 	return 0;
 }
 
+static void f2fs_dio_write_submit_io(const struct iomap_iter *iter,
+					struct bio *bio, loff_t file_offset)
+{
+	struct inode *inode = iter->inode;
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+	int seg_type = f2fs_rw_hint_to_seg_type(sbi, inode->i_write_hint);
+	enum temp_type temp = f2fs_get_segment_temp(seg_type);
+
+	bio->bi_write_hint = f2fs_io_type_to_rw_hint(sbi, DATA, temp);
+	submit_bio(bio);
+}
+
 static const struct iomap_dio_ops f2fs_iomap_dio_write_ops = {
-	.end_io = f2fs_dio_write_end_io,
+	.end_io		= f2fs_dio_write_end_io,
+	.submit_io	= f2fs_dio_write_submit_io,
 };
 
 static void f2fs_flush_buffered_write(struct address_space *mapping,
@@ -4773,6 +4903,8 @@ static ssize_t f2fs_file_write_iter(struct kiocb *iocb, struct iov_iter *from)
 	bool dio;
 	bool may_need_sync = true;
 	int preallocated;
+	const loff_t pos = iocb->ki_pos;
+	const ssize_t count = iov_iter_count(from);
 	ssize_t ret;
 
 	if (unlikely(f2fs_cp_error(F2FS_I_SB(inode)))) {
@@ -4794,6 +4926,12 @@ static ssize_t f2fs_file_write_iter(struct kiocb *iocb, struct iov_iter *from)
 		inode_lock(inode);
 	}
 
+	if (f2fs_is_pinned_file(inode) &&
+	    !f2fs_overwrite_io(inode, pos, count)) {
+		ret = -EIO;
+		goto out_unlock;
+	}
+
 	ret = f2fs_write_checks(iocb, from);
 	if (ret <= 0)
 		goto out_unlock;
@@ -5020,4 +5158,5 @@ const struct file_operations f2fs_file_operations = {
 	.splice_read	= f2fs_file_splice_read,
 	.splice_write	= iter_file_splice_write,
 	.fadvise	= f2fs_file_fadvise,
+	.fop_flags	= FOP_BUFFER_RASYNC,
 };
diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c
index a079eebfb0..448c75e80b 100644
--- a/fs/f2fs/gc.c
+++ b/fs/f2fs/gc.c
@@ -259,7 +259,7 @@ static void select_policy(struct f2fs_sb_info *sbi, int gc_type,
 		p->ofs_unit = 1;
 	} else {
 		p->gc_mode = select_gc_type(sbi, gc_type);
-		p->ofs_unit = sbi->segs_per_sec;
+		p->ofs_unit = SEGS_PER_SEC(sbi);
 		if (__is_large_section(sbi)) {
 			p->dirty_bitmap = dirty_i->dirty_secmap;
 			p->max_search = count_bits(p->dirty_bitmap,
@@ -280,11 +280,11 @@ static void select_policy(struct f2fs_sb_info *sbi, int gc_type,
 			p->max_search > sbi->max_victim_search)
 		p->max_search = sbi->max_victim_search;
 
-	/* let's select beginning hot/small space first in no_heap mode*/
+	/* let's select beginning hot/small space first. */
 	if (f2fs_need_rand_seg(sbi))
-		p->offset = get_random_u32_below(MAIN_SECS(sbi) * sbi->segs_per_sec);
-	else if (test_opt(sbi, NOHEAP) &&
-		(type == CURSEG_HOT_DATA || IS_NODESEG(type)))
+		p->offset = get_random_u32_below(MAIN_SECS(sbi) *
+						SEGS_PER_SEC(sbi));
+	else if (type == CURSEG_HOT_DATA || IS_NODESEG(type))
 		p->offset = 0;
 	else
 		p->offset = SIT_I(sbi)->last_victim[p->gc_mode];
@@ -295,13 +295,13 @@ static unsigned int get_max_cost(struct f2fs_sb_info *sbi,
 {
 	/* SSR allocates in a segment unit */
 	if (p->alloc_mode == SSR)
-		return sbi->blocks_per_seg;
+		return BLKS_PER_SEG(sbi);
 	else if (p->alloc_mode == AT_SSR)
 		return UINT_MAX;
 
 	/* LFS */
 	if (p->gc_mode == GC_GREEDY)
-		return 2 * sbi->blocks_per_seg * p->ofs_unit;
+		return SEGS_TO_BLKS(sbi, 2 * p->ofs_unit);
 	else if (p->gc_mode == GC_CB)
 		return UINT_MAX;
 	else if (p->gc_mode == GC_AT)
@@ -348,7 +348,7 @@ static unsigned int get_cb_cost(struct f2fs_sb_info *sbi, unsigned int segno)
 	mtime = div_u64(mtime, usable_segs_per_sec);
 	vblocks = div_u64(vblocks, usable_segs_per_sec);
 
-	u = (vblocks * 100) >> sbi->log_blocks_per_seg;
+	u = BLKS_TO_SEGS(sbi, vblocks * 100);
 
 	/* Handle if the system time has changed by the user */
 	if (mtime < sit_i->min_mtime)
@@ -496,9 +496,9 @@ static void add_victim_entry(struct f2fs_sb_info *sbi,
 			return;
 	}
 
-	for (i = 0; i < sbi->segs_per_sec; i++)
+	for (i = 0; i < SEGS_PER_SEC(sbi); i++)
 		mtime += get_seg_entry(sbi, start + i)->mtime;
-	mtime = div_u64(mtime, sbi->segs_per_sec);
+	mtime = div_u64(mtime, SEGS_PER_SEC(sbi));
 
 	/* Handle if the system time has changed by the user */
 	if (mtime < sit_i->min_mtime)
@@ -599,7 +599,6 @@ static void atssr_lookup_victim(struct f2fs_sb_info *sbi,
 	unsigned long long age;
 	unsigned long long max_mtime = sit_i->dirty_max_mtime;
 	unsigned long long min_mtime = sit_i->dirty_min_mtime;
-	unsigned int seg_blocks = sbi->blocks_per_seg;
 	unsigned int vblocks;
 	unsigned int dirty_threshold = max(am->max_candidate_count,
 					am->candidate_ratio *
@@ -629,7 +628,7 @@ next_node:
 	f2fs_bug_on(sbi, !vblocks);
 
 	/* rare case */
-	if (vblocks == seg_blocks)
+	if (vblocks == BLKS_PER_SEG(sbi))
 		goto skip_node;
 
 	iter++;
@@ -755,7 +754,7 @@ int f2fs_get_victim(struct f2fs_sb_info *sbi, unsigned int *result,
 	int ret = 0;
 
 	mutex_lock(&dirty_i->seglist_lock);
-	last_segment = MAIN_SECS(sbi) * sbi->segs_per_sec;
+	last_segment = MAIN_SECS(sbi) * SEGS_PER_SEC(sbi);
 
 	p.alloc_mode = alloc_mode;
 	p.age = age;
@@ -896,7 +895,7 @@ next:
 			else
 				sm->last_victim[p.gc_mode] = segno + p.ofs_unit;
 			sm->last_victim[p.gc_mode] %=
-				(MAIN_SECS(sbi) * sbi->segs_per_sec);
+				(MAIN_SECS(sbi) * SEGS_PER_SEC(sbi));
 			break;
 		}
 	}
@@ -1172,7 +1171,8 @@ static bool is_alive(struct f2fs_sb_info *sbi, struct f2fs_summary *sum,
 static int ra_data_block(struct inode *inode, pgoff_t index)
 {
 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
-	struct address_space *mapping = inode->i_mapping;
+	struct address_space *mapping = f2fs_is_cow_file(inode) ?
+				F2FS_I(inode)->atomic_inode->i_mapping : inode->i_mapping;
 	struct dnode_of_data dn;
 	struct page *page;
 	struct f2fs_io_info fio = {
@@ -1184,7 +1184,6 @@ static int ra_data_block(struct inode *inode, pgoff_t index)
 		.op_flags = 0,
 		.encrypted_page = NULL,
 		.in_list = 0,
-		.retry = 0,
 	};
 	int err;
 
@@ -1197,7 +1196,6 @@ static int ra_data_block(struct inode *inode, pgoff_t index)
 		if (unlikely(!f2fs_is_valid_blkaddr(sbi, dn.data_blkaddr,
 						DATA_GENERIC_ENHANCE_READ))) {
 			err = -EFSCORRUPTED;
-			f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR);
 			goto put_page;
 		}
 		goto got_it;
@@ -1216,7 +1214,6 @@ static int ra_data_block(struct inode *inode, pgoff_t index)
 	if (unlikely(!f2fs_is_valid_blkaddr(sbi, dn.data_blkaddr,
 						DATA_GENERIC_ENHANCE))) {
 		err = -EFSCORRUPTED;
-		f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR);
 		goto put_page;
 	}
 got_it:
@@ -1264,6 +1261,8 @@ put_page:
 static int move_data_block(struct inode *inode, block_t bidx,
 				int gc_type, unsigned int segno, int off)
 {
+	struct address_space *mapping = f2fs_is_cow_file(inode) ?
+				F2FS_I(inode)->atomic_inode->i_mapping : inode->i_mapping;
 	struct f2fs_io_info fio = {
 		.sbi = F2FS_I_SB(inode),
 		.ino = inode->i_ino,
@@ -1273,7 +1272,6 @@ static int move_data_block(struct inode *inode, block_t bidx,
 		.op_flags = 0,
 		.encrypted_page = NULL,
 		.in_list = 0,
-		.retry = 0,
 	};
 	struct dnode_of_data dn;
 	struct f2fs_summary sum;
@@ -1287,7 +1285,7 @@ static int move_data_block(struct inode *inode, block_t bidx,
 				CURSEG_ALL_DATA_ATGC : CURSEG_COLD_DATA;
 
 	/* do not read out */
-	page = f2fs_grab_cache_page(inode->i_mapping, bidx, false);
+	page = f2fs_grab_cache_page(mapping, bidx, false);
 	if (!page)
 		return -ENOMEM;
 
@@ -1364,8 +1362,13 @@ static int move_data_block(struct inode *inode, block_t bidx,
 	set_summary(&sum, dn.nid, dn.ofs_in_node, ni.version);
 
 	/* allocate block address */
-	f2fs_allocate_data_block(fio.sbi, NULL, fio.old_blkaddr, &newaddr,
+	err = f2fs_allocate_data_block(fio.sbi, NULL, fio.old_blkaddr, &newaddr,
 				&sum, type, NULL);
+	if (err) {
+		f2fs_put_page(mpage, 1);
+		/* filesystem should shutdown, no need to recovery block */
+		goto up_out;
+	}
 
 	fio.encrypted_page = f2fs_pagecache_get_page(META_MAPPING(fio.sbi),
 				newaddr, FGP_LOCK | FGP_CREAT, GFP_NOFS);
@@ -1393,18 +1396,12 @@ static int move_data_block(struct inode *inode, block_t bidx,
 	fio.op_flags = REQ_SYNC;
 	fio.new_blkaddr = newaddr;
 	f2fs_submit_page_write(&fio);
-	if (fio.retry) {
-		err = -EAGAIN;
-		if (PageWriteback(fio.encrypted_page))
-			end_page_writeback(fio.encrypted_page);
-		goto put_page_out;
-	}
 
 	f2fs_update_iostat(fio.sbi, NULL, FS_GC_DATA_IO, F2FS_BLKSIZE);
 
 	f2fs_update_data_blkaddr(&dn, newaddr);
 	set_inode_flag(inode, FI_APPEND_WRITE);
-put_page_out:
+
 	f2fs_put_page(fio.encrypted_page, 1);
 recover_block:
 	if (err)
@@ -1440,7 +1437,7 @@ static int move_data_page(struct inode *inode, block_t bidx, int gc_type,
 		goto out;
 
 	if (gc_type == BG_GC) {
-		if (PageWriteback(page)) {
+		if (folio_test_writeback(page_folio(page))) {
 			err = -EAGAIN;
 			goto out;
 		}
@@ -1560,10 +1557,25 @@ next_step:
 			int err;
 
 			inode = f2fs_iget(sb, dni.ino);
-			if (IS_ERR(inode) || is_bad_inode(inode) ||
-					special_file(inode->i_mode))
+			if (IS_ERR(inode))
 				continue;
 
+			if (is_bad_inode(inode) ||
+					special_file(inode->i_mode)) {
+				iput(inode);
+				continue;
+			}
+
+			if (f2fs_has_inline_data(inode)) {
+				iput(inode);
+				set_sbi_flag(sbi, SBI_NEED_FSCK);
+				f2fs_err_ratelimited(sbi,
+					"inode %lx has both inline_data flag and "
+					"data block, nid=%u, ofs_in_node=%u",
+					inode->i_ino, dni.nid, ofs_in_node);
+				continue;
+			}
+
 			err = f2fs_gc_pinned_control(inode, gc_type, segno);
 			if (err == -EAGAIN) {
 				iput(inode);
@@ -1580,7 +1592,7 @@ next_step:
 			start_bidx = f2fs_start_bidx_of_node(nofs, inode) +
 								ofs_in_node;
 
-			if (f2fs_post_read_required(inode)) {
+			if (f2fs_meta_inode_gc_required(inode)) {
 				int err = ra_data_block(inode, start_bidx);
 
 				f2fs_up_write(&F2FS_I(inode)->i_gc_rwsem[WRITE]);
@@ -1631,7 +1643,7 @@ next_step:
 
 			start_bidx = f2fs_start_bidx_of_node(nofs, inode)
 								+ ofs_in_node;
-			if (f2fs_post_read_required(inode))
+			if (f2fs_meta_inode_gc_required(inode))
 				err = move_data_block(inode, start_bidx,
 							gc_type, segno, off);
 			else
@@ -1639,7 +1651,7 @@ next_step:
 								segno, off);
 
 			if (!err && (gc_type == FG_GC ||
-					f2fs_post_read_required(inode)))
+					f2fs_meta_inode_gc_required(inode)))
 				submitted++;
 
 			if (locked) {
@@ -1678,7 +1690,7 @@ static int do_garbage_collect(struct f2fs_sb_info *sbi,
 	struct f2fs_summary_block *sum;
 	struct blk_plug plug;
 	unsigned int segno = start_segno;
-	unsigned int end_segno = start_segno + sbi->segs_per_sec;
+	unsigned int end_segno = start_segno + SEGS_PER_SEC(sbi);
 	int seg_freed = 0, migrated = 0;
 	unsigned char type = IS_DATASEG(get_seg_entry(sbi, segno)->type) ?
 						SUM_TYPE_DATA : SUM_TYPE_NODE;
@@ -1686,7 +1698,7 @@ static int do_garbage_collect(struct f2fs_sb_info *sbi,
 	int submitted = 0;
 
 	if (__is_large_section(sbi))
-		end_segno = rounddown(end_segno, sbi->segs_per_sec);
+		end_segno = rounddown(end_segno, SEGS_PER_SEC(sbi));
 
 	/*
 	 * zone-capacity can be less than zone-size in zoned devices,
@@ -1694,7 +1706,7 @@ static int do_garbage_collect(struct f2fs_sb_info *sbi,
 	 * calculate the end segno in the zone which can be garbage collected
 	 */
 	if (f2fs_sb_has_blkzoned(sbi))
-		end_segno -= sbi->segs_per_sec -
+		end_segno -= SEGS_PER_SEC(sbi) -
 					f2fs_usable_segs_in_sec(sbi, segno);
 
 	sanity_check_seg_type(sbi, get_seg_entry(sbi, segno)->type);
@@ -1983,10 +1995,43 @@ void f2fs_build_gc_manager(struct f2fs_sb_info *sbi)
 	init_atgc_management(sbi);
 }
 
+int f2fs_gc_range(struct f2fs_sb_info *sbi,
+		unsigned int start_seg, unsigned int end_seg,
+		bool dry_run, unsigned int dry_run_sections)
+{
+	unsigned int segno;
+	unsigned int gc_secs = dry_run_sections;
+
+	if (unlikely(f2fs_cp_error(sbi)))
+		return -EIO;
+
+	for (segno = start_seg; segno <= end_seg; segno += SEGS_PER_SEC(sbi)) {
+		struct gc_inode_list gc_list = {
+			.ilist = LIST_HEAD_INIT(gc_list.ilist),
+			.iroot = RADIX_TREE_INIT(gc_list.iroot, GFP_NOFS),
+		};
+
+		do_garbage_collect(sbi, segno, &gc_list, FG_GC,
+						dry_run_sections == 0);
+		put_gc_inode(&gc_list);
+
+		if (!dry_run && get_valid_blocks(sbi, segno, true))
+			return -EAGAIN;
+		if (dry_run && dry_run_sections &&
+		    !get_valid_blocks(sbi, segno, true) && --gc_secs == 0)
+			break;
+
+		if (fatal_signal_pending(current))
+			return -ERESTARTSYS;
+	}
+
+	return 0;
+}
+
 static int free_segment_range(struct f2fs_sb_info *sbi,
-				unsigned int secs, bool gc_only)
+				unsigned int secs, bool dry_run)
 {
-	unsigned int segno, next_inuse, start, end;
+	unsigned int next_inuse, start, end;
 	struct cp_control cpc = { CP_RESIZE, 0, 0, 0 };
 	int gc_mode, gc_type;
 	int err = 0;
@@ -1994,7 +2039,7 @@ static int free_segment_range(struct f2fs_sb_info *sbi,
 
 	/* Force block allocation for GC */
 	MAIN_SECS(sbi) -= secs;
-	start = MAIN_SECS(sbi) * sbi->segs_per_sec;
+	start = MAIN_SECS(sbi) * SEGS_PER_SEC(sbi);
 	end = MAIN_SEGS(sbi) - 1;
 
 	mutex_lock(&DIRTY_I(sbi)->seglist_lock);
@@ -2008,29 +2053,15 @@ static int free_segment_range(struct f2fs_sb_info *sbi,
 	mutex_unlock(&DIRTY_I(sbi)->seglist_lock);
 
 	/* Move out cursegs from the target range */
-	for (type = CURSEG_HOT_DATA; type < NR_CURSEG_PERSIST_TYPE; type++)
-		f2fs_allocate_segment_for_resize(sbi, type, start, end);
-
-	/* do GC to move out valid blocks in the range */
-	for (segno = start; segno <= end; segno += sbi->segs_per_sec) {
-		struct gc_inode_list gc_list = {
-			.ilist = LIST_HEAD_INIT(gc_list.ilist),
-			.iroot = RADIX_TREE_INIT(gc_list.iroot, GFP_NOFS),
-		};
-
-		do_garbage_collect(sbi, segno, &gc_list, FG_GC, true);
-		put_gc_inode(&gc_list);
-
-		if (!gc_only && get_valid_blocks(sbi, segno, true)) {
-			err = -EAGAIN;
-			goto out;
-		}
-		if (fatal_signal_pending(current)) {
-			err = -ERESTARTSYS;
+	for (type = CURSEG_HOT_DATA; type < NR_CURSEG_PERSIST_TYPE; type++) {
+		err = f2fs_allocate_segment_for_resize(sbi, type, start, end);
+		if (err)
 			goto out;
-		}
 	}
-	if (gc_only)
+
+	/* do GC to move out valid blocks in the range */
+	err = f2fs_gc_range(sbi, start, end, dry_run, 0);
+	if (err || dry_run)
 		goto out;
 
 	stat_inc_cp_call_count(sbi, TOTAL_CALL);
@@ -2056,7 +2087,7 @@ static void update_sb_metadata(struct f2fs_sb_info *sbi, int secs)
 	int segment_count;
 	int segment_count_main;
 	long long block_count;
-	int segs = secs * sbi->segs_per_sec;
+	int segs = secs * SEGS_PER_SEC(sbi);
 
 	f2fs_down_write(&sbi->sb_lock);
 
@@ -2069,7 +2100,7 @@ static void update_sb_metadata(struct f2fs_sb_info *sbi, int secs)
 	raw_sb->segment_count = cpu_to_le32(segment_count + segs);
 	raw_sb->segment_count_main = cpu_to_le32(segment_count_main + segs);
 	raw_sb->block_count = cpu_to_le64(block_count +
-					(long long)segs * sbi->blocks_per_seg);
+			(long long)SEGS_TO_BLKS(sbi, segs));
 	if (f2fs_is_multi_device(sbi)) {
 		int last_dev = sbi->s_ndevs - 1;
 		int dev_segs =
@@ -2084,8 +2115,8 @@ static void update_sb_metadata(struct f2fs_sb_info *sbi, int secs)
 
 static void update_fs_metadata(struct f2fs_sb_info *sbi, int secs)
 {
-	int segs = secs * sbi->segs_per_sec;
-	long long blks = (long long)segs * sbi->blocks_per_seg;
+	int segs = secs * SEGS_PER_SEC(sbi);
+	long long blks = SEGS_TO_BLKS(sbi, segs);
 	long long user_block_count =
 				le64_to_cpu(F2FS_CKPT(sbi)->user_block_count);
 
@@ -2127,7 +2158,7 @@ int f2fs_resize_fs(struct file *filp, __u64 block_count)
 		int last_dev = sbi->s_ndevs - 1;
 		__u64 last_segs = FDEV(last_dev).total_segments;
 
-		if (block_count + last_segs * sbi->blocks_per_seg <=
+		if (block_count + SEGS_TO_BLKS(sbi, last_segs) <=
 								old_block_count)
 			return -EINVAL;
 	}
diff --git a/fs/f2fs/gc.h b/fs/f2fs/gc.h
index 28a0094280..a8ea3301b8 100644
--- a/fs/f2fs/gc.h
+++ b/fs/f2fs/gc.h
@@ -26,6 +26,7 @@
 #define LIMIT_FREE_BLOCK	40 /* percentage over invalid + free space */
 
 #define DEF_GC_FAILED_PINNED_FILES	2048
+#define MAX_GC_FAILED_PINNED_FILES	USHRT_MAX
 
 /* Search max. number of dirty segments to select a victim segment */
 #define DEF_MAX_VICTIM_SEARCH 4096 /* covers 8GB */
@@ -96,7 +97,7 @@ static inline block_t free_segs_blk_count(struct f2fs_sb_info *sbi)
 	if (f2fs_sb_has_blkzoned(sbi))
 		return free_segs_blk_count_zoned(sbi);
 
-	return free_segments(sbi) << sbi->log_blocks_per_seg;
+	return SEGS_TO_BLKS(sbi, free_segments(sbi));
 }
 
 static inline block_t free_user_blocks(struct f2fs_sb_info *sbi)
@@ -104,7 +105,7 @@ static inline block_t free_user_blocks(struct f2fs_sb_info *sbi)
 	block_t free_blks, ovp_blks;
 
 	free_blks = free_segs_blk_count(sbi);
-	ovp_blks = overprovision_segments(sbi) << sbi->log_blocks_per_seg;
+	ovp_blks = SEGS_TO_BLKS(sbi, overprovision_segments(sbi));
 
 	if (free_blks < ovp_blks)
 		return 0;
diff --git a/fs/f2fs/inline.c b/fs/f2fs/inline.c
index ac00423f11..215daa71dc 100644
--- a/fs/f2fs/inline.c
+++ b/fs/f2fs/inline.c
@@ -16,7 +16,7 @@
 
 static bool support_inline_data(struct inode *inode)
 {
-	if (f2fs_is_atomic_file(inode))
+	if (f2fs_used_in_atomic_write(inode))
 		return false;
 	if (!S_ISREG(inode->i_mode) && !S_ISLNK(inode->i_mode))
 		return false;
@@ -61,22 +61,22 @@ bool f2fs_may_inline_dentry(struct inode *inode)
 	return true;
 }
 
-void f2fs_do_read_inline_data(struct page *page, struct page *ipage)
+void f2fs_do_read_inline_data(struct folio *folio, struct page *ipage)
 {
-	struct inode *inode = page->mapping->host;
+	struct inode *inode = folio_file_mapping(folio)->host;
 
-	if (PageUptodate(page))
+	if (folio_test_uptodate(folio))
 		return;
 
-	f2fs_bug_on(F2FS_P_SB(page), page->index);
+	f2fs_bug_on(F2FS_I_SB(inode), folio_index(folio));
 
-	zero_user_segment(page, MAX_INLINE_DATA(inode), PAGE_SIZE);
+	folio_zero_segment(folio, MAX_INLINE_DATA(inode), folio_size(folio));
 
 	/* Copy the whole inline data block */
-	memcpy_to_page(page, 0, inline_data_addr(inode, ipage),
+	memcpy_to_folio(folio, 0, inline_data_addr(inode, ipage),
 		       MAX_INLINE_DATA(inode));
-	if (!PageUptodate(page))
-		SetPageUptodate(page);
+	if (!folio_test_uptodate(folio))
+		folio_mark_uptodate(folio);
 }
 
 void f2fs_truncate_inline_inode(struct inode *inode,
@@ -97,13 +97,13 @@ void f2fs_truncate_inline_inode(struct inode *inode,
 		clear_inode_flag(inode, FI_DATA_EXIST);
 }
 
-int f2fs_read_inline_data(struct inode *inode, struct page *page)
+int f2fs_read_inline_data(struct inode *inode, struct folio *folio)
 {
 	struct page *ipage;
 
 	ipage = f2fs_get_node_page(F2FS_I_SB(inode), inode->i_ino);
 	if (IS_ERR(ipage)) {
-		unlock_page(page);
+		folio_unlock(folio);
 		return PTR_ERR(ipage);
 	}
 
@@ -112,15 +112,15 @@ int f2fs_read_inline_data(struct inode *inode, struct page *page)
 		return -EAGAIN;
 	}
 
-	if (page->index)
-		zero_user_segment(page, 0, PAGE_SIZE);
+	if (folio_index(folio))
+		folio_zero_segment(folio, 0, folio_size(folio));
 	else
-		f2fs_do_read_inline_data(page, ipage);
+		f2fs_do_read_inline_data(folio, ipage);
 
-	if (!PageUptodate(page))
-		SetPageUptodate(page);
+	if (!folio_test_uptodate(folio))
+		folio_mark_uptodate(folio);
 	f2fs_put_page(ipage, 1);
-	unlock_page(page);
+	folio_unlock(folio);
 	return 0;
 }
 
@@ -164,9 +164,9 @@ int f2fs_convert_inline_page(struct dnode_of_data *dn, struct page *page)
 		return -EFSCORRUPTED;
 	}
 
-	f2fs_bug_on(F2FS_P_SB(page), PageWriteback(page));
+	f2fs_bug_on(F2FS_P_SB(page), folio_test_writeback(page_folio(page)));
 
-	f2fs_do_read_inline_data(page, dn->inode_page);
+	f2fs_do_read_inline_data(page_folio(page), dn->inode_page);
 	set_page_dirty(page);
 
 	/* clear dirty state */
@@ -203,8 +203,10 @@ int f2fs_convert_inline_inode(struct inode *inode)
 	struct page *ipage, *page;
 	int err = 0;
 
-	if (!f2fs_has_inline_data(inode) ||
-			f2fs_hw_is_readonly(sbi) || f2fs_readonly(sbi->sb))
+	if (f2fs_hw_is_readonly(sbi) || f2fs_readonly(sbi->sb))
+		return -EROFS;
+
+	if (!f2fs_has_inline_data(inode))
 		return 0;
 
 	err = f2fs_dquot_initialize(inode);
diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
index c26effdce9..ed629dabbf 100644
--- a/fs/f2fs/inode.c
+++ b/fs/f2fs/inode.c
@@ -29,6 +29,9 @@ void f2fs_mark_inode_dirty_sync(struct inode *inode, bool sync)
 	if (is_inode_flag_set(inode, FI_NEW_INODE))
 		return;
 
+	if (f2fs_readonly(F2FS_I_SB(inode)->sb))
+		return;
+
 	if (f2fs_inode_dirtied(inode, sync))
 		return;
 
@@ -161,7 +164,8 @@ bool f2fs_inode_chksum_verify(struct f2fs_sb_info *sbi, struct page *page)
 	if (!f2fs_enable_inode_chksum(sbi, page))
 #else
 	if (!f2fs_enable_inode_chksum(sbi, page) ||
-			PageDirty(page) || PageWriteback(page))
+			PageDirty(page) ||
+			folio_test_writeback(page_folio(page)))
 #endif
 		return true;
 
@@ -361,6 +365,12 @@ static bool sanity_check_inode(struct inode *inode, struct page *node_page)
 		return false;
 	}
 
+	if (fi->i_xattr_nid && f2fs_check_nid_range(sbi, fi->i_xattr_nid)) {
+		f2fs_warn(sbi, "%s: inode (ino=%lx) has corrupted i_xattr_nid: %u, run fsck to fix.",
+			  __func__, inode->i_ino, fi->i_xattr_nid);
+		return false;
+	}
+
 	return true;
 }
 
@@ -408,8 +418,7 @@ static int do_read_inode(struct inode *inode)
 	if (S_ISDIR(inode->i_mode))
 		fi->i_current_depth = le32_to_cpu(ri->i_current_depth);
 	else if (S_ISREG(inode->i_mode))
-		fi->i_gc_failures[GC_FAILURE_PIN] =
-					le16_to_cpu(ri->i_gc_failures);
+		fi->i_gc_failures = le16_to_cpu(ri->i_gc_failures);
 	fi->i_xattr_nid = le32_to_cpu(ri->i_xattr_nid);
 	fi->i_flags = le32_to_cpu(ri->i_flags);
 	if (S_ISREG(inode->i_mode))
@@ -502,16 +511,16 @@ static int do_read_inode(struct inode *inode)
 
 	init_idisk_time(inode);
 
-	/* Need all the flag bits */
-	f2fs_init_read_extent_tree(inode, node_page);
-	f2fs_init_age_extent_tree(inode);
-
-	if (!sanity_check_extent_cache(inode)) {
+	if (!sanity_check_extent_cache(inode, node_page)) {
 		f2fs_put_page(node_page, 1);
 		f2fs_handle_error(sbi, ERROR_CORRUPTED_INODE);
 		return -EFSCORRUPTED;
 	}
 
+	/* Need all the flag bits */
+	f2fs_init_read_extent_tree(inode, node_page);
+	f2fs_init_age_extent_tree(inode);
+
 	f2fs_put_page(node_page, 1);
 
 	stat_inc_inline_xattr(inode);
@@ -604,14 +613,6 @@ make_now:
 	}
 	f2fs_set_inode_flags(inode);
 
-	if (file_should_truncate(inode) &&
-			!is_sbi_flag_set(sbi, SBI_POR_DOING)) {
-		ret = f2fs_truncate(inode);
-		if (ret)
-			goto bad_inode;
-		file_dont_truncate(inode);
-	}
-
 	unlock_new_inode(inode);
 	trace_f2fs_iget(inode);
 	return inode;
@@ -679,8 +680,7 @@ void f2fs_update_inode(struct inode *inode, struct page *node_page)
 		ri->i_current_depth =
 			cpu_to_le32(F2FS_I(inode)->i_current_depth);
 	else if (S_ISREG(inode->i_mode))
-		ri->i_gc_failures =
-			cpu_to_le16(F2FS_I(inode)->i_gc_failures[GC_FAILURE_PIN]);
+		ri->i_gc_failures = cpu_to_le16(F2FS_I(inode)->i_gc_failures);
 	ri->i_xattr_nid = cpu_to_le32(F2FS_I(inode)->i_xattr_nid);
 	ri->i_flags = cpu_to_le32(F2FS_I(inode)->i_flags);
 	ri->i_pino = cpu_to_le32(F2FS_I(inode)->i_pino);
@@ -804,11 +804,13 @@ void f2fs_evict_inode(struct inode *inode)
 	struct f2fs_inode_info *fi = F2FS_I(inode);
 	nid_t xnid = fi->i_xattr_nid;
 	int err = 0;
+	bool freeze_protected = false;
 
 	f2fs_abort_atomic_write(inode, true);
 
-	if (fi->cow_inode) {
+	if (fi->cow_inode && f2fs_is_cow_file(fi->cow_inode)) {
 		clear_inode_flag(fi->cow_inode, FI_COW_FILE);
+		F2FS_I(fi->cow_inode)->atomic_inode = NULL;
 		iput(fi->cow_inode);
 		fi->cow_inode = NULL;
 	}
@@ -843,8 +845,10 @@ void f2fs_evict_inode(struct inode *inode)
 	f2fs_remove_ino_entry(sbi, inode->i_ino, UPDATE_INO);
 	f2fs_remove_ino_entry(sbi, inode->i_ino, FLUSH_INO);
 
-	if (!is_sbi_flag_set(sbi, SBI_IS_FREEZING))
+	if (!is_sbi_flag_set(sbi, SBI_IS_FREEZING)) {
 		sb_start_intwrite(inode->i_sb);
+		freeze_protected = true;
+	}
 	set_inode_flag(inode, FI_NO_ALLOC);
 	i_size_write(inode, 0);
 retry:
@@ -887,7 +891,7 @@ retry:
 		if (dquot_initialize_needed(inode))
 			set_sbi_flag(sbi, SBI_QUOTA_NEED_REPAIR);
 	}
-	if (!is_sbi_flag_set(sbi, SBI_IS_FREEZING))
+	if (freeze_protected)
 		sb_end_intwrite(inode->i_sb);
 no_delete:
 	dquot_drop(inode);
diff --git a/fs/f2fs/namei.c b/fs/f2fs/namei.c
index 550cc811e0..e54f8c08bd 100644
--- a/fs/f2fs/namei.c
+++ b/fs/f2fs/namei.c
@@ -531,7 +531,6 @@ static struct dentry *f2fs_lookup(struct inode *dir, struct dentry *dentry,
 	}
 
 	err = f2fs_prepare_lookup(dir, dentry, &fname);
-	generic_set_encrypted_ci_d_ops(dentry);
 	if (err == -ENOENT)
 		goto out_splice;
 	if (err)
@@ -1114,14 +1113,11 @@ static int f2fs_rename(struct mnt_idmap *idmap, struct inode *old_dir,
 		iput(whiteout);
 	}
 
-	if (old_is_dir) {
-		if (old_dir_entry)
-			f2fs_set_link(old_inode, old_dir_entry,
-						old_dir_page, new_dir);
-		else
-			f2fs_put_page(old_dir_page, 0);
+	if (old_dir_entry)
+		f2fs_set_link(old_inode, old_dir_entry, old_dir_page, new_dir);
+	if (old_is_dir)
 		f2fs_i_links_write(old_dir, false);
-	}
+
 	if (F2FS_OPTION(sbi).fsync_mode == FSYNC_MODE_STRICT) {
 		f2fs_add_ino_entry(sbi, new_dir->i_ino, TRANS_DIR_INO);
 		if (S_ISDIR(old_inode->i_mode))
diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
index 2ea9c99e7d..b72ef96f7e 100644
--- a/fs/f2fs/node.c
+++ b/fs/f2fs/node.c
@@ -1187,7 +1187,17 @@ skip_partial:
 		default:
 			BUG();
 		}
-		if (err < 0 && err != -ENOENT)
+		if (err == -ENOENT) {
+			set_sbi_flag(F2FS_P_SB(page), SBI_NEED_FSCK);
+			f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR);
+			f2fs_err_ratelimited(sbi,
+				"truncate node fail, ino:%lu, nid:%u, "
+				"offset[0]:%d, offset[1]:%d, nofs:%d",
+				inode->i_ino, dn.nid, offset[0],
+				offset[1], nofs);
+			err = 0;
+		}
+		if (err < 0)
 			goto fail;
 		if (offset[1] == 0 &&
 				ri->i_nid[offset[0] - NODE_DIR1_BLOCK]) {
@@ -1319,6 +1329,7 @@ struct page *f2fs_new_node_page(struct dnode_of_data *dn, unsigned int ofs)
 	}
 	if (unlikely(new_ni.blk_addr != NULL_ADDR)) {
 		err = -EFSCORRUPTED;
+		dec_valid_node_count(sbi, dn->inode, !ofs);
 		set_sbi_flag(sbi, SBI_NEED_FSCK);
 		f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR);
 		goto fail;
@@ -1345,7 +1356,6 @@ struct page *f2fs_new_node_page(struct dnode_of_data *dn, unsigned int ofs)
 	if (ofs == 0)
 		inc_valid_inode_count(sbi);
 	return page;
-
 fail:
 	clear_node_page_dirty(page);
 	f2fs_put_page(page, 1);
@@ -1614,7 +1624,7 @@ static int __write_node_page(struct page *page, bool atomic, bool *submitted,
 	};
 	unsigned int seq;
 
-	trace_f2fs_writepage(page, NODE);
+	trace_f2fs_writepage(page_folio(page), NODE);
 
 	if (unlikely(f2fs_cp_error(sbi))) {
 		/* keep node pages in remount-ro mode */
@@ -1733,7 +1743,7 @@ int f2fs_move_node_page(struct page *node_page, int gc_type)
 		goto release_page;
 	} else {
 		/* set page dirty and write it */
-		if (!PageWriteback(node_page))
+		if (!folio_test_writeback(page_folio(node_page)))
 			set_page_dirty(node_page);
 	}
 out_page:
@@ -1927,7 +1937,7 @@ void f2fs_flush_inline_data(struct f2fs_sb_info *sbi)
 		for (i = 0; i < nr_folios; i++) {
 			struct page *page = &fbatch.folios[i]->page;
 
-			if (!IS_DNODE(page))
+			if (!IS_INODE(page))
 				continue;
 
 			lock_page(page);
@@ -2161,7 +2171,7 @@ skip_write:
 static bool f2fs_dirty_node_folio(struct address_space *mapping,
 		struct folio *folio)
 {
-	trace_f2fs_set_page_dirty(&folio->page, NODE);
+	trace_f2fs_set_page_dirty(folio, NODE);
 
 	if (!folio_test_uptodate(folio))
 		folio_mark_uptodate(folio);
@@ -2849,7 +2859,7 @@ int f2fs_restore_node_summary(struct f2fs_sb_info *sbi,
 	int i, idx, last_offset, nrpages;
 
 	/* scan the node segment */
-	last_offset = sbi->blocks_per_seg;
+	last_offset = BLKS_PER_SEG(sbi);
 	addr = START_BLOCK(sbi, segno);
 	sum_entry = &sum->entries[0];
 
@@ -3166,7 +3176,7 @@ static int __get_nat_bitmaps(struct f2fs_sb_info *sbi)
 	if (!is_set_ckpt_flags(sbi, CP_NAT_BITS_FLAG))
 		return 0;
 
-	nat_bits_addr = __start_cp_addr(sbi) + sbi->blocks_per_seg -
+	nat_bits_addr = __start_cp_addr(sbi) + BLKS_PER_SEG(sbi) -
 						nm_i->nat_bits_blocks;
 	for (i = 0; i < nm_i->nat_bits_blocks; i++) {
 		struct page *page;
diff --git a/fs/f2fs/node.h b/fs/f2fs/node.h
index 5bd16a95ee..6aea13024a 100644
--- a/fs/f2fs/node.h
+++ b/fs/f2fs/node.h
@@ -208,10 +208,10 @@ static inline pgoff_t current_nat_addr(struct f2fs_sb_info *sbi, nid_t start)
 
 	block_addr = (pgoff_t)(nm_i->nat_blkaddr +
 		(block_off << 1) -
-		(block_off & (sbi->blocks_per_seg - 1)));
+		(block_off & (BLKS_PER_SEG(sbi) - 1)));
 
 	if (f2fs_test_bit(block_off, nm_i->nat_bitmap))
-		block_addr += sbi->blocks_per_seg;
+		block_addr += BLKS_PER_SEG(sbi);
 
 	return block_addr;
 }
diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c
index aad1d1a9b3..496aee53c3 100644
--- a/fs/f2fs/recovery.c
+++ b/fs/f2fs/recovery.c
@@ -330,8 +330,7 @@ static int recover_inode(struct inode *inode, struct page *page)
 	F2FS_I(inode)->i_advise = raw->i_advise;
 	F2FS_I(inode)->i_flags = le32_to_cpu(raw->i_flags);
 	f2fs_set_inode_flags(inode);
-	F2FS_I(inode)->i_gc_failures[GC_FAILURE_PIN] =
-				le16_to_cpu(raw->i_gc_failures);
+	F2FS_I(inode)->i_gc_failures = le16_to_cpu(raw->i_gc_failures);
 
 	recover_inline_flags(inode, raw);
 
@@ -354,7 +353,7 @@ static unsigned int adjust_por_ra_blocks(struct f2fs_sb_info *sbi,
 	if (blkaddr + 1 == next_blkaddr)
 		ra_blocks = min_t(unsigned int, RECOVERY_MAX_RA_BLOCKS,
 							ra_blocks * 2);
-	else if (next_blkaddr % sbi->blocks_per_seg)
+	else if (next_blkaddr % BLKS_PER_SEG(sbi))
 		ra_blocks = max_t(unsigned int, RECOVERY_MIN_RA_BLOCKS,
 							ra_blocks / 2);
 	return ra_blocks;
@@ -693,14 +692,12 @@ retry_dn:
 		if (__is_valid_data_blkaddr(src) &&
 			!f2fs_is_valid_blkaddr(sbi, src, META_POR)) {
 			err = -EFSCORRUPTED;
-			f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR);
 			goto err;
 		}
 
 		if (__is_valid_data_blkaddr(dest) &&
 			!f2fs_is_valid_blkaddr(sbi, dest, META_POR)) {
 			err = -EFSCORRUPTED;
-			f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR);
 			goto err;
 		}
 
@@ -755,8 +752,6 @@ retry_prev:
 				f2fs_err(sbi, "Inconsistent dest blkaddr:%u, ino:%lu, ofs:%u",
 					dest, inode->i_ino, dn.ofs_in_node);
 				err = -EFSCORRUPTED;
-				f2fs_handle_error(sbi,
-						ERROR_INVALID_BLKADDR);
 				goto err;
 			}
 
@@ -851,7 +846,7 @@ next:
 		f2fs_ra_meta_pages_cond(sbi, blkaddr, ra_blocks);
 	}
 	if (!err)
-		f2fs_allocate_new_segments(sbi);
+		err = f2fs_allocate_new_segments(sbi);
 	return err;
 }
 
@@ -863,7 +858,6 @@ int f2fs_recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only)
 	int ret = 0;
 	unsigned long s_flags = sbi->sb->s_flags;
 	bool need_writecp = false;
-	bool fix_curseg_write_pointer = false;
 
 	if (is_sbi_flag_set(sbi, SBI_IS_WRITABLE))
 		f2fs_info(sbi, "recover fsync data on readonly fs");
@@ -894,8 +888,6 @@ int f2fs_recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only)
 	else
 		f2fs_bug_on(sbi, sbi->sb->s_flags & SB_ACTIVE);
 skip:
-	fix_curseg_write_pointer = !check_only || list_empty(&inode_list);
-
 	destroy_fsync_dnodes(&inode_list, err);
 	destroy_fsync_dnodes(&tmp_inode_list, err);
 
@@ -913,11 +905,13 @@ skip:
 	 * and the f2fs is not read only, check and fix zoned block devices'
 	 * write pointer consistency.
 	 */
-	if (!err && fix_curseg_write_pointer && !f2fs_readonly(sbi->sb) &&
-			f2fs_sb_has_blkzoned(sbi)) {
-		err = f2fs_fix_curseg_write_pointer(sbi);
-		if (!err)
-			err = f2fs_check_write_pointer(sbi);
+	if (f2fs_sb_has_blkzoned(sbi) && !f2fs_readonly(sbi->sb)) {
+		int err2 = f2fs_fix_curseg_write_pointer(sbi);
+
+		if (!err2)
+			err2 = f2fs_check_write_pointer(sbi);
+		if (err2)
+			err = err2;
 		ret = err;
 	}
 
diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index 6d57740934..6018257852 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -336,8 +336,6 @@ static int __f2fs_commit_atomic_write(struct inode *inode)
 					DATA_GENERIC_ENHANCE)) {
 				f2fs_put_dnode(&dn);
 				ret = -EFSCORRUPTED;
-				f2fs_handle_error(sbi,
-						ERROR_INVALID_BLKADDR);
 				goto out;
 			}
 
@@ -402,6 +400,9 @@ int f2fs_commit_atomic_write(struct inode *inode)
  */
 void f2fs_balance_fs(struct f2fs_sb_info *sbi, bool need)
 {
+	if (f2fs_cp_error(sbi))
+		return;
+
 	if (time_to_inject(sbi, FAULT_CHECKPOINT))
 		f2fs_stop_checkpoint(sbi, false, STOP_CP_REASON_FAULT_INJECT);
 
@@ -450,8 +451,8 @@ static inline bool excess_dirty_threshold(struct f2fs_sb_info *sbi)
 	unsigned int nodes = get_pages(sbi, F2FS_DIRTY_NODES);
 	unsigned int meta = get_pages(sbi, F2FS_DIRTY_META);
 	unsigned int imeta = get_pages(sbi, F2FS_DIRTY_IMETA);
-	unsigned int threshold = sbi->blocks_per_seg * factor *
-					DEFAULT_DIRTY_THRESHOLD;
+	unsigned int threshold =
+		SEGS_TO_BLKS(sbi, (factor * DEFAULT_DIRTY_THRESHOLD));
 	unsigned int global_threshold = threshold * 3 / 2;
 
 	if (dents >= threshold || qdata >= threshold ||
@@ -770,8 +771,10 @@ static void __locate_dirty_segment(struct f2fs_sb_info *sbi, unsigned int segno,
 			block_t valid_blocks =
 				get_valid_blocks(sbi, segno, true);
 
-			f2fs_bug_on(sbi, unlikely(!valid_blocks ||
-					valid_blocks == CAP_BLKS_PER_SEC(sbi)));
+			f2fs_bug_on(sbi,
+				(!is_sbi_flag_set(sbi, SBI_CP_DISABLED) &&
+				!valid_blocks) ||
+				valid_blocks == CAP_BLKS_PER_SEC(sbi));
 
 			if (!IS_CURSEC(sbi, secno))
 				set_bit(secno, dirty_i->dirty_secmap);
@@ -874,7 +877,7 @@ block_t f2fs_get_unusable_blocks(struct f2fs_sb_info *sbi)
 {
 	int ovp_hole_segs =
 		(overprovision_segments(sbi) - reserved_segments(sbi));
-	block_t ovp_holes = ovp_hole_segs << sbi->log_blocks_per_seg;
+	block_t ovp_holes = SEGS_TO_BLKS(sbi, ovp_hole_segs);
 	struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
 	block_t holes[2] = {0, 0};	/* DATA and NODE */
 	block_t unusable;
@@ -903,11 +906,16 @@ int f2fs_disable_cp_again(struct f2fs_sb_info *sbi, block_t unusable)
 {
 	int ovp_hole_segs =
 		(overprovision_segments(sbi) - reserved_segments(sbi));
+
+	if (F2FS_OPTION(sbi).unusable_cap_perc == 100)
+		return 0;
 	if (unusable > F2FS_OPTION(sbi).unusable_cap)
 		return -EAGAIN;
 	if (is_sbi_flag_set(sbi, SBI_CP_DISABLED_QUICK) &&
 		dirty_segments(sbi) > ovp_hole_segs)
 		return -EAGAIN;
+	if (has_not_enough_free_secs(sbi, 0, 0))
+		return -EAGAIN;
 	return 0;
 }
 
@@ -1103,9 +1111,8 @@ static void __remove_discard_cmd(struct f2fs_sb_info *sbi,
 		dc->error = 0;
 
 	if (dc->error)
-		printk_ratelimited(
-			"%sF2FS-fs (%s): Issue discard(%u, %u, %u) failed, ret: %d",
-			KERN_INFO, sbi->sb->s_id,
+		f2fs_info_ratelimited(sbi,
+			"Issue discard(%u, %u, %u) failed, ret: %d",
 			dc->di.lstart, dc->di.start, dc->di.len, dc->error);
 	__detach_discard_cmd(dcc, dc);
 }
@@ -1134,8 +1141,7 @@ static void __check_sit_bitmap(struct f2fs_sb_info *sbi,
 	struct seg_entry *sentry;
 	unsigned int segno;
 	block_t blk = start;
-	unsigned long offset, size, max_blocks = sbi->blocks_per_seg;
-	unsigned long *map;
+	unsigned long offset, size, *map;
 
 	while (blk < end) {
 		segno = GET_SEGNO(sbi, blk);
@@ -1145,7 +1151,7 @@ static void __check_sit_bitmap(struct f2fs_sb_info *sbi,
 		if (end < START_BLOCK(sbi, segno + 1))
 			size = GET_BLKOFF_FROM_SEG0(sbi, end);
 		else
-			size = max_blocks;
+			size = BLKS_PER_SEG(sbi);
 		map = (unsigned long *)(sentry->cur_valid_map);
 		offset = __find_rev_next_bit(map, size, offset);
 		f2fs_bug_on(sbi, offset != size);
@@ -1973,9 +1979,15 @@ static int __f2fs_issue_discard_zone(struct f2fs_sb_info *sbi,
 		}
 
 		if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING))) {
+			unsigned int nofs_flags;
+			int ret;
+
 			trace_f2fs_issue_reset_zone(bdev, blkstart);
-			return blkdev_zone_mgmt(bdev, REQ_OP_ZONE_RESET,
-						sector, nr_sects, GFP_NOFS);
+			nofs_flags = memalloc_nofs_save();
+			ret = blkdev_zone_mgmt(bdev, REQ_OP_ZONE_RESET,
+						sector, nr_sects);
+			memalloc_nofs_restore(nofs_flags);
+			return ret;
 		}
 
 		__queue_zone_reset_cmd(sbi, bdev, blkstart, lblkstart, blklen);
@@ -2044,7 +2056,6 @@ static bool add_discard_addrs(struct f2fs_sb_info *sbi, struct cp_control *cpc,
 							bool check_only)
 {
 	int entries = SIT_VBLOCK_MAP_SIZE / sizeof(unsigned long);
-	int max_blocks = sbi->blocks_per_seg;
 	struct seg_entry *se = get_seg_entry(sbi, cpc->trim_start);
 	unsigned long *cur_map = (unsigned long *)se->cur_valid_map;
 	unsigned long *ckpt_map = (unsigned long *)se->ckpt_valid_map;
@@ -2056,8 +2067,9 @@ static bool add_discard_addrs(struct f2fs_sb_info *sbi, struct cp_control *cpc,
 	struct list_head *head = &SM_I(sbi)->dcc_info->entry_list;
 	int i;
 
-	if (se->valid_blocks == max_blocks || !f2fs_hw_support_discard(sbi) ||
-			!f2fs_block_unit_discard(sbi))
+	if (se->valid_blocks == BLKS_PER_SEG(sbi) ||
+	    !f2fs_hw_support_discard(sbi) ||
+	    !f2fs_block_unit_discard(sbi))
 		return false;
 
 	if (!force) {
@@ -2074,13 +2086,14 @@ static bool add_discard_addrs(struct f2fs_sb_info *sbi, struct cp_control *cpc,
 
 	while (force || SM_I(sbi)->dcc_info->nr_discards <=
 				SM_I(sbi)->dcc_info->max_discards) {
-		start = __find_rev_next_bit(dmap, max_blocks, end + 1);
-		if (start >= max_blocks)
+		start = __find_rev_next_bit(dmap, BLKS_PER_SEG(sbi), end + 1);
+		if (start >= BLKS_PER_SEG(sbi))
 			break;
 
-		end = __find_rev_next_zero_bit(dmap, max_blocks, start + 1);
-		if (force && start && end != max_blocks
-					&& (end - start) < cpc->trim_minlen)
+		end = __find_rev_next_zero_bit(dmap,
+						BLKS_PER_SEG(sbi), start + 1);
+		if (force && start && end != BLKS_PER_SEG(sbi) &&
+		    (end - start) < cpc->trim_minlen)
 			continue;
 
 		if (check_only)
@@ -2162,8 +2175,8 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
 								start + 1);
 
 		if (section_alignment) {
-			start = rounddown(start, sbi->segs_per_sec);
-			end = roundup(end, sbi->segs_per_sec);
+			start = rounddown(start, SEGS_PER_SEC(sbi));
+			end = roundup(end, SEGS_PER_SEC(sbi));
 		}
 
 		for (i = start; i < end; i++) {
@@ -2182,7 +2195,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
 		if (!f2fs_sb_has_blkzoned(sbi) &&
 		    (!f2fs_lfs_mode(sbi) || !__is_large_section(sbi))) {
 			f2fs_issue_discard(sbi, START_BLOCK(sbi, start),
-				(end - start) << sbi->log_blocks_per_seg);
+				SEGS_TO_BLKS(sbi, end - start));
 			continue;
 		}
 next:
@@ -2191,9 +2204,9 @@ next:
 		if (!IS_CURSEC(sbi, secno) &&
 			!get_valid_blocks(sbi, start, true))
 			f2fs_issue_discard(sbi, START_BLOCK(sbi, start_segno),
-				sbi->segs_per_sec << sbi->log_blocks_per_seg);
+						BLKS_PER_SEC(sbi));
 
-		start = start_segno + sbi->segs_per_sec;
+		start = start_segno + SEGS_PER_SEC(sbi);
 		if (start < end)
 			goto next;
 		else
@@ -2212,7 +2225,7 @@ next:
 find_next:
 		if (is_valid) {
 			next_pos = find_next_zero_bit_le(entry->discard_map,
-					sbi->blocks_per_seg, cur_pos);
+						BLKS_PER_SEG(sbi), cur_pos);
 			len = next_pos - cur_pos;
 
 			if (f2fs_sb_has_blkzoned(sbi) ||
@@ -2224,13 +2237,13 @@ find_next:
 			total_len += len;
 		} else {
 			next_pos = find_next_bit_le(entry->discard_map,
-					sbi->blocks_per_seg, cur_pos);
+						BLKS_PER_SEG(sbi), cur_pos);
 		}
 skip:
 		cur_pos = next_pos;
 		is_valid = !is_valid;
 
-		if (cur_pos < sbi->blocks_per_seg)
+		if (cur_pos < BLKS_PER_SEG(sbi))
 			goto find_next;
 
 		release_discard_addr(entry);
@@ -2247,6 +2260,12 @@ int f2fs_start_discard_thread(struct f2fs_sb_info *sbi)
 	struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
 	int err = 0;
 
+	if (f2fs_sb_has_readonly(sbi)) {
+		f2fs_info(sbi,
+			"Skip to start discard thread for readonly image");
+		return 0;
+	}
+
 	if (!f2fs_realtime_discard_enable(sbi))
 		return 0;
 
@@ -2279,7 +2298,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
 	dcc->max_ordered_discard = DEFAULT_MAX_ORDERED_DISCARD_GRANULARITY;
 	dcc->discard_io_aware = DPOLICY_IO_AWARE_ENABLE;
 	if (F2FS_OPTION(sbi).discard_unit == DISCARD_UNIT_SEGMENT)
-		dcc->discard_granularity = sbi->blocks_per_seg;
+		dcc->discard_granularity = BLKS_PER_SEG(sbi);
 	else if (F2FS_OPTION(sbi).discard_unit == DISCARD_UNIT_SECTION)
 		dcc->discard_granularity = BLKS_PER_SEC(sbi);
 
@@ -2293,7 +2312,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
 	atomic_set(&dcc->queued_discard, 0);
 	atomic_set(&dcc->discard_cmd_cnt, 0);
 	dcc->nr_discards = 0;
-	dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
+	dcc->max_discards = SEGS_TO_BLKS(sbi, MAIN_SEGS(sbi));
 	dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
 	dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
 	dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
@@ -2401,6 +2420,8 @@ static void update_sit_entry(struct f2fs_sb_info *sbi, block_t blkaddr, int del)
 #endif
 
 	segno = GET_SEGNO(sbi, blkaddr);
+	if (segno == NULL_SEGNO)
+		return;
 
 	se = get_seg_entry(sbi, segno);
 	new_vblocks = se->valid_blocks + del;
@@ -2542,7 +2563,7 @@ static unsigned short f2fs_curseg_valid_blocks(struct f2fs_sb_info *sbi, int typ
 	struct curseg_info *curseg = CURSEG_I(sbi, type);
 
 	if (sbi->ckpt->alloc_type[type] == SSR)
-		return sbi->blocks_per_seg;
+		return BLKS_PER_SEG(sbi);
 	return curseg->next_blkoff;
 }
 
@@ -2625,12 +2646,12 @@ static void write_current_sum_page(struct f2fs_sb_info *sbi,
 }
 
 static int is_next_segment_free(struct f2fs_sb_info *sbi,
-				struct curseg_info *curseg, int type)
+				struct curseg_info *curseg)
 {
 	unsigned int segno = curseg->segno + 1;
 	struct free_segmap_info *free_i = FREE_I(sbi);
 
-	if (segno < MAIN_SEGS(sbi) && segno % sbi->segs_per_sec)
+	if (segno < MAIN_SEGS(sbi) && segno % SEGS_PER_SEC(sbi))
 		return !test_bit(segno, free_i->free_segmap);
 	return 0;
 }
@@ -2639,54 +2660,51 @@ static int is_next_segment_free(struct f2fs_sb_info *sbi,
  * Find a new segment from the free segments bitmap to right order
  * This function should be returned with success, otherwise BUG
  */
-static void get_new_segment(struct f2fs_sb_info *sbi,
-			unsigned int *newseg, bool new_sec, int dir)
+static int get_new_segment(struct f2fs_sb_info *sbi,
+			unsigned int *newseg, bool new_sec, bool pinning)
 {
 	struct free_segmap_info *free_i = FREE_I(sbi);
 	unsigned int segno, secno, zoneno;
 	unsigned int total_zones = MAIN_SECS(sbi) / sbi->secs_per_zone;
 	unsigned int hint = GET_SEC_FROM_SEG(sbi, *newseg);
 	unsigned int old_zoneno = GET_ZONE_FROM_SEG(sbi, *newseg);
-	unsigned int left_start = hint;
 	bool init = true;
-	int go_left = 0;
 	int i;
+	int ret = 0;
 
 	spin_lock(&free_i->segmap_lock);
 
-	if (!new_sec && ((*newseg + 1) % sbi->segs_per_sec)) {
+	if (time_to_inject(sbi, FAULT_NO_SEGMENT)) {
+		ret = -ENOSPC;
+		goto out_unlock;
+	}
+
+	if (!new_sec && ((*newseg + 1) % SEGS_PER_SEC(sbi))) {
 		segno = find_next_zero_bit(free_i->free_segmap,
 			GET_SEG_FROM_SEC(sbi, hint + 1), *newseg + 1);
 		if (segno < GET_SEG_FROM_SEC(sbi, hint + 1))
 			goto got_it;
 	}
+
+	/*
+	 * If we format f2fs on zoned storage, let's try to get pinned sections
+	 * from beginning of the storage, which should be a conventional one.
+	 */
+	if (f2fs_sb_has_blkzoned(sbi)) {
+		segno = pinning ? 0 : max(first_zoned_segno(sbi), *newseg);
+		hint = GET_SEC_FROM_SEG(sbi, segno);
+	}
+
 find_other_zone:
 	secno = find_next_zero_bit(free_i->free_secmap, MAIN_SECS(sbi), hint);
 	if (secno >= MAIN_SECS(sbi)) {
-		if (dir == ALLOC_RIGHT) {
-			secno = find_first_zero_bit(free_i->free_secmap,
+		secno = find_first_zero_bit(free_i->free_secmap,
 							MAIN_SECS(sbi));
-			f2fs_bug_on(sbi, secno >= MAIN_SECS(sbi));
-		} else {
-			go_left = 1;
-			left_start = hint - 1;
+		if (secno >= MAIN_SECS(sbi)) {
+			ret = -ENOSPC;
+			goto out_unlock;
 		}
 	}
-	if (go_left == 0)
-		goto skip_left;
-
-	while (test_bit(left_start, free_i->free_secmap)) {
-		if (left_start > 0) {
-			left_start--;
-			continue;
-		}
-		left_start = find_first_zero_bit(free_i->free_secmap,
-							MAIN_SECS(sbi));
-		f2fs_bug_on(sbi, left_start >= MAIN_SECS(sbi));
-		break;
-	}
-	secno = left_start;
-skip_left:
 	segno = GET_SEG_FROM_SEC(sbi, secno);
 	zoneno = GET_ZONE_FROM_SEC(sbi, secno);
 
@@ -2697,21 +2715,13 @@ skip_left:
 		goto got_it;
 	if (zoneno == old_zoneno)
 		goto got_it;
-	if (dir == ALLOC_LEFT) {
-		if (!go_left && zoneno + 1 >= total_zones)
-			goto got_it;
-		if (go_left && zoneno == 0)
-			goto got_it;
-	}
 	for (i = 0; i < NR_CURSEG_TYPE; i++)
 		if (CURSEG_I(sbi, i)->zone == zoneno)
 			break;
 
 	if (i < NR_CURSEG_TYPE) {
 		/* zone is in user, try another */
-		if (go_left)
-			hint = zoneno * sbi->secs_per_zone - 1;
-		else if (zoneno + 1 >= total_zones)
+		if (zoneno + 1 >= total_zones)
 			hint = 0;
 		else
 			hint = (zoneno + 1) * sbi->secs_per_zone;
@@ -2721,9 +2731,23 @@ skip_left:
 got_it:
 	/* set it as dirty segment in free segmap */
 	f2fs_bug_on(sbi, test_bit(segno, free_i->free_segmap));
+
+	/* no free section in conventional zone */
+	if (new_sec && pinning &&
+		!f2fs_valid_pinned_area(sbi, START_BLOCK(sbi, segno))) {
+		ret = -EAGAIN;
+		goto out_unlock;
+	}
 	__set_inuse(sbi, segno);
 	*newseg = segno;
+out_unlock:
 	spin_unlock(&free_i->segmap_lock);
+
+	if (ret == -ENOSPC) {
+		f2fs_stop_checkpoint(sbi, false, STOP_CP_REASON_NO_SEGMENT);
+		f2fs_bug_on(sbi, 1);
+	}
+	return ret;
 }
 
 static void reset_curseg(struct f2fs_sb_info *sbi, int type, int modified)
@@ -2732,6 +2756,10 @@ static void reset_curseg(struct f2fs_sb_info *sbi, int type, int modified)
 	struct summary_footer *sum_footer;
 	unsigned short seg_type = curseg->seg_type;
 
+	/* only happen when get_new_segment() fails */
+	if (curseg->next_segno == NULL_SEGNO)
+		return;
+
 	curseg->inited = true;
 	curseg->segno = curseg->next_segno;
 	curseg->zone = GET_ZONE_FROM_SEG(sbi, curseg->segno);
@@ -2757,9 +2785,8 @@ static unsigned int __get_next_segno(struct f2fs_sb_info *sbi, int type)
 
 	sanity_check_seg_type(sbi, seg_type);
 	if (f2fs_need_rand_seg(sbi))
-		return get_random_u32_below(MAIN_SECS(sbi) * sbi->segs_per_sec);
+		return get_random_u32_below(MAIN_SECS(sbi) * SEGS_PER_SEC(sbi));
 
-	/* if segs_per_sec is large than 1, we need to keep original policy. */
 	if (__is_large_section(sbi))
 		return curseg->segno;
 
@@ -2770,8 +2797,7 @@ static unsigned int __get_next_segno(struct f2fs_sb_info *sbi, int type)
 	if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED)))
 		return 0;
 
-	if (test_opt(sbi, NOHEAP) &&
-		(seg_type == CURSEG_HOT_DATA || IS_NODESEG(seg_type)))
+	if (seg_type == CURSEG_HOT_DATA || IS_NODESEG(seg_type))
 		return 0;
 
 	if (SIT_I(sbi)->last_victim[ALLOC_NEXT])
@@ -2788,30 +2814,31 @@ static unsigned int __get_next_segno(struct f2fs_sb_info *sbi, int type)
  * Allocate a current working segment.
  * This function always allocates a free segment in LFS manner.
  */
-static void new_curseg(struct f2fs_sb_info *sbi, int type, bool new_sec)
+static int new_curseg(struct f2fs_sb_info *sbi, int type, bool new_sec)
 {
 	struct curseg_info *curseg = CURSEG_I(sbi, type);
-	unsigned short seg_type = curseg->seg_type;
 	unsigned int segno = curseg->segno;
-	int dir = ALLOC_LEFT;
+	bool pinning = type == CURSEG_COLD_DATA_PINNED;
+	int ret;
 
 	if (curseg->inited)
-		write_sum_page(sbi, curseg->sum_blk,
-				GET_SUM_BLOCK(sbi, segno));
-	if (seg_type == CURSEG_WARM_DATA || seg_type == CURSEG_COLD_DATA)
-		dir = ALLOC_RIGHT;
-
-	if (test_opt(sbi, NOHEAP))
-		dir = ALLOC_RIGHT;
+		write_sum_page(sbi, curseg->sum_blk, GET_SUM_BLOCK(sbi, segno));
 
 	segno = __get_next_segno(sbi, type);
-	get_new_segment(sbi, &segno, new_sec, dir);
+	ret = get_new_segment(sbi, &segno, new_sec, pinning);
+	if (ret) {
+		if (ret == -ENOSPC)
+			curseg->segno = NULL_SEGNO;
+		return ret;
+	}
+
 	curseg->next_segno = segno;
 	reset_curseg(sbi, type, 1);
 	curseg->alloc_type = LFS;
 	if (F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_BLK)
 		curseg->fragment_remained_chunk =
 				get_random_u32_inclusive(1, sbi->max_fragment_chunk);
+	return 0;
 }
 
 static int __next_free_blkoff(struct f2fs_sb_info *sbi,
@@ -2827,7 +2854,7 @@ static int __next_free_blkoff(struct f2fs_sb_info *sbi,
 	for (i = 0; i < entries; i++)
 		target_map[i] = ckpt_map[i] | cur_map[i];
 
-	return __find_rev_next_zero_bit(target_map, sbi->blocks_per_seg, start);
+	return __find_rev_next_zero_bit(target_map, BLKS_PER_SEG(sbi), start);
 }
 
 static int f2fs_find_next_ssr_block(struct f2fs_sb_info *sbi,
@@ -2838,14 +2865,14 @@ static int f2fs_find_next_ssr_block(struct f2fs_sb_info *sbi,
 
 bool f2fs_segment_has_free_slot(struct f2fs_sb_info *sbi, int segno)
 {
-	return __next_free_blkoff(sbi, segno, 0) < sbi->blocks_per_seg;
+	return __next_free_blkoff(sbi, segno, 0) < BLKS_PER_SEG(sbi);
 }
 
 /*
  * This function always allocates a used segment(from dirty seglist) by SSR
  * manner, so it should recover the existing segment information of valid blocks
  */
-static void change_curseg(struct f2fs_sb_info *sbi, int type)
+static int change_curseg(struct f2fs_sb_info *sbi, int type)
 {
 	struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
 	struct curseg_info *curseg = CURSEG_I(sbi, type);
@@ -2870,21 +2897,23 @@ static void change_curseg(struct f2fs_sb_info *sbi, int type)
 	if (IS_ERR(sum_page)) {
 		/* GC won't be able to use stale summary pages by cp_error */
 		memset(curseg->sum_blk, 0, SUM_ENTRY_SIZE);
-		return;
+		return PTR_ERR(sum_page);
 	}
 	sum_node = (struct f2fs_summary_block *)page_address(sum_page);
 	memcpy(curseg->sum_blk, sum_node, SUM_ENTRY_SIZE);
 	f2fs_put_page(sum_page, 1);
+	return 0;
 }
 
 static int get_ssr_segment(struct f2fs_sb_info *sbi, int type,
 				int alloc_mode, unsigned long long age);
 
-static void get_atssr_segment(struct f2fs_sb_info *sbi, int type,
+static int get_atssr_segment(struct f2fs_sb_info *sbi, int type,
 					int target_type, int alloc_mode,
 					unsigned long long age)
 {
 	struct curseg_info *curseg = CURSEG_I(sbi, type);
+	int ret = 0;
 
 	curseg->seg_type = target_type;
 
@@ -2892,38 +2921,41 @@ static void get_atssr_segment(struct f2fs_sb_info *sbi, int type,
 		struct seg_entry *se = get_seg_entry(sbi, curseg->next_segno);
 
 		curseg->seg_type = se->type;
-		change_curseg(sbi, type);
+		ret = change_curseg(sbi, type);
 	} else {
 		/* allocate cold segment by default */
 		curseg->seg_type = CURSEG_COLD_DATA;
-		new_curseg(sbi, type, true);
+		ret = new_curseg(sbi, type, true);
 	}
 	stat_inc_seg_type(sbi, curseg);
+	return ret;
 }
 
-static void __f2fs_init_atgc_curseg(struct f2fs_sb_info *sbi)
+static int __f2fs_init_atgc_curseg(struct f2fs_sb_info *sbi)
 {
 	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_ALL_DATA_ATGC);
+	int ret = 0;
 
 	if (!sbi->am.atgc_enabled)
-		return;
+		return 0;
 
 	f2fs_down_read(&SM_I(sbi)->curseg_lock);
 
 	mutex_lock(&curseg->curseg_mutex);
 	down_write(&SIT_I(sbi)->sentry_lock);
 
-	get_atssr_segment(sbi, CURSEG_ALL_DATA_ATGC, CURSEG_COLD_DATA, SSR, 0);
+	ret = get_atssr_segment(sbi, CURSEG_ALL_DATA_ATGC,
+					CURSEG_COLD_DATA, SSR, 0);
 
 	up_write(&SIT_I(sbi)->sentry_lock);
 	mutex_unlock(&curseg->curseg_mutex);
 
 	f2fs_up_read(&SM_I(sbi)->curseg_lock);
-
+	return ret;
 }
-void f2fs_init_inmem_curseg(struct f2fs_sb_info *sbi)
+int f2fs_init_inmem_curseg(struct f2fs_sb_info *sbi)
 {
-	__f2fs_init_atgc_curseg(sbi);
+	return __f2fs_init_atgc_curseg(sbi);
 }
 
 static void __f2fs_save_inmem_curseg(struct f2fs_sb_info *sbi, int type)
@@ -3042,8 +3074,7 @@ static bool need_new_seg(struct f2fs_sb_info *sbi, int type)
 	if (!is_set_ckpt_flags(sbi, CP_CRC_RECOVERY_FLAG) &&
 	    curseg->seg_type == CURSEG_WARM_NODE)
 		return true;
-	if (curseg->alloc_type == LFS &&
-	    is_next_segment_free(sbi, curseg, type) &&
+	if (curseg->alloc_type == LFS && is_next_segment_free(sbi, curseg) &&
 	    likely(!is_sbi_flag_set(sbi, SBI_CP_DISABLED)))
 		return true;
 	if (!f2fs_need_SSR(sbi) || !get_ssr_segment(sbi, type, SSR, 0))
@@ -3051,11 +3082,12 @@ static bool need_new_seg(struct f2fs_sb_info *sbi, int type)
 	return false;
 }
 
-void f2fs_allocate_segment_for_resize(struct f2fs_sb_info *sbi, int type,
+int f2fs_allocate_segment_for_resize(struct f2fs_sb_info *sbi, int type,
 					unsigned int start, unsigned int end)
 {
 	struct curseg_info *curseg = CURSEG_I(sbi, type);
 	unsigned int segno;
+	int ret = 0;
 
 	f2fs_down_read(&SM_I(sbi)->curseg_lock);
 	mutex_lock(&curseg->curseg_mutex);
@@ -3066,9 +3098,9 @@ void f2fs_allocate_segment_for_resize(struct f2fs_sb_info *sbi, int type,
 		goto unlock;
 
 	if (f2fs_need_SSR(sbi) && get_ssr_segment(sbi, type, SSR, 0))
-		change_curseg(sbi, type);
+		ret = change_curseg(sbi, type);
 	else
-		new_curseg(sbi, type, true);
+		ret = new_curseg(sbi, type, true);
 
 	stat_inc_seg_type(sbi, curseg);
 
@@ -3082,45 +3114,84 @@ unlock:
 
 	mutex_unlock(&curseg->curseg_mutex);
 	f2fs_up_read(&SM_I(sbi)->curseg_lock);
+	return ret;
 }
 
-static void __allocate_new_segment(struct f2fs_sb_info *sbi, int type,
+static int __allocate_new_segment(struct f2fs_sb_info *sbi, int type,
 						bool new_sec, bool force)
 {
 	struct curseg_info *curseg = CURSEG_I(sbi, type);
 	unsigned int old_segno;
+	int err = 0;
+
+	if (type == CURSEG_COLD_DATA_PINNED && !curseg->inited)
+		goto allocate;
 
 	if (!force && curseg->inited &&
 	    !curseg->next_blkoff &&
 	    !get_valid_blocks(sbi, curseg->segno, new_sec) &&
 	    !get_ckpt_valid_blocks(sbi, curseg->segno, new_sec))
-		return;
+		return 0;
 
+allocate:
 	old_segno = curseg->segno;
-	new_curseg(sbi, type, true);
+	err = new_curseg(sbi, type, true);
+	if (err)
+		return err;
 	stat_inc_seg_type(sbi, curseg);
 	locate_dirty_segment(sbi, old_segno);
+	return 0;
 }
 
-void f2fs_allocate_new_section(struct f2fs_sb_info *sbi, int type, bool force)
+int f2fs_allocate_new_section(struct f2fs_sb_info *sbi, int type, bool force)
 {
+	int ret;
+
 	f2fs_down_read(&SM_I(sbi)->curseg_lock);
 	down_write(&SIT_I(sbi)->sentry_lock);
-	__allocate_new_segment(sbi, type, true, force);
+	ret = __allocate_new_segment(sbi, type, true, force);
 	up_write(&SIT_I(sbi)->sentry_lock);
 	f2fs_up_read(&SM_I(sbi)->curseg_lock);
+
+	return ret;
 }
 
-void f2fs_allocate_new_segments(struct f2fs_sb_info *sbi)
+int f2fs_allocate_pinning_section(struct f2fs_sb_info *sbi)
+{
+	int err;
+	bool gc_required = true;
+
+retry:
+	f2fs_lock_op(sbi);
+	err = f2fs_allocate_new_section(sbi, CURSEG_COLD_DATA_PINNED, false);
+	f2fs_unlock_op(sbi);
+
+	if (f2fs_sb_has_blkzoned(sbi) && err == -EAGAIN && gc_required) {
+		f2fs_down_write(&sbi->gc_lock);
+		err = f2fs_gc_range(sbi, 0, GET_SEGNO(sbi, FDEV(0).end_blk), true, 1);
+		f2fs_up_write(&sbi->gc_lock);
+
+		gc_required = false;
+		if (!err)
+			goto retry;
+	}
+
+	return err;
+}
+
+int f2fs_allocate_new_segments(struct f2fs_sb_info *sbi)
 {
 	int i;
+	int err = 0;
 
 	f2fs_down_read(&SM_I(sbi)->curseg_lock);
 	down_write(&SIT_I(sbi)->sentry_lock);
 	for (i = CURSEG_HOT_DATA; i <= CURSEG_COLD_DATA; i++)
-		__allocate_new_segment(sbi, i, false, false);
+		err += __allocate_new_segment(sbi, i, false, false);
 	up_write(&SIT_I(sbi)->sentry_lock);
 	f2fs_up_read(&SM_I(sbi)->curseg_lock);
+
+	return err;
 }
 
 bool f2fs_exist_trim_candidates(struct f2fs_sb_info *sbi,
@@ -3238,8 +3309,8 @@ int f2fs_trim_fs(struct f2fs_sb_info *sbi, struct fstrim_range *range)
 	end_segno = (end >= MAX_BLKADDR(sbi)) ? MAIN_SEGS(sbi) - 1 :
 						GET_SEGNO(sbi, end);
 	if (need_align) {
-		start_segno = rounddown(start_segno, sbi->segs_per_sec);
-		end_segno = roundup(end_segno + 1, sbi->segs_per_sec) - 1;
+		start_segno = rounddown(start_segno, SEGS_PER_SEC(sbi));
+		end_segno = roundup(end_segno + 1, SEGS_PER_SEC(sbi)) - 1;
 	}
 
 	cpc.reason = CP_DISCARD;
@@ -3281,8 +3352,14 @@ out:
 	return err;
 }
 
-int f2fs_rw_hint_to_seg_type(enum rw_hint hint)
+int f2fs_rw_hint_to_seg_type(struct f2fs_sb_info *sbi, enum rw_hint hint)
 {
+	if (F2FS_OPTION(sbi).active_logs == 2)
+		return CURSEG_HOT_DATA;
+	else if (F2FS_OPTION(sbi).active_logs == 4)
+		return CURSEG_COLD_DATA;
+
+	/* active_log == 6 */
 	switch (hint) {
 	case WRITE_LIFE_SHORT:
 		return CURSEG_HOT_DATA;
@@ -3293,6 +3370,65 @@ int f2fs_rw_hint_to_seg_type(enum rw_hint hint)
 	}
 }
 
+/*
+ * This returns write hints for each segment type. This hints will be
+ * passed down to block layer as below by default.
+ *
+ * User                  F2FS                     Block
+ * ----                  ----                     -----
+ *                       META                     WRITE_LIFE_NONE|REQ_META
+ *                       HOT_NODE                 WRITE_LIFE_NONE
+ *                       WARM_NODE                WRITE_LIFE_MEDIUM
+ *                       COLD_NODE                WRITE_LIFE_LONG
+ * ioctl(COLD)           COLD_DATA                WRITE_LIFE_EXTREME
+ * extension list        "                        "
+ *
+ * -- buffered io
+ *                       COLD_DATA                WRITE_LIFE_EXTREME
+ *                       HOT_DATA                 WRITE_LIFE_SHORT
+ *                       WARM_DATA                WRITE_LIFE_NOT_SET
+ *
+ * -- direct io
+ * WRITE_LIFE_EXTREME    COLD_DATA                WRITE_LIFE_EXTREME
+ * WRITE_LIFE_SHORT      HOT_DATA                 WRITE_LIFE_SHORT
+ * WRITE_LIFE_NOT_SET    WARM_DATA                WRITE_LIFE_NOT_SET
+ * WRITE_LIFE_NONE       "                        WRITE_LIFE_NONE
+ * WRITE_LIFE_MEDIUM     "                        WRITE_LIFE_MEDIUM
+ * WRITE_LIFE_LONG       "                        WRITE_LIFE_LONG
+ */
+enum rw_hint f2fs_io_type_to_rw_hint(struct f2fs_sb_info *sbi,
+				enum page_type type, enum temp_type temp)
+{
+	switch (type) {
+	case DATA:
+		switch (temp) {
+		case WARM:
+			return WRITE_LIFE_NOT_SET;
+		case HOT:
+			return WRITE_LIFE_SHORT;
+		case COLD:
+			return WRITE_LIFE_EXTREME;
+		default:
+			return WRITE_LIFE_NONE;
+		}
+	case NODE:
+		switch (temp) {
+		case WARM:
+			return WRITE_LIFE_MEDIUM;
+		case HOT:
+			return WRITE_LIFE_NONE;
+		case COLD:
+			return WRITE_LIFE_LONG;
+		default:
+			return WRITE_LIFE_NONE;
+		}
+	case META:
+		return WRITE_LIFE_NONE;
+	default:
+		return WRITE_LIFE_NONE;
+	}
+}
+
 static int __get_segment_type_2(struct f2fs_io_info *fio)
 {
 	if (fio->type == DATA)
@@ -3347,7 +3483,9 @@ static int __get_segment_type_6(struct f2fs_io_info *fio)
 		if (page_private_gcing(fio->page)) {
 			if (fio->sbi->am.atgc_enabled &&
 				(fio->io_type == FS_DATA_IO) &&
-				(fio->sbi->gc_mode != GC_URGENT_HIGH))
+				(fio->sbi->gc_mode != GC_URGENT_HIGH) &&
+				__is_valid_data_blkaddr(fio->old_blkaddr) &&
+				!is_inode_flag_set(inode, FI_OPU_WRITE))
 				return CURSEG_ALL_DATA_ATGC;
 			else
 				return CURSEG_COLD_DATA;
@@ -3363,7 +3501,8 @@ static int __get_segment_type_6(struct f2fs_io_info *fio)
 				is_inode_flag_set(inode, FI_HOT_DATA) ||
 				f2fs_is_cow_file(inode))
 			return CURSEG_HOT_DATA;
-		return f2fs_rw_hint_to_seg_type(inode->i_write_hint);
+		return f2fs_rw_hint_to_seg_type(F2FS_I_SB(inode),
+						inode->i_write_hint);
 	} else {
 		if (IS_DNODE(fio->page))
 			return is_cold_node(fio->page) ? CURSEG_WARM_NODE :
@@ -3372,6 +3511,15 @@ static int __get_segment_type_6(struct f2fs_io_info *fio)
 	}
 }
 
+int f2fs_get_segment_temp(int seg_type)
+{
+	if (IS_HOT(seg_type))
+		return HOT;
+	else if (IS_WARM(seg_type))
+		return WARM;
+	return COLD;
+}
+
 static int __get_segment_type(struct f2fs_io_info *fio)
 {
 	int type = 0;
@@ -3390,12 +3538,8 @@ static int __get_segment_type(struct f2fs_io_info *fio)
 		f2fs_bug_on(fio->sbi, true);
 	}
 
-	if (IS_HOT(type))
-		fio->temp = HOT;
-	else if (IS_WARM(type))
-		fio->temp = WARM;
-	else
-		fio->temp = COLD;
+	fio->temp = f2fs_get_segment_temp(type);
+
 	return type;
 }
 
@@ -3412,7 +3556,14 @@ static void f2fs_randomize_chunk(struct f2fs_sb_info *sbi,
 		get_random_u32_inclusive(1, sbi->max_fragment_hole);
 }
 
-void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
+static void reset_curseg_fields(struct curseg_info *curseg)
+{
+	curseg->inited = false;
+	curseg->segno = NULL_SEGNO;
+	curseg->next_segno = 0;
+}
+
+int f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
 		block_t old_blkaddr, block_t *new_blkaddr,
 		struct f2fs_summary *sum, int type,
 		struct f2fs_io_info *fio)
@@ -3423,12 +3574,18 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
 	bool from_gc = (type == CURSEG_ALL_DATA_ATGC);
 	struct seg_entry *se = NULL;
 	bool segment_full = false;
+	int ret = 0;
 
 	f2fs_down_read(&SM_I(sbi)->curseg_lock);
 
 	mutex_lock(&curseg->curseg_mutex);
 	down_write(&sit_i->sentry_lock);
 
+	if (curseg->segno == NULL_SEGNO) {
+		ret = -ENOSPC;
+		goto out_err;
+	}
+
 	if (from_gc) {
 		f2fs_bug_on(sbi, GET_SEGNO(sbi, old_blkaddr) == NULL_SEGNO);
 		se = get_seg_entry(sbi, GET_SEGNO(sbi, old_blkaddr));
@@ -3437,7 +3594,7 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
 	}
 	*new_blkaddr = NEXT_FREE_BLKADDR(sbi, curseg);
 
-	f2fs_bug_on(sbi, curseg->next_blkoff >= sbi->blocks_per_seg);
+	f2fs_bug_on(sbi, curseg->next_blkoff >= BLKS_PER_SEG(sbi));
 
 	f2fs_wait_discard_bio(sbi, *new_blkaddr);
 
@@ -3466,25 +3623,37 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
 	 * since SSR needs latest valid block information.
 	 */
 	update_sit_entry(sbi, *new_blkaddr, 1);
-	if (GET_SEGNO(sbi, old_blkaddr) != NULL_SEGNO)
-		update_sit_entry(sbi, old_blkaddr, -1);
+	update_sit_entry(sbi, old_blkaddr, -1);
 
 	/*
 	 * If the current segment is full, flush it out and replace it with a
 	 * new segment.
 	 */
 	if (segment_full) {
+		if (type == CURSEG_COLD_DATA_PINNED &&
+		    !((curseg->segno + 1) % sbi->segs_per_sec)) {
+			write_sum_page(sbi, curseg->sum_blk,
+					GET_SUM_BLOCK(sbi, curseg->segno));
+			reset_curseg_fields(curseg);
+			goto skip_new_segment;
+		}
+
 		if (from_gc) {
-			get_atssr_segment(sbi, type, se->type,
+			ret = get_atssr_segment(sbi, type, se->type,
 						AT_SSR, se->mtime);
 		} else {
 			if (need_new_seg(sbi, type))
-				new_curseg(sbi, type, false);
+				ret = new_curseg(sbi, type, false);
 			else
-				change_curseg(sbi, type);
+				ret = change_curseg(sbi, type);
 			stat_inc_seg_type(sbi, curseg);
 		}
+
+		if (ret)
+			goto out_err;
 	}
+
+skip_new_segment:
 	/*
 	 * segment dirty status should be updated after segment allocation,
 	 * so we just need to update status only one time after previous
@@ -3507,9 +3676,6 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
 	if (fio) {
 		struct f2fs_bio_info *io;
 
-		if (F2FS_IO_ALIGNED(sbi))
-			fio->retry = 0;
-
 		INIT_LIST_HEAD(&fio->list);
 		fio->in_list = 1;
 		io = sbi->write_io[fio->type] + fio->temp;
@@ -3519,8 +3685,15 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
 	}
 
 	mutex_unlock(&curseg->curseg_mutex);
+	f2fs_up_read(&SM_I(sbi)->curseg_lock);
+	return 0;
 
+out_err:
+	*new_blkaddr = NULL_ADDR;
+	up_write(&sit_i->sentry_lock);
+	mutex_unlock(&curseg->curseg_mutex);
 	f2fs_up_read(&SM_I(sbi)->curseg_lock);
+	return ret;
 }
 
 void f2fs_update_device_state(struct f2fs_sb_info *sbi, nid_t ino,
@@ -3557,21 +3730,24 @@ static void do_write_page(struct f2fs_summary *sum, struct f2fs_io_info *fio)
 
 	if (keep_order)
 		f2fs_down_read(&fio->sbi->io_order_lock);
-reallocate:
-	f2fs_allocate_data_block(fio->sbi, fio->page, fio->old_blkaddr,
-			&fio->new_blkaddr, sum, type, fio);
+
+	if (f2fs_allocate_data_block(fio->sbi, fio->page, fio->old_blkaddr,
+			&fio->new_blkaddr, sum, type, fio)) {
+		if (fscrypt_inode_uses_fs_layer_crypto(fio->page->mapping->host))
+			fscrypt_finalize_bounce_page(&fio->encrypted_page);
+		end_page_writeback(fio->page);
+		if (f2fs_in_warm_node_list(fio->sbi, fio->page))
+			f2fs_del_fsync_node_entry(fio->sbi, fio->page);
+		goto out;
+	}
 	if (GET_SEGNO(fio->sbi, fio->old_blkaddr) != NULL_SEGNO)
 		f2fs_invalidate_internal_cache(fio->sbi, fio->old_blkaddr);
 
 	/* writeout dirty page into bdev */
 	f2fs_submit_page_write(fio);
-	if (fio->retry) {
-		fio->old_blkaddr = fio->new_blkaddr;
-		goto reallocate;
-	}
 
 	f2fs_update_device_state(fio->sbi, fio->ino, fio->new_blkaddr, 1);
-
+out:
 	if (keep_order)
 		f2fs_up_read(&fio->sbi->io_order_lock);
 }
@@ -3654,7 +3830,7 @@ int f2fs_inplace_write_data(struct f2fs_io_info *fio)
 		goto drop_bio;
 	}
 
-	if (fio->post_read)
+	if (fio->meta_gc)
 		f2fs_truncate_meta_inode_pages(sbi, fio->new_blkaddr, 1);
 
 	stat_inc_inplace_blocks(fio->sbi);
@@ -3744,7 +3920,8 @@ void f2fs_do_replace_block(struct f2fs_sb_info *sbi, struct f2fs_summary *sum,
 	/* change the current segment */
 	if (segno != curseg->segno) {
 		curseg->next_segno = segno;
-		change_curseg(sbi, type);
+		if (change_curseg(sbi, type))
+			goto out_unlock;
 	}
 
 	curseg->next_blkoff = GET_BLKOFF_FROM_SEG0(sbi, new_blkaddr);
@@ -3770,12 +3947,14 @@ void f2fs_do_replace_block(struct f2fs_sb_info *sbi, struct f2fs_summary *sum,
 	if (recover_curseg) {
 		if (old_cursegno != curseg->segno) {
 			curseg->next_segno = old_cursegno;
-			change_curseg(sbi, type);
+			if (change_curseg(sbi, type))
+				goto out_unlock;
 		}
 		curseg->next_blkoff = old_blkoff;
 		curseg->alloc_type = old_alloc_type;
 	}
 
+out_unlock:
 	up_write(&sit_i->sentry_lock);
 	mutex_unlock(&curseg->curseg_mutex);
 	f2fs_up_write(&SM_I(sbi)->curseg_lock);
@@ -3799,7 +3978,7 @@ void f2fs_replace_block(struct f2fs_sb_info *sbi, struct dnode_of_data *dn,
 void f2fs_wait_on_page_writeback(struct page *page,
 				enum page_type type, bool ordered, bool locked)
 {
-	if (PageWriteback(page)) {
+	if (folio_test_writeback(page_folio(page))) {
 		struct f2fs_sb_info *sbi = F2FS_P_SB(page);
 
 		/* submit cached LFS IO */
@@ -3808,7 +3987,8 @@ void f2fs_wait_on_page_writeback(struct page *page,
 		f2fs_submit_merged_ipu_write(sbi, NULL, page);
 		if (ordered) {
 			wait_on_page_writeback(page);
-			f2fs_bug_on(sbi, locked && PageWriteback(page));
+			f2fs_bug_on(sbi, locked &&
+				folio_test_writeback(page_folio(page)));
 		} else {
 			wait_for_stable_page(page);
 		}
@@ -3820,7 +4000,7 @@ void f2fs_wait_on_block_writeback(struct inode *inode, block_t blkaddr)
 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 	struct page *cpage;
 
-	if (!f2fs_post_read_required(inode))
+	if (!f2fs_meta_inode_gc_required(inode))
 		return;
 
 	if (!__is_valid_data_blkaddr(blkaddr))
@@ -3839,7 +4019,7 @@ void f2fs_wait_on_block_writeback_range(struct inode *inode, block_t blkaddr,
 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 	block_t i;
 
-	if (!f2fs_post_read_required(inode))
+	if (!f2fs_meta_inode_gc_required(inode))
 		return;
 
 	for (i = 0; i < len; i++)
@@ -3887,7 +4067,7 @@ static int read_compacted_summaries(struct f2fs_sb_info *sbi)
 		seg_i->next_blkoff = blk_off;
 
 		if (seg_i->alloc_type == SSR)
-			blk_off = sbi->blocks_per_seg;
+			blk_off = BLKS_PER_SEG(sbi);
 
 		for (j = 0; j < blk_off; j++) {
 			struct f2fs_summary *s;
@@ -3955,7 +4135,7 @@ static int read_normal_summaries(struct f2fs_sb_info *sbi, int type)
 			struct f2fs_summary *ns = &sum->entries[0];
 			int i;
 
-			for (i = 0; i < sbi->blocks_per_seg; i++, ns++) {
+			for (i = 0; i < BLKS_PER_SEG(sbi); i++, ns++) {
 				ns->version = 0;
 				ns->ofs_in_node = 0;
 			}
@@ -4461,7 +4641,7 @@ static int build_sit_info(struct f2fs_sb_info *sbi)
 #endif
 
 	sit_i->sit_base_addr = le32_to_cpu(raw_super->sit_blkaddr);
-	sit_i->sit_blocks = sit_segs << sbi->log_blocks_per_seg;
+	sit_i->sit_blocks = SEGS_TO_BLKS(sbi, sit_segs);
 	sit_i->written_valid_blocks = 0;
 	sit_i->bitmap_size = sit_bitmap_size;
 	sit_i->dirty_sentries = 0;
@@ -4534,9 +4714,7 @@ static int build_curseg(struct f2fs_sb_info *sbi)
 			array[i].seg_type = CURSEG_COLD_DATA;
 		else if (i == CURSEG_ALL_DATA_ATGC)
 			array[i].seg_type = CURSEG_COLD_DATA;
-		array[i].segno = NULL_SEGNO;
-		array[i].next_blkoff = 0;
-		array[i].inited = false;
+		reset_curseg_fields(&array[i]);
 	}
 	return restore_curseg_summaries(sbi);
 }
@@ -4588,21 +4766,20 @@ static int build_sit_entries(struct f2fs_sb_info *sbi)
 
 			sit_valid_blocks[SE_PAGETYPE(se)] += se->valid_blocks;
 
-			if (f2fs_block_unit_discard(sbi)) {
-				/* build discard map only one time */
-				if (is_set_ckpt_flags(sbi, CP_TRIMMED_FLAG)) {
-					memset(se->discard_map, 0xff,
+			if (!f2fs_block_unit_discard(sbi))
+				goto init_discard_map_done;
+
+			/* build discard map only one time */
+			if (is_set_ckpt_flags(sbi, CP_TRIMMED_FLAG)) {
+				memset(se->discard_map, 0xff,
 						SIT_VBLOCK_MAP_SIZE);
-				} else {
-					memcpy(se->discard_map,
-						se->cur_valid_map,
+				goto init_discard_map_done;
+			}
+			memcpy(se->discard_map, se->cur_valid_map,
 						SIT_VBLOCK_MAP_SIZE);
-					sbi->discard_blks +=
-						sbi->blocks_per_seg -
+			sbi->discard_blks += BLKS_PER_SEG(sbi) -
 						se->valid_blocks;
-				}
-			}
-
+init_discard_map_done:
 			if (__is_large_section(sbi))
 				get_sec_entry(sbi, start)->valid_blocks +=
 							se->valid_blocks;
@@ -4742,7 +4919,7 @@ static void init_dirty_segmap(struct f2fs_sb_info *sbi)
 		return;
 
 	mutex_lock(&dirty_i->seglist_lock);
-	for (segno = 0; segno < MAIN_SEGS(sbi); segno += sbi->segs_per_sec) {
+	for (segno = 0; segno < MAIN_SEGS(sbi); segno += SEGS_PER_SEC(sbi)) {
 		valid_blocks = get_valid_blocks(sbi, segno, true);
 		secno = GET_SEC_FROM_SEG(sbi, segno);
 
@@ -4841,7 +5018,7 @@ static int sanity_check_curseg(struct f2fs_sb_info *sbi)
 		if (curseg->alloc_type == SSR)
 			continue;
 
-		for (blkofs += 1; blkofs < sbi->blocks_per_seg; blkofs++) {
+		for (blkofs += 1; blkofs < BLKS_PER_SEG(sbi); blkofs++) {
 			if (!f2fs_test_bit(blkofs, se->cur_valid_map))
 				continue;
 out:
@@ -4857,7 +5034,6 @@ out:
 }
 
 #ifdef CONFIG_BLK_DEV_ZONED
-
 static int check_zone_write_pointer(struct f2fs_sb_info *sbi,
 				    struct f2fs_dev_info *fdev,
 				    struct blk_zone *zone)
@@ -4866,6 +5042,7 @@ static int check_zone_write_pointer(struct f2fs_sb_info *sbi,
 	block_t zone_block, valid_block_cnt;
 	unsigned int log_sectors_per_block = sbi->log_blocksize - SECTOR_SHIFT;
 	int ret;
+	unsigned int nofs_flags;
 
 	if (zone->type != BLK_ZONE_TYPE_SEQWRITE_REQ)
 		return 0;
@@ -4877,14 +5054,19 @@ static int check_zone_write_pointer(struct f2fs_sb_info *sbi,
 	 * Skip check of zones cursegs point to, since
 	 * fix_curseg_write_pointer() checks them.
 	 */
-	if (zone_segno >= MAIN_SEGS(sbi) ||
-	    IS_CURSEC(sbi, GET_SEC_FROM_SEG(sbi, zone_segno)))
+	if (zone_segno >= MAIN_SEGS(sbi))
 		return 0;
 
 	/*
 	 * Get # of valid block of the zone.
 	 */
 	valid_block_cnt = get_valid_blocks(sbi, zone_segno, true);
+	if (IS_CURSEC(sbi, GET_SEC_FROM_SEG(sbi, zone_segno))) {
+		f2fs_notice(sbi, "Open zones: valid block[0x%x,0x%x] cond[%s]",
+				zone_segno, valid_block_cnt,
+				blk_zone_cond_str(zone->cond));
+		return 0;
+	}
 
 	if ((!valid_block_cnt && zone->cond == BLK_ZONE_COND_EMPTY) ||
 	    (valid_block_cnt && zone->cond == BLK_ZONE_COND_FULL))
@@ -4892,8 +5074,8 @@ static int check_zone_write_pointer(struct f2fs_sb_info *sbi,
 
 	if (!valid_block_cnt) {
 		f2fs_notice(sbi, "Zone without valid block has non-zero write "
-			    "pointer. Reset the write pointer: cond[0x%x]",
-			    zone->cond);
+			    "pointer. Reset the write pointer: cond[%s]",
+			    blk_zone_cond_str(zone->cond));
 		ret = __f2fs_issue_discard_zone(sbi, fdev->bdev, zone_block,
 					zone->len >> log_sectors_per_block);
 		if (ret)
@@ -4910,11 +5092,13 @@ static int check_zone_write_pointer(struct f2fs_sb_info *sbi,
 	 * selected for write operation until it get discarded.
 	 */
 	f2fs_notice(sbi, "Valid blocks are not aligned with write "
-		    "pointer: valid block[0x%x,0x%x] cond[0x%x]",
-		    zone_segno, valid_block_cnt, zone->cond);
+		    "pointer: valid block[0x%x,0x%x] cond[%s]",
+		    zone_segno, valid_block_cnt, blk_zone_cond_str(zone->cond));
 
+	nofs_flags = memalloc_nofs_save();
 	ret = blkdev_zone_mgmt(fdev->bdev, REQ_OP_ZONE_FINISH,
-				zone->start, zone->len, GFP_NOFS);
+				zone->start, zone->len);
+	memalloc_nofs_restore(nofs_flags);
 	if (ret == -EOPNOTSUPP) {
 		ret = blkdev_issue_zeroout(fdev->bdev, zone->wp,
 					zone->len - (zone->wp - zone->start),
@@ -5120,7 +5304,7 @@ static inline unsigned int f2fs_usable_zone_blks_in_seg(
 	unsigned int secno;
 
 	if (!sbi->unusable_blocks_per_sec)
-		return sbi->blocks_per_seg;
+		return BLKS_PER_SEG(sbi);
 
 	secno = GET_SEC_FROM_SEG(sbi, segno);
 	seg_start = START_BLOCK(sbi, segno);
@@ -5135,10 +5319,10 @@ static inline unsigned int f2fs_usable_zone_blks_in_seg(
 	 */
 	if (seg_start >= sec_cap_blkaddr)
 		return 0;
-	if (seg_start + sbi->blocks_per_seg > sec_cap_blkaddr)
+	if (seg_start + BLKS_PER_SEG(sbi) > sec_cap_blkaddr)
 		return sec_cap_blkaddr - seg_start;
 
-	return sbi->blocks_per_seg;
+	return BLKS_PER_SEG(sbi);
 }
 #else
 int f2fs_fix_curseg_write_pointer(struct f2fs_sb_info *sbi)
@@ -5164,7 +5348,7 @@ unsigned int f2fs_usable_blks_in_seg(struct f2fs_sb_info *sbi,
 	if (f2fs_sb_has_blkzoned(sbi))
 		return f2fs_usable_zone_blks_in_seg(sbi, segno);
 
-	return sbi->blocks_per_seg;
+	return BLKS_PER_SEG(sbi);
 }
 
 unsigned int f2fs_usable_segs_in_sec(struct f2fs_sb_info *sbi,
@@ -5173,7 +5357,7 @@ unsigned int f2fs_usable_segs_in_sec(struct f2fs_sb_info *sbi,
 	if (f2fs_sb_has_blkzoned(sbi))
 		return CAP_SEGS_PER_SEC(sbi);
 
-	return sbi->segs_per_sec;
+	return SEGS_PER_SEC(sbi);
 }
 
 /*
@@ -5188,14 +5372,14 @@ static void init_min_max_mtime(struct f2fs_sb_info *sbi)
 
 	sit_i->min_mtime = ULLONG_MAX;
 
-	for (segno = 0; segno < MAIN_SEGS(sbi); segno += sbi->segs_per_sec) {
+	for (segno = 0; segno < MAIN_SEGS(sbi); segno += SEGS_PER_SEC(sbi)) {
 		unsigned int i;
 		unsigned long long mtime = 0;
 
-		for (i = 0; i < sbi->segs_per_sec; i++)
+		for (i = 0; i < SEGS_PER_SEC(sbi); i++)
 			mtime += get_seg_entry(sbi, segno + i)->mtime;
 
-		mtime = div_u64(mtime, sbi->segs_per_sec);
+		mtime = div_u64(mtime, SEGS_PER_SEC(sbi));
 
 		if (sit_i->min_mtime > mtime)
 			sit_i->min_mtime = mtime;
@@ -5234,7 +5418,7 @@ int f2fs_build_segment_manager(struct f2fs_sb_info *sbi)
 		sm_info->ipu_policy = BIT(F2FS_IPU_FSYNC);
 	sm_info->min_ipu_util = DEF_MIN_IPU_UTIL;
 	sm_info->min_fsync_blocks = DEF_MIN_FSYNC_BLOCKS;
-	sm_info->min_seq_blocks = sbi->blocks_per_seg;
+	sm_info->min_seq_blocks = BLKS_PER_SEG(sbi);
 	sm_info->min_hot_blocks = DEF_MIN_HOT_BLOCKS;
 	sm_info->min_ssr_sections = reserved_sections(sbi);
 
diff --git a/fs/f2fs/segment.h b/fs/f2fs/segment.h
index c77a562831..bfc01a521c 100644
--- a/fs/f2fs/segment.h
+++ b/fs/f2fs/segment.h
@@ -48,21 +48,21 @@ static inline void sanity_check_seg_type(struct f2fs_sb_info *sbi,
 
 #define IS_CURSEC(sbi, secno)						\
 	(((secno) == CURSEG_I(sbi, CURSEG_HOT_DATA)->segno /		\
-	  (sbi)->segs_per_sec) ||	\
+	  SEGS_PER_SEC(sbi)) ||	\
 	 ((secno) == CURSEG_I(sbi, CURSEG_WARM_DATA)->segno /		\
-	  (sbi)->segs_per_sec) ||	\
+	  SEGS_PER_SEC(sbi)) ||	\
 	 ((secno) == CURSEG_I(sbi, CURSEG_COLD_DATA)->segno /		\
-	  (sbi)->segs_per_sec) ||	\
+	  SEGS_PER_SEC(sbi)) ||	\
 	 ((secno) == CURSEG_I(sbi, CURSEG_HOT_NODE)->segno /		\
-	  (sbi)->segs_per_sec) ||	\
+	  SEGS_PER_SEC(sbi)) ||	\
 	 ((secno) == CURSEG_I(sbi, CURSEG_WARM_NODE)->segno /		\
-	  (sbi)->segs_per_sec) ||	\
+	  SEGS_PER_SEC(sbi)) ||	\
 	 ((secno) == CURSEG_I(sbi, CURSEG_COLD_NODE)->segno /		\
-	  (sbi)->segs_per_sec) ||	\
+	  SEGS_PER_SEC(sbi)) ||	\
 	 ((secno) == CURSEG_I(sbi, CURSEG_COLD_DATA_PINNED)->segno /	\
-	  (sbi)->segs_per_sec) ||	\
+	  SEGS_PER_SEC(sbi)) ||	\
 	 ((secno) == CURSEG_I(sbi, CURSEG_ALL_DATA_ATGC)->segno /	\
-	  (sbi)->segs_per_sec))
+	  SEGS_PER_SEC(sbi)))
 
 #define MAIN_BLKADDR(sbi)						\
 	(SM_I(sbi) ? SM_I(sbi)->main_blkaddr : 				\
@@ -77,40 +77,37 @@ static inline void sanity_check_seg_type(struct f2fs_sb_info *sbi,
 #define TOTAL_SEGS(sbi)							\
 	(SM_I(sbi) ? SM_I(sbi)->segment_count : 				\
 		le32_to_cpu(F2FS_RAW_SUPER(sbi)->segment_count))
-#define TOTAL_BLKS(sbi)	(TOTAL_SEGS(sbi) << (sbi)->log_blocks_per_seg)
+#define TOTAL_BLKS(sbi)	(SEGS_TO_BLKS(sbi, TOTAL_SEGS(sbi)))
 
 #define MAX_BLKADDR(sbi)	(SEG0_BLKADDR(sbi) + TOTAL_BLKS(sbi))
 #define SEGMENT_SIZE(sbi)	(1ULL << ((sbi)->log_blocksize +	\
 					(sbi)->log_blocks_per_seg))
 
 #define START_BLOCK(sbi, segno)	(SEG0_BLKADDR(sbi) +			\
-	 (GET_R2L_SEGNO(FREE_I(sbi), segno) << (sbi)->log_blocks_per_seg))
+	 (SEGS_TO_BLKS(sbi, GET_R2L_SEGNO(FREE_I(sbi), segno))))
 
 #define NEXT_FREE_BLKADDR(sbi, curseg)					\
 	(START_BLOCK(sbi, (curseg)->segno) + (curseg)->next_blkoff)
 
 #define GET_SEGOFF_FROM_SEG0(sbi, blk_addr)	((blk_addr) - SEG0_BLKADDR(sbi))
 #define GET_SEGNO_FROM_SEG0(sbi, blk_addr)				\
-	(GET_SEGOFF_FROM_SEG0(sbi, blk_addr) >> (sbi)->log_blocks_per_seg)
+	(BLKS_TO_SEGS(sbi, GET_SEGOFF_FROM_SEG0(sbi, blk_addr)))
 #define GET_BLKOFF_FROM_SEG0(sbi, blk_addr)				\
-	(GET_SEGOFF_FROM_SEG0(sbi, blk_addr) & ((sbi)->blocks_per_seg - 1))
+	(GET_SEGOFF_FROM_SEG0(sbi, blk_addr) & (BLKS_PER_SEG(sbi) - 1))
 
 #define GET_SEGNO(sbi, blk_addr)					\
 	((!__is_valid_data_blkaddr(blk_addr)) ?			\
 	NULL_SEGNO : GET_L2R_SEGNO(FREE_I(sbi),			\
 		GET_SEGNO_FROM_SEG0(sbi, blk_addr)))
-#define BLKS_PER_SEC(sbi)					\
-	((sbi)->segs_per_sec * (sbi)->blocks_per_seg)
 #define CAP_BLKS_PER_SEC(sbi)					\
-	((sbi)->segs_per_sec * (sbi)->blocks_per_seg -		\
-	 (sbi)->unusable_blocks_per_sec)
+	(BLKS_PER_SEC(sbi) - (sbi)->unusable_blocks_per_sec)
 #define CAP_SEGS_PER_SEC(sbi)					\
-	((sbi)->segs_per_sec - ((sbi)->unusable_blocks_per_sec >>\
-	(sbi)->log_blocks_per_seg))
+	(SEGS_PER_SEC(sbi) -					\
+	BLKS_TO_SEGS(sbi, (sbi)->unusable_blocks_per_sec))
 #define GET_SEC_FROM_SEG(sbi, segno)				\
-	(((segno) == -1) ? -1 : (segno) / (sbi)->segs_per_sec)
+	(((segno) == -1) ? -1 : (segno) / SEGS_PER_SEC(sbi))
 #define GET_SEG_FROM_SEC(sbi, secno)				\
-	((secno) * (sbi)->segs_per_sec)
+	((secno) * SEGS_PER_SEC(sbi))
 #define GET_ZONE_FROM_SEC(sbi, secno)				\
 	(((secno) == -1) ? -1 : (secno) / (sbi)->secs_per_zone)
 #define GET_ZONE_FROM_SEG(sbi, segno)				\
@@ -139,16 +136,6 @@ static inline void sanity_check_seg_type(struct f2fs_sb_info *sbi,
 	((sectors) >> F2FS_LOG_SECTORS_PER_BLOCK)
 
 /*
- * indicate a block allocation direction: RIGHT and LEFT.
- * RIGHT means allocating new sections towards the end of volume.
- * LEFT means the opposite direction.
- */
-enum {
-	ALLOC_RIGHT = 0,
-	ALLOC_LEFT
-};
-
-/*
  * In the victim_sel_policy->alloc_mode, there are three block allocation modes.
  * LFS writes data sequentially with cleaning operations.
  * SSR (Slack Space Recycle) reuses obsolete space without cleaning operations.
@@ -360,11 +347,12 @@ static inline unsigned int get_ckpt_valid_blocks(struct f2fs_sb_info *sbi,
 				unsigned int segno, bool use_section)
 {
 	if (use_section && __is_large_section(sbi)) {
-		unsigned int start_segno = START_SEGNO(segno);
+		unsigned int secno = GET_SEC_FROM_SEG(sbi, segno);
+		unsigned int start_segno = GET_SEG_FROM_SEC(sbi, secno);
 		unsigned int blocks = 0;
 		int i;
 
-		for (i = 0; i < sbi->segs_per_sec; i++, start_segno++) {
+		for (i = 0; i < SEGS_PER_SEC(sbi); i++, start_segno++) {
 			struct seg_entry *se = get_seg_entry(sbi, start_segno);
 
 			blocks += se->ckpt_valid_blocks;
@@ -449,7 +437,7 @@ static inline void __set_free(struct f2fs_sb_info *sbi, unsigned int segno)
 	free_i->free_segments++;
 
 	next = find_next_bit(free_i->free_segmap,
-			start_segno + sbi->segs_per_sec, start_segno);
+			start_segno + SEGS_PER_SEC(sbi), start_segno);
 	if (next >= start_segno + usable_segs) {
 		clear_bit(secno, free_i->free_secmap);
 		free_i->free_sections++;
@@ -485,7 +473,7 @@ static inline void __set_test_and_free(struct f2fs_sb_info *sbi,
 		if (!inmem && IS_CURSEC(sbi, secno))
 			goto skip_free;
 		next = find_next_bit(free_i->free_segmap,
-				start_segno + sbi->segs_per_sec, start_segno);
+				start_segno + SEGS_PER_SEC(sbi), start_segno);
 		if (next >= start_segno + usable_segs) {
 			if (test_and_clear_bit(secno, free_i->free_secmap))
 				free_i->free_sections++;
@@ -792,10 +780,10 @@ static inline int check_block_count(struct f2fs_sb_info *sbi,
 		return -EFSCORRUPTED;
 	}
 
-	if (usable_blks_per_seg < sbi->blocks_per_seg)
+	if (usable_blks_per_seg < BLKS_PER_SEG(sbi))
 		f2fs_bug_on(sbi, find_next_bit_le(&raw_sit->valid_map,
-				sbi->blocks_per_seg,
-				usable_blks_per_seg) != sbi->blocks_per_seg);
+				BLKS_PER_SEG(sbi),
+				usable_blks_per_seg) != BLKS_PER_SEG(sbi));
 
 	/* check segment usage, and check boundary of a given segment number */
 	if (unlikely(GET_SIT_VBLOCKS(raw_sit) > usable_blks_per_seg
@@ -914,9 +902,9 @@ static inline int nr_pages_to_skip(struct f2fs_sb_info *sbi, int type)
 		return 0;
 
 	if (type == DATA)
-		return sbi->blocks_per_seg;
+		return BLKS_PER_SEG(sbi);
 	else if (type == NODE)
-		return 8 * sbi->blocks_per_seg;
+		return SEGS_TO_BLKS(sbi, 8);
 	else if (type == META)
 		return 8 * BIO_MAX_VECS;
 	else
@@ -968,3 +956,13 @@ wake_up:
 	dcc->discard_wake = true;
 	wake_up_interruptible_all(&dcc->discard_wait_queue);
 }
+
+static inline unsigned int first_zoned_segno(struct f2fs_sb_info *sbi)
+{
+	int devi;
+
+	for (devi = 0; devi < sbi->s_ndevs; devi++)
+		if (bdev_is_zoned(FDEV(devi).bdev))
+			return GET_SEGNO(sbi, FDEV(devi).start_blk);
+	return 0;
+}
diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index 4ba613ed3a..1f1b3647a9 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -44,41 +44,53 @@ static struct kmem_cache *f2fs_inode_cachep;
 #ifdef CONFIG_F2FS_FAULT_INJECTION
 
 const char *f2fs_fault_name[FAULT_MAX] = {
-	[FAULT_KMALLOC]		= "kmalloc",
-	[FAULT_KVMALLOC]	= "kvmalloc",
-	[FAULT_PAGE_ALLOC]	= "page alloc",
-	[FAULT_PAGE_GET]	= "page get",
-	[FAULT_ALLOC_NID]	= "alloc nid",
-	[FAULT_ORPHAN]		= "orphan",
-	[FAULT_BLOCK]		= "no more block",
-	[FAULT_DIR_DEPTH]	= "too big dir depth",
-	[FAULT_EVICT_INODE]	= "evict_inode fail",
-	[FAULT_TRUNCATE]	= "truncate fail",
-	[FAULT_READ_IO]		= "read IO error",
-	[FAULT_CHECKPOINT]	= "checkpoint error",
-	[FAULT_DISCARD]		= "discard error",
-	[FAULT_WRITE_IO]	= "write IO error",
-	[FAULT_SLAB_ALLOC]	= "slab alloc",
-	[FAULT_DQUOT_INIT]	= "dquot initialize",
-	[FAULT_LOCK_OP]		= "lock_op",
-	[FAULT_BLKADDR]		= "invalid blkaddr",
+	[FAULT_KMALLOC]			= "kmalloc",
+	[FAULT_KVMALLOC]		= "kvmalloc",
+	[FAULT_PAGE_ALLOC]		= "page alloc",
+	[FAULT_PAGE_GET]		= "page get",
+	[FAULT_ALLOC_NID]		= "alloc nid",
+	[FAULT_ORPHAN]			= "orphan",
+	[FAULT_BLOCK]			= "no more block",
+	[FAULT_DIR_DEPTH]		= "too big dir depth",
+	[FAULT_EVICT_INODE]		= "evict_inode fail",
+	[FAULT_TRUNCATE]		= "truncate fail",
+	[FAULT_READ_IO]			= "read IO error",
+	[FAULT_CHECKPOINT]		= "checkpoint error",
+	[FAULT_DISCARD]			= "discard error",
+	[FAULT_WRITE_IO]		= "write IO error",
+	[FAULT_SLAB_ALLOC]		= "slab alloc",
+	[FAULT_DQUOT_INIT]		= "dquot initialize",
+	[FAULT_LOCK_OP]			= "lock_op",
+	[FAULT_BLKADDR_VALIDITY]	= "invalid blkaddr",
+	[FAULT_BLKADDR_CONSISTENCE]	= "inconsistent blkaddr",
+	[FAULT_NO_SEGMENT]		= "no free segment",
 };
 
-void f2fs_build_fault_attr(struct f2fs_sb_info *sbi, unsigned int rate,
-							unsigned int type)
+int f2fs_build_fault_attr(struct f2fs_sb_info *sbi, unsigned long rate,
+							unsigned long type)
 {
 	struct f2fs_fault_info *ffi = &F2FS_OPTION(sbi).fault_info;
 
 	if (rate) {
+		if (rate > INT_MAX)
+			return -EINVAL;
 		atomic_set(&ffi->inject_ops, 0);
-		ffi->inject_rate = rate;
+		ffi->inject_rate = (int)rate;
 	}
 
-	if (type)
-		ffi->inject_type = type;
+	if (type) {
+		if (type >= BIT(FAULT_MAX))
+			return -EINVAL;
+		ffi->inject_type = (unsigned int)type;
+	}
 
 	if (!rate && !type)
 		memset(ffi, 0, sizeof(struct f2fs_fault_info));
+	else
+		f2fs_info(sbi,
+			"build fault injection attr: rate: %lu, type: 0x%lx",
+								rate, type);
+	return 0;
 }
 #endif
 
@@ -137,7 +149,6 @@ enum {
 	Opt_resgid,
 	Opt_resuid,
 	Opt_mode,
-	Opt_io_size_bits,
 	Opt_fault_injection,
 	Opt_fault_type,
 	Opt_lazytime,
@@ -216,7 +227,6 @@ static match_table_t f2fs_tokens = {
 	{Opt_resgid, "resgid=%u"},
 	{Opt_resuid, "resuid=%u"},
 	{Opt_mode, "mode=%s"},
-	{Opt_io_size_bits, "io_bits=%u"},
 	{Opt_fault_injection, "fault_injection=%u"},
 	{Opt_fault_type, "fault_type=%u"},
 	{Opt_lazytime, "lazytime"},
@@ -263,7 +273,8 @@ static match_table_t f2fs_tokens = {
 	{Opt_err, NULL},
 };
 
-void f2fs_printk(struct f2fs_sb_info *sbi, const char *fmt, ...)
+void f2fs_printk(struct f2fs_sb_info *sbi, bool limit_rate,
+						const char *fmt, ...)
 {
 	struct va_format vaf;
 	va_list args;
@@ -274,8 +285,12 @@ void f2fs_printk(struct f2fs_sb_info *sbi, const char *fmt, ...)
 	level = printk_get_level(fmt);
 	vaf.fmt = printk_skip_level(fmt);
 	vaf.va = &args;
-	printk("%c%cF2FS-fs (%s): %pV\n",
-	       KERN_SOH_ASCII, level, sbi->sb->s_id, &vaf);
+	if (limit_rate)
+		printk_ratelimited("%c%cF2FS-fs (%s): %pV\n",
+			KERN_SOH_ASCII, level, sbi->sb->s_id, &vaf);
+	else
+		printk("%c%cF2FS-fs (%s): %pV\n",
+			KERN_SOH_ASCII, level, sbi->sb->s_id, &vaf);
 
 	va_end(args);
 }
@@ -343,46 +358,6 @@ static inline void limit_reserve_root(struct f2fs_sb_info *sbi)
 					   F2FS_OPTION(sbi).s_resgid));
 }
 
-static inline int adjust_reserved_segment(struct f2fs_sb_info *sbi)
-{
-	unsigned int sec_blks = sbi->blocks_per_seg * sbi->segs_per_sec;
-	unsigned int avg_vblocks;
-	unsigned int wanted_reserved_segments;
-	block_t avail_user_block_count;
-
-	if (!F2FS_IO_ALIGNED(sbi))
-		return 0;
-
-	/* average valid block count in section in worst case */
-	avg_vblocks = sec_blks / F2FS_IO_SIZE(sbi);
-
-	/*
-	 * we need enough free space when migrating one section in worst case
-	 */
-	wanted_reserved_segments = (F2FS_IO_SIZE(sbi) / avg_vblocks) *
-						reserved_segments(sbi);
-	wanted_reserved_segments -= reserved_segments(sbi);
-
-	avail_user_block_count = sbi->user_block_count -
-				sbi->current_reserved_blocks -
-				F2FS_OPTION(sbi).root_reserved_blocks;
-
-	if (wanted_reserved_segments * sbi->blocks_per_seg >
-					avail_user_block_count) {
-		f2fs_err(sbi, "IO align feature can't grab additional reserved segment: %u, available segments: %u",
-			wanted_reserved_segments,
-			avail_user_block_count >> sbi->log_blocks_per_seg);
-		return -ENOSPC;
-	}
-
-	SM_I(sbi)->additional_reserved_segments = wanted_reserved_segments;
-
-	f2fs_info(sbi, "IO align feature needs additional reserved segment: %u",
-			 wanted_reserved_segments);
-
-	return 0;
-}
-
 static inline void adjust_unusable_cap_perc(struct f2fs_sb_info *sbi)
 {
 	if (!F2FS_OPTION(sbi).unusable_cap_perc)
@@ -769,10 +744,8 @@ static int parse_options(struct super_block *sb, char *options, bool is_remount)
 			clear_opt(sbi, DISCARD);
 			break;
 		case Opt_noheap:
-			set_opt(sbi, NOHEAP);
-			break;
 		case Opt_heap:
-			clear_opt(sbi, NOHEAP);
+			f2fs_warn(sbi, "heap/no_heap options were deprecated");
 			break;
 #ifdef CONFIG_F2FS_FS_XATTR
 		case Opt_user_xattr:
@@ -919,28 +892,21 @@ static int parse_options(struct super_block *sb, char *options, bool is_remount)
 			}
 			kfree(name);
 			break;
-		case Opt_io_size_bits:
-			if (args->from && match_int(args, &arg))
-				return -EINVAL;
-			if (arg <= 0 || arg > __ilog2_u32(BIO_MAX_VECS)) {
-				f2fs_warn(sbi, "Not support %ld, larger than %d",
-					BIT(arg), BIO_MAX_VECS);
-				return -EINVAL;
-			}
-			F2FS_OPTION(sbi).write_io_size_bits = arg;
-			break;
 #ifdef CONFIG_F2FS_FAULT_INJECTION
 		case Opt_fault_injection:
 			if (args->from && match_int(args, &arg))
 				return -EINVAL;
-			f2fs_build_fault_attr(sbi, arg, F2FS_ALL_FAULT_TYPE);
+			if (f2fs_build_fault_attr(sbi, arg,
+					F2FS_ALL_FAULT_TYPE))
+				return -EINVAL;
 			set_opt(sbi, FAULT_INJECTION);
 			break;
 
 		case Opt_fault_type:
 			if (args->from && match_int(args, &arg))
 				return -EINVAL;
-			f2fs_build_fault_attr(sbi, 0, arg);
+			if (f2fs_build_fault_attr(sbi, 0, arg))
+				return -EINVAL;
 			set_opt(sbi, FAULT_INJECTION);
 			break;
 #else
@@ -1398,12 +1364,6 @@ default_check:
 	}
 #endif
 
-	if (F2FS_IO_SIZE_BITS(sbi) && !f2fs_lfs_mode(sbi)) {
-		f2fs_err(sbi, "Should set mode=lfs with %luKB-sized IO",
-			 F2FS_IO_SIZE_KB(sbi));
-		return -EINVAL;
-	}
-
 	if (test_opt(sbi, INLINE_XATTR_SIZE)) {
 		int min_size, max_size;
 
@@ -1611,7 +1571,7 @@ static void destroy_device_list(struct f2fs_sb_info *sbi)
 
 	for (i = 0; i < sbi->s_ndevs; i++) {
 		if (i > 0)
-			bdev_release(FDEV(i).bdev_handle);
+			bdev_fput(FDEV(i).bdev_file);
 #ifdef CONFIG_BLK_DEV_ZONED
 		kvfree(FDEV(i).blkz_seq);
 #endif
@@ -1724,7 +1684,6 @@ static void f2fs_put_super(struct super_block *sb)
 
 	f2fs_destroy_page_array_cache(sbi);
 	f2fs_destroy_xattr_caches(sbi);
-	mempool_destroy(sbi->write_io_dummy);
 #ifdef CONFIG_QUOTA
 	for (i = 0; i < MAXQUOTAS; i++)
 		kfree(F2FS_OPTION(sbi).s_qf_names[i]);
@@ -2015,10 +1974,6 @@ static int f2fs_show_options(struct seq_file *seq, struct dentry *root)
 	} else {
 		seq_puts(seq, ",nodiscard");
 	}
-	if (test_opt(sbi, NOHEAP))
-		seq_puts(seq, ",no_heap");
-	else
-		seq_puts(seq, ",heap");
 #ifdef CONFIG_F2FS_FS_XATTR
 	if (test_opt(sbi, XATTR_USER))
 		seq_puts(seq, ",user_xattr");
@@ -2084,9 +2039,6 @@ static int f2fs_show_options(struct seq_file *seq, struct dentry *root)
 					F2FS_OPTION(sbi).s_resuid),
 				from_kgid_munged(&init_user_ns,
 					F2FS_OPTION(sbi).s_resgid));
-	if (F2FS_IO_SIZE_BITS(sbi))
-		seq_printf(seq, ",io_bits=%u",
-				F2FS_OPTION(sbi).write_io_size_bits);
 #ifdef CONFIG_F2FS_FAULT_INJECTION
 	if (test_opt(sbi, FAULT_INJECTION)) {
 		seq_printf(seq, ",fault_injection=%u",
@@ -2193,12 +2145,9 @@ static void default_options(struct f2fs_sb_info *sbi, bool remount)
 	F2FS_OPTION(sbi).memory_mode = MEMORY_MODE_NORMAL;
 	F2FS_OPTION(sbi).errors = MOUNT_ERRORS_CONTINUE;
 
-	sbi->sb->s_flags &= ~SB_INLINECRYPT;
-
 	set_opt(sbi, INLINE_XATTR);
 	set_opt(sbi, INLINE_DATA);
 	set_opt(sbi, INLINE_DENTRY);
-	set_opt(sbi, NOHEAP);
 	set_opt(sbi, MERGE_CHECKPOINT);
 	F2FS_OPTION(sbi).unusable_cap = 0;
 	sbi->sb->s_flags |= SB_LAZYTIME;
@@ -2253,6 +2202,7 @@ static int f2fs_disable_checkpoint(struct f2fs_sb_info *sbi)
 			.init_gc_type = FG_GC,
 			.should_migrate_blocks = false,
 			.err_gc_skipped = true,
+			.no_bg_gc = true,
 			.nr_free_secs = 1 };
 
 		f2fs_down_write(&sbi->gc_lock);
@@ -2338,7 +2288,6 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
 	bool no_read_extent_cache = !test_opt(sbi, READ_EXTENT_CACHE);
 	bool no_age_extent_cache = !test_opt(sbi, AGE_EXTENT_CACHE);
 	bool enable_checkpoint = !test_opt(sbi, DISABLE_CHECKPOINT);
-	bool no_io_align = !F2FS_IO_ALIGNED(sbi);
 	bool no_atgc = !test_opt(sbi, ATGC);
 	bool no_discard = !test_opt(sbi, DISCARD);
 	bool no_compress_cache = !test_opt(sbi, COMPRESS_CACHE);
@@ -2388,6 +2337,17 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
 	if (err)
 		goto restore_opts;
 
+#ifdef CONFIG_BLK_DEV_ZONED
+	if (f2fs_sb_has_blkzoned(sbi) &&
+		sbi->max_open_zones < F2FS_OPTION(sbi).active_logs) {
+		f2fs_err(sbi,
+			"zoned: max open zones %u is too small, need at least %u open zones",
+				 sbi->max_open_zones, F2FS_OPTION(sbi).active_logs);
+		err = -EINVAL;
+		goto restore_opts;
+	}
+#endif
+
 	/* flush outstanding errors before changing fs state */
 	flush_work(&sbi->s_error_work);
 
@@ -2446,12 +2406,6 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
 		goto restore_opts;
 	}
 
-	if (no_io_align == !!F2FS_IO_ALIGNED(sbi)) {
-		err = -EINVAL;
-		f2fs_warn(sbi, "switch io_bits option is not allowed");
-		goto restore_opts;
-	}
-
 	if (no_compress_cache == !!test_opt(sbi, COMPRESS_CACHE)) {
 		err = -EINVAL;
 		f2fs_warn(sbi, "switch compress_cache option is not allowed");
@@ -2615,6 +2569,11 @@ restore_opts:
 	return err;
 }
 
+static void f2fs_shutdown(struct super_block *sb)
+{
+	f2fs_do_shutdown(F2FS_SB(sb), F2FS_GOING_DOWN_NOSYNC, false);
+}
+
 #ifdef CONFIG_QUOTA
 static bool f2fs_need_recovery(struct f2fs_sb_info *sbi)
 {
@@ -3214,6 +3173,7 @@ static const struct super_operations f2fs_sops = {
 	.unfreeze_fs	= f2fs_unfreeze,
 	.statfs		= f2fs_statfs,
 	.remount_fs	= f2fs_remount,
+	.shutdown	= f2fs_shutdown,
 };
 
 #ifdef CONFIG_FS_ENCRYPTION
@@ -3509,7 +3469,7 @@ static int sanity_check_raw_super(struct f2fs_sb_info *sbi,
 		}
 	}
 
-	/* Currently, support only 4KB block size */
+	/* only support block_size equals to PAGE_SIZE */
 	if (le32_to_cpu(raw_super->log_blocksize) != F2FS_BLKSIZE_BITS) {
 		f2fs_info(sbi, "Invalid log_blocksize (%u), supports only %u",
 			  le32_to_cpu(raw_super->log_blocksize),
@@ -3712,7 +3672,7 @@ int f2fs_sanity_check_ckpt(struct f2fs_sb_info *sbi)
 	}
 
 	main_segs = le32_to_cpu(raw_super->segment_count_main);
-	blocks_per_seg = sbi->blocks_per_seg;
+	blocks_per_seg = BLKS_PER_SEG(sbi);
 
 	for (i = 0; i < NR_CURSEG_NODE_TYPE; i++) {
 		if (le32_to_cpu(ckpt->cur_node_segno[i]) >= main_segs ||
@@ -3824,9 +3784,9 @@ static void init_sb_info(struct f2fs_sb_info *sbi)
 	sbi->segs_per_sec = le32_to_cpu(raw_super->segs_per_sec);
 	sbi->secs_per_zone = le32_to_cpu(raw_super->secs_per_zone);
 	sbi->total_sections = le32_to_cpu(raw_super->section_count);
-	sbi->total_node_count =
-		(le32_to_cpu(raw_super->segment_count_nat) / 2)
-			* sbi->blocks_per_seg * NAT_ENTRY_PER_BLOCK;
+	sbi->total_node_count = SEGS_TO_BLKS(sbi,
+			((le32_to_cpu(raw_super->segment_count_nat) / 2) *
+			NAT_ENTRY_PER_BLOCK));
 	F2FS_ROOT_INO(sbi) = le32_to_cpu(raw_super->root_ino);
 	F2FS_NODE_INO(sbi) = le32_to_cpu(raw_super->node_ino);
 	F2FS_META_INO(sbi) = le32_to_cpu(raw_super->meta_ino);
@@ -3835,7 +3795,7 @@ static void init_sb_info(struct f2fs_sb_info *sbi)
 	sbi->next_victim_seg[BG_GC] = NULL_SEGNO;
 	sbi->next_victim_seg[FG_GC] = NULL_SEGNO;
 	sbi->max_victim_search = DEF_MAX_VICTIM_SEARCH;
-	sbi->migration_granularity = sbi->segs_per_sec;
+	sbi->migration_granularity = SEGS_PER_SEC(sbi);
 	sbi->seq_file_ra_mul = MIN_RA_MUL;
 	sbi->max_fragment_chunk = DEF_FRAGMENT_SIZE;
 	sbi->max_fragment_hole = DEF_FRAGMENT_SIZE;
@@ -3930,11 +3890,24 @@ static int init_blkz_info(struct f2fs_sb_info *sbi, int devi)
 	sector_t nr_sectors = bdev_nr_sectors(bdev);
 	struct f2fs_report_zones_args rep_zone_arg;
 	u64 zone_sectors;
+	unsigned int max_open_zones;
 	int ret;
 
 	if (!f2fs_sb_has_blkzoned(sbi))
 		return 0;
 
+	if (bdev_is_zoned(FDEV(devi).bdev)) {
+		max_open_zones = bdev_max_open_zones(bdev);
+		if (max_open_zones && (max_open_zones < sbi->max_open_zones))
+			sbi->max_open_zones = max_open_zones;
+		if (sbi->max_open_zones < F2FS_OPTION(sbi).active_logs) {
+			f2fs_err(sbi,
+				"zoned: max open zones %u is too small, need at least %u open zones",
+				sbi->max_open_zones, F2FS_OPTION(sbi).active_logs);
+			return -EINVAL;
+		}
+	}
+
 	zone_sectors = bdev_zone_sectors(bdev);
 	if (sbi->blocks_per_blkz && sbi->blocks_per_blkz !=
 				SECTOR_TO_BLOCK(zone_sectors))
@@ -4091,7 +4064,9 @@ static void f2fs_record_stop_reason(struct f2fs_sb_info *sbi)
 
 	f2fs_up_write(&sbi->sb_lock);
 	if (err)
-		f2fs_err(sbi, "f2fs_commit_super fails to record err:%d", err);
+		f2fs_err_ratelimited(sbi,
+			"f2fs_commit_super fails to record stop_reason, err:%d",
+			err);
 }
 
 void f2fs_save_errors(struct f2fs_sb_info *sbi, unsigned char flag)
@@ -4134,8 +4109,9 @@ static void f2fs_record_errors(struct f2fs_sb_info *sbi, unsigned char error)
 
 	err = f2fs_commit_super(sbi, false);
 	if (err)
-		f2fs_err(sbi, "f2fs_commit_super fails to record errors:%u, err:%d",
-								error, err);
+		f2fs_err_ratelimited(sbi,
+			"f2fs_commit_super fails to record errors:%u, err:%d",
+			error, err);
 out_unlock:
 	f2fs_up_write(&sbi->sb_lock);
 }
@@ -4196,9 +4172,15 @@ void f2fs_handle_critical_error(struct f2fs_sb_info *sbi, unsigned char reason,
 	if (shutdown)
 		set_sbi_flag(sbi, SBI_IS_SHUTDOWN);
 
-	/* continue filesystem operators if errors=continue */
-	if (continue_fs || f2fs_readonly(sb))
+	/*
+	 * Continue filesystem operators if errors=continue. Should not set
+	 * RO by shutdown, since RO bypasses thaw_super which can hang the
+	 * system.
+	 */
+	if (continue_fs || f2fs_readonly(sb) || shutdown) {
+		f2fs_warn(sbi, "Stopped filesystem due to reason: %d", reason);
 		return;
+	}
 
 	f2fs_warn(sbi, "Remounting filesystem read-only");
 	/*
@@ -4245,10 +4227,13 @@ static int f2fs_scan_devices(struct f2fs_sb_info *sbi)
 
 	logical_blksize = bdev_logical_block_size(sbi->sb->s_bdev);
 	sbi->aligned_blksize = true;
+#ifdef CONFIG_BLK_DEV_ZONED
+	sbi->max_open_zones = UINT_MAX;
+#endif
 
 	for (i = 0; i < max_devices; i++) {
 		if (i == 0)
-			FDEV(0).bdev_handle = sbi->sb->s_bdev_handle;
+			FDEV(0).bdev_file = sbi->sb->s_bdev_file;
 		else if (!RDEV(i).path[0])
 			break;
 
@@ -4260,22 +4245,22 @@ static int f2fs_scan_devices(struct f2fs_sb_info *sbi)
 			if (i == 0) {
 				FDEV(i).start_blk = 0;
 				FDEV(i).end_blk = FDEV(i).start_blk +
-				    (FDEV(i).total_segments <<
-				    sbi->log_blocks_per_seg) - 1 +
-				    le32_to_cpu(raw_super->segment0_blkaddr);
+					SEGS_TO_BLKS(sbi,
+					FDEV(i).total_segments) - 1 +
+					le32_to_cpu(raw_super->segment0_blkaddr);
 			} else {
 				FDEV(i).start_blk = FDEV(i - 1).end_blk + 1;
 				FDEV(i).end_blk = FDEV(i).start_blk +
-					(FDEV(i).total_segments <<
-					sbi->log_blocks_per_seg) - 1;
-				FDEV(i).bdev_handle = bdev_open_by_path(
+						SEGS_TO_BLKS(sbi,
+						FDEV(i).total_segments) - 1;
+				FDEV(i).bdev_file = bdev_file_open_by_path(
 					FDEV(i).path, mode, sbi->sb, NULL);
 			}
 		}
-		if (IS_ERR(FDEV(i).bdev_handle))
-			return PTR_ERR(FDEV(i).bdev_handle);
+		if (IS_ERR(FDEV(i).bdev_file))
+			return PTR_ERR(FDEV(i).bdev_file);
 
-		FDEV(i).bdev = FDEV(i).bdev_handle->bdev;
+		FDEV(i).bdev = file_bdev(FDEV(i).bdev_file);
 		/* to release errored devices */
 		sbi->s_ndevs = i + 1;
 
@@ -4306,8 +4291,6 @@ static int f2fs_scan_devices(struct f2fs_sb_info *sbi)
 			  FDEV(i).total_segments,
 			  FDEV(i).start_blk, FDEV(i).end_blk);
 	}
-	f2fs_info(sbi,
-		  "IO Block Size: %8ld KB", F2FS_IO_SIZE_KB(sbi));
 	return 0;
 }
 
@@ -4497,7 +4480,7 @@ try_onemore:
 	sb->s_time_gran = 1;
 	sb->s_flags = (sb->s_flags & ~SB_POSIXACL) |
 		(test_opt(sbi, POSIX_ACL) ? SB_POSIXACL : 0);
-	memcpy(&sb->s_uuid, raw_super->uuid, sizeof(raw_super->uuid));
+	super_set_uuid(sb, (void *) raw_super->uuid, sizeof(raw_super->uuid));
 	sb->s_iflags |= SB_I_CGROUPWB;
 
 	/* init f2fs-specific super block info */
@@ -4520,19 +4503,10 @@ try_onemore:
 	if (err)
 		goto free_iostat;
 
-	if (F2FS_IO_ALIGNED(sbi)) {
-		sbi->write_io_dummy =
-			mempool_create_page_pool(2 * (F2FS_IO_SIZE(sbi) - 1), 0);
-		if (!sbi->write_io_dummy) {
-			err = -ENOMEM;
-			goto free_percpu;
-		}
-	}
-
 	/* init per sbi slab cache */
 	err = f2fs_init_xattr_caches(sbi);
 	if (err)
-		goto free_io_dummy;
+		goto free_percpu;
 	err = f2fs_init_page_array_cache(sbi);
 	if (err)
 		goto free_xattr_cache;
@@ -4620,10 +4594,6 @@ try_onemore:
 		goto free_nm;
 	}
 
-	err = adjust_reserved_segment(sbi);
-	if (err)
-		goto free_nm;
-
 	/* For write statistics */
 	sbi->sectors_written_start = f2fs_get_sectors_written(sbi);
 
@@ -4661,6 +4631,7 @@ try_onemore:
 		goto free_node_inode;
 	}
 
+	generic_set_sb_d_ops(sb);
 	sb->s_root = d_make_root(root); /* allocate root dentry */
 	if (!sb->s_root) {
 		err = -ENOMEM;
@@ -4749,13 +4720,20 @@ reset_checkpoint:
 	 * If the f2fs is not readonly and fsync data recovery succeeds,
 	 * check zoned block devices' write pointer consistency.
 	 */
-	if (!err && !f2fs_readonly(sb) && f2fs_sb_has_blkzoned(sbi)) {
-		err = f2fs_check_write_pointer(sbi);
-		if (err)
-			goto free_meta;
+	if (f2fs_sb_has_blkzoned(sbi) && !f2fs_readonly(sb)) {
+		int err2;
+
+		f2fs_notice(sbi, "Checking entire write pointers");
+		err2 = f2fs_check_write_pointer(sbi);
+		if (err2)
+			err = err2;
 	}
+	if (err)
+		goto free_meta;
 
-	f2fs_init_inmem_curseg(sbi);
+	err = f2fs_init_inmem_curseg(sbi);
+	if (err)
+		goto sync_free_meta;
 
 	/* f2fs_recover_fsync_data() cleared this already */
 	clear_sbi_flag(sbi, SBI_POR_DOING);
@@ -4854,8 +4832,6 @@ free_page_array_cache:
 	f2fs_destroy_page_array_cache(sbi);
 free_xattr_cache:
 	f2fs_destroy_xattr_caches(sbi);
-free_io_dummy:
-	mempool_destroy(sbi->write_io_dummy);
 free_percpu:
 	destroy_percpu_info(sbi);
 free_iostat:
@@ -4968,12 +4944,6 @@ static int __init init_f2fs_fs(void)
 {
 	int err;
 
-	if (PAGE_SIZE != F2FS_BLKSIZE) {
-		printk("F2FS not supported on PAGE_SIZE(%lu) != BLOCK_SIZE(%lu)\n",
-				PAGE_SIZE, F2FS_BLKSIZE);
-		return -EINVAL;
-	}
-
 	err = init_inodecache();
 	if (err)
 		goto fail;
diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c
index a7ec55c7bb..09d3ecfaa4 100644
--- a/fs/f2fs/sysfs.c
+++ b/fs/f2fs/sysfs.c
@@ -484,17 +484,23 @@ out:
 	if (ret < 0)
 		return ret;
 #ifdef CONFIG_F2FS_FAULT_INJECTION
-	if (a->struct_type == FAULT_INFO_TYPE && t >= BIT(FAULT_MAX))
-		return -EINVAL;
-	if (a->struct_type == FAULT_INFO_RATE && t >= UINT_MAX)
-		return -EINVAL;
+	if (a->struct_type == FAULT_INFO_TYPE) {
+		if (f2fs_build_fault_attr(sbi, 0, t))
+			return -EINVAL;
+		return count;
+	}
+	if (a->struct_type == FAULT_INFO_RATE) {
+		if (f2fs_build_fault_attr(sbi, t, 0))
+			return -EINVAL;
+		return count;
+	}
 #endif
 	if (a->struct_type == RESERVED_BLOCKS) {
 		spin_lock(&sbi->stat_lock);
 		if (t > (unsigned long)(sbi->user_block_count -
 				F2FS_OPTION(sbi).root_reserved_blocks -
-				sbi->blocks_per_seg *
-				SM_I(sbi)->additional_reserved_segments)) {
+				SEGS_TO_BLKS(sbi,
+				SM_I(sbi)->additional_reserved_segments))) {
 			spin_unlock(&sbi->stat_lock);
 			return -EINVAL;
 		}
@@ -551,7 +557,7 @@ out:
 	}
 
 	if (!strcmp(a->attr.name, "migration_granularity")) {
-		if (t == 0 || t > sbi->segs_per_sec)
+		if (t == 0 || t > SEGS_PER_SEC(sbi))
 			return -EINVAL;
 	}
 
@@ -675,6 +681,13 @@ out:
 		return count;
 	}
 
+	if (!strcmp(a->attr.name, "gc_pin_file_threshold")) {
+		if (t > MAX_GC_FAILED_PINNED_FILES)
+			return -EINVAL;
+		sbi->gc_pin_file_threshold = t;
+		return count;
+	}
+
 	if (!strcmp(a->attr.name, "gc_reclaimed_segments")) {
 		if (t != 0)
 			return -EINVAL;
@@ -1492,6 +1505,50 @@ static int __maybe_unused discard_plist_seq_show(struct seq_file *seq,
 	return 0;
 }
 
+static int __maybe_unused disk_map_seq_show(struct seq_file *seq,
+						void *offset)
+{
+	struct super_block *sb = seq->private;
+	struct f2fs_sb_info *sbi = F2FS_SB(sb);
+	int i;
+
+	seq_printf(seq, "Address Layout   : %5luB Block address (# of Segments)\n",
+					F2FS_BLKSIZE);
+	seq_printf(seq, " SB            : %12s\n", "0/1024B");
+	seq_printf(seq, " seg0_blkaddr  : 0x%010x\n", SEG0_BLKADDR(sbi));
+	seq_printf(seq, " Checkpoint    : 0x%010x (%10d)\n",
+			le32_to_cpu(F2FS_RAW_SUPER(sbi)->cp_blkaddr), 2);
+	seq_printf(seq, " SIT           : 0x%010x (%10d)\n",
+			SIT_I(sbi)->sit_base_addr,
+			le32_to_cpu(F2FS_RAW_SUPER(sbi)->segment_count_sit));
+	seq_printf(seq, " NAT           : 0x%010x (%10d)\n",
+			NM_I(sbi)->nat_blkaddr,
+			le32_to_cpu(F2FS_RAW_SUPER(sbi)->segment_count_nat));
+	seq_printf(seq, " SSA           : 0x%010x (%10d)\n",
+			SM_I(sbi)->ssa_blkaddr,
+			le32_to_cpu(F2FS_RAW_SUPER(sbi)->segment_count_ssa));
+	seq_printf(seq, " Main          : 0x%010x (%10d)\n",
+			SM_I(sbi)->main_blkaddr,
+			le32_to_cpu(F2FS_RAW_SUPER(sbi)->segment_count_main));
+	seq_printf(seq, " # of Sections : %12d\n",
+			le32_to_cpu(F2FS_RAW_SUPER(sbi)->section_count));
+	seq_printf(seq, " Segs/Sections : %12d\n",
+			SEGS_PER_SEC(sbi));
+	seq_printf(seq, " Section size  : %12d MB\n",
+			SEGS_PER_SEC(sbi) << 1);
+
+	if (!f2fs_is_multi_device(sbi))
+		return 0;
+
+	seq_puts(seq, "\nDisk Map for multi devices:\n");
+	for (i = 0; i < sbi->s_ndevs; i++)
+		seq_printf(seq, "Disk:%2d (zoned=%d): 0x%010x - 0x%010x on %s\n",
+			i, bdev_is_zoned(FDEV(i).bdev),
+			FDEV(i).start_blk, FDEV(i).end_blk,
+			FDEV(i).path);
+	return 0;
+}
+
 int __init f2fs_init_sysfs(void)
 {
 	int ret;
@@ -1573,6 +1630,8 @@ int f2fs_register_sysfs(struct f2fs_sb_info *sbi)
 				victim_bits_seq_show, sb);
 	proc_create_single_data("discard_plist_info", 0444, sbi->s_proc,
 				discard_plist_seq_show, sb);
+	proc_create_single_data("disk_map", 0444, sbi->s_proc,
+				disk_map_seq_show, sb);
 	return 0;
 put_feature_list_kobj:
 	kobject_put(&sbi->s_feature_list_kobj);
diff --git a/fs/f2fs/verity.c b/fs/f2fs/verity.c
index 4fc95f353a..f7bb0c5450 100644
--- a/fs/f2fs/verity.c
+++ b/fs/f2fs/verity.c
@@ -258,21 +258,23 @@ static struct page *f2fs_read_merkle_tree_page(struct inode *inode,
 					       pgoff_t index,
 					       unsigned long num_ra_pages)
 {
-	struct page *page;
+	struct folio *folio;
 
 	index += f2fs_verity_metadata_pos(inode) >> PAGE_SHIFT;
 
-	page = find_get_page_flags(inode->i_mapping, index, FGP_ACCESSED);
-	if (!page || !PageUptodate(page)) {
+	folio = __filemap_get_folio(inode->i_mapping, index, FGP_ACCESSED, 0);
+	if (IS_ERR(folio) || !folio_test_uptodate(folio)) {
 		DEFINE_READAHEAD(ractl, NULL, NULL, inode->i_mapping, index);
 
-		if (page)
-			put_page(page);
+		if (!IS_ERR(folio))
+			folio_put(folio);
 		else if (num_ra_pages > 1)
 			page_cache_ra_unbounded(&ractl, num_ra_pages, 0);
-		page = read_mapping_page(inode->i_mapping, index, NULL);
+		folio = read_mapping_folio(inode->i_mapping, index, NULL);
+		if (IS_ERR(folio))
+			return ERR_CAST(folio);
 	}
-	return page;
+	return folio_file_page(folio, index);
 }
 
 static int f2fs_write_merkle_tree_block(struct inode *inode, const void *buf,
diff --git a/fs/fat/cache.c b/fs/fat/cache.c
index 738e427e2d..2af424e200 100644
--- a/fs/fat/cache.c
+++ b/fs/fat/cache.c
@@ -47,7 +47,7 @@ int __init fat_cache_init(void)
 {
 	fat_cache_cachep = kmem_cache_create("fat_cache",
 				sizeof(struct fat_cache),
-				0, SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD,
+				0, SLAB_RECLAIM_ACCOUNT,
 				init_once);
 	if (fat_cache_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/fat/dir.c b/fs/fat/dir.c
index 00235b8a18..acbec5bdd5 100644
--- a/fs/fat/dir.c
+++ b/fs/fat/dir.c
@@ -269,6 +269,18 @@ enum { PARSE_INVALID = 1, PARSE_NOT_LONGNAME, PARSE_EOF, };
 /**
  * fat_parse_long - Parse extended directory entry.
  *
+ * @dir: Pointer to the inode that represents the directory.
+ * @pos: On input, contains the starting position to read from.
+ *       On output, updated with the new position.
+ * @bh: Pointer to the buffer head that may be used for reading directory
+ *	 entries. May be updated.
+ * @de: On input, points to the current directory entry.
+ *      On output, points to the next directory entry.
+ * @unicode: Pointer to a buffer where the parsed Unicode long filename will be
+ *	      stored.
+ * @nr_slots: Pointer to a variable that will store the number of longname
+ *	       slots found.
+ *
  * This function returns zero on success, negative value on error, or one of
  * the following:
  *
diff --git a/fs/fat/inode.c b/fs/fat/inode.c
index 1fac3dabf1..d9e6fbb6f2 100644
--- a/fs/fat/inode.c
+++ b/fs/fat/inode.c
@@ -787,7 +787,7 @@ static int __init fat_init_inodecache(void)
 	fat_inode_cachep = kmem_cache_create("fat_inode_cache",
 					     sizeof(struct msdos_inode_info),
 					     0, (SLAB_RECLAIM_ACCOUNT|
-						SLAB_MEM_SPREAD|SLAB_ACCOUNT),
+						SLAB_ACCOUNT),
 					     init_once);
 	if (fat_inode_cachep == NULL)
 		return -ENOMEM;
@@ -1762,6 +1762,9 @@ int fat_fill_super(struct super_block *sb, void *data, int silent, int isvfat,
 	else /* fat 16 or 12 */
 		sbi->vol_id = bpb.fat16_vol_id;
 
+	__le32 vol_id_le = cpu_to_le32(sbi->vol_id);
+	super_set_uuid(sb, (void *) &vol_id_le, sizeof(vol_id_le));
+
 	sbi->dir_per_block = sb->s_blocksize / sizeof(struct msdos_dir_entry);
 	sbi->dir_per_block_bits = ffs(sbi->dir_per_block) - 1;
 
diff --git a/fs/fcntl.c b/fs/fcntl.c
index 3ff707bf27..300e5d9ad9 100644
--- a/fs/fcntl.c
+++ b/fs/fcntl.c
@@ -27,6 +27,7 @@
 #include <linux/memfd.h>
 #include <linux/compat.h>
 #include <linux/mount.h>
+#include <linux/rw_hint.h>
 
 #include <linux/poll.h>
 #include <asm/siginfo.h>
@@ -270,6 +271,13 @@ static int f_getowner_uids(struct file *filp, unsigned long arg)
 
 static bool rw_hint_valid(u64 hint)
 {
+	BUILD_BUG_ON(WRITE_LIFE_NOT_SET != RWH_WRITE_LIFE_NOT_SET);
+	BUILD_BUG_ON(WRITE_LIFE_NONE != RWH_WRITE_LIFE_NONE);
+	BUILD_BUG_ON(WRITE_LIFE_SHORT != RWH_WRITE_LIFE_SHORT);
+	BUILD_BUG_ON(WRITE_LIFE_MEDIUM != RWH_WRITE_LIFE_MEDIUM);
+	BUILD_BUG_ON(WRITE_LIFE_LONG != RWH_WRITE_LIFE_LONG);
+	BUILD_BUG_ON(WRITE_LIFE_EXTREME != RWH_WRITE_LIFE_EXTREME);
+
 	switch (hint) {
 	case RWH_WRITE_LIFE_NOT_SET:
 	case RWH_WRITE_LIFE_NONE:
@@ -283,32 +291,56 @@ static bool rw_hint_valid(u64 hint)
 	}
 }
 
-static long fcntl_rw_hint(struct file *file, unsigned int cmd,
-			  unsigned long arg)
+static long fcntl_get_rw_hint(struct file *file, unsigned int cmd,
+			      unsigned long arg)
 {
 	struct inode *inode = file_inode(file);
 	u64 __user *argp = (u64 __user *)arg;
-	u64 hint;
+	u64 hint = READ_ONCE(inode->i_write_hint);
 
-	switch (cmd) {
-	case F_GET_RW_HINT:
-		hint = inode->i_write_hint;
-		if (copy_to_user(argp, &hint, sizeof(*argp)))
-			return -EFAULT;
-		return 0;
-	case F_SET_RW_HINT:
-		if (copy_from_user(&hint, argp, sizeof(hint)))
-			return -EFAULT;
-		if (!rw_hint_valid(hint))
-			return -EINVAL;
+	if (copy_to_user(argp, &hint, sizeof(*argp)))
+		return -EFAULT;
+	return 0;
+}
 
-		inode_lock(inode);
-		inode->i_write_hint = hint;
-		inode_unlock(inode);
-		return 0;
-	default:
+static long fcntl_set_rw_hint(struct file *file, unsigned int cmd,
+			      unsigned long arg)
+{
+	struct inode *inode = file_inode(file);
+	u64 __user *argp = (u64 __user *)arg;
+	u64 hint;
+
+	if (copy_from_user(&hint, argp, sizeof(hint)))
+		return -EFAULT;
+	if (!rw_hint_valid(hint))
 		return -EINVAL;
-	}
+
+	WRITE_ONCE(inode->i_write_hint, hint);
+
+	/*
+	 * file->f_mapping->host may differ from inode. As an example,
+	 * blkdev_open() modifies file->f_mapping.
+	 */
+	if (file->f_mapping->host != inode)
+		WRITE_ONCE(file->f_mapping->host->i_write_hint, hint);
+
+	return 0;
+}
+
+/* Is the file descriptor a dup of the file? */
+static long f_dupfd_query(int fd, struct file *filp)
+{
+	CLASS(fd_raw, f)(fd);
+
+	/*
+	 * We can do the 'fdput()' immediately, as the only thing that
+	 * matters is the pointer value which isn't changed by the fdput.
+	 *
+	 * Technically we didn't need a ref at all, and 'fdget()' was
+	 * overkill, but given our lockless file pointer lookup, the
+	 * alternatives are complicated.
+	 */
+	return f.file == filp;
 }
 
 static long do_fcntl(int fd, unsigned int cmd, unsigned long arg,
@@ -326,6 +358,9 @@ static long do_fcntl(int fd, unsigned int cmd, unsigned long arg,
 	case F_DUPFD_CLOEXEC:
 		err = f_dupfd(argi, filp, O_CLOEXEC);
 		break;
+	case F_DUPFD_QUERY:
+		err = f_dupfd_query(argi, filp);
+		break;
 	case F_GETFD:
 		err = get_close_on_exec(fd) ? FD_CLOEXEC : 0;
 		break;
@@ -414,8 +449,10 @@ static long do_fcntl(int fd, unsigned int cmd, unsigned long arg,
 		err = memfd_fcntl(filp, cmd, argi);
 		break;
 	case F_GET_RW_HINT:
+		err = fcntl_get_rw_hint(filp, cmd, arg);
+		break;
 	case F_SET_RW_HINT:
-		err = fcntl_rw_hint(filp, cmd, arg);
+		err = fcntl_set_rw_hint(filp, cmd, arg);
 		break;
 	default:
 		break;
@@ -428,6 +465,7 @@ static int check_fcntl_cmd(unsigned cmd)
 	switch (cmd) {
 	case F_DUPFD:
 	case F_DUPFD_CLOEXEC:
+	case F_DUPFD_QUERY:
 	case F_GETFD:
 	case F_SETFD:
 	case F_GETFL:
@@ -844,12 +882,6 @@ int send_sigurg(struct fown_struct *fown)
 static DEFINE_SPINLOCK(fasync_lock);
 static struct kmem_cache *fasync_cache __ro_after_init;
 
-static void fasync_free_rcu(struct rcu_head *head)
-{
-	kmem_cache_free(fasync_cache,
-			container_of(head, struct fasync_struct, fa_rcu));
-}
-
 /*
  * Remove a fasync entry. If successfully removed, return
  * positive and clear the FASYNC flag. If no entry exists,
@@ -875,7 +907,7 @@ int fasync_remove_entry(struct file *filp, struct fasync_struct **fapp)
 		write_unlock_irq(&fa->fa_lock);
 
 		*fp = fa->fa_next;
-		call_rcu(&fa->fa_rcu, fasync_free_rcu);
+		kfree_rcu(fa, fa_rcu);
 		filp->f_flags &= ~FASYNC;
 		result = 1;
 		break;
diff --git a/fs/fhandle.c b/fs/fhandle.c
index 57a12614ad..8a7f86c213 100644
--- a/fs/fhandle.c
+++ b/fs/fhandle.c
@@ -36,7 +36,7 @@ static long do_sys_name_to_handle(const struct path *path,
 	if (f_handle.handle_bytes > MAX_HANDLE_SZ)
 		return -EINVAL;
 
-	handle = kzalloc(sizeof(struct file_handle) + f_handle.handle_bytes,
+	handle = kzalloc(struct_size(handle, f_handle, f_handle.handle_bytes),
 			 GFP_KERNEL);
 	if (!handle)
 		return -ENOMEM;
@@ -71,7 +71,7 @@ static long do_sys_name_to_handle(const struct path *path,
 	/* copy the mount id */
 	if (put_user(real_mount(path->mnt)->mnt_id, mnt_id) ||
 	    copy_to_user(ufh, handle,
-			 sizeof(struct file_handle) + handle_bytes))
+			 struct_size(handle, f_handle, handle_bytes)))
 		retval = -EFAULT;
 	kfree(handle);
 	return retval;
@@ -192,7 +192,7 @@ static int handle_to_path(int mountdirfd, struct file_handle __user *ufh,
 		retval = -EINVAL;
 		goto out_err;
 	}
-	handle = kmalloc(sizeof(struct file_handle) + f_handle.handle_bytes,
+	handle = kmalloc(struct_size(handle, f_handle, f_handle.handle_bytes),
 			 GFP_KERNEL);
 	if (!handle) {
 		retval = -ENOMEM;
diff --git a/fs/file.c b/fs/file.c
index 3b683b9101..a11e59b5d6 100644
--- a/fs/file.c
+++ b/fs/file.c
@@ -271,6 +271,11 @@ static inline void __clear_open_fd(unsigned int fd, struct fdtable *fdt)
 	__clear_bit(fd / BITS_PER_LONG, fdt->full_fds_bits);
 }
 
+static inline bool fd_is_open(unsigned int fd, const struct fdtable *fdt)
+{
+	return test_bit(fd, fdt->open_fds);
+}
+
 static unsigned int count_open_files(struct fdtable *fdt)
 {
 	unsigned int size = fdt->max_fds;
@@ -481,12 +486,12 @@ struct files_struct init_files = {
 
 static unsigned int find_next_fd(struct fdtable *fdt, unsigned int start)
 {
-	unsigned int maxfd = fdt->max_fds;
+	unsigned int maxfd = fdt->max_fds; /* always multiple of BITS_PER_LONG */
 	unsigned int maxbit = maxfd / BITS_PER_LONG;
 	unsigned int bitbit = start / BITS_PER_LONG;
 
 	bitbit = find_next_zero_bit(fdt->full_fds_bits, maxbit, bitbit) * BITS_PER_LONG;
-	if (bitbit > maxfd)
+	if (bitbit >= maxfd)
 		return maxfd;
 	if (bitbit > start)
 		start = bitbit;
@@ -915,13 +920,8 @@ struct file *get_file_rcu(struct file __rcu **f)
 		struct file __rcu *file;
 
 		file = __get_file_rcu(f);
-		if (unlikely(!file))
-			return NULL;
-
-		if (unlikely(IS_ERR(file)))
-			continue;
-
-		return file;
+		if (!IS_ERR(file))
+			return file;
 	}
 }
 EXPORT_SYMBOL_GPL(get_file_rcu);
@@ -1219,12 +1219,9 @@ void set_close_on_exec(unsigned int fd, int flag)
 
 bool get_close_on_exec(unsigned int fd)
 {
-	struct files_struct *files = current->files;
-	struct fdtable *fdt;
 	bool res;
 	rcu_read_lock();
-	fdt = files_fdtable(files);
-	res = close_on_exec(fd, fdt);
+	res = close_on_exec(fd, current->files);
 	rcu_read_unlock();
 	return res;
 }
@@ -1251,6 +1248,7 @@ __releases(&files->file_lock)
 	 * tables and this condition does not arise without those.
 	 */
 	fdt = files_fdtable(files);
+	fd = array_index_nospec(fd, fdt->max_fds);
 	tofree = fdt->fd[fd];
 	if (!tofree && fd_is_open(fd, fdt))
 		goto Ebusy;
diff --git a/fs/file_table.c b/fs/file_table.c
index b991f90571..4f03beed47 100644
--- a/fs/file_table.c
+++ b/fs/file_table.c
@@ -26,7 +26,6 @@
 #include <linux/percpu_counter.h>
 #include <linux/percpu.h>
 #include <linux/task_work.h>
-#include <linux/ima.h>
 #include <linux/swap.h>
 #include <linux/kmemleak.h>
 
@@ -276,21 +275,15 @@ struct file *alloc_empty_backing_file(int flags, const struct cred *cred)
 }
 
 /**
- * alloc_file - allocate and initialize a 'struct file'
+ * file_init_path - initialize a 'struct file' based on path
  *
+ * @file: the file to set up
  * @path: the (dentry, vfsmount) pair for the new file
- * @flags: O_... flags with which the new file will be opened
  * @fop: the 'struct file_operations' for the new file
  */
-static struct file *alloc_file(const struct path *path, int flags,
-		const struct file_operations *fop)
+static void file_init_path(struct file *file, const struct path *path,
+			   const struct file_operations *fop)
 {
-	struct file *file;
-
-	file = alloc_empty_file(flags, current_cred());
-	if (IS_ERR(file))
-		return file;
-
 	file->f_path = *path;
 	file->f_inode = path->dentry->d_inode;
 	file->f_mapping = path->dentry->d_inode->i_mapping;
@@ -309,22 +302,51 @@ static struct file *alloc_file(const struct path *path, int flags,
 	file->f_op = fop;
 	if ((file->f_mode & (FMODE_READ | FMODE_WRITE)) == FMODE_READ)
 		i_readcount_inc(path->dentry->d_inode);
+}
+
+/**
+ * alloc_file - allocate and initialize a 'struct file'
+ *
+ * @path: the (dentry, vfsmount) pair for the new file
+ * @flags: O_... flags with which the new file will be opened
+ * @fop: the 'struct file_operations' for the new file
+ */
+static struct file *alloc_file(const struct path *path, int flags,
+		const struct file_operations *fop)
+{
+	struct file *file;
+
+	file = alloc_empty_file(flags, current_cred());
+	if (!IS_ERR(file))
+		file_init_path(file, path, fop);
 	return file;
 }
 
-struct file *alloc_file_pseudo(struct inode *inode, struct vfsmount *mnt,
-				const char *name, int flags,
-				const struct file_operations *fops)
+static inline int alloc_path_pseudo(const char *name, struct inode *inode,
+				    struct vfsmount *mnt, struct path *path)
 {
 	struct qstr this = QSTR_INIT(name, strlen(name));
+
+	path->dentry = d_alloc_pseudo(mnt->mnt_sb, &this);
+	if (!path->dentry)
+		return -ENOMEM;
+	path->mnt = mntget(mnt);
+	d_instantiate(path->dentry, inode);
+	return 0;
+}
+
+struct file *alloc_file_pseudo(struct inode *inode, struct vfsmount *mnt,
+			       const char *name, int flags,
+			       const struct file_operations *fops)
+{
+	int ret;
 	struct path path;
 	struct file *file;
 
-	path.dentry = d_alloc_pseudo(mnt->mnt_sb, &this);
-	if (!path.dentry)
-		return ERR_PTR(-ENOMEM);
-	path.mnt = mntget(mnt);
-	d_instantiate(path.dentry, inode);
+	ret = alloc_path_pseudo(name, inode, mnt, &path);
+	if (ret)
+		return ERR_PTR(ret);
+
 	file = alloc_file(&path, flags, fops);
 	if (IS_ERR(file)) {
 		ihold(inode);
@@ -334,6 +356,30 @@ struct file *alloc_file_pseudo(struct inode *inode, struct vfsmount *mnt,
 }
 EXPORT_SYMBOL(alloc_file_pseudo);
 
+struct file *alloc_file_pseudo_noaccount(struct inode *inode,
+					 struct vfsmount *mnt, const char *name,
+					 int flags,
+					 const struct file_operations *fops)
+{
+	int ret;
+	struct path path;
+	struct file *file;
+
+	ret = alloc_path_pseudo(name, inode, mnt, &path);
+	if (ret)
+		return ERR_PTR(ret);
+
+	file = alloc_empty_file_noaccount(flags, current_cred());
+	if (IS_ERR(file)) {
+		ihold(inode);
+		path_put(&path);
+		return file;
+	}
+	file_init_path(file, &path, fops);
+	return file;
+}
+EXPORT_SYMBOL_GPL(alloc_file_pseudo_noaccount);
+
 struct file *alloc_file_clone(struct file *base, int flags,
 				const struct file_operations *fops)
 {
@@ -367,7 +413,7 @@ static void __fput(struct file *file)
 	eventpoll_release(file);
 	locks_remove_file(file);
 
-	ima_file_free(file);
+	security_file_release(file);
 	if (unlikely(file->f_flags & FASYNC)) {
 		if (file->f_op->fasync)
 			file->f_op->fasync(-1, file, 0);
diff --git a/fs/freevxfs/vxfs_super.c b/fs/freevxfs/vxfs_super.c
index e6e2a2185e..fabe607786 100644
--- a/fs/freevxfs/vxfs_super.c
+++ b/fs/freevxfs/vxfs_super.c
@@ -17,7 +17,7 @@
 #include <linux/slab.h>
 #include <linux/stat.h>
 #include <linux/vfs.h>
-#include <linux/mount.h>
+#include <linux/fs_context.h>
 
 #include "vxfs.h"
 #include "vxfs_extern.h"
@@ -91,10 +91,10 @@ vxfs_statfs(struct dentry *dentry, struct kstatfs *bufp)
 	return 0;
 }
 
-static int vxfs_remount(struct super_block *sb, int *flags, char *data)
+static int vxfs_reconfigure(struct fs_context *fc)
 {
-	sync_filesystem(sb);
-	*flags |= SB_RDONLY;
+	sync_filesystem(fc->root->d_sb);
+	fc->sb_flags |= SB_RDONLY;
 	return 0;
 }
 
@@ -120,24 +120,24 @@ static const struct super_operations vxfs_super_ops = {
 	.evict_inode		= vxfs_evict_inode,
 	.put_super		= vxfs_put_super,
 	.statfs			= vxfs_statfs,
-	.remount_fs		= vxfs_remount,
 };
 
-static int vxfs_try_sb_magic(struct super_block *sbp, int silent,
+static int vxfs_try_sb_magic(struct super_block *sbp, struct fs_context *fc,
 		unsigned blk, __fs32 magic)
 {
 	struct buffer_head *bp;
 	struct vxfs_sb *rsbp;
 	struct vxfs_sb_info *infp = VXFS_SBI(sbp);
+	int silent = fc->sb_flags & SB_SILENT;
 	int rc = -ENOMEM;
 
 	bp = sb_bread(sbp, blk);
 	do {
 		if (!bp || !buffer_mapped(bp)) {
 			if (!silent) {
-				printk(KERN_WARNING
-					"vxfs: unable to read disk superblock at %u\n",
-					blk);
+				warnf(fc,
+				      "vxfs: unable to read disk superblock at %u",
+				      blk);
 			}
 			break;
 		}
@@ -146,9 +146,9 @@ static int vxfs_try_sb_magic(struct super_block *sbp, int silent,
 		rsbp = (struct vxfs_sb *)bp->b_data;
 		if (rsbp->vs_magic != magic) {
 			if (!silent)
-				printk(KERN_NOTICE
-					"vxfs: WRONG superblock magic %08x at %u\n",
-					rsbp->vs_magic, blk);
+				infof(fc,
+				      "vxfs: WRONG superblock magic %08x at %u",
+				      rsbp->vs_magic, blk);
 			break;
 		}
 
@@ -169,8 +169,7 @@ static int vxfs_try_sb_magic(struct super_block *sbp, int silent,
 /**
  * vxfs_fill_super - read superblock into memory and initialize filesystem
  * @sbp:		VFS superblock (to fill)
- * @dp:			fs private mount data
- * @silent:		do not complain loudly when sth is wrong
+ * @fc:			filesytem context
  *
  * Description:
  *   We are called on the first mount of a filesystem to read the
@@ -182,26 +181,27 @@ static int vxfs_try_sb_magic(struct super_block *sbp, int silent,
  * Locking:
  *   We are under @sbp->s_lock.
  */
-static int vxfs_fill_super(struct super_block *sbp, void *dp, int silent)
+static int vxfs_fill_super(struct super_block *sbp, struct fs_context *fc)
 {
 	struct vxfs_sb_info	*infp;
 	struct vxfs_sb		*rsbp;
 	u_long			bsize;
 	struct inode *root;
 	int ret = -EINVAL;
+	int silent = fc->sb_flags & SB_SILENT;
 	u32 j;
 
 	sbp->s_flags |= SB_RDONLY;
 
 	infp = kzalloc(sizeof(*infp), GFP_KERNEL);
 	if (!infp) {
-		printk(KERN_WARNING "vxfs: unable to allocate incore superblock\n");
+		warnf(fc, "vxfs: unable to allocate incore superblock");
 		return -ENOMEM;
 	}
 
 	bsize = sb_min_blocksize(sbp, BLOCK_SIZE);
 	if (!bsize) {
-		printk(KERN_WARNING "vxfs: unable to set blocksize\n");
+		warnf(fc, "vxfs: unable to set blocksize");
 		goto out;
 	}
 
@@ -210,24 +210,24 @@ static int vxfs_fill_super(struct super_block *sbp, void *dp, int silent)
 	sbp->s_time_min = 0;
 	sbp->s_time_max = U32_MAX;
 
-	if (!vxfs_try_sb_magic(sbp, silent, 1,
+	if (!vxfs_try_sb_magic(sbp, fc, 1,
 			(__force __fs32)cpu_to_le32(VXFS_SUPER_MAGIC))) {
 		/* Unixware, x86 */
 		infp->byte_order = VXFS_BO_LE;
-	} else if (!vxfs_try_sb_magic(sbp, silent, 8,
+	} else if (!vxfs_try_sb_magic(sbp, fc, 8,
 			(__force __fs32)cpu_to_be32(VXFS_SUPER_MAGIC))) {
 		/* HP-UX, parisc */
 		infp->byte_order = VXFS_BO_BE;
 	} else {
 		if (!silent)
-			printk(KERN_NOTICE "vxfs: can't find superblock.\n");
+			infof(fc, "vxfs: can't find superblock.");
 		goto out;
 	}
 
 	rsbp = infp->vsi_raw;
 	j = fs32_to_cpu(infp, rsbp->vs_version);
 	if ((j < 2 || j > 4) && !silent) {
-		printk(KERN_NOTICE "vxfs: unsupported VxFS version (%d)\n", j);
+		infof(fc, "vxfs: unsupported VxFS version (%d)", j);
 		goto out;
 	}
 
@@ -244,17 +244,17 @@ static int vxfs_fill_super(struct super_block *sbp, void *dp, int silent)
 
 	j = fs32_to_cpu(infp, rsbp->vs_bsize);
 	if (!sb_set_blocksize(sbp, j)) {
-		printk(KERN_WARNING "vxfs: unable to set final block size\n");
+		warnf(fc, "vxfs: unable to set final block size");
 		goto out;
 	}
 
 	if (vxfs_read_olt(sbp, bsize)) {
-		printk(KERN_WARNING "vxfs: unable to read olt\n");
+		warnf(fc, "vxfs: unable to read olt");
 		goto out;
 	}
 
 	if (vxfs_read_fshead(sbp)) {
-		printk(KERN_WARNING "vxfs: unable to read fshead\n");
+		warnf(fc, "vxfs: unable to read fshead");
 		goto out;
 	}
 
@@ -265,7 +265,7 @@ static int vxfs_fill_super(struct super_block *sbp, void *dp, int silent)
 	}
 	sbp->s_root = d_make_root(root);
 	if (!sbp->s_root) {
-		printk(KERN_WARNING "vxfs: unable to get root dentry.\n");
+		warnf(fc, "vxfs: unable to get root dentry.");
 		goto out_free_ilist;
 	}
 
@@ -284,18 +284,29 @@ out:
 /*
  * The usual module blurb.
  */
-static struct dentry *vxfs_mount(struct file_system_type *fs_type,
-	int flags, const char *dev_name, void *data)
+static int vxfs_get_tree(struct fs_context *fc)
 {
-	return mount_bdev(fs_type, flags, dev_name, data, vxfs_fill_super);
+	return get_tree_bdev(fc, vxfs_fill_super);
+}
+
+static const struct fs_context_operations vxfs_context_ops = {
+	.get_tree	= vxfs_get_tree,
+	.reconfigure	= vxfs_reconfigure,
+};
+
+static int vxfs_init_fs_context(struct fs_context *fc)
+{
+	fc->ops = &vxfs_context_ops;
+
+	return 0;
 }
 
 static struct file_system_type vxfs_fs_type = {
 	.owner		= THIS_MODULE,
 	.name		= "vxfs",
-	.mount		= vxfs_mount,
 	.kill_sb	= kill_block_super,
 	.fs_flags	= FS_REQUIRES_DEV,
+	.init_fs_context = vxfs_init_fs_context,
 };
 MODULE_ALIAS_FS("vxfs"); /* makes mount -t vxfs autoload the module */
 MODULE_ALIAS("vxfs");
@@ -307,7 +318,7 @@ vxfs_init(void)
 
 	vxfs_inode_cachep = kmem_cache_create_usercopy("vxfs_inode",
 			sizeof(struct vxfs_inode_info), 0,
-			SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD,
+			SLAB_RECLAIM_ACCOUNT,
 			offsetof(struct vxfs_inode_info, vii_immed.vi_immed),
 			sizeof_field(struct vxfs_inode_info,
 				vii_immed.vi_immed),
diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
index 3d84fcc471..92a5b82835 100644
--- a/fs/fs-writeback.c
+++ b/fs/fs-writeback.c
@@ -141,8 +141,32 @@ static void wb_wakeup(struct bdi_writeback *wb)
 	spin_unlock_irq(&wb->work_lock);
 }
 
-static void finish_writeback_work(struct bdi_writeback *wb,
-				  struct wb_writeback_work *work)
+/*
+ * This function is used when the first inode for this wb is marked dirty. It
+ * wakes-up the corresponding bdi thread which should then take care of the
+ * periodic background write-out of dirty inodes. Since the write-out would
+ * starts only 'dirty_writeback_interval' centisecs from now anyway, we just
+ * set up a timer which wakes the bdi thread up later.
+ *
+ * Note, we wouldn't bother setting up the timer, but this function is on the
+ * fast-path (used by '__mark_inode_dirty()'), so we save few context switches
+ * by delaying the wake-up.
+ *
+ * We have to be careful not to postpone flush work if it is scheduled for
+ * earlier. Thus we use queue_delayed_work().
+ */
+static void wb_wakeup_delayed(struct bdi_writeback *wb)
+{
+	unsigned long timeout;
+
+	timeout = msecs_to_jiffies(dirty_writeback_interval * 10);
+	spin_lock_irq(&wb->work_lock);
+	if (test_bit(WB_registered, &wb->state))
+		queue_delayed_work(bdi_wq, &wb->dwork, timeout);
+	spin_unlock_irq(&wb->work_lock);
+}
+
+static void finish_writeback_work(struct wb_writeback_work *work)
 {
 	struct wb_completion *done = work->done;
 
@@ -171,7 +195,7 @@ static void wb_queue_work(struct bdi_writeback *wb,
 		list_add_tail(&work->list, &wb->work_list);
 		mod_delayed_work(bdi_wq, &wb->dwork, 0);
 	} else
-		finish_writeback_work(wb, work);
+		finish_writeback_work(work);
 
 	spin_unlock_irq(&wb->work_lock);
 }
@@ -1536,7 +1560,8 @@ static void inode_sleep_on_writeback(struct inode *inode)
  * thread's back can have unexpected consequences.
  */
 static void requeue_inode(struct inode *inode, struct bdi_writeback *wb,
-			  struct writeback_control *wbc)
+			  struct writeback_control *wbc,
+			  unsigned long dirtied_before)
 {
 	if (inode->i_state & I_FREEING)
 		return;
@@ -1569,7 +1594,8 @@ static void requeue_inode(struct inode *inode, struct bdi_writeback *wb,
 		 * We didn't write back all the pages.  nfs_writepages()
 		 * sometimes bales out without doing anything.
 		 */
-		if (wbc->nr_to_write <= 0) {
+		if (wbc->nr_to_write <= 0 &&
+		    !inode_dirtied_after(inode, dirtied_before)) {
 			/* Slice used up. Queue for next turn. */
 			requeue_io(inode, wb);
 		} else {
@@ -1837,6 +1863,11 @@ static long writeback_sb_inodes(struct super_block *sb,
 	unsigned long start_time = jiffies;
 	long write_chunk;
 	long total_wrote = 0;  /* count both pages and inodes */
+	unsigned long dirtied_before = jiffies;
+
+	if (work->for_kupdate)
+		dirtied_before = jiffies -
+			msecs_to_jiffies(dirty_expire_interval * 10);
 
 	while (!list_empty(&wb->b_io)) {
 		struct inode *inode = wb_inode(wb->b_io.prev);
@@ -1942,7 +1973,7 @@ static long writeback_sb_inodes(struct super_block *sb,
 		spin_lock(&inode->i_lock);
 		if (!(inode->i_state & I_DIRTY_ALL))
 			total_wrote++;
-		requeue_inode(inode, tmp_wb, &wbc);
+		requeue_inode(inode, tmp_wb, &wbc, dirtied_before);
 		inode_sync_complete(inode);
 		spin_unlock(&inode->i_lock);
 
@@ -2044,6 +2075,7 @@ static long wb_writeback(struct bdi_writeback *wb,
 	struct inode *inode;
 	long progress;
 	struct blk_plug plug;
+	bool queued = false;
 
 	blk_start_plug(&plug);
 	for (;;) {
@@ -2073,21 +2105,24 @@ static long wb_writeback(struct bdi_writeback *wb,
 
 		spin_lock(&wb->list_lock);
 
-		/*
-		 * Kupdate and background works are special and we want to
-		 * include all inodes that need writing. Livelock avoidance is
-		 * handled by these works yielding to any other work so we are
-		 * safe.
-		 */
-		if (work->for_kupdate) {
-			dirtied_before = jiffies -
-				msecs_to_jiffies(dirty_expire_interval * 10);
-		} else if (work->for_background)
-			dirtied_before = jiffies;
-
 		trace_writeback_start(wb, work);
-		if (list_empty(&wb->b_io))
+		if (list_empty(&wb->b_io)) {
+			/*
+			 * Kupdate and background works are special and we want
+			 * to include all inodes that need writing. Livelock
+			 * avoidance is handled by these works yielding to any
+			 * other work so we are safe.
+			 */
+			if (work->for_kupdate) {
+				dirtied_before = jiffies -
+					msecs_to_jiffies(dirty_expire_interval *
+							 10);
+			} else if (work->for_background)
+				dirtied_before = jiffies;
+
 			queue_io(wb, work, dirtied_before);
+			queued = true;
+		}
 		if (work->sb)
 			progress = writeback_sb_inodes(work->sb, wb, work);
 		else
@@ -2102,7 +2137,7 @@ static long wb_writeback(struct bdi_writeback *wb,
 		 * mean the overall work is done. So we keep looping as long
 		 * as made some progress on cleaning pages or inodes.
 		 */
-		if (progress) {
+		if (progress || !queued) {
 			spin_unlock(&wb->list_lock);
 			continue;
 		}
@@ -2237,7 +2272,7 @@ static long wb_do_writeback(struct bdi_writeback *wb)
 	while ((work = get_next_work_item(wb)) != NULL) {
 		trace_writeback_exec(wb, work);
 		wrote += wb_writeback(wb, work);
-		finish_writeback_work(wb, work);
+		finish_writeback_work(work);
 	}
 
 	/*
@@ -2297,8 +2332,7 @@ void wb_workfn(struct work_struct *work)
 }
 
 /*
- * Start writeback of `nr_pages' pages on this bdi. If `nr_pages' is zero,
- * write back the whole world.
+ * Start writeback of all dirty pages on this bdi.
  */
 static void __wakeup_flusher_threads_bdi(struct backing_dev_info *bdi,
 					 enum wb_reason reason)
@@ -2701,7 +2735,7 @@ EXPORT_SYMBOL(writeback_inodes_sb_nr);
  */
 void writeback_inodes_sb(struct super_block *sb, enum wb_reason reason)
 {
-	return writeback_inodes_sb_nr(sb, get_nr_dirty_pages(), reason);
+	writeback_inodes_sb_nr(sb, get_nr_dirty_pages(), reason);
 }
 EXPORT_SYMBOL(writeback_inodes_sb);
 
diff --git a/fs/fs_parser.c b/fs/fs_parser.c
index edb3712dcf..a4d6ca0b89 100644
--- a/fs/fs_parser.c
+++ b/fs/fs_parser.c
@@ -83,8 +83,8 @@ static const struct fs_parameter_spec *fs_lookup_key(
 }
 
 /*
- * fs_parse - Parse a filesystem configuration parameter
- * @fc: The filesystem context to log errors through.
+ * __fs_parse - Parse a filesystem configuration parameter
+ * @log: The filesystem context to log errors through.
  * @desc: The parameter description to use.
  * @param: The parameter.
  * @result: Where to place the result of the parse
diff --git a/fs/fuse/Kconfig b/fs/fuse/Kconfig
index 038ed0b9aa..8674dbfbe5 100644
--- a/fs/fuse/Kconfig
+++ b/fs/fuse/Kconfig
@@ -52,3 +52,14 @@ config FUSE_DAX
 
 	  If you want to allow mounting a Virtio Filesystem with the "dax"
 	  option, answer Y.
+
+config FUSE_PASSTHROUGH
+	bool "FUSE passthrough operations support"
+	default y
+	depends on FUSE_FS
+	select FS_STACK
+	help
+	  This allows bypassing FUSE server by mapping specific FUSE operations
+	  to be performed directly on a backing file.
+
+	  If you want to allow passthrough operations, answer Y.
diff --git a/fs/fuse/Makefile b/fs/fuse/Makefile
index 0c48b35c05..6e0228c6d0 100644
--- a/fs/fuse/Makefile
+++ b/fs/fuse/Makefile
@@ -8,6 +8,8 @@ obj-$(CONFIG_CUSE) += cuse.o
 obj-$(CONFIG_VIRTIO_FS) += virtiofs.o
 
 fuse-y := dev.o dir.o file.o inode.o control.o xattr.o acl.o readdir.o ioctl.o
+fuse-y += iomode.o
 fuse-$(CONFIG_FUSE_DAX) += dax.o
+fuse-$(CONFIG_FUSE_PASSTHROUGH) += passthrough.o
 
 virtiofs-y := virtio_fs.o
diff --git a/fs/fuse/control.c b/fs/fuse/control.c
index 284a350064..97ac994ff7 100644
--- a/fs/fuse/control.c
+++ b/fs/fuse/control.c
@@ -174,11 +174,7 @@ static ssize_t fuse_conn_congestion_threshold_write(struct file *file,
 	if (!fc)
 		goto out;
 
-	down_read(&fc->killsb);
-	spin_lock(&fc->bg_lock);
-	fc->congestion_threshold = val;
-	spin_unlock(&fc->bg_lock);
-	up_read(&fc->killsb);
+	WRITE_ONCE(fc->congestion_threshold, val);
 	fuse_conn_put(fc);
 out:
 	return ret;
diff --git a/fs/fuse/cuse.c b/fs/fuse/cuse.c
index b6cad106c3..0b2da7b7e2 100644
--- a/fs/fuse/cuse.c
+++ b/fs/fuse/cuse.c
@@ -310,6 +310,10 @@ struct cuse_init_args {
 /**
  * cuse_process_init_reply - finish initializing CUSE channel
  *
+ * @fm: The fuse mount information containing the CUSE connection.
+ * @args: The arguments passed to the init reply.
+ * @error: The error code signifying if any error occurred during the process.
+ *
  * This function creates the character device and sets up all the
  * required data structures for it.  Please read the comment at the
  * top of this file for high level overview.
diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c
index 1a8f82f478..9eb191b5c4 100644
--- a/fs/fuse/dev.c
+++ b/fs/fuse/dev.c
@@ -1775,6 +1775,62 @@ copy_finish:
 	return err;
 }
 
+/*
+ * Resending all processing queue requests.
+ *
+ * During a FUSE daemon panics and failover, it is possible for some inflight
+ * requests to be lost and never returned. As a result, applications awaiting
+ * replies would become stuck forever. To address this, we can use notification
+ * to trigger resending of these pending requests to the FUSE daemon, ensuring
+ * they are properly processed again.
+ *
+ * Please note that this strategy is applicable only to idempotent requests or
+ * if the FUSE daemon takes careful measures to avoid processing duplicated
+ * non-idempotent requests.
+ */
+static void fuse_resend(struct fuse_conn *fc)
+{
+	struct fuse_dev *fud;
+	struct fuse_req *req, *next;
+	struct fuse_iqueue *fiq = &fc->iq;
+	LIST_HEAD(to_queue);
+	unsigned int i;
+
+	spin_lock(&fc->lock);
+	if (!fc->connected) {
+		spin_unlock(&fc->lock);
+		return;
+	}
+
+	list_for_each_entry(fud, &fc->devices, entry) {
+		struct fuse_pqueue *fpq = &fud->pq;
+
+		spin_lock(&fpq->lock);
+		for (i = 0; i < FUSE_PQ_HASH_SIZE; i++)
+			list_splice_tail_init(&fpq->processing[i], &to_queue);
+		spin_unlock(&fpq->lock);
+	}
+	spin_unlock(&fc->lock);
+
+	list_for_each_entry_safe(req, next, &to_queue, list) {
+		set_bit(FR_PENDING, &req->flags);
+		clear_bit(FR_SENT, &req->flags);
+		/* mark the request as resend request */
+		req->in.h.unique |= FUSE_UNIQUE_RESEND;
+	}
+
+	spin_lock(&fiq->lock);
+	/* iq and pq requests are both oldest to newest */
+	list_splice(&to_queue, &fiq->pending);
+	fiq->ops->wake_pending_and_unlock(fiq);
+}
+
+static int fuse_notify_resend(struct fuse_conn *fc)
+{
+	fuse_resend(fc);
+	return 0;
+}
+
 static int fuse_notify(struct fuse_conn *fc, enum fuse_notify_code code,
 		       unsigned int size, struct fuse_copy_state *cs)
 {
@@ -1800,6 +1856,9 @@ static int fuse_notify(struct fuse_conn *fc, enum fuse_notify_code code,
 	case FUSE_NOTIFY_DELETE:
 		return fuse_notify_delete(fc, size, cs);
 
+	case FUSE_NOTIFY_RESEND:
+		return fuse_notify_resend(fc);
+
 	default:
 		fuse_copy_finish(cs);
 		return -EINVAL;
@@ -2251,43 +2310,91 @@ static int fuse_device_clone(struct fuse_conn *fc, struct file *new)
 	return 0;
 }
 
-static long fuse_dev_ioctl(struct file *file, unsigned int cmd,
-			   unsigned long arg)
+static long fuse_dev_ioctl_clone(struct file *file, __u32 __user *argp)
 {
 	int res;
 	int oldfd;
 	struct fuse_dev *fud = NULL;
 	struct fd f;
 
+	if (get_user(oldfd, argp))
+		return -EFAULT;
+
+	f = fdget(oldfd);
+	if (!f.file)
+		return -EINVAL;
+
+	/*
+	 * Check against file->f_op because CUSE
+	 * uses the same ioctl handler.
+	 */
+	if (f.file->f_op == file->f_op)
+		fud = fuse_get_dev(f.file);
+
+	res = -EINVAL;
+	if (fud) {
+		mutex_lock(&fuse_mutex);
+		res = fuse_device_clone(fud->fc, file);
+		mutex_unlock(&fuse_mutex);
+	}
+
+	fdput(f);
+	return res;
+}
+
+static long fuse_dev_ioctl_backing_open(struct file *file,
+					struct fuse_backing_map __user *argp)
+{
+	struct fuse_dev *fud = fuse_get_dev(file);
+	struct fuse_backing_map map;
+
+	if (!fud)
+		return -EPERM;
+
+	if (!IS_ENABLED(CONFIG_FUSE_PASSTHROUGH))
+		return -EOPNOTSUPP;
+
+	if (copy_from_user(&map, argp, sizeof(map)))
+		return -EFAULT;
+
+	return fuse_backing_open(fud->fc, &map);
+}
+
+static long fuse_dev_ioctl_backing_close(struct file *file, __u32 __user *argp)
+{
+	struct fuse_dev *fud = fuse_get_dev(file);
+	int backing_id;
+
+	if (!fud)
+		return -EPERM;
+
+	if (!IS_ENABLED(CONFIG_FUSE_PASSTHROUGH))
+		return -EOPNOTSUPP;
+
+	if (get_user(backing_id, argp))
+		return -EFAULT;
+
+	return fuse_backing_close(fud->fc, backing_id);
+}
+
+static long fuse_dev_ioctl(struct file *file, unsigned int cmd,
+			   unsigned long arg)
+{
+	void __user *argp = (void __user *)arg;
+
 	switch (cmd) {
 	case FUSE_DEV_IOC_CLONE:
-		if (get_user(oldfd, (__u32 __user *)arg))
-			return -EFAULT;
+		return fuse_dev_ioctl_clone(file, argp);
 
-		f = fdget(oldfd);
-		if (!f.file)
-			return -EINVAL;
+	case FUSE_DEV_IOC_BACKING_OPEN:
+		return fuse_dev_ioctl_backing_open(file, argp);
+
+	case FUSE_DEV_IOC_BACKING_CLOSE:
+		return fuse_dev_ioctl_backing_close(file, argp);
 
-		/*
-		 * Check against file->f_op because CUSE
-		 * uses the same ioctl handler.
-		 */
-		if (f.file->f_op == file->f_op)
-			fud = fuse_get_dev(f.file);
-
-		res = -EINVAL;
-		if (fud) {
-			mutex_lock(&fuse_mutex);
-			res = fuse_device_clone(fud->fc, file);
-			mutex_unlock(&fuse_mutex);
-		}
-		fdput(f);
-		break;
 	default:
-		res = -ENOTTY;
-		break;
+		return -ENOTTY;
 	}
-	return res;
 }
 
 const struct file_operations fuse_dev_operations = {
diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c
index bd3b21d485..2b0d4781f3 100644
--- a/fs/fuse/dir.c
+++ b/fs/fuse/dir.c
@@ -619,7 +619,7 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry,
 	FUSE_ARGS(args);
 	struct fuse_forget_link *forget;
 	struct fuse_create_in inarg;
-	struct fuse_open_out outopen;
+	struct fuse_open_out *outopenp;
 	struct fuse_entry_out outentry;
 	struct fuse_inode *fi;
 	struct fuse_file *ff;
@@ -634,7 +634,7 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry,
 		goto out_err;
 
 	err = -ENOMEM;
-	ff = fuse_file_alloc(fm);
+	ff = fuse_file_alloc(fm, true);
 	if (!ff)
 		goto out_put_forget_req;
 
@@ -663,8 +663,10 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry,
 	args.out_numargs = 2;
 	args.out_args[0].size = sizeof(outentry);
 	args.out_args[0].value = &outentry;
-	args.out_args[1].size = sizeof(outopen);
-	args.out_args[1].value = &outopen;
+	/* Store outarg for fuse_finish_open() */
+	outopenp = &ff->args->open_outarg;
+	args.out_args[1].size = sizeof(*outopenp);
+	args.out_args[1].value = outopenp;
 
 	err = get_create_ext(&args, dir, entry, mode);
 	if (err)
@@ -680,9 +682,9 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry,
 	    fuse_invalid_attr(&outentry.attr))
 		goto out_free_ff;
 
-	ff->fh = outopen.fh;
+	ff->fh = outopenp->fh;
 	ff->nodeid = outentry.nodeid;
-	ff->open_flags = outopen.open_flags;
+	ff->open_flags = outopenp->open_flags;
 	inode = fuse_iget(dir->i_sb, outentry.nodeid, outentry.generation,
 			  &outentry.attr, ATTR_TIMEOUT(&outentry), 0);
 	if (!inode) {
@@ -696,13 +698,15 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry,
 	d_instantiate(entry, inode);
 	fuse_change_entry_timeout(entry, &outentry);
 	fuse_dir_changed(dir);
-	err = finish_open(file, entry, generic_file_open);
+	err = generic_file_open(inode, file);
+	if (!err) {
+		file->private_data = ff;
+		err = finish_open(file, entry, fuse_finish_open);
+	}
 	if (err) {
 		fi = get_fuse_inode(inode);
 		fuse_sync_release(fi, ff, flags);
 	} else {
-		file->private_data = ff;
-		fuse_finish_open(inode, file);
 		if (fm->fc->atomic_o_trunc && trunc)
 			truncate_pagecache(inode, 0);
 		else if (!(ff->open_flags & FOPEN_KEEP_CACHE))
@@ -1490,7 +1494,7 @@ static int fuse_perm_getattr(struct inode *inode, int mask)
  *
  * 1) Local access checking ('default_permissions' mount option) based
  * on file mode.  This is the plain old disk filesystem permission
- * modell.
+ * model.
  *
  * 2) "Remote" access checking, where server is responsible for
  * checking permission in each inode operation.  An exception to this
@@ -1635,7 +1639,30 @@ out_err:
 
 static int fuse_dir_open(struct inode *inode, struct file *file)
 {
-	return fuse_open_common(inode, file, true);
+	struct fuse_mount *fm = get_fuse_mount(inode);
+	int err;
+
+	if (fuse_is_bad(inode))
+		return -EIO;
+
+	err = generic_file_open(inode, file);
+	if (err)
+		return err;
+
+	err = fuse_do_open(fm, get_node_id(inode), file, true);
+	if (!err) {
+		struct fuse_file *ff = file->private_data;
+
+		/*
+		 * Keep handling FOPEN_STREAM and FOPEN_NONSEEKABLE for
+		 * directories for backward compatibility, though it's unlikely
+		 * to be useful.
+		 */
+		if (ff->open_flags & (FOPEN_STREAM | FOPEN_NONSEEKABLE))
+			nonseekable_open(inode, file);
+	}
+
+	return err;
 }
 
 static int fuse_dir_release(struct inode *inode, struct file *file)
diff --git a/fs/fuse/file.c b/fs/fuse/file.c
index b9cff9b6ca..f39456c65e 100644
--- a/fs/fuse/file.c
+++ b/fs/fuse/file.c
@@ -20,6 +20,7 @@
 #include <linux/fs.h>
 #include <linux/filelock.h>
 #include <linux/splice.h>
+#include <linux/task_io_accounting_ops.h>
 
 static int fuse_send_open(struct fuse_mount *fm, u64 nodeid,
 			  unsigned int open_flags, int opcode,
@@ -50,13 +51,7 @@ static int fuse_send_open(struct fuse_mount *fm, u64 nodeid,
 	return fuse_simple_request(fm, &args);
 }
 
-struct fuse_release_args {
-	struct fuse_args args;
-	struct fuse_release_in inarg;
-	struct inode *inode;
-};
-
-struct fuse_file *fuse_file_alloc(struct fuse_mount *fm)
+struct fuse_file *fuse_file_alloc(struct fuse_mount *fm, bool release)
 {
 	struct fuse_file *ff;
 
@@ -65,15 +60,15 @@ struct fuse_file *fuse_file_alloc(struct fuse_mount *fm)
 		return NULL;
 
 	ff->fm = fm;
-	ff->release_args = kzalloc(sizeof(*ff->release_args),
-				   GFP_KERNEL_ACCOUNT);
-	if (!ff->release_args) {
-		kfree(ff);
-		return NULL;
+	if (release) {
+		ff->args = kzalloc(sizeof(*ff->args), GFP_KERNEL_ACCOUNT);
+		if (!ff->args) {
+			kfree(ff);
+			return NULL;
+		}
 	}
 
 	INIT_LIST_HEAD(&ff->write_entry);
-	mutex_init(&ff->readdir.lock);
 	refcount_set(&ff->count, 1);
 	RB_CLEAR_NODE(&ff->polled_node);
 	init_waitqueue_head(&ff->poll_wait);
@@ -85,8 +80,7 @@ struct fuse_file *fuse_file_alloc(struct fuse_mount *fm)
 
 void fuse_file_free(struct fuse_file *ff)
 {
-	kfree(ff->release_args);
-	mutex_destroy(&ff->readdir.lock);
+	kfree(ff->args);
 	kfree(ff);
 }
 
@@ -105,14 +99,17 @@ static void fuse_release_end(struct fuse_mount *fm, struct fuse_args *args,
 	kfree(ra);
 }
 
-static void fuse_file_put(struct fuse_file *ff, bool sync, bool isdir)
+static void fuse_file_put(struct fuse_file *ff, bool sync)
 {
 	if (refcount_dec_and_test(&ff->count)) {
-		struct fuse_args *args = &ff->release_args->args;
+		struct fuse_release_args *ra = &ff->args->release_args;
+		struct fuse_args *args = (ra ? &ra->args : NULL);
 
-		if (isdir ? ff->fm->fc->no_opendir : ff->fm->fc->no_open) {
-			/* Do nothing when client does not implement 'open' */
-			fuse_release_end(ff->fm, args, 0);
+		if (ra && ra->inode)
+			fuse_file_io_release(ff, ra->inode);
+
+		if (!args) {
+			/* Do nothing when server does not implement 'open' */
 		} else if (sync) {
 			fuse_simple_request(ff->fm, args);
 			fuse_release_end(ff->fm, args, 0);
@@ -132,27 +129,31 @@ struct fuse_file *fuse_file_open(struct fuse_mount *fm, u64 nodeid,
 	struct fuse_conn *fc = fm->fc;
 	struct fuse_file *ff;
 	int opcode = isdir ? FUSE_OPENDIR : FUSE_OPEN;
+	bool open = isdir ? !fc->no_opendir : !fc->no_open;
 
-	ff = fuse_file_alloc(fm);
+	ff = fuse_file_alloc(fm, open);
 	if (!ff)
 		return ERR_PTR(-ENOMEM);
 
 	ff->fh = 0;
 	/* Default for no-open */
 	ff->open_flags = FOPEN_KEEP_CACHE | (isdir ? FOPEN_CACHE_DIR : 0);
-	if (isdir ? !fc->no_opendir : !fc->no_open) {
-		struct fuse_open_out outarg;
+	if (open) {
+		/* Store outarg for fuse_finish_open() */
+		struct fuse_open_out *outargp = &ff->args->open_outarg;
 		int err;
 
-		err = fuse_send_open(fm, nodeid, open_flags, opcode, &outarg);
+		err = fuse_send_open(fm, nodeid, open_flags, opcode, outargp);
 		if (!err) {
-			ff->fh = outarg.fh;
-			ff->open_flags = outarg.open_flags;
-
+			ff->fh = outargp->fh;
+			ff->open_flags = outargp->open_flags;
 		} else if (err != -ENOSYS) {
 			fuse_file_free(ff);
 			return ERR_PTR(err);
 		} else {
+			/* No release needed */
+			kfree(ff->args);
+			ff->args = NULL;
 			if (isdir)
 				fc->no_opendir = 1;
 			else
@@ -195,40 +196,50 @@ static void fuse_link_write_file(struct file *file)
 	spin_unlock(&fi->lock);
 }
 
-void fuse_finish_open(struct inode *inode, struct file *file)
+int fuse_finish_open(struct inode *inode, struct file *file)
 {
 	struct fuse_file *ff = file->private_data;
 	struct fuse_conn *fc = get_fuse_conn(inode);
+	int err;
+
+	err = fuse_file_io_open(file, inode);
+	if (err)
+		return err;
 
 	if (ff->open_flags & FOPEN_STREAM)
 		stream_open(inode, file);
 	else if (ff->open_flags & FOPEN_NONSEEKABLE)
 		nonseekable_open(inode, file);
 
-	if (fc->atomic_o_trunc && (file->f_flags & O_TRUNC)) {
-		struct fuse_inode *fi = get_fuse_inode(inode);
-
-		spin_lock(&fi->lock);
-		fi->attr_version = atomic64_inc_return(&fc->attr_version);
-		i_size_write(inode, 0);
-		spin_unlock(&fi->lock);
-		file_update_time(file);
-		fuse_invalidate_attr_mask(inode, FUSE_STATX_MODSIZE);
-	}
 	if ((file->f_mode & FMODE_WRITE) && fc->writeback_cache)
 		fuse_link_write_file(file);
+
+	return 0;
 }
 
-int fuse_open_common(struct inode *inode, struct file *file, bool isdir)
+static void fuse_truncate_update_attr(struct inode *inode, struct file *file)
+{
+	struct fuse_conn *fc = get_fuse_conn(inode);
+	struct fuse_inode *fi = get_fuse_inode(inode);
+
+	spin_lock(&fi->lock);
+	fi->attr_version = atomic64_inc_return(&fc->attr_version);
+	i_size_write(inode, 0);
+	spin_unlock(&fi->lock);
+	file_update_time(file);
+	fuse_invalidate_attr_mask(inode, FUSE_STATX_MODSIZE);
+}
+
+static int fuse_open(struct inode *inode, struct file *file)
 {
 	struct fuse_mount *fm = get_fuse_mount(inode);
+	struct fuse_inode *fi = get_fuse_inode(inode);
 	struct fuse_conn *fc = fm->fc;
+	struct fuse_file *ff;
 	int err;
-	bool is_wb_truncate = (file->f_flags & O_TRUNC) &&
-			  fc->atomic_o_trunc &&
-			  fc->writeback_cache;
-	bool dax_truncate = (file->f_flags & O_TRUNC) &&
-			  fc->atomic_o_trunc && FUSE_IS_DAX(inode);
+	bool is_truncate = (file->f_flags & O_TRUNC) && fc->atomic_o_trunc;
+	bool is_wb_truncate = is_truncate && fc->writeback_cache;
+	bool dax_truncate = is_truncate && FUSE_IS_DAX(inode);
 
 	if (fuse_is_bad(inode))
 		return -EIO;
@@ -250,16 +261,20 @@ int fuse_open_common(struct inode *inode, struct file *file, bool isdir)
 	if (is_wb_truncate || dax_truncate)
 		fuse_set_nowrite(inode);
 
-	err = fuse_do_open(fm, get_node_id(inode), file, isdir);
-	if (!err)
-		fuse_finish_open(inode, file);
+	err = fuse_do_open(fm, get_node_id(inode), file, false);
+	if (!err) {
+		ff = file->private_data;
+		err = fuse_finish_open(inode, file);
+		if (err)
+			fuse_sync_release(fi, ff, file->f_flags);
+		else if (is_truncate)
+			fuse_truncate_update_attr(inode, file);
+	}
 
 	if (is_wb_truncate || dax_truncate)
 		fuse_release_nowrite(inode);
 	if (!err) {
-		struct fuse_file *ff = file->private_data;
-
-		if (fc->atomic_o_trunc && (file->f_flags & O_TRUNC))
+		if (is_truncate)
 			truncate_pagecache(inode, 0);
 		else if (!(ff->open_flags & FOPEN_KEEP_CACHE))
 			invalidate_inode_pages2(inode->i_mapping);
@@ -274,10 +289,13 @@ out_inode_unlock:
 }
 
 static void fuse_prepare_release(struct fuse_inode *fi, struct fuse_file *ff,
-				 unsigned int flags, int opcode)
+				 unsigned int flags, int opcode, bool sync)
 {
 	struct fuse_conn *fc = ff->fm->fc;
-	struct fuse_release_args *ra = ff->release_args;
+	struct fuse_release_args *ra = &ff->args->release_args;
+
+	if (fuse_file_passthrough(ff))
+		fuse_passthrough_release(ff, fuse_inode_backing(fi));
 
 	/* Inode is NULL on error path of fuse_create_open() */
 	if (likely(fi)) {
@@ -292,6 +310,11 @@ static void fuse_prepare_release(struct fuse_inode *fi, struct fuse_file *ff,
 
 	wake_up_interruptible_all(&ff->poll_wait);
 
+	if (!ra)
+		return;
+
+	/* ff->args was used for open outarg */
+	memset(ff->args, 0, sizeof(*ff->args));
 	ra->inarg.fh = ff->fh;
 	ra->inarg.flags = flags;
 	ra->args.in_numargs = 1;
@@ -301,23 +324,28 @@ static void fuse_prepare_release(struct fuse_inode *fi, struct fuse_file *ff,
 	ra->args.nodeid = ff->nodeid;
 	ra->args.force = true;
 	ra->args.nocreds = true;
+
+	/*
+	 * Hold inode until release is finished.
+	 * From fuse_sync_release() the refcount is 1 and everything's
+	 * synchronous, so we are fine with not doing igrab() here.
+	 */
+	ra->inode = sync ? NULL : igrab(&fi->inode);
 }
 
 void fuse_file_release(struct inode *inode, struct fuse_file *ff,
 		       unsigned int open_flags, fl_owner_t id, bool isdir)
 {
 	struct fuse_inode *fi = get_fuse_inode(inode);
-	struct fuse_release_args *ra = ff->release_args;
+	struct fuse_release_args *ra = &ff->args->release_args;
 	int opcode = isdir ? FUSE_RELEASEDIR : FUSE_RELEASE;
 
-	fuse_prepare_release(fi, ff, open_flags, opcode);
+	fuse_prepare_release(fi, ff, open_flags, opcode, false);
 
-	if (ff->flock) {
+	if (ra && ff->flock) {
 		ra->inarg.release_flags |= FUSE_RELEASE_FLOCK_UNLOCK;
 		ra->inarg.lock_owner = fuse_lock_owner_id(ff->fm->fc, id);
 	}
-	/* Hold inode until release is finished */
-	ra->inode = igrab(inode);
 
 	/*
 	 * Normally this will send the RELEASE request, however if
@@ -328,7 +356,7 @@ void fuse_file_release(struct inode *inode, struct fuse_file *ff,
 	 * synchronous RELEASE is allowed (and desirable) in this case
 	 * because the server can be trusted not to screw up.
 	 */
-	fuse_file_put(ff, ff->fm->fc->destroy, isdir);
+	fuse_file_put(ff, ff->fm->fc->destroy);
 }
 
 void fuse_release_common(struct file *file, bool isdir)
@@ -337,11 +365,6 @@ void fuse_release_common(struct file *file, bool isdir)
 			  (fl_owner_t) file, isdir);
 }
 
-static int fuse_open(struct inode *inode, struct file *file)
-{
-	return fuse_open_common(inode, file, false);
-}
-
 static int fuse_release(struct inode *inode, struct file *file)
 {
 	struct fuse_conn *fc = get_fuse_conn(inode);
@@ -363,12 +386,8 @@ void fuse_sync_release(struct fuse_inode *fi, struct fuse_file *ff,
 		       unsigned int flags)
 {
 	WARN_ON(refcount_read(&ff->count) > 1);
-	fuse_prepare_release(fi, ff, flags, FUSE_RELEASE);
-	/*
-	 * iput(NULL) is a no-op and since the refcount is 1 and everything's
-	 * synchronous, we are fine with not doing igrab() here"
-	 */
-	fuse_file_put(ff, true, false);
+	fuse_prepare_release(fi, ff, flags, FUSE_RELEASE, true);
+	fuse_file_put(ff, true);
 }
 EXPORT_SYMBOL_GPL(fuse_sync_release);
 
@@ -634,7 +653,8 @@ static void fuse_release_user_pages(struct fuse_args_pages *ap,
 	for (i = 0; i < ap->num_pages; i++) {
 		if (should_dirty)
 			set_page_dirty_lock(ap->pages[i]);
-		put_page(ap->pages[i]);
+		if (ap->args.is_pinned)
+			unpin_user_page(ap->pages[i]);
 	}
 }
 
@@ -915,17 +935,13 @@ static void fuse_readpages_end(struct fuse_mount *fm, struct fuse_args *args,
 	}
 
 	for (i = 0; i < ap->num_pages; i++) {
-		struct page *page = ap->pages[i];
+		struct folio *folio = page_folio(ap->pages[i]);
 
-		if (!err)
-			SetPageUptodate(page);
-		else
-			SetPageError(page);
-		unlock_page(page);
-		put_page(page);
+		folio_end_read(folio, !err);
+		folio_put(folio);
 	}
 	if (ia->ff)
-		fuse_file_put(ia->ff, false, false);
+		fuse_file_put(ia->ff, false);
 
 	fuse_io_free(ia);
 }
@@ -1299,13 +1315,93 @@ static ssize_t fuse_perform_write(struct kiocb *iocb, struct iov_iter *ii)
 	return res;
 }
 
+static bool fuse_io_past_eof(struct kiocb *iocb, struct iov_iter *iter)
+{
+	struct inode *inode = file_inode(iocb->ki_filp);
+
+	return iocb->ki_pos + iov_iter_count(iter) > i_size_read(inode);
+}
+
+/*
+ * @return true if an exclusive lock for direct IO writes is needed
+ */
+static bool fuse_dio_wr_exclusive_lock(struct kiocb *iocb, struct iov_iter *from)
+{
+	struct file *file = iocb->ki_filp;
+	struct fuse_file *ff = file->private_data;
+	struct inode *inode = file_inode(iocb->ki_filp);
+	struct fuse_inode *fi = get_fuse_inode(inode);
+
+	/* Server side has to advise that it supports parallel dio writes. */
+	if (!(ff->open_flags & FOPEN_PARALLEL_DIRECT_WRITES))
+		return true;
+
+	/*
+	 * Append will need to know the eventual EOF - always needs an
+	 * exclusive lock.
+	 */
+	if (iocb->ki_flags & IOCB_APPEND)
+		return true;
+
+	/* shared locks are not allowed with parallel page cache IO */
+	if (test_bit(FUSE_I_CACHE_IO_MODE, &fi->state))
+		return false;
+
+	/* Parallel dio beyond EOF is not supported, at least for now. */
+	if (fuse_io_past_eof(iocb, from))
+		return true;
+
+	return false;
+}
+
+static void fuse_dio_lock(struct kiocb *iocb, struct iov_iter *from,
+			  bool *exclusive)
+{
+	struct inode *inode = file_inode(iocb->ki_filp);
+	struct fuse_inode *fi = get_fuse_inode(inode);
+
+	*exclusive = fuse_dio_wr_exclusive_lock(iocb, from);
+	if (*exclusive) {
+		inode_lock(inode);
+	} else {
+		inode_lock_shared(inode);
+		/*
+		 * New parallal dio allowed only if inode is not in caching
+		 * mode and denies new opens in caching mode. This check
+		 * should be performed only after taking shared inode lock.
+		 * Previous past eof check was without inode lock and might
+		 * have raced, so check it again.
+		 */
+		if (fuse_io_past_eof(iocb, from) ||
+		    fuse_inode_uncached_io_start(fi, NULL) != 0) {
+			inode_unlock_shared(inode);
+			inode_lock(inode);
+			*exclusive = true;
+		}
+	}
+}
+
+static void fuse_dio_unlock(struct kiocb *iocb, bool exclusive)
+{
+	struct inode *inode = file_inode(iocb->ki_filp);
+	struct fuse_inode *fi = get_fuse_inode(inode);
+
+	if (exclusive) {
+		inode_unlock(inode);
+	} else {
+		/* Allow opens in caching mode after last parallel dio end */
+		fuse_inode_uncached_io_end(fi);
+		inode_unlock_shared(inode);
+	}
+}
+
 static ssize_t fuse_cache_write_iter(struct kiocb *iocb, struct iov_iter *from)
 {
 	struct file *file = iocb->ki_filp;
 	struct address_space *mapping = file->f_mapping;
 	ssize_t written = 0;
 	struct inode *inode = mapping->host;
-	ssize_t err;
+	ssize_t err, count;
 	struct fuse_conn *fc = get_fuse_conn(inode);
 
 	if (fc->writeback_cache) {
@@ -1327,10 +1423,12 @@ static ssize_t fuse_cache_write_iter(struct kiocb *iocb, struct iov_iter *from)
 writethrough:
 	inode_lock(inode);
 
-	err = generic_write_checks(iocb, from);
+	err = count = generic_write_checks(iocb, from);
 	if (err <= 0)
 		goto out;
 
+	task_io_account_write(count);
+
 	err = file_remove_privs(file);
 	if (err)
 		goto out;
@@ -1392,10 +1490,13 @@ static int fuse_get_user_pages(struct fuse_args_pages *ap, struct iov_iter *ii,
 	while (nbytes < *nbytesp && ap->num_pages < max_pages) {
 		unsigned npages;
 		size_t start;
-		ret = iov_iter_get_pages2(ii, &ap->pages[ap->num_pages],
-					*nbytesp - nbytes,
-					max_pages - ap->num_pages,
-					&start);
+		struct page **pt_pages;
+
+		pt_pages = &ap->pages[ap->num_pages];
+		ret = iov_iter_extract_pages(ii, &pt_pages,
+					     *nbytesp - nbytes,
+					     max_pages - ap->num_pages,
+					     0, &start);
 		if (ret < 0)
 			break;
 
@@ -1412,6 +1513,7 @@ static int fuse_get_user_pages(struct fuse_args_pages *ap, struct iov_iter *ii,
 			(PAGE_SIZE - ret) & (PAGE_SIZE - 1);
 	}
 
+	ap->args.is_pinned = iov_iter_extract_will_pin(ii);
 	ap->args.user_pages = true;
 	if (write)
 		ap->args.in_pages = true;
@@ -1558,51 +1660,17 @@ static ssize_t fuse_direct_read_iter(struct kiocb *iocb, struct iov_iter *to)
 	return res;
 }
 
-static bool fuse_direct_write_extending_i_size(struct kiocb *iocb,
-					       struct iov_iter *iter)
-{
-	struct inode *inode = file_inode(iocb->ki_filp);
-
-	return iocb->ki_pos + iov_iter_count(iter) > i_size_read(inode);
-}
-
 static ssize_t fuse_direct_write_iter(struct kiocb *iocb, struct iov_iter *from)
 {
 	struct inode *inode = file_inode(iocb->ki_filp);
-	struct file *file = iocb->ki_filp;
-	struct fuse_file *ff = file->private_data;
 	struct fuse_io_priv io = FUSE_IO_PRIV_SYNC(iocb);
 	ssize_t res;
-	bool exclusive_lock =
-		!(ff->open_flags & FOPEN_PARALLEL_DIRECT_WRITES) ||
-		get_fuse_conn(inode)->direct_io_allow_mmap ||
-		iocb->ki_flags & IOCB_APPEND ||
-		fuse_direct_write_extending_i_size(iocb, from);
-
-	/*
-	 * Take exclusive lock if
-	 * - Parallel direct writes are disabled - a user space decision
-	 * - Parallel direct writes are enabled and i_size is being extended.
-	 * - Shared mmap on direct_io file is supported (FUSE_DIRECT_IO_ALLOW_MMAP).
-	 *   This might not be needed at all, but needs further investigation.
-	 */
-	if (exclusive_lock)
-		inode_lock(inode);
-	else {
-		inode_lock_shared(inode);
-
-		/* A race with truncate might have come up as the decision for
-		 * the lock type was done without holding the lock, check again.
-		 */
-		if (fuse_direct_write_extending_i_size(iocb, from)) {
-			inode_unlock_shared(inode);
-			inode_lock(inode);
-			exclusive_lock = true;
-		}
-	}
+	bool exclusive;
 
+	fuse_dio_lock(iocb, from, &exclusive);
 	res = generic_write_checks(iocb, from);
 	if (res > 0) {
+		task_io_account_write(res);
 		if (!is_sync_kiocb(iocb) && iocb->ki_flags & IOCB_DIRECT) {
 			res = fuse_direct_IO(iocb, from);
 		} else {
@@ -1611,10 +1679,7 @@ static ssize_t fuse_direct_write_iter(struct kiocb *iocb, struct iov_iter *from)
 			fuse_write_update_attr(inode, iocb->ki_pos, res);
 		}
 	}
-	if (exclusive_lock)
-		inode_unlock(inode);
-	else
-		inode_unlock_shared(inode);
+	fuse_dio_unlock(iocb, exclusive);
 
 	return res;
 }
@@ -1631,10 +1696,13 @@ static ssize_t fuse_file_read_iter(struct kiocb *iocb, struct iov_iter *to)
 	if (FUSE_IS_DAX(inode))
 		return fuse_dax_read_iter(iocb, to);
 
-	if (!(ff->open_flags & FOPEN_DIRECT_IO))
-		return fuse_cache_read_iter(iocb, to);
-	else
+	/* FOPEN_DIRECT_IO overrides FOPEN_PASSTHROUGH */
+	if (ff->open_flags & FOPEN_DIRECT_IO)
 		return fuse_direct_read_iter(iocb, to);
+	else if (fuse_file_passthrough(ff))
+		return fuse_passthrough_read_iter(iocb, to);
+	else
+		return fuse_cache_read_iter(iocb, to);
 }
 
 static ssize_t fuse_file_write_iter(struct kiocb *iocb, struct iov_iter *from)
@@ -1649,10 +1717,38 @@ static ssize_t fuse_file_write_iter(struct kiocb *iocb, struct iov_iter *from)
 	if (FUSE_IS_DAX(inode))
 		return fuse_dax_write_iter(iocb, from);
 
-	if (!(ff->open_flags & FOPEN_DIRECT_IO))
+	/* FOPEN_DIRECT_IO overrides FOPEN_PASSTHROUGH */
+	if (ff->open_flags & FOPEN_DIRECT_IO)
+		return fuse_direct_write_iter(iocb, from);
+	else if (fuse_file_passthrough(ff))
+		return fuse_passthrough_write_iter(iocb, from);
+	else
 		return fuse_cache_write_iter(iocb, from);
+}
+
+static ssize_t fuse_splice_read(struct file *in, loff_t *ppos,
+				struct pipe_inode_info *pipe, size_t len,
+				unsigned int flags)
+{
+	struct fuse_file *ff = in->private_data;
+
+	/* FOPEN_DIRECT_IO overrides FOPEN_PASSTHROUGH */
+	if (fuse_file_passthrough(ff) && !(ff->open_flags & FOPEN_DIRECT_IO))
+		return fuse_passthrough_splice_read(in, ppos, pipe, len, flags);
 	else
-		return fuse_direct_write_iter(iocb, from);
+		return filemap_splice_read(in, ppos, pipe, len, flags);
+}
+
+static ssize_t fuse_splice_write(struct pipe_inode_info *pipe, struct file *out,
+				 loff_t *ppos, size_t len, unsigned int flags)
+{
+	struct fuse_file *ff = out->private_data;
+
+	/* FOPEN_DIRECT_IO overrides FOPEN_PASSTHROUGH */
+	if (fuse_file_passthrough(ff) && !(ff->open_flags & FOPEN_DIRECT_IO))
+		return fuse_passthrough_splice_write(pipe, out, ppos, len, flags);
+	else
+		return iter_file_splice_write(pipe, out, ppos, len, flags);
 }
 
 static void fuse_writepage_free(struct fuse_writepage_args *wpa)
@@ -1667,7 +1763,7 @@ static void fuse_writepage_free(struct fuse_writepage_args *wpa)
 		__free_page(ap->pages[i]);
 
 	if (wpa->ia.ff)
-		fuse_file_put(wpa->ia.ff, false, false);
+		fuse_file_put(wpa->ia.ff, false);
 
 	kfree(ap->pages);
 	kfree(wpa);
@@ -1909,7 +2005,7 @@ int fuse_write_inode(struct inode *inode, struct writeback_control *wbc)
 	ff = __fuse_write_file_get(fi);
 	err = fuse_flush_times(inode, ff);
 	if (ff)
-		fuse_file_put(ff, false, false);
+		fuse_file_put(ff, false);
 
 	return err;
 }
@@ -1947,26 +2043,26 @@ static void fuse_writepage_add_to_bucket(struct fuse_conn *fc,
 	rcu_read_unlock();
 }
 
-static int fuse_writepage_locked(struct page *page)
+static int fuse_writepage_locked(struct folio *folio)
 {
-	struct address_space *mapping = page->mapping;
+	struct address_space *mapping = folio->mapping;
 	struct inode *inode = mapping->host;
 	struct fuse_conn *fc = get_fuse_conn(inode);
 	struct fuse_inode *fi = get_fuse_inode(inode);
 	struct fuse_writepage_args *wpa;
 	struct fuse_args_pages *ap;
-	struct page *tmp_page;
+	struct folio *tmp_folio;
 	int error = -ENOMEM;
 
-	set_page_writeback(page);
+	folio_start_writeback(folio);
 
 	wpa = fuse_writepage_args_alloc();
 	if (!wpa)
 		goto err;
 	ap = &wpa->ia.ap;
 
-	tmp_page = alloc_page(GFP_NOFS | __GFP_HIGHMEM);
-	if (!tmp_page)
+	tmp_folio = folio_alloc(GFP_NOFS | __GFP_HIGHMEM, 0);
+	if (!tmp_folio)
 		goto err_free;
 
 	error = -EIO;
@@ -1975,21 +2071,21 @@ static int fuse_writepage_locked(struct page *page)
 		goto err_nofile;
 
 	fuse_writepage_add_to_bucket(fc, wpa);
-	fuse_write_args_fill(&wpa->ia, wpa->ia.ff, page_offset(page), 0);
+	fuse_write_args_fill(&wpa->ia, wpa->ia.ff, folio_pos(folio), 0);
 
-	copy_highpage(tmp_page, page);
+	folio_copy(tmp_folio, folio);
 	wpa->ia.write.in.write_flags |= FUSE_WRITE_CACHE;
 	wpa->next = NULL;
 	ap->args.in_pages = true;
 	ap->num_pages = 1;
-	ap->pages[0] = tmp_page;
+	ap->pages[0] = &tmp_folio->page;
 	ap->descs[0].offset = 0;
 	ap->descs[0].length = PAGE_SIZE;
 	ap->args.end = fuse_writepage_end;
 	wpa->inode = inode;
 
 	inc_wb_stat(&inode_to_bdi(inode)->wb, WB_WRITEBACK);
-	inc_node_page_state(tmp_page, NR_WRITEBACK_TEMP);
+	node_stat_add_folio(tmp_folio, NR_WRITEBACK_TEMP);
 
 	spin_lock(&fi->lock);
 	tree_insert(&fi->writepages, wpa);
@@ -1997,48 +2093,20 @@ static int fuse_writepage_locked(struct page *page)
 	fuse_flush_writepages(inode);
 	spin_unlock(&fi->lock);
 
-	end_page_writeback(page);
+	folio_end_writeback(folio);
 
 	return 0;
 
 err_nofile:
-	__free_page(tmp_page);
+	folio_put(tmp_folio);
 err_free:
 	kfree(wpa);
 err:
-	mapping_set_error(page->mapping, error);
-	end_page_writeback(page);
+	mapping_set_error(folio->mapping, error);
+	folio_end_writeback(folio);
 	return error;
 }
 
-static int fuse_writepage(struct page *page, struct writeback_control *wbc)
-{
-	struct fuse_conn *fc = get_fuse_conn(page->mapping->host);
-	int err;
-
-	if (fuse_page_is_writeback(page->mapping->host, page->index)) {
-		/*
-		 * ->writepages() should be called for sync() and friends.  We
-		 * should only get here on direct reclaim and then we are
-		 * allowed to skip a page which is already in flight
-		 */
-		WARN_ON(wbc->sync_mode == WB_SYNC_ALL);
-
-		redirty_page_for_writepage(wbc, page);
-		unlock_page(page);
-		return 0;
-	}
-
-	if (wbc->sync_mode == WB_SYNC_NONE &&
-	    fc->num_background >= fc->congestion_threshold)
-		return AOP_WRITEPAGE_ACTIVATE;
-
-	err = fuse_writepage_locked(page);
-	unlock_page(page);
-
-	return err;
-}
-
 struct fuse_fill_wb_data {
 	struct fuse_writepage_args *wpa;
 	struct fuse_file *ff;
@@ -2307,7 +2375,7 @@ static int fuse_writepages(struct address_space *mapping,
 		fuse_writepages_send(&data);
 	}
 	if (data.ff)
-		fuse_file_put(data.ff, false, false);
+		fuse_file_put(data.ff, false);
 
 	kfree(data.orig_pages);
 out:
@@ -2401,7 +2469,7 @@ static int fuse_launder_folio(struct folio *folio)
 
 		/* Serialize with pending writeback for the same page */
 		fuse_wait_on_page_writeback(inode, folio->index);
-		err = fuse_writepage_locked(&folio->page);
+		err = fuse_writepage_locked(folio);
 		if (!err)
 			fuse_wait_on_page_writeback(inode, folio->index);
 	}
@@ -2462,11 +2530,27 @@ static int fuse_file_mmap(struct file *file, struct vm_area_struct *vma)
 {
 	struct fuse_file *ff = file->private_data;
 	struct fuse_conn *fc = ff->fm->fc;
+	struct inode *inode = file_inode(file);
+	int rc;
 
 	/* DAX mmap is superior to direct_io mmap */
-	if (FUSE_IS_DAX(file_inode(file)))
+	if (FUSE_IS_DAX(inode))
 		return fuse_dax_mmap(file, vma);
 
+	/*
+	 * If inode is in passthrough io mode, because it has some file open
+	 * in passthrough mode, either mmap to backing file or fail mmap,
+	 * because mixing cached mmap and passthrough io mode is not allowed.
+	 */
+	if (fuse_file_passthrough(ff))
+		return fuse_passthrough_mmap(file, vma);
+	else if (fuse_inode_backing(get_fuse_inode(inode)))
+		return -ENODEV;
+
+	/*
+	 * FOPEN_DIRECT_IO handling is special compared to O_DIRECT,
+	 * as does not allow MAP_SHARED mmap without FUSE_DIRECT_IO_ALLOW_MMAP.
+	 */
 	if (ff->open_flags & FOPEN_DIRECT_IO) {
 		/*
 		 * Can't provide the coherency needed for MAP_SHARED
@@ -2481,6 +2565,17 @@ static int fuse_file_mmap(struct file *file, struct vm_area_struct *vma)
 			/* MAP_PRIVATE */
 			return generic_file_mmap(file, vma);
 		}
+
+		/*
+		 * First mmap of direct_io file enters caching inode io mode.
+		 * Also waits for parallel dio writers to go into serial mode
+		 * (exclusive instead of shared lock).
+		 * After first mmap, the inode stays in caching io mode until
+		 * the direct_io file release.
+		 */
+		rc = fuse_file_cached_io_open(inode, ff);
+		if (rc)
+			return rc;
 	}
 
 	if ((vma->vm_flags & VM_SHARED) && (vma->vm_flags & VM_MAYWRITE))
@@ -2513,14 +2608,14 @@ static int convert_fuse_file_lock(struct fuse_conn *fc,
 		 * translate it into the caller's pid namespace.
 		 */
 		rcu_read_lock();
-		fl->fl_pid = pid_nr_ns(find_pid_ns(ffl->pid, fc->pid_ns), &init_pid_ns);
+		fl->c.flc_pid = pid_nr_ns(find_pid_ns(ffl->pid, fc->pid_ns), &init_pid_ns);
 		rcu_read_unlock();
 		break;
 
 	default:
 		return -EIO;
 	}
-	fl->fl_type = ffl->type;
+	fl->c.flc_type = ffl->type;
 	return 0;
 }
 
@@ -2534,10 +2629,10 @@ static void fuse_lk_fill(struct fuse_args *args, struct file *file,
 
 	memset(inarg, 0, sizeof(*inarg));
 	inarg->fh = ff->fh;
-	inarg->owner = fuse_lock_owner_id(fc, fl->fl_owner);
+	inarg->owner = fuse_lock_owner_id(fc, fl->c.flc_owner);
 	inarg->lk.start = fl->fl_start;
 	inarg->lk.end = fl->fl_end;
-	inarg->lk.type = fl->fl_type;
+	inarg->lk.type = fl->c.flc_type;
 	inarg->lk.pid = pid;
 	if (flock)
 		inarg->lk_flags |= FUSE_LK_FLOCK;
@@ -2574,8 +2669,8 @@ static int fuse_setlk(struct file *file, struct file_lock *fl, int flock)
 	struct fuse_mount *fm = get_fuse_mount(inode);
 	FUSE_ARGS(args);
 	struct fuse_lk_in inarg;
-	int opcode = (fl->fl_flags & FL_SLEEP) ? FUSE_SETLKW : FUSE_SETLK;
-	struct pid *pid = fl->fl_type != F_UNLCK ? task_tgid(current) : NULL;
+	int opcode = (fl->c.flc_flags & FL_SLEEP) ? FUSE_SETLKW : FUSE_SETLK;
+	struct pid *pid = fl->c.flc_type != F_UNLCK ? task_tgid(current) : NULL;
 	pid_t pid_nr = pid_nr_ns(pid, fm->fc->pid_ns);
 	int err;
 
@@ -2584,10 +2679,6 @@ static int fuse_setlk(struct file *file, struct file_lock *fl, int flock)
 		return -ENOLCK;
 	}
 
-	/* Unlock on close is handled by the flush method */
-	if ((fl->fl_flags & FL_CLOSE_POSIX) == FL_CLOSE_POSIX)
-		return 0;
-
 	fuse_lk_fill(&args, file, fl, opcode, pid_nr, flock, &inarg);
 	err = fuse_simple_request(fm, &args);
 
@@ -3217,8 +3308,8 @@ static const struct file_operations fuse_file_operations = {
 	.lock		= fuse_file_lock,
 	.get_unmapped_area = thp_get_unmapped_area,
 	.flock		= fuse_file_flock,
-	.splice_read	= filemap_splice_read,
-	.splice_write	= iter_file_splice_write,
+	.splice_read	= fuse_splice_read,
+	.splice_write	= fuse_splice_write,
 	.unlocked_ioctl	= fuse_file_ioctl,
 	.compat_ioctl	= fuse_file_compat_ioctl,
 	.poll		= fuse_file_poll,
@@ -3229,10 +3320,10 @@ static const struct file_operations fuse_file_operations = {
 static const struct address_space_operations fuse_file_aops  = {
 	.read_folio	= fuse_read_folio,
 	.readahead	= fuse_readahead,
-	.writepage	= fuse_writepage,
 	.writepages	= fuse_writepages,
 	.launder_folio	= fuse_launder_folio,
 	.dirty_folio	= filemap_dirty_folio,
+	.migrate_folio	= filemap_migrate_folio,
 	.bmap		= fuse_bmap,
 	.direct_IO	= fuse_direct_IO,
 	.write_begin	= fuse_write_begin,
@@ -3249,7 +3340,9 @@ void fuse_init_file_inode(struct inode *inode, unsigned int flags)
 	INIT_LIST_HEAD(&fi->write_files);
 	INIT_LIST_HEAD(&fi->queued_writes);
 	fi->writectr = 0;
+	fi->iocachectr = 0;
 	init_waitqueue_head(&fi->page_waitq);
+	init_waitqueue_head(&fi->direct_io_waitq);
 	fi->writepages = RB_ROOT;
 
 	if (IS_ENABLED(CONFIG_FUSE_DAX))
diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h
index 20924e799c..f239196103 100644
--- a/fs/fuse/fuse_i.h
+++ b/fs/fuse/fuse_i.h
@@ -76,6 +76,16 @@ struct fuse_submount_lookup {
 	struct fuse_forget_link *forget;
 };
 
+/** Container for data related to mapping to backing file */
+struct fuse_backing {
+	struct file *file;
+	struct cred *cred;
+
+	/** refcount */
+	refcount_t count;
+	struct rcu_head rcu;
+};
+
 /** FUSE inode */
 struct fuse_inode {
 	/** Inode data */
@@ -111,7 +121,7 @@ struct fuse_inode {
 	u64 attr_version;
 
 	union {
-		/* Write related fields (regular file only) */
+		/* read/write io cache (regular file only) */
 		struct {
 			/* Files usable in writepage.  Protected by fi->lock */
 			struct list_head write_files;
@@ -123,9 +133,15 @@ struct fuse_inode {
 			 * (FUSE_NOWRITE) means more writes are blocked */
 			int writectr;
 
+			/** Number of files/maps using page cache */
+			int iocachectr;
+
 			/* Waitq for writepage completion */
 			wait_queue_head_t page_waitq;
 
+			/* waitq for direct-io completion */
+			wait_queue_head_t direct_io_waitq;
+
 			/* List of writepage requestst (pending or sent) */
 			struct rb_root writepages;
 		};
@@ -173,6 +189,10 @@ struct fuse_inode {
 #endif
 	/** Submount specific lookup tracking */
 	struct fuse_submount_lookup *submount_lookup;
+#ifdef CONFIG_FUSE_PASSTHROUGH
+	/** Reference to backing file in passthrough mode */
+	struct fuse_backing *fb;
+#endif
 };
 
 /** FUSE inode state bits */
@@ -187,19 +207,21 @@ enum {
 	FUSE_I_BAD,
 	/* Has btime */
 	FUSE_I_BTIME,
+	/* Wants or already has page cache IO */
+	FUSE_I_CACHE_IO_MODE,
 };
 
 struct fuse_conn;
 struct fuse_mount;
-struct fuse_release_args;
+union fuse_file_args;
 
 /** FUSE specific file data */
 struct fuse_file {
 	/** Fuse connection for this file */
 	struct fuse_mount *fm;
 
-	/* Argument space reserved for release */
-	struct fuse_release_args *release_args;
+	/* Argument space reserved for open/release */
+	union fuse_file_args *args;
 
 	/** Kernel file handle guaranteed to be unique */
 	u64 kh;
@@ -221,12 +243,6 @@ struct fuse_file {
 
 	/* Readdir related */
 	struct {
-		/*
-		 * Protects below fields against (crazy) parallel readdir on
-		 * same open file.  Uncontended in the normal case.
-		 */
-		struct mutex lock;
-
 		/* Dir stream position */
 		loff_t pos;
 
@@ -244,6 +260,15 @@ struct fuse_file {
 	/** Wait queue head for poll */
 	wait_queue_head_t poll_wait;
 
+	/** Does file hold a fi->iocachectr refcount? */
+	enum { IOM_NONE, IOM_CACHED, IOM_UNCACHED } iomode;
+
+#ifdef CONFIG_FUSE_PASSTHROUGH
+	/** Reference to backing file in passthrough mode */
+	struct file *passthrough;
+	const struct cred *cred;
+#endif
+
 	/** Has flock been performed on this file? */
 	bool flock:1;
 };
@@ -283,6 +308,7 @@ struct fuse_args {
 	bool page_replace:1;
 	bool may_block:1;
 	bool is_ext:1;
+	bool is_pinned:1;
 	struct fuse_in_arg in_args[3];
 	struct fuse_arg out_args[2];
 	void (*end)(struct fuse_mount *fm, struct fuse_args *args, int error);
@@ -295,6 +321,19 @@ struct fuse_args_pages {
 	unsigned int num_pages;
 };
 
+struct fuse_release_args {
+	struct fuse_args args;
+	struct fuse_release_in inarg;
+	struct inode *inode;
+};
+
+union fuse_file_args {
+	/* Used during open() */
+	struct fuse_open_out open_outarg;
+	/* Used during release() */
+	struct fuse_release_args release_args;
+};
+
 #define FUSE_ARGS(args) struct fuse_args args = {}
 
 /** The request IO state (for asynchronous processing) */
@@ -818,6 +857,12 @@ struct fuse_conn {
 	/* Is statx not implemented by fs? */
 	unsigned int no_statx:1;
 
+	/** Passthrough support for read/write IO */
+	unsigned int passthrough:1;
+
+	/** Maximum stack depth for passthrough backing files */
+	int max_stack_depth;
+
 	/** The number of requests waiting for completion */
 	atomic_t num_waiting;
 
@@ -867,6 +912,11 @@ struct fuse_conn {
 
 	/* New writepages go into this bucket */
 	struct fuse_sync_bucket __rcu *curr_bucket;
+
+#ifdef CONFIG_FUSE_PASSTHROUGH
+	/** IDR for backing files ids */
+	struct idr backing_files_map;
+#endif
 };
 
 /*
@@ -1031,14 +1081,9 @@ void fuse_read_args_fill(struct fuse_io_args *ia, struct file *file, loff_t pos,
 			 size_t count, int opcode);
 
 
-/**
- * Send OPEN or OPENDIR request
- */
-int fuse_open_common(struct inode *inode, struct file *file, bool isdir);
-
-struct fuse_file *fuse_file_alloc(struct fuse_mount *fm);
+struct fuse_file *fuse_file_alloc(struct fuse_mount *fm, bool release);
 void fuse_file_free(struct fuse_file *ff);
-void fuse_finish_open(struct inode *inode, struct file *file);
+int fuse_finish_open(struct inode *inode, struct file *file);
 
 void fuse_sync_release(struct fuse_inode *fi, struct fuse_file *ff,
 		       unsigned int flags);
@@ -1348,11 +1393,83 @@ int fuse_fileattr_get(struct dentry *dentry, struct fileattr *fa);
 int fuse_fileattr_set(struct mnt_idmap *idmap,
 		      struct dentry *dentry, struct fileattr *fa);
 
-/* file.c */
+/* iomode.c */
+int fuse_file_cached_io_open(struct inode *inode, struct fuse_file *ff);
+int fuse_inode_uncached_io_start(struct fuse_inode *fi,
+				 struct fuse_backing *fb);
+void fuse_inode_uncached_io_end(struct fuse_inode *fi);
+
+int fuse_file_io_open(struct file *file, struct inode *inode);
+void fuse_file_io_release(struct fuse_file *ff, struct inode *inode);
 
+/* file.c */
 struct fuse_file *fuse_file_open(struct fuse_mount *fm, u64 nodeid,
 				 unsigned int open_flags, bool isdir);
 void fuse_file_release(struct inode *inode, struct fuse_file *ff,
 		       unsigned int open_flags, fl_owner_t id, bool isdir);
 
+/* passthrough.c */
+static inline struct fuse_backing *fuse_inode_backing(struct fuse_inode *fi)
+{
+#ifdef CONFIG_FUSE_PASSTHROUGH
+	return READ_ONCE(fi->fb);
+#else
+	return NULL;
+#endif
+}
+
+static inline struct fuse_backing *fuse_inode_backing_set(struct fuse_inode *fi,
+							  struct fuse_backing *fb)
+{
+#ifdef CONFIG_FUSE_PASSTHROUGH
+	return xchg(&fi->fb, fb);
+#else
+	return NULL;
+#endif
+}
+
+#ifdef CONFIG_FUSE_PASSTHROUGH
+struct fuse_backing *fuse_backing_get(struct fuse_backing *fb);
+void fuse_backing_put(struct fuse_backing *fb);
+#else
+
+static inline struct fuse_backing *fuse_backing_get(struct fuse_backing *fb)
+{
+	return NULL;
+}
+
+static inline void fuse_backing_put(struct fuse_backing *fb)
+{
+}
+#endif
+
+void fuse_backing_files_init(struct fuse_conn *fc);
+void fuse_backing_files_free(struct fuse_conn *fc);
+int fuse_backing_open(struct fuse_conn *fc, struct fuse_backing_map *map);
+int fuse_backing_close(struct fuse_conn *fc, int backing_id);
+
+struct fuse_backing *fuse_passthrough_open(struct file *file,
+					   struct inode *inode,
+					   int backing_id);
+void fuse_passthrough_release(struct fuse_file *ff, struct fuse_backing *fb);
+
+static inline struct file *fuse_file_passthrough(struct fuse_file *ff)
+{
+#ifdef CONFIG_FUSE_PASSTHROUGH
+	return ff->passthrough;
+#else
+	return NULL;
+#endif
+}
+
+ssize_t fuse_passthrough_read_iter(struct kiocb *iocb, struct iov_iter *iter);
+ssize_t fuse_passthrough_write_iter(struct kiocb *iocb, struct iov_iter *iter);
+ssize_t fuse_passthrough_splice_read(struct file *in, loff_t *ppos,
+				     struct pipe_inode_info *pipe,
+				     size_t len, unsigned int flags);
+ssize_t fuse_passthrough_splice_write(struct pipe_inode_info *pipe,
+				      struct file *out, loff_t *ppos,
+				      size_t len, unsigned int flags);
+ssize_t fuse_passthrough_mmap(struct file *file, struct vm_area_struct *vma);
+
 #endif /* _FS_FUSE_I_H */
diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c
index f73865751c..32fe6fa72f 100644
--- a/fs/fuse/inode.c
+++ b/fs/fuse/inode.c
@@ -111,6 +111,9 @@ static struct inode *fuse_alloc_inode(struct super_block *sb)
 	if (IS_ENABLED(CONFIG_FUSE_DAX) && !fuse_dax_inode_alloc(sb, fi))
 		goto out_free_forget;
 
+	if (IS_ENABLED(CONFIG_FUSE_PASSTHROUGH))
+		fuse_inode_backing_set(fi, NULL);
+
 	return &fi->inode;
 
 out_free_forget:
@@ -129,6 +132,9 @@ static void fuse_free_inode(struct inode *inode)
 #ifdef CONFIG_FUSE_DAX
 	kfree(fi->dax);
 #endif
+	if (IS_ENABLED(CONFIG_FUSE_PASSTHROUGH))
+		fuse_backing_put(fuse_inode_backing(fi));
+
 	kmem_cache_free(fuse_inode_cachep, fi);
 }
 
@@ -169,6 +175,7 @@ static void fuse_evict_inode(struct inode *inode)
 		}
 	}
 	if (S_ISREG(inode->i_mode) && !fuse_is_bad(inode)) {
+		WARN_ON(fi->iocachectr != 0);
 		WARN_ON(!list_empty(&fi->write_files));
 		WARN_ON(!list_empty(&fi->queued_writes));
 	}
@@ -748,6 +755,8 @@ static int fuse_parse_param(struct fs_context *fsc, struct fs_parameter *param)
 	struct fs_parse_result result;
 	struct fuse_fs_context *ctx = fsc->fs_private;
 	int opt;
+	kuid_t kuid;
+	kgid_t kgid;
 
 	if (fsc->purpose == FS_CONTEXT_FOR_RECONFIGURE) {
 		/*
@@ -792,16 +801,30 @@ static int fuse_parse_param(struct fs_context *fsc, struct fs_parameter *param)
 		break;
 
 	case OPT_USER_ID:
-		ctx->user_id = make_kuid(fsc->user_ns, result.uint_32);
-		if (!uid_valid(ctx->user_id))
+		kuid =  make_kuid(fsc->user_ns, result.uint_32);
+		if (!uid_valid(kuid))
+			return invalfc(fsc, "Invalid user_id");
+		/*
+		 * The requested uid must be representable in the
+		 * filesystem's idmapping.
+		 */
+		if (!kuid_has_mapping(fsc->user_ns, kuid))
 			return invalfc(fsc, "Invalid user_id");
+		ctx->user_id = kuid;
 		ctx->user_id_present = true;
 		break;
 
 	case OPT_GROUP_ID:
-		ctx->group_id = make_kgid(fsc->user_ns, result.uint_32);
-		if (!gid_valid(ctx->group_id))
+		kgid = make_kgid(fsc->user_ns, result.uint_32);;
+		if (!gid_valid(kgid))
+			return invalfc(fsc, "Invalid group_id");
+		/*
+		 * The requested gid must be representable in the
+		 * filesystem's idmapping.
+		 */
+		if (!kgid_has_mapping(fsc->user_ns, kgid))
 			return invalfc(fsc, "Invalid group_id");
+		ctx->group_id = kgid;
 		ctx->group_id_present = true;
 		break;
 
@@ -927,6 +950,9 @@ void fuse_conn_init(struct fuse_conn *fc, struct fuse_mount *fm,
 	fc->max_pages = FUSE_DEFAULT_MAX_PAGES_PER_REQ;
 	fc->max_pages_limit = FUSE_MAX_MAX_PAGES;
 
+	if (IS_ENABLED(CONFIG_FUSE_PASSTHROUGH))
+		fuse_backing_files_init(fc);
+
 	INIT_LIST_HEAD(&fc->mounts);
 	list_add(&fm->fc_entry, &fc->mounts);
 	fm->fc = fc;
@@ -957,6 +983,8 @@ void fuse_conn_put(struct fuse_conn *fc)
 			WARN_ON(atomic_read(&bucket->count) != 1);
 			kfree(bucket);
 		}
+		if (IS_ENABLED(CONFIG_FUSE_PASSTHROUGH))
+			fuse_backing_files_free(fc);
 		call_rcu(&fc->rcu, delayed_release);
 	}
 }
@@ -977,7 +1005,7 @@ static struct inode *fuse_get_root_inode(struct super_block *sb, unsigned mode)
 	attr.mode = mode;
 	attr.ino = FUSE_ROOT_ID;
 	attr.nlink = 1;
-	return fuse_iget(sb, 1, 0, &attr, 0, 0);
+	return fuse_iget(sb, FUSE_ROOT_ID, 0, &attr, 0, 0);
 }
 
 struct fuse_inode_handle {
@@ -1120,6 +1148,11 @@ static struct dentry *fuse_get_parent(struct dentry *child)
 	return parent;
 }
 
+/* only for fid encoding; no support for file handle */
+static const struct export_operations fuse_export_fid_operations = {
+	.encode_fh	= fuse_encode_fh,
+};
+
 static const struct export_operations fuse_export_operations = {
 	.fh_to_dentry	= fuse_fh_to_dentry,
 	.fh_to_parent	= fuse_fh_to_parent,
@@ -1294,6 +1327,26 @@ static void process_init_reply(struct fuse_mount *fm, struct fuse_args *args,
 				fc->create_supp_group = 1;
 			if (flags & FUSE_DIRECT_IO_ALLOW_MMAP)
 				fc->direct_io_allow_mmap = 1;
+			/*
+			 * max_stack_depth is the max stack depth of FUSE fs,
+			 * so it has to be at least 1 to support passthrough
+			 * to backing files.
+			 *
+			 * with max_stack_depth > 1, the backing files can be
+			 * on a stacked fs (e.g. overlayfs) themselves and with
+			 * max_stack_depth == 1, FUSE fs can be stacked as the
+			 * underlying fs of a stacked fs (e.g. overlayfs).
+			 */
+			if (IS_ENABLED(CONFIG_FUSE_PASSTHROUGH) &&
+			    (flags & FUSE_PASSTHROUGH) &&
+			    arg->max_stack_depth > 0 &&
+			    arg->max_stack_depth <= FILESYSTEM_MAX_STACK_DEPTH) {
+				fc->passthrough = 1;
+				fc->max_stack_depth = arg->max_stack_depth;
+				fm->sb->s_stack_depth = arg->max_stack_depth;
+			}
+			if (flags & FUSE_NO_EXPORT_SUPPORT)
+				fm->sb->s_export_op = &fuse_export_fid_operations;
 		} else {
 			ra_pages = fc->max_read / PAGE_SIZE;
 			fc->no_lock = 1;
@@ -1340,7 +1393,8 @@ void fuse_send_init(struct fuse_mount *fm)
 		FUSE_NO_OPENDIR_SUPPORT | FUSE_EXPLICIT_INVAL_DATA |
 		FUSE_HANDLE_KILLPRIV_V2 | FUSE_SETXATTR_EXT | FUSE_INIT_EXT |
 		FUSE_SECURITY_CTX | FUSE_CREATE_SUPP_GROUP |
-		FUSE_HAS_EXPIRE_ONLY | FUSE_DIRECT_IO_ALLOW_MMAP;
+		FUSE_HAS_EXPIRE_ONLY | FUSE_DIRECT_IO_ALLOW_MMAP |
+		FUSE_NO_EXPORT_SUPPORT | FUSE_HAS_RESEND;
 #ifdef CONFIG_FUSE_DAX
 	if (fm->fc->dax)
 		flags |= FUSE_MAP_ALIGNMENT;
@@ -1349,6 +1403,8 @@ void fuse_send_init(struct fuse_mount *fm)
 #endif
 	if (fm->fc->auto_submounts)
 		flags |= FUSE_SUBMOUNTS;
+	if (IS_ENABLED(CONFIG_FUSE_PASSTHROUGH))
+		flags |= FUSE_PASSTHROUGH;
 
 	ia->in.flags = flags;
 	ia->in.flags2 = flags >> 32;
@@ -1499,8 +1555,8 @@ static void fuse_fill_attr_from_inode(struct fuse_attr *attr,
 		.ctimensec	= ctime.tv_nsec,
 		.mode		= fi->inode.i_mode,
 		.nlink		= fi->inode.i_nlink,
-		.uid		= fi->inode.i_uid.val,
-		.gid		= fi->inode.i_gid.val,
+		.uid		= __kuid_val(fi->inode.i_uid),
+		.gid		= __kgid_val(fi->inode.i_gid),
 		.rdev		= fi->inode.i_rdev,
 		.blksize	= 1u << fi->inode.i_blkbits,
 	};
@@ -1537,6 +1593,7 @@ static int fuse_fill_super_submount(struct super_block *sb,
 	sb->s_bdi = bdi_get(parent_sb->s_bdi);
 
 	sb->s_xattr = parent_sb->s_xattr;
+	sb->s_export_op = parent_sb->s_export_op;
 	sb->s_time_gran = parent_sb->s_time_gran;
 	sb->s_blocksize = parent_sb->s_blocksize;
 	sb->s_blocksize_bits = parent_sb->s_blocksize_bits;
diff --git a/fs/fuse/ioctl.c b/fs/fuse/ioctl.c
index 726640fa43..572ce8a82c 100644
--- a/fs/fuse/ioctl.c
+++ b/fs/fuse/ioctl.c
@@ -8,6 +8,7 @@
 #include <linux/uio.h>
 #include <linux/compat.h>
 #include <linux/fileattr.h>
+#include <linux/fsverity.h>
 
 static ssize_t fuse_send_ioctl(struct fuse_mount *fm, struct fuse_args *args,
 			       struct fuse_ioctl_out *outarg)
@@ -117,6 +118,53 @@ static int fuse_copy_ioctl_iovec(struct fuse_conn *fc, struct iovec *dst,
 	return 0;
 }
 
+/* For fs-verity, determine iov lengths from input */
+static int fuse_setup_measure_verity(unsigned long arg, struct iovec *iov)
+{
+	__u16 digest_size;
+	struct fsverity_digest __user *uarg = (void __user *)arg;
+
+	if (copy_from_user(&digest_size, &uarg->digest_size, sizeof(digest_size)))
+		return -EFAULT;
+
+	if (digest_size > SIZE_MAX - sizeof(struct fsverity_digest))
+		return -EINVAL;
+
+	iov->iov_len = sizeof(struct fsverity_digest) + digest_size;
+
+	return 0;
+}
+
+static int fuse_setup_enable_verity(unsigned long arg, struct iovec *iov,
+				    unsigned int *in_iovs)
+{
+	struct fsverity_enable_arg enable;
+	struct fsverity_enable_arg __user *uarg = (void __user *)arg;
+	const __u32 max_buffer_len = FUSE_MAX_MAX_PAGES * PAGE_SIZE;
+
+	if (copy_from_user(&enable, uarg, sizeof(enable)))
+		return -EFAULT;
+
+	if (enable.salt_size > max_buffer_len || enable.sig_size > max_buffer_len)
+		return -ENOMEM;
+
+	if (enable.salt_size > 0) {
+		iov++;
+		(*in_iovs)++;
+
+		iov->iov_base = u64_to_user_ptr(enable.salt_ptr);
+		iov->iov_len = enable.salt_size;
+	}
+
+	if (enable.sig_size > 0) {
+		iov++;
+		(*in_iovs)++;
+
+		iov->iov_base = u64_to_user_ptr(enable.sig_ptr);
+		iov->iov_len = enable.sig_size;
+	}
+	return 0;
+}
 
 /*
  * For ioctls, there is no generic way to determine how much memory
@@ -227,6 +275,18 @@ long fuse_do_ioctl(struct file *file, unsigned int cmd, unsigned long arg,
 			out_iov = iov;
 			out_iovs = 1;
 		}
+
+		err = 0;
+		switch (cmd) {
+		case FS_IOC_MEASURE_VERITY:
+			err = fuse_setup_measure_verity(arg, iov);
+			break;
+		case FS_IOC_ENABLE_VERITY:
+			err = fuse_setup_enable_verity(arg, iov, &in_iovs);
+			break;
+		}
+		if (err)
+			goto out;
 	}
 
  retry:
diff --git a/fs/fuse/iomode.c b/fs/fuse/iomode.c
new file mode 100644
index 0000000000..c99e285f31
--- /dev/null
+++ b/fs/fuse/iomode.c
@@ -0,0 +1,276 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * FUSE inode io modes.
+ *
+ * Copyright (c) 2024 CTERA Networks.
+ */
+
+#include "fuse_i.h"
+
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/file.h>
+#include <linux/fs.h>
+
+/*
+ * Return true if need to wait for new opens in caching mode.
+ */
+static inline bool fuse_is_io_cache_wait(struct fuse_inode *fi)
+{
+	return READ_ONCE(fi->iocachectr) < 0 && !fuse_inode_backing(fi);
+}
+
+/*
+ * Called on cached file open() and on first mmap() of direct_io file.
+ * Takes cached_io inode mode reference to be dropped on file release.
+ *
+ * Blocks new parallel dio writes and waits for the in-progress parallel dio
+ * writes to complete.
+ */
+int fuse_file_cached_io_open(struct inode *inode, struct fuse_file *ff)
+{
+	struct fuse_inode *fi = get_fuse_inode(inode);
+
+	/* There are no io modes if server does not implement open */
+	if (!ff->args)
+		return 0;
+
+	spin_lock(&fi->lock);
+	/*
+	 * Setting the bit advises new direct-io writes to use an exclusive
+	 * lock - without it the wait below might be forever.
+	 */
+	while (fuse_is_io_cache_wait(fi)) {
+		set_bit(FUSE_I_CACHE_IO_MODE, &fi->state);
+		spin_unlock(&fi->lock);
+		wait_event(fi->direct_io_waitq, !fuse_is_io_cache_wait(fi));
+		spin_lock(&fi->lock);
+	}
+
+	/*
+	 * Check if inode entered passthrough io mode while waiting for parallel
+	 * dio write completion.
+	 */
+	if (fuse_inode_backing(fi)) {
+		clear_bit(FUSE_I_CACHE_IO_MODE, &fi->state);
+		spin_unlock(&fi->lock);
+		return -ETXTBSY;
+	}
+
+	WARN_ON(ff->iomode == IOM_UNCACHED);
+	if (ff->iomode == IOM_NONE) {
+		ff->iomode = IOM_CACHED;
+		if (fi->iocachectr == 0)
+			set_bit(FUSE_I_CACHE_IO_MODE, &fi->state);
+		fi->iocachectr++;
+	}
+	spin_unlock(&fi->lock);
+	return 0;
+}
+
+static void fuse_file_cached_io_release(struct fuse_file *ff,
+					struct fuse_inode *fi)
+{
+	spin_lock(&fi->lock);
+	WARN_ON(fi->iocachectr <= 0);
+	WARN_ON(ff->iomode != IOM_CACHED);
+	ff->iomode = IOM_NONE;
+	fi->iocachectr--;
+	if (fi->iocachectr == 0)
+		clear_bit(FUSE_I_CACHE_IO_MODE, &fi->state);
+	spin_unlock(&fi->lock);
+}
+
+/* Start strictly uncached io mode where cache access is not allowed */
+int fuse_inode_uncached_io_start(struct fuse_inode *fi, struct fuse_backing *fb)
+{
+	struct fuse_backing *oldfb;
+	int err = 0;
+
+	spin_lock(&fi->lock);
+	/* deny conflicting backing files on same fuse inode */
+	oldfb = fuse_inode_backing(fi);
+	if (fb && oldfb && oldfb != fb) {
+		err = -EBUSY;
+		goto unlock;
+	}
+	if (fi->iocachectr > 0) {
+		err = -ETXTBSY;
+		goto unlock;
+	}
+	fi->iocachectr--;
+
+	/* fuse inode holds a single refcount of backing file */
+	if (fb && !oldfb) {
+		oldfb = fuse_inode_backing_set(fi, fb);
+		WARN_ON_ONCE(oldfb != NULL);
+	} else {
+		fuse_backing_put(fb);
+	}
+unlock:
+	spin_unlock(&fi->lock);
+	return err;
+}
+
+/* Takes uncached_io inode mode reference to be dropped on file release */
+static int fuse_file_uncached_io_open(struct inode *inode,
+				      struct fuse_file *ff,
+				      struct fuse_backing *fb)
+{
+	struct fuse_inode *fi = get_fuse_inode(inode);
+	int err;
+
+	err = fuse_inode_uncached_io_start(fi, fb);
+	if (err)
+		return err;
+
+	WARN_ON(ff->iomode != IOM_NONE);
+	ff->iomode = IOM_UNCACHED;
+	return 0;
+}
+
+void fuse_inode_uncached_io_end(struct fuse_inode *fi)
+{
+	struct fuse_backing *oldfb = NULL;
+
+	spin_lock(&fi->lock);
+	WARN_ON(fi->iocachectr >= 0);
+	fi->iocachectr++;
+	if (!fi->iocachectr) {
+		wake_up(&fi->direct_io_waitq);
+		oldfb = fuse_inode_backing_set(fi, NULL);
+	}
+	spin_unlock(&fi->lock);
+	if (oldfb)
+		fuse_backing_put(oldfb);
+}
+
+/* Drop uncached_io reference from passthrough open */
+static void fuse_file_uncached_io_release(struct fuse_file *ff,
+					  struct fuse_inode *fi)
+{
+	WARN_ON(ff->iomode != IOM_UNCACHED);
+	ff->iomode = IOM_NONE;
+	fuse_inode_uncached_io_end(fi);
+}
+
+/*
+ * Open flags that are allowed in combination with FOPEN_PASSTHROUGH.
+ * A combination of FOPEN_PASSTHROUGH and FOPEN_DIRECT_IO means that read/write
+ * operations go directly to the server, but mmap is done on the backing file.
+ * FOPEN_PASSTHROUGH mode should not co-exist with any users of the fuse inode
+ * page cache, so FOPEN_KEEP_CACHE is a strange and undesired combination.
+ */
+#define FOPEN_PASSTHROUGH_MASK \
+	(FOPEN_PASSTHROUGH | FOPEN_DIRECT_IO | FOPEN_PARALLEL_DIRECT_WRITES | \
+	 FOPEN_NOFLUSH)
+
+static int fuse_file_passthrough_open(struct inode *inode, struct file *file)
+{
+	struct fuse_file *ff = file->private_data;
+	struct fuse_conn *fc = get_fuse_conn(inode);
+	struct fuse_backing *fb;
+	int err;
+
+	/* Check allowed conditions for file open in passthrough mode */
+	if (!IS_ENABLED(CONFIG_FUSE_PASSTHROUGH) || !fc->passthrough ||
+	    (ff->open_flags & ~FOPEN_PASSTHROUGH_MASK))
+		return -EINVAL;
+
+	fb = fuse_passthrough_open(file, inode,
+				   ff->args->open_outarg.backing_id);
+	if (IS_ERR(fb))
+		return PTR_ERR(fb);
+
+	/* First passthrough file open denies caching inode io mode */
+	err = fuse_file_uncached_io_open(inode, ff, fb);
+	if (!err)
+		return 0;
+
+	fuse_passthrough_release(ff, fb);
+	fuse_backing_put(fb);
+
+	return err;
+}
+
+/* Request access to submit new io to inode via open file */
+int fuse_file_io_open(struct file *file, struct inode *inode)
+{
+	struct fuse_file *ff = file->private_data;
+	struct fuse_inode *fi = get_fuse_inode(inode);
+	int err;
+
+	/*
+	 * io modes are not relevant with DAX and with server that does not
+	 * implement open.
+	 */
+	if (FUSE_IS_DAX(inode) || !ff->args)
+		return 0;
+
+	/*
+	 * Server is expected to use FOPEN_PASSTHROUGH for all opens of an inode
+	 * which is already open for passthrough.
+	 */
+	err = -EINVAL;
+	if (fuse_inode_backing(fi) && !(ff->open_flags & FOPEN_PASSTHROUGH))
+		goto fail;
+
+	/*
+	 * FOPEN_PARALLEL_DIRECT_WRITES requires FOPEN_DIRECT_IO.
+	 */
+	if (!(ff->open_flags & FOPEN_DIRECT_IO))
+		ff->open_flags &= ~FOPEN_PARALLEL_DIRECT_WRITES;
+
+	/*
+	 * First passthrough file open denies caching inode io mode.
+	 * First caching file open enters caching inode io mode.
+	 *
+	 * Note that if user opens a file open with O_DIRECT, but server did
+	 * not specify FOPEN_DIRECT_IO, a later fcntl() could remove O_DIRECT,
+	 * so we put the inode in caching mode to prevent parallel dio.
+	 */
+	if ((ff->open_flags & FOPEN_DIRECT_IO) &&
+	    !(ff->open_flags & FOPEN_PASSTHROUGH))
+		return 0;
+
+	if (ff->open_flags & FOPEN_PASSTHROUGH)
+		err = fuse_file_passthrough_open(inode, file);
+	else
+		err = fuse_file_cached_io_open(inode, ff);
+	if (err)
+		goto fail;
+
+	return 0;
+
+fail:
+	pr_debug("failed to open file in requested io mode (open_flags=0x%x, err=%i).\n",
+		 ff->open_flags, err);
+	/*
+	 * The file open mode determines the inode io mode.
+	 * Using incorrect open mode is a server mistake, which results in
+	 * user visible failure of open() with EIO error.
+	 */
+	return -EIO;
+}
+
+/* No more pending io and no new io possible to inode via open/mmapped file */
+void fuse_file_io_release(struct fuse_file *ff, struct inode *inode)
+{
+	struct fuse_inode *fi = get_fuse_inode(inode);
+
+	/*
+	 * Last passthrough file close allows caching inode io mode.
+	 * Last caching file close exits caching inode io mode.
+	 */
+	switch (ff->iomode) {
+	case IOM_NONE:
+		/* Nothing to do */
+		break;
+	case IOM_UNCACHED:
+		fuse_file_uncached_io_release(ff, fi);
+		break;
+	case IOM_CACHED:
+		fuse_file_cached_io_release(ff, fi);
+		break;
+	}
+}
diff --git a/fs/fuse/passthrough.c b/fs/fuse/passthrough.c
new file mode 100644
index 0000000000..9666d13884
--- /dev/null
+++ b/fs/fuse/passthrough.c
@@ -0,0 +1,355 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * FUSE passthrough to backing file.
+ *
+ * Copyright (c) 2023 CTERA Networks.
+ */
+
+#include "fuse_i.h"
+
+#include <linux/file.h>
+#include <linux/backing-file.h>
+#include <linux/splice.h>
+
+static void fuse_file_accessed(struct file *file)
+{
+	struct inode *inode = file_inode(file);
+
+	fuse_invalidate_atime(inode);
+}
+
+static void fuse_file_modified(struct file *file)
+{
+	struct inode *inode = file_inode(file);
+
+	fuse_invalidate_attr_mask(inode, FUSE_STATX_MODSIZE);
+}
+
+ssize_t fuse_passthrough_read_iter(struct kiocb *iocb, struct iov_iter *iter)
+{
+	struct file *file = iocb->ki_filp;
+	struct fuse_file *ff = file->private_data;
+	struct file *backing_file = fuse_file_passthrough(ff);
+	size_t count = iov_iter_count(iter);
+	ssize_t ret;
+	struct backing_file_ctx ctx = {
+		.cred = ff->cred,
+		.user_file = file,
+		.accessed = fuse_file_accessed,
+	};
+
+
+	pr_debug("%s: backing_file=0x%p, pos=%lld, len=%zu\n", __func__,
+		 backing_file, iocb->ki_pos, count);
+
+	if (!count)
+		return 0;
+
+	ret = backing_file_read_iter(backing_file, iter, iocb, iocb->ki_flags,
+				     &ctx);
+
+	return ret;
+}
+
+ssize_t fuse_passthrough_write_iter(struct kiocb *iocb,
+				    struct iov_iter *iter)
+{
+	struct file *file = iocb->ki_filp;
+	struct inode *inode = file_inode(file);
+	struct fuse_file *ff = file->private_data;
+	struct file *backing_file = fuse_file_passthrough(ff);
+	size_t count = iov_iter_count(iter);
+	ssize_t ret;
+	struct backing_file_ctx ctx = {
+		.cred = ff->cred,
+		.user_file = file,
+		.end_write = fuse_file_modified,
+	};
+
+	pr_debug("%s: backing_file=0x%p, pos=%lld, len=%zu\n", __func__,
+		 backing_file, iocb->ki_pos, count);
+
+	if (!count)
+		return 0;
+
+	inode_lock(inode);
+	ret = backing_file_write_iter(backing_file, iter, iocb, iocb->ki_flags,
+				      &ctx);
+	inode_unlock(inode);
+
+	return ret;
+}
+
+ssize_t fuse_passthrough_splice_read(struct file *in, loff_t *ppos,
+				     struct pipe_inode_info *pipe,
+				     size_t len, unsigned int flags)
+{
+	struct fuse_file *ff = in->private_data;
+	struct file *backing_file = fuse_file_passthrough(ff);
+	struct backing_file_ctx ctx = {
+		.cred = ff->cred,
+		.user_file = in,
+		.accessed = fuse_file_accessed,
+	};
+
+	pr_debug("%s: backing_file=0x%p, pos=%lld, len=%zu, flags=0x%x\n", __func__,
+		 backing_file, ppos ? *ppos : 0, len, flags);
+
+	return backing_file_splice_read(backing_file, ppos, pipe, len, flags,
+					&ctx);
+}
+
+ssize_t fuse_passthrough_splice_write(struct pipe_inode_info *pipe,
+				      struct file *out, loff_t *ppos,
+				      size_t len, unsigned int flags)
+{
+	struct fuse_file *ff = out->private_data;
+	struct file *backing_file = fuse_file_passthrough(ff);
+	struct inode *inode = file_inode(out);
+	ssize_t ret;
+	struct backing_file_ctx ctx = {
+		.cred = ff->cred,
+		.user_file = out,
+		.end_write = fuse_file_modified,
+	};
+
+	pr_debug("%s: backing_file=0x%p, pos=%lld, len=%zu, flags=0x%x\n", __func__,
+		 backing_file, ppos ? *ppos : 0, len, flags);
+
+	inode_lock(inode);
+	ret = backing_file_splice_write(pipe, backing_file, ppos, len, flags,
+					&ctx);
+	inode_unlock(inode);
+
+	return ret;
+}
+
+ssize_t fuse_passthrough_mmap(struct file *file, struct vm_area_struct *vma)
+{
+	struct fuse_file *ff = file->private_data;
+	struct file *backing_file = fuse_file_passthrough(ff);
+	struct backing_file_ctx ctx = {
+		.cred = ff->cred,
+		.user_file = file,
+		.accessed = fuse_file_accessed,
+	};
+
+	pr_debug("%s: backing_file=0x%p, start=%lu, end=%lu\n", __func__,
+		 backing_file, vma->vm_start, vma->vm_end);
+
+	return backing_file_mmap(backing_file, vma, &ctx);
+}
+
+struct fuse_backing *fuse_backing_get(struct fuse_backing *fb)
+{
+	if (fb && refcount_inc_not_zero(&fb->count))
+		return fb;
+	return NULL;
+}
+
+static void fuse_backing_free(struct fuse_backing *fb)
+{
+	pr_debug("%s: fb=0x%p\n", __func__, fb);
+
+	if (fb->file)
+		fput(fb->file);
+	put_cred(fb->cred);
+	kfree_rcu(fb, rcu);
+}
+
+void fuse_backing_put(struct fuse_backing *fb)
+{
+	if (fb && refcount_dec_and_test(&fb->count))
+		fuse_backing_free(fb);
+}
+
+void fuse_backing_files_init(struct fuse_conn *fc)
+{
+	idr_init(&fc->backing_files_map);
+}
+
+static int fuse_backing_id_alloc(struct fuse_conn *fc, struct fuse_backing *fb)
+{
+	int id;
+
+	idr_preload(GFP_KERNEL);
+	spin_lock(&fc->lock);
+	/* FIXME: xarray might be space inefficient */
+	id = idr_alloc_cyclic(&fc->backing_files_map, fb, 1, 0, GFP_ATOMIC);
+	spin_unlock(&fc->lock);
+	idr_preload_end();
+
+	WARN_ON_ONCE(id == 0);
+	return id;
+}
+
+static struct fuse_backing *fuse_backing_id_remove(struct fuse_conn *fc,
+						   int id)
+{
+	struct fuse_backing *fb;
+
+	spin_lock(&fc->lock);
+	fb = idr_remove(&fc->backing_files_map, id);
+	spin_unlock(&fc->lock);
+
+	return fb;
+}
+
+static int fuse_backing_id_free(int id, void *p, void *data)
+{
+	struct fuse_backing *fb = p;
+
+	WARN_ON_ONCE(refcount_read(&fb->count) != 1);
+	fuse_backing_free(fb);
+	return 0;
+}
+
+void fuse_backing_files_free(struct fuse_conn *fc)
+{
+	idr_for_each(&fc->backing_files_map, fuse_backing_id_free, NULL);
+	idr_destroy(&fc->backing_files_map);
+}
+
+int fuse_backing_open(struct fuse_conn *fc, struct fuse_backing_map *map)
+{
+	struct file *file;
+	struct super_block *backing_sb;
+	struct fuse_backing *fb = NULL;
+	int res;
+
+	pr_debug("%s: fd=%d flags=0x%x\n", __func__, map->fd, map->flags);
+
+	/* TODO: relax CAP_SYS_ADMIN once backing files are visible to lsof */
+	res = -EPERM;
+	if (!fc->passthrough || !capable(CAP_SYS_ADMIN))
+		goto out;
+
+	res = -EINVAL;
+	if (map->flags || map->padding)
+		goto out;
+
+	file = fget(map->fd);
+	res = -EBADF;
+	if (!file)
+		goto out;
+
+	res = -EOPNOTSUPP;
+	if (!file->f_op->read_iter || !file->f_op->write_iter)
+		goto out_fput;
+
+	backing_sb = file_inode(file)->i_sb;
+	res = -ELOOP;
+	if (backing_sb->s_stack_depth >= fc->max_stack_depth)
+		goto out_fput;
+
+	fb = kmalloc(sizeof(struct fuse_backing), GFP_KERNEL);
+	res = -ENOMEM;
+	if (!fb)
+		goto out_fput;
+
+	fb->file = file;
+	fb->cred = prepare_creds();
+	refcount_set(&fb->count, 1);
+
+	res = fuse_backing_id_alloc(fc, fb);
+	if (res < 0) {
+		fuse_backing_free(fb);
+		fb = NULL;
+	}
+
+out:
+	pr_debug("%s: fb=0x%p, ret=%i\n", __func__, fb, res);
+
+	return res;
+
+out_fput:
+	fput(file);
+	goto out;
+}
+
+int fuse_backing_close(struct fuse_conn *fc, int backing_id)
+{
+	struct fuse_backing *fb = NULL;
+	int err;
+
+	pr_debug("%s: backing_id=%d\n", __func__, backing_id);
+
+	/* TODO: relax CAP_SYS_ADMIN once backing files are visible to lsof */
+	err = -EPERM;
+	if (!fc->passthrough || !capable(CAP_SYS_ADMIN))
+		goto out;
+
+	err = -EINVAL;
+	if (backing_id <= 0)
+		goto out;
+
+	err = -ENOENT;
+	fb = fuse_backing_id_remove(fc, backing_id);
+	if (!fb)
+		goto out;
+
+	fuse_backing_put(fb);
+	err = 0;
+out:
+	pr_debug("%s: fb=0x%p, err=%i\n", __func__, fb, err);
+
+	return err;
+}
+
+/*
+ * Setup passthrough to a backing file.
+ *
+ * Returns an fb object with elevated refcount to be stored in fuse inode.
+ */
+struct fuse_backing *fuse_passthrough_open(struct file *file,
+					   struct inode *inode,
+					   int backing_id)
+{
+	struct fuse_file *ff = file->private_data;
+	struct fuse_conn *fc = ff->fm->fc;
+	struct fuse_backing *fb = NULL;
+	struct file *backing_file;
+	int err;
+
+	err = -EINVAL;
+	if (backing_id <= 0)
+		goto out;
+
+	rcu_read_lock();
+	fb = idr_find(&fc->backing_files_map, backing_id);
+	fb = fuse_backing_get(fb);
+	rcu_read_unlock();
+
+	err = -ENOENT;
+	if (!fb)
+		goto out;
+
+	/* Allocate backing file per fuse file to store fuse path */
+	backing_file = backing_file_open(&file->f_path, file->f_flags,
+					 &fb->file->f_path, fb->cred);
+	err = PTR_ERR(backing_file);
+	if (IS_ERR(backing_file)) {
+		fuse_backing_put(fb);
+		goto out;
+	}
+
+	err = 0;
+	ff->passthrough = backing_file;
+	ff->cred = get_cred(fb->cred);
+out:
+	pr_debug("%s: backing_id=%d, fb=0x%p, backing_file=0x%p, err=%i\n", __func__,
+		 backing_id, fb, ff->passthrough, err);
+
+	return err ? ERR_PTR(err) : fb;
+}
+
+void fuse_passthrough_release(struct fuse_file *ff, struct fuse_backing *fb)
+{
+	pr_debug("%s: fb=0x%p, backing_file=0x%p\n", __func__,
+		 fb, ff->passthrough);
+
+	fput(ff->passthrough);
+	ff->passthrough = NULL;
+	put_cred(ff->cred);
+	ff->cred = NULL;
+}
diff --git a/fs/fuse/readdir.c b/fs/fuse/readdir.c
index c66a54d6c7..0377b6dc24 100644
--- a/fs/fuse/readdir.c
+++ b/fs/fuse/readdir.c
@@ -592,15 +592,11 @@ int fuse_readdir(struct file *file, struct dir_context *ctx)
 	if (fuse_is_bad(inode))
 		return -EIO;
 
-	mutex_lock(&ff->readdir.lock);
-
 	err = UNCACHED;
 	if (ff->open_flags & FOPEN_CACHE_DIR)
 		err = fuse_readdir_cached(file, ctx);
 	if (err == UNCACHED)
 		err = fuse_readdir_uncached(file, ctx);
 
-	mutex_unlock(&ff->readdir.lock);
-
 	return err;
 }
diff --git a/fs/fuse/virtio_fs.c b/fs/fuse/virtio_fs.c
index 5f1be1da92..1a52a51b6b 100644
--- a/fs/fuse/virtio_fs.c
+++ b/fs/fuse/virtio_fs.c
@@ -7,6 +7,8 @@
 #include <linux/fs.h>
 #include <linux/dax.h>
 #include <linux/pci.h>
+#include <linux/interrupt.h>
+#include <linux/group_cpus.h>
 #include <linux/pfn_t.h>
 #include <linux/memremap.h>
 #include <linux/module.h>
@@ -16,6 +18,7 @@
 #include <linux/fs_context.h>
 #include <linux/fs_parser.h>
 #include <linux/highmem.h>
+#include <linux/cleanup.h>
 #include <linux/uio.h>
 #include "fuse_i.h"
 
@@ -31,6 +34,9 @@
 static DEFINE_MUTEX(virtio_fs_mutex);
 static LIST_HEAD(virtio_fs_instances);
 
+/* The /sys/fs/virtio_fs/ kset */
+static struct kset *virtio_fs_kset;
+
 enum {
 	VQ_HIPRIO,
 	VQ_REQUEST
@@ -55,7 +61,7 @@ struct virtio_fs_vq {
 
 /* A virtio-fs device instance */
 struct virtio_fs {
-	struct kref refcount;
+	struct kobject kobj;
 	struct list_head list;    /* on virtio_fs_instances */
 	char *tag;
 	struct virtio_fs_vq *vqs;
@@ -63,6 +69,8 @@ struct virtio_fs {
 	unsigned int num_request_queues; /* number of request queues */
 	struct dax_device *dax_dev;
 
+	unsigned int *mq_map; /* index = cpu id, value = request vq id */
+
 	/* DAX memory window where file contents are mapped */
 	void *window_kaddr;
 	phys_addr_t window_phys_addr;
@@ -161,18 +169,41 @@ static inline void dec_in_flight_req(struct virtio_fs_vq *fsvq)
 		complete(&fsvq->in_flight_zero);
 }
 
-static void release_virtio_fs_obj(struct kref *ref)
+static ssize_t tag_show(struct kobject *kobj,
+		struct kobj_attribute *attr, char *buf)
+{
+	struct virtio_fs *fs = container_of(kobj, struct virtio_fs, kobj);
+
+	return sysfs_emit(buf, "%s\n", fs->tag);
+}
+
+static struct kobj_attribute virtio_fs_tag_attr = __ATTR_RO(tag);
+
+static struct attribute *virtio_fs_attrs[] = {
+	&virtio_fs_tag_attr.attr,
+	NULL
+};
+ATTRIBUTE_GROUPS(virtio_fs);
+
+static void virtio_fs_ktype_release(struct kobject *kobj)
 {
-	struct virtio_fs *vfs = container_of(ref, struct virtio_fs, refcount);
+	struct virtio_fs *vfs = container_of(kobj, struct virtio_fs, kobj);
 
+	kfree(vfs->mq_map);
 	kfree(vfs->vqs);
 	kfree(vfs);
 }
 
+static const struct kobj_type virtio_fs_ktype = {
+	.release = virtio_fs_ktype_release,
+	.sysfs_ops = &kobj_sysfs_ops,
+	.default_groups = virtio_fs_groups,
+};
+
 /* Make sure virtiofs_mutex is held */
 static void virtio_fs_put(struct virtio_fs *fs)
 {
-	kref_put(&fs->refcount, release_virtio_fs_obj);
+	kobject_put(&fs->kobj);
 }
 
 static void virtio_fs_fiq_release(struct fuse_iqueue *fiq)
@@ -243,25 +274,46 @@ static void virtio_fs_start_all_queues(struct virtio_fs *fs)
 }
 
 /* Add a new instance to the list or return -EEXIST if tag name exists*/
-static int virtio_fs_add_instance(struct virtio_fs *fs)
+static int virtio_fs_add_instance(struct virtio_device *vdev,
+				  struct virtio_fs *fs)
 {
 	struct virtio_fs *fs2;
-	bool duplicate = false;
+	int ret;
 
 	mutex_lock(&virtio_fs_mutex);
 
 	list_for_each_entry(fs2, &virtio_fs_instances, list) {
-		if (strcmp(fs->tag, fs2->tag) == 0)
-			duplicate = true;
+		if (strcmp(fs->tag, fs2->tag) == 0) {
+			mutex_unlock(&virtio_fs_mutex);
+			return -EEXIST;
+		}
 	}
 
-	if (!duplicate)
-		list_add_tail(&fs->list, &virtio_fs_instances);
+	/* Use the virtio_device's index as a unique identifier, there is no
+	 * need to allocate our own identifiers because the virtio_fs instance
+	 * is only visible to userspace as long as the underlying virtio_device
+	 * exists.
+	 */
+	fs->kobj.kset = virtio_fs_kset;
+	ret = kobject_add(&fs->kobj, NULL, "%d", vdev->index);
+	if (ret < 0) {
+		mutex_unlock(&virtio_fs_mutex);
+		return ret;
+	}
+
+	ret = sysfs_create_link(&fs->kobj, &vdev->dev.kobj, "device");
+	if (ret < 0) {
+		kobject_del(&fs->kobj);
+		mutex_unlock(&virtio_fs_mutex);
+		return ret;
+	}
+
+	list_add_tail(&fs->list, &virtio_fs_instances);
 
 	mutex_unlock(&virtio_fs_mutex);
 
-	if (duplicate)
-		return -EEXIST;
+	kobject_uevent(&fs->kobj, KOBJ_ADD);
+
 	return 0;
 }
 
@@ -274,7 +326,7 @@ static struct virtio_fs *virtio_fs_find_instance(const char *tag)
 
 	list_for_each_entry(fs, &virtio_fs_instances, list) {
 		if (strcmp(fs->tag, tag) == 0) {
-			kref_get(&fs->refcount);
+			kobject_get(&fs->kobj);
 			goto found;
 		}
 	}
@@ -323,6 +375,16 @@ static int virtio_fs_read_tag(struct virtio_device *vdev, struct virtio_fs *fs)
 		return -ENOMEM;
 	memcpy(fs->tag, tag_buf, len);
 	fs->tag[len] = '\0';
+
+	/* While the VIRTIO specification allows any character, newlines are
+	 * awkward on mount(8) command-lines and cause problems in the sysfs
+	 * "tag" attr and uevent TAG= properties. Forbid them.
+	 */
+	if (strchr(fs->tag, '\n')) {
+		dev_dbg(&vdev->dev, "refusing virtiofs tag with newline character\n");
+		return -EINVAL;
+	}
+
 	return 0;
 }
 
@@ -345,7 +407,7 @@ static void virtio_fs_hiprio_done_work(struct work_struct *work)
 			kfree(req);
 			dec_in_flight_req(fsvq);
 		}
-	} while (!virtqueue_enable_cb(vq) && likely(!virtqueue_is_broken(vq)));
+	} while (!virtqueue_enable_cb(vq));
 	spin_unlock(&fsvq->lock);
 }
 
@@ -627,7 +689,7 @@ static void virtio_fs_requests_done_work(struct work_struct *work)
 			list_move_tail(&req->list, &reqs);
 			spin_unlock(&fpq->lock);
 		}
-	} while (!virtqueue_enable_cb(vq) && likely(!virtqueue_is_broken(vq)));
+	} while (!virtqueue_enable_cb(vq));
 	spin_unlock(&fsvq->lock);
 
 	/* End requests */
@@ -649,6 +711,44 @@ static void virtio_fs_requests_done_work(struct work_struct *work)
 	}
 }
 
+static void virtio_fs_map_queues(struct virtio_device *vdev, struct virtio_fs *fs)
+{
+	const struct cpumask *mask, *masks;
+	unsigned int q, cpu;
+
+	/* First attempt to map using existing transport layer affinities
+	 * e.g. PCIe MSI-X
+	 */
+	if (!vdev->config->get_vq_affinity)
+		goto fallback;
+
+	for (q = 0; q < fs->num_request_queues; q++) {
+		mask = vdev->config->get_vq_affinity(vdev, VQ_REQUEST + q);
+		if (!mask)
+			goto fallback;
+
+		for_each_cpu(cpu, mask)
+			fs->mq_map[cpu] = q;
+	}
+
+	return;
+fallback:
+	/* Attempt to map evenly in groups over the CPUs */
+	masks = group_cpus_evenly(fs->num_request_queues);
+	/* If even this fails we default to all CPUs use queue zero */
+	if (!masks) {
+		for_each_possible_cpu(cpu)
+			fs->mq_map[cpu] = 0;
+		return;
+	}
+
+	for (q = 0; q < fs->num_request_queues; q++) {
+		for_each_cpu(cpu, &masks[q])
+			fs->mq_map[cpu] = q;
+	}
+	kfree(masks);
+}
+
 /* Virtqueue interrupt handler */
 static void virtio_fs_vq_done(struct virtqueue *vq)
 {
@@ -685,6 +785,11 @@ static int virtio_fs_setup_vqs(struct virtio_device *vdev,
 {
 	struct virtqueue **vqs;
 	vq_callback_t **callbacks;
+	/* Specify pre_vectors to ensure that the queues before the
+	 * request queues (e.g. hiprio) don't claim any of the CPUs in
+	 * the multi-queue mapping and interrupt affinities
+	 */
+	struct irq_affinity desc = { .pre_vectors = VQ_REQUEST };
 	const char **names;
 	unsigned int i;
 	int ret = 0;
@@ -694,6 +799,9 @@ static int virtio_fs_setup_vqs(struct virtio_device *vdev,
 	if (fs->num_request_queues == 0)
 		return -EINVAL;
 
+	/* Truncate nr of request queues to nr_cpu_id */
+	fs->num_request_queues = min_t(unsigned int, fs->num_request_queues,
+					nr_cpu_ids);
 	fs->nvqs = VQ_REQUEST + fs->num_request_queues;
 	fs->vqs = kcalloc(fs->nvqs, sizeof(fs->vqs[VQ_HIPRIO]), GFP_KERNEL);
 	if (!fs->vqs)
@@ -703,7 +811,9 @@ static int virtio_fs_setup_vqs(struct virtio_device *vdev,
 	callbacks = kmalloc_array(fs->nvqs, sizeof(callbacks[VQ_HIPRIO]),
 					GFP_KERNEL);
 	names = kmalloc_array(fs->nvqs, sizeof(names[VQ_HIPRIO]), GFP_KERNEL);
-	if (!vqs || !callbacks || !names) {
+	fs->mq_map = kcalloc_node(nr_cpu_ids, sizeof(*fs->mq_map), GFP_KERNEL,
+					dev_to_node(&vdev->dev));
+	if (!vqs || !callbacks || !names || !fs->mq_map) {
 		ret = -ENOMEM;
 		goto out;
 	}
@@ -723,7 +833,7 @@ static int virtio_fs_setup_vqs(struct virtio_device *vdev,
 		names[i] = fs->vqs[i].name;
 	}
 
-	ret = virtio_find_vqs(vdev, fs->nvqs, vqs, callbacks, names, NULL);
+	ret = virtio_find_vqs(vdev, fs->nvqs, vqs, callbacks, names, &desc);
 	if (ret < 0)
 		goto out;
 
@@ -735,8 +845,10 @@ out:
 	kfree(names);
 	kfree(callbacks);
 	kfree(vqs);
-	if (ret)
+	if (ret) {
 		kfree(fs->vqs);
+		kfree(fs->mq_map);
+	}
 	return ret;
 }
 
@@ -795,8 +907,11 @@ static void virtio_fs_cleanup_dax(void *data)
 	put_dax(dax_dev);
 }
 
+DEFINE_FREE(cleanup_dax, struct dax_dev *, if (!IS_ERR_OR_NULL(_T)) virtio_fs_cleanup_dax(_T))
+
 static int virtio_fs_setup_dax(struct virtio_device *vdev, struct virtio_fs *fs)
 {
+	struct dax_device *dax_dev __free(cleanup_dax) = NULL;
 	struct virtio_shm_region cache_reg;
 	struct dev_pagemap *pgmap;
 	bool have_cache;
@@ -804,6 +919,12 @@ static int virtio_fs_setup_dax(struct virtio_device *vdev, struct virtio_fs *fs)
 	if (!IS_ENABLED(CONFIG_FUSE_DAX))
 		return 0;
 
+	dax_dev = alloc_dax(fs, &virtio_fs_dax_ops);
+	if (IS_ERR(dax_dev)) {
+		int rc = PTR_ERR(dax_dev);
+		return rc == -EOPNOTSUPP ? 0 : rc;
+	}
+
 	/* Get cache region */
 	have_cache = virtio_get_shm_region(vdev, &cache_reg,
 					   (u8)VIRTIO_FS_SHMCAP_ID_CACHE);
@@ -849,10 +970,7 @@ static int virtio_fs_setup_dax(struct virtio_device *vdev, struct virtio_fs *fs)
 	dev_dbg(&vdev->dev, "%s: window kaddr 0x%px phys_addr 0x%llx len 0x%llx\n",
 		__func__, fs->window_kaddr, cache_reg.addr, cache_reg.len);
 
-	fs->dax_dev = alloc_dax(fs, &virtio_fs_dax_ops);
-	if (IS_ERR(fs->dax_dev))
-		return PTR_ERR(fs->dax_dev);
-
+	fs->dax_dev = no_free_ptr(dax_dev);
 	return devm_add_action_or_reset(&vdev->dev, virtio_fs_cleanup_dax,
 					fs->dax_dev);
 }
@@ -865,7 +983,7 @@ static int virtio_fs_probe(struct virtio_device *vdev)
 	fs = kzalloc(sizeof(*fs), GFP_KERNEL);
 	if (!fs)
 		return -ENOMEM;
-	kref_init(&fs->refcount);
+	kobject_init(&fs->kobj, &virtio_fs_ktype);
 	vdev->priv = fs;
 
 	ret = virtio_fs_read_tag(vdev, fs);
@@ -876,7 +994,7 @@ static int virtio_fs_probe(struct virtio_device *vdev)
 	if (ret < 0)
 		goto out;
 
-	/* TODO vq affinity */
+	virtio_fs_map_queues(vdev, fs);
 
 	ret = virtio_fs_setup_dax(vdev, fs);
 	if (ret < 0)
@@ -887,7 +1005,7 @@ static int virtio_fs_probe(struct virtio_device *vdev)
 	 */
 	virtio_device_ready(vdev);
 
-	ret = virtio_fs_add_instance(fs);
+	ret = virtio_fs_add_instance(vdev, fs);
 	if (ret < 0)
 		goto out_vqs;
 
@@ -896,11 +1014,10 @@ static int virtio_fs_probe(struct virtio_device *vdev)
 out_vqs:
 	virtio_reset_device(vdev);
 	virtio_fs_cleanup_vqs(vdev);
-	kfree(fs->vqs);
 
 out:
 	vdev->priv = NULL;
-	kfree(fs);
+	kobject_put(&fs->kobj);
 	return ret;
 }
 
@@ -924,6 +1041,8 @@ static void virtio_fs_remove(struct virtio_device *vdev)
 	mutex_lock(&virtio_fs_mutex);
 	/* This device is going away. No one should get new reference */
 	list_del_init(&fs->list);
+	sysfs_remove_link(&fs->kobj, "device");
+	kobject_del(&fs->kobj);
 	virtio_fs_stop_all_queues(fs);
 	virtio_fs_drain_all_queues_locked(fs);
 	virtio_reset_device(vdev);
@@ -959,7 +1078,6 @@ static const unsigned int feature_table[] = {};
 
 static struct virtio_driver virtio_fs_driver = {
 	.driver.name		= KBUILD_MODNAME,
-	.driver.owner		= THIS_MODULE,
 	.id_table		= id_table,
 	.feature_table		= feature_table,
 	.feature_table_size	= ARRAY_SIZE(feature_table),
@@ -1224,7 +1342,7 @@ out:
 static void virtio_fs_wake_pending_and_unlock(struct fuse_iqueue *fiq)
 __releases(fiq->lock)
 {
-	unsigned int queue_id = VQ_REQUEST; /* TODO multiqueue */
+	unsigned int queue_id;
 	struct virtio_fs *fs;
 	struct fuse_req *req;
 	struct virtio_fs_vq *fsvq;
@@ -1238,11 +1356,13 @@ __releases(fiq->lock)
 	spin_unlock(&fiq->lock);
 
 	fs = fiq->priv;
+	queue_id = VQ_REQUEST + fs->mq_map[raw_smp_processor_id()];
 
-	pr_debug("%s: opcode %u unique %#llx nodeid %#llx in.len %u out.len %u\n",
-		  __func__, req->in.h.opcode, req->in.h.unique,
+	pr_debug("%s: opcode %u unique %#llx nodeid %#llx in.len %u out.len %u queue_id %u\n",
+		 __func__, req->in.h.opcode, req->in.h.unique,
 		 req->in.h.nodeid, req->in.h.len,
-		 fuse_len_args(req->args->out_numargs, req->args->out_args));
+		 fuse_len_args(req->args->out_numargs, req->args->out_args),
+		 queue_id);
 
 	fsvq = &fs->vqs[queue_id];
 	ret = virtio_fs_enqueue_req(fsvq, req, false);
@@ -1510,21 +1630,56 @@ static struct file_system_type virtio_fs_type = {
 	.kill_sb	= virtio_kill_sb,
 };
 
+static int virtio_fs_uevent(const struct kobject *kobj, struct kobj_uevent_env *env)
+{
+	const struct virtio_fs *fs = container_of(kobj, struct virtio_fs, kobj);
+
+	add_uevent_var(env, "TAG=%s", fs->tag);
+	return 0;
+}
+
+static const struct kset_uevent_ops virtio_fs_uevent_ops = {
+	.uevent = virtio_fs_uevent,
+};
+
+static int __init virtio_fs_sysfs_init(void)
+{
+	virtio_fs_kset = kset_create_and_add("virtiofs", &virtio_fs_uevent_ops,
+					     fs_kobj);
+	if (!virtio_fs_kset)
+		return -ENOMEM;
+	return 0;
+}
+
+static void virtio_fs_sysfs_exit(void)
+{
+	kset_unregister(virtio_fs_kset);
+	virtio_fs_kset = NULL;
+}
+
 static int __init virtio_fs_init(void)
 {
 	int ret;
 
-	ret = register_virtio_driver(&virtio_fs_driver);
+	ret = virtio_fs_sysfs_init();
 	if (ret < 0)
 		return ret;
 
+	ret = register_virtio_driver(&virtio_fs_driver);
+	if (ret < 0)
+		goto sysfs_exit;
+
 	ret = register_filesystem(&virtio_fs_type);
-	if (ret < 0) {
-		unregister_virtio_driver(&virtio_fs_driver);
-		return ret;
-	}
+	if (ret < 0)
+		goto unregister_virtio_driver;
 
 	return 0;
+
+unregister_virtio_driver:
+	unregister_virtio_driver(&virtio_fs_driver);
+sysfs_exit:
+	virtio_fs_sysfs_exit();
+	return ret;
 }
 module_init(virtio_fs_init);
 
@@ -1532,6 +1687,7 @@ static void __exit virtio_fs_exit(void)
 {
 	unregister_filesystem(&virtio_fs_type);
 	unregister_virtio_driver(&virtio_fs_driver);
+	virtio_fs_sysfs_exit();
 }
 module_exit(virtio_fs_exit);
 
diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c
index 974aca9c8e..10d5acd3f7 100644
--- a/fs/gfs2/aops.c
+++ b/fs/gfs2/aops.c
@@ -116,8 +116,7 @@ static int gfs2_write_jdata_folio(struct folio *folio,
  * @folio: The folio to write
  * @wbc: The writeback control
  *
- * This is shared between writepage and writepages and implements the
- * core of the writepage operation. If a transaction is required then
+ * Implements the core of write back. If a transaction is required then
  * the checked flag will have been set and the transaction will have
  * already been started before this is called.
  */
@@ -755,6 +754,7 @@ static const struct address_space_operations gfs2_jdata_aops = {
 	.readahead = gfs2_readahead,
 	.dirty_folio = jdata_dirty_folio,
 	.bmap = gfs2_bmap,
+	.migrate_folio = buffer_migrate_folio,
 	.invalidate_folio = gfs2_invalidate_folio,
 	.release_folio = gfs2_release_folio,
 	.is_partially_uptodate = block_is_partially_uptodate,
diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c
index 643175498d..1795c4e8db 100644
--- a/fs/gfs2/bmap.c
+++ b/fs/gfs2/bmap.c
@@ -1827,7 +1827,7 @@ static int punch_hole(struct gfs2_inode *ip, u64 offset, u64 length)
 			gfs2_assert_withdraw(sdp, bh);
 			if (gfs2_assert_withdraw(sdp,
 						 prev_bnr != bh->b_blocknr)) {
-				fs_emerg(sdp, "inode %llu, block:%llu, i_h:%u,"
+				fs_emerg(sdp, "inode %llu, block:%llu, i_h:%u, "
 					 "s_h:%u, mp_h:%u\n",
 				       (unsigned long long)ip->i_no_addr,
 				       prev_bnr, ip->i_height, strip_h, mp_h);
@@ -2466,7 +2466,7 @@ out:
 }
 
 static int gfs2_map_blocks(struct iomap_writepage_ctx *wpc, struct inode *inode,
-		loff_t offset)
+		loff_t offset, unsigned int len)
 {
 	int ret;
 
diff --git a/fs/gfs2/dir.c b/fs/gfs2/dir.c
index 560e4624c0..dbf1aede74 100644
--- a/fs/gfs2/dir.c
+++ b/fs/gfs2/dir.c
@@ -562,15 +562,18 @@ static struct gfs2_dirent *gfs2_dirent_scan(struct inode *inode, void *buf,
 	int ret = 0;
 
 	ret = gfs2_dirent_offset(GFS2_SB(inode), buf);
-	if (ret < 0)
-		goto consist_inode;
-
+	if (ret < 0) {
+		gfs2_consist_inode(GFS2_I(inode));
+		return ERR_PTR(-EIO);
+	}
 	offset = ret;
 	prev = NULL;
 	dent = buf + offset;
 	size = be16_to_cpu(dent->de_rec_len);
-	if (gfs2_check_dirent(GFS2_SB(inode), dent, offset, size, len, 1))
-		goto consist_inode;
+	if (gfs2_check_dirent(GFS2_SB(inode), dent, offset, size, len, 1)) {
+		gfs2_consist_inode(GFS2_I(inode));
+		return ERR_PTR(-EIO);
+	}
 	do {
 		ret = scan(dent, name, opaque);
 		if (ret)
@@ -582,8 +585,10 @@ static struct gfs2_dirent *gfs2_dirent_scan(struct inode *inode, void *buf,
 		dent = buf + offset;
 		size = be16_to_cpu(dent->de_rec_len);
 		if (gfs2_check_dirent(GFS2_SB(inode), dent, offset, size,
-				      len, 0))
-			goto consist_inode;
+				      len, 0)) {
+			gfs2_consist_inode(GFS2_I(inode));
+			return ERR_PTR(-EIO);
+		}
 	} while(1);
 
 	switch(ret) {
@@ -597,10 +602,6 @@ static struct gfs2_dirent *gfs2_dirent_scan(struct inode *inode, void *buf,
 		BUG_ON(ret > 0);
 		return ERR_PTR(ret);
 	}
-
-consist_inode:
-	gfs2_consist_inode(GFS2_I(inode));
-	return ERR_PTR(-EIO);
 }
 
 static int dirent_check_reclen(struct gfs2_inode *dip,
@@ -609,14 +610,16 @@ static int dirent_check_reclen(struct gfs2_inode *dip,
 	const void *ptr = d;
 	u16 rec_len = be16_to_cpu(d->de_rec_len);
 
-	if (unlikely(rec_len < sizeof(struct gfs2_dirent)))
-		goto broken;
+	if (unlikely(rec_len < sizeof(struct gfs2_dirent))) {
+		gfs2_consist_inode(dip);
+		return -EIO;
+	}
 	ptr += rec_len;
 	if (ptr < end_p)
 		return rec_len;
 	if (ptr == end_p)
 		return -ENOENT;
-broken:
+
 	gfs2_consist_inode(dip);
 	return -EIO;
 }
diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c
index 992ca4effb..08982937b5 100644
--- a/fs/gfs2/file.c
+++ b/fs/gfs2/file.c
@@ -376,23 +376,23 @@ static void gfs2_size_hint(struct file *filep, loff_t offset, size_t size)
 }
 
 /**
- * gfs2_allocate_page_backing - Allocate blocks for a write fault
- * @page: The (locked) page to allocate backing for
+ * gfs2_allocate_folio_backing - Allocate blocks for a write fault
+ * @folio: The (locked) folio to allocate backing for
  * @length: Size of the allocation
  *
- * We try to allocate all the blocks required for the page in one go.  This
+ * We try to allocate all the blocks required for the folio in one go.  This
  * might fail for various reasons, so we keep trying until all the blocks to
- * back this page are allocated.  If some of the blocks are already allocated,
+ * back this folio are allocated.  If some of the blocks are already allocated,
  * that is ok too.
  */
-static int gfs2_allocate_page_backing(struct page *page, unsigned int length)
+static int gfs2_allocate_folio_backing(struct folio *folio, size_t length)
 {
-	u64 pos = page_offset(page);
+	u64 pos = folio_pos(folio);
 
 	do {
 		struct iomap iomap = { };
 
-		if (gfs2_iomap_alloc(page->mapping->host, pos, length, &iomap))
+		if (gfs2_iomap_alloc(folio->mapping->host, pos, length, &iomap))
 			return -EIO;
 
 		if (length < iomap.length)
@@ -414,16 +414,16 @@ static int gfs2_allocate_page_backing(struct page *page, unsigned int length)
 
 static vm_fault_t gfs2_page_mkwrite(struct vm_fault *vmf)
 {
-	struct page *page = vmf->page;
+	struct folio *folio = page_folio(vmf->page);
 	struct inode *inode = file_inode(vmf->vma->vm_file);
 	struct gfs2_inode *ip = GFS2_I(inode);
 	struct gfs2_sbd *sdp = GFS2_SB(inode);
 	struct gfs2_alloc_parms ap = {};
-	u64 offset = page_offset(page);
+	u64 pos = folio_pos(folio);
 	unsigned int data_blocks, ind_blocks, rblocks;
 	vm_fault_t ret = VM_FAULT_LOCKED;
 	struct gfs2_holder gh;
-	unsigned int length;
+	size_t length;
 	loff_t size;
 	int err;
 
@@ -436,23 +436,23 @@ static vm_fault_t gfs2_page_mkwrite(struct vm_fault *vmf)
 		goto out_uninit;
 	}
 
-	/* Check page index against inode size */
+	/* Check folio index against inode size */
 	size = i_size_read(inode);
-	if (offset >= size) {
+	if (pos >= size) {
 		ret = VM_FAULT_SIGBUS;
 		goto out_unlock;
 	}
 
-	/* Update file times before taking page lock */
+	/* Update file times before taking folio lock */
 	file_update_time(vmf->vma->vm_file);
 
-	/* page is wholly or partially inside EOF */
-	if (size - offset < PAGE_SIZE)
-		length = size - offset;
+	/* folio is wholly or partially inside EOF */
+	if (size - pos < folio_size(folio))
+		length = size - pos;
 	else
-		length = PAGE_SIZE;
+		length = folio_size(folio);
 
-	gfs2_size_hint(vmf->vma->vm_file, offset, length);
+	gfs2_size_hint(vmf->vma->vm_file, pos, length);
 
 	set_bit(GLF_DIRTY, &ip->i_gl->gl_flags);
 	set_bit(GIF_SW_PAGED, &ip->i_flags);
@@ -463,11 +463,12 @@ static vm_fault_t gfs2_page_mkwrite(struct vm_fault *vmf)
 	 */
 
 	if (!gfs2_is_stuffed(ip) &&
-	    !gfs2_write_alloc_required(ip, offset, length)) {
-		lock_page(page);
-		if (!PageUptodate(page) || page->mapping != inode->i_mapping) {
+	    !gfs2_write_alloc_required(ip, pos, length)) {
+		folio_lock(folio);
+		if (!folio_test_uptodate(folio) ||
+		    folio->mapping != inode->i_mapping) {
 			ret = VM_FAULT_NOPAGE;
-			unlock_page(page);
+			folio_unlock(folio);
 		}
 		goto out_unlock;
 	}
@@ -504,7 +505,7 @@ static vm_fault_t gfs2_page_mkwrite(struct vm_fault *vmf)
 		goto out_trans_fail;
 	}
 
-	/* Unstuff, if required, and allocate backing blocks for page */
+	/* Unstuff, if required, and allocate backing blocks for folio */
 	if (gfs2_is_stuffed(ip)) {
 		err = gfs2_unstuff_dinode(ip);
 		if (err) {
@@ -513,22 +514,22 @@ static vm_fault_t gfs2_page_mkwrite(struct vm_fault *vmf)
 		}
 	}
 
-	lock_page(page);
+	folio_lock(folio);
 	/* If truncated, we must retry the operation, we may have raced
 	 * with the glock demotion code.
 	 */
-	if (!PageUptodate(page) || page->mapping != inode->i_mapping) {
+	if (!folio_test_uptodate(folio) || folio->mapping != inode->i_mapping) {
 		ret = VM_FAULT_NOPAGE;
 		goto out_page_locked;
 	}
 
-	err = gfs2_allocate_page_backing(page, length);
+	err = gfs2_allocate_folio_backing(folio, length);
 	if (err)
 		ret = vmf_fs_error(err);
 
 out_page_locked:
 	if (ret != VM_FAULT_LOCKED)
-		unlock_page(page);
+		folio_unlock(folio);
 out_trans_end:
 	gfs2_trans_end(sdp);
 out_trans_fail:
@@ -540,8 +541,8 @@ out_unlock:
 out_uninit:
 	gfs2_holder_uninit(&gh);
 	if (ret == VM_FAULT_LOCKED) {
-		set_page_dirty(page);
-		wait_for_stable_page(page);
+		folio_mark_dirty(folio);
+		folio_wait_stable(folio);
 	}
 	sb_end_pagefault(inode->i_sb);
 	return ret;
@@ -1440,10 +1441,10 @@ static int gfs2_lock(struct file *file, int cmd, struct file_lock *fl)
 	struct gfs2_sbd *sdp = GFS2_SB(file->f_mapping->host);
 	struct lm_lockstruct *ls = &sdp->sd_lockstruct;
 
-	if (!(fl->fl_flags & FL_POSIX))
+	if (!(fl->c.flc_flags & FL_POSIX))
 		return -ENOLCK;
 	if (gfs2_withdrawing_or_withdrawn(sdp)) {
-		if (fl->fl_type == F_UNLCK)
+		if (lock_is_unlock(fl))
 			locks_lock_file_wait(file, fl);
 		return -EIO;
 	}
@@ -1451,7 +1452,7 @@ static int gfs2_lock(struct file *file, int cmd, struct file_lock *fl)
 		return dlm_posix_cancel(ls->ls_dlm, ip->i_no_addr, file, fl);
 	else if (IS_GETLK(cmd))
 		return dlm_posix_get(ls->ls_dlm, ip->i_no_addr, file, fl);
-	else if (fl->fl_type == F_UNLCK)
+	else if (lock_is_unlock(fl))
 		return dlm_posix_unlock(ls->ls_dlm, ip->i_no_addr, file, fl);
 	else
 		return dlm_posix_lock(ls->ls_dlm, ip->i_no_addr, file, cmd, fl);
@@ -1483,7 +1484,7 @@ static int do_flock(struct file *file, int cmd, struct file_lock *fl)
 	int error = 0;
 	int sleeptime;
 
-	state = (fl->fl_type == F_WRLCK) ? LM_ST_EXCLUSIVE : LM_ST_SHARED;
+	state = lock_is_write(fl) ? LM_ST_EXCLUSIVE : LM_ST_SHARED;
 	flags = GL_EXACT | GL_NOPID;
 	if (!IS_SETLKW(cmd))
 		flags |= LM_FLAG_TRY_1CB;
@@ -1495,8 +1496,8 @@ static int do_flock(struct file *file, int cmd, struct file_lock *fl)
 		if (fl_gh->gh_state == state)
 			goto out;
 		locks_init_lock(&request);
-		request.fl_type = F_UNLCK;
-		request.fl_flags = FL_FLOCK;
+		request.c.flc_type = F_UNLCK;
+		request.c.flc_flags = FL_FLOCK;
 		locks_lock_file_wait(file, &request);
 		gfs2_glock_dq(fl_gh);
 		gfs2_holder_reinit(state, flags, fl_gh);
@@ -1557,10 +1558,10 @@ static void do_unflock(struct file *file, struct file_lock *fl)
 
 static int gfs2_flock(struct file *file, int cmd, struct file_lock *fl)
 {
-	if (!(fl->fl_flags & FL_FLOCK))
+	if (!(fl->c.flc_flags & FL_FLOCK))
 		return -ENOLCK;
 
-	if (fl->fl_type == F_UNLCK) {
+	if (lock_is_unlock(fl)) {
 		do_unflock(file, fl);
 		return 0;
 	} else {
diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c
index 2507fe34cb..4ea6c8bfb4 100644
--- a/fs/gfs2/glock.c
+++ b/fs/gfs2/glock.c
@@ -274,7 +274,7 @@ static void gfs2_glock_remove_from_lru(struct gfs2_glock *gl)
  * Enqueue the glock on the work queue.  Passes one glock reference on to the
  * work queue.
  */
-static void __gfs2_glock_queue_work(struct gfs2_glock *gl, unsigned long delay) {
+static void gfs2_glock_queue_work(struct gfs2_glock *gl, unsigned long delay) {
 	if (!queue_delayed_work(glock_workqueue, &gl->gl_work, delay)) {
 		/*
 		 * We are holding the lockref spinlock, and the work was still
@@ -287,12 +287,6 @@ static void __gfs2_glock_queue_work(struct gfs2_glock *gl, unsigned long delay)
 	}
 }
 
-static void gfs2_glock_queue_work(struct gfs2_glock *gl, unsigned long delay) {
-	spin_lock(&gl->gl_lockref.lock);
-	__gfs2_glock_queue_work(gl, delay);
-	spin_unlock(&gl->gl_lockref.lock);
-}
-
 static void __gfs2_glock_put(struct gfs2_glock *gl)
 {
 	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
@@ -311,14 +305,6 @@ static void __gfs2_glock_put(struct gfs2_glock *gl)
 	sdp->sd_lockstruct.ls_ops->lm_put_lock(gl);
 }
 
-/*
- * Cause the glock to be put in work queue context.
- */
-void gfs2_glock_queue_put(struct gfs2_glock *gl)
-{
-	gfs2_glock_queue_work(gl, 0);
-}
-
 /**
  * gfs2_glock_put() - Decrement reference count on glock
  * @gl: The glock to put
@@ -333,6 +319,23 @@ void gfs2_glock_put(struct gfs2_glock *gl)
 	__gfs2_glock_put(gl);
 }
 
+/*
+ * gfs2_glock_put_async - Decrement reference count without sleeping
+ * @gl: The glock to put
+ *
+ * Decrement the reference count on glock immediately unless it is the last
+ * reference.  Defer putting the last reference to work queue context.
+ */
+void gfs2_glock_put_async(struct gfs2_glock *gl)
+{
+	if (lockref_put_or_lock(&gl->gl_lockref))
+		return;
+
+	GLOCK_BUG_ON(gl, gl->gl_lockref.count != 1);
+	gfs2_glock_queue_work(gl, 0);
+	spin_unlock(&gl->gl_lockref.lock);
+}
+
 /**
  * may_grant - check if it's ok to grant a new lock
  * @gl: The glock
@@ -823,7 +826,7 @@ skip_inval:
 			 */
 			clear_bit(GLF_LOCK, &gl->gl_flags);
 			clear_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags);
-			__gfs2_glock_queue_work(gl, GL_GLOCK_DFT_HOLD);
+			gfs2_glock_queue_work(gl, GL_GLOCK_DFT_HOLD);
 			return;
 		} else {
 			clear_bit(GLF_INVALIDATE_IN_PROGRESS, &gl->gl_flags);
@@ -853,7 +856,7 @@ skip_inval:
 
 	/* Complete the operation now. */
 	finish_xmote(gl, target);
-	__gfs2_glock_queue_work(gl, 0);
+	gfs2_glock_queue_work(gl, 0);
 }
 
 /**
@@ -869,8 +872,9 @@ __acquires(&gl->gl_lockref.lock)
 {
 	struct gfs2_holder *gh = NULL;
 
-	if (test_and_set_bit(GLF_LOCK, &gl->gl_flags))
+	if (test_bit(GLF_LOCK, &gl->gl_flags))
 		return;
+	set_bit(GLF_LOCK, &gl->gl_flags);
 
 	GLOCK_BUG_ON(gl, test_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags));
 
@@ -900,7 +904,7 @@ out_sched:
 	clear_bit(GLF_LOCK, &gl->gl_flags);
 	smp_mb__after_atomic();
 	gl->gl_lockref.count++;
-	__gfs2_glock_queue_work(gl, 0);
+	gfs2_glock_queue_work(gl, 0);
 	return;
 
 out_unlock:
@@ -1132,12 +1136,12 @@ static void glock_work_func(struct work_struct *work)
 		drop_refs--;
 		if (gl->gl_name.ln_type != LM_TYPE_INODE)
 			delay = 0;
-		__gfs2_glock_queue_work(gl, delay);
+		gfs2_glock_queue_work(gl, delay);
 	}
 
 	/*
 	 * Drop the remaining glock references manually here. (Mind that
-	 * __gfs2_glock_queue_work depends on the lockref spinlock begin held
+	 * gfs2_glock_queue_work depends on the lockref spinlock begin held
 	 * here as well.)
 	 */
 	gl->gl_lockref.count -= drop_refs;
@@ -1263,7 +1267,7 @@ int gfs2_glock_get(struct gfs2_sbd *sdp, u64 number,
 	mapping = gfs2_glock2aspace(gl);
 	if (mapping) {
                 mapping->a_ops = &gfs2_meta_aops;
-		mapping->host = s->s_bdev->bd_inode;
+		mapping->host = s->s_bdev->bd_mapping->host;
 		mapping->flags = 0;
 		mapping_set_gfp_mask(mapping, GFP_NOFS);
 		mapping->i_private_data = NULL;
@@ -1642,7 +1646,7 @@ unlock:
 		     test_and_clear_bit(GLF_FROZEN, &gl->gl_flags))) {
 		set_bit(GLF_REPLY_PENDING, &gl->gl_flags);
 		gl->gl_lockref.count++;
-		__gfs2_glock_queue_work(gl, 0);
+		gfs2_glock_queue_work(gl, 0);
 	}
 	run_queue(gl, 1);
 	spin_unlock(&gl->gl_lockref.lock);
@@ -1708,7 +1712,7 @@ static void __gfs2_glock_dq(struct gfs2_holder *gh)
 		    !test_bit(GLF_DEMOTE, &gl->gl_flags) &&
 		    gl->gl_name.ln_type == LM_TYPE_INODE)
 			delay = gl->gl_hold_time;
-		__gfs2_glock_queue_work(gl, delay);
+		gfs2_glock_queue_work(gl, delay);
 	}
 }
 
@@ -1932,7 +1936,7 @@ void gfs2_glock_cb(struct gfs2_glock *gl, unsigned int state)
 			delay = gl->gl_hold_time;
 	}
 	handle_callback(gl, state, delay, true);
-	__gfs2_glock_queue_work(gl, delay);
+	gfs2_glock_queue_work(gl, delay);
 	spin_unlock(&gl->gl_lockref.lock);
 }
 
@@ -1992,7 +1996,7 @@ void gfs2_glock_complete(struct gfs2_glock *gl, int ret)
 
 	gl->gl_lockref.count++;
 	set_bit(GLF_REPLY_PENDING, &gl->gl_flags);
-	__gfs2_glock_queue_work(gl, 0);
+	gfs2_glock_queue_work(gl, 0);
 	spin_unlock(&gl->gl_lockref.lock);
 }
 
@@ -2012,6 +2016,14 @@ static int glock_cmp(void *priv, const struct list_head *a,
 	return 0;
 }
 
+static bool can_free_glock(struct gfs2_glock *gl)
+{
+	bool held = gl->gl_state != LM_ST_UNLOCKED;
+
+	return !test_bit(GLF_LOCK, &gl->gl_flags) &&
+	       gl->gl_lockref.count == held;
+}
+
 /**
  * gfs2_dispose_glock_lru - Demote a list of glocks
  * @list: The list to dispose of
@@ -2026,37 +2038,38 @@ static int glock_cmp(void *priv, const struct list_head *a,
  * private)
  */
 
-static void gfs2_dispose_glock_lru(struct list_head *list)
+static unsigned long gfs2_dispose_glock_lru(struct list_head *list)
 __releases(&lru_lock)
 __acquires(&lru_lock)
 {
 	struct gfs2_glock *gl;
+	unsigned long freed = 0;
 
 	list_sort(NULL, list, glock_cmp);
 
 	while(!list_empty(list)) {
 		gl = list_first_entry(list, struct gfs2_glock, gl_lru);
-		list_del_init(&gl->gl_lru);
-		clear_bit(GLF_LRU, &gl->gl_flags);
 		if (!spin_trylock(&gl->gl_lockref.lock)) {
 add_back_to_lru:
-			list_add(&gl->gl_lru, &lru_list);
-			set_bit(GLF_LRU, &gl->gl_flags);
-			atomic_inc(&lru_count);
+			list_move(&gl->gl_lru, &lru_list);
 			continue;
 		}
-		if (test_and_set_bit(GLF_LOCK, &gl->gl_flags)) {
+		if (!can_free_glock(gl)) {
 			spin_unlock(&gl->gl_lockref.lock);
 			goto add_back_to_lru;
 		}
+		list_del_init(&gl->gl_lru);
+		atomic_dec(&lru_count);
+		clear_bit(GLF_LRU, &gl->gl_flags);
+		freed++;
 		gl->gl_lockref.count++;
 		if (demote_ok(gl))
 			handle_callback(gl, LM_ST_UNLOCKED, 0, false);
-		WARN_ON(!test_and_clear_bit(GLF_LOCK, &gl->gl_flags));
-		__gfs2_glock_queue_work(gl, 0);
+		gfs2_glock_queue_work(gl, 0);
 		spin_unlock(&gl->gl_lockref.lock);
 		cond_resched_lock(&lru_lock);
 	}
+	return freed;
 }
 
 /**
@@ -2068,32 +2081,21 @@ add_back_to_lru:
  * gfs2_dispose_glock_lru() above.
  */
 
-static long gfs2_scan_glock_lru(int nr)
+static unsigned long gfs2_scan_glock_lru(unsigned long nr)
 {
 	struct gfs2_glock *gl, *next;
 	LIST_HEAD(dispose);
-	long freed = 0;
+	unsigned long freed = 0;
 
 	spin_lock(&lru_lock);
 	list_for_each_entry_safe(gl, next, &lru_list, gl_lru) {
-		if (nr-- <= 0)
+		if (!nr--)
 			break;
-		/* Test for being demotable */
-		if (!test_bit(GLF_LOCK, &gl->gl_flags)) {
-			if (!spin_trylock(&gl->gl_lockref.lock))
-				continue;
-			if (gl->gl_lockref.count <= 1 &&
-			    (gl->gl_state == LM_ST_UNLOCKED ||
-			     demote_ok(gl))) {
-				list_move(&gl->gl_lru, &dispose);
-				atomic_dec(&lru_count);
-				freed++;
-			}
-			spin_unlock(&gl->gl_lockref.lock);
-		}
+		if (can_free_glock(gl))
+			list_move(&gl->gl_lru, &dispose);
 	}
 	if (!list_empty(&dispose))
-		gfs2_dispose_glock_lru(&dispose);
+		freed = gfs2_dispose_glock_lru(&dispose);
 	spin_unlock(&lru_lock);
 
 	return freed;
@@ -2187,7 +2189,7 @@ static void thaw_glock(struct gfs2_glock *gl)
 
 	spin_lock(&gl->gl_lockref.lock);
 	set_bit(GLF_REPLY_PENDING, &gl->gl_flags);
-	__gfs2_glock_queue_work(gl, 0);
+	gfs2_glock_queue_work(gl, 0);
 	spin_unlock(&gl->gl_lockref.lock);
 }
 
@@ -2206,7 +2208,7 @@ static void clear_glock(struct gfs2_glock *gl)
 		gl->gl_lockref.count++;
 		if (gl->gl_state != LM_ST_UNLOCKED)
 			handle_callback(gl, LM_ST_UNLOCKED, 0, false);
-		__gfs2_glock_queue_work(gl, 0);
+		gfs2_glock_queue_work(gl, 0);
 	}
 	spin_unlock(&gl->gl_lockref.lock);
 }
@@ -2570,8 +2572,7 @@ static void gfs2_glock_iter_next(struct gfs2_glock_iter *gi, loff_t n)
 	if (gl) {
 		if (n == 0)
 			return;
-		if (!lockref_put_not_zero(&gl->gl_lockref))
-			gfs2_glock_queue_put(gl);
+		gfs2_glock_put_async(gl);
 	}
 	for (;;) {
 		gl = rhashtable_walk_next(&gi->hti);
diff --git a/fs/gfs2/glock.h b/fs/gfs2/glock.h
index 86987a59a0..19aef6d532 100644
--- a/fs/gfs2/glock.h
+++ b/fs/gfs2/glock.h
@@ -172,7 +172,7 @@ int gfs2_glock_get(struct gfs2_sbd *sdp, u64 number,
 		   int create, struct gfs2_glock **glp);
 struct gfs2_glock *gfs2_glock_hold(struct gfs2_glock *gl);
 void gfs2_glock_put(struct gfs2_glock *gl);
-void gfs2_glock_queue_put(struct gfs2_glock *gl);
+void gfs2_glock_put_async(struct gfs2_glock *gl);
 
 void __gfs2_holder_init(struct gfs2_glock *gl, unsigned int state,
 		        u16 flags, struct gfs2_holder *gh,
diff --git a/fs/gfs2/glops.c b/fs/gfs2/glops.c
index e0e8dfeee7..68677fb69a 100644
--- a/fs/gfs2/glops.c
+++ b/fs/gfs2/glops.c
@@ -412,10 +412,14 @@ static int gfs2_dinode_in(struct gfs2_inode *ip, const void *buf)
 	struct inode *inode = &ip->i_inode;
 	bool is_new = inode->i_state & I_NEW;
 
-	if (unlikely(ip->i_no_addr != be64_to_cpu(str->di_num.no_addr)))
-		goto corrupt;
-	if (unlikely(!is_new && inode_wrong_type(inode, mode)))
-		goto corrupt;
+	if (unlikely(ip->i_no_addr != be64_to_cpu(str->di_num.no_addr))) {
+		gfs2_consist_inode(ip);
+		return -EIO;
+	}
+	if (unlikely(!is_new && inode_wrong_type(inode, mode))) {
+		gfs2_consist_inode(ip);
+		return -EIO;
+	}
 	ip->i_no_formal_ino = be64_to_cpu(str->di_num.no_formal_ino);
 	inode->i_mode = mode;
 	if (is_new) {
@@ -452,26 +456,28 @@ static int gfs2_dinode_in(struct gfs2_inode *ip, const void *buf)
 	/* i_diskflags and i_eattr must be set before gfs2_set_inode_flags() */
 	gfs2_set_inode_flags(inode);
 	height = be16_to_cpu(str->di_height);
-	if (unlikely(height > sdp->sd_max_height))
-		goto corrupt;
+	if (unlikely(height > sdp->sd_max_height)) {
+		gfs2_consist_inode(ip);
+		return -EIO;
+	}
 	ip->i_height = (u8)height;
 
 	depth = be16_to_cpu(str->di_depth);
-	if (unlikely(depth > GFS2_DIR_MAX_DEPTH))
-		goto corrupt;
+	if (unlikely(depth > GFS2_DIR_MAX_DEPTH)) {
+		gfs2_consist_inode(ip);
+		return -EIO;
+	}
 	ip->i_depth = (u8)depth;
 	ip->i_entries = be32_to_cpu(str->di_entries);
 
-	if (gfs2_is_stuffed(ip) && inode->i_size > gfs2_max_stuffed_size(ip))
-		goto corrupt;
-
+	if (gfs2_is_stuffed(ip) && inode->i_size > gfs2_max_stuffed_size(ip)) {
+		gfs2_consist_inode(ip);
+		return -EIO;
+	}
 	if (S_ISREG(inode->i_mode))
 		gfs2_set_aops(inode);
 
 	return 0;
-corrupt:
-	gfs2_consist_inode(ip);
-	return -EIO;
 }
 
 /**
diff --git a/fs/gfs2/lock_dlm.c b/fs/gfs2/lock_dlm.c
index e028e55e67..49059274a5 100644
--- a/fs/gfs2/lock_dlm.c
+++ b/fs/gfs2/lock_dlm.c
@@ -316,10 +316,16 @@ static void gdlm_put_lock(struct gfs2_glock *gl)
 		gfs2_glock_free(gl);
 		return;
 	}
-	/* don't want to skip dlm_unlock writing the lvb when lock has one */
+
+	/*
+	 * When the lockspace is released, all remaining glocks will be
+	 * unlocked automatically.  This is more efficient than unlocking them
+	 * individually, but when the lock is held in DLM_LOCK_EX or
+	 * DLM_LOCK_PW mode, the lock value block (LVB) will be lost.
+	 */
 
 	if (test_bit(SDF_SKIP_DLM_UNLOCK, &sdp->sd_flags) &&
-	    !gl->gl_lksb.sb_lvbptr) {
+	    (!gl->gl_lksb.sb_lvbptr || gl->gl_state != LM_ST_EXCLUSIVE)) {
 		gfs2_glock_free_later(gl);
 		return;
 	}
diff --git a/fs/gfs2/log.c b/fs/gfs2/log.c
index 8cddf955eb..6ee6013fb8 100644
--- a/fs/gfs2/log.c
+++ b/fs/gfs2/log.c
@@ -786,7 +786,7 @@ void gfs2_glock_remove_revoke(struct gfs2_glock *gl)
 {
 	if (atomic_dec_return(&gl->gl_revokes) == 0) {
 		clear_bit(GLF_LFLUSH, &gl->gl_flags);
-		gfs2_glock_queue_put(gl);
+		gfs2_glock_put_async(gl);
 	}
 }
 
@@ -1108,7 +1108,8 @@ repeat:
 	lops_before_commit(sdp, tr);
 	if (gfs2_withdrawing_or_withdrawn(sdp))
 		goto out_withdraw;
-	gfs2_log_submit_bio(&sdp->sd_jdesc->jd_log_bio, REQ_OP_WRITE);
+	if (sdp->sd_jdesc)
+		gfs2_log_submit_bio(&sdp->sd_jdesc->jd_log_bio, REQ_OP_WRITE);
 	if (gfs2_withdrawing_or_withdrawn(sdp))
 		goto out_withdraw;
 
diff --git a/fs/gfs2/main.c b/fs/gfs2/main.c
index 79be0cdc73..04cadc02e5 100644
--- a/fs/gfs2/main.c
+++ b/fs/gfs2/main.c
@@ -111,7 +111,6 @@ static int __init init_gfs2_fs(void)
 	gfs2_inode_cachep = kmem_cache_create("gfs2_inode",
 					      sizeof(struct gfs2_inode),
 					      0,  SLAB_RECLAIM_ACCOUNT|
-						  SLAB_MEM_SPREAD|
 						  SLAB_ACCOUNT,
 					      gfs2_init_inode_once);
 	if (!gfs2_inode_cachep)
diff --git a/fs/gfs2/meta_io.c b/fs/gfs2/meta_io.c
index f814054c8c..2b26e8d529 100644
--- a/fs/gfs2/meta_io.c
+++ b/fs/gfs2/meta_io.c
@@ -32,14 +32,14 @@
 
 static int gfs2_aspace_writepage(struct page *page, struct writeback_control *wbc)
 {
+	struct folio *folio = page_folio(page);
 	struct buffer_head *bh, *head;
 	int nr_underway = 0;
 	blk_opf_t write_flags = REQ_META | REQ_PRIO | wbc_to_write_flags(wbc);
 
-	BUG_ON(!PageLocked(page));
-	BUG_ON(!page_has_buffers(page));
+	BUG_ON(!folio_test_locked(folio));
 
-	head = page_buffers(page);
+	head = folio_buffers(folio);
 	bh = head;
 
 	do {
@@ -55,7 +55,7 @@ static int gfs2_aspace_writepage(struct page *page, struct writeback_control *wb
 		if (wbc->sync_mode != WB_SYNC_NONE) {
 			lock_buffer(bh);
 		} else if (!trylock_buffer(bh)) {
-			redirty_page_for_writepage(wbc, page);
+			folio_redirty_for_writepage(wbc, folio);
 			continue;
 		}
 		if (test_clear_buffer_dirty(bh)) {
@@ -69,8 +69,8 @@ static int gfs2_aspace_writepage(struct page *page, struct writeback_control *wb
 	 * The page and its buffers are protected by PageWriteback(), so we can
 	 * drop the bh refcounts early.
 	 */
-	BUG_ON(PageWriteback(page));
-	set_page_writeback(page);
+	BUG_ON(folio_test_writeback(folio));
+	folio_start_writeback(folio);
 
 	do {
 		struct buffer_head *next = bh->b_this_page;
@@ -80,10 +80,10 @@ static int gfs2_aspace_writepage(struct page *page, struct writeback_control *wb
 		}
 		bh = next;
 	} while (bh != head);
-	unlock_page(page);
+	folio_unlock(folio);
 
 	if (nr_underway == 0)
-		end_page_writeback(page);
+		folio_end_writeback(folio);
 
 	return 0;
 }
diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c
index db0df091a6..05975ec76d 100644
--- a/fs/gfs2/ops_fstype.c
+++ b/fs/gfs2/ops_fstype.c
@@ -114,7 +114,7 @@ static struct gfs2_sbd *init_sbd(struct super_block *sb)
 
 	address_space_init_once(mapping);
 	mapping->a_ops = &gfs2_rgrp_aops;
-	mapping->host = sb->s_bdev->bd_inode;
+	mapping->host = sb->s_bdev->bd_mapping->host;
 	mapping->flags = 0;
 	mapping_set_gfp_mask(mapping, GFP_NOFS);
 	mapping->i_private_data = NULL;
@@ -185,22 +185,10 @@ static int gfs2_check_sb(struct gfs2_sbd *sdp, int silent)
 	return 0;
 }
 
-static void end_bio_io_page(struct bio *bio)
-{
-	struct page *page = bio->bi_private;
-
-	if (!bio->bi_status)
-		SetPageUptodate(page);
-	else
-		pr_warn("error %d reading superblock\n", bio->bi_status);
-	unlock_page(page);
-}
-
-static void gfs2_sb_in(struct gfs2_sbd *sdp, const void *buf)
+static void gfs2_sb_in(struct gfs2_sbd *sdp, const struct gfs2_sb *str)
 {
 	struct gfs2_sb_host *sb = &sdp->sd_sb;
 	struct super_block *s = sdp->sd_vfs;
-	const struct gfs2_sb *str = buf;
 
 	sb->sb_magic = be32_to_cpu(str->sb_header.mh_magic);
 	sb->sb_type = be32_to_cpu(str->sb_header.mh_type);
@@ -215,7 +203,7 @@ static void gfs2_sb_in(struct gfs2_sbd *sdp, const void *buf)
 
 	memcpy(sb->sb_lockproto, str->sb_lockproto, GFS2_LOCKNAME_LEN);
 	memcpy(sb->sb_locktable, str->sb_locktable, GFS2_LOCKNAME_LEN);
-	memcpy(&s->s_uuid, str->sb_uuid, 16);
+	super_set_uuid(s, str->sb_uuid, 16);
 }
 
 /**
@@ -240,34 +228,26 @@ static void gfs2_sb_in(struct gfs2_sbd *sdp, const void *buf)
 static int gfs2_read_super(struct gfs2_sbd *sdp, sector_t sector, int silent)
 {
 	struct super_block *sb = sdp->sd_vfs;
-	struct gfs2_sb *p;
 	struct page *page;
-	struct bio *bio;
+	struct bio_vec bvec;
+	struct bio bio;
+	int err;
 
-	page = alloc_page(GFP_NOFS);
+	page = alloc_page(GFP_KERNEL);
 	if (unlikely(!page))
 		return -ENOMEM;
 
-	ClearPageUptodate(page);
-	ClearPageDirty(page);
-	lock_page(page);
-
-	bio = bio_alloc(sb->s_bdev, 1, REQ_OP_READ | REQ_META, GFP_NOFS);
-	bio->bi_iter.bi_sector = sector * (sb->s_blocksize >> 9);
-	__bio_add_page(bio, page, PAGE_SIZE, 0);
+	bio_init(&bio, sb->s_bdev, &bvec, 1, REQ_OP_READ | REQ_META);
+	bio.bi_iter.bi_sector = sector * (sb->s_blocksize >> 9);
+	__bio_add_page(&bio, page, PAGE_SIZE, 0);
 
-	bio->bi_end_io = end_bio_io_page;
-	bio->bi_private = page;
-	submit_bio(bio);
-	wait_on_page_locked(page);
-	bio_put(bio);
-	if (!PageUptodate(page)) {
+	err = submit_bio_wait(&bio);
+	if (err) {
+		pr_warn("error %d reading superblock\n", err);
 		__free_page(page);
-		return -EIO;
+		return err;
 	}
-	p = kmap(page);
-	gfs2_sb_in(sdp, p);
-	kunmap(page);
+	gfs2_sb_in(sdp, page_address(page));
 	__free_page(page);
 	return gfs2_check_sb(sdp, silent);
 }
@@ -1289,7 +1269,7 @@ static int gfs2_fill_super(struct super_block *sb, struct fs_context *fc)
 		error = gfs2_make_fs_rw(sdp);
 
 	if (error) {
-		gfs2_freeze_unlock(&sdp->sd_freeze_gh);
+		gfs2_freeze_unlock(sdp);
 		gfs2_destroy_threads(sdp);
 		fs_err(sdp, "can't make FS RW: %d\n", error);
 		goto fail_per_node;
diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c
index 26d6c1eea5..29c7728167 100644
--- a/fs/gfs2/rgrp.c
+++ b/fs/gfs2/rgrp.c
@@ -814,11 +814,11 @@ static int compute_bitstructs(struct gfs2_rgrpd *rgd)
 	bi = rgd->rd_bits + (length - 1);
 	if ((bi->bi_start + bi->bi_bytes) * GFS2_NBBY != rgd->rd_data) {
 		gfs2_lm(sdp,
-			"ri_addr = %llu\n"
-			"ri_length = %u\n"
-			"ri_data0 = %llu\n"
-			"ri_data = %u\n"
-			"ri_bitbytes = %u\n"
+			"ri_addr=%llu "
+			"ri_length=%u "
+			"ri_data0=%llu "
+			"ri_data=%u "
+			"ri_bitbytes=%u "
 			"start=%u len=%u offset=%u\n",
 			(unsigned long long)rgd->rd_addr,
 			rgd->rd_length,
diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c
index 2d780b4701..7a5aedfcd5 100644
--- a/fs/gfs2/super.c
+++ b/fs/gfs2/super.c
@@ -67,9 +67,13 @@ void gfs2_jindex_free(struct gfs2_sbd *sdp)
 	sdp->sd_journals = 0;
 	spin_unlock(&sdp->sd_jindex_spin);
 
+	down_write(&sdp->sd_log_flush_lock);
 	sdp->sd_jdesc = NULL;
+	up_write(&sdp->sd_log_flush_lock);
+
 	while (!list_empty(&list)) {
 		jd = list_first_entry(&list, struct gfs2_jdesc, jd_list);
+		BUG_ON(jd->jd_log_bio);
 		gfs2_free_journal_extents(jd);
 		list_del(&jd->jd_list);
 		iput(jd->jd_inode);
@@ -354,7 +358,7 @@ static int gfs2_lock_fs_check_clean(struct gfs2_sbd *sdp)
 		list_add(&lfcc->list, &list);
 	}
 
-	gfs2_freeze_unlock(&sdp->sd_freeze_gh);
+	gfs2_freeze_unlock(sdp);
 
 	error = gfs2_glock_nq_init(sdp->sd_freeze_gl, LM_ST_EXCLUSIVE,
 				   LM_FLAG_NOEXP | GL_NOPID,
@@ -378,7 +382,7 @@ static int gfs2_lock_fs_check_clean(struct gfs2_sbd *sdp)
 	if (!error)
 		goto out;  /* success */
 
-	gfs2_freeze_unlock(&sdp->sd_freeze_gh);
+	gfs2_freeze_unlock(sdp);
 
 relock_shared:
 	error2 = gfs2_freeze_lock_shared(sdp);
@@ -617,7 +621,7 @@ restart:
 
 	/*  Release stuff  */
 
-	gfs2_freeze_unlock(&sdp->sd_freeze_gh);
+	gfs2_freeze_unlock(sdp);
 
 	iput(sdp->sd_jindex);
 	iput(sdp->sd_statfs_inode);
@@ -703,7 +707,7 @@ void gfs2_freeze_func(struct work_struct *work)
 	if (error)
 		goto freeze_failed;
 
-	gfs2_freeze_unlock(&sdp->sd_freeze_gh);
+	gfs2_freeze_unlock(sdp);
 	set_bit(SDF_FROZEN, &sdp->sd_flags);
 
 	error = gfs2_do_thaw(sdp);
@@ -808,7 +812,7 @@ static int gfs2_thaw_super(struct super_block *sb, enum freeze_holder who)
 	}
 
 	atomic_inc(&sb->s_active);
-	gfs2_freeze_unlock(&sdp->sd_freeze_gh);
+	gfs2_freeze_unlock(sdp);
 
 	error = gfs2_do_thaw(sdp);
 
@@ -829,7 +833,7 @@ void gfs2_thaw_freeze_initiator(struct super_block *sb)
 	if (!test_bit(SDF_FREEZE_INITIATOR, &sdp->sd_flags))
 		goto out;
 
-	gfs2_freeze_unlock(&sdp->sd_freeze_gh);
+	gfs2_freeze_unlock(sdp);
 
 out:
 	mutex_unlock(&sdp->sd_freeze_mutex);
@@ -1042,7 +1046,7 @@ static int gfs2_drop_inode(struct inode *inode)
 
 		gfs2_glock_hold(gl);
 		if (!gfs2_queue_try_to_evict(gl))
-			gfs2_glock_queue_put(gl);
+			gfs2_glock_put_async(gl);
 		return 0;
 	}
 
@@ -1248,7 +1252,7 @@ out_qs:
 static void gfs2_glock_put_eventually(struct gfs2_glock *gl)
 {
 	if (current->flags & PF_MEMALLOC)
-		gfs2_glock_queue_put(gl);
+		gfs2_glock_put_async(gl);
 	else
 		gfs2_glock_put(gl);
 }
@@ -1258,7 +1262,6 @@ static bool gfs2_upgrade_iopen_glock(struct inode *inode)
 	struct gfs2_inode *ip = GFS2_I(inode);
 	struct gfs2_sbd *sdp = GFS2_SB(inode);
 	struct gfs2_holder *gh = &ip->i_iopen_gh;
-	long timeout = 5 * HZ;
 	int error;
 
 	gh->gh_flags |= GL_NOCACHE;
@@ -1289,10 +1292,10 @@ static bool gfs2_upgrade_iopen_glock(struct inode *inode)
 	if (error)
 		return false;
 
-	timeout = wait_event_interruptible_timeout(sdp->sd_async_glock_wait,
+	wait_event_interruptible_timeout(sdp->sd_async_glock_wait,
 		!test_bit(HIF_WAIT, &gh->gh_iflags) ||
 		test_bit(GLF_DEMOTE, &ip->i_gl->gl_flags),
-		timeout);
+		5 * HZ);
 	if (!test_bit(HIF_HOLDER, &gh->gh_iflags)) {
 		gfs2_glock_dq(gh);
 		return false;
diff --git a/fs/gfs2/sys.c b/fs/gfs2/sys.c
index 250f340cb4..ecc699f8d9 100644
--- a/fs/gfs2/sys.c
+++ b/fs/gfs2/sys.c
@@ -88,7 +88,7 @@ static ssize_t status_show(struct gfs2_sbd *sdp, char *buf)
 		     "Withdraw In Prog:         %d\n"
 		     "Remote Withdraw:          %d\n"
 		     "Withdraw Recovery:        %d\n"
-		     "Deactivating:             %d\n"
+		     "Killing:                  %d\n"
 		     "sd_log_error:             %d\n"
 		     "sd_log_flush_lock:        %d\n"
 		     "sd_log_num_revoke:        %u\n"
@@ -336,7 +336,7 @@ static ssize_t demote_rq_store(struct gfs2_sbd *sdp, const char *buf, size_t len
 		return -EINVAL;
 	if (!test_and_set_bit(SDF_DEMOTE, &sdp->sd_flags))
 		fs_info(sdp, "demote interface used\n");
-	rv = gfs2_glock_get(sdp, glnum, glops, 0, &gl);
+	rv = gfs2_glock_get(sdp, glnum, glops, NO_CREATE, &gl);
 	if (rv)
 		return rv;
 	gfs2_glock_cb(gl, glmode);
diff --git a/fs/gfs2/util.c b/fs/gfs2/util.c
index fc3ecb180a..af4758d8d8 100644
--- a/fs/gfs2/util.c
+++ b/fs/gfs2/util.c
@@ -109,10 +109,10 @@ int gfs2_freeze_lock_shared(struct gfs2_sbd *sdp)
 	return error;
 }
 
-void gfs2_freeze_unlock(struct gfs2_holder *freeze_gh)
+void gfs2_freeze_unlock(struct gfs2_sbd *sdp)
 {
-	if (gfs2_holder_initialized(freeze_gh))
-		gfs2_glock_dq_uninit(freeze_gh);
+	if (gfs2_holder_initialized(&sdp->sd_freeze_gh))
+		gfs2_glock_dq_uninit(&sdp->sd_freeze_gh);
 }
 
 static void signal_our_withdraw(struct gfs2_sbd *sdp)
@@ -255,7 +255,7 @@ static void signal_our_withdraw(struct gfs2_sbd *sdp)
 		gfs2_glock_nq(&sdp->sd_live_gh);
 	}
 
-	gfs2_glock_queue_put(live_gl); /* drop extra reference we acquired */
+	gfs2_glock_put(live_gl); /* drop extra reference we acquired */
 	clear_bit(SDF_WITHDRAW_RECOVERY, &sdp->sd_flags);
 
 	/*
@@ -375,8 +375,8 @@ void gfs2_assert_withdraw_i(struct gfs2_sbd *sdp, char *assertion,
 		return;
 
 	fs_err(sdp,
-	       "fatal: assertion \"%s\" failed\n"
-	       "   function = %s, file = %s, line = %u\n",
+	       "fatal: assertion \"%s\" failed - "
+	       "function = %s, file = %s, line = %u\n",
 	       assertion, function, file, line);
 
 	/*
@@ -406,7 +406,8 @@ void gfs2_assert_warn_i(struct gfs2_sbd *sdp, char *assertion,
 		return;
 
 	if (sdp->sd_args.ar_errors == GFS2_ERRORS_WITHDRAW)
-		fs_warn(sdp, "warning: assertion \"%s\" failed at function = %s, file = %s, line = %u\n",
+		fs_warn(sdp, "warning: assertion \"%s\" failed - "
+			"function = %s, file = %s, line = %u\n",
 			assertion, function, file, line);
 
 	if (sdp->sd_args.ar_debug)
@@ -415,10 +416,10 @@ void gfs2_assert_warn_i(struct gfs2_sbd *sdp, char *assertion,
 		dump_stack();
 
 	if (sdp->sd_args.ar_errors == GFS2_ERRORS_PANIC)
-		panic("GFS2: fsid=%s: warning: assertion \"%s\" failed\n"
-		      "GFS2: fsid=%s:   function = %s, file = %s, line = %u\n",
+		panic("GFS2: fsid=%s: warning: assertion \"%s\" failed - "
+		      "function = %s, file = %s, line = %u\n",
 		      sdp->sd_fsname, assertion,
-		      sdp->sd_fsname, function, file, line);
+		      function, file, line);
 
 	sdp->sd_last_warning = jiffies;
 }
@@ -431,7 +432,8 @@ void gfs2_consist_i(struct gfs2_sbd *sdp, const char *function,
 		    char *file, unsigned int line)
 {
 	gfs2_lm(sdp,
-		"fatal: filesystem consistency error - function = %s, file = %s, line = %u\n",
+		"fatal: filesystem consistency error - "
+		"function = %s, file = %s, line = %u\n",
 		function, file, line);
 	gfs2_withdraw(sdp);
 }
@@ -446,9 +448,9 @@ void gfs2_consist_inode_i(struct gfs2_inode *ip,
 	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
 
 	gfs2_lm(sdp,
-		"fatal: filesystem consistency error\n"
-		"  inode = %llu %llu\n"
-		"  function = %s, file = %s, line = %u\n",
+		"fatal: filesystem consistency error - "
+		"inode = %llu %llu, "
+		"function = %s, file = %s, line = %u\n",
 		(unsigned long long)ip->i_no_formal_ino,
 		(unsigned long long)ip->i_no_addr,
 		function, file, line);
@@ -469,9 +471,9 @@ void gfs2_consist_rgrpd_i(struct gfs2_rgrpd *rgd,
 	sprintf(fs_id_buf, "fsid=%s: ", sdp->sd_fsname);
 	gfs2_rgrp_dump(NULL, rgd, fs_id_buf);
 	gfs2_lm(sdp,
-		"fatal: filesystem consistency error\n"
-		"  RG = %llu\n"
-		"  function = %s, file = %s, line = %u\n",
+		"fatal: filesystem consistency error - "
+		"RG = %llu, "
+		"function = %s, file = %s, line = %u\n",
 		(unsigned long long)rgd->rd_addr,
 		function, file, line);
 	gfs2_dump_glock(NULL, rgd->rd_gl, 1);
@@ -485,16 +487,16 @@ void gfs2_consist_rgrpd_i(struct gfs2_rgrpd *rgd,
  */
 
 int gfs2_meta_check_ii(struct gfs2_sbd *sdp, struct buffer_head *bh,
-		       const char *type, const char *function, char *file,
+		       const char *function, char *file,
 		       unsigned int line)
 {
 	int me;
 
 	gfs2_lm(sdp,
-		"fatal: invalid metadata block\n"
-		"  bh = %llu (%s)\n"
-		"  function = %s, file = %s, line = %u\n",
-		(unsigned long long)bh->b_blocknr, type,
+		"fatal: invalid metadata block - "
+		"bh = %llu (bad magic number), "
+		"function = %s, file = %s, line = %u\n",
+		(unsigned long long)bh->b_blocknr,
 		function, file, line);
 	me = gfs2_withdraw(sdp);
 	return (me) ? -1 : -2;
@@ -513,9 +515,9 @@ int gfs2_metatype_check_ii(struct gfs2_sbd *sdp, struct buffer_head *bh,
 	int me;
 
 	gfs2_lm(sdp,
-		"fatal: invalid metadata block\n"
-		"  bh = %llu (type: exp=%u, found=%u)\n"
-		"  function = %s, file = %s, line = %u\n",
+		"fatal: invalid metadata block - "
+		"bh = %llu (type: exp=%u, found=%u), "
+		"function = %s, file = %s, line = %u\n",
 		(unsigned long long)bh->b_blocknr, type, t,
 		function, file, line);
 	me = gfs2_withdraw(sdp);
@@ -532,8 +534,8 @@ int gfs2_io_error_i(struct gfs2_sbd *sdp, const char *function, char *file,
 		    unsigned int line)
 {
 	gfs2_lm(sdp,
-		"fatal: I/O error\n"
-		"  function = %s, file = %s, line = %u\n",
+		"fatal: I/O error - "
+		"function = %s, file = %s, line = %u\n",
 		function, file, line);
 	return gfs2_withdraw(sdp);
 }
@@ -550,9 +552,9 @@ void gfs2_io_error_bh_i(struct gfs2_sbd *sdp, struct buffer_head *bh,
 	if (gfs2_withdrawing_or_withdrawn(sdp))
 		return;
 
-	fs_err(sdp, "fatal: I/O error\n"
-	       "  block = %llu\n"
-	       "  function = %s, file = %s, line = %u\n",
+	fs_err(sdp, "fatal: I/O error - "
+	       "block = %llu, "
+	       "function = %s, file = %s, line = %u\n",
 	       (unsigned long long)bh->b_blocknr, function, file, line);
 	if (withdraw)
 		gfs2_withdraw(sdp);
diff --git a/fs/gfs2/util.h b/fs/gfs2/util.h
index ba07199846..27d03b6410 100644
--- a/fs/gfs2/util.h
+++ b/fs/gfs2/util.h
@@ -92,7 +92,7 @@ gfs2_consist_rgrpd_i((rgd), __func__, __FILE__, __LINE__)
 
 
 int gfs2_meta_check_ii(struct gfs2_sbd *sdp, struct buffer_head *bh,
-		       const char *type, const char *function,
+		       const char *function,
 		       char *file, unsigned int line);
 
 static inline int gfs2_meta_check(struct gfs2_sbd *sdp,
@@ -123,7 +123,7 @@ static inline int gfs2_metatype_check_i(struct gfs2_sbd *sdp,
 	u32 magic = be32_to_cpu(mh->mh_magic);
 	u16 t = be32_to_cpu(mh->mh_type);
 	if (unlikely(magic != GFS2_MAGIC))
-		return gfs2_meta_check_ii(sdp, bh, "magic number", function,
+		return gfs2_meta_check_ii(sdp, bh, function,
 					  file, line);
         if (unlikely(t != type))
 		return gfs2_metatype_check_ii(sdp, bh, type, t, function,
@@ -150,7 +150,7 @@ int gfs2_io_error_i(struct gfs2_sbd *sdp, const char *function,
 int check_journal_clean(struct gfs2_sbd *sdp, struct gfs2_jdesc *jd,
 		        bool verbose);
 int gfs2_freeze_lock_shared(struct gfs2_sbd *sdp);
-void gfs2_freeze_unlock(struct gfs2_holder *freeze_gh);
+void gfs2_freeze_unlock(struct gfs2_sbd *sdp);
 
 #define gfs2_io_error(sdp) \
 gfs2_io_error_i((sdp), __func__, __FILE__, __LINE__)
diff --git a/fs/gfs2/xattr.c b/fs/gfs2/xattr.c
index 8c96ba6230..17ae5070a9 100644
--- a/fs/gfs2/xattr.c
+++ b/fs/gfs2/xattr.c
@@ -96,30 +96,34 @@ static int ea_foreach_i(struct gfs2_inode *ip, struct buffer_head *bh,
 		return -EIO;
 
 	for (ea = GFS2_EA_BH2FIRST(bh);; prev = ea, ea = GFS2_EA2NEXT(ea)) {
-		if (!GFS2_EA_REC_LEN(ea))
-			goto fail;
+		if (!GFS2_EA_REC_LEN(ea)) {
+			gfs2_consist_inode(ip);
+			return -EIO;
+		}
 		if (!(bh->b_data <= (char *)ea && (char *)GFS2_EA2NEXT(ea) <=
-						  bh->b_data + bh->b_size))
-			goto fail;
-		if (!gfs2_eatype_valid(sdp, ea->ea_type))
-			goto fail;
+						  bh->b_data + bh->b_size)) {
+			gfs2_consist_inode(ip);
+			return -EIO;
+		}
+		if (!gfs2_eatype_valid(sdp, ea->ea_type)) {
+			gfs2_consist_inode(ip);
+			return -EIO;
+		}
 		error = ea_call(ip, bh, ea, prev, data);
 		if (error)
 			return error;
 
 		if (GFS2_EA_IS_LAST(ea)) {
 			if ((char *)GFS2_EA2NEXT(ea) !=
-			    bh->b_data + bh->b_size)
-				goto fail;
+			    bh->b_data + bh->b_size) {
+				gfs2_consist_inode(ip);
+				return -EIO;
+			}
 			break;
 		}
 	}
 
 	return error;
-
-fail:
-	gfs2_consist_inode(ip);
-	return -EIO;
 }
 
 static int ea_foreach(struct gfs2_inode *ip, ea_call_t ea_call, void *data)
diff --git a/fs/hfs/inode.c b/fs/hfs/inode.c
index 8c34798a07..744e10b469 100644
--- a/fs/hfs/inode.c
+++ b/fs/hfs/inode.c
@@ -200,6 +200,7 @@ struct inode *hfs_new_inode(struct inode *dir, const struct qstr *name, umode_t
 	HFS_I(inode)->flags = 0;
 	HFS_I(inode)->rsrc_inode = NULL;
 	HFS_I(inode)->fs_blocks = 0;
+	HFS_I(inode)->tz_secondswest = sys_tz.tz_minuteswest * 60;
 	if (S_ISDIR(mode)) {
 		inode->i_size = 2;
 		HFS_SB(sb)->folder_count++;
@@ -275,6 +276,8 @@ void hfs_inode_read_fork(struct inode *inode, struct hfs_extent *ext,
 	for (count = 0, i = 0; i < 3; i++)
 		count += be16_to_cpu(ext[i].count);
 	HFS_I(inode)->first_blocks = count;
+	HFS_I(inode)->cached_start = 0;
+	HFS_I(inode)->cached_blocks = 0;
 
 	inode->i_size = HFS_I(inode)->phys_size = log_size;
 	HFS_I(inode)->fs_blocks = (log_size + sb->s_blocksize - 1) >> sb->s_blocksize_bits;
diff --git a/fs/hfsplus/bfind.c b/fs/hfsplus/bfind.c
index ca2ba8c9f8..901e83d65d 100644
--- a/fs/hfsplus/bfind.c
+++ b/fs/hfsplus/bfind.c
@@ -25,19 +25,8 @@ int hfs_find_init(struct hfs_btree *tree, struct hfs_find_data *fd)
 	fd->key = ptr + tree->max_key_len + 2;
 	hfs_dbg(BNODE_REFS, "find_init: %d (%p)\n",
 		tree->cnid, __builtin_return_address(0));
-	switch (tree->cnid) {
-	case HFSPLUS_CAT_CNID:
-		mutex_lock_nested(&tree->tree_lock, CATALOG_BTREE_MUTEX);
-		break;
-	case HFSPLUS_EXT_CNID:
-		mutex_lock_nested(&tree->tree_lock, EXTENTS_BTREE_MUTEX);
-		break;
-	case HFSPLUS_ATTR_CNID:
-		mutex_lock_nested(&tree->tree_lock, ATTR_BTREE_MUTEX);
-		break;
-	default:
-		BUG();
-	}
+	mutex_lock_nested(&tree->tree_lock,
+			hfsplus_btree_lock_class(tree));
 	return 0;
 }
 
diff --git a/fs/hfsplus/extents.c b/fs/hfsplus/extents.c
index 3c572e44f2..9c51867ddd 100644
--- a/fs/hfsplus/extents.c
+++ b/fs/hfsplus/extents.c
@@ -430,7 +430,8 @@ int hfsplus_free_fork(struct super_block *sb, u32 cnid,
 		hfsplus_free_extents(sb, ext_entry, total_blocks - start,
 				     total_blocks);
 		total_blocks = start;
-		mutex_lock(&fd.tree->tree_lock);
+		mutex_lock_nested(&fd.tree->tree_lock,
+			hfsplus_btree_lock_class(fd.tree));
 	} while (total_blocks > blocks);
 	hfs_find_exit(&fd);
 
@@ -592,7 +593,8 @@ void hfsplus_file_truncate(struct inode *inode)
 					     alloc_cnt, alloc_cnt - blk_cnt);
 			hfsplus_dump_extent(hip->first_extents);
 			hip->first_blocks = blk_cnt;
-			mutex_lock(&fd.tree->tree_lock);
+			mutex_lock_nested(&fd.tree->tree_lock,
+				hfsplus_btree_lock_class(fd.tree));
 			break;
 		}
 		res = __hfsplus_ext_cache_extent(&fd, inode, alloc_cnt);
@@ -606,7 +608,8 @@ void hfsplus_file_truncate(struct inode *inode)
 		hfsplus_free_extents(sb, hip->cached_extents,
 				     alloc_cnt - start, alloc_cnt - blk_cnt);
 		hfsplus_dump_extent(hip->cached_extents);
-		mutex_lock(&fd.tree->tree_lock);
+		mutex_lock_nested(&fd.tree->tree_lock,
+				hfsplus_btree_lock_class(fd.tree));
 		if (blk_cnt > start) {
 			hip->extent_state |= HFSPLUS_EXT_DIRTY;
 			break;
diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h
index 012a3d003f..9e78f181c2 100644
--- a/fs/hfsplus/hfsplus_fs.h
+++ b/fs/hfsplus/hfsplus_fs.h
@@ -553,6 +553,27 @@ static inline __be32 __hfsp_ut2mt(time64_t ut)
 	return cpu_to_be32(lower_32_bits(ut) + HFSPLUS_UTC_OFFSET);
 }
 
+static inline enum hfsplus_btree_mutex_classes
+hfsplus_btree_lock_class(struct hfs_btree *tree)
+{
+	enum hfsplus_btree_mutex_classes class;
+
+	switch (tree->cnid) {
+	case HFSPLUS_CAT_CNID:
+		class = CATALOG_BTREE_MUTEX;
+		break;
+	case HFSPLUS_EXT_CNID:
+		class = EXTENTS_BTREE_MUTEX;
+		break;
+	case HFSPLUS_ATTR_CNID:
+		class = ATTR_BTREE_MUTEX;
+		break;
+	default:
+		BUG();
+	}
+	return class;
+}
+
 /* compatibility */
 #define hfsp_mt2ut(t)		(struct timespec64){ .tv_sec = __hfsp_mt2ut(t) }
 #define hfsp_ut2mt(t)		__hfsp_ut2mt((t).tv_sec)
diff --git a/fs/hfsplus/xattr.c b/fs/hfsplus/xattr.c
index 9c9ff6b8c6..9a1a93e388 100644
--- a/fs/hfsplus/xattr.c
+++ b/fs/hfsplus/xattr.c
@@ -400,21 +400,19 @@ static int name_len(const char *xattr_name, int xattr_name_len)
 	return len;
 }
 
-static int copy_name(char *buffer, const char *xattr_name, int name_len)
+static ssize_t copy_name(char *buffer, const char *xattr_name, int name_len)
 {
-	int len = name_len;
-	int offset = 0;
+	ssize_t len;
 
-	if (!is_known_namespace(xattr_name)) {
-		memcpy(buffer, XATTR_MAC_OSX_PREFIX, XATTR_MAC_OSX_PREFIX_LEN);
-		offset += XATTR_MAC_OSX_PREFIX_LEN;
-		len += XATTR_MAC_OSX_PREFIX_LEN;
-	}
-
-	strncpy(buffer + offset, xattr_name, name_len);
-	memset(buffer + offset + name_len, 0, 1);
-	len += 1;
+	if (!is_known_namespace(xattr_name))
+		len = scnprintf(buffer, name_len + XATTR_MAC_OSX_PREFIX_LEN,
+				 "%s%s", XATTR_MAC_OSX_PREFIX, xattr_name);
+	else
+		len = strscpy(buffer, xattr_name, name_len + 1);
 
+	/* include NUL-byte in length for non-empty name */
+	if (len >= 0)
+		len++;
 	return len;
 }
 
@@ -698,7 +696,7 @@ ssize_t hfsplus_listxattr(struct dentry *dentry, char *buffer, size_t size)
 		return err;
 	}
 
-	strbuf = kmalloc(NLS_MAX_CHARSET_SIZE * HFSPLUS_ATTR_MAX_STRLEN +
+	strbuf = kzalloc(NLS_MAX_CHARSET_SIZE * HFSPLUS_ATTR_MAX_STRLEN +
 			XATTR_MAC_OSX_PREFIX_LEN + 1, GFP_KERNEL);
 	if (!strbuf) {
 		res = -ENOMEM;
diff --git a/fs/hostfs/hostfs.h b/fs/hostfs/hostfs.h
index 0239e3af39..8b39c15c40 100644
--- a/fs/hostfs/hostfs.h
+++ b/fs/hostfs/hostfs.h
@@ -63,9 +63,10 @@ struct hostfs_stat {
 	struct hostfs_timespec atime, mtime, ctime;
 	unsigned int blksize;
 	unsigned long long blocks;
-	unsigned int maj;
-	unsigned int min;
-	dev_t dev;
+	struct {
+		unsigned int maj;
+		unsigned int min;
+	} rdev, dev;
 };
 
 extern int stat_file(const char *path, struct hostfs_stat *p, int fd);
diff --git a/fs/hostfs/hostfs_kern.c b/fs/hostfs/hostfs_kern.c
index a73d27c4dd..2c4d503a62 100644
--- a/fs/hostfs/hostfs_kern.c
+++ b/fs/hostfs/hostfs_kern.c
@@ -530,10 +530,11 @@ static int hostfs_inode_update(struct inode *ino, const struct hostfs_stat *st)
 static int hostfs_inode_set(struct inode *ino, void *data)
 {
 	struct hostfs_stat *st = data;
-	dev_t rdev;
+	dev_t dev, rdev;
 
 	/* Reencode maj and min with the kernel encoding.*/
-	rdev = MKDEV(st->maj, st->min);
+	rdev = MKDEV(st->rdev.maj, st->rdev.min);
+	dev = MKDEV(st->dev.maj, st->dev.min);
 
 	switch (st->mode & S_IFMT) {
 	case S_IFLNK:
@@ -559,7 +560,7 @@ static int hostfs_inode_set(struct inode *ino, void *data)
 		return -EIO;
 	}
 
-	HOSTFS_I(ino)->dev = st->dev;
+	HOSTFS_I(ino)->dev = dev;
 	ino->i_ino = st->ino;
 	ino->i_mode = st->mode;
 	return hostfs_inode_update(ino, st);
@@ -568,8 +569,9 @@ static int hostfs_inode_set(struct inode *ino, void *data)
 static int hostfs_inode_test(struct inode *inode, void *data)
 {
 	const struct hostfs_stat *st = data;
+	dev_t dev = MKDEV(st->dev.maj, st->dev.min);
 
-	return inode->i_ino == st->ino && HOSTFS_I(inode)->dev == st->dev;
+	return inode->i_ino == st->ino && HOSTFS_I(inode)->dev == dev;
 }
 
 static struct inode *hostfs_iget(struct super_block *sb, char *name)
diff --git a/fs/hostfs/hostfs_user.c b/fs/hostfs/hostfs_user.c
index 840619e39a..97e9c40a94 100644
--- a/fs/hostfs/hostfs_user.c
+++ b/fs/hostfs/hostfs_user.c
@@ -34,9 +34,10 @@ static void stat64_to_hostfs(const struct stat64 *buf, struct hostfs_stat *p)
 	p->mtime.tv_nsec = 0;
 	p->blksize = buf->st_blksize;
 	p->blocks = buf->st_blocks;
-	p->maj = os_major(buf->st_rdev);
-	p->min = os_minor(buf->st_rdev);
-	p->dev = buf->st_dev;
+	p->rdev.maj = os_major(buf->st_rdev);
+	p->rdev.min = os_minor(buf->st_rdev);
+	p->dev.maj = os_major(buf->st_dev);
+	p->dev.min = os_minor(buf->st_dev);
 }
 
 int stat_file(const char *path, struct hostfs_stat *p, int fd)
diff --git a/fs/hpfs/super.c b/fs/hpfs/super.c
index 6b0ba3c1ef..314834a078 100644
--- a/fs/hpfs/super.c
+++ b/fs/hpfs/super.c
@@ -255,7 +255,7 @@ static int init_inodecache(void)
 	hpfs_inode_cachep = kmem_cache_create("hpfs_inode_cache",
 					     sizeof(struct hpfs_inode_info),
 					     0, (SLAB_RECLAIM_ACCOUNT|
-						SLAB_MEM_SPREAD|SLAB_ACCOUNT),
+						SLAB_ACCOUNT),
 					     init_once);
 	if (hpfs_inode_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c
index d746866ae3..412f295ace 100644
--- a/fs/hugetlbfs/inode.c
+++ b/fs/hugetlbfs/inode.c
@@ -40,7 +40,7 @@
 #include <linux/sched/mm.h>
 
 static const struct address_space_operations hugetlbfs_aops;
-const struct file_operations hugetlbfs_file_operations;
+static const struct file_operations hugetlbfs_file_operations;
 static const struct inode_operations hugetlbfs_dir_inode_operations;
 static const struct inode_operations hugetlbfs_inode_operations;
 
@@ -176,14 +176,12 @@ hugetlb_get_unmapped_area_bottomup(struct file *file, unsigned long addr,
 		unsigned long len, unsigned long pgoff, unsigned long flags)
 {
 	struct hstate *h = hstate_file(file);
-	struct vm_unmapped_area_info info;
+	struct vm_unmapped_area_info info = {};
 
-	info.flags = 0;
 	info.length = len;
 	info.low_limit = current->mm->mmap_base;
 	info.high_limit = arch_get_mmap_end(addr, len, flags);
 	info.align_mask = PAGE_MASK & ~huge_page_mask(h);
-	info.align_offset = 0;
 	return vm_unmapped_area(&info);
 }
 
@@ -192,14 +190,13 @@ hugetlb_get_unmapped_area_topdown(struct file *file, unsigned long addr,
 		unsigned long len, unsigned long pgoff, unsigned long flags)
 {
 	struct hstate *h = hstate_file(file);
-	struct vm_unmapped_area_info info;
+	struct vm_unmapped_area_info info = {};
 
 	info.flags = VM_UNMAPPED_AREA_TOPDOWN;
 	info.length = len;
 	info.low_limit = PAGE_SIZE;
 	info.high_limit = arch_get_mmap_base(addr, current->mm->mmap_base);
 	info.align_mask = PAGE_MASK & ~huge_page_mask(h);
-	info.align_offset = 0;
 	addr = vm_unmapped_area(&info);
 
 	/*
@@ -249,11 +246,11 @@ generic_hugetlb_get_unmapped_area(struct file *file, unsigned long addr,
 	}
 
 	/*
-	 * Use mm->get_unmapped_area value as a hint to use topdown routine.
+	 * Use MMF_TOPDOWN flag as a hint to use topdown routine.
 	 * If architectures have special needs, they should define their own
 	 * version of hugetlb_get_unmapped_area.
 	 */
-	if (mm->get_unmapped_area == arch_get_unmapped_area_topdown)
+	if (test_bit(MMF_TOPDOWN, &mm->flags))
 		return hugetlb_get_unmapped_area_topdown(file, addr, len,
 				pgoff, flags);
 	return hugetlb_get_unmapped_area_bottomup(file, addr, len,
@@ -933,7 +930,7 @@ static int hugetlbfs_setattr(struct mnt_idmap *idmap,
 	unsigned int ia_valid = attr->ia_valid;
 	struct hugetlbfs_inode_info *info = HUGETLBFS_I(inode);
 
-	error = setattr_prepare(&nop_mnt_idmap, dentry, attr);
+	error = setattr_prepare(idmap, dentry, attr);
 	if (error)
 		return error;
 
@@ -950,7 +947,7 @@ static int hugetlbfs_setattr(struct mnt_idmap *idmap,
 		hugetlb_vmtruncate(inode, newsize);
 	}
 
-	setattr_copy(&nop_mnt_idmap, inode, attr);
+	setattr_copy(idmap, inode, attr);
 	mark_inode_dirty(inode);
 	return 0;
 }
@@ -985,6 +982,7 @@ static struct inode *hugetlbfs_get_root(struct super_block *sb,
 static struct lock_class_key hugetlbfs_i_mmap_rwsem_key;
 
 static struct inode *hugetlbfs_get_inode(struct super_block *sb,
+					struct mnt_idmap *idmap,
 					struct inode *dir,
 					umode_t mode, dev_t dev)
 {
@@ -1006,7 +1004,7 @@ static struct inode *hugetlbfs_get_inode(struct super_block *sb,
 		struct hugetlbfs_inode_info *info = HUGETLBFS_I(inode);
 
 		inode->i_ino = get_next_ino();
-		inode_init_owner(&nop_mnt_idmap, inode, dir, mode);
+		inode_init_owner(idmap, inode, dir, mode);
 		lockdep_set_class(&inode->i_mapping->i_mmap_rwsem,
 				&hugetlbfs_i_mmap_rwsem_key);
 		inode->i_mapping->a_ops = &hugetlbfs_aops;
@@ -1050,7 +1048,7 @@ static int hugetlbfs_mknod(struct mnt_idmap *idmap, struct inode *dir,
 {
 	struct inode *inode;
 
-	inode = hugetlbfs_get_inode(dir->i_sb, dir, mode, dev);
+	inode = hugetlbfs_get_inode(dir->i_sb, idmap, dir, mode, dev);
 	if (!inode)
 		return -ENOSPC;
 	inode_set_mtime_to_ts(dir, inode_set_ctime_current(dir));
@@ -1062,7 +1060,7 @@ static int hugetlbfs_mknod(struct mnt_idmap *idmap, struct inode *dir,
 static int hugetlbfs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
 			   struct dentry *dentry, umode_t mode)
 {
-	int retval = hugetlbfs_mknod(&nop_mnt_idmap, dir, dentry,
+	int retval = hugetlbfs_mknod(idmap, dir, dentry,
 				     mode | S_IFDIR, 0);
 	if (!retval)
 		inc_nlink(dir);
@@ -1073,7 +1071,7 @@ static int hugetlbfs_create(struct mnt_idmap *idmap,
 			    struct inode *dir, struct dentry *dentry,
 			    umode_t mode, bool excl)
 {
-	return hugetlbfs_mknod(&nop_mnt_idmap, dir, dentry, mode | S_IFREG, 0);
+	return hugetlbfs_mknod(idmap, dir, dentry, mode | S_IFREG, 0);
 }
 
 static int hugetlbfs_tmpfile(struct mnt_idmap *idmap,
@@ -1082,7 +1080,7 @@ static int hugetlbfs_tmpfile(struct mnt_idmap *idmap,
 {
 	struct inode *inode;
 
-	inode = hugetlbfs_get_inode(dir->i_sb, dir, mode | S_IFREG, 0);
+	inode = hugetlbfs_get_inode(dir->i_sb, idmap, dir, mode | S_IFREG, 0);
 	if (!inode)
 		return -ENOSPC;
 	inode_set_mtime_to_ts(dir, inode_set_ctime_current(dir));
@@ -1094,10 +1092,11 @@ static int hugetlbfs_symlink(struct mnt_idmap *idmap,
 			     struct inode *dir, struct dentry *dentry,
 			     const char *symname)
 {
+	const umode_t mode = S_IFLNK|S_IRWXUGO;
 	struct inode *inode;
 	int error = -ENOSPC;
 
-	inode = hugetlbfs_get_inode(dir->i_sb, dir, S_IFLNK|S_IRWXUGO, 0);
+	inode = hugetlbfs_get_inode(dir->i_sb, idmap, dir, mode, 0);
 	if (inode) {
 		int l = strlen(symname)+1;
 		error = page_symlink(inode, symname, l);
@@ -1299,13 +1298,14 @@ static void init_once(void *foo)
 	inode_init_once(&ei->vfs_inode);
 }
 
-const struct file_operations hugetlbfs_file_operations = {
+static const struct file_operations hugetlbfs_file_operations = {
 	.read_iter		= hugetlbfs_read_iter,
 	.mmap			= hugetlbfs_file_mmap,
 	.fsync			= noop_fsync,
 	.get_unmapped_area	= hugetlb_get_unmapped_area,
 	.llseek			= default_llseek,
 	.fallocate		= hugetlbfs_fallocate,
+	.fop_flags		= FOP_HUGE_PAGES,
 };
 
 static const struct inode_operations hugetlbfs_dir_inode_operations = {
@@ -1566,6 +1566,7 @@ static struct file_system_type hugetlbfs_fs_type = {
 	.init_fs_context	= hugetlbfs_init_fs_context,
 	.parameters		= hugetlb_fs_parameters,
 	.kill_sb		= kill_litter_super,
+	.fs_flags               = FS_ALLOW_IDMAP,
 };
 
 static struct vfsmount *hugetlbfs_vfsmount[HUGE_MAX_HSTATE];
@@ -1619,7 +1620,9 @@ struct file *hugetlb_file_setup(const char *name, size_t size,
 	}
 
 	file = ERR_PTR(-ENOSPC);
-	inode = hugetlbfs_get_inode(mnt->mnt_sb, NULL, S_IFREG | S_IRWXUGO, 0);
+	/* hugetlbfs_vfsmount[] mounts do not use idmapped mounts.  */
+	inode = hugetlbfs_get_inode(mnt->mnt_sb, &nop_mnt_idmap, NULL,
+				    S_IFREG | S_IRWXUGO, 0);
 	if (!inode)
 		goto out;
 	if (creat_flags == HUGETLB_SHMFS_INODE)
diff --git a/fs/inode.c b/fs/inode.c
index 91048c4c9c..3a41f83a4b 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -20,6 +20,7 @@
 #include <linux/ratelimit.h>
 #include <linux/list_lru.h>
 #include <linux/iversion.h>
+#include <linux/rw_hint.h>
 #include <trace/events/writeback.h>
 #include "internal.h"
 
@@ -588,7 +589,8 @@ void dump_mapping(const struct address_space *mapping)
 	}
 
 	dentry_ptr = container_of(dentry_first, struct dentry, d_u.d_alias);
-	if (get_kernel_nofault(dentry, dentry_ptr)) {
+	if (get_kernel_nofault(dentry, dentry_ptr) ||
+	    !dentry.d_parent || !dentry.d_name.name) {
 		pr_warn("aops:%ps ino:%lx invalid dentry:%px\n",
 				a_ops, ino, dentry_ptr);
 		return;
@@ -2031,7 +2033,7 @@ static int __remove_privs(struct mnt_idmap *idmap,
 	return notify_change(idmap, dentry, &newattrs, NULL);
 }
 
-static int __file_remove_privs(struct file *file, unsigned int flags)
+int file_remove_privs_flags(struct file *file, unsigned int flags)
 {
 	struct dentry *dentry = file_dentry(file);
 	struct inode *inode = file_inode(file);
@@ -2056,6 +2058,7 @@ static int __file_remove_privs(struct file *file, unsigned int flags)
 		inode_has_no_xattr(inode);
 	return error;
 }
+EXPORT_SYMBOL_GPL(file_remove_privs_flags);
 
 /**
  * file_remove_privs - remove special file privileges (suid, capabilities)
@@ -2068,7 +2071,7 @@ static int __file_remove_privs(struct file *file, unsigned int flags)
  */
 int file_remove_privs(struct file *file)
 {
-	return __file_remove_privs(file, 0);
+	return file_remove_privs_flags(file, 0);
 }
 EXPORT_SYMBOL(file_remove_privs);
 
@@ -2161,7 +2164,7 @@ static int file_modified_flags(struct file *file, int flags)
 	 * Clear the security bits if the process is not being run by root.
 	 * This keeps people from modifying setuid and setgid binaries.
 	 */
-	ret = __file_remove_privs(file, flags);
+	ret = file_remove_privs_flags(file, flags);
 	if (ret)
 		return ret;
 
@@ -2285,7 +2288,7 @@ void __init inode_init(void)
 					 sizeof(struct inode),
 					 0,
 					 (SLAB_RECLAIM_ACCOUNT|SLAB_PANIC|
-					 SLAB_MEM_SPREAD|SLAB_ACCOUNT),
+					 SLAB_ACCOUNT),
 					 init_once);
 
 	/* Hash may have been set up in inode_init_early */
@@ -2509,7 +2512,7 @@ struct timespec64 inode_set_ctime_current(struct inode *inode)
 {
 	struct timespec64 now = current_time(inode);
 
-	inode_set_ctime(inode, now.tv_sec, now.tv_nsec);
+	inode_set_ctime_to_ts(inode, now);
 	return now;
 }
 EXPORT_SYMBOL(inode_set_ctime_current);
diff --git a/fs/internal.h b/fs/internal.h
index b67406435f..ab2225136f 100644
--- a/fs/internal.h
+++ b/fs/internal.h
@@ -62,6 +62,9 @@ int do_mkdirat(int dfd, struct filename *name, umode_t mode);
 int do_symlinkat(struct filename *from, int newdfd, struct filename *to);
 int do_linkat(int olddfd, struct filename *old, int newdfd,
 			struct filename *new, int flags);
+int vfs_tmpfile(struct mnt_idmap *idmap,
+		const struct path *parentpath,
+		struct file *file, umode_t mode);
 
 /*
  * namespace.c
@@ -183,6 +186,7 @@ extern struct open_how build_open_how(int flags, umode_t mode);
 extern int build_open_flags(const struct open_how *how, struct open_flags *op);
 struct file *file_close_fd_locked(struct files_struct *files, unsigned fd);
 
+long do_ftruncate(struct file *file, loff_t length, int small);
 long do_sys_ftruncate(unsigned int fd, loff_t length, int small);
 int chmod_common(const struct path *path, umode_t mode);
 int do_fchownat(int dfd, const char __user *filename, uid_t user, gid_t group,
@@ -310,3 +314,10 @@ ssize_t __kernel_write_iter(struct file *file, struct iov_iter *from, loff_t *po
 struct mnt_idmap *alloc_mnt_idmap(struct user_namespace *mnt_userns);
 struct mnt_idmap *mnt_idmap_get(struct mnt_idmap *idmap);
 void mnt_idmap_put(struct mnt_idmap *idmap);
+struct stashed_operations {
+	void (*put_data)(void *data);
+	int (*init_inode)(struct inode *inode, void *data);
+};
+int path_from_stashed(struct dentry **stashed, struct vfsmount *mnt, void *data,
+		      struct path *path);
+void stashed_dentry_prune(struct dentry *dentry);
diff --git a/fs/ioctl.c b/fs/ioctl.c
index 76cf22ac97..6477689112 100644
--- a/fs/ioctl.c
+++ b/fs/ioctl.c
@@ -763,12 +763,42 @@ static int ioctl_fssetxattr(struct file *file, void __user *argp)
 	return err;
 }
 
+static int ioctl_getfsuuid(struct file *file, void __user *argp)
+{
+	struct super_block *sb = file_inode(file)->i_sb;
+	struct fsuuid2 u = { .len = sb->s_uuid_len, };
+
+	if (!sb->s_uuid_len)
+		return -ENOTTY;
+
+	memcpy(&u.uuid[0], &sb->s_uuid, sb->s_uuid_len);
+
+	return copy_to_user(argp, &u, sizeof(u)) ? -EFAULT : 0;
+}
+
+static int ioctl_get_fs_sysfs_path(struct file *file, void __user *argp)
+{
+	struct super_block *sb = file_inode(file)->i_sb;
+
+	if (!strlen(sb->s_sysfs_name))
+		return -ENOTTY;
+
+	struct fs_sysfs_path u = {};
+
+	u.len = scnprintf(u.name, sizeof(u.name), "%s/%s", sb->s_type->name, sb->s_sysfs_name);
+
+	return copy_to_user(argp, &u, sizeof(u)) ? -EFAULT : 0;
+}
+
 /*
  * do_vfs_ioctl() is not for drivers and not intended to be EXPORT_SYMBOL()'d.
  * It's just a simple helper for sys_ioctl and compat_sys_ioctl.
  *
  * When you add any new common ioctls to the switches above and below,
  * please ensure they have compatible arguments in compat mode.
+ *
+ * The LSM mailing list should also be notified of any command additions or
+ * changes, as specific LSMs may be affected.
  */
 static int do_vfs_ioctl(struct file *filp, unsigned int fd,
 			unsigned int cmd, unsigned long arg)
@@ -845,6 +875,12 @@ static int do_vfs_ioctl(struct file *filp, unsigned int fd,
 	case FS_IOC_FSSETXATTR:
 		return ioctl_fssetxattr(filp, argp);
 
+	case FS_IOC_GETFSUUID:
+		return ioctl_getfsuuid(filp, argp);
+
+	case FS_IOC_GETFSSYSFSPATH:
+		return ioctl_get_fs_sysfs_path(filp, argp);
+
 	default:
 		if (S_ISREG(inode->i_mode))
 			return file_ioctl(filp, cmd, argp);
diff --git a/fs/iomap/Makefile b/fs/iomap/Makefile
index fc070184b7..381d76c5c2 100644
--- a/fs/iomap/Makefile
+++ b/fs/iomap/Makefile
@@ -4,7 +4,7 @@
 # All Rights Reserved.
 #
 
-ccflags-y += -I $(srctree)/$(src)		# needed for trace events
+ccflags-y += -I $(src)		# needed for trace events
 
 obj-$(CONFIG_FS_IOMAP)		+= iomap.o
 
diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
index 228fd2e05e..d465589902 100644
--- a/fs/iomap/buffered-io.c
+++ b/fs/iomap/buffered-io.c
@@ -1,7 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0
 /*
  * Copyright (C) 2010 Red Hat, Inc.
- * Copyright (C) 2016-2019 Christoph Hellwig.
+ * Copyright (C) 2016-2023 Christoph Hellwig.
  */
 #include <linux/module.h>
 #include <linux/compiler.h>
@@ -95,6 +95,44 @@ static inline bool ifs_block_is_dirty(struct folio *folio,
 	return test_bit(block + blks_per_folio, ifs->state);
 }
 
+static unsigned ifs_find_dirty_range(struct folio *folio,
+		struct iomap_folio_state *ifs, u64 *range_start, u64 range_end)
+{
+	struct inode *inode = folio->mapping->host;
+	unsigned start_blk =
+		offset_in_folio(folio, *range_start) >> inode->i_blkbits;
+	unsigned end_blk = min_not_zero(
+		offset_in_folio(folio, range_end) >> inode->i_blkbits,
+		i_blocks_per_folio(inode, folio));
+	unsigned nblks = 1;
+
+	while (!ifs_block_is_dirty(folio, ifs, start_blk))
+		if (++start_blk == end_blk)
+			return 0;
+
+	while (start_blk + nblks < end_blk) {
+		if (!ifs_block_is_dirty(folio, ifs, start_blk + nblks))
+			break;
+		nblks++;
+	}
+
+	*range_start = folio_pos(folio) + (start_blk << inode->i_blkbits);
+	return nblks << inode->i_blkbits;
+}
+
+static unsigned iomap_find_dirty_range(struct folio *folio, u64 *range_start,
+		u64 range_end)
+{
+	struct iomap_folio_state *ifs = folio->private;
+
+	if (*range_start >= range_end)
+		return 0;
+
+	if (ifs)
+		return ifs_find_dirty_range(folio, ifs, range_start, range_end);
+	return range_end - *range_start;
+}
+
 static void ifs_clear_range_dirty(struct folio *folio,
 		struct iomap_folio_state *ifs, size_t off, size_t len)
 {
@@ -203,6 +241,7 @@ static void iomap_adjust_read_range(struct inode *inode, struct folio *folio,
 	unsigned block_size = (1 << block_bits);
 	size_t poff = offset_in_folio(folio, *pos);
 	size_t plen = min_t(loff_t, folio_size(folio) - poff, length);
+	size_t orig_plen = plen;
 	unsigned first = poff >> block_bits;
 	unsigned last = (poff + plen - 1) >> block_bits;
 
@@ -239,7 +278,7 @@ static void iomap_adjust_read_range(struct inode *inode, struct folio *folio,
 	 * handle both halves separately so that we properly zero data in the
 	 * page cache for blocks that are entirely outside of i_size.
 	 */
-	if (orig_pos <= isize && orig_pos + length > isize) {
+	if (orig_pos <= isize && orig_pos + orig_plen > isize) {
 		unsigned end = offset_in_folio(folio, isize - 1) >> block_bits;
 
 		if (first <= end && last > end)
@@ -786,12 +825,11 @@ static int iomap_write_begin(struct iomap_iter *iter, loff_t pos,
 
 out_unlock:
 	__iomap_put_folio(iter, pos, 0, folio);
-	iomap_write_failed(iter->inode, pos, len);
 
 	return status;
 }
 
-static size_t __iomap_write_end(struct inode *inode, loff_t pos, size_t len,
+static bool __iomap_write_end(struct inode *inode, loff_t pos, size_t len,
 		size_t copied, struct folio *folio)
 {
 	flush_dcache_folio(folio);
@@ -808,14 +846,14 @@ static size_t __iomap_write_end(struct inode *inode, loff_t pos, size_t len,
 	 * redo the whole thing.
 	 */
 	if (unlikely(copied < len && !folio_test_uptodate(folio)))
-		return 0;
+		return false;
 	iomap_set_range_uptodate(folio, offset_in_folio(folio, pos), len);
 	iomap_set_range_dirty(folio, offset_in_folio(folio, pos), copied);
 	filemap_dirty_folio(inode->i_mapping, folio);
-	return copied;
+	return true;
 }
 
-static size_t iomap_write_end_inline(const struct iomap_iter *iter,
+static void iomap_write_end_inline(const struct iomap_iter *iter,
 		struct folio *folio, loff_t pos, size_t copied)
 {
 	const struct iomap *iomap = &iter->iomap;
@@ -830,52 +868,57 @@ static size_t iomap_write_end_inline(const struct iomap_iter *iter,
 	kunmap_local(addr);
 
 	mark_inode_dirty(iter->inode);
-	return copied;
 }
 
-/* Returns the number of bytes copied.  May be 0.  Cannot be an errno. */
-static size_t iomap_write_end(struct iomap_iter *iter, loff_t pos, size_t len,
+/*
+ * Returns true if all copied bytes have been written to the pagecache,
+ * otherwise return false.
+ */
+static bool iomap_write_end(struct iomap_iter *iter, loff_t pos, size_t len,
 		size_t copied, struct folio *folio)
 {
 	const struct iomap *srcmap = iomap_iter_srcmap(iter);
 	loff_t old_size = iter->inode->i_size;
-	size_t ret;
+	size_t written;
 
 	if (srcmap->type == IOMAP_INLINE) {
-		ret = iomap_write_end_inline(iter, folio, pos, copied);
+		iomap_write_end_inline(iter, folio, pos, copied);
+		written = copied;
 	} else if (srcmap->flags & IOMAP_F_BUFFER_HEAD) {
-		ret = block_write_end(NULL, iter->inode->i_mapping, pos, len,
-				copied, &folio->page, NULL);
+		written = block_write_end(NULL, iter->inode->i_mapping, pos,
+					len, copied, &folio->page, NULL);
+		WARN_ON_ONCE(written != copied && written != 0);
 	} else {
-		ret = __iomap_write_end(iter->inode, pos, len, copied, folio);
+		written = __iomap_write_end(iter->inode, pos, len, copied,
+					    folio) ? copied : 0;
 	}
 
 	/*
 	 * Update the in-memory inode size after copying the data into the page
 	 * cache.  It's up to the file system to write the updated size to disk,
 	 * preferably after I/O completion so that no stale data is exposed.
+	 * Only once that's done can we unlock and release the folio.
 	 */
-	if (pos + ret > old_size) {
-		i_size_write(iter->inode, pos + ret);
+	if (pos + written > old_size) {
+		i_size_write(iter->inode, pos + written);
 		iter->iomap.flags |= IOMAP_F_SIZE_CHANGED;
 	}
-	__iomap_put_folio(iter, pos, ret, folio);
+	__iomap_put_folio(iter, pos, written, folio);
 
 	if (old_size < pos)
 		pagecache_isize_extended(iter->inode, old_size, pos);
-	if (ret < len)
-		iomap_write_failed(iter->inode, pos + ret, len - ret);
-	return ret;
+
+	return written == copied;
 }
 
 static loff_t iomap_write_iter(struct iomap_iter *iter, struct iov_iter *i)
 {
 	loff_t length = iomap_length(iter);
-	size_t chunk = PAGE_SIZE << MAX_PAGECACHE_ORDER;
 	loff_t pos = iter->pos;
-	ssize_t written = 0;
+	ssize_t total_written = 0;
 	long status = 0;
 	struct address_space *mapping = iter->inode->i_mapping;
+	size_t chunk = mapping_max_folio_size(mapping);
 	unsigned int bdp_flags = (iter->flags & IOMAP_NOWAIT) ? BDP_ASYNC : 0;
 
 	do {
@@ -883,6 +926,7 @@ static loff_t iomap_write_iter(struct iomap_iter *iter, struct iov_iter *i)
 		size_t offset;		/* Offset into folio */
 		size_t bytes;		/* Bytes to write to folio */
 		size_t copied;		/* Bytes copied from user */
+		size_t written;		/* Bytes have been written */
 
 		bytes = iov_iter_count(i);
 retry:
@@ -912,8 +956,10 @@ retry:
 		}
 
 		status = iomap_write_begin(iter, pos, bytes, &folio);
-		if (unlikely(status))
+		if (unlikely(status)) {
+			iomap_write_failed(iter->inode, pos, bytes);
 			break;
+		}
 		if (iter->iomap.flags & IOMAP_F_STALE)
 			break;
 
@@ -925,19 +971,20 @@ retry:
 			flush_dcache_folio(folio);
 
 		copied = copy_folio_from_iter_atomic(folio, offset, bytes, i);
-		status = iomap_write_end(iter, pos, bytes, copied, folio);
-
-		if (unlikely(copied != status))
-			iov_iter_revert(i, copied - status);
+		written = iomap_write_end(iter, pos, bytes, copied, folio) ?
+			  copied : 0;
 
 		cond_resched();
-		if (unlikely(status == 0)) {
+		if (unlikely(written == 0)) {
 			/*
 			 * A short copy made iomap_write_end() reject the
 			 * thing entirely.  Might be memory poisoning
 			 * halfway through, might be a race with munmap,
 			 * might be severe memory pressure.
 			 */
+			iomap_write_failed(iter->inode, pos, bytes);
+			iov_iter_revert(i, copied);
+
 			if (chunk > PAGE_SIZE)
 				chunk /= 2;
 			if (copied) {
@@ -945,17 +992,17 @@ retry:
 				goto retry;
 			}
 		} else {
-			pos += status;
-			written += status;
-			length -= status;
+			pos += written;
+			total_written += written;
+			length -= written;
 		}
 	} while (iov_iter_count(i) && length);
 
 	if (status == -EAGAIN) {
-		iov_iter_revert(i, written);
+		iov_iter_revert(i, total_written);
 		return -EAGAIN;
 	}
-	return written ? written : status;
+	return total_written ? total_written : status;
 }
 
 ssize_t
@@ -1284,6 +1331,7 @@ static loff_t iomap_unshare_iter(struct iomap_iter *iter)
 		int status;
 		size_t offset;
 		size_t bytes = min_t(u64, SIZE_MAX, length);
+		bool ret;
 
 		status = iomap_write_begin(iter, pos, bytes, &folio);
 		if (unlikely(status))
@@ -1295,8 +1343,8 @@ static loff_t iomap_unshare_iter(struct iomap_iter *iter)
 		if (bytes > folio_size(folio) - offset)
 			bytes = folio_size(folio) - offset;
 
-		bytes = iomap_write_end(iter, pos, bytes, bytes, folio);
-		if (WARN_ON_ONCE(bytes == 0))
+		ret = iomap_write_end(iter, pos, bytes, bytes, folio);
+		if (WARN_ON_ONCE(!ret))
 			return -EIO;
 
 		cond_resched();
@@ -1345,6 +1393,7 @@ static loff_t iomap_zero_iter(struct iomap_iter *iter, bool *did_zero)
 		int status;
 		size_t offset;
 		size_t bytes = min_t(u64, SIZE_MAX, length);
+		bool ret;
 
 		status = iomap_write_begin(iter, pos, bytes, &folio);
 		if (status)
@@ -1359,8 +1408,8 @@ static loff_t iomap_zero_iter(struct iomap_iter *iter, bool *did_zero)
 		folio_zero_range(folio, offset, bytes);
 		folio_mark_accessed(folio);
 
-		bytes = iomap_write_end(iter, pos, bytes, bytes, folio);
-		if (WARN_ON_ONCE(bytes == 0))
+		ret = iomap_write_end(iter, pos, bytes, bytes, folio);
+		if (WARN_ON_ONCE(!ret))
 			return -EIO;
 
 		pos += bytes;
@@ -1454,15 +1503,10 @@ out_unlock:
 EXPORT_SYMBOL_GPL(iomap_page_mkwrite);
 
 static void iomap_finish_folio_write(struct inode *inode, struct folio *folio,
-		size_t len, int error)
+		size_t len)
 {
 	struct iomap_folio_state *ifs = folio->private;
 
-	if (error) {
-		folio_set_error(folio);
-		mapping_set_error(inode->i_mapping, error);
-	}
-
 	WARN_ON_ONCE(i_blocks_per_folio(inode, folio) > 1 && !ifs);
 	WARN_ON_ONCE(ifs && atomic_read(&ifs->write_bytes_pending) <= 0);
 
@@ -1479,40 +1523,29 @@ static u32
 iomap_finish_ioend(struct iomap_ioend *ioend, int error)
 {
 	struct inode *inode = ioend->io_inode;
-	struct bio *bio = &ioend->io_inline_bio;
-	struct bio *last = ioend->io_bio, *next;
-	u64 start = bio->bi_iter.bi_sector;
-	loff_t offset = ioend->io_offset;
-	bool quiet = bio_flagged(bio, BIO_QUIET);
+	struct bio *bio = &ioend->io_bio;
+	struct folio_iter fi;
 	u32 folio_count = 0;
 
-	for (bio = &ioend->io_inline_bio; bio; bio = next) {
-		struct folio_iter fi;
-
-		/*
-		 * For the last bio, bi_private points to the ioend, so we
-		 * need to explicitly end the iteration here.
-		 */
-		if (bio == last)
-			next = NULL;
-		else
-			next = bio->bi_private;
-
-		/* walk all folios in bio, ending page IO on them */
-		bio_for_each_folio_all(fi, bio) {
-			iomap_finish_folio_write(inode, fi.folio, fi.length,
-					error);
-			folio_count++;
+	if (error) {
+		mapping_set_error(inode->i_mapping, error);
+		if (!bio_flagged(bio, BIO_QUIET)) {
+			pr_err_ratelimited(
+"%s: writeback error on inode %lu, offset %lld, sector %llu",
+				inode->i_sb->s_id, inode->i_ino,
+				ioend->io_offset, ioend->io_sector);
 		}
-		bio_put(bio);
 	}
-	/* The ioend has been freed by bio_put() */
 
-	if (unlikely(error && !quiet)) {
-		printk_ratelimited(KERN_ERR
-"%s: writeback error on inode %lu, offset %lld, sector %llu",
-			inode->i_sb->s_id, inode->i_ino, offset, start);
+	/* walk all folios in bio, ending page IO on them */
+	bio_for_each_folio_all(fi, bio) {
+		if (error)
+			folio_set_error(fi.folio);
+		iomap_finish_folio_write(inode, fi.folio, fi.length);
+		folio_count++;
 	}
+
+	bio_put(bio);	/* frees the ioend */
 	return folio_count;
 }
 
@@ -1553,7 +1586,7 @@ EXPORT_SYMBOL_GPL(iomap_finish_ioends);
 static bool
 iomap_ioend_can_merge(struct iomap_ioend *ioend, struct iomap_ioend *next)
 {
-	if (ioend->io_bio->bi_status != next->io_bio->bi_status)
+	if (ioend->io_bio.bi_status != next->io_bio.bi_status)
 		return false;
 	if ((ioend->io_flags & IOMAP_F_SHARED) ^
 	    (next->io_flags & IOMAP_F_SHARED))
@@ -1618,47 +1651,46 @@ EXPORT_SYMBOL_GPL(iomap_sort_ioends);
 
 static void iomap_writepage_end_bio(struct bio *bio)
 {
-	struct iomap_ioend *ioend = bio->bi_private;
-
-	iomap_finish_ioend(ioend, blk_status_to_errno(bio->bi_status));
+	iomap_finish_ioend(iomap_ioend_from_bio(bio),
+			blk_status_to_errno(bio->bi_status));
 }
 
 /*
  * Submit the final bio for an ioend.
  *
  * If @error is non-zero, it means that we have a situation where some part of
- * the submission process has failed after we've marked pages for writeback
- * and unlocked them.  In this situation, we need to fail the bio instead of
- * submitting it.  This typically only happens on a filesystem shutdown.
+ * the submission process has failed after we've marked pages for writeback.
+ * We cannot cancel ioend directly in that case, so call the bio end I/O handler
+ * with the error status here to run the normal I/O completion handler to clear
+ * the writeback bit and let the file system proess the errors.
  */
-static int
-iomap_submit_ioend(struct iomap_writepage_ctx *wpc, struct iomap_ioend *ioend,
-		int error)
+static int iomap_submit_ioend(struct iomap_writepage_ctx *wpc, int error)
 {
-	ioend->io_bio->bi_private = ioend;
-	ioend->io_bio->bi_end_io = iomap_writepage_end_bio;
+	if (!wpc->ioend)
+		return error;
 
+	/*
+	 * Let the file systems prepare the I/O submission and hook in an I/O
+	 * comletion handler.  This also needs to happen in case after a
+	 * failure happened so that the file system end I/O handler gets called
+	 * to clean up.
+	 */
 	if (wpc->ops->prepare_ioend)
-		error = wpc->ops->prepare_ioend(ioend, error);
+		error = wpc->ops->prepare_ioend(wpc->ioend, error);
+
 	if (error) {
-		/*
-		 * If we're failing the IO now, just mark the ioend with an
-		 * error and finish it.  This will run IO completion immediately
-		 * as there is only one reference to the ioend at this point in
-		 * time.
-		 */
-		ioend->io_bio->bi_status = errno_to_blk_status(error);
-		bio_endio(ioend->io_bio);
-		return error;
+		wpc->ioend->io_bio.bi_status = errno_to_blk_status(error);
+		bio_endio(&wpc->ioend->io_bio);
+	} else {
+		submit_bio(&wpc->ioend->io_bio);
 	}
 
-	submit_bio(ioend->io_bio);
-	return 0;
+	wpc->ioend = NULL;
+	return error;
 }
 
-static struct iomap_ioend *
-iomap_alloc_ioend(struct inode *inode, struct iomap_writepage_ctx *wpc,
-		loff_t offset, sector_t sector, struct writeback_control *wbc)
+static struct iomap_ioend *iomap_alloc_ioend(struct iomap_writepage_ctx *wpc,
+		struct writeback_control *wbc, struct inode *inode, loff_t pos)
 {
 	struct iomap_ioend *ioend;
 	struct bio *bio;
@@ -1666,63 +1698,42 @@ iomap_alloc_ioend(struct inode *inode, struct iomap_writepage_ctx *wpc,
 	bio = bio_alloc_bioset(wpc->iomap.bdev, BIO_MAX_VECS,
 			       REQ_OP_WRITE | wbc_to_write_flags(wbc),
 			       GFP_NOFS, &iomap_ioend_bioset);
-	bio->bi_iter.bi_sector = sector;
+	bio->bi_iter.bi_sector = iomap_sector(&wpc->iomap, pos);
+	bio->bi_end_io = iomap_writepage_end_bio;
 	wbc_init_bio(wbc, bio);
+	bio->bi_write_hint = inode->i_write_hint;
 
-	ioend = container_of(bio, struct iomap_ioend, io_inline_bio);
+	ioend = iomap_ioend_from_bio(bio);
 	INIT_LIST_HEAD(&ioend->io_list);
 	ioend->io_type = wpc->iomap.type;
 	ioend->io_flags = wpc->iomap.flags;
 	ioend->io_inode = inode;
 	ioend->io_size = 0;
-	ioend->io_folios = 0;
-	ioend->io_offset = offset;
-	ioend->io_bio = bio;
-	ioend->io_sector = sector;
-	return ioend;
-}
-
-/*
- * Allocate a new bio, and chain the old bio to the new one.
- *
- * Note that we have to perform the chaining in this unintuitive order
- * so that the bi_private linkage is set up in the right direction for the
- * traversal in iomap_finish_ioend().
- */
-static struct bio *
-iomap_chain_bio(struct bio *prev)
-{
-	struct bio *new;
+	ioend->io_offset = pos;
+	ioend->io_sector = bio->bi_iter.bi_sector;
 
-	new = bio_alloc(prev->bi_bdev, BIO_MAX_VECS, prev->bi_opf, GFP_NOFS);
-	bio_clone_blkg_association(new, prev);
-	new->bi_iter.bi_sector = bio_end_sector(prev);
-
-	bio_chain(prev, new);
-	bio_get(prev);		/* for iomap_finish_ioend */
-	submit_bio(prev);
-	return new;
+	wpc->nr_folios = 0;
+	return ioend;
 }
 
-static bool
-iomap_can_add_to_ioend(struct iomap_writepage_ctx *wpc, loff_t offset,
-		sector_t sector)
+static bool iomap_can_add_to_ioend(struct iomap_writepage_ctx *wpc, loff_t pos)
 {
 	if ((wpc->iomap.flags & IOMAP_F_SHARED) !=
 	    (wpc->ioend->io_flags & IOMAP_F_SHARED))
 		return false;
 	if (wpc->iomap.type != wpc->ioend->io_type)
 		return false;
-	if (offset != wpc->ioend->io_offset + wpc->ioend->io_size)
+	if (pos != wpc->ioend->io_offset + wpc->ioend->io_size)
 		return false;
-	if (sector != bio_end_sector(wpc->ioend->io_bio))
+	if (iomap_sector(&wpc->iomap, pos) !=
+	    bio_end_sector(&wpc->ioend->io_bio))
 		return false;
 	/*
 	 * Limit ioend bio chain lengths to minimise IO completion latency. This
 	 * also prevents long tight loops ending page writeback on all the
 	 * folios in the ioend.
 	 */
-	if (wpc->ioend->io_folios >= IOEND_BATCH_SIZE)
+	if (wpc->nr_folios >= IOEND_BATCH_SIZE)
 		return false;
 	return true;
 }
@@ -1730,259 +1741,232 @@ iomap_can_add_to_ioend(struct iomap_writepage_ctx *wpc, loff_t offset,
 /*
  * Test to see if we have an existing ioend structure that we could append to
  * first; otherwise finish off the current ioend and start another.
+ *
+ * If a new ioend is created and cached, the old ioend is submitted to the block
+ * layer instantly.  Batching optimisations are provided by higher level block
+ * plugging.
+ *
+ * At the end of a writeback pass, there will be a cached ioend remaining on the
+ * writepage context that the caller will need to submit.
  */
-static void
-iomap_add_to_ioend(struct inode *inode, loff_t pos, struct folio *folio,
-		struct iomap_folio_state *ifs, struct iomap_writepage_ctx *wpc,
-		struct writeback_control *wbc, struct list_head *iolist)
+static int iomap_add_to_ioend(struct iomap_writepage_ctx *wpc,
+		struct writeback_control *wbc, struct folio *folio,
+		struct inode *inode, loff_t pos, unsigned len)
 {
-	sector_t sector = iomap_sector(&wpc->iomap, pos);
-	unsigned len = i_blocksize(inode);
+	struct iomap_folio_state *ifs = folio->private;
 	size_t poff = offset_in_folio(folio, pos);
+	int error;
 
-	if (!wpc->ioend || !iomap_can_add_to_ioend(wpc, pos, sector)) {
-		if (wpc->ioend)
-			list_add(&wpc->ioend->io_list, iolist);
-		wpc->ioend = iomap_alloc_ioend(inode, wpc, pos, sector, wbc);
+	if (!wpc->ioend || !iomap_can_add_to_ioend(wpc, pos)) {
+new_ioend:
+		error = iomap_submit_ioend(wpc, 0);
+		if (error)
+			return error;
+		wpc->ioend = iomap_alloc_ioend(wpc, wbc, inode, pos);
 	}
 
-	if (!bio_add_folio(wpc->ioend->io_bio, folio, len, poff)) {
-		wpc->ioend->io_bio = iomap_chain_bio(wpc->ioend->io_bio);
-		bio_add_folio_nofail(wpc->ioend->io_bio, folio, len, poff);
-	}
+	if (!bio_add_folio(&wpc->ioend->io_bio, folio, len, poff))
+		goto new_ioend;
 
 	if (ifs)
 		atomic_add(len, &ifs->write_bytes_pending);
 	wpc->ioend->io_size += len;
 	wbc_account_cgroup_owner(wbc, &folio->page, len);
+	return 0;
 }
 
-/*
- * We implement an immediate ioend submission policy here to avoid needing to
- * chain multiple ioends and hence nest mempool allocations which can violate
- * the forward progress guarantees we need to provide. The current ioend we're
- * adding blocks to is cached in the writepage context, and if the new block
- * doesn't append to the cached ioend, it will create a new ioend and cache that
- * instead.
- *
- * If a new ioend is created and cached, the old ioend is returned and queued
- * locally for submission once the entire page is processed or an error has been
- * detected.  While ioends are submitted immediately after they are completed,
- * batching optimisations are provided by higher level block plugging.
- *
- * At the end of a writeback pass, there will be a cached ioend remaining on the
- * writepage context that the caller will need to submit.
- */
-static int
-iomap_writepage_map(struct iomap_writepage_ctx *wpc,
-		struct writeback_control *wbc, struct inode *inode,
-		struct folio *folio, u64 end_pos)
+static int iomap_writepage_map_blocks(struct iomap_writepage_ctx *wpc,
+		struct writeback_control *wbc, struct folio *folio,
+		struct inode *inode, u64 pos, unsigned dirty_len,
+		unsigned *count)
 {
-	struct iomap_folio_state *ifs = folio->private;
-	struct iomap_ioend *ioend, *next;
-	unsigned len = i_blocksize(inode);
-	unsigned nblocks = i_blocks_per_folio(inode, folio);
-	u64 pos = folio_pos(folio);
-	int error = 0, count = 0, i;
-	LIST_HEAD(submit_list);
-
-	WARN_ON_ONCE(end_pos <= pos);
-
-	if (!ifs && nblocks > 1) {
-		ifs = ifs_alloc(inode, folio, 0);
-		iomap_set_range_dirty(folio, 0, end_pos - pos);
-	}
+	int error;
 
-	WARN_ON_ONCE(ifs && atomic_read(&ifs->write_bytes_pending) != 0);
-
-	/*
-	 * Walk through the folio to find areas to write back. If we
-	 * run off the end of the current map or find the current map
-	 * invalid, grab a new one.
-	 */
-	for (i = 0; i < nblocks && pos < end_pos; i++, pos += len) {
-		if (ifs && !ifs_block_is_dirty(folio, ifs, i))
-			continue;
+	do {
+		unsigned map_len;
 
-		error = wpc->ops->map_blocks(wpc, inode, pos);
+		error = wpc->ops->map_blocks(wpc, inode, pos, dirty_len);
 		if (error)
 			break;
-		trace_iomap_writepage_map(inode, &wpc->iomap);
-		if (WARN_ON_ONCE(wpc->iomap.type == IOMAP_INLINE))
-			continue;
-		if (wpc->iomap.type == IOMAP_HOLE)
-			continue;
-		iomap_add_to_ioend(inode, pos, folio, ifs, wpc, wbc,
-				 &submit_list);
-		count++;
-	}
-	if (count)
-		wpc->ioend->io_folios++;
+		trace_iomap_writepage_map(inode, pos, dirty_len, &wpc->iomap);
 
-	WARN_ON_ONCE(!wpc->ioend && !list_empty(&submit_list));
-	WARN_ON_ONCE(!folio_test_locked(folio));
-	WARN_ON_ONCE(folio_test_writeback(folio));
-	WARN_ON_ONCE(folio_test_dirty(folio));
+		map_len = min_t(u64, dirty_len,
+			wpc->iomap.offset + wpc->iomap.length - pos);
+		WARN_ON_ONCE(!folio->private && map_len < dirty_len);
+
+		switch (wpc->iomap.type) {
+		case IOMAP_INLINE:
+			WARN_ON_ONCE(1);
+			error = -EIO;
+			break;
+		case IOMAP_HOLE:
+			break;
+		default:
+			error = iomap_add_to_ioend(wpc, wbc, folio, inode, pos,
+					map_len);
+			if (!error)
+				(*count)++;
+			break;
+		}
+		dirty_len -= map_len;
+		pos += map_len;
+	} while (dirty_len && !error);
 
 	/*
 	 * We cannot cancel the ioend directly here on error.  We may have
 	 * already set other pages under writeback and hence we have to run I/O
 	 * completion to mark the error state of the pages under writeback
 	 * appropriately.
+	 *
+	 * Just let the file system know what portion of the folio failed to
+	 * map.
 	 */
-	if (unlikely(error)) {
-		/*
-		 * Let the filesystem know what portion of the current page
-		 * failed to map.
-		 */
-		if (wpc->ops->discard_folio)
-			wpc->ops->discard_folio(folio, pos);
-	}
-
-	/*
-	 * We can have dirty bits set past end of file in page_mkwrite path
-	 * while mapping the last partial folio. Hence it's better to clear
-	 * all the dirty bits in the folio here.
-	 */
-	iomap_clear_range_dirty(folio, 0, folio_size(folio));
-
-	/*
-	 * If the page hasn't been added to the ioend, it won't be affected by
-	 * I/O completion and we must unlock it now.
-	 */
-	if (error && !count) {
-		folio_unlock(folio);
-		goto done;
-	}
-
-	folio_start_writeback(folio);
-	folio_unlock(folio);
-
-	/*
-	 * Preserve the original error if there was one; catch
-	 * submission errors here and propagate into subsequent ioend
-	 * submissions.
-	 */
-	list_for_each_entry_safe(ioend, next, &submit_list, io_list) {
-		int error2;
-
-		list_del_init(&ioend->io_list);
-		error2 = iomap_submit_ioend(wpc, ioend, error);
-		if (error2 && !error)
-			error = error2;
-	}
-
-	/*
-	 * We can end up here with no error and nothing to write only if we race
-	 * with a partial page truncate on a sub-page block sized filesystem.
-	 */
-	if (!count)
-		folio_end_writeback(folio);
-done:
-	mapping_set_error(inode->i_mapping, error);
+	if (error && wpc->ops->discard_folio)
+		wpc->ops->discard_folio(folio, pos);
 	return error;
 }
 
 /*
- * Write out a dirty page.
+ * Check interaction of the folio with the file end.
  *
- * For delalloc space on the page, we need to allocate space and flush it.
- * For unwritten space on the page, we need to start the conversion to
- * regular allocated space.
+ * If the folio is entirely beyond i_size, return false.  If it straddles
+ * i_size, adjust end_pos and zero all data beyond i_size.
  */
-static int iomap_do_writepage(struct folio *folio,
-		struct writeback_control *wbc, void *data)
+static bool iomap_writepage_handle_eof(struct folio *folio, struct inode *inode,
+		u64 *end_pos)
 {
-	struct iomap_writepage_ctx *wpc = data;
-	struct inode *inode = folio->mapping->host;
-	u64 end_pos, isize;
+	u64 isize = i_size_read(inode);
 
-	trace_iomap_writepage(inode, folio_pos(folio), folio_size(folio));
-
-	/*
-	 * Refuse to write the folio out if we're called from reclaim context.
-	 *
-	 * This avoids stack overflows when called from deeply used stacks in
-	 * random callers for direct reclaim or memcg reclaim.  We explicitly
-	 * allow reclaim from kswapd as the stack usage there is relatively low.
-	 *
-	 * This should never happen except in the case of a VM regression so
-	 * warn about it.
-	 */
-	if (WARN_ON_ONCE((current->flags & (PF_MEMALLOC|PF_KSWAPD)) ==
-			PF_MEMALLOC))
-		goto redirty;
-
-	/*
-	 * Is this folio beyond the end of the file?
-	 *
-	 * The folio index is less than the end_index, adjust the end_pos
-	 * to the highest offset that this folio should represent.
-	 * -----------------------------------------------------
-	 * |			file mapping	       | <EOF> |
-	 * -----------------------------------------------------
-	 * | Page ... | Page N-2 | Page N-1 |  Page N  |       |
-	 * ^--------------------------------^----------|--------
-	 * |     desired writeback range    |      see else    |
-	 * ---------------------------------^------------------|
-	 */
-	isize = i_size_read(inode);
-	end_pos = folio_pos(folio) + folio_size(folio);
-	if (end_pos > isize) {
-		/*
-		 * Check whether the page to write out is beyond or straddles
-		 * i_size or not.
-		 * -------------------------------------------------------
-		 * |		file mapping		        | <EOF>  |
-		 * -------------------------------------------------------
-		 * | Page ... | Page N-2 | Page N-1 |  Page N   | Beyond |
-		 * ^--------------------------------^-----------|---------
-		 * |				    |      Straddles     |
-		 * ---------------------------------^-----------|--------|
-		 */
+	if (*end_pos > isize) {
 		size_t poff = offset_in_folio(folio, isize);
 		pgoff_t end_index = isize >> PAGE_SHIFT;
 
 		/*
-		 * Skip the page if it's fully outside i_size, e.g.
-		 * due to a truncate operation that's in progress.  We've
-		 * cleaned this page and truncate will finish things off for
-		 * us.
+		 * If the folio is entirely ouside of i_size, skip it.
+		 *
+		 * This can happen due to a truncate operation that is in
+		 * progress and in that case truncate will finish it off once
+		 * we've dropped the folio lock.
 		 *
-		 * Note that the end_index is unsigned long.  If the given
-		 * offset is greater than 16TB on a 32-bit system then if we
-		 * checked if the page is fully outside i_size with
-		 * "if (page->index >= end_index + 1)", "end_index + 1" would
-		 * overflow and evaluate to 0.  Hence this page would be
+		 * Note that the pgoff_t used for end_index is an unsigned long.
+		 * If the given offset is greater than 16TB on a 32-bit system,
+		 * then if we checked if the folio is fully outside i_size with
+		 * "if (folio->index >= end_index + 1)", "end_index + 1" would
+		 * overflow and evaluate to 0.  Hence this folio would be
 		 * redirtied and written out repeatedly, which would result in
 		 * an infinite loop; the user program performing this operation
 		 * would hang.  Instead, we can detect this situation by
-		 * checking if the page is totally beyond i_size or if its
+		 * checking if the folio is totally beyond i_size or if its
 		 * offset is just equal to the EOF.
 		 */
 		if (folio->index > end_index ||
 		    (folio->index == end_index && poff == 0))
-			goto unlock;
+			return false;
 
 		/*
-		 * The page straddles i_size.  It must be zeroed out on each
-		 * and every writepage invocation because it may be mmapped.
-		 * "A file is mapped in multiples of the page size.  For a file
-		 * that is not a multiple of the page size, the remaining
-		 * memory is zeroed when mapped, and writes to that region are
-		 * not written out to the file."
+		 * The folio straddles i_size.
+		 *
+		 * It must be zeroed out on each and every writepage invocation
+		 * because it may be mmapped:
+		 *
+		 *    A file is mapped in multiples of the page size.  For a
+		 *    file that is not a multiple of the page size, the
+		 *    remaining memory is zeroed when mapped, and writes to that
+		 *    region are not written out to the file.
+		 *
+		 * Also adjust the writeback range to skip all blocks entirely
+		 * beyond i_size.
 		 */
 		folio_zero_segment(folio, poff, folio_size(folio));
-		end_pos = isize;
+		*end_pos = round_up(isize, i_blocksize(inode));
+	}
+
+	return true;
+}
+
+static int iomap_writepage_map(struct iomap_writepage_ctx *wpc,
+		struct writeback_control *wbc, struct folio *folio)
+{
+	struct iomap_folio_state *ifs = folio->private;
+	struct inode *inode = folio->mapping->host;
+	u64 pos = folio_pos(folio);
+	u64 end_pos = pos + folio_size(folio);
+	unsigned count = 0;
+	int error = 0;
+	u32 rlen;
+
+	WARN_ON_ONCE(!folio_test_locked(folio));
+	WARN_ON_ONCE(folio_test_dirty(folio));
+	WARN_ON_ONCE(folio_test_writeback(folio));
+
+	trace_iomap_writepage(inode, pos, folio_size(folio));
+
+	if (!iomap_writepage_handle_eof(folio, inode, &end_pos)) {
+		folio_unlock(folio);
+		return 0;
 	}
+	WARN_ON_ONCE(end_pos <= pos);
 
-	return iomap_writepage_map(wpc, wbc, inode, folio, end_pos);
+	if (i_blocks_per_folio(inode, folio) > 1) {
+		if (!ifs) {
+			ifs = ifs_alloc(inode, folio, 0);
+			iomap_set_range_dirty(folio, 0, end_pos - pos);
+		}
 
-redirty:
-	folio_redirty_for_writepage(wbc, folio);
-unlock:
+		/*
+		 * Keep the I/O completion handler from clearing the writeback
+		 * bit until we have submitted all blocks by adding a bias to
+		 * ifs->write_bytes_pending, which is dropped after submitting
+		 * all blocks.
+		 */
+		WARN_ON_ONCE(atomic_read(&ifs->write_bytes_pending) != 0);
+		atomic_inc(&ifs->write_bytes_pending);
+	}
+
+	/*
+	 * Set the writeback bit ASAP, as the I/O completion for the single
+	 * block per folio case happen hit as soon as we're submitting the bio.
+	 */
+	folio_start_writeback(folio);
+
+	/*
+	 * Walk through the folio to find dirty areas to write back.
+	 */
+	while ((rlen = iomap_find_dirty_range(folio, &pos, end_pos))) {
+		error = iomap_writepage_map_blocks(wpc, wbc, folio, inode,
+				pos, rlen, &count);
+		if (error)
+			break;
+		pos += rlen;
+	}
+
+	if (count)
+		wpc->nr_folios++;
+
+	/*
+	 * We can have dirty bits set past end of file in page_mkwrite path
+	 * while mapping the last partial folio. Hence it's better to clear
+	 * all the dirty bits in the folio here.
+	 */
+	iomap_clear_range_dirty(folio, 0, folio_size(folio));
+
+	/*
+	 * Usually the writeback bit is cleared by the I/O completion handler.
+	 * But we may end up either not actually writing any blocks, or (when
+	 * there are multiple blocks in a folio) all I/O might have finished
+	 * already at this point.  In that case we need to clear the writeback
+	 * bit ourselves right after unlocking the page.
+	 */
 	folio_unlock(folio);
-	return 0;
+	if (ifs) {
+		if (atomic_dec_and_test(&ifs->write_bytes_pending))
+			folio_end_writeback(folio);
+	} else {
+		if (!count)
+			folio_end_writeback(folio);
+	}
+	mapping_set_error(inode->i_mapping, error);
+	return error;
 }
 
 int
@@ -1990,20 +1974,28 @@ iomap_writepages(struct address_space *mapping, struct writeback_control *wbc,
 		struct iomap_writepage_ctx *wpc,
 		const struct iomap_writeback_ops *ops)
 {
-	int			ret;
+	struct folio *folio = NULL;
+	int error;
+
+	/*
+	 * Writeback from reclaim context should never happen except in the case
+	 * of a VM regression so warn about it and refuse to write the data.
+	 */
+	if (WARN_ON_ONCE((current->flags & (PF_MEMALLOC | PF_KSWAPD)) ==
+			PF_MEMALLOC))
+		return -EIO;
 
 	wpc->ops = ops;
-	ret = write_cache_pages(mapping, wbc, iomap_do_writepage, wpc);
-	if (!wpc->ioend)
-		return ret;
-	return iomap_submit_ioend(wpc, wpc->ioend, ret);
+	while ((folio = writeback_iter(mapping, wbc, folio, &error)))
+		error = iomap_writepage_map(wpc, wbc, folio);
+	return iomap_submit_ioend(wpc, error);
 }
 EXPORT_SYMBOL_GPL(iomap_writepages);
 
 static int __init iomap_init(void)
 {
 	return bioset_init(&iomap_ioend_bioset, 4 * (PAGE_SIZE / SECTOR_SIZE),
-			   offsetof(struct iomap_ioend, io_inline_bio),
+			   offsetof(struct iomap_ioend, io_bio),
 			   BIOSET_NEED_BVECS);
 }
 fs_initcall(iomap_init);
diff --git a/fs/iomap/direct-io.c b/fs/iomap/direct-io.c
index bcd3f8cf5e..f3b43d223a 100644
--- a/fs/iomap/direct-io.c
+++ b/fs/iomap/direct-io.c
@@ -380,6 +380,7 @@ static loff_t iomap_dio_bio_iter(const struct iomap_iter *iter,
 		fscrypt_set_bio_crypt_ctx(bio, inode, pos >> inode->i_blkbits,
 					  GFP_KERNEL);
 		bio->bi_iter.bi_sector = iomap_sector(iomap, pos);
+		bio->bi_write_hint = inode->i_write_hint;
 		bio->bi_ioprio = dio->iocb->ki_ioprio;
 		bio->bi_private = dio;
 		bio->bi_end_io = iomap_dio_bio_end_io;
diff --git a/fs/iomap/trace.h b/fs/iomap/trace.h
index c16fd55f55..0a991c4ce8 100644
--- a/fs/iomap/trace.h
+++ b/fs/iomap/trace.h
@@ -154,7 +154,48 @@ DEFINE_EVENT(iomap_class, name,	\
 	TP_ARGS(inode, iomap))
 DEFINE_IOMAP_EVENT(iomap_iter_dstmap);
 DEFINE_IOMAP_EVENT(iomap_iter_srcmap);
-DEFINE_IOMAP_EVENT(iomap_writepage_map);
+
+TRACE_EVENT(iomap_writepage_map,
+	TP_PROTO(struct inode *inode, u64 pos, unsigned int dirty_len,
+		 struct iomap *iomap),
+	TP_ARGS(inode, pos, dirty_len, iomap),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(u64, ino)
+		__field(u64, pos)
+		__field(u64, dirty_len)
+		__field(u64, addr)
+		__field(loff_t, offset)
+		__field(u64, length)
+		__field(u16, type)
+		__field(u16, flags)
+		__field(dev_t, bdev)
+	),
+	TP_fast_assign(
+		__entry->dev = inode->i_sb->s_dev;
+		__entry->ino = inode->i_ino;
+		__entry->pos = pos;
+		__entry->dirty_len = dirty_len;
+		__entry->addr = iomap->addr;
+		__entry->offset = iomap->offset;
+		__entry->length = iomap->length;
+		__entry->type = iomap->type;
+		__entry->flags = iomap->flags;
+		__entry->bdev = iomap->bdev ? iomap->bdev->bd_dev : 0;
+	),
+	TP_printk("dev %d:%d ino 0x%llx bdev %d:%d pos 0x%llx dirty len 0x%llx "
+		  "addr 0x%llx offset 0x%llx length 0x%llx type %s flags %s",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  MAJOR(__entry->bdev), MINOR(__entry->bdev),
+		  __entry->pos,
+		  __entry->dirty_len,
+		  __entry->addr,
+		  __entry->offset,
+		  __entry->length,
+		  __print_symbolic(__entry->type, IOMAP_TYPE_STRINGS),
+		  __print_flags(__entry->flags, "|", IOMAP_F_FLAGS_STRINGS))
+);
 
 TRACE_EVENT(iomap_iter,
 	TP_PROTO(struct iomap_iter *iter, const void *ops,
@@ -165,6 +206,7 @@ TRACE_EVENT(iomap_iter,
 		__field(u64, ino)
 		__field(loff_t, pos)
 		__field(u64, length)
+		__field(s64, processed)
 		__field(unsigned int, flags)
 		__field(const void *, ops)
 		__field(unsigned long, caller)
@@ -174,15 +216,17 @@ TRACE_EVENT(iomap_iter,
 		__entry->ino = iter->inode->i_ino;
 		__entry->pos = iter->pos;
 		__entry->length = iomap_length(iter);
+		__entry->processed = iter->processed;
 		__entry->flags = iter->flags;
 		__entry->ops = ops;
 		__entry->caller = caller;
 	),
-	TP_printk("dev %d:%d ino 0x%llx pos 0x%llx length 0x%llx flags %s (0x%x) ops %ps caller %pS",
+	TP_printk("dev %d:%d ino 0x%llx pos 0x%llx length 0x%llx processed %lld flags %s (0x%x) ops %ps caller %pS",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		   __entry->ino,
 		   __entry->pos,
 		   __entry->length,
+		   __entry->processed,
 		   __print_flags(__entry->flags, "|", IOMAP_FLAGS_STRINGS),
 		   __entry->flags,
 		   __entry->ops,
diff --git a/fs/isofs/Makefile b/fs/isofs/Makefile
index 6498fd2b0f..b25bc542a2 100644
--- a/fs/isofs/Makefile
+++ b/fs/isofs/Makefile
@@ -5,7 +5,6 @@
 
 obj-$(CONFIG_ISO9660_FS) += isofs.o
 
-isofs-objs-y 			:= namei.o inode.o dir.o util.o rock.o export.o
-isofs-objs-$(CONFIG_JOLIET)	+= joliet.o
-isofs-objs-$(CONFIG_ZISOFS)	+= compress.o
-isofs-objs			:= $(isofs-objs-y)
+isofs-y 		:= namei.o inode.o dir.o util.o rock.o export.o
+isofs-$(CONFIG_JOLIET)	+= joliet.o
+isofs-$(CONFIG_ZISOFS)	+= compress.o
diff --git a/fs/isofs/compress.c b/fs/isofs/compress.c
index c4da3f634b..34d5baa5d8 100644
--- a/fs/isofs/compress.c
+++ b/fs/isofs/compress.c
@@ -346,8 +346,6 @@ static int zisofs_read_folio(struct file *file, struct folio *folio)
 	for (i = 0; i < pcount; i++, index++) {
 		if (i != full_page)
 			pages[i] = grab_cache_page_nowait(mapping, index);
-		if (pages[i])
-			ClearPageError(pages[i]);
 	}
 
 	err = zisofs_fill_pages(inode, full_page, pcount, pages);
@@ -356,8 +354,6 @@ static int zisofs_read_folio(struct file *file, struct folio *folio)
 	for (i = 0; i < pcount; i++) {
 		if (pages[i]) {
 			flush_dcache_page(pages[i]);
-			if (i == full_page && err)
-				SetPageError(pages[i]);
 			unlock_page(pages[i]);
 			if (i != full_page)
 				put_page(pages[i]);
diff --git a/fs/isofs/inode.c b/fs/isofs/inode.c
index 379c9edc90..93b1077a38 100644
--- a/fs/isofs/inode.c
+++ b/fs/isofs/inode.c
@@ -21,11 +21,12 @@
 #include <linux/ctype.h>
 #include <linux/statfs.h>
 #include <linux/cdrom.h>
-#include <linux/parser.h>
 #include <linux/mpage.h>
 #include <linux/user_namespace.h>
 #include <linux/seq_file.h>
 #include <linux/blkdev.h>
+#include <linux/fs_context.h>
+#include <linux/fs_parser.h>
 
 #include "isofs.h"
 #include "zisofs.h"
@@ -93,7 +94,7 @@ static int __init init_inodecache(void)
 	isofs_inode_cachep = kmem_cache_create("isofs_inode_cache",
 					sizeof(struct iso_inode_info),
 					0, (SLAB_RECLAIM_ACCOUNT|
-					SLAB_MEM_SPREAD|SLAB_ACCOUNT),
+					SLAB_ACCOUNT),
 					init_once);
 	if (!isofs_inode_cachep)
 		return -ENOMEM;
@@ -110,10 +111,10 @@ static void destroy_inodecache(void)
 	kmem_cache_destroy(isofs_inode_cachep);
 }
 
-static int isofs_remount(struct super_block *sb, int *flags, char *data)
+static int isofs_reconfigure(struct fs_context *fc)
 {
-	sync_filesystem(sb);
-	if (!(*flags & SB_RDONLY))
+	sync_filesystem(fc->root->d_sb);
+	if (!(fc->sb_flags & SB_RDONLY))
 		return -EROFS;
 	return 0;
 }
@@ -123,7 +124,6 @@ static const struct super_operations isofs_sops = {
 	.free_inode	= isofs_free_inode,
 	.put_super	= isofs_put_super,
 	.statfs		= isofs_statfs,
-	.remount_fs	= isofs_remount,
 	.show_options	= isofs_show_options,
 };
 
@@ -145,7 +145,7 @@ static const struct dentry_operations isofs_dentry_ops[] = {
 #endif
 };
 
-struct iso9660_options{
+struct isofs_options{
 	unsigned int rock:1;
 	unsigned int joliet:1;
 	unsigned int cruft:1;
@@ -289,197 +289,161 @@ isofs_dentry_cmpi_ms(const struct dentry *dentry,
 #endif
 
 enum {
-	Opt_block, Opt_check_r, Opt_check_s, Opt_cruft, Opt_gid, Opt_ignore,
-	Opt_iocharset, Opt_map_a, Opt_map_n, Opt_map_o, Opt_mode, Opt_nojoliet,
-	Opt_norock, Opt_sb, Opt_session, Opt_uid, Opt_unhide, Opt_utf8, Opt_err,
-	Opt_nocompress, Opt_hide, Opt_showassoc, Opt_dmode, Opt_overriderockperm,
+	Opt_block, Opt_check, Opt_cruft, Opt_gid, Opt_ignore, Opt_iocharset,
+	Opt_map, Opt_mode, Opt_nojoliet, Opt_norock, Opt_sb, Opt_session,
+	Opt_uid, Opt_unhide, Opt_utf8, Opt_err, Opt_nocompress, Opt_hide,
+	Opt_showassoc, Opt_dmode, Opt_overriderockperm,
 };
 
-static const match_table_t tokens = {
-	{Opt_norock, "norock"},
-	{Opt_nojoliet, "nojoliet"},
-	{Opt_unhide, "unhide"},
-	{Opt_hide, "hide"},
-	{Opt_showassoc, "showassoc"},
-	{Opt_cruft, "cruft"},
-	{Opt_utf8, "utf8"},
-	{Opt_iocharset, "iocharset=%s"},
-	{Opt_map_a, "map=acorn"},
-	{Opt_map_a, "map=a"},
-	{Opt_map_n, "map=normal"},
-	{Opt_map_n, "map=n"},
-	{Opt_map_o, "map=off"},
-	{Opt_map_o, "map=o"},
-	{Opt_session, "session=%u"},
-	{Opt_sb, "sbsector=%u"},
-	{Opt_check_r, "check=relaxed"},
-	{Opt_check_r, "check=r"},
-	{Opt_check_s, "check=strict"},
-	{Opt_check_s, "check=s"},
-	{Opt_uid, "uid=%u"},
-	{Opt_gid, "gid=%u"},
-	{Opt_mode, "mode=%u"},
-	{Opt_dmode, "dmode=%u"},
-	{Opt_overriderockperm, "overriderockperm"},
-	{Opt_block, "block=%u"},
-	{Opt_ignore, "conv=binary"},
-	{Opt_ignore, "conv=b"},
-	{Opt_ignore, "conv=text"},
-	{Opt_ignore, "conv=t"},
-	{Opt_ignore, "conv=mtext"},
-	{Opt_ignore, "conv=m"},
-	{Opt_ignore, "conv=auto"},
-	{Opt_ignore, "conv=a"},
-	{Opt_nocompress, "nocompress"},
-	{Opt_err, NULL}
+static const struct constant_table isofs_param_map[] = {
+	{"acorn",	'a'},
+	{"a",		'a'},
+	{"normal",	'n'},
+	{"n",		'n'},
+	{"off",		'o'},
+	{"o",		'o'},
+	{}
 };
 
-static int parse_options(char *options, struct iso9660_options *popt)
-{
-	char *p;
-	int option;
-	unsigned int uv;
-
-	popt->map = 'n';
-	popt->rock = 1;
-	popt->joliet = 1;
-	popt->cruft = 0;
-	popt->hide = 0;
-	popt->showassoc = 0;
-	popt->check = 'u';		/* unset */
-	popt->nocompress = 0;
-	popt->blocksize = 1024;
-	popt->fmode = popt->dmode = ISOFS_INVALID_MODE;
-	popt->uid_set = 0;
-	popt->gid_set = 0;
-	popt->gid = GLOBAL_ROOT_GID;
-	popt->uid = GLOBAL_ROOT_UID;
-	popt->iocharset = NULL;
-	popt->overriderockperm = 0;
-	popt->session=-1;
-	popt->sbsector=-1;
-	if (!options)
-		return 1;
-
-	while ((p = strsep(&options, ",")) != NULL) {
-		int token;
-		substring_t args[MAX_OPT_ARGS];
-		unsigned n;
-
-		if (!*p)
-			continue;
+static const struct constant_table isofs_param_check[] = {
+	{"relaxed",	'r'},
+	{"r",		'r'},
+	{"strict",	's'},
+	{"s",		's'},
+	{}
+};
 
-		token = match_token(p, tokens, args);
-		switch (token) {
-		case Opt_norock:
-			popt->rock = 0;
-			break;
-		case Opt_nojoliet:
-			popt->joliet = 0;
-			break;
-		case Opt_hide:
-			popt->hide = 1;
-			break;
-		case Opt_unhide:
-		case Opt_showassoc:
-			popt->showassoc = 1;
-			break;
-		case Opt_cruft:
-			popt->cruft = 1;
-			break;
+static const struct fs_parameter_spec isofs_param_spec[] = {
+	fsparam_flag	("norock",		Opt_norock),
+	fsparam_flag	("nojoliet",		Opt_nojoliet),
+	fsparam_flag	("unhide",		Opt_unhide),
+	fsparam_flag	("hide",		Opt_hide),
+	fsparam_flag	("showassoc",		Opt_showassoc),
+	fsparam_flag	("cruft",		Opt_cruft),
+	fsparam_flag	("utf8",		Opt_utf8),
+	fsparam_string	("iocharset",		Opt_iocharset),
+	fsparam_enum	("map",			Opt_map, isofs_param_map),
+	fsparam_u32	("session",		Opt_session),
+	fsparam_u32	("sbsector",		Opt_sb),
+	fsparam_enum	("check",		Opt_check, isofs_param_check),
+	fsparam_u32	("uid",			Opt_uid),
+	fsparam_u32	("gid",			Opt_gid),
+	/* Note: mode/dmode historically accepted %u not strictly %o */
+	fsparam_u32	("mode",		Opt_mode),
+	fsparam_u32	("dmode",		Opt_dmode),
+	fsparam_flag	("overriderockperm",	Opt_overriderockperm),
+	fsparam_u32	("block",		Opt_block),
+	fsparam_string	("conv",		Opt_ignore),
+	fsparam_flag	("nocompress",		Opt_nocompress),
+	{}
+};
+
+static int isofs_parse_param(struct fs_context *fc,
+			       struct fs_parameter *param)
+{
+	struct isofs_options *popt = fc->fs_private;
+	struct fs_parse_result result;
+	int opt;
+	kuid_t uid;
+	kgid_t gid;
+	unsigned int n;
+
+	/* There are no remountable options */
+	if (fc->purpose == FS_CONTEXT_FOR_RECONFIGURE)
+		return 0;
+
+	opt = fs_parse(fc, isofs_param_spec, param, &result);
+	if (opt < 0)
+		return opt;
+
+	switch (opt) {
+	case Opt_norock:
+		popt->rock = 0;
+		break;
+	case Opt_nojoliet:
+		popt->joliet = 0;
+		break;
+	case Opt_hide:
+		popt->hide = 1;
+		break;
+	case Opt_unhide:
+	case Opt_showassoc:
+		popt->showassoc = 1;
+		break;
+	case Opt_cruft:
+		popt->cruft = 1;
+		break;
 #ifdef CONFIG_JOLIET
-		case Opt_utf8:
-			kfree(popt->iocharset);
-			popt->iocharset = kstrdup("utf8", GFP_KERNEL);
-			if (!popt->iocharset)
-				return 0;
-			break;
-		case Opt_iocharset:
-			kfree(popt->iocharset);
-			popt->iocharset = match_strdup(&args[0]);
-			if (!popt->iocharset)
-				return 0;
-			break;
+	case Opt_utf8:
+		kfree(popt->iocharset);
+		popt->iocharset = kstrdup("utf8", GFP_KERNEL);
+		if (!popt->iocharset)
+			return -ENOMEM;
+		break;
+	case Opt_iocharset:
+		kfree(popt->iocharset);
+		popt->iocharset = kstrdup(param->string, GFP_KERNEL);
+		if (!popt->iocharset)
+			return -ENOMEM;
+		break;
 #endif
-		case Opt_map_a:
-			popt->map = 'a';
-			break;
-		case Opt_map_o:
-			popt->map = 'o';
-			break;
-		case Opt_map_n:
-			popt->map = 'n';
-			break;
-		case Opt_session:
-			if (match_int(&args[0], &option))
-				return 0;
-			n = option;
-			/*
-			 * Track numbers are supposed to be in range 1-99, the
-			 * mount option starts indexing at 0.
-			 */
-			if (n >= 99)
-				return 0;
-			popt->session = n + 1;
-			break;
-		case Opt_sb:
-			if (match_int(&args[0], &option))
-				return 0;
-			popt->sbsector = option;
-			break;
-		case Opt_check_r:
-			popt->check = 'r';
-			break;
-		case Opt_check_s:
-			popt->check = 's';
-			break;
-		case Opt_ignore:
-			break;
-		case Opt_uid:
-			if (match_uint(&args[0], &uv))
-				return 0;
-			popt->uid = make_kuid(current_user_ns(), uv);
-			if (!uid_valid(popt->uid))
-				return 0;
-			popt->uid_set = 1;
-			break;
-		case Opt_gid:
-			if (match_uint(&args[0], &uv))
-				return 0;
-			popt->gid = make_kgid(current_user_ns(), uv);
-			if (!gid_valid(popt->gid))
-				return 0;
-			popt->gid_set = 1;
-			break;
-		case Opt_mode:
-			if (match_int(&args[0], &option))
-				return 0;
-			popt->fmode = option;
-			break;
-		case Opt_dmode:
-			if (match_int(&args[0], &option))
-				return 0;
-			popt->dmode = option;
-			break;
-		case Opt_overriderockperm:
-			popt->overriderockperm = 1;
-			break;
-		case Opt_block:
-			if (match_int(&args[0], &option))
-				return 0;
-			n = option;
-			if (n != 512 && n != 1024 && n != 2048)
-				return 0;
-			popt->blocksize = n;
-			break;
-		case Opt_nocompress:
-			popt->nocompress = 1;
-			break;
-		default:
-			return 0;
-		}
+	case Opt_map:
+		popt->map = result.uint_32;
+		break;
+	case Opt_session:
+		n = result.uint_32;
+		/*
+		 * Track numbers are supposed to be in range 1-99, the
+		 * mount option starts indexing at 0.
+		 */
+		if (n >= 99)
+			return -EINVAL;
+		popt->session = n + 1;
+		break;
+	case Opt_sb:
+		popt->sbsector = result.uint_32;
+		break;
+	case Opt_check:
+		popt->check = result.uint_32;
+		break;
+	case Opt_ignore:
+		break;
+	case Opt_uid:
+		uid = make_kuid(current_user_ns(), result.uint_32);
+		if (!uid_valid(uid))
+			return -EINVAL;
+		popt->uid = uid;
+		popt->uid_set = 1;
+		break;
+	case Opt_gid:
+		gid = make_kgid(current_user_ns(), result.uint_32);
+		if (!gid_valid(gid))
+			return -EINVAL;
+		popt->gid = gid;
+		popt->gid_set = 1;
+		break;
+	case Opt_mode:
+		popt->fmode = result.uint_32;
+		break;
+	case Opt_dmode:
+		popt->dmode = result.uint_32;
+		break;
+	case Opt_overriderockperm:
+		popt->overriderockperm = 1;
+		break;
+	case Opt_block:
+		n = result.uint_32;
+		if (n != 512 && n != 1024 && n != 2048)
+			return -EINVAL;
+		popt->blocksize = n;
+		break;
+	case Opt_nocompress:
+		popt->nocompress = 1;
+		break;
+	default:
+		return -EINVAL;
 	}
-	return 1;
+	return 0;
 }
 
 /*
@@ -615,7 +579,7 @@ static bool rootdir_empty(struct super_block *sb, unsigned long block)
 /*
  * Initialize the superblock and read the root inode.
  */
-static int isofs_fill_super(struct super_block *s, void *data, int silent)
+static int isofs_fill_super(struct super_block *s, struct fs_context *fc)
 {
 	struct buffer_head *bh = NULL, *pri_bh = NULL;
 	struct hs_primary_descriptor *h_pri = NULL;
@@ -623,7 +587,7 @@ static int isofs_fill_super(struct super_block *s, void *data, int silent)
 	struct iso_supplementary_descriptor *sec = NULL;
 	struct iso_directory_record *rootp;
 	struct inode *inode;
-	struct iso9660_options opt;
+	struct isofs_options *opt = fc->fs_private;
 	struct isofs_sb_info *sbi;
 	unsigned long first_data_zone;
 	int joliet_level = 0;
@@ -631,15 +595,13 @@ static int isofs_fill_super(struct super_block *s, void *data, int silent)
 	int orig_zonesize;
 	int table, error = -EINVAL;
 	unsigned int vol_desc_start;
+	int silent = fc->sb_flags & SB_SILENT;
 
 	sbi = kzalloc(sizeof(*sbi), GFP_KERNEL);
 	if (!sbi)
 		return -ENOMEM;
 	s->s_fs_info = sbi;
 
-	if (!parse_options((char *)data, &opt))
-		goto out_freesbi;
-
 	/*
 	 * First of all, get the hardware blocksize for this device.
 	 * If we don't know what it is, or the hardware blocksize is
@@ -655,14 +617,14 @@ static int isofs_fill_super(struct super_block *s, void *data, int silent)
 			bdev_logical_block_size(s->s_bdev));
 		goto out_freesbi;
 	}
-	opt.blocksize = sb_min_blocksize(s, opt.blocksize);
+	opt->blocksize = sb_min_blocksize(s, opt->blocksize);
 
 	sbi->s_high_sierra = 0; /* default is iso9660 */
-	sbi->s_session = opt.session;
-	sbi->s_sbsector = opt.sbsector;
+	sbi->s_session = opt->session;
+	sbi->s_sbsector = opt->sbsector;
 
-	vol_desc_start = (opt.sbsector != -1) ?
-		opt.sbsector : isofs_get_last_session(s,opt.session);
+	vol_desc_start = (opt->sbsector != -1) ?
+		opt->sbsector : isofs_get_last_session(s, opt->session);
 
 	for (iso_blknum = vol_desc_start+16;
 		iso_blknum < vol_desc_start+100; iso_blknum++) {
@@ -696,7 +658,7 @@ static int isofs_fill_super(struct super_block *s, void *data, int silent)
 			else if (isonum_711(vdp->type) == ISO_VD_SUPPLEMENTARY) {
 				sec = (struct iso_supplementary_descriptor *)vdp;
 				if (sec->escape[0] == 0x25 && sec->escape[1] == 0x2f) {
-					if (opt.joliet) {
+					if (opt->joliet) {
 						if (sec->escape[2] == 0x40)
 							joliet_level = 1;
 						else if (sec->escape[2] == 0x43)
@@ -721,7 +683,7 @@ static int isofs_fill_super(struct super_block *s, void *data, int silent)
 					goto out_freebh;
 
 				sbi->s_high_sierra = 1;
-				opt.rock = 0;
+				opt->rock = 0;
 				h_pri = (struct hs_primary_descriptor *)vdp;
 				goto root_found;
 			}
@@ -749,7 +711,7 @@ root_found:
 		goto out_freebh;
 	}
 
-	if (joliet_level && (!pri || !opt.rock)) {
+	if (joliet_level && (!pri || !opt->rock)) {
 		/* This is the case of Joliet with the norock mount flag.
 		 * A disc with both Joliet and Rock Ridge is handled later
 		 */
@@ -780,7 +742,7 @@ root_found:
 	 * blocks that were 512 bytes (which should only very rarely
 	 * happen.)
 	 */
-	if (orig_zonesize < opt.blocksize)
+	if (orig_zonesize < opt->blocksize)
 		goto out_bad_size;
 
 	/* RDE: convert log zone size to bit shift */
@@ -865,10 +827,10 @@ root_found:
 
 #ifdef CONFIG_JOLIET
 	if (joliet_level) {
-		char *p = opt.iocharset ? opt.iocharset : CONFIG_NLS_DEFAULT;
+		char *p = opt->iocharset ? opt->iocharset : CONFIG_NLS_DEFAULT;
 		if (strcmp(p, "utf8") != 0) {
-			sbi->s_nls_iocharset = opt.iocharset ?
-				load_nls(opt.iocharset) : load_nls_default();
+			sbi->s_nls_iocharset = opt->iocharset ?
+				load_nls(opt->iocharset) : load_nls_default();
 			if (!sbi->s_nls_iocharset)
 				goto out_freesbi;
 		}
@@ -876,29 +838,29 @@ root_found:
 #endif
 	s->s_op = &isofs_sops;
 	s->s_export_op = &isofs_export_ops;
-	sbi->s_mapping = opt.map;
-	sbi->s_rock = (opt.rock ? 2 : 0);
+	sbi->s_mapping = opt->map;
+	sbi->s_rock = (opt->rock ? 2 : 0);
 	sbi->s_rock_offset = -1; /* initial offset, will guess until SP is found*/
-	sbi->s_cruft = opt.cruft;
-	sbi->s_hide = opt.hide;
-	sbi->s_showassoc = opt.showassoc;
-	sbi->s_uid = opt.uid;
-	sbi->s_gid = opt.gid;
-	sbi->s_uid_set = opt.uid_set;
-	sbi->s_gid_set = opt.gid_set;
-	sbi->s_nocompress = opt.nocompress;
-	sbi->s_overriderockperm = opt.overriderockperm;
+	sbi->s_cruft = opt->cruft;
+	sbi->s_hide = opt->hide;
+	sbi->s_showassoc = opt->showassoc;
+	sbi->s_uid = opt->uid;
+	sbi->s_gid = opt->gid;
+	sbi->s_uid_set = opt->uid_set;
+	sbi->s_gid_set = opt->gid_set;
+	sbi->s_nocompress = opt->nocompress;
+	sbi->s_overriderockperm = opt->overriderockperm;
 	/*
 	 * It would be incredibly stupid to allow people to mark every file
 	 * on the disk as suid, so we merely allow them to set the default
 	 * permissions.
 	 */
-	if (opt.fmode != ISOFS_INVALID_MODE)
-		sbi->s_fmode = opt.fmode & 0777;
+	if (opt->fmode != ISOFS_INVALID_MODE)
+		sbi->s_fmode = opt->fmode & 0777;
 	else
 		sbi->s_fmode = ISOFS_INVALID_MODE;
-	if (opt.dmode != ISOFS_INVALID_MODE)
-		sbi->s_dmode = opt.dmode & 0777;
+	if (opt->dmode != ISOFS_INVALID_MODE)
+		sbi->s_dmode = opt->dmode & 0777;
 	else
 		sbi->s_dmode = ISOFS_INVALID_MODE;
 
@@ -960,12 +922,12 @@ root_found:
 		}
 	}
 
-	if (opt.check == 'u') {
+	if (opt->check == 'u') {
 		/* Only Joliet is case insensitive by default */
 		if (joliet_level)
-			opt.check = 'r';
+			opt->check = 'r';
 		else
-			opt.check = 's';
+			opt->check = 's';
 	}
 	sbi->s_joliet_level = joliet_level;
 
@@ -980,9 +942,9 @@ root_found:
 	table = 0;
 	if (joliet_level)
 		table += 2;
-	if (opt.check == 'r')
+	if (opt->check == 'r')
 		table++;
-	sbi->s_check = opt.check;
+	sbi->s_check = opt->check;
 
 	if (table)
 		s->s_d_op = &isofs_dentry_ops[table - 1];
@@ -994,7 +956,7 @@ root_found:
 		goto out_no_inode;
 	}
 
-	kfree(opt.iocharset);
+	kfree(opt->iocharset);
 
 	return 0;
 
@@ -1023,7 +985,7 @@ out_bad_zone_size:
 	goto out_freebh;
 out_bad_size:
 	printk(KERN_WARNING "ISOFS: Logical zone size(%d) < hardware blocksize(%u)\n",
-		orig_zonesize, opt.blocksize);
+		orig_zonesize, opt->blocksize);
 	goto out_freebh;
 out_unknown_format:
 	if (!silent)
@@ -1033,7 +995,7 @@ out_freebh:
 	brelse(bh);
 	brelse(pri_bh);
 out_freesbi:
-	kfree(opt.iocharset);
+	kfree(opt->iocharset);
 	kfree(sbi);
 	s->s_fs_info = NULL;
 	return error;
@@ -1567,18 +1529,63 @@ struct inode *__isofs_iget(struct super_block *sb,
 	return inode;
 }
 
-static struct dentry *isofs_mount(struct file_system_type *fs_type,
-	int flags, const char *dev_name, void *data)
+static int isofs_get_tree(struct fs_context *fc)
 {
-	return mount_bdev(fs_type, flags, dev_name, data, isofs_fill_super);
+	return get_tree_bdev(fc, isofs_fill_super);
+}
+
+static void isofs_free_fc(struct fs_context *fc)
+{
+	kfree(fc->fs_private);
+}
+
+static const struct fs_context_operations isofs_context_ops = {
+	.parse_param	= isofs_parse_param,
+	.get_tree	= isofs_get_tree,
+	.reconfigure	= isofs_reconfigure,
+	.free		= isofs_free_fc,
+};
+
+static int isofs_init_fs_context(struct fs_context *fc)
+{
+	struct isofs_options *opt;
+
+	opt = kzalloc(sizeof(*opt), GFP_KERNEL);
+	if (!opt)
+		return -ENOMEM;
+
+	opt->map = 'n';
+	opt->rock = 1;
+	opt->joliet = 1;
+	opt->cruft = 0;
+	opt->hide = 0;
+	opt->showassoc = 0;
+	opt->check = 'u';		/* unset */
+	opt->nocompress = 0;
+	opt->blocksize = 1024;
+	opt->fmode = opt->dmode = ISOFS_INVALID_MODE;
+	opt->uid_set = 0;
+	opt->gid_set = 0;
+	opt->gid = GLOBAL_ROOT_GID;
+	opt->uid = GLOBAL_ROOT_UID;
+	opt->iocharset = NULL;
+	opt->overriderockperm = 0;
+	opt->session = -1;
+	opt->sbsector = -1;
+
+	fc->fs_private = opt;
+	fc->ops = &isofs_context_ops;
+
+	return 0;
 }
 
 static struct file_system_type iso9660_fs_type = {
 	.owner		= THIS_MODULE,
 	.name		= "iso9660",
-	.mount		= isofs_mount,
 	.kill_sb	= kill_block_super,
 	.fs_flags	= FS_REQUIRES_DEV,
+	.init_fs_context = isofs_init_fs_context,
+	.parameters	= isofs_param_spec,
 };
 MODULE_ALIAS_FS("iso9660");
 MODULE_ALIAS("iso9660");
diff --git a/fs/jbd2/checkpoint.c b/fs/jbd2/checkpoint.c
index 1c97e64c47..951f78634a 100644
--- a/fs/jbd2/checkpoint.c
+++ b/fs/jbd2/checkpoint.c
@@ -337,8 +337,6 @@ int jbd2_cleanup_journal_tail(journal_t *journal)
 
 /* Checkpoint list management */
 
-enum shrink_type {SHRINK_DESTROY, SHRINK_BUSY_STOP, SHRINK_BUSY_SKIP};
-
 /*
  * journal_shrink_one_cp_list
  *
@@ -350,7 +348,7 @@ enum shrink_type {SHRINK_DESTROY, SHRINK_BUSY_STOP, SHRINK_BUSY_SKIP};
  * Called with j_list_lock held.
  */
 static unsigned long journal_shrink_one_cp_list(struct journal_head *jh,
-						enum shrink_type type,
+						enum jbd2_shrink_type type,
 						bool *released)
 {
 	struct journal_head *last_jh;
@@ -367,12 +365,12 @@ static unsigned long journal_shrink_one_cp_list(struct journal_head *jh,
 		jh = next_jh;
 		next_jh = jh->b_cpnext;
 
-		if (type == SHRINK_DESTROY) {
+		if (type == JBD2_SHRINK_DESTROY) {
 			ret = __jbd2_journal_remove_checkpoint(jh);
 		} else {
 			ret = jbd2_journal_try_remove_checkpoint(jh);
 			if (ret < 0) {
-				if (type == SHRINK_BUSY_SKIP)
+				if (type == JBD2_SHRINK_BUSY_SKIP)
 					continue;
 				break;
 			}
@@ -439,7 +437,7 @@ again:
 		tid = transaction->t_tid;
 
 		freed = journal_shrink_one_cp_list(transaction->t_checkpoint_list,
-						   SHRINK_BUSY_SKIP, &released);
+						   JBD2_SHRINK_BUSY_SKIP, &released);
 		nr_freed += freed;
 		(*nr_to_scan) -= min(*nr_to_scan, freed);
 		if (*nr_to_scan == 0)
@@ -472,21 +470,25 @@ out:
  * journal_clean_checkpoint_list
  *
  * Find all the written-back checkpoint buffers in the journal and release them.
- * If 'destroy' is set, release all buffers unconditionally.
+ * If 'type' is JBD2_SHRINK_DESTROY, release all buffers unconditionally. If
+ * 'type' is JBD2_SHRINK_BUSY_STOP, will stop release buffers if encounters a
+ * busy buffer. To avoid wasting CPU cycles scanning the buffer list in some
+ * cases, don't pass JBD2_SHRINK_BUSY_SKIP 'type' for this function.
  *
  * Called with j_list_lock held.
  */
-void __jbd2_journal_clean_checkpoint_list(journal_t *journal, bool destroy)
+void __jbd2_journal_clean_checkpoint_list(journal_t *journal,
+					  enum jbd2_shrink_type type)
 {
 	transaction_t *transaction, *last_transaction, *next_transaction;
-	enum shrink_type type;
 	bool released;
 
+	WARN_ON_ONCE(type == JBD2_SHRINK_BUSY_SKIP);
+
 	transaction = journal->j_checkpoint_transactions;
 	if (!transaction)
 		return;
 
-	type = destroy ? SHRINK_DESTROY : SHRINK_BUSY_STOP;
 	last_transaction = transaction->t_cpprev;
 	next_transaction = transaction;
 	do {
@@ -527,7 +529,7 @@ void jbd2_journal_destroy_checkpoint(journal_t *journal)
 			spin_unlock(&journal->j_list_lock);
 			break;
 		}
-		__jbd2_journal_clean_checkpoint_list(journal, true);
+		__jbd2_journal_clean_checkpoint_list(journal, JBD2_SHRINK_DESTROY);
 		spin_unlock(&journal->j_list_lock);
 		cond_resched();
 	}
diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c
index 5e122586e0..e7fc912693 100644
--- a/fs/jbd2/commit.c
+++ b/fs/jbd2/commit.c
@@ -501,7 +501,7 @@ void jbd2_journal_commit_transaction(journal_t *journal)
 	 * frees some memory
 	 */
 	spin_lock(&journal->j_list_lock);
-	__jbd2_journal_clean_checkpoint_list(journal, false);
+	__jbd2_journal_clean_checkpoint_list(journal, JBD2_SHRINK_BUSY_STOP);
 	spin_unlock(&journal->j_list_lock);
 
 	jbd2_debug(3, "JBD2: commit phase 1\n");
@@ -571,7 +571,6 @@ void jbd2_journal_commit_transaction(journal_t *journal)
 	J_ASSERT(commit_transaction->t_nr_buffers <=
 		 atomic_read(&commit_transaction->t_outstanding_credits));
 
-	err = 0;
 	bufs = 0;
 	descriptor = NULL;
 	while (commit_transaction->t_buffers) {
@@ -767,7 +766,7 @@ start_journal_io:
 		if (first_block < journal->j_tail)
 			freed += journal->j_last - journal->j_first;
 		/* Update tail only if we free significant amount of space */
-		if (freed < jbd2_journal_get_max_txn_bufs(journal))
+		if (freed < journal->j_max_transaction_buffers)
 			update_tail = 0;
 	}
 	J_ASSERT(commit_transaction->t_state == T_COMMIT);
diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index b6c114c11b..c8d9d85e0e 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -399,6 +399,7 @@ repeat:
 		tmp = jbd2_alloc(bh_in->b_size, GFP_NOFS);
 		if (!tmp) {
 			brelse(new_bh);
+			free_buffer_head(new_bh);
 			return -ENOMEM;
 		}
 		spin_lock(&jh_in->b_state_lock);
@@ -1451,6 +1452,48 @@ static int journal_revoke_records_per_block(journal_t *journal)
 	return space / record_size;
 }
 
+static int jbd2_journal_get_max_txn_bufs(journal_t *journal)
+{
+	return (journal->j_total_len - journal->j_fc_wbufsize) / 4;
+}
+
+/*
+ * Base amount of descriptor blocks we reserve for each transaction.
+ */
+static int jbd2_descriptor_blocks_per_trans(journal_t *journal)
+{
+	int tag_space = journal->j_blocksize - sizeof(journal_header_t);
+	int tags_per_block;
+
+	/* Subtract UUID */
+	tag_space -= 16;
+	if (jbd2_journal_has_csum_v2or3(journal))
+		tag_space -= sizeof(struct jbd2_journal_block_tail);
+	/* Commit code leaves a slack space of 16 bytes at the end of block */
+	tags_per_block = (tag_space - 16) / journal_tag_bytes(journal);
+	/*
+	 * Revoke descriptors are accounted separately so we need to reserve
+	 * space for commit block and normal transaction descriptor blocks.
+	 */
+	return 1 + DIV_ROUND_UP(jbd2_journal_get_max_txn_bufs(journal),
+				tags_per_block);
+}
+
+/*
+ * Initialize number of blocks each transaction reserves for its bookkeeping
+ * and maximum number of blocks a transaction can use. This needs to be called
+ * after the journal size and the fastcommit area size are initialized.
+ */
+static void jbd2_journal_init_transaction_limits(journal_t *journal)
+{
+	journal->j_revoke_records_per_block =
+				journal_revoke_records_per_block(journal);
+	journal->j_transaction_overhead_buffers =
+				jbd2_descriptor_blocks_per_trans(journal);
+	journal->j_max_transaction_buffers =
+				jbd2_journal_get_max_txn_bufs(journal);
+}
+
 /*
  * Load the on-disk journal superblock and read the key fields into the
  * journal_t.
@@ -1492,8 +1535,8 @@ static int journal_load_superblock(journal_t *journal)
 	if (jbd2_journal_has_csum_v2or3(journal))
 		journal->j_csum_seed = jbd2_chksum(journal, ~0, sb->s_uuid,
 						   sizeof(sb->s_uuid));
-	journal->j_revoke_records_per_block =
-				journal_revoke_records_per_block(journal);
+	/* After journal features are set, we can compute transaction limits */
+	jbd2_journal_init_transaction_limits(journal);
 
 	if (jbd2_has_feature_fast_commit(journal)) {
 		journal->j_fc_last = be32_to_cpu(sb->s_maxlen);
@@ -1743,8 +1786,6 @@ static int journal_reset(journal_t *journal)
 	journal->j_commit_sequence = journal->j_transaction_sequence - 1;
 	journal->j_commit_request = journal->j_commit_sequence;
 
-	journal->j_max_transaction_buffers = jbd2_journal_get_max_txn_bufs(journal);
-
 	/*
 	 * Now that journal recovery is done, turn fast commits off here. This
 	 * way, if fast commit was enabled before the crash but if now FS has
@@ -2009,7 +2050,7 @@ static int __jbd2_journal_erase(journal_t *journal, unsigned int flags)
 		byte_count = (block_stop - block_start + 1) *
 				journal->j_blocksize;
 
-		truncate_inode_pages_range(journal->j_dev->bd_inode->i_mapping,
+		truncate_inode_pages_range(journal->j_dev->bd_mapping,
 				byte_start, byte_stop);
 
 		if (flags & JBD2_JOURNAL_FLUSH_DISCARD) {
@@ -2285,8 +2326,6 @@ jbd2_journal_initialize_fast_commit(journal_t *journal)
 	journal->j_fc_first = journal->j_last + 1;
 	journal->j_fc_off = 0;
 	journal->j_free = journal->j_last - journal->j_first;
-	journal->j_max_transaction_buffers =
-		jbd2_journal_get_max_txn_bufs(journal);
 
 	return 0;
 }
@@ -2374,8 +2413,7 @@ int jbd2_journal_set_features(journal_t *journal, unsigned long compat,
 	sb->s_feature_ro_compat |= cpu_to_be32(ro);
 	sb->s_feature_incompat  |= cpu_to_be32(incompat);
 	unlock_buffer(journal->j_sb_buffer);
-	journal->j_revoke_records_per_block =
-				journal_revoke_records_per_block(journal);
+	jbd2_journal_init_transaction_limits(journal);
 
 	return 1;
 #undef COMPAT_FEATURE_ON
@@ -2406,8 +2444,7 @@ void jbd2_journal_clear_features(journal_t *journal, unsigned long compat,
 	sb->s_feature_compat    &= ~cpu_to_be32(compat);
 	sb->s_feature_ro_compat &= ~cpu_to_be32(ro);
 	sb->s_feature_incompat  &= ~cpu_to_be32(incompat);
-	journal->j_revoke_records_per_block =
-				journal_revoke_records_per_block(journal);
+	jbd2_journal_init_transaction_limits(journal);
 }
 EXPORT_SYMBOL(jbd2_journal_clear_features);
 
diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c
index cb0b8d6fc0..66513c18ca 100644
--- a/fs/jbd2/transaction.c
+++ b/fs/jbd2/transaction.c
@@ -63,28 +63,6 @@ void jbd2_journal_free_transaction(transaction_t *transaction)
 }
 
 /*
- * Base amount of descriptor blocks we reserve for each transaction.
- */
-static int jbd2_descriptor_blocks_per_trans(journal_t *journal)
-{
-	int tag_space = journal->j_blocksize - sizeof(journal_header_t);
-	int tags_per_block;
-
-	/* Subtract UUID */
-	tag_space -= 16;
-	if (jbd2_journal_has_csum_v2or3(journal))
-		tag_space -= sizeof(struct jbd2_journal_block_tail);
-	/* Commit code leaves a slack space of 16 bytes at the end of block */
-	tags_per_block = (tag_space - 16) / journal_tag_bytes(journal);
-	/*
-	 * Revoke descriptors are accounted separately so we need to reserve
-	 * space for commit block and normal transaction descriptor blocks.
-	 */
-	return 1 + DIV_ROUND_UP(journal->j_max_transaction_buffers,
-				tags_per_block);
-}
-
-/*
  * jbd2_get_transaction: obtain a new transaction_t object.
  *
  * Simply initialise a new transaction. Initialize it in
@@ -109,7 +87,7 @@ static void jbd2_get_transaction(journal_t *journal,
 	transaction->t_expires = jiffies + journal->j_commit_interval;
 	atomic_set(&transaction->t_updates, 0);
 	atomic_set(&transaction->t_outstanding_credits,
-		   jbd2_descriptor_blocks_per_trans(journal) +
+		   journal->j_transaction_overhead_buffers +
 		   atomic_read(&journal->j_reserved_credits));
 	atomic_set(&transaction->t_outstanding_revokes, 0);
 	atomic_set(&transaction->t_handle_count, 0);
@@ -213,6 +191,13 @@ static void sub_reserved_credits(journal_t *journal, int blocks)
 	wake_up(&journal->j_wait_reserved);
 }
 
+/* Maximum number of blocks for user transaction payload */
+static int jbd2_max_user_trans_buffers(journal_t *journal)
+{
+	return journal->j_max_transaction_buffers -
+				journal->j_transaction_overhead_buffers;
+}
+
 /*
  * Wait until we can add credits for handle to the running transaction.  Called
  * with j_state_lock held for reading. Returns 0 if handle joined the running
@@ -262,12 +247,12 @@ __must_hold(&journal->j_state_lock)
 		 * big to fit this handle? Wait until reserved credits are freed.
 		 */
 		if (atomic_read(&journal->j_reserved_credits) + total >
-		    journal->j_max_transaction_buffers) {
+		    jbd2_max_user_trans_buffers(journal)) {
 			read_unlock(&journal->j_state_lock);
 			jbd2_might_wait_for_commit(journal);
 			wait_event(journal->j_wait_reserved,
 				   atomic_read(&journal->j_reserved_credits) + total <=
-				   journal->j_max_transaction_buffers);
+				   jbd2_max_user_trans_buffers(journal));
 			__acquire(&journal->j_state_lock); /* fake out sparse */
 			return 1;
 		}
@@ -307,14 +292,14 @@ __must_hold(&journal->j_state_lock)
 
 	needed = atomic_add_return(rsv_blocks, &journal->j_reserved_credits);
 	/* We allow at most half of a transaction to be reserved */
-	if (needed > journal->j_max_transaction_buffers / 2) {
+	if (needed > jbd2_max_user_trans_buffers(journal) / 2) {
 		sub_reserved_credits(journal, rsv_blocks);
 		atomic_sub(total, &t->t_outstanding_credits);
 		read_unlock(&journal->j_state_lock);
 		jbd2_might_wait_for_commit(journal);
 		wait_event(journal->j_wait_reserved,
 			 atomic_read(&journal->j_reserved_credits) + rsv_blocks
-			 <= journal->j_max_transaction_buffers / 2);
+			 <= jbd2_max_user_trans_buffers(journal) / 2);
 		__acquire(&journal->j_state_lock); /* fake out sparse */
 		return 1;
 	}
@@ -344,12 +329,12 @@ static int start_this_handle(journal_t *journal, handle_t *handle,
 	 * size and limit the number of total credits to not exceed maximum
 	 * transaction size per operation.
 	 */
-	if ((rsv_blocks > journal->j_max_transaction_buffers / 2) ||
-	    (rsv_blocks + blocks > journal->j_max_transaction_buffers)) {
+	if (rsv_blocks > jbd2_max_user_trans_buffers(journal) / 2 ||
+	    rsv_blocks + blocks > jbd2_max_user_trans_buffers(journal)) {
 		printk(KERN_ERR "JBD2: %s wants too many credits "
 		       "credits:%d rsv_credits:%d max:%d\n",
 		       current->comm, blocks, rsv_blocks,
-		       journal->j_max_transaction_buffers);
+		       jbd2_max_user_trans_buffers(journal));
 		WARN_ON(1);
 		return -ENOSPC;
 	}
diff --git a/fs/jffs2/background.c b/fs/jffs2/background.c
index 6da92ecaf6..bb0ee1a59e 100644
--- a/fs/jffs2/background.c
+++ b/fs/jffs2/background.c
@@ -44,8 +44,8 @@ int jffs2_start_garbage_collect_thread(struct jffs2_sb_info *c)
 
 	tsk = kthread_run(jffs2_garbage_collect_thread, c, "jffs2_gcd_mtd%d", c->mtd->index);
 	if (IS_ERR(tsk)) {
-		pr_warn("fork failed for JFFS2 garbage collect thread: %ld\n",
-			-PTR_ERR(tsk));
+		pr_warn("fork failed for JFFS2 garbage collect thread: %pe\n",
+			tsk);
 		complete(&c->gc_thread_exit);
 		ret = PTR_ERR(tsk);
 	} else {
diff --git a/fs/jffs2/malloc.c b/fs/jffs2/malloc.c
index ce11897932..411de8b361 100644
--- a/fs/jffs2/malloc.c
+++ b/fs/jffs2/malloc.c
@@ -33,27 +33,19 @@ static struct kmem_cache *xattr_ref_cache;
 
 int __init jffs2_create_slab_caches(void)
 {
-	full_dnode_slab = kmem_cache_create("jffs2_full_dnode",
-					    sizeof(struct jffs2_full_dnode),
-					    0, 0, NULL);
+	full_dnode_slab = KMEM_CACHE(jffs2_full_dnode, 0);
 	if (!full_dnode_slab)
 		goto err;
 
-	raw_dirent_slab = kmem_cache_create("jffs2_raw_dirent",
-					    sizeof(struct jffs2_raw_dirent),
-					    0, SLAB_HWCACHE_ALIGN, NULL);
+	raw_dirent_slab = KMEM_CACHE(jffs2_raw_dirent, SLAB_HWCACHE_ALIGN);
 	if (!raw_dirent_slab)
 		goto err;
 
-	raw_inode_slab = kmem_cache_create("jffs2_raw_inode",
-					   sizeof(struct jffs2_raw_inode),
-					   0, SLAB_HWCACHE_ALIGN, NULL);
+	raw_inode_slab = KMEM_CACHE(jffs2_raw_inode, SLAB_HWCACHE_ALIGN);
 	if (!raw_inode_slab)
 		goto err;
 
-	tmp_dnode_info_slab = kmem_cache_create("jffs2_tmp_dnode",
-						sizeof(struct jffs2_tmp_dnode_info),
-						0, 0, NULL);
+	tmp_dnode_info_slab = KMEM_CACHE(jffs2_tmp_dnode_info, 0);
 	if (!tmp_dnode_info_slab)
 		goto err;
 
@@ -63,28 +55,20 @@ int __init jffs2_create_slab_caches(void)
 	if (!raw_node_ref_slab)
 		goto err;
 
-	node_frag_slab = kmem_cache_create("jffs2_node_frag",
-					   sizeof(struct jffs2_node_frag),
-					   0, 0, NULL);
+	node_frag_slab = KMEM_CACHE(jffs2_node_frag, 0);
 	if (!node_frag_slab)
 		goto err;
 
-	inode_cache_slab = kmem_cache_create("jffs2_inode_cache",
-					     sizeof(struct jffs2_inode_cache),
-					     0, 0, NULL);
+	inode_cache_slab = KMEM_CACHE(jffs2_inode_cache, 0);
 	if (!inode_cache_slab)
 		goto err;
 
 #ifdef CONFIG_JFFS2_FS_XATTR
-	xattr_datum_cache = kmem_cache_create("jffs2_xattr_datum",
-					     sizeof(struct jffs2_xattr_datum),
-					     0, 0, NULL);
+	xattr_datum_cache = KMEM_CACHE(jffs2_xattr_datum, 0);
 	if (!xattr_datum_cache)
 		goto err;
 
-	xattr_ref_cache = kmem_cache_create("jffs2_xattr_ref",
-					   sizeof(struct jffs2_xattr_ref),
-					   0, 0, NULL);
+	xattr_ref_cache = KMEM_CACHE(jffs2_xattr_ref, 0);
 	if (!xattr_ref_cache)
 		goto err;
 #endif
diff --git a/fs/jffs2/nodemgmt.c b/fs/jffs2/nodemgmt.c
index a7bbe879cf..bbab2bdc71 100644
--- a/fs/jffs2/nodemgmt.c
+++ b/fs/jffs2/nodemgmt.c
@@ -49,28 +49,31 @@ static int jffs2_rp_can_write(struct jffs2_sb_info *c)
 	return 0;
 }
 
+static int jffs2_do_reserve_space(struct jffs2_sb_info *c,  uint32_t minsize,
+				  uint32_t *len, uint32_t sumsize);
+
 /**
  *	jffs2_reserve_space - request physical space to write nodes to flash
  *	@c: superblock info
  *	@minsize: Minimum acceptable size of allocation
  *	@len: Returned value of allocation length
  *	@prio: Allocation type - ALLOC_{NORMAL,DELETION}
+ *	@sumsize: summary size requested or JFFS2_SUMMARY_NOSUM_SIZE for no summary
+ *
+ *	Requests a block of physical space on the flash.
  *
- *	Requests a block of physical space on the flash. Returns zero for success
- *	and puts 'len' into the appropriate place, or returns -ENOSPC or other 
- *	error if appropriate. Doesn't return len since that's 
+ *	Returns: %0 for success	and puts 'len' into the appropriate place,
+ *	or returns -ENOSPC or other error if appropriate.
+ *	Doesn't return len since that's already returned in @len.
  *
- *	If it returns zero, jffs2_reserve_space() also downs the per-filesystem
+ *	If it returns %0, jffs2_reserve_space() also downs the per-filesystem
  *	allocation semaphore, to prevent more than one allocation from being
- *	active at any time. The semaphore is later released by jffs2_commit_allocation()
+ *	active at any time. The semaphore is later released by jffs2_commit_allocation().
  *
  *	jffs2_reserve_space() may trigger garbage collection in order to make room
  *	for the requested allocation.
  */
 
-static int jffs2_do_reserve_space(struct jffs2_sb_info *c,  uint32_t minsize,
-				  uint32_t *len, uint32_t sumsize);
-
 int jffs2_reserve_space(struct jffs2_sb_info *c, uint32_t minsize,
 			uint32_t *len, int prio, uint32_t sumsize)
 {
@@ -488,13 +491,16 @@ static int jffs2_do_reserve_space(struct jffs2_sb_info *c, uint32_t minsize,
 /**
  *	jffs2_add_physical_node_ref - add a physical node reference to the list
  *	@c: superblock info
- *	@new: new node reference to add
+ *	@ofs: offset in the block
  *	@len: length of this physical node
+ *	@ic: inode cache pointer
  *
  *	Should only be used to report nodes for which space has been allocated
  *	by jffs2_reserve_space.
  *
  *	Must be called with the alloc_sem held.
+ *
+ *	Returns: pointer to new node on success or -errno code on error
  */
 
 struct jffs2_raw_node_ref *jffs2_add_physical_node_ref(struct jffs2_sb_info *c,
diff --git a/fs/jffs2/super.c b/fs/jffs2/super.c
index f99591a634..4545f885c4 100644
--- a/fs/jffs2/super.c
+++ b/fs/jffs2/super.c
@@ -58,6 +58,7 @@ static void jffs2_i_init_once(void *foo)
 	struct jffs2_inode_info *f = foo;
 
 	mutex_init(&f->sem);
+	f->target = NULL;
 	inode_init_once(&f->vfs_inode);
 }
 
@@ -387,7 +388,7 @@ static int __init init_jffs2_fs(void)
 	jffs2_inode_cachep = kmem_cache_create("jffs2_i",
 					     sizeof(struct jffs2_inode_info),
 					     0, (SLAB_RECLAIM_ACCOUNT|
-						SLAB_MEM_SPREAD|SLAB_ACCOUNT),
+						SLAB_ACCOUNT),
 					     jffs2_i_init_once);
 	if (!jffs2_inode_cachep) {
 		pr_err("error: Failed to initialise inode cache\n");
diff --git a/fs/jfs/jfs_dmap.c b/fs/jfs/jfs_dmap.c
index cb3cda1390..5713994328 100644
--- a/fs/jfs/jfs_dmap.c
+++ b/fs/jfs/jfs_dmap.c
@@ -1626,6 +1626,8 @@ s64 dbDiscardAG(struct inode *ip, int agno, s64 minlen)
 		} else if (rc == -ENOSPC) {
 			/* search for next smaller log2 block */
 			l2nb = BLKSTOL2(nblocks) - 1;
+			if (unlikely(l2nb < 0))
+				break;
 			nblocks = 1LL << l2nb;
 		} else {
 			/* Trim any already allocated blocks */
diff --git a/fs/jfs/jfs_dtree.c b/fs/jfs/jfs_dtree.c
index 031d8f570f..5d3127ca68 100644
--- a/fs/jfs/jfs_dtree.c
+++ b/fs/jfs/jfs_dtree.c
@@ -834,6 +834,8 @@ int dtInsert(tid_t tid, struct inode *ip,
 	 * the full page.
 	 */
 	DT_GETSEARCH(ip, btstack->top, bn, mp, p, index);
+	if (p->header.freelist == 0)
+		return -EINVAL;
 
 	/*
 	 *	insert entry for new key
diff --git a/fs/jfs/jfs_imap.c b/fs/jfs/jfs_imap.c
index 2ec35889ad..1407feccbc 100644
--- a/fs/jfs/jfs_imap.c
+++ b/fs/jfs/jfs_imap.c
@@ -290,7 +290,7 @@ int diSync(struct inode *ipimap)
 int diRead(struct inode *ip)
 {
 	struct jfs_sb_info *sbi = JFS_SBI(ip->i_sb);
-	int iagno, ino, extno, rc;
+	int iagno, ino, extno, rc, agno;
 	struct inode *ipimap;
 	struct dinode *dp;
 	struct iag *iagp;
@@ -339,8 +339,11 @@ int diRead(struct inode *ip)
 
 	/* get the ag for the iag */
 	agstart = le64_to_cpu(iagp->agstart);
+	agno = BLKTOAG(agstart, JFS_SBI(ip->i_sb));
 
 	release_metapage(mp);
+	if (agno >= MAXAG || agno < 0)
+		return -EIO;
 
 	rel_inode = (ino & (INOSPERPAGE - 1));
 	pageno = blkno >> sbi->l2nbperpage;
diff --git a/fs/jfs/jfs_logmgr.c b/fs/jfs/jfs_logmgr.c
index cb6d1fda66..9609349e92 100644
--- a/fs/jfs/jfs_logmgr.c
+++ b/fs/jfs/jfs_logmgr.c
@@ -1058,7 +1058,7 @@ void jfs_syncpt(struct jfs_log *log, int hard_sync)
 int lmLogOpen(struct super_block *sb)
 {
 	int rc;
-	struct bdev_handle *bdev_handle;
+	struct file *bdev_file;
 	struct jfs_log *log;
 	struct jfs_sb_info *sbi = JFS_SBI(sb);
 
@@ -1070,7 +1070,7 @@ int lmLogOpen(struct super_block *sb)
 
 	mutex_lock(&jfs_log_mutex);
 	list_for_each_entry(log, &jfs_external_logs, journal_list) {
-		if (log->bdev_handle->bdev->bd_dev == sbi->logdev) {
+		if (file_bdev(log->bdev_file)->bd_dev == sbi->logdev) {
 			if (!uuid_equal(&log->uuid, &sbi->loguuid)) {
 				jfs_warn("wrong uuid on JFS journal");
 				mutex_unlock(&jfs_log_mutex);
@@ -1100,14 +1100,14 @@ int lmLogOpen(struct super_block *sb)
 	 * file systems to log may have n-to-1 relationship;
 	 */
 
-	bdev_handle = bdev_open_by_dev(sbi->logdev,
+	bdev_file = bdev_file_open_by_dev(sbi->logdev,
 			BLK_OPEN_READ | BLK_OPEN_WRITE, log, NULL);
-	if (IS_ERR(bdev_handle)) {
-		rc = PTR_ERR(bdev_handle);
+	if (IS_ERR(bdev_file)) {
+		rc = PTR_ERR(bdev_file);
 		goto free;
 	}
 
-	log->bdev_handle = bdev_handle;
+	log->bdev_file = bdev_file;
 	uuid_copy(&log->uuid, &sbi->loguuid);
 
 	/*
@@ -1141,7 +1141,7 @@ journal_found:
 	lbmLogShutdown(log);
 
       close:		/* close external log device */
-	bdev_release(bdev_handle);
+	bdev_fput(bdev_file);
 
       free:		/* free log descriptor */
 	mutex_unlock(&jfs_log_mutex);
@@ -1162,7 +1162,7 @@ static int open_inline_log(struct super_block *sb)
 	init_waitqueue_head(&log->syncwait);
 
 	set_bit(log_INLINELOG, &log->flag);
-	log->bdev_handle = sb->s_bdev_handle;
+	log->bdev_file = sb->s_bdev_file;
 	log->base = addressPXD(&JFS_SBI(sb)->logpxd);
 	log->size = lengthPXD(&JFS_SBI(sb)->logpxd) >>
 	    (L2LOGPSIZE - sb->s_blocksize_bits);
@@ -1436,7 +1436,7 @@ int lmLogClose(struct super_block *sb)
 {
 	struct jfs_sb_info *sbi = JFS_SBI(sb);
 	struct jfs_log *log = sbi->log;
-	struct bdev_handle *bdev_handle;
+	struct file *bdev_file;
 	int rc = 0;
 
 	jfs_info("lmLogClose: log:0x%p", log);
@@ -1482,10 +1482,10 @@ int lmLogClose(struct super_block *sb)
 	 *	external log as separate logical volume
 	 */
 	list_del(&log->journal_list);
-	bdev_handle = log->bdev_handle;
+	bdev_file = log->bdev_file;
 	rc = lmLogShutdown(log);
 
-	bdev_release(bdev_handle);
+	bdev_fput(bdev_file);
 
 	kfree(log);
 
@@ -1972,7 +1972,7 @@ static int lbmRead(struct jfs_log * log, int pn, struct lbuf ** bpp)
 
 	bp->l_flag |= lbmREAD;
 
-	bio = bio_alloc(log->bdev_handle->bdev, 1, REQ_OP_READ, GFP_NOFS);
+	bio = bio_alloc(file_bdev(log->bdev_file), 1, REQ_OP_READ, GFP_NOFS);
 	bio->bi_iter.bi_sector = bp->l_blkno << (log->l2bsize - 9);
 	__bio_add_page(bio, bp->l_page, LOGPSIZE, bp->l_offset);
 	BUG_ON(bio->bi_iter.bi_size != LOGPSIZE);
@@ -2115,7 +2115,7 @@ static void lbmStartIO(struct lbuf * bp)
 	jfs_info("lbmStartIO");
 
 	if (!log->no_integrity)
-		bdev = log->bdev_handle->bdev;
+		bdev = file_bdev(log->bdev_file);
 
 	bio = bio_alloc(bdev, 1, REQ_OP_WRITE | REQ_SYNC,
 			GFP_NOFS);
diff --git a/fs/jfs/jfs_logmgr.h b/fs/jfs/jfs_logmgr.h
index 84aa2d2539..8b8994e48c 100644
--- a/fs/jfs/jfs_logmgr.h
+++ b/fs/jfs/jfs_logmgr.h
@@ -356,7 +356,7 @@ struct jfs_log {
 				 *    before writing syncpt.
 				 */
 	struct list_head journal_list; /* Global list */
-	struct bdev_handle *bdev_handle; /* 4: log lv pointer */
+	struct file *bdev_file;	/* 4: log lv pointer */
 	int serial;		/* 4: log mount serial number */
 
 	s64 base;		/* @8: log extent address (inline log ) */
diff --git a/fs/jfs/jfs_mount.c b/fs/jfs/jfs_mount.c
index 9b5c6a20b3..98f9a432c3 100644
--- a/fs/jfs/jfs_mount.c
+++ b/fs/jfs/jfs_mount.c
@@ -431,7 +431,7 @@ int updateSuper(struct super_block *sb, uint state)
 	if (state == FM_MOUNT) {
 		/* record log's dev_t and mount serial number */
 		j_sb->s_logdev = cpu_to_le32(
-			new_encode_dev(sbi->log->bdev_handle->bdev->bd_dev));
+			new_encode_dev(file_bdev(sbi->log->bdev_file)->bd_dev));
 		j_sb->s_logserial = cpu_to_le32(sbi->log->serial);
 	} else if (state == FM_CLEAN) {
 		/*
diff --git a/fs/jfs/super.c b/fs/jfs/super.c
index ff135a43b5..e1be21ca5d 100644
--- a/fs/jfs/super.c
+++ b/fs/jfs/super.c
@@ -932,7 +932,7 @@ static int __init init_jfs_fs(void)
 
 	jfs_inode_cachep =
 	    kmem_cache_create_usercopy("jfs_ip", sizeof(struct jfs_inode_info),
-			0, SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD|SLAB_ACCOUNT,
+			0, SLAB_RECLAIM_ACCOUNT|SLAB_ACCOUNT,
 			offsetof(struct jfs_inode_info, i_inline_all),
 			sizeof_field(struct jfs_inode_info, i_inline_all),
 			init_once);
diff --git a/fs/jfs/xattr.c b/fs/jfs/xattr.c
index 0fb7afac29..2999ed5d83 100644
--- a/fs/jfs/xattr.c
+++ b/fs/jfs/xattr.c
@@ -557,9 +557,11 @@ static int ea_get(struct inode *inode, struct ea_buffer *ea_buf, int min_size)
 
       size_check:
 	if (EALIST_SIZE(ea_buf->xattr) != ea_size) {
+		int size = min_t(int, EALIST_SIZE(ea_buf->xattr), ea_size);
+
 		printk(KERN_ERR "ea_get: invalid extended attribute\n");
 		print_hex_dump(KERN_ERR, "", DUMP_PREFIX_ADDRESS, 16, 1,
-				     ea_buf->xattr, ea_size, 1);
+				     ea_buf->xattr, size, 1);
 		ea_release(inode, ea_buf);
 		rc = -EIO;
 		goto clean_up;
@@ -795,7 +797,7 @@ ssize_t __jfs_getxattr(struct inode *inode, const char *name, void *data,
 		       size_t buf_size)
 {
 	struct jfs_ea_list *ealist;
-	struct jfs_ea *ea;
+	struct jfs_ea *ea, *ealist_end;
 	struct ea_buffer ea_buf;
 	int xattr_size;
 	ssize_t size;
@@ -815,9 +817,16 @@ ssize_t __jfs_getxattr(struct inode *inode, const char *name, void *data,
 		goto not_found;
 
 	ealist = (struct jfs_ea_list *) ea_buf.xattr;
+	ealist_end = END_EALIST(ealist);
 
 	/* Find the named attribute */
-	for (ea = FIRST_EA(ealist); ea < END_EALIST(ealist); ea = NEXT_EA(ea))
+	for (ea = FIRST_EA(ealist); ea < ealist_end; ea = NEXT_EA(ea)) {
+		if (unlikely(ea + 1 > ealist_end) ||
+		    unlikely(NEXT_EA(ea) > ealist_end)) {
+			size = -EUCLEAN;
+			goto release;
+		}
+
 		if ((namelen == ea->namelen) &&
 		    memcmp(name, ea->name, namelen) == 0) {
 			/* Found it */
@@ -832,6 +841,7 @@ ssize_t __jfs_getxattr(struct inode *inode, const char *name, void *data,
 			memcpy(data, value, size);
 			goto release;
 		}
+	}
       not_found:
 	size = -ENODATA;
       release:
@@ -859,7 +869,7 @@ ssize_t jfs_listxattr(struct dentry * dentry, char *data, size_t buf_size)
 	ssize_t size = 0;
 	int xattr_size;
 	struct jfs_ea_list *ealist;
-	struct jfs_ea *ea;
+	struct jfs_ea *ea, *ealist_end;
 	struct ea_buffer ea_buf;
 
 	down_read(&JFS_IP(inode)->xattr_sem);
@@ -874,9 +884,16 @@ ssize_t jfs_listxattr(struct dentry * dentry, char *data, size_t buf_size)
 		goto release;
 
 	ealist = (struct jfs_ea_list *) ea_buf.xattr;
+	ealist_end = END_EALIST(ealist);
 
 	/* compute required size of list */
-	for (ea = FIRST_EA(ealist); ea < END_EALIST(ealist); ea = NEXT_EA(ea)) {
+	for (ea = FIRST_EA(ealist); ea < ealist_end; ea = NEXT_EA(ea)) {
+		if (unlikely(ea + 1 > ealist_end) ||
+		    unlikely(NEXT_EA(ea) > ealist_end)) {
+			size = -EUCLEAN;
+			goto release;
+		}
+
 		if (can_list(ea))
 			size += name_size(ea) + 1;
 	}
diff --git a/fs/kernfs/file.c b/fs/kernfs/file.c
index b7f943f333..8502ef6845 100644
--- a/fs/kernfs/file.c
+++ b/fs/kernfs/file.c
@@ -483,9 +483,11 @@ static int kernfs_fop_mmap(struct file *file, struct vm_area_struct *vma)
 		goto out_put;
 
 	rc = 0;
-	of->mmapped = true;
-	of_on(of)->nr_mmapped++;
-	of->vm_ops = vma->vm_ops;
+	if (!of->mmapped) {
+		of->mmapped = true;
+		of_on(of)->nr_mmapped++;
+		of->vm_ops = vma->vm_ops;
+	}
 	vma->vm_ops = &kernfs_vm_ops;
 out_put:
 	kernfs_put_active(of->kn);
diff --git a/fs/kernfs/mount.c b/fs/kernfs/mount.c
index 0c93cad0f0..1358c21837 100644
--- a/fs/kernfs/mount.c
+++ b/fs/kernfs/mount.c
@@ -206,7 +206,7 @@ struct dentry *kernfs_node_dentry(struct kernfs_node *kn,
 				  struct super_block *sb)
 {
 	struct dentry *dentry;
-	struct kernfs_node *knparent = NULL;
+	struct kernfs_node *knparent;
 
 	BUG_ON(sb->s_op != &kernfs_sops);
 
@@ -358,7 +358,9 @@ int kernfs_get_tree(struct fs_context *fc)
 		}
 		sb->s_flags |= SB_ACTIVE;
 
-		uuid_gen(&sb->s_uuid);
+		uuid_t uuid;
+		uuid_gen(&uuid);
+		super_set_uuid(sb, uuid.b, sizeof(uuid));
 
 		down_write(&root->kernfs_supers_rwsem);
 		list_add(&info->node, &info->root->supers);
diff --git a/fs/libfs.c b/fs/libfs.c
index ef700d39f0..b635ee5adb 100644
--- a/fs/libfs.c
+++ b/fs/libfs.c
@@ -23,6 +23,7 @@
 #include <linux/fsnotify.h>
 #include <linux/unicode.h>
 #include <linux/fscrypt.h>
+#include <linux/pidfs.h>
 
 #include <linux/uaccess.h>
 
@@ -1657,7 +1658,7 @@ EXPORT_SYMBOL(alloc_anon_inode);
  * All arguments are ignored and it just returns -EINVAL.
  */
 int
-simple_nosetlease(struct file *filp, int arg, struct file_lock **flp,
+simple_nosetlease(struct file *filp, int arg, struct file_lease **flp,
 		  void **priv)
 {
 	return -EINVAL;
@@ -1781,16 +1782,28 @@ bool is_empty_dir_inode(struct inode *inode)
 static int generic_ci_d_compare(const struct dentry *dentry, unsigned int len,
 				const char *str, const struct qstr *name)
 {
-	const struct dentry *parent = READ_ONCE(dentry->d_parent);
-	const struct inode *dir = READ_ONCE(parent->d_inode);
-	const struct super_block *sb = dentry->d_sb;
-	const struct unicode_map *um = sb->s_encoding;
-	struct qstr qstr = QSTR_INIT(str, len);
+	const struct dentry *parent;
+	const struct inode *dir;
 	char strbuf[DNAME_INLINE_LEN];
-	int ret;
+	struct qstr qstr;
+
+	/*
+	 * Attempt a case-sensitive match first. It is cheaper and
+	 * should cover most lookups, including all the sane
+	 * applications that expect a case-sensitive filesystem.
+	 *
+	 * This comparison is safe under RCU because the caller
+	 * guarantees the consistency between str and len. See
+	 * __d_lookup_rcu_op_compare() for details.
+	 */
+	if (len == name->len && !memcmp(str, name->name, len))
+		return 0;
 
+	parent = READ_ONCE(dentry->d_parent);
+	dir = READ_ONCE(parent->d_inode);
 	if (!dir || !IS_CASEFOLDED(dir))
-		goto fallback;
+		return 1;
+
 	/*
 	 * If the dentry name is stored in-line, then it may be concurrently
 	 * modified by a rename.  If this happens, the VFS will eventually retry
@@ -1801,20 +1814,14 @@ static int generic_ci_d_compare(const struct dentry *dentry, unsigned int len,
 	if (len <= DNAME_INLINE_LEN - 1) {
 		memcpy(strbuf, str, len);
 		strbuf[len] = 0;
-		qstr.name = strbuf;
+		str = strbuf;
 		/* prevent compiler from optimizing out the temporary buffer */
 		barrier();
 	}
-	ret = utf8_strncasecmp(um, name, &qstr);
-	if (ret >= 0)
-		return ret;
+	qstr.len = len;
+	qstr.name = str;
 
-	if (sb_has_strict_encoding(sb))
-		return -EINVAL;
-fallback:
-	if (len != name->len)
-		return 1;
-	return !!memcmp(str, name->name, len);
+	return utf8_strncasecmp(dentry->d_sb->s_encoding, name, &qstr);
 }
 
 /**
@@ -1829,7 +1836,7 @@ static int generic_ci_d_hash(const struct dentry *dentry, struct qstr *str)
 	const struct inode *dir = READ_ONCE(dentry->d_inode);
 	struct super_block *sb = dentry->d_sb;
 	const struct unicode_map *um = sb->s_encoding;
-	int ret = 0;
+	int ret;
 
 	if (!dir || !IS_CASEFOLDED(dir))
 		return 0;
@@ -1843,73 +1850,45 @@ static int generic_ci_d_hash(const struct dentry *dentry, struct qstr *str)
 static const struct dentry_operations generic_ci_dentry_ops = {
 	.d_hash = generic_ci_d_hash,
 	.d_compare = generic_ci_d_compare,
-};
-#endif
-
 #ifdef CONFIG_FS_ENCRYPTION
-static const struct dentry_operations generic_encrypted_dentry_ops = {
 	.d_revalidate = fscrypt_d_revalidate,
+#endif
 };
 #endif
 
-#if defined(CONFIG_FS_ENCRYPTION) && IS_ENABLED(CONFIG_UNICODE)
-static const struct dentry_operations generic_encrypted_ci_dentry_ops = {
-	.d_hash = generic_ci_d_hash,
-	.d_compare = generic_ci_d_compare,
+#ifdef CONFIG_FS_ENCRYPTION
+static const struct dentry_operations generic_encrypted_dentry_ops = {
 	.d_revalidate = fscrypt_d_revalidate,
 };
 #endif
 
 /**
- * generic_set_encrypted_ci_d_ops - helper for setting d_ops for given dentry
- * @dentry:	dentry to set ops on
- *
- * Casefolded directories need d_hash and d_compare set, so that the dentries
- * contained in them are handled case-insensitively.  Note that these operations
- * are needed on the parent directory rather than on the dentries in it, and
- * while the casefolding flag can be toggled on and off on an empty directory,
- * dentry_operations can't be changed later.  As a result, if the filesystem has
- * casefolding support enabled at all, we have to give all dentries the
- * casefolding operations even if their inode doesn't have the casefolding flag
- * currently (and thus the casefolding ops would be no-ops for now).
+ * generic_set_sb_d_ops - helper for choosing the set of
+ * filesystem-wide dentry operations for the enabled features
+ * @sb: superblock to be configured
  *
- * Encryption works differently in that the only dentry operation it needs is
- * d_revalidate, which it only needs on dentries that have the no-key name flag.
- * The no-key flag can't be set "later", so we don't have to worry about that.
- *
- * Finally, to maximize compatibility with overlayfs (which isn't compatible
- * with certain dentry operations) and to avoid taking an unnecessary
- * performance hit, we use custom dentry_operations for each possible
- * combination rather than always installing all operations.
+ * Filesystems supporting casefolding and/or fscrypt can call this
+ * helper at mount-time to configure sb->s_d_op to best set of dentry
+ * operations required for the enabled features. The helper must be
+ * called after these have been configured, but before the root dentry
+ * is created.
  */
-void generic_set_encrypted_ci_d_ops(struct dentry *dentry)
+void generic_set_sb_d_ops(struct super_block *sb)
 {
-#ifdef CONFIG_FS_ENCRYPTION
-	bool needs_encrypt_ops = dentry->d_flags & DCACHE_NOKEY_NAME;
-#endif
 #if IS_ENABLED(CONFIG_UNICODE)
-	bool needs_ci_ops = dentry->d_sb->s_encoding;
-#endif
-#if defined(CONFIG_FS_ENCRYPTION) && IS_ENABLED(CONFIG_UNICODE)
-	if (needs_encrypt_ops && needs_ci_ops) {
-		d_set_d_op(dentry, &generic_encrypted_ci_dentry_ops);
+	if (sb->s_encoding) {
+		sb->s_d_op = &generic_ci_dentry_ops;
 		return;
 	}
 #endif
 #ifdef CONFIG_FS_ENCRYPTION
-	if (needs_encrypt_ops) {
-		d_set_d_op(dentry, &generic_encrypted_dentry_ops);
-		return;
-	}
-#endif
-#if IS_ENABLED(CONFIG_UNICODE)
-	if (needs_ci_ops) {
-		d_set_d_op(dentry, &generic_ci_dentry_ops);
+	if (sb->s_cop) {
+		sb->s_d_op = &generic_encrypted_dentry_ops;
 		return;
 	}
 #endif
 }
-EXPORT_SYMBOL(generic_set_encrypted_ci_d_ops);
+EXPORT_SYMBOL(generic_set_sb_d_ops);
 
 /**
  * inode_maybe_inc_iversion - increments i_version
@@ -2050,3 +2029,147 @@ struct timespec64 simple_inode_init_ts(struct inode *inode)
 	return ts;
 }
 EXPORT_SYMBOL(simple_inode_init_ts);
+
+static inline struct dentry *get_stashed_dentry(struct dentry *stashed)
+{
+	struct dentry *dentry;
+
+	guard(rcu)();
+	dentry = READ_ONCE(stashed);
+	if (!dentry)
+		return NULL;
+	if (!lockref_get_not_dead(&dentry->d_lockref))
+		return NULL;
+	return dentry;
+}
+
+static struct dentry *prepare_anon_dentry(struct dentry **stashed,
+					  struct super_block *sb,
+					  void *data)
+{
+	struct dentry *dentry;
+	struct inode *inode;
+	const struct stashed_operations *sops = sb->s_fs_info;
+	int ret;
+
+	inode = new_inode_pseudo(sb);
+	if (!inode) {
+		sops->put_data(data);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	inode->i_flags |= S_IMMUTABLE;
+	inode->i_mode = S_IFREG;
+	simple_inode_init_ts(inode);
+
+	ret = sops->init_inode(inode, data);
+	if (ret < 0) {
+		iput(inode);
+		return ERR_PTR(ret);
+	}
+
+	/* Notice when this is changed. */
+	WARN_ON_ONCE(!S_ISREG(inode->i_mode));
+	WARN_ON_ONCE(!IS_IMMUTABLE(inode));
+
+	dentry = d_alloc_anon(sb);
+	if (!dentry) {
+		iput(inode);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	/* Store address of location where dentry's supposed to be stashed. */
+	dentry->d_fsdata = stashed;
+
+	/* @data is now owned by the fs */
+	d_instantiate(dentry, inode);
+	return dentry;
+}
+
+static struct dentry *stash_dentry(struct dentry **stashed,
+				   struct dentry *dentry)
+{
+	guard(rcu)();
+	for (;;) {
+		struct dentry *old;
+
+		/* Assume any old dentry was cleared out. */
+		old = cmpxchg(stashed, NULL, dentry);
+		if (likely(!old))
+			return dentry;
+
+		/* Check if somebody else installed a reusable dentry. */
+		if (lockref_get_not_dead(&old->d_lockref))
+			return old;
+
+		/* There's an old dead dentry there, try to take it over. */
+		if (likely(try_cmpxchg(stashed, &old, dentry)))
+			return dentry;
+	}
+}
+
+/**
+ * path_from_stashed - create path from stashed or new dentry
+ * @stashed:    where to retrieve or stash dentry
+ * @mnt:        mnt of the filesystems to use
+ * @data:       data to store in inode->i_private
+ * @path:       path to create
+ *
+ * The function tries to retrieve a stashed dentry from @stashed. If the dentry
+ * is still valid then it will be reused. If the dentry isn't able the function
+ * will allocate a new dentry and inode. It will then check again whether it
+ * can reuse an existing dentry in case one has been added in the meantime or
+ * update @stashed with the newly added dentry.
+ *
+ * Special-purpose helper for nsfs and pidfs.
+ *
+ * Return: On success zero and on failure a negative error is returned.
+ */
+int path_from_stashed(struct dentry **stashed, struct vfsmount *mnt, void *data,
+		      struct path *path)
+{
+	struct dentry *dentry;
+	const struct stashed_operations *sops = mnt->mnt_sb->s_fs_info;
+
+	/* See if dentry can be reused. */
+	path->dentry = get_stashed_dentry(*stashed);
+	if (path->dentry) {
+		sops->put_data(data);
+		goto out_path;
+	}
+
+	/* Allocate a new dentry. */
+	dentry = prepare_anon_dentry(stashed, mnt->mnt_sb, data);
+	if (IS_ERR(dentry))
+		return PTR_ERR(dentry);
+
+	/* Added a new dentry. @data is now owned by the filesystem. */
+	path->dentry = stash_dentry(stashed, dentry);
+	if (path->dentry != dentry)
+		dput(dentry);
+
+out_path:
+	WARN_ON_ONCE(path->dentry->d_fsdata != stashed);
+	WARN_ON_ONCE(d_inode(path->dentry)->i_private != data);
+	path->mnt = mntget(mnt);
+	return 0;
+}
+
+void stashed_dentry_prune(struct dentry *dentry)
+{
+	struct dentry **stashed = dentry->d_fsdata;
+	struct inode *inode = d_inode(dentry);
+
+	if (WARN_ON_ONCE(!stashed))
+		return;
+
+	if (!inode)
+		return;
+
+	/*
+	 * Only replace our own @dentry as someone else might've
+	 * already cleared out @dentry and stashed their own
+	 * dentry in there.
+	 */
+	cmpxchg(stashed, dentry, NULL);
+}
diff --git a/fs/lockd/clnt4xdr.c b/fs/lockd/clnt4xdr.c
index 8161667c97..527458db45 100644
--- a/fs/lockd/clnt4xdr.c
+++ b/fs/lockd/clnt4xdr.c
@@ -243,7 +243,7 @@ static void encode_nlm4_holder(struct xdr_stream *xdr,
 	u64 l_offset, l_len;
 	__be32 *p;
 
-	encode_bool(xdr, lock->fl.fl_type == F_RDLCK);
+	encode_bool(xdr, lock->fl.c.flc_type == F_RDLCK);
 	encode_int32(xdr, lock->svid);
 	encode_netobj(xdr, lock->oh.data, lock->oh.len);
 
@@ -270,7 +270,7 @@ static int decode_nlm4_holder(struct xdr_stream *xdr, struct nlm_res *result)
 		goto out_overflow;
 	exclusive = be32_to_cpup(p++);
 	lock->svid = be32_to_cpup(p);
-	fl->fl_pid = (pid_t)lock->svid;
+	fl->c.flc_pid = (pid_t)lock->svid;
 
 	error = decode_netobj(xdr, &lock->oh);
 	if (unlikely(error))
@@ -280,8 +280,8 @@ static int decode_nlm4_holder(struct xdr_stream *xdr, struct nlm_res *result)
 	if (unlikely(p == NULL))
 		goto out_overflow;
 
-	fl->fl_flags = FL_POSIX;
-	fl->fl_type  = exclusive != 0 ? F_WRLCK : F_RDLCK;
+	fl->c.flc_flags = FL_POSIX;
+	fl->c.flc_type  = exclusive != 0 ? F_WRLCK : F_RDLCK;
 	p = xdr_decode_hyper(p, &l_offset);
 	xdr_decode_hyper(p, &l_len);
 	nlm4svc_set_file_lock_range(fl, l_offset, l_len);
@@ -357,7 +357,7 @@ static void nlm4_xdr_enc_testargs(struct rpc_rqst *req,
 	const struct nlm_lock *lock = &args->lock;
 
 	encode_cookie(xdr, &args->cookie);
-	encode_bool(xdr, lock->fl.fl_type == F_WRLCK);
+	encode_bool(xdr, lock->fl.c.flc_type == F_WRLCK);
 	encode_nlm4_lock(xdr, lock);
 }
 
@@ -380,7 +380,7 @@ static void nlm4_xdr_enc_lockargs(struct rpc_rqst *req,
 
 	encode_cookie(xdr, &args->cookie);
 	encode_bool(xdr, args->block);
-	encode_bool(xdr, lock->fl.fl_type == F_WRLCK);
+	encode_bool(xdr, lock->fl.c.flc_type == F_WRLCK);
 	encode_nlm4_lock(xdr, lock);
 	encode_bool(xdr, args->reclaim);
 	encode_int32(xdr, args->state);
@@ -403,7 +403,7 @@ static void nlm4_xdr_enc_cancargs(struct rpc_rqst *req,
 
 	encode_cookie(xdr, &args->cookie);
 	encode_bool(xdr, args->block);
-	encode_bool(xdr, lock->fl.fl_type == F_WRLCK);
+	encode_bool(xdr, lock->fl.c.flc_type == F_WRLCK);
 	encode_nlm4_lock(xdr, lock);
 }
 
diff --git a/fs/lockd/clntlock.c b/fs/lockd/clntlock.c
index 5d85715be7..a7e0519ec0 100644
--- a/fs/lockd/clntlock.c
+++ b/fs/lockd/clntlock.c
@@ -185,7 +185,7 @@ __be32 nlmclnt_grant(const struct sockaddr *addr, const struct nlm_lock *lock)
 			continue;
 		if (!rpc_cmp_addr(nlm_addr(block->b_host), addr))
 			continue;
-		if (nfs_compare_fh(NFS_FH(file_inode(fl_blocked->fl_file)), fh) != 0)
+		if (nfs_compare_fh(NFS_FH(file_inode(fl_blocked->c.flc_file)), fh) != 0)
 			continue;
 		/* Alright, we found a lock. Set the return status
 		 * and wake up the caller
diff --git a/fs/lockd/clntproc.c b/fs/lockd/clntproc.c
index fba6c7fa74..cebcc283b7 100644
--- a/fs/lockd/clntproc.c
+++ b/fs/lockd/clntproc.c
@@ -133,7 +133,8 @@ static void nlmclnt_setlockargs(struct nlm_rqst *req, struct file_lock *fl)
 	char *nodename = req->a_host->h_rpcclnt->cl_nodename;
 
 	nlmclnt_next_cookie(&argp->cookie);
-	memcpy(&lock->fh, NFS_FH(file_inode(fl->fl_file)), sizeof(struct nfs_fh));
+	memcpy(&lock->fh, NFS_FH(file_inode(fl->c.flc_file)),
+	       sizeof(struct nfs_fh));
 	lock->caller  = nodename;
 	lock->oh.data = req->a_owner;
 	lock->oh.len  = snprintf(req->a_owner, sizeof(req->a_owner), "%u@%s",
@@ -142,7 +143,7 @@ static void nlmclnt_setlockargs(struct nlm_rqst *req, struct file_lock *fl)
 	lock->svid = fl->fl_u.nfs_fl.owner->pid;
 	lock->fl.fl_start = fl->fl_start;
 	lock->fl.fl_end = fl->fl_end;
-	lock->fl.fl_type = fl->fl_type;
+	lock->fl.c.flc_type = fl->c.flc_type;
 }
 
 static void nlmclnt_release_lockargs(struct nlm_rqst *req)
@@ -182,7 +183,7 @@ int nlmclnt_proc(struct nlm_host *host, int cmd, struct file_lock *fl, void *dat
 	call->a_callback_data = data;
 
 	if (IS_SETLK(cmd) || IS_SETLKW(cmd)) {
-		if (fl->fl_type != F_UNLCK) {
+		if (fl->c.flc_type != F_UNLCK) {
 			call->a_args.block = IS_SETLKW(cmd) ? 1 : 0;
 			status = nlmclnt_lock(call, fl);
 		} else
@@ -432,13 +433,14 @@ nlmclnt_test(struct nlm_rqst *req, struct file_lock *fl)
 {
 	int	status;
 
-	status = nlmclnt_call(nfs_file_cred(fl->fl_file), req, NLMPROC_TEST);
+	status = nlmclnt_call(nfs_file_cred(fl->c.flc_file), req,
+			      NLMPROC_TEST);
 	if (status < 0)
 		goto out;
 
 	switch (req->a_res.status) {
 		case nlm_granted:
-			fl->fl_type = F_UNLCK;
+			fl->c.flc_type = F_UNLCK;
 			break;
 		case nlm_lck_denied:
 			/*
@@ -446,8 +448,8 @@ nlmclnt_test(struct nlm_rqst *req, struct file_lock *fl)
 			 */
 			fl->fl_start = req->a_res.lock.fl.fl_start;
 			fl->fl_end = req->a_res.lock.fl.fl_end;
-			fl->fl_type = req->a_res.lock.fl.fl_type;
-			fl->fl_pid = -req->a_res.lock.fl.fl_pid;
+			fl->c.flc_type = req->a_res.lock.fl.c.flc_type;
+			fl->c.flc_pid = -req->a_res.lock.fl.c.flc_pid;
 			break;
 		default:
 			status = nlm_stat_to_errno(req->a_res.status);
@@ -485,14 +487,15 @@ static const struct file_lock_operations nlmclnt_lock_ops = {
 static void nlmclnt_locks_init_private(struct file_lock *fl, struct nlm_host *host)
 {
 	fl->fl_u.nfs_fl.state = 0;
-	fl->fl_u.nfs_fl.owner = nlmclnt_find_lockowner(host, fl->fl_owner);
+	fl->fl_u.nfs_fl.owner = nlmclnt_find_lockowner(host,
+						       fl->c.flc_owner);
 	INIT_LIST_HEAD(&fl->fl_u.nfs_fl.list);
 	fl->fl_ops = &nlmclnt_lock_ops;
 }
 
 static int do_vfs_lock(struct file_lock *fl)
 {
-	return locks_lock_file_wait(fl->fl_file, fl);
+	return locks_lock_file_wait(fl->c.flc_file, fl);
 }
 
 /*
@@ -518,12 +521,12 @@ static int do_vfs_lock(struct file_lock *fl)
 static int
 nlmclnt_lock(struct nlm_rqst *req, struct file_lock *fl)
 {
-	const struct cred *cred = nfs_file_cred(fl->fl_file);
+	const struct cred *cred = nfs_file_cred(fl->c.flc_file);
 	struct nlm_host	*host = req->a_host;
 	struct nlm_res	*resp = &req->a_res;
 	struct nlm_wait block;
-	unsigned char fl_flags = fl->fl_flags;
-	unsigned char fl_type;
+	unsigned char flags = fl->c.flc_flags;
+	unsigned char type;
 	__be32 b_status;
 	int status = -ENOLCK;
 
@@ -531,9 +534,9 @@ nlmclnt_lock(struct nlm_rqst *req, struct file_lock *fl)
 		goto out;
 	req->a_args.state = nsm_local_state;
 
-	fl->fl_flags |= FL_ACCESS;
+	fl->c.flc_flags |= FL_ACCESS;
 	status = do_vfs_lock(fl);
-	fl->fl_flags = fl_flags;
+	fl->c.flc_flags = flags;
 	if (status < 0)
 		goto out;
 
@@ -591,11 +594,11 @@ again:
 			goto again;
 		}
 		/* Ensure the resulting lock will get added to granted list */
-		fl->fl_flags |= FL_SLEEP;
+		fl->c.flc_flags |= FL_SLEEP;
 		if (do_vfs_lock(fl) < 0)
 			printk(KERN_WARNING "%s: VFS is out of sync with lock manager!\n", __func__);
 		up_read(&host->h_rwsem);
-		fl->fl_flags = fl_flags;
+		fl->c.flc_flags = flags;
 		status = 0;
 	}
 	if (status < 0)
@@ -605,7 +608,7 @@ again:
 	 * cases NLM_LCK_DENIED is returned for a permanent error.  So
 	 * turn it into an ENOLCK.
 	 */
-	if (resp->status == nlm_lck_denied && (fl_flags & FL_SLEEP))
+	if (resp->status == nlm_lck_denied && (flags & FL_SLEEP))
 		status = -ENOLCK;
 	else
 		status = nlm_stat_to_errno(resp->status);
@@ -622,13 +625,13 @@ out_unlock:
 			   req->a_host->h_addrlen, req->a_res.status);
 	dprintk("lockd: lock attempt ended in fatal error.\n"
 		"       Attempting to unlock.\n");
-	fl_type = fl->fl_type;
-	fl->fl_type = F_UNLCK;
+	type = fl->c.flc_type;
+	fl->c.flc_type = F_UNLCK;
 	down_read(&host->h_rwsem);
 	do_vfs_lock(fl);
 	up_read(&host->h_rwsem);
-	fl->fl_type = fl_type;
-	fl->fl_flags = fl_flags;
+	fl->c.flc_type = type;
+	fl->c.flc_flags = flags;
 	nlmclnt_async_call(cred, req, NLMPROC_UNLOCK, &nlmclnt_unlock_ops);
 	return status;
 }
@@ -651,12 +654,14 @@ nlmclnt_reclaim(struct nlm_host *host, struct file_lock *fl,
 	nlmclnt_setlockargs(req, fl);
 	req->a_args.reclaim = 1;
 
-	status = nlmclnt_call(nfs_file_cred(fl->fl_file), req, NLMPROC_LOCK);
+	status = nlmclnt_call(nfs_file_cred(fl->c.flc_file), req,
+			      NLMPROC_LOCK);
 	if (status >= 0 && req->a_res.status == nlm_granted)
 		return 0;
 
 	printk(KERN_WARNING "lockd: failed to reclaim lock for pid %d "
-				"(errno %d, status %d)\n", fl->fl_pid,
+				"(errno %d, status %d)\n",
+				fl->c.flc_pid,
 				status, ntohl(req->a_res.status));
 
 	/*
@@ -683,26 +688,26 @@ nlmclnt_unlock(struct nlm_rqst *req, struct file_lock *fl)
 	struct nlm_host	*host = req->a_host;
 	struct nlm_res	*resp = &req->a_res;
 	int status;
-	unsigned char fl_flags = fl->fl_flags;
+	unsigned char flags = fl->c.flc_flags;
 
 	/*
 	 * Note: the server is supposed to either grant us the unlock
 	 * request, or to deny it with NLM_LCK_DENIED_GRACE_PERIOD. In either
 	 * case, we want to unlock.
 	 */
-	fl->fl_flags |= FL_EXISTS;
+	fl->c.flc_flags |= FL_EXISTS;
 	down_read(&host->h_rwsem);
 	status = do_vfs_lock(fl);
 	up_read(&host->h_rwsem);
-	fl->fl_flags = fl_flags;
+	fl->c.flc_flags = flags;
 	if (status == -ENOENT) {
 		status = 0;
 		goto out;
 	}
 
 	refcount_inc(&req->a_count);
-	status = nlmclnt_async_call(nfs_file_cred(fl->fl_file), req,
-			NLMPROC_UNLOCK, &nlmclnt_unlock_ops);
+	status = nlmclnt_async_call(nfs_file_cred(fl->c.flc_file), req,
+				    NLMPROC_UNLOCK, &nlmclnt_unlock_ops);
 	if (status < 0)
 		goto out;
 
@@ -795,8 +800,8 @@ static int nlmclnt_cancel(struct nlm_host *host, int block, struct file_lock *fl
 	req->a_args.block = block;
 
 	refcount_inc(&req->a_count);
-	status = nlmclnt_async_call(nfs_file_cred(fl->fl_file), req,
-			NLMPROC_CANCEL, &nlmclnt_cancel_ops);
+	status = nlmclnt_async_call(nfs_file_cred(fl->c.flc_file), req,
+				    NLMPROC_CANCEL, &nlmclnt_cancel_ops);
 	if (status == 0 && req->a_res.status == nlm_lck_denied)
 		status = -ENOLCK;
 	nlmclnt_release_call(req);
diff --git a/fs/lockd/clntxdr.c b/fs/lockd/clntxdr.c
index 4df62f6355..a3e97278b9 100644
--- a/fs/lockd/clntxdr.c
+++ b/fs/lockd/clntxdr.c
@@ -238,7 +238,7 @@ static void encode_nlm_holder(struct xdr_stream *xdr,
 	u32 l_offset, l_len;
 	__be32 *p;
 
-	encode_bool(xdr, lock->fl.fl_type == F_RDLCK);
+	encode_bool(xdr, lock->fl.c.flc_type == F_RDLCK);
 	encode_int32(xdr, lock->svid);
 	encode_netobj(xdr, lock->oh.data, lock->oh.len);
 
@@ -265,7 +265,7 @@ static int decode_nlm_holder(struct xdr_stream *xdr, struct nlm_res *result)
 		goto out_overflow;
 	exclusive = be32_to_cpup(p++);
 	lock->svid = be32_to_cpup(p);
-	fl->fl_pid = (pid_t)lock->svid;
+	fl->c.flc_pid = (pid_t)lock->svid;
 
 	error = decode_netobj(xdr, &lock->oh);
 	if (unlikely(error))
@@ -275,8 +275,8 @@ static int decode_nlm_holder(struct xdr_stream *xdr, struct nlm_res *result)
 	if (unlikely(p == NULL))
 		goto out_overflow;
 
-	fl->fl_flags = FL_POSIX;
-	fl->fl_type  = exclusive != 0 ? F_WRLCK : F_RDLCK;
+	fl->c.flc_flags = FL_POSIX;
+	fl->c.flc_type  = exclusive != 0 ? F_WRLCK : F_RDLCK;
 	l_offset = be32_to_cpup(p++);
 	l_len = be32_to_cpup(p);
 	end = l_offset + l_len - 1;
@@ -357,7 +357,7 @@ static void nlm_xdr_enc_testargs(struct rpc_rqst *req,
 	const struct nlm_lock *lock = &args->lock;
 
 	encode_cookie(xdr, &args->cookie);
-	encode_bool(xdr, lock->fl.fl_type == F_WRLCK);
+	encode_bool(xdr, lock->fl.c.flc_type == F_WRLCK);
 	encode_nlm_lock(xdr, lock);
 }
 
@@ -380,7 +380,7 @@ static void nlm_xdr_enc_lockargs(struct rpc_rqst *req,
 
 	encode_cookie(xdr, &args->cookie);
 	encode_bool(xdr, args->block);
-	encode_bool(xdr, lock->fl.fl_type == F_WRLCK);
+	encode_bool(xdr, lock->fl.c.flc_type == F_WRLCK);
 	encode_nlm_lock(xdr, lock);
 	encode_bool(xdr, args->reclaim);
 	encode_int32(xdr, args->state);
@@ -403,7 +403,7 @@ static void nlm_xdr_enc_cancargs(struct rpc_rqst *req,
 
 	encode_cookie(xdr, &args->cookie);
 	encode_bool(xdr, args->block);
-	encode_bool(xdr, lock->fl.fl_type == F_WRLCK);
+	encode_bool(xdr, lock->fl.c.flc_type == F_WRLCK);
 	encode_nlm_lock(xdr, lock);
 }
 
diff --git a/fs/lockd/host.c b/fs/lockd/host.c
index 127a728fcb..c115168017 100644
--- a/fs/lockd/host.c
+++ b/fs/lockd/host.c
@@ -117,7 +117,6 @@ static struct nlm_host *nlm_alloc_host(struct nlm_lookup_host_info *ni,
 	if (nsm != NULL)
 		refcount_inc(&nsm->sm_count);
 	else {
-		host = NULL;
 		nsm = nsm_get_handle(ni->net, ni->sap, ni->salen,
 					ni->hostname, ni->hostname_len);
 		if (unlikely(nsm == NULL)) {
diff --git a/fs/lockd/svc.c b/fs/lockd/svc.c
index ce58624820..ab8042a5b8 100644
--- a/fs/lockd/svc.c
+++ b/fs/lockd/svc.c
@@ -710,8 +710,6 @@ static const struct svc_version *nlmsvc_version[] = {
 #endif
 };
 
-static struct svc_stat		nlmsvc_stats;
-
 #define NLM_NRVERS	ARRAY_SIZE(nlmsvc_version)
 static struct svc_program	nlmsvc_program = {
 	.pg_prog		= NLM_PROGRAM,		/* program number */
@@ -719,7 +717,6 @@ static struct svc_program	nlmsvc_program = {
 	.pg_vers		= nlmsvc_version,	/* version table */
 	.pg_name		= "lockd",		/* service name */
 	.pg_class		= "nfsd",		/* share authentication with nfsd */
-	.pg_stats		= &nlmsvc_stats,	/* stats table */
 	.pg_authenticate	= &lockd_authenticate,	/* export authentication */
 	.pg_init_request	= svc_generic_init_request,
 	.pg_rpcbind_set		= svc_generic_rpcbind_set,
diff --git a/fs/lockd/svc4proc.c b/fs/lockd/svc4proc.c
index b72023a6b4..8a72c418cd 100644
--- a/fs/lockd/svc4proc.c
+++ b/fs/lockd/svc4proc.c
@@ -52,16 +52,16 @@ nlm4svc_retrieve_args(struct svc_rqst *rqstp, struct nlm_args *argp,
 		*filp = file;
 
 		/* Set up the missing parts of the file_lock structure */
-		lock->fl.fl_flags = FL_POSIX;
-		lock->fl.fl_file  = file->f_file[mode];
-		lock->fl.fl_pid = current->tgid;
+		lock->fl.c.flc_flags = FL_POSIX;
+		lock->fl.c.flc_file  = file->f_file[mode];
+		lock->fl.c.flc_pid = current->tgid;
 		lock->fl.fl_start = (loff_t)lock->lock_start;
 		lock->fl.fl_end = lock->lock_len ?
 				   (loff_t)(lock->lock_start + lock->lock_len - 1) :
 				   OFFSET_MAX;
 		lock->fl.fl_lmops = &nlmsvc_lock_operations;
 		nlmsvc_locks_init_private(&lock->fl, host, (pid_t)lock->svid);
-		if (!lock->fl.fl_owner) {
+		if (!lock->fl.c.flc_owner) {
 			/* lockowner allocation has failed */
 			nlmsvc_release_host(host);
 			return nlm_lck_denied_nolocks;
@@ -106,7 +106,7 @@ __nlm4svc_proc_test(struct svc_rqst *rqstp, struct nlm_res *resp)
 	if ((resp->status = nlm4svc_retrieve_args(rqstp, argp, &host, &file)))
 		return resp->status == nlm_drop_reply ? rpc_drop_reply :rpc_success;
 
-	test_owner = argp->lock.fl.fl_owner;
+	test_owner = argp->lock.fl.c.flc_owner;
 	/* Now check for conflicting locks */
 	resp->status = nlmsvc_testlock(rqstp, file, host, &argp->lock, &resp->lock, &resp->cookie);
 	if (resp->status == nlm_drop_reply)
diff --git a/fs/lockd/svclock.c b/fs/lockd/svclock.c
index 2dc10900ad..1f2149db10 100644
--- a/fs/lockd/svclock.c
+++ b/fs/lockd/svclock.c
@@ -150,16 +150,17 @@ nlmsvc_lookup_block(struct nlm_file *file, struct nlm_lock *lock)
 	struct file_lock	*fl;
 
 	dprintk("lockd: nlmsvc_lookup_block f=%p pd=%d %Ld-%Ld ty=%d\n",
-				file, lock->fl.fl_pid,
+				file, lock->fl.c.flc_pid,
 				(long long)lock->fl.fl_start,
-				(long long)lock->fl.fl_end, lock->fl.fl_type);
+				(long long)lock->fl.fl_end,
+				lock->fl.c.flc_type);
 	spin_lock(&nlm_blocked_lock);
 	list_for_each_entry(block, &nlm_blocked, b_list) {
 		fl = &block->b_call->a_args.lock.fl;
 		dprintk("lockd: check f=%p pd=%d %Ld-%Ld ty=%d cookie=%s\n",
-				block->b_file, fl->fl_pid,
+				block->b_file, fl->c.flc_pid,
 				(long long)fl->fl_start,
-				(long long)fl->fl_end, fl->fl_type,
+				(long long)fl->fl_end, fl->c.flc_type,
 				nlmdbg_cookie2a(&block->b_call->a_args.cookie));
 		if (block->b_file == file && nlm_compare_locks(fl, &lock->fl)) {
 			kref_get(&block->b_count);
@@ -244,7 +245,7 @@ nlmsvc_create_block(struct svc_rqst *rqstp, struct nlm_host *host,
 		goto failed_free;
 
 	/* Set notifier function for VFS, and init args */
-	call->a_args.lock.fl.fl_flags |= FL_SLEEP;
+	call->a_args.lock.fl.c.flc_flags |= FL_SLEEP;
 	call->a_args.lock.fl.fl_lmops = &nlmsvc_lock_operations;
 	nlmclnt_next_cookie(&call->a_args.cookie);
 
@@ -402,14 +403,14 @@ static struct nlm_lockowner *nlmsvc_find_lockowner(struct nlm_host *host, pid_t
 void
 nlmsvc_release_lockowner(struct nlm_lock *lock)
 {
-	if (lock->fl.fl_owner)
-		nlmsvc_put_lockowner(lock->fl.fl_owner);
+	if (lock->fl.c.flc_owner)
+		nlmsvc_put_lockowner(lock->fl.c.flc_owner);
 }
 
 void nlmsvc_locks_init_private(struct file_lock *fl, struct nlm_host *host,
 						pid_t pid)
 {
-	fl->fl_owner = nlmsvc_find_lockowner(host, pid);
+	fl->c.flc_owner = nlmsvc_find_lockowner(host, pid);
 }
 
 /*
@@ -425,7 +426,7 @@ static int nlmsvc_setgrantargs(struct nlm_rqst *call, struct nlm_lock *lock)
 
 	/* set default data area */
 	call->a_args.lock.oh.data = call->a_owner;
-	call->a_args.lock.svid = ((struct nlm_lockowner *)lock->fl.fl_owner)->pid;
+	call->a_args.lock.svid = ((struct nlm_lockowner *) lock->fl.c.flc_owner)->pid;
 
 	if (lock->oh.len > NLMCLNT_OHSIZE) {
 		void *data = kmalloc(lock->oh.len, GFP_KERNEL);
@@ -489,7 +490,8 @@ nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file,
 
 	dprintk("lockd: nlmsvc_lock(%s/%ld, ty=%d, pi=%d, %Ld-%Ld, bl=%d)\n",
 				inode->i_sb->s_id, inode->i_ino,
-				lock->fl.fl_type, lock->fl.fl_pid,
+				lock->fl.c.flc_type,
+				lock->fl.c.flc_pid,
 				(long long)lock->fl.fl_start,
 				(long long)lock->fl.fl_end,
 				wait);
@@ -512,7 +514,7 @@ nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file,
 			goto out;
 		lock = &block->b_call->a_args.lock;
 	} else
-		lock->fl.fl_flags &= ~FL_SLEEP;
+		lock->fl.c.flc_flags &= ~FL_SLEEP;
 
 	if (block->b_flags & B_QUEUED) {
 		dprintk("lockd: nlmsvc_lock deferred block %p flags %d\n",
@@ -560,10 +562,10 @@ nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file,
 	spin_unlock(&nlm_blocked_lock);
 
 	if (!wait)
-		lock->fl.fl_flags &= ~FL_SLEEP;
+		lock->fl.c.flc_flags &= ~FL_SLEEP;
 	mode = lock_to_openmode(&lock->fl);
 	error = vfs_lock_file(file->f_file[mode], F_SETLK, &lock->fl, NULL);
-	lock->fl.fl_flags &= ~FL_SLEEP;
+	lock->fl.c.flc_flags &= ~FL_SLEEP;
 
 	dprintk("lockd: vfs_lock_file returned %d\n", error);
 	switch (error) {
@@ -616,7 +618,7 @@ nlmsvc_testlock(struct svc_rqst *rqstp, struct nlm_file *file,
 	dprintk("lockd: nlmsvc_testlock(%s/%ld, ty=%d, %Ld-%Ld)\n",
 				nlmsvc_file_inode(file)->i_sb->s_id,
 				nlmsvc_file_inode(file)->i_ino,
-				lock->fl.fl_type,
+				lock->fl.c.flc_type,
 				(long long)lock->fl.fl_start,
 				(long long)lock->fl.fl_end);
 
@@ -636,19 +638,19 @@ nlmsvc_testlock(struct svc_rqst *rqstp, struct nlm_file *file,
 		goto out;
 	}
 
-	if (lock->fl.fl_type == F_UNLCK) {
+	if (lock->fl.c.flc_type == F_UNLCK) {
 		ret = nlm_granted;
 		goto out;
 	}
 
 	dprintk("lockd: conflicting lock(ty=%d, %Ld-%Ld)\n",
-		lock->fl.fl_type, (long long)lock->fl.fl_start,
+		lock->fl.c.flc_type, (long long)lock->fl.fl_start,
 		(long long)lock->fl.fl_end);
 	conflock->caller = "somehost";	/* FIXME */
 	conflock->len = strlen(conflock->caller);
 	conflock->oh.len = 0;		/* don't return OH info */
-	conflock->svid = lock->fl.fl_pid;
-	conflock->fl.fl_type = lock->fl.fl_type;
+	conflock->svid = lock->fl.c.flc_pid;
+	conflock->fl.c.flc_type = lock->fl.c.flc_type;
 	conflock->fl.fl_start = lock->fl.fl_start;
 	conflock->fl.fl_end = lock->fl.fl_end;
 	locks_release_private(&lock->fl);
@@ -673,21 +675,21 @@ nlmsvc_unlock(struct net *net, struct nlm_file *file, struct nlm_lock *lock)
 	dprintk("lockd: nlmsvc_unlock(%s/%ld, pi=%d, %Ld-%Ld)\n",
 				nlmsvc_file_inode(file)->i_sb->s_id,
 				nlmsvc_file_inode(file)->i_ino,
-				lock->fl.fl_pid,
+				lock->fl.c.flc_pid,
 				(long long)lock->fl.fl_start,
 				(long long)lock->fl.fl_end);
 
 	/* First, cancel any lock that might be there */
 	nlmsvc_cancel_blocked(net, file, lock);
 
-	lock->fl.fl_type = F_UNLCK;
-	lock->fl.fl_file = file->f_file[O_RDONLY];
-	if (lock->fl.fl_file)
-		error = vfs_lock_file(lock->fl.fl_file, F_SETLK,
+	lock->fl.c.flc_type = F_UNLCK;
+	lock->fl.c.flc_file = file->f_file[O_RDONLY];
+	if (lock->fl.c.flc_file)
+		error = vfs_lock_file(lock->fl.c.flc_file, F_SETLK,
 					&lock->fl, NULL);
-	lock->fl.fl_file = file->f_file[O_WRONLY];
-	if (lock->fl.fl_file)
-		error |= vfs_lock_file(lock->fl.fl_file, F_SETLK,
+	lock->fl.c.flc_file = file->f_file[O_WRONLY];
+	if (lock->fl.c.flc_file)
+		error |= vfs_lock_file(lock->fl.c.flc_file, F_SETLK,
 					&lock->fl, NULL);
 
 	return (error < 0)? nlm_lck_denied_nolocks : nlm_granted;
@@ -710,7 +712,7 @@ nlmsvc_cancel_blocked(struct net *net, struct nlm_file *file, struct nlm_lock *l
 	dprintk("lockd: nlmsvc_cancel(%s/%ld, pi=%d, %Ld-%Ld)\n",
 				nlmsvc_file_inode(file)->i_sb->s_id,
 				nlmsvc_file_inode(file)->i_ino,
-				lock->fl.fl_pid,
+				lock->fl.c.flc_pid,
 				(long long)lock->fl.fl_start,
 				(long long)lock->fl.fl_end);
 
@@ -863,12 +865,12 @@ nlmsvc_grant_blocked(struct nlm_block *block)
 	/* vfs_lock_file() can mangle fl_start and fl_end, but we need
 	 * them unchanged for the GRANT_MSG
 	 */
-	lock->fl.fl_flags |= FL_SLEEP;
+	lock->fl.c.flc_flags |= FL_SLEEP;
 	fl_start = lock->fl.fl_start;
 	fl_end = lock->fl.fl_end;
 	mode = lock_to_openmode(&lock->fl);
 	error = vfs_lock_file(file->f_file[mode], F_SETLK, &lock->fl, NULL);
-	lock->fl.fl_flags &= ~FL_SLEEP;
+	lock->fl.c.flc_flags &= ~FL_SLEEP;
 	lock->fl.fl_start = fl_start;
 	lock->fl.fl_end = fl_end;
 
@@ -993,8 +995,8 @@ nlmsvc_grant_reply(struct nlm_cookie *cookie, __be32 status)
 		/* Client doesn't want it, just unlock it */
 		nlmsvc_unlink_block(block);
 		fl = &block->b_call->a_args.lock.fl;
-		fl->fl_type = F_UNLCK;
-		error = vfs_lock_file(fl->fl_file, F_SETLK, fl, NULL);
+		fl->c.flc_type = F_UNLCK;
+		error = vfs_lock_file(fl->c.flc_file, F_SETLK, fl, NULL);
 		if (error)
 			pr_warn("lockd: unable to unlock lock rejected by client!\n");
 		break;
diff --git a/fs/lockd/svcproc.c b/fs/lockd/svcproc.c
index 32784f508c..a03220e66c 100644
--- a/fs/lockd/svcproc.c
+++ b/fs/lockd/svcproc.c
@@ -77,12 +77,12 @@ nlmsvc_retrieve_args(struct svc_rqst *rqstp, struct nlm_args *argp,
 
 		/* Set up the missing parts of the file_lock structure */
 		mode = lock_to_openmode(&lock->fl);
-		lock->fl.fl_flags = FL_POSIX;
-		lock->fl.fl_file  = file->f_file[mode];
-		lock->fl.fl_pid = current->tgid;
+		lock->fl.c.flc_flags = FL_POSIX;
+		lock->fl.c.flc_file  = file->f_file[mode];
+		lock->fl.c.flc_pid = current->tgid;
 		lock->fl.fl_lmops = &nlmsvc_lock_operations;
 		nlmsvc_locks_init_private(&lock->fl, host, (pid_t)lock->svid);
-		if (!lock->fl.fl_owner) {
+		if (!lock->fl.c.flc_owner) {
 			/* lockowner allocation has failed */
 			nlmsvc_release_host(host);
 			return nlm_lck_denied_nolocks;
@@ -127,7 +127,7 @@ __nlmsvc_proc_test(struct svc_rqst *rqstp, struct nlm_res *resp)
 	if ((resp->status = nlmsvc_retrieve_args(rqstp, argp, &host, &file)))
 		return resp->status == nlm_drop_reply ? rpc_drop_reply :rpc_success;
 
-	test_owner = argp->lock.fl.fl_owner;
+	test_owner = argp->lock.fl.c.flc_owner;
 
 	/* Now check for conflicting locks */
 	resp->status = cast_status(nlmsvc_testlock(rqstp, file, host, &argp->lock, &resp->lock, &resp->cookie));
diff --git a/fs/lockd/svcsubs.c b/fs/lockd/svcsubs.c
index e3b6229e7a..9103896164 100644
--- a/fs/lockd/svcsubs.c
+++ b/fs/lockd/svcsubs.c
@@ -73,7 +73,7 @@ static inline unsigned int file_hash(struct nfs_fh *f)
 
 int lock_to_openmode(struct file_lock *lock)
 {
-	return (lock->fl_type == F_WRLCK) ? O_WRONLY : O_RDONLY;
+	return lock_is_write(lock) ? O_WRONLY : O_RDONLY;
 }
 
 /*
@@ -181,18 +181,18 @@ static int nlm_unlock_files(struct nlm_file *file, const struct file_lock *fl)
 	struct file_lock lock;
 
 	locks_init_lock(&lock);
-	lock.fl_type  = F_UNLCK;
+	lock.c.flc_type  = F_UNLCK;
 	lock.fl_start = 0;
 	lock.fl_end   = OFFSET_MAX;
-	lock.fl_owner = fl->fl_owner;
-	lock.fl_pid   = fl->fl_pid;
-	lock.fl_flags = FL_POSIX;
+	lock.c.flc_owner = fl->c.flc_owner;
+	lock.c.flc_pid   = fl->c.flc_pid;
+	lock.c.flc_flags = FL_POSIX;
 
-	lock.fl_file = file->f_file[O_RDONLY];
-	if (lock.fl_file && vfs_lock_file(lock.fl_file, F_SETLK, &lock, NULL))
+	lock.c.flc_file = file->f_file[O_RDONLY];
+	if (lock.c.flc_file && vfs_lock_file(lock.c.flc_file, F_SETLK, &lock, NULL))
 		goto out_err;
-	lock.fl_file = file->f_file[O_WRONLY];
-	if (lock.fl_file && vfs_lock_file(lock.fl_file, F_SETLK, &lock, NULL))
+	lock.c.flc_file = file->f_file[O_WRONLY];
+	if (lock.c.flc_file && vfs_lock_file(lock.c.flc_file, F_SETLK, &lock, NULL))
 		goto out_err;
 	return 0;
 out_err:
@@ -218,14 +218,14 @@ nlm_traverse_locks(struct nlm_host *host, struct nlm_file *file,
 again:
 	file->f_locks = 0;
 	spin_lock(&flctx->flc_lock);
-	list_for_each_entry(fl, &flctx->flc_posix, fl_list) {
+	for_each_file_lock(fl, &flctx->flc_posix) {
 		if (fl->fl_lmops != &nlmsvc_lock_operations)
 			continue;
 
 		/* update current lock count */
 		file->f_locks++;
 
-		lockhost = ((struct nlm_lockowner *)fl->fl_owner)->host;
+		lockhost = ((struct nlm_lockowner *) fl->c.flc_owner)->host;
 		if (match(lockhost, host)) {
 
 			spin_unlock(&flctx->flc_lock);
@@ -272,7 +272,7 @@ nlm_file_inuse(struct nlm_file *file)
 
 	if (flctx && !list_empty_careful(&flctx->flc_posix)) {
 		spin_lock(&flctx->flc_lock);
-		list_for_each_entry(fl, &flctx->flc_posix, fl_list) {
+		for_each_file_lock(fl, &flctx->flc_posix) {
 			if (fl->fl_lmops == &nlmsvc_lock_operations) {
 				spin_unlock(&flctx->flc_lock);
 				return 1;
diff --git a/fs/lockd/xdr.c b/fs/lockd/xdr.c
index 2fb5748dae..adfcce2bf1 100644
--- a/fs/lockd/xdr.c
+++ b/fs/lockd/xdr.c
@@ -88,8 +88,8 @@ svcxdr_decode_lock(struct xdr_stream *xdr, struct nlm_lock *lock)
 		return false;
 
 	locks_init_lock(fl);
-	fl->fl_flags = FL_POSIX;
-	fl->fl_type  = F_RDLCK;
+	fl->c.flc_flags = FL_POSIX;
+	fl->c.flc_type  = F_RDLCK;
 	end = start + len - 1;
 	fl->fl_start = s32_to_loff_t(start);
 	if (len == 0 || end < 0)
@@ -107,7 +107,7 @@ svcxdr_encode_holder(struct xdr_stream *xdr, const struct nlm_lock *lock)
 	s32 start, len;
 
 	/* exclusive */
-	if (xdr_stream_encode_bool(xdr, fl->fl_type != F_RDLCK) < 0)
+	if (xdr_stream_encode_bool(xdr, fl->c.flc_type != F_RDLCK) < 0)
 		return false;
 	if (xdr_stream_encode_u32(xdr, lock->svid) < 0)
 		return false;
@@ -164,7 +164,7 @@ nlmsvc_decode_testargs(struct svc_rqst *rqstp, struct xdr_stream *xdr)
 	if (!svcxdr_decode_lock(xdr, &argp->lock))
 		return false;
 	if (exclusive)
-		argp->lock.fl.fl_type = F_WRLCK;
+		argp->lock.fl.c.flc_type = F_WRLCK;
 
 	return true;
 }
@@ -184,7 +184,7 @@ nlmsvc_decode_lockargs(struct svc_rqst *rqstp, struct xdr_stream *xdr)
 	if (!svcxdr_decode_lock(xdr, &argp->lock))
 		return false;
 	if (exclusive)
-		argp->lock.fl.fl_type = F_WRLCK;
+		argp->lock.fl.c.flc_type = F_WRLCK;
 	if (xdr_stream_decode_bool(xdr, &argp->reclaim) < 0)
 		return false;
 	if (xdr_stream_decode_u32(xdr, &argp->state) < 0)
@@ -209,7 +209,7 @@ nlmsvc_decode_cancargs(struct svc_rqst *rqstp, struct xdr_stream *xdr)
 	if (!svcxdr_decode_lock(xdr, &argp->lock))
 		return false;
 	if (exclusive)
-		argp->lock.fl.fl_type = F_WRLCK;
+		argp->lock.fl.c.flc_type = F_WRLCK;
 
 	return true;
 }
@@ -223,7 +223,7 @@ nlmsvc_decode_unlockargs(struct svc_rqst *rqstp, struct xdr_stream *xdr)
 		return false;
 	if (!svcxdr_decode_lock(xdr, &argp->lock))
 		return false;
-	argp->lock.fl.fl_type = F_UNLCK;
+	argp->lock.fl.c.flc_type = F_UNLCK;
 
 	return true;
 }
diff --git a/fs/lockd/xdr4.c b/fs/lockd/xdr4.c
index 5fcbf30cd2..3d28b9c3ed 100644
--- a/fs/lockd/xdr4.c
+++ b/fs/lockd/xdr4.c
@@ -89,8 +89,8 @@ svcxdr_decode_lock(struct xdr_stream *xdr, struct nlm_lock *lock)
 		return false;
 
 	locks_init_lock(fl);
-	fl->fl_flags = FL_POSIX;
-	fl->fl_type  = F_RDLCK;
+	fl->c.flc_flags = FL_POSIX;
+	fl->c.flc_type  = F_RDLCK;
 	nlm4svc_set_file_lock_range(fl, lock->lock_start, lock->lock_len);
 	return true;
 }
@@ -102,7 +102,7 @@ svcxdr_encode_holder(struct xdr_stream *xdr, const struct nlm_lock *lock)
 	s64 start, len;
 
 	/* exclusive */
-	if (xdr_stream_encode_bool(xdr, fl->fl_type != F_RDLCK) < 0)
+	if (xdr_stream_encode_bool(xdr, fl->c.flc_type != F_RDLCK) < 0)
 		return false;
 	if (xdr_stream_encode_u32(xdr, lock->svid) < 0)
 		return false;
@@ -159,7 +159,7 @@ nlm4svc_decode_testargs(struct svc_rqst *rqstp, struct xdr_stream *xdr)
 	if (!svcxdr_decode_lock(xdr, &argp->lock))
 		return false;
 	if (exclusive)
-		argp->lock.fl.fl_type = F_WRLCK;
+		argp->lock.fl.c.flc_type = F_WRLCK;
 
 	return true;
 }
@@ -179,7 +179,7 @@ nlm4svc_decode_lockargs(struct svc_rqst *rqstp, struct xdr_stream *xdr)
 	if (!svcxdr_decode_lock(xdr, &argp->lock))
 		return false;
 	if (exclusive)
-		argp->lock.fl.fl_type = F_WRLCK;
+		argp->lock.fl.c.flc_type = F_WRLCK;
 	if (xdr_stream_decode_bool(xdr, &argp->reclaim) < 0)
 		return false;
 	if (xdr_stream_decode_u32(xdr, &argp->state) < 0)
@@ -204,7 +204,7 @@ nlm4svc_decode_cancargs(struct svc_rqst *rqstp, struct xdr_stream *xdr)
 	if (!svcxdr_decode_lock(xdr, &argp->lock))
 		return false;
 	if (exclusive)
-		argp->lock.fl.fl_type = F_WRLCK;
+		argp->lock.fl.c.flc_type = F_WRLCK;
 
 	return true;
 }
@@ -218,7 +218,7 @@ nlm4svc_decode_unlockargs(struct svc_rqst *rqstp, struct xdr_stream *xdr)
 		return false;
 	if (!svcxdr_decode_lock(xdr, &argp->lock))
 		return false;
-	argp->lock.fl.fl_type = F_UNLCK;
+	argp->lock.fl.c.flc_type = F_UNLCK;
 
 	return true;
 }
diff --git a/fs/locks.c b/fs/locks.c
index cc7c117ee1..9afb16e068 100644
--- a/fs/locks.c
+++ b/fs/locks.c
@@ -48,7 +48,6 @@
  * children.
  *
  */
-
 #include <linux/capability.h>
 #include <linux/file.h>
 #include <linux/fdtable.h>
@@ -70,24 +69,28 @@
 
 #include <linux/uaccess.h>
 
-#define IS_POSIX(fl)	(fl->fl_flags & FL_POSIX)
-#define IS_FLOCK(fl)	(fl->fl_flags & FL_FLOCK)
-#define IS_LEASE(fl)	(fl->fl_flags & (FL_LEASE|FL_DELEG|FL_LAYOUT))
-#define IS_OFDLCK(fl)	(fl->fl_flags & FL_OFDLCK)
-#define IS_REMOTELCK(fl)	(fl->fl_pid <= 0)
+static struct file_lock *file_lock(struct file_lock_core *flc)
+{
+	return container_of(flc, struct file_lock, c);
+}
+
+static struct file_lease *file_lease(struct file_lock_core *flc)
+{
+	return container_of(flc, struct file_lease, c);
+}
 
-static bool lease_breaking(struct file_lock *fl)
+static bool lease_breaking(struct file_lease *fl)
 {
-	return fl->fl_flags & (FL_UNLOCK_PENDING | FL_DOWNGRADE_PENDING);
+	return fl->c.flc_flags & (FL_UNLOCK_PENDING | FL_DOWNGRADE_PENDING);
 }
 
-static int target_leasetype(struct file_lock *fl)
+static int target_leasetype(struct file_lease *fl)
 {
-	if (fl->fl_flags & FL_UNLOCK_PENDING)
+	if (fl->c.flc_flags & FL_UNLOCK_PENDING)
 		return F_UNLCK;
-	if (fl->fl_flags & FL_DOWNGRADE_PENDING)
+	if (fl->c.flc_flags & FL_DOWNGRADE_PENDING)
 		return F_RDLCK;
-	return fl->fl_type;
+	return fl->c.flc_type;
 }
 
 static int leases_enable = 1;
@@ -168,6 +171,7 @@ static DEFINE_SPINLOCK(blocked_lock_lock);
 
 static struct kmem_cache *flctx_cache __ro_after_init;
 static struct kmem_cache *filelock_cache __ro_after_init;
+static struct kmem_cache *filelease_cache __ro_after_init;
 
 static struct file_lock_context *
 locks_get_lock_context(struct inode *inode, int type)
@@ -204,11 +208,12 @@ out:
 static void
 locks_dump_ctx_list(struct list_head *list, char *list_type)
 {
-	struct file_lock *fl;
+	struct file_lock_core *flc;
 
-	list_for_each_entry(fl, list, fl_list) {
-		pr_warn("%s: fl_owner=%p fl_flags=0x%x fl_type=0x%x fl_pid=%u\n", list_type, fl->fl_owner, fl->fl_flags, fl->fl_type, fl->fl_pid);
-	}
+	list_for_each_entry(flc, list, flc_list)
+		pr_warn("%s: fl_owner=%p fl_flags=0x%x fl_type=0x%x fl_pid=%u\n",
+			list_type, flc->flc_owner, flc->flc_flags,
+			flc->flc_type, flc->flc_pid);
 }
 
 static void
@@ -229,19 +234,19 @@ locks_check_ctx_lists(struct inode *inode)
 }
 
 static void
-locks_check_ctx_file_list(struct file *filp, struct list_head *list,
-				char *list_type)
+locks_check_ctx_file_list(struct file *filp, struct list_head *list, char *list_type)
 {
-	struct file_lock *fl;
+	struct file_lock_core *flc;
 	struct inode *inode = file_inode(filp);
 
-	list_for_each_entry(fl, list, fl_list)
-		if (fl->fl_file == filp)
+	list_for_each_entry(flc, list, flc_list)
+		if (flc->flc_file == filp)
 			pr_warn("Leaked %s lock on dev=0x%x:0x%x ino=0x%lx "
 				" fl_owner=%p fl_flags=0x%x fl_type=0x%x fl_pid=%u\n",
 				list_type, MAJOR(inode->i_sb->s_dev),
 				MINOR(inode->i_sb->s_dev), inode->i_ino,
-				fl->fl_owner, fl->fl_flags, fl->fl_type, fl->fl_pid);
+				flc->flc_owner, flc->flc_flags,
+				flc->flc_type, flc->flc_pid);
 }
 
 void
@@ -255,13 +260,13 @@ locks_free_lock_context(struct inode *inode)
 	}
 }
 
-static void locks_init_lock_heads(struct file_lock *fl)
+static void locks_init_lock_heads(struct file_lock_core *flc)
 {
-	INIT_HLIST_NODE(&fl->fl_link);
-	INIT_LIST_HEAD(&fl->fl_list);
-	INIT_LIST_HEAD(&fl->fl_blocked_requests);
-	INIT_LIST_HEAD(&fl->fl_blocked_member);
-	init_waitqueue_head(&fl->fl_wait);
+	INIT_HLIST_NODE(&flc->flc_link);
+	INIT_LIST_HEAD(&flc->flc_list);
+	INIT_LIST_HEAD(&flc->flc_blocked_requests);
+	INIT_LIST_HEAD(&flc->flc_blocked_member);
+	init_waitqueue_head(&flc->flc_wait);
 }
 
 /* Allocate an empty lock structure. */
@@ -270,19 +275,33 @@ struct file_lock *locks_alloc_lock(void)
 	struct file_lock *fl = kmem_cache_zalloc(filelock_cache, GFP_KERNEL);
 
 	if (fl)
-		locks_init_lock_heads(fl);
+		locks_init_lock_heads(&fl->c);
 
 	return fl;
 }
 EXPORT_SYMBOL_GPL(locks_alloc_lock);
 
+/* Allocate an empty lock structure. */
+struct file_lease *locks_alloc_lease(void)
+{
+	struct file_lease *fl = kmem_cache_zalloc(filelease_cache, GFP_KERNEL);
+
+	if (fl)
+		locks_init_lock_heads(&fl->c);
+
+	return fl;
+}
+EXPORT_SYMBOL_GPL(locks_alloc_lease);
+
 void locks_release_private(struct file_lock *fl)
 {
-	BUG_ON(waitqueue_active(&fl->fl_wait));
-	BUG_ON(!list_empty(&fl->fl_list));
-	BUG_ON(!list_empty(&fl->fl_blocked_requests));
-	BUG_ON(!list_empty(&fl->fl_blocked_member));
-	BUG_ON(!hlist_unhashed(&fl->fl_link));
+	struct file_lock_core *flc = &fl->c;
+
+	BUG_ON(waitqueue_active(&flc->flc_wait));
+	BUG_ON(!list_empty(&flc->flc_list));
+	BUG_ON(!list_empty(&flc->flc_blocked_requests));
+	BUG_ON(!list_empty(&flc->flc_blocked_member));
+	BUG_ON(!hlist_unhashed(&flc->flc_link));
 
 	if (fl->fl_ops) {
 		if (fl->fl_ops->fl_release_private)
@@ -292,8 +311,8 @@ void locks_release_private(struct file_lock *fl)
 
 	if (fl->fl_lmops) {
 		if (fl->fl_lmops->lm_put_owner) {
-			fl->fl_lmops->lm_put_owner(fl->fl_owner);
-			fl->fl_owner = NULL;
+			fl->fl_lmops->lm_put_owner(flc->flc_owner);
+			flc->flc_owner = NULL;
 		}
 		fl->fl_lmops = NULL;
 	}
@@ -309,16 +328,15 @@ EXPORT_SYMBOL_GPL(locks_release_private);
  *   %true: @owner has at least one blocker
  *   %false: @owner has no blockers
  */
-bool locks_owner_has_blockers(struct file_lock_context *flctx,
-		fl_owner_t owner)
+bool locks_owner_has_blockers(struct file_lock_context *flctx, fl_owner_t owner)
 {
-	struct file_lock *fl;
+	struct file_lock_core *flc;
 
 	spin_lock(&flctx->flc_lock);
-	list_for_each_entry(fl, &flctx->flc_posix, fl_list) {
-		if (fl->fl_owner != owner)
+	list_for_each_entry(flc, &flctx->flc_posix, flc_list) {
+		if (flc->flc_owner != owner)
 			continue;
-		if (!list_empty(&fl->fl_blocked_requests)) {
+		if (!list_empty(&flc->flc_blocked_requests)) {
 			spin_unlock(&flctx->flc_lock);
 			return true;
 		}
@@ -336,35 +354,52 @@ void locks_free_lock(struct file_lock *fl)
 }
 EXPORT_SYMBOL(locks_free_lock);
 
+/* Free a lease which is not in use. */
+void locks_free_lease(struct file_lease *fl)
+{
+	kmem_cache_free(filelease_cache, fl);
+}
+EXPORT_SYMBOL(locks_free_lease);
+
 static void
 locks_dispose_list(struct list_head *dispose)
 {
-	struct file_lock *fl;
+	struct file_lock_core *flc;
 
 	while (!list_empty(dispose)) {
-		fl = list_first_entry(dispose, struct file_lock, fl_list);
-		list_del_init(&fl->fl_list);
-		locks_free_lock(fl);
+		flc = list_first_entry(dispose, struct file_lock_core, flc_list);
+		list_del_init(&flc->flc_list);
+		if (flc->flc_flags & (FL_LEASE|FL_DELEG|FL_LAYOUT))
+			locks_free_lease(file_lease(flc));
+		else
+			locks_free_lock(file_lock(flc));
 	}
 }
 
 void locks_init_lock(struct file_lock *fl)
 {
 	memset(fl, 0, sizeof(struct file_lock));
-	locks_init_lock_heads(fl);
+	locks_init_lock_heads(&fl->c);
 }
 EXPORT_SYMBOL(locks_init_lock);
 
+void locks_init_lease(struct file_lease *fl)
+{
+	memset(fl, 0, sizeof(*fl));
+	locks_init_lock_heads(&fl->c);
+}
+EXPORT_SYMBOL(locks_init_lease);
+
 /*
  * Initialize a new lock from an existing file_lock structure.
  */
 void locks_copy_conflock(struct file_lock *new, struct file_lock *fl)
 {
-	new->fl_owner = fl->fl_owner;
-	new->fl_pid = fl->fl_pid;
-	new->fl_file = NULL;
-	new->fl_flags = fl->fl_flags;
-	new->fl_type = fl->fl_type;
+	new->c.flc_owner = fl->c.flc_owner;
+	new->c.flc_pid = fl->c.flc_pid;
+	new->c.flc_file = NULL;
+	new->c.flc_flags = fl->c.flc_flags;
+	new->c.flc_type = fl->c.flc_type;
 	new->fl_start = fl->fl_start;
 	new->fl_end = fl->fl_end;
 	new->fl_lmops = fl->fl_lmops;
@@ -372,7 +407,7 @@ void locks_copy_conflock(struct file_lock *new, struct file_lock *fl)
 
 	if (fl->fl_lmops) {
 		if (fl->fl_lmops->lm_get_owner)
-			fl->fl_lmops->lm_get_owner(fl->fl_owner);
+			fl->fl_lmops->lm_get_owner(fl->c.flc_owner);
 	}
 }
 EXPORT_SYMBOL(locks_copy_conflock);
@@ -384,7 +419,7 @@ void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
 
 	locks_copy_conflock(new, fl);
 
-	new->fl_file = fl->fl_file;
+	new->c.flc_file = fl->c.flc_file;
 	new->fl_ops = fl->fl_ops;
 
 	if (fl->fl_ops) {
@@ -400,15 +435,17 @@ static void locks_move_blocks(struct file_lock *new, struct file_lock *fl)
 
 	/*
 	 * As ctx->flc_lock is held, new requests cannot be added to
-	 * ->fl_blocked_requests, so we don't need a lock to check if it
+	 * ->flc_blocked_requests, so we don't need a lock to check if it
 	 * is empty.
 	 */
-	if (list_empty(&fl->fl_blocked_requests))
+	if (list_empty(&fl->c.flc_blocked_requests))
 		return;
 	spin_lock(&blocked_lock_lock);
-	list_splice_init(&fl->fl_blocked_requests, &new->fl_blocked_requests);
-	list_for_each_entry(f, &new->fl_blocked_requests, fl_blocked_member)
-		f->fl_blocker = new;
+	list_splice_init(&fl->c.flc_blocked_requests,
+			 &new->c.flc_blocked_requests);
+	list_for_each_entry(f, &new->c.flc_blocked_requests,
+			    c.flc_blocked_member)
+		f->c.flc_blocker = &new->c;
 	spin_unlock(&blocked_lock_lock);
 }
 
@@ -429,21 +466,21 @@ static void flock_make_lock(struct file *filp, struct file_lock *fl, int type)
 {
 	locks_init_lock(fl);
 
-	fl->fl_file = filp;
-	fl->fl_owner = filp;
-	fl->fl_pid = current->tgid;
-	fl->fl_flags = FL_FLOCK;
-	fl->fl_type = type;
+	fl->c.flc_file = filp;
+	fl->c.flc_owner = filp;
+	fl->c.flc_pid = current->tgid;
+	fl->c.flc_flags = FL_FLOCK;
+	fl->c.flc_type = type;
 	fl->fl_end = OFFSET_MAX;
 }
 
-static int assign_type(struct file_lock *fl, int type)
+static int assign_type(struct file_lock_core *flc, int type)
 {
 	switch (type) {
 	case F_RDLCK:
 	case F_WRLCK:
 	case F_UNLCK:
-		fl->fl_type = type;
+		flc->flc_type = type;
 		break;
 	default:
 		return -EINVAL;
@@ -488,14 +525,14 @@ static int flock64_to_posix_lock(struct file *filp, struct file_lock *fl,
 	} else
 		fl->fl_end = OFFSET_MAX;
 
-	fl->fl_owner = current->files;
-	fl->fl_pid = current->tgid;
-	fl->fl_file = filp;
-	fl->fl_flags = FL_POSIX;
+	fl->c.flc_owner = current->files;
+	fl->c.flc_pid = current->tgid;
+	fl->c.flc_file = filp;
+	fl->c.flc_flags = FL_POSIX;
 	fl->fl_ops = NULL;
 	fl->fl_lmops = NULL;
 
-	return assign_type(fl, l->l_type);
+	return assign_type(&fl->c, l->l_type);
 }
 
 /* Verify a "struct flock" and copy it to a "struct file_lock" as a POSIX
@@ -516,16 +553,16 @@ static int flock_to_posix_lock(struct file *filp, struct file_lock *fl,
 
 /* default lease lock manager operations */
 static bool
-lease_break_callback(struct file_lock *fl)
+lease_break_callback(struct file_lease *fl)
 {
 	kill_fasync(&fl->fl_fasync, SIGIO, POLL_MSG);
 	return false;
 }
 
 static void
-lease_setup(struct file_lock *fl, void **priv)
+lease_setup(struct file_lease *fl, void **priv)
 {
-	struct file *filp = fl->fl_file;
+	struct file *filp = fl->c.flc_file;
 	struct fasync_struct *fa = *priv;
 
 	/*
@@ -539,7 +576,7 @@ lease_setup(struct file_lock *fl, void **priv)
 	__f_setown(filp, task_pid(current), PIDTYPE_TGID, 0);
 }
 
-static const struct lock_manager_operations lease_manager_ops = {
+static const struct lease_manager_operations lease_manager_ops = {
 	.lm_break = lease_break_callback,
 	.lm_change = lease_modify,
 	.lm_setup = lease_setup,
@@ -548,27 +585,24 @@ static const struct lock_manager_operations lease_manager_ops = {
 /*
  * Initialize a lease, use the default lock manager operations
  */
-static int lease_init(struct file *filp, int type, struct file_lock *fl)
+static int lease_init(struct file *filp, int type, struct file_lease *fl)
 {
-	if (assign_type(fl, type) != 0)
+	if (assign_type(&fl->c, type) != 0)
 		return -EINVAL;
 
-	fl->fl_owner = filp;
-	fl->fl_pid = current->tgid;
+	fl->c.flc_owner = filp;
+	fl->c.flc_pid = current->tgid;
 
-	fl->fl_file = filp;
-	fl->fl_flags = FL_LEASE;
-	fl->fl_start = 0;
-	fl->fl_end = OFFSET_MAX;
-	fl->fl_ops = NULL;
+	fl->c.flc_file = filp;
+	fl->c.flc_flags = FL_LEASE;
 	fl->fl_lmops = &lease_manager_ops;
 	return 0;
 }
 
 /* Allocate a file_lock initialised to this type of lease */
-static struct file_lock *lease_alloc(struct file *filp, int type)
+static struct file_lease *lease_alloc(struct file *filp, int type)
 {
-	struct file_lock *fl = locks_alloc_lock();
+	struct file_lease *fl = locks_alloc_lease();
 	int error = -ENOMEM;
 
 	if (fl == NULL)
@@ -576,7 +610,7 @@ static struct file_lock *lease_alloc(struct file *filp, int type)
 
 	error = lease_init(filp, type, fl);
 	if (error) {
-		locks_free_lock(fl);
+		locks_free_lease(fl);
 		return ERR_PTR(error);
 	}
 	return fl;
@@ -593,26 +627,26 @@ static inline int locks_overlap(struct file_lock *fl1, struct file_lock *fl2)
 /*
  * Check whether two locks have the same owner.
  */
-static int posix_same_owner(struct file_lock *fl1, struct file_lock *fl2)
+static int posix_same_owner(struct file_lock_core *fl1, struct file_lock_core *fl2)
 {
-	return fl1->fl_owner == fl2->fl_owner;
+	return fl1->flc_owner == fl2->flc_owner;
 }
 
 /* Must be called with the flc_lock held! */
-static void locks_insert_global_locks(struct file_lock *fl)
+static void locks_insert_global_locks(struct file_lock_core *flc)
 {
 	struct file_lock_list_struct *fll = this_cpu_ptr(&file_lock_list);
 
 	percpu_rwsem_assert_held(&file_rwsem);
 
 	spin_lock(&fll->lock);
-	fl->fl_link_cpu = smp_processor_id();
-	hlist_add_head(&fl->fl_link, &fll->hlist);
+	flc->flc_link_cpu = smp_processor_id();
+	hlist_add_head(&flc->flc_link, &fll->hlist);
 	spin_unlock(&fll->lock);
 }
 
 /* Must be called with the flc_lock held! */
-static void locks_delete_global_locks(struct file_lock *fl)
+static void locks_delete_global_locks(struct file_lock_core *flc)
 {
 	struct file_lock_list_struct *fll;
 
@@ -623,33 +657,33 @@ static void locks_delete_global_locks(struct file_lock *fl)
 	 * is done while holding the flc_lock, and new insertions into the list
 	 * also require that it be held.
 	 */
-	if (hlist_unhashed(&fl->fl_link))
+	if (hlist_unhashed(&flc->flc_link))
 		return;
 
-	fll = per_cpu_ptr(&file_lock_list, fl->fl_link_cpu);
+	fll = per_cpu_ptr(&file_lock_list, flc->flc_link_cpu);
 	spin_lock(&fll->lock);
-	hlist_del_init(&fl->fl_link);
+	hlist_del_init(&flc->flc_link);
 	spin_unlock(&fll->lock);
 }
 
 static unsigned long
-posix_owner_key(struct file_lock *fl)
+posix_owner_key(struct file_lock_core *flc)
 {
-	return (unsigned long)fl->fl_owner;
+	return (unsigned long) flc->flc_owner;
 }
 
-static void locks_insert_global_blocked(struct file_lock *waiter)
+static void locks_insert_global_blocked(struct file_lock_core *waiter)
 {
 	lockdep_assert_held(&blocked_lock_lock);
 
-	hash_add(blocked_hash, &waiter->fl_link, posix_owner_key(waiter));
+	hash_add(blocked_hash, &waiter->flc_link, posix_owner_key(waiter));
 }
 
-static void locks_delete_global_blocked(struct file_lock *waiter)
+static void locks_delete_global_blocked(struct file_lock_core *waiter)
 {
 	lockdep_assert_held(&blocked_lock_lock);
 
-	hash_del(&waiter->fl_link);
+	hash_del(&waiter->flc_link);
 }
 
 /* Remove waiter from blocker's block list.
@@ -657,41 +691,39 @@ static void locks_delete_global_blocked(struct file_lock *waiter)
  *
  * Must be called with blocked_lock_lock held.
  */
-static void __locks_delete_block(struct file_lock *waiter)
+static void __locks_unlink_block(struct file_lock_core *waiter)
 {
 	locks_delete_global_blocked(waiter);
-	list_del_init(&waiter->fl_blocked_member);
+	list_del_init(&waiter->flc_blocked_member);
 }
 
-static void __locks_wake_up_blocks(struct file_lock *blocker)
+static void __locks_wake_up_blocks(struct file_lock_core *blocker)
 {
-	while (!list_empty(&blocker->fl_blocked_requests)) {
-		struct file_lock *waiter;
+	while (!list_empty(&blocker->flc_blocked_requests)) {
+		struct file_lock_core *waiter;
+		struct file_lock *fl;
 
-		waiter = list_first_entry(&blocker->fl_blocked_requests,
-					  struct file_lock, fl_blocked_member);
-		__locks_delete_block(waiter);
-		if (waiter->fl_lmops && waiter->fl_lmops->lm_notify)
-			waiter->fl_lmops->lm_notify(waiter);
+		waiter = list_first_entry(&blocker->flc_blocked_requests,
+					  struct file_lock_core, flc_blocked_member);
+
+		fl = file_lock(waiter);
+		__locks_unlink_block(waiter);
+		if ((waiter->flc_flags & (FL_POSIX | FL_FLOCK)) &&
+		    fl->fl_lmops && fl->fl_lmops->lm_notify)
+			fl->fl_lmops->lm_notify(fl);
 		else
-			wake_up(&waiter->fl_wait);
+			locks_wake_up(fl);
 
 		/*
-		 * The setting of fl_blocker to NULL marks the "done"
+		 * The setting of flc_blocker to NULL marks the "done"
 		 * point in deleting a block. Paired with acquire at the top
 		 * of locks_delete_block().
 		 */
-		smp_store_release(&waiter->fl_blocker, NULL);
+		smp_store_release(&waiter->flc_blocker, NULL);
 	}
 }
 
-/**
- *	locks_delete_block - stop waiting for a file lock
- *	@waiter: the lock which was waiting
- *
- *	lockd/nfsd need to disconnect the lock while working on it.
- */
-int locks_delete_block(struct file_lock *waiter)
+static int __locks_delete_block(struct file_lock_core *waiter)
 {
 	int status = -ENOENT;
 
@@ -716,24 +748,35 @@ int locks_delete_block(struct file_lock *waiter)
 	 * no new locks can be inserted into its fl_blocked_requests list, and
 	 * can avoid doing anything further if the list is empty.
 	 */
-	if (!smp_load_acquire(&waiter->fl_blocker) &&
-	    list_empty(&waiter->fl_blocked_requests))
+	if (!smp_load_acquire(&waiter->flc_blocker) &&
+	    list_empty(&waiter->flc_blocked_requests))
 		return status;
 
 	spin_lock(&blocked_lock_lock);
-	if (waiter->fl_blocker)
+	if (waiter->flc_blocker)
 		status = 0;
 	__locks_wake_up_blocks(waiter);
-	__locks_delete_block(waiter);
+	__locks_unlink_block(waiter);
 
 	/*
 	 * The setting of fl_blocker to NULL marks the "done" point in deleting
 	 * a block. Paired with acquire at the top of this function.
 	 */
-	smp_store_release(&waiter->fl_blocker, NULL);
+	smp_store_release(&waiter->flc_blocker, NULL);
 	spin_unlock(&blocked_lock_lock);
 	return status;
 }
+
+/**
+ *	locks_delete_block - stop waiting for a file lock
+ *	@waiter: the lock which was waiting
+ *
+ *	lockd/nfsd need to disconnect the lock while working on it.
+ */
+int locks_delete_block(struct file_lock *waiter)
+{
+	return __locks_delete_block(&waiter->c);
+}
 EXPORT_SYMBOL(locks_delete_block);
 
 /* Insert waiter into blocker's block list.
@@ -751,26 +794,28 @@ EXPORT_SYMBOL(locks_delete_block);
  * waiters, and add beneath any waiter that blocks the new waiter.
  * Thus wakeups don't happen until needed.
  */
-static void __locks_insert_block(struct file_lock *blocker,
-				 struct file_lock *waiter,
-				 bool conflict(struct file_lock *,
-					       struct file_lock *))
+static void __locks_insert_block(struct file_lock_core *blocker,
+				 struct file_lock_core *waiter,
+				 bool conflict(struct file_lock_core *,
+					       struct file_lock_core *))
 {
-	struct file_lock *fl;
-	BUG_ON(!list_empty(&waiter->fl_blocked_member));
+	struct file_lock_core *flc;
 
+	BUG_ON(!list_empty(&waiter->flc_blocked_member));
 new_blocker:
-	list_for_each_entry(fl, &blocker->fl_blocked_requests, fl_blocked_member)
-		if (conflict(fl, waiter)) {
-			blocker =  fl;
+	list_for_each_entry(flc, &blocker->flc_blocked_requests, flc_blocked_member)
+		if (conflict(flc, waiter)) {
+			blocker =  flc;
 			goto new_blocker;
 		}
-	waiter->fl_blocker = blocker;
-	list_add_tail(&waiter->fl_blocked_member, &blocker->fl_blocked_requests);
-	if (IS_POSIX(blocker) && !IS_OFDLCK(blocker))
+	waiter->flc_blocker = blocker;
+	list_add_tail(&waiter->flc_blocked_member,
+		      &blocker->flc_blocked_requests);
+
+	if ((blocker->flc_flags & (FL_POSIX|FL_OFDLCK)) == FL_POSIX)
 		locks_insert_global_blocked(waiter);
 
-	/* The requests in waiter->fl_blocked are known to conflict with
+	/* The requests in waiter->flc_blocked are known to conflict with
 	 * waiter, but might not conflict with blocker, or the requests
 	 * and lock which block it.  So they all need to be woken.
 	 */
@@ -778,10 +823,10 @@ new_blocker:
 }
 
 /* Must be called with flc_lock held. */
-static void locks_insert_block(struct file_lock *blocker,
-			       struct file_lock *waiter,
-			       bool conflict(struct file_lock *,
-					     struct file_lock *))
+static void locks_insert_block(struct file_lock_core *blocker,
+			       struct file_lock_core *waiter,
+			       bool conflict(struct file_lock_core *,
+					     struct file_lock_core *))
 {
 	spin_lock(&blocked_lock_lock);
 	__locks_insert_block(blocker, waiter, conflict);
@@ -793,7 +838,7 @@ static void locks_insert_block(struct file_lock *blocker,
  *
  * Must be called with the inode->flc_lock held!
  */
-static void locks_wake_up_blocks(struct file_lock *blocker)
+static void locks_wake_up_blocks(struct file_lock_core *blocker)
 {
 	/*
 	 * Avoid taking global lock if list is empty. This is safe since new
@@ -802,7 +847,7 @@ static void locks_wake_up_blocks(struct file_lock *blocker)
 	 * fl_blocked_requests list does not require the flc_lock, so we must
 	 * recheck list_empty() after acquiring the blocked_lock_lock.
 	 */
-	if (list_empty(&blocker->fl_blocked_requests))
+	if (list_empty(&blocker->flc_blocked_requests))
 		return;
 
 	spin_lock(&blocked_lock_lock);
@@ -811,39 +856,39 @@ static void locks_wake_up_blocks(struct file_lock *blocker)
 }
 
 static void
-locks_insert_lock_ctx(struct file_lock *fl, struct list_head *before)
+locks_insert_lock_ctx(struct file_lock_core *fl, struct list_head *before)
 {
-	list_add_tail(&fl->fl_list, before);
+	list_add_tail(&fl->flc_list, before);
 	locks_insert_global_locks(fl);
 }
 
 static void
-locks_unlink_lock_ctx(struct file_lock *fl)
+locks_unlink_lock_ctx(struct file_lock_core *fl)
 {
 	locks_delete_global_locks(fl);
-	list_del_init(&fl->fl_list);
+	list_del_init(&fl->flc_list);
 	locks_wake_up_blocks(fl);
 }
 
 static void
-locks_delete_lock_ctx(struct file_lock *fl, struct list_head *dispose)
+locks_delete_lock_ctx(struct file_lock_core *fl, struct list_head *dispose)
 {
 	locks_unlink_lock_ctx(fl);
 	if (dispose)
-		list_add(&fl->fl_list, dispose);
+		list_add(&fl->flc_list, dispose);
 	else
-		locks_free_lock(fl);
+		locks_free_lock(file_lock(fl));
 }
 
 /* Determine if lock sys_fl blocks lock caller_fl. Common functionality
  * checks for shared/exclusive status of overlapping locks.
  */
-static bool locks_conflict(struct file_lock *caller_fl,
-			   struct file_lock *sys_fl)
+static bool locks_conflict(struct file_lock_core *caller_flc,
+			   struct file_lock_core *sys_flc)
 {
-	if (sys_fl->fl_type == F_WRLCK)
+	if (sys_flc->flc_type == F_WRLCK)
 		return true;
-	if (caller_fl->fl_type == F_WRLCK)
+	if (caller_flc->flc_type == F_WRLCK)
 		return true;
 	return false;
 }
@@ -851,20 +896,23 @@ static bool locks_conflict(struct file_lock *caller_fl,
 /* Determine if lock sys_fl blocks lock caller_fl. POSIX specific
  * checking before calling the locks_conflict().
  */
-static bool posix_locks_conflict(struct file_lock *caller_fl,
-				 struct file_lock *sys_fl)
+static bool posix_locks_conflict(struct file_lock_core *caller_flc,
+				 struct file_lock_core *sys_flc)
 {
+	struct file_lock *caller_fl = file_lock(caller_flc);
+	struct file_lock *sys_fl = file_lock(sys_flc);
+
 	/* POSIX locks owned by the same process do not conflict with
 	 * each other.
 	 */
-	if (posix_same_owner(caller_fl, sys_fl))
+	if (posix_same_owner(caller_flc, sys_flc))
 		return false;
 
 	/* Check whether they overlap */
 	if (!locks_overlap(caller_fl, sys_fl))
 		return false;
 
-	return locks_conflict(caller_fl, sys_fl);
+	return locks_conflict(caller_flc, sys_flc);
 }
 
 /* Determine if lock sys_fl blocks lock caller_fl. Used on xx_GETLK
@@ -873,28 +921,31 @@ static bool posix_locks_conflict(struct file_lock *caller_fl,
 static bool posix_test_locks_conflict(struct file_lock *caller_fl,
 				      struct file_lock *sys_fl)
 {
+	struct file_lock_core *caller = &caller_fl->c;
+	struct file_lock_core *sys = &sys_fl->c;
+
 	/* F_UNLCK checks any locks on the same fd. */
-	if (caller_fl->fl_type == F_UNLCK) {
-		if (!posix_same_owner(caller_fl, sys_fl))
+	if (lock_is_unlock(caller_fl)) {
+		if (!posix_same_owner(caller, sys))
 			return false;
 		return locks_overlap(caller_fl, sys_fl);
 	}
-	return posix_locks_conflict(caller_fl, sys_fl);
+	return posix_locks_conflict(caller, sys);
 }
 
 /* Determine if lock sys_fl blocks lock caller_fl. FLOCK specific
  * checking before calling the locks_conflict().
  */
-static bool flock_locks_conflict(struct file_lock *caller_fl,
-				 struct file_lock *sys_fl)
+static bool flock_locks_conflict(struct file_lock_core *caller_flc,
+				 struct file_lock_core *sys_flc)
 {
 	/* FLOCK locks referring to the same filp do not conflict with
 	 * each other.
 	 */
-	if (caller_fl->fl_file == sys_fl->fl_file)
+	if (caller_flc->flc_file == sys_flc->flc_file)
 		return false;
 
-	return locks_conflict(caller_fl, sys_fl);
+	return locks_conflict(caller_flc, sys_flc);
 }
 
 void
@@ -908,13 +959,13 @@ posix_test_lock(struct file *filp, struct file_lock *fl)
 
 	ctx = locks_inode_context(inode);
 	if (!ctx || list_empty_careful(&ctx->flc_posix)) {
-		fl->fl_type = F_UNLCK;
+		fl->c.flc_type = F_UNLCK;
 		return;
 	}
 
 retry:
 	spin_lock(&ctx->flc_lock);
-	list_for_each_entry(cfl, &ctx->flc_posix, fl_list) {
+	list_for_each_entry(cfl, &ctx->flc_posix, c.flc_list) {
 		if (!posix_test_locks_conflict(fl, cfl))
 			continue;
 		if (cfl->fl_lmops && cfl->fl_lmops->lm_lock_expirable
@@ -930,7 +981,7 @@ retry:
 		locks_copy_conflock(fl, cfl);
 		goto out;
 	}
-	fl->fl_type = F_UNLCK;
+	fl->c.flc_type = F_UNLCK;
 out:
 	spin_unlock(&ctx->flc_lock);
 	return;
@@ -972,25 +1023,27 @@ EXPORT_SYMBOL(posix_test_lock);
 
 #define MAX_DEADLK_ITERATIONS 10
 
-/* Find a lock that the owner of the given block_fl is blocking on. */
-static struct file_lock *what_owner_is_waiting_for(struct file_lock *block_fl)
+/* Find a lock that the owner of the given @blocker is blocking on. */
+static struct file_lock_core *what_owner_is_waiting_for(struct file_lock_core *blocker)
 {
-	struct file_lock *fl;
+	struct file_lock_core *flc;
 
-	hash_for_each_possible(blocked_hash, fl, fl_link, posix_owner_key(block_fl)) {
-		if (posix_same_owner(fl, block_fl)) {
-			while (fl->fl_blocker)
-				fl = fl->fl_blocker;
-			return fl;
+	hash_for_each_possible(blocked_hash, flc, flc_link, posix_owner_key(blocker)) {
+		if (posix_same_owner(flc, blocker)) {
+			while (flc->flc_blocker)
+				flc = flc->flc_blocker;
+			return flc;
 		}
 	}
 	return NULL;
 }
 
 /* Must be called with the blocked_lock_lock held! */
-static int posix_locks_deadlock(struct file_lock *caller_fl,
-				struct file_lock *block_fl)
+static bool posix_locks_deadlock(struct file_lock *caller_fl,
+				 struct file_lock *block_fl)
 {
+	struct file_lock_core *caller = &caller_fl->c;
+	struct file_lock_core *blocker = &block_fl->c;
 	int i = 0;
 
 	lockdep_assert_held(&blocked_lock_lock);
@@ -999,16 +1052,16 @@ static int posix_locks_deadlock(struct file_lock *caller_fl,
 	 * This deadlock detector can't reasonably detect deadlocks with
 	 * FL_OFDLCK locks, since they aren't owned by a process, per-se.
 	 */
-	if (IS_OFDLCK(caller_fl))
-		return 0;
+	if (caller->flc_flags & FL_OFDLCK)
+		return false;
 
-	while ((block_fl = what_owner_is_waiting_for(block_fl))) {
+	while ((blocker = what_owner_is_waiting_for(blocker))) {
 		if (i++ > MAX_DEADLK_ITERATIONS)
-			return 0;
-		if (posix_same_owner(caller_fl, block_fl))
-			return 1;
+			return false;
+		if (posix_same_owner(caller, blocker))
+			return true;
 	}
-	return 0;
+	return false;
 }
 
 /* Try to create a FLOCK lock on filp. We always insert new FLOCK locks
@@ -1027,14 +1080,14 @@ static int flock_lock_inode(struct inode *inode, struct file_lock *request)
 	bool found = false;
 	LIST_HEAD(dispose);
 
-	ctx = locks_get_lock_context(inode, request->fl_type);
+	ctx = locks_get_lock_context(inode, request->c.flc_type);
 	if (!ctx) {
-		if (request->fl_type != F_UNLCK)
+		if (request->c.flc_type != F_UNLCK)
 			return -ENOMEM;
-		return (request->fl_flags & FL_EXISTS) ? -ENOENT : 0;
+		return (request->c.flc_flags & FL_EXISTS) ? -ENOENT : 0;
 	}
 
-	if (!(request->fl_flags & FL_ACCESS) && (request->fl_type != F_UNLCK)) {
+	if (!(request->c.flc_flags & FL_ACCESS) && (request->c.flc_type != F_UNLCK)) {
 		new_fl = locks_alloc_lock();
 		if (!new_fl)
 			return -ENOMEM;
@@ -1042,41 +1095,41 @@ static int flock_lock_inode(struct inode *inode, struct file_lock *request)
 
 	percpu_down_read(&file_rwsem);
 	spin_lock(&ctx->flc_lock);
-	if (request->fl_flags & FL_ACCESS)
+	if (request->c.flc_flags & FL_ACCESS)
 		goto find_conflict;
 
-	list_for_each_entry(fl, &ctx->flc_flock, fl_list) {
-		if (request->fl_file != fl->fl_file)
+	list_for_each_entry(fl, &ctx->flc_flock, c.flc_list) {
+		if (request->c.flc_file != fl->c.flc_file)
 			continue;
-		if (request->fl_type == fl->fl_type)
+		if (request->c.flc_type == fl->c.flc_type)
 			goto out;
 		found = true;
-		locks_delete_lock_ctx(fl, &dispose);
+		locks_delete_lock_ctx(&fl->c, &dispose);
 		break;
 	}
 
-	if (request->fl_type == F_UNLCK) {
-		if ((request->fl_flags & FL_EXISTS) && !found)
+	if (lock_is_unlock(request)) {
+		if ((request->c.flc_flags & FL_EXISTS) && !found)
 			error = -ENOENT;
 		goto out;
 	}
 
 find_conflict:
-	list_for_each_entry(fl, &ctx->flc_flock, fl_list) {
-		if (!flock_locks_conflict(request, fl))
+	list_for_each_entry(fl, &ctx->flc_flock, c.flc_list) {
+		if (!flock_locks_conflict(&request->c, &fl->c))
 			continue;
 		error = -EAGAIN;
-		if (!(request->fl_flags & FL_SLEEP))
+		if (!(request->c.flc_flags & FL_SLEEP))
 			goto out;
 		error = FILE_LOCK_DEFERRED;
-		locks_insert_block(fl, request, flock_locks_conflict);
+		locks_insert_block(&fl->c, &request->c, flock_locks_conflict);
 		goto out;
 	}
-	if (request->fl_flags & FL_ACCESS)
+	if (request->c.flc_flags & FL_ACCESS)
 		goto out;
 	locks_copy_lock(new_fl, request);
 	locks_move_blocks(new_fl, request);
-	locks_insert_lock_ctx(new_fl, &ctx->flc_flock);
+	locks_insert_lock_ctx(&new_fl->c, &ctx->flc_flock);
 	new_fl = NULL;
 	error = 0;
 
@@ -1105,9 +1158,9 @@ static int posix_lock_inode(struct inode *inode, struct file_lock *request,
 	void *owner;
 	void (*func)(void);
 
-	ctx = locks_get_lock_context(inode, request->fl_type);
+	ctx = locks_get_lock_context(inode, request->c.flc_type);
 	if (!ctx)
-		return (request->fl_type == F_UNLCK) ? 0 : -ENOMEM;
+		return lock_is_unlock(request) ? 0 : -ENOMEM;
 
 	/*
 	 * We may need two file_lock structures for this operation,
@@ -1115,8 +1168,8 @@ static int posix_lock_inode(struct inode *inode, struct file_lock *request,
 	 *
 	 * In some cases we can be sure, that no new locks will be needed
 	 */
-	if (!(request->fl_flags & FL_ACCESS) &&
-	    (request->fl_type != F_UNLCK ||
+	if (!(request->c.flc_flags & FL_ACCESS) &&
+	    (request->c.flc_type != F_UNLCK ||
 	     request->fl_start != 0 || request->fl_end != OFFSET_MAX)) {
 		new_fl = locks_alloc_lock();
 		new_fl2 = locks_alloc_lock();
@@ -1130,9 +1183,9 @@ retry:
 	 * there are any, either return error or put the request on the
 	 * blocker's list of waiters and the global blocked_hash.
 	 */
-	if (request->fl_type != F_UNLCK) {
-		list_for_each_entry(fl, &ctx->flc_posix, fl_list) {
-			if (!posix_locks_conflict(request, fl))
+	if (request->c.flc_type != F_UNLCK) {
+		list_for_each_entry(fl, &ctx->flc_posix, c.flc_list) {
+			if (!posix_locks_conflict(&request->c, &fl->c))
 				continue;
 			if (fl->fl_lmops && fl->fl_lmops->lm_lock_expirable
 				&& (*fl->fl_lmops->lm_lock_expirable)(fl)) {
@@ -1148,7 +1201,7 @@ retry:
 			if (conflock)
 				locks_copy_conflock(conflock, fl);
 			error = -EAGAIN;
-			if (!(request->fl_flags & FL_SLEEP))
+			if (!(request->c.flc_flags & FL_SLEEP))
 				goto out;
 			/*
 			 * Deadlock detection and insertion into the blocked
@@ -1160,10 +1213,10 @@ retry:
 			 * Ensure that we don't find any locks blocked on this
 			 * request during deadlock detection.
 			 */
-			__locks_wake_up_blocks(request);
+			__locks_wake_up_blocks(&request->c);
 			if (likely(!posix_locks_deadlock(request, fl))) {
 				error = FILE_LOCK_DEFERRED;
-				__locks_insert_block(fl, request,
+				__locks_insert_block(&fl->c, &request->c,
 						     posix_locks_conflict);
 			}
 			spin_unlock(&blocked_lock_lock);
@@ -1173,22 +1226,22 @@ retry:
 
 	/* If we're just looking for a conflict, we're done. */
 	error = 0;
-	if (request->fl_flags & FL_ACCESS)
+	if (request->c.flc_flags & FL_ACCESS)
 		goto out;
 
 	/* Find the first old lock with the same owner as the new lock */
-	list_for_each_entry(fl, &ctx->flc_posix, fl_list) {
-		if (posix_same_owner(request, fl))
+	list_for_each_entry(fl, &ctx->flc_posix, c.flc_list) {
+		if (posix_same_owner(&request->c, &fl->c))
 			break;
 	}
 
 	/* Process locks with this owner. */
-	list_for_each_entry_safe_from(fl, tmp, &ctx->flc_posix, fl_list) {
-		if (!posix_same_owner(request, fl))
+	list_for_each_entry_safe_from(fl, tmp, &ctx->flc_posix, c.flc_list) {
+		if (!posix_same_owner(&request->c, &fl->c))
 			break;
 
 		/* Detect adjacent or overlapping regions (if same lock type) */
-		if (request->fl_type == fl->fl_type) {
+		if (request->c.flc_type == fl->c.flc_type) {
 			/* In all comparisons of start vs end, use
 			 * "start - 1" rather than "end + 1". If end
 			 * is OFFSET_MAX, end + 1 will become negative.
@@ -1215,7 +1268,7 @@ retry:
 			else
 				request->fl_end = fl->fl_end;
 			if (added) {
-				locks_delete_lock_ctx(fl, &dispose);
+				locks_delete_lock_ctx(&fl->c, &dispose);
 				continue;
 			}
 			request = fl;
@@ -1228,7 +1281,7 @@ retry:
 				continue;
 			if (fl->fl_start > request->fl_end)
 				break;
-			if (request->fl_type == F_UNLCK)
+			if (lock_is_unlock(request))
 				added = true;
 			if (fl->fl_start < request->fl_start)
 				left = fl;
@@ -1244,7 +1297,7 @@ retry:
 				 * one (This may happen several times).
 				 */
 				if (added) {
-					locks_delete_lock_ctx(fl, &dispose);
+					locks_delete_lock_ctx(&fl->c, &dispose);
 					continue;
 				}
 				/*
@@ -1261,8 +1314,9 @@ retry:
 				locks_move_blocks(new_fl, request);
 				request = new_fl;
 				new_fl = NULL;
-				locks_insert_lock_ctx(request, &fl->fl_list);
-				locks_delete_lock_ctx(fl, &dispose);
+				locks_insert_lock_ctx(&request->c,
+						      &fl->c.flc_list);
+				locks_delete_lock_ctx(&fl->c, &dispose);
 				added = true;
 			}
 		}
@@ -1279,8 +1333,8 @@ retry:
 
 	error = 0;
 	if (!added) {
-		if (request->fl_type == F_UNLCK) {
-			if (request->fl_flags & FL_EXISTS)
+		if (lock_is_unlock(request)) {
+			if (request->c.flc_flags & FL_EXISTS)
 				error = -ENOENT;
 			goto out;
 		}
@@ -1291,7 +1345,7 @@ retry:
 		}
 		locks_copy_lock(new_fl, request);
 		locks_move_blocks(new_fl, request);
-		locks_insert_lock_ctx(new_fl, &fl->fl_list);
+		locks_insert_lock_ctx(&new_fl->c, &fl->c.flc_list);
 		fl = new_fl;
 		new_fl = NULL;
 	}
@@ -1303,19 +1357,19 @@ retry:
 			left = new_fl2;
 			new_fl2 = NULL;
 			locks_copy_lock(left, right);
-			locks_insert_lock_ctx(left, &fl->fl_list);
+			locks_insert_lock_ctx(&left->c, &fl->c.flc_list);
 		}
 		right->fl_start = request->fl_end + 1;
-		locks_wake_up_blocks(right);
+		locks_wake_up_blocks(&right->c);
 	}
 	if (left) {
 		left->fl_end = request->fl_start - 1;
-		locks_wake_up_blocks(left);
+		locks_wake_up_blocks(&left->c);
 	}
  out:
+	trace_posix_lock_inode(inode, request, error);
 	spin_unlock(&ctx->flc_lock);
 	percpu_up_read(&file_rwsem);
-	trace_posix_lock_inode(inode, request, error);
 	/*
 	 * Free any unused locks.
 	 */
@@ -1364,8 +1418,8 @@ static int posix_lock_inode_wait(struct inode *inode, struct file_lock *fl)
 		error = posix_lock_inode(inode, fl, NULL);
 		if (error != FILE_LOCK_DEFERRED)
 			break;
-		error = wait_event_interruptible(fl->fl_wait,
-					list_empty(&fl->fl_blocked_member));
+		error = wait_event_interruptible(fl->c.flc_wait,
+						 list_empty(&fl->c.flc_blocked_member));
 		if (error)
 			break;
 	}
@@ -1373,37 +1427,37 @@ static int posix_lock_inode_wait(struct inode *inode, struct file_lock *fl)
 	return error;
 }
 
-static void lease_clear_pending(struct file_lock *fl, int arg)
+static void lease_clear_pending(struct file_lease *fl, int arg)
 {
 	switch (arg) {
 	case F_UNLCK:
-		fl->fl_flags &= ~FL_UNLOCK_PENDING;
+		fl->c.flc_flags &= ~FL_UNLOCK_PENDING;
 		fallthrough;
 	case F_RDLCK:
-		fl->fl_flags &= ~FL_DOWNGRADE_PENDING;
+		fl->c.flc_flags &= ~FL_DOWNGRADE_PENDING;
 	}
 }
 
 /* We already had a lease on this file; just change its type */
-int lease_modify(struct file_lock *fl, int arg, struct list_head *dispose)
+int lease_modify(struct file_lease *fl, int arg, struct list_head *dispose)
 {
-	int error = assign_type(fl, arg);
+	int error = assign_type(&fl->c, arg);
 
 	if (error)
 		return error;
 	lease_clear_pending(fl, arg);
-	locks_wake_up_blocks(fl);
+	locks_wake_up_blocks(&fl->c);
 	if (arg == F_UNLCK) {
-		struct file *filp = fl->fl_file;
+		struct file *filp = fl->c.flc_file;
 
 		f_delown(filp);
 		filp->f_owner.signum = 0;
-		fasync_helper(0, fl->fl_file, 0, &fl->fl_fasync);
+		fasync_helper(0, fl->c.flc_file, 0, &fl->fl_fasync);
 		if (fl->fl_fasync != NULL) {
 			printk(KERN_ERR "locks_delete_lock: fasync == %p\n", fl->fl_fasync);
 			fl->fl_fasync = NULL;
 		}
-		locks_delete_lock_ctx(fl, dispose);
+		locks_delete_lock_ctx(&fl->c, dispose);
 	}
 	return 0;
 }
@@ -1420,11 +1474,11 @@ static bool past_time(unsigned long then)
 static void time_out_leases(struct inode *inode, struct list_head *dispose)
 {
 	struct file_lock_context *ctx = inode->i_flctx;
-	struct file_lock *fl, *tmp;
+	struct file_lease *fl, *tmp;
 
 	lockdep_assert_held(&ctx->flc_lock);
 
-	list_for_each_entry_safe(fl, tmp, &ctx->flc_lease, fl_list) {
+	list_for_each_entry_safe(fl, tmp, &ctx->flc_lease, c.flc_list) {
 		trace_time_out_leases(inode, fl);
 		if (past_time(fl->fl_downgrade_time))
 			lease_modify(fl, F_RDLCK, dispose);
@@ -1433,38 +1487,40 @@ static void time_out_leases(struct inode *inode, struct list_head *dispose)
 	}
 }
 
-static bool leases_conflict(struct file_lock *lease, struct file_lock *breaker)
+static bool leases_conflict(struct file_lock_core *lc, struct file_lock_core *bc)
 {
 	bool rc;
+	struct file_lease *lease = file_lease(lc);
+	struct file_lease *breaker = file_lease(bc);
 
 	if (lease->fl_lmops->lm_breaker_owns_lease
 			&& lease->fl_lmops->lm_breaker_owns_lease(lease))
 		return false;
-	if ((breaker->fl_flags & FL_LAYOUT) != (lease->fl_flags & FL_LAYOUT)) {
+	if ((bc->flc_flags & FL_LAYOUT) != (lc->flc_flags & FL_LAYOUT)) {
 		rc = false;
 		goto trace;
 	}
-	if ((breaker->fl_flags & FL_DELEG) && (lease->fl_flags & FL_LEASE)) {
+	if ((bc->flc_flags & FL_DELEG) && (lc->flc_flags & FL_LEASE)) {
 		rc = false;
 		goto trace;
 	}
 
-	rc = locks_conflict(breaker, lease);
+	rc = locks_conflict(bc, lc);
 trace:
 	trace_leases_conflict(rc, lease, breaker);
 	return rc;
 }
 
 static bool
-any_leases_conflict(struct inode *inode, struct file_lock *breaker)
+any_leases_conflict(struct inode *inode, struct file_lease *breaker)
 {
 	struct file_lock_context *ctx = inode->i_flctx;
-	struct file_lock *fl;
+	struct file_lock_core *flc;
 
 	lockdep_assert_held(&ctx->flc_lock);
 
-	list_for_each_entry(fl, &ctx->flc_lease, fl_list) {
-		if (leases_conflict(fl, breaker))
+	list_for_each_entry(flc, &ctx->flc_lease, flc_list) {
+		if (leases_conflict(flc, &breaker->c))
 			return true;
 	}
 	return false;
@@ -1487,7 +1543,7 @@ int __break_lease(struct inode *inode, unsigned int mode, unsigned int type)
 {
 	int error = 0;
 	struct file_lock_context *ctx;
-	struct file_lock *new_fl, *fl, *tmp;
+	struct file_lease *new_fl, *fl, *tmp;
 	unsigned long break_time;
 	int want_write = (mode & O_ACCMODE) != O_RDONLY;
 	LIST_HEAD(dispose);
@@ -1495,7 +1551,7 @@ int __break_lease(struct inode *inode, unsigned int mode, unsigned int type)
 	new_fl = lease_alloc(NULL, want_write ? F_WRLCK : F_RDLCK);
 	if (IS_ERR(new_fl))
 		return PTR_ERR(new_fl);
-	new_fl->fl_flags = type;
+	new_fl->c.flc_flags = type;
 
 	/* typically we will check that ctx is non-NULL before calling */
 	ctx = locks_inode_context(inode);
@@ -1519,22 +1575,22 @@ int __break_lease(struct inode *inode, unsigned int mode, unsigned int type)
 			break_time++;	/* so that 0 means no break time */
 	}
 
-	list_for_each_entry_safe(fl, tmp, &ctx->flc_lease, fl_list) {
-		if (!leases_conflict(fl, new_fl))
+	list_for_each_entry_safe(fl, tmp, &ctx->flc_lease, c.flc_list) {
+		if (!leases_conflict(&fl->c, &new_fl->c))
 			continue;
 		if (want_write) {
-			if (fl->fl_flags & FL_UNLOCK_PENDING)
+			if (fl->c.flc_flags & FL_UNLOCK_PENDING)
 				continue;
-			fl->fl_flags |= FL_UNLOCK_PENDING;
+			fl->c.flc_flags |= FL_UNLOCK_PENDING;
 			fl->fl_break_time = break_time;
 		} else {
 			if (lease_breaking(fl))
 				continue;
-			fl->fl_flags |= FL_DOWNGRADE_PENDING;
+			fl->c.flc_flags |= FL_DOWNGRADE_PENDING;
 			fl->fl_downgrade_time = break_time;
 		}
 		if (fl->fl_lmops->lm_break(fl))
-			locks_delete_lock_ctx(fl, &dispose);
+			locks_delete_lock_ctx(&fl->c, &dispose);
 	}
 
 	if (list_empty(&ctx->flc_lease))
@@ -1547,26 +1603,26 @@ int __break_lease(struct inode *inode, unsigned int mode, unsigned int type)
 	}
 
 restart:
-	fl = list_first_entry(&ctx->flc_lease, struct file_lock, fl_list);
+	fl = list_first_entry(&ctx->flc_lease, struct file_lease, c.flc_list);
 	break_time = fl->fl_break_time;
 	if (break_time != 0)
 		break_time -= jiffies;
 	if (break_time == 0)
 		break_time++;
-	locks_insert_block(fl, new_fl, leases_conflict);
+	locks_insert_block(&fl->c, &new_fl->c, leases_conflict);
 	trace_break_lease_block(inode, new_fl);
 	spin_unlock(&ctx->flc_lock);
 	percpu_up_read(&file_rwsem);
 
 	locks_dispose_list(&dispose);
-	error = wait_event_interruptible_timeout(new_fl->fl_wait,
-					list_empty(&new_fl->fl_blocked_member),
-					break_time);
+	error = wait_event_interruptible_timeout(new_fl->c.flc_wait,
+						 list_empty(&new_fl->c.flc_blocked_member),
+						 break_time);
 
 	percpu_down_read(&file_rwsem);
 	spin_lock(&ctx->flc_lock);
 	trace_break_lease_unblock(inode, new_fl);
-	locks_delete_block(new_fl);
+	__locks_delete_block(&new_fl->c);
 	if (error >= 0) {
 		/*
 		 * Wait for the next conflicting lease that has not been
@@ -1583,7 +1639,7 @@ out:
 	percpu_up_read(&file_rwsem);
 	locks_dispose_list(&dispose);
 free_lock:
-	locks_free_lock(new_fl);
+	locks_free_lease(new_fl);
 	return error;
 }
 EXPORT_SYMBOL(__break_lease);
@@ -1601,14 +1657,14 @@ void lease_get_mtime(struct inode *inode, struct timespec64 *time)
 {
 	bool has_lease = false;
 	struct file_lock_context *ctx;
-	struct file_lock *fl;
+	struct file_lock_core *flc;
 
 	ctx = locks_inode_context(inode);
 	if (ctx && !list_empty_careful(&ctx->flc_lease)) {
 		spin_lock(&ctx->flc_lock);
-		fl = list_first_entry_or_null(&ctx->flc_lease,
-					      struct file_lock, fl_list);
-		if (fl && (fl->fl_type == F_WRLCK))
+		flc = list_first_entry_or_null(&ctx->flc_lease,
+					       struct file_lock_core, flc_list);
+		if (flc && flc->flc_type == F_WRLCK)
 			has_lease = true;
 		spin_unlock(&ctx->flc_lock);
 	}
@@ -1643,7 +1699,7 @@ EXPORT_SYMBOL(lease_get_mtime);
  */
 int fcntl_getlease(struct file *filp)
 {
-	struct file_lock *fl;
+	struct file_lease *fl;
 	struct inode *inode = file_inode(filp);
 	struct file_lock_context *ctx;
 	int type = F_UNLCK;
@@ -1654,8 +1710,8 @@ int fcntl_getlease(struct file *filp)
 		percpu_down_read(&file_rwsem);
 		spin_lock(&ctx->flc_lock);
 		time_out_leases(inode, &dispose);
-		list_for_each_entry(fl, &ctx->flc_lease, fl_list) {
-			if (fl->fl_file != filp)
+		list_for_each_entry(fl, &ctx->flc_lease, c.flc_list) {
+			if (fl->c.flc_file != filp)
 				continue;
 			type = target_leasetype(fl);
 			break;
@@ -1715,12 +1771,12 @@ check_conflicting_open(struct file *filp, const int arg, int flags)
 }
 
 static int
-generic_add_lease(struct file *filp, int arg, struct file_lock **flp, void **priv)
+generic_add_lease(struct file *filp, int arg, struct file_lease **flp, void **priv)
 {
-	struct file_lock *fl, *my_fl = NULL, *lease;
+	struct file_lease *fl, *my_fl = NULL, *lease;
 	struct inode *inode = file_inode(filp);
 	struct file_lock_context *ctx;
-	bool is_deleg = (*flp)->fl_flags & FL_DELEG;
+	bool is_deleg = (*flp)->c.flc_flags & FL_DELEG;
 	int error;
 	LIST_HEAD(dispose);
 
@@ -1746,7 +1802,7 @@ generic_add_lease(struct file *filp, int arg, struct file_lock **flp, void **pri
 	percpu_down_read(&file_rwsem);
 	spin_lock(&ctx->flc_lock);
 	time_out_leases(inode, &dispose);
-	error = check_conflicting_open(filp, arg, lease->fl_flags);
+	error = check_conflicting_open(filp, arg, lease->c.flc_flags);
 	if (error)
 		goto out;
 
@@ -1759,9 +1815,9 @@ generic_add_lease(struct file *filp, int arg, struct file_lock **flp, void **pri
 	 * except for this filp.
 	 */
 	error = -EAGAIN;
-	list_for_each_entry(fl, &ctx->flc_lease, fl_list) {
-		if (fl->fl_file == filp &&
-		    fl->fl_owner == lease->fl_owner) {
+	list_for_each_entry(fl, &ctx->flc_lease, c.flc_list) {
+		if (fl->c.flc_file == filp &&
+		    fl->c.flc_owner == lease->c.flc_owner) {
 			my_fl = fl;
 			continue;
 		}
@@ -1776,7 +1832,7 @@ generic_add_lease(struct file *filp, int arg, struct file_lock **flp, void **pri
 		 * Modifying our existing lease is OK, but no getting a
 		 * new lease if someone else is opening for write:
 		 */
-		if (fl->fl_flags & FL_UNLOCK_PENDING)
+		if (fl->c.flc_flags & FL_UNLOCK_PENDING)
 			goto out;
 	}
 
@@ -1792,7 +1848,7 @@ generic_add_lease(struct file *filp, int arg, struct file_lock **flp, void **pri
 	if (!leases_enable)
 		goto out;
 
-	locks_insert_lock_ctx(lease, &ctx->flc_lease);
+	locks_insert_lock_ctx(&lease->c, &ctx->flc_lease);
 	/*
 	 * The check in break_lease() is lockless. It's possible for another
 	 * open to race in after we did the earlier check for a conflicting
@@ -1803,9 +1859,9 @@ generic_add_lease(struct file *filp, int arg, struct file_lock **flp, void **pri
 	 * precedes these checks.
 	 */
 	smp_mb();
-	error = check_conflicting_open(filp, arg, lease->fl_flags);
+	error = check_conflicting_open(filp, arg, lease->c.flc_flags);
 	if (error) {
-		locks_unlink_lock_ctx(lease);
+		locks_unlink_lock_ctx(&lease->c);
 		goto out;
 	}
 
@@ -1826,7 +1882,7 @@ out:
 static int generic_delete_lease(struct file *filp, void *owner)
 {
 	int error = -EAGAIN;
-	struct file_lock *fl, *victim = NULL;
+	struct file_lease *fl, *victim = NULL;
 	struct inode *inode = file_inode(filp);
 	struct file_lock_context *ctx;
 	LIST_HEAD(dispose);
@@ -1839,9 +1895,9 @@ static int generic_delete_lease(struct file *filp, void *owner)
 
 	percpu_down_read(&file_rwsem);
 	spin_lock(&ctx->flc_lock);
-	list_for_each_entry(fl, &ctx->flc_lease, fl_list) {
-		if (fl->fl_file == filp &&
-		    fl->fl_owner == owner) {
+	list_for_each_entry(fl, &ctx->flc_lease, c.flc_list) {
+		if (fl->c.flc_file == filp &&
+		    fl->c.flc_owner == owner) {
 			victim = fl;
 			break;
 		}
@@ -1866,21 +1922,9 @@ static int generic_delete_lease(struct file *filp, void *owner)
  *	The (input) flp->fl_lmops->lm_break function is required
  *	by break_lease().
  */
-int generic_setlease(struct file *filp, int arg, struct file_lock **flp,
+int generic_setlease(struct file *filp, int arg, struct file_lease **flp,
 			void **priv)
 {
-	struct inode *inode = file_inode(filp);
-	vfsuid_t vfsuid = i_uid_into_vfsuid(file_mnt_idmap(filp), inode);
-	int error;
-
-	if ((!vfsuid_eq_kuid(vfsuid, current_fsuid())) && !capable(CAP_LEASE))
-		return -EACCES;
-	if (!S_ISREG(inode->i_mode))
-		return -EINVAL;
-	error = security_file_lock(filp, arg);
-	if (error)
-		return error;
-
 	switch (arg) {
 	case F_UNLCK:
 		return generic_delete_lease(filp, *priv);
@@ -1913,7 +1957,7 @@ lease_notifier_chain_init(void)
 }
 
 static inline void
-setlease_notifier(int arg, struct file_lock *lease)
+setlease_notifier(int arg, struct file_lease *lease)
 {
 	if (arg != F_UNLCK)
 		srcu_notifier_call_chain(&lease_notifier_chain, arg, lease);
@@ -1931,6 +1975,19 @@ void lease_unregister_notifier(struct notifier_block *nb)
 }
 EXPORT_SYMBOL_GPL(lease_unregister_notifier);
 
+
+int
+kernel_setlease(struct file *filp, int arg, struct file_lease **lease, void **priv)
+{
+	if (lease)
+		setlease_notifier(arg, *lease);
+	if (filp->f_op->setlease)
+		return filp->f_op->setlease(filp, arg, lease, priv);
+	else
+		return generic_setlease(filp, arg, lease, priv);
+}
+EXPORT_SYMBOL_GPL(kernel_setlease);
+
 /**
  * vfs_setlease        -       sets a lease on an open file
  * @filp:	file pointer
@@ -1949,20 +2006,26 @@ EXPORT_SYMBOL_GPL(lease_unregister_notifier);
  * may be NULL if the lm_setup operation doesn't require it.
  */
 int
-vfs_setlease(struct file *filp, int arg, struct file_lock **lease, void **priv)
+vfs_setlease(struct file *filp, int arg, struct file_lease **lease, void **priv)
 {
-	if (lease)
-		setlease_notifier(arg, *lease);
-	if (filp->f_op->setlease)
-		return filp->f_op->setlease(filp, arg, lease, priv);
-	else
-		return generic_setlease(filp, arg, lease, priv);
+	struct inode *inode = file_inode(filp);
+	vfsuid_t vfsuid = i_uid_into_vfsuid(file_mnt_idmap(filp), inode);
+	int error;
+
+	if ((!vfsuid_eq_kuid(vfsuid, current_fsuid())) && !capable(CAP_LEASE))
+		return -EACCES;
+	if (!S_ISREG(inode->i_mode))
+		return -EINVAL;
+	error = security_file_lock(filp, arg);
+	if (error)
+		return error;
+	return kernel_setlease(filp, arg, lease, priv);
 }
 EXPORT_SYMBOL_GPL(vfs_setlease);
 
 static int do_fcntl_add_lease(unsigned int fd, struct file *filp, int arg)
 {
-	struct file_lock *fl;
+	struct file_lease *fl;
 	struct fasync_struct *new;
 	int error;
 
@@ -1972,14 +2035,14 @@ static int do_fcntl_add_lease(unsigned int fd, struct file *filp, int arg)
 
 	new = fasync_alloc();
 	if (!new) {
-		locks_free_lock(fl);
+		locks_free_lease(fl);
 		return -ENOMEM;
 	}
 	new->fa_fd = fd;
 
 	error = vfs_setlease(filp, arg, &fl, (void **)&new);
 	if (fl)
-		locks_free_lock(fl);
+		locks_free_lease(fl);
 	if (new)
 		fasync_free(new);
 	return error;
@@ -2017,8 +2080,8 @@ static int flock_lock_inode_wait(struct inode *inode, struct file_lock *fl)
 		error = flock_lock_inode(inode, fl);
 		if (error != FILE_LOCK_DEFERRED)
 			break;
-		error = wait_event_interruptible(fl->fl_wait,
-				list_empty(&fl->fl_blocked_member));
+		error = wait_event_interruptible(fl->c.flc_wait,
+						 list_empty(&fl->c.flc_blocked_member));
 		if (error)
 			break;
 	}
@@ -2036,7 +2099,7 @@ static int flock_lock_inode_wait(struct inode *inode, struct file_lock *fl)
 int locks_lock_inode_wait(struct inode *inode, struct file_lock *fl)
 {
 	int res = 0;
-	switch (fl->fl_flags & (FL_POSIX|FL_FLOCK)) {
+	switch (fl->c.flc_flags & (FL_POSIX|FL_FLOCK)) {
 		case FL_POSIX:
 			res = posix_lock_inode_wait(inode, fl);
 			break;
@@ -2098,13 +2161,13 @@ SYSCALL_DEFINE2(flock, unsigned int, fd, unsigned int, cmd)
 
 	flock_make_lock(f.file, &fl, type);
 
-	error = security_file_lock(f.file, fl.fl_type);
+	error = security_file_lock(f.file, fl.c.flc_type);
 	if (error)
 		goto out_putf;
 
 	can_sleep = !(cmd & LOCK_NB);
 	if (can_sleep)
-		fl.fl_flags |= FL_SLEEP;
+		fl.c.flc_flags |= FL_SLEEP;
 
 	if (f.file->f_op->flock)
 		error = f.file->f_op->flock(f.file,
@@ -2130,7 +2193,7 @@ SYSCALL_DEFINE2(flock, unsigned int, fd, unsigned int, cmd)
  */
 int vfs_test_lock(struct file *filp, struct file_lock *fl)
 {
-	WARN_ON_ONCE(filp != fl->fl_file);
+	WARN_ON_ONCE(filp != fl->c.flc_file);
 	if (filp->f_op->lock)
 		return filp->f_op->lock(filp, F_GETLK, fl);
 	posix_test_lock(filp, fl);
@@ -2145,25 +2208,28 @@ EXPORT_SYMBOL_GPL(vfs_test_lock);
  *
  * Used to translate a fl_pid into a namespace virtual pid number
  */
-static pid_t locks_translate_pid(struct file_lock *fl, struct pid_namespace *ns)
+static pid_t locks_translate_pid(struct file_lock_core *fl, struct pid_namespace *ns)
 {
 	pid_t vnr;
 	struct pid *pid;
 
-	if (IS_OFDLCK(fl))
+	if (fl->flc_flags & FL_OFDLCK)
 		return -1;
-	if (IS_REMOTELCK(fl))
-		return fl->fl_pid;
+
+	/* Remote locks report a negative pid value */
+	if (fl->flc_pid <= 0)
+		return fl->flc_pid;
+
 	/*
 	 * If the flock owner process is dead and its pid has been already
 	 * freed, the translation below won't work, but we still want to show
 	 * flock owner pid number in init pidns.
 	 */
 	if (ns == &init_pid_ns)
-		return (pid_t)fl->fl_pid;
+		return (pid_t) fl->flc_pid;
 
 	rcu_read_lock();
-	pid = find_pid_ns(fl->fl_pid, &init_pid_ns);
+	pid = find_pid_ns(fl->flc_pid, &init_pid_ns);
 	vnr = pid_nr_ns(pid, ns);
 	rcu_read_unlock();
 	return vnr;
@@ -2171,7 +2237,7 @@ static pid_t locks_translate_pid(struct file_lock *fl, struct pid_namespace *ns)
 
 static int posix_lock_to_flock(struct flock *flock, struct file_lock *fl)
 {
-	flock->l_pid = locks_translate_pid(fl, task_active_pid_ns(current));
+	flock->l_pid = locks_translate_pid(&fl->c, task_active_pid_ns(current));
 #if BITS_PER_LONG == 32
 	/*
 	 * Make sure we can represent the posix lock via
@@ -2186,19 +2252,19 @@ static int posix_lock_to_flock(struct flock *flock, struct file_lock *fl)
 	flock->l_len = fl->fl_end == OFFSET_MAX ? 0 :
 		fl->fl_end - fl->fl_start + 1;
 	flock->l_whence = 0;
-	flock->l_type = fl->fl_type;
+	flock->l_type = fl->c.flc_type;
 	return 0;
 }
 
 #if BITS_PER_LONG == 32
 static void posix_lock_to_flock64(struct flock64 *flock, struct file_lock *fl)
 {
-	flock->l_pid = locks_translate_pid(fl, task_active_pid_ns(current));
+	flock->l_pid = locks_translate_pid(&fl->c, task_active_pid_ns(current));
 	flock->l_start = fl->fl_start;
 	flock->l_len = fl->fl_end == OFFSET_MAX ? 0 :
 		fl->fl_end - fl->fl_start + 1;
 	flock->l_whence = 0;
-	flock->l_type = fl->fl_type;
+	flock->l_type = fl->c.flc_type;
 }
 #endif
 
@@ -2227,16 +2293,16 @@ int fcntl_getlk(struct file *filp, unsigned int cmd, struct flock *flock)
 		if (flock->l_pid != 0)
 			goto out;
 
-		fl->fl_flags |= FL_OFDLCK;
-		fl->fl_owner = filp;
+		fl->c.flc_flags |= FL_OFDLCK;
+		fl->c.flc_owner = filp;
 	}
 
 	error = vfs_test_lock(filp, fl);
 	if (error)
 		goto out;
 
-	flock->l_type = fl->fl_type;
-	if (fl->fl_type != F_UNLCK) {
+	flock->l_type = fl->c.flc_type;
+	if (fl->c.flc_type != F_UNLCK) {
 		error = posix_lock_to_flock(flock, fl);
 		if (error)
 			goto out;
@@ -2283,7 +2349,7 @@ out:
  */
 int vfs_lock_file(struct file *filp, unsigned int cmd, struct file_lock *fl, struct file_lock *conf)
 {
-	WARN_ON_ONCE(filp != fl->fl_file);
+	WARN_ON_ONCE(filp != fl->c.flc_file);
 	if (filp->f_op->lock)
 		return filp->f_op->lock(filp, cmd, fl);
 	else
@@ -2296,7 +2362,7 @@ static int do_lock_file_wait(struct file *filp, unsigned int cmd,
 {
 	int error;
 
-	error = security_file_lock(filp, fl->fl_type);
+	error = security_file_lock(filp, fl->c.flc_type);
 	if (error)
 		return error;
 
@@ -2304,8 +2370,8 @@ static int do_lock_file_wait(struct file *filp, unsigned int cmd,
 		error = vfs_lock_file(filp, cmd, fl, NULL);
 		if (error != FILE_LOCK_DEFERRED)
 			break;
-		error = wait_event_interruptible(fl->fl_wait,
-					list_empty(&fl->fl_blocked_member));
+		error = wait_event_interruptible(fl->c.flc_wait,
+						 list_empty(&fl->c.flc_blocked_member));
 		if (error)
 			break;
 	}
@@ -2318,13 +2384,13 @@ static int do_lock_file_wait(struct file *filp, unsigned int cmd,
 static int
 check_fmode_for_setlk(struct file_lock *fl)
 {
-	switch (fl->fl_type) {
+	switch (fl->c.flc_type) {
 	case F_RDLCK:
-		if (!(fl->fl_file->f_mode & FMODE_READ))
+		if (!(fl->c.flc_file->f_mode & FMODE_READ))
 			return -EBADF;
 		break;
 	case F_WRLCK:
-		if (!(fl->fl_file->f_mode & FMODE_WRITE))
+		if (!(fl->c.flc_file->f_mode & FMODE_WRITE))
 			return -EBADF;
 	}
 	return 0;
@@ -2363,8 +2429,8 @@ int fcntl_setlk(unsigned int fd, struct file *filp, unsigned int cmd,
 			goto out;
 
 		cmd = F_SETLK;
-		file_lock->fl_flags |= FL_OFDLCK;
-		file_lock->fl_owner = filp;
+		file_lock->c.flc_flags |= FL_OFDLCK;
+		file_lock->c.flc_owner = filp;
 		break;
 	case F_OFD_SETLKW:
 		error = -EINVAL;
@@ -2372,22 +2438,23 @@ int fcntl_setlk(unsigned int fd, struct file *filp, unsigned int cmd,
 			goto out;
 
 		cmd = F_SETLKW;
-		file_lock->fl_flags |= FL_OFDLCK;
-		file_lock->fl_owner = filp;
+		file_lock->c.flc_flags |= FL_OFDLCK;
+		file_lock->c.flc_owner = filp;
 		fallthrough;
 	case F_SETLKW:
-		file_lock->fl_flags |= FL_SLEEP;
+		file_lock->c.flc_flags |= FL_SLEEP;
 	}
 
 	error = do_lock_file_wait(filp, cmd, file_lock);
 
 	/*
-	 * Attempt to detect a close/fcntl race and recover by releasing the
-	 * lock that was just acquired. There is no need to do that when we're
+	 * Detect close/fcntl races and recover by zapping all POSIX locks
+	 * associated with this file and our files_struct, just like on
+	 * filp_flush(). There is no need to do that when we're
 	 * unlocking though, or for OFD locks.
 	 */
-	if (!error && file_lock->fl_type != F_UNLCK &&
-	    !(file_lock->fl_flags & FL_OFDLCK)) {
+	if (!error && file_lock->c.flc_type != F_UNLCK &&
+	    !(file_lock->c.flc_flags & FL_OFDLCK)) {
 		struct files_struct *files = current->files;
 		/*
 		 * We need that spin_lock here - it prevents reordering between
@@ -2398,9 +2465,7 @@ int fcntl_setlk(unsigned int fd, struct file *filp, unsigned int cmd,
 		f = files_lookup_fd_locked(files, fd);
 		spin_unlock(&files->file_lock);
 		if (f != filp) {
-			file_lock->fl_type = F_UNLCK;
-			error = do_lock_file_wait(filp, cmd, file_lock);
-			WARN_ON_ONCE(error);
+			locks_remove_posix(filp, files);
 			error = -EBADF;
 		}
 	}
@@ -2437,16 +2502,16 @@ int fcntl_getlk64(struct file *filp, unsigned int cmd, struct flock64 *flock)
 		if (flock->l_pid != 0)
 			goto out;
 
-		fl->fl_flags |= FL_OFDLCK;
-		fl->fl_owner = filp;
+		fl->c.flc_flags |= FL_OFDLCK;
+		fl->c.flc_owner = filp;
 	}
 
 	error = vfs_test_lock(filp, fl);
 	if (error)
 		goto out;
 
-	flock->l_type = fl->fl_type;
-	if (fl->fl_type != F_UNLCK)
+	flock->l_type = fl->c.flc_type;
+	if (fl->c.flc_type != F_UNLCK)
 		posix_lock_to_flock64(flock, fl);
 
 out:
@@ -2486,8 +2551,8 @@ int fcntl_setlk64(unsigned int fd, struct file *filp, unsigned int cmd,
 			goto out;
 
 		cmd = F_SETLK64;
-		file_lock->fl_flags |= FL_OFDLCK;
-		file_lock->fl_owner = filp;
+		file_lock->c.flc_flags |= FL_OFDLCK;
+		file_lock->c.flc_owner = filp;
 		break;
 	case F_OFD_SETLKW:
 		error = -EINVAL;
@@ -2495,22 +2560,23 @@ int fcntl_setlk64(unsigned int fd, struct file *filp, unsigned int cmd,
 			goto out;
 
 		cmd = F_SETLKW64;
-		file_lock->fl_flags |= FL_OFDLCK;
-		file_lock->fl_owner = filp;
+		file_lock->c.flc_flags |= FL_OFDLCK;
+		file_lock->c.flc_owner = filp;
 		fallthrough;
 	case F_SETLKW64:
-		file_lock->fl_flags |= FL_SLEEP;
+		file_lock->c.flc_flags |= FL_SLEEP;
 	}
 
 	error = do_lock_file_wait(filp, cmd, file_lock);
 
 	/*
-	 * Attempt to detect a close/fcntl race and recover by releasing the
-	 * lock that was just acquired. There is no need to do that when we're
+	 * Detect close/fcntl races and recover by zapping all POSIX locks
+	 * associated with this file and our files_struct, just like on
+	 * filp_flush(). There is no need to do that when we're
 	 * unlocking though, or for OFD locks.
 	 */
-	if (!error && file_lock->fl_type != F_UNLCK &&
-	    !(file_lock->fl_flags & FL_OFDLCK)) {
+	if (!error && file_lock->c.flc_type != F_UNLCK &&
+	    !(file_lock->c.flc_flags & FL_OFDLCK)) {
 		struct files_struct *files = current->files;
 		/*
 		 * We need that spin_lock here - it prevents reordering between
@@ -2521,9 +2587,7 @@ int fcntl_setlk64(unsigned int fd, struct file *filp, unsigned int cmd,
 		f = files_lookup_fd_locked(files, fd);
 		spin_unlock(&files->file_lock);
 		if (f != filp) {
-			file_lock->fl_type = F_UNLCK;
-			error = do_lock_file_wait(filp, cmd, file_lock);
-			WARN_ON_ONCE(error);
+			locks_remove_posix(filp, files);
 			error = -EBADF;
 		}
 	}
@@ -2555,13 +2619,13 @@ void locks_remove_posix(struct file *filp, fl_owner_t owner)
 		return;
 
 	locks_init_lock(&lock);
-	lock.fl_type = F_UNLCK;
-	lock.fl_flags = FL_POSIX | FL_CLOSE;
+	lock.c.flc_type = F_UNLCK;
+	lock.c.flc_flags = FL_POSIX | FL_CLOSE;
 	lock.fl_start = 0;
 	lock.fl_end = OFFSET_MAX;
-	lock.fl_owner = owner;
-	lock.fl_pid = current->tgid;
-	lock.fl_file = filp;
+	lock.c.flc_owner = owner;
+	lock.c.flc_pid = current->tgid;
+	lock.c.flc_file = filp;
 	lock.fl_ops = NULL;
 	lock.fl_lmops = NULL;
 
@@ -2584,7 +2648,7 @@ locks_remove_flock(struct file *filp, struct file_lock_context *flctx)
 		return;
 
 	flock_make_lock(filp, &fl, F_UNLCK);
-	fl.fl_flags |= FL_CLOSE;
+	fl.c.flc_flags |= FL_CLOSE;
 
 	if (filp->f_op->flock)
 		filp->f_op->flock(filp, F_SETLKW, &fl);
@@ -2599,7 +2663,7 @@ locks_remove_flock(struct file *filp, struct file_lock_context *flctx)
 static void
 locks_remove_lease(struct file *filp, struct file_lock_context *ctx)
 {
-	struct file_lock *fl, *tmp;
+	struct file_lease *fl, *tmp;
 	LIST_HEAD(dispose);
 
 	if (list_empty(&ctx->flc_lease))
@@ -2607,8 +2671,8 @@ locks_remove_lease(struct file *filp, struct file_lock_context *ctx)
 
 	percpu_down_read(&file_rwsem);
 	spin_lock(&ctx->flc_lock);
-	list_for_each_entry_safe(fl, tmp, &ctx->flc_lease, fl_list)
-		if (filp == fl->fl_file)
+	list_for_each_entry_safe(fl, tmp, &ctx->flc_lease, c.flc_list)
+		if (filp == fl->c.flc_file)
 			lease_modify(fl, F_UNLCK, &dispose);
 	spin_unlock(&ctx->flc_lock);
 	percpu_up_read(&file_rwsem);
@@ -2652,7 +2716,7 @@ void locks_remove_file(struct file *filp)
  */
 int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
 {
-	WARN_ON_ONCE(filp != fl->fl_file);
+	WARN_ON_ONCE(filp != fl->c.flc_file);
 	if (filp->f_op->lock)
 		return filp->f_op->lock(filp, F_CANCELLK, fl);
 	return 0;
@@ -2691,69 +2755,73 @@ struct locks_iterator {
 	loff_t	li_pos;
 };
 
-static void lock_get_status(struct seq_file *f, struct file_lock *fl,
+static void lock_get_status(struct seq_file *f, struct file_lock_core *flc,
 			    loff_t id, char *pfx, int repeat)
 {
 	struct inode *inode = NULL;
-	unsigned int fl_pid;
+	unsigned int pid;
 	struct pid_namespace *proc_pidns = proc_pid_ns(file_inode(f->file)->i_sb);
-	int type;
+	int type = flc->flc_type;
+	struct file_lock *fl = file_lock(flc);
+
+	pid = locks_translate_pid(flc, proc_pidns);
 
-	fl_pid = locks_translate_pid(fl, proc_pidns);
 	/*
 	 * If lock owner is dead (and pid is freed) or not visible in current
 	 * pidns, zero is shown as a pid value. Check lock info from
 	 * init_pid_ns to get saved lock pid value.
 	 */
-
-	if (fl->fl_file != NULL)
-		inode = file_inode(fl->fl_file);
+	if (flc->flc_file != NULL)
+		inode = file_inode(flc->flc_file);
 
 	seq_printf(f, "%lld: ", id);
 
 	if (repeat)
 		seq_printf(f, "%*s", repeat - 1 + (int)strlen(pfx), pfx);
 
-	if (IS_POSIX(fl)) {
-		if (fl->fl_flags & FL_ACCESS)
+	if (flc->flc_flags & FL_POSIX) {
+		if (flc->flc_flags & FL_ACCESS)
 			seq_puts(f, "ACCESS");
-		else if (IS_OFDLCK(fl))
+		else if (flc->flc_flags & FL_OFDLCK)
 			seq_puts(f, "OFDLCK");
 		else
 			seq_puts(f, "POSIX ");
 
 		seq_printf(f, " %s ",
 			     (inode == NULL) ? "*NOINODE*" : "ADVISORY ");
-	} else if (IS_FLOCK(fl)) {
+	} else if (flc->flc_flags & FL_FLOCK) {
 		seq_puts(f, "FLOCK  ADVISORY  ");
-	} else if (IS_LEASE(fl)) {
-		if (fl->fl_flags & FL_DELEG)
+	} else if (flc->flc_flags & (FL_LEASE|FL_DELEG|FL_LAYOUT)) {
+		struct file_lease *lease = file_lease(flc);
+
+		type = target_leasetype(lease);
+
+		if (flc->flc_flags & FL_DELEG)
 			seq_puts(f, "DELEG  ");
 		else
 			seq_puts(f, "LEASE  ");
 
-		if (lease_breaking(fl))
+		if (lease_breaking(lease))
 			seq_puts(f, "BREAKING  ");
-		else if (fl->fl_file)
+		else if (flc->flc_file)
 			seq_puts(f, "ACTIVE    ");
 		else
 			seq_puts(f, "BREAKER   ");
 	} else {
 		seq_puts(f, "UNKNOWN UNKNOWN  ");
 	}
-	type = IS_LEASE(fl) ? target_leasetype(fl) : fl->fl_type;
 
 	seq_printf(f, "%s ", (type == F_WRLCK) ? "WRITE" :
 			     (type == F_RDLCK) ? "READ" : "UNLCK");
 	if (inode) {
 		/* userspace relies on this representation of dev_t */
-		seq_printf(f, "%d %02x:%02x:%lu ", fl_pid,
+		seq_printf(f, "%d %02x:%02x:%lu ", pid,
 				MAJOR(inode->i_sb->s_dev),
 				MINOR(inode->i_sb->s_dev), inode->i_ino);
 	} else {
-		seq_printf(f, "%d <none>:0 ", fl_pid);
+		seq_printf(f, "%d <none>:0 ", pid);
 	}
-	if (IS_POSIX(fl)) {
+	if (flc->flc_flags & FL_POSIX) {
 		if (fl->fl_end == OFFSET_MAX)
 			seq_printf(f, "%Ld EOF\n", fl->fl_start);
 		else
@@ -2763,17 +2831,18 @@ static void lock_get_status(struct seq_file *f, struct file_lock *fl,
 	}
 }
 
-static struct file_lock *get_next_blocked_member(struct file_lock *node)
+static struct file_lock_core *get_next_blocked_member(struct file_lock_core *node)
 {
-	struct file_lock *tmp;
+	struct file_lock_core *tmp;
 
 	/* NULL node or root node */
-	if (node == NULL || node->fl_blocker == NULL)
+	if (node == NULL || node->flc_blocker == NULL)
 		return NULL;
 
 	/* Next member in the linked list could be itself */
-	tmp = list_next_entry(node, fl_blocked_member);
-	if (list_entry_is_head(tmp, &node->fl_blocker->fl_blocked_requests, fl_blocked_member)
+	tmp = list_next_entry(node, flc_blocked_member);
+	if (list_entry_is_head(tmp, &node->flc_blocker->flc_blocked_requests,
+			       flc_blocked_member)
 		|| tmp == node) {
 		return NULL;
 	}
@@ -2784,18 +2853,18 @@ static struct file_lock *get_next_blocked_member(struct file_lock *node)
 static int locks_show(struct seq_file *f, void *v)
 {
 	struct locks_iterator *iter = f->private;
-	struct file_lock *cur, *tmp;
+	struct file_lock_core *cur, *tmp;
 	struct pid_namespace *proc_pidns = proc_pid_ns(file_inode(f->file)->i_sb);
 	int level = 0;
 
-	cur = hlist_entry(v, struct file_lock, fl_link);
+	cur = hlist_entry(v, struct file_lock_core, flc_link);
 
 	if (locks_translate_pid(cur, proc_pidns) == 0)
 		return 0;
 
-	/* View this crossed linked list as a binary tree, the first member of fl_blocked_requests
-	 * is the left child of current node, the next silibing in fl_blocked_member is the
-	 * right child, we can alse get the parent of current node from fl_blocker, so this
+	/* View this crossed linked list as a binary tree, the first member of flc_blocked_requests
+	 * is the left child of current node, the next silibing in flc_blocked_member is the
+	 * right child, we can alse get the parent of current node from flc_blocker, so this
 	 * question becomes traversal of a binary tree
 	 */
 	while (cur != NULL) {
@@ -2804,17 +2873,18 @@ static int locks_show(struct seq_file *f, void *v)
 		else
 			lock_get_status(f, cur, iter->li_pos, "", level);
 
-		if (!list_empty(&cur->fl_blocked_requests)) {
+		if (!list_empty(&cur->flc_blocked_requests)) {
 			/* Turn left */
-			cur = list_first_entry_or_null(&cur->fl_blocked_requests,
-				struct file_lock, fl_blocked_member);
+			cur = list_first_entry_or_null(&cur->flc_blocked_requests,
+						       struct file_lock_core,
+						       flc_blocked_member);
 			level++;
 		} else {
 			/* Turn right */
 			tmp = get_next_blocked_member(cur);
 			/* Fall back to parent node */
-			while (tmp == NULL && cur->fl_blocker != NULL) {
-				cur = cur->fl_blocker;
+			while (tmp == NULL && cur->flc_blocker != NULL) {
+				cur = cur->flc_blocker;
 				level--;
 				tmp = get_next_blocked_member(cur);
 			}
@@ -2829,14 +2899,13 @@ static void __show_fd_locks(struct seq_file *f,
 			struct list_head *head, int *id,
 			struct file *filp, struct files_struct *files)
 {
-	struct file_lock *fl;
+	struct file_lock_core *fl;
 
-	list_for_each_entry(fl, head, fl_list) {
+	list_for_each_entry(fl, head, flc_list) {
 
-		if (filp != fl->fl_file)
+		if (filp != fl->flc_file)
 			continue;
-		if (fl->fl_owner != files &&
-		    fl->fl_owner != filp)
+		if (fl->flc_owner != files && fl->flc_owner != filp)
 			continue;
 
 		(*id)++;
@@ -2915,6 +2984,9 @@ static int __init filelock_init(void)
 	filelock_cache = kmem_cache_create("file_lock_cache",
 			sizeof(struct file_lock), 0, SLAB_PANIC, NULL);
 
+	filelease_cache = kmem_cache_create("file_lock_cache",
+			sizeof(struct file_lease), 0, SLAB_PANIC, NULL);
+
 	for_each_possible_cpu(i) {
 		struct file_lock_list_struct *fll = per_cpu_ptr(&file_lock_list, i);
 
diff --git a/fs/mbcache.c b/fs/mbcache.c
index 82aa7a35db..e60a840999 100644
--- a/fs/mbcache.c
+++ b/fs/mbcache.c
@@ -426,9 +426,7 @@ EXPORT_SYMBOL(mb_cache_destroy);
 
 static int __init mbcache_init(void)
 {
-	mb_entry_cache = kmem_cache_create("mbcache",
-				sizeof(struct mb_cache_entry), 0,
-				SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD, NULL);
+	mb_entry_cache = KMEM_CACHE(mb_cache_entry, SLAB_RECLAIM_ACCOUNT);
 	if (!mb_entry_cache)
 		return -ENOMEM;
 	return 0;
diff --git a/fs/minix/inode.c b/fs/minix/inode.c
index 73f37f2980..7f9a2d8aa4 100644
--- a/fs/minix/inode.c
+++ b/fs/minix/inode.c
@@ -20,11 +20,11 @@
 #include <linux/mpage.h>
 #include <linux/vfs.h>
 #include <linux/writeback.h>
+#include <linux/fs_context.h>
 
 static int minix_write_inode(struct inode *inode,
 		struct writeback_control *wbc);
 static int minix_statfs(struct dentry *dentry, struct kstatfs *buf);
-static int minix_remount (struct super_block * sb, int * flags, char * data);
 
 static void minix_evict_inode(struct inode *inode)
 {
@@ -87,7 +87,7 @@ static int __init init_inodecache(void)
 	minix_inode_cachep = kmem_cache_create("minix_inode_cache",
 					     sizeof(struct minix_inode_info),
 					     0, (SLAB_RECLAIM_ACCOUNT|
-						SLAB_MEM_SPREAD|SLAB_ACCOUNT),
+						SLAB_ACCOUNT),
 					     init_once);
 	if (minix_inode_cachep == NULL)
 		return -ENOMEM;
@@ -111,19 +111,19 @@ static const struct super_operations minix_sops = {
 	.evict_inode	= minix_evict_inode,
 	.put_super	= minix_put_super,
 	.statfs		= minix_statfs,
-	.remount_fs	= minix_remount,
 };
 
-static int minix_remount (struct super_block * sb, int * flags, char * data)
+static int minix_reconfigure(struct fs_context *fc)
 {
-	struct minix_sb_info * sbi = minix_sb(sb);
 	struct minix_super_block * ms;
+	struct super_block *sb = fc->root->d_sb;
+	struct minix_sb_info * sbi = sb->s_fs_info;
 
 	sync_filesystem(sb);
 	ms = sbi->s_ms;
-	if ((bool)(*flags & SB_RDONLY) == sb_rdonly(sb))
+	if ((bool)(fc->sb_flags & SB_RDONLY) == sb_rdonly(sb))
 		return 0;
-	if (*flags & SB_RDONLY) {
+	if (fc->sb_flags & SB_RDONLY) {
 		if (ms->s_state & MINIX_VALID_FS ||
 		    !(sbi->s_mount_state & MINIX_VALID_FS))
 			return 0;
@@ -170,7 +170,7 @@ static bool minix_check_superblock(struct super_block *sb)
 	return true;
 }
 
-static int minix_fill_super(struct super_block *s, void *data, int silent)
+static int minix_fill_super(struct super_block *s, struct fs_context *fc)
 {
 	struct buffer_head *bh;
 	struct buffer_head **map;
@@ -180,6 +180,7 @@ static int minix_fill_super(struct super_block *s, void *data, int silent)
 	struct inode *root_inode;
 	struct minix_sb_info *sbi;
 	int ret = -EINVAL;
+	int silent = fc->sb_flags & SB_SILENT;
 
 	sbi = kzalloc(sizeof(struct minix_sb_info), GFP_KERNEL);
 	if (!sbi)
@@ -371,6 +372,23 @@ out:
 	return ret;
 }
 
+static int minix_get_tree(struct fs_context *fc)
+{
+	 return get_tree_bdev(fc, minix_fill_super);
+}
+
+static const struct fs_context_operations minix_context_ops = {
+	.get_tree	= minix_get_tree,
+	.reconfigure	= minix_reconfigure,
+};
+
+static int minix_init_fs_context(struct fs_context *fc)
+{
+	fc->ops = &minix_context_ops;
+
+	return 0;
+}
+
 static int minix_statfs(struct dentry *dentry, struct kstatfs *buf)
 {
 	struct super_block *sb = dentry->d_sb;
@@ -680,18 +698,12 @@ void minix_truncate(struct inode * inode)
 		V2_minix_truncate(inode);
 }
 
-static struct dentry *minix_mount(struct file_system_type *fs_type,
-	int flags, const char *dev_name, void *data)
-{
-	return mount_bdev(fs_type, flags, dev_name, data, minix_fill_super);
-}
-
 static struct file_system_type minix_fs_type = {
-	.owner		= THIS_MODULE,
-	.name		= "minix",
-	.mount		= minix_mount,
-	.kill_sb	= kill_block_super,
-	.fs_flags	= FS_REQUIRES_DEV,
+	.owner			= THIS_MODULE,
+	.name			= "minix",
+	.kill_sb		= kill_block_super,
+	.fs_flags		= FS_REQUIRES_DEV,
+	.init_fs_context	= minix_init_fs_context,
 };
 MODULE_ALIAS_FS("minix");
 
diff --git a/fs/minix/namei.c b/fs/minix/namei.c
index d6031acc34..a944a0f17b 100644
--- a/fs/minix/namei.c
+++ b/fs/minix/namei.c
@@ -213,8 +213,7 @@ static int minix_rename(struct mnt_idmap *idmap,
 		if (!new_de)
 			goto out_dir;
 		err = minix_set_link(new_de, new_page, old_inode);
-		kunmap(new_page);
-		put_page(new_page);
+		unmap_and_put_page(new_page, new_de);
 		if (err)
 			goto out_dir;
 		inode_set_ctime_current(new_inode);
diff --git a/fs/mnt_idmapping.c b/fs/mnt_idmapping.c
index 64c5205e2b..3c60f1eaca 100644
--- a/fs/mnt_idmapping.c
+++ b/fs/mnt_idmapping.c
@@ -214,7 +214,7 @@ static int copy_mnt_idmap(struct uid_gid_map *map_from,
 	 * anything at all.
 	 */
 	if (nr_extents == 0)
-		return 0;
+		return -EINVAL;
 
 	/*
 	 * Here we know that nr_extents is greater than zero which means
diff --git a/fs/mpage.c b/fs/mpage.c
index 738882e076..fa8b99a199 100644
--- a/fs/mpage.c
+++ b/fs/mpage.c
@@ -605,6 +605,7 @@ alloc_new:
 				GFP_NOFS);
 		bio->bi_iter.bi_sector = first_block << (blkbits - 9);
 		wbc_init_bio(wbc, bio);
+		bio->bi_write_hint = inode->i_write_hint;
 	}
 
 	/*
diff --git a/fs/namei.c b/fs/namei.c
index 9342fa6a38..1e05a0f3f0 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -17,8 +17,8 @@
 
 #include <linux/init.h>
 #include <linux/export.h>
-#include <linux/kernel.h>
 #include <linux/slab.h>
+#include <linux/wordpart.h>
 #include <linux/fs.h>
 #include <linux/filelock.h>
 #include <linux/namei.h>
@@ -27,7 +27,6 @@
 #include <linux/fsnotify.h>
 #include <linux/personality.h>
 #include <linux/security.h>
-#include <linux/ima.h>
 #include <linux/syscalls.h>
 #include <linux/mount.h>
 #include <linux/audit.h>
@@ -2423,6 +2422,14 @@ static const char *path_init(struct nameidata *nd, unsigned flags)
 		if (!f.file)
 			return ERR_PTR(-EBADF);
 
+		if (flags & LOOKUP_LINKAT_EMPTY) {
+			if (f.file->f_cred != current_cred() &&
+			    !ns_capable(f.file->f_cred->user_ns, CAP_DAC_READ_SEARCH)) {
+				fdput(f);
+				return ERR_PTR(-ENOENT);
+			}
+		}
+
 		dentry = f.file->f_path.dentry;
 
 		if (*s && unlikely(!d_can_lookup(dentry))) {
@@ -2680,10 +2687,8 @@ static int lookup_one_common(struct mnt_idmap *idmap,
 	if (!len)
 		return -EACCES;
 
-	if (unlikely(name[0] == '.')) {
-		if (len < 2 || (len == 2 && name[1] == '.'))
-			return -EACCES;
-	}
+	if (is_dot_dotdot(name, len))
+		return -EACCES;
 
 	while (len--) {
 		unsigned int c = *(const unsigned char *)name++;
@@ -3567,8 +3572,12 @@ static const char *open_last_lookups(struct nameidata *nd,
 	else
 		inode_lock_shared(dir->d_inode);
 	dentry = lookup_open(nd, file, op, got_write);
-	if (!IS_ERR(dentry) && (file->f_mode & FMODE_CREATED))
-		fsnotify_create(dir->d_inode, dentry);
+	if (!IS_ERR(dentry)) {
+		if (file->f_mode & FMODE_CREATED)
+			fsnotify_create(dir->d_inode, dentry);
+		if (file->f_mode & FMODE_OPENED)
+			fsnotify_open(file);
+	}
 	if (open_flag & O_CREAT)
 		inode_unlock(dir->d_inode);
 	else
@@ -3644,7 +3653,7 @@ static int do_open(struct nameidata *nd,
 	if (!error && !(file->f_mode & FMODE_OPENED))
 		error = vfs_open(&nd->path, file);
 	if (!error)
-		error = ima_file_check(file, op->acc_mode);
+		error = security_file_post_open(file, op->acc_mode);
 	if (!error && do_truncate)
 		error = handle_truncate(idmap, file);
 	if (unlikely(error > 0)) {
@@ -3671,9 +3680,9 @@ static int do_open(struct nameidata *nd,
  * On non-idmapped mounts or if permission checking is to be performed on the
  * raw inode simply pass @nop_mnt_idmap.
  */
-static int vfs_tmpfile(struct mnt_idmap *idmap,
-		       const struct path *parentpath,
-		       struct file *file, umode_t mode)
+int vfs_tmpfile(struct mnt_idmap *idmap,
+		const struct path *parentpath,
+		struct file *file, umode_t mode)
 {
 	struct dentry *child;
 	struct inode *dir = d_inode(parentpath->dentry);
@@ -3695,6 +3704,8 @@ static int vfs_tmpfile(struct mnt_idmap *idmap,
 	mode = vfs_prepare_mode(idmap, dir, mode, mode, mode);
 	error = dir->i_op->tmpfile(idmap, dir, file, mode);
 	dput(child);
+	if (file->f_mode & FMODE_OPENED)
+		fsnotify_open(file);
 	if (error)
 		return error;
 	/* Don't check for other permissions, the inode was just created */
@@ -3707,7 +3718,7 @@ static int vfs_tmpfile(struct mnt_idmap *idmap,
 		inode->i_state |= I_LINKABLE;
 		spin_unlock(&inode->i_lock);
 	}
-	ima_post_create_tmpfile(idmap, inode);
+	security_inode_post_create_tmpfile(idmap, inode);
 	return 0;
 }
 
@@ -4054,7 +4065,7 @@ retry:
 			error = vfs_create(idmap, path.dentry->d_inode,
 					   dentry, mode, true);
 			if (!error)
-				ima_post_path_mknod(idmap, dentry);
+				security_path_post_mknod(idmap, dentry);
 			break;
 		case S_IFCHR: case S_IFBLK:
 			error = vfs_mknod(idmap, path.dentry->d_inode,
@@ -4644,14 +4655,13 @@ int do_linkat(int olddfd, struct filename *old, int newdfd,
 		goto out_putnames;
 	}
 	/*
-	 * To use null names we require CAP_DAC_READ_SEARCH
+	 * To use null names we require CAP_DAC_READ_SEARCH or
+	 * that the open-time creds of the dfd matches current.
 	 * This ensures that not everyone will be able to create
-	 * handlink using the passed filedescriptor.
+	 * a hardlink using the passed file descriptor.
 	 */
-	if (flags & AT_EMPTY_PATH && !capable(CAP_DAC_READ_SEARCH)) {
-		error = -ENOENT;
-		goto out_putnames;
-	}
+	if (flags & AT_EMPTY_PATH)
+		how |= LOOKUP_LINKAT_EMPTY;
 
 	if (flags & AT_SYMLINK_FOLLOW)
 		how |= LOOKUP_FOLLOW;
diff --git a/fs/netfs/Makefile b/fs/netfs/Makefile
index d4d1d79981..8e6781e0b1 100644
--- a/fs/netfs/Makefile
+++ b/fs/netfs/Makefile
@@ -11,7 +11,8 @@ netfs-y := \
 	main.o \
 	misc.o \
 	objects.o \
-	output.o
+	write_collect.o \
+	write_issue.o
 
 netfs-$(CONFIG_NETFS_STATS) += stats.o
 
diff --git a/fs/netfs/buffered_read.c b/fs/netfs/buffered_read.c
index 3298c29b55..4c0401dbbf 100644
--- a/fs/netfs/buffered_read.c
+++ b/fs/netfs/buffered_read.c
@@ -10,8 +10,11 @@
 #include "internal.h"
 
 /*
- * Unlock the folios in a read operation.  We need to set PG_fscache on any
+ * Unlock the folios in a read operation.  We need to set PG_writeback on any
  * folios we're going to write back before we unlock them.
+ *
+ * Note that if the deprecated NETFS_RREQ_USE_PGPRIV2 is set then we use
+ * PG_private_2 and do a direct write to the cache from here instead.
  */
 void netfs_rreq_unlock_folios(struct netfs_io_request *rreq)
 {
@@ -48,14 +51,14 @@ void netfs_rreq_unlock_folios(struct netfs_io_request *rreq)
 	xas_for_each(&xas, folio, last_page) {
 		loff_t pg_end;
 		bool pg_failed = false;
-		bool folio_started;
+		bool wback_to_cache = false;
+		bool folio_started = false;
 
 		if (xas_retry(&xas, folio))
 			continue;
 
 		pg_end = folio_pos(folio) + folio_size(folio) - 1;
 
-		folio_started = false;
 		for (;;) {
 			loff_t sreq_end;
 
@@ -63,10 +66,16 @@ void netfs_rreq_unlock_folios(struct netfs_io_request *rreq)
 				pg_failed = true;
 				break;
 			}
-			if (!folio_started && test_bit(NETFS_SREQ_COPY_TO_CACHE, &subreq->flags)) {
-				trace_netfs_folio(folio, netfs_folio_trace_copy_to_cache);
-				folio_start_fscache(folio);
-				folio_started = true;
+			if (test_bit(NETFS_RREQ_USE_PGPRIV2, &rreq->flags)) {
+				if (!folio_started && test_bit(NETFS_SREQ_COPY_TO_CACHE,
+							       &subreq->flags)) {
+					trace_netfs_folio(folio, netfs_folio_trace_copy_to_cache);
+					folio_start_private_2(folio);
+					folio_started = true;
+				}
+			} else {
+				wback_to_cache |=
+					test_bit(NETFS_SREQ_COPY_TO_CACHE, &subreq->flags);
 			}
 			pg_failed |= subreq_failed;
 			sreq_end = subreq->start + subreq->len - 1;
@@ -98,12 +107,17 @@ void netfs_rreq_unlock_folios(struct netfs_io_request *rreq)
 				kfree(finfo);
 			}
 			folio_mark_uptodate(folio);
+			if (wback_to_cache && !WARN_ON_ONCE(folio_get_private(folio) != NULL)) {
+				trace_netfs_folio(folio, netfs_folio_trace_copy_to_cache);
+				folio_attach_private(folio, NETFS_FOLIO_COPY_TO_CACHE);
+				filemap_dirty_folio(folio->mapping, folio);
+			}
 		}
 
 		if (!test_bit(NETFS_RREQ_DONT_UNLOCK_FOLIOS, &rreq->flags)) {
 			if (folio->index == rreq->no_unlock_folio &&
 			    test_bit(NETFS_RREQ_NO_UNLOCK_FOLIO, &rreq->flags))
-				_debug("no unlock");
+				kdebug("no unlock");
 			else
 				folio_unlock(folio);
 		}
@@ -116,7 +130,9 @@ void netfs_rreq_unlock_folios(struct netfs_io_request *rreq)
 }
 
 static void netfs_cache_expand_readahead(struct netfs_io_request *rreq,
-					 loff_t *_start, size_t *_len, loff_t i_size)
+					 unsigned long long *_start,
+					 unsigned long long *_len,
+					 unsigned long long i_size)
 {
 	struct netfs_cache_resources *cres = &rreq->cache_resources;
 
@@ -188,7 +204,7 @@ void netfs_readahead(struct readahead_control *ractl)
 	struct netfs_inode *ctx = netfs_inode(ractl->mapping->host);
 	int ret;
 
-	_enter("%lx,%x", readahead_index(ractl), readahead_count(ractl));
+	kenter("%lx,%x", readahead_index(ractl), readahead_count(ractl));
 
 	if (readahead_count(ractl) == 0)
 		return;
@@ -252,7 +268,7 @@ int netfs_read_folio(struct file *file, struct folio *folio)
 	struct folio *sink = NULL;
 	int ret;
 
-	_enter("%lx", folio->index);
+	kenter("%lx", folio->index);
 
 	rreq = netfs_alloc_request(mapping, file,
 				   folio_file_pos(folio), folio_size(folio),
@@ -266,7 +282,7 @@ int netfs_read_folio(struct file *file, struct folio *folio)
 	if (ret == -ENOMEM || ret == -EINTR || ret == -ERESTARTSYS)
 		goto discard;
 
-	netfs_stat(&netfs_n_rh_readpage);
+	netfs_stat(&netfs_n_rh_read_folio);
 	trace_netfs_read(rreq, rreq->start, rreq->len, netfs_read_trace_readpage);
 
 	/* Set up the output buffer */
@@ -450,7 +466,7 @@ retry:
 	if (!netfs_is_cache_enabled(ctx) &&
 	    netfs_skip_folio_read(folio, pos, len, false)) {
 		netfs_stat(&netfs_n_rh_write_zskip);
-		goto have_folio_no_wait;
+		goto have_folio;
 	}
 
 	rreq = netfs_alloc_request(mapping, file,
@@ -491,12 +507,8 @@ retry:
 	netfs_put_request(rreq, false, netfs_rreq_trace_put_return);
 
 have_folio:
-	ret = folio_wait_fscache_killable(folio);
-	if (ret < 0)
-		goto error;
-have_folio_no_wait:
 	*_folio = folio;
-	_leave(" = 0");
+	kleave(" = 0");
 	return 0;
 
 error_put:
@@ -506,7 +518,7 @@ error:
 		folio_unlock(folio);
 		folio_put(folio);
 	}
-	_leave(" = %d", ret);
+	kleave(" = %d", ret);
 	return ret;
 }
 EXPORT_SYMBOL(netfs_write_begin);
@@ -524,7 +536,7 @@ int netfs_prefetch_for_write(struct file *file, struct folio *folio,
 	size_t flen = folio_size(folio);
 	int ret;
 
-	_enter("%zx @%llx", flen, start);
+	kenter("%zx @%llx", flen, start);
 
 	ret = -ENOMEM;
 
@@ -555,7 +567,7 @@ int netfs_prefetch_for_write(struct file *file, struct folio *folio,
 error_put:
 	netfs_put_request(rreq, false, netfs_rreq_trace_put_discard);
 error:
-	_leave(" = %d", ret);
+	kleave(" = %d", ret);
 	return ret;
 }
 
diff --git a/fs/netfs/buffered_write.c b/fs/netfs/buffered_write.c
index 267b622d92..ecbc99ec7d 100644
--- a/fs/netfs/buffered_write.c
+++ b/fs/netfs/buffered_write.c
@@ -1,5 +1,5 @@
 // SPDX-License-Identifier: GPL-2.0-only
-/* Network filesystem high-level write support.
+/* Network filesystem high-level buffered write support.
  *
  * Copyright (C) 2023 Red Hat, Inc. All Rights Reserved.
  * Written by David Howells (dhowells@redhat.com)
@@ -26,25 +26,15 @@ enum netfs_how_to_modify {
 	NETFS_FLUSH_CONTENT,		/* Flush incompatible content. */
 };
 
-static void netfs_cleanup_buffered_write(struct netfs_io_request *wreq);
-
 static void netfs_set_group(struct folio *folio, struct netfs_group *netfs_group)
 {
-	if (netfs_group && !folio_get_private(folio))
-		folio_attach_private(folio, netfs_get_group(netfs_group));
-}
+	void *priv = folio_get_private(folio);
 
-#if IS_ENABLED(CONFIG_FSCACHE)
-static void netfs_folio_start_fscache(bool caching, struct folio *folio)
-{
-	if (caching)
-		folio_start_fscache(folio);
-}
-#else
-static void netfs_folio_start_fscache(bool caching, struct folio *folio)
-{
+	if (netfs_group && (!priv || priv == NETFS_FOLIO_COPY_TO_CACHE))
+		folio_attach_private(folio, netfs_get_group(netfs_group));
+	else if (!netfs_group && priv == NETFS_FOLIO_COPY_TO_CACHE)
+		folio_detach_private(folio);
 }
-#endif
 
 /*
  * Decide how we should modify a folio.  We might be attempting to do
@@ -63,11 +53,12 @@ static enum netfs_how_to_modify netfs_how_to_modify(struct netfs_inode *ctx,
 						    bool maybe_trouble)
 {
 	struct netfs_folio *finfo = netfs_folio_info(folio);
+	struct netfs_group *group = netfs_folio_group(folio);
 	loff_t pos = folio_file_pos(folio);
 
-	_enter("");
+	kenter("");
 
-	if (netfs_folio_group(folio) != netfs_group)
+	if (group != netfs_group && group != NETFS_FOLIO_COPY_TO_CACHE)
 		return NETFS_FLUSH_CONTENT;
 
 	if (folio_test_uptodate(folio))
@@ -81,16 +72,12 @@ static enum netfs_how_to_modify netfs_how_to_modify(struct netfs_inode *ctx,
 
 	if (file->f_mode & FMODE_READ)
 		goto no_write_streaming;
-	if (test_bit(NETFS_ICTX_NO_WRITE_STREAMING, &ctx->flags))
-		goto no_write_streaming;
 
 	if (netfs_is_cache_enabled(ctx)) {
 		/* We don't want to get a streaming write on a file that loses
 		 * caching service temporarily because the backing store got
 		 * culled.
 		 */
-		if (!test_bit(NETFS_ICTX_NO_WRITE_STREAMING, &ctx->flags))
-			set_bit(NETFS_ICTX_NO_WRITE_STREAMING, &ctx->flags);
 		goto no_write_streaming;
 	}
 
@@ -130,6 +117,37 @@ static struct folio *netfs_grab_folio_for_write(struct address_space *mapping,
 				   mapping_gfp_mask(mapping));
 }
 
+/*
+ * Update i_size and estimate the update to i_blocks to reflect the additional
+ * data written into the pagecache until we can find out from the server what
+ * the values actually are.
+ */
+static void netfs_update_i_size(struct netfs_inode *ctx, struct inode *inode,
+				loff_t i_size, loff_t pos, size_t copied)
+{
+	blkcnt_t add;
+	size_t gap;
+
+	if (ctx->ops->update_i_size) {
+		ctx->ops->update_i_size(inode, pos);
+		return;
+	}
+
+	i_size_write(inode, pos);
+#if IS_ENABLED(CONFIG_FSCACHE)
+	fscache_update_cookie(ctx->cache, NULL, &pos);
+#endif
+
+	gap = SECTOR_SIZE - (i_size & (SECTOR_SIZE - 1));
+	if (copied > gap) {
+		add = DIV_ROUND_UP(copied - gap, SECTOR_SIZE);
+
+		inode->i_blocks = min_t(blkcnt_t,
+					DIV_ROUND_UP(pos, SECTOR_SIZE),
+					inode->i_blocks + add);
+	}
+}
+
 /**
  * netfs_perform_write - Copy data into the pagecache.
  * @iocb: The operation parameters
@@ -160,10 +178,10 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
 	};
 	struct netfs_io_request *wreq = NULL;
 	struct netfs_folio *finfo;
-	struct folio *folio;
+	struct folio *folio, *writethrough = NULL;
 	enum netfs_how_to_modify howto;
 	enum netfs_folio_trace trace;
-	unsigned int bdp_flags = (iocb->ki_flags & IOCB_SYNC) ? 0: BDP_ASYNC;
+	unsigned int bdp_flags = (iocb->ki_flags & IOCB_NOWAIT) ? BDP_ASYNC : 0;
 	ssize_t written = 0, ret, ret2;
 	loff_t i_size, pos = iocb->ki_pos, from, to;
 	size_t max_chunk = PAGE_SIZE << MAX_PAGECACHE_ORDER;
@@ -189,7 +207,9 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
 		}
 		if (!is_sync_kiocb(iocb))
 			wreq->iocb = iocb;
-		wreq->cleanup = netfs_cleanup_buffered_write;
+		netfs_stat(&netfs_n_wh_writethrough);
+	} else {
+		netfs_stat(&netfs_n_wh_buffered_write);
 	}
 
 	do {
@@ -230,6 +250,16 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
 		offset = pos & (flen - 1);
 		part = min_t(size_t, flen - offset, part);
 
+		/* Wait for writeback to complete.  The writeback engine owns
+		 * the info in folio->private and may change it until it
+		 * removes the WB mark.
+		 */
+		if (folio_get_private(folio) &&
+		    folio_wait_writeback_killable(folio)) {
+			ret = written ? -EINTR : -ERESTARTSYS;
+			goto error_folio_unlock;
+		}
+
 		if (signal_pending(current)) {
 			ret = written ? -EINTR : -ERESTARTSYS;
 			goto error_folio_unlock;
@@ -242,12 +272,12 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
 		 */
 		howto = netfs_how_to_modify(ctx, file, folio, netfs_group,
 					    flen, offset, part, maybe_trouble);
-		_debug("howto %u", howto);
+		kdebug("howto %u", howto);
 		switch (howto) {
 		case NETFS_JUST_PREFETCH:
 			ret = netfs_prefetch_for_write(file, folio, offset, part);
 			if (ret < 0) {
-				_debug("prefetch = %zd", ret);
+				kdebug("prefetch = %zd", ret);
 				goto error_folio_unlock;
 			}
 			break;
@@ -304,6 +334,7 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
 				maybe_trouble = true;
 				iov_iter_revert(iter, copied);
 				copied = 0;
+				folio_unlock(folio);
 				goto retry;
 			}
 			netfs_set_group(folio, netfs_group);
@@ -351,41 +382,22 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
 		trace_netfs_folio(folio, trace);
 
 		/* Update the inode size if we moved the EOF marker */
-		i_size = i_size_read(inode);
 		pos += copied;
-		if (pos > i_size) {
-			if (ctx->ops->update_i_size) {
-				ctx->ops->update_i_size(inode, pos);
-			} else {
-				i_size_write(inode, pos);
-#if IS_ENABLED(CONFIG_FSCACHE)
-				fscache_update_cookie(ctx->cache, NULL, &pos);
-#endif
-			}
-		}
+		i_size = i_size_read(inode);
+		if (pos > i_size)
+			netfs_update_i_size(ctx, inode, i_size, pos, copied);
 		written += copied;
 
 		if (likely(!wreq)) {
 			folio_mark_dirty(folio);
+			folio_unlock(folio);
 		} else {
-			if (folio_test_dirty(folio))
-				/* Sigh.  mmap. */
-				folio_clear_dirty_for_io(folio);
-			/* We make multiple writes to the folio... */
-			if (!folio_test_writeback(folio)) {
-				folio_wait_fscache(folio);
-				folio_start_writeback(folio);
-				folio_start_fscache(folio);
-				if (wreq->iter.count == 0)
-					trace_netfs_folio(folio, netfs_folio_trace_wthru);
-				else
-					trace_netfs_folio(folio, netfs_folio_trace_wthru_plus);
-			}
-			netfs_advance_writethrough(wreq, copied,
-						   offset + copied == flen);
+			netfs_advance_writethrough(wreq, &wbc, folio, copied,
+						   offset + copied == flen,
+						   &writethrough);
+			/* Folio unlocked */
 		}
 	retry:
-		folio_unlock(folio);
 		folio_put(folio);
 		folio = NULL;
 
@@ -393,8 +405,11 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
 	} while (iov_iter_count(iter));
 
 out:
+	if (likely(written) && ctx->ops->post_modify)
+		ctx->ops->post_modify(inode);
+
 	if (unlikely(wreq)) {
-		ret2 = netfs_end_writethrough(wreq, iocb);
+		ret2 = netfs_end_writethrough(wreq, &wbc, writethrough);
 		wbc_detach_inode(&wbc);
 		if (ret2 == -EIOCBQUEUED)
 			return ret2;
@@ -403,7 +418,7 @@ out:
 	}
 
 	iocb->ki_pos += written;
-	_leave(" = %zd [%zd]", written, ret);
+	kleave(" = %zd [%zd]", written, ret);
 	return written ? written : ret;
 
 error_folio_unlock:
@@ -476,7 +491,7 @@ ssize_t netfs_file_write_iter(struct kiocb *iocb, struct iov_iter *from)
 	struct netfs_inode *ictx = netfs_inode(inode);
 	ssize_t ret;
 
-	_enter("%llx,%zx,%llx", iocb->ki_pos, iov_iter_count(from), i_size_read(inode));
+	kenter("%llx,%zx,%llx", iocb->ki_pos, iov_iter_count(from), i_size_read(inode));
 
 	if (!iov_iter_count(from))
 		return 0;
@@ -505,21 +520,31 @@ EXPORT_SYMBOL(netfs_file_write_iter);
  */
 vm_fault_t netfs_page_mkwrite(struct vm_fault *vmf, struct netfs_group *netfs_group)
 {
+	struct netfs_group *group;
 	struct folio *folio = page_folio(vmf->page);
 	struct file *file = vmf->vma->vm_file;
+	struct address_space *mapping = file->f_mapping;
 	struct inode *inode = file_inode(file);
+	struct netfs_inode *ictx = netfs_inode(inode);
 	vm_fault_t ret = VM_FAULT_RETRY;
 	int err;
 
-	_enter("%lx", folio->index);
+	kenter("%lx", folio->index);
 
 	sb_start_pagefault(inode->i_sb);
 
-	if (folio_wait_writeback_killable(folio))
+	if (folio_lock_killable(folio) < 0)
+		goto out;
+	if (folio->mapping != mapping) {
+		folio_unlock(folio);
+		ret = VM_FAULT_NOPAGE;
 		goto out;
+	}
 
-	if (folio_lock_killable(folio) < 0)
+	if (folio_wait_writeback_killable(folio)) {
+		ret = VM_FAULT_LOCKED;
 		goto out;
+	}
 
 	/* Can we see a streaming write here? */
 	if (WARN_ON(!folio_test_uptodate(folio))) {
@@ -527,11 +552,12 @@ vm_fault_t netfs_page_mkwrite(struct vm_fault *vmf, struct netfs_group *netfs_gr
 		goto out;
 	}
 
-	if (netfs_folio_group(folio) != netfs_group) {
+	group = netfs_folio_group(folio);
+	if (group != netfs_group && group != NETFS_FOLIO_COPY_TO_CACHE) {
 		folio_unlock(folio);
-		err = filemap_fdatawait_range(inode->i_mapping,
-					      folio_pos(folio),
-					      folio_pos(folio) + folio_size(folio));
+		err = filemap_fdatawrite_range(mapping,
+					       folio_pos(folio),
+					       folio_pos(folio) + folio_size(folio));
 		switch (err) {
 		case 0:
 			ret = VM_FAULT_RETRY;
@@ -551,708 +577,11 @@ vm_fault_t netfs_page_mkwrite(struct vm_fault *vmf, struct netfs_group *netfs_gr
 		trace_netfs_folio(folio, netfs_folio_trace_mkwrite);
 	netfs_set_group(folio, netfs_group);
 	file_update_time(file);
+	if (ictx->ops->post_modify)
+		ictx->ops->post_modify(inode);
 	ret = VM_FAULT_LOCKED;
 out:
 	sb_end_pagefault(inode->i_sb);
 	return ret;
 }
 EXPORT_SYMBOL(netfs_page_mkwrite);
-
-/*
- * Kill all the pages in the given range
- */
-static void netfs_kill_pages(struct address_space *mapping,
-			     loff_t start, loff_t len)
-{
-	struct folio *folio;
-	pgoff_t index = start / PAGE_SIZE;
-	pgoff_t last = (start + len - 1) / PAGE_SIZE, next;
-
-	_enter("%llx-%llx", start, start + len - 1);
-
-	do {
-		_debug("kill %lx (to %lx)", index, last);
-
-		folio = filemap_get_folio(mapping, index);
-		if (IS_ERR(folio)) {
-			next = index + 1;
-			continue;
-		}
-
-		next = folio_next_index(folio);
-
-		trace_netfs_folio(folio, netfs_folio_trace_kill);
-		folio_clear_uptodate(folio);
-		if (folio_test_fscache(folio))
-			folio_end_fscache(folio);
-		folio_end_writeback(folio);
-		folio_lock(folio);
-		generic_error_remove_folio(mapping, folio);
-		folio_unlock(folio);
-		folio_put(folio);
-
-	} while (index = next, index <= last);
-
-	_leave("");
-}
-
-/*
- * Redirty all the pages in a given range.
- */
-static void netfs_redirty_pages(struct address_space *mapping,
-				loff_t start, loff_t len)
-{
-	struct folio *folio;
-	pgoff_t index = start / PAGE_SIZE;
-	pgoff_t last = (start + len - 1) / PAGE_SIZE, next;
-
-	_enter("%llx-%llx", start, start + len - 1);
-
-	do {
-		_debug("redirty %llx @%llx", len, start);
-
-		folio = filemap_get_folio(mapping, index);
-		if (IS_ERR(folio)) {
-			next = index + 1;
-			continue;
-		}
-
-		next = folio_next_index(folio);
-		trace_netfs_folio(folio, netfs_folio_trace_redirty);
-		filemap_dirty_folio(mapping, folio);
-		if (folio_test_fscache(folio))
-			folio_end_fscache(folio);
-		folio_end_writeback(folio);
-		folio_put(folio);
-	} while (index = next, index <= last);
-
-	balance_dirty_pages_ratelimited(mapping);
-
-	_leave("");
-}
-
-/*
- * Completion of write to server
- */
-static void netfs_pages_written_back(struct netfs_io_request *wreq)
-{
-	struct address_space *mapping = wreq->mapping;
-	struct netfs_folio *finfo;
-	struct netfs_group *group = NULL;
-	struct folio *folio;
-	pgoff_t last;
-	int gcount = 0;
-
-	XA_STATE(xas, &mapping->i_pages, wreq->start / PAGE_SIZE);
-
-	_enter("%llx-%llx", wreq->start, wreq->start + wreq->len);
-
-	rcu_read_lock();
-
-	last = (wreq->start + wreq->len - 1) / PAGE_SIZE;
-	xas_for_each(&xas, folio, last) {
-		WARN(!folio_test_writeback(folio),
-		     "bad %zx @%llx page %lx %lx\n",
-		     wreq->len, wreq->start, folio->index, last);
-
-		if ((finfo = netfs_folio_info(folio))) {
-			/* Streaming writes cannot be redirtied whilst under
-			 * writeback, so discard the streaming record.
-			 */
-			folio_detach_private(folio);
-			group = finfo->netfs_group;
-			gcount++;
-			trace_netfs_folio(folio, netfs_folio_trace_clear_s);
-			kfree(finfo);
-		} else if ((group = netfs_folio_group(folio))) {
-			/* Need to detach the group pointer if the page didn't
-			 * get redirtied.  If it has been redirtied, then it
-			 * must be within the same group.
-			 */
-			if (folio_test_dirty(folio)) {
-				trace_netfs_folio(folio, netfs_folio_trace_redirtied);
-				goto end_wb;
-			}
-			if (folio_trylock(folio)) {
-				if (!folio_test_dirty(folio)) {
-					folio_detach_private(folio);
-					gcount++;
-					trace_netfs_folio(folio, netfs_folio_trace_clear_g);
-				} else {
-					trace_netfs_folio(folio, netfs_folio_trace_redirtied);
-				}
-				folio_unlock(folio);
-				goto end_wb;
-			}
-
-			xas_pause(&xas);
-			rcu_read_unlock();
-			folio_lock(folio);
-			if (!folio_test_dirty(folio)) {
-				folio_detach_private(folio);
-				gcount++;
-				trace_netfs_folio(folio, netfs_folio_trace_clear_g);
-			} else {
-				trace_netfs_folio(folio, netfs_folio_trace_redirtied);
-			}
-			folio_unlock(folio);
-			rcu_read_lock();
-		} else {
-			trace_netfs_folio(folio, netfs_folio_trace_clear);
-		}
-	end_wb:
-		if (folio_test_fscache(folio))
-			folio_end_fscache(folio);
-		xas_advance(&xas, folio_next_index(folio) - 1);
-		folio_end_writeback(folio);
-	}
-
-	rcu_read_unlock();
-	netfs_put_group_many(group, gcount);
-	_leave("");
-}
-
-/*
- * Deal with the disposition of the folios that are under writeback to close
- * out the operation.
- */
-static void netfs_cleanup_buffered_write(struct netfs_io_request *wreq)
-{
-	struct address_space *mapping = wreq->mapping;
-
-	_enter("");
-
-	switch (wreq->error) {
-	case 0:
-		netfs_pages_written_back(wreq);
-		break;
-
-	default:
-		pr_notice("R=%08x Unexpected error %d\n", wreq->debug_id, wreq->error);
-		fallthrough;
-	case -EACCES:
-	case -EPERM:
-	case -ENOKEY:
-	case -EKEYEXPIRED:
-	case -EKEYREJECTED:
-	case -EKEYREVOKED:
-	case -ENETRESET:
-	case -EDQUOT:
-	case -ENOSPC:
-		netfs_redirty_pages(mapping, wreq->start, wreq->len);
-		break;
-
-	case -EROFS:
-	case -EIO:
-	case -EREMOTEIO:
-	case -EFBIG:
-	case -ENOENT:
-	case -ENOMEDIUM:
-	case -ENXIO:
-		netfs_kill_pages(mapping, wreq->start, wreq->len);
-		break;
-	}
-
-	if (wreq->error)
-		mapping_set_error(mapping, wreq->error);
-	if (wreq->netfs_ops->done)
-		wreq->netfs_ops->done(wreq);
-}
-
-/*
- * Extend the region to be written back to include subsequent contiguously
- * dirty pages if possible, but don't sleep while doing so.
- *
- * If this page holds new content, then we can include filler zeros in the
- * writeback.
- */
-static void netfs_extend_writeback(struct address_space *mapping,
-				   struct netfs_group *group,
-				   struct xa_state *xas,
-				   long *_count,
-				   loff_t start,
-				   loff_t max_len,
-				   bool caching,
-				   size_t *_len,
-				   size_t *_top)
-{
-	struct netfs_folio *finfo;
-	struct folio_batch fbatch;
-	struct folio *folio;
-	unsigned int i;
-	pgoff_t index = (start + *_len) / PAGE_SIZE;
-	size_t len;
-	void *priv;
-	bool stop = true;
-
-	folio_batch_init(&fbatch);
-
-	do {
-		/* Firstly, we gather up a batch of contiguous dirty pages
-		 * under the RCU read lock - but we can't clear the dirty flags
-		 * there if any of those pages are mapped.
-		 */
-		rcu_read_lock();
-
-		xas_for_each(xas, folio, ULONG_MAX) {
-			stop = true;
-			if (xas_retry(xas, folio))
-				continue;
-			if (xa_is_value(folio))
-				break;
-			if (folio->index != index) {
-				xas_reset(xas);
-				break;
-			}
-
-			if (!folio_try_get_rcu(folio)) {
-				xas_reset(xas);
-				continue;
-			}
-
-			/* Has the folio moved or been split? */
-			if (unlikely(folio != xas_reload(xas))) {
-				folio_put(folio);
-				xas_reset(xas);
-				break;
-			}
-
-			if (!folio_trylock(folio)) {
-				folio_put(folio);
-				xas_reset(xas);
-				break;
-			}
-			if (!folio_test_dirty(folio) ||
-			    folio_test_writeback(folio) ||
-			    folio_test_fscache(folio)) {
-				folio_unlock(folio);
-				folio_put(folio);
-				xas_reset(xas);
-				break;
-			}
-
-			stop = false;
-			len = folio_size(folio);
-			priv = folio_get_private(folio);
-			if ((const struct netfs_group *)priv != group) {
-				stop = true;
-				finfo = netfs_folio_info(folio);
-				if (finfo->netfs_group != group ||
-				    finfo->dirty_offset > 0) {
-					folio_unlock(folio);
-					folio_put(folio);
-					xas_reset(xas);
-					break;
-				}
-				len = finfo->dirty_len;
-			}
-
-			*_top += folio_size(folio);
-			index += folio_nr_pages(folio);
-			*_count -= folio_nr_pages(folio);
-			*_len += len;
-			if (*_len >= max_len || *_count <= 0)
-				stop = true;
-
-			if (!folio_batch_add(&fbatch, folio))
-				break;
-			if (stop)
-				break;
-		}
-
-		xas_pause(xas);
-		rcu_read_unlock();
-
-		/* Now, if we obtained any folios, we can shift them to being
-		 * writable and mark them for caching.
-		 */
-		if (!folio_batch_count(&fbatch))
-			break;
-
-		for (i = 0; i < folio_batch_count(&fbatch); i++) {
-			folio = fbatch.folios[i];
-			trace_netfs_folio(folio, netfs_folio_trace_store_plus);
-
-			if (!folio_clear_dirty_for_io(folio))
-				BUG();
-			folio_start_writeback(folio);
-			netfs_folio_start_fscache(caching, folio);
-			folio_unlock(folio);
-		}
-
-		folio_batch_release(&fbatch);
-		cond_resched();
-	} while (!stop);
-}
-
-/*
- * Synchronously write back the locked page and any subsequent non-locked dirty
- * pages.
- */
-static ssize_t netfs_write_back_from_locked_folio(struct address_space *mapping,
-						  struct writeback_control *wbc,
-						  struct netfs_group *group,
-						  struct xa_state *xas,
-						  struct folio *folio,
-						  unsigned long long start,
-						  unsigned long long end)
-{
-	struct netfs_io_request *wreq;
-	struct netfs_folio *finfo;
-	struct netfs_inode *ctx = netfs_inode(mapping->host);
-	unsigned long long i_size = i_size_read(&ctx->inode);
-	size_t len, max_len;
-	bool caching = netfs_is_cache_enabled(ctx);
-	long count = wbc->nr_to_write;
-	int ret;
-
-	_enter(",%lx,%llx-%llx,%u", folio->index, start, end, caching);
-
-	wreq = netfs_alloc_request(mapping, NULL, start, folio_size(folio),
-				   NETFS_WRITEBACK);
-	if (IS_ERR(wreq)) {
-		folio_unlock(folio);
-		return PTR_ERR(wreq);
-	}
-
-	if (!folio_clear_dirty_for_io(folio))
-		BUG();
-	folio_start_writeback(folio);
-	netfs_folio_start_fscache(caching, folio);
-
-	count -= folio_nr_pages(folio);
-
-	/* Find all consecutive lockable dirty pages that have contiguous
-	 * written regions, stopping when we find a page that is not
-	 * immediately lockable, is not dirty or is missing, or we reach the
-	 * end of the range.
-	 */
-	trace_netfs_folio(folio, netfs_folio_trace_store);
-
-	len = wreq->len;
-	finfo = netfs_folio_info(folio);
-	if (finfo) {
-		start += finfo->dirty_offset;
-		if (finfo->dirty_offset + finfo->dirty_len != len) {
-			len = finfo->dirty_len;
-			goto cant_expand;
-		}
-		len = finfo->dirty_len;
-	}
-
-	if (start < i_size) {
-		/* Trim the write to the EOF; the extra data is ignored.  Also
-		 * put an upper limit on the size of a single storedata op.
-		 */
-		max_len = 65536 * 4096;
-		max_len = min_t(unsigned long long, max_len, end - start + 1);
-		max_len = min_t(unsigned long long, max_len, i_size - start);
-
-		if (len < max_len)
-			netfs_extend_writeback(mapping, group, xas, &count, start,
-					       max_len, caching, &len, &wreq->upper_len);
-	}
-
-cant_expand:
-	len = min_t(unsigned long long, len, i_size - start);
-
-	/* We now have a contiguous set of dirty pages, each with writeback
-	 * set; the first page is still locked at this point, but all the rest
-	 * have been unlocked.
-	 */
-	folio_unlock(folio);
-	wreq->start = start;
-	wreq->len = len;
-
-	if (start < i_size) {
-		_debug("write back %zx @%llx [%llx]", len, start, i_size);
-
-		/* Speculatively write to the cache.  We have to fix this up
-		 * later if the store fails.
-		 */
-		wreq->cleanup = netfs_cleanup_buffered_write;
-
-		iov_iter_xarray(&wreq->iter, ITER_SOURCE, &mapping->i_pages, start,
-				wreq->upper_len);
-		__set_bit(NETFS_RREQ_UPLOAD_TO_SERVER, &wreq->flags);
-		ret = netfs_begin_write(wreq, true, netfs_write_trace_writeback);
-		if (ret == 0 || ret == -EIOCBQUEUED)
-			wbc->nr_to_write -= len / PAGE_SIZE;
-	} else {
-		_debug("write discard %zx @%llx [%llx]", len, start, i_size);
-
-		/* The dirty region was entirely beyond the EOF. */
-		fscache_clear_page_bits(mapping, start, len, caching);
-		netfs_pages_written_back(wreq);
-		ret = 0;
-	}
-
-	netfs_put_request(wreq, false, netfs_rreq_trace_put_return);
-	_leave(" = 1");
-	return 1;
-}
-
-/*
- * Write a region of pages back to the server
- */
-static ssize_t netfs_writepages_begin(struct address_space *mapping,
-				      struct writeback_control *wbc,
-				      struct netfs_group *group,
-				      struct xa_state *xas,
-				      unsigned long long *_start,
-				      unsigned long long end)
-{
-	const struct netfs_folio *finfo;
-	struct folio *folio;
-	unsigned long long start = *_start;
-	ssize_t ret;
-	void *priv;
-	int skips = 0;
-
-	_enter("%llx,%llx,", start, end);
-
-search_again:
-	/* Find the first dirty page in the group. */
-	rcu_read_lock();
-
-	for (;;) {
-		folio = xas_find_marked(xas, end / PAGE_SIZE, PAGECACHE_TAG_DIRTY);
-		if (xas_retry(xas, folio) || xa_is_value(folio))
-			continue;
-		if (!folio)
-			break;
-
-		if (!folio_try_get_rcu(folio)) {
-			xas_reset(xas);
-			continue;
-		}
-
-		if (unlikely(folio != xas_reload(xas))) {
-			folio_put(folio);
-			xas_reset(xas);
-			continue;
-		}
-
-		/* Skip any dirty folio that's not in the group of interest. */
-		priv = folio_get_private(folio);
-		if ((const struct netfs_group *)priv != group) {
-			finfo = netfs_folio_info(folio);
-			if (finfo->netfs_group != group) {
-				folio_put(folio);
-				continue;
-			}
-		}
-
-		xas_pause(xas);
-		break;
-	}
-	rcu_read_unlock();
-	if (!folio)
-		return 0;
-
-	start = folio_pos(folio); /* May regress with THPs */
-
-	_debug("wback %lx", folio->index);
-
-	/* At this point we hold neither the i_pages lock nor the page lock:
-	 * the page may be truncated or invalidated (changing page->mapping to
-	 * NULL), or even swizzled back from swapper_space to tmpfs file
-	 * mapping
-	 */
-lock_again:
-	if (wbc->sync_mode != WB_SYNC_NONE) {
-		ret = folio_lock_killable(folio);
-		if (ret < 0)
-			return ret;
-	} else {
-		if (!folio_trylock(folio))
-			goto search_again;
-	}
-
-	if (folio->mapping != mapping ||
-	    !folio_test_dirty(folio)) {
-		start += folio_size(folio);
-		folio_unlock(folio);
-		goto search_again;
-	}
-
-	if (folio_test_writeback(folio) ||
-	    folio_test_fscache(folio)) {
-		folio_unlock(folio);
-		if (wbc->sync_mode != WB_SYNC_NONE) {
-			folio_wait_writeback(folio);
-#ifdef CONFIG_FSCACHE
-			folio_wait_fscache(folio);
-#endif
-			goto lock_again;
-		}
-
-		start += folio_size(folio);
-		if (wbc->sync_mode == WB_SYNC_NONE) {
-			if (skips >= 5 || need_resched()) {
-				ret = 0;
-				goto out;
-			}
-			skips++;
-		}
-		goto search_again;
-	}
-
-	ret = netfs_write_back_from_locked_folio(mapping, wbc, group, xas,
-						 folio, start, end);
-out:
-	if (ret > 0)
-		*_start = start + ret;
-	_leave(" = %zd [%llx]", ret, *_start);
-	return ret;
-}
-
-/*
- * Write a region of pages back to the server
- */
-static int netfs_writepages_region(struct address_space *mapping,
-				   struct writeback_control *wbc,
-				   struct netfs_group *group,
-				   unsigned long long *_start,
-				   unsigned long long end)
-{
-	ssize_t ret;
-
-	XA_STATE(xas, &mapping->i_pages, *_start / PAGE_SIZE);
-
-	do {
-		ret = netfs_writepages_begin(mapping, wbc, group, &xas,
-					     _start, end);
-		if (ret > 0 && wbc->nr_to_write > 0)
-			cond_resched();
-	} while (ret > 0 && wbc->nr_to_write > 0);
-
-	return ret > 0 ? 0 : ret;
-}
-
-/*
- * write some of the pending data back to the server
- */
-int netfs_writepages(struct address_space *mapping,
-		     struct writeback_control *wbc)
-{
-	struct netfs_group *group = NULL;
-	loff_t start, end;
-	int ret;
-
-	_enter("");
-
-	/* We have to be careful as we can end up racing with setattr()
-	 * truncating the pagecache since the caller doesn't take a lock here
-	 * to prevent it.
-	 */
-
-	if (wbc->range_cyclic && mapping->writeback_index) {
-		start = mapping->writeback_index * PAGE_SIZE;
-		ret = netfs_writepages_region(mapping, wbc, group,
-					      &start, LLONG_MAX);
-		if (ret < 0)
-			goto out;
-
-		if (wbc->nr_to_write <= 0) {
-			mapping->writeback_index = start / PAGE_SIZE;
-			goto out;
-		}
-
-		start = 0;
-		end = mapping->writeback_index * PAGE_SIZE;
-		mapping->writeback_index = 0;
-		ret = netfs_writepages_region(mapping, wbc, group, &start, end);
-		if (ret == 0)
-			mapping->writeback_index = start / PAGE_SIZE;
-	} else if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX) {
-		start = 0;
-		ret = netfs_writepages_region(mapping, wbc, group,
-					      &start, LLONG_MAX);
-		if (wbc->nr_to_write > 0 && ret == 0)
-			mapping->writeback_index = start / PAGE_SIZE;
-	} else {
-		start = wbc->range_start;
-		ret = netfs_writepages_region(mapping, wbc, group,
-					      &start, wbc->range_end);
-	}
-
-out:
-	_leave(" = %d", ret);
-	return ret;
-}
-EXPORT_SYMBOL(netfs_writepages);
-
-/*
- * Deal with the disposition of a laundered folio.
- */
-static void netfs_cleanup_launder_folio(struct netfs_io_request *wreq)
-{
-	if (wreq->error) {
-		pr_notice("R=%08x Laundering error %d\n", wreq->debug_id, wreq->error);
-		mapping_set_error(wreq->mapping, wreq->error);
-	}
-}
-
-/**
- * netfs_launder_folio - Clean up a dirty folio that's being invalidated
- * @folio: The folio to clean
- *
- * This is called to write back a folio that's being invalidated when an inode
- * is getting torn down.  Ideally, writepages would be used instead.
- */
-int netfs_launder_folio(struct folio *folio)
-{
-	struct netfs_io_request *wreq;
-	struct address_space *mapping = folio->mapping;
-	struct netfs_folio *finfo = netfs_folio_info(folio);
-	struct netfs_group *group = netfs_folio_group(folio);
-	struct bio_vec bvec;
-	unsigned long long i_size = i_size_read(mapping->host);
-	unsigned long long start = folio_pos(folio);
-	size_t offset = 0, len;
-	int ret = 0;
-
-	if (finfo) {
-		offset = finfo->dirty_offset;
-		start += offset;
-		len = finfo->dirty_len;
-	} else {
-		len = folio_size(folio);
-	}
-	len = min_t(unsigned long long, len, i_size - start);
-
-	wreq = netfs_alloc_request(mapping, NULL, start, len, NETFS_LAUNDER_WRITE);
-	if (IS_ERR(wreq)) {
-		ret = PTR_ERR(wreq);
-		goto out;
-	}
-
-	if (!folio_clear_dirty_for_io(folio))
-		goto out_put;
-
-	trace_netfs_folio(folio, netfs_folio_trace_launder);
-
-	_debug("launder %llx-%llx", start, start + len - 1);
-
-	/* Speculatively write to the cache.  We have to fix this up later if
-	 * the store fails.
-	 */
-	wreq->cleanup = netfs_cleanup_launder_folio;
-
-	bvec_set_folio(&bvec, folio, len, offset);
-	iov_iter_bvec(&wreq->iter, ITER_SOURCE, &bvec, 1, len);
-	__set_bit(NETFS_RREQ_UPLOAD_TO_SERVER, &wreq->flags);
-	ret = netfs_begin_write(wreq, true, netfs_write_trace_launder);
-
-out_put:
-	folio_detach_private(folio);
-	netfs_put_group(group);
-	kfree(finfo);
-	netfs_put_request(wreq, false, netfs_rreq_trace_put_return);
-out:
-	folio_wait_fscache(folio);
-	_leave(" = %d", ret);
-	return ret;
-}
-EXPORT_SYMBOL(netfs_launder_folio);
diff --git a/fs/netfs/direct_read.c b/fs/netfs/direct_read.c
index ad4370b393..b6debac620 100644
--- a/fs/netfs/direct_read.c
+++ b/fs/netfs/direct_read.c
@@ -26,14 +26,14 @@
  *
  * The caller must hold any appropriate locks.
  */
-static ssize_t netfs_unbuffered_read_iter_locked(struct kiocb *iocb, struct iov_iter *iter)
+ssize_t netfs_unbuffered_read_iter_locked(struct kiocb *iocb, struct iov_iter *iter)
 {
 	struct netfs_io_request *rreq;
 	ssize_t ret;
 	size_t orig_count = iov_iter_count(iter);
 	bool async = !is_sync_kiocb(iocb);
 
-	_enter("");
+	kenter("");
 
 	if (!orig_count)
 		return 0; /* Don't update atime */
@@ -98,6 +98,7 @@ out:
 		iov_iter_revert(iter, orig_count - iov_iter_count(iter));
 	return ret;
 }
+EXPORT_SYMBOL(netfs_unbuffered_read_iter_locked);
 
 /**
  * netfs_unbuffered_read_iter - Perform an unbuffered or direct I/O read
diff --git a/fs/netfs/direct_write.c b/fs/netfs/direct_write.c
index bee047e20f..792ef17bae 100644
--- a/fs/netfs/direct_write.c
+++ b/fs/netfs/direct_write.c
@@ -12,7 +12,7 @@
 static void netfs_cleanup_dio_write(struct netfs_io_request *wreq)
 {
 	struct inode *inode = wreq->inode;
-	unsigned long long end = wreq->start + wreq->len;
+	unsigned long long end = wreq->start + wreq->transferred;
 
 	if (!wreq->error &&
 	    i_size_read(inode) < end) {
@@ -27,16 +27,17 @@ static void netfs_cleanup_dio_write(struct netfs_io_request *wreq)
  * Perform an unbuffered write where we may have to do an RMW operation on an
  * encrypted file.  This can also be used for direct I/O writes.
  */
-static ssize_t netfs_unbuffered_write_iter_locked(struct kiocb *iocb, struct iov_iter *iter,
+ssize_t netfs_unbuffered_write_iter_locked(struct kiocb *iocb, struct iov_iter *iter,
 						  struct netfs_group *netfs_group)
 {
 	struct netfs_io_request *wreq;
 	unsigned long long start = iocb->ki_pos;
 	unsigned long long end = start + iov_iter_count(iter);
 	ssize_t ret, n;
+	size_t len = iov_iter_count(iter);
 	bool async = !is_sync_kiocb(iocb);
 
-	_enter("");
+	kenter("");
 
 	/* We're going to need a bounce buffer if what we transmit is going to
 	 * be different in some way to the source buffer, e.g. because it gets
@@ -44,15 +45,19 @@ static ssize_t netfs_unbuffered_write_iter_locked(struct kiocb *iocb, struct iov
 	 */
 	// TODO
 
-	_debug("uw %llx-%llx", start, end);
+	kdebug("uw %llx-%llx", start, end);
 
-	wreq = netfs_alloc_request(iocb->ki_filp->f_mapping, iocb->ki_filp,
-				   start, end - start,
-				   iocb->ki_flags & IOCB_DIRECT ?
-				   NETFS_DIO_WRITE : NETFS_UNBUFFERED_WRITE);
+	wreq = netfs_create_write_req(iocb->ki_filp->f_mapping, iocb->ki_filp, start,
+				      iocb->ki_flags & IOCB_DIRECT ?
+				      NETFS_DIO_WRITE : NETFS_UNBUFFERED_WRITE);
 	if (IS_ERR(wreq))
 		return PTR_ERR(wreq);
 
+	wreq->io_streams[0].avail = true;
+	trace_netfs_write(wreq, (iocb->ki_flags & IOCB_DIRECT ?
+				 netfs_write_trace_dio_write :
+				 netfs_write_trace_unbuffered_write));
+
 	{
 		/* If this is an async op and we're not using a bounce buffer,
 		 * we have to save the source buffer as the iterator is only
@@ -63,7 +68,7 @@ static ssize_t netfs_unbuffered_write_iter_locked(struct kiocb *iocb, struct iov
 		 * request.
 		 */
 		if (async || user_backed_iter(iter)) {
-			n = netfs_extract_user_iter(iter, wreq->len, &wreq->iter, 0);
+			n = netfs_extract_user_iter(iter, len, &wreq->iter, 0);
 			if (n < 0) {
 				ret = n;
 				goto out;
@@ -71,7 +76,6 @@ static ssize_t netfs_unbuffered_write_iter_locked(struct kiocb *iocb, struct iov
 			wreq->direct_bv = (struct bio_vec *)wreq->iter.bvec;
 			wreq->direct_bv_count = n;
 			wreq->direct_bv_unpin = iov_iter_extract_will_pin(iter);
-			wreq->len = iov_iter_count(&wreq->iter);
 		} else {
 			wreq->iter = *iter;
 		}
@@ -79,6 +83,8 @@ static ssize_t netfs_unbuffered_write_iter_locked(struct kiocb *iocb, struct iov
 		wreq->io_iter = wreq->iter;
 	}
 
+	__set_bit(NETFS_RREQ_USE_IO_ITER, &wreq->flags);
+
 	/* Copy the data into the bounce buffer and encrypt it. */
 	// TODO
 
@@ -86,13 +92,11 @@ static ssize_t netfs_unbuffered_write_iter_locked(struct kiocb *iocb, struct iov
 	__set_bit(NETFS_RREQ_UPLOAD_TO_SERVER, &wreq->flags);
 	if (async)
 		wreq->iocb = iocb;
+	wreq->len = iov_iter_count(&wreq->io_iter);
 	wreq->cleanup = netfs_cleanup_dio_write;
-	ret = netfs_begin_write(wreq, is_sync_kiocb(iocb),
-				iocb->ki_flags & IOCB_DIRECT ?
-				netfs_write_trace_dio_write :
-				netfs_write_trace_unbuffered_write);
+	ret = netfs_unbuffered_write(wreq, is_sync_kiocb(iocb), wreq->len);
 	if (ret < 0) {
-		_debug("begin = %zd", ret);
+		kdebug("begin = %zd", ret);
 		goto out;
 	}
 
@@ -100,9 +104,8 @@ static ssize_t netfs_unbuffered_write_iter_locked(struct kiocb *iocb, struct iov
 		trace_netfs_rreq(wreq, netfs_rreq_trace_wait_ip);
 		wait_on_bit(&wreq->flags, NETFS_RREQ_IN_PROGRESS,
 			    TASK_UNINTERRUPTIBLE);
-
+		smp_rmb(); /* Read error/transferred after RIP flag */
 		ret = wreq->error;
-		_debug("waited = %zd", ret);
 		if (ret == 0) {
 			ret = wreq->transferred;
 			iocb->ki_pos += ret;
@@ -115,6 +118,7 @@ out:
 	netfs_put_request(wreq, false, netfs_rreq_trace_put_return);
 	return ret;
 }
+EXPORT_SYMBOL(netfs_unbuffered_write_iter_locked);
 
 /**
  * netfs_unbuffered_write_iter - Unbuffered write to a file
@@ -132,18 +136,20 @@ out:
 ssize_t netfs_unbuffered_write_iter(struct kiocb *iocb, struct iov_iter *from)
 {
 	struct file *file = iocb->ki_filp;
-	struct inode *inode = file->f_mapping->host;
+	struct address_space *mapping = file->f_mapping;
+	struct inode *inode = mapping->host;
 	struct netfs_inode *ictx = netfs_inode(inode);
-	unsigned long long end;
 	ssize_t ret;
+	loff_t pos = iocb->ki_pos;
+	unsigned long long end = pos + iov_iter_count(from) - 1;
 
-	_enter("%llx,%zx,%llx", iocb->ki_pos, iov_iter_count(from), i_size_read(inode));
+	kenter("%llx,%zx,%llx", pos, iov_iter_count(from), i_size_read(inode));
 
 	if (!iov_iter_count(from))
 		return 0;
 
 	trace_netfs_write_iter(iocb, from);
-	netfs_stat(&netfs_n_rh_dio_write);
+	netfs_stat(&netfs_n_wh_dio_write);
 
 	ret = netfs_start_io_direct(inode);
 	if (ret < 0)
@@ -157,7 +163,25 @@ ssize_t netfs_unbuffered_write_iter(struct kiocb *iocb, struct iov_iter *from)
 	ret = file_update_time(file);
 	if (ret < 0)
 		goto out;
-	ret = kiocb_invalidate_pages(iocb, iov_iter_count(from));
+	if (iocb->ki_flags & IOCB_NOWAIT) {
+		/* We could block if there are any pages in the range. */
+		ret = -EAGAIN;
+		if (filemap_range_has_page(mapping, pos, end))
+			if (filemap_invalidate_inode(inode, true, pos, end))
+				goto out;
+	} else {
+		ret = filemap_write_and_wait_range(mapping, pos, end);
+		if (ret < 0)
+			goto out;
+	}
+
+	/*
+	 * After a write we want buffered reads to be sure to go to disk to get
+	 * the new data.  We invalidate clean cached page from the region we're
+	 * about to write.  We do this *before* the write so that we can return
+	 * without clobbering -EIOCBQUEUED from ->direct_IO().
+	 */
+	ret = filemap_invalidate_inode(inode, true, pos, end);
 	if (ret < 0)
 		goto out;
 	end = iocb->ki_pos + iov_iter_count(from);
diff --git a/fs/netfs/fscache_cache.c b/fs/netfs/fscache_cache.c
index 9397ed39b0..288a73c307 100644
--- a/fs/netfs/fscache_cache.c
+++ b/fs/netfs/fscache_cache.c
@@ -237,7 +237,7 @@ int fscache_add_cache(struct fscache_cache *cache,
 {
 	int n_accesses;
 
-	_enter("{%s,%s}", ops->name, cache->name);
+	kenter("{%s,%s}", ops->name, cache->name);
 
 	BUG_ON(fscache_cache_state(cache) != FSCACHE_CACHE_IS_PREPARING);
 
@@ -257,7 +257,7 @@ int fscache_add_cache(struct fscache_cache *cache,
 
 	up_write(&fscache_addremove_sem);
 	pr_notice("Cache \"%s\" added (type %s)\n", cache->name, ops->name);
-	_leave(" = 0 [%s]", cache->name);
+	kleave(" = 0 [%s]", cache->name);
 	return 0;
 }
 EXPORT_SYMBOL(fscache_add_cache);
diff --git a/fs/netfs/fscache_cookie.c b/fs/netfs/fscache_cookie.c
index bce2492186..4d1e8bf4c6 100644
--- a/fs/netfs/fscache_cookie.c
+++ b/fs/netfs/fscache_cookie.c
@@ -456,7 +456,7 @@ struct fscache_cookie *__fscache_acquire_cookie(
 {
 	struct fscache_cookie *cookie;
 
-	_enter("V=%x", volume->debug_id);
+	kenter("V=%x", volume->debug_id);
 
 	if (!index_key || !index_key_len || index_key_len > 255 || aux_data_len > 255)
 		return NULL;
@@ -484,7 +484,7 @@ struct fscache_cookie *__fscache_acquire_cookie(
 
 	trace_fscache_acquire(cookie);
 	fscache_stat(&fscache_n_acquires_ok);
-	_leave(" = c=%08x", cookie->debug_id);
+	kleave(" = c=%08x", cookie->debug_id);
 	return cookie;
 }
 EXPORT_SYMBOL(__fscache_acquire_cookie);
@@ -505,7 +505,7 @@ static void fscache_perform_lookup(struct fscache_cookie *cookie)
 	enum fscache_access_trace trace = fscache_access_lookup_cookie_end_failed;
 	bool need_withdraw = false;
 
-	_enter("");
+	kenter("");
 
 	if (!cookie->volume->cache_priv) {
 		fscache_create_volume(cookie->volume, true);
@@ -519,7 +519,7 @@ static void fscache_perform_lookup(struct fscache_cookie *cookie)
 		if (cookie->state != FSCACHE_COOKIE_STATE_FAILED)
 			fscache_set_cookie_state(cookie, FSCACHE_COOKIE_STATE_QUIESCENT);
 		need_withdraw = true;
-		_leave(" [fail]");
+		kleave(" [fail]");
 		goto out;
 	}
 
@@ -572,7 +572,7 @@ void __fscache_use_cookie(struct fscache_cookie *cookie, bool will_modify)
 	bool queue = false;
 	int n_active;
 
-	_enter("c=%08x", cookie->debug_id);
+	kenter("c=%08x", cookie->debug_id);
 
 	if (WARN(test_bit(FSCACHE_COOKIE_RELINQUISHED, &cookie->flags),
 		 "Trying to use relinquished cookie\n"))
@@ -636,7 +636,7 @@ again:
 	spin_unlock(&cookie->lock);
 	if (queue)
 		fscache_queue_cookie(cookie, fscache_cookie_get_use_work);
-	_leave("");
+	kleave("");
 }
 EXPORT_SYMBOL(__fscache_use_cookie);
 
@@ -702,7 +702,7 @@ static void fscache_cookie_state_machine(struct fscache_cookie *cookie)
 	enum fscache_cookie_state state;
 	bool wake = false;
 
-	_enter("c=%x", cookie->debug_id);
+	kenter("c=%x", cookie->debug_id);
 
 again:
 	spin_lock(&cookie->lock);
@@ -820,7 +820,7 @@ out:
 	spin_unlock(&cookie->lock);
 	if (wake)
 		wake_up_cookie_state(cookie);
-	_leave("");
+	kleave("");
 }
 
 static void fscache_cookie_worker(struct work_struct *work)
@@ -867,7 +867,7 @@ static void fscache_cookie_lru_do_one(struct fscache_cookie *cookie)
 		set_bit(FSCACHE_COOKIE_DO_LRU_DISCARD, &cookie->flags);
 		spin_unlock(&cookie->lock);
 		fscache_stat(&fscache_n_cookies_lru_expired);
-		_debug("lru c=%x", cookie->debug_id);
+		kdebug("lru c=%x", cookie->debug_id);
 		__fscache_withdraw_cookie(cookie);
 	}
 
@@ -971,7 +971,7 @@ void __fscache_relinquish_cookie(struct fscache_cookie *cookie, bool retire)
 	if (retire)
 		fscache_stat(&fscache_n_relinquishes_retire);
 
-	_enter("c=%08x{%d},%d",
+	kenter("c=%08x{%d},%d",
 	       cookie->debug_id, atomic_read(&cookie->n_active), retire);
 
 	if (WARN(test_and_set_bit(FSCACHE_COOKIE_RELINQUISHED, &cookie->flags),
@@ -1050,7 +1050,7 @@ void __fscache_invalidate(struct fscache_cookie *cookie,
 {
 	bool is_caching;
 
-	_enter("c=%x", cookie->debug_id);
+	kenter("c=%x", cookie->debug_id);
 
 	fscache_stat(&fscache_n_invalidates);
 
@@ -1072,7 +1072,7 @@ void __fscache_invalidate(struct fscache_cookie *cookie,
 	case FSCACHE_COOKIE_STATE_INVALIDATING: /* is_still_valid will catch it */
 	default:
 		spin_unlock(&cookie->lock);
-		_leave(" [no %u]", cookie->state);
+		kleave(" [no %u]", cookie->state);
 		return;
 
 	case FSCACHE_COOKIE_STATE_LOOKING_UP:
@@ -1081,7 +1081,7 @@ void __fscache_invalidate(struct fscache_cookie *cookie,
 		fallthrough;
 	case FSCACHE_COOKIE_STATE_CREATING:
 		spin_unlock(&cookie->lock);
-		_leave(" [look %x]", cookie->inval_counter);
+		kleave(" [look %x]", cookie->inval_counter);
 		return;
 
 	case FSCACHE_COOKIE_STATE_ACTIVE:
@@ -1094,7 +1094,7 @@ void __fscache_invalidate(struct fscache_cookie *cookie,
 
 		if (is_caching)
 			fscache_queue_cookie(cookie, fscache_cookie_get_inval_work);
-		_leave(" [inv]");
+		kleave(" [inv]");
 		return;
 	}
 }
diff --git a/fs/netfs/fscache_io.c b/fs/netfs/fscache_io.c
index ad572f7ee8..bf4eaeec44 100644
--- a/fs/netfs/fscache_io.c
+++ b/fs/netfs/fscache_io.c
@@ -28,12 +28,12 @@ bool fscache_wait_for_operation(struct netfs_cache_resources *cres,
 
 again:
 	if (!fscache_cache_is_live(cookie->volume->cache)) {
-		_leave(" [broken]");
+		kleave(" [broken]");
 		return false;
 	}
 
 	state = fscache_cookie_state(cookie);
-	_enter("c=%08x{%u},%x", cookie->debug_id, state, want_state);
+	kenter("c=%08x{%u},%x", cookie->debug_id, state, want_state);
 
 	switch (state) {
 	case FSCACHE_COOKIE_STATE_CREATING:
@@ -52,7 +52,7 @@ again:
 	case FSCACHE_COOKIE_STATE_DROPPED:
 	case FSCACHE_COOKIE_STATE_RELINQUISHING:
 	default:
-		_leave(" [not live]");
+		kleave(" [not live]");
 		return false;
 	}
 
@@ -83,14 +83,16 @@ static int fscache_begin_operation(struct netfs_cache_resources *cres,
 	cres->debug_id		= cookie->debug_id;
 	cres->inval_counter	= cookie->inval_counter;
 
-	if (!fscache_begin_cookie_access(cookie, why))
+	if (!fscache_begin_cookie_access(cookie, why)) {
+		cres->cache_priv = NULL;
 		return -ENOBUFS;
+	}
 
 again:
 	spin_lock(&cookie->lock);
 
 	state = fscache_cookie_state(cookie);
-	_enter("c=%08x{%u},%x", cookie->debug_id, state, want_state);
+	kenter("c=%08x{%u},%x", cookie->debug_id, state, want_state);
 
 	switch (state) {
 	case FSCACHE_COOKIE_STATE_LOOKING_UP:
@@ -138,7 +140,7 @@ failed:
 	cres->cache_priv = NULL;
 	cres->ops = NULL;
 	fscache_end_cookie_access(cookie, fscache_access_io_not_live);
-	_leave(" = -ENOBUFS");
+	kleave(" = -ENOBUFS");
 	return -ENOBUFS;
 }
 
@@ -164,6 +166,7 @@ struct fscache_write_request {
 	loff_t			start;
 	size_t			len;
 	bool			set_bits;
+	bool			using_pgpriv2;
 	netfs_io_terminated_t	term_func;
 	void			*term_func_priv;
 };
@@ -180,7 +183,7 @@ void __fscache_clear_page_bits(struct address_space *mapping,
 
 		rcu_read_lock();
 		xas_for_each(&xas, page, last) {
-			end_page_fscache(page);
+			folio_end_private_2(page_folio(page));
 		}
 		rcu_read_unlock();
 	}
@@ -195,8 +198,9 @@ static void fscache_wreq_done(void *priv, ssize_t transferred_or_error,
 {
 	struct fscache_write_request *wreq = priv;
 
-	fscache_clear_page_bits(wreq->mapping, wreq->start, wreq->len,
-				wreq->set_bits);
+	if (wreq->using_pgpriv2)
+		fscache_clear_page_bits(wreq->mapping, wreq->start, wreq->len,
+					wreq->set_bits);
 
 	if (wreq->term_func)
 		wreq->term_func(wreq->term_func_priv, transferred_or_error,
@@ -210,7 +214,7 @@ void __fscache_write_to_cache(struct fscache_cookie *cookie,
 			      loff_t start, size_t len, loff_t i_size,
 			      netfs_io_terminated_t term_func,
 			      void *term_func_priv,
-			      bool cond)
+			      bool using_pgpriv2, bool cond)
 {
 	struct fscache_write_request *wreq;
 	struct netfs_cache_resources *cres;
@@ -220,7 +224,7 @@ void __fscache_write_to_cache(struct fscache_cookie *cookie,
 	if (len == 0)
 		goto abandon;
 
-	_enter("%llx,%zx", start, len);
+	kenter("%llx,%zx", start, len);
 
 	wreq = kzalloc(sizeof(struct fscache_write_request), GFP_NOFS);
 	if (!wreq)
@@ -228,6 +232,7 @@ void __fscache_write_to_cache(struct fscache_cookie *cookie,
 	wreq->mapping		= mapping;
 	wreq->start		= start;
 	wreq->len		= len;
+	wreq->using_pgpriv2	= using_pgpriv2;
 	wreq->set_bits		= cond;
 	wreq->term_func		= term_func;
 	wreq->term_func_priv	= term_func_priv;
@@ -255,7 +260,8 @@ abandon_end:
 abandon_free:
 	kfree(wreq);
 abandon:
-	fscache_clear_page_bits(mapping, start, len, cond);
+	if (using_pgpriv2)
+		fscache_clear_page_bits(mapping, start, len, cond);
 	if (term_func)
 		term_func(term_func_priv, ret, false);
 }
diff --git a/fs/netfs/fscache_main.c b/fs/netfs/fscache_main.c
index 42e98bb523..bf9b33d26e 100644
--- a/fs/netfs/fscache_main.c
+++ b/fs/netfs/fscache_main.c
@@ -99,7 +99,7 @@ error_wq:
  */
 void __exit fscache_exit(void)
 {
-	_enter("");
+	kenter("");
 
 	kmem_cache_destroy(fscache_cookie_jar);
 	fscache_proc_cleanup();
diff --git a/fs/netfs/fscache_volume.c b/fs/netfs/fscache_volume.c
index cdf991bdd9..2e2a405ca9 100644
--- a/fs/netfs/fscache_volume.c
+++ b/fs/netfs/fscache_volume.c
@@ -27,6 +27,19 @@ struct fscache_volume *fscache_get_volume(struct fscache_volume *volume,
 	return volume;
 }
 
+struct fscache_volume *fscache_try_get_volume(struct fscache_volume *volume,
+					      enum fscache_volume_trace where)
+{
+	int ref;
+
+	if (!__refcount_inc_not_zero(&volume->ref, &ref))
+		return NULL;
+
+	trace_fscache_volume(volume->debug_id, ref + 1, where);
+	return volume;
+}
+EXPORT_SYMBOL(fscache_try_get_volume);
+
 static void fscache_see_volume(struct fscache_volume *volume,
 			       enum fscache_volume_trace where)
 {
@@ -251,7 +264,7 @@ static struct fscache_volume *fscache_alloc_volume(const char *volume_key,
 	fscache_see_volume(volume, fscache_volume_new_acquire);
 	fscache_stat(&fscache_n_volumes);
 	up_write(&fscache_addremove_sem);
-	_leave(" = v=%x", volume->debug_id);
+	kleave(" = v=%x", volume->debug_id);
 	return volume;
 
 err_vol:
@@ -420,6 +433,7 @@ void fscache_put_volume(struct fscache_volume *volume,
 			fscache_free_volume(volume);
 	}
 }
+EXPORT_SYMBOL(fscache_put_volume);
 
 /*
  * Relinquish a volume representation cookie.
@@ -452,7 +466,7 @@ void fscache_withdraw_volume(struct fscache_volume *volume)
 {
 	int n_accesses;
 
-	_debug("withdraw V=%x", volume->debug_id);
+	kdebug("withdraw V=%x", volume->debug_id);
 
 	/* Allow wakeups on dec-to-0 */
 	n_accesses = atomic_dec_return(&volume->n_accesses);
diff --git a/fs/netfs/internal.h b/fs/netfs/internal.h
index ec7045d244..21e46bc9aa 100644
--- a/fs/netfs/internal.h
+++ b/fs/netfs/internal.h
@@ -34,9 +34,10 @@ int netfs_begin_read(struct netfs_io_request *rreq, bool sync);
 /*
  * main.c
  */
-extern unsigned int netfs_debug;
 extern struct list_head netfs_io_requests;
 extern spinlock_t netfs_proc_lock;
+extern mempool_t netfs_request_pool;
+extern mempool_t netfs_subrequest_pool;
 
 #ifdef CONFIG_PROC_FS
 static inline void netfs_proc_add_rreq(struct netfs_io_request *rreq)
@@ -61,15 +62,6 @@ static inline void netfs_proc_del_rreq(struct netfs_io_request *rreq) {}
 /*
  * misc.c
  */
-#define NETFS_FLAG_PUT_MARK		BIT(0)
-#define NETFS_FLAG_PAGECACHE_MARK	BIT(1)
-int netfs_xa_store_and_mark(struct xarray *xa, unsigned long index,
-			    struct folio *folio, unsigned int flags,
-			    gfp_t gfp_mask);
-int netfs_add_folios_to_buffer(struct xarray *buffer,
-			       struct address_space *mapping,
-			       pgoff_t index, pgoff_t to, gfp_t gfp_mask);
-void netfs_clear_buffer(struct xarray *buffer);
 
 /*
  * objects.c
@@ -91,22 +83,12 @@ static inline void netfs_see_request(struct netfs_io_request *rreq,
 }
 
 /*
- * output.c
- */
-int netfs_begin_write(struct netfs_io_request *wreq, bool may_wait,
-		      enum netfs_write_trace what);
-struct netfs_io_request *netfs_begin_writethrough(struct kiocb *iocb, size_t len);
-int netfs_advance_writethrough(struct netfs_io_request *wreq, size_t copied, bool to_page_end);
-int netfs_end_writethrough(struct netfs_io_request *wreq, struct kiocb *iocb);
-
-/*
  * stats.c
  */
 #ifdef CONFIG_NETFS_STATS
 extern atomic_t netfs_n_rh_dio_read;
-extern atomic_t netfs_n_rh_dio_write;
 extern atomic_t netfs_n_rh_readahead;
-extern atomic_t netfs_n_rh_readpage;
+extern atomic_t netfs_n_rh_read_folio;
 extern atomic_t netfs_n_rh_rreq;
 extern atomic_t netfs_n_rh_sreq;
 extern atomic_t netfs_n_rh_download;
@@ -123,6 +105,10 @@ extern atomic_t netfs_n_rh_write_begin;
 extern atomic_t netfs_n_rh_write_done;
 extern atomic_t netfs_n_rh_write_failed;
 extern atomic_t netfs_n_rh_write_zskip;
+extern atomic_t netfs_n_wh_buffered_write;
+extern atomic_t netfs_n_wh_writethrough;
+extern atomic_t netfs_n_wh_dio_write;
+extern atomic_t netfs_n_wh_writepages;
 extern atomic_t netfs_n_wh_wstream_conflict;
 extern atomic_t netfs_n_wh_upload;
 extern atomic_t netfs_n_wh_upload_done;
@@ -149,6 +135,33 @@ static inline void netfs_stat_d(atomic_t *stat)
 #endif
 
 /*
+ * write_collect.c
+ */
+int netfs_folio_written_back(struct folio *folio);
+void netfs_write_collection_worker(struct work_struct *work);
+void netfs_wake_write_collector(struct netfs_io_request *wreq, bool was_async);
+
+/*
+ * write_issue.c
+ */
+struct netfs_io_request *netfs_create_write_req(struct address_space *mapping,
+						struct file *file,
+						loff_t start,
+						enum netfs_io_origin origin);
+void netfs_reissue_write(struct netfs_io_stream *stream,
+			 struct netfs_io_subrequest *subreq);
+int netfs_advance_write(struct netfs_io_request *wreq,
+			struct netfs_io_stream *stream,
+			loff_t start, size_t len, bool to_eof);
+struct netfs_io_request *netfs_begin_writethrough(struct kiocb *iocb, size_t len);
+int netfs_advance_writethrough(struct netfs_io_request *wreq, struct writeback_control *wbc,
+			       struct folio *folio, size_t copied, bool to_page_end,
+			       struct folio **writethrough_cache);
+int netfs_end_writethrough(struct netfs_io_request *wreq, struct writeback_control *wbc,
+			   struct folio *writethrough_cache);
+int netfs_unbuffered_write(struct netfs_io_request *wreq, bool may_wait, size_t len);
+
+/*
  * Miscellaneous functions.
  */
 static inline bool netfs_is_cache_enabled(struct netfs_inode *ctx)
@@ -168,7 +181,7 @@ static inline bool netfs_is_cache_enabled(struct netfs_inode *ctx)
  */
 static inline struct netfs_group *netfs_get_group(struct netfs_group *netfs_group)
 {
-	if (netfs_group)
+	if (netfs_group && netfs_group != NETFS_FOLIO_COPY_TO_CACHE)
 		refcount_inc(&netfs_group->ref);
 	return netfs_group;
 }
@@ -178,7 +191,9 @@ static inline struct netfs_group *netfs_get_group(struct netfs_group *netfs_grou
  */
 static inline void netfs_put_group(struct netfs_group *netfs_group)
 {
-	if (netfs_group && refcount_dec_and_test(&netfs_group->ref))
+	if (netfs_group &&
+	    netfs_group != NETFS_FOLIO_COPY_TO_CACHE &&
+	    refcount_dec_and_test(&netfs_group->ref))
 		netfs_group->free(netfs_group);
 }
 
@@ -187,7 +202,9 @@ static inline void netfs_put_group(struct netfs_group *netfs_group)
  */
 static inline void netfs_put_group_many(struct netfs_group *netfs_group, int nr)
 {
-	if (netfs_group && refcount_sub_and_test(nr, &netfs_group->ref))
+	if (netfs_group &&
+	    netfs_group != NETFS_FOLIO_COPY_TO_CACHE &&
+	    refcount_sub_and_test(nr, &netfs_group->ref))
 		netfs_group->free(netfs_group);
 }
 
@@ -326,8 +343,6 @@ extern const struct seq_operations fscache_volumes_seq_ops;
 
 struct fscache_volume *fscache_get_volume(struct fscache_volume *volume,
 					  enum fscache_volume_trace where);
-void fscache_put_volume(struct fscache_volume *volume,
-			enum fscache_volume_trace where);
 bool fscache_begin_volume_access(struct fscache_volume *volume,
 				 struct fscache_cookie *cookie,
 				 enum fscache_access_trace why);
@@ -338,42 +353,12 @@ void fscache_create_volume(struct fscache_volume *volume, bool wait);
  * debug tracing
  */
 #define dbgprintk(FMT, ...) \
-	printk("[%-6.6s] "FMT"\n", current->comm, ##__VA_ARGS__)
+	pr_debug("[%-6.6s] "FMT"\n", current->comm, ##__VA_ARGS__)
 
 #define kenter(FMT, ...) dbgprintk("==> %s("FMT")", __func__, ##__VA_ARGS__)
 #define kleave(FMT, ...) dbgprintk("<== %s()"FMT"", __func__, ##__VA_ARGS__)
 #define kdebug(FMT, ...) dbgprintk(FMT, ##__VA_ARGS__)
 
-#ifdef __KDEBUG
-#define _enter(FMT, ...) kenter(FMT, ##__VA_ARGS__)
-#define _leave(FMT, ...) kleave(FMT, ##__VA_ARGS__)
-#define _debug(FMT, ...) kdebug(FMT, ##__VA_ARGS__)
-
-#elif defined(CONFIG_NETFS_DEBUG)
-#define _enter(FMT, ...)			\
-do {						\
-	if (netfs_debug)			\
-		kenter(FMT, ##__VA_ARGS__);	\
-} while (0)
-
-#define _leave(FMT, ...)			\
-do {						\
-	if (netfs_debug)			\
-		kleave(FMT, ##__VA_ARGS__);	\
-} while (0)
-
-#define _debug(FMT, ...)			\
-do {						\
-	if (netfs_debug)			\
-		kdebug(FMT, ##__VA_ARGS__);	\
-} while (0)
-
-#else
-#define _enter(FMT, ...) no_printk("==> %s("FMT")", __func__, ##__VA_ARGS__)
-#define _leave(FMT, ...) no_printk("<== %s()"FMT"", __func__, ##__VA_ARGS__)
-#define _debug(FMT, ...) no_printk(FMT, ##__VA_ARGS__)
-#endif
-
 /*
  * assertions
  */
diff --git a/fs/netfs/io.c b/fs/netfs/io.c
index 4261ad6c55..c7576481c3 100644
--- a/fs/netfs/io.c
+++ b/fs/netfs/io.c
@@ -99,145 +99,6 @@ static void netfs_rreq_completed(struct netfs_io_request *rreq, bool was_async)
 }
 
 /*
- * Deal with the completion of writing the data to the cache.  We have to clear
- * the PG_fscache bits on the folios involved and release the caller's ref.
- *
- * May be called in softirq mode and we inherit a ref from the caller.
- */
-static void netfs_rreq_unmark_after_write(struct netfs_io_request *rreq,
-					  bool was_async)
-{
-	struct netfs_io_subrequest *subreq;
-	struct folio *folio;
-	pgoff_t unlocked = 0;
-	bool have_unlocked = false;
-
-	rcu_read_lock();
-
-	list_for_each_entry(subreq, &rreq->subrequests, rreq_link) {
-		XA_STATE(xas, &rreq->mapping->i_pages, subreq->start / PAGE_SIZE);
-
-		xas_for_each(&xas, folio, (subreq->start + subreq->len - 1) / PAGE_SIZE) {
-			if (xas_retry(&xas, folio))
-				continue;
-
-			/* We might have multiple writes from the same huge
-			 * folio, but we mustn't unlock a folio more than once.
-			 */
-			if (have_unlocked && folio->index <= unlocked)
-				continue;
-			unlocked = folio_next_index(folio) - 1;
-			trace_netfs_folio(folio, netfs_folio_trace_end_copy);
-			folio_end_fscache(folio);
-			have_unlocked = true;
-		}
-	}
-
-	rcu_read_unlock();
-	netfs_rreq_completed(rreq, was_async);
-}
-
-static void netfs_rreq_copy_terminated(void *priv, ssize_t transferred_or_error,
-				       bool was_async)
-{
-	struct netfs_io_subrequest *subreq = priv;
-	struct netfs_io_request *rreq = subreq->rreq;
-
-	if (IS_ERR_VALUE(transferred_or_error)) {
-		netfs_stat(&netfs_n_rh_write_failed);
-		trace_netfs_failure(rreq, subreq, transferred_or_error,
-				    netfs_fail_copy_to_cache);
-	} else {
-		netfs_stat(&netfs_n_rh_write_done);
-	}
-
-	trace_netfs_sreq(subreq, netfs_sreq_trace_write_term);
-
-	/* If we decrement nr_copy_ops to 0, the ref belongs to us. */
-	if (atomic_dec_and_test(&rreq->nr_copy_ops))
-		netfs_rreq_unmark_after_write(rreq, was_async);
-
-	netfs_put_subrequest(subreq, was_async, netfs_sreq_trace_put_terminated);
-}
-
-/*
- * Perform any outstanding writes to the cache.  We inherit a ref from the
- * caller.
- */
-static void netfs_rreq_do_write_to_cache(struct netfs_io_request *rreq)
-{
-	struct netfs_cache_resources *cres = &rreq->cache_resources;
-	struct netfs_io_subrequest *subreq, *next, *p;
-	struct iov_iter iter;
-	int ret;
-
-	trace_netfs_rreq(rreq, netfs_rreq_trace_copy);
-
-	/* We don't want terminating writes trying to wake us up whilst we're
-	 * still going through the list.
-	 */
-	atomic_inc(&rreq->nr_copy_ops);
-
-	list_for_each_entry_safe(subreq, p, &rreq->subrequests, rreq_link) {
-		if (!test_bit(NETFS_SREQ_COPY_TO_CACHE, &subreq->flags)) {
-			list_del_init(&subreq->rreq_link);
-			netfs_put_subrequest(subreq, false,
-					     netfs_sreq_trace_put_no_copy);
-		}
-	}
-
-	list_for_each_entry(subreq, &rreq->subrequests, rreq_link) {
-		/* Amalgamate adjacent writes */
-		while (!list_is_last(&subreq->rreq_link, &rreq->subrequests)) {
-			next = list_next_entry(subreq, rreq_link);
-			if (next->start != subreq->start + subreq->len)
-				break;
-			subreq->len += next->len;
-			list_del_init(&next->rreq_link);
-			netfs_put_subrequest(next, false,
-					     netfs_sreq_trace_put_merged);
-		}
-
-		ret = cres->ops->prepare_write(cres, &subreq->start, &subreq->len,
-					       subreq->len, rreq->i_size, true);
-		if (ret < 0) {
-			trace_netfs_failure(rreq, subreq, ret, netfs_fail_prepare_write);
-			trace_netfs_sreq(subreq, netfs_sreq_trace_write_skip);
-			continue;
-		}
-
-		iov_iter_xarray(&iter, ITER_SOURCE, &rreq->mapping->i_pages,
-				subreq->start, subreq->len);
-
-		atomic_inc(&rreq->nr_copy_ops);
-		netfs_stat(&netfs_n_rh_write);
-		netfs_get_subrequest(subreq, netfs_sreq_trace_get_copy_to_cache);
-		trace_netfs_sreq(subreq, netfs_sreq_trace_write);
-		cres->ops->write(cres, subreq->start, &iter,
-				 netfs_rreq_copy_terminated, subreq);
-	}
-
-	/* If we decrement nr_copy_ops to 0, the usage ref belongs to us. */
-	if (atomic_dec_and_test(&rreq->nr_copy_ops))
-		netfs_rreq_unmark_after_write(rreq, false);
-}
-
-static void netfs_rreq_write_to_cache_work(struct work_struct *work)
-{
-	struct netfs_io_request *rreq =
-		container_of(work, struct netfs_io_request, work);
-
-	netfs_rreq_do_write_to_cache(rreq);
-}
-
-static void netfs_rreq_write_to_cache(struct netfs_io_request *rreq)
-{
-	rreq->work.func = netfs_rreq_write_to_cache_work;
-	if (!queue_work(system_unbound_wq, &rreq->work))
-		BUG();
-}
-
-/*
  * Handle a short read.
  */
 static void netfs_rreq_short_read(struct netfs_io_request *rreq,
@@ -269,7 +130,7 @@ static void netfs_reset_subreq_iter(struct netfs_io_request *rreq,
 	if (count == remaining)
 		return;
 
-	_debug("R=%08x[%u] ITER RESUB-MISMATCH %zx != %zx-%zx-%llx %x\n",
+	kdebug("R=%08x[%u] ITER RESUB-MISMATCH %zx != %zx-%zx-%llx %x\n",
 	       rreq->debug_id, subreq->debug_index,
 	       iov_iter_count(&subreq->io_iter), subreq->transferred,
 	       subreq->len, rreq->i_size,
@@ -352,8 +213,13 @@ static void netfs_rreq_assess_dio(struct netfs_io_request *rreq)
 	unsigned int i;
 	size_t transferred = 0;
 
-	for (i = 0; i < rreq->direct_bv_count; i++)
+	for (i = 0; i < rreq->direct_bv_count; i++) {
 		flush_dcache_page(rreq->direct_bv[i].bv_page);
+		// TODO: cifs marks pages in the destination buffer
+		// dirty under some circumstances after a read.  Do we
+		// need to do that too?
+		set_page_dirty(rreq->direct_bv[i].bv_page);
+	}
 
 	list_for_each_entry(subreq, &rreq->subrequests, rreq_link) {
 		if (subreq->error || subreq->transferred == 0)
@@ -409,9 +275,6 @@ again:
 	clear_bit_unlock(NETFS_RREQ_IN_PROGRESS, &rreq->flags);
 	wake_up_bit(&rreq->flags, NETFS_RREQ_IN_PROGRESS);
 
-	if (test_bit(NETFS_RREQ_COPY_TO_CACHE, &rreq->flags))
-		return netfs_rreq_write_to_cache(rreq);
-
 	netfs_rreq_completed(rreq, was_async);
 }
 
@@ -463,7 +326,7 @@ void netfs_subreq_terminated(struct netfs_io_subrequest *subreq,
 	struct netfs_io_request *rreq = subreq->rreq;
 	int u;
 
-	_enter("R=%x[%x]{%llx,%lx},%zd",
+	kenter("R=%x[%x]{%llx,%lx},%zd",
 	       rreq->debug_id, subreq->debug_index,
 	       subreq->start, subreq->flags, transferred_or_error);
 
@@ -572,7 +435,7 @@ netfs_rreq_prepare_read(struct netfs_io_request *rreq,
 	struct netfs_inode *ictx = netfs_inode(rreq->inode);
 	size_t lsize;
 
-	_enter("%llx-%llx,%llx", subreq->start, subreq->start + subreq->len, rreq->i_size);
+	kenter("%llx-%llx,%llx", subreq->start, subreq->start + subreq->len, rreq->i_size);
 
 	if (rreq->origin != NETFS_DIO_READ) {
 		source = netfs_cache_prepare_read(subreq, rreq->i_size);
@@ -618,7 +481,7 @@ netfs_rreq_prepare_read(struct netfs_io_request *rreq,
 
 set:
 	if (subreq->len > rreq->len)
-		pr_warn("R=%08x[%u] SREQ>RREQ %zx > %zx\n",
+		pr_warn("R=%08x[%u] SREQ>RREQ %zx > %llx\n",
 			rreq->debug_id, subreq->debug_index,
 			subreq->len, rreq->len);
 
@@ -643,8 +506,7 @@ out:
  * Slice off a piece of a read request and submit an I/O request for it.
  */
 static bool netfs_rreq_submit_slice(struct netfs_io_request *rreq,
-				    struct iov_iter *io_iter,
-				    unsigned int *_debug_index)
+				    struct iov_iter *io_iter)
 {
 	struct netfs_io_subrequest *subreq;
 	enum netfs_io_source source;
@@ -653,11 +515,10 @@ static bool netfs_rreq_submit_slice(struct netfs_io_request *rreq,
 	if (!subreq)
 		return false;
 
-	subreq->debug_index	= (*_debug_index)++;
 	subreq->start		= rreq->start + rreq->submitted;
 	subreq->len		= io_iter->count;
 
-	_debug("slice %llx,%zx,%zx", subreq->start, subreq->len, rreq->submitted);
+	kdebug("slice %llx,%zx,%llx", subreq->start, subreq->len, rreq->submitted);
 	list_add_tail(&subreq->rreq_link, &rreq->subrequests);
 
 	/* Call out to the cache to find out what it can do with the remaining
@@ -707,10 +568,9 @@ subreq_failed:
 int netfs_begin_read(struct netfs_io_request *rreq, bool sync)
 {
 	struct iov_iter io_iter;
-	unsigned int debug_index = 0;
 	int ret;
 
-	_enter("R=%x %llx-%llx",
+	kenter("R=%x %llx-%llx",
 	       rreq->debug_id, rreq->start, rreq->start + rreq->len - 1);
 
 	if (rreq->len == 0) {
@@ -733,12 +593,12 @@ int netfs_begin_read(struct netfs_io_request *rreq, bool sync)
 	atomic_set(&rreq->nr_outstanding, 1);
 	io_iter = rreq->io_iter;
 	do {
-		_debug("submit %llx + %zx >= %llx",
+		kdebug("submit %llx + %llx >= %llx",
 		       rreq->start, rreq->submitted, rreq->i_size);
 		if (rreq->origin == NETFS_DIO_READ &&
 		    rreq->start + rreq->submitted >= rreq->i_size)
 			break;
-		if (!netfs_rreq_submit_slice(rreq, &io_iter, &debug_index))
+		if (!netfs_rreq_submit_slice(rreq, &io_iter))
 			break;
 		if (test_bit(NETFS_RREQ_BLOCKED, &rreq->flags) &&
 		    test_bit(NETFS_RREQ_NONBLOCK, &rreq->flags))
diff --git a/fs/netfs/main.c b/fs/netfs/main.c
index 5e77618a79..db824c3728 100644
--- a/fs/netfs/main.c
+++ b/fs/netfs/main.c
@@ -7,6 +7,7 @@
 
 #include <linux/module.h>
 #include <linux/export.h>
+#include <linux/mempool.h>
 #include <linux/proc_fs.h>
 #include <linux/seq_file.h>
 #include "internal.h"
@@ -19,9 +20,10 @@ MODULE_LICENSE("GPL");
 
 EXPORT_TRACEPOINT_SYMBOL(netfs_sreq);
 
-unsigned netfs_debug;
-module_param_named(debug, netfs_debug, uint, S_IWUSR | S_IRUGO);
-MODULE_PARM_DESC(netfs_debug, "Netfs support debugging mask");
+static struct kmem_cache *netfs_request_slab;
+static struct kmem_cache *netfs_subrequest_slab;
+mempool_t netfs_request_pool;
+mempool_t netfs_subrequest_pool;
 
 #ifdef CONFIG_PROC_FS
 LIST_HEAD(netfs_io_requests);
@@ -31,9 +33,9 @@ static const char *netfs_origins[nr__netfs_io_origin] = {
 	[NETFS_READAHEAD]		= "RA",
 	[NETFS_READPAGE]		= "RP",
 	[NETFS_READ_FOR_WRITE]		= "RW",
+	[NETFS_COPY_TO_CACHE]		= "CC",
 	[NETFS_WRITEBACK]		= "WB",
 	[NETFS_WRITETHROUGH]		= "WT",
-	[NETFS_LAUNDER_WRITE]		= "LW",
 	[NETFS_UNBUFFERED_WRITE]	= "UW",
 	[NETFS_DIO_READ]		= "DR",
 	[NETFS_DIO_WRITE]		= "DW",
@@ -56,7 +58,7 @@ static int netfs_requests_seq_show(struct seq_file *m, void *v)
 
 	rreq = list_entry(v, struct netfs_io_request, proc_link);
 	seq_printf(m,
-		   "%08x %s %3d %2lx %4d %3d @%04llx %zx/%zx",
+		   "%08x %s %3d %2lx %4d %3d @%04llx %llx/%llx",
 		   rreq->debug_id,
 		   netfs_origins[rreq->origin],
 		   refcount_read(&rreq->ref),
@@ -98,25 +100,54 @@ static int __init netfs_init(void)
 {
 	int ret = -ENOMEM;
 
+	netfs_request_slab = kmem_cache_create("netfs_request",
+					       sizeof(struct netfs_io_request), 0,
+					       SLAB_HWCACHE_ALIGN | SLAB_ACCOUNT,
+					       NULL);
+	if (!netfs_request_slab)
+		goto error_req;
+
+	if (mempool_init_slab_pool(&netfs_request_pool, 100, netfs_request_slab) < 0)
+		goto error_reqpool;
+
+	netfs_subrequest_slab = kmem_cache_create("netfs_subrequest",
+						  sizeof(struct netfs_io_subrequest), 0,
+						  SLAB_HWCACHE_ALIGN | SLAB_ACCOUNT,
+						  NULL);
+	if (!netfs_subrequest_slab)
+		goto error_subreq;
+
+	if (mempool_init_slab_pool(&netfs_subrequest_pool, 100, netfs_subrequest_slab) < 0)
+		goto error_subreqpool;
+
 	if (!proc_mkdir("fs/netfs", NULL))
-		goto error;
+		goto error_proc;
 	if (!proc_create_seq("fs/netfs/requests", S_IFREG | 0444, NULL,
 			     &netfs_requests_seq_ops))
-		goto error_proc;
+		goto error_procfile;
 #ifdef CONFIG_FSCACHE_STATS
 	if (!proc_create_single("fs/netfs/stats", S_IFREG | 0444, NULL,
 				netfs_stats_show))
-		goto error_proc;
+		goto error_procfile;
 #endif
 
 	ret = fscache_init();
 	if (ret < 0)
-		goto error_proc;
+		goto error_fscache;
 	return 0;
 
-error_proc:
+error_fscache:
+error_procfile:
 	remove_proc_entry("fs/netfs", NULL);
-error:
+error_proc:
+	mempool_exit(&netfs_subrequest_pool);
+error_subreqpool:
+	kmem_cache_destroy(netfs_subrequest_slab);
+error_subreq:
+	mempool_exit(&netfs_request_pool);
+error_reqpool:
+	kmem_cache_destroy(netfs_request_slab);
+error_req:
 	return ret;
 }
 fs_initcall(netfs_init);
@@ -125,5 +156,9 @@ static void __exit netfs_exit(void)
 {
 	fscache_exit();
 	remove_proc_entry("fs/netfs", NULL);
+	mempool_exit(&netfs_subrequest_pool);
+	kmem_cache_destroy(netfs_subrequest_slab);
+	mempool_exit(&netfs_request_pool);
+	kmem_cache_destroy(netfs_request_slab);
 }
 module_exit(netfs_exit);
diff --git a/fs/netfs/misc.c b/fs/netfs/misc.c
index 90051ced8e..172808e83c 100644
--- a/fs/netfs/misc.c
+++ b/fs/netfs/misc.c
@@ -8,87 +8,6 @@
 #include <linux/swap.h>
 #include "internal.h"
 
-/*
- * Attach a folio to the buffer and maybe set marks on it to say that we need
- * to put the folio later and twiddle the pagecache flags.
- */
-int netfs_xa_store_and_mark(struct xarray *xa, unsigned long index,
-			    struct folio *folio, unsigned int flags,
-			    gfp_t gfp_mask)
-{
-	XA_STATE_ORDER(xas, xa, index, folio_order(folio));
-
-retry:
-	xas_lock(&xas);
-	for (;;) {
-		xas_store(&xas, folio);
-		if (!xas_error(&xas))
-			break;
-		xas_unlock(&xas);
-		if (!xas_nomem(&xas, gfp_mask))
-			return xas_error(&xas);
-		goto retry;
-	}
-
-	if (flags & NETFS_FLAG_PUT_MARK)
-		xas_set_mark(&xas, NETFS_BUF_PUT_MARK);
-	if (flags & NETFS_FLAG_PAGECACHE_MARK)
-		xas_set_mark(&xas, NETFS_BUF_PAGECACHE_MARK);
-	xas_unlock(&xas);
-	return xas_error(&xas);
-}
-
-/*
- * Create the specified range of folios in the buffer attached to the read
- * request.  The folios are marked with NETFS_BUF_PUT_MARK so that we know that
- * these need freeing later.
- */
-int netfs_add_folios_to_buffer(struct xarray *buffer,
-			       struct address_space *mapping,
-			       pgoff_t index, pgoff_t to, gfp_t gfp_mask)
-{
-	struct folio *folio;
-	int ret;
-
-	if (to + 1 == index) /* Page range is inclusive */
-		return 0;
-
-	do {
-		/* TODO: Figure out what order folio can be allocated here */
-		folio = filemap_alloc_folio(readahead_gfp_mask(mapping), 0);
-		if (!folio)
-			return -ENOMEM;
-		folio->index = index;
-		ret = netfs_xa_store_and_mark(buffer, index, folio,
-					      NETFS_FLAG_PUT_MARK, gfp_mask);
-		if (ret < 0) {
-			folio_put(folio);
-			return ret;
-		}
-
-		index += folio_nr_pages(folio);
-	} while (index <= to && index != 0);
-
-	return 0;
-}
-
-/*
- * Clear an xarray buffer, putting a ref on the folios that have
- * NETFS_BUF_PUT_MARK set.
- */
-void netfs_clear_buffer(struct xarray *buffer)
-{
-	struct folio *folio;
-	XA_STATE(xas, buffer, 0);
-
-	rcu_read_lock();
-	xas_for_each_marked(&xas, folio, ULONG_MAX, NETFS_BUF_PUT_MARK) {
-		folio_put(folio);
-	}
-	rcu_read_unlock();
-	xa_destroy(buffer);
-}
-
 /**
  * netfs_dirty_folio - Mark folio dirty and pin a cache object for writeback
  * @mapping: The mapping the folio belongs to.
@@ -107,7 +26,7 @@ bool netfs_dirty_folio(struct address_space *mapping, struct folio *folio)
 	struct fscache_cookie *cookie = netfs_i_cookie(ictx);
 	bool need_use = false;
 
-	_enter("");
+	kenter("");
 
 	if (!filemap_dirty_folio(mapping, folio))
 		return false;
@@ -177,12 +96,10 @@ EXPORT_SYMBOL(netfs_clear_inode_writeback);
  */
 void netfs_invalidate_folio(struct folio *folio, size_t offset, size_t length)
 {
-	struct netfs_folio *finfo = NULL;
+	struct netfs_folio *finfo;
 	size_t flen = folio_size(folio);
 
-	_enter("{%lx},%zx,%zx", folio->index, offset, length);
-
-	folio_wait_fscache(folio);
+	kenter("{%lx},%zx,%zx", folio->index, offset, length);
 
 	if (!folio_test_private(folio))
 		return;
@@ -248,12 +165,6 @@ bool netfs_release_folio(struct folio *folio, gfp_t gfp)
 
 	if (folio_test_private(folio))
 		return false;
-	if (folio_test_fscache(folio)) {
-		if (current_is_kswapd() || !(gfp & __GFP_FS))
-			return false;
-		folio_wait_fscache(folio);
-	}
-
 	fscache_note_page_release(netfs_i_cookie(ctx));
 	return true;
 }
diff --git a/fs/netfs/objects.c b/fs/netfs/objects.c
index 610ceb5bd8..f4a6427274 100644
--- a/fs/netfs/objects.c
+++ b/fs/netfs/objects.c
@@ -6,6 +6,8 @@
  */
 
 #include <linux/slab.h>
+#include <linux/mempool.h>
+#include <linux/delay.h>
 #include "internal.h"
 
 /*
@@ -20,17 +22,22 @@ struct netfs_io_request *netfs_alloc_request(struct address_space *mapping,
 	struct inode *inode = file ? file_inode(file) : mapping->host;
 	struct netfs_inode *ctx = netfs_inode(inode);
 	struct netfs_io_request *rreq;
+	mempool_t *mempool = ctx->ops->request_pool ?: &netfs_request_pool;
+	struct kmem_cache *cache = mempool->pool_data;
 	bool is_unbuffered = (origin == NETFS_UNBUFFERED_WRITE ||
 			      origin == NETFS_DIO_READ ||
 			      origin == NETFS_DIO_WRITE);
 	bool cached = !is_unbuffered && netfs_is_cache_enabled(ctx);
 	int ret;
 
-	rreq = kzalloc(ctx->ops->io_request_size ?: sizeof(struct netfs_io_request),
-		       GFP_KERNEL);
-	if (!rreq)
-		return ERR_PTR(-ENOMEM);
+	for (;;) {
+		rreq = mempool_alloc(mempool, GFP_KERNEL);
+		if (rreq)
+			break;
+		msleep(10);
+	}
 
+	memset(rreq, 0, kmem_cache_size(cache));
 	rreq->start	= start;
 	rreq->len	= len;
 	rreq->upper_len	= len;
@@ -40,23 +47,32 @@ struct netfs_io_request *netfs_alloc_request(struct address_space *mapping,
 	rreq->inode	= inode;
 	rreq->i_size	= i_size_read(inode);
 	rreq->debug_id	= atomic_inc_return(&debug_ids);
+	rreq->wsize	= INT_MAX;
+	spin_lock_init(&rreq->lock);
+	INIT_LIST_HEAD(&rreq->io_streams[0].subrequests);
+	INIT_LIST_HEAD(&rreq->io_streams[1].subrequests);
 	INIT_LIST_HEAD(&rreq->subrequests);
 	INIT_WORK(&rreq->work, NULL);
 	refcount_set(&rreq->ref, 1);
 
 	__set_bit(NETFS_RREQ_IN_PROGRESS, &rreq->flags);
-	if (cached)
+	if (cached) {
 		__set_bit(NETFS_RREQ_WRITE_TO_CACHE, &rreq->flags);
+		if (test_bit(NETFS_ICTX_USE_PGPRIV2, &ctx->flags))
+			/* Filesystem uses deprecated PG_private_2 marking. */
+			__set_bit(NETFS_RREQ_USE_PGPRIV2, &rreq->flags);
+	}
 	if (file && file->f_flags & O_NONBLOCK)
 		__set_bit(NETFS_RREQ_NONBLOCK, &rreq->flags);
 	if (rreq->netfs_ops->init_request) {
 		ret = rreq->netfs_ops->init_request(rreq, file);
 		if (ret < 0) {
-			kfree(rreq);
+			mempool_free(rreq, rreq->netfs_ops->request_pool ?: &netfs_request_pool);
 			return ERR_PTR(ret);
 		}
 	}
 
+	atomic_inc(&ctx->io_count);
 	trace_netfs_rreq_ref(rreq->debug_id, 1, netfs_rreq_trace_new);
 	netfs_proc_add_rreq(rreq);
 	netfs_stat(&netfs_n_rh_rreq);
@@ -74,6 +90,8 @@ void netfs_get_request(struct netfs_io_request *rreq, enum netfs_rreq_ref_trace
 void netfs_clear_subrequests(struct netfs_io_request *rreq, bool was_async)
 {
 	struct netfs_io_subrequest *subreq;
+	struct netfs_io_stream *stream;
+	int s;
 
 	while (!list_empty(&rreq->subrequests)) {
 		subreq = list_first_entry(&rreq->subrequests,
@@ -82,12 +100,32 @@ void netfs_clear_subrequests(struct netfs_io_request *rreq, bool was_async)
 		netfs_put_subrequest(subreq, was_async,
 				     netfs_sreq_trace_put_clear);
 	}
+
+	for (s = 0; s < ARRAY_SIZE(rreq->io_streams); s++) {
+		stream = &rreq->io_streams[s];
+		while (!list_empty(&stream->subrequests)) {
+			subreq = list_first_entry(&stream->subrequests,
+						  struct netfs_io_subrequest, rreq_link);
+			list_del(&subreq->rreq_link);
+			netfs_put_subrequest(subreq, was_async,
+					     netfs_sreq_trace_put_clear);
+		}
+	}
+}
+
+static void netfs_free_request_rcu(struct rcu_head *rcu)
+{
+	struct netfs_io_request *rreq = container_of(rcu, struct netfs_io_request, rcu);
+
+	mempool_free(rreq, rreq->netfs_ops->request_pool ?: &netfs_request_pool);
+	netfs_stat_d(&netfs_n_rh_rreq);
 }
 
 static void netfs_free_request(struct work_struct *work)
 {
 	struct netfs_io_request *rreq =
 		container_of(work, struct netfs_io_request, work);
+	struct netfs_inode *ictx = netfs_inode(rreq->inode);
 	unsigned int i;
 
 	trace_netfs_rreq(rreq, netfs_rreq_trace_free);
@@ -106,8 +144,10 @@ static void netfs_free_request(struct work_struct *work)
 		}
 		kvfree(rreq->direct_bv);
 	}
-	kfree_rcu(rreq, rcu);
-	netfs_stat_d(&netfs_n_rh_rreq);
+
+	if (atomic_dec_and_test(&ictx->io_count))
+		wake_up_var(&ictx->io_count);
+	call_rcu(&rreq->rcu, netfs_free_request_rcu);
 }
 
 void netfs_put_request(struct netfs_io_request *rreq, bool was_async,
@@ -139,19 +179,25 @@ void netfs_put_request(struct netfs_io_request *rreq, bool was_async,
 struct netfs_io_subrequest *netfs_alloc_subrequest(struct netfs_io_request *rreq)
 {
 	struct netfs_io_subrequest *subreq;
-
-	subreq = kzalloc(rreq->netfs_ops->io_subrequest_size ?:
-			 sizeof(struct netfs_io_subrequest),
-			 GFP_KERNEL);
-	if (subreq) {
-		INIT_WORK(&subreq->work, NULL);
-		INIT_LIST_HEAD(&subreq->rreq_link);
-		refcount_set(&subreq->ref, 2);
-		subreq->rreq = rreq;
-		netfs_get_request(rreq, netfs_rreq_trace_get_subreq);
-		netfs_stat(&netfs_n_rh_sreq);
+	mempool_t *mempool = rreq->netfs_ops->subrequest_pool ?: &netfs_subrequest_pool;
+	struct kmem_cache *cache = mempool->pool_data;
+
+	for (;;) {
+		subreq = mempool_alloc(rreq->netfs_ops->subrequest_pool ?: &netfs_subrequest_pool,
+				       GFP_KERNEL);
+		if (subreq)
+			break;
+		msleep(10);
 	}
 
+	memset(subreq, 0, kmem_cache_size(cache));
+	INIT_WORK(&subreq->work, NULL);
+	INIT_LIST_HEAD(&subreq->rreq_link);
+	refcount_set(&subreq->ref, 2);
+	subreq->rreq = rreq;
+	subreq->debug_index = atomic_inc_return(&rreq->subreq_counter);
+	netfs_get_request(rreq, netfs_rreq_trace_get_subreq);
+	netfs_stat(&netfs_n_rh_sreq);
 	return subreq;
 }
 
@@ -173,7 +219,7 @@ static void netfs_free_subrequest(struct netfs_io_subrequest *subreq,
 	trace_netfs_sreq(subreq, netfs_sreq_trace_free);
 	if (rreq->netfs_ops->free_subrequest)
 		rreq->netfs_ops->free_subrequest(subreq);
-	kfree(subreq);
+	mempool_free(subreq, rreq->netfs_ops->subrequest_pool ?: &netfs_subrequest_pool);
 	netfs_stat_d(&netfs_n_rh_sreq);
 	netfs_put_request(rreq, was_async, netfs_rreq_trace_put_subreq);
 }
diff --git a/fs/netfs/output.c b/fs/netfs/output.c
deleted file mode 100644
index 625eb68f3e..0000000000
--- a/fs/netfs/output.c
+++ /dev/null
@@ -1,478 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-only
-/* Network filesystem high-level write support.
- *
- * Copyright (C) 2023 Red Hat, Inc. All Rights Reserved.
- * Written by David Howells (dhowells@redhat.com)
- */
-
-#include <linux/fs.h>
-#include <linux/mm.h>
-#include <linux/pagemap.h>
-#include <linux/slab.h>
-#include <linux/writeback.h>
-#include <linux/pagevec.h>
-#include "internal.h"
-
-/**
- * netfs_create_write_request - Create a write operation.
- * @wreq: The write request this is storing from.
- * @dest: The destination type
- * @start: Start of the region this write will modify
- * @len: Length of the modification
- * @worker: The worker function to handle the write(s)
- *
- * Allocate a write operation, set it up and add it to the list on a write
- * request.
- */
-struct netfs_io_subrequest *netfs_create_write_request(struct netfs_io_request *wreq,
-						       enum netfs_io_source dest,
-						       loff_t start, size_t len,
-						       work_func_t worker)
-{
-	struct netfs_io_subrequest *subreq;
-
-	subreq = netfs_alloc_subrequest(wreq);
-	if (subreq) {
-		INIT_WORK(&subreq->work, worker);
-		subreq->source	= dest;
-		subreq->start	= start;
-		subreq->len	= len;
-		subreq->debug_index = wreq->subreq_counter++;
-
-		switch (subreq->source) {
-		case NETFS_UPLOAD_TO_SERVER:
-			netfs_stat(&netfs_n_wh_upload);
-			break;
-		case NETFS_WRITE_TO_CACHE:
-			netfs_stat(&netfs_n_wh_write);
-			break;
-		default:
-			BUG();
-		}
-
-		subreq->io_iter = wreq->io_iter;
-		iov_iter_advance(&subreq->io_iter, subreq->start - wreq->start);
-		iov_iter_truncate(&subreq->io_iter, subreq->len);
-
-		trace_netfs_sreq_ref(wreq->debug_id, subreq->debug_index,
-				     refcount_read(&subreq->ref),
-				     netfs_sreq_trace_new);
-		atomic_inc(&wreq->nr_outstanding);
-		list_add_tail(&subreq->rreq_link, &wreq->subrequests);
-		trace_netfs_sreq(subreq, netfs_sreq_trace_prepare);
-	}
-
-	return subreq;
-}
-EXPORT_SYMBOL(netfs_create_write_request);
-
-/*
- * Process a completed write request once all the component operations have
- * been completed.
- */
-static void netfs_write_terminated(struct netfs_io_request *wreq, bool was_async)
-{
-	struct netfs_io_subrequest *subreq;
-	struct netfs_inode *ctx = netfs_inode(wreq->inode);
-	size_t transferred = 0;
-
-	_enter("R=%x[]", wreq->debug_id);
-
-	trace_netfs_rreq(wreq, netfs_rreq_trace_write_done);
-
-	list_for_each_entry(subreq, &wreq->subrequests, rreq_link) {
-		if (subreq->error || subreq->transferred == 0)
-			break;
-		transferred += subreq->transferred;
-		if (subreq->transferred < subreq->len)
-			break;
-	}
-	wreq->transferred = transferred;
-
-	list_for_each_entry(subreq, &wreq->subrequests, rreq_link) {
-		if (!subreq->error)
-			continue;
-		switch (subreq->source) {
-		case NETFS_UPLOAD_TO_SERVER:
-			/* Depending on the type of failure, this may prevent
-			 * writeback completion unless we're in disconnected
-			 * mode.
-			 */
-			if (!wreq->error)
-				wreq->error = subreq->error;
-			break;
-
-		case NETFS_WRITE_TO_CACHE:
-			/* Failure doesn't prevent writeback completion unless
-			 * we're in disconnected mode.
-			 */
-			if (subreq->error != -ENOBUFS)
-				ctx->ops->invalidate_cache(wreq);
-			break;
-
-		default:
-			WARN_ON_ONCE(1);
-			if (!wreq->error)
-				wreq->error = -EIO;
-			return;
-		}
-	}
-
-	wreq->cleanup(wreq);
-
-	if (wreq->origin == NETFS_DIO_WRITE &&
-	    wreq->mapping->nrpages) {
-		pgoff_t first = wreq->start >> PAGE_SHIFT;
-		pgoff_t last = (wreq->start + wreq->transferred - 1) >> PAGE_SHIFT;
-		invalidate_inode_pages2_range(wreq->mapping, first, last);
-	}
-
-	if (wreq->origin == NETFS_DIO_WRITE)
-		inode_dio_end(wreq->inode);
-
-	_debug("finished");
-	trace_netfs_rreq(wreq, netfs_rreq_trace_wake_ip);
-	clear_bit_unlock(NETFS_RREQ_IN_PROGRESS, &wreq->flags);
-	wake_up_bit(&wreq->flags, NETFS_RREQ_IN_PROGRESS);
-
-	if (wreq->iocb) {
-		wreq->iocb->ki_pos += transferred;
-		if (wreq->iocb->ki_complete)
-			wreq->iocb->ki_complete(
-				wreq->iocb, wreq->error ? wreq->error : transferred);
-	}
-
-	netfs_clear_subrequests(wreq, was_async);
-	netfs_put_request(wreq, was_async, netfs_rreq_trace_put_complete);
-}
-
-/*
- * Deal with the completion of writing the data to the cache.
- */
-void netfs_write_subrequest_terminated(void *_op, ssize_t transferred_or_error,
-				       bool was_async)
-{
-	struct netfs_io_subrequest *subreq = _op;
-	struct netfs_io_request *wreq = subreq->rreq;
-	unsigned int u;
-
-	_enter("%x[%x] %zd", wreq->debug_id, subreq->debug_index, transferred_or_error);
-
-	switch (subreq->source) {
-	case NETFS_UPLOAD_TO_SERVER:
-		netfs_stat(&netfs_n_wh_upload_done);
-		break;
-	case NETFS_WRITE_TO_CACHE:
-		netfs_stat(&netfs_n_wh_write_done);
-		break;
-	case NETFS_INVALID_WRITE:
-		break;
-	default:
-		BUG();
-	}
-
-	if (IS_ERR_VALUE(transferred_or_error)) {
-		subreq->error = transferred_or_error;
-		trace_netfs_failure(wreq, subreq, transferred_or_error,
-				    netfs_fail_write);
-		goto failed;
-	}
-
-	if (WARN(transferred_or_error > subreq->len - subreq->transferred,
-		 "Subreq excess write: R%x[%x] %zd > %zu - %zu",
-		 wreq->debug_id, subreq->debug_index,
-		 transferred_or_error, subreq->len, subreq->transferred))
-		transferred_or_error = subreq->len - subreq->transferred;
-
-	subreq->error = 0;
-	subreq->transferred += transferred_or_error;
-
-	if (iov_iter_count(&subreq->io_iter) != subreq->len - subreq->transferred)
-		pr_warn("R=%08x[%u] ITER POST-MISMATCH %zx != %zx-%zx %x\n",
-			wreq->debug_id, subreq->debug_index,
-			iov_iter_count(&subreq->io_iter), subreq->len,
-			subreq->transferred, subreq->io_iter.iter_type);
-
-	if (subreq->transferred < subreq->len)
-		goto incomplete;
-
-	__clear_bit(NETFS_SREQ_NO_PROGRESS, &subreq->flags);
-out:
-	trace_netfs_sreq(subreq, netfs_sreq_trace_terminated);
-
-	/* If we decrement nr_outstanding to 0, the ref belongs to us. */
-	u = atomic_dec_return(&wreq->nr_outstanding);
-	if (u == 0)
-		netfs_write_terminated(wreq, was_async);
-	else if (u == 1)
-		wake_up_var(&wreq->nr_outstanding);
-
-	netfs_put_subrequest(subreq, was_async, netfs_sreq_trace_put_terminated);
-	return;
-
-incomplete:
-	if (transferred_or_error == 0) {
-		if (__test_and_set_bit(NETFS_SREQ_NO_PROGRESS, &subreq->flags)) {
-			subreq->error = -ENODATA;
-			goto failed;
-		}
-	} else {
-		__clear_bit(NETFS_SREQ_NO_PROGRESS, &subreq->flags);
-	}
-
-	__set_bit(NETFS_SREQ_SHORT_IO, &subreq->flags);
-	set_bit(NETFS_RREQ_INCOMPLETE_IO, &wreq->flags);
-	goto out;
-
-failed:
-	switch (subreq->source) {
-	case NETFS_WRITE_TO_CACHE:
-		netfs_stat(&netfs_n_wh_write_failed);
-		set_bit(NETFS_RREQ_INCOMPLETE_IO, &wreq->flags);
-		break;
-	case NETFS_UPLOAD_TO_SERVER:
-		netfs_stat(&netfs_n_wh_upload_failed);
-		set_bit(NETFS_RREQ_FAILED, &wreq->flags);
-		wreq->error = subreq->error;
-		break;
-	default:
-		break;
-	}
-	goto out;
-}
-EXPORT_SYMBOL(netfs_write_subrequest_terminated);
-
-static void netfs_write_to_cache_op(struct netfs_io_subrequest *subreq)
-{
-	struct netfs_io_request *wreq = subreq->rreq;
-	struct netfs_cache_resources *cres = &wreq->cache_resources;
-
-	trace_netfs_sreq(subreq, netfs_sreq_trace_submit);
-
-	cres->ops->write(cres, subreq->start, &subreq->io_iter,
-			 netfs_write_subrequest_terminated, subreq);
-}
-
-static void netfs_write_to_cache_op_worker(struct work_struct *work)
-{
-	struct netfs_io_subrequest *subreq =
-		container_of(work, struct netfs_io_subrequest, work);
-
-	netfs_write_to_cache_op(subreq);
-}
-
-/**
- * netfs_queue_write_request - Queue a write request for attention
- * @subreq: The write request to be queued
- *
- * Queue the specified write request for processing by a worker thread.  We
- * pass the caller's ref on the request to the worker thread.
- */
-void netfs_queue_write_request(struct netfs_io_subrequest *subreq)
-{
-	if (!queue_work(system_unbound_wq, &subreq->work))
-		netfs_put_subrequest(subreq, false, netfs_sreq_trace_put_wip);
-}
-EXPORT_SYMBOL(netfs_queue_write_request);
-
-/*
- * Set up a op for writing to the cache.
- */
-static void netfs_set_up_write_to_cache(struct netfs_io_request *wreq)
-{
-	struct netfs_cache_resources *cres = &wreq->cache_resources;
-	struct netfs_io_subrequest *subreq;
-	struct netfs_inode *ctx = netfs_inode(wreq->inode);
-	struct fscache_cookie *cookie = netfs_i_cookie(ctx);
-	loff_t start = wreq->start;
-	size_t len = wreq->len;
-	int ret;
-
-	if (!fscache_cookie_enabled(cookie)) {
-		clear_bit(NETFS_RREQ_WRITE_TO_CACHE, &wreq->flags);
-		return;
-	}
-
-	_debug("write to cache");
-	ret = fscache_begin_write_operation(cres, cookie);
-	if (ret < 0)
-		return;
-
-	ret = cres->ops->prepare_write(cres, &start, &len, wreq->upper_len,
-				       i_size_read(wreq->inode), true);
-	if (ret < 0)
-		return;
-
-	subreq = netfs_create_write_request(wreq, NETFS_WRITE_TO_CACHE, start, len,
-					    netfs_write_to_cache_op_worker);
-	if (!subreq)
-		return;
-
-	netfs_write_to_cache_op(subreq);
-}
-
-/*
- * Begin the process of writing out a chunk of data.
- *
- * We are given a write request that holds a series of dirty regions and
- * (partially) covers a sequence of folios, all of which are present.  The
- * pages must have been marked as writeback as appropriate.
- *
- * We need to perform the following steps:
- *
- * (1) If encrypting, create an output buffer and encrypt each block of the
- *     data into it, otherwise the output buffer will point to the original
- *     folios.
- *
- * (2) If the data is to be cached, set up a write op for the entire output
- *     buffer to the cache, if the cache wants to accept it.
- *
- * (3) If the data is to be uploaded (ie. not merely cached):
- *
- *     (a) If the data is to be compressed, create a compression buffer and
- *         compress the data into it.
- *
- *     (b) For each destination we want to upload to, set up write ops to write
- *         to that destination.  We may need multiple writes if the data is not
- *         contiguous or the span exceeds wsize for a server.
- */
-int netfs_begin_write(struct netfs_io_request *wreq, bool may_wait,
-		      enum netfs_write_trace what)
-{
-	struct netfs_inode *ctx = netfs_inode(wreq->inode);
-
-	_enter("R=%x %llx-%llx f=%lx",
-	       wreq->debug_id, wreq->start, wreq->start + wreq->len - 1,
-	       wreq->flags);
-
-	trace_netfs_write(wreq, what);
-	if (wreq->len == 0 || wreq->iter.count == 0) {
-		pr_err("Zero-sized write [R=%x]\n", wreq->debug_id);
-		return -EIO;
-	}
-
-	if (wreq->origin == NETFS_DIO_WRITE)
-		inode_dio_begin(wreq->inode);
-
-	wreq->io_iter = wreq->iter;
-
-	/* ->outstanding > 0 carries a ref */
-	netfs_get_request(wreq, netfs_rreq_trace_get_for_outstanding);
-	atomic_set(&wreq->nr_outstanding, 1);
-
-	/* Start the encryption/compression going.  We can do that in the
-	 * background whilst we generate a list of write ops that we want to
-	 * perform.
-	 */
-	// TODO: Encrypt or compress the region as appropriate
-
-	/* We need to write all of the region to the cache */
-	if (test_bit(NETFS_RREQ_WRITE_TO_CACHE, &wreq->flags))
-		netfs_set_up_write_to_cache(wreq);
-
-	/* However, we don't necessarily write all of the region to the server.
-	 * Caching of reads is being managed this way also.
-	 */
-	if (test_bit(NETFS_RREQ_UPLOAD_TO_SERVER, &wreq->flags))
-		ctx->ops->create_write_requests(wreq, wreq->start, wreq->len);
-
-	if (atomic_dec_and_test(&wreq->nr_outstanding))
-		netfs_write_terminated(wreq, false);
-
-	if (!may_wait)
-		return -EIOCBQUEUED;
-
-	wait_on_bit(&wreq->flags, NETFS_RREQ_IN_PROGRESS,
-		    TASK_UNINTERRUPTIBLE);
-	return wreq->error;
-}
-
-/*
- * Begin a write operation for writing through the pagecache.
- */
-struct netfs_io_request *netfs_begin_writethrough(struct kiocb *iocb, size_t len)
-{
-	struct netfs_io_request *wreq;
-	struct file *file = iocb->ki_filp;
-
-	wreq = netfs_alloc_request(file->f_mapping, file, iocb->ki_pos, len,
-				   NETFS_WRITETHROUGH);
-	if (IS_ERR(wreq))
-		return wreq;
-
-	trace_netfs_write(wreq, netfs_write_trace_writethrough);
-
-	__set_bit(NETFS_RREQ_UPLOAD_TO_SERVER, &wreq->flags);
-	iov_iter_xarray(&wreq->iter, ITER_SOURCE, &wreq->mapping->i_pages, wreq->start, 0);
-	wreq->io_iter = wreq->iter;
-
-	/* ->outstanding > 0 carries a ref */
-	netfs_get_request(wreq, netfs_rreq_trace_get_for_outstanding);
-	atomic_set(&wreq->nr_outstanding, 1);
-	return wreq;
-}
-
-static void netfs_submit_writethrough(struct netfs_io_request *wreq, bool final)
-{
-	struct netfs_inode *ictx = netfs_inode(wreq->inode);
-	unsigned long long start;
-	size_t len;
-
-	if (!test_bit(NETFS_RREQ_UPLOAD_TO_SERVER, &wreq->flags))
-		return;
-
-	start = wreq->start + wreq->submitted;
-	len = wreq->iter.count - wreq->submitted;
-	if (!final) {
-		len /= wreq->wsize; /* Round to number of maximum packets */
-		len *= wreq->wsize;
-	}
-
-	ictx->ops->create_write_requests(wreq, start, len);
-	wreq->submitted += len;
-}
-
-/*
- * Advance the state of the write operation used when writing through the
- * pagecache.  Data has been copied into the pagecache that we need to append
- * to the request.  If we've added more than wsize then we need to create a new
- * subrequest.
- */
-int netfs_advance_writethrough(struct netfs_io_request *wreq, size_t copied, bool to_page_end)
-{
-	_enter("ic=%zu sb=%zu ws=%u cp=%zu tp=%u",
-	       wreq->iter.count, wreq->submitted, wreq->wsize, copied, to_page_end);
-
-	wreq->iter.count += copied;
-	wreq->io_iter.count += copied;
-	if (to_page_end && wreq->io_iter.count - wreq->submitted >= wreq->wsize)
-		netfs_submit_writethrough(wreq, false);
-
-	return wreq->error;
-}
-
-/*
- * End a write operation used when writing through the pagecache.
- */
-int netfs_end_writethrough(struct netfs_io_request *wreq, struct kiocb *iocb)
-{
-	int ret = -EIOCBQUEUED;
-
-	_enter("ic=%zu sb=%zu ws=%u",
-	       wreq->iter.count, wreq->submitted, wreq->wsize);
-
-	if (wreq->submitted < wreq->io_iter.count)
-		netfs_submit_writethrough(wreq, true);
-
-	if (atomic_dec_and_test(&wreq->nr_outstanding))
-		netfs_write_terminated(wreq, false);
-
-	if (is_sync_kiocb(iocb)) {
-		wait_on_bit(&wreq->flags, NETFS_RREQ_IN_PROGRESS,
-			    TASK_UNINTERRUPTIBLE);
-		ret = wreq->error;
-	}
-
-	netfs_put_request(wreq, false, netfs_rreq_trace_put_return);
-	return ret;
-}
diff --git a/fs/netfs/stats.c b/fs/netfs/stats.c
index deeba9f9dc..0892768eea 100644
--- a/fs/netfs/stats.c
+++ b/fs/netfs/stats.c
@@ -10,9 +10,8 @@
 #include "internal.h"
 
 atomic_t netfs_n_rh_dio_read;
-atomic_t netfs_n_rh_dio_write;
 atomic_t netfs_n_rh_readahead;
-atomic_t netfs_n_rh_readpage;
+atomic_t netfs_n_rh_read_folio;
 atomic_t netfs_n_rh_rreq;
 atomic_t netfs_n_rh_sreq;
 atomic_t netfs_n_rh_download;
@@ -29,6 +28,10 @@ atomic_t netfs_n_rh_write_begin;
 atomic_t netfs_n_rh_write_done;
 atomic_t netfs_n_rh_write_failed;
 atomic_t netfs_n_rh_write_zskip;
+atomic_t netfs_n_wh_buffered_write;
+atomic_t netfs_n_wh_writethrough;
+atomic_t netfs_n_wh_dio_write;
+atomic_t netfs_n_wh_writepages;
 atomic_t netfs_n_wh_wstream_conflict;
 atomic_t netfs_n_wh_upload;
 atomic_t netfs_n_wh_upload_done;
@@ -39,13 +42,17 @@ atomic_t netfs_n_wh_write_failed;
 
 int netfs_stats_show(struct seq_file *m, void *v)
 {
-	seq_printf(m, "Netfs  : DR=%u DW=%u RA=%u RP=%u WB=%u WBZ=%u\n",
+	seq_printf(m, "Netfs  : DR=%u RA=%u RF=%u WB=%u WBZ=%u\n",
 		   atomic_read(&netfs_n_rh_dio_read),
-		   atomic_read(&netfs_n_rh_dio_write),
 		   atomic_read(&netfs_n_rh_readahead),
-		   atomic_read(&netfs_n_rh_readpage),
+		   atomic_read(&netfs_n_rh_read_folio),
 		   atomic_read(&netfs_n_rh_write_begin),
 		   atomic_read(&netfs_n_rh_write_zskip));
+	seq_printf(m, "Netfs  : BW=%u WT=%u DW=%u WP=%u\n",
+		   atomic_read(&netfs_n_wh_buffered_write),
+		   atomic_read(&netfs_n_wh_writethrough),
+		   atomic_read(&netfs_n_wh_dio_write),
+		   atomic_read(&netfs_n_wh_writepages));
 	seq_printf(m, "Netfs  : ZR=%u sh=%u sk=%u\n",
 		   atomic_read(&netfs_n_rh_zero),
 		   atomic_read(&netfs_n_rh_short_read),
diff --git a/fs/netfs/write_collect.c b/fs/netfs/write_collect.c
new file mode 100644
index 0000000000..488147439f
--- /dev/null
+++ b/fs/netfs/write_collect.c
@@ -0,0 +1,809 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Network filesystem write subrequest result collection, assessment
+ * and retrying.
+ *
+ * Copyright (C) 2024 Red Hat, Inc. All Rights Reserved.
+ * Written by David Howells (dhowells@redhat.com)
+ */
+
+#include <linux/export.h>
+#include <linux/fs.h>
+#include <linux/mm.h>
+#include <linux/pagemap.h>
+#include <linux/slab.h>
+#include "internal.h"
+
+/* Notes made in the collector */
+#define HIT_PENDING		0x01	/* A front op was still pending */
+#define SOME_EMPTY		0x02	/* One of more streams are empty */
+#define ALL_EMPTY		0x04	/* All streams are empty */
+#define MAYBE_DISCONTIG		0x08	/* A front op may be discontiguous (rounded to PAGE_SIZE) */
+#define NEED_REASSESS		0x10	/* Need to loop round and reassess */
+#define REASSESS_DISCONTIG	0x20	/* Reassess discontiguity if contiguity advances */
+#define MADE_PROGRESS		0x40	/* Made progress cleaning up a stream or the folio set */
+#define BUFFERED		0x80	/* The pagecache needs cleaning up */
+#define NEED_RETRY		0x100	/* A front op requests retrying */
+#define SAW_FAILURE		0x200	/* One stream or hit a permanent failure */
+
+/*
+ * Successful completion of write of a folio to the server and/or cache.  Note
+ * that we are not allowed to lock the folio here on pain of deadlocking with
+ * truncate.
+ */
+int netfs_folio_written_back(struct folio *folio)
+{
+	enum netfs_folio_trace why = netfs_folio_trace_clear;
+	struct netfs_folio *finfo;
+	struct netfs_group *group = NULL;
+	int gcount = 0;
+
+	if ((finfo = netfs_folio_info(folio))) {
+		/* Streaming writes cannot be redirtied whilst under writeback,
+		 * so discard the streaming record.
+		 */
+		folio_detach_private(folio);
+		group = finfo->netfs_group;
+		gcount++;
+		kfree(finfo);
+		why = netfs_folio_trace_clear_s;
+		goto end_wb;
+	}
+
+	if ((group = netfs_folio_group(folio))) {
+		if (group == NETFS_FOLIO_COPY_TO_CACHE) {
+			why = netfs_folio_trace_clear_cc;
+			folio_detach_private(folio);
+			goto end_wb;
+		}
+
+		/* Need to detach the group pointer if the page didn't get
+		 * redirtied.  If it has been redirtied, then it must be within
+		 * the same group.
+		 */
+		why = netfs_folio_trace_redirtied;
+		if (!folio_test_dirty(folio)) {
+			folio_detach_private(folio);
+			gcount++;
+			why = netfs_folio_trace_clear_g;
+		}
+	}
+
+end_wb:
+	trace_netfs_folio(folio, why);
+	folio_end_writeback(folio);
+	return gcount;
+}
+
+/*
+ * Get hold of a folio we have under writeback.  We don't want to get the
+ * refcount on it.
+ */
+static struct folio *netfs_writeback_lookup_folio(struct netfs_io_request *wreq, loff_t pos)
+{
+	XA_STATE(xas, &wreq->mapping->i_pages, pos / PAGE_SIZE);
+	struct folio *folio;
+
+	rcu_read_lock();
+
+	for (;;) {
+		xas_reset(&xas);
+		folio = xas_load(&xas);
+		if (xas_retry(&xas, folio))
+			continue;
+
+		if (!folio || xa_is_value(folio))
+			kdebug("R=%08x: folio %lx (%llx) not present",
+			       wreq->debug_id, xas.xa_index, pos / PAGE_SIZE);
+		BUG_ON(!folio || xa_is_value(folio));
+
+		if (folio == xas_reload(&xas))
+			break;
+	}
+
+	rcu_read_unlock();
+
+	if (WARN_ONCE(!folio_test_writeback(folio),
+		      "R=%08x: folio %lx is not under writeback\n",
+		      wreq->debug_id, folio->index)) {
+		trace_netfs_folio(folio, netfs_folio_trace_not_under_wback);
+	}
+	return folio;
+}
+
+/*
+ * Unlock any folios we've finished with.
+ */
+static void netfs_writeback_unlock_folios(struct netfs_io_request *wreq,
+					  unsigned long long collected_to,
+					  unsigned int *notes)
+{
+	for (;;) {
+		struct folio *folio;
+		struct netfs_folio *finfo;
+		unsigned long long fpos, fend;
+		size_t fsize, flen;
+
+		folio = netfs_writeback_lookup_folio(wreq, wreq->cleaned_to);
+
+		fpos = folio_pos(folio);
+		fsize = folio_size(folio);
+		finfo = netfs_folio_info(folio);
+		flen = finfo ? finfo->dirty_offset + finfo->dirty_len : fsize;
+
+		fend = min_t(unsigned long long, fpos + flen, wreq->i_size);
+
+		trace_netfs_collect_folio(wreq, folio, fend, collected_to);
+
+		if (fpos + fsize > wreq->contiguity) {
+			trace_netfs_collect_contig(wreq, fpos + fsize,
+						   netfs_contig_trace_unlock);
+			wreq->contiguity = fpos + fsize;
+		}
+
+		/* Unlock any folio we've transferred all of. */
+		if (collected_to < fend)
+			break;
+
+		wreq->nr_group_rel += netfs_folio_written_back(folio);
+		wreq->cleaned_to = fpos + fsize;
+		*notes |= MADE_PROGRESS;
+
+		if (fpos + fsize >= collected_to)
+			break;
+	}
+}
+
+/*
+ * Perform retries on the streams that need it.
+ */
+static void netfs_retry_write_stream(struct netfs_io_request *wreq,
+				     struct netfs_io_stream *stream)
+{
+	struct list_head *next;
+
+	kenter("R=%x[%x:]", wreq->debug_id, stream->stream_nr);
+
+	if (list_empty(&stream->subrequests))
+		return;
+
+	if (stream->source == NETFS_UPLOAD_TO_SERVER &&
+	    wreq->netfs_ops->retry_request)
+		wreq->netfs_ops->retry_request(wreq, stream);
+
+	if (unlikely(stream->failed))
+		return;
+
+	/* If there's no renegotiation to do, just resend each failed subreq. */
+	if (!stream->prepare_write) {
+		struct netfs_io_subrequest *subreq;
+
+		list_for_each_entry(subreq, &stream->subrequests, rreq_link) {
+			if (test_bit(NETFS_SREQ_FAILED, &subreq->flags))
+				break;
+			if (__test_and_clear_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags)) {
+				__set_bit(NETFS_SREQ_RETRYING, &subreq->flags);
+				netfs_get_subrequest(subreq, netfs_sreq_trace_get_resubmit);
+				netfs_reissue_write(stream, subreq);
+			}
+		}
+		return;
+	}
+
+	next = stream->subrequests.next;
+
+	do {
+		struct netfs_io_subrequest *subreq = NULL, *from, *to, *tmp;
+		unsigned long long start, len;
+		size_t part;
+		bool boundary = false;
+
+		/* Go through the stream and find the next span of contiguous
+		 * data that we then rejig (cifs, for example, needs the wsize
+		 * renegotiating) and reissue.
+		 */
+		from = list_entry(next, struct netfs_io_subrequest, rreq_link);
+		to = from;
+		start = from->start + from->transferred;
+		len   = from->len   - from->transferred;
+
+		if (test_bit(NETFS_SREQ_FAILED, &from->flags) ||
+		    !test_bit(NETFS_SREQ_NEED_RETRY, &from->flags))
+			return;
+
+		list_for_each_continue(next, &stream->subrequests) {
+			subreq = list_entry(next, struct netfs_io_subrequest, rreq_link);
+			if (subreq->start + subreq->transferred != start + len ||
+			    test_bit(NETFS_SREQ_BOUNDARY, &subreq->flags) ||
+			    !test_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags))
+				break;
+			to = subreq;
+			len += to->len;
+		}
+
+		/* Work through the sublist. */
+		subreq = from;
+		list_for_each_entry_from(subreq, &stream->subrequests, rreq_link) {
+			if (!len)
+				break;
+			/* Renegotiate max_len (wsize) */
+			trace_netfs_sreq(subreq, netfs_sreq_trace_retry);
+			__clear_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags);
+			__set_bit(NETFS_SREQ_RETRYING, &subreq->flags);
+			stream->prepare_write(subreq);
+
+			part = min(len, subreq->max_len);
+			subreq->len = part;
+			subreq->start = start;
+			subreq->transferred = 0;
+			len -= part;
+			start += part;
+			if (len && subreq == to &&
+			    __test_and_clear_bit(NETFS_SREQ_BOUNDARY, &to->flags))
+				boundary = true;
+
+			netfs_get_subrequest(subreq, netfs_sreq_trace_get_resubmit);
+			netfs_reissue_write(stream, subreq);
+			if (subreq == to)
+				break;
+		}
+
+		/* If we managed to use fewer subreqs, we can discard the
+		 * excess; if we used the same number, then we're done.
+		 */
+		if (!len) {
+			if (subreq == to)
+				continue;
+			list_for_each_entry_safe_from(subreq, tmp,
+						      &stream->subrequests, rreq_link) {
+				trace_netfs_sreq(subreq, netfs_sreq_trace_discard);
+				list_del(&subreq->rreq_link);
+				netfs_put_subrequest(subreq, false, netfs_sreq_trace_put_done);
+				if (subreq == to)
+					break;
+			}
+			continue;
+		}
+
+		/* We ran out of subrequests, so we need to allocate some more
+		 * and insert them after.
+		 */
+		do {
+			subreq = netfs_alloc_subrequest(wreq);
+			subreq->source		= to->source;
+			subreq->start		= start;
+			subreq->max_len		= len;
+			subreq->max_nr_segs	= INT_MAX;
+			subreq->debug_index	= atomic_inc_return(&wreq->subreq_counter);
+			subreq->stream_nr	= to->stream_nr;
+			__set_bit(NETFS_SREQ_RETRYING, &subreq->flags);
+
+			trace_netfs_sreq_ref(wreq->debug_id, subreq->debug_index,
+					     refcount_read(&subreq->ref),
+					     netfs_sreq_trace_new);
+			netfs_get_subrequest(subreq, netfs_sreq_trace_get_resubmit);
+
+			list_add(&subreq->rreq_link, &to->rreq_link);
+			to = list_next_entry(to, rreq_link);
+			trace_netfs_sreq(subreq, netfs_sreq_trace_retry);
+
+			switch (stream->source) {
+			case NETFS_UPLOAD_TO_SERVER:
+				netfs_stat(&netfs_n_wh_upload);
+				subreq->max_len = min(len, wreq->wsize);
+				break;
+			case NETFS_WRITE_TO_CACHE:
+				netfs_stat(&netfs_n_wh_write);
+				break;
+			default:
+				WARN_ON_ONCE(1);
+			}
+
+			stream->prepare_write(subreq);
+
+			part = min(len, subreq->max_len);
+			subreq->len = subreq->transferred + part;
+			len -= part;
+			start += part;
+			if (!len && boundary) {
+				__set_bit(NETFS_SREQ_BOUNDARY, &to->flags);
+				boundary = false;
+			}
+
+			netfs_reissue_write(stream, subreq);
+			if (!len)
+				break;
+
+		} while (len);
+
+	} while (!list_is_head(next, &stream->subrequests));
+}
+
+/*
+ * Perform retries on the streams that need it.  If we're doing content
+ * encryption and the server copy changed due to a third-party write, we may
+ * need to do an RMW cycle and also rewrite the data to the cache.
+ */
+static void netfs_retry_writes(struct netfs_io_request *wreq)
+{
+	struct netfs_io_subrequest *subreq;
+	struct netfs_io_stream *stream;
+	int s;
+
+	/* Wait for all outstanding I/O to quiesce before performing retries as
+	 * we may need to renegotiate the I/O sizes.
+	 */
+	for (s = 0; s < NR_IO_STREAMS; s++) {
+		stream = &wreq->io_streams[s];
+		if (!stream->active)
+			continue;
+
+		list_for_each_entry(subreq, &stream->subrequests, rreq_link) {
+			wait_on_bit(&subreq->flags, NETFS_SREQ_IN_PROGRESS,
+				    TASK_UNINTERRUPTIBLE);
+		}
+	}
+
+	// TODO: Enc: Fetch changed partial pages
+	// TODO: Enc: Reencrypt content if needed.
+	// TODO: Enc: Wind back transferred point.
+	// TODO: Enc: Mark cache pages for retry.
+
+	for (s = 0; s < NR_IO_STREAMS; s++) {
+		stream = &wreq->io_streams[s];
+		if (stream->need_retry) {
+			stream->need_retry = false;
+			netfs_retry_write_stream(wreq, stream);
+		}
+	}
+}
+
+/*
+ * Collect and assess the results of various write subrequests.  We may need to
+ * retry some of the results - or even do an RMW cycle for content crypto.
+ *
+ * Note that we have a number of parallel, overlapping lists of subrequests,
+ * one to the server and one to the local cache for example, which may not be
+ * the same size or starting position and may not even correspond in boundary
+ * alignment.
+ */
+static void netfs_collect_write_results(struct netfs_io_request *wreq)
+{
+	struct netfs_io_subrequest *front, *remove;
+	struct netfs_io_stream *stream;
+	unsigned long long collected_to;
+	unsigned int notes;
+	int s;
+
+	kenter("%llx-%llx", wreq->start, wreq->start + wreq->len);
+	trace_netfs_collect(wreq);
+	trace_netfs_rreq(wreq, netfs_rreq_trace_collect);
+
+reassess_streams:
+	smp_rmb();
+	collected_to = ULLONG_MAX;
+	if (wreq->origin == NETFS_WRITEBACK)
+		notes = ALL_EMPTY | BUFFERED | MAYBE_DISCONTIG;
+	else if (wreq->origin == NETFS_WRITETHROUGH)
+		notes = ALL_EMPTY | BUFFERED;
+	else
+		notes = ALL_EMPTY;
+
+	/* Remove completed subrequests from the front of the streams and
+	 * advance the completion point on each stream.  We stop when we hit
+	 * something that's in progress.  The issuer thread may be adding stuff
+	 * to the tail whilst we're doing this.
+	 *
+	 * We must not, however, merge in discontiguities that span whole
+	 * folios that aren't under writeback.  This is made more complicated
+	 * by the folios in the gap being of unpredictable sizes - if they even
+	 * exist - but we don't want to look them up.
+	 */
+	for (s = 0; s < NR_IO_STREAMS; s++) {
+		loff_t rstart, rend;
+
+		stream = &wreq->io_streams[s];
+		/* Read active flag before list pointers */
+		if (!smp_load_acquire(&stream->active))
+			continue;
+
+		front = stream->front;
+		while (front) {
+			trace_netfs_collect_sreq(wreq, front);
+			//kdebug("sreq [%x] %llx %zx/%zx",
+			//       front->debug_index, front->start, front->transferred, front->len);
+
+			/* Stall if there may be a discontinuity. */
+			rstart = round_down(front->start, PAGE_SIZE);
+			if (rstart > wreq->contiguity) {
+				if (wreq->contiguity > stream->collected_to) {
+					trace_netfs_collect_gap(wreq, stream,
+								wreq->contiguity, 'D');
+					stream->collected_to = wreq->contiguity;
+				}
+				notes |= REASSESS_DISCONTIG;
+				break;
+			}
+			rend = round_up(front->start + front->len, PAGE_SIZE);
+			if (rend > wreq->contiguity) {
+				trace_netfs_collect_contig(wreq, rend,
+							   netfs_contig_trace_collect);
+				wreq->contiguity = rend;
+				if (notes & REASSESS_DISCONTIG)
+					notes |= NEED_REASSESS;
+			}
+			notes &= ~MAYBE_DISCONTIG;
+
+			/* Stall if the front is still undergoing I/O. */
+			if (test_bit(NETFS_SREQ_IN_PROGRESS, &front->flags)) {
+				notes |= HIT_PENDING;
+				break;
+			}
+			smp_rmb(); /* Read counters after I-P flag. */
+
+			if (stream->failed) {
+				stream->collected_to = front->start + front->len;
+				notes |= MADE_PROGRESS | SAW_FAILURE;
+				goto cancel;
+			}
+			if (front->start + front->transferred > stream->collected_to) {
+				stream->collected_to = front->start + front->transferred;
+				stream->transferred = stream->collected_to - wreq->start;
+				notes |= MADE_PROGRESS;
+			}
+			if (test_bit(NETFS_SREQ_FAILED, &front->flags)) {
+				stream->failed = true;
+				stream->error = front->error;
+				if (stream->source == NETFS_UPLOAD_TO_SERVER)
+					mapping_set_error(wreq->mapping, front->error);
+				notes |= NEED_REASSESS | SAW_FAILURE;
+				break;
+			}
+			if (front->transferred < front->len) {
+				stream->need_retry = true;
+				notes |= NEED_RETRY | MADE_PROGRESS;
+				break;
+			}
+
+		cancel:
+			/* Remove if completely consumed. */
+			spin_lock(&wreq->lock);
+
+			remove = front;
+			list_del_init(&front->rreq_link);
+			front = list_first_entry_or_null(&stream->subrequests,
+							 struct netfs_io_subrequest, rreq_link);
+			stream->front = front;
+			if (!front) {
+				unsigned long long jump_to = atomic64_read(&wreq->issued_to);
+
+				if (stream->collected_to < jump_to) {
+					trace_netfs_collect_gap(wreq, stream, jump_to, 'A');
+					stream->collected_to = jump_to;
+				}
+			}
+
+			spin_unlock(&wreq->lock);
+			netfs_put_subrequest(remove, false,
+					     notes & SAW_FAILURE ?
+					     netfs_sreq_trace_put_cancel :
+					     netfs_sreq_trace_put_done);
+		}
+
+		if (front)
+			notes &= ~ALL_EMPTY;
+		else
+			notes |= SOME_EMPTY;
+
+		if (stream->collected_to < collected_to)
+			collected_to = stream->collected_to;
+	}
+
+	if (collected_to != ULLONG_MAX && collected_to > wreq->collected_to)
+		wreq->collected_to = collected_to;
+
+	/* If we have an empty stream, we need to jump it forward over any gap
+	 * otherwise the collection point will never advance.
+	 *
+	 * Note that the issuer always adds to the stream with the lowest
+	 * so-far submitted start, so if we see two consecutive subreqs in one
+	 * stream with nothing between then in another stream, then the second
+	 * stream has a gap that can be jumped.
+	 */
+	if (notes & SOME_EMPTY) {
+		unsigned long long jump_to = wreq->start + READ_ONCE(wreq->submitted);
+
+		for (s = 0; s < NR_IO_STREAMS; s++) {
+			stream = &wreq->io_streams[s];
+			if (stream->active &&
+			    stream->front &&
+			    stream->front->start < jump_to)
+				jump_to = stream->front->start;
+		}
+
+		for (s = 0; s < NR_IO_STREAMS; s++) {
+			stream = &wreq->io_streams[s];
+			if (stream->active &&
+			    !stream->front &&
+			    stream->collected_to < jump_to) {
+				trace_netfs_collect_gap(wreq, stream, jump_to, 'B');
+				stream->collected_to = jump_to;
+			}
+		}
+	}
+
+	for (s = 0; s < NR_IO_STREAMS; s++) {
+		stream = &wreq->io_streams[s];
+		if (stream->active)
+			trace_netfs_collect_stream(wreq, stream);
+	}
+
+	trace_netfs_collect_state(wreq, wreq->collected_to, notes);
+
+	/* Unlock any folios that we have now finished with. */
+	if (notes & BUFFERED) {
+		unsigned long long clean_to = min(wreq->collected_to, wreq->contiguity);
+
+		if (wreq->cleaned_to < clean_to)
+			netfs_writeback_unlock_folios(wreq, clean_to, &notes);
+	} else {
+		wreq->cleaned_to = wreq->collected_to;
+	}
+
+	// TODO: Discard encryption buffers
+
+	/* If all streams are discontiguous with the last folio we cleared, we
+	 * may need to skip a set of folios.
+	 */
+	if ((notes & (MAYBE_DISCONTIG | ALL_EMPTY)) == MAYBE_DISCONTIG) {
+		unsigned long long jump_to = ULLONG_MAX;
+
+		for (s = 0; s < NR_IO_STREAMS; s++) {
+			stream = &wreq->io_streams[s];
+			if (stream->active && stream->front &&
+			    stream->front->start < jump_to)
+				jump_to = stream->front->start;
+		}
+
+		trace_netfs_collect_contig(wreq, jump_to, netfs_contig_trace_jump);
+		wreq->contiguity = jump_to;
+		wreq->cleaned_to = jump_to;
+		wreq->collected_to = jump_to;
+		for (s = 0; s < NR_IO_STREAMS; s++) {
+			stream = &wreq->io_streams[s];
+			if (stream->collected_to < jump_to)
+				stream->collected_to = jump_to;
+		}
+		//cond_resched();
+		notes |= MADE_PROGRESS;
+		goto reassess_streams;
+	}
+
+	if (notes & NEED_RETRY)
+		goto need_retry;
+	if ((notes & MADE_PROGRESS) && test_bit(NETFS_RREQ_PAUSE, &wreq->flags)) {
+		trace_netfs_rreq(wreq, netfs_rreq_trace_unpause);
+		clear_bit_unlock(NETFS_RREQ_PAUSE, &wreq->flags);
+		wake_up_bit(&wreq->flags, NETFS_RREQ_PAUSE);
+	}
+
+	if (notes & NEED_REASSESS) {
+		//cond_resched();
+		goto reassess_streams;
+	}
+	if (notes & MADE_PROGRESS) {
+		//cond_resched();
+		goto reassess_streams;
+	}
+
+out:
+	netfs_put_group_many(wreq->group, wreq->nr_group_rel);
+	wreq->nr_group_rel = 0;
+	kleave(" = %x", notes);
+	return;
+
+need_retry:
+	/* Okay...  We're going to have to retry one or both streams.  Note
+	 * that any partially completed op will have had any wholly transferred
+	 * folios removed from it.
+	 */
+	kdebug("retry");
+	netfs_retry_writes(wreq);
+	goto out;
+}
+
+/*
+ * Perform the collection of subrequests, folios and encryption buffers.
+ */
+void netfs_write_collection_worker(struct work_struct *work)
+{
+	struct netfs_io_request *wreq = container_of(work, struct netfs_io_request, work);
+	struct netfs_inode *ictx = netfs_inode(wreq->inode);
+	size_t transferred;
+	int s;
+
+	kenter("R=%x", wreq->debug_id);
+
+	netfs_see_request(wreq, netfs_rreq_trace_see_work);
+	if (!test_bit(NETFS_RREQ_IN_PROGRESS, &wreq->flags)) {
+		netfs_put_request(wreq, false, netfs_rreq_trace_put_work);
+		return;
+	}
+
+	netfs_collect_write_results(wreq);
+
+	/* We're done when the app thread has finished posting subreqs and all
+	 * the queues in all the streams are empty.
+	 */
+	if (!test_bit(NETFS_RREQ_ALL_QUEUED, &wreq->flags)) {
+		netfs_put_request(wreq, false, netfs_rreq_trace_put_work);
+		return;
+	}
+	smp_rmb(); /* Read ALL_QUEUED before lists. */
+
+	transferred = LONG_MAX;
+	for (s = 0; s < NR_IO_STREAMS; s++) {
+		struct netfs_io_stream *stream = &wreq->io_streams[s];
+		if (!stream->active)
+			continue;
+		if (!list_empty(&stream->subrequests)) {
+			netfs_put_request(wreq, false, netfs_rreq_trace_put_work);
+			return;
+		}
+		if (stream->transferred < transferred)
+			transferred = stream->transferred;
+	}
+
+	/* Okay, declare that all I/O is complete. */
+	wreq->transferred = transferred;
+	trace_netfs_rreq(wreq, netfs_rreq_trace_write_done);
+
+	if (wreq->io_streams[1].active &&
+	    wreq->io_streams[1].failed) {
+		/* Cache write failure doesn't prevent writeback completion
+		 * unless we're in disconnected mode.
+		 */
+		ictx->ops->invalidate_cache(wreq);
+	}
+
+	if (wreq->cleanup)
+		wreq->cleanup(wreq);
+
+	if (wreq->origin == NETFS_DIO_WRITE &&
+	    wreq->mapping->nrpages) {
+		/* mmap may have got underfoot and we may now have folios
+		 * locally covering the region we just wrote.  Attempt to
+		 * discard the folios, but leave in place any modified locally.
+		 * ->write_iter() is prevented from interfering by the DIO
+		 * counter.
+		 */
+		pgoff_t first = wreq->start >> PAGE_SHIFT;
+		pgoff_t last = (wreq->start + wreq->transferred - 1) >> PAGE_SHIFT;
+		invalidate_inode_pages2_range(wreq->mapping, first, last);
+	}
+
+	if (wreq->origin == NETFS_DIO_WRITE)
+		inode_dio_end(wreq->inode);
+
+	kdebug("finished");
+	trace_netfs_rreq(wreq, netfs_rreq_trace_wake_ip);
+	clear_bit_unlock(NETFS_RREQ_IN_PROGRESS, &wreq->flags);
+	wake_up_bit(&wreq->flags, NETFS_RREQ_IN_PROGRESS);
+
+	if (wreq->iocb) {
+		size_t written = min(wreq->transferred, wreq->len);
+		wreq->iocb->ki_pos += written;
+		if (wreq->iocb->ki_complete)
+			wreq->iocb->ki_complete(
+				wreq->iocb, wreq->error ? wreq->error : written);
+		wreq->iocb = VFS_PTR_POISON;
+	}
+
+	netfs_clear_subrequests(wreq, false);
+	netfs_put_request(wreq, false, netfs_rreq_trace_put_work_complete);
+}
+
+/*
+ * Wake the collection work item.
+ */
+void netfs_wake_write_collector(struct netfs_io_request *wreq, bool was_async)
+{
+	if (!work_pending(&wreq->work)) {
+		netfs_get_request(wreq, netfs_rreq_trace_get_work);
+		if (!queue_work(system_unbound_wq, &wreq->work))
+			netfs_put_request(wreq, was_async, netfs_rreq_trace_put_work_nq);
+	}
+}
+
+/**
+ * netfs_write_subrequest_terminated - Note the termination of a write operation.
+ * @_op: The I/O request that has terminated.
+ * @transferred_or_error: The amount of data transferred or an error code.
+ * @was_async: The termination was asynchronous
+ *
+ * This tells the library that a contributory write I/O operation has
+ * terminated, one way or another, and that it should collect the results.
+ *
+ * The caller indicates in @transferred_or_error the outcome of the operation,
+ * supplying a positive value to indicate the number of bytes transferred or a
+ * negative error code.  The library will look after reissuing I/O operations
+ * as appropriate and writing downloaded data to the cache.
+ *
+ * If @was_async is true, the caller might be running in softirq or interrupt
+ * context and we can't sleep.
+ *
+ * When this is called, ownership of the subrequest is transferred back to the
+ * library, along with a ref.
+ *
+ * Note that %_op is a void* so that the function can be passed to
+ * kiocb::term_func without the need for a casting wrapper.
+ */
+void netfs_write_subrequest_terminated(void *_op, ssize_t transferred_or_error,
+				       bool was_async)
+{
+	struct netfs_io_subrequest *subreq = _op;
+	struct netfs_io_request *wreq = subreq->rreq;
+	struct netfs_io_stream *stream = &wreq->io_streams[subreq->stream_nr];
+
+	kenter("%x[%x] %zd", wreq->debug_id, subreq->debug_index, transferred_or_error);
+
+	switch (subreq->source) {
+	case NETFS_UPLOAD_TO_SERVER:
+		netfs_stat(&netfs_n_wh_upload_done);
+		break;
+	case NETFS_WRITE_TO_CACHE:
+		netfs_stat(&netfs_n_wh_write_done);
+		break;
+	case NETFS_INVALID_WRITE:
+		break;
+	default:
+		BUG();
+	}
+
+	if (IS_ERR_VALUE(transferred_or_error)) {
+		subreq->error = transferred_or_error;
+		if (subreq->error == -EAGAIN)
+			set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags);
+		else
+			set_bit(NETFS_SREQ_FAILED, &subreq->flags);
+		trace_netfs_failure(wreq, subreq, transferred_or_error, netfs_fail_write);
+
+		switch (subreq->source) {
+		case NETFS_WRITE_TO_CACHE:
+			netfs_stat(&netfs_n_wh_write_failed);
+			break;
+		case NETFS_UPLOAD_TO_SERVER:
+			netfs_stat(&netfs_n_wh_upload_failed);
+			break;
+		default:
+			break;
+		}
+		trace_netfs_rreq(wreq, netfs_rreq_trace_set_pause);
+		set_bit(NETFS_RREQ_PAUSE, &wreq->flags);
+	} else {
+		if (WARN(transferred_or_error > subreq->len - subreq->transferred,
+			 "Subreq excess write: R=%x[%x] %zd > %zu - %zu",
+			 wreq->debug_id, subreq->debug_index,
+			 transferred_or_error, subreq->len, subreq->transferred))
+			transferred_or_error = subreq->len - subreq->transferred;
+
+		subreq->error = 0;
+		subreq->transferred += transferred_or_error;
+
+		if (subreq->transferred < subreq->len)
+			set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags);
+	}
+
+	trace_netfs_sreq(subreq, netfs_sreq_trace_terminated);
+
+	clear_bit_unlock(NETFS_SREQ_IN_PROGRESS, &subreq->flags);
+	wake_up_bit(&subreq->flags, NETFS_SREQ_IN_PROGRESS);
+
+	/* If we are at the head of the queue, wake up the collector,
+	 * transferring a ref to it if we were the ones to do so.
+	 */
+	if (list_is_first(&subreq->rreq_link, &stream->subrequests))
+		netfs_wake_write_collector(wreq, was_async);
+
+	netfs_put_subrequest(subreq, was_async, netfs_sreq_trace_put_terminated);
+}
+EXPORT_SYMBOL(netfs_write_subrequest_terminated);
diff --git a/fs/netfs/write_issue.c b/fs/netfs/write_issue.c
new file mode 100644
index 0000000000..32bc88bee5
--- /dev/null
+++ b/fs/netfs/write_issue.c
@@ -0,0 +1,690 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Network filesystem high-level (buffered) writeback.
+ *
+ * Copyright (C) 2024 Red Hat, Inc. All Rights Reserved.
+ * Written by David Howells (dhowells@redhat.com)
+ *
+ *
+ * To support network filesystems with local caching, we manage a situation
+ * that can be envisioned like the following:
+ *
+ *               +---+---+-----+-----+---+----------+
+ *    Folios:    |   |   |     |     |   |          |
+ *               +---+---+-----+-----+---+----------+
+ *
+ *                 +------+------+     +----+----+
+ *    Upload:      |      |      |.....|    |    |
+ *  (Stream 0)     +------+------+     +----+----+
+ *
+ *               +------+------+------+------+------+
+ *    Cache:     |      |      |      |      |      |
+ *  (Stream 1)   +------+------+------+------+------+
+ *
+ * Where we have a sequence of folios of varying sizes that we need to overlay
+ * with multiple parallel streams of I/O requests, where the I/O requests in a
+ * stream may also be of various sizes (in cifs, for example, the sizes are
+ * negotiated with the server; in something like ceph, they may represent the
+ * sizes of storage objects).
+ *
+ * The sequence in each stream may contain gaps and noncontiguous subrequests
+ * may be glued together into single vectored write RPCs.
+ */
+
+#include <linux/export.h>
+#include <linux/fs.h>
+#include <linux/mm.h>
+#include <linux/pagemap.h>
+#include "internal.h"
+
+/*
+ * Kill all dirty folios in the event of an unrecoverable error, starting with
+ * a locked folio we've already obtained from writeback_iter().
+ */
+static void netfs_kill_dirty_pages(struct address_space *mapping,
+				   struct writeback_control *wbc,
+				   struct folio *folio)
+{
+	int error = 0;
+
+	do {
+		enum netfs_folio_trace why = netfs_folio_trace_kill;
+		struct netfs_group *group = NULL;
+		struct netfs_folio *finfo = NULL;
+		void *priv;
+
+		priv = folio_detach_private(folio);
+		if (priv) {
+			finfo = __netfs_folio_info(priv);
+			if (finfo) {
+				/* Kill folio from streaming write. */
+				group = finfo->netfs_group;
+				why = netfs_folio_trace_kill_s;
+			} else {
+				group = priv;
+				if (group == NETFS_FOLIO_COPY_TO_CACHE) {
+					/* Kill copy-to-cache folio */
+					why = netfs_folio_trace_kill_cc;
+					group = NULL;
+				} else {
+					/* Kill folio with group */
+					why = netfs_folio_trace_kill_g;
+				}
+			}
+		}
+
+		trace_netfs_folio(folio, why);
+
+		folio_start_writeback(folio);
+		folio_unlock(folio);
+		folio_end_writeback(folio);
+
+		netfs_put_group(group);
+		kfree(finfo);
+
+	} while ((folio = writeback_iter(mapping, wbc, folio, &error)));
+}
+
+/*
+ * Create a write request and set it up appropriately for the origin type.
+ */
+struct netfs_io_request *netfs_create_write_req(struct address_space *mapping,
+						struct file *file,
+						loff_t start,
+						enum netfs_io_origin origin)
+{
+	struct netfs_io_request *wreq;
+	struct netfs_inode *ictx;
+
+	wreq = netfs_alloc_request(mapping, file, start, 0, origin);
+	if (IS_ERR(wreq))
+		return wreq;
+
+	kenter("R=%x", wreq->debug_id);
+
+	ictx = netfs_inode(wreq->inode);
+	if (test_bit(NETFS_RREQ_WRITE_TO_CACHE, &wreq->flags))
+		fscache_begin_write_operation(&wreq->cache_resources, netfs_i_cookie(ictx));
+
+	wreq->contiguity = wreq->start;
+	wreq->cleaned_to = wreq->start;
+	INIT_WORK(&wreq->work, netfs_write_collection_worker);
+
+	wreq->io_streams[0].stream_nr		= 0;
+	wreq->io_streams[0].source		= NETFS_UPLOAD_TO_SERVER;
+	wreq->io_streams[0].prepare_write	= ictx->ops->prepare_write;
+	wreq->io_streams[0].issue_write		= ictx->ops->issue_write;
+	wreq->io_streams[0].collected_to	= start;
+	wreq->io_streams[0].transferred		= LONG_MAX;
+
+	wreq->io_streams[1].stream_nr		= 1;
+	wreq->io_streams[1].source		= NETFS_WRITE_TO_CACHE;
+	wreq->io_streams[1].collected_to	= start;
+	wreq->io_streams[1].transferred		= LONG_MAX;
+	if (fscache_resources_valid(&wreq->cache_resources)) {
+		wreq->io_streams[1].avail	= true;
+		wreq->io_streams[1].active	= true;
+		wreq->io_streams[1].prepare_write = wreq->cache_resources.ops->prepare_write_subreq;
+		wreq->io_streams[1].issue_write = wreq->cache_resources.ops->issue_write;
+	}
+
+	return wreq;
+}
+
+/**
+ * netfs_prepare_write_failed - Note write preparation failed
+ * @subreq: The subrequest to mark
+ *
+ * Mark a subrequest to note that preparation for write failed.
+ */
+void netfs_prepare_write_failed(struct netfs_io_subrequest *subreq)
+{
+	__set_bit(NETFS_SREQ_FAILED, &subreq->flags);
+	trace_netfs_sreq(subreq, netfs_sreq_trace_prep_failed);
+}
+EXPORT_SYMBOL(netfs_prepare_write_failed);
+
+/*
+ * Prepare a write subrequest.  We need to allocate a new subrequest
+ * if we don't have one.
+ */
+static void netfs_prepare_write(struct netfs_io_request *wreq,
+				struct netfs_io_stream *stream,
+				loff_t start)
+{
+	struct netfs_io_subrequest *subreq;
+
+	subreq = netfs_alloc_subrequest(wreq);
+	subreq->source		= stream->source;
+	subreq->start		= start;
+	subreq->max_len		= ULONG_MAX;
+	subreq->max_nr_segs	= INT_MAX;
+	subreq->stream_nr	= stream->stream_nr;
+
+	kenter("R=%x[%x]", wreq->debug_id, subreq->debug_index);
+
+	trace_netfs_sreq_ref(wreq->debug_id, subreq->debug_index,
+			     refcount_read(&subreq->ref),
+			     netfs_sreq_trace_new);
+
+	trace_netfs_sreq(subreq, netfs_sreq_trace_prepare);
+
+	switch (stream->source) {
+	case NETFS_UPLOAD_TO_SERVER:
+		netfs_stat(&netfs_n_wh_upload);
+		subreq->max_len = wreq->wsize;
+		break;
+	case NETFS_WRITE_TO_CACHE:
+		netfs_stat(&netfs_n_wh_write);
+		break;
+	default:
+		WARN_ON_ONCE(1);
+		break;
+	}
+
+	if (stream->prepare_write)
+		stream->prepare_write(subreq);
+
+	__set_bit(NETFS_SREQ_IN_PROGRESS, &subreq->flags);
+
+	/* We add to the end of the list whilst the collector may be walking
+	 * the list.  The collector only goes nextwards and uses the lock to
+	 * remove entries off of the front.
+	 */
+	spin_lock(&wreq->lock);
+	list_add_tail(&subreq->rreq_link, &stream->subrequests);
+	if (list_is_first(&subreq->rreq_link, &stream->subrequests)) {
+		stream->front = subreq;
+		if (!stream->active) {
+			stream->collected_to = stream->front->start;
+			/* Write list pointers before active flag */
+			smp_store_release(&stream->active, true);
+		}
+	}
+
+	spin_unlock(&wreq->lock);
+
+	stream->construct = subreq;
+}
+
+/*
+ * Set the I/O iterator for the filesystem/cache to use and dispatch the I/O
+ * operation.  The operation may be asynchronous and should call
+ * netfs_write_subrequest_terminated() when complete.
+ */
+static void netfs_do_issue_write(struct netfs_io_stream *stream,
+				 struct netfs_io_subrequest *subreq)
+{
+	struct netfs_io_request *wreq = subreq->rreq;
+
+	kenter("R=%x[%x],%zx", wreq->debug_id, subreq->debug_index, subreq->len);
+
+	if (test_bit(NETFS_SREQ_FAILED, &subreq->flags))
+		return netfs_write_subrequest_terminated(subreq, subreq->error, false);
+
+	// TODO: Use encrypted buffer
+	if (test_bit(NETFS_RREQ_USE_IO_ITER, &wreq->flags)) {
+		subreq->io_iter = wreq->io_iter;
+		iov_iter_advance(&subreq->io_iter,
+				 subreq->start + subreq->transferred - wreq->start);
+		iov_iter_truncate(&subreq->io_iter,
+				 subreq->len - subreq->transferred);
+	} else {
+		iov_iter_xarray(&subreq->io_iter, ITER_SOURCE, &wreq->mapping->i_pages,
+				subreq->start + subreq->transferred,
+				subreq->len   - subreq->transferred);
+	}
+
+	trace_netfs_sreq(subreq, netfs_sreq_trace_submit);
+	stream->issue_write(subreq);
+}
+
+void netfs_reissue_write(struct netfs_io_stream *stream,
+			 struct netfs_io_subrequest *subreq)
+{
+	__set_bit(NETFS_SREQ_IN_PROGRESS, &subreq->flags);
+	netfs_do_issue_write(stream, subreq);
+}
+
+static void netfs_issue_write(struct netfs_io_request *wreq,
+			      struct netfs_io_stream *stream)
+{
+	struct netfs_io_subrequest *subreq = stream->construct;
+
+	if (!subreq)
+		return;
+	stream->construct = NULL;
+
+	if (subreq->start + subreq->len > wreq->start + wreq->submitted)
+		WRITE_ONCE(wreq->submitted, subreq->start + subreq->len - wreq->start);
+	netfs_do_issue_write(stream, subreq);
+}
+
+/*
+ * Add data to the write subrequest, dispatching each as we fill it up or if it
+ * is discontiguous with the previous.  We only fill one part at a time so that
+ * we can avoid overrunning the credits obtained (cifs) and try to parallelise
+ * content-crypto preparation with network writes.
+ */
+int netfs_advance_write(struct netfs_io_request *wreq,
+			struct netfs_io_stream *stream,
+			loff_t start, size_t len, bool to_eof)
+{
+	struct netfs_io_subrequest *subreq = stream->construct;
+	size_t part;
+
+	if (!stream->avail) {
+		kleave("no write");
+		return len;
+	}
+
+	kenter("R=%x[%x]", wreq->debug_id, subreq ? subreq->debug_index : 0);
+
+	if (subreq && start != subreq->start + subreq->len) {
+		netfs_issue_write(wreq, stream);
+		subreq = NULL;
+	}
+
+	if (!stream->construct)
+		netfs_prepare_write(wreq, stream, start);
+	subreq = stream->construct;
+
+	part = min(subreq->max_len - subreq->len, len);
+	kdebug("part %zx/%zx %zx/%zx", subreq->len, subreq->max_len, part, len);
+	subreq->len += part;
+	subreq->nr_segs++;
+
+	if (subreq->len >= subreq->max_len ||
+	    subreq->nr_segs >= subreq->max_nr_segs ||
+	    to_eof) {
+		netfs_issue_write(wreq, stream);
+		subreq = NULL;
+	}
+
+	return part;
+}
+
+/*
+ * Write some of a pending folio data back to the server.
+ */
+static int netfs_write_folio(struct netfs_io_request *wreq,
+			     struct writeback_control *wbc,
+			     struct folio *folio)
+{
+	struct netfs_io_stream *upload = &wreq->io_streams[0];
+	struct netfs_io_stream *cache  = &wreq->io_streams[1];
+	struct netfs_io_stream *stream;
+	struct netfs_group *fgroup; /* TODO: Use this with ceph */
+	struct netfs_folio *finfo;
+	size_t fsize = folio_size(folio), flen = fsize, foff = 0;
+	loff_t fpos = folio_pos(folio), i_size;
+	bool to_eof = false, streamw = false;
+	bool debug = false;
+
+	kenter("");
+
+	/* netfs_perform_write() may shift i_size around the page or from out
+	 * of the page to beyond it, but cannot move i_size into or through the
+	 * page since we have it locked.
+	 */
+	i_size = i_size_read(wreq->inode);
+
+	if (fpos >= i_size) {
+		/* mmap beyond eof. */
+		kdebug("beyond eof");
+		folio_start_writeback(folio);
+		folio_unlock(folio);
+		wreq->nr_group_rel += netfs_folio_written_back(folio);
+		netfs_put_group_many(wreq->group, wreq->nr_group_rel);
+		wreq->nr_group_rel = 0;
+		return 0;
+	}
+
+	if (fpos + fsize > wreq->i_size)
+		wreq->i_size = i_size;
+
+	fgroup = netfs_folio_group(folio);
+	finfo = netfs_folio_info(folio);
+	if (finfo) {
+		foff = finfo->dirty_offset;
+		flen = foff + finfo->dirty_len;
+		streamw = true;
+	}
+
+	if (wreq->origin == NETFS_WRITETHROUGH) {
+		to_eof = false;
+		if (flen > i_size - fpos)
+			flen = i_size - fpos;
+	} else if (flen > i_size - fpos) {
+		flen = i_size - fpos;
+		if (!streamw)
+			folio_zero_segment(folio, flen, fsize);
+		to_eof = true;
+	} else if (flen == i_size - fpos) {
+		to_eof = true;
+	}
+	flen -= foff;
+
+	kdebug("folio %zx %zx %zx", foff, flen, fsize);
+
+	/* Deal with discontinuities in the stream of dirty pages.  These can
+	 * arise from a number of sources:
+	 *
+	 * (1) Intervening non-dirty pages from random-access writes, multiple
+	 *     flushers writing back different parts simultaneously and manual
+	 *     syncing.
+	 *
+	 * (2) Partially-written pages from write-streaming.
+	 *
+	 * (3) Pages that belong to a different write-back group (eg.  Ceph
+	 *     snapshots).
+	 *
+	 * (4) Actually-clean pages that were marked for write to the cache
+	 *     when they were read.  Note that these appear as a special
+	 *     write-back group.
+	 */
+	if (fgroup == NETFS_FOLIO_COPY_TO_CACHE) {
+		netfs_issue_write(wreq, upload);
+	} else if (fgroup != wreq->group) {
+		/* We can't write this page to the server yet. */
+		kdebug("wrong group");
+		folio_redirty_for_writepage(wbc, folio);
+		folio_unlock(folio);
+		netfs_issue_write(wreq, upload);
+		netfs_issue_write(wreq, cache);
+		return 0;
+	}
+
+	if (foff > 0)
+		netfs_issue_write(wreq, upload);
+	if (streamw)
+		netfs_issue_write(wreq, cache);
+
+	/* Flip the page to the writeback state and unlock.  If we're called
+	 * from write-through, then the page has already been put into the wb
+	 * state.
+	 */
+	if (wreq->origin == NETFS_WRITEBACK)
+		folio_start_writeback(folio);
+	folio_unlock(folio);
+
+	if (fgroup == NETFS_FOLIO_COPY_TO_CACHE) {
+		if (!fscache_resources_valid(&wreq->cache_resources)) {
+			trace_netfs_folio(folio, netfs_folio_trace_cancel_copy);
+			netfs_issue_write(wreq, upload);
+			netfs_folio_written_back(folio);
+			return 0;
+		}
+		trace_netfs_folio(folio, netfs_folio_trace_store_copy);
+	} else if (!upload->construct) {
+		trace_netfs_folio(folio, netfs_folio_trace_store);
+	} else {
+		trace_netfs_folio(folio, netfs_folio_trace_store_plus);
+	}
+
+	/* Move the submission point forward to allow for write-streaming data
+	 * not starting at the front of the page.  We don't do write-streaming
+	 * with the cache as the cache requires DIO alignment.
+	 *
+	 * Also skip uploading for data that's been read and just needs copying
+	 * to the cache.
+	 */
+	for (int s = 0; s < NR_IO_STREAMS; s++) {
+		stream = &wreq->io_streams[s];
+		stream->submit_max_len = fsize;
+		stream->submit_off = foff;
+		stream->submit_len = flen;
+		if ((stream->source == NETFS_WRITE_TO_CACHE && streamw) ||
+		    (stream->source == NETFS_UPLOAD_TO_SERVER &&
+		     fgroup == NETFS_FOLIO_COPY_TO_CACHE)) {
+			stream->submit_off = UINT_MAX;
+			stream->submit_len = 0;
+			stream->submit_max_len = 0;
+		}
+	}
+
+	/* Attach the folio to one or more subrequests.  For a big folio, we
+	 * could end up with thousands of subrequests if the wsize is small -
+	 * but we might need to wait during the creation of subrequests for
+	 * network resources (eg. SMB credits).
+	 */
+	for (;;) {
+		ssize_t part;
+		size_t lowest_off = ULONG_MAX;
+		int choose_s = -1;
+
+		/* Always add to the lowest-submitted stream first. */
+		for (int s = 0; s < NR_IO_STREAMS; s++) {
+			stream = &wreq->io_streams[s];
+			if (stream->submit_len > 0 &&
+			    stream->submit_off < lowest_off) {
+				lowest_off = stream->submit_off;
+				choose_s = s;
+			}
+		}
+
+		if (choose_s < 0)
+			break;
+		stream = &wreq->io_streams[choose_s];
+
+		part = netfs_advance_write(wreq, stream, fpos + stream->submit_off,
+					   stream->submit_len, to_eof);
+		atomic64_set(&wreq->issued_to, fpos + stream->submit_off);
+		stream->submit_off += part;
+		stream->submit_max_len -= part;
+		if (part > stream->submit_len)
+			stream->submit_len = 0;
+		else
+			stream->submit_len -= part;
+		if (part > 0)
+			debug = true;
+	}
+
+	atomic64_set(&wreq->issued_to, fpos + fsize);
+
+	if (!debug)
+		kdebug("R=%x: No submit", wreq->debug_id);
+
+	if (foff + flen < fsize)
+		for (int s = 0; s < NR_IO_STREAMS; s++)
+			netfs_issue_write(wreq, &wreq->io_streams[s]);
+
+	kleave(" = 0");
+	return 0;
+}
+
+/*
+ * Write some of the pending data back to the server
+ */
+int netfs_writepages(struct address_space *mapping,
+		     struct writeback_control *wbc)
+{
+	struct netfs_inode *ictx = netfs_inode(mapping->host);
+	struct netfs_io_request *wreq = NULL;
+	struct folio *folio;
+	int error = 0;
+
+	if (wbc->sync_mode == WB_SYNC_ALL)
+		mutex_lock(&ictx->wb_lock);
+	else if (!mutex_trylock(&ictx->wb_lock))
+		return 0;
+
+	/* Need the first folio to be able to set up the op. */
+	folio = writeback_iter(mapping, wbc, NULL, &error);
+	if (!folio)
+		goto out;
+
+	wreq = netfs_create_write_req(mapping, NULL, folio_pos(folio), NETFS_WRITEBACK);
+	if (IS_ERR(wreq)) {
+		error = PTR_ERR(wreq);
+		goto couldnt_start;
+	}
+
+	trace_netfs_write(wreq, netfs_write_trace_writeback);
+	netfs_stat(&netfs_n_wh_writepages);
+
+	do {
+		kdebug("wbiter %lx %llx", folio->index, wreq->start + wreq->submitted);
+
+		/* It appears we don't have to handle cyclic writeback wrapping. */
+		WARN_ON_ONCE(wreq && folio_pos(folio) < wreq->start + wreq->submitted);
+
+		if (netfs_folio_group(folio) != NETFS_FOLIO_COPY_TO_CACHE &&
+		    unlikely(!test_bit(NETFS_RREQ_UPLOAD_TO_SERVER, &wreq->flags))) {
+			set_bit(NETFS_RREQ_UPLOAD_TO_SERVER, &wreq->flags);
+			wreq->netfs_ops->begin_writeback(wreq);
+		}
+
+		error = netfs_write_folio(wreq, wbc, folio);
+		if (error < 0)
+			break;
+	} while ((folio = writeback_iter(mapping, wbc, folio, &error)));
+
+	for (int s = 0; s < NR_IO_STREAMS; s++)
+		netfs_issue_write(wreq, &wreq->io_streams[s]);
+	smp_wmb(); /* Write lists before ALL_QUEUED. */
+	set_bit(NETFS_RREQ_ALL_QUEUED, &wreq->flags);
+
+	mutex_unlock(&ictx->wb_lock);
+
+	netfs_put_request(wreq, false, netfs_rreq_trace_put_return);
+	kleave(" = %d", error);
+	return error;
+
+couldnt_start:
+	netfs_kill_dirty_pages(mapping, wbc, folio);
+out:
+	mutex_unlock(&ictx->wb_lock);
+	kleave(" = %d", error);
+	return error;
+}
+EXPORT_SYMBOL(netfs_writepages);
+
+/*
+ * Begin a write operation for writing through the pagecache.
+ */
+struct netfs_io_request *netfs_begin_writethrough(struct kiocb *iocb, size_t len)
+{
+	struct netfs_io_request *wreq = NULL;
+	struct netfs_inode *ictx = netfs_inode(file_inode(iocb->ki_filp));
+
+	mutex_lock(&ictx->wb_lock);
+
+	wreq = netfs_create_write_req(iocb->ki_filp->f_mapping, iocb->ki_filp,
+				      iocb->ki_pos, NETFS_WRITETHROUGH);
+	if (IS_ERR(wreq)) {
+		mutex_unlock(&ictx->wb_lock);
+		return wreq;
+	}
+
+	wreq->io_streams[0].avail = true;
+	trace_netfs_write(wreq, netfs_write_trace_writethrough);
+	return wreq;
+}
+
+/*
+ * Advance the state of the write operation used when writing through the
+ * pagecache.  Data has been copied into the pagecache that we need to append
+ * to the request.  If we've added more than wsize then we need to create a new
+ * subrequest.
+ */
+int netfs_advance_writethrough(struct netfs_io_request *wreq, struct writeback_control *wbc,
+			       struct folio *folio, size_t copied, bool to_page_end,
+			       struct folio **writethrough_cache)
+{
+	kenter("R=%x ic=%zu ws=%u cp=%zu tp=%u",
+	       wreq->debug_id, wreq->iter.count, wreq->wsize, copied, to_page_end);
+
+	if (!*writethrough_cache) {
+		if (folio_test_dirty(folio))
+			/* Sigh.  mmap. */
+			folio_clear_dirty_for_io(folio);
+
+		/* We can make multiple writes to the folio... */
+		folio_start_writeback(folio);
+		if (wreq->len == 0)
+			trace_netfs_folio(folio, netfs_folio_trace_wthru);
+		else
+			trace_netfs_folio(folio, netfs_folio_trace_wthru_plus);
+		*writethrough_cache = folio;
+	}
+
+	wreq->len += copied;
+	if (!to_page_end)
+		return 0;
+
+	*writethrough_cache = NULL;
+	return netfs_write_folio(wreq, wbc, folio);
+}
+
+/*
+ * End a write operation used when writing through the pagecache.
+ */
+int netfs_end_writethrough(struct netfs_io_request *wreq, struct writeback_control *wbc,
+			   struct folio *writethrough_cache)
+{
+	struct netfs_inode *ictx = netfs_inode(wreq->inode);
+	int ret;
+
+	kenter("R=%x", wreq->debug_id);
+
+	if (writethrough_cache)
+		netfs_write_folio(wreq, wbc, writethrough_cache);
+
+	netfs_issue_write(wreq, &wreq->io_streams[0]);
+	netfs_issue_write(wreq, &wreq->io_streams[1]);
+	smp_wmb(); /* Write lists before ALL_QUEUED. */
+	set_bit(NETFS_RREQ_ALL_QUEUED, &wreq->flags);
+
+	mutex_unlock(&ictx->wb_lock);
+
+	if (wreq->iocb) {
+		ret = -EIOCBQUEUED;
+	} else {
+		wait_on_bit(&wreq->flags, NETFS_RREQ_IN_PROGRESS, TASK_UNINTERRUPTIBLE);
+		ret = wreq->error;
+	}
+	netfs_put_request(wreq, false, netfs_rreq_trace_put_return);
+	return ret;
+}
+
+/*
+ * Write data to the server without going through the pagecache and without
+ * writing it to the local cache.
+ */
+int netfs_unbuffered_write(struct netfs_io_request *wreq, bool may_wait, size_t len)
+{
+	struct netfs_io_stream *upload = &wreq->io_streams[0];
+	ssize_t part;
+	loff_t start = wreq->start;
+	int error = 0;
+
+	kenter("%zx", len);
+
+	if (wreq->origin == NETFS_DIO_WRITE)
+		inode_dio_begin(wreq->inode);
+
+	while (len) {
+		// TODO: Prepare content encryption
+
+		kdebug("unbuffered %zx", len);
+		part = netfs_advance_write(wreq, upload, start, len, false);
+		start += part;
+		len -= part;
+		if (test_bit(NETFS_RREQ_PAUSE, &wreq->flags)) {
+			trace_netfs_rreq(wreq, netfs_rreq_trace_wait_pause);
+			wait_on_bit(&wreq->flags, NETFS_RREQ_PAUSE, TASK_UNINTERRUPTIBLE);
+		}
+		if (test_bit(NETFS_RREQ_FAILED, &wreq->flags))
+			break;
+	}
+
+	netfs_issue_write(wreq, upload);
+
+	smp_wmb(); /* Write lists before ALL_QUEUED. */
+	set_bit(NETFS_RREQ_ALL_QUEUED, &wreq->flags);
+	if (list_empty(&upload->subrequests))
+		netfs_wake_write_collector(wreq, false);
+
+	kleave(" = %d", error);
+	return error;
+}
diff --git a/fs/nfs/Kconfig b/fs/nfs/Kconfig
index f7e32d76e3..57249f040d 100644
--- a/fs/nfs/Kconfig
+++ b/fs/nfs/Kconfig
@@ -33,12 +33,12 @@ config NFS_FS
 config NFS_V2
 	tristate "NFS client support for NFS version 2"
 	depends on NFS_FS
-	default y
+	default n
 	help
 	  This option enables support for version 2 of the NFS protocol
 	  (RFC 1094) in the kernel's NFS client.
 
-	  If unsure, say Y.
+	  If unsure, say N.
 
 config NFS_V3
 	tristate "NFS client support for NFS version 3"
diff --git a/fs/nfs/blocklayout/blocklayout.h b/fs/nfs/blocklayout/blocklayout.h
index b4294a8aa2..f1eeb49141 100644
--- a/fs/nfs/blocklayout/blocklayout.h
+++ b/fs/nfs/blocklayout/blocklayout.h
@@ -108,7 +108,7 @@ struct pnfs_block_dev {
 	struct pnfs_block_dev		*children;
 	u64				chunk_size;
 
-	struct bdev_handle		*bdev_handle;
+	struct file			*bdev_file;
 	u64				disk_offset;
 
 	u64				pr_key;
diff --git a/fs/nfs/blocklayout/dev.c b/fs/nfs/blocklayout/dev.c
index c97ebc42ec..93ef7f8649 100644
--- a/fs/nfs/blocklayout/dev.c
+++ b/fs/nfs/blocklayout/dev.c
@@ -25,17 +25,17 @@ bl_free_device(struct pnfs_block_dev *dev)
 	} else {
 		if (dev->pr_registered) {
 			const struct pr_ops *ops =
-				dev->bdev_handle->bdev->bd_disk->fops->pr_ops;
+				file_bdev(dev->bdev_file)->bd_disk->fops->pr_ops;
 			int error;
 
-			error = ops->pr_register(dev->bdev_handle->bdev,
+			error = ops->pr_register(file_bdev(dev->bdev_file),
 				dev->pr_key, 0, false);
 			if (error)
 				pr_err("failed to unregister PR key.\n");
 		}
 
-		if (dev->bdev_handle)
-			bdev_release(dev->bdev_handle);
+		if (dev->bdev_file)
+			fput(dev->bdev_file);
 	}
 }
 
@@ -169,7 +169,7 @@ static bool bl_map_simple(struct pnfs_block_dev *dev, u64 offset,
 	map->start = dev->start;
 	map->len = dev->len;
 	map->disk_offset = dev->disk_offset;
-	map->bdev = dev->bdev_handle->bdev;
+	map->bdev = file_bdev(dev->bdev_file);
 	return true;
 }
 
@@ -236,26 +236,26 @@ bl_parse_simple(struct nfs_server *server, struct pnfs_block_dev *d,
 		struct pnfs_block_volume *volumes, int idx, gfp_t gfp_mask)
 {
 	struct pnfs_block_volume *v = &volumes[idx];
-	struct bdev_handle *bdev_handle;
+	struct file *bdev_file;
 	dev_t dev;
 
 	dev = bl_resolve_deviceid(server, v, gfp_mask);
 	if (!dev)
 		return -EIO;
 
-	bdev_handle = bdev_open_by_dev(dev, BLK_OPEN_READ | BLK_OPEN_WRITE,
+	bdev_file = bdev_file_open_by_dev(dev, BLK_OPEN_READ | BLK_OPEN_WRITE,
 				       NULL, NULL);
-	if (IS_ERR(bdev_handle)) {
+	if (IS_ERR(bdev_file)) {
 		printk(KERN_WARNING "pNFS: failed to open device %d:%d (%ld)\n",
-			MAJOR(dev), MINOR(dev), PTR_ERR(bdev_handle));
-		return PTR_ERR(bdev_handle);
+			MAJOR(dev), MINOR(dev), PTR_ERR(bdev_file));
+		return PTR_ERR(bdev_file);
 	}
-	d->bdev_handle = bdev_handle;
-	d->len = bdev_nr_bytes(bdev_handle->bdev);
+	d->bdev_file = bdev_file;
+	d->len = bdev_nr_bytes(file_bdev(bdev_file));
 	d->map = bl_map_simple;
 
 	printk(KERN_INFO "pNFS: using block device %s\n",
-		bdev_handle->bdev->bd_disk->disk_name);
+		file_bdev(bdev_file)->bd_disk->disk_name);
 	return 0;
 }
 
@@ -300,10 +300,10 @@ bl_validate_designator(struct pnfs_block_volume *v)
 	}
 }
 
-static struct bdev_handle *
+static struct file *
 bl_open_path(struct pnfs_block_volume *v, const char *prefix)
 {
-	struct bdev_handle *bdev_handle;
+	struct file *bdev_file;
 	const char *devname;
 
 	devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/%s%*phN",
@@ -311,15 +311,15 @@ bl_open_path(struct pnfs_block_volume *v, const char *prefix)
 	if (!devname)
 		return ERR_PTR(-ENOMEM);
 
-	bdev_handle = bdev_open_by_path(devname, BLK_OPEN_READ | BLK_OPEN_WRITE,
+	bdev_file = bdev_file_open_by_path(devname, BLK_OPEN_READ | BLK_OPEN_WRITE,
 					NULL, NULL);
-	if (IS_ERR(bdev_handle)) {
+	if (IS_ERR(bdev_file)) {
 		pr_warn("pNFS: failed to open device %s (%ld)\n",
-			devname, PTR_ERR(bdev_handle));
+			devname, PTR_ERR(bdev_file));
 	}
 
 	kfree(devname);
-	return bdev_handle;
+	return bdev_file;
 }
 
 static int
@@ -327,7 +327,7 @@ bl_parse_scsi(struct nfs_server *server, struct pnfs_block_dev *d,
 		struct pnfs_block_volume *volumes, int idx, gfp_t gfp_mask)
 {
 	struct pnfs_block_volume *v = &volumes[idx];
-	struct bdev_handle *bdev_handle;
+	struct file *bdev_file;
 	const struct pr_ops *ops;
 	int error;
 
@@ -340,14 +340,14 @@ bl_parse_scsi(struct nfs_server *server, struct pnfs_block_dev *d,
 	 * On other distributions like Debian, the default SCSI by-id path will
 	 * point to the dm-multipath device if one exists.
 	 */
-	bdev_handle = bl_open_path(v, "dm-uuid-mpath-0x");
-	if (IS_ERR(bdev_handle))
-		bdev_handle = bl_open_path(v, "wwn-0x");
-	if (IS_ERR(bdev_handle))
-		return PTR_ERR(bdev_handle);
-	d->bdev_handle = bdev_handle;
-
-	d->len = bdev_nr_bytes(d->bdev_handle->bdev);
+	bdev_file = bl_open_path(v, "dm-uuid-mpath-0x");
+	if (IS_ERR(bdev_file))
+		bdev_file = bl_open_path(v, "wwn-0x");
+	if (IS_ERR(bdev_file))
+		return PTR_ERR(bdev_file);
+	d->bdev_file = bdev_file;
+
+	d->len = bdev_nr_bytes(file_bdev(d->bdev_file));
 	d->map = bl_map_simple;
 	d->pr_key = v->scsi.pr_key;
 
@@ -355,20 +355,20 @@ bl_parse_scsi(struct nfs_server *server, struct pnfs_block_dev *d,
 		return -ENODEV;
 
 	pr_info("pNFS: using block device %s (reservation key 0x%llx)\n",
-		d->bdev_handle->bdev->bd_disk->disk_name, d->pr_key);
+		file_bdev(d->bdev_file)->bd_disk->disk_name, d->pr_key);
 
-	ops = d->bdev_handle->bdev->bd_disk->fops->pr_ops;
+	ops = file_bdev(d->bdev_file)->bd_disk->fops->pr_ops;
 	if (!ops) {
 		pr_err("pNFS: block device %s does not support reservations.",
-				d->bdev_handle->bdev->bd_disk->disk_name);
+				file_bdev(d->bdev_file)->bd_disk->disk_name);
 		error = -EINVAL;
 		goto out_blkdev_put;
 	}
 
-	error = ops->pr_register(d->bdev_handle->bdev, 0, d->pr_key, true);
+	error = ops->pr_register(file_bdev(d->bdev_file), 0, d->pr_key, true);
 	if (error) {
 		pr_err("pNFS: failed to register key for block device %s.",
-				d->bdev_handle->bdev->bd_disk->disk_name);
+				file_bdev(d->bdev_file)->bd_disk->disk_name);
 		goto out_blkdev_put;
 	}
 
@@ -376,7 +376,7 @@ bl_parse_scsi(struct nfs_server *server, struct pnfs_block_dev *d,
 	return 0;
 
 out_blkdev_put:
-	bdev_release(d->bdev_handle);
+	fput(d->bdev_file);
 	return error;
 }
 
diff --git a/fs/nfs/callback.c b/fs/nfs/callback.c
index 760d27dd72..8adfcd4c8c 100644
--- a/fs/nfs/callback.c
+++ b/fs/nfs/callback.c
@@ -356,15 +356,12 @@ static const struct svc_version *nfs4_callback_version[] = {
 	[4] = &nfs4_callback_version4,
 };
 
-static struct svc_stat nfs4_callback_stats;
-
 static struct svc_program nfs4_callback_program = {
 	.pg_prog = NFS4_CALLBACK,			/* RPC service number */
 	.pg_nvers = ARRAY_SIZE(nfs4_callback_version),	/* Number of entries */
 	.pg_vers = nfs4_callback_version,		/* version table */
 	.pg_name = "NFSv4 callback",			/* service name */
 	.pg_class = "nfs",				/* authentication class */
-	.pg_stats = &nfs4_callback_stats,
 	.pg_authenticate = nfs_callback_authenticate,
 	.pg_init_request = svc_generic_init_request,
 	.pg_rpcbind_set	= svc_generic_rpcbind_set,
diff --git a/fs/nfs/client.c b/fs/nfs/client.c
index a8fad331df..de77848ae6 100644
--- a/fs/nfs/client.c
+++ b/fs/nfs/client.c
@@ -76,10 +76,6 @@ const struct rpc_program nfs_program = {
 	.pipe_dir_name		= NFS_PIPE_DIRNAME,
 };
 
-struct rpc_stat nfs_rpcstat = {
-	.program		= &nfs_program
-};
-
 static struct nfs_subversion *find_nfs_version(unsigned int version)
 {
 	struct nfs_subversion *nfs;
diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c
index fa1a14def4..6bace5fece 100644
--- a/fs/nfs/delegation.c
+++ b/fs/nfs/delegation.c
@@ -156,8 +156,8 @@ static int nfs_delegation_claim_locks(struct nfs4_state *state, const nfs4_state
 	list = &flctx->flc_posix;
 	spin_lock(&flctx->flc_lock);
 restart:
-	list_for_each_entry(fl, list, fl_list) {
-		if (nfs_file_open_context(fl->fl_file)->state != state)
+	for_each_file_lock(fl, list) {
+		if (nfs_file_open_context(fl->c.flc_file)->state != state)
 			continue;
 		spin_unlock(&flctx->flc_lock);
 		status = nfs4_lock_delegation_recall(fl, state, stateid);
@@ -181,7 +181,6 @@ static int nfs_delegation_claim_opens(struct inode *inode,
 	struct nfs_open_context *ctx;
 	struct nfs4_state_owner *sp;
 	struct nfs4_state *state;
-	unsigned int seq;
 	int err;
 
 again:
@@ -202,12 +201,9 @@ again:
 		sp = state->owner;
 		/* Block nfs4_proc_unlck */
 		mutex_lock(&sp->so_delegreturn_mutex);
-		seq = raw_seqcount_begin(&sp->so_reclaim_seqcount);
 		err = nfs4_open_delegation_recall(ctx, state, stateid);
 		if (!err)
 			err = nfs_delegation_claim_locks(state, stateid);
-		if (!err && read_seqcount_retry(&sp->so_reclaim_seqcount, seq))
-			err = -EAGAIN;
 		mutex_unlock(&sp->so_delegreturn_mutex);
 		put_nfs_open_context(ctx);
 		if (err != 0)
diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index ac505671ef..07a7be2718 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -56,6 +56,8 @@ static int nfs_readdir(struct file *, struct dir_context *);
 static int nfs_fsync_dir(struct file *, loff_t, loff_t, int);
 static loff_t nfs_llseek_dir(struct file *, loff_t, int);
 static void nfs_readdir_clear_array(struct folio *);
+static int nfs_do_create(struct inode *dir, struct dentry *dentry,
+			 umode_t mode, int open_flags);
 
 const struct file_operations nfs_dir_operations = {
 	.llseek		= nfs_llseek_dir,
@@ -1625,7 +1627,16 @@ nfs_lookup_revalidate_done(struct inode *dir, struct dentry *dentry,
 	switch (error) {
 	case 1:
 		break;
-	case 0:
+	case -ETIMEDOUT:
+		if (inode && (IS_ROOT(dentry) ||
+			      NFS_SERVER(inode)->flags & NFS_MOUNT_SOFTREVAL))
+			error = 1;
+		break;
+	case -ESTALE:
+	case -ENOENT:
+		error = 0;
+		fallthrough;
+	default:
 		/*
 		 * We can't d_drop the root of a disconnected tree:
 		 * its d_hash is on the s_anon list and d_drop() would hide
@@ -1680,18 +1691,8 @@ static int nfs_lookup_revalidate_dentry(struct inode *dir,
 
 	dir_verifier = nfs_save_change_attribute(dir);
 	ret = NFS_PROTO(dir)->lookup(dir, dentry, fhandle, fattr);
-	if (ret < 0) {
-		switch (ret) {
-		case -ESTALE:
-		case -ENOENT:
-			ret = 0;
-			break;
-		case -ETIMEDOUT:
-			if (NFS_SERVER(inode)->flags & NFS_MOUNT_SOFTREVAL)
-				ret = 1;
-		}
+	if (ret < 0)
 		goto out;
-	}
 
 	/* Request help from readdirplus */
 	nfs_lookup_advise_force_readdirplus(dir, flags);
@@ -1735,7 +1736,7 @@ nfs_do_lookup_revalidate(struct inode *dir, struct dentry *dentry,
 			 unsigned int flags)
 {
 	struct inode *inode;
-	int error;
+	int error = 0;
 
 	nfs_inc_stats(dir, NFSIOS_DENTRYREVALIDATE);
 	inode = d_inode(dentry);
@@ -1780,7 +1781,7 @@ out_valid:
 out_bad:
 	if (flags & LOOKUP_RCU)
 		return -ECHILD;
-	return nfs_lookup_revalidate_done(dir, dentry, inode, 0);
+	return nfs_lookup_revalidate_done(dir, dentry, inode, error);
 }
 
 static int
@@ -1802,9 +1803,10 @@ __nfs_lookup_revalidate(struct dentry *dentry, unsigned int flags,
 		if (parent != READ_ONCE(dentry->d_parent))
 			return -ECHILD;
 	} else {
-		/* Wait for unlink to complete */
+		/* Wait for unlink to complete - see unblock_revalidate() */
 		wait_var_event(&dentry->d_fsdata,
-			       dentry->d_fsdata != NFS_FSDATA_BLOCKED);
+			       smp_load_acquire(&dentry->d_fsdata)
+			       != NFS_FSDATA_BLOCKED);
 		parent = dget_parent(dentry);
 		ret = reval(d_inode(parent), dentry, flags);
 		dput(parent);
@@ -1817,6 +1819,29 @@ static int nfs_lookup_revalidate(struct dentry *dentry, unsigned int flags)
 	return __nfs_lookup_revalidate(dentry, flags, nfs_do_lookup_revalidate);
 }
 
+static void block_revalidate(struct dentry *dentry)
+{
+	/* old devname - just in case */
+	kfree(dentry->d_fsdata);
+
+	/* Any new reference that could lead to an open
+	 * will take ->d_lock in lookup_open() -> d_lookup().
+	 * Holding this lock ensures we cannot race with
+	 * __nfs_lookup_revalidate() and removes and need
+	 * for further barriers.
+	 */
+	lockdep_assert_held(&dentry->d_lock);
+
+	dentry->d_fsdata = NFS_FSDATA_BLOCKED;
+}
+
+static void unblock_revalidate(struct dentry *dentry)
+{
+	/* store_release ensures wait_var_event() sees the update */
+	smp_store_release(&dentry->d_fsdata, NULL);
+	wake_up_var(&dentry->d_fsdata);
+}
+
 /*
  * A weaker form of d_revalidate for revalidating just the d_inode(dentry)
  * when we don't really care about the dentry name. This is called when a
@@ -2243,6 +2268,44 @@ static int nfs4_lookup_revalidate(struct dentry *dentry, unsigned int flags)
 
 #endif /* CONFIG_NFSV4 */
 
+int nfs_atomic_open_v23(struct inode *dir, struct dentry *dentry,
+			struct file *file, unsigned int open_flags,
+			umode_t mode)
+{
+
+	/* Same as look+open from lookup_open(), but with different O_TRUNC
+	 * handling.
+	 */
+	int error = 0;
+
+	if (dentry->d_name.len > NFS_SERVER(dir)->namelen)
+		return -ENAMETOOLONG;
+
+	if (open_flags & O_CREAT) {
+		file->f_mode |= FMODE_CREATED;
+		error = nfs_do_create(dir, dentry, mode, open_flags);
+		if (error)
+			return error;
+		return finish_open(file, dentry, NULL);
+	} else if (d_in_lookup(dentry)) {
+		/* The only flags nfs_lookup considers are
+		 * LOOKUP_EXCL and LOOKUP_RENAME_TARGET, and
+		 * we want those to be zero so the lookup isn't skipped.
+		 */
+		struct dentry *res = nfs_lookup(dir, dentry, 0);
+
+		d_lookup_done(dentry);
+		if (unlikely(res)) {
+			if (IS_ERR(res))
+				return PTR_ERR(res);
+			return finish_no_open(file, res);
+		}
+	}
+	return finish_no_open(file, NULL);
+
+}
+EXPORT_SYMBOL_GPL(nfs_atomic_open_v23);
+
 struct dentry *
 nfs_add_or_obtain(struct dentry *dentry, struct nfs_fh *fhandle,
 				struct nfs_fattr *fattr)
@@ -2303,18 +2366,23 @@ EXPORT_SYMBOL_GPL(nfs_instantiate);
  * that the operation succeeded on the server, but an error in the
  * reply path made it appear to have failed.
  */
-int nfs_create(struct mnt_idmap *idmap, struct inode *dir,
-	       struct dentry *dentry, umode_t mode, bool excl)
+static int nfs_do_create(struct inode *dir, struct dentry *dentry,
+			 umode_t mode, int open_flags)
 {
 	struct iattr attr;
-	int open_flags = excl ? O_CREAT | O_EXCL : O_CREAT;
 	int error;
 
+	open_flags |= O_CREAT;
+
 	dfprintk(VFS, "NFS: create(%s/%lu), %pd\n",
 			dir->i_sb->s_id, dir->i_ino, dentry);
 
 	attr.ia_mode = mode;
 	attr.ia_valid = ATTR_MODE;
+	if (open_flags & O_TRUNC) {
+		attr.ia_size = 0;
+		attr.ia_valid |= ATTR_SIZE;
+	}
 
 	trace_nfs_create_enter(dir, dentry, open_flags);
 	error = NFS_PROTO(dir)->create(dir, dentry, &attr, open_flags);
@@ -2326,6 +2394,12 @@ out_err:
 	d_drop(dentry);
 	return error;
 }
+
+int nfs_create(struct mnt_idmap *idmap, struct inode *dir,
+	       struct dentry *dentry, umode_t mode, bool excl)
+{
+	return nfs_do_create(dir, dentry, mode, excl ? O_EXCL : 0);
+}
 EXPORT_SYMBOL_GPL(nfs_create);
 
 /*
@@ -2501,15 +2575,12 @@ int nfs_unlink(struct inode *dir, struct dentry *dentry)
 		spin_unlock(&dentry->d_lock);
 		goto out;
 	}
-	/* old devname */
-	kfree(dentry->d_fsdata);
-	dentry->d_fsdata = NFS_FSDATA_BLOCKED;
+	block_revalidate(dentry);
 
 	spin_unlock(&dentry->d_lock);
 	error = nfs_safe_remove(dentry);
 	nfs_dentry_remove_handle_error(dir, dentry, error);
-	dentry->d_fsdata = NULL;
-	wake_up_var(&dentry->d_fsdata);
+	unblock_revalidate(dentry);
 out:
 	trace_nfs_unlink_exit(dir, dentry, error);
 	return error;
@@ -2616,8 +2687,7 @@ nfs_unblock_rename(struct rpc_task *task, struct nfs_renamedata *data)
 {
 	struct dentry *new_dentry = data->new_dentry;
 
-	new_dentry->d_fsdata = NULL;
-	wake_up_var(&new_dentry->d_fsdata);
+	unblock_revalidate(new_dentry);
 }
 
 /*
@@ -2679,11 +2749,6 @@ int nfs_rename(struct mnt_idmap *idmap, struct inode *old_dir,
 		if (WARN_ON(new_dentry->d_flags & DCACHE_NFSFS_RENAMED) ||
 		    WARN_ON(new_dentry->d_fsdata == NFS_FSDATA_BLOCKED))
 			goto out;
-		if (new_dentry->d_fsdata) {
-			/* old devname */
-			kfree(new_dentry->d_fsdata);
-			new_dentry->d_fsdata = NULL;
-		}
 
 		spin_lock(&new_dentry->d_lock);
 		if (d_count(new_dentry) > 2) {
@@ -2705,7 +2770,7 @@ int nfs_rename(struct mnt_idmap *idmap, struct inode *old_dir,
 			new_dentry = dentry;
 			new_inode = NULL;
 		} else {
-			new_dentry->d_fsdata = NFS_FSDATA_BLOCKED;
+			block_revalidate(new_dentry);
 			must_unblock = true;
 			spin_unlock(&new_dentry->d_lock);
 		}
@@ -2717,6 +2782,8 @@ int nfs_rename(struct mnt_idmap *idmap, struct inode *old_dir,
 	task = nfs_async_rename(old_dir, new_dir, old_dentry, new_dentry,
 				must_unblock ? nfs_unblock_rename : NULL);
 	if (IS_ERR(task)) {
+		if (must_unblock)
+			unblock_revalidate(new_dentry);
 		error = PTR_ERR(task);
 		goto out;
 	}
diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c
index d8cd4f69e3..90079ca134 100644
--- a/fs/nfs/direct.c
+++ b/fs/nfs/direct.c
@@ -141,8 +141,6 @@ int nfs_swap_rw(struct kiocb *iocb, struct iov_iter *iter)
 {
 	ssize_t ret;
 
-	VM_BUG_ON(iov_iter_count(iter) != PAGE_SIZE);
-
 	if (iov_iter_rw(iter) == READ)
 		ret = nfs_file_direct_read(iocb, iter, true);
 	else
@@ -606,6 +604,7 @@ static void nfs_direct_commit_complete(struct nfs_commit_data *data)
 
 	trace_nfs_direct_commit_complete(dreq);
 
+	spin_lock(&dreq->lock);
 	if (status < 0) {
 		/* Errors in commit are fatal */
 		dreq->error = status;
@@ -613,6 +612,7 @@ static void nfs_direct_commit_complete(struct nfs_commit_data *data)
 	} else {
 		status = dreq->error;
 	}
+	spin_unlock(&dreq->lock);
 
 	nfs_init_cinfo_from_dreq(&cinfo, dreq);
 
@@ -625,7 +625,10 @@ static void nfs_direct_commit_complete(struct nfs_commit_data *data)
 			spin_unlock(&dreq->lock);
 			nfs_release_request(req);
 		} else if (!nfs_write_match_verf(verf, req)) {
-			dreq->flags = NFS_ODIRECT_RESCHED_WRITES;
+			spin_lock(&dreq->lock);
+			if (dreq->flags == 0)
+				dreq->flags = NFS_ODIRECT_RESCHED_WRITES;
+			spin_unlock(&dreq->lock);
 			/*
 			 * Despite the reboot, the write was successful,
 			 * so reset wb_nio.
@@ -1044,8 +1047,7 @@ int __init nfs_init_directcache(void)
 {
 	nfs_direct_cachep = kmem_cache_create("nfs_direct_cache",
 						sizeof(struct nfs_direct_req),
-						0, (SLAB_RECLAIM_ACCOUNT|
-							SLAB_MEM_SPREAD),
+						0, SLAB_RECLAIM_ACCOUNT,
 						NULL);
 	if (nfs_direct_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/nfs/file.c b/fs/nfs/file.c
index 8577ccf621..445db17f1b 100644
--- a/fs/nfs/file.c
+++ b/fs/nfs/file.c
@@ -433,8 +433,8 @@ static void nfs_invalidate_folio(struct folio *folio, size_t offset,
 		return;
 	/* Cancel any unstarted writes on this page */
 	nfs_wb_folio_cancel(inode, folio);
-	folio_wait_fscache(folio);
-	trace_nfs_invalidate_folio(inode, folio);
+	folio_wait_private_2(folio); /* [DEPRECATED] */
+	trace_nfs_invalidate_folio(inode, folio_pos(folio) + offset, length);
 }
 
 /*
@@ -500,9 +500,10 @@ static int nfs_launder_folio(struct folio *folio)
 	dfprintk(PAGECACHE, "NFS: launder_folio(%ld, %llu)\n",
 		inode->i_ino, folio_pos(folio));
 
-	folio_wait_fscache(folio);
+	folio_wait_private_2(folio); /* [DEPRECATED] */
 	ret = nfs_wb_folio(inode, folio);
-	trace_nfs_launder_folio_done(inode, folio, ret);
+	trace_nfs_launder_folio_done(inode, folio_pos(folio),
+			folio_size(folio), ret);
 	return ret;
 }
 
@@ -593,8 +594,8 @@ static vm_fault_t nfs_vm_page_mkwrite(struct vm_fault *vmf)
 	sb_start_pagefault(inode->i_sb);
 
 	/* make sure the cache has finished storing the page */
-	if (folio_test_fscache(folio) &&
-	    folio_wait_fscache_killable(folio) < 0) {
+	if (folio_test_private_2(folio) && /* [DEPRECATED] */
+	    folio_wait_private_2_killable(folio) < 0) {
 		ret = VM_FAULT_RETRY;
 		goto out;
 	}
@@ -720,15 +721,15 @@ do_getlk(struct file *filp, int cmd, struct file_lock *fl, int is_local)
 {
 	struct inode *inode = filp->f_mapping->host;
 	int status = 0;
-	unsigned int saved_type = fl->fl_type;
+	unsigned int saved_type = fl->c.flc_type;
 
 	/* Try local locking first */
 	posix_test_lock(filp, fl);
-	if (fl->fl_type != F_UNLCK) {
+	if (fl->c.flc_type != F_UNLCK) {
 		/* found a conflict */
 		goto out;
 	}
-	fl->fl_type = saved_type;
+	fl->c.flc_type = saved_type;
 
 	if (NFS_PROTO(inode)->have_delegation(inode, FMODE_READ))
 		goto out_noconflict;
@@ -740,7 +741,7 @@ do_getlk(struct file *filp, int cmd, struct file_lock *fl, int is_local)
 out:
 	return status;
 out_noconflict:
-	fl->fl_type = F_UNLCK;
+	fl->c.flc_type = F_UNLCK;
 	goto out;
 }
 
@@ -765,7 +766,7 @@ do_unlk(struct file *filp, int cmd, struct file_lock *fl, int is_local)
 		 * 	If we're signalled while cleaning up locks on process exit, we
 		 * 	still need to complete the unlock.
 		 */
-		if (status < 0 && !(fl->fl_flags & FL_CLOSE))
+		if (status < 0 && !(fl->c.flc_flags & FL_CLOSE))
 			return status;
 	}
 
@@ -832,12 +833,12 @@ int nfs_lock(struct file *filp, int cmd, struct file_lock *fl)
 	int is_local = 0;
 
 	dprintk("NFS: lock(%pD2, t=%x, fl=%x, r=%lld:%lld)\n",
-			filp, fl->fl_type, fl->fl_flags,
+			filp, fl->c.flc_type, fl->c.flc_flags,
 			(long long)fl->fl_start, (long long)fl->fl_end);
 
 	nfs_inc_stats(inode, NFSIOS_VFSLOCK);
 
-	if (fl->fl_flags & FL_RECLAIM)
+	if (fl->c.flc_flags & FL_RECLAIM)
 		return -ENOGRACE;
 
 	if (NFS_SERVER(inode)->flags & NFS_MOUNT_LOCAL_FCNTL)
@@ -851,7 +852,7 @@ int nfs_lock(struct file *filp, int cmd, struct file_lock *fl)
 
 	if (IS_GETLK(cmd))
 		ret = do_getlk(filp, cmd, fl, is_local);
-	else if (fl->fl_type == F_UNLCK)
+	else if (lock_is_unlock(fl))
 		ret = do_unlk(filp, cmd, fl, is_local);
 	else
 		ret = do_setlk(filp, cmd, fl, is_local);
@@ -869,16 +870,16 @@ int nfs_flock(struct file *filp, int cmd, struct file_lock *fl)
 	int is_local = 0;
 
 	dprintk("NFS: flock(%pD2, t=%x, fl=%x)\n",
-			filp, fl->fl_type, fl->fl_flags);
+			filp, fl->c.flc_type, fl->c.flc_flags);
 
-	if (!(fl->fl_flags & FL_FLOCK))
+	if (!(fl->c.flc_flags & FL_FLOCK))
 		return -ENOLCK;
 
 	if (NFS_SERVER(inode)->flags & NFS_MOUNT_LOCAL_FLOCK)
 		is_local = 1;
 
 	/* We're simulating flock() locks using posix locks on the server */
-	if (fl->fl_type == F_UNLCK)
+	if (lock_is_unlock(fl))
 		return do_unlk(filp, cmd, fl, is_local);
 	return do_setlk(filp, cmd, fl, is_local);
 }
diff --git a/fs/nfs/filelayout/filelayout.c b/fs/nfs/filelayout/filelayout.c
index ce8f8934bc..29d84dc66c 100644
--- a/fs/nfs/filelayout/filelayout.c
+++ b/fs/nfs/filelayout/filelayout.c
@@ -605,14 +605,6 @@ filelayout_check_layout(struct pnfs_layout_hdr *lo,
 
 	dprintk("--> %s\n", __func__);
 
-	/* FIXME: remove this check when layout segment support is added */
-	if (lgr->range.offset != 0 ||
-	    lgr->range.length != NFS4_MAX_UINT64) {
-		dprintk("%s Only whole file layouts supported. Use MDS i/o\n",
-			__func__);
-		goto out;
-	}
-
 	if (fl->pattern_offset > lgr->range.offset) {
 		dprintk("%s pattern_offset %lld too large\n",
 				__func__, fl->pattern_offset);
@@ -875,15 +867,15 @@ static void
 filelayout_pg_init_read(struct nfs_pageio_descriptor *pgio,
 			struct nfs_page *req)
 {
-	pnfs_generic_pg_check_layout(pgio);
+	pnfs_generic_pg_check_layout(pgio, req);
 	if (!pgio->pg_lseg) {
 		pgio->pg_lseg = fl_pnfs_update_layout(pgio->pg_inode,
 						      nfs_req_openctx(req),
-						      0,
-						      NFS4_MAX_UINT64,
+						      req_offset(req),
+						      req->wb_bytes,
 						      IOMODE_READ,
 						      false,
-						      GFP_KERNEL);
+						      nfs_io_gfp_mask());
 		if (IS_ERR(pgio->pg_lseg)) {
 			pgio->pg_error = PTR_ERR(pgio->pg_lseg);
 			pgio->pg_lseg = NULL;
@@ -899,15 +891,15 @@ static void
 filelayout_pg_init_write(struct nfs_pageio_descriptor *pgio,
 			 struct nfs_page *req)
 {
-	pnfs_generic_pg_check_layout(pgio);
+	pnfs_generic_pg_check_layout(pgio, req);
 	if (!pgio->pg_lseg) {
 		pgio->pg_lseg = fl_pnfs_update_layout(pgio->pg_inode,
 						      nfs_req_openctx(req),
-						      0,
-						      NFS4_MAX_UINT64,
+						      req_offset(req),
+						      req->wb_bytes,
 						      IOMODE_RW,
 						      false,
-						      GFP_NOFS);
+						      nfs_io_gfp_mask());
 		if (IS_ERR(pgio->pg_lseg)) {
 			pgio->pg_error = PTR_ERR(pgio->pg_lseg);
 			pgio->pg_lseg = NULL;
diff --git a/fs/nfs/filelayout/filelayoutdev.c b/fs/nfs/filelayout/filelayoutdev.c
index acf4b88889..4fa304fa5b 100644
--- a/fs/nfs/filelayout/filelayoutdev.c
+++ b/fs/nfs/filelayout/filelayoutdev.c
@@ -35,6 +35,7 @@
 #include "../internal.h"
 #include "../nfs4session.h"
 #include "filelayout.h"
+#include "../nfs4trace.h"
 
 #define NFSDBG_FACILITY		NFSDBG_PNFS_LD
 
@@ -172,6 +173,7 @@ nfs4_fl_alloc_deviceid_node(struct nfs_server *server, struct pnfs_device *pdev,
 		dsaddr->ds_list[i] = nfs4_pnfs_ds_add(&dsaddrs, gfp_flags);
 		if (!dsaddr->ds_list[i])
 			goto out_err_drain_dsaddrs;
+		trace_fl_getdevinfo(server, &pdev->dev_id, dsaddr->ds_list[i]->ds_remotestr);
 
 		/* If DS was already in cache, free ds addrs */
 		while (!list_empty(&dsaddrs)) {
diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c
index 3e724cb7ef..24188af56d 100644
--- a/fs/nfs/flexfilelayout/flexfilelayout.c
+++ b/fs/nfs/flexfilelayout/flexfilelayout.c
@@ -823,14 +823,6 @@ ff_layout_pg_get_read(struct nfs_pageio_descriptor *pgio,
 }
 
 static void
-ff_layout_pg_check_layout(struct nfs_pageio_descriptor *pgio,
-			  struct nfs_page *req)
-{
-	pnfs_generic_pg_check_layout(pgio);
-	pnfs_generic_pg_check_range(pgio, req);
-}
-
-static void
 ff_layout_pg_init_read(struct nfs_pageio_descriptor *pgio,
 			struct nfs_page *req)
 {
@@ -840,7 +832,7 @@ ff_layout_pg_init_read(struct nfs_pageio_descriptor *pgio,
 	u32 ds_idx;
 
 retry:
-	ff_layout_pg_check_layout(pgio, req);
+	pnfs_generic_pg_check_layout(pgio, req);
 	/* Use full layout for now */
 	if (!pgio->pg_lseg) {
 		ff_layout_pg_get_read(pgio, req, false);
@@ -895,7 +887,7 @@ ff_layout_pg_init_write(struct nfs_pageio_descriptor *pgio,
 	u32 i;
 
 retry:
-	ff_layout_pg_check_layout(pgio, req);
+	pnfs_generic_pg_check_layout(pgio, req);
 	if (!pgio->pg_lseg) {
 		pgio->pg_lseg =
 			pnfs_update_layout(pgio->pg_inode, nfs_req_openctx(req),
diff --git a/fs/nfs/fs_context.c b/fs/nfs/fs_context.c
index 853e8d609b..6c9f3f6645 100644
--- a/fs/nfs/fs_context.c
+++ b/fs/nfs/fs_context.c
@@ -600,9 +600,11 @@ static int nfs_fs_context_parse_param(struct fs_context *fc,
 		break;
 	case Opt_lock:
 		if (result.negated) {
+			ctx->lock_status = NFS_LOCK_NOLOCK;
 			ctx->flags |= NFS_MOUNT_NONLM;
 			ctx->flags |= (NFS_MOUNT_LOCAL_FLOCK | NFS_MOUNT_LOCAL_FCNTL);
 		} else {
+			ctx->lock_status = NFS_LOCK_LOCK;
 			ctx->flags &= ~NFS_MOUNT_NONLM;
 			ctx->flags &= ~(NFS_MOUNT_LOCAL_FLOCK | NFS_MOUNT_LOCAL_FCNTL);
 		}
@@ -652,6 +654,7 @@ static int nfs_fs_context_parse_param(struct fs_context *fc,
 		ctx->fscache_uniq = NULL;
 		break;
 	case Opt_fscache:
+		trace_nfs_mount_assign(param->key, param->string);
 		ctx->options |= NFS_OPTION_FSCACHE;
 		kfree(ctx->fscache_uniq);
 		ctx->fscache_uniq = param->string;
@@ -1111,9 +1114,12 @@ static int nfs23_parse_monolithic(struct fs_context *fc,
 		ctx->acdirmax	= data->acdirmax;
 		ctx->need_mount	= false;
 
-		memcpy(sap, &data->addr, sizeof(data->addr));
-		ctx->nfs_server.addrlen = sizeof(data->addr);
-		ctx->nfs_server.port = ntohs(data->addr.sin_port);
+		if (!is_remount_fc(fc)) {
+			memcpy(sap, &data->addr, sizeof(data->addr));
+			ctx->nfs_server.addrlen = sizeof(data->addr);
+			ctx->nfs_server.port = ntohs(data->addr.sin_port);
+		}
+
 		if (sap->ss_family != AF_INET ||
 		    !nfs_verify_server_address(sap))
 			goto out_no_address;
diff --git a/fs/nfs/fscache.h b/fs/nfs/fscache.h
index e3cb492331..fbed002799 100644
--- a/fs/nfs/fscache.h
+++ b/fs/nfs/fscache.h
@@ -81,6 +81,8 @@ static inline void nfs_netfs_put(struct nfs_netfs_io_data *netfs)
 static inline void nfs_netfs_inode_init(struct nfs_inode *nfsi)
 {
 	netfs_inode_init(&nfsi->netfs, &nfs_netfs_ops, false);
+	/* [DEPRECATED] Use PG_private_2 to mark folio being written to the cache. */
+	__set_bit(NETFS_ICTX_USE_PGPRIV2, &nfsi->netfs.flags);
 }
 extern void nfs_netfs_initiate_read(struct nfs_pgio_header *hdr);
 extern void nfs_netfs_read_completion(struct nfs_pgio_header *hdr);
@@ -101,10 +103,10 @@ extern int nfs_netfs_read_folio(struct file *file, struct folio *folio);
 
 static inline bool nfs_fscache_release_folio(struct folio *folio, gfp_t gfp)
 {
-	if (folio_test_fscache(folio)) {
+	if (folio_test_private_2(folio)) { /* [DEPRECATED] */
 		if (current_is_kswapd() || !(gfp & __GFP_FS))
 			return false;
-		folio_wait_fscache(folio);
+		folio_wait_private_2(folio);
 	}
 	fscache_note_page_release(netfs_i_cookie(netfs_inode(folio->mapping->host)));
 	return true;
diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index 6fe4b47c39..acef52ecb1 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -2372,7 +2372,7 @@ static int __init nfs_init_inodecache(void)
 	nfs_inode_cachep = kmem_cache_create("nfs_inode_cache",
 					     sizeof(struct nfs_inode),
 					     0, (SLAB_RECLAIM_ACCOUNT|
-						SLAB_MEM_SPREAD|SLAB_ACCOUNT),
+						SLAB_ACCOUNT),
 					     init_once);
 	if (nfs_inode_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h
index 06253695fe..9f0f453474 100644
--- a/fs/nfs/internal.h
+++ b/fs/nfs/internal.h
@@ -112,6 +112,7 @@ struct nfs_fs_context {
 	unsigned short		protofamily;
 	unsigned short		mountfamily;
 	bool			has_sec_mnt_opts;
+	int			lock_status;
 
 	struct {
 		union {
@@ -153,6 +154,12 @@ struct nfs_fs_context {
 	} clone_data;
 };
 
+enum nfs_lock_status {
+	NFS_LOCK_NOT_SET	= 0,
+	NFS_LOCK_LOCK		= 1,
+	NFS_LOCK_NOLOCK		= 2,
+};
+
 #define nfs_errorf(fc, fmt, ...) ((fc)->log.log ?		\
 	errorf(fc, fmt, ## __VA_ARGS__) :			\
 	({ dprintk(fmt "\n", ## __VA_ARGS__); }))
@@ -710,9 +717,9 @@ unsigned long nfs_block_bits(unsigned long bsize, unsigned char *nrbitsp)
 	if ((bsize & (bsize - 1)) || nrbitsp) {
 		unsigned char	nrbits;
 
-		for (nrbits = 31; nrbits && !(bsize & (1 << nrbits)); nrbits--)
+		for (nrbits = 31; nrbits && !(bsize & (1UL << nrbits)); nrbits--)
 			;
-		bsize = 1 << nrbits;
+		bsize = 1UL << nrbits;
 		if (nrbitsp)
 			*nrbitsp = nrbits;
 	}
diff --git a/fs/nfs/iostat.h b/fs/nfs/iostat.h
index 5aa776b5a3..b17a9eb9b1 100644
--- a/fs/nfs/iostat.h
+++ b/fs/nfs/iostat.h
@@ -46,10 +46,7 @@ static inline void nfs_add_stats(const struct inode *inode,
 	nfs_add_server_stats(NFS_SERVER(inode), stat, addend);
 }
 
-static inline struct nfs_iostats __percpu *nfs_alloc_iostats(void)
-{
-	return alloc_percpu(struct nfs_iostats);
-}
+#define nfs_alloc_iostats()	alloc_percpu(struct nfs_iostats)
 
 static inline void nfs_free_iostats(struct nfs_iostats __percpu *stats)
 {
diff --git a/fs/nfs/nfs3client.c b/fs/nfs/nfs3client.c
index 674c012868..b0c8a39c2b 100644
--- a/fs/nfs/nfs3client.c
+++ b/fs/nfs/nfs3client.c
@@ -111,6 +111,7 @@ struct nfs_client *nfs3_set_ds_client(struct nfs_server *mds_srv,
 	cl_init.hostname = buf;
 
 	switch (ds_proto) {
+	case XPRT_TRANSPORT_RDMA:
 	case XPRT_TRANSPORT_TCP:
 	case XPRT_TRANSPORT_TCP_TLS:
 		if (mds_clp->cl_nconnect > 1)
diff --git a/fs/nfs/nfs3proc.c b/fs/nfs/nfs3proc.c
index 2de66e4e82..74bda639a7 100644
--- a/fs/nfs/nfs3proc.c
+++ b/fs/nfs/nfs3proc.c
@@ -963,7 +963,7 @@ nfs3_proc_lock(struct file *filp, int cmd, struct file_lock *fl)
 	struct nfs_open_context *ctx = nfs_file_open_context(filp);
 	int status;
 
-	if (fl->fl_flags & FL_CLOSE) {
+	if (fl->c.flc_flags & FL_CLOSE) {
 		l_ctx = nfs_get_lock_context(ctx);
 		if (IS_ERR(l_ctx))
 			l_ctx = NULL;
@@ -986,6 +986,7 @@ static int nfs3_have_delegation(struct inode *inode, fmode_t flags)
 
 static const struct inode_operations nfs3_dir_inode_operations = {
 	.create		= nfs_create,
+	.atomic_open	= nfs_atomic_open_v23,
 	.lookup		= nfs_lookup,
 	.link		= nfs_link,
 	.unlink		= nfs_unlink,
diff --git a/fs/nfs/nfs42xattr.c b/fs/nfs/nfs42xattr.c
index 49aaf28a69..b6e3d8f77b 100644
--- a/fs/nfs/nfs42xattr.c
+++ b/fs/nfs/nfs42xattr.c
@@ -1017,7 +1017,7 @@ int __init nfs4_xattr_cache_init(void)
 
 	nfs4_xattr_cache_cachep = kmem_cache_create("nfs4_xattr_cache_cache",
 	    sizeof(struct nfs4_xattr_cache), 0,
-	    (SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD),
+	    (SLAB_RECLAIM_ACCOUNT),
 	    nfs4_xattr_cache_init_once);
 	if (nfs4_xattr_cache_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h
index 581698f1b7..7024230f0d 100644
--- a/fs/nfs/nfs4_fs.h
+++ b/fs/nfs/nfs4_fs.h
@@ -120,7 +120,6 @@ struct nfs4_state_owner {
 	unsigned long	     so_flags;
 	struct list_head     so_states;
 	struct nfs_seqid_counter so_seqid;
-	seqcount_spinlock_t  so_reclaim_seqcount;
 	struct mutex	     so_delegreturn_mutex;
 };
 
@@ -330,7 +329,7 @@ extern int update_open_stateid(struct nfs4_state *state,
 				const nfs4_stateid *deleg_stateid,
 				fmode_t fmode);
 extern int nfs4_proc_setlease(struct file *file, int arg,
-			      struct file_lock **lease, void **priv);
+			      struct file_lease **lease, void **priv);
 extern int nfs4_proc_get_lease_time(struct nfs_client *clp,
 		struct nfs_fsinfo *fsinfo);
 extern void nfs4_update_changeattr(struct inode *dir,
diff --git a/fs/nfs/nfs4client.c b/fs/nfs/nfs4client.c
index 11e3a28559..83378f69b3 100644
--- a/fs/nfs/nfs4client.c
+++ b/fs/nfs/nfs4client.c
@@ -231,9 +231,8 @@ struct nfs_client *nfs4_alloc_client(const struct nfs_client_initdata *cl_init)
 		__set_bit(NFS_CS_INFINITE_SLOTS, &clp->cl_flags);
 	__set_bit(NFS_CS_DISCRTRY, &clp->cl_flags);
 	__set_bit(NFS_CS_NO_RETRANS_TIMEOUT, &clp->cl_flags);
-
-	if (test_bit(NFS_CS_DS, &cl_init->init_flags))
-		__set_bit(NFS_CS_DS, &clp->cl_flags);
+	if (test_bit(NFS_CS_PNFS, &cl_init->init_flags))
+		__set_bit(NFS_CS_PNFS, &clp->cl_flags);
 	/*
 	 * Set up the connection to the server before we add add to the
 	 * global list.
@@ -924,6 +923,7 @@ static int nfs4_set_client(struct nfs_server *server,
 	else
 		cl_init.max_connect = max_connect;
 	switch (proto) {
+	case XPRT_TRANSPORT_RDMA:
 	case XPRT_TRANSPORT_TCP:
 	case XPRT_TRANSPORT_TCP_TLS:
 		cl_init.nconnect = nconnect;
@@ -1000,6 +1000,7 @@ struct nfs_client *nfs4_set_ds_client(struct nfs_server *mds_srv,
 	cl_init.hostname = buf;
 
 	switch (ds_proto) {
+	case XPRT_TRANSPORT_RDMA:
 	case XPRT_TRANSPORT_TCP:
 	case XPRT_TRANSPORT_TCP_TLS:
 		if (mds_clp->cl_nconnect > 1) {
@@ -1011,7 +1012,6 @@ struct nfs_client *nfs4_set_ds_client(struct nfs_server *mds_srv,
 	if (mds_srv->flags & NFS_MOUNT_NORESVPORT)
 		__set_bit(NFS_CS_NORESVPORT, &cl_init.init_flags);
 
-	__set_bit(NFS_CS_DS, &cl_init.init_flags);
 	__set_bit(NFS_CS_PNFS, &cl_init.init_flags);
 	cl_init.max_connect = NFS_MAX_TRANSPORTS;
 	/*
diff --git a/fs/nfs/nfs4file.c b/fs/nfs/nfs4file.c
index e238abc78a..1cd9652f3c 100644
--- a/fs/nfs/nfs4file.c
+++ b/fs/nfs/nfs4file.c
@@ -439,7 +439,7 @@ void nfs42_ssc_unregister_ops(void)
 }
 #endif /* CONFIG_NFS_V4_2 */
 
-static int nfs4_setlease(struct file *file, int arg, struct file_lock **lease,
+static int nfs4_setlease(struct file *file, int arg, struct file_lease **lease,
 			 void **priv)
 {
 	return nfs4_proc_setlease(file, arg, lease, priv);
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index b2ff8c7a21..bff9d66007 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -3069,10 +3069,8 @@ static int _nfs4_open_and_get_state(struct nfs4_opendata *opendata,
 	fmode_t acc_mode = _nfs4_ctx_to_accessmode(ctx);
 	struct inode *dir = d_inode(opendata->dir);
 	unsigned long dir_verifier;
-	unsigned int seq;
 	int ret;
 
-	seq = raw_seqcount_begin(&sp->so_reclaim_seqcount);
 	dir_verifier = nfs_save_change_attribute(dir);
 
 	ret = _nfs4_proc_open(opendata, ctx);
@@ -3125,11 +3123,8 @@ static int _nfs4_open_and_get_state(struct nfs4_opendata *opendata,
 	if (ret != 0)
 		goto out;
 
-	if (d_inode(dentry) == state->inode) {
+	if (d_inode(dentry) == state->inode)
 		nfs_inode_attach_open_context(ctx);
-		if (read_seqcount_retry(&sp->so_reclaim_seqcount, seq))
-			nfs4_schedule_stateid_recovery(server, state);
-	}
 
 out:
 	if (!opendata->cancelled) {
@@ -4028,6 +4023,23 @@ static void test_fs_location_for_trunking(struct nfs4_fs_location *location,
 	}
 }
 
+static bool _is_same_nfs4_pathname(struct nfs4_pathname *path1,
+				   struct nfs4_pathname *path2)
+{
+	int i;
+
+	if (path1->ncomponents != path2->ncomponents)
+		return false;
+	for (i = 0; i < path1->ncomponents; i++) {
+		if (path1->components[i].len != path2->components[i].len)
+			return false;
+		if (memcmp(path1->components[i].data, path2->components[i].data,
+				path1->components[i].len))
+			return false;
+	}
+	return true;
+}
+
 static int _nfs4_discover_trunking(struct nfs_server *server,
 				   struct nfs_fh *fhandle)
 {
@@ -4061,9 +4073,13 @@ static int _nfs4_discover_trunking(struct nfs_server *server,
 	if (status)
 		goto out_free_3;
 
-	for (i = 0; i < locations->nlocations; i++)
+	for (i = 0; i < locations->nlocations; i++) {
+		if (!_is_same_nfs4_pathname(&locations->fs_path,
+					&locations->locations[i].rootpath))
+			continue;
 		test_fs_location_for_trunking(&locations->locations[i], clp,
 					      server);
+	}
 out_free_3:
 	kfree(locations->fattr);
 out_free_2:
@@ -5461,7 +5477,7 @@ static bool nfs4_read_plus_not_supported(struct rpc_task *task,
 	struct rpc_message *msg = &task->tk_msg;
 
 	if (msg->rpc_proc == &nfs4_procedures[NFSPROC4_CLNT_READ_PLUS] &&
-	    server->caps & NFS_CAP_READ_PLUS && task->tk_status == -ENOTSUPP) {
+	    task->tk_status == -ENOTSUPP) {
 		server->caps &= ~NFS_CAP_READ_PLUS;
 		msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READ];
 		rpc_restart_call_prepare(task);
@@ -6273,6 +6289,7 @@ nfs4_set_security_label(struct inode *inode, const void *buf, size_t buflen)
 	if (status == 0)
 		nfs_setsecurity(inode, fattr);
 
+	nfs_free_fattr(fattr);
 	return status;
 }
 #endif	/* CONFIG_NFS_V4_SECURITY_LABEL */
@@ -6800,7 +6817,7 @@ static int _nfs4_proc_getlk(struct nfs4_state *state, int cmd, struct file_lock
 	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
 	switch (status) {
 		case 0:
-			request->fl_type = F_UNLCK;
+			request->c.flc_type = F_UNLCK;
 			break;
 		case -NFS4ERR_DENIED:
 			status = 0;
@@ -7018,8 +7035,8 @@ static struct rpc_task *nfs4_do_unlck(struct file_lock *fl,
 	/* Ensure this is an unlock - when canceling a lock, the
 	 * canceled lock is passed in, and it won't be an unlock.
 	 */
-	fl->fl_type = F_UNLCK;
-	if (fl->fl_flags & FL_CLOSE)
+	fl->c.flc_type = F_UNLCK;
+	if (fl->c.flc_flags & FL_CLOSE)
 		set_bit(NFS_CONTEXT_UNLOCK, &ctx->flags);
 
 	data = nfs4_alloc_unlockdata(fl, ctx, lsp, seqid);
@@ -7045,11 +7062,11 @@ static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock *
 	struct rpc_task *task;
 	struct nfs_seqid *(*alloc_seqid)(struct nfs_seqid_counter *, gfp_t);
 	int status = 0;
-	unsigned char fl_flags = request->fl_flags;
+	unsigned char saved_flags = request->c.flc_flags;
 
 	status = nfs4_set_lock_state(state, request);
 	/* Unlock _before_ we do the RPC call */
-	request->fl_flags |= FL_EXISTS;
+	request->c.flc_flags |= FL_EXISTS;
 	/* Exclude nfs_delegation_claim_locks() */
 	mutex_lock(&sp->so_delegreturn_mutex);
 	/* Exclude nfs4_reclaim_open_stateid() - note nesting! */
@@ -7073,14 +7090,16 @@ static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock *
 	status = -ENOMEM;
 	if (IS_ERR(seqid))
 		goto out;
-	task = nfs4_do_unlck(request, nfs_file_open_context(request->fl_file), lsp, seqid);
+	task = nfs4_do_unlck(request,
+			     nfs_file_open_context(request->c.flc_file),
+			     lsp, seqid);
 	status = PTR_ERR(task);
 	if (IS_ERR(task))
 		goto out;
 	status = rpc_wait_for_completion_task(task);
 	rpc_put_task(task);
 out:
-	request->fl_flags = fl_flags;
+	request->c.flc_flags = saved_flags;
 	trace_nfs4_unlock(request, state, F_SETLK, status);
 	return status;
 }
@@ -7191,7 +7210,7 @@ static void nfs4_lock_done(struct rpc_task *task, void *calldata)
 		renew_lease(NFS_SERVER(d_inode(data->ctx->dentry)),
 				data->timestamp);
 		if (data->arg.new_lock && !data->cancelled) {
-			data->fl.fl_flags &= ~(FL_SLEEP | FL_ACCESS);
+			data->fl.c.flc_flags &= ~(FL_SLEEP | FL_ACCESS);
 			if (locks_lock_inode_wait(lsp->ls_state->inode, &data->fl) < 0)
 				goto out_restart;
 		}
@@ -7292,7 +7311,8 @@ static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *f
 	if (nfs_server_capable(state->inode, NFS_CAP_MOVEABLE))
 		task_setup_data.flags |= RPC_TASK_MOVEABLE;
 
-	data = nfs4_alloc_lockdata(fl, nfs_file_open_context(fl->fl_file),
+	data = nfs4_alloc_lockdata(fl,
+				   nfs_file_open_context(fl->c.flc_file),
 				   fl->fl_u.nfs4_fl.owner, GFP_KERNEL);
 	if (data == NULL)
 		return -ENOMEM;
@@ -7398,10 +7418,10 @@ static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock
 {
 	struct nfs_inode *nfsi = NFS_I(state->inode);
 	struct nfs4_state_owner *sp = state->owner;
-	unsigned char fl_flags = request->fl_flags;
+	unsigned char flags = request->c.flc_flags;
 	int status;
 
-	request->fl_flags |= FL_ACCESS;
+	request->c.flc_flags |= FL_ACCESS;
 	status = locks_lock_inode_wait(state->inode, request);
 	if (status < 0)
 		goto out;
@@ -7410,7 +7430,7 @@ static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock
 	if (test_bit(NFS_DELEGATED_STATE, &state->flags)) {
 		/* Yes: cache locks! */
 		/* ...but avoid races with delegation recall... */
-		request->fl_flags = fl_flags & ~FL_SLEEP;
+		request->c.flc_flags = flags & ~FL_SLEEP;
 		status = locks_lock_inode_wait(state->inode, request);
 		up_read(&nfsi->rwsem);
 		mutex_unlock(&sp->so_delegreturn_mutex);
@@ -7420,7 +7440,7 @@ static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock
 	mutex_unlock(&sp->so_delegreturn_mutex);
 	status = _nfs4_do_setlk(state, cmd, request, NFS_LOCK_NEW);
 out:
-	request->fl_flags = fl_flags;
+	request->c.flc_flags = flags;
 	return status;
 }
 
@@ -7562,7 +7582,7 @@ nfs4_proc_lock(struct file *filp, int cmd, struct file_lock *request)
 	if (!(IS_SETLK(cmd) || IS_SETLKW(cmd)))
 		return -EINVAL;
 
-	if (request->fl_type == F_UNLCK) {
+	if (lock_is_unlock(request)) {
 		if (state != NULL)
 			return nfs4_proc_unlck(state, cmd, request);
 		return 0;
@@ -7571,7 +7591,7 @@ nfs4_proc_lock(struct file *filp, int cmd, struct file_lock *request)
 	if (state == NULL)
 		return -ENOLCK;
 
-	if ((request->fl_flags & FL_POSIX) &&
+	if ((request->c.flc_flags & FL_POSIX) &&
 	    !test_bit(NFS_STATE_POSIX_LOCKS, &state->flags))
 		return -ENOLCK;
 
@@ -7579,7 +7599,7 @@ nfs4_proc_lock(struct file *filp, int cmd, struct file_lock *request)
 	 * Don't rely on the VFS having checked the file open mode,
 	 * since it won't do this for flock() locks.
 	 */
-	switch (request->fl_type) {
+	switch (request->c.flc_type) {
 	case F_RDLCK:
 		if (!(filp->f_mode & FMODE_READ))
 			return -EBADF;
@@ -7601,7 +7621,7 @@ static int nfs4_delete_lease(struct file *file, void **priv)
 	return generic_setlease(file, F_UNLCK, NULL, priv);
 }
 
-static int nfs4_add_lease(struct file *file, int arg, struct file_lock **lease,
+static int nfs4_add_lease(struct file *file, int arg, struct file_lease **lease,
 			  void **priv)
 {
 	struct inode *inode = file_inode(file);
@@ -7619,7 +7639,7 @@ static int nfs4_add_lease(struct file *file, int arg, struct file_lock **lease,
 	return -EAGAIN;
 }
 
-int nfs4_proc_setlease(struct file *file, int arg, struct file_lock **lease,
+int nfs4_proc_setlease(struct file *file, int arg, struct file_lease **lease,
 		       void **priv)
 {
 	switch (arg) {
@@ -8820,7 +8840,7 @@ nfs4_run_exchange_id(struct nfs_client *clp, const struct cred *cred,
 #ifdef CONFIG_NFS_V4_1_MIGRATION
 	calldata->args.flags |= EXCHGID4_FLAG_SUPP_MOVED_MIGR;
 #endif
-	if (test_bit(NFS_CS_DS, &clp->cl_flags))
+	if (test_bit(NFS_CS_PNFS, &clp->cl_flags))
 		calldata->args.flags |= EXCHGID4_FLAG_USE_PNFS_DS;
 	msg.rpc_argp = &calldata->args;
 	msg.rpc_resp = &calldata->res;
@@ -8970,10 +8990,12 @@ try_again:
 		return;
 
 	status = task->tk_status;
-	if (status == 0)
+	if (status == 0) {
 		status = nfs4_detect_session_trunking(adata->clp,
 				task->tk_msg.rpc_resp, xprt);
-
+		trace_nfs4_trunked_exchange_id(adata->clp,
+			xprt->address_strings[RPC_DISPLAY_ADDR], status);
+	}
 	if (status == 0)
 		rpc_clnt_xprt_switch_add_xprt(clnt, xprt);
 	else if (status != -NFS4ERR_DELAY && rpc_clnt_xprt_switch_has_addr(clnt,
diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c
index 9a5d911a7e..5b452411e8 100644
--- a/fs/nfs/nfs4state.c
+++ b/fs/nfs/nfs4state.c
@@ -513,7 +513,6 @@ nfs4_alloc_state_owner(struct nfs_server *server,
 	nfs4_init_seqid_counter(&sp->so_seqid);
 	atomic_set(&sp->so_count, 1);
 	INIT_LIST_HEAD(&sp->so_lru);
-	seqcount_spinlock_init(&sp->so_reclaim_seqcount, &sp->so_lock);
 	mutex_init(&sp->so_delegreturn_mutex);
 	return sp;
 }
@@ -847,15 +846,15 @@ void nfs4_close_sync(struct nfs4_state *state, fmode_t fmode)
  */
 static struct nfs4_lock_state *
 __nfs4_find_lock_state(struct nfs4_state *state,
-		       fl_owner_t fl_owner, fl_owner_t fl_owner2)
+		       fl_owner_t owner, fl_owner_t owner2)
 {
 	struct nfs4_lock_state *pos, *ret = NULL;
 	list_for_each_entry(pos, &state->lock_states, ls_locks) {
-		if (pos->ls_owner == fl_owner) {
+		if (pos->ls_owner == owner) {
 			ret = pos;
 			break;
 		}
-		if (pos->ls_owner == fl_owner2)
+		if (pos->ls_owner == owner2)
 			ret = pos;
 	}
 	if (ret)
@@ -868,7 +867,7 @@ __nfs4_find_lock_state(struct nfs4_state *state,
  * exists, return an uninitialized one.
  *
  */
-static struct nfs4_lock_state *nfs4_alloc_lock_state(struct nfs4_state *state, fl_owner_t fl_owner)
+static struct nfs4_lock_state *nfs4_alloc_lock_state(struct nfs4_state *state, fl_owner_t owner)
 {
 	struct nfs4_lock_state *lsp;
 	struct nfs_server *server = state->owner->so_server;
@@ -879,7 +878,7 @@ static struct nfs4_lock_state *nfs4_alloc_lock_state(struct nfs4_state *state, f
 	nfs4_init_seqid_counter(&lsp->ls_seqid);
 	refcount_set(&lsp->ls_count, 1);
 	lsp->ls_state = state;
-	lsp->ls_owner = fl_owner;
+	lsp->ls_owner = owner;
 	lsp->ls_seqid.owner_id = ida_alloc(&server->lockowner_id, GFP_KERNEL_ACCOUNT);
 	if (lsp->ls_seqid.owner_id < 0)
 		goto out_free;
@@ -980,7 +979,7 @@ int nfs4_set_lock_state(struct nfs4_state *state, struct file_lock *fl)
 
 	if (fl->fl_ops != NULL)
 		return 0;
-	lsp = nfs4_get_lock_state(state, fl->fl_owner);
+	lsp = nfs4_get_lock_state(state, fl->c.flc_owner);
 	if (lsp == NULL)
 		return -ENOMEM;
 	fl->fl_u.nfs4_fl.owner = lsp;
@@ -993,7 +992,7 @@ static int nfs4_copy_lock_stateid(nfs4_stateid *dst,
 		const struct nfs_lock_context *l_ctx)
 {
 	struct nfs4_lock_state *lsp;
-	fl_owner_t fl_owner, fl_flock_owner;
+	fl_owner_t owner, fl_flock_owner;
 	int ret = -ENOENT;
 
 	if (l_ctx == NULL)
@@ -1002,11 +1001,11 @@ static int nfs4_copy_lock_stateid(nfs4_stateid *dst,
 	if (test_bit(LK_STATE_IN_USE, &state->flags) == 0)
 		goto out;
 
-	fl_owner = l_ctx->lockowner;
+	owner = l_ctx->lockowner;
 	fl_flock_owner = l_ctx->open_context->flock_owner;
 
 	spin_lock(&state->state_lock);
-	lsp = __nfs4_find_lock_state(state, fl_owner, fl_flock_owner);
+	lsp = __nfs4_find_lock_state(state, owner, fl_flock_owner);
 	if (lsp && test_bit(NFS_LOCK_LOST, &lsp->ls_flags))
 		ret = -EIO;
 	else if (lsp != NULL && test_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags) != 0) {
@@ -1529,8 +1528,8 @@ static int nfs4_reclaim_locks(struct nfs4_state *state, const struct nfs4_state_
 	down_write(&nfsi->rwsem);
 	spin_lock(&flctx->flc_lock);
 restart:
-	list_for_each_entry(fl, list, fl_list) {
-		if (nfs_file_open_context(fl->fl_file)->state != state)
+	for_each_file_lock(fl, list) {
+		if (nfs_file_open_context(fl->c.flc_file)->state != state)
 			continue;
 		spin_unlock(&flctx->flc_lock);
 		status = ops->recover_lock(state, fl);
@@ -1667,7 +1666,6 @@ static int nfs4_reclaim_open_state(struct nfs4_state_owner *sp,
 	 * server that doesn't support a grace period.
 	 */
 	spin_lock(&sp->so_lock);
-	raw_write_seqcount_begin(&sp->so_reclaim_seqcount);
 restart:
 	list_for_each_entry(state, &sp->so_states, open_states) {
 		if (!test_and_clear_bit(ops->state_flag_bit, &state->flags))
@@ -1735,7 +1733,6 @@ restart:
 		spin_lock(&sp->so_lock);
 		goto restart;
 	}
-	raw_write_seqcount_end(&sp->so_reclaim_seqcount);
 	spin_unlock(&sp->so_lock);
 #ifdef CONFIG_NFS_V4_2
 	if (found_ssc_copy_state)
@@ -1745,7 +1742,6 @@ restart:
 out_err:
 	nfs4_put_open_state(state);
 	spin_lock(&sp->so_lock);
-	raw_write_seqcount_end(&sp->so_reclaim_seqcount);
 	spin_unlock(&sp->so_lock);
 	return status;
 }
@@ -1928,9 +1924,12 @@ static int nfs4_do_reclaim(struct nfs_client *clp, const struct nfs4_state_recov
 	struct nfs_server *server;
 	struct rb_node *pos;
 	LIST_HEAD(freeme);
-	int status = 0;
 	int lost_locks = 0;
+	int status;
 
+	status = nfs4_begin_drain_session(clp);
+	if (status < 0)
+		return status;
 restart:
 	rcu_read_lock();
 	list_for_each_entry_rcu(server, &clp->cl_superblocks, client_link) {
@@ -2117,6 +2116,7 @@ static int nfs4_try_migration(struct nfs_server *server, const struct cred *cred
 {
 	struct nfs_client *clp = server->nfs_client;
 	struct nfs4_fs_locations *locations = NULL;
+	struct nfs_fattr *fattr;
 	struct inode *inode;
 	struct page *page;
 	int status, result;
@@ -2126,19 +2126,16 @@ static int nfs4_try_migration(struct nfs_server *server, const struct cred *cred
 			(unsigned long long)server->fsid.minor,
 			clp->cl_hostname);
 
-	result = 0;
 	page = alloc_page(GFP_KERNEL);
 	locations = kmalloc(sizeof(struct nfs4_fs_locations), GFP_KERNEL);
-	if (page == NULL || locations == NULL) {
-		dprintk("<-- %s: no memory\n", __func__);
-		goto out;
-	}
-	locations->fattr = nfs_alloc_fattr();
-	if (locations->fattr == NULL) {
+	fattr = nfs_alloc_fattr();
+	if (page == NULL || locations == NULL || fattr == NULL) {
 		dprintk("<-- %s: no memory\n", __func__);
+		result = 0;
 		goto out;
 	}
 
+	locations->fattr = fattr;
 	inode = d_inode(server->super->s_root);
 	result = nfs4_proc_get_locations(server, NFS_FH(inode), locations,
 					 page, cred);
@@ -2694,6 +2691,9 @@ static void nfs4_state_manager(struct nfs_client *clp)
 		/* Detect expired delegations... */
 		if (test_and_clear_bit(NFS4CLNT_DELEGATION_EXPIRED, &clp->cl_state)) {
 			section = "detect expired delegations";
+			status = nfs4_begin_drain_session(clp);
+			if (status < 0)
+				goto out_error;
 			nfs_reap_expired_delegations(clp);
 			continue;
 		}
diff --git a/fs/nfs/nfs4super.c b/fs/nfs/nfs4super.c
index d09bcfd7db..8da5a9c000 100644
--- a/fs/nfs/nfs4super.c
+++ b/fs/nfs/nfs4super.c
@@ -145,6 +145,7 @@ static int do_nfs4_mount(struct nfs_server *server,
 			 const char *export_path)
 {
 	struct nfs_fs_context *root_ctx;
+	struct nfs_fs_context *ctx;
 	struct fs_context *root_fc;
 	struct vfsmount *root_mnt;
 	struct dentry *dentry;
@@ -157,6 +158,12 @@ static int do_nfs4_mount(struct nfs_server *server,
 		.dirfd	= -1,
 	};
 
+	struct fs_parameter param_fsc = {
+		.key	= "fsc",
+		.type	= fs_value_is_string,
+		.dirfd	= -1,
+	};
+
 	if (IS_ERR(server))
 		return PTR_ERR(server);
 
@@ -168,9 +175,26 @@ static int do_nfs4_mount(struct nfs_server *server,
 	kfree(root_fc->source);
 	root_fc->source = NULL;
 
+	ctx = nfs_fc2context(fc);
 	root_ctx = nfs_fc2context(root_fc);
 	root_ctx->internal = true;
 	root_ctx->server = server;
+
+	if (ctx->fscache_uniq) {
+		len = strlen(ctx->fscache_uniq);
+		param_fsc.size = len;
+		param_fsc.string = kmemdup_nul(ctx->fscache_uniq, len, GFP_KERNEL);
+		if (param_fsc.string == NULL) {
+			put_fs_context(root_fc);
+			return -ENOMEM;
+		}
+		ret = vfs_parse_fs_param(root_fc, &param_fsc);
+		kfree(param_fsc.string);
+		if (ret < 0) {
+			put_fs_context(root_fc);
+			return ret;
+		}
+	}
 	/* We leave export_path unset as it's not used to find the root. */
 
 	len = strlen(hostname) + 5;
diff --git a/fs/nfs/nfs4trace.c b/fs/nfs/nfs4trace.c
index d9ac556beb..d22c6670f7 100644
--- a/fs/nfs/nfs4trace.c
+++ b/fs/nfs/nfs4trace.c
@@ -28,4 +28,6 @@ EXPORT_TRACEPOINT_SYMBOL_GPL(pnfs_mds_fallback_write_pagelist);
 EXPORT_TRACEPOINT_SYMBOL_GPL(ff_layout_read_error);
 EXPORT_TRACEPOINT_SYMBOL_GPL(ff_layout_write_error);
 EXPORT_TRACEPOINT_SYMBOL_GPL(ff_layout_commit_error);
+
+EXPORT_TRACEPOINT_SYMBOL_GPL(fl_getdevinfo);
 #endif
diff --git a/fs/nfs/nfs4trace.h b/fs/nfs/nfs4trace.h
index d27919d724..4de8780a7c 100644
--- a/fs/nfs/nfs4trace.h
+++ b/fs/nfs/nfs4trace.h
@@ -47,7 +47,7 @@ DECLARE_EVENT_CLASS(nfs4_clientid_event,
 
 		TP_fast_assign(
 			__entry->error = error < 0 ? -error : 0;
-			__assign_str(dstaddr, clp->cl_hostname);
+			__assign_str(dstaddr);
 		),
 
 		TP_printk(
@@ -77,6 +77,36 @@ DEFINE_NFS4_CLIENTID_EVENT(nfs4_bind_conn_to_session);
 DEFINE_NFS4_CLIENTID_EVENT(nfs4_sequence);
 DEFINE_NFS4_CLIENTID_EVENT(nfs4_reclaim_complete);
 
+TRACE_EVENT(nfs4_trunked_exchange_id,
+		TP_PROTO(
+			const struct nfs_client *clp,
+			const char *addr,
+			int error
+		),
+
+		TP_ARGS(clp, addr, error),
+
+		TP_STRUCT__entry(
+			__string(main_addr, clp->cl_hostname)
+			__string(trunk_addr, addr)
+			__field(unsigned long, error)
+		),
+
+		TP_fast_assign(
+			__entry->error = error < 0 ? -error : 0;
+			__assign_str(main_addr);
+			__assign_str(trunk_addr);
+		),
+
+		TP_printk(
+			"error=%ld (%s) main_addr=%s trunk_addr=%s",
+			-__entry->error,
+			show_nfs4_status(__entry->error),
+			__get_str(main_addr),
+			__get_str(trunk_addr)
+		)
+);
+
 TRACE_EVENT(nfs4_sequence_done,
 		TP_PROTO(
 			const struct nfs4_session *session,
@@ -335,7 +365,7 @@ TRACE_EVENT(nfs4_state_mgr,
 
 		TP_fast_assign(
 			__entry->state = clp->cl_state;
-			__assign_str(hostname, clp->cl_hostname);
+			__assign_str(hostname);
 		),
 
 		TP_printk(
@@ -363,8 +393,8 @@ TRACE_EVENT(nfs4_state_mgr_failed,
 		TP_fast_assign(
 			__entry->error = status < 0 ? -status : 0;
 			__entry->state = clp->cl_state;
-			__assign_str(hostname, clp->cl_hostname);
-			__assign_str(section, section);
+			__assign_str(hostname);
+			__assign_str(section);
 		),
 
 		TP_printk(
@@ -548,7 +578,7 @@ DECLARE_EVENT_CLASS(nfs4_open_event,
 				__entry->fhandle = 0;
 			}
 			__entry->dir = NFS_FILEID(d_inode(ctx->dentry->d_parent));
-			__assign_str(name, ctx->dentry->d_name.name);
+			__assign_str(name);
 		),
 
 		TP_printk(
@@ -699,7 +729,7 @@ DECLARE_EVENT_CLASS(nfs4_lock_event,
 
 			__entry->error = error < 0 ? -error : 0;
 			__entry->cmd = cmd;
-			__entry->type = request->fl_type;
+			__entry->type = request->c.flc_type;
 			__entry->start = request->fl_start;
 			__entry->end = request->fl_end;
 			__entry->dev = inode->i_sb->s_dev;
@@ -771,7 +801,7 @@ TRACE_EVENT(nfs4_set_lock,
 
 			__entry->error = error < 0 ? -error : 0;
 			__entry->cmd = cmd;
-			__entry->type = request->fl_type;
+			__entry->type = request->c.flc_type;
 			__entry->start = request->fl_start;
 			__entry->end = request->fl_end;
 			__entry->dev = inode->i_sb->s_dev;
@@ -1042,7 +1072,7 @@ DECLARE_EVENT_CLASS(nfs4_lookup_event,
 			__entry->dev = dir->i_sb->s_dev;
 			__entry->dir = NFS_FILEID(dir);
 			__entry->error = -error;
-			__assign_str(name, name->name);
+			__assign_str(name);
 		),
 
 		TP_printk(
@@ -1126,8 +1156,8 @@ TRACE_EVENT(nfs4_rename,
 			__entry->olddir = NFS_FILEID(olddir);
 			__entry->newdir = NFS_FILEID(newdir);
 			__entry->error = error < 0 ? -error : 0;
-			__assign_str(oldname, oldname->name);
-			__assign_str(newname, newname->name);
+			__assign_str(oldname);
+			__assign_str(newname);
 		),
 
 		TP_printk(
@@ -1329,7 +1359,7 @@ DECLARE_EVENT_CLASS(nfs4_inode_callback_event,
 				__entry->fileid = 0;
 				__entry->dev = 0;
 			}
-			__assign_str(dstaddr, clp ? clp->cl_hostname : "unknown");
+			__assign_str(dstaddr);
 		),
 
 		TP_printk(
@@ -1386,7 +1416,7 @@ DECLARE_EVENT_CLASS(nfs4_inode_stateid_callback_event,
 				__entry->fileid = 0;
 				__entry->dev = 0;
 			}
-			__assign_str(dstaddr, clp ? clp->cl_hostname : "unknown");
+			__assign_str(dstaddr);
 			__entry->stateid_seq =
 				be32_to_cpu(stateid->seqid);
 			__entry->stateid_hash =
@@ -1930,7 +1960,7 @@ DECLARE_EVENT_CLASS(nfs4_deviceid_event,
 		),
 
 		TP_fast_assign(
-			__assign_str(dstaddr, clp->cl_hostname);
+			__assign_str(dstaddr);
 			memcpy(__entry->deviceid, deviceid->data,
 			       NFS4_DEVICEID4_SIZE);
 		),
@@ -1968,7 +1998,7 @@ DECLARE_EVENT_CLASS(nfs4_deviceid_status,
 		TP_fast_assign(
 			__entry->dev = server->s_dev;
 			__entry->status = status;
-			__assign_str(dstaddr, server->nfs_client->cl_hostname);
+			__assign_str(dstaddr);
 			memcpy(__entry->deviceid, deviceid->data,
 			       NFS4_DEVICEID4_SIZE);
 		),
@@ -1991,6 +2021,34 @@ DECLARE_EVENT_CLASS(nfs4_deviceid_status,
 DEFINE_PNFS_DEVICEID_STATUS(nfs4_getdeviceinfo);
 DEFINE_PNFS_DEVICEID_STATUS(nfs4_find_deviceid);
 
+TRACE_EVENT(fl_getdevinfo,
+		TP_PROTO(
+			const struct nfs_server *server,
+			const struct nfs4_deviceid *deviceid,
+			char *ds_remotestr
+		),
+		TP_ARGS(server, deviceid, ds_remotestr),
+
+		TP_STRUCT__entry(
+			__string(mds_addr, server->nfs_client->cl_hostname)
+			__array(unsigned char, deviceid, NFS4_DEVICEID4_SIZE)
+			__string(ds_ips, ds_remotestr)
+		),
+
+		TP_fast_assign(
+			__assign_str(mds_addr);
+			__assign_str(ds_ips);
+			memcpy(__entry->deviceid, deviceid->data,
+			       NFS4_DEVICEID4_SIZE);
+		),
+		TP_printk(
+			"deviceid=%s, mds_addr=%s, ds_ips=%s",
+			__print_hex(__entry->deviceid, NFS4_DEVICEID4_SIZE),
+			__get_str(mds_addr),
+			__get_str(ds_ips)
+		)
+);
+
 DECLARE_EVENT_CLASS(nfs4_flexfiles_io_event,
 		TP_PROTO(
 			const struct nfs_pgio_header *hdr
@@ -2025,9 +2083,7 @@ DECLARE_EVENT_CLASS(nfs4_flexfiles_io_event,
 				be32_to_cpu(hdr->args.stateid.seqid);
 			__entry->stateid_hash =
 				nfs_stateid_hash(&hdr->args.stateid);
-			__assign_str(dstaddr, hdr->ds_clp ?
-				rpc_peeraddr2str(hdr->ds_clp->cl_rpcclient,
-					RPC_DISPLAY_ADDR) : "unknown");
+			__assign_str(dstaddr);
 		),
 
 		TP_printk(
@@ -2081,9 +2137,7 @@ TRACE_EVENT(ff_layout_commit_error,
 			__entry->dev = inode->i_sb->s_dev;
 			__entry->offset = data->args.offset;
 			__entry->count = data->args.count;
-			__assign_str(dstaddr, data->ds_clp ?
-				rpc_peeraddr2str(data->ds_clp->cl_rpcclient,
-					RPC_DISPLAY_ADDR) : "unknown");
+			__assign_str(dstaddr);
 		),
 
 		TP_printk(
@@ -2521,7 +2575,7 @@ DECLARE_EVENT_CLASS(nfs4_xattr_event,
 			__entry->dev = inode->i_sb->s_dev;
 			__entry->fileid = NFS_FILEID(inode);
 			__entry->fhandle = nfs_fhandle_hash(NFS_FH(inode));
-			__assign_str(name, name);
+			__assign_str(name);
 		),
 
 		TP_printk(
diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c
index 69406e60f3..1416099dfc 100644
--- a/fs/nfs/nfs4xdr.c
+++ b/fs/nfs/nfs4xdr.c
@@ -1305,7 +1305,7 @@ static void encode_link(struct xdr_stream *xdr, const struct qstr *name, struct
 
 static inline int nfs4_lock_type(struct file_lock *fl, int block)
 {
-	if (fl->fl_type == F_RDLCK)
+	if (lock_is_read(fl))
 		return block ? NFS4_READW_LT : NFS4_READ_LT;
 	return block ? NFS4_WRITEW_LT : NFS4_WRITE_LT;
 }
@@ -5052,10 +5052,10 @@ static int decode_lock_denied (struct xdr_stream *xdr, struct file_lock *fl)
 		fl->fl_end = fl->fl_start + (loff_t)length - 1;
 		if (length == ~(uint64_t)0)
 			fl->fl_end = OFFSET_MAX;
-		fl->fl_type = F_WRLCK;
+		fl->c.flc_type = F_WRLCK;
 		if (type & 1)
-			fl->fl_type = F_RDLCK;
-		fl->fl_pid = 0;
+			fl->c.flc_type = F_RDLCK;
+		fl->c.flc_pid = 0;
 	}
 	p = xdr_decode_hyper(p, &clientid); /* read 8 bytes */
 	namelen = be32_to_cpup(p); /* read 4 bytes */  /* have read all 32 bytes now */
diff --git a/fs/nfs/nfstrace.h b/fs/nfs/nfstrace.h
index afedb449b5..352fdaed40 100644
--- a/fs/nfs/nfstrace.h
+++ b/fs/nfs/nfstrace.h
@@ -409,7 +409,7 @@ DECLARE_EVENT_CLASS(nfs_lookup_event,
 			__entry->dir = NFS_FILEID(dir);
 			__entry->flags = flags;
 			__entry->fileid = d_is_negative(dentry) ? 0 : NFS_FILEID(d_inode(dentry));
-			__assign_str(name, dentry->d_name.name);
+			__assign_str(name);
 		),
 
 		TP_printk(
@@ -457,7 +457,7 @@ DECLARE_EVENT_CLASS(nfs_lookup_event_done,
 			__entry->error = error < 0 ? -error : 0;
 			__entry->flags = flags;
 			__entry->fileid = d_is_negative(dentry) ? 0 : NFS_FILEID(d_inode(dentry));
-			__assign_str(name, dentry->d_name.name);
+			__assign_str(name);
 		),
 
 		TP_printk(
@@ -512,7 +512,7 @@ TRACE_EVENT(nfs_atomic_open_enter,
 			__entry->dir = NFS_FILEID(dir);
 			__entry->flags = flags;
 			__entry->fmode = (__force unsigned long)ctx->mode;
-			__assign_str(name, ctx->dentry->d_name.name);
+			__assign_str(name);
 		),
 
 		TP_printk(
@@ -551,7 +551,7 @@ TRACE_EVENT(nfs_atomic_open_exit,
 			__entry->dir = NFS_FILEID(dir);
 			__entry->flags = flags;
 			__entry->fmode = (__force unsigned long)ctx->mode;
-			__assign_str(name, ctx->dentry->d_name.name);
+			__assign_str(name);
 		),
 
 		TP_printk(
@@ -587,7 +587,7 @@ TRACE_EVENT(nfs_create_enter,
 			__entry->dev = dir->i_sb->s_dev;
 			__entry->dir = NFS_FILEID(dir);
 			__entry->flags = flags;
-			__assign_str(name, dentry->d_name.name);
+			__assign_str(name);
 		),
 
 		TP_printk(
@@ -623,7 +623,7 @@ TRACE_EVENT(nfs_create_exit,
 			__entry->dev = dir->i_sb->s_dev;
 			__entry->dir = NFS_FILEID(dir);
 			__entry->flags = flags;
-			__assign_str(name, dentry->d_name.name);
+			__assign_str(name);
 		),
 
 		TP_printk(
@@ -654,7 +654,7 @@ DECLARE_EVENT_CLASS(nfs_directory_event,
 		TP_fast_assign(
 			__entry->dev = dir->i_sb->s_dev;
 			__entry->dir = NFS_FILEID(dir);
-			__assign_str(name, dentry->d_name.name);
+			__assign_str(name);
 		),
 
 		TP_printk(
@@ -693,7 +693,7 @@ DECLARE_EVENT_CLASS(nfs_directory_event_done,
 			__entry->dev = dir->i_sb->s_dev;
 			__entry->dir = NFS_FILEID(dir);
 			__entry->error = error < 0 ? -error : 0;
-			__assign_str(name, dentry->d_name.name);
+			__assign_str(name);
 		),
 
 		TP_printk(
@@ -747,7 +747,7 @@ TRACE_EVENT(nfs_link_enter,
 			__entry->dev = inode->i_sb->s_dev;
 			__entry->fileid = NFS_FILEID(inode);
 			__entry->dir = NFS_FILEID(dir);
-			__assign_str(name, dentry->d_name.name);
+			__assign_str(name);
 		),
 
 		TP_printk(
@@ -783,7 +783,7 @@ TRACE_EVENT(nfs_link_exit,
 			__entry->fileid = NFS_FILEID(inode);
 			__entry->dir = NFS_FILEID(dir);
 			__entry->error = error < 0 ? -error : 0;
-			__assign_str(name, dentry->d_name.name);
+			__assign_str(name);
 		),
 
 		TP_printk(
@@ -819,8 +819,8 @@ DECLARE_EVENT_CLASS(nfs_rename_event,
 			__entry->dev = old_dir->i_sb->s_dev;
 			__entry->old_dir = NFS_FILEID(old_dir);
 			__entry->new_dir = NFS_FILEID(new_dir);
-			__assign_str(old_name, old_dentry->d_name.name);
-			__assign_str(new_name, new_dentry->d_name.name);
+			__assign_str(old_name);
+			__assign_str(new_name);
 		),
 
 		TP_printk(
@@ -868,8 +868,8 @@ DECLARE_EVENT_CLASS(nfs_rename_event_done,
 			__entry->error = -error;
 			__entry->old_dir = NFS_FILEID(old_dir);
 			__entry->new_dir = NFS_FILEID(new_dir);
-			__assign_str(old_name, old_dentry->d_name.name);
-			__assign_str(new_name, new_dentry->d_name.name);
+			__assign_str(old_name);
+			__assign_str(new_name);
 		),
 
 		TP_printk(
@@ -939,10 +939,11 @@ TRACE_EVENT(nfs_sillyrename_unlink,
 DECLARE_EVENT_CLASS(nfs_folio_event,
 		TP_PROTO(
 			const struct inode *inode,
-			struct folio *folio
+			loff_t offset,
+			size_t count
 		),
 
-		TP_ARGS(inode, folio),
+		TP_ARGS(inode, offset, count),
 
 		TP_STRUCT__entry(
 			__field(dev_t, dev)
@@ -950,7 +951,7 @@ DECLARE_EVENT_CLASS(nfs_folio_event,
 			__field(u64, fileid)
 			__field(u64, version)
 			__field(loff_t, offset)
-			__field(u32, count)
+			__field(size_t, count)
 		),
 
 		TP_fast_assign(
@@ -960,13 +961,13 @@ DECLARE_EVENT_CLASS(nfs_folio_event,
 			__entry->fileid = nfsi->fileid;
 			__entry->fhandle = nfs_fhandle_hash(&nfsi->fh);
 			__entry->version = inode_peek_iversion_raw(inode);
-			__entry->offset = folio_file_pos(folio);
-			__entry->count = nfs_folio_length(folio);
+			__entry->offset = offset,
+			__entry->count = count;
 		),
 
 		TP_printk(
 			"fileid=%02x:%02x:%llu fhandle=0x%08x version=%llu "
-			"offset=%lld count=%u",
+			"offset=%lld count=%zu",
 			MAJOR(__entry->dev), MINOR(__entry->dev),
 			(unsigned long long)__entry->fileid,
 			__entry->fhandle, __entry->version,
@@ -978,18 +979,20 @@ DECLARE_EVENT_CLASS(nfs_folio_event,
 	DEFINE_EVENT(nfs_folio_event, name, \
 			TP_PROTO( \
 				const struct inode *inode, \
-				struct folio *folio \
+				loff_t offset, \
+				size_t count \
 			), \
-			TP_ARGS(inode, folio))
+			TP_ARGS(inode, offset, count))
 
 DECLARE_EVENT_CLASS(nfs_folio_event_done,
 		TP_PROTO(
 			const struct inode *inode,
-			struct folio *folio,
+			loff_t offset,
+			size_t count,
 			int ret
 		),
 
-		TP_ARGS(inode, folio, ret),
+		TP_ARGS(inode, offset, count, ret),
 
 		TP_STRUCT__entry(
 			__field(dev_t, dev)
@@ -998,7 +1001,7 @@ DECLARE_EVENT_CLASS(nfs_folio_event_done,
 			__field(u64, fileid)
 			__field(u64, version)
 			__field(loff_t, offset)
-			__field(u32, count)
+			__field(size_t, count)
 		),
 
 		TP_fast_assign(
@@ -1008,14 +1011,14 @@ DECLARE_EVENT_CLASS(nfs_folio_event_done,
 			__entry->fileid = nfsi->fileid;
 			__entry->fhandle = nfs_fhandle_hash(&nfsi->fh);
 			__entry->version = inode_peek_iversion_raw(inode);
-			__entry->offset = folio_file_pos(folio);
-			__entry->count = nfs_folio_length(folio);
+			__entry->offset = offset,
+			__entry->count = count,
 			__entry->ret = ret;
 		),
 
 		TP_printk(
 			"fileid=%02x:%02x:%llu fhandle=0x%08x version=%llu "
-			"offset=%lld count=%u ret=%d",
+			"offset=%lld count=%zu ret=%d",
 			MAJOR(__entry->dev), MINOR(__entry->dev),
 			(unsigned long long)__entry->fileid,
 			__entry->fhandle, __entry->version,
@@ -1027,10 +1030,11 @@ DECLARE_EVENT_CLASS(nfs_folio_event_done,
 	DEFINE_EVENT(nfs_folio_event_done, name, \
 			TP_PROTO( \
 				const struct inode *inode, \
-				struct folio *folio, \
+				loff_t offset, \
+				size_t count, \
 				int ret \
 			), \
-			TP_ARGS(inode, folio, ret))
+			TP_ARGS(inode, offset, count, ret))
 
 DEFINE_NFS_FOLIO_EVENT(nfs_aop_readpage);
 DEFINE_NFS_FOLIO_EVENT_DONE(nfs_aop_readpage_done);
@@ -1636,8 +1640,8 @@ TRACE_EVENT(nfs_mount_assign,
 	),
 
 	TP_fast_assign(
-		__assign_str(option, option);
-		__assign_str(value, value);
+		__assign_str(option);
+		__assign_str(value);
 	),
 
 	TP_printk("option %s=%s",
@@ -1657,7 +1661,7 @@ TRACE_EVENT(nfs_mount_option,
 	),
 
 	TP_fast_assign(
-		__assign_str(option, param->key);
+		__assign_str(option);
 	),
 
 	TP_printk("option %s", __get_str(option))
@@ -1675,7 +1679,7 @@ TRACE_EVENT(nfs_mount_path,
 	),
 
 	TP_fast_assign(
-		__assign_str(path, path);
+		__assign_str(path);
 	),
 
 	TP_printk("path='%s'", __get_str(path))
@@ -1710,9 +1714,8 @@ DECLARE_EVENT_CLASS(nfs_xdr_event,
 			__entry->xid = be32_to_cpu(rqstp->rq_xid);
 			__entry->version = task->tk_client->cl_vers;
 			__entry->error = error;
-			__assign_str(program,
-				     task->tk_client->cl_program->name);
-			__assign_str(procedure, task->tk_msg.rpc_proc->p_name);
+			__assign_str(program);
+			__assign_str(procedure);
 		),
 
 		TP_printk(SUNRPC_TRACE_TASK_SPECIFIER
diff --git a/fs/nfs/pagelist.c b/fs/nfs/pagelist.c
index 6efb5068c1..040b6b79c7 100644
--- a/fs/nfs/pagelist.c
+++ b/fs/nfs/pagelist.c
@@ -1545,6 +1545,11 @@ void nfs_pageio_cond_complete(struct nfs_pageio_descriptor *desc, pgoff_t index)
 					continue;
 			} else if (index == prev->wb_index + 1)
 				continue;
+			/*
+			 * We will submit more requests after these. Indicate
+			 * this to the underlying layers.
+			 */
+			desc->pg_moreio = 1;
 			nfs_pageio_complete(desc);
 			break;
 		}
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 0c0fed1ecd..b5834728f3 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -1999,6 +1999,14 @@ pnfs_update_layout(struct inode *ino,
 	}
 
 lookup_again:
+	if (!nfs4_valid_open_stateid(ctx->state)) {
+		trace_pnfs_update_layout(ino, pos, count,
+					 iomode, lo, lseg,
+					 PNFS_UPDATE_LAYOUT_INVALID_OPEN);
+		lseg = ERR_PTR(-EIO);
+		goto out;
+	}
+
 	lseg = ERR_PTR(nfs4_client_recover_expired_lease(clp));
 	if (IS_ERR(lseg))
 		goto out;
@@ -2697,38 +2705,28 @@ pnfs_layout_return_unused_byclid(struct nfs_client *clp,
 			&range);
 }
 
+/* Check if we have we have a valid layout but if there isn't an intersection
+ * between the request and the pgio->pg_lseg, put this pgio->pg_lseg away.
+ */
 void
-pnfs_generic_pg_check_layout(struct nfs_pageio_descriptor *pgio)
+pnfs_generic_pg_check_layout(struct nfs_pageio_descriptor *pgio,
+			     struct nfs_page *req)
 {
 	if (pgio->pg_lseg == NULL ||
-	    test_bit(NFS_LSEG_VALID, &pgio->pg_lseg->pls_flags))
+	    (test_bit(NFS_LSEG_VALID, &pgio->pg_lseg->pls_flags) &&
+	    pnfs_lseg_request_intersecting(pgio->pg_lseg, req)))
 		return;
 	pnfs_put_lseg(pgio->pg_lseg);
 	pgio->pg_lseg = NULL;
 }
 EXPORT_SYMBOL_GPL(pnfs_generic_pg_check_layout);
 
-/*
- * Check for any intersection between the request and the pgio->pg_lseg,
- * and if none, put this pgio->pg_lseg away.
- */
-void
-pnfs_generic_pg_check_range(struct nfs_pageio_descriptor *pgio, struct nfs_page *req)
-{
-	if (pgio->pg_lseg && !pnfs_lseg_request_intersecting(pgio->pg_lseg, req)) {
-		pnfs_put_lseg(pgio->pg_lseg);
-		pgio->pg_lseg = NULL;
-	}
-}
-EXPORT_SYMBOL_GPL(pnfs_generic_pg_check_range);
-
 void
 pnfs_generic_pg_init_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *req)
 {
 	u64 rd_size;
 
-	pnfs_generic_pg_check_layout(pgio);
-	pnfs_generic_pg_check_range(pgio, req);
+	pnfs_generic_pg_check_layout(pgio, req);
 	if (pgio->pg_lseg == NULL) {
 		if (pgio->pg_dreq == NULL)
 			rd_size = i_size_read(pgio->pg_inode) - req_offset(req);
@@ -2758,8 +2756,7 @@ void
 pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *pgio,
 			   struct nfs_page *req, u64 wb_size)
 {
-	pnfs_generic_pg_check_layout(pgio);
-	pnfs_generic_pg_check_range(pgio, req);
+	pnfs_generic_pg_check_layout(pgio, req);
 	if (pgio->pg_lseg == NULL) {
 		pgio->pg_lseg =
 			pnfs_update_layout(pgio->pg_inode, nfs_req_openctx(req),
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index db57a85500..fa5beeaaf5 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -257,8 +257,7 @@ void pnfs_put_lseg(struct pnfs_layout_segment *lseg);
 
 void set_pnfs_layoutdriver(struct nfs_server *, const struct nfs_fh *, struct nfs_fsinfo *);
 void unset_pnfs_layoutdriver(struct nfs_server *);
-void pnfs_generic_pg_check_layout(struct nfs_pageio_descriptor *pgio);
-void pnfs_generic_pg_check_range(struct nfs_pageio_descriptor *pgio, struct nfs_page *req);
+void pnfs_generic_pg_check_layout(struct nfs_pageio_descriptor *pgio, struct nfs_page *req);
 void pnfs_generic_pg_init_read(struct nfs_pageio_descriptor *, struct nfs_page *);
 int pnfs_generic_pg_readpages(struct nfs_pageio_descriptor *desc);
 void pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *pgio,
diff --git a/fs/nfs/proc.c b/fs/nfs/proc.c
index ad3a321ae9..d105e5b265 100644
--- a/fs/nfs/proc.c
+++ b/fs/nfs/proc.c
@@ -695,6 +695,7 @@ static int nfs_have_delegation(struct inode *inode, fmode_t flags)
 static const struct inode_operations nfs_dir_inode_operations = {
 	.create		= nfs_create,
 	.lookup		= nfs_lookup,
+	.atomic_open	= nfs_atomic_open_v23,
 	.link		= nfs_link,
 	.unlink		= nfs_unlink,
 	.symlink	= nfs_symlink,
diff --git a/fs/nfs/read.c b/fs/nfs/read.c
index a142287d86..88e6a78d37 100644
--- a/fs/nfs/read.c
+++ b/fs/nfs/read.c
@@ -332,13 +332,15 @@ out:
 int nfs_read_folio(struct file *file, struct folio *folio)
 {
 	struct inode *inode = file_inode(file);
+	loff_t pos = folio_pos(folio);
+	size_t len = folio_size(folio);
 	struct nfs_pageio_descriptor pgio;
 	struct nfs_open_context *ctx;
 	int ret;
 
-	trace_nfs_aop_readpage(inode, folio);
+	trace_nfs_aop_readpage(inode, pos, len);
 	nfs_inc_stats(inode, NFSIOS_VFSREADPAGE);
-	task_io_account_read(folio_size(folio));
+	task_io_account_read(len);
 
 	/*
 	 * Try to flush any pending writes to the file..
@@ -381,7 +383,7 @@ int nfs_read_folio(struct file *file, struct folio *folio)
 out_put:
 	put_nfs_open_context(ctx);
 out:
-	trace_nfs_aop_readpage_done(inode, folio, ret);
+	trace_nfs_aop_readpage_done(inode, pos, len, ret);
 	return ret;
 out_unlock:
 	folio_unlock(folio);
diff --git a/fs/nfs/super.c b/fs/nfs/super.c
index 075b31c93f..cbbd4866b0 100644
--- a/fs/nfs/super.c
+++ b/fs/nfs/super.c
@@ -516,8 +516,16 @@ static void nfs_show_mount_options(struct seq_file *m, struct nfs_server *nfss,
 	else
 		nfs_show_nfsv4_options(m, nfss, showdefaults);
 
-	if (nfss->options & NFS_OPTION_FSCACHE)
+	if (nfss->options & NFS_OPTION_FSCACHE) {
+#ifdef CONFIG_NFS_FSCACHE
+		if (nfss->fscache_uniq)
+			seq_printf(m, ",fsc=%s", nfss->fscache_uniq);
+		else
+			seq_puts(m, ",fsc");
+#else
 		seq_puts(m, ",fsc");
+#endif
+	}
 
 	if (nfss->options & NFS_OPTION_MIGRATION)
 		seq_puts(m, ",migration");
@@ -893,6 +901,16 @@ static struct nfs_server *nfs_try_mount_request(struct fs_context *fc)
 	rpc_authflavor_t authlist[NFS_MAX_SECFLAVORS];
 	unsigned int authlist_len = ARRAY_SIZE(authlist);
 
+	/* make sure 'nolock'/'lock' override the 'local_lock' mount option */
+	if (ctx->lock_status) {
+		if (ctx->lock_status == NFS_LOCK_NOLOCK) {
+			ctx->flags |= NFS_MOUNT_NONLM;
+			ctx->flags |= (NFS_MOUNT_LOCAL_FLOCK | NFS_MOUNT_LOCAL_FCNTL);
+		} else {
+			ctx->flags &= ~NFS_MOUNT_NONLM;
+			ctx->flags &= ~(NFS_MOUNT_LOCAL_FLOCK | NFS_MOUNT_LOCAL_FCNTL);
+		}
+	}
 	status = nfs_request_mount(fc, ctx->mntfh, authlist, &authlist_len);
 	if (status)
 		return ERR_PTR(status);
diff --git a/fs/nfs/symlink.c b/fs/nfs/symlink.c
index 0e27a2e4e6..13818129d2 100644
--- a/fs/nfs/symlink.c
+++ b/fs/nfs/symlink.c
@@ -41,7 +41,7 @@ static int nfs_symlink_filler(struct file *file, struct folio *folio)
 error:
 	folio_set_error(folio);
 	folio_unlock(folio);
-	return -EIO;
+	return error;
 }
 
 static const char *nfs_get_link(struct dentry *dentry,
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index 5d9dc6c053..75c1b3c7fa 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -667,10 +667,6 @@ static int nfs_writepage_locked(struct folio *folio,
 	struct inode *inode = folio_file_mapping(folio)->host;
 	int err;
 
-	if (wbc->sync_mode == WB_SYNC_NONE &&
-	    NFS_SERVER(inode)->write_congested)
-		return AOP_WRITEPAGE_ACTIVATE;
-
 	nfs_inc_stats(inode, NFSIOS_VFSWRITEPAGE);
 	nfs_pageio_init_write(&pgio, inode, 0, false,
 			      &nfs_async_write_completion_ops);
@@ -1301,7 +1297,7 @@ static bool
 is_whole_file_wrlock(struct file_lock *fl)
 {
 	return fl->fl_start == 0 && fl->fl_end == OFFSET_MAX &&
-			fl->fl_type == F_WRLCK;
+			lock_is_write(fl);
 }
 
 /* If we know the page is up to date, and we're not using byte range locks (or
@@ -1335,13 +1331,13 @@ static int nfs_can_extend_write(struct file *file, struct folio *folio,
 	spin_lock(&flctx->flc_lock);
 	if (!list_empty(&flctx->flc_posix)) {
 		fl = list_first_entry(&flctx->flc_posix, struct file_lock,
-					fl_list);
+					c.flc_list);
 		if (is_whole_file_wrlock(fl))
 			ret = 1;
 	} else if (!list_empty(&flctx->flc_flock)) {
 		fl = list_first_entry(&flctx->flc_flock, struct file_lock,
-					fl_list);
-		if (fl->fl_type == F_WRLCK)
+					c.flc_list);
+		if (lock_is_write(fl))
 			ret = 1;
 	}
 	spin_unlock(&flctx->flc_lock);
@@ -2077,17 +2073,17 @@ int nfs_wb_folio_cancel(struct inode *inode, struct folio *folio)
  */
 int nfs_wb_folio(struct inode *inode, struct folio *folio)
 {
-	loff_t range_start = folio_file_pos(folio);
-	loff_t range_end = range_start + (loff_t)folio_size(folio) - 1;
+	loff_t range_start = folio_pos(folio);
+	size_t len = folio_size(folio);
 	struct writeback_control wbc = {
 		.sync_mode = WB_SYNC_ALL,
 		.nr_to_write = 0,
 		.range_start = range_start,
-		.range_end = range_end,
+		.range_end = range_start + len - 1,
 	};
 	int ret;
 
-	trace_nfs_writeback_folio(inode, folio);
+	trace_nfs_writeback_folio(inode, range_start, len);
 
 	for (;;) {
 		folio_wait_writeback(folio);
@@ -2105,7 +2101,7 @@ int nfs_wb_folio(struct inode *inode, struct folio *folio)
 			goto out_error;
 	}
 out_error:
-	trace_nfs_writeback_folio_done(inode, folio, ret);
+	trace_nfs_writeback_folio_done(inode, range_start, len, ret);
 	return ret;
 }
 
@@ -2124,10 +2120,10 @@ int nfs_migrate_folio(struct address_space *mapping, struct folio *dst,
 	if (folio_test_private(src))
 		return -EBUSY;
 
-	if (folio_test_fscache(src)) {
+	if (folio_test_private_2(src)) { /* [DEPRECATED] */
 		if (mode == MIGRATE_ASYNC)
 			return -EBUSY;
-		folio_wait_fscache(src);
+		folio_wait_private_2(src);
 	}
 
 	return migrate_folio(mapping, dst, src, mode);
diff --git a/fs/nfsd/Kconfig b/fs/nfsd/Kconfig
index 272ab8d5c4..ec2ab6429e 100644
--- a/fs/nfsd/Kconfig
+++ b/fs/nfsd/Kconfig
@@ -162,7 +162,7 @@ config NFSD_V4_SECURITY_LABEL
 config NFSD_LEGACY_CLIENT_TRACKING
 	bool "Support legacy NFSv4 client tracking methods (DEPRECATED)"
 	depends on NFSD_V4
-	default n
+	default y
 	help
 	  The NFSv4 server needs to store a small amount of information on
 	  stable storage in order to handle state recovery after reboot. Most
diff --git a/fs/nfsd/blocklayout.c b/fs/nfsd/blocklayout.c
index 46fd74d91e..3c040c81c7 100644
--- a/fs/nfsd/blocklayout.c
+++ b/fs/nfsd/blocklayout.c
@@ -328,10 +328,10 @@ nfsd4_scsi_proc_layoutcommit(struct inode *inode,
 }
 
 static void
-nfsd4_scsi_fence_client(struct nfs4_layout_stateid *ls)
+nfsd4_scsi_fence_client(struct nfs4_layout_stateid *ls, struct nfsd_file *file)
 {
 	struct nfs4_client *clp = ls->ls_stid.sc_client;
-	struct block_device *bdev = ls->ls_file->nf_file->f_path.mnt->mnt_sb->s_bdev;
+	struct block_device *bdev = file->nf_file->f_path.mnt->mnt_sb->s_bdev;
 
 	bdev->bd_disk->fops->pr_ops->pr_preempt(bdev, NFSD_MDS_PR_KEY,
 			nfsd4_scsi_pr_key(clp), 0, true);
diff --git a/fs/nfsd/export.c b/fs/nfsd/export.c
index 7b641095a6..50b3135d07 100644
--- a/fs/nfsd/export.c
+++ b/fs/nfsd/export.c
@@ -334,21 +334,25 @@ static void nfsd4_fslocs_free(struct nfsd4_fs_locations *fsloc)
 static int export_stats_init(struct export_stats *stats)
 {
 	stats->start_time = ktime_get_seconds();
-	return nfsd_percpu_counters_init(stats->counter, EXP_STATS_COUNTERS_NUM);
+	return percpu_counter_init_many(stats->counter, 0, GFP_KERNEL,
+					EXP_STATS_COUNTERS_NUM);
 }
 
 static void export_stats_reset(struct export_stats *stats)
 {
-	if (stats)
-		nfsd_percpu_counters_reset(stats->counter,
-					   EXP_STATS_COUNTERS_NUM);
+	if (stats) {
+		int i;
+
+		for (i = 0; i < EXP_STATS_COUNTERS_NUM; i++)
+			percpu_counter_set(&stats->counter[i], 0);
+	}
 }
 
 static void export_stats_destroy(struct export_stats *stats)
 {
 	if (stats)
-		nfsd_percpu_counters_destroy(stats->counter,
-					     EXP_STATS_COUNTERS_NUM);
+		percpu_counter_destroy_many(stats->counter,
+					    EXP_STATS_COUNTERS_NUM);
 }
 
 static void svc_export_put(struct kref *ref)
diff --git a/fs/nfsd/filecache.c b/fs/nfsd/filecache.c
index 9cb7f0c33d..f4704f5d40 100644
--- a/fs/nfsd/filecache.c
+++ b/fs/nfsd/filecache.c
@@ -61,13 +61,10 @@ static DEFINE_PER_CPU(unsigned long, nfsd_file_total_age);
 static DEFINE_PER_CPU(unsigned long, nfsd_file_evictions);
 
 struct nfsd_fcache_disposal {
-	struct work_struct work;
 	spinlock_t lock;
 	struct list_head freeme;
 };
 
-static struct workqueue_struct *nfsd_filecache_wq __read_mostly;
-
 static struct kmem_cache		*nfsd_file_slab;
 static struct kmem_cache		*nfsd_file_mark_slab;
 static struct list_lru			nfsd_file_lru;
@@ -162,8 +159,8 @@ nfsd_file_mark_find_or_create(struct nfsd_file *nf, struct inode *inode)
 
 	do {
 		fsnotify_group_lock(nfsd_file_fsnotify_group);
-		mark = fsnotify_find_mark(&inode->i_fsnotify_marks,
-					  nfsd_file_fsnotify_group);
+		mark = fsnotify_find_inode_mark(inode,
+						nfsd_file_fsnotify_group);
 		if (mark) {
 			nfm = nfsd_file_mark_get(container_of(mark,
 						 struct nfsd_file_mark,
@@ -283,7 +280,7 @@ nfsd_file_free(struct nfsd_file *nf)
 		nfsd_file_mark_put(nf->nf_mark);
 	if (nf->nf_file) {
 		nfsd_file_check_write_error(nf);
-		filp_close(nf->nf_file, NULL);
+		nfsd_filp_close(nf->nf_file);
 	}
 
 	/*
@@ -421,7 +418,37 @@ nfsd_file_dispose_list_delayed(struct list_head *dispose)
 		spin_lock(&l->lock);
 		list_move_tail(&nf->nf_lru, &l->freeme);
 		spin_unlock(&l->lock);
-		queue_work(nfsd_filecache_wq, &l->work);
+		svc_wake_up(nn->nfsd_serv);
+	}
+}
+
+/**
+ * nfsd_file_net_dispose - deal with nfsd_files waiting to be disposed.
+ * @nn: nfsd_net in which to find files to be disposed.
+ *
+ * When files held open for nfsv3 are removed from the filecache, whether
+ * due to memory pressure or garbage collection, they are queued to
+ * a per-net-ns queue.  This function completes the disposal, either
+ * directly or by waking another nfsd thread to help with the work.
+ */
+void nfsd_file_net_dispose(struct nfsd_net *nn)
+{
+	struct nfsd_fcache_disposal *l = nn->fcache_disposal;
+
+	if (!list_empty(&l->freeme)) {
+		LIST_HEAD(dispose);
+		int i;
+
+		spin_lock(&l->lock);
+		for (i = 0; i < 8 && !list_empty(&l->freeme); i++)
+			list_move(l->freeme.next, &dispose);
+		spin_unlock(&l->lock);
+		if (!list_empty(&l->freeme))
+			/* Wake up another thread to share the work
+			 * *before* doing any actual disposing.
+			 */
+			svc_wake_up(nn->nfsd_serv);
+		nfsd_file_dispose_list(&dispose);
 	}
 }
 
@@ -631,39 +658,17 @@ nfsd_file_close_inode_sync(struct inode *inode)
 		list_del_init(&nf->nf_lru);
 		nfsd_file_free(nf);
 	}
-	flush_delayed_fput();
-}
-
-/**
- * nfsd_file_delayed_close - close unused nfsd_files
- * @work: dummy
- *
- * Scrape the freeme list for this nfsd_net, and then dispose of them
- * all.
- */
-static void
-nfsd_file_delayed_close(struct work_struct *work)
-{
-	LIST_HEAD(head);
-	struct nfsd_fcache_disposal *l = container_of(work,
-			struct nfsd_fcache_disposal, work);
-
-	spin_lock(&l->lock);
-	list_splice_init(&l->freeme, &head);
-	spin_unlock(&l->lock);
-
-	nfsd_file_dispose_list(&head);
 }
 
 static int
 nfsd_file_lease_notifier_call(struct notifier_block *nb, unsigned long arg,
 			    void *data)
 {
-	struct file_lock *fl = data;
+	struct file_lease *fl = data;
 
 	/* Only close files for F_SETLEASE leases */
-	if (fl->fl_flags & FL_LEASE)
-		nfsd_file_close_inode(file_inode(fl->fl_file));
+	if (fl->c.flc_flags & FL_LEASE)
+		nfsd_file_close_inode(file_inode(fl->c.flc_file));
 	return 0;
 }
 
@@ -717,25 +722,18 @@ nfsd_file_cache_init(void)
 		return ret;
 
 	ret = -ENOMEM;
-	nfsd_filecache_wq = alloc_workqueue("nfsd_filecache", WQ_UNBOUND, 0);
-	if (!nfsd_filecache_wq)
-		goto out;
-
-	nfsd_file_slab = kmem_cache_create("nfsd_file",
-				sizeof(struct nfsd_file), 0, 0, NULL);
+	nfsd_file_slab = KMEM_CACHE(nfsd_file, 0);
 	if (!nfsd_file_slab) {
 		pr_err("nfsd: unable to create nfsd_file_slab\n");
 		goto out_err;
 	}
 
-	nfsd_file_mark_slab = kmem_cache_create("nfsd_file_mark",
-					sizeof(struct nfsd_file_mark), 0, 0, NULL);
+	nfsd_file_mark_slab = KMEM_CACHE(nfsd_file_mark, 0);
 	if (!nfsd_file_mark_slab) {
 		pr_err("nfsd: unable to create nfsd_file_mark_slab\n");
 		goto out_err;
 	}
 
-
 	ret = list_lru_init(&nfsd_file_lru);
 	if (ret) {
 		pr_err("nfsd: failed to init nfsd_file_lru: %d\n", ret);
@@ -785,8 +783,6 @@ out_err:
 	nfsd_file_slab = NULL;
 	kmem_cache_destroy(nfsd_file_mark_slab);
 	nfsd_file_mark_slab = NULL;
-	destroy_workqueue(nfsd_filecache_wq);
-	nfsd_filecache_wq = NULL;
 	rhltable_destroy(&nfsd_file_rhltable);
 	goto out;
 }
@@ -832,7 +828,6 @@ nfsd_alloc_fcache_disposal(void)
 	l = kmalloc(sizeof(*l), GFP_KERNEL);
 	if (!l)
 		return NULL;
-	INIT_WORK(&l->work, nfsd_file_delayed_close);
 	spin_lock_init(&l->lock);
 	INIT_LIST_HEAD(&l->freeme);
 	return l;
@@ -841,7 +836,6 @@ nfsd_alloc_fcache_disposal(void)
 static void
 nfsd_free_fcache_disposal(struct nfsd_fcache_disposal *l)
 {
-	cancel_work_sync(&l->work);
 	nfsd_file_dispose_list(&l->freeme);
 	kfree(l);
 }
@@ -910,8 +904,6 @@ nfsd_file_cache_shutdown(void)
 	fsnotify_wait_marks_destroyed();
 	kmem_cache_destroy(nfsd_file_mark_slab);
 	nfsd_file_mark_slab = NULL;
-	destroy_workqueue(nfsd_filecache_wq);
-	nfsd_filecache_wq = NULL;
 	rhltable_destroy(&nfsd_file_rhltable);
 
 	for_each_possible_cpu(i) {
diff --git a/fs/nfsd/filecache.h b/fs/nfsd/filecache.h
index e54165a322..c61884def9 100644
--- a/fs/nfsd/filecache.h
+++ b/fs/nfsd/filecache.h
@@ -56,6 +56,7 @@ void nfsd_file_cache_shutdown_net(struct net *net);
 void nfsd_file_put(struct nfsd_file *nf);
 struct nfsd_file *nfsd_file_get(struct nfsd_file *nf);
 void nfsd_file_close_inode_sync(struct inode *inode);
+void nfsd_file_net_dispose(struct nfsd_net *nn);
 bool nfsd_file_is_cached(struct inode *inode);
 __be32 nfsd_file_acquire_gc(struct svc_rqst *rqstp, struct svc_fh *fhp,
 		  unsigned int may_flags, struct nfsd_file **nfp);
diff --git a/fs/nfsd/netlink.c b/fs/nfsd/netlink.c
index 0e1d635ec5..529a75ecf2 100644
--- a/fs/nfsd/netlink.c
+++ b/fs/nfsd/netlink.c
@@ -10,15 +10,79 @@
 
 #include <uapi/linux/nfsd_netlink.h>
 
+/* Common nested types */
+const struct nla_policy nfsd_sock_nl_policy[NFSD_A_SOCK_TRANSPORT_NAME + 1] = {
+	[NFSD_A_SOCK_ADDR] = { .type = NLA_BINARY, },
+	[NFSD_A_SOCK_TRANSPORT_NAME] = { .type = NLA_NUL_STRING, },
+};
+
+const struct nla_policy nfsd_version_nl_policy[NFSD_A_VERSION_ENABLED + 1] = {
+	[NFSD_A_VERSION_MAJOR] = { .type = NLA_U32, },
+	[NFSD_A_VERSION_MINOR] = { .type = NLA_U32, },
+	[NFSD_A_VERSION_ENABLED] = { .type = NLA_FLAG, },
+};
+
+/* NFSD_CMD_THREADS_SET - do */
+static const struct nla_policy nfsd_threads_set_nl_policy[NFSD_A_SERVER_SCOPE + 1] = {
+	[NFSD_A_SERVER_THREADS] = { .type = NLA_U32, },
+	[NFSD_A_SERVER_GRACETIME] = { .type = NLA_U32, },
+	[NFSD_A_SERVER_LEASETIME] = { .type = NLA_U32, },
+	[NFSD_A_SERVER_SCOPE] = { .type = NLA_NUL_STRING, },
+};
+
+/* NFSD_CMD_VERSION_SET - do */
+static const struct nla_policy nfsd_version_set_nl_policy[NFSD_A_SERVER_PROTO_VERSION + 1] = {
+	[NFSD_A_SERVER_PROTO_VERSION] = NLA_POLICY_NESTED(nfsd_version_nl_policy),
+};
+
+/* NFSD_CMD_LISTENER_SET - do */
+static const struct nla_policy nfsd_listener_set_nl_policy[NFSD_A_SERVER_SOCK_ADDR + 1] = {
+	[NFSD_A_SERVER_SOCK_ADDR] = NLA_POLICY_NESTED(nfsd_sock_nl_policy),
+};
+
 /* Ops table for nfsd */
 static const struct genl_split_ops nfsd_nl_ops[] = {
 	{
 		.cmd	= NFSD_CMD_RPC_STATUS_GET,
-		.start	= nfsd_nl_rpc_status_get_start,
 		.dumpit	= nfsd_nl_rpc_status_get_dumpit,
-		.done	= nfsd_nl_rpc_status_get_done,
 		.flags	= GENL_CMD_CAP_DUMP,
 	},
+	{
+		.cmd		= NFSD_CMD_THREADS_SET,
+		.doit		= nfsd_nl_threads_set_doit,
+		.policy		= nfsd_threads_set_nl_policy,
+		.maxattr	= NFSD_A_SERVER_SCOPE,
+		.flags		= GENL_ADMIN_PERM | GENL_CMD_CAP_DO,
+	},
+	{
+		.cmd	= NFSD_CMD_THREADS_GET,
+		.doit	= nfsd_nl_threads_get_doit,
+		.flags	= GENL_CMD_CAP_DO,
+	},
+	{
+		.cmd		= NFSD_CMD_VERSION_SET,
+		.doit		= nfsd_nl_version_set_doit,
+		.policy		= nfsd_version_set_nl_policy,
+		.maxattr	= NFSD_A_SERVER_PROTO_VERSION,
+		.flags		= GENL_ADMIN_PERM | GENL_CMD_CAP_DO,
+	},
+	{
+		.cmd	= NFSD_CMD_VERSION_GET,
+		.doit	= nfsd_nl_version_get_doit,
+		.flags	= GENL_CMD_CAP_DO,
+	},
+	{
+		.cmd		= NFSD_CMD_LISTENER_SET,
+		.doit		= nfsd_nl_listener_set_doit,
+		.policy		= nfsd_listener_set_nl_policy,
+		.maxattr	= NFSD_A_SERVER_SOCK_ADDR,
+		.flags		= GENL_ADMIN_PERM | GENL_CMD_CAP_DO,
+	},
+	{
+		.cmd	= NFSD_CMD_LISTENER_GET,
+		.doit	= nfsd_nl_listener_get_doit,
+		.flags	= GENL_CMD_CAP_DO,
+	},
 };
 
 struct genl_family nfsd_nl_family __ro_after_init = {
diff --git a/fs/nfsd/netlink.h b/fs/nfsd/netlink.h
index d83dd6bdee..2e132ef328 100644
--- a/fs/nfsd/netlink.h
+++ b/fs/nfsd/netlink.h
@@ -11,11 +11,18 @@
 
 #include <uapi/linux/nfsd_netlink.h>
 
-int nfsd_nl_rpc_status_get_start(struct netlink_callback *cb);
-int nfsd_nl_rpc_status_get_done(struct netlink_callback *cb);
+/* Common nested types */
+extern const struct nla_policy nfsd_sock_nl_policy[NFSD_A_SOCK_TRANSPORT_NAME + 1];
+extern const struct nla_policy nfsd_version_nl_policy[NFSD_A_VERSION_ENABLED + 1];
 
 int nfsd_nl_rpc_status_get_dumpit(struct sk_buff *skb,
 				  struct netlink_callback *cb);
+int nfsd_nl_threads_set_doit(struct sk_buff *skb, struct genl_info *info);
+int nfsd_nl_threads_get_doit(struct sk_buff *skb, struct genl_info *info);
+int nfsd_nl_version_set_doit(struct sk_buff *skb, struct genl_info *info);
+int nfsd_nl_version_get_doit(struct sk_buff *skb, struct genl_info *info);
+int nfsd_nl_listener_set_doit(struct sk_buff *skb, struct genl_info *info);
+int nfsd_nl_listener_get_doit(struct sk_buff *skb, struct genl_info *info);
 
 extern struct genl_family nfsd_nl_family;
 
diff --git a/fs/nfsd/netns.h b/fs/nfsd/netns.h
index 0cef4bb407..14ec156563 100644
--- a/fs/nfsd/netns.h
+++ b/fs/nfsd/netns.h
@@ -14,6 +14,7 @@
 #include <linux/nfs4.h>
 #include <linux/percpu_counter.h>
 #include <linux/siphash.h>
+#include <linux/sunrpc/stats.h>
 
 /* Hash tables for nfs4_clientid state */
 #define CLIENT_HASH_BITS                 4
@@ -179,6 +180,9 @@ struct nfsd_net {
 	/* Per-netns stats counters */
 	struct percpu_counter    counter[NFSD_STATS_COUNTERS_NUM];
 
+	/* sunrpc svc stats */
+	struct svc_stat          nfsd_svcstats;
+
 	/* longest hash chain seen */
 	unsigned int             longest_chain;
 
@@ -205,11 +209,16 @@ struct nfsd_net {
 	atomic_t		nfsd_courtesy_clients;
 	struct shrinker		*nfsd_client_shrinker;
 	struct work_struct	nfsd_shrinker_work;
+
+	/* last time an admin-revoke happened for NFSv4.0 */
+	time64_t		nfs40_last_revoke;
+
 };
 
 /* Simple check to find out if a given net was properly initialized */
 #define nfsd_netns_ready(nn) ((nn)->sessionid_hashtbl)
 
+extern bool nfsd_support_version(int vers);
 extern void nfsd_netns_free_versions(struct nfsd_net *nn);
 
 extern unsigned int nfsd_net_id;
diff --git a/fs/nfsd/nfs3proc.c b/fs/nfsd/nfs3proc.c
index b78eceebd9..dfcc957e46 100644
--- a/fs/nfsd/nfs3proc.c
+++ b/fs/nfsd/nfs3proc.c
@@ -71,13 +71,15 @@ nfsd3_proc_setattr(struct svc_rqst *rqstp)
 	struct nfsd_attrs attrs = {
 		.na_iattr	= &argp->attrs,
 	};
+	const struct timespec64 *guardtime = NULL;
 
 	dprintk("nfsd: SETATTR(3)  %s\n",
 				SVCFH_fmt(&argp->fh));
 
 	fh_copy(&resp->fh, &argp->fh);
-	resp->status = nfsd_setattr(rqstp, &resp->fh, &attrs,
-				    argp->check_guard, argp->guardtime);
+	if (argp->check_guard)
+		guardtime = &argp->guardtime;
+	resp->status = nfsd_setattr(rqstp, &resp->fh, &attrs, guardtime);
 	return rpc_success;
 }
 
diff --git a/fs/nfsd/nfs3xdr.c b/fs/nfsd/nfs3xdr.c
index f32128955e..a7a07470c1 100644
--- a/fs/nfsd/nfs3xdr.c
+++ b/fs/nfsd/nfs3xdr.c
@@ -295,17 +295,14 @@ svcxdr_decode_sattr3(struct svc_rqst *rqstp, struct xdr_stream *xdr,
 static bool
 svcxdr_decode_sattrguard3(struct xdr_stream *xdr, struct nfsd3_sattrargs *args)
 {
-	__be32 *p;
 	u32 check;
 
 	if (xdr_stream_decode_bool(xdr, &check) < 0)
 		return false;
 	if (check) {
-		p = xdr_inline_decode(xdr, XDR_UNIT * 2);
-		if (!p)
+		if (!svcxdr_decode_nfstime3(xdr, &args->guardtime))
 			return false;
 		args->check_guard = 1;
-		args->guardtime = be32_to_cpup(p);
 	} else
 		args->check_guard = 0;
 
diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c
index 30aa241038..d756f443fc 100644
--- a/fs/nfsd/nfs4callback.c
+++ b/fs/nfsd/nfs4callback.c
@@ -45,7 +45,7 @@
 
 #define NFSDDBG_FACILITY                NFSDDBG_PROC
 
-static void nfsd4_mark_cb_fault(struct nfs4_client *, int reason);
+static void nfsd4_mark_cb_fault(struct nfs4_client *clp);
 
 #define NFSPROC4_CB_NULL 0
 #define NFSPROC4_CB_COMPOUND 1
@@ -768,7 +768,7 @@ static void nfs4_xdr_enc_cb_notify_lock(struct rpc_rqst *req,
 	const struct nfsd4_callback *cb = data;
 	const struct nfsd4_blocked_lock *nbl =
 		container_of(cb, struct nfsd4_blocked_lock, nbl_cb);
-	struct nfs4_lockowner *lo = (struct nfs4_lockowner *)nbl->nbl_lock.fl_owner;
+	struct nfs4_lockowner *lo = (struct nfs4_lockowner *)nbl->nbl_lock.c.flc_owner;
 	struct nfs4_cb_compound_hdr hdr = {
 		.ident = 0,
 		.minorversion = cb->cb_clp->cl_minorversion,
@@ -978,11 +978,12 @@ static int max_cb_time(struct net *net)
 	return max(((u32)nn->nfsd4_lease)/10, 1u) * HZ;
 }
 
-static struct workqueue_struct *callback_wq;
-
 static bool nfsd4_queue_cb(struct nfsd4_callback *cb)
 {
-	return queue_work(callback_wq, &cb->cb_work);
+	struct nfs4_client *clp = cb->cb_clp;
+
+	trace_nfsd_cb_queue(clp, cb);
+	return queue_work(clp->cl_callback_wq, &cb->cb_work);
 }
 
 static void nfsd41_cb_inflight_begin(struct nfs4_client *clp)
@@ -1094,18 +1095,18 @@ static void nfsd4_mark_cb_state(struct nfs4_client *clp, int newstate)
 {
 	if (clp->cl_cb_state != newstate) {
 		clp->cl_cb_state = newstate;
-		trace_nfsd_cb_state(clp);
+		trace_nfsd_cb_new_state(clp);
 	}
 }
 
-static void nfsd4_mark_cb_down(struct nfs4_client *clp, int reason)
+static void nfsd4_mark_cb_down(struct nfs4_client *clp)
 {
 	if (test_bit(NFSD4_CLIENT_CB_UPDATE, &clp->cl_flags))
 		return;
 	nfsd4_mark_cb_state(clp, NFSD4_CB_DOWN);
 }
 
-static void nfsd4_mark_cb_fault(struct nfs4_client *clp, int reason)
+static void nfsd4_mark_cb_fault(struct nfs4_client *clp)
 {
 	if (test_bit(NFSD4_CLIENT_CB_UPDATE, &clp->cl_flags))
 		return;
@@ -1117,7 +1118,7 @@ static void nfsd4_cb_probe_done(struct rpc_task *task, void *calldata)
 	struct nfs4_client *clp = container_of(calldata, struct nfs4_client, cl_cb_null);
 
 	if (task->tk_status)
-		nfsd4_mark_cb_down(clp, task->tk_status);
+		nfsd4_mark_cb_down(clp);
 	else
 		nfsd4_mark_cb_state(clp, NFSD4_CB_UP);
 }
@@ -1152,7 +1153,7 @@ void nfsd4_probe_callback(struct nfs4_client *clp)
 void nfsd4_probe_callback_sync(struct nfs4_client *clp)
 {
 	nfsd4_probe_callback(clp);
-	flush_workqueue(callback_wq);
+	flush_workqueue(clp->cl_callback_wq);
 }
 
 void nfsd4_change_callback(struct nfs4_client *clp, struct nfs4_cb_conn *conn)
@@ -1201,6 +1202,7 @@ static void nfsd41_destroy_cb(struct nfsd4_callback *cb)
 {
 	struct nfs4_client *clp = cb->cb_clp;
 
+	trace_nfsd_cb_destroy(clp, cb);
 	nfsd41_cb_release_slot(cb);
 	if (cb->cb_ops && cb->cb_ops->release)
 		cb->cb_ops->release(cb);
@@ -1253,6 +1255,8 @@ static bool nfsd4_cb_sequence_done(struct rpc_task *task, struct nfsd4_callback
 	if (!cb->cb_holds_slot)
 		goto need_restart;
 
+	/* This is the operation status code for CB_SEQUENCE */
+	trace_nfsd_cb_seq_status(task, cb);
 	switch (cb->cb_seq_status) {
 	case 0:
 		/*
@@ -1266,13 +1270,23 @@ static bool nfsd4_cb_sequence_done(struct rpc_task *task, struct nfsd4_callback
 		break;
 	case -ESERVERFAULT:
 		++session->se_cb_seq_nr;
-		fallthrough;
+		nfsd4_mark_cb_fault(cb->cb_clp);
+		ret = false;
+		break;
 	case 1:
+		/*
+		 * cb_seq_status remains 1 if an RPC Reply was never
+		 * received. NFSD can't know if the client processed
+		 * the CB_SEQUENCE operation. Ask the client to send a
+		 * DESTROY_SESSION to recover.
+		 */
+		fallthrough;
 	case -NFS4ERR_BADSESSION:
-		nfsd4_mark_cb_fault(cb->cb_clp, cb->cb_seq_status);
+		nfsd4_mark_cb_fault(cb->cb_clp);
 		ret = false;
-		break;
+		goto need_restart;
 	case -NFS4ERR_DELAY:
+		cb->cb_seq_status = 1;
 		if (!rpc_restart_call(task))
 			goto out;
 
@@ -1287,14 +1301,11 @@ static bool nfsd4_cb_sequence_done(struct rpc_task *task, struct nfsd4_callback
 		}
 		break;
 	default:
-		nfsd4_mark_cb_fault(cb->cb_clp, cb->cb_seq_status);
-		dprintk("%s: unprocessed error %d\n", __func__,
-			cb->cb_seq_status);
+		nfsd4_mark_cb_fault(cb->cb_clp);
 	}
-
 	nfsd41_cb_release_slot(cb);
-	dprintk("%s: freed slot, new seqid=%d\n", __func__,
-		clp->cl_cb_session->se_cb_seq_nr);
+
+	trace_nfsd_cb_free_slot(task, cb);
 
 	if (RPC_SIGNALLED(task))
 		goto need_restart;
@@ -1306,6 +1317,7 @@ retry_nowait:
 	goto out;
 need_restart:
 	if (!test_bit(NFSD4_CLIENT_CB_KILL, &clp->cl_flags)) {
+		trace_nfsd_cb_restart(clp, cb);
 		task->tk_status = 0;
 		cb->cb_need_restart = true;
 	}
@@ -1335,7 +1347,7 @@ static void nfsd4_cb_done(struct rpc_task *task, void *calldata)
 		case -EIO:
 		case -ETIMEDOUT:
 		case -EACCES:
-			nfsd4_mark_cb_down(clp, task->tk_status);
+			nfsd4_mark_cb_down(clp);
 		}
 		break;
 	default:
@@ -1360,19 +1372,6 @@ static const struct rpc_call_ops nfsd4_cb_ops = {
 	.rpc_release = nfsd4_cb_release,
 };
 
-int nfsd4_create_callback_queue(void)
-{
-	callback_wq = alloc_ordered_workqueue("nfsd4_callbacks", 0);
-	if (!callback_wq)
-		return -ENOMEM;
-	return 0;
-}
-
-void nfsd4_destroy_callback_queue(void)
-{
-	destroy_workqueue(callback_wq);
-}
-
 /* must be called under the state lock */
 void nfsd4_shutdown_callback(struct nfs4_client *clp)
 {
@@ -1386,16 +1385,17 @@ void nfsd4_shutdown_callback(struct nfs4_client *clp)
 	 * client, destroy the rpc client, and stop:
 	 */
 	nfsd4_run_cb(&clp->cl_cb_null);
-	flush_workqueue(callback_wq);
+	flush_workqueue(clp->cl_callback_wq);
 	nfsd41_cb_inflight_wait_complete(clp);
 }
 
-/* requires cl_lock: */
 static struct nfsd4_conn * __nfsd4_find_backchannel(struct nfs4_client *clp)
 {
 	struct nfsd4_session *s;
 	struct nfsd4_conn *c;
 
+	lockdep_assert_held(&clp->cl_lock);
+
 	list_for_each_entry(s, &clp->cl_sessions, se_perclnt) {
 		list_for_each_entry(c, &s->se_conns, cn_persession) {
 			if (c->cn_flags & NFS4_CDFC4_BACK)
@@ -1407,9 +1407,9 @@ static struct nfsd4_conn * __nfsd4_find_backchannel(struct nfs4_client *clp)
 
 /*
  * Note there isn't a lot of locking in this code; instead we depend on
- * the fact that it is run from the callback_wq, which won't run two
- * work items at once.  So, for example, callback_wq handles all access
- * of cl_cb_client and all calls to rpc_create or rpc_shutdown_client.
+ * the fact that it is run from clp->cl_callback_wq, which won't run two
+ * work items at once.  So, for example, clp->cl_callback_wq handles all
+ * access of cl_cb_client and all calls to rpc_create or rpc_shutdown_client.
  */
 static void nfsd4_process_cb_update(struct nfsd4_callback *cb)
 {
@@ -1419,11 +1419,14 @@ static void nfsd4_process_cb_update(struct nfsd4_callback *cb)
 	struct nfsd4_conn *c;
 	int err;
 
+	trace_nfsd_cb_bc_update(clp, cb);
+
 	/*
 	 * This is either an update, or the client dying; in either case,
 	 * kill the old client:
 	 */
 	if (clp->cl_cb_client) {
+		trace_nfsd_cb_bc_shutdown(clp, cb);
 		rpc_shutdown_client(clp->cl_cb_client);
 		clp->cl_cb_client = NULL;
 		put_cred(clp->cl_cb_cred);
@@ -1435,13 +1438,15 @@ static void nfsd4_process_cb_update(struct nfsd4_callback *cb)
 	}
 	if (test_bit(NFSD4_CLIENT_CB_KILL, &clp->cl_flags))
 		return;
+
 	spin_lock(&clp->cl_lock);
 	/*
 	 * Only serialized callback code is allowed to clear these
 	 * flags; main nfsd code can only set them:
 	 */
-	BUG_ON(!(clp->cl_flags & NFSD4_CLIENT_CB_FLAG_MASK));
+	WARN_ON(!(clp->cl_flags & NFSD4_CLIENT_CB_FLAG_MASK));
 	clear_bit(NFSD4_CLIENT_CB_UPDATE, &clp->cl_flags);
+
 	memcpy(&conn, &cb->cb_clp->cl_cb_conn, sizeof(struct nfs4_cb_conn));
 	c = __nfsd4_find_backchannel(clp);
 	if (c) {
@@ -1453,7 +1458,7 @@ static void nfsd4_process_cb_update(struct nfsd4_callback *cb)
 
 	err = setup_callback_client(clp, &conn, ses);
 	if (err) {
-		nfsd4_mark_cb_down(clp, err);
+		nfsd4_mark_cb_down(clp);
 		if (c)
 			svc_xprt_put(c->cn_xprt);
 		return;
@@ -1469,12 +1474,7 @@ nfsd4_run_cb_work(struct work_struct *work)
 	struct rpc_clnt *clnt;
 	int flags;
 
-	if (cb->cb_need_restart) {
-		cb->cb_need_restart = false;
-	} else {
-		if (cb->cb_ops && cb->cb_ops->prepare)
-			cb->cb_ops->prepare(cb);
-	}
+	trace_nfsd_cb_start(clp);
 
 	if (clp->cl_flags & NFSD4_CLIENT_CB_FLAG_MASK)
 		nfsd4_process_cb_update(cb);
@@ -1495,6 +1495,12 @@ nfsd4_run_cb_work(struct work_struct *work)
 		return;
 	}
 
+	if (cb->cb_need_restart) {
+		cb->cb_need_restart = false;
+	} else {
+		if (cb->cb_ops && cb->cb_ops->prepare)
+			cb->cb_ops->prepare(cb);
+	}
 	cb->cb_msg.rpc_cred = clp->cl_cb_cred;
 	flags = clp->cl_minorversion ? RPC_TASK_NOCONNECT : RPC_TASK_SOFTCONN;
 	rpc_call_async(clnt, &cb->cb_msg, RPC_TASK_SOFT | flags,
@@ -1510,7 +1516,6 @@ void nfsd4_init_cb(struct nfsd4_callback *cb, struct nfs4_client *clp,
 	cb->cb_msg.rpc_resp = cb;
 	cb->cb_ops = ops;
 	INIT_WORK(&cb->cb_work, nfsd4_run_cb_work);
-	cb->cb_seq_status = 1;
 	cb->cb_status = 0;
 	cb->cb_need_restart = false;
 	cb->cb_holds_slot = false;
diff --git a/fs/nfsd/nfs4layouts.c b/fs/nfsd/nfs4layouts.c
index 5e8096bc5e..4f3072b597 100644
--- a/fs/nfsd/nfs4layouts.c
+++ b/fs/nfsd/nfs4layouts.c
@@ -25,7 +25,7 @@ static struct kmem_cache *nfs4_layout_cache;
 static struct kmem_cache *nfs4_layout_stateid_cache;
 
 static const struct nfsd4_callback_ops nfsd4_cb_layout_ops;
-static const struct lock_manager_operations nfsd4_layouts_lm_ops;
+static const struct lease_manager_operations nfsd4_layouts_lm_ops;
 
 const struct nfsd4_layout_ops *nfsd4_layout_ops[LAYOUT_TYPE_MAX] =  {
 #ifdef CONFIG_NFSD_FLEXFILELAYOUT
@@ -152,6 +152,23 @@ void nfsd4_setup_layout_type(struct svc_export *exp)
 #endif
 }
 
+void nfsd4_close_layout(struct nfs4_layout_stateid *ls)
+{
+	struct nfsd_file *fl;
+
+	spin_lock(&ls->ls_stid.sc_file->fi_lock);
+	fl = ls->ls_file;
+	ls->ls_file = NULL;
+	spin_unlock(&ls->ls_stid.sc_file->fi_lock);
+
+	if (fl) {
+		if (!nfsd4_layout_ops[ls->ls_layout_type]->disable_recalls)
+			kernel_setlease(fl->nf_file, F_UNLCK, NULL,
+					(void **)&ls);
+		nfsd_file_put(fl);
+	}
+}
+
 static void
 nfsd4_free_layout_stateid(struct nfs4_stid *stid)
 {
@@ -169,9 +186,7 @@ nfsd4_free_layout_stateid(struct nfs4_stid *stid)
 	list_del_init(&ls->ls_perfile);
 	spin_unlock(&fp->fi_lock);
 
-	if (!nfsd4_layout_ops[ls->ls_layout_type]->disable_recalls)
-		vfs_setlease(ls->ls_file->nf_file, F_UNLCK, NULL, (void **)&ls);
-	nfsd_file_put(ls->ls_file);
+	nfsd4_close_layout(ls);
 
 	if (ls->ls_recalled)
 		atomic_dec(&ls->ls_stid.sc_file->fi_lo_recalls);
@@ -182,27 +197,26 @@ nfsd4_free_layout_stateid(struct nfs4_stid *stid)
 static int
 nfsd4_layout_setlease(struct nfs4_layout_stateid *ls)
 {
-	struct file_lock *fl;
+	struct file_lease *fl;
 	int status;
 
 	if (nfsd4_layout_ops[ls->ls_layout_type]->disable_recalls)
 		return 0;
 
-	fl = locks_alloc_lock();
+	fl = locks_alloc_lease();
 	if (!fl)
 		return -ENOMEM;
-	locks_init_lock(fl);
+	locks_init_lease(fl);
 	fl->fl_lmops = &nfsd4_layouts_lm_ops;
-	fl->fl_flags = FL_LAYOUT;
-	fl->fl_type = F_RDLCK;
-	fl->fl_end = OFFSET_MAX;
-	fl->fl_owner = ls;
-	fl->fl_pid = current->tgid;
-	fl->fl_file = ls->ls_file->nf_file;
-
-	status = vfs_setlease(fl->fl_file, fl->fl_type, &fl, NULL);
+	fl->c.flc_flags = FL_LAYOUT;
+	fl->c.flc_type = F_RDLCK;
+	fl->c.flc_owner = ls;
+	fl->c.flc_pid = current->tgid;
+	fl->c.flc_file = ls->ls_file->nf_file;
+
+	status = kernel_setlease(fl->c.flc_file, fl->c.flc_type, &fl, NULL);
 	if (status) {
-		locks_free_lock(fl);
+		locks_free_lease(fl);
 		return status;
 	}
 	BUG_ON(fl != NULL);
@@ -236,7 +250,7 @@ nfsd4_alloc_layout_stateid(struct nfsd4_compound_state *cstate,
 	nfsd4_init_cb(&ls->ls_recall, clp, &nfsd4_cb_layout_ops,
 			NFSPROC4_CLNT_CB_LAYOUT);
 
-	if (parent->sc_type == NFS4_DELEG_STID)
+	if (parent->sc_type == SC_TYPE_DELEG)
 		ls->ls_file = nfsd_file_get(fp->fi_deleg_file);
 	else
 		ls->ls_file = find_any_file(fp);
@@ -250,7 +264,7 @@ nfsd4_alloc_layout_stateid(struct nfsd4_compound_state *cstate,
 	}
 
 	spin_lock(&clp->cl_lock);
-	stp->sc_type = NFS4_LAYOUT_STID;
+	stp->sc_type = SC_TYPE_LAYOUT;
 	list_add(&ls->ls_perclnt, &clp->cl_lo_states);
 	spin_unlock(&clp->cl_lock);
 
@@ -269,13 +283,13 @@ nfsd4_preprocess_layout_stateid(struct svc_rqst *rqstp,
 {
 	struct nfs4_layout_stateid *ls;
 	struct nfs4_stid *stid;
-	unsigned char typemask = NFS4_LAYOUT_STID;
+	unsigned short typemask = SC_TYPE_LAYOUT;
 	__be32 status;
 
 	if (create)
-		typemask |= (NFS4_OPEN_STID | NFS4_LOCK_STID | NFS4_DELEG_STID);
+		typemask |= (SC_TYPE_OPEN | SC_TYPE_LOCK | SC_TYPE_DELEG);
 
-	status = nfsd4_lookup_stateid(cstate, stateid, typemask, &stid,
+	status = nfsd4_lookup_stateid(cstate, stateid, typemask, 0, &stid,
 			net_generic(SVC_NET(rqstp), nfsd_net_id));
 	if (status)
 		goto out;
@@ -286,7 +300,7 @@ nfsd4_preprocess_layout_stateid(struct svc_rqst *rqstp,
 		goto out_put_stid;
 	}
 
-	if (stid->sc_type != NFS4_LAYOUT_STID) {
+	if (stid->sc_type != SC_TYPE_LAYOUT) {
 		ls = nfsd4_alloc_layout_stateid(cstate, stid, layout_type);
 		nfs4_put_stid(stid);
 
@@ -518,7 +532,7 @@ nfsd4_return_file_layouts(struct svc_rqst *rqstp,
 		lrp->lrs_present = true;
 	} else {
 		trace_nfsd_layoutstate_unhash(&ls->ls_stid.sc_stateid);
-		nfs4_unhash_stid(&ls->ls_stid);
+		ls->ls_stid.sc_status |= SC_STATUS_CLOSED;
 		lrp->lrs_present = false;
 	}
 	spin_unlock(&ls->ls_lock);
@@ -605,7 +619,7 @@ nfsd4_return_all_file_layouts(struct nfs4_client *clp, struct nfs4_file *fp)
 }
 
 static void
-nfsd4_cb_layout_fail(struct nfs4_layout_stateid *ls)
+nfsd4_cb_layout_fail(struct nfs4_layout_stateid *ls, struct nfsd_file *file)
 {
 	struct nfs4_client *clp = ls->ls_stid.sc_client;
 	char addr_str[INET6_ADDRSTRLEN];
@@ -627,7 +641,7 @@ nfsd4_cb_layout_fail(struct nfs4_layout_stateid *ls)
 
 	argv[0] = (char *)nfsd_recall_failed;
 	argv[1] = addr_str;
-	argv[2] = ls->ls_file->nf_file->f_path.mnt->mnt_sb->s_id;
+	argv[2] = file->nf_file->f_path.mnt->mnt_sb->s_id;
 	argv[3] = NULL;
 
 	error = call_usermodehelper(nfsd_recall_failed, argv, envp,
@@ -657,6 +671,7 @@ nfsd4_cb_layout_done(struct nfsd4_callback *cb, struct rpc_task *task)
 	struct nfsd_net *nn;
 	ktime_t now, cutoff;
 	const struct nfsd4_layout_ops *ops;
+	struct nfsd_file *fl;
 
 	trace_nfsd_cb_layout_done(&ls->ls_stid.sc_stateid, task);
 	switch (task->tk_status) {
@@ -688,12 +703,17 @@ nfsd4_cb_layout_done(struct nfsd4_callback *cb, struct rpc_task *task)
 		 * Unknown error or non-responding client, we'll need to fence.
 		 */
 		trace_nfsd_layout_recall_fail(&ls->ls_stid.sc_stateid);
-
-		ops = nfsd4_layout_ops[ls->ls_layout_type];
-		if (ops->fence_client)
-			ops->fence_client(ls);
-		else
-			nfsd4_cb_layout_fail(ls);
+		rcu_read_lock();
+		fl = nfsd_file_get(ls->ls_file);
+		rcu_read_unlock();
+		if (fl) {
+			ops = nfsd4_layout_ops[ls->ls_layout_type];
+			if (ops->fence_client)
+				ops->fence_client(ls, fl);
+			else
+				nfsd4_cb_layout_fail(ls, fl);
+			nfsd_file_put(fl);
+		}
 		return 1;
 	case -NFS4ERR_NOMATCHING_LAYOUT:
 		trace_nfsd_layout_recall_done(&ls->ls_stid.sc_stateid);
@@ -723,7 +743,7 @@ static const struct nfsd4_callback_ops nfsd4_cb_layout_ops = {
 };
 
 static bool
-nfsd4_layout_lm_break(struct file_lock *fl)
+nfsd4_layout_lm_break(struct file_lease *fl)
 {
 	/*
 	 * We don't want the locks code to timeout the lease for us;
@@ -731,19 +751,19 @@ nfsd4_layout_lm_break(struct file_lock *fl)
 	 * in time:
 	 */
 	fl->fl_break_time = 0;
-	nfsd4_recall_file_layout(fl->fl_owner);
+	nfsd4_recall_file_layout(fl->c.flc_owner);
 	return false;
 }
 
 static int
-nfsd4_layout_lm_change(struct file_lock *onlist, int arg,
+nfsd4_layout_lm_change(struct file_lease *onlist, int arg,
 		struct list_head *dispose)
 {
 	BUG_ON(!(arg & F_UNLCK));
 	return lease_modify(onlist, arg, dispose);
 }
 
-static const struct lock_manager_operations nfsd4_layouts_lm_ops = {
+static const struct lease_manager_operations nfsd4_layouts_lm_ops = {
 	.lm_break	= nfsd4_layout_lm_break,
 	.lm_change	= nfsd4_layout_lm_change,
 };
@@ -756,13 +776,11 @@ nfsd4_init_pnfs(void)
 	for (i = 0; i < DEVID_HASH_SIZE; i++)
 		INIT_LIST_HEAD(&nfsd_devid_hash[i]);
 
-	nfs4_layout_cache = kmem_cache_create("nfs4_layout",
-			sizeof(struct nfs4_layout), 0, 0, NULL);
+	nfs4_layout_cache = KMEM_CACHE(nfs4_layout, 0);
 	if (!nfs4_layout_cache)
 		return -ENOMEM;
 
-	nfs4_layout_stateid_cache = kmem_cache_create("nfs4_layout_stateid",
-			sizeof(struct nfs4_layout_stateid), 0, 0, NULL);
+	nfs4_layout_stateid_cache = KMEM_CACHE(nfs4_layout_stateid, 0);
 	if (!nfs4_layout_stateid_cache) {
 		kmem_cache_destroy(nfs4_layout_cache);
 		return -ENOMEM;
diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c
index 648ff42700..2e39cf2e50 100644
--- a/fs/nfsd/nfs4proc.c
+++ b/fs/nfsd/nfs4proc.c
@@ -1143,6 +1143,7 @@ nfsd4_setattr(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	};
 	struct inode *inode;
 	__be32 status = nfs_ok;
+	bool save_no_wcc;
 	int err;
 
 	if (setattr->sa_iattr.ia_valid & ATTR_SIZE) {
@@ -1168,8 +1169,10 @@ nfsd4_setattr(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 
 	if (status)
 		goto out;
-	status = nfsd_setattr(rqstp, &cstate->current_fh, &attrs,
-				0, (time64_t)0);
+	save_no_wcc = cstate->current_fh.fh_no_wcc;
+	cstate->current_fh.fh_no_wcc = true;
+	status = nfsd_setattr(rqstp, &cstate->current_fh, &attrs, NULL);
+	cstate->current_fh.fh_no_wcc = save_no_wcc;
 	if (!status)
 		status = nfserrno(attrs.na_labelerr);
 	if (!status)
@@ -1734,7 +1737,7 @@ static void cleanup_async_copy(struct nfsd4_copy *copy)
 	nfs4_put_copy(copy);
 }
 
-static void nfsd4_send_cb_offload(struct nfsd4_copy *copy, __be32 nfserr)
+static void nfsd4_send_cb_offload(struct nfsd4_copy *copy)
 {
 	struct nfsd4_cb_offload *cbo;
 
@@ -1744,12 +1747,12 @@ static void nfsd4_send_cb_offload(struct nfsd4_copy *copy, __be32 nfserr)
 
 	memcpy(&cbo->co_res, &copy->cp_res, sizeof(copy->cp_res));
 	memcpy(&cbo->co_fh, &copy->fh, sizeof(copy->fh));
-	cbo->co_nfserr = nfserr;
+	cbo->co_nfserr = copy->nfserr;
 
 	nfsd4_init_cb(&cbo->co_cb, copy->cp_clp, &nfsd4_cb_offload_ops,
 		      NFSPROC4_CLNT_CB_OFFLOAD);
 	trace_nfsd_cb_offload(copy->cp_clp, &cbo->co_res.cb_stateid,
-			      &cbo->co_fh, copy->cp_count, nfserr);
+			      &cbo->co_fh, copy->cp_count, copy->nfserr);
 	nfsd4_run_cb(&cbo->co_cb);
 }
 
@@ -1763,7 +1766,6 @@ static void nfsd4_send_cb_offload(struct nfsd4_copy *copy, __be32 nfserr)
 static int nfsd4_do_async_copy(void *data)
 {
 	struct nfsd4_copy *copy = (struct nfsd4_copy *)data;
-	__be32 nfserr;
 
 	trace_nfsd_copy_do_async(copy);
 	if (nfsd4_ssc_is_inter(copy)) {
@@ -1774,24 +1776,25 @@ static int nfsd4_do_async_copy(void *data)
 		if (IS_ERR(filp)) {
 			switch (PTR_ERR(filp)) {
 			case -EBADF:
-				nfserr = nfserr_wrong_type;
+				copy->nfserr = nfserr_wrong_type;
 				break;
 			default:
-				nfserr = nfserr_offload_denied;
+				copy->nfserr = nfserr_offload_denied;
 			}
 			/* ss_mnt will be unmounted by the laundromat */
 			goto do_callback;
 		}
-		nfserr = nfsd4_do_copy(copy, filp, copy->nf_dst->nf_file,
-				       false);
+		copy->nfserr = nfsd4_do_copy(copy, filp, copy->nf_dst->nf_file,
+					     false);
 		nfsd4_cleanup_inter_ssc(copy->ss_nsui, filp, copy->nf_dst);
 	} else {
-		nfserr = nfsd4_do_copy(copy, copy->nf_src->nf_file,
-				       copy->nf_dst->nf_file, false);
+		copy->nfserr = nfsd4_do_copy(copy, copy->nf_src->nf_file,
+					     copy->nf_dst->nf_file, false);
 	}
 
 do_callback:
-	nfsd4_send_cb_offload(copy, nfserr);
+	set_bit(NFSD4_COPY_F_COMPLETED, &copy->cp_flags);
+	nfsd4_send_cb_offload(copy);
 	cleanup_async_copy(copy);
 	return 0;
 }
@@ -1804,6 +1807,13 @@ nfsd4_copy(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	__be32 status;
 	struct nfsd4_copy *async_copy = NULL;
 
+	/*
+	 * Currently, async COPY is not reliable. Force all COPY
+	 * requests to be synchronous to avoid client application
+	 * hangs waiting for COPY completion.
+	 */
+	nfsd4_copy_set_sync(copy, true);
+
 	copy->cp_clp = cstate->clp;
 	if (nfsd4_ssc_is_inter(copy)) {
 		trace_nfsd_copy_inter(copy);
@@ -2000,11 +2010,16 @@ nfsd4_offload_status(struct svc_rqst *rqstp,
 	struct nfsd4_copy *copy;
 	struct nfs4_client *clp = cstate->clp;
 
+	os->completed = false;
 	spin_lock(&clp->async_lock);
 	copy = find_async_copy_locked(clp, &os->stateid);
-	if (copy)
+	if (copy) {
 		os->count = copy->cp_res.wr_bytes_written;
-	else
+		if (test_bit(NFSD4_COPY_F_COMPLETED, &copy->cp_flags)) {
+			os->completed = true;
+			os->status = copy->nfserr;
+		}
+	} else
 		status = nfserr_bad_stateid;
 	spin_unlock(&clp->async_lock);
 
@@ -2151,6 +2166,29 @@ nfsd4_verify(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	return status == nfserr_same ? nfs_ok : status;
 }
 
+static __be32
+nfsd4_get_dir_delegation(struct svc_rqst *rqstp,
+			 struct nfsd4_compound_state *cstate,
+			 union nfsd4_op_u *u)
+{
+	struct nfsd4_get_dir_delegation *gdd = &u->get_dir_delegation;
+
+	/*
+	 * RFC 8881, section 18.39.3 says:
+	 *
+	 * "The server may refuse to grant the delegation. In that case, the
+	 *  server will return NFS4ERR_DIRDELEG_UNAVAIL."
+	 *
+	 * This is sub-optimal, since it means that the server would need to
+	 * abort compound processing just because the delegation wasn't
+	 * available. RFC8881bis should change this to allow the server to
+	 * return NFS4_OK with a non-fatal status of GDD4_UNAVAIL in this
+	 * situation.
+	 */
+	gdd->gddrnf_status = GDD4_UNAVAIL;
+	return nfs_ok;
+}
+
 #ifdef CONFIG_NFSD_PNFS
 static const struct nfsd4_layout_ops *
 nfsd4_layout_verify(struct svc_export *exp, unsigned int layout_type)
@@ -2231,7 +2269,7 @@ nfsd4_layoutget(struct svc_rqst *rqstp,
 	const struct nfsd4_layout_ops *ops;
 	struct nfs4_layout_stateid *ls;
 	__be32 nfserr;
-	int accmode = NFSD_MAY_READ_IF_EXEC;
+	int accmode = NFSD_MAY_READ_IF_EXEC | NFSD_MAY_OWNER_OVERRIDE;
 
 	switch (lgp->lg_seg.iomode) {
 	case IOMODE_READ:
@@ -2321,7 +2359,8 @@ nfsd4_layoutcommit(struct svc_rqst *rqstp,
 	struct nfs4_layout_stateid *ls;
 	__be32 nfserr;
 
-	nfserr = fh_verify(rqstp, current_fh, 0, NFSD_MAY_WRITE);
+	nfserr = fh_verify(rqstp, current_fh, 0,
+			   NFSD_MAY_WRITE | NFSD_MAY_OWNER_OVERRIDE);
 	if (nfserr)
 		goto out;
 
@@ -3079,6 +3118,18 @@ static u32 nfsd4_copy_notify_rsize(const struct svc_rqst *rqstp,
 		* sizeof(__be32);
 }
 
+static u32 nfsd4_get_dir_delegation_rsize(const struct svc_rqst *rqstp,
+					  const struct nfsd4_op *op)
+{
+	return (op_encode_hdr_size +
+		1 /* gddr_status */ +
+		op_encode_verifier_maxsz +
+		op_encode_stateid_maxsz +
+		2 /* gddr_notification */ +
+		2 /* gddr_child_attributes */ +
+		2 /* gddr_dir_attributes */);
+}
+
 #ifdef CONFIG_NFSD_PNFS
 static u32 nfsd4_getdeviceinfo_rsize(const struct svc_rqst *rqstp,
 				     const struct nfsd4_op *op)
@@ -3467,6 +3518,12 @@ static const struct nfsd4_operation nfsd4_ops[] = {
 		.op_get_currentstateid = nfsd4_get_freestateid,
 		.op_rsize_bop = nfsd4_only_status_rsize,
 	},
+	[OP_GET_DIR_DELEGATION] = {
+		.op_func = nfsd4_get_dir_delegation,
+		.op_flags = OP_MODIFIES_SOMETHING,
+		.op_name = "OP_GET_DIR_DELEGATION",
+		.op_rsize_bop = nfsd4_get_dir_delegation_rsize,
+	},
 #ifdef CONFIG_NFSD_PNFS
 	[OP_GETDEVICEINFO] = {
 		.op_func = nfsd4_getdeviceinfo,
diff --git a/fs/nfsd/nfs4recover.c b/fs/nfsd/nfs4recover.c
index 2c060e0b16..67d8673a93 100644
--- a/fs/nfsd/nfs4recover.c
+++ b/fs/nfsd/nfs4recover.c
@@ -2086,8 +2086,8 @@ do_init:
 	status = nn->client_tracking_ops->init(net);
 out:
 	if (status) {
-		printk(KERN_WARNING "NFSD: Unable to initialize client "
-				    "recovery tracking! (%d)\n", status);
+		pr_warn("NFSD: Unable to initialize client recovery tracking! (%d)\n", status);
+		pr_warn("NFSD: Is nfsdcld running? If not, enable CONFIG_NFSD_LEGACY_CLIENT_TRACKING.\n");
 		nn->client_tracking_ops = NULL;
 	}
 	return status;
diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
index 71f9442c5c..a20c2c9d7d 100644
--- a/fs/nfsd/nfs4state.c
+++ b/fs/nfsd/nfs4state.c
@@ -87,6 +87,7 @@ static void nfs4_free_ol_stateid(struct nfs4_stid *stid);
 void nfsd4_end_grace(struct nfsd_net *nn);
 static void _free_cpntf_state_locked(struct nfsd_net *nn, struct nfs4_cpntf_state *cps);
 static void nfsd4_file_hash_remove(struct nfs4_file *fi);
+static void deleg_reaper(struct nfsd_net *nn);
 
 /* Locking: */
 
@@ -127,6 +128,7 @@ static void free_session(struct nfsd4_session *);
 
 static const struct nfsd4_callback_ops nfsd4_cb_recall_ops;
 static const struct nfsd4_callback_ops nfsd4_cb_notify_lock_ops;
+static const struct nfsd4_callback_ops nfsd4_cb_getattr_ops;
 
 static struct workqueue_struct *laundry_wq;
 
@@ -318,6 +320,7 @@ free_nbl(struct kref *kref)
 	struct nfsd4_blocked_lock *nbl;
 
 	nbl = container_of(kref, struct nfsd4_blocked_lock, nbl_kref);
+	locks_release_private(&nbl->nbl_lock);
 	kfree(nbl);
 }
 
@@ -325,7 +328,6 @@ static void
 free_blocked_lock(struct nfsd4_blocked_lock *nbl)
 {
 	locks_delete_block(&nbl->nbl_lock);
-	locks_release_private(&nbl->nbl_lock);
 	kref_put(&nbl->nbl_kref, free_nbl);
 }
 
@@ -539,7 +541,7 @@ same_owner_str(struct nfs4_stateowner *sop, struct xdr_netobj *owner)
 }
 
 static struct nfs4_openowner *
-find_openstateowner_str_locked(unsigned int hashval, struct nfsd4_open *open,
+find_openstateowner_str(unsigned int hashval, struct nfsd4_open *open,
 			struct nfs4_client *clp)
 {
 	struct nfs4_stateowner *so;
@@ -556,18 +558,6 @@ find_openstateowner_str_locked(unsigned int hashval, struct nfsd4_open *open,
 	return NULL;
 }
 
-static struct nfs4_openowner *
-find_openstateowner_str(unsigned int hashval, struct nfsd4_open *open,
-			struct nfs4_client *clp)
-{
-	struct nfs4_openowner *oo;
-
-	spin_lock(&clp->cl_lock);
-	oo = find_openstateowner_str_locked(hashval, open, clp);
-	spin_unlock(&clp->cl_lock);
-	return oo;
-}
-
 static inline u32
 opaque_hashval(const void *ptr, int nbytes)
 {
@@ -1189,6 +1179,10 @@ alloc_init_deleg(struct nfs4_client *clp, struct nfs4_file *fp,
 	dp->dl_recalled = false;
 	nfsd4_init_cb(&dp->dl_recall, dp->dl_stid.sc_client,
 		      &nfsd4_cb_recall_ops, NFSPROC4_CLNT_CB_RECALL);
+	nfsd4_init_cb(&dp->dl_cb_fattr.ncf_getattr, dp->dl_stid.sc_client,
+			&nfsd4_cb_getattr_ops, NFSPROC4_CLNT_CB_GETATTR);
+	dp->dl_cb_fattr.ncf_file_modified = false;
+	dp->dl_cb_fattr.ncf_cb_bmap[0] = FATTR4_WORD0_CHANGE | FATTR4_WORD0_SIZE;
 	get_nfs4_file(fp);
 	dp->dl_stid.sc_file = fp;
 	return dp;
@@ -1210,6 +1204,8 @@ nfs4_put_stid(struct nfs4_stid *s)
 		return;
 	}
 	idr_remove(&clp->cl_stateids, s->sc_stateid.si_opaque.so_id);
+	if (s->sc_status & SC_STATUS_ADMIN_REVOKED)
+		atomic_dec(&s->sc_client->cl_admin_revoked);
 	nfs4_free_cpntf_statelist(clp->net, s);
 	spin_unlock(&clp->cl_lock);
 	s->sc_free(s);
@@ -1249,7 +1245,7 @@ static void nfs4_unlock_deleg_lease(struct nfs4_delegation *dp)
 
 	WARN_ON_ONCE(!fp->fi_delegees);
 
-	vfs_setlease(nf->nf_file, F_UNLCK, NULL, (void **)&dp);
+	kernel_setlease(nf->nf_file, F_UNLCK, NULL, (void **)&dp);
 	put_deleg_file(fp);
 }
 
@@ -1260,11 +1256,6 @@ static void destroy_unhashed_deleg(struct nfs4_delegation *dp)
 	nfs4_put_stid(&dp->dl_stid);
 }
 
-void nfs4_unhash_stid(struct nfs4_stid *s)
-{
-	s->sc_type = 0;
-}
-
 /**
  * nfs4_delegation_exists - Discover if this delegation already exists
  * @clp:     a pointer to the nfs4_client we're granting a delegation to
@@ -1312,11 +1303,12 @@ hash_delegation_locked(struct nfs4_delegation *dp, struct nfs4_file *fp)
 
 	lockdep_assert_held(&state_lock);
 	lockdep_assert_held(&fp->fi_lock);
+	lockdep_assert_held(&clp->cl_lock);
 
 	if (nfs4_delegation_exists(clp, fp))
 		return -EAGAIN;
 	refcount_inc(&dp->dl_stid.sc_count);
-	dp->dl_stid.sc_type = NFS4_DELEG_STID;
+	dp->dl_stid.sc_type = SC_TYPE_DELEG;
 	list_add(&dp->dl_perfile, &fp->fi_delegations);
 	list_add(&dp->dl_perclnt, &clp->cl_delegations);
 	return 0;
@@ -1328,7 +1320,7 @@ static bool delegation_hashed(struct nfs4_delegation *dp)
 }
 
 static bool
-unhash_delegation_locked(struct nfs4_delegation *dp)
+unhash_delegation_locked(struct nfs4_delegation *dp, unsigned short statusmask)
 {
 	struct nfs4_file *fp = dp->dl_stid.sc_file;
 
@@ -1337,7 +1329,13 @@ unhash_delegation_locked(struct nfs4_delegation *dp)
 	if (!delegation_hashed(dp))
 		return false;
 
-	dp->dl_stid.sc_type = NFS4_CLOSED_DELEG_STID;
+	if (statusmask == SC_STATUS_REVOKED &&
+	    dp->dl_stid.sc_client->cl_minorversion == 0)
+		statusmask = SC_STATUS_CLOSED;
+	dp->dl_stid.sc_status |= statusmask;
+	if (statusmask & SC_STATUS_ADMIN_REVOKED)
+		atomic_inc(&dp->dl_stid.sc_client->cl_admin_revoked);
+
 	/* Ensure that deleg break won't try to requeue it */
 	++dp->dl_time;
 	spin_lock(&fp->fi_lock);
@@ -1353,7 +1351,7 @@ static void destroy_delegation(struct nfs4_delegation *dp)
 	bool unhashed;
 
 	spin_lock(&state_lock);
-	unhashed = unhash_delegation_locked(dp);
+	unhashed = unhash_delegation_locked(dp, SC_STATUS_CLOSED);
 	spin_unlock(&state_lock);
 	if (unhashed)
 		destroy_unhashed_deleg(dp);
@@ -1367,9 +1365,9 @@ static void revoke_delegation(struct nfs4_delegation *dp)
 
 	trace_nfsd_stid_revoke(&dp->dl_stid);
 
-	if (clp->cl_minorversion) {
+	if (dp->dl_stid.sc_status &
+	    (SC_STATUS_REVOKED | SC_STATUS_ADMIN_REVOKED)) {
 		spin_lock(&clp->cl_lock);
-		dp->dl_stid.sc_type = NFS4_REVOKED_DELEG_STID;
 		refcount_inc(&dp->dl_stid.sc_count);
 		list_add(&dp->dl_recall_lru, &clp->cl_revoked);
 		spin_unlock(&clp->cl_lock);
@@ -1377,8 +1375,8 @@ static void revoke_delegation(struct nfs4_delegation *dp)
 	destroy_unhashed_deleg(dp);
 }
 
-/* 
- * SETCLIENTID state 
+/*
+ * SETCLIENTID state
  */
 
 static unsigned int clientid_hashval(u32 id)
@@ -1399,11 +1397,16 @@ static void
 recalculate_deny_mode(struct nfs4_file *fp)
 {
 	struct nfs4_ol_stateid *stp;
+	u32 old_deny;
 
 	spin_lock(&fp->fi_lock);
+	old_deny = fp->fi_share_deny;
 	fp->fi_share_deny = 0;
-	list_for_each_entry(stp, &fp->fi_stateids, st_perfile)
+	list_for_each_entry(stp, &fp->fi_stateids, st_perfile) {
 		fp->fi_share_deny |= bmap_to_share_mode(stp->st_deny_bmap);
+		if (fp->fi_share_deny == old_deny)
+			break;
+	}
 	spin_unlock(&fp->fi_lock);
 }
 
@@ -1531,6 +1534,8 @@ static void put_ol_stateid_locked(struct nfs4_ol_stateid *stp,
 	}
 
 	idr_remove(&clp->cl_stateids, s->sc_stateid.si_opaque.so_id);
+	if (s->sc_status & SC_STATUS_ADMIN_REVOKED)
+		atomic_dec(&s->sc_client->cl_admin_revoked);
 	list_add(&stp->st_locks, reaplist);
 }
 
@@ -1541,7 +1546,7 @@ static bool unhash_lock_stateid(struct nfs4_ol_stateid *stp)
 	if (!unhash_ol_stateid(stp))
 		return false;
 	list_del_init(&stp->st_locks);
-	nfs4_unhash_stid(&stp->st_stid);
+	stp->st_stid.sc_status |= SC_STATUS_CLOSED;
 	return true;
 }
 
@@ -1599,7 +1604,7 @@ static void release_open_stateid_locks(struct nfs4_ol_stateid *open_stp,
 	while (!list_empty(&open_stp->st_locks)) {
 		stp = list_entry(open_stp->st_locks.next,
 				struct nfs4_ol_stateid, st_locks);
-		WARN_ON(!unhash_lock_stateid(stp));
+		unhash_lock_stateid(stp);
 		put_ol_stateid_locked(stp, reaplist);
 	}
 }
@@ -1620,6 +1625,7 @@ static void release_open_stateid(struct nfs4_ol_stateid *stp)
 	LIST_HEAD(reaplist);
 
 	spin_lock(&stp->st_stid.sc_client->cl_lock);
+	stp->st_stid.sc_status |= SC_STATUS_CLOSED;
 	if (unhash_open_stateid(stp, &reaplist))
 		put_ol_stateid_locked(stp, &reaplist);
 	spin_unlock(&stp->st_stid.sc_client->cl_lock);
@@ -1675,6 +1681,136 @@ static void release_openowner(struct nfs4_openowner *oo)
 	nfs4_put_stateowner(&oo->oo_owner);
 }
 
+static struct nfs4_stid *find_one_sb_stid(struct nfs4_client *clp,
+					  struct super_block *sb,
+					  unsigned int sc_types)
+{
+	unsigned long id, tmp;
+	struct nfs4_stid *stid;
+
+	spin_lock(&clp->cl_lock);
+	idr_for_each_entry_ul(&clp->cl_stateids, stid, tmp, id)
+		if ((stid->sc_type & sc_types) &&
+		    stid->sc_status == 0 &&
+		    stid->sc_file->fi_inode->i_sb == sb) {
+			refcount_inc(&stid->sc_count);
+			break;
+		}
+	spin_unlock(&clp->cl_lock);
+	return stid;
+}
+
+/**
+ * nfsd4_revoke_states - revoke all nfsv4 states associated with given filesystem
+ * @net:  used to identify instance of nfsd (there is one per net namespace)
+ * @sb:   super_block used to identify target filesystem
+ *
+ * All nfs4 states (open, lock, delegation, layout) held by the server instance
+ * and associated with a file on the given filesystem will be revoked resulting
+ * in any files being closed and so all references from nfsd to the filesystem
+ * being released.  Thus nfsd will no longer prevent the filesystem from being
+ * unmounted.
+ *
+ * The clients which own the states will subsequently being notified that the
+ * states have been "admin-revoked".
+ */
+void nfsd4_revoke_states(struct net *net, struct super_block *sb)
+{
+	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
+	unsigned int idhashval;
+	unsigned int sc_types;
+
+	sc_types = SC_TYPE_OPEN | SC_TYPE_LOCK | SC_TYPE_DELEG | SC_TYPE_LAYOUT;
+
+	spin_lock(&nn->client_lock);
+	for (idhashval = 0; idhashval < CLIENT_HASH_MASK; idhashval++) {
+		struct list_head *head = &nn->conf_id_hashtbl[idhashval];
+		struct nfs4_client *clp;
+	retry:
+		list_for_each_entry(clp, head, cl_idhash) {
+			struct nfs4_stid *stid = find_one_sb_stid(clp, sb,
+								  sc_types);
+			if (stid) {
+				struct nfs4_ol_stateid *stp;
+				struct nfs4_delegation *dp;
+				struct nfs4_layout_stateid *ls;
+
+				spin_unlock(&nn->client_lock);
+				switch (stid->sc_type) {
+				case SC_TYPE_OPEN:
+					stp = openlockstateid(stid);
+					mutex_lock_nested(&stp->st_mutex,
+							  OPEN_STATEID_MUTEX);
+
+					spin_lock(&clp->cl_lock);
+					if (stid->sc_status == 0) {
+						stid->sc_status |=
+							SC_STATUS_ADMIN_REVOKED;
+						atomic_inc(&clp->cl_admin_revoked);
+						spin_unlock(&clp->cl_lock);
+						release_all_access(stp);
+					} else
+						spin_unlock(&clp->cl_lock);
+					mutex_unlock(&stp->st_mutex);
+					break;
+				case SC_TYPE_LOCK:
+					stp = openlockstateid(stid);
+					mutex_lock_nested(&stp->st_mutex,
+							  LOCK_STATEID_MUTEX);
+					spin_lock(&clp->cl_lock);
+					if (stid->sc_status == 0) {
+						struct nfs4_lockowner *lo =
+							lockowner(stp->st_stateowner);
+						struct nfsd_file *nf;
+
+						stid->sc_status |=
+							SC_STATUS_ADMIN_REVOKED;
+						atomic_inc(&clp->cl_admin_revoked);
+						spin_unlock(&clp->cl_lock);
+						nf = find_any_file(stp->st_stid.sc_file);
+						if (nf) {
+							get_file(nf->nf_file);
+							filp_close(nf->nf_file,
+								   (fl_owner_t)lo);
+							nfsd_file_put(nf);
+						}
+						release_all_access(stp);
+					} else
+						spin_unlock(&clp->cl_lock);
+					mutex_unlock(&stp->st_mutex);
+					break;
+				case SC_TYPE_DELEG:
+					dp = delegstateid(stid);
+					spin_lock(&state_lock);
+					if (!unhash_delegation_locked(
+						    dp, SC_STATUS_ADMIN_REVOKED))
+						dp = NULL;
+					spin_unlock(&state_lock);
+					if (dp)
+						revoke_delegation(dp);
+					break;
+				case SC_TYPE_LAYOUT:
+					ls = layoutstateid(stid);
+					nfsd4_close_layout(ls);
+					break;
+				}
+				nfs4_put_stid(stid);
+				spin_lock(&nn->client_lock);
+				if (clp->cl_minorversion == 0)
+					/* Allow cleanup after a lease period.
+					 * store_release ensures cleanup will
+					 * see any newly revoked states if it
+					 * sees the time updated.
+					 */
+					nn->nfs40_last_revoke =
+						ktime_get_boottime_seconds();
+				goto retry;
+			}
+		}
+	}
+	spin_unlock(&nn->client_lock);
+}
+
 static inline int
 hash_sessionid(struct nfs4_sessionid *sessionid)
 {
@@ -2102,6 +2238,10 @@ static struct nfs4_client *alloc_client(struct xdr_netobj name,
 						 GFP_KERNEL);
 	if (!clp->cl_ownerstr_hashtbl)
 		goto err_no_hashtbl;
+	clp->cl_callback_wq = alloc_ordered_workqueue("nfsd4_callbacks", 0);
+	if (!clp->cl_callback_wq)
+		goto err_no_callback_wq;
+
 	for (i = 0; i < OWNER_HASH_SIZE; i++)
 		INIT_LIST_HEAD(&clp->cl_ownerstr_hashtbl[i]);
 	INIT_LIST_HEAD(&clp->cl_sessions);
@@ -2124,6 +2264,8 @@ static struct nfs4_client *alloc_client(struct xdr_netobj name,
 	spin_lock_init(&clp->cl_lock);
 	rpc_init_wait_queue(&clp->cl_cb_waitq, "Backchannel slot table");
 	return clp;
+err_no_callback_wq:
+	kfree(clp->cl_ownerstr_hashtbl);
 err_no_hashtbl:
 	kfree(clp->cl_name.data);
 err_no_name:
@@ -2137,6 +2279,7 @@ static void __free_client(struct kref *k)
 	struct nfs4_client *clp = container_of(c, struct nfs4_client, cl_nfsdfs);
 
 	free_svc_cred(&clp->cl_cred);
+	destroy_workqueue(clp->cl_callback_wq);
 	kfree(clp->cl_ownerstr_hashtbl);
 	kfree(clp->cl_name.data);
 	kfree(clp->cl_nii_domain.data);
@@ -2209,7 +2352,11 @@ unhash_client(struct nfs4_client *clp)
 
 static __be32 mark_client_expired_locked(struct nfs4_client *clp)
 {
-	if (atomic_read(&clp->cl_rpc_users))
+	int users = atomic_read(&clp->cl_rpc_users);
+
+	trace_nfsd_mark_client_expired(clp, users);
+
+	if (users)
 		return nfserr_jukebox;
 	unhash_client_locked(clp);
 	return nfs_ok;
@@ -2228,7 +2375,7 @@ __destroy_client(struct nfs4_client *clp)
 	spin_lock(&state_lock);
 	while (!list_empty(&clp->cl_delegations)) {
 		dp = list_entry(clp->cl_delegations.next, struct nfs4_delegation, dl_perclnt);
-		WARN_ON(!unhash_delegation_locked(dp));
+		unhash_delegation_locked(dp, SC_STATUS_CLOSED);
 		list_add(&dp->dl_recall_lru, &reaplist);
 	}
 	spin_unlock(&state_lock);
@@ -2460,14 +2607,16 @@ find_stateid_locked(struct nfs4_client *cl, stateid_t *t)
 }
 
 static struct nfs4_stid *
-find_stateid_by_type(struct nfs4_client *cl, stateid_t *t, char typemask)
+find_stateid_by_type(struct nfs4_client *cl, stateid_t *t,
+		     unsigned short typemask, unsigned short ok_states)
 {
 	struct nfs4_stid *s;
 
 	spin_lock(&cl->cl_lock);
 	s = find_stateid_locked(cl, t);
 	if (s != NULL) {
-		if (typemask & s->sc_type)
+		if ((s->sc_status & ~ok_states) == 0 &&
+		    (typemask & s->sc_type))
 			refcount_inc(&s->sc_count);
 		else
 			s = NULL;
@@ -2487,9 +2636,9 @@ static struct nfs4_client *get_nfsdfs_clp(struct inode *inode)
 
 static void seq_quote_mem(struct seq_file *m, char *data, int len)
 {
-	seq_printf(m, "\"");
+	seq_puts(m, "\"");
 	seq_escape_mem(m, data, len, ESCAPE_HEX | ESCAPE_NAP | ESCAPE_APPEND, "\"\\");
-	seq_printf(m, "\"");
+	seq_puts(m, "\"");
 }
 
 static const char *cb_state2str(int state)
@@ -2530,20 +2679,22 @@ static int client_info_show(struct seq_file *m, void *v)
 		seq_puts(m, "status: unconfirmed\n");
 	seq_printf(m, "seconds from last renew: %lld\n",
 		ktime_get_boottime_seconds() - clp->cl_time);
-	seq_printf(m, "name: ");
+	seq_puts(m, "name: ");
 	seq_quote_mem(m, clp->cl_name.data, clp->cl_name.len);
 	seq_printf(m, "\nminor version: %d\n", clp->cl_minorversion);
 	if (clp->cl_nii_domain.data) {
-		seq_printf(m, "Implementation domain: ");
+		seq_puts(m, "Implementation domain: ");
 		seq_quote_mem(m, clp->cl_nii_domain.data,
 					clp->cl_nii_domain.len);
-		seq_printf(m, "\nImplementation name: ");
+		seq_puts(m, "\nImplementation name: ");
 		seq_quote_mem(m, clp->cl_nii_name.data, clp->cl_nii_name.len);
 		seq_printf(m, "\nImplementation time: [%lld, %ld]\n",
 			clp->cl_nii_time.tv_sec, clp->cl_nii_time.tv_nsec);
 	}
 	seq_printf(m, "callback state: %s\n", cb_state2str(clp->cl_cb_state));
 	seq_printf(m, "callback address: %pISpc\n", &clp->cl_cb_conn.cb_addr);
+	seq_printf(m, "admin-revoked states: %d\n",
+		   atomic_read(&clp->cl_admin_revoked));
 	drop_client(clp);
 
 	return 0;
@@ -2602,7 +2753,7 @@ static void nfs4_show_superblock(struct seq_file *s, struct nfsd_file *f)
 
 static void nfs4_show_owner(struct seq_file *s, struct nfs4_stateowner *oo)
 {
-	seq_printf(s, "owner: ");
+	seq_puts(s, "owner: ");
 	seq_quote_mem(s, oo->so_owner.data, oo->so_owner.len);
 }
 
@@ -2620,20 +2771,13 @@ static int nfs4_show_open(struct seq_file *s, struct nfs4_stid *st)
 	struct nfs4_stateowner *oo;
 	unsigned int access, deny;
 
-	if (st->sc_type != NFS4_OPEN_STID && st->sc_type != NFS4_LOCK_STID)
-		return 0; /* XXX: or SEQ_SKIP? */
 	ols = openlockstateid(st);
 	oo = ols->st_stateowner;
 	nf = st->sc_file;
 
-	spin_lock(&nf->fi_lock);
-	file = find_any_file_locked(nf);
-	if (!file)
-		goto out;
-
-	seq_printf(s, "- ");
+	seq_puts(s, "- ");
 	nfs4_show_stateid(s, &st->sc_stateid);
-	seq_printf(s, ": { type: open, ");
+	seq_puts(s, ": { type: open, ");
 
 	access = bmap_to_share_mode(ols->st_access_bmap);
 	deny   = bmap_to_share_mode(ols->st_deny_bmap);
@@ -2645,14 +2789,19 @@ static int nfs4_show_open(struct seq_file *s, struct nfs4_stid *st)
 		deny & NFS4_SHARE_ACCESS_READ ? "r" : "-",
 		deny & NFS4_SHARE_ACCESS_WRITE ? "w" : "-");
 
-	nfs4_show_superblock(s, file);
-	seq_printf(s, ", ");
-	nfs4_show_fname(s, file);
-	seq_printf(s, ", ");
-	nfs4_show_owner(s, oo);
-	seq_printf(s, " }\n");
-out:
+	spin_lock(&nf->fi_lock);
+	file = find_any_file_locked(nf);
+	if (file) {
+		nfs4_show_superblock(s, file);
+		seq_puts(s, ", ");
+		nfs4_show_fname(s, file);
+		seq_puts(s, ", ");
+	}
 	spin_unlock(&nf->fi_lock);
+	nfs4_show_owner(s, oo);
+	if (st->sc_status & SC_STATUS_ADMIN_REVOKED)
+		seq_puts(s, ", admin-revoked");
+	seq_puts(s, " }\n");
 	return 0;
 }
 
@@ -2666,30 +2815,31 @@ static int nfs4_show_lock(struct seq_file *s, struct nfs4_stid *st)
 	ols = openlockstateid(st);
 	oo = ols->st_stateowner;
 	nf = st->sc_file;
-	spin_lock(&nf->fi_lock);
-	file = find_any_file_locked(nf);
-	if (!file)
-		goto out;
 
-	seq_printf(s, "- ");
+	seq_puts(s, "- ");
 	nfs4_show_stateid(s, &st->sc_stateid);
-	seq_printf(s, ": { type: lock, ");
+	seq_puts(s, ": { type: lock, ");
 
-	/*
-	 * Note: a lock stateid isn't really the same thing as a lock,
-	 * it's the locking state held by one owner on a file, and there
-	 * may be multiple (or no) lock ranges associated with it.
-	 * (Same for the matter is true of open stateids.)
-	 */
+	spin_lock(&nf->fi_lock);
+	file = find_any_file_locked(nf);
+	if (file) {
+		/*
+		 * Note: a lock stateid isn't really the same thing as a lock,
+		 * it's the locking state held by one owner on a file, and there
+		 * may be multiple (or no) lock ranges associated with it.
+		 * (Same for the matter is true of open stateids.)
+		 */
 
-	nfs4_show_superblock(s, file);
-	/* XXX: open stateid? */
-	seq_printf(s, ", ");
-	nfs4_show_fname(s, file);
-	seq_printf(s, ", ");
+		nfs4_show_superblock(s, file);
+		/* XXX: open stateid? */
+		seq_puts(s, ", ");
+		nfs4_show_fname(s, file);
+		seq_puts(s, ", ");
+	}
 	nfs4_show_owner(s, oo);
-	seq_printf(s, " }\n");
-out:
+	if (st->sc_status & SC_STATUS_ADMIN_REVOKED)
+		seq_puts(s, ", admin-revoked");
+	seq_puts(s, " }\n");
 	spin_unlock(&nf->fi_lock);
 	return 0;
 }
@@ -2702,27 +2852,28 @@ static int nfs4_show_deleg(struct seq_file *s, struct nfs4_stid *st)
 
 	ds = delegstateid(st);
 	nf = st->sc_file;
-	spin_lock(&nf->fi_lock);
-	file = nf->fi_deleg_file;
-	if (!file)
-		goto out;
 
-	seq_printf(s, "- ");
+	seq_puts(s, "- ");
 	nfs4_show_stateid(s, &st->sc_stateid);
-	seq_printf(s, ": { type: deleg, ");
+	seq_puts(s, ": { type: deleg, ");
 
-	/* Kinda dead code as long as we only support read delegs: */
-	seq_printf(s, "access: %s, ",
-		ds->dl_type == NFS4_OPEN_DELEGATE_READ ? "r" : "w");
+	seq_printf(s, "access: %s",
+		   ds->dl_type == NFS4_OPEN_DELEGATE_READ ? "r" : "w");
 
 	/* XXX: lease time, whether it's being recalled. */
 
-	nfs4_show_superblock(s, file);
-	seq_printf(s, ", ");
-	nfs4_show_fname(s, file);
-	seq_printf(s, " }\n");
-out:
+	spin_lock(&nf->fi_lock);
+	file = nf->fi_deleg_file;
+	if (file) {
+		seq_puts(s, ", ");
+		nfs4_show_superblock(s, file);
+		seq_puts(s, ", ");
+		nfs4_show_fname(s, file);
+	}
 	spin_unlock(&nf->fi_lock);
+	if (st->sc_status & SC_STATUS_ADMIN_REVOKED)
+		seq_puts(s, ", admin-revoked");
+	seq_puts(s, " }\n");
 	return 0;
 }
 
@@ -2732,18 +2883,25 @@ static int nfs4_show_layout(struct seq_file *s, struct nfs4_stid *st)
 	struct nfsd_file *file;
 
 	ls = container_of(st, struct nfs4_layout_stateid, ls_stid);
-	file = ls->ls_file;
 
-	seq_printf(s, "- ");
+	seq_puts(s, "- ");
 	nfs4_show_stateid(s, &st->sc_stateid);
-	seq_printf(s, ": { type: layout, ");
+	seq_puts(s, ": { type: layout");
 
 	/* XXX: What else would be useful? */
 
-	nfs4_show_superblock(s, file);
-	seq_printf(s, ", ");
-	nfs4_show_fname(s, file);
-	seq_printf(s, " }\n");
+	spin_lock(&ls->ls_stid.sc_file->fi_lock);
+	file = ls->ls_file;
+	if (file) {
+		seq_puts(s, ", ");
+		nfs4_show_superblock(s, file);
+		seq_puts(s, ", ");
+		nfs4_show_fname(s, file);
+	}
+	spin_unlock(&ls->ls_stid.sc_file->fi_lock);
+	if (st->sc_status & SC_STATUS_ADMIN_REVOKED)
+		seq_puts(s, ", admin-revoked");
+	seq_puts(s, " }\n");
 
 	return 0;
 }
@@ -2753,13 +2911,13 @@ static int states_show(struct seq_file *s, void *v)
 	struct nfs4_stid *st = v;
 
 	switch (st->sc_type) {
-	case NFS4_OPEN_STID:
+	case SC_TYPE_OPEN:
 		return nfs4_show_open(s, st);
-	case NFS4_LOCK_STID:
+	case SC_TYPE_LOCK:
 		return nfs4_show_lock(s, st);
-	case NFS4_DELEG_STID:
+	case SC_TYPE_DELEG:
 		return nfs4_show_deleg(s, st);
-	case NFS4_LAYOUT_STID:
+	case SC_TYPE_LAYOUT:
 		return nfs4_show_layout(s, st);
 	default:
 		return 0; /* XXX: or SEQ_SKIP? */
@@ -2893,11 +3051,59 @@ nfsd4_cb_recall_any_release(struct nfsd4_callback *cb)
 	drop_client(clp);
 }
 
+static int
+nfsd4_cb_getattr_done(struct nfsd4_callback *cb, struct rpc_task *task)
+{
+	struct nfs4_cb_fattr *ncf =
+			container_of(cb, struct nfs4_cb_fattr, ncf_getattr);
+
+	ncf->ncf_cb_status = task->tk_status;
+	switch (task->tk_status) {
+	case -NFS4ERR_DELAY:
+		rpc_delay(task, 2 * HZ);
+		return 0;
+	default:
+		return 1;
+	}
+}
+
+static void
+nfsd4_cb_getattr_release(struct nfsd4_callback *cb)
+{
+	struct nfs4_cb_fattr *ncf =
+			container_of(cb, struct nfs4_cb_fattr, ncf_getattr);
+	struct nfs4_delegation *dp =
+			container_of(ncf, struct nfs4_delegation, dl_cb_fattr);
+
+	nfs4_put_stid(&dp->dl_stid);
+	clear_bit(CB_GETATTR_BUSY, &ncf->ncf_cb_flags);
+	wake_up_bit(&ncf->ncf_cb_flags, CB_GETATTR_BUSY);
+}
+
 static const struct nfsd4_callback_ops nfsd4_cb_recall_any_ops = {
 	.done		= nfsd4_cb_recall_any_done,
 	.release	= nfsd4_cb_recall_any_release,
 };
 
+static const struct nfsd4_callback_ops nfsd4_cb_getattr_ops = {
+	.done		= nfsd4_cb_getattr_done,
+	.release	= nfsd4_cb_getattr_release,
+};
+
+static void nfs4_cb_getattr(struct nfs4_cb_fattr *ncf)
+{
+	struct nfs4_delegation *dp =
+			container_of(ncf, struct nfs4_delegation, dl_cb_fattr);
+
+	if (test_and_set_bit(CB_GETATTR_BUSY, &ncf->ncf_cb_flags))
+		return;
+	/* set to proper status when nfsd4_cb_getattr_done runs */
+	ncf->ncf_cb_status = NFS4ERR_IO;
+
+	refcount_inc(&dp->dl_stid.sc_count);
+	nfsd4_run_cb(&ncf->ncf_getattr);
+}
+
 static struct nfs4_client *create_client(struct xdr_netobj name,
 		struct svc_rqst *rqstp, nfs4_verifier *verf)
 {
@@ -3411,6 +3617,9 @@ out_new:
 	new->cl_spo_must_allow.u.words[0] = exid->spo_must_allow[0];
 	new->cl_spo_must_allow.u.words[1] = exid->spo_must_allow[1];
 
+	/* Contrived initial CREATE_SESSION response */
+	new->cl_cs_slot.sl_status = nfserr_seq_misordered;
+
 	add_to_unconfirmed(new);
 	swap(new, conf);
 out_copy:
@@ -3436,12 +3645,8 @@ out_nolock:
 	return status;
 }
 
-static __be32
-check_slot_seqid(u32 seqid, u32 slot_seqid, int slot_inuse)
+static __be32 check_slot_seqid(u32 seqid, u32 slot_seqid, bool slot_inuse)
 {
-	dprintk("%s enter. seqid %d slot_seqid %d\n", __func__, seqid,
-		slot_seqid);
-
 	/* The slot is in use, and no response has been sent. */
 	if (slot_inuse) {
 		if (seqid == slot_seqid)
@@ -3581,10 +3786,10 @@ nfsd4_create_session(struct svc_rqst *rqstp,
 	struct nfsd4_create_session *cr_ses = &u->create_session;
 	struct sockaddr *sa = svc_addr(rqstp);
 	struct nfs4_client *conf, *unconf;
+	struct nfsd4_clid_slot *cs_slot;
 	struct nfs4_client *old = NULL;
 	struct nfsd4_session *new;
 	struct nfsd4_conn *conn;
-	struct nfsd4_clid_slot *cs_slot = NULL;
 	__be32 status = 0;
 	struct nfsd_net *nn = net_generic(SVC_NET(rqstp), nfsd_net_id);
 
@@ -3608,53 +3813,66 @@ nfsd4_create_session(struct svc_rqst *rqstp,
 		goto out_free_session;
 
 	spin_lock(&nn->client_lock);
+
+	/* RFC 8881 Section 18.36.4 Phase 1: Client record look-up. */
 	unconf = find_unconfirmed_client(&cr_ses->clientid, true, nn);
 	conf = find_confirmed_client(&cr_ses->clientid, true, nn);
-	WARN_ON_ONCE(conf && unconf);
+	if (!conf && !unconf) {
+		status = nfserr_stale_clientid;
+		goto out_free_conn;
+	}
+
+	/* RFC 8881 Section 18.36.4 Phase 2: Sequence ID processing. */
+	if (conf) {
+		cs_slot = &conf->cl_cs_slot;
+		trace_nfsd_slot_seqid_conf(conf, cr_ses);
+	} else {
+		cs_slot = &unconf->cl_cs_slot;
+		trace_nfsd_slot_seqid_unconf(unconf, cr_ses);
+	}
+	status = check_slot_seqid(cr_ses->seqid, cs_slot->sl_seqid, 0);
+	switch (status) {
+	case nfs_ok:
+		cs_slot->sl_seqid++;
+		cr_ses->seqid = cs_slot->sl_seqid;
+		break;
+	case nfserr_replay_cache:
+		status = nfsd4_replay_create_session(cr_ses, cs_slot);
+		fallthrough;
+	case nfserr_jukebox:
+		/* The server MUST NOT cache NFS4ERR_DELAY */
+		goto out_free_conn;
+	default:
+		goto out_cache_error;
+	}
 
+	/* RFC 8881 Section 18.36.4 Phase 3: Client ID confirmation. */
 	if (conf) {
 		status = nfserr_wrong_cred;
 		if (!nfsd4_mach_creds_match(conf, rqstp))
-			goto out_free_conn;
-		cs_slot = &conf->cl_cs_slot;
-		status = check_slot_seqid(cr_ses->seqid, cs_slot->sl_seqid, 0);
-		if (status) {
-			if (status == nfserr_replay_cache)
-				status = nfsd4_replay_create_session(cr_ses, cs_slot);
-			goto out_free_conn;
-		}
-	} else if (unconf) {
+			goto out_cache_error;
+	} else {
 		status = nfserr_clid_inuse;
 		if (!same_creds(&unconf->cl_cred, &rqstp->rq_cred) ||
 		    !rpc_cmp_addr(sa, (struct sockaddr *) &unconf->cl_addr)) {
 			trace_nfsd_clid_cred_mismatch(unconf, rqstp);
-			goto out_free_conn;
+			goto out_cache_error;
 		}
 		status = nfserr_wrong_cred;
 		if (!nfsd4_mach_creds_match(unconf, rqstp))
-			goto out_free_conn;
-		cs_slot = &unconf->cl_cs_slot;
-		status = check_slot_seqid(cr_ses->seqid, cs_slot->sl_seqid, 0);
-		if (status) {
-			/* an unconfirmed replay returns misordered */
-			status = nfserr_seq_misordered;
-			goto out_free_conn;
-		}
+			goto out_cache_error;
 		old = find_confirmed_client_by_name(&unconf->cl_name, nn);
 		if (old) {
 			status = mark_client_expired_locked(old);
-			if (status) {
-				old = NULL;
-				goto out_free_conn;
-			}
+			if (status)
+				goto out_expired_error;
 			trace_nfsd_clid_replaced(&old->cl_clientid);
 		}
 		move_to_confirmed(unconf);
 		conf = unconf;
-	} else {
-		status = nfserr_stale_clientid;
-		goto out_free_conn;
 	}
+
+	/* RFC 8881 Section 18.36.4 Phase 4: Session creation. */
 	status = nfs_ok;
 	/* Persistent sessions are not supported */
 	cr_ses->flags &= ~SESSION4_PERSIST;
@@ -3666,8 +3884,6 @@ nfsd4_create_session(struct svc_rqst *rqstp,
 
 	memcpy(cr_ses->sessionid.data, new->se_sessionid.data,
 	       NFS4_MAX_SESSIONID_LEN);
-	cs_slot->sl_seqid++;
-	cr_ses->seqid = cs_slot->sl_seqid;
 
 	/* cache solo and embedded create sessions under the client_lock */
 	nfsd4_cache_create_session(cr_ses, cs_slot, status);
@@ -3680,6 +3896,20 @@ nfsd4_create_session(struct svc_rqst *rqstp,
 	if (old)
 		expire_client(old);
 	return status;
+
+out_expired_error:
+	old = NULL;
+	/*
+	 * Revert the slot seq_nr change so the server will process
+	 * the client's resend instead of returning a cached response.
+	 */
+	if (status == nfserr_jukebox) {
+		cs_slot->sl_seqid--;
+		cr_ses->seqid = cs_slot->sl_seqid;
+		goto out_free_conn;
+	}
+out_cache_error:
+	nfsd4_cache_create_session(cr_ses, cs_slot, status);
 out_free_conn:
 	spin_unlock(&nn->client_lock);
 	free_conn(conn);
@@ -3994,6 +4224,7 @@ nfsd4_sequence(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	 * sr_highest_slotid and the sr_target_slot id to maxslots */
 	seq->maxslots = session->se_fchannel.maxreqs;
 
+	trace_nfsd_slot_seqid_sequence(clp, seq, slot);
 	status = check_slot_seqid(seq->seqid, slot->sl_seqid,
 					slot->sl_flags & NFSD4_SLOT_INUSE);
 	if (status == nfserr_replay_cache) {
@@ -4055,6 +4286,9 @@ out:
 	}
 	if (!list_empty(&clp->cl_revoked))
 		seq->status_flags |= SEQ4_STATUS_RECALLABLE_STATE_REVOKED;
+	if (atomic_read(&clp->cl_admin_revoked))
+		seq->status_flags |= SEQ4_STATUS_ADMIN_STATE_REVOKED;
+	trace_nfsd_seq4_status(rqstp, seq);
 out_no_session:
 	if (conn)
 		free_conn(conn);
@@ -4349,32 +4583,25 @@ nfsd4_free_slabs(void)
 int
 nfsd4_init_slabs(void)
 {
-	client_slab = kmem_cache_create("nfsd4_clients",
-			sizeof(struct nfs4_client), 0, 0, NULL);
+	client_slab = KMEM_CACHE(nfs4_client, 0);
 	if (client_slab == NULL)
 		goto out;
-	openowner_slab = kmem_cache_create("nfsd4_openowners",
-			sizeof(struct nfs4_openowner), 0, 0, NULL);
+	openowner_slab = KMEM_CACHE(nfs4_openowner, 0);
 	if (openowner_slab == NULL)
 		goto out_free_client_slab;
-	lockowner_slab = kmem_cache_create("nfsd4_lockowners",
-			sizeof(struct nfs4_lockowner), 0, 0, NULL);
+	lockowner_slab = KMEM_CACHE(nfs4_lockowner, 0);
 	if (lockowner_slab == NULL)
 		goto out_free_openowner_slab;
-	file_slab = kmem_cache_create("nfsd4_files",
-			sizeof(struct nfs4_file), 0, 0, NULL);
+	file_slab = KMEM_CACHE(nfs4_file, 0);
 	if (file_slab == NULL)
 		goto out_free_lockowner_slab;
-	stateid_slab = kmem_cache_create("nfsd4_stateids",
-			sizeof(struct nfs4_ol_stateid), 0, 0, NULL);
+	stateid_slab = KMEM_CACHE(nfs4_ol_stateid, 0);
 	if (stateid_slab == NULL)
 		goto out_free_file_slab;
-	deleg_slab = kmem_cache_create("nfsd4_delegations",
-			sizeof(struct nfs4_delegation), 0, 0, NULL);
+	deleg_slab = KMEM_CACHE(nfs4_delegation, 0);
 	if (deleg_slab == NULL)
 		goto out_free_stateid_slab;
-	odstate_slab = kmem_cache_create("nfsd4_odstate",
-			sizeof(struct nfs4_clnt_odstate), 0, 0, NULL);
+	odstate_slab = KMEM_CACHE(nfs4_clnt_odstate, 0);
 	if (odstate_slab == NULL)
 		goto out_free_deleg_slab;
 	return 0;
@@ -4439,21 +4666,32 @@ nfsd4_init_leases_net(struct nfsd_net *nn)
 	atomic_set(&nn->nfsd_courtesy_clients, 0);
 }
 
+enum rp_lock {
+	RP_UNLOCKED,
+	RP_LOCKED,
+	RP_UNHASHED,
+};
+
 static void init_nfs4_replay(struct nfs4_replay *rp)
 {
 	rp->rp_status = nfserr_serverfault;
 	rp->rp_buflen = 0;
 	rp->rp_buf = rp->rp_ibuf;
-	mutex_init(&rp->rp_mutex);
+	atomic_set(&rp->rp_locked, RP_UNLOCKED);
 }
 
-static void nfsd4_cstate_assign_replay(struct nfsd4_compound_state *cstate,
-		struct nfs4_stateowner *so)
+static int nfsd4_cstate_assign_replay(struct nfsd4_compound_state *cstate,
+				      struct nfs4_stateowner *so)
 {
 	if (!nfsd4_has_session(cstate)) {
-		mutex_lock(&so->so_replay.rp_mutex);
+		wait_var_event(&so->so_replay.rp_locked,
+			       atomic_cmpxchg(&so->so_replay.rp_locked,
+					      RP_UNLOCKED, RP_LOCKED) != RP_LOCKED);
+		if (atomic_read(&so->so_replay.rp_locked) == RP_UNHASHED)
+			return -EAGAIN;
 		cstate->replay_owner = nfs4_get_stateowner(so);
 	}
+	return 0;
 }
 
 void nfsd4_cstate_clear_replay(struct nfsd4_compound_state *cstate)
@@ -4462,7 +4700,8 @@ void nfsd4_cstate_clear_replay(struct nfsd4_compound_state *cstate)
 
 	if (so != NULL) {
 		cstate->replay_owner = NULL;
-		mutex_unlock(&so->so_replay.rp_mutex);
+		atomic_set(&so->so_replay.rp_locked, RP_UNLOCKED);
+		wake_up_var(&so->so_replay.rp_locked);
 		nfs4_put_stateowner(so);
 	}
 }
@@ -4528,7 +4767,8 @@ nfsd4_find_existing_open(struct nfs4_file *fp, struct nfsd4_open *open)
 			continue;
 		if (local->st_stateowner != &oo->oo_owner)
 			continue;
-		if (local->st_stid.sc_type == NFS4_OPEN_STID) {
+		if (local->st_stid.sc_type == SC_TYPE_OPEN &&
+		    !local->st_stid.sc_status) {
 			ret = local;
 			refcount_inc(&ret->st_stid.sc_count);
 			break;
@@ -4537,22 +4777,75 @@ nfsd4_find_existing_open(struct nfs4_file *fp, struct nfsd4_open *open)
 	return ret;
 }
 
-static __be32
-nfsd4_verify_open_stid(struct nfs4_stid *s)
+static void nfsd4_drop_revoked_stid(struct nfs4_stid *s)
+	__releases(&s->sc_client->cl_lock)
 {
-	__be32 ret = nfs_ok;
+	struct nfs4_client *cl = s->sc_client;
+	LIST_HEAD(reaplist);
+	struct nfs4_ol_stateid *stp;
+	struct nfs4_delegation *dp;
+	bool unhashed;
 
 	switch (s->sc_type) {
-	default:
+	case SC_TYPE_OPEN:
+		stp = openlockstateid(s);
+		if (unhash_open_stateid(stp, &reaplist))
+			put_ol_stateid_locked(stp, &reaplist);
+		spin_unlock(&cl->cl_lock);
+		free_ol_stateid_reaplist(&reaplist);
 		break;
-	case 0:
-	case NFS4_CLOSED_STID:
-	case NFS4_CLOSED_DELEG_STID:
-		ret = nfserr_bad_stateid;
+	case SC_TYPE_LOCK:
+		stp = openlockstateid(s);
+		unhashed = unhash_lock_stateid(stp);
+		spin_unlock(&cl->cl_lock);
+		if (unhashed)
+			nfs4_put_stid(s);
 		break;
-	case NFS4_REVOKED_DELEG_STID:
-		ret = nfserr_deleg_revoked;
+	case SC_TYPE_DELEG:
+		dp = delegstateid(s);
+		list_del_init(&dp->dl_recall_lru);
+		spin_unlock(&cl->cl_lock);
+		nfs4_put_stid(s);
+		break;
+	default:
+		spin_unlock(&cl->cl_lock);
+	}
+}
+
+static void nfsd40_drop_revoked_stid(struct nfs4_client *cl,
+				    stateid_t *stid)
+{
+	/* NFSv4.0 has no way for the client to tell the server
+	 * that it can forget an admin-revoked stateid.
+	 * So we keep it around until the first time that the
+	 * client uses it, and drop it the first time
+	 * nfserr_admin_revoked is returned.
+	 * For v4.1 and later we wait until explicitly told
+	 * to free the stateid.
+	 */
+	if (cl->cl_minorversion == 0) {
+		struct nfs4_stid *st;
+
+		spin_lock(&cl->cl_lock);
+		st = find_stateid_locked(cl, stid);
+		if (st)
+			nfsd4_drop_revoked_stid(st);
+		else
+			spin_unlock(&cl->cl_lock);
 	}
+}
+
+static __be32
+nfsd4_verify_open_stid(struct nfs4_stid *s)
+{
+	__be32 ret = nfs_ok;
+
+	if (s->sc_status & SC_STATUS_ADMIN_REVOKED)
+		ret = nfserr_admin_revoked;
+	else if (s->sc_status & SC_STATUS_REVOKED)
+		ret = nfserr_deleg_revoked;
+	else if (s->sc_status & SC_STATUS_CLOSED)
+		ret = nfserr_bad_stateid;
 	return ret;
 }
 
@@ -4564,6 +4857,10 @@ nfsd4_lock_ol_stateid(struct nfs4_ol_stateid *stp)
 
 	mutex_lock_nested(&stp->st_mutex, LOCK_STATEID_MUTEX);
 	ret = nfsd4_verify_open_stid(&stp->st_stid);
+	if (ret == nfserr_admin_revoked)
+		nfsd40_drop_revoked_stid(stp->st_stid.sc_client,
+					&stp->st_stid.sc_stateid);
+
 	if (ret != nfs_ok)
 		mutex_unlock(&stp->st_mutex);
 	return ret;
@@ -4585,34 +4882,46 @@ nfsd4_find_and_lock_existing_open(struct nfs4_file *fp, struct nfsd4_open *open)
 }
 
 static struct nfs4_openowner *
-alloc_init_open_stateowner(unsigned int strhashval, struct nfsd4_open *open,
-			   struct nfsd4_compound_state *cstate)
+find_or_alloc_open_stateowner(unsigned int strhashval, struct nfsd4_open *open,
+			      struct nfsd4_compound_state *cstate)
 {
 	struct nfs4_client *clp = cstate->clp;
-	struct nfs4_openowner *oo, *ret;
+	struct nfs4_openowner *oo, *new = NULL;
 
-	oo = alloc_stateowner(openowner_slab, &open->op_owner, clp);
-	if (!oo)
-		return NULL;
-	oo->oo_owner.so_ops = &openowner_ops;
-	oo->oo_owner.so_is_open_owner = 1;
-	oo->oo_owner.so_seqid = open->op_seqid;
-	oo->oo_flags = 0;
-	if (nfsd4_has_session(cstate))
-		oo->oo_flags |= NFS4_OO_CONFIRMED;
-	oo->oo_time = 0;
-	oo->oo_last_closed_stid = NULL;
-	INIT_LIST_HEAD(&oo->oo_close_lru);
+retry:
 	spin_lock(&clp->cl_lock);
-	ret = find_openstateowner_str_locked(strhashval, open, clp);
-	if (ret == NULL) {
-		hash_openowner(oo, clp, strhashval);
-		ret = oo;
-	} else
-		nfs4_free_stateowner(&oo->oo_owner);
-
+	oo = find_openstateowner_str(strhashval, open, clp);
+	if (!oo && new) {
+		hash_openowner(new, clp, strhashval);
+		spin_unlock(&clp->cl_lock);
+		return new;
+	}
 	spin_unlock(&clp->cl_lock);
-	return ret;
+
+	if (oo && !(oo->oo_flags & NFS4_OO_CONFIRMED)) {
+		/* Replace unconfirmed owners without checking for replay. */
+		release_openowner(oo);
+		oo = NULL;
+	}
+	if (oo) {
+		if (new)
+			nfs4_free_stateowner(&new->oo_owner);
+		return oo;
+	}
+
+	new = alloc_stateowner(openowner_slab, &open->op_owner, clp);
+	if (!new)
+		return NULL;
+	new->oo_owner.so_ops = &openowner_ops;
+	new->oo_owner.so_is_open_owner = 1;
+	new->oo_owner.so_seqid = open->op_seqid;
+	new->oo_flags = 0;
+	if (nfsd4_has_session(cstate))
+		new->oo_flags |= NFS4_OO_CONFIRMED;
+	new->oo_time = 0;
+	new->oo_last_closed_stid = NULL;
+	INIT_LIST_HEAD(&new->oo_close_lru);
+	goto retry;
 }
 
 static struct nfs4_ol_stateid *
@@ -4638,7 +4947,7 @@ retry:
 
 	open->op_stp = NULL;
 	refcount_inc(&stp->st_stid.sc_count);
-	stp->st_stid.sc_type = NFS4_OPEN_STID;
+	stp->st_stid.sc_type = SC_TYPE_OPEN;
 	INIT_LIST_HEAD(&stp->st_locks);
 	stp->st_stateowner = nfs4_get_stateowner(&oo->oo_owner);
 	get_nfs4_file(fp);
@@ -4688,7 +4997,11 @@ move_to_close_lru(struct nfs4_ol_stateid *s, struct net *net)
 	 * Wait for the refcount to drop to 2. Since it has been unhashed,
 	 * there should be no danger of the refcount going back up again at
 	 * this point.
+	 * Some threads with a reference might be waiting for rp_locked,
+	 * so tell them to stop waiting.
 	 */
+	atomic_set(&oo->oo_owner.so_replay.rp_locked, RP_UNHASHED);
+	wake_up_var(&oo->oo_owner.so_replay.rp_locked);
 	wait_event(close_wq, refcount_read(&s->st_stid.sc_count) == 2);
 
 	release_all_access(s);
@@ -4865,9 +5178,9 @@ static int nfsd4_cb_recall_done(struct nfsd4_callback *cb,
 
 	trace_nfsd_cb_recall_done(&dp->dl_stid.sc_stateid, task);
 
-	if (dp->dl_stid.sc_type == NFS4_CLOSED_DELEG_STID ||
-	    dp->dl_stid.sc_type == NFS4_REVOKED_DELEG_STID)
-	        return 1;
+	if (dp->dl_stid.sc_status)
+		/* CLOSED or REVOKED */
+		return 1;
 
 	switch (task->tk_status) {
 	case 0:
@@ -4919,9 +5232,9 @@ static void nfsd_break_one_deleg(struct nfs4_delegation *dp)
 
 /* Called from break_lease() with flc_lock held. */
 static bool
-nfsd_break_deleg_cb(struct file_lock *fl)
+nfsd_break_deleg_cb(struct file_lease *fl)
 {
-	struct nfs4_delegation *dp = (struct nfs4_delegation *)fl->fl_owner;
+	struct nfs4_delegation *dp = (struct nfs4_delegation *) fl->c.flc_owner;
 	struct nfs4_file *fp = dp->dl_stid.sc_file;
 	struct nfs4_client *clp = dp->dl_stid.sc_client;
 	struct nfsd_net *nn;
@@ -4955,9 +5268,9 @@ nfsd_break_deleg_cb(struct file_lock *fl)
  *   %true: Lease conflict was resolved
  *   %false: Lease conflict was not resolved.
  */
-static bool nfsd_breaker_owns_lease(struct file_lock *fl)
+static bool nfsd_breaker_owns_lease(struct file_lease *fl)
 {
-	struct nfs4_delegation *dl = fl->fl_owner;
+	struct nfs4_delegation *dl = fl->c.flc_owner;
 	struct svc_rqst *rqst;
 	struct nfs4_client *clp;
 
@@ -4972,10 +5285,10 @@ static bool nfsd_breaker_owns_lease(struct file_lock *fl)
 }
 
 static int
-nfsd_change_deleg_cb(struct file_lock *onlist, int arg,
+nfsd_change_deleg_cb(struct file_lease *onlist, int arg,
 		     struct list_head *dispose)
 {
-	struct nfs4_delegation *dp = (struct nfs4_delegation *)onlist->fl_owner;
+	struct nfs4_delegation *dp = (struct nfs4_delegation *) onlist->c.flc_owner;
 	struct nfs4_client *clp = dp->dl_stid.sc_client;
 
 	if (arg & F_UNLCK) {
@@ -4986,7 +5299,7 @@ nfsd_change_deleg_cb(struct file_lock *onlist, int arg,
 		return -EAGAIN;
 }
 
-static const struct lock_manager_operations nfsd_lease_mng_ops = {
+static const struct lease_manager_operations nfsd_lease_mng_ops = {
 	.lm_breaker_owns_lease = nfsd_breaker_owns_lease,
 	.lm_break = nfsd_break_deleg_cb,
 	.lm_change = nfsd_change_deleg_cb,
@@ -5061,27 +5374,19 @@ nfsd4_process_open1(struct nfsd4_compound_state *cstate,
 	clp = cstate->clp;
 
 	strhashval = ownerstr_hashval(&open->op_owner);
-	oo = find_openstateowner_str(strhashval, open, clp);
+retry:
+	oo = find_or_alloc_open_stateowner(strhashval, open, cstate);
 	open->op_openowner = oo;
-	if (!oo) {
-		goto new_owner;
-	}
-	if (!(oo->oo_flags & NFS4_OO_CONFIRMED)) {
-		/* Replace unconfirmed owners without checking for replay. */
-		release_openowner(oo);
-		open->op_openowner = NULL;
-		goto new_owner;
+	if (!oo)
+		return nfserr_jukebox;
+	if (nfsd4_cstate_assign_replay(cstate, &oo->oo_owner) == -EAGAIN) {
+		nfs4_put_stateowner(&oo->oo_owner);
+		goto retry;
 	}
 	status = nfsd4_check_seqid(cstate, &oo->oo_owner, open->op_seqid);
 	if (status)
 		return status;
-	goto alloc_stateid;
-new_owner:
-	oo = alloc_init_open_stateowner(strhashval, open, cstate);
-	if (oo == NULL)
-		return nfserr_jukebox;
-	open->op_openowner = oo;
-alloc_stateid:
+
 	open->op_stp = nfs4_alloc_open_stateid(clp);
 	if (!open->op_stp)
 		return nfserr_jukebox;
@@ -5110,12 +5415,12 @@ static int share_access_to_flags(u32 share_access)
 	return share_access == NFS4_SHARE_ACCESS_READ ? RD_STATE : WR_STATE;
 }
 
-static struct nfs4_delegation *find_deleg_stateid(struct nfs4_client *cl, stateid_t *s)
+static struct nfs4_delegation *find_deleg_stateid(struct nfs4_client *cl,
+						  stateid_t *s)
 {
 	struct nfs4_stid *ret;
 
-	ret = find_stateid_by_type(cl, s,
-				NFS4_DELEG_STID|NFS4_REVOKED_DELEG_STID);
+	ret = find_stateid_by_type(cl, s, SC_TYPE_DELEG, SC_STATUS_REVOKED);
 	if (!ret)
 		return NULL;
 	return delegstateid(ret);
@@ -5138,10 +5443,15 @@ nfs4_check_deleg(struct nfs4_client *cl, struct nfsd4_open *open,
 	deleg = find_deleg_stateid(cl, &open->op_delegate_stateid);
 	if (deleg == NULL)
 		goto out;
-	if (deleg->dl_stid.sc_type == NFS4_REVOKED_DELEG_STID) {
+	if (deleg->dl_stid.sc_status & SC_STATUS_ADMIN_REVOKED) {
 		nfs4_put_stid(&deleg->dl_stid);
-		if (cl->cl_minorversion)
-			status = nfserr_deleg_revoked;
+		status = nfserr_admin_revoked;
+		goto out;
+	}
+	if (deleg->dl_stid.sc_status & SC_STATUS_REVOKED) {
+		nfs4_put_stid(&deleg->dl_stid);
+		nfsd40_drop_revoked_stid(cl, &open->op_delegate_stateid);
+		status = nfserr_deleg_revoked;
 		goto out;
 	}
 	flags = share_access_to_flags(open->op_share_access);
@@ -5186,7 +5496,7 @@ nfsd4_truncate(struct svc_rqst *rqstp, struct svc_fh *fh,
 		return 0;
 	if (!(open->op_share_access & NFS4_SHARE_ACCESS_WRITE))
 		return nfserr_inval;
-	return nfsd_setattr(rqstp, fh, &attrs, 0, (time64_t)0);
+	return nfsd_setattr(rqstp, fh, &attrs, NULL);
 }
 
 static __be32 nfs4_get_vfs_file(struct svc_rqst *rqstp, struct nfs4_file *fp,
@@ -5326,21 +5636,20 @@ static bool nfsd4_cb_channel_good(struct nfs4_client *clp)
 	return clp->cl_minorversion && clp->cl_cb_state == NFSD4_CB_UNKNOWN;
 }
 
-static struct file_lock *nfs4_alloc_init_lease(struct nfs4_delegation *dp,
+static struct file_lease *nfs4_alloc_init_lease(struct nfs4_delegation *dp,
 						int flag)
 {
-	struct file_lock *fl;
+	struct file_lease *fl;
 
-	fl = locks_alloc_lock();
+	fl = locks_alloc_lease();
 	if (!fl)
 		return NULL;
 	fl->fl_lmops = &nfsd_lease_mng_ops;
-	fl->fl_flags = FL_DELEG;
-	fl->fl_type = flag == NFS4_OPEN_DELEGATE_READ? F_RDLCK: F_WRLCK;
-	fl->fl_end = OFFSET_MAX;
-	fl->fl_owner = (fl_owner_t)dp;
-	fl->fl_pid = current->tgid;
-	fl->fl_file = dp->dl_stid.sc_file->fi_deleg_file->nf_file;
+	fl->c.flc_flags = FL_DELEG;
+	fl->c.flc_type = flag == NFS4_OPEN_DELEGATE_READ? F_RDLCK: F_WRLCK;
+	fl->c.flc_owner = (fl_owner_t)dp;
+	fl->c.flc_pid = current->tgid;
+	fl->c.flc_file = dp->dl_stid.sc_file->fi_deleg_file->nf_file;
 	return fl;
 }
 
@@ -5458,7 +5767,7 @@ nfs4_set_delegation(struct nfsd4_open *open, struct nfs4_ol_stateid *stp,
 	struct nfs4_clnt_odstate *odstate = stp->st_clnt_odstate;
 	struct nfs4_delegation *dp;
 	struct nfsd_file *nf = NULL;
-	struct file_lock *fl;
+	struct file_lease *fl;
 	u32 dl_type;
 
 	/*
@@ -5528,9 +5837,10 @@ nfs4_set_delegation(struct nfsd4_open *open, struct nfs4_ol_stateid *stp,
 	if (!fl)
 		goto out_clnt_odstate;
 
-	status = vfs_setlease(fp->fi_deleg_file->nf_file, fl->fl_type, &fl, NULL);
+	status = kernel_setlease(fp->fi_deleg_file->nf_file,
+				      fl->c.flc_type, &fl, NULL);
 	if (fl)
-		locks_free_lock(fl);
+		locks_free_lease(fl);
 	if (status)
 		goto out_clnt_odstate;
 
@@ -5557,9 +5867,11 @@ nfs4_set_delegation(struct nfsd4_open *open, struct nfs4_ol_stateid *stp,
 		goto out_unlock;
 
 	spin_lock(&state_lock);
+	spin_lock(&clp->cl_lock);
 	spin_lock(&fp->fi_lock);
 	status = hash_delegation_locked(dp, fp);
 	spin_unlock(&fp->fi_lock);
+	spin_unlock(&clp->cl_lock);
 	spin_unlock(&state_lock);
 
 	if (status)
@@ -5567,7 +5879,7 @@ nfs4_set_delegation(struct nfsd4_open *open, struct nfs4_ol_stateid *stp,
 
 	return dp;
 out_unlock:
-	vfs_setlease(fp->fi_deleg_file->nf_file, F_UNLCK, NULL, (void **)&dp);
+	kernel_setlease(fp->fi_deleg_file->nf_file, F_UNLCK, NULL, (void **)&dp);
 out_clnt_odstate:
 	put_clnt_odstate(dp->dl_clnt_odstate);
 	nfs4_put_stid(&dp->dl_stid);
@@ -5631,6 +5943,8 @@ nfs4_open_delegation(struct nfsd4_open *open, struct nfs4_ol_stateid *stp,
 	struct svc_fh *parent = NULL;
 	int cb_up;
 	int status = 0;
+	struct kstat stat;
+	struct path path;
 
 	cb_up = nfsd4_cb_channel_good(oo->oo_owner.so_client);
 	open->op_recall = false;
@@ -5668,6 +5982,18 @@ nfs4_open_delegation(struct nfsd4_open *open, struct nfs4_ol_stateid *stp,
 	if (open->op_share_access & NFS4_SHARE_ACCESS_WRITE) {
 		open->op_delegate_type = NFS4_OPEN_DELEGATE_WRITE;
 		trace_nfsd_deleg_write(&dp->dl_stid.sc_stateid);
+		path.mnt = currentfh->fh_export->ex_path.mnt;
+		path.dentry = currentfh->fh_dentry;
+		if (vfs_getattr(&path, &stat,
+				(STATX_SIZE | STATX_CTIME | STATX_CHANGE_COOKIE),
+				AT_STATX_SYNC_AS_STAT)) {
+			nfs4_put_stid(&dp->dl_stid);
+			destroy_delegation(dp);
+			goto out_no_deleg;
+		}
+		dp->dl_cb_fattr.ncf_cur_fsize = stat.size;
+		dp->dl_cb_fattr.ncf_initial_cinfo =
+			nfsd4_change_attribute(&stat, d_inode(currentfh->fh_dentry));
 	} else {
 		open->op_delegate_type = NFS4_OPEN_DELEGATE_READ;
 		trace_nfsd_deleg_read(&dp->dl_stid.sc_stateid);
@@ -5771,7 +6097,6 @@ nfsd4_process_open2(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nf
 	} else {
 		status = nfs4_get_vfs_file(rqstp, fp, current_fh, stp, open, true);
 		if (status) {
-			stp->st_stid.sc_type = NFS4_CLOSED_STID;
 			release_open_stateid(stp);
 			mutex_unlock(&stp->st_mutex);
 			goto out;
@@ -5832,12 +6157,8 @@ out:
 void nfsd4_cleanup_open_state(struct nfsd4_compound_state *cstate,
 			      struct nfsd4_open *open)
 {
-	if (open->op_openowner) {
-		struct nfs4_stateowner *so = &open->op_openowner->oo_owner;
-
-		nfsd4_cstate_assign_replay(cstate, so);
-		nfs4_put_stateowner(so);
-	}
+	if (open->op_openowner)
+		nfs4_put_stateowner(&open->op_openowner->oo_owner);
 	if (open->op_file)
 		kmem_cache_free(file_slab, open->op_file);
 	if (open->op_stp)
@@ -6125,6 +6446,43 @@ nfs4_process_client_reaplist(struct list_head *reaplist)
 	}
 }
 
+static void nfs40_clean_admin_revoked(struct nfsd_net *nn,
+				      struct laundry_time *lt)
+{
+	struct nfs4_client *clp;
+
+	spin_lock(&nn->client_lock);
+	if (nn->nfs40_last_revoke == 0 ||
+	    nn->nfs40_last_revoke > lt->cutoff) {
+		spin_unlock(&nn->client_lock);
+		return;
+	}
+	nn->nfs40_last_revoke = 0;
+
+retry:
+	list_for_each_entry(clp, &nn->client_lru, cl_lru) {
+		unsigned long id, tmp;
+		struct nfs4_stid *stid;
+
+		if (atomic_read(&clp->cl_admin_revoked) == 0)
+			continue;
+
+		spin_lock(&clp->cl_lock);
+		idr_for_each_entry_ul(&clp->cl_stateids, stid, tmp, id)
+			if (stid->sc_status & SC_STATUS_ADMIN_REVOKED) {
+				refcount_inc(&stid->sc_count);
+				spin_unlock(&nn->client_lock);
+				/* this function drops ->cl_lock */
+				nfsd4_drop_revoked_stid(stid);
+				nfs4_put_stid(stid);
+				spin_lock(&nn->client_lock);
+				goto retry;
+			}
+		spin_unlock(&clp->cl_lock);
+	}
+	spin_unlock(&nn->client_lock);
+}
+
 static time64_t
 nfs4_laundromat(struct nfsd_net *nn)
 {
@@ -6158,12 +6516,14 @@ nfs4_laundromat(struct nfsd_net *nn)
 	nfs4_get_client_reaplist(nn, &reaplist, &lt);
 	nfs4_process_client_reaplist(&reaplist);
 
+	nfs40_clean_admin_revoked(nn, &lt);
+
 	spin_lock(&state_lock);
 	list_for_each_safe(pos, next, &nn->del_recall_lru) {
 		dp = list_entry (pos, struct nfs4_delegation, dl_recall_lru);
 		if (!state_expired(&lt, dp->dl_time))
 			break;
-		WARN_ON(!unhash_delegation_locked(dp));
+		unhash_delegation_locked(dp, SC_STATUS_REVOKED);
 		list_add(&dp->dl_recall_lru, &reaplist);
 	}
 	spin_unlock(&state_lock);
@@ -6222,6 +6582,8 @@ nfs4_laundromat(struct nfsd_net *nn)
 	/* service the server-to-server copy delayed unmount list */
 	nfsd4_ssc_expire_umount(nn);
 #endif
+	if (atomic_long_read(&num_delegations) >= max_delegations)
+		deleg_reaper(nn);
 out:
 	return max_t(time64_t, lt.new_timeo, NFSD_LAUNDROMAT_MINTIMEOUT);
 }
@@ -6283,6 +6645,8 @@ deleg_reaper(struct nfsd_net *nn)
 		list_del_init(&clp->cl_ra_cblist);
 		clp->cl_ra->ra_keep = 0;
 		clp->cl_ra->ra_bmval[0] = BIT(RCA4_TYPE_MASK_RDATA_DLG);
+		clp->cl_ra->ra_bmval[0] = BIT(RCA4_TYPE_MASK_RDATA_DLG) |
+						BIT(RCA4_TYPE_MASK_WDATA_DLG);
 		trace_nfsd_cb_recall_any(clp->cl_ra);
 		nfsd4_run_cb(&clp->cl_ra->ra_cb);
 	}
@@ -6376,6 +6740,9 @@ static __be32 nfsd4_stid_check_stateid_generation(stateid_t *in, struct nfs4_sti
 	if (ret == nfs_ok)
 		ret = check_stateid_generation(in, &s->sc_stateid, has_session);
 	spin_unlock(&s->sc_lock);
+	if (ret == nfserr_admin_revoked)
+		nfsd40_drop_revoked_stid(s->sc_client,
+					&s->sc_stateid);
 	return ret;
 }
 
@@ -6402,32 +6769,33 @@ static __be32 nfsd4_validate_stateid(struct nfs4_client *cl, stateid_t *stateid)
 	status = nfsd4_stid_check_stateid_generation(stateid, s, 1);
 	if (status)
 		goto out_unlock;
+	status = nfsd4_verify_open_stid(s);
+	if (status)
+		goto out_unlock;
+
 	switch (s->sc_type) {
-	case NFS4_DELEG_STID:
+	case SC_TYPE_DELEG:
 		status = nfs_ok;
 		break;
-	case NFS4_REVOKED_DELEG_STID:
-		status = nfserr_deleg_revoked;
-		break;
-	case NFS4_OPEN_STID:
-	case NFS4_LOCK_STID:
+	case SC_TYPE_OPEN:
+	case SC_TYPE_LOCK:
 		status = nfsd4_check_openowner_confirmed(openlockstateid(s));
 		break;
 	default:
 		printk("unknown stateid type %x\n", s->sc_type);
-		fallthrough;
-	case NFS4_CLOSED_STID:
-	case NFS4_CLOSED_DELEG_STID:
 		status = nfserr_bad_stateid;
 	}
 out_unlock:
 	spin_unlock(&cl->cl_lock);
+	if (status == nfserr_admin_revoked)
+		nfsd40_drop_revoked_stid(cl, stateid);
 	return status;
 }
 
 __be32
 nfsd4_lookup_stateid(struct nfsd4_compound_state *cstate,
-		     stateid_t *stateid, unsigned char typemask,
+		     stateid_t *stateid,
+		     unsigned short typemask, unsigned short statusmask,
 		     struct nfs4_stid **s, struct nfsd_net *nn)
 {
 	__be32 status;
@@ -6438,10 +6806,15 @@ nfsd4_lookup_stateid(struct nfsd4_compound_state *cstate,
 	 *  only return revoked delegations if explicitly asked.
 	 *  otherwise we report revoked or bad_stateid status.
 	 */
-	if (typemask & NFS4_REVOKED_DELEG_STID)
+	if (statusmask & SC_STATUS_REVOKED)
 		return_revoked = true;
-	else if (typemask & NFS4_DELEG_STID)
-		typemask |= NFS4_REVOKED_DELEG_STID;
+	if (typemask & SC_TYPE_DELEG)
+		/* Always allow REVOKED for DELEG so we can
+		 * retturn the appropriate error.
+		 */
+		statusmask |= SC_STATUS_REVOKED;
+
+	statusmask |= SC_STATUS_ADMIN_REVOKED;
 
 	if (ZERO_STATEID(stateid) || ONE_STATEID(stateid) ||
 		CLOSE_STATEID(stateid))
@@ -6454,14 +6827,17 @@ nfsd4_lookup_stateid(struct nfsd4_compound_state *cstate,
 	}
 	if (status)
 		return status;
-	stid = find_stateid_by_type(cstate->clp, stateid, typemask);
+	stid = find_stateid_by_type(cstate->clp, stateid, typemask, statusmask);
 	if (!stid)
 		return nfserr_bad_stateid;
-	if ((stid->sc_type == NFS4_REVOKED_DELEG_STID) && !return_revoked) {
+	if ((stid->sc_status & SC_STATUS_REVOKED) && !return_revoked) {
 		nfs4_put_stid(stid);
-		if (cstate->minorversion)
-			return nfserr_deleg_revoked;
-		return nfserr_bad_stateid;
+		return nfserr_deleg_revoked;
+	}
+	if (stid->sc_status & SC_STATUS_ADMIN_REVOKED) {
+		nfsd40_drop_revoked_stid(cstate->clp, stateid);
+		nfs4_put_stid(stid);
+		return nfserr_admin_revoked;
 	}
 	*s = stid;
 	return nfs_ok;
@@ -6472,17 +6848,17 @@ nfs4_find_file(struct nfs4_stid *s, int flags)
 {
 	struct nfsd_file *ret = NULL;
 
-	if (!s)
+	if (!s || s->sc_status)
 		return NULL;
 
 	switch (s->sc_type) {
-	case NFS4_DELEG_STID:
+	case SC_TYPE_DELEG:
 		spin_lock(&s->sc_file->fi_lock);
 		ret = nfsd_file_get(s->sc_file->fi_deleg_file);
 		spin_unlock(&s->sc_file->fi_lock);
 		break;
-	case NFS4_OPEN_STID:
-	case NFS4_LOCK_STID:
+	case SC_TYPE_OPEN:
+	case SC_TYPE_LOCK:
 		if (flags & RD_STATE)
 			ret = find_readable_file(s->sc_file);
 		else
@@ -6595,7 +6971,8 @@ static __be32 find_cpntf_state(struct nfsd_net *nn, stateid_t *st,
 		goto out;
 
 	*stid = find_stateid_by_type(found, &cps->cp_p_stateid,
-			NFS4_DELEG_STID|NFS4_OPEN_STID|NFS4_LOCK_STID);
+				     SC_TYPE_DELEG|SC_TYPE_OPEN|SC_TYPE_LOCK,
+				     0);
 	if (*stid)
 		status = nfs_ok;
 	else
@@ -6652,8 +7029,8 @@ nfs4_preprocess_stateid_op(struct svc_rqst *rqstp,
 	}
 
 	status = nfsd4_lookup_stateid(cstate, stateid,
-				NFS4_DELEG_STID|NFS4_OPEN_STID|NFS4_LOCK_STID,
-				&s, nn);
+				SC_TYPE_DELEG|SC_TYPE_OPEN|SC_TYPE_LOCK,
+				0, &s, nn);
 	if (status == nfserr_bad_stateid)
 		status = find_cpntf_state(nn, stateid, &s);
 	if (status)
@@ -6664,16 +7041,13 @@ nfs4_preprocess_stateid_op(struct svc_rqst *rqstp,
 		goto out;
 
 	switch (s->sc_type) {
-	case NFS4_DELEG_STID:
+	case SC_TYPE_DELEG:
 		status = nfs4_check_delegmode(delegstateid(s), flags);
 		break;
-	case NFS4_OPEN_STID:
-	case NFS4_LOCK_STID:
+	case SC_TYPE_OPEN:
+	case SC_TYPE_LOCK:
 		status = nfs4_check_olstateid(openlockstateid(s), flags);
 		break;
-	default:
-		status = nfserr_bad_stateid;
-		break;
 	}
 	if (status)
 		goto out;
@@ -6752,34 +7126,39 @@ nfsd4_free_stateid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 
 	spin_lock(&cl->cl_lock);
 	s = find_stateid_locked(cl, stateid);
-	if (!s)
+	if (!s || s->sc_status & SC_STATUS_CLOSED)
 		goto out_unlock;
+	if (s->sc_status & SC_STATUS_ADMIN_REVOKED) {
+		nfsd4_drop_revoked_stid(s);
+		ret = nfs_ok;
+		goto out;
+	}
 	spin_lock(&s->sc_lock);
 	switch (s->sc_type) {
-	case NFS4_DELEG_STID:
+	case SC_TYPE_DELEG:
+		if (s->sc_status & SC_STATUS_REVOKED) {
+			spin_unlock(&s->sc_lock);
+			dp = delegstateid(s);
+			list_del_init(&dp->dl_recall_lru);
+			spin_unlock(&cl->cl_lock);
+			nfs4_put_stid(s);
+			ret = nfs_ok;
+			goto out;
+		}
 		ret = nfserr_locks_held;
 		break;
-	case NFS4_OPEN_STID:
+	case SC_TYPE_OPEN:
 		ret = check_stateid_generation(stateid, &s->sc_stateid, 1);
 		if (ret)
 			break;
 		ret = nfserr_locks_held;
 		break;
-	case NFS4_LOCK_STID:
+	case SC_TYPE_LOCK:
 		spin_unlock(&s->sc_lock);
 		refcount_inc(&s->sc_count);
 		spin_unlock(&cl->cl_lock);
 		ret = nfsd4_free_lock_stateid(stateid, s);
 		goto out;
-	case NFS4_REVOKED_DELEG_STID:
-		spin_unlock(&s->sc_lock);
-		dp = delegstateid(s);
-		list_del_init(&dp->dl_recall_lru);
-		spin_unlock(&cl->cl_lock);
-		nfs4_put_stid(s);
-		ret = nfs_ok;
-		goto out;
-	/* Default falls through and returns nfserr_bad_stateid */
 	}
 	spin_unlock(&s->sc_lock);
 out_unlock:
@@ -6821,6 +7200,7 @@ static __be32 nfs4_seqid_op_checks(struct nfsd4_compound_state *cstate, stateid_
  * @seqid: seqid (provided by client)
  * @stateid: stateid (provided by client)
  * @typemask: mask of allowable types for this operation
+ * @statusmask: mask of allowed states: 0 or STID_CLOSED
  * @stpp: return pointer for the stateid found
  * @nn: net namespace for request
  *
@@ -6830,7 +7210,8 @@ static __be32 nfs4_seqid_op_checks(struct nfsd4_compound_state *cstate, stateid_
  */
 static __be32
 nfs4_preprocess_seqid_op(struct nfsd4_compound_state *cstate, u32 seqid,
-			 stateid_t *stateid, char typemask,
+			 stateid_t *stateid,
+			 unsigned short typemask, unsigned short statusmask,
 			 struct nfs4_ol_stateid **stpp,
 			 struct nfsd_net *nn)
 {
@@ -6841,11 +7222,16 @@ nfs4_preprocess_seqid_op(struct nfsd4_compound_state *cstate, u32 seqid,
 	trace_nfsd_preprocess(seqid, stateid);
 
 	*stpp = NULL;
-	status = nfsd4_lookup_stateid(cstate, stateid, typemask, &s, nn);
+retry:
+	status = nfsd4_lookup_stateid(cstate, stateid,
+				      typemask, statusmask, &s, nn);
 	if (status)
 		return status;
 	stp = openlockstateid(s);
-	nfsd4_cstate_assign_replay(cstate, stp->st_stateowner);
+	if (nfsd4_cstate_assign_replay(cstate, stp->st_stateowner) == -EAGAIN) {
+		nfs4_put_stateowner(stp->st_stateowner);
+		goto retry;
+	}
 
 	status = nfs4_seqid_op_checks(cstate, stateid, seqid, stp);
 	if (!status)
@@ -6863,7 +7249,7 @@ static __be32 nfs4_preprocess_confirmed_seqid_op(struct nfsd4_compound_state *cs
 	struct nfs4_ol_stateid *stp;
 
 	status = nfs4_preprocess_seqid_op(cstate, seqid, stateid,
-						NFS4_OPEN_STID, &stp, nn);
+					  SC_TYPE_OPEN, 0, &stp, nn);
 	if (status)
 		return status;
 	oo = openowner(stp->st_stateowner);
@@ -6894,8 +7280,8 @@ nfsd4_open_confirm(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 		return status;
 
 	status = nfs4_preprocess_seqid_op(cstate,
-					oc->oc_seqid, &oc->oc_req_stateid,
-					NFS4_OPEN_STID, &stp, nn);
+					  oc->oc_seqid, &oc->oc_req_stateid,
+					  SC_TYPE_OPEN, 0, &stp, nn);
 	if (status)
 		goto out;
 	oo = openowner(stp->st_stateowner);
@@ -6987,7 +7373,7 @@ out:
 	return status;
 }
 
-static void nfsd4_close_open_stateid(struct nfs4_ol_stateid *s)
+static bool nfsd4_close_open_stateid(struct nfs4_ol_stateid *s)
 {
 	struct nfs4_client *clp = s->st_stid.sc_client;
 	bool unhashed;
@@ -7004,11 +7390,11 @@ static void nfsd4_close_open_stateid(struct nfs4_ol_stateid *s)
 		list_for_each_entry(stp, &reaplist, st_locks)
 			nfs4_free_cpntf_statelist(clp->net, &stp->st_stid);
 		free_ol_stateid_reaplist(&reaplist);
+		return false;
 	} else {
 		spin_unlock(&clp->cl_lock);
 		free_ol_stateid_reaplist(&reaplist);
-		if (unhashed)
-			move_to_close_lru(s, clp->net);
+		return unhashed;
 	}
 }
 
@@ -7024,19 +7410,22 @@ nfsd4_close(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	struct nfs4_ol_stateid *stp;
 	struct net *net = SVC_NET(rqstp);
 	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
+	bool need_move_to_close_list;
 
-	dprintk("NFSD: nfsd4_close on file %pd\n", 
+	dprintk("NFSD: nfsd4_close on file %pd\n",
 			cstate->current_fh.fh_dentry);
 
 	status = nfs4_preprocess_seqid_op(cstate, close->cl_seqid,
-					&close->cl_stateid,
-					NFS4_OPEN_STID|NFS4_CLOSED_STID,
-					&stp, nn);
+					  &close->cl_stateid,
+					  SC_TYPE_OPEN, SC_STATUS_CLOSED,
+					  &stp, nn);
 	nfsd4_bump_seqid(cstate, status);
 	if (status)
-		goto out; 
+		goto out;
 
-	stp->st_stid.sc_type = NFS4_CLOSED_STID;
+	spin_lock(&stp->st_stid.sc_client->cl_lock);
+	stp->st_stid.sc_status |= SC_STATUS_CLOSED;
+	spin_unlock(&stp->st_stid.sc_client->cl_lock);
 
 	/*
 	 * Technically we don't _really_ have to increment or copy it, since
@@ -7046,8 +7435,10 @@ nfsd4_close(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	 */
 	nfs4_inc_and_copy_stateid(&close->cl_stateid, &stp->st_stid);
 
-	nfsd4_close_open_stateid(stp);
+	need_move_to_close_list = nfsd4_close_open_stateid(stp);
 	mutex_unlock(&stp->st_mutex);
+	if (need_move_to_close_list)
+		move_to_close_lru(stp, net);
 
 	/* v4.1+ suggests that we send a special stateid in here, since the
 	 * clients should just ignore this anyway. Since this is not useful
@@ -7078,7 +7469,7 @@ nfsd4_delegreturn(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	if ((status = fh_verify(rqstp, &cstate->current_fh, S_IFREG, 0)))
 		return status;
 
-	status = nfsd4_lookup_stateid(cstate, stateid, NFS4_DELEG_STID, &s, nn);
+	status = nfsd4_lookup_stateid(cstate, stateid, SC_TYPE_DELEG, 0, &s, nn);
 	if (status)
 		goto out;
 	dp = delegstateid(s);
@@ -7145,7 +7536,7 @@ nfsd4_lm_put_owner(fl_owner_t owner)
 static bool
 nfsd4_lm_lock_expirable(struct file_lock *cfl)
 {
-	struct nfs4_lockowner *lo = (struct nfs4_lockowner *)cfl->fl_owner;
+	struct nfs4_lockowner *lo = (struct nfs4_lockowner *) cfl->c.flc_owner;
 	struct nfs4_client *clp = lo->lo_owner.so_client;
 	struct nfsd_net *nn;
 
@@ -7167,7 +7558,7 @@ nfsd4_lm_expire_lock(void)
 static void
 nfsd4_lm_notify(struct file_lock *fl)
 {
-	struct nfs4_lockowner		*lo = (struct nfs4_lockowner *)fl->fl_owner;
+	struct nfs4_lockowner		*lo = (struct nfs4_lockowner *) fl->c.flc_owner;
 	struct net			*net = lo->lo_owner.so_client->net;
 	struct nfsd_net			*nn = net_generic(net, nfsd_net_id);
 	struct nfsd4_blocked_lock	*nbl = container_of(fl,
@@ -7204,7 +7595,7 @@ nfs4_set_lock_denied(struct file_lock *fl, struct nfsd4_lock_denied *deny)
 	struct nfs4_lockowner *lo;
 
 	if (fl->fl_lmops == &nfsd_posix_mng_ops) {
-		lo = (struct nfs4_lockowner *) fl->fl_owner;
+		lo = (struct nfs4_lockowner *) fl->c.flc_owner;
 		xdr_netobj_dup(&deny->ld_owner, &lo->lo_owner.so_owner,
 						GFP_KERNEL);
 		if (!deny->ld_owner.data)
@@ -7223,7 +7614,7 @@ nevermind:
 	if (fl->fl_end != NFS4_MAX_UINT64)
 		deny->ld_length = fl->fl_end - fl->fl_start + 1;        
 	deny->ld_type = NFS4_READ_LT;
-	if (fl->fl_type != F_RDLCK)
+	if (fl->c.flc_type != F_RDLCK)
 		deny->ld_type = NFS4_WRITE_LT;
 }
 
@@ -7345,7 +7736,7 @@ retry:
 	if (retstp)
 		goto out_found;
 	refcount_inc(&stp->st_stid.sc_count);
-	stp->st_stid.sc_type = NFS4_LOCK_STID;
+	stp->st_stid.sc_type = SC_TYPE_LOCK;
 	stp->st_stateowner = nfs4_get_stateowner(&lo->lo_owner);
 	get_nfs4_file(fp);
 	stp->st_stid.sc_file = fp;
@@ -7489,8 +7880,8 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	int lkflg;
 	int err;
 	bool new = false;
-	unsigned char fl_type;
-	unsigned int fl_flags = FL_POSIX;
+	unsigned char type;
+	unsigned int flags = FL_POSIX;
 	struct net *net = SVC_NET(rqstp);
 	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
 
@@ -7532,9 +7923,10 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 							&lock_stp, &new);
 	} else {
 		status = nfs4_preprocess_seqid_op(cstate,
-				       lock->lk_old_lock_seqid,
-				       &lock->lk_old_lock_stateid,
-				       NFS4_LOCK_STID, &lock_stp, nn);
+						  lock->lk_old_lock_seqid,
+						  &lock->lk_old_lock_stateid,
+						  SC_TYPE_LOCK, 0, &lock_stp,
+						  nn);
 	}
 	if (status)
 		goto out;
@@ -7553,14 +7945,14 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 		goto out;
 
 	if (lock->lk_reclaim)
-		fl_flags |= FL_RECLAIM;
+		flags |= FL_RECLAIM;
 
 	fp = lock_stp->st_stid.sc_file;
 	switch (lock->lk_type) {
 		case NFS4_READW_LT:
 			if (nfsd4_has_session(cstate) ||
 			    exportfs_lock_op_is_async(sb->s_export_op))
-				fl_flags |= FL_SLEEP;
+				flags |= FL_SLEEP;
 			fallthrough;
 		case NFS4_READ_LT:
 			spin_lock(&fp->fi_lock);
@@ -7568,12 +7960,12 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 			if (nf)
 				get_lock_access(lock_stp, NFS4_SHARE_ACCESS_READ);
 			spin_unlock(&fp->fi_lock);
-			fl_type = F_RDLCK;
+			type = F_RDLCK;
 			break;
 		case NFS4_WRITEW_LT:
 			if (nfsd4_has_session(cstate) ||
 			    exportfs_lock_op_is_async(sb->s_export_op))
-				fl_flags |= FL_SLEEP;
+				flags |= FL_SLEEP;
 			fallthrough;
 		case NFS4_WRITE_LT:
 			spin_lock(&fp->fi_lock);
@@ -7581,7 +7973,7 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 			if (nf)
 				get_lock_access(lock_stp, NFS4_SHARE_ACCESS_WRITE);
 			spin_unlock(&fp->fi_lock);
-			fl_type = F_WRLCK;
+			type = F_WRLCK;
 			break;
 		default:
 			status = nfserr_inval;
@@ -7601,7 +7993,7 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	 * on those filesystems:
 	 */
 	if (!exportfs_lock_op_is_async(sb->s_export_op))
-		fl_flags &= ~FL_SLEEP;
+		flags &= ~FL_SLEEP;
 
 	nbl = find_or_allocate_block(lock_sop, &fp->fi_fhandle, nn);
 	if (!nbl) {
@@ -7611,11 +8003,11 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	}
 
 	file_lock = &nbl->nbl_lock;
-	file_lock->fl_type = fl_type;
-	file_lock->fl_owner = (fl_owner_t)lockowner(nfs4_get_stateowner(&lock_sop->lo_owner));
-	file_lock->fl_pid = current->tgid;
-	file_lock->fl_file = nf->nf_file;
-	file_lock->fl_flags = fl_flags;
+	file_lock->c.flc_type = type;
+	file_lock->c.flc_owner = (fl_owner_t)lockowner(nfs4_get_stateowner(&lock_sop->lo_owner));
+	file_lock->c.flc_pid = current->tgid;
+	file_lock->c.flc_file = nf->nf_file;
+	file_lock->c.flc_flags = flags;
 	file_lock->fl_lmops = &nfsd_posix_mng_ops;
 	file_lock->fl_start = lock->lk_offset;
 	file_lock->fl_end = last_byte_offset(lock->lk_offset, lock->lk_length);
@@ -7628,7 +8020,7 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 		goto out;
 	}
 
-	if (fl_flags & FL_SLEEP) {
+	if (flags & FL_SLEEP) {
 		nbl->nbl_time = ktime_get_boottime_seconds();
 		spin_lock(&nn->blocked_locks_lock);
 		list_add_tail(&nbl->nbl_list, &lock_sop->lo_blocked);
@@ -7665,7 +8057,7 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 out:
 	if (nbl) {
 		/* dequeue it if we queued it before */
-		if (fl_flags & FL_SLEEP) {
+		if (flags & FL_SLEEP) {
 			spin_lock(&nn->blocked_locks_lock);
 			if (!list_empty(&nbl->nbl_list) &&
 			    !list_empty(&nbl->nbl_lru)) {
@@ -7733,9 +8125,9 @@ static __be32 nfsd_test_lock(struct svc_rqst *rqstp, struct svc_fh *fhp, struct
 	err = nfserrno(nfsd_open_break_lease(inode, NFSD_MAY_READ));
 	if (err)
 		goto out;
-	lock->fl_file = nf->nf_file;
+	lock->c.flc_file = nf->nf_file;
 	err = nfserrno(vfs_test_lock(nf->nf_file, lock));
-	lock->fl_file = NULL;
+	lock->c.flc_file = NULL;
 out:
 	inode_unlock(inode);
 	nfsd_file_put(nf);
@@ -7780,11 +8172,11 @@ nfsd4_lockt(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	switch (lockt->lt_type) {
 		case NFS4_READ_LT:
 		case NFS4_READW_LT:
-			file_lock->fl_type = F_RDLCK;
+			file_lock->c.flc_type = F_RDLCK;
 			break;
 		case NFS4_WRITE_LT:
 		case NFS4_WRITEW_LT:
-			file_lock->fl_type = F_WRLCK;
+			file_lock->c.flc_type = F_WRLCK;
 			break;
 		default:
 			dprintk("NFSD: nfs4_lockt: bad lock type!\n");
@@ -7794,9 +8186,9 @@ nfsd4_lockt(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 
 	lo = find_lockowner_str(cstate->clp, &lockt->lt_owner);
 	if (lo)
-		file_lock->fl_owner = (fl_owner_t)lo;
-	file_lock->fl_pid = current->tgid;
-	file_lock->fl_flags = FL_POSIX;
+		file_lock->c.flc_owner = (fl_owner_t)lo;
+	file_lock->c.flc_pid = current->tgid;
+	file_lock->c.flc_flags = FL_POSIX;
 
 	file_lock->fl_start = lockt->lt_offset;
 	file_lock->fl_end = last_byte_offset(lockt->lt_offset, lockt->lt_length);
@@ -7807,7 +8199,7 @@ nfsd4_lockt(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	if (status)
 		goto out;
 
-	if (file_lock->fl_type != F_UNLCK) {
+	if (file_lock->c.flc_type != F_UNLCK) {
 		status = nfserr_denied;
 		nfs4_set_lock_denied(file_lock, &lockt->lt_denied);
 	}
@@ -7847,8 +8239,8 @@ nfsd4_locku(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 		 return nfserr_inval;
 
 	status = nfs4_preprocess_seqid_op(cstate, locku->lu_seqid,
-					&locku->lu_stateid, NFS4_LOCK_STID,
-					&stp, nn);
+					  &locku->lu_stateid, SC_TYPE_LOCK, 0,
+					  &stp, nn);
 	if (status)
 		goto out;
 	nf = find_any_file(stp->st_stid.sc_file);
@@ -7863,11 +8255,11 @@ nfsd4_locku(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 		goto put_file;
 	}
 
-	file_lock->fl_type = F_UNLCK;
-	file_lock->fl_owner = (fl_owner_t)lockowner(nfs4_get_stateowner(stp->st_stateowner));
-	file_lock->fl_pid = current->tgid;
-	file_lock->fl_file = nf->nf_file;
-	file_lock->fl_flags = FL_POSIX;
+	file_lock->c.flc_type = F_UNLCK;
+	file_lock->c.flc_owner = (fl_owner_t)lockowner(nfs4_get_stateowner(stp->st_stateowner));
+	file_lock->c.flc_pid = current->tgid;
+	file_lock->c.flc_file = nf->nf_file;
+	file_lock->c.flc_flags = FL_POSIX;
 	file_lock->fl_lmops = &nfsd_posix_mng_ops;
 	file_lock->fl_start = locku->lu_offset;
 
@@ -7924,8 +8316,8 @@ check_for_locks(struct nfs4_file *fp, struct nfs4_lockowner *lowner)
 
 	if (flctx && !list_empty_careful(&flctx->flc_posix)) {
 		spin_lock(&flctx->flc_lock);
-		list_for_each_entry(fl, &flctx->flc_posix, fl_list) {
-			if (fl->fl_owner == (fl_owner_t)lowner) {
+		for_each_file_lock(fl, &flctx->flc_posix) {
+			if (fl->c.flc_owner == (fl_owner_t)lowner) {
 				status = true;
 				break;
 			}
@@ -7993,7 +8385,7 @@ nfsd4_release_lockowner(struct svc_rqst *rqstp,
 		stp = list_first_entry(&lo->lo_owner.so_stateids,
 				       struct nfs4_ol_stateid,
 				       st_perstateowner);
-		WARN_ON(!unhash_lock_stateid(stp));
+		unhash_lock_stateid(stp);
 		put_ol_stateid_locked(stp, &reaplist);
 	}
 	spin_unlock(&clp->cl_lock);
@@ -8260,12 +8652,6 @@ nfs4_state_start(void)
 	if (ret)
 		return ret;
 
-	ret = nfsd4_create_callback_queue();
-	if (ret) {
-		rhltable_destroy(&nfs4_file_rhltable);
-		return ret;
-	}
-
 	set_max_delegations();
 	return 0;
 }
@@ -8286,7 +8672,7 @@ nfs4_state_shutdown_net(struct net *net)
 	spin_lock(&state_lock);
 	list_for_each_safe(pos, next, &nn->del_recall_lru) {
 		dp = list_entry (pos, struct nfs4_delegation, dl_recall_lru);
-		WARN_ON(!unhash_delegation_locked(dp));
+		unhash_delegation_locked(dp, SC_STATUS_CLOSED);
 		list_add(&dp->dl_recall_lru, &reaplist);
 	}
 	spin_unlock(&state_lock);
@@ -8306,7 +8692,6 @@ nfs4_state_shutdown_net(struct net *net)
 void
 nfs4_state_shutdown(void)
 {
-	nfsd4_destroy_callback_queue();
 	rhltable_destroy(&nfs4_file_rhltable);
 }
 
@@ -8428,6 +8813,8 @@ nfsd4_get_writestateid(struct nfsd4_compound_state *cstate,
  * nfsd4_deleg_getattr_conflict - Recall if GETATTR causes conflict
  * @rqstp: RPC transaction context
  * @inode: file to be checked for a conflict
+ * @modified: return true if file was modified
+ * @size: new size of file if modified is true
  *
  * This function is called when there is a conflict between a write
  * delegation and a change/size GETATTR from another client. The server
@@ -8436,28 +8823,30 @@ nfsd4_get_writestateid(struct nfsd4_compound_state *cstate,
  * delegation before replying to the GETATTR. See RFC 8881 section
  * 18.7.4.
  *
- * The current implementation does not support CB_GETATTR yet. However
- * this can avoid recalling the delegation could be added in follow up
- * work.
- *
  * Returns 0 if there is no conflict; otherwise an nfs_stat
  * code is returned.
  */
 __be32
-nfsd4_deleg_getattr_conflict(struct svc_rqst *rqstp, struct inode *inode)
+nfsd4_deleg_getattr_conflict(struct svc_rqst *rqstp, struct inode *inode,
+				bool *modified, u64 *size)
 {
 	__be32 status;
 	struct nfsd_net *nn = net_generic(SVC_NET(rqstp), nfsd_net_id);
 	struct file_lock_context *ctx;
-	struct file_lock *fl;
+	struct file_lease *fl;
 	struct nfs4_delegation *dp;
+	struct iattr attrs;
+	struct nfs4_cb_fattr *ncf;
 
+	*modified = false;
 	ctx = locks_inode_context(inode);
 	if (!ctx)
 		return 0;
 	spin_lock(&ctx->flc_lock);
-	list_for_each_entry(fl, &ctx->flc_lease, fl_list) {
-		if (fl->fl_flags == FL_LAYOUT)
+	for_each_file_lock(fl, &ctx->flc_lease) {
+		unsigned char type = fl->c.flc_type;
+
+		if (fl->c.flc_flags == FL_LAYOUT)
 			continue;
 		if (fl->fl_lmops != &nfsd_lease_mng_ops) {
 			/*
@@ -8465,23 +8854,49 @@ nfsd4_deleg_getattr_conflict(struct svc_rqst *rqstp, struct inode *inode)
 			 * we are done; there isn't any write delegation
 			 * on this inode
 			 */
-			if (fl->fl_type == F_RDLCK)
+			if (type == F_RDLCK)
 				break;
 			goto break_lease;
 		}
-		if (fl->fl_type == F_WRLCK) {
-			dp = fl->fl_owner;
+		if (type == F_WRLCK) {
+			dp = fl->c.flc_owner;
 			if (dp->dl_recall.cb_clp == *(rqstp->rq_lease_breaker)) {
 				spin_unlock(&ctx->flc_lock);
 				return 0;
 			}
 break_lease:
-			spin_unlock(&ctx->flc_lock);
 			nfsd_stats_wdeleg_getattr_inc(nn);
-			status = nfserrno(nfsd_open_break_lease(inode, NFSD_MAY_READ));
-			if (status != nfserr_jukebox ||
-					!nfsd_wait_for_delegreturn(rqstp, inode))
-				return status;
+			dp = fl->c.flc_owner;
+			ncf = &dp->dl_cb_fattr;
+			nfs4_cb_getattr(&dp->dl_cb_fattr);
+			spin_unlock(&ctx->flc_lock);
+			wait_on_bit_timeout(&ncf->ncf_cb_flags, CB_GETATTR_BUSY,
+					TASK_INTERRUPTIBLE, NFSD_CB_GETATTR_TIMEOUT);
+			if (ncf->ncf_cb_status) {
+				/* Recall delegation only if client didn't respond */
+				status = nfserrno(nfsd_open_break_lease(inode, NFSD_MAY_READ));
+				if (status != nfserr_jukebox ||
+						!nfsd_wait_for_delegreturn(rqstp, inode))
+					return status;
+			}
+			if (!ncf->ncf_file_modified &&
+					(ncf->ncf_initial_cinfo != ncf->ncf_cb_change ||
+					ncf->ncf_cur_fsize != ncf->ncf_cb_fsize))
+				ncf->ncf_file_modified = true;
+			if (ncf->ncf_file_modified) {
+				/*
+				 * Per section 10.4.3 of RFC 8881, the server would
+				 * not update the file's metadata with the client's
+				 * modified size
+				 */
+				attrs.ia_mtime = attrs.ia_ctime = current_time(inode);
+				attrs.ia_valid = ATTR_MTIME | ATTR_CTIME;
+				setattr_copy(&nop_mnt_idmap, inode, &attrs);
+				mark_inode_dirty(inode);
+				ncf->ncf_cur_fsize = ncf->ncf_cb_fsize;
+				*size = ncf->ncf_cur_fsize;
+				*modified = true;
+			}
 			return 0;
 		}
 		break;
diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c
index 24db9f9ea8..c7bfd2180e 100644
--- a/fs/nfsd/nfs4xdr.c
+++ b/fs/nfsd/nfs4xdr.c
@@ -1732,6 +1732,35 @@ nfsd4_decode_free_stateid(struct nfsd4_compoundargs *argp,
 	return nfsd4_decode_stateid4(argp, &free_stateid->fr_stateid);
 }
 
+static __be32
+nfsd4_decode_get_dir_delegation(struct nfsd4_compoundargs *argp,
+		union nfsd4_op_u *u)
+{
+	struct nfsd4_get_dir_delegation *gdd = &u->get_dir_delegation;
+	__be32 status;
+
+	memset(gdd, 0, sizeof(*gdd));
+
+	if (xdr_stream_decode_bool(argp->xdr, &gdd->gdda_signal_deleg_avail) < 0)
+		return nfserr_bad_xdr;
+	status = nfsd4_decode_bitmap4(argp, gdd->gdda_notification_types,
+				      ARRAY_SIZE(gdd->gdda_notification_types));
+	if (status)
+		return status;
+	status = nfsd4_decode_nfstime4(argp, &gdd->gdda_child_attr_delay);
+	if (status)
+		return status;
+	status = nfsd4_decode_nfstime4(argp, &gdd->gdda_dir_attr_delay);
+	if (status)
+		return status;
+	status = nfsd4_decode_bitmap4(argp, gdd->gdda_child_attributes,
+					ARRAY_SIZE(gdd->gdda_child_attributes));
+	if (status)
+		return status;
+	return nfsd4_decode_bitmap4(argp, gdd->gdda_dir_attributes,
+					ARRAY_SIZE(gdd->gdda_dir_attributes));
+}
+
 #ifdef CONFIG_NFSD_PNFS
 static __be32
 nfsd4_decode_getdeviceinfo(struct nfsd4_compoundargs *argp,
@@ -2370,7 +2399,7 @@ static const nfsd4_dec nfsd4_dec_ops[] = {
 	[OP_CREATE_SESSION]	= nfsd4_decode_create_session,
 	[OP_DESTROY_SESSION]	= nfsd4_decode_destroy_session,
 	[OP_FREE_STATEID]	= nfsd4_decode_free_stateid,
-	[OP_GET_DIR_DELEGATION]	= nfsd4_decode_notsupp,
+	[OP_GET_DIR_DELEGATION]	= nfsd4_decode_get_dir_delegation,
 #ifdef CONFIG_NFSD_PNFS
 	[OP_GETDEVICEINFO]	= nfsd4_decode_getdeviceinfo,
 	[OP_GETDEVICELIST]	= nfsd4_decode_notsupp,
@@ -3505,6 +3534,8 @@ nfsd4_encode_fattr4(struct svc_rqst *rqstp, struct xdr_stream *xdr,
 		.dentry	= dentry,
 	};
 	unsigned long bit;
+	bool file_modified = false;
+	u64 size = 0;
 
 	WARN_ON_ONCE(bmval[1] & NFSD_WRITEONLY_ATTRS_WORD1);
 	WARN_ON_ONCE(!nfsd_attrs_supported(minorversion, bmval));
@@ -3531,7 +3562,8 @@ nfsd4_encode_fattr4(struct svc_rqst *rqstp, struct xdr_stream *xdr,
 	}
 	args.size = 0;
 	if (attrmask[0] & (FATTR4_WORD0_CHANGE | FATTR4_WORD0_SIZE)) {
-		status = nfsd4_deleg_getattr_conflict(rqstp, d_inode(dentry));
+		status = nfsd4_deleg_getattr_conflict(rqstp, d_inode(dentry),
+					&file_modified, &size);
 		if (status)
 			goto out;
 	}
@@ -3541,7 +3573,10 @@ nfsd4_encode_fattr4(struct svc_rqst *rqstp, struct xdr_stream *xdr,
 			  AT_STATX_SYNC_AS_STAT);
 	if (err)
 		goto out_nfserr;
-	args.size = args.stat.size;
+	if (file_modified)
+		args.size = size;
+	else
+		args.size = args.stat.size;
 
 	if (!(args.stat.result_mask & STATX_BTIME))
 		/* underlying FS does not offer btime so we can't share it */
@@ -4957,6 +4992,49 @@ nfsd4_encode_test_stateid(struct nfsd4_compoundres *resp, __be32 nfserr,
 	return nfs_ok;
 }
 
+static __be32
+nfsd4_encode_get_dir_delegation(struct nfsd4_compoundres *resp, __be32 nfserr,
+				union nfsd4_op_u *u)
+{
+	struct nfsd4_get_dir_delegation *gdd = &u->get_dir_delegation;
+	struct xdr_stream *xdr = resp->xdr;
+	__be32 status = nfserr_resource;
+
+	switch(gdd->gddrnf_status) {
+	case GDD4_OK:
+		if (xdr_stream_encode_u32(xdr, GDD4_OK) != XDR_UNIT)
+			break;
+		status = nfsd4_encode_verifier4(xdr, &gdd->gddr_cookieverf);
+		if (status)
+			break;
+		status = nfsd4_encode_stateid4(xdr, &gdd->gddr_stateid);
+		if (status)
+			break;
+		status = nfsd4_encode_bitmap4(xdr, gdd->gddr_notification[0], 0, 0);
+		if (status)
+			break;
+		status = nfsd4_encode_bitmap4(xdr, gdd->gddr_child_attributes[0],
+						   gdd->gddr_child_attributes[1],
+						   gdd->gddr_child_attributes[2]);
+		if (status)
+			break;
+		status = nfsd4_encode_bitmap4(xdr, gdd->gddr_dir_attributes[0],
+						   gdd->gddr_dir_attributes[1],
+						   gdd->gddr_dir_attributes[2]);
+		break;
+	default:
+		pr_warn("nfsd: bad gddrnf_status (%u)\n", gdd->gddrnf_status);
+		gdd->gddrnf_will_signal_deleg_avail = 0;
+		fallthrough;
+	case GDD4_UNAVAIL:
+		if (xdr_stream_encode_u32(xdr, GDD4_UNAVAIL) != XDR_UNIT)
+			break;
+		status = nfsd4_encode_bool(xdr, gdd->gddrnf_will_signal_deleg_avail);
+		break;
+	}
+	return status;
+}
+
 #ifdef CONFIG_NFSD_PNFS
 static __be32
 nfsd4_encode_device_addr4(struct xdr_stream *xdr,
@@ -5193,7 +5271,12 @@ nfsd4_encode_offload_status(struct nfsd4_compoundres *resp, __be32 nfserr,
 	if (nfserr != nfs_ok)
 		return nfserr;
 	/* osr_complete<1> */
-	if (xdr_stream_encode_u32(xdr, 0) != XDR_UNIT)
+	if (os->completed) {
+		if (xdr_stream_encode_u32(xdr, 1) != XDR_UNIT)
+			return nfserr_resource;
+		if (xdr_stream_encode_be32(xdr, os->status) != XDR_UNIT)
+			return nfserr_resource;
+	} else if (xdr_stream_encode_u32(xdr, 0) != XDR_UNIT)
 		return nfserr_resource;
 	return nfs_ok;
 }
@@ -5385,16 +5468,11 @@ nfsd4_listxattr_validate_cookie(struct nfsd4_listxattrs *listxattrs,
 
 	/*
 	 * If the cookie is larger than the maximum number we can fit
-	 * in either the buffer we just got back from vfs_listxattr, or,
-	 * XDR-encoded, in the return buffer, it's invalid.
+	 * in the buffer we just got back from vfs_listxattr, it's invalid.
 	 */
 	if (cookie > (listxattrs->lsxa_len) / (XATTR_USER_PREFIX_LEN + 2))
 		return nfserr_badcookie;
 
-	if (cookie > (listxattrs->lsxa_maxcount /
-		      (XDR_QUADLEN(XATTR_USER_PREFIX_LEN + 2) + 4)))
-		return nfserr_badcookie;
-
 	*offsetp = (u32)cookie;
 	return 0;
 }
@@ -5411,6 +5489,7 @@ nfsd4_encode_listxattrs(struct nfsd4_compoundres *resp, __be32 nfserr,
 	u64 cookie;
 	char *sp;
 	__be32 status, tmp;
+	__be64 wire_cookie;
 	__be32 *p;
 	u32 nuser;
 
@@ -5426,7 +5505,7 @@ nfsd4_encode_listxattrs(struct nfsd4_compoundres *resp, __be32 nfserr,
 	 */
 	cookie_offset = xdr->buf->len;
 	count_offset = cookie_offset + 8;
-	p = xdr_reserve_space(xdr, 12);
+	p = xdr_reserve_space(xdr, XDR_UNIT * 3);
 	if (!p) {
 		status = nfserr_resource;
 		goto out;
@@ -5437,7 +5516,8 @@ nfsd4_encode_listxattrs(struct nfsd4_compoundres *resp, __be32 nfserr,
 	sp = listxattrs->lsxa_buf;
 	nuser = 0;
 
-	xdrleft = listxattrs->lsxa_maxcount;
+	/* Bytes left is maxcount - 8 (cookie) - 4 (array count) */
+	xdrleft = listxattrs->lsxa_maxcount - XDR_UNIT * 3;
 
 	while (left > 0 && xdrleft > 0) {
 		slen = strlen(sp);
@@ -5450,7 +5530,8 @@ nfsd4_encode_listxattrs(struct nfsd4_compoundres *resp, __be32 nfserr,
 
 		slen -= XATTR_USER_PREFIX_LEN;
 		xdrlen = 4 + ((slen + 3) & ~3);
-		if (xdrlen > xdrleft) {
+		/* Check if both entry and eof can fit in the XDR buffer */
+		if (xdrlen + XDR_UNIT > xdrleft) {
 			if (count == 0) {
 				/*
 				 * Can't even fit the first attribute name.
@@ -5502,7 +5583,8 @@ wreof:
 
 	cookie = offset + count;
 
-	write_bytes_to_xdr_buf(xdr->buf, cookie_offset, &cookie, 8);
+	wire_cookie = cpu_to_be64(cookie);
+	write_bytes_to_xdr_buf(xdr->buf, cookie_offset, &wire_cookie, 8);
 	tmp = cpu_to_be32(count);
 	write_bytes_to_xdr_buf(xdr->buf, count_offset, &tmp, 4);
 out:
@@ -5574,7 +5656,7 @@ static const nfsd4_enc nfsd4_enc_ops[] = {
 	[OP_CREATE_SESSION]	= nfsd4_encode_create_session,
 	[OP_DESTROY_SESSION]	= nfsd4_encode_noop,
 	[OP_FREE_STATEID]	= nfsd4_encode_noop,
-	[OP_GET_DIR_DELEGATION]	= nfsd4_encode_noop,
+	[OP_GET_DIR_DELEGATION]	= nfsd4_encode_get_dir_delegation,
 #ifdef CONFIG_NFSD_PNFS
 	[OP_GETDEVICEINFO]	= nfsd4_encode_getdeviceinfo,
 	[OP_GETDEVICELIST]	= nfsd4_encode_noop,
@@ -5726,27 +5808,24 @@ release:
 	rqstp->rq_next_page = xdr->page_ptr + 1;
 }
 
-/* 
- * Encode the reply stored in the stateowner reply cache 
- * 
- * XDR note: do not encode rp->rp_buflen: the buffer contains the
- * previously sent already encoded operation.
+/**
+ * nfsd4_encode_replay - encode a result stored in the stateowner reply cache
+ * @xdr: send buffer's XDR stream
+ * @op: operation being replayed
+ *
+ * @op->replay->rp_buf contains the previously-sent already-encoded result.
  */
-void
-nfsd4_encode_replay(struct xdr_stream *xdr, struct nfsd4_op *op)
+void nfsd4_encode_replay(struct xdr_stream *xdr, struct nfsd4_op *op)
 {
-	__be32 *p;
 	struct nfs4_replay *rp = op->replay;
 
-	p = xdr_reserve_space(xdr, 8 + rp->rp_buflen);
-	if (!p) {
-		WARN_ON_ONCE(1);
-		return;
-	}
-	*p++ = cpu_to_be32(op->opnum);
-	*p++ = rp->rp_status;  /* already xdr'ed */
+	trace_nfsd_stateowner_replay(op->opnum, rp);
 
-	p = xdr_encode_opaque_fixed(p, rp->rp_buf, rp->rp_buflen);
+	if (xdr_stream_encode_u32(xdr, op->opnum) != XDR_UNIT)
+		return;
+	if (xdr_stream_encode_be32(xdr, rp->rp_status) != XDR_UNIT)
+		return;
+	xdr_stream_encode_opaque_fixed(xdr, rp->rp_buf, rp->rp_buflen);
 }
 
 void nfsd4_release_compoundargs(struct svc_rqst *rqstp)
diff --git a/fs/nfsd/nfscache.c b/fs/nfsd/nfscache.c
index cfcc6ac8f2..ba9d326b3d 100644
--- a/fs/nfsd/nfscache.c
+++ b/fs/nfsd/nfscache.c
@@ -166,8 +166,7 @@ nfsd_reply_cache_free(struct nfsd_drc_bucket *b, struct nfsd_cacherep *rp,
 
 int nfsd_drc_slab_create(void)
 {
-	drc_slab = kmem_cache_create("nfsd_drc",
-				sizeof(struct nfsd_cacherep), 0, 0, NULL);
+	drc_slab = KMEM_CACHE(nfsd_cacherep, 0);
 	return drc_slab ? 0: -ENOMEM;
 }
 
diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c
index d8f54eb745..0f9b4f7b56 100644
--- a/fs/nfsd/nfsctl.c
+++ b/fs/nfsd/nfsctl.c
@@ -15,6 +15,7 @@
 #include <linux/sunrpc/addr.h>
 #include <linux/sunrpc/gss_api.h>
 #include <linux/sunrpc/rpc_pipe_fs.h>
+#include <linux/sunrpc/svc.h>
 #include <linux/module.h>
 #include <linux/fsnotify.h>
 
@@ -279,6 +280,7 @@ static ssize_t write_unlock_fs(struct file *file, char *buf, size_t size)
 	 * 3.  Is that directory the root of an exported file system?
 	 */
 	error = nlmsvc_unlock_all_by_sb(path.dentry->d_sb);
+	nfsd4_revoke_states(netns(file), path.dentry->d_sb);
 
 	path_put(&path);
 	return error;
@@ -403,7 +405,9 @@ static ssize_t write_threads(struct file *file, char *buf, size_t size)
 		if (newthreads < 0)
 			return -EINVAL;
 		trace_nfsd_ctl_threads(net, newthreads);
-		rv = nfsd_svc(newthreads, net, file->f_cred);
+		mutex_lock(&nfsd_mutex);
+		rv = nfsd_svc(newthreads, net, file->f_cred, NULL);
+		mutex_unlock(&nfsd_mutex);
 		if (rv < 0)
 			return rv;
 	} else
@@ -1456,28 +1460,6 @@ static int create_proc_exports_entry(void)
 
 unsigned int nfsd_net_id;
 
-/**
- * nfsd_nl_rpc_status_get_start - Prepare rpc_status_get dumpit
- * @cb: netlink metadata and command arguments
- *
- * Return values:
- *   %0: The rpc_status_get command may proceed
- *   %-ENODEV: There is no NFSD running in this namespace
- */
-int nfsd_nl_rpc_status_get_start(struct netlink_callback *cb)
-{
-	struct nfsd_net *nn = net_generic(sock_net(cb->skb->sk), nfsd_net_id);
-	int ret = -ENODEV;
-
-	mutex_lock(&nfsd_mutex);
-	if (nn->nfsd_serv)
-		ret = 0;
-	else
-		mutex_unlock(&nfsd_mutex);
-
-	return ret;
-}
-
 static int nfsd_genl_rpc_status_compose_msg(struct sk_buff *skb,
 					    struct netlink_callback *cb,
 					    struct nfsd_genl_rqstp *rqstp)
@@ -1554,8 +1536,16 @@ static int nfsd_genl_rpc_status_compose_msg(struct sk_buff *skb,
 int nfsd_nl_rpc_status_get_dumpit(struct sk_buff *skb,
 				  struct netlink_callback *cb)
 {
-	struct nfsd_net *nn = net_generic(sock_net(skb->sk), nfsd_net_id);
 	int i, ret, rqstp_index = 0;
+	struct nfsd_net *nn;
+
+	mutex_lock(&nfsd_mutex);
+
+	nn = net_generic(sock_net(skb->sk), nfsd_net_id);
+	if (!nn->nfsd_serv) {
+		ret = -ENODEV;
+		goto out_unlock;
+	}
 
 	rcu_read_lock();
 
@@ -1632,25 +1622,524 @@ int nfsd_nl_rpc_status_get_dumpit(struct sk_buff *skb,
 	ret = skb->len;
 out:
 	rcu_read_unlock();
+out_unlock:
+	mutex_unlock(&nfsd_mutex);
 
 	return ret;
 }
 
 /**
- * nfsd_nl_rpc_status_get_done - rpc_status_get dumpit post-processing
- * @cb: netlink metadata and command arguments
+ * nfsd_nl_threads_set_doit - set the number of running threads
+ * @skb: reply buffer
+ * @info: netlink metadata and command arguments
  *
- * Return values:
- *   %0: Success
+ * Return 0 on success or a negative errno.
  */
-int nfsd_nl_rpc_status_get_done(struct netlink_callback *cb)
+int nfsd_nl_threads_set_doit(struct sk_buff *skb, struct genl_info *info)
 {
+	int nthreads = 0, count = 0, nrpools, ret = -EOPNOTSUPP, rem;
+	struct net *net = genl_info_net(info);
+	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
+	const struct nlattr *attr;
+	const char *scope = NULL;
+
+	if (GENL_REQ_ATTR_CHECK(info, NFSD_A_SERVER_THREADS))
+		return -EINVAL;
+
+	/* count number of SERVER_THREADS values */
+	nlmsg_for_each_attr(attr, info->nlhdr, GENL_HDRLEN, rem) {
+		if (nla_type(attr) == NFSD_A_SERVER_THREADS)
+			count++;
+	}
+
+	mutex_lock(&nfsd_mutex);
+
+	nrpools = nfsd_nrpools(net);
+	if (nrpools && count > nrpools)
+		count = nrpools;
+
+	/* XXX: make this handle non-global pool-modes */
+	if (count > 1)
+		goto out_unlock;
+
+	nthreads = nla_get_u32(info->attrs[NFSD_A_SERVER_THREADS]);
+	if (info->attrs[NFSD_A_SERVER_GRACETIME] ||
+	    info->attrs[NFSD_A_SERVER_LEASETIME] ||
+	    info->attrs[NFSD_A_SERVER_SCOPE]) {
+		ret = -EBUSY;
+		if (nn->nfsd_serv && nn->nfsd_serv->sv_nrthreads)
+			goto out_unlock;
+
+		ret = -EINVAL;
+		attr = info->attrs[NFSD_A_SERVER_GRACETIME];
+		if (attr) {
+			u32 gracetime = nla_get_u32(attr);
+
+			if (gracetime < 10 || gracetime > 3600)
+				goto out_unlock;
+
+			nn->nfsd4_grace = gracetime;
+		}
+
+		attr = info->attrs[NFSD_A_SERVER_LEASETIME];
+		if (attr) {
+			u32 leasetime = nla_get_u32(attr);
+
+			if (leasetime < 10 || leasetime > 3600)
+				goto out_unlock;
+
+			nn->nfsd4_lease = leasetime;
+		}
+
+		attr = info->attrs[NFSD_A_SERVER_SCOPE];
+		if (attr)
+			scope = nla_data(attr);
+	}
+
+	ret = nfsd_svc(nthreads, net, get_current_cred(), scope);
+
+out_unlock:
+	mutex_unlock(&nfsd_mutex);
+
+	return ret == nthreads ? 0 : ret;
+}
+
+/**
+ * nfsd_nl_threads_get_doit - get the number of running threads
+ * @skb: reply buffer
+ * @info: netlink metadata and command arguments
+ *
+ * Return 0 on success or a negative errno.
+ */
+int nfsd_nl_threads_get_doit(struct sk_buff *skb, struct genl_info *info)
+{
+	struct net *net = genl_info_net(info);
+	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
+	void *hdr;
+	int err;
+
+	skb = genlmsg_new(GENLMSG_DEFAULT_SIZE, GFP_KERNEL);
+	if (!skb)
+		return -ENOMEM;
+
+	hdr = genlmsg_iput(skb, info);
+	if (!hdr) {
+		err = -EMSGSIZE;
+		goto err_free_msg;
+	}
+
+	mutex_lock(&nfsd_mutex);
+
+	err = nla_put_u32(skb, NFSD_A_SERVER_GRACETIME,
+			  nn->nfsd4_grace) ||
+	      nla_put_u32(skb, NFSD_A_SERVER_LEASETIME,
+			  nn->nfsd4_lease) ||
+	      nla_put_string(skb, NFSD_A_SERVER_SCOPE,
+			  nn->nfsd_name);
+	if (err)
+		goto err_unlock;
+
+	if (nn->nfsd_serv) {
+		int i;
+
+		for (i = 0; i < nfsd_nrpools(net); ++i) {
+			struct svc_pool *sp = &nn->nfsd_serv->sv_pools[i];
+
+			err = nla_put_u32(skb, NFSD_A_SERVER_THREADS,
+					  atomic_read(&sp->sp_nrthreads));
+			if (err)
+				goto err_unlock;
+		}
+	} else {
+		err = nla_put_u32(skb, NFSD_A_SERVER_THREADS, 0);
+		if (err)
+			goto err_unlock;
+	}
+
+	mutex_unlock(&nfsd_mutex);
+
+	genlmsg_end(skb, hdr);
+
+	return genlmsg_reply(skb, info);
+
+err_unlock:
+	mutex_unlock(&nfsd_mutex);
+err_free_msg:
+	nlmsg_free(skb);
+
+	return err;
+}
+
+/**
+ * nfsd_nl_version_set_doit - set the nfs enabled versions
+ * @skb: reply buffer
+ * @info: netlink metadata and command arguments
+ *
+ * Return 0 on success or a negative errno.
+ */
+int nfsd_nl_version_set_doit(struct sk_buff *skb, struct genl_info *info)
+{
+	const struct nlattr *attr;
+	struct nfsd_net *nn;
+	int i, rem;
+
+	if (GENL_REQ_ATTR_CHECK(info, NFSD_A_SERVER_PROTO_VERSION))
+		return -EINVAL;
+
+	mutex_lock(&nfsd_mutex);
+
+	nn = net_generic(genl_info_net(info), nfsd_net_id);
+	if (nn->nfsd_serv) {
+		mutex_unlock(&nfsd_mutex);
+		return -EBUSY;
+	}
+
+	/* clear current supported versions. */
+	nfsd_vers(nn, 2, NFSD_CLEAR);
+	nfsd_vers(nn, 3, NFSD_CLEAR);
+	for (i = 0; i <= NFSD_SUPPORTED_MINOR_VERSION; i++)
+		nfsd_minorversion(nn, i, NFSD_CLEAR);
+
+	nlmsg_for_each_attr(attr, info->nlhdr, GENL_HDRLEN, rem) {
+		struct nlattr *tb[NFSD_A_VERSION_MAX + 1];
+		u32 major, minor = 0;
+		bool enabled;
+
+		if (nla_type(attr) != NFSD_A_SERVER_PROTO_VERSION)
+			continue;
+
+		if (nla_parse_nested(tb, NFSD_A_VERSION_MAX, attr,
+				     nfsd_version_nl_policy, info->extack) < 0)
+			continue;
+
+		if (!tb[NFSD_A_VERSION_MAJOR])
+			continue;
+
+		major = nla_get_u32(tb[NFSD_A_VERSION_MAJOR]);
+		if (tb[NFSD_A_VERSION_MINOR])
+			minor = nla_get_u32(tb[NFSD_A_VERSION_MINOR]);
+
+		enabled = nla_get_flag(tb[NFSD_A_VERSION_ENABLED]);
+
+		switch (major) {
+		case 4:
+			nfsd_minorversion(nn, minor, enabled ? NFSD_SET : NFSD_CLEAR);
+			break;
+		case 3:
+		case 2:
+			if (!minor)
+				nfsd_vers(nn, major, enabled ? NFSD_SET : NFSD_CLEAR);
+			break;
+		default:
+			break;
+		}
+	}
+
 	mutex_unlock(&nfsd_mutex);
 
 	return 0;
 }
 
 /**
+ * nfsd_nl_version_get_doit - get the enabled status for all supported nfs versions
+ * @skb: reply buffer
+ * @info: netlink metadata and command arguments
+ *
+ * Return 0 on success or a negative errno.
+ */
+int nfsd_nl_version_get_doit(struct sk_buff *skb, struct genl_info *info)
+{
+	struct nfsd_net *nn;
+	int i, err;
+	void *hdr;
+
+	skb = genlmsg_new(GENLMSG_DEFAULT_SIZE, GFP_KERNEL);
+	if (!skb)
+		return -ENOMEM;
+
+	hdr = genlmsg_iput(skb, info);
+	if (!hdr) {
+		err = -EMSGSIZE;
+		goto err_free_msg;
+	}
+
+	mutex_lock(&nfsd_mutex);
+	nn = net_generic(genl_info_net(info), nfsd_net_id);
+
+	for (i = 2; i <= 4; i++) {
+		int j;
+
+		for (j = 0; j <= NFSD_SUPPORTED_MINOR_VERSION; j++) {
+			struct nlattr *attr;
+
+			/* Don't record any versions the kernel doesn't have
+			 * compiled in
+			 */
+			if (!nfsd_support_version(i))
+				continue;
+
+			/* NFSv{2,3} does not support minor numbers */
+			if (i < 4 && j)
+				continue;
+
+			attr = nla_nest_start(skb,
+					      NFSD_A_SERVER_PROTO_VERSION);
+			if (!attr) {
+				err = -EINVAL;
+				goto err_nfsd_unlock;
+			}
+
+			if (nla_put_u32(skb, NFSD_A_VERSION_MAJOR, i) ||
+			    nla_put_u32(skb, NFSD_A_VERSION_MINOR, j)) {
+				err = -EINVAL;
+				goto err_nfsd_unlock;
+			}
+
+			/* Set the enabled flag if the version is enabled */
+			if (nfsd_vers(nn, i, NFSD_TEST) &&
+			    (i < 4 || nfsd_minorversion(nn, j, NFSD_TEST)) &&
+			    nla_put_flag(skb, NFSD_A_VERSION_ENABLED)) {
+				err = -EINVAL;
+				goto err_nfsd_unlock;
+			}
+
+			nla_nest_end(skb, attr);
+		}
+	}
+
+	mutex_unlock(&nfsd_mutex);
+	genlmsg_end(skb, hdr);
+
+	return genlmsg_reply(skb, info);
+
+err_nfsd_unlock:
+	mutex_unlock(&nfsd_mutex);
+err_free_msg:
+	nlmsg_free(skb);
+
+	return err;
+}
+
+/**
+ * nfsd_nl_listener_set_doit - set the nfs running sockets
+ * @skb: reply buffer
+ * @info: netlink metadata and command arguments
+ *
+ * Return 0 on success or a negative errno.
+ */
+int nfsd_nl_listener_set_doit(struct sk_buff *skb, struct genl_info *info)
+{
+	struct net *net = genl_info_net(info);
+	struct svc_xprt *xprt, *tmp;
+	const struct nlattr *attr;
+	struct svc_serv *serv;
+	LIST_HEAD(permsocks);
+	struct nfsd_net *nn;
+	int err, rem;
+
+	mutex_lock(&nfsd_mutex);
+
+	err = nfsd_create_serv(net);
+	if (err) {
+		mutex_unlock(&nfsd_mutex);
+		return err;
+	}
+
+	nn = net_generic(net, nfsd_net_id);
+	serv = nn->nfsd_serv;
+
+	spin_lock_bh(&serv->sv_lock);
+
+	/* Move all of the old listener sockets to a temp list */
+	list_splice_init(&serv->sv_permsocks, &permsocks);
+
+	/*
+	 * Walk the list of server_socks from userland and move any that match
+	 * back to sv_permsocks
+	 */
+	nlmsg_for_each_attr(attr, info->nlhdr, GENL_HDRLEN, rem) {
+		struct nlattr *tb[NFSD_A_SOCK_MAX + 1];
+		const char *xcl_name;
+		struct sockaddr *sa;
+
+		if (nla_type(attr) != NFSD_A_SERVER_SOCK_ADDR)
+			continue;
+
+		if (nla_parse_nested(tb, NFSD_A_SOCK_MAX, attr,
+				     nfsd_sock_nl_policy, info->extack) < 0)
+			continue;
+
+		if (!tb[NFSD_A_SOCK_ADDR] || !tb[NFSD_A_SOCK_TRANSPORT_NAME])
+			continue;
+
+		if (nla_len(tb[NFSD_A_SOCK_ADDR]) < sizeof(*sa))
+			continue;
+
+		xcl_name = nla_data(tb[NFSD_A_SOCK_TRANSPORT_NAME]);
+		sa = nla_data(tb[NFSD_A_SOCK_ADDR]);
+
+		/* Put back any matching sockets */
+		list_for_each_entry_safe(xprt, tmp, &permsocks, xpt_list) {
+			/* This shouldn't be possible */
+			if (WARN_ON_ONCE(xprt->xpt_net != net)) {
+				list_move(&xprt->xpt_list, &serv->sv_permsocks);
+				continue;
+			}
+
+			/* If everything matches, put it back */
+			if (!strcmp(xprt->xpt_class->xcl_name, xcl_name) &&
+			    rpc_cmp_addr_port(sa, (struct sockaddr *)&xprt->xpt_local)) {
+				list_move(&xprt->xpt_list, &serv->sv_permsocks);
+				break;
+			}
+		}
+	}
+
+	/* For now, no removing old sockets while server is running */
+	if (serv->sv_nrthreads && !list_empty(&permsocks)) {
+		list_splice_init(&permsocks, &serv->sv_permsocks);
+		spin_unlock_bh(&serv->sv_lock);
+		err = -EBUSY;
+		goto out_unlock_mtx;
+	}
+
+	/* Close the remaining sockets on the permsocks list */
+	while (!list_empty(&permsocks)) {
+		xprt = list_first_entry(&permsocks, struct svc_xprt, xpt_list);
+		list_move(&xprt->xpt_list, &serv->sv_permsocks);
+
+		/*
+		 * Newly-created sockets are born with the BUSY bit set. Clear
+		 * it if there are no threads, since nothing can pick it up
+		 * in that case.
+		 */
+		if (!serv->sv_nrthreads)
+			clear_bit(XPT_BUSY, &xprt->xpt_flags);
+
+		set_bit(XPT_CLOSE, &xprt->xpt_flags);
+		spin_unlock_bh(&serv->sv_lock);
+		svc_xprt_close(xprt);
+		spin_lock_bh(&serv->sv_lock);
+	}
+
+	spin_unlock_bh(&serv->sv_lock);
+
+	/* walk list of addrs again, open any that still don't exist */
+	nlmsg_for_each_attr(attr, info->nlhdr, GENL_HDRLEN, rem) {
+		struct nlattr *tb[NFSD_A_SOCK_MAX + 1];
+		const char *xcl_name;
+		struct sockaddr *sa;
+		int ret;
+
+		if (nla_type(attr) != NFSD_A_SERVER_SOCK_ADDR)
+			continue;
+
+		if (nla_parse_nested(tb, NFSD_A_SOCK_MAX, attr,
+				     nfsd_sock_nl_policy, info->extack) < 0)
+			continue;
+
+		if (!tb[NFSD_A_SOCK_ADDR] || !tb[NFSD_A_SOCK_TRANSPORT_NAME])
+			continue;
+
+		if (nla_len(tb[NFSD_A_SOCK_ADDR]) < sizeof(*sa))
+			continue;
+
+		xcl_name = nla_data(tb[NFSD_A_SOCK_TRANSPORT_NAME]);
+		sa = nla_data(tb[NFSD_A_SOCK_ADDR]);
+
+		xprt = svc_find_listener(serv, xcl_name, net, sa);
+		if (xprt) {
+			svc_xprt_put(xprt);
+			continue;
+		}
+
+		ret = svc_xprt_create_from_sa(serv, xcl_name, net, sa, 0,
+					      get_current_cred());
+		/* always save the latest error */
+		if (ret < 0)
+			err = ret;
+	}
+
+	if (!serv->sv_nrthreads && list_empty(&nn->nfsd_serv->sv_permsocks))
+		nfsd_destroy_serv(net);
+
+out_unlock_mtx:
+	mutex_unlock(&nfsd_mutex);
+
+	return err;
+}
+
+/**
+ * nfsd_nl_listener_get_doit - get the nfs running listeners
+ * @skb: reply buffer
+ * @info: netlink metadata and command arguments
+ *
+ * Return 0 on success or a negative errno.
+ */
+int nfsd_nl_listener_get_doit(struct sk_buff *skb, struct genl_info *info)
+{
+	struct svc_xprt *xprt;
+	struct svc_serv *serv;
+	struct nfsd_net *nn;
+	void *hdr;
+	int err;
+
+	skb = genlmsg_new(GENLMSG_DEFAULT_SIZE, GFP_KERNEL);
+	if (!skb)
+		return -ENOMEM;
+
+	hdr = genlmsg_iput(skb, info);
+	if (!hdr) {
+		err = -EMSGSIZE;
+		goto err_free_msg;
+	}
+
+	mutex_lock(&nfsd_mutex);
+	nn = net_generic(genl_info_net(info), nfsd_net_id);
+
+	/* no nfs server? Just send empty socket list */
+	if (!nn->nfsd_serv)
+		goto out_unlock_mtx;
+
+	serv = nn->nfsd_serv;
+	spin_lock_bh(&serv->sv_lock);
+	list_for_each_entry(xprt, &serv->sv_permsocks, xpt_list) {
+		struct nlattr *attr;
+
+		attr = nla_nest_start(skb, NFSD_A_SERVER_SOCK_ADDR);
+		if (!attr) {
+			err = -EINVAL;
+			goto err_serv_unlock;
+		}
+
+		if (nla_put_string(skb, NFSD_A_SOCK_TRANSPORT_NAME,
+				   xprt->xpt_class->xcl_name) ||
+		    nla_put(skb, NFSD_A_SOCK_ADDR,
+			    sizeof(struct sockaddr_storage),
+			    &xprt->xpt_local)) {
+			err = -EINVAL;
+			goto err_serv_unlock;
+		}
+
+		nla_nest_end(skb, attr);
+	}
+	spin_unlock_bh(&serv->sv_lock);
+out_unlock_mtx:
+	mutex_unlock(&nfsd_mutex);
+	genlmsg_end(skb, hdr);
+
+	return genlmsg_reply(skb, info);
+
+err_serv_unlock:
+	spin_unlock_bh(&serv->sv_lock);
+	mutex_unlock(&nfsd_mutex);
+err_free_msg:
+	nlmsg_free(skb);
+
+	return err;
+}
+
+/**
  * nfsd_net_init - Prepare the nfsd_net portion of a new net namespace
  * @net: a freshly-created network namespace
  *
@@ -1671,11 +2160,16 @@ static __net_init int nfsd_net_init(struct net *net)
 	retval = nfsd_idmap_init(net);
 	if (retval)
 		goto out_idmap_error;
-	retval = nfsd_stat_counters_init(nn);
+	retval = percpu_counter_init_many(nn->counter, 0, GFP_KERNEL,
+					  NFSD_STATS_COUNTERS_NUM);
 	if (retval)
 		goto out_repcache_error;
+	memset(&nn->nfsd_svcstats, 0, sizeof(nn->nfsd_svcstats));
+	nn->nfsd_svcstats.program = &nfsd_program;
 	nn->nfsd_versions = NULL;
 	nn->nfsd4_minorversions = NULL;
+	nn->nfsd_info.mutex = &nfsd_mutex;
+	nn->nfsd_serv = NULL;
 	nfsd4_init_leases_net(nn);
 	get_random_bytes(&nn->siphash_key, sizeof(nn->siphash_key));
 	seqlock_init(&nn->writeverf_lock);
@@ -1701,7 +2195,7 @@ static __net_exit void nfsd_net_exit(struct net *net)
 	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
 
 	nfsd_proc_stat_shutdown(net);
-	nfsd_stat_counters_destroy(nn);
+	percpu_counter_destroy_many(nn->counter, NFSD_STATS_COUNTERS_NUM);
 	nfsd_idmap_shutdown(net);
 	nfsd_export_shutdown(net);
 	nfsd_netns_free_versions(nn);
diff --git a/fs/nfsd/nfsd.h b/fs/nfsd/nfsd.h
index 304e9728b9..8f4f239d9f 100644
--- a/fs/nfsd/nfsd.h
+++ b/fs/nfsd/nfsd.h
@@ -86,6 +86,7 @@ extern struct mutex		nfsd_mutex;
 extern spinlock_t		nfsd_drc_lock;
 extern unsigned long		nfsd_drc_max_mem;
 extern unsigned long		nfsd_drc_mem_used;
+extern atomic_t			nfsd_th_cnt;		/* number of available threads */
 
 extern const struct seq_operations nfs_exports_op;
 
@@ -102,7 +103,7 @@ bool		nfssvc_encode_voidres(struct svc_rqst *rqstp,
 /*
  * Function prototypes.
  */
-int		nfsd_svc(int nrservs, struct net *net, const struct cred *cred);
+int		nfsd_svc(int nrservs, struct net *net, const struct cred *cred, const char *scope);
 int		nfsd_dispatch(struct svc_rqst *rqstp);
 
 int		nfsd_nrthreads(struct net *);
@@ -229,7 +230,6 @@ void		nfsd_lockd_shutdown(void);
 #define	nfserr_nospc		cpu_to_be32(NFSERR_NOSPC)
 #define	nfserr_rofs		cpu_to_be32(NFSERR_ROFS)
 #define	nfserr_mlink		cpu_to_be32(NFSERR_MLINK)
-#define	nfserr_opnotsupp	cpu_to_be32(NFSERR_OPNOTSUPP)
 #define	nfserr_nametoolong	cpu_to_be32(NFSERR_NAMETOOLONG)
 #define	nfserr_notempty		cpu_to_be32(NFSERR_NOTEMPTY)
 #define	nfserr_dquot		cpu_to_be32(NFSERR_DQUOT)
@@ -274,6 +274,7 @@ void		nfsd_lockd_shutdown(void);
 #define	nfserr_no_grace		cpu_to_be32(NFSERR_NO_GRACE)
 #define	nfserr_reclaim_bad	cpu_to_be32(NFSERR_RECLAIM_BAD)
 #define	nfserr_badname		cpu_to_be32(NFSERR_BADNAME)
+#define	nfserr_admin_revoked	cpu_to_be32(NFS4ERR_ADMIN_REVOKED)
 #define	nfserr_cb_path_down	cpu_to_be32(NFSERR_CB_PATH_DOWN)
 #define	nfserr_locked		cpu_to_be32(NFSERR_LOCKED)
 #define	nfserr_wrongsec		cpu_to_be32(NFSERR_WRONGSEC)
@@ -365,6 +366,7 @@ void		nfsd_lockd_shutdown(void);
 #define	NFSD_CLIENT_MAX_TRIM_PER_RUN	128
 #define	NFS4_CLIENTS_PER_GB		1024
 #define NFSD_DELEGRETURN_TIMEOUT	(HZ / 34)	/* 30ms */
+#define	NFSD_CB_GETATTR_TIMEOUT		NFSD_DELEGRETURN_TIMEOUT
 
 /*
  * The following attributes are currently not supported by the NFSv4 server:
diff --git a/fs/nfsd/nfsfh.c b/fs/nfsd/nfsfh.c
index 40fecf7b22..0b75305fb5 100644
--- a/fs/nfsd/nfsfh.c
+++ b/fs/nfsd/nfsfh.c
@@ -573,7 +573,7 @@ fh_compose(struct svc_fh *fhp, struct svc_export *exp, struct dentry *dentry,
 		_fh_update(fhp, exp, dentry);
 	if (fhp->fh_handle.fh_fileid_type == FILEID_INVALID) {
 		fh_put(fhp);
-		return nfserr_opnotsupp;
+		return nfserr_stale;
 	}
 
 	return 0;
@@ -599,7 +599,7 @@ fh_update(struct svc_fh *fhp)
 
 	_fh_update(fhp, fhp->fh_export, dentry);
 	if (fhp->fh_handle.fh_fileid_type == FILEID_INVALID)
-		return nfserr_opnotsupp;
+		return nfserr_stale;
 	return 0;
 out_bad:
 	printk(KERN_ERR "fh_update: fh not verified!\n");
diff --git a/fs/nfsd/nfsproc.c b/fs/nfsd/nfsproc.c
index a7315928a7..36370b957b 100644
--- a/fs/nfsd/nfsproc.c
+++ b/fs/nfsd/nfsproc.c
@@ -103,7 +103,7 @@ nfsd_proc_setattr(struct svc_rqst *rqstp)
 		}
 	}
 
-	resp->status = nfsd_setattr(rqstp, fhp, &attrs, 0, (time64_t)0);
+	resp->status = nfsd_setattr(rqstp, fhp, &attrs, NULL);
 	if (resp->status != nfs_ok)
 		goto out;
 
@@ -390,8 +390,8 @@ nfsd_proc_create(struct svc_rqst *rqstp)
 		 */
 		attr->ia_valid &= ATTR_SIZE;
 		if (attr->ia_valid)
-			resp->status = nfsd_setattr(rqstp, newfhp, &attrs, 0,
-						    (time64_t)0);
+			resp->status = nfsd_setattr(rqstp, newfhp, &attrs,
+						    NULL);
 	}
 
 out_unlock:
diff --git a/fs/nfsd/nfssvc.c b/fs/nfsd/nfssvc.c
index a667802e08..89d7918de7 100644
--- a/fs/nfsd/nfssvc.c
+++ b/fs/nfsd/nfssvc.c
@@ -34,6 +34,7 @@
 
 #define NFSDDBG_FACILITY	NFSDDBG_SVC
 
+atomic_t			nfsd_th_cnt = ATOMIC_INIT(0);
 extern struct svc_program	nfsd_program;
 static int			nfsd(void *vrqstp);
 #if defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL)
@@ -80,7 +81,6 @@ unsigned long	nfsd_drc_max_mem;
 unsigned long	nfsd_drc_mem_used;
 
 #if defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL)
-static struct svc_stat	nfsd_acl_svcstats;
 static const struct svc_version *nfsd_acl_version[] = {
 # if defined(CONFIG_NFSD_V2_ACL)
 	[2] = &nfsd_acl_version2,
@@ -99,15 +99,11 @@ static struct svc_program	nfsd_acl_program = {
 	.pg_vers		= nfsd_acl_version,
 	.pg_name		= "nfsacl",
 	.pg_class		= "nfsd",
-	.pg_stats		= &nfsd_acl_svcstats,
 	.pg_authenticate	= &svc_set_client,
 	.pg_init_request	= nfsd_acl_init_request,
 	.pg_rpcbind_set		= nfsd_acl_rpcbind_set,
 };
 
-static struct svc_stat	nfsd_acl_svcstats = {
-	.program	= &nfsd_acl_program,
-};
 #endif /* defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL) */
 
 static const struct svc_version *nfsd_version[] = {
@@ -132,14 +128,12 @@ struct svc_program		nfsd_program = {
 	.pg_vers		= nfsd_version,		/* version table */
 	.pg_name		= "nfsd",		/* program name */
 	.pg_class		= "nfsd",		/* authentication class */
-	.pg_stats		= &nfsd_svcstats,	/* version table */
 	.pg_authenticate	= &svc_set_client,	/* export authentication */
 	.pg_init_request	= nfsd_init_request,
 	.pg_rpcbind_set		= nfsd_rpcbind_set,
 };
 
-static bool
-nfsd_support_version(int vers)
+bool nfsd_support_version(int vers)
 {
 	if (vers >= NFSD_MINVERS && vers < NFSD_NRVERS)
 		return nfsd_version[vers] != NULL;
@@ -666,7 +660,8 @@ int nfsd_create_serv(struct net *net)
 	if (nfsd_max_blksize == 0)
 		nfsd_max_blksize = nfsd_get_default_max_blksize();
 	nfsd_reset_versions(nn);
-	serv = svc_create_pooled(&nfsd_program, nfsd_max_blksize, nfsd);
+	serv = svc_create_pooled(&nfsd_program, &nn->nfsd_svcstats,
+				 nfsd_max_blksize, nfsd);
 	if (serv == NULL)
 		return -ENOMEM;
 
@@ -677,7 +672,6 @@ int nfsd_create_serv(struct net *net)
 		return error;
 	}
 	spin_lock(&nfsd_notifier_lock);
-	nn->nfsd_info.mutex = &nfsd_mutex;
 	nn->nfsd_serv = serv;
 	spin_unlock(&nfsd_notifier_lock);
 
@@ -773,13 +767,14 @@ int nfsd_set_nrthreads(int n, int *nthreads, struct net *net)
  * this is the first time nrservs is nonzero.
  */
 int
-nfsd_svc(int nrservs, struct net *net, const struct cred *cred)
+nfsd_svc(int nrservs, struct net *net, const struct cred *cred, const char *scope)
 {
 	int	error;
 	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
 	struct svc_serv *serv;
 
-	mutex_lock(&nfsd_mutex);
+	lockdep_assert_held(&nfsd_mutex);
+
 	dprintk("nfsd: creating service\n");
 
 	nrservs = max(nrservs, 0);
@@ -789,7 +784,7 @@ nfsd_svc(int nrservs, struct net *net, const struct cred *cred)
 	if (nrservs == 0 && nn->nfsd_serv == NULL)
 		goto out;
 
-	strscpy(nn->nfsd_name, utsname()->nodename,
+	strscpy(nn->nfsd_name, scope ? scope : utsname()->nodename,
 		sizeof(nn->nfsd_name));
 
 	error = nfsd_create_serv(net);
@@ -808,7 +803,6 @@ out_put:
 	if (serv->sv_nrthreads == 0)
 		nfsd_destroy_serv(net);
 out:
-	mutex_unlock(&nfsd_mutex);
 	return error;
 }
 
@@ -929,7 +923,7 @@ nfsd(void *vrqstp)
 
 	current->fs->umask = 0;
 
-	atomic_inc(&nfsdstats.th_cnt);
+	atomic_inc(&nfsd_th_cnt);
 
 	set_freezable();
 
@@ -941,9 +935,11 @@ nfsd(void *vrqstp)
 		rqstp->rq_server->sv_maxconn = nn->max_connections;
 
 		svc_recv(rqstp);
+
+		nfsd_file_net_dispose(nn);
 	}
 
-	atomic_dec(&nfsdstats.th_cnt);
+	atomic_dec(&nfsd_th_cnt);
 
 out:
 	/* Release the thread */
diff --git a/fs/nfsd/pnfs.h b/fs/nfsd/pnfs.h
index de1e0dfed0..925817f669 100644
--- a/fs/nfsd/pnfs.h
+++ b/fs/nfsd/pnfs.h
@@ -37,7 +37,8 @@ struct nfsd4_layout_ops {
 	__be32 (*proc_layoutcommit)(struct inode *inode,
 			struct nfsd4_layoutcommit *lcp);
 
-	void (*fence_client)(struct nfs4_layout_stateid *ls);
+	void (*fence_client)(struct nfs4_layout_stateid *ls,
+			     struct nfsd_file *file);
 };
 
 extern const struct nfsd4_layout_ops *nfsd4_layout_ops[];
@@ -72,11 +73,13 @@ void nfsd4_setup_layout_type(struct svc_export *exp);
 void nfsd4_return_all_client_layouts(struct nfs4_client *);
 void nfsd4_return_all_file_layouts(struct nfs4_client *clp,
 		struct nfs4_file *fp);
+void nfsd4_close_layout(struct nfs4_layout_stateid *ls);
 int nfsd4_init_pnfs(void);
 void nfsd4_exit_pnfs(void);
 #else
 struct nfs4_client;
 struct nfs4_file;
+struct nfs4_layout_stateid;
 
 static inline void nfsd4_setup_layout_type(struct svc_export *exp)
 {
@@ -89,6 +92,9 @@ static inline void nfsd4_return_all_file_layouts(struct nfs4_client *clp,
 		struct nfs4_file *fp)
 {
 }
+static inline void nfsd4_close_layout(struct nfs4_layout_stateid *ls)
+{
+}
 static inline void nfsd4_exit_pnfs(void)
 {
 }
diff --git a/fs/nfsd/state.h b/fs/nfsd/state.h
index 0bbbe57e02..ffc217099d 100644
--- a/fs/nfsd/state.h
+++ b/fs/nfsd/state.h
@@ -88,17 +88,34 @@ struct nfsd4_callback_ops {
  */
 struct nfs4_stid {
 	refcount_t		sc_count;
-#define NFS4_OPEN_STID 1
-#define NFS4_LOCK_STID 2
-#define NFS4_DELEG_STID 4
-/* For an open stateid kept around *only* to process close replays: */
-#define NFS4_CLOSED_STID 8
+
+	/* A new stateid is added to the cl_stateids idr early before it
+	 * is fully initialised.  Its sc_type is then zero.  After
+	 * initialisation the sc_type it set under cl_lock, and then
+	 * never changes.
+	 */
+#define SC_TYPE_OPEN		BIT(0)
+#define SC_TYPE_LOCK		BIT(1)
+#define SC_TYPE_DELEG		BIT(2)
+#define SC_TYPE_LAYOUT		BIT(3)
+	unsigned short		sc_type;
+
+/* state_lock protects sc_status for delegation stateids.
+ * ->cl_lock protects sc_status for open and lock stateids.
+ * ->st_mutex also protect sc_status for open stateids.
+ * ->ls_lock protects sc_status for layout stateids.
+ */
+/*
+ * For an open stateid kept around *only* to process close replays.
+ * For deleg stateid, kept in idr until last reference is dropped.
+ */
+#define SC_STATUS_CLOSED	BIT(0)
 /* For a deleg stateid kept around only to process free_stateid's: */
-#define NFS4_REVOKED_DELEG_STID 16
-#define NFS4_CLOSED_DELEG_STID 32
-#define NFS4_LAYOUT_STID 64
+#define SC_STATUS_REVOKED	BIT(1)
+#define SC_STATUS_ADMIN_REVOKED	BIT(2)
+	unsigned short		sc_status;
+
 	struct list_head	sc_cp_list;
-	unsigned char		sc_type;
 	stateid_t		sc_stateid;
 	spinlock_t		sc_lock;
 	struct nfs4_client	*sc_client;
@@ -125,8 +142,16 @@ struct nfs4_cb_fattr {
 	/* from CB_GETATTR reply */
 	u64 ncf_cb_change;
 	u64 ncf_cb_fsize;
+
+	unsigned long ncf_cb_flags;
+	bool ncf_file_modified;
+	u64 ncf_initial_cinfo;
+	u64 ncf_cur_fsize;
 };
 
+/* bits for ncf_cb_flags */
+#define	CB_GETATTR_BUSY		0
+
 /*
  * Represents a delegation stateid. The nfs4_client holds references to these
  * and they are put when it is being destroyed or when the delegation is
@@ -330,8 +355,9 @@ enum {
  * 0. If they are not renewed within a lease period, they become eligible for
  * destruction by the laundromat.
  *
- * These objects can also be destroyed prematurely by the fault injection code,
- * or if the client sends certain forms of SETCLIENTID or EXCHANGE_ID updates.
+ * These objects can also be destroyed if the client sends certain forms of
+ * SETCLIENTID or EXCHANGE_ID operations.
+ *
  * Care is taken *not* to do this however when the objects have an elevated
  * refcount.
  *
@@ -339,7 +365,7 @@ enum {
  *
  * o Each nfs4_clients is also hashed by name (the opaque quantity initially
  *   sent by the client to identify itself).
- * 	  
+ *
  * o cl_perclient list is used to ensure no dangling stateowner references
  *   when we expire the nfs4_client
  */
@@ -364,6 +390,7 @@ struct nfs4_client {
 	clientid_t		cl_clientid;	/* generated by server */
 	nfs4_verifier		cl_confirm;	/* generated by server */
 	u32			cl_minorversion;
+	atomic_t		cl_admin_revoked; /* count of admin-revoked states */
 	/* NFSv4.1 client implementation id: */
 	struct xdr_netobj	cl_nii_domain;
 	struct xdr_netobj	cl_nii_name;
@@ -381,6 +408,8 @@ struct nfs4_client {
 					 1 << NFSD4_CLIENT_CB_KILL)
 #define NFSD4_CLIENT_CB_RECALL_ANY	(6)
 	unsigned long		cl_flags;
+
+	struct workqueue_struct *cl_callback_wq;
 	const struct cred	*cl_cb_cred;
 	struct rpc_clnt		*cl_cb_client;
 	u32			cl_cb_ident;
@@ -459,7 +488,7 @@ struct nfs4_replay {
 	unsigned int		rp_buflen;
 	char			*rp_buf;
 	struct knfsd_fh		rp_openfh;
-	struct mutex		rp_mutex;
+	atomic_t		rp_locked;
 	char			rp_ibuf[NFSD4_REPLAY_ISIZE];
 };
 
@@ -686,15 +715,15 @@ extern __be32 nfs4_preprocess_stateid_op(struct svc_rqst *rqstp,
 		stateid_t *stateid, int flags, struct nfsd_file **filp,
 		struct nfs4_stid **cstid);
 __be32 nfsd4_lookup_stateid(struct nfsd4_compound_state *cstate,
-		     stateid_t *stateid, unsigned char typemask,
-		     struct nfs4_stid **s, struct nfsd_net *nn);
+			    stateid_t *stateid, unsigned short typemask,
+			    unsigned short statusmask,
+			    struct nfs4_stid **s, struct nfsd_net *nn);
 struct nfs4_stid *nfs4_alloc_stid(struct nfs4_client *cl, struct kmem_cache *slab,
 				  void (*sc_free)(struct nfs4_stid *));
 int nfs4_init_copy_state(struct nfsd_net *nn, struct nfsd4_copy *copy);
 void nfs4_free_copy_state(struct nfsd4_copy *copy);
 struct nfs4_cpntf_state *nfs4_alloc_init_cpntf_state(struct nfsd_net *nn,
 			struct nfs4_stid *p_stid);
-void nfs4_unhash_stid(struct nfs4_stid *s);
 void nfs4_put_stid(struct nfs4_stid *s);
 void nfs4_inc_and_copy_stateid(stateid_t *dst, struct nfs4_stid *stid);
 void nfs4_remove_reclaim_record(struct nfs4_client_reclaim *, struct nfsd_net *);
@@ -708,8 +737,6 @@ extern void nfsd4_change_callback(struct nfs4_client *clp, struct nfs4_cb_conn *
 extern void nfsd4_init_cb(struct nfsd4_callback *cb, struct nfs4_client *clp,
 		const struct nfsd4_callback_ops *ops, enum nfsd4_cb_op op);
 extern bool nfsd4_run_cb(struct nfsd4_callback *cb);
-extern int nfsd4_create_callback_queue(void);
-extern void nfsd4_destroy_callback_queue(void);
 extern void nfsd4_shutdown_callback(struct nfs4_client *);
 extern void nfsd4_shutdown_copy(struct nfs4_client *clp);
 extern struct nfs4_client_reclaim *nfs4_client_to_reclaim(struct xdr_netobj name,
@@ -728,6 +755,14 @@ static inline void get_nfs4_file(struct nfs4_file *fi)
 }
 struct nfsd_file *find_any_file(struct nfs4_file *f);
 
+#ifdef CONFIG_NFSD_V4
+void nfsd4_revoke_states(struct net *net, struct super_block *sb);
+#else
+static inline void nfsd4_revoke_states(struct net *net, struct super_block *sb)
+{
+}
+#endif
+
 /* grace period management */
 void nfsd4_end_grace(struct nfsd_net *nn);
 
@@ -746,5 +781,5 @@ static inline bool try_to_expire_client(struct nfs4_client *clp)
 }
 
 extern __be32 nfsd4_deleg_getattr_conflict(struct svc_rqst *rqstp,
-				struct inode *inode);
+		struct inode *inode, bool *file_modified, u64 *size);
 #endif   /* NFSD4_STATE_H */
diff --git a/fs/nfsd/stats.c b/fs/nfsd/stats.c
index 44e275324b..bb22893f11 100644
--- a/fs/nfsd/stats.c
+++ b/fs/nfsd/stats.c
@@ -27,11 +27,6 @@
 
 #include "nfsd.h"
 
-struct nfsd_stats	nfsdstats;
-struct svc_stat		nfsd_svcstats = {
-	.program	= &nfsd_program,
-};
-
 static int nfsd_show(struct seq_file *seq, void *v)
 {
 	struct net *net = pde_data(file_inode(seq->file));
@@ -47,7 +42,7 @@ static int nfsd_show(struct seq_file *seq, void *v)
 		   percpu_counter_sum_positive(&nn->counter[NFSD_STATS_IO_WRITE]));
 
 	/* thread usage: */
-	seq_printf(seq, "th %u 0", atomic_read(&nfsdstats.th_cnt));
+	seq_printf(seq, "th %u 0", atomic_read(&nfsd_th_cnt));
 
 	/* deprecated thread usage histogram stats */
 	for (i = 0; i < 10; i++)
@@ -57,7 +52,7 @@ static int nfsd_show(struct seq_file *seq, void *v)
 	seq_puts(seq, "\nra 0 0 0 0 0 0 0 0 0 0 0 0\n");
 
 	/* show my rpc info */
-	svc_seq_show(seq, &nfsd_svcstats);
+	svc_seq_show(seq, &nn->nfsd_svcstats);
 
 #ifdef CONFIG_NFSD_V4
 	/* Show count for individual nfsv4 operations */
@@ -78,51 +73,11 @@ static int nfsd_show(struct seq_file *seq, void *v)
 
 DEFINE_PROC_SHOW_ATTRIBUTE(nfsd);
 
-int nfsd_percpu_counters_init(struct percpu_counter *counters, int num)
-{
-	int i, err = 0;
-
-	for (i = 0; !err && i < num; i++)
-		err = percpu_counter_init(&counters[i], 0, GFP_KERNEL);
-
-	if (!err)
-		return 0;
-
-	for (; i > 0; i--)
-		percpu_counter_destroy(&counters[i-1]);
-
-	return err;
-}
-
-void nfsd_percpu_counters_reset(struct percpu_counter counters[], int num)
-{
-	int i;
-
-	for (i = 0; i < num; i++)
-		percpu_counter_set(&counters[i], 0);
-}
-
-void nfsd_percpu_counters_destroy(struct percpu_counter counters[], int num)
-{
-	int i;
-
-	for (i = 0; i < num; i++)
-		percpu_counter_destroy(&counters[i]);
-}
-
-int nfsd_stat_counters_init(struct nfsd_net *nn)
-{
-	return nfsd_percpu_counters_init(nn->counter, NFSD_STATS_COUNTERS_NUM);
-}
-
-void nfsd_stat_counters_destroy(struct nfsd_net *nn)
-{
-	nfsd_percpu_counters_destroy(nn->counter, NFSD_STATS_COUNTERS_NUM);
-}
-
 void nfsd_proc_stat_init(struct net *net)
 {
-	svc_proc_register(net, &nfsd_svcstats, &nfsd_proc_ops);
+	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
+
+	svc_proc_register(net, &nn->nfsd_svcstats, &nfsd_proc_ops);
 }
 
 void nfsd_proc_stat_shutdown(struct net *net)
diff --git a/fs/nfsd/stats.h b/fs/nfsd/stats.h
index c24be4ddbe..04aacb6c36 100644
--- a/fs/nfsd/stats.h
+++ b/fs/nfsd/stats.h
@@ -10,19 +10,6 @@
 #include <uapi/linux/nfsd/stats.h>
 #include <linux/percpu_counter.h>
 
-struct nfsd_stats {
-	atomic_t	th_cnt;		/* number of available threads */
-};
-
-extern struct nfsd_stats	nfsdstats;
-
-extern struct svc_stat		nfsd_svcstats;
-
-int nfsd_percpu_counters_init(struct percpu_counter *counters, int num);
-void nfsd_percpu_counters_reset(struct percpu_counter *counters, int num);
-void nfsd_percpu_counters_destroy(struct percpu_counter *counters, int num);
-int nfsd_stat_counters_init(struct nfsd_net *nn);
-void nfsd_stat_counters_destroy(struct nfsd_net *nn);
 void nfsd_proc_stat_init(struct net *net);
 void nfsd_proc_stat_shutdown(struct net *net);
 
diff --git a/fs/nfsd/trace.h b/fs/nfsd/trace.h
index 2cd5703379..77bbd23aa1 100644
--- a/fs/nfsd/trace.h
+++ b/fs/nfsd/trace.h
@@ -9,8 +9,10 @@
 #define _NFSD_TRACE_H
 
 #include <linux/tracepoint.h>
+#include <linux/sunrpc/clnt.h>
 #include <linux/sunrpc/xprt.h>
 #include <trace/misc/nfs.h>
+#include <trace/misc/sunrpc.h>
 
 #include "export.h"
 #include "nfsfh.h"
@@ -102,7 +104,7 @@ TRACE_EVENT(nfsd_compound,
 	TP_fast_assign(
 		__entry->xid = be32_to_cpu(rqst->rq_xid);
 		__entry->opcnt = opcnt;
-		__assign_str_len(tag, tag, taglen);
+		__assign_str(tag);
 	),
 	TP_printk("xid=0x%08x opcnt=%u tag=%s",
 		__entry->xid, __entry->opcnt, __get_str(tag)
@@ -125,7 +127,7 @@ TRACE_EVENT(nfsd_compound_status,
 		__entry->args_opcnt = args_opcnt;
 		__entry->resp_opcnt = resp_opcnt;
 		__entry->status = be32_to_cpu(status);
-		__assign_str(name, name);
+		__assign_str(name);
 	),
 	TP_printk("op=%u/%u %s status=%d",
 		__entry->resp_opcnt, __entry->args_opcnt,
@@ -316,7 +318,7 @@ TRACE_EVENT(nfsd_exp_find_key,
 	TP_fast_assign(
 		__entry->fsidtype = key->ek_fsidtype;
 		memcpy(__entry->fsid, key->ek_fsid, 4*6);
-		__assign_str(auth_domain, key->ek_client->name);
+		__assign_str(auth_domain);
 		__entry->status = status;
 	),
 	TP_printk("fsid=%x::%s domain=%s status=%d",
@@ -340,8 +342,8 @@ TRACE_EVENT(nfsd_expkey_update,
 	TP_fast_assign(
 		__entry->fsidtype = key->ek_fsidtype;
 		memcpy(__entry->fsid, key->ek_fsid, 4*6);
-		__assign_str(auth_domain, key->ek_client->name);
-		__assign_str(path, exp_path);
+		__assign_str(auth_domain);
+		__assign_str(path);
 		__entry->cache = !test_bit(CACHE_NEGATIVE, &key->h.flags);
 	),
 	TP_printk("fsid=%x::%s domain=%s path=%s cache=%s",
@@ -363,8 +365,8 @@ TRACE_EVENT(nfsd_exp_get_by_name,
 		__field(int, status)
 	),
 	TP_fast_assign(
-		__assign_str(path, key->ex_path.dentry->d_name.name);
-		__assign_str(auth_domain, key->ex_client->name);
+		__assign_str(path);
+		__assign_str(auth_domain);
 		__entry->status = status;
 	),
 	TP_printk("path=%s domain=%s status=%d",
@@ -383,8 +385,8 @@ TRACE_EVENT(nfsd_export_update,
 		__field(bool, cache)
 	),
 	TP_fast_assign(
-		__assign_str(path, key->ex_path.dentry->d_name.name);
-		__assign_str(auth_domain, key->ex_client->name);
+		__assign_str(path);
+		__assign_str(auth_domain);
 		__entry->cache = !test_bit(CACHE_NEGATIVE, &key->h.flags);
 	),
 	TP_printk("path=%s domain=%s cache=%s",
@@ -483,7 +485,7 @@ TRACE_EVENT(nfsd_dirent,
 	TP_fast_assign(
 		__entry->fh_hash = fhp ? knfsd_fh_hash(&fhp->fh_handle) : 0;
 		__entry->ino = ino;
-		__assign_str_len(name, name, namlen)
+		__assign_str(name);
 	),
 	TP_printk("fh_hash=0x%08x ino=%llu name=%s",
 		__entry->fh_hash, __entry->ino, __get_str(name)
@@ -641,23 +643,18 @@ DEFINE_EVENT(nfsd_stateseqid_class, nfsd_##name, \
 DEFINE_STATESEQID_EVENT(preprocess);
 DEFINE_STATESEQID_EVENT(open_confirm);
 
-TRACE_DEFINE_ENUM(NFS4_OPEN_STID);
-TRACE_DEFINE_ENUM(NFS4_LOCK_STID);
-TRACE_DEFINE_ENUM(NFS4_DELEG_STID);
-TRACE_DEFINE_ENUM(NFS4_CLOSED_STID);
-TRACE_DEFINE_ENUM(NFS4_REVOKED_DELEG_STID);
-TRACE_DEFINE_ENUM(NFS4_CLOSED_DELEG_STID);
-TRACE_DEFINE_ENUM(NFS4_LAYOUT_STID);
-
 #define show_stid_type(x)						\
 	__print_flags(x, "|",						\
-		{ NFS4_OPEN_STID,		"OPEN" },		\
-		{ NFS4_LOCK_STID,		"LOCK" },		\
-		{ NFS4_DELEG_STID,		"DELEG" },		\
-		{ NFS4_CLOSED_STID,		"CLOSED" },		\
-		{ NFS4_REVOKED_DELEG_STID,	"REVOKED" },		\
-		{ NFS4_CLOSED_DELEG_STID,	"CLOSED_DELEG" },	\
-		{ NFS4_LAYOUT_STID,		"LAYOUT" })
+		{ SC_TYPE_OPEN,		"OPEN" },		\
+		{ SC_TYPE_LOCK,		"LOCK" },		\
+		{ SC_TYPE_DELEG,		"DELEG" },		\
+		{ SC_TYPE_LAYOUT,		"LAYOUT" })
+
+#define show_stid_status(x)						\
+	__print_flags(x, "|",						\
+		{ SC_STATUS_CLOSED,		"CLOSED" },		\
+		{ SC_STATUS_REVOKED,		"REVOKED" },		\
+		{ SC_STATUS_ADMIN_REVOKED,	"ADMIN_REVOKED" })
 
 DECLARE_EVENT_CLASS(nfsd_stid_class,
 	TP_PROTO(
@@ -666,6 +663,7 @@ DECLARE_EVENT_CLASS(nfsd_stid_class,
 	TP_ARGS(stid),
 	TP_STRUCT__entry(
 		__field(unsigned long, sc_type)
+		__field(unsigned long, sc_status)
 		__field(int, sc_count)
 		__field(u32, cl_boot)
 		__field(u32, cl_id)
@@ -676,16 +674,18 @@ DECLARE_EVENT_CLASS(nfsd_stid_class,
 		const stateid_t *stp = &stid->sc_stateid;
 
 		__entry->sc_type = stid->sc_type;
+		__entry->sc_status = stid->sc_status;
 		__entry->sc_count = refcount_read(&stid->sc_count);
 		__entry->cl_boot = stp->si_opaque.so_clid.cl_boot;
 		__entry->cl_id = stp->si_opaque.so_clid.cl_id;
 		__entry->si_id = stp->si_opaque.so_id;
 		__entry->si_generation = stp->si_generation;
 	),
-	TP_printk("client %08x:%08x stateid %08x:%08x ref=%d type=%s",
+	TP_printk("client %08x:%08x stateid %08x:%08x ref=%d type=%s state=%s",
 		__entry->cl_boot, __entry->cl_id,
 		__entry->si_id, __entry->si_generation,
-		__entry->sc_count, show_stid_type(__entry->sc_type)
+		__entry->sc_count, show_stid_type(__entry->sc_type),
+		show_stid_status(__entry->sc_status)
 	)
 );
 
@@ -696,6 +696,129 @@ DEFINE_EVENT(nfsd_stid_class, nfsd_stid_##name,			\
 
 DEFINE_STID_EVENT(revoke);
 
+TRACE_EVENT(nfsd_stateowner_replay,
+	TP_PROTO(
+		u32 opnum,
+		const struct nfs4_replay *rp
+	),
+	TP_ARGS(opnum, rp),
+	TP_STRUCT__entry(
+		__field(unsigned long, status)
+		__field(u32, opnum)
+	),
+	TP_fast_assign(
+		__entry->status = be32_to_cpu(rp->rp_status);
+		__entry->opnum = opnum;
+	),
+	TP_printk("opnum=%u status=%lu",
+		__entry->opnum, __entry->status)
+);
+
+TRACE_EVENT_CONDITION(nfsd_seq4_status,
+	TP_PROTO(
+		const struct svc_rqst *rqstp,
+		const struct nfsd4_sequence *sequence
+	),
+	TP_ARGS(rqstp, sequence),
+	TP_CONDITION(sequence->status_flags),
+	TP_STRUCT__entry(
+		__field(unsigned int, netns_ino)
+		__field(u32, xid)
+		__field(u32, cl_boot)
+		__field(u32, cl_id)
+		__field(u32, seqno)
+		__field(u32, reserved)
+		__field(unsigned long, status_flags)
+	),
+	TP_fast_assign(
+		const struct nfsd4_sessionid *sid =
+			(struct nfsd4_sessionid *)&sequence->sessionid;
+
+		__entry->netns_ino = SVC_NET(rqstp)->ns.inum;
+		__entry->xid = be32_to_cpu(rqstp->rq_xid);
+		__entry->cl_boot = sid->clientid.cl_boot;
+		__entry->cl_id = sid->clientid.cl_id;
+		__entry->seqno = sid->sequence;
+		__entry->reserved = sid->reserved;
+		__entry->status_flags = sequence->status_flags;
+	),
+	TP_printk("xid=0x%08x sessionid=%08x:%08x:%08x:%08x status_flags=%s",
+		__entry->xid, __entry->cl_boot, __entry->cl_id,
+		__entry->seqno, __entry->reserved,
+		show_nfs4_seq4_status(__entry->status_flags)
+	)
+);
+
+DECLARE_EVENT_CLASS(nfsd_cs_slot_class,
+	TP_PROTO(
+		const struct nfs4_client *clp,
+		const struct nfsd4_create_session *cs
+	),
+	TP_ARGS(clp, cs),
+	TP_STRUCT__entry(
+		__field(u32, seqid)
+		__field(u32, slot_seqid)
+		__field(u32, cl_boot)
+		__field(u32, cl_id)
+		__sockaddr(addr, clp->cl_cb_conn.cb_addrlen)
+	),
+	TP_fast_assign(
+		const struct nfsd4_clid_slot *slot = &clp->cl_cs_slot;
+
+		__entry->cl_boot = clp->cl_clientid.cl_boot;
+		__entry->cl_id = clp->cl_clientid.cl_id;
+		__assign_sockaddr(addr, &clp->cl_cb_conn.cb_addr,
+				  clp->cl_cb_conn.cb_addrlen);
+		__entry->seqid = cs->seqid;
+		__entry->slot_seqid = slot->sl_seqid;
+	),
+	TP_printk("addr=%pISpc client %08x:%08x seqid=%u slot_seqid=%u",
+		__get_sockaddr(addr), __entry->cl_boot, __entry->cl_id,
+		__entry->seqid, __entry->slot_seqid
+	)
+);
+
+#define DEFINE_CS_SLOT_EVENT(name) \
+DEFINE_EVENT(nfsd_cs_slot_class, nfsd_##name, \
+	TP_PROTO( \
+		const struct nfs4_client *clp, \
+		const struct nfsd4_create_session *cs \
+	), \
+	TP_ARGS(clp, cs))
+
+DEFINE_CS_SLOT_EVENT(slot_seqid_conf);
+DEFINE_CS_SLOT_EVENT(slot_seqid_unconf);
+
+TRACE_EVENT(nfsd_slot_seqid_sequence,
+	TP_PROTO(
+		const struct nfs4_client *clp,
+		const struct nfsd4_sequence *seq,
+		const struct nfsd4_slot *slot
+	),
+	TP_ARGS(clp, seq, slot),
+	TP_STRUCT__entry(
+		__field(u32, seqid)
+		__field(u32, slot_seqid)
+		__field(u32, cl_boot)
+		__field(u32, cl_id)
+		__sockaddr(addr, clp->cl_cb_conn.cb_addrlen)
+		__field(bool, in_use)
+	),
+	TP_fast_assign(
+		__entry->cl_boot = clp->cl_clientid.cl_boot;
+		__entry->cl_id = clp->cl_clientid.cl_id;
+		__assign_sockaddr(addr, &clp->cl_cb_conn.cb_addr,
+				  clp->cl_cb_conn.cb_addrlen);
+		__entry->seqid = seq->seqid;
+		__entry->slot_seqid = slot->sl_seqid;
+	),
+	TP_printk("addr=%pISpc client %08x:%08x seqid=%u slot_seqid=%u (%sin use)",
+		__get_sockaddr(addr), __entry->cl_boot, __entry->cl_id,
+		__entry->seqid, __entry->slot_seqid,
+		__entry->in_use ? "" : "not "
+	)
+);
+
 DECLARE_EVENT_CLASS(nfsd_clientid_class,
 	TP_PROTO(const clientid_t *clid),
 	TP_ARGS(clid),
@@ -725,6 +848,30 @@ DEFINE_CLIENTID_EVENT(purged);
 DEFINE_CLIENTID_EVENT(renew);
 DEFINE_CLIENTID_EVENT(stale);
 
+TRACE_EVENT(nfsd_mark_client_expired,
+	TP_PROTO(
+		const struct nfs4_client *clp,
+		int cl_rpc_users
+	),
+	TP_ARGS(clp, cl_rpc_users),
+	TP_STRUCT__entry(
+		__field(int, cl_rpc_users)
+		__field(u32, cl_boot)
+		__field(u32, cl_id)
+		__sockaddr(addr, clp->cl_cb_conn.cb_addrlen)
+	),
+	TP_fast_assign(
+		__entry->cl_rpc_users = cl_rpc_users;
+		__entry->cl_boot = clp->cl_clientid.cl_boot;
+		__entry->cl_id = clp->cl_clientid.cl_id;
+		__assign_sockaddr(addr, &clp->cl_cb_conn.cb_addr,
+				  clp->cl_cb_conn.cb_addrlen)
+	),
+	TP_printk("addr=%pISpc client %08x:%08x cl_rpc_users=%d",
+		__get_sockaddr(addr), __entry->cl_boot, __entry->cl_id,
+		__entry->cl_rpc_users)
+);
+
 DECLARE_EVENT_CLASS(nfsd_net_class,
 	TP_PROTO(const struct nfsd_net *nn),
 	TP_ARGS(nn),
@@ -853,7 +1000,7 @@ DECLARE_EVENT_CLASS(nfsd_clid_class,
 		__entry->flavor = clp->cl_cred.cr_flavor;
 		memcpy(__entry->verifier, (void *)&clp->cl_verifier,
 		       NFS4_VERIFIER_SIZE);
-		__assign_str_len(name, clp->cl_name.data, clp->cl_name.len);
+		__assign_str(name);
 	),
 	TP_printk("addr=%pISpc name='%s' verifier=0x%s flavor=%s client=%08x:%08x",
 		__entry->addr, __get_str(name),
@@ -1334,7 +1481,8 @@ DEFINE_EVENT(nfsd_cb_class, nfsd_cb_##name,		\
 	TP_PROTO(const struct nfs4_client *clp),	\
 	TP_ARGS(clp))
 
-DEFINE_NFSD_CB_EVENT(state);
+DEFINE_NFSD_CB_EVENT(start);
+DEFINE_NFSD_CB_EVENT(new_state);
 DEFINE_NFSD_CB_EVENT(probe);
 DEFINE_NFSD_CB_EVENT(lost);
 DEFINE_NFSD_CB_EVENT(shutdown);
@@ -1371,7 +1519,7 @@ TRACE_EVENT(nfsd_cb_setup,
 	TP_fast_assign(
 		__entry->cl_boot = clp->cl_clientid.cl_boot;
 		__entry->cl_id = clp->cl_clientid.cl_id;
-		__assign_str(netid, netid);
+		__assign_str(netid);
 		__entry->authflavor = authflavor;
 		__assign_sockaddr(addr, &clp->cl_cb_conn.cb_addr,
 				  clp->cl_cb_conn.cb_addrlen)
@@ -1405,6 +1553,128 @@ TRACE_EVENT(nfsd_cb_setup_err,
 		__entry->error)
 );
 
+DECLARE_EVENT_CLASS(nfsd_cb_lifetime_class,
+	TP_PROTO(
+		const struct nfs4_client *clp,
+		const struct nfsd4_callback *cb
+	),
+	TP_ARGS(clp, cb),
+	TP_STRUCT__entry(
+		__field(u32, cl_boot)
+		__field(u32, cl_id)
+		__field(const void *, cb)
+		__field(bool, need_restart)
+		__sockaddr(addr, clp->cl_cb_conn.cb_addrlen)
+	),
+	TP_fast_assign(
+		__entry->cl_boot = clp->cl_clientid.cl_boot;
+		__entry->cl_id = clp->cl_clientid.cl_id;
+		__entry->cb = cb;
+		__entry->need_restart = cb->cb_need_restart;
+		__assign_sockaddr(addr, &clp->cl_cb_conn.cb_addr,
+				  clp->cl_cb_conn.cb_addrlen)
+	),
+	TP_printk("addr=%pISpc client %08x:%08x cb=%p%s",
+		__get_sockaddr(addr), __entry->cl_boot, __entry->cl_id,
+		__entry->cb, __entry->need_restart ?
+			" (need restart)" : " (first try)"
+	)
+);
+
+#define DEFINE_NFSD_CB_LIFETIME_EVENT(name)		\
+DEFINE_EVENT(nfsd_cb_lifetime_class, nfsd_cb_##name,	\
+	TP_PROTO(					\
+		const struct nfs4_client *clp,		\
+		const struct nfsd4_callback *cb		\
+	),						\
+	TP_ARGS(clp, cb))
+
+DEFINE_NFSD_CB_LIFETIME_EVENT(queue);
+DEFINE_NFSD_CB_LIFETIME_EVENT(destroy);
+DEFINE_NFSD_CB_LIFETIME_EVENT(restart);
+DEFINE_NFSD_CB_LIFETIME_EVENT(bc_update);
+DEFINE_NFSD_CB_LIFETIME_EVENT(bc_shutdown);
+
+TRACE_EVENT(nfsd_cb_seq_status,
+	TP_PROTO(
+		const struct rpc_task *task,
+		const struct nfsd4_callback *cb
+	),
+	TP_ARGS(task, cb),
+	TP_STRUCT__entry(
+		__field(unsigned int, task_id)
+		__field(unsigned int, client_id)
+		__field(u32, cl_boot)
+		__field(u32, cl_id)
+		__field(u32, seqno)
+		__field(u32, reserved)
+		__field(int, tk_status)
+		__field(int, seq_status)
+	),
+	TP_fast_assign(
+		const struct nfs4_client *clp = cb->cb_clp;
+		const struct nfsd4_session *session = clp->cl_cb_session;
+		const struct nfsd4_sessionid *sid =
+			(struct nfsd4_sessionid *)&session->se_sessionid;
+
+		__entry->task_id = task->tk_pid;
+		__entry->client_id = task->tk_client ?
+				     task->tk_client->cl_clid : -1;
+		__entry->cl_boot = sid->clientid.cl_boot;
+		__entry->cl_id = sid->clientid.cl_id;
+		__entry->seqno = sid->sequence;
+		__entry->reserved = sid->reserved;
+		__entry->tk_status = task->tk_status;
+		__entry->seq_status = cb->cb_seq_status;
+	),
+	TP_printk(SUNRPC_TRACE_TASK_SPECIFIER
+		" sessionid=%08x:%08x:%08x:%08x tk_status=%d seq_status=%d",
+		__entry->task_id, __entry->client_id,
+		__entry->cl_boot, __entry->cl_id,
+		__entry->seqno, __entry->reserved,
+		__entry->tk_status, __entry->seq_status
+	)
+);
+
+TRACE_EVENT(nfsd_cb_free_slot,
+	TP_PROTO(
+		const struct rpc_task *task,
+		const struct nfsd4_callback *cb
+	),
+	TP_ARGS(task, cb),
+	TP_STRUCT__entry(
+		__field(unsigned int, task_id)
+		__field(unsigned int, client_id)
+		__field(u32, cl_boot)
+		__field(u32, cl_id)
+		__field(u32, seqno)
+		__field(u32, reserved)
+		__field(u32, slot_seqno)
+	),
+	TP_fast_assign(
+		const struct nfs4_client *clp = cb->cb_clp;
+		const struct nfsd4_session *session = clp->cl_cb_session;
+		const struct nfsd4_sessionid *sid =
+			(struct nfsd4_sessionid *)&session->se_sessionid;
+
+		__entry->task_id = task->tk_pid;
+		__entry->client_id = task->tk_client ?
+				     task->tk_client->cl_clid : -1;
+		__entry->cl_boot = sid->clientid.cl_boot;
+		__entry->cl_id = sid->clientid.cl_id;
+		__entry->seqno = sid->sequence;
+		__entry->reserved = sid->reserved;
+		__entry->slot_seqno = session->se_cb_seq_nr;
+	),
+	TP_printk(SUNRPC_TRACE_TASK_SPECIFIER
+		" sessionid=%08x:%08x:%08x:%08x new slot seqno=%u",
+		__entry->task_id, __entry->client_id,
+		__entry->cl_boot, __entry->cl_id,
+		__entry->seqno, __entry->reserved,
+		__entry->slot_seqno
+	)
+);
+
 TRACE_EVENT_CONDITION(nfsd_cb_recall,
 	TP_PROTO(
 		const struct nfs4_stid *stid
@@ -1594,7 +1864,7 @@ TRACE_EVENT(nfsd_ctl_unlock_ip,
 	),
 	TP_fast_assign(
 		__entry->netns_ino = net->ns.inum;
-		__assign_str(address, address);
+		__assign_str(address);
 	),
 	TP_printk("address=%s",
 		__get_str(address)
@@ -1613,7 +1883,7 @@ TRACE_EVENT(nfsd_ctl_unlock_fs,
 	),
 	TP_fast_assign(
 		__entry->netns_ino = net->ns.inum;
-		__assign_str(path, path);
+		__assign_str(path);
 	),
 	TP_printk("path=%s",
 		__get_str(path)
@@ -1637,8 +1907,8 @@ TRACE_EVENT(nfsd_ctl_filehandle,
 	TP_fast_assign(
 		__entry->netns_ino = net->ns.inum;
 		__entry->maxsize = maxsize;
-		__assign_str(domain, domain);
-		__assign_str(path, path);
+		__assign_str(domain);
+		__assign_str(path);
 	),
 	TP_printk("domain=%s path=%s maxsize=%d",
 		__get_str(domain), __get_str(path), __entry->maxsize
@@ -1698,7 +1968,7 @@ TRACE_EVENT(nfsd_ctl_version,
 	),
 	TP_fast_assign(
 		__entry->netns_ino = net->ns.inum;
-		__assign_str(mesg, mesg);
+		__assign_str(mesg);
 	),
 	TP_printk("%s",
 		__get_str(mesg)
@@ -1739,7 +2009,7 @@ TRACE_EVENT(nfsd_ctl_ports_addxprt,
 	TP_fast_assign(
 		__entry->netns_ino = net->ns.inum;
 		__entry->port = port;
-		__assign_str(transport, transport);
+		__assign_str(transport);
 	),
 	TP_printk("transport=%s port=%d",
 		__get_str(transport), __entry->port
@@ -1800,9 +2070,9 @@ TRACE_EVENT(nfsd_ctl_time,
 	TP_fast_assign(
 		__entry->netns_ino = net->ns.inum;
 		__entry->time = time;
-		__assign_str_len(name, name, namelen);
+		__assign_str(name);
 	),
-	TP_printk("file=%s time=%d\n",
+	TP_printk("file=%s time=%d",
 		__get_str(name), __entry->time
 	)
 );
@@ -1819,7 +2089,7 @@ TRACE_EVENT(nfsd_ctl_recoverydir,
 	),
 	TP_fast_assign(
 		__entry->netns_ino = net->ns.inum;
-		__assign_str(recdir, recdir);
+		__assign_str(recdir);
 	),
 	TP_printk("recdir=%s",
 		__get_str(recdir)
diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c
index 9b5e20b2d0..29b1f36138 100644
--- a/fs/nfsd/vfs.c
+++ b/fs/nfsd/vfs.c
@@ -25,7 +25,6 @@
 #include <linux/posix_acl_xattr.h>
 #include <linux/xattr.h>
 #include <linux/jhash.h>
-#include <linux/ima.h>
 #include <linux/pagemap.h>
 #include <linux/slab.h>
 #include <linux/uaccess.h>
@@ -476,7 +475,6 @@ static int __nfsd_setattr(struct dentry *dentry, struct iattr *iap)
  * @rqstp: controlling RPC transaction
  * @fhp: filehandle of target
  * @attr: attributes to set
- * @check_guard: set to 1 if guardtime is a valid timestamp
  * @guardtime: do not act if ctime.tv_sec does not match this timestamp
  *
  * This call may adjust the contents of @attr (in particular, this
@@ -488,8 +486,7 @@ static int __nfsd_setattr(struct dentry *dentry, struct iattr *iap)
  */
 __be32
 nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
-	     struct nfsd_attrs *attr,
-	     int check_guard, time64_t guardtime)
+	     struct nfsd_attrs *attr, const struct timespec64 *guardtime)
 {
 	struct dentry	*dentry;
 	struct inode	*inode;
@@ -497,7 +494,7 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	int		accmode = NFSD_MAY_SATTR;
 	umode_t		ftype = 0;
 	__be32		err;
-	int		host_err;
+	int		host_err = 0;
 	bool		get_write_count;
 	bool		size_change = (iap->ia_valid & ATTR_SIZE);
 	int		retries;
@@ -538,9 +535,6 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 
 	nfsd_sanitize_attrs(inode, iap);
 
-	if (check_guard && guardtime != inode_get_ctime_sec(inode))
-		return nfserr_notsync;
-
 	/*
 	 * The size case is special, it changes the file in addition to the
 	 * attributes, and file systems don't expect it to be mixed with
@@ -555,6 +549,19 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	}
 
 	inode_lock(inode);
+	err = fh_fill_pre_attrs(fhp);
+	if (err)
+		goto out_unlock;
+
+	if (guardtime) {
+		struct timespec64 ctime = inode_get_ctime(inode);
+		if ((u32)guardtime->tv_sec != (u32)ctime.tv_sec ||
+		    guardtime->tv_nsec != ctime.tv_nsec) {
+			err = nfserr_notsync;
+			goto out_fill_attrs;
+		}
+	}
+
 	for (retries = 1;;) {
 		struct iattr attrs;
 
@@ -582,13 +589,23 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 		attr->na_aclerr = set_posix_acl(&nop_mnt_idmap,
 						dentry, ACL_TYPE_DEFAULT,
 						attr->na_dpacl);
+out_fill_attrs:
+	/*
+	 * RFC 1813 Section 3.3.2 does not mandate that an NFS server
+	 * returns wcc_data for SETATTR. Some client implementations
+	 * depend on receiving wcc_data, however, to sort out partial
+	 * updates (eg., the client requested that size and mode be
+	 * modified, but the server changed only the file mode).
+	 */
+	fh_fill_post_attrs(fhp);
+out_unlock:
 	inode_unlock(inode);
 	if (size_change)
 		put_write_access(inode);
 out:
 	if (!host_err)
 		host_err = commit_metadata(fhp);
-	return nfserrno(host_err);
+	return err != 0 ? err : nfserrno(host_err);
 }
 
 #if defined(CONFIG_NFSD_V4)
@@ -877,7 +894,7 @@ __nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, umode_t type,
 		goto out;
 	}
 
-	host_err = ima_file_check(file, may_flags);
+	host_err = security_file_post_open(file, may_flags);
 	if (host_err) {
 		fput(file);
 		goto out;
@@ -1405,8 +1422,8 @@ nfsd_create_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	 * Callers expect new file metadata to be committed even
 	 * if the attributes have not changed.
 	 */
-	if (iap->ia_valid)
-		status = nfsd_setattr(rqstp, resfhp, attrs, 0, (time64_t)0);
+	if (nfsd_attrs_valid(attrs))
+		status = nfsd_setattr(rqstp, resfhp, attrs, NULL);
 	else
 		status = nfserrno(commit_metadata(resfhp));
 
@@ -1909,10 +1926,10 @@ out_want_write:
 	fh_drop_write(ffhp);
 
 	/*
-	 * If the target dentry has cached open files, then we need to try to
-	 * close them prior to doing the rename. Flushing delayed fput
-	 * shouldn't be done with locks held however, so we delay it until this
-	 * point and then reattempt the whole shebang.
+	 * If the target dentry has cached open files, then we need to
+	 * try to close them prior to doing the rename.  Final fput
+	 * shouldn't be done with locks held however, so we delay it
+	 * until this point and then reattempt the whole shebang.
 	 */
 	if (close_cached) {
 		close_cached = false;
@@ -2180,11 +2197,43 @@ nfsd_readdir(struct svc_rqst *rqstp, struct svc_fh *fhp, loff_t *offsetp,
 	if (err == nfserr_eof || err == nfserr_toosmall)
 		err = nfs_ok; /* can still be found in ->err */
 out_close:
-	fput(file);
+	nfsd_filp_close(file);
 out:
 	return err;
 }
 
+/**
+ * nfsd_filp_close: close a file synchronously
+ * @fp: the file to close
+ *
+ * nfsd_filp_close() is similar in behaviour to filp_close().
+ * The difference is that if this is the final close on the
+ * file, the that finalisation happens immediately, rather then
+ * being handed over to a work_queue, as it the case for
+ * filp_close().
+ * When a user-space process closes a file (even when using
+ * filp_close() the finalisation happens before returning to
+ * userspace, so it is effectively synchronous.  When a kernel thread
+ * uses file_close(), on the other hand, the handling is completely
+ * asynchronous.  This means that any cost imposed by that finalisation
+ * is not imposed on the nfsd thread, and nfsd could potentually
+ * close files more quickly than the work queue finalises the close,
+ * which would lead to unbounded growth in the queue.
+ *
+ * In some contexts is it not safe to synchronously wait for
+ * close finalisation (see comment for __fput_sync()), but nfsd
+ * does not match those contexts.  In partcilarly it does not, at the
+ * time that this function is called, hold and locks and no finalisation
+ * of any file, socket, or device driver would have any cause to wait
+ * for nfsd to make progress.
+ */
+void nfsd_filp_close(struct file *fp)
+{
+	get_file(fp);
+	filp_close(fp, NULL);
+	__fput_sync(fp);
+}
+
 /*
  * Get file system stats
  * N.B. After this call fhp needs an fh_put
diff --git a/fs/nfsd/vfs.h b/fs/nfsd/vfs.h
index 702fbc4483..57cd700620 100644
--- a/fs/nfsd/vfs.h
+++ b/fs/nfsd/vfs.h
@@ -60,6 +60,14 @@ static inline void nfsd_attrs_free(struct nfsd_attrs *attrs)
 	posix_acl_release(attrs->na_dpacl);
 }
 
+static inline bool nfsd_attrs_valid(struct nfsd_attrs *attrs)
+{
+	struct iattr *iap = attrs->na_iattr;
+
+	return (iap->ia_valid || (attrs->na_seclabel &&
+		attrs->na_seclabel->len));
+}
+
 __be32		nfserrno (int errno);
 int		nfsd_cross_mnt(struct svc_rqst *rqstp, struct dentry **dpp,
 		                struct svc_export **expp);
@@ -69,7 +77,7 @@ __be32		 nfsd_lookup_dentry(struct svc_rqst *, struct svc_fh *,
 				const char *, unsigned int,
 				struct svc_export **, struct dentry **);
 __be32		nfsd_setattr(struct svc_rqst *, struct svc_fh *,
-				struct nfsd_attrs *, int, time64_t);
+			     struct nfsd_attrs *, const struct timespec64 *);
 int nfsd_mountpoint(struct dentry *, struct svc_export *);
 #ifdef CONFIG_NFSD_V4
 __be32		nfsd4_vfs_fallocate(struct svc_rqst *, struct svc_fh *,
@@ -148,6 +156,8 @@ __be32		nfsd_statfs(struct svc_rqst *, struct svc_fh *,
 __be32		nfsd_permission(struct svc_rqst *, struct svc_export *,
 				struct dentry *, int);
 
+void		nfsd_filp_close(struct file *fp);
+
 static inline int fh_want_write(struct svc_fh *fh)
 {
 	int ret;
diff --git a/fs/nfsd/xdr3.h b/fs/nfsd/xdr3.h
index 03fe4e2130..522067b7fd 100644
--- a/fs/nfsd/xdr3.h
+++ b/fs/nfsd/xdr3.h
@@ -14,7 +14,7 @@ struct nfsd3_sattrargs {
 	struct svc_fh		fh;
 	struct iattr		attrs;
 	int			check_guard;
-	time64_t		guardtime;
+	struct timespec64	guardtime;
 };
 
 struct nfsd3_diropargs {
diff --git a/fs/nfsd/xdr4.h b/fs/nfsd/xdr4.h
index 415516c1b2..fbdd42cde1 100644
--- a/fs/nfsd/xdr4.h
+++ b/fs/nfsd/xdr4.h
@@ -518,6 +518,24 @@ struct nfsd4_free_stateid {
 	stateid_t	fr_stateid;         /* request */
 };
 
+struct nfsd4_get_dir_delegation {
+	/* request */
+	u32			gdda_signal_deleg_avail;
+	u32			gdda_notification_types[1];
+	struct timespec64	gdda_child_attr_delay;
+	struct timespec64	gdda_dir_attr_delay;
+	u32			gdda_child_attributes[3];
+	u32			gdda_dir_attributes[3];
+	/* response */
+	u32			gddrnf_status;
+	nfs4_verifier		gddr_cookieverf;
+	stateid_t		gddr_stateid;
+	u32			gddr_notification[1];
+	u32			gddr_child_attributes[3];
+	u32			gddr_dir_attributes[3];
+	bool			gddrnf_will_signal_deleg_avail;
+};
+
 /* also used for NVERIFY */
 struct nfsd4_verify {
 	u32		ve_bmval[3];        /* request */
@@ -674,8 +692,10 @@ struct nfsd4_copy {
 #define NFSD4_COPY_F_INTRA		(1)
 #define NFSD4_COPY_F_SYNCHRONOUS	(2)
 #define NFSD4_COPY_F_COMMITTED		(3)
+#define NFSD4_COPY_F_COMPLETED		(4)
 
 	/* response */
+	__be32			nfserr;
 	struct nfsd42_write_res	cp_res;
 	struct knfsd_fh		fh;
 
@@ -735,7 +755,8 @@ struct nfsd4_offload_status {
 
 	/* response */
 	u64		count;
-	u32		status;
+	__be32		status;
+	bool		completed;
 };
 
 struct nfsd4_copy_notify {
@@ -797,6 +818,7 @@ struct nfsd4_op {
 		struct nfsd4_reclaim_complete	reclaim_complete;
 		struct nfsd4_test_stateid	test_stateid;
 		struct nfsd4_free_stateid	free_stateid;
+		struct nfsd4_get_dir_delegation	get_dir_delegation;
 		struct nfsd4_getdeviceinfo	getdeviceinfo;
 		struct nfsd4_layoutget		layoutget;
 		struct nfsd4_layoutcommit	layoutcommit;
diff --git a/fs/nilfs2/alloc.c b/fs/nilfs2/alloc.c
index 7342de296e..ba50388ee4 100644
--- a/fs/nilfs2/alloc.c
+++ b/fs/nilfs2/alloc.c
@@ -377,11 +377,12 @@ void *nilfs_palloc_block_get_entry(const struct inode *inode, __u64 nr,
  * @target: offset number of an entry in the group (start point)
  * @bsize: size in bits
  * @lock: spin lock protecting @bitmap
+ * @wrap: whether to wrap around
  */
 static int nilfs_palloc_find_available_slot(unsigned char *bitmap,
 					    unsigned long target,
 					    unsigned int bsize,
-					    spinlock_t *lock)
+					    spinlock_t *lock, bool wrap)
 {
 	int pos, end = bsize;
 
@@ -397,6 +398,8 @@ static int nilfs_palloc_find_available_slot(unsigned char *bitmap,
 
 		end = target;
 	}
+	if (!wrap)
+		return -ENOSPC;
 
 	/* wrap around */
 	for (pos = 0; pos < end; pos++) {
@@ -495,9 +498,10 @@ int nilfs_palloc_count_max_entries(struct inode *inode, u64 nused, u64 *nmaxp)
  * nilfs_palloc_prepare_alloc_entry - prepare to allocate a persistent object
  * @inode: inode of metadata file using this allocator
  * @req: nilfs_palloc_req structure exchanged for the allocation
+ * @wrap: whether to wrap around
  */
 int nilfs_palloc_prepare_alloc_entry(struct inode *inode,
-				     struct nilfs_palloc_req *req)
+				     struct nilfs_palloc_req *req, bool wrap)
 {
 	struct buffer_head *desc_bh, *bitmap_bh;
 	struct nilfs_palloc_group_desc *desc;
@@ -516,7 +520,7 @@ int nilfs_palloc_prepare_alloc_entry(struct inode *inode,
 	entries_per_group = nilfs_palloc_entries_per_group(inode);
 
 	for (i = 0; i < ngroups; i += n) {
-		if (group >= ngroups) {
+		if (group >= ngroups && wrap) {
 			/* wrap around */
 			group = 0;
 			maxgroup = nilfs_palloc_group(inode, req->pr_entry_nr,
@@ -525,54 +529,62 @@ int nilfs_palloc_prepare_alloc_entry(struct inode *inode,
 		ret = nilfs_palloc_get_desc_block(inode, group, 1, &desc_bh);
 		if (ret < 0)
 			return ret;
-		desc_kaddr = kmap(desc_bh->b_page);
+		desc_kaddr = kmap_local_page(desc_bh->b_page);
 		desc = nilfs_palloc_block_get_group_desc(
 			inode, group, desc_bh, desc_kaddr);
 		n = nilfs_palloc_rest_groups_in_desc_block(inode, group,
 							   maxgroup);
-		for (j = 0; j < n; j++, desc++, group++) {
+		for (j = 0; j < n; j++, desc++, group++, group_offset = 0) {
 			lock = nilfs_mdt_bgl_lock(inode, group);
-			if (nilfs_palloc_group_desc_nfrees(desc, lock) > 0) {
-				ret = nilfs_palloc_get_bitmap_block(
-					inode, group, 1, &bitmap_bh);
-				if (ret < 0)
-					goto out_desc;
-				bitmap_kaddr = kmap(bitmap_bh->b_page);
-				bitmap = bitmap_kaddr + bh_offset(bitmap_bh);
-				pos = nilfs_palloc_find_available_slot(
-					bitmap, group_offset,
-					entries_per_group, lock);
-				if (pos >= 0) {
-					/* found a free entry */
-					nilfs_palloc_group_desc_add_entries(
-						desc, lock, -1);
-					req->pr_entry_nr =
-						entries_per_group * group + pos;
-					kunmap(desc_bh->b_page);
-					kunmap(bitmap_bh->b_page);
-
-					req->pr_desc_bh = desc_bh;
-					req->pr_bitmap_bh = bitmap_bh;
-					return 0;
-				}
-				kunmap(bitmap_bh->b_page);
-				brelse(bitmap_bh);
+			if (nilfs_palloc_group_desc_nfrees(desc, lock) == 0)
+				continue;
+
+			kunmap_local(desc_kaddr);
+			ret = nilfs_palloc_get_bitmap_block(inode, group, 1,
+							    &bitmap_bh);
+			if (unlikely(ret < 0)) {
+				brelse(desc_bh);
+				return ret;
 			}
 
-			group_offset = 0;
+			desc_kaddr = kmap_local_page(desc_bh->b_page);
+			desc = nilfs_palloc_block_get_group_desc(
+				inode, group, desc_bh, desc_kaddr);
+
+			bitmap_kaddr = kmap_local_page(bitmap_bh->b_page);
+			bitmap = bitmap_kaddr + bh_offset(bitmap_bh);
+			pos = nilfs_palloc_find_available_slot(
+				bitmap, group_offset, entries_per_group, lock,
+				wrap);
+			/*
+			 * Since the search for a free slot in the second and
+			 * subsequent bitmap blocks always starts from the
+			 * beginning, the wrap flag only has an effect on the
+			 * first search.
+			 */
+			kunmap_local(bitmap_kaddr);
+			if (pos >= 0)
+				goto found;
+
+			brelse(bitmap_bh);
 		}
 
-		kunmap(desc_bh->b_page);
+		kunmap_local(desc_kaddr);
 		brelse(desc_bh);
 	}
 
 	/* no entries left */
 	return -ENOSPC;
 
- out_desc:
-	kunmap(desc_bh->b_page);
-	brelse(desc_bh);
-	return ret;
+found:
+	/* found a free entry */
+	nilfs_palloc_group_desc_add_entries(desc, lock, -1);
+	req->pr_entry_nr = entries_per_group * group + pos;
+	kunmap_local(desc_kaddr);
+
+	req->pr_desc_bh = desc_bh;
+	req->pr_bitmap_bh = bitmap_bh;
+	return 0;
 }
 
 /**
@@ -606,10 +618,10 @@ void nilfs_palloc_commit_free_entry(struct inode *inode,
 	spinlock_t *lock;
 
 	group = nilfs_palloc_group(inode, req->pr_entry_nr, &group_offset);
-	desc_kaddr = kmap(req->pr_desc_bh->b_page);
+	desc_kaddr = kmap_local_page(req->pr_desc_bh->b_page);
 	desc = nilfs_palloc_block_get_group_desc(inode, group,
 						 req->pr_desc_bh, desc_kaddr);
-	bitmap_kaddr = kmap(req->pr_bitmap_bh->b_page);
+	bitmap_kaddr = kmap_local_page(req->pr_bitmap_bh->b_page);
 	bitmap = bitmap_kaddr + bh_offset(req->pr_bitmap_bh);
 	lock = nilfs_mdt_bgl_lock(inode, group);
 
@@ -621,8 +633,8 @@ void nilfs_palloc_commit_free_entry(struct inode *inode,
 	else
 		nilfs_palloc_group_desc_add_entries(desc, lock, 1);
 
-	kunmap(req->pr_bitmap_bh->b_page);
-	kunmap(req->pr_desc_bh->b_page);
+	kunmap_local(bitmap_kaddr);
+	kunmap_local(desc_kaddr);
 
 	mark_buffer_dirty(req->pr_desc_bh);
 	mark_buffer_dirty(req->pr_bitmap_bh);
@@ -647,10 +659,10 @@ void nilfs_palloc_abort_alloc_entry(struct inode *inode,
 	spinlock_t *lock;
 
 	group = nilfs_palloc_group(inode, req->pr_entry_nr, &group_offset);
-	desc_kaddr = kmap(req->pr_desc_bh->b_page);
+	desc_kaddr = kmap_local_page(req->pr_desc_bh->b_page);
 	desc = nilfs_palloc_block_get_group_desc(inode, group,
 						 req->pr_desc_bh, desc_kaddr);
-	bitmap_kaddr = kmap(req->pr_bitmap_bh->b_page);
+	bitmap_kaddr = kmap_local_page(req->pr_bitmap_bh->b_page);
 	bitmap = bitmap_kaddr + bh_offset(req->pr_bitmap_bh);
 	lock = nilfs_mdt_bgl_lock(inode, group);
 
@@ -662,8 +674,8 @@ void nilfs_palloc_abort_alloc_entry(struct inode *inode,
 	else
 		nilfs_palloc_group_desc_add_entries(desc, lock, 1);
 
-	kunmap(req->pr_bitmap_bh->b_page);
-	kunmap(req->pr_desc_bh->b_page);
+	kunmap_local(bitmap_kaddr);
+	kunmap_local(desc_kaddr);
 
 	brelse(req->pr_bitmap_bh);
 	brelse(req->pr_desc_bh);
@@ -755,7 +767,7 @@ int nilfs_palloc_freev(struct inode *inode, __u64 *entry_nrs, size_t nitems)
 		/* Get the first entry number of the group */
 		group_min_nr = (__u64)group * epg;
 
-		bitmap_kaddr = kmap(bitmap_bh->b_page);
+		bitmap_kaddr = kmap_local_page(bitmap_bh->b_page);
 		bitmap = bitmap_kaddr + bh_offset(bitmap_bh);
 		lock = nilfs_mdt_bgl_lock(inode, group);
 
@@ -801,7 +813,7 @@ int nilfs_palloc_freev(struct inode *inode, __u64 *entry_nrs, size_t nitems)
 			entry_start = rounddown(group_offset, epb);
 		} while (true);
 
-		kunmap(bitmap_bh->b_page);
+		kunmap_local(bitmap_kaddr);
 		mark_buffer_dirty(bitmap_bh);
 		brelse(bitmap_bh);
 
@@ -815,11 +827,11 @@ int nilfs_palloc_freev(struct inode *inode, __u64 *entry_nrs, size_t nitems)
 					   inode->i_ino);
 		}
 
-		desc_kaddr = kmap_atomic(desc_bh->b_page);
+		desc_kaddr = kmap_local_page(desc_bh->b_page);
 		desc = nilfs_palloc_block_get_group_desc(
 			inode, group, desc_bh, desc_kaddr);
 		nfree = nilfs_palloc_group_desc_add_entries(desc, lock, n);
-		kunmap_atomic(desc_kaddr);
+		kunmap_local(desc_kaddr);
 		mark_buffer_dirty(desc_bh);
 		nilfs_mdt_mark_dirty(inode);
 		brelse(desc_bh);
diff --git a/fs/nilfs2/alloc.h b/fs/nilfs2/alloc.h
index b667e869ac..d825a9faca 100644
--- a/fs/nilfs2/alloc.h
+++ b/fs/nilfs2/alloc.h
@@ -50,8 +50,8 @@ struct nilfs_palloc_req {
 	struct buffer_head *pr_entry_bh;
 };
 
-int nilfs_palloc_prepare_alloc_entry(struct inode *,
-				     struct nilfs_palloc_req *);
+int nilfs_palloc_prepare_alloc_entry(struct inode *inode,
+				     struct nilfs_palloc_req *req, bool wrap);
 void nilfs_palloc_commit_alloc_entry(struct inode *,
 				     struct nilfs_palloc_req *);
 void nilfs_palloc_abort_alloc_entry(struct inode *, struct nilfs_palloc_req *);
diff --git a/fs/nilfs2/bmap.c b/fs/nilfs2/bmap.c
index 7a8f166f2c..383f0afa2c 100644
--- a/fs/nilfs2/bmap.c
+++ b/fs/nilfs2/bmap.c
@@ -548,13 +548,10 @@ int nilfs_bmap_read(struct nilfs_bmap *bmap, struct nilfs_inode *raw_inode)
  */
 void nilfs_bmap_write(struct nilfs_bmap *bmap, struct nilfs_inode *raw_inode)
 {
-	down_write(&bmap->b_sem);
 	memcpy(raw_inode->i_bmap, bmap->b_u.u_data,
 	       NILFS_INODE_BMAP_SIZE * sizeof(__le64));
 	if (bmap->b_inode->i_ino == NILFS_DAT_INO)
 		bmap->b_last_allocated_ptr = NILFS_BMAP_NEW_PTR_INIT;
-
-	up_write(&bmap->b_sem);
 }
 
 void nilfs_bmap_init_gc(struct nilfs_bmap *bmap)
diff --git a/fs/nilfs2/btnode.c b/fs/nilfs2/btnode.c
index 0131d83b91..c034080c33 100644
--- a/fs/nilfs2/btnode.c
+++ b/fs/nilfs2/btnode.c
@@ -51,12 +51,21 @@ nilfs_btnode_create_block(struct address_space *btnc, __u64 blocknr)
 
 	bh = nilfs_grab_buffer(inode, btnc, blocknr, BIT(BH_NILFS_Node));
 	if (unlikely(!bh))
-		return NULL;
+		return ERR_PTR(-ENOMEM);
 
 	if (unlikely(buffer_mapped(bh) || buffer_uptodate(bh) ||
 		     buffer_dirty(bh))) {
-		brelse(bh);
-		BUG();
+		/*
+		 * The block buffer at the specified new address was already
+		 * in use.  This can happen if it is a virtual block number
+		 * and has been reallocated due to corruption of the bitmap
+		 * used to manage its allocation state (if not, the buffer
+		 * clearing of an abandoned b-tree node is missing somewhere).
+		 */
+		nilfs_error(inode->i_sb,
+			    "state inconsistency probably due to duplicate use of b-tree node block address %llu (ino=%lu)",
+			    (unsigned long long)blocknr, inode->i_ino);
+		goto failed;
 	}
 	memset(bh->b_data, 0, i_blocksize(inode));
 	bh->b_bdev = inode->i_sb->s_bdev;
@@ -67,6 +76,12 @@ nilfs_btnode_create_block(struct address_space *btnc, __u64 blocknr)
 	folio_unlock(bh->b_folio);
 	folio_put(bh->b_folio);
 	return bh;
+
+failed:
+	folio_unlock(bh->b_folio);
+	folio_put(bh->b_folio);
+	brelse(bh);
+	return ERR_PTR(-EIO);
 }
 
 int nilfs_btnode_submit_block(struct address_space *btnc, __u64 blocknr,
@@ -217,8 +232,8 @@ retry:
 	}
 
 	nbh = nilfs_btnode_create_block(btnc, newkey);
-	if (!nbh)
-		return -ENOMEM;
+	if (IS_ERR(nbh))
+		return PTR_ERR(nbh);
 
 	BUG_ON(nbh == obh);
 	ctxt->newbh = nbh;
diff --git a/fs/nilfs2/btree.c b/fs/nilfs2/btree.c
index 65659fa037..862bdf2312 100644
--- a/fs/nilfs2/btree.c
+++ b/fs/nilfs2/btree.c
@@ -63,8 +63,8 @@ static int nilfs_btree_get_new_block(const struct nilfs_bmap *btree,
 	struct buffer_head *bh;
 
 	bh = nilfs_btnode_create_block(btnc, ptr);
-	if (!bh)
-		return -ENOMEM;
+	if (IS_ERR(bh))
+		return PTR_ERR(bh);
 
 	set_buffer_nilfs_volatile(bh);
 	*bhp = bh;
@@ -1857,13 +1857,22 @@ nilfs_btree_commit_convert_and_insert(struct nilfs_bmap *btree,
 }
 
 /**
- * nilfs_btree_convert_and_insert -
- * @bmap:
- * @key:
- * @ptr:
- * @keys:
- * @ptrs:
- * @n:
+ * nilfs_btree_convert_and_insert - Convert and insert entries into a B-tree
+ * @btree: NILFS B-tree structure
+ * @key: Key of the new entry to be inserted
+ * @ptr: Pointer (block number) associated with the key to be inserted
+ * @keys: Array of keys to be inserted in addition to @key
+ * @ptrs: Array of pointers associated with @keys
+ * @n: Number of keys and pointers in @keys and @ptrs
+ *
+ * This function is used to insert a new entry specified by @key and @ptr,
+ * along with additional entries specified by @keys and @ptrs arrays, into a
+ * NILFS B-tree.
+ * It prepares the necessary changes by allocating the required blocks and any
+ * necessary intermediate nodes. It converts configurations from other forms of
+ * block mapping (the one that currently exists is direct mapping) to a B-tree.
+ *
+ * Return: 0 on success or a negative error code on failure.
  */
 int nilfs_btree_convert_and_insert(struct nilfs_bmap *btree,
 				   __u64 key, __u64 ptr,
diff --git a/fs/nilfs2/cpfile.c b/fs/nilfs2/cpfile.c
index 39136637f7..69a5cced1e 100644
--- a/fs/nilfs2/cpfile.c
+++ b/fs/nilfs2/cpfile.c
@@ -28,7 +28,7 @@ nilfs_cpfile_get_blkoff(const struct inode *cpfile, __u64 cno)
 {
 	__u64 tcno = cno + NILFS_MDT(cpfile)->mi_first_entry_offset - 1;
 
-	do_div(tcno, nilfs_cpfile_checkpoints_per_block(cpfile));
+	tcno = div64_ul(tcno, nilfs_cpfile_checkpoints_per_block(cpfile));
 	return (unsigned long)tcno;
 }
 
@@ -187,35 +187,90 @@ static inline int nilfs_cpfile_delete_checkpoint_block(struct inode *cpfile,
 }
 
 /**
- * nilfs_cpfile_get_checkpoint - get a checkpoint
- * @cpfile: inode of checkpoint file
- * @cno: checkpoint number
- * @create: create flag
- * @cpp: pointer to a checkpoint
- * @bhp: pointer to a buffer head
- *
- * Description: nilfs_cpfile_get_checkpoint() acquires the checkpoint
- * specified by @cno. A new checkpoint will be created if @cno is the current
- * checkpoint number and @create is nonzero.
- *
- * Return Value: On success, 0 is returned, and the checkpoint and the
- * buffer head of the buffer on which the checkpoint is located are stored in
- * the place pointed by @cpp and @bhp, respectively. On error, one of the
- * following negative error codes is returned.
+ * nilfs_cpfile_read_checkpoint - read a checkpoint entry in cpfile
+ * @cpfile: checkpoint file inode
+ * @cno:    number of checkpoint entry to read
+ * @root:   nilfs root object
+ * @ifile:  ifile's inode to read and attach to @root
  *
- * %-EIO - I/O error.
+ * This function imports checkpoint information from the checkpoint file and
+ * stores it to the inode file given by @ifile and the nilfs root object
+ * given by @root.
  *
- * %-ENOMEM - Insufficient amount of memory available.
+ * Return: 0 on success, or the following negative error code on failure.
+ * * %-EINVAL	- Invalid checkpoint.
+ * * %-ENOMEM	- Insufficient memory available.
+ * * %-EIO	- I/O error (including metadata corruption).
+ */
+int nilfs_cpfile_read_checkpoint(struct inode *cpfile, __u64 cno,
+				 struct nilfs_root *root, struct inode *ifile)
+{
+	struct buffer_head *cp_bh;
+	struct nilfs_checkpoint *cp;
+	void *kaddr;
+	int ret;
+
+	if (cno < 1 || cno > nilfs_mdt_cno(cpfile))
+		return -EINVAL;
+
+	down_read(&NILFS_MDT(cpfile)->mi_sem);
+	ret = nilfs_cpfile_get_checkpoint_block(cpfile, cno, 0, &cp_bh);
+	if (unlikely(ret < 0)) {
+		if (ret == -ENOENT)
+			ret = -EINVAL;
+		goto out_sem;
+	}
+
+	kaddr = kmap_local_page(cp_bh->b_page);
+	cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, cp_bh, kaddr);
+	if (nilfs_checkpoint_invalid(cp)) {
+		ret = -EINVAL;
+		goto put_cp;
+	}
+
+	ret = nilfs_read_inode_common(ifile, &cp->cp_ifile_inode);
+	if (unlikely(ret)) {
+		/*
+		 * Since this inode is on a checkpoint entry, treat errors
+		 * as metadata corruption.
+		 */
+		nilfs_err(cpfile->i_sb,
+			  "ifile inode (checkpoint number=%llu) corrupted",
+			  (unsigned long long)cno);
+		ret = -EIO;
+		goto put_cp;
+	}
+
+	/* Configure the nilfs root object */
+	atomic64_set(&root->inodes_count, le64_to_cpu(cp->cp_inodes_count));
+	atomic64_set(&root->blocks_count, le64_to_cpu(cp->cp_blocks_count));
+	root->ifile = ifile;
+
+put_cp:
+	kunmap_local(kaddr);
+	brelse(cp_bh);
+out_sem:
+	up_read(&NILFS_MDT(cpfile)->mi_sem);
+	return ret;
+}
+
+/**
+ * nilfs_cpfile_create_checkpoint - create a checkpoint entry on cpfile
+ * @cpfile: checkpoint file inode
+ * @cno:    number of checkpoint to set up
  *
- * %-ENOENT - No such checkpoint.
+ * This function creates a checkpoint with the number specified by @cno on
+ * cpfile.  If the specified checkpoint entry already exists due to a past
+ * failure, it will be reused without returning an error.
+ * In either case, the buffer of the block containing the checkpoint entry
+ * and the cpfile inode are made dirty for inclusion in the write log.
  *
- * %-EINVAL - invalid checkpoint.
+ * Return: 0 on success, or the following negative error code on failure.
+ * * %-ENOMEM	- Insufficient memory available.
+ * * %-EIO	- I/O error (including metadata corruption).
+ * * %-EROFS	- Read only filesystem
  */
-int nilfs_cpfile_get_checkpoint(struct inode *cpfile,
-				__u64 cno,
-				int create,
-				struct nilfs_checkpoint **cpp,
-				struct buffer_head **bhp)
+int nilfs_cpfile_create_checkpoint(struct inode *cpfile, __u64 cno)
 {
 	struct buffer_head *header_bh, *cp_bh;
 	struct nilfs_cpfile_header *header;
@@ -223,70 +278,128 @@ int nilfs_cpfile_get_checkpoint(struct inode *cpfile,
 	void *kaddr;
 	int ret;
 
-	if (unlikely(cno < 1 || cno > nilfs_mdt_cno(cpfile) ||
-		     (cno < nilfs_mdt_cno(cpfile) && create)))
-		return -EINVAL;
+	if (WARN_ON_ONCE(cno < 1))
+		return -EIO;
 
 	down_write(&NILFS_MDT(cpfile)->mi_sem);
-
 	ret = nilfs_cpfile_get_header_block(cpfile, &header_bh);
-	if (ret < 0)
+	if (unlikely(ret < 0)) {
+		if (ret == -ENOENT) {
+			nilfs_error(cpfile->i_sb,
+				    "checkpoint creation failed due to metadata corruption.");
+			ret = -EIO;
+		}
 		goto out_sem;
-	ret = nilfs_cpfile_get_checkpoint_block(cpfile, cno, create, &cp_bh);
-	if (ret < 0)
+	}
+	ret = nilfs_cpfile_get_checkpoint_block(cpfile, cno, 1, &cp_bh);
+	if (unlikely(ret < 0))
 		goto out_header;
-	kaddr = kmap(cp_bh->b_page);
+
+	kaddr = kmap_local_page(cp_bh->b_page);
 	cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, cp_bh, kaddr);
 	if (nilfs_checkpoint_invalid(cp)) {
-		if (!create) {
-			kunmap(cp_bh->b_page);
-			brelse(cp_bh);
-			ret = -ENOENT;
-			goto out_header;
-		}
 		/* a newly-created checkpoint */
 		nilfs_checkpoint_clear_invalid(cp);
 		if (!nilfs_cpfile_is_in_first(cpfile, cno))
 			nilfs_cpfile_block_add_valid_checkpoints(cpfile, cp_bh,
 								 kaddr, 1);
-		mark_buffer_dirty(cp_bh);
+		kunmap_local(kaddr);
 
-		kaddr = kmap_atomic(header_bh->b_page);
+		kaddr = kmap_local_page(header_bh->b_page);
 		header = nilfs_cpfile_block_get_header(cpfile, header_bh,
 						       kaddr);
 		le64_add_cpu(&header->ch_ncheckpoints, 1);
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		mark_buffer_dirty(header_bh);
-		nilfs_mdt_mark_dirty(cpfile);
+	} else {
+		kunmap_local(kaddr);
 	}
 
-	if (cpp != NULL)
-		*cpp = cp;
-	*bhp = cp_bh;
+	/* Force the buffer and the inode to become dirty */
+	mark_buffer_dirty(cp_bh);
+	brelse(cp_bh);
+	nilfs_mdt_mark_dirty(cpfile);
 
- out_header:
+out_header:
 	brelse(header_bh);
 
- out_sem:
+out_sem:
 	up_write(&NILFS_MDT(cpfile)->mi_sem);
 	return ret;
 }
 
 /**
- * nilfs_cpfile_put_checkpoint - put a checkpoint
- * @cpfile: inode of checkpoint file
- * @cno: checkpoint number
- * @bh: buffer head
+ * nilfs_cpfile_finalize_checkpoint - fill in a checkpoint entry in cpfile
+ * @cpfile: checkpoint file inode
+ * @cno:    checkpoint number
+ * @root:   nilfs root object
+ * @blkinc: number of blocks added by this checkpoint
+ * @ctime:  checkpoint creation time
+ * @minor:  minor checkpoint flag
+ *
+ * This function completes the checkpoint entry numbered by @cno in the
+ * cpfile with the data given by the arguments @root, @blkinc, @ctime, and
+ * @minor.
  *
- * Description: nilfs_cpfile_put_checkpoint() releases the checkpoint
- * specified by @cno. @bh must be the buffer head which has been returned by
- * a previous call to nilfs_cpfile_get_checkpoint() with @cno.
+ * Return: 0 on success, or the following negative error code on failure.
+ * * %-ENOMEM	- Insufficient memory available.
+ * * %-EIO	- I/O error (including metadata corruption).
  */
-void nilfs_cpfile_put_checkpoint(struct inode *cpfile, __u64 cno,
-				 struct buffer_head *bh)
+int nilfs_cpfile_finalize_checkpoint(struct inode *cpfile, __u64 cno,
+				     struct nilfs_root *root, __u64 blkinc,
+				     time64_t ctime, bool minor)
 {
-	kunmap(bh->b_page);
-	brelse(bh);
+	struct buffer_head *cp_bh;
+	struct nilfs_checkpoint *cp;
+	void *kaddr;
+	int ret;
+
+	if (WARN_ON_ONCE(cno < 1))
+		return -EIO;
+
+	down_write(&NILFS_MDT(cpfile)->mi_sem);
+	ret = nilfs_cpfile_get_checkpoint_block(cpfile, cno, 0, &cp_bh);
+	if (unlikely(ret < 0)) {
+		if (ret == -ENOENT)
+			goto error;
+		goto out_sem;
+	}
+
+	kaddr = kmap_local_page(cp_bh->b_page);
+	cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, cp_bh, kaddr);
+	if (unlikely(nilfs_checkpoint_invalid(cp))) {
+		kunmap_local(kaddr);
+		brelse(cp_bh);
+		goto error;
+	}
+
+	cp->cp_snapshot_list.ssl_next = 0;
+	cp->cp_snapshot_list.ssl_prev = 0;
+	cp->cp_inodes_count = cpu_to_le64(atomic64_read(&root->inodes_count));
+	cp->cp_blocks_count = cpu_to_le64(atomic64_read(&root->blocks_count));
+	cp->cp_nblk_inc = cpu_to_le64(blkinc);
+	cp->cp_create = cpu_to_le64(ctime);
+	cp->cp_cno = cpu_to_le64(cno);
+
+	if (minor)
+		nilfs_checkpoint_set_minor(cp);
+	else
+		nilfs_checkpoint_clear_minor(cp);
+
+	nilfs_write_inode_common(root->ifile, &cp->cp_ifile_inode);
+	nilfs_bmap_write(NILFS_I(root->ifile)->i_bmap, &cp->cp_ifile_inode);
+
+	kunmap_local(kaddr);
+	brelse(cp_bh);
+out_sem:
+	up_write(&NILFS_MDT(cpfile)->mi_sem);
+	return ret;
+
+error:
+	nilfs_error(cpfile->i_sb,
+		    "checkpoint finalization failed due to metadata corruption.");
+	ret = -EIO;
+	goto out_sem;
 }
 
 /**
@@ -347,7 +460,7 @@ int nilfs_cpfile_delete_checkpoints(struct inode *cpfile,
 			continue;
 		}
 
-		kaddr = kmap_atomic(cp_bh->b_page);
+		kaddr = kmap_local_page(cp_bh->b_page);
 		cp = nilfs_cpfile_block_get_checkpoint(
 			cpfile, cno, cp_bh, kaddr);
 		nicps = 0;
@@ -369,7 +482,7 @@ int nilfs_cpfile_delete_checkpoints(struct inode *cpfile,
 						cpfile, cp_bh, kaddr, nicps);
 				if (count == 0) {
 					/* make hole */
-					kunmap_atomic(kaddr);
+					kunmap_local(kaddr);
 					brelse(cp_bh);
 					ret =
 					  nilfs_cpfile_delete_checkpoint_block(
@@ -384,18 +497,18 @@ int nilfs_cpfile_delete_checkpoints(struct inode *cpfile,
 			}
 		}
 
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		brelse(cp_bh);
 	}
 
 	if (tnicps > 0) {
-		kaddr = kmap_atomic(header_bh->b_page);
+		kaddr = kmap_local_page(header_bh->b_page);
 		header = nilfs_cpfile_block_get_header(cpfile, header_bh,
 						       kaddr);
 		le64_add_cpu(&header->ch_ncheckpoints, -(u64)tnicps);
 		mark_buffer_dirty(header_bh);
 		nilfs_mdt_mark_dirty(cpfile);
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 	}
 
 	brelse(header_bh);
@@ -447,7 +560,7 @@ static ssize_t nilfs_cpfile_do_get_cpinfo(struct inode *cpfile, __u64 *cnop,
 		}
 		ncps = nilfs_cpfile_checkpoints_in_block(cpfile, cno, cur_cno);
 
-		kaddr = kmap_atomic(bh->b_page);
+		kaddr = kmap_local_page(bh->b_page);
 		cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, bh, kaddr);
 		for (i = 0; i < ncps && n < nci; i++, cp = (void *)cp + cpsz) {
 			if (!nilfs_checkpoint_invalid(cp)) {
@@ -457,7 +570,7 @@ static ssize_t nilfs_cpfile_do_get_cpinfo(struct inode *cpfile, __u64 *cnop,
 				n++;
 			}
 		}
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		brelse(bh);
 	}
 
@@ -491,10 +604,10 @@ static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 *cnop,
 		ret = nilfs_cpfile_get_header_block(cpfile, &bh);
 		if (ret < 0)
 			goto out;
-		kaddr = kmap_atomic(bh->b_page);
+		kaddr = kmap_local_page(bh->b_page);
 		header = nilfs_cpfile_block_get_header(cpfile, bh, kaddr);
 		curr = le64_to_cpu(header->ch_snapshot_list.ssl_next);
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		brelse(bh);
 		if (curr == 0) {
 			ret = 0;
@@ -512,7 +625,7 @@ static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 *cnop,
 			ret = 0; /* No snapshots (started from a hole block) */
 		goto out;
 	}
-	kaddr = kmap_atomic(bh->b_page);
+	kaddr = kmap_local_page(bh->b_page);
 	while (n < nci) {
 		cp = nilfs_cpfile_block_get_checkpoint(cpfile, curr, bh, kaddr);
 		curr = ~(__u64)0; /* Terminator */
@@ -528,7 +641,7 @@ static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 *cnop,
 
 		next_blkoff = nilfs_cpfile_get_blkoff(cpfile, next);
 		if (curr_blkoff != next_blkoff) {
-			kunmap_atomic(kaddr);
+			kunmap_local(kaddr);
 			brelse(bh);
 			ret = nilfs_cpfile_get_checkpoint_block(cpfile, next,
 								0, &bh);
@@ -536,12 +649,12 @@ static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 *cnop,
 				WARN_ON(ret == -ENOENT);
 				goto out;
 			}
-			kaddr = kmap_atomic(bh->b_page);
+			kaddr = kmap_local_page(bh->b_page);
 		}
 		curr = next;
 		curr_blkoff = next_blkoff;
 	}
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 	brelse(bh);
 	*cnop = curr;
 	ret = n;
@@ -650,24 +763,24 @@ static int nilfs_cpfile_set_snapshot(struct inode *cpfile, __u64 cno)
 	ret = nilfs_cpfile_get_checkpoint_block(cpfile, cno, 0, &cp_bh);
 	if (ret < 0)
 		goto out_sem;
-	kaddr = kmap_atomic(cp_bh->b_page);
+	kaddr = kmap_local_page(cp_bh->b_page);
 	cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, cp_bh, kaddr);
 	if (nilfs_checkpoint_invalid(cp)) {
 		ret = -ENOENT;
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		goto out_cp;
 	}
 	if (nilfs_checkpoint_snapshot(cp)) {
 		ret = 0;
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		goto out_cp;
 	}
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	ret = nilfs_cpfile_get_header_block(cpfile, &header_bh);
 	if (ret < 0)
 		goto out_cp;
-	kaddr = kmap_atomic(header_bh->b_page);
+	kaddr = kmap_local_page(header_bh->b_page);
 	header = nilfs_cpfile_block_get_header(cpfile, header_bh, kaddr);
 	list = &header->ch_snapshot_list;
 	curr_bh = header_bh;
@@ -679,13 +792,13 @@ static int nilfs_cpfile_set_snapshot(struct inode *cpfile, __u64 cno)
 		prev_blkoff = nilfs_cpfile_get_blkoff(cpfile, prev);
 		curr = prev;
 		if (curr_blkoff != prev_blkoff) {
-			kunmap_atomic(kaddr);
+			kunmap_local(kaddr);
 			brelse(curr_bh);
 			ret = nilfs_cpfile_get_checkpoint_block(cpfile, curr,
 								0, &curr_bh);
 			if (ret < 0)
 				goto out_header;
-			kaddr = kmap_atomic(curr_bh->b_page);
+			kaddr = kmap_local_page(curr_bh->b_page);
 		}
 		curr_blkoff = prev_blkoff;
 		cp = nilfs_cpfile_block_get_checkpoint(
@@ -693,7 +806,7 @@ static int nilfs_cpfile_set_snapshot(struct inode *cpfile, __u64 cno)
 		list = &cp->cp_snapshot_list;
 		prev = le64_to_cpu(list->ssl_prev);
 	}
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	if (prev != 0) {
 		ret = nilfs_cpfile_get_checkpoint_block(cpfile, prev, 0,
@@ -705,29 +818,29 @@ static int nilfs_cpfile_set_snapshot(struct inode *cpfile, __u64 cno)
 		get_bh(prev_bh);
 	}
 
-	kaddr = kmap_atomic(curr_bh->b_page);
+	kaddr = kmap_local_page(curr_bh->b_page);
 	list = nilfs_cpfile_block_get_snapshot_list(
 		cpfile, curr, curr_bh, kaddr);
 	list->ssl_prev = cpu_to_le64(cno);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
-	kaddr = kmap_atomic(cp_bh->b_page);
+	kaddr = kmap_local_page(cp_bh->b_page);
 	cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, cp_bh, kaddr);
 	cp->cp_snapshot_list.ssl_next = cpu_to_le64(curr);
 	cp->cp_snapshot_list.ssl_prev = cpu_to_le64(prev);
 	nilfs_checkpoint_set_snapshot(cp);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
-	kaddr = kmap_atomic(prev_bh->b_page);
+	kaddr = kmap_local_page(prev_bh->b_page);
 	list = nilfs_cpfile_block_get_snapshot_list(
 		cpfile, prev, prev_bh, kaddr);
 	list->ssl_next = cpu_to_le64(cno);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
-	kaddr = kmap_atomic(header_bh->b_page);
+	kaddr = kmap_local_page(header_bh->b_page);
 	header = nilfs_cpfile_block_get_header(cpfile, header_bh, kaddr);
 	le64_add_cpu(&header->ch_nsnapshots, 1);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	mark_buffer_dirty(prev_bh);
 	mark_buffer_dirty(curr_bh);
@@ -768,23 +881,23 @@ static int nilfs_cpfile_clear_snapshot(struct inode *cpfile, __u64 cno)
 	ret = nilfs_cpfile_get_checkpoint_block(cpfile, cno, 0, &cp_bh);
 	if (ret < 0)
 		goto out_sem;
-	kaddr = kmap_atomic(cp_bh->b_page);
+	kaddr = kmap_local_page(cp_bh->b_page);
 	cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, cp_bh, kaddr);
 	if (nilfs_checkpoint_invalid(cp)) {
 		ret = -ENOENT;
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		goto out_cp;
 	}
 	if (!nilfs_checkpoint_snapshot(cp)) {
 		ret = 0;
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		goto out_cp;
 	}
 
 	list = &cp->cp_snapshot_list;
 	next = le64_to_cpu(list->ssl_next);
 	prev = le64_to_cpu(list->ssl_prev);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	ret = nilfs_cpfile_get_header_block(cpfile, &header_bh);
 	if (ret < 0)
@@ -808,29 +921,29 @@ static int nilfs_cpfile_clear_snapshot(struct inode *cpfile, __u64 cno)
 		get_bh(prev_bh);
 	}
 
-	kaddr = kmap_atomic(next_bh->b_page);
+	kaddr = kmap_local_page(next_bh->b_page);
 	list = nilfs_cpfile_block_get_snapshot_list(
 		cpfile, next, next_bh, kaddr);
 	list->ssl_prev = cpu_to_le64(prev);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
-	kaddr = kmap_atomic(prev_bh->b_page);
+	kaddr = kmap_local_page(prev_bh->b_page);
 	list = nilfs_cpfile_block_get_snapshot_list(
 		cpfile, prev, prev_bh, kaddr);
 	list->ssl_next = cpu_to_le64(next);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
-	kaddr = kmap_atomic(cp_bh->b_page);
+	kaddr = kmap_local_page(cp_bh->b_page);
 	cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, cp_bh, kaddr);
 	cp->cp_snapshot_list.ssl_next = cpu_to_le64(0);
 	cp->cp_snapshot_list.ssl_prev = cpu_to_le64(0);
 	nilfs_checkpoint_clear_snapshot(cp);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
-	kaddr = kmap_atomic(header_bh->b_page);
+	kaddr = kmap_local_page(header_bh->b_page);
 	header = nilfs_cpfile_block_get_header(cpfile, header_bh, kaddr);
 	le64_add_cpu(&header->ch_nsnapshots, -1);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	mark_buffer_dirty(next_bh);
 	mark_buffer_dirty(prev_bh);
@@ -889,13 +1002,13 @@ int nilfs_cpfile_is_snapshot(struct inode *cpfile, __u64 cno)
 	ret = nilfs_cpfile_get_checkpoint_block(cpfile, cno, 0, &bh);
 	if (ret < 0)
 		goto out;
-	kaddr = kmap_atomic(bh->b_page);
+	kaddr = kmap_local_page(bh->b_page);
 	cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, bh, kaddr);
 	if (nilfs_checkpoint_invalid(cp))
 		ret = -ENOENT;
 	else
 		ret = nilfs_checkpoint_snapshot(cp);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 	brelse(bh);
 
  out:
@@ -972,12 +1085,12 @@ int nilfs_cpfile_get_stat(struct inode *cpfile, struct nilfs_cpstat *cpstat)
 	ret = nilfs_cpfile_get_header_block(cpfile, &bh);
 	if (ret < 0)
 		goto out_sem;
-	kaddr = kmap_atomic(bh->b_page);
+	kaddr = kmap_local_page(bh->b_page);
 	header = nilfs_cpfile_block_get_header(cpfile, bh, kaddr);
 	cpstat->cs_cno = nilfs_mdt_cno(cpfile);
 	cpstat->cs_ncps = le64_to_cpu(header->ch_ncheckpoints);
 	cpstat->cs_nsss = le64_to_cpu(header->ch_nsnapshots);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 	brelse(bh);
 
  out_sem:
diff --git a/fs/nilfs2/cpfile.h b/fs/nilfs2/cpfile.h
index edabb2dc57..f5b1d59289 100644
--- a/fs/nilfs2/cpfile.h
+++ b/fs/nilfs2/cpfile.h
@@ -16,10 +16,12 @@
 #include <linux/nilfs2_ondisk.h>	/* nilfs_inode, nilfs_checkpoint */
 
 
-int nilfs_cpfile_get_checkpoint(struct inode *, __u64, int,
-				struct nilfs_checkpoint **,
-				struct buffer_head **);
-void nilfs_cpfile_put_checkpoint(struct inode *, __u64, struct buffer_head *);
+int nilfs_cpfile_read_checkpoint(struct inode *cpfile, __u64 cno,
+				 struct nilfs_root *root, struct inode *ifile);
+int nilfs_cpfile_create_checkpoint(struct inode *cpfile, __u64 cno);
+int nilfs_cpfile_finalize_checkpoint(struct inode *cpfile, __u64 cno,
+				     struct nilfs_root *root, __u64 blkinc,
+				     time64_t ctime, bool minor);
 int nilfs_cpfile_delete_checkpoints(struct inode *, __u64, __u64);
 int nilfs_cpfile_delete_checkpoint(struct inode *, __u64);
 int nilfs_cpfile_change_cpmode(struct inode *, __u64, int);
diff --git a/fs/nilfs2/dat.c b/fs/nilfs2/dat.c
index 9cf6ba58f5..fc1caf63a4 100644
--- a/fs/nilfs2/dat.c
+++ b/fs/nilfs2/dat.c
@@ -75,7 +75,7 @@ int nilfs_dat_prepare_alloc(struct inode *dat, struct nilfs_palloc_req *req)
 {
 	int ret;
 
-	ret = nilfs_palloc_prepare_alloc_entry(dat, req);
+	ret = nilfs_palloc_prepare_alloc_entry(dat, req, true);
 	if (ret < 0)
 		return ret;
 
@@ -91,13 +91,13 @@ void nilfs_dat_commit_alloc(struct inode *dat, struct nilfs_palloc_req *req)
 	struct nilfs_dat_entry *entry;
 	void *kaddr;
 
-	kaddr = kmap_atomic(req->pr_entry_bh->b_page);
+	kaddr = kmap_local_page(req->pr_entry_bh->b_page);
 	entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr,
 					     req->pr_entry_bh, kaddr);
 	entry->de_start = cpu_to_le64(NILFS_CNO_MIN);
 	entry->de_end = cpu_to_le64(NILFS_CNO_MAX);
 	entry->de_blocknr = cpu_to_le64(0);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	nilfs_palloc_commit_alloc_entry(dat, req);
 	nilfs_dat_commit_entry(dat, req);
@@ -115,13 +115,13 @@ static void nilfs_dat_commit_free(struct inode *dat,
 	struct nilfs_dat_entry *entry;
 	void *kaddr;
 
-	kaddr = kmap_atomic(req->pr_entry_bh->b_page);
+	kaddr = kmap_local_page(req->pr_entry_bh->b_page);
 	entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr,
 					     req->pr_entry_bh, kaddr);
 	entry->de_start = cpu_to_le64(NILFS_CNO_MIN);
 	entry->de_end = cpu_to_le64(NILFS_CNO_MIN);
 	entry->de_blocknr = cpu_to_le64(0);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	nilfs_dat_commit_entry(dat, req);
 
@@ -145,12 +145,12 @@ void nilfs_dat_commit_start(struct inode *dat, struct nilfs_palloc_req *req,
 	struct nilfs_dat_entry *entry;
 	void *kaddr;
 
-	kaddr = kmap_atomic(req->pr_entry_bh->b_page);
+	kaddr = kmap_local_page(req->pr_entry_bh->b_page);
 	entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr,
 					     req->pr_entry_bh, kaddr);
 	entry->de_start = cpu_to_le64(nilfs_mdt_cno(dat));
 	entry->de_blocknr = cpu_to_le64(blocknr);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	nilfs_dat_commit_entry(dat, req);
 }
@@ -167,12 +167,12 @@ int nilfs_dat_prepare_end(struct inode *dat, struct nilfs_palloc_req *req)
 	if (ret < 0)
 		return ret;
 
-	kaddr = kmap_atomic(req->pr_entry_bh->b_page);
+	kaddr = kmap_local_page(req->pr_entry_bh->b_page);
 	entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr,
 					     req->pr_entry_bh, kaddr);
 	start = le64_to_cpu(entry->de_start);
 	blocknr = le64_to_cpu(entry->de_blocknr);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	if (blocknr == 0) {
 		ret = nilfs_palloc_prepare_free_entry(dat, req);
@@ -202,7 +202,7 @@ void nilfs_dat_commit_end(struct inode *dat, struct nilfs_palloc_req *req,
 	sector_t blocknr;
 	void *kaddr;
 
-	kaddr = kmap_atomic(req->pr_entry_bh->b_page);
+	kaddr = kmap_local_page(req->pr_entry_bh->b_page);
 	entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr,
 					     req->pr_entry_bh, kaddr);
 	end = start = le64_to_cpu(entry->de_start);
@@ -212,7 +212,7 @@ void nilfs_dat_commit_end(struct inode *dat, struct nilfs_palloc_req *req,
 	}
 	entry->de_end = cpu_to_le64(end);
 	blocknr = le64_to_cpu(entry->de_blocknr);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	if (blocknr == 0)
 		nilfs_dat_commit_free(dat, req);
@@ -227,12 +227,12 @@ void nilfs_dat_abort_end(struct inode *dat, struct nilfs_palloc_req *req)
 	sector_t blocknr;
 	void *kaddr;
 
-	kaddr = kmap_atomic(req->pr_entry_bh->b_page);
+	kaddr = kmap_local_page(req->pr_entry_bh->b_page);
 	entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr,
 					     req->pr_entry_bh, kaddr);
 	start = le64_to_cpu(entry->de_start);
 	blocknr = le64_to_cpu(entry->de_blocknr);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	if (start == nilfs_mdt_cno(dat) && blocknr == 0)
 		nilfs_palloc_abort_free_entry(dat, req);
@@ -362,7 +362,7 @@ int nilfs_dat_move(struct inode *dat, __u64 vblocknr, sector_t blocknr)
 		}
 	}
 
-	kaddr = kmap_atomic(entry_bh->b_page);
+	kaddr = kmap_local_page(entry_bh->b_page);
 	entry = nilfs_palloc_block_get_entry(dat, vblocknr, entry_bh, kaddr);
 	if (unlikely(entry->de_blocknr == cpu_to_le64(0))) {
 		nilfs_crit(dat->i_sb,
@@ -370,13 +370,13 @@ int nilfs_dat_move(struct inode *dat, __u64 vblocknr, sector_t blocknr)
 			   __func__, (unsigned long long)vblocknr,
 			   (unsigned long long)le64_to_cpu(entry->de_start),
 			   (unsigned long long)le64_to_cpu(entry->de_end));
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		brelse(entry_bh);
 		return -EINVAL;
 	}
 	WARN_ON(blocknr == 0);
 	entry->de_blocknr = cpu_to_le64(blocknr);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	mark_buffer_dirty(entry_bh);
 	nilfs_mdt_mark_dirty(dat);
@@ -426,7 +426,7 @@ int nilfs_dat_translate(struct inode *dat, __u64 vblocknr, sector_t *blocknrp)
 		}
 	}
 
-	kaddr = kmap_atomic(entry_bh->b_page);
+	kaddr = kmap_local_page(entry_bh->b_page);
 	entry = nilfs_palloc_block_get_entry(dat, vblocknr, entry_bh, kaddr);
 	blocknr = le64_to_cpu(entry->de_blocknr);
 	if (blocknr == 0) {
@@ -436,7 +436,7 @@ int nilfs_dat_translate(struct inode *dat, __u64 vblocknr, sector_t *blocknrp)
 	*blocknrp = blocknr;
 
  out:
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 	brelse(entry_bh);
 	return ret;
 }
@@ -457,10 +457,10 @@ ssize_t nilfs_dat_get_vinfo(struct inode *dat, void *buf, unsigned int visz,
 						   0, &entry_bh);
 		if (ret < 0)
 			return ret;
-		kaddr = kmap_atomic(entry_bh->b_page);
+		kaddr = kmap_local_page(entry_bh->b_page);
 		/* last virtual block number in this block */
 		first = vinfo->vi_vblocknr;
-		do_div(first, entries_per_block);
+		first = div64_ul(first, entries_per_block);
 		first *= entries_per_block;
 		last = first + entries_per_block - 1;
 		for (j = i, n = 0;
@@ -473,7 +473,7 @@ ssize_t nilfs_dat_get_vinfo(struct inode *dat, void *buf, unsigned int visz,
 			vinfo->vi_end = le64_to_cpu(entry->de_end);
 			vinfo->vi_blocknr = le64_to_cpu(entry->de_blocknr);
 		}
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		brelse(entry_bh);
 	}
 
diff --git a/fs/nilfs2/dir.c b/fs/nilfs2/dir.c
index aee40db7a0..4a29b0138d 100644
--- a/fs/nilfs2/dir.c
+++ b/fs/nilfs2/dir.c
@@ -135,6 +135,9 @@ static bool nilfs_check_folio(struct folio *folio, char *kaddr)
 			goto Enamelen;
 		if (((offs + rec_len - 1) ^ offs) & ~(chunk_size-1))
 			goto Espan;
+		if (unlikely(p->inode &&
+			     NILFS_PRIVATE_INODE(le64_to_cpu(p->inode))))
+			goto Einumber;
 	}
 	if (offs != limit)
 		goto Eend;
@@ -160,6 +163,9 @@ Enamelen:
 	goto bad_entry;
 Espan:
 	error = "directory entry across blocks";
+	goto bad_entry;
+Einumber:
+	error = "disallowed inode number";
 bad_entry:
 	nilfs_error(sb,
 		    "bad entry in directory #%lu: %s - offset=%lu, inode=%lu, rec_len=%zd, name_len=%d",
@@ -174,7 +180,6 @@ Eend:
 		    dir->i_ino, (folio->index << PAGE_SHIFT) + offs,
 		    (unsigned long)le64_to_cpu(p->inode));
 fail:
-	folio_set_error(folio);
 	return false;
 }
 
@@ -378,11 +383,39 @@ found:
 
 struct nilfs_dir_entry *nilfs_dotdot(struct inode *dir, struct folio **foliop)
 {
-	struct nilfs_dir_entry *de = nilfs_get_folio(dir, 0, foliop);
+	struct folio *folio;
+	struct nilfs_dir_entry *de, *next_de;
+	size_t limit;
+	char *msg;
 
+	de = nilfs_get_folio(dir, 0, &folio);
 	if (IS_ERR(de))
 		return NULL;
-	return nilfs_next_entry(de);
+
+	limit = nilfs_last_byte(dir, 0);  /* is a multiple of chunk size */
+	if (unlikely(!limit || le64_to_cpu(de->inode) != dir->i_ino ||
+		     !nilfs_match(1, ".", de))) {
+		msg = "missing '.'";
+		goto fail;
+	}
+
+	next_de = nilfs_next_entry(de);
+	/*
+	 * If "next_de" has not reached the end of the chunk, there is
+	 * at least one more record.  Check whether it matches "..".
+	 */
+	if (unlikely((char *)next_de == (char *)de + nilfs_chunk_size(dir) ||
+		     !nilfs_match(2, "..", next_de))) {
+		msg = "missing '..'";
+		goto fail;
+	}
+	*foliop = folio;
+	return next_de;
+
+fail:
+	nilfs_error(dir->i_sb, "directory #%lu %s", dir->i_ino, msg);
+	folio_release_kmap(folio, de);
+	return NULL;
 }
 
 ino_t nilfs_inode_by_name(struct inode *dir, const struct qstr *qstr)
@@ -608,7 +641,7 @@ int nilfs_empty_dir(struct inode *inode)
 
 		kaddr = nilfs_get_folio(inode, i, &folio);
 		if (IS_ERR(kaddr))
-			continue;
+			return 0;
 
 		de = (struct nilfs_dir_entry *)kaddr;
 		kaddr += nilfs_last_byte(inode, i) - NILFS_DIR_REC_LEN(1);
diff --git a/fs/nilfs2/gcinode.c b/fs/nilfs2/gcinode.c
index bf9a11d588..1c9ae36a03 100644
--- a/fs/nilfs2/gcinode.c
+++ b/fs/nilfs2/gcinode.c
@@ -175,6 +175,7 @@ int nilfs_init_gcinode(struct inode *inode)
 
 /**
  * nilfs_remove_all_gcinodes() - remove all unprocessed gc inodes
+ * @nilfs: NILFS filesystem instance
  */
 void nilfs_remove_all_gcinodes(struct the_nilfs *nilfs)
 {
diff --git a/fs/nilfs2/ifile.c b/fs/nilfs2/ifile.c
index a8a4bc8490..1e86b9303b 100644
--- a/fs/nilfs2/ifile.c
+++ b/fs/nilfs2/ifile.c
@@ -15,6 +15,7 @@
 #include "mdt.h"
 #include "alloc.h"
 #include "ifile.h"
+#include "cpfile.h"
 
 /**
  * struct nilfs_ifile_info - on-memory private data of ifile
@@ -55,13 +56,10 @@ int nilfs_ifile_create_inode(struct inode *ifile, ino_t *out_ino,
 	struct nilfs_palloc_req req;
 	int ret;
 
-	req.pr_entry_nr = 0;  /*
-			       * 0 says find free inode from beginning
-			       * of a group. dull code!!
-			       */
+	req.pr_entry_nr = NILFS_FIRST_INO(ifile->i_sb);
 	req.pr_entry_bh = NULL;
 
-	ret = nilfs_palloc_prepare_alloc_entry(ifile, &req);
+	ret = nilfs_palloc_prepare_alloc_entry(ifile, &req, false);
 	if (!ret) {
 		ret = nilfs_palloc_get_entry_block(ifile, req.pr_entry_nr, 1,
 						   &req.pr_entry_bh);
@@ -115,11 +113,11 @@ int nilfs_ifile_delete_inode(struct inode *ifile, ino_t ino)
 		return ret;
 	}
 
-	kaddr = kmap_atomic(req.pr_entry_bh->b_page);
+	kaddr = kmap_local_page(req.pr_entry_bh->b_page);
 	raw_inode = nilfs_palloc_block_get_entry(ifile, req.pr_entry_nr,
 						 req.pr_entry_bh, kaddr);
 	raw_inode->i_flags = 0;
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	mark_buffer_dirty(req.pr_entry_bh);
 	brelse(req.pr_entry_bh);
@@ -173,14 +171,18 @@ int nilfs_ifile_count_free_inodes(struct inode *ifile,
  * nilfs_ifile_read - read or get ifile inode
  * @sb: super block instance
  * @root: root object
+ * @cno: number of checkpoint entry to read
  * @inode_size: size of an inode
- * @raw_inode: on-disk ifile inode
- * @inodep: buffer to store the inode
+ *
+ * Return: 0 on success, or the following negative error code on failure.
+ * * %-EINVAL	- Invalid checkpoint.
+ * * %-ENOMEM	- Insufficient memory available.
+ * * %-EIO	- I/O error (including metadata corruption).
  */
 int nilfs_ifile_read(struct super_block *sb, struct nilfs_root *root,
-		     size_t inode_size, struct nilfs_inode *raw_inode,
-		     struct inode **inodep)
+		     __u64 cno, size_t inode_size)
 {
+	struct the_nilfs *nilfs;
 	struct inode *ifile;
 	int err;
 
@@ -201,13 +203,13 @@ int nilfs_ifile_read(struct super_block *sb, struct nilfs_root *root,
 
 	nilfs_palloc_setup_cache(ifile, &NILFS_IFILE_I(ifile)->palloc_cache);
 
-	err = nilfs_read_inode_common(ifile, raw_inode);
+	nilfs = sb->s_fs_info;
+	err = nilfs_cpfile_read_checkpoint(nilfs->ns_cpfile, cno, root, ifile);
 	if (err)
 		goto failed;
 
 	unlock_new_inode(ifile);
  out:
-	*inodep = ifile;
 	return 0;
  failed:
 	iget_failed(ifile);
diff --git a/fs/nilfs2/ifile.h b/fs/nilfs2/ifile.h
index 35c5273f48..625545cc2a 100644
--- a/fs/nilfs2/ifile.h
+++ b/fs/nilfs2/ifile.h
@@ -21,15 +21,14 @@
 static inline struct nilfs_inode *
 nilfs_ifile_map_inode(struct inode *ifile, ino_t ino, struct buffer_head *ibh)
 {
-	void *kaddr = kmap(ibh->b_page);
+	void *kaddr = kmap_local_page(ibh->b_page);
 
 	return nilfs_palloc_block_get_entry(ifile, ino, ibh, kaddr);
 }
 
-static inline void nilfs_ifile_unmap_inode(struct inode *ifile, ino_t ino,
-					   struct buffer_head *ibh)
+static inline void nilfs_ifile_unmap_inode(struct nilfs_inode *raw_inode)
 {
-	kunmap(ibh->b_page);
+	kunmap_local(raw_inode);
 }
 
 int nilfs_ifile_create_inode(struct inode *, ino_t *, struct buffer_head **);
@@ -39,7 +38,6 @@ int nilfs_ifile_get_inode_block(struct inode *, ino_t, struct buffer_head **);
 int nilfs_ifile_count_free_inodes(struct inode *, u64 *, u64 *);
 
 int nilfs_ifile_read(struct super_block *sb, struct nilfs_root *root,
-		     size_t inode_size, struct nilfs_inode *raw_inode,
-		     struct inode **inodep);
+		     __u64 cno, size_t inode_size);
 
 #endif	/* _NILFS_IFILE_H */
diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c
index 5a888b2c18..7340a01d80 100644
--- a/fs/nilfs2/inode.c
+++ b/fs/nilfs2/inode.c
@@ -520,7 +520,7 @@ static int __nilfs_read_inode(struct super_block *sb,
 			inode, inode->i_mode,
 			huge_decode_dev(le64_to_cpu(raw_inode->i_device_code)));
 	}
-	nilfs_ifile_unmap_inode(root->ifile, ino, bh);
+	nilfs_ifile_unmap_inode(raw_inode);
 	brelse(bh);
 	up_read(&NILFS_MDT(nilfs->ns_dat)->mi_sem);
 	nilfs_set_inode_flags(inode);
@@ -529,7 +529,7 @@ static int __nilfs_read_inode(struct super_block *sb,
 	return 0;
 
  failed_unmap:
-	nilfs_ifile_unmap_inode(root->ifile, ino, bh);
+	nilfs_ifile_unmap_inode(raw_inode);
 	brelse(bh);
 
  bad_inode:
@@ -759,8 +759,18 @@ struct inode *nilfs_iget_for_shadow(struct inode *inode)
 	return s_inode;
 }
 
+/**
+ * nilfs_write_inode_common - export common inode information to on-disk inode
+ * @inode:     inode object
+ * @raw_inode: on-disk inode
+ *
+ * This function writes standard information from the on-memory inode @inode
+ * to @raw_inode on ifile, cpfile or a super root block.  Since inode bmap
+ * data is not exported, nilfs_bmap_write() must be called separately during
+ * log writing.
+ */
 void nilfs_write_inode_common(struct inode *inode,
-			      struct nilfs_inode *raw_inode, int has_bmap)
+			      struct nilfs_inode *raw_inode)
 {
 	struct nilfs_inode_info *ii = NILFS_I(inode);
 
@@ -778,21 +788,6 @@ void nilfs_write_inode_common(struct inode *inode,
 	raw_inode->i_flags = cpu_to_le32(ii->i_flags);
 	raw_inode->i_generation = cpu_to_le32(inode->i_generation);
 
-	if (NILFS_ROOT_METADATA_FILE(inode->i_ino)) {
-		struct the_nilfs *nilfs = inode->i_sb->s_fs_info;
-
-		/* zero-fill unused portion in the case of super root block */
-		raw_inode->i_xattr = 0;
-		raw_inode->i_pad = 0;
-		memset((void *)raw_inode + sizeof(*raw_inode), 0,
-		       nilfs->ns_inode_size - sizeof(*raw_inode));
-	}
-
-	if (has_bmap)
-		nilfs_bmap_write(ii->i_bmap, raw_inode);
-	else if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode))
-		raw_inode->i_device_code =
-			cpu_to_le64(huge_encode_dev(inode->i_rdev));
 	/*
 	 * When extending inode, nilfs->ns_inode_size should be checked
 	 * for substitutions of appended fields.
@@ -813,14 +808,13 @@ void nilfs_update_inode(struct inode *inode, struct buffer_head *ibh, int flags)
 	if (flags & I_DIRTY_DATASYNC)
 		set_bit(NILFS_I_INODE_SYNC, &ii->i_state);
 
-	nilfs_write_inode_common(inode, raw_inode, 0);
-		/*
-		 * XXX: call with has_bmap = 0 is a workaround to avoid
-		 * deadlock of bmap.  This delays update of i_bmap to just
-		 * before writing.
-		 */
+	nilfs_write_inode_common(inode, raw_inode);
+
+	if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode))
+		raw_inode->i_device_code =
+			cpu_to_le64(huge_encode_dev(inode->i_rdev));
 
-	nilfs_ifile_unmap_inode(ifile, ino, ibh);
+	nilfs_ifile_unmap_inode(raw_inode);
 }
 
 #define NILFS_MAX_TRUNCATE_BLOCKS	16384  /* 64MB for 4KB block */
diff --git a/fs/nilfs2/ioctl.c b/fs/nilfs2/ioctl.c
index d6e3f20e88..8be471ce4f 100644
--- a/fs/nilfs2/ioctl.c
+++ b/fs/nilfs2/ioctl.c
@@ -1111,7 +1111,7 @@ static int nilfs_ioctl_set_alloc_range(struct inode *inode, void __user *argp)
 	segbytes = nilfs->ns_blocks_per_segment * nilfs->ns_blocksize;
 
 	minseg = range[0] + segbytes - 1;
-	do_div(minseg, segbytes);
+	minseg = div64_ul(minseg, segbytes);
 
 	if (range[1] < 4096)
 		goto out;
@@ -1120,7 +1120,7 @@ static int nilfs_ioctl_set_alloc_range(struct inode *inode, void __user *argp)
 	if (maxseg < segbytes)
 		goto out;
 
-	do_div(maxseg, segbytes);
+	maxseg = div64_ul(maxseg, segbytes);
 	maxseg--;
 
 	ret = nilfs_sufile_set_alloc_range(nilfs->ns_sufile, minseg, maxseg);
diff --git a/fs/nilfs2/mdt.c b/fs/nilfs2/mdt.c
index e45c01a559..4f792a0ad0 100644
--- a/fs/nilfs2/mdt.c
+++ b/fs/nilfs2/mdt.c
@@ -47,12 +47,12 @@ nilfs_mdt_insert_new_block(struct inode *inode, unsigned long block,
 
 	set_buffer_mapped(bh);
 
-	kaddr = kmap_atomic(bh->b_page);
+	kaddr = kmap_local_page(bh->b_page);
 	memset(kaddr + bh_offset(bh), 0, i_blocksize(inode));
 	if (init_block)
 		init_block(inode, bh, kaddr);
 	flush_dcache_page(bh->b_page);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	set_buffer_uptodate(bh);
 	mark_buffer_dirty(bh);
diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h
index 98cffaf0ac..4017f78564 100644
--- a/fs/nilfs2/nilfs.h
+++ b/fs/nilfs2/nilfs.h
@@ -116,9 +116,15 @@ enum {
 #define NILFS_FIRST_INO(sb) (((struct the_nilfs *)sb->s_fs_info)->ns_first_ino)
 
 #define NILFS_MDT_INODE(sb, ino) \
-	((ino) < NILFS_FIRST_INO(sb) && (NILFS_MDT_INO_BITS & BIT(ino)))
+	((ino) < NILFS_USER_INO && (NILFS_MDT_INO_BITS & BIT(ino)))
 #define NILFS_VALID_INODE(sb, ino) \
-	((ino) >= NILFS_FIRST_INO(sb) || (NILFS_SYS_INO_BITS & BIT(ino)))
+	((ino) >= NILFS_FIRST_INO(sb) ||				\
+	 ((ino) < NILFS_USER_INO && (NILFS_SYS_INO_BITS & BIT(ino))))
+
+#define NILFS_PRIVATE_INODE(ino) ({					\
+	ino_t __ino = (ino);						\
+	((__ino) < NILFS_USER_INO && (__ino) != NILFS_ROOT_INO &&	\
+	 (__ino) != NILFS_SKETCH_INO); })
 
 /**
  * struct nilfs_transaction_info: context information for synchronization
@@ -256,7 +262,8 @@ extern struct inode *nilfs_new_inode(struct inode *, umode_t);
 extern int nilfs_get_block(struct inode *, sector_t, struct buffer_head *, int);
 extern void nilfs_set_inode_flags(struct inode *);
 extern int nilfs_read_inode_common(struct inode *, struct nilfs_inode *);
-extern void nilfs_write_inode_common(struct inode *, struct nilfs_inode *, int);
+void nilfs_write_inode_common(struct inode *inode,
+			      struct nilfs_inode *raw_inode);
 struct inode *nilfs_ilookup(struct super_block *sb, struct nilfs_root *root,
 			    unsigned long ino);
 struct inode *nilfs_iget_locked(struct super_block *sb, struct nilfs_root *root,
@@ -334,8 +341,8 @@ void __nilfs_error(struct super_block *sb, const char *function,
 
 extern struct nilfs_super_block *
 nilfs_read_super_block(struct super_block *, u64, int, struct buffer_head **);
-extern int nilfs_store_magic_and_option(struct super_block *,
-					struct nilfs_super_block *, char *);
+extern int nilfs_store_magic(struct super_block *sb,
+			     struct nilfs_super_block *sbp);
 extern int nilfs_check_feature_compatibility(struct super_block *,
 					     struct nilfs_super_block *);
 extern void nilfs_set_log_cursor(struct nilfs_super_block *,
diff --git a/fs/nilfs2/page.c b/fs/nilfs2/page.c
index 5c2eba1987..14e470fb88 100644
--- a/fs/nilfs2/page.c
+++ b/fs/nilfs2/page.c
@@ -103,11 +103,11 @@ void nilfs_copy_buffer(struct buffer_head *dbh, struct buffer_head *sbh)
 	struct page *spage = sbh->b_page, *dpage = dbh->b_page;
 	struct buffer_head *bh;
 
-	kaddr0 = kmap_atomic(spage);
-	kaddr1 = kmap_atomic(dpage);
+	kaddr0 = kmap_local_page(spage);
+	kaddr1 = kmap_local_page(dpage);
 	memcpy(kaddr1 + bh_offset(dbh), kaddr0 + bh_offset(sbh), sbh->b_size);
-	kunmap_atomic(kaddr1);
-	kunmap_atomic(kaddr0);
+	kunmap_local(kaddr1);
+	kunmap_local(kaddr0);
 
 	dbh->b_state = sbh->b_state & NILFS_BUFFER_INHERENT_BITS;
 	dbh->b_blocknr = sbh->b_blocknr;
diff --git a/fs/nilfs2/recovery.c b/fs/nilfs2/recovery.c
index a9b8d77c8c..b638dc06df 100644
--- a/fs/nilfs2/recovery.c
+++ b/fs/nilfs2/recovery.c
@@ -482,9 +482,9 @@ static int nilfs_recovery_copy_block(struct the_nilfs *nilfs,
 	if (unlikely(!bh_org))
 		return -EIO;
 
-	kaddr = kmap_atomic(page);
+	kaddr = kmap_local_page(page);
 	memcpy(kaddr + from, bh_org->b_data, bh_org->b_size);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 	brelse(bh_org);
 	return 0;
 }
@@ -563,6 +563,7 @@ static int nilfs_recover_dsync_blocks(struct the_nilfs *nilfs,
  * checkpoint
  * @nilfs: nilfs object
  * @sb: super block instance
+ * @root: NILFS root instance
  * @ri: pointer to a nilfs_recovery_info
  */
 static int nilfs_do_roll_forward(struct the_nilfs *nilfs,
@@ -698,9 +699,15 @@ static void nilfs_finish_roll_forward(struct the_nilfs *nilfs,
 		return;
 
 	bh = __getblk(nilfs->ns_bdev, ri->ri_lsegs_start, nilfs->ns_blocksize);
-	BUG_ON(!bh);
+	if (WARN_ON(!bh))
+		return;  /* should never happen */
+
+	lock_buffer(bh);
 	memset(bh->b_data, 0, bh->b_size);
+	set_buffer_uptodate(bh);
 	set_buffer_dirty(bh);
+	unlock_buffer(bh);
+
 	err = sync_dirty_buffer(bh);
 	if (unlikely(err))
 		nilfs_warn(nilfs->ns_sb,
diff --git a/fs/nilfs2/segbuf.c b/fs/nilfs2/segbuf.c
index 6e59dc19a7..dc431b4c34 100644
--- a/fs/nilfs2/segbuf.c
+++ b/fs/nilfs2/segbuf.c
@@ -220,9 +220,9 @@ static void nilfs_segbuf_fill_in_data_crc(struct nilfs_segment_buffer *segbuf,
 		crc = crc32_le(crc, bh->b_data, bh->b_size);
 	}
 	list_for_each_entry(bh, &segbuf->sb_payload_buffers, b_assoc_buffers) {
-		kaddr = kmap_atomic(bh->b_page);
+		kaddr = kmap_local_page(bh->b_page);
 		crc = crc32_le(crc, kaddr + bh_offset(bh), bh->b_size);
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 	}
 	raw_sum->ss_datasum = cpu_to_le32(crc);
 }
diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c
index 5b3f530016..d02fd92cdb 100644
--- a/fs/nilfs2/segment.c
+++ b/fs/nilfs2/segment.c
@@ -136,7 +136,7 @@ static void nilfs_dispose_list(struct the_nilfs *, struct list_head *, int);
 
 #define nilfs_cnt32_ge(a, b)   \
 	(typecheck(__u32, a) && typecheck(__u32, b) && \
-	 ((__s32)(a) - (__s32)(b) >= 0))
+	 ((__s32)((a) - (b)) >= 0))
 
 static int nilfs_prepare_segment_lock(struct super_block *sb,
 				      struct nilfs_transaction_info *ti)
@@ -880,76 +880,6 @@ static void nilfs_segctor_clear_metadata_dirty(struct nilfs_sc_info *sci)
 	nilfs_mdt_clear_dirty(nilfs->ns_dat);
 }
 
-static int nilfs_segctor_create_checkpoint(struct nilfs_sc_info *sci)
-{
-	struct the_nilfs *nilfs = sci->sc_super->s_fs_info;
-	struct buffer_head *bh_cp;
-	struct nilfs_checkpoint *raw_cp;
-	int err;
-
-	/* XXX: this interface will be changed */
-	err = nilfs_cpfile_get_checkpoint(nilfs->ns_cpfile, nilfs->ns_cno, 1,
-					  &raw_cp, &bh_cp);
-	if (likely(!err)) {
-		/*
-		 * The following code is duplicated with cpfile.  But, it is
-		 * needed to collect the checkpoint even if it was not newly
-		 * created.
-		 */
-		mark_buffer_dirty(bh_cp);
-		nilfs_mdt_mark_dirty(nilfs->ns_cpfile);
-		nilfs_cpfile_put_checkpoint(
-			nilfs->ns_cpfile, nilfs->ns_cno, bh_cp);
-	} else if (err == -EINVAL || err == -ENOENT) {
-		nilfs_error(sci->sc_super,
-			    "checkpoint creation failed due to metadata corruption.");
-		err = -EIO;
-	}
-	return err;
-}
-
-static int nilfs_segctor_fill_in_checkpoint(struct nilfs_sc_info *sci)
-{
-	struct the_nilfs *nilfs = sci->sc_super->s_fs_info;
-	struct buffer_head *bh_cp;
-	struct nilfs_checkpoint *raw_cp;
-	int err;
-
-	err = nilfs_cpfile_get_checkpoint(nilfs->ns_cpfile, nilfs->ns_cno, 0,
-					  &raw_cp, &bh_cp);
-	if (unlikely(err)) {
-		if (err == -EINVAL || err == -ENOENT) {
-			nilfs_error(sci->sc_super,
-				    "checkpoint finalization failed due to metadata corruption.");
-			err = -EIO;
-		}
-		goto failed_ibh;
-	}
-	raw_cp->cp_snapshot_list.ssl_next = 0;
-	raw_cp->cp_snapshot_list.ssl_prev = 0;
-	raw_cp->cp_inodes_count =
-		cpu_to_le64(atomic64_read(&sci->sc_root->inodes_count));
-	raw_cp->cp_blocks_count =
-		cpu_to_le64(atomic64_read(&sci->sc_root->blocks_count));
-	raw_cp->cp_nblk_inc =
-		cpu_to_le64(sci->sc_nblk_inc + sci->sc_nblk_this_inc);
-	raw_cp->cp_create = cpu_to_le64(sci->sc_seg_ctime);
-	raw_cp->cp_cno = cpu_to_le64(nilfs->ns_cno);
-
-	if (test_bit(NILFS_SC_HAVE_DELTA, &sci->sc_flags))
-		nilfs_checkpoint_clear_minor(raw_cp);
-	else
-		nilfs_checkpoint_set_minor(raw_cp);
-
-	nilfs_write_inode_common(sci->sc_root->ifile,
-				 &raw_cp->cp_ifile_inode, 1);
-	nilfs_cpfile_put_checkpoint(nilfs->ns_cpfile, nilfs->ns_cno, bh_cp);
-	return 0;
-
- failed_ibh:
-	return err;
-}
-
 static void nilfs_fill_in_file_bmap(struct inode *ifile,
 				    struct nilfs_inode_info *ii)
 
@@ -963,7 +893,7 @@ static void nilfs_fill_in_file_bmap(struct inode *ifile,
 		raw_inode = nilfs_ifile_map_inode(ifile, ii->vfs_inode.i_ino,
 						  ibh);
 		nilfs_bmap_write(ii->i_bmap, raw_inode);
-		nilfs_ifile_unmap_inode(ifile, ii->vfs_inode.i_ino, ibh);
+		nilfs_ifile_unmap_inode(raw_inode);
 	}
 }
 
@@ -977,6 +907,33 @@ static void nilfs_segctor_fill_in_file_bmap(struct nilfs_sc_info *sci)
 	}
 }
 
+/**
+ * nilfs_write_root_mdt_inode - export root metadata inode information to
+ *                              the on-disk inode
+ * @inode:     inode object of the root metadata file
+ * @raw_inode: on-disk inode
+ *
+ * nilfs_write_root_mdt_inode() writes inode information and bmap data of
+ * @inode to the inode area of the metadata file allocated on the super root
+ * block created to finalize the log.  Since super root blocks are configured
+ * each time, this function zero-fills the unused area of @raw_inode.
+ */
+static void nilfs_write_root_mdt_inode(struct inode *inode,
+				       struct nilfs_inode *raw_inode)
+{
+	struct the_nilfs *nilfs = inode->i_sb->s_fs_info;
+
+	nilfs_write_inode_common(inode, raw_inode);
+
+	/* zero-fill unused portion of raw_inode */
+	raw_inode->i_xattr = 0;
+	raw_inode->i_pad = 0;
+	memset((void *)raw_inode + sizeof(*raw_inode), 0,
+	       nilfs->ns_inode_size - sizeof(*raw_inode));
+
+	nilfs_bmap_write(NILFS_I(inode)->i_bmap, raw_inode);
+}
+
 static void nilfs_segctor_fill_in_super_root(struct nilfs_sc_info *sci,
 					     struct the_nilfs *nilfs)
 {
@@ -998,12 +955,13 @@ static void nilfs_segctor_fill_in_super_root(struct nilfs_sc_info *sci,
 			      nilfs->ns_nongc_ctime : sci->sc_seg_ctime);
 	raw_sr->sr_flags = 0;
 
-	nilfs_write_inode_common(nilfs->ns_dat, (void *)raw_sr +
-				 NILFS_SR_DAT_OFFSET(isz), 1);
-	nilfs_write_inode_common(nilfs->ns_cpfile, (void *)raw_sr +
-				 NILFS_SR_CPFILE_OFFSET(isz), 1);
-	nilfs_write_inode_common(nilfs->ns_sufile, (void *)raw_sr +
-				 NILFS_SR_SUFILE_OFFSET(isz), 1);
+	nilfs_write_root_mdt_inode(nilfs->ns_dat, (void *)raw_sr +
+				   NILFS_SR_DAT_OFFSET(isz));
+	nilfs_write_root_mdt_inode(nilfs->ns_cpfile, (void *)raw_sr +
+				   NILFS_SR_CPFILE_OFFSET(isz));
+	nilfs_write_root_mdt_inode(nilfs->ns_sufile, (void *)raw_sr +
+				   NILFS_SR_SUFILE_OFFSET(isz));
+
 	memset((void *)raw_sr + srsz, 0, nilfs->ns_blocksize - srsz);
 	set_buffer_uptodate(bh_sr);
 	unlock_buffer(bh_sr);
@@ -1230,7 +1188,8 @@ static int nilfs_segctor_collect_blocks(struct nilfs_sc_info *sci, int mode)
 			break;
 		nilfs_sc_cstage_inc(sci);
 		/* Creating a checkpoint */
-		err = nilfs_segctor_create_checkpoint(sci);
+		err = nilfs_cpfile_create_checkpoint(nilfs->ns_cpfile,
+						     nilfs->ns_cno);
 		if (unlikely(err))
 			break;
 		fallthrough;
@@ -1693,6 +1652,7 @@ static void nilfs_segctor_prepare_write(struct nilfs_sc_info *sci)
 			if (bh->b_folio != bd_folio) {
 				if (bd_folio) {
 					folio_lock(bd_folio);
+					folio_wait_writeback(bd_folio);
 					folio_clear_dirty_for_io(bd_folio);
 					folio_start_writeback(bd_folio);
 					folio_unlock(bd_folio);
@@ -1706,6 +1666,7 @@ static void nilfs_segctor_prepare_write(struct nilfs_sc_info *sci)
 			if (bh == segbuf->sb_super_root) {
 				if (bh->b_folio != bd_folio) {
 					folio_lock(bd_folio);
+					folio_wait_writeback(bd_folio);
 					folio_clear_dirty_for_io(bd_folio);
 					folio_start_writeback(bd_folio);
 					folio_unlock(bd_folio);
@@ -1722,6 +1683,7 @@ static void nilfs_segctor_prepare_write(struct nilfs_sc_info *sci)
 	}
 	if (bd_folio) {
 		folio_lock(bd_folio);
+		folio_wait_writeback(bd_folio);
 		folio_clear_dirty_for_io(bd_folio);
 		folio_start_writeback(bd_folio);
 		folio_unlock(bd_folio);
@@ -1766,14 +1728,8 @@ static void nilfs_end_folio_io(struct folio *folio, int err)
 		return;
 	}
 
-	if (!err) {
-		if (!nilfs_folio_buffers_clean(folio))
-			filemap_dirty_folio(folio->mapping, folio);
-		folio_clear_error(folio);
-	} else {
+	if (err || !nilfs_folio_buffers_clean(folio))
 		filemap_dirty_folio(folio->mapping, folio);
-		folio_set_error(folio);
-	}
 
 	folio_end_writeback(folio);
 }
@@ -2101,7 +2057,11 @@ static int nilfs_segctor_do_construct(struct nilfs_sc_info *sci, int mode)
 
 		if (mode == SC_LSEG_SR &&
 		    nilfs_sc_cstage_get(sci) >= NILFS_ST_CPFILE) {
-			err = nilfs_segctor_fill_in_checkpoint(sci);
+			err = nilfs_cpfile_finalize_checkpoint(
+				nilfs->ns_cpfile, nilfs->ns_cno, sci->sc_root,
+				sci->sc_nblk_inc + sci->sc_nblk_this_inc,
+				sci->sc_seg_ctime,
+				!test_bit(NILFS_SC_HAVE_DELTA, &sci->sc_flags));
 			if (unlikely(err))
 				goto failed_to_write;
 
@@ -2864,7 +2824,7 @@ int nilfs_attach_log_writer(struct super_block *sb, struct nilfs_root *root)
 	if (!nilfs->ns_writer)
 		return -ENOMEM;
 
-	inode_attach_wb(nilfs->ns_bdev->bd_inode, NULL);
+	inode_attach_wb(nilfs->ns_bdev->bd_mapping->host, NULL);
 
 	err = nilfs_segctor_start_thread(nilfs->ns_writer);
 	if (unlikely(err))
diff --git a/fs/nilfs2/sufile.c b/fs/nilfs2/sufile.c
index 0a8119456c..6748218be7 100644
--- a/fs/nilfs2/sufile.c
+++ b/fs/nilfs2/sufile.c
@@ -48,7 +48,7 @@ nilfs_sufile_get_blkoff(const struct inode *sufile, __u64 segnum)
 {
 	__u64 t = segnum + NILFS_MDT(sufile)->mi_first_entry_offset;
 
-	do_div(t, nilfs_sufile_segment_usages_per_block(sufile));
+	t = div64_ul(t, nilfs_sufile_segment_usages_per_block(sufile));
 	return (unsigned long)t;
 }
 
@@ -107,11 +107,11 @@ static void nilfs_sufile_mod_counter(struct buffer_head *header_bh,
 	struct nilfs_sufile_header *header;
 	void *kaddr;
 
-	kaddr = kmap_atomic(header_bh->b_page);
+	kaddr = kmap_local_page(header_bh->b_page);
 	header = kaddr + bh_offset(header_bh);
 	le64_add_cpu(&header->sh_ncleansegs, ncleanadd);
 	le64_add_cpu(&header->sh_ndirtysegs, ndirtyadd);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	mark_buffer_dirty(header_bh);
 }
@@ -315,10 +315,10 @@ int nilfs_sufile_alloc(struct inode *sufile, __u64 *segnump)
 	ret = nilfs_sufile_get_header_block(sufile, &header_bh);
 	if (ret < 0)
 		goto out_sem;
-	kaddr = kmap_atomic(header_bh->b_page);
+	kaddr = kmap_local_page(header_bh->b_page);
 	header = kaddr + bh_offset(header_bh);
 	last_alloc = le64_to_cpu(header->sh_last_alloc);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	nsegments = nilfs_sufile_get_nsegments(sufile);
 	maxsegnum = sui->allocmax;
@@ -352,7 +352,7 @@ int nilfs_sufile_alloc(struct inode *sufile, __u64 *segnump)
 							   &su_bh);
 		if (ret < 0)
 			goto out_header;
-		kaddr = kmap_atomic(su_bh->b_page);
+		kaddr = kmap_local_page(su_bh->b_page);
 		su = nilfs_sufile_block_get_segment_usage(
 			sufile, segnum, su_bh, kaddr);
 
@@ -363,14 +363,14 @@ int nilfs_sufile_alloc(struct inode *sufile, __u64 *segnump)
 				continue;
 			/* found a clean segment */
 			nilfs_segment_usage_set_dirty(su);
-			kunmap_atomic(kaddr);
+			kunmap_local(kaddr);
 
-			kaddr = kmap_atomic(header_bh->b_page);
+			kaddr = kmap_local_page(header_bh->b_page);
 			header = kaddr + bh_offset(header_bh);
 			le64_add_cpu(&header->sh_ncleansegs, -1);
 			le64_add_cpu(&header->sh_ndirtysegs, 1);
 			header->sh_last_alloc = cpu_to_le64(segnum);
-			kunmap_atomic(kaddr);
+			kunmap_local(kaddr);
 
 			sui->ncleansegs--;
 			mark_buffer_dirty(header_bh);
@@ -384,7 +384,7 @@ int nilfs_sufile_alloc(struct inode *sufile, __u64 *segnump)
 			goto out_header;
 		}
 
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		brelse(su_bh);
 	}
 
@@ -406,16 +406,16 @@ void nilfs_sufile_do_cancel_free(struct inode *sufile, __u64 segnum,
 	struct nilfs_segment_usage *su;
 	void *kaddr;
 
-	kaddr = kmap_atomic(su_bh->b_page);
+	kaddr = kmap_local_page(su_bh->b_page);
 	su = nilfs_sufile_block_get_segment_usage(sufile, segnum, su_bh, kaddr);
 	if (unlikely(!nilfs_segment_usage_clean(su))) {
 		nilfs_warn(sufile->i_sb, "%s: segment %llu must be clean",
 			   __func__, (unsigned long long)segnum);
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		return;
 	}
 	nilfs_segment_usage_set_dirty(su);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	nilfs_sufile_mod_counter(header_bh, -1, 1);
 	NILFS_SUI(sufile)->ncleansegs--;
@@ -432,11 +432,11 @@ void nilfs_sufile_do_scrap(struct inode *sufile, __u64 segnum,
 	void *kaddr;
 	int clean, dirty;
 
-	kaddr = kmap_atomic(su_bh->b_page);
+	kaddr = kmap_local_page(su_bh->b_page);
 	su = nilfs_sufile_block_get_segment_usage(sufile, segnum, su_bh, kaddr);
 	if (su->su_flags == cpu_to_le32(BIT(NILFS_SEGMENT_USAGE_DIRTY)) &&
 	    su->su_nblocks == cpu_to_le32(0)) {
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		return;
 	}
 	clean = nilfs_segment_usage_clean(su);
@@ -446,7 +446,7 @@ void nilfs_sufile_do_scrap(struct inode *sufile, __u64 segnum,
 	su->su_lastmod = cpu_to_le64(0);
 	su->su_nblocks = cpu_to_le32(0);
 	su->su_flags = cpu_to_le32(BIT(NILFS_SEGMENT_USAGE_DIRTY));
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	nilfs_sufile_mod_counter(header_bh, clean ? (u64)-1 : 0, dirty ? 0 : 1);
 	NILFS_SUI(sufile)->ncleansegs -= clean;
@@ -463,12 +463,12 @@ void nilfs_sufile_do_free(struct inode *sufile, __u64 segnum,
 	void *kaddr;
 	int sudirty;
 
-	kaddr = kmap_atomic(su_bh->b_page);
+	kaddr = kmap_local_page(su_bh->b_page);
 	su = nilfs_sufile_block_get_segment_usage(sufile, segnum, su_bh, kaddr);
 	if (nilfs_segment_usage_clean(su)) {
 		nilfs_warn(sufile->i_sb, "%s: segment %llu is already clean",
 			   __func__, (unsigned long long)segnum);
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		return;
 	}
 	if (unlikely(nilfs_segment_usage_error(su)))
@@ -481,7 +481,7 @@ void nilfs_sufile_do_free(struct inode *sufile, __u64 segnum,
 			   (unsigned long long)segnum);
 
 	nilfs_segment_usage_set_clean(su);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 	mark_buffer_dirty(su_bh);
 
 	nilfs_sufile_mod_counter(header_bh, 1, sudirty ? (u64)-1 : 0);
@@ -509,12 +509,12 @@ int nilfs_sufile_mark_dirty(struct inode *sufile, __u64 segnum)
 	if (ret)
 		goto out_sem;
 
-	kaddr = kmap_atomic(bh->b_page);
+	kaddr = kmap_local_page(bh->b_page);
 	su = nilfs_sufile_block_get_segment_usage(sufile, segnum, bh, kaddr);
 	if (unlikely(nilfs_segment_usage_error(su))) {
 		struct the_nilfs *nilfs = sufile->i_sb->s_fs_info;
 
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		brelse(bh);
 		if (nilfs_segment_is_active(nilfs, segnum)) {
 			nilfs_error(sufile->i_sb,
@@ -532,7 +532,7 @@ int nilfs_sufile_mark_dirty(struct inode *sufile, __u64 segnum)
 		ret = -EIO;
 	} else {
 		nilfs_segment_usage_set_dirty(su);
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		mark_buffer_dirty(bh);
 		nilfs_mdt_mark_dirty(sufile);
 		brelse(bh);
@@ -562,7 +562,7 @@ int nilfs_sufile_set_segment_usage(struct inode *sufile, __u64 segnum,
 	if (ret < 0)
 		goto out_sem;
 
-	kaddr = kmap_atomic(bh->b_page);
+	kaddr = kmap_local_page(bh->b_page);
 	su = nilfs_sufile_block_get_segment_usage(sufile, segnum, bh, kaddr);
 	if (modtime) {
 		/*
@@ -573,7 +573,7 @@ int nilfs_sufile_set_segment_usage(struct inode *sufile, __u64 segnum,
 		su->su_lastmod = cpu_to_le64(modtime);
 	}
 	su->su_nblocks = cpu_to_le32(nblocks);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	mark_buffer_dirty(bh);
 	nilfs_mdt_mark_dirty(sufile);
@@ -614,7 +614,7 @@ int nilfs_sufile_get_stat(struct inode *sufile, struct nilfs_sustat *sustat)
 	if (ret < 0)
 		goto out_sem;
 
-	kaddr = kmap_atomic(header_bh->b_page);
+	kaddr = kmap_local_page(header_bh->b_page);
 	header = kaddr + bh_offset(header_bh);
 	sustat->ss_nsegs = nilfs_sufile_get_nsegments(sufile);
 	sustat->ss_ncleansegs = le64_to_cpu(header->sh_ncleansegs);
@@ -624,7 +624,7 @@ int nilfs_sufile_get_stat(struct inode *sufile, struct nilfs_sustat *sustat)
 	spin_lock(&nilfs->ns_last_segment_lock);
 	sustat->ss_prot_seq = nilfs->ns_prot_seq;
 	spin_unlock(&nilfs->ns_last_segment_lock);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 	brelse(header_bh);
 
  out_sem:
@@ -640,15 +640,15 @@ void nilfs_sufile_do_set_error(struct inode *sufile, __u64 segnum,
 	void *kaddr;
 	int suclean;
 
-	kaddr = kmap_atomic(su_bh->b_page);
+	kaddr = kmap_local_page(su_bh->b_page);
 	su = nilfs_sufile_block_get_segment_usage(sufile, segnum, su_bh, kaddr);
 	if (nilfs_segment_usage_error(su)) {
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		return;
 	}
 	suclean = nilfs_segment_usage_clean(su);
 	nilfs_segment_usage_set_error(su);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	if (suclean) {
 		nilfs_sufile_mod_counter(header_bh, -1, 0);
@@ -717,7 +717,7 @@ static int nilfs_sufile_truncate_range(struct inode *sufile,
 			/* hole */
 			continue;
 		}
-		kaddr = kmap_atomic(su_bh->b_page);
+		kaddr = kmap_local_page(su_bh->b_page);
 		su = nilfs_sufile_block_get_segment_usage(
 			sufile, segnum, su_bh, kaddr);
 		su2 = su;
@@ -726,7 +726,7 @@ static int nilfs_sufile_truncate_range(struct inode *sufile,
 			     ~BIT(NILFS_SEGMENT_USAGE_ERROR)) ||
 			    nilfs_segment_is_active(nilfs, segnum + j)) {
 				ret = -EBUSY;
-				kunmap_atomic(kaddr);
+				kunmap_local(kaddr);
 				brelse(su_bh);
 				goto out_header;
 			}
@@ -738,7 +738,7 @@ static int nilfs_sufile_truncate_range(struct inode *sufile,
 				nc++;
 			}
 		}
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		if (nc > 0) {
 			mark_buffer_dirty(su_bh);
 			ncleaned += nc;
@@ -823,10 +823,10 @@ int nilfs_sufile_resize(struct inode *sufile, __u64 newnsegs)
 		sui->allocmin = 0;
 	}
 
-	kaddr = kmap_atomic(header_bh->b_page);
+	kaddr = kmap_local_page(header_bh->b_page);
 	header = kaddr + bh_offset(header_bh);
 	header->sh_ncleansegs = cpu_to_le64(sui->ncleansegs);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 
 	mark_buffer_dirty(header_bh);
 	nilfs_mdt_mark_dirty(sufile);
@@ -891,7 +891,7 @@ ssize_t nilfs_sufile_get_suinfo(struct inode *sufile, __u64 segnum, void *buf,
 			continue;
 		}
 
-		kaddr = kmap_atomic(su_bh->b_page);
+		kaddr = kmap_local_page(su_bh->b_page);
 		su = nilfs_sufile_block_get_segment_usage(
 			sufile, segnum, su_bh, kaddr);
 		for (j = 0; j < n;
@@ -904,7 +904,7 @@ ssize_t nilfs_sufile_get_suinfo(struct inode *sufile, __u64 segnum, void *buf,
 				si->sui_flags |=
 					BIT(NILFS_SEGMENT_USAGE_ACTIVE);
 		}
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		brelse(su_bh);
 	}
 	ret = nsegs;
@@ -973,7 +973,7 @@ ssize_t nilfs_sufile_set_suinfo(struct inode *sufile, void *buf,
 		goto out_header;
 
 	for (;;) {
-		kaddr = kmap_atomic(bh->b_page);
+		kaddr = kmap_local_page(bh->b_page);
 		su = nilfs_sufile_block_get_segment_usage(
 			sufile, sup->sup_segnum, bh, kaddr);
 
@@ -1010,7 +1010,7 @@ ssize_t nilfs_sufile_set_suinfo(struct inode *sufile, void *buf,
 			su->su_flags = cpu_to_le32(sup->sup_sui.sui_flags);
 		}
 
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 
 		sup = (void *)sup + supsz;
 		if (sup >= supend)
@@ -1115,7 +1115,7 @@ int nilfs_sufile_trim_fs(struct inode *sufile, struct fstrim_range *range)
 			continue;
 		}
 
-		kaddr = kmap_atomic(su_bh->b_page);
+		kaddr = kmap_local_page(su_bh->b_page);
 		su = nilfs_sufile_block_get_segment_usage(sufile, segnum,
 				su_bh, kaddr);
 		for (i = 0; i < n; ++i, ++segnum, su = (void *)su + susz) {
@@ -1145,7 +1145,7 @@ int nilfs_sufile_trim_fs(struct inode *sufile, struct fstrim_range *range)
 			}
 
 			if (nblocks >= minlen) {
-				kunmap_atomic(kaddr);
+				kunmap_local(kaddr);
 
 				ret = blkdev_issue_discard(nilfs->ns_bdev,
 						start * sects_per_block,
@@ -1157,7 +1157,7 @@ int nilfs_sufile_trim_fs(struct inode *sufile, struct fstrim_range *range)
 				}
 
 				ndiscarded += nblocks;
-				kaddr = kmap_atomic(su_bh->b_page);
+				kaddr = kmap_local_page(su_bh->b_page);
 				su = nilfs_sufile_block_get_segment_usage(
 					sufile, segnum, su_bh, kaddr);
 			}
@@ -1166,7 +1166,7 @@ int nilfs_sufile_trim_fs(struct inode *sufile, struct fstrim_range *range)
 			start = seg_start;
 			nblocks = seg_end - seg_start + 1;
 		}
-		kunmap_atomic(kaddr);
+		kunmap_local(kaddr);
 		put_bh(su_bh);
 	}
 
@@ -1246,10 +1246,10 @@ int nilfs_sufile_read(struct super_block *sb, size_t susize,
 		goto failed;
 
 	sui = NILFS_SUI(sufile);
-	kaddr = kmap_atomic(header_bh->b_page);
+	kaddr = kmap_local_page(header_bh->b_page);
 	header = kaddr + bh_offset(header_bh);
 	sui->ncleansegs = le64_to_cpu(header->sh_ncleansegs);
-	kunmap_atomic(kaddr);
+	kunmap_local(kaddr);
 	brelse(header_bh);
 
 	sui->allocmax = nilfs_sufile_get_nsegments(sufile) - 1;
diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c
index df8674173b..e835e1f5a7 100644
--- a/fs/nilfs2/super.c
+++ b/fs/nilfs2/super.c
@@ -29,13 +29,13 @@
 #include <linux/slab.h>
 #include <linux/init.h>
 #include <linux/blkdev.h>
-#include <linux/parser.h>
 #include <linux/crc32.h>
 #include <linux/vfs.h>
 #include <linux/writeback.h>
 #include <linux/seq_file.h>
 #include <linux/mount.h>
 #include <linux/fs_context.h>
+#include <linux/fs_parser.h>
 #include "nilfs.h"
 #include "export.h"
 #include "mdt.h"
@@ -61,7 +61,6 @@ struct kmem_cache *nilfs_segbuf_cachep;
 struct kmem_cache *nilfs_btree_path_cache;
 
 static int nilfs_setup_super(struct super_block *sb, int is_mount);
-static int nilfs_remount(struct super_block *sb, int *flags, char *data);
 
 void __nilfs_msg(struct super_block *sb, const char *fmt, ...)
 {
@@ -448,7 +447,7 @@ int nilfs_resize_fs(struct super_block *sb, __u64 newsize)
 
 	sb2off = NILFS_SB2_OFFSET_BYTES(newsize);
 	newnsegs = sb2off >> nilfs->ns_blocksize_bits;
-	do_div(newnsegs, nilfs->ns_blocks_per_segment);
+	newnsegs = div64_ul(newnsegs, nilfs->ns_blocks_per_segment);
 
 	ret = nilfs_sufile_resize(nilfs->ns_sufile, newnsegs);
 	up_write(&nilfs->ns_segctor_sem);
@@ -544,8 +543,6 @@ int nilfs_attach_checkpoint(struct super_block *sb, __u64 cno, int curr_mnt,
 {
 	struct the_nilfs *nilfs = sb->s_fs_info;
 	struct nilfs_root *root;
-	struct nilfs_checkpoint *raw_cp;
-	struct buffer_head *bh_cp;
 	int err = -ENOMEM;
 
 	root = nilfs_find_or_create_root(
@@ -557,38 +554,19 @@ int nilfs_attach_checkpoint(struct super_block *sb, __u64 cno, int curr_mnt,
 		goto reuse; /* already attached checkpoint */
 
 	down_read(&nilfs->ns_segctor_sem);
-	err = nilfs_cpfile_get_checkpoint(nilfs->ns_cpfile, cno, 0, &raw_cp,
-					  &bh_cp);
+	err = nilfs_ifile_read(sb, root, cno, nilfs->ns_inode_size);
 	up_read(&nilfs->ns_segctor_sem);
-	if (unlikely(err)) {
-		if (err == -ENOENT || err == -EINVAL) {
-			nilfs_err(sb,
-				  "Invalid checkpoint (checkpoint number=%llu)",
-				  (unsigned long long)cno);
-			err = -EINVAL;
-		}
+	if (unlikely(err))
 		goto failed;
-	}
-
-	err = nilfs_ifile_read(sb, root, nilfs->ns_inode_size,
-			       &raw_cp->cp_ifile_inode, &root->ifile);
-	if (err)
-		goto failed_bh;
-
-	atomic64_set(&root->inodes_count,
-			le64_to_cpu(raw_cp->cp_inodes_count));
-	atomic64_set(&root->blocks_count,
-			le64_to_cpu(raw_cp->cp_blocks_count));
-
-	nilfs_cpfile_put_checkpoint(nilfs->ns_cpfile, cno, bh_cp);
 
  reuse:
 	*rootp = root;
 	return 0;
 
- failed_bh:
-	nilfs_cpfile_put_checkpoint(nilfs->ns_cpfile, cno, bh_cp);
  failed:
+	if (err == -EINVAL)
+		nilfs_err(sb, "Invalid checkpoint (checkpoint number=%llu)",
+			  (unsigned long long)cno);
 	nilfs_put_root(root);
 
 	return err;
@@ -723,105 +701,98 @@ static const struct super_operations nilfs_sops = {
 	.freeze_fs	= nilfs_freeze,
 	.unfreeze_fs	= nilfs_unfreeze,
 	.statfs         = nilfs_statfs,
-	.remount_fs     = nilfs_remount,
 	.show_options = nilfs_show_options
 };
 
 enum {
-	Opt_err_cont, Opt_err_panic, Opt_err_ro,
-	Opt_barrier, Opt_nobarrier, Opt_snapshot, Opt_order, Opt_norecovery,
-	Opt_discard, Opt_nodiscard, Opt_err,
+	Opt_err, Opt_barrier, Opt_snapshot, Opt_order, Opt_norecovery,
+	Opt_discard,
 };
 
-static match_table_t tokens = {
-	{Opt_err_cont, "errors=continue"},
-	{Opt_err_panic, "errors=panic"},
-	{Opt_err_ro, "errors=remount-ro"},
-	{Opt_barrier, "barrier"},
-	{Opt_nobarrier, "nobarrier"},
-	{Opt_snapshot, "cp=%u"},
-	{Opt_order, "order=%s"},
-	{Opt_norecovery, "norecovery"},
-	{Opt_discard, "discard"},
-	{Opt_nodiscard, "nodiscard"},
-	{Opt_err, NULL}
+static const struct constant_table nilfs_param_err[] = {
+	{"continue",	NILFS_MOUNT_ERRORS_CONT},
+	{"panic",	NILFS_MOUNT_ERRORS_PANIC},
+	{"remount-ro",	NILFS_MOUNT_ERRORS_RO},
+	{}
 };
 
-static int parse_options(char *options, struct super_block *sb, int is_remount)
-{
-	struct the_nilfs *nilfs = sb->s_fs_info;
-	char *p;
-	substring_t args[MAX_OPT_ARGS];
-
-	if (!options)
-		return 1;
-
-	while ((p = strsep(&options, ",")) != NULL) {
-		int token;
+static const struct fs_parameter_spec nilfs_param_spec[] = {
+	fsparam_enum	("errors", Opt_err, nilfs_param_err),
+	fsparam_flag_no	("barrier", Opt_barrier),
+	fsparam_u64	("cp", Opt_snapshot),
+	fsparam_string	("order", Opt_order),
+	fsparam_flag	("norecovery", Opt_norecovery),
+	fsparam_flag_no	("discard", Opt_discard),
+	{}
+};
 
-		if (!*p)
-			continue;
+struct nilfs_fs_context {
+	unsigned long ns_mount_opt;
+	__u64 cno;
+};
 
-		token = match_token(p, tokens, args);
-		switch (token) {
-		case Opt_barrier:
-			nilfs_set_opt(nilfs, BARRIER);
-			break;
-		case Opt_nobarrier:
+static int nilfs_parse_param(struct fs_context *fc, struct fs_parameter *param)
+{
+	struct nilfs_fs_context *nilfs = fc->fs_private;
+	int is_remount = fc->purpose == FS_CONTEXT_FOR_RECONFIGURE;
+	struct fs_parse_result result;
+	int opt;
+
+	opt = fs_parse(fc, nilfs_param_spec, param, &result);
+	if (opt < 0)
+		return opt;
+
+	switch (opt) {
+	case Opt_barrier:
+		if (result.negated)
 			nilfs_clear_opt(nilfs, BARRIER);
-			break;
-		case Opt_order:
-			if (strcmp(args[0].from, "relaxed") == 0)
-				/* Ordered data semantics */
-				nilfs_clear_opt(nilfs, STRICT_ORDER);
-			else if (strcmp(args[0].from, "strict") == 0)
-				/* Strict in-order semantics */
-				nilfs_set_opt(nilfs, STRICT_ORDER);
-			else
-				return 0;
-			break;
-		case Opt_err_panic:
-			nilfs_write_opt(nilfs, ERROR_MODE, ERRORS_PANIC);
-			break;
-		case Opt_err_ro:
-			nilfs_write_opt(nilfs, ERROR_MODE, ERRORS_RO);
-			break;
-		case Opt_err_cont:
-			nilfs_write_opt(nilfs, ERROR_MODE, ERRORS_CONT);
-			break;
-		case Opt_snapshot:
-			if (is_remount) {
-				nilfs_err(sb,
-					  "\"%s\" option is invalid for remount",
-					  p);
-				return 0;
-			}
-			break;
-		case Opt_norecovery:
-			nilfs_set_opt(nilfs, NORECOVERY);
-			break;
-		case Opt_discard:
-			nilfs_set_opt(nilfs, DISCARD);
-			break;
-		case Opt_nodiscard:
-			nilfs_clear_opt(nilfs, DISCARD);
-			break;
-		default:
-			nilfs_err(sb, "unrecognized mount option \"%s\"", p);
-			return 0;
+		else
+			nilfs_set_opt(nilfs, BARRIER);
+		break;
+	case Opt_order:
+		if (strcmp(param->string, "relaxed") == 0)
+			/* Ordered data semantics */
+			nilfs_clear_opt(nilfs, STRICT_ORDER);
+		else if (strcmp(param->string, "strict") == 0)
+			/* Strict in-order semantics */
+			nilfs_set_opt(nilfs, STRICT_ORDER);
+		else
+			return -EINVAL;
+		break;
+	case Opt_err:
+		nilfs->ns_mount_opt &= ~NILFS_MOUNT_ERROR_MODE;
+		nilfs->ns_mount_opt |= result.uint_32;
+		break;
+	case Opt_snapshot:
+		if (is_remount) {
+			struct super_block *sb = fc->root->d_sb;
+
+			nilfs_err(sb,
+				  "\"%s\" option is invalid for remount",
+				  param->key);
+			return -EINVAL;
 		}
+		if (result.uint_64 == 0) {
+			nilfs_err(NULL,
+				  "invalid option \"cp=0\": invalid checkpoint number 0");
+			return -EINVAL;
+		}
+		nilfs->cno = result.uint_64;
+		break;
+	case Opt_norecovery:
+		nilfs_set_opt(nilfs, NORECOVERY);
+		break;
+	case Opt_discard:
+		if (result.negated)
+			nilfs_clear_opt(nilfs, DISCARD);
+		else
+			nilfs_set_opt(nilfs, DISCARD);
+		break;
+	default:
+		return -EINVAL;
 	}
-	return 1;
-}
-
-static inline void
-nilfs_set_default_options(struct super_block *sb,
-			  struct nilfs_super_block *sbp)
-{
-	struct the_nilfs *nilfs = sb->s_fs_info;
 
-	nilfs->ns_mount_opt =
-		NILFS_MOUNT_ERRORS_RO | NILFS_MOUNT_BARRIER;
+	return 0;
 }
 
 static int nilfs_setup_super(struct super_block *sb, int is_mount)
@@ -878,9 +849,8 @@ struct nilfs_super_block *nilfs_read_super_block(struct super_block *sb,
 	return (struct nilfs_super_block *)((char *)(*pbh)->b_data + offset);
 }
 
-int nilfs_store_magic_and_option(struct super_block *sb,
-				 struct nilfs_super_block *sbp,
-				 char *data)
+int nilfs_store_magic(struct super_block *sb,
+		      struct nilfs_super_block *sbp)
 {
 	struct the_nilfs *nilfs = sb->s_fs_info;
 
@@ -891,14 +861,12 @@ int nilfs_store_magic_and_option(struct super_block *sb,
 	sb->s_flags |= SB_NOATIME;
 #endif
 
-	nilfs_set_default_options(sb, sbp);
-
 	nilfs->ns_resuid = le16_to_cpu(sbp->s_def_resuid);
 	nilfs->ns_resgid = le16_to_cpu(sbp->s_def_resgid);
 	nilfs->ns_interval = le32_to_cpu(sbp->s_c_interval);
 	nilfs->ns_watermark = le32_to_cpu(sbp->s_c_block_max);
 
-	return !parse_options(data, sb, 0) ? -EINVAL : 0;
+	return 0;
 }
 
 int nilfs_check_feature_compatibility(struct super_block *sb,
@@ -1056,17 +1024,17 @@ int nilfs_checkpoint_is_mounted(struct super_block *sb, __u64 cno)
 /**
  * nilfs_fill_super() - initialize a super block instance
  * @sb: super_block
- * @data: mount options
- * @silent: silent mode flag
+ * @fc: filesystem context
  *
  * This function is called exclusively by nilfs->ns_mount_mutex.
  * So, the recovery process is protected from other simultaneous mounts.
  */
 static int
-nilfs_fill_super(struct super_block *sb, void *data, int silent)
+nilfs_fill_super(struct super_block *sb, struct fs_context *fc)
 {
 	struct the_nilfs *nilfs;
 	struct nilfs_root *fsroot;
+	struct nilfs_fs_context *ctx = fc->fs_private;
 	__u64 cno;
 	int err;
 
@@ -1076,10 +1044,13 @@ nilfs_fill_super(struct super_block *sb, void *data, int silent)
 
 	sb->s_fs_info = nilfs;
 
-	err = init_nilfs(nilfs, sb, (char *)data);
+	err = init_nilfs(nilfs, sb);
 	if (err)
 		goto failed_nilfs;
 
+	/* Copy in parsed mount options */
+	nilfs->ns_mount_opt = ctx->ns_mount_opt;
+
 	sb->s_op = &nilfs_sops;
 	sb->s_export_op = &nilfs_export_ops;
 	sb->s_root = NULL;
@@ -1138,34 +1109,25 @@ nilfs_fill_super(struct super_block *sb, void *data, int silent)
 	return err;
 }
 
-static int nilfs_remount(struct super_block *sb, int *flags, char *data)
+static int nilfs_reconfigure(struct fs_context *fc)
 {
+	struct nilfs_fs_context *ctx = fc->fs_private;
+	struct super_block *sb = fc->root->d_sb;
 	struct the_nilfs *nilfs = sb->s_fs_info;
-	unsigned long old_sb_flags;
-	unsigned long old_mount_opt;
 	int err;
 
 	sync_filesystem(sb);
-	old_sb_flags = sb->s_flags;
-	old_mount_opt = nilfs->ns_mount_opt;
-
-	if (!parse_options(data, sb, 1)) {
-		err = -EINVAL;
-		goto restore_opts;
-	}
-	sb->s_flags = (sb->s_flags & ~SB_POSIXACL);
 
 	err = -EINVAL;
 
 	if (!nilfs_valid_fs(nilfs)) {
 		nilfs_warn(sb,
 			   "couldn't remount because the filesystem is in an incomplete recovery state");
-		goto restore_opts;
+		goto ignore_opts;
 	}
-
-	if ((bool)(*flags & SB_RDONLY) == sb_rdonly(sb))
+	if ((bool)(fc->sb_flags & SB_RDONLY) == sb_rdonly(sb))
 		goto out;
-	if (*flags & SB_RDONLY) {
+	if (fc->sb_flags & SB_RDONLY) {
 		sb->s_flags |= SB_RDONLY;
 
 		/*
@@ -1193,138 +1155,67 @@ static int nilfs_remount(struct super_block *sb, int *flags, char *data)
 				   "couldn't remount RDWR because of unsupported optional features (%llx)",
 				   (unsigned long long)features);
 			err = -EROFS;
-			goto restore_opts;
+			goto ignore_opts;
 		}
 
 		sb->s_flags &= ~SB_RDONLY;
 
 		root = NILFS_I(d_inode(sb->s_root))->i_root;
 		err = nilfs_attach_log_writer(sb, root);
-		if (err)
-			goto restore_opts;
+		if (err) {
+			sb->s_flags |= SB_RDONLY;
+			goto ignore_opts;
+		}
 
 		down_write(&nilfs->ns_sem);
 		nilfs_setup_super(sb, true);
 		up_write(&nilfs->ns_sem);
 	}
  out:
-	return 0;
-
- restore_opts:
-	sb->s_flags = old_sb_flags;
-	nilfs->ns_mount_opt = old_mount_opt;
-	return err;
-}
-
-struct nilfs_super_data {
-	__u64 cno;
-	int flags;
-};
-
-static int nilfs_parse_snapshot_option(const char *option,
-				       const substring_t *arg,
-				       struct nilfs_super_data *sd)
-{
-	unsigned long long val;
-	const char *msg = NULL;
-	int err;
-
-	if (!(sd->flags & SB_RDONLY)) {
-		msg = "read-only option is not specified";
-		goto parse_error;
-	}
+	sb->s_flags = (sb->s_flags & ~SB_POSIXACL);
+	/* Copy over parsed remount options */
+	nilfs->ns_mount_opt = ctx->ns_mount_opt;
 
-	err = kstrtoull(arg->from, 0, &val);
-	if (err) {
-		if (err == -ERANGE)
-			msg = "too large checkpoint number";
-		else
-			msg = "malformed argument";
-		goto parse_error;
-	} else if (val == 0) {
-		msg = "invalid checkpoint number 0";
-		goto parse_error;
-	}
-	sd->cno = val;
 	return 0;
 
-parse_error:
-	nilfs_err(NULL, "invalid option \"%s\": %s", option, msg);
-	return 1;
-}
-
-/**
- * nilfs_identify - pre-read mount options needed to identify mount instance
- * @data: mount options
- * @sd: nilfs_super_data
- */
-static int nilfs_identify(char *data, struct nilfs_super_data *sd)
-{
-	char *p, *options = data;
-	substring_t args[MAX_OPT_ARGS];
-	int token;
-	int ret = 0;
-
-	do {
-		p = strsep(&options, ",");
-		if (p != NULL && *p) {
-			token = match_token(p, tokens, args);
-			if (token == Opt_snapshot)
-				ret = nilfs_parse_snapshot_option(p, &args[0],
-								  sd);
-		}
-		if (!options)
-			break;
-		BUG_ON(options == data);
-		*(options - 1) = ',';
-	} while (!ret);
-	return ret;
-}
-
-static int nilfs_set_bdev_super(struct super_block *s, void *data)
-{
-	s->s_dev = *(dev_t *)data;
-	return 0;
-}
-
-static int nilfs_test_bdev_super(struct super_block *s, void *data)
-{
-	return !(s->s_iflags & SB_I_RETIRED) && s->s_dev == *(dev_t *)data;
+ ignore_opts:
+	return err;
 }
 
-static struct dentry *
-nilfs_mount(struct file_system_type *fs_type, int flags,
-	     const char *dev_name, void *data)
+static int
+nilfs_get_tree(struct fs_context *fc)
 {
-	struct nilfs_super_data sd = { .flags = flags };
+	struct nilfs_fs_context *ctx = fc->fs_private;
 	struct super_block *s;
 	dev_t dev;
 	int err;
 
-	if (nilfs_identify(data, &sd))
-		return ERR_PTR(-EINVAL);
+	if (ctx->cno && !(fc->sb_flags & SB_RDONLY)) {
+		nilfs_err(NULL,
+			  "invalid option \"cp=%llu\": read-only option is not specified",
+			  ctx->cno);
+		return -EINVAL;
+	}
 
-	err = lookup_bdev(dev_name, &dev);
+	err = lookup_bdev(fc->source, &dev);
 	if (err)
-		return ERR_PTR(err);
+		return err;
 
-	s = sget(fs_type, nilfs_test_bdev_super, nilfs_set_bdev_super, flags,
-		 &dev);
+	s = sget_dev(fc, dev);
 	if (IS_ERR(s))
-		return ERR_CAST(s);
+		return PTR_ERR(s);
 
 	if (!s->s_root) {
-		err = setup_bdev_super(s, flags, NULL);
+		err = setup_bdev_super(s, fc->sb_flags, fc);
 		if (!err)
-			err = nilfs_fill_super(s, data,
-					       flags & SB_SILENT ? 1 : 0);
+			err = nilfs_fill_super(s, fc);
 		if (err)
 			goto failed_super;
 
 		s->s_flags |= SB_ACTIVE;
-	} else if (!sd.cno) {
+	} else if (!ctx->cno) {
 		if (nilfs_tree_is_busy(s->s_root)) {
-			if ((flags ^ s->s_flags) & SB_RDONLY) {
+			if ((fc->sb_flags ^ s->s_flags) & SB_RDONLY) {
 				nilfs_err(s,
 					  "the device already has a %s mount.",
 					  sb_rdonly(s) ? "read-only" : "read/write");
@@ -1333,37 +1224,75 @@ nilfs_mount(struct file_system_type *fs_type, int flags,
 			}
 		} else {
 			/*
-			 * Try remount to setup mount states if the current
+			 * Try reconfigure to setup mount states if the current
 			 * tree is not mounted and only snapshots use this sb.
+			 *
+			 * Since nilfs_reconfigure() requires fc->root to be
+			 * set, set it first and release it on failure.
 			 */
-			err = nilfs_remount(s, &flags, data);
-			if (err)
+			fc->root = dget(s->s_root);
+			err = nilfs_reconfigure(fc);
+			if (err) {
+				dput(fc->root);
+				fc->root = NULL;  /* prevent double release */
 				goto failed_super;
+			}
+			return 0;
 		}
 	}
 
-	if (sd.cno) {
+	if (ctx->cno) {
 		struct dentry *root_dentry;
 
-		err = nilfs_attach_snapshot(s, sd.cno, &root_dentry);
+		err = nilfs_attach_snapshot(s, ctx->cno, &root_dentry);
 		if (err)
 			goto failed_super;
-		return root_dentry;
+		fc->root = root_dentry;
+		return 0;
 	}
 
-	return dget(s->s_root);
+	fc->root = dget(s->s_root);
+	return 0;
 
  failed_super:
 	deactivate_locked_super(s);
-	return ERR_PTR(err);
+	return err;
+}
+
+static void nilfs_free_fc(struct fs_context *fc)
+{
+	kfree(fc->fs_private);
+}
+
+static const struct fs_context_operations nilfs_context_ops = {
+	.parse_param	= nilfs_parse_param,
+	.get_tree	= nilfs_get_tree,
+	.reconfigure	= nilfs_reconfigure,
+	.free		= nilfs_free_fc,
+};
+
+static int nilfs_init_fs_context(struct fs_context *fc)
+{
+	struct nilfs_fs_context *ctx;
+
+	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
+	if (!ctx)
+		return -ENOMEM;
+
+	ctx->ns_mount_opt = NILFS_MOUNT_ERRORS_RO | NILFS_MOUNT_BARRIER;
+	fc->fs_private = ctx;
+	fc->ops = &nilfs_context_ops;
+
+	return 0;
 }
 
 struct file_system_type nilfs_fs_type = {
 	.owner    = THIS_MODULE,
 	.name     = "nilfs2",
-	.mount    = nilfs_mount,
 	.kill_sb  = kill_block_super,
 	.fs_flags = FS_REQUIRES_DEV,
+	.init_fs_context = nilfs_init_fs_context,
+	.parameters = nilfs_param_spec,
 };
 MODULE_ALIAS_FS("nilfs2");
 
diff --git a/fs/nilfs2/the_nilfs.c b/fs/nilfs2/the_nilfs.c
index 71400496ed..e44dde57ab 100644
--- a/fs/nilfs2/the_nilfs.c
+++ b/fs/nilfs2/the_nilfs.c
@@ -413,7 +413,7 @@ static u64 nilfs_max_segment_count(struct the_nilfs *nilfs)
 {
 	u64 max_count = U64_MAX;
 
-	do_div(max_count, nilfs->ns_blocks_per_segment);
+	max_count = div64_ul(max_count, nilfs->ns_blocks_per_segment);
 	return min_t(u64, max_count, ULONG_MAX);
 }
 
@@ -452,6 +452,12 @@ static int nilfs_store_disk_layout(struct the_nilfs *nilfs,
 	}
 
 	nilfs->ns_first_ino = le32_to_cpu(sbp->s_first_ino);
+	if (nilfs->ns_first_ino < NILFS_USER_INO) {
+		nilfs_err(nilfs->ns_sb,
+			  "too small lower limit for non-reserved inode numbers: %u",
+			  nilfs->ns_first_ino);
+		return -EINVAL;
+	}
 
 	nilfs->ns_blocks_per_segment = le32_to_cpu(sbp->s_blocks_per_segment);
 	if (nilfs->ns_blocks_per_segment < NILFS_SEG_MIN_BLOCKS) {
@@ -592,7 +598,7 @@ static int nilfs_load_super_block(struct the_nilfs *nilfs,
 	struct nilfs_super_block **sbp = nilfs->ns_sbp;
 	struct buffer_head **sbh = nilfs->ns_sbh;
 	u64 sb2off, devsize = bdev_nr_bytes(nilfs->ns_bdev);
-	int valid[2], swp = 0;
+	int valid[2], swp = 0, older;
 
 	if (devsize < NILFS_SEG_MIN_BLOCKS * NILFS_MIN_BLOCK_SIZE + 4096) {
 		nilfs_err(sb, "device size too small");
@@ -648,9 +654,25 @@ static int nilfs_load_super_block(struct the_nilfs *nilfs,
 	if (swp)
 		nilfs_swap_super_block(nilfs);
 
+	/*
+	 * Calculate the array index of the older superblock data.
+	 * If one has been dropped, set index 0 pointing to the remaining one,
+	 * otherwise set index 1 pointing to the old one (including if both
+	 * are the same).
+	 *
+	 *  Divided case             valid[0]  valid[1]  swp  ->  older
+	 *  -------------------------------------------------------------
+	 *  Both SBs are invalid        0         0       N/A (Error)
+	 *  SB1 is invalid              0         1       1         0
+	 *  SB2 is invalid              1         0       0         0
+	 *  SB2 is newer                1         1       1         0
+	 *  SB2 is older or the same    1         1       0         1
+	 */
+	older = valid[1] ^ swp;
+
 	nilfs->ns_sbwcount = 0;
 	nilfs->ns_sbwtime = le64_to_cpu(sbp[0]->s_wtime);
-	nilfs->ns_prot_seq = le64_to_cpu(sbp[valid[1] & !swp]->s_last_seq);
+	nilfs->ns_prot_seq = le64_to_cpu(sbp[older]->s_last_seq);
 	*sbpp = sbp[0];
 	return 0;
 }
@@ -659,7 +681,6 @@ static int nilfs_load_super_block(struct the_nilfs *nilfs,
  * init_nilfs - initialize a NILFS instance.
  * @nilfs: the_nilfs structure
  * @sb: super block
- * @data: mount options
  *
  * init_nilfs() performs common initialization per block device (e.g.
  * reading the super block, getting disk layout information, initializing
@@ -668,7 +689,7 @@ static int nilfs_load_super_block(struct the_nilfs *nilfs,
  * Return Value: On success, 0 is returned. On error, a negative error
  * code is returned.
  */
-int init_nilfs(struct the_nilfs *nilfs, struct super_block *sb, char *data)
+int init_nilfs(struct the_nilfs *nilfs, struct super_block *sb)
 {
 	struct nilfs_super_block *sbp;
 	int blocksize;
@@ -686,7 +707,7 @@ int init_nilfs(struct the_nilfs *nilfs, struct super_block *sb, char *data)
 	if (err)
 		goto out;
 
-	err = nilfs_store_magic_and_option(sb, sbp, data);
+	err = nilfs_store_magic(sb, sbp);
 	if (err)
 		goto failed_sbh;
 
diff --git a/fs/nilfs2/the_nilfs.h b/fs/nilfs2/the_nilfs.h
index cd4ae1b8ae..1e829ed7b0 100644
--- a/fs/nilfs2/the_nilfs.h
+++ b/fs/nilfs2/the_nilfs.h
@@ -182,7 +182,7 @@ struct the_nilfs {
 	unsigned long		ns_nrsvsegs;
 	unsigned long		ns_first_data_block;
 	int			ns_inode_size;
-	int			ns_first_ino;
+	unsigned int		ns_first_ino;
 	u32			ns_crc_seed;
 
 	/* /sys/fs/<nilfs>/<device> */
@@ -219,10 +219,6 @@ THE_NILFS_FNS(PURGING, purging)
 #define nilfs_set_opt(nilfs, opt)  \
 	((nilfs)->ns_mount_opt |= NILFS_MOUNT_##opt)
 #define nilfs_test_opt(nilfs, opt) ((nilfs)->ns_mount_opt & NILFS_MOUNT_##opt)
-#define nilfs_write_opt(nilfs, mask, opt)				\
-	((nilfs)->ns_mount_opt =					\
-		(((nilfs)->ns_mount_opt & ~NILFS_MOUNT_##mask) |	\
-		 NILFS_MOUNT_##opt))					\
 
 /**
  * struct nilfs_root - nilfs root object
@@ -276,7 +272,7 @@ static inline int nilfs_sb_will_flip(struct the_nilfs *nilfs)
 void nilfs_set_last_segment(struct the_nilfs *, sector_t, u64, __u64);
 struct the_nilfs *alloc_nilfs(struct super_block *sb);
 void destroy_nilfs(struct the_nilfs *nilfs);
-int init_nilfs(struct the_nilfs *nilfs, struct super_block *sb, char *data);
+int init_nilfs(struct the_nilfs *nilfs, struct super_block *sb);
 int load_nilfs(struct the_nilfs *nilfs, struct super_block *sb);
 unsigned long nilfs_nrsvsegs(struct the_nilfs *nilfs, unsigned long nsegs);
 void nilfs_set_nsegments(struct the_nilfs *nilfs, unsigned long nsegs);
diff --git a/fs/notify/dnotify/dnotify.c b/fs/notify/dnotify/dnotify.c
index 3464fa7e85..f3669403fa 100644
--- a/fs/notify/dnotify/dnotify.c
+++ b/fs/notify/dnotify/dnotify.c
@@ -162,7 +162,7 @@ void dnotify_flush(struct file *filp, fl_owner_t id)
 	if (!S_ISDIR(inode->i_mode))
 		return;
 
-	fsn_mark = fsnotify_find_mark(&inode->i_fsnotify_marks, dnotify_group);
+	fsn_mark = fsnotify_find_inode_mark(inode, dnotify_group);
 	if (!fsn_mark)
 		return;
 	dn_mark = container_of(fsn_mark, struct dnotify_mark, fsn_mark);
@@ -326,7 +326,7 @@ int fcntl_dirnotify(int fd, struct file *filp, unsigned int arg)
 	fsnotify_group_lock(dnotify_group);
 
 	/* add the new_fsn_mark or find an old one. */
-	fsn_mark = fsnotify_find_mark(&inode->i_fsnotify_marks, dnotify_group);
+	fsn_mark = fsnotify_find_inode_mark(inode, dnotify_group);
 	if (fsn_mark) {
 		dn_mark = container_of(fsn_mark, struct dnotify_mark, fsn_mark);
 		spin_lock(&fsn_mark->lock);
diff --git a/fs/notify/fanotify/fanotify.c b/fs/notify/fanotify/fanotify.c
index 1e4def2181..224bccaab4 100644
--- a/fs/notify/fanotify/fanotify.c
+++ b/fs/notify/fanotify/fanotify.c
@@ -228,8 +228,10 @@ static int fanotify_get_response(struct fsnotify_group *group,
 
 	pr_debug("%s: group=%p event=%p\n", __func__, group, event);
 
-	ret = wait_event_killable(group->fanotify_data.access_waitq,
-				  event->state == FAN_EVENT_ANSWERED);
+	ret = wait_event_state(group->fanotify_data.access_waitq,
+				  event->state == FAN_EVENT_ANSWERED,
+				  (TASK_KILLABLE|TASK_FREEZABLE));
+
 	/* Signal pending? */
 	if (ret < 0) {
 		spin_lock(&group->notification_lock);
diff --git a/fs/notify/fanotify/fanotify_user.c b/fs/notify/fanotify/fanotify_user.c
index fbdc63cc10..9ec313e9f6 100644
--- a/fs/notify/fanotify/fanotify_user.c
+++ b/fs/notify/fanotify/fanotify_user.c
@@ -1076,7 +1076,7 @@ static __u32 fanotify_mark_remove_from_mask(struct fsnotify_mark *fsn_mark,
 }
 
 static int fanotify_remove_mark(struct fsnotify_group *group,
-				fsnotify_connp_t *connp, __u32 mask,
+				void *obj, unsigned int obj_type, __u32 mask,
 				unsigned int flags, __u32 umask)
 {
 	struct fsnotify_mark *fsn_mark = NULL;
@@ -1084,7 +1084,7 @@ static int fanotify_remove_mark(struct fsnotify_group *group,
 	int destroy_mark;
 
 	fsnotify_group_lock(group);
-	fsn_mark = fsnotify_find_mark(connp, group);
+	fsn_mark = fsnotify_find_mark(obj, obj_type, group);
 	if (!fsn_mark) {
 		fsnotify_group_unlock(group);
 		return -ENOENT;
@@ -1105,30 +1105,6 @@ static int fanotify_remove_mark(struct fsnotify_group *group,
 	return 0;
 }
 
-static int fanotify_remove_vfsmount_mark(struct fsnotify_group *group,
-					 struct vfsmount *mnt, __u32 mask,
-					 unsigned int flags, __u32 umask)
-{
-	return fanotify_remove_mark(group, &real_mount(mnt)->mnt_fsnotify_marks,
-				    mask, flags, umask);
-}
-
-static int fanotify_remove_sb_mark(struct fsnotify_group *group,
-				   struct super_block *sb, __u32 mask,
-				   unsigned int flags, __u32 umask)
-{
-	return fanotify_remove_mark(group, &sb->s_fsnotify_marks, mask,
-				    flags, umask);
-}
-
-static int fanotify_remove_inode_mark(struct fsnotify_group *group,
-				      struct inode *inode, __u32 mask,
-				      unsigned int flags, __u32 umask)
-{
-	return fanotify_remove_mark(group, &inode->i_fsnotify_marks, mask,
-				    flags, umask);
-}
-
 static bool fanotify_mark_update_flags(struct fsnotify_mark *fsn_mark,
 				       unsigned int fan_flags)
 {
@@ -1249,7 +1225,7 @@ static int fanotify_set_mark_fsid(struct fsnotify_group *group,
 }
 
 static struct fsnotify_mark *fanotify_add_new_mark(struct fsnotify_group *group,
-						   fsnotify_connp_t *connp,
+						   void *obj,
 						   unsigned int obj_type,
 						   unsigned int fan_flags,
 						   struct fan_fsid *fsid)
@@ -1288,7 +1264,7 @@ static struct fsnotify_mark *fanotify_add_new_mark(struct fsnotify_group *group,
 		fan_mark->fsid.val[0] = fan_mark->fsid.val[1] = 0;
 	}
 
-	ret = fsnotify_add_mark_locked(mark, connp, obj_type, 0);
+	ret = fsnotify_add_mark_locked(mark, obj, obj_type, 0);
 	if (ret)
 		goto out_put_mark;
 
@@ -1344,7 +1320,7 @@ static int fanotify_may_update_existing_mark(struct fsnotify_mark *fsn_mark,
 }
 
 static int fanotify_add_mark(struct fsnotify_group *group,
-			     fsnotify_connp_t *connp, unsigned int obj_type,
+			     void *obj, unsigned int obj_type,
 			     __u32 mask, unsigned int fan_flags,
 			     struct fan_fsid *fsid)
 {
@@ -1353,9 +1329,9 @@ static int fanotify_add_mark(struct fsnotify_group *group,
 	int ret = 0;
 
 	fsnotify_group_lock(group);
-	fsn_mark = fsnotify_find_mark(connp, group);
+	fsn_mark = fsnotify_find_mark(obj, obj_type, group);
 	if (!fsn_mark) {
-		fsn_mark = fanotify_add_new_mark(group, connp, obj_type,
+		fsn_mark = fanotify_add_new_mark(group, obj, obj_type,
 						 fan_flags, fsid);
 		if (IS_ERR(fsn_mark)) {
 			fsnotify_group_unlock(group);
@@ -1392,42 +1368,6 @@ out:
 	return ret;
 }
 
-static int fanotify_add_vfsmount_mark(struct fsnotify_group *group,
-				      struct vfsmount *mnt, __u32 mask,
-				      unsigned int flags, struct fan_fsid *fsid)
-{
-	return fanotify_add_mark(group, &real_mount(mnt)->mnt_fsnotify_marks,
-				 FSNOTIFY_OBJ_TYPE_VFSMOUNT, mask, flags, fsid);
-}
-
-static int fanotify_add_sb_mark(struct fsnotify_group *group,
-				struct super_block *sb, __u32 mask,
-				unsigned int flags, struct fan_fsid *fsid)
-{
-	return fanotify_add_mark(group, &sb->s_fsnotify_marks,
-				 FSNOTIFY_OBJ_TYPE_SB, mask, flags, fsid);
-}
-
-static int fanotify_add_inode_mark(struct fsnotify_group *group,
-				   struct inode *inode, __u32 mask,
-				   unsigned int flags, struct fan_fsid *fsid)
-{
-	pr_debug("%s: group=%p inode=%p\n", __func__, group, inode);
-
-	/*
-	 * If some other task has this inode open for write we should not add
-	 * an ignore mask, unless that ignore mask is supposed to survive
-	 * modification changes anyway.
-	 */
-	if ((flags & FANOTIFY_MARK_IGNORE_BITS) &&
-	    !(flags & FAN_MARK_IGNORED_SURV_MODIFY) &&
-	    inode_is_open_for_write(inode))
-		return 0;
-
-	return fanotify_add_mark(group, &inode->i_fsnotify_marks,
-				 FSNOTIFY_OBJ_TYPE_INODE, mask, flags, fsid);
-}
-
 static struct fsnotify_event *fanotify_alloc_overflow_event(void)
 {
 	struct fanotify_event *oevent;
@@ -1576,13 +1516,13 @@ SYSCALL_DEFINE2(fanotify_init, unsigned int, flags, unsigned int, event_f_flags)
 	INIT_LIST_HEAD(&group->fanotify_data.access_list);
 	switch (class) {
 	case FAN_CLASS_NOTIF:
-		group->priority = FS_PRIO_0;
+		group->priority = FSNOTIFY_PRIO_NORMAL;
 		break;
 	case FAN_CLASS_CONTENT:
-		group->priority = FS_PRIO_1;
+		group->priority = FSNOTIFY_PRIO_CONTENT;
 		break;
 	case FAN_CLASS_PRE_CONTENT:
-		group->priority = FS_PRIO_2;
+		group->priority = FSNOTIFY_PRIO_PRE_CONTENT;
 		break;
 	default:
 		fd = -EINVAL;
@@ -1750,6 +1690,7 @@ static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 	unsigned int mark_cmd = flags & FANOTIFY_MARK_CMD_BITS;
 	unsigned int ignore = flags & FANOTIFY_MARK_IGNORE_BITS;
 	unsigned int obj_type, fid_mode;
+	void *obj;
 	u32 umask = 0;
 	int ret;
 
@@ -1833,12 +1774,11 @@ static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 		goto fput_and_out;
 
 	/*
-	 * group->priority == FS_PRIO_0 == FAN_CLASS_NOTIF.  These are not
-	 * allowed to set permissions events.
+	 * Permission events require minimum priority FAN_CLASS_CONTENT.
 	 */
 	ret = -EINVAL;
 	if (mask & FANOTIFY_PERM_EVENTS &&
-	    group->priority == FS_PRIO_0)
+	    group->priority < FSNOTIFY_PRIO_CONTENT)
 		goto fput_and_out;
 
 	if (mask & FAN_FS_ERROR &&
@@ -1908,17 +1848,34 @@ static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 	}
 
 	/* inode held in place by reference to path; group by fget on fd */
-	if (mark_type == FAN_MARK_INODE)
+	if (mark_type == FAN_MARK_INODE) {
 		inode = path.dentry->d_inode;
-	else
+		obj = inode;
+	} else {
 		mnt = path.mnt;
+		if (mark_type == FAN_MARK_MOUNT)
+			obj = mnt;
+		else
+			obj = mnt->mnt_sb;
+	}
 
-	ret = mnt ? -EINVAL : -EISDIR;
-	/* FAN_MARK_IGNORE requires SURV_MODIFY for sb/mount/dir marks */
-	if (mark_cmd == FAN_MARK_ADD && ignore == FAN_MARK_IGNORE &&
-	    (mnt || S_ISDIR(inode->i_mode)) &&
-	    !(flags & FAN_MARK_IGNORED_SURV_MODIFY))
-		goto path_put_and_out;
+	/*
+	 * If some other task has this inode open for write we should not add
+	 * an ignore mask, unless that ignore mask is supposed to survive
+	 * modification changes anyway.
+	 */
+	if (mark_cmd == FAN_MARK_ADD && (flags & FANOTIFY_MARK_IGNORE_BITS) &&
+	    !(flags & FAN_MARK_IGNORED_SURV_MODIFY)) {
+		ret = mnt ? -EINVAL : -EISDIR;
+		/* FAN_MARK_IGNORE requires SURV_MODIFY for sb/mount/dir marks */
+		if (ignore == FAN_MARK_IGNORE &&
+		    (mnt || S_ISDIR(inode->i_mode)))
+			goto path_put_and_out;
+
+		ret = 0;
+		if (inode && inode_is_open_for_write(inode))
+			goto path_put_and_out;
+	}
 
 	/* Mask out FAN_EVENT_ON_CHILD flag for sb/mount/non-dir marks */
 	if (mnt || !S_ISDIR(inode->i_mode)) {
@@ -1936,26 +1893,12 @@ static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 	/* create/update an inode mark */
 	switch (mark_cmd) {
 	case FAN_MARK_ADD:
-		if (mark_type == FAN_MARK_MOUNT)
-			ret = fanotify_add_vfsmount_mark(group, mnt, mask,
-							 flags, fsid);
-		else if (mark_type == FAN_MARK_FILESYSTEM)
-			ret = fanotify_add_sb_mark(group, mnt->mnt_sb, mask,
-						   flags, fsid);
-		else
-			ret = fanotify_add_inode_mark(group, inode, mask,
-						      flags, fsid);
+		ret = fanotify_add_mark(group, obj, obj_type, mask, flags,
+					fsid);
 		break;
 	case FAN_MARK_REMOVE:
-		if (mark_type == FAN_MARK_MOUNT)
-			ret = fanotify_remove_vfsmount_mark(group, mnt, mask,
-							    flags, umask);
-		else if (mark_type == FAN_MARK_FILESYSTEM)
-			ret = fanotify_remove_sb_mark(group, mnt->mnt_sb, mask,
-						      flags, umask);
-		else
-			ret = fanotify_remove_inode_mark(group, inode, mask,
-							 flags, umask);
+		ret = fanotify_remove_mark(group, obj, obj_type, mask, flags,
+					   umask);
 		break;
 	default:
 		ret = -EINVAL;
diff --git a/fs/notify/fdinfo.c b/fs/notify/fdinfo.c
index 5c430736ec..dec5530340 100644
--- a/fs/notify/fdinfo.c
+++ b/fs/notify/fdinfo.c
@@ -41,29 +41,25 @@ static void show_fdinfo(struct seq_file *m, struct file *f,
 #if defined(CONFIG_EXPORTFS)
 static void show_mark_fhandle(struct seq_file *m, struct inode *inode)
 {
-	struct {
-		struct file_handle handle;
-		u8 pad[MAX_HANDLE_SZ];
-	} f;
+	DEFINE_FLEX(struct file_handle, f, f_handle, handle_bytes, MAX_HANDLE_SZ);
 	int size, ret, i;
 
-	f.handle.handle_bytes = sizeof(f.pad);
-	size = f.handle.handle_bytes >> 2;
+	size = f->handle_bytes >> 2;
 
-	ret = exportfs_encode_fid(inode, (struct fid *)f.handle.f_handle, &size);
+	ret = exportfs_encode_fid(inode, (struct fid *)f->f_handle, &size);
 	if ((ret == FILEID_INVALID) || (ret < 0)) {
 		WARN_ONCE(1, "Can't encode file handler for inotify: %d\n", ret);
 		return;
 	}
 
-	f.handle.handle_type = ret;
-	f.handle.handle_bytes = size * sizeof(u32);
+	f->handle_type = ret;
+	f->handle_bytes = size * sizeof(u32);
 
 	seq_printf(m, "fhandle-bytes:%x fhandle-type:%x f_handle:",
-		   f.handle.handle_bytes, f.handle.handle_type);
+		   f->handle_bytes, f->handle_type);
 
-	for (i = 0; i < f.handle.handle_bytes; i++)
-		seq_printf(m, "%02x", (int)f.handle.f_handle[i]);
+	for (i = 0; i < f->handle_bytes; i++)
+		seq_printf(m, "%02x", (int)f->f_handle[i]);
 }
 #else
 static void show_mark_fhandle(struct seq_file *m, struct inode *inode)
diff --git a/fs/notify/fsnotify.c b/fs/notify/fsnotify.c
index 8bfd690e9f..ff69ae24c4 100644
--- a/fs/notify/fsnotify.c
+++ b/fs/notify/fsnotify.c
@@ -89,11 +89,25 @@ static void fsnotify_unmount_inodes(struct super_block *sb)
 
 void fsnotify_sb_delete(struct super_block *sb)
 {
+	struct fsnotify_sb_info *sbinfo = fsnotify_sb_info(sb);
+
+	/* Were any marks ever added to any object on this sb? */
+	if (!sbinfo)
+		return;
+
 	fsnotify_unmount_inodes(sb);
 	fsnotify_clear_marks_by_sb(sb);
 	/* Wait for outstanding object references from connectors */
-	wait_var_event(&sb->s_fsnotify_connectors,
-		       !atomic_long_read(&sb->s_fsnotify_connectors));
+	wait_var_event(fsnotify_sb_watched_objects(sb),
+		       !atomic_long_read(fsnotify_sb_watched_objects(sb)));
+	WARN_ON(fsnotify_sb_has_priority_watchers(sb, FSNOTIFY_PRIO_CONTENT));
+	WARN_ON(fsnotify_sb_has_priority_watchers(sb,
+						  FSNOTIFY_PRIO_PRE_CONTENT));
+}
+
+void fsnotify_sb_free(struct super_block *sb)
+{
+	kfree(sb->s_fsnotify_info);
 }
 
 /*
@@ -141,7 +155,7 @@ void __fsnotify_update_child_dentry_flags(struct inode *inode)
 }
 
 /* Are inode/sb/mount interested in parent and name info with this event? */
-static bool fsnotify_event_needs_parent(struct inode *inode, struct mount *mnt,
+static bool fsnotify_event_needs_parent(struct inode *inode, __u32 mnt_mask,
 					__u32 mask)
 {
 	__u32 marks_mask = 0;
@@ -160,13 +174,22 @@ static bool fsnotify_event_needs_parent(struct inode *inode, struct mount *mnt,
 	/* Did either inode/sb/mount subscribe for events with parent/name? */
 	marks_mask |= fsnotify_parent_needed_mask(inode->i_fsnotify_mask);
 	marks_mask |= fsnotify_parent_needed_mask(inode->i_sb->s_fsnotify_mask);
-	if (mnt)
-		marks_mask |= fsnotify_parent_needed_mask(mnt->mnt_fsnotify_mask);
+	marks_mask |= fsnotify_parent_needed_mask(mnt_mask);
 
 	/* Did they subscribe for this event with parent/name info? */
 	return mask & marks_mask;
 }
 
+/* Are there any inode/mount/sb objects that are interested in this event? */
+static inline bool fsnotify_object_watched(struct inode *inode, __u32 mnt_mask,
+					   __u32 mask)
+{
+	__u32 marks_mask = inode->i_fsnotify_mask | mnt_mask |
+			   inode->i_sb->s_fsnotify_mask;
+
+	return mask & marks_mask & ALL_FSNOTIFY_EVENTS;
+}
+
 /*
  * Notify this dentry's parent about a child's events with child name info
  * if parent is watching or if inode/sb/mount are interested in events with
@@ -179,7 +202,7 @@ int __fsnotify_parent(struct dentry *dentry, __u32 mask, const void *data,
 		      int data_type)
 {
 	const struct path *path = fsnotify_data_path(data, data_type);
-	struct mount *mnt = path ? real_mount(path->mnt) : NULL;
+	__u32 mnt_mask = path ? real_mount(path->mnt)->mnt_fsnotify_mask : 0;
 	struct inode *inode = d_inode(dentry);
 	struct dentry *parent;
 	bool parent_watched = dentry->d_flags & DCACHE_FSNOTIFY_PARENT_WATCHED;
@@ -190,16 +213,13 @@ int __fsnotify_parent(struct dentry *dentry, __u32 mask, const void *data,
 	struct qstr *file_name = NULL;
 	int ret = 0;
 
-	/*
-	 * Do inode/sb/mount care about parent and name info on non-dir?
-	 * Do they care about any event at all?
-	 */
-	if (!inode->i_fsnotify_marks && !inode->i_sb->s_fsnotify_marks &&
-	    (!mnt || !mnt->mnt_fsnotify_marks) && !parent_watched)
+	/* Optimize the likely case of nobody watching this path */
+	if (likely(!parent_watched &&
+		   !fsnotify_object_watched(inode, mnt_mask, mask)))
 		return 0;
 
 	parent = NULL;
-	parent_needed = fsnotify_event_needs_parent(inode, mnt, mask);
+	parent_needed = fsnotify_event_needs_parent(inode, mnt_mask, mask);
 	if (!parent_watched && !parent_needed)
 		goto notify;
 
@@ -483,6 +503,7 @@ int fsnotify(__u32 mask, const void *data, int data_type, struct inode *dir,
 {
 	const struct path *path = fsnotify_data_path(data, data_type);
 	struct super_block *sb = fsnotify_data_sb(data, data_type);
+	struct fsnotify_sb_info *sbinfo = fsnotify_sb_info(sb);
 	struct fsnotify_iter_info iter_info = {};
 	struct mount *mnt = NULL;
 	struct inode *inode2 = NULL;
@@ -519,7 +540,7 @@ int fsnotify(__u32 mask, const void *data, int data_type, struct inode *dir,
 	 * SRCU because we have no references to any objects and do not
 	 * need SRCU to keep them "alive".
 	 */
-	if (!sb->s_fsnotify_marks &&
+	if ((!sbinfo || !sbinfo->sb_marks) &&
 	    (!mnt || !mnt->mnt_fsnotify_marks) &&
 	    (!inode || !inode->i_fsnotify_marks) &&
 	    (!inode2 || !inode2->i_fsnotify_marks))
@@ -546,8 +567,10 @@ int fsnotify(__u32 mask, const void *data, int data_type, struct inode *dir,
 
 	iter_info.srcu_idx = srcu_read_lock(&fsnotify_mark_srcu);
 
-	iter_info.marks[FSNOTIFY_ITER_TYPE_SB] =
-		fsnotify_first_mark(&sb->s_fsnotify_marks);
+	if (sbinfo) {
+		iter_info.marks[FSNOTIFY_ITER_TYPE_SB] =
+			fsnotify_first_mark(&sbinfo->sb_marks);
+	}
 	if (mnt) {
 		iter_info.marks[FSNOTIFY_ITER_TYPE_VFSMOUNT] =
 			fsnotify_first_mark(&mnt->mnt_fsnotify_marks);
diff --git a/fs/notify/fsnotify.h b/fs/notify/fsnotify.h
index fde74eb333..2d059f789e 100644
--- a/fs/notify/fsnotify.h
+++ b/fs/notify/fsnotify.h
@@ -9,39 +9,58 @@
 
 #include "../mount.h"
 
+/*
+ * fsnotify_connp_t is what we embed in objects which connector can be attached
+ * to.
+ */
+typedef struct fsnotify_mark_connector __rcu *fsnotify_connp_t;
+
 static inline struct inode *fsnotify_conn_inode(
 				struct fsnotify_mark_connector *conn)
 {
-	return container_of(conn->obj, struct inode, i_fsnotify_marks);
+	return conn->obj;
 }
 
 static inline struct mount *fsnotify_conn_mount(
 				struct fsnotify_mark_connector *conn)
 {
-	return container_of(conn->obj, struct mount, mnt_fsnotify_marks);
+	return real_mount(conn->obj);
 }
 
 static inline struct super_block *fsnotify_conn_sb(
 				struct fsnotify_mark_connector *conn)
 {
-	return container_of(conn->obj, struct super_block, s_fsnotify_marks);
+	return conn->obj;
 }
 
-static inline struct super_block *fsnotify_connector_sb(
-				struct fsnotify_mark_connector *conn)
+static inline struct super_block *fsnotify_object_sb(void *obj,
+			enum fsnotify_obj_type obj_type)
 {
-	switch (conn->type) {
+	switch (obj_type) {
 	case FSNOTIFY_OBJ_TYPE_INODE:
-		return fsnotify_conn_inode(conn)->i_sb;
+		return ((struct inode *)obj)->i_sb;
 	case FSNOTIFY_OBJ_TYPE_VFSMOUNT:
-		return fsnotify_conn_mount(conn)->mnt.mnt_sb;
+		return ((struct vfsmount *)obj)->mnt_sb;
 	case FSNOTIFY_OBJ_TYPE_SB:
-		return fsnotify_conn_sb(conn);
+		return (struct super_block *)obj;
 	default:
 		return NULL;
 	}
 }
 
+static inline struct super_block *fsnotify_connector_sb(
+				struct fsnotify_mark_connector *conn)
+{
+	return fsnotify_object_sb(conn->obj, conn->type);
+}
+
+static inline fsnotify_connp_t *fsnotify_sb_marks(struct super_block *sb)
+{
+	struct fsnotify_sb_info *sbinfo = fsnotify_sb_info(sb);
+
+	return sbinfo ? &sbinfo->sb_marks : NULL;
+}
+
 /* destroy all events sitting in this groups notification queue */
 extern void fsnotify_flush_notify(struct fsnotify_group *group);
 
@@ -67,7 +86,7 @@ static inline void fsnotify_clear_marks_by_mount(struct vfsmount *mnt)
 /* run the list of all marks associated with sb and destroy them */
 static inline void fsnotify_clear_marks_by_sb(struct super_block *sb)
 {
-	fsnotify_destroy_marks(&sb->s_fsnotify_marks);
+	fsnotify_destroy_marks(fsnotify_sb_marks(sb));
 }
 
 /*
diff --git a/fs/notify/inotify/inotify_user.c b/fs/notify/inotify/inotify_user.c
index 85d8fdd553..4ffc30606e 100644
--- a/fs/notify/inotify/inotify_user.c
+++ b/fs/notify/inotify/inotify_user.c
@@ -544,7 +544,7 @@ static int inotify_update_existing_watch(struct fsnotify_group *group,
 	int create = (arg & IN_MASK_CREATE);
 	int ret;
 
-	fsn_mark = fsnotify_find_mark(&inode->i_fsnotify_marks, group);
+	fsn_mark = fsnotify_find_inode_mark(inode, group);
 	if (!fsn_mark)
 		return -ENOENT;
 	else if (create) {
diff --git a/fs/notify/mark.c b/fs/notify/mark.c
index d6944ff86f..c3eefa7063 100644
--- a/fs/notify/mark.c
+++ b/fs/notify/mark.c
@@ -97,6 +97,21 @@ void fsnotify_get_mark(struct fsnotify_mark *mark)
 	refcount_inc(&mark->refcnt);
 }
 
+static fsnotify_connp_t *fsnotify_object_connp(void *obj,
+				enum fsnotify_obj_type obj_type)
+{
+	switch (obj_type) {
+	case FSNOTIFY_OBJ_TYPE_INODE:
+		return &((struct inode *)obj)->i_fsnotify_marks;
+	case FSNOTIFY_OBJ_TYPE_VFSMOUNT:
+		return &real_mount(obj)->mnt_fsnotify_marks;
+	case FSNOTIFY_OBJ_TYPE_SB:
+		return fsnotify_sb_marks(obj);
+	default:
+		return NULL;
+	}
+}
+
 static __u32 *fsnotify_conn_mask_p(struct fsnotify_mark_connector *conn)
 {
 	if (conn->type == FSNOTIFY_OBJ_TYPE_INODE)
@@ -116,10 +131,69 @@ __u32 fsnotify_conn_mask(struct fsnotify_mark_connector *conn)
 	return *fsnotify_conn_mask_p(conn);
 }
 
+static void fsnotify_get_sb_watched_objects(struct super_block *sb)
+{
+	atomic_long_inc(fsnotify_sb_watched_objects(sb));
+}
+
+static void fsnotify_put_sb_watched_objects(struct super_block *sb)
+{
+	if (atomic_long_dec_and_test(fsnotify_sb_watched_objects(sb)))
+		wake_up_var(fsnotify_sb_watched_objects(sb));
+}
+
 static void fsnotify_get_inode_ref(struct inode *inode)
 {
 	ihold(inode);
-	atomic_long_inc(&inode->i_sb->s_fsnotify_connectors);
+	fsnotify_get_sb_watched_objects(inode->i_sb);
+}
+
+static void fsnotify_put_inode_ref(struct inode *inode)
+{
+	fsnotify_put_sb_watched_objects(inode->i_sb);
+	iput(inode);
+}
+
+/*
+ * Grab or drop watched objects reference depending on whether the connector
+ * is attached and has any marks attached.
+ */
+static void fsnotify_update_sb_watchers(struct super_block *sb,
+					struct fsnotify_mark_connector *conn)
+{
+	struct fsnotify_sb_info *sbinfo = fsnotify_sb_info(sb);
+	bool is_watched = conn->flags & FSNOTIFY_CONN_FLAG_IS_WATCHED;
+	struct fsnotify_mark *first_mark = NULL;
+	unsigned int highest_prio = 0;
+
+	if (conn->obj)
+		first_mark = hlist_entry_safe(conn->list.first,
+					      struct fsnotify_mark, obj_list);
+	if (first_mark)
+		highest_prio = first_mark->group->priority;
+	if (WARN_ON(highest_prio >= __FSNOTIFY_PRIO_NUM))
+		highest_prio = 0;
+
+	/*
+	 * If the highest priority of group watching this object is prio,
+	 * then watched object has a reference on counters [0..prio].
+	 * Update priority >= 1 watched objects counters.
+	 */
+	for (unsigned int p = conn->prio + 1; p <= highest_prio; p++)
+		atomic_long_inc(&sbinfo->watched_objects[p]);
+	for (unsigned int p = conn->prio; p > highest_prio; p--)
+		atomic_long_dec(&sbinfo->watched_objects[p]);
+	conn->prio = highest_prio;
+
+	/* Update priority >= 0 (a.k.a total) watched objects counter */
+	BUILD_BUG_ON(FSNOTIFY_PRIO_NORMAL != 0);
+	if (first_mark && !is_watched) {
+		conn->flags |= FSNOTIFY_CONN_FLAG_IS_WATCHED;
+		fsnotify_get_sb_watched_objects(sb);
+	} else if (!first_mark && is_watched) {
+		conn->flags &= ~FSNOTIFY_CONN_FLAG_IS_WATCHED;
+		fsnotify_put_sb_watched_objects(sb);
+	}
 }
 
 /*
@@ -213,35 +287,12 @@ static void fsnotify_connector_destroy_workfn(struct work_struct *work)
 	}
 }
 
-static void fsnotify_put_inode_ref(struct inode *inode)
-{
-	struct super_block *sb = inode->i_sb;
-
-	iput(inode);
-	if (atomic_long_dec_and_test(&sb->s_fsnotify_connectors))
-		wake_up_var(&sb->s_fsnotify_connectors);
-}
-
-static void fsnotify_get_sb_connectors(struct fsnotify_mark_connector *conn)
-{
-	struct super_block *sb = fsnotify_connector_sb(conn);
-
-	if (sb)
-		atomic_long_inc(&sb->s_fsnotify_connectors);
-}
-
-static void fsnotify_put_sb_connectors(struct fsnotify_mark_connector *conn)
-{
-	struct super_block *sb = fsnotify_connector_sb(conn);
-
-	if (sb && atomic_long_dec_and_test(&sb->s_fsnotify_connectors))
-		wake_up_var(&sb->s_fsnotify_connectors);
-}
-
 static void *fsnotify_detach_connector_from_object(
 					struct fsnotify_mark_connector *conn,
 					unsigned int *type)
 {
+	fsnotify_connp_t *connp = fsnotify_object_connp(conn->obj, conn->type);
+	struct super_block *sb = fsnotify_connector_sb(conn);
 	struct inode *inode = NULL;
 
 	*type = conn->type;
@@ -261,10 +312,10 @@ static void *fsnotify_detach_connector_from_object(
 		fsnotify_conn_sb(conn)->s_fsnotify_mask = 0;
 	}
 
-	fsnotify_put_sb_connectors(conn);
-	rcu_assign_pointer(*(conn->obj), NULL);
+	rcu_assign_pointer(*connp, NULL);
 	conn->obj = NULL;
 	conn->type = FSNOTIFY_OBJ_TYPE_DETACHED;
+	fsnotify_update_sb_watchers(sb, conn);
 
 	return inode;
 }
@@ -316,6 +367,11 @@ void fsnotify_put_mark(struct fsnotify_mark *mark)
 		objp = fsnotify_detach_connector_from_object(conn, &type);
 		free_conn = true;
 	} else {
+		struct super_block *sb = fsnotify_connector_sb(conn);
+
+		/* Update watched objects after detaching mark */
+		if (sb)
+			fsnotify_update_sb_watchers(sb, conn);
 		objp = __fsnotify_recalc_mask(conn);
 		type = conn->type;
 	}
@@ -536,8 +592,28 @@ int fsnotify_compare_groups(struct fsnotify_group *a, struct fsnotify_group *b)
 	return -1;
 }
 
+static int fsnotify_attach_info_to_sb(struct super_block *sb)
+{
+	struct fsnotify_sb_info *sbinfo;
+
+	/* sb info is freed on fsnotify_sb_delete() */
+	sbinfo = kzalloc(sizeof(*sbinfo), GFP_KERNEL);
+	if (!sbinfo)
+		return -ENOMEM;
+
+	/*
+	 * cmpxchg() provides the barrier so that callers of fsnotify_sb_info()
+	 * will observe an initialized structure
+	 */
+	if (cmpxchg(&sb->s_fsnotify_info, NULL, sbinfo)) {
+		/* Someone else created sbinfo for us */
+		kfree(sbinfo);
+	}
+	return 0;
+}
+
 static int fsnotify_attach_connector_to_object(fsnotify_connp_t *connp,
-					       unsigned int obj_type)
+					       void *obj, unsigned int obj_type)
 {
 	struct fsnotify_mark_connector *conn;
 
@@ -547,10 +623,9 @@ static int fsnotify_attach_connector_to_object(fsnotify_connp_t *connp,
 	spin_lock_init(&conn->lock);
 	INIT_HLIST_HEAD(&conn->list);
 	conn->flags = 0;
+	conn->prio = 0;
 	conn->type = obj_type;
-	conn->obj = connp;
-	conn->flags = 0;
-	fsnotify_get_sb_connectors(conn);
+	conn->obj = obj;
 
 	/*
 	 * cmpxchg() provides the barrier so that readers of *connp can see
@@ -558,10 +633,8 @@ static int fsnotify_attach_connector_to_object(fsnotify_connp_t *connp,
 	 */
 	if (cmpxchg(connp, NULL, conn)) {
 		/* Someone else created list structure for us */
-		fsnotify_put_sb_connectors(conn);
 		kmem_cache_free(fsnotify_mark_connector_cachep, conn);
 	}
-
 	return 0;
 }
 
@@ -598,24 +671,36 @@ out:
  * to which group and for which inodes. These marks are ordered according to
  * priority, highest number first, and then by the group's location in memory.
  */
-static int fsnotify_add_mark_list(struct fsnotify_mark *mark,
-				  fsnotify_connp_t *connp,
+static int fsnotify_add_mark_list(struct fsnotify_mark *mark, void *obj,
 				  unsigned int obj_type, int add_flags)
 {
+	struct super_block *sb = fsnotify_object_sb(obj, obj_type);
 	struct fsnotify_mark *lmark, *last = NULL;
 	struct fsnotify_mark_connector *conn;
+	fsnotify_connp_t *connp;
 	int cmp;
 	int err = 0;
 
 	if (WARN_ON(!fsnotify_valid_obj_type(obj_type)))
 		return -EINVAL;
 
+	/*
+	 * Attach the sb info before attaching a connector to any object on sb.
+	 * The sb info will remain attached as long as sb lives.
+	 */
+	if (!fsnotify_sb_info(sb)) {
+		err = fsnotify_attach_info_to_sb(sb);
+		if (err)
+			return err;
+	}
+
+	connp = fsnotify_object_connp(obj, obj_type);
 restart:
 	spin_lock(&mark->lock);
 	conn = fsnotify_grab_connector(connp);
 	if (!conn) {
 		spin_unlock(&mark->lock);
-		err = fsnotify_attach_connector_to_object(connp, obj_type);
+		err = fsnotify_attach_connector_to_object(connp, obj, obj_type);
 		if (err)
 			return err;
 		goto restart;
@@ -649,6 +734,7 @@ restart:
 	/* mark should be the last entry.  last is the current last entry */
 	hlist_add_behind_rcu(&mark->obj_list, &last->obj_list);
 added:
+	fsnotify_update_sb_watchers(sb, conn);
 	/*
 	 * Since connector is attached to object using cmpxchg() we are
 	 * guaranteed that connector initialization is fully visible by anyone
@@ -667,7 +753,7 @@ out_err:
  * event types should be delivered to which group.
  */
 int fsnotify_add_mark_locked(struct fsnotify_mark *mark,
-			     fsnotify_connp_t *connp, unsigned int obj_type,
+			     void *obj, unsigned int obj_type,
 			     int add_flags)
 {
 	struct fsnotify_group *group = mark->group;
@@ -688,7 +774,7 @@ int fsnotify_add_mark_locked(struct fsnotify_mark *mark,
 	fsnotify_get_mark(mark); /* for g_list */
 	spin_unlock(&mark->lock);
 
-	ret = fsnotify_add_mark_list(mark, connp, obj_type, add_flags);
+	ret = fsnotify_add_mark_list(mark, obj, obj_type, add_flags);
 	if (ret)
 		goto err;
 
@@ -706,14 +792,14 @@ err:
 	return ret;
 }
 
-int fsnotify_add_mark(struct fsnotify_mark *mark, fsnotify_connp_t *connp,
+int fsnotify_add_mark(struct fsnotify_mark *mark, void *obj,
 		      unsigned int obj_type, int add_flags)
 {
 	int ret;
 	struct fsnotify_group *group = mark->group;
 
 	fsnotify_group_lock(group);
-	ret = fsnotify_add_mark_locked(mark, connp, obj_type, add_flags);
+	ret = fsnotify_add_mark_locked(mark, obj, obj_type, add_flags);
 	fsnotify_group_unlock(group);
 	return ret;
 }
@@ -723,12 +809,16 @@ EXPORT_SYMBOL_GPL(fsnotify_add_mark);
  * Given a list of marks, find the mark associated with given group. If found
  * take a reference to that mark and return it, else return NULL.
  */
-struct fsnotify_mark *fsnotify_find_mark(fsnotify_connp_t *connp,
+struct fsnotify_mark *fsnotify_find_mark(void *obj, unsigned int obj_type,
 					 struct fsnotify_group *group)
 {
+	fsnotify_connp_t *connp = fsnotify_object_connp(obj, obj_type);
 	struct fsnotify_mark_connector *conn;
 	struct fsnotify_mark *mark;
 
+	if (!connp)
+		return NULL;
+
 	conn = fsnotify_grab_connector(connp);
 	if (!conn)
 		return NULL;
diff --git a/fs/nsfs.c b/fs/nsfs.c
index 34e1e3e367..07e22a15ef 100644
--- a/fs/nsfs.c
+++ b/fs/nsfs.c
@@ -27,26 +27,17 @@ static const struct file_operations ns_file_operations = {
 static char *ns_dname(struct dentry *dentry, char *buffer, int buflen)
 {
 	struct inode *inode = d_inode(dentry);
-	const struct proc_ns_operations *ns_ops = dentry->d_fsdata;
+	struct ns_common *ns = inode->i_private;
+	const struct proc_ns_operations *ns_ops = ns->ops;
 
 	return dynamic_dname(buffer, buflen, "%s:[%lu]",
 		ns_ops->name, inode->i_ino);
 }
 
-static void ns_prune_dentry(struct dentry *dentry)
-{
-	struct inode *inode = d_inode(dentry);
-	if (inode) {
-		struct ns_common *ns = inode->i_private;
-		atomic_long_set(&ns->stashed, 0);
-	}
-}
-
-const struct dentry_operations ns_dentry_operations =
-{
-	.d_prune	= ns_prune_dentry,
+const struct dentry_operations ns_dentry_operations = {
 	.d_delete	= always_delete_dentry,
 	.d_dname	= ns_dname,
+	.d_prune	= stashed_dentry_prune,
 };
 
 static void nsfs_evict(struct inode *inode)
@@ -56,67 +47,16 @@ static void nsfs_evict(struct inode *inode)
 	ns->ops->put(ns);
 }
 
-static int __ns_get_path(struct path *path, struct ns_common *ns)
-{
-	struct vfsmount *mnt = nsfs_mnt;
-	struct dentry *dentry;
-	struct inode *inode;
-	unsigned long d;
-
-	rcu_read_lock();
-	d = atomic_long_read(&ns->stashed);
-	if (!d)
-		goto slow;
-	dentry = (struct dentry *)d;
-	if (!lockref_get_not_dead(&dentry->d_lockref))
-		goto slow;
-	rcu_read_unlock();
-	ns->ops->put(ns);
-got_it:
-	path->mnt = mntget(mnt);
-	path->dentry = dentry;
-	return 0;
-slow:
-	rcu_read_unlock();
-	inode = new_inode_pseudo(mnt->mnt_sb);
-	if (!inode) {
-		ns->ops->put(ns);
-		return -ENOMEM;
-	}
-	inode->i_ino = ns->inum;
-	simple_inode_init_ts(inode);
-	inode->i_flags |= S_IMMUTABLE;
-	inode->i_mode = S_IFREG | S_IRUGO;
-	inode->i_fop = &ns_file_operations;
-	inode->i_private = ns;
-
-	dentry = d_make_root(inode);	/* not the normal use, but... */
-	if (!dentry)
-		return -ENOMEM;
-	dentry->d_fsdata = (void *)ns->ops;
-	d = atomic_long_cmpxchg(&ns->stashed, 0, (unsigned long)dentry);
-	if (d) {
-		d_delete(dentry);	/* make sure ->d_prune() does nothing */
-		dput(dentry);
-		cpu_relax();
-		return -EAGAIN;
-	}
-	goto got_it;
-}
-
 int ns_get_path_cb(struct path *path, ns_get_path_helper_t *ns_get_cb,
 		     void *private_data)
 {
-	int ret;
+	struct ns_common *ns;
 
-	do {
-		struct ns_common *ns = ns_get_cb(private_data);
-		if (!ns)
-			return -ENOENT;
-		ret = __ns_get_path(path, ns);
-	} while (ret == -EAGAIN);
+	ns = ns_get_cb(private_data);
+	if (!ns)
+		return -ENOENT;
 
-	return ret;
+	return path_from_stashed(&ns->stashed, nsfs_mnt, ns, path);
 }
 
 struct ns_get_path_task_args {
@@ -146,6 +86,7 @@ int open_related_ns(struct ns_common *ns,
 		   struct ns_common *(*get_ns)(struct ns_common *ns))
 {
 	struct path path = {};
+	struct ns_common *relative;
 	struct file *f;
 	int err;
 	int fd;
@@ -154,19 +95,14 @@ int open_related_ns(struct ns_common *ns,
 	if (fd < 0)
 		return fd;
 
-	do {
-		struct ns_common *relative;
-
-		relative = get_ns(ns);
-		if (IS_ERR(relative)) {
-			put_unused_fd(fd);
-			return PTR_ERR(relative);
-		}
-
-		err = __ns_get_path(&path, relative);
-	} while (err == -EAGAIN);
+	relative = get_ns(ns);
+	if (IS_ERR(relative)) {
+		put_unused_fd(fd);
+		return PTR_ERR(relative);
+	}
 
-	if (err) {
+	err = path_from_stashed(&relative->stashed, nsfs_mnt, relative, &path);
+	if (err < 0) {
 		put_unused_fd(fd);
 		return err;
 	}
@@ -249,7 +185,8 @@ bool ns_match(const struct ns_common *ns, dev_t dev, ino_t ino)
 static int nsfs_show_path(struct seq_file *seq, struct dentry *dentry)
 {
 	struct inode *inode = d_inode(dentry);
-	const struct proc_ns_operations *ns_ops = dentry->d_fsdata;
+	const struct ns_common *ns = inode->i_private;
+	const struct proc_ns_operations *ns_ops = ns->ops;
 
 	seq_printf(seq, "%s:[%lu]", ns_ops->name, inode->i_ino);
 	return 0;
@@ -261,6 +198,28 @@ static const struct super_operations nsfs_ops = {
 	.show_path = nsfs_show_path,
 };
 
+static int nsfs_init_inode(struct inode *inode, void *data)
+{
+	struct ns_common *ns = data;
+
+	inode->i_private = data;
+	inode->i_mode |= S_IRUGO;
+	inode->i_fop = &ns_file_operations;
+	inode->i_ino = ns->inum;
+	return 0;
+}
+
+static void nsfs_put_data(void *data)
+{
+	struct ns_common *ns = data;
+	ns->ops->put(ns);
+}
+
+static const struct stashed_operations nsfs_stashed_ops = {
+	.init_inode = nsfs_init_inode,
+	.put_data = nsfs_put_data,
+};
+
 static int nsfs_init_fs_context(struct fs_context *fc)
 {
 	struct pseudo_fs_context *ctx = init_pseudo(fc, NSFS_MAGIC);
@@ -268,6 +227,7 @@ static int nsfs_init_fs_context(struct fs_context *fc)
 		return -ENOMEM;
 	ctx->ops = &nsfs_ops;
 	ctx->dops = &ns_dentry_operations;
+	fc->s_fs_info = (void *)&nsfs_stashed_ops;
 	return 0;
 }
 
diff --git a/fs/ntfs/Kconfig b/fs/ntfs/Kconfig
deleted file mode 100644
index 7b25097417..0000000000
--- a/fs/ntfs/Kconfig
+++ /dev/null
@@ -1,81 +0,0 @@
-# SPDX-License-Identifier: GPL-2.0-only
-config NTFS_FS
-	tristate "NTFS file system support"
-	select BUFFER_HEAD
-	select NLS
-	help
-	  NTFS is the file system of Microsoft Windows NT, 2000, XP and 2003.
-
-	  Saying Y or M here enables read support.  There is partial, but
-	  safe, write support available.  For write support you must also
-	  say Y to "NTFS write support" below.
-
-	  There are also a number of user-space tools available, called
-	  ntfsprogs.  These include ntfsundelete and ntfsresize, that work
-	  without NTFS support enabled in the kernel.
-
-	  This is a rewrite from scratch of Linux NTFS support and replaced
-	  the old NTFS code starting with Linux 2.5.11.  A backport to
-	  the Linux 2.4 kernel series is separately available as a patch
-	  from the project web site.
-
-	  For more information see <file:Documentation/filesystems/ntfs.rst>
-	  and <http://www.linux-ntfs.org/>.
-
-	  To compile this file system support as a module, choose M here: the
-	  module will be called ntfs.
-
-	  If you are not using Windows NT, 2000, XP or 2003 in addition to
-	  Linux on your computer it is safe to say N.
-
-config NTFS_DEBUG
-	bool "NTFS debugging support"
-	depends on NTFS_FS
-	help
-	  If you are experiencing any problems with the NTFS file system, say
-	  Y here.  This will result in additional consistency checks to be
-	  performed by the driver as well as additional debugging messages to
-	  be written to the system log.  Note that debugging messages are
-	  disabled by default.  To enable them, supply the option debug_msgs=1
-	  at the kernel command line when booting the kernel or as an option
-	  to insmod when loading the ntfs module.  Once the driver is active,
-	  you can enable debugging messages by doing (as root):
-	  echo 1 > /proc/sys/fs/ntfs-debug
-	  Replacing the "1" with "0" would disable debug messages.
-
-	  If you leave debugging messages disabled, this results in little
-	  overhead, but enabling debug messages results in very significant
-	  slowdown of the system.
-
-	  When reporting bugs, please try to have available a full dump of
-	  debugging messages while the misbehaviour was occurring.
-
-config NTFS_RW
-	bool "NTFS write support"
-	depends on NTFS_FS
-	depends on PAGE_SIZE_LESS_THAN_64KB
-	help
-	  This enables the partial, but safe, write support in the NTFS driver.
-
-	  The only supported operation is overwriting existing files, without
-	  changing the file length.  No file or directory creation, deletion or
-	  renaming is possible.  Note only non-resident files can be written to
-	  so you may find that some very small files (<500 bytes or so) cannot
-	  be written to.
-
-	  While we cannot guarantee that it will not damage any data, we have
-	  so far not received a single report where the driver would have
-	  damaged someones data so we assume it is perfectly safe to use.
-
-	  Note:  While write support is safe in this version (a rewrite from
-	  scratch of the NTFS support), it should be noted that the old NTFS
-	  write support, included in Linux 2.5.10 and before (since 1997),
-	  is not safe.
-
-	  This is currently useful with TopologiLinux.  TopologiLinux is run
-	  on top of any DOS/Microsoft Windows system without partitioning your
-	  hard disk.  Unlike other Linux distributions TopologiLinux does not
-	  need its own partition.  For more information see
-	  <http://topologi-linux.sourceforge.net/>
-
-	  It is perfectly safe to say N here.
diff --git a/fs/ntfs/Makefile b/fs/ntfs/Makefile
deleted file mode 100644
index 3e736572ed..0000000000
--- a/fs/ntfs/Makefile
+++ /dev/null
@@ -1,15 +0,0 @@
-# SPDX-License-Identifier: GPL-2.0
-# Rules for making the NTFS driver.
-
-obj-$(CONFIG_NTFS_FS) += ntfs.o
-
-ntfs-y := aops.o attrib.o collate.o compress.o debug.o dir.o file.o \
-	  index.o inode.o mft.o mst.o namei.o runlist.o super.o sysctl.o \
-	  unistr.o upcase.o
-
-ntfs-$(CONFIG_NTFS_RW) += bitmap.o lcnalloc.o logfile.o quota.o usnjrnl.o
-
-ccflags-y := -DNTFS_VERSION=\"2.1.32\"
-ccflags-$(CONFIG_NTFS_DEBUG)	+= -DDEBUG
-ccflags-$(CONFIG_NTFS_RW)	+= -DNTFS_RW
-
diff --git a/fs/ntfs/aops.c b/fs/ntfs/aops.c
deleted file mode 100644
index 2d01517a2d..0000000000
--- a/fs/ntfs/aops.c
+++ /dev/null
@@ -1,1744 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * aops.c - NTFS kernel address space operations and page cache handling.
- *
- * Copyright (c) 2001-2014 Anton Altaparmakov and Tuxera Inc.
- * Copyright (c) 2002 Richard Russon
- */
-
-#include <linux/errno.h>
-#include <linux/fs.h>
-#include <linux/gfp.h>
-#include <linux/mm.h>
-#include <linux/pagemap.h>
-#include <linux/swap.h>
-#include <linux/buffer_head.h>
-#include <linux/writeback.h>
-#include <linux/bit_spinlock.h>
-#include <linux/bio.h>
-
-#include "aops.h"
-#include "attrib.h"
-#include "debug.h"
-#include "inode.h"
-#include "mft.h"
-#include "runlist.h"
-#include "types.h"
-#include "ntfs.h"
-
-/**
- * ntfs_end_buffer_async_read - async io completion for reading attributes
- * @bh:		buffer head on which io is completed
- * @uptodate:	whether @bh is now uptodate or not
- *
- * Asynchronous I/O completion handler for reading pages belonging to the
- * attribute address space of an inode.  The inodes can either be files or
- * directories or they can be fake inodes describing some attribute.
- *
- * If NInoMstProtected(), perform the post read mst fixups when all IO on the
- * page has been completed and mark the page uptodate or set the error bit on
- * the page.  To determine the size of the records that need fixing up, we
- * cheat a little bit by setting the index_block_size in ntfs_inode to the ntfs
- * record size, and index_block_size_bits, to the log(base 2) of the ntfs
- * record size.
- */
-static void ntfs_end_buffer_async_read(struct buffer_head *bh, int uptodate)
-{
-	unsigned long flags;
-	struct buffer_head *first, *tmp;
-	struct page *page;
-	struct inode *vi;
-	ntfs_inode *ni;
-	int page_uptodate = 1;
-
-	page = bh->b_page;
-	vi = page->mapping->host;
-	ni = NTFS_I(vi);
-
-	if (likely(uptodate)) {
-		loff_t i_size;
-		s64 file_ofs, init_size;
-
-		set_buffer_uptodate(bh);
-
-		file_ofs = ((s64)page->index << PAGE_SHIFT) +
-				bh_offset(bh);
-		read_lock_irqsave(&ni->size_lock, flags);
-		init_size = ni->initialized_size;
-		i_size = i_size_read(vi);
-		read_unlock_irqrestore(&ni->size_lock, flags);
-		if (unlikely(init_size > i_size)) {
-			/* Race with shrinking truncate. */
-			init_size = i_size;
-		}
-		/* Check for the current buffer head overflowing. */
-		if (unlikely(file_ofs + bh->b_size > init_size)) {
-			int ofs;
-			void *kaddr;
-
-			ofs = 0;
-			if (file_ofs < init_size)
-				ofs = init_size - file_ofs;
-			kaddr = kmap_atomic(page);
-			memset(kaddr + bh_offset(bh) + ofs, 0,
-					bh->b_size - ofs);
-			flush_dcache_page(page);
-			kunmap_atomic(kaddr);
-		}
-	} else {
-		clear_buffer_uptodate(bh);
-		SetPageError(page);
-		ntfs_error(ni->vol->sb, "Buffer I/O error, logical block "
-				"0x%llx.", (unsigned long long)bh->b_blocknr);
-	}
-	first = page_buffers(page);
-	spin_lock_irqsave(&first->b_uptodate_lock, flags);
-	clear_buffer_async_read(bh);
-	unlock_buffer(bh);
-	tmp = bh;
-	do {
-		if (!buffer_uptodate(tmp))
-			page_uptodate = 0;
-		if (buffer_async_read(tmp)) {
-			if (likely(buffer_locked(tmp)))
-				goto still_busy;
-			/* Async buffers must be locked. */
-			BUG();
-		}
-		tmp = tmp->b_this_page;
-	} while (tmp != bh);
-	spin_unlock_irqrestore(&first->b_uptodate_lock, flags);
-	/*
-	 * If none of the buffers had errors then we can set the page uptodate,
-	 * but we first have to perform the post read mst fixups, if the
-	 * attribute is mst protected, i.e. if NInoMstProteced(ni) is true.
-	 * Note we ignore fixup errors as those are detected when
-	 * map_mft_record() is called which gives us per record granularity
-	 * rather than per page granularity.
-	 */
-	if (!NInoMstProtected(ni)) {
-		if (likely(page_uptodate && !PageError(page)))
-			SetPageUptodate(page);
-	} else {
-		u8 *kaddr;
-		unsigned int i, recs;
-		u32 rec_size;
-
-		rec_size = ni->itype.index.block_size;
-		recs = PAGE_SIZE / rec_size;
-		/* Should have been verified before we got here... */
-		BUG_ON(!recs);
-		kaddr = kmap_atomic(page);
-		for (i = 0; i < recs; i++)
-			post_read_mst_fixup((NTFS_RECORD*)(kaddr +
-					i * rec_size), rec_size);
-		kunmap_atomic(kaddr);
-		flush_dcache_page(page);
-		if (likely(page_uptodate && !PageError(page)))
-			SetPageUptodate(page);
-	}
-	unlock_page(page);
-	return;
-still_busy:
-	spin_unlock_irqrestore(&first->b_uptodate_lock, flags);
-	return;
-}
-
-/**
- * ntfs_read_block - fill a @folio of an address space with data
- * @folio:	page cache folio to fill with data
- *
- * We read each buffer asynchronously and when all buffers are read in, our io
- * completion handler ntfs_end_buffer_read_async(), if required, automatically
- * applies the mst fixups to the folio before finally marking it uptodate and
- * unlocking it.
- *
- * We only enforce allocated_size limit because i_size is checked for in
- * generic_file_read().
- *
- * Return 0 on success and -errno on error.
- *
- * Contains an adapted version of fs/buffer.c::block_read_full_folio().
- */
-static int ntfs_read_block(struct folio *folio)
-{
-	loff_t i_size;
-	VCN vcn;
-	LCN lcn;
-	s64 init_size;
-	struct inode *vi;
-	ntfs_inode *ni;
-	ntfs_volume *vol;
-	runlist_element *rl;
-	struct buffer_head *bh, *head, *arr[MAX_BUF_PER_PAGE];
-	sector_t iblock, lblock, zblock;
-	unsigned long flags;
-	unsigned int blocksize, vcn_ofs;
-	int i, nr;
-	unsigned char blocksize_bits;
-
-	vi = folio->mapping->host;
-	ni = NTFS_I(vi);
-	vol = ni->vol;
-
-	/* $MFT/$DATA must have its complete runlist in memory at all times. */
-	BUG_ON(!ni->runlist.rl && !ni->mft_no && !NInoAttr(ni));
-
-	blocksize = vol->sb->s_blocksize;
-	blocksize_bits = vol->sb->s_blocksize_bits;
-
-	head = folio_buffers(folio);
-	if (!head)
-		head = create_empty_buffers(folio, blocksize, 0);
-	bh = head;
-
-	/*
-	 * We may be racing with truncate.  To avoid some of the problems we
-	 * now take a snapshot of the various sizes and use those for the whole
-	 * of the function.  In case of an extending truncate it just means we
-	 * may leave some buffers unmapped which are now allocated.  This is
-	 * not a problem since these buffers will just get mapped when a write
-	 * occurs.  In case of a shrinking truncate, we will detect this later
-	 * on due to the runlist being incomplete and if the folio is being
-	 * fully truncated, truncate will throw it away as soon as we unlock
-	 * it so no need to worry what we do with it.
-	 */
-	iblock = (s64)folio->index << (PAGE_SHIFT - blocksize_bits);
-	read_lock_irqsave(&ni->size_lock, flags);
-	lblock = (ni->allocated_size + blocksize - 1) >> blocksize_bits;
-	init_size = ni->initialized_size;
-	i_size = i_size_read(vi);
-	read_unlock_irqrestore(&ni->size_lock, flags);
-	if (unlikely(init_size > i_size)) {
-		/* Race with shrinking truncate. */
-		init_size = i_size;
-	}
-	zblock = (init_size + blocksize - 1) >> blocksize_bits;
-
-	/* Loop through all the buffers in the folio. */
-	rl = NULL;
-	nr = i = 0;
-	do {
-		int err = 0;
-
-		if (unlikely(buffer_uptodate(bh)))
-			continue;
-		if (unlikely(buffer_mapped(bh))) {
-			arr[nr++] = bh;
-			continue;
-		}
-		bh->b_bdev = vol->sb->s_bdev;
-		/* Is the block within the allowed limits? */
-		if (iblock < lblock) {
-			bool is_retry = false;
-
-			/* Convert iblock into corresponding vcn and offset. */
-			vcn = (VCN)iblock << blocksize_bits >>
-					vol->cluster_size_bits;
-			vcn_ofs = ((VCN)iblock << blocksize_bits) &
-					vol->cluster_size_mask;
-			if (!rl) {
-lock_retry_remap:
-				down_read(&ni->runlist.lock);
-				rl = ni->runlist.rl;
-			}
-			if (likely(rl != NULL)) {
-				/* Seek to element containing target vcn. */
-				while (rl->length && rl[1].vcn <= vcn)
-					rl++;
-				lcn = ntfs_rl_vcn_to_lcn(rl, vcn);
-			} else
-				lcn = LCN_RL_NOT_MAPPED;
-			/* Successful remap. */
-			if (lcn >= 0) {
-				/* Setup buffer head to correct block. */
-				bh->b_blocknr = ((lcn << vol->cluster_size_bits)
-						+ vcn_ofs) >> blocksize_bits;
-				set_buffer_mapped(bh);
-				/* Only read initialized data blocks. */
-				if (iblock < zblock) {
-					arr[nr++] = bh;
-					continue;
-				}
-				/* Fully non-initialized data block, zero it. */
-				goto handle_zblock;
-			}
-			/* It is a hole, need to zero it. */
-			if (lcn == LCN_HOLE)
-				goto handle_hole;
-			/* If first try and runlist unmapped, map and retry. */
-			if (!is_retry && lcn == LCN_RL_NOT_MAPPED) {
-				is_retry = true;
-				/*
-				 * Attempt to map runlist, dropping lock for
-				 * the duration.
-				 */
-				up_read(&ni->runlist.lock);
-				err = ntfs_map_runlist(ni, vcn);
-				if (likely(!err))
-					goto lock_retry_remap;
-				rl = NULL;
-			} else if (!rl)
-				up_read(&ni->runlist.lock);
-			/*
-			 * If buffer is outside the runlist, treat it as a
-			 * hole.  This can happen due to concurrent truncate
-			 * for example.
-			 */
-			if (err == -ENOENT || lcn == LCN_ENOENT) {
-				err = 0;
-				goto handle_hole;
-			}
-			/* Hard error, zero out region. */
-			if (!err)
-				err = -EIO;
-			bh->b_blocknr = -1;
-			folio_set_error(folio);
-			ntfs_error(vol->sb, "Failed to read from inode 0x%lx, "
-					"attribute type 0x%x, vcn 0x%llx, "
-					"offset 0x%x because its location on "
-					"disk could not be determined%s "
-					"(error code %i).", ni->mft_no,
-					ni->type, (unsigned long long)vcn,
-					vcn_ofs, is_retry ? " even after "
-					"retrying" : "", err);
-		}
-		/*
-		 * Either iblock was outside lblock limits or
-		 * ntfs_rl_vcn_to_lcn() returned error.  Just zero that portion
-		 * of the folio and set the buffer uptodate.
-		 */
-handle_hole:
-		bh->b_blocknr = -1UL;
-		clear_buffer_mapped(bh);
-handle_zblock:
-		folio_zero_range(folio, i * blocksize, blocksize);
-		if (likely(!err))
-			set_buffer_uptodate(bh);
-	} while (i++, iblock++, (bh = bh->b_this_page) != head);
-
-	/* Release the lock if we took it. */
-	if (rl)
-		up_read(&ni->runlist.lock);
-
-	/* Check we have at least one buffer ready for i/o. */
-	if (nr) {
-		struct buffer_head *tbh;
-
-		/* Lock the buffers. */
-		for (i = 0; i < nr; i++) {
-			tbh = arr[i];
-			lock_buffer(tbh);
-			tbh->b_end_io = ntfs_end_buffer_async_read;
-			set_buffer_async_read(tbh);
-		}
-		/* Finally, start i/o on the buffers. */
-		for (i = 0; i < nr; i++) {
-			tbh = arr[i];
-			if (likely(!buffer_uptodate(tbh)))
-				submit_bh(REQ_OP_READ, tbh);
-			else
-				ntfs_end_buffer_async_read(tbh, 1);
-		}
-		return 0;
-	}
-	/* No i/o was scheduled on any of the buffers. */
-	if (likely(!folio_test_error(folio)))
-		folio_mark_uptodate(folio);
-	else /* Signal synchronous i/o error. */
-		nr = -EIO;
-	folio_unlock(folio);
-	return nr;
-}
-
-/**
- * ntfs_read_folio - fill a @folio of a @file with data from the device
- * @file:	open file to which the folio @folio belongs or NULL
- * @folio:	page cache folio to fill with data
- *
- * For non-resident attributes, ntfs_read_folio() fills the @folio of the open
- * file @file by calling the ntfs version of the generic block_read_full_folio()
- * function, ntfs_read_block(), which in turn creates and reads in the buffers
- * associated with the folio asynchronously.
- *
- * For resident attributes, OTOH, ntfs_read_folio() fills @folio by copying the
- * data from the mft record (which at this stage is most likely in memory) and
- * fills the remainder with zeroes. Thus, in this case, I/O is synchronous, as
- * even if the mft record is not cached at this point in time, we need to wait
- * for it to be read in before we can do the copy.
- *
- * Return 0 on success and -errno on error.
- */
-static int ntfs_read_folio(struct file *file, struct folio *folio)
-{
-	struct page *page = &folio->page;
-	loff_t i_size;
-	struct inode *vi;
-	ntfs_inode *ni, *base_ni;
-	u8 *addr;
-	ntfs_attr_search_ctx *ctx;
-	MFT_RECORD *mrec;
-	unsigned long flags;
-	u32 attr_len;
-	int err = 0;
-
-retry_readpage:
-	BUG_ON(!PageLocked(page));
-	vi = page->mapping->host;
-	i_size = i_size_read(vi);
-	/* Is the page fully outside i_size? (truncate in progress) */
-	if (unlikely(page->index >= (i_size + PAGE_SIZE - 1) >>
-			PAGE_SHIFT)) {
-		zero_user(page, 0, PAGE_SIZE);
-		ntfs_debug("Read outside i_size - truncated?");
-		goto done;
-	}
-	/*
-	 * This can potentially happen because we clear PageUptodate() during
-	 * ntfs_writepage() of MstProtected() attributes.
-	 */
-	if (PageUptodate(page)) {
-		unlock_page(page);
-		return 0;
-	}
-	ni = NTFS_I(vi);
-	/*
-	 * Only $DATA attributes can be encrypted and only unnamed $DATA
-	 * attributes can be compressed.  Index root can have the flags set but
-	 * this means to create compressed/encrypted files, not that the
-	 * attribute is compressed/encrypted.  Note we need to check for
-	 * AT_INDEX_ALLOCATION since this is the type of both directory and
-	 * index inodes.
-	 */
-	if (ni->type != AT_INDEX_ALLOCATION) {
-		/* If attribute is encrypted, deny access, just like NT4. */
-		if (NInoEncrypted(ni)) {
-			BUG_ON(ni->type != AT_DATA);
-			err = -EACCES;
-			goto err_out;
-		}
-		/* Compressed data streams are handled in compress.c. */
-		if (NInoNonResident(ni) && NInoCompressed(ni)) {
-			BUG_ON(ni->type != AT_DATA);
-			BUG_ON(ni->name_len);
-			return ntfs_read_compressed_block(page);
-		}
-	}
-	/* NInoNonResident() == NInoIndexAllocPresent() */
-	if (NInoNonResident(ni)) {
-		/* Normal, non-resident data stream. */
-		return ntfs_read_block(folio);
-	}
-	/*
-	 * Attribute is resident, implying it is not compressed or encrypted.
-	 * This also means the attribute is smaller than an mft record and
-	 * hence smaller than a page, so can simply zero out any pages with
-	 * index above 0.  Note the attribute can actually be marked compressed
-	 * but if it is resident the actual data is not compressed so we are
-	 * ok to ignore the compressed flag here.
-	 */
-	if (unlikely(page->index > 0)) {
-		zero_user(page, 0, PAGE_SIZE);
-		goto done;
-	}
-	if (!NInoAttr(ni))
-		base_ni = ni;
-	else
-		base_ni = ni->ext.base_ntfs_ino;
-	/* Map, pin, and lock the mft record. */
-	mrec = map_mft_record(base_ni);
-	if (IS_ERR(mrec)) {
-		err = PTR_ERR(mrec);
-		goto err_out;
-	}
-	/*
-	 * If a parallel write made the attribute non-resident, drop the mft
-	 * record and retry the read_folio.
-	 */
-	if (unlikely(NInoNonResident(ni))) {
-		unmap_mft_record(base_ni);
-		goto retry_readpage;
-	}
-	ctx = ntfs_attr_get_search_ctx(base_ni, mrec);
-	if (unlikely(!ctx)) {
-		err = -ENOMEM;
-		goto unm_err_out;
-	}
-	err = ntfs_attr_lookup(ni->type, ni->name, ni->name_len,
-			CASE_SENSITIVE, 0, NULL, 0, ctx);
-	if (unlikely(err))
-		goto put_unm_err_out;
-	attr_len = le32_to_cpu(ctx->attr->data.resident.value_length);
-	read_lock_irqsave(&ni->size_lock, flags);
-	if (unlikely(attr_len > ni->initialized_size))
-		attr_len = ni->initialized_size;
-	i_size = i_size_read(vi);
-	read_unlock_irqrestore(&ni->size_lock, flags);
-	if (unlikely(attr_len > i_size)) {
-		/* Race with shrinking truncate. */
-		attr_len = i_size;
-	}
-	addr = kmap_atomic(page);
-	/* Copy the data to the page. */
-	memcpy(addr, (u8*)ctx->attr +
-			le16_to_cpu(ctx->attr->data.resident.value_offset),
-			attr_len);
-	/* Zero the remainder of the page. */
-	memset(addr + attr_len, 0, PAGE_SIZE - attr_len);
-	flush_dcache_page(page);
-	kunmap_atomic(addr);
-put_unm_err_out:
-	ntfs_attr_put_search_ctx(ctx);
-unm_err_out:
-	unmap_mft_record(base_ni);
-done:
-	SetPageUptodate(page);
-err_out:
-	unlock_page(page);
-	return err;
-}
-
-#ifdef NTFS_RW
-
-/**
- * ntfs_write_block - write a @folio to the backing store
- * @folio:	page cache folio to write out
- * @wbc:	writeback control structure
- *
- * This function is for writing folios belonging to non-resident, non-mst
- * protected attributes to their backing store.
- *
- * For a folio with buffers, map and write the dirty buffers asynchronously
- * under folio writeback. For a folio without buffers, create buffers for the
- * folio, then proceed as above.
- *
- * If a folio doesn't have buffers the folio dirty state is definitive. If
- * a folio does have buffers, the folio dirty state is just a hint,
- * and the buffer dirty state is definitive. (A hint which has rules:
- * dirty buffers against a clean folio is illegal. Other combinations are
- * legal and need to be handled. In particular a dirty folio containing
- * clean buffers for example.)
- *
- * Return 0 on success and -errno on error.
- *
- * Based on ntfs_read_block() and __block_write_full_folio().
- */
-static int ntfs_write_block(struct folio *folio, struct writeback_control *wbc)
-{
-	VCN vcn;
-	LCN lcn;
-	s64 initialized_size;
-	loff_t i_size;
-	sector_t block, dblock, iblock;
-	struct inode *vi;
-	ntfs_inode *ni;
-	ntfs_volume *vol;
-	runlist_element *rl;
-	struct buffer_head *bh, *head;
-	unsigned long flags;
-	unsigned int blocksize, vcn_ofs;
-	int err;
-	bool need_end_writeback;
-	unsigned char blocksize_bits;
-
-	vi = folio->mapping->host;
-	ni = NTFS_I(vi);
-	vol = ni->vol;
-
-	ntfs_debug("Entering for inode 0x%lx, attribute type 0x%x, page index "
-			"0x%lx.", ni->mft_no, ni->type, folio->index);
-
-	BUG_ON(!NInoNonResident(ni));
-	BUG_ON(NInoMstProtected(ni));
-	blocksize = vol->sb->s_blocksize;
-	blocksize_bits = vol->sb->s_blocksize_bits;
-	head = folio_buffers(folio);
-	if (!head) {
-		BUG_ON(!folio_test_uptodate(folio));
-		head = create_empty_buffers(folio, blocksize,
-				(1 << BH_Uptodate) | (1 << BH_Dirty));
-	}
-	bh = head;
-
-	/* NOTE: Different naming scheme to ntfs_read_block()! */
-
-	/* The first block in the folio. */
-	block = (s64)folio->index << (PAGE_SHIFT - blocksize_bits);
-
-	read_lock_irqsave(&ni->size_lock, flags);
-	i_size = i_size_read(vi);
-	initialized_size = ni->initialized_size;
-	read_unlock_irqrestore(&ni->size_lock, flags);
-
-	/* The first out of bounds block for the data size. */
-	dblock = (i_size + blocksize - 1) >> blocksize_bits;
-
-	/* The last (fully or partially) initialized block. */
-	iblock = initialized_size >> blocksize_bits;
-
-	/*
-	 * Be very careful.  We have no exclusion from block_dirty_folio
-	 * here, and the (potentially unmapped) buffers may become dirty at
-	 * any time.  If a buffer becomes dirty here after we've inspected it
-	 * then we just miss that fact, and the folio stays dirty.
-	 *
-	 * Buffers outside i_size may be dirtied by block_dirty_folio;
-	 * handle that here by just cleaning them.
-	 */
-
-	/*
-	 * Loop through all the buffers in the folio, mapping all the dirty
-	 * buffers to disk addresses and handling any aliases from the
-	 * underlying block device's mapping.
-	 */
-	rl = NULL;
-	err = 0;
-	do {
-		bool is_retry = false;
-
-		if (unlikely(block >= dblock)) {
-			/*
-			 * Mapped buffers outside i_size will occur, because
-			 * this folio can be outside i_size when there is a
-			 * truncate in progress. The contents of such buffers
-			 * were zeroed by ntfs_writepage().
-			 *
-			 * FIXME: What about the small race window where
-			 * ntfs_writepage() has not done any clearing because
-			 * the folio was within i_size but before we get here,
-			 * vmtruncate() modifies i_size?
-			 */
-			clear_buffer_dirty(bh);
-			set_buffer_uptodate(bh);
-			continue;
-		}
-
-		/* Clean buffers are not written out, so no need to map them. */
-		if (!buffer_dirty(bh))
-			continue;
-
-		/* Make sure we have enough initialized size. */
-		if (unlikely((block >= iblock) &&
-				(initialized_size < i_size))) {
-			/*
-			 * If this folio is fully outside initialized
-			 * size, zero out all folios between the current
-			 * initialized size and the current folio. Just
-			 * use ntfs_read_folio() to do the zeroing
-			 * transparently.
-			 */
-			if (block > iblock) {
-				// TODO:
-				// For each folio do:
-				// - read_cache_folio()
-				// Again for each folio do:
-				// - wait_on_folio_locked()
-				// - Check (folio_test_uptodate(folio) &&
-				//		!folio_test_error(folio))
-				// Update initialized size in the attribute and
-				// in the inode.
-				// Again, for each folio do:
-				//	block_dirty_folio();
-				// folio_put()
-				// We don't need to wait on the writes.
-				// Update iblock.
-			}
-			/*
-			 * The current folio straddles initialized size. Zero
-			 * all non-uptodate buffers and set them uptodate (and
-			 * dirty?). Note, there aren't any non-uptodate buffers
-			 * if the folio is uptodate.
-			 * FIXME: For an uptodate folio, the buffers may need to
-			 * be written out because they were not initialized on
-			 * disk before.
-			 */
-			if (!folio_test_uptodate(folio)) {
-				// TODO:
-				// Zero any non-uptodate buffers up to i_size.
-				// Set them uptodate and dirty.
-			}
-			// TODO:
-			// Update initialized size in the attribute and in the
-			// inode (up to i_size).
-			// Update iblock.
-			// FIXME: This is inefficient. Try to batch the two
-			// size changes to happen in one go.
-			ntfs_error(vol->sb, "Writing beyond initialized size "
-					"is not supported yet. Sorry.");
-			err = -EOPNOTSUPP;
-			break;
-			// Do NOT set_buffer_new() BUT DO clear buffer range
-			// outside write request range.
-			// set_buffer_uptodate() on complete buffers as well as
-			// set_buffer_dirty().
-		}
-
-		/* No need to map buffers that are already mapped. */
-		if (buffer_mapped(bh))
-			continue;
-
-		/* Unmapped, dirty buffer. Need to map it. */
-		bh->b_bdev = vol->sb->s_bdev;
-
-		/* Convert block into corresponding vcn and offset. */
-		vcn = (VCN)block << blocksize_bits;
-		vcn_ofs = vcn & vol->cluster_size_mask;
-		vcn >>= vol->cluster_size_bits;
-		if (!rl) {
-lock_retry_remap:
-			down_read(&ni->runlist.lock);
-			rl = ni->runlist.rl;
-		}
-		if (likely(rl != NULL)) {
-			/* Seek to element containing target vcn. */
-			while (rl->length && rl[1].vcn <= vcn)
-				rl++;
-			lcn = ntfs_rl_vcn_to_lcn(rl, vcn);
-		} else
-			lcn = LCN_RL_NOT_MAPPED;
-		/* Successful remap. */
-		if (lcn >= 0) {
-			/* Setup buffer head to point to correct block. */
-			bh->b_blocknr = ((lcn << vol->cluster_size_bits) +
-					vcn_ofs) >> blocksize_bits;
-			set_buffer_mapped(bh);
-			continue;
-		}
-		/* It is a hole, need to instantiate it. */
-		if (lcn == LCN_HOLE) {
-			u8 *kaddr;
-			unsigned long *bpos, *bend;
-
-			/* Check if the buffer is zero. */
-			kaddr = kmap_local_folio(folio, bh_offset(bh));
-			bpos = (unsigned long *)kaddr;
-			bend = (unsigned long *)(kaddr + blocksize);
-			do {
-				if (unlikely(*bpos))
-					break;
-			} while (likely(++bpos < bend));
-			kunmap_local(kaddr);
-			if (bpos == bend) {
-				/*
-				 * Buffer is zero and sparse, no need to write
-				 * it.
-				 */
-				bh->b_blocknr = -1;
-				clear_buffer_dirty(bh);
-				continue;
-			}
-			// TODO: Instantiate the hole.
-			// clear_buffer_new(bh);
-			// clean_bdev_bh_alias(bh);
-			ntfs_error(vol->sb, "Writing into sparse regions is "
-					"not supported yet. Sorry.");
-			err = -EOPNOTSUPP;
-			break;
-		}
-		/* If first try and runlist unmapped, map and retry. */
-		if (!is_retry && lcn == LCN_RL_NOT_MAPPED) {
-			is_retry = true;
-			/*
-			 * Attempt to map runlist, dropping lock for
-			 * the duration.
-			 */
-			up_read(&ni->runlist.lock);
-			err = ntfs_map_runlist(ni, vcn);
-			if (likely(!err))
-				goto lock_retry_remap;
-			rl = NULL;
-		} else if (!rl)
-			up_read(&ni->runlist.lock);
-		/*
-		 * If buffer is outside the runlist, truncate has cut it out
-		 * of the runlist.  Just clean and clear the buffer and set it
-		 * uptodate so it can get discarded by the VM.
-		 */
-		if (err == -ENOENT || lcn == LCN_ENOENT) {
-			bh->b_blocknr = -1;
-			clear_buffer_dirty(bh);
-			folio_zero_range(folio, bh_offset(bh), blocksize);
-			set_buffer_uptodate(bh);
-			err = 0;
-			continue;
-		}
-		/* Failed to map the buffer, even after retrying. */
-		if (!err)
-			err = -EIO;
-		bh->b_blocknr = -1;
-		ntfs_error(vol->sb, "Failed to write to inode 0x%lx, "
-				"attribute type 0x%x, vcn 0x%llx, offset 0x%x "
-				"because its location on disk could not be "
-				"determined%s (error code %i).", ni->mft_no,
-				ni->type, (unsigned long long)vcn,
-				vcn_ofs, is_retry ? " even after "
-				"retrying" : "", err);
-		break;
-	} while (block++, (bh = bh->b_this_page) != head);
-
-	/* Release the lock if we took it. */
-	if (rl)
-		up_read(&ni->runlist.lock);
-
-	/* For the error case, need to reset bh to the beginning. */
-	bh = head;
-
-	/* Just an optimization, so ->read_folio() is not called later. */
-	if (unlikely(!folio_test_uptodate(folio))) {
-		int uptodate = 1;
-		do {
-			if (!buffer_uptodate(bh)) {
-				uptodate = 0;
-				bh = head;
-				break;
-			}
-		} while ((bh = bh->b_this_page) != head);
-		if (uptodate)
-			folio_mark_uptodate(folio);
-	}
-
-	/* Setup all mapped, dirty buffers for async write i/o. */
-	do {
-		if (buffer_mapped(bh) && buffer_dirty(bh)) {
-			lock_buffer(bh);
-			if (test_clear_buffer_dirty(bh)) {
-				BUG_ON(!buffer_uptodate(bh));
-				mark_buffer_async_write(bh);
-			} else
-				unlock_buffer(bh);
-		} else if (unlikely(err)) {
-			/*
-			 * For the error case. The buffer may have been set
-			 * dirty during attachment to a dirty folio.
-			 */
-			if (err != -ENOMEM)
-				clear_buffer_dirty(bh);
-		}
-	} while ((bh = bh->b_this_page) != head);
-
-	if (unlikely(err)) {
-		// TODO: Remove the -EOPNOTSUPP check later on...
-		if (unlikely(err == -EOPNOTSUPP))
-			err = 0;
-		else if (err == -ENOMEM) {
-			ntfs_warning(vol->sb, "Error allocating memory. "
-					"Redirtying folio so we try again "
-					"later.");
-			/*
-			 * Put the folio back on mapping->dirty_pages, but
-			 * leave its buffer's dirty state as-is.
-			 */
-			folio_redirty_for_writepage(wbc, folio);
-			err = 0;
-		} else
-			folio_set_error(folio);
-	}
-
-	BUG_ON(folio_test_writeback(folio));
-	folio_start_writeback(folio);	/* Keeps try_to_free_buffers() away. */
-
-	/* Submit the prepared buffers for i/o. */
-	need_end_writeback = true;
-	do {
-		struct buffer_head *next = bh->b_this_page;
-		if (buffer_async_write(bh)) {
-			submit_bh(REQ_OP_WRITE, bh);
-			need_end_writeback = false;
-		}
-		bh = next;
-	} while (bh != head);
-	folio_unlock(folio);
-
-	/* If no i/o was started, need to end writeback here. */
-	if (unlikely(need_end_writeback))
-		folio_end_writeback(folio);
-
-	ntfs_debug("Done.");
-	return err;
-}
-
-/**
- * ntfs_write_mst_block - write a @page to the backing store
- * @page:	page cache page to write out
- * @wbc:	writeback control structure
- *
- * This function is for writing pages belonging to non-resident, mst protected
- * attributes to their backing store.  The only supported attributes are index
- * allocation and $MFT/$DATA.  Both directory inodes and index inodes are
- * supported for the index allocation case.
- *
- * The page must remain locked for the duration of the write because we apply
- * the mst fixups, write, and then undo the fixups, so if we were to unlock the
- * page before undoing the fixups, any other user of the page will see the
- * page contents as corrupt.
- *
- * We clear the page uptodate flag for the duration of the function to ensure
- * exclusion for the $MFT/$DATA case against someone mapping an mft record we
- * are about to apply the mst fixups to.
- *
- * Return 0 on success and -errno on error.
- *
- * Based on ntfs_write_block(), ntfs_mft_writepage(), and
- * write_mft_record_nolock().
- */
-static int ntfs_write_mst_block(struct page *page,
-		struct writeback_control *wbc)
-{
-	sector_t block, dblock, rec_block;
-	struct inode *vi = page->mapping->host;
-	ntfs_inode *ni = NTFS_I(vi);
-	ntfs_volume *vol = ni->vol;
-	u8 *kaddr;
-	unsigned int rec_size = ni->itype.index.block_size;
-	ntfs_inode *locked_nis[PAGE_SIZE / NTFS_BLOCK_SIZE];
-	struct buffer_head *bh, *head, *tbh, *rec_start_bh;
-	struct buffer_head *bhs[MAX_BUF_PER_PAGE];
-	runlist_element *rl;
-	int i, nr_locked_nis, nr_recs, nr_bhs, max_bhs, bhs_per_rec, err, err2;
-	unsigned bh_size, rec_size_bits;
-	bool sync, is_mft, page_is_dirty, rec_is_dirty;
-	unsigned char bh_size_bits;
-
-	if (WARN_ON(rec_size < NTFS_BLOCK_SIZE))
-		return -EINVAL;
-
-	ntfs_debug("Entering for inode 0x%lx, attribute type 0x%x, page index "
-			"0x%lx.", vi->i_ino, ni->type, page->index);
-	BUG_ON(!NInoNonResident(ni));
-	BUG_ON(!NInoMstProtected(ni));
-	is_mft = (S_ISREG(vi->i_mode) && !vi->i_ino);
-	/*
-	 * NOTE: ntfs_write_mst_block() would be called for $MFTMirr if a page
-	 * in its page cache were to be marked dirty.  However this should
-	 * never happen with the current driver and considering we do not
-	 * handle this case here we do want to BUG(), at least for now.
-	 */
-	BUG_ON(!(is_mft || S_ISDIR(vi->i_mode) ||
-			(NInoAttr(ni) && ni->type == AT_INDEX_ALLOCATION)));
-	bh_size = vol->sb->s_blocksize;
-	bh_size_bits = vol->sb->s_blocksize_bits;
-	max_bhs = PAGE_SIZE / bh_size;
-	BUG_ON(!max_bhs);
-	BUG_ON(max_bhs > MAX_BUF_PER_PAGE);
-
-	/* Were we called for sync purposes? */
-	sync = (wbc->sync_mode == WB_SYNC_ALL);
-
-	/* Make sure we have mapped buffers. */
-	bh = head = page_buffers(page);
-	BUG_ON(!bh);
-
-	rec_size_bits = ni->itype.index.block_size_bits;
-	BUG_ON(!(PAGE_SIZE >> rec_size_bits));
-	bhs_per_rec = rec_size >> bh_size_bits;
-	BUG_ON(!bhs_per_rec);
-
-	/* The first block in the page. */
-	rec_block = block = (sector_t)page->index <<
-			(PAGE_SHIFT - bh_size_bits);
-
-	/* The first out of bounds block for the data size. */
-	dblock = (i_size_read(vi) + bh_size - 1) >> bh_size_bits;
-
-	rl = NULL;
-	err = err2 = nr_bhs = nr_recs = nr_locked_nis = 0;
-	page_is_dirty = rec_is_dirty = false;
-	rec_start_bh = NULL;
-	do {
-		bool is_retry = false;
-
-		if (likely(block < rec_block)) {
-			if (unlikely(block >= dblock)) {
-				clear_buffer_dirty(bh);
-				set_buffer_uptodate(bh);
-				continue;
-			}
-			/*
-			 * This block is not the first one in the record.  We
-			 * ignore the buffer's dirty state because we could
-			 * have raced with a parallel mark_ntfs_record_dirty().
-			 */
-			if (!rec_is_dirty)
-				continue;
-			if (unlikely(err2)) {
-				if (err2 != -ENOMEM)
-					clear_buffer_dirty(bh);
-				continue;
-			}
-		} else /* if (block == rec_block) */ {
-			BUG_ON(block > rec_block);
-			/* This block is the first one in the record. */
-			rec_block += bhs_per_rec;
-			err2 = 0;
-			if (unlikely(block >= dblock)) {
-				clear_buffer_dirty(bh);
-				continue;
-			}
-			if (!buffer_dirty(bh)) {
-				/* Clean records are not written out. */
-				rec_is_dirty = false;
-				continue;
-			}
-			rec_is_dirty = true;
-			rec_start_bh = bh;
-		}
-		/* Need to map the buffer if it is not mapped already. */
-		if (unlikely(!buffer_mapped(bh))) {
-			VCN vcn;
-			LCN lcn;
-			unsigned int vcn_ofs;
-
-			bh->b_bdev = vol->sb->s_bdev;
-			/* Obtain the vcn and offset of the current block. */
-			vcn = (VCN)block << bh_size_bits;
-			vcn_ofs = vcn & vol->cluster_size_mask;
-			vcn >>= vol->cluster_size_bits;
-			if (!rl) {
-lock_retry_remap:
-				down_read(&ni->runlist.lock);
-				rl = ni->runlist.rl;
-			}
-			if (likely(rl != NULL)) {
-				/* Seek to element containing target vcn. */
-				while (rl->length && rl[1].vcn <= vcn)
-					rl++;
-				lcn = ntfs_rl_vcn_to_lcn(rl, vcn);
-			} else
-				lcn = LCN_RL_NOT_MAPPED;
-			/* Successful remap. */
-			if (likely(lcn >= 0)) {
-				/* Setup buffer head to correct block. */
-				bh->b_blocknr = ((lcn <<
-						vol->cluster_size_bits) +
-						vcn_ofs) >> bh_size_bits;
-				set_buffer_mapped(bh);
-			} else {
-				/*
-				 * Remap failed.  Retry to map the runlist once
-				 * unless we are working on $MFT which always
-				 * has the whole of its runlist in memory.
-				 */
-				if (!is_mft && !is_retry &&
-						lcn == LCN_RL_NOT_MAPPED) {
-					is_retry = true;
-					/*
-					 * Attempt to map runlist, dropping
-					 * lock for the duration.
-					 */
-					up_read(&ni->runlist.lock);
-					err2 = ntfs_map_runlist(ni, vcn);
-					if (likely(!err2))
-						goto lock_retry_remap;
-					if (err2 == -ENOMEM)
-						page_is_dirty = true;
-					lcn = err2;
-				} else {
-					err2 = -EIO;
-					if (!rl)
-						up_read(&ni->runlist.lock);
-				}
-				/* Hard error.  Abort writing this record. */
-				if (!err || err == -ENOMEM)
-					err = err2;
-				bh->b_blocknr = -1;
-				ntfs_error(vol->sb, "Cannot write ntfs record "
-						"0x%llx (inode 0x%lx, "
-						"attribute type 0x%x) because "
-						"its location on disk could "
-						"not be determined (error "
-						"code %lli).",
-						(long long)block <<
-						bh_size_bits >>
-						vol->mft_record_size_bits,
-						ni->mft_no, ni->type,
-						(long long)lcn);
-				/*
-				 * If this is not the first buffer, remove the
-				 * buffers in this record from the list of
-				 * buffers to write and clear their dirty bit
-				 * if not error -ENOMEM.
-				 */
-				if (rec_start_bh != bh) {
-					while (bhs[--nr_bhs] != rec_start_bh)
-						;
-					if (err2 != -ENOMEM) {
-						do {
-							clear_buffer_dirty(
-								rec_start_bh);
-						} while ((rec_start_bh =
-								rec_start_bh->
-								b_this_page) !=
-								bh);
-					}
-				}
-				continue;
-			}
-		}
-		BUG_ON(!buffer_uptodate(bh));
-		BUG_ON(nr_bhs >= max_bhs);
-		bhs[nr_bhs++] = bh;
-	} while (block++, (bh = bh->b_this_page) != head);
-	if (unlikely(rl))
-		up_read(&ni->runlist.lock);
-	/* If there were no dirty buffers, we are done. */
-	if (!nr_bhs)
-		goto done;
-	/* Map the page so we can access its contents. */
-	kaddr = kmap(page);
-	/* Clear the page uptodate flag whilst the mst fixups are applied. */
-	BUG_ON(!PageUptodate(page));
-	ClearPageUptodate(page);
-	for (i = 0; i < nr_bhs; i++) {
-		unsigned int ofs;
-
-		/* Skip buffers which are not at the beginning of records. */
-		if (i % bhs_per_rec)
-			continue;
-		tbh = bhs[i];
-		ofs = bh_offset(tbh);
-		if (is_mft) {
-			ntfs_inode *tni;
-			unsigned long mft_no;
-
-			/* Get the mft record number. */
-			mft_no = (((s64)page->index << PAGE_SHIFT) + ofs)
-					>> rec_size_bits;
-			/* Check whether to write this mft record. */
-			tni = NULL;
-			if (!ntfs_may_write_mft_record(vol, mft_no,
-					(MFT_RECORD*)(kaddr + ofs), &tni)) {
-				/*
-				 * The record should not be written.  This
-				 * means we need to redirty the page before
-				 * returning.
-				 */
-				page_is_dirty = true;
-				/*
-				 * Remove the buffers in this mft record from
-				 * the list of buffers to write.
-				 */
-				do {
-					bhs[i] = NULL;
-				} while (++i % bhs_per_rec);
-				continue;
-			}
-			/*
-			 * The record should be written.  If a locked ntfs
-			 * inode was returned, add it to the array of locked
-			 * ntfs inodes.
-			 */
-			if (tni)
-				locked_nis[nr_locked_nis++] = tni;
-		}
-		/* Apply the mst protection fixups. */
-		err2 = pre_write_mst_fixup((NTFS_RECORD*)(kaddr + ofs),
-				rec_size);
-		if (unlikely(err2)) {
-			if (!err || err == -ENOMEM)
-				err = -EIO;
-			ntfs_error(vol->sb, "Failed to apply mst fixups "
-					"(inode 0x%lx, attribute type 0x%x, "
-					"page index 0x%lx, page offset 0x%x)!"
-					"  Unmount and run chkdsk.", vi->i_ino,
-					ni->type, page->index, ofs);
-			/*
-			 * Mark all the buffers in this record clean as we do
-			 * not want to write corrupt data to disk.
-			 */
-			do {
-				clear_buffer_dirty(bhs[i]);
-				bhs[i] = NULL;
-			} while (++i % bhs_per_rec);
-			continue;
-		}
-		nr_recs++;
-	}
-	/* If no records are to be written out, we are done. */
-	if (!nr_recs)
-		goto unm_done;
-	flush_dcache_page(page);
-	/* Lock buffers and start synchronous write i/o on them. */
-	for (i = 0; i < nr_bhs; i++) {
-		tbh = bhs[i];
-		if (!tbh)
-			continue;
-		if (!trylock_buffer(tbh))
-			BUG();
-		/* The buffer dirty state is now irrelevant, just clean it. */
-		clear_buffer_dirty(tbh);
-		BUG_ON(!buffer_uptodate(tbh));
-		BUG_ON(!buffer_mapped(tbh));
-		get_bh(tbh);
-		tbh->b_end_io = end_buffer_write_sync;
-		submit_bh(REQ_OP_WRITE, tbh);
-	}
-	/* Synchronize the mft mirror now if not @sync. */
-	if (is_mft && !sync)
-		goto do_mirror;
-do_wait:
-	/* Wait on i/o completion of buffers. */
-	for (i = 0; i < nr_bhs; i++) {
-		tbh = bhs[i];
-		if (!tbh)
-			continue;
-		wait_on_buffer(tbh);
-		if (unlikely(!buffer_uptodate(tbh))) {
-			ntfs_error(vol->sb, "I/O error while writing ntfs "
-					"record buffer (inode 0x%lx, "
-					"attribute type 0x%x, page index "
-					"0x%lx, page offset 0x%lx)!  Unmount "
-					"and run chkdsk.", vi->i_ino, ni->type,
-					page->index, bh_offset(tbh));
-			if (!err || err == -ENOMEM)
-				err = -EIO;
-			/*
-			 * Set the buffer uptodate so the page and buffer
-			 * states do not become out of sync.
-			 */
-			set_buffer_uptodate(tbh);
-		}
-	}
-	/* If @sync, now synchronize the mft mirror. */
-	if (is_mft && sync) {
-do_mirror:
-		for (i = 0; i < nr_bhs; i++) {
-			unsigned long mft_no;
-			unsigned int ofs;
-
-			/*
-			 * Skip buffers which are not at the beginning of
-			 * records.
-			 */
-			if (i % bhs_per_rec)
-				continue;
-			tbh = bhs[i];
-			/* Skip removed buffers (and hence records). */
-			if (!tbh)
-				continue;
-			ofs = bh_offset(tbh);
-			/* Get the mft record number. */
-			mft_no = (((s64)page->index << PAGE_SHIFT) + ofs)
-					>> rec_size_bits;
-			if (mft_no < vol->mftmirr_size)
-				ntfs_sync_mft_mirror(vol, mft_no,
-						(MFT_RECORD*)(kaddr + ofs),
-						sync);
-		}
-		if (!sync)
-			goto do_wait;
-	}
-	/* Remove the mst protection fixups again. */
-	for (i = 0; i < nr_bhs; i++) {
-		if (!(i % bhs_per_rec)) {
-			tbh = bhs[i];
-			if (!tbh)
-				continue;
-			post_write_mst_fixup((NTFS_RECORD*)(kaddr +
-					bh_offset(tbh)));
-		}
-	}
-	flush_dcache_page(page);
-unm_done:
-	/* Unlock any locked inodes. */
-	while (nr_locked_nis-- > 0) {
-		ntfs_inode *tni, *base_tni;
-		
-		tni = locked_nis[nr_locked_nis];
-		/* Get the base inode. */
-		mutex_lock(&tni->extent_lock);
-		if (tni->nr_extents >= 0)
-			base_tni = tni;
-		else {
-			base_tni = tni->ext.base_ntfs_ino;
-			BUG_ON(!base_tni);
-		}
-		mutex_unlock(&tni->extent_lock);
-		ntfs_debug("Unlocking %s inode 0x%lx.",
-				tni == base_tni ? "base" : "extent",
-				tni->mft_no);
-		mutex_unlock(&tni->mrec_lock);
-		atomic_dec(&tni->count);
-		iput(VFS_I(base_tni));
-	}
-	SetPageUptodate(page);
-	kunmap(page);
-done:
-	if (unlikely(err && err != -ENOMEM)) {
-		/*
-		 * Set page error if there is only one ntfs record in the page.
-		 * Otherwise we would loose per-record granularity.
-		 */
-		if (ni->itype.index.block_size == PAGE_SIZE)
-			SetPageError(page);
-		NVolSetErrors(vol);
-	}
-	if (page_is_dirty) {
-		ntfs_debug("Page still contains one or more dirty ntfs "
-				"records.  Redirtying the page starting at "
-				"record 0x%lx.", page->index <<
-				(PAGE_SHIFT - rec_size_bits));
-		redirty_page_for_writepage(wbc, page);
-		unlock_page(page);
-	} else {
-		/*
-		 * Keep the VM happy.  This must be done otherwise the
-		 * radix-tree tag PAGECACHE_TAG_DIRTY remains set even though
-		 * the page is clean.
-		 */
-		BUG_ON(PageWriteback(page));
-		set_page_writeback(page);
-		unlock_page(page);
-		end_page_writeback(page);
-	}
-	if (likely(!err))
-		ntfs_debug("Done.");
-	return err;
-}
-
-/**
- * ntfs_writepage - write a @page to the backing store
- * @page:	page cache page to write out
- * @wbc:	writeback control structure
- *
- * This is called from the VM when it wants to have a dirty ntfs page cache
- * page cleaned.  The VM has already locked the page and marked it clean.
- *
- * For non-resident attributes, ntfs_writepage() writes the @page by calling
- * the ntfs version of the generic block_write_full_folio() function,
- * ntfs_write_block(), which in turn if necessary creates and writes the
- * buffers associated with the page asynchronously.
- *
- * For resident attributes, OTOH, ntfs_writepage() writes the @page by copying
- * the data to the mft record (which at this stage is most likely in memory).
- * The mft record is then marked dirty and written out asynchronously via the
- * vfs inode dirty code path for the inode the mft record belongs to or via the
- * vm page dirty code path for the page the mft record is in.
- *
- * Based on ntfs_read_folio() and fs/buffer.c::block_write_full_folio().
- *
- * Return 0 on success and -errno on error.
- */
-static int ntfs_writepage(struct page *page, struct writeback_control *wbc)
-{
-	struct folio *folio = page_folio(page);
-	loff_t i_size;
-	struct inode *vi = folio->mapping->host;
-	ntfs_inode *base_ni = NULL, *ni = NTFS_I(vi);
-	char *addr;
-	ntfs_attr_search_ctx *ctx = NULL;
-	MFT_RECORD *m = NULL;
-	u32 attr_len;
-	int err;
-
-retry_writepage:
-	BUG_ON(!folio_test_locked(folio));
-	i_size = i_size_read(vi);
-	/* Is the folio fully outside i_size? (truncate in progress) */
-	if (unlikely(folio->index >= (i_size + PAGE_SIZE - 1) >>
-			PAGE_SHIFT)) {
-		/*
-		 * The folio may have dirty, unmapped buffers.  Make them
-		 * freeable here, so the page does not leak.
-		 */
-		block_invalidate_folio(folio, 0, folio_size(folio));
-		folio_unlock(folio);
-		ntfs_debug("Write outside i_size - truncated?");
-		return 0;
-	}
-	/*
-	 * Only $DATA attributes can be encrypted and only unnamed $DATA
-	 * attributes can be compressed.  Index root can have the flags set but
-	 * this means to create compressed/encrypted files, not that the
-	 * attribute is compressed/encrypted.  Note we need to check for
-	 * AT_INDEX_ALLOCATION since this is the type of both directory and
-	 * index inodes.
-	 */
-	if (ni->type != AT_INDEX_ALLOCATION) {
-		/* If file is encrypted, deny access, just like NT4. */
-		if (NInoEncrypted(ni)) {
-			folio_unlock(folio);
-			BUG_ON(ni->type != AT_DATA);
-			ntfs_debug("Denying write access to encrypted file.");
-			return -EACCES;
-		}
-		/* Compressed data streams are handled in compress.c. */
-		if (NInoNonResident(ni) && NInoCompressed(ni)) {
-			BUG_ON(ni->type != AT_DATA);
-			BUG_ON(ni->name_len);
-			// TODO: Implement and replace this with
-			// return ntfs_write_compressed_block(page);
-			folio_unlock(folio);
-			ntfs_error(vi->i_sb, "Writing to compressed files is "
-					"not supported yet.  Sorry.");
-			return -EOPNOTSUPP;
-		}
-		// TODO: Implement and remove this check.
-		if (NInoNonResident(ni) && NInoSparse(ni)) {
-			folio_unlock(folio);
-			ntfs_error(vi->i_sb, "Writing to sparse files is not "
-					"supported yet.  Sorry.");
-			return -EOPNOTSUPP;
-		}
-	}
-	/* NInoNonResident() == NInoIndexAllocPresent() */
-	if (NInoNonResident(ni)) {
-		/* We have to zero every time due to mmap-at-end-of-file. */
-		if (folio->index >= (i_size >> PAGE_SHIFT)) {
-			/* The folio straddles i_size. */
-			unsigned int ofs = i_size & (folio_size(folio) - 1);
-			folio_zero_segment(folio, ofs, folio_size(folio));
-		}
-		/* Handle mst protected attributes. */
-		if (NInoMstProtected(ni))
-			return ntfs_write_mst_block(page, wbc);
-		/* Normal, non-resident data stream. */
-		return ntfs_write_block(folio, wbc);
-	}
-	/*
-	 * Attribute is resident, implying it is not compressed, encrypted, or
-	 * mst protected.  This also means the attribute is smaller than an mft
-	 * record and hence smaller than a folio, so can simply return error on
-	 * any folios with index above 0.  Note the attribute can actually be
-	 * marked compressed but if it is resident the actual data is not
-	 * compressed so we are ok to ignore the compressed flag here.
-	 */
-	BUG_ON(folio_buffers(folio));
-	BUG_ON(!folio_test_uptodate(folio));
-	if (unlikely(folio->index > 0)) {
-		ntfs_error(vi->i_sb, "BUG()! folio->index (0x%lx) > 0.  "
-				"Aborting write.", folio->index);
-		BUG_ON(folio_test_writeback(folio));
-		folio_start_writeback(folio);
-		folio_unlock(folio);
-		folio_end_writeback(folio);
-		return -EIO;
-	}
-	if (!NInoAttr(ni))
-		base_ni = ni;
-	else
-		base_ni = ni->ext.base_ntfs_ino;
-	/* Map, pin, and lock the mft record. */
-	m = map_mft_record(base_ni);
-	if (IS_ERR(m)) {
-		err = PTR_ERR(m);
-		m = NULL;
-		ctx = NULL;
-		goto err_out;
-	}
-	/*
-	 * If a parallel write made the attribute non-resident, drop the mft
-	 * record and retry the writepage.
-	 */
-	if (unlikely(NInoNonResident(ni))) {
-		unmap_mft_record(base_ni);
-		goto retry_writepage;
-	}
-	ctx = ntfs_attr_get_search_ctx(base_ni, m);
-	if (unlikely(!ctx)) {
-		err = -ENOMEM;
-		goto err_out;
-	}
-	err = ntfs_attr_lookup(ni->type, ni->name, ni->name_len,
-			CASE_SENSITIVE, 0, NULL, 0, ctx);
-	if (unlikely(err))
-		goto err_out;
-	/*
-	 * Keep the VM happy.  This must be done otherwise
-	 * PAGECACHE_TAG_DIRTY remains set even though the folio is clean.
-	 */
-	BUG_ON(folio_test_writeback(folio));
-	folio_start_writeback(folio);
-	folio_unlock(folio);
-	attr_len = le32_to_cpu(ctx->attr->data.resident.value_length);
-	i_size = i_size_read(vi);
-	if (unlikely(attr_len > i_size)) {
-		/* Race with shrinking truncate or a failed truncate. */
-		attr_len = i_size;
-		/*
-		 * If the truncate failed, fix it up now.  If a concurrent
-		 * truncate, we do its job, so it does not have to do anything.
-		 */
-		err = ntfs_resident_attr_value_resize(ctx->mrec, ctx->attr,
-				attr_len);
-		/* Shrinking cannot fail. */
-		BUG_ON(err);
-	}
-	addr = kmap_local_folio(folio, 0);
-	/* Copy the data from the folio to the mft record. */
-	memcpy((u8*)ctx->attr +
-			le16_to_cpu(ctx->attr->data.resident.value_offset),
-			addr, attr_len);
-	/* Zero out of bounds area in the page cache folio. */
-	memset(addr + attr_len, 0, folio_size(folio) - attr_len);
-	kunmap_local(addr);
-	flush_dcache_folio(folio);
-	flush_dcache_mft_record_page(ctx->ntfs_ino);
-	/* We are done with the folio. */
-	folio_end_writeback(folio);
-	/* Finally, mark the mft record dirty, so it gets written back. */
-	mark_mft_record_dirty(ctx->ntfs_ino);
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(base_ni);
-	return 0;
-err_out:
-	if (err == -ENOMEM) {
-		ntfs_warning(vi->i_sb, "Error allocating memory. Redirtying "
-				"page so we try again later.");
-		/*
-		 * Put the folio back on mapping->dirty_pages, but leave its
-		 * buffers' dirty state as-is.
-		 */
-		folio_redirty_for_writepage(wbc, folio);
-		err = 0;
-	} else {
-		ntfs_error(vi->i_sb, "Resident attribute write failed with "
-				"error %i.", err);
-		folio_set_error(folio);
-		NVolSetErrors(ni->vol);
-	}
-	folio_unlock(folio);
-	if (ctx)
-		ntfs_attr_put_search_ctx(ctx);
-	if (m)
-		unmap_mft_record(base_ni);
-	return err;
-}
-
-#endif	/* NTFS_RW */
-
-/**
- * ntfs_bmap - map logical file block to physical device block
- * @mapping:	address space mapping to which the block to be mapped belongs
- * @block:	logical block to map to its physical device block
- *
- * For regular, non-resident files (i.e. not compressed and not encrypted), map
- * the logical @block belonging to the file described by the address space
- * mapping @mapping to its physical device block.
- *
- * The size of the block is equal to the @s_blocksize field of the super block
- * of the mounted file system which is guaranteed to be smaller than or equal
- * to the cluster size thus the block is guaranteed to fit entirely inside the
- * cluster which means we do not need to care how many contiguous bytes are
- * available after the beginning of the block.
- *
- * Return the physical device block if the mapping succeeded or 0 if the block
- * is sparse or there was an error.
- *
- * Note: This is a problem if someone tries to run bmap() on $Boot system file
- * as that really is in block zero but there is nothing we can do.  bmap() is
- * just broken in that respect (just like it cannot distinguish sparse from
- * not available or error).
- */
-static sector_t ntfs_bmap(struct address_space *mapping, sector_t block)
-{
-	s64 ofs, size;
-	loff_t i_size;
-	LCN lcn;
-	unsigned long blocksize, flags;
-	ntfs_inode *ni = NTFS_I(mapping->host);
-	ntfs_volume *vol = ni->vol;
-	unsigned delta;
-	unsigned char blocksize_bits, cluster_size_shift;
-
-	ntfs_debug("Entering for mft_no 0x%lx, logical block 0x%llx.",
-			ni->mft_no, (unsigned long long)block);
-	if (ni->type != AT_DATA || !NInoNonResident(ni) || NInoEncrypted(ni)) {
-		ntfs_error(vol->sb, "BMAP does not make sense for %s "
-				"attributes, returning 0.",
-				(ni->type != AT_DATA) ? "non-data" :
-				(!NInoNonResident(ni) ? "resident" :
-				"encrypted"));
-		return 0;
-	}
-	/* None of these can happen. */
-	BUG_ON(NInoCompressed(ni));
-	BUG_ON(NInoMstProtected(ni));
-	blocksize = vol->sb->s_blocksize;
-	blocksize_bits = vol->sb->s_blocksize_bits;
-	ofs = (s64)block << blocksize_bits;
-	read_lock_irqsave(&ni->size_lock, flags);
-	size = ni->initialized_size;
-	i_size = i_size_read(VFS_I(ni));
-	read_unlock_irqrestore(&ni->size_lock, flags);
-	/*
-	 * If the offset is outside the initialized size or the block straddles
-	 * the initialized size then pretend it is a hole unless the
-	 * initialized size equals the file size.
-	 */
-	if (unlikely(ofs >= size || (ofs + blocksize > size && size < i_size)))
-		goto hole;
-	cluster_size_shift = vol->cluster_size_bits;
-	down_read(&ni->runlist.lock);
-	lcn = ntfs_attr_vcn_to_lcn_nolock(ni, ofs >> cluster_size_shift, false);
-	up_read(&ni->runlist.lock);
-	if (unlikely(lcn < LCN_HOLE)) {
-		/*
-		 * Step down to an integer to avoid gcc doing a long long
-		 * comparision in the switch when we know @lcn is between
-		 * LCN_HOLE and LCN_EIO (i.e. -1 to -5).
-		 *
-		 * Otherwise older gcc (at least on some architectures) will
-		 * try to use __cmpdi2() which is of course not available in
-		 * the kernel.
-		 */
-		switch ((int)lcn) {
-		case LCN_ENOENT:
-			/*
-			 * If the offset is out of bounds then pretend it is a
-			 * hole.
-			 */
-			goto hole;
-		case LCN_ENOMEM:
-			ntfs_error(vol->sb, "Not enough memory to complete "
-					"mapping for inode 0x%lx.  "
-					"Returning 0.", ni->mft_no);
-			break;
-		default:
-			ntfs_error(vol->sb, "Failed to complete mapping for "
-					"inode 0x%lx.  Run chkdsk.  "
-					"Returning 0.", ni->mft_no);
-			break;
-		}
-		return 0;
-	}
-	if (lcn < 0) {
-		/* It is a hole. */
-hole:
-		ntfs_debug("Done (returning hole).");
-		return 0;
-	}
-	/*
-	 * The block is really allocated and fullfils all our criteria.
-	 * Convert the cluster to units of block size and return the result.
-	 */
-	delta = ofs & vol->cluster_size_mask;
-	if (unlikely(sizeof(block) < sizeof(lcn))) {
-		block = lcn = ((lcn << cluster_size_shift) + delta) >>
-				blocksize_bits;
-		/* If the block number was truncated return 0. */
-		if (unlikely(block != lcn)) {
-			ntfs_error(vol->sb, "Physical block 0x%llx is too "
-					"large to be returned, returning 0.",
-					(long long)lcn);
-			return 0;
-		}
-	} else
-		block = ((lcn << cluster_size_shift) + delta) >>
-				blocksize_bits;
-	ntfs_debug("Done (returning block 0x%llx).", (unsigned long long)lcn);
-	return block;
-}
-
-/*
- * ntfs_normal_aops - address space operations for normal inodes and attributes
- *
- * Note these are not used for compressed or mst protected inodes and
- * attributes.
- */
-const struct address_space_operations ntfs_normal_aops = {
-	.read_folio	= ntfs_read_folio,
-#ifdef NTFS_RW
-	.writepage	= ntfs_writepage,
-	.dirty_folio	= block_dirty_folio,
-#endif /* NTFS_RW */
-	.bmap		= ntfs_bmap,
-	.migrate_folio	= buffer_migrate_folio,
-	.is_partially_uptodate = block_is_partially_uptodate,
-	.error_remove_folio = generic_error_remove_folio,
-};
-
-/*
- * ntfs_compressed_aops - address space operations for compressed inodes
- */
-const struct address_space_operations ntfs_compressed_aops = {
-	.read_folio	= ntfs_read_folio,
-#ifdef NTFS_RW
-	.writepage	= ntfs_writepage,
-	.dirty_folio	= block_dirty_folio,
-#endif /* NTFS_RW */
-	.migrate_folio	= buffer_migrate_folio,
-	.is_partially_uptodate = block_is_partially_uptodate,
-	.error_remove_folio = generic_error_remove_folio,
-};
-
-/*
- * ntfs_mst_aops - general address space operations for mst protecteed inodes
- *			  and attributes
- */
-const struct address_space_operations ntfs_mst_aops = {
-	.read_folio	= ntfs_read_folio,	/* Fill page with data. */
-#ifdef NTFS_RW
-	.writepage	= ntfs_writepage,	/* Write dirty page to disk. */
-	.dirty_folio	= filemap_dirty_folio,
-#endif /* NTFS_RW */
-	.migrate_folio	= buffer_migrate_folio,
-	.is_partially_uptodate	= block_is_partially_uptodate,
-	.error_remove_folio = generic_error_remove_folio,
-};
-
-#ifdef NTFS_RW
-
-/**
- * mark_ntfs_record_dirty - mark an ntfs record dirty
- * @page:	page containing the ntfs record to mark dirty
- * @ofs:	byte offset within @page at which the ntfs record begins
- *
- * Set the buffers and the page in which the ntfs record is located dirty.
- *
- * The latter also marks the vfs inode the ntfs record belongs to dirty
- * (I_DIRTY_PAGES only).
- *
- * If the page does not have buffers, we create them and set them uptodate.
- * The page may not be locked which is why we need to handle the buffers under
- * the mapping->i_private_lock.  Once the buffers are marked dirty we no longer
- * need the lock since try_to_free_buffers() does not free dirty buffers.
- */
-void mark_ntfs_record_dirty(struct page *page, const unsigned int ofs) {
-	struct address_space *mapping = page->mapping;
-	ntfs_inode *ni = NTFS_I(mapping->host);
-	struct buffer_head *bh, *head, *buffers_to_free = NULL;
-	unsigned int end, bh_size, bh_ofs;
-
-	BUG_ON(!PageUptodate(page));
-	end = ofs + ni->itype.index.block_size;
-	bh_size = VFS_I(ni)->i_sb->s_blocksize;
-	spin_lock(&mapping->i_private_lock);
-	if (unlikely(!page_has_buffers(page))) {
-		spin_unlock(&mapping->i_private_lock);
-		bh = head = alloc_page_buffers(page, bh_size, true);
-		spin_lock(&mapping->i_private_lock);
-		if (likely(!page_has_buffers(page))) {
-			struct buffer_head *tail;
-
-			do {
-				set_buffer_uptodate(bh);
-				tail = bh;
-				bh = bh->b_this_page;
-			} while (bh);
-			tail->b_this_page = head;
-			attach_page_private(page, head);
-		} else
-			buffers_to_free = bh;
-	}
-	bh = head = page_buffers(page);
-	BUG_ON(!bh);
-	do {
-		bh_ofs = bh_offset(bh);
-		if (bh_ofs + bh_size <= ofs)
-			continue;
-		if (unlikely(bh_ofs >= end))
-			break;
-		set_buffer_dirty(bh);
-	} while ((bh = bh->b_this_page) != head);
-	spin_unlock(&mapping->i_private_lock);
-	filemap_dirty_folio(mapping, page_folio(page));
-	if (unlikely(buffers_to_free)) {
-		do {
-			bh = buffers_to_free->b_this_page;
-			free_buffer_head(buffers_to_free);
-			buffers_to_free = bh;
-		} while (buffers_to_free);
-	}
-}
-
-#endif /* NTFS_RW */
diff --git a/fs/ntfs/aops.h b/fs/ntfs/aops.h
deleted file mode 100644
index 8d0958a149..0000000000
--- a/fs/ntfs/aops.h
+++ /dev/null
@@ -1,88 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * aops.h - Defines for NTFS kernel address space operations and page cache
- *	    handling.  Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2004 Anton Altaparmakov
- * Copyright (c) 2002 Richard Russon
- */
-
-#ifndef _LINUX_NTFS_AOPS_H
-#define _LINUX_NTFS_AOPS_H
-
-#include <linux/mm.h>
-#include <linux/highmem.h>
-#include <linux/pagemap.h>
-#include <linux/fs.h>
-
-#include "inode.h"
-
-/**
- * ntfs_unmap_page - release a page that was mapped using ntfs_map_page()
- * @page:	the page to release
- *
- * Unpin, unmap and release a page that was obtained from ntfs_map_page().
- */
-static inline void ntfs_unmap_page(struct page *page)
-{
-	kunmap(page);
-	put_page(page);
-}
-
-/**
- * ntfs_map_page - map a page into accessible memory, reading it if necessary
- * @mapping:	address space for which to obtain the page
- * @index:	index into the page cache for @mapping of the page to map
- *
- * Read a page from the page cache of the address space @mapping at position
- * @index, where @index is in units of PAGE_SIZE, and not in bytes.
- *
- * If the page is not in memory it is loaded from disk first using the
- * read_folio method defined in the address space operations of @mapping
- * and the page is added to the page cache of @mapping in the process.
- *
- * If the page belongs to an mst protected attribute and it is marked as such
- * in its ntfs inode (NInoMstProtected()) the mst fixups are applied but no
- * error checking is performed.  This means the caller has to verify whether
- * the ntfs record(s) contained in the page are valid or not using one of the
- * ntfs_is_XXXX_record{,p}() macros, where XXXX is the record type you are
- * expecting to see.  (For details of the macros, see fs/ntfs/layout.h.)
- *
- * If the page is in high memory it is mapped into memory directly addressible
- * by the kernel.
- *
- * Finally the page count is incremented, thus pinning the page into place.
- *
- * The above means that page_address(page) can be used on all pages obtained
- * with ntfs_map_page() to get the kernel virtual address of the page.
- *
- * When finished with the page, the caller has to call ntfs_unmap_page() to
- * unpin, unmap and release the page.
- *
- * Note this does not grant exclusive access. If such is desired, the caller
- * must provide it independently of the ntfs_{un}map_page() calls by using
- * a {rw_}semaphore or other means of serialization. A spin lock cannot be
- * used as ntfs_map_page() can block.
- *
- * The unlocked and uptodate page is returned on success or an encoded error
- * on failure. Caller has to test for error using the IS_ERR() macro on the
- * return value. If that evaluates to 'true', the negative error code can be
- * obtained using PTR_ERR() on the return value of ntfs_map_page().
- */
-static inline struct page *ntfs_map_page(struct address_space *mapping,
-		unsigned long index)
-{
-	struct page *page = read_mapping_page(mapping, index, NULL);
-
-	if (!IS_ERR(page))
-		kmap(page);
-	return page;
-}
-
-#ifdef NTFS_RW
-
-extern void mark_ntfs_record_dirty(struct page *page, const unsigned int ofs);
-
-#endif /* NTFS_RW */
-
-#endif /* _LINUX_NTFS_AOPS_H */
diff --git a/fs/ntfs/attrib.c b/fs/ntfs/attrib.c
deleted file mode 100644
index f79408f912..0000000000
--- a/fs/ntfs/attrib.c
+++ /dev/null
@@ -1,2624 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * attrib.c - NTFS attribute operations.  Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2012 Anton Altaparmakov and Tuxera Inc.
- * Copyright (c) 2002 Richard Russon
- */
-
-#include <linux/buffer_head.h>
-#include <linux/sched.h>
-#include <linux/slab.h>
-#include <linux/swap.h>
-#include <linux/writeback.h>
-
-#include "attrib.h"
-#include "debug.h"
-#include "layout.h"
-#include "lcnalloc.h"
-#include "malloc.h"
-#include "mft.h"
-#include "ntfs.h"
-#include "types.h"
-
-/**
- * ntfs_map_runlist_nolock - map (a part of) a runlist of an ntfs inode
- * @ni:		ntfs inode for which to map (part of) a runlist
- * @vcn:	map runlist part containing this vcn
- * @ctx:	active attribute search context if present or NULL if not
- *
- * Map the part of a runlist containing the @vcn of the ntfs inode @ni.
- *
- * If @ctx is specified, it is an active search context of @ni and its base mft
- * record.  This is needed when ntfs_map_runlist_nolock() encounters unmapped
- * runlist fragments and allows their mapping.  If you do not have the mft
- * record mapped, you can specify @ctx as NULL and ntfs_map_runlist_nolock()
- * will perform the necessary mapping and unmapping.
- *
- * Note, ntfs_map_runlist_nolock() saves the state of @ctx on entry and
- * restores it before returning.  Thus, @ctx will be left pointing to the same
- * attribute on return as on entry.  However, the actual pointers in @ctx may
- * point to different memory locations on return, so you must remember to reset
- * any cached pointers from the @ctx, i.e. after the call to
- * ntfs_map_runlist_nolock(), you will probably want to do:
- *	m = ctx->mrec;
- *	a = ctx->attr;
- * Assuming you cache ctx->attr in a variable @a of type ATTR_RECORD * and that
- * you cache ctx->mrec in a variable @m of type MFT_RECORD *.
- *
- * Return 0 on success and -errno on error.  There is one special error code
- * which is not an error as such.  This is -ENOENT.  It means that @vcn is out
- * of bounds of the runlist.
- *
- * Note the runlist can be NULL after this function returns if @vcn is zero and
- * the attribute has zero allocated size, i.e. there simply is no runlist.
- *
- * WARNING: If @ctx is supplied, regardless of whether success or failure is
- *	    returned, you need to check IS_ERR(@ctx->mrec) and if 'true' the @ctx
- *	    is no longer valid, i.e. you need to either call
- *	    ntfs_attr_reinit_search_ctx() or ntfs_attr_put_search_ctx() on it.
- *	    In that case PTR_ERR(@ctx->mrec) will give you the error code for
- *	    why the mapping of the old inode failed.
- *
- * Locking: - The runlist described by @ni must be locked for writing on entry
- *	      and is locked on return.  Note the runlist will be modified.
- *	    - If @ctx is NULL, the base mft record of @ni must not be mapped on
- *	      entry and it will be left unmapped on return.
- *	    - If @ctx is not NULL, the base mft record must be mapped on entry
- *	      and it will be left mapped on return.
- */
-int ntfs_map_runlist_nolock(ntfs_inode *ni, VCN vcn, ntfs_attr_search_ctx *ctx)
-{
-	VCN end_vcn;
-	unsigned long flags;
-	ntfs_inode *base_ni;
-	MFT_RECORD *m;
-	ATTR_RECORD *a;
-	runlist_element *rl;
-	struct page *put_this_page = NULL;
-	int err = 0;
-	bool ctx_is_temporary, ctx_needs_reset;
-	ntfs_attr_search_ctx old_ctx = { NULL, };
-
-	ntfs_debug("Mapping runlist part containing vcn 0x%llx.",
-			(unsigned long long)vcn);
-	if (!NInoAttr(ni))
-		base_ni = ni;
-	else
-		base_ni = ni->ext.base_ntfs_ino;
-	if (!ctx) {
-		ctx_is_temporary = ctx_needs_reset = true;
-		m = map_mft_record(base_ni);
-		if (IS_ERR(m))
-			return PTR_ERR(m);
-		ctx = ntfs_attr_get_search_ctx(base_ni, m);
-		if (unlikely(!ctx)) {
-			err = -ENOMEM;
-			goto err_out;
-		}
-	} else {
-		VCN allocated_size_vcn;
-
-		BUG_ON(IS_ERR(ctx->mrec));
-		a = ctx->attr;
-		BUG_ON(!a->non_resident);
-		ctx_is_temporary = false;
-		end_vcn = sle64_to_cpu(a->data.non_resident.highest_vcn);
-		read_lock_irqsave(&ni->size_lock, flags);
-		allocated_size_vcn = ni->allocated_size >>
-				ni->vol->cluster_size_bits;
-		read_unlock_irqrestore(&ni->size_lock, flags);
-		if (!a->data.non_resident.lowest_vcn && end_vcn <= 0)
-			end_vcn = allocated_size_vcn - 1;
-		/*
-		 * If we already have the attribute extent containing @vcn in
-		 * @ctx, no need to look it up again.  We slightly cheat in
-		 * that if vcn exceeds the allocated size, we will refuse to
-		 * map the runlist below, so there is definitely no need to get
-		 * the right attribute extent.
-		 */
-		if (vcn >= allocated_size_vcn || (a->type == ni->type &&
-				a->name_length == ni->name_len &&
-				!memcmp((u8*)a + le16_to_cpu(a->name_offset),
-				ni->name, ni->name_len) &&
-				sle64_to_cpu(a->data.non_resident.lowest_vcn)
-				<= vcn && end_vcn >= vcn))
-			ctx_needs_reset = false;
-		else {
-			/* Save the old search context. */
-			old_ctx = *ctx;
-			/*
-			 * If the currently mapped (extent) inode is not the
-			 * base inode we will unmap it when we reinitialize the
-			 * search context which means we need to get a
-			 * reference to the page containing the mapped mft
-			 * record so we do not accidentally drop changes to the
-			 * mft record when it has not been marked dirty yet.
-			 */
-			if (old_ctx.base_ntfs_ino && old_ctx.ntfs_ino !=
-					old_ctx.base_ntfs_ino) {
-				put_this_page = old_ctx.ntfs_ino->page;
-				get_page(put_this_page);
-			}
-			/*
-			 * Reinitialize the search context so we can lookup the
-			 * needed attribute extent.
-			 */
-			ntfs_attr_reinit_search_ctx(ctx);
-			ctx_needs_reset = true;
-		}
-	}
-	if (ctx_needs_reset) {
-		err = ntfs_attr_lookup(ni->type, ni->name, ni->name_len,
-				CASE_SENSITIVE, vcn, NULL, 0, ctx);
-		if (unlikely(err)) {
-			if (err == -ENOENT)
-				err = -EIO;
-			goto err_out;
-		}
-		BUG_ON(!ctx->attr->non_resident);
-	}
-	a = ctx->attr;
-	/*
-	 * Only decompress the mapping pairs if @vcn is inside it.  Otherwise
-	 * we get into problems when we try to map an out of bounds vcn because
-	 * we then try to map the already mapped runlist fragment and
-	 * ntfs_mapping_pairs_decompress() fails.
-	 */
-	end_vcn = sle64_to_cpu(a->data.non_resident.highest_vcn) + 1;
-	if (unlikely(vcn && vcn >= end_vcn)) {
-		err = -ENOENT;
-		goto err_out;
-	}
-	rl = ntfs_mapping_pairs_decompress(ni->vol, a, ni->runlist.rl);
-	if (IS_ERR(rl))
-		err = PTR_ERR(rl);
-	else
-		ni->runlist.rl = rl;
-err_out:
-	if (ctx_is_temporary) {
-		if (likely(ctx))
-			ntfs_attr_put_search_ctx(ctx);
-		unmap_mft_record(base_ni);
-	} else if (ctx_needs_reset) {
-		/*
-		 * If there is no attribute list, restoring the search context
-		 * is accomplished simply by copying the saved context back over
-		 * the caller supplied context.  If there is an attribute list,
-		 * things are more complicated as we need to deal with mapping
-		 * of mft records and resulting potential changes in pointers.
-		 */
-		if (NInoAttrList(base_ni)) {
-			/*
-			 * If the currently mapped (extent) inode is not the
-			 * one we had before, we need to unmap it and map the
-			 * old one.
-			 */
-			if (ctx->ntfs_ino != old_ctx.ntfs_ino) {
-				/*
-				 * If the currently mapped inode is not the
-				 * base inode, unmap it.
-				 */
-				if (ctx->base_ntfs_ino && ctx->ntfs_ino !=
-						ctx->base_ntfs_ino) {
-					unmap_extent_mft_record(ctx->ntfs_ino);
-					ctx->mrec = ctx->base_mrec;
-					BUG_ON(!ctx->mrec);
-				}
-				/*
-				 * If the old mapped inode is not the base
-				 * inode, map it.
-				 */
-				if (old_ctx.base_ntfs_ino &&
-						old_ctx.ntfs_ino !=
-						old_ctx.base_ntfs_ino) {
-retry_map:
-					ctx->mrec = map_mft_record(
-							old_ctx.ntfs_ino);
-					/*
-					 * Something bad has happened.  If out
-					 * of memory retry till it succeeds.
-					 * Any other errors are fatal and we
-					 * return the error code in ctx->mrec.
-					 * Let the caller deal with it...  We
-					 * just need to fudge things so the
-					 * caller can reinit and/or put the
-					 * search context safely.
-					 */
-					if (IS_ERR(ctx->mrec)) {
-						if (PTR_ERR(ctx->mrec) ==
-								-ENOMEM) {
-							schedule();
-							goto retry_map;
-						} else
-							old_ctx.ntfs_ino =
-								old_ctx.
-								base_ntfs_ino;
-					}
-				}
-			}
-			/* Update the changed pointers in the saved context. */
-			if (ctx->mrec != old_ctx.mrec) {
-				if (!IS_ERR(ctx->mrec))
-					old_ctx.attr = (ATTR_RECORD*)(
-							(u8*)ctx->mrec +
-							((u8*)old_ctx.attr -
-							(u8*)old_ctx.mrec));
-				old_ctx.mrec = ctx->mrec;
-			}
-		}
-		/* Restore the search context to the saved one. */
-		*ctx = old_ctx;
-		/*
-		 * We drop the reference on the page we took earlier.  In the
-		 * case that IS_ERR(ctx->mrec) is true this means we might lose
-		 * some changes to the mft record that had been made between
-		 * the last time it was marked dirty/written out and now.  This
-		 * at this stage is not a problem as the mapping error is fatal
-		 * enough that the mft record cannot be written out anyway and
-		 * the caller is very likely to shutdown the whole inode
-		 * immediately and mark the volume dirty for chkdsk to pick up
-		 * the pieces anyway.
-		 */
-		if (put_this_page)
-			put_page(put_this_page);
-	}
-	return err;
-}
-
-/**
- * ntfs_map_runlist - map (a part of) a runlist of an ntfs inode
- * @ni:		ntfs inode for which to map (part of) a runlist
- * @vcn:	map runlist part containing this vcn
- *
- * Map the part of a runlist containing the @vcn of the ntfs inode @ni.
- *
- * Return 0 on success and -errno on error.  There is one special error code
- * which is not an error as such.  This is -ENOENT.  It means that @vcn is out
- * of bounds of the runlist.
- *
- * Locking: - The runlist must be unlocked on entry and is unlocked on return.
- *	    - This function takes the runlist lock for writing and may modify
- *	      the runlist.
- */
-int ntfs_map_runlist(ntfs_inode *ni, VCN vcn)
-{
-	int err = 0;
-
-	down_write(&ni->runlist.lock);
-	/* Make sure someone else didn't do the work while we were sleeping. */
-	if (likely(ntfs_rl_vcn_to_lcn(ni->runlist.rl, vcn) <=
-			LCN_RL_NOT_MAPPED))
-		err = ntfs_map_runlist_nolock(ni, vcn, NULL);
-	up_write(&ni->runlist.lock);
-	return err;
-}
-
-/**
- * ntfs_attr_vcn_to_lcn_nolock - convert a vcn into a lcn given an ntfs inode
- * @ni:			ntfs inode of the attribute whose runlist to search
- * @vcn:		vcn to convert
- * @write_locked:	true if the runlist is locked for writing
- *
- * Find the virtual cluster number @vcn in the runlist of the ntfs attribute
- * described by the ntfs inode @ni and return the corresponding logical cluster
- * number (lcn).
- *
- * If the @vcn is not mapped yet, the attempt is made to map the attribute
- * extent containing the @vcn and the vcn to lcn conversion is retried.
- *
- * If @write_locked is true the caller has locked the runlist for writing and
- * if false for reading.
- *
- * Since lcns must be >= 0, we use negative return codes with special meaning:
- *
- * Return code	Meaning / Description
- * ==========================================
- *  LCN_HOLE	Hole / not allocated on disk.
- *  LCN_ENOENT	There is no such vcn in the runlist, i.e. @vcn is out of bounds.
- *  LCN_ENOMEM	Not enough memory to map runlist.
- *  LCN_EIO	Critical error (runlist/file is corrupt, i/o error, etc).
- *
- * Locking: - The runlist must be locked on entry and is left locked on return.
- *	    - If @write_locked is 'false', i.e. the runlist is locked for reading,
- *	      the lock may be dropped inside the function so you cannot rely on
- *	      the runlist still being the same when this function returns.
- */
-LCN ntfs_attr_vcn_to_lcn_nolock(ntfs_inode *ni, const VCN vcn,
-		const bool write_locked)
-{
-	LCN lcn;
-	unsigned long flags;
-	bool is_retry = false;
-
-	BUG_ON(!ni);
-	ntfs_debug("Entering for i_ino 0x%lx, vcn 0x%llx, %s_locked.",
-			ni->mft_no, (unsigned long long)vcn,
-			write_locked ? "write" : "read");
-	BUG_ON(!NInoNonResident(ni));
-	BUG_ON(vcn < 0);
-	if (!ni->runlist.rl) {
-		read_lock_irqsave(&ni->size_lock, flags);
-		if (!ni->allocated_size) {
-			read_unlock_irqrestore(&ni->size_lock, flags);
-			return LCN_ENOENT;
-		}
-		read_unlock_irqrestore(&ni->size_lock, flags);
-	}
-retry_remap:
-	/* Convert vcn to lcn.  If that fails map the runlist and retry once. */
-	lcn = ntfs_rl_vcn_to_lcn(ni->runlist.rl, vcn);
-	if (likely(lcn >= LCN_HOLE)) {
-		ntfs_debug("Done, lcn 0x%llx.", (long long)lcn);
-		return lcn;
-	}
-	if (lcn != LCN_RL_NOT_MAPPED) {
-		if (lcn != LCN_ENOENT)
-			lcn = LCN_EIO;
-	} else if (!is_retry) {
-		int err;
-
-		if (!write_locked) {
-			up_read(&ni->runlist.lock);
-			down_write(&ni->runlist.lock);
-			if (unlikely(ntfs_rl_vcn_to_lcn(ni->runlist.rl, vcn) !=
-					LCN_RL_NOT_MAPPED)) {
-				up_write(&ni->runlist.lock);
-				down_read(&ni->runlist.lock);
-				goto retry_remap;
-			}
-		}
-		err = ntfs_map_runlist_nolock(ni, vcn, NULL);
-		if (!write_locked) {
-			up_write(&ni->runlist.lock);
-			down_read(&ni->runlist.lock);
-		}
-		if (likely(!err)) {
-			is_retry = true;
-			goto retry_remap;
-		}
-		if (err == -ENOENT)
-			lcn = LCN_ENOENT;
-		else if (err == -ENOMEM)
-			lcn = LCN_ENOMEM;
-		else
-			lcn = LCN_EIO;
-	}
-	if (lcn != LCN_ENOENT)
-		ntfs_error(ni->vol->sb, "Failed with error code %lli.",
-				(long long)lcn);
-	return lcn;
-}
-
-/**
- * ntfs_attr_find_vcn_nolock - find a vcn in the runlist of an ntfs inode
- * @ni:		ntfs inode describing the runlist to search
- * @vcn:	vcn to find
- * @ctx:	active attribute search context if present or NULL if not
- *
- * Find the virtual cluster number @vcn in the runlist described by the ntfs
- * inode @ni and return the address of the runlist element containing the @vcn.
- *
- * If the @vcn is not mapped yet, the attempt is made to map the attribute
- * extent containing the @vcn and the vcn to lcn conversion is retried.
- *
- * If @ctx is specified, it is an active search context of @ni and its base mft
- * record.  This is needed when ntfs_attr_find_vcn_nolock() encounters unmapped
- * runlist fragments and allows their mapping.  If you do not have the mft
- * record mapped, you can specify @ctx as NULL and ntfs_attr_find_vcn_nolock()
- * will perform the necessary mapping and unmapping.
- *
- * Note, ntfs_attr_find_vcn_nolock() saves the state of @ctx on entry and
- * restores it before returning.  Thus, @ctx will be left pointing to the same
- * attribute on return as on entry.  However, the actual pointers in @ctx may
- * point to different memory locations on return, so you must remember to reset
- * any cached pointers from the @ctx, i.e. after the call to
- * ntfs_attr_find_vcn_nolock(), you will probably want to do:
- *	m = ctx->mrec;
- *	a = ctx->attr;
- * Assuming you cache ctx->attr in a variable @a of type ATTR_RECORD * and that
- * you cache ctx->mrec in a variable @m of type MFT_RECORD *.
- * Note you need to distinguish between the lcn of the returned runlist element
- * being >= 0 and LCN_HOLE.  In the later case you have to return zeroes on
- * read and allocate clusters on write.
- *
- * Return the runlist element containing the @vcn on success and
- * ERR_PTR(-errno) on error.  You need to test the return value with IS_ERR()
- * to decide if the return is success or failure and PTR_ERR() to get to the
- * error code if IS_ERR() is true.
- *
- * The possible error return codes are:
- *	-ENOENT - No such vcn in the runlist, i.e. @vcn is out of bounds.
- *	-ENOMEM - Not enough memory to map runlist.
- *	-EIO	- Critical error (runlist/file is corrupt, i/o error, etc).
- *
- * WARNING: If @ctx is supplied, regardless of whether success or failure is
- *	    returned, you need to check IS_ERR(@ctx->mrec) and if 'true' the @ctx
- *	    is no longer valid, i.e. you need to either call
- *	    ntfs_attr_reinit_search_ctx() or ntfs_attr_put_search_ctx() on it.
- *	    In that case PTR_ERR(@ctx->mrec) will give you the error code for
- *	    why the mapping of the old inode failed.
- *
- * Locking: - The runlist described by @ni must be locked for writing on entry
- *	      and is locked on return.  Note the runlist may be modified when
- *	      needed runlist fragments need to be mapped.
- *	    - If @ctx is NULL, the base mft record of @ni must not be mapped on
- *	      entry and it will be left unmapped on return.
- *	    - If @ctx is not NULL, the base mft record must be mapped on entry
- *	      and it will be left mapped on return.
- */
-runlist_element *ntfs_attr_find_vcn_nolock(ntfs_inode *ni, const VCN vcn,
-		ntfs_attr_search_ctx *ctx)
-{
-	unsigned long flags;
-	runlist_element *rl;
-	int err = 0;
-	bool is_retry = false;
-
-	BUG_ON(!ni);
-	ntfs_debug("Entering for i_ino 0x%lx, vcn 0x%llx, with%s ctx.",
-			ni->mft_no, (unsigned long long)vcn, ctx ? "" : "out");
-	BUG_ON(!NInoNonResident(ni));
-	BUG_ON(vcn < 0);
-	if (!ni->runlist.rl) {
-		read_lock_irqsave(&ni->size_lock, flags);
-		if (!ni->allocated_size) {
-			read_unlock_irqrestore(&ni->size_lock, flags);
-			return ERR_PTR(-ENOENT);
-		}
-		read_unlock_irqrestore(&ni->size_lock, flags);
-	}
-retry_remap:
-	rl = ni->runlist.rl;
-	if (likely(rl && vcn >= rl[0].vcn)) {
-		while (likely(rl->length)) {
-			if (unlikely(vcn < rl[1].vcn)) {
-				if (likely(rl->lcn >= LCN_HOLE)) {
-					ntfs_debug("Done.");
-					return rl;
-				}
-				break;
-			}
-			rl++;
-		}
-		if (likely(rl->lcn != LCN_RL_NOT_MAPPED)) {
-			if (likely(rl->lcn == LCN_ENOENT))
-				err = -ENOENT;
-			else
-				err = -EIO;
-		}
-	}
-	if (!err && !is_retry) {
-		/*
-		 * If the search context is invalid we cannot map the unmapped
-		 * region.
-		 */
-		if (IS_ERR(ctx->mrec))
-			err = PTR_ERR(ctx->mrec);
-		else {
-			/*
-			 * The @vcn is in an unmapped region, map the runlist
-			 * and retry.
-			 */
-			err = ntfs_map_runlist_nolock(ni, vcn, ctx);
-			if (likely(!err)) {
-				is_retry = true;
-				goto retry_remap;
-			}
-		}
-		if (err == -EINVAL)
-			err = -EIO;
-	} else if (!err)
-		err = -EIO;
-	if (err != -ENOENT)
-		ntfs_error(ni->vol->sb, "Failed with error code %i.", err);
-	return ERR_PTR(err);
-}
-
-/**
- * ntfs_attr_find - find (next) attribute in mft record
- * @type:	attribute type to find
- * @name:	attribute name to find (optional, i.e. NULL means don't care)
- * @name_len:	attribute name length (only needed if @name present)
- * @ic:		IGNORE_CASE or CASE_SENSITIVE (ignored if @name not present)
- * @val:	attribute value to find (optional, resident attributes only)
- * @val_len:	attribute value length
- * @ctx:	search context with mft record and attribute to search from
- *
- * You should not need to call this function directly.  Use ntfs_attr_lookup()
- * instead.
- *
- * ntfs_attr_find() takes a search context @ctx as parameter and searches the
- * mft record specified by @ctx->mrec, beginning at @ctx->attr, for an
- * attribute of @type, optionally @name and @val.
- *
- * If the attribute is found, ntfs_attr_find() returns 0 and @ctx->attr will
- * point to the found attribute.
- *
- * If the attribute is not found, ntfs_attr_find() returns -ENOENT and
- * @ctx->attr will point to the attribute before which the attribute being
- * searched for would need to be inserted if such an action were to be desired.
- *
- * On actual error, ntfs_attr_find() returns -EIO.  In this case @ctx->attr is
- * undefined and in particular do not rely on it not changing.
- *
- * If @ctx->is_first is 'true', the search begins with @ctx->attr itself.  If it
- * is 'false', the search begins after @ctx->attr.
- *
- * If @ic is IGNORE_CASE, the @name comparisson is not case sensitive and
- * @ctx->ntfs_ino must be set to the ntfs inode to which the mft record
- * @ctx->mrec belongs.  This is so we can get at the ntfs volume and hence at
- * the upcase table.  If @ic is CASE_SENSITIVE, the comparison is case
- * sensitive.  When @name is present, @name_len is the @name length in Unicode
- * characters.
- *
- * If @name is not present (NULL), we assume that the unnamed attribute is
- * being searched for.
- *
- * Finally, the resident attribute value @val is looked for, if present.  If
- * @val is not present (NULL), @val_len is ignored.
- *
- * ntfs_attr_find() only searches the specified mft record and it ignores the
- * presence of an attribute list attribute (unless it is the one being searched
- * for, obviously).  If you need to take attribute lists into consideration,
- * use ntfs_attr_lookup() instead (see below).  This also means that you cannot
- * use ntfs_attr_find() to search for extent records of non-resident
- * attributes, as extents with lowest_vcn != 0 are usually described by the
- * attribute list attribute only. - Note that it is possible that the first
- * extent is only in the attribute list while the last extent is in the base
- * mft record, so do not rely on being able to find the first extent in the
- * base mft record.
- *
- * Warning: Never use @val when looking for attribute types which can be
- *	    non-resident as this most likely will result in a crash!
- */
-static int ntfs_attr_find(const ATTR_TYPE type, const ntfschar *name,
-		const u32 name_len, const IGNORE_CASE_BOOL ic,
-		const u8 *val, const u32 val_len, ntfs_attr_search_ctx *ctx)
-{
-	ATTR_RECORD *a;
-	ntfs_volume *vol = ctx->ntfs_ino->vol;
-	ntfschar *upcase = vol->upcase;
-	u32 upcase_len = vol->upcase_len;
-
-	/*
-	 * Iterate over attributes in mft record starting at @ctx->attr, or the
-	 * attribute following that, if @ctx->is_first is 'true'.
-	 */
-	if (ctx->is_first) {
-		a = ctx->attr;
-		ctx->is_first = false;
-	} else
-		a = (ATTR_RECORD*)((u8*)ctx->attr +
-				le32_to_cpu(ctx->attr->length));
-	for (;;	a = (ATTR_RECORD*)((u8*)a + le32_to_cpu(a->length))) {
-		u8 *mrec_end = (u8 *)ctx->mrec +
-		               le32_to_cpu(ctx->mrec->bytes_allocated);
-		u8 *name_end;
-
-		/* check whether ATTR_RECORD wrap */
-		if ((u8 *)a < (u8 *)ctx->mrec)
-			break;
-
-		/* check whether Attribute Record Header is within bounds */
-		if ((u8 *)a > mrec_end ||
-		    (u8 *)a + sizeof(ATTR_RECORD) > mrec_end)
-			break;
-
-		/* check whether ATTR_RECORD's name is within bounds */
-		name_end = (u8 *)a + le16_to_cpu(a->name_offset) +
-			   a->name_length * sizeof(ntfschar);
-		if (name_end > mrec_end)
-			break;
-
-		ctx->attr = a;
-		if (unlikely(le32_to_cpu(a->type) > le32_to_cpu(type) ||
-				a->type == AT_END))
-			return -ENOENT;
-		if (unlikely(!a->length))
-			break;
-
-		/* check whether ATTR_RECORD's length wrap */
-		if ((u8 *)a + le32_to_cpu(a->length) < (u8 *)a)
-			break;
-		/* check whether ATTR_RECORD's length is within bounds */
-		if ((u8 *)a + le32_to_cpu(a->length) > mrec_end)
-			break;
-
-		if (a->type != type)
-			continue;
-		/*
-		 * If @name is present, compare the two names.  If @name is
-		 * missing, assume we want an unnamed attribute.
-		 */
-		if (!name) {
-			/* The search failed if the found attribute is named. */
-			if (a->name_length)
-				return -ENOENT;
-		} else if (!ntfs_are_names_equal(name, name_len,
-			    (ntfschar*)((u8*)a + le16_to_cpu(a->name_offset)),
-			    a->name_length, ic, upcase, upcase_len)) {
-			register int rc;
-
-			rc = ntfs_collate_names(name, name_len,
-					(ntfschar*)((u8*)a +
-					le16_to_cpu(a->name_offset)),
-					a->name_length, 1, IGNORE_CASE,
-					upcase, upcase_len);
-			/*
-			 * If @name collates before a->name, there is no
-			 * matching attribute.
-			 */
-			if (rc == -1)
-				return -ENOENT;
-			/* If the strings are not equal, continue search. */
-			if (rc)
-				continue;
-			rc = ntfs_collate_names(name, name_len,
-					(ntfschar*)((u8*)a +
-					le16_to_cpu(a->name_offset)),
-					a->name_length, 1, CASE_SENSITIVE,
-					upcase, upcase_len);
-			if (rc == -1)
-				return -ENOENT;
-			if (rc)
-				continue;
-		}
-		/*
-		 * The names match or @name not present and attribute is
-		 * unnamed.  If no @val specified, we have found the attribute
-		 * and are done.
-		 */
-		if (!val)
-			return 0;
-		/* @val is present; compare values. */
-		else {
-			register int rc;
-
-			rc = memcmp(val, (u8*)a + le16_to_cpu(
-					a->data.resident.value_offset),
-					min_t(u32, val_len, le32_to_cpu(
-					a->data.resident.value_length)));
-			/*
-			 * If @val collates before the current attribute's
-			 * value, there is no matching attribute.
-			 */
-			if (!rc) {
-				register u32 avl;
-
-				avl = le32_to_cpu(
-						a->data.resident.value_length);
-				if (val_len == avl)
-					return 0;
-				if (val_len < avl)
-					return -ENOENT;
-			} else if (rc < 0)
-				return -ENOENT;
-		}
-	}
-	ntfs_error(vol->sb, "Inode is corrupt.  Run chkdsk.");
-	NVolSetErrors(vol);
-	return -EIO;
-}
-
-/**
- * load_attribute_list - load an attribute list into memory
- * @vol:		ntfs volume from which to read
- * @runlist:		runlist of the attribute list
- * @al_start:		destination buffer
- * @size:		size of the destination buffer in bytes
- * @initialized_size:	initialized size of the attribute list
- *
- * Walk the runlist @runlist and load all clusters from it copying them into
- * the linear buffer @al. The maximum number of bytes copied to @al is @size
- * bytes. Note, @size does not need to be a multiple of the cluster size. If
- * @initialized_size is less than @size, the region in @al between
- * @initialized_size and @size will be zeroed and not read from disk.
- *
- * Return 0 on success or -errno on error.
- */
-int load_attribute_list(ntfs_volume *vol, runlist *runlist, u8 *al_start,
-		const s64 size, const s64 initialized_size)
-{
-	LCN lcn;
-	u8 *al = al_start;
-	u8 *al_end = al + initialized_size;
-	runlist_element *rl;
-	struct buffer_head *bh;
-	struct super_block *sb;
-	unsigned long block_size;
-	unsigned long block, max_block;
-	int err = 0;
-	unsigned char block_size_bits;
-
-	ntfs_debug("Entering.");
-	if (!vol || !runlist || !al || size <= 0 || initialized_size < 0 ||
-			initialized_size > size)
-		return -EINVAL;
-	if (!initialized_size) {
-		memset(al, 0, size);
-		return 0;
-	}
-	sb = vol->sb;
-	block_size = sb->s_blocksize;
-	block_size_bits = sb->s_blocksize_bits;
-	down_read(&runlist->lock);
-	rl = runlist->rl;
-	if (!rl) {
-		ntfs_error(sb, "Cannot read attribute list since runlist is "
-				"missing.");
-		goto err_out;	
-	}
-	/* Read all clusters specified by the runlist one run at a time. */
-	while (rl->length) {
-		lcn = ntfs_rl_vcn_to_lcn(rl, rl->vcn);
-		ntfs_debug("Reading vcn = 0x%llx, lcn = 0x%llx.",
-				(unsigned long long)rl->vcn,
-				(unsigned long long)lcn);
-		/* The attribute list cannot be sparse. */
-		if (lcn < 0) {
-			ntfs_error(sb, "ntfs_rl_vcn_to_lcn() failed.  Cannot "
-					"read attribute list.");
-			goto err_out;
-		}
-		block = lcn << vol->cluster_size_bits >> block_size_bits;
-		/* Read the run from device in chunks of block_size bytes. */
-		max_block = block + (rl->length << vol->cluster_size_bits >>
-				block_size_bits);
-		ntfs_debug("max_block = 0x%lx.", max_block);
-		do {
-			ntfs_debug("Reading block = 0x%lx.", block);
-			bh = sb_bread(sb, block);
-			if (!bh) {
-				ntfs_error(sb, "sb_bread() failed. Cannot "
-						"read attribute list.");
-				goto err_out;
-			}
-			if (al + block_size >= al_end)
-				goto do_final;
-			memcpy(al, bh->b_data, block_size);
-			brelse(bh);
-			al += block_size;
-		} while (++block < max_block);
-		rl++;
-	}
-	if (initialized_size < size) {
-initialize:
-		memset(al_start + initialized_size, 0, size - initialized_size);
-	}
-done:
-	up_read(&runlist->lock);
-	return err;
-do_final:
-	if (al < al_end) {
-		/*
-		 * Partial block.
-		 *
-		 * Note: The attribute list can be smaller than its allocation
-		 * by multiple clusters.  This has been encountered by at least
-		 * two people running Windows XP, thus we cannot do any
-		 * truncation sanity checking here. (AIA)
-		 */
-		memcpy(al, bh->b_data, al_end - al);
-		brelse(bh);
-		if (initialized_size < size)
-			goto initialize;
-		goto done;
-	}
-	brelse(bh);
-	/* Real overflow! */
-	ntfs_error(sb, "Attribute list buffer overflow. Read attribute list "
-			"is truncated.");
-err_out:
-	err = -EIO;
-	goto done;
-}
-
-/**
- * ntfs_external_attr_find - find an attribute in the attribute list of an inode
- * @type:	attribute type to find
- * @name:	attribute name to find (optional, i.e. NULL means don't care)
- * @name_len:	attribute name length (only needed if @name present)
- * @ic:		IGNORE_CASE or CASE_SENSITIVE (ignored if @name not present)
- * @lowest_vcn:	lowest vcn to find (optional, non-resident attributes only)
- * @val:	attribute value to find (optional, resident attributes only)
- * @val_len:	attribute value length
- * @ctx:	search context with mft record and attribute to search from
- *
- * You should not need to call this function directly.  Use ntfs_attr_lookup()
- * instead.
- *
- * Find an attribute by searching the attribute list for the corresponding
- * attribute list entry.  Having found the entry, map the mft record if the
- * attribute is in a different mft record/inode, ntfs_attr_find() the attribute
- * in there and return it.
- *
- * On first search @ctx->ntfs_ino must be the base mft record and @ctx must
- * have been obtained from a call to ntfs_attr_get_search_ctx().  On subsequent
- * calls @ctx->ntfs_ino can be any extent inode, too (@ctx->base_ntfs_ino is
- * then the base inode).
- *
- * After finishing with the attribute/mft record you need to call
- * ntfs_attr_put_search_ctx() to cleanup the search context (unmapping any
- * mapped inodes, etc).
- *
- * If the attribute is found, ntfs_external_attr_find() returns 0 and
- * @ctx->attr will point to the found attribute.  @ctx->mrec will point to the
- * mft record in which @ctx->attr is located and @ctx->al_entry will point to
- * the attribute list entry for the attribute.
- *
- * If the attribute is not found, ntfs_external_attr_find() returns -ENOENT and
- * @ctx->attr will point to the attribute in the base mft record before which
- * the attribute being searched for would need to be inserted if such an action
- * were to be desired.  @ctx->mrec will point to the mft record in which
- * @ctx->attr is located and @ctx->al_entry will point to the attribute list
- * entry of the attribute before which the attribute being searched for would
- * need to be inserted if such an action were to be desired.
- *
- * Thus to insert the not found attribute, one wants to add the attribute to
- * @ctx->mrec (the base mft record) and if there is not enough space, the
- * attribute should be placed in a newly allocated extent mft record.  The
- * attribute list entry for the inserted attribute should be inserted in the
- * attribute list attribute at @ctx->al_entry.
- *
- * On actual error, ntfs_external_attr_find() returns -EIO.  In this case
- * @ctx->attr is undefined and in particular do not rely on it not changing.
- */
-static int ntfs_external_attr_find(const ATTR_TYPE type,
-		const ntfschar *name, const u32 name_len,
-		const IGNORE_CASE_BOOL ic, const VCN lowest_vcn,
-		const u8 *val, const u32 val_len, ntfs_attr_search_ctx *ctx)
-{
-	ntfs_inode *base_ni, *ni;
-	ntfs_volume *vol;
-	ATTR_LIST_ENTRY *al_entry, *next_al_entry;
-	u8 *al_start, *al_end;
-	ATTR_RECORD *a;
-	ntfschar *al_name;
-	u32 al_name_len;
-	int err = 0;
-	static const char *es = " Unmount and run chkdsk.";
-
-	ni = ctx->ntfs_ino;
-	base_ni = ctx->base_ntfs_ino;
-	ntfs_debug("Entering for inode 0x%lx, type 0x%x.", ni->mft_no, type);
-	if (!base_ni) {
-		/* First call happens with the base mft record. */
-		base_ni = ctx->base_ntfs_ino = ctx->ntfs_ino;
-		ctx->base_mrec = ctx->mrec;
-	}
-	if (ni == base_ni)
-		ctx->base_attr = ctx->attr;
-	if (type == AT_END)
-		goto not_found;
-	vol = base_ni->vol;
-	al_start = base_ni->attr_list;
-	al_end = al_start + base_ni->attr_list_size;
-	if (!ctx->al_entry)
-		ctx->al_entry = (ATTR_LIST_ENTRY*)al_start;
-	/*
-	 * Iterate over entries in attribute list starting at @ctx->al_entry,
-	 * or the entry following that, if @ctx->is_first is 'true'.
-	 */
-	if (ctx->is_first) {
-		al_entry = ctx->al_entry;
-		ctx->is_first = false;
-	} else
-		al_entry = (ATTR_LIST_ENTRY*)((u8*)ctx->al_entry +
-				le16_to_cpu(ctx->al_entry->length));
-	for (;; al_entry = next_al_entry) {
-		/* Out of bounds check. */
-		if ((u8*)al_entry < base_ni->attr_list ||
-				(u8*)al_entry > al_end)
-			break;	/* Inode is corrupt. */
-		ctx->al_entry = al_entry;
-		/* Catch the end of the attribute list. */
-		if ((u8*)al_entry == al_end)
-			goto not_found;
-		if (!al_entry->length)
-			break;
-		if ((u8*)al_entry + 6 > al_end || (u8*)al_entry +
-				le16_to_cpu(al_entry->length) > al_end)
-			break;
-		next_al_entry = (ATTR_LIST_ENTRY*)((u8*)al_entry +
-				le16_to_cpu(al_entry->length));
-		if (le32_to_cpu(al_entry->type) > le32_to_cpu(type))
-			goto not_found;
-		if (type != al_entry->type)
-			continue;
-		/*
-		 * If @name is present, compare the two names.  If @name is
-		 * missing, assume we want an unnamed attribute.
-		 */
-		al_name_len = al_entry->name_length;
-		al_name = (ntfschar*)((u8*)al_entry + al_entry->name_offset);
-		if (!name) {
-			if (al_name_len)
-				goto not_found;
-		} else if (!ntfs_are_names_equal(al_name, al_name_len, name,
-				name_len, ic, vol->upcase, vol->upcase_len)) {
-			register int rc;
-
-			rc = ntfs_collate_names(name, name_len, al_name,
-					al_name_len, 1, IGNORE_CASE,
-					vol->upcase, vol->upcase_len);
-			/*
-			 * If @name collates before al_name, there is no
-			 * matching attribute.
-			 */
-			if (rc == -1)
-				goto not_found;
-			/* If the strings are not equal, continue search. */
-			if (rc)
-				continue;
-			/*
-			 * FIXME: Reverse engineering showed 0, IGNORE_CASE but
-			 * that is inconsistent with ntfs_attr_find().  The
-			 * subsequent rc checks were also different.  Perhaps I
-			 * made a mistake in one of the two.  Need to recheck
-			 * which is correct or at least see what is going on...
-			 * (AIA)
-			 */
-			rc = ntfs_collate_names(name, name_len, al_name,
-					al_name_len, 1, CASE_SENSITIVE,
-					vol->upcase, vol->upcase_len);
-			if (rc == -1)
-				goto not_found;
-			if (rc)
-				continue;
-		}
-		/*
-		 * The names match or @name not present and attribute is
-		 * unnamed.  Now check @lowest_vcn.  Continue search if the
-		 * next attribute list entry still fits @lowest_vcn.  Otherwise
-		 * we have reached the right one or the search has failed.
-		 */
-		if (lowest_vcn && (u8*)next_al_entry >= al_start	    &&
-				(u8*)next_al_entry + 6 < al_end		    &&
-				(u8*)next_al_entry + le16_to_cpu(
-					next_al_entry->length) <= al_end    &&
-				sle64_to_cpu(next_al_entry->lowest_vcn) <=
-					lowest_vcn			    &&
-				next_al_entry->type == al_entry->type	    &&
-				next_al_entry->name_length == al_name_len   &&
-				ntfs_are_names_equal((ntfschar*)((u8*)
-					next_al_entry +
-					next_al_entry->name_offset),
-					next_al_entry->name_length,
-					al_name, al_name_len, CASE_SENSITIVE,
-					vol->upcase, vol->upcase_len))
-			continue;
-		if (MREF_LE(al_entry->mft_reference) == ni->mft_no) {
-			if (MSEQNO_LE(al_entry->mft_reference) != ni->seq_no) {
-				ntfs_error(vol->sb, "Found stale mft "
-						"reference in attribute list "
-						"of base inode 0x%lx.%s",
-						base_ni->mft_no, es);
-				err = -EIO;
-				break;
-			}
-		} else { /* Mft references do not match. */
-			/* If there is a mapped record unmap it first. */
-			if (ni != base_ni)
-				unmap_extent_mft_record(ni);
-			/* Do we want the base record back? */
-			if (MREF_LE(al_entry->mft_reference) ==
-					base_ni->mft_no) {
-				ni = ctx->ntfs_ino = base_ni;
-				ctx->mrec = ctx->base_mrec;
-			} else {
-				/* We want an extent record. */
-				ctx->mrec = map_extent_mft_record(base_ni,
-						le64_to_cpu(
-						al_entry->mft_reference), &ni);
-				if (IS_ERR(ctx->mrec)) {
-					ntfs_error(vol->sb, "Failed to map "
-							"extent mft record "
-							"0x%lx of base inode "
-							"0x%lx.%s",
-							MREF_LE(al_entry->
-							mft_reference),
-							base_ni->mft_no, es);
-					err = PTR_ERR(ctx->mrec);
-					if (err == -ENOENT)
-						err = -EIO;
-					/* Cause @ctx to be sanitized below. */
-					ni = NULL;
-					break;
-				}
-				ctx->ntfs_ino = ni;
-			}
-			ctx->attr = (ATTR_RECORD*)((u8*)ctx->mrec +
-					le16_to_cpu(ctx->mrec->attrs_offset));
-		}
-		/*
-		 * ctx->vfs_ino, ctx->mrec, and ctx->attr now point to the
-		 * mft record containing the attribute represented by the
-		 * current al_entry.
-		 */
-		/*
-		 * We could call into ntfs_attr_find() to find the right
-		 * attribute in this mft record but this would be less
-		 * efficient and not quite accurate as ntfs_attr_find() ignores
-		 * the attribute instance numbers for example which become
-		 * important when one plays with attribute lists.  Also,
-		 * because a proper match has been found in the attribute list
-		 * entry above, the comparison can now be optimized.  So it is
-		 * worth re-implementing a simplified ntfs_attr_find() here.
-		 */
-		a = ctx->attr;
-		/*
-		 * Use a manual loop so we can still use break and continue
-		 * with the same meanings as above.
-		 */
-do_next_attr_loop:
-		if ((u8*)a < (u8*)ctx->mrec || (u8*)a > (u8*)ctx->mrec +
-				le32_to_cpu(ctx->mrec->bytes_allocated))
-			break;
-		if (a->type == AT_END)
-			break;
-		if (!a->length)
-			break;
-		if (al_entry->instance != a->instance)
-			goto do_next_attr;
-		/*
-		 * If the type and/or the name are mismatched between the
-		 * attribute list entry and the attribute record, there is
-		 * corruption so we break and return error EIO.
-		 */
-		if (al_entry->type != a->type)
-			break;
-		if (!ntfs_are_names_equal((ntfschar*)((u8*)a +
-				le16_to_cpu(a->name_offset)), a->name_length,
-				al_name, al_name_len, CASE_SENSITIVE,
-				vol->upcase, vol->upcase_len))
-			break;
-		ctx->attr = a;
-		/*
-		 * If no @val specified or @val specified and it matches, we
-		 * have found it!
-		 */
-		if (!val || (!a->non_resident && le32_to_cpu(
-				a->data.resident.value_length) == val_len &&
-				!memcmp((u8*)a +
-				le16_to_cpu(a->data.resident.value_offset),
-				val, val_len))) {
-			ntfs_debug("Done, found.");
-			return 0;
-		}
-do_next_attr:
-		/* Proceed to the next attribute in the current mft record. */
-		a = (ATTR_RECORD*)((u8*)a + le32_to_cpu(a->length));
-		goto do_next_attr_loop;
-	}
-	if (!err) {
-		ntfs_error(vol->sb, "Base inode 0x%lx contains corrupt "
-				"attribute list attribute.%s", base_ni->mft_no,
-				es);
-		err = -EIO;
-	}
-	if (ni != base_ni) {
-		if (ni)
-			unmap_extent_mft_record(ni);
-		ctx->ntfs_ino = base_ni;
-		ctx->mrec = ctx->base_mrec;
-		ctx->attr = ctx->base_attr;
-	}
-	if (err != -ENOMEM)
-		NVolSetErrors(vol);
-	return err;
-not_found:
-	/*
-	 * If we were looking for AT_END, we reset the search context @ctx and
-	 * use ntfs_attr_find() to seek to the end of the base mft record.
-	 */
-	if (type == AT_END) {
-		ntfs_attr_reinit_search_ctx(ctx);
-		return ntfs_attr_find(AT_END, name, name_len, ic, val, val_len,
-				ctx);
-	}
-	/*
-	 * The attribute was not found.  Before we return, we want to ensure
-	 * @ctx->mrec and @ctx->attr indicate the position at which the
-	 * attribute should be inserted in the base mft record.  Since we also
-	 * want to preserve @ctx->al_entry we cannot reinitialize the search
-	 * context using ntfs_attr_reinit_search_ctx() as this would set
-	 * @ctx->al_entry to NULL.  Thus we do the necessary bits manually (see
-	 * ntfs_attr_init_search_ctx() below).  Note, we _only_ preserve
-	 * @ctx->al_entry as the remaining fields (base_*) are identical to
-	 * their non base_ counterparts and we cannot set @ctx->base_attr
-	 * correctly yet as we do not know what @ctx->attr will be set to by
-	 * the call to ntfs_attr_find() below.
-	 */
-	if (ni != base_ni)
-		unmap_extent_mft_record(ni);
-	ctx->mrec = ctx->base_mrec;
-	ctx->attr = (ATTR_RECORD*)((u8*)ctx->mrec +
-			le16_to_cpu(ctx->mrec->attrs_offset));
-	ctx->is_first = true;
-	ctx->ntfs_ino = base_ni;
-	ctx->base_ntfs_ino = NULL;
-	ctx->base_mrec = NULL;
-	ctx->base_attr = NULL;
-	/*
-	 * In case there are multiple matches in the base mft record, need to
-	 * keep enumerating until we get an attribute not found response (or
-	 * another error), otherwise we would keep returning the same attribute
-	 * over and over again and all programs using us for enumeration would
-	 * lock up in a tight loop.
-	 */
-	do {
-		err = ntfs_attr_find(type, name, name_len, ic, val, val_len,
-				ctx);
-	} while (!err);
-	ntfs_debug("Done, not found.");
-	return err;
-}
-
-/**
- * ntfs_attr_lookup - find an attribute in an ntfs inode
- * @type:	attribute type to find
- * @name:	attribute name to find (optional, i.e. NULL means don't care)
- * @name_len:	attribute name length (only needed if @name present)
- * @ic:		IGNORE_CASE or CASE_SENSITIVE (ignored if @name not present)
- * @lowest_vcn:	lowest vcn to find (optional, non-resident attributes only)
- * @val:	attribute value to find (optional, resident attributes only)
- * @val_len:	attribute value length
- * @ctx:	search context with mft record and attribute to search from
- *
- * Find an attribute in an ntfs inode.  On first search @ctx->ntfs_ino must
- * be the base mft record and @ctx must have been obtained from a call to
- * ntfs_attr_get_search_ctx().
- *
- * This function transparently handles attribute lists and @ctx is used to
- * continue searches where they were left off at.
- *
- * After finishing with the attribute/mft record you need to call
- * ntfs_attr_put_search_ctx() to cleanup the search context (unmapping any
- * mapped inodes, etc).
- *
- * Return 0 if the search was successful and -errno if not.
- *
- * When 0, @ctx->attr is the found attribute and it is in mft record
- * @ctx->mrec.  If an attribute list attribute is present, @ctx->al_entry is
- * the attribute list entry of the found attribute.
- *
- * When -ENOENT, @ctx->attr is the attribute which collates just after the
- * attribute being searched for, i.e. if one wants to add the attribute to the
- * mft record this is the correct place to insert it into.  If an attribute
- * list attribute is present, @ctx->al_entry is the attribute list entry which
- * collates just after the attribute list entry of the attribute being searched
- * for, i.e. if one wants to add the attribute to the mft record this is the
- * correct place to insert its attribute list entry into.
- *
- * When -errno != -ENOENT, an error occurred during the lookup.  @ctx->attr is
- * then undefined and in particular you should not rely on it not changing.
- */
-int ntfs_attr_lookup(const ATTR_TYPE type, const ntfschar *name,
-		const u32 name_len, const IGNORE_CASE_BOOL ic,
-		const VCN lowest_vcn, const u8 *val, const u32 val_len,
-		ntfs_attr_search_ctx *ctx)
-{
-	ntfs_inode *base_ni;
-
-	ntfs_debug("Entering.");
-	BUG_ON(IS_ERR(ctx->mrec));
-	if (ctx->base_ntfs_ino)
-		base_ni = ctx->base_ntfs_ino;
-	else
-		base_ni = ctx->ntfs_ino;
-	/* Sanity check, just for debugging really. */
-	BUG_ON(!base_ni);
-	if (!NInoAttrList(base_ni) || type == AT_ATTRIBUTE_LIST)
-		return ntfs_attr_find(type, name, name_len, ic, val, val_len,
-				ctx);
-	return ntfs_external_attr_find(type, name, name_len, ic, lowest_vcn,
-			val, val_len, ctx);
-}
-
-/**
- * ntfs_attr_init_search_ctx - initialize an attribute search context
- * @ctx:	attribute search context to initialize
- * @ni:		ntfs inode with which to initialize the search context
- * @mrec:	mft record with which to initialize the search context
- *
- * Initialize the attribute search context @ctx with @ni and @mrec.
- */
-static inline void ntfs_attr_init_search_ctx(ntfs_attr_search_ctx *ctx,
-		ntfs_inode *ni, MFT_RECORD *mrec)
-{
-	*ctx = (ntfs_attr_search_ctx) {
-		.mrec = mrec,
-		/* Sanity checks are performed elsewhere. */
-		.attr = (ATTR_RECORD*)((u8*)mrec +
-				le16_to_cpu(mrec->attrs_offset)),
-		.is_first = true,
-		.ntfs_ino = ni,
-	};
-}
-
-/**
- * ntfs_attr_reinit_search_ctx - reinitialize an attribute search context
- * @ctx:	attribute search context to reinitialize
- *
- * Reinitialize the attribute search context @ctx, unmapping an associated
- * extent mft record if present, and initialize the search context again.
- *
- * This is used when a search for a new attribute is being started to reset
- * the search context to the beginning.
- */
-void ntfs_attr_reinit_search_ctx(ntfs_attr_search_ctx *ctx)
-{
-	if (likely(!ctx->base_ntfs_ino)) {
-		/* No attribute list. */
-		ctx->is_first = true;
-		/* Sanity checks are performed elsewhere. */
-		ctx->attr = (ATTR_RECORD*)((u8*)ctx->mrec +
-				le16_to_cpu(ctx->mrec->attrs_offset));
-		/*
-		 * This needs resetting due to ntfs_external_attr_find() which
-		 * can leave it set despite having zeroed ctx->base_ntfs_ino.
-		 */
-		ctx->al_entry = NULL;
-		return;
-	} /* Attribute list. */
-	if (ctx->ntfs_ino != ctx->base_ntfs_ino)
-		unmap_extent_mft_record(ctx->ntfs_ino);
-	ntfs_attr_init_search_ctx(ctx, ctx->base_ntfs_ino, ctx->base_mrec);
-	return;
-}
-
-/**
- * ntfs_attr_get_search_ctx - allocate/initialize a new attribute search context
- * @ni:		ntfs inode with which to initialize the search context
- * @mrec:	mft record with which to initialize the search context
- *
- * Allocate a new attribute search context, initialize it with @ni and @mrec,
- * and return it. Return NULL if allocation failed.
- */
-ntfs_attr_search_ctx *ntfs_attr_get_search_ctx(ntfs_inode *ni, MFT_RECORD *mrec)
-{
-	ntfs_attr_search_ctx *ctx;
-
-	ctx = kmem_cache_alloc(ntfs_attr_ctx_cache, GFP_NOFS);
-	if (ctx)
-		ntfs_attr_init_search_ctx(ctx, ni, mrec);
-	return ctx;
-}
-
-/**
- * ntfs_attr_put_search_ctx - release an attribute search context
- * @ctx:	attribute search context to free
- *
- * Release the attribute search context @ctx, unmapping an associated extent
- * mft record if present.
- */
-void ntfs_attr_put_search_ctx(ntfs_attr_search_ctx *ctx)
-{
-	if (ctx->base_ntfs_ino && ctx->ntfs_ino != ctx->base_ntfs_ino)
-		unmap_extent_mft_record(ctx->ntfs_ino);
-	kmem_cache_free(ntfs_attr_ctx_cache, ctx);
-	return;
-}
-
-#ifdef NTFS_RW
-
-/**
- * ntfs_attr_find_in_attrdef - find an attribute in the $AttrDef system file
- * @vol:	ntfs volume to which the attribute belongs
- * @type:	attribute type which to find
- *
- * Search for the attribute definition record corresponding to the attribute
- * @type in the $AttrDef system file.
- *
- * Return the attribute type definition record if found and NULL if not found.
- */
-static ATTR_DEF *ntfs_attr_find_in_attrdef(const ntfs_volume *vol,
-		const ATTR_TYPE type)
-{
-	ATTR_DEF *ad;
-
-	BUG_ON(!vol->attrdef);
-	BUG_ON(!type);
-	for (ad = vol->attrdef; (u8*)ad - (u8*)vol->attrdef <
-			vol->attrdef_size && ad->type; ++ad) {
-		/* We have not found it yet, carry on searching. */
-		if (likely(le32_to_cpu(ad->type) < le32_to_cpu(type)))
-			continue;
-		/* We found the attribute; return it. */
-		if (likely(ad->type == type))
-			return ad;
-		/* We have gone too far already.  No point in continuing. */
-		break;
-	}
-	/* Attribute not found. */
-	ntfs_debug("Attribute type 0x%x not found in $AttrDef.",
-			le32_to_cpu(type));
-	return NULL;
-}
-
-/**
- * ntfs_attr_size_bounds_check - check a size of an attribute type for validity
- * @vol:	ntfs volume to which the attribute belongs
- * @type:	attribute type which to check
- * @size:	size which to check
- *
- * Check whether the @size in bytes is valid for an attribute of @type on the
- * ntfs volume @vol.  This information is obtained from $AttrDef system file.
- *
- * Return 0 if valid, -ERANGE if not valid, or -ENOENT if the attribute is not
- * listed in $AttrDef.
- */
-int ntfs_attr_size_bounds_check(const ntfs_volume *vol, const ATTR_TYPE type,
-		const s64 size)
-{
-	ATTR_DEF *ad;
-
-	BUG_ON(size < 0);
-	/*
-	 * $ATTRIBUTE_LIST has a maximum size of 256kiB, but this is not
-	 * listed in $AttrDef.
-	 */
-	if (unlikely(type == AT_ATTRIBUTE_LIST && size > 256 * 1024))
-		return -ERANGE;
-	/* Get the $AttrDef entry for the attribute @type. */
-	ad = ntfs_attr_find_in_attrdef(vol, type);
-	if (unlikely(!ad))
-		return -ENOENT;
-	/* Do the bounds check. */
-	if (((sle64_to_cpu(ad->min_size) > 0) &&
-			size < sle64_to_cpu(ad->min_size)) ||
-			((sle64_to_cpu(ad->max_size) > 0) && size >
-			sle64_to_cpu(ad->max_size)))
-		return -ERANGE;
-	return 0;
-}
-
-/**
- * ntfs_attr_can_be_non_resident - check if an attribute can be non-resident
- * @vol:	ntfs volume to which the attribute belongs
- * @type:	attribute type which to check
- *
- * Check whether the attribute of @type on the ntfs volume @vol is allowed to
- * be non-resident.  This information is obtained from $AttrDef system file.
- *
- * Return 0 if the attribute is allowed to be non-resident, -EPERM if not, and
- * -ENOENT if the attribute is not listed in $AttrDef.
- */
-int ntfs_attr_can_be_non_resident(const ntfs_volume *vol, const ATTR_TYPE type)
-{
-	ATTR_DEF *ad;
-
-	/* Find the attribute definition record in $AttrDef. */
-	ad = ntfs_attr_find_in_attrdef(vol, type);
-	if (unlikely(!ad))
-		return -ENOENT;
-	/* Check the flags and return the result. */
-	if (ad->flags & ATTR_DEF_RESIDENT)
-		return -EPERM;
-	return 0;
-}
-
-/**
- * ntfs_attr_can_be_resident - check if an attribute can be resident
- * @vol:	ntfs volume to which the attribute belongs
- * @type:	attribute type which to check
- *
- * Check whether the attribute of @type on the ntfs volume @vol is allowed to
- * be resident.  This information is derived from our ntfs knowledge and may
- * not be completely accurate, especially when user defined attributes are
- * present.  Basically we allow everything to be resident except for index
- * allocation and $EA attributes.
- *
- * Return 0 if the attribute is allowed to be non-resident and -EPERM if not.
- *
- * Warning: In the system file $MFT the attribute $Bitmap must be non-resident
- *	    otherwise windows will not boot (blue screen of death)!  We cannot
- *	    check for this here as we do not know which inode's $Bitmap is
- *	    being asked about so the caller needs to special case this.
- */
-int ntfs_attr_can_be_resident(const ntfs_volume *vol, const ATTR_TYPE type)
-{
-	if (type == AT_INDEX_ALLOCATION)
-		return -EPERM;
-	return 0;
-}
-
-/**
- * ntfs_attr_record_resize - resize an attribute record
- * @m:		mft record containing attribute record
- * @a:		attribute record to resize
- * @new_size:	new size in bytes to which to resize the attribute record @a
- *
- * Resize the attribute record @a, i.e. the resident part of the attribute, in
- * the mft record @m to @new_size bytes.
- *
- * Return 0 on success and -errno on error.  The following error codes are
- * defined:
- *	-ENOSPC	- Not enough space in the mft record @m to perform the resize.
- *
- * Note: On error, no modifications have been performed whatsoever.
- *
- * Warning: If you make a record smaller without having copied all the data you
- *	    are interested in the data may be overwritten.
- */
-int ntfs_attr_record_resize(MFT_RECORD *m, ATTR_RECORD *a, u32 new_size)
-{
-	ntfs_debug("Entering for new_size %u.", new_size);
-	/* Align to 8 bytes if it is not already done. */
-	if (new_size & 7)
-		new_size = (new_size + 7) & ~7;
-	/* If the actual attribute length has changed, move things around. */
-	if (new_size != le32_to_cpu(a->length)) {
-		u32 new_muse = le32_to_cpu(m->bytes_in_use) -
-				le32_to_cpu(a->length) + new_size;
-		/* Not enough space in this mft record. */
-		if (new_muse > le32_to_cpu(m->bytes_allocated))
-			return -ENOSPC;
-		/* Move attributes following @a to their new location. */
-		memmove((u8*)a + new_size, (u8*)a + le32_to_cpu(a->length),
-				le32_to_cpu(m->bytes_in_use) - ((u8*)a -
-				(u8*)m) - le32_to_cpu(a->length));
-		/* Adjust @m to reflect the change in used space. */
-		m->bytes_in_use = cpu_to_le32(new_muse);
-		/* Adjust @a to reflect the new size. */
-		if (new_size >= offsetof(ATTR_REC, length) + sizeof(a->length))
-			a->length = cpu_to_le32(new_size);
-	}
-	return 0;
-}
-
-/**
- * ntfs_resident_attr_value_resize - resize the value of a resident attribute
- * @m:		mft record containing attribute record
- * @a:		attribute record whose value to resize
- * @new_size:	new size in bytes to which to resize the attribute value of @a
- *
- * Resize the value of the attribute @a in the mft record @m to @new_size bytes.
- * If the value is made bigger, the newly allocated space is cleared.
- *
- * Return 0 on success and -errno on error.  The following error codes are
- * defined:
- *	-ENOSPC	- Not enough space in the mft record @m to perform the resize.
- *
- * Note: On error, no modifications have been performed whatsoever.
- *
- * Warning: If you make a record smaller without having copied all the data you
- *	    are interested in the data may be overwritten.
- */
-int ntfs_resident_attr_value_resize(MFT_RECORD *m, ATTR_RECORD *a,
-		const u32 new_size)
-{
-	u32 old_size;
-
-	/* Resize the resident part of the attribute record. */
-	if (ntfs_attr_record_resize(m, a,
-			le16_to_cpu(a->data.resident.value_offset) + new_size))
-		return -ENOSPC;
-	/*
-	 * The resize succeeded!  If we made the attribute value bigger, clear
-	 * the area between the old size and @new_size.
-	 */
-	old_size = le32_to_cpu(a->data.resident.value_length);
-	if (new_size > old_size)
-		memset((u8*)a + le16_to_cpu(a->data.resident.value_offset) +
-				old_size, 0, new_size - old_size);
-	/* Finally update the length of the attribute value. */
-	a->data.resident.value_length = cpu_to_le32(new_size);
-	return 0;
-}
-
-/**
- * ntfs_attr_make_non_resident - convert a resident to a non-resident attribute
- * @ni:		ntfs inode describing the attribute to convert
- * @data_size:	size of the resident data to copy to the non-resident attribute
- *
- * Convert the resident ntfs attribute described by the ntfs inode @ni to a
- * non-resident one.
- *
- * @data_size must be equal to the attribute value size.  This is needed since
- * we need to know the size before we can map the mft record and our callers
- * always know it.  The reason we cannot simply read the size from the vfs
- * inode i_size is that this is not necessarily uptodate.  This happens when
- * ntfs_attr_make_non_resident() is called in the ->truncate call path(s).
- *
- * Return 0 on success and -errno on error.  The following error return codes
- * are defined:
- *	-EPERM	- The attribute is not allowed to be non-resident.
- *	-ENOMEM	- Not enough memory.
- *	-ENOSPC	- Not enough disk space.
- *	-EINVAL	- Attribute not defined on the volume.
- *	-EIO	- I/o error or other error.
- * Note that -ENOSPC is also returned in the case that there is not enough
- * space in the mft record to do the conversion.  This can happen when the mft
- * record is already very full.  The caller is responsible for trying to make
- * space in the mft record and trying again.  FIXME: Do we need a separate
- * error return code for this kind of -ENOSPC or is it always worth trying
- * again in case the attribute may then fit in a resident state so no need to
- * make it non-resident at all?  Ho-hum...  (AIA)
- *
- * NOTE to self: No changes in the attribute list are required to move from
- *		 a resident to a non-resident attribute.
- *
- * Locking: - The caller must hold i_mutex on the inode.
- */
-int ntfs_attr_make_non_resident(ntfs_inode *ni, const u32 data_size)
-{
-	s64 new_size;
-	struct inode *vi = VFS_I(ni);
-	ntfs_volume *vol = ni->vol;
-	ntfs_inode *base_ni;
-	MFT_RECORD *m;
-	ATTR_RECORD *a;
-	ntfs_attr_search_ctx *ctx;
-	struct page *page;
-	runlist_element *rl;
-	u8 *kaddr;
-	unsigned long flags;
-	int mp_size, mp_ofs, name_ofs, arec_size, err, err2;
-	u32 attr_size;
-	u8 old_res_attr_flags;
-
-	/* Check that the attribute is allowed to be non-resident. */
-	err = ntfs_attr_can_be_non_resident(vol, ni->type);
-	if (unlikely(err)) {
-		if (err == -EPERM)
-			ntfs_debug("Attribute is not allowed to be "
-					"non-resident.");
-		else
-			ntfs_debug("Attribute not defined on the NTFS "
-					"volume!");
-		return err;
-	}
-	/*
-	 * FIXME: Compressed and encrypted attributes are not supported when
-	 * writing and we should never have gotten here for them.
-	 */
-	BUG_ON(NInoCompressed(ni));
-	BUG_ON(NInoEncrypted(ni));
-	/*
-	 * The size needs to be aligned to a cluster boundary for allocation
-	 * purposes.
-	 */
-	new_size = (data_size + vol->cluster_size - 1) &
-			~(vol->cluster_size - 1);
-	if (new_size > 0) {
-		/*
-		 * Will need the page later and since the page lock nests
-		 * outside all ntfs locks, we need to get the page now.
-		 */
-		page = find_or_create_page(vi->i_mapping, 0,
-				mapping_gfp_mask(vi->i_mapping));
-		if (unlikely(!page))
-			return -ENOMEM;
-		/* Start by allocating clusters to hold the attribute value. */
-		rl = ntfs_cluster_alloc(vol, 0, new_size >>
-				vol->cluster_size_bits, -1, DATA_ZONE, true);
-		if (IS_ERR(rl)) {
-			err = PTR_ERR(rl);
-			ntfs_debug("Failed to allocate cluster%s, error code "
-					"%i.", (new_size >>
-					vol->cluster_size_bits) > 1 ? "s" : "",
-					err);
-			goto page_err_out;
-		}
-	} else {
-		rl = NULL;
-		page = NULL;
-	}
-	/* Determine the size of the mapping pairs array. */
-	mp_size = ntfs_get_size_for_mapping_pairs(vol, rl, 0, -1);
-	if (unlikely(mp_size < 0)) {
-		err = mp_size;
-		ntfs_debug("Failed to get size for mapping pairs array, error "
-				"code %i.", err);
-		goto rl_err_out;
-	}
-	down_write(&ni->runlist.lock);
-	if (!NInoAttr(ni))
-		base_ni = ni;
-	else
-		base_ni = ni->ext.base_ntfs_ino;
-	m = map_mft_record(base_ni);
-	if (IS_ERR(m)) {
-		err = PTR_ERR(m);
-		m = NULL;
-		ctx = NULL;
-		goto err_out;
-	}
-	ctx = ntfs_attr_get_search_ctx(base_ni, m);
-	if (unlikely(!ctx)) {
-		err = -ENOMEM;
-		goto err_out;
-	}
-	err = ntfs_attr_lookup(ni->type, ni->name, ni->name_len,
-			CASE_SENSITIVE, 0, NULL, 0, ctx);
-	if (unlikely(err)) {
-		if (err == -ENOENT)
-			err = -EIO;
-		goto err_out;
-	}
-	m = ctx->mrec;
-	a = ctx->attr;
-	BUG_ON(NInoNonResident(ni));
-	BUG_ON(a->non_resident);
-	/*
-	 * Calculate new offsets for the name and the mapping pairs array.
-	 */
-	if (NInoSparse(ni) || NInoCompressed(ni))
-		name_ofs = (offsetof(ATTR_REC,
-				data.non_resident.compressed_size) +
-				sizeof(a->data.non_resident.compressed_size) +
-				7) & ~7;
-	else
-		name_ofs = (offsetof(ATTR_REC,
-				data.non_resident.compressed_size) + 7) & ~7;
-	mp_ofs = (name_ofs + a->name_length * sizeof(ntfschar) + 7) & ~7;
-	/*
-	 * Determine the size of the resident part of the now non-resident
-	 * attribute record.
-	 */
-	arec_size = (mp_ofs + mp_size + 7) & ~7;
-	/*
-	 * If the page is not uptodate bring it uptodate by copying from the
-	 * attribute value.
-	 */
-	attr_size = le32_to_cpu(a->data.resident.value_length);
-	BUG_ON(attr_size != data_size);
-	if (page && !PageUptodate(page)) {
-		kaddr = kmap_atomic(page);
-		memcpy(kaddr, (u8*)a +
-				le16_to_cpu(a->data.resident.value_offset),
-				attr_size);
-		memset(kaddr + attr_size, 0, PAGE_SIZE - attr_size);
-		kunmap_atomic(kaddr);
-		flush_dcache_page(page);
-		SetPageUptodate(page);
-	}
-	/* Backup the attribute flag. */
-	old_res_attr_flags = a->data.resident.flags;
-	/* Resize the resident part of the attribute record. */
-	err = ntfs_attr_record_resize(m, a, arec_size);
-	if (unlikely(err))
-		goto err_out;
-	/*
-	 * Convert the resident part of the attribute record to describe a
-	 * non-resident attribute.
-	 */
-	a->non_resident = 1;
-	/* Move the attribute name if it exists and update the offset. */
-	if (a->name_length)
-		memmove((u8*)a + name_ofs, (u8*)a + le16_to_cpu(a->name_offset),
-				a->name_length * sizeof(ntfschar));
-	a->name_offset = cpu_to_le16(name_ofs);
-	/* Setup the fields specific to non-resident attributes. */
-	a->data.non_resident.lowest_vcn = 0;
-	a->data.non_resident.highest_vcn = cpu_to_sle64((new_size - 1) >>
-			vol->cluster_size_bits);
-	a->data.non_resident.mapping_pairs_offset = cpu_to_le16(mp_ofs);
-	memset(&a->data.non_resident.reserved, 0,
-			sizeof(a->data.non_resident.reserved));
-	a->data.non_resident.allocated_size = cpu_to_sle64(new_size);
-	a->data.non_resident.data_size =
-			a->data.non_resident.initialized_size =
-			cpu_to_sle64(attr_size);
-	if (NInoSparse(ni) || NInoCompressed(ni)) {
-		a->data.non_resident.compression_unit = 0;
-		if (NInoCompressed(ni) || vol->major_ver < 3)
-			a->data.non_resident.compression_unit = 4;
-		a->data.non_resident.compressed_size =
-				a->data.non_resident.allocated_size;
-	} else
-		a->data.non_resident.compression_unit = 0;
-	/* Generate the mapping pairs array into the attribute record. */
-	err = ntfs_mapping_pairs_build(vol, (u8*)a + mp_ofs,
-			arec_size - mp_ofs, rl, 0, -1, NULL);
-	if (unlikely(err)) {
-		ntfs_debug("Failed to build mapping pairs, error code %i.",
-				err);
-		goto undo_err_out;
-	}
-	/* Setup the in-memory attribute structure to be non-resident. */
-	ni->runlist.rl = rl;
-	write_lock_irqsave(&ni->size_lock, flags);
-	ni->allocated_size = new_size;
-	if (NInoSparse(ni) || NInoCompressed(ni)) {
-		ni->itype.compressed.size = ni->allocated_size;
-		if (a->data.non_resident.compression_unit) {
-			ni->itype.compressed.block_size = 1U << (a->data.
-					non_resident.compression_unit +
-					vol->cluster_size_bits);
-			ni->itype.compressed.block_size_bits =
-					ffs(ni->itype.compressed.block_size) -
-					1;
-			ni->itype.compressed.block_clusters = 1U <<
-					a->data.non_resident.compression_unit;
-		} else {
-			ni->itype.compressed.block_size = 0;
-			ni->itype.compressed.block_size_bits = 0;
-			ni->itype.compressed.block_clusters = 0;
-		}
-		vi->i_blocks = ni->itype.compressed.size >> 9;
-	} else
-		vi->i_blocks = ni->allocated_size >> 9;
-	write_unlock_irqrestore(&ni->size_lock, flags);
-	/*
-	 * This needs to be last since the address space operations ->read_folio
-	 * and ->writepage can run concurrently with us as they are not
-	 * serialized on i_mutex.  Note, we are not allowed to fail once we flip
-	 * this switch, which is another reason to do this last.
-	 */
-	NInoSetNonResident(ni);
-	/* Mark the mft record dirty, so it gets written back. */
-	flush_dcache_mft_record_page(ctx->ntfs_ino);
-	mark_mft_record_dirty(ctx->ntfs_ino);
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(base_ni);
-	up_write(&ni->runlist.lock);
-	if (page) {
-		set_page_dirty(page);
-		unlock_page(page);
-		put_page(page);
-	}
-	ntfs_debug("Done.");
-	return 0;
-undo_err_out:
-	/* Convert the attribute back into a resident attribute. */
-	a->non_resident = 0;
-	/* Move the attribute name if it exists and update the offset. */
-	name_ofs = (offsetof(ATTR_RECORD, data.resident.reserved) +
-			sizeof(a->data.resident.reserved) + 7) & ~7;
-	if (a->name_length)
-		memmove((u8*)a + name_ofs, (u8*)a + le16_to_cpu(a->name_offset),
-				a->name_length * sizeof(ntfschar));
-	mp_ofs = (name_ofs + a->name_length * sizeof(ntfschar) + 7) & ~7;
-	a->name_offset = cpu_to_le16(name_ofs);
-	arec_size = (mp_ofs + attr_size + 7) & ~7;
-	/* Resize the resident part of the attribute record. */
-	err2 = ntfs_attr_record_resize(m, a, arec_size);
-	if (unlikely(err2)) {
-		/*
-		 * This cannot happen (well if memory corruption is at work it
-		 * could happen in theory), but deal with it as well as we can.
-		 * If the old size is too small, truncate the attribute,
-		 * otherwise simply give it a larger allocated size.
-		 * FIXME: Should check whether chkdsk complains when the
-		 * allocated size is much bigger than the resident value size.
-		 */
-		arec_size = le32_to_cpu(a->length);
-		if ((mp_ofs + attr_size) > arec_size) {
-			err2 = attr_size;
-			attr_size = arec_size - mp_ofs;
-			ntfs_error(vol->sb, "Failed to undo partial resident "
-					"to non-resident attribute "
-					"conversion.  Truncating inode 0x%lx, "
-					"attribute type 0x%x from %i bytes to "
-					"%i bytes to maintain metadata "
-					"consistency.  THIS MEANS YOU ARE "
-					"LOSING %i BYTES DATA FROM THIS %s.",
-					vi->i_ino,
-					(unsigned)le32_to_cpu(ni->type),
-					err2, attr_size, err2 - attr_size,
-					((ni->type == AT_DATA) &&
-					!ni->name_len) ? "FILE": "ATTRIBUTE");
-			write_lock_irqsave(&ni->size_lock, flags);
-			ni->initialized_size = attr_size;
-			i_size_write(vi, attr_size);
-			write_unlock_irqrestore(&ni->size_lock, flags);
-		}
-	}
-	/* Setup the fields specific to resident attributes. */
-	a->data.resident.value_length = cpu_to_le32(attr_size);
-	a->data.resident.value_offset = cpu_to_le16(mp_ofs);
-	a->data.resident.flags = old_res_attr_flags;
-	memset(&a->data.resident.reserved, 0,
-			sizeof(a->data.resident.reserved));
-	/* Copy the data from the page back to the attribute value. */
-	if (page) {
-		kaddr = kmap_atomic(page);
-		memcpy((u8*)a + mp_ofs, kaddr, attr_size);
-		kunmap_atomic(kaddr);
-	}
-	/* Setup the allocated size in the ntfs inode in case it changed. */
-	write_lock_irqsave(&ni->size_lock, flags);
-	ni->allocated_size = arec_size - mp_ofs;
-	write_unlock_irqrestore(&ni->size_lock, flags);
-	/* Mark the mft record dirty, so it gets written back. */
-	flush_dcache_mft_record_page(ctx->ntfs_ino);
-	mark_mft_record_dirty(ctx->ntfs_ino);
-err_out:
-	if (ctx)
-		ntfs_attr_put_search_ctx(ctx);
-	if (m)
-		unmap_mft_record(base_ni);
-	ni->runlist.rl = NULL;
-	up_write(&ni->runlist.lock);
-rl_err_out:
-	if (rl) {
-		if (ntfs_cluster_free_from_rl(vol, rl) < 0) {
-			ntfs_error(vol->sb, "Failed to release allocated "
-					"cluster(s) in error code path.  Run "
-					"chkdsk to recover the lost "
-					"cluster(s).");
-			NVolSetErrors(vol);
-		}
-		ntfs_free(rl);
-page_err_out:
-		unlock_page(page);
-		put_page(page);
-	}
-	if (err == -EINVAL)
-		err = -EIO;
-	return err;
-}
-
-/**
- * ntfs_attr_extend_allocation - extend the allocated space of an attribute
- * @ni:			ntfs inode of the attribute whose allocation to extend
- * @new_alloc_size:	new size in bytes to which to extend the allocation to
- * @new_data_size:	new size in bytes to which to extend the data to
- * @data_start:		beginning of region which is required to be non-sparse
- *
- * Extend the allocated space of an attribute described by the ntfs inode @ni
- * to @new_alloc_size bytes.  If @data_start is -1, the whole extension may be
- * implemented as a hole in the file (as long as both the volume and the ntfs
- * inode @ni have sparse support enabled).  If @data_start is >= 0, then the
- * region between the old allocated size and @data_start - 1 may be made sparse
- * but the regions between @data_start and @new_alloc_size must be backed by
- * actual clusters.
- *
- * If @new_data_size is -1, it is ignored.  If it is >= 0, then the data size
- * of the attribute is extended to @new_data_size.  Note that the i_size of the
- * vfs inode is not updated.  Only the data size in the base attribute record
- * is updated.  The caller has to update i_size separately if this is required.
- * WARNING: It is a BUG() for @new_data_size to be smaller than the old data
- * size as well as for @new_data_size to be greater than @new_alloc_size.
- *
- * For resident attributes this involves resizing the attribute record and if
- * necessary moving it and/or other attributes into extent mft records and/or
- * converting the attribute to a non-resident attribute which in turn involves
- * extending the allocation of a non-resident attribute as described below.
- *
- * For non-resident attributes this involves allocating clusters in the data
- * zone on the volume (except for regions that are being made sparse) and
- * extending the run list to describe the allocated clusters as well as
- * updating the mapping pairs array of the attribute.  This in turn involves
- * resizing the attribute record and if necessary moving it and/or other
- * attributes into extent mft records and/or splitting the attribute record
- * into multiple extent attribute records.
- *
- * Also, the attribute list attribute is updated if present and in some of the
- * above cases (the ones where extent mft records/attributes come into play),
- * an attribute list attribute is created if not already present.
- *
- * Return the new allocated size on success and -errno on error.  In the case
- * that an error is encountered but a partial extension at least up to
- * @data_start (if present) is possible, the allocation is partially extended
- * and this is returned.  This means the caller must check the returned size to
- * determine if the extension was partial.  If @data_start is -1 then partial
- * allocations are not performed.
- *
- * WARNING: Do not call ntfs_attr_extend_allocation() for $MFT/$DATA.
- *
- * Locking: This function takes the runlist lock of @ni for writing as well as
- * locking the mft record of the base ntfs inode.  These locks are maintained
- * throughout execution of the function.  These locks are required so that the
- * attribute can be resized safely and so that it can for example be converted
- * from resident to non-resident safely.
- *
- * TODO: At present attribute list attribute handling is not implemented.
- *
- * TODO: At present it is not safe to call this function for anything other
- * than the $DATA attribute(s) of an uncompressed and unencrypted file.
- */
-s64 ntfs_attr_extend_allocation(ntfs_inode *ni, s64 new_alloc_size,
-		const s64 new_data_size, const s64 data_start)
-{
-	VCN vcn;
-	s64 ll, allocated_size, start = data_start;
-	struct inode *vi = VFS_I(ni);
-	ntfs_volume *vol = ni->vol;
-	ntfs_inode *base_ni;
-	MFT_RECORD *m;
-	ATTR_RECORD *a;
-	ntfs_attr_search_ctx *ctx;
-	runlist_element *rl, *rl2;
-	unsigned long flags;
-	int err, mp_size;
-	u32 attr_len = 0; /* Silence stupid gcc warning. */
-	bool mp_rebuilt;
-
-#ifdef DEBUG
-	read_lock_irqsave(&ni->size_lock, flags);
-	allocated_size = ni->allocated_size;
-	read_unlock_irqrestore(&ni->size_lock, flags);
-	ntfs_debug("Entering for i_ino 0x%lx, attribute type 0x%x, "
-			"old_allocated_size 0x%llx, "
-			"new_allocated_size 0x%llx, new_data_size 0x%llx, "
-			"data_start 0x%llx.", vi->i_ino,
-			(unsigned)le32_to_cpu(ni->type),
-			(unsigned long long)allocated_size,
-			(unsigned long long)new_alloc_size,
-			(unsigned long long)new_data_size,
-			(unsigned long long)start);
-#endif
-retry_extend:
-	/*
-	 * For non-resident attributes, @start and @new_size need to be aligned
-	 * to cluster boundaries for allocation purposes.
-	 */
-	if (NInoNonResident(ni)) {
-		if (start > 0)
-			start &= ~(s64)vol->cluster_size_mask;
-		new_alloc_size = (new_alloc_size + vol->cluster_size - 1) &
-				~(s64)vol->cluster_size_mask;
-	}
-	BUG_ON(new_data_size >= 0 && new_data_size > new_alloc_size);
-	/* Check if new size is allowed in $AttrDef. */
-	err = ntfs_attr_size_bounds_check(vol, ni->type, new_alloc_size);
-	if (unlikely(err)) {
-		/* Only emit errors when the write will fail completely. */
-		read_lock_irqsave(&ni->size_lock, flags);
-		allocated_size = ni->allocated_size;
-		read_unlock_irqrestore(&ni->size_lock, flags);
-		if (start < 0 || start >= allocated_size) {
-			if (err == -ERANGE) {
-				ntfs_error(vol->sb, "Cannot extend allocation "
-						"of inode 0x%lx, attribute "
-						"type 0x%x, because the new "
-						"allocation would exceed the "
-						"maximum allowed size for "
-						"this attribute type.",
-						vi->i_ino, (unsigned)
-						le32_to_cpu(ni->type));
-			} else {
-				ntfs_error(vol->sb, "Cannot extend allocation "
-						"of inode 0x%lx, attribute "
-						"type 0x%x, because this "
-						"attribute type is not "
-						"defined on the NTFS volume.  "
-						"Possible corruption!  You "
-						"should run chkdsk!",
-						vi->i_ino, (unsigned)
-						le32_to_cpu(ni->type));
-			}
-		}
-		/* Translate error code to be POSIX conformant for write(2). */
-		if (err == -ERANGE)
-			err = -EFBIG;
-		else
-			err = -EIO;
-		return err;
-	}
-	if (!NInoAttr(ni))
-		base_ni = ni;
-	else
-		base_ni = ni->ext.base_ntfs_ino;
-	/*
-	 * We will be modifying both the runlist (if non-resident) and the mft
-	 * record so lock them both down.
-	 */
-	down_write(&ni->runlist.lock);
-	m = map_mft_record(base_ni);
-	if (IS_ERR(m)) {
-		err = PTR_ERR(m);
-		m = NULL;
-		ctx = NULL;
-		goto err_out;
-	}
-	ctx = ntfs_attr_get_search_ctx(base_ni, m);
-	if (unlikely(!ctx)) {
-		err = -ENOMEM;
-		goto err_out;
-	}
-	read_lock_irqsave(&ni->size_lock, flags);
-	allocated_size = ni->allocated_size;
-	read_unlock_irqrestore(&ni->size_lock, flags);
-	/*
-	 * If non-resident, seek to the last extent.  If resident, there is
-	 * only one extent, so seek to that.
-	 */
-	vcn = NInoNonResident(ni) ? allocated_size >> vol->cluster_size_bits :
-			0;
-	/*
-	 * Abort if someone did the work whilst we waited for the locks.  If we
-	 * just converted the attribute from resident to non-resident it is
-	 * likely that exactly this has happened already.  We cannot quite
-	 * abort if we need to update the data size.
-	 */
-	if (unlikely(new_alloc_size <= allocated_size)) {
-		ntfs_debug("Allocated size already exceeds requested size.");
-		new_alloc_size = allocated_size;
-		if (new_data_size < 0)
-			goto done;
-		/*
-		 * We want the first attribute extent so that we can update the
-		 * data size.
-		 */
-		vcn = 0;
-	}
-	err = ntfs_attr_lookup(ni->type, ni->name, ni->name_len,
-			CASE_SENSITIVE, vcn, NULL, 0, ctx);
-	if (unlikely(err)) {
-		if (err == -ENOENT)
-			err = -EIO;
-		goto err_out;
-	}
-	m = ctx->mrec;
-	a = ctx->attr;
-	/* Use goto to reduce indentation. */
-	if (a->non_resident)
-		goto do_non_resident_extend;
-	BUG_ON(NInoNonResident(ni));
-	/* The total length of the attribute value. */
-	attr_len = le32_to_cpu(a->data.resident.value_length);
-	/*
-	 * Extend the attribute record to be able to store the new attribute
-	 * size.  ntfs_attr_record_resize() will not do anything if the size is
-	 * not changing.
-	 */
-	if (new_alloc_size < vol->mft_record_size &&
-			!ntfs_attr_record_resize(m, a,
-			le16_to_cpu(a->data.resident.value_offset) +
-			new_alloc_size)) {
-		/* The resize succeeded! */
-		write_lock_irqsave(&ni->size_lock, flags);
-		ni->allocated_size = le32_to_cpu(a->length) -
-				le16_to_cpu(a->data.resident.value_offset);
-		write_unlock_irqrestore(&ni->size_lock, flags);
-		if (new_data_size >= 0) {
-			BUG_ON(new_data_size < attr_len);
-			a->data.resident.value_length =
-					cpu_to_le32((u32)new_data_size);
-		}
-		goto flush_done;
-	}
-	/*
-	 * We have to drop all the locks so we can call
-	 * ntfs_attr_make_non_resident().  This could be optimised by try-
-	 * locking the first page cache page and only if that fails dropping
-	 * the locks, locking the page, and redoing all the locking and
-	 * lookups.  While this would be a huge optimisation, it is not worth
-	 * it as this is definitely a slow code path.
-	 */
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(base_ni);
-	up_write(&ni->runlist.lock);
-	/*
-	 * Not enough space in the mft record, try to make the attribute
-	 * non-resident and if successful restart the extension process.
-	 */
-	err = ntfs_attr_make_non_resident(ni, attr_len);
-	if (likely(!err))
-		goto retry_extend;
-	/*
-	 * Could not make non-resident.  If this is due to this not being
-	 * permitted for this attribute type or there not being enough space,
-	 * try to make other attributes non-resident.  Otherwise fail.
-	 */
-	if (unlikely(err != -EPERM && err != -ENOSPC)) {
-		/* Only emit errors when the write will fail completely. */
-		read_lock_irqsave(&ni->size_lock, flags);
-		allocated_size = ni->allocated_size;
-		read_unlock_irqrestore(&ni->size_lock, flags);
-		if (start < 0 || start >= allocated_size)
-			ntfs_error(vol->sb, "Cannot extend allocation of "
-					"inode 0x%lx, attribute type 0x%x, "
-					"because the conversion from resident "
-					"to non-resident attribute failed "
-					"with error code %i.", vi->i_ino,
-					(unsigned)le32_to_cpu(ni->type), err);
-		if (err != -ENOMEM)
-			err = -EIO;
-		goto conv_err_out;
-	}
-	/* TODO: Not implemented from here, abort. */
-	read_lock_irqsave(&ni->size_lock, flags);
-	allocated_size = ni->allocated_size;
-	read_unlock_irqrestore(&ni->size_lock, flags);
-	if (start < 0 || start >= allocated_size) {
-		if (err == -ENOSPC)
-			ntfs_error(vol->sb, "Not enough space in the mft "
-					"record/on disk for the non-resident "
-					"attribute value.  This case is not "
-					"implemented yet.");
-		else /* if (err == -EPERM) */
-			ntfs_error(vol->sb, "This attribute type may not be "
-					"non-resident.  This case is not "
-					"implemented yet.");
-	}
-	err = -EOPNOTSUPP;
-	goto conv_err_out;
-#if 0
-	// TODO: Attempt to make other attributes non-resident.
-	if (!err)
-		goto do_resident_extend;
-	/*
-	 * Both the attribute list attribute and the standard information
-	 * attribute must remain in the base inode.  Thus, if this is one of
-	 * these attributes, we have to try to move other attributes out into
-	 * extent mft records instead.
-	 */
-	if (ni->type == AT_ATTRIBUTE_LIST ||
-			ni->type == AT_STANDARD_INFORMATION) {
-		// TODO: Attempt to move other attributes into extent mft
-		// records.
-		err = -EOPNOTSUPP;
-		if (!err)
-			goto do_resident_extend;
-		goto err_out;
-	}
-	// TODO: Attempt to move this attribute to an extent mft record, but
-	// only if it is not already the only attribute in an mft record in
-	// which case there would be nothing to gain.
-	err = -EOPNOTSUPP;
-	if (!err)
-		goto do_resident_extend;
-	/* There is nothing we can do to make enough space. )-: */
-	goto err_out;
-#endif
-do_non_resident_extend:
-	BUG_ON(!NInoNonResident(ni));
-	if (new_alloc_size == allocated_size) {
-		BUG_ON(vcn);
-		goto alloc_done;
-	}
-	/*
-	 * If the data starts after the end of the old allocation, this is a
-	 * $DATA attribute and sparse attributes are enabled on the volume and
-	 * for this inode, then create a sparse region between the old
-	 * allocated size and the start of the data.  Otherwise simply proceed
-	 * with filling the whole space between the old allocated size and the
-	 * new allocated size with clusters.
-	 */
-	if ((start >= 0 && start <= allocated_size) || ni->type != AT_DATA ||
-			!NVolSparseEnabled(vol) || NInoSparseDisabled(ni))
-		goto skip_sparse;
-	// TODO: This is not implemented yet.  We just fill in with real
-	// clusters for now...
-	ntfs_debug("Inserting holes is not-implemented yet.  Falling back to "
-			"allocating real clusters instead.");
-skip_sparse:
-	rl = ni->runlist.rl;
-	if (likely(rl)) {
-		/* Seek to the end of the runlist. */
-		while (rl->length)
-			rl++;
-	}
-	/* If this attribute extent is not mapped, map it now. */
-	if (unlikely(!rl || rl->lcn == LCN_RL_NOT_MAPPED ||
-			(rl->lcn == LCN_ENOENT && rl > ni->runlist.rl &&
-			(rl-1)->lcn == LCN_RL_NOT_MAPPED))) {
-		if (!rl && !allocated_size)
-			goto first_alloc;
-		rl = ntfs_mapping_pairs_decompress(vol, a, ni->runlist.rl);
-		if (IS_ERR(rl)) {
-			err = PTR_ERR(rl);
-			if (start < 0 || start >= allocated_size)
-				ntfs_error(vol->sb, "Cannot extend allocation "
-						"of inode 0x%lx, attribute "
-						"type 0x%x, because the "
-						"mapping of a runlist "
-						"fragment failed with error "
-						"code %i.", vi->i_ino,
-						(unsigned)le32_to_cpu(ni->type),
-						err);
-			if (err != -ENOMEM)
-				err = -EIO;
-			goto err_out;
-		}
-		ni->runlist.rl = rl;
-		/* Seek to the end of the runlist. */
-		while (rl->length)
-			rl++;
-	}
-	/*
-	 * We now know the runlist of the last extent is mapped and @rl is at
-	 * the end of the runlist.  We want to begin allocating clusters
-	 * starting at the last allocated cluster to reduce fragmentation.  If
-	 * there are no valid LCNs in the attribute we let the cluster
-	 * allocator choose the starting cluster.
-	 */
-	/* If the last LCN is a hole or simillar seek back to last real LCN. */
-	while (rl->lcn < 0 && rl > ni->runlist.rl)
-		rl--;
-first_alloc:
-	// FIXME: Need to implement partial allocations so at least part of the
-	// write can be performed when start >= 0.  (Needed for POSIX write(2)
-	// conformance.)
-	rl2 = ntfs_cluster_alloc(vol, allocated_size >> vol->cluster_size_bits,
-			(new_alloc_size - allocated_size) >>
-			vol->cluster_size_bits, (rl && (rl->lcn >= 0)) ?
-			rl->lcn + rl->length : -1, DATA_ZONE, true);
-	if (IS_ERR(rl2)) {
-		err = PTR_ERR(rl2);
-		if (start < 0 || start >= allocated_size)
-			ntfs_error(vol->sb, "Cannot extend allocation of "
-					"inode 0x%lx, attribute type 0x%x, "
-					"because the allocation of clusters "
-					"failed with error code %i.", vi->i_ino,
-					(unsigned)le32_to_cpu(ni->type), err);
-		if (err != -ENOMEM && err != -ENOSPC)
-			err = -EIO;
-		goto err_out;
-	}
-	rl = ntfs_runlists_merge(ni->runlist.rl, rl2);
-	if (IS_ERR(rl)) {
-		err = PTR_ERR(rl);
-		if (start < 0 || start >= allocated_size)
-			ntfs_error(vol->sb, "Cannot extend allocation of "
-					"inode 0x%lx, attribute type 0x%x, "
-					"because the runlist merge failed "
-					"with error code %i.", vi->i_ino,
-					(unsigned)le32_to_cpu(ni->type), err);
-		if (err != -ENOMEM)
-			err = -EIO;
-		if (ntfs_cluster_free_from_rl(vol, rl2)) {
-			ntfs_error(vol->sb, "Failed to release allocated "
-					"cluster(s) in error code path.  Run "
-					"chkdsk to recover the lost "
-					"cluster(s).");
-			NVolSetErrors(vol);
-		}
-		ntfs_free(rl2);
-		goto err_out;
-	}
-	ni->runlist.rl = rl;
-	ntfs_debug("Allocated 0x%llx clusters.", (long long)(new_alloc_size -
-			allocated_size) >> vol->cluster_size_bits);
-	/* Find the runlist element with which the attribute extent starts. */
-	ll = sle64_to_cpu(a->data.non_resident.lowest_vcn);
-	rl2 = ntfs_rl_find_vcn_nolock(rl, ll);
-	BUG_ON(!rl2);
-	BUG_ON(!rl2->length);
-	BUG_ON(rl2->lcn < LCN_HOLE);
-	mp_rebuilt = false;
-	/* Get the size for the new mapping pairs array for this extent. */
-	mp_size = ntfs_get_size_for_mapping_pairs(vol, rl2, ll, -1);
-	if (unlikely(mp_size <= 0)) {
-		err = mp_size;
-		if (start < 0 || start >= allocated_size)
-			ntfs_error(vol->sb, "Cannot extend allocation of "
-					"inode 0x%lx, attribute type 0x%x, "
-					"because determining the size for the "
-					"mapping pairs failed with error code "
-					"%i.", vi->i_ino,
-					(unsigned)le32_to_cpu(ni->type), err);
-		err = -EIO;
-		goto undo_alloc;
-	}
-	/* Extend the attribute record to fit the bigger mapping pairs array. */
-	attr_len = le32_to_cpu(a->length);
-	err = ntfs_attr_record_resize(m, a, mp_size +
-			le16_to_cpu(a->data.non_resident.mapping_pairs_offset));
-	if (unlikely(err)) {
-		BUG_ON(err != -ENOSPC);
-		// TODO: Deal with this by moving this extent to a new mft
-		// record or by starting a new extent in a new mft record,
-		// possibly by extending this extent partially and filling it
-		// and creating a new extent for the remainder, or by making
-		// other attributes non-resident and/or by moving other
-		// attributes out of this mft record.
-		if (start < 0 || start >= allocated_size)
-			ntfs_error(vol->sb, "Not enough space in the mft "
-					"record for the extended attribute "
-					"record.  This case is not "
-					"implemented yet.");
-		err = -EOPNOTSUPP;
-		goto undo_alloc;
-	}
-	mp_rebuilt = true;
-	/* Generate the mapping pairs array directly into the attr record. */
-	err = ntfs_mapping_pairs_build(vol, (u8*)a +
-			le16_to_cpu(a->data.non_resident.mapping_pairs_offset),
-			mp_size, rl2, ll, -1, NULL);
-	if (unlikely(err)) {
-		if (start < 0 || start >= allocated_size)
-			ntfs_error(vol->sb, "Cannot extend allocation of "
-					"inode 0x%lx, attribute type 0x%x, "
-					"because building the mapping pairs "
-					"failed with error code %i.", vi->i_ino,
-					(unsigned)le32_to_cpu(ni->type), err);
-		err = -EIO;
-		goto undo_alloc;
-	}
-	/* Update the highest_vcn. */
-	a->data.non_resident.highest_vcn = cpu_to_sle64((new_alloc_size >>
-			vol->cluster_size_bits) - 1);
-	/*
-	 * We now have extended the allocated size of the attribute.  Reflect
-	 * this in the ntfs_inode structure and the attribute record.
-	 */
-	if (a->data.non_resident.lowest_vcn) {
-		/*
-		 * We are not in the first attribute extent, switch to it, but
-		 * first ensure the changes will make it to disk later.
-		 */
-		flush_dcache_mft_record_page(ctx->ntfs_ino);
-		mark_mft_record_dirty(ctx->ntfs_ino);
-		ntfs_attr_reinit_search_ctx(ctx);
-		err = ntfs_attr_lookup(ni->type, ni->name, ni->name_len,
-				CASE_SENSITIVE, 0, NULL, 0, ctx);
-		if (unlikely(err))
-			goto restore_undo_alloc;
-		/* @m is not used any more so no need to set it. */
-		a = ctx->attr;
-	}
-	write_lock_irqsave(&ni->size_lock, flags);
-	ni->allocated_size = new_alloc_size;
-	a->data.non_resident.allocated_size = cpu_to_sle64(new_alloc_size);
-	/*
-	 * FIXME: This would fail if @ni is a directory, $MFT, or an index,
-	 * since those can have sparse/compressed set.  For example can be
-	 * set compressed even though it is not compressed itself and in that
-	 * case the bit means that files are to be created compressed in the
-	 * directory...  At present this is ok as this code is only called for
-	 * regular files, and only for their $DATA attribute(s).
-	 * FIXME: The calculation is wrong if we created a hole above.  For now
-	 * it does not matter as we never create holes.
-	 */
-	if (NInoSparse(ni) || NInoCompressed(ni)) {
-		ni->itype.compressed.size += new_alloc_size - allocated_size;
-		a->data.non_resident.compressed_size =
-				cpu_to_sle64(ni->itype.compressed.size);
-		vi->i_blocks = ni->itype.compressed.size >> 9;
-	} else
-		vi->i_blocks = new_alloc_size >> 9;
-	write_unlock_irqrestore(&ni->size_lock, flags);
-alloc_done:
-	if (new_data_size >= 0) {
-		BUG_ON(new_data_size <
-				sle64_to_cpu(a->data.non_resident.data_size));
-		a->data.non_resident.data_size = cpu_to_sle64(new_data_size);
-	}
-flush_done:
-	/* Ensure the changes make it to disk. */
-	flush_dcache_mft_record_page(ctx->ntfs_ino);
-	mark_mft_record_dirty(ctx->ntfs_ino);
-done:
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(base_ni);
-	up_write(&ni->runlist.lock);
-	ntfs_debug("Done, new_allocated_size 0x%llx.",
-			(unsigned long long)new_alloc_size);
-	return new_alloc_size;
-restore_undo_alloc:
-	if (start < 0 || start >= allocated_size)
-		ntfs_error(vol->sb, "Cannot complete extension of allocation "
-				"of inode 0x%lx, attribute type 0x%x, because "
-				"lookup of first attribute extent failed with "
-				"error code %i.", vi->i_ino,
-				(unsigned)le32_to_cpu(ni->type), err);
-	if (err == -ENOENT)
-		err = -EIO;
-	ntfs_attr_reinit_search_ctx(ctx);
-	if (ntfs_attr_lookup(ni->type, ni->name, ni->name_len, CASE_SENSITIVE,
-			allocated_size >> vol->cluster_size_bits, NULL, 0,
-			ctx)) {
-		ntfs_error(vol->sb, "Failed to find last attribute extent of "
-				"attribute in error code path.  Run chkdsk to "
-				"recover.");
-		write_lock_irqsave(&ni->size_lock, flags);
-		ni->allocated_size = new_alloc_size;
-		/*
-		 * FIXME: This would fail if @ni is a directory...  See above.
-		 * FIXME: The calculation is wrong if we created a hole above.
-		 * For now it does not matter as we never create holes.
-		 */
-		if (NInoSparse(ni) || NInoCompressed(ni)) {
-			ni->itype.compressed.size += new_alloc_size -
-					allocated_size;
-			vi->i_blocks = ni->itype.compressed.size >> 9;
-		} else
-			vi->i_blocks = new_alloc_size >> 9;
-		write_unlock_irqrestore(&ni->size_lock, flags);
-		ntfs_attr_put_search_ctx(ctx);
-		unmap_mft_record(base_ni);
-		up_write(&ni->runlist.lock);
-		/*
-		 * The only thing that is now wrong is the allocated size of the
-		 * base attribute extent which chkdsk should be able to fix.
-		 */
-		NVolSetErrors(vol);
-		return err;
-	}
-	ctx->attr->data.non_resident.highest_vcn = cpu_to_sle64(
-			(allocated_size >> vol->cluster_size_bits) - 1);
-undo_alloc:
-	ll = allocated_size >> vol->cluster_size_bits;
-	if (ntfs_cluster_free(ni, ll, -1, ctx) < 0) {
-		ntfs_error(vol->sb, "Failed to release allocated cluster(s) "
-				"in error code path.  Run chkdsk to recover "
-				"the lost cluster(s).");
-		NVolSetErrors(vol);
-	}
-	m = ctx->mrec;
-	a = ctx->attr;
-	/*
-	 * If the runlist truncation fails and/or the search context is no
-	 * longer valid, we cannot resize the attribute record or build the
-	 * mapping pairs array thus we mark the inode bad so that no access to
-	 * the freed clusters can happen.
-	 */
-	if (ntfs_rl_truncate_nolock(vol, &ni->runlist, ll) || IS_ERR(m)) {
-		ntfs_error(vol->sb, "Failed to %s in error code path.  Run "
-				"chkdsk to recover.", IS_ERR(m) ?
-				"restore attribute search context" :
-				"truncate attribute runlist");
-		NVolSetErrors(vol);
-	} else if (mp_rebuilt) {
-		if (ntfs_attr_record_resize(m, a, attr_len)) {
-			ntfs_error(vol->sb, "Failed to restore attribute "
-					"record in error code path.  Run "
-					"chkdsk to recover.");
-			NVolSetErrors(vol);
-		} else /* if (success) */ {
-			if (ntfs_mapping_pairs_build(vol, (u8*)a + le16_to_cpu(
-					a->data.non_resident.
-					mapping_pairs_offset), attr_len -
-					le16_to_cpu(a->data.non_resident.
-					mapping_pairs_offset), rl2, ll, -1,
-					NULL)) {
-				ntfs_error(vol->sb, "Failed to restore "
-						"mapping pairs array in error "
-						"code path.  Run chkdsk to "
-						"recover.");
-				NVolSetErrors(vol);
-			}
-			flush_dcache_mft_record_page(ctx->ntfs_ino);
-			mark_mft_record_dirty(ctx->ntfs_ino);
-		}
-	}
-err_out:
-	if (ctx)
-		ntfs_attr_put_search_ctx(ctx);
-	if (m)
-		unmap_mft_record(base_ni);
-	up_write(&ni->runlist.lock);
-conv_err_out:
-	ntfs_debug("Failed.  Returning error code %i.", err);
-	return err;
-}
-
-/**
- * ntfs_attr_set - fill (a part of) an attribute with a byte
- * @ni:		ntfs inode describing the attribute to fill
- * @ofs:	offset inside the attribute at which to start to fill
- * @cnt:	number of bytes to fill
- * @val:	the unsigned 8-bit value with which to fill the attribute
- *
- * Fill @cnt bytes of the attribute described by the ntfs inode @ni starting at
- * byte offset @ofs inside the attribute with the constant byte @val.
- *
- * This function is effectively like memset() applied to an ntfs attribute.
- * Note this function actually only operates on the page cache pages belonging
- * to the ntfs attribute and it marks them dirty after doing the memset().
- * Thus it relies on the vm dirty page write code paths to cause the modified
- * pages to be written to the mft record/disk.
- *
- * Return 0 on success and -errno on error.  An error code of -ESPIPE means
- * that @ofs + @cnt were outside the end of the attribute and no write was
- * performed.
- */
-int ntfs_attr_set(ntfs_inode *ni, const s64 ofs, const s64 cnt, const u8 val)
-{
-	ntfs_volume *vol = ni->vol;
-	struct address_space *mapping;
-	struct page *page;
-	u8 *kaddr;
-	pgoff_t idx, end;
-	unsigned start_ofs, end_ofs, size;
-
-	ntfs_debug("Entering for ofs 0x%llx, cnt 0x%llx, val 0x%hx.",
-			(long long)ofs, (long long)cnt, val);
-	BUG_ON(ofs < 0);
-	BUG_ON(cnt < 0);
-	if (!cnt)
-		goto done;
-	/*
-	 * FIXME: Compressed and encrypted attributes are not supported when
-	 * writing and we should never have gotten here for them.
-	 */
-	BUG_ON(NInoCompressed(ni));
-	BUG_ON(NInoEncrypted(ni));
-	mapping = VFS_I(ni)->i_mapping;
-	/* Work out the starting index and page offset. */
-	idx = ofs >> PAGE_SHIFT;
-	start_ofs = ofs & ~PAGE_MASK;
-	/* Work out the ending index and page offset. */
-	end = ofs + cnt;
-	end_ofs = end & ~PAGE_MASK;
-	/* If the end is outside the inode size return -ESPIPE. */
-	if (unlikely(end > i_size_read(VFS_I(ni)))) {
-		ntfs_error(vol->sb, "Request exceeds end of attribute.");
-		return -ESPIPE;
-	}
-	end >>= PAGE_SHIFT;
-	/* If there is a first partial page, need to do it the slow way. */
-	if (start_ofs) {
-		page = read_mapping_page(mapping, idx, NULL);
-		if (IS_ERR(page)) {
-			ntfs_error(vol->sb, "Failed to read first partial "
-					"page (error, index 0x%lx).", idx);
-			return PTR_ERR(page);
-		}
-		/*
-		 * If the last page is the same as the first page, need to
-		 * limit the write to the end offset.
-		 */
-		size = PAGE_SIZE;
-		if (idx == end)
-			size = end_ofs;
-		kaddr = kmap_atomic(page);
-		memset(kaddr + start_ofs, val, size - start_ofs);
-		flush_dcache_page(page);
-		kunmap_atomic(kaddr);
-		set_page_dirty(page);
-		put_page(page);
-		balance_dirty_pages_ratelimited(mapping);
-		cond_resched();
-		if (idx == end)
-			goto done;
-		idx++;
-	}
-	/* Do the whole pages the fast way. */
-	for (; idx < end; idx++) {
-		/* Find or create the current page.  (The page is locked.) */
-		page = grab_cache_page(mapping, idx);
-		if (unlikely(!page)) {
-			ntfs_error(vol->sb, "Insufficient memory to grab "
-					"page (index 0x%lx).", idx);
-			return -ENOMEM;
-		}
-		kaddr = kmap_atomic(page);
-		memset(kaddr, val, PAGE_SIZE);
-		flush_dcache_page(page);
-		kunmap_atomic(kaddr);
-		/*
-		 * If the page has buffers, mark them uptodate since buffer
-		 * state and not page state is definitive in 2.6 kernels.
-		 */
-		if (page_has_buffers(page)) {
-			struct buffer_head *bh, *head;
-
-			bh = head = page_buffers(page);
-			do {
-				set_buffer_uptodate(bh);
-			} while ((bh = bh->b_this_page) != head);
-		}
-		/* Now that buffers are uptodate, set the page uptodate, too. */
-		SetPageUptodate(page);
-		/*
-		 * Set the page and all its buffers dirty and mark the inode
-		 * dirty, too.  The VM will write the page later on.
-		 */
-		set_page_dirty(page);
-		/* Finally unlock and release the page. */
-		unlock_page(page);
-		put_page(page);
-		balance_dirty_pages_ratelimited(mapping);
-		cond_resched();
-	}
-	/* If there is a last partial page, need to do it the slow way. */
-	if (end_ofs) {
-		page = read_mapping_page(mapping, idx, NULL);
-		if (IS_ERR(page)) {
-			ntfs_error(vol->sb, "Failed to read last partial page "
-					"(error, index 0x%lx).", idx);
-			return PTR_ERR(page);
-		}
-		kaddr = kmap_atomic(page);
-		memset(kaddr, val, end_ofs);
-		flush_dcache_page(page);
-		kunmap_atomic(kaddr);
-		set_page_dirty(page);
-		put_page(page);
-		balance_dirty_pages_ratelimited(mapping);
-		cond_resched();
-	}
-done:
-	ntfs_debug("Done.");
-	return 0;
-}
-
-#endif /* NTFS_RW */
diff --git a/fs/ntfs/attrib.h b/fs/ntfs/attrib.h
deleted file mode 100644
index fe0890d3d0..0000000000
--- a/fs/ntfs/attrib.h
+++ /dev/null
@@ -1,102 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * attrib.h - Defines for attribute handling in NTFS Linux kernel driver.
- *	      Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2005 Anton Altaparmakov
- * Copyright (c) 2002 Richard Russon
- */
-
-#ifndef _LINUX_NTFS_ATTRIB_H
-#define _LINUX_NTFS_ATTRIB_H
-
-#include "endian.h"
-#include "types.h"
-#include "layout.h"
-#include "inode.h"
-#include "runlist.h"
-#include "volume.h"
-
-/**
- * ntfs_attr_search_ctx - used in attribute search functions
- * @mrec:	buffer containing mft record to search
- * @attr:	attribute record in @mrec where to begin/continue search
- * @is_first:	if true ntfs_attr_lookup() begins search with @attr, else after
- *
- * Structure must be initialized to zero before the first call to one of the
- * attribute search functions. Initialize @mrec to point to the mft record to
- * search, and @attr to point to the first attribute within @mrec (not necessary
- * if calling the _first() functions), and set @is_first to 'true' (not necessary
- * if calling the _first() functions).
- *
- * If @is_first is 'true', the search begins with @attr. If @is_first is 'false',
- * the search begins after @attr. This is so that, after the first call to one
- * of the search attribute functions, we can call the function again, without
- * any modification of the search context, to automagically get the next
- * matching attribute.
- */
-typedef struct {
-	MFT_RECORD *mrec;
-	ATTR_RECORD *attr;
-	bool is_first;
-	ntfs_inode *ntfs_ino;
-	ATTR_LIST_ENTRY *al_entry;
-	ntfs_inode *base_ntfs_ino;
-	MFT_RECORD *base_mrec;
-	ATTR_RECORD *base_attr;
-} ntfs_attr_search_ctx;
-
-extern int ntfs_map_runlist_nolock(ntfs_inode *ni, VCN vcn,
-		ntfs_attr_search_ctx *ctx);
-extern int ntfs_map_runlist(ntfs_inode *ni, VCN vcn);
-
-extern LCN ntfs_attr_vcn_to_lcn_nolock(ntfs_inode *ni, const VCN vcn,
-		const bool write_locked);
-
-extern runlist_element *ntfs_attr_find_vcn_nolock(ntfs_inode *ni,
-		const VCN vcn, ntfs_attr_search_ctx *ctx);
-
-int ntfs_attr_lookup(const ATTR_TYPE type, const ntfschar *name,
-		const u32 name_len, const IGNORE_CASE_BOOL ic,
-		const VCN lowest_vcn, const u8 *val, const u32 val_len,
-		ntfs_attr_search_ctx *ctx);
-
-extern int load_attribute_list(ntfs_volume *vol, runlist *rl, u8 *al_start,
-		const s64 size, const s64 initialized_size);
-
-static inline s64 ntfs_attr_size(const ATTR_RECORD *a)
-{
-	if (!a->non_resident)
-		return (s64)le32_to_cpu(a->data.resident.value_length);
-	return sle64_to_cpu(a->data.non_resident.data_size);
-}
-
-extern void ntfs_attr_reinit_search_ctx(ntfs_attr_search_ctx *ctx);
-extern ntfs_attr_search_ctx *ntfs_attr_get_search_ctx(ntfs_inode *ni,
-		MFT_RECORD *mrec);
-extern void ntfs_attr_put_search_ctx(ntfs_attr_search_ctx *ctx);
-
-#ifdef NTFS_RW
-
-extern int ntfs_attr_size_bounds_check(const ntfs_volume *vol,
-		const ATTR_TYPE type, const s64 size);
-extern int ntfs_attr_can_be_non_resident(const ntfs_volume *vol,
-		const ATTR_TYPE type);
-extern int ntfs_attr_can_be_resident(const ntfs_volume *vol,
-		const ATTR_TYPE type);
-
-extern int ntfs_attr_record_resize(MFT_RECORD *m, ATTR_RECORD *a, u32 new_size);
-extern int ntfs_resident_attr_value_resize(MFT_RECORD *m, ATTR_RECORD *a,
-		const u32 new_size);
-
-extern int ntfs_attr_make_non_resident(ntfs_inode *ni, const u32 data_size);
-
-extern s64 ntfs_attr_extend_allocation(ntfs_inode *ni, s64 new_alloc_size,
-		const s64 new_data_size, const s64 data_start);
-
-extern int ntfs_attr_set(ntfs_inode *ni, const s64 ofs, const s64 cnt,
-		const u8 val);
-
-#endif /* NTFS_RW */
-
-#endif /* _LINUX_NTFS_ATTRIB_H */
diff --git a/fs/ntfs/bitmap.c b/fs/ntfs/bitmap.c
deleted file mode 100644
index 0675b24008..0000000000
--- a/fs/ntfs/bitmap.c
+++ /dev/null
@@ -1,179 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * bitmap.c - NTFS kernel bitmap handling.  Part of the Linux-NTFS project.
- *
- * Copyright (c) 2004-2005 Anton Altaparmakov
- */
-
-#ifdef NTFS_RW
-
-#include <linux/pagemap.h>
-
-#include "bitmap.h"
-#include "debug.h"
-#include "aops.h"
-#include "ntfs.h"
-
-/**
- * __ntfs_bitmap_set_bits_in_run - set a run of bits in a bitmap to a value
- * @vi:			vfs inode describing the bitmap
- * @start_bit:		first bit to set
- * @count:		number of bits to set
- * @value:		value to set the bits to (i.e. 0 or 1)
- * @is_rollback:	if 'true' this is a rollback operation
- *
- * Set @count bits starting at bit @start_bit in the bitmap described by the
- * vfs inode @vi to @value, where @value is either 0 or 1.
- *
- * @is_rollback should always be 'false', it is for internal use to rollback
- * errors.  You probably want to use ntfs_bitmap_set_bits_in_run() instead.
- *
- * Return 0 on success and -errno on error.
- */
-int __ntfs_bitmap_set_bits_in_run(struct inode *vi, const s64 start_bit,
-		const s64 count, const u8 value, const bool is_rollback)
-{
-	s64 cnt = count;
-	pgoff_t index, end_index;
-	struct address_space *mapping;
-	struct page *page;
-	u8 *kaddr;
-	int pos, len;
-	u8 bit;
-
-	BUG_ON(!vi);
-	ntfs_debug("Entering for i_ino 0x%lx, start_bit 0x%llx, count 0x%llx, "
-			"value %u.%s", vi->i_ino, (unsigned long long)start_bit,
-			(unsigned long long)cnt, (unsigned int)value,
-			is_rollback ? " (rollback)" : "");
-	BUG_ON(start_bit < 0);
-	BUG_ON(cnt < 0);
-	BUG_ON(value > 1);
-	/*
-	 * Calculate the indices for the pages containing the first and last
-	 * bits, i.e. @start_bit and @start_bit + @cnt - 1, respectively.
-	 */
-	index = start_bit >> (3 + PAGE_SHIFT);
-	end_index = (start_bit + cnt - 1) >> (3 + PAGE_SHIFT);
-
-	/* Get the page containing the first bit (@start_bit). */
-	mapping = vi->i_mapping;
-	page = ntfs_map_page(mapping, index);
-	if (IS_ERR(page)) {
-		if (!is_rollback)
-			ntfs_error(vi->i_sb, "Failed to map first page (error "
-					"%li), aborting.", PTR_ERR(page));
-		return PTR_ERR(page);
-	}
-	kaddr = page_address(page);
-
-	/* Set @pos to the position of the byte containing @start_bit. */
-	pos = (start_bit >> 3) & ~PAGE_MASK;
-
-	/* Calculate the position of @start_bit in the first byte. */
-	bit = start_bit & 7;
-
-	/* If the first byte is partial, modify the appropriate bits in it. */
-	if (bit) {
-		u8 *byte = kaddr + pos;
-		while ((bit & 7) && cnt) {
-			cnt--;
-			if (value)
-				*byte |= 1 << bit++;
-			else
-				*byte &= ~(1 << bit++);
-		}
-		/* If we are done, unmap the page and return success. */
-		if (!cnt)
-			goto done;
-
-		/* Update @pos to the new position. */
-		pos++;
-	}
-	/*
-	 * Depending on @value, modify all remaining whole bytes in the page up
-	 * to @cnt.
-	 */
-	len = min_t(s64, cnt >> 3, PAGE_SIZE - pos);
-	memset(kaddr + pos, value ? 0xff : 0, len);
-	cnt -= len << 3;
-
-	/* Update @len to point to the first not-done byte in the page. */
-	if (cnt < 8)
-		len += pos;
-
-	/* If we are not in the last page, deal with all subsequent pages. */
-	while (index < end_index) {
-		BUG_ON(cnt <= 0);
-
-		/* Update @index and get the next page. */
-		flush_dcache_page(page);
-		set_page_dirty(page);
-		ntfs_unmap_page(page);
-		page = ntfs_map_page(mapping, ++index);
-		if (IS_ERR(page))
-			goto rollback;
-		kaddr = page_address(page);
-		/*
-		 * Depending on @value, modify all remaining whole bytes in the
-		 * page up to @cnt.
-		 */
-		len = min_t(s64, cnt >> 3, PAGE_SIZE);
-		memset(kaddr, value ? 0xff : 0, len);
-		cnt -= len << 3;
-	}
-	/*
-	 * The currently mapped page is the last one.  If the last byte is
-	 * partial, modify the appropriate bits in it.  Note, @len is the
-	 * position of the last byte inside the page.
-	 */
-	if (cnt) {
-		u8 *byte;
-
-		BUG_ON(cnt > 7);
-
-		bit = cnt;
-		byte = kaddr + len;
-		while (bit--) {
-			if (value)
-				*byte |= 1 << bit;
-			else
-				*byte &= ~(1 << bit);
-		}
-	}
-done:
-	/* We are done.  Unmap the page and return success. */
-	flush_dcache_page(page);
-	set_page_dirty(page);
-	ntfs_unmap_page(page);
-	ntfs_debug("Done.");
-	return 0;
-rollback:
-	/*
-	 * Current state:
-	 *	- no pages are mapped
-	 *	- @count - @cnt is the number of bits that have been modified
-	 */
-	if (is_rollback)
-		return PTR_ERR(page);
-	if (count != cnt)
-		pos = __ntfs_bitmap_set_bits_in_run(vi, start_bit, count - cnt,
-				value ? 0 : 1, true);
-	else
-		pos = 0;
-	if (!pos) {
-		/* Rollback was successful. */
-		ntfs_error(vi->i_sb, "Failed to map subsequent page (error "
-				"%li), aborting.", PTR_ERR(page));
-	} else {
-		/* Rollback failed. */
-		ntfs_error(vi->i_sb, "Failed to map subsequent page (error "
-				"%li) and rollback failed (error %i).  "
-				"Aborting and leaving inconsistent metadata.  "
-				"Unmount and run chkdsk.", PTR_ERR(page), pos);
-		NVolSetErrors(NTFS_SB(vi->i_sb));
-	}
-	return PTR_ERR(page);
-}
-
-#endif /* NTFS_RW */
diff --git a/fs/ntfs/bitmap.h b/fs/ntfs/bitmap.h
deleted file mode 100644
index 9dd2224ca9..0000000000
--- a/fs/ntfs/bitmap.h
+++ /dev/null
@@ -1,104 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * bitmap.h - Defines for NTFS kernel bitmap handling.  Part of the Linux-NTFS
- *	      project.
- *
- * Copyright (c) 2004 Anton Altaparmakov
- */
-
-#ifndef _LINUX_NTFS_BITMAP_H
-#define _LINUX_NTFS_BITMAP_H
-
-#ifdef NTFS_RW
-
-#include <linux/fs.h>
-
-#include "types.h"
-
-extern int __ntfs_bitmap_set_bits_in_run(struct inode *vi, const s64 start_bit,
-		const s64 count, const u8 value, const bool is_rollback);
-
-/**
- * ntfs_bitmap_set_bits_in_run - set a run of bits in a bitmap to a value
- * @vi:			vfs inode describing the bitmap
- * @start_bit:		first bit to set
- * @count:		number of bits to set
- * @value:		value to set the bits to (i.e. 0 or 1)
- *
- * Set @count bits starting at bit @start_bit in the bitmap described by the
- * vfs inode @vi to @value, where @value is either 0 or 1.
- *
- * Return 0 on success and -errno on error.
- */
-static inline int ntfs_bitmap_set_bits_in_run(struct inode *vi,
-		const s64 start_bit, const s64 count, const u8 value)
-{
-	return __ntfs_bitmap_set_bits_in_run(vi, start_bit, count, value,
-			false);
-}
-
-/**
- * ntfs_bitmap_set_run - set a run of bits in a bitmap
- * @vi:		vfs inode describing the bitmap
- * @start_bit:	first bit to set
- * @count:	number of bits to set
- *
- * Set @count bits starting at bit @start_bit in the bitmap described by the
- * vfs inode @vi.
- *
- * Return 0 on success and -errno on error.
- */
-static inline int ntfs_bitmap_set_run(struct inode *vi, const s64 start_bit,
-		const s64 count)
-{
-	return ntfs_bitmap_set_bits_in_run(vi, start_bit, count, 1);
-}
-
-/**
- * ntfs_bitmap_clear_run - clear a run of bits in a bitmap
- * @vi:		vfs inode describing the bitmap
- * @start_bit:	first bit to clear
- * @count:	number of bits to clear
- *
- * Clear @count bits starting at bit @start_bit in the bitmap described by the
- * vfs inode @vi.
- *
- * Return 0 on success and -errno on error.
- */
-static inline int ntfs_bitmap_clear_run(struct inode *vi, const s64 start_bit,
-		const s64 count)
-{
-	return ntfs_bitmap_set_bits_in_run(vi, start_bit, count, 0);
-}
-
-/**
- * ntfs_bitmap_set_bit - set a bit in a bitmap
- * @vi:		vfs inode describing the bitmap
- * @bit:	bit to set
- *
- * Set bit @bit in the bitmap described by the vfs inode @vi.
- *
- * Return 0 on success and -errno on error.
- */
-static inline int ntfs_bitmap_set_bit(struct inode *vi, const s64 bit)
-{
-	return ntfs_bitmap_set_run(vi, bit, 1);
-}
-
-/**
- * ntfs_bitmap_clear_bit - clear a bit in a bitmap
- * @vi:		vfs inode describing the bitmap
- * @bit:	bit to clear
- *
- * Clear bit @bit in the bitmap described by the vfs inode @vi.
- *
- * Return 0 on success and -errno on error.
- */
-static inline int ntfs_bitmap_clear_bit(struct inode *vi, const s64 bit)
-{
-	return ntfs_bitmap_clear_run(vi, bit, 1);
-}
-
-#endif /* NTFS_RW */
-
-#endif /* defined _LINUX_NTFS_BITMAP_H */
diff --git a/fs/ntfs/collate.c b/fs/ntfs/collate.c
deleted file mode 100644
index 3ab6ec96ab..0000000000
--- a/fs/ntfs/collate.c
+++ /dev/null
@@ -1,110 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * collate.c - NTFS kernel collation handling.  Part of the Linux-NTFS project.
- *
- * Copyright (c) 2004 Anton Altaparmakov
- */
-
-#include "collate.h"
-#include "debug.h"
-#include "ntfs.h"
-
-static int ntfs_collate_binary(ntfs_volume *vol,
-		const void *data1, const int data1_len,
-		const void *data2, const int data2_len)
-{
-	int rc;
-
-	ntfs_debug("Entering.");
-	rc = memcmp(data1, data2, min(data1_len, data2_len));
-	if (!rc && (data1_len != data2_len)) {
-		if (data1_len < data2_len)
-			rc = -1;
-		else
-			rc = 1;
-	}
-	ntfs_debug("Done, returning %i", rc);
-	return rc;
-}
-
-static int ntfs_collate_ntofs_ulong(ntfs_volume *vol,
-		const void *data1, const int data1_len,
-		const void *data2, const int data2_len)
-{
-	int rc;
-	u32 d1, d2;
-
-	ntfs_debug("Entering.");
-	// FIXME:  We don't really want to bug here.
-	BUG_ON(data1_len != data2_len);
-	BUG_ON(data1_len != 4);
-	d1 = le32_to_cpup(data1);
-	d2 = le32_to_cpup(data2);
-	if (d1 < d2)
-		rc = -1;
-	else {
-		if (d1 == d2)
-			rc = 0;
-		else
-			rc = 1;
-	}
-	ntfs_debug("Done, returning %i", rc);
-	return rc;
-}
-
-typedef int (*ntfs_collate_func_t)(ntfs_volume *, const void *, const int,
-		const void *, const int);
-
-static ntfs_collate_func_t ntfs_do_collate0x0[3] = {
-	ntfs_collate_binary,
-	NULL/*ntfs_collate_file_name*/,
-	NULL/*ntfs_collate_unicode_string*/,
-};
-
-static ntfs_collate_func_t ntfs_do_collate0x1[4] = {
-	ntfs_collate_ntofs_ulong,
-	NULL/*ntfs_collate_ntofs_sid*/,
-	NULL/*ntfs_collate_ntofs_security_hash*/,
-	NULL/*ntfs_collate_ntofs_ulongs*/,
-};
-
-/**
- * ntfs_collate - collate two data items using a specified collation rule
- * @vol:	ntfs volume to which the data items belong
- * @cr:		collation rule to use when comparing the items
- * @data1:	first data item to collate
- * @data1_len:	length in bytes of @data1
- * @data2:	second data item to collate
- * @data2_len:	length in bytes of @data2
- *
- * Collate the two data items @data1 and @data2 using the collation rule @cr
- * and return -1, 0, ir 1 if @data1 is found, respectively, to collate before,
- * to match, or to collate after @data2.
- *
- * For speed we use the collation rule @cr as an index into two tables of
- * function pointers to call the appropriate collation function.
- */
-int ntfs_collate(ntfs_volume *vol, COLLATION_RULE cr,
-		const void *data1, const int data1_len,
-		const void *data2, const int data2_len) {
-	int i;
-
-	ntfs_debug("Entering.");
-	/*
-	 * FIXME:  At the moment we only support COLLATION_BINARY and
-	 * COLLATION_NTOFS_ULONG, so we BUG() for everything else for now.
-	 */
-	BUG_ON(cr != COLLATION_BINARY && cr != COLLATION_NTOFS_ULONG);
-	i = le32_to_cpu(cr);
-	BUG_ON(i < 0);
-	if (i <= 0x02)
-		return ntfs_do_collate0x0[i](vol, data1, data1_len,
-				data2, data2_len);
-	BUG_ON(i < 0x10);
-	i -= 0x10;
-	if (likely(i <= 3))
-		return ntfs_do_collate0x1[i](vol, data1, data1_len,
-				data2, data2_len);
-	BUG();
-	return 0;
-}
diff --git a/fs/ntfs/collate.h b/fs/ntfs/collate.h
deleted file mode 100644
index f2255619b4..0000000000
--- a/fs/ntfs/collate.h
+++ /dev/null
@@ -1,36 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * collate.h - Defines for NTFS kernel collation handling.  Part of the
- *	       Linux-NTFS project.
- *
- * Copyright (c) 2004 Anton Altaparmakov
- */
-
-#ifndef _LINUX_NTFS_COLLATE_H
-#define _LINUX_NTFS_COLLATE_H
-
-#include "types.h"
-#include "volume.h"
-
-static inline bool ntfs_is_collation_rule_supported(COLLATION_RULE cr) {
-	int i;
-
-	/*
-	 * FIXME:  At the moment we only support COLLATION_BINARY and
-	 * COLLATION_NTOFS_ULONG, so we return false for everything else for
-	 * now.
-	 */
-	if (unlikely(cr != COLLATION_BINARY && cr != COLLATION_NTOFS_ULONG))
-		return false;
-	i = le32_to_cpu(cr);
-	if (likely(((i >= 0) && (i <= 0x02)) ||
-			((i >= 0x10) && (i <= 0x13))))
-		return true;
-	return false;
-}
-
-extern int ntfs_collate(ntfs_volume *vol, COLLATION_RULE cr,
-		const void *data1, const int data1_len,
-		const void *data2, const int data2_len);
-
-#endif /* _LINUX_NTFS_COLLATE_H */
diff --git a/fs/ntfs/compress.c b/fs/ntfs/compress.c
deleted file mode 100644
index 761aaa0195..0000000000
--- a/fs/ntfs/compress.c
+++ /dev/null
@@ -1,950 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * compress.c - NTFS kernel compressed attributes handling.
- *		Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2004 Anton Altaparmakov
- * Copyright (c) 2002 Richard Russon
- */
-
-#include <linux/fs.h>
-#include <linux/buffer_head.h>
-#include <linux/blkdev.h>
-#include <linux/vmalloc.h>
-#include <linux/slab.h>
-
-#include "attrib.h"
-#include "inode.h"
-#include "debug.h"
-#include "ntfs.h"
-
-/**
- * ntfs_compression_constants - enum of constants used in the compression code
- */
-typedef enum {
-	/* Token types and access mask. */
-	NTFS_SYMBOL_TOKEN	=	0,
-	NTFS_PHRASE_TOKEN	=	1,
-	NTFS_TOKEN_MASK		=	1,
-
-	/* Compression sub-block constants. */
-	NTFS_SB_SIZE_MASK	=	0x0fff,
-	NTFS_SB_SIZE		=	0x1000,
-	NTFS_SB_IS_COMPRESSED	=	0x8000,
-
-	/*
-	 * The maximum compression block size is by definition 16 * the cluster
-	 * size, with the maximum supported cluster size being 4kiB. Thus the
-	 * maximum compression buffer size is 64kiB, so we use this when
-	 * initializing the compression buffer.
-	 */
-	NTFS_MAX_CB_SIZE	= 64 * 1024,
-} ntfs_compression_constants;
-
-/*
- * ntfs_compression_buffer - one buffer for the decompression engine
- */
-static u8 *ntfs_compression_buffer;
-
-/*
- * ntfs_cb_lock - spinlock which protects ntfs_compression_buffer
- */
-static DEFINE_SPINLOCK(ntfs_cb_lock);
-
-/**
- * allocate_compression_buffers - allocate the decompression buffers
- *
- * Caller has to hold the ntfs_lock mutex.
- *
- * Return 0 on success or -ENOMEM if the allocations failed.
- */
-int allocate_compression_buffers(void)
-{
-	BUG_ON(ntfs_compression_buffer);
-
-	ntfs_compression_buffer = vmalloc(NTFS_MAX_CB_SIZE);
-	if (!ntfs_compression_buffer)
-		return -ENOMEM;
-	return 0;
-}
-
-/**
- * free_compression_buffers - free the decompression buffers
- *
- * Caller has to hold the ntfs_lock mutex.
- */
-void free_compression_buffers(void)
-{
-	BUG_ON(!ntfs_compression_buffer);
-	vfree(ntfs_compression_buffer);
-	ntfs_compression_buffer = NULL;
-}
-
-/**
- * zero_partial_compressed_page - zero out of bounds compressed page region
- */
-static void zero_partial_compressed_page(struct page *page,
-		const s64 initialized_size)
-{
-	u8 *kp = page_address(page);
-	unsigned int kp_ofs;
-
-	ntfs_debug("Zeroing page region outside initialized size.");
-	if (((s64)page->index << PAGE_SHIFT) >= initialized_size) {
-		clear_page(kp);
-		return;
-	}
-	kp_ofs = initialized_size & ~PAGE_MASK;
-	memset(kp + kp_ofs, 0, PAGE_SIZE - kp_ofs);
-	return;
-}
-
-/**
- * handle_bounds_compressed_page - test for&handle out of bounds compressed page
- */
-static inline void handle_bounds_compressed_page(struct page *page,
-		const loff_t i_size, const s64 initialized_size)
-{
-	if ((page->index >= (initialized_size >> PAGE_SHIFT)) &&
-			(initialized_size < i_size))
-		zero_partial_compressed_page(page, initialized_size);
-	return;
-}
-
-/**
- * ntfs_decompress - decompress a compression block into an array of pages
- * @dest_pages:		destination array of pages
- * @completed_pages:	scratch space to track completed pages
- * @dest_index:		current index into @dest_pages (IN/OUT)
- * @dest_ofs:		current offset within @dest_pages[@dest_index] (IN/OUT)
- * @dest_max_index:	maximum index into @dest_pages (IN)
- * @dest_max_ofs:	maximum offset within @dest_pages[@dest_max_index] (IN)
- * @xpage:		the target page (-1 if none) (IN)
- * @xpage_done:		set to 1 if xpage was completed successfully (IN/OUT)
- * @cb_start:		compression block to decompress (IN)
- * @cb_size:		size of compression block @cb_start in bytes (IN)
- * @i_size:		file size when we started the read (IN)
- * @initialized_size:	initialized file size when we started the read (IN)
- *
- * The caller must have disabled preemption. ntfs_decompress() reenables it when
- * the critical section is finished.
- *
- * This decompresses the compression block @cb_start into the array of
- * destination pages @dest_pages starting at index @dest_index into @dest_pages
- * and at offset @dest_pos into the page @dest_pages[@dest_index].
- *
- * When the page @dest_pages[@xpage] is completed, @xpage_done is set to 1.
- * If xpage is -1 or @xpage has not been completed, @xpage_done is not modified.
- *
- * @cb_start is a pointer to the compression block which needs decompressing
- * and @cb_size is the size of @cb_start in bytes (8-64kiB).
- *
- * Return 0 if success or -EOVERFLOW on error in the compressed stream.
- * @xpage_done indicates whether the target page (@dest_pages[@xpage]) was
- * completed during the decompression of the compression block (@cb_start).
- *
- * Warning: This function *REQUIRES* PAGE_SIZE >= 4096 or it will blow up
- * unpredicatbly! You have been warned!
- *
- * Note to hackers: This function may not sleep until it has finished accessing
- * the compression block @cb_start as it is a per-CPU buffer.
- */
-static int ntfs_decompress(struct page *dest_pages[], int completed_pages[],
-		int *dest_index, int *dest_ofs, const int dest_max_index,
-		const int dest_max_ofs, const int xpage, char *xpage_done,
-		u8 *const cb_start, const u32 cb_size, const loff_t i_size,
-		const s64 initialized_size)
-{
-	/*
-	 * Pointers into the compressed data, i.e. the compression block (cb),
-	 * and the therein contained sub-blocks (sb).
-	 */
-	u8 *cb_end = cb_start + cb_size; /* End of cb. */
-	u8 *cb = cb_start;	/* Current position in cb. */
-	u8 *cb_sb_start;	/* Beginning of the current sb in the cb. */
-	u8 *cb_sb_end;		/* End of current sb / beginning of next sb. */
-
-	/* Variables for uncompressed data / destination. */
-	struct page *dp;	/* Current destination page being worked on. */
-	u8 *dp_addr;		/* Current pointer into dp. */
-	u8 *dp_sb_start;	/* Start of current sub-block in dp. */
-	u8 *dp_sb_end;		/* End of current sb in dp (dp_sb_start +
-				   NTFS_SB_SIZE). */
-	u16 do_sb_start;	/* @dest_ofs when starting this sub-block. */
-	u16 do_sb_end;		/* @dest_ofs of end of this sb (do_sb_start +
-				   NTFS_SB_SIZE). */
-
-	/* Variables for tag and token parsing. */
-	u8 tag;			/* Current tag. */
-	int token;		/* Loop counter for the eight tokens in tag. */
-	int nr_completed_pages = 0;
-
-	/* Default error code. */
-	int err = -EOVERFLOW;
-
-	ntfs_debug("Entering, cb_size = 0x%x.", cb_size);
-do_next_sb:
-	ntfs_debug("Beginning sub-block at offset = 0x%zx in the cb.",
-			cb - cb_start);
-	/*
-	 * Have we reached the end of the compression block or the end of the
-	 * decompressed data?  The latter can happen for example if the current
-	 * position in the compression block is one byte before its end so the
-	 * first two checks do not detect it.
-	 */
-	if (cb == cb_end || !le16_to_cpup((le16*)cb) ||
-			(*dest_index == dest_max_index &&
-			*dest_ofs == dest_max_ofs)) {
-		int i;
-
-		ntfs_debug("Completed. Returning success (0).");
-		err = 0;
-return_error:
-		/* We can sleep from now on, so we drop lock. */
-		spin_unlock(&ntfs_cb_lock);
-		/* Second stage: finalize completed pages. */
-		if (nr_completed_pages > 0) {
-			for (i = 0; i < nr_completed_pages; i++) {
-				int di = completed_pages[i];
-
-				dp = dest_pages[di];
-				/*
-				 * If we are outside the initialized size, zero
-				 * the out of bounds page range.
-				 */
-				handle_bounds_compressed_page(dp, i_size,
-						initialized_size);
-				flush_dcache_page(dp);
-				kunmap(dp);
-				SetPageUptodate(dp);
-				unlock_page(dp);
-				if (di == xpage)
-					*xpage_done = 1;
-				else
-					put_page(dp);
-				dest_pages[di] = NULL;
-			}
-		}
-		return err;
-	}
-
-	/* Setup offsets for the current sub-block destination. */
-	do_sb_start = *dest_ofs;
-	do_sb_end = do_sb_start + NTFS_SB_SIZE;
-
-	/* Check that we are still within allowed boundaries. */
-	if (*dest_index == dest_max_index && do_sb_end > dest_max_ofs)
-		goto return_overflow;
-
-	/* Does the minimum size of a compressed sb overflow valid range? */
-	if (cb + 6 > cb_end)
-		goto return_overflow;
-
-	/* Setup the current sub-block source pointers and validate range. */
-	cb_sb_start = cb;
-	cb_sb_end = cb_sb_start + (le16_to_cpup((le16*)cb) & NTFS_SB_SIZE_MASK)
-			+ 3;
-	if (cb_sb_end > cb_end)
-		goto return_overflow;
-
-	/* Get the current destination page. */
-	dp = dest_pages[*dest_index];
-	if (!dp) {
-		/* No page present. Skip decompression of this sub-block. */
-		cb = cb_sb_end;
-
-		/* Advance destination position to next sub-block. */
-		*dest_ofs = (*dest_ofs + NTFS_SB_SIZE) & ~PAGE_MASK;
-		if (!*dest_ofs && (++*dest_index > dest_max_index))
-			goto return_overflow;
-		goto do_next_sb;
-	}
-
-	/* We have a valid destination page. Setup the destination pointers. */
-	dp_addr = (u8*)page_address(dp) + do_sb_start;
-
-	/* Now, we are ready to process the current sub-block (sb). */
-	if (!(le16_to_cpup((le16*)cb) & NTFS_SB_IS_COMPRESSED)) {
-		ntfs_debug("Found uncompressed sub-block.");
-		/* This sb is not compressed, just copy it into destination. */
-
-		/* Advance source position to first data byte. */
-		cb += 2;
-
-		/* An uncompressed sb must be full size. */
-		if (cb_sb_end - cb != NTFS_SB_SIZE)
-			goto return_overflow;
-
-		/* Copy the block and advance the source position. */
-		memcpy(dp_addr, cb, NTFS_SB_SIZE);
-		cb += NTFS_SB_SIZE;
-
-		/* Advance destination position to next sub-block. */
-		*dest_ofs += NTFS_SB_SIZE;
-		if (!(*dest_ofs &= ~PAGE_MASK)) {
-finalize_page:
-			/*
-			 * First stage: add current page index to array of
-			 * completed pages.
-			 */
-			completed_pages[nr_completed_pages++] = *dest_index;
-			if (++*dest_index > dest_max_index)
-				goto return_overflow;
-		}
-		goto do_next_sb;
-	}
-	ntfs_debug("Found compressed sub-block.");
-	/* This sb is compressed, decompress it into destination. */
-
-	/* Setup destination pointers. */
-	dp_sb_start = dp_addr;
-	dp_sb_end = dp_sb_start + NTFS_SB_SIZE;
-
-	/* Forward to the first tag in the sub-block. */
-	cb += 2;
-do_next_tag:
-	if (cb == cb_sb_end) {
-		/* Check if the decompressed sub-block was not full-length. */
-		if (dp_addr < dp_sb_end) {
-			int nr_bytes = do_sb_end - *dest_ofs;
-
-			ntfs_debug("Filling incomplete sub-block with "
-					"zeroes.");
-			/* Zero remainder and update destination position. */
-			memset(dp_addr, 0, nr_bytes);
-			*dest_ofs += nr_bytes;
-		}
-		/* We have finished the current sub-block. */
-		if (!(*dest_ofs &= ~PAGE_MASK))
-			goto finalize_page;
-		goto do_next_sb;
-	}
-
-	/* Check we are still in range. */
-	if (cb > cb_sb_end || dp_addr > dp_sb_end)
-		goto return_overflow;
-
-	/* Get the next tag and advance to first token. */
-	tag = *cb++;
-
-	/* Parse the eight tokens described by the tag. */
-	for (token = 0; token < 8; token++, tag >>= 1) {
-		u16 lg, pt, length, max_non_overlap;
-		register u16 i;
-		u8 *dp_back_addr;
-
-		/* Check if we are done / still in range. */
-		if (cb >= cb_sb_end || dp_addr > dp_sb_end)
-			break;
-
-		/* Determine token type and parse appropriately.*/
-		if ((tag & NTFS_TOKEN_MASK) == NTFS_SYMBOL_TOKEN) {
-			/*
-			 * We have a symbol token, copy the symbol across, and
-			 * advance the source and destination positions.
-			 */
-			*dp_addr++ = *cb++;
-			++*dest_ofs;
-
-			/* Continue with the next token. */
-			continue;
-		}
-
-		/*
-		 * We have a phrase token. Make sure it is not the first tag in
-		 * the sb as this is illegal and would confuse the code below.
-		 */
-		if (dp_addr == dp_sb_start)
-			goto return_overflow;
-
-		/*
-		 * Determine the number of bytes to go back (p) and the number
-		 * of bytes to copy (l). We use an optimized algorithm in which
-		 * we first calculate log2(current destination position in sb),
-		 * which allows determination of l and p in O(1) rather than
-		 * O(n). We just need an arch-optimized log2() function now.
-		 */
-		lg = 0;
-		for (i = *dest_ofs - do_sb_start - 1; i >= 0x10; i >>= 1)
-			lg++;
-
-		/* Get the phrase token into i. */
-		pt = le16_to_cpup((le16*)cb);
-
-		/*
-		 * Calculate starting position of the byte sequence in
-		 * the destination using the fact that p = (pt >> (12 - lg)) + 1
-		 * and make sure we don't go too far back.
-		 */
-		dp_back_addr = dp_addr - (pt >> (12 - lg)) - 1;
-		if (dp_back_addr < dp_sb_start)
-			goto return_overflow;
-
-		/* Now calculate the length of the byte sequence. */
-		length = (pt & (0xfff >> lg)) + 3;
-
-		/* Advance destination position and verify it is in range. */
-		*dest_ofs += length;
-		if (*dest_ofs > do_sb_end)
-			goto return_overflow;
-
-		/* The number of non-overlapping bytes. */
-		max_non_overlap = dp_addr - dp_back_addr;
-
-		if (length <= max_non_overlap) {
-			/* The byte sequence doesn't overlap, just copy it. */
-			memcpy(dp_addr, dp_back_addr, length);
-
-			/* Advance destination pointer. */
-			dp_addr += length;
-		} else {
-			/*
-			 * The byte sequence does overlap, copy non-overlapping
-			 * part and then do a slow byte by byte copy for the
-			 * overlapping part. Also, advance the destination
-			 * pointer.
-			 */
-			memcpy(dp_addr, dp_back_addr, max_non_overlap);
-			dp_addr += max_non_overlap;
-			dp_back_addr += max_non_overlap;
-			length -= max_non_overlap;
-			while (length--)
-				*dp_addr++ = *dp_back_addr++;
-		}
-
-		/* Advance source position and continue with the next token. */
-		cb += 2;
-	}
-
-	/* No tokens left in the current tag. Continue with the next tag. */
-	goto do_next_tag;
-
-return_overflow:
-	ntfs_error(NULL, "Failed. Returning -EOVERFLOW.");
-	goto return_error;
-}
-
-/**
- * ntfs_read_compressed_block - read a compressed block into the page cache
- * @page:	locked page in the compression block(s) we need to read
- *
- * When we are called the page has already been verified to be locked and the
- * attribute is known to be non-resident, not encrypted, but compressed.
- *
- * 1. Determine which compression block(s) @page is in.
- * 2. Get hold of all pages corresponding to this/these compression block(s).
- * 3. Read the (first) compression block.
- * 4. Decompress it into the corresponding pages.
- * 5. Throw the compressed data away and proceed to 3. for the next compression
- *    block or return success if no more compression blocks left.
- *
- * Warning: We have to be careful what we do about existing pages. They might
- * have been written to so that we would lose data if we were to just overwrite
- * them with the out-of-date uncompressed data.
- *
- * FIXME: For PAGE_SIZE > cb_size we are not doing the Right Thing(TM) at
- * the end of the file I think. We need to detect this case and zero the out
- * of bounds remainder of the page in question and mark it as handled. At the
- * moment we would just return -EIO on such a page. This bug will only become
- * apparent if pages are above 8kiB and the NTFS volume only uses 512 byte
- * clusters so is probably not going to be seen by anyone. Still this should
- * be fixed. (AIA)
- *
- * FIXME: Again for PAGE_SIZE > cb_size we are screwing up both in
- * handling sparse and compressed cbs. (AIA)
- *
- * FIXME: At the moment we don't do any zeroing out in the case that
- * initialized_size is less than data_size. This should be safe because of the
- * nature of the compression algorithm used. Just in case we check and output
- * an error message in read inode if the two sizes are not equal for a
- * compressed file. (AIA)
- */
-int ntfs_read_compressed_block(struct page *page)
-{
-	loff_t i_size;
-	s64 initialized_size;
-	struct address_space *mapping = page->mapping;
-	ntfs_inode *ni = NTFS_I(mapping->host);
-	ntfs_volume *vol = ni->vol;
-	struct super_block *sb = vol->sb;
-	runlist_element *rl;
-	unsigned long flags, block_size = sb->s_blocksize;
-	unsigned char block_size_bits = sb->s_blocksize_bits;
-	u8 *cb, *cb_pos, *cb_end;
-	struct buffer_head **bhs;
-	unsigned long offset, index = page->index;
-	u32 cb_size = ni->itype.compressed.block_size;
-	u64 cb_size_mask = cb_size - 1UL;
-	VCN vcn;
-	LCN lcn;
-	/* The first wanted vcn (minimum alignment is PAGE_SIZE). */
-	VCN start_vcn = (((s64)index << PAGE_SHIFT) & ~cb_size_mask) >>
-			vol->cluster_size_bits;
-	/*
-	 * The first vcn after the last wanted vcn (minimum alignment is again
-	 * PAGE_SIZE.
-	 */
-	VCN end_vcn = ((((s64)(index + 1UL) << PAGE_SHIFT) + cb_size - 1)
-			& ~cb_size_mask) >> vol->cluster_size_bits;
-	/* Number of compression blocks (cbs) in the wanted vcn range. */
-	unsigned int nr_cbs = (end_vcn - start_vcn) << vol->cluster_size_bits
-			>> ni->itype.compressed.block_size_bits;
-	/*
-	 * Number of pages required to store the uncompressed data from all
-	 * compression blocks (cbs) overlapping @page. Due to alignment
-	 * guarantees of start_vcn and end_vcn, no need to round up here.
-	 */
-	unsigned int nr_pages = (end_vcn - start_vcn) <<
-			vol->cluster_size_bits >> PAGE_SHIFT;
-	unsigned int xpage, max_page, cur_page, cur_ofs, i;
-	unsigned int cb_clusters, cb_max_ofs;
-	int block, max_block, cb_max_page, bhs_size, nr_bhs, err = 0;
-	struct page **pages;
-	int *completed_pages;
-	unsigned char xpage_done = 0;
-
-	ntfs_debug("Entering, page->index = 0x%lx, cb_size = 0x%x, nr_pages = "
-			"%i.", index, cb_size, nr_pages);
-	/*
-	 * Bad things happen if we get here for anything that is not an
-	 * unnamed $DATA attribute.
-	 */
-	BUG_ON(ni->type != AT_DATA);
-	BUG_ON(ni->name_len);
-
-	pages = kmalloc_array(nr_pages, sizeof(struct page *), GFP_NOFS);
-	completed_pages = kmalloc_array(nr_pages + 1, sizeof(int), GFP_NOFS);
-
-	/* Allocate memory to store the buffer heads we need. */
-	bhs_size = cb_size / block_size * sizeof(struct buffer_head *);
-	bhs = kmalloc(bhs_size, GFP_NOFS);
-
-	if (unlikely(!pages || !bhs || !completed_pages)) {
-		kfree(bhs);
-		kfree(pages);
-		kfree(completed_pages);
-		unlock_page(page);
-		ntfs_error(vol->sb, "Failed to allocate internal buffers.");
-		return -ENOMEM;
-	}
-
-	/*
-	 * We have already been given one page, this is the one we must do.
-	 * Once again, the alignment guarantees keep it simple.
-	 */
-	offset = start_vcn << vol->cluster_size_bits >> PAGE_SHIFT;
-	xpage = index - offset;
-	pages[xpage] = page;
-	/*
-	 * The remaining pages need to be allocated and inserted into the page
-	 * cache, alignment guarantees keep all the below much simpler. (-8
-	 */
-	read_lock_irqsave(&ni->size_lock, flags);
-	i_size = i_size_read(VFS_I(ni));
-	initialized_size = ni->initialized_size;
-	read_unlock_irqrestore(&ni->size_lock, flags);
-	max_page = ((i_size + PAGE_SIZE - 1) >> PAGE_SHIFT) -
-			offset;
-	/* Is the page fully outside i_size? (truncate in progress) */
-	if (xpage >= max_page) {
-		kfree(bhs);
-		kfree(pages);
-		kfree(completed_pages);
-		zero_user(page, 0, PAGE_SIZE);
-		ntfs_debug("Compressed read outside i_size - truncated?");
-		SetPageUptodate(page);
-		unlock_page(page);
-		return 0;
-	}
-	if (nr_pages < max_page)
-		max_page = nr_pages;
-	for (i = 0; i < max_page; i++, offset++) {
-		if (i != xpage)
-			pages[i] = grab_cache_page_nowait(mapping, offset);
-		page = pages[i];
-		if (page) {
-			/*
-			 * We only (re)read the page if it isn't already read
-			 * in and/or dirty or we would be losing data or at
-			 * least wasting our time.
-			 */
-			if (!PageDirty(page) && (!PageUptodate(page) ||
-					PageError(page))) {
-				ClearPageError(page);
-				kmap(page);
-				continue;
-			}
-			unlock_page(page);
-			put_page(page);
-			pages[i] = NULL;
-		}
-	}
-
-	/*
-	 * We have the runlist, and all the destination pages we need to fill.
-	 * Now read the first compression block.
-	 */
-	cur_page = 0;
-	cur_ofs = 0;
-	cb_clusters = ni->itype.compressed.block_clusters;
-do_next_cb:
-	nr_cbs--;
-	nr_bhs = 0;
-
-	/* Read all cb buffer heads one cluster at a time. */
-	rl = NULL;
-	for (vcn = start_vcn, start_vcn += cb_clusters; vcn < start_vcn;
-			vcn++) {
-		bool is_retry = false;
-
-		if (!rl) {
-lock_retry_remap:
-			down_read(&ni->runlist.lock);
-			rl = ni->runlist.rl;
-		}
-		if (likely(rl != NULL)) {
-			/* Seek to element containing target vcn. */
-			while (rl->length && rl[1].vcn <= vcn)
-				rl++;
-			lcn = ntfs_rl_vcn_to_lcn(rl, vcn);
-		} else
-			lcn = LCN_RL_NOT_MAPPED;
-		ntfs_debug("Reading vcn = 0x%llx, lcn = 0x%llx.",
-				(unsigned long long)vcn,
-				(unsigned long long)lcn);
-		if (lcn < 0) {
-			/*
-			 * When we reach the first sparse cluster we have
-			 * finished with the cb.
-			 */
-			if (lcn == LCN_HOLE)
-				break;
-			if (is_retry || lcn != LCN_RL_NOT_MAPPED)
-				goto rl_err;
-			is_retry = true;
-			/*
-			 * Attempt to map runlist, dropping lock for the
-			 * duration.
-			 */
-			up_read(&ni->runlist.lock);
-			if (!ntfs_map_runlist(ni, vcn))
-				goto lock_retry_remap;
-			goto map_rl_err;
-		}
-		block = lcn << vol->cluster_size_bits >> block_size_bits;
-		/* Read the lcn from device in chunks of block_size bytes. */
-		max_block = block + (vol->cluster_size >> block_size_bits);
-		do {
-			ntfs_debug("block = 0x%x.", block);
-			if (unlikely(!(bhs[nr_bhs] = sb_getblk(sb, block))))
-				goto getblk_err;
-			nr_bhs++;
-		} while (++block < max_block);
-	}
-
-	/* Release the lock if we took it. */
-	if (rl)
-		up_read(&ni->runlist.lock);
-
-	/* Setup and initiate io on all buffer heads. */
-	for (i = 0; i < nr_bhs; i++) {
-		struct buffer_head *tbh = bhs[i];
-
-		if (!trylock_buffer(tbh))
-			continue;
-		if (unlikely(buffer_uptodate(tbh))) {
-			unlock_buffer(tbh);
-			continue;
-		}
-		get_bh(tbh);
-		tbh->b_end_io = end_buffer_read_sync;
-		submit_bh(REQ_OP_READ, tbh);
-	}
-
-	/* Wait for io completion on all buffer heads. */
-	for (i = 0; i < nr_bhs; i++) {
-		struct buffer_head *tbh = bhs[i];
-
-		if (buffer_uptodate(tbh))
-			continue;
-		wait_on_buffer(tbh);
-		/*
-		 * We need an optimization barrier here, otherwise we start
-		 * hitting the below fixup code when accessing a loopback
-		 * mounted ntfs partition. This indicates either there is a
-		 * race condition in the loop driver or, more likely, gcc
-		 * overoptimises the code without the barrier and it doesn't
-		 * do the Right Thing(TM).
-		 */
-		barrier();
-		if (unlikely(!buffer_uptodate(tbh))) {
-			ntfs_warning(vol->sb, "Buffer is unlocked but not "
-					"uptodate! Unplugging the disk queue "
-					"and rescheduling.");
-			get_bh(tbh);
-			io_schedule();
-			put_bh(tbh);
-			if (unlikely(!buffer_uptodate(tbh)))
-				goto read_err;
-			ntfs_warning(vol->sb, "Buffer is now uptodate. Good.");
-		}
-	}
-
-	/*
-	 * Get the compression buffer. We must not sleep any more
-	 * until we are finished with it.
-	 */
-	spin_lock(&ntfs_cb_lock);
-	cb = ntfs_compression_buffer;
-
-	BUG_ON(!cb);
-
-	cb_pos = cb;
-	cb_end = cb + cb_size;
-
-	/* Copy the buffer heads into the contiguous buffer. */
-	for (i = 0; i < nr_bhs; i++) {
-		memcpy(cb_pos, bhs[i]->b_data, block_size);
-		cb_pos += block_size;
-	}
-
-	/* Just a precaution. */
-	if (cb_pos + 2 <= cb + cb_size)
-		*(u16*)cb_pos = 0;
-
-	/* Reset cb_pos back to the beginning. */
-	cb_pos = cb;
-
-	/* We now have both source (if present) and destination. */
-	ntfs_debug("Successfully read the compression block.");
-
-	/* The last page and maximum offset within it for the current cb. */
-	cb_max_page = (cur_page << PAGE_SHIFT) + cur_ofs + cb_size;
-	cb_max_ofs = cb_max_page & ~PAGE_MASK;
-	cb_max_page >>= PAGE_SHIFT;
-
-	/* Catch end of file inside a compression block. */
-	if (cb_max_page > max_page)
-		cb_max_page = max_page;
-
-	if (vcn == start_vcn - cb_clusters) {
-		/* Sparse cb, zero out page range overlapping the cb. */
-		ntfs_debug("Found sparse compression block.");
-		/* We can sleep from now on, so we drop lock. */
-		spin_unlock(&ntfs_cb_lock);
-		if (cb_max_ofs)
-			cb_max_page--;
-		for (; cur_page < cb_max_page; cur_page++) {
-			page = pages[cur_page];
-			if (page) {
-				if (likely(!cur_ofs))
-					clear_page(page_address(page));
-				else
-					memset(page_address(page) + cur_ofs, 0,
-							PAGE_SIZE -
-							cur_ofs);
-				flush_dcache_page(page);
-				kunmap(page);
-				SetPageUptodate(page);
-				unlock_page(page);
-				if (cur_page == xpage)
-					xpage_done = 1;
-				else
-					put_page(page);
-				pages[cur_page] = NULL;
-			}
-			cb_pos += PAGE_SIZE - cur_ofs;
-			cur_ofs = 0;
-			if (cb_pos >= cb_end)
-				break;
-		}
-		/* If we have a partial final page, deal with it now. */
-		if (cb_max_ofs && cb_pos < cb_end) {
-			page = pages[cur_page];
-			if (page)
-				memset(page_address(page) + cur_ofs, 0,
-						cb_max_ofs - cur_ofs);
-			/*
-			 * No need to update cb_pos at this stage:
-			 *	cb_pos += cb_max_ofs - cur_ofs;
-			 */
-			cur_ofs = cb_max_ofs;
-		}
-	} else if (vcn == start_vcn) {
-		/* We can't sleep so we need two stages. */
-		unsigned int cur2_page = cur_page;
-		unsigned int cur_ofs2 = cur_ofs;
-		u8 *cb_pos2 = cb_pos;
-
-		ntfs_debug("Found uncompressed compression block.");
-		/* Uncompressed cb, copy it to the destination pages. */
-		/*
-		 * TODO: As a big optimization, we could detect this case
-		 * before we read all the pages and use block_read_full_folio()
-		 * on all full pages instead (we still have to treat partial
-		 * pages especially but at least we are getting rid of the
-		 * synchronous io for the majority of pages.
-		 * Or if we choose not to do the read-ahead/-behind stuff, we
-		 * could just return block_read_full_folio(pages[xpage]) as long
-		 * as PAGE_SIZE <= cb_size.
-		 */
-		if (cb_max_ofs)
-			cb_max_page--;
-		/* First stage: copy data into destination pages. */
-		for (; cur_page < cb_max_page; cur_page++) {
-			page = pages[cur_page];
-			if (page)
-				memcpy(page_address(page) + cur_ofs, cb_pos,
-						PAGE_SIZE - cur_ofs);
-			cb_pos += PAGE_SIZE - cur_ofs;
-			cur_ofs = 0;
-			if (cb_pos >= cb_end)
-				break;
-		}
-		/* If we have a partial final page, deal with it now. */
-		if (cb_max_ofs && cb_pos < cb_end) {
-			page = pages[cur_page];
-			if (page)
-				memcpy(page_address(page) + cur_ofs, cb_pos,
-						cb_max_ofs - cur_ofs);
-			cb_pos += cb_max_ofs - cur_ofs;
-			cur_ofs = cb_max_ofs;
-		}
-		/* We can sleep from now on, so drop lock. */
-		spin_unlock(&ntfs_cb_lock);
-		/* Second stage: finalize pages. */
-		for (; cur2_page < cb_max_page; cur2_page++) {
-			page = pages[cur2_page];
-			if (page) {
-				/*
-				 * If we are outside the initialized size, zero
-				 * the out of bounds page range.
-				 */
-				handle_bounds_compressed_page(page, i_size,
-						initialized_size);
-				flush_dcache_page(page);
-				kunmap(page);
-				SetPageUptodate(page);
-				unlock_page(page);
-				if (cur2_page == xpage)
-					xpage_done = 1;
-				else
-					put_page(page);
-				pages[cur2_page] = NULL;
-			}
-			cb_pos2 += PAGE_SIZE - cur_ofs2;
-			cur_ofs2 = 0;
-			if (cb_pos2 >= cb_end)
-				break;
-		}
-	} else {
-		/* Compressed cb, decompress it into the destination page(s). */
-		unsigned int prev_cur_page = cur_page;
-
-		ntfs_debug("Found compressed compression block.");
-		err = ntfs_decompress(pages, completed_pages, &cur_page,
-				&cur_ofs, cb_max_page, cb_max_ofs, xpage,
-				&xpage_done, cb_pos, cb_size - (cb_pos - cb),
-				i_size, initialized_size);
-		/*
-		 * We can sleep from now on, lock already dropped by
-		 * ntfs_decompress().
-		 */
-		if (err) {
-			ntfs_error(vol->sb, "ntfs_decompress() failed in inode "
-					"0x%lx with error code %i. Skipping "
-					"this compression block.",
-					ni->mft_no, -err);
-			/* Release the unfinished pages. */
-			for (; prev_cur_page < cur_page; prev_cur_page++) {
-				page = pages[prev_cur_page];
-				if (page) {
-					flush_dcache_page(page);
-					kunmap(page);
-					unlock_page(page);
-					if (prev_cur_page != xpage)
-						put_page(page);
-					pages[prev_cur_page] = NULL;
-				}
-			}
-		}
-	}
-
-	/* Release the buffer heads. */
-	for (i = 0; i < nr_bhs; i++)
-		brelse(bhs[i]);
-
-	/* Do we have more work to do? */
-	if (nr_cbs)
-		goto do_next_cb;
-
-	/* We no longer need the list of buffer heads. */
-	kfree(bhs);
-
-	/* Clean up if we have any pages left. Should never happen. */
-	for (cur_page = 0; cur_page < max_page; cur_page++) {
-		page = pages[cur_page];
-		if (page) {
-			ntfs_error(vol->sb, "Still have pages left! "
-					"Terminating them with extreme "
-					"prejudice.  Inode 0x%lx, page index "
-					"0x%lx.", ni->mft_no, page->index);
-			flush_dcache_page(page);
-			kunmap(page);
-			unlock_page(page);
-			if (cur_page != xpage)
-				put_page(page);
-			pages[cur_page] = NULL;
-		}
-	}
-
-	/* We no longer need the list of pages. */
-	kfree(pages);
-	kfree(completed_pages);
-
-	/* If we have completed the requested page, we return success. */
-	if (likely(xpage_done))
-		return 0;
-
-	ntfs_debug("Failed. Returning error code %s.", err == -EOVERFLOW ?
-			"EOVERFLOW" : (!err ? "EIO" : "unknown error"));
-	return err < 0 ? err : -EIO;
-
-read_err:
-	ntfs_error(vol->sb, "IO error while reading compressed data.");
-	/* Release the buffer heads. */
-	for (i = 0; i < nr_bhs; i++)
-		brelse(bhs[i]);
-	goto err_out;
-
-map_rl_err:
-	ntfs_error(vol->sb, "ntfs_map_runlist() failed. Cannot read "
-			"compression block.");
-	goto err_out;
-
-rl_err:
-	up_read(&ni->runlist.lock);
-	ntfs_error(vol->sb, "ntfs_rl_vcn_to_lcn() failed. Cannot read "
-			"compression block.");
-	goto err_out;
-
-getblk_err:
-	up_read(&ni->runlist.lock);
-	ntfs_error(vol->sb, "getblk() failed. Cannot read compression block.");
-
-err_out:
-	kfree(bhs);
-	for (i = cur_page; i < max_page; i++) {
-		page = pages[i];
-		if (page) {
-			flush_dcache_page(page);
-			kunmap(page);
-			unlock_page(page);
-			if (i != xpage)
-				put_page(page);
-		}
-	}
-	kfree(pages);
-	kfree(completed_pages);
-	return -EIO;
-}
diff --git a/fs/ntfs/debug.c b/fs/ntfs/debug.c
deleted file mode 100644
index a3c1c5656f..0000000000
--- a/fs/ntfs/debug.c
+++ /dev/null
@@ -1,159 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * debug.c - NTFS kernel debug support. Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2004 Anton Altaparmakov
- */
-#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
-#include "debug.h"
-
-/**
- * __ntfs_warning - output a warning to the syslog
- * @function:	name of function outputting the warning
- * @sb:		super block of mounted ntfs filesystem
- * @fmt:	warning string containing format specifications
- * @...:	a variable number of arguments specified in @fmt
- *
- * Outputs a warning to the syslog for the mounted ntfs filesystem described
- * by @sb.
- *
- * @fmt and the corresponding @... is printf style format string containing
- * the warning string and the corresponding format arguments, respectively.
- *
- * @function is the name of the function from which __ntfs_warning is being
- * called.
- *
- * Note, you should be using debug.h::ntfs_warning(@sb, @fmt, @...) instead
- * as this provides the @function parameter automatically.
- */
-void __ntfs_warning(const char *function, const struct super_block *sb,
-		const char *fmt, ...)
-{
-	struct va_format vaf;
-	va_list args;
-	int flen = 0;
-
-#ifndef DEBUG
-	if (!printk_ratelimit())
-		return;
-#endif
-	if (function)
-		flen = strlen(function);
-	va_start(args, fmt);
-	vaf.fmt = fmt;
-	vaf.va = &args;
-	if (sb)
-		pr_warn("(device %s): %s(): %pV\n",
-			sb->s_id, flen ? function : "", &vaf);
-	else
-		pr_warn("%s(): %pV\n", flen ? function : "", &vaf);
-	va_end(args);
-}
-
-/**
- * __ntfs_error - output an error to the syslog
- * @function:	name of function outputting the error
- * @sb:		super block of mounted ntfs filesystem
- * @fmt:	error string containing format specifications
- * @...:	a variable number of arguments specified in @fmt
- *
- * Outputs an error to the syslog for the mounted ntfs filesystem described
- * by @sb.
- *
- * @fmt and the corresponding @... is printf style format string containing
- * the error string and the corresponding format arguments, respectively.
- *
- * @function is the name of the function from which __ntfs_error is being
- * called.
- *
- * Note, you should be using debug.h::ntfs_error(@sb, @fmt, @...) instead
- * as this provides the @function parameter automatically.
- */
-void __ntfs_error(const char *function, const struct super_block *sb,
-		const char *fmt, ...)
-{
-	struct va_format vaf;
-	va_list args;
-	int flen = 0;
-
-#ifndef DEBUG
-	if (!printk_ratelimit())
-		return;
-#endif
-	if (function)
-		flen = strlen(function);
-	va_start(args, fmt);
-	vaf.fmt = fmt;
-	vaf.va = &args;
-	if (sb)
-		pr_err("(device %s): %s(): %pV\n",
-		       sb->s_id, flen ? function : "", &vaf);
-	else
-		pr_err("%s(): %pV\n", flen ? function : "", &vaf);
-	va_end(args);
-}
-
-#ifdef DEBUG
-
-/* If 1, output debug messages, and if 0, don't. */
-int debug_msgs = 0;
-
-void __ntfs_debug(const char *file, int line, const char *function,
-		const char *fmt, ...)
-{
-	struct va_format vaf;
-	va_list args;
-	int flen = 0;
-
-	if (!debug_msgs)
-		return;
-	if (function)
-		flen = strlen(function);
-	va_start(args, fmt);
-	vaf.fmt = fmt;
-	vaf.va = &args;
-	pr_debug("(%s, %d): %s(): %pV", file, line, flen ? function : "", &vaf);
-	va_end(args);
-}
-
-/* Dump a runlist. Caller has to provide synchronisation for @rl. */
-void ntfs_debug_dump_runlist(const runlist_element *rl)
-{
-	int i;
-	const char *lcn_str[5] = { "LCN_HOLE         ", "LCN_RL_NOT_MAPPED",
-				   "LCN_ENOENT       ", "LCN_unknown      " };
-
-	if (!debug_msgs)
-		return;
-	pr_debug("Dumping runlist (values in hex):\n");
-	if (!rl) {
-		pr_debug("Run list not present.\n");
-		return;
-	}
-	pr_debug("VCN              LCN               Run length\n");
-	for (i = 0; ; i++) {
-		LCN lcn = (rl + i)->lcn;
-
-		if (lcn < (LCN)0) {
-			int index = -lcn - 1;
-
-			if (index > -LCN_ENOENT - 1)
-				index = 3;
-			pr_debug("%-16Lx %s %-16Lx%s\n",
-					(long long)(rl + i)->vcn, lcn_str[index],
-					(long long)(rl + i)->length,
-					(rl + i)->length ? "" :
-						" (runlist end)");
-		} else
-			pr_debug("%-16Lx %-16Lx  %-16Lx%s\n",
-					(long long)(rl + i)->vcn,
-					(long long)(rl + i)->lcn,
-					(long long)(rl + i)->length,
-					(rl + i)->length ? "" :
-						" (runlist end)");
-		if (!(rl + i)->length)
-			break;
-	}
-}
-
-#endif
diff --git a/fs/ntfs/debug.h b/fs/ntfs/debug.h
deleted file mode 100644
index 6fdef388f1..0000000000
--- a/fs/ntfs/debug.h
+++ /dev/null
@@ -1,57 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * debug.h - NTFS kernel debug support. Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2004 Anton Altaparmakov
- */
-
-#ifndef _LINUX_NTFS_DEBUG_H
-#define _LINUX_NTFS_DEBUG_H
-
-#include <linux/fs.h>
-
-#include "runlist.h"
-
-#ifdef DEBUG
-
-extern int debug_msgs;
-
-extern __printf(4, 5)
-void __ntfs_debug(const char *file, int line, const char *function,
-		  const char *format, ...);
-/**
- * ntfs_debug - write a debug level message to syslog
- * @f:		a printf format string containing the message
- * @...:	the variables to substitute into @f
- *
- * ntfs_debug() writes a DEBUG level message to the syslog but only if the
- * driver was compiled with -DDEBUG. Otherwise, the call turns into a NOP.
- */
-#define ntfs_debug(f, a...)						\
-	__ntfs_debug(__FILE__, __LINE__, __func__, f, ##a)
-
-extern void ntfs_debug_dump_runlist(const runlist_element *rl);
-
-#else	/* !DEBUG */
-
-#define ntfs_debug(fmt, ...)						\
-do {									\
-	if (0)								\
-		no_printk(fmt, ##__VA_ARGS__);				\
-} while (0)
-
-#define ntfs_debug_dump_runlist(rl)	do {} while (0)
-
-#endif	/* !DEBUG */
-
-extern  __printf(3, 4)
-void __ntfs_warning(const char *function, const struct super_block *sb,
-		    const char *fmt, ...);
-#define ntfs_warning(sb, f, a...)	__ntfs_warning(__func__, sb, f, ##a)
-
-extern  __printf(3, 4)
-void __ntfs_error(const char *function, const struct super_block *sb,
-		  const char *fmt, ...);
-#define ntfs_error(sb, f, a...)		__ntfs_error(__func__, sb, f, ##a)
-
-#endif /* _LINUX_NTFS_DEBUG_H */
diff --git a/fs/ntfs/dir.c b/fs/ntfs/dir.c
deleted file mode 100644
index 629723a8d7..0000000000
--- a/fs/ntfs/dir.c
+++ /dev/null
@@ -1,1540 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * dir.c - NTFS kernel directory operations. Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2007 Anton Altaparmakov
- * Copyright (c) 2002 Richard Russon
- */
-
-#include <linux/buffer_head.h>
-#include <linux/slab.h>
-#include <linux/blkdev.h>
-
-#include "dir.h"
-#include "aops.h"
-#include "attrib.h"
-#include "mft.h"
-#include "debug.h"
-#include "ntfs.h"
-
-/*
- * The little endian Unicode string $I30 as a global constant.
- */
-ntfschar I30[5] = { cpu_to_le16('$'), cpu_to_le16('I'),
-		cpu_to_le16('3'),	cpu_to_le16('0'), 0 };
-
-/**
- * ntfs_lookup_inode_by_name - find an inode in a directory given its name
- * @dir_ni:	ntfs inode of the directory in which to search for the name
- * @uname:	Unicode name for which to search in the directory
- * @uname_len:	length of the name @uname in Unicode characters
- * @res:	return the found file name if necessary (see below)
- *
- * Look for an inode with name @uname in the directory with inode @dir_ni.
- * ntfs_lookup_inode_by_name() walks the contents of the directory looking for
- * the Unicode name. If the name is found in the directory, the corresponding
- * inode number (>= 0) is returned as a mft reference in cpu format, i.e. it
- * is a 64-bit number containing the sequence number.
- *
- * On error, a negative value is returned corresponding to the error code. In
- * particular if the inode is not found -ENOENT is returned. Note that you
- * can't just check the return value for being negative, you have to check the
- * inode number for being negative which you can extract using MREC(return
- * value).
- *
- * Note, @uname_len does not include the (optional) terminating NULL character.
- *
- * Note, we look for a case sensitive match first but we also look for a case
- * insensitive match at the same time. If we find a case insensitive match, we
- * save that for the case that we don't find an exact match, where we return
- * the case insensitive match and setup @res (which we allocate!) with the mft
- * reference, the file name type, length and with a copy of the little endian
- * Unicode file name itself. If we match a file name which is in the DOS name
- * space, we only return the mft reference and file name type in @res.
- * ntfs_lookup() then uses this to find the long file name in the inode itself.
- * This is to avoid polluting the dcache with short file names. We want them to
- * work but we don't care for how quickly one can access them. This also fixes
- * the dcache aliasing issues.
- *
- * Locking:  - Caller must hold i_mutex on the directory.
- *	     - Each page cache page in the index allocation mapping must be
- *	       locked whilst being accessed otherwise we may find a corrupt
- *	       page due to it being under ->writepage at the moment which
- *	       applies the mst protection fixups before writing out and then
- *	       removes them again after the write is complete after which it 
- *	       unlocks the page.
- */
-MFT_REF ntfs_lookup_inode_by_name(ntfs_inode *dir_ni, const ntfschar *uname,
-		const int uname_len, ntfs_name **res)
-{
-	ntfs_volume *vol = dir_ni->vol;
-	struct super_block *sb = vol->sb;
-	MFT_RECORD *m;
-	INDEX_ROOT *ir;
-	INDEX_ENTRY *ie;
-	INDEX_ALLOCATION *ia;
-	u8 *index_end;
-	u64 mref;
-	ntfs_attr_search_ctx *ctx;
-	int err, rc;
-	VCN vcn, old_vcn;
-	struct address_space *ia_mapping;
-	struct page *page;
-	u8 *kaddr;
-	ntfs_name *name = NULL;
-
-	BUG_ON(!S_ISDIR(VFS_I(dir_ni)->i_mode));
-	BUG_ON(NInoAttr(dir_ni));
-	/* Get hold of the mft record for the directory. */
-	m = map_mft_record(dir_ni);
-	if (IS_ERR(m)) {
-		ntfs_error(sb, "map_mft_record() failed with error code %ld.",
-				-PTR_ERR(m));
-		return ERR_MREF(PTR_ERR(m));
-	}
-	ctx = ntfs_attr_get_search_ctx(dir_ni, m);
-	if (unlikely(!ctx)) {
-		err = -ENOMEM;
-		goto err_out;
-	}
-	/* Find the index root attribute in the mft record. */
-	err = ntfs_attr_lookup(AT_INDEX_ROOT, I30, 4, CASE_SENSITIVE, 0, NULL,
-			0, ctx);
-	if (unlikely(err)) {
-		if (err == -ENOENT) {
-			ntfs_error(sb, "Index root attribute missing in "
-					"directory inode 0x%lx.",
-					dir_ni->mft_no);
-			err = -EIO;
-		}
-		goto err_out;
-	}
-	/* Get to the index root value (it's been verified in read_inode). */
-	ir = (INDEX_ROOT*)((u8*)ctx->attr +
-			le16_to_cpu(ctx->attr->data.resident.value_offset));
-	index_end = (u8*)&ir->index + le32_to_cpu(ir->index.index_length);
-	/* The first index entry. */
-	ie = (INDEX_ENTRY*)((u8*)&ir->index +
-			le32_to_cpu(ir->index.entries_offset));
-	/*
-	 * Loop until we exceed valid memory (corruption case) or until we
-	 * reach the last entry.
-	 */
-	for (;; ie = (INDEX_ENTRY*)((u8*)ie + le16_to_cpu(ie->length))) {
-		/* Bounds checks. */
-		if ((u8*)ie < (u8*)ctx->mrec || (u8*)ie +
-				sizeof(INDEX_ENTRY_HEADER) > index_end ||
-				(u8*)ie + le16_to_cpu(ie->key_length) >
-				index_end)
-			goto dir_err_out;
-		/*
-		 * The last entry cannot contain a name. It can however contain
-		 * a pointer to a child node in the B+tree so we just break out.
-		 */
-		if (ie->flags & INDEX_ENTRY_END)
-			break;
-		/*
-		 * We perform a case sensitive comparison and if that matches
-		 * we are done and return the mft reference of the inode (i.e.
-		 * the inode number together with the sequence number for
-		 * consistency checking). We convert it to cpu format before
-		 * returning.
-		 */
-		if (ntfs_are_names_equal(uname, uname_len,
-				(ntfschar*)&ie->key.file_name.file_name,
-				ie->key.file_name.file_name_length,
-				CASE_SENSITIVE, vol->upcase, vol->upcase_len)) {
-found_it:
-			/*
-			 * We have a perfect match, so we don't need to care
-			 * about having matched imperfectly before, so we can
-			 * free name and set *res to NULL.
-			 * However, if the perfect match is a short file name,
-			 * we need to signal this through *res, so that
-			 * ntfs_lookup() can fix dcache aliasing issues.
-			 * As an optimization we just reuse an existing
-			 * allocation of *res.
-			 */
-			if (ie->key.file_name.file_name_type == FILE_NAME_DOS) {
-				if (!name) {
-					name = kmalloc(sizeof(ntfs_name),
-							GFP_NOFS);
-					if (!name) {
-						err = -ENOMEM;
-						goto err_out;
-					}
-				}
-				name->mref = le64_to_cpu(
-						ie->data.dir.indexed_file);
-				name->type = FILE_NAME_DOS;
-				name->len = 0;
-				*res = name;
-			} else {
-				kfree(name);
-				*res = NULL;
-			}
-			mref = le64_to_cpu(ie->data.dir.indexed_file);
-			ntfs_attr_put_search_ctx(ctx);
-			unmap_mft_record(dir_ni);
-			return mref;
-		}
-		/*
-		 * For a case insensitive mount, we also perform a case
-		 * insensitive comparison (provided the file name is not in the
-		 * POSIX namespace). If the comparison matches, and the name is
-		 * in the WIN32 namespace, we cache the filename in *res so
-		 * that the caller, ntfs_lookup(), can work on it. If the
-		 * comparison matches, and the name is in the DOS namespace, we
-		 * only cache the mft reference and the file name type (we set
-		 * the name length to zero for simplicity).
-		 */
-		if (!NVolCaseSensitive(vol) &&
-				ie->key.file_name.file_name_type &&
-				ntfs_are_names_equal(uname, uname_len,
-				(ntfschar*)&ie->key.file_name.file_name,
-				ie->key.file_name.file_name_length,
-				IGNORE_CASE, vol->upcase, vol->upcase_len)) {
-			int name_size = sizeof(ntfs_name);
-			u8 type = ie->key.file_name.file_name_type;
-			u8 len = ie->key.file_name.file_name_length;
-
-			/* Only one case insensitive matching name allowed. */
-			if (name) {
-				ntfs_error(sb, "Found already allocated name "
-						"in phase 1. Please run chkdsk "
-						"and if that doesn't find any "
-						"errors please report you saw "
-						"this message to "
-						"linux-ntfs-dev@lists."
-						"sourceforge.net.");
-				goto dir_err_out;
-			}
-
-			if (type != FILE_NAME_DOS)
-				name_size += len * sizeof(ntfschar);
-			name = kmalloc(name_size, GFP_NOFS);
-			if (!name) {
-				err = -ENOMEM;
-				goto err_out;
-			}
-			name->mref = le64_to_cpu(ie->data.dir.indexed_file);
-			name->type = type;
-			if (type != FILE_NAME_DOS) {
-				name->len = len;
-				memcpy(name->name, ie->key.file_name.file_name,
-						len * sizeof(ntfschar));
-			} else
-				name->len = 0;
-			*res = name;
-		}
-		/*
-		 * Not a perfect match, need to do full blown collation so we
-		 * know which way in the B+tree we have to go.
-		 */
-		rc = ntfs_collate_names(uname, uname_len,
-				(ntfschar*)&ie->key.file_name.file_name,
-				ie->key.file_name.file_name_length, 1,
-				IGNORE_CASE, vol->upcase, vol->upcase_len);
-		/*
-		 * If uname collates before the name of the current entry, there
-		 * is definitely no such name in this index but we might need to
-		 * descend into the B+tree so we just break out of the loop.
-		 */
-		if (rc == -1)
-			break;
-		/* The names are not equal, continue the search. */
-		if (rc)
-			continue;
-		/*
-		 * Names match with case insensitive comparison, now try the
-		 * case sensitive comparison, which is required for proper
-		 * collation.
-		 */
-		rc = ntfs_collate_names(uname, uname_len,
-				(ntfschar*)&ie->key.file_name.file_name,
-				ie->key.file_name.file_name_length, 1,
-				CASE_SENSITIVE, vol->upcase, vol->upcase_len);
-		if (rc == -1)
-			break;
-		if (rc)
-			continue;
-		/*
-		 * Perfect match, this will never happen as the
-		 * ntfs_are_names_equal() call will have gotten a match but we
-		 * still treat it correctly.
-		 */
-		goto found_it;
-	}
-	/*
-	 * We have finished with this index without success. Check for the
-	 * presence of a child node and if not present return -ENOENT, unless
-	 * we have got a matching name cached in name in which case return the
-	 * mft reference associated with it.
-	 */
-	if (!(ie->flags & INDEX_ENTRY_NODE)) {
-		if (name) {
-			ntfs_attr_put_search_ctx(ctx);
-			unmap_mft_record(dir_ni);
-			return name->mref;
-		}
-		ntfs_debug("Entry not found.");
-		err = -ENOENT;
-		goto err_out;
-	} /* Child node present, descend into it. */
-	/* Consistency check: Verify that an index allocation exists. */
-	if (!NInoIndexAllocPresent(dir_ni)) {
-		ntfs_error(sb, "No index allocation attribute but index entry "
-				"requires one. Directory inode 0x%lx is "
-				"corrupt or driver bug.", dir_ni->mft_no);
-		goto err_out;
-	}
-	/* Get the starting vcn of the index_block holding the child node. */
-	vcn = sle64_to_cpup((sle64*)((u8*)ie + le16_to_cpu(ie->length) - 8));
-	ia_mapping = VFS_I(dir_ni)->i_mapping;
-	/*
-	 * We are done with the index root and the mft record. Release them,
-	 * otherwise we deadlock with ntfs_map_page().
-	 */
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(dir_ni);
-	m = NULL;
-	ctx = NULL;
-descend_into_child_node:
-	/*
-	 * Convert vcn to index into the index allocation attribute in units
-	 * of PAGE_SIZE and map the page cache page, reading it from
-	 * disk if necessary.
-	 */
-	page = ntfs_map_page(ia_mapping, vcn <<
-			dir_ni->itype.index.vcn_size_bits >> PAGE_SHIFT);
-	if (IS_ERR(page)) {
-		ntfs_error(sb, "Failed to map directory index page, error %ld.",
-				-PTR_ERR(page));
-		err = PTR_ERR(page);
-		goto err_out;
-	}
-	lock_page(page);
-	kaddr = (u8*)page_address(page);
-fast_descend_into_child_node:
-	/* Get to the index allocation block. */
-	ia = (INDEX_ALLOCATION*)(kaddr + ((vcn <<
-			dir_ni->itype.index.vcn_size_bits) & ~PAGE_MASK));
-	/* Bounds checks. */
-	if ((u8*)ia < kaddr || (u8*)ia > kaddr + PAGE_SIZE) {
-		ntfs_error(sb, "Out of bounds check failed. Corrupt directory "
-				"inode 0x%lx or driver bug.", dir_ni->mft_no);
-		goto unm_err_out;
-	}
-	/* Catch multi sector transfer fixup errors. */
-	if (unlikely(!ntfs_is_indx_record(ia->magic))) {
-		ntfs_error(sb, "Directory index record with vcn 0x%llx is "
-				"corrupt.  Corrupt inode 0x%lx.  Run chkdsk.",
-				(unsigned long long)vcn, dir_ni->mft_no);
-		goto unm_err_out;
-	}
-	if (sle64_to_cpu(ia->index_block_vcn) != vcn) {
-		ntfs_error(sb, "Actual VCN (0x%llx) of index buffer is "
-				"different from expected VCN (0x%llx). "
-				"Directory inode 0x%lx is corrupt or driver "
-				"bug.", (unsigned long long)
-				sle64_to_cpu(ia->index_block_vcn),
-				(unsigned long long)vcn, dir_ni->mft_no);
-		goto unm_err_out;
-	}
-	if (le32_to_cpu(ia->index.allocated_size) + 0x18 !=
-			dir_ni->itype.index.block_size) {
-		ntfs_error(sb, "Index buffer (VCN 0x%llx) of directory inode "
-				"0x%lx has a size (%u) differing from the "
-				"directory specified size (%u). Directory "
-				"inode is corrupt or driver bug.",
-				(unsigned long long)vcn, dir_ni->mft_no,
-				le32_to_cpu(ia->index.allocated_size) + 0x18,
-				dir_ni->itype.index.block_size);
-		goto unm_err_out;
-	}
-	index_end = (u8*)ia + dir_ni->itype.index.block_size;
-	if (index_end > kaddr + PAGE_SIZE) {
-		ntfs_error(sb, "Index buffer (VCN 0x%llx) of directory inode "
-				"0x%lx crosses page boundary. Impossible! "
-				"Cannot access! This is probably a bug in the "
-				"driver.", (unsigned long long)vcn,
-				dir_ni->mft_no);
-		goto unm_err_out;
-	}
-	index_end = (u8*)&ia->index + le32_to_cpu(ia->index.index_length);
-	if (index_end > (u8*)ia + dir_ni->itype.index.block_size) {
-		ntfs_error(sb, "Size of index buffer (VCN 0x%llx) of directory "
-				"inode 0x%lx exceeds maximum size.",
-				(unsigned long long)vcn, dir_ni->mft_no);
-		goto unm_err_out;
-	}
-	/* The first index entry. */
-	ie = (INDEX_ENTRY*)((u8*)&ia->index +
-			le32_to_cpu(ia->index.entries_offset));
-	/*
-	 * Iterate similar to above big loop but applied to index buffer, thus
-	 * loop until we exceed valid memory (corruption case) or until we
-	 * reach the last entry.
-	 */
-	for (;; ie = (INDEX_ENTRY*)((u8*)ie + le16_to_cpu(ie->length))) {
-		/* Bounds check. */
-		if ((u8*)ie < (u8*)ia || (u8*)ie +
-				sizeof(INDEX_ENTRY_HEADER) > index_end ||
-				(u8*)ie + le16_to_cpu(ie->key_length) >
-				index_end) {
-			ntfs_error(sb, "Index entry out of bounds in "
-					"directory inode 0x%lx.",
-					dir_ni->mft_no);
-			goto unm_err_out;
-		}
-		/*
-		 * The last entry cannot contain a name. It can however contain
-		 * a pointer to a child node in the B+tree so we just break out.
-		 */
-		if (ie->flags & INDEX_ENTRY_END)
-			break;
-		/*
-		 * We perform a case sensitive comparison and if that matches
-		 * we are done and return the mft reference of the inode (i.e.
-		 * the inode number together with the sequence number for
-		 * consistency checking). We convert it to cpu format before
-		 * returning.
-		 */
-		if (ntfs_are_names_equal(uname, uname_len,
-				(ntfschar*)&ie->key.file_name.file_name,
-				ie->key.file_name.file_name_length,
-				CASE_SENSITIVE, vol->upcase, vol->upcase_len)) {
-found_it2:
-			/*
-			 * We have a perfect match, so we don't need to care
-			 * about having matched imperfectly before, so we can
-			 * free name and set *res to NULL.
-			 * However, if the perfect match is a short file name,
-			 * we need to signal this through *res, so that
-			 * ntfs_lookup() can fix dcache aliasing issues.
-			 * As an optimization we just reuse an existing
-			 * allocation of *res.
-			 */
-			if (ie->key.file_name.file_name_type == FILE_NAME_DOS) {
-				if (!name) {
-					name = kmalloc(sizeof(ntfs_name),
-							GFP_NOFS);
-					if (!name) {
-						err = -ENOMEM;
-						goto unm_err_out;
-					}
-				}
-				name->mref = le64_to_cpu(
-						ie->data.dir.indexed_file);
-				name->type = FILE_NAME_DOS;
-				name->len = 0;
-				*res = name;
-			} else {
-				kfree(name);
-				*res = NULL;
-			}
-			mref = le64_to_cpu(ie->data.dir.indexed_file);
-			unlock_page(page);
-			ntfs_unmap_page(page);
-			return mref;
-		}
-		/*
-		 * For a case insensitive mount, we also perform a case
-		 * insensitive comparison (provided the file name is not in the
-		 * POSIX namespace). If the comparison matches, and the name is
-		 * in the WIN32 namespace, we cache the filename in *res so
-		 * that the caller, ntfs_lookup(), can work on it. If the
-		 * comparison matches, and the name is in the DOS namespace, we
-		 * only cache the mft reference and the file name type (we set
-		 * the name length to zero for simplicity).
-		 */
-		if (!NVolCaseSensitive(vol) &&
-				ie->key.file_name.file_name_type &&
-				ntfs_are_names_equal(uname, uname_len,
-				(ntfschar*)&ie->key.file_name.file_name,
-				ie->key.file_name.file_name_length,
-				IGNORE_CASE, vol->upcase, vol->upcase_len)) {
-			int name_size = sizeof(ntfs_name);
-			u8 type = ie->key.file_name.file_name_type;
-			u8 len = ie->key.file_name.file_name_length;
-
-			/* Only one case insensitive matching name allowed. */
-			if (name) {
-				ntfs_error(sb, "Found already allocated name "
-						"in phase 2. Please run chkdsk "
-						"and if that doesn't find any "
-						"errors please report you saw "
-						"this message to "
-						"linux-ntfs-dev@lists."
-						"sourceforge.net.");
-				unlock_page(page);
-				ntfs_unmap_page(page);
-				goto dir_err_out;
-			}
-
-			if (type != FILE_NAME_DOS)
-				name_size += len * sizeof(ntfschar);
-			name = kmalloc(name_size, GFP_NOFS);
-			if (!name) {
-				err = -ENOMEM;
-				goto unm_err_out;
-			}
-			name->mref = le64_to_cpu(ie->data.dir.indexed_file);
-			name->type = type;
-			if (type != FILE_NAME_DOS) {
-				name->len = len;
-				memcpy(name->name, ie->key.file_name.file_name,
-						len * sizeof(ntfschar));
-			} else
-				name->len = 0;
-			*res = name;
-		}
-		/*
-		 * Not a perfect match, need to do full blown collation so we
-		 * know which way in the B+tree we have to go.
-		 */
-		rc = ntfs_collate_names(uname, uname_len,
-				(ntfschar*)&ie->key.file_name.file_name,
-				ie->key.file_name.file_name_length, 1,
-				IGNORE_CASE, vol->upcase, vol->upcase_len);
-		/*
-		 * If uname collates before the name of the current entry, there
-		 * is definitely no such name in this index but we might need to
-		 * descend into the B+tree so we just break out of the loop.
-		 */
-		if (rc == -1)
-			break;
-		/* The names are not equal, continue the search. */
-		if (rc)
-			continue;
-		/*
-		 * Names match with case insensitive comparison, now try the
-		 * case sensitive comparison, which is required for proper
-		 * collation.
-		 */
-		rc = ntfs_collate_names(uname, uname_len,
-				(ntfschar*)&ie->key.file_name.file_name,
-				ie->key.file_name.file_name_length, 1,
-				CASE_SENSITIVE, vol->upcase, vol->upcase_len);
-		if (rc == -1)
-			break;
-		if (rc)
-			continue;
-		/*
-		 * Perfect match, this will never happen as the
-		 * ntfs_are_names_equal() call will have gotten a match but we
-		 * still treat it correctly.
-		 */
-		goto found_it2;
-	}
-	/*
-	 * We have finished with this index buffer without success. Check for
-	 * the presence of a child node.
-	 */
-	if (ie->flags & INDEX_ENTRY_NODE) {
-		if ((ia->index.flags & NODE_MASK) == LEAF_NODE) {
-			ntfs_error(sb, "Index entry with child node found in "
-					"a leaf node in directory inode 0x%lx.",
-					dir_ni->mft_no);
-			goto unm_err_out;
-		}
-		/* Child node present, descend into it. */
-		old_vcn = vcn;
-		vcn = sle64_to_cpup((sle64*)((u8*)ie +
-				le16_to_cpu(ie->length) - 8));
-		if (vcn >= 0) {
-			/* If vcn is in the same page cache page as old_vcn we
-			 * recycle the mapped page. */
-			if (old_vcn << vol->cluster_size_bits >>
-					PAGE_SHIFT == vcn <<
-					vol->cluster_size_bits >>
-					PAGE_SHIFT)
-				goto fast_descend_into_child_node;
-			unlock_page(page);
-			ntfs_unmap_page(page);
-			goto descend_into_child_node;
-		}
-		ntfs_error(sb, "Negative child node vcn in directory inode "
-				"0x%lx.", dir_ni->mft_no);
-		goto unm_err_out;
-	}
-	/*
-	 * No child node present, return -ENOENT, unless we have got a matching
-	 * name cached in name in which case return the mft reference
-	 * associated with it.
-	 */
-	if (name) {
-		unlock_page(page);
-		ntfs_unmap_page(page);
-		return name->mref;
-	}
-	ntfs_debug("Entry not found.");
-	err = -ENOENT;
-unm_err_out:
-	unlock_page(page);
-	ntfs_unmap_page(page);
-err_out:
-	if (!err)
-		err = -EIO;
-	if (ctx)
-		ntfs_attr_put_search_ctx(ctx);
-	if (m)
-		unmap_mft_record(dir_ni);
-	if (name) {
-		kfree(name);
-		*res = NULL;
-	}
-	return ERR_MREF(err);
-dir_err_out:
-	ntfs_error(sb, "Corrupt directory.  Aborting lookup.");
-	goto err_out;
-}
-
-#if 0
-
-// TODO: (AIA)
-// The algorithm embedded in this code will be required for the time when we
-// want to support adding of entries to directories, where we require correct
-// collation of file names in order not to cause corruption of the filesystem.
-
-/**
- * ntfs_lookup_inode_by_name - find an inode in a directory given its name
- * @dir_ni:	ntfs inode of the directory in which to search for the name
- * @uname:	Unicode name for which to search in the directory
- * @uname_len:	length of the name @uname in Unicode characters
- *
- * Look for an inode with name @uname in the directory with inode @dir_ni.
- * ntfs_lookup_inode_by_name() walks the contents of the directory looking for
- * the Unicode name. If the name is found in the directory, the corresponding
- * inode number (>= 0) is returned as a mft reference in cpu format, i.e. it
- * is a 64-bit number containing the sequence number.
- *
- * On error, a negative value is returned corresponding to the error code. In
- * particular if the inode is not found -ENOENT is returned. Note that you
- * can't just check the return value for being negative, you have to check the
- * inode number for being negative which you can extract using MREC(return
- * value).
- *
- * Note, @uname_len does not include the (optional) terminating NULL character.
- */
-u64 ntfs_lookup_inode_by_name(ntfs_inode *dir_ni, const ntfschar *uname,
-		const int uname_len)
-{
-	ntfs_volume *vol = dir_ni->vol;
-	struct super_block *sb = vol->sb;
-	MFT_RECORD *m;
-	INDEX_ROOT *ir;
-	INDEX_ENTRY *ie;
-	INDEX_ALLOCATION *ia;
-	u8 *index_end;
-	u64 mref;
-	ntfs_attr_search_ctx *ctx;
-	int err, rc;
-	IGNORE_CASE_BOOL ic;
-	VCN vcn, old_vcn;
-	struct address_space *ia_mapping;
-	struct page *page;
-	u8 *kaddr;
-
-	/* Get hold of the mft record for the directory. */
-	m = map_mft_record(dir_ni);
-	if (IS_ERR(m)) {
-		ntfs_error(sb, "map_mft_record() failed with error code %ld.",
-				-PTR_ERR(m));
-		return ERR_MREF(PTR_ERR(m));
-	}
-	ctx = ntfs_attr_get_search_ctx(dir_ni, m);
-	if (!ctx) {
-		err = -ENOMEM;
-		goto err_out;
-	}
-	/* Find the index root attribute in the mft record. */
-	err = ntfs_attr_lookup(AT_INDEX_ROOT, I30, 4, CASE_SENSITIVE, 0, NULL,
-			0, ctx);
-	if (unlikely(err)) {
-		if (err == -ENOENT) {
-			ntfs_error(sb, "Index root attribute missing in "
-					"directory inode 0x%lx.",
-					dir_ni->mft_no);
-			err = -EIO;
-		}
-		goto err_out;
-	}
-	/* Get to the index root value (it's been verified in read_inode). */
-	ir = (INDEX_ROOT*)((u8*)ctx->attr +
-			le16_to_cpu(ctx->attr->data.resident.value_offset));
-	index_end = (u8*)&ir->index + le32_to_cpu(ir->index.index_length);
-	/* The first index entry. */
-	ie = (INDEX_ENTRY*)((u8*)&ir->index +
-			le32_to_cpu(ir->index.entries_offset));
-	/*
-	 * Loop until we exceed valid memory (corruption case) or until we
-	 * reach the last entry.
-	 */
-	for (;; ie = (INDEX_ENTRY*)((u8*)ie + le16_to_cpu(ie->length))) {
-		/* Bounds checks. */
-		if ((u8*)ie < (u8*)ctx->mrec || (u8*)ie +
-				sizeof(INDEX_ENTRY_HEADER) > index_end ||
-				(u8*)ie + le16_to_cpu(ie->key_length) >
-				index_end)
-			goto dir_err_out;
-		/*
-		 * The last entry cannot contain a name. It can however contain
-		 * a pointer to a child node in the B+tree so we just break out.
-		 */
-		if (ie->flags & INDEX_ENTRY_END)
-			break;
-		/*
-		 * If the current entry has a name type of POSIX, the name is
-		 * case sensitive and not otherwise. This has the effect of us
-		 * not being able to access any POSIX file names which collate
-		 * after the non-POSIX one when they only differ in case, but
-		 * anyone doing screwy stuff like that deserves to burn in
-		 * hell... Doing that kind of stuff on NT4 actually causes
-		 * corruption on the partition even when using SP6a and Linux
-		 * is not involved at all.
-		 */
-		ic = ie->key.file_name.file_name_type ? IGNORE_CASE :
-				CASE_SENSITIVE;
-		/*
-		 * If the names match perfectly, we are done and return the
-		 * mft reference of the inode (i.e. the inode number together
-		 * with the sequence number for consistency checking. We
-		 * convert it to cpu format before returning.
-		 */
-		if (ntfs_are_names_equal(uname, uname_len,
-				(ntfschar*)&ie->key.file_name.file_name,
-				ie->key.file_name.file_name_length, ic,
-				vol->upcase, vol->upcase_len)) {
-found_it:
-			mref = le64_to_cpu(ie->data.dir.indexed_file);
-			ntfs_attr_put_search_ctx(ctx);
-			unmap_mft_record(dir_ni);
-			return mref;
-		}
-		/*
-		 * Not a perfect match, need to do full blown collation so we
-		 * know which way in the B+tree we have to go.
-		 */
-		rc = ntfs_collate_names(uname, uname_len,
-				(ntfschar*)&ie->key.file_name.file_name,
-				ie->key.file_name.file_name_length, 1,
-				IGNORE_CASE, vol->upcase, vol->upcase_len);
-		/*
-		 * If uname collates before the name of the current entry, there
-		 * is definitely no such name in this index but we might need to
-		 * descend into the B+tree so we just break out of the loop.
-		 */
-		if (rc == -1)
-			break;
-		/* The names are not equal, continue the search. */
-		if (rc)
-			continue;
-		/*
-		 * Names match with case insensitive comparison, now try the
-		 * case sensitive comparison, which is required for proper
-		 * collation.
-		 */
-		rc = ntfs_collate_names(uname, uname_len,
-				(ntfschar*)&ie->key.file_name.file_name,
-				ie->key.file_name.file_name_length, 1,
-				CASE_SENSITIVE, vol->upcase, vol->upcase_len);
-		if (rc == -1)
-			break;
-		if (rc)
-			continue;
-		/*
-		 * Perfect match, this will never happen as the
-		 * ntfs_are_names_equal() call will have gotten a match but we
-		 * still treat it correctly.
-		 */
-		goto found_it;
-	}
-	/*
-	 * We have finished with this index without success. Check for the
-	 * presence of a child node.
-	 */
-	if (!(ie->flags & INDEX_ENTRY_NODE)) {
-		/* No child node, return -ENOENT. */
-		err = -ENOENT;
-		goto err_out;
-	} /* Child node present, descend into it. */
-	/* Consistency check: Verify that an index allocation exists. */
-	if (!NInoIndexAllocPresent(dir_ni)) {
-		ntfs_error(sb, "No index allocation attribute but index entry "
-				"requires one. Directory inode 0x%lx is "
-				"corrupt or driver bug.", dir_ni->mft_no);
-		goto err_out;
-	}
-	/* Get the starting vcn of the index_block holding the child node. */
-	vcn = sle64_to_cpup((u8*)ie + le16_to_cpu(ie->length) - 8);
-	ia_mapping = VFS_I(dir_ni)->i_mapping;
-	/*
-	 * We are done with the index root and the mft record. Release them,
-	 * otherwise we deadlock with ntfs_map_page().
-	 */
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(dir_ni);
-	m = NULL;
-	ctx = NULL;
-descend_into_child_node:
-	/*
-	 * Convert vcn to index into the index allocation attribute in units
-	 * of PAGE_SIZE and map the page cache page, reading it from
-	 * disk if necessary.
-	 */
-	page = ntfs_map_page(ia_mapping, vcn <<
-			dir_ni->itype.index.vcn_size_bits >> PAGE_SHIFT);
-	if (IS_ERR(page)) {
-		ntfs_error(sb, "Failed to map directory index page, error %ld.",
-				-PTR_ERR(page));
-		err = PTR_ERR(page);
-		goto err_out;
-	}
-	lock_page(page);
-	kaddr = (u8*)page_address(page);
-fast_descend_into_child_node:
-	/* Get to the index allocation block. */
-	ia = (INDEX_ALLOCATION*)(kaddr + ((vcn <<
-			dir_ni->itype.index.vcn_size_bits) & ~PAGE_MASK));
-	/* Bounds checks. */
-	if ((u8*)ia < kaddr || (u8*)ia > kaddr + PAGE_SIZE) {
-		ntfs_error(sb, "Out of bounds check failed. Corrupt directory "
-				"inode 0x%lx or driver bug.", dir_ni->mft_no);
-		goto unm_err_out;
-	}
-	/* Catch multi sector transfer fixup errors. */
-	if (unlikely(!ntfs_is_indx_record(ia->magic))) {
-		ntfs_error(sb, "Directory index record with vcn 0x%llx is "
-				"corrupt.  Corrupt inode 0x%lx.  Run chkdsk.",
-				(unsigned long long)vcn, dir_ni->mft_no);
-		goto unm_err_out;
-	}
-	if (sle64_to_cpu(ia->index_block_vcn) != vcn) {
-		ntfs_error(sb, "Actual VCN (0x%llx) of index buffer is "
-				"different from expected VCN (0x%llx). "
-				"Directory inode 0x%lx is corrupt or driver "
-				"bug.", (unsigned long long)
-				sle64_to_cpu(ia->index_block_vcn),
-				(unsigned long long)vcn, dir_ni->mft_no);
-		goto unm_err_out;
-	}
-	if (le32_to_cpu(ia->index.allocated_size) + 0x18 !=
-			dir_ni->itype.index.block_size) {
-		ntfs_error(sb, "Index buffer (VCN 0x%llx) of directory inode "
-				"0x%lx has a size (%u) differing from the "
-				"directory specified size (%u). Directory "
-				"inode is corrupt or driver bug.",
-				(unsigned long long)vcn, dir_ni->mft_no,
-				le32_to_cpu(ia->index.allocated_size) + 0x18,
-				dir_ni->itype.index.block_size);
-		goto unm_err_out;
-	}
-	index_end = (u8*)ia + dir_ni->itype.index.block_size;
-	if (index_end > kaddr + PAGE_SIZE) {
-		ntfs_error(sb, "Index buffer (VCN 0x%llx) of directory inode "
-				"0x%lx crosses page boundary. Impossible! "
-				"Cannot access! This is probably a bug in the "
-				"driver.", (unsigned long long)vcn,
-				dir_ni->mft_no);
-		goto unm_err_out;
-	}
-	index_end = (u8*)&ia->index + le32_to_cpu(ia->index.index_length);
-	if (index_end > (u8*)ia + dir_ni->itype.index.block_size) {
-		ntfs_error(sb, "Size of index buffer (VCN 0x%llx) of directory "
-				"inode 0x%lx exceeds maximum size.",
-				(unsigned long long)vcn, dir_ni->mft_no);
-		goto unm_err_out;
-	}
-	/* The first index entry. */
-	ie = (INDEX_ENTRY*)((u8*)&ia->index +
-			le32_to_cpu(ia->index.entries_offset));
-	/*
-	 * Iterate similar to above big loop but applied to index buffer, thus
-	 * loop until we exceed valid memory (corruption case) or until we
-	 * reach the last entry.
-	 */
-	for (;; ie = (INDEX_ENTRY*)((u8*)ie + le16_to_cpu(ie->length))) {
-		/* Bounds check. */
-		if ((u8*)ie < (u8*)ia || (u8*)ie +
-				sizeof(INDEX_ENTRY_HEADER) > index_end ||
-				(u8*)ie + le16_to_cpu(ie->key_length) >
-				index_end) {
-			ntfs_error(sb, "Index entry out of bounds in "
-					"directory inode 0x%lx.",
-					dir_ni->mft_no);
-			goto unm_err_out;
-		}
-		/*
-		 * The last entry cannot contain a name. It can however contain
-		 * a pointer to a child node in the B+tree so we just break out.
-		 */
-		if (ie->flags & INDEX_ENTRY_END)
-			break;
-		/*
-		 * If the current entry has a name type of POSIX, the name is
-		 * case sensitive and not otherwise. This has the effect of us
-		 * not being able to access any POSIX file names which collate
-		 * after the non-POSIX one when they only differ in case, but
-		 * anyone doing screwy stuff like that deserves to burn in
-		 * hell... Doing that kind of stuff on NT4 actually causes
-		 * corruption on the partition even when using SP6a and Linux
-		 * is not involved at all.
-		 */
-		ic = ie->key.file_name.file_name_type ? IGNORE_CASE :
-				CASE_SENSITIVE;
-		/*
-		 * If the names match perfectly, we are done and return the
-		 * mft reference of the inode (i.e. the inode number together
-		 * with the sequence number for consistency checking. We
-		 * convert it to cpu format before returning.
-		 */
-		if (ntfs_are_names_equal(uname, uname_len,
-				(ntfschar*)&ie->key.file_name.file_name,
-				ie->key.file_name.file_name_length, ic,
-				vol->upcase, vol->upcase_len)) {
-found_it2:
-			mref = le64_to_cpu(ie->data.dir.indexed_file);
-			unlock_page(page);
-			ntfs_unmap_page(page);
-			return mref;
-		}
-		/*
-		 * Not a perfect match, need to do full blown collation so we
-		 * know which way in the B+tree we have to go.
-		 */
-		rc = ntfs_collate_names(uname, uname_len,
-				(ntfschar*)&ie->key.file_name.file_name,
-				ie->key.file_name.file_name_length, 1,
-				IGNORE_CASE, vol->upcase, vol->upcase_len);
-		/*
-		 * If uname collates before the name of the current entry, there
-		 * is definitely no such name in this index but we might need to
-		 * descend into the B+tree so we just break out of the loop.
-		 */
-		if (rc == -1)
-			break;
-		/* The names are not equal, continue the search. */
-		if (rc)
-			continue;
-		/*
-		 * Names match with case insensitive comparison, now try the
-		 * case sensitive comparison, which is required for proper
-		 * collation.
-		 */
-		rc = ntfs_collate_names(uname, uname_len,
-				(ntfschar*)&ie->key.file_name.file_name,
-				ie->key.file_name.file_name_length, 1,
-				CASE_SENSITIVE, vol->upcase, vol->upcase_len);
-		if (rc == -1)
-			break;
-		if (rc)
-			continue;
-		/*
-		 * Perfect match, this will never happen as the
-		 * ntfs_are_names_equal() call will have gotten a match but we
-		 * still treat it correctly.
-		 */
-		goto found_it2;
-	}
-	/*
-	 * We have finished with this index buffer without success. Check for
-	 * the presence of a child node.
-	 */
-	if (ie->flags & INDEX_ENTRY_NODE) {
-		if ((ia->index.flags & NODE_MASK) == LEAF_NODE) {
-			ntfs_error(sb, "Index entry with child node found in "
-					"a leaf node in directory inode 0x%lx.",
-					dir_ni->mft_no);
-			goto unm_err_out;
-		}
-		/* Child node present, descend into it. */
-		old_vcn = vcn;
-		vcn = sle64_to_cpup((u8*)ie + le16_to_cpu(ie->length) - 8);
-		if (vcn >= 0) {
-			/* If vcn is in the same page cache page as old_vcn we
-			 * recycle the mapped page. */
-			if (old_vcn << vol->cluster_size_bits >>
-					PAGE_SHIFT == vcn <<
-					vol->cluster_size_bits >>
-					PAGE_SHIFT)
-				goto fast_descend_into_child_node;
-			unlock_page(page);
-			ntfs_unmap_page(page);
-			goto descend_into_child_node;
-		}
-		ntfs_error(sb, "Negative child node vcn in directory inode "
-				"0x%lx.", dir_ni->mft_no);
-		goto unm_err_out;
-	}
-	/* No child node, return -ENOENT. */
-	ntfs_debug("Entry not found.");
-	err = -ENOENT;
-unm_err_out:
-	unlock_page(page);
-	ntfs_unmap_page(page);
-err_out:
-	if (!err)
-		err = -EIO;
-	if (ctx)
-		ntfs_attr_put_search_ctx(ctx);
-	if (m)
-		unmap_mft_record(dir_ni);
-	return ERR_MREF(err);
-dir_err_out:
-	ntfs_error(sb, "Corrupt directory. Aborting lookup.");
-	goto err_out;
-}
-
-#endif
-
-/**
- * ntfs_filldir - ntfs specific filldir method
- * @vol:	current ntfs volume
- * @ndir:	ntfs inode of current directory
- * @ia_page:	page in which the index allocation buffer @ie is in resides
- * @ie:		current index entry
- * @name:	buffer to use for the converted name
- * @actor:	what to feed the entries to
- *
- * Convert the Unicode @name to the loaded NLS and pass it to the @filldir
- * callback.
- *
- * If @ia_page is not NULL it is the locked page containing the index
- * allocation block containing the index entry @ie.
- *
- * Note, we drop (and then reacquire) the page lock on @ia_page across the
- * @filldir() call otherwise we would deadlock with NFSd when it calls ->lookup
- * since ntfs_lookup() will lock the same page.  As an optimization, we do not
- * retake the lock if we are returning a non-zero value as ntfs_readdir()
- * would need to drop the lock immediately anyway.
- */
-static inline int ntfs_filldir(ntfs_volume *vol,
-		ntfs_inode *ndir, struct page *ia_page, INDEX_ENTRY *ie,
-		u8 *name, struct dir_context *actor)
-{
-	unsigned long mref;
-	int name_len;
-	unsigned dt_type;
-	FILE_NAME_TYPE_FLAGS name_type;
-
-	name_type = ie->key.file_name.file_name_type;
-	if (name_type == FILE_NAME_DOS) {
-		ntfs_debug("Skipping DOS name space entry.");
-		return 0;
-	}
-	if (MREF_LE(ie->data.dir.indexed_file) == FILE_root) {
-		ntfs_debug("Skipping root directory self reference entry.");
-		return 0;
-	}
-	if (MREF_LE(ie->data.dir.indexed_file) < FILE_first_user &&
-			!NVolShowSystemFiles(vol)) {
-		ntfs_debug("Skipping system file.");
-		return 0;
-	}
-	name_len = ntfs_ucstonls(vol, (ntfschar*)&ie->key.file_name.file_name,
-			ie->key.file_name.file_name_length, &name,
-			NTFS_MAX_NAME_LEN * NLS_MAX_CHARSET_SIZE + 1);
-	if (name_len <= 0) {
-		ntfs_warning(vol->sb, "Skipping unrepresentable inode 0x%llx.",
-				(long long)MREF_LE(ie->data.dir.indexed_file));
-		return 0;
-	}
-	if (ie->key.file_name.file_attributes &
-			FILE_ATTR_DUP_FILE_NAME_INDEX_PRESENT)
-		dt_type = DT_DIR;
-	else
-		dt_type = DT_REG;
-	mref = MREF_LE(ie->data.dir.indexed_file);
-	/*
-	 * Drop the page lock otherwise we deadlock with NFS when it calls
-	 * ->lookup since ntfs_lookup() will lock the same page.
-	 */
-	if (ia_page)
-		unlock_page(ia_page);
-	ntfs_debug("Calling filldir for %s with len %i, fpos 0x%llx, inode "
-			"0x%lx, DT_%s.", name, name_len, actor->pos, mref,
-			dt_type == DT_DIR ? "DIR" : "REG");
-	if (!dir_emit(actor, name, name_len, mref, dt_type))
-		return 1;
-	/* Relock the page but not if we are aborting ->readdir. */
-	if (ia_page)
-		lock_page(ia_page);
-	return 0;
-}
-
-/*
- * We use the same basic approach as the old NTFS driver, i.e. we parse the
- * index root entries and then the index allocation entries that are marked
- * as in use in the index bitmap.
- *
- * While this will return the names in random order this doesn't matter for
- * ->readdir but OTOH results in a faster ->readdir.
- *
- * VFS calls ->readdir without BKL but with i_mutex held. This protects the VFS
- * parts (e.g. ->f_pos and ->i_size, and it also protects against directory
- * modifications).
- *
- * Locking:  - Caller must hold i_mutex on the directory.
- *	     - Each page cache page in the index allocation mapping must be
- *	       locked whilst being accessed otherwise we may find a corrupt
- *	       page due to it being under ->writepage at the moment which
- *	       applies the mst protection fixups before writing out and then
- *	       removes them again after the write is complete after which it 
- *	       unlocks the page.
- */
-static int ntfs_readdir(struct file *file, struct dir_context *actor)
-{
-	s64 ia_pos, ia_start, prev_ia_pos, bmp_pos;
-	loff_t i_size;
-	struct inode *bmp_vi, *vdir = file_inode(file);
-	struct super_block *sb = vdir->i_sb;
-	ntfs_inode *ndir = NTFS_I(vdir);
-	ntfs_volume *vol = NTFS_SB(sb);
-	MFT_RECORD *m;
-	INDEX_ROOT *ir = NULL;
-	INDEX_ENTRY *ie;
-	INDEX_ALLOCATION *ia;
-	u8 *name = NULL;
-	int rc, err, ir_pos, cur_bmp_pos;
-	struct address_space *ia_mapping, *bmp_mapping;
-	struct page *bmp_page = NULL, *ia_page = NULL;
-	u8 *kaddr, *bmp, *index_end;
-	ntfs_attr_search_ctx *ctx;
-
-	ntfs_debug("Entering for inode 0x%lx, fpos 0x%llx.",
-			vdir->i_ino, actor->pos);
-	rc = err = 0;
-	/* Are we at end of dir yet? */
-	i_size = i_size_read(vdir);
-	if (actor->pos >= i_size + vol->mft_record_size)
-		return 0;
-	/* Emulate . and .. for all directories. */
-	if (!dir_emit_dots(file, actor))
-		return 0;
-	m = NULL;
-	ctx = NULL;
-	/*
-	 * Allocate a buffer to store the current name being processed
-	 * converted to format determined by current NLS.
-	 */
-	name = kmalloc(NTFS_MAX_NAME_LEN * NLS_MAX_CHARSET_SIZE + 1, GFP_NOFS);
-	if (unlikely(!name)) {
-		err = -ENOMEM;
-		goto err_out;
-	}
-	/* Are we jumping straight into the index allocation attribute? */
-	if (actor->pos >= vol->mft_record_size)
-		goto skip_index_root;
-	/* Get hold of the mft record for the directory. */
-	m = map_mft_record(ndir);
-	if (IS_ERR(m)) {
-		err = PTR_ERR(m);
-		m = NULL;
-		goto err_out;
-	}
-	ctx = ntfs_attr_get_search_ctx(ndir, m);
-	if (unlikely(!ctx)) {
-		err = -ENOMEM;
-		goto err_out;
-	}
-	/* Get the offset into the index root attribute. */
-	ir_pos = (s64)actor->pos;
-	/* Find the index root attribute in the mft record. */
-	err = ntfs_attr_lookup(AT_INDEX_ROOT, I30, 4, CASE_SENSITIVE, 0, NULL,
-			0, ctx);
-	if (unlikely(err)) {
-		ntfs_error(sb, "Index root attribute missing in directory "
-				"inode 0x%lx.", vdir->i_ino);
-		goto err_out;
-	}
-	/*
-	 * Copy the index root attribute value to a buffer so that we can put
-	 * the search context and unmap the mft record before calling the
-	 * filldir() callback.  We need to do this because of NFSd which calls
-	 * ->lookup() from its filldir callback() and this causes NTFS to
-	 * deadlock as ntfs_lookup() maps the mft record of the directory and
-	 * we have got it mapped here already.  The only solution is for us to
-	 * unmap the mft record here so that a call to ntfs_lookup() is able to
-	 * map the mft record without deadlocking.
-	 */
-	rc = le32_to_cpu(ctx->attr->data.resident.value_length);
-	ir = kmalloc(rc, GFP_NOFS);
-	if (unlikely(!ir)) {
-		err = -ENOMEM;
-		goto err_out;
-	}
-	/* Copy the index root value (it has been verified in read_inode). */
-	memcpy(ir, (u8*)ctx->attr +
-			le16_to_cpu(ctx->attr->data.resident.value_offset), rc);
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(ndir);
-	ctx = NULL;
-	m = NULL;
-	index_end = (u8*)&ir->index + le32_to_cpu(ir->index.index_length);
-	/* The first index entry. */
-	ie = (INDEX_ENTRY*)((u8*)&ir->index +
-			le32_to_cpu(ir->index.entries_offset));
-	/*
-	 * Loop until we exceed valid memory (corruption case) or until we
-	 * reach the last entry or until filldir tells us it has had enough
-	 * or signals an error (both covered by the rc test).
-	 */
-	for (;; ie = (INDEX_ENTRY*)((u8*)ie + le16_to_cpu(ie->length))) {
-		ntfs_debug("In index root, offset 0x%zx.", (u8*)ie - (u8*)ir);
-		/* Bounds checks. */
-		if (unlikely((u8*)ie < (u8*)ir || (u8*)ie +
-				sizeof(INDEX_ENTRY_HEADER) > index_end ||
-				(u8*)ie + le16_to_cpu(ie->key_length) >
-				index_end))
-			goto err_out;
-		/* The last entry cannot contain a name. */
-		if (ie->flags & INDEX_ENTRY_END)
-			break;
-		/* Skip index root entry if continuing previous readdir. */
-		if (ir_pos > (u8*)ie - (u8*)ir)
-			continue;
-		/* Advance the position even if going to skip the entry. */
-		actor->pos = (u8*)ie - (u8*)ir;
-		/* Submit the name to the filldir callback. */
-		rc = ntfs_filldir(vol, ndir, NULL, ie, name, actor);
-		if (rc) {
-			kfree(ir);
-			goto abort;
-		}
-	}
-	/* We are done with the index root and can free the buffer. */
-	kfree(ir);
-	ir = NULL;
-	/* If there is no index allocation attribute we are finished. */
-	if (!NInoIndexAllocPresent(ndir))
-		goto EOD;
-	/* Advance fpos to the beginning of the index allocation. */
-	actor->pos = vol->mft_record_size;
-skip_index_root:
-	kaddr = NULL;
-	prev_ia_pos = -1LL;
-	/* Get the offset into the index allocation attribute. */
-	ia_pos = (s64)actor->pos - vol->mft_record_size;
-	ia_mapping = vdir->i_mapping;
-	ntfs_debug("Inode 0x%lx, getting index bitmap.", vdir->i_ino);
-	bmp_vi = ntfs_attr_iget(vdir, AT_BITMAP, I30, 4);
-	if (IS_ERR(bmp_vi)) {
-		ntfs_error(sb, "Failed to get bitmap attribute.");
-		err = PTR_ERR(bmp_vi);
-		goto err_out;
-	}
-	bmp_mapping = bmp_vi->i_mapping;
-	/* Get the starting bitmap bit position and sanity check it. */
-	bmp_pos = ia_pos >> ndir->itype.index.block_size_bits;
-	if (unlikely(bmp_pos >> 3 >= i_size_read(bmp_vi))) {
-		ntfs_error(sb, "Current index allocation position exceeds "
-				"index bitmap size.");
-		goto iput_err_out;
-	}
-	/* Get the starting bit position in the current bitmap page. */
-	cur_bmp_pos = bmp_pos & ((PAGE_SIZE * 8) - 1);
-	bmp_pos &= ~(u64)((PAGE_SIZE * 8) - 1);
-get_next_bmp_page:
-	ntfs_debug("Reading bitmap with page index 0x%llx, bit ofs 0x%llx",
-			(unsigned long long)bmp_pos >> (3 + PAGE_SHIFT),
-			(unsigned long long)bmp_pos &
-			(unsigned long long)((PAGE_SIZE * 8) - 1));
-	bmp_page = ntfs_map_page(bmp_mapping,
-			bmp_pos >> (3 + PAGE_SHIFT));
-	if (IS_ERR(bmp_page)) {
-		ntfs_error(sb, "Reading index bitmap failed.");
-		err = PTR_ERR(bmp_page);
-		bmp_page = NULL;
-		goto iput_err_out;
-	}
-	bmp = (u8*)page_address(bmp_page);
-	/* Find next index block in use. */
-	while (!(bmp[cur_bmp_pos >> 3] & (1 << (cur_bmp_pos & 7)))) {
-find_next_index_buffer:
-		cur_bmp_pos++;
-		/*
-		 * If we have reached the end of the bitmap page, get the next
-		 * page, and put away the old one.
-		 */
-		if (unlikely((cur_bmp_pos >> 3) >= PAGE_SIZE)) {
-			ntfs_unmap_page(bmp_page);
-			bmp_pos += PAGE_SIZE * 8;
-			cur_bmp_pos = 0;
-			goto get_next_bmp_page;
-		}
-		/* If we have reached the end of the bitmap, we are done. */
-		if (unlikely(((bmp_pos + cur_bmp_pos) >> 3) >= i_size))
-			goto unm_EOD;
-		ia_pos = (bmp_pos + cur_bmp_pos) <<
-				ndir->itype.index.block_size_bits;
-	}
-	ntfs_debug("Handling index buffer 0x%llx.",
-			(unsigned long long)bmp_pos + cur_bmp_pos);
-	/* If the current index buffer is in the same page we reuse the page. */
-	if ((prev_ia_pos & (s64)PAGE_MASK) !=
-			(ia_pos & (s64)PAGE_MASK)) {
-		prev_ia_pos = ia_pos;
-		if (likely(ia_page != NULL)) {
-			unlock_page(ia_page);
-			ntfs_unmap_page(ia_page);
-		}
-		/*
-		 * Map the page cache page containing the current ia_pos,
-		 * reading it from disk if necessary.
-		 */
-		ia_page = ntfs_map_page(ia_mapping, ia_pos >> PAGE_SHIFT);
-		if (IS_ERR(ia_page)) {
-			ntfs_error(sb, "Reading index allocation data failed.");
-			err = PTR_ERR(ia_page);
-			ia_page = NULL;
-			goto err_out;
-		}
-		lock_page(ia_page);
-		kaddr = (u8*)page_address(ia_page);
-	}
-	/* Get the current index buffer. */
-	ia = (INDEX_ALLOCATION*)(kaddr + (ia_pos & ~PAGE_MASK &
-					  ~(s64)(ndir->itype.index.block_size - 1)));
-	/* Bounds checks. */
-	if (unlikely((u8*)ia < kaddr || (u8*)ia > kaddr + PAGE_SIZE)) {
-		ntfs_error(sb, "Out of bounds check failed. Corrupt directory "
-				"inode 0x%lx or driver bug.", vdir->i_ino);
-		goto err_out;
-	}
-	/* Catch multi sector transfer fixup errors. */
-	if (unlikely(!ntfs_is_indx_record(ia->magic))) {
-		ntfs_error(sb, "Directory index record with vcn 0x%llx is "
-				"corrupt.  Corrupt inode 0x%lx.  Run chkdsk.",
-				(unsigned long long)ia_pos >>
-				ndir->itype.index.vcn_size_bits, vdir->i_ino);
-		goto err_out;
-	}
-	if (unlikely(sle64_to_cpu(ia->index_block_vcn) != (ia_pos &
-			~(s64)(ndir->itype.index.block_size - 1)) >>
-			ndir->itype.index.vcn_size_bits)) {
-		ntfs_error(sb, "Actual VCN (0x%llx) of index buffer is "
-				"different from expected VCN (0x%llx). "
-				"Directory inode 0x%lx is corrupt or driver "
-				"bug. ", (unsigned long long)
-				sle64_to_cpu(ia->index_block_vcn),
-				(unsigned long long)ia_pos >>
-				ndir->itype.index.vcn_size_bits, vdir->i_ino);
-		goto err_out;
-	}
-	if (unlikely(le32_to_cpu(ia->index.allocated_size) + 0x18 !=
-			ndir->itype.index.block_size)) {
-		ntfs_error(sb, "Index buffer (VCN 0x%llx) of directory inode "
-				"0x%lx has a size (%u) differing from the "
-				"directory specified size (%u). Directory "
-				"inode is corrupt or driver bug.",
-				(unsigned long long)ia_pos >>
-				ndir->itype.index.vcn_size_bits, vdir->i_ino,
-				le32_to_cpu(ia->index.allocated_size) + 0x18,
-				ndir->itype.index.block_size);
-		goto err_out;
-	}
-	index_end = (u8*)ia + ndir->itype.index.block_size;
-	if (unlikely(index_end > kaddr + PAGE_SIZE)) {
-		ntfs_error(sb, "Index buffer (VCN 0x%llx) of directory inode "
-				"0x%lx crosses page boundary. Impossible! "
-				"Cannot access! This is probably a bug in the "
-				"driver.", (unsigned long long)ia_pos >>
-				ndir->itype.index.vcn_size_bits, vdir->i_ino);
-		goto err_out;
-	}
-	ia_start = ia_pos & ~(s64)(ndir->itype.index.block_size - 1);
-	index_end = (u8*)&ia->index + le32_to_cpu(ia->index.index_length);
-	if (unlikely(index_end > (u8*)ia + ndir->itype.index.block_size)) {
-		ntfs_error(sb, "Size of index buffer (VCN 0x%llx) of directory "
-				"inode 0x%lx exceeds maximum size.",
-				(unsigned long long)ia_pos >>
-				ndir->itype.index.vcn_size_bits, vdir->i_ino);
-		goto err_out;
-	}
-	/* The first index entry in this index buffer. */
-	ie = (INDEX_ENTRY*)((u8*)&ia->index +
-			le32_to_cpu(ia->index.entries_offset));
-	/*
-	 * Loop until we exceed valid memory (corruption case) or until we
-	 * reach the last entry or until filldir tells us it has had enough
-	 * or signals an error (both covered by the rc test).
-	 */
-	for (;; ie = (INDEX_ENTRY*)((u8*)ie + le16_to_cpu(ie->length))) {
-		ntfs_debug("In index allocation, offset 0x%llx.",
-				(unsigned long long)ia_start +
-				(unsigned long long)((u8*)ie - (u8*)ia));
-		/* Bounds checks. */
-		if (unlikely((u8*)ie < (u8*)ia || (u8*)ie +
-				sizeof(INDEX_ENTRY_HEADER) > index_end ||
-				(u8*)ie + le16_to_cpu(ie->key_length) >
-				index_end))
-			goto err_out;
-		/* The last entry cannot contain a name. */
-		if (ie->flags & INDEX_ENTRY_END)
-			break;
-		/* Skip index block entry if continuing previous readdir. */
-		if (ia_pos - ia_start > (u8*)ie - (u8*)ia)
-			continue;
-		/* Advance the position even if going to skip the entry. */
-		actor->pos = (u8*)ie - (u8*)ia +
-				(sle64_to_cpu(ia->index_block_vcn) <<
-				ndir->itype.index.vcn_size_bits) +
-				vol->mft_record_size;
-		/*
-		 * Submit the name to the @filldir callback.  Note,
-		 * ntfs_filldir() drops the lock on @ia_page but it retakes it
-		 * before returning, unless a non-zero value is returned in
-		 * which case the page is left unlocked.
-		 */
-		rc = ntfs_filldir(vol, ndir, ia_page, ie, name, actor);
-		if (rc) {
-			/* @ia_page is already unlocked in this case. */
-			ntfs_unmap_page(ia_page);
-			ntfs_unmap_page(bmp_page);
-			iput(bmp_vi);
-			goto abort;
-		}
-	}
-	goto find_next_index_buffer;
-unm_EOD:
-	if (ia_page) {
-		unlock_page(ia_page);
-		ntfs_unmap_page(ia_page);
-	}
-	ntfs_unmap_page(bmp_page);
-	iput(bmp_vi);
-EOD:
-	/* We are finished, set fpos to EOD. */
-	actor->pos = i_size + vol->mft_record_size;
-abort:
-	kfree(name);
-	return 0;
-err_out:
-	if (bmp_page) {
-		ntfs_unmap_page(bmp_page);
-iput_err_out:
-		iput(bmp_vi);
-	}
-	if (ia_page) {
-		unlock_page(ia_page);
-		ntfs_unmap_page(ia_page);
-	}
-	kfree(ir);
-	kfree(name);
-	if (ctx)
-		ntfs_attr_put_search_ctx(ctx);
-	if (m)
-		unmap_mft_record(ndir);
-	if (!err)
-		err = -EIO;
-	ntfs_debug("Failed. Returning error code %i.", -err);
-	return err;
-}
-
-/**
- * ntfs_dir_open - called when an inode is about to be opened
- * @vi:		inode to be opened
- * @filp:	file structure describing the inode
- *
- * Limit directory size to the page cache limit on architectures where unsigned
- * long is 32-bits. This is the most we can do for now without overflowing the
- * page cache page index. Doing it this way means we don't run into problems
- * because of existing too large directories. It would be better to allow the
- * user to read the accessible part of the directory but I doubt very much
- * anyone is going to hit this check on a 32-bit architecture, so there is no
- * point in adding the extra complexity required to support this.
- *
- * On 64-bit architectures, the check is hopefully optimized away by the
- * compiler.
- */
-static int ntfs_dir_open(struct inode *vi, struct file *filp)
-{
-	if (sizeof(unsigned long) < 8) {
-		if (i_size_read(vi) > MAX_LFS_FILESIZE)
-			return -EFBIG;
-	}
-	return 0;
-}
-
-#ifdef NTFS_RW
-
-/**
- * ntfs_dir_fsync - sync a directory to disk
- * @filp:	directory to be synced
- * @start:	offset in bytes of the beginning of data range to sync
- * @end:	offset in bytes of the end of data range (inclusive)
- * @datasync:	if non-zero only flush user data and not metadata
- *
- * Data integrity sync of a directory to disk.  Used for fsync, fdatasync, and
- * msync system calls.  This function is based on file.c::ntfs_file_fsync().
- *
- * Write the mft record and all associated extent mft records as well as the
- * $INDEX_ALLOCATION and $BITMAP attributes and then sync the block device.
- *
- * If @datasync is true, we do not wait on the inode(s) to be written out
- * but we always wait on the page cache pages to be written out.
- *
- * Note: In the past @filp could be NULL so we ignore it as we don't need it
- * anyway.
- *
- * Locking: Caller must hold i_mutex on the inode.
- *
- * TODO: We should probably also write all attribute/index inodes associated
- * with this inode but since we have no simple way of getting to them we ignore
- * this problem for now.  We do write the $BITMAP attribute if it is present
- * which is the important one for a directory so things are not too bad.
- */
-static int ntfs_dir_fsync(struct file *filp, loff_t start, loff_t end,
-			  int datasync)
-{
-	struct inode *bmp_vi, *vi = filp->f_mapping->host;
-	int err, ret;
-	ntfs_attr na;
-
-	ntfs_debug("Entering for inode 0x%lx.", vi->i_ino);
-
-	err = file_write_and_wait_range(filp, start, end);
-	if (err)
-		return err;
-	inode_lock(vi);
-
-	BUG_ON(!S_ISDIR(vi->i_mode));
-	/* If the bitmap attribute inode is in memory sync it, too. */
-	na.mft_no = vi->i_ino;
-	na.type = AT_BITMAP;
-	na.name = I30;
-	na.name_len = 4;
-	bmp_vi = ilookup5(vi->i_sb, vi->i_ino, ntfs_test_inode, &na);
-	if (bmp_vi) {
- 		write_inode_now(bmp_vi, !datasync);
-		iput(bmp_vi);
-	}
-	ret = __ntfs_write_inode(vi, 1);
-	write_inode_now(vi, !datasync);
-	err = sync_blockdev(vi->i_sb->s_bdev);
-	if (unlikely(err && !ret))
-		ret = err;
-	if (likely(!ret))
-		ntfs_debug("Done.");
-	else
-		ntfs_warning(vi->i_sb, "Failed to f%ssync inode 0x%lx.  Error "
-				"%u.", datasync ? "data" : "", vi->i_ino, -ret);
-	inode_unlock(vi);
-	return ret;
-}
-
-#endif /* NTFS_RW */
-
-WRAP_DIR_ITER(ntfs_readdir) // FIXME!
-const struct file_operations ntfs_dir_ops = {
-	.llseek		= generic_file_llseek,	/* Seek inside directory. */
-	.read		= generic_read_dir,	/* Return -EISDIR. */
-	.iterate_shared	= shared_ntfs_readdir,	/* Read directory contents. */
-#ifdef NTFS_RW
-	.fsync		= ntfs_dir_fsync,	/* Sync a directory to disk. */
-#endif /* NTFS_RW */
-	/*.ioctl	= ,*/			/* Perform function on the
-						   mounted filesystem. */
-	.open		= ntfs_dir_open,	/* Open directory. */
-};
diff --git a/fs/ntfs/dir.h b/fs/ntfs/dir.h
deleted file mode 100644
index 0e326753df..0000000000
--- a/fs/ntfs/dir.h
+++ /dev/null
@@ -1,34 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * dir.h - Defines for directory handling in NTFS Linux kernel driver. Part of
- *	   the Linux-NTFS project.
- *
- * Copyright (c) 2002-2004 Anton Altaparmakov
- */
-
-#ifndef _LINUX_NTFS_DIR_H
-#define _LINUX_NTFS_DIR_H
-
-#include "layout.h"
-#include "inode.h"
-#include "types.h"
-
-/*
- * ntfs_name is used to return the file name to the caller of
- * ntfs_lookup_inode_by_name() in order for the caller (namei.c::ntfs_lookup())
- * to be able to deal with dcache aliasing issues.
- */
-typedef struct {
-	MFT_REF mref;
-	FILE_NAME_TYPE_FLAGS type;
-	u8 len;
-	ntfschar name[0];
-} __attribute__ ((__packed__)) ntfs_name;
-
-/* The little endian Unicode string $I30 as a global constant. */
-extern ntfschar I30[5];
-
-extern MFT_REF ntfs_lookup_inode_by_name(ntfs_inode *dir_ni,
-		const ntfschar *uname, const int uname_len, ntfs_name **res);
-
-#endif /* _LINUX_NTFS_FS_DIR_H */
diff --git a/fs/ntfs/endian.h b/fs/ntfs/endian.h
deleted file mode 100644
index f30c139bf9..0000000000
--- a/fs/ntfs/endian.h
+++ /dev/null
@@ -1,79 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * endian.h - Defines for endianness handling in NTFS Linux kernel driver.
- *	      Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2004 Anton Altaparmakov
- */
-
-#ifndef _LINUX_NTFS_ENDIAN_H
-#define _LINUX_NTFS_ENDIAN_H
-
-#include <asm/byteorder.h>
-#include "types.h"
-
-/*
- * Signed endianness conversion functions.
- */
-
-static inline s16 sle16_to_cpu(sle16 x)
-{
-	return le16_to_cpu((__force le16)x);
-}
-
-static inline s32 sle32_to_cpu(sle32 x)
-{
-	return le32_to_cpu((__force le32)x);
-}
-
-static inline s64 sle64_to_cpu(sle64 x)
-{
-	return le64_to_cpu((__force le64)x);
-}
-
-static inline s16 sle16_to_cpup(sle16 *x)
-{
-	return le16_to_cpu(*(__force le16*)x);
-}
-
-static inline s32 sle32_to_cpup(sle32 *x)
-{
-	return le32_to_cpu(*(__force le32*)x);
-}
-
-static inline s64 sle64_to_cpup(sle64 *x)
-{
-	return le64_to_cpu(*(__force le64*)x);
-}
-
-static inline sle16 cpu_to_sle16(s16 x)
-{
-	return (__force sle16)cpu_to_le16(x);
-}
-
-static inline sle32 cpu_to_sle32(s32 x)
-{
-	return (__force sle32)cpu_to_le32(x);
-}
-
-static inline sle64 cpu_to_sle64(s64 x)
-{
-	return (__force sle64)cpu_to_le64(x);
-}
-
-static inline sle16 cpu_to_sle16p(s16 *x)
-{
-	return (__force sle16)cpu_to_le16(*x);
-}
-
-static inline sle32 cpu_to_sle32p(s32 *x)
-{
-	return (__force sle32)cpu_to_le32(*x);
-}
-
-static inline sle64 cpu_to_sle64p(s64 *x)
-{
-	return (__force sle64)cpu_to_le64(*x);
-}
-
-#endif /* _LINUX_NTFS_ENDIAN_H */
diff --git a/fs/ntfs/file.c b/fs/ntfs/file.c
deleted file mode 100644
index 297c0b9db6..0000000000
--- a/fs/ntfs/file.c
+++ /dev/null
@@ -1,1997 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * file.c - NTFS kernel file operations.  Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2015 Anton Altaparmakov and Tuxera Inc.
- */
-
-#include <linux/blkdev.h>
-#include <linux/backing-dev.h>
-#include <linux/buffer_head.h>
-#include <linux/gfp.h>
-#include <linux/pagemap.h>
-#include <linux/pagevec.h>
-#include <linux/sched/signal.h>
-#include <linux/swap.h>
-#include <linux/uio.h>
-#include <linux/writeback.h>
-
-#include <asm/page.h>
-#include <linux/uaccess.h>
-
-#include "attrib.h"
-#include "bitmap.h"
-#include "inode.h"
-#include "debug.h"
-#include "lcnalloc.h"
-#include "malloc.h"
-#include "mft.h"
-#include "ntfs.h"
-
-/**
- * ntfs_file_open - called when an inode is about to be opened
- * @vi:		inode to be opened
- * @filp:	file structure describing the inode
- *
- * Limit file size to the page cache limit on architectures where unsigned long
- * is 32-bits. This is the most we can do for now without overflowing the page
- * cache page index. Doing it this way means we don't run into problems because
- * of existing too large files. It would be better to allow the user to read
- * the beginning of the file but I doubt very much anyone is going to hit this
- * check on a 32-bit architecture, so there is no point in adding the extra
- * complexity required to support this.
- *
- * On 64-bit architectures, the check is hopefully optimized away by the
- * compiler.
- *
- * After the check passes, just call generic_file_open() to do its work.
- */
-static int ntfs_file_open(struct inode *vi, struct file *filp)
-{
-	if (sizeof(unsigned long) < 8) {
-		if (i_size_read(vi) > MAX_LFS_FILESIZE)
-			return -EOVERFLOW;
-	}
-	return generic_file_open(vi, filp);
-}
-
-#ifdef NTFS_RW
-
-/**
- * ntfs_attr_extend_initialized - extend the initialized size of an attribute
- * @ni:			ntfs inode of the attribute to extend
- * @new_init_size:	requested new initialized size in bytes
- *
- * Extend the initialized size of an attribute described by the ntfs inode @ni
- * to @new_init_size bytes.  This involves zeroing any non-sparse space between
- * the old initialized size and @new_init_size both in the page cache and on
- * disk (if relevant complete pages are already uptodate in the page cache then
- * these are simply marked dirty).
- *
- * As a side-effect, the file size (vfs inode->i_size) may be incremented as,
- * in the resident attribute case, it is tied to the initialized size and, in
- * the non-resident attribute case, it may not fall below the initialized size.
- *
- * Note that if the attribute is resident, we do not need to touch the page
- * cache at all.  This is because if the page cache page is not uptodate we
- * bring it uptodate later, when doing the write to the mft record since we
- * then already have the page mapped.  And if the page is uptodate, the
- * non-initialized region will already have been zeroed when the page was
- * brought uptodate and the region may in fact already have been overwritten
- * with new data via mmap() based writes, so we cannot just zero it.  And since
- * POSIX specifies that the behaviour of resizing a file whilst it is mmap()ped
- * is unspecified, we choose not to do zeroing and thus we do not need to touch
- * the page at all.  For a more detailed explanation see ntfs_truncate() in
- * fs/ntfs/inode.c.
- *
- * Return 0 on success and -errno on error.  In the case that an error is
- * encountered it is possible that the initialized size will already have been
- * incremented some way towards @new_init_size but it is guaranteed that if
- * this is the case, the necessary zeroing will also have happened and that all
- * metadata is self-consistent.
- *
- * Locking: i_mutex on the vfs inode corrseponsind to the ntfs inode @ni must be
- *	    held by the caller.
- */
-static int ntfs_attr_extend_initialized(ntfs_inode *ni, const s64 new_init_size)
-{
-	s64 old_init_size;
-	loff_t old_i_size;
-	pgoff_t index, end_index;
-	unsigned long flags;
-	struct inode *vi = VFS_I(ni);
-	ntfs_inode *base_ni;
-	MFT_RECORD *m = NULL;
-	ATTR_RECORD *a;
-	ntfs_attr_search_ctx *ctx = NULL;
-	struct address_space *mapping;
-	struct page *page = NULL;
-	u8 *kattr;
-	int err;
-	u32 attr_len;
-
-	read_lock_irqsave(&ni->size_lock, flags);
-	old_init_size = ni->initialized_size;
-	old_i_size = i_size_read(vi);
-	BUG_ON(new_init_size > ni->allocated_size);
-	read_unlock_irqrestore(&ni->size_lock, flags);
-	ntfs_debug("Entering for i_ino 0x%lx, attribute type 0x%x, "
-			"old_initialized_size 0x%llx, "
-			"new_initialized_size 0x%llx, i_size 0x%llx.",
-			vi->i_ino, (unsigned)le32_to_cpu(ni->type),
-			(unsigned long long)old_init_size,
-			(unsigned long long)new_init_size, old_i_size);
-	if (!NInoAttr(ni))
-		base_ni = ni;
-	else
-		base_ni = ni->ext.base_ntfs_ino;
-	/* Use goto to reduce indentation and we need the label below anyway. */
-	if (NInoNonResident(ni))
-		goto do_non_resident_extend;
-	BUG_ON(old_init_size != old_i_size);
-	m = map_mft_record(base_ni);
-	if (IS_ERR(m)) {
-		err = PTR_ERR(m);
-		m = NULL;
-		goto err_out;
-	}
-	ctx = ntfs_attr_get_search_ctx(base_ni, m);
-	if (unlikely(!ctx)) {
-		err = -ENOMEM;
-		goto err_out;
-	}
-	err = ntfs_attr_lookup(ni->type, ni->name, ni->name_len,
-			CASE_SENSITIVE, 0, NULL, 0, ctx);
-	if (unlikely(err)) {
-		if (err == -ENOENT)
-			err = -EIO;
-		goto err_out;
-	}
-	m = ctx->mrec;
-	a = ctx->attr;
-	BUG_ON(a->non_resident);
-	/* The total length of the attribute value. */
-	attr_len = le32_to_cpu(a->data.resident.value_length);
-	BUG_ON(old_i_size != (loff_t)attr_len);
-	/*
-	 * Do the zeroing in the mft record and update the attribute size in
-	 * the mft record.
-	 */
-	kattr = (u8*)a + le16_to_cpu(a->data.resident.value_offset);
-	memset(kattr + attr_len, 0, new_init_size - attr_len);
-	a->data.resident.value_length = cpu_to_le32((u32)new_init_size);
-	/* Finally, update the sizes in the vfs and ntfs inodes. */
-	write_lock_irqsave(&ni->size_lock, flags);
-	i_size_write(vi, new_init_size);
-	ni->initialized_size = new_init_size;
-	write_unlock_irqrestore(&ni->size_lock, flags);
-	goto done;
-do_non_resident_extend:
-	/*
-	 * If the new initialized size @new_init_size exceeds the current file
-	 * size (vfs inode->i_size), we need to extend the file size to the
-	 * new initialized size.
-	 */
-	if (new_init_size > old_i_size) {
-		m = map_mft_record(base_ni);
-		if (IS_ERR(m)) {
-			err = PTR_ERR(m);
-			m = NULL;
-			goto err_out;
-		}
-		ctx = ntfs_attr_get_search_ctx(base_ni, m);
-		if (unlikely(!ctx)) {
-			err = -ENOMEM;
-			goto err_out;
-		}
-		err = ntfs_attr_lookup(ni->type, ni->name, ni->name_len,
-				CASE_SENSITIVE, 0, NULL, 0, ctx);
-		if (unlikely(err)) {
-			if (err == -ENOENT)
-				err = -EIO;
-			goto err_out;
-		}
-		m = ctx->mrec;
-		a = ctx->attr;
-		BUG_ON(!a->non_resident);
-		BUG_ON(old_i_size != (loff_t)
-				sle64_to_cpu(a->data.non_resident.data_size));
-		a->data.non_resident.data_size = cpu_to_sle64(new_init_size);
-		flush_dcache_mft_record_page(ctx->ntfs_ino);
-		mark_mft_record_dirty(ctx->ntfs_ino);
-		/* Update the file size in the vfs inode. */
-		i_size_write(vi, new_init_size);
-		ntfs_attr_put_search_ctx(ctx);
-		ctx = NULL;
-		unmap_mft_record(base_ni);
-		m = NULL;
-	}
-	mapping = vi->i_mapping;
-	index = old_init_size >> PAGE_SHIFT;
-	end_index = (new_init_size + PAGE_SIZE - 1) >> PAGE_SHIFT;
-	do {
-		/*
-		 * Read the page.  If the page is not present, this will zero
-		 * the uninitialized regions for us.
-		 */
-		page = read_mapping_page(mapping, index, NULL);
-		if (IS_ERR(page)) {
-			err = PTR_ERR(page);
-			goto init_err_out;
-		}
-		/*
-		 * Update the initialized size in the ntfs inode.  This is
-		 * enough to make ntfs_writepage() work.
-		 */
-		write_lock_irqsave(&ni->size_lock, flags);
-		ni->initialized_size = (s64)(index + 1) << PAGE_SHIFT;
-		if (ni->initialized_size > new_init_size)
-			ni->initialized_size = new_init_size;
-		write_unlock_irqrestore(&ni->size_lock, flags);
-		/* Set the page dirty so it gets written out. */
-		set_page_dirty(page);
-		put_page(page);
-		/*
-		 * Play nice with the vm and the rest of the system.  This is
-		 * very much needed as we can potentially be modifying the
-		 * initialised size from a very small value to a really huge
-		 * value, e.g.
-		 *	f = open(somefile, O_TRUNC);
-		 *	truncate(f, 10GiB);
-		 *	seek(f, 10GiB);
-		 *	write(f, 1);
-		 * And this would mean we would be marking dirty hundreds of
-		 * thousands of pages or as in the above example more than
-		 * two and a half million pages!
-		 *
-		 * TODO: For sparse pages could optimize this workload by using
-		 * the FsMisc / MiscFs page bit as a "PageIsSparse" bit.  This
-		 * would be set in read_folio for sparse pages and here we would
-		 * not need to mark dirty any pages which have this bit set.
-		 * The only caveat is that we have to clear the bit everywhere
-		 * where we allocate any clusters that lie in the page or that
-		 * contain the page.
-		 *
-		 * TODO: An even greater optimization would be for us to only
-		 * call read_folio() on pages which are not in sparse regions as
-		 * determined from the runlist.  This would greatly reduce the
-		 * number of pages we read and make dirty in the case of sparse
-		 * files.
-		 */
-		balance_dirty_pages_ratelimited(mapping);
-		cond_resched();
-	} while (++index < end_index);
-	read_lock_irqsave(&ni->size_lock, flags);
-	BUG_ON(ni->initialized_size != new_init_size);
-	read_unlock_irqrestore(&ni->size_lock, flags);
-	/* Now bring in sync the initialized_size in the mft record. */
-	m = map_mft_record(base_ni);
-	if (IS_ERR(m)) {
-		err = PTR_ERR(m);
-		m = NULL;
-		goto init_err_out;
-	}
-	ctx = ntfs_attr_get_search_ctx(base_ni, m);
-	if (unlikely(!ctx)) {
-		err = -ENOMEM;
-		goto init_err_out;
-	}
-	err = ntfs_attr_lookup(ni->type, ni->name, ni->name_len,
-			CASE_SENSITIVE, 0, NULL, 0, ctx);
-	if (unlikely(err)) {
-		if (err == -ENOENT)
-			err = -EIO;
-		goto init_err_out;
-	}
-	m = ctx->mrec;
-	a = ctx->attr;
-	BUG_ON(!a->non_resident);
-	a->data.non_resident.initialized_size = cpu_to_sle64(new_init_size);
-done:
-	flush_dcache_mft_record_page(ctx->ntfs_ino);
-	mark_mft_record_dirty(ctx->ntfs_ino);
-	if (ctx)
-		ntfs_attr_put_search_ctx(ctx);
-	if (m)
-		unmap_mft_record(base_ni);
-	ntfs_debug("Done, initialized_size 0x%llx, i_size 0x%llx.",
-			(unsigned long long)new_init_size, i_size_read(vi));
-	return 0;
-init_err_out:
-	write_lock_irqsave(&ni->size_lock, flags);
-	ni->initialized_size = old_init_size;
-	write_unlock_irqrestore(&ni->size_lock, flags);
-err_out:
-	if (ctx)
-		ntfs_attr_put_search_ctx(ctx);
-	if (m)
-		unmap_mft_record(base_ni);
-	ntfs_debug("Failed.  Returning error code %i.", err);
-	return err;
-}
-
-static ssize_t ntfs_prepare_file_for_write(struct kiocb *iocb,
-		struct iov_iter *from)
-{
-	loff_t pos;
-	s64 end, ll;
-	ssize_t err;
-	unsigned long flags;
-	struct file *file = iocb->ki_filp;
-	struct inode *vi = file_inode(file);
-	ntfs_inode *ni = NTFS_I(vi);
-	ntfs_volume *vol = ni->vol;
-
-	ntfs_debug("Entering for i_ino 0x%lx, attribute type 0x%x, pos "
-			"0x%llx, count 0x%zx.", vi->i_ino,
-			(unsigned)le32_to_cpu(ni->type),
-			(unsigned long long)iocb->ki_pos,
-			iov_iter_count(from));
-	err = generic_write_checks(iocb, from);
-	if (unlikely(err <= 0))
-		goto out;
-	/*
-	 * All checks have passed.  Before we start doing any writing we want
-	 * to abort any totally illegal writes.
-	 */
-	BUG_ON(NInoMstProtected(ni));
-	BUG_ON(ni->type != AT_DATA);
-	/* If file is encrypted, deny access, just like NT4. */
-	if (NInoEncrypted(ni)) {
-		/* Only $DATA attributes can be encrypted. */
-		/*
-		 * Reminder for later: Encrypted files are _always_
-		 * non-resident so that the content can always be encrypted.
-		 */
-		ntfs_debug("Denying write access to encrypted file.");
-		err = -EACCES;
-		goto out;
-	}
-	if (NInoCompressed(ni)) {
-		/* Only unnamed $DATA attribute can be compressed. */
-		BUG_ON(ni->name_len);
-		/*
-		 * Reminder for later: If resident, the data is not actually
-		 * compressed.  Only on the switch to non-resident does
-		 * compression kick in.  This is in contrast to encrypted files
-		 * (see above).
-		 */
-		ntfs_error(vi->i_sb, "Writing to compressed files is not "
-				"implemented yet.  Sorry.");
-		err = -EOPNOTSUPP;
-		goto out;
-	}
-	err = file_remove_privs(file);
-	if (unlikely(err))
-		goto out;
-	/*
-	 * Our ->update_time method always succeeds thus file_update_time()
-	 * cannot fail either so there is no need to check the return code.
-	 */
-	file_update_time(file);
-	pos = iocb->ki_pos;
-	/* The first byte after the last cluster being written to. */
-	end = (pos + iov_iter_count(from) + vol->cluster_size_mask) &
-			~(u64)vol->cluster_size_mask;
-	/*
-	 * If the write goes beyond the allocated size, extend the allocation
-	 * to cover the whole of the write, rounded up to the nearest cluster.
-	 */
-	read_lock_irqsave(&ni->size_lock, flags);
-	ll = ni->allocated_size;
-	read_unlock_irqrestore(&ni->size_lock, flags);
-	if (end > ll) {
-		/*
-		 * Extend the allocation without changing the data size.
-		 *
-		 * Note we ensure the allocation is big enough to at least
-		 * write some data but we do not require the allocation to be
-		 * complete, i.e. it may be partial.
-		 */
-		ll = ntfs_attr_extend_allocation(ni, end, -1, pos);
-		if (likely(ll >= 0)) {
-			BUG_ON(pos >= ll);
-			/* If the extension was partial truncate the write. */
-			if (end > ll) {
-				ntfs_debug("Truncating write to inode 0x%lx, "
-						"attribute type 0x%x, because "
-						"the allocation was only "
-						"partially extended.",
-						vi->i_ino, (unsigned)
-						le32_to_cpu(ni->type));
-				iov_iter_truncate(from, ll - pos);
-			}
-		} else {
-			err = ll;
-			read_lock_irqsave(&ni->size_lock, flags);
-			ll = ni->allocated_size;
-			read_unlock_irqrestore(&ni->size_lock, flags);
-			/* Perform a partial write if possible or fail. */
-			if (pos < ll) {
-				ntfs_debug("Truncating write to inode 0x%lx "
-						"attribute type 0x%x, because "
-						"extending the allocation "
-						"failed (error %d).",
-						vi->i_ino, (unsigned)
-						le32_to_cpu(ni->type),
-						(int)-err);
-				iov_iter_truncate(from, ll - pos);
-			} else {
-				if (err != -ENOSPC)
-					ntfs_error(vi->i_sb, "Cannot perform "
-							"write to inode "
-							"0x%lx, attribute "
-							"type 0x%x, because "
-							"extending the "
-							"allocation failed "
-							"(error %ld).",
-							vi->i_ino, (unsigned)
-							le32_to_cpu(ni->type),
-							(long)-err);
-				else
-					ntfs_debug("Cannot perform write to "
-							"inode 0x%lx, "
-							"attribute type 0x%x, "
-							"because there is not "
-							"space left.",
-							vi->i_ino, (unsigned)
-							le32_to_cpu(ni->type));
-				goto out;
-			}
-		}
-	}
-	/*
-	 * If the write starts beyond the initialized size, extend it up to the
-	 * beginning of the write and initialize all non-sparse space between
-	 * the old initialized size and the new one.  This automatically also
-	 * increments the vfs inode->i_size to keep it above or equal to the
-	 * initialized_size.
-	 */
-	read_lock_irqsave(&ni->size_lock, flags);
-	ll = ni->initialized_size;
-	read_unlock_irqrestore(&ni->size_lock, flags);
-	if (pos > ll) {
-		/*
-		 * Wait for ongoing direct i/o to complete before proceeding.
-		 * New direct i/o cannot start as we hold i_mutex.
-		 */
-		inode_dio_wait(vi);
-		err = ntfs_attr_extend_initialized(ni, pos);
-		if (unlikely(err < 0))
-			ntfs_error(vi->i_sb, "Cannot perform write to inode "
-					"0x%lx, attribute type 0x%x, because "
-					"extending the initialized size "
-					"failed (error %d).", vi->i_ino,
-					(unsigned)le32_to_cpu(ni->type),
-					(int)-err);
-	}
-out:
-	return err;
-}
-
-/**
- * __ntfs_grab_cache_pages - obtain a number of locked pages
- * @mapping:	address space mapping from which to obtain page cache pages
- * @index:	starting index in @mapping at which to begin obtaining pages
- * @nr_pages:	number of page cache pages to obtain
- * @pages:	array of pages in which to return the obtained page cache pages
- * @cached_page: allocated but as yet unused page
- *
- * Obtain @nr_pages locked page cache pages from the mapping @mapping and
- * starting at index @index.
- *
- * If a page is newly created, add it to lru list
- *
- * Note, the page locks are obtained in ascending page index order.
- */
-static inline int __ntfs_grab_cache_pages(struct address_space *mapping,
-		pgoff_t index, const unsigned nr_pages, struct page **pages,
-		struct page **cached_page)
-{
-	int err, nr;
-
-	BUG_ON(!nr_pages);
-	err = nr = 0;
-	do {
-		pages[nr] = find_get_page_flags(mapping, index, FGP_LOCK |
-				FGP_ACCESSED);
-		if (!pages[nr]) {
-			if (!*cached_page) {
-				*cached_page = page_cache_alloc(mapping);
-				if (unlikely(!*cached_page)) {
-					err = -ENOMEM;
-					goto err_out;
-				}
-			}
-			err = add_to_page_cache_lru(*cached_page, mapping,
-				   index,
-				   mapping_gfp_constraint(mapping, GFP_KERNEL));
-			if (unlikely(err)) {
-				if (err == -EEXIST)
-					continue;
-				goto err_out;
-			}
-			pages[nr] = *cached_page;
-			*cached_page = NULL;
-		}
-		index++;
-		nr++;
-	} while (nr < nr_pages);
-out:
-	return err;
-err_out:
-	while (nr > 0) {
-		unlock_page(pages[--nr]);
-		put_page(pages[nr]);
-	}
-	goto out;
-}
-
-static inline void ntfs_submit_bh_for_read(struct buffer_head *bh)
-{
-	lock_buffer(bh);
-	get_bh(bh);
-	bh->b_end_io = end_buffer_read_sync;
-	submit_bh(REQ_OP_READ, bh);
-}
-
-/**
- * ntfs_prepare_pages_for_non_resident_write - prepare pages for receiving data
- * @pages:	array of destination pages
- * @nr_pages:	number of pages in @pages
- * @pos:	byte position in file at which the write begins
- * @bytes:	number of bytes to be written
- *
- * This is called for non-resident attributes from ntfs_file_buffered_write()
- * with i_mutex held on the inode (@pages[0]->mapping->host).  There are
- * @nr_pages pages in @pages which are locked but not kmap()ped.  The source
- * data has not yet been copied into the @pages.
- * 
- * Need to fill any holes with actual clusters, allocate buffers if necessary,
- * ensure all the buffers are mapped, and bring uptodate any buffers that are
- * only partially being written to.
- *
- * If @nr_pages is greater than one, we are guaranteed that the cluster size is
- * greater than PAGE_SIZE, that all pages in @pages are entirely inside
- * the same cluster and that they are the entirety of that cluster, and that
- * the cluster is sparse, i.e. we need to allocate a cluster to fill the hole.
- *
- * i_size is not to be modified yet.
- *
- * Return 0 on success or -errno on error.
- */
-static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
-		unsigned nr_pages, s64 pos, size_t bytes)
-{
-	VCN vcn, highest_vcn = 0, cpos, cend, bh_cpos, bh_cend;
-	LCN lcn;
-	s64 bh_pos, vcn_len, end, initialized_size;
-	sector_t lcn_block;
-	struct folio *folio;
-	struct inode *vi;
-	ntfs_inode *ni, *base_ni = NULL;
-	ntfs_volume *vol;
-	runlist_element *rl, *rl2;
-	struct buffer_head *bh, *head, *wait[2], **wait_bh = wait;
-	ntfs_attr_search_ctx *ctx = NULL;
-	MFT_RECORD *m = NULL;
-	ATTR_RECORD *a = NULL;
-	unsigned long flags;
-	u32 attr_rec_len = 0;
-	unsigned blocksize, u;
-	int err, mp_size;
-	bool rl_write_locked, was_hole, is_retry;
-	unsigned char blocksize_bits;
-	struct {
-		u8 runlist_merged:1;
-		u8 mft_attr_mapped:1;
-		u8 mp_rebuilt:1;
-		u8 attr_switched:1;
-	} status = { 0, 0, 0, 0 };
-
-	BUG_ON(!nr_pages);
-	BUG_ON(!pages);
-	BUG_ON(!*pages);
-	vi = pages[0]->mapping->host;
-	ni = NTFS_I(vi);
-	vol = ni->vol;
-	ntfs_debug("Entering for inode 0x%lx, attribute type 0x%x, start page "
-			"index 0x%lx, nr_pages 0x%x, pos 0x%llx, bytes 0x%zx.",
-			vi->i_ino, ni->type, pages[0]->index, nr_pages,
-			(long long)pos, bytes);
-	blocksize = vol->sb->s_blocksize;
-	blocksize_bits = vol->sb->s_blocksize_bits;
-	rl_write_locked = false;
-	rl = NULL;
-	err = 0;
-	vcn = lcn = -1;
-	vcn_len = 0;
-	lcn_block = -1;
-	was_hole = false;
-	cpos = pos >> vol->cluster_size_bits;
-	end = pos + bytes;
-	cend = (end + vol->cluster_size - 1) >> vol->cluster_size_bits;
-	/*
-	 * Loop over each buffer in each folio.  Use goto to
-	 * reduce indentation.
-	 */
-	u = 0;
-do_next_folio:
-	folio = page_folio(pages[u]);
-	bh_pos = folio_pos(folio);
-	head = folio_buffers(folio);
-	if (!head)
-		/*
-		 * create_empty_buffers() will create uptodate/dirty
-		 * buffers if the folio is uptodate/dirty.
-		 */
-		head = create_empty_buffers(folio, blocksize, 0);
-	bh = head;
-	do {
-		VCN cdelta;
-		s64 bh_end;
-		unsigned bh_cofs;
-
-		/* Clear buffer_new on all buffers to reinitialise state. */
-		if (buffer_new(bh))
-			clear_buffer_new(bh);
-		bh_end = bh_pos + blocksize;
-		bh_cpos = bh_pos >> vol->cluster_size_bits;
-		bh_cofs = bh_pos & vol->cluster_size_mask;
-		if (buffer_mapped(bh)) {
-			/*
-			 * The buffer is already mapped.  If it is uptodate,
-			 * ignore it.
-			 */
-			if (buffer_uptodate(bh))
-				continue;
-			/*
-			 * The buffer is not uptodate.  If the folio is uptodate
-			 * set the buffer uptodate and otherwise ignore it.
-			 */
-			if (folio_test_uptodate(folio)) {
-				set_buffer_uptodate(bh);
-				continue;
-			}
-			/*
-			 * Neither the folio nor the buffer are uptodate.  If
-			 * the buffer is only partially being written to, we
-			 * need to read it in before the write, i.e. now.
-			 */
-			if ((bh_pos < pos && bh_end > pos) ||
-					(bh_pos < end && bh_end > end)) {
-				/*
-				 * If the buffer is fully or partially within
-				 * the initialized size, do an actual read.
-				 * Otherwise, simply zero the buffer.
-				 */
-				read_lock_irqsave(&ni->size_lock, flags);
-				initialized_size = ni->initialized_size;
-				read_unlock_irqrestore(&ni->size_lock, flags);
-				if (bh_pos < initialized_size) {
-					ntfs_submit_bh_for_read(bh);
-					*wait_bh++ = bh;
-				} else {
-					folio_zero_range(folio, bh_offset(bh),
-							blocksize);
-					set_buffer_uptodate(bh);
-				}
-			}
-			continue;
-		}
-		/* Unmapped buffer.  Need to map it. */
-		bh->b_bdev = vol->sb->s_bdev;
-		/*
-		 * If the current buffer is in the same clusters as the map
-		 * cache, there is no need to check the runlist again.  The
-		 * map cache is made up of @vcn, which is the first cached file
-		 * cluster, @vcn_len which is the number of cached file
-		 * clusters, @lcn is the device cluster corresponding to @vcn,
-		 * and @lcn_block is the block number corresponding to @lcn.
-		 */
-		cdelta = bh_cpos - vcn;
-		if (likely(!cdelta || (cdelta > 0 && cdelta < vcn_len))) {
-map_buffer_cached:
-			BUG_ON(lcn < 0);
-			bh->b_blocknr = lcn_block +
-					(cdelta << (vol->cluster_size_bits -
-					blocksize_bits)) +
-					(bh_cofs >> blocksize_bits);
-			set_buffer_mapped(bh);
-			/*
-			 * If the folio is uptodate so is the buffer.  If the
-			 * buffer is fully outside the write, we ignore it if
-			 * it was already allocated and we mark it dirty so it
-			 * gets written out if we allocated it.  On the other
-			 * hand, if we allocated the buffer but we are not
-			 * marking it dirty we set buffer_new so we can do
-			 * error recovery.
-			 */
-			if (folio_test_uptodate(folio)) {
-				if (!buffer_uptodate(bh))
-					set_buffer_uptodate(bh);
-				if (unlikely(was_hole)) {
-					/* We allocated the buffer. */
-					clean_bdev_bh_alias(bh);
-					if (bh_end <= pos || bh_pos >= end)
-						mark_buffer_dirty(bh);
-					else
-						set_buffer_new(bh);
-				}
-				continue;
-			}
-			/* Page is _not_ uptodate. */
-			if (likely(!was_hole)) {
-				/*
-				 * Buffer was already allocated.  If it is not
-				 * uptodate and is only partially being written
-				 * to, we need to read it in before the write,
-				 * i.e. now.
-				 */
-				if (!buffer_uptodate(bh) && bh_pos < end &&
-						bh_end > pos &&
-						(bh_pos < pos ||
-						bh_end > end)) {
-					/*
-					 * If the buffer is fully or partially
-					 * within the initialized size, do an
-					 * actual read.  Otherwise, simply zero
-					 * the buffer.
-					 */
-					read_lock_irqsave(&ni->size_lock,
-							flags);
-					initialized_size = ni->initialized_size;
-					read_unlock_irqrestore(&ni->size_lock,
-							flags);
-					if (bh_pos < initialized_size) {
-						ntfs_submit_bh_for_read(bh);
-						*wait_bh++ = bh;
-					} else {
-						folio_zero_range(folio,
-								bh_offset(bh),
-								blocksize);
-						set_buffer_uptodate(bh);
-					}
-				}
-				continue;
-			}
-			/* We allocated the buffer. */
-			clean_bdev_bh_alias(bh);
-			/*
-			 * If the buffer is fully outside the write, zero it,
-			 * set it uptodate, and mark it dirty so it gets
-			 * written out.  If it is partially being written to,
-			 * zero region surrounding the write but leave it to
-			 * commit write to do anything else.  Finally, if the
-			 * buffer is fully being overwritten, do nothing.
-			 */
-			if (bh_end <= pos || bh_pos >= end) {
-				if (!buffer_uptodate(bh)) {
-					folio_zero_range(folio, bh_offset(bh),
-							blocksize);
-					set_buffer_uptodate(bh);
-				}
-				mark_buffer_dirty(bh);
-				continue;
-			}
-			set_buffer_new(bh);
-			if (!buffer_uptodate(bh) &&
-					(bh_pos < pos || bh_end > end)) {
-				u8 *kaddr;
-				unsigned pofs;
-					
-				kaddr = kmap_local_folio(folio, 0);
-				if (bh_pos < pos) {
-					pofs = bh_pos & ~PAGE_MASK;
-					memset(kaddr + pofs, 0, pos - bh_pos);
-				}
-				if (bh_end > end) {
-					pofs = end & ~PAGE_MASK;
-					memset(kaddr + pofs, 0, bh_end - end);
-				}
-				kunmap_local(kaddr);
-				flush_dcache_folio(folio);
-			}
-			continue;
-		}
-		/*
-		 * Slow path: this is the first buffer in the cluster.  If it
-		 * is outside allocated size and is not uptodate, zero it and
-		 * set it uptodate.
-		 */
-		read_lock_irqsave(&ni->size_lock, flags);
-		initialized_size = ni->allocated_size;
-		read_unlock_irqrestore(&ni->size_lock, flags);
-		if (bh_pos > initialized_size) {
-			if (folio_test_uptodate(folio)) {
-				if (!buffer_uptodate(bh))
-					set_buffer_uptodate(bh);
-			} else if (!buffer_uptodate(bh)) {
-				folio_zero_range(folio, bh_offset(bh),
-						blocksize);
-				set_buffer_uptodate(bh);
-			}
-			continue;
-		}
-		is_retry = false;
-		if (!rl) {
-			down_read(&ni->runlist.lock);
-retry_remap:
-			rl = ni->runlist.rl;
-		}
-		if (likely(rl != NULL)) {
-			/* Seek to element containing target cluster. */
-			while (rl->length && rl[1].vcn <= bh_cpos)
-				rl++;
-			lcn = ntfs_rl_vcn_to_lcn(rl, bh_cpos);
-			if (likely(lcn >= 0)) {
-				/*
-				 * Successful remap, setup the map cache and
-				 * use that to deal with the buffer.
-				 */
-				was_hole = false;
-				vcn = bh_cpos;
-				vcn_len = rl[1].vcn - vcn;
-				lcn_block = lcn << (vol->cluster_size_bits -
-						blocksize_bits);
-				cdelta = 0;
-				/*
-				 * If the number of remaining clusters touched
-				 * by the write is smaller or equal to the
-				 * number of cached clusters, unlock the
-				 * runlist as the map cache will be used from
-				 * now on.
-				 */
-				if (likely(vcn + vcn_len >= cend)) {
-					if (rl_write_locked) {
-						up_write(&ni->runlist.lock);
-						rl_write_locked = false;
-					} else
-						up_read(&ni->runlist.lock);
-					rl = NULL;
-				}
-				goto map_buffer_cached;
-			}
-		} else
-			lcn = LCN_RL_NOT_MAPPED;
-		/*
-		 * If it is not a hole and not out of bounds, the runlist is
-		 * probably unmapped so try to map it now.
-		 */
-		if (unlikely(lcn != LCN_HOLE && lcn != LCN_ENOENT)) {
-			if (likely(!is_retry && lcn == LCN_RL_NOT_MAPPED)) {
-				/* Attempt to map runlist. */
-				if (!rl_write_locked) {
-					/*
-					 * We need the runlist locked for
-					 * writing, so if it is locked for
-					 * reading relock it now and retry in
-					 * case it changed whilst we dropped
-					 * the lock.
-					 */
-					up_read(&ni->runlist.lock);
-					down_write(&ni->runlist.lock);
-					rl_write_locked = true;
-					goto retry_remap;
-				}
-				err = ntfs_map_runlist_nolock(ni, bh_cpos,
-						NULL);
-				if (likely(!err)) {
-					is_retry = true;
-					goto retry_remap;
-				}
-				/*
-				 * If @vcn is out of bounds, pretend @lcn is
-				 * LCN_ENOENT.  As long as the buffer is out
-				 * of bounds this will work fine.
-				 */
-				if (err == -ENOENT) {
-					lcn = LCN_ENOENT;
-					err = 0;
-					goto rl_not_mapped_enoent;
-				}
-			} else
-				err = -EIO;
-			/* Failed to map the buffer, even after retrying. */
-			bh->b_blocknr = -1;
-			ntfs_error(vol->sb, "Failed to write to inode 0x%lx, "
-					"attribute type 0x%x, vcn 0x%llx, "
-					"vcn offset 0x%x, because its "
-					"location on disk could not be "
-					"determined%s (error code %i).",
-					ni->mft_no, ni->type,
-					(unsigned long long)bh_cpos,
-					(unsigned)bh_pos &
-					vol->cluster_size_mask,
-					is_retry ? " even after retrying" : "",
-					err);
-			break;
-		}
-rl_not_mapped_enoent:
-		/*
-		 * The buffer is in a hole or out of bounds.  We need to fill
-		 * the hole, unless the buffer is in a cluster which is not
-		 * touched by the write, in which case we just leave the buffer
-		 * unmapped.  This can only happen when the cluster size is
-		 * less than the page cache size.
-		 */
-		if (unlikely(vol->cluster_size < PAGE_SIZE)) {
-			bh_cend = (bh_end + vol->cluster_size - 1) >>
-					vol->cluster_size_bits;
-			if ((bh_cend <= cpos || bh_cpos >= cend)) {
-				bh->b_blocknr = -1;
-				/*
-				 * If the buffer is uptodate we skip it.  If it
-				 * is not but the folio is uptodate, we can set
-				 * the buffer uptodate.  If the folio is not
-				 * uptodate, we can clear the buffer and set it
-				 * uptodate.  Whether this is worthwhile is
-				 * debatable and this could be removed.
-				 */
-				if (folio_test_uptodate(folio)) {
-					if (!buffer_uptodate(bh))
-						set_buffer_uptodate(bh);
-				} else if (!buffer_uptodate(bh)) {
-					folio_zero_range(folio, bh_offset(bh),
-						blocksize);
-					set_buffer_uptodate(bh);
-				}
-				continue;
-			}
-		}
-		/*
-		 * Out of bounds buffer is invalid if it was not really out of
-		 * bounds.
-		 */
-		BUG_ON(lcn != LCN_HOLE);
-		/*
-		 * We need the runlist locked for writing, so if it is locked
-		 * for reading relock it now and retry in case it changed
-		 * whilst we dropped the lock.
-		 */
-		BUG_ON(!rl);
-		if (!rl_write_locked) {
-			up_read(&ni->runlist.lock);
-			down_write(&ni->runlist.lock);
-			rl_write_locked = true;
-			goto retry_remap;
-		}
-		/* Find the previous last allocated cluster. */
-		BUG_ON(rl->lcn != LCN_HOLE);
-		lcn = -1;
-		rl2 = rl;
-		while (--rl2 >= ni->runlist.rl) {
-			if (rl2->lcn >= 0) {
-				lcn = rl2->lcn + rl2->length;
-				break;
-			}
-		}
-		rl2 = ntfs_cluster_alloc(vol, bh_cpos, 1, lcn, DATA_ZONE,
-				false);
-		if (IS_ERR(rl2)) {
-			err = PTR_ERR(rl2);
-			ntfs_debug("Failed to allocate cluster, error code %i.",
-					err);
-			break;
-		}
-		lcn = rl2->lcn;
-		rl = ntfs_runlists_merge(ni->runlist.rl, rl2);
-		if (IS_ERR(rl)) {
-			err = PTR_ERR(rl);
-			if (err != -ENOMEM)
-				err = -EIO;
-			if (ntfs_cluster_free_from_rl(vol, rl2)) {
-				ntfs_error(vol->sb, "Failed to release "
-						"allocated cluster in error "
-						"code path.  Run chkdsk to "
-						"recover the lost cluster.");
-				NVolSetErrors(vol);
-			}
-			ntfs_free(rl2);
-			break;
-		}
-		ni->runlist.rl = rl;
-		status.runlist_merged = 1;
-		ntfs_debug("Allocated cluster, lcn 0x%llx.",
-				(unsigned long long)lcn);
-		/* Map and lock the mft record and get the attribute record. */
-		if (!NInoAttr(ni))
-			base_ni = ni;
-		else
-			base_ni = ni->ext.base_ntfs_ino;
-		m = map_mft_record(base_ni);
-		if (IS_ERR(m)) {
-			err = PTR_ERR(m);
-			break;
-		}
-		ctx = ntfs_attr_get_search_ctx(base_ni, m);
-		if (unlikely(!ctx)) {
-			err = -ENOMEM;
-			unmap_mft_record(base_ni);
-			break;
-		}
-		status.mft_attr_mapped = 1;
-		err = ntfs_attr_lookup(ni->type, ni->name, ni->name_len,
-				CASE_SENSITIVE, bh_cpos, NULL, 0, ctx);
-		if (unlikely(err)) {
-			if (err == -ENOENT)
-				err = -EIO;
-			break;
-		}
-		m = ctx->mrec;
-		a = ctx->attr;
-		/*
-		 * Find the runlist element with which the attribute extent
-		 * starts.  Note, we cannot use the _attr_ version because we
-		 * have mapped the mft record.  That is ok because we know the
-		 * runlist fragment must be mapped already to have ever gotten
-		 * here, so we can just use the _rl_ version.
-		 */
-		vcn = sle64_to_cpu(a->data.non_resident.lowest_vcn);
-		rl2 = ntfs_rl_find_vcn_nolock(rl, vcn);
-		BUG_ON(!rl2);
-		BUG_ON(!rl2->length);
-		BUG_ON(rl2->lcn < LCN_HOLE);
-		highest_vcn = sle64_to_cpu(a->data.non_resident.highest_vcn);
-		/*
-		 * If @highest_vcn is zero, calculate the real highest_vcn
-		 * (which can really be zero).
-		 */
-		if (!highest_vcn)
-			highest_vcn = (sle64_to_cpu(
-					a->data.non_resident.allocated_size) >>
-					vol->cluster_size_bits) - 1;
-		/*
-		 * Determine the size of the mapping pairs array for the new
-		 * extent, i.e. the old extent with the hole filled.
-		 */
-		mp_size = ntfs_get_size_for_mapping_pairs(vol, rl2, vcn,
-				highest_vcn);
-		if (unlikely(mp_size <= 0)) {
-			if (!(err = mp_size))
-				err = -EIO;
-			ntfs_debug("Failed to get size for mapping pairs "
-					"array, error code %i.", err);
-			break;
-		}
-		/*
-		 * Resize the attribute record to fit the new mapping pairs
-		 * array.
-		 */
-		attr_rec_len = le32_to_cpu(a->length);
-		err = ntfs_attr_record_resize(m, a, mp_size + le16_to_cpu(
-				a->data.non_resident.mapping_pairs_offset));
-		if (unlikely(err)) {
-			BUG_ON(err != -ENOSPC);
-			// TODO: Deal with this by using the current attribute
-			// and fill it with as much of the mapping pairs
-			// array as possible.  Then loop over each attribute
-			// extent rewriting the mapping pairs arrays as we go
-			// along and if when we reach the end we have not
-			// enough space, try to resize the last attribute
-			// extent and if even that fails, add a new attribute
-			// extent.
-			// We could also try to resize at each step in the hope
-			// that we will not need to rewrite every single extent.
-			// Note, we may need to decompress some extents to fill
-			// the runlist as we are walking the extents...
-			ntfs_error(vol->sb, "Not enough space in the mft "
-					"record for the extended attribute "
-					"record.  This case is not "
-					"implemented yet.");
-			err = -EOPNOTSUPP;
-			break ;
-		}
-		status.mp_rebuilt = 1;
-		/*
-		 * Generate the mapping pairs array directly into the attribute
-		 * record.
-		 */
-		err = ntfs_mapping_pairs_build(vol, (u8*)a + le16_to_cpu(
-				a->data.non_resident.mapping_pairs_offset),
-				mp_size, rl2, vcn, highest_vcn, NULL);
-		if (unlikely(err)) {
-			ntfs_error(vol->sb, "Cannot fill hole in inode 0x%lx, "
-					"attribute type 0x%x, because building "
-					"the mapping pairs failed with error "
-					"code %i.", vi->i_ino,
-					(unsigned)le32_to_cpu(ni->type), err);
-			err = -EIO;
-			break;
-		}
-		/* Update the highest_vcn but only if it was not set. */
-		if (unlikely(!a->data.non_resident.highest_vcn))
-			a->data.non_resident.highest_vcn =
-					cpu_to_sle64(highest_vcn);
-		/*
-		 * If the attribute is sparse/compressed, update the compressed
-		 * size in the ntfs_inode structure and the attribute record.
-		 */
-		if (likely(NInoSparse(ni) || NInoCompressed(ni))) {
-			/*
-			 * If we are not in the first attribute extent, switch
-			 * to it, but first ensure the changes will make it to
-			 * disk later.
-			 */
-			if (a->data.non_resident.lowest_vcn) {
-				flush_dcache_mft_record_page(ctx->ntfs_ino);
-				mark_mft_record_dirty(ctx->ntfs_ino);
-				ntfs_attr_reinit_search_ctx(ctx);
-				err = ntfs_attr_lookup(ni->type, ni->name,
-						ni->name_len, CASE_SENSITIVE,
-						0, NULL, 0, ctx);
-				if (unlikely(err)) {
-					status.attr_switched = 1;
-					break;
-				}
-				/* @m is not used any more so do not set it. */
-				a = ctx->attr;
-			}
-			write_lock_irqsave(&ni->size_lock, flags);
-			ni->itype.compressed.size += vol->cluster_size;
-			a->data.non_resident.compressed_size =
-					cpu_to_sle64(ni->itype.compressed.size);
-			write_unlock_irqrestore(&ni->size_lock, flags);
-		}
-		/* Ensure the changes make it to disk. */
-		flush_dcache_mft_record_page(ctx->ntfs_ino);
-		mark_mft_record_dirty(ctx->ntfs_ino);
-		ntfs_attr_put_search_ctx(ctx);
-		unmap_mft_record(base_ni);
-		/* Successfully filled the hole. */
-		status.runlist_merged = 0;
-		status.mft_attr_mapped = 0;
-		status.mp_rebuilt = 0;
-		/* Setup the map cache and use that to deal with the buffer. */
-		was_hole = true;
-		vcn = bh_cpos;
-		vcn_len = 1;
-		lcn_block = lcn << (vol->cluster_size_bits - blocksize_bits);
-		cdelta = 0;
-		/*
-		 * If the number of remaining clusters in the @pages is smaller
-		 * or equal to the number of cached clusters, unlock the
-		 * runlist as the map cache will be used from now on.
-		 */
-		if (likely(vcn + vcn_len >= cend)) {
-			up_write(&ni->runlist.lock);
-			rl_write_locked = false;
-			rl = NULL;
-		}
-		goto map_buffer_cached;
-	} while (bh_pos += blocksize, (bh = bh->b_this_page) != head);
-	/* If there are no errors, do the next page. */
-	if (likely(!err && ++u < nr_pages))
-		goto do_next_folio;
-	/* If there are no errors, release the runlist lock if we took it. */
-	if (likely(!err)) {
-		if (unlikely(rl_write_locked)) {
-			up_write(&ni->runlist.lock);
-			rl_write_locked = false;
-		} else if (unlikely(rl))
-			up_read(&ni->runlist.lock);
-		rl = NULL;
-	}
-	/* If we issued read requests, let them complete. */
-	read_lock_irqsave(&ni->size_lock, flags);
-	initialized_size = ni->initialized_size;
-	read_unlock_irqrestore(&ni->size_lock, flags);
-	while (wait_bh > wait) {
-		bh = *--wait_bh;
-		wait_on_buffer(bh);
-		if (likely(buffer_uptodate(bh))) {
-			folio = bh->b_folio;
-			bh_pos = folio_pos(folio) + bh_offset(bh);
-			/*
-			 * If the buffer overflows the initialized size, need
-			 * to zero the overflowing region.
-			 */
-			if (unlikely(bh_pos + blocksize > initialized_size)) {
-				int ofs = 0;
-
-				if (likely(bh_pos < initialized_size))
-					ofs = initialized_size - bh_pos;
-				folio_zero_segment(folio, bh_offset(bh) + ofs,
-						blocksize);
-			}
-		} else /* if (unlikely(!buffer_uptodate(bh))) */
-			err = -EIO;
-	}
-	if (likely(!err)) {
-		/* Clear buffer_new on all buffers. */
-		u = 0;
-		do {
-			bh = head = page_buffers(pages[u]);
-			do {
-				if (buffer_new(bh))
-					clear_buffer_new(bh);
-			} while ((bh = bh->b_this_page) != head);
-		} while (++u < nr_pages);
-		ntfs_debug("Done.");
-		return err;
-	}
-	if (status.attr_switched) {
-		/* Get back to the attribute extent we modified. */
-		ntfs_attr_reinit_search_ctx(ctx);
-		if (ntfs_attr_lookup(ni->type, ni->name, ni->name_len,
-				CASE_SENSITIVE, bh_cpos, NULL, 0, ctx)) {
-			ntfs_error(vol->sb, "Failed to find required "
-					"attribute extent of attribute in "
-					"error code path.  Run chkdsk to "
-					"recover.");
-			write_lock_irqsave(&ni->size_lock, flags);
-			ni->itype.compressed.size += vol->cluster_size;
-			write_unlock_irqrestore(&ni->size_lock, flags);
-			flush_dcache_mft_record_page(ctx->ntfs_ino);
-			mark_mft_record_dirty(ctx->ntfs_ino);
-			/*
-			 * The only thing that is now wrong is the compressed
-			 * size of the base attribute extent which chkdsk
-			 * should be able to fix.
-			 */
-			NVolSetErrors(vol);
-		} else {
-			m = ctx->mrec;
-			a = ctx->attr;
-			status.attr_switched = 0;
-		}
-	}
-	/*
-	 * If the runlist has been modified, need to restore it by punching a
-	 * hole into it and we then need to deallocate the on-disk cluster as
-	 * well.  Note, we only modify the runlist if we are able to generate a
-	 * new mapping pairs array, i.e. only when the mapped attribute extent
-	 * is not switched.
-	 */
-	if (status.runlist_merged && !status.attr_switched) {
-		BUG_ON(!rl_write_locked);
-		/* Make the file cluster we allocated sparse in the runlist. */
-		if (ntfs_rl_punch_nolock(vol, &ni->runlist, bh_cpos, 1)) {
-			ntfs_error(vol->sb, "Failed to punch hole into "
-					"attribute runlist in error code "
-					"path.  Run chkdsk to recover the "
-					"lost cluster.");
-			NVolSetErrors(vol);
-		} else /* if (success) */ {
-			status.runlist_merged = 0;
-			/*
-			 * Deallocate the on-disk cluster we allocated but only
-			 * if we succeeded in punching its vcn out of the
-			 * runlist.
-			 */
-			down_write(&vol->lcnbmp_lock);
-			if (ntfs_bitmap_clear_bit(vol->lcnbmp_ino, lcn)) {
-				ntfs_error(vol->sb, "Failed to release "
-						"allocated cluster in error "
-						"code path.  Run chkdsk to "
-						"recover the lost cluster.");
-				NVolSetErrors(vol);
-			}
-			up_write(&vol->lcnbmp_lock);
-		}
-	}
-	/*
-	 * Resize the attribute record to its old size and rebuild the mapping
-	 * pairs array.  Note, we only can do this if the runlist has been
-	 * restored to its old state which also implies that the mapped
-	 * attribute extent is not switched.
-	 */
-	if (status.mp_rebuilt && !status.runlist_merged) {
-		if (ntfs_attr_record_resize(m, a, attr_rec_len)) {
-			ntfs_error(vol->sb, "Failed to restore attribute "
-					"record in error code path.  Run "
-					"chkdsk to recover.");
-			NVolSetErrors(vol);
-		} else /* if (success) */ {
-			if (ntfs_mapping_pairs_build(vol, (u8*)a +
-					le16_to_cpu(a->data.non_resident.
-					mapping_pairs_offset), attr_rec_len -
-					le16_to_cpu(a->data.non_resident.
-					mapping_pairs_offset), ni->runlist.rl,
-					vcn, highest_vcn, NULL)) {
-				ntfs_error(vol->sb, "Failed to restore "
-						"mapping pairs array in error "
-						"code path.  Run chkdsk to "
-						"recover.");
-				NVolSetErrors(vol);
-			}
-			flush_dcache_mft_record_page(ctx->ntfs_ino);
-			mark_mft_record_dirty(ctx->ntfs_ino);
-		}
-	}
-	/* Release the mft record and the attribute. */
-	if (status.mft_attr_mapped) {
-		ntfs_attr_put_search_ctx(ctx);
-		unmap_mft_record(base_ni);
-	}
-	/* Release the runlist lock. */
-	if (rl_write_locked)
-		up_write(&ni->runlist.lock);
-	else if (rl)
-		up_read(&ni->runlist.lock);
-	/*
-	 * Zero out any newly allocated blocks to avoid exposing stale data.
-	 * If BH_New is set, we know that the block was newly allocated above
-	 * and that it has not been fully zeroed and marked dirty yet.
-	 */
-	nr_pages = u;
-	u = 0;
-	end = bh_cpos << vol->cluster_size_bits;
-	do {
-		folio = page_folio(pages[u]);
-		bh = head = folio_buffers(folio);
-		do {
-			if (u == nr_pages &&
-			    folio_pos(folio) + bh_offset(bh) >= end)
-				break;
-			if (!buffer_new(bh))
-				continue;
-			clear_buffer_new(bh);
-			if (!buffer_uptodate(bh)) {
-				if (folio_test_uptodate(folio))
-					set_buffer_uptodate(bh);
-				else {
-					folio_zero_range(folio, bh_offset(bh),
-							blocksize);
-					set_buffer_uptodate(bh);
-				}
-			}
-			mark_buffer_dirty(bh);
-		} while ((bh = bh->b_this_page) != head);
-	} while (++u <= nr_pages);
-	ntfs_error(vol->sb, "Failed.  Returning error code %i.", err);
-	return err;
-}
-
-static inline void ntfs_flush_dcache_pages(struct page **pages,
-		unsigned nr_pages)
-{
-	BUG_ON(!nr_pages);
-	/*
-	 * Warning: Do not do the decrement at the same time as the call to
-	 * flush_dcache_page() because it is a NULL macro on i386 and hence the
-	 * decrement never happens so the loop never terminates.
-	 */
-	do {
-		--nr_pages;
-		flush_dcache_page(pages[nr_pages]);
-	} while (nr_pages > 0);
-}
-
-/**
- * ntfs_commit_pages_after_non_resident_write - commit the received data
- * @pages:	array of destination pages
- * @nr_pages:	number of pages in @pages
- * @pos:	byte position in file at which the write begins
- * @bytes:	number of bytes to be written
- *
- * See description of ntfs_commit_pages_after_write(), below.
- */
-static inline int ntfs_commit_pages_after_non_resident_write(
-		struct page **pages, const unsigned nr_pages,
-		s64 pos, size_t bytes)
-{
-	s64 end, initialized_size;
-	struct inode *vi;
-	ntfs_inode *ni, *base_ni;
-	struct buffer_head *bh, *head;
-	ntfs_attr_search_ctx *ctx;
-	MFT_RECORD *m;
-	ATTR_RECORD *a;
-	unsigned long flags;
-	unsigned blocksize, u;
-	int err;
-
-	vi = pages[0]->mapping->host;
-	ni = NTFS_I(vi);
-	blocksize = vi->i_sb->s_blocksize;
-	end = pos + bytes;
-	u = 0;
-	do {
-		s64 bh_pos;
-		struct page *page;
-		bool partial;
-
-		page = pages[u];
-		bh_pos = (s64)page->index << PAGE_SHIFT;
-		bh = head = page_buffers(page);
-		partial = false;
-		do {
-			s64 bh_end;
-
-			bh_end = bh_pos + blocksize;
-			if (bh_end <= pos || bh_pos >= end) {
-				if (!buffer_uptodate(bh))
-					partial = true;
-			} else {
-				set_buffer_uptodate(bh);
-				mark_buffer_dirty(bh);
-			}
-		} while (bh_pos += blocksize, (bh = bh->b_this_page) != head);
-		/*
-		 * If all buffers are now uptodate but the page is not, set the
-		 * page uptodate.
-		 */
-		if (!partial && !PageUptodate(page))
-			SetPageUptodate(page);
-	} while (++u < nr_pages);
-	/*
-	 * Finally, if we do not need to update initialized_size or i_size we
-	 * are finished.
-	 */
-	read_lock_irqsave(&ni->size_lock, flags);
-	initialized_size = ni->initialized_size;
-	read_unlock_irqrestore(&ni->size_lock, flags);
-	if (end <= initialized_size) {
-		ntfs_debug("Done.");
-		return 0;
-	}
-	/*
-	 * Update initialized_size/i_size as appropriate, both in the inode and
-	 * the mft record.
-	 */
-	if (!NInoAttr(ni))
-		base_ni = ni;
-	else
-		base_ni = ni->ext.base_ntfs_ino;
-	/* Map, pin, and lock the mft record. */
-	m = map_mft_record(base_ni);
-	if (IS_ERR(m)) {
-		err = PTR_ERR(m);
-		m = NULL;
-		ctx = NULL;
-		goto err_out;
-	}
-	BUG_ON(!NInoNonResident(ni));
-	ctx = ntfs_attr_get_search_ctx(base_ni, m);
-	if (unlikely(!ctx)) {
-		err = -ENOMEM;
-		goto err_out;
-	}
-	err = ntfs_attr_lookup(ni->type, ni->name, ni->name_len,
-			CASE_SENSITIVE, 0, NULL, 0, ctx);
-	if (unlikely(err)) {
-		if (err == -ENOENT)
-			err = -EIO;
-		goto err_out;
-	}
-	a = ctx->attr;
-	BUG_ON(!a->non_resident);
-	write_lock_irqsave(&ni->size_lock, flags);
-	BUG_ON(end > ni->allocated_size);
-	ni->initialized_size = end;
-	a->data.non_resident.initialized_size = cpu_to_sle64(end);
-	if (end > i_size_read(vi)) {
-		i_size_write(vi, end);
-		a->data.non_resident.data_size =
-				a->data.non_resident.initialized_size;
-	}
-	write_unlock_irqrestore(&ni->size_lock, flags);
-	/* Mark the mft record dirty, so it gets written back. */
-	flush_dcache_mft_record_page(ctx->ntfs_ino);
-	mark_mft_record_dirty(ctx->ntfs_ino);
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(base_ni);
-	ntfs_debug("Done.");
-	return 0;
-err_out:
-	if (ctx)
-		ntfs_attr_put_search_ctx(ctx);
-	if (m)
-		unmap_mft_record(base_ni);
-	ntfs_error(vi->i_sb, "Failed to update initialized_size/i_size (error "
-			"code %i).", err);
-	if (err != -ENOMEM)
-		NVolSetErrors(ni->vol);
-	return err;
-}
-
-/**
- * ntfs_commit_pages_after_write - commit the received data
- * @pages:	array of destination pages
- * @nr_pages:	number of pages in @pages
- * @pos:	byte position in file at which the write begins
- * @bytes:	number of bytes to be written
- *
- * This is called from ntfs_file_buffered_write() with i_mutex held on the inode
- * (@pages[0]->mapping->host).  There are @nr_pages pages in @pages which are
- * locked but not kmap()ped.  The source data has already been copied into the
- * @page.  ntfs_prepare_pages_for_non_resident_write() has been called before
- * the data was copied (for non-resident attributes only) and it returned
- * success.
- *
- * Need to set uptodate and mark dirty all buffers within the boundary of the
- * write.  If all buffers in a page are uptodate we set the page uptodate, too.
- *
- * Setting the buffers dirty ensures that they get written out later when
- * ntfs_writepage() is invoked by the VM.
- *
- * Finally, we need to update i_size and initialized_size as appropriate both
- * in the inode and the mft record.
- *
- * This is modelled after fs/buffer.c::generic_commit_write(), which marks
- * buffers uptodate and dirty, sets the page uptodate if all buffers in the
- * page are uptodate, and updates i_size if the end of io is beyond i_size.  In
- * that case, it also marks the inode dirty.
- *
- * If things have gone as outlined in
- * ntfs_prepare_pages_for_non_resident_write(), we do not need to do any page
- * content modifications here for non-resident attributes.  For resident
- * attributes we need to do the uptodate bringing here which we combine with
- * the copying into the mft record which means we save one atomic kmap.
- *
- * Return 0 on success or -errno on error.
- */
-static int ntfs_commit_pages_after_write(struct page **pages,
-		const unsigned nr_pages, s64 pos, size_t bytes)
-{
-	s64 end, initialized_size;
-	loff_t i_size;
-	struct inode *vi;
-	ntfs_inode *ni, *base_ni;
-	struct page *page;
-	ntfs_attr_search_ctx *ctx;
-	MFT_RECORD *m;
-	ATTR_RECORD *a;
-	char *kattr, *kaddr;
-	unsigned long flags;
-	u32 attr_len;
-	int err;
-
-	BUG_ON(!nr_pages);
-	BUG_ON(!pages);
-	page = pages[0];
-	BUG_ON(!page);
-	vi = page->mapping->host;
-	ni = NTFS_I(vi);
-	ntfs_debug("Entering for inode 0x%lx, attribute type 0x%x, start page "
-			"index 0x%lx, nr_pages 0x%x, pos 0x%llx, bytes 0x%zx.",
-			vi->i_ino, ni->type, page->index, nr_pages,
-			(long long)pos, bytes);
-	if (NInoNonResident(ni))
-		return ntfs_commit_pages_after_non_resident_write(pages,
-				nr_pages, pos, bytes);
-	BUG_ON(nr_pages > 1);
-	/*
-	 * Attribute is resident, implying it is not compressed, encrypted, or
-	 * sparse.
-	 */
-	if (!NInoAttr(ni))
-		base_ni = ni;
-	else
-		base_ni = ni->ext.base_ntfs_ino;
-	BUG_ON(NInoNonResident(ni));
-	/* Map, pin, and lock the mft record. */
-	m = map_mft_record(base_ni);
-	if (IS_ERR(m)) {
-		err = PTR_ERR(m);
-		m = NULL;
-		ctx = NULL;
-		goto err_out;
-	}
-	ctx = ntfs_attr_get_search_ctx(base_ni, m);
-	if (unlikely(!ctx)) {
-		err = -ENOMEM;
-		goto err_out;
-	}
-	err = ntfs_attr_lookup(ni->type, ni->name, ni->name_len,
-			CASE_SENSITIVE, 0, NULL, 0, ctx);
-	if (unlikely(err)) {
-		if (err == -ENOENT)
-			err = -EIO;
-		goto err_out;
-	}
-	a = ctx->attr;
-	BUG_ON(a->non_resident);
-	/* The total length of the attribute value. */
-	attr_len = le32_to_cpu(a->data.resident.value_length);
-	i_size = i_size_read(vi);
-	BUG_ON(attr_len != i_size);
-	BUG_ON(pos > attr_len);
-	end = pos + bytes;
-	BUG_ON(end > le32_to_cpu(a->length) -
-			le16_to_cpu(a->data.resident.value_offset));
-	kattr = (u8*)a + le16_to_cpu(a->data.resident.value_offset);
-	kaddr = kmap_atomic(page);
-	/* Copy the received data from the page to the mft record. */
-	memcpy(kattr + pos, kaddr + pos, bytes);
-	/* Update the attribute length if necessary. */
-	if (end > attr_len) {
-		attr_len = end;
-		a->data.resident.value_length = cpu_to_le32(attr_len);
-	}
-	/*
-	 * If the page is not uptodate, bring the out of bounds area(s)
-	 * uptodate by copying data from the mft record to the page.
-	 */
-	if (!PageUptodate(page)) {
-		if (pos > 0)
-			memcpy(kaddr, kattr, pos);
-		if (end < attr_len)
-			memcpy(kaddr + end, kattr + end, attr_len - end);
-		/* Zero the region outside the end of the attribute value. */
-		memset(kaddr + attr_len, 0, PAGE_SIZE - attr_len);
-		flush_dcache_page(page);
-		SetPageUptodate(page);
-	}
-	kunmap_atomic(kaddr);
-	/* Update initialized_size/i_size if necessary. */
-	read_lock_irqsave(&ni->size_lock, flags);
-	initialized_size = ni->initialized_size;
-	BUG_ON(end > ni->allocated_size);
-	read_unlock_irqrestore(&ni->size_lock, flags);
-	BUG_ON(initialized_size != i_size);
-	if (end > initialized_size) {
-		write_lock_irqsave(&ni->size_lock, flags);
-		ni->initialized_size = end;
-		i_size_write(vi, end);
-		write_unlock_irqrestore(&ni->size_lock, flags);
-	}
-	/* Mark the mft record dirty, so it gets written back. */
-	flush_dcache_mft_record_page(ctx->ntfs_ino);
-	mark_mft_record_dirty(ctx->ntfs_ino);
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(base_ni);
-	ntfs_debug("Done.");
-	return 0;
-err_out:
-	if (err == -ENOMEM) {
-		ntfs_warning(vi->i_sb, "Error allocating memory required to "
-				"commit the write.");
-		if (PageUptodate(page)) {
-			ntfs_warning(vi->i_sb, "Page is uptodate, setting "
-					"dirty so the write will be retried "
-					"later on by the VM.");
-			/*
-			 * Put the page on mapping->dirty_pages, but leave its
-			 * buffers' dirty state as-is.
-			 */
-			__set_page_dirty_nobuffers(page);
-			err = 0;
-		} else
-			ntfs_error(vi->i_sb, "Page is not uptodate.  Written "
-					"data has been lost.");
-	} else {
-		ntfs_error(vi->i_sb, "Resident attribute commit write failed "
-				"with error %i.", err);
-		NVolSetErrors(ni->vol);
-	}
-	if (ctx)
-		ntfs_attr_put_search_ctx(ctx);
-	if (m)
-		unmap_mft_record(base_ni);
-	return err;
-}
-
-/*
- * Copy as much as we can into the pages and return the number of bytes which
- * were successfully copied.  If a fault is encountered then clear the pages
- * out to (ofs + bytes) and return the number of bytes which were copied.
- */
-static size_t ntfs_copy_from_user_iter(struct page **pages, unsigned nr_pages,
-		unsigned ofs, struct iov_iter *i, size_t bytes)
-{
-	struct page **last_page = pages + nr_pages;
-	size_t total = 0;
-	unsigned len, copied;
-
-	do {
-		len = PAGE_SIZE - ofs;
-		if (len > bytes)
-			len = bytes;
-		copied = copy_page_from_iter_atomic(*pages, ofs, len, i);
-		total += copied;
-		bytes -= copied;
-		if (!bytes)
-			break;
-		if (copied < len)
-			goto err;
-		ofs = 0;
-	} while (++pages < last_page);
-out:
-	return total;
-err:
-	/* Zero the rest of the target like __copy_from_user(). */
-	len = PAGE_SIZE - copied;
-	do {
-		if (len > bytes)
-			len = bytes;
-		zero_user(*pages, copied, len);
-		bytes -= len;
-		copied = 0;
-		len = PAGE_SIZE;
-	} while (++pages < last_page);
-	goto out;
-}
-
-/**
- * ntfs_perform_write - perform buffered write to a file
- * @file:	file to write to
- * @i:		iov_iter with data to write
- * @pos:	byte offset in file at which to begin writing to
- */
-static ssize_t ntfs_perform_write(struct file *file, struct iov_iter *i,
-		loff_t pos)
-{
-	struct address_space *mapping = file->f_mapping;
-	struct inode *vi = mapping->host;
-	ntfs_inode *ni = NTFS_I(vi);
-	ntfs_volume *vol = ni->vol;
-	struct page *pages[NTFS_MAX_PAGES_PER_CLUSTER];
-	struct page *cached_page = NULL;
-	VCN last_vcn;
-	LCN lcn;
-	size_t bytes;
-	ssize_t status, written = 0;
-	unsigned nr_pages;
-
-	ntfs_debug("Entering for i_ino 0x%lx, attribute type 0x%x, pos "
-			"0x%llx, count 0x%lx.", vi->i_ino,
-			(unsigned)le32_to_cpu(ni->type),
-			(unsigned long long)pos,
-			(unsigned long)iov_iter_count(i));
-	/*
-	 * If a previous ntfs_truncate() failed, repeat it and abort if it
-	 * fails again.
-	 */
-	if (unlikely(NInoTruncateFailed(ni))) {
-		int err;
-
-		inode_dio_wait(vi);
-		err = ntfs_truncate(vi);
-		if (err || NInoTruncateFailed(ni)) {
-			if (!err)
-				err = -EIO;
-			ntfs_error(vol->sb, "Cannot perform write to inode "
-					"0x%lx, attribute type 0x%x, because "
-					"ntfs_truncate() failed (error code "
-					"%i).", vi->i_ino,
-					(unsigned)le32_to_cpu(ni->type), err);
-			return err;
-		}
-	}
-	/*
-	 * Determine the number of pages per cluster for non-resident
-	 * attributes.
-	 */
-	nr_pages = 1;
-	if (vol->cluster_size > PAGE_SIZE && NInoNonResident(ni))
-		nr_pages = vol->cluster_size >> PAGE_SHIFT;
-	last_vcn = -1;
-	do {
-		VCN vcn;
-		pgoff_t start_idx;
-		unsigned ofs, do_pages, u;
-		size_t copied;
-
-		start_idx = pos >> PAGE_SHIFT;
-		ofs = pos & ~PAGE_MASK;
-		bytes = PAGE_SIZE - ofs;
-		do_pages = 1;
-		if (nr_pages > 1) {
-			vcn = pos >> vol->cluster_size_bits;
-			if (vcn != last_vcn) {
-				last_vcn = vcn;
-				/*
-				 * Get the lcn of the vcn the write is in.  If
-				 * it is a hole, need to lock down all pages in
-				 * the cluster.
-				 */
-				down_read(&ni->runlist.lock);
-				lcn = ntfs_attr_vcn_to_lcn_nolock(ni, pos >>
-						vol->cluster_size_bits, false);
-				up_read(&ni->runlist.lock);
-				if (unlikely(lcn < LCN_HOLE)) {
-					if (lcn == LCN_ENOMEM)
-						status = -ENOMEM;
-					else {
-						status = -EIO;
-						ntfs_error(vol->sb, "Cannot "
-							"perform write to "
-							"inode 0x%lx, "
-							"attribute type 0x%x, "
-							"because the attribute "
-							"is corrupt.",
-							vi->i_ino, (unsigned)
-							le32_to_cpu(ni->type));
-					}
-					break;
-				}
-				if (lcn == LCN_HOLE) {
-					start_idx = (pos & ~(s64)
-							vol->cluster_size_mask)
-							>> PAGE_SHIFT;
-					bytes = vol->cluster_size - (pos &
-							vol->cluster_size_mask);
-					do_pages = nr_pages;
-				}
-			}
-		}
-		if (bytes > iov_iter_count(i))
-			bytes = iov_iter_count(i);
-again:
-		/*
-		 * Bring in the user page(s) that we will copy from _first_.
-		 * Otherwise there is a nasty deadlock on copying from the same
-		 * page(s) as we are writing to, without it/them being marked
-		 * up-to-date.  Note, at present there is nothing to stop the
-		 * pages being swapped out between us bringing them into memory
-		 * and doing the actual copying.
-		 */
-		if (unlikely(fault_in_iov_iter_readable(i, bytes))) {
-			status = -EFAULT;
-			break;
-		}
-		/* Get and lock @do_pages starting at index @start_idx. */
-		status = __ntfs_grab_cache_pages(mapping, start_idx, do_pages,
-				pages, &cached_page);
-		if (unlikely(status))
-			break;
-		/*
-		 * For non-resident attributes, we need to fill any holes with
-		 * actual clusters and ensure all bufferes are mapped.  We also
-		 * need to bring uptodate any buffers that are only partially
-		 * being written to.
-		 */
-		if (NInoNonResident(ni)) {
-			status = ntfs_prepare_pages_for_non_resident_write(
-					pages, do_pages, pos, bytes);
-			if (unlikely(status)) {
-				do {
-					unlock_page(pages[--do_pages]);
-					put_page(pages[do_pages]);
-				} while (do_pages);
-				break;
-			}
-		}
-		u = (pos >> PAGE_SHIFT) - pages[0]->index;
-		copied = ntfs_copy_from_user_iter(pages + u, do_pages - u, ofs,
-					i, bytes);
-		ntfs_flush_dcache_pages(pages + u, do_pages - u);
-		status = 0;
-		if (likely(copied == bytes)) {
-			status = ntfs_commit_pages_after_write(pages, do_pages,
-					pos, bytes);
-		}
-		do {
-			unlock_page(pages[--do_pages]);
-			put_page(pages[do_pages]);
-		} while (do_pages);
-		if (unlikely(status < 0)) {
-			iov_iter_revert(i, copied);
-			break;
-		}
-		cond_resched();
-		if (unlikely(copied < bytes)) {
-			iov_iter_revert(i, copied);
-			if (copied)
-				bytes = copied;
-			else if (bytes > PAGE_SIZE - ofs)
-				bytes = PAGE_SIZE - ofs;
-			goto again;
-		}
-		pos += copied;
-		written += copied;
-		balance_dirty_pages_ratelimited(mapping);
-		if (fatal_signal_pending(current)) {
-			status = -EINTR;
-			break;
-		}
-	} while (iov_iter_count(i));
-	if (cached_page)
-		put_page(cached_page);
-	ntfs_debug("Done.  Returning %s (written 0x%lx, status %li).",
-			written ? "written" : "status", (unsigned long)written,
-			(long)status);
-	return written ? written : status;
-}
-
-/**
- * ntfs_file_write_iter - simple wrapper for ntfs_file_write_iter_nolock()
- * @iocb:	IO state structure
- * @from:	iov_iter with data to write
- *
- * Basically the same as generic_file_write_iter() except that it ends up
- * up calling ntfs_perform_write() instead of generic_perform_write() and that
- * O_DIRECT is not implemented.
- */
-static ssize_t ntfs_file_write_iter(struct kiocb *iocb, struct iov_iter *from)
-{
-	struct file *file = iocb->ki_filp;
-	struct inode *vi = file_inode(file);
-	ssize_t written = 0;
-	ssize_t err;
-
-	inode_lock(vi);
-	/* We can write back this queue in page reclaim. */
-	err = ntfs_prepare_file_for_write(iocb, from);
-	if (iov_iter_count(from) && !err)
-		written = ntfs_perform_write(file, from, iocb->ki_pos);
-	inode_unlock(vi);
-	iocb->ki_pos += written;
-	if (likely(written > 0))
-		written = generic_write_sync(iocb, written);
-	return written ? written : err;
-}
-
-/**
- * ntfs_file_fsync - sync a file to disk
- * @filp:	file to be synced
- * @datasync:	if non-zero only flush user data and not metadata
- *
- * Data integrity sync of a file to disk.  Used for fsync, fdatasync, and msync
- * system calls.  This function is inspired by fs/buffer.c::file_fsync().
- *
- * If @datasync is false, write the mft record and all associated extent mft
- * records as well as the $DATA attribute and then sync the block device.
- *
- * If @datasync is true and the attribute is non-resident, we skip the writing
- * of the mft record and all associated extent mft records (this might still
- * happen due to the write_inode_now() call).
- *
- * Also, if @datasync is true, we do not wait on the inode to be written out
- * but we always wait on the page cache pages to be written out.
- *
- * Locking: Caller must hold i_mutex on the inode.
- *
- * TODO: We should probably also write all attribute/index inodes associated
- * with this inode but since we have no simple way of getting to them we ignore
- * this problem for now.
- */
-static int ntfs_file_fsync(struct file *filp, loff_t start, loff_t end,
-			   int datasync)
-{
-	struct inode *vi = filp->f_mapping->host;
-	int err, ret = 0;
-
-	ntfs_debug("Entering for inode 0x%lx.", vi->i_ino);
-
-	err = file_write_and_wait_range(filp, start, end);
-	if (err)
-		return err;
-	inode_lock(vi);
-
-	BUG_ON(S_ISDIR(vi->i_mode));
-	if (!datasync || !NInoNonResident(NTFS_I(vi)))
-		ret = __ntfs_write_inode(vi, 1);
-	write_inode_now(vi, !datasync);
-	/*
-	 * NOTE: If we were to use mapping->private_list (see ext2 and
-	 * fs/buffer.c) for dirty blocks then we could optimize the below to be
-	 * sync_mapping_buffers(vi->i_mapping).
-	 */
-	err = sync_blockdev(vi->i_sb->s_bdev);
-	if (unlikely(err && !ret))
-		ret = err;
-	if (likely(!ret))
-		ntfs_debug("Done.");
-	else
-		ntfs_warning(vi->i_sb, "Failed to f%ssync inode 0x%lx.  Error "
-				"%u.", datasync ? "data" : "", vi->i_ino, -ret);
-	inode_unlock(vi);
-	return ret;
-}
-
-#endif /* NTFS_RW */
-
-const struct file_operations ntfs_file_ops = {
-	.llseek		= generic_file_llseek,
-	.read_iter	= generic_file_read_iter,
-#ifdef NTFS_RW
-	.write_iter	= ntfs_file_write_iter,
-	.fsync		= ntfs_file_fsync,
-#endif /* NTFS_RW */
-	.mmap		= generic_file_mmap,
-	.open		= ntfs_file_open,
-	.splice_read	= filemap_splice_read,
-};
-
-const struct inode_operations ntfs_file_inode_ops = {
-#ifdef NTFS_RW
-	.setattr	= ntfs_setattr,
-#endif /* NTFS_RW */
-};
-
-const struct file_operations ntfs_empty_file_ops = {};
-
-const struct inode_operations ntfs_empty_inode_ops = {};
diff --git a/fs/ntfs/index.c b/fs/ntfs/index.c
deleted file mode 100644
index d46c2c03a0..0000000000
--- a/fs/ntfs/index.c
+++ /dev/null
@@ -1,440 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * index.c - NTFS kernel index handling.  Part of the Linux-NTFS project.
- *
- * Copyright (c) 2004-2005 Anton Altaparmakov
- */
-
-#include <linux/slab.h>
-
-#include "aops.h"
-#include "collate.h"
-#include "debug.h"
-#include "index.h"
-#include "ntfs.h"
-
-/**
- * ntfs_index_ctx_get - allocate and initialize a new index context
- * @idx_ni:	ntfs index inode with which to initialize the context
- *
- * Allocate a new index context, initialize it with @idx_ni and return it.
- * Return NULL if allocation failed.
- *
- * Locking:  Caller must hold i_mutex on the index inode.
- */
-ntfs_index_context *ntfs_index_ctx_get(ntfs_inode *idx_ni)
-{
-	ntfs_index_context *ictx;
-
-	ictx = kmem_cache_alloc(ntfs_index_ctx_cache, GFP_NOFS);
-	if (ictx)
-		*ictx = (ntfs_index_context){ .idx_ni = idx_ni };
-	return ictx;
-}
-
-/**
- * ntfs_index_ctx_put - release an index context
- * @ictx:	index context to free
- *
- * Release the index context @ictx, releasing all associated resources.
- *
- * Locking:  Caller must hold i_mutex on the index inode.
- */
-void ntfs_index_ctx_put(ntfs_index_context *ictx)
-{
-	if (ictx->entry) {
-		if (ictx->is_in_root) {
-			if (ictx->actx)
-				ntfs_attr_put_search_ctx(ictx->actx);
-			if (ictx->base_ni)
-				unmap_mft_record(ictx->base_ni);
-		} else {
-			struct page *page = ictx->page;
-			if (page) {
-				BUG_ON(!PageLocked(page));
-				unlock_page(page);
-				ntfs_unmap_page(page);
-			}
-		}
-	}
-	kmem_cache_free(ntfs_index_ctx_cache, ictx);
-	return;
-}
-
-/**
- * ntfs_index_lookup - find a key in an index and return its index entry
- * @key:	[IN] key for which to search in the index
- * @key_len:	[IN] length of @key in bytes
- * @ictx:	[IN/OUT] context describing the index and the returned entry
- *
- * Before calling ntfs_index_lookup(), @ictx must have been obtained from a
- * call to ntfs_index_ctx_get().
- *
- * Look for the @key in the index specified by the index lookup context @ictx.
- * ntfs_index_lookup() walks the contents of the index looking for the @key.
- *
- * If the @key is found in the index, 0 is returned and @ictx is setup to
- * describe the index entry containing the matching @key.  @ictx->entry is the
- * index entry and @ictx->data and @ictx->data_len are the index entry data and
- * its length in bytes, respectively.
- *
- * If the @key is not found in the index, -ENOENT is returned and @ictx is
- * setup to describe the index entry whose key collates immediately after the
- * search @key, i.e. this is the position in the index at which an index entry
- * with a key of @key would need to be inserted.
- *
- * If an error occurs return the negative error code and @ictx is left
- * untouched.
- *
- * When finished with the entry and its data, call ntfs_index_ctx_put() to free
- * the context and other associated resources.
- *
- * If the index entry was modified, call flush_dcache_index_entry_page()
- * immediately after the modification and either ntfs_index_entry_mark_dirty()
- * or ntfs_index_entry_write() before the call to ntfs_index_ctx_put() to
- * ensure that the changes are written to disk.
- *
- * Locking:  - Caller must hold i_mutex on the index inode.
- *	     - Each page cache page in the index allocation mapping must be
- *	       locked whilst being accessed otherwise we may find a corrupt
- *	       page due to it being under ->writepage at the moment which
- *	       applies the mst protection fixups before writing out and then
- *	       removes them again after the write is complete after which it 
- *	       unlocks the page.
- */
-int ntfs_index_lookup(const void *key, const int key_len,
-		ntfs_index_context *ictx)
-{
-	VCN vcn, old_vcn;
-	ntfs_inode *idx_ni = ictx->idx_ni;
-	ntfs_volume *vol = idx_ni->vol;
-	struct super_block *sb = vol->sb;
-	ntfs_inode *base_ni = idx_ni->ext.base_ntfs_ino;
-	MFT_RECORD *m;
-	INDEX_ROOT *ir;
-	INDEX_ENTRY *ie;
-	INDEX_ALLOCATION *ia;
-	u8 *index_end, *kaddr;
-	ntfs_attr_search_ctx *actx;
-	struct address_space *ia_mapping;
-	struct page *page;
-	int rc, err = 0;
-
-	ntfs_debug("Entering.");
-	BUG_ON(!NInoAttr(idx_ni));
-	BUG_ON(idx_ni->type != AT_INDEX_ALLOCATION);
-	BUG_ON(idx_ni->nr_extents != -1);
-	BUG_ON(!base_ni);
-	BUG_ON(!key);
-	BUG_ON(key_len <= 0);
-	if (!ntfs_is_collation_rule_supported(
-			idx_ni->itype.index.collation_rule)) {
-		ntfs_error(sb, "Index uses unsupported collation rule 0x%x.  "
-				"Aborting lookup.", le32_to_cpu(
-				idx_ni->itype.index.collation_rule));
-		return -EOPNOTSUPP;
-	}
-	/* Get hold of the mft record for the index inode. */
-	m = map_mft_record(base_ni);
-	if (IS_ERR(m)) {
-		ntfs_error(sb, "map_mft_record() failed with error code %ld.",
-				-PTR_ERR(m));
-		return PTR_ERR(m);
-	}
-	actx = ntfs_attr_get_search_ctx(base_ni, m);
-	if (unlikely(!actx)) {
-		err = -ENOMEM;
-		goto err_out;
-	}
-	/* Find the index root attribute in the mft record. */
-	err = ntfs_attr_lookup(AT_INDEX_ROOT, idx_ni->name, idx_ni->name_len,
-			CASE_SENSITIVE, 0, NULL, 0, actx);
-	if (unlikely(err)) {
-		if (err == -ENOENT) {
-			ntfs_error(sb, "Index root attribute missing in inode "
-					"0x%lx.", idx_ni->mft_no);
-			err = -EIO;
-		}
-		goto err_out;
-	}
-	/* Get to the index root value (it has been verified in read_inode). */
-	ir = (INDEX_ROOT*)((u8*)actx->attr +
-			le16_to_cpu(actx->attr->data.resident.value_offset));
-	index_end = (u8*)&ir->index + le32_to_cpu(ir->index.index_length);
-	/* The first index entry. */
-	ie = (INDEX_ENTRY*)((u8*)&ir->index +
-			le32_to_cpu(ir->index.entries_offset));
-	/*
-	 * Loop until we exceed valid memory (corruption case) or until we
-	 * reach the last entry.
-	 */
-	for (;; ie = (INDEX_ENTRY*)((u8*)ie + le16_to_cpu(ie->length))) {
-		/* Bounds checks. */
-		if ((u8*)ie < (u8*)actx->mrec || (u8*)ie +
-				sizeof(INDEX_ENTRY_HEADER) > index_end ||
-				(u8*)ie + le16_to_cpu(ie->length) > index_end)
-			goto idx_err_out;
-		/*
-		 * The last entry cannot contain a key.  It can however contain
-		 * a pointer to a child node in the B+tree so we just break out.
-		 */
-		if (ie->flags & INDEX_ENTRY_END)
-			break;
-		/* Further bounds checks. */
-		if ((u32)sizeof(INDEX_ENTRY_HEADER) +
-				le16_to_cpu(ie->key_length) >
-				le16_to_cpu(ie->data.vi.data_offset) ||
-				(u32)le16_to_cpu(ie->data.vi.data_offset) +
-				le16_to_cpu(ie->data.vi.data_length) >
-				le16_to_cpu(ie->length))
-			goto idx_err_out;
-		/* If the keys match perfectly, we setup @ictx and return 0. */
-		if ((key_len == le16_to_cpu(ie->key_length)) && !memcmp(key,
-				&ie->key, key_len)) {
-ir_done:
-			ictx->is_in_root = true;
-			ictx->ir = ir;
-			ictx->actx = actx;
-			ictx->base_ni = base_ni;
-			ictx->ia = NULL;
-			ictx->page = NULL;
-done:
-			ictx->entry = ie;
-			ictx->data = (u8*)ie +
-					le16_to_cpu(ie->data.vi.data_offset);
-			ictx->data_len = le16_to_cpu(ie->data.vi.data_length);
-			ntfs_debug("Done.");
-			return err;
-		}
-		/*
-		 * Not a perfect match, need to do full blown collation so we
-		 * know which way in the B+tree we have to go.
-		 */
-		rc = ntfs_collate(vol, idx_ni->itype.index.collation_rule, key,
-				key_len, &ie->key, le16_to_cpu(ie->key_length));
-		/*
-		 * If @key collates before the key of the current entry, there
-		 * is definitely no such key in this index but we might need to
-		 * descend into the B+tree so we just break out of the loop.
-		 */
-		if (rc == -1)
-			break;
-		/*
-		 * A match should never happen as the memcmp() call should have
-		 * cought it, but we still treat it correctly.
-		 */
-		if (!rc)
-			goto ir_done;
-		/* The keys are not equal, continue the search. */
-	}
-	/*
-	 * We have finished with this index without success.  Check for the
-	 * presence of a child node and if not present setup @ictx and return
-	 * -ENOENT.
-	 */
-	if (!(ie->flags & INDEX_ENTRY_NODE)) {
-		ntfs_debug("Entry not found.");
-		err = -ENOENT;
-		goto ir_done;
-	} /* Child node present, descend into it. */
-	/* Consistency check: Verify that an index allocation exists. */
-	if (!NInoIndexAllocPresent(idx_ni)) {
-		ntfs_error(sb, "No index allocation attribute but index entry "
-				"requires one.  Inode 0x%lx is corrupt or "
-				"driver bug.", idx_ni->mft_no);
-		goto err_out;
-	}
-	/* Get the starting vcn of the index_block holding the child node. */
-	vcn = sle64_to_cpup((sle64*)((u8*)ie + le16_to_cpu(ie->length) - 8));
-	ia_mapping = VFS_I(idx_ni)->i_mapping;
-	/*
-	 * We are done with the index root and the mft record.  Release them,
-	 * otherwise we deadlock with ntfs_map_page().
-	 */
-	ntfs_attr_put_search_ctx(actx);
-	unmap_mft_record(base_ni);
-	m = NULL;
-	actx = NULL;
-descend_into_child_node:
-	/*
-	 * Convert vcn to index into the index allocation attribute in units
-	 * of PAGE_SIZE and map the page cache page, reading it from
-	 * disk if necessary.
-	 */
-	page = ntfs_map_page(ia_mapping, vcn <<
-			idx_ni->itype.index.vcn_size_bits >> PAGE_SHIFT);
-	if (IS_ERR(page)) {
-		ntfs_error(sb, "Failed to map index page, error %ld.",
-				-PTR_ERR(page));
-		err = PTR_ERR(page);
-		goto err_out;
-	}
-	lock_page(page);
-	kaddr = (u8*)page_address(page);
-fast_descend_into_child_node:
-	/* Get to the index allocation block. */
-	ia = (INDEX_ALLOCATION*)(kaddr + ((vcn <<
-			idx_ni->itype.index.vcn_size_bits) & ~PAGE_MASK));
-	/* Bounds checks. */
-	if ((u8*)ia < kaddr || (u8*)ia > kaddr + PAGE_SIZE) {
-		ntfs_error(sb, "Out of bounds check failed.  Corrupt inode "
-				"0x%lx or driver bug.", idx_ni->mft_no);
-		goto unm_err_out;
-	}
-	/* Catch multi sector transfer fixup errors. */
-	if (unlikely(!ntfs_is_indx_record(ia->magic))) {
-		ntfs_error(sb, "Index record with vcn 0x%llx is corrupt.  "
-				"Corrupt inode 0x%lx.  Run chkdsk.",
-				(long long)vcn, idx_ni->mft_no);
-		goto unm_err_out;
-	}
-	if (sle64_to_cpu(ia->index_block_vcn) != vcn) {
-		ntfs_error(sb, "Actual VCN (0x%llx) of index buffer is "
-				"different from expected VCN (0x%llx).  Inode "
-				"0x%lx is corrupt or driver bug.",
-				(unsigned long long)
-				sle64_to_cpu(ia->index_block_vcn),
-				(unsigned long long)vcn, idx_ni->mft_no);
-		goto unm_err_out;
-	}
-	if (le32_to_cpu(ia->index.allocated_size) + 0x18 !=
-			idx_ni->itype.index.block_size) {
-		ntfs_error(sb, "Index buffer (VCN 0x%llx) of inode 0x%lx has "
-				"a size (%u) differing from the index "
-				"specified size (%u).  Inode is corrupt or "
-				"driver bug.", (unsigned long long)vcn,
-				idx_ni->mft_no,
-				le32_to_cpu(ia->index.allocated_size) + 0x18,
-				idx_ni->itype.index.block_size);
-		goto unm_err_out;
-	}
-	index_end = (u8*)ia + idx_ni->itype.index.block_size;
-	if (index_end > kaddr + PAGE_SIZE) {
-		ntfs_error(sb, "Index buffer (VCN 0x%llx) of inode 0x%lx "
-				"crosses page boundary.  Impossible!  Cannot "
-				"access!  This is probably a bug in the "
-				"driver.", (unsigned long long)vcn,
-				idx_ni->mft_no);
-		goto unm_err_out;
-	}
-	index_end = (u8*)&ia->index + le32_to_cpu(ia->index.index_length);
-	if (index_end > (u8*)ia + idx_ni->itype.index.block_size) {
-		ntfs_error(sb, "Size of index buffer (VCN 0x%llx) of inode "
-				"0x%lx exceeds maximum size.",
-				(unsigned long long)vcn, idx_ni->mft_no);
-		goto unm_err_out;
-	}
-	/* The first index entry. */
-	ie = (INDEX_ENTRY*)((u8*)&ia->index +
-			le32_to_cpu(ia->index.entries_offset));
-	/*
-	 * Iterate similar to above big loop but applied to index buffer, thus
-	 * loop until we exceed valid memory (corruption case) or until we
-	 * reach the last entry.
-	 */
-	for (;; ie = (INDEX_ENTRY*)((u8*)ie + le16_to_cpu(ie->length))) {
-		/* Bounds checks. */
-		if ((u8*)ie < (u8*)ia || (u8*)ie +
-				sizeof(INDEX_ENTRY_HEADER) > index_end ||
-				(u8*)ie + le16_to_cpu(ie->length) > index_end) {
-			ntfs_error(sb, "Index entry out of bounds in inode "
-					"0x%lx.", idx_ni->mft_no);
-			goto unm_err_out;
-		}
-		/*
-		 * The last entry cannot contain a key.  It can however contain
-		 * a pointer to a child node in the B+tree so we just break out.
-		 */
-		if (ie->flags & INDEX_ENTRY_END)
-			break;
-		/* Further bounds checks. */
-		if ((u32)sizeof(INDEX_ENTRY_HEADER) +
-				le16_to_cpu(ie->key_length) >
-				le16_to_cpu(ie->data.vi.data_offset) ||
-				(u32)le16_to_cpu(ie->data.vi.data_offset) +
-				le16_to_cpu(ie->data.vi.data_length) >
-				le16_to_cpu(ie->length)) {
-			ntfs_error(sb, "Index entry out of bounds in inode "
-					"0x%lx.", idx_ni->mft_no);
-			goto unm_err_out;
-		}
-		/* If the keys match perfectly, we setup @ictx and return 0. */
-		if ((key_len == le16_to_cpu(ie->key_length)) && !memcmp(key,
-				&ie->key, key_len)) {
-ia_done:
-			ictx->is_in_root = false;
-			ictx->actx = NULL;
-			ictx->base_ni = NULL;
-			ictx->ia = ia;
-			ictx->page = page;
-			goto done;
-		}
-		/*
-		 * Not a perfect match, need to do full blown collation so we
-		 * know which way in the B+tree we have to go.
-		 */
-		rc = ntfs_collate(vol, idx_ni->itype.index.collation_rule, key,
-				key_len, &ie->key, le16_to_cpu(ie->key_length));
-		/*
-		 * If @key collates before the key of the current entry, there
-		 * is definitely no such key in this index but we might need to
-		 * descend into the B+tree so we just break out of the loop.
-		 */
-		if (rc == -1)
-			break;
-		/*
-		 * A match should never happen as the memcmp() call should have
-		 * cought it, but we still treat it correctly.
-		 */
-		if (!rc)
-			goto ia_done;
-		/* The keys are not equal, continue the search. */
-	}
-	/*
-	 * We have finished with this index buffer without success.  Check for
-	 * the presence of a child node and if not present return -ENOENT.
-	 */
-	if (!(ie->flags & INDEX_ENTRY_NODE)) {
-		ntfs_debug("Entry not found.");
-		err = -ENOENT;
-		goto ia_done;
-	}
-	if ((ia->index.flags & NODE_MASK) == LEAF_NODE) {
-		ntfs_error(sb, "Index entry with child node found in a leaf "
-				"node in inode 0x%lx.", idx_ni->mft_no);
-		goto unm_err_out;
-	}
-	/* Child node present, descend into it. */
-	old_vcn = vcn;
-	vcn = sle64_to_cpup((sle64*)((u8*)ie + le16_to_cpu(ie->length) - 8));
-	if (vcn >= 0) {
-		/*
-		 * If vcn is in the same page cache page as old_vcn we recycle
-		 * the mapped page.
-		 */
-		if (old_vcn << vol->cluster_size_bits >>
-				PAGE_SHIFT == vcn <<
-				vol->cluster_size_bits >>
-				PAGE_SHIFT)
-			goto fast_descend_into_child_node;
-		unlock_page(page);
-		ntfs_unmap_page(page);
-		goto descend_into_child_node;
-	}
-	ntfs_error(sb, "Negative child node vcn in inode 0x%lx.",
-			idx_ni->mft_no);
-unm_err_out:
-	unlock_page(page);
-	ntfs_unmap_page(page);
-err_out:
-	if (!err)
-		err = -EIO;
-	if (actx)
-		ntfs_attr_put_search_ctx(actx);
-	if (m)
-		unmap_mft_record(base_ni);
-	return err;
-idx_err_out:
-	ntfs_error(sb, "Corrupt index.  Aborting lookup.");
-	goto err_out;
-}
diff --git a/fs/ntfs/index.h b/fs/ntfs/index.h
deleted file mode 100644
index bb3c3ae551..0000000000
--- a/fs/ntfs/index.h
+++ /dev/null
@@ -1,134 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * index.h - Defines for NTFS kernel index handling.  Part of the Linux-NTFS
- *	     project.
- *
- * Copyright (c) 2004 Anton Altaparmakov
- */
-
-#ifndef _LINUX_NTFS_INDEX_H
-#define _LINUX_NTFS_INDEX_H
-
-#include <linux/fs.h>
-
-#include "types.h"
-#include "layout.h"
-#include "inode.h"
-#include "attrib.h"
-#include "mft.h"
-#include "aops.h"
-
-/**
- * @idx_ni:	index inode containing the @entry described by this context
- * @entry:	index entry (points into @ir or @ia)
- * @data:	index entry data (points into @entry)
- * @data_len:	length in bytes of @data
- * @is_in_root:	'true' if @entry is in @ir and 'false' if it is in @ia
- * @ir:		index root if @is_in_root and NULL otherwise
- * @actx:	attribute search context if @is_in_root and NULL otherwise
- * @base_ni:	base inode if @is_in_root and NULL otherwise
- * @ia:		index block if @is_in_root is 'false' and NULL otherwise
- * @page:	page if @is_in_root is 'false' and NULL otherwise
- *
- * @idx_ni is the index inode this context belongs to.
- *
- * @entry is the index entry described by this context.  @data and @data_len
- * are the index entry data and its length in bytes, respectively.  @data
- * simply points into @entry.  This is probably what the user is interested in.
- *
- * If @is_in_root is 'true', @entry is in the index root attribute @ir described
- * by the attribute search context @actx and the base inode @base_ni.  @ia and
- * @page are NULL in this case.
- *
- * If @is_in_root is 'false', @entry is in the index allocation attribute and @ia
- * and @page point to the index allocation block and the mapped, locked page it
- * is in, respectively.  @ir, @actx and @base_ni are NULL in this case.
- *
- * To obtain a context call ntfs_index_ctx_get().
- *
- * We use this context to allow ntfs_index_lookup() to return the found index
- * @entry and its @data without having to allocate a buffer and copy the @entry
- * and/or its @data into it.
- *
- * When finished with the @entry and its @data, call ntfs_index_ctx_put() to
- * free the context and other associated resources.
- *
- * If the index entry was modified, call flush_dcache_index_entry_page()
- * immediately after the modification and either ntfs_index_entry_mark_dirty()
- * or ntfs_index_entry_write() before the call to ntfs_index_ctx_put() to
- * ensure that the changes are written to disk.
- */
-typedef struct {
-	ntfs_inode *idx_ni;
-	INDEX_ENTRY *entry;
-	void *data;
-	u16 data_len;
-	bool is_in_root;
-	INDEX_ROOT *ir;
-	ntfs_attr_search_ctx *actx;
-	ntfs_inode *base_ni;
-	INDEX_ALLOCATION *ia;
-	struct page *page;
-} ntfs_index_context;
-
-extern ntfs_index_context *ntfs_index_ctx_get(ntfs_inode *idx_ni);
-extern void ntfs_index_ctx_put(ntfs_index_context *ictx);
-
-extern int ntfs_index_lookup(const void *key, const int key_len,
-		ntfs_index_context *ictx);
-
-#ifdef NTFS_RW
-
-/**
- * ntfs_index_entry_flush_dcache_page - flush_dcache_page() for index entries
- * @ictx:	ntfs index context describing the index entry
- *
- * Call flush_dcache_page() for the page in which an index entry resides.
- *
- * This must be called every time an index entry is modified, just after the
- * modification.
- *
- * If the index entry is in the index root attribute, simply flush the page
- * containing the mft record containing the index root attribute.
- *
- * If the index entry is in an index block belonging to the index allocation
- * attribute, simply flush the page cache page containing the index block.
- */
-static inline void ntfs_index_entry_flush_dcache_page(ntfs_index_context *ictx)
-{
-	if (ictx->is_in_root)
-		flush_dcache_mft_record_page(ictx->actx->ntfs_ino);
-	else
-		flush_dcache_page(ictx->page);
-}
-
-/**
- * ntfs_index_entry_mark_dirty - mark an index entry dirty
- * @ictx:	ntfs index context describing the index entry
- *
- * Mark the index entry described by the index entry context @ictx dirty.
- *
- * If the index entry is in the index root attribute, simply mark the mft
- * record containing the index root attribute dirty.  This ensures the mft
- * record, and hence the index root attribute, will be written out to disk
- * later.
- *
- * If the index entry is in an index block belonging to the index allocation
- * attribute, mark the buffers belonging to the index record as well as the
- * page cache page the index block is in dirty.  This automatically marks the
- * VFS inode of the ntfs index inode to which the index entry belongs dirty,
- * too (I_DIRTY_PAGES) and this in turn ensures the page buffers, and hence the
- * dirty index block, will be written out to disk later.
- */
-static inline void ntfs_index_entry_mark_dirty(ntfs_index_context *ictx)
-{
-	if (ictx->is_in_root)
-		mark_mft_record_dirty(ictx->actx->ntfs_ino);
-	else
-		mark_ntfs_record_dirty(ictx->page,
-				(u8*)ictx->ia - (u8*)page_address(ictx->page));
-}
-
-#endif /* NTFS_RW */
-
-#endif /* _LINUX_NTFS_INDEX_H */
diff --git a/fs/ntfs/inode.c b/fs/ntfs/inode.c
deleted file mode 100644
index aba1e22db4..0000000000
--- a/fs/ntfs/inode.c
+++ /dev/null
@@ -1,3102 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * inode.c - NTFS kernel inode handling.
- *
- * Copyright (c) 2001-2014 Anton Altaparmakov and Tuxera Inc.
- */
-
-#include <linux/buffer_head.h>
-#include <linux/fs.h>
-#include <linux/mm.h>
-#include <linux/mount.h>
-#include <linux/mutex.h>
-#include <linux/pagemap.h>
-#include <linux/quotaops.h>
-#include <linux/slab.h>
-#include <linux/log2.h>
-
-#include "aops.h"
-#include "attrib.h"
-#include "bitmap.h"
-#include "dir.h"
-#include "debug.h"
-#include "inode.h"
-#include "lcnalloc.h"
-#include "malloc.h"
-#include "mft.h"
-#include "time.h"
-#include "ntfs.h"
-
-/**
- * ntfs_test_inode - compare two (possibly fake) inodes for equality
- * @vi:		vfs inode which to test
- * @data:	data which is being tested with
- *
- * Compare the ntfs attribute embedded in the ntfs specific part of the vfs
- * inode @vi for equality with the ntfs attribute @data.
- *
- * If searching for the normal file/directory inode, set @na->type to AT_UNUSED.
- * @na->name and @na->name_len are then ignored.
- *
- * Return 1 if the attributes match and 0 if not.
- *
- * NOTE: This function runs with the inode_hash_lock spin lock held so it is not
- * allowed to sleep.
- */
-int ntfs_test_inode(struct inode *vi, void *data)
-{
-	ntfs_attr *na = (ntfs_attr *)data;
-	ntfs_inode *ni;
-
-	if (vi->i_ino != na->mft_no)
-		return 0;
-	ni = NTFS_I(vi);
-	/* If !NInoAttr(ni), @vi is a normal file or directory inode. */
-	if (likely(!NInoAttr(ni))) {
-		/* If not looking for a normal inode this is a mismatch. */
-		if (unlikely(na->type != AT_UNUSED))
-			return 0;
-	} else {
-		/* A fake inode describing an attribute. */
-		if (ni->type != na->type)
-			return 0;
-		if (ni->name_len != na->name_len)
-			return 0;
-		if (na->name_len && memcmp(ni->name, na->name,
-				na->name_len * sizeof(ntfschar)))
-			return 0;
-	}
-	/* Match! */
-	return 1;
-}
-
-/**
- * ntfs_init_locked_inode - initialize an inode
- * @vi:		vfs inode to initialize
- * @data:	data which to initialize @vi to
- *
- * Initialize the vfs inode @vi with the values from the ntfs attribute @data in
- * order to enable ntfs_test_inode() to do its work.
- *
- * If initializing the normal file/directory inode, set @na->type to AT_UNUSED.
- * In that case, @na->name and @na->name_len should be set to NULL and 0,
- * respectively. Although that is not strictly necessary as
- * ntfs_read_locked_inode() will fill them in later.
- *
- * Return 0 on success and -errno on error.
- *
- * NOTE: This function runs with the inode->i_lock spin lock held so it is not
- * allowed to sleep. (Hence the GFP_ATOMIC allocation.)
- */
-static int ntfs_init_locked_inode(struct inode *vi, void *data)
-{
-	ntfs_attr *na = (ntfs_attr *)data;
-	ntfs_inode *ni = NTFS_I(vi);
-
-	vi->i_ino = na->mft_no;
-
-	ni->type = na->type;
-	if (na->type == AT_INDEX_ALLOCATION)
-		NInoSetMstProtected(ni);
-
-	ni->name = na->name;
-	ni->name_len = na->name_len;
-
-	/* If initializing a normal inode, we are done. */
-	if (likely(na->type == AT_UNUSED)) {
-		BUG_ON(na->name);
-		BUG_ON(na->name_len);
-		return 0;
-	}
-
-	/* It is a fake inode. */
-	NInoSetAttr(ni);
-
-	/*
-	 * We have I30 global constant as an optimization as it is the name
-	 * in >99.9% of named attributes! The other <0.1% incur a GFP_ATOMIC
-	 * allocation but that is ok. And most attributes are unnamed anyway,
-	 * thus the fraction of named attributes with name != I30 is actually
-	 * absolutely tiny.
-	 */
-	if (na->name_len && na->name != I30) {
-		unsigned int i;
-
-		BUG_ON(!na->name);
-		i = na->name_len * sizeof(ntfschar);
-		ni->name = kmalloc(i + sizeof(ntfschar), GFP_ATOMIC);
-		if (!ni->name)
-			return -ENOMEM;
-		memcpy(ni->name, na->name, i);
-		ni->name[na->name_len] = 0;
-	}
-	return 0;
-}
-
-static int ntfs_read_locked_inode(struct inode *vi);
-static int ntfs_read_locked_attr_inode(struct inode *base_vi, struct inode *vi);
-static int ntfs_read_locked_index_inode(struct inode *base_vi,
-		struct inode *vi);
-
-/**
- * ntfs_iget - obtain a struct inode corresponding to a specific normal inode
- * @sb:		super block of mounted volume
- * @mft_no:	mft record number / inode number to obtain
- *
- * Obtain the struct inode corresponding to a specific normal inode (i.e. a
- * file or directory).
- *
- * If the inode is in the cache, it is just returned with an increased
- * reference count. Otherwise, a new struct inode is allocated and initialized,
- * and finally ntfs_read_locked_inode() is called to read in the inode and
- * fill in the remainder of the inode structure.
- *
- * Return the struct inode on success. Check the return value with IS_ERR() and
- * if true, the function failed and the error code is obtained from PTR_ERR().
- */
-struct inode *ntfs_iget(struct super_block *sb, unsigned long mft_no)
-{
-	struct inode *vi;
-	int err;
-	ntfs_attr na;
-
-	na.mft_no = mft_no;
-	na.type = AT_UNUSED;
-	na.name = NULL;
-	na.name_len = 0;
-
-	vi = iget5_locked(sb, mft_no, ntfs_test_inode,
-			ntfs_init_locked_inode, &na);
-	if (unlikely(!vi))
-		return ERR_PTR(-ENOMEM);
-
-	err = 0;
-
-	/* If this is a freshly allocated inode, need to read it now. */
-	if (vi->i_state & I_NEW) {
-		err = ntfs_read_locked_inode(vi);
-		unlock_new_inode(vi);
-	}
-	/*
-	 * There is no point in keeping bad inodes around if the failure was
-	 * due to ENOMEM. We want to be able to retry again later.
-	 */
-	if (unlikely(err == -ENOMEM)) {
-		iput(vi);
-		vi = ERR_PTR(err);
-	}
-	return vi;
-}
-
-/**
- * ntfs_attr_iget - obtain a struct inode corresponding to an attribute
- * @base_vi:	vfs base inode containing the attribute
- * @type:	attribute type
- * @name:	Unicode name of the attribute (NULL if unnamed)
- * @name_len:	length of @name in Unicode characters (0 if unnamed)
- *
- * Obtain the (fake) struct inode corresponding to the attribute specified by
- * @type, @name, and @name_len, which is present in the base mft record
- * specified by the vfs inode @base_vi.
- *
- * If the attribute inode is in the cache, it is just returned with an
- * increased reference count. Otherwise, a new struct inode is allocated and
- * initialized, and finally ntfs_read_locked_attr_inode() is called to read the
- * attribute and fill in the inode structure.
- *
- * Note, for index allocation attributes, you need to use ntfs_index_iget()
- * instead of ntfs_attr_iget() as working with indices is a lot more complex.
- *
- * Return the struct inode of the attribute inode on success. Check the return
- * value with IS_ERR() and if true, the function failed and the error code is
- * obtained from PTR_ERR().
- */
-struct inode *ntfs_attr_iget(struct inode *base_vi, ATTR_TYPE type,
-		ntfschar *name, u32 name_len)
-{
-	struct inode *vi;
-	int err;
-	ntfs_attr na;
-
-	/* Make sure no one calls ntfs_attr_iget() for indices. */
-	BUG_ON(type == AT_INDEX_ALLOCATION);
-
-	na.mft_no = base_vi->i_ino;
-	na.type = type;
-	na.name = name;
-	na.name_len = name_len;
-
-	vi = iget5_locked(base_vi->i_sb, na.mft_no, ntfs_test_inode,
-			ntfs_init_locked_inode, &na);
-	if (unlikely(!vi))
-		return ERR_PTR(-ENOMEM);
-
-	err = 0;
-
-	/* If this is a freshly allocated inode, need to read it now. */
-	if (vi->i_state & I_NEW) {
-		err = ntfs_read_locked_attr_inode(base_vi, vi);
-		unlock_new_inode(vi);
-	}
-	/*
-	 * There is no point in keeping bad attribute inodes around. This also
-	 * simplifies things in that we never need to check for bad attribute
-	 * inodes elsewhere.
-	 */
-	if (unlikely(err)) {
-		iput(vi);
-		vi = ERR_PTR(err);
-	}
-	return vi;
-}
-
-/**
- * ntfs_index_iget - obtain a struct inode corresponding to an index
- * @base_vi:	vfs base inode containing the index related attributes
- * @name:	Unicode name of the index
- * @name_len:	length of @name in Unicode characters
- *
- * Obtain the (fake) struct inode corresponding to the index specified by @name
- * and @name_len, which is present in the base mft record specified by the vfs
- * inode @base_vi.
- *
- * If the index inode is in the cache, it is just returned with an increased
- * reference count.  Otherwise, a new struct inode is allocated and
- * initialized, and finally ntfs_read_locked_index_inode() is called to read
- * the index related attributes and fill in the inode structure.
- *
- * Return the struct inode of the index inode on success. Check the return
- * value with IS_ERR() and if true, the function failed and the error code is
- * obtained from PTR_ERR().
- */
-struct inode *ntfs_index_iget(struct inode *base_vi, ntfschar *name,
-		u32 name_len)
-{
-	struct inode *vi;
-	int err;
-	ntfs_attr na;
-
-	na.mft_no = base_vi->i_ino;
-	na.type = AT_INDEX_ALLOCATION;
-	na.name = name;
-	na.name_len = name_len;
-
-	vi = iget5_locked(base_vi->i_sb, na.mft_no, ntfs_test_inode,
-			ntfs_init_locked_inode, &na);
-	if (unlikely(!vi))
-		return ERR_PTR(-ENOMEM);
-
-	err = 0;
-
-	/* If this is a freshly allocated inode, need to read it now. */
-	if (vi->i_state & I_NEW) {
-		err = ntfs_read_locked_index_inode(base_vi, vi);
-		unlock_new_inode(vi);
-	}
-	/*
-	 * There is no point in keeping bad index inodes around.  This also
-	 * simplifies things in that we never need to check for bad index
-	 * inodes elsewhere.
-	 */
-	if (unlikely(err)) {
-		iput(vi);
-		vi = ERR_PTR(err);
-	}
-	return vi;
-}
-
-struct inode *ntfs_alloc_big_inode(struct super_block *sb)
-{
-	ntfs_inode *ni;
-
-	ntfs_debug("Entering.");
-	ni = alloc_inode_sb(sb, ntfs_big_inode_cache, GFP_NOFS);
-	if (likely(ni != NULL)) {
-		ni->state = 0;
-		return VFS_I(ni);
-	}
-	ntfs_error(sb, "Allocation of NTFS big inode structure failed.");
-	return NULL;
-}
-
-void ntfs_free_big_inode(struct inode *inode)
-{
-	kmem_cache_free(ntfs_big_inode_cache, NTFS_I(inode));
-}
-
-static inline ntfs_inode *ntfs_alloc_extent_inode(void)
-{
-	ntfs_inode *ni;
-
-	ntfs_debug("Entering.");
-	ni = kmem_cache_alloc(ntfs_inode_cache, GFP_NOFS);
-	if (likely(ni != NULL)) {
-		ni->state = 0;
-		return ni;
-	}
-	ntfs_error(NULL, "Allocation of NTFS inode structure failed.");
-	return NULL;
-}
-
-static void ntfs_destroy_extent_inode(ntfs_inode *ni)
-{
-	ntfs_debug("Entering.");
-	BUG_ON(ni->page);
-	if (!atomic_dec_and_test(&ni->count))
-		BUG();
-	kmem_cache_free(ntfs_inode_cache, ni);
-}
-
-/*
- * The attribute runlist lock has separate locking rules from the
- * normal runlist lock, so split the two lock-classes:
- */
-static struct lock_class_key attr_list_rl_lock_class;
-
-/**
- * __ntfs_init_inode - initialize ntfs specific part of an inode
- * @sb:		super block of mounted volume
- * @ni:		freshly allocated ntfs inode which to initialize
- *
- * Initialize an ntfs inode to defaults.
- *
- * NOTE: ni->mft_no, ni->state, ni->type, ni->name, and ni->name_len are left
- * untouched. Make sure to initialize them elsewhere.
- *
- * Return zero on success and -ENOMEM on error.
- */
-void __ntfs_init_inode(struct super_block *sb, ntfs_inode *ni)
-{
-	ntfs_debug("Entering.");
-	rwlock_init(&ni->size_lock);
-	ni->initialized_size = ni->allocated_size = 0;
-	ni->seq_no = 0;
-	atomic_set(&ni->count, 1);
-	ni->vol = NTFS_SB(sb);
-	ntfs_init_runlist(&ni->runlist);
-	mutex_init(&ni->mrec_lock);
-	ni->page = NULL;
-	ni->page_ofs = 0;
-	ni->attr_list_size = 0;
-	ni->attr_list = NULL;
-	ntfs_init_runlist(&ni->attr_list_rl);
-	lockdep_set_class(&ni->attr_list_rl.lock,
-				&attr_list_rl_lock_class);
-	ni->itype.index.block_size = 0;
-	ni->itype.index.vcn_size = 0;
-	ni->itype.index.collation_rule = 0;
-	ni->itype.index.block_size_bits = 0;
-	ni->itype.index.vcn_size_bits = 0;
-	mutex_init(&ni->extent_lock);
-	ni->nr_extents = 0;
-	ni->ext.base_ntfs_ino = NULL;
-}
-
-/*
- * Extent inodes get MFT-mapped in a nested way, while the base inode
- * is still mapped. Teach this nesting to the lock validator by creating
- * a separate class for nested inode's mrec_lock's:
- */
-static struct lock_class_key extent_inode_mrec_lock_key;
-
-inline ntfs_inode *ntfs_new_extent_inode(struct super_block *sb,
-		unsigned long mft_no)
-{
-	ntfs_inode *ni = ntfs_alloc_extent_inode();
-
-	ntfs_debug("Entering.");
-	if (likely(ni != NULL)) {
-		__ntfs_init_inode(sb, ni);
-		lockdep_set_class(&ni->mrec_lock, &extent_inode_mrec_lock_key);
-		ni->mft_no = mft_no;
-		ni->type = AT_UNUSED;
-		ni->name = NULL;
-		ni->name_len = 0;
-	}
-	return ni;
-}
-
-/**
- * ntfs_is_extended_system_file - check if a file is in the $Extend directory
- * @ctx:	initialized attribute search context
- *
- * Search all file name attributes in the inode described by the attribute
- * search context @ctx and check if any of the names are in the $Extend system
- * directory.
- *
- * Return values:
- *	   1: file is in $Extend directory
- *	   0: file is not in $Extend directory
- *    -errno: failed to determine if the file is in the $Extend directory
- */
-static int ntfs_is_extended_system_file(ntfs_attr_search_ctx *ctx)
-{
-	int nr_links, err;
-
-	/* Restart search. */
-	ntfs_attr_reinit_search_ctx(ctx);
-
-	/* Get number of hard links. */
-	nr_links = le16_to_cpu(ctx->mrec->link_count);
-
-	/* Loop through all hard links. */
-	while (!(err = ntfs_attr_lookup(AT_FILE_NAME, NULL, 0, 0, 0, NULL, 0,
-			ctx))) {
-		FILE_NAME_ATTR *file_name_attr;
-		ATTR_RECORD *attr = ctx->attr;
-		u8 *p, *p2;
-
-		nr_links--;
-		/*
-		 * Maximum sanity checking as we are called on an inode that
-		 * we suspect might be corrupt.
-		 */
-		p = (u8*)attr + le32_to_cpu(attr->length);
-		if (p < (u8*)ctx->mrec || (u8*)p > (u8*)ctx->mrec +
-				le32_to_cpu(ctx->mrec->bytes_in_use)) {
-err_corrupt_attr:
-			ntfs_error(ctx->ntfs_ino->vol->sb, "Corrupt file name "
-					"attribute. You should run chkdsk.");
-			return -EIO;
-		}
-		if (attr->non_resident) {
-			ntfs_error(ctx->ntfs_ino->vol->sb, "Non-resident file "
-					"name. You should run chkdsk.");
-			return -EIO;
-		}
-		if (attr->flags) {
-			ntfs_error(ctx->ntfs_ino->vol->sb, "File name with "
-					"invalid flags. You should run "
-					"chkdsk.");
-			return -EIO;
-		}
-		if (!(attr->data.resident.flags & RESIDENT_ATTR_IS_INDEXED)) {
-			ntfs_error(ctx->ntfs_ino->vol->sb, "Unindexed file "
-					"name. You should run chkdsk.");
-			return -EIO;
-		}
-		file_name_attr = (FILE_NAME_ATTR*)((u8*)attr +
-				le16_to_cpu(attr->data.resident.value_offset));
-		p2 = (u8 *)file_name_attr + le32_to_cpu(attr->data.resident.value_length);
-		if (p2 < (u8*)attr || p2 > p)
-			goto err_corrupt_attr;
-		/* This attribute is ok, but is it in the $Extend directory? */
-		if (MREF_LE(file_name_attr->parent_directory) == FILE_Extend)
-			return 1;	/* YES, it's an extended system file. */
-	}
-	if (unlikely(err != -ENOENT))
-		return err;
-	if (unlikely(nr_links)) {
-		ntfs_error(ctx->ntfs_ino->vol->sb, "Inode hard link count "
-				"doesn't match number of name attributes. You "
-				"should run chkdsk.");
-		return -EIO;
-	}
-	return 0;	/* NO, it is not an extended system file. */
-}
-
-/**
- * ntfs_read_locked_inode - read an inode from its device
- * @vi:		inode to read
- *
- * ntfs_read_locked_inode() is called from ntfs_iget() to read the inode
- * described by @vi into memory from the device.
- *
- * The only fields in @vi that we need to/can look at when the function is
- * called are i_sb, pointing to the mounted device's super block, and i_ino,
- * the number of the inode to load.
- *
- * ntfs_read_locked_inode() maps, pins and locks the mft record number i_ino
- * for reading and sets up the necessary @vi fields as well as initializing
- * the ntfs inode.
- *
- * Q: What locks are held when the function is called?
- * A: i_state has I_NEW set, hence the inode is locked, also
- *    i_count is set to 1, so it is not going to go away
- *    i_flags is set to 0 and we have no business touching it.  Only an ioctl()
- *    is allowed to write to them. We should of course be honouring them but
- *    we need to do that using the IS_* macros defined in include/linux/fs.h.
- *    In any case ntfs_read_locked_inode() has nothing to do with i_flags.
- *
- * Return 0 on success and -errno on error.  In the error case, the inode will
- * have had make_bad_inode() executed on it.
- */
-static int ntfs_read_locked_inode(struct inode *vi)
-{
-	ntfs_volume *vol = NTFS_SB(vi->i_sb);
-	ntfs_inode *ni;
-	struct inode *bvi;
-	MFT_RECORD *m;
-	ATTR_RECORD *a;
-	STANDARD_INFORMATION *si;
-	ntfs_attr_search_ctx *ctx;
-	int err = 0;
-
-	ntfs_debug("Entering for i_ino 0x%lx.", vi->i_ino);
-
-	/* Setup the generic vfs inode parts now. */
-	vi->i_uid = vol->uid;
-	vi->i_gid = vol->gid;
-	vi->i_mode = 0;
-
-	/*
-	 * Initialize the ntfs specific part of @vi special casing
-	 * FILE_MFT which we need to do at mount time.
-	 */
-	if (vi->i_ino != FILE_MFT)
-		ntfs_init_big_inode(vi);
-	ni = NTFS_I(vi);
-
-	m = map_mft_record(ni);
-	if (IS_ERR(m)) {
-		err = PTR_ERR(m);
-		goto err_out;
-	}
-	ctx = ntfs_attr_get_search_ctx(ni, m);
-	if (!ctx) {
-		err = -ENOMEM;
-		goto unm_err_out;
-	}
-
-	if (!(m->flags & MFT_RECORD_IN_USE)) {
-		ntfs_error(vi->i_sb, "Inode is not in use!");
-		goto unm_err_out;
-	}
-	if (m->base_mft_record) {
-		ntfs_error(vi->i_sb, "Inode is an extent inode!");
-		goto unm_err_out;
-	}
-
-	/* Transfer information from mft record into vfs and ntfs inodes. */
-	vi->i_generation = ni->seq_no = le16_to_cpu(m->sequence_number);
-
-	/*
-	 * FIXME: Keep in mind that link_count is two for files which have both
-	 * a long file name and a short file name as separate entries, so if
-	 * we are hiding short file names this will be too high. Either we need
-	 * to account for the short file names by subtracting them or we need
-	 * to make sure we delete files even though i_nlink is not zero which
-	 * might be tricky due to vfs interactions. Need to think about this
-	 * some more when implementing the unlink command.
-	 */
-	set_nlink(vi, le16_to_cpu(m->link_count));
-	/*
-	 * FIXME: Reparse points can have the directory bit set even though
-	 * they would be S_IFLNK. Need to deal with this further below when we
-	 * implement reparse points / symbolic links but it will do for now.
-	 * Also if not a directory, it could be something else, rather than
-	 * a regular file. But again, will do for now.
-	 */
-	/* Everyone gets all permissions. */
-	vi->i_mode |= S_IRWXUGO;
-	/* If read-only, no one gets write permissions. */
-	if (IS_RDONLY(vi))
-		vi->i_mode &= ~S_IWUGO;
-	if (m->flags & MFT_RECORD_IS_DIRECTORY) {
-		vi->i_mode |= S_IFDIR;
-		/*
-		 * Apply the directory permissions mask set in the mount
-		 * options.
-		 */
-		vi->i_mode &= ~vol->dmask;
-		/* Things break without this kludge! */
-		if (vi->i_nlink > 1)
-			set_nlink(vi, 1);
-	} else {
-		vi->i_mode |= S_IFREG;
-		/* Apply the file permissions mask set in the mount options. */
-		vi->i_mode &= ~vol->fmask;
-	}
-	/*
-	 * Find the standard information attribute in the mft record. At this
-	 * stage we haven't setup the attribute list stuff yet, so this could
-	 * in fact fail if the standard information is in an extent record, but
-	 * I don't think this actually ever happens.
-	 */
-	err = ntfs_attr_lookup(AT_STANDARD_INFORMATION, NULL, 0, 0, 0, NULL, 0,
-			ctx);
-	if (unlikely(err)) {
-		if (err == -ENOENT) {
-			/*
-			 * TODO: We should be performing a hot fix here (if the
-			 * recover mount option is set) by creating a new
-			 * attribute.
-			 */
-			ntfs_error(vi->i_sb, "$STANDARD_INFORMATION attribute "
-					"is missing.");
-		}
-		goto unm_err_out;
-	}
-	a = ctx->attr;
-	/* Get the standard information attribute value. */
-	if ((u8 *)a + le16_to_cpu(a->data.resident.value_offset)
-			+ le32_to_cpu(a->data.resident.value_length) >
-			(u8 *)ctx->mrec + vol->mft_record_size) {
-		ntfs_error(vi->i_sb, "Corrupt standard information attribute in inode.");
-		goto unm_err_out;
-	}
-	si = (STANDARD_INFORMATION*)((u8*)a +
-			le16_to_cpu(a->data.resident.value_offset));
-
-	/* Transfer information from the standard information into vi. */
-	/*
-	 * Note: The i_?times do not quite map perfectly onto the NTFS times,
-	 * but they are close enough, and in the end it doesn't really matter
-	 * that much...
-	 */
-	/*
-	 * mtime is the last change of the data within the file. Not changed
-	 * when only metadata is changed, e.g. a rename doesn't affect mtime.
-	 */
-	inode_set_mtime_to_ts(vi, ntfs2utc(si->last_data_change_time));
-	/*
-	 * ctime is the last change of the metadata of the file. This obviously
-	 * always changes, when mtime is changed. ctime can be changed on its
-	 * own, mtime is then not changed, e.g. when a file is renamed.
-	 */
-	inode_set_ctime_to_ts(vi, ntfs2utc(si->last_mft_change_time));
-	/*
-	 * Last access to the data within the file. Not changed during a rename
-	 * for example but changed whenever the file is written to.
-	 */
-	inode_set_atime_to_ts(vi, ntfs2utc(si->last_access_time));
-
-	/* Find the attribute list attribute if present. */
-	ntfs_attr_reinit_search_ctx(ctx);
-	err = ntfs_attr_lookup(AT_ATTRIBUTE_LIST, NULL, 0, 0, 0, NULL, 0, ctx);
-	if (err) {
-		if (unlikely(err != -ENOENT)) {
-			ntfs_error(vi->i_sb, "Failed to lookup attribute list "
-					"attribute.");
-			goto unm_err_out;
-		}
-	} else /* if (!err) */ {
-		if (vi->i_ino == FILE_MFT)
-			goto skip_attr_list_load;
-		ntfs_debug("Attribute list found in inode 0x%lx.", vi->i_ino);
-		NInoSetAttrList(ni);
-		a = ctx->attr;
-		if (a->flags & ATTR_COMPRESSION_MASK) {
-			ntfs_error(vi->i_sb, "Attribute list attribute is "
-					"compressed.");
-			goto unm_err_out;
-		}
-		if (a->flags & ATTR_IS_ENCRYPTED ||
-				a->flags & ATTR_IS_SPARSE) {
-			if (a->non_resident) {
-				ntfs_error(vi->i_sb, "Non-resident attribute "
-						"list attribute is encrypted/"
-						"sparse.");
-				goto unm_err_out;
-			}
-			ntfs_warning(vi->i_sb, "Resident attribute list "
-					"attribute in inode 0x%lx is marked "
-					"encrypted/sparse which is not true.  "
-					"However, Windows allows this and "
-					"chkdsk does not detect or correct it "
-					"so we will just ignore the invalid "
-					"flags and pretend they are not set.",
-					vi->i_ino);
-		}
-		/* Now allocate memory for the attribute list. */
-		ni->attr_list_size = (u32)ntfs_attr_size(a);
-		ni->attr_list = ntfs_malloc_nofs(ni->attr_list_size);
-		if (!ni->attr_list) {
-			ntfs_error(vi->i_sb, "Not enough memory to allocate "
-					"buffer for attribute list.");
-			err = -ENOMEM;
-			goto unm_err_out;
-		}
-		if (a->non_resident) {
-			NInoSetAttrListNonResident(ni);
-			if (a->data.non_resident.lowest_vcn) {
-				ntfs_error(vi->i_sb, "Attribute list has non "
-						"zero lowest_vcn.");
-				goto unm_err_out;
-			}
-			/*
-			 * Setup the runlist. No need for locking as we have
-			 * exclusive access to the inode at this time.
-			 */
-			ni->attr_list_rl.rl = ntfs_mapping_pairs_decompress(vol,
-					a, NULL);
-			if (IS_ERR(ni->attr_list_rl.rl)) {
-				err = PTR_ERR(ni->attr_list_rl.rl);
-				ni->attr_list_rl.rl = NULL;
-				ntfs_error(vi->i_sb, "Mapping pairs "
-						"decompression failed.");
-				goto unm_err_out;
-			}
-			/* Now load the attribute list. */
-			if ((err = load_attribute_list(vol, &ni->attr_list_rl,
-					ni->attr_list, ni->attr_list_size,
-					sle64_to_cpu(a->data.non_resident.
-					initialized_size)))) {
-				ntfs_error(vi->i_sb, "Failed to load "
-						"attribute list attribute.");
-				goto unm_err_out;
-			}
-		} else /* if (!a->non_resident) */ {
-			if ((u8*)a + le16_to_cpu(a->data.resident.value_offset)
-					+ le32_to_cpu(
-					a->data.resident.value_length) >
-					(u8*)ctx->mrec + vol->mft_record_size) {
-				ntfs_error(vi->i_sb, "Corrupt attribute list "
-						"in inode.");
-				goto unm_err_out;
-			}
-			/* Now copy the attribute list. */
-			memcpy(ni->attr_list, (u8*)a + le16_to_cpu(
-					a->data.resident.value_offset),
-					le32_to_cpu(
-					a->data.resident.value_length));
-		}
-	}
-skip_attr_list_load:
-	/*
-	 * If an attribute list is present we now have the attribute list value
-	 * in ntfs_ino->attr_list and it is ntfs_ino->attr_list_size bytes.
-	 */
-	if (S_ISDIR(vi->i_mode)) {
-		loff_t bvi_size;
-		ntfs_inode *bni;
-		INDEX_ROOT *ir;
-		u8 *ir_end, *index_end;
-
-		/* It is a directory, find index root attribute. */
-		ntfs_attr_reinit_search_ctx(ctx);
-		err = ntfs_attr_lookup(AT_INDEX_ROOT, I30, 4, CASE_SENSITIVE,
-				0, NULL, 0, ctx);
-		if (unlikely(err)) {
-			if (err == -ENOENT) {
-				// FIXME: File is corrupt! Hot-fix with empty
-				// index root attribute if recovery option is
-				// set.
-				ntfs_error(vi->i_sb, "$INDEX_ROOT attribute "
-						"is missing.");
-			}
-			goto unm_err_out;
-		}
-		a = ctx->attr;
-		/* Set up the state. */
-		if (unlikely(a->non_resident)) {
-			ntfs_error(vol->sb, "$INDEX_ROOT attribute is not "
-					"resident.");
-			goto unm_err_out;
-		}
-		/* Ensure the attribute name is placed before the value. */
-		if (unlikely(a->name_length && (le16_to_cpu(a->name_offset) >=
-				le16_to_cpu(a->data.resident.value_offset)))) {
-			ntfs_error(vol->sb, "$INDEX_ROOT attribute name is "
-					"placed after the attribute value.");
-			goto unm_err_out;
-		}
-		/*
-		 * Compressed/encrypted index root just means that the newly
-		 * created files in that directory should be created compressed/
-		 * encrypted. However index root cannot be both compressed and
-		 * encrypted.
-		 */
-		if (a->flags & ATTR_COMPRESSION_MASK)
-			NInoSetCompressed(ni);
-		if (a->flags & ATTR_IS_ENCRYPTED) {
-			if (a->flags & ATTR_COMPRESSION_MASK) {
-				ntfs_error(vi->i_sb, "Found encrypted and "
-						"compressed attribute.");
-				goto unm_err_out;
-			}
-			NInoSetEncrypted(ni);
-		}
-		if (a->flags & ATTR_IS_SPARSE)
-			NInoSetSparse(ni);
-		ir = (INDEX_ROOT*)((u8*)a +
-				le16_to_cpu(a->data.resident.value_offset));
-		ir_end = (u8*)ir + le32_to_cpu(a->data.resident.value_length);
-		if (ir_end > (u8*)ctx->mrec + vol->mft_record_size) {
-			ntfs_error(vi->i_sb, "$INDEX_ROOT attribute is "
-					"corrupt.");
-			goto unm_err_out;
-		}
-		index_end = (u8*)&ir->index +
-				le32_to_cpu(ir->index.index_length);
-		if (index_end > ir_end) {
-			ntfs_error(vi->i_sb, "Directory index is corrupt.");
-			goto unm_err_out;
-		}
-		if (ir->type != AT_FILE_NAME) {
-			ntfs_error(vi->i_sb, "Indexed attribute is not "
-					"$FILE_NAME.");
-			goto unm_err_out;
-		}
-		if (ir->collation_rule != COLLATION_FILE_NAME) {
-			ntfs_error(vi->i_sb, "Index collation rule is not "
-					"COLLATION_FILE_NAME.");
-			goto unm_err_out;
-		}
-		ni->itype.index.collation_rule = ir->collation_rule;
-		ni->itype.index.block_size = le32_to_cpu(ir->index_block_size);
-		if (ni->itype.index.block_size &
-				(ni->itype.index.block_size - 1)) {
-			ntfs_error(vi->i_sb, "Index block size (%u) is not a "
-					"power of two.",
-					ni->itype.index.block_size);
-			goto unm_err_out;
-		}
-		if (ni->itype.index.block_size > PAGE_SIZE) {
-			ntfs_error(vi->i_sb, "Index block size (%u) > "
-					"PAGE_SIZE (%ld) is not "
-					"supported.  Sorry.",
-					ni->itype.index.block_size,
-					PAGE_SIZE);
-			err = -EOPNOTSUPP;
-			goto unm_err_out;
-		}
-		if (ni->itype.index.block_size < NTFS_BLOCK_SIZE) {
-			ntfs_error(vi->i_sb, "Index block size (%u) < "
-					"NTFS_BLOCK_SIZE (%i) is not "
-					"supported.  Sorry.",
-					ni->itype.index.block_size,
-					NTFS_BLOCK_SIZE);
-			err = -EOPNOTSUPP;
-			goto unm_err_out;
-		}
-		ni->itype.index.block_size_bits =
-				ffs(ni->itype.index.block_size) - 1;
-		/* Determine the size of a vcn in the directory index. */
-		if (vol->cluster_size <= ni->itype.index.block_size) {
-			ni->itype.index.vcn_size = vol->cluster_size;
-			ni->itype.index.vcn_size_bits = vol->cluster_size_bits;
-		} else {
-			ni->itype.index.vcn_size = vol->sector_size;
-			ni->itype.index.vcn_size_bits = vol->sector_size_bits;
-		}
-
-		/* Setup the index allocation attribute, even if not present. */
-		NInoSetMstProtected(ni);
-		ni->type = AT_INDEX_ALLOCATION;
-		ni->name = I30;
-		ni->name_len = 4;
-
-		if (!(ir->index.flags & LARGE_INDEX)) {
-			/* No index allocation. */
-			vi->i_size = ni->initialized_size =
-					ni->allocated_size = 0;
-			/* We are done with the mft record, so we release it. */
-			ntfs_attr_put_search_ctx(ctx);
-			unmap_mft_record(ni);
-			m = NULL;
-			ctx = NULL;
-			goto skip_large_dir_stuff;
-		} /* LARGE_INDEX: Index allocation present. Setup state. */
-		NInoSetIndexAllocPresent(ni);
-		/* Find index allocation attribute. */
-		ntfs_attr_reinit_search_ctx(ctx);
-		err = ntfs_attr_lookup(AT_INDEX_ALLOCATION, I30, 4,
-				CASE_SENSITIVE, 0, NULL, 0, ctx);
-		if (unlikely(err)) {
-			if (err == -ENOENT)
-				ntfs_error(vi->i_sb, "$INDEX_ALLOCATION "
-						"attribute is not present but "
-						"$INDEX_ROOT indicated it is.");
-			else
-				ntfs_error(vi->i_sb, "Failed to lookup "
-						"$INDEX_ALLOCATION "
-						"attribute.");
-			goto unm_err_out;
-		}
-		a = ctx->attr;
-		if (!a->non_resident) {
-			ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute "
-					"is resident.");
-			goto unm_err_out;
-		}
-		/*
-		 * Ensure the attribute name is placed before the mapping pairs
-		 * array.
-		 */
-		if (unlikely(a->name_length && (le16_to_cpu(a->name_offset) >=
-				le16_to_cpu(
-				a->data.non_resident.mapping_pairs_offset)))) {
-			ntfs_error(vol->sb, "$INDEX_ALLOCATION attribute name "
-					"is placed after the mapping pairs "
-					"array.");
-			goto unm_err_out;
-		}
-		if (a->flags & ATTR_IS_ENCRYPTED) {
-			ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute "
-					"is encrypted.");
-			goto unm_err_out;
-		}
-		if (a->flags & ATTR_IS_SPARSE) {
-			ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute "
-					"is sparse.");
-			goto unm_err_out;
-		}
-		if (a->flags & ATTR_COMPRESSION_MASK) {
-			ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute "
-					"is compressed.");
-			goto unm_err_out;
-		}
-		if (a->data.non_resident.lowest_vcn) {
-			ntfs_error(vi->i_sb, "First extent of "
-					"$INDEX_ALLOCATION attribute has non "
-					"zero lowest_vcn.");
-			goto unm_err_out;
-		}
-		vi->i_size = sle64_to_cpu(a->data.non_resident.data_size);
-		ni->initialized_size = sle64_to_cpu(
-				a->data.non_resident.initialized_size);
-		ni->allocated_size = sle64_to_cpu(
-				a->data.non_resident.allocated_size);
-		/*
-		 * We are done with the mft record, so we release it. Otherwise
-		 * we would deadlock in ntfs_attr_iget().
-		 */
-		ntfs_attr_put_search_ctx(ctx);
-		unmap_mft_record(ni);
-		m = NULL;
-		ctx = NULL;
-		/* Get the index bitmap attribute inode. */
-		bvi = ntfs_attr_iget(vi, AT_BITMAP, I30, 4);
-		if (IS_ERR(bvi)) {
-			ntfs_error(vi->i_sb, "Failed to get bitmap attribute.");
-			err = PTR_ERR(bvi);
-			goto unm_err_out;
-		}
-		bni = NTFS_I(bvi);
-		if (NInoCompressed(bni) || NInoEncrypted(bni) ||
-				NInoSparse(bni)) {
-			ntfs_error(vi->i_sb, "$BITMAP attribute is compressed "
-					"and/or encrypted and/or sparse.");
-			goto iput_unm_err_out;
-		}
-		/* Consistency check bitmap size vs. index allocation size. */
-		bvi_size = i_size_read(bvi);
-		if ((bvi_size << 3) < (vi->i_size >>
-				ni->itype.index.block_size_bits)) {
-			ntfs_error(vi->i_sb, "Index bitmap too small (0x%llx) "
-					"for index allocation (0x%llx).",
-					bvi_size << 3, vi->i_size);
-			goto iput_unm_err_out;
-		}
-		/* No longer need the bitmap attribute inode. */
-		iput(bvi);
-skip_large_dir_stuff:
-		/* Setup the operations for this inode. */
-		vi->i_op = &ntfs_dir_inode_ops;
-		vi->i_fop = &ntfs_dir_ops;
-		vi->i_mapping->a_ops = &ntfs_mst_aops;
-	} else {
-		/* It is a file. */
-		ntfs_attr_reinit_search_ctx(ctx);
-
-		/* Setup the data attribute, even if not present. */
-		ni->type = AT_DATA;
-		ni->name = NULL;
-		ni->name_len = 0;
-
-		/* Find first extent of the unnamed data attribute. */
-		err = ntfs_attr_lookup(AT_DATA, NULL, 0, 0, 0, NULL, 0, ctx);
-		if (unlikely(err)) {
-			vi->i_size = ni->initialized_size =
-					ni->allocated_size = 0;
-			if (err != -ENOENT) {
-				ntfs_error(vi->i_sb, "Failed to lookup $DATA "
-						"attribute.");
-				goto unm_err_out;
-			}
-			/*
-			 * FILE_Secure does not have an unnamed $DATA
-			 * attribute, so we special case it here.
-			 */
-			if (vi->i_ino == FILE_Secure)
-				goto no_data_attr_special_case;
-			/*
-			 * Most if not all the system files in the $Extend
-			 * system directory do not have unnamed data
-			 * attributes so we need to check if the parent
-			 * directory of the file is FILE_Extend and if it is
-			 * ignore this error. To do this we need to get the
-			 * name of this inode from the mft record as the name
-			 * contains the back reference to the parent directory.
-			 */
-			if (ntfs_is_extended_system_file(ctx) > 0)
-				goto no_data_attr_special_case;
-			// FIXME: File is corrupt! Hot-fix with empty data
-			// attribute if recovery option is set.
-			ntfs_error(vi->i_sb, "$DATA attribute is missing.");
-			goto unm_err_out;
-		}
-		a = ctx->attr;
-		/* Setup the state. */
-		if (a->flags & (ATTR_COMPRESSION_MASK | ATTR_IS_SPARSE)) {
-			if (a->flags & ATTR_COMPRESSION_MASK) {
-				NInoSetCompressed(ni);
-				if (vol->cluster_size > 4096) {
-					ntfs_error(vi->i_sb, "Found "
-							"compressed data but "
-							"compression is "
-							"disabled due to "
-							"cluster size (%i) > "
-							"4kiB.",
-							vol->cluster_size);
-					goto unm_err_out;
-				}
-				if ((a->flags & ATTR_COMPRESSION_MASK)
-						!= ATTR_IS_COMPRESSED) {
-					ntfs_error(vi->i_sb, "Found unknown "
-							"compression method "
-							"or corrupt file.");
-					goto unm_err_out;
-				}
-			}
-			if (a->flags & ATTR_IS_SPARSE)
-				NInoSetSparse(ni);
-		}
-		if (a->flags & ATTR_IS_ENCRYPTED) {
-			if (NInoCompressed(ni)) {
-				ntfs_error(vi->i_sb, "Found encrypted and "
-						"compressed data.");
-				goto unm_err_out;
-			}
-			NInoSetEncrypted(ni);
-		}
-		if (a->non_resident) {
-			NInoSetNonResident(ni);
-			if (NInoCompressed(ni) || NInoSparse(ni)) {
-				if (NInoCompressed(ni) && a->data.non_resident.
-						compression_unit != 4) {
-					ntfs_error(vi->i_sb, "Found "
-							"non-standard "
-							"compression unit (%u "
-							"instead of 4).  "
-							"Cannot handle this.",
-							a->data.non_resident.
-							compression_unit);
-					err = -EOPNOTSUPP;
-					goto unm_err_out;
-				}
-				if (a->data.non_resident.compression_unit) {
-					ni->itype.compressed.block_size = 1U <<
-							(a->data.non_resident.
-							compression_unit +
-							vol->cluster_size_bits);
-					ni->itype.compressed.block_size_bits =
-							ffs(ni->itype.
-							compressed.
-							block_size) - 1;
-					ni->itype.compressed.block_clusters =
-							1U << a->data.
-							non_resident.
-							compression_unit;
-				} else {
-					ni->itype.compressed.block_size = 0;
-					ni->itype.compressed.block_size_bits =
-							0;
-					ni->itype.compressed.block_clusters =
-							0;
-				}
-				ni->itype.compressed.size = sle64_to_cpu(
-						a->data.non_resident.
-						compressed_size);
-			}
-			if (a->data.non_resident.lowest_vcn) {
-				ntfs_error(vi->i_sb, "First extent of $DATA "
-						"attribute has non zero "
-						"lowest_vcn.");
-				goto unm_err_out;
-			}
-			vi->i_size = sle64_to_cpu(
-					a->data.non_resident.data_size);
-			ni->initialized_size = sle64_to_cpu(
-					a->data.non_resident.initialized_size);
-			ni->allocated_size = sle64_to_cpu(
-					a->data.non_resident.allocated_size);
-		} else { /* Resident attribute. */
-			vi->i_size = ni->initialized_size = le32_to_cpu(
-					a->data.resident.value_length);
-			ni->allocated_size = le32_to_cpu(a->length) -
-					le16_to_cpu(
-					a->data.resident.value_offset);
-			if (vi->i_size > ni->allocated_size) {
-				ntfs_error(vi->i_sb, "Resident data attribute "
-						"is corrupt (size exceeds "
-						"allocation).");
-				goto unm_err_out;
-			}
-		}
-no_data_attr_special_case:
-		/* We are done with the mft record, so we release it. */
-		ntfs_attr_put_search_ctx(ctx);
-		unmap_mft_record(ni);
-		m = NULL;
-		ctx = NULL;
-		/* Setup the operations for this inode. */
-		vi->i_op = &ntfs_file_inode_ops;
-		vi->i_fop = &ntfs_file_ops;
-		vi->i_mapping->a_ops = &ntfs_normal_aops;
-		if (NInoMstProtected(ni))
-			vi->i_mapping->a_ops = &ntfs_mst_aops;
-		else if (NInoCompressed(ni))
-			vi->i_mapping->a_ops = &ntfs_compressed_aops;
-	}
-	/*
-	 * The number of 512-byte blocks used on disk (for stat). This is in so
-	 * far inaccurate as it doesn't account for any named streams or other
-	 * special non-resident attributes, but that is how Windows works, too,
-	 * so we are at least consistent with Windows, if not entirely
-	 * consistent with the Linux Way. Doing it the Linux Way would cause a
-	 * significant slowdown as it would involve iterating over all
-	 * attributes in the mft record and adding the allocated/compressed
-	 * sizes of all non-resident attributes present to give us the Linux
-	 * correct size that should go into i_blocks (after division by 512).
-	 */
-	if (S_ISREG(vi->i_mode) && (NInoCompressed(ni) || NInoSparse(ni)))
-		vi->i_blocks = ni->itype.compressed.size >> 9;
-	else
-		vi->i_blocks = ni->allocated_size >> 9;
-	ntfs_debug("Done.");
-	return 0;
-iput_unm_err_out:
-	iput(bvi);
-unm_err_out:
-	if (!err)
-		err = -EIO;
-	if (ctx)
-		ntfs_attr_put_search_ctx(ctx);
-	if (m)
-		unmap_mft_record(ni);
-err_out:
-	ntfs_error(vol->sb, "Failed with error code %i.  Marking corrupt "
-			"inode 0x%lx as bad.  Run chkdsk.", err, vi->i_ino);
-	make_bad_inode(vi);
-	if (err != -EOPNOTSUPP && err != -ENOMEM)
-		NVolSetErrors(vol);
-	return err;
-}
-
-/**
- * ntfs_read_locked_attr_inode - read an attribute inode from its base inode
- * @base_vi:	base inode
- * @vi:		attribute inode to read
- *
- * ntfs_read_locked_attr_inode() is called from ntfs_attr_iget() to read the
- * attribute inode described by @vi into memory from the base mft record
- * described by @base_ni.
- *
- * ntfs_read_locked_attr_inode() maps, pins and locks the base inode for
- * reading and looks up the attribute described by @vi before setting up the
- * necessary fields in @vi as well as initializing the ntfs inode.
- *
- * Q: What locks are held when the function is called?
- * A: i_state has I_NEW set, hence the inode is locked, also
- *    i_count is set to 1, so it is not going to go away
- *
- * Return 0 on success and -errno on error.  In the error case, the inode will
- * have had make_bad_inode() executed on it.
- *
- * Note this cannot be called for AT_INDEX_ALLOCATION.
- */
-static int ntfs_read_locked_attr_inode(struct inode *base_vi, struct inode *vi)
-{
-	ntfs_volume *vol = NTFS_SB(vi->i_sb);
-	ntfs_inode *ni, *base_ni;
-	MFT_RECORD *m;
-	ATTR_RECORD *a;
-	ntfs_attr_search_ctx *ctx;
-	int err = 0;
-
-	ntfs_debug("Entering for i_ino 0x%lx.", vi->i_ino);
-
-	ntfs_init_big_inode(vi);
-
-	ni	= NTFS_I(vi);
-	base_ni = NTFS_I(base_vi);
-
-	/* Just mirror the values from the base inode. */
-	vi->i_uid	= base_vi->i_uid;
-	vi->i_gid	= base_vi->i_gid;
-	set_nlink(vi, base_vi->i_nlink);
-	inode_set_mtime_to_ts(vi, inode_get_mtime(base_vi));
-	inode_set_ctime_to_ts(vi, inode_get_ctime(base_vi));
-	inode_set_atime_to_ts(vi, inode_get_atime(base_vi));
-	vi->i_generation = ni->seq_no = base_ni->seq_no;
-
-	/* Set inode type to zero but preserve permissions. */
-	vi->i_mode	= base_vi->i_mode & ~S_IFMT;
-
-	m = map_mft_record(base_ni);
-	if (IS_ERR(m)) {
-		err = PTR_ERR(m);
-		goto err_out;
-	}
-	ctx = ntfs_attr_get_search_ctx(base_ni, m);
-	if (!ctx) {
-		err = -ENOMEM;
-		goto unm_err_out;
-	}
-	/* Find the attribute. */
-	err = ntfs_attr_lookup(ni->type, ni->name, ni->name_len,
-			CASE_SENSITIVE, 0, NULL, 0, ctx);
-	if (unlikely(err))
-		goto unm_err_out;
-	a = ctx->attr;
-	if (a->flags & (ATTR_COMPRESSION_MASK | ATTR_IS_SPARSE)) {
-		if (a->flags & ATTR_COMPRESSION_MASK) {
-			NInoSetCompressed(ni);
-			if ((ni->type != AT_DATA) || (ni->type == AT_DATA &&
-					ni->name_len)) {
-				ntfs_error(vi->i_sb, "Found compressed "
-						"non-data or named data "
-						"attribute.  Please report "
-						"you saw this message to "
-						"linux-ntfs-dev@lists."
-						"sourceforge.net");
-				goto unm_err_out;
-			}
-			if (vol->cluster_size > 4096) {
-				ntfs_error(vi->i_sb, "Found compressed "
-						"attribute but compression is "
-						"disabled due to cluster size "
-						"(%i) > 4kiB.",
-						vol->cluster_size);
-				goto unm_err_out;
-			}
-			if ((a->flags & ATTR_COMPRESSION_MASK) !=
-					ATTR_IS_COMPRESSED) {
-				ntfs_error(vi->i_sb, "Found unknown "
-						"compression method.");
-				goto unm_err_out;
-			}
-		}
-		/*
-		 * The compressed/sparse flag set in an index root just means
-		 * to compress all files.
-		 */
-		if (NInoMstProtected(ni) && ni->type != AT_INDEX_ROOT) {
-			ntfs_error(vi->i_sb, "Found mst protected attribute "
-					"but the attribute is %s.  Please "
-					"report you saw this message to "
-					"linux-ntfs-dev@lists.sourceforge.net",
-					NInoCompressed(ni) ? "compressed" :
-					"sparse");
-			goto unm_err_out;
-		}
-		if (a->flags & ATTR_IS_SPARSE)
-			NInoSetSparse(ni);
-	}
-	if (a->flags & ATTR_IS_ENCRYPTED) {
-		if (NInoCompressed(ni)) {
-			ntfs_error(vi->i_sb, "Found encrypted and compressed "
-					"data.");
-			goto unm_err_out;
-		}
-		/*
-		 * The encryption flag set in an index root just means to
-		 * encrypt all files.
-		 */
-		if (NInoMstProtected(ni) && ni->type != AT_INDEX_ROOT) {
-			ntfs_error(vi->i_sb, "Found mst protected attribute "
-					"but the attribute is encrypted.  "
-					"Please report you saw this message "
-					"to linux-ntfs-dev@lists.sourceforge."
-					"net");
-			goto unm_err_out;
-		}
-		if (ni->type != AT_DATA) {
-			ntfs_error(vi->i_sb, "Found encrypted non-data "
-					"attribute.");
-			goto unm_err_out;
-		}
-		NInoSetEncrypted(ni);
-	}
-	if (!a->non_resident) {
-		/* Ensure the attribute name is placed before the value. */
-		if (unlikely(a->name_length && (le16_to_cpu(a->name_offset) >=
-				le16_to_cpu(a->data.resident.value_offset)))) {
-			ntfs_error(vol->sb, "Attribute name is placed after "
-					"the attribute value.");
-			goto unm_err_out;
-		}
-		if (NInoMstProtected(ni)) {
-			ntfs_error(vi->i_sb, "Found mst protected attribute "
-					"but the attribute is resident.  "
-					"Please report you saw this message to "
-					"linux-ntfs-dev@lists.sourceforge.net");
-			goto unm_err_out;
-		}
-		vi->i_size = ni->initialized_size = le32_to_cpu(
-				a->data.resident.value_length);
-		ni->allocated_size = le32_to_cpu(a->length) -
-				le16_to_cpu(a->data.resident.value_offset);
-		if (vi->i_size > ni->allocated_size) {
-			ntfs_error(vi->i_sb, "Resident attribute is corrupt "
-					"(size exceeds allocation).");
-			goto unm_err_out;
-		}
-	} else {
-		NInoSetNonResident(ni);
-		/*
-		 * Ensure the attribute name is placed before the mapping pairs
-		 * array.
-		 */
-		if (unlikely(a->name_length && (le16_to_cpu(a->name_offset) >=
-				le16_to_cpu(
-				a->data.non_resident.mapping_pairs_offset)))) {
-			ntfs_error(vol->sb, "Attribute name is placed after "
-					"the mapping pairs array.");
-			goto unm_err_out;
-		}
-		if (NInoCompressed(ni) || NInoSparse(ni)) {
-			if (NInoCompressed(ni) && a->data.non_resident.
-					compression_unit != 4) {
-				ntfs_error(vi->i_sb, "Found non-standard "
-						"compression unit (%u instead "
-						"of 4).  Cannot handle this.",
-						a->data.non_resident.
-						compression_unit);
-				err = -EOPNOTSUPP;
-				goto unm_err_out;
-			}
-			if (a->data.non_resident.compression_unit) {
-				ni->itype.compressed.block_size = 1U <<
-						(a->data.non_resident.
-						compression_unit +
-						vol->cluster_size_bits);
-				ni->itype.compressed.block_size_bits =
-						ffs(ni->itype.compressed.
-						block_size) - 1;
-				ni->itype.compressed.block_clusters = 1U <<
-						a->data.non_resident.
-						compression_unit;
-			} else {
-				ni->itype.compressed.block_size = 0;
-				ni->itype.compressed.block_size_bits = 0;
-				ni->itype.compressed.block_clusters = 0;
-			}
-			ni->itype.compressed.size = sle64_to_cpu(
-					a->data.non_resident.compressed_size);
-		}
-		if (a->data.non_resident.lowest_vcn) {
-			ntfs_error(vi->i_sb, "First extent of attribute has "
-					"non-zero lowest_vcn.");
-			goto unm_err_out;
-		}
-		vi->i_size = sle64_to_cpu(a->data.non_resident.data_size);
-		ni->initialized_size = sle64_to_cpu(
-				a->data.non_resident.initialized_size);
-		ni->allocated_size = sle64_to_cpu(
-				a->data.non_resident.allocated_size);
-	}
-	vi->i_mapping->a_ops = &ntfs_normal_aops;
-	if (NInoMstProtected(ni))
-		vi->i_mapping->a_ops = &ntfs_mst_aops;
-	else if (NInoCompressed(ni))
-		vi->i_mapping->a_ops = &ntfs_compressed_aops;
-	if ((NInoCompressed(ni) || NInoSparse(ni)) && ni->type != AT_INDEX_ROOT)
-		vi->i_blocks = ni->itype.compressed.size >> 9;
-	else
-		vi->i_blocks = ni->allocated_size >> 9;
-	/*
-	 * Make sure the base inode does not go away and attach it to the
-	 * attribute inode.
-	 */
-	igrab(base_vi);
-	ni->ext.base_ntfs_ino = base_ni;
-	ni->nr_extents = -1;
-
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(base_ni);
-
-	ntfs_debug("Done.");
-	return 0;
-
-unm_err_out:
-	if (!err)
-		err = -EIO;
-	if (ctx)
-		ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(base_ni);
-err_out:
-	ntfs_error(vol->sb, "Failed with error code %i while reading attribute "
-			"inode (mft_no 0x%lx, type 0x%x, name_len %i).  "
-			"Marking corrupt inode and base inode 0x%lx as bad.  "
-			"Run chkdsk.", err, vi->i_ino, ni->type, ni->name_len,
-			base_vi->i_ino);
-	make_bad_inode(vi);
-	if (err != -ENOMEM)
-		NVolSetErrors(vol);
-	return err;
-}
-
-/**
- * ntfs_read_locked_index_inode - read an index inode from its base inode
- * @base_vi:	base inode
- * @vi:		index inode to read
- *
- * ntfs_read_locked_index_inode() is called from ntfs_index_iget() to read the
- * index inode described by @vi into memory from the base mft record described
- * by @base_ni.
- *
- * ntfs_read_locked_index_inode() maps, pins and locks the base inode for
- * reading and looks up the attributes relating to the index described by @vi
- * before setting up the necessary fields in @vi as well as initializing the
- * ntfs inode.
- *
- * Note, index inodes are essentially attribute inodes (NInoAttr() is true)
- * with the attribute type set to AT_INDEX_ALLOCATION.  Apart from that, they
- * are setup like directory inodes since directories are a special case of
- * indices ao they need to be treated in much the same way.  Most importantly,
- * for small indices the index allocation attribute might not actually exist.
- * However, the index root attribute always exists but this does not need to
- * have an inode associated with it and this is why we define a new inode type
- * index.  Also, like for directories, we need to have an attribute inode for
- * the bitmap attribute corresponding to the index allocation attribute and we
- * can store this in the appropriate field of the inode, just like we do for
- * normal directory inodes.
- *
- * Q: What locks are held when the function is called?
- * A: i_state has I_NEW set, hence the inode is locked, also
- *    i_count is set to 1, so it is not going to go away
- *
- * Return 0 on success and -errno on error.  In the error case, the inode will
- * have had make_bad_inode() executed on it.
- */
-static int ntfs_read_locked_index_inode(struct inode *base_vi, struct inode *vi)
-{
-	loff_t bvi_size;
-	ntfs_volume *vol = NTFS_SB(vi->i_sb);
-	ntfs_inode *ni, *base_ni, *bni;
-	struct inode *bvi;
-	MFT_RECORD *m;
-	ATTR_RECORD *a;
-	ntfs_attr_search_ctx *ctx;
-	INDEX_ROOT *ir;
-	u8 *ir_end, *index_end;
-	int err = 0;
-
-	ntfs_debug("Entering for i_ino 0x%lx.", vi->i_ino);
-	ntfs_init_big_inode(vi);
-	ni	= NTFS_I(vi);
-	base_ni = NTFS_I(base_vi);
-	/* Just mirror the values from the base inode. */
-	vi->i_uid	= base_vi->i_uid;
-	vi->i_gid	= base_vi->i_gid;
-	set_nlink(vi, base_vi->i_nlink);
-	inode_set_mtime_to_ts(vi, inode_get_mtime(base_vi));
-	inode_set_ctime_to_ts(vi, inode_get_ctime(base_vi));
-	inode_set_atime_to_ts(vi, inode_get_atime(base_vi));
-	vi->i_generation = ni->seq_no = base_ni->seq_no;
-	/* Set inode type to zero but preserve permissions. */
-	vi->i_mode	= base_vi->i_mode & ~S_IFMT;
-	/* Map the mft record for the base inode. */
-	m = map_mft_record(base_ni);
-	if (IS_ERR(m)) {
-		err = PTR_ERR(m);
-		goto err_out;
-	}
-	ctx = ntfs_attr_get_search_ctx(base_ni, m);
-	if (!ctx) {
-		err = -ENOMEM;
-		goto unm_err_out;
-	}
-	/* Find the index root attribute. */
-	err = ntfs_attr_lookup(AT_INDEX_ROOT, ni->name, ni->name_len,
-			CASE_SENSITIVE, 0, NULL, 0, ctx);
-	if (unlikely(err)) {
-		if (err == -ENOENT)
-			ntfs_error(vi->i_sb, "$INDEX_ROOT attribute is "
-					"missing.");
-		goto unm_err_out;
-	}
-	a = ctx->attr;
-	/* Set up the state. */
-	if (unlikely(a->non_resident)) {
-		ntfs_error(vol->sb, "$INDEX_ROOT attribute is not resident.");
-		goto unm_err_out;
-	}
-	/* Ensure the attribute name is placed before the value. */
-	if (unlikely(a->name_length && (le16_to_cpu(a->name_offset) >=
-			le16_to_cpu(a->data.resident.value_offset)))) {
-		ntfs_error(vol->sb, "$INDEX_ROOT attribute name is placed "
-				"after the attribute value.");
-		goto unm_err_out;
-	}
-	/*
-	 * Compressed/encrypted/sparse index root is not allowed, except for
-	 * directories of course but those are not dealt with here.
-	 */
-	if (a->flags & (ATTR_COMPRESSION_MASK | ATTR_IS_ENCRYPTED |
-			ATTR_IS_SPARSE)) {
-		ntfs_error(vi->i_sb, "Found compressed/encrypted/sparse index "
-				"root attribute.");
-		goto unm_err_out;
-	}
-	ir = (INDEX_ROOT*)((u8*)a + le16_to_cpu(a->data.resident.value_offset));
-	ir_end = (u8*)ir + le32_to_cpu(a->data.resident.value_length);
-	if (ir_end > (u8*)ctx->mrec + vol->mft_record_size) {
-		ntfs_error(vi->i_sb, "$INDEX_ROOT attribute is corrupt.");
-		goto unm_err_out;
-	}
-	index_end = (u8*)&ir->index + le32_to_cpu(ir->index.index_length);
-	if (index_end > ir_end) {
-		ntfs_error(vi->i_sb, "Index is corrupt.");
-		goto unm_err_out;
-	}
-	if (ir->type) {
-		ntfs_error(vi->i_sb, "Index type is not 0 (type is 0x%x).",
-				le32_to_cpu(ir->type));
-		goto unm_err_out;
-	}
-	ni->itype.index.collation_rule = ir->collation_rule;
-	ntfs_debug("Index collation rule is 0x%x.",
-			le32_to_cpu(ir->collation_rule));
-	ni->itype.index.block_size = le32_to_cpu(ir->index_block_size);
-	if (!is_power_of_2(ni->itype.index.block_size)) {
-		ntfs_error(vi->i_sb, "Index block size (%u) is not a power of "
-				"two.", ni->itype.index.block_size);
-		goto unm_err_out;
-	}
-	if (ni->itype.index.block_size > PAGE_SIZE) {
-		ntfs_error(vi->i_sb, "Index block size (%u) > PAGE_SIZE "
-				"(%ld) is not supported.  Sorry.",
-				ni->itype.index.block_size, PAGE_SIZE);
-		err = -EOPNOTSUPP;
-		goto unm_err_out;
-	}
-	if (ni->itype.index.block_size < NTFS_BLOCK_SIZE) {
-		ntfs_error(vi->i_sb, "Index block size (%u) < NTFS_BLOCK_SIZE "
-				"(%i) is not supported.  Sorry.",
-				ni->itype.index.block_size, NTFS_BLOCK_SIZE);
-		err = -EOPNOTSUPP;
-		goto unm_err_out;
-	}
-	ni->itype.index.block_size_bits = ffs(ni->itype.index.block_size) - 1;
-	/* Determine the size of a vcn in the index. */
-	if (vol->cluster_size <= ni->itype.index.block_size) {
-		ni->itype.index.vcn_size = vol->cluster_size;
-		ni->itype.index.vcn_size_bits = vol->cluster_size_bits;
-	} else {
-		ni->itype.index.vcn_size = vol->sector_size;
-		ni->itype.index.vcn_size_bits = vol->sector_size_bits;
-	}
-	/* Check for presence of index allocation attribute. */
-	if (!(ir->index.flags & LARGE_INDEX)) {
-		/* No index allocation. */
-		vi->i_size = ni->initialized_size = ni->allocated_size = 0;
-		/* We are done with the mft record, so we release it. */
-		ntfs_attr_put_search_ctx(ctx);
-		unmap_mft_record(base_ni);
-		m = NULL;
-		ctx = NULL;
-		goto skip_large_index_stuff;
-	} /* LARGE_INDEX:  Index allocation present.  Setup state. */
-	NInoSetIndexAllocPresent(ni);
-	/* Find index allocation attribute. */
-	ntfs_attr_reinit_search_ctx(ctx);
-	err = ntfs_attr_lookup(AT_INDEX_ALLOCATION, ni->name, ni->name_len,
-			CASE_SENSITIVE, 0, NULL, 0, ctx);
-	if (unlikely(err)) {
-		if (err == -ENOENT)
-			ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute is "
-					"not present but $INDEX_ROOT "
-					"indicated it is.");
-		else
-			ntfs_error(vi->i_sb, "Failed to lookup "
-					"$INDEX_ALLOCATION attribute.");
-		goto unm_err_out;
-	}
-	a = ctx->attr;
-	if (!a->non_resident) {
-		ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute is "
-				"resident.");
-		goto unm_err_out;
-	}
-	/*
-	 * Ensure the attribute name is placed before the mapping pairs array.
-	 */
-	if (unlikely(a->name_length && (le16_to_cpu(a->name_offset) >=
-			le16_to_cpu(
-			a->data.non_resident.mapping_pairs_offset)))) {
-		ntfs_error(vol->sb, "$INDEX_ALLOCATION attribute name is "
-				"placed after the mapping pairs array.");
-		goto unm_err_out;
-	}
-	if (a->flags & ATTR_IS_ENCRYPTED) {
-		ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute is "
-				"encrypted.");
-		goto unm_err_out;
-	}
-	if (a->flags & ATTR_IS_SPARSE) {
-		ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute is sparse.");
-		goto unm_err_out;
-	}
-	if (a->flags & ATTR_COMPRESSION_MASK) {
-		ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute is "
-				"compressed.");
-		goto unm_err_out;
-	}
-	if (a->data.non_resident.lowest_vcn) {
-		ntfs_error(vi->i_sb, "First extent of $INDEX_ALLOCATION "
-				"attribute has non zero lowest_vcn.");
-		goto unm_err_out;
-	}
-	vi->i_size = sle64_to_cpu(a->data.non_resident.data_size);
-	ni->initialized_size = sle64_to_cpu(
-			a->data.non_resident.initialized_size);
-	ni->allocated_size = sle64_to_cpu(a->data.non_resident.allocated_size);
-	/*
-	 * We are done with the mft record, so we release it.  Otherwise
-	 * we would deadlock in ntfs_attr_iget().
-	 */
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(base_ni);
-	m = NULL;
-	ctx = NULL;
-	/* Get the index bitmap attribute inode. */
-	bvi = ntfs_attr_iget(base_vi, AT_BITMAP, ni->name, ni->name_len);
-	if (IS_ERR(bvi)) {
-		ntfs_error(vi->i_sb, "Failed to get bitmap attribute.");
-		err = PTR_ERR(bvi);
-		goto unm_err_out;
-	}
-	bni = NTFS_I(bvi);
-	if (NInoCompressed(bni) || NInoEncrypted(bni) ||
-			NInoSparse(bni)) {
-		ntfs_error(vi->i_sb, "$BITMAP attribute is compressed and/or "
-				"encrypted and/or sparse.");
-		goto iput_unm_err_out;
-	}
-	/* Consistency check bitmap size vs. index allocation size. */
-	bvi_size = i_size_read(bvi);
-	if ((bvi_size << 3) < (vi->i_size >> ni->itype.index.block_size_bits)) {
-		ntfs_error(vi->i_sb, "Index bitmap too small (0x%llx) for "
-				"index allocation (0x%llx).", bvi_size << 3,
-				vi->i_size);
-		goto iput_unm_err_out;
-	}
-	iput(bvi);
-skip_large_index_stuff:
-	/* Setup the operations for this index inode. */
-	vi->i_mapping->a_ops = &ntfs_mst_aops;
-	vi->i_blocks = ni->allocated_size >> 9;
-	/*
-	 * Make sure the base inode doesn't go away and attach it to the
-	 * index inode.
-	 */
-	igrab(base_vi);
-	ni->ext.base_ntfs_ino = base_ni;
-	ni->nr_extents = -1;
-
-	ntfs_debug("Done.");
-	return 0;
-iput_unm_err_out:
-	iput(bvi);
-unm_err_out:
-	if (!err)
-		err = -EIO;
-	if (ctx)
-		ntfs_attr_put_search_ctx(ctx);
-	if (m)
-		unmap_mft_record(base_ni);
-err_out:
-	ntfs_error(vi->i_sb, "Failed with error code %i while reading index "
-			"inode (mft_no 0x%lx, name_len %i.", err, vi->i_ino,
-			ni->name_len);
-	make_bad_inode(vi);
-	if (err != -EOPNOTSUPP && err != -ENOMEM)
-		NVolSetErrors(vol);
-	return err;
-}
-
-/*
- * The MFT inode has special locking, so teach the lock validator
- * about this by splitting off the locking rules of the MFT from
- * the locking rules of other inodes. The MFT inode can never be
- * accessed from the VFS side (or even internally), only by the
- * map_mft functions.
- */
-static struct lock_class_key mft_ni_runlist_lock_key, mft_ni_mrec_lock_key;
-
-/**
- * ntfs_read_inode_mount - special read_inode for mount time use only
- * @vi:		inode to read
- *
- * Read inode FILE_MFT at mount time, only called with super_block lock
- * held from within the read_super() code path.
- *
- * This function exists because when it is called the page cache for $MFT/$DATA
- * is not initialized and hence we cannot get at the contents of mft records
- * by calling map_mft_record*().
- *
- * Further it needs to cope with the circular references problem, i.e. cannot
- * load any attributes other than $ATTRIBUTE_LIST until $DATA is loaded, because
- * we do not know where the other extent mft records are yet and again, because
- * we cannot call map_mft_record*() yet.  Obviously this applies only when an
- * attribute list is actually present in $MFT inode.
- *
- * We solve these problems by starting with the $DATA attribute before anything
- * else and iterating using ntfs_attr_lookup($DATA) over all extents.  As each
- * extent is found, we ntfs_mapping_pairs_decompress() including the implied
- * ntfs_runlists_merge().  Each step of the iteration necessarily provides
- * sufficient information for the next step to complete.
- *
- * This should work but there are two possible pit falls (see inline comments
- * below), but only time will tell if they are real pits or just smoke...
- */
-int ntfs_read_inode_mount(struct inode *vi)
-{
-	VCN next_vcn, last_vcn, highest_vcn;
-	s64 block;
-	struct super_block *sb = vi->i_sb;
-	ntfs_volume *vol = NTFS_SB(sb);
-	struct buffer_head *bh;
-	ntfs_inode *ni;
-	MFT_RECORD *m = NULL;
-	ATTR_RECORD *a;
-	ntfs_attr_search_ctx *ctx;
-	unsigned int i, nr_blocks;
-	int err;
-
-	ntfs_debug("Entering.");
-
-	/* Initialize the ntfs specific part of @vi. */
-	ntfs_init_big_inode(vi);
-
-	ni = NTFS_I(vi);
-
-	/* Setup the data attribute. It is special as it is mst protected. */
-	NInoSetNonResident(ni);
-	NInoSetMstProtected(ni);
-	NInoSetSparseDisabled(ni);
-	ni->type = AT_DATA;
-	ni->name = NULL;
-	ni->name_len = 0;
-	/*
-	 * This sets up our little cheat allowing us to reuse the async read io
-	 * completion handler for directories.
-	 */
-	ni->itype.index.block_size = vol->mft_record_size;
-	ni->itype.index.block_size_bits = vol->mft_record_size_bits;
-
-	/* Very important! Needed to be able to call map_mft_record*(). */
-	vol->mft_ino = vi;
-
-	/* Allocate enough memory to read the first mft record. */
-	if (vol->mft_record_size > 64 * 1024) {
-		ntfs_error(sb, "Unsupported mft record size %i (max 64kiB).",
-				vol->mft_record_size);
-		goto err_out;
-	}
-	i = vol->mft_record_size;
-	if (i < sb->s_blocksize)
-		i = sb->s_blocksize;
-	m = (MFT_RECORD*)ntfs_malloc_nofs(i);
-	if (!m) {
-		ntfs_error(sb, "Failed to allocate buffer for $MFT record 0.");
-		goto err_out;
-	}
-
-	/* Determine the first block of the $MFT/$DATA attribute. */
-	block = vol->mft_lcn << vol->cluster_size_bits >>
-			sb->s_blocksize_bits;
-	nr_blocks = vol->mft_record_size >> sb->s_blocksize_bits;
-	if (!nr_blocks)
-		nr_blocks = 1;
-
-	/* Load $MFT/$DATA's first mft record. */
-	for (i = 0; i < nr_blocks; i++) {
-		bh = sb_bread(sb, block++);
-		if (!bh) {
-			ntfs_error(sb, "Device read failed.");
-			goto err_out;
-		}
-		memcpy((char*)m + (i << sb->s_blocksize_bits), bh->b_data,
-				sb->s_blocksize);
-		brelse(bh);
-	}
-
-	if (le32_to_cpu(m->bytes_allocated) != vol->mft_record_size) {
-		ntfs_error(sb, "Incorrect mft record size %u in superblock, should be %u.",
-				le32_to_cpu(m->bytes_allocated), vol->mft_record_size);
-		goto err_out;
-	}
-
-	/* Apply the mst fixups. */
-	if (post_read_mst_fixup((NTFS_RECORD*)m, vol->mft_record_size)) {
-		/* FIXME: Try to use the $MFTMirr now. */
-		ntfs_error(sb, "MST fixup failed. $MFT is corrupt.");
-		goto err_out;
-	}
-
-	/* Sanity check offset to the first attribute */
-	if (le16_to_cpu(m->attrs_offset) >= le32_to_cpu(m->bytes_allocated)) {
-		ntfs_error(sb, "Incorrect mft offset to the first attribute %u in superblock.",
-			       le16_to_cpu(m->attrs_offset));
-		goto err_out;
-	}
-
-	/* Need this to sanity check attribute list references to $MFT. */
-	vi->i_generation = ni->seq_no = le16_to_cpu(m->sequence_number);
-
-	/* Provides read_folio() for map_mft_record(). */
-	vi->i_mapping->a_ops = &ntfs_mst_aops;
-
-	ctx = ntfs_attr_get_search_ctx(ni, m);
-	if (!ctx) {
-		err = -ENOMEM;
-		goto err_out;
-	}
-
-	/* Find the attribute list attribute if present. */
-	err = ntfs_attr_lookup(AT_ATTRIBUTE_LIST, NULL, 0, 0, 0, NULL, 0, ctx);
-	if (err) {
-		if (unlikely(err != -ENOENT)) {
-			ntfs_error(sb, "Failed to lookup attribute list "
-					"attribute. You should run chkdsk.");
-			goto put_err_out;
-		}
-	} else /* if (!err) */ {
-		ATTR_LIST_ENTRY *al_entry, *next_al_entry;
-		u8 *al_end;
-		static const char *es = "  Not allowed.  $MFT is corrupt.  "
-				"You should run chkdsk.";
-
-		ntfs_debug("Attribute list attribute found in $MFT.");
-		NInoSetAttrList(ni);
-		a = ctx->attr;
-		if (a->flags & ATTR_COMPRESSION_MASK) {
-			ntfs_error(sb, "Attribute list attribute is "
-					"compressed.%s", es);
-			goto put_err_out;
-		}
-		if (a->flags & ATTR_IS_ENCRYPTED ||
-				a->flags & ATTR_IS_SPARSE) {
-			if (a->non_resident) {
-				ntfs_error(sb, "Non-resident attribute list "
-						"attribute is encrypted/"
-						"sparse.%s", es);
-				goto put_err_out;
-			}
-			ntfs_warning(sb, "Resident attribute list attribute "
-					"in $MFT system file is marked "
-					"encrypted/sparse which is not true.  "
-					"However, Windows allows this and "
-					"chkdsk does not detect or correct it "
-					"so we will just ignore the invalid "
-					"flags and pretend they are not set.");
-		}
-		/* Now allocate memory for the attribute list. */
-		ni->attr_list_size = (u32)ntfs_attr_size(a);
-		if (!ni->attr_list_size) {
-			ntfs_error(sb, "Attr_list_size is zero");
-			goto put_err_out;
-		}
-		ni->attr_list = ntfs_malloc_nofs(ni->attr_list_size);
-		if (!ni->attr_list) {
-			ntfs_error(sb, "Not enough memory to allocate buffer "
-					"for attribute list.");
-			goto put_err_out;
-		}
-		if (a->non_resident) {
-			NInoSetAttrListNonResident(ni);
-			if (a->data.non_resident.lowest_vcn) {
-				ntfs_error(sb, "Attribute list has non zero "
-						"lowest_vcn. $MFT is corrupt. "
-						"You should run chkdsk.");
-				goto put_err_out;
-			}
-			/* Setup the runlist. */
-			ni->attr_list_rl.rl = ntfs_mapping_pairs_decompress(vol,
-					a, NULL);
-			if (IS_ERR(ni->attr_list_rl.rl)) {
-				err = PTR_ERR(ni->attr_list_rl.rl);
-				ni->attr_list_rl.rl = NULL;
-				ntfs_error(sb, "Mapping pairs decompression "
-						"failed with error code %i.",
-						-err);
-				goto put_err_out;
-			}
-			/* Now load the attribute list. */
-			if ((err = load_attribute_list(vol, &ni->attr_list_rl,
-					ni->attr_list, ni->attr_list_size,
-					sle64_to_cpu(a->data.
-					non_resident.initialized_size)))) {
-				ntfs_error(sb, "Failed to load attribute list "
-						"attribute with error code %i.",
-						-err);
-				goto put_err_out;
-			}
-		} else /* if (!ctx.attr->non_resident) */ {
-			if ((u8*)a + le16_to_cpu(
-					a->data.resident.value_offset) +
-					le32_to_cpu(
-					a->data.resident.value_length) >
-					(u8*)ctx->mrec + vol->mft_record_size) {
-				ntfs_error(sb, "Corrupt attribute list "
-						"attribute.");
-				goto put_err_out;
-			}
-			/* Now copy the attribute list. */
-			memcpy(ni->attr_list, (u8*)a + le16_to_cpu(
-					a->data.resident.value_offset),
-					le32_to_cpu(
-					a->data.resident.value_length));
-		}
-		/* The attribute list is now setup in memory. */
-		/*
-		 * FIXME: I don't know if this case is actually possible.
-		 * According to logic it is not possible but I have seen too
-		 * many weird things in MS software to rely on logic... Thus we
-		 * perform a manual search and make sure the first $MFT/$DATA
-		 * extent is in the base inode. If it is not we abort with an
-		 * error and if we ever see a report of this error we will need
-		 * to do some magic in order to have the necessary mft record
-		 * loaded and in the right place in the page cache. But
-		 * hopefully logic will prevail and this never happens...
-		 */
-		al_entry = (ATTR_LIST_ENTRY*)ni->attr_list;
-		al_end = (u8*)al_entry + ni->attr_list_size;
-		for (;; al_entry = next_al_entry) {
-			/* Out of bounds check. */
-			if ((u8*)al_entry < ni->attr_list ||
-					(u8*)al_entry > al_end)
-				goto em_put_err_out;
-			/* Catch the end of the attribute list. */
-			if ((u8*)al_entry == al_end)
-				goto em_put_err_out;
-			if (!al_entry->length)
-				goto em_put_err_out;
-			if ((u8*)al_entry + 6 > al_end || (u8*)al_entry +
-					le16_to_cpu(al_entry->length) > al_end)
-				goto em_put_err_out;
-			next_al_entry = (ATTR_LIST_ENTRY*)((u8*)al_entry +
-					le16_to_cpu(al_entry->length));
-			if (le32_to_cpu(al_entry->type) > le32_to_cpu(AT_DATA))
-				goto em_put_err_out;
-			if (AT_DATA != al_entry->type)
-				continue;
-			/* We want an unnamed attribute. */
-			if (al_entry->name_length)
-				goto em_put_err_out;
-			/* Want the first entry, i.e. lowest_vcn == 0. */
-			if (al_entry->lowest_vcn)
-				goto em_put_err_out;
-			/* First entry has to be in the base mft record. */
-			if (MREF_LE(al_entry->mft_reference) != vi->i_ino) {
-				/* MFT references do not match, logic fails. */
-				ntfs_error(sb, "BUG: The first $DATA extent "
-						"of $MFT is not in the base "
-						"mft record. Please report "
-						"you saw this message to "
-						"linux-ntfs-dev@lists."
-						"sourceforge.net");
-				goto put_err_out;
-			} else {
-				/* Sequence numbers must match. */
-				if (MSEQNO_LE(al_entry->mft_reference) !=
-						ni->seq_no)
-					goto em_put_err_out;
-				/* Got it. All is ok. We can stop now. */
-				break;
-			}
-		}
-	}
-
-	ntfs_attr_reinit_search_ctx(ctx);
-
-	/* Now load all attribute extents. */
-	a = NULL;
-	next_vcn = last_vcn = highest_vcn = 0;
-	while (!(err = ntfs_attr_lookup(AT_DATA, NULL, 0, 0, next_vcn, NULL, 0,
-			ctx))) {
-		runlist_element *nrl;
-
-		/* Cache the current attribute. */
-		a = ctx->attr;
-		/* $MFT must be non-resident. */
-		if (!a->non_resident) {
-			ntfs_error(sb, "$MFT must be non-resident but a "
-					"resident extent was found. $MFT is "
-					"corrupt. Run chkdsk.");
-			goto put_err_out;
-		}
-		/* $MFT must be uncompressed and unencrypted. */
-		if (a->flags & ATTR_COMPRESSION_MASK ||
-				a->flags & ATTR_IS_ENCRYPTED ||
-				a->flags & ATTR_IS_SPARSE) {
-			ntfs_error(sb, "$MFT must be uncompressed, "
-					"non-sparse, and unencrypted but a "
-					"compressed/sparse/encrypted extent "
-					"was found. $MFT is corrupt. Run "
-					"chkdsk.");
-			goto put_err_out;
-		}
-		/*
-		 * Decompress the mapping pairs array of this extent and merge
-		 * the result into the existing runlist. No need for locking
-		 * as we have exclusive access to the inode at this time and we
-		 * are a mount in progress task, too.
-		 */
-		nrl = ntfs_mapping_pairs_decompress(vol, a, ni->runlist.rl);
-		if (IS_ERR(nrl)) {
-			ntfs_error(sb, "ntfs_mapping_pairs_decompress() "
-					"failed with error code %ld.  $MFT is "
-					"corrupt.", PTR_ERR(nrl));
-			goto put_err_out;
-		}
-		ni->runlist.rl = nrl;
-
-		/* Are we in the first extent? */
-		if (!next_vcn) {
-			if (a->data.non_resident.lowest_vcn) {
-				ntfs_error(sb, "First extent of $DATA "
-						"attribute has non zero "
-						"lowest_vcn. $MFT is corrupt. "
-						"You should run chkdsk.");
-				goto put_err_out;
-			}
-			/* Get the last vcn in the $DATA attribute. */
-			last_vcn = sle64_to_cpu(
-					a->data.non_resident.allocated_size)
-					>> vol->cluster_size_bits;
-			/* Fill in the inode size. */
-			vi->i_size = sle64_to_cpu(
-					a->data.non_resident.data_size);
-			ni->initialized_size = sle64_to_cpu(
-					a->data.non_resident.initialized_size);
-			ni->allocated_size = sle64_to_cpu(
-					a->data.non_resident.allocated_size);
-			/*
-			 * Verify the number of mft records does not exceed
-			 * 2^32 - 1.
-			 */
-			if ((vi->i_size >> vol->mft_record_size_bits) >=
-					(1ULL << 32)) {
-				ntfs_error(sb, "$MFT is too big! Aborting.");
-				goto put_err_out;
-			}
-			/*
-			 * We have got the first extent of the runlist for
-			 * $MFT which means it is now relatively safe to call
-			 * the normal ntfs_read_inode() function.
-			 * Complete reading the inode, this will actually
-			 * re-read the mft record for $MFT, this time entering
-			 * it into the page cache with which we complete the
-			 * kick start of the volume. It should be safe to do
-			 * this now as the first extent of $MFT/$DATA is
-			 * already known and we would hope that we don't need
-			 * further extents in order to find the other
-			 * attributes belonging to $MFT. Only time will tell if
-			 * this is really the case. If not we will have to play
-			 * magic at this point, possibly duplicating a lot of
-			 * ntfs_read_inode() at this point. We will need to
-			 * ensure we do enough of its work to be able to call
-			 * ntfs_read_inode() on extents of $MFT/$DATA. But lets
-			 * hope this never happens...
-			 */
-			ntfs_read_locked_inode(vi);
-			if (is_bad_inode(vi)) {
-				ntfs_error(sb, "ntfs_read_inode() of $MFT "
-						"failed. BUG or corrupt $MFT. "
-						"Run chkdsk and if no errors "
-						"are found, please report you "
-						"saw this message to "
-						"linux-ntfs-dev@lists."
-						"sourceforge.net");
-				ntfs_attr_put_search_ctx(ctx);
-				/* Revert to the safe super operations. */
-				ntfs_free(m);
-				return -1;
-			}
-			/*
-			 * Re-initialize some specifics about $MFT's inode as
-			 * ntfs_read_inode() will have set up the default ones.
-			 */
-			/* Set uid and gid to root. */
-			vi->i_uid = GLOBAL_ROOT_UID;
-			vi->i_gid = GLOBAL_ROOT_GID;
-			/* Regular file. No access for anyone. */
-			vi->i_mode = S_IFREG;
-			/* No VFS initiated operations allowed for $MFT. */
-			vi->i_op = &ntfs_empty_inode_ops;
-			vi->i_fop = &ntfs_empty_file_ops;
-		}
-
-		/* Get the lowest vcn for the next extent. */
-		highest_vcn = sle64_to_cpu(a->data.non_resident.highest_vcn);
-		next_vcn = highest_vcn + 1;
-
-		/* Only one extent or error, which we catch below. */
-		if (next_vcn <= 0)
-			break;
-
-		/* Avoid endless loops due to corruption. */
-		if (next_vcn < sle64_to_cpu(
-				a->data.non_resident.lowest_vcn)) {
-			ntfs_error(sb, "$MFT has corrupt attribute list "
-					"attribute. Run chkdsk.");
-			goto put_err_out;
-		}
-	}
-	if (err != -ENOENT) {
-		ntfs_error(sb, "Failed to lookup $MFT/$DATA attribute extent. "
-				"$MFT is corrupt. Run chkdsk.");
-		goto put_err_out;
-	}
-	if (!a) {
-		ntfs_error(sb, "$MFT/$DATA attribute not found. $MFT is "
-				"corrupt. Run chkdsk.");
-		goto put_err_out;
-	}
-	if (highest_vcn && highest_vcn != last_vcn - 1) {
-		ntfs_error(sb, "Failed to load the complete runlist for "
-				"$MFT/$DATA. Driver bug or corrupt $MFT. "
-				"Run chkdsk.");
-		ntfs_debug("highest_vcn = 0x%llx, last_vcn - 1 = 0x%llx",
-				(unsigned long long)highest_vcn,
-				(unsigned long long)last_vcn - 1);
-		goto put_err_out;
-	}
-	ntfs_attr_put_search_ctx(ctx);
-	ntfs_debug("Done.");
-	ntfs_free(m);
-
-	/*
-	 * Split the locking rules of the MFT inode from the
-	 * locking rules of other inodes:
-	 */
-	lockdep_set_class(&ni->runlist.lock, &mft_ni_runlist_lock_key);
-	lockdep_set_class(&ni->mrec_lock, &mft_ni_mrec_lock_key);
-
-	return 0;
-
-em_put_err_out:
-	ntfs_error(sb, "Couldn't find first extent of $DATA attribute in "
-			"attribute list. $MFT is corrupt. Run chkdsk.");
-put_err_out:
-	ntfs_attr_put_search_ctx(ctx);
-err_out:
-	ntfs_error(sb, "Failed. Marking inode as bad.");
-	make_bad_inode(vi);
-	ntfs_free(m);
-	return -1;
-}
-
-static void __ntfs_clear_inode(ntfs_inode *ni)
-{
-	/* Free all alocated memory. */
-	down_write(&ni->runlist.lock);
-	if (ni->runlist.rl) {
-		ntfs_free(ni->runlist.rl);
-		ni->runlist.rl = NULL;
-	}
-	up_write(&ni->runlist.lock);
-
-	if (ni->attr_list) {
-		ntfs_free(ni->attr_list);
-		ni->attr_list = NULL;
-	}
-
-	down_write(&ni->attr_list_rl.lock);
-	if (ni->attr_list_rl.rl) {
-		ntfs_free(ni->attr_list_rl.rl);
-		ni->attr_list_rl.rl = NULL;
-	}
-	up_write(&ni->attr_list_rl.lock);
-
-	if (ni->name_len && ni->name != I30) {
-		/* Catch bugs... */
-		BUG_ON(!ni->name);
-		kfree(ni->name);
-	}
-}
-
-void ntfs_clear_extent_inode(ntfs_inode *ni)
-{
-	ntfs_debug("Entering for inode 0x%lx.", ni->mft_no);
-
-	BUG_ON(NInoAttr(ni));
-	BUG_ON(ni->nr_extents != -1);
-
-#ifdef NTFS_RW
-	if (NInoDirty(ni)) {
-		if (!is_bad_inode(VFS_I(ni->ext.base_ntfs_ino)))
-			ntfs_error(ni->vol->sb, "Clearing dirty extent inode!  "
-					"Losing data!  This is a BUG!!!");
-		// FIXME:  Do something!!!
-	}
-#endif /* NTFS_RW */
-
-	__ntfs_clear_inode(ni);
-
-	/* Bye, bye... */
-	ntfs_destroy_extent_inode(ni);
-}
-
-/**
- * ntfs_evict_big_inode - clean up the ntfs specific part of an inode
- * @vi:		vfs inode pending annihilation
- *
- * When the VFS is going to remove an inode from memory, ntfs_clear_big_inode()
- * is called, which deallocates all memory belonging to the NTFS specific part
- * of the inode and returns.
- *
- * If the MFT record is dirty, we commit it before doing anything else.
- */
-void ntfs_evict_big_inode(struct inode *vi)
-{
-	ntfs_inode *ni = NTFS_I(vi);
-
-	truncate_inode_pages_final(&vi->i_data);
-	clear_inode(vi);
-
-#ifdef NTFS_RW
-	if (NInoDirty(ni)) {
-		bool was_bad = (is_bad_inode(vi));
-
-		/* Committing the inode also commits all extent inodes. */
-		ntfs_commit_inode(vi);
-
-		if (!was_bad && (is_bad_inode(vi) || NInoDirty(ni))) {
-			ntfs_error(vi->i_sb, "Failed to commit dirty inode "
-					"0x%lx.  Losing data!", vi->i_ino);
-			// FIXME:  Do something!!!
-		}
-	}
-#endif /* NTFS_RW */
-
-	/* No need to lock at this stage as no one else has a reference. */
-	if (ni->nr_extents > 0) {
-		int i;
-
-		for (i = 0; i < ni->nr_extents; i++)
-			ntfs_clear_extent_inode(ni->ext.extent_ntfs_inos[i]);
-		kfree(ni->ext.extent_ntfs_inos);
-	}
-
-	__ntfs_clear_inode(ni);
-
-	if (NInoAttr(ni)) {
-		/* Release the base inode if we are holding it. */
-		if (ni->nr_extents == -1) {
-			iput(VFS_I(ni->ext.base_ntfs_ino));
-			ni->nr_extents = 0;
-			ni->ext.base_ntfs_ino = NULL;
-		}
-	}
-	BUG_ON(ni->page);
-	if (!atomic_dec_and_test(&ni->count))
-		BUG();
-	return;
-}
-
-/**
- * ntfs_show_options - show mount options in /proc/mounts
- * @sf:		seq_file in which to write our mount options
- * @root:	root of the mounted tree whose mount options to display
- *
- * Called by the VFS once for each mounted ntfs volume when someone reads
- * /proc/mounts in order to display the NTFS specific mount options of each
- * mount. The mount options of fs specified by @root are written to the seq file
- * @sf and success is returned.
- */
-int ntfs_show_options(struct seq_file *sf, struct dentry *root)
-{
-	ntfs_volume *vol = NTFS_SB(root->d_sb);
-	int i;
-
-	seq_printf(sf, ",uid=%i", from_kuid_munged(&init_user_ns, vol->uid));
-	seq_printf(sf, ",gid=%i", from_kgid_munged(&init_user_ns, vol->gid));
-	if (vol->fmask == vol->dmask)
-		seq_printf(sf, ",umask=0%o", vol->fmask);
-	else {
-		seq_printf(sf, ",fmask=0%o", vol->fmask);
-		seq_printf(sf, ",dmask=0%o", vol->dmask);
-	}
-	seq_printf(sf, ",nls=%s", vol->nls_map->charset);
-	if (NVolCaseSensitive(vol))
-		seq_printf(sf, ",case_sensitive");
-	if (NVolShowSystemFiles(vol))
-		seq_printf(sf, ",show_sys_files");
-	if (!NVolSparseEnabled(vol))
-		seq_printf(sf, ",disable_sparse");
-	for (i = 0; on_errors_arr[i].val; i++) {
-		if (on_errors_arr[i].val & vol->on_errors)
-			seq_printf(sf, ",errors=%s", on_errors_arr[i].str);
-	}
-	seq_printf(sf, ",mft_zone_multiplier=%i", vol->mft_zone_multiplier);
-	return 0;
-}
-
-#ifdef NTFS_RW
-
-static const char *es = "  Leaving inconsistent metadata.  Unmount and run "
-		"chkdsk.";
-
-/**
- * ntfs_truncate - called when the i_size of an ntfs inode is changed
- * @vi:		inode for which the i_size was changed
- *
- * We only support i_size changes for normal files at present, i.e. not
- * compressed and not encrypted.  This is enforced in ntfs_setattr(), see
- * below.
- *
- * The kernel guarantees that @vi is a regular file (S_ISREG() is true) and
- * that the change is allowed.
- *
- * This implies for us that @vi is a file inode rather than a directory, index,
- * or attribute inode as well as that @vi is a base inode.
- *
- * Returns 0 on success or -errno on error.
- *
- * Called with ->i_mutex held.
- */
-int ntfs_truncate(struct inode *vi)
-{
-	s64 new_size, old_size, nr_freed, new_alloc_size, old_alloc_size;
-	VCN highest_vcn;
-	unsigned long flags;
-	ntfs_inode *base_ni, *ni = NTFS_I(vi);
-	ntfs_volume *vol = ni->vol;
-	ntfs_attr_search_ctx *ctx;
-	MFT_RECORD *m;
-	ATTR_RECORD *a;
-	const char *te = "  Leaving file length out of sync with i_size.";
-	int err, mp_size, size_change, alloc_change;
-
-	ntfs_debug("Entering for inode 0x%lx.", vi->i_ino);
-	BUG_ON(NInoAttr(ni));
-	BUG_ON(S_ISDIR(vi->i_mode));
-	BUG_ON(NInoMstProtected(ni));
-	BUG_ON(ni->nr_extents < 0);
-retry_truncate:
-	/*
-	 * Lock the runlist for writing and map the mft record to ensure it is
-	 * safe to mess with the attribute runlist and sizes.
-	 */
-	down_write(&ni->runlist.lock);
-	if (!NInoAttr(ni))
-		base_ni = ni;
-	else
-		base_ni = ni->ext.base_ntfs_ino;
-	m = map_mft_record(base_ni);
-	if (IS_ERR(m)) {
-		err = PTR_ERR(m);
-		ntfs_error(vi->i_sb, "Failed to map mft record for inode 0x%lx "
-				"(error code %d).%s", vi->i_ino, err, te);
-		ctx = NULL;
-		m = NULL;
-		goto old_bad_out;
-	}
-	ctx = ntfs_attr_get_search_ctx(base_ni, m);
-	if (unlikely(!ctx)) {
-		ntfs_error(vi->i_sb, "Failed to allocate a search context for "
-				"inode 0x%lx (not enough memory).%s",
-				vi->i_ino, te);
-		err = -ENOMEM;
-		goto old_bad_out;
-	}
-	err = ntfs_attr_lookup(ni->type, ni->name, ni->name_len,
-			CASE_SENSITIVE, 0, NULL, 0, ctx);
-	if (unlikely(err)) {
-		if (err == -ENOENT) {
-			ntfs_error(vi->i_sb, "Open attribute is missing from "
-					"mft record.  Inode 0x%lx is corrupt.  "
-					"Run chkdsk.%s", vi->i_ino, te);
-			err = -EIO;
-		} else
-			ntfs_error(vi->i_sb, "Failed to lookup attribute in "
-					"inode 0x%lx (error code %d).%s",
-					vi->i_ino, err, te);
-		goto old_bad_out;
-	}
-	m = ctx->mrec;
-	a = ctx->attr;
-	/*
-	 * The i_size of the vfs inode is the new size for the attribute value.
-	 */
-	new_size = i_size_read(vi);
-	/* The current size of the attribute value is the old size. */
-	old_size = ntfs_attr_size(a);
-	/* Calculate the new allocated size. */
-	if (NInoNonResident(ni))
-		new_alloc_size = (new_size + vol->cluster_size - 1) &
-				~(s64)vol->cluster_size_mask;
-	else
-		new_alloc_size = (new_size + 7) & ~7;
-	/* The current allocated size is the old allocated size. */
-	read_lock_irqsave(&ni->size_lock, flags);
-	old_alloc_size = ni->allocated_size;
-	read_unlock_irqrestore(&ni->size_lock, flags);
-	/*
-	 * The change in the file size.  This will be 0 if no change, >0 if the
-	 * size is growing, and <0 if the size is shrinking.
-	 */
-	size_change = -1;
-	if (new_size - old_size >= 0) {
-		size_change = 1;
-		if (new_size == old_size)
-			size_change = 0;
-	}
-	/* As above for the allocated size. */
-	alloc_change = -1;
-	if (new_alloc_size - old_alloc_size >= 0) {
-		alloc_change = 1;
-		if (new_alloc_size == old_alloc_size)
-			alloc_change = 0;
-	}
-	/*
-	 * If neither the size nor the allocation are being changed there is
-	 * nothing to do.
-	 */
-	if (!size_change && !alloc_change)
-		goto unm_done;
-	/* If the size is changing, check if new size is allowed in $AttrDef. */
-	if (size_change) {
-		err = ntfs_attr_size_bounds_check(vol, ni->type, new_size);
-		if (unlikely(err)) {
-			if (err == -ERANGE) {
-				ntfs_error(vol->sb, "Truncate would cause the "
-						"inode 0x%lx to %simum size "
-						"for its attribute type "
-						"(0x%x).  Aborting truncate.",
-						vi->i_ino,
-						new_size > old_size ? "exceed "
-						"the max" : "go under the min",
-						le32_to_cpu(ni->type));
-				err = -EFBIG;
-			} else {
-				ntfs_error(vol->sb, "Inode 0x%lx has unknown "
-						"attribute type 0x%x.  "
-						"Aborting truncate.",
-						vi->i_ino,
-						le32_to_cpu(ni->type));
-				err = -EIO;
-			}
-			/* Reset the vfs inode size to the old size. */
-			i_size_write(vi, old_size);
-			goto err_out;
-		}
-	}
-	if (NInoCompressed(ni) || NInoEncrypted(ni)) {
-		ntfs_warning(vi->i_sb, "Changes in inode size are not "
-				"supported yet for %s files, ignoring.",
-				NInoCompressed(ni) ? "compressed" :
-				"encrypted");
-		err = -EOPNOTSUPP;
-		goto bad_out;
-	}
-	if (a->non_resident)
-		goto do_non_resident_truncate;
-	BUG_ON(NInoNonResident(ni));
-	/* Resize the attribute record to best fit the new attribute size. */
-	if (new_size < vol->mft_record_size &&
-			!ntfs_resident_attr_value_resize(m, a, new_size)) {
-		/* The resize succeeded! */
-		flush_dcache_mft_record_page(ctx->ntfs_ino);
-		mark_mft_record_dirty(ctx->ntfs_ino);
-		write_lock_irqsave(&ni->size_lock, flags);
-		/* Update the sizes in the ntfs inode and all is done. */
-		ni->allocated_size = le32_to_cpu(a->length) -
-				le16_to_cpu(a->data.resident.value_offset);
-		/*
-		 * Note ntfs_resident_attr_value_resize() has already done any
-		 * necessary data clearing in the attribute record.  When the
-		 * file is being shrunk vmtruncate() will already have cleared
-		 * the top part of the last partial page, i.e. since this is
-		 * the resident case this is the page with index 0.  However,
-		 * when the file is being expanded, the page cache page data
-		 * between the old data_size, i.e. old_size, and the new_size
-		 * has not been zeroed.  Fortunately, we do not need to zero it
-		 * either since on one hand it will either already be zero due
-		 * to both read_folio and writepage clearing partial page data
-		 * beyond i_size in which case there is nothing to do or in the
-		 * case of the file being mmap()ped at the same time, POSIX
-		 * specifies that the behaviour is unspecified thus we do not
-		 * have to do anything.  This means that in our implementation
-		 * in the rare case that the file is mmap()ped and a write
-		 * occurred into the mmap()ped region just beyond the file size
-		 * and writepage has not yet been called to write out the page
-		 * (which would clear the area beyond the file size) and we now
-		 * extend the file size to incorporate this dirty region
-		 * outside the file size, a write of the page would result in
-		 * this data being written to disk instead of being cleared.
-		 * Given both POSIX and the Linux mmap(2) man page specify that
-		 * this corner case is undefined, we choose to leave it like
-		 * that as this is much simpler for us as we cannot lock the
-		 * relevant page now since we are holding too many ntfs locks
-		 * which would result in a lock reversal deadlock.
-		 */
-		ni->initialized_size = new_size;
-		write_unlock_irqrestore(&ni->size_lock, flags);
-		goto unm_done;
-	}
-	/* If the above resize failed, this must be an attribute extension. */
-	BUG_ON(size_change < 0);
-	/*
-	 * We have to drop all the locks so we can call
-	 * ntfs_attr_make_non_resident().  This could be optimised by try-
-	 * locking the first page cache page and only if that fails dropping
-	 * the locks, locking the page, and redoing all the locking and
-	 * lookups.  While this would be a huge optimisation, it is not worth
-	 * it as this is definitely a slow code path as it only ever can happen
-	 * once for any given file.
-	 */
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(base_ni);
-	up_write(&ni->runlist.lock);
-	/*
-	 * Not enough space in the mft record, try to make the attribute
-	 * non-resident and if successful restart the truncation process.
-	 */
-	err = ntfs_attr_make_non_resident(ni, old_size);
-	if (likely(!err))
-		goto retry_truncate;
-	/*
-	 * Could not make non-resident.  If this is due to this not being
-	 * permitted for this attribute type or there not being enough space,
-	 * try to make other attributes non-resident.  Otherwise fail.
-	 */
-	if (unlikely(err != -EPERM && err != -ENOSPC)) {
-		ntfs_error(vol->sb, "Cannot truncate inode 0x%lx, attribute "
-				"type 0x%x, because the conversion from "
-				"resident to non-resident attribute failed "
-				"with error code %i.", vi->i_ino,
-				(unsigned)le32_to_cpu(ni->type), err);
-		if (err != -ENOMEM)
-			err = -EIO;
-		goto conv_err_out;
-	}
-	/* TODO: Not implemented from here, abort. */
-	if (err == -ENOSPC)
-		ntfs_error(vol->sb, "Not enough space in the mft record/on "
-				"disk for the non-resident attribute value.  "
-				"This case is not implemented yet.");
-	else /* if (err == -EPERM) */
-		ntfs_error(vol->sb, "This attribute type may not be "
-				"non-resident.  This case is not implemented "
-				"yet.");
-	err = -EOPNOTSUPP;
-	goto conv_err_out;
-#if 0
-	// TODO: Attempt to make other attributes non-resident.
-	if (!err)
-		goto do_resident_extend;
-	/*
-	 * Both the attribute list attribute and the standard information
-	 * attribute must remain in the base inode.  Thus, if this is one of
-	 * these attributes, we have to try to move other attributes out into
-	 * extent mft records instead.
-	 */
-	if (ni->type == AT_ATTRIBUTE_LIST ||
-			ni->type == AT_STANDARD_INFORMATION) {
-		// TODO: Attempt to move other attributes into extent mft
-		// records.
-		err = -EOPNOTSUPP;
-		if (!err)
-			goto do_resident_extend;
-		goto err_out;
-	}
-	// TODO: Attempt to move this attribute to an extent mft record, but
-	// only if it is not already the only attribute in an mft record in
-	// which case there would be nothing to gain.
-	err = -EOPNOTSUPP;
-	if (!err)
-		goto do_resident_extend;
-	/* There is nothing we can do to make enough space. )-: */
-	goto err_out;
-#endif
-do_non_resident_truncate:
-	BUG_ON(!NInoNonResident(ni));
-	if (alloc_change < 0) {
-		highest_vcn = sle64_to_cpu(a->data.non_resident.highest_vcn);
-		if (highest_vcn > 0 &&
-				old_alloc_size >> vol->cluster_size_bits >
-				highest_vcn + 1) {
-			/*
-			 * This attribute has multiple extents.  Not yet
-			 * supported.
-			 */
-			ntfs_error(vol->sb, "Cannot truncate inode 0x%lx, "
-					"attribute type 0x%x, because the "
-					"attribute is highly fragmented (it "
-					"consists of multiple extents) and "
-					"this case is not implemented yet.",
-					vi->i_ino,
-					(unsigned)le32_to_cpu(ni->type));
-			err = -EOPNOTSUPP;
-			goto bad_out;
-		}
-	}
-	/*
-	 * If the size is shrinking, need to reduce the initialized_size and
-	 * the data_size before reducing the allocation.
-	 */
-	if (size_change < 0) {
-		/*
-		 * Make the valid size smaller (i_size is already up-to-date).
-		 */
-		write_lock_irqsave(&ni->size_lock, flags);
-		if (new_size < ni->initialized_size) {
-			ni->initialized_size = new_size;
-			a->data.non_resident.initialized_size =
-					cpu_to_sle64(new_size);
-		}
-		a->data.non_resident.data_size = cpu_to_sle64(new_size);
-		write_unlock_irqrestore(&ni->size_lock, flags);
-		flush_dcache_mft_record_page(ctx->ntfs_ino);
-		mark_mft_record_dirty(ctx->ntfs_ino);
-		/* If the allocated size is not changing, we are done. */
-		if (!alloc_change)
-			goto unm_done;
-		/*
-		 * If the size is shrinking it makes no sense for the
-		 * allocation to be growing.
-		 */
-		BUG_ON(alloc_change > 0);
-	} else /* if (size_change >= 0) */ {
-		/*
-		 * The file size is growing or staying the same but the
-		 * allocation can be shrinking, growing or staying the same.
-		 */
-		if (alloc_change > 0) {
-			/*
-			 * We need to extend the allocation and possibly update
-			 * the data size.  If we are updating the data size,
-			 * since we are not touching the initialized_size we do
-			 * not need to worry about the actual data on disk.
-			 * And as far as the page cache is concerned, there
-			 * will be no pages beyond the old data size and any
-			 * partial region in the last page between the old and
-			 * new data size (or the end of the page if the new
-			 * data size is outside the page) does not need to be
-			 * modified as explained above for the resident
-			 * attribute truncate case.  To do this, we simply drop
-			 * the locks we hold and leave all the work to our
-			 * friendly helper ntfs_attr_extend_allocation().
-			 */
-			ntfs_attr_put_search_ctx(ctx);
-			unmap_mft_record(base_ni);
-			up_write(&ni->runlist.lock);
-			err = ntfs_attr_extend_allocation(ni, new_size,
-					size_change > 0 ? new_size : -1, -1);
-			/*
-			 * ntfs_attr_extend_allocation() will have done error
-			 * output already.
-			 */
-			goto done;
-		}
-		if (!alloc_change)
-			goto alloc_done;
-	}
-	/* alloc_change < 0 */
-	/* Free the clusters. */
-	nr_freed = ntfs_cluster_free(ni, new_alloc_size >>
-			vol->cluster_size_bits, -1, ctx);
-	m = ctx->mrec;
-	a = ctx->attr;
-	if (unlikely(nr_freed < 0)) {
-		ntfs_error(vol->sb, "Failed to release cluster(s) (error code "
-				"%lli).  Unmount and run chkdsk to recover "
-				"the lost cluster(s).", (long long)nr_freed);
-		NVolSetErrors(vol);
-		nr_freed = 0;
-	}
-	/* Truncate the runlist. */
-	err = ntfs_rl_truncate_nolock(vol, &ni->runlist,
-			new_alloc_size >> vol->cluster_size_bits);
-	/*
-	 * If the runlist truncation failed and/or the search context is no
-	 * longer valid, we cannot resize the attribute record or build the
-	 * mapping pairs array thus we mark the inode bad so that no access to
-	 * the freed clusters can happen.
-	 */
-	if (unlikely(err || IS_ERR(m))) {
-		ntfs_error(vol->sb, "Failed to %s (error code %li).%s",
-				IS_ERR(m) ?
-				"restore attribute search context" :
-				"truncate attribute runlist",
-				IS_ERR(m) ? PTR_ERR(m) : err, es);
-		err = -EIO;
-		goto bad_out;
-	}
-	/* Get the size for the shrunk mapping pairs array for the runlist. */
-	mp_size = ntfs_get_size_for_mapping_pairs(vol, ni->runlist.rl, 0, -1);
-	if (unlikely(mp_size <= 0)) {
-		ntfs_error(vol->sb, "Cannot shrink allocation of inode 0x%lx, "
-				"attribute type 0x%x, because determining the "
-				"size for the mapping pairs failed with error "
-				"code %i.%s", vi->i_ino,
-				(unsigned)le32_to_cpu(ni->type), mp_size, es);
-		err = -EIO;
-		goto bad_out;
-	}
-	/*
-	 * Shrink the attribute record for the new mapping pairs array.  Note,
-	 * this cannot fail since we are making the attribute smaller thus by
-	 * definition there is enough space to do so.
-	 */
-	err = ntfs_attr_record_resize(m, a, mp_size +
-			le16_to_cpu(a->data.non_resident.mapping_pairs_offset));
-	BUG_ON(err);
-	/*
-	 * Generate the mapping pairs array directly into the attribute record.
-	 */
-	err = ntfs_mapping_pairs_build(vol, (u8*)a +
-			le16_to_cpu(a->data.non_resident.mapping_pairs_offset),
-			mp_size, ni->runlist.rl, 0, -1, NULL);
-	if (unlikely(err)) {
-		ntfs_error(vol->sb, "Cannot shrink allocation of inode 0x%lx, "
-				"attribute type 0x%x, because building the "
-				"mapping pairs failed with error code %i.%s",
-				vi->i_ino, (unsigned)le32_to_cpu(ni->type),
-				err, es);
-		err = -EIO;
-		goto bad_out;
-	}
-	/* Update the allocated/compressed size as well as the highest vcn. */
-	a->data.non_resident.highest_vcn = cpu_to_sle64((new_alloc_size >>
-			vol->cluster_size_bits) - 1);
-	write_lock_irqsave(&ni->size_lock, flags);
-	ni->allocated_size = new_alloc_size;
-	a->data.non_resident.allocated_size = cpu_to_sle64(new_alloc_size);
-	if (NInoSparse(ni) || NInoCompressed(ni)) {
-		if (nr_freed) {
-			ni->itype.compressed.size -= nr_freed <<
-					vol->cluster_size_bits;
-			BUG_ON(ni->itype.compressed.size < 0);
-			a->data.non_resident.compressed_size = cpu_to_sle64(
-					ni->itype.compressed.size);
-			vi->i_blocks = ni->itype.compressed.size >> 9;
-		}
-	} else
-		vi->i_blocks = new_alloc_size >> 9;
-	write_unlock_irqrestore(&ni->size_lock, flags);
-	/*
-	 * We have shrunk the allocation.  If this is a shrinking truncate we
-	 * have already dealt with the initialized_size and the data_size above
-	 * and we are done.  If the truncate is only changing the allocation
-	 * and not the data_size, we are also done.  If this is an extending
-	 * truncate, need to extend the data_size now which is ensured by the
-	 * fact that @size_change is positive.
-	 */
-alloc_done:
-	/*
-	 * If the size is growing, need to update it now.  If it is shrinking,
-	 * we have already updated it above (before the allocation change).
-	 */
-	if (size_change > 0)
-		a->data.non_resident.data_size = cpu_to_sle64(new_size);
-	/* Ensure the modified mft record is written out. */
-	flush_dcache_mft_record_page(ctx->ntfs_ino);
-	mark_mft_record_dirty(ctx->ntfs_ino);
-unm_done:
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(base_ni);
-	up_write(&ni->runlist.lock);
-done:
-	/* Update the mtime and ctime on the base inode. */
-	/* normally ->truncate shouldn't update ctime or mtime,
-	 * but ntfs did before so it got a copy & paste version
-	 * of file_update_time.  one day someone should fix this
-	 * for real.
-	 */
-	if (!IS_NOCMTIME(VFS_I(base_ni)) && !IS_RDONLY(VFS_I(base_ni))) {
-		struct timespec64 now = current_time(VFS_I(base_ni));
-		struct timespec64 ctime = inode_get_ctime(VFS_I(base_ni));
-		struct timespec64 mtime = inode_get_mtime(VFS_I(base_ni));
-		int sync_it = 0;
-
-		if (!timespec64_equal(&mtime, &now) ||
-		    !timespec64_equal(&ctime, &now))
-			sync_it = 1;
-		inode_set_ctime_to_ts(VFS_I(base_ni), now);
-		inode_set_mtime_to_ts(VFS_I(base_ni), now);
-
-		if (sync_it)
-			mark_inode_dirty_sync(VFS_I(base_ni));
-	}
-
-	if (likely(!err)) {
-		NInoClearTruncateFailed(ni);
-		ntfs_debug("Done.");
-	}
-	return err;
-old_bad_out:
-	old_size = -1;
-bad_out:
-	if (err != -ENOMEM && err != -EOPNOTSUPP)
-		NVolSetErrors(vol);
-	if (err != -EOPNOTSUPP)
-		NInoSetTruncateFailed(ni);
-	else if (old_size >= 0)
-		i_size_write(vi, old_size);
-err_out:
-	if (ctx)
-		ntfs_attr_put_search_ctx(ctx);
-	if (m)
-		unmap_mft_record(base_ni);
-	up_write(&ni->runlist.lock);
-out:
-	ntfs_debug("Failed.  Returning error code %i.", err);
-	return err;
-conv_err_out:
-	if (err != -ENOMEM && err != -EOPNOTSUPP)
-		NVolSetErrors(vol);
-	if (err != -EOPNOTSUPP)
-		NInoSetTruncateFailed(ni);
-	else
-		i_size_write(vi, old_size);
-	goto out;
-}
-
-/**
- * ntfs_truncate_vfs - wrapper for ntfs_truncate() that has no return value
- * @vi:		inode for which the i_size was changed
- *
- * Wrapper for ntfs_truncate() that has no return value.
- *
- * See ntfs_truncate() description above for details.
- */
-#ifdef NTFS_RW
-void ntfs_truncate_vfs(struct inode *vi) {
-	ntfs_truncate(vi);
-}
-#endif
-
-/**
- * ntfs_setattr - called from notify_change() when an attribute is being changed
- * @idmap:	idmap of the mount the inode was found from
- * @dentry:	dentry whose attributes to change
- * @attr:	structure describing the attributes and the changes
- *
- * We have to trap VFS attempts to truncate the file described by @dentry as
- * soon as possible, because we do not implement changes in i_size yet.  So we
- * abort all i_size changes here.
- *
- * We also abort all changes of user, group, and mode as we do not implement
- * the NTFS ACLs yet.
- *
- * Called with ->i_mutex held.
- */
-int ntfs_setattr(struct mnt_idmap *idmap, struct dentry *dentry,
-		 struct iattr *attr)
-{
-	struct inode *vi = d_inode(dentry);
-	int err;
-	unsigned int ia_valid = attr->ia_valid;
-
-	err = setattr_prepare(&nop_mnt_idmap, dentry, attr);
-	if (err)
-		goto out;
-	/* We do not support NTFS ACLs yet. */
-	if (ia_valid & (ATTR_UID | ATTR_GID | ATTR_MODE)) {
-		ntfs_warning(vi->i_sb, "Changes in user/group/mode are not "
-				"supported yet, ignoring.");
-		err = -EOPNOTSUPP;
-		goto out;
-	}
-	if (ia_valid & ATTR_SIZE) {
-		if (attr->ia_size != i_size_read(vi)) {
-			ntfs_inode *ni = NTFS_I(vi);
-			/*
-			 * FIXME: For now we do not support resizing of
-			 * compressed or encrypted files yet.
-			 */
-			if (NInoCompressed(ni) || NInoEncrypted(ni)) {
-				ntfs_warning(vi->i_sb, "Changes in inode size "
-						"are not supported yet for "
-						"%s files, ignoring.",
-						NInoCompressed(ni) ?
-						"compressed" : "encrypted");
-				err = -EOPNOTSUPP;
-			} else {
-				truncate_setsize(vi, attr->ia_size);
-				ntfs_truncate_vfs(vi);
-			}
-			if (err || ia_valid == ATTR_SIZE)
-				goto out;
-		} else {
-			/*
-			 * We skipped the truncate but must still update
-			 * timestamps.
-			 */
-			ia_valid |= ATTR_MTIME | ATTR_CTIME;
-		}
-	}
-	if (ia_valid & ATTR_ATIME)
-		inode_set_atime_to_ts(vi, attr->ia_atime);
-	if (ia_valid & ATTR_MTIME)
-		inode_set_mtime_to_ts(vi, attr->ia_mtime);
-	if (ia_valid & ATTR_CTIME)
-		inode_set_ctime_to_ts(vi, attr->ia_ctime);
-	mark_inode_dirty(vi);
-out:
-	return err;
-}
-
-/**
- * __ntfs_write_inode - write out a dirty inode
- * @vi:		inode to write out
- * @sync:	if true, write out synchronously
- *
- * Write out a dirty inode to disk including any extent inodes if present.
- *
- * If @sync is true, commit the inode to disk and wait for io completion.  This
- * is done using write_mft_record().
- *
- * If @sync is false, just schedule the write to happen but do not wait for i/o
- * completion.  In 2.6 kernels, scheduling usually happens just by virtue of
- * marking the page (and in this case mft record) dirty but we do not implement
- * this yet as write_mft_record() largely ignores the @sync parameter and
- * always performs synchronous writes.
- *
- * Return 0 on success and -errno on error.
- */
-int __ntfs_write_inode(struct inode *vi, int sync)
-{
-	sle64 nt;
-	ntfs_inode *ni = NTFS_I(vi);
-	ntfs_attr_search_ctx *ctx;
-	MFT_RECORD *m;
-	STANDARD_INFORMATION *si;
-	int err = 0;
-	bool modified = false;
-
-	ntfs_debug("Entering for %sinode 0x%lx.", NInoAttr(ni) ? "attr " : "",
-			vi->i_ino);
-	/*
-	 * Dirty attribute inodes are written via their real inodes so just
-	 * clean them here.  Access time updates are taken care off when the
-	 * real inode is written.
-	 */
-	if (NInoAttr(ni)) {
-		NInoClearDirty(ni);
-		ntfs_debug("Done.");
-		return 0;
-	}
-	/* Map, pin, and lock the mft record belonging to the inode. */
-	m = map_mft_record(ni);
-	if (IS_ERR(m)) {
-		err = PTR_ERR(m);
-		goto err_out;
-	}
-	/* Update the access times in the standard information attribute. */
-	ctx = ntfs_attr_get_search_ctx(ni, m);
-	if (unlikely(!ctx)) {
-		err = -ENOMEM;
-		goto unm_err_out;
-	}
-	err = ntfs_attr_lookup(AT_STANDARD_INFORMATION, NULL, 0,
-			CASE_SENSITIVE, 0, NULL, 0, ctx);
-	if (unlikely(err)) {
-		ntfs_attr_put_search_ctx(ctx);
-		goto unm_err_out;
-	}
-	si = (STANDARD_INFORMATION*)((u8*)ctx->attr +
-			le16_to_cpu(ctx->attr->data.resident.value_offset));
-	/* Update the access times if they have changed. */
-	nt = utc2ntfs(inode_get_mtime(vi));
-	if (si->last_data_change_time != nt) {
-		ntfs_debug("Updating mtime for inode 0x%lx: old = 0x%llx, "
-				"new = 0x%llx", vi->i_ino, (long long)
-				sle64_to_cpu(si->last_data_change_time),
-				(long long)sle64_to_cpu(nt));
-		si->last_data_change_time = nt;
-		modified = true;
-	}
-	nt = utc2ntfs(inode_get_ctime(vi));
-	if (si->last_mft_change_time != nt) {
-		ntfs_debug("Updating ctime for inode 0x%lx: old = 0x%llx, "
-				"new = 0x%llx", vi->i_ino, (long long)
-				sle64_to_cpu(si->last_mft_change_time),
-				(long long)sle64_to_cpu(nt));
-		si->last_mft_change_time = nt;
-		modified = true;
-	}
-	nt = utc2ntfs(inode_get_atime(vi));
-	if (si->last_access_time != nt) {
-		ntfs_debug("Updating atime for inode 0x%lx: old = 0x%llx, "
-				"new = 0x%llx", vi->i_ino,
-				(long long)sle64_to_cpu(si->last_access_time),
-				(long long)sle64_to_cpu(nt));
-		si->last_access_time = nt;
-		modified = true;
-	}
-	/*
-	 * If we just modified the standard information attribute we need to
-	 * mark the mft record it is in dirty.  We do this manually so that
-	 * mark_inode_dirty() is not called which would redirty the inode and
-	 * hence result in an infinite loop of trying to write the inode.
-	 * There is no need to mark the base inode nor the base mft record
-	 * dirty, since we are going to write this mft record below in any case
-	 * and the base mft record may actually not have been modified so it
-	 * might not need to be written out.
-	 * NOTE: It is not a problem when the inode for $MFT itself is being
-	 * written out as mark_ntfs_record_dirty() will only set I_DIRTY_PAGES
-	 * on the $MFT inode and hence __ntfs_write_inode() will not be
-	 * re-invoked because of it which in turn is ok since the dirtied mft
-	 * record will be cleaned and written out to disk below, i.e. before
-	 * this function returns.
-	 */
-	if (modified) {
-		flush_dcache_mft_record_page(ctx->ntfs_ino);
-		if (!NInoTestSetDirty(ctx->ntfs_ino))
-			mark_ntfs_record_dirty(ctx->ntfs_ino->page,
-					ctx->ntfs_ino->page_ofs);
-	}
-	ntfs_attr_put_search_ctx(ctx);
-	/* Now the access times are updated, write the base mft record. */
-	if (NInoDirty(ni))
-		err = write_mft_record(ni, m, sync);
-	/* Write all attached extent mft records. */
-	mutex_lock(&ni->extent_lock);
-	if (ni->nr_extents > 0) {
-		ntfs_inode **extent_nis = ni->ext.extent_ntfs_inos;
-		int i;
-
-		ntfs_debug("Writing %i extent inodes.", ni->nr_extents);
-		for (i = 0; i < ni->nr_extents; i++) {
-			ntfs_inode *tni = extent_nis[i];
-
-			if (NInoDirty(tni)) {
-				MFT_RECORD *tm = map_mft_record(tni);
-				int ret;
-
-				if (IS_ERR(tm)) {
-					if (!err || err == -ENOMEM)
-						err = PTR_ERR(tm);
-					continue;
-				}
-				ret = write_mft_record(tni, tm, sync);
-				unmap_mft_record(tni);
-				if (unlikely(ret)) {
-					if (!err || err == -ENOMEM)
-						err = ret;
-				}
-			}
-		}
-	}
-	mutex_unlock(&ni->extent_lock);
-	unmap_mft_record(ni);
-	if (unlikely(err))
-		goto err_out;
-	ntfs_debug("Done.");
-	return 0;
-unm_err_out:
-	unmap_mft_record(ni);
-err_out:
-	if (err == -ENOMEM) {
-		ntfs_warning(vi->i_sb, "Not enough memory to write inode.  "
-				"Marking the inode dirty again, so the VFS "
-				"retries later.");
-		mark_inode_dirty(vi);
-	} else {
-		ntfs_error(vi->i_sb, "Failed (error %i):  Run chkdsk.", -err);
-		NVolSetErrors(ni->vol);
-	}
-	return err;
-}
-
-#endif /* NTFS_RW */
diff --git a/fs/ntfs/inode.h b/fs/ntfs/inode.h
deleted file mode 100644
index 147ef4ddb6..0000000000
--- a/fs/ntfs/inode.h
+++ /dev/null
@@ -1,310 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * inode.h - Defines for inode structures NTFS Linux kernel driver. Part of
- *	     the Linux-NTFS project.
- *
- * Copyright (c) 2001-2007 Anton Altaparmakov
- * Copyright (c) 2002 Richard Russon
- */
-
-#ifndef _LINUX_NTFS_INODE_H
-#define _LINUX_NTFS_INODE_H
-
-#include <linux/atomic.h>
-
-#include <linux/fs.h>
-#include <linux/list.h>
-#include <linux/mm.h>
-#include <linux/mutex.h>
-#include <linux/seq_file.h>
-
-#include "layout.h"
-#include "volume.h"
-#include "types.h"
-#include "runlist.h"
-#include "debug.h"
-
-typedef struct _ntfs_inode ntfs_inode;
-
-/*
- * The NTFS in-memory inode structure. It is just used as an extension to the
- * fields already provided in the VFS inode.
- */
-struct _ntfs_inode {
-	rwlock_t size_lock;	/* Lock serializing access to inode sizes. */
-	s64 initialized_size;	/* Copy from the attribute record. */
-	s64 allocated_size;	/* Copy from the attribute record. */
-	unsigned long state;	/* NTFS specific flags describing this inode.
-				   See ntfs_inode_state_bits below. */
-	unsigned long mft_no;	/* Number of the mft record / inode. */
-	u16 seq_no;		/* Sequence number of the mft record. */
-	atomic_t count;		/* Inode reference count for book keeping. */
-	ntfs_volume *vol;	/* Pointer to the ntfs volume of this inode. */
-	/*
-	 * If NInoAttr() is true, the below fields describe the attribute which
-	 * this fake inode belongs to. The actual inode of this attribute is
-	 * pointed to by base_ntfs_ino and nr_extents is always set to -1 (see
-	 * below). For real inodes, we also set the type (AT_DATA for files and
-	 * AT_INDEX_ALLOCATION for directories), with the name = NULL and
-	 * name_len = 0 for files and name = I30 (global constant) and
-	 * name_len = 4 for directories.
-	 */
-	ATTR_TYPE type;	/* Attribute type of this fake inode. */
-	ntfschar *name;		/* Attribute name of this fake inode. */
-	u32 name_len;		/* Attribute name length of this fake inode. */
-	runlist runlist;	/* If state has the NI_NonResident bit set,
-				   the runlist of the unnamed data attribute
-				   (if a file) or of the index allocation
-				   attribute (directory) or of the attribute
-				   described by the fake inode (if NInoAttr()).
-				   If runlist.rl is NULL, the runlist has not
-				   been read in yet or has been unmapped. If
-				   NI_NonResident is clear, the attribute is
-				   resident (file and fake inode) or there is
-				   no $I30 index allocation attribute
-				   (small directory). In the latter case
-				   runlist.rl is always NULL.*/
-	/*
-	 * The following fields are only valid for real inodes and extent
-	 * inodes.
-	 */
-	struct mutex mrec_lock;	/* Lock for serializing access to the
-				   mft record belonging to this inode. */
-	struct page *page;	/* The page containing the mft record of the
-				   inode. This should only be touched by the
-				   (un)map_mft_record*() functions. */
-	int page_ofs;		/* Offset into the page at which the mft record
-				   begins. This should only be touched by the
-				   (un)map_mft_record*() functions. */
-	/*
-	 * Attribute list support (only for use by the attribute lookup
-	 * functions). Setup during read_inode for all inodes with attribute
-	 * lists. Only valid if NI_AttrList is set in state, and attr_list_rl is
-	 * further only valid if NI_AttrListNonResident is set.
-	 */
-	u32 attr_list_size;	/* Length of attribute list value in bytes. */
-	u8 *attr_list;		/* Attribute list value itself. */
-	runlist attr_list_rl;	/* Run list for the attribute list value. */
-	union {
-		struct { /* It is a directory, $MFT, or an index inode. */
-			u32 block_size;		/* Size of an index block. */
-			u32 vcn_size;		/* Size of a vcn in this
-						   index. */
-			COLLATION_RULE collation_rule; /* The collation rule
-						   for the index. */
-			u8 block_size_bits; 	/* Log2 of the above. */
-			u8 vcn_size_bits;	/* Log2 of the above. */
-		} index;
-		struct { /* It is a compressed/sparse file/attribute inode. */
-			s64 size;		/* Copy of compressed_size from
-						   $DATA. */
-			u32 block_size;		/* Size of a compression block
-						   (cb). */
-			u8 block_size_bits;	/* Log2 of the size of a cb. */
-			u8 block_clusters;	/* Number of clusters per cb. */
-		} compressed;
-	} itype;
-	struct mutex extent_lock;	/* Lock for accessing/modifying the
-					   below . */
-	s32 nr_extents;	/* For a base mft record, the number of attached extent
-			   inodes (0 if none), for extent records and for fake
-			   inodes describing an attribute this is -1. */
-	union {		/* This union is only used if nr_extents != 0. */
-		ntfs_inode **extent_ntfs_inos;	/* For nr_extents > 0, array of
-						   the ntfs inodes of the extent
-						   mft records belonging to
-						   this base inode which have
-						   been loaded. */
-		ntfs_inode *base_ntfs_ino;	/* For nr_extents == -1, the
-						   ntfs inode of the base mft
-						   record. For fake inodes, the
-						   real (base) inode to which
-						   the attribute belongs. */
-	} ext;
-};
-
-/*
- * Defined bits for the state field in the ntfs_inode structure.
- * (f) = files only, (d) = directories only, (a) = attributes/fake inodes only
- */
-typedef enum {
-	NI_Dirty,		/* 1: Mft record needs to be written to disk. */
-	NI_AttrList,		/* 1: Mft record contains an attribute list. */
-	NI_AttrListNonResident,	/* 1: Attribute list is non-resident. Implies
-				      NI_AttrList is set. */
-
-	NI_Attr,		/* 1: Fake inode for attribute i/o.
-				   0: Real inode or extent inode. */
-
-	NI_MstProtected,	/* 1: Attribute is protected by MST fixups.
-				   0: Attribute is not protected by fixups. */
-	NI_NonResident,		/* 1: Unnamed data attr is non-resident (f).
-				   1: Attribute is non-resident (a). */
-	NI_IndexAllocPresent = NI_NonResident,	/* 1: $I30 index alloc attr is
-						   present (d). */
-	NI_Compressed,		/* 1: Unnamed data attr is compressed (f).
-				   1: Create compressed files by default (d).
-				   1: Attribute is compressed (a). */
-	NI_Encrypted,		/* 1: Unnamed data attr is encrypted (f).
-				   1: Create encrypted files by default (d).
-				   1: Attribute is encrypted (a). */
-	NI_Sparse,		/* 1: Unnamed data attr is sparse (f).
-				   1: Create sparse files by default (d).
-				   1: Attribute is sparse (a). */
-	NI_SparseDisabled,	/* 1: May not create sparse regions. */
-	NI_TruncateFailed,	/* 1: Last ntfs_truncate() call failed. */
-} ntfs_inode_state_bits;
-
-/*
- * NOTE: We should be adding dirty mft records to a list somewhere and they
- * should be independent of the (ntfs/vfs) inode structure so that an inode can
- * be removed but the record can be left dirty for syncing later.
- */
-
-/*
- * Macro tricks to expand the NInoFoo(), NInoSetFoo(), and NInoClearFoo()
- * functions.
- */
-#define NINO_FNS(flag)					\
-static inline int NIno##flag(ntfs_inode *ni)		\
-{							\
-	return test_bit(NI_##flag, &(ni)->state);	\
-}							\
-static inline void NInoSet##flag(ntfs_inode *ni)	\
-{							\
-	set_bit(NI_##flag, &(ni)->state);		\
-}							\
-static inline void NInoClear##flag(ntfs_inode *ni)	\
-{							\
-	clear_bit(NI_##flag, &(ni)->state);		\
-}
-
-/*
- * As above for NInoTestSetFoo() and NInoTestClearFoo().
- */
-#define TAS_NINO_FNS(flag)					\
-static inline int NInoTestSet##flag(ntfs_inode *ni)		\
-{								\
-	return test_and_set_bit(NI_##flag, &(ni)->state);	\
-}								\
-static inline int NInoTestClear##flag(ntfs_inode *ni)		\
-{								\
-	return test_and_clear_bit(NI_##flag, &(ni)->state);	\
-}
-
-/* Emit the ntfs inode bitops functions. */
-NINO_FNS(Dirty)
-TAS_NINO_FNS(Dirty)
-NINO_FNS(AttrList)
-NINO_FNS(AttrListNonResident)
-NINO_FNS(Attr)
-NINO_FNS(MstProtected)
-NINO_FNS(NonResident)
-NINO_FNS(IndexAllocPresent)
-NINO_FNS(Compressed)
-NINO_FNS(Encrypted)
-NINO_FNS(Sparse)
-NINO_FNS(SparseDisabled)
-NINO_FNS(TruncateFailed)
-
-/*
- * The full structure containing a ntfs_inode and a vfs struct inode. Used for
- * all real and fake inodes but not for extent inodes which lack the vfs struct
- * inode.
- */
-typedef struct {
-	ntfs_inode ntfs_inode;
-	struct inode vfs_inode;		/* The vfs inode structure. */
-} big_ntfs_inode;
-
-/**
- * NTFS_I - return the ntfs inode given a vfs inode
- * @inode:	VFS inode
- *
- * NTFS_I() returns the ntfs inode associated with the VFS @inode.
- */
-static inline ntfs_inode *NTFS_I(struct inode *inode)
-{
-	return (ntfs_inode *)container_of(inode, big_ntfs_inode, vfs_inode);
-}
-
-static inline struct inode *VFS_I(ntfs_inode *ni)
-{
-	return &((big_ntfs_inode *)ni)->vfs_inode;
-}
-
-/**
- * ntfs_attr - ntfs in memory attribute structure
- * @mft_no:	mft record number of the base mft record of this attribute
- * @name:	Unicode name of the attribute (NULL if unnamed)
- * @name_len:	length of @name in Unicode characters (0 if unnamed)
- * @type:	attribute type (see layout.h)
- *
- * This structure exists only to provide a small structure for the
- * ntfs_{attr_}iget()/ntfs_test_inode()/ntfs_init_locked_inode() mechanism.
- *
- * NOTE: Elements are ordered by size to make the structure as compact as
- * possible on all architectures.
- */
-typedef struct {
-	unsigned long mft_no;
-	ntfschar *name;
-	u32 name_len;
-	ATTR_TYPE type;
-} ntfs_attr;
-
-extern int ntfs_test_inode(struct inode *vi, void *data);
-
-extern struct inode *ntfs_iget(struct super_block *sb, unsigned long mft_no);
-extern struct inode *ntfs_attr_iget(struct inode *base_vi, ATTR_TYPE type,
-		ntfschar *name, u32 name_len);
-extern struct inode *ntfs_index_iget(struct inode *base_vi, ntfschar *name,
-		u32 name_len);
-
-extern struct inode *ntfs_alloc_big_inode(struct super_block *sb);
-extern void ntfs_free_big_inode(struct inode *inode);
-extern void ntfs_evict_big_inode(struct inode *vi);
-
-extern void __ntfs_init_inode(struct super_block *sb, ntfs_inode *ni);
-
-static inline void ntfs_init_big_inode(struct inode *vi)
-{
-	ntfs_inode *ni = NTFS_I(vi);
-
-	ntfs_debug("Entering.");
-	__ntfs_init_inode(vi->i_sb, ni);
-	ni->mft_no = vi->i_ino;
-}
-
-extern ntfs_inode *ntfs_new_extent_inode(struct super_block *sb,
-		unsigned long mft_no);
-extern void ntfs_clear_extent_inode(ntfs_inode *ni);
-
-extern int ntfs_read_inode_mount(struct inode *vi);
-
-extern int ntfs_show_options(struct seq_file *sf, struct dentry *root);
-
-#ifdef NTFS_RW
-
-extern int ntfs_truncate(struct inode *vi);
-extern void ntfs_truncate_vfs(struct inode *vi);
-
-extern int ntfs_setattr(struct mnt_idmap *idmap,
-			struct dentry *dentry, struct iattr *attr);
-
-extern int __ntfs_write_inode(struct inode *vi, int sync);
-
-static inline void ntfs_commit_inode(struct inode *vi)
-{
-	if (!is_bad_inode(vi))
-		__ntfs_write_inode(vi, 1);
-	return;
-}
-
-#else
-
-static inline void ntfs_truncate_vfs(struct inode *vi) {}
-
-#endif /* NTFS_RW */
-
-#endif /* _LINUX_NTFS_INODE_H */
diff --git a/fs/ntfs/layout.h b/fs/ntfs/layout.h
deleted file mode 100644
index 5d4bf7a325..0000000000
--- a/fs/ntfs/layout.h
+++ /dev/null
@@ -1,2421 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * layout.h - All NTFS associated on-disk structures. Part of the Linux-NTFS
- *	      project.
- *
- * Copyright (c) 2001-2005 Anton Altaparmakov
- * Copyright (c) 2002 Richard Russon
- */
-
-#ifndef _LINUX_NTFS_LAYOUT_H
-#define _LINUX_NTFS_LAYOUT_H
-
-#include <linux/types.h>
-#include <linux/bitops.h>
-#include <linux/list.h>
-#include <asm/byteorder.h>
-
-#include "types.h"
-
-/* The NTFS oem_id "NTFS    " */
-#define magicNTFS	cpu_to_le64(0x202020205346544eULL)
-
-/*
- * Location of bootsector on partition:
- *	The standard NTFS_BOOT_SECTOR is on sector 0 of the partition.
- *	On NT4 and above there is one backup copy of the boot sector to
- *	be found on the last sector of the partition (not normally accessible
- *	from within Windows as the bootsector contained number of sectors
- *	value is one less than the actual value!).
- *	On versions of NT 3.51 and earlier, the backup copy was located at
- *	number of sectors/2 (integer divide), i.e. in the middle of the volume.
- */
-
-/*
- * BIOS parameter block (bpb) structure.
- */
-typedef struct {
-	le16 bytes_per_sector;		/* Size of a sector in bytes. */
-	u8  sectors_per_cluster;	/* Size of a cluster in sectors. */
-	le16 reserved_sectors;		/* zero */
-	u8  fats;			/* zero */
-	le16 root_entries;		/* zero */
-	le16 sectors;			/* zero */
-	u8  media_type;			/* 0xf8 = hard disk */
-	le16 sectors_per_fat;		/* zero */
-	le16 sectors_per_track;		/* irrelevant */
-	le16 heads;			/* irrelevant */
-	le32 hidden_sectors;		/* zero */
-	le32 large_sectors;		/* zero */
-} __attribute__ ((__packed__)) BIOS_PARAMETER_BLOCK;
-
-/*
- * NTFS boot sector structure.
- */
-typedef struct {
-	u8  jump[3];			/* Irrelevant (jump to boot up code).*/
-	le64 oem_id;			/* Magic "NTFS    ". */
-	BIOS_PARAMETER_BLOCK bpb;	/* See BIOS_PARAMETER_BLOCK. */
-	u8  unused[4];			/* zero, NTFS diskedit.exe states that
-					   this is actually:
-						__u8 physical_drive;	// 0x80
-						__u8 current_head;	// zero
-						__u8 extended_boot_signature;
-									// 0x80
-						__u8 unused;		// zero
-					 */
-/*0x28*/sle64 number_of_sectors;	/* Number of sectors in volume. Gives
-					   maximum volume size of 2^63 sectors.
-					   Assuming standard sector size of 512
-					   bytes, the maximum byte size is
-					   approx. 4.7x10^21 bytes. (-; */
-	sle64 mft_lcn;			/* Cluster location of mft data. */
-	sle64 mftmirr_lcn;		/* Cluster location of copy of mft. */
-	s8  clusters_per_mft_record;	/* Mft record size in clusters. */
-	u8  reserved0[3];		/* zero */
-	s8  clusters_per_index_record;	/* Index block size in clusters. */
-	u8  reserved1[3];		/* zero */
-	le64 volume_serial_number;	/* Irrelevant (serial number). */
-	le32 checksum;			/* Boot sector checksum. */
-/*0x54*/u8  bootstrap[426];		/* Irrelevant (boot up code). */
-	le16 end_of_sector_marker;	/* End of bootsector magic. Always is
-					   0xaa55 in little endian. */
-/* sizeof() = 512 (0x200) bytes */
-} __attribute__ ((__packed__)) NTFS_BOOT_SECTOR;
-
-/*
- * Magic identifiers present at the beginning of all ntfs record containing
- * records (like mft records for example).
- */
-enum {
-	/* Found in $MFT/$DATA. */
-	magic_FILE = cpu_to_le32(0x454c4946), /* Mft entry. */
-	magic_INDX = cpu_to_le32(0x58444e49), /* Index buffer. */
-	magic_HOLE = cpu_to_le32(0x454c4f48), /* ? (NTFS 3.0+?) */
-
-	/* Found in $LogFile/$DATA. */
-	magic_RSTR = cpu_to_le32(0x52545352), /* Restart page. */
-	magic_RCRD = cpu_to_le32(0x44524352), /* Log record page. */
-
-	/* Found in $LogFile/$DATA.  (May be found in $MFT/$DATA, also?) */
-	magic_CHKD = cpu_to_le32(0x444b4843), /* Modified by chkdsk. */
-
-	/* Found in all ntfs record containing records. */
-	magic_BAAD = cpu_to_le32(0x44414142), /* Failed multi sector
-						       transfer was detected. */
-	/*
-	 * Found in $LogFile/$DATA when a page is full of 0xff bytes and is
-	 * thus not initialized.  Page must be initialized before using it.
-	 */
-	magic_empty = cpu_to_le32(0xffffffff) /* Record is empty. */
-};
-
-typedef le32 NTFS_RECORD_TYPE;
-
-/*
- * Generic magic comparison macros. Finally found a use for the ## preprocessor
- * operator! (-8
- */
-
-static inline bool __ntfs_is_magic(le32 x, NTFS_RECORD_TYPE r)
-{
-	return (x == r);
-}
-#define ntfs_is_magic(x, m)	__ntfs_is_magic(x, magic_##m)
-
-static inline bool __ntfs_is_magicp(le32 *p, NTFS_RECORD_TYPE r)
-{
-	return (*p == r);
-}
-#define ntfs_is_magicp(p, m)	__ntfs_is_magicp(p, magic_##m)
-
-/*
- * Specialised magic comparison macros for the NTFS_RECORD_TYPEs defined above.
- */
-#define ntfs_is_file_record(x)		( ntfs_is_magic (x, FILE) )
-#define ntfs_is_file_recordp(p)		( ntfs_is_magicp(p, FILE) )
-#define ntfs_is_mft_record(x)		( ntfs_is_file_record (x) )
-#define ntfs_is_mft_recordp(p)		( ntfs_is_file_recordp(p) )
-#define ntfs_is_indx_record(x)		( ntfs_is_magic (x, INDX) )
-#define ntfs_is_indx_recordp(p)		( ntfs_is_magicp(p, INDX) )
-#define ntfs_is_hole_record(x)		( ntfs_is_magic (x, HOLE) )
-#define ntfs_is_hole_recordp(p)		( ntfs_is_magicp(p, HOLE) )
-
-#define ntfs_is_rstr_record(x)		( ntfs_is_magic (x, RSTR) )
-#define ntfs_is_rstr_recordp(p)		( ntfs_is_magicp(p, RSTR) )
-#define ntfs_is_rcrd_record(x)		( ntfs_is_magic (x, RCRD) )
-#define ntfs_is_rcrd_recordp(p)		( ntfs_is_magicp(p, RCRD) )
-
-#define ntfs_is_chkd_record(x)		( ntfs_is_magic (x, CHKD) )
-#define ntfs_is_chkd_recordp(p)		( ntfs_is_magicp(p, CHKD) )
-
-#define ntfs_is_baad_record(x)		( ntfs_is_magic (x, BAAD) )
-#define ntfs_is_baad_recordp(p)		( ntfs_is_magicp(p, BAAD) )
-
-#define ntfs_is_empty_record(x)		( ntfs_is_magic (x, empty) )
-#define ntfs_is_empty_recordp(p)	( ntfs_is_magicp(p, empty) )
-
-/*
- * The Update Sequence Array (usa) is an array of the le16 values which belong
- * to the end of each sector protected by the update sequence record in which
- * this array is contained. Note that the first entry is the Update Sequence
- * Number (usn), a cyclic counter of how many times the protected record has
- * been written to disk. The values 0 and -1 (ie. 0xffff) are not used. All
- * last le16's of each sector have to be equal to the usn (during reading) or
- * are set to it (during writing). If they are not, an incomplete multi sector
- * transfer has occurred when the data was written.
- * The maximum size for the update sequence array is fixed to:
- *	maximum size = usa_ofs + (usa_count * 2) = 510 bytes
- * The 510 bytes comes from the fact that the last le16 in the array has to
- * (obviously) finish before the last le16 of the first 512-byte sector.
- * This formula can be used as a consistency check in that usa_ofs +
- * (usa_count * 2) has to be less than or equal to 510.
- */
-typedef struct {
-	NTFS_RECORD_TYPE magic;	/* A four-byte magic identifying the record
-				   type and/or status. */
-	le16 usa_ofs;		/* Offset to the Update Sequence Array (usa)
-				   from the start of the ntfs record. */
-	le16 usa_count;		/* Number of le16 sized entries in the usa
-				   including the Update Sequence Number (usn),
-				   thus the number of fixups is the usa_count
-				   minus 1. */
-} __attribute__ ((__packed__)) NTFS_RECORD;
-
-/*
- * System files mft record numbers. All these files are always marked as used
- * in the bitmap attribute of the mft; presumably in order to avoid accidental
- * allocation for random other mft records. Also, the sequence number for each
- * of the system files is always equal to their mft record number and it is
- * never modified.
- */
-typedef enum {
-	FILE_MFT       = 0,	/* Master file table (mft). Data attribute
-				   contains the entries and bitmap attribute
-				   records which ones are in use (bit==1). */
-	FILE_MFTMirr   = 1,	/* Mft mirror: copy of first four mft records
-				   in data attribute. If cluster size > 4kiB,
-				   copy of first N mft records, with
-					N = cluster_size / mft_record_size. */
-	FILE_LogFile   = 2,	/* Journalling log in data attribute. */
-	FILE_Volume    = 3,	/* Volume name attribute and volume information
-				   attribute (flags and ntfs version). Windows
-				   refers to this file as volume DASD (Direct
-				   Access Storage Device). */
-	FILE_AttrDef   = 4,	/* Array of attribute definitions in data
-				   attribute. */
-	FILE_root      = 5,	/* Root directory. */
-	FILE_Bitmap    = 6,	/* Allocation bitmap of all clusters (lcns) in
-				   data attribute. */
-	FILE_Boot      = 7,	/* Boot sector (always at cluster 0) in data
-				   attribute. */
-	FILE_BadClus   = 8,	/* Contains all bad clusters in the non-resident
-				   data attribute. */
-	FILE_Secure    = 9,	/* Shared security descriptors in data attribute
-				   and two indexes into the descriptors.
-				   Appeared in Windows 2000. Before that, this
-				   file was named $Quota but was unused. */
-	FILE_UpCase    = 10,	/* Uppercase equivalents of all 65536 Unicode
-				   characters in data attribute. */
-	FILE_Extend    = 11,	/* Directory containing other system files (eg.
-				   $ObjId, $Quota, $Reparse and $UsnJrnl). This
-				   is new to NTFS3.0. */
-	FILE_reserved12 = 12,	/* Reserved for future use (records 12-15). */
-	FILE_reserved13 = 13,
-	FILE_reserved14 = 14,
-	FILE_reserved15 = 15,
-	FILE_first_user = 16,	/* First user file, used as test limit for
-				   whether to allow opening a file or not. */
-} NTFS_SYSTEM_FILES;
-
-/*
- * These are the so far known MFT_RECORD_* flags (16-bit) which contain
- * information about the mft record in which they are present.
- */
-enum {
-	MFT_RECORD_IN_USE	= cpu_to_le16(0x0001),
-	MFT_RECORD_IS_DIRECTORY = cpu_to_le16(0x0002),
-} __attribute__ ((__packed__));
-
-typedef le16 MFT_RECORD_FLAGS;
-
-/*
- * mft references (aka file references or file record segment references) are
- * used whenever a structure needs to refer to a record in the mft.
- *
- * A reference consists of a 48-bit index into the mft and a 16-bit sequence
- * number used to detect stale references.
- *
- * For error reporting purposes we treat the 48-bit index as a signed quantity.
- *
- * The sequence number is a circular counter (skipping 0) describing how many
- * times the referenced mft record has been (re)used. This has to match the
- * sequence number of the mft record being referenced, otherwise the reference
- * is considered stale and removed (FIXME: only ntfsck or the driver itself?).
- *
- * If the sequence number is zero it is assumed that no sequence number
- * consistency checking should be performed.
- *
- * FIXME: Since inodes are 32-bit as of now, the driver needs to always check
- * for high_part being 0 and if not either BUG(), cause a panic() or handle
- * the situation in some other way. This shouldn't be a problem as a volume has
- * to become HUGE in order to need more than 32-bits worth of mft records.
- * Assuming the standard mft record size of 1kb only the records (never mind
- * the non-resident attributes, etc.) would require 4Tb of space on their own
- * for the first 32 bits worth of records. This is only if some strange person
- * doesn't decide to foul play and make the mft sparse which would be a really
- * horrible thing to do as it would trash our current driver implementation. )-:
- * Do I hear screams "we want 64-bit inodes!" ?!? (-;
- *
- * FIXME: The mft zone is defined as the first 12% of the volume. This space is
- * reserved so that the mft can grow contiguously and hence doesn't become
- * fragmented. Volume free space includes the empty part of the mft zone and
- * when the volume's free 88% are used up, the mft zone is shrunk by a factor
- * of 2, thus making more space available for more files/data. This process is
- * repeated every time there is no more free space except for the mft zone until
- * there really is no more free space.
- */
-
-/*
- * Typedef the MFT_REF as a 64-bit value for easier handling.
- * Also define two unpacking macros to get to the reference (MREF) and
- * sequence number (MSEQNO) respectively.
- * The _LE versions are to be applied on little endian MFT_REFs.
- * Note: The _LE versions will return a CPU endian formatted value!
- */
-#define MFT_REF_MASK_CPU 0x0000ffffffffffffULL
-#define MFT_REF_MASK_LE cpu_to_le64(MFT_REF_MASK_CPU)
-
-typedef u64 MFT_REF;
-typedef le64 leMFT_REF;
-
-#define MK_MREF(m, s)	((MFT_REF)(((MFT_REF)(s) << 48) |		\
-					((MFT_REF)(m) & MFT_REF_MASK_CPU)))
-#define MK_LE_MREF(m, s) cpu_to_le64(MK_MREF(m, s))
-
-#define MREF(x)		((unsigned long)((x) & MFT_REF_MASK_CPU))
-#define MSEQNO(x)	((u16)(((x) >> 48) & 0xffff))
-#define MREF_LE(x)	((unsigned long)(le64_to_cpu(x) & MFT_REF_MASK_CPU))
-#define MSEQNO_LE(x)	((u16)((le64_to_cpu(x) >> 48) & 0xffff))
-
-#define IS_ERR_MREF(x)	(((x) & 0x0000800000000000ULL) ? true : false)
-#define ERR_MREF(x)	((u64)((s64)(x)))
-#define MREF_ERR(x)	((int)((s64)(x)))
-
-/*
- * The mft record header present at the beginning of every record in the mft.
- * This is followed by a sequence of variable length attribute records which
- * is terminated by an attribute of type AT_END which is a truncated attribute
- * in that it only consists of the attribute type code AT_END and none of the
- * other members of the attribute structure are present.
- */
-typedef struct {
-/*Ofs*/
-/*  0	NTFS_RECORD; -- Unfolded here as gcc doesn't like unnamed structs. */
-	NTFS_RECORD_TYPE magic;	/* Usually the magic is "FILE". */
-	le16 usa_ofs;		/* See NTFS_RECORD definition above. */
-	le16 usa_count;		/* See NTFS_RECORD definition above. */
-
-/*  8*/	le64 lsn;		/* $LogFile sequence number for this record.
-				   Changed every time the record is modified. */
-/* 16*/	le16 sequence_number;	/* Number of times this mft record has been
-				   reused. (See description for MFT_REF
-				   above.) NOTE: The increment (skipping zero)
-				   is done when the file is deleted. NOTE: If
-				   this is zero it is left zero. */
-/* 18*/	le16 link_count;	/* Number of hard links, i.e. the number of
-				   directory entries referencing this record.
-				   NOTE: Only used in mft base records.
-				   NOTE: When deleting a directory entry we
-				   check the link_count and if it is 1 we
-				   delete the file. Otherwise we delete the
-				   FILE_NAME_ATTR being referenced by the
-				   directory entry from the mft record and
-				   decrement the link_count.
-				   FIXME: Careful with Win32 + DOS names! */
-/* 20*/	le16 attrs_offset;	/* Byte offset to the first attribute in this
-				   mft record from the start of the mft record.
-				   NOTE: Must be aligned to 8-byte boundary. */
-/* 22*/	MFT_RECORD_FLAGS flags;	/* Bit array of MFT_RECORD_FLAGS. When a file
-				   is deleted, the MFT_RECORD_IN_USE flag is
-				   set to zero. */
-/* 24*/	le32 bytes_in_use;	/* Number of bytes used in this mft record.
-				   NOTE: Must be aligned to 8-byte boundary. */
-/* 28*/	le32 bytes_allocated;	/* Number of bytes allocated for this mft
-				   record. This should be equal to the mft
-				   record size. */
-/* 32*/	leMFT_REF base_mft_record;/* This is zero for base mft records.
-				   When it is not zero it is a mft reference
-				   pointing to the base mft record to which
-				   this record belongs (this is then used to
-				   locate the attribute list attribute present
-				   in the base record which describes this
-				   extension record and hence might need
-				   modification when the extension record
-				   itself is modified, also locating the
-				   attribute list also means finding the other
-				   potential extents, belonging to the non-base
-				   mft record). */
-/* 40*/	le16 next_attr_instance;/* The instance number that will be assigned to
-				   the next attribute added to this mft record.
-				   NOTE: Incremented each time after it is used.
-				   NOTE: Every time the mft record is reused
-				   this number is set to zero.  NOTE: The first
-				   instance number is always 0. */
-/* The below fields are specific to NTFS 3.1+ (Windows XP and above): */
-/* 42*/ le16 reserved;		/* Reserved/alignment. */
-/* 44*/ le32 mft_record_number;	/* Number of this mft record. */
-/* sizeof() = 48 bytes */
-/*
- * When (re)using the mft record, we place the update sequence array at this
- * offset, i.e. before we start with the attributes.  This also makes sense,
- * otherwise we could run into problems with the update sequence array
- * containing in itself the last two bytes of a sector which would mean that
- * multi sector transfer protection wouldn't work.  As you can't protect data
- * by overwriting it since you then can't get it back...
- * When reading we obviously use the data from the ntfs record header.
- */
-} __attribute__ ((__packed__)) MFT_RECORD;
-
-/* This is the version without the NTFS 3.1+ specific fields. */
-typedef struct {
-/*Ofs*/
-/*  0	NTFS_RECORD; -- Unfolded here as gcc doesn't like unnamed structs. */
-	NTFS_RECORD_TYPE magic;	/* Usually the magic is "FILE". */
-	le16 usa_ofs;		/* See NTFS_RECORD definition above. */
-	le16 usa_count;		/* See NTFS_RECORD definition above. */
-
-/*  8*/	le64 lsn;		/* $LogFile sequence number for this record.
-				   Changed every time the record is modified. */
-/* 16*/	le16 sequence_number;	/* Number of times this mft record has been
-				   reused. (See description for MFT_REF
-				   above.) NOTE: The increment (skipping zero)
-				   is done when the file is deleted. NOTE: If
-				   this is zero it is left zero. */
-/* 18*/	le16 link_count;	/* Number of hard links, i.e. the number of
-				   directory entries referencing this record.
-				   NOTE: Only used in mft base records.
-				   NOTE: When deleting a directory entry we
-				   check the link_count and if it is 1 we
-				   delete the file. Otherwise we delete the
-				   FILE_NAME_ATTR being referenced by the
-				   directory entry from the mft record and
-				   decrement the link_count.
-				   FIXME: Careful with Win32 + DOS names! */
-/* 20*/	le16 attrs_offset;	/* Byte offset to the first attribute in this
-				   mft record from the start of the mft record.
-				   NOTE: Must be aligned to 8-byte boundary. */
-/* 22*/	MFT_RECORD_FLAGS flags;	/* Bit array of MFT_RECORD_FLAGS. When a file
-				   is deleted, the MFT_RECORD_IN_USE flag is
-				   set to zero. */
-/* 24*/	le32 bytes_in_use;	/* Number of bytes used in this mft record.
-				   NOTE: Must be aligned to 8-byte boundary. */
-/* 28*/	le32 bytes_allocated;	/* Number of bytes allocated for this mft
-				   record. This should be equal to the mft
-				   record size. */
-/* 32*/	leMFT_REF base_mft_record;/* This is zero for base mft records.
-				   When it is not zero it is a mft reference
-				   pointing to the base mft record to which
-				   this record belongs (this is then used to
-				   locate the attribute list attribute present
-				   in the base record which describes this
-				   extension record and hence might need
-				   modification when the extension record
-				   itself is modified, also locating the
-				   attribute list also means finding the other
-				   potential extents, belonging to the non-base
-				   mft record). */
-/* 40*/	le16 next_attr_instance;/* The instance number that will be assigned to
-				   the next attribute added to this mft record.
-				   NOTE: Incremented each time after it is used.
-				   NOTE: Every time the mft record is reused
-				   this number is set to zero.  NOTE: The first
-				   instance number is always 0. */
-/* sizeof() = 42 bytes */
-/*
- * When (re)using the mft record, we place the update sequence array at this
- * offset, i.e. before we start with the attributes.  This also makes sense,
- * otherwise we could run into problems with the update sequence array
- * containing in itself the last two bytes of a sector which would mean that
- * multi sector transfer protection wouldn't work.  As you can't protect data
- * by overwriting it since you then can't get it back...
- * When reading we obviously use the data from the ntfs record header.
- */
-} __attribute__ ((__packed__)) MFT_RECORD_OLD;
-
-/*
- * System defined attributes (32-bit).  Each attribute type has a corresponding
- * attribute name (Unicode string of maximum 64 character length) as described
- * by the attribute definitions present in the data attribute of the $AttrDef
- * system file.  On NTFS 3.0 volumes the names are just as the types are named
- * in the below defines exchanging AT_ for the dollar sign ($).  If that is not
- * a revealing choice of symbol I do not know what is... (-;
- */
-enum {
-	AT_UNUSED			= cpu_to_le32(         0),
-	AT_STANDARD_INFORMATION		= cpu_to_le32(      0x10),
-	AT_ATTRIBUTE_LIST		= cpu_to_le32(      0x20),
-	AT_FILE_NAME			= cpu_to_le32(      0x30),
-	AT_OBJECT_ID			= cpu_to_le32(      0x40),
-	AT_SECURITY_DESCRIPTOR		= cpu_to_le32(      0x50),
-	AT_VOLUME_NAME			= cpu_to_le32(      0x60),
-	AT_VOLUME_INFORMATION		= cpu_to_le32(      0x70),
-	AT_DATA				= cpu_to_le32(      0x80),
-	AT_INDEX_ROOT			= cpu_to_le32(      0x90),
-	AT_INDEX_ALLOCATION		= cpu_to_le32(      0xa0),
-	AT_BITMAP			= cpu_to_le32(      0xb0),
-	AT_REPARSE_POINT		= cpu_to_le32(      0xc0),
-	AT_EA_INFORMATION		= cpu_to_le32(      0xd0),
-	AT_EA				= cpu_to_le32(      0xe0),
-	AT_PROPERTY_SET			= cpu_to_le32(      0xf0),
-	AT_LOGGED_UTILITY_STREAM	= cpu_to_le32(     0x100),
-	AT_FIRST_USER_DEFINED_ATTRIBUTE	= cpu_to_le32(    0x1000),
-	AT_END				= cpu_to_le32(0xffffffff)
-};
-
-typedef le32 ATTR_TYPE;
-
-/*
- * The collation rules for sorting views/indexes/etc (32-bit).
- *
- * COLLATION_BINARY - Collate by binary compare where the first byte is most
- *	significant.
- * COLLATION_UNICODE_STRING - Collate Unicode strings by comparing their binary
- *	Unicode values, except that when a character can be uppercased, the
- *	upper case value collates before the lower case one.
- * COLLATION_FILE_NAME - Collate file names as Unicode strings. The collation
- *	is done very much like COLLATION_UNICODE_STRING. In fact I have no idea
- *	what the difference is. Perhaps the difference is that file names
- *	would treat some special characters in an odd way (see
- *	unistr.c::ntfs_collate_names() and unistr.c::legal_ansi_char_array[]
- *	for what I mean but COLLATION_UNICODE_STRING would not give any special
- *	treatment to any characters at all, but this is speculation.
- * COLLATION_NTOFS_ULONG - Sorting is done according to ascending le32 key
- *	values. E.g. used for $SII index in FILE_Secure, which sorts by
- *	security_id (le32).
- * COLLATION_NTOFS_SID - Sorting is done according to ascending SID values.
- *	E.g. used for $O index in FILE_Extend/$Quota.
- * COLLATION_NTOFS_SECURITY_HASH - Sorting is done first by ascending hash
- *	values and second by ascending security_id values. E.g. used for $SDH
- *	index in FILE_Secure.
- * COLLATION_NTOFS_ULONGS - Sorting is done according to a sequence of ascending
- *	le32 key values. E.g. used for $O index in FILE_Extend/$ObjId, which
- *	sorts by object_id (16-byte), by splitting up the object_id in four
- *	le32 values and using them as individual keys. E.g. take the following
- *	two security_ids, stored as follows on disk:
- *		1st: a1 61 65 b7 65 7b d4 11 9e 3d 00 e0 81 10 42 59
- *		2nd: 38 14 37 d2 d2 f3 d4 11 a5 21 c8 6b 79 b1 97 45
- *	To compare them, they are split into four le32 values each, like so:
- *		1st: 0xb76561a1 0x11d47b65 0xe0003d9e 0x59421081
- *		2nd: 0xd2371438 0x11d4f3d2 0x6bc821a5 0x4597b179
- *	Now, it is apparent why the 2nd object_id collates after the 1st: the
- *	first le32 value of the 1st object_id is less than the first le32 of
- *	the 2nd object_id. If the first le32 values of both object_ids were
- *	equal then the second le32 values would be compared, etc.
- */
-enum {
-	COLLATION_BINARY		= cpu_to_le32(0x00),
-	COLLATION_FILE_NAME		= cpu_to_le32(0x01),
-	COLLATION_UNICODE_STRING	= cpu_to_le32(0x02),
-	COLLATION_NTOFS_ULONG		= cpu_to_le32(0x10),
-	COLLATION_NTOFS_SID		= cpu_to_le32(0x11),
-	COLLATION_NTOFS_SECURITY_HASH	= cpu_to_le32(0x12),
-	COLLATION_NTOFS_ULONGS		= cpu_to_le32(0x13),
-};
-
-typedef le32 COLLATION_RULE;
-
-/*
- * The flags (32-bit) describing attribute properties in the attribute
- * definition structure.  FIXME: This information is based on Regis's
- * information and, according to him, it is not certain and probably
- * incomplete.  The INDEXABLE flag is fairly certainly correct as only the file
- * name attribute has this flag set and this is the only attribute indexed in
- * NT4.
- */
-enum {
-	ATTR_DEF_INDEXABLE	= cpu_to_le32(0x02), /* Attribute can be
-					indexed. */
-	ATTR_DEF_MULTIPLE	= cpu_to_le32(0x04), /* Attribute type
-					can be present multiple times in the
-					mft records of an inode. */
-	ATTR_DEF_NOT_ZERO	= cpu_to_le32(0x08), /* Attribute value
-					must contain at least one non-zero
-					byte. */
-	ATTR_DEF_INDEXED_UNIQUE	= cpu_to_le32(0x10), /* Attribute must be
-					indexed and the attribute value must be
-					unique for the attribute type in all of
-					the mft records of an inode. */
-	ATTR_DEF_NAMED_UNIQUE	= cpu_to_le32(0x20), /* Attribute must be
-					named and the name must be unique for
-					the attribute type in all of the mft
-					records of an inode. */
-	ATTR_DEF_RESIDENT	= cpu_to_le32(0x40), /* Attribute must be
-					resident. */
-	ATTR_DEF_ALWAYS_LOG	= cpu_to_le32(0x80), /* Always log
-					modifications to this attribute,
-					regardless of whether it is resident or
-					non-resident.  Without this, only log
-					modifications if the attribute is
-					resident. */
-};
-
-typedef le32 ATTR_DEF_FLAGS;
-
-/*
- * The data attribute of FILE_AttrDef contains a sequence of attribute
- * definitions for the NTFS volume. With this, it is supposed to be safe for an
- * older NTFS driver to mount a volume containing a newer NTFS version without
- * damaging it (that's the theory. In practice it's: not damaging it too much).
- * Entries are sorted by attribute type. The flags describe whether the
- * attribute can be resident/non-resident and possibly other things, but the
- * actual bits are unknown.
- */
-typedef struct {
-/*hex ofs*/
-/*  0*/	ntfschar name[0x40];		/* Unicode name of the attribute. Zero
-					   terminated. */
-/* 80*/	ATTR_TYPE type;			/* Type of the attribute. */
-/* 84*/	le32 display_rule;		/* Default display rule.
-					   FIXME: What does it mean? (AIA) */
-/* 88*/ COLLATION_RULE collation_rule;	/* Default collation rule. */
-/* 8c*/	ATTR_DEF_FLAGS flags;		/* Flags describing the attribute. */
-/* 90*/	sle64 min_size;			/* Optional minimum attribute size. */
-/* 98*/	sle64 max_size;			/* Maximum size of attribute. */
-/* sizeof() = 0xa0 or 160 bytes */
-} __attribute__ ((__packed__)) ATTR_DEF;
-
-/*
- * Attribute flags (16-bit).
- */
-enum {
-	ATTR_IS_COMPRESSED    = cpu_to_le16(0x0001),
-	ATTR_COMPRESSION_MASK = cpu_to_le16(0x00ff), /* Compression method
-							      mask.  Also, first
-							      illegal value. */
-	ATTR_IS_ENCRYPTED     = cpu_to_le16(0x4000),
-	ATTR_IS_SPARSE	      = cpu_to_le16(0x8000),
-} __attribute__ ((__packed__));
-
-typedef le16 ATTR_FLAGS;
-
-/*
- * Attribute compression.
- *
- * Only the data attribute is ever compressed in the current ntfs driver in
- * Windows. Further, compression is only applied when the data attribute is
- * non-resident. Finally, to use compression, the maximum allowed cluster size
- * on a volume is 4kib.
- *
- * The compression method is based on independently compressing blocks of X
- * clusters, where X is determined from the compression_unit value found in the
- * non-resident attribute record header (more precisely: X = 2^compression_unit
- * clusters). On Windows NT/2k, X always is 16 clusters (compression_unit = 4).
- *
- * There are three different cases of how a compression block of X clusters
- * can be stored:
- *
- *   1) The data in the block is all zero (a sparse block):
- *	  This is stored as a sparse block in the runlist, i.e. the runlist
- *	  entry has length = X and lcn = -1. The mapping pairs array actually
- *	  uses a delta_lcn value length of 0, i.e. delta_lcn is not present at
- *	  all, which is then interpreted by the driver as lcn = -1.
- *	  NOTE: Even uncompressed files can be sparse on NTFS 3.0 volumes, then
- *	  the same principles apply as above, except that the length is not
- *	  restricted to being any particular value.
- *
- *   2) The data in the block is not compressed:
- *	  This happens when compression doesn't reduce the size of the block
- *	  in clusters. I.e. if compression has a small effect so that the
- *	  compressed data still occupies X clusters, then the uncompressed data
- *	  is stored in the block.
- *	  This case is recognised by the fact that the runlist entry has
- *	  length = X and lcn >= 0. The mapping pairs array stores this as
- *	  normal with a run length of X and some specific delta_lcn, i.e.
- *	  delta_lcn has to be present.
- *
- *   3) The data in the block is compressed:
- *	  The common case. This case is recognised by the fact that the run
- *	  list entry has length L < X and lcn >= 0. The mapping pairs array
- *	  stores this as normal with a run length of X and some specific
- *	  delta_lcn, i.e. delta_lcn has to be present. This runlist entry is
- *	  immediately followed by a sparse entry with length = X - L and
- *	  lcn = -1. The latter entry is to make up the vcn counting to the
- *	  full compression block size X.
- *
- * In fact, life is more complicated because adjacent entries of the same type
- * can be coalesced. This means that one has to keep track of the number of
- * clusters handled and work on a basis of X clusters at a time being one
- * block. An example: if length L > X this means that this particular runlist
- * entry contains a block of length X and part of one or more blocks of length
- * L - X. Another example: if length L < X, this does not necessarily mean that
- * the block is compressed as it might be that the lcn changes inside the block
- * and hence the following runlist entry describes the continuation of the
- * potentially compressed block. The block would be compressed if the
- * following runlist entry describes at least X - L sparse clusters, thus
- * making up the compression block length as described in point 3 above. (Of
- * course, there can be several runlist entries with small lengths so that the
- * sparse entry does not follow the first data containing entry with
- * length < X.)
- *
- * NOTE: At the end of the compressed attribute value, there most likely is not
- * just the right amount of data to make up a compression block, thus this data
- * is not even attempted to be compressed. It is just stored as is, unless
- * the number of clusters it occupies is reduced when compressed in which case
- * it is stored as a compressed compression block, complete with sparse
- * clusters at the end.
- */
-
-/*
- * Flags of resident attributes (8-bit).
- */
-enum {
-	RESIDENT_ATTR_IS_INDEXED = 0x01, /* Attribute is referenced in an index
-					    (has implications for deleting and
-					    modifying the attribute). */
-} __attribute__ ((__packed__));
-
-typedef u8 RESIDENT_ATTR_FLAGS;
-
-/*
- * Attribute record header. Always aligned to 8-byte boundary.
- */
-typedef struct {
-/*Ofs*/
-/*  0*/	ATTR_TYPE type;		/* The (32-bit) type of the attribute. */
-/*  4*/	le32 length;		/* Byte size of the resident part of the
-				   attribute (aligned to 8-byte boundary).
-				   Used to get to the next attribute. */
-/*  8*/	u8 non_resident;	/* If 0, attribute is resident.
-				   If 1, attribute is non-resident. */
-/*  9*/	u8 name_length;		/* Unicode character size of name of attribute.
-				   0 if unnamed. */
-/* 10*/	le16 name_offset;	/* If name_length != 0, the byte offset to the
-				   beginning of the name from the attribute
-				   record. Note that the name is stored as a
-				   Unicode string. When creating, place offset
-				   just at the end of the record header. Then,
-				   follow with attribute value or mapping pairs
-				   array, resident and non-resident attributes
-				   respectively, aligning to an 8-byte
-				   boundary. */
-/* 12*/	ATTR_FLAGS flags;	/* Flags describing the attribute. */
-/* 14*/	le16 instance;		/* The instance of this attribute record. This
-				   number is unique within this mft record (see
-				   MFT_RECORD/next_attribute_instance notes in
-				   mft.h for more details). */
-/* 16*/	union {
-		/* Resident attributes. */
-		struct {
-/* 16 */		le32 value_length;/* Byte size of attribute value. */
-/* 20 */		le16 value_offset;/* Byte offset of the attribute
-					     value from the start of the
-					     attribute record. When creating,
-					     align to 8-byte boundary if we
-					     have a name present as this might
-					     not have a length of a multiple
-					     of 8-bytes. */
-/* 22 */		RESIDENT_ATTR_FLAGS flags; /* See above. */
-/* 23 */		s8 reserved;	  /* Reserved/alignment to 8-byte
-					     boundary. */
-		} __attribute__ ((__packed__)) resident;
-		/* Non-resident attributes. */
-		struct {
-/* 16*/			leVCN lowest_vcn;/* Lowest valid virtual cluster number
-				for this portion of the attribute value or
-				0 if this is the only extent (usually the
-				case). - Only when an attribute list is used
-				does lowest_vcn != 0 ever occur. */
-/* 24*/			leVCN highest_vcn;/* Highest valid vcn of this extent of
-				the attribute value. - Usually there is only one
-				portion, so this usually equals the attribute
-				value size in clusters minus 1. Can be -1 for
-				zero length files. Can be 0 for "single extent"
-				attributes. */
-/* 32*/			le16 mapping_pairs_offset; /* Byte offset from the
-				beginning of the structure to the mapping pairs
-				array which contains the mappings between the
-				vcns and the logical cluster numbers (lcns).
-				When creating, place this at the end of this
-				record header aligned to 8-byte boundary. */
-/* 34*/			u8 compression_unit; /* The compression unit expressed
-				as the log to the base 2 of the number of
-				clusters in a compression unit.  0 means not
-				compressed.  (This effectively limits the
-				compression unit size to be a power of two
-				clusters.)  WinNT4 only uses a value of 4.
-				Sparse files have this set to 0 on XPSP2. */
-/* 35*/			u8 reserved[5];		/* Align to 8-byte boundary. */
-/* The sizes below are only used when lowest_vcn is zero, as otherwise it would
-   be difficult to keep them up-to-date.*/
-/* 40*/			sle64 allocated_size;	/* Byte size of disk space
-				allocated to hold the attribute value. Always
-				is a multiple of the cluster size. When a file
-				is compressed, this field is a multiple of the
-				compression block size (2^compression_unit) and
-				it represents the logically allocated space
-				rather than the actual on disk usage. For this
-				use the compressed_size (see below). */
-/* 48*/			sle64 data_size;	/* Byte size of the attribute
-				value. Can be larger than allocated_size if
-				attribute value is compressed or sparse. */
-/* 56*/			sle64 initialized_size;	/* Byte size of initialized
-				portion of the attribute value. Usually equals
-				data_size. */
-/* sizeof(uncompressed attr) = 64*/
-/* 64*/			sle64 compressed_size;	/* Byte size of the attribute
-				value after compression.  Only present when
-				compressed or sparse.  Always is a multiple of
-				the cluster size.  Represents the actual amount
-				of disk space being used on the disk. */
-/* sizeof(compressed attr) = 72*/
-		} __attribute__ ((__packed__)) non_resident;
-	} __attribute__ ((__packed__)) data;
-} __attribute__ ((__packed__)) ATTR_RECORD;
-
-typedef ATTR_RECORD ATTR_REC;
-
-/*
- * File attribute flags (32-bit) appearing in the file_attributes fields of the
- * STANDARD_INFORMATION attribute of MFT_RECORDs and the FILENAME_ATTR
- * attributes of MFT_RECORDs and directory index entries.
- *
- * All of the below flags appear in the directory index entries but only some
- * appear in the STANDARD_INFORMATION attribute whilst only some others appear
- * in the FILENAME_ATTR attribute of MFT_RECORDs.  Unless otherwise stated the
- * flags appear in all of the above.
- */
-enum {
-	FILE_ATTR_READONLY		= cpu_to_le32(0x00000001),
-	FILE_ATTR_HIDDEN		= cpu_to_le32(0x00000002),
-	FILE_ATTR_SYSTEM		= cpu_to_le32(0x00000004),
-	/* Old DOS volid. Unused in NT.	= cpu_to_le32(0x00000008), */
-
-	FILE_ATTR_DIRECTORY		= cpu_to_le32(0x00000010),
-	/* Note, FILE_ATTR_DIRECTORY is not considered valid in NT.  It is
-	   reserved for the DOS SUBDIRECTORY flag. */
-	FILE_ATTR_ARCHIVE		= cpu_to_le32(0x00000020),
-	FILE_ATTR_DEVICE		= cpu_to_le32(0x00000040),
-	FILE_ATTR_NORMAL		= cpu_to_le32(0x00000080),
-
-	FILE_ATTR_TEMPORARY		= cpu_to_le32(0x00000100),
-	FILE_ATTR_SPARSE_FILE		= cpu_to_le32(0x00000200),
-	FILE_ATTR_REPARSE_POINT		= cpu_to_le32(0x00000400),
-	FILE_ATTR_COMPRESSED		= cpu_to_le32(0x00000800),
-
-	FILE_ATTR_OFFLINE		= cpu_to_le32(0x00001000),
-	FILE_ATTR_NOT_CONTENT_INDEXED	= cpu_to_le32(0x00002000),
-	FILE_ATTR_ENCRYPTED		= cpu_to_le32(0x00004000),
-
-	FILE_ATTR_VALID_FLAGS		= cpu_to_le32(0x00007fb7),
-	/* Note, FILE_ATTR_VALID_FLAGS masks out the old DOS VolId and the
-	   FILE_ATTR_DEVICE and preserves everything else.  This mask is used
-	   to obtain all flags that are valid for reading. */
-	FILE_ATTR_VALID_SET_FLAGS	= cpu_to_le32(0x000031a7),
-	/* Note, FILE_ATTR_VALID_SET_FLAGS masks out the old DOS VolId, the
-	   F_A_DEVICE, F_A_DIRECTORY, F_A_SPARSE_FILE, F_A_REPARSE_POINT,
-	   F_A_COMPRESSED, and F_A_ENCRYPTED and preserves the rest.  This mask
-	   is used to obtain all flags that are valid for setting. */
-	/*
-	 * The flag FILE_ATTR_DUP_FILENAME_INDEX_PRESENT is present in all
-	 * FILENAME_ATTR attributes but not in the STANDARD_INFORMATION
-	 * attribute of an mft record.
-	 */
-	FILE_ATTR_DUP_FILE_NAME_INDEX_PRESENT	= cpu_to_le32(0x10000000),
-	/* Note, this is a copy of the corresponding bit from the mft record,
-	   telling us whether this is a directory or not, i.e. whether it has
-	   an index root attribute or not. */
-	FILE_ATTR_DUP_VIEW_INDEX_PRESENT	= cpu_to_le32(0x20000000),
-	/* Note, this is a copy of the corresponding bit from the mft record,
-	   telling us whether this file has a view index present (eg. object id
-	   index, quota index, one of the security indexes or the encrypting
-	   filesystem related indexes). */
-};
-
-typedef le32 FILE_ATTR_FLAGS;
-
-/*
- * NOTE on times in NTFS: All times are in MS standard time format, i.e. they
- * are the number of 100-nanosecond intervals since 1st January 1601, 00:00:00
- * universal coordinated time (UTC). (In Linux time starts 1st January 1970,
- * 00:00:00 UTC and is stored as the number of 1-second intervals since then.)
- */
-
-/*
- * Attribute: Standard information (0x10).
- *
- * NOTE: Always resident.
- * NOTE: Present in all base file records on a volume.
- * NOTE: There is conflicting information about the meaning of each of the time
- *	 fields but the meaning as defined below has been verified to be
- *	 correct by practical experimentation on Windows NT4 SP6a and is hence
- *	 assumed to be the one and only correct interpretation.
- */
-typedef struct {
-/*Ofs*/
-/*  0*/	sle64 creation_time;		/* Time file was created. Updated when
-					   a filename is changed(?). */
-/*  8*/	sle64 last_data_change_time;	/* Time the data attribute was last
-					   modified. */
-/* 16*/	sle64 last_mft_change_time;	/* Time this mft record was last
-					   modified. */
-/* 24*/	sle64 last_access_time;		/* Approximate time when the file was
-					   last accessed (obviously this is not
-					   updated on read-only volumes). In
-					   Windows this is only updated when
-					   accessed if some time delta has
-					   passed since the last update. Also,
-					   last access time updates can be
-					   disabled altogether for speed. */
-/* 32*/	FILE_ATTR_FLAGS file_attributes; /* Flags describing the file. */
-/* 36*/	union {
-	/* NTFS 1.2 */
-		struct {
-		/* 36*/	u8 reserved12[12];	/* Reserved/alignment to 8-byte
-						   boundary. */
-		} __attribute__ ((__packed__)) v1;
-	/* sizeof() = 48 bytes */
-	/* NTFS 3.x */
-		struct {
-/*
- * If a volume has been upgraded from a previous NTFS version, then these
- * fields are present only if the file has been accessed since the upgrade.
- * Recognize the difference by comparing the length of the resident attribute
- * value. If it is 48, then the following fields are missing. If it is 72 then
- * the fields are present. Maybe just check like this:
- *	if (resident.ValueLength < sizeof(STANDARD_INFORMATION)) {
- *		Assume NTFS 1.2- format.
- *		If (volume version is 3.x)
- *			Upgrade attribute to NTFS 3.x format.
- *		else
- *			Use NTFS 1.2- format for access.
- *	} else
- *		Use NTFS 3.x format for access.
- * Only problem is that it might be legal to set the length of the value to
- * arbitrarily large values thus spoiling this check. - But chkdsk probably
- * views that as a corruption, assuming that it behaves like this for all
- * attributes.
- */
-		/* 36*/	le32 maximum_versions;	/* Maximum allowed versions for
-				file. Zero if version numbering is disabled. */
-		/* 40*/	le32 version_number;	/* This file's version (if any).
-				Set to zero if maximum_versions is zero. */
-		/* 44*/	le32 class_id;		/* Class id from bidirectional
-				class id index (?). */
-		/* 48*/	le32 owner_id;		/* Owner_id of the user owning
-				the file. Translate via $Q index in FILE_Extend
-				/$Quota to the quota control entry for the user
-				owning the file. Zero if quotas are disabled. */
-		/* 52*/	le32 security_id;	/* Security_id for the file.
-				Translate via $SII index and $SDS data stream
-				in FILE_Secure to the security descriptor. */
-		/* 56*/	le64 quota_charged;	/* Byte size of the charge to
-				the quota for all streams of the file. Note: Is
-				zero if quotas are disabled. */
-		/* 64*/	leUSN usn;		/* Last update sequence number
-				of the file.  This is a direct index into the
-				transaction log file ($UsnJrnl).  It is zero if
-				the usn journal is disabled or this file has
-				not been subject to logging yet.  See usnjrnl.h
-				for details. */
-		} __attribute__ ((__packed__)) v3;
-	/* sizeof() = 72 bytes (NTFS 3.x) */
-	} __attribute__ ((__packed__)) ver;
-} __attribute__ ((__packed__)) STANDARD_INFORMATION;
-
-/*
- * Attribute: Attribute list (0x20).
- *
- * - Can be either resident or non-resident.
- * - Value consists of a sequence of variable length, 8-byte aligned,
- * ATTR_LIST_ENTRY records.
- * - The list is not terminated by anything at all! The only way to know when
- * the end is reached is to keep track of the current offset and compare it to
- * the attribute value size.
- * - The attribute list attribute contains one entry for each attribute of
- * the file in which the list is located, except for the list attribute
- * itself. The list is sorted: first by attribute type, second by attribute
- * name (if present), third by instance number. The extents of one
- * non-resident attribute (if present) immediately follow after the initial
- * extent. They are ordered by lowest_vcn and have their instace set to zero.
- * It is not allowed to have two attributes with all sorting keys equal.
- * - Further restrictions:
- *	- If not resident, the vcn to lcn mapping array has to fit inside the
- *	  base mft record.
- *	- The attribute list attribute value has a maximum size of 256kb. This
- *	  is imposed by the Windows cache manager.
- * - Attribute lists are only used when the attributes of mft record do not
- * fit inside the mft record despite all attributes (that can be made
- * non-resident) having been made non-resident. This can happen e.g. when:
- *	- File has a large number of hard links (lots of file name
- *	  attributes present).
- *	- The mapping pairs array of some non-resident attribute becomes so
- *	  large due to fragmentation that it overflows the mft record.
- *	- The security descriptor is very complex (not applicable to
- *	  NTFS 3.0 volumes).
- *	- There are many named streams.
- */
-typedef struct {
-/*Ofs*/
-/*  0*/	ATTR_TYPE type;		/* Type of referenced attribute. */
-/*  4*/	le16 length;		/* Byte size of this entry (8-byte aligned). */
-/*  6*/	u8 name_length;		/* Size in Unicode chars of the name of the
-				   attribute or 0 if unnamed. */
-/*  7*/	u8 name_offset;		/* Byte offset to beginning of attribute name
-				   (always set this to where the name would
-				   start even if unnamed). */
-/*  8*/	leVCN lowest_vcn;	/* Lowest virtual cluster number of this portion
-				   of the attribute value. This is usually 0. It
-				   is non-zero for the case where one attribute
-				   does not fit into one mft record and thus
-				   several mft records are allocated to hold
-				   this attribute. In the latter case, each mft
-				   record holds one extent of the attribute and
-				   there is one attribute list entry for each
-				   extent. NOTE: This is DEFINITELY a signed
-				   value! The windows driver uses cmp, followed
-				   by jg when comparing this, thus it treats it
-				   as signed. */
-/* 16*/	leMFT_REF mft_reference;/* The reference of the mft record holding
-				   the ATTR_RECORD for this portion of the
-				   attribute value. */
-/* 24*/	le16 instance;		/* If lowest_vcn = 0, the instance of the
-				   attribute being referenced; otherwise 0. */
-/* 26*/	ntfschar name[0];	/* Use when creating only. When reading use
-				   name_offset to determine the location of the
-				   name. */
-/* sizeof() = 26 + (attribute_name_length * 2) bytes */
-} __attribute__ ((__packed__)) ATTR_LIST_ENTRY;
-
-/*
- * The maximum allowed length for a file name.
- */
-#define MAXIMUM_FILE_NAME_LENGTH	255
-
-/*
- * Possible namespaces for filenames in ntfs (8-bit).
- */
-enum {
-	FILE_NAME_POSIX		= 0x00,
-	/* This is the largest namespace. It is case sensitive and allows all
-	   Unicode characters except for: '\0' and '/'.  Beware that in
-	   WinNT/2k/2003 by default files which eg have the same name except
-	   for their case will not be distinguished by the standard utilities
-	   and thus a "del filename" will delete both "filename" and "fileName"
-	   without warning.  However if for example Services For Unix (SFU) are
-	   installed and the case sensitive option was enabled at installation
-	   time, then you can create/access/delete such files.
-	   Note that even SFU places restrictions on the filenames beyond the
-	   '\0' and '/' and in particular the following set of characters is
-	   not allowed: '"', '/', '<', '>', '\'.  All other characters,
-	   including the ones no allowed in WIN32 namespace are allowed.
-	   Tested with SFU 3.5 (this is now free) running on Windows XP. */
-	FILE_NAME_WIN32		= 0x01,
-	/* The standard WinNT/2k NTFS long filenames. Case insensitive.  All
-	   Unicode chars except: '\0', '"', '*', '/', ':', '<', '>', '?', '\',
-	   and '|'.  Further, names cannot end with a '.' or a space. */
-	FILE_NAME_DOS		= 0x02,
-	/* The standard DOS filenames (8.3 format). Uppercase only.  All 8-bit
-	   characters greater space, except: '"', '*', '+', ',', '/', ':', ';',
-	   '<', '=', '>', '?', and '\'. */
-	FILE_NAME_WIN32_AND_DOS	= 0x03,
-	/* 3 means that both the Win32 and the DOS filenames are identical and
-	   hence have been saved in this single filename record. */
-} __attribute__ ((__packed__));
-
-typedef u8 FILE_NAME_TYPE_FLAGS;
-
-/*
- * Attribute: Filename (0x30).
- *
- * NOTE: Always resident.
- * NOTE: All fields, except the parent_directory, are only updated when the
- *	 filename is changed. Until then, they just become out of sync with
- *	 reality and the more up to date values are present in the standard
- *	 information attribute.
- * NOTE: There is conflicting information about the meaning of each of the time
- *	 fields but the meaning as defined below has been verified to be
- *	 correct by practical experimentation on Windows NT4 SP6a and is hence
- *	 assumed to be the one and only correct interpretation.
- */
-typedef struct {
-/*hex ofs*/
-/*  0*/	leMFT_REF parent_directory;	/* Directory this filename is
-					   referenced from. */
-/*  8*/	sle64 creation_time;		/* Time file was created. */
-/* 10*/	sle64 last_data_change_time;	/* Time the data attribute was last
-					   modified. */
-/* 18*/	sle64 last_mft_change_time;	/* Time this mft record was last
-					   modified. */
-/* 20*/	sle64 last_access_time;		/* Time this mft record was last
-					   accessed. */
-/* 28*/	sle64 allocated_size;		/* Byte size of on-disk allocated space
-					   for the unnamed data attribute.  So
-					   for normal $DATA, this is the
-					   allocated_size from the unnamed
-					   $DATA attribute and for compressed
-					   and/or sparse $DATA, this is the
-					   compressed_size from the unnamed
-					   $DATA attribute.  For a directory or
-					   other inode without an unnamed $DATA
-					   attribute, this is always 0.  NOTE:
-					   This is a multiple of the cluster
-					   size. */
-/* 30*/	sle64 data_size;		/* Byte size of actual data in unnamed
-					   data attribute.  For a directory or
-					   other inode without an unnamed $DATA
-					   attribute, this is always 0. */
-/* 38*/	FILE_ATTR_FLAGS file_attributes;	/* Flags describing the file. */
-/* 3c*/	union {
-	/* 3c*/	struct {
-		/* 3c*/	le16 packed_ea_size;	/* Size of the buffer needed to
-						   pack the extended attributes
-						   (EAs), if such are present.*/
-		/* 3e*/	le16 reserved;		/* Reserved for alignment. */
-		} __attribute__ ((__packed__)) ea;
-	/* 3c*/	struct {
-		/* 3c*/	le32 reparse_point_tag;	/* Type of reparse point,
-						   present only in reparse
-						   points and only if there are
-						   no EAs. */
-		} __attribute__ ((__packed__)) rp;
-	} __attribute__ ((__packed__)) type;
-/* 40*/	u8 file_name_length;			/* Length of file name in
-						   (Unicode) characters. */
-/* 41*/	FILE_NAME_TYPE_FLAGS file_name_type;	/* Namespace of the file name.*/
-/* 42*/	ntfschar file_name[0];			/* File name in Unicode. */
-} __attribute__ ((__packed__)) FILE_NAME_ATTR;
-
-/*
- * GUID structures store globally unique identifiers (GUID). A GUID is a
- * 128-bit value consisting of one group of eight hexadecimal digits, followed
- * by three groups of four hexadecimal digits each, followed by one group of
- * twelve hexadecimal digits. GUIDs are Microsoft's implementation of the
- * distributed computing environment (DCE) universally unique identifier (UUID).
- * Example of a GUID:
- *	1F010768-5A73-BC91-0010A52216A7
- */
-typedef struct {
-	le32 data1;	/* The first eight hexadecimal digits of the GUID. */
-	le16 data2;	/* The first group of four hexadecimal digits. */
-	le16 data3;	/* The second group of four hexadecimal digits. */
-	u8 data4[8];	/* The first two bytes are the third group of four
-			   hexadecimal digits. The remaining six bytes are the
-			   final 12 hexadecimal digits. */
-} __attribute__ ((__packed__)) GUID;
-
-/*
- * FILE_Extend/$ObjId contains an index named $O. This index contains all
- * object_ids present on the volume as the index keys and the corresponding
- * mft_record numbers as the index entry data parts. The data part (defined
- * below) also contains three other object_ids:
- *	birth_volume_id - object_id of FILE_Volume on which the file was first
- *			  created. Optional (i.e. can be zero).
- *	birth_object_id - object_id of file when it was first created. Usually
- *			  equals the object_id. Optional (i.e. can be zero).
- *	domain_id	- Reserved (always zero).
- */
-typedef struct {
-	leMFT_REF mft_reference;/* Mft record containing the object_id in
-				   the index entry key. */
-	union {
-		struct {
-			GUID birth_volume_id;
-			GUID birth_object_id;
-			GUID domain_id;
-		} __attribute__ ((__packed__)) origin;
-		u8 extended_info[48];
-	} __attribute__ ((__packed__)) opt;
-} __attribute__ ((__packed__)) OBJ_ID_INDEX_DATA;
-
-/*
- * Attribute: Object id (NTFS 3.0+) (0x40).
- *
- * NOTE: Always resident.
- */
-typedef struct {
-	GUID object_id;				/* Unique id assigned to the
-						   file.*/
-	/* The following fields are optional. The attribute value size is 16
-	   bytes, i.e. sizeof(GUID), if these are not present at all. Note,
-	   the entries can be present but one or more (or all) can be zero
-	   meaning that that particular value(s) is(are) not defined. */
-	union {
-		struct {
-			GUID birth_volume_id;	/* Unique id of volume on which
-						   the file was first created.*/
-			GUID birth_object_id;	/* Unique id of file when it was
-						   first created. */
-			GUID domain_id;		/* Reserved, zero. */
-		} __attribute__ ((__packed__)) origin;
-		u8 extended_info[48];
-	} __attribute__ ((__packed__)) opt;
-} __attribute__ ((__packed__)) OBJECT_ID_ATTR;
-
-/*
- * The pre-defined IDENTIFIER_AUTHORITIES used as SID_IDENTIFIER_AUTHORITY in
- * the SID structure (see below).
- */
-//typedef enum {					/* SID string prefix. */
-//	SECURITY_NULL_SID_AUTHORITY	= {0, 0, 0, 0, 0, 0},	/* S-1-0 */
-//	SECURITY_WORLD_SID_AUTHORITY	= {0, 0, 0, 0, 0, 1},	/* S-1-1 */
-//	SECURITY_LOCAL_SID_AUTHORITY	= {0, 0, 0, 0, 0, 2},	/* S-1-2 */
-//	SECURITY_CREATOR_SID_AUTHORITY	= {0, 0, 0, 0, 0, 3},	/* S-1-3 */
-//	SECURITY_NON_UNIQUE_AUTHORITY	= {0, 0, 0, 0, 0, 4},	/* S-1-4 */
-//	SECURITY_NT_SID_AUTHORITY	= {0, 0, 0, 0, 0, 5},	/* S-1-5 */
-//} IDENTIFIER_AUTHORITIES;
-
-/*
- * These relative identifiers (RIDs) are used with the above identifier
- * authorities to make up universal well-known SIDs.
- *
- * Note: The relative identifier (RID) refers to the portion of a SID, which
- * identifies a user or group in relation to the authority that issued the SID.
- * For example, the universal well-known SID Creator Owner ID (S-1-3-0) is
- * made up of the identifier authority SECURITY_CREATOR_SID_AUTHORITY (3) and
- * the relative identifier SECURITY_CREATOR_OWNER_RID (0).
- */
-typedef enum {					/* Identifier authority. */
-	SECURITY_NULL_RID		  = 0,	/* S-1-0 */
-	SECURITY_WORLD_RID		  = 0,	/* S-1-1 */
-	SECURITY_LOCAL_RID		  = 0,	/* S-1-2 */
-
-	SECURITY_CREATOR_OWNER_RID	  = 0,	/* S-1-3 */
-	SECURITY_CREATOR_GROUP_RID	  = 1,	/* S-1-3 */
-
-	SECURITY_CREATOR_OWNER_SERVER_RID = 2,	/* S-1-3 */
-	SECURITY_CREATOR_GROUP_SERVER_RID = 3,	/* S-1-3 */
-
-	SECURITY_DIALUP_RID		  = 1,
-	SECURITY_NETWORK_RID		  = 2,
-	SECURITY_BATCH_RID		  = 3,
-	SECURITY_INTERACTIVE_RID	  = 4,
-	SECURITY_SERVICE_RID		  = 6,
-	SECURITY_ANONYMOUS_LOGON_RID	  = 7,
-	SECURITY_PROXY_RID		  = 8,
-	SECURITY_ENTERPRISE_CONTROLLERS_RID=9,
-	SECURITY_SERVER_LOGON_RID	  = 9,
-	SECURITY_PRINCIPAL_SELF_RID	  = 0xa,
-	SECURITY_AUTHENTICATED_USER_RID	  = 0xb,
-	SECURITY_RESTRICTED_CODE_RID	  = 0xc,
-	SECURITY_TERMINAL_SERVER_RID	  = 0xd,
-
-	SECURITY_LOGON_IDS_RID		  = 5,
-	SECURITY_LOGON_IDS_RID_COUNT	  = 3,
-
-	SECURITY_LOCAL_SYSTEM_RID	  = 0x12,
-
-	SECURITY_NT_NON_UNIQUE		  = 0x15,
-
-	SECURITY_BUILTIN_DOMAIN_RID	  = 0x20,
-
-	/*
-	 * Well-known domain relative sub-authority values (RIDs).
-	 */
-
-	/* Users. */
-	DOMAIN_USER_RID_ADMIN		  = 0x1f4,
-	DOMAIN_USER_RID_GUEST		  = 0x1f5,
-	DOMAIN_USER_RID_KRBTGT		  = 0x1f6,
-
-	/* Groups. */
-	DOMAIN_GROUP_RID_ADMINS		  = 0x200,
-	DOMAIN_GROUP_RID_USERS		  = 0x201,
-	DOMAIN_GROUP_RID_GUESTS		  = 0x202,
-	DOMAIN_GROUP_RID_COMPUTERS	  = 0x203,
-	DOMAIN_GROUP_RID_CONTROLLERS	  = 0x204,
-	DOMAIN_GROUP_RID_CERT_ADMINS	  = 0x205,
-	DOMAIN_GROUP_RID_SCHEMA_ADMINS	  = 0x206,
-	DOMAIN_GROUP_RID_ENTERPRISE_ADMINS= 0x207,
-	DOMAIN_GROUP_RID_POLICY_ADMINS	  = 0x208,
-
-	/* Aliases. */
-	DOMAIN_ALIAS_RID_ADMINS		  = 0x220,
-	DOMAIN_ALIAS_RID_USERS		  = 0x221,
-	DOMAIN_ALIAS_RID_GUESTS		  = 0x222,
-	DOMAIN_ALIAS_RID_POWER_USERS	  = 0x223,
-
-	DOMAIN_ALIAS_RID_ACCOUNT_OPS	  = 0x224,
-	DOMAIN_ALIAS_RID_SYSTEM_OPS	  = 0x225,
-	DOMAIN_ALIAS_RID_PRINT_OPS	  = 0x226,
-	DOMAIN_ALIAS_RID_BACKUP_OPS	  = 0x227,
-
-	DOMAIN_ALIAS_RID_REPLICATOR	  = 0x228,
-	DOMAIN_ALIAS_RID_RAS_SERVERS	  = 0x229,
-	DOMAIN_ALIAS_RID_PREW2KCOMPACCESS = 0x22a,
-} RELATIVE_IDENTIFIERS;
-
-/*
- * The universal well-known SIDs:
- *
- *	NULL_SID			S-1-0-0
- *	WORLD_SID			S-1-1-0
- *	LOCAL_SID			S-1-2-0
- *	CREATOR_OWNER_SID		S-1-3-0
- *	CREATOR_GROUP_SID		S-1-3-1
- *	CREATOR_OWNER_SERVER_SID	S-1-3-2
- *	CREATOR_GROUP_SERVER_SID	S-1-3-3
- *
- *	(Non-unique IDs)		S-1-4
- *
- * NT well-known SIDs:
- *
- *	NT_AUTHORITY_SID	S-1-5
- *	DIALUP_SID		S-1-5-1
- *
- *	NETWORD_SID		S-1-5-2
- *	BATCH_SID		S-1-5-3
- *	INTERACTIVE_SID		S-1-5-4
- *	SERVICE_SID		S-1-5-6
- *	ANONYMOUS_LOGON_SID	S-1-5-7		(aka null logon session)
- *	PROXY_SID		S-1-5-8
- *	SERVER_LOGON_SID	S-1-5-9		(aka domain controller account)
- *	SELF_SID		S-1-5-10	(self RID)
- *	AUTHENTICATED_USER_SID	S-1-5-11
- *	RESTRICTED_CODE_SID	S-1-5-12	(running restricted code)
- *	TERMINAL_SERVER_SID	S-1-5-13	(running on terminal server)
- *
- *	(Logon IDs)		S-1-5-5-X-Y
- *
- *	(NT non-unique IDs)	S-1-5-0x15-...
- *
- *	(Built-in domain)	S-1-5-0x20
- */
-
-/*
- * The SID_IDENTIFIER_AUTHORITY is a 48-bit value used in the SID structure.
- *
- * NOTE: This is stored as a big endian number, hence the high_part comes
- * before the low_part.
- */
-typedef union {
-	struct {
-		u16 high_part;	/* High 16-bits. */
-		u32 low_part;	/* Low 32-bits. */
-	} __attribute__ ((__packed__)) parts;
-	u8 value[6];		/* Value as individual bytes. */
-} __attribute__ ((__packed__)) SID_IDENTIFIER_AUTHORITY;
-
-/*
- * The SID structure is a variable-length structure used to uniquely identify
- * users or groups. SID stands for security identifier.
- *
- * The standard textual representation of the SID is of the form:
- *	S-R-I-S-S...
- * Where:
- *    - The first "S" is the literal character 'S' identifying the following
- *	digits as a SID.
- *    - R is the revision level of the SID expressed as a sequence of digits
- *	either in decimal or hexadecimal (if the later, prefixed by "0x").
- *    - I is the 48-bit identifier_authority, expressed as digits as R above.
- *    - S... is one or more sub_authority values, expressed as digits as above.
- *
- * Example SID; the domain-relative SID of the local Administrators group on
- * Windows NT/2k:
- *	S-1-5-32-544
- * This translates to a SID with:
- *	revision = 1,
- *	sub_authority_count = 2,
- *	identifier_authority = {0,0,0,0,0,5},	// SECURITY_NT_AUTHORITY
- *	sub_authority[0] = 32,			// SECURITY_BUILTIN_DOMAIN_RID
- *	sub_authority[1] = 544			// DOMAIN_ALIAS_RID_ADMINS
- */
-typedef struct {
-	u8 revision;
-	u8 sub_authority_count;
-	SID_IDENTIFIER_AUTHORITY identifier_authority;
-	le32 sub_authority[1];		/* At least one sub_authority. */
-} __attribute__ ((__packed__)) SID;
-
-/*
- * Current constants for SIDs.
- */
-typedef enum {
-	SID_REVISION			=  1,	/* Current revision level. */
-	SID_MAX_SUB_AUTHORITIES		= 15,	/* Maximum number of those. */
-	SID_RECOMMENDED_SUB_AUTHORITIES	=  1,	/* Will change to around 6 in
-						   a future revision. */
-} SID_CONSTANTS;
-
-/*
- * The predefined ACE types (8-bit, see below).
- */
-enum {
-	ACCESS_MIN_MS_ACE_TYPE		= 0,
-	ACCESS_ALLOWED_ACE_TYPE		= 0,
-	ACCESS_DENIED_ACE_TYPE		= 1,
-	SYSTEM_AUDIT_ACE_TYPE		= 2,
-	SYSTEM_ALARM_ACE_TYPE		= 3, /* Not implemented as of Win2k. */
-	ACCESS_MAX_MS_V2_ACE_TYPE	= 3,
-
-	ACCESS_ALLOWED_COMPOUND_ACE_TYPE= 4,
-	ACCESS_MAX_MS_V3_ACE_TYPE	= 4,
-
-	/* The following are Win2k only. */
-	ACCESS_MIN_MS_OBJECT_ACE_TYPE	= 5,
-	ACCESS_ALLOWED_OBJECT_ACE_TYPE	= 5,
-	ACCESS_DENIED_OBJECT_ACE_TYPE	= 6,
-	SYSTEM_AUDIT_OBJECT_ACE_TYPE	= 7,
-	SYSTEM_ALARM_OBJECT_ACE_TYPE	= 8,
-	ACCESS_MAX_MS_OBJECT_ACE_TYPE	= 8,
-
-	ACCESS_MAX_MS_V4_ACE_TYPE	= 8,
-
-	/* This one is for WinNT/2k. */
-	ACCESS_MAX_MS_ACE_TYPE		= 8,
-} __attribute__ ((__packed__));
-
-typedef u8 ACE_TYPES;
-
-/*
- * The ACE flags (8-bit) for audit and inheritance (see below).
- *
- * SUCCESSFUL_ACCESS_ACE_FLAG is only used with system audit and alarm ACE
- * types to indicate that a message is generated (in Windows!) for successful
- * accesses.
- *
- * FAILED_ACCESS_ACE_FLAG is only used with system audit and alarm ACE types
- * to indicate that a message is generated (in Windows!) for failed accesses.
- */
-enum {
-	/* The inheritance flags. */
-	OBJECT_INHERIT_ACE		= 0x01,
-	CONTAINER_INHERIT_ACE		= 0x02,
-	NO_PROPAGATE_INHERIT_ACE	= 0x04,
-	INHERIT_ONLY_ACE		= 0x08,
-	INHERITED_ACE			= 0x10,	/* Win2k only. */
-	VALID_INHERIT_FLAGS		= 0x1f,
-
-	/* The audit flags. */
-	SUCCESSFUL_ACCESS_ACE_FLAG	= 0x40,
-	FAILED_ACCESS_ACE_FLAG		= 0x80,
-} __attribute__ ((__packed__));
-
-typedef u8 ACE_FLAGS;
-
-/*
- * An ACE is an access-control entry in an access-control list (ACL).
- * An ACE defines access to an object for a specific user or group or defines
- * the types of access that generate system-administration messages or alarms
- * for a specific user or group. The user or group is identified by a security
- * identifier (SID).
- *
- * Each ACE starts with an ACE_HEADER structure (aligned on 4-byte boundary),
- * which specifies the type and size of the ACE. The format of the subsequent
- * data depends on the ACE type.
- */
-typedef struct {
-/*Ofs*/
-/*  0*/	ACE_TYPES type;		/* Type of the ACE. */
-/*  1*/	ACE_FLAGS flags;	/* Flags describing the ACE. */
-/*  2*/	le16 size;		/* Size in bytes of the ACE. */
-} __attribute__ ((__packed__)) ACE_HEADER;
-
-/*
- * The access mask (32-bit). Defines the access rights.
- *
- * The specific rights (bits 0 to 15).  These depend on the type of the object
- * being secured by the ACE.
- */
-enum {
-	/* Specific rights for files and directories are as follows: */
-
-	/* Right to read data from the file. (FILE) */
-	FILE_READ_DATA			= cpu_to_le32(0x00000001),
-	/* Right to list contents of a directory. (DIRECTORY) */
-	FILE_LIST_DIRECTORY		= cpu_to_le32(0x00000001),
-
-	/* Right to write data to the file. (FILE) */
-	FILE_WRITE_DATA			= cpu_to_le32(0x00000002),
-	/* Right to create a file in the directory. (DIRECTORY) */
-	FILE_ADD_FILE			= cpu_to_le32(0x00000002),
-
-	/* Right to append data to the file. (FILE) */
-	FILE_APPEND_DATA		= cpu_to_le32(0x00000004),
-	/* Right to create a subdirectory. (DIRECTORY) */
-	FILE_ADD_SUBDIRECTORY		= cpu_to_le32(0x00000004),
-
-	/* Right to read extended attributes. (FILE/DIRECTORY) */
-	FILE_READ_EA			= cpu_to_le32(0x00000008),
-
-	/* Right to write extended attributes. (FILE/DIRECTORY) */
-	FILE_WRITE_EA			= cpu_to_le32(0x00000010),
-
-	/* Right to execute a file. (FILE) */
-	FILE_EXECUTE			= cpu_to_le32(0x00000020),
-	/* Right to traverse the directory. (DIRECTORY) */
-	FILE_TRAVERSE			= cpu_to_le32(0x00000020),
-
-	/*
-	 * Right to delete a directory and all the files it contains (its
-	 * children), even if the files are read-only. (DIRECTORY)
-	 */
-	FILE_DELETE_CHILD		= cpu_to_le32(0x00000040),
-
-	/* Right to read file attributes. (FILE/DIRECTORY) */
-	FILE_READ_ATTRIBUTES		= cpu_to_le32(0x00000080),
-
-	/* Right to change file attributes. (FILE/DIRECTORY) */
-	FILE_WRITE_ATTRIBUTES		= cpu_to_le32(0x00000100),
-
-	/*
-	 * The standard rights (bits 16 to 23).  These are independent of the
-	 * type of object being secured.
-	 */
-
-	/* Right to delete the object. */
-	DELETE				= cpu_to_le32(0x00010000),
-
-	/*
-	 * Right to read the information in the object's security descriptor,
-	 * not including the information in the SACL, i.e. right to read the
-	 * security descriptor and owner.
-	 */
-	READ_CONTROL			= cpu_to_le32(0x00020000),
-
-	/* Right to modify the DACL in the object's security descriptor. */
-	WRITE_DAC			= cpu_to_le32(0x00040000),
-
-	/* Right to change the owner in the object's security descriptor. */
-	WRITE_OWNER			= cpu_to_le32(0x00080000),
-
-	/*
-	 * Right to use the object for synchronization.  Enables a process to
-	 * wait until the object is in the signalled state.  Some object types
-	 * do not support this access right.
-	 */
-	SYNCHRONIZE			= cpu_to_le32(0x00100000),
-
-	/*
-	 * The following STANDARD_RIGHTS_* are combinations of the above for
-	 * convenience and are defined by the Win32 API.
-	 */
-
-	/* These are currently defined to READ_CONTROL. */
-	STANDARD_RIGHTS_READ		= cpu_to_le32(0x00020000),
-	STANDARD_RIGHTS_WRITE		= cpu_to_le32(0x00020000),
-	STANDARD_RIGHTS_EXECUTE		= cpu_to_le32(0x00020000),
-
-	/* Combines DELETE, READ_CONTROL, WRITE_DAC, and WRITE_OWNER access. */
-	STANDARD_RIGHTS_REQUIRED	= cpu_to_le32(0x000f0000),
-
-	/*
-	 * Combines DELETE, READ_CONTROL, WRITE_DAC, WRITE_OWNER, and
-	 * SYNCHRONIZE access.
-	 */
-	STANDARD_RIGHTS_ALL		= cpu_to_le32(0x001f0000),
-
-	/*
-	 * The access system ACL and maximum allowed access types (bits 24 to
-	 * 25, bits 26 to 27 are reserved).
-	 */
-	ACCESS_SYSTEM_SECURITY		= cpu_to_le32(0x01000000),
-	MAXIMUM_ALLOWED			= cpu_to_le32(0x02000000),
-
-	/*
-	 * The generic rights (bits 28 to 31).  These map onto the standard and
-	 * specific rights.
-	 */
-
-	/* Read, write, and execute access. */
-	GENERIC_ALL			= cpu_to_le32(0x10000000),
-
-	/* Execute access. */
-	GENERIC_EXECUTE			= cpu_to_le32(0x20000000),
-
-	/*
-	 * Write access.  For files, this maps onto:
-	 *	FILE_APPEND_DATA | FILE_WRITE_ATTRIBUTES | FILE_WRITE_DATA |
-	 *	FILE_WRITE_EA | STANDARD_RIGHTS_WRITE | SYNCHRONIZE
-	 * For directories, the mapping has the same numerical value.  See
-	 * above for the descriptions of the rights granted.
-	 */
-	GENERIC_WRITE			= cpu_to_le32(0x40000000),
-
-	/*
-	 * Read access.  For files, this maps onto:
-	 *	FILE_READ_ATTRIBUTES | FILE_READ_DATA | FILE_READ_EA |
-	 *	STANDARD_RIGHTS_READ | SYNCHRONIZE
-	 * For directories, the mapping has the same numberical value.  See
-	 * above for the descriptions of the rights granted.
-	 */
-	GENERIC_READ			= cpu_to_le32(0x80000000),
-};
-
-typedef le32 ACCESS_MASK;
-
-/*
- * The generic mapping array. Used to denote the mapping of each generic
- * access right to a specific access mask.
- *
- * FIXME: What exactly is this and what is it for? (AIA)
- */
-typedef struct {
-	ACCESS_MASK generic_read;
-	ACCESS_MASK generic_write;
-	ACCESS_MASK generic_execute;
-	ACCESS_MASK generic_all;
-} __attribute__ ((__packed__)) GENERIC_MAPPING;
-
-/*
- * The predefined ACE type structures are as defined below.
- */
-
-/*
- * ACCESS_ALLOWED_ACE, ACCESS_DENIED_ACE, SYSTEM_AUDIT_ACE, SYSTEM_ALARM_ACE
- */
-typedef struct {
-/*  0	ACE_HEADER; -- Unfolded here as gcc doesn't like unnamed structs. */
-	ACE_TYPES type;		/* Type of the ACE. */
-	ACE_FLAGS flags;	/* Flags describing the ACE. */
-	le16 size;		/* Size in bytes of the ACE. */
-/*  4*/	ACCESS_MASK mask;	/* Access mask associated with the ACE. */
-
-/*  8*/	SID sid;		/* The SID associated with the ACE. */
-} __attribute__ ((__packed__)) ACCESS_ALLOWED_ACE, ACCESS_DENIED_ACE,
-			       SYSTEM_AUDIT_ACE, SYSTEM_ALARM_ACE;
-
-/*
- * The object ACE flags (32-bit).
- */
-enum {
-	ACE_OBJECT_TYPE_PRESENT			= cpu_to_le32(1),
-	ACE_INHERITED_OBJECT_TYPE_PRESENT	= cpu_to_le32(2),
-};
-
-typedef le32 OBJECT_ACE_FLAGS;
-
-typedef struct {
-/*  0	ACE_HEADER; -- Unfolded here as gcc doesn't like unnamed structs. */
-	ACE_TYPES type;		/* Type of the ACE. */
-	ACE_FLAGS flags;	/* Flags describing the ACE. */
-	le16 size;		/* Size in bytes of the ACE. */
-/*  4*/	ACCESS_MASK mask;	/* Access mask associated with the ACE. */
-
-/*  8*/	OBJECT_ACE_FLAGS object_flags;	/* Flags describing the object ACE. */
-/* 12*/	GUID object_type;
-/* 28*/	GUID inherited_object_type;
-
-/* 44*/	SID sid;		/* The SID associated with the ACE. */
-} __attribute__ ((__packed__)) ACCESS_ALLOWED_OBJECT_ACE,
-			       ACCESS_DENIED_OBJECT_ACE,
-			       SYSTEM_AUDIT_OBJECT_ACE,
-			       SYSTEM_ALARM_OBJECT_ACE;
-
-/*
- * An ACL is an access-control list (ACL).
- * An ACL starts with an ACL header structure, which specifies the size of
- * the ACL and the number of ACEs it contains. The ACL header is followed by
- * zero or more access control entries (ACEs). The ACL as well as each ACE
- * are aligned on 4-byte boundaries.
- */
-typedef struct {
-	u8 revision;	/* Revision of this ACL. */
-	u8 alignment1;
-	le16 size;	/* Allocated space in bytes for ACL. Includes this
-			   header, the ACEs and the remaining free space. */
-	le16 ace_count;	/* Number of ACEs in the ACL. */
-	le16 alignment2;
-/* sizeof() = 8 bytes */
-} __attribute__ ((__packed__)) ACL;
-
-/*
- * Current constants for ACLs.
- */
-typedef enum {
-	/* Current revision. */
-	ACL_REVISION		= 2,
-	ACL_REVISION_DS		= 4,
-
-	/* History of revisions. */
-	ACL_REVISION1		= 1,
-	MIN_ACL_REVISION	= 2,
-	ACL_REVISION2		= 2,
-	ACL_REVISION3		= 3,
-	ACL_REVISION4		= 4,
-	MAX_ACL_REVISION	= 4,
-} ACL_CONSTANTS;
-
-/*
- * The security descriptor control flags (16-bit).
- *
- * SE_OWNER_DEFAULTED - This boolean flag, when set, indicates that the SID
- *	pointed to by the Owner field was provided by a defaulting mechanism
- *	rather than explicitly provided by the original provider of the
- *	security descriptor.  This may affect the treatment of the SID with
- *	respect to inheritance of an owner.
- *
- * SE_GROUP_DEFAULTED - This boolean flag, when set, indicates that the SID in
- *	the Group field was provided by a defaulting mechanism rather than
- *	explicitly provided by the original provider of the security
- *	descriptor.  This may affect the treatment of the SID with respect to
- *	inheritance of a primary group.
- *
- * SE_DACL_PRESENT - This boolean flag, when set, indicates that the security
- *	descriptor contains a discretionary ACL.  If this flag is set and the
- *	Dacl field of the SECURITY_DESCRIPTOR is null, then a null ACL is
- *	explicitly being specified.
- *
- * SE_DACL_DEFAULTED - This boolean flag, when set, indicates that the ACL
- *	pointed to by the Dacl field was provided by a defaulting mechanism
- *	rather than explicitly provided by the original provider of the
- *	security descriptor.  This may affect the treatment of the ACL with
- *	respect to inheritance of an ACL.  This flag is ignored if the
- *	DaclPresent flag is not set.
- *
- * SE_SACL_PRESENT - This boolean flag, when set,  indicates that the security
- *	descriptor contains a system ACL pointed to by the Sacl field.  If this
- *	flag is set and the Sacl field of the SECURITY_DESCRIPTOR is null, then
- *	an empty (but present) ACL is being specified.
- *
- * SE_SACL_DEFAULTED - This boolean flag, when set, indicates that the ACL
- *	pointed to by the Sacl field was provided by a defaulting mechanism
- *	rather than explicitly provided by the original provider of the
- *	security descriptor.  This may affect the treatment of the ACL with
- *	respect to inheritance of an ACL.  This flag is ignored if the
- *	SaclPresent flag is not set.
- *
- * SE_SELF_RELATIVE - This boolean flag, when set, indicates that the security
- *	descriptor is in self-relative form.  In this form, all fields of the
- *	security descriptor are contiguous in memory and all pointer fields are
- *	expressed as offsets from the beginning of the security descriptor.
- */
-enum {
-	SE_OWNER_DEFAULTED		= cpu_to_le16(0x0001),
-	SE_GROUP_DEFAULTED		= cpu_to_le16(0x0002),
-	SE_DACL_PRESENT			= cpu_to_le16(0x0004),
-	SE_DACL_DEFAULTED		= cpu_to_le16(0x0008),
-
-	SE_SACL_PRESENT			= cpu_to_le16(0x0010),
-	SE_SACL_DEFAULTED		= cpu_to_le16(0x0020),
-
-	SE_DACL_AUTO_INHERIT_REQ	= cpu_to_le16(0x0100),
-	SE_SACL_AUTO_INHERIT_REQ	= cpu_to_le16(0x0200),
-	SE_DACL_AUTO_INHERITED		= cpu_to_le16(0x0400),
-	SE_SACL_AUTO_INHERITED		= cpu_to_le16(0x0800),
-
-	SE_DACL_PROTECTED		= cpu_to_le16(0x1000),
-	SE_SACL_PROTECTED		= cpu_to_le16(0x2000),
-	SE_RM_CONTROL_VALID		= cpu_to_le16(0x4000),
-	SE_SELF_RELATIVE		= cpu_to_le16(0x8000)
-} __attribute__ ((__packed__));
-
-typedef le16 SECURITY_DESCRIPTOR_CONTROL;
-
-/*
- * Self-relative security descriptor. Contains the owner and group SIDs as well
- * as the sacl and dacl ACLs inside the security descriptor itself.
- */
-typedef struct {
-	u8 revision;	/* Revision level of the security descriptor. */
-	u8 alignment;
-	SECURITY_DESCRIPTOR_CONTROL control; /* Flags qualifying the type of
-			   the descriptor as well as the following fields. */
-	le32 owner;	/* Byte offset to a SID representing an object's
-			   owner. If this is NULL, no owner SID is present in
-			   the descriptor. */
-	le32 group;	/* Byte offset to a SID representing an object's
-			   primary group. If this is NULL, no primary group
-			   SID is present in the descriptor. */
-	le32 sacl;	/* Byte offset to a system ACL. Only valid, if
-			   SE_SACL_PRESENT is set in the control field. If
-			   SE_SACL_PRESENT is set but sacl is NULL, a NULL ACL
-			   is specified. */
-	le32 dacl;	/* Byte offset to a discretionary ACL. Only valid, if
-			   SE_DACL_PRESENT is set in the control field. If
-			   SE_DACL_PRESENT is set but dacl is NULL, a NULL ACL
-			   (unconditionally granting access) is specified. */
-/* sizeof() = 0x14 bytes */
-} __attribute__ ((__packed__)) SECURITY_DESCRIPTOR_RELATIVE;
-
-/*
- * Absolute security descriptor. Does not contain the owner and group SIDs, nor
- * the sacl and dacl ACLs inside the security descriptor. Instead, it contains
- * pointers to these structures in memory. Obviously, absolute security
- * descriptors are only useful for in memory representations of security
- * descriptors. On disk, a self-relative security descriptor is used.
- */
-typedef struct {
-	u8 revision;	/* Revision level of the security descriptor. */
-	u8 alignment;
-	SECURITY_DESCRIPTOR_CONTROL control;	/* Flags qualifying the type of
-			   the descriptor as well as the following fields. */
-	SID *owner;	/* Points to a SID representing an object's owner. If
-			   this is NULL, no owner SID is present in the
-			   descriptor. */
-	SID *group;	/* Points to a SID representing an object's primary
-			   group. If this is NULL, no primary group SID is
-			   present in the descriptor. */
-	ACL *sacl;	/* Points to a system ACL. Only valid, if
-			   SE_SACL_PRESENT is set in the control field. If
-			   SE_SACL_PRESENT is set but sacl is NULL, a NULL ACL
-			   is specified. */
-	ACL *dacl;	/* Points to a discretionary ACL. Only valid, if
-			   SE_DACL_PRESENT is set in the control field. If
-			   SE_DACL_PRESENT is set but dacl is NULL, a NULL ACL
-			   (unconditionally granting access) is specified. */
-} __attribute__ ((__packed__)) SECURITY_DESCRIPTOR;
-
-/*
- * Current constants for security descriptors.
- */
-typedef enum {
-	/* Current revision. */
-	SECURITY_DESCRIPTOR_REVISION	= 1,
-	SECURITY_DESCRIPTOR_REVISION1	= 1,
-
-	/* The sizes of both the absolute and relative security descriptors is
-	   the same as pointers, at least on ia32 architecture are 32-bit. */
-	SECURITY_DESCRIPTOR_MIN_LENGTH	= sizeof(SECURITY_DESCRIPTOR),
-} SECURITY_DESCRIPTOR_CONSTANTS;
-
-/*
- * Attribute: Security descriptor (0x50). A standard self-relative security
- * descriptor.
- *
- * NOTE: Can be resident or non-resident.
- * NOTE: Not used in NTFS 3.0+, as security descriptors are stored centrally
- * in FILE_Secure and the correct descriptor is found using the security_id
- * from the standard information attribute.
- */
-typedef SECURITY_DESCRIPTOR_RELATIVE SECURITY_DESCRIPTOR_ATTR;
-
-/*
- * On NTFS 3.0+, all security descriptors are stored in FILE_Secure. Only one
- * referenced instance of each unique security descriptor is stored.
- *
- * FILE_Secure contains no unnamed data attribute, i.e. it has zero length. It
- * does, however, contain two indexes ($SDH and $SII) as well as a named data
- * stream ($SDS).
- *
- * Every unique security descriptor is assigned a unique security identifier
- * (security_id, not to be confused with a SID). The security_id is unique for
- * the NTFS volume and is used as an index into the $SII index, which maps
- * security_ids to the security descriptor's storage location within the $SDS
- * data attribute. The $SII index is sorted by ascending security_id.
- *
- * A simple hash is computed from each security descriptor. This hash is used
- * as an index into the $SDH index, which maps security descriptor hashes to
- * the security descriptor's storage location within the $SDS data attribute.
- * The $SDH index is sorted by security descriptor hash and is stored in a B+
- * tree. When searching $SDH (with the intent of determining whether or not a
- * new security descriptor is already present in the $SDS data stream), if a
- * matching hash is found, but the security descriptors do not match, the
- * search in the $SDH index is continued, searching for a next matching hash.
- *
- * When a precise match is found, the security_id coresponding to the security
- * descriptor in the $SDS attribute is read from the found $SDH index entry and
- * is stored in the $STANDARD_INFORMATION attribute of the file/directory to
- * which the security descriptor is being applied. The $STANDARD_INFORMATION
- * attribute is present in all base mft records (i.e. in all files and
- * directories).
- *
- * If a match is not found, the security descriptor is assigned a new unique
- * security_id and is added to the $SDS data attribute. Then, entries
- * referencing the this security descriptor in the $SDS data attribute are
- * added to the $SDH and $SII indexes.
- *
- * Note: Entries are never deleted from FILE_Secure, even if nothing
- * references an entry any more.
- */
-
-/*
- * This header precedes each security descriptor in the $SDS data stream.
- * This is also the index entry data part of both the $SII and $SDH indexes.
- */
-typedef struct {
-	le32 hash;	  /* Hash of the security descriptor. */
-	le32 security_id; /* The security_id assigned to the descriptor. */
-	le64 offset;	  /* Byte offset of this entry in the $SDS stream. */
-	le32 length;	  /* Size in bytes of this entry in $SDS stream. */
-} __attribute__ ((__packed__)) SECURITY_DESCRIPTOR_HEADER;
-
-/*
- * The $SDS data stream contains the security descriptors, aligned on 16-byte
- * boundaries, sorted by security_id in a B+ tree. Security descriptors cannot
- * cross 256kib boundaries (this restriction is imposed by the Windows cache
- * manager). Each security descriptor is contained in a SDS_ENTRY structure.
- * Also, each security descriptor is stored twice in the $SDS stream with a
- * fixed offset of 0x40000 bytes (256kib, the Windows cache manager's max size)
- * between them; i.e. if a SDS_ENTRY specifies an offset of 0x51d0, then the
- * first copy of the security descriptor will be at offset 0x51d0 in the
- * $SDS data stream and the second copy will be at offset 0x451d0.
- */
-typedef struct {
-/*Ofs*/
-/*  0	SECURITY_DESCRIPTOR_HEADER; -- Unfolded here as gcc doesn't like
-				       unnamed structs. */
-	le32 hash;	  /* Hash of the security descriptor. */
-	le32 security_id; /* The security_id assigned to the descriptor. */
-	le64 offset;	  /* Byte offset of this entry in the $SDS stream. */
-	le32 length;	  /* Size in bytes of this entry in $SDS stream. */
-/* 20*/	SECURITY_DESCRIPTOR_RELATIVE sid; /* The self-relative security
-					     descriptor. */
-} __attribute__ ((__packed__)) SDS_ENTRY;
-
-/*
- * The index entry key used in the $SII index. The collation type is
- * COLLATION_NTOFS_ULONG.
- */
-typedef struct {
-	le32 security_id; /* The security_id assigned to the descriptor. */
-} __attribute__ ((__packed__)) SII_INDEX_KEY;
-
-/*
- * The index entry key used in the $SDH index. The keys are sorted first by
- * hash and then by security_id. The collation rule is
- * COLLATION_NTOFS_SECURITY_HASH.
- */
-typedef struct {
-	le32 hash;	  /* Hash of the security descriptor. */
-	le32 security_id; /* The security_id assigned to the descriptor. */
-} __attribute__ ((__packed__)) SDH_INDEX_KEY;
-
-/*
- * Attribute: Volume name (0x60).
- *
- * NOTE: Always resident.
- * NOTE: Present only in FILE_Volume.
- */
-typedef struct {
-	ntfschar name[0];	/* The name of the volume in Unicode. */
-} __attribute__ ((__packed__)) VOLUME_NAME;
-
-/*
- * Possible flags for the volume (16-bit).
- */
-enum {
-	VOLUME_IS_DIRTY			= cpu_to_le16(0x0001),
-	VOLUME_RESIZE_LOG_FILE		= cpu_to_le16(0x0002),
-	VOLUME_UPGRADE_ON_MOUNT		= cpu_to_le16(0x0004),
-	VOLUME_MOUNTED_ON_NT4		= cpu_to_le16(0x0008),
-
-	VOLUME_DELETE_USN_UNDERWAY	= cpu_to_le16(0x0010),
-	VOLUME_REPAIR_OBJECT_ID		= cpu_to_le16(0x0020),
-
-	VOLUME_CHKDSK_UNDERWAY		= cpu_to_le16(0x4000),
-	VOLUME_MODIFIED_BY_CHKDSK	= cpu_to_le16(0x8000),
-
-	VOLUME_FLAGS_MASK		= cpu_to_le16(0xc03f),
-
-	/* To make our life easier when checking if we must mount read-only. */
-	VOLUME_MUST_MOUNT_RO_MASK	= cpu_to_le16(0xc027),
-} __attribute__ ((__packed__));
-
-typedef le16 VOLUME_FLAGS;
-
-/*
- * Attribute: Volume information (0x70).
- *
- * NOTE: Always resident.
- * NOTE: Present only in FILE_Volume.
- * NOTE: Windows 2000 uses NTFS 3.0 while Windows NT4 service pack 6a uses
- *	 NTFS 1.2. I haven't personally seen other values yet.
- */
-typedef struct {
-	le64 reserved;		/* Not used (yet?). */
-	u8 major_ver;		/* Major version of the ntfs format. */
-	u8 minor_ver;		/* Minor version of the ntfs format. */
-	VOLUME_FLAGS flags;	/* Bit array of VOLUME_* flags. */
-} __attribute__ ((__packed__)) VOLUME_INFORMATION;
-
-/*
- * Attribute: Data attribute (0x80).
- *
- * NOTE: Can be resident or non-resident.
- *
- * Data contents of a file (i.e. the unnamed stream) or of a named stream.
- */
-typedef struct {
-	u8 data[0];		/* The file's data contents. */
-} __attribute__ ((__packed__)) DATA_ATTR;
-
-/*
- * Index header flags (8-bit).
- */
-enum {
-	/*
-	 * When index header is in an index root attribute:
-	 */
-	SMALL_INDEX = 0, /* The index is small enough to fit inside the index
-			    root attribute and there is no index allocation
-			    attribute present. */
-	LARGE_INDEX = 1, /* The index is too large to fit in the index root
-			    attribute and/or an index allocation attribute is
-			    present. */
-	/*
-	 * When index header is in an index block, i.e. is part of index
-	 * allocation attribute:
-	 */
-	LEAF_NODE  = 0, /* This is a leaf node, i.e. there are no more nodes
-			   branching off it. */
-	INDEX_NODE = 1, /* This node indexes other nodes, i.e. it is not a leaf
-			   node. */
-	NODE_MASK  = 1, /* Mask for accessing the *_NODE bits. */
-} __attribute__ ((__packed__));
-
-typedef u8 INDEX_HEADER_FLAGS;
-
-/*
- * This is the header for indexes, describing the INDEX_ENTRY records, which
- * follow the INDEX_HEADER. Together the index header and the index entries
- * make up a complete index.
- *
- * IMPORTANT NOTE: The offset, length and size structure members are counted
- * relative to the start of the index header structure and not relative to the
- * start of the index root or index allocation structures themselves.
- */
-typedef struct {
-	le32 entries_offset;		/* Byte offset to first INDEX_ENTRY
-					   aligned to 8-byte boundary. */
-	le32 index_length;		/* Data size of the index in bytes,
-					   i.e. bytes used from allocated
-					   size, aligned to 8-byte boundary. */
-	le32 allocated_size;		/* Byte size of this index (block),
-					   multiple of 8 bytes. */
-	/* NOTE: For the index root attribute, the above two numbers are always
-	   equal, as the attribute is resident and it is resized as needed. In
-	   the case of the index allocation attribute the attribute is not
-	   resident and hence the allocated_size is a fixed value and must
-	   equal the index_block_size specified by the INDEX_ROOT attribute
-	   corresponding to the INDEX_ALLOCATION attribute this INDEX_BLOCK
-	   belongs to. */
-	INDEX_HEADER_FLAGS flags;	/* Bit field of INDEX_HEADER_FLAGS. */
-	u8 reserved[3];			/* Reserved/align to 8-byte boundary. */
-} __attribute__ ((__packed__)) INDEX_HEADER;
-
-/*
- * Attribute: Index root (0x90).
- *
- * NOTE: Always resident.
- *
- * This is followed by a sequence of index entries (INDEX_ENTRY structures)
- * as described by the index header.
- *
- * When a directory is small enough to fit inside the index root then this
- * is the only attribute describing the directory. When the directory is too
- * large to fit in the index root, on the other hand, two additional attributes
- * are present: an index allocation attribute, containing sub-nodes of the B+
- * directory tree (see below), and a bitmap attribute, describing which virtual
- * cluster numbers (vcns) in the index allocation attribute are in use by an
- * index block.
- *
- * NOTE: The root directory (FILE_root) contains an entry for itself. Other
- * directories do not contain entries for themselves, though.
- */
-typedef struct {
-	ATTR_TYPE type;			/* Type of the indexed attribute. Is
-					   $FILE_NAME for directories, zero
-					   for view indexes. No other values
-					   allowed. */
-	COLLATION_RULE collation_rule;	/* Collation rule used to sort the
-					   index entries. If type is $FILE_NAME,
-					   this must be COLLATION_FILE_NAME. */
-	le32 index_block_size;		/* Size of each index block in bytes (in
-					   the index allocation attribute). */
-	u8 clusters_per_index_block;	/* Cluster size of each index block (in
-					   the index allocation attribute), when
-					   an index block is >= than a cluster,
-					   otherwise this will be the log of
-					   the size (like how the encoding of
-					   the mft record size and the index
-					   record size found in the boot sector
-					   work). Has to be a power of 2. */
-	u8 reserved[3];			/* Reserved/align to 8-byte boundary. */
-	INDEX_HEADER index;		/* Index header describing the
-					   following index entries. */
-} __attribute__ ((__packed__)) INDEX_ROOT;
-
-/*
- * Attribute: Index allocation (0xa0).
- *
- * NOTE: Always non-resident (doesn't make sense to be resident anyway!).
- *
- * This is an array of index blocks. Each index block starts with an
- * INDEX_BLOCK structure containing an index header, followed by a sequence of
- * index entries (INDEX_ENTRY structures), as described by the INDEX_HEADER.
- */
-typedef struct {
-/*  0	NTFS_RECORD; -- Unfolded here as gcc doesn't like unnamed structs. */
-	NTFS_RECORD_TYPE magic;	/* Magic is "INDX". */
-	le16 usa_ofs;		/* See NTFS_RECORD definition. */
-	le16 usa_count;		/* See NTFS_RECORD definition. */
-
-/*  8*/	sle64 lsn;		/* $LogFile sequence number of the last
-				   modification of this index block. */
-/* 16*/	leVCN index_block_vcn;	/* Virtual cluster number of the index block.
-				   If the cluster_size on the volume is <= the
-				   index_block_size of the directory,
-				   index_block_vcn counts in units of clusters,
-				   and in units of sectors otherwise. */
-/* 24*/	INDEX_HEADER index;	/* Describes the following index entries. */
-/* sizeof()= 40 (0x28) bytes */
-/*
- * When creating the index block, we place the update sequence array at this
- * offset, i.e. before we start with the index entries. This also makes sense,
- * otherwise we could run into problems with the update sequence array
- * containing in itself the last two bytes of a sector which would mean that
- * multi sector transfer protection wouldn't work. As you can't protect data
- * by overwriting it since you then can't get it back...
- * When reading use the data from the ntfs record header.
- */
-} __attribute__ ((__packed__)) INDEX_BLOCK;
-
-typedef INDEX_BLOCK INDEX_ALLOCATION;
-
-/*
- * The system file FILE_Extend/$Reparse contains an index named $R listing
- * all reparse points on the volume. The index entry keys are as defined
- * below. Note, that there is no index data associated with the index entries.
- *
- * The index entries are sorted by the index key file_id. The collation rule is
- * COLLATION_NTOFS_ULONGS. FIXME: Verify whether the reparse_tag is not the
- * primary key / is not a key at all. (AIA)
- */
-typedef struct {
-	le32 reparse_tag;	/* Reparse point type (inc. flags). */
-	leMFT_REF file_id;	/* Mft record of the file containing the
-				   reparse point attribute. */
-} __attribute__ ((__packed__)) REPARSE_INDEX_KEY;
-
-/*
- * Quota flags (32-bit).
- *
- * The user quota flags.  Names explain meaning.
- */
-enum {
-	QUOTA_FLAG_DEFAULT_LIMITS	= cpu_to_le32(0x00000001),
-	QUOTA_FLAG_LIMIT_REACHED	= cpu_to_le32(0x00000002),
-	QUOTA_FLAG_ID_DELETED		= cpu_to_le32(0x00000004),
-
-	QUOTA_FLAG_USER_MASK		= cpu_to_le32(0x00000007),
-	/* This is a bit mask for the user quota flags. */
-
-	/*
-	 * These flags are only present in the quota defaults index entry, i.e.
-	 * in the entry where owner_id = QUOTA_DEFAULTS_ID.
-	 */
-	QUOTA_FLAG_TRACKING_ENABLED	= cpu_to_le32(0x00000010),
-	QUOTA_FLAG_ENFORCEMENT_ENABLED	= cpu_to_le32(0x00000020),
-	QUOTA_FLAG_TRACKING_REQUESTED	= cpu_to_le32(0x00000040),
-	QUOTA_FLAG_LOG_THRESHOLD	= cpu_to_le32(0x00000080),
-
-	QUOTA_FLAG_LOG_LIMIT		= cpu_to_le32(0x00000100),
-	QUOTA_FLAG_OUT_OF_DATE		= cpu_to_le32(0x00000200),
-	QUOTA_FLAG_CORRUPT		= cpu_to_le32(0x00000400),
-	QUOTA_FLAG_PENDING_DELETES	= cpu_to_le32(0x00000800),
-};
-
-typedef le32 QUOTA_FLAGS;
-
-/*
- * The system file FILE_Extend/$Quota contains two indexes $O and $Q. Quotas
- * are on a per volume and per user basis.
- *
- * The $Q index contains one entry for each existing user_id on the volume. The
- * index key is the user_id of the user/group owning this quota control entry,
- * i.e. the key is the owner_id. The user_id of the owner of a file, i.e. the
- * owner_id, is found in the standard information attribute. The collation rule
- * for $Q is COLLATION_NTOFS_ULONG.
- *
- * The $O index contains one entry for each user/group who has been assigned
- * a quota on that volume. The index key holds the SID of the user_id the
- * entry belongs to, i.e. the owner_id. The collation rule for $O is
- * COLLATION_NTOFS_SID.
- *
- * The $O index entry data is the user_id of the user corresponding to the SID.
- * This user_id is used as an index into $Q to find the quota control entry
- * associated with the SID.
- *
- * The $Q index entry data is the quota control entry and is defined below.
- */
-typedef struct {
-	le32 version;		/* Currently equals 2. */
-	QUOTA_FLAGS flags;	/* Flags describing this quota entry. */
-	le64 bytes_used;	/* How many bytes of the quota are in use. */
-	sle64 change_time;	/* Last time this quota entry was changed. */
-	sle64 threshold;	/* Soft quota (-1 if not limited). */
-	sle64 limit;		/* Hard quota (-1 if not limited). */
-	sle64 exceeded_time;	/* How long the soft quota has been exceeded. */
-	SID sid;		/* The SID of the user/object associated with
-				   this quota entry.  Equals zero for the quota
-				   defaults entry (and in fact on a WinXP
-				   volume, it is not present at all). */
-} __attribute__ ((__packed__)) QUOTA_CONTROL_ENTRY;
-
-/*
- * Predefined owner_id values (32-bit).
- */
-enum {
-	QUOTA_INVALID_ID	= cpu_to_le32(0x00000000),
-	QUOTA_DEFAULTS_ID	= cpu_to_le32(0x00000001),
-	QUOTA_FIRST_USER_ID	= cpu_to_le32(0x00000100),
-};
-
-/*
- * Current constants for quota control entries.
- */
-typedef enum {
-	/* Current version. */
-	QUOTA_VERSION	= 2,
-} QUOTA_CONTROL_ENTRY_CONSTANTS;
-
-/*
- * Index entry flags (16-bit).
- */
-enum {
-	INDEX_ENTRY_NODE = cpu_to_le16(1), /* This entry contains a
-			sub-node, i.e. a reference to an index block in form of
-			a virtual cluster number (see below). */
-	INDEX_ENTRY_END  = cpu_to_le16(2), /* This signifies the last
-			entry in an index block.  The index entry does not
-			represent a file but it can point to a sub-node. */
-
-	INDEX_ENTRY_SPACE_FILLER = cpu_to_le16(0xffff), /* gcc: Force
-			enum bit width to 16-bit. */
-} __attribute__ ((__packed__));
-
-typedef le16 INDEX_ENTRY_FLAGS;
-
-/*
- * This the index entry header (see below).
- */
-typedef struct {
-/*  0*/	union {
-		struct { /* Only valid when INDEX_ENTRY_END is not set. */
-			leMFT_REF indexed_file;	/* The mft reference of the file
-						   described by this index
-						   entry. Used for directory
-						   indexes. */
-		} __attribute__ ((__packed__)) dir;
-		struct { /* Used for views/indexes to find the entry's data. */
-			le16 data_offset;	/* Data byte offset from this
-						   INDEX_ENTRY. Follows the
-						   index key. */
-			le16 data_length;	/* Data length in bytes. */
-			le32 reservedV;		/* Reserved (zero). */
-		} __attribute__ ((__packed__)) vi;
-	} __attribute__ ((__packed__)) data;
-/*  8*/	le16 length;		 /* Byte size of this index entry, multiple of
-				    8-bytes. */
-/* 10*/	le16 key_length;	 /* Byte size of the key value, which is in the
-				    index entry. It follows field reserved. Not
-				    multiple of 8-bytes. */
-/* 12*/	INDEX_ENTRY_FLAGS flags; /* Bit field of INDEX_ENTRY_* flags. */
-/* 14*/	le16 reserved;		 /* Reserved/align to 8-byte boundary. */
-/* sizeof() = 16 bytes */
-} __attribute__ ((__packed__)) INDEX_ENTRY_HEADER;
-
-/*
- * This is an index entry. A sequence of such entries follows each INDEX_HEADER
- * structure. Together they make up a complete index. The index follows either
- * an index root attribute or an index allocation attribute.
- *
- * NOTE: Before NTFS 3.0 only filename attributes were indexed.
- */
-typedef struct {
-/*Ofs*/
-/*  0	INDEX_ENTRY_HEADER; -- Unfolded here as gcc dislikes unnamed structs. */
-	union {
-		struct { /* Only valid when INDEX_ENTRY_END is not set. */
-			leMFT_REF indexed_file;	/* The mft reference of the file
-						   described by this index
-						   entry. Used for directory
-						   indexes. */
-		} __attribute__ ((__packed__)) dir;
-		struct { /* Used for views/indexes to find the entry's data. */
-			le16 data_offset;	/* Data byte offset from this
-						   INDEX_ENTRY. Follows the
-						   index key. */
-			le16 data_length;	/* Data length in bytes. */
-			le32 reservedV;		/* Reserved (zero). */
-		} __attribute__ ((__packed__)) vi;
-	} __attribute__ ((__packed__)) data;
-	le16 length;		 /* Byte size of this index entry, multiple of
-				    8-bytes. */
-	le16 key_length;	 /* Byte size of the key value, which is in the
-				    index entry. It follows field reserved. Not
-				    multiple of 8-bytes. */
-	INDEX_ENTRY_FLAGS flags; /* Bit field of INDEX_ENTRY_* flags. */
-	le16 reserved;		 /* Reserved/align to 8-byte boundary. */
-
-/* 16*/	union {		/* The key of the indexed attribute. NOTE: Only present
-			   if INDEX_ENTRY_END bit in flags is not set. NOTE: On
-			   NTFS versions before 3.0 the only valid key is the
-			   FILE_NAME_ATTR. On NTFS 3.0+ the following
-			   additional index keys are defined: */
-		FILE_NAME_ATTR file_name;/* $I30 index in directories. */
-		SII_INDEX_KEY sii;	/* $SII index in $Secure. */
-		SDH_INDEX_KEY sdh;	/* $SDH index in $Secure. */
-		GUID object_id;		/* $O index in FILE_Extend/$ObjId: The
-					   object_id of the mft record found in
-					   the data part of the index. */
-		REPARSE_INDEX_KEY reparse;	/* $R index in
-						   FILE_Extend/$Reparse. */
-		SID sid;		/* $O index in FILE_Extend/$Quota:
-					   SID of the owner of the user_id. */
-		le32 owner_id;		/* $Q index in FILE_Extend/$Quota:
-					   user_id of the owner of the quota
-					   control entry in the data part of
-					   the index. */
-	} __attribute__ ((__packed__)) key;
-	/* The (optional) index data is inserted here when creating. */
-	// leVCN vcn;	/* If INDEX_ENTRY_NODE bit in flags is set, the last
-	//		   eight bytes of this index entry contain the virtual
-	//		   cluster number of the index block that holds the
-	//		   entries immediately preceding the current entry (the
-	//		   vcn references the corresponding cluster in the data
-	//		   of the non-resident index allocation attribute). If
-	//		   the key_length is zero, then the vcn immediately
-	//		   follows the INDEX_ENTRY_HEADER. Regardless of
-	//		   key_length, the address of the 8-byte boundary
-	//		   aligned vcn of INDEX_ENTRY{_HEADER} *ie is given by
-	//		   (char*)ie + le16_to_cpu(ie*)->length) - sizeof(VCN),
-	//		   where sizeof(VCN) can be hardcoded as 8 if wanted. */
-} __attribute__ ((__packed__)) INDEX_ENTRY;
-
-/*
- * Attribute: Bitmap (0xb0).
- *
- * Contains an array of bits (aka a bitfield).
- *
- * When used in conjunction with the index allocation attribute, each bit
- * corresponds to one index block within the index allocation attribute. Thus
- * the number of bits in the bitmap * index block size / cluster size is the
- * number of clusters in the index allocation attribute.
- */
-typedef struct {
-	u8 bitmap[0];			/* Array of bits. */
-} __attribute__ ((__packed__)) BITMAP_ATTR;
-
-/*
- * The reparse point tag defines the type of the reparse point. It also
- * includes several flags, which further describe the reparse point.
- *
- * The reparse point tag is an unsigned 32-bit value divided in three parts:
- *
- * 1. The least significant 16 bits (i.e. bits 0 to 15) specifiy the type of
- *    the reparse point.
- * 2. The 13 bits after this (i.e. bits 16 to 28) are reserved for future use.
- * 3. The most significant three bits are flags describing the reparse point.
- *    They are defined as follows:
- *	bit 29: Name surrogate bit. If set, the filename is an alias for
- *		another object in the system.
- *	bit 30: High-latency bit. If set, accessing the first byte of data will
- *		be slow. (E.g. the data is stored on a tape drive.)
- *	bit 31: Microsoft bit. If set, the tag is owned by Microsoft. User
- *		defined tags have to use zero here.
- *
- * These are the predefined reparse point tags:
- */
-enum {
-	IO_REPARSE_TAG_IS_ALIAS		= cpu_to_le32(0x20000000),
-	IO_REPARSE_TAG_IS_HIGH_LATENCY	= cpu_to_le32(0x40000000),
-	IO_REPARSE_TAG_IS_MICROSOFT	= cpu_to_le32(0x80000000),
-
-	IO_REPARSE_TAG_RESERVED_ZERO	= cpu_to_le32(0x00000000),
-	IO_REPARSE_TAG_RESERVED_ONE	= cpu_to_le32(0x00000001),
-	IO_REPARSE_TAG_RESERVED_RANGE	= cpu_to_le32(0x00000001),
-
-	IO_REPARSE_TAG_NSS		= cpu_to_le32(0x68000005),
-	IO_REPARSE_TAG_NSS_RECOVER	= cpu_to_le32(0x68000006),
-	IO_REPARSE_TAG_SIS		= cpu_to_le32(0x68000007),
-	IO_REPARSE_TAG_DFS		= cpu_to_le32(0x68000008),
-
-	IO_REPARSE_TAG_MOUNT_POINT	= cpu_to_le32(0x88000003),
-
-	IO_REPARSE_TAG_HSM		= cpu_to_le32(0xa8000004),
-
-	IO_REPARSE_TAG_SYMBOLIC_LINK	= cpu_to_le32(0xe8000000),
-
-	IO_REPARSE_TAG_VALID_VALUES	= cpu_to_le32(0xe000ffff),
-};
-
-/*
- * Attribute: Reparse point (0xc0).
- *
- * NOTE: Can be resident or non-resident.
- */
-typedef struct {
-	le32 reparse_tag;		/* Reparse point type (inc. flags). */
-	le16 reparse_data_length;	/* Byte size of reparse data. */
-	le16 reserved;			/* Align to 8-byte boundary. */
-	u8 reparse_data[0];		/* Meaning depends on reparse_tag. */
-} __attribute__ ((__packed__)) REPARSE_POINT;
-
-/*
- * Attribute: Extended attribute (EA) information (0xd0).
- *
- * NOTE: Always resident. (Is this true???)
- */
-typedef struct {
-	le16 ea_length;		/* Byte size of the packed extended
-				   attributes. */
-	le16 need_ea_count;	/* The number of extended attributes which have
-				   the NEED_EA bit set. */
-	le32 ea_query_length;	/* Byte size of the buffer required to query
-				   the extended attributes when calling
-				   ZwQueryEaFile() in Windows NT/2k. I.e. the
-				   byte size of the unpacked extended
-				   attributes. */
-} __attribute__ ((__packed__)) EA_INFORMATION;
-
-/*
- * Extended attribute flags (8-bit).
- */
-enum {
-	NEED_EA	= 0x80		/* If set the file to which the EA belongs
-				   cannot be interpreted without understanding
-				   the associates extended attributes. */
-} __attribute__ ((__packed__));
-
-typedef u8 EA_FLAGS;
-
-/*
- * Attribute: Extended attribute (EA) (0xe0).
- *
- * NOTE: Can be resident or non-resident.
- *
- * Like the attribute list and the index buffer list, the EA attribute value is
- * a sequence of EA_ATTR variable length records.
- */
-typedef struct {
-	le32 next_entry_offset;	/* Offset to the next EA_ATTR. */
-	EA_FLAGS flags;		/* Flags describing the EA. */
-	u8 ea_name_length;	/* Length of the name of the EA in bytes
-				   excluding the '\0' byte terminator. */
-	le16 ea_value_length;	/* Byte size of the EA's value. */
-	u8 ea_name[0];		/* Name of the EA.  Note this is ASCII, not
-				   Unicode and it is zero terminated. */
-	u8 ea_value[0];		/* The value of the EA.  Immediately follows
-				   the name. */
-} __attribute__ ((__packed__)) EA_ATTR;
-
-/*
- * Attribute: Property set (0xf0).
- *
- * Intended to support Native Structure Storage (NSS) - a feature removed from
- * NTFS 3.0 during beta testing.
- */
-typedef struct {
-	/* Irrelevant as feature unused. */
-} __attribute__ ((__packed__)) PROPERTY_SET;
-
-/*
- * Attribute: Logged utility stream (0x100).
- *
- * NOTE: Can be resident or non-resident.
- *
- * Operations on this attribute are logged to the journal ($LogFile) like
- * normal metadata changes.
- *
- * Used by the Encrypting File System (EFS). All encrypted files have this
- * attribute with the name $EFS.
- */
-typedef struct {
-	/* Can be anything the creator chooses. */
-	/* EFS uses it as follows: */
-	// FIXME: Type this info, verifying it along the way. (AIA)
-} __attribute__ ((__packed__)) LOGGED_UTILITY_STREAM, EFS_ATTR;
-
-#endif /* _LINUX_NTFS_LAYOUT_H */
diff --git a/fs/ntfs/lcnalloc.c b/fs/ntfs/lcnalloc.c
deleted file mode 100644
index eda9972e61..0000000000
--- a/fs/ntfs/lcnalloc.c
+++ /dev/null
@@ -1,1000 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * lcnalloc.c - Cluster (de)allocation code.  Part of the Linux-NTFS project.
- *
- * Copyright (c) 2004-2005 Anton Altaparmakov
- */
-
-#ifdef NTFS_RW
-
-#include <linux/pagemap.h>
-
-#include "lcnalloc.h"
-#include "debug.h"
-#include "bitmap.h"
-#include "inode.h"
-#include "volume.h"
-#include "attrib.h"
-#include "malloc.h"
-#include "aops.h"
-#include "ntfs.h"
-
-/**
- * ntfs_cluster_free_from_rl_nolock - free clusters from runlist
- * @vol:	mounted ntfs volume on which to free the clusters
- * @rl:		runlist describing the clusters to free
- *
- * Free all the clusters described by the runlist @rl on the volume @vol.  In
- * the case of an error being returned, at least some of the clusters were not
- * freed.
- *
- * Return 0 on success and -errno on error.
- *
- * Locking: - The volume lcn bitmap must be locked for writing on entry and is
- *	      left locked on return.
- */
-int ntfs_cluster_free_from_rl_nolock(ntfs_volume *vol,
-		const runlist_element *rl)
-{
-	struct inode *lcnbmp_vi = vol->lcnbmp_ino;
-	int ret = 0;
-
-	ntfs_debug("Entering.");
-	if (!rl)
-		return 0;
-	for (; rl->length; rl++) {
-		int err;
-
-		if (rl->lcn < 0)
-			continue;
-		err = ntfs_bitmap_clear_run(lcnbmp_vi, rl->lcn, rl->length);
-		if (unlikely(err && (!ret || ret == -ENOMEM) && ret != err))
-			ret = err;
-	}
-	ntfs_debug("Done.");
-	return ret;
-}
-
-/**
- * ntfs_cluster_alloc - allocate clusters on an ntfs volume
- * @vol:	mounted ntfs volume on which to allocate the clusters
- * @start_vcn:	vcn to use for the first allocated cluster
- * @count:	number of clusters to allocate
- * @start_lcn:	starting lcn at which to allocate the clusters (or -1 if none)
- * @zone:	zone from which to allocate the clusters
- * @is_extension:	if 'true', this is an attribute extension
- *
- * Allocate @count clusters preferably starting at cluster @start_lcn or at the
- * current allocator position if @start_lcn is -1, on the mounted ntfs volume
- * @vol. @zone is either DATA_ZONE for allocation of normal clusters or
- * MFT_ZONE for allocation of clusters for the master file table, i.e. the
- * $MFT/$DATA attribute.
- *
- * @start_vcn specifies the vcn of the first allocated cluster.  This makes
- * merging the resulting runlist with the old runlist easier.
- *
- * If @is_extension is 'true', the caller is allocating clusters to extend an
- * attribute and if it is 'false', the caller is allocating clusters to fill a
- * hole in an attribute.  Practically the difference is that if @is_extension
- * is 'true' the returned runlist will be terminated with LCN_ENOENT and if
- * @is_extension is 'false' the runlist will be terminated with
- * LCN_RL_NOT_MAPPED.
- *
- * You need to check the return value with IS_ERR().  If this is false, the
- * function was successful and the return value is a runlist describing the
- * allocated cluster(s).  If IS_ERR() is true, the function failed and
- * PTR_ERR() gives you the error code.
- *
- * Notes on the allocation algorithm
- * =================================
- *
- * There are two data zones.  First is the area between the end of the mft zone
- * and the end of the volume, and second is the area between the start of the
- * volume and the start of the mft zone.  On unmodified/standard NTFS 1.x
- * volumes, the second data zone does not exist due to the mft zone being
- * expanded to cover the start of the volume in order to reserve space for the
- * mft bitmap attribute.
- *
- * This is not the prettiest function but the complexity stems from the need of
- * implementing the mft vs data zoned approach and from the fact that we have
- * access to the lcn bitmap in portions of up to 8192 bytes at a time, so we
- * need to cope with crossing over boundaries of two buffers.  Further, the
- * fact that the allocator allows for caller supplied hints as to the location
- * of where allocation should begin and the fact that the allocator keeps track
- * of where in the data zones the next natural allocation should occur,
- * contribute to the complexity of the function.  But it should all be
- * worthwhile, because this allocator should: 1) be a full implementation of
- * the MFT zone approach used by Windows NT, 2) cause reduction in
- * fragmentation, and 3) be speedy in allocations (the code is not optimized
- * for speed, but the algorithm is, so further speed improvements are probably
- * possible).
- *
- * FIXME: We should be monitoring cluster allocation and increment the MFT zone
- * size dynamically but this is something for the future.  We will just cause
- * heavier fragmentation by not doing it and I am not even sure Windows would
- * grow the MFT zone dynamically, so it might even be correct not to do this.
- * The overhead in doing dynamic MFT zone expansion would be very large and
- * unlikely worth the effort. (AIA)
- *
- * TODO: I have added in double the required zone position pointer wrap around
- * logic which can be optimized to having only one of the two logic sets.
- * However, having the double logic will work fine, but if we have only one of
- * the sets and we get it wrong somewhere, then we get into trouble, so
- * removing the duplicate logic requires _very_ careful consideration of _all_
- * possible code paths.  So at least for now, I am leaving the double logic -
- * better safe than sorry... (AIA)
- *
- * Locking: - The volume lcn bitmap must be unlocked on entry and is unlocked
- *	      on return.
- *	    - This function takes the volume lcn bitmap lock for writing and
- *	      modifies the bitmap contents.
- */
-runlist_element *ntfs_cluster_alloc(ntfs_volume *vol, const VCN start_vcn,
-		const s64 count, const LCN start_lcn,
-		const NTFS_CLUSTER_ALLOCATION_ZONES zone,
-		const bool is_extension)
-{
-	LCN zone_start, zone_end, bmp_pos, bmp_initial_pos, last_read_pos, lcn;
-	LCN prev_lcn = 0, prev_run_len = 0, mft_zone_size;
-	s64 clusters;
-	loff_t i_size;
-	struct inode *lcnbmp_vi;
-	runlist_element *rl = NULL;
-	struct address_space *mapping;
-	struct page *page = NULL;
-	u8 *buf, *byte;
-	int err = 0, rlpos, rlsize, buf_size;
-	u8 pass, done_zones, search_zone, need_writeback = 0, bit;
-
-	ntfs_debug("Entering for start_vcn 0x%llx, count 0x%llx, start_lcn "
-			"0x%llx, zone %s_ZONE.", (unsigned long long)start_vcn,
-			(unsigned long long)count,
-			(unsigned long long)start_lcn,
-			zone == MFT_ZONE ? "MFT" : "DATA");
-	BUG_ON(!vol);
-	lcnbmp_vi = vol->lcnbmp_ino;
-	BUG_ON(!lcnbmp_vi);
-	BUG_ON(start_vcn < 0);
-	BUG_ON(count < 0);
-	BUG_ON(start_lcn < -1);
-	BUG_ON(zone < FIRST_ZONE);
-	BUG_ON(zone > LAST_ZONE);
-
-	/* Return NULL if @count is zero. */
-	if (!count)
-		return NULL;
-	/* Take the lcnbmp lock for writing. */
-	down_write(&vol->lcnbmp_lock);
-	/*
-	 * If no specific @start_lcn was requested, use the current data zone
-	 * position, otherwise use the requested @start_lcn but make sure it
-	 * lies outside the mft zone.  Also set done_zones to 0 (no zones done)
-	 * and pass depending on whether we are starting inside a zone (1) or
-	 * at the beginning of a zone (2).  If requesting from the MFT_ZONE,
-	 * we either start at the current position within the mft zone or at
-	 * the specified position.  If the latter is out of bounds then we start
-	 * at the beginning of the MFT_ZONE.
-	 */
-	done_zones = 0;
-	pass = 1;
-	/*
-	 * zone_start and zone_end are the current search range.  search_zone
-	 * is 1 for mft zone, 2 for data zone 1 (end of mft zone till end of
-	 * volume) and 4 for data zone 2 (start of volume till start of mft
-	 * zone).
-	 */
-	zone_start = start_lcn;
-	if (zone_start < 0) {
-		if (zone == DATA_ZONE)
-			zone_start = vol->data1_zone_pos;
-		else
-			zone_start = vol->mft_zone_pos;
-		if (!zone_start) {
-			/*
-			 * Zone starts at beginning of volume which means a
-			 * single pass is sufficient.
-			 */
-			pass = 2;
-		}
-	} else if (zone == DATA_ZONE && zone_start >= vol->mft_zone_start &&
-			zone_start < vol->mft_zone_end) {
-		zone_start = vol->mft_zone_end;
-		/*
-		 * Starting at beginning of data1_zone which means a single
-		 * pass in this zone is sufficient.
-		 */
-		pass = 2;
-	} else if (zone == MFT_ZONE && (zone_start < vol->mft_zone_start ||
-			zone_start >= vol->mft_zone_end)) {
-		zone_start = vol->mft_lcn;
-		if (!vol->mft_zone_end)
-			zone_start = 0;
-		/*
-		 * Starting at beginning of volume which means a single pass
-		 * is sufficient.
-		 */
-		pass = 2;
-	}
-	if (zone == MFT_ZONE) {
-		zone_end = vol->mft_zone_end;
-		search_zone = 1;
-	} else /* if (zone == DATA_ZONE) */ {
-		/* Skip searching the mft zone. */
-		done_zones |= 1;
-		if (zone_start >= vol->mft_zone_end) {
-			zone_end = vol->nr_clusters;
-			search_zone = 2;
-		} else {
-			zone_end = vol->mft_zone_start;
-			search_zone = 4;
-		}
-	}
-	/*
-	 * bmp_pos is the current bit position inside the bitmap.  We use
-	 * bmp_initial_pos to determine whether or not to do a zone switch.
-	 */
-	bmp_pos = bmp_initial_pos = zone_start;
-
-	/* Loop until all clusters are allocated, i.e. clusters == 0. */
-	clusters = count;
-	rlpos = rlsize = 0;
-	mapping = lcnbmp_vi->i_mapping;
-	i_size = i_size_read(lcnbmp_vi);
-	while (1) {
-		ntfs_debug("Start of outer while loop: done_zones 0x%x, "
-				"search_zone %i, pass %i, zone_start 0x%llx, "
-				"zone_end 0x%llx, bmp_initial_pos 0x%llx, "
-				"bmp_pos 0x%llx, rlpos %i, rlsize %i.",
-				done_zones, search_zone, pass,
-				(unsigned long long)zone_start,
-				(unsigned long long)zone_end,
-				(unsigned long long)bmp_initial_pos,
-				(unsigned long long)bmp_pos, rlpos, rlsize);
-		/* Loop until we run out of free clusters. */
-		last_read_pos = bmp_pos >> 3;
-		ntfs_debug("last_read_pos 0x%llx.",
-				(unsigned long long)last_read_pos);
-		if (last_read_pos > i_size) {
-			ntfs_debug("End of attribute reached.  "
-					"Skipping to zone_pass_done.");
-			goto zone_pass_done;
-		}
-		if (likely(page)) {
-			if (need_writeback) {
-				ntfs_debug("Marking page dirty.");
-				flush_dcache_page(page);
-				set_page_dirty(page);
-				need_writeback = 0;
-			}
-			ntfs_unmap_page(page);
-		}
-		page = ntfs_map_page(mapping, last_read_pos >>
-				PAGE_SHIFT);
-		if (IS_ERR(page)) {
-			err = PTR_ERR(page);
-			ntfs_error(vol->sb, "Failed to map page.");
-			goto out;
-		}
-		buf_size = last_read_pos & ~PAGE_MASK;
-		buf = page_address(page) + buf_size;
-		buf_size = PAGE_SIZE - buf_size;
-		if (unlikely(last_read_pos + buf_size > i_size))
-			buf_size = i_size - last_read_pos;
-		buf_size <<= 3;
-		lcn = bmp_pos & 7;
-		bmp_pos &= ~(LCN)7;
-		ntfs_debug("Before inner while loop: buf_size %i, lcn 0x%llx, "
-				"bmp_pos 0x%llx, need_writeback %i.", buf_size,
-				(unsigned long long)lcn,
-				(unsigned long long)bmp_pos, need_writeback);
-		while (lcn < buf_size && lcn + bmp_pos < zone_end) {
-			byte = buf + (lcn >> 3);
-			ntfs_debug("In inner while loop: buf_size %i, "
-					"lcn 0x%llx, bmp_pos 0x%llx, "
-					"need_writeback %i, byte ofs 0x%x, "
-					"*byte 0x%x.", buf_size,
-					(unsigned long long)lcn,
-					(unsigned long long)bmp_pos,
-					need_writeback,
-					(unsigned int)(lcn >> 3),
-					(unsigned int)*byte);
-			/* Skip full bytes. */
-			if (*byte == 0xff) {
-				lcn = (lcn + 8) & ~(LCN)7;
-				ntfs_debug("Continuing while loop 1.");
-				continue;
-			}
-			bit = 1 << (lcn & 7);
-			ntfs_debug("bit 0x%x.", bit);
-			/* If the bit is already set, go onto the next one. */
-			if (*byte & bit) {
-				lcn++;
-				ntfs_debug("Continuing while loop 2.");
-				continue;
-			}
-			/*
-			 * Allocate more memory if needed, including space for
-			 * the terminator element.
-			 * ntfs_malloc_nofs() operates on whole pages only.
-			 */
-			if ((rlpos + 2) * sizeof(*rl) > rlsize) {
-				runlist_element *rl2;
-
-				ntfs_debug("Reallocating memory.");
-				if (!rl)
-					ntfs_debug("First free bit is at LCN "
-							"0x%llx.",
-							(unsigned long long)
-							(lcn + bmp_pos));
-				rl2 = ntfs_malloc_nofs(rlsize + (int)PAGE_SIZE);
-				if (unlikely(!rl2)) {
-					err = -ENOMEM;
-					ntfs_error(vol->sb, "Failed to "
-							"allocate memory.");
-					goto out;
-				}
-				memcpy(rl2, rl, rlsize);
-				ntfs_free(rl);
-				rl = rl2;
-				rlsize += PAGE_SIZE;
-				ntfs_debug("Reallocated memory, rlsize 0x%x.",
-						rlsize);
-			}
-			/* Allocate the bitmap bit. */
-			*byte |= bit;
-			/* We need to write this bitmap page to disk. */
-			need_writeback = 1;
-			ntfs_debug("*byte 0x%x, need_writeback is set.",
-					(unsigned int)*byte);
-			/*
-			 * Coalesce with previous run if adjacent LCNs.
-			 * Otherwise, append a new run.
-			 */
-			ntfs_debug("Adding run (lcn 0x%llx, len 0x%llx), "
-					"prev_lcn 0x%llx, lcn 0x%llx, "
-					"bmp_pos 0x%llx, prev_run_len 0x%llx, "
-					"rlpos %i.",
-					(unsigned long long)(lcn + bmp_pos),
-					1ULL, (unsigned long long)prev_lcn,
-					(unsigned long long)lcn,
-					(unsigned long long)bmp_pos,
-					(unsigned long long)prev_run_len,
-					rlpos);
-			if (prev_lcn == lcn + bmp_pos - prev_run_len && rlpos) {
-				ntfs_debug("Coalescing to run (lcn 0x%llx, "
-						"len 0x%llx).",
-						(unsigned long long)
-						rl[rlpos - 1].lcn,
-						(unsigned long long)
-						rl[rlpos - 1].length);
-				rl[rlpos - 1].length = ++prev_run_len;
-				ntfs_debug("Run now (lcn 0x%llx, len 0x%llx), "
-						"prev_run_len 0x%llx.",
-						(unsigned long long)
-						rl[rlpos - 1].lcn,
-						(unsigned long long)
-						rl[rlpos - 1].length,
-						(unsigned long long)
-						prev_run_len);
-			} else {
-				if (likely(rlpos)) {
-					ntfs_debug("Adding new run, (previous "
-							"run lcn 0x%llx, "
-							"len 0x%llx).",
-							(unsigned long long)
-							rl[rlpos - 1].lcn,
-							(unsigned long long)
-							rl[rlpos - 1].length);
-					rl[rlpos].vcn = rl[rlpos - 1].vcn +
-							prev_run_len;
-				} else {
-					ntfs_debug("Adding new run, is first "
-							"run.");
-					rl[rlpos].vcn = start_vcn;
-				}
-				rl[rlpos].lcn = prev_lcn = lcn + bmp_pos;
-				rl[rlpos].length = prev_run_len = 1;
-				rlpos++;
-			}
-			/* Done? */
-			if (!--clusters) {
-				LCN tc;
-				/*
-				 * Update the current zone position.  Positions
-				 * of already scanned zones have been updated
-				 * during the respective zone switches.
-				 */
-				tc = lcn + bmp_pos + 1;
-				ntfs_debug("Done. Updating current zone "
-						"position, tc 0x%llx, "
-						"search_zone %i.",
-						(unsigned long long)tc,
-						search_zone);
-				switch (search_zone) {
-				case 1:
-					ntfs_debug("Before checks, "
-							"vol->mft_zone_pos "
-							"0x%llx.",
-							(unsigned long long)
-							vol->mft_zone_pos);
-					if (tc >= vol->mft_zone_end) {
-						vol->mft_zone_pos =
-								vol->mft_lcn;
-						if (!vol->mft_zone_end)
-							vol->mft_zone_pos = 0;
-					} else if ((bmp_initial_pos >=
-							vol->mft_zone_pos ||
-							tc > vol->mft_zone_pos)
-							&& tc >= vol->mft_lcn)
-						vol->mft_zone_pos = tc;
-					ntfs_debug("After checks, "
-							"vol->mft_zone_pos "
-							"0x%llx.",
-							(unsigned long long)
-							vol->mft_zone_pos);
-					break;
-				case 2:
-					ntfs_debug("Before checks, "
-							"vol->data1_zone_pos "
-							"0x%llx.",
-							(unsigned long long)
-							vol->data1_zone_pos);
-					if (tc >= vol->nr_clusters)
-						vol->data1_zone_pos =
-							     vol->mft_zone_end;
-					else if ((bmp_initial_pos >=
-						    vol->data1_zone_pos ||
-						    tc > vol->data1_zone_pos)
-						    && tc >= vol->mft_zone_end)
-						vol->data1_zone_pos = tc;
-					ntfs_debug("After checks, "
-							"vol->data1_zone_pos "
-							"0x%llx.",
-							(unsigned long long)
-							vol->data1_zone_pos);
-					break;
-				case 4:
-					ntfs_debug("Before checks, "
-							"vol->data2_zone_pos "
-							"0x%llx.",
-							(unsigned long long)
-							vol->data2_zone_pos);
-					if (tc >= vol->mft_zone_start)
-						vol->data2_zone_pos = 0;
-					else if (bmp_initial_pos >=
-						      vol->data2_zone_pos ||
-						      tc > vol->data2_zone_pos)
-						vol->data2_zone_pos = tc;
-					ntfs_debug("After checks, "
-							"vol->data2_zone_pos "
-							"0x%llx.",
-							(unsigned long long)
-							vol->data2_zone_pos);
-					break;
-				default:
-					BUG();
-				}
-				ntfs_debug("Finished.  Going to out.");
-				goto out;
-			}
-			lcn++;
-		}
-		bmp_pos += buf_size;
-		ntfs_debug("After inner while loop: buf_size 0x%x, lcn "
-				"0x%llx, bmp_pos 0x%llx, need_writeback %i.",
-				buf_size, (unsigned long long)lcn,
-				(unsigned long long)bmp_pos, need_writeback);
-		if (bmp_pos < zone_end) {
-			ntfs_debug("Continuing outer while loop, "
-					"bmp_pos 0x%llx, zone_end 0x%llx.",
-					(unsigned long long)bmp_pos,
-					(unsigned long long)zone_end);
-			continue;
-		}
-zone_pass_done:	/* Finished with the current zone pass. */
-		ntfs_debug("At zone_pass_done, pass %i.", pass);
-		if (pass == 1) {
-			/*
-			 * Now do pass 2, scanning the first part of the zone
-			 * we omitted in pass 1.
-			 */
-			pass = 2;
-			zone_end = zone_start;
-			switch (search_zone) {
-			case 1: /* mft_zone */
-				zone_start = vol->mft_zone_start;
-				break;
-			case 2: /* data1_zone */
-				zone_start = vol->mft_zone_end;
-				break;
-			case 4: /* data2_zone */
-				zone_start = 0;
-				break;
-			default:
-				BUG();
-			}
-			/* Sanity check. */
-			if (zone_end < zone_start)
-				zone_end = zone_start;
-			bmp_pos = zone_start;
-			ntfs_debug("Continuing outer while loop, pass 2, "
-					"zone_start 0x%llx, zone_end 0x%llx, "
-					"bmp_pos 0x%llx.",
-					(unsigned long long)zone_start,
-					(unsigned long long)zone_end,
-					(unsigned long long)bmp_pos);
-			continue;
-		} /* pass == 2 */
-done_zones_check:
-		ntfs_debug("At done_zones_check, search_zone %i, done_zones "
-				"before 0x%x, done_zones after 0x%x.",
-				search_zone, done_zones,
-				done_zones | search_zone);
-		done_zones |= search_zone;
-		if (done_zones < 7) {
-			ntfs_debug("Switching zone.");
-			/* Now switch to the next zone we haven't done yet. */
-			pass = 1;
-			switch (search_zone) {
-			case 1:
-				ntfs_debug("Switching from mft zone to data1 "
-						"zone.");
-				/* Update mft zone position. */
-				if (rlpos) {
-					LCN tc;
-
-					ntfs_debug("Before checks, "
-							"vol->mft_zone_pos "
-							"0x%llx.",
-							(unsigned long long)
-							vol->mft_zone_pos);
-					tc = rl[rlpos - 1].lcn +
-							rl[rlpos - 1].length;
-					if (tc >= vol->mft_zone_end) {
-						vol->mft_zone_pos =
-								vol->mft_lcn;
-						if (!vol->mft_zone_end)
-							vol->mft_zone_pos = 0;
-					} else if ((bmp_initial_pos >=
-							vol->mft_zone_pos ||
-							tc > vol->mft_zone_pos)
-							&& tc >= vol->mft_lcn)
-						vol->mft_zone_pos = tc;
-					ntfs_debug("After checks, "
-							"vol->mft_zone_pos "
-							"0x%llx.",
-							(unsigned long long)
-							vol->mft_zone_pos);
-				}
-				/* Switch from mft zone to data1 zone. */
-switch_to_data1_zone:		search_zone = 2;
-				zone_start = bmp_initial_pos =
-						vol->data1_zone_pos;
-				zone_end = vol->nr_clusters;
-				if (zone_start == vol->mft_zone_end)
-					pass = 2;
-				if (zone_start >= zone_end) {
-					vol->data1_zone_pos = zone_start =
-							vol->mft_zone_end;
-					pass = 2;
-				}
-				break;
-			case 2:
-				ntfs_debug("Switching from data1 zone to "
-						"data2 zone.");
-				/* Update data1 zone position. */
-				if (rlpos) {
-					LCN tc;
-
-					ntfs_debug("Before checks, "
-							"vol->data1_zone_pos "
-							"0x%llx.",
-							(unsigned long long)
-							vol->data1_zone_pos);
-					tc = rl[rlpos - 1].lcn +
-							rl[rlpos - 1].length;
-					if (tc >= vol->nr_clusters)
-						vol->data1_zone_pos =
-							     vol->mft_zone_end;
-					else if ((bmp_initial_pos >=
-						    vol->data1_zone_pos ||
-						    tc > vol->data1_zone_pos)
-						    && tc >= vol->mft_zone_end)
-						vol->data1_zone_pos = tc;
-					ntfs_debug("After checks, "
-							"vol->data1_zone_pos "
-							"0x%llx.",
-							(unsigned long long)
-							vol->data1_zone_pos);
-				}
-				/* Switch from data1 zone to data2 zone. */
-				search_zone = 4;
-				zone_start = bmp_initial_pos =
-						vol->data2_zone_pos;
-				zone_end = vol->mft_zone_start;
-				if (!zone_start)
-					pass = 2;
-				if (zone_start >= zone_end) {
-					vol->data2_zone_pos = zone_start =
-							bmp_initial_pos = 0;
-					pass = 2;
-				}
-				break;
-			case 4:
-				ntfs_debug("Switching from data2 zone to "
-						"data1 zone.");
-				/* Update data2 zone position. */
-				if (rlpos) {
-					LCN tc;
-
-					ntfs_debug("Before checks, "
-							"vol->data2_zone_pos "
-							"0x%llx.",
-							(unsigned long long)
-							vol->data2_zone_pos);
-					tc = rl[rlpos - 1].lcn +
-							rl[rlpos - 1].length;
-					if (tc >= vol->mft_zone_start)
-						vol->data2_zone_pos = 0;
-					else if (bmp_initial_pos >=
-						      vol->data2_zone_pos ||
-						      tc > vol->data2_zone_pos)
-						vol->data2_zone_pos = tc;
-					ntfs_debug("After checks, "
-							"vol->data2_zone_pos "
-							"0x%llx.",
-							(unsigned long long)
-							vol->data2_zone_pos);
-				}
-				/* Switch from data2 zone to data1 zone. */
-				goto switch_to_data1_zone;
-			default:
-				BUG();
-			}
-			ntfs_debug("After zone switch, search_zone %i, "
-					"pass %i, bmp_initial_pos 0x%llx, "
-					"zone_start 0x%llx, zone_end 0x%llx.",
-					search_zone, pass,
-					(unsigned long long)bmp_initial_pos,
-					(unsigned long long)zone_start,
-					(unsigned long long)zone_end);
-			bmp_pos = zone_start;
-			if (zone_start == zone_end) {
-				ntfs_debug("Empty zone, going to "
-						"done_zones_check.");
-				/* Empty zone. Don't bother searching it. */
-				goto done_zones_check;
-			}
-			ntfs_debug("Continuing outer while loop.");
-			continue;
-		} /* done_zones == 7 */
-		ntfs_debug("All zones are finished.");
-		/*
-		 * All zones are finished!  If DATA_ZONE, shrink mft zone.  If
-		 * MFT_ZONE, we have really run out of space.
-		 */
-		mft_zone_size = vol->mft_zone_end - vol->mft_zone_start;
-		ntfs_debug("vol->mft_zone_start 0x%llx, vol->mft_zone_end "
-				"0x%llx, mft_zone_size 0x%llx.",
-				(unsigned long long)vol->mft_zone_start,
-				(unsigned long long)vol->mft_zone_end,
-				(unsigned long long)mft_zone_size);
-		if (zone == MFT_ZONE || mft_zone_size <= 0) {
-			ntfs_debug("No free clusters left, going to out.");
-			/* Really no more space left on device. */
-			err = -ENOSPC;
-			goto out;
-		} /* zone == DATA_ZONE && mft_zone_size > 0 */
-		ntfs_debug("Shrinking mft zone.");
-		zone_end = vol->mft_zone_end;
-		mft_zone_size >>= 1;
-		if (mft_zone_size > 0)
-			vol->mft_zone_end = vol->mft_zone_start + mft_zone_size;
-		else /* mft zone and data2 zone no longer exist. */
-			vol->data2_zone_pos = vol->mft_zone_start =
-					vol->mft_zone_end = 0;
-		if (vol->mft_zone_pos >= vol->mft_zone_end) {
-			vol->mft_zone_pos = vol->mft_lcn;
-			if (!vol->mft_zone_end)
-				vol->mft_zone_pos = 0;
-		}
-		bmp_pos = zone_start = bmp_initial_pos =
-				vol->data1_zone_pos = vol->mft_zone_end;
-		search_zone = 2;
-		pass = 2;
-		done_zones &= ~2;
-		ntfs_debug("After shrinking mft zone, mft_zone_size 0x%llx, "
-				"vol->mft_zone_start 0x%llx, "
-				"vol->mft_zone_end 0x%llx, "
-				"vol->mft_zone_pos 0x%llx, search_zone 2, "
-				"pass 2, dones_zones 0x%x, zone_start 0x%llx, "
-				"zone_end 0x%llx, vol->data1_zone_pos 0x%llx, "
-				"continuing outer while loop.",
-				(unsigned long long)mft_zone_size,
-				(unsigned long long)vol->mft_zone_start,
-				(unsigned long long)vol->mft_zone_end,
-				(unsigned long long)vol->mft_zone_pos,
-				done_zones, (unsigned long long)zone_start,
-				(unsigned long long)zone_end,
-				(unsigned long long)vol->data1_zone_pos);
-	}
-	ntfs_debug("After outer while loop.");
-out:
-	ntfs_debug("At out.");
-	/* Add runlist terminator element. */
-	if (likely(rl)) {
-		rl[rlpos].vcn = rl[rlpos - 1].vcn + rl[rlpos - 1].length;
-		rl[rlpos].lcn = is_extension ? LCN_ENOENT : LCN_RL_NOT_MAPPED;
-		rl[rlpos].length = 0;
-	}
-	if (likely(page && !IS_ERR(page))) {
-		if (need_writeback) {
-			ntfs_debug("Marking page dirty.");
-			flush_dcache_page(page);
-			set_page_dirty(page);
-			need_writeback = 0;
-		}
-		ntfs_unmap_page(page);
-	}
-	if (likely(!err)) {
-		up_write(&vol->lcnbmp_lock);
-		ntfs_debug("Done.");
-		return rl;
-	}
-	ntfs_error(vol->sb, "Failed to allocate clusters, aborting "
-			"(error %i).", err);
-	if (rl) {
-		int err2;
-
-		if (err == -ENOSPC)
-			ntfs_debug("Not enough space to complete allocation, "
-					"err -ENOSPC, first free lcn 0x%llx, "
-					"could allocate up to 0x%llx "
-					"clusters.",
-					(unsigned long long)rl[0].lcn,
-					(unsigned long long)(count - clusters));
-		/* Deallocate all allocated clusters. */
-		ntfs_debug("Attempting rollback...");
-		err2 = ntfs_cluster_free_from_rl_nolock(vol, rl);
-		if (err2) {
-			ntfs_error(vol->sb, "Failed to rollback (error %i).  "
-					"Leaving inconsistent metadata!  "
-					"Unmount and run chkdsk.", err2);
-			NVolSetErrors(vol);
-		}
-		/* Free the runlist. */
-		ntfs_free(rl);
-	} else if (err == -ENOSPC)
-		ntfs_debug("No space left at all, err = -ENOSPC, first free "
-				"lcn = 0x%llx.",
-				(long long)vol->data1_zone_pos);
-	up_write(&vol->lcnbmp_lock);
-	return ERR_PTR(err);
-}
-
-/**
- * __ntfs_cluster_free - free clusters on an ntfs volume
- * @ni:		ntfs inode whose runlist describes the clusters to free
- * @start_vcn:	vcn in the runlist of @ni at which to start freeing clusters
- * @count:	number of clusters to free or -1 for all clusters
- * @ctx:	active attribute search context if present or NULL if not
- * @is_rollback:	true if this is a rollback operation
- *
- * Free @count clusters starting at the cluster @start_vcn in the runlist
- * described by the vfs inode @ni.
- *
- * If @count is -1, all clusters from @start_vcn to the end of the runlist are
- * deallocated.  Thus, to completely free all clusters in a runlist, use
- * @start_vcn = 0 and @count = -1.
- *
- * If @ctx is specified, it is an active search context of @ni and its base mft
- * record.  This is needed when __ntfs_cluster_free() encounters unmapped
- * runlist fragments and allows their mapping.  If you do not have the mft
- * record mapped, you can specify @ctx as NULL and __ntfs_cluster_free() will
- * perform the necessary mapping and unmapping.
- *
- * Note, __ntfs_cluster_free() saves the state of @ctx on entry and restores it
- * before returning.  Thus, @ctx will be left pointing to the same attribute on
- * return as on entry.  However, the actual pointers in @ctx may point to
- * different memory locations on return, so you must remember to reset any
- * cached pointers from the @ctx, i.e. after the call to __ntfs_cluster_free(),
- * you will probably want to do:
- *	m = ctx->mrec;
- *	a = ctx->attr;
- * Assuming you cache ctx->attr in a variable @a of type ATTR_RECORD * and that
- * you cache ctx->mrec in a variable @m of type MFT_RECORD *.
- *
- * @is_rollback should always be 'false', it is for internal use to rollback
- * errors.  You probably want to use ntfs_cluster_free() instead.
- *
- * Note, __ntfs_cluster_free() does not modify the runlist, so you have to
- * remove from the runlist or mark sparse the freed runs later.
- *
- * Return the number of deallocated clusters (not counting sparse ones) on
- * success and -errno on error.
- *
- * WARNING: If @ctx is supplied, regardless of whether success or failure is
- *	    returned, you need to check IS_ERR(@ctx->mrec) and if 'true' the @ctx
- *	    is no longer valid, i.e. you need to either call
- *	    ntfs_attr_reinit_search_ctx() or ntfs_attr_put_search_ctx() on it.
- *	    In that case PTR_ERR(@ctx->mrec) will give you the error code for
- *	    why the mapping of the old inode failed.
- *
- * Locking: - The runlist described by @ni must be locked for writing on entry
- *	      and is locked on return.  Note the runlist may be modified when
- *	      needed runlist fragments need to be mapped.
- *	    - The volume lcn bitmap must be unlocked on entry and is unlocked
- *	      on return.
- *	    - This function takes the volume lcn bitmap lock for writing and
- *	      modifies the bitmap contents.
- *	    - If @ctx is NULL, the base mft record of @ni must not be mapped on
- *	      entry and it will be left unmapped on return.
- *	    - If @ctx is not NULL, the base mft record must be mapped on entry
- *	      and it will be left mapped on return.
- */
-s64 __ntfs_cluster_free(ntfs_inode *ni, const VCN start_vcn, s64 count,
-		ntfs_attr_search_ctx *ctx, const bool is_rollback)
-{
-	s64 delta, to_free, total_freed, real_freed;
-	ntfs_volume *vol;
-	struct inode *lcnbmp_vi;
-	runlist_element *rl;
-	int err;
-
-	BUG_ON(!ni);
-	ntfs_debug("Entering for i_ino 0x%lx, start_vcn 0x%llx, count "
-			"0x%llx.%s", ni->mft_no, (unsigned long long)start_vcn,
-			(unsigned long long)count,
-			is_rollback ? " (rollback)" : "");
-	vol = ni->vol;
-	lcnbmp_vi = vol->lcnbmp_ino;
-	BUG_ON(!lcnbmp_vi);
-	BUG_ON(start_vcn < 0);
-	BUG_ON(count < -1);
-	/*
-	 * Lock the lcn bitmap for writing but only if not rolling back.  We
-	 * must hold the lock all the way including through rollback otherwise
-	 * rollback is not possible because once we have cleared a bit and
-	 * dropped the lock, anyone could have set the bit again, thus
-	 * allocating the cluster for another use.
-	 */
-	if (likely(!is_rollback))
-		down_write(&vol->lcnbmp_lock);
-
-	total_freed = real_freed = 0;
-
-	rl = ntfs_attr_find_vcn_nolock(ni, start_vcn, ctx);
-	if (IS_ERR(rl)) {
-		if (!is_rollback)
-			ntfs_error(vol->sb, "Failed to find first runlist "
-					"element (error %li), aborting.",
-					PTR_ERR(rl));
-		err = PTR_ERR(rl);
-		goto err_out;
-	}
-	if (unlikely(rl->lcn < LCN_HOLE)) {
-		if (!is_rollback)
-			ntfs_error(vol->sb, "First runlist element has "
-					"invalid lcn, aborting.");
-		err = -EIO;
-		goto err_out;
-	}
-	/* Find the starting cluster inside the run that needs freeing. */
-	delta = start_vcn - rl->vcn;
-
-	/* The number of clusters in this run that need freeing. */
-	to_free = rl->length - delta;
-	if (count >= 0 && to_free > count)
-		to_free = count;
-
-	if (likely(rl->lcn >= 0)) {
-		/* Do the actual freeing of the clusters in this run. */
-		err = ntfs_bitmap_set_bits_in_run(lcnbmp_vi, rl->lcn + delta,
-				to_free, likely(!is_rollback) ? 0 : 1);
-		if (unlikely(err)) {
-			if (!is_rollback)
-				ntfs_error(vol->sb, "Failed to clear first run "
-						"(error %i), aborting.", err);
-			goto err_out;
-		}
-		/* We have freed @to_free real clusters. */
-		real_freed = to_free;
-	};
-	/* Go to the next run and adjust the number of clusters left to free. */
-	++rl;
-	if (count >= 0)
-		count -= to_free;
-
-	/* Keep track of the total "freed" clusters, including sparse ones. */
-	total_freed = to_free;
-	/*
-	 * Loop over the remaining runs, using @count as a capping value, and
-	 * free them.
-	 */
-	for (; rl->length && count != 0; ++rl) {
-		if (unlikely(rl->lcn < LCN_HOLE)) {
-			VCN vcn;
-
-			/* Attempt to map runlist. */
-			vcn = rl->vcn;
-			rl = ntfs_attr_find_vcn_nolock(ni, vcn, ctx);
-			if (IS_ERR(rl)) {
-				err = PTR_ERR(rl);
-				if (!is_rollback)
-					ntfs_error(vol->sb, "Failed to map "
-							"runlist fragment or "
-							"failed to find "
-							"subsequent runlist "
-							"element.");
-				goto err_out;
-			}
-			if (unlikely(rl->lcn < LCN_HOLE)) {
-				if (!is_rollback)
-					ntfs_error(vol->sb, "Runlist element "
-							"has invalid lcn "
-							"(0x%llx).",
-							(unsigned long long)
-							rl->lcn);
-				err = -EIO;
-				goto err_out;
-			}
-		}
-		/* The number of clusters in this run that need freeing. */
-		to_free = rl->length;
-		if (count >= 0 && to_free > count)
-			to_free = count;
-
-		if (likely(rl->lcn >= 0)) {
-			/* Do the actual freeing of the clusters in the run. */
-			err = ntfs_bitmap_set_bits_in_run(lcnbmp_vi, rl->lcn,
-					to_free, likely(!is_rollback) ? 0 : 1);
-			if (unlikely(err)) {
-				if (!is_rollback)
-					ntfs_error(vol->sb, "Failed to clear "
-							"subsequent run.");
-				goto err_out;
-			}
-			/* We have freed @to_free real clusters. */
-			real_freed += to_free;
-		}
-		/* Adjust the number of clusters left to free. */
-		if (count >= 0)
-			count -= to_free;
-	
-		/* Update the total done clusters. */
-		total_freed += to_free;
-	}
-	if (likely(!is_rollback))
-		up_write(&vol->lcnbmp_lock);
-
-	BUG_ON(count > 0);
-
-	/* We are done.  Return the number of actually freed clusters. */
-	ntfs_debug("Done.");
-	return real_freed;
-err_out:
-	if (is_rollback)
-		return err;
-	/* If no real clusters were freed, no need to rollback. */
-	if (!real_freed) {
-		up_write(&vol->lcnbmp_lock);
-		return err;
-	}
-	/*
-	 * Attempt to rollback and if that succeeds just return the error code.
-	 * If rollback fails, set the volume errors flag, emit an error
-	 * message, and return the error code.
-	 */
-	delta = __ntfs_cluster_free(ni, start_vcn, total_freed, ctx, true);
-	if (delta < 0) {
-		ntfs_error(vol->sb, "Failed to rollback (error %i).  Leaving "
-				"inconsistent metadata!  Unmount and run "
-				"chkdsk.", (int)delta);
-		NVolSetErrors(vol);
-	}
-	up_write(&vol->lcnbmp_lock);
-	ntfs_error(vol->sb, "Aborting (error %i).", err);
-	return err;
-}
-
-#endif /* NTFS_RW */
diff --git a/fs/ntfs/lcnalloc.h b/fs/ntfs/lcnalloc.h
deleted file mode 100644
index 1589a6d843..0000000000
--- a/fs/ntfs/lcnalloc.h
+++ /dev/null
@@ -1,131 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * lcnalloc.h - Exports for NTFS kernel cluster (de)allocation.  Part of the
- *		Linux-NTFS project.
- *
- * Copyright (c) 2004-2005 Anton Altaparmakov
- */
-
-#ifndef _LINUX_NTFS_LCNALLOC_H
-#define _LINUX_NTFS_LCNALLOC_H
-
-#ifdef NTFS_RW
-
-#include <linux/fs.h>
-
-#include "attrib.h"
-#include "types.h"
-#include "inode.h"
-#include "runlist.h"
-#include "volume.h"
-
-typedef enum {
-	FIRST_ZONE	= 0,	/* For sanity checking. */
-	MFT_ZONE	= 0,	/* Allocate from $MFT zone. */
-	DATA_ZONE	= 1,	/* Allocate from $DATA zone. */
-	LAST_ZONE	= 1,	/* For sanity checking. */
-} NTFS_CLUSTER_ALLOCATION_ZONES;
-
-extern runlist_element *ntfs_cluster_alloc(ntfs_volume *vol,
-		const VCN start_vcn, const s64 count, const LCN start_lcn,
-		const NTFS_CLUSTER_ALLOCATION_ZONES zone,
-		const bool is_extension);
-
-extern s64 __ntfs_cluster_free(ntfs_inode *ni, const VCN start_vcn,
-		s64 count, ntfs_attr_search_ctx *ctx, const bool is_rollback);
-
-/**
- * ntfs_cluster_free - free clusters on an ntfs volume
- * @ni:		ntfs inode whose runlist describes the clusters to free
- * @start_vcn:	vcn in the runlist of @ni at which to start freeing clusters
- * @count:	number of clusters to free or -1 for all clusters
- * @ctx:	active attribute search context if present or NULL if not
- *
- * Free @count clusters starting at the cluster @start_vcn in the runlist
- * described by the ntfs inode @ni.
- *
- * If @count is -1, all clusters from @start_vcn to the end of the runlist are
- * deallocated.  Thus, to completely free all clusters in a runlist, use
- * @start_vcn = 0 and @count = -1.
- *
- * If @ctx is specified, it is an active search context of @ni and its base mft
- * record.  This is needed when ntfs_cluster_free() encounters unmapped runlist
- * fragments and allows their mapping.  If you do not have the mft record
- * mapped, you can specify @ctx as NULL and ntfs_cluster_free() will perform
- * the necessary mapping and unmapping.
- *
- * Note, ntfs_cluster_free() saves the state of @ctx on entry and restores it
- * before returning.  Thus, @ctx will be left pointing to the same attribute on
- * return as on entry.  However, the actual pointers in @ctx may point to
- * different memory locations on return, so you must remember to reset any
- * cached pointers from the @ctx, i.e. after the call to ntfs_cluster_free(),
- * you will probably want to do:
- *	m = ctx->mrec;
- *	a = ctx->attr;
- * Assuming you cache ctx->attr in a variable @a of type ATTR_RECORD * and that
- * you cache ctx->mrec in a variable @m of type MFT_RECORD *.
- *
- * Note, ntfs_cluster_free() does not modify the runlist, so you have to remove
- * from the runlist or mark sparse the freed runs later.
- *
- * Return the number of deallocated clusters (not counting sparse ones) on
- * success and -errno on error.
- *
- * WARNING: If @ctx is supplied, regardless of whether success or failure is
- *	    returned, you need to check IS_ERR(@ctx->mrec) and if 'true' the @ctx
- *	    is no longer valid, i.e. you need to either call
- *	    ntfs_attr_reinit_search_ctx() or ntfs_attr_put_search_ctx() on it.
- *	    In that case PTR_ERR(@ctx->mrec) will give you the error code for
- *	    why the mapping of the old inode failed.
- *
- * Locking: - The runlist described by @ni must be locked for writing on entry
- *	      and is locked on return.  Note the runlist may be modified when
- *	      needed runlist fragments need to be mapped.
- *	    - The volume lcn bitmap must be unlocked on entry and is unlocked
- *	      on return.
- *	    - This function takes the volume lcn bitmap lock for writing and
- *	      modifies the bitmap contents.
- *	    - If @ctx is NULL, the base mft record of @ni must not be mapped on
- *	      entry and it will be left unmapped on return.
- *	    - If @ctx is not NULL, the base mft record must be mapped on entry
- *	      and it will be left mapped on return.
- */
-static inline s64 ntfs_cluster_free(ntfs_inode *ni, const VCN start_vcn,
-		s64 count, ntfs_attr_search_ctx *ctx)
-{
-	return __ntfs_cluster_free(ni, start_vcn, count, ctx, false);
-}
-
-extern int ntfs_cluster_free_from_rl_nolock(ntfs_volume *vol,
-		const runlist_element *rl);
-
-/**
- * ntfs_cluster_free_from_rl - free clusters from runlist
- * @vol:	mounted ntfs volume on which to free the clusters
- * @rl:		runlist describing the clusters to free
- *
- * Free all the clusters described by the runlist @rl on the volume @vol.  In
- * the case of an error being returned, at least some of the clusters were not
- * freed.
- *
- * Return 0 on success and -errno on error.
- *
- * Locking: - This function takes the volume lcn bitmap lock for writing and
- *	      modifies the bitmap contents.
- *	    - The caller must have locked the runlist @rl for reading or
- *	      writing.
- */
-static inline int ntfs_cluster_free_from_rl(ntfs_volume *vol,
-		const runlist_element *rl)
-{
-	int ret;
-
-	down_write(&vol->lcnbmp_lock);
-	ret = ntfs_cluster_free_from_rl_nolock(vol, rl);
-	up_write(&vol->lcnbmp_lock);
-	return ret;
-}
-
-#endif /* NTFS_RW */
-
-#endif /* defined _LINUX_NTFS_LCNALLOC_H */
diff --git a/fs/ntfs/logfile.c b/fs/ntfs/logfile.c
deleted file mode 100644
index 6ce60ffc6a..0000000000
--- a/fs/ntfs/logfile.c
+++ /dev/null
@@ -1,849 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * logfile.c - NTFS kernel journal handling. Part of the Linux-NTFS project.
- *
- * Copyright (c) 2002-2007 Anton Altaparmakov
- */
-
-#ifdef NTFS_RW
-
-#include <linux/types.h>
-#include <linux/fs.h>
-#include <linux/highmem.h>
-#include <linux/buffer_head.h>
-#include <linux/bitops.h>
-#include <linux/log2.h>
-#include <linux/bio.h>
-
-#include "attrib.h"
-#include "aops.h"
-#include "debug.h"
-#include "logfile.h"
-#include "malloc.h"
-#include "volume.h"
-#include "ntfs.h"
-
-/**
- * ntfs_check_restart_page_header - check the page header for consistency
- * @vi:		$LogFile inode to which the restart page header belongs
- * @rp:		restart page header to check
- * @pos:	position in @vi at which the restart page header resides
- *
- * Check the restart page header @rp for consistency and return 'true' if it is
- * consistent and 'false' otherwise.
- *
- * This function only needs NTFS_BLOCK_SIZE bytes in @rp, i.e. it does not
- * require the full restart page.
- */
-static bool ntfs_check_restart_page_header(struct inode *vi,
-		RESTART_PAGE_HEADER *rp, s64 pos)
-{
-	u32 logfile_system_page_size, logfile_log_page_size;
-	u16 ra_ofs, usa_count, usa_ofs, usa_end = 0;
-	bool have_usa = true;
-
-	ntfs_debug("Entering.");
-	/*
-	 * If the system or log page sizes are smaller than the ntfs block size
-	 * or either is not a power of 2 we cannot handle this log file.
-	 */
-	logfile_system_page_size = le32_to_cpu(rp->system_page_size);
-	logfile_log_page_size = le32_to_cpu(rp->log_page_size);
-	if (logfile_system_page_size < NTFS_BLOCK_SIZE ||
-			logfile_log_page_size < NTFS_BLOCK_SIZE ||
-			logfile_system_page_size &
-			(logfile_system_page_size - 1) ||
-			!is_power_of_2(logfile_log_page_size)) {
-		ntfs_error(vi->i_sb, "$LogFile uses unsupported page size.");
-		return false;
-	}
-	/*
-	 * We must be either at !pos (1st restart page) or at pos = system page
-	 * size (2nd restart page).
-	 */
-	if (pos && pos != logfile_system_page_size) {
-		ntfs_error(vi->i_sb, "Found restart area in incorrect "
-				"position in $LogFile.");
-		return false;
-	}
-	/* We only know how to handle version 1.1. */
-	if (sle16_to_cpu(rp->major_ver) != 1 ||
-			sle16_to_cpu(rp->minor_ver) != 1) {
-		ntfs_error(vi->i_sb, "$LogFile version %i.%i is not "
-				"supported.  (This driver supports version "
-				"1.1 only.)", (int)sle16_to_cpu(rp->major_ver),
-				(int)sle16_to_cpu(rp->minor_ver));
-		return false;
-	}
-	/*
-	 * If chkdsk has been run the restart page may not be protected by an
-	 * update sequence array.
-	 */
-	if (ntfs_is_chkd_record(rp->magic) && !le16_to_cpu(rp->usa_count)) {
-		have_usa = false;
-		goto skip_usa_checks;
-	}
-	/* Verify the size of the update sequence array. */
-	usa_count = 1 + (logfile_system_page_size >> NTFS_BLOCK_SIZE_BITS);
-	if (usa_count != le16_to_cpu(rp->usa_count)) {
-		ntfs_error(vi->i_sb, "$LogFile restart page specifies "
-				"inconsistent update sequence array count.");
-		return false;
-	}
-	/* Verify the position of the update sequence array. */
-	usa_ofs = le16_to_cpu(rp->usa_ofs);
-	usa_end = usa_ofs + usa_count * sizeof(u16);
-	if (usa_ofs < sizeof(RESTART_PAGE_HEADER) ||
-			usa_end > NTFS_BLOCK_SIZE - sizeof(u16)) {
-		ntfs_error(vi->i_sb, "$LogFile restart page specifies "
-				"inconsistent update sequence array offset.");
-		return false;
-	}
-skip_usa_checks:
-	/*
-	 * Verify the position of the restart area.  It must be:
-	 *	- aligned to 8-byte boundary,
-	 *	- after the update sequence array, and
-	 *	- within the system page size.
-	 */
-	ra_ofs = le16_to_cpu(rp->restart_area_offset);
-	if (ra_ofs & 7 || (have_usa ? ra_ofs < usa_end :
-			ra_ofs < sizeof(RESTART_PAGE_HEADER)) ||
-			ra_ofs > logfile_system_page_size) {
-		ntfs_error(vi->i_sb, "$LogFile restart page specifies "
-				"inconsistent restart area offset.");
-		return false;
-	}
-	/*
-	 * Only restart pages modified by chkdsk are allowed to have chkdsk_lsn
-	 * set.
-	 */
-	if (!ntfs_is_chkd_record(rp->magic) && sle64_to_cpu(rp->chkdsk_lsn)) {
-		ntfs_error(vi->i_sb, "$LogFile restart page is not modified "
-				"by chkdsk but a chkdsk LSN is specified.");
-		return false;
-	}
-	ntfs_debug("Done.");
-	return true;
-}
-
-/**
- * ntfs_check_restart_area - check the restart area for consistency
- * @vi:		$LogFile inode to which the restart page belongs
- * @rp:		restart page whose restart area to check
- *
- * Check the restart area of the restart page @rp for consistency and return
- * 'true' if it is consistent and 'false' otherwise.
- *
- * This function assumes that the restart page header has already been
- * consistency checked.
- *
- * This function only needs NTFS_BLOCK_SIZE bytes in @rp, i.e. it does not
- * require the full restart page.
- */
-static bool ntfs_check_restart_area(struct inode *vi, RESTART_PAGE_HEADER *rp)
-{
-	u64 file_size;
-	RESTART_AREA *ra;
-	u16 ra_ofs, ra_len, ca_ofs;
-	u8 fs_bits;
-
-	ntfs_debug("Entering.");
-	ra_ofs = le16_to_cpu(rp->restart_area_offset);
-	ra = (RESTART_AREA*)((u8*)rp + ra_ofs);
-	/*
-	 * Everything before ra->file_size must be before the first word
-	 * protected by an update sequence number.  This ensures that it is
-	 * safe to access ra->client_array_offset.
-	 */
-	if (ra_ofs + offsetof(RESTART_AREA, file_size) >
-			NTFS_BLOCK_SIZE - sizeof(u16)) {
-		ntfs_error(vi->i_sb, "$LogFile restart area specifies "
-				"inconsistent file offset.");
-		return false;
-	}
-	/*
-	 * Now that we can access ra->client_array_offset, make sure everything
-	 * up to the log client array is before the first word protected by an
-	 * update sequence number.  This ensures we can access all of the
-	 * restart area elements safely.  Also, the client array offset must be
-	 * aligned to an 8-byte boundary.
-	 */
-	ca_ofs = le16_to_cpu(ra->client_array_offset);
-	if (((ca_ofs + 7) & ~7) != ca_ofs ||
-			ra_ofs + ca_ofs > NTFS_BLOCK_SIZE - sizeof(u16)) {
-		ntfs_error(vi->i_sb, "$LogFile restart area specifies "
-				"inconsistent client array offset.");
-		return false;
-	}
-	/*
-	 * The restart area must end within the system page size both when
-	 * calculated manually and as specified by ra->restart_area_length.
-	 * Also, the calculated length must not exceed the specified length.
-	 */
-	ra_len = ca_ofs + le16_to_cpu(ra->log_clients) *
-			sizeof(LOG_CLIENT_RECORD);
-	if (ra_ofs + ra_len > le32_to_cpu(rp->system_page_size) ||
-			ra_ofs + le16_to_cpu(ra->restart_area_length) >
-			le32_to_cpu(rp->system_page_size) ||
-			ra_len > le16_to_cpu(ra->restart_area_length)) {
-		ntfs_error(vi->i_sb, "$LogFile restart area is out of bounds "
-				"of the system page size specified by the "
-				"restart page header and/or the specified "
-				"restart area length is inconsistent.");
-		return false;
-	}
-	/*
-	 * The ra->client_free_list and ra->client_in_use_list must be either
-	 * LOGFILE_NO_CLIENT or less than ra->log_clients or they are
-	 * overflowing the client array.
-	 */
-	if ((ra->client_free_list != LOGFILE_NO_CLIENT &&
-			le16_to_cpu(ra->client_free_list) >=
-			le16_to_cpu(ra->log_clients)) ||
-			(ra->client_in_use_list != LOGFILE_NO_CLIENT &&
-			le16_to_cpu(ra->client_in_use_list) >=
-			le16_to_cpu(ra->log_clients))) {
-		ntfs_error(vi->i_sb, "$LogFile restart area specifies "
-				"overflowing client free and/or in use lists.");
-		return false;
-	}
-	/*
-	 * Check ra->seq_number_bits against ra->file_size for consistency.
-	 * We cannot just use ffs() because the file size is not a power of 2.
-	 */
-	file_size = (u64)sle64_to_cpu(ra->file_size);
-	fs_bits = 0;
-	while (file_size) {
-		file_size >>= 1;
-		fs_bits++;
-	}
-	if (le32_to_cpu(ra->seq_number_bits) != 67 - fs_bits) {
-		ntfs_error(vi->i_sb, "$LogFile restart area specifies "
-				"inconsistent sequence number bits.");
-		return false;
-	}
-	/* The log record header length must be a multiple of 8. */
-	if (((le16_to_cpu(ra->log_record_header_length) + 7) & ~7) !=
-			le16_to_cpu(ra->log_record_header_length)) {
-		ntfs_error(vi->i_sb, "$LogFile restart area specifies "
-				"inconsistent log record header length.");
-		return false;
-	}
-	/* Dito for the log page data offset. */
-	if (((le16_to_cpu(ra->log_page_data_offset) + 7) & ~7) !=
-			le16_to_cpu(ra->log_page_data_offset)) {
-		ntfs_error(vi->i_sb, "$LogFile restart area specifies "
-				"inconsistent log page data offset.");
-		return false;
-	}
-	ntfs_debug("Done.");
-	return true;
-}
-
-/**
- * ntfs_check_log_client_array - check the log client array for consistency
- * @vi:		$LogFile inode to which the restart page belongs
- * @rp:		restart page whose log client array to check
- *
- * Check the log client array of the restart page @rp for consistency and
- * return 'true' if it is consistent and 'false' otherwise.
- *
- * This function assumes that the restart page header and the restart area have
- * already been consistency checked.
- *
- * Unlike ntfs_check_restart_page_header() and ntfs_check_restart_area(), this
- * function needs @rp->system_page_size bytes in @rp, i.e. it requires the full
- * restart page and the page must be multi sector transfer deprotected.
- */
-static bool ntfs_check_log_client_array(struct inode *vi,
-		RESTART_PAGE_HEADER *rp)
-{
-	RESTART_AREA *ra;
-	LOG_CLIENT_RECORD *ca, *cr;
-	u16 nr_clients, idx;
-	bool in_free_list, idx_is_first;
-
-	ntfs_debug("Entering.");
-	ra = (RESTART_AREA*)((u8*)rp + le16_to_cpu(rp->restart_area_offset));
-	ca = (LOG_CLIENT_RECORD*)((u8*)ra +
-			le16_to_cpu(ra->client_array_offset));
-	/*
-	 * Check the ra->client_free_list first and then check the
-	 * ra->client_in_use_list.  Check each of the log client records in
-	 * each of the lists and check that the array does not overflow the
-	 * ra->log_clients value.  Also keep track of the number of records
-	 * visited as there cannot be more than ra->log_clients records and
-	 * that way we detect eventual loops in within a list.
-	 */
-	nr_clients = le16_to_cpu(ra->log_clients);
-	idx = le16_to_cpu(ra->client_free_list);
-	in_free_list = true;
-check_list:
-	for (idx_is_first = true; idx != LOGFILE_NO_CLIENT_CPU; nr_clients--,
-			idx = le16_to_cpu(cr->next_client)) {
-		if (!nr_clients || idx >= le16_to_cpu(ra->log_clients))
-			goto err_out;
-		/* Set @cr to the current log client record. */
-		cr = ca + idx;
-		/* The first log client record must not have a prev_client. */
-		if (idx_is_first) {
-			if (cr->prev_client != LOGFILE_NO_CLIENT)
-				goto err_out;
-			idx_is_first = false;
-		}
-	}
-	/* Switch to and check the in use list if we just did the free list. */
-	if (in_free_list) {
-		in_free_list = false;
-		idx = le16_to_cpu(ra->client_in_use_list);
-		goto check_list;
-	}
-	ntfs_debug("Done.");
-	return true;
-err_out:
-	ntfs_error(vi->i_sb, "$LogFile log client array is corrupt.");
-	return false;
-}
-
-/**
- * ntfs_check_and_load_restart_page - check the restart page for consistency
- * @vi:		$LogFile inode to which the restart page belongs
- * @rp:		restart page to check
- * @pos:	position in @vi at which the restart page resides
- * @wrp:	[OUT] copy of the multi sector transfer deprotected restart page
- * @lsn:	[OUT] set to the current logfile lsn on success
- *
- * Check the restart page @rp for consistency and return 0 if it is consistent
- * and -errno otherwise.  The restart page may have been modified by chkdsk in
- * which case its magic is CHKD instead of RSTR.
- *
- * This function only needs NTFS_BLOCK_SIZE bytes in @rp, i.e. it does not
- * require the full restart page.
- *
- * If @wrp is not NULL, on success, *@wrp will point to a buffer containing a
- * copy of the complete multi sector transfer deprotected page.  On failure,
- * *@wrp is undefined.
- *
- * Simillarly, if @lsn is not NULL, on success *@lsn will be set to the current
- * logfile lsn according to this restart page.  On failure, *@lsn is undefined.
- *
- * The following error codes are defined:
- *	-EINVAL	- The restart page is inconsistent.
- *	-ENOMEM	- Not enough memory to load the restart page.
- *	-EIO	- Failed to reading from $LogFile.
- */
-static int ntfs_check_and_load_restart_page(struct inode *vi,
-		RESTART_PAGE_HEADER *rp, s64 pos, RESTART_PAGE_HEADER **wrp,
-		LSN *lsn)
-{
-	RESTART_AREA *ra;
-	RESTART_PAGE_HEADER *trp;
-	int size, err;
-
-	ntfs_debug("Entering.");
-	/* Check the restart page header for consistency. */
-	if (!ntfs_check_restart_page_header(vi, rp, pos)) {
-		/* Error output already done inside the function. */
-		return -EINVAL;
-	}
-	/* Check the restart area for consistency. */
-	if (!ntfs_check_restart_area(vi, rp)) {
-		/* Error output already done inside the function. */
-		return -EINVAL;
-	}
-	ra = (RESTART_AREA*)((u8*)rp + le16_to_cpu(rp->restart_area_offset));
-	/*
-	 * Allocate a buffer to store the whole restart page so we can multi
-	 * sector transfer deprotect it.
-	 */
-	trp = ntfs_malloc_nofs(le32_to_cpu(rp->system_page_size));
-	if (!trp) {
-		ntfs_error(vi->i_sb, "Failed to allocate memory for $LogFile "
-				"restart page buffer.");
-		return -ENOMEM;
-	}
-	/*
-	 * Read the whole of the restart page into the buffer.  If it fits
-	 * completely inside @rp, just copy it from there.  Otherwise map all
-	 * the required pages and copy the data from them.
-	 */
-	size = PAGE_SIZE - (pos & ~PAGE_MASK);
-	if (size >= le32_to_cpu(rp->system_page_size)) {
-		memcpy(trp, rp, le32_to_cpu(rp->system_page_size));
-	} else {
-		pgoff_t idx;
-		struct page *page;
-		int have_read, to_read;
-
-		/* First copy what we already have in @rp. */
-		memcpy(trp, rp, size);
-		/* Copy the remaining data one page at a time. */
-		have_read = size;
-		to_read = le32_to_cpu(rp->system_page_size) - size;
-		idx = (pos + size) >> PAGE_SHIFT;
-		BUG_ON((pos + size) & ~PAGE_MASK);
-		do {
-			page = ntfs_map_page(vi->i_mapping, idx);
-			if (IS_ERR(page)) {
-				ntfs_error(vi->i_sb, "Error mapping $LogFile "
-						"page (index %lu).", idx);
-				err = PTR_ERR(page);
-				if (err != -EIO && err != -ENOMEM)
-					err = -EIO;
-				goto err_out;
-			}
-			size = min_t(int, to_read, PAGE_SIZE);
-			memcpy((u8*)trp + have_read, page_address(page), size);
-			ntfs_unmap_page(page);
-			have_read += size;
-			to_read -= size;
-			idx++;
-		} while (to_read > 0);
-	}
-	/*
-	 * Perform the multi sector transfer deprotection on the buffer if the
-	 * restart page is protected.
-	 */
-	if ((!ntfs_is_chkd_record(trp->magic) || le16_to_cpu(trp->usa_count))
-			&& post_read_mst_fixup((NTFS_RECORD*)trp,
-			le32_to_cpu(rp->system_page_size))) {
-		/*
-		 * A multi sector tranfer error was detected.  We only need to
-		 * abort if the restart page contents exceed the multi sector
-		 * transfer fixup of the first sector.
-		 */
-		if (le16_to_cpu(rp->restart_area_offset) +
-				le16_to_cpu(ra->restart_area_length) >
-				NTFS_BLOCK_SIZE - sizeof(u16)) {
-			ntfs_error(vi->i_sb, "Multi sector transfer error "
-					"detected in $LogFile restart page.");
-			err = -EINVAL;
-			goto err_out;
-		}
-	}
-	/*
-	 * If the restart page is modified by chkdsk or there are no active
-	 * logfile clients, the logfile is consistent.  Otherwise, need to
-	 * check the log client records for consistency, too.
-	 */
-	err = 0;
-	if (ntfs_is_rstr_record(rp->magic) &&
-			ra->client_in_use_list != LOGFILE_NO_CLIENT) {
-		if (!ntfs_check_log_client_array(vi, trp)) {
-			err = -EINVAL;
-			goto err_out;
-		}
-	}
-	if (lsn) {
-		if (ntfs_is_rstr_record(rp->magic))
-			*lsn = sle64_to_cpu(ra->current_lsn);
-		else /* if (ntfs_is_chkd_record(rp->magic)) */
-			*lsn = sle64_to_cpu(rp->chkdsk_lsn);
-	}
-	ntfs_debug("Done.");
-	if (wrp)
-		*wrp = trp;
-	else {
-err_out:
-		ntfs_free(trp);
-	}
-	return err;
-}
-
-/**
- * ntfs_check_logfile - check the journal for consistency
- * @log_vi:	struct inode of loaded journal $LogFile to check
- * @rp:		[OUT] on success this is a copy of the current restart page
- *
- * Check the $LogFile journal for consistency and return 'true' if it is
- * consistent and 'false' if not.  On success, the current restart page is
- * returned in *@rp.  Caller must call ntfs_free(*@rp) when finished with it.
- *
- * At present we only check the two restart pages and ignore the log record
- * pages.
- *
- * Note that the MstProtected flag is not set on the $LogFile inode and hence
- * when reading pages they are not deprotected.  This is because we do not know
- * if the $LogFile was created on a system with a different page size to ours
- * yet and mst deprotection would fail if our page size is smaller.
- */
-bool ntfs_check_logfile(struct inode *log_vi, RESTART_PAGE_HEADER **rp)
-{
-	s64 size, pos;
-	LSN rstr1_lsn, rstr2_lsn;
-	ntfs_volume *vol = NTFS_SB(log_vi->i_sb);
-	struct address_space *mapping = log_vi->i_mapping;
-	struct page *page = NULL;
-	u8 *kaddr = NULL;
-	RESTART_PAGE_HEADER *rstr1_ph = NULL;
-	RESTART_PAGE_HEADER *rstr2_ph = NULL;
-	int log_page_size, err;
-	bool logfile_is_empty = true;
-	u8 log_page_bits;
-
-	ntfs_debug("Entering.");
-	/* An empty $LogFile must have been clean before it got emptied. */
-	if (NVolLogFileEmpty(vol))
-		goto is_empty;
-	size = i_size_read(log_vi);
-	/* Make sure the file doesn't exceed the maximum allowed size. */
-	if (size > MaxLogFileSize)
-		size = MaxLogFileSize;
-	/*
-	 * Truncate size to a multiple of the page cache size or the default
-	 * log page size if the page cache size is between the default log page
-	 * log page size if the page cache size is between the default log page
-	 * size and twice that.
-	 */
-	if (PAGE_SIZE >= DefaultLogPageSize && PAGE_SIZE <=
-			DefaultLogPageSize * 2)
-		log_page_size = DefaultLogPageSize;
-	else
-		log_page_size = PAGE_SIZE;
-	/*
-	 * Use ntfs_ffs() instead of ffs() to enable the compiler to
-	 * optimize log_page_size and log_page_bits into constants.
-	 */
-	log_page_bits = ntfs_ffs(log_page_size) - 1;
-	size &= ~(s64)(log_page_size - 1);
-	/*
-	 * Ensure the log file is big enough to store at least the two restart
-	 * pages and the minimum number of log record pages.
-	 */
-	if (size < log_page_size * 2 || (size - log_page_size * 2) >>
-			log_page_bits < MinLogRecordPages) {
-		ntfs_error(vol->sb, "$LogFile is too small.");
-		return false;
-	}
-	/*
-	 * Read through the file looking for a restart page.  Since the restart
-	 * page header is at the beginning of a page we only need to search at
-	 * what could be the beginning of a page (for each page size) rather
-	 * than scanning the whole file byte by byte.  If all potential places
-	 * contain empty and uninitialzed records, the log file can be assumed
-	 * to be empty.
-	 */
-	for (pos = 0; pos < size; pos <<= 1) {
-		pgoff_t idx = pos >> PAGE_SHIFT;
-		if (!page || page->index != idx) {
-			if (page)
-				ntfs_unmap_page(page);
-			page = ntfs_map_page(mapping, idx);
-			if (IS_ERR(page)) {
-				ntfs_error(vol->sb, "Error mapping $LogFile "
-						"page (index %lu).", idx);
-				goto err_out;
-			}
-		}
-		kaddr = (u8*)page_address(page) + (pos & ~PAGE_MASK);
-		/*
-		 * A non-empty block means the logfile is not empty while an
-		 * empty block after a non-empty block has been encountered
-		 * means we are done.
-		 */
-		if (!ntfs_is_empty_recordp((le32*)kaddr))
-			logfile_is_empty = false;
-		else if (!logfile_is_empty)
-			break;
-		/*
-		 * A log record page means there cannot be a restart page after
-		 * this so no need to continue searching.
-		 */
-		if (ntfs_is_rcrd_recordp((le32*)kaddr))
-			break;
-		/* If not a (modified by chkdsk) restart page, continue. */
-		if (!ntfs_is_rstr_recordp((le32*)kaddr) &&
-				!ntfs_is_chkd_recordp((le32*)kaddr)) {
-			if (!pos)
-				pos = NTFS_BLOCK_SIZE >> 1;
-			continue;
-		}
-		/*
-		 * Check the (modified by chkdsk) restart page for consistency
-		 * and get a copy of the complete multi sector transfer
-		 * deprotected restart page.
-		 */
-		err = ntfs_check_and_load_restart_page(log_vi,
-				(RESTART_PAGE_HEADER*)kaddr, pos,
-				!rstr1_ph ? &rstr1_ph : &rstr2_ph,
-				!rstr1_ph ? &rstr1_lsn : &rstr2_lsn);
-		if (!err) {
-			/*
-			 * If we have now found the first (modified by chkdsk)
-			 * restart page, continue looking for the second one.
-			 */
-			if (!pos) {
-				pos = NTFS_BLOCK_SIZE >> 1;
-				continue;
-			}
-			/*
-			 * We have now found the second (modified by chkdsk)
-			 * restart page, so we can stop looking.
-			 */
-			break;
-		}
-		/*
-		 * Error output already done inside the function.  Note, we do
-		 * not abort if the restart page was invalid as we might still
-		 * find a valid one further in the file.
-		 */
-		if (err != -EINVAL) {
-			ntfs_unmap_page(page);
-			goto err_out;
-		}
-		/* Continue looking. */
-		if (!pos)
-			pos = NTFS_BLOCK_SIZE >> 1;
-	}
-	if (page)
-		ntfs_unmap_page(page);
-	if (logfile_is_empty) {
-		NVolSetLogFileEmpty(vol);
-is_empty:
-		ntfs_debug("Done.  ($LogFile is empty.)");
-		return true;
-	}
-	if (!rstr1_ph) {
-		BUG_ON(rstr2_ph);
-		ntfs_error(vol->sb, "Did not find any restart pages in "
-				"$LogFile and it was not empty.");
-		return false;
-	}
-	/* If both restart pages were found, use the more recent one. */
-	if (rstr2_ph) {
-		/*
-		 * If the second restart area is more recent, switch to it.
-		 * Otherwise just throw it away.
-		 */
-		if (rstr2_lsn > rstr1_lsn) {
-			ntfs_debug("Using second restart page as it is more "
-					"recent.");
-			ntfs_free(rstr1_ph);
-			rstr1_ph = rstr2_ph;
-			/* rstr1_lsn = rstr2_lsn; */
-		} else {
-			ntfs_debug("Using first restart page as it is more "
-					"recent.");
-			ntfs_free(rstr2_ph);
-		}
-		rstr2_ph = NULL;
-	}
-	/* All consistency checks passed. */
-	if (rp)
-		*rp = rstr1_ph;
-	else
-		ntfs_free(rstr1_ph);
-	ntfs_debug("Done.");
-	return true;
-err_out:
-	if (rstr1_ph)
-		ntfs_free(rstr1_ph);
-	return false;
-}
-
-/**
- * ntfs_is_logfile_clean - check in the journal if the volume is clean
- * @log_vi:	struct inode of loaded journal $LogFile to check
- * @rp:		copy of the current restart page
- *
- * Analyze the $LogFile journal and return 'true' if it indicates the volume was
- * shutdown cleanly and 'false' if not.
- *
- * At present we only look at the two restart pages and ignore the log record
- * pages.  This is a little bit crude in that there will be a very small number
- * of cases where we think that a volume is dirty when in fact it is clean.
- * This should only affect volumes that have not been shutdown cleanly but did
- * not have any pending, non-check-pointed i/o, i.e. they were completely idle
- * at least for the five seconds preceding the unclean shutdown.
- *
- * This function assumes that the $LogFile journal has already been consistency
- * checked by a call to ntfs_check_logfile() and in particular if the $LogFile
- * is empty this function requires that NVolLogFileEmpty() is true otherwise an
- * empty volume will be reported as dirty.
- */
-bool ntfs_is_logfile_clean(struct inode *log_vi, const RESTART_PAGE_HEADER *rp)
-{
-	ntfs_volume *vol = NTFS_SB(log_vi->i_sb);
-	RESTART_AREA *ra;
-
-	ntfs_debug("Entering.");
-	/* An empty $LogFile must have been clean before it got emptied. */
-	if (NVolLogFileEmpty(vol)) {
-		ntfs_debug("Done.  ($LogFile is empty.)");
-		return true;
-	}
-	BUG_ON(!rp);
-	if (!ntfs_is_rstr_record(rp->magic) &&
-			!ntfs_is_chkd_record(rp->magic)) {
-		ntfs_error(vol->sb, "Restart page buffer is invalid.  This is "
-				"probably a bug in that the $LogFile should "
-				"have been consistency checked before calling "
-				"this function.");
-		return false;
-	}
-	ra = (RESTART_AREA*)((u8*)rp + le16_to_cpu(rp->restart_area_offset));
-	/*
-	 * If the $LogFile has active clients, i.e. it is open, and we do not
-	 * have the RESTART_VOLUME_IS_CLEAN bit set in the restart area flags,
-	 * we assume there was an unclean shutdown.
-	 */
-	if (ra->client_in_use_list != LOGFILE_NO_CLIENT &&
-			!(ra->flags & RESTART_VOLUME_IS_CLEAN)) {
-		ntfs_debug("Done.  $LogFile indicates a dirty shutdown.");
-		return false;
-	}
-	/* $LogFile indicates a clean shutdown. */
-	ntfs_debug("Done.  $LogFile indicates a clean shutdown.");
-	return true;
-}
-
-/**
- * ntfs_empty_logfile - empty the contents of the $LogFile journal
- * @log_vi:	struct inode of loaded journal $LogFile to empty
- *
- * Empty the contents of the $LogFile journal @log_vi and return 'true' on
- * success and 'false' on error.
- *
- * This function assumes that the $LogFile journal has already been consistency
- * checked by a call to ntfs_check_logfile() and that ntfs_is_logfile_clean()
- * has been used to ensure that the $LogFile is clean.
- */
-bool ntfs_empty_logfile(struct inode *log_vi)
-{
-	VCN vcn, end_vcn;
-	ntfs_inode *log_ni = NTFS_I(log_vi);
-	ntfs_volume *vol = log_ni->vol;
-	struct super_block *sb = vol->sb;
-	runlist_element *rl;
-	unsigned long flags;
-	unsigned block_size, block_size_bits;
-	int err;
-	bool should_wait = true;
-
-	ntfs_debug("Entering.");
-	if (NVolLogFileEmpty(vol)) {
-		ntfs_debug("Done.");
-		return true;
-	}
-	/*
-	 * We cannot use ntfs_attr_set() because we may be still in the middle
-	 * of a mount operation.  Thus we do the emptying by hand by first
-	 * zapping the page cache pages for the $LogFile/$DATA attribute and
-	 * then emptying each of the buffers in each of the clusters specified
-	 * by the runlist by hand.
-	 */
-	block_size = sb->s_blocksize;
-	block_size_bits = sb->s_blocksize_bits;
-	vcn = 0;
-	read_lock_irqsave(&log_ni->size_lock, flags);
-	end_vcn = (log_ni->initialized_size + vol->cluster_size_mask) >>
-			vol->cluster_size_bits;
-	read_unlock_irqrestore(&log_ni->size_lock, flags);
-	truncate_inode_pages(log_vi->i_mapping, 0);
-	down_write(&log_ni->runlist.lock);
-	rl = log_ni->runlist.rl;
-	if (unlikely(!rl || vcn < rl->vcn || !rl->length)) {
-map_vcn:
-		err = ntfs_map_runlist_nolock(log_ni, vcn, NULL);
-		if (err) {
-			ntfs_error(sb, "Failed to map runlist fragment (error "
-					"%d).", -err);
-			goto err;
-		}
-		rl = log_ni->runlist.rl;
-		BUG_ON(!rl || vcn < rl->vcn || !rl->length);
-	}
-	/* Seek to the runlist element containing @vcn. */
-	while (rl->length && vcn >= rl[1].vcn)
-		rl++;
-	do {
-		LCN lcn;
-		sector_t block, end_block;
-		s64 len;
-
-		/*
-		 * If this run is not mapped map it now and start again as the
-		 * runlist will have been updated.
-		 */
-		lcn = rl->lcn;
-		if (unlikely(lcn == LCN_RL_NOT_MAPPED)) {
-			vcn = rl->vcn;
-			goto map_vcn;
-		}
-		/* If this run is not valid abort with an error. */
-		if (unlikely(!rl->length || lcn < LCN_HOLE))
-			goto rl_err;
-		/* Skip holes. */
-		if (lcn == LCN_HOLE)
-			continue;
-		block = lcn << vol->cluster_size_bits >> block_size_bits;
-		len = rl->length;
-		if (rl[1].vcn > end_vcn)
-			len = end_vcn - rl->vcn;
-		end_block = (lcn + len) << vol->cluster_size_bits >>
-				block_size_bits;
-		/* Iterate over the blocks in the run and empty them. */
-		do {
-			struct buffer_head *bh;
-
-			/* Obtain the buffer, possibly not uptodate. */
-			bh = sb_getblk(sb, block);
-			BUG_ON(!bh);
-			/* Setup buffer i/o submission. */
-			lock_buffer(bh);
-			bh->b_end_io = end_buffer_write_sync;
-			get_bh(bh);
-			/* Set the entire contents of the buffer to 0xff. */
-			memset(bh->b_data, -1, block_size);
-			if (!buffer_uptodate(bh))
-				set_buffer_uptodate(bh);
-			if (buffer_dirty(bh))
-				clear_buffer_dirty(bh);
-			/*
-			 * Submit the buffer and wait for i/o to complete but
-			 * only for the first buffer so we do not miss really
-			 * serious i/o errors.  Once the first buffer has
-			 * completed ignore errors afterwards as we can assume
-			 * that if one buffer worked all of them will work.
-			 */
-			submit_bh(REQ_OP_WRITE, bh);
-			if (should_wait) {
-				should_wait = false;
-				wait_on_buffer(bh);
-				if (unlikely(!buffer_uptodate(bh)))
-					goto io_err;
-			}
-			brelse(bh);
-		} while (++block < end_block);
-	} while ((++rl)->vcn < end_vcn);
-	up_write(&log_ni->runlist.lock);
-	/*
-	 * Zap the pages again just in case any got instantiated whilst we were
-	 * emptying the blocks by hand.  FIXME: We may not have completed
-	 * writing to all the buffer heads yet so this may happen too early.
-	 * We really should use a kernel thread to do the emptying
-	 * asynchronously and then we can also set the volume dirty and output
-	 * an error message if emptying should fail.
-	 */
-	truncate_inode_pages(log_vi->i_mapping, 0);
-	/* Set the flag so we do not have to do it again on remount. */
-	NVolSetLogFileEmpty(vol);
-	ntfs_debug("Done.");
-	return true;
-io_err:
-	ntfs_error(sb, "Failed to write buffer.  Unmount and run chkdsk.");
-	goto dirty_err;
-rl_err:
-	ntfs_error(sb, "Runlist is corrupt.  Unmount and run chkdsk.");
-dirty_err:
-	NVolSetErrors(vol);
-	err = -EIO;
-err:
-	up_write(&log_ni->runlist.lock);
-	ntfs_error(sb, "Failed to fill $LogFile with 0xff bytes (error %d).",
-			-err);
-	return false;
-}
-
-#endif /* NTFS_RW */
diff --git a/fs/ntfs/logfile.h b/fs/ntfs/logfile.h
deleted file mode 100644
index 429d4909cc..0000000000
--- a/fs/ntfs/logfile.h
+++ /dev/null
@@ -1,295 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * logfile.h - Defines for NTFS kernel journal ($LogFile) handling.  Part of
- *	       the Linux-NTFS project.
- *
- * Copyright (c) 2000-2005 Anton Altaparmakov
- */
-
-#ifndef _LINUX_NTFS_LOGFILE_H
-#define _LINUX_NTFS_LOGFILE_H
-
-#ifdef NTFS_RW
-
-#include <linux/fs.h>
-
-#include "types.h"
-#include "endian.h"
-#include "layout.h"
-
-/*
- * Journal ($LogFile) organization:
- *
- * Two restart areas present in the first two pages (restart pages, one restart
- * area in each page).  When the volume is dismounted they should be identical,
- * except for the update sequence array which usually has a different update
- * sequence number.
- *
- * These are followed by log records organized in pages headed by a log record
- * header going up to log file size.  Not all pages contain log records when a
- * volume is first formatted, but as the volume ages, all records will be used.
- * When the log file fills up, the records at the beginning are purged (by
- * modifying the oldest_lsn to a higher value presumably) and writing begins
- * at the beginning of the file.  Effectively, the log file is viewed as a
- * circular entity.
- *
- * NOTE: Windows NT, 2000, and XP all use log file version 1.1 but they accept
- * versions <= 1.x, including 0.-1.  (Yes, that is a minus one in there!)  We
- * probably only want to support 1.1 as this seems to be the current version
- * and we don't know how that differs from the older versions.  The only
- * exception is if the journal is clean as marked by the two restart pages
- * then it doesn't matter whether we are on an earlier version.  We can just
- * reinitialize the logfile and start again with version 1.1.
- */
-
-/* Some $LogFile related constants. */
-#define MaxLogFileSize		0x100000000ULL
-#define DefaultLogPageSize	4096
-#define MinLogRecordPages	48
-
-/*
- * Log file restart page header (begins the restart area).
- */
-typedef struct {
-/*Ofs*/
-/*  0	NTFS_RECORD; -- Unfolded here as gcc doesn't like unnamed structs. */
-/*  0*/	NTFS_RECORD_TYPE magic;	/* The magic is "RSTR". */
-/*  4*/	le16 usa_ofs;		/* See NTFS_RECORD definition in layout.h.
-				   When creating, set this to be immediately
-				   after this header structure (without any
-				   alignment). */
-/*  6*/	le16 usa_count;		/* See NTFS_RECORD definition in layout.h. */
-
-/*  8*/	leLSN chkdsk_lsn;	/* The last log file sequence number found by
-				   chkdsk.  Only used when the magic is changed
-				   to "CHKD".  Otherwise this is zero. */
-/* 16*/	le32 system_page_size;	/* Byte size of system pages when the log file
-				   was created, has to be >= 512 and a power of
-				   2.  Use this to calculate the required size
-				   of the usa (usa_count) and add it to usa_ofs.
-				   Then verify that the result is less than the
-				   value of the restart_area_offset. */
-/* 20*/	le32 log_page_size;	/* Byte size of log file pages, has to be >=
-				   512 and a power of 2.  The default is 4096
-				   and is used when the system page size is
-				   between 4096 and 8192.  Otherwise this is
-				   set to the system page size instead. */
-/* 24*/	le16 restart_area_offset;/* Byte offset from the start of this header to
-				   the RESTART_AREA.  Value has to be aligned
-				   to 8-byte boundary.  When creating, set this
-				   to be after the usa. */
-/* 26*/	sle16 minor_ver;	/* Log file minor version.  Only check if major
-				   version is 1. */
-/* 28*/	sle16 major_ver;	/* Log file major version.  We only support
-				   version 1.1. */
-/* sizeof() = 30 (0x1e) bytes */
-} __attribute__ ((__packed__)) RESTART_PAGE_HEADER;
-
-/*
- * Constant for the log client indices meaning that there are no client records
- * in this particular client array.  Also inside the client records themselves,
- * this means that there are no client records preceding or following this one.
- */
-#define LOGFILE_NO_CLIENT	cpu_to_le16(0xffff)
-#define LOGFILE_NO_CLIENT_CPU	0xffff
-
-/*
- * These are the so far known RESTART_AREA_* flags (16-bit) which contain
- * information about the log file in which they are present.
- */
-enum {
-	RESTART_VOLUME_IS_CLEAN	= cpu_to_le16(0x0002),
-	RESTART_SPACE_FILLER	= cpu_to_le16(0xffff), /* gcc: Force enum bit width to 16. */
-} __attribute__ ((__packed__));
-
-typedef le16 RESTART_AREA_FLAGS;
-
-/*
- * Log file restart area record.  The offset of this record is found by adding
- * the offset of the RESTART_PAGE_HEADER to the restart_area_offset value found
- * in it.  See notes at restart_area_offset above.
- */
-typedef struct {
-/*Ofs*/
-/*  0*/	leLSN current_lsn;	/* The current, i.e. last LSN inside the log
-				   when the restart area was last written.
-				   This happens often but what is the interval?
-				   Is it just fixed time or is it every time a
-				   check point is written or somethine else?
-				   On create set to 0. */
-/*  8*/	le16 log_clients;	/* Number of log client records in the array of
-				   log client records which follows this
-				   restart area.  Must be 1.  */
-/* 10*/	le16 client_free_list;	/* The index of the first free log client record
-				   in the array of log client records.
-				   LOGFILE_NO_CLIENT means that there are no
-				   free log client records in the array.
-				   If != LOGFILE_NO_CLIENT, check that
-				   log_clients > client_free_list.  On Win2k
-				   and presumably earlier, on a clean volume
-				   this is != LOGFILE_NO_CLIENT, and it should
-				   be 0, i.e. the first (and only) client
-				   record is free and thus the logfile is
-				   closed and hence clean.  A dirty volume
-				   would have left the logfile open and hence
-				   this would be LOGFILE_NO_CLIENT.  On WinXP
-				   and presumably later, the logfile is always
-				   open, even on clean shutdown so this should
-				   always be LOGFILE_NO_CLIENT. */
-/* 12*/	le16 client_in_use_list;/* The index of the first in-use log client
-				   record in the array of log client records.
-				   LOGFILE_NO_CLIENT means that there are no
-				   in-use log client records in the array.  If
-				   != LOGFILE_NO_CLIENT check that log_clients
-				   > client_in_use_list.  On Win2k and
-				   presumably earlier, on a clean volume this
-				   is LOGFILE_NO_CLIENT, i.e. there are no
-				   client records in use and thus the logfile
-				   is closed and hence clean.  A dirty volume
-				   would have left the logfile open and hence
-				   this would be != LOGFILE_NO_CLIENT, and it
-				   should be 0, i.e. the first (and only)
-				   client record is in use.  On WinXP and
-				   presumably later, the logfile is always
-				   open, even on clean shutdown so this should
-				   always be 0. */
-/* 14*/	RESTART_AREA_FLAGS flags;/* Flags modifying LFS behaviour.  On Win2k
-				   and presumably earlier this is always 0.  On
-				   WinXP and presumably later, if the logfile
-				   was shutdown cleanly, the second bit,
-				   RESTART_VOLUME_IS_CLEAN, is set.  This bit
-				   is cleared when the volume is mounted by
-				   WinXP and set when the volume is dismounted,
-				   thus if the logfile is dirty, this bit is
-				   clear.  Thus we don't need to check the
-				   Windows version to determine if the logfile
-				   is clean.  Instead if the logfile is closed,
-				   we know it must be clean.  If it is open and
-				   this bit is set, we also know it must be
-				   clean.  If on the other hand the logfile is
-				   open and this bit is clear, we can be almost
-				   certain that the logfile is dirty. */
-/* 16*/	le32 seq_number_bits;	/* How many bits to use for the sequence
-				   number.  This is calculated as 67 - the
-				   number of bits required to store the logfile
-				   size in bytes and this can be used in with
-				   the specified file_size as a consistency
-				   check. */
-/* 20*/	le16 restart_area_length;/* Length of the restart area including the
-				   client array.  Following checks required if
-				   version matches.  Otherwise, skip them.
-				   restart_area_offset + restart_area_length
-				   has to be <= system_page_size.  Also,
-				   restart_area_length has to be >=
-				   client_array_offset + (log_clients *
-				   sizeof(log client record)). */
-/* 22*/	le16 client_array_offset;/* Offset from the start of this record to
-				   the first log client record if versions are
-				   matched.  When creating, set this to be
-				   after this restart area structure, aligned
-				   to 8-bytes boundary.  If the versions do not
-				   match, this is ignored and the offset is
-				   assumed to be (sizeof(RESTART_AREA) + 7) &
-				   ~7, i.e. rounded up to first 8-byte
-				   boundary.  Either way, client_array_offset
-				   has to be aligned to an 8-byte boundary.
-				   Also, restart_area_offset +
-				   client_array_offset has to be <= 510.
-				   Finally, client_array_offset + (log_clients
-				   * sizeof(log client record)) has to be <=
-				   system_page_size.  On Win2k and presumably
-				   earlier, this is 0x30, i.e. immediately
-				   following this record.  On WinXP and
-				   presumably later, this is 0x40, i.e. there
-				   are 16 extra bytes between this record and
-				   the client array.  This probably means that
-				   the RESTART_AREA record is actually bigger
-				   in WinXP and later. */
-/* 24*/	sle64 file_size;	/* Usable byte size of the log file.  If the
-				   restart_area_offset + the offset of the
-				   file_size are > 510 then corruption has
-				   occurred.  This is the very first check when
-				   starting with the restart_area as if it
-				   fails it means that some of the above values
-				   will be corrupted by the multi sector
-				   transfer protection.  The file_size has to
-				   be rounded down to be a multiple of the
-				   log_page_size in the RESTART_PAGE_HEADER and
-				   then it has to be at least big enough to
-				   store the two restart pages and 48 (0x30)
-				   log record pages. */
-/* 32*/	le32 last_lsn_data_length;/* Length of data of last LSN, not including
-				   the log record header.  On create set to
-				   0. */
-/* 36*/	le16 log_record_header_length;/* Byte size of the log record header.
-				   If the version matches then check that the
-				   value of log_record_header_length is a
-				   multiple of 8, i.e.
-				   (log_record_header_length + 7) & ~7 ==
-				   log_record_header_length.  When creating set
-				   it to sizeof(LOG_RECORD_HEADER), aligned to
-				   8 bytes. */
-/* 38*/	le16 log_page_data_offset;/* Offset to the start of data in a log record
-				   page.  Must be a multiple of 8.  On create
-				   set it to immediately after the update
-				   sequence array of the log record page. */
-/* 40*/	le32 restart_log_open_count;/* A counter that gets incremented every
-				   time the logfile is restarted which happens
-				   at mount time when the logfile is opened.
-				   When creating set to a random value.  Win2k
-				   sets it to the low 32 bits of the current
-				   system time in NTFS format (see time.h). */
-/* 44*/	le32 reserved;		/* Reserved/alignment to 8-byte boundary. */
-/* sizeof() = 48 (0x30) bytes */
-} __attribute__ ((__packed__)) RESTART_AREA;
-
-/*
- * Log client record.  The offset of this record is found by adding the offset
- * of the RESTART_AREA to the client_array_offset value found in it.
- */
-typedef struct {
-/*Ofs*/
-/*  0*/	leLSN oldest_lsn;	/* Oldest LSN needed by this client.  On create
-				   set to 0. */
-/*  8*/	leLSN client_restart_lsn;/* LSN at which this client needs to restart
-				   the volume, i.e. the current position within
-				   the log file.  At present, if clean this
-				   should = current_lsn in restart area but it
-				   probably also = current_lsn when dirty most
-				   of the time.  At create set to 0. */
-/* 16*/	le16 prev_client;	/* The offset to the previous log client record
-				   in the array of log client records.
-				   LOGFILE_NO_CLIENT means there is no previous
-				   client record, i.e. this is the first one.
-				   This is always LOGFILE_NO_CLIENT. */
-/* 18*/	le16 next_client;	/* The offset to the next log client record in
-				   the array of log client records.
-				   LOGFILE_NO_CLIENT means there are no next
-				   client records, i.e. this is the last one.
-				   This is always LOGFILE_NO_CLIENT. */
-/* 20*/	le16 seq_number;	/* On Win2k and presumably earlier, this is set
-				   to zero every time the logfile is restarted
-				   and it is incremented when the logfile is
-				   closed at dismount time.  Thus it is 0 when
-				   dirty and 1 when clean.  On WinXP and
-				   presumably later, this is always 0. */
-/* 22*/	u8 reserved[6];		/* Reserved/alignment. */
-/* 28*/	le32 client_name_length;/* Length of client name in bytes.  Should
-				   always be 8. */
-/* 32*/	ntfschar client_name[64];/* Name of the client in Unicode.  Should
-				   always be "NTFS" with the remaining bytes
-				   set to 0. */
-/* sizeof() = 160 (0xa0) bytes */
-} __attribute__ ((__packed__)) LOG_CLIENT_RECORD;
-
-extern bool ntfs_check_logfile(struct inode *log_vi,
-		RESTART_PAGE_HEADER **rp);
-
-extern bool ntfs_is_logfile_clean(struct inode *log_vi,
-		const RESTART_PAGE_HEADER *rp);
-
-extern bool ntfs_empty_logfile(struct inode *log_vi);
-
-#endif /* NTFS_RW */
-
-#endif /* _LINUX_NTFS_LOGFILE_H */
diff --git a/fs/ntfs/malloc.h b/fs/ntfs/malloc.h
deleted file mode 100644
index 7068425735..0000000000
--- a/fs/ntfs/malloc.h
+++ /dev/null
@@ -1,77 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * malloc.h - NTFS kernel memory handling. Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2005 Anton Altaparmakov
- */
-
-#ifndef _LINUX_NTFS_MALLOC_H
-#define _LINUX_NTFS_MALLOC_H
-
-#include <linux/vmalloc.h>
-#include <linux/slab.h>
-#include <linux/highmem.h>
-
-/**
- * __ntfs_malloc - allocate memory in multiples of pages
- * @size:	number of bytes to allocate
- * @gfp_mask:	extra flags for the allocator
- *
- * Internal function.  You probably want ntfs_malloc_nofs()...
- *
- * Allocates @size bytes of memory, rounded up to multiples of PAGE_SIZE and
- * returns a pointer to the allocated memory.
- *
- * If there was insufficient memory to complete the request, return NULL.
- * Depending on @gfp_mask the allocation may be guaranteed to succeed.
- */
-static inline void *__ntfs_malloc(unsigned long size, gfp_t gfp_mask)
-{
-	if (likely(size <= PAGE_SIZE)) {
-		BUG_ON(!size);
-		/* kmalloc() has per-CPU caches so is faster for now. */
-		return kmalloc(PAGE_SIZE, gfp_mask & ~__GFP_HIGHMEM);
-		/* return (void *)__get_free_page(gfp_mask); */
-	}
-	if (likely((size >> PAGE_SHIFT) < totalram_pages()))
-		return __vmalloc(size, gfp_mask);
-	return NULL;
-}
-
-/**
- * ntfs_malloc_nofs - allocate memory in multiples of pages
- * @size:	number of bytes to allocate
- *
- * Allocates @size bytes of memory, rounded up to multiples of PAGE_SIZE and
- * returns a pointer to the allocated memory.
- *
- * If there was insufficient memory to complete the request, return NULL.
- */
-static inline void *ntfs_malloc_nofs(unsigned long size)
-{
-	return __ntfs_malloc(size, GFP_NOFS | __GFP_HIGHMEM);
-}
-
-/**
- * ntfs_malloc_nofs_nofail - allocate memory in multiples of pages
- * @size:	number of bytes to allocate
- *
- * Allocates @size bytes of memory, rounded up to multiples of PAGE_SIZE and
- * returns a pointer to the allocated memory.
- *
- * This function guarantees that the allocation will succeed.  It will sleep
- * for as long as it takes to complete the allocation.
- *
- * If there was insufficient memory to complete the request, return NULL.
- */
-static inline void *ntfs_malloc_nofs_nofail(unsigned long size)
-{
-	return __ntfs_malloc(size, GFP_NOFS | __GFP_HIGHMEM | __GFP_NOFAIL);
-}
-
-static inline void ntfs_free(void *addr)
-{
-	kvfree(addr);
-}
-
-#endif /* _LINUX_NTFS_MALLOC_H */
diff --git a/fs/ntfs/mft.c b/fs/ntfs/mft.c
deleted file mode 100644
index 6fd1dc4b08..0000000000
--- a/fs/ntfs/mft.c
+++ /dev/null
@@ -1,2907 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * mft.c - NTFS kernel mft record operations. Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2012 Anton Altaparmakov and Tuxera Inc.
- * Copyright (c) 2002 Richard Russon
- */
-
-#include <linux/buffer_head.h>
-#include <linux/slab.h>
-#include <linux/swap.h>
-#include <linux/bio.h>
-
-#include "attrib.h"
-#include "aops.h"
-#include "bitmap.h"
-#include "debug.h"
-#include "dir.h"
-#include "lcnalloc.h"
-#include "malloc.h"
-#include "mft.h"
-#include "ntfs.h"
-
-#define MAX_BHS	(PAGE_SIZE / NTFS_BLOCK_SIZE)
-
-/**
- * map_mft_record_page - map the page in which a specific mft record resides
- * @ni:		ntfs inode whose mft record page to map
- *
- * This maps the page in which the mft record of the ntfs inode @ni is situated
- * and returns a pointer to the mft record within the mapped page.
- *
- * Return value needs to be checked with IS_ERR() and if that is true PTR_ERR()
- * contains the negative error code returned.
- */
-static inline MFT_RECORD *map_mft_record_page(ntfs_inode *ni)
-{
-	loff_t i_size;
-	ntfs_volume *vol = ni->vol;
-	struct inode *mft_vi = vol->mft_ino;
-	struct page *page;
-	unsigned long index, end_index;
-	unsigned ofs;
-
-	BUG_ON(ni->page);
-	/*
-	 * The index into the page cache and the offset within the page cache
-	 * page of the wanted mft record. FIXME: We need to check for
-	 * overflowing the unsigned long, but I don't think we would ever get
-	 * here if the volume was that big...
-	 */
-	index = (u64)ni->mft_no << vol->mft_record_size_bits >>
-			PAGE_SHIFT;
-	ofs = (ni->mft_no << vol->mft_record_size_bits) & ~PAGE_MASK;
-
-	i_size = i_size_read(mft_vi);
-	/* The maximum valid index into the page cache for $MFT's data. */
-	end_index = i_size >> PAGE_SHIFT;
-
-	/* If the wanted index is out of bounds the mft record doesn't exist. */
-	if (unlikely(index >= end_index)) {
-		if (index > end_index || (i_size & ~PAGE_MASK) < ofs +
-				vol->mft_record_size) {
-			page = ERR_PTR(-ENOENT);
-			ntfs_error(vol->sb, "Attempt to read mft record 0x%lx, "
-					"which is beyond the end of the mft.  "
-					"This is probably a bug in the ntfs "
-					"driver.", ni->mft_no);
-			goto err_out;
-		}
-	}
-	/* Read, map, and pin the page. */
-	page = ntfs_map_page(mft_vi->i_mapping, index);
-	if (!IS_ERR(page)) {
-		/* Catch multi sector transfer fixup errors. */
-		if (likely(ntfs_is_mft_recordp((le32*)(page_address(page) +
-				ofs)))) {
-			ni->page = page;
-			ni->page_ofs = ofs;
-			return page_address(page) + ofs;
-		}
-		ntfs_error(vol->sb, "Mft record 0x%lx is corrupt.  "
-				"Run chkdsk.", ni->mft_no);
-		ntfs_unmap_page(page);
-		page = ERR_PTR(-EIO);
-		NVolSetErrors(vol);
-	}
-err_out:
-	ni->page = NULL;
-	ni->page_ofs = 0;
-	return (void*)page;
-}
-
-/**
- * map_mft_record - map, pin and lock an mft record
- * @ni:		ntfs inode whose MFT record to map
- *
- * First, take the mrec_lock mutex.  We might now be sleeping, while waiting
- * for the mutex if it was already locked by someone else.
- *
- * The page of the record is mapped using map_mft_record_page() before being
- * returned to the caller.
- *
- * This in turn uses ntfs_map_page() to get the page containing the wanted mft
- * record (it in turn calls read_cache_page() which reads it in from disk if
- * necessary, increments the use count on the page so that it cannot disappear
- * under us and returns a reference to the page cache page).
- *
- * If read_cache_page() invokes ntfs_readpage() to load the page from disk, it
- * sets PG_locked and clears PG_uptodate on the page. Once I/O has completed
- * and the post-read mst fixups on each mft record in the page have been
- * performed, the page gets PG_uptodate set and PG_locked cleared (this is done
- * in our asynchronous I/O completion handler end_buffer_read_mft_async()).
- * ntfs_map_page() waits for PG_locked to become clear and checks if
- * PG_uptodate is set and returns an error code if not. This provides
- * sufficient protection against races when reading/using the page.
- *
- * However there is the write mapping to think about. Doing the above described
- * checking here will be fine, because when initiating the write we will set
- * PG_locked and clear PG_uptodate making sure nobody is touching the page
- * contents. Doing the locking this way means that the commit to disk code in
- * the page cache code paths is automatically sufficiently locked with us as
- * we will not touch a page that has been locked or is not uptodate. The only
- * locking problem then is them locking the page while we are accessing it.
- *
- * So that code will end up having to own the mrec_lock of all mft
- * records/inodes present in the page before I/O can proceed. In that case we
- * wouldn't need to bother with PG_locked and PG_uptodate as nobody will be
- * accessing anything without owning the mrec_lock mutex.  But we do need to
- * use them because of the read_cache_page() invocation and the code becomes so
- * much simpler this way that it is well worth it.
- *
- * The mft record is now ours and we return a pointer to it. You need to check
- * the returned pointer with IS_ERR() and if that is true, PTR_ERR() will return
- * the error code.
- *
- * NOTE: Caller is responsible for setting the mft record dirty before calling
- * unmap_mft_record(). This is obviously only necessary if the caller really
- * modified the mft record...
- * Q: Do we want to recycle one of the VFS inode state bits instead?
- * A: No, the inode ones mean we want to change the mft record, not we want to
- * write it out.
- */
-MFT_RECORD *map_mft_record(ntfs_inode *ni)
-{
-	MFT_RECORD *m;
-
-	ntfs_debug("Entering for mft_no 0x%lx.", ni->mft_no);
-
-	/* Make sure the ntfs inode doesn't go away. */
-	atomic_inc(&ni->count);
-
-	/* Serialize access to this mft record. */
-	mutex_lock(&ni->mrec_lock);
-
-	m = map_mft_record_page(ni);
-	if (!IS_ERR(m))
-		return m;
-
-	mutex_unlock(&ni->mrec_lock);
-	atomic_dec(&ni->count);
-	ntfs_error(ni->vol->sb, "Failed with error code %lu.", -PTR_ERR(m));
-	return m;
-}
-
-/**
- * unmap_mft_record_page - unmap the page in which a specific mft record resides
- * @ni:		ntfs inode whose mft record page to unmap
- *
- * This unmaps the page in which the mft record of the ntfs inode @ni is
- * situated and returns. This is a NOOP if highmem is not configured.
- *
- * The unmap happens via ntfs_unmap_page() which in turn decrements the use
- * count on the page thus releasing it from the pinned state.
- *
- * We do not actually unmap the page from memory of course, as that will be
- * done by the page cache code itself when memory pressure increases or
- * whatever.
- */
-static inline void unmap_mft_record_page(ntfs_inode *ni)
-{
-	BUG_ON(!ni->page);
-
-	// TODO: If dirty, blah...
-	ntfs_unmap_page(ni->page);
-	ni->page = NULL;
-	ni->page_ofs = 0;
-	return;
-}
-
-/**
- * unmap_mft_record - release a mapped mft record
- * @ni:		ntfs inode whose MFT record to unmap
- *
- * We release the page mapping and the mrec_lock mutex which unmaps the mft
- * record and releases it for others to get hold of. We also release the ntfs
- * inode by decrementing the ntfs inode reference count.
- *
- * NOTE: If caller has modified the mft record, it is imperative to set the mft
- * record dirty BEFORE calling unmap_mft_record().
- */
-void unmap_mft_record(ntfs_inode *ni)
-{
-	struct page *page = ni->page;
-
-	BUG_ON(!page);
-
-	ntfs_debug("Entering for mft_no 0x%lx.", ni->mft_no);
-
-	unmap_mft_record_page(ni);
-	mutex_unlock(&ni->mrec_lock);
-	atomic_dec(&ni->count);
-	/*
-	 * If pure ntfs_inode, i.e. no vfs inode attached, we leave it to
-	 * ntfs_clear_extent_inode() in the extent inode case, and to the
-	 * caller in the non-extent, yet pure ntfs inode case, to do the actual
-	 * tear down of all structures and freeing of all allocated memory.
-	 */
-	return;
-}
-
-/**
- * map_extent_mft_record - load an extent inode and attach it to its base
- * @base_ni:	base ntfs inode
- * @mref:	mft reference of the extent inode to load
- * @ntfs_ino:	on successful return, pointer to the ntfs_inode structure
- *
- * Load the extent mft record @mref and attach it to its base inode @base_ni.
- * Return the mapped extent mft record if IS_ERR(result) is false.  Otherwise
- * PTR_ERR(result) gives the negative error code.
- *
- * On successful return, @ntfs_ino contains a pointer to the ntfs_inode
- * structure of the mapped extent inode.
- */
-MFT_RECORD *map_extent_mft_record(ntfs_inode *base_ni, MFT_REF mref,
-		ntfs_inode **ntfs_ino)
-{
-	MFT_RECORD *m;
-	ntfs_inode *ni = NULL;
-	ntfs_inode **extent_nis = NULL;
-	int i;
-	unsigned long mft_no = MREF(mref);
-	u16 seq_no = MSEQNO(mref);
-	bool destroy_ni = false;
-
-	ntfs_debug("Mapping extent mft record 0x%lx (base mft record 0x%lx).",
-			mft_no, base_ni->mft_no);
-	/* Make sure the base ntfs inode doesn't go away. */
-	atomic_inc(&base_ni->count);
-	/*
-	 * Check if this extent inode has already been added to the base inode,
-	 * in which case just return it. If not found, add it to the base
-	 * inode before returning it.
-	 */
-	mutex_lock(&base_ni->extent_lock);
-	if (base_ni->nr_extents > 0) {
-		extent_nis = base_ni->ext.extent_ntfs_inos;
-		for (i = 0; i < base_ni->nr_extents; i++) {
-			if (mft_no != extent_nis[i]->mft_no)
-				continue;
-			ni = extent_nis[i];
-			/* Make sure the ntfs inode doesn't go away. */
-			atomic_inc(&ni->count);
-			break;
-		}
-	}
-	if (likely(ni != NULL)) {
-		mutex_unlock(&base_ni->extent_lock);
-		atomic_dec(&base_ni->count);
-		/* We found the record; just have to map and return it. */
-		m = map_mft_record(ni);
-		/* map_mft_record() has incremented this on success. */
-		atomic_dec(&ni->count);
-		if (!IS_ERR(m)) {
-			/* Verify the sequence number. */
-			if (likely(le16_to_cpu(m->sequence_number) == seq_no)) {
-				ntfs_debug("Done 1.");
-				*ntfs_ino = ni;
-				return m;
-			}
-			unmap_mft_record(ni);
-			ntfs_error(base_ni->vol->sb, "Found stale extent mft "
-					"reference! Corrupt filesystem. "
-					"Run chkdsk.");
-			return ERR_PTR(-EIO);
-		}
-map_err_out:
-		ntfs_error(base_ni->vol->sb, "Failed to map extent "
-				"mft record, error code %ld.", -PTR_ERR(m));
-		return m;
-	}
-	/* Record wasn't there. Get a new ntfs inode and initialize it. */
-	ni = ntfs_new_extent_inode(base_ni->vol->sb, mft_no);
-	if (unlikely(!ni)) {
-		mutex_unlock(&base_ni->extent_lock);
-		atomic_dec(&base_ni->count);
-		return ERR_PTR(-ENOMEM);
-	}
-	ni->vol = base_ni->vol;
-	ni->seq_no = seq_no;
-	ni->nr_extents = -1;
-	ni->ext.base_ntfs_ino = base_ni;
-	/* Now map the record. */
-	m = map_mft_record(ni);
-	if (IS_ERR(m)) {
-		mutex_unlock(&base_ni->extent_lock);
-		atomic_dec(&base_ni->count);
-		ntfs_clear_extent_inode(ni);
-		goto map_err_out;
-	}
-	/* Verify the sequence number if it is present. */
-	if (seq_no && (le16_to_cpu(m->sequence_number) != seq_no)) {
-		ntfs_error(base_ni->vol->sb, "Found stale extent mft "
-				"reference! Corrupt filesystem. Run chkdsk.");
-		destroy_ni = true;
-		m = ERR_PTR(-EIO);
-		goto unm_err_out;
-	}
-	/* Attach extent inode to base inode, reallocating memory if needed. */
-	if (!(base_ni->nr_extents & 3)) {
-		ntfs_inode **tmp;
-		int new_size = (base_ni->nr_extents + 4) * sizeof(ntfs_inode *);
-
-		tmp = kmalloc(new_size, GFP_NOFS);
-		if (unlikely(!tmp)) {
-			ntfs_error(base_ni->vol->sb, "Failed to allocate "
-					"internal buffer.");
-			destroy_ni = true;
-			m = ERR_PTR(-ENOMEM);
-			goto unm_err_out;
-		}
-		if (base_ni->nr_extents) {
-			BUG_ON(!base_ni->ext.extent_ntfs_inos);
-			memcpy(tmp, base_ni->ext.extent_ntfs_inos, new_size -
-					4 * sizeof(ntfs_inode *));
-			kfree(base_ni->ext.extent_ntfs_inos);
-		}
-		base_ni->ext.extent_ntfs_inos = tmp;
-	}
-	base_ni->ext.extent_ntfs_inos[base_ni->nr_extents++] = ni;
-	mutex_unlock(&base_ni->extent_lock);
-	atomic_dec(&base_ni->count);
-	ntfs_debug("Done 2.");
-	*ntfs_ino = ni;
-	return m;
-unm_err_out:
-	unmap_mft_record(ni);
-	mutex_unlock(&base_ni->extent_lock);
-	atomic_dec(&base_ni->count);
-	/*
-	 * If the extent inode was not attached to the base inode we need to
-	 * release it or we will leak memory.
-	 */
-	if (destroy_ni)
-		ntfs_clear_extent_inode(ni);
-	return m;
-}
-
-#ifdef NTFS_RW
-
-/**
- * __mark_mft_record_dirty - set the mft record and the page containing it dirty
- * @ni:		ntfs inode describing the mapped mft record
- *
- * Internal function.  Users should call mark_mft_record_dirty() instead.
- *
- * Set the mapped (extent) mft record of the (base or extent) ntfs inode @ni,
- * as well as the page containing the mft record, dirty.  Also, mark the base
- * vfs inode dirty.  This ensures that any changes to the mft record are
- * written out to disk.
- *
- * NOTE:  We only set I_DIRTY_DATASYNC (and not I_DIRTY_PAGES)
- * on the base vfs inode, because even though file data may have been modified,
- * it is dirty in the inode meta data rather than the data page cache of the
- * inode, and thus there are no data pages that need writing out.  Therefore, a
- * full mark_inode_dirty() is overkill.  A mark_inode_dirty_sync(), on the
- * other hand, is not sufficient, because ->write_inode needs to be called even
- * in case of fdatasync. This needs to happen or the file data would not
- * necessarily hit the device synchronously, even though the vfs inode has the
- * O_SYNC flag set.  Also, I_DIRTY_DATASYNC simply "feels" better than just
- * I_DIRTY_SYNC, since the file data has not actually hit the block device yet,
- * which is not what I_DIRTY_SYNC on its own would suggest.
- */
-void __mark_mft_record_dirty(ntfs_inode *ni)
-{
-	ntfs_inode *base_ni;
-
-	ntfs_debug("Entering for inode 0x%lx.", ni->mft_no);
-	BUG_ON(NInoAttr(ni));
-	mark_ntfs_record_dirty(ni->page, ni->page_ofs);
-	/* Determine the base vfs inode and mark it dirty, too. */
-	mutex_lock(&ni->extent_lock);
-	if (likely(ni->nr_extents >= 0))
-		base_ni = ni;
-	else
-		base_ni = ni->ext.base_ntfs_ino;
-	mutex_unlock(&ni->extent_lock);
-	__mark_inode_dirty(VFS_I(base_ni), I_DIRTY_DATASYNC);
-}
-
-static const char *ntfs_please_email = "Please email "
-		"linux-ntfs-dev@lists.sourceforge.net and say that you saw "
-		"this message.  Thank you.";
-
-/**
- * ntfs_sync_mft_mirror_umount - synchronise an mft record to the mft mirror
- * @vol:	ntfs volume on which the mft record to synchronize resides
- * @mft_no:	mft record number of mft record to synchronize
- * @m:		mapped, mst protected (extent) mft record to synchronize
- *
- * Write the mapped, mst protected (extent) mft record @m with mft record
- * number @mft_no to the mft mirror ($MFTMirr) of the ntfs volume @vol,
- * bypassing the page cache and the $MFTMirr inode itself.
- *
- * This function is only for use at umount time when the mft mirror inode has
- * already been disposed off.  We BUG() if we are called while the mft mirror
- * inode is still attached to the volume.
- *
- * On success return 0.  On error return -errno.
- *
- * NOTE:  This function is not implemented yet as I am not convinced it can
- * actually be triggered considering the sequence of commits we do in super.c::
- * ntfs_put_super().  But just in case we provide this place holder as the
- * alternative would be either to BUG() or to get a NULL pointer dereference
- * and Oops.
- */
-static int ntfs_sync_mft_mirror_umount(ntfs_volume *vol,
-		const unsigned long mft_no, MFT_RECORD *m)
-{
-	BUG_ON(vol->mftmirr_ino);
-	ntfs_error(vol->sb, "Umount time mft mirror syncing is not "
-			"implemented yet.  %s", ntfs_please_email);
-	return -EOPNOTSUPP;
-}
-
-/**
- * ntfs_sync_mft_mirror - synchronize an mft record to the mft mirror
- * @vol:	ntfs volume on which the mft record to synchronize resides
- * @mft_no:	mft record number of mft record to synchronize
- * @m:		mapped, mst protected (extent) mft record to synchronize
- * @sync:	if true, wait for i/o completion
- *
- * Write the mapped, mst protected (extent) mft record @m with mft record
- * number @mft_no to the mft mirror ($MFTMirr) of the ntfs volume @vol.
- *
- * On success return 0.  On error return -errno and set the volume errors flag
- * in the ntfs volume @vol.
- *
- * NOTE:  We always perform synchronous i/o and ignore the @sync parameter.
- *
- * TODO:  If @sync is false, want to do truly asynchronous i/o, i.e. just
- * schedule i/o via ->writepage or do it via kntfsd or whatever.
- */
-int ntfs_sync_mft_mirror(ntfs_volume *vol, const unsigned long mft_no,
-		MFT_RECORD *m, int sync)
-{
-	struct page *page;
-	unsigned int blocksize = vol->sb->s_blocksize;
-	int max_bhs = vol->mft_record_size / blocksize;
-	struct buffer_head *bhs[MAX_BHS];
-	struct buffer_head *bh, *head;
-	u8 *kmirr;
-	runlist_element *rl;
-	unsigned int block_start, block_end, m_start, m_end, page_ofs;
-	int i_bhs, nr_bhs, err = 0;
-	unsigned char blocksize_bits = vol->sb->s_blocksize_bits;
-
-	ntfs_debug("Entering for inode 0x%lx.", mft_no);
-	BUG_ON(!max_bhs);
-	if (WARN_ON(max_bhs > MAX_BHS))
-		return -EINVAL;
-	if (unlikely(!vol->mftmirr_ino)) {
-		/* This could happen during umount... */
-		err = ntfs_sync_mft_mirror_umount(vol, mft_no, m);
-		if (likely(!err))
-			return err;
-		goto err_out;
-	}
-	/* Get the page containing the mirror copy of the mft record @m. */
-	page = ntfs_map_page(vol->mftmirr_ino->i_mapping, mft_no >>
-			(PAGE_SHIFT - vol->mft_record_size_bits));
-	if (IS_ERR(page)) {
-		ntfs_error(vol->sb, "Failed to map mft mirror page.");
-		err = PTR_ERR(page);
-		goto err_out;
-	}
-	lock_page(page);
-	BUG_ON(!PageUptodate(page));
-	ClearPageUptodate(page);
-	/* Offset of the mft mirror record inside the page. */
-	page_ofs = (mft_no << vol->mft_record_size_bits) & ~PAGE_MASK;
-	/* The address in the page of the mirror copy of the mft record @m. */
-	kmirr = page_address(page) + page_ofs;
-	/* Copy the mst protected mft record to the mirror. */
-	memcpy(kmirr, m, vol->mft_record_size);
-	/* Create uptodate buffers if not present. */
-	if (unlikely(!page_has_buffers(page))) {
-		struct buffer_head *tail;
-
-		bh = head = alloc_page_buffers(page, blocksize, true);
-		do {
-			set_buffer_uptodate(bh);
-			tail = bh;
-			bh = bh->b_this_page;
-		} while (bh);
-		tail->b_this_page = head;
-		attach_page_private(page, head);
-	}
-	bh = head = page_buffers(page);
-	BUG_ON(!bh);
-	rl = NULL;
-	nr_bhs = 0;
-	block_start = 0;
-	m_start = kmirr - (u8*)page_address(page);
-	m_end = m_start + vol->mft_record_size;
-	do {
-		block_end = block_start + blocksize;
-		/* If the buffer is outside the mft record, skip it. */
-		if (block_end <= m_start)
-			continue;
-		if (unlikely(block_start >= m_end))
-			break;
-		/* Need to map the buffer if it is not mapped already. */
-		if (unlikely(!buffer_mapped(bh))) {
-			VCN vcn;
-			LCN lcn;
-			unsigned int vcn_ofs;
-
-			bh->b_bdev = vol->sb->s_bdev;
-			/* Obtain the vcn and offset of the current block. */
-			vcn = ((VCN)mft_no << vol->mft_record_size_bits) +
-					(block_start - m_start);
-			vcn_ofs = vcn & vol->cluster_size_mask;
-			vcn >>= vol->cluster_size_bits;
-			if (!rl) {
-				down_read(&NTFS_I(vol->mftmirr_ino)->
-						runlist.lock);
-				rl = NTFS_I(vol->mftmirr_ino)->runlist.rl;
-				/*
-				 * $MFTMirr always has the whole of its runlist
-				 * in memory.
-				 */
-				BUG_ON(!rl);
-			}
-			/* Seek to element containing target vcn. */
-			while (rl->length && rl[1].vcn <= vcn)
-				rl++;
-			lcn = ntfs_rl_vcn_to_lcn(rl, vcn);
-			/* For $MFTMirr, only lcn >= 0 is a successful remap. */
-			if (likely(lcn >= 0)) {
-				/* Setup buffer head to correct block. */
-				bh->b_blocknr = ((lcn <<
-						vol->cluster_size_bits) +
-						vcn_ofs) >> blocksize_bits;
-				set_buffer_mapped(bh);
-			} else {
-				bh->b_blocknr = -1;
-				ntfs_error(vol->sb, "Cannot write mft mirror "
-						"record 0x%lx because its "
-						"location on disk could not "
-						"be determined (error code "
-						"%lli).", mft_no,
-						(long long)lcn);
-				err = -EIO;
-			}
-		}
-		BUG_ON(!buffer_uptodate(bh));
-		BUG_ON(!nr_bhs && (m_start != block_start));
-		BUG_ON(nr_bhs >= max_bhs);
-		bhs[nr_bhs++] = bh;
-		BUG_ON((nr_bhs >= max_bhs) && (m_end != block_end));
-	} while (block_start = block_end, (bh = bh->b_this_page) != head);
-	if (unlikely(rl))
-		up_read(&NTFS_I(vol->mftmirr_ino)->runlist.lock);
-	if (likely(!err)) {
-		/* Lock buffers and start synchronous write i/o on them. */
-		for (i_bhs = 0; i_bhs < nr_bhs; i_bhs++) {
-			struct buffer_head *tbh = bhs[i_bhs];
-
-			if (!trylock_buffer(tbh))
-				BUG();
-			BUG_ON(!buffer_uptodate(tbh));
-			clear_buffer_dirty(tbh);
-			get_bh(tbh);
-			tbh->b_end_io = end_buffer_write_sync;
-			submit_bh(REQ_OP_WRITE, tbh);
-		}
-		/* Wait on i/o completion of buffers. */
-		for (i_bhs = 0; i_bhs < nr_bhs; i_bhs++) {
-			struct buffer_head *tbh = bhs[i_bhs];
-
-			wait_on_buffer(tbh);
-			if (unlikely(!buffer_uptodate(tbh))) {
-				err = -EIO;
-				/*
-				 * Set the buffer uptodate so the page and
-				 * buffer states do not become out of sync.
-				 */
-				set_buffer_uptodate(tbh);
-			}
-		}
-	} else /* if (unlikely(err)) */ {
-		/* Clean the buffers. */
-		for (i_bhs = 0; i_bhs < nr_bhs; i_bhs++)
-			clear_buffer_dirty(bhs[i_bhs]);
-	}
-	/* Current state: all buffers are clean, unlocked, and uptodate. */
-	/* Remove the mst protection fixups again. */
-	post_write_mst_fixup((NTFS_RECORD*)kmirr);
-	flush_dcache_page(page);
-	SetPageUptodate(page);
-	unlock_page(page);
-	ntfs_unmap_page(page);
-	if (likely(!err)) {
-		ntfs_debug("Done.");
-	} else {
-		ntfs_error(vol->sb, "I/O error while writing mft mirror "
-				"record 0x%lx!", mft_no);
-err_out:
-		ntfs_error(vol->sb, "Failed to synchronize $MFTMirr (error "
-				"code %i).  Volume will be left marked dirty "
-				"on umount.  Run ntfsfix on the partition "
-				"after umounting to correct this.", -err);
-		NVolSetErrors(vol);
-	}
-	return err;
-}
-
-/**
- * write_mft_record_nolock - write out a mapped (extent) mft record
- * @ni:		ntfs inode describing the mapped (extent) mft record
- * @m:		mapped (extent) mft record to write
- * @sync:	if true, wait for i/o completion
- *
- * Write the mapped (extent) mft record @m described by the (regular or extent)
- * ntfs inode @ni to backing store.  If the mft record @m has a counterpart in
- * the mft mirror, that is also updated.
- *
- * We only write the mft record if the ntfs inode @ni is dirty and the first
- * buffer belonging to its mft record is dirty, too.  We ignore the dirty state
- * of subsequent buffers because we could have raced with
- * fs/ntfs/aops.c::mark_ntfs_record_dirty().
- *
- * On success, clean the mft record and return 0.  On error, leave the mft
- * record dirty and return -errno.
- *
- * NOTE:  We always perform synchronous i/o and ignore the @sync parameter.
- * However, if the mft record has a counterpart in the mft mirror and @sync is
- * true, we write the mft record, wait for i/o completion, and only then write
- * the mft mirror copy.  This ensures that if the system crashes either the mft
- * or the mft mirror will contain a self-consistent mft record @m.  If @sync is
- * false on the other hand, we start i/o on both and then wait for completion
- * on them.  This provides a speedup but no longer guarantees that you will end
- * up with a self-consistent mft record in the case of a crash but if you asked
- * for asynchronous writing you probably do not care about that anyway.
- *
- * TODO:  If @sync is false, want to do truly asynchronous i/o, i.e. just
- * schedule i/o via ->writepage or do it via kntfsd or whatever.
- */
-int write_mft_record_nolock(ntfs_inode *ni, MFT_RECORD *m, int sync)
-{
-	ntfs_volume *vol = ni->vol;
-	struct page *page = ni->page;
-	unsigned int blocksize = vol->sb->s_blocksize;
-	unsigned char blocksize_bits = vol->sb->s_blocksize_bits;
-	int max_bhs = vol->mft_record_size / blocksize;
-	struct buffer_head *bhs[MAX_BHS];
-	struct buffer_head *bh, *head;
-	runlist_element *rl;
-	unsigned int block_start, block_end, m_start, m_end;
-	int i_bhs, nr_bhs, err = 0;
-
-	ntfs_debug("Entering for inode 0x%lx.", ni->mft_no);
-	BUG_ON(NInoAttr(ni));
-	BUG_ON(!max_bhs);
-	BUG_ON(!PageLocked(page));
-	if (WARN_ON(max_bhs > MAX_BHS)) {
-		err = -EINVAL;
-		goto err_out;
-	}
-	/*
-	 * If the ntfs_inode is clean no need to do anything.  If it is dirty,
-	 * mark it as clean now so that it can be redirtied later on if needed.
-	 * There is no danger of races since the caller is holding the locks
-	 * for the mft record @m and the page it is in.
-	 */
-	if (!NInoTestClearDirty(ni))
-		goto done;
-	bh = head = page_buffers(page);
-	BUG_ON(!bh);
-	rl = NULL;
-	nr_bhs = 0;
-	block_start = 0;
-	m_start = ni->page_ofs;
-	m_end = m_start + vol->mft_record_size;
-	do {
-		block_end = block_start + blocksize;
-		/* If the buffer is outside the mft record, skip it. */
-		if (block_end <= m_start)
-			continue;
-		if (unlikely(block_start >= m_end))
-			break;
-		/*
-		 * If this block is not the first one in the record, we ignore
-		 * the buffer's dirty state because we could have raced with a
-		 * parallel mark_ntfs_record_dirty().
-		 */
-		if (block_start == m_start) {
-			/* This block is the first one in the record. */
-			if (!buffer_dirty(bh)) {
-				BUG_ON(nr_bhs);
-				/* Clean records are not written out. */
-				break;
-			}
-		}
-		/* Need to map the buffer if it is not mapped already. */
-		if (unlikely(!buffer_mapped(bh))) {
-			VCN vcn;
-			LCN lcn;
-			unsigned int vcn_ofs;
-
-			bh->b_bdev = vol->sb->s_bdev;
-			/* Obtain the vcn and offset of the current block. */
-			vcn = ((VCN)ni->mft_no << vol->mft_record_size_bits) +
-					(block_start - m_start);
-			vcn_ofs = vcn & vol->cluster_size_mask;
-			vcn >>= vol->cluster_size_bits;
-			if (!rl) {
-				down_read(&NTFS_I(vol->mft_ino)->runlist.lock);
-				rl = NTFS_I(vol->mft_ino)->runlist.rl;
-				BUG_ON(!rl);
-			}
-			/* Seek to element containing target vcn. */
-			while (rl->length && rl[1].vcn <= vcn)
-				rl++;
-			lcn = ntfs_rl_vcn_to_lcn(rl, vcn);
-			/* For $MFT, only lcn >= 0 is a successful remap. */
-			if (likely(lcn >= 0)) {
-				/* Setup buffer head to correct block. */
-				bh->b_blocknr = ((lcn <<
-						vol->cluster_size_bits) +
-						vcn_ofs) >> blocksize_bits;
-				set_buffer_mapped(bh);
-			} else {
-				bh->b_blocknr = -1;
-				ntfs_error(vol->sb, "Cannot write mft record "
-						"0x%lx because its location "
-						"on disk could not be "
-						"determined (error code %lli).",
-						ni->mft_no, (long long)lcn);
-				err = -EIO;
-			}
-		}
-		BUG_ON(!buffer_uptodate(bh));
-		BUG_ON(!nr_bhs && (m_start != block_start));
-		BUG_ON(nr_bhs >= max_bhs);
-		bhs[nr_bhs++] = bh;
-		BUG_ON((nr_bhs >= max_bhs) && (m_end != block_end));
-	} while (block_start = block_end, (bh = bh->b_this_page) != head);
-	if (unlikely(rl))
-		up_read(&NTFS_I(vol->mft_ino)->runlist.lock);
-	if (!nr_bhs)
-		goto done;
-	if (unlikely(err))
-		goto cleanup_out;
-	/* Apply the mst protection fixups. */
-	err = pre_write_mst_fixup((NTFS_RECORD*)m, vol->mft_record_size);
-	if (err) {
-		ntfs_error(vol->sb, "Failed to apply mst fixups!");
-		goto cleanup_out;
-	}
-	flush_dcache_mft_record_page(ni);
-	/* Lock buffers and start synchronous write i/o on them. */
-	for (i_bhs = 0; i_bhs < nr_bhs; i_bhs++) {
-		struct buffer_head *tbh = bhs[i_bhs];
-
-		if (!trylock_buffer(tbh))
-			BUG();
-		BUG_ON(!buffer_uptodate(tbh));
-		clear_buffer_dirty(tbh);
-		get_bh(tbh);
-		tbh->b_end_io = end_buffer_write_sync;
-		submit_bh(REQ_OP_WRITE, tbh);
-	}
-	/* Synchronize the mft mirror now if not @sync. */
-	if (!sync && ni->mft_no < vol->mftmirr_size)
-		ntfs_sync_mft_mirror(vol, ni->mft_no, m, sync);
-	/* Wait on i/o completion of buffers. */
-	for (i_bhs = 0; i_bhs < nr_bhs; i_bhs++) {
-		struct buffer_head *tbh = bhs[i_bhs];
-
-		wait_on_buffer(tbh);
-		if (unlikely(!buffer_uptodate(tbh))) {
-			err = -EIO;
-			/*
-			 * Set the buffer uptodate so the page and buffer
-			 * states do not become out of sync.
-			 */
-			if (PageUptodate(page))
-				set_buffer_uptodate(tbh);
-		}
-	}
-	/* If @sync, now synchronize the mft mirror. */
-	if (sync && ni->mft_no < vol->mftmirr_size)
-		ntfs_sync_mft_mirror(vol, ni->mft_no, m, sync);
-	/* Remove the mst protection fixups again. */
-	post_write_mst_fixup((NTFS_RECORD*)m);
-	flush_dcache_mft_record_page(ni);
-	if (unlikely(err)) {
-		/* I/O error during writing.  This is really bad! */
-		ntfs_error(vol->sb, "I/O error while writing mft record "
-				"0x%lx!  Marking base inode as bad.  You "
-				"should unmount the volume and run chkdsk.",
-				ni->mft_no);
-		goto err_out;
-	}
-done:
-	ntfs_debug("Done.");
-	return 0;
-cleanup_out:
-	/* Clean the buffers. */
-	for (i_bhs = 0; i_bhs < nr_bhs; i_bhs++)
-		clear_buffer_dirty(bhs[i_bhs]);
-err_out:
-	/*
-	 * Current state: all buffers are clean, unlocked, and uptodate.
-	 * The caller should mark the base inode as bad so that no more i/o
-	 * happens.  ->clear_inode() will still be invoked so all extent inodes
-	 * and other allocated memory will be freed.
-	 */
-	if (err == -ENOMEM) {
-		ntfs_error(vol->sb, "Not enough memory to write mft record.  "
-				"Redirtying so the write is retried later.");
-		mark_mft_record_dirty(ni);
-		err = 0;
-	} else
-		NVolSetErrors(vol);
-	return err;
-}
-
-/**
- * ntfs_may_write_mft_record - check if an mft record may be written out
- * @vol:	[IN]  ntfs volume on which the mft record to check resides
- * @mft_no:	[IN]  mft record number of the mft record to check
- * @m:		[IN]  mapped mft record to check
- * @locked_ni:	[OUT] caller has to unlock this ntfs inode if one is returned
- *
- * Check if the mapped (base or extent) mft record @m with mft record number
- * @mft_no belonging to the ntfs volume @vol may be written out.  If necessary
- * and possible the ntfs inode of the mft record is locked and the base vfs
- * inode is pinned.  The locked ntfs inode is then returned in @locked_ni.  The
- * caller is responsible for unlocking the ntfs inode and unpinning the base
- * vfs inode.
- *
- * Return 'true' if the mft record may be written out and 'false' if not.
- *
- * The caller has locked the page and cleared the uptodate flag on it which
- * means that we can safely write out any dirty mft records that do not have
- * their inodes in icache as determined by ilookup5() as anyone
- * opening/creating such an inode would block when attempting to map the mft
- * record in read_cache_page() until we are finished with the write out.
- *
- * Here is a description of the tests we perform:
- *
- * If the inode is found in icache we know the mft record must be a base mft
- * record.  If it is dirty, we do not write it and return 'false' as the vfs
- * inode write paths will result in the access times being updated which would
- * cause the base mft record to be redirtied and written out again.  (We know
- * the access time update will modify the base mft record because Windows
- * chkdsk complains if the standard information attribute is not in the base
- * mft record.)
- *
- * If the inode is in icache and not dirty, we attempt to lock the mft record
- * and if we find the lock was already taken, it is not safe to write the mft
- * record and we return 'false'.
- *
- * If we manage to obtain the lock we have exclusive access to the mft record,
- * which also allows us safe writeout of the mft record.  We then set
- * @locked_ni to the locked ntfs inode and return 'true'.
- *
- * Note we cannot just lock the mft record and sleep while waiting for the lock
- * because this would deadlock due to lock reversal (normally the mft record is
- * locked before the page is locked but we already have the page locked here
- * when we try to lock the mft record).
- *
- * If the inode is not in icache we need to perform further checks.
- *
- * If the mft record is not a FILE record or it is a base mft record, we can
- * safely write it and return 'true'.
- *
- * We now know the mft record is an extent mft record.  We check if the inode
- * corresponding to its base mft record is in icache and obtain a reference to
- * it if it is.  If it is not, we can safely write it and return 'true'.
- *
- * We now have the base inode for the extent mft record.  We check if it has an
- * ntfs inode for the extent mft record attached and if not it is safe to write
- * the extent mft record and we return 'true'.
- *
- * The ntfs inode for the extent mft record is attached to the base inode so we
- * attempt to lock the extent mft record and if we find the lock was already
- * taken, it is not safe to write the extent mft record and we return 'false'.
- *
- * If we manage to obtain the lock we have exclusive access to the extent mft
- * record, which also allows us safe writeout of the extent mft record.  We
- * set the ntfs inode of the extent mft record clean and then set @locked_ni to
- * the now locked ntfs inode and return 'true'.
- *
- * Note, the reason for actually writing dirty mft records here and not just
- * relying on the vfs inode dirty code paths is that we can have mft records
- * modified without them ever having actual inodes in memory.  Also we can have
- * dirty mft records with clean ntfs inodes in memory.  None of the described
- * cases would result in the dirty mft records being written out if we only
- * relied on the vfs inode dirty code paths.  And these cases can really occur
- * during allocation of new mft records and in particular when the
- * initialized_size of the $MFT/$DATA attribute is extended and the new space
- * is initialized using ntfs_mft_record_format().  The clean inode can then
- * appear if the mft record is reused for a new inode before it got written
- * out.
- */
-bool ntfs_may_write_mft_record(ntfs_volume *vol, const unsigned long mft_no,
-		const MFT_RECORD *m, ntfs_inode **locked_ni)
-{
-	struct super_block *sb = vol->sb;
-	struct inode *mft_vi = vol->mft_ino;
-	struct inode *vi;
-	ntfs_inode *ni, *eni, **extent_nis;
-	int i;
-	ntfs_attr na;
-
-	ntfs_debug("Entering for inode 0x%lx.", mft_no);
-	/*
-	 * Normally we do not return a locked inode so set @locked_ni to NULL.
-	 */
-	BUG_ON(!locked_ni);
-	*locked_ni = NULL;
-	/*
-	 * Check if the inode corresponding to this mft record is in the VFS
-	 * inode cache and obtain a reference to it if it is.
-	 */
-	ntfs_debug("Looking for inode 0x%lx in icache.", mft_no);
-	na.mft_no = mft_no;
-	na.name = NULL;
-	na.name_len = 0;
-	na.type = AT_UNUSED;
-	/*
-	 * Optimize inode 0, i.e. $MFT itself, since we have it in memory and
-	 * we get here for it rather often.
-	 */
-	if (!mft_no) {
-		/* Balance the below iput(). */
-		vi = igrab(mft_vi);
-		BUG_ON(vi != mft_vi);
-	} else {
-		/*
-		 * Have to use ilookup5_nowait() since ilookup5() waits for the
-		 * inode lock which causes ntfs to deadlock when a concurrent
-		 * inode write via the inode dirty code paths and the page
-		 * dirty code path of the inode dirty code path when writing
-		 * $MFT occurs.
-		 */
-		vi = ilookup5_nowait(sb, mft_no, ntfs_test_inode, &na);
-	}
-	if (vi) {
-		ntfs_debug("Base inode 0x%lx is in icache.", mft_no);
-		/* The inode is in icache. */
-		ni = NTFS_I(vi);
-		/* Take a reference to the ntfs inode. */
-		atomic_inc(&ni->count);
-		/* If the inode is dirty, do not write this record. */
-		if (NInoDirty(ni)) {
-			ntfs_debug("Inode 0x%lx is dirty, do not write it.",
-					mft_no);
-			atomic_dec(&ni->count);
-			iput(vi);
-			return false;
-		}
-		ntfs_debug("Inode 0x%lx is not dirty.", mft_no);
-		/* The inode is not dirty, try to take the mft record lock. */
-		if (unlikely(!mutex_trylock(&ni->mrec_lock))) {
-			ntfs_debug("Mft record 0x%lx is already locked, do "
-					"not write it.", mft_no);
-			atomic_dec(&ni->count);
-			iput(vi);
-			return false;
-		}
-		ntfs_debug("Managed to lock mft record 0x%lx, write it.",
-				mft_no);
-		/*
-		 * The write has to occur while we hold the mft record lock so
-		 * return the locked ntfs inode.
-		 */
-		*locked_ni = ni;
-		return true;
-	}
-	ntfs_debug("Inode 0x%lx is not in icache.", mft_no);
-	/* The inode is not in icache. */
-	/* Write the record if it is not a mft record (type "FILE"). */
-	if (!ntfs_is_mft_record(m->magic)) {
-		ntfs_debug("Mft record 0x%lx is not a FILE record, write it.",
-				mft_no);
-		return true;
-	}
-	/* Write the mft record if it is a base inode. */
-	if (!m->base_mft_record) {
-		ntfs_debug("Mft record 0x%lx is a base record, write it.",
-				mft_no);
-		return true;
-	}
-	/*
-	 * This is an extent mft record.  Check if the inode corresponding to
-	 * its base mft record is in icache and obtain a reference to it if it
-	 * is.
-	 */
-	na.mft_no = MREF_LE(m->base_mft_record);
-	ntfs_debug("Mft record 0x%lx is an extent record.  Looking for base "
-			"inode 0x%lx in icache.", mft_no, na.mft_no);
-	if (!na.mft_no) {
-		/* Balance the below iput(). */
-		vi = igrab(mft_vi);
-		BUG_ON(vi != mft_vi);
-	} else
-		vi = ilookup5_nowait(sb, na.mft_no, ntfs_test_inode,
-				&na);
-	if (!vi) {
-		/*
-		 * The base inode is not in icache, write this extent mft
-		 * record.
-		 */
-		ntfs_debug("Base inode 0x%lx is not in icache, write the "
-				"extent record.", na.mft_no);
-		return true;
-	}
-	ntfs_debug("Base inode 0x%lx is in icache.", na.mft_no);
-	/*
-	 * The base inode is in icache.  Check if it has the extent inode
-	 * corresponding to this extent mft record attached.
-	 */
-	ni = NTFS_I(vi);
-	mutex_lock(&ni->extent_lock);
-	if (ni->nr_extents <= 0) {
-		/*
-		 * The base inode has no attached extent inodes, write this
-		 * extent mft record.
-		 */
-		mutex_unlock(&ni->extent_lock);
-		iput(vi);
-		ntfs_debug("Base inode 0x%lx has no attached extent inodes, "
-				"write the extent record.", na.mft_no);
-		return true;
-	}
-	/* Iterate over the attached extent inodes. */
-	extent_nis = ni->ext.extent_ntfs_inos;
-	for (eni = NULL, i = 0; i < ni->nr_extents; ++i) {
-		if (mft_no == extent_nis[i]->mft_no) {
-			/*
-			 * Found the extent inode corresponding to this extent
-			 * mft record.
-			 */
-			eni = extent_nis[i];
-			break;
-		}
-	}
-	/*
-	 * If the extent inode was not attached to the base inode, write this
-	 * extent mft record.
-	 */
-	if (!eni) {
-		mutex_unlock(&ni->extent_lock);
-		iput(vi);
-		ntfs_debug("Extent inode 0x%lx is not attached to its base "
-				"inode 0x%lx, write the extent record.",
-				mft_no, na.mft_no);
-		return true;
-	}
-	ntfs_debug("Extent inode 0x%lx is attached to its base inode 0x%lx.",
-			mft_no, na.mft_no);
-	/* Take a reference to the extent ntfs inode. */
-	atomic_inc(&eni->count);
-	mutex_unlock(&ni->extent_lock);
-	/*
-	 * Found the extent inode coresponding to this extent mft record.
-	 * Try to take the mft record lock.
-	 */
-	if (unlikely(!mutex_trylock(&eni->mrec_lock))) {
-		atomic_dec(&eni->count);
-		iput(vi);
-		ntfs_debug("Extent mft record 0x%lx is already locked, do "
-				"not write it.", mft_no);
-		return false;
-	}
-	ntfs_debug("Managed to lock extent mft record 0x%lx, write it.",
-			mft_no);
-	if (NInoTestClearDirty(eni))
-		ntfs_debug("Extent inode 0x%lx is dirty, marking it clean.",
-				mft_no);
-	/*
-	 * The write has to occur while we hold the mft record lock so return
-	 * the locked extent ntfs inode.
-	 */
-	*locked_ni = eni;
-	return true;
-}
-
-static const char *es = "  Leaving inconsistent metadata.  Unmount and run "
-		"chkdsk.";
-
-/**
- * ntfs_mft_bitmap_find_and_alloc_free_rec_nolock - see name
- * @vol:	volume on which to search for a free mft record
- * @base_ni:	open base inode if allocating an extent mft record or NULL
- *
- * Search for a free mft record in the mft bitmap attribute on the ntfs volume
- * @vol.
- *
- * If @base_ni is NULL start the search at the default allocator position.
- *
- * If @base_ni is not NULL start the search at the mft record after the base
- * mft record @base_ni.
- *
- * Return the free mft record on success and -errno on error.  An error code of
- * -ENOSPC means that there are no free mft records in the currently
- * initialized mft bitmap.
- *
- * Locking: Caller must hold vol->mftbmp_lock for writing.
- */
-static int ntfs_mft_bitmap_find_and_alloc_free_rec_nolock(ntfs_volume *vol,
-		ntfs_inode *base_ni)
-{
-	s64 pass_end, ll, data_pos, pass_start, ofs, bit;
-	unsigned long flags;
-	struct address_space *mftbmp_mapping;
-	u8 *buf, *byte;
-	struct page *page;
-	unsigned int page_ofs, size;
-	u8 pass, b;
-
-	ntfs_debug("Searching for free mft record in the currently "
-			"initialized mft bitmap.");
-	mftbmp_mapping = vol->mftbmp_ino->i_mapping;
-	/*
-	 * Set the end of the pass making sure we do not overflow the mft
-	 * bitmap.
-	 */
-	read_lock_irqsave(&NTFS_I(vol->mft_ino)->size_lock, flags);
-	pass_end = NTFS_I(vol->mft_ino)->allocated_size >>
-			vol->mft_record_size_bits;
-	read_unlock_irqrestore(&NTFS_I(vol->mft_ino)->size_lock, flags);
-	read_lock_irqsave(&NTFS_I(vol->mftbmp_ino)->size_lock, flags);
-	ll = NTFS_I(vol->mftbmp_ino)->initialized_size << 3;
-	read_unlock_irqrestore(&NTFS_I(vol->mftbmp_ino)->size_lock, flags);
-	if (pass_end > ll)
-		pass_end = ll;
-	pass = 1;
-	if (!base_ni)
-		data_pos = vol->mft_data_pos;
-	else
-		data_pos = base_ni->mft_no + 1;
-	if (data_pos < 24)
-		data_pos = 24;
-	if (data_pos >= pass_end) {
-		data_pos = 24;
-		pass = 2;
-		/* This happens on a freshly formatted volume. */
-		if (data_pos >= pass_end)
-			return -ENOSPC;
-	}
-	pass_start = data_pos;
-	ntfs_debug("Starting bitmap search: pass %u, pass_start 0x%llx, "
-			"pass_end 0x%llx, data_pos 0x%llx.", pass,
-			(long long)pass_start, (long long)pass_end,
-			(long long)data_pos);
-	/* Loop until a free mft record is found. */
-	for (; pass <= 2;) {
-		/* Cap size to pass_end. */
-		ofs = data_pos >> 3;
-		page_ofs = ofs & ~PAGE_MASK;
-		size = PAGE_SIZE - page_ofs;
-		ll = ((pass_end + 7) >> 3) - ofs;
-		if (size > ll)
-			size = ll;
-		size <<= 3;
-		/*
-		 * If we are still within the active pass, search the next page
-		 * for a zero bit.
-		 */
-		if (size) {
-			page = ntfs_map_page(mftbmp_mapping,
-					ofs >> PAGE_SHIFT);
-			if (IS_ERR(page)) {
-				ntfs_error(vol->sb, "Failed to read mft "
-						"bitmap, aborting.");
-				return PTR_ERR(page);
-			}
-			buf = (u8*)page_address(page) + page_ofs;
-			bit = data_pos & 7;
-			data_pos &= ~7ull;
-			ntfs_debug("Before inner for loop: size 0x%x, "
-					"data_pos 0x%llx, bit 0x%llx", size,
-					(long long)data_pos, (long long)bit);
-			for (; bit < size && data_pos + bit < pass_end;
-					bit &= ~7ull, bit += 8) {
-				byte = buf + (bit >> 3);
-				if (*byte == 0xff)
-					continue;
-				b = ffz((unsigned long)*byte);
-				if (b < 8 && b >= (bit & 7)) {
-					ll = data_pos + (bit & ~7ull) + b;
-					if (unlikely(ll > (1ll << 32))) {
-						ntfs_unmap_page(page);
-						return -ENOSPC;
-					}
-					*byte |= 1 << b;
-					flush_dcache_page(page);
-					set_page_dirty(page);
-					ntfs_unmap_page(page);
-					ntfs_debug("Done.  (Found and "
-							"allocated mft record "
-							"0x%llx.)",
-							(long long)ll);
-					return ll;
-				}
-			}
-			ntfs_debug("After inner for loop: size 0x%x, "
-					"data_pos 0x%llx, bit 0x%llx", size,
-					(long long)data_pos, (long long)bit);
-			data_pos += size;
-			ntfs_unmap_page(page);
-			/*
-			 * If the end of the pass has not been reached yet,
-			 * continue searching the mft bitmap for a zero bit.
-			 */
-			if (data_pos < pass_end)
-				continue;
-		}
-		/* Do the next pass. */
-		if (++pass == 2) {
-			/*
-			 * Starting the second pass, in which we scan the first
-			 * part of the zone which we omitted earlier.
-			 */
-			pass_end = pass_start;
-			data_pos = pass_start = 24;
-			ntfs_debug("pass %i, pass_start 0x%llx, pass_end "
-					"0x%llx.", pass, (long long)pass_start,
-					(long long)pass_end);
-			if (data_pos >= pass_end)
-				break;
-		}
-	}
-	/* No free mft records in currently initialized mft bitmap. */
-	ntfs_debug("Done.  (No free mft records left in currently initialized "
-			"mft bitmap.)");
-	return -ENOSPC;
-}
-
-/**
- * ntfs_mft_bitmap_extend_allocation_nolock - extend mft bitmap by a cluster
- * @vol:	volume on which to extend the mft bitmap attribute
- *
- * Extend the mft bitmap attribute on the ntfs volume @vol by one cluster.
- *
- * Note: Only changes allocated_size, i.e. does not touch initialized_size or
- * data_size.
- *
- * Return 0 on success and -errno on error.
- *
- * Locking: - Caller must hold vol->mftbmp_lock for writing.
- *	    - This function takes NTFS_I(vol->mftbmp_ino)->runlist.lock for
- *	      writing and releases it before returning.
- *	    - This function takes vol->lcnbmp_lock for writing and releases it
- *	      before returning.
- */
-static int ntfs_mft_bitmap_extend_allocation_nolock(ntfs_volume *vol)
-{
-	LCN lcn;
-	s64 ll;
-	unsigned long flags;
-	struct page *page;
-	ntfs_inode *mft_ni, *mftbmp_ni;
-	runlist_element *rl, *rl2 = NULL;
-	ntfs_attr_search_ctx *ctx = NULL;
-	MFT_RECORD *mrec;
-	ATTR_RECORD *a = NULL;
-	int ret, mp_size;
-	u32 old_alen = 0;
-	u8 *b, tb;
-	struct {
-		u8 added_cluster:1;
-		u8 added_run:1;
-		u8 mp_rebuilt:1;
-	} status = { 0, 0, 0 };
-
-	ntfs_debug("Extending mft bitmap allocation.");
-	mft_ni = NTFS_I(vol->mft_ino);
-	mftbmp_ni = NTFS_I(vol->mftbmp_ino);
-	/*
-	 * Determine the last lcn of the mft bitmap.  The allocated size of the
-	 * mft bitmap cannot be zero so we are ok to do this.
-	 */
-	down_write(&mftbmp_ni->runlist.lock);
-	read_lock_irqsave(&mftbmp_ni->size_lock, flags);
-	ll = mftbmp_ni->allocated_size;
-	read_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
-	rl = ntfs_attr_find_vcn_nolock(mftbmp_ni,
-			(ll - 1) >> vol->cluster_size_bits, NULL);
-	if (IS_ERR(rl) || unlikely(!rl->length || rl->lcn < 0)) {
-		up_write(&mftbmp_ni->runlist.lock);
-		ntfs_error(vol->sb, "Failed to determine last allocated "
-				"cluster of mft bitmap attribute.");
-		if (!IS_ERR(rl))
-			ret = -EIO;
-		else
-			ret = PTR_ERR(rl);
-		return ret;
-	}
-	lcn = rl->lcn + rl->length;
-	ntfs_debug("Last lcn of mft bitmap attribute is 0x%llx.",
-			(long long)lcn);
-	/*
-	 * Attempt to get the cluster following the last allocated cluster by
-	 * hand as it may be in the MFT zone so the allocator would not give it
-	 * to us.
-	 */
-	ll = lcn >> 3;
-	page = ntfs_map_page(vol->lcnbmp_ino->i_mapping,
-			ll >> PAGE_SHIFT);
-	if (IS_ERR(page)) {
-		up_write(&mftbmp_ni->runlist.lock);
-		ntfs_error(vol->sb, "Failed to read from lcn bitmap.");
-		return PTR_ERR(page);
-	}
-	b = (u8*)page_address(page) + (ll & ~PAGE_MASK);
-	tb = 1 << (lcn & 7ull);
-	down_write(&vol->lcnbmp_lock);
-	if (*b != 0xff && !(*b & tb)) {
-		/* Next cluster is free, allocate it. */
-		*b |= tb;
-		flush_dcache_page(page);
-		set_page_dirty(page);
-		up_write(&vol->lcnbmp_lock);
-		ntfs_unmap_page(page);
-		/* Update the mft bitmap runlist. */
-		rl->length++;
-		rl[1].vcn++;
-		status.added_cluster = 1;
-		ntfs_debug("Appending one cluster to mft bitmap.");
-	} else {
-		up_write(&vol->lcnbmp_lock);
-		ntfs_unmap_page(page);
-		/* Allocate a cluster from the DATA_ZONE. */
-		rl2 = ntfs_cluster_alloc(vol, rl[1].vcn, 1, lcn, DATA_ZONE,
-				true);
-		if (IS_ERR(rl2)) {
-			up_write(&mftbmp_ni->runlist.lock);
-			ntfs_error(vol->sb, "Failed to allocate a cluster for "
-					"the mft bitmap.");
-			return PTR_ERR(rl2);
-		}
-		rl = ntfs_runlists_merge(mftbmp_ni->runlist.rl, rl2);
-		if (IS_ERR(rl)) {
-			up_write(&mftbmp_ni->runlist.lock);
-			ntfs_error(vol->sb, "Failed to merge runlists for mft "
-					"bitmap.");
-			if (ntfs_cluster_free_from_rl(vol, rl2)) {
-				ntfs_error(vol->sb, "Failed to deallocate "
-						"allocated cluster.%s", es);
-				NVolSetErrors(vol);
-			}
-			ntfs_free(rl2);
-			return PTR_ERR(rl);
-		}
-		mftbmp_ni->runlist.rl = rl;
-		status.added_run = 1;
-		ntfs_debug("Adding one run to mft bitmap.");
-		/* Find the last run in the new runlist. */
-		for (; rl[1].length; rl++)
-			;
-	}
-	/*
-	 * Update the attribute record as well.  Note: @rl is the last
-	 * (non-terminator) runlist element of mft bitmap.
-	 */
-	mrec = map_mft_record(mft_ni);
-	if (IS_ERR(mrec)) {
-		ntfs_error(vol->sb, "Failed to map mft record.");
-		ret = PTR_ERR(mrec);
-		goto undo_alloc;
-	}
-	ctx = ntfs_attr_get_search_ctx(mft_ni, mrec);
-	if (unlikely(!ctx)) {
-		ntfs_error(vol->sb, "Failed to get search context.");
-		ret = -ENOMEM;
-		goto undo_alloc;
-	}
-	ret = ntfs_attr_lookup(mftbmp_ni->type, mftbmp_ni->name,
-			mftbmp_ni->name_len, CASE_SENSITIVE, rl[1].vcn, NULL,
-			0, ctx);
-	if (unlikely(ret)) {
-		ntfs_error(vol->sb, "Failed to find last attribute extent of "
-				"mft bitmap attribute.");
-		if (ret == -ENOENT)
-			ret = -EIO;
-		goto undo_alloc;
-	}
-	a = ctx->attr;
-	ll = sle64_to_cpu(a->data.non_resident.lowest_vcn);
-	/* Search back for the previous last allocated cluster of mft bitmap. */
-	for (rl2 = rl; rl2 > mftbmp_ni->runlist.rl; rl2--) {
-		if (ll >= rl2->vcn)
-			break;
-	}
-	BUG_ON(ll < rl2->vcn);
-	BUG_ON(ll >= rl2->vcn + rl2->length);
-	/* Get the size for the new mapping pairs array for this extent. */
-	mp_size = ntfs_get_size_for_mapping_pairs(vol, rl2, ll, -1);
-	if (unlikely(mp_size <= 0)) {
-		ntfs_error(vol->sb, "Get size for mapping pairs failed for "
-				"mft bitmap attribute extent.");
-		ret = mp_size;
-		if (!ret)
-			ret = -EIO;
-		goto undo_alloc;
-	}
-	/* Expand the attribute record if necessary. */
-	old_alen = le32_to_cpu(a->length);
-	ret = ntfs_attr_record_resize(ctx->mrec, a, mp_size +
-			le16_to_cpu(a->data.non_resident.mapping_pairs_offset));
-	if (unlikely(ret)) {
-		if (ret != -ENOSPC) {
-			ntfs_error(vol->sb, "Failed to resize attribute "
-					"record for mft bitmap attribute.");
-			goto undo_alloc;
-		}
-		// TODO: Deal with this by moving this extent to a new mft
-		// record or by starting a new extent in a new mft record or by
-		// moving other attributes out of this mft record.
-		// Note: It will need to be a special mft record and if none of
-		// those are available it gets rather complicated...
-		ntfs_error(vol->sb, "Not enough space in this mft record to "
-				"accommodate extended mft bitmap attribute "
-				"extent.  Cannot handle this yet.");
-		ret = -EOPNOTSUPP;
-		goto undo_alloc;
-	}
-	status.mp_rebuilt = 1;
-	/* Generate the mapping pairs array directly into the attr record. */
-	ret = ntfs_mapping_pairs_build(vol, (u8*)a +
-			le16_to_cpu(a->data.non_resident.mapping_pairs_offset),
-			mp_size, rl2, ll, -1, NULL);
-	if (unlikely(ret)) {
-		ntfs_error(vol->sb, "Failed to build mapping pairs array for "
-				"mft bitmap attribute.");
-		goto undo_alloc;
-	}
-	/* Update the highest_vcn. */
-	a->data.non_resident.highest_vcn = cpu_to_sle64(rl[1].vcn - 1);
-	/*
-	 * We now have extended the mft bitmap allocated_size by one cluster.
-	 * Reflect this in the ntfs_inode structure and the attribute record.
-	 */
-	if (a->data.non_resident.lowest_vcn) {
-		/*
-		 * We are not in the first attribute extent, switch to it, but
-		 * first ensure the changes will make it to disk later.
-		 */
-		flush_dcache_mft_record_page(ctx->ntfs_ino);
-		mark_mft_record_dirty(ctx->ntfs_ino);
-		ntfs_attr_reinit_search_ctx(ctx);
-		ret = ntfs_attr_lookup(mftbmp_ni->type, mftbmp_ni->name,
-				mftbmp_ni->name_len, CASE_SENSITIVE, 0, NULL,
-				0, ctx);
-		if (unlikely(ret)) {
-			ntfs_error(vol->sb, "Failed to find first attribute "
-					"extent of mft bitmap attribute.");
-			goto restore_undo_alloc;
-		}
-		a = ctx->attr;
-	}
-	write_lock_irqsave(&mftbmp_ni->size_lock, flags);
-	mftbmp_ni->allocated_size += vol->cluster_size;
-	a->data.non_resident.allocated_size =
-			cpu_to_sle64(mftbmp_ni->allocated_size);
-	write_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
-	/* Ensure the changes make it to disk. */
-	flush_dcache_mft_record_page(ctx->ntfs_ino);
-	mark_mft_record_dirty(ctx->ntfs_ino);
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(mft_ni);
-	up_write(&mftbmp_ni->runlist.lock);
-	ntfs_debug("Done.");
-	return 0;
-restore_undo_alloc:
-	ntfs_attr_reinit_search_ctx(ctx);
-	if (ntfs_attr_lookup(mftbmp_ni->type, mftbmp_ni->name,
-			mftbmp_ni->name_len, CASE_SENSITIVE, rl[1].vcn, NULL,
-			0, ctx)) {
-		ntfs_error(vol->sb, "Failed to find last attribute extent of "
-				"mft bitmap attribute.%s", es);
-		write_lock_irqsave(&mftbmp_ni->size_lock, flags);
-		mftbmp_ni->allocated_size += vol->cluster_size;
-		write_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
-		ntfs_attr_put_search_ctx(ctx);
-		unmap_mft_record(mft_ni);
-		up_write(&mftbmp_ni->runlist.lock);
-		/*
-		 * The only thing that is now wrong is ->allocated_size of the
-		 * base attribute extent which chkdsk should be able to fix.
-		 */
-		NVolSetErrors(vol);
-		return ret;
-	}
-	a = ctx->attr;
-	a->data.non_resident.highest_vcn = cpu_to_sle64(rl[1].vcn - 2);
-undo_alloc:
-	if (status.added_cluster) {
-		/* Truncate the last run in the runlist by one cluster. */
-		rl->length--;
-		rl[1].vcn--;
-	} else if (status.added_run) {
-		lcn = rl->lcn;
-		/* Remove the last run from the runlist. */
-		rl->lcn = rl[1].lcn;
-		rl->length = 0;
-	}
-	/* Deallocate the cluster. */
-	down_write(&vol->lcnbmp_lock);
-	if (ntfs_bitmap_clear_bit(vol->lcnbmp_ino, lcn)) {
-		ntfs_error(vol->sb, "Failed to free allocated cluster.%s", es);
-		NVolSetErrors(vol);
-	}
-	up_write(&vol->lcnbmp_lock);
-	if (status.mp_rebuilt) {
-		if (ntfs_mapping_pairs_build(vol, (u8*)a + le16_to_cpu(
-				a->data.non_resident.mapping_pairs_offset),
-				old_alen - le16_to_cpu(
-				a->data.non_resident.mapping_pairs_offset),
-				rl2, ll, -1, NULL)) {
-			ntfs_error(vol->sb, "Failed to restore mapping pairs "
-					"array.%s", es);
-			NVolSetErrors(vol);
-		}
-		if (ntfs_attr_record_resize(ctx->mrec, a, old_alen)) {
-			ntfs_error(vol->sb, "Failed to restore attribute "
-					"record.%s", es);
-			NVolSetErrors(vol);
-		}
-		flush_dcache_mft_record_page(ctx->ntfs_ino);
-		mark_mft_record_dirty(ctx->ntfs_ino);
-	}
-	if (ctx)
-		ntfs_attr_put_search_ctx(ctx);
-	if (!IS_ERR(mrec))
-		unmap_mft_record(mft_ni);
-	up_write(&mftbmp_ni->runlist.lock);
-	return ret;
-}
-
-/**
- * ntfs_mft_bitmap_extend_initialized_nolock - extend mftbmp initialized data
- * @vol:	volume on which to extend the mft bitmap attribute
- *
- * Extend the initialized portion of the mft bitmap attribute on the ntfs
- * volume @vol by 8 bytes.
- *
- * Note:  Only changes initialized_size and data_size, i.e. requires that
- * allocated_size is big enough to fit the new initialized_size.
- *
- * Return 0 on success and -error on error.
- *
- * Locking: Caller must hold vol->mftbmp_lock for writing.
- */
-static int ntfs_mft_bitmap_extend_initialized_nolock(ntfs_volume *vol)
-{
-	s64 old_data_size, old_initialized_size;
-	unsigned long flags;
-	struct inode *mftbmp_vi;
-	ntfs_inode *mft_ni, *mftbmp_ni;
-	ntfs_attr_search_ctx *ctx;
-	MFT_RECORD *mrec;
-	ATTR_RECORD *a;
-	int ret;
-
-	ntfs_debug("Extending mft bitmap initiailized (and data) size.");
-	mft_ni = NTFS_I(vol->mft_ino);
-	mftbmp_vi = vol->mftbmp_ino;
-	mftbmp_ni = NTFS_I(mftbmp_vi);
-	/* Get the attribute record. */
-	mrec = map_mft_record(mft_ni);
-	if (IS_ERR(mrec)) {
-		ntfs_error(vol->sb, "Failed to map mft record.");
-		return PTR_ERR(mrec);
-	}
-	ctx = ntfs_attr_get_search_ctx(mft_ni, mrec);
-	if (unlikely(!ctx)) {
-		ntfs_error(vol->sb, "Failed to get search context.");
-		ret = -ENOMEM;
-		goto unm_err_out;
-	}
-	ret = ntfs_attr_lookup(mftbmp_ni->type, mftbmp_ni->name,
-			mftbmp_ni->name_len, CASE_SENSITIVE, 0, NULL, 0, ctx);
-	if (unlikely(ret)) {
-		ntfs_error(vol->sb, "Failed to find first attribute extent of "
-				"mft bitmap attribute.");
-		if (ret == -ENOENT)
-			ret = -EIO;
-		goto put_err_out;
-	}
-	a = ctx->attr;
-	write_lock_irqsave(&mftbmp_ni->size_lock, flags);
-	old_data_size = i_size_read(mftbmp_vi);
-	old_initialized_size = mftbmp_ni->initialized_size;
-	/*
-	 * We can simply update the initialized_size before filling the space
-	 * with zeroes because the caller is holding the mft bitmap lock for
-	 * writing which ensures that no one else is trying to access the data.
-	 */
-	mftbmp_ni->initialized_size += 8;
-	a->data.non_resident.initialized_size =
-			cpu_to_sle64(mftbmp_ni->initialized_size);
-	if (mftbmp_ni->initialized_size > old_data_size) {
-		i_size_write(mftbmp_vi, mftbmp_ni->initialized_size);
-		a->data.non_resident.data_size =
-				cpu_to_sle64(mftbmp_ni->initialized_size);
-	}
-	write_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
-	/* Ensure the changes make it to disk. */
-	flush_dcache_mft_record_page(ctx->ntfs_ino);
-	mark_mft_record_dirty(ctx->ntfs_ino);
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(mft_ni);
-	/* Initialize the mft bitmap attribute value with zeroes. */
-	ret = ntfs_attr_set(mftbmp_ni, old_initialized_size, 8, 0);
-	if (likely(!ret)) {
-		ntfs_debug("Done.  (Wrote eight initialized bytes to mft "
-				"bitmap.");
-		return 0;
-	}
-	ntfs_error(vol->sb, "Failed to write to mft bitmap.");
-	/* Try to recover from the error. */
-	mrec = map_mft_record(mft_ni);
-	if (IS_ERR(mrec)) {
-		ntfs_error(vol->sb, "Failed to map mft record.%s", es);
-		NVolSetErrors(vol);
-		return ret;
-	}
-	ctx = ntfs_attr_get_search_ctx(mft_ni, mrec);
-	if (unlikely(!ctx)) {
-		ntfs_error(vol->sb, "Failed to get search context.%s", es);
-		NVolSetErrors(vol);
-		goto unm_err_out;
-	}
-	if (ntfs_attr_lookup(mftbmp_ni->type, mftbmp_ni->name,
-			mftbmp_ni->name_len, CASE_SENSITIVE, 0, NULL, 0, ctx)) {
-		ntfs_error(vol->sb, "Failed to find first attribute extent of "
-				"mft bitmap attribute.%s", es);
-		NVolSetErrors(vol);
-put_err_out:
-		ntfs_attr_put_search_ctx(ctx);
-unm_err_out:
-		unmap_mft_record(mft_ni);
-		goto err_out;
-	}
-	a = ctx->attr;
-	write_lock_irqsave(&mftbmp_ni->size_lock, flags);
-	mftbmp_ni->initialized_size = old_initialized_size;
-	a->data.non_resident.initialized_size =
-			cpu_to_sle64(old_initialized_size);
-	if (i_size_read(mftbmp_vi) != old_data_size) {
-		i_size_write(mftbmp_vi, old_data_size);
-		a->data.non_resident.data_size = cpu_to_sle64(old_data_size);
-	}
-	write_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
-	flush_dcache_mft_record_page(ctx->ntfs_ino);
-	mark_mft_record_dirty(ctx->ntfs_ino);
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(mft_ni);
-#ifdef DEBUG
-	read_lock_irqsave(&mftbmp_ni->size_lock, flags);
-	ntfs_debug("Restored status of mftbmp: allocated_size 0x%llx, "
-			"data_size 0x%llx, initialized_size 0x%llx.",
-			(long long)mftbmp_ni->allocated_size,
-			(long long)i_size_read(mftbmp_vi),
-			(long long)mftbmp_ni->initialized_size);
-	read_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
-#endif /* DEBUG */
-err_out:
-	return ret;
-}
-
-/**
- * ntfs_mft_data_extend_allocation_nolock - extend mft data attribute
- * @vol:	volume on which to extend the mft data attribute
- *
- * Extend the mft data attribute on the ntfs volume @vol by 16 mft records
- * worth of clusters or if not enough space for this by one mft record worth
- * of clusters.
- *
- * Note:  Only changes allocated_size, i.e. does not touch initialized_size or
- * data_size.
- *
- * Return 0 on success and -errno on error.
- *
- * Locking: - Caller must hold vol->mftbmp_lock for writing.
- *	    - This function takes NTFS_I(vol->mft_ino)->runlist.lock for
- *	      writing and releases it before returning.
- *	    - This function calls functions which take vol->lcnbmp_lock for
- *	      writing and release it before returning.
- */
-static int ntfs_mft_data_extend_allocation_nolock(ntfs_volume *vol)
-{
-	LCN lcn;
-	VCN old_last_vcn;
-	s64 min_nr, nr, ll;
-	unsigned long flags;
-	ntfs_inode *mft_ni;
-	runlist_element *rl, *rl2;
-	ntfs_attr_search_ctx *ctx = NULL;
-	MFT_RECORD *mrec;
-	ATTR_RECORD *a = NULL;
-	int ret, mp_size;
-	u32 old_alen = 0;
-	bool mp_rebuilt = false;
-
-	ntfs_debug("Extending mft data allocation.");
-	mft_ni = NTFS_I(vol->mft_ino);
-	/*
-	 * Determine the preferred allocation location, i.e. the last lcn of
-	 * the mft data attribute.  The allocated size of the mft data
-	 * attribute cannot be zero so we are ok to do this.
-	 */
-	down_write(&mft_ni->runlist.lock);
-	read_lock_irqsave(&mft_ni->size_lock, flags);
-	ll = mft_ni->allocated_size;
-	read_unlock_irqrestore(&mft_ni->size_lock, flags);
-	rl = ntfs_attr_find_vcn_nolock(mft_ni,
-			(ll - 1) >> vol->cluster_size_bits, NULL);
-	if (IS_ERR(rl) || unlikely(!rl->length || rl->lcn < 0)) {
-		up_write(&mft_ni->runlist.lock);
-		ntfs_error(vol->sb, "Failed to determine last allocated "
-				"cluster of mft data attribute.");
-		if (!IS_ERR(rl))
-			ret = -EIO;
-		else
-			ret = PTR_ERR(rl);
-		return ret;
-	}
-	lcn = rl->lcn + rl->length;
-	ntfs_debug("Last lcn of mft data attribute is 0x%llx.", (long long)lcn);
-	/* Minimum allocation is one mft record worth of clusters. */
-	min_nr = vol->mft_record_size >> vol->cluster_size_bits;
-	if (!min_nr)
-		min_nr = 1;
-	/* Want to allocate 16 mft records worth of clusters. */
-	nr = vol->mft_record_size << 4 >> vol->cluster_size_bits;
-	if (!nr)
-		nr = min_nr;
-	/* Ensure we do not go above 2^32-1 mft records. */
-	read_lock_irqsave(&mft_ni->size_lock, flags);
-	ll = mft_ni->allocated_size;
-	read_unlock_irqrestore(&mft_ni->size_lock, flags);
-	if (unlikely((ll + (nr << vol->cluster_size_bits)) >>
-			vol->mft_record_size_bits >= (1ll << 32))) {
-		nr = min_nr;
-		if (unlikely((ll + (nr << vol->cluster_size_bits)) >>
-				vol->mft_record_size_bits >= (1ll << 32))) {
-			ntfs_warning(vol->sb, "Cannot allocate mft record "
-					"because the maximum number of inodes "
-					"(2^32) has already been reached.");
-			up_write(&mft_ni->runlist.lock);
-			return -ENOSPC;
-		}
-	}
-	ntfs_debug("Trying mft data allocation with %s cluster count %lli.",
-			nr > min_nr ? "default" : "minimal", (long long)nr);
-	old_last_vcn = rl[1].vcn;
-	do {
-		rl2 = ntfs_cluster_alloc(vol, old_last_vcn, nr, lcn, MFT_ZONE,
-				true);
-		if (!IS_ERR(rl2))
-			break;
-		if (PTR_ERR(rl2) != -ENOSPC || nr == min_nr) {
-			ntfs_error(vol->sb, "Failed to allocate the minimal "
-					"number of clusters (%lli) for the "
-					"mft data attribute.", (long long)nr);
-			up_write(&mft_ni->runlist.lock);
-			return PTR_ERR(rl2);
-		}
-		/*
-		 * There is not enough space to do the allocation, but there
-		 * might be enough space to do a minimal allocation so try that
-		 * before failing.
-		 */
-		nr = min_nr;
-		ntfs_debug("Retrying mft data allocation with minimal cluster "
-				"count %lli.", (long long)nr);
-	} while (1);
-	rl = ntfs_runlists_merge(mft_ni->runlist.rl, rl2);
-	if (IS_ERR(rl)) {
-		up_write(&mft_ni->runlist.lock);
-		ntfs_error(vol->sb, "Failed to merge runlists for mft data "
-				"attribute.");
-		if (ntfs_cluster_free_from_rl(vol, rl2)) {
-			ntfs_error(vol->sb, "Failed to deallocate clusters "
-					"from the mft data attribute.%s", es);
-			NVolSetErrors(vol);
-		}
-		ntfs_free(rl2);
-		return PTR_ERR(rl);
-	}
-	mft_ni->runlist.rl = rl;
-	ntfs_debug("Allocated %lli clusters.", (long long)nr);
-	/* Find the last run in the new runlist. */
-	for (; rl[1].length; rl++)
-		;
-	/* Update the attribute record as well. */
-	mrec = map_mft_record(mft_ni);
-	if (IS_ERR(mrec)) {
-		ntfs_error(vol->sb, "Failed to map mft record.");
-		ret = PTR_ERR(mrec);
-		goto undo_alloc;
-	}
-	ctx = ntfs_attr_get_search_ctx(mft_ni, mrec);
-	if (unlikely(!ctx)) {
-		ntfs_error(vol->sb, "Failed to get search context.");
-		ret = -ENOMEM;
-		goto undo_alloc;
-	}
-	ret = ntfs_attr_lookup(mft_ni->type, mft_ni->name, mft_ni->name_len,
-			CASE_SENSITIVE, rl[1].vcn, NULL, 0, ctx);
-	if (unlikely(ret)) {
-		ntfs_error(vol->sb, "Failed to find last attribute extent of "
-				"mft data attribute.");
-		if (ret == -ENOENT)
-			ret = -EIO;
-		goto undo_alloc;
-	}
-	a = ctx->attr;
-	ll = sle64_to_cpu(a->data.non_resident.lowest_vcn);
-	/* Search back for the previous last allocated cluster of mft bitmap. */
-	for (rl2 = rl; rl2 > mft_ni->runlist.rl; rl2--) {
-		if (ll >= rl2->vcn)
-			break;
-	}
-	BUG_ON(ll < rl2->vcn);
-	BUG_ON(ll >= rl2->vcn + rl2->length);
-	/* Get the size for the new mapping pairs array for this extent. */
-	mp_size = ntfs_get_size_for_mapping_pairs(vol, rl2, ll, -1);
-	if (unlikely(mp_size <= 0)) {
-		ntfs_error(vol->sb, "Get size for mapping pairs failed for "
-				"mft data attribute extent.");
-		ret = mp_size;
-		if (!ret)
-			ret = -EIO;
-		goto undo_alloc;
-	}
-	/* Expand the attribute record if necessary. */
-	old_alen = le32_to_cpu(a->length);
-	ret = ntfs_attr_record_resize(ctx->mrec, a, mp_size +
-			le16_to_cpu(a->data.non_resident.mapping_pairs_offset));
-	if (unlikely(ret)) {
-		if (ret != -ENOSPC) {
-			ntfs_error(vol->sb, "Failed to resize attribute "
-					"record for mft data attribute.");
-			goto undo_alloc;
-		}
-		// TODO: Deal with this by moving this extent to a new mft
-		// record or by starting a new extent in a new mft record or by
-		// moving other attributes out of this mft record.
-		// Note: Use the special reserved mft records and ensure that
-		// this extent is not required to find the mft record in
-		// question.  If no free special records left we would need to
-		// move an existing record away, insert ours in its place, and
-		// then place the moved record into the newly allocated space
-		// and we would then need to update all references to this mft
-		// record appropriately.  This is rather complicated...
-		ntfs_error(vol->sb, "Not enough space in this mft record to "
-				"accommodate extended mft data attribute "
-				"extent.  Cannot handle this yet.");
-		ret = -EOPNOTSUPP;
-		goto undo_alloc;
-	}
-	mp_rebuilt = true;
-	/* Generate the mapping pairs array directly into the attr record. */
-	ret = ntfs_mapping_pairs_build(vol, (u8*)a +
-			le16_to_cpu(a->data.non_resident.mapping_pairs_offset),
-			mp_size, rl2, ll, -1, NULL);
-	if (unlikely(ret)) {
-		ntfs_error(vol->sb, "Failed to build mapping pairs array of "
-				"mft data attribute.");
-		goto undo_alloc;
-	}
-	/* Update the highest_vcn. */
-	a->data.non_resident.highest_vcn = cpu_to_sle64(rl[1].vcn - 1);
-	/*
-	 * We now have extended the mft data allocated_size by nr clusters.
-	 * Reflect this in the ntfs_inode structure and the attribute record.
-	 * @rl is the last (non-terminator) runlist element of mft data
-	 * attribute.
-	 */
-	if (a->data.non_resident.lowest_vcn) {
-		/*
-		 * We are not in the first attribute extent, switch to it, but
-		 * first ensure the changes will make it to disk later.
-		 */
-		flush_dcache_mft_record_page(ctx->ntfs_ino);
-		mark_mft_record_dirty(ctx->ntfs_ino);
-		ntfs_attr_reinit_search_ctx(ctx);
-		ret = ntfs_attr_lookup(mft_ni->type, mft_ni->name,
-				mft_ni->name_len, CASE_SENSITIVE, 0, NULL, 0,
-				ctx);
-		if (unlikely(ret)) {
-			ntfs_error(vol->sb, "Failed to find first attribute "
-					"extent of mft data attribute.");
-			goto restore_undo_alloc;
-		}
-		a = ctx->attr;
-	}
-	write_lock_irqsave(&mft_ni->size_lock, flags);
-	mft_ni->allocated_size += nr << vol->cluster_size_bits;
-	a->data.non_resident.allocated_size =
-			cpu_to_sle64(mft_ni->allocated_size);
-	write_unlock_irqrestore(&mft_ni->size_lock, flags);
-	/* Ensure the changes make it to disk. */
-	flush_dcache_mft_record_page(ctx->ntfs_ino);
-	mark_mft_record_dirty(ctx->ntfs_ino);
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(mft_ni);
-	up_write(&mft_ni->runlist.lock);
-	ntfs_debug("Done.");
-	return 0;
-restore_undo_alloc:
-	ntfs_attr_reinit_search_ctx(ctx);
-	if (ntfs_attr_lookup(mft_ni->type, mft_ni->name, mft_ni->name_len,
-			CASE_SENSITIVE, rl[1].vcn, NULL, 0, ctx)) {
-		ntfs_error(vol->sb, "Failed to find last attribute extent of "
-				"mft data attribute.%s", es);
-		write_lock_irqsave(&mft_ni->size_lock, flags);
-		mft_ni->allocated_size += nr << vol->cluster_size_bits;
-		write_unlock_irqrestore(&mft_ni->size_lock, flags);
-		ntfs_attr_put_search_ctx(ctx);
-		unmap_mft_record(mft_ni);
-		up_write(&mft_ni->runlist.lock);
-		/*
-		 * The only thing that is now wrong is ->allocated_size of the
-		 * base attribute extent which chkdsk should be able to fix.
-		 */
-		NVolSetErrors(vol);
-		return ret;
-	}
-	ctx->attr->data.non_resident.highest_vcn =
-			cpu_to_sle64(old_last_vcn - 1);
-undo_alloc:
-	if (ntfs_cluster_free(mft_ni, old_last_vcn, -1, ctx) < 0) {
-		ntfs_error(vol->sb, "Failed to free clusters from mft data "
-				"attribute.%s", es);
-		NVolSetErrors(vol);
-	}
-
-	if (ntfs_rl_truncate_nolock(vol, &mft_ni->runlist, old_last_vcn)) {
-		ntfs_error(vol->sb, "Failed to truncate mft data attribute "
-				"runlist.%s", es);
-		NVolSetErrors(vol);
-	}
-	if (ctx) {
-		a = ctx->attr;
-		if (mp_rebuilt && !IS_ERR(ctx->mrec)) {
-			if (ntfs_mapping_pairs_build(vol, (u8 *)a + le16_to_cpu(
-				a->data.non_resident.mapping_pairs_offset),
-				old_alen - le16_to_cpu(
-					a->data.non_resident.mapping_pairs_offset),
-				rl2, ll, -1, NULL)) {
-				ntfs_error(vol->sb, "Failed to restore mapping pairs "
-					"array.%s", es);
-				NVolSetErrors(vol);
-			}
-			if (ntfs_attr_record_resize(ctx->mrec, a, old_alen)) {
-				ntfs_error(vol->sb, "Failed to restore attribute "
-					"record.%s", es);
-				NVolSetErrors(vol);
-			}
-			flush_dcache_mft_record_page(ctx->ntfs_ino);
-			mark_mft_record_dirty(ctx->ntfs_ino);
-		} else if (IS_ERR(ctx->mrec)) {
-			ntfs_error(vol->sb, "Failed to restore attribute search "
-				"context.%s", es);
-			NVolSetErrors(vol);
-		}
-		ntfs_attr_put_search_ctx(ctx);
-	}
-	if (!IS_ERR(mrec))
-		unmap_mft_record(mft_ni);
-	up_write(&mft_ni->runlist.lock);
-	return ret;
-}
-
-/**
- * ntfs_mft_record_layout - layout an mft record into a memory buffer
- * @vol:	volume to which the mft record will belong
- * @mft_no:	mft reference specifying the mft record number
- * @m:		destination buffer of size >= @vol->mft_record_size bytes
- *
- * Layout an empty, unused mft record with the mft record number @mft_no into
- * the buffer @m.  The volume @vol is needed because the mft record structure
- * was modified in NTFS 3.1 so we need to know which volume version this mft
- * record will be used on.
- *
- * Return 0 on success and -errno on error.
- */
-static int ntfs_mft_record_layout(const ntfs_volume *vol, const s64 mft_no,
-		MFT_RECORD *m)
-{
-	ATTR_RECORD *a;
-
-	ntfs_debug("Entering for mft record 0x%llx.", (long long)mft_no);
-	if (mft_no >= (1ll << 32)) {
-		ntfs_error(vol->sb, "Mft record number 0x%llx exceeds "
-				"maximum of 2^32.", (long long)mft_no);
-		return -ERANGE;
-	}
-	/* Start by clearing the whole mft record to gives us a clean slate. */
-	memset(m, 0, vol->mft_record_size);
-	/* Aligned to 2-byte boundary. */
-	if (vol->major_ver < 3 || (vol->major_ver == 3 && !vol->minor_ver))
-		m->usa_ofs = cpu_to_le16((sizeof(MFT_RECORD_OLD) + 1) & ~1);
-	else {
-		m->usa_ofs = cpu_to_le16((sizeof(MFT_RECORD) + 1) & ~1);
-		/*
-		 * Set the NTFS 3.1+ specific fields while we know that the
-		 * volume version is 3.1+.
-		 */
-		m->reserved = 0;
-		m->mft_record_number = cpu_to_le32((u32)mft_no);
-	}
-	m->magic = magic_FILE;
-	if (vol->mft_record_size >= NTFS_BLOCK_SIZE)
-		m->usa_count = cpu_to_le16(vol->mft_record_size /
-				NTFS_BLOCK_SIZE + 1);
-	else {
-		m->usa_count = cpu_to_le16(1);
-		ntfs_warning(vol->sb, "Sector size is bigger than mft record "
-				"size.  Setting usa_count to 1.  If chkdsk "
-				"reports this as corruption, please email "
-				"linux-ntfs-dev@lists.sourceforge.net stating "
-				"that you saw this message and that the "
-				"modified filesystem created was corrupt.  "
-				"Thank you.");
-	}
-	/* Set the update sequence number to 1. */
-	*(le16*)((u8*)m + le16_to_cpu(m->usa_ofs)) = cpu_to_le16(1);
-	m->lsn = 0;
-	m->sequence_number = cpu_to_le16(1);
-	m->link_count = 0;
-	/*
-	 * Place the attributes straight after the update sequence array,
-	 * aligned to 8-byte boundary.
-	 */
-	m->attrs_offset = cpu_to_le16((le16_to_cpu(m->usa_ofs) +
-			(le16_to_cpu(m->usa_count) << 1) + 7) & ~7);
-	m->flags = 0;
-	/*
-	 * Using attrs_offset plus eight bytes (for the termination attribute).
-	 * attrs_offset is already aligned to 8-byte boundary, so no need to
-	 * align again.
-	 */
-	m->bytes_in_use = cpu_to_le32(le16_to_cpu(m->attrs_offset) + 8);
-	m->bytes_allocated = cpu_to_le32(vol->mft_record_size);
-	m->base_mft_record = 0;
-	m->next_attr_instance = 0;
-	/* Add the termination attribute. */
-	a = (ATTR_RECORD*)((u8*)m + le16_to_cpu(m->attrs_offset));
-	a->type = AT_END;
-	a->length = 0;
-	ntfs_debug("Done.");
-	return 0;
-}
-
-/**
- * ntfs_mft_record_format - format an mft record on an ntfs volume
- * @vol:	volume on which to format the mft record
- * @mft_no:	mft record number to format
- *
- * Format the mft record @mft_no in $MFT/$DATA, i.e. lay out an empty, unused
- * mft record into the appropriate place of the mft data attribute.  This is
- * used when extending the mft data attribute.
- *
- * Return 0 on success and -errno on error.
- */
-static int ntfs_mft_record_format(const ntfs_volume *vol, const s64 mft_no)
-{
-	loff_t i_size;
-	struct inode *mft_vi = vol->mft_ino;
-	struct page *page;
-	MFT_RECORD *m;
-	pgoff_t index, end_index;
-	unsigned int ofs;
-	int err;
-
-	ntfs_debug("Entering for mft record 0x%llx.", (long long)mft_no);
-	/*
-	 * The index into the page cache and the offset within the page cache
-	 * page of the wanted mft record.
-	 */
-	index = mft_no << vol->mft_record_size_bits >> PAGE_SHIFT;
-	ofs = (mft_no << vol->mft_record_size_bits) & ~PAGE_MASK;
-	/* The maximum valid index into the page cache for $MFT's data. */
-	i_size = i_size_read(mft_vi);
-	end_index = i_size >> PAGE_SHIFT;
-	if (unlikely(index >= end_index)) {
-		if (unlikely(index > end_index || ofs + vol->mft_record_size >=
-				(i_size & ~PAGE_MASK))) {
-			ntfs_error(vol->sb, "Tried to format non-existing mft "
-					"record 0x%llx.", (long long)mft_no);
-			return -ENOENT;
-		}
-	}
-	/* Read, map, and pin the page containing the mft record. */
-	page = ntfs_map_page(mft_vi->i_mapping, index);
-	if (IS_ERR(page)) {
-		ntfs_error(vol->sb, "Failed to map page containing mft record "
-				"to format 0x%llx.", (long long)mft_no);
-		return PTR_ERR(page);
-	}
-	lock_page(page);
-	BUG_ON(!PageUptodate(page));
-	ClearPageUptodate(page);
-	m = (MFT_RECORD*)((u8*)page_address(page) + ofs);
-	err = ntfs_mft_record_layout(vol, mft_no, m);
-	if (unlikely(err)) {
-		ntfs_error(vol->sb, "Failed to layout mft record 0x%llx.",
-				(long long)mft_no);
-		SetPageUptodate(page);
-		unlock_page(page);
-		ntfs_unmap_page(page);
-		return err;
-	}
-	flush_dcache_page(page);
-	SetPageUptodate(page);
-	unlock_page(page);
-	/*
-	 * Make sure the mft record is written out to disk.  We could use
-	 * ilookup5() to check if an inode is in icache and so on but this is
-	 * unnecessary as ntfs_writepage() will write the dirty record anyway.
-	 */
-	mark_ntfs_record_dirty(page, ofs);
-	ntfs_unmap_page(page);
-	ntfs_debug("Done.");
-	return 0;
-}
-
-/**
- * ntfs_mft_record_alloc - allocate an mft record on an ntfs volume
- * @vol:	[IN]  volume on which to allocate the mft record
- * @mode:	[IN]  mode if want a file or directory, i.e. base inode or 0
- * @base_ni:	[IN]  open base inode if allocating an extent mft record or NULL
- * @mrec:	[OUT] on successful return this is the mapped mft record
- *
- * Allocate an mft record in $MFT/$DATA of an open ntfs volume @vol.
- *
- * If @base_ni is NULL make the mft record a base mft record, i.e. a file or
- * direvctory inode, and allocate it at the default allocator position.  In
- * this case @mode is the file mode as given to us by the caller.  We in
- * particular use @mode to distinguish whether a file or a directory is being
- * created (S_IFDIR(mode) and S_IFREG(mode), respectively).
- *
- * If @base_ni is not NULL make the allocated mft record an extent record,
- * allocate it starting at the mft record after the base mft record and attach
- * the allocated and opened ntfs inode to the base inode @base_ni.  In this
- * case @mode must be 0 as it is meaningless for extent inodes.
- *
- * You need to check the return value with IS_ERR().  If false, the function
- * was successful and the return value is the now opened ntfs inode of the
- * allocated mft record.  *@mrec is then set to the allocated, mapped, pinned,
- * and locked mft record.  If IS_ERR() is true, the function failed and the
- * error code is obtained from PTR_ERR(return value).  *@mrec is undefined in
- * this case.
- *
- * Allocation strategy:
- *
- * To find a free mft record, we scan the mft bitmap for a zero bit.  To
- * optimize this we start scanning at the place specified by @base_ni or if
- * @base_ni is NULL we start where we last stopped and we perform wrap around
- * when we reach the end.  Note, we do not try to allocate mft records below
- * number 24 because numbers 0 to 15 are the defined system files anyway and 16
- * to 24 are special in that they are used for storing extension mft records
- * for the $DATA attribute of $MFT.  This is required to avoid the possibility
- * of creating a runlist with a circular dependency which once written to disk
- * can never be read in again.  Windows will only use records 16 to 24 for
- * normal files if the volume is completely out of space.  We never use them
- * which means that when the volume is really out of space we cannot create any
- * more files while Windows can still create up to 8 small files.  We can start
- * doing this at some later time, it does not matter much for now.
- *
- * When scanning the mft bitmap, we only search up to the last allocated mft
- * record.  If there are no free records left in the range 24 to number of
- * allocated mft records, then we extend the $MFT/$DATA attribute in order to
- * create free mft records.  We extend the allocated size of $MFT/$DATA by 16
- * records at a time or one cluster, if cluster size is above 16kiB.  If there
- * is not sufficient space to do this, we try to extend by a single mft record
- * or one cluster, if cluster size is above the mft record size.
- *
- * No matter how many mft records we allocate, we initialize only the first
- * allocated mft record, incrementing mft data size and initialized size
- * accordingly, open an ntfs_inode for it and return it to the caller, unless
- * there are less than 24 mft records, in which case we allocate and initialize
- * mft records until we reach record 24 which we consider as the first free mft
- * record for use by normal files.
- *
- * If during any stage we overflow the initialized data in the mft bitmap, we
- * extend the initialized size (and data size) by 8 bytes, allocating another
- * cluster if required.  The bitmap data size has to be at least equal to the
- * number of mft records in the mft, but it can be bigger, in which case the
- * superflous bits are padded with zeroes.
- *
- * Thus, when we return successfully (IS_ERR() is false), we will have:
- *	- initialized / extended the mft bitmap if necessary,
- *	- initialized / extended the mft data if necessary,
- *	- set the bit corresponding to the mft record being allocated in the
- *	  mft bitmap,
- *	- opened an ntfs_inode for the allocated mft record, and we will have
- *	- returned the ntfs_inode as well as the allocated mapped, pinned, and
- *	  locked mft record.
- *
- * On error, the volume will be left in a consistent state and no record will
- * be allocated.  If rolling back a partial operation fails, we may leave some
- * inconsistent metadata in which case we set NVolErrors() so the volume is
- * left dirty when unmounted.
- *
- * Note, this function cannot make use of most of the normal functions, like
- * for example for attribute resizing, etc, because when the run list overflows
- * the base mft record and an attribute list is used, it is very important that
- * the extension mft records used to store the $DATA attribute of $MFT can be
- * reached without having to read the information contained inside them, as
- * this would make it impossible to find them in the first place after the
- * volume is unmounted.  $MFT/$BITMAP probably does not need to follow this
- * rule because the bitmap is not essential for finding the mft records, but on
- * the other hand, handling the bitmap in this special way would make life
- * easier because otherwise there might be circular invocations of functions
- * when reading the bitmap.
- */
-ntfs_inode *ntfs_mft_record_alloc(ntfs_volume *vol, const int mode,
-		ntfs_inode *base_ni, MFT_RECORD **mrec)
-{
-	s64 ll, bit, old_data_initialized, old_data_size;
-	unsigned long flags;
-	struct inode *vi;
-	struct page *page;
-	ntfs_inode *mft_ni, *mftbmp_ni, *ni;
-	ntfs_attr_search_ctx *ctx;
-	MFT_RECORD *m;
-	ATTR_RECORD *a;
-	pgoff_t index;
-	unsigned int ofs;
-	int err;
-	le16 seq_no, usn;
-	bool record_formatted = false;
-
-	if (base_ni) {
-		ntfs_debug("Entering (allocating an extent mft record for "
-				"base mft record 0x%llx).",
-				(long long)base_ni->mft_no);
-		/* @mode and @base_ni are mutually exclusive. */
-		BUG_ON(mode);
-	} else
-		ntfs_debug("Entering (allocating a base mft record).");
-	if (mode) {
-		/* @mode and @base_ni are mutually exclusive. */
-		BUG_ON(base_ni);
-		/* We only support creation of normal files and directories. */
-		if (!S_ISREG(mode) && !S_ISDIR(mode))
-			return ERR_PTR(-EOPNOTSUPP);
-	}
-	BUG_ON(!mrec);
-	mft_ni = NTFS_I(vol->mft_ino);
-	mftbmp_ni = NTFS_I(vol->mftbmp_ino);
-	down_write(&vol->mftbmp_lock);
-	bit = ntfs_mft_bitmap_find_and_alloc_free_rec_nolock(vol, base_ni);
-	if (bit >= 0) {
-		ntfs_debug("Found and allocated free record (#1), bit 0x%llx.",
-				(long long)bit);
-		goto have_alloc_rec;
-	}
-	if (bit != -ENOSPC) {
-		up_write(&vol->mftbmp_lock);
-		return ERR_PTR(bit);
-	}
-	/*
-	 * No free mft records left.  If the mft bitmap already covers more
-	 * than the currently used mft records, the next records are all free,
-	 * so we can simply allocate the first unused mft record.
-	 * Note: We also have to make sure that the mft bitmap at least covers
-	 * the first 24 mft records as they are special and whilst they may not
-	 * be in use, we do not allocate from them.
-	 */
-	read_lock_irqsave(&mft_ni->size_lock, flags);
-	ll = mft_ni->initialized_size >> vol->mft_record_size_bits;
-	read_unlock_irqrestore(&mft_ni->size_lock, flags);
-	read_lock_irqsave(&mftbmp_ni->size_lock, flags);
-	old_data_initialized = mftbmp_ni->initialized_size;
-	read_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
-	if (old_data_initialized << 3 > ll && old_data_initialized > 3) {
-		bit = ll;
-		if (bit < 24)
-			bit = 24;
-		if (unlikely(bit >= (1ll << 32)))
-			goto max_err_out;
-		ntfs_debug("Found free record (#2), bit 0x%llx.",
-				(long long)bit);
-		goto found_free_rec;
-	}
-	/*
-	 * The mft bitmap needs to be expanded until it covers the first unused
-	 * mft record that we can allocate.
-	 * Note: The smallest mft record we allocate is mft record 24.
-	 */
-	bit = old_data_initialized << 3;
-	if (unlikely(bit >= (1ll << 32)))
-		goto max_err_out;
-	read_lock_irqsave(&mftbmp_ni->size_lock, flags);
-	old_data_size = mftbmp_ni->allocated_size;
-	ntfs_debug("Status of mftbmp before extension: allocated_size 0x%llx, "
-			"data_size 0x%llx, initialized_size 0x%llx.",
-			(long long)old_data_size,
-			(long long)i_size_read(vol->mftbmp_ino),
-			(long long)old_data_initialized);
-	read_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
-	if (old_data_initialized + 8 > old_data_size) {
-		/* Need to extend bitmap by one more cluster. */
-		ntfs_debug("mftbmp: initialized_size + 8 > allocated_size.");
-		err = ntfs_mft_bitmap_extend_allocation_nolock(vol);
-		if (unlikely(err)) {
-			up_write(&vol->mftbmp_lock);
-			goto err_out;
-		}
-#ifdef DEBUG
-		read_lock_irqsave(&mftbmp_ni->size_lock, flags);
-		ntfs_debug("Status of mftbmp after allocation extension: "
-				"allocated_size 0x%llx, data_size 0x%llx, "
-				"initialized_size 0x%llx.",
-				(long long)mftbmp_ni->allocated_size,
-				(long long)i_size_read(vol->mftbmp_ino),
-				(long long)mftbmp_ni->initialized_size);
-		read_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
-#endif /* DEBUG */
-	}
-	/*
-	 * We now have sufficient allocated space, extend the initialized_size
-	 * as well as the data_size if necessary and fill the new space with
-	 * zeroes.
-	 */
-	err = ntfs_mft_bitmap_extend_initialized_nolock(vol);
-	if (unlikely(err)) {
-		up_write(&vol->mftbmp_lock);
-		goto err_out;
-	}
-#ifdef DEBUG
-	read_lock_irqsave(&mftbmp_ni->size_lock, flags);
-	ntfs_debug("Status of mftbmp after initialized extension: "
-			"allocated_size 0x%llx, data_size 0x%llx, "
-			"initialized_size 0x%llx.",
-			(long long)mftbmp_ni->allocated_size,
-			(long long)i_size_read(vol->mftbmp_ino),
-			(long long)mftbmp_ni->initialized_size);
-	read_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
-#endif /* DEBUG */
-	ntfs_debug("Found free record (#3), bit 0x%llx.", (long long)bit);
-found_free_rec:
-	/* @bit is the found free mft record, allocate it in the mft bitmap. */
-	ntfs_debug("At found_free_rec.");
-	err = ntfs_bitmap_set_bit(vol->mftbmp_ino, bit);
-	if (unlikely(err)) {
-		ntfs_error(vol->sb, "Failed to allocate bit in mft bitmap.");
-		up_write(&vol->mftbmp_lock);
-		goto err_out;
-	}
-	ntfs_debug("Set bit 0x%llx in mft bitmap.", (long long)bit);
-have_alloc_rec:
-	/*
-	 * The mft bitmap is now uptodate.  Deal with mft data attribute now.
-	 * Note, we keep hold of the mft bitmap lock for writing until all
-	 * modifications to the mft data attribute are complete, too, as they
-	 * will impact decisions for mft bitmap and mft record allocation done
-	 * by a parallel allocation and if the lock is not maintained a
-	 * parallel allocation could allocate the same mft record as this one.
-	 */
-	ll = (bit + 1) << vol->mft_record_size_bits;
-	read_lock_irqsave(&mft_ni->size_lock, flags);
-	old_data_initialized = mft_ni->initialized_size;
-	read_unlock_irqrestore(&mft_ni->size_lock, flags);
-	if (ll <= old_data_initialized) {
-		ntfs_debug("Allocated mft record already initialized.");
-		goto mft_rec_already_initialized;
-	}
-	ntfs_debug("Initializing allocated mft record.");
-	/*
-	 * The mft record is outside the initialized data.  Extend the mft data
-	 * attribute until it covers the allocated record.  The loop is only
-	 * actually traversed more than once when a freshly formatted volume is
-	 * first written to so it optimizes away nicely in the common case.
-	 */
-	read_lock_irqsave(&mft_ni->size_lock, flags);
-	ntfs_debug("Status of mft data before extension: "
-			"allocated_size 0x%llx, data_size 0x%llx, "
-			"initialized_size 0x%llx.",
-			(long long)mft_ni->allocated_size,
-			(long long)i_size_read(vol->mft_ino),
-			(long long)mft_ni->initialized_size);
-	while (ll > mft_ni->allocated_size) {
-		read_unlock_irqrestore(&mft_ni->size_lock, flags);
-		err = ntfs_mft_data_extend_allocation_nolock(vol);
-		if (unlikely(err)) {
-			ntfs_error(vol->sb, "Failed to extend mft data "
-					"allocation.");
-			goto undo_mftbmp_alloc_nolock;
-		}
-		read_lock_irqsave(&mft_ni->size_lock, flags);
-		ntfs_debug("Status of mft data after allocation extension: "
-				"allocated_size 0x%llx, data_size 0x%llx, "
-				"initialized_size 0x%llx.",
-				(long long)mft_ni->allocated_size,
-				(long long)i_size_read(vol->mft_ino),
-				(long long)mft_ni->initialized_size);
-	}
-	read_unlock_irqrestore(&mft_ni->size_lock, flags);
-	/*
-	 * Extend mft data initialized size (and data size of course) to reach
-	 * the allocated mft record, formatting the mft records allong the way.
-	 * Note: We only modify the ntfs_inode structure as that is all that is
-	 * needed by ntfs_mft_record_format().  We will update the attribute
-	 * record itself in one fell swoop later on.
-	 */
-	write_lock_irqsave(&mft_ni->size_lock, flags);
-	old_data_initialized = mft_ni->initialized_size;
-	old_data_size = vol->mft_ino->i_size;
-	while (ll > mft_ni->initialized_size) {
-		s64 new_initialized_size, mft_no;
-		
-		new_initialized_size = mft_ni->initialized_size +
-				vol->mft_record_size;
-		mft_no = mft_ni->initialized_size >> vol->mft_record_size_bits;
-		if (new_initialized_size > i_size_read(vol->mft_ino))
-			i_size_write(vol->mft_ino, new_initialized_size);
-		write_unlock_irqrestore(&mft_ni->size_lock, flags);
-		ntfs_debug("Initializing mft record 0x%llx.",
-				(long long)mft_no);
-		err = ntfs_mft_record_format(vol, mft_no);
-		if (unlikely(err)) {
-			ntfs_error(vol->sb, "Failed to format mft record.");
-			goto undo_data_init;
-		}
-		write_lock_irqsave(&mft_ni->size_lock, flags);
-		mft_ni->initialized_size = new_initialized_size;
-	}
-	write_unlock_irqrestore(&mft_ni->size_lock, flags);
-	record_formatted = true;
-	/* Update the mft data attribute record to reflect the new sizes. */
-	m = map_mft_record(mft_ni);
-	if (IS_ERR(m)) {
-		ntfs_error(vol->sb, "Failed to map mft record.");
-		err = PTR_ERR(m);
-		goto undo_data_init;
-	}
-	ctx = ntfs_attr_get_search_ctx(mft_ni, m);
-	if (unlikely(!ctx)) {
-		ntfs_error(vol->sb, "Failed to get search context.");
-		err = -ENOMEM;
-		unmap_mft_record(mft_ni);
-		goto undo_data_init;
-	}
-	err = ntfs_attr_lookup(mft_ni->type, mft_ni->name, mft_ni->name_len,
-			CASE_SENSITIVE, 0, NULL, 0, ctx);
-	if (unlikely(err)) {
-		ntfs_error(vol->sb, "Failed to find first attribute extent of "
-				"mft data attribute.");
-		ntfs_attr_put_search_ctx(ctx);
-		unmap_mft_record(mft_ni);
-		goto undo_data_init;
-	}
-	a = ctx->attr;
-	read_lock_irqsave(&mft_ni->size_lock, flags);
-	a->data.non_resident.initialized_size =
-			cpu_to_sle64(mft_ni->initialized_size);
-	a->data.non_resident.data_size =
-			cpu_to_sle64(i_size_read(vol->mft_ino));
-	read_unlock_irqrestore(&mft_ni->size_lock, flags);
-	/* Ensure the changes make it to disk. */
-	flush_dcache_mft_record_page(ctx->ntfs_ino);
-	mark_mft_record_dirty(ctx->ntfs_ino);
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(mft_ni);
-	read_lock_irqsave(&mft_ni->size_lock, flags);
-	ntfs_debug("Status of mft data after mft record initialization: "
-			"allocated_size 0x%llx, data_size 0x%llx, "
-			"initialized_size 0x%llx.",
-			(long long)mft_ni->allocated_size,
-			(long long)i_size_read(vol->mft_ino),
-			(long long)mft_ni->initialized_size);
-	BUG_ON(i_size_read(vol->mft_ino) > mft_ni->allocated_size);
-	BUG_ON(mft_ni->initialized_size > i_size_read(vol->mft_ino));
-	read_unlock_irqrestore(&mft_ni->size_lock, flags);
-mft_rec_already_initialized:
-	/*
-	 * We can finally drop the mft bitmap lock as the mft data attribute
-	 * has been fully updated.  The only disparity left is that the
-	 * allocated mft record still needs to be marked as in use to match the
-	 * set bit in the mft bitmap but this is actually not a problem since
-	 * this mft record is not referenced from anywhere yet and the fact
-	 * that it is allocated in the mft bitmap means that no-one will try to
-	 * allocate it either.
-	 */
-	up_write(&vol->mftbmp_lock);
-	/*
-	 * We now have allocated and initialized the mft record.  Calculate the
-	 * index of and the offset within the page cache page the record is in.
-	 */
-	index = bit << vol->mft_record_size_bits >> PAGE_SHIFT;
-	ofs = (bit << vol->mft_record_size_bits) & ~PAGE_MASK;
-	/* Read, map, and pin the page containing the mft record. */
-	page = ntfs_map_page(vol->mft_ino->i_mapping, index);
-	if (IS_ERR(page)) {
-		ntfs_error(vol->sb, "Failed to map page containing allocated "
-				"mft record 0x%llx.", (long long)bit);
-		err = PTR_ERR(page);
-		goto undo_mftbmp_alloc;
-	}
-	lock_page(page);
-	BUG_ON(!PageUptodate(page));
-	ClearPageUptodate(page);
-	m = (MFT_RECORD*)((u8*)page_address(page) + ofs);
-	/* If we just formatted the mft record no need to do it again. */
-	if (!record_formatted) {
-		/* Sanity check that the mft record is really not in use. */
-		if (ntfs_is_file_record(m->magic) &&
-				(m->flags & MFT_RECORD_IN_USE)) {
-			ntfs_error(vol->sb, "Mft record 0x%llx was marked "
-					"free in mft bitmap but is marked "
-					"used itself.  Corrupt filesystem.  "
-					"Unmount and run chkdsk.",
-					(long long)bit);
-			err = -EIO;
-			SetPageUptodate(page);
-			unlock_page(page);
-			ntfs_unmap_page(page);
-			NVolSetErrors(vol);
-			goto undo_mftbmp_alloc;
-		}
-		/*
-		 * We need to (re-)format the mft record, preserving the
-		 * sequence number if it is not zero as well as the update
-		 * sequence number if it is not zero or -1 (0xffff).  This
-		 * means we do not need to care whether or not something went
-		 * wrong with the previous mft record.
-		 */
-		seq_no = m->sequence_number;
-		usn = *(le16*)((u8*)m + le16_to_cpu(m->usa_ofs));
-		err = ntfs_mft_record_layout(vol, bit, m);
-		if (unlikely(err)) {
-			ntfs_error(vol->sb, "Failed to layout allocated mft "
-					"record 0x%llx.", (long long)bit);
-			SetPageUptodate(page);
-			unlock_page(page);
-			ntfs_unmap_page(page);
-			goto undo_mftbmp_alloc;
-		}
-		if (seq_no)
-			m->sequence_number = seq_no;
-		if (usn && le16_to_cpu(usn) != 0xffff)
-			*(le16*)((u8*)m + le16_to_cpu(m->usa_ofs)) = usn;
-	}
-	/* Set the mft record itself in use. */
-	m->flags |= MFT_RECORD_IN_USE;
-	if (S_ISDIR(mode))
-		m->flags |= MFT_RECORD_IS_DIRECTORY;
-	flush_dcache_page(page);
-	SetPageUptodate(page);
-	if (base_ni) {
-		MFT_RECORD *m_tmp;
-
-		/*
-		 * Setup the base mft record in the extent mft record.  This
-		 * completes initialization of the allocated extent mft record
-		 * and we can simply use it with map_extent_mft_record().
-		 */
-		m->base_mft_record = MK_LE_MREF(base_ni->mft_no,
-				base_ni->seq_no);
-		/*
-		 * Allocate an extent inode structure for the new mft record,
-		 * attach it to the base inode @base_ni and map, pin, and lock
-		 * its, i.e. the allocated, mft record.
-		 */
-		m_tmp = map_extent_mft_record(base_ni, bit, &ni);
-		if (IS_ERR(m_tmp)) {
-			ntfs_error(vol->sb, "Failed to map allocated extent "
-					"mft record 0x%llx.", (long long)bit);
-			err = PTR_ERR(m_tmp);
-			/* Set the mft record itself not in use. */
-			m->flags &= cpu_to_le16(
-					~le16_to_cpu(MFT_RECORD_IN_USE));
-			flush_dcache_page(page);
-			/* Make sure the mft record is written out to disk. */
-			mark_ntfs_record_dirty(page, ofs);
-			unlock_page(page);
-			ntfs_unmap_page(page);
-			goto undo_mftbmp_alloc;
-		}
-		BUG_ON(m != m_tmp);
-		/*
-		 * Make sure the allocated mft record is written out to disk.
-		 * No need to set the inode dirty because the caller is going
-		 * to do that anyway after finishing with the new extent mft
-		 * record (e.g. at a minimum a new attribute will be added to
-		 * the mft record.
-		 */
-		mark_ntfs_record_dirty(page, ofs);
-		unlock_page(page);
-		/*
-		 * Need to unmap the page since map_extent_mft_record() mapped
-		 * it as well so we have it mapped twice at the moment.
-		 */
-		ntfs_unmap_page(page);
-	} else {
-		/*
-		 * Allocate a new VFS inode and set it up.  NOTE: @vi->i_nlink
-		 * is set to 1 but the mft record->link_count is 0.  The caller
-		 * needs to bear this in mind.
-		 */
-		vi = new_inode(vol->sb);
-		if (unlikely(!vi)) {
-			err = -ENOMEM;
-			/* Set the mft record itself not in use. */
-			m->flags &= cpu_to_le16(
-					~le16_to_cpu(MFT_RECORD_IN_USE));
-			flush_dcache_page(page);
-			/* Make sure the mft record is written out to disk. */
-			mark_ntfs_record_dirty(page, ofs);
-			unlock_page(page);
-			ntfs_unmap_page(page);
-			goto undo_mftbmp_alloc;
-		}
-		vi->i_ino = bit;
-
-		/* The owner and group come from the ntfs volume. */
-		vi->i_uid = vol->uid;
-		vi->i_gid = vol->gid;
-
-		/* Initialize the ntfs specific part of @vi. */
-		ntfs_init_big_inode(vi);
-		ni = NTFS_I(vi);
-		/*
-		 * Set the appropriate mode, attribute type, and name.  For
-		 * directories, also setup the index values to the defaults.
-		 */
-		if (S_ISDIR(mode)) {
-			vi->i_mode = S_IFDIR | S_IRWXUGO;
-			vi->i_mode &= ~vol->dmask;
-
-			NInoSetMstProtected(ni);
-			ni->type = AT_INDEX_ALLOCATION;
-			ni->name = I30;
-			ni->name_len = 4;
-
-			ni->itype.index.block_size = 4096;
-			ni->itype.index.block_size_bits = ntfs_ffs(4096) - 1;
-			ni->itype.index.collation_rule = COLLATION_FILE_NAME;
-			if (vol->cluster_size <= ni->itype.index.block_size) {
-				ni->itype.index.vcn_size = vol->cluster_size;
-				ni->itype.index.vcn_size_bits =
-						vol->cluster_size_bits;
-			} else {
-				ni->itype.index.vcn_size = vol->sector_size;
-				ni->itype.index.vcn_size_bits =
-						vol->sector_size_bits;
-			}
-		} else {
-			vi->i_mode = S_IFREG | S_IRWXUGO;
-			vi->i_mode &= ~vol->fmask;
-
-			ni->type = AT_DATA;
-			ni->name = NULL;
-			ni->name_len = 0;
-		}
-		if (IS_RDONLY(vi))
-			vi->i_mode &= ~S_IWUGO;
-
-		/* Set the inode times to the current time. */
-		simple_inode_init_ts(vi);
-		/*
-		 * Set the file size to 0, the ntfs inode sizes are set to 0 by
-		 * the call to ntfs_init_big_inode() below.
-		 */
-		vi->i_size = 0;
-		vi->i_blocks = 0;
-
-		/* Set the sequence number. */
-		vi->i_generation = ni->seq_no = le16_to_cpu(m->sequence_number);
-		/*
-		 * Manually map, pin, and lock the mft record as we already
-		 * have its page mapped and it is very easy to do.
-		 */
-		atomic_inc(&ni->count);
-		mutex_lock(&ni->mrec_lock);
-		ni->page = page;
-		ni->page_ofs = ofs;
-		/*
-		 * Make sure the allocated mft record is written out to disk.
-		 * NOTE: We do not set the ntfs inode dirty because this would
-		 * fail in ntfs_write_inode() because the inode does not have a
-		 * standard information attribute yet.  Also, there is no need
-		 * to set the inode dirty because the caller is going to do
-		 * that anyway after finishing with the new mft record (e.g. at
-		 * a minimum some new attributes will be added to the mft
-		 * record.
-		 */
-		mark_ntfs_record_dirty(page, ofs);
-		unlock_page(page);
-
-		/* Add the inode to the inode hash for the superblock. */
-		insert_inode_hash(vi);
-
-		/* Update the default mft allocation position. */
-		vol->mft_data_pos = bit + 1;
-	}
-	/*
-	 * Return the opened, allocated inode of the allocated mft record as
-	 * well as the mapped, pinned, and locked mft record.
-	 */
-	ntfs_debug("Returning opened, allocated %sinode 0x%llx.",
-			base_ni ? "extent " : "", (long long)bit);
-	*mrec = m;
-	return ni;
-undo_data_init:
-	write_lock_irqsave(&mft_ni->size_lock, flags);
-	mft_ni->initialized_size = old_data_initialized;
-	i_size_write(vol->mft_ino, old_data_size);
-	write_unlock_irqrestore(&mft_ni->size_lock, flags);
-	goto undo_mftbmp_alloc_nolock;
-undo_mftbmp_alloc:
-	down_write(&vol->mftbmp_lock);
-undo_mftbmp_alloc_nolock:
-	if (ntfs_bitmap_clear_bit(vol->mftbmp_ino, bit)) {
-		ntfs_error(vol->sb, "Failed to clear bit in mft bitmap.%s", es);
-		NVolSetErrors(vol);
-	}
-	up_write(&vol->mftbmp_lock);
-err_out:
-	return ERR_PTR(err);
-max_err_out:
-	ntfs_warning(vol->sb, "Cannot allocate mft record because the maximum "
-			"number of inodes (2^32) has already been reached.");
-	up_write(&vol->mftbmp_lock);
-	return ERR_PTR(-ENOSPC);
-}
-
-/**
- * ntfs_extent_mft_record_free - free an extent mft record on an ntfs volume
- * @ni:		ntfs inode of the mapped extent mft record to free
- * @m:		mapped extent mft record of the ntfs inode @ni
- *
- * Free the mapped extent mft record @m of the extent ntfs inode @ni.
- *
- * Note that this function unmaps the mft record and closes and destroys @ni
- * internally and hence you cannot use either @ni nor @m any more after this
- * function returns success.
- *
- * On success return 0 and on error return -errno.  @ni and @m are still valid
- * in this case and have not been freed.
- *
- * For some errors an error message is displayed and the success code 0 is
- * returned and the volume is then left dirty on umount.  This makes sense in
- * case we could not rollback the changes that were already done since the
- * caller no longer wants to reference this mft record so it does not matter to
- * the caller if something is wrong with it as long as it is properly detached
- * from the base inode.
- */
-int ntfs_extent_mft_record_free(ntfs_inode *ni, MFT_RECORD *m)
-{
-	unsigned long mft_no = ni->mft_no;
-	ntfs_volume *vol = ni->vol;
-	ntfs_inode *base_ni;
-	ntfs_inode **extent_nis;
-	int i, err;
-	le16 old_seq_no;
-	u16 seq_no;
-	
-	BUG_ON(NInoAttr(ni));
-	BUG_ON(ni->nr_extents != -1);
-
-	mutex_lock(&ni->extent_lock);
-	base_ni = ni->ext.base_ntfs_ino;
-	mutex_unlock(&ni->extent_lock);
-
-	BUG_ON(base_ni->nr_extents <= 0);
-
-	ntfs_debug("Entering for extent inode 0x%lx, base inode 0x%lx.\n",
-			mft_no, base_ni->mft_no);
-
-	mutex_lock(&base_ni->extent_lock);
-
-	/* Make sure we are holding the only reference to the extent inode. */
-	if (atomic_read(&ni->count) > 2) {
-		ntfs_error(vol->sb, "Tried to free busy extent inode 0x%lx, "
-				"not freeing.", base_ni->mft_no);
-		mutex_unlock(&base_ni->extent_lock);
-		return -EBUSY;
-	}
-
-	/* Dissociate the ntfs inode from the base inode. */
-	extent_nis = base_ni->ext.extent_ntfs_inos;
-	err = -ENOENT;
-	for (i = 0; i < base_ni->nr_extents; i++) {
-		if (ni != extent_nis[i])
-			continue;
-		extent_nis += i;
-		base_ni->nr_extents--;
-		memmove(extent_nis, extent_nis + 1, (base_ni->nr_extents - i) *
-				sizeof(ntfs_inode*));
-		err = 0;
-		break;
-	}
-
-	mutex_unlock(&base_ni->extent_lock);
-
-	if (unlikely(err)) {
-		ntfs_error(vol->sb, "Extent inode 0x%lx is not attached to "
-				"its base inode 0x%lx.", mft_no,
-				base_ni->mft_no);
-		BUG();
-	}
-
-	/*
-	 * The extent inode is no longer attached to the base inode so no one
-	 * can get a reference to it any more.
-	 */
-
-	/* Mark the mft record as not in use. */
-	m->flags &= ~MFT_RECORD_IN_USE;
-
-	/* Increment the sequence number, skipping zero, if it is not zero. */
-	old_seq_no = m->sequence_number;
-	seq_no = le16_to_cpu(old_seq_no);
-	if (seq_no == 0xffff)
-		seq_no = 1;
-	else if (seq_no)
-		seq_no++;
-	m->sequence_number = cpu_to_le16(seq_no);
-
-	/*
-	 * Set the ntfs inode dirty and write it out.  We do not need to worry
-	 * about the base inode here since whatever caused the extent mft
-	 * record to be freed is guaranteed to do it already.
-	 */
-	NInoSetDirty(ni);
-	err = write_mft_record(ni, m, 0);
-	if (unlikely(err)) {
-		ntfs_error(vol->sb, "Failed to write mft record 0x%lx, not "
-				"freeing.", mft_no);
-		goto rollback;
-	}
-rollback_error:
-	/* Unmap and throw away the now freed extent inode. */
-	unmap_extent_mft_record(ni);
-	ntfs_clear_extent_inode(ni);
-
-	/* Clear the bit in the $MFT/$BITMAP corresponding to this record. */
-	down_write(&vol->mftbmp_lock);
-	err = ntfs_bitmap_clear_bit(vol->mftbmp_ino, mft_no);
-	up_write(&vol->mftbmp_lock);
-	if (unlikely(err)) {
-		/*
-		 * The extent inode is gone but we failed to deallocate it in
-		 * the mft bitmap.  Just emit a warning and leave the volume
-		 * dirty on umount.
-		 */
-		ntfs_error(vol->sb, "Failed to clear bit in mft bitmap.%s", es);
-		NVolSetErrors(vol);
-	}
-	return 0;
-rollback:
-	/* Rollback what we did... */
-	mutex_lock(&base_ni->extent_lock);
-	extent_nis = base_ni->ext.extent_ntfs_inos;
-	if (!(base_ni->nr_extents & 3)) {
-		int new_size = (base_ni->nr_extents + 4) * sizeof(ntfs_inode*);
-
-		extent_nis = kmalloc(new_size, GFP_NOFS);
-		if (unlikely(!extent_nis)) {
-			ntfs_error(vol->sb, "Failed to allocate internal "
-					"buffer during rollback.%s", es);
-			mutex_unlock(&base_ni->extent_lock);
-			NVolSetErrors(vol);
-			goto rollback_error;
-		}
-		if (base_ni->nr_extents) {
-			BUG_ON(!base_ni->ext.extent_ntfs_inos);
-			memcpy(extent_nis, base_ni->ext.extent_ntfs_inos,
-					new_size - 4 * sizeof(ntfs_inode*));
-			kfree(base_ni->ext.extent_ntfs_inos);
-		}
-		base_ni->ext.extent_ntfs_inos = extent_nis;
-	}
-	m->flags |= MFT_RECORD_IN_USE;
-	m->sequence_number = old_seq_no;
-	extent_nis[base_ni->nr_extents++] = ni;
-	mutex_unlock(&base_ni->extent_lock);
-	mark_mft_record_dirty(ni);
-	return err;
-}
-#endif /* NTFS_RW */
diff --git a/fs/ntfs/mft.h b/fs/ntfs/mft.h
deleted file mode 100644
index 49c001af16..0000000000
--- a/fs/ntfs/mft.h
+++ /dev/null
@@ -1,110 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * mft.h - Defines for mft record handling in NTFS Linux kernel driver.
- *	   Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2004 Anton Altaparmakov
- */
-
-#ifndef _LINUX_NTFS_MFT_H
-#define _LINUX_NTFS_MFT_H
-
-#include <linux/fs.h>
-#include <linux/highmem.h>
-#include <linux/pagemap.h>
-
-#include "inode.h"
-
-extern MFT_RECORD *map_mft_record(ntfs_inode *ni);
-extern void unmap_mft_record(ntfs_inode *ni);
-
-extern MFT_RECORD *map_extent_mft_record(ntfs_inode *base_ni, MFT_REF mref,
-		ntfs_inode **ntfs_ino);
-
-static inline void unmap_extent_mft_record(ntfs_inode *ni)
-{
-	unmap_mft_record(ni);
-	return;
-}
-
-#ifdef NTFS_RW
-
-/**
- * flush_dcache_mft_record_page - flush_dcache_page() for mft records
- * @ni:		ntfs inode structure of mft record
- *
- * Call flush_dcache_page() for the page in which an mft record resides.
- *
- * This must be called every time an mft record is modified, just after the
- * modification.
- */
-static inline void flush_dcache_mft_record_page(ntfs_inode *ni)
-{
-	flush_dcache_page(ni->page);
-}
-
-extern void __mark_mft_record_dirty(ntfs_inode *ni);
-
-/**
- * mark_mft_record_dirty - set the mft record and the page containing it dirty
- * @ni:		ntfs inode describing the mapped mft record
- *
- * Set the mapped (extent) mft record of the (base or extent) ntfs inode @ni,
- * as well as the page containing the mft record, dirty.  Also, mark the base
- * vfs inode dirty.  This ensures that any changes to the mft record are
- * written out to disk.
- *
- * NOTE:  Do not do anything if the mft record is already marked dirty.
- */
-static inline void mark_mft_record_dirty(ntfs_inode *ni)
-{
-	if (!NInoTestSetDirty(ni))
-		__mark_mft_record_dirty(ni);
-}
-
-extern int ntfs_sync_mft_mirror(ntfs_volume *vol, const unsigned long mft_no,
-		MFT_RECORD *m, int sync);
-
-extern int write_mft_record_nolock(ntfs_inode *ni, MFT_RECORD *m, int sync);
-
-/**
- * write_mft_record - write out a mapped (extent) mft record
- * @ni:		ntfs inode describing the mapped (extent) mft record
- * @m:		mapped (extent) mft record to write
- * @sync:	if true, wait for i/o completion
- *
- * This is just a wrapper for write_mft_record_nolock() (see mft.c), which
- * locks the page for the duration of the write.  This ensures that there are
- * no race conditions between writing the mft record via the dirty inode code
- * paths and via the page cache write back code paths or between writing
- * neighbouring mft records residing in the same page.
- *
- * Locking the page also serializes us against ->read_folio() if the page is not
- * uptodate.
- *
- * On success, clean the mft record and return 0.  On error, leave the mft
- * record dirty and return -errno.
- */
-static inline int write_mft_record(ntfs_inode *ni, MFT_RECORD *m, int sync)
-{
-	struct page *page = ni->page;
-	int err;
-
-	BUG_ON(!page);
-	lock_page(page);
-	err = write_mft_record_nolock(ni, m, sync);
-	unlock_page(page);
-	return err;
-}
-
-extern bool ntfs_may_write_mft_record(ntfs_volume *vol,
-		const unsigned long mft_no, const MFT_RECORD *m,
-		ntfs_inode **locked_ni);
-
-extern ntfs_inode *ntfs_mft_record_alloc(ntfs_volume *vol, const int mode,
-		ntfs_inode *base_ni, MFT_RECORD **mrec);
-extern int ntfs_extent_mft_record_free(ntfs_inode *ni, MFT_RECORD *m);
-
-#endif /* NTFS_RW */
-
-#endif /* _LINUX_NTFS_MFT_H */
diff --git a/fs/ntfs/mst.c b/fs/ntfs/mst.c
deleted file mode 100644
index 16b3c884ab..0000000000
--- a/fs/ntfs/mst.c
+++ /dev/null
@@ -1,189 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * mst.c - NTFS multi sector transfer protection handling code. Part of the
- *	   Linux-NTFS project.
- *
- * Copyright (c) 2001-2004 Anton Altaparmakov
- */
-
-#include "ntfs.h"
-
-/**
- * post_read_mst_fixup - deprotect multi sector transfer protected data
- * @b:		pointer to the data to deprotect
- * @size:	size in bytes of @b
- *
- * Perform the necessary post read multi sector transfer fixup and detect the
- * presence of incomplete multi sector transfers. - In that case, overwrite the
- * magic of the ntfs record header being processed with "BAAD" (in memory only!)
- * and abort processing.
- *
- * Return 0 on success and -EINVAL on error ("BAAD" magic will be present).
- *
- * NOTE: We consider the absence / invalidity of an update sequence array to
- * mean that the structure is not protected at all and hence doesn't need to
- * be fixed up. Thus, we return success and not failure in this case. This is
- * in contrast to pre_write_mst_fixup(), see below.
- */
-int post_read_mst_fixup(NTFS_RECORD *b, const u32 size)
-{
-	u16 usa_ofs, usa_count, usn;
-	u16 *usa_pos, *data_pos;
-
-	/* Setup the variables. */
-	usa_ofs = le16_to_cpu(b->usa_ofs);
-	/* Decrement usa_count to get number of fixups. */
-	usa_count = le16_to_cpu(b->usa_count) - 1;
-	/* Size and alignment checks. */
-	if ( size & (NTFS_BLOCK_SIZE - 1)	||
-	     usa_ofs & 1			||
-	     usa_ofs + (usa_count * 2) > size	||
-	     (size >> NTFS_BLOCK_SIZE_BITS) != usa_count)
-		return 0;
-	/* Position of usn in update sequence array. */
-	usa_pos = (u16*)b + usa_ofs/sizeof(u16);
-	/*
-	 * The update sequence number which has to be equal to each of the
-	 * u16 values before they are fixed up. Note no need to care for
-	 * endianness since we are comparing and moving data for on disk
-	 * structures which means the data is consistent. - If it is
-	 * consistenty the wrong endianness it doesn't make any difference.
-	 */
-	usn = *usa_pos;
-	/*
-	 * Position in protected data of first u16 that needs fixing up.
-	 */
-	data_pos = (u16*)b + NTFS_BLOCK_SIZE/sizeof(u16) - 1;
-	/*
-	 * Check for incomplete multi sector transfer(s).
-	 */
-	while (usa_count--) {
-		if (*data_pos != usn) {
-			/*
-			 * Incomplete multi sector transfer detected! )-:
-			 * Set the magic to "BAAD" and return failure.
-			 * Note that magic_BAAD is already converted to le32.
-			 */
-			b->magic = magic_BAAD;
-			return -EINVAL;
-		}
-		data_pos += NTFS_BLOCK_SIZE/sizeof(u16);
-	}
-	/* Re-setup the variables. */
-	usa_count = le16_to_cpu(b->usa_count) - 1;
-	data_pos = (u16*)b + NTFS_BLOCK_SIZE/sizeof(u16) - 1;
-	/* Fixup all sectors. */
-	while (usa_count--) {
-		/*
-		 * Increment position in usa and restore original data from
-		 * the usa into the data buffer.
-		 */
-		*data_pos = *(++usa_pos);
-		/* Increment position in data as well. */
-		data_pos += NTFS_BLOCK_SIZE/sizeof(u16);
-	}
-	return 0;
-}
-
-/**
- * pre_write_mst_fixup - apply multi sector transfer protection
- * @b:		pointer to the data to protect
- * @size:	size in bytes of @b
- *
- * Perform the necessary pre write multi sector transfer fixup on the data
- * pointer to by @b of @size.
- *
- * Return 0 if fixup applied (success) or -EINVAL if no fixup was performed
- * (assumed not needed). This is in contrast to post_read_mst_fixup() above.
- *
- * NOTE: We consider the absence / invalidity of an update sequence array to
- * mean that the structure is not subject to protection and hence doesn't need
- * to be fixed up. This means that you have to create a valid update sequence
- * array header in the ntfs record before calling this function, otherwise it
- * will fail (the header needs to contain the position of the update sequence
- * array together with the number of elements in the array). You also need to
- * initialise the update sequence number before calling this function
- * otherwise a random word will be used (whatever was in the record at that
- * position at that time).
- */
-int pre_write_mst_fixup(NTFS_RECORD *b, const u32 size)
-{
-	le16 *usa_pos, *data_pos;
-	u16 usa_ofs, usa_count, usn;
-	le16 le_usn;
-
-	/* Sanity check + only fixup if it makes sense. */
-	if (!b || ntfs_is_baad_record(b->magic) ||
-			ntfs_is_hole_record(b->magic))
-		return -EINVAL;
-	/* Setup the variables. */
-	usa_ofs = le16_to_cpu(b->usa_ofs);
-	/* Decrement usa_count to get number of fixups. */
-	usa_count = le16_to_cpu(b->usa_count) - 1;
-	/* Size and alignment checks. */
-	if ( size & (NTFS_BLOCK_SIZE - 1)	||
-	     usa_ofs & 1			||
-	     usa_ofs + (usa_count * 2) > size	||
-	     (size >> NTFS_BLOCK_SIZE_BITS) != usa_count)
-		return -EINVAL;
-	/* Position of usn in update sequence array. */
-	usa_pos = (le16*)((u8*)b + usa_ofs);
-	/*
-	 * Cyclically increment the update sequence number
-	 * (skipping 0 and -1, i.e. 0xffff).
-	 */
-	usn = le16_to_cpup(usa_pos) + 1;
-	if (usn == 0xffff || !usn)
-		usn = 1;
-	le_usn = cpu_to_le16(usn);
-	*usa_pos = le_usn;
-	/* Position in data of first u16 that needs fixing up. */
-	data_pos = (le16*)b + NTFS_BLOCK_SIZE/sizeof(le16) - 1;
-	/* Fixup all sectors. */
-	while (usa_count--) {
-		/*
-		 * Increment the position in the usa and save the
-		 * original data from the data buffer into the usa.
-		 */
-		*(++usa_pos) = *data_pos;
-		/* Apply fixup to data. */
-		*data_pos = le_usn;
-		/* Increment position in data as well. */
-		data_pos += NTFS_BLOCK_SIZE/sizeof(le16);
-	}
-	return 0;
-}
-
-/**
- * post_write_mst_fixup - fast deprotect multi sector transfer protected data
- * @b:		pointer to the data to deprotect
- *
- * Perform the necessary post write multi sector transfer fixup, not checking
- * for any errors, because we assume we have just used pre_write_mst_fixup(),
- * thus the data will be fine or we would never have gotten here.
- */
-void post_write_mst_fixup(NTFS_RECORD *b)
-{
-	le16 *usa_pos, *data_pos;
-
-	u16 usa_ofs = le16_to_cpu(b->usa_ofs);
-	u16 usa_count = le16_to_cpu(b->usa_count) - 1;
-
-	/* Position of usn in update sequence array. */
-	usa_pos = (le16*)b + usa_ofs/sizeof(le16);
-
-	/* Position in protected data of first u16 that needs fixing up. */
-	data_pos = (le16*)b + NTFS_BLOCK_SIZE/sizeof(le16) - 1;
-
-	/* Fixup all sectors. */
-	while (usa_count--) {
-		/*
-		 * Increment position in usa and restore original data from
-		 * the usa into the data buffer.
-		 */
-		*data_pos = *(++usa_pos);
-
-		/* Increment position in data as well. */
-		data_pos += NTFS_BLOCK_SIZE/sizeof(le16);
-	}
-}
diff --git a/fs/ntfs/namei.c b/fs/ntfs/namei.c
deleted file mode 100644
index d7498ddc4a..0000000000
--- a/fs/ntfs/namei.c
+++ /dev/null
@@ -1,392 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * namei.c - NTFS kernel directory inode operations. Part of the Linux-NTFS
- *	     project.
- *
- * Copyright (c) 2001-2006 Anton Altaparmakov
- */
-
-#include <linux/dcache.h>
-#include <linux/exportfs.h>
-#include <linux/security.h>
-#include <linux/slab.h>
-
-#include "attrib.h"
-#include "debug.h"
-#include "dir.h"
-#include "mft.h"
-#include "ntfs.h"
-
-/**
- * ntfs_lookup - find the inode represented by a dentry in a directory inode
- * @dir_ino:	directory inode in which to look for the inode
- * @dent:	dentry representing the inode to look for
- * @flags:	lookup flags
- *
- * In short, ntfs_lookup() looks for the inode represented by the dentry @dent
- * in the directory inode @dir_ino and if found attaches the inode to the
- * dentry @dent.
- *
- * In more detail, the dentry @dent specifies which inode to look for by
- * supplying the name of the inode in @dent->d_name.name. ntfs_lookup()
- * converts the name to Unicode and walks the contents of the directory inode
- * @dir_ino looking for the converted Unicode name. If the name is found in the
- * directory, the corresponding inode is loaded by calling ntfs_iget() on its
- * inode number and the inode is associated with the dentry @dent via a call to
- * d_splice_alias().
- *
- * If the name is not found in the directory, a NULL inode is inserted into the
- * dentry @dent via a call to d_add(). The dentry is then termed a negative
- * dentry.
- *
- * Only if an actual error occurs, do we return an error via ERR_PTR().
- *
- * In order to handle the case insensitivity issues of NTFS with regards to the
- * dcache and the dcache requiring only one dentry per directory, we deal with
- * dentry aliases that only differ in case in ->ntfs_lookup() while maintaining
- * a case sensitive dcache. This means that we get the full benefit of dcache
- * speed when the file/directory is looked up with the same case as returned by
- * ->ntfs_readdir() but that a lookup for any other case (or for the short file
- * name) will not find anything in dcache and will enter ->ntfs_lookup()
- * instead, where we search the directory for a fully matching file name
- * (including case) and if that is not found, we search for a file name that
- * matches with different case and if that has non-POSIX semantics we return
- * that. We actually do only one search (case sensitive) and keep tabs on
- * whether we have found a case insensitive match in the process.
- *
- * To simplify matters for us, we do not treat the short vs long filenames as
- * two hard links but instead if the lookup matches a short filename, we
- * return the dentry for the corresponding long filename instead.
- *
- * There are three cases we need to distinguish here:
- *
- * 1) @dent perfectly matches (i.e. including case) a directory entry with a
- *    file name in the WIN32 or POSIX namespaces. In this case
- *    ntfs_lookup_inode_by_name() will return with name set to NULL and we
- *    just d_splice_alias() @dent.
- * 2) @dent matches (not including case) a directory entry with a file name in
- *    the WIN32 namespace. In this case ntfs_lookup_inode_by_name() will return
- *    with name set to point to a kmalloc()ed ntfs_name structure containing
- *    the properly cased little endian Unicode name. We convert the name to the
- *    current NLS code page, search if a dentry with this name already exists
- *    and if so return that instead of @dent.  At this point things are
- *    complicated by the possibility of 'disconnected' dentries due to NFS
- *    which we deal with appropriately (see the code comments).  The VFS will
- *    then destroy the old @dent and use the one we returned.  If a dentry is
- *    not found, we allocate a new one, d_splice_alias() it, and return it as
- *    above.
- * 3) @dent matches either perfectly or not (i.e. we don't care about case) a
- *    directory entry with a file name in the DOS namespace. In this case
- *    ntfs_lookup_inode_by_name() will return with name set to point to a
- *    kmalloc()ed ntfs_name structure containing the mft reference (cpu endian)
- *    of the inode. We use the mft reference to read the inode and to find the
- *    file name in the WIN32 namespace corresponding to the matched short file
- *    name. We then convert the name to the current NLS code page, and proceed
- *    searching for a dentry with this name, etc, as in case 2), above.
- *
- * Locking: Caller must hold i_mutex on the directory.
- */
-static struct dentry *ntfs_lookup(struct inode *dir_ino, struct dentry *dent,
-		unsigned int flags)
-{
-	ntfs_volume *vol = NTFS_SB(dir_ino->i_sb);
-	struct inode *dent_inode;
-	ntfschar *uname;
-	ntfs_name *name = NULL;
-	MFT_REF mref;
-	unsigned long dent_ino;
-	int uname_len;
-
-	ntfs_debug("Looking up %pd in directory inode 0x%lx.",
-			dent, dir_ino->i_ino);
-	/* Convert the name of the dentry to Unicode. */
-	uname_len = ntfs_nlstoucs(vol, dent->d_name.name, dent->d_name.len,
-			&uname);
-	if (uname_len < 0) {
-		if (uname_len != -ENAMETOOLONG)
-			ntfs_error(vol->sb, "Failed to convert name to "
-					"Unicode.");
-		return ERR_PTR(uname_len);
-	}
-	mref = ntfs_lookup_inode_by_name(NTFS_I(dir_ino), uname, uname_len,
-			&name);
-	kmem_cache_free(ntfs_name_cache, uname);
-	if (!IS_ERR_MREF(mref)) {
-		dent_ino = MREF(mref);
-		ntfs_debug("Found inode 0x%lx. Calling ntfs_iget.", dent_ino);
-		dent_inode = ntfs_iget(vol->sb, dent_ino);
-		if (!IS_ERR(dent_inode)) {
-			/* Consistency check. */
-			if (is_bad_inode(dent_inode) || MSEQNO(mref) ==
-					NTFS_I(dent_inode)->seq_no ||
-					dent_ino == FILE_MFT) {
-				/* Perfect WIN32/POSIX match. -- Case 1. */
-				if (!name) {
-					ntfs_debug("Done.  (Case 1.)");
-					return d_splice_alias(dent_inode, dent);
-				}
-				/*
-				 * We are too indented.  Handle imperfect
-				 * matches and short file names further below.
-				 */
-				goto handle_name;
-			}
-			ntfs_error(vol->sb, "Found stale reference to inode "
-					"0x%lx (reference sequence number = "
-					"0x%x, inode sequence number = 0x%x), "
-					"returning -EIO. Run chkdsk.",
-					dent_ino, MSEQNO(mref),
-					NTFS_I(dent_inode)->seq_no);
-			iput(dent_inode);
-			dent_inode = ERR_PTR(-EIO);
-		} else
-			ntfs_error(vol->sb, "ntfs_iget(0x%lx) failed with "
-					"error code %li.", dent_ino,
-					PTR_ERR(dent_inode));
-		kfree(name);
-		/* Return the error code. */
-		return ERR_CAST(dent_inode);
-	}
-	/* It is guaranteed that @name is no longer allocated at this point. */
-	if (MREF_ERR(mref) == -ENOENT) {
-		ntfs_debug("Entry was not found, adding negative dentry.");
-		/* The dcache will handle negative entries. */
-		d_add(dent, NULL);
-		ntfs_debug("Done.");
-		return NULL;
-	}
-	ntfs_error(vol->sb, "ntfs_lookup_ino_by_name() failed with error "
-			"code %i.", -MREF_ERR(mref));
-	return ERR_PTR(MREF_ERR(mref));
-	// TODO: Consider moving this lot to a separate function! (AIA)
-handle_name:
-   {
-	MFT_RECORD *m;
-	ntfs_attr_search_ctx *ctx;
-	ntfs_inode *ni = NTFS_I(dent_inode);
-	int err;
-	struct qstr nls_name;
-
-	nls_name.name = NULL;
-	if (name->type != FILE_NAME_DOS) {			/* Case 2. */
-		ntfs_debug("Case 2.");
-		nls_name.len = (unsigned)ntfs_ucstonls(vol,
-				(ntfschar*)&name->name, name->len,
-				(unsigned char**)&nls_name.name, 0);
-		kfree(name);
-	} else /* if (name->type == FILE_NAME_DOS) */ {		/* Case 3. */
-		FILE_NAME_ATTR *fn;
-
-		ntfs_debug("Case 3.");
-		kfree(name);
-
-		/* Find the WIN32 name corresponding to the matched DOS name. */
-		ni = NTFS_I(dent_inode);
-		m = map_mft_record(ni);
-		if (IS_ERR(m)) {
-			err = PTR_ERR(m);
-			m = NULL;
-			ctx = NULL;
-			goto err_out;
-		}
-		ctx = ntfs_attr_get_search_ctx(ni, m);
-		if (unlikely(!ctx)) {
-			err = -ENOMEM;
-			goto err_out;
-		}
-		do {
-			ATTR_RECORD *a;
-			u32 val_len;
-
-			err = ntfs_attr_lookup(AT_FILE_NAME, NULL, 0, 0, 0,
-					NULL, 0, ctx);
-			if (unlikely(err)) {
-				ntfs_error(vol->sb, "Inode corrupt: No WIN32 "
-						"namespace counterpart to DOS "
-						"file name. Run chkdsk.");
-				if (err == -ENOENT)
-					err = -EIO;
-				goto err_out;
-			}
-			/* Consistency checks. */
-			a = ctx->attr;
-			if (a->non_resident || a->flags)
-				goto eio_err_out;
-			val_len = le32_to_cpu(a->data.resident.value_length);
-			if (le16_to_cpu(a->data.resident.value_offset) +
-					val_len > le32_to_cpu(a->length))
-				goto eio_err_out;
-			fn = (FILE_NAME_ATTR*)((u8*)ctx->attr + le16_to_cpu(
-					ctx->attr->data.resident.value_offset));
-			if ((u32)(fn->file_name_length * sizeof(ntfschar) +
-					sizeof(FILE_NAME_ATTR)) > val_len)
-				goto eio_err_out;
-		} while (fn->file_name_type != FILE_NAME_WIN32);
-
-		/* Convert the found WIN32 name to current NLS code page. */
-		nls_name.len = (unsigned)ntfs_ucstonls(vol,
-				(ntfschar*)&fn->file_name, fn->file_name_length,
-				(unsigned char**)&nls_name.name, 0);
-
-		ntfs_attr_put_search_ctx(ctx);
-		unmap_mft_record(ni);
-	}
-	m = NULL;
-	ctx = NULL;
-
-	/* Check if a conversion error occurred. */
-	if ((signed)nls_name.len < 0) {
-		err = (signed)nls_name.len;
-		goto err_out;
-	}
-	nls_name.hash = full_name_hash(dent, nls_name.name, nls_name.len);
-
-	dent = d_add_ci(dent, dent_inode, &nls_name);
-	kfree(nls_name.name);
-	return dent;
-
-eio_err_out:
-	ntfs_error(vol->sb, "Illegal file name attribute. Run chkdsk.");
-	err = -EIO;
-err_out:
-	if (ctx)
-		ntfs_attr_put_search_ctx(ctx);
-	if (m)
-		unmap_mft_record(ni);
-	iput(dent_inode);
-	ntfs_error(vol->sb, "Failed, returning error code %i.", err);
-	return ERR_PTR(err);
-   }
-}
-
-/*
- * Inode operations for directories.
- */
-const struct inode_operations ntfs_dir_inode_ops = {
-	.lookup	= ntfs_lookup,	/* VFS: Lookup directory. */
-};
-
-/**
- * ntfs_get_parent - find the dentry of the parent of a given directory dentry
- * @child_dent:		dentry of the directory whose parent directory to find
- *
- * Find the dentry for the parent directory of the directory specified by the
- * dentry @child_dent.  This function is called from
- * fs/exportfs/expfs.c::find_exported_dentry() which in turn is called from the
- * default ->decode_fh() which is export_decode_fh() in the same file.
- *
- * The code is based on the ext3 ->get_parent() implementation found in
- * fs/ext3/namei.c::ext3_get_parent().
- *
- * Note: ntfs_get_parent() is called with @d_inode(child_dent)->i_mutex down.
- *
- * Return the dentry of the parent directory on success or the error code on
- * error (IS_ERR() is true).
- */
-static struct dentry *ntfs_get_parent(struct dentry *child_dent)
-{
-	struct inode *vi = d_inode(child_dent);
-	ntfs_inode *ni = NTFS_I(vi);
-	MFT_RECORD *mrec;
-	ntfs_attr_search_ctx *ctx;
-	ATTR_RECORD *attr;
-	FILE_NAME_ATTR *fn;
-	unsigned long parent_ino;
-	int err;
-
-	ntfs_debug("Entering for inode 0x%lx.", vi->i_ino);
-	/* Get the mft record of the inode belonging to the child dentry. */
-	mrec = map_mft_record(ni);
-	if (IS_ERR(mrec))
-		return ERR_CAST(mrec);
-	/* Find the first file name attribute in the mft record. */
-	ctx = ntfs_attr_get_search_ctx(ni, mrec);
-	if (unlikely(!ctx)) {
-		unmap_mft_record(ni);
-		return ERR_PTR(-ENOMEM);
-	}
-try_next:
-	err = ntfs_attr_lookup(AT_FILE_NAME, NULL, 0, CASE_SENSITIVE, 0, NULL,
-			0, ctx);
-	if (unlikely(err)) {
-		ntfs_attr_put_search_ctx(ctx);
-		unmap_mft_record(ni);
-		if (err == -ENOENT)
-			ntfs_error(vi->i_sb, "Inode 0x%lx does not have a "
-					"file name attribute.  Run chkdsk.",
-					vi->i_ino);
-		return ERR_PTR(err);
-	}
-	attr = ctx->attr;
-	if (unlikely(attr->non_resident))
-		goto try_next;
-	fn = (FILE_NAME_ATTR *)((u8 *)attr +
-			le16_to_cpu(attr->data.resident.value_offset));
-	if (unlikely((u8 *)fn + le32_to_cpu(attr->data.resident.value_length) >
-			(u8*)attr + le32_to_cpu(attr->length)))
-		goto try_next;
-	/* Get the inode number of the parent directory. */
-	parent_ino = MREF_LE(fn->parent_directory);
-	/* Release the search context and the mft record of the child. */
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(ni);
-
-	return d_obtain_alias(ntfs_iget(vi->i_sb, parent_ino));
-}
-
-static struct inode *ntfs_nfs_get_inode(struct super_block *sb,
-		u64 ino, u32 generation)
-{
-	struct inode *inode;
-
-	inode = ntfs_iget(sb, ino);
-	if (!IS_ERR(inode)) {
-		if (is_bad_inode(inode) || inode->i_generation != generation) {
-			iput(inode);
-			inode = ERR_PTR(-ESTALE);
-		}
-	}
-
-	return inode;
-}
-
-static struct dentry *ntfs_fh_to_dentry(struct super_block *sb, struct fid *fid,
-		int fh_len, int fh_type)
-{
-	return generic_fh_to_dentry(sb, fid, fh_len, fh_type,
-				    ntfs_nfs_get_inode);
-}
-
-static struct dentry *ntfs_fh_to_parent(struct super_block *sb, struct fid *fid,
-		int fh_len, int fh_type)
-{
-	return generic_fh_to_parent(sb, fid, fh_len, fh_type,
-				    ntfs_nfs_get_inode);
-}
-
-/*
- * Export operations allowing NFS exporting of mounted NTFS partitions.
- *
- * We use the default ->encode_fh() for now.  Note that they
- * use 32 bits to store the inode number which is an unsigned long so on 64-bit
- * architectures is usually 64 bits so it would all fail horribly on huge
- * volumes.  I guess we need to define our own encode and decode fh functions
- * that store 64-bit inode numbers at some point but for now we will ignore the
- * problem...
- *
- * We also use the default ->get_name() helper (used by ->decode_fh() via
- * fs/exportfs/expfs.c::find_exported_dentry()) as that is completely fs
- * independent.
- *
- * The default ->get_parent() just returns -EACCES so we have to provide our
- * own and the default ->get_dentry() is incompatible with NTFS due to not
- * allowing the inode number 0 which is used in NTFS for the system file $MFT
- * and due to using iget() whereas NTFS needs ntfs_iget().
- */
-const struct export_operations ntfs_export_ops = {
-	.encode_fh	= generic_encode_ino32_fh,
-	.get_parent	= ntfs_get_parent,	/* Find the parent of a given
-						   directory. */
-	.fh_to_dentry	= ntfs_fh_to_dentry,
-	.fh_to_parent	= ntfs_fh_to_parent,
-};
diff --git a/fs/ntfs/ntfs.h b/fs/ntfs/ntfs.h
deleted file mode 100644
index e81376ea91..0000000000
--- a/fs/ntfs/ntfs.h
+++ /dev/null
@@ -1,150 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * ntfs.h - Defines for NTFS Linux kernel driver.
- *
- * Copyright (c) 2001-2014 Anton Altaparmakov and Tuxera Inc.
- * Copyright (C) 2002 Richard Russon
- */
-
-#ifndef _LINUX_NTFS_H
-#define _LINUX_NTFS_H
-
-#include <linux/stddef.h>
-#include <linux/kernel.h>
-#include <linux/module.h>
-#include <linux/compiler.h>
-#include <linux/fs.h>
-#include <linux/nls.h>
-#include <linux/smp.h>
-#include <linux/pagemap.h>
-
-#include "types.h"
-#include "volume.h"
-#include "layout.h"
-
-typedef enum {
-	NTFS_BLOCK_SIZE		= 512,
-	NTFS_BLOCK_SIZE_BITS	= 9,
-	NTFS_SB_MAGIC		= 0x5346544e,	/* 'NTFS' */
-	NTFS_MAX_NAME_LEN	= 255,
-	NTFS_MAX_ATTR_NAME_LEN	= 255,
-	NTFS_MAX_CLUSTER_SIZE	= 64 * 1024,	/* 64kiB */
-	NTFS_MAX_PAGES_PER_CLUSTER = NTFS_MAX_CLUSTER_SIZE / PAGE_SIZE,
-} NTFS_CONSTANTS;
-
-/* Global variables. */
-
-/* Slab caches (from super.c). */
-extern struct kmem_cache *ntfs_name_cache;
-extern struct kmem_cache *ntfs_inode_cache;
-extern struct kmem_cache *ntfs_big_inode_cache;
-extern struct kmem_cache *ntfs_attr_ctx_cache;
-extern struct kmem_cache *ntfs_index_ctx_cache;
-
-/* The various operations structs defined throughout the driver files. */
-extern const struct address_space_operations ntfs_normal_aops;
-extern const struct address_space_operations ntfs_compressed_aops;
-extern const struct address_space_operations ntfs_mst_aops;
-
-extern const struct  file_operations ntfs_file_ops;
-extern const struct inode_operations ntfs_file_inode_ops;
-
-extern const struct  file_operations ntfs_dir_ops;
-extern const struct inode_operations ntfs_dir_inode_ops;
-
-extern const struct  file_operations ntfs_empty_file_ops;
-extern const struct inode_operations ntfs_empty_inode_ops;
-
-extern const struct export_operations ntfs_export_ops;
-
-/**
- * NTFS_SB - return the ntfs volume given a vfs super block
- * @sb:		VFS super block
- *
- * NTFS_SB() returns the ntfs volume associated with the VFS super block @sb.
- */
-static inline ntfs_volume *NTFS_SB(struct super_block *sb)
-{
-	return sb->s_fs_info;
-}
-
-/* Declarations of functions and global variables. */
-
-/* From fs/ntfs/compress.c */
-extern int ntfs_read_compressed_block(struct page *page);
-extern int allocate_compression_buffers(void);
-extern void free_compression_buffers(void);
-
-/* From fs/ntfs/super.c */
-#define default_upcase_len 0x10000
-extern struct mutex ntfs_lock;
-
-typedef struct {
-	int val;
-	char *str;
-} option_t;
-extern const option_t on_errors_arr[];
-
-/* From fs/ntfs/mst.c */
-extern int post_read_mst_fixup(NTFS_RECORD *b, const u32 size);
-extern int pre_write_mst_fixup(NTFS_RECORD *b, const u32 size);
-extern void post_write_mst_fixup(NTFS_RECORD *b);
-
-/* From fs/ntfs/unistr.c */
-extern bool ntfs_are_names_equal(const ntfschar *s1, size_t s1_len,
-		const ntfschar *s2, size_t s2_len,
-		const IGNORE_CASE_BOOL ic,
-		const ntfschar *upcase, const u32 upcase_size);
-extern int ntfs_collate_names(const ntfschar *name1, const u32 name1_len,
-		const ntfschar *name2, const u32 name2_len,
-		const int err_val, const IGNORE_CASE_BOOL ic,
-		const ntfschar *upcase, const u32 upcase_len);
-extern int ntfs_ucsncmp(const ntfschar *s1, const ntfschar *s2, size_t n);
-extern int ntfs_ucsncasecmp(const ntfschar *s1, const ntfschar *s2, size_t n,
-		const ntfschar *upcase, const u32 upcase_size);
-extern void ntfs_upcase_name(ntfschar *name, u32 name_len,
-		const ntfschar *upcase, const u32 upcase_len);
-extern void ntfs_file_upcase_value(FILE_NAME_ATTR *file_name_attr,
-		const ntfschar *upcase, const u32 upcase_len);
-extern int ntfs_file_compare_values(FILE_NAME_ATTR *file_name_attr1,
-		FILE_NAME_ATTR *file_name_attr2,
-		const int err_val, const IGNORE_CASE_BOOL ic,
-		const ntfschar *upcase, const u32 upcase_len);
-extern int ntfs_nlstoucs(const ntfs_volume *vol, const char *ins,
-		const int ins_len, ntfschar **outs);
-extern int ntfs_ucstonls(const ntfs_volume *vol, const ntfschar *ins,
-		const int ins_len, unsigned char **outs, int outs_len);
-
-/* From fs/ntfs/upcase.c */
-extern ntfschar *generate_default_upcase(void);
-
-static inline int ntfs_ffs(int x)
-{
-	int r = 1;
-
-	if (!x)
-		return 0;
-	if (!(x & 0xffff)) {
-		x >>= 16;
-		r += 16;
-	}
-	if (!(x & 0xff)) {
-		x >>= 8;
-		r += 8;
-	}
-	if (!(x & 0xf)) {
-		x >>= 4;
-		r += 4;
-	}
-	if (!(x & 3)) {
-		x >>= 2;
-		r += 2;
-	}
-	if (!(x & 1)) {
-		x >>= 1;
-		r += 1;
-	}
-	return r;
-}
-
-#endif /* _LINUX_NTFS_H */
diff --git a/fs/ntfs/quota.c b/fs/ntfs/quota.c
deleted file mode 100644
index 9160480222..0000000000
--- a/fs/ntfs/quota.c
+++ /dev/null
@@ -1,103 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * quota.c - NTFS kernel quota ($Quota) handling.  Part of the Linux-NTFS
- *	     project.
- *
- * Copyright (c) 2004 Anton Altaparmakov
- */
-
-#ifdef NTFS_RW
-
-#include "index.h"
-#include "quota.h"
-#include "debug.h"
-#include "ntfs.h"
-
-/**
- * ntfs_mark_quotas_out_of_date - mark the quotas out of date on an ntfs volume
- * @vol:	ntfs volume on which to mark the quotas out of date
- *
- * Mark the quotas out of date on the ntfs volume @vol and return 'true' on
- * success and 'false' on error.
- */
-bool ntfs_mark_quotas_out_of_date(ntfs_volume *vol)
-{
-	ntfs_index_context *ictx;
-	QUOTA_CONTROL_ENTRY *qce;
-	const le32 qid = QUOTA_DEFAULTS_ID;
-	int err;
-
-	ntfs_debug("Entering.");
-	if (NVolQuotaOutOfDate(vol))
-		goto done;
-	if (!vol->quota_ino || !vol->quota_q_ino) {
-		ntfs_error(vol->sb, "Quota inodes are not open.");
-		return false;
-	}
-	inode_lock(vol->quota_q_ino);
-	ictx = ntfs_index_ctx_get(NTFS_I(vol->quota_q_ino));
-	if (!ictx) {
-		ntfs_error(vol->sb, "Failed to get index context.");
-		goto err_out;
-	}
-	err = ntfs_index_lookup(&qid, sizeof(qid), ictx);
-	if (err) {
-		if (err == -ENOENT)
-			ntfs_error(vol->sb, "Quota defaults entry is not "
-					"present.");
-		else
-			ntfs_error(vol->sb, "Lookup of quota defaults entry "
-					"failed.");
-		goto err_out;
-	}
-	if (ictx->data_len < offsetof(QUOTA_CONTROL_ENTRY, sid)) {
-		ntfs_error(vol->sb, "Quota defaults entry size is invalid.  "
-				"Run chkdsk.");
-		goto err_out;
-	}
-	qce = (QUOTA_CONTROL_ENTRY*)ictx->data;
-	if (le32_to_cpu(qce->version) != QUOTA_VERSION) {
-		ntfs_error(vol->sb, "Quota defaults entry version 0x%x is not "
-				"supported.", le32_to_cpu(qce->version));
-		goto err_out;
-	}
-	ntfs_debug("Quota defaults flags = 0x%x.", le32_to_cpu(qce->flags));
-	/* If quotas are already marked out of date, no need to do anything. */
-	if (qce->flags & QUOTA_FLAG_OUT_OF_DATE)
-		goto set_done;
-	/*
-	 * If quota tracking is neither requested, nor enabled and there are no
-	 * pending deletes, no need to mark the quotas out of date.
-	 */
-	if (!(qce->flags & (QUOTA_FLAG_TRACKING_ENABLED |
-			QUOTA_FLAG_TRACKING_REQUESTED |
-			QUOTA_FLAG_PENDING_DELETES)))
-		goto set_done;
-	/*
-	 * Set the QUOTA_FLAG_OUT_OF_DATE bit thus marking quotas out of date.
-	 * This is verified on WinXP to be sufficient to cause windows to
-	 * rescan the volume on boot and update all quota entries.
-	 */
-	qce->flags |= QUOTA_FLAG_OUT_OF_DATE;
-	/* Ensure the modified flags are written to disk. */
-	ntfs_index_entry_flush_dcache_page(ictx);
-	ntfs_index_entry_mark_dirty(ictx);
-set_done:
-	ntfs_index_ctx_put(ictx);
-	inode_unlock(vol->quota_q_ino);
-	/*
-	 * We set the flag so we do not try to mark the quotas out of date
-	 * again on remount.
-	 */
-	NVolSetQuotaOutOfDate(vol);
-done:
-	ntfs_debug("Done.");
-	return true;
-err_out:
-	if (ictx)
-		ntfs_index_ctx_put(ictx);
-	inode_unlock(vol->quota_q_ino);
-	return false;
-}
-
-#endif /* NTFS_RW */
diff --git a/fs/ntfs/quota.h b/fs/ntfs/quota.h
deleted file mode 100644
index fe3132a3d6..0000000000
--- a/fs/ntfs/quota.h
+++ /dev/null
@@ -1,21 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * quota.h - Defines for NTFS kernel quota ($Quota) handling.  Part of the
- *	     Linux-NTFS project.
- *
- * Copyright (c) 2004 Anton Altaparmakov
- */
-
-#ifndef _LINUX_NTFS_QUOTA_H
-#define _LINUX_NTFS_QUOTA_H
-
-#ifdef NTFS_RW
-
-#include "types.h"
-#include "volume.h"
-
-extern bool ntfs_mark_quotas_out_of_date(ntfs_volume *vol);
-
-#endif /* NTFS_RW */
-
-#endif /* _LINUX_NTFS_QUOTA_H */
diff --git a/fs/ntfs/runlist.c b/fs/ntfs/runlist.c
deleted file mode 100644
index 0d448e9881..0000000000
--- a/fs/ntfs/runlist.c
+++ /dev/null
@@ -1,1893 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * runlist.c - NTFS runlist handling code.  Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2007 Anton Altaparmakov
- * Copyright (c) 2002-2005 Richard Russon
- */
-
-#include "debug.h"
-#include "dir.h"
-#include "endian.h"
-#include "malloc.h"
-#include "ntfs.h"
-
-/**
- * ntfs_rl_mm - runlist memmove
- *
- * It is up to the caller to serialize access to the runlist @base.
- */
-static inline void ntfs_rl_mm(runlist_element *base, int dst, int src,
-		int size)
-{
-	if (likely((dst != src) && (size > 0)))
-		memmove(base + dst, base + src, size * sizeof(*base));
-}
-
-/**
- * ntfs_rl_mc - runlist memory copy
- *
- * It is up to the caller to serialize access to the runlists @dstbase and
- * @srcbase.
- */
-static inline void ntfs_rl_mc(runlist_element *dstbase, int dst,
-		runlist_element *srcbase, int src, int size)
-{
-	if (likely(size > 0))
-		memcpy(dstbase + dst, srcbase + src, size * sizeof(*dstbase));
-}
-
-/**
- * ntfs_rl_realloc - Reallocate memory for runlists
- * @rl:		original runlist
- * @old_size:	number of runlist elements in the original runlist @rl
- * @new_size:	number of runlist elements we need space for
- *
- * As the runlists grow, more memory will be required.  To prevent the
- * kernel having to allocate and reallocate large numbers of small bits of
- * memory, this function returns an entire page of memory.
- *
- * It is up to the caller to serialize access to the runlist @rl.
- *
- * N.B.  If the new allocation doesn't require a different number of pages in
- *       memory, the function will return the original pointer.
- *
- * On success, return a pointer to the newly allocated, or recycled, memory.
- * On error, return -errno. The following error codes are defined:
- *	-ENOMEM	- Not enough memory to allocate runlist array.
- *	-EINVAL	- Invalid parameters were passed in.
- */
-static inline runlist_element *ntfs_rl_realloc(runlist_element *rl,
-		int old_size, int new_size)
-{
-	runlist_element *new_rl;
-
-	old_size = PAGE_ALIGN(old_size * sizeof(*rl));
-	new_size = PAGE_ALIGN(new_size * sizeof(*rl));
-	if (old_size == new_size)
-		return rl;
-
-	new_rl = ntfs_malloc_nofs(new_size);
-	if (unlikely(!new_rl))
-		return ERR_PTR(-ENOMEM);
-
-	if (likely(rl != NULL)) {
-		if (unlikely(old_size > new_size))
-			old_size = new_size;
-		memcpy(new_rl, rl, old_size);
-		ntfs_free(rl);
-	}
-	return new_rl;
-}
-
-/**
- * ntfs_rl_realloc_nofail - Reallocate memory for runlists
- * @rl:		original runlist
- * @old_size:	number of runlist elements in the original runlist @rl
- * @new_size:	number of runlist elements we need space for
- *
- * As the runlists grow, more memory will be required.  To prevent the
- * kernel having to allocate and reallocate large numbers of small bits of
- * memory, this function returns an entire page of memory.
- *
- * This function guarantees that the allocation will succeed.  It will sleep
- * for as long as it takes to complete the allocation.
- *
- * It is up to the caller to serialize access to the runlist @rl.
- *
- * N.B.  If the new allocation doesn't require a different number of pages in
- *       memory, the function will return the original pointer.
- *
- * On success, return a pointer to the newly allocated, or recycled, memory.
- * On error, return -errno. The following error codes are defined:
- *	-ENOMEM	- Not enough memory to allocate runlist array.
- *	-EINVAL	- Invalid parameters were passed in.
- */
-static inline runlist_element *ntfs_rl_realloc_nofail(runlist_element *rl,
-		int old_size, int new_size)
-{
-	runlist_element *new_rl;
-
-	old_size = PAGE_ALIGN(old_size * sizeof(*rl));
-	new_size = PAGE_ALIGN(new_size * sizeof(*rl));
-	if (old_size == new_size)
-		return rl;
-
-	new_rl = ntfs_malloc_nofs_nofail(new_size);
-	BUG_ON(!new_rl);
-
-	if (likely(rl != NULL)) {
-		if (unlikely(old_size > new_size))
-			old_size = new_size;
-		memcpy(new_rl, rl, old_size);
-		ntfs_free(rl);
-	}
-	return new_rl;
-}
-
-/**
- * ntfs_are_rl_mergeable - test if two runlists can be joined together
- * @dst:	original runlist
- * @src:	new runlist to test for mergeability with @dst
- *
- * Test if two runlists can be joined together. For this, their VCNs and LCNs
- * must be adjacent.
- *
- * It is up to the caller to serialize access to the runlists @dst and @src.
- *
- * Return: true   Success, the runlists can be merged.
- *	   false  Failure, the runlists cannot be merged.
- */
-static inline bool ntfs_are_rl_mergeable(runlist_element *dst,
-		runlist_element *src)
-{
-	BUG_ON(!dst);
-	BUG_ON(!src);
-
-	/* We can merge unmapped regions even if they are misaligned. */
-	if ((dst->lcn == LCN_RL_NOT_MAPPED) && (src->lcn == LCN_RL_NOT_MAPPED))
-		return true;
-	/* If the runs are misaligned, we cannot merge them. */
-	if ((dst->vcn + dst->length) != src->vcn)
-		return false;
-	/* If both runs are non-sparse and contiguous, we can merge them. */
-	if ((dst->lcn >= 0) && (src->lcn >= 0) &&
-			((dst->lcn + dst->length) == src->lcn))
-		return true;
-	/* If we are merging two holes, we can merge them. */
-	if ((dst->lcn == LCN_HOLE) && (src->lcn == LCN_HOLE))
-		return true;
-	/* Cannot merge. */
-	return false;
-}
-
-/**
- * __ntfs_rl_merge - merge two runlists without testing if they can be merged
- * @dst:	original, destination runlist
- * @src:	new runlist to merge with @dst
- *
- * Merge the two runlists, writing into the destination runlist @dst. The
- * caller must make sure the runlists can be merged or this will corrupt the
- * destination runlist.
- *
- * It is up to the caller to serialize access to the runlists @dst and @src.
- */
-static inline void __ntfs_rl_merge(runlist_element *dst, runlist_element *src)
-{
-	dst->length += src->length;
-}
-
-/**
- * ntfs_rl_append - append a runlist after a given element
- * @dst:	original runlist to be worked on
- * @dsize:	number of elements in @dst (including end marker)
- * @src:	runlist to be inserted into @dst
- * @ssize:	number of elements in @src (excluding end marker)
- * @loc:	append the new runlist @src after this element in @dst
- *
- * Append the runlist @src after element @loc in @dst.  Merge the right end of
- * the new runlist, if necessary. Adjust the size of the hole before the
- * appended runlist.
- *
- * It is up to the caller to serialize access to the runlists @dst and @src.
- *
- * On success, return a pointer to the new, combined, runlist. Note, both
- * runlists @dst and @src are deallocated before returning so you cannot use
- * the pointers for anything any more. (Strictly speaking the returned runlist
- * may be the same as @dst but this is irrelevant.)
- *
- * On error, return -errno. Both runlists are left unmodified. The following
- * error codes are defined:
- *	-ENOMEM	- Not enough memory to allocate runlist array.
- *	-EINVAL	- Invalid parameters were passed in.
- */
-static inline runlist_element *ntfs_rl_append(runlist_element *dst,
-		int dsize, runlist_element *src, int ssize, int loc)
-{
-	bool right = false;	/* Right end of @src needs merging. */
-	int marker;		/* End of the inserted runs. */
-
-	BUG_ON(!dst);
-	BUG_ON(!src);
-
-	/* First, check if the right hand end needs merging. */
-	if ((loc + 1) < dsize)
-		right = ntfs_are_rl_mergeable(src + ssize - 1, dst + loc + 1);
-
-	/* Space required: @dst size + @src size, less one if we merged. */
-	dst = ntfs_rl_realloc(dst, dsize, dsize + ssize - right);
-	if (IS_ERR(dst))
-		return dst;
-	/*
-	 * We are guaranteed to succeed from here so can start modifying the
-	 * original runlists.
-	 */
-
-	/* First, merge the right hand end, if necessary. */
-	if (right)
-		__ntfs_rl_merge(src + ssize - 1, dst + loc + 1);
-
-	/* First run after the @src runs that have been inserted. */
-	marker = loc + ssize + 1;
-
-	/* Move the tail of @dst out of the way, then copy in @src. */
-	ntfs_rl_mm(dst, marker, loc + 1 + right, dsize - (loc + 1 + right));
-	ntfs_rl_mc(dst, loc + 1, src, 0, ssize);
-
-	/* Adjust the size of the preceding hole. */
-	dst[loc].length = dst[loc + 1].vcn - dst[loc].vcn;
-
-	/* We may have changed the length of the file, so fix the end marker */
-	if (dst[marker].lcn == LCN_ENOENT)
-		dst[marker].vcn = dst[marker - 1].vcn + dst[marker - 1].length;
-
-	return dst;
-}
-
-/**
- * ntfs_rl_insert - insert a runlist into another
- * @dst:	original runlist to be worked on
- * @dsize:	number of elements in @dst (including end marker)
- * @src:	new runlist to be inserted
- * @ssize:	number of elements in @src (excluding end marker)
- * @loc:	insert the new runlist @src before this element in @dst
- *
- * Insert the runlist @src before element @loc in the runlist @dst. Merge the
- * left end of the new runlist, if necessary. Adjust the size of the hole
- * after the inserted runlist.
- *
- * It is up to the caller to serialize access to the runlists @dst and @src.
- *
- * On success, return a pointer to the new, combined, runlist. Note, both
- * runlists @dst and @src are deallocated before returning so you cannot use
- * the pointers for anything any more. (Strictly speaking the returned runlist
- * may be the same as @dst but this is irrelevant.)
- *
- * On error, return -errno. Both runlists are left unmodified. The following
- * error codes are defined:
- *	-ENOMEM	- Not enough memory to allocate runlist array.
- *	-EINVAL	- Invalid parameters were passed in.
- */
-static inline runlist_element *ntfs_rl_insert(runlist_element *dst,
-		int dsize, runlist_element *src, int ssize, int loc)
-{
-	bool left = false;	/* Left end of @src needs merging. */
-	bool disc = false;	/* Discontinuity between @dst and @src. */
-	int marker;		/* End of the inserted runs. */
-
-	BUG_ON(!dst);
-	BUG_ON(!src);
-
-	/*
-	 * disc => Discontinuity between the end of @dst and the start of @src.
-	 *	   This means we might need to insert a "not mapped" run.
-	 */
-	if (loc == 0)
-		disc = (src[0].vcn > 0);
-	else {
-		s64 merged_length;
-
-		left = ntfs_are_rl_mergeable(dst + loc - 1, src);
-
-		merged_length = dst[loc - 1].length;
-		if (left)
-			merged_length += src->length;
-
-		disc = (src[0].vcn > dst[loc - 1].vcn + merged_length);
-	}
-	/*
-	 * Space required: @dst size + @src size, less one if we merged, plus
-	 * one if there was a discontinuity.
-	 */
-	dst = ntfs_rl_realloc(dst, dsize, dsize + ssize - left + disc);
-	if (IS_ERR(dst))
-		return dst;
-	/*
-	 * We are guaranteed to succeed from here so can start modifying the
-	 * original runlist.
-	 */
-	if (left)
-		__ntfs_rl_merge(dst + loc - 1, src);
-	/*
-	 * First run after the @src runs that have been inserted.
-	 * Nominally,  @marker equals @loc + @ssize, i.e. location + number of
-	 * runs in @src.  However, if @left, then the first run in @src has
-	 * been merged with one in @dst.  And if @disc, then @dst and @src do
-	 * not meet and we need an extra run to fill the gap.
-	 */
-	marker = loc + ssize - left + disc;
-
-	/* Move the tail of @dst out of the way, then copy in @src. */
-	ntfs_rl_mm(dst, marker, loc, dsize - loc);
-	ntfs_rl_mc(dst, loc + disc, src, left, ssize - left);
-
-	/* Adjust the VCN of the first run after the insertion... */
-	dst[marker].vcn = dst[marker - 1].vcn + dst[marker - 1].length;
-	/* ... and the length. */
-	if (dst[marker].lcn == LCN_HOLE || dst[marker].lcn == LCN_RL_NOT_MAPPED)
-		dst[marker].length = dst[marker + 1].vcn - dst[marker].vcn;
-
-	/* Writing beyond the end of the file and there is a discontinuity. */
-	if (disc) {
-		if (loc > 0) {
-			dst[loc].vcn = dst[loc - 1].vcn + dst[loc - 1].length;
-			dst[loc].length = dst[loc + 1].vcn - dst[loc].vcn;
-		} else {
-			dst[loc].vcn = 0;
-			dst[loc].length = dst[loc + 1].vcn;
-		}
-		dst[loc].lcn = LCN_RL_NOT_MAPPED;
-	}
-	return dst;
-}
-
-/**
- * ntfs_rl_replace - overwrite a runlist element with another runlist
- * @dst:	original runlist to be worked on
- * @dsize:	number of elements in @dst (including end marker)
- * @src:	new runlist to be inserted
- * @ssize:	number of elements in @src (excluding end marker)
- * @loc:	index in runlist @dst to overwrite with @src
- *
- * Replace the runlist element @dst at @loc with @src. Merge the left and
- * right ends of the inserted runlist, if necessary.
- *
- * It is up to the caller to serialize access to the runlists @dst and @src.
- *
- * On success, return a pointer to the new, combined, runlist. Note, both
- * runlists @dst and @src are deallocated before returning so you cannot use
- * the pointers for anything any more. (Strictly speaking the returned runlist
- * may be the same as @dst but this is irrelevant.)
- *
- * On error, return -errno. Both runlists are left unmodified. The following
- * error codes are defined:
- *	-ENOMEM	- Not enough memory to allocate runlist array.
- *	-EINVAL	- Invalid parameters were passed in.
- */
-static inline runlist_element *ntfs_rl_replace(runlist_element *dst,
-		int dsize, runlist_element *src, int ssize, int loc)
-{
-	signed delta;
-	bool left = false;	/* Left end of @src needs merging. */
-	bool right = false;	/* Right end of @src needs merging. */
-	int tail;		/* Start of tail of @dst. */
-	int marker;		/* End of the inserted runs. */
-
-	BUG_ON(!dst);
-	BUG_ON(!src);
-
-	/* First, see if the left and right ends need merging. */
-	if ((loc + 1) < dsize)
-		right = ntfs_are_rl_mergeable(src + ssize - 1, dst + loc + 1);
-	if (loc > 0)
-		left = ntfs_are_rl_mergeable(dst + loc - 1, src);
-	/*
-	 * Allocate some space.  We will need less if the left, right, or both
-	 * ends get merged.  The -1 accounts for the run being replaced.
-	 */
-	delta = ssize - 1 - left - right;
-	if (delta > 0) {
-		dst = ntfs_rl_realloc(dst, dsize, dsize + delta);
-		if (IS_ERR(dst))
-			return dst;
-	}
-	/*
-	 * We are guaranteed to succeed from here so can start modifying the
-	 * original runlists.
-	 */
-
-	/* First, merge the left and right ends, if necessary. */
-	if (right)
-		__ntfs_rl_merge(src + ssize - 1, dst + loc + 1);
-	if (left)
-		__ntfs_rl_merge(dst + loc - 1, src);
-	/*
-	 * Offset of the tail of @dst.  This needs to be moved out of the way
-	 * to make space for the runs to be copied from @src, i.e. the first
-	 * run of the tail of @dst.
-	 * Nominally, @tail equals @loc + 1, i.e. location, skipping the
-	 * replaced run.  However, if @right, then one of @dst's runs is
-	 * already merged into @src.
-	 */
-	tail = loc + right + 1;
-	/*
-	 * First run after the @src runs that have been inserted, i.e. where
-	 * the tail of @dst needs to be moved to.
-	 * Nominally, @marker equals @loc + @ssize, i.e. location + number of
-	 * runs in @src.  However, if @left, then the first run in @src has
-	 * been merged with one in @dst.
-	 */
-	marker = loc + ssize - left;
-
-	/* Move the tail of @dst out of the way, then copy in @src. */
-	ntfs_rl_mm(dst, marker, tail, dsize - tail);
-	ntfs_rl_mc(dst, loc, src, left, ssize - left);
-
-	/* We may have changed the length of the file, so fix the end marker. */
-	if (dsize - tail > 0 && dst[marker].lcn == LCN_ENOENT)
-		dst[marker].vcn = dst[marker - 1].vcn + dst[marker - 1].length;
-	return dst;
-}
-
-/**
- * ntfs_rl_split - insert a runlist into the centre of a hole
- * @dst:	original runlist to be worked on
- * @dsize:	number of elements in @dst (including end marker)
- * @src:	new runlist to be inserted
- * @ssize:	number of elements in @src (excluding end marker)
- * @loc:	index in runlist @dst at which to split and insert @src
- *
- * Split the runlist @dst at @loc into two and insert @new in between the two
- * fragments. No merging of runlists is necessary. Adjust the size of the
- * holes either side.
- *
- * It is up to the caller to serialize access to the runlists @dst and @src.
- *
- * On success, return a pointer to the new, combined, runlist. Note, both
- * runlists @dst and @src are deallocated before returning so you cannot use
- * the pointers for anything any more. (Strictly speaking the returned runlist
- * may be the same as @dst but this is irrelevant.)
- *
- * On error, return -errno. Both runlists are left unmodified. The following
- * error codes are defined:
- *	-ENOMEM	- Not enough memory to allocate runlist array.
- *	-EINVAL	- Invalid parameters were passed in.
- */
-static inline runlist_element *ntfs_rl_split(runlist_element *dst, int dsize,
-		runlist_element *src, int ssize, int loc)
-{
-	BUG_ON(!dst);
-	BUG_ON(!src);
-
-	/* Space required: @dst size + @src size + one new hole. */
-	dst = ntfs_rl_realloc(dst, dsize, dsize + ssize + 1);
-	if (IS_ERR(dst))
-		return dst;
-	/*
-	 * We are guaranteed to succeed from here so can start modifying the
-	 * original runlists.
-	 */
-
-	/* Move the tail of @dst out of the way, then copy in @src. */
-	ntfs_rl_mm(dst, loc + 1 + ssize, loc, dsize - loc);
-	ntfs_rl_mc(dst, loc + 1, src, 0, ssize);
-
-	/* Adjust the size of the holes either size of @src. */
-	dst[loc].length		= dst[loc+1].vcn       - dst[loc].vcn;
-	dst[loc+ssize+1].vcn    = dst[loc+ssize].vcn   + dst[loc+ssize].length;
-	dst[loc+ssize+1].length = dst[loc+ssize+2].vcn - dst[loc+ssize+1].vcn;
-
-	return dst;
-}
-
-/**
- * ntfs_runlists_merge - merge two runlists into one
- * @drl:	original runlist to be worked on
- * @srl:	new runlist to be merged into @drl
- *
- * First we sanity check the two runlists @srl and @drl to make sure that they
- * are sensible and can be merged. The runlist @srl must be either after the
- * runlist @drl or completely within a hole (or unmapped region) in @drl.
- *
- * It is up to the caller to serialize access to the runlists @drl and @srl.
- *
- * Merging of runlists is necessary in two cases:
- *   1. When attribute lists are used and a further extent is being mapped.
- *   2. When new clusters are allocated to fill a hole or extend a file.
- *
- * There are four possible ways @srl can be merged. It can:
- *	- be inserted at the beginning of a hole,
- *	- split the hole in two and be inserted between the two fragments,
- *	- be appended at the end of a hole, or it can
- *	- replace the whole hole.
- * It can also be appended to the end of the runlist, which is just a variant
- * of the insert case.
- *
- * On success, return a pointer to the new, combined, runlist. Note, both
- * runlists @drl and @srl are deallocated before returning so you cannot use
- * the pointers for anything any more. (Strictly speaking the returned runlist
- * may be the same as @dst but this is irrelevant.)
- *
- * On error, return -errno. Both runlists are left unmodified. The following
- * error codes are defined:
- *	-ENOMEM	- Not enough memory to allocate runlist array.
- *	-EINVAL	- Invalid parameters were passed in.
- *	-ERANGE	- The runlists overlap and cannot be merged.
- */
-runlist_element *ntfs_runlists_merge(runlist_element *drl,
-		runlist_element *srl)
-{
-	int di, si;		/* Current index into @[ds]rl. */
-	int sstart;		/* First index with lcn > LCN_RL_NOT_MAPPED. */
-	int dins;		/* Index into @drl at which to insert @srl. */
-	int dend, send;		/* Last index into @[ds]rl. */
-	int dfinal, sfinal;	/* The last index into @[ds]rl with
-				   lcn >= LCN_HOLE. */
-	int marker = 0;
-	VCN marker_vcn = 0;
-
-#ifdef DEBUG
-	ntfs_debug("dst:");
-	ntfs_debug_dump_runlist(drl);
-	ntfs_debug("src:");
-	ntfs_debug_dump_runlist(srl);
-#endif
-
-	/* Check for silly calling... */
-	if (unlikely(!srl))
-		return drl;
-	if (IS_ERR(srl) || IS_ERR(drl))
-		return ERR_PTR(-EINVAL);
-
-	/* Check for the case where the first mapping is being done now. */
-	if (unlikely(!drl)) {
-		drl = srl;
-		/* Complete the source runlist if necessary. */
-		if (unlikely(drl[0].vcn)) {
-			/* Scan to the end of the source runlist. */
-			for (dend = 0; likely(drl[dend].length); dend++)
-				;
-			dend++;
-			drl = ntfs_rl_realloc(drl, dend, dend + 1);
-			if (IS_ERR(drl))
-				return drl;
-			/* Insert start element at the front of the runlist. */
-			ntfs_rl_mm(drl, 1, 0, dend);
-			drl[0].vcn = 0;
-			drl[0].lcn = LCN_RL_NOT_MAPPED;
-			drl[0].length = drl[1].vcn;
-		}
-		goto finished;
-	}
-
-	si = di = 0;
-
-	/* Skip any unmapped start element(s) in the source runlist. */
-	while (srl[si].length && srl[si].lcn < LCN_HOLE)
-		si++;
-
-	/* Can't have an entirely unmapped source runlist. */
-	BUG_ON(!srl[si].length);
-
-	/* Record the starting points. */
-	sstart = si;
-
-	/*
-	 * Skip forward in @drl until we reach the position where @srl needs to
-	 * be inserted. If we reach the end of @drl, @srl just needs to be
-	 * appended to @drl.
-	 */
-	for (; drl[di].length; di++) {
-		if (drl[di].vcn + drl[di].length > srl[sstart].vcn)
-			break;
-	}
-	dins = di;
-
-	/* Sanity check for illegal overlaps. */
-	if ((drl[di].vcn == srl[si].vcn) && (drl[di].lcn >= 0) &&
-			(srl[si].lcn >= 0)) {
-		ntfs_error(NULL, "Run lists overlap. Cannot merge!");
-		return ERR_PTR(-ERANGE);
-	}
-
-	/* Scan to the end of both runlists in order to know their sizes. */
-	for (send = si; srl[send].length; send++)
-		;
-	for (dend = di; drl[dend].length; dend++)
-		;
-
-	if (srl[send].lcn == LCN_ENOENT)
-		marker_vcn = srl[marker = send].vcn;
-
-	/* Scan to the last element with lcn >= LCN_HOLE. */
-	for (sfinal = send; sfinal >= 0 && srl[sfinal].lcn < LCN_HOLE; sfinal--)
-		;
-	for (dfinal = dend; dfinal >= 0 && drl[dfinal].lcn < LCN_HOLE; dfinal--)
-		;
-
-	{
-	bool start;
-	bool finish;
-	int ds = dend + 1;		/* Number of elements in drl & srl */
-	int ss = sfinal - sstart + 1;
-
-	start  = ((drl[dins].lcn <  LCN_RL_NOT_MAPPED) ||    /* End of file   */
-		  (drl[dins].vcn == srl[sstart].vcn));	     /* Start of hole */
-	finish = ((drl[dins].lcn >= LCN_RL_NOT_MAPPED) &&    /* End of file   */
-		 ((drl[dins].vcn + drl[dins].length) <=      /* End of hole   */
-		  (srl[send - 1].vcn + srl[send - 1].length)));
-
-	/* Or we will lose an end marker. */
-	if (finish && !drl[dins].length)
-		ss++;
-	if (marker && (drl[dins].vcn + drl[dins].length > srl[send - 1].vcn))
-		finish = false;
-#if 0
-	ntfs_debug("dfinal = %i, dend = %i", dfinal, dend);
-	ntfs_debug("sstart = %i, sfinal = %i, send = %i", sstart, sfinal, send);
-	ntfs_debug("start = %i, finish = %i", start, finish);
-	ntfs_debug("ds = %i, ss = %i, dins = %i", ds, ss, dins);
-#endif
-	if (start) {
-		if (finish)
-			drl = ntfs_rl_replace(drl, ds, srl + sstart, ss, dins);
-		else
-			drl = ntfs_rl_insert(drl, ds, srl + sstart, ss, dins);
-	} else {
-		if (finish)
-			drl = ntfs_rl_append(drl, ds, srl + sstart, ss, dins);
-		else
-			drl = ntfs_rl_split(drl, ds, srl + sstart, ss, dins);
-	}
-	if (IS_ERR(drl)) {
-		ntfs_error(NULL, "Merge failed.");
-		return drl;
-	}
-	ntfs_free(srl);
-	if (marker) {
-		ntfs_debug("Triggering marker code.");
-		for (ds = dend; drl[ds].length; ds++)
-			;
-		/* We only need to care if @srl ended after @drl. */
-		if (drl[ds].vcn <= marker_vcn) {
-			int slots = 0;
-
-			if (drl[ds].vcn == marker_vcn) {
-				ntfs_debug("Old marker = 0x%llx, replacing "
-						"with LCN_ENOENT.",
-						(unsigned long long)
-						drl[ds].lcn);
-				drl[ds].lcn = LCN_ENOENT;
-				goto finished;
-			}
-			/*
-			 * We need to create an unmapped runlist element in
-			 * @drl or extend an existing one before adding the
-			 * ENOENT terminator.
-			 */
-			if (drl[ds].lcn == LCN_ENOENT) {
-				ds--;
-				slots = 1;
-			}
-			if (drl[ds].lcn != LCN_RL_NOT_MAPPED) {
-				/* Add an unmapped runlist element. */
-				if (!slots) {
-					drl = ntfs_rl_realloc_nofail(drl, ds,
-							ds + 2);
-					slots = 2;
-				}
-				ds++;
-				/* Need to set vcn if it isn't set already. */
-				if (slots != 1)
-					drl[ds].vcn = drl[ds - 1].vcn +
-							drl[ds - 1].length;
-				drl[ds].lcn = LCN_RL_NOT_MAPPED;
-				/* We now used up a slot. */
-				slots--;
-			}
-			drl[ds].length = marker_vcn - drl[ds].vcn;
-			/* Finally add the ENOENT terminator. */
-			ds++;
-			if (!slots)
-				drl = ntfs_rl_realloc_nofail(drl, ds, ds + 1);
-			drl[ds].vcn = marker_vcn;
-			drl[ds].lcn = LCN_ENOENT;
-			drl[ds].length = (s64)0;
-		}
-	}
-	}
-
-finished:
-	/* The merge was completed successfully. */
-	ntfs_debug("Merged runlist:");
-	ntfs_debug_dump_runlist(drl);
-	return drl;
-}
-
-/**
- * ntfs_mapping_pairs_decompress - convert mapping pairs array to runlist
- * @vol:	ntfs volume on which the attribute resides
- * @attr:	attribute record whose mapping pairs array to decompress
- * @old_rl:	optional runlist in which to insert @attr's runlist
- *
- * It is up to the caller to serialize access to the runlist @old_rl.
- *
- * Decompress the attribute @attr's mapping pairs array into a runlist. On
- * success, return the decompressed runlist.
- *
- * If @old_rl is not NULL, decompressed runlist is inserted into the
- * appropriate place in @old_rl and the resultant, combined runlist is
- * returned. The original @old_rl is deallocated.
- *
- * On error, return -errno. @old_rl is left unmodified in that case.
- *
- * The following error codes are defined:
- *	-ENOMEM	- Not enough memory to allocate runlist array.
- *	-EIO	- Corrupt runlist.
- *	-EINVAL	- Invalid parameters were passed in.
- *	-ERANGE	- The two runlists overlap.
- *
- * FIXME: For now we take the conceptionally simplest approach of creating the
- * new runlist disregarding the already existing one and then splicing the
- * two into one, if that is possible (we check for overlap and discard the new
- * runlist if overlap present before returning ERR_PTR(-ERANGE)).
- */
-runlist_element *ntfs_mapping_pairs_decompress(const ntfs_volume *vol,
-		const ATTR_RECORD *attr, runlist_element *old_rl)
-{
-	VCN vcn;		/* Current vcn. */
-	LCN lcn;		/* Current lcn. */
-	s64 deltaxcn;		/* Change in [vl]cn. */
-	runlist_element *rl;	/* The output runlist. */
-	u8 *buf;		/* Current position in mapping pairs array. */
-	u8 *attr_end;		/* End of attribute. */
-	int rlsize;		/* Size of runlist buffer. */
-	u16 rlpos;		/* Current runlist position in units of
-				   runlist_elements. */
-	u8 b;			/* Current byte offset in buf. */
-
-#ifdef DEBUG
-	/* Make sure attr exists and is non-resident. */
-	if (!attr || !attr->non_resident || sle64_to_cpu(
-			attr->data.non_resident.lowest_vcn) < (VCN)0) {
-		ntfs_error(vol->sb, "Invalid arguments.");
-		return ERR_PTR(-EINVAL);
-	}
-#endif
-	/* Start at vcn = lowest_vcn and lcn 0. */
-	vcn = sle64_to_cpu(attr->data.non_resident.lowest_vcn);
-	lcn = 0;
-	/* Get start of the mapping pairs array. */
-	buf = (u8*)attr + le16_to_cpu(
-			attr->data.non_resident.mapping_pairs_offset);
-	attr_end = (u8*)attr + le32_to_cpu(attr->length);
-	if (unlikely(buf < (u8*)attr || buf > attr_end)) {
-		ntfs_error(vol->sb, "Corrupt attribute.");
-		return ERR_PTR(-EIO);
-	}
-	/* If the mapping pairs array is valid but empty, nothing to do. */
-	if (!vcn && !*buf)
-		return old_rl;
-	/* Current position in runlist array. */
-	rlpos = 0;
-	/* Allocate first page and set current runlist size to one page. */
-	rl = ntfs_malloc_nofs(rlsize = PAGE_SIZE);
-	if (unlikely(!rl))
-		return ERR_PTR(-ENOMEM);
-	/* Insert unmapped starting element if necessary. */
-	if (vcn) {
-		rl->vcn = 0;
-		rl->lcn = LCN_RL_NOT_MAPPED;
-		rl->length = vcn;
-		rlpos++;
-	}
-	while (buf < attr_end && *buf) {
-		/*
-		 * Allocate more memory if needed, including space for the
-		 * not-mapped and terminator elements. ntfs_malloc_nofs()
-		 * operates on whole pages only.
-		 */
-		if (((rlpos + 3) * sizeof(*old_rl)) > rlsize) {
-			runlist_element *rl2;
-
-			rl2 = ntfs_malloc_nofs(rlsize + (int)PAGE_SIZE);
-			if (unlikely(!rl2)) {
-				ntfs_free(rl);
-				return ERR_PTR(-ENOMEM);
-			}
-			memcpy(rl2, rl, rlsize);
-			ntfs_free(rl);
-			rl = rl2;
-			rlsize += PAGE_SIZE;
-		}
-		/* Enter the current vcn into the current runlist element. */
-		rl[rlpos].vcn = vcn;
-		/*
-		 * Get the change in vcn, i.e. the run length in clusters.
-		 * Doing it this way ensures that we signextend negative values.
-		 * A negative run length doesn't make any sense, but hey, I
-		 * didn't make up the NTFS specs and Windows NT4 treats the run
-		 * length as a signed value so that's how it is...
-		 */
-		b = *buf & 0xf;
-		if (b) {
-			if (unlikely(buf + b > attr_end))
-				goto io_error;
-			for (deltaxcn = (s8)buf[b--]; b; b--)
-				deltaxcn = (deltaxcn << 8) + buf[b];
-		} else { /* The length entry is compulsory. */
-			ntfs_error(vol->sb, "Missing length entry in mapping "
-					"pairs array.");
-			deltaxcn = (s64)-1;
-		}
-		/*
-		 * Assume a negative length to indicate data corruption and
-		 * hence clean-up and return NULL.
-		 */
-		if (unlikely(deltaxcn < 0)) {
-			ntfs_error(vol->sb, "Invalid length in mapping pairs "
-					"array.");
-			goto err_out;
-		}
-		/*
-		 * Enter the current run length into the current runlist
-		 * element.
-		 */
-		rl[rlpos].length = deltaxcn;
-		/* Increment the current vcn by the current run length. */
-		vcn += deltaxcn;
-		/*
-		 * There might be no lcn change at all, as is the case for
-		 * sparse clusters on NTFS 3.0+, in which case we set the lcn
-		 * to LCN_HOLE.
-		 */
-		if (!(*buf & 0xf0))
-			rl[rlpos].lcn = LCN_HOLE;
-		else {
-			/* Get the lcn change which really can be negative. */
-			u8 b2 = *buf & 0xf;
-			b = b2 + ((*buf >> 4) & 0xf);
-			if (buf + b > attr_end)
-				goto io_error;
-			for (deltaxcn = (s8)buf[b--]; b > b2; b--)
-				deltaxcn = (deltaxcn << 8) + buf[b];
-			/* Change the current lcn to its new value. */
-			lcn += deltaxcn;
-#ifdef DEBUG
-			/*
-			 * On NTFS 1.2-, apparently can have lcn == -1 to
-			 * indicate a hole. But we haven't verified ourselves
-			 * whether it is really the lcn or the deltaxcn that is
-			 * -1. So if either is found give us a message so we
-			 * can investigate it further!
-			 */
-			if (vol->major_ver < 3) {
-				if (unlikely(deltaxcn == (LCN)-1))
-					ntfs_error(vol->sb, "lcn delta == -1");
-				if (unlikely(lcn == (LCN)-1))
-					ntfs_error(vol->sb, "lcn == -1");
-			}
-#endif
-			/* Check lcn is not below -1. */
-			if (unlikely(lcn < (LCN)-1)) {
-				ntfs_error(vol->sb, "Invalid LCN < -1 in "
-						"mapping pairs array.");
-				goto err_out;
-			}
-			/* Enter the current lcn into the runlist element. */
-			rl[rlpos].lcn = lcn;
-		}
-		/* Get to the next runlist element. */
-		rlpos++;
-		/* Increment the buffer position to the next mapping pair. */
-		buf += (*buf & 0xf) + ((*buf >> 4) & 0xf) + 1;
-	}
-	if (unlikely(buf >= attr_end))
-		goto io_error;
-	/*
-	 * If there is a highest_vcn specified, it must be equal to the final
-	 * vcn in the runlist - 1, or something has gone badly wrong.
-	 */
-	deltaxcn = sle64_to_cpu(attr->data.non_resident.highest_vcn);
-	if (unlikely(deltaxcn && vcn - 1 != deltaxcn)) {
-mpa_err:
-		ntfs_error(vol->sb, "Corrupt mapping pairs array in "
-				"non-resident attribute.");
-		goto err_out;
-	}
-	/* Setup not mapped runlist element if this is the base extent. */
-	if (!attr->data.non_resident.lowest_vcn) {
-		VCN max_cluster;
-
-		max_cluster = ((sle64_to_cpu(
-				attr->data.non_resident.allocated_size) +
-				vol->cluster_size - 1) >>
-				vol->cluster_size_bits) - 1;
-		/*
-		 * A highest_vcn of zero means this is a single extent
-		 * attribute so simply terminate the runlist with LCN_ENOENT).
-		 */
-		if (deltaxcn) {
-			/*
-			 * If there is a difference between the highest_vcn and
-			 * the highest cluster, the runlist is either corrupt
-			 * or, more likely, there are more extents following
-			 * this one.
-			 */
-			if (deltaxcn < max_cluster) {
-				ntfs_debug("More extents to follow; deltaxcn "
-						"= 0x%llx, max_cluster = "
-						"0x%llx",
-						(unsigned long long)deltaxcn,
-						(unsigned long long)
-						max_cluster);
-				rl[rlpos].vcn = vcn;
-				vcn += rl[rlpos].length = max_cluster -
-						deltaxcn;
-				rl[rlpos].lcn = LCN_RL_NOT_MAPPED;
-				rlpos++;
-			} else if (unlikely(deltaxcn > max_cluster)) {
-				ntfs_error(vol->sb, "Corrupt attribute.  "
-						"deltaxcn = 0x%llx, "
-						"max_cluster = 0x%llx",
-						(unsigned long long)deltaxcn,
-						(unsigned long long)
-						max_cluster);
-				goto mpa_err;
-			}
-		}
-		rl[rlpos].lcn = LCN_ENOENT;
-	} else /* Not the base extent. There may be more extents to follow. */
-		rl[rlpos].lcn = LCN_RL_NOT_MAPPED;
-
-	/* Setup terminating runlist element. */
-	rl[rlpos].vcn = vcn;
-	rl[rlpos].length = (s64)0;
-	/* If no existing runlist was specified, we are done. */
-	if (!old_rl) {
-		ntfs_debug("Mapping pairs array successfully decompressed:");
-		ntfs_debug_dump_runlist(rl);
-		return rl;
-	}
-	/* Now combine the new and old runlists checking for overlaps. */
-	old_rl = ntfs_runlists_merge(old_rl, rl);
-	if (!IS_ERR(old_rl))
-		return old_rl;
-	ntfs_free(rl);
-	ntfs_error(vol->sb, "Failed to merge runlists.");
-	return old_rl;
-io_error:
-	ntfs_error(vol->sb, "Corrupt attribute.");
-err_out:
-	ntfs_free(rl);
-	return ERR_PTR(-EIO);
-}
-
-/**
- * ntfs_rl_vcn_to_lcn - convert a vcn into a lcn given a runlist
- * @rl:		runlist to use for conversion
- * @vcn:	vcn to convert
- *
- * Convert the virtual cluster number @vcn of an attribute into a logical
- * cluster number (lcn) of a device using the runlist @rl to map vcns to their
- * corresponding lcns.
- *
- * It is up to the caller to serialize access to the runlist @rl.
- *
- * Since lcns must be >= 0, we use negative return codes with special meaning:
- *
- * Return code		Meaning / Description
- * ==================================================
- *  LCN_HOLE		Hole / not allocated on disk.
- *  LCN_RL_NOT_MAPPED	This is part of the runlist which has not been
- *			inserted into the runlist yet.
- *  LCN_ENOENT		There is no such vcn in the attribute.
- *
- * Locking: - The caller must have locked the runlist (for reading or writing).
- *	    - This function does not touch the lock, nor does it modify the
- *	      runlist.
- */
-LCN ntfs_rl_vcn_to_lcn(const runlist_element *rl, const VCN vcn)
-{
-	int i;
-
-	BUG_ON(vcn < 0);
-	/*
-	 * If rl is NULL, assume that we have found an unmapped runlist. The
-	 * caller can then attempt to map it and fail appropriately if
-	 * necessary.
-	 */
-	if (unlikely(!rl))
-		return LCN_RL_NOT_MAPPED;
-
-	/* Catch out of lower bounds vcn. */
-	if (unlikely(vcn < rl[0].vcn))
-		return LCN_ENOENT;
-
-	for (i = 0; likely(rl[i].length); i++) {
-		if (unlikely(vcn < rl[i+1].vcn)) {
-			if (likely(rl[i].lcn >= (LCN)0))
-				return rl[i].lcn + (vcn - rl[i].vcn);
-			return rl[i].lcn;
-		}
-	}
-	/*
-	 * The terminator element is setup to the correct value, i.e. one of
-	 * LCN_HOLE, LCN_RL_NOT_MAPPED, or LCN_ENOENT.
-	 */
-	if (likely(rl[i].lcn < (LCN)0))
-		return rl[i].lcn;
-	/* Just in case... We could replace this with BUG() some day. */
-	return LCN_ENOENT;
-}
-
-#ifdef NTFS_RW
-
-/**
- * ntfs_rl_find_vcn_nolock - find a vcn in a runlist
- * @rl:		runlist to search
- * @vcn:	vcn to find
- *
- * Find the virtual cluster number @vcn in the runlist @rl and return the
- * address of the runlist element containing the @vcn on success.
- *
- * Return NULL if @rl is NULL or @vcn is in an unmapped part/out of bounds of
- * the runlist.
- *
- * Locking: The runlist must be locked on entry.
- */
-runlist_element *ntfs_rl_find_vcn_nolock(runlist_element *rl, const VCN vcn)
-{
-	BUG_ON(vcn < 0);
-	if (unlikely(!rl || vcn < rl[0].vcn))
-		return NULL;
-	while (likely(rl->length)) {
-		if (unlikely(vcn < rl[1].vcn)) {
-			if (likely(rl->lcn >= LCN_HOLE))
-				return rl;
-			return NULL;
-		}
-		rl++;
-	}
-	if (likely(rl->lcn == LCN_ENOENT))
-		return rl;
-	return NULL;
-}
-
-/**
- * ntfs_get_nr_significant_bytes - get number of bytes needed to store a number
- * @n:		number for which to get the number of bytes for
- *
- * Return the number of bytes required to store @n unambiguously as
- * a signed number.
- *
- * This is used in the context of the mapping pairs array to determine how
- * many bytes will be needed in the array to store a given logical cluster
- * number (lcn) or a specific run length.
- *
- * Return the number of bytes written.  This function cannot fail.
- */
-static inline int ntfs_get_nr_significant_bytes(const s64 n)
-{
-	s64 l = n;
-	int i;
-	s8 j;
-
-	i = 0;
-	do {
-		l >>= 8;
-		i++;
-	} while (l != 0 && l != -1);
-	j = (n >> 8 * (i - 1)) & 0xff;
-	/* If the sign bit is wrong, we need an extra byte. */
-	if ((n < 0 && j >= 0) || (n > 0 && j < 0))
-		i++;
-	return i;
-}
-
-/**
- * ntfs_get_size_for_mapping_pairs - get bytes needed for mapping pairs array
- * @vol:	ntfs volume (needed for the ntfs version)
- * @rl:		locked runlist to determine the size of the mapping pairs of
- * @first_vcn:	first vcn which to include in the mapping pairs array
- * @last_vcn:	last vcn which to include in the mapping pairs array
- *
- * Walk the locked runlist @rl and calculate the size in bytes of the mapping
- * pairs array corresponding to the runlist @rl, starting at vcn @first_vcn and
- * finishing with vcn @last_vcn.
- *
- * A @last_vcn of -1 means end of runlist and in that case the size of the
- * mapping pairs array corresponding to the runlist starting at vcn @first_vcn
- * and finishing at the end of the runlist is determined.
- *
- * This for example allows us to allocate a buffer of the right size when
- * building the mapping pairs array.
- *
- * If @rl is NULL, just return 1 (for the single terminator byte).
- *
- * Return the calculated size in bytes on success.  On error, return -errno.
- * The following error codes are defined:
- *	-EINVAL	- Run list contains unmapped elements.  Make sure to only pass
- *		  fully mapped runlists to this function.
- *	-EIO	- The runlist is corrupt.
- *
- * Locking: @rl must be locked on entry (either for reading or writing), it
- *	    remains locked throughout, and is left locked upon return.
- */
-int ntfs_get_size_for_mapping_pairs(const ntfs_volume *vol,
-		const runlist_element *rl, const VCN first_vcn,
-		const VCN last_vcn)
-{
-	LCN prev_lcn;
-	int rls;
-	bool the_end = false;
-
-	BUG_ON(first_vcn < 0);
-	BUG_ON(last_vcn < -1);
-	BUG_ON(last_vcn >= 0 && first_vcn > last_vcn);
-	if (!rl) {
-		BUG_ON(first_vcn);
-		BUG_ON(last_vcn > 0);
-		return 1;
-	}
-	/* Skip to runlist element containing @first_vcn. */
-	while (rl->length && first_vcn >= rl[1].vcn)
-		rl++;
-	if (unlikely((!rl->length && first_vcn > rl->vcn) ||
-			first_vcn < rl->vcn))
-		return -EINVAL;
-	prev_lcn = 0;
-	/* Always need the termining zero byte. */
-	rls = 1;
-	/* Do the first partial run if present. */
-	if (first_vcn > rl->vcn) {
-		s64 delta, length = rl->length;
-
-		/* We know rl->length != 0 already. */
-		if (unlikely(length < 0 || rl->lcn < LCN_HOLE))
-			goto err_out;
-		/*
-		 * If @stop_vcn is given and finishes inside this run, cap the
-		 * run length.
-		 */
-		if (unlikely(last_vcn >= 0 && rl[1].vcn > last_vcn)) {
-			s64 s1 = last_vcn + 1;
-			if (unlikely(rl[1].vcn > s1))
-				length = s1 - rl->vcn;
-			the_end = true;
-		}
-		delta = first_vcn - rl->vcn;
-		/* Header byte + length. */
-		rls += 1 + ntfs_get_nr_significant_bytes(length - delta);
-		/*
-		 * If the logical cluster number (lcn) denotes a hole and we
-		 * are on NTFS 3.0+, we don't store it at all, i.e. we need
-		 * zero space.  On earlier NTFS versions we just store the lcn.
-		 * Note: this assumes that on NTFS 1.2-, holes are stored with
-		 * an lcn of -1 and not a delta_lcn of -1 (unless both are -1).
-		 */
-		if (likely(rl->lcn >= 0 || vol->major_ver < 3)) {
-			prev_lcn = rl->lcn;
-			if (likely(rl->lcn >= 0))
-				prev_lcn += delta;
-			/* Change in lcn. */
-			rls += ntfs_get_nr_significant_bytes(prev_lcn);
-		}
-		/* Go to next runlist element. */
-		rl++;
-	}
-	/* Do the full runs. */
-	for (; rl->length && !the_end; rl++) {
-		s64 length = rl->length;
-
-		if (unlikely(length < 0 || rl->lcn < LCN_HOLE))
-			goto err_out;
-		/*
-		 * If @stop_vcn is given and finishes inside this run, cap the
-		 * run length.
-		 */
-		if (unlikely(last_vcn >= 0 && rl[1].vcn > last_vcn)) {
-			s64 s1 = last_vcn + 1;
-			if (unlikely(rl[1].vcn > s1))
-				length = s1 - rl->vcn;
-			the_end = true;
-		}
-		/* Header byte + length. */
-		rls += 1 + ntfs_get_nr_significant_bytes(length);
-		/*
-		 * If the logical cluster number (lcn) denotes a hole and we
-		 * are on NTFS 3.0+, we don't store it at all, i.e. we need
-		 * zero space.  On earlier NTFS versions we just store the lcn.
-		 * Note: this assumes that on NTFS 1.2-, holes are stored with
-		 * an lcn of -1 and not a delta_lcn of -1 (unless both are -1).
-		 */
-		if (likely(rl->lcn >= 0 || vol->major_ver < 3)) {
-			/* Change in lcn. */
-			rls += ntfs_get_nr_significant_bytes(rl->lcn -
-					prev_lcn);
-			prev_lcn = rl->lcn;
-		}
-	}
-	return rls;
-err_out:
-	if (rl->lcn == LCN_RL_NOT_MAPPED)
-		rls = -EINVAL;
-	else
-		rls = -EIO;
-	return rls;
-}
-
-/**
- * ntfs_write_significant_bytes - write the significant bytes of a number
- * @dst:	destination buffer to write to
- * @dst_max:	pointer to last byte of destination buffer for bounds checking
- * @n:		number whose significant bytes to write
- *
- * Store in @dst, the minimum bytes of the number @n which are required to
- * identify @n unambiguously as a signed number, taking care not to exceed
- * @dest_max, the maximum position within @dst to which we are allowed to
- * write.
- *
- * This is used when building the mapping pairs array of a runlist to compress
- * a given logical cluster number (lcn) or a specific run length to the minimum
- * size possible.
- *
- * Return the number of bytes written on success.  On error, i.e. the
- * destination buffer @dst is too small, return -ENOSPC.
- */
-static inline int ntfs_write_significant_bytes(s8 *dst, const s8 *dst_max,
-		const s64 n)
-{
-	s64 l = n;
-	int i;
-	s8 j;
-
-	i = 0;
-	do {
-		if (unlikely(dst > dst_max))
-			goto err_out;
-		*dst++ = l & 0xffll;
-		l >>= 8;
-		i++;
-	} while (l != 0 && l != -1);
-	j = (n >> 8 * (i - 1)) & 0xff;
-	/* If the sign bit is wrong, we need an extra byte. */
-	if (n < 0 && j >= 0) {
-		if (unlikely(dst > dst_max))
-			goto err_out;
-		i++;
-		*dst = (s8)-1;
-	} else if (n > 0 && j < 0) {
-		if (unlikely(dst > dst_max))
-			goto err_out;
-		i++;
-		*dst = (s8)0;
-	}
-	return i;
-err_out:
-	return -ENOSPC;
-}
-
-/**
- * ntfs_mapping_pairs_build - build the mapping pairs array from a runlist
- * @vol:	ntfs volume (needed for the ntfs version)
- * @dst:	destination buffer to which to write the mapping pairs array
- * @dst_len:	size of destination buffer @dst in bytes
- * @rl:		locked runlist for which to build the mapping pairs array
- * @first_vcn:	first vcn which to include in the mapping pairs array
- * @last_vcn:	last vcn which to include in the mapping pairs array
- * @stop_vcn:	first vcn outside destination buffer on success or -ENOSPC
- *
- * Create the mapping pairs array from the locked runlist @rl, starting at vcn
- * @first_vcn and finishing with vcn @last_vcn and save the array in @dst.
- * @dst_len is the size of @dst in bytes and it should be at least equal to the
- * value obtained by calling ntfs_get_size_for_mapping_pairs().
- *
- * A @last_vcn of -1 means end of runlist and in that case the mapping pairs
- * array corresponding to the runlist starting at vcn @first_vcn and finishing
- * at the end of the runlist is created.
- *
- * If @rl is NULL, just write a single terminator byte to @dst.
- *
- * On success or -ENOSPC error, if @stop_vcn is not NULL, *@stop_vcn is set to
- * the first vcn outside the destination buffer.  Note that on error, @dst has
- * been filled with all the mapping pairs that will fit, thus it can be treated
- * as partial success, in that a new attribute extent needs to be created or
- * the next extent has to be used and the mapping pairs build has to be
- * continued with @first_vcn set to *@stop_vcn.
- *
- * Return 0 on success and -errno on error.  The following error codes are
- * defined:
- *	-EINVAL	- Run list contains unmapped elements.  Make sure to only pass
- *		  fully mapped runlists to this function.
- *	-EIO	- The runlist is corrupt.
- *	-ENOSPC	- The destination buffer is too small.
- *
- * Locking: @rl must be locked on entry (either for reading or writing), it
- *	    remains locked throughout, and is left locked upon return.
- */
-int ntfs_mapping_pairs_build(const ntfs_volume *vol, s8 *dst,
-		const int dst_len, const runlist_element *rl,
-		const VCN first_vcn, const VCN last_vcn, VCN *const stop_vcn)
-{
-	LCN prev_lcn;
-	s8 *dst_max, *dst_next;
-	int err = -ENOSPC;
-	bool the_end = false;
-	s8 len_len, lcn_len;
-
-	BUG_ON(first_vcn < 0);
-	BUG_ON(last_vcn < -1);
-	BUG_ON(last_vcn >= 0 && first_vcn > last_vcn);
-	BUG_ON(dst_len < 1);
-	if (!rl) {
-		BUG_ON(first_vcn);
-		BUG_ON(last_vcn > 0);
-		if (stop_vcn)
-			*stop_vcn = 0;
-		/* Terminator byte. */
-		*dst = 0;
-		return 0;
-	}
-	/* Skip to runlist element containing @first_vcn. */
-	while (rl->length && first_vcn >= rl[1].vcn)
-		rl++;
-	if (unlikely((!rl->length && first_vcn > rl->vcn) ||
-			first_vcn < rl->vcn))
-		return -EINVAL;
-	/*
-	 * @dst_max is used for bounds checking in
-	 * ntfs_write_significant_bytes().
-	 */
-	dst_max = dst + dst_len - 1;
-	prev_lcn = 0;
-	/* Do the first partial run if present. */
-	if (first_vcn > rl->vcn) {
-		s64 delta, length = rl->length;
-
-		/* We know rl->length != 0 already. */
-		if (unlikely(length < 0 || rl->lcn < LCN_HOLE))
-			goto err_out;
-		/*
-		 * If @stop_vcn is given and finishes inside this run, cap the
-		 * run length.
-		 */
-		if (unlikely(last_vcn >= 0 && rl[1].vcn > last_vcn)) {
-			s64 s1 = last_vcn + 1;
-			if (unlikely(rl[1].vcn > s1))
-				length = s1 - rl->vcn;
-			the_end = true;
-		}
-		delta = first_vcn - rl->vcn;
-		/* Write length. */
-		len_len = ntfs_write_significant_bytes(dst + 1, dst_max,
-				length - delta);
-		if (unlikely(len_len < 0))
-			goto size_err;
-		/*
-		 * If the logical cluster number (lcn) denotes a hole and we
-		 * are on NTFS 3.0+, we don't store it at all, i.e. we need
-		 * zero space.  On earlier NTFS versions we just write the lcn
-		 * change.  FIXME: Do we need to write the lcn change or just
-		 * the lcn in that case?  Not sure as I have never seen this
-		 * case on NT4. - We assume that we just need to write the lcn
-		 * change until someone tells us otherwise... (AIA)
-		 */
-		if (likely(rl->lcn >= 0 || vol->major_ver < 3)) {
-			prev_lcn = rl->lcn;
-			if (likely(rl->lcn >= 0))
-				prev_lcn += delta;
-			/* Write change in lcn. */
-			lcn_len = ntfs_write_significant_bytes(dst + 1 +
-					len_len, dst_max, prev_lcn);
-			if (unlikely(lcn_len < 0))
-				goto size_err;
-		} else
-			lcn_len = 0;
-		dst_next = dst + len_len + lcn_len + 1;
-		if (unlikely(dst_next > dst_max))
-			goto size_err;
-		/* Update header byte. */
-		*dst = lcn_len << 4 | len_len;
-		/* Position at next mapping pairs array element. */
-		dst = dst_next;
-		/* Go to next runlist element. */
-		rl++;
-	}
-	/* Do the full runs. */
-	for (; rl->length && !the_end; rl++) {
-		s64 length = rl->length;
-
-		if (unlikely(length < 0 || rl->lcn < LCN_HOLE))
-			goto err_out;
-		/*
-		 * If @stop_vcn is given and finishes inside this run, cap the
-		 * run length.
-		 */
-		if (unlikely(last_vcn >= 0 && rl[1].vcn > last_vcn)) {
-			s64 s1 = last_vcn + 1;
-			if (unlikely(rl[1].vcn > s1))
-				length = s1 - rl->vcn;
-			the_end = true;
-		}
-		/* Write length. */
-		len_len = ntfs_write_significant_bytes(dst + 1, dst_max,
-				length);
-		if (unlikely(len_len < 0))
-			goto size_err;
-		/*
-		 * If the logical cluster number (lcn) denotes a hole and we
-		 * are on NTFS 3.0+, we don't store it at all, i.e. we need
-		 * zero space.  On earlier NTFS versions we just write the lcn
-		 * change.  FIXME: Do we need to write the lcn change or just
-		 * the lcn in that case?  Not sure as I have never seen this
-		 * case on NT4. - We assume that we just need to write the lcn
-		 * change until someone tells us otherwise... (AIA)
-		 */
-		if (likely(rl->lcn >= 0 || vol->major_ver < 3)) {
-			/* Write change in lcn. */
-			lcn_len = ntfs_write_significant_bytes(dst + 1 +
-					len_len, dst_max, rl->lcn - prev_lcn);
-			if (unlikely(lcn_len < 0))
-				goto size_err;
-			prev_lcn = rl->lcn;
-		} else
-			lcn_len = 0;
-		dst_next = dst + len_len + lcn_len + 1;
-		if (unlikely(dst_next > dst_max))
-			goto size_err;
-		/* Update header byte. */
-		*dst = lcn_len << 4 | len_len;
-		/* Position at next mapping pairs array element. */
-		dst = dst_next;
-	}
-	/* Success. */
-	err = 0;
-size_err:
-	/* Set stop vcn. */
-	if (stop_vcn)
-		*stop_vcn = rl->vcn;
-	/* Add terminator byte. */
-	*dst = 0;
-	return err;
-err_out:
-	if (rl->lcn == LCN_RL_NOT_MAPPED)
-		err = -EINVAL;
-	else
-		err = -EIO;
-	return err;
-}
-
-/**
- * ntfs_rl_truncate_nolock - truncate a runlist starting at a specified vcn
- * @vol:	ntfs volume (needed for error output)
- * @runlist:	runlist to truncate
- * @new_length:	the new length of the runlist in VCNs
- *
- * Truncate the runlist described by @runlist as well as the memory buffer
- * holding the runlist elements to a length of @new_length VCNs.
- *
- * If @new_length lies within the runlist, the runlist elements with VCNs of
- * @new_length and above are discarded.  As a special case if @new_length is
- * zero, the runlist is discarded and set to NULL.
- *
- * If @new_length lies beyond the runlist, a sparse runlist element is added to
- * the end of the runlist @runlist or if the last runlist element is a sparse
- * one already, this is extended.
- *
- * Note, no checking is done for unmapped runlist elements.  It is assumed that
- * the caller has mapped any elements that need to be mapped already.
- *
- * Return 0 on success and -errno on error.
- *
- * Locking: The caller must hold @runlist->lock for writing.
- */
-int ntfs_rl_truncate_nolock(const ntfs_volume *vol, runlist *const runlist,
-		const s64 new_length)
-{
-	runlist_element *rl;
-	int old_size;
-
-	ntfs_debug("Entering for new_length 0x%llx.", (long long)new_length);
-	BUG_ON(!runlist);
-	BUG_ON(new_length < 0);
-	rl = runlist->rl;
-	if (!new_length) {
-		ntfs_debug("Freeing runlist.");
-		runlist->rl = NULL;
-		if (rl)
-			ntfs_free(rl);
-		return 0;
-	}
-	if (unlikely(!rl)) {
-		/*
-		 * Create a runlist consisting of a sparse runlist element of
-		 * length @new_length followed by a terminator runlist element.
-		 */
-		rl = ntfs_malloc_nofs(PAGE_SIZE);
-		if (unlikely(!rl)) {
-			ntfs_error(vol->sb, "Not enough memory to allocate "
-					"runlist element buffer.");
-			return -ENOMEM;
-		}
-		runlist->rl = rl;
-		rl[1].length = rl->vcn = 0;
-		rl->lcn = LCN_HOLE;
-		rl[1].vcn = rl->length = new_length;
-		rl[1].lcn = LCN_ENOENT;
-		return 0;
-	}
-	BUG_ON(new_length < rl->vcn);
-	/* Find @new_length in the runlist. */
-	while (likely(rl->length && new_length >= rl[1].vcn))
-		rl++;
-	/*
-	 * If not at the end of the runlist we need to shrink it.
-	 * If at the end of the runlist we need to expand it.
-	 */
-	if (rl->length) {
-		runlist_element *trl;
-		bool is_end;
-
-		ntfs_debug("Shrinking runlist.");
-		/* Determine the runlist size. */
-		trl = rl + 1;
-		while (likely(trl->length))
-			trl++;
-		old_size = trl - runlist->rl + 1;
-		/* Truncate the run. */
-		rl->length = new_length - rl->vcn;
-		/*
-		 * If a run was partially truncated, make the following runlist
-		 * element a terminator.
-		 */
-		is_end = false;
-		if (rl->length) {
-			rl++;
-			if (!rl->length)
-				is_end = true;
-			rl->vcn = new_length;
-			rl->length = 0;
-		}
-		rl->lcn = LCN_ENOENT;
-		/* Reallocate memory if necessary. */
-		if (!is_end) {
-			int new_size = rl - runlist->rl + 1;
-			rl = ntfs_rl_realloc(runlist->rl, old_size, new_size);
-			if (IS_ERR(rl))
-				ntfs_warning(vol->sb, "Failed to shrink "
-						"runlist buffer.  This just "
-						"wastes a bit of memory "
-						"temporarily so we ignore it "
-						"and return success.");
-			else
-				runlist->rl = rl;
-		}
-	} else if (likely(/* !rl->length && */ new_length > rl->vcn)) {
-		ntfs_debug("Expanding runlist.");
-		/*
-		 * If there is a previous runlist element and it is a sparse
-		 * one, extend it.  Otherwise need to add a new, sparse runlist
-		 * element.
-		 */
-		if ((rl > runlist->rl) && ((rl - 1)->lcn == LCN_HOLE))
-			(rl - 1)->length = new_length - (rl - 1)->vcn;
-		else {
-			/* Determine the runlist size. */
-			old_size = rl - runlist->rl + 1;
-			/* Reallocate memory if necessary. */
-			rl = ntfs_rl_realloc(runlist->rl, old_size,
-					old_size + 1);
-			if (IS_ERR(rl)) {
-				ntfs_error(vol->sb, "Failed to expand runlist "
-						"buffer, aborting.");
-				return PTR_ERR(rl);
-			}
-			runlist->rl = rl;
-			/*
-			 * Set @rl to the same runlist element in the new
-			 * runlist as before in the old runlist.
-			 */
-			rl += old_size - 1;
-			/* Add a new, sparse runlist element. */
-			rl->lcn = LCN_HOLE;
-			rl->length = new_length - rl->vcn;
-			/* Add a new terminator runlist element. */
-			rl++;
-			rl->length = 0;
-		}
-		rl->vcn = new_length;
-		rl->lcn = LCN_ENOENT;
-	} else /* if (unlikely(!rl->length && new_length == rl->vcn)) */ {
-		/* Runlist already has same size as requested. */
-		rl->lcn = LCN_ENOENT;
-	}
-	ntfs_debug("Done.");
-	return 0;
-}
-
-/**
- * ntfs_rl_punch_nolock - punch a hole into a runlist
- * @vol:	ntfs volume (needed for error output)
- * @runlist:	runlist to punch a hole into
- * @start:	starting VCN of the hole to be created
- * @length:	size of the hole to be created in units of clusters
- *
- * Punch a hole into the runlist @runlist starting at VCN @start and of size
- * @length clusters.
- *
- * Return 0 on success and -errno on error, in which case @runlist has not been
- * modified.
- *
- * If @start and/or @start + @length are outside the runlist return error code
- * -ENOENT.
- *
- * If the runlist contains unmapped or error elements between @start and @start
- * + @length return error code -EINVAL.
- *
- * Locking: The caller must hold @runlist->lock for writing.
- */
-int ntfs_rl_punch_nolock(const ntfs_volume *vol, runlist *const runlist,
-		const VCN start, const s64 length)
-{
-	const VCN end = start + length;
-	s64 delta;
-	runlist_element *rl, *rl_end, *rl_real_end, *trl;
-	int old_size;
-	bool lcn_fixup = false;
-
-	ntfs_debug("Entering for start 0x%llx, length 0x%llx.",
-			(long long)start, (long long)length);
-	BUG_ON(!runlist);
-	BUG_ON(start < 0);
-	BUG_ON(length < 0);
-	BUG_ON(end < 0);
-	rl = runlist->rl;
-	if (unlikely(!rl)) {
-		if (likely(!start && !length))
-			return 0;
-		return -ENOENT;
-	}
-	/* Find @start in the runlist. */
-	while (likely(rl->length && start >= rl[1].vcn))
-		rl++;
-	rl_end = rl;
-	/* Find @end in the runlist. */
-	while (likely(rl_end->length && end >= rl_end[1].vcn)) {
-		/* Verify there are no unmapped or error elements. */
-		if (unlikely(rl_end->lcn < LCN_HOLE))
-			return -EINVAL;
-		rl_end++;
-	}
-	/* Check the last element. */
-	if (unlikely(rl_end->length && rl_end->lcn < LCN_HOLE))
-		return -EINVAL;
-	/* This covers @start being out of bounds, too. */
-	if (!rl_end->length && end > rl_end->vcn)
-		return -ENOENT;
-	if (!length)
-		return 0;
-	if (!rl->length)
-		return -ENOENT;
-	rl_real_end = rl_end;
-	/* Determine the runlist size. */
-	while (likely(rl_real_end->length))
-		rl_real_end++;
-	old_size = rl_real_end - runlist->rl + 1;
-	/* If @start is in a hole simply extend the hole. */
-	if (rl->lcn == LCN_HOLE) {
-		/*
-		 * If both @start and @end are in the same sparse run, we are
-		 * done.
-		 */
-		if (end <= rl[1].vcn) {
-			ntfs_debug("Done (requested hole is already sparse).");
-			return 0;
-		}
-extend_hole:
-		/* Extend the hole. */
-		rl->length = end - rl->vcn;
-		/* If @end is in a hole, merge it with the current one. */
-		if (rl_end->lcn == LCN_HOLE) {
-			rl_end++;
-			rl->length = rl_end->vcn - rl->vcn;
-		}
-		/* We have done the hole.  Now deal with the remaining tail. */
-		rl++;
-		/* Cut out all runlist elements up to @end. */
-		if (rl < rl_end)
-			memmove(rl, rl_end, (rl_real_end - rl_end + 1) *
-					sizeof(*rl));
-		/* Adjust the beginning of the tail if necessary. */
-		if (end > rl->vcn) {
-			delta = end - rl->vcn;
-			rl->vcn = end;
-			rl->length -= delta;
-			/* Only adjust the lcn if it is real. */
-			if (rl->lcn >= 0)
-				rl->lcn += delta;
-		}
-shrink_allocation:
-		/* Reallocate memory if the allocation changed. */
-		if (rl < rl_end) {
-			rl = ntfs_rl_realloc(runlist->rl, old_size,
-					old_size - (rl_end - rl));
-			if (IS_ERR(rl))
-				ntfs_warning(vol->sb, "Failed to shrink "
-						"runlist buffer.  This just "
-						"wastes a bit of memory "
-						"temporarily so we ignore it "
-						"and return success.");
-			else
-				runlist->rl = rl;
-		}
-		ntfs_debug("Done (extend hole).");
-		return 0;
-	}
-	/*
-	 * If @start is at the beginning of a run things are easier as there is
-	 * no need to split the first run.
-	 */
-	if (start == rl->vcn) {
-		/*
-		 * @start is at the beginning of a run.
-		 *
-		 * If the previous run is sparse, extend its hole.
-		 *
-		 * If @end is not in the same run, switch the run to be sparse
-		 * and extend the newly created hole.
-		 *
-		 * Thus both of these cases reduce the problem to the above
-		 * case of "@start is in a hole".
-		 */
-		if (rl > runlist->rl && (rl - 1)->lcn == LCN_HOLE) {
-			rl--;
-			goto extend_hole;
-		}
-		if (end >= rl[1].vcn) {
-			rl->lcn = LCN_HOLE;
-			goto extend_hole;
-		}
-		/*
-		 * The final case is when @end is in the same run as @start.
-		 * For this need to split the run into two.  One run for the
-		 * sparse region between the beginning of the old run, i.e.
-		 * @start, and @end and one for the remaining non-sparse
-		 * region, i.e. between @end and the end of the old run.
-		 */
-		trl = ntfs_rl_realloc(runlist->rl, old_size, old_size + 1);
-		if (IS_ERR(trl))
-			goto enomem_out;
-		old_size++;
-		if (runlist->rl != trl) {
-			rl = trl + (rl - runlist->rl);
-			rl_end = trl + (rl_end - runlist->rl);
-			rl_real_end = trl + (rl_real_end - runlist->rl);
-			runlist->rl = trl;
-		}
-split_end:
-		/* Shift all the runs up by one. */
-		memmove(rl + 1, rl, (rl_real_end - rl + 1) * sizeof(*rl));
-		/* Finally, setup the two split runs. */
-		rl->lcn = LCN_HOLE;
-		rl->length = length;
-		rl++;
-		rl->vcn += length;
-		/* Only adjust the lcn if it is real. */
-		if (rl->lcn >= 0 || lcn_fixup)
-			rl->lcn += length;
-		rl->length -= length;
-		ntfs_debug("Done (split one).");
-		return 0;
-	}
-	/*
-	 * @start is neither in a hole nor at the beginning of a run.
-	 *
-	 * If @end is in a hole, things are easier as simply truncating the run
-	 * @start is in to end at @start - 1, deleting all runs after that up
-	 * to @end, and finally extending the beginning of the run @end is in
-	 * to be @start is all that is needed.
-	 */
-	if (rl_end->lcn == LCN_HOLE) {
-		/* Truncate the run containing @start. */
-		rl->length = start - rl->vcn;
-		rl++;
-		/* Cut out all runlist elements up to @end. */
-		if (rl < rl_end)
-			memmove(rl, rl_end, (rl_real_end - rl_end + 1) *
-					sizeof(*rl));
-		/* Extend the beginning of the run @end is in to be @start. */
-		rl->vcn = start;
-		rl->length = rl[1].vcn - start;
-		goto shrink_allocation;
-	}
-	/* 
-	 * If @end is not in a hole there are still two cases to distinguish.
-	 * Either @end is or is not in the same run as @start.
-	 *
-	 * The second case is easier as it can be reduced to an already solved
-	 * problem by truncating the run @start is in to end at @start - 1.
-	 * Then, if @end is in the next run need to split the run into a sparse
-	 * run followed by a non-sparse run (already covered above) and if @end
-	 * is not in the next run switching it to be sparse, again reduces the
-	 * problem to the already covered case of "@start is in a hole".
-	 */
-	if (end >= rl[1].vcn) {
-		/*
-		 * If @end is not in the next run, reduce the problem to the
-		 * case of "@start is in a hole".
-		 */
-		if (rl[1].length && end >= rl[2].vcn) {
-			/* Truncate the run containing @start. */
-			rl->length = start - rl->vcn;
-			rl++;
-			rl->vcn = start;
-			rl->lcn = LCN_HOLE;
-			goto extend_hole;
-		}
-		trl = ntfs_rl_realloc(runlist->rl, old_size, old_size + 1);
-		if (IS_ERR(trl))
-			goto enomem_out;
-		old_size++;
-		if (runlist->rl != trl) {
-			rl = trl + (rl - runlist->rl);
-			rl_end = trl + (rl_end - runlist->rl);
-			rl_real_end = trl + (rl_real_end - runlist->rl);
-			runlist->rl = trl;
-		}
-		/* Truncate the run containing @start. */
-		rl->length = start - rl->vcn;
-		rl++;
-		/*
-		 * @end is in the next run, reduce the problem to the case
-		 * where "@start is at the beginning of a run and @end is in
-		 * the same run as @start".
-		 */
-		delta = rl->vcn - start;
-		rl->vcn = start;
-		if (rl->lcn >= 0) {
-			rl->lcn -= delta;
-			/* Need this in case the lcn just became negative. */
-			lcn_fixup = true;
-		}
-		rl->length += delta;
-		goto split_end;
-	}
-	/*
-	 * The first case from above, i.e. @end is in the same run as @start.
-	 * We need to split the run into three.  One run for the non-sparse
-	 * region between the beginning of the old run and @start, one for the
-	 * sparse region between @start and @end, and one for the remaining
-	 * non-sparse region, i.e. between @end and the end of the old run.
-	 */
-	trl = ntfs_rl_realloc(runlist->rl, old_size, old_size + 2);
-	if (IS_ERR(trl))
-		goto enomem_out;
-	old_size += 2;
-	if (runlist->rl != trl) {
-		rl = trl + (rl - runlist->rl);
-		rl_end = trl + (rl_end - runlist->rl);
-		rl_real_end = trl + (rl_real_end - runlist->rl);
-		runlist->rl = trl;
-	}
-	/* Shift all the runs up by two. */
-	memmove(rl + 2, rl, (rl_real_end - rl + 1) * sizeof(*rl));
-	/* Finally, setup the three split runs. */
-	rl->length = start - rl->vcn;
-	rl++;
-	rl->vcn = start;
-	rl->lcn = LCN_HOLE;
-	rl->length = length;
-	rl++;
-	delta = end - rl->vcn;
-	rl->vcn = end;
-	rl->lcn += delta;
-	rl->length -= delta;
-	ntfs_debug("Done (split both).");
-	return 0;
-enomem_out:
-	ntfs_error(vol->sb, "Not enough memory to extend runlist buffer.");
-	return -ENOMEM;
-}
-
-#endif /* NTFS_RW */
diff --git a/fs/ntfs/runlist.h b/fs/ntfs/runlist.h
deleted file mode 100644
index 38de0a375f..0000000000
--- a/fs/ntfs/runlist.h
+++ /dev/null
@@ -1,88 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * runlist.h - Defines for runlist handling in NTFS Linux kernel driver.
- *	       Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2005 Anton Altaparmakov
- * Copyright (c) 2002 Richard Russon
- */
-
-#ifndef _LINUX_NTFS_RUNLIST_H
-#define _LINUX_NTFS_RUNLIST_H
-
-#include "types.h"
-#include "layout.h"
-#include "volume.h"
-
-/**
- * runlist_element - in memory vcn to lcn mapping array element
- * @vcn:	starting vcn of the current array element
- * @lcn:	starting lcn of the current array element
- * @length:	length in clusters of the current array element
- *
- * The last vcn (in fact the last vcn + 1) is reached when length == 0.
- *
- * When lcn == -1 this means that the count vcns starting at vcn are not
- * physically allocated (i.e. this is a hole / data is sparse).
- */
-typedef struct {	/* In memory vcn to lcn mapping structure element. */
-	VCN vcn;	/* vcn = Starting virtual cluster number. */
-	LCN lcn;	/* lcn = Starting logical cluster number. */
-	s64 length;	/* Run length in clusters. */
-} runlist_element;
-
-/**
- * runlist - in memory vcn to lcn mapping array including a read/write lock
- * @rl:		pointer to an array of runlist elements
- * @lock:	read/write spinlock for serializing access to @rl
- *
- */
-typedef struct {
-	runlist_element *rl;
-	struct rw_semaphore lock;
-} runlist;
-
-static inline void ntfs_init_runlist(runlist *rl)
-{
-	rl->rl = NULL;
-	init_rwsem(&rl->lock);
-}
-
-typedef enum {
-	LCN_HOLE		= -1,	/* Keep this as highest value or die! */
-	LCN_RL_NOT_MAPPED	= -2,
-	LCN_ENOENT		= -3,
-	LCN_ENOMEM		= -4,
-	LCN_EIO			= -5,
-} LCN_SPECIAL_VALUES;
-
-extern runlist_element *ntfs_runlists_merge(runlist_element *drl,
-		runlist_element *srl);
-
-extern runlist_element *ntfs_mapping_pairs_decompress(const ntfs_volume *vol,
-		const ATTR_RECORD *attr, runlist_element *old_rl);
-
-extern LCN ntfs_rl_vcn_to_lcn(const runlist_element *rl, const VCN vcn);
-
-#ifdef NTFS_RW
-
-extern runlist_element *ntfs_rl_find_vcn_nolock(runlist_element *rl,
-		const VCN vcn);
-
-extern int ntfs_get_size_for_mapping_pairs(const ntfs_volume *vol,
-		const runlist_element *rl, const VCN first_vcn,
-		const VCN last_vcn);
-
-extern int ntfs_mapping_pairs_build(const ntfs_volume *vol, s8 *dst,
-		const int dst_len, const runlist_element *rl,
-		const VCN first_vcn, const VCN last_vcn, VCN *const stop_vcn);
-
-extern int ntfs_rl_truncate_nolock(const ntfs_volume *vol,
-		runlist *const runlist, const s64 new_length);
-
-int ntfs_rl_punch_nolock(const ntfs_volume *vol, runlist *const runlist,
-		const VCN start, const s64 length);
-
-#endif /* NTFS_RW */
-
-#endif /* _LINUX_NTFS_RUNLIST_H */
diff --git a/fs/ntfs/super.c b/fs/ntfs/super.c
deleted file mode 100644
index 56a7d5bd33..0000000000
--- a/fs/ntfs/super.c
+++ /dev/null
@@ -1,3202 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * super.c - NTFS kernel super block handling. Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2012 Anton Altaparmakov and Tuxera Inc.
- * Copyright (c) 2001,2002 Richard Russon
- */
-#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
-
-#include <linux/stddef.h>
-#include <linux/init.h>
-#include <linux/slab.h>
-#include <linux/string.h>
-#include <linux/spinlock.h>
-#include <linux/blkdev.h>	/* For bdev_logical_block_size(). */
-#include <linux/backing-dev.h>
-#include <linux/buffer_head.h>
-#include <linux/vfs.h>
-#include <linux/moduleparam.h>
-#include <linux/bitmap.h>
-
-#include "sysctl.h"
-#include "logfile.h"
-#include "quota.h"
-#include "usnjrnl.h"
-#include "dir.h"
-#include "debug.h"
-#include "index.h"
-#include "inode.h"
-#include "aops.h"
-#include "layout.h"
-#include "malloc.h"
-#include "ntfs.h"
-
-/* Number of mounted filesystems which have compression enabled. */
-static unsigned long ntfs_nr_compression_users;
-
-/* A global default upcase table and a corresponding reference count. */
-static ntfschar *default_upcase;
-static unsigned long ntfs_nr_upcase_users;
-
-/* Error constants/strings used in inode.c::ntfs_show_options(). */
-typedef enum {
-	/* One of these must be present, default is ON_ERRORS_CONTINUE. */
-	ON_ERRORS_PANIC			= 0x01,
-	ON_ERRORS_REMOUNT_RO		= 0x02,
-	ON_ERRORS_CONTINUE		= 0x04,
-	/* Optional, can be combined with any of the above. */
-	ON_ERRORS_RECOVER		= 0x10,
-} ON_ERRORS_ACTIONS;
-
-const option_t on_errors_arr[] = {
-	{ ON_ERRORS_PANIC,	"panic" },
-	{ ON_ERRORS_REMOUNT_RO,	"remount-ro", },
-	{ ON_ERRORS_CONTINUE,	"continue", },
-	{ ON_ERRORS_RECOVER,	"recover" },
-	{ 0,			NULL }
-};
-
-/**
- * simple_getbool - convert input string to a boolean value
- * @s: input string to convert
- * @setval: where to store the output boolean value
- *
- * Copied from old ntfs driver (which copied from vfat driver).
- *
- * "1", "yes", "true", or an empty string are converted to %true.
- * "0", "no", and "false" are converted to %false.
- *
- * Return: %1 if the string is converted or was empty and *setval contains it;
- *	   %0 if the string was not valid.
- */
-static int simple_getbool(char *s, bool *setval)
-{
-	if (s) {
-		if (!strcmp(s, "1") || !strcmp(s, "yes") || !strcmp(s, "true"))
-			*setval = true;
-		else if (!strcmp(s, "0") || !strcmp(s, "no") ||
-							!strcmp(s, "false"))
-			*setval = false;
-		else
-			return 0;
-	} else
-		*setval = true;
-	return 1;
-}
-
-/**
- * parse_options - parse the (re)mount options
- * @vol:	ntfs volume
- * @opt:	string containing the (re)mount options
- *
- * Parse the recognized options in @opt for the ntfs volume described by @vol.
- */
-static bool parse_options(ntfs_volume *vol, char *opt)
-{
-	char *p, *v, *ov;
-	static char *utf8 = "utf8";
-	int errors = 0, sloppy = 0;
-	kuid_t uid = INVALID_UID;
-	kgid_t gid = INVALID_GID;
-	umode_t fmask = (umode_t)-1, dmask = (umode_t)-1;
-	int mft_zone_multiplier = -1, on_errors = -1;
-	int show_sys_files = -1, case_sensitive = -1, disable_sparse = -1;
-	struct nls_table *nls_map = NULL, *old_nls;
-
-	/* I am lazy... (-8 */
-#define NTFS_GETOPT_WITH_DEFAULT(option, variable, default_value)	\
-	if (!strcmp(p, option)) {					\
-		if (!v || !*v)						\
-			variable = default_value;			\
-		else {							\
-			variable = simple_strtoul(ov = v, &v, 0);	\
-			if (*v)						\
-				goto needs_val;				\
-		}							\
-	}
-#define NTFS_GETOPT(option, variable)					\
-	if (!strcmp(p, option)) {					\
-		if (!v || !*v)						\
-			goto needs_arg;					\
-		variable = simple_strtoul(ov = v, &v, 0);		\
-		if (*v)							\
-			goto needs_val;					\
-	}
-#define NTFS_GETOPT_UID(option, variable)				\
-	if (!strcmp(p, option)) {					\
-		uid_t uid_value;					\
-		if (!v || !*v)						\
-			goto needs_arg;					\
-		uid_value = simple_strtoul(ov = v, &v, 0);		\
-		if (*v)							\
-			goto needs_val;					\
-		variable = make_kuid(current_user_ns(), uid_value);	\
-		if (!uid_valid(variable))				\
-			goto needs_val;					\
-	}
-#define NTFS_GETOPT_GID(option, variable)				\
-	if (!strcmp(p, option)) {					\
-		gid_t gid_value;					\
-		if (!v || !*v)						\
-			goto needs_arg;					\
-		gid_value = simple_strtoul(ov = v, &v, 0);		\
-		if (*v)							\
-			goto needs_val;					\
-		variable = make_kgid(current_user_ns(), gid_value);	\
-		if (!gid_valid(variable))				\
-			goto needs_val;					\
-	}
-#define NTFS_GETOPT_OCTAL(option, variable)				\
-	if (!strcmp(p, option)) {					\
-		if (!v || !*v)						\
-			goto needs_arg;					\
-		variable = simple_strtoul(ov = v, &v, 8);		\
-		if (*v)							\
-			goto needs_val;					\
-	}
-#define NTFS_GETOPT_BOOL(option, variable)				\
-	if (!strcmp(p, option)) {					\
-		bool val;						\
-		if (!simple_getbool(v, &val))				\
-			goto needs_bool;				\
-		variable = val;						\
-	}
-#define NTFS_GETOPT_OPTIONS_ARRAY(option, variable, opt_array)		\
-	if (!strcmp(p, option)) {					\
-		int _i;							\
-		if (!v || !*v)						\
-			goto needs_arg;					\
-		ov = v;							\
-		if (variable == -1)					\
-			variable = 0;					\
-		for (_i = 0; opt_array[_i].str && *opt_array[_i].str; _i++) \
-			if (!strcmp(opt_array[_i].str, v)) {		\
-				variable |= opt_array[_i].val;		\
-				break;					\
-			}						\
-		if (!opt_array[_i].str || !*opt_array[_i].str)		\
-			goto needs_val;					\
-	}
-	if (!opt || !*opt)
-		goto no_mount_options;
-	ntfs_debug("Entering with mount options string: %s", opt);
-	while ((p = strsep(&opt, ","))) {
-		if ((v = strchr(p, '=')))
-			*v++ = 0;
-		NTFS_GETOPT_UID("uid", uid)
-		else NTFS_GETOPT_GID("gid", gid)
-		else NTFS_GETOPT_OCTAL("umask", fmask = dmask)
-		else NTFS_GETOPT_OCTAL("fmask", fmask)
-		else NTFS_GETOPT_OCTAL("dmask", dmask)
-		else NTFS_GETOPT("mft_zone_multiplier", mft_zone_multiplier)
-		else NTFS_GETOPT_WITH_DEFAULT("sloppy", sloppy, true)
-		else NTFS_GETOPT_BOOL("show_sys_files", show_sys_files)
-		else NTFS_GETOPT_BOOL("case_sensitive", case_sensitive)
-		else NTFS_GETOPT_BOOL("disable_sparse", disable_sparse)
-		else NTFS_GETOPT_OPTIONS_ARRAY("errors", on_errors,
-				on_errors_arr)
-		else if (!strcmp(p, "posix") || !strcmp(p, "show_inodes"))
-			ntfs_warning(vol->sb, "Ignoring obsolete option %s.",
-					p);
-		else if (!strcmp(p, "nls") || !strcmp(p, "iocharset")) {
-			if (!strcmp(p, "iocharset"))
-				ntfs_warning(vol->sb, "Option iocharset is "
-						"deprecated. Please use "
-						"option nls=<charsetname> in "
-						"the future.");
-			if (!v || !*v)
-				goto needs_arg;
-use_utf8:
-			old_nls = nls_map;
-			nls_map = load_nls(v);
-			if (!nls_map) {
-				if (!old_nls) {
-					ntfs_error(vol->sb, "NLS character set "
-							"%s not found.", v);
-					return false;
-				}
-				ntfs_error(vol->sb, "NLS character set %s not "
-						"found. Using previous one %s.",
-						v, old_nls->charset);
-				nls_map = old_nls;
-			} else /* nls_map */ {
-				unload_nls(old_nls);
-			}
-		} else if (!strcmp(p, "utf8")) {
-			bool val = false;
-			ntfs_warning(vol->sb, "Option utf8 is no longer "
-				   "supported, using option nls=utf8. Please "
-				   "use option nls=utf8 in the future and "
-				   "make sure utf8 is compiled either as a "
-				   "module or into the kernel.");
-			if (!v || !*v)
-				val = true;
-			else if (!simple_getbool(v, &val))
-				goto needs_bool;
-			if (val) {
-				v = utf8;
-				goto use_utf8;
-			}
-		} else {
-			ntfs_error(vol->sb, "Unrecognized mount option %s.", p);
-			if (errors < INT_MAX)
-				errors++;
-		}
-#undef NTFS_GETOPT_OPTIONS_ARRAY
-#undef NTFS_GETOPT_BOOL
-#undef NTFS_GETOPT
-#undef NTFS_GETOPT_WITH_DEFAULT
-	}
-no_mount_options:
-	if (errors && !sloppy)
-		return false;
-	if (sloppy)
-		ntfs_warning(vol->sb, "Sloppy option given. Ignoring "
-				"unrecognized mount option(s) and continuing.");
-	/* Keep this first! */
-	if (on_errors != -1) {
-		if (!on_errors) {
-			ntfs_error(vol->sb, "Invalid errors option argument "
-					"or bug in options parser.");
-			return false;
-		}
-	}
-	if (nls_map) {
-		if (vol->nls_map && vol->nls_map != nls_map) {
-			ntfs_error(vol->sb, "Cannot change NLS character set "
-					"on remount.");
-			return false;
-		} /* else (!vol->nls_map) */
-		ntfs_debug("Using NLS character set %s.", nls_map->charset);
-		vol->nls_map = nls_map;
-	} else /* (!nls_map) */ {
-		if (!vol->nls_map) {
-			vol->nls_map = load_nls_default();
-			if (!vol->nls_map) {
-				ntfs_error(vol->sb, "Failed to load default "
-						"NLS character set.");
-				return false;
-			}
-			ntfs_debug("Using default NLS character set (%s).",
-					vol->nls_map->charset);
-		}
-	}
-	if (mft_zone_multiplier != -1) {
-		if (vol->mft_zone_multiplier && vol->mft_zone_multiplier !=
-				mft_zone_multiplier) {
-			ntfs_error(vol->sb, "Cannot change mft_zone_multiplier "
-					"on remount.");
-			return false;
-		}
-		if (mft_zone_multiplier < 1 || mft_zone_multiplier > 4) {
-			ntfs_error(vol->sb, "Invalid mft_zone_multiplier. "
-					"Using default value, i.e. 1.");
-			mft_zone_multiplier = 1;
-		}
-		vol->mft_zone_multiplier = mft_zone_multiplier;
-	}
-	if (!vol->mft_zone_multiplier)
-		vol->mft_zone_multiplier = 1;
-	if (on_errors != -1)
-		vol->on_errors = on_errors;
-	if (!vol->on_errors || vol->on_errors == ON_ERRORS_RECOVER)
-		vol->on_errors |= ON_ERRORS_CONTINUE;
-	if (uid_valid(uid))
-		vol->uid = uid;
-	if (gid_valid(gid))
-		vol->gid = gid;
-	if (fmask != (umode_t)-1)
-		vol->fmask = fmask;
-	if (dmask != (umode_t)-1)
-		vol->dmask = dmask;
-	if (show_sys_files != -1) {
-		if (show_sys_files)
-			NVolSetShowSystemFiles(vol);
-		else
-			NVolClearShowSystemFiles(vol);
-	}
-	if (case_sensitive != -1) {
-		if (case_sensitive)
-			NVolSetCaseSensitive(vol);
-		else
-			NVolClearCaseSensitive(vol);
-	}
-	if (disable_sparse != -1) {
-		if (disable_sparse)
-			NVolClearSparseEnabled(vol);
-		else {
-			if (!NVolSparseEnabled(vol) &&
-					vol->major_ver && vol->major_ver < 3)
-				ntfs_warning(vol->sb, "Not enabling sparse "
-						"support due to NTFS volume "
-						"version %i.%i (need at least "
-						"version 3.0).", vol->major_ver,
-						vol->minor_ver);
-			else
-				NVolSetSparseEnabled(vol);
-		}
-	}
-	return true;
-needs_arg:
-	ntfs_error(vol->sb, "The %s option requires an argument.", p);
-	return false;
-needs_bool:
-	ntfs_error(vol->sb, "The %s option requires a boolean argument.", p);
-	return false;
-needs_val:
-	ntfs_error(vol->sb, "Invalid %s option argument: %s", p, ov);
-	return false;
-}
-
-#ifdef NTFS_RW
-
-/**
- * ntfs_write_volume_flags - write new flags to the volume information flags
- * @vol:	ntfs volume on which to modify the flags
- * @flags:	new flags value for the volume information flags
- *
- * Internal function.  You probably want to use ntfs_{set,clear}_volume_flags()
- * instead (see below).
- *
- * Replace the volume information flags on the volume @vol with the value
- * supplied in @flags.  Note, this overwrites the volume information flags, so
- * make sure to combine the flags you want to modify with the old flags and use
- * the result when calling ntfs_write_volume_flags().
- *
- * Return 0 on success and -errno on error.
- */
-static int ntfs_write_volume_flags(ntfs_volume *vol, const VOLUME_FLAGS flags)
-{
-	ntfs_inode *ni = NTFS_I(vol->vol_ino);
-	MFT_RECORD *m;
-	VOLUME_INFORMATION *vi;
-	ntfs_attr_search_ctx *ctx;
-	int err;
-
-	ntfs_debug("Entering, old flags = 0x%x, new flags = 0x%x.",
-			le16_to_cpu(vol->vol_flags), le16_to_cpu(flags));
-	if (vol->vol_flags == flags)
-		goto done;
-	BUG_ON(!ni);
-	m = map_mft_record(ni);
-	if (IS_ERR(m)) {
-		err = PTR_ERR(m);
-		goto err_out;
-	}
-	ctx = ntfs_attr_get_search_ctx(ni, m);
-	if (!ctx) {
-		err = -ENOMEM;
-		goto put_unm_err_out;
-	}
-	err = ntfs_attr_lookup(AT_VOLUME_INFORMATION, NULL, 0, 0, 0, NULL, 0,
-			ctx);
-	if (err)
-		goto put_unm_err_out;
-	vi = (VOLUME_INFORMATION*)((u8*)ctx->attr +
-			le16_to_cpu(ctx->attr->data.resident.value_offset));
-	vol->vol_flags = vi->flags = flags;
-	flush_dcache_mft_record_page(ctx->ntfs_ino);
-	mark_mft_record_dirty(ctx->ntfs_ino);
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(ni);
-done:
-	ntfs_debug("Done.");
-	return 0;
-put_unm_err_out:
-	if (ctx)
-		ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(ni);
-err_out:
-	ntfs_error(vol->sb, "Failed with error code %i.", -err);
-	return err;
-}
-
-/**
- * ntfs_set_volume_flags - set bits in the volume information flags
- * @vol:	ntfs volume on which to modify the flags
- * @flags:	flags to set on the volume
- *
- * Set the bits in @flags in the volume information flags on the volume @vol.
- *
- * Return 0 on success and -errno on error.
- */
-static inline int ntfs_set_volume_flags(ntfs_volume *vol, VOLUME_FLAGS flags)
-{
-	flags &= VOLUME_FLAGS_MASK;
-	return ntfs_write_volume_flags(vol, vol->vol_flags | flags);
-}
-
-/**
- * ntfs_clear_volume_flags - clear bits in the volume information flags
- * @vol:	ntfs volume on which to modify the flags
- * @flags:	flags to clear on the volume
- *
- * Clear the bits in @flags in the volume information flags on the volume @vol.
- *
- * Return 0 on success and -errno on error.
- */
-static inline int ntfs_clear_volume_flags(ntfs_volume *vol, VOLUME_FLAGS flags)
-{
-	flags &= VOLUME_FLAGS_MASK;
-	flags = vol->vol_flags & cpu_to_le16(~le16_to_cpu(flags));
-	return ntfs_write_volume_flags(vol, flags);
-}
-
-#endif /* NTFS_RW */
-
-/**
- * ntfs_remount - change the mount options of a mounted ntfs filesystem
- * @sb:		superblock of mounted ntfs filesystem
- * @flags:	remount flags
- * @opt:	remount options string
- *
- * Change the mount options of an already mounted ntfs filesystem.
- *
- * NOTE:  The VFS sets the @sb->s_flags remount flags to @flags after
- * ntfs_remount() returns successfully (i.e. returns 0).  Otherwise,
- * @sb->s_flags are not changed.
- */
-static int ntfs_remount(struct super_block *sb, int *flags, char *opt)
-{
-	ntfs_volume *vol = NTFS_SB(sb);
-
-	ntfs_debug("Entering with remount options string: %s", opt);
-
-	sync_filesystem(sb);
-
-#ifndef NTFS_RW
-	/* For read-only compiled driver, enforce read-only flag. */
-	*flags |= SB_RDONLY;
-#else /* NTFS_RW */
-	/*
-	 * For the read-write compiled driver, if we are remounting read-write,
-	 * make sure there are no volume errors and that no unsupported volume
-	 * flags are set.  Also, empty the logfile journal as it would become
-	 * stale as soon as something is written to the volume and mark the
-	 * volume dirty so that chkdsk is run if the volume is not umounted
-	 * cleanly.  Finally, mark the quotas out of date so Windows rescans
-	 * the volume on boot and updates them.
-	 *
-	 * When remounting read-only, mark the volume clean if no volume errors
-	 * have occurred.
-	 */
-	if (sb_rdonly(sb) && !(*flags & SB_RDONLY)) {
-		static const char *es = ".  Cannot remount read-write.";
-
-		/* Remounting read-write. */
-		if (NVolErrors(vol)) {
-			ntfs_error(sb, "Volume has errors and is read-only%s",
-					es);
-			return -EROFS;
-		}
-		if (vol->vol_flags & VOLUME_IS_DIRTY) {
-			ntfs_error(sb, "Volume is dirty and read-only%s", es);
-			return -EROFS;
-		}
-		if (vol->vol_flags & VOLUME_MODIFIED_BY_CHKDSK) {
-			ntfs_error(sb, "Volume has been modified by chkdsk "
-					"and is read-only%s", es);
-			return -EROFS;
-		}
-		if (vol->vol_flags & VOLUME_MUST_MOUNT_RO_MASK) {
-			ntfs_error(sb, "Volume has unsupported flags set "
-					"(0x%x) and is read-only%s",
-					(unsigned)le16_to_cpu(vol->vol_flags),
-					es);
-			return -EROFS;
-		}
-		if (ntfs_set_volume_flags(vol, VOLUME_IS_DIRTY)) {
-			ntfs_error(sb, "Failed to set dirty bit in volume "
-					"information flags%s", es);
-			return -EROFS;
-		}
-#if 0
-		// TODO: Enable this code once we start modifying anything that
-		//	 is different between NTFS 1.2 and 3.x...
-		/* Set NT4 compatibility flag on newer NTFS version volumes. */
-		if ((vol->major_ver > 1)) {
-			if (ntfs_set_volume_flags(vol, VOLUME_MOUNTED_ON_NT4)) {
-				ntfs_error(sb, "Failed to set NT4 "
-						"compatibility flag%s", es);
-				NVolSetErrors(vol);
-				return -EROFS;
-			}
-		}
-#endif
-		if (!ntfs_empty_logfile(vol->logfile_ino)) {
-			ntfs_error(sb, "Failed to empty journal $LogFile%s",
-					es);
-			NVolSetErrors(vol);
-			return -EROFS;
-		}
-		if (!ntfs_mark_quotas_out_of_date(vol)) {
-			ntfs_error(sb, "Failed to mark quotas out of date%s",
-					es);
-			NVolSetErrors(vol);
-			return -EROFS;
-		}
-		if (!ntfs_stamp_usnjrnl(vol)) {
-			ntfs_error(sb, "Failed to stamp transaction log "
-					"($UsnJrnl)%s", es);
-			NVolSetErrors(vol);
-			return -EROFS;
-		}
-	} else if (!sb_rdonly(sb) && (*flags & SB_RDONLY)) {
-		/* Remounting read-only. */
-		if (!NVolErrors(vol)) {
-			if (ntfs_clear_volume_flags(vol, VOLUME_IS_DIRTY))
-				ntfs_warning(sb, "Failed to clear dirty bit "
-						"in volume information "
-						"flags.  Run chkdsk.");
-		}
-	}
-#endif /* NTFS_RW */
-
-	// TODO: Deal with *flags.
-
-	if (!parse_options(vol, opt))
-		return -EINVAL;
-
-	ntfs_debug("Done.");
-	return 0;
-}
-
-/**
- * is_boot_sector_ntfs - check whether a boot sector is a valid NTFS boot sector
- * @sb:		Super block of the device to which @b belongs.
- * @b:		Boot sector of device @sb to check.
- * @silent:	If 'true', all output will be silenced.
- *
- * is_boot_sector_ntfs() checks whether the boot sector @b is a valid NTFS boot
- * sector. Returns 'true' if it is valid and 'false' if not.
- *
- * @sb is only needed for warning/error output, i.e. it can be NULL when silent
- * is 'true'.
- */
-static bool is_boot_sector_ntfs(const struct super_block *sb,
-		const NTFS_BOOT_SECTOR *b, const bool silent)
-{
-	/*
-	 * Check that checksum == sum of u32 values from b to the checksum
-	 * field.  If checksum is zero, no checking is done.  We will work when
-	 * the checksum test fails, since some utilities update the boot sector
-	 * ignoring the checksum which leaves the checksum out-of-date.  We
-	 * report a warning if this is the case.
-	 */
-	if ((void*)b < (void*)&b->checksum && b->checksum && !silent) {
-		le32 *u;
-		u32 i;
-
-		for (i = 0, u = (le32*)b; u < (le32*)(&b->checksum); ++u)
-			i += le32_to_cpup(u);
-		if (le32_to_cpu(b->checksum) != i)
-			ntfs_warning(sb, "Invalid boot sector checksum.");
-	}
-	/* Check OEMidentifier is "NTFS    " */
-	if (b->oem_id != magicNTFS)
-		goto not_ntfs;
-	/* Check bytes per sector value is between 256 and 4096. */
-	if (le16_to_cpu(b->bpb.bytes_per_sector) < 0x100 ||
-			le16_to_cpu(b->bpb.bytes_per_sector) > 0x1000)
-		goto not_ntfs;
-	/* Check sectors per cluster value is valid. */
-	switch (b->bpb.sectors_per_cluster) {
-	case 1: case 2: case 4: case 8: case 16: case 32: case 64: case 128:
-		break;
-	default:
-		goto not_ntfs;
-	}
-	/* Check the cluster size is not above the maximum (64kiB). */
-	if ((u32)le16_to_cpu(b->bpb.bytes_per_sector) *
-			b->bpb.sectors_per_cluster > NTFS_MAX_CLUSTER_SIZE)
-		goto not_ntfs;
-	/* Check reserved/unused fields are really zero. */
-	if (le16_to_cpu(b->bpb.reserved_sectors) ||
-			le16_to_cpu(b->bpb.root_entries) ||
-			le16_to_cpu(b->bpb.sectors) ||
-			le16_to_cpu(b->bpb.sectors_per_fat) ||
-			le32_to_cpu(b->bpb.large_sectors) || b->bpb.fats)
-		goto not_ntfs;
-	/* Check clusters per file mft record value is valid. */
-	if ((u8)b->clusters_per_mft_record < 0xe1 ||
-			(u8)b->clusters_per_mft_record > 0xf7)
-		switch (b->clusters_per_mft_record) {
-		case 1: case 2: case 4: case 8: case 16: case 32: case 64:
-			break;
-		default:
-			goto not_ntfs;
-		}
-	/* Check clusters per index block value is valid. */
-	if ((u8)b->clusters_per_index_record < 0xe1 ||
-			(u8)b->clusters_per_index_record > 0xf7)
-		switch (b->clusters_per_index_record) {
-		case 1: case 2: case 4: case 8: case 16: case 32: case 64:
-			break;
-		default:
-			goto not_ntfs;
-		}
-	/*
-	 * Check for valid end of sector marker. We will work without it, but
-	 * many BIOSes will refuse to boot from a bootsector if the magic is
-	 * incorrect, so we emit a warning.
-	 */
-	if (!silent && b->end_of_sector_marker != cpu_to_le16(0xaa55))
-		ntfs_warning(sb, "Invalid end of sector marker.");
-	return true;
-not_ntfs:
-	return false;
-}
-
-/**
- * read_ntfs_boot_sector - read the NTFS boot sector of a device
- * @sb:		super block of device to read the boot sector from
- * @silent:	if true, suppress all output
- *
- * Reads the boot sector from the device and validates it. If that fails, tries
- * to read the backup boot sector, first from the end of the device a-la NT4 and
- * later and then from the middle of the device a-la NT3.51 and before.
- *
- * If a valid boot sector is found but it is not the primary boot sector, we
- * repair the primary boot sector silently (unless the device is read-only or
- * the primary boot sector is not accessible).
- *
- * NOTE: To call this function, @sb must have the fields s_dev, the ntfs super
- * block (u.ntfs_sb), nr_blocks and the device flags (s_flags) initialized
- * to their respective values.
- *
- * Return the unlocked buffer head containing the boot sector or NULL on error.
- */
-static struct buffer_head *read_ntfs_boot_sector(struct super_block *sb,
-		const int silent)
-{
-	const char *read_err_str = "Unable to read %s boot sector.";
-	struct buffer_head *bh_primary, *bh_backup;
-	sector_t nr_blocks = NTFS_SB(sb)->nr_blocks;
-
-	/* Try to read primary boot sector. */
-	if ((bh_primary = sb_bread(sb, 0))) {
-		if (is_boot_sector_ntfs(sb, (NTFS_BOOT_SECTOR*)
-				bh_primary->b_data, silent))
-			return bh_primary;
-		if (!silent)
-			ntfs_error(sb, "Primary boot sector is invalid.");
-	} else if (!silent)
-		ntfs_error(sb, read_err_str, "primary");
-	if (!(NTFS_SB(sb)->on_errors & ON_ERRORS_RECOVER)) {
-		if (bh_primary)
-			brelse(bh_primary);
-		if (!silent)
-			ntfs_error(sb, "Mount option errors=recover not used. "
-					"Aborting without trying to recover.");
-		return NULL;
-	}
-	/* Try to read NT4+ backup boot sector. */
-	if ((bh_backup = sb_bread(sb, nr_blocks - 1))) {
-		if (is_boot_sector_ntfs(sb, (NTFS_BOOT_SECTOR*)
-				bh_backup->b_data, silent))
-			goto hotfix_primary_boot_sector;
-		brelse(bh_backup);
-	} else if (!silent)
-		ntfs_error(sb, read_err_str, "backup");
-	/* Try to read NT3.51- backup boot sector. */
-	if ((bh_backup = sb_bread(sb, nr_blocks >> 1))) {
-		if (is_boot_sector_ntfs(sb, (NTFS_BOOT_SECTOR*)
-				bh_backup->b_data, silent))
-			goto hotfix_primary_boot_sector;
-		if (!silent)
-			ntfs_error(sb, "Could not find a valid backup boot "
-					"sector.");
-		brelse(bh_backup);
-	} else if (!silent)
-		ntfs_error(sb, read_err_str, "backup");
-	/* We failed. Cleanup and return. */
-	if (bh_primary)
-		brelse(bh_primary);
-	return NULL;
-hotfix_primary_boot_sector:
-	if (bh_primary) {
-		/*
-		 * If we managed to read sector zero and the volume is not
-		 * read-only, copy the found, valid backup boot sector to the
-		 * primary boot sector.  Note we only copy the actual boot
-		 * sector structure, not the actual whole device sector as that
-		 * may be bigger and would potentially damage the $Boot system
-		 * file (FIXME: Would be nice to know if the backup boot sector
-		 * on a large sector device contains the whole boot loader or
-		 * just the first 512 bytes).
-		 */
-		if (!sb_rdonly(sb)) {
-			ntfs_warning(sb, "Hot-fix: Recovering invalid primary "
-					"boot sector from backup copy.");
-			memcpy(bh_primary->b_data, bh_backup->b_data,
-					NTFS_BLOCK_SIZE);
-			mark_buffer_dirty(bh_primary);
-			sync_dirty_buffer(bh_primary);
-			if (buffer_uptodate(bh_primary)) {
-				brelse(bh_backup);
-				return bh_primary;
-			}
-			ntfs_error(sb, "Hot-fix: Device write error while "
-					"recovering primary boot sector.");
-		} else {
-			ntfs_warning(sb, "Hot-fix: Recovery of primary boot "
-					"sector failed: Read-only mount.");
-		}
-		brelse(bh_primary);
-	}
-	ntfs_warning(sb, "Using backup boot sector.");
-	return bh_backup;
-}
-
-/**
- * parse_ntfs_boot_sector - parse the boot sector and store the data in @vol
- * @vol:	volume structure to initialise with data from boot sector
- * @b:		boot sector to parse
- *
- * Parse the ntfs boot sector @b and store all imporant information therein in
- * the ntfs super block @vol.  Return 'true' on success and 'false' on error.
- */
-static bool parse_ntfs_boot_sector(ntfs_volume *vol, const NTFS_BOOT_SECTOR *b)
-{
-	unsigned int sectors_per_cluster_bits, nr_hidden_sects;
-	int clusters_per_mft_record, clusters_per_index_record;
-	s64 ll;
-
-	vol->sector_size = le16_to_cpu(b->bpb.bytes_per_sector);
-	vol->sector_size_bits = ffs(vol->sector_size) - 1;
-	ntfs_debug("vol->sector_size = %i (0x%x)", vol->sector_size,
-			vol->sector_size);
-	ntfs_debug("vol->sector_size_bits = %i (0x%x)", vol->sector_size_bits,
-			vol->sector_size_bits);
-	if (vol->sector_size < vol->sb->s_blocksize) {
-		ntfs_error(vol->sb, "Sector size (%i) is smaller than the "
-				"device block size (%lu).  This is not "
-				"supported.  Sorry.", vol->sector_size,
-				vol->sb->s_blocksize);
-		return false;
-	}
-	ntfs_debug("sectors_per_cluster = 0x%x", b->bpb.sectors_per_cluster);
-	sectors_per_cluster_bits = ffs(b->bpb.sectors_per_cluster) - 1;
-	ntfs_debug("sectors_per_cluster_bits = 0x%x",
-			sectors_per_cluster_bits);
-	nr_hidden_sects = le32_to_cpu(b->bpb.hidden_sectors);
-	ntfs_debug("number of hidden sectors = 0x%x", nr_hidden_sects);
-	vol->cluster_size = vol->sector_size << sectors_per_cluster_bits;
-	vol->cluster_size_mask = vol->cluster_size - 1;
-	vol->cluster_size_bits = ffs(vol->cluster_size) - 1;
-	ntfs_debug("vol->cluster_size = %i (0x%x)", vol->cluster_size,
-			vol->cluster_size);
-	ntfs_debug("vol->cluster_size_mask = 0x%x", vol->cluster_size_mask);
-	ntfs_debug("vol->cluster_size_bits = %i", vol->cluster_size_bits);
-	if (vol->cluster_size < vol->sector_size) {
-		ntfs_error(vol->sb, "Cluster size (%i) is smaller than the "
-				"sector size (%i).  This is not supported.  "
-				"Sorry.", vol->cluster_size, vol->sector_size);
-		return false;
-	}
-	clusters_per_mft_record = b->clusters_per_mft_record;
-	ntfs_debug("clusters_per_mft_record = %i (0x%x)",
-			clusters_per_mft_record, clusters_per_mft_record);
-	if (clusters_per_mft_record > 0)
-		vol->mft_record_size = vol->cluster_size <<
-				(ffs(clusters_per_mft_record) - 1);
-	else
-		/*
-		 * When mft_record_size < cluster_size, clusters_per_mft_record
-		 * = -log2(mft_record_size) bytes. mft_record_size normaly is
-		 * 1024 bytes, which is encoded as 0xF6 (-10 in decimal).
-		 */
-		vol->mft_record_size = 1 << -clusters_per_mft_record;
-	vol->mft_record_size_mask = vol->mft_record_size - 1;
-	vol->mft_record_size_bits = ffs(vol->mft_record_size) - 1;
-	ntfs_debug("vol->mft_record_size = %i (0x%x)", vol->mft_record_size,
-			vol->mft_record_size);
-	ntfs_debug("vol->mft_record_size_mask = 0x%x",
-			vol->mft_record_size_mask);
-	ntfs_debug("vol->mft_record_size_bits = %i (0x%x)",
-			vol->mft_record_size_bits, vol->mft_record_size_bits);
-	/*
-	 * We cannot support mft record sizes above the PAGE_SIZE since
-	 * we store $MFT/$DATA, the table of mft records in the page cache.
-	 */
-	if (vol->mft_record_size > PAGE_SIZE) {
-		ntfs_error(vol->sb, "Mft record size (%i) exceeds the "
-				"PAGE_SIZE on your system (%lu).  "
-				"This is not supported.  Sorry.",
-				vol->mft_record_size, PAGE_SIZE);
-		return false;
-	}
-	/* We cannot support mft record sizes below the sector size. */
-	if (vol->mft_record_size < vol->sector_size) {
-		ntfs_error(vol->sb, "Mft record size (%i) is smaller than the "
-				"sector size (%i).  This is not supported.  "
-				"Sorry.", vol->mft_record_size,
-				vol->sector_size);
-		return false;
-	}
-	clusters_per_index_record = b->clusters_per_index_record;
-	ntfs_debug("clusters_per_index_record = %i (0x%x)",
-			clusters_per_index_record, clusters_per_index_record);
-	if (clusters_per_index_record > 0)
-		vol->index_record_size = vol->cluster_size <<
-				(ffs(clusters_per_index_record) - 1);
-	else
-		/*
-		 * When index_record_size < cluster_size,
-		 * clusters_per_index_record = -log2(index_record_size) bytes.
-		 * index_record_size normaly equals 4096 bytes, which is
-		 * encoded as 0xF4 (-12 in decimal).
-		 */
-		vol->index_record_size = 1 << -clusters_per_index_record;
-	vol->index_record_size_mask = vol->index_record_size - 1;
-	vol->index_record_size_bits = ffs(vol->index_record_size) - 1;
-	ntfs_debug("vol->index_record_size = %i (0x%x)",
-			vol->index_record_size, vol->index_record_size);
-	ntfs_debug("vol->index_record_size_mask = 0x%x",
-			vol->index_record_size_mask);
-	ntfs_debug("vol->index_record_size_bits = %i (0x%x)",
-			vol->index_record_size_bits,
-			vol->index_record_size_bits);
-	/* We cannot support index record sizes below the sector size. */
-	if (vol->index_record_size < vol->sector_size) {
-		ntfs_error(vol->sb, "Index record size (%i) is smaller than "
-				"the sector size (%i).  This is not "
-				"supported.  Sorry.", vol->index_record_size,
-				vol->sector_size);
-		return false;
-	}
-	/*
-	 * Get the size of the volume in clusters and check for 64-bit-ness.
-	 * Windows currently only uses 32 bits to save the clusters so we do
-	 * the same as it is much faster on 32-bit CPUs.
-	 */
-	ll = sle64_to_cpu(b->number_of_sectors) >> sectors_per_cluster_bits;
-	if ((u64)ll >= 1ULL << 32) {
-		ntfs_error(vol->sb, "Cannot handle 64-bit clusters.  Sorry.");
-		return false;
-	}
-	vol->nr_clusters = ll;
-	ntfs_debug("vol->nr_clusters = 0x%llx", (long long)vol->nr_clusters);
-	/*
-	 * On an architecture where unsigned long is 32-bits, we restrict the
-	 * volume size to 2TiB (2^41). On a 64-bit architecture, the compiler
-	 * will hopefully optimize the whole check away.
-	 */
-	if (sizeof(unsigned long) < 8) {
-		if ((ll << vol->cluster_size_bits) >= (1ULL << 41)) {
-			ntfs_error(vol->sb, "Volume size (%lluTiB) is too "
-					"large for this architecture.  "
-					"Maximum supported is 2TiB.  Sorry.",
-					(unsigned long long)ll >> (40 -
-					vol->cluster_size_bits));
-			return false;
-		}
-	}
-	ll = sle64_to_cpu(b->mft_lcn);
-	if (ll >= vol->nr_clusters) {
-		ntfs_error(vol->sb, "MFT LCN (%lli, 0x%llx) is beyond end of "
-				"volume.  Weird.", (unsigned long long)ll,
-				(unsigned long long)ll);
-		return false;
-	}
-	vol->mft_lcn = ll;
-	ntfs_debug("vol->mft_lcn = 0x%llx", (long long)vol->mft_lcn);
-	ll = sle64_to_cpu(b->mftmirr_lcn);
-	if (ll >= vol->nr_clusters) {
-		ntfs_error(vol->sb, "MFTMirr LCN (%lli, 0x%llx) is beyond end "
-				"of volume.  Weird.", (unsigned long long)ll,
-				(unsigned long long)ll);
-		return false;
-	}
-	vol->mftmirr_lcn = ll;
-	ntfs_debug("vol->mftmirr_lcn = 0x%llx", (long long)vol->mftmirr_lcn);
-#ifdef NTFS_RW
-	/*
-	 * Work out the size of the mft mirror in number of mft records. If the
-	 * cluster size is less than or equal to the size taken by four mft
-	 * records, the mft mirror stores the first four mft records. If the
-	 * cluster size is bigger than the size taken by four mft records, the
-	 * mft mirror contains as many mft records as will fit into one
-	 * cluster.
-	 */
-	if (vol->cluster_size <= (4 << vol->mft_record_size_bits))
-		vol->mftmirr_size = 4;
-	else
-		vol->mftmirr_size = vol->cluster_size >>
-				vol->mft_record_size_bits;
-	ntfs_debug("vol->mftmirr_size = %i", vol->mftmirr_size);
-#endif /* NTFS_RW */
-	vol->serial_no = le64_to_cpu(b->volume_serial_number);
-	ntfs_debug("vol->serial_no = 0x%llx",
-			(unsigned long long)vol->serial_no);
-	return true;
-}
-
-/**
- * ntfs_setup_allocators - initialize the cluster and mft allocators
- * @vol:	volume structure for which to setup the allocators
- *
- * Setup the cluster (lcn) and mft allocators to the starting values.
- */
-static void ntfs_setup_allocators(ntfs_volume *vol)
-{
-#ifdef NTFS_RW
-	LCN mft_zone_size, mft_lcn;
-#endif /* NTFS_RW */
-
-	ntfs_debug("vol->mft_zone_multiplier = 0x%x",
-			vol->mft_zone_multiplier);
-#ifdef NTFS_RW
-	/* Determine the size of the MFT zone. */
-	mft_zone_size = vol->nr_clusters;
-	switch (vol->mft_zone_multiplier) {  /* % of volume size in clusters */
-	case 4:
-		mft_zone_size >>= 1;			/* 50%   */
-		break;
-	case 3:
-		mft_zone_size = (mft_zone_size +
-				(mft_zone_size >> 1)) >> 2;	/* 37.5% */
-		break;
-	case 2:
-		mft_zone_size >>= 2;			/* 25%   */
-		break;
-	/* case 1: */
-	default:
-		mft_zone_size >>= 3;			/* 12.5% */
-		break;
-	}
-	/* Setup the mft zone. */
-	vol->mft_zone_start = vol->mft_zone_pos = vol->mft_lcn;
-	ntfs_debug("vol->mft_zone_pos = 0x%llx",
-			(unsigned long long)vol->mft_zone_pos);
-	/*
-	 * Calculate the mft_lcn for an unmodified NTFS volume (see mkntfs
-	 * source) and if the actual mft_lcn is in the expected place or even
-	 * further to the front of the volume, extend the mft_zone to cover the
-	 * beginning of the volume as well.  This is in order to protect the
-	 * area reserved for the mft bitmap as well within the mft_zone itself.
-	 * On non-standard volumes we do not protect it as the overhead would
-	 * be higher than the speed increase we would get by doing it.
-	 */
-	mft_lcn = (8192 + 2 * vol->cluster_size - 1) / vol->cluster_size;
-	if (mft_lcn * vol->cluster_size < 16 * 1024)
-		mft_lcn = (16 * 1024 + vol->cluster_size - 1) /
-				vol->cluster_size;
-	if (vol->mft_zone_start <= mft_lcn)
-		vol->mft_zone_start = 0;
-	ntfs_debug("vol->mft_zone_start = 0x%llx",
-			(unsigned long long)vol->mft_zone_start);
-	/*
-	 * Need to cap the mft zone on non-standard volumes so that it does
-	 * not point outside the boundaries of the volume.  We do this by
-	 * halving the zone size until we are inside the volume.
-	 */
-	vol->mft_zone_end = vol->mft_lcn + mft_zone_size;
-	while (vol->mft_zone_end >= vol->nr_clusters) {
-		mft_zone_size >>= 1;
-		vol->mft_zone_end = vol->mft_lcn + mft_zone_size;
-	}
-	ntfs_debug("vol->mft_zone_end = 0x%llx",
-			(unsigned long long)vol->mft_zone_end);
-	/*
-	 * Set the current position within each data zone to the start of the
-	 * respective zone.
-	 */
-	vol->data1_zone_pos = vol->mft_zone_end;
-	ntfs_debug("vol->data1_zone_pos = 0x%llx",
-			(unsigned long long)vol->data1_zone_pos);
-	vol->data2_zone_pos = 0;
-	ntfs_debug("vol->data2_zone_pos = 0x%llx",
-			(unsigned long long)vol->data2_zone_pos);
-
-	/* Set the mft data allocation position to mft record 24. */
-	vol->mft_data_pos = 24;
-	ntfs_debug("vol->mft_data_pos = 0x%llx",
-			(unsigned long long)vol->mft_data_pos);
-#endif /* NTFS_RW */
-}
-
-#ifdef NTFS_RW
-
-/**
- * load_and_init_mft_mirror - load and setup the mft mirror inode for a volume
- * @vol:	ntfs super block describing device whose mft mirror to load
- *
- * Return 'true' on success or 'false' on error.
- */
-static bool load_and_init_mft_mirror(ntfs_volume *vol)
-{
-	struct inode *tmp_ino;
-	ntfs_inode *tmp_ni;
-
-	ntfs_debug("Entering.");
-	/* Get mft mirror inode. */
-	tmp_ino = ntfs_iget(vol->sb, FILE_MFTMirr);
-	if (IS_ERR(tmp_ino) || is_bad_inode(tmp_ino)) {
-		if (!IS_ERR(tmp_ino))
-			iput(tmp_ino);
-		/* Caller will display error message. */
-		return false;
-	}
-	/*
-	 * Re-initialize some specifics about $MFTMirr's inode as
-	 * ntfs_read_inode() will have set up the default ones.
-	 */
-	/* Set uid and gid to root. */
-	tmp_ino->i_uid = GLOBAL_ROOT_UID;
-	tmp_ino->i_gid = GLOBAL_ROOT_GID;
-	/* Regular file.  No access for anyone. */
-	tmp_ino->i_mode = S_IFREG;
-	/* No VFS initiated operations allowed for $MFTMirr. */
-	tmp_ino->i_op = &ntfs_empty_inode_ops;
-	tmp_ino->i_fop = &ntfs_empty_file_ops;
-	/* Put in our special address space operations. */
-	tmp_ino->i_mapping->a_ops = &ntfs_mst_aops;
-	tmp_ni = NTFS_I(tmp_ino);
-	/* The $MFTMirr, like the $MFT is multi sector transfer protected. */
-	NInoSetMstProtected(tmp_ni);
-	NInoSetSparseDisabled(tmp_ni);
-	/*
-	 * Set up our little cheat allowing us to reuse the async read io
-	 * completion handler for directories.
-	 */
-	tmp_ni->itype.index.block_size = vol->mft_record_size;
-	tmp_ni->itype.index.block_size_bits = vol->mft_record_size_bits;
-	vol->mftmirr_ino = tmp_ino;
-	ntfs_debug("Done.");
-	return true;
-}
-
-/**
- * check_mft_mirror - compare contents of the mft mirror with the mft
- * @vol:	ntfs super block describing device whose mft mirror to check
- *
- * Return 'true' on success or 'false' on error.
- *
- * Note, this function also results in the mft mirror runlist being completely
- * mapped into memory.  The mft mirror write code requires this and will BUG()
- * should it find an unmapped runlist element.
- */
-static bool check_mft_mirror(ntfs_volume *vol)
-{
-	struct super_block *sb = vol->sb;
-	ntfs_inode *mirr_ni;
-	struct page *mft_page, *mirr_page;
-	u8 *kmft, *kmirr;
-	runlist_element *rl, rl2[2];
-	pgoff_t index;
-	int mrecs_per_page, i;
-
-	ntfs_debug("Entering.");
-	/* Compare contents of $MFT and $MFTMirr. */
-	mrecs_per_page = PAGE_SIZE / vol->mft_record_size;
-	BUG_ON(!mrecs_per_page);
-	BUG_ON(!vol->mftmirr_size);
-	mft_page = mirr_page = NULL;
-	kmft = kmirr = NULL;
-	index = i = 0;
-	do {
-		u32 bytes;
-
-		/* Switch pages if necessary. */
-		if (!(i % mrecs_per_page)) {
-			if (index) {
-				ntfs_unmap_page(mft_page);
-				ntfs_unmap_page(mirr_page);
-			}
-			/* Get the $MFT page. */
-			mft_page = ntfs_map_page(vol->mft_ino->i_mapping,
-					index);
-			if (IS_ERR(mft_page)) {
-				ntfs_error(sb, "Failed to read $MFT.");
-				return false;
-			}
-			kmft = page_address(mft_page);
-			/* Get the $MFTMirr page. */
-			mirr_page = ntfs_map_page(vol->mftmirr_ino->i_mapping,
-					index);
-			if (IS_ERR(mirr_page)) {
-				ntfs_error(sb, "Failed to read $MFTMirr.");
-				goto mft_unmap_out;
-			}
-			kmirr = page_address(mirr_page);
-			++index;
-		}
-		/* Do not check the record if it is not in use. */
-		if (((MFT_RECORD*)kmft)->flags & MFT_RECORD_IN_USE) {
-			/* Make sure the record is ok. */
-			if (ntfs_is_baad_recordp((le32*)kmft)) {
-				ntfs_error(sb, "Incomplete multi sector "
-						"transfer detected in mft "
-						"record %i.", i);
-mm_unmap_out:
-				ntfs_unmap_page(mirr_page);
-mft_unmap_out:
-				ntfs_unmap_page(mft_page);
-				return false;
-			}
-		}
-		/* Do not check the mirror record if it is not in use. */
-		if (((MFT_RECORD*)kmirr)->flags & MFT_RECORD_IN_USE) {
-			if (ntfs_is_baad_recordp((le32*)kmirr)) {
-				ntfs_error(sb, "Incomplete multi sector "
-						"transfer detected in mft "
-						"mirror record %i.", i);
-				goto mm_unmap_out;
-			}
-		}
-		/* Get the amount of data in the current record. */
-		bytes = le32_to_cpu(((MFT_RECORD*)kmft)->bytes_in_use);
-		if (bytes < sizeof(MFT_RECORD_OLD) ||
-				bytes > vol->mft_record_size ||
-				ntfs_is_baad_recordp((le32*)kmft)) {
-			bytes = le32_to_cpu(((MFT_RECORD*)kmirr)->bytes_in_use);
-			if (bytes < sizeof(MFT_RECORD_OLD) ||
-					bytes > vol->mft_record_size ||
-					ntfs_is_baad_recordp((le32*)kmirr))
-				bytes = vol->mft_record_size;
-		}
-		/* Compare the two records. */
-		if (memcmp(kmft, kmirr, bytes)) {
-			ntfs_error(sb, "$MFT and $MFTMirr (record %i) do not "
-					"match.  Run ntfsfix or chkdsk.", i);
-			goto mm_unmap_out;
-		}
-		kmft += vol->mft_record_size;
-		kmirr += vol->mft_record_size;
-	} while (++i < vol->mftmirr_size);
-	/* Release the last pages. */
-	ntfs_unmap_page(mft_page);
-	ntfs_unmap_page(mirr_page);
-
-	/* Construct the mft mirror runlist by hand. */
-	rl2[0].vcn = 0;
-	rl2[0].lcn = vol->mftmirr_lcn;
-	rl2[0].length = (vol->mftmirr_size * vol->mft_record_size +
-			vol->cluster_size - 1) / vol->cluster_size;
-	rl2[1].vcn = rl2[0].length;
-	rl2[1].lcn = LCN_ENOENT;
-	rl2[1].length = 0;
-	/*
-	 * Because we have just read all of the mft mirror, we know we have
-	 * mapped the full runlist for it.
-	 */
-	mirr_ni = NTFS_I(vol->mftmirr_ino);
-	down_read(&mirr_ni->runlist.lock);
-	rl = mirr_ni->runlist.rl;
-	/* Compare the two runlists.  They must be identical. */
-	i = 0;
-	do {
-		if (rl2[i].vcn != rl[i].vcn || rl2[i].lcn != rl[i].lcn ||
-				rl2[i].length != rl[i].length) {
-			ntfs_error(sb, "$MFTMirr location mismatch.  "
-					"Run chkdsk.");
-			up_read(&mirr_ni->runlist.lock);
-			return false;
-		}
-	} while (rl2[i++].length);
-	up_read(&mirr_ni->runlist.lock);
-	ntfs_debug("Done.");
-	return true;
-}
-
-/**
- * load_and_check_logfile - load and check the logfile inode for a volume
- * @vol:	ntfs super block describing device whose logfile to load
- *
- * Return 'true' on success or 'false' on error.
- */
-static bool load_and_check_logfile(ntfs_volume *vol,
-		RESTART_PAGE_HEADER **rp)
-{
-	struct inode *tmp_ino;
-
-	ntfs_debug("Entering.");
-	tmp_ino = ntfs_iget(vol->sb, FILE_LogFile);
-	if (IS_ERR(tmp_ino) || is_bad_inode(tmp_ino)) {
-		if (!IS_ERR(tmp_ino))
-			iput(tmp_ino);
-		/* Caller will display error message. */
-		return false;
-	}
-	if (!ntfs_check_logfile(tmp_ino, rp)) {
-		iput(tmp_ino);
-		/* ntfs_check_logfile() will have displayed error output. */
-		return false;
-	}
-	NInoSetSparseDisabled(NTFS_I(tmp_ino));
-	vol->logfile_ino = tmp_ino;
-	ntfs_debug("Done.");
-	return true;
-}
-
-#define NTFS_HIBERFIL_HEADER_SIZE	4096
-
-/**
- * check_windows_hibernation_status - check if Windows is suspended on a volume
- * @vol:	ntfs super block of device to check
- *
- * Check if Windows is hibernated on the ntfs volume @vol.  This is done by
- * looking for the file hiberfil.sys in the root directory of the volume.  If
- * the file is not present Windows is definitely not suspended.
- *
- * If hiberfil.sys exists and is less than 4kiB in size it means Windows is
- * definitely suspended (this volume is not the system volume).  Caveat:  on a
- * system with many volumes it is possible that the < 4kiB check is bogus but
- * for now this should do fine.
- *
- * If hiberfil.sys exists and is larger than 4kiB in size, we need to read the
- * hiberfil header (which is the first 4kiB).  If this begins with "hibr",
- * Windows is definitely suspended.  If it is completely full of zeroes,
- * Windows is definitely not hibernated.  Any other case is treated as if
- * Windows is suspended.  This caters for the above mentioned caveat of a
- * system with many volumes where no "hibr" magic would be present and there is
- * no zero header.
- *
- * Return 0 if Windows is not hibernated on the volume, >0 if Windows is
- * hibernated on the volume, and -errno on error.
- */
-static int check_windows_hibernation_status(ntfs_volume *vol)
-{
-	MFT_REF mref;
-	struct inode *vi;
-	struct page *page;
-	u32 *kaddr, *kend;
-	ntfs_name *name = NULL;
-	int ret = 1;
-	static const ntfschar hiberfil[13] = { cpu_to_le16('h'),
-			cpu_to_le16('i'), cpu_to_le16('b'),
-			cpu_to_le16('e'), cpu_to_le16('r'),
-			cpu_to_le16('f'), cpu_to_le16('i'),
-			cpu_to_le16('l'), cpu_to_le16('.'),
-			cpu_to_le16('s'), cpu_to_le16('y'),
-			cpu_to_le16('s'), 0 };
-
-	ntfs_debug("Entering.");
-	/*
-	 * Find the inode number for the hibernation file by looking up the
-	 * filename hiberfil.sys in the root directory.
-	 */
-	inode_lock(vol->root_ino);
-	mref = ntfs_lookup_inode_by_name(NTFS_I(vol->root_ino), hiberfil, 12,
-			&name);
-	inode_unlock(vol->root_ino);
-	if (IS_ERR_MREF(mref)) {
-		ret = MREF_ERR(mref);
-		/* If the file does not exist, Windows is not hibernated. */
-		if (ret == -ENOENT) {
-			ntfs_debug("hiberfil.sys not present.  Windows is not "
-					"hibernated on the volume.");
-			return 0;
-		}
-		/* A real error occurred. */
-		ntfs_error(vol->sb, "Failed to find inode number for "
-				"hiberfil.sys.");
-		return ret;
-	}
-	/* We do not care for the type of match that was found. */
-	kfree(name);
-	/* Get the inode. */
-	vi = ntfs_iget(vol->sb, MREF(mref));
-	if (IS_ERR(vi) || is_bad_inode(vi)) {
-		if (!IS_ERR(vi))
-			iput(vi);
-		ntfs_error(vol->sb, "Failed to load hiberfil.sys.");
-		return IS_ERR(vi) ? PTR_ERR(vi) : -EIO;
-	}
-	if (unlikely(i_size_read(vi) < NTFS_HIBERFIL_HEADER_SIZE)) {
-		ntfs_debug("hiberfil.sys is smaller than 4kiB (0x%llx).  "
-				"Windows is hibernated on the volume.  This "
-				"is not the system volume.", i_size_read(vi));
-		goto iput_out;
-	}
-	page = ntfs_map_page(vi->i_mapping, 0);
-	if (IS_ERR(page)) {
-		ntfs_error(vol->sb, "Failed to read from hiberfil.sys.");
-		ret = PTR_ERR(page);
-		goto iput_out;
-	}
-	kaddr = (u32*)page_address(page);
-	if (*(le32*)kaddr == cpu_to_le32(0x72626968)/*'hibr'*/) {
-		ntfs_debug("Magic \"hibr\" found in hiberfil.sys.  Windows is "
-				"hibernated on the volume.  This is the "
-				"system volume.");
-		goto unm_iput_out;
-	}
-	kend = kaddr + NTFS_HIBERFIL_HEADER_SIZE/sizeof(*kaddr);
-	do {
-		if (unlikely(*kaddr)) {
-			ntfs_debug("hiberfil.sys is larger than 4kiB "
-					"(0x%llx), does not contain the "
-					"\"hibr\" magic, and does not have a "
-					"zero header.  Windows is hibernated "
-					"on the volume.  This is not the "
-					"system volume.", i_size_read(vi));
-			goto unm_iput_out;
-		}
-	} while (++kaddr < kend);
-	ntfs_debug("hiberfil.sys contains a zero header.  Windows is not "
-			"hibernated on the volume.  This is the system "
-			"volume.");
-	ret = 0;
-unm_iput_out:
-	ntfs_unmap_page(page);
-iput_out:
-	iput(vi);
-	return ret;
-}
-
-/**
- * load_and_init_quota - load and setup the quota file for a volume if present
- * @vol:	ntfs super block describing device whose quota file to load
- *
- * Return 'true' on success or 'false' on error.  If $Quota is not present, we
- * leave vol->quota_ino as NULL and return success.
- */
-static bool load_and_init_quota(ntfs_volume *vol)
-{
-	MFT_REF mref;
-	struct inode *tmp_ino;
-	ntfs_name *name = NULL;
-	static const ntfschar Quota[7] = { cpu_to_le16('$'),
-			cpu_to_le16('Q'), cpu_to_le16('u'),
-			cpu_to_le16('o'), cpu_to_le16('t'),
-			cpu_to_le16('a'), 0 };
-	static ntfschar Q[3] = { cpu_to_le16('$'),
-			cpu_to_le16('Q'), 0 };
-
-	ntfs_debug("Entering.");
-	/*
-	 * Find the inode number for the quota file by looking up the filename
-	 * $Quota in the extended system files directory $Extend.
-	 */
-	inode_lock(vol->extend_ino);
-	mref = ntfs_lookup_inode_by_name(NTFS_I(vol->extend_ino), Quota, 6,
-			&name);
-	inode_unlock(vol->extend_ino);
-	if (IS_ERR_MREF(mref)) {
-		/*
-		 * If the file does not exist, quotas are disabled and have
-		 * never been enabled on this volume, just return success.
-		 */
-		if (MREF_ERR(mref) == -ENOENT) {
-			ntfs_debug("$Quota not present.  Volume does not have "
-					"quotas enabled.");
-			/*
-			 * No need to try to set quotas out of date if they are
-			 * not enabled.
-			 */
-			NVolSetQuotaOutOfDate(vol);
-			return true;
-		}
-		/* A real error occurred. */
-		ntfs_error(vol->sb, "Failed to find inode number for $Quota.");
-		return false;
-	}
-	/* We do not care for the type of match that was found. */
-	kfree(name);
-	/* Get the inode. */
-	tmp_ino = ntfs_iget(vol->sb, MREF(mref));
-	if (IS_ERR(tmp_ino) || is_bad_inode(tmp_ino)) {
-		if (!IS_ERR(tmp_ino))
-			iput(tmp_ino);
-		ntfs_error(vol->sb, "Failed to load $Quota.");
-		return false;
-	}
-	vol->quota_ino = tmp_ino;
-	/* Get the $Q index allocation attribute. */
-	tmp_ino = ntfs_index_iget(vol->quota_ino, Q, 2);
-	if (IS_ERR(tmp_ino)) {
-		ntfs_error(vol->sb, "Failed to load $Quota/$Q index.");
-		return false;
-	}
-	vol->quota_q_ino = tmp_ino;
-	ntfs_debug("Done.");
-	return true;
-}
-
-/**
- * load_and_init_usnjrnl - load and setup the transaction log if present
- * @vol:	ntfs super block describing device whose usnjrnl file to load
- *
- * Return 'true' on success or 'false' on error.
- *
- * If $UsnJrnl is not present or in the process of being disabled, we set
- * NVolUsnJrnlStamped() and return success.
- *
- * If the $UsnJrnl $DATA/$J attribute has a size equal to the lowest valid usn,
- * i.e. transaction logging has only just been enabled or the journal has been
- * stamped and nothing has been logged since, we also set NVolUsnJrnlStamped()
- * and return success.
- */
-static bool load_and_init_usnjrnl(ntfs_volume *vol)
-{
-	MFT_REF mref;
-	struct inode *tmp_ino;
-	ntfs_inode *tmp_ni;
-	struct page *page;
-	ntfs_name *name = NULL;
-	USN_HEADER *uh;
-	static const ntfschar UsnJrnl[9] = { cpu_to_le16('$'),
-			cpu_to_le16('U'), cpu_to_le16('s'),
-			cpu_to_le16('n'), cpu_to_le16('J'),
-			cpu_to_le16('r'), cpu_to_le16('n'),
-			cpu_to_le16('l'), 0 };
-	static ntfschar Max[5] = { cpu_to_le16('$'),
-			cpu_to_le16('M'), cpu_to_le16('a'),
-			cpu_to_le16('x'), 0 };
-	static ntfschar J[3] = { cpu_to_le16('$'),
-			cpu_to_le16('J'), 0 };
-
-	ntfs_debug("Entering.");
-	/*
-	 * Find the inode number for the transaction log file by looking up the
-	 * filename $UsnJrnl in the extended system files directory $Extend.
-	 */
-	inode_lock(vol->extend_ino);
-	mref = ntfs_lookup_inode_by_name(NTFS_I(vol->extend_ino), UsnJrnl, 8,
-			&name);
-	inode_unlock(vol->extend_ino);
-	if (IS_ERR_MREF(mref)) {
-		/*
-		 * If the file does not exist, transaction logging is disabled,
-		 * just return success.
-		 */
-		if (MREF_ERR(mref) == -ENOENT) {
-			ntfs_debug("$UsnJrnl not present.  Volume does not "
-					"have transaction logging enabled.");
-not_enabled:
-			/*
-			 * No need to try to stamp the transaction log if
-			 * transaction logging is not enabled.
-			 */
-			NVolSetUsnJrnlStamped(vol);
-			return true;
-		}
-		/* A real error occurred. */
-		ntfs_error(vol->sb, "Failed to find inode number for "
-				"$UsnJrnl.");
-		return false;
-	}
-	/* We do not care for the type of match that was found. */
-	kfree(name);
-	/* Get the inode. */
-	tmp_ino = ntfs_iget(vol->sb, MREF(mref));
-	if (IS_ERR(tmp_ino) || unlikely(is_bad_inode(tmp_ino))) {
-		if (!IS_ERR(tmp_ino))
-			iput(tmp_ino);
-		ntfs_error(vol->sb, "Failed to load $UsnJrnl.");
-		return false;
-	}
-	vol->usnjrnl_ino = tmp_ino;
-	/*
-	 * If the transaction log is in the process of being deleted, we can
-	 * ignore it.
-	 */
-	if (unlikely(vol->vol_flags & VOLUME_DELETE_USN_UNDERWAY)) {
-		ntfs_debug("$UsnJrnl in the process of being disabled.  "
-				"Volume does not have transaction logging "
-				"enabled.");
-		goto not_enabled;
-	}
-	/* Get the $DATA/$Max attribute. */
-	tmp_ino = ntfs_attr_iget(vol->usnjrnl_ino, AT_DATA, Max, 4);
-	if (IS_ERR(tmp_ino)) {
-		ntfs_error(vol->sb, "Failed to load $UsnJrnl/$DATA/$Max "
-				"attribute.");
-		return false;
-	}
-	vol->usnjrnl_max_ino = tmp_ino;
-	if (unlikely(i_size_read(tmp_ino) < sizeof(USN_HEADER))) {
-		ntfs_error(vol->sb, "Found corrupt $UsnJrnl/$DATA/$Max "
-				"attribute (size is 0x%llx but should be at "
-				"least 0x%zx bytes).", i_size_read(tmp_ino),
-				sizeof(USN_HEADER));
-		return false;
-	}
-	/* Get the $DATA/$J attribute. */
-	tmp_ino = ntfs_attr_iget(vol->usnjrnl_ino, AT_DATA, J, 2);
-	if (IS_ERR(tmp_ino)) {
-		ntfs_error(vol->sb, "Failed to load $UsnJrnl/$DATA/$J "
-				"attribute.");
-		return false;
-	}
-	vol->usnjrnl_j_ino = tmp_ino;
-	/* Verify $J is non-resident and sparse. */
-	tmp_ni = NTFS_I(vol->usnjrnl_j_ino);
-	if (unlikely(!NInoNonResident(tmp_ni) || !NInoSparse(tmp_ni))) {
-		ntfs_error(vol->sb, "$UsnJrnl/$DATA/$J attribute is resident "
-				"and/or not sparse.");
-		return false;
-	}
-	/* Read the USN_HEADER from $DATA/$Max. */
-	page = ntfs_map_page(vol->usnjrnl_max_ino->i_mapping, 0);
-	if (IS_ERR(page)) {
-		ntfs_error(vol->sb, "Failed to read from $UsnJrnl/$DATA/$Max "
-				"attribute.");
-		return false;
-	}
-	uh = (USN_HEADER*)page_address(page);
-	/* Sanity check the $Max. */
-	if (unlikely(sle64_to_cpu(uh->allocation_delta) >
-			sle64_to_cpu(uh->maximum_size))) {
-		ntfs_error(vol->sb, "Allocation delta (0x%llx) exceeds "
-				"maximum size (0x%llx).  $UsnJrnl is corrupt.",
-				(long long)sle64_to_cpu(uh->allocation_delta),
-				(long long)sle64_to_cpu(uh->maximum_size));
-		ntfs_unmap_page(page);
-		return false;
-	}
-	/*
-	 * If the transaction log has been stamped and nothing has been written
-	 * to it since, we do not need to stamp it.
-	 */
-	if (unlikely(sle64_to_cpu(uh->lowest_valid_usn) >=
-			i_size_read(vol->usnjrnl_j_ino))) {
-		if (likely(sle64_to_cpu(uh->lowest_valid_usn) ==
-				i_size_read(vol->usnjrnl_j_ino))) {
-			ntfs_unmap_page(page);
-			ntfs_debug("$UsnJrnl is enabled but nothing has been "
-					"logged since it was last stamped.  "
-					"Treating this as if the volume does "
-					"not have transaction logging "
-					"enabled.");
-			goto not_enabled;
-		}
-		ntfs_error(vol->sb, "$UsnJrnl has lowest valid usn (0x%llx) "
-				"which is out of bounds (0x%llx).  $UsnJrnl "
-				"is corrupt.",
-				(long long)sle64_to_cpu(uh->lowest_valid_usn),
-				i_size_read(vol->usnjrnl_j_ino));
-		ntfs_unmap_page(page);
-		return false;
-	}
-	ntfs_unmap_page(page);
-	ntfs_debug("Done.");
-	return true;
-}
-
-/**
- * load_and_init_attrdef - load the attribute definitions table for a volume
- * @vol:	ntfs super block describing device whose attrdef to load
- *
- * Return 'true' on success or 'false' on error.
- */
-static bool load_and_init_attrdef(ntfs_volume *vol)
-{
-	loff_t i_size;
-	struct super_block *sb = vol->sb;
-	struct inode *ino;
-	struct page *page;
-	pgoff_t index, max_index;
-	unsigned int size;
-
-	ntfs_debug("Entering.");
-	/* Read attrdef table and setup vol->attrdef and vol->attrdef_size. */
-	ino = ntfs_iget(sb, FILE_AttrDef);
-	if (IS_ERR(ino) || is_bad_inode(ino)) {
-		if (!IS_ERR(ino))
-			iput(ino);
-		goto failed;
-	}
-	NInoSetSparseDisabled(NTFS_I(ino));
-	/* The size of FILE_AttrDef must be above 0 and fit inside 31 bits. */
-	i_size = i_size_read(ino);
-	if (i_size <= 0 || i_size > 0x7fffffff)
-		goto iput_failed;
-	vol->attrdef = (ATTR_DEF*)ntfs_malloc_nofs(i_size);
-	if (!vol->attrdef)
-		goto iput_failed;
-	index = 0;
-	max_index = i_size >> PAGE_SHIFT;
-	size = PAGE_SIZE;
-	while (index < max_index) {
-		/* Read the attrdef table and copy it into the linear buffer. */
-read_partial_attrdef_page:
-		page = ntfs_map_page(ino->i_mapping, index);
-		if (IS_ERR(page))
-			goto free_iput_failed;
-		memcpy((u8*)vol->attrdef + (index++ << PAGE_SHIFT),
-				page_address(page), size);
-		ntfs_unmap_page(page);
-	}
-	if (size == PAGE_SIZE) {
-		size = i_size & ~PAGE_MASK;
-		if (size)
-			goto read_partial_attrdef_page;
-	}
-	vol->attrdef_size = i_size;
-	ntfs_debug("Read %llu bytes from $AttrDef.", i_size);
-	iput(ino);
-	return true;
-free_iput_failed:
-	ntfs_free(vol->attrdef);
-	vol->attrdef = NULL;
-iput_failed:
-	iput(ino);
-failed:
-	ntfs_error(sb, "Failed to initialize attribute definition table.");
-	return false;
-}
-
-#endif /* NTFS_RW */
-
-/**
- * load_and_init_upcase - load the upcase table for an ntfs volume
- * @vol:	ntfs super block describing device whose upcase to load
- *
- * Return 'true' on success or 'false' on error.
- */
-static bool load_and_init_upcase(ntfs_volume *vol)
-{
-	loff_t i_size;
-	struct super_block *sb = vol->sb;
-	struct inode *ino;
-	struct page *page;
-	pgoff_t index, max_index;
-	unsigned int size;
-	int i, max;
-
-	ntfs_debug("Entering.");
-	/* Read upcase table and setup vol->upcase and vol->upcase_len. */
-	ino = ntfs_iget(sb, FILE_UpCase);
-	if (IS_ERR(ino) || is_bad_inode(ino)) {
-		if (!IS_ERR(ino))
-			iput(ino);
-		goto upcase_failed;
-	}
-	/*
-	 * The upcase size must not be above 64k Unicode characters, must not
-	 * be zero and must be a multiple of sizeof(ntfschar).
-	 */
-	i_size = i_size_read(ino);
-	if (!i_size || i_size & (sizeof(ntfschar) - 1) ||
-			i_size > 64ULL * 1024 * sizeof(ntfschar))
-		goto iput_upcase_failed;
-	vol->upcase = (ntfschar*)ntfs_malloc_nofs(i_size);
-	if (!vol->upcase)
-		goto iput_upcase_failed;
-	index = 0;
-	max_index = i_size >> PAGE_SHIFT;
-	size = PAGE_SIZE;
-	while (index < max_index) {
-		/* Read the upcase table and copy it into the linear buffer. */
-read_partial_upcase_page:
-		page = ntfs_map_page(ino->i_mapping, index);
-		if (IS_ERR(page))
-			goto iput_upcase_failed;
-		memcpy((char*)vol->upcase + (index++ << PAGE_SHIFT),
-				page_address(page), size);
-		ntfs_unmap_page(page);
-	}
-	if (size == PAGE_SIZE) {
-		size = i_size & ~PAGE_MASK;
-		if (size)
-			goto read_partial_upcase_page;
-	}
-	vol->upcase_len = i_size >> UCHAR_T_SIZE_BITS;
-	ntfs_debug("Read %llu bytes from $UpCase (expected %zu bytes).",
-			i_size, 64 * 1024 * sizeof(ntfschar));
-	iput(ino);
-	mutex_lock(&ntfs_lock);
-	if (!default_upcase) {
-		ntfs_debug("Using volume specified $UpCase since default is "
-				"not present.");
-		mutex_unlock(&ntfs_lock);
-		return true;
-	}
-	max = default_upcase_len;
-	if (max > vol->upcase_len)
-		max = vol->upcase_len;
-	for (i = 0; i < max; i++)
-		if (vol->upcase[i] != default_upcase[i])
-			break;
-	if (i == max) {
-		ntfs_free(vol->upcase);
-		vol->upcase = default_upcase;
-		vol->upcase_len = max;
-		ntfs_nr_upcase_users++;
-		mutex_unlock(&ntfs_lock);
-		ntfs_debug("Volume specified $UpCase matches default. Using "
-				"default.");
-		return true;
-	}
-	mutex_unlock(&ntfs_lock);
-	ntfs_debug("Using volume specified $UpCase since it does not match "
-			"the default.");
-	return true;
-iput_upcase_failed:
-	iput(ino);
-	ntfs_free(vol->upcase);
-	vol->upcase = NULL;
-upcase_failed:
-	mutex_lock(&ntfs_lock);
-	if (default_upcase) {
-		vol->upcase = default_upcase;
-		vol->upcase_len = default_upcase_len;
-		ntfs_nr_upcase_users++;
-		mutex_unlock(&ntfs_lock);
-		ntfs_error(sb, "Failed to load $UpCase from the volume. Using "
-				"default.");
-		return true;
-	}
-	mutex_unlock(&ntfs_lock);
-	ntfs_error(sb, "Failed to initialize upcase table.");
-	return false;
-}
-
-/*
- * The lcn and mft bitmap inodes are NTFS-internal inodes with
- * their own special locking rules:
- */
-static struct lock_class_key
-	lcnbmp_runlist_lock_key, lcnbmp_mrec_lock_key,
-	mftbmp_runlist_lock_key, mftbmp_mrec_lock_key;
-
-/**
- * load_system_files - open the system files using normal functions
- * @vol:	ntfs super block describing device whose system files to load
- *
- * Open the system files with normal access functions and complete setting up
- * the ntfs super block @vol.
- *
- * Return 'true' on success or 'false' on error.
- */
-static bool load_system_files(ntfs_volume *vol)
-{
-	struct super_block *sb = vol->sb;
-	MFT_RECORD *m;
-	VOLUME_INFORMATION *vi;
-	ntfs_attr_search_ctx *ctx;
-#ifdef NTFS_RW
-	RESTART_PAGE_HEADER *rp;
-	int err;
-#endif /* NTFS_RW */
-
-	ntfs_debug("Entering.");
-#ifdef NTFS_RW
-	/* Get mft mirror inode compare the contents of $MFT and $MFTMirr. */
-	if (!load_and_init_mft_mirror(vol) || !check_mft_mirror(vol)) {
-		static const char *es1 = "Failed to load $MFTMirr";
-		static const char *es2 = "$MFTMirr does not match $MFT";
-		static const char *es3 = ".  Run ntfsfix and/or chkdsk.";
-
-		/* If a read-write mount, convert it to a read-only mount. */
-		if (!sb_rdonly(sb)) {
-			if (!(vol->on_errors & (ON_ERRORS_REMOUNT_RO |
-					ON_ERRORS_CONTINUE))) {
-				ntfs_error(sb, "%s and neither on_errors="
-						"continue nor on_errors="
-						"remount-ro was specified%s",
-						!vol->mftmirr_ino ? es1 : es2,
-						es3);
-				goto iput_mirr_err_out;
-			}
-			sb->s_flags |= SB_RDONLY;
-			ntfs_error(sb, "%s.  Mounting read-only%s",
-					!vol->mftmirr_ino ? es1 : es2, es3);
-		} else
-			ntfs_warning(sb, "%s.  Will not be able to remount "
-					"read-write%s",
-					!vol->mftmirr_ino ? es1 : es2, es3);
-		/* This will prevent a read-write remount. */
-		NVolSetErrors(vol);
-	}
-#endif /* NTFS_RW */
-	/* Get mft bitmap attribute inode. */
-	vol->mftbmp_ino = ntfs_attr_iget(vol->mft_ino, AT_BITMAP, NULL, 0);
-	if (IS_ERR(vol->mftbmp_ino)) {
-		ntfs_error(sb, "Failed to load $MFT/$BITMAP attribute.");
-		goto iput_mirr_err_out;
-	}
-	lockdep_set_class(&NTFS_I(vol->mftbmp_ino)->runlist.lock,
-			   &mftbmp_runlist_lock_key);
-	lockdep_set_class(&NTFS_I(vol->mftbmp_ino)->mrec_lock,
-			   &mftbmp_mrec_lock_key);
-	/* Read upcase table and setup @vol->upcase and @vol->upcase_len. */
-	if (!load_and_init_upcase(vol))
-		goto iput_mftbmp_err_out;
-#ifdef NTFS_RW
-	/*
-	 * Read attribute definitions table and setup @vol->attrdef and
-	 * @vol->attrdef_size.
-	 */
-	if (!load_and_init_attrdef(vol))
-		goto iput_upcase_err_out;
-#endif /* NTFS_RW */
-	/*
-	 * Get the cluster allocation bitmap inode and verify the size, no
-	 * need for any locking at this stage as we are already running
-	 * exclusively as we are mount in progress task.
-	 */
-	vol->lcnbmp_ino = ntfs_iget(sb, FILE_Bitmap);
-	if (IS_ERR(vol->lcnbmp_ino) || is_bad_inode(vol->lcnbmp_ino)) {
-		if (!IS_ERR(vol->lcnbmp_ino))
-			iput(vol->lcnbmp_ino);
-		goto bitmap_failed;
-	}
-	lockdep_set_class(&NTFS_I(vol->lcnbmp_ino)->runlist.lock,
-			   &lcnbmp_runlist_lock_key);
-	lockdep_set_class(&NTFS_I(vol->lcnbmp_ino)->mrec_lock,
-			   &lcnbmp_mrec_lock_key);
-
-	NInoSetSparseDisabled(NTFS_I(vol->lcnbmp_ino));
-	if ((vol->nr_clusters + 7) >> 3 > i_size_read(vol->lcnbmp_ino)) {
-		iput(vol->lcnbmp_ino);
-bitmap_failed:
-		ntfs_error(sb, "Failed to load $Bitmap.");
-		goto iput_attrdef_err_out;
-	}
-	/*
-	 * Get the volume inode and setup our cache of the volume flags and
-	 * version.
-	 */
-	vol->vol_ino = ntfs_iget(sb, FILE_Volume);
-	if (IS_ERR(vol->vol_ino) || is_bad_inode(vol->vol_ino)) {
-		if (!IS_ERR(vol->vol_ino))
-			iput(vol->vol_ino);
-volume_failed:
-		ntfs_error(sb, "Failed to load $Volume.");
-		goto iput_lcnbmp_err_out;
-	}
-	m = map_mft_record(NTFS_I(vol->vol_ino));
-	if (IS_ERR(m)) {
-iput_volume_failed:
-		iput(vol->vol_ino);
-		goto volume_failed;
-	}
-	if (!(ctx = ntfs_attr_get_search_ctx(NTFS_I(vol->vol_ino), m))) {
-		ntfs_error(sb, "Failed to get attribute search context.");
-		goto get_ctx_vol_failed;
-	}
-	if (ntfs_attr_lookup(AT_VOLUME_INFORMATION, NULL, 0, 0, 0, NULL, 0,
-			ctx) || ctx->attr->non_resident || ctx->attr->flags) {
-err_put_vol:
-		ntfs_attr_put_search_ctx(ctx);
-get_ctx_vol_failed:
-		unmap_mft_record(NTFS_I(vol->vol_ino));
-		goto iput_volume_failed;
-	}
-	vi = (VOLUME_INFORMATION*)((char*)ctx->attr +
-			le16_to_cpu(ctx->attr->data.resident.value_offset));
-	/* Some bounds checks. */
-	if ((u8*)vi < (u8*)ctx->attr || (u8*)vi +
-			le32_to_cpu(ctx->attr->data.resident.value_length) >
-			(u8*)ctx->attr + le32_to_cpu(ctx->attr->length))
-		goto err_put_vol;
-	/* Copy the volume flags and version to the ntfs_volume structure. */
-	vol->vol_flags = vi->flags;
-	vol->major_ver = vi->major_ver;
-	vol->minor_ver = vi->minor_ver;
-	ntfs_attr_put_search_ctx(ctx);
-	unmap_mft_record(NTFS_I(vol->vol_ino));
-	pr_info("volume version %i.%i.\n", vol->major_ver,
-			vol->minor_ver);
-	if (vol->major_ver < 3 && NVolSparseEnabled(vol)) {
-		ntfs_warning(vol->sb, "Disabling sparse support due to NTFS "
-				"volume version %i.%i (need at least version "
-				"3.0).", vol->major_ver, vol->minor_ver);
-		NVolClearSparseEnabled(vol);
-	}
-#ifdef NTFS_RW
-	/* Make sure that no unsupported volume flags are set. */
-	if (vol->vol_flags & VOLUME_MUST_MOUNT_RO_MASK) {
-		static const char *es1a = "Volume is dirty";
-		static const char *es1b = "Volume has been modified by chkdsk";
-		static const char *es1c = "Volume has unsupported flags set";
-		static const char *es2a = ".  Run chkdsk and mount in Windows.";
-		static const char *es2b = ".  Mount in Windows.";
-		const char *es1, *es2;
-
-		es2 = es2a;
-		if (vol->vol_flags & VOLUME_IS_DIRTY)
-			es1 = es1a;
-		else if (vol->vol_flags & VOLUME_MODIFIED_BY_CHKDSK) {
-			es1 = es1b;
-			es2 = es2b;
-		} else {
-			es1 = es1c;
-			ntfs_warning(sb, "Unsupported volume flags 0x%x "
-					"encountered.",
-					(unsigned)le16_to_cpu(vol->vol_flags));
-		}
-		/* If a read-write mount, convert it to a read-only mount. */
-		if (!sb_rdonly(sb)) {
-			if (!(vol->on_errors & (ON_ERRORS_REMOUNT_RO |
-					ON_ERRORS_CONTINUE))) {
-				ntfs_error(sb, "%s and neither on_errors="
-						"continue nor on_errors="
-						"remount-ro was specified%s",
-						es1, es2);
-				goto iput_vol_err_out;
-			}
-			sb->s_flags |= SB_RDONLY;
-			ntfs_error(sb, "%s.  Mounting read-only%s", es1, es2);
-		} else
-			ntfs_warning(sb, "%s.  Will not be able to remount "
-					"read-write%s", es1, es2);
-		/*
-		 * Do not set NVolErrors() because ntfs_remount() re-checks the
-		 * flags which we need to do in case any flags have changed.
-		 */
-	}
-	/*
-	 * Get the inode for the logfile, check it and determine if the volume
-	 * was shutdown cleanly.
-	 */
-	rp = NULL;
-	if (!load_and_check_logfile(vol, &rp) ||
-			!ntfs_is_logfile_clean(vol->logfile_ino, rp)) {
-		static const char *es1a = "Failed to load $LogFile";
-		static const char *es1b = "$LogFile is not clean";
-		static const char *es2 = ".  Mount in Windows.";
-		const char *es1;
-
-		es1 = !vol->logfile_ino ? es1a : es1b;
-		/* If a read-write mount, convert it to a read-only mount. */
-		if (!sb_rdonly(sb)) {
-			if (!(vol->on_errors & (ON_ERRORS_REMOUNT_RO |
-					ON_ERRORS_CONTINUE))) {
-				ntfs_error(sb, "%s and neither on_errors="
-						"continue nor on_errors="
-						"remount-ro was specified%s",
-						es1, es2);
-				if (vol->logfile_ino) {
-					BUG_ON(!rp);
-					ntfs_free(rp);
-				}
-				goto iput_logfile_err_out;
-			}
-			sb->s_flags |= SB_RDONLY;
-			ntfs_error(sb, "%s.  Mounting read-only%s", es1, es2);
-		} else
-			ntfs_warning(sb, "%s.  Will not be able to remount "
-					"read-write%s", es1, es2);
-		/* This will prevent a read-write remount. */
-		NVolSetErrors(vol);
-	}
-	ntfs_free(rp);
-#endif /* NTFS_RW */
-	/* Get the root directory inode so we can do path lookups. */
-	vol->root_ino = ntfs_iget(sb, FILE_root);
-	if (IS_ERR(vol->root_ino) || is_bad_inode(vol->root_ino)) {
-		if (!IS_ERR(vol->root_ino))
-			iput(vol->root_ino);
-		ntfs_error(sb, "Failed to load root directory.");
-		goto iput_logfile_err_out;
-	}
-#ifdef NTFS_RW
-	/*
-	 * Check if Windows is suspended to disk on the target volume.  If it
-	 * is hibernated, we must not write *anything* to the disk so set
-	 * NVolErrors() without setting the dirty volume flag and mount
-	 * read-only.  This will prevent read-write remounting and it will also
-	 * prevent all writes.
-	 */
-	err = check_windows_hibernation_status(vol);
-	if (unlikely(err)) {
-		static const char *es1a = "Failed to determine if Windows is "
-				"hibernated";
-		static const char *es1b = "Windows is hibernated";
-		static const char *es2 = ".  Run chkdsk.";
-		const char *es1;
-
-		es1 = err < 0 ? es1a : es1b;
-		/* If a read-write mount, convert it to a read-only mount. */
-		if (!sb_rdonly(sb)) {
-			if (!(vol->on_errors & (ON_ERRORS_REMOUNT_RO |
-					ON_ERRORS_CONTINUE))) {
-				ntfs_error(sb, "%s and neither on_errors="
-						"continue nor on_errors="
-						"remount-ro was specified%s",
-						es1, es2);
-				goto iput_root_err_out;
-			}
-			sb->s_flags |= SB_RDONLY;
-			ntfs_error(sb, "%s.  Mounting read-only%s", es1, es2);
-		} else
-			ntfs_warning(sb, "%s.  Will not be able to remount "
-					"read-write%s", es1, es2);
-		/* This will prevent a read-write remount. */
-		NVolSetErrors(vol);
-	}
-	/* If (still) a read-write mount, mark the volume dirty. */
-	if (!sb_rdonly(sb) && ntfs_set_volume_flags(vol, VOLUME_IS_DIRTY)) {
-		static const char *es1 = "Failed to set dirty bit in volume "
-				"information flags";
-		static const char *es2 = ".  Run chkdsk.";
-
-		/* Convert to a read-only mount. */
-		if (!(vol->on_errors & (ON_ERRORS_REMOUNT_RO |
-				ON_ERRORS_CONTINUE))) {
-			ntfs_error(sb, "%s and neither on_errors=continue nor "
-					"on_errors=remount-ro was specified%s",
-					es1, es2);
-			goto iput_root_err_out;
-		}
-		ntfs_error(sb, "%s.  Mounting read-only%s", es1, es2);
-		sb->s_flags |= SB_RDONLY;
-		/*
-		 * Do not set NVolErrors() because ntfs_remount() might manage
-		 * to set the dirty flag in which case all would be well.
-		 */
-	}
-#if 0
-	// TODO: Enable this code once we start modifying anything that is
-	//	 different between NTFS 1.2 and 3.x...
-	/*
-	 * If (still) a read-write mount, set the NT4 compatibility flag on
-	 * newer NTFS version volumes.
-	 */
-	if (!(sb->s_flags & SB_RDONLY) && (vol->major_ver > 1) &&
-			ntfs_set_volume_flags(vol, VOLUME_MOUNTED_ON_NT4)) {
-		static const char *es1 = "Failed to set NT4 compatibility flag";
-		static const char *es2 = ".  Run chkdsk.";
-
-		/* Convert to a read-only mount. */
-		if (!(vol->on_errors & (ON_ERRORS_REMOUNT_RO |
-				ON_ERRORS_CONTINUE))) {
-			ntfs_error(sb, "%s and neither on_errors=continue nor "
-					"on_errors=remount-ro was specified%s",
-					es1, es2);
-			goto iput_root_err_out;
-		}
-		ntfs_error(sb, "%s.  Mounting read-only%s", es1, es2);
-		sb->s_flags |= SB_RDONLY;
-		NVolSetErrors(vol);
-	}
-#endif
-	/* If (still) a read-write mount, empty the logfile. */
-	if (!sb_rdonly(sb) && !ntfs_empty_logfile(vol->logfile_ino)) {
-		static const char *es1 = "Failed to empty $LogFile";
-		static const char *es2 = ".  Mount in Windows.";
-
-		/* Convert to a read-only mount. */
-		if (!(vol->on_errors & (ON_ERRORS_REMOUNT_RO |
-				ON_ERRORS_CONTINUE))) {
-			ntfs_error(sb, "%s and neither on_errors=continue nor "
-					"on_errors=remount-ro was specified%s",
-					es1, es2);
-			goto iput_root_err_out;
-		}
-		ntfs_error(sb, "%s.  Mounting read-only%s", es1, es2);
-		sb->s_flags |= SB_RDONLY;
-		NVolSetErrors(vol);
-	}
-#endif /* NTFS_RW */
-	/* If on NTFS versions before 3.0, we are done. */
-	if (unlikely(vol->major_ver < 3))
-		return true;
-	/* NTFS 3.0+ specific initialization. */
-	/* Get the security descriptors inode. */
-	vol->secure_ino = ntfs_iget(sb, FILE_Secure);
-	if (IS_ERR(vol->secure_ino) || is_bad_inode(vol->secure_ino)) {
-		if (!IS_ERR(vol->secure_ino))
-			iput(vol->secure_ino);
-		ntfs_error(sb, "Failed to load $Secure.");
-		goto iput_root_err_out;
-	}
-	// TODO: Initialize security.
-	/* Get the extended system files' directory inode. */
-	vol->extend_ino = ntfs_iget(sb, FILE_Extend);
-	if (IS_ERR(vol->extend_ino) || is_bad_inode(vol->extend_ino) ||
-	    !S_ISDIR(vol->extend_ino->i_mode)) {
-		if (!IS_ERR(vol->extend_ino))
-			iput(vol->extend_ino);
-		ntfs_error(sb, "Failed to load $Extend.");
-		goto iput_sec_err_out;
-	}
-#ifdef NTFS_RW
-	/* Find the quota file, load it if present, and set it up. */
-	if (!load_and_init_quota(vol)) {
-		static const char *es1 = "Failed to load $Quota";
-		static const char *es2 = ".  Run chkdsk.";
-
-		/* If a read-write mount, convert it to a read-only mount. */
-		if (!sb_rdonly(sb)) {
-			if (!(vol->on_errors & (ON_ERRORS_REMOUNT_RO |
-					ON_ERRORS_CONTINUE))) {
-				ntfs_error(sb, "%s and neither on_errors="
-						"continue nor on_errors="
-						"remount-ro was specified%s",
-						es1, es2);
-				goto iput_quota_err_out;
-			}
-			sb->s_flags |= SB_RDONLY;
-			ntfs_error(sb, "%s.  Mounting read-only%s", es1, es2);
-		} else
-			ntfs_warning(sb, "%s.  Will not be able to remount "
-					"read-write%s", es1, es2);
-		/* This will prevent a read-write remount. */
-		NVolSetErrors(vol);
-	}
-	/* If (still) a read-write mount, mark the quotas out of date. */
-	if (!sb_rdonly(sb) && !ntfs_mark_quotas_out_of_date(vol)) {
-		static const char *es1 = "Failed to mark quotas out of date";
-		static const char *es2 = ".  Run chkdsk.";
-
-		/* Convert to a read-only mount. */
-		if (!(vol->on_errors & (ON_ERRORS_REMOUNT_RO |
-				ON_ERRORS_CONTINUE))) {
-			ntfs_error(sb, "%s and neither on_errors=continue nor "
-					"on_errors=remount-ro was specified%s",
-					es1, es2);
-			goto iput_quota_err_out;
-		}
-		ntfs_error(sb, "%s.  Mounting read-only%s", es1, es2);
-		sb->s_flags |= SB_RDONLY;
-		NVolSetErrors(vol);
-	}
-	/*
-	 * Find the transaction log file ($UsnJrnl), load it if present, check
-	 * it, and set it up.
-	 */
-	if (!load_and_init_usnjrnl(vol)) {
-		static const char *es1 = "Failed to load $UsnJrnl";
-		static const char *es2 = ".  Run chkdsk.";
-
-		/* If a read-write mount, convert it to a read-only mount. */
-		if (!sb_rdonly(sb)) {
-			if (!(vol->on_errors & (ON_ERRORS_REMOUNT_RO |
-					ON_ERRORS_CONTINUE))) {
-				ntfs_error(sb, "%s and neither on_errors="
-						"continue nor on_errors="
-						"remount-ro was specified%s",
-						es1, es2);
-				goto iput_usnjrnl_err_out;
-			}
-			sb->s_flags |= SB_RDONLY;
-			ntfs_error(sb, "%s.  Mounting read-only%s", es1, es2);
-		} else
-			ntfs_warning(sb, "%s.  Will not be able to remount "
-					"read-write%s", es1, es2);
-		/* This will prevent a read-write remount. */
-		NVolSetErrors(vol);
-	}
-	/* If (still) a read-write mount, stamp the transaction log. */
-	if (!sb_rdonly(sb) && !ntfs_stamp_usnjrnl(vol)) {
-		static const char *es1 = "Failed to stamp transaction log "
-				"($UsnJrnl)";
-		static const char *es2 = ".  Run chkdsk.";
-
-		/* Convert to a read-only mount. */
-		if (!(vol->on_errors & (ON_ERRORS_REMOUNT_RO |
-				ON_ERRORS_CONTINUE))) {
-			ntfs_error(sb, "%s and neither on_errors=continue nor "
-					"on_errors=remount-ro was specified%s",
-					es1, es2);
-			goto iput_usnjrnl_err_out;
-		}
-		ntfs_error(sb, "%s.  Mounting read-only%s", es1, es2);
-		sb->s_flags |= SB_RDONLY;
-		NVolSetErrors(vol);
-	}
-#endif /* NTFS_RW */
-	return true;
-#ifdef NTFS_RW
-iput_usnjrnl_err_out:
-	iput(vol->usnjrnl_j_ino);
-	iput(vol->usnjrnl_max_ino);
-	iput(vol->usnjrnl_ino);
-iput_quota_err_out:
-	iput(vol->quota_q_ino);
-	iput(vol->quota_ino);
-	iput(vol->extend_ino);
-#endif /* NTFS_RW */
-iput_sec_err_out:
-	iput(vol->secure_ino);
-iput_root_err_out:
-	iput(vol->root_ino);
-iput_logfile_err_out:
-#ifdef NTFS_RW
-	iput(vol->logfile_ino);
-iput_vol_err_out:
-#endif /* NTFS_RW */
-	iput(vol->vol_ino);
-iput_lcnbmp_err_out:
-	iput(vol->lcnbmp_ino);
-iput_attrdef_err_out:
-	vol->attrdef_size = 0;
-	if (vol->attrdef) {
-		ntfs_free(vol->attrdef);
-		vol->attrdef = NULL;
-	}
-#ifdef NTFS_RW
-iput_upcase_err_out:
-#endif /* NTFS_RW */
-	vol->upcase_len = 0;
-	mutex_lock(&ntfs_lock);
-	if (vol->upcase == default_upcase) {
-		ntfs_nr_upcase_users--;
-		vol->upcase = NULL;
-	}
-	mutex_unlock(&ntfs_lock);
-	if (vol->upcase) {
-		ntfs_free(vol->upcase);
-		vol->upcase = NULL;
-	}
-iput_mftbmp_err_out:
-	iput(vol->mftbmp_ino);
-iput_mirr_err_out:
-#ifdef NTFS_RW
-	iput(vol->mftmirr_ino);
-#endif /* NTFS_RW */
-	return false;
-}
-
-/**
- * ntfs_put_super - called by the vfs to unmount a volume
- * @sb:		vfs superblock of volume to unmount
- *
- * ntfs_put_super() is called by the VFS (from fs/super.c::do_umount()) when
- * the volume is being unmounted (umount system call has been invoked) and it
- * releases all inodes and memory belonging to the NTFS specific part of the
- * super block.
- */
-static void ntfs_put_super(struct super_block *sb)
-{
-	ntfs_volume *vol = NTFS_SB(sb);
-
-	ntfs_debug("Entering.");
-
-#ifdef NTFS_RW
-	/*
-	 * Commit all inodes while they are still open in case some of them
-	 * cause others to be dirtied.
-	 */
-	ntfs_commit_inode(vol->vol_ino);
-
-	/* NTFS 3.0+ specific. */
-	if (vol->major_ver >= 3) {
-		if (vol->usnjrnl_j_ino)
-			ntfs_commit_inode(vol->usnjrnl_j_ino);
-		if (vol->usnjrnl_max_ino)
-			ntfs_commit_inode(vol->usnjrnl_max_ino);
-		if (vol->usnjrnl_ino)
-			ntfs_commit_inode(vol->usnjrnl_ino);
-		if (vol->quota_q_ino)
-			ntfs_commit_inode(vol->quota_q_ino);
-		if (vol->quota_ino)
-			ntfs_commit_inode(vol->quota_ino);
-		if (vol->extend_ino)
-			ntfs_commit_inode(vol->extend_ino);
-		if (vol->secure_ino)
-			ntfs_commit_inode(vol->secure_ino);
-	}
-
-	ntfs_commit_inode(vol->root_ino);
-
-	down_write(&vol->lcnbmp_lock);
-	ntfs_commit_inode(vol->lcnbmp_ino);
-	up_write(&vol->lcnbmp_lock);
-
-	down_write(&vol->mftbmp_lock);
-	ntfs_commit_inode(vol->mftbmp_ino);
-	up_write(&vol->mftbmp_lock);
-
-	if (vol->logfile_ino)
-		ntfs_commit_inode(vol->logfile_ino);
-
-	if (vol->mftmirr_ino)
-		ntfs_commit_inode(vol->mftmirr_ino);
-	ntfs_commit_inode(vol->mft_ino);
-
-	/*
-	 * If a read-write mount and no volume errors have occurred, mark the
-	 * volume clean.  Also, re-commit all affected inodes.
-	 */
-	if (!sb_rdonly(sb)) {
-		if (!NVolErrors(vol)) {
-			if (ntfs_clear_volume_flags(vol, VOLUME_IS_DIRTY))
-				ntfs_warning(sb, "Failed to clear dirty bit "
-						"in volume information "
-						"flags.  Run chkdsk.");
-			ntfs_commit_inode(vol->vol_ino);
-			ntfs_commit_inode(vol->root_ino);
-			if (vol->mftmirr_ino)
-				ntfs_commit_inode(vol->mftmirr_ino);
-			ntfs_commit_inode(vol->mft_ino);
-		} else {
-			ntfs_warning(sb, "Volume has errors.  Leaving volume "
-					"marked dirty.  Run chkdsk.");
-		}
-	}
-#endif /* NTFS_RW */
-
-	iput(vol->vol_ino);
-	vol->vol_ino = NULL;
-
-	/* NTFS 3.0+ specific clean up. */
-	if (vol->major_ver >= 3) {
-#ifdef NTFS_RW
-		if (vol->usnjrnl_j_ino) {
-			iput(vol->usnjrnl_j_ino);
-			vol->usnjrnl_j_ino = NULL;
-		}
-		if (vol->usnjrnl_max_ino) {
-			iput(vol->usnjrnl_max_ino);
-			vol->usnjrnl_max_ino = NULL;
-		}
-		if (vol->usnjrnl_ino) {
-			iput(vol->usnjrnl_ino);
-			vol->usnjrnl_ino = NULL;
-		}
-		if (vol->quota_q_ino) {
-			iput(vol->quota_q_ino);
-			vol->quota_q_ino = NULL;
-		}
-		if (vol->quota_ino) {
-			iput(vol->quota_ino);
-			vol->quota_ino = NULL;
-		}
-#endif /* NTFS_RW */
-		if (vol->extend_ino) {
-			iput(vol->extend_ino);
-			vol->extend_ino = NULL;
-		}
-		if (vol->secure_ino) {
-			iput(vol->secure_ino);
-			vol->secure_ino = NULL;
-		}
-	}
-
-	iput(vol->root_ino);
-	vol->root_ino = NULL;
-
-	down_write(&vol->lcnbmp_lock);
-	iput(vol->lcnbmp_ino);
-	vol->lcnbmp_ino = NULL;
-	up_write(&vol->lcnbmp_lock);
-
-	down_write(&vol->mftbmp_lock);
-	iput(vol->mftbmp_ino);
-	vol->mftbmp_ino = NULL;
-	up_write(&vol->mftbmp_lock);
-
-#ifdef NTFS_RW
-	if (vol->logfile_ino) {
-		iput(vol->logfile_ino);
-		vol->logfile_ino = NULL;
-	}
-	if (vol->mftmirr_ino) {
-		/* Re-commit the mft mirror and mft just in case. */
-		ntfs_commit_inode(vol->mftmirr_ino);
-		ntfs_commit_inode(vol->mft_ino);
-		iput(vol->mftmirr_ino);
-		vol->mftmirr_ino = NULL;
-	}
-	/*
-	 * We should have no dirty inodes left, due to
-	 * mft.c::ntfs_mft_writepage() cleaning all the dirty pages as
-	 * the underlying mft records are written out and cleaned.
-	 */
-	ntfs_commit_inode(vol->mft_ino);
-	write_inode_now(vol->mft_ino, 1);
-#endif /* NTFS_RW */
-
-	iput(vol->mft_ino);
-	vol->mft_ino = NULL;
-
-	/* Throw away the table of attribute definitions. */
-	vol->attrdef_size = 0;
-	if (vol->attrdef) {
-		ntfs_free(vol->attrdef);
-		vol->attrdef = NULL;
-	}
-	vol->upcase_len = 0;
-	/*
-	 * Destroy the global default upcase table if necessary.  Also decrease
-	 * the number of upcase users if we are a user.
-	 */
-	mutex_lock(&ntfs_lock);
-	if (vol->upcase == default_upcase) {
-		ntfs_nr_upcase_users--;
-		vol->upcase = NULL;
-	}
-	if (!ntfs_nr_upcase_users && default_upcase) {
-		ntfs_free(default_upcase);
-		default_upcase = NULL;
-	}
-	if (vol->cluster_size <= 4096 && !--ntfs_nr_compression_users)
-		free_compression_buffers();
-	mutex_unlock(&ntfs_lock);
-	if (vol->upcase) {
-		ntfs_free(vol->upcase);
-		vol->upcase = NULL;
-	}
-
-	unload_nls(vol->nls_map);
-
-	sb->s_fs_info = NULL;
-	kfree(vol);
-}
-
-/**
- * get_nr_free_clusters - return the number of free clusters on a volume
- * @vol:	ntfs volume for which to obtain free cluster count
- *
- * Calculate the number of free clusters on the mounted NTFS volume @vol. We
- * actually calculate the number of clusters in use instead because this
- * allows us to not care about partial pages as these will be just zero filled
- * and hence not be counted as allocated clusters.
- *
- * The only particularity is that clusters beyond the end of the logical ntfs
- * volume will be marked as allocated to prevent errors which means we have to
- * discount those at the end. This is important as the cluster bitmap always
- * has a size in multiples of 8 bytes, i.e. up to 63 clusters could be outside
- * the logical volume and marked in use when they are not as they do not exist.
- *
- * If any pages cannot be read we assume all clusters in the erroring pages are
- * in use. This means we return an underestimate on errors which is better than
- * an overestimate.
- */
-static s64 get_nr_free_clusters(ntfs_volume *vol)
-{
-	s64 nr_free = vol->nr_clusters;
-	struct address_space *mapping = vol->lcnbmp_ino->i_mapping;
-	struct page *page;
-	pgoff_t index, max_index;
-
-	ntfs_debug("Entering.");
-	/* Serialize accesses to the cluster bitmap. */
-	down_read(&vol->lcnbmp_lock);
-	/*
-	 * Convert the number of bits into bytes rounded up, then convert into
-	 * multiples of PAGE_SIZE, rounding up so that if we have one
-	 * full and one partial page max_index = 2.
-	 */
-	max_index = (((vol->nr_clusters + 7) >> 3) + PAGE_SIZE - 1) >>
-			PAGE_SHIFT;
-	/* Use multiples of 4 bytes, thus max_size is PAGE_SIZE / 4. */
-	ntfs_debug("Reading $Bitmap, max_index = 0x%lx, max_size = 0x%lx.",
-			max_index, PAGE_SIZE / 4);
-	for (index = 0; index < max_index; index++) {
-		unsigned long *kaddr;
-
-		/*
-		 * Read the page from page cache, getting it from backing store
-		 * if necessary, and increment the use count.
-		 */
-		page = read_mapping_page(mapping, index, NULL);
-		/* Ignore pages which errored synchronously. */
-		if (IS_ERR(page)) {
-			ntfs_debug("read_mapping_page() error. Skipping "
-					"page (index 0x%lx).", index);
-			nr_free -= PAGE_SIZE * 8;
-			continue;
-		}
-		kaddr = kmap_atomic(page);
-		/*
-		 * Subtract the number of set bits. If this
-		 * is the last page and it is partial we don't really care as
-		 * it just means we do a little extra work but it won't affect
-		 * the result as all out of range bytes are set to zero by
-		 * ntfs_readpage().
-		 */
-		nr_free -= bitmap_weight(kaddr,
-					PAGE_SIZE * BITS_PER_BYTE);
-		kunmap_atomic(kaddr);
-		put_page(page);
-	}
-	ntfs_debug("Finished reading $Bitmap, last index = 0x%lx.", index - 1);
-	/*
-	 * Fixup for eventual bits outside logical ntfs volume (see function
-	 * description above).
-	 */
-	if (vol->nr_clusters & 63)
-		nr_free += 64 - (vol->nr_clusters & 63);
-	up_read(&vol->lcnbmp_lock);
-	/* If errors occurred we may well have gone below zero, fix this. */
-	if (nr_free < 0)
-		nr_free = 0;
-	ntfs_debug("Exiting.");
-	return nr_free;
-}
-
-/**
- * __get_nr_free_mft_records - return the number of free inodes on a volume
- * @vol:	ntfs volume for which to obtain free inode count
- * @nr_free:	number of mft records in filesystem
- * @max_index:	maximum number of pages containing set bits
- *
- * Calculate the number of free mft records (inodes) on the mounted NTFS
- * volume @vol. We actually calculate the number of mft records in use instead
- * because this allows us to not care about partial pages as these will be just
- * zero filled and hence not be counted as allocated mft record.
- *
- * If any pages cannot be read we assume all mft records in the erroring pages
- * are in use. This means we return an underestimate on errors which is better
- * than an overestimate.
- *
- * NOTE: Caller must hold mftbmp_lock rw_semaphore for reading or writing.
- */
-static unsigned long __get_nr_free_mft_records(ntfs_volume *vol,
-		s64 nr_free, const pgoff_t max_index)
-{
-	struct address_space *mapping = vol->mftbmp_ino->i_mapping;
-	struct page *page;
-	pgoff_t index;
-
-	ntfs_debug("Entering.");
-	/* Use multiples of 4 bytes, thus max_size is PAGE_SIZE / 4. */
-	ntfs_debug("Reading $MFT/$BITMAP, max_index = 0x%lx, max_size = "
-			"0x%lx.", max_index, PAGE_SIZE / 4);
-	for (index = 0; index < max_index; index++) {
-		unsigned long *kaddr;
-
-		/*
-		 * Read the page from page cache, getting it from backing store
-		 * if necessary, and increment the use count.
-		 */
-		page = read_mapping_page(mapping, index, NULL);
-		/* Ignore pages which errored synchronously. */
-		if (IS_ERR(page)) {
-			ntfs_debug("read_mapping_page() error. Skipping "
-					"page (index 0x%lx).", index);
-			nr_free -= PAGE_SIZE * 8;
-			continue;
-		}
-		kaddr = kmap_atomic(page);
-		/*
-		 * Subtract the number of set bits. If this
-		 * is the last page and it is partial we don't really care as
-		 * it just means we do a little extra work but it won't affect
-		 * the result as all out of range bytes are set to zero by
-		 * ntfs_readpage().
-		 */
-		nr_free -= bitmap_weight(kaddr,
-					PAGE_SIZE * BITS_PER_BYTE);
-		kunmap_atomic(kaddr);
-		put_page(page);
-	}
-	ntfs_debug("Finished reading $MFT/$BITMAP, last index = 0x%lx.",
-			index - 1);
-	/* If errors occurred we may well have gone below zero, fix this. */
-	if (nr_free < 0)
-		nr_free = 0;
-	ntfs_debug("Exiting.");
-	return nr_free;
-}
-
-/**
- * ntfs_statfs - return information about mounted NTFS volume
- * @dentry:	dentry from mounted volume
- * @sfs:	statfs structure in which to return the information
- *
- * Return information about the mounted NTFS volume @dentry in the statfs structure
- * pointed to by @sfs (this is initialized with zeros before ntfs_statfs is
- * called). We interpret the values to be correct of the moment in time at
- * which we are called. Most values are variable otherwise and this isn't just
- * the free values but the totals as well. For example we can increase the
- * total number of file nodes if we run out and we can keep doing this until
- * there is no more space on the volume left at all.
- *
- * Called from vfs_statfs which is used to handle the statfs, fstatfs, and
- * ustat system calls.
- *
- * Return 0 on success or -errno on error.
- */
-static int ntfs_statfs(struct dentry *dentry, struct kstatfs *sfs)
-{
-	struct super_block *sb = dentry->d_sb;
-	s64 size;
-	ntfs_volume *vol = NTFS_SB(sb);
-	ntfs_inode *mft_ni = NTFS_I(vol->mft_ino);
-	pgoff_t max_index;
-	unsigned long flags;
-
-	ntfs_debug("Entering.");
-	/* Type of filesystem. */
-	sfs->f_type   = NTFS_SB_MAGIC;
-	/* Optimal transfer block size. */
-	sfs->f_bsize  = PAGE_SIZE;
-	/*
-	 * Total data blocks in filesystem in units of f_bsize and since
-	 * inodes are also stored in data blocs ($MFT is a file) this is just
-	 * the total clusters.
-	 */
-	sfs->f_blocks = vol->nr_clusters << vol->cluster_size_bits >>
-				PAGE_SHIFT;
-	/* Free data blocks in filesystem in units of f_bsize. */
-	size	      = get_nr_free_clusters(vol) << vol->cluster_size_bits >>
-				PAGE_SHIFT;
-	if (size < 0LL)
-		size = 0LL;
-	/* Free blocks avail to non-superuser, same as above on NTFS. */
-	sfs->f_bavail = sfs->f_bfree = size;
-	/* Serialize accesses to the inode bitmap. */
-	down_read(&vol->mftbmp_lock);
-	read_lock_irqsave(&mft_ni->size_lock, flags);
-	size = i_size_read(vol->mft_ino) >> vol->mft_record_size_bits;
-	/*
-	 * Convert the maximum number of set bits into bytes rounded up, then
-	 * convert into multiples of PAGE_SIZE, rounding up so that if we
-	 * have one full and one partial page max_index = 2.
-	 */
-	max_index = ((((mft_ni->initialized_size >> vol->mft_record_size_bits)
-			+ 7) >> 3) + PAGE_SIZE - 1) >> PAGE_SHIFT;
-	read_unlock_irqrestore(&mft_ni->size_lock, flags);
-	/* Number of inodes in filesystem (at this point in time). */
-	sfs->f_files = size;
-	/* Free inodes in fs (based on current total count). */
-	sfs->f_ffree = __get_nr_free_mft_records(vol, size, max_index);
-	up_read(&vol->mftbmp_lock);
-	/*
-	 * File system id. This is extremely *nix flavour dependent and even
-	 * within Linux itself all fs do their own thing. I interpret this to
-	 * mean a unique id associated with the mounted fs and not the id
-	 * associated with the filesystem driver, the latter is already given
-	 * by the filesystem type in sfs->f_type. Thus we use the 64-bit
-	 * volume serial number splitting it into two 32-bit parts. We enter
-	 * the least significant 32-bits in f_fsid[0] and the most significant
-	 * 32-bits in f_fsid[1].
-	 */
-	sfs->f_fsid = u64_to_fsid(vol->serial_no);
-	/* Maximum length of filenames. */
-	sfs->f_namelen	   = NTFS_MAX_NAME_LEN;
-	return 0;
-}
-
-#ifdef NTFS_RW
-static int ntfs_write_inode(struct inode *vi, struct writeback_control *wbc)
-{
-	return __ntfs_write_inode(vi, wbc->sync_mode == WB_SYNC_ALL);
-}
-#endif
-
-/*
- * The complete super operations.
- */
-static const struct super_operations ntfs_sops = {
-	.alloc_inode	= ntfs_alloc_big_inode,	  /* VFS: Allocate new inode. */
-	.free_inode	= ntfs_free_big_inode, /* VFS: Deallocate inode. */
-#ifdef NTFS_RW
-	.write_inode	= ntfs_write_inode,	/* VFS: Write dirty inode to
-						   disk. */
-#endif /* NTFS_RW */
-	.put_super	= ntfs_put_super,	/* Syscall: umount. */
-	.statfs		= ntfs_statfs,		/* Syscall: statfs */
-	.remount_fs	= ntfs_remount,		/* Syscall: mount -o remount. */
-	.evict_inode	= ntfs_evict_big_inode,	/* VFS: Called when an inode is
-						   removed from memory. */
-	.show_options	= ntfs_show_options,	/* Show mount options in
-						   proc. */
-};
-
-/**
- * ntfs_fill_super - mount an ntfs filesystem
- * @sb:		super block of ntfs filesystem to mount
- * @opt:	string containing the mount options
- * @silent:	silence error output
- *
- * ntfs_fill_super() is called by the VFS to mount the device described by @sb
- * with the mount otions in @data with the NTFS filesystem.
- *
- * If @silent is true, remain silent even if errors are detected. This is used
- * during bootup, when the kernel tries to mount the root filesystem with all
- * registered filesystems one after the other until one succeeds. This implies
- * that all filesystems except the correct one will quite correctly and
- * expectedly return an error, but nobody wants to see error messages when in
- * fact this is what is supposed to happen.
- *
- * NOTE: @sb->s_flags contains the mount options flags.
- */
-static int ntfs_fill_super(struct super_block *sb, void *opt, const int silent)
-{
-	ntfs_volume *vol;
-	struct buffer_head *bh;
-	struct inode *tmp_ino;
-	int blocksize, result;
-
-	/*
-	 * We do a pretty difficult piece of bootstrap by reading the
-	 * MFT (and other metadata) from disk into memory. We'll only
-	 * release this metadata during umount, so the locking patterns
-	 * observed during bootstrap do not count. So turn off the
-	 * observation of locking patterns (strictly for this context
-	 * only) while mounting NTFS. [The validator is still active
-	 * otherwise, even for this context: it will for example record
-	 * lock class registrations.]
-	 */
-	lockdep_off();
-	ntfs_debug("Entering.");
-#ifndef NTFS_RW
-	sb->s_flags |= SB_RDONLY;
-#endif /* ! NTFS_RW */
-	/* Allocate a new ntfs_volume and place it in sb->s_fs_info. */
-	sb->s_fs_info = kmalloc(sizeof(ntfs_volume), GFP_NOFS);
-	vol = NTFS_SB(sb);
-	if (!vol) {
-		if (!silent)
-			ntfs_error(sb, "Allocation of NTFS volume structure "
-					"failed. Aborting mount...");
-		lockdep_on();
-		return -ENOMEM;
-	}
-	/* Initialize ntfs_volume structure. */
-	*vol = (ntfs_volume) {
-		.sb = sb,
-		/*
-		 * Default is group and other don't have any access to files or
-		 * directories while owner has full access. Further, files by
-		 * default are not executable but directories are of course
-		 * browseable.
-		 */
-		.fmask = 0177,
-		.dmask = 0077,
-	};
-	init_rwsem(&vol->mftbmp_lock);
-	init_rwsem(&vol->lcnbmp_lock);
-
-	/* By default, enable sparse support. */
-	NVolSetSparseEnabled(vol);
-
-	/* Important to get the mount options dealt with now. */
-	if (!parse_options(vol, (char*)opt))
-		goto err_out_now;
-
-	/* We support sector sizes up to the PAGE_SIZE. */
-	if (bdev_logical_block_size(sb->s_bdev) > PAGE_SIZE) {
-		if (!silent)
-			ntfs_error(sb, "Device has unsupported sector size "
-					"(%i).  The maximum supported sector "
-					"size on this architecture is %lu "
-					"bytes.",
-					bdev_logical_block_size(sb->s_bdev),
-					PAGE_SIZE);
-		goto err_out_now;
-	}
-	/*
-	 * Setup the device access block size to NTFS_BLOCK_SIZE or the hard
-	 * sector size, whichever is bigger.
-	 */
-	blocksize = sb_min_blocksize(sb, NTFS_BLOCK_SIZE);
-	if (blocksize < NTFS_BLOCK_SIZE) {
-		if (!silent)
-			ntfs_error(sb, "Unable to set device block size.");
-		goto err_out_now;
-	}
-	BUG_ON(blocksize != sb->s_blocksize);
-	ntfs_debug("Set device block size to %i bytes (block size bits %i).",
-			blocksize, sb->s_blocksize_bits);
-	/* Determine the size of the device in units of block_size bytes. */
-	vol->nr_blocks = sb_bdev_nr_blocks(sb);
-	if (!vol->nr_blocks) {
-		if (!silent)
-			ntfs_error(sb, "Unable to determine device size.");
-		goto err_out_now;
-	}
-	/* Read the boot sector and return unlocked buffer head to it. */
-	if (!(bh = read_ntfs_boot_sector(sb, silent))) {
-		if (!silent)
-			ntfs_error(sb, "Not an NTFS volume.");
-		goto err_out_now;
-	}
-	/*
-	 * Extract the data from the boot sector and setup the ntfs volume
-	 * using it.
-	 */
-	result = parse_ntfs_boot_sector(vol, (NTFS_BOOT_SECTOR*)bh->b_data);
-	brelse(bh);
-	if (!result) {
-		if (!silent)
-			ntfs_error(sb, "Unsupported NTFS filesystem.");
-		goto err_out_now;
-	}
-	/*
-	 * If the boot sector indicates a sector size bigger than the current
-	 * device block size, switch the device block size to the sector size.
-	 * TODO: It may be possible to support this case even when the set
-	 * below fails, we would just be breaking up the i/o for each sector
-	 * into multiple blocks for i/o purposes but otherwise it should just
-	 * work.  However it is safer to leave disabled until someone hits this
-	 * error message and then we can get them to try it without the setting
-	 * so we know for sure that it works.
-	 */
-	if (vol->sector_size > blocksize) {
-		blocksize = sb_set_blocksize(sb, vol->sector_size);
-		if (blocksize != vol->sector_size) {
-			if (!silent)
-				ntfs_error(sb, "Unable to set device block "
-						"size to sector size (%i).",
-						vol->sector_size);
-			goto err_out_now;
-		}
-		BUG_ON(blocksize != sb->s_blocksize);
-		vol->nr_blocks = sb_bdev_nr_blocks(sb);
-		ntfs_debug("Changed device block size to %i bytes (block size "
-				"bits %i) to match volume sector size.",
-				blocksize, sb->s_blocksize_bits);
-	}
-	/* Initialize the cluster and mft allocators. */
-	ntfs_setup_allocators(vol);
-	/* Setup remaining fields in the super block. */
-	sb->s_magic = NTFS_SB_MAGIC;
-	/*
-	 * Ntfs allows 63 bits for the file size, i.e. correct would be:
-	 *	sb->s_maxbytes = ~0ULL >> 1;
-	 * But the kernel uses a long as the page cache page index which on
-	 * 32-bit architectures is only 32-bits. MAX_LFS_FILESIZE is kernel
-	 * defined to the maximum the page cache page index can cope with
-	 * without overflowing the index or to 2^63 - 1, whichever is smaller.
-	 */
-	sb->s_maxbytes = MAX_LFS_FILESIZE;
-	/* Ntfs measures time in 100ns intervals. */
-	sb->s_time_gran = 100;
-	/*
-	 * Now load the metadata required for the page cache and our address
-	 * space operations to function. We do this by setting up a specialised
-	 * read_inode method and then just calling the normal iget() to obtain
-	 * the inode for $MFT which is sufficient to allow our normal inode
-	 * operations and associated address space operations to function.
-	 */
-	sb->s_op = &ntfs_sops;
-	tmp_ino = new_inode(sb);
-	if (!tmp_ino) {
-		if (!silent)
-			ntfs_error(sb, "Failed to load essential metadata.");
-		goto err_out_now;
-	}
-	tmp_ino->i_ino = FILE_MFT;
-	insert_inode_hash(tmp_ino);
-	if (ntfs_read_inode_mount(tmp_ino) < 0) {
-		if (!silent)
-			ntfs_error(sb, "Failed to load essential metadata.");
-		goto iput_tmp_ino_err_out_now;
-	}
-	mutex_lock(&ntfs_lock);
-	/*
-	 * The current mount is a compression user if the cluster size is
-	 * less than or equal 4kiB.
-	 */
-	if (vol->cluster_size <= 4096 && !ntfs_nr_compression_users++) {
-		result = allocate_compression_buffers();
-		if (result) {
-			ntfs_error(NULL, "Failed to allocate buffers "
-					"for compression engine.");
-			ntfs_nr_compression_users--;
-			mutex_unlock(&ntfs_lock);
-			goto iput_tmp_ino_err_out_now;
-		}
-	}
-	/*
-	 * Generate the global default upcase table if necessary.  Also
-	 * temporarily increment the number of upcase users to avoid race
-	 * conditions with concurrent (u)mounts.
-	 */
-	if (!default_upcase)
-		default_upcase = generate_default_upcase();
-	ntfs_nr_upcase_users++;
-	mutex_unlock(&ntfs_lock);
-	/*
-	 * From now on, ignore @silent parameter. If we fail below this line,
-	 * it will be due to a corrupt fs or a system error, so we report it.
-	 */
-	/*
-	 * Open the system files with normal access functions and complete
-	 * setting up the ntfs super block.
-	 */
-	if (!load_system_files(vol)) {
-		ntfs_error(sb, "Failed to load system files.");
-		goto unl_upcase_iput_tmp_ino_err_out_now;
-	}
-
-	/* We grab a reference, simulating an ntfs_iget(). */
-	ihold(vol->root_ino);
-	if ((sb->s_root = d_make_root(vol->root_ino))) {
-		ntfs_debug("Exiting, status successful.");
-		/* Release the default upcase if it has no users. */
-		mutex_lock(&ntfs_lock);
-		if (!--ntfs_nr_upcase_users && default_upcase) {
-			ntfs_free(default_upcase);
-			default_upcase = NULL;
-		}
-		mutex_unlock(&ntfs_lock);
-		sb->s_export_op = &ntfs_export_ops;
-		lockdep_on();
-		return 0;
-	}
-	ntfs_error(sb, "Failed to allocate root directory.");
-	/* Clean up after the successful load_system_files() call from above. */
-	// TODO: Use ntfs_put_super() instead of repeating all this code...
-	// FIXME: Should mark the volume clean as the error is most likely
-	// 	  -ENOMEM.
-	iput(vol->vol_ino);
-	vol->vol_ino = NULL;
-	/* NTFS 3.0+ specific clean up. */
-	if (vol->major_ver >= 3) {
-#ifdef NTFS_RW
-		if (vol->usnjrnl_j_ino) {
-			iput(vol->usnjrnl_j_ino);
-			vol->usnjrnl_j_ino = NULL;
-		}
-		if (vol->usnjrnl_max_ino) {
-			iput(vol->usnjrnl_max_ino);
-			vol->usnjrnl_max_ino = NULL;
-		}
-		if (vol->usnjrnl_ino) {
-			iput(vol->usnjrnl_ino);
-			vol->usnjrnl_ino = NULL;
-		}
-		if (vol->quota_q_ino) {
-			iput(vol->quota_q_ino);
-			vol->quota_q_ino = NULL;
-		}
-		if (vol->quota_ino) {
-			iput(vol->quota_ino);
-			vol->quota_ino = NULL;
-		}
-#endif /* NTFS_RW */
-		if (vol->extend_ino) {
-			iput(vol->extend_ino);
-			vol->extend_ino = NULL;
-		}
-		if (vol->secure_ino) {
-			iput(vol->secure_ino);
-			vol->secure_ino = NULL;
-		}
-	}
-	iput(vol->root_ino);
-	vol->root_ino = NULL;
-	iput(vol->lcnbmp_ino);
-	vol->lcnbmp_ino = NULL;
-	iput(vol->mftbmp_ino);
-	vol->mftbmp_ino = NULL;
-#ifdef NTFS_RW
-	if (vol->logfile_ino) {
-		iput(vol->logfile_ino);
-		vol->logfile_ino = NULL;
-	}
-	if (vol->mftmirr_ino) {
-		iput(vol->mftmirr_ino);
-		vol->mftmirr_ino = NULL;
-	}
-#endif /* NTFS_RW */
-	/* Throw away the table of attribute definitions. */
-	vol->attrdef_size = 0;
-	if (vol->attrdef) {
-		ntfs_free(vol->attrdef);
-		vol->attrdef = NULL;
-	}
-	vol->upcase_len = 0;
-	mutex_lock(&ntfs_lock);
-	if (vol->upcase == default_upcase) {
-		ntfs_nr_upcase_users--;
-		vol->upcase = NULL;
-	}
-	mutex_unlock(&ntfs_lock);
-	if (vol->upcase) {
-		ntfs_free(vol->upcase);
-		vol->upcase = NULL;
-	}
-	if (vol->nls_map) {
-		unload_nls(vol->nls_map);
-		vol->nls_map = NULL;
-	}
-	/* Error exit code path. */
-unl_upcase_iput_tmp_ino_err_out_now:
-	/*
-	 * Decrease the number of upcase users and destroy the global default
-	 * upcase table if necessary.
-	 */
-	mutex_lock(&ntfs_lock);
-	if (!--ntfs_nr_upcase_users && default_upcase) {
-		ntfs_free(default_upcase);
-		default_upcase = NULL;
-	}
-	if (vol->cluster_size <= 4096 && !--ntfs_nr_compression_users)
-		free_compression_buffers();
-	mutex_unlock(&ntfs_lock);
-iput_tmp_ino_err_out_now:
-	iput(tmp_ino);
-	if (vol->mft_ino && vol->mft_ino != tmp_ino)
-		iput(vol->mft_ino);
-	vol->mft_ino = NULL;
-	/* Errors at this stage are irrelevant. */
-err_out_now:
-	sb->s_fs_info = NULL;
-	kfree(vol);
-	ntfs_debug("Failed, returning -EINVAL.");
-	lockdep_on();
-	return -EINVAL;
-}
-
-/*
- * This is a slab cache to optimize allocations and deallocations of Unicode
- * strings of the maximum length allowed by NTFS, which is NTFS_MAX_NAME_LEN
- * (255) Unicode characters + a terminating NULL Unicode character.
- */
-struct kmem_cache *ntfs_name_cache;
-
-/* Slab caches for efficient allocation/deallocation of inodes. */
-struct kmem_cache *ntfs_inode_cache;
-struct kmem_cache *ntfs_big_inode_cache;
-
-/* Init once constructor for the inode slab cache. */
-static void ntfs_big_inode_init_once(void *foo)
-{
-	ntfs_inode *ni = (ntfs_inode *)foo;
-
-	inode_init_once(VFS_I(ni));
-}
-
-/*
- * Slab caches to optimize allocations and deallocations of attribute search
- * contexts and index contexts, respectively.
- */
-struct kmem_cache *ntfs_attr_ctx_cache;
-struct kmem_cache *ntfs_index_ctx_cache;
-
-/* Driver wide mutex. */
-DEFINE_MUTEX(ntfs_lock);
-
-static struct dentry *ntfs_mount(struct file_system_type *fs_type,
-	int flags, const char *dev_name, void *data)
-{
-	return mount_bdev(fs_type, flags, dev_name, data, ntfs_fill_super);
-}
-
-static struct file_system_type ntfs_fs_type = {
-	.owner		= THIS_MODULE,
-	.name		= "ntfs",
-	.mount		= ntfs_mount,
-	.kill_sb	= kill_block_super,
-	.fs_flags	= FS_REQUIRES_DEV,
-};
-MODULE_ALIAS_FS("ntfs");
-
-/* Stable names for the slab caches. */
-static const char ntfs_index_ctx_cache_name[] = "ntfs_index_ctx_cache";
-static const char ntfs_attr_ctx_cache_name[] = "ntfs_attr_ctx_cache";
-static const char ntfs_name_cache_name[] = "ntfs_name_cache";
-static const char ntfs_inode_cache_name[] = "ntfs_inode_cache";
-static const char ntfs_big_inode_cache_name[] = "ntfs_big_inode_cache";
-
-static int __init init_ntfs_fs(void)
-{
-	int err = 0;
-
-	/* This may be ugly but it results in pretty output so who cares. (-8 */
-	pr_info("driver " NTFS_VERSION " [Flags: R/"
-#ifdef NTFS_RW
-			"W"
-#else
-			"O"
-#endif
-#ifdef DEBUG
-			" DEBUG"
-#endif
-#ifdef MODULE
-			" MODULE"
-#endif
-			"].\n");
-
-	ntfs_debug("Debug messages are enabled.");
-
-	ntfs_index_ctx_cache = kmem_cache_create(ntfs_index_ctx_cache_name,
-			sizeof(ntfs_index_context), 0 /* offset */,
-			SLAB_HWCACHE_ALIGN, NULL /* ctor */);
-	if (!ntfs_index_ctx_cache) {
-		pr_crit("Failed to create %s!\n", ntfs_index_ctx_cache_name);
-		goto ictx_err_out;
-	}
-	ntfs_attr_ctx_cache = kmem_cache_create(ntfs_attr_ctx_cache_name,
-			sizeof(ntfs_attr_search_ctx), 0 /* offset */,
-			SLAB_HWCACHE_ALIGN, NULL /* ctor */);
-	if (!ntfs_attr_ctx_cache) {
-		pr_crit("NTFS: Failed to create %s!\n",
-			ntfs_attr_ctx_cache_name);
-		goto actx_err_out;
-	}
-
-	ntfs_name_cache = kmem_cache_create(ntfs_name_cache_name,
-			(NTFS_MAX_NAME_LEN+1) * sizeof(ntfschar), 0,
-			SLAB_HWCACHE_ALIGN, NULL);
-	if (!ntfs_name_cache) {
-		pr_crit("Failed to create %s!\n", ntfs_name_cache_name);
-		goto name_err_out;
-	}
-
-	ntfs_inode_cache = kmem_cache_create(ntfs_inode_cache_name,
-			sizeof(ntfs_inode), 0,
-			SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD, NULL);
-	if (!ntfs_inode_cache) {
-		pr_crit("Failed to create %s!\n", ntfs_inode_cache_name);
-		goto inode_err_out;
-	}
-
-	ntfs_big_inode_cache = kmem_cache_create(ntfs_big_inode_cache_name,
-			sizeof(big_ntfs_inode), 0,
-			SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD|
-			SLAB_ACCOUNT, ntfs_big_inode_init_once);
-	if (!ntfs_big_inode_cache) {
-		pr_crit("Failed to create %s!\n", ntfs_big_inode_cache_name);
-		goto big_inode_err_out;
-	}
-
-	/* Register the ntfs sysctls. */
-	err = ntfs_sysctl(1);
-	if (err) {
-		pr_crit("Failed to register NTFS sysctls!\n");
-		goto sysctl_err_out;
-	}
-
-	err = register_filesystem(&ntfs_fs_type);
-	if (!err) {
-		ntfs_debug("NTFS driver registered successfully.");
-		return 0; /* Success! */
-	}
-	pr_crit("Failed to register NTFS filesystem driver!\n");
-
-	/* Unregister the ntfs sysctls. */
-	ntfs_sysctl(0);
-sysctl_err_out:
-	kmem_cache_destroy(ntfs_big_inode_cache);
-big_inode_err_out:
-	kmem_cache_destroy(ntfs_inode_cache);
-inode_err_out:
-	kmem_cache_destroy(ntfs_name_cache);
-name_err_out:
-	kmem_cache_destroy(ntfs_attr_ctx_cache);
-actx_err_out:
-	kmem_cache_destroy(ntfs_index_ctx_cache);
-ictx_err_out:
-	if (!err) {
-		pr_crit("Aborting NTFS filesystem driver registration...\n");
-		err = -ENOMEM;
-	}
-	return err;
-}
-
-static void __exit exit_ntfs_fs(void)
-{
-	ntfs_debug("Unregistering NTFS driver.");
-
-	unregister_filesystem(&ntfs_fs_type);
-
-	/*
-	 * Make sure all delayed rcu free inodes are flushed before we
-	 * destroy cache.
-	 */
-	rcu_barrier();
-	kmem_cache_destroy(ntfs_big_inode_cache);
-	kmem_cache_destroy(ntfs_inode_cache);
-	kmem_cache_destroy(ntfs_name_cache);
-	kmem_cache_destroy(ntfs_attr_ctx_cache);
-	kmem_cache_destroy(ntfs_index_ctx_cache);
-	/* Unregister the ntfs sysctls. */
-	ntfs_sysctl(0);
-}
-
-MODULE_AUTHOR("Anton Altaparmakov <anton@tuxera.com>");
-MODULE_DESCRIPTION("NTFS 1.2/3.x driver - Copyright (c) 2001-2014 Anton Altaparmakov and Tuxera Inc.");
-MODULE_VERSION(NTFS_VERSION);
-MODULE_LICENSE("GPL");
-#ifdef DEBUG
-module_param(debug_msgs, bint, 0);
-MODULE_PARM_DESC(debug_msgs, "Enable debug messages.");
-#endif
-
-module_init(init_ntfs_fs)
-module_exit(exit_ntfs_fs)
diff --git a/fs/ntfs/sysctl.c b/fs/ntfs/sysctl.c
deleted file mode 100644
index 4e980170d8..0000000000
--- a/fs/ntfs/sysctl.c
+++ /dev/null
@@ -1,58 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * sysctl.c - Code for sysctl handling in NTFS Linux kernel driver. Part of
- *	      the Linux-NTFS project. Adapted from the old NTFS driver,
- *	      Copyright (C) 1997 Martin von Löwis, Régis Duchesne
- *
- * Copyright (c) 2002-2005 Anton Altaparmakov
- */
-
-#ifdef DEBUG
-
-#include <linux/module.h>
-
-#ifdef CONFIG_SYSCTL
-
-#include <linux/proc_fs.h>
-#include <linux/sysctl.h>
-
-#include "sysctl.h"
-#include "debug.h"
-
-/* Definition of the ntfs sysctl. */
-static struct ctl_table ntfs_sysctls[] = {
-	{
-		.procname	= "ntfs-debug",
-		.data		= &debug_msgs,		/* Data pointer and size. */
-		.maxlen		= sizeof(debug_msgs),
-		.mode		= 0644,			/* Mode, proc handler. */
-		.proc_handler	= proc_dointvec
-	},
-};
-
-/* Storage for the sysctls header. */
-static struct ctl_table_header *sysctls_root_table;
-
-/**
- * ntfs_sysctl - add or remove the debug sysctl
- * @add:	add (1) or remove (0) the sysctl
- *
- * Add or remove the debug sysctl. Return 0 on success or -errno on error.
- */
-int ntfs_sysctl(int add)
-{
-	if (add) {
-		BUG_ON(sysctls_root_table);
-		sysctls_root_table = register_sysctl("fs", ntfs_sysctls);
-		if (!sysctls_root_table)
-			return -ENOMEM;
-	} else {
-		BUG_ON(!sysctls_root_table);
-		unregister_sysctl_table(sysctls_root_table);
-		sysctls_root_table = NULL;
-	}
-	return 0;
-}
-
-#endif /* CONFIG_SYSCTL */
-#endif /* DEBUG */
diff --git a/fs/ntfs/sysctl.h b/fs/ntfs/sysctl.h
deleted file mode 100644
index 96bb2299d2..0000000000
--- a/fs/ntfs/sysctl.h
+++ /dev/null
@@ -1,27 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * sysctl.h - Defines for sysctl handling in NTFS Linux kernel driver. Part of
- *	      the Linux-NTFS project. Adapted from the old NTFS driver,
- *	      Copyright (C) 1997 Martin von Löwis, Régis Duchesne
- *
- * Copyright (c) 2002-2004 Anton Altaparmakov
- */
-
-#ifndef _LINUX_NTFS_SYSCTL_H
-#define _LINUX_NTFS_SYSCTL_H
-
-
-#if defined(DEBUG) && defined(CONFIG_SYSCTL)
-
-extern int ntfs_sysctl(int add);
-
-#else
-
-/* Just return success. */
-static inline int ntfs_sysctl(int add)
-{
-	return 0;
-}
-
-#endif /* DEBUG && CONFIG_SYSCTL */
-#endif /* _LINUX_NTFS_SYSCTL_H */
diff --git a/fs/ntfs/time.h b/fs/ntfs/time.h
deleted file mode 100644
index 6b63261300..0000000000
--- a/fs/ntfs/time.h
+++ /dev/null
@@ -1,89 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * time.h - NTFS time conversion functions.  Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2005 Anton Altaparmakov
- */
-
-#ifndef _LINUX_NTFS_TIME_H
-#define _LINUX_NTFS_TIME_H
-
-#include <linux/time.h>		/* For current_kernel_time(). */
-#include <asm/div64.h>		/* For do_div(). */
-
-#include "endian.h"
-
-#define NTFS_TIME_OFFSET ((s64)(369 * 365 + 89) * 24 * 3600 * 10000000)
-
-/**
- * utc2ntfs - convert Linux UTC time to NTFS time
- * @ts:		Linux UTC time to convert to NTFS time
- *
- * Convert the Linux UTC time @ts to its corresponding NTFS time and return
- * that in little endian format.
- *
- * Linux stores time in a struct timespec64 consisting of a time64_t tv_sec
- * and a long tv_nsec where tv_sec is the number of 1-second intervals since
- * 1st January 1970, 00:00:00 UTC and tv_nsec is the number of 1-nano-second
- * intervals since the value of tv_sec.
- *
- * NTFS uses Microsoft's standard time format which is stored in a s64 and is
- * measured as the number of 100-nano-second intervals since 1st January 1601,
- * 00:00:00 UTC.
- */
-static inline sle64 utc2ntfs(const struct timespec64 ts)
-{
-	/*
-	 * Convert the seconds to 100ns intervals, add the nano-seconds
-	 * converted to 100ns intervals, and then add the NTFS time offset.
-	 */
-	return cpu_to_sle64((s64)ts.tv_sec * 10000000 + ts.tv_nsec / 100 +
-			NTFS_TIME_OFFSET);
-}
-
-/**
- * get_current_ntfs_time - get the current time in little endian NTFS format
- *
- * Get the current time from the Linux kernel, convert it to its corresponding
- * NTFS time and return that in little endian format.
- */
-static inline sle64 get_current_ntfs_time(void)
-{
-	struct timespec64 ts;
-
-	ktime_get_coarse_real_ts64(&ts);
-	return utc2ntfs(ts);
-}
-
-/**
- * ntfs2utc - convert NTFS time to Linux time
- * @time:	NTFS time (little endian) to convert to Linux UTC
- *
- * Convert the little endian NTFS time @time to its corresponding Linux UTC
- * time and return that in cpu format.
- *
- * Linux stores time in a struct timespec64 consisting of a time64_t tv_sec
- * and a long tv_nsec where tv_sec is the number of 1-second intervals since
- * 1st January 1970, 00:00:00 UTC and tv_nsec is the number of 1-nano-second
- * intervals since the value of tv_sec.
- *
- * NTFS uses Microsoft's standard time format which is stored in a s64 and is
- * measured as the number of 100 nano-second intervals since 1st January 1601,
- * 00:00:00 UTC.
- */
-static inline struct timespec64 ntfs2utc(const sle64 time)
-{
-	struct timespec64 ts;
-
-	/* Subtract the NTFS time offset. */
-	u64 t = (u64)(sle64_to_cpu(time) - NTFS_TIME_OFFSET);
-	/*
-	 * Convert the time to 1-second intervals and the remainder to
-	 * 1-nano-second intervals.
-	 */
-	ts.tv_nsec = do_div(t, 10000000) * 100;
-	ts.tv_sec = t;
-	return ts;
-}
-
-#endif /* _LINUX_NTFS_TIME_H */
diff --git a/fs/ntfs/types.h b/fs/ntfs/types.h
deleted file mode 100644
index 9a47859e7a..0000000000
--- a/fs/ntfs/types.h
+++ /dev/null
@@ -1,55 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * types.h - Defines for NTFS Linux kernel driver specific types.
- *	     Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2005 Anton Altaparmakov
- */
-
-#ifndef _LINUX_NTFS_TYPES_H
-#define _LINUX_NTFS_TYPES_H
-
-#include <linux/types.h>
-
-typedef __le16 le16;
-typedef __le32 le32;
-typedef __le64 le64;
-typedef __u16 __bitwise sle16;
-typedef __u32 __bitwise sle32;
-typedef __u64 __bitwise sle64;
-
-/* 2-byte Unicode character type. */
-typedef le16 ntfschar;
-#define UCHAR_T_SIZE_BITS 1
-
-/*
- * Clusters are signed 64-bit values on NTFS volumes. We define two types, LCN
- * and VCN, to allow for type checking and better code readability.
- */
-typedef s64 VCN;
-typedef sle64 leVCN;
-typedef s64 LCN;
-typedef sle64 leLCN;
-
-/*
- * The NTFS journal $LogFile uses log sequence numbers which are signed 64-bit
- * values.  We define our own type LSN, to allow for type checking and better
- * code readability.
- */
-typedef s64 LSN;
-typedef sle64 leLSN;
-
-/*
- * The NTFS transaction log $UsnJrnl uses usn which are signed 64-bit values.
- * We define our own type USN, to allow for type checking and better code
- * readability.
- */
-typedef s64 USN;
-typedef sle64 leUSN;
-
-typedef enum {
-	CASE_SENSITIVE = 0,
-	IGNORE_CASE = 1,
-} IGNORE_CASE_BOOL;
-
-#endif /* _LINUX_NTFS_TYPES_H */
diff --git a/fs/ntfs/unistr.c b/fs/ntfs/unistr.c
deleted file mode 100644
index a6b6c64f14..0000000000
--- a/fs/ntfs/unistr.c
+++ /dev/null
@@ -1,384 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * unistr.c - NTFS Unicode string handling. Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2006 Anton Altaparmakov
- */
-
-#include <linux/slab.h>
-
-#include "types.h"
-#include "debug.h"
-#include "ntfs.h"
-
-/*
- * IMPORTANT
- * =========
- *
- * All these routines assume that the Unicode characters are in little endian
- * encoding inside the strings!!!
- */
-
-/*
- * This is used by the name collation functions to quickly determine what
- * characters are (in)valid.
- */
-static const u8 legal_ansi_char_array[0x40] = {
-	0x00, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
-	0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
-
-	0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
-	0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
-
-	0x17, 0x07, 0x18, 0x17, 0x17, 0x17, 0x17, 0x17,
-	0x17, 0x17, 0x18, 0x16, 0x16, 0x17, 0x07, 0x00,
-
-	0x17, 0x17, 0x17, 0x17, 0x17, 0x17, 0x17, 0x17,
-	0x17, 0x17, 0x04, 0x16, 0x18, 0x16, 0x18, 0x18,
-};
-
-/**
- * ntfs_are_names_equal - compare two Unicode names for equality
- * @s1:			name to compare to @s2
- * @s1_len:		length in Unicode characters of @s1
- * @s2:			name to compare to @s1
- * @s2_len:		length in Unicode characters of @s2
- * @ic:			ignore case bool
- * @upcase:		upcase table (only if @ic == IGNORE_CASE)
- * @upcase_size:	length in Unicode characters of @upcase (if present)
- *
- * Compare the names @s1 and @s2 and return 'true' (1) if the names are
- * identical, or 'false' (0) if they are not identical. If @ic is IGNORE_CASE,
- * the @upcase table is used to performa a case insensitive comparison.
- */
-bool ntfs_are_names_equal(const ntfschar *s1, size_t s1_len,
-		const ntfschar *s2, size_t s2_len, const IGNORE_CASE_BOOL ic,
-		const ntfschar *upcase, const u32 upcase_size)
-{
-	if (s1_len != s2_len)
-		return false;
-	if (ic == CASE_SENSITIVE)
-		return !ntfs_ucsncmp(s1, s2, s1_len);
-	return !ntfs_ucsncasecmp(s1, s2, s1_len, upcase, upcase_size);
-}
-
-/**
- * ntfs_collate_names - collate two Unicode names
- * @name1:	first Unicode name to compare
- * @name2:	second Unicode name to compare
- * @err_val:	if @name1 contains an invalid character return this value
- * @ic:		either CASE_SENSITIVE or IGNORE_CASE
- * @upcase:	upcase table (ignored if @ic is CASE_SENSITIVE)
- * @upcase_len:	upcase table size (ignored if @ic is CASE_SENSITIVE)
- *
- * ntfs_collate_names collates two Unicode names and returns:
- *
- *  -1 if the first name collates before the second one,
- *   0 if the names match,
- *   1 if the second name collates before the first one, or
- * @err_val if an invalid character is found in @name1 during the comparison.
- *
- * The following characters are considered invalid: '"', '*', '<', '>' and '?'.
- */
-int ntfs_collate_names(const ntfschar *name1, const u32 name1_len,
-		const ntfschar *name2, const u32 name2_len,
-		const int err_val, const IGNORE_CASE_BOOL ic,
-		const ntfschar *upcase, const u32 upcase_len)
-{
-	u32 cnt, min_len;
-	u16 c1, c2;
-
-	min_len = name1_len;
-	if (name1_len > name2_len)
-		min_len = name2_len;
-	for (cnt = 0; cnt < min_len; ++cnt) {
-		c1 = le16_to_cpu(*name1++);
-		c2 = le16_to_cpu(*name2++);
-		if (ic) {
-			if (c1 < upcase_len)
-				c1 = le16_to_cpu(upcase[c1]);
-			if (c2 < upcase_len)
-				c2 = le16_to_cpu(upcase[c2]);
-		}
-		if (c1 < 64 && legal_ansi_char_array[c1] & 8)
-			return err_val;
-		if (c1 < c2)
-			return -1;
-		if (c1 > c2)
-			return 1;
-	}
-	if (name1_len < name2_len)
-		return -1;
-	if (name1_len == name2_len)
-		return 0;
-	/* name1_len > name2_len */
-	c1 = le16_to_cpu(*name1);
-	if (c1 < 64 && legal_ansi_char_array[c1] & 8)
-		return err_val;
-	return 1;
-}
-
-/**
- * ntfs_ucsncmp - compare two little endian Unicode strings
- * @s1:		first string
- * @s2:		second string
- * @n:		maximum unicode characters to compare
- *
- * Compare the first @n characters of the Unicode strings @s1 and @s2,
- * The strings in little endian format and appropriate le16_to_cpu()
- * conversion is performed on non-little endian machines.
- *
- * The function returns an integer less than, equal to, or greater than zero
- * if @s1 (or the first @n Unicode characters thereof) is found, respectively,
- * to be less than, to match, or be greater than @s2.
- */
-int ntfs_ucsncmp(const ntfschar *s1, const ntfschar *s2, size_t n)
-{
-	u16 c1, c2;
-	size_t i;
-
-	for (i = 0; i < n; ++i) {
-		c1 = le16_to_cpu(s1[i]);
-		c2 = le16_to_cpu(s2[i]);
-		if (c1 < c2)
-			return -1;
-		if (c1 > c2)
-			return 1;
-		if (!c1)
-			break;
-	}
-	return 0;
-}
-
-/**
- * ntfs_ucsncasecmp - compare two little endian Unicode strings, ignoring case
- * @s1:			first string
- * @s2:			second string
- * @n:			maximum unicode characters to compare
- * @upcase:		upcase table
- * @upcase_size:	upcase table size in Unicode characters
- *
- * Compare the first @n characters of the Unicode strings @s1 and @s2,
- * ignoring case. The strings in little endian format and appropriate
- * le16_to_cpu() conversion is performed on non-little endian machines.
- *
- * Each character is uppercased using the @upcase table before the comparison.
- *
- * The function returns an integer less than, equal to, or greater than zero
- * if @s1 (or the first @n Unicode characters thereof) is found, respectively,
- * to be less than, to match, or be greater than @s2.
- */
-int ntfs_ucsncasecmp(const ntfschar *s1, const ntfschar *s2, size_t n,
-		const ntfschar *upcase, const u32 upcase_size)
-{
-	size_t i;
-	u16 c1, c2;
-
-	for (i = 0; i < n; ++i) {
-		if ((c1 = le16_to_cpu(s1[i])) < upcase_size)
-			c1 = le16_to_cpu(upcase[c1]);
-		if ((c2 = le16_to_cpu(s2[i])) < upcase_size)
-			c2 = le16_to_cpu(upcase[c2]);
-		if (c1 < c2)
-			return -1;
-		if (c1 > c2)
-			return 1;
-		if (!c1)
-			break;
-	}
-	return 0;
-}
-
-void ntfs_upcase_name(ntfschar *name, u32 name_len, const ntfschar *upcase,
-		const u32 upcase_len)
-{
-	u32 i;
-	u16 u;
-
-	for (i = 0; i < name_len; i++)
-		if ((u = le16_to_cpu(name[i])) < upcase_len)
-			name[i] = upcase[u];
-}
-
-void ntfs_file_upcase_value(FILE_NAME_ATTR *file_name_attr,
-		const ntfschar *upcase, const u32 upcase_len)
-{
-	ntfs_upcase_name((ntfschar*)&file_name_attr->file_name,
-			file_name_attr->file_name_length, upcase, upcase_len);
-}
-
-int ntfs_file_compare_values(FILE_NAME_ATTR *file_name_attr1,
-		FILE_NAME_ATTR *file_name_attr2,
-		const int err_val, const IGNORE_CASE_BOOL ic,
-		const ntfschar *upcase, const u32 upcase_len)
-{
-	return ntfs_collate_names((ntfschar*)&file_name_attr1->file_name,
-			file_name_attr1->file_name_length,
-			(ntfschar*)&file_name_attr2->file_name,
-			file_name_attr2->file_name_length,
-			err_val, ic, upcase, upcase_len);
-}
-
-/**
- * ntfs_nlstoucs - convert NLS string to little endian Unicode string
- * @vol:	ntfs volume which we are working with
- * @ins:	input NLS string buffer
- * @ins_len:	length of input string in bytes
- * @outs:	on return contains the allocated output Unicode string buffer
- *
- * Convert the input string @ins, which is in whatever format the loaded NLS
- * map dictates, into a little endian, 2-byte Unicode string.
- *
- * This function allocates the string and the caller is responsible for
- * calling kmem_cache_free(ntfs_name_cache, *@outs); when finished with it.
- *
- * On success the function returns the number of Unicode characters written to
- * the output string *@outs (>= 0), not counting the terminating Unicode NULL
- * character. *@outs is set to the allocated output string buffer.
- *
- * On error, a negative number corresponding to the error code is returned. In
- * that case the output string is not allocated. Both *@outs and *@outs_len
- * are then undefined.
- *
- * This might look a bit odd due to fast path optimization...
- */
-int ntfs_nlstoucs(const ntfs_volume *vol, const char *ins,
-		const int ins_len, ntfschar **outs)
-{
-	struct nls_table *nls = vol->nls_map;
-	ntfschar *ucs;
-	wchar_t wc;
-	int i, o, wc_len;
-
-	/* We do not trust outside sources. */
-	if (likely(ins)) {
-		ucs = kmem_cache_alloc(ntfs_name_cache, GFP_NOFS);
-		if (likely(ucs)) {
-			for (i = o = 0; i < ins_len; i += wc_len) {
-				wc_len = nls->char2uni(ins + i, ins_len - i,
-						&wc);
-				if (likely(wc_len >= 0 &&
-						o < NTFS_MAX_NAME_LEN)) {
-					if (likely(wc)) {
-						ucs[o++] = cpu_to_le16(wc);
-						continue;
-					} /* else if (!wc) */
-					break;
-				} /* else if (wc_len < 0 ||
-						o >= NTFS_MAX_NAME_LEN) */
-				goto name_err;
-			}
-			ucs[o] = 0;
-			*outs = ucs;
-			return o;
-		} /* else if (!ucs) */
-		ntfs_error(vol->sb, "Failed to allocate buffer for converted "
-				"name from ntfs_name_cache.");
-		return -ENOMEM;
-	} /* else if (!ins) */
-	ntfs_error(vol->sb, "Received NULL pointer.");
-	return -EINVAL;
-name_err:
-	kmem_cache_free(ntfs_name_cache, ucs);
-	if (wc_len < 0) {
-		ntfs_error(vol->sb, "Name using character set %s contains "
-				"characters that cannot be converted to "
-				"Unicode.", nls->charset);
-		i = -EILSEQ;
-	} else /* if (o >= NTFS_MAX_NAME_LEN) */ {
-		ntfs_error(vol->sb, "Name is too long (maximum length for a "
-				"name on NTFS is %d Unicode characters.",
-				NTFS_MAX_NAME_LEN);
-		i = -ENAMETOOLONG;
-	}
-	return i;
-}
-
-/**
- * ntfs_ucstonls - convert little endian Unicode string to NLS string
- * @vol:	ntfs volume which we are working with
- * @ins:	input Unicode string buffer
- * @ins_len:	length of input string in Unicode characters
- * @outs:	on return contains the (allocated) output NLS string buffer
- * @outs_len:	length of output string buffer in bytes
- *
- * Convert the input little endian, 2-byte Unicode string @ins, of length
- * @ins_len into the string format dictated by the loaded NLS.
- *
- * If *@outs is NULL, this function allocates the string and the caller is
- * responsible for calling kfree(*@outs); when finished with it. In this case
- * @outs_len is ignored and can be 0.
- *
- * On success the function returns the number of bytes written to the output
- * string *@outs (>= 0), not counting the terminating NULL byte. If the output
- * string buffer was allocated, *@outs is set to it.
- *
- * On error, a negative number corresponding to the error code is returned. In
- * that case the output string is not allocated. The contents of *@outs are
- * then undefined.
- *
- * This might look a bit odd due to fast path optimization...
- */
-int ntfs_ucstonls(const ntfs_volume *vol, const ntfschar *ins,
-		const int ins_len, unsigned char **outs, int outs_len)
-{
-	struct nls_table *nls = vol->nls_map;
-	unsigned char *ns;
-	int i, o, ns_len, wc;
-
-	/* We don't trust outside sources. */
-	if (ins) {
-		ns = *outs;
-		ns_len = outs_len;
-		if (ns && !ns_len) {
-			wc = -ENAMETOOLONG;
-			goto conversion_err;
-		}
-		if (!ns) {
-			ns_len = ins_len * NLS_MAX_CHARSET_SIZE;
-			ns = kmalloc(ns_len + 1, GFP_NOFS);
-			if (!ns)
-				goto mem_err_out;
-		}
-		for (i = o = 0; i < ins_len; i++) {
-retry:			wc = nls->uni2char(le16_to_cpu(ins[i]), ns + o,
-					ns_len - o);
-			if (wc > 0) {
-				o += wc;
-				continue;
-			} else if (!wc)
-				break;
-			else if (wc == -ENAMETOOLONG && ns != *outs) {
-				unsigned char *tc;
-				/* Grow in multiples of 64 bytes. */
-				tc = kmalloc((ns_len + 64) &
-						~63, GFP_NOFS);
-				if (tc) {
-					memcpy(tc, ns, ns_len);
-					ns_len = ((ns_len + 64) & ~63) - 1;
-					kfree(ns);
-					ns = tc;
-					goto retry;
-				} /* No memory so goto conversion_error; */
-			} /* wc < 0, real error. */
-			goto conversion_err;
-		}
-		ns[o] = 0;
-		*outs = ns;
-		return o;
-	} /* else (!ins) */
-	ntfs_error(vol->sb, "Received NULL pointer.");
-	return -EINVAL;
-conversion_err:
-	ntfs_error(vol->sb, "Unicode name contains characters that cannot be "
-			"converted to character set %s.  You might want to "
-			"try to use the mount option nls=utf8.", nls->charset);
-	if (ns != *outs)
-		kfree(ns);
-	if (wc != -ENAMETOOLONG)
-		wc = -EILSEQ;
-	return wc;
-mem_err_out:
-	ntfs_error(vol->sb, "Failed to allocate name!");
-	return -ENOMEM;
-}
diff --git a/fs/ntfs/upcase.c b/fs/ntfs/upcase.c
deleted file mode 100644
index 4ebe84a78d..0000000000
--- a/fs/ntfs/upcase.c
+++ /dev/null
@@ -1,73 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * upcase.c - Generate the full NTFS Unicode upcase table in little endian.
- *	      Part of the Linux-NTFS project.
- *
- * Copyright (c) 2001 Richard Russon <ntfs@flatcap.org>
- * Copyright (c) 2001-2006 Anton Altaparmakov
- */
-
-#include "malloc.h"
-#include "ntfs.h"
-
-ntfschar *generate_default_upcase(void)
-{
-	static const int uc_run_table[][3] = { /* Start, End, Add */
-	{0x0061, 0x007B,  -32}, {0x0451, 0x045D, -80}, {0x1F70, 0x1F72,  74},
-	{0x00E0, 0x00F7,  -32}, {0x045E, 0x0460, -80}, {0x1F72, 0x1F76,  86},
-	{0x00F8, 0x00FF,  -32}, {0x0561, 0x0587, -48}, {0x1F76, 0x1F78, 100},
-	{0x0256, 0x0258, -205}, {0x1F00, 0x1F08,   8}, {0x1F78, 0x1F7A, 128},
-	{0x028A, 0x028C, -217}, {0x1F10, 0x1F16,   8}, {0x1F7A, 0x1F7C, 112},
-	{0x03AC, 0x03AD,  -38}, {0x1F20, 0x1F28,   8}, {0x1F7C, 0x1F7E, 126},
-	{0x03AD, 0x03B0,  -37}, {0x1F30, 0x1F38,   8}, {0x1FB0, 0x1FB2,   8},
-	{0x03B1, 0x03C2,  -32}, {0x1F40, 0x1F46,   8}, {0x1FD0, 0x1FD2,   8},
-	{0x03C2, 0x03C3,  -31}, {0x1F51, 0x1F52,   8}, {0x1FE0, 0x1FE2,   8},
-	{0x03C3, 0x03CC,  -32}, {0x1F53, 0x1F54,   8}, {0x1FE5, 0x1FE6,   7},
-	{0x03CC, 0x03CD,  -64}, {0x1F55, 0x1F56,   8}, {0x2170, 0x2180, -16},
-	{0x03CD, 0x03CF,  -63}, {0x1F57, 0x1F58,   8}, {0x24D0, 0x24EA, -26},
-	{0x0430, 0x0450,  -32}, {0x1F60, 0x1F68,   8}, {0xFF41, 0xFF5B, -32},
-	{0}
-	};
-
-	static const int uc_dup_table[][2] = { /* Start, End */
-	{0x0100, 0x012F}, {0x01A0, 0x01A6}, {0x03E2, 0x03EF}, {0x04CB, 0x04CC},
-	{0x0132, 0x0137}, {0x01B3, 0x01B7}, {0x0460, 0x0481}, {0x04D0, 0x04EB},
-	{0x0139, 0x0149}, {0x01CD, 0x01DD}, {0x0490, 0x04BF}, {0x04EE, 0x04F5},
-	{0x014A, 0x0178}, {0x01DE, 0x01EF}, {0x04BF, 0x04BF}, {0x04F8, 0x04F9},
-	{0x0179, 0x017E}, {0x01F4, 0x01F5}, {0x04C1, 0x04C4}, {0x1E00, 0x1E95},
-	{0x018B, 0x018B}, {0x01FA, 0x0218}, {0x04C7, 0x04C8}, {0x1EA0, 0x1EF9},
-	{0}
-	};
-
-	static const int uc_word_table[][2] = { /* Offset, Value */
-	{0x00FF, 0x0178}, {0x01AD, 0x01AC}, {0x01F3, 0x01F1}, {0x0269, 0x0196},
-	{0x0183, 0x0182}, {0x01B0, 0x01AF}, {0x0253, 0x0181}, {0x026F, 0x019C},
-	{0x0185, 0x0184}, {0x01B9, 0x01B8}, {0x0254, 0x0186}, {0x0272, 0x019D},
-	{0x0188, 0x0187}, {0x01BD, 0x01BC}, {0x0259, 0x018F}, {0x0275, 0x019F},
-	{0x018C, 0x018B}, {0x01C6, 0x01C4}, {0x025B, 0x0190}, {0x0283, 0x01A9},
-	{0x0192, 0x0191}, {0x01C9, 0x01C7}, {0x0260, 0x0193}, {0x0288, 0x01AE},
-	{0x0199, 0x0198}, {0x01CC, 0x01CA}, {0x0263, 0x0194}, {0x0292, 0x01B7},
-	{0x01A8, 0x01A7}, {0x01DD, 0x018E}, {0x0268, 0x0197},
-	{0}
-	};
-
-	int i, r;
-	ntfschar *uc;
-
-	uc = ntfs_malloc_nofs(default_upcase_len * sizeof(ntfschar));
-	if (!uc)
-		return uc;
-	memset(uc, 0, default_upcase_len * sizeof(ntfschar));
-	/* Generate the little endian Unicode upcase table used by ntfs. */
-	for (i = 0; i < default_upcase_len; i++)
-		uc[i] = cpu_to_le16(i);
-	for (r = 0; uc_run_table[r][0]; r++)
-		for (i = uc_run_table[r][0]; i < uc_run_table[r][1]; i++)
-			le16_add_cpu(&uc[i], uc_run_table[r][2]);
-	for (r = 0; uc_dup_table[r][0]; r++)
-		for (i = uc_dup_table[r][0]; i < uc_dup_table[r][1]; i += 2)
-			le16_add_cpu(&uc[i + 1], -1);
-	for (r = 0; uc_word_table[r][0]; r++)
-		uc[uc_word_table[r][0]] = cpu_to_le16(uc_word_table[r][1]);
-	return uc;
-}
diff --git a/fs/ntfs/usnjrnl.c b/fs/ntfs/usnjrnl.c
deleted file mode 100644
index 9097a0b4ef..0000000000
--- a/fs/ntfs/usnjrnl.c
+++ /dev/null
@@ -1,70 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * usnjrnl.h - NTFS kernel transaction log ($UsnJrnl) handling.  Part of the
- *	       Linux-NTFS project.
- *
- * Copyright (c) 2005 Anton Altaparmakov
- */
-
-#ifdef NTFS_RW
-
-#include <linux/fs.h>
-#include <linux/highmem.h>
-#include <linux/mm.h>
-
-#include "aops.h"
-#include "debug.h"
-#include "endian.h"
-#include "time.h"
-#include "types.h"
-#include "usnjrnl.h"
-#include "volume.h"
-
-/**
- * ntfs_stamp_usnjrnl - stamp the transaction log ($UsnJrnl) on an ntfs volume
- * @vol:	ntfs volume on which to stamp the transaction log
- *
- * Stamp the transaction log ($UsnJrnl) on the ntfs volume @vol and return
- * 'true' on success and 'false' on error.
- *
- * This function assumes that the transaction log has already been loaded and
- * consistency checked by a call to fs/ntfs/super.c::load_and_init_usnjrnl().
- */
-bool ntfs_stamp_usnjrnl(ntfs_volume *vol)
-{
-	ntfs_debug("Entering.");
-	if (likely(!NVolUsnJrnlStamped(vol))) {
-		sle64 stamp;
-		struct page *page;
-		USN_HEADER *uh;
-
-		page = ntfs_map_page(vol->usnjrnl_max_ino->i_mapping, 0);
-		if (IS_ERR(page)) {
-			ntfs_error(vol->sb, "Failed to read from "
-					"$UsnJrnl/$DATA/$Max attribute.");
-			return false;
-		}
-		uh = (USN_HEADER*)page_address(page);
-		stamp = get_current_ntfs_time();
-		ntfs_debug("Stamping transaction log ($UsnJrnl): old "
-				"journal_id 0x%llx, old lowest_valid_usn "
-				"0x%llx, new journal_id 0x%llx, new "
-				"lowest_valid_usn 0x%llx.",
-				(long long)sle64_to_cpu(uh->journal_id),
-				(long long)sle64_to_cpu(uh->lowest_valid_usn),
-				(long long)sle64_to_cpu(stamp),
-				i_size_read(vol->usnjrnl_j_ino));
-		uh->lowest_valid_usn =
-				cpu_to_sle64(i_size_read(vol->usnjrnl_j_ino));
-		uh->journal_id = stamp;
-		flush_dcache_page(page);
-		set_page_dirty(page);
-		ntfs_unmap_page(page);
-		/* Set the flag so we do not have to do it again on remount. */
-		NVolSetUsnJrnlStamped(vol);
-	}
-	ntfs_debug("Done.");
-	return true;
-}
-
-#endif /* NTFS_RW */
diff --git a/fs/ntfs/usnjrnl.h b/fs/ntfs/usnjrnl.h
deleted file mode 100644
index 85f531b593..0000000000
--- a/fs/ntfs/usnjrnl.h
+++ /dev/null
@@ -1,191 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * usnjrnl.h - Defines for NTFS kernel transaction log ($UsnJrnl) handling.
- *	       Part of the Linux-NTFS project.
- *
- * Copyright (c) 2005 Anton Altaparmakov
- */
-
-#ifndef _LINUX_NTFS_USNJRNL_H
-#define _LINUX_NTFS_USNJRNL_H
-
-#ifdef NTFS_RW
-
-#include "types.h"
-#include "endian.h"
-#include "layout.h"
-#include "volume.h"
-
-/*
- * Transaction log ($UsnJrnl) organization:
- *
- * The transaction log records whenever a file is modified in any way.  So for
- * example it will record that file "blah" was written to at a particular time
- * but not what was written.  If will record that a file was deleted or
- * created, that a file was truncated, etc.  See below for all the reason
- * codes used.
- *
- * The transaction log is in the $Extend directory which is in the root
- * directory of each volume.  If it is not present it means transaction
- * logging is disabled.  If it is present it means transaction logging is
- * either enabled or in the process of being disabled in which case we can
- * ignore it as it will go away as soon as Windows gets its hands on it.
- *
- * To determine whether the transaction logging is enabled or in the process
- * of being disabled, need to check the volume flags in the
- * $VOLUME_INFORMATION attribute in the $Volume system file (which is present
- * in the root directory and has a fixed mft record number, see layout.h).
- * If the flag VOLUME_DELETE_USN_UNDERWAY is set it means the transaction log
- * is in the process of being disabled and if this flag is clear it means the
- * transaction log is enabled.
- *
- * The transaction log consists of two parts; the $DATA/$Max attribute as well
- * as the $DATA/$J attribute.  $Max is a header describing the transaction
- * log whilst $J is the transaction log data itself as a sequence of variable
- * sized USN_RECORDs (see below for all the structures).
- *
- * We do not care about transaction logging at this point in time but we still
- * need to let windows know that the transaction log is out of date.  To do
- * this we need to stamp the transaction log.  This involves setting the
- * lowest_valid_usn field in the $DATA/$Max attribute to the usn to be used
- * for the next added USN_RECORD to the $DATA/$J attribute as well as
- * generating a new journal_id in $DATA/$Max.
- *
- * The journal_id is as of the current version (2.0) of the transaction log
- * simply the 64-bit timestamp of when the journal was either created or last
- * stamped.
- *
- * To determine the next usn there are two ways.  The first is to parse
- * $DATA/$J and to find the last USN_RECORD in it and to add its record_length
- * to its usn (which is the byte offset in the $DATA/$J attribute).  The
- * second is simply to take the data size of the attribute.  Since the usns
- * are simply byte offsets into $DATA/$J, this is exactly the next usn.  For
- * obvious reasons we use the second method as it is much simpler and faster.
- *
- * As an aside, note that to actually disable the transaction log, one would
- * need to set the VOLUME_DELETE_USN_UNDERWAY flag (see above), then go
- * through all the mft records on the volume and set the usn field in their
- * $STANDARD_INFORMATION attribute to zero.  Once that is done, one would need
- * to delete the transaction log file, i.e. \$Extent\$UsnJrnl, and finally,
- * one would need to clear the VOLUME_DELETE_USN_UNDERWAY flag.
- *
- * Note that if a volume is unmounted whilst the transaction log is being
- * disabled, the process will continue the next time the volume is mounted.
- * This is why we can safely mount read-write when we see a transaction log
- * in the process of being deleted.
- */
-
-/* Some $UsnJrnl related constants. */
-#define UsnJrnlMajorVer		2
-#define UsnJrnlMinorVer		0
-
-/*
- * $DATA/$Max attribute.  This is (always?) resident and has a fixed size of
- * 32 bytes.  It contains the header describing the transaction log.
- */
-typedef struct {
-/*Ofs*/
-/*   0*/sle64 maximum_size;	/* The maximum on-disk size of the $DATA/$J
-				   attribute. */
-/*   8*/sle64 allocation_delta;	/* Number of bytes by which to increase the
-				   size of the $DATA/$J attribute. */
-/*0x10*/sle64 journal_id;	/* Current id of the transaction log. */
-/*0x18*/leUSN lowest_valid_usn;	/* Lowest valid usn in $DATA/$J for the
-				   current journal_id. */
-/* sizeof() = 32 (0x20) bytes */
-} __attribute__ ((__packed__)) USN_HEADER;
-
-/*
- * Reason flags (32-bit).  Cumulative flags describing the change(s) to the
- * file since it was last opened.  I think the names speak for themselves but
- * if you disagree check out the descriptions in the Linux NTFS project NTFS
- * documentation: http://www.linux-ntfs.org/
- */
-enum {
-	USN_REASON_DATA_OVERWRITE	= cpu_to_le32(0x00000001),
-	USN_REASON_DATA_EXTEND		= cpu_to_le32(0x00000002),
-	USN_REASON_DATA_TRUNCATION	= cpu_to_le32(0x00000004),
-	USN_REASON_NAMED_DATA_OVERWRITE	= cpu_to_le32(0x00000010),
-	USN_REASON_NAMED_DATA_EXTEND	= cpu_to_le32(0x00000020),
-	USN_REASON_NAMED_DATA_TRUNCATION= cpu_to_le32(0x00000040),
-	USN_REASON_FILE_CREATE		= cpu_to_le32(0x00000100),
-	USN_REASON_FILE_DELETE		= cpu_to_le32(0x00000200),
-	USN_REASON_EA_CHANGE		= cpu_to_le32(0x00000400),
-	USN_REASON_SECURITY_CHANGE	= cpu_to_le32(0x00000800),
-	USN_REASON_RENAME_OLD_NAME	= cpu_to_le32(0x00001000),
-	USN_REASON_RENAME_NEW_NAME	= cpu_to_le32(0x00002000),
-	USN_REASON_INDEXABLE_CHANGE	= cpu_to_le32(0x00004000),
-	USN_REASON_BASIC_INFO_CHANGE	= cpu_to_le32(0x00008000),
-	USN_REASON_HARD_LINK_CHANGE	= cpu_to_le32(0x00010000),
-	USN_REASON_COMPRESSION_CHANGE	= cpu_to_le32(0x00020000),
-	USN_REASON_ENCRYPTION_CHANGE	= cpu_to_le32(0x00040000),
-	USN_REASON_OBJECT_ID_CHANGE	= cpu_to_le32(0x00080000),
-	USN_REASON_REPARSE_POINT_CHANGE	= cpu_to_le32(0x00100000),
-	USN_REASON_STREAM_CHANGE	= cpu_to_le32(0x00200000),
-	USN_REASON_CLOSE		= cpu_to_le32(0x80000000),
-};
-
-typedef le32 USN_REASON_FLAGS;
-
-/*
- * Source info flags (32-bit).  Information about the source of the change(s)
- * to the file.  For detailed descriptions of what these mean, see the Linux
- * NTFS project NTFS documentation:
- *	http://www.linux-ntfs.org/
- */
-enum {
-	USN_SOURCE_DATA_MANAGEMENT	  = cpu_to_le32(0x00000001),
-	USN_SOURCE_AUXILIARY_DATA	  = cpu_to_le32(0x00000002),
-	USN_SOURCE_REPLICATION_MANAGEMENT = cpu_to_le32(0x00000004),
-};
-
-typedef le32 USN_SOURCE_INFO_FLAGS;
-
-/*
- * $DATA/$J attribute.  This is always non-resident, is marked as sparse, and
- * is of variabled size.  It consists of a sequence of variable size
- * USN_RECORDS.  The minimum allocated_size is allocation_delta as
- * specified in $DATA/$Max.  When the maximum_size specified in $DATA/$Max is
- * exceeded by more than allocation_delta bytes, allocation_delta bytes are
- * allocated and appended to the $DATA/$J attribute and an equal number of
- * bytes at the beginning of the attribute are freed and made sparse.  Note the
- * making sparse only happens at volume checkpoints and hence the actual
- * $DATA/$J size can exceed maximum_size + allocation_delta temporarily.
- */
-typedef struct {
-/*Ofs*/
-/*   0*/le32 length;		/* Byte size of this record (8-byte
-				   aligned). */
-/*   4*/le16 major_ver;		/* Major version of the transaction log used
-				   for this record. */
-/*   6*/le16 minor_ver;		/* Minor version of the transaction log used
-				   for this record. */
-/*   8*/leMFT_REF mft_reference;/* The mft reference of the file (or
-				   directory) described by this record. */
-/*0x10*/leMFT_REF parent_directory;/* The mft reference of the parent
-				   directory of the file described by this
-				   record. */
-/*0x18*/leUSN usn;		/* The usn of this record.  Equals the offset
-				   within the $DATA/$J attribute. */
-/*0x20*/sle64 time;		/* Time when this record was created. */
-/*0x28*/USN_REASON_FLAGS reason;/* Reason flags (see above). */
-/*0x2c*/USN_SOURCE_INFO_FLAGS source_info;/* Source info flags (see above). */
-/*0x30*/le32 security_id;	/* File security_id copied from
-				   $STANDARD_INFORMATION. */
-/*0x34*/FILE_ATTR_FLAGS file_attributes;	/* File attributes copied from
-				   $STANDARD_INFORMATION or $FILE_NAME (not
-				   sure which). */
-/*0x38*/le16 file_name_size;	/* Size of the file name in bytes. */
-/*0x3a*/le16 file_name_offset;	/* Offset to the file name in bytes from the
-				   start of this record. */
-/*0x3c*/ntfschar file_name[0];	/* Use when creating only.  When reading use
-				   file_name_offset to determine the location
-				   of the name. */
-/* sizeof() = 60 (0x3c) bytes */
-} __attribute__ ((__packed__)) USN_RECORD;
-
-extern bool ntfs_stamp_usnjrnl(ntfs_volume *vol);
-
-#endif /* NTFS_RW */
-
-#endif /* _LINUX_NTFS_USNJRNL_H */
diff --git a/fs/ntfs/volume.h b/fs/ntfs/volume.h
deleted file mode 100644
index 930a9ae8a0..0000000000
--- a/fs/ntfs/volume.h
+++ /dev/null
@@ -1,164 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * volume.h - Defines for volume structures in NTFS Linux kernel driver. Part
- *	      of the Linux-NTFS project.
- *
- * Copyright (c) 2001-2006 Anton Altaparmakov
- * Copyright (c) 2002 Richard Russon
- */
-
-#ifndef _LINUX_NTFS_VOLUME_H
-#define _LINUX_NTFS_VOLUME_H
-
-#include <linux/rwsem.h>
-#include <linux/uidgid.h>
-
-#include "types.h"
-#include "layout.h"
-
-/*
- * The NTFS in memory super block structure.
- */
-typedef struct {
-	/*
-	 * FIXME: Reorder to have commonly used together element within the
-	 * same cache line, aiming at a cache line size of 32 bytes. Aim for
-	 * 64 bytes for less commonly used together elements. Put most commonly
-	 * used elements to front of structure. Obviously do this only when the
-	 * structure has stabilized... (AIA)
-	 */
-	/* Device specifics. */
-	struct super_block *sb;		/* Pointer back to the super_block. */
-	LCN nr_blocks;			/* Number of sb->s_blocksize bytes
-					   sized blocks on the device. */
-	/* Configuration provided by user at mount time. */
-	unsigned long flags;		/* Miscellaneous flags, see below. */
-	kuid_t uid;			/* uid that files will be mounted as. */
-	kgid_t gid;			/* gid that files will be mounted as. */
-	umode_t fmask;			/* The mask for file permissions. */
-	umode_t dmask;			/* The mask for directory
-					   permissions. */
-	u8 mft_zone_multiplier;		/* Initial mft zone multiplier. */
-	u8 on_errors;			/* What to do on filesystem errors. */
-	/* NTFS bootsector provided information. */
-	u16 sector_size;		/* in bytes */
-	u8 sector_size_bits;		/* log2(sector_size) */
-	u32 cluster_size;		/* in bytes */
-	u32 cluster_size_mask;		/* cluster_size - 1 */
-	u8 cluster_size_bits;		/* log2(cluster_size) */
-	u32 mft_record_size;		/* in bytes */
-	u32 mft_record_size_mask;	/* mft_record_size - 1 */
-	u8 mft_record_size_bits;	/* log2(mft_record_size) */
-	u32 index_record_size;		/* in bytes */
-	u32 index_record_size_mask;	/* index_record_size - 1 */
-	u8 index_record_size_bits;	/* log2(index_record_size) */
-	LCN nr_clusters;		/* Volume size in clusters == number of
-					   bits in lcn bitmap. */
-	LCN mft_lcn;			/* Cluster location of mft data. */
-	LCN mftmirr_lcn;		/* Cluster location of copy of mft. */
-	u64 serial_no;			/* The volume serial number. */
-	/* Mount specific NTFS information. */
-	u32 upcase_len;			/* Number of entries in upcase[]. */
-	ntfschar *upcase;		/* The upcase table. */
-
-	s32 attrdef_size;		/* Size of the attribute definition
-					   table in bytes. */
-	ATTR_DEF *attrdef;		/* Table of attribute definitions.
-					   Obtained from FILE_AttrDef. */
-
-#ifdef NTFS_RW
-	/* Variables used by the cluster and mft allocators. */
-	s64 mft_data_pos;		/* Mft record number at which to
-					   allocate the next mft record. */
-	LCN mft_zone_start;		/* First cluster of the mft zone. */
-	LCN mft_zone_end;		/* First cluster beyond the mft zone. */
-	LCN mft_zone_pos;		/* Current position in the mft zone. */
-	LCN data1_zone_pos;		/* Current position in the first data
-					   zone. */
-	LCN data2_zone_pos;		/* Current position in the second data
-					   zone. */
-#endif /* NTFS_RW */
-
-	struct inode *mft_ino;		/* The VFS inode of $MFT. */
-
-	struct inode *mftbmp_ino;	/* Attribute inode for $MFT/$BITMAP. */
-	struct rw_semaphore mftbmp_lock; /* Lock for serializing accesses to the
-					    mft record bitmap ($MFT/$BITMAP). */
-#ifdef NTFS_RW
-	struct inode *mftmirr_ino;	/* The VFS inode of $MFTMirr. */
-	int mftmirr_size;		/* Size of mft mirror in mft records. */
-
-	struct inode *logfile_ino;	/* The VFS inode of $LogFile. */
-#endif /* NTFS_RW */
-
-	struct inode *lcnbmp_ino;	/* The VFS inode of $Bitmap. */
-	struct rw_semaphore lcnbmp_lock; /* Lock for serializing accesses to the
-					    cluster bitmap ($Bitmap/$DATA). */
-
-	struct inode *vol_ino;		/* The VFS inode of $Volume. */
-	VOLUME_FLAGS vol_flags;		/* Volume flags. */
-	u8 major_ver;			/* Ntfs major version of volume. */
-	u8 minor_ver;			/* Ntfs minor version of volume. */
-
-	struct inode *root_ino;		/* The VFS inode of the root
-					   directory. */
-	struct inode *secure_ino;	/* The VFS inode of $Secure (NTFS3.0+
-					   only, otherwise NULL). */
-	struct inode *extend_ino;	/* The VFS inode of $Extend (NTFS3.0+
-					   only, otherwise NULL). */
-#ifdef NTFS_RW
-	/* $Quota stuff is NTFS3.0+ specific.  Unused/NULL otherwise. */
-	struct inode *quota_ino;	/* The VFS inode of $Quota. */
-	struct inode *quota_q_ino;	/* Attribute inode for $Quota/$Q. */
-	/* $UsnJrnl stuff is NTFS3.0+ specific.  Unused/NULL otherwise. */
-	struct inode *usnjrnl_ino;	/* The VFS inode of $UsnJrnl. */
-	struct inode *usnjrnl_max_ino;	/* Attribute inode for $UsnJrnl/$Max. */
-	struct inode *usnjrnl_j_ino;	/* Attribute inode for $UsnJrnl/$J. */
-#endif /* NTFS_RW */
-	struct nls_table *nls_map;
-} ntfs_volume;
-
-/*
- * Defined bits for the flags field in the ntfs_volume structure.
- */
-typedef enum {
-	NV_Errors,		/* 1: Volume has errors, prevent remount rw. */
-	NV_ShowSystemFiles,	/* 1: Return system files in ntfs_readdir(). */
-	NV_CaseSensitive,	/* 1: Treat file names as case sensitive and
-				      create filenames in the POSIX namespace.
-				      Otherwise be case insensitive but still
-				      create file names in POSIX namespace. */
-	NV_LogFileEmpty,	/* 1: $LogFile journal is empty. */
-	NV_QuotaOutOfDate,	/* 1: $Quota is out of date. */
-	NV_UsnJrnlStamped,	/* 1: $UsnJrnl has been stamped. */
-	NV_SparseEnabled,	/* 1: May create sparse files. */
-} ntfs_volume_flags;
-
-/*
- * Macro tricks to expand the NVolFoo(), NVolSetFoo(), and NVolClearFoo()
- * functions.
- */
-#define DEFINE_NVOL_BIT_OPS(flag)					\
-static inline int NVol##flag(ntfs_volume *vol)		\
-{							\
-	return test_bit(NV_##flag, &(vol)->flags);	\
-}							\
-static inline void NVolSet##flag(ntfs_volume *vol)	\
-{							\
-	set_bit(NV_##flag, &(vol)->flags);		\
-}							\
-static inline void NVolClear##flag(ntfs_volume *vol)	\
-{							\
-	clear_bit(NV_##flag, &(vol)->flags);		\
-}
-
-/* Emit the ntfs volume bitops functions. */
-DEFINE_NVOL_BIT_OPS(Errors)
-DEFINE_NVOL_BIT_OPS(ShowSystemFiles)
-DEFINE_NVOL_BIT_OPS(CaseSensitive)
-DEFINE_NVOL_BIT_OPS(LogFileEmpty)
-DEFINE_NVOL_BIT_OPS(QuotaOutOfDate)
-DEFINE_NVOL_BIT_OPS(UsnJrnlStamped)
-DEFINE_NVOL_BIT_OPS(SparseEnabled)
-
-#endif /* _LINUX_NTFS_VOLUME_H */
diff --git a/fs/ntfs3/Kconfig b/fs/ntfs3/Kconfig
index cdfdf51e55..7bc31d69f6 100644
--- a/fs/ntfs3/Kconfig
+++ b/fs/ntfs3/Kconfig
@@ -46,3 +46,12 @@ config NTFS3_FS_POSIX_ACL
 	  NOTE: this is linux only feature. Windows will ignore these ACLs.
 
 	  If you don't know what Access Control Lists are, say N.
+
+config NTFS_FS
+	tristate "NTFS file system support"
+	select NTFS3_FS
+	select BUFFER_HEAD
+	select NLS
+	help
+	  This config option is here only for backward compatibility. NTFS
+	  filesystem is now handled by the NTFS3 driver.
diff --git a/fs/ntfs3/attrib.c b/fs/ntfs3/attrib.c
index 7aadf50109..a810ef5475 100644
--- a/fs/ntfs3/attrib.c
+++ b/fs/ntfs3/attrib.c
@@ -231,7 +231,7 @@ int attr_make_nonresident(struct ntfs_inode *ni, struct ATTRIB *attr,
 	struct ntfs_sb_info *sbi;
 	struct ATTRIB *attr_s;
 	struct MFT_REC *rec;
-	u32 used, asize, rsize, aoff, align;
+	u32 used, asize, rsize, aoff;
 	bool is_data;
 	CLST len, alen;
 	char *next;
@@ -252,10 +252,13 @@ int attr_make_nonresident(struct ntfs_inode *ni, struct ATTRIB *attr,
 	rsize = le32_to_cpu(attr->res.data_size);
 	is_data = attr->type == ATTR_DATA && !attr->name_len;
 
-	align = sbi->cluster_size;
-	if (is_attr_compressed(attr))
-		align <<= COMPRESSION_UNIT;
-	len = (rsize + align - 1) >> sbi->cluster_bits;
+	/* len - how many clusters required to store 'rsize' bytes */
+	if (is_attr_compressed(attr)) {
+		u8 shift = sbi->cluster_bits + NTFS_LZNT_CUNIT;
+		len = ((rsize + (1u << shift) - 1) >> shift) << NTFS_LZNT_CUNIT;
+	} else {
+		len = bytes_to_cluster(sbi, rsize);
+	}
 
 	run_init(run);
 
@@ -670,7 +673,8 @@ pack_runs:
 			goto undo_2;
 		}
 
-		if (!is_mft)
+		/* keep runs for $MFT::$ATTR_DATA and $MFT::$ATTR_BITMAP. */
+		if (ni->mi.rno != MFT_REC_MFT)
 			run_truncate_head(run, evcn + 1);
 
 		svcn = le64_to_cpu(attr->nres.svcn);
@@ -972,6 +976,19 @@ int attr_data_get_block(struct ntfs_inode *ni, CLST vcn, CLST clen, CLST *lcn,
 	if (err)
 		goto out;
 
+	/* Check for compressed frame. */
+	err = attr_is_frame_compressed(ni, attr, vcn >> NTFS_LZNT_CUNIT, &hint);
+	if (err)
+		goto out;
+
+	if (hint) {
+		/* if frame is compressed - don't touch it. */
+		*lcn = COMPRESSED_LCN;
+		*len = hint;
+		err = -EOPNOTSUPP;
+		goto out;
+	}
+
 	if (!*len) {
 		if (run_lookup_entry(run, vcn, lcn, len, NULL)) {
 			if (*lcn != SPARSE_LCN || !new)
@@ -1722,6 +1739,7 @@ repack:
 
 	attr_b->nres.total_size = cpu_to_le64(total_size);
 	inode_set_bytes(&ni->vfs_inode, total_size);
+	ni->ni_flags |= NI_FLAG_UPDATE_PARENT;
 
 	mi_b->dirty = true;
 	mark_inode_dirty(&ni->vfs_inode);
@@ -2558,3 +2576,35 @@ undo_insert_range:
 
 	goto out;
 }
+
+/*
+ * attr_force_nonresident
+ *
+ * Convert default data attribute into non resident form.
+ */
+int attr_force_nonresident(struct ntfs_inode *ni)
+{
+	int err;
+	struct ATTRIB *attr;
+	struct ATTR_LIST_ENTRY *le = NULL;
+	struct mft_inode *mi;
+
+	attr = ni_find_attr(ni, NULL, &le, ATTR_DATA, NULL, 0, NULL, &mi);
+	if (!attr) {
+		ntfs_bad_inode(&ni->vfs_inode, "no data attribute");
+		return -ENOENT;
+	}
+
+	if (attr->non_res) {
+		/* Already non resident. */
+		return 0;
+	}
+
+	down_write(&ni->file.run_lock);
+	err = attr_make_nonresident(ni, attr, le, mi,
+				    le32_to_cpu(attr->res.data_size),
+				    &ni->file.run, &attr, NULL);
+	up_write(&ni->file.run_lock);
+
+	return err;
+}
diff --git a/fs/ntfs3/bitmap.c b/fs/ntfs3/bitmap.c
index 845f9b22de..cf4fe21a50 100644
--- a/fs/ntfs3/bitmap.c
+++ b/fs/ntfs3/bitmap.c
@@ -654,7 +654,7 @@ int wnd_init(struct wnd_bitmap *wnd, struct super_block *sb, size_t nbits)
 	wnd->total_zeroes = nbits;
 	wnd->extent_max = MINUS_ONE_T;
 	wnd->zone_bit = wnd->zone_end = 0;
-	wnd->nwnd = bytes_to_block(sb, bitmap_size(nbits));
+	wnd->nwnd = bytes_to_block(sb, ntfs3_bitmap_size(nbits));
 	wnd->bits_last = nbits & (wbits - 1);
 	if (!wnd->bits_last)
 		wnd->bits_last = wbits;
@@ -1347,7 +1347,7 @@ int wnd_extend(struct wnd_bitmap *wnd, size_t new_bits)
 		return -EINVAL;
 
 	/* Align to 8 byte boundary. */
-	new_wnd = bytes_to_block(sb, bitmap_size(new_bits));
+	new_wnd = bytes_to_block(sb, ntfs3_bitmap_size(new_bits));
 	new_last = new_bits & (wbits - 1);
 	if (!new_last)
 		new_last = wbits;
@@ -1382,7 +1382,7 @@ int wnd_extend(struct wnd_bitmap *wnd, size_t new_bits)
 
 		err = ntfs_vbo_to_lbo(sbi, &wnd->run, vbo, &lbo, &bytes);
 		if (err)
-			break;
+			return err;
 
 		bh = ntfs_bread(sb, lbo >> sb->s_blocksize_bits);
 		if (!bh)
diff --git a/fs/ntfs3/dir.c b/fs/ntfs3/dir.c
index 45e556fd7c..858efe255f 100644
--- a/fs/ntfs3/dir.c
+++ b/fs/ntfs3/dir.c
@@ -326,7 +326,8 @@ static inline int ntfs_filldir(struct ntfs_sb_info *sbi, struct ntfs_inode *ni,
 	 * It does additional locks/reads just to get the type of name.
 	 * Should we use additional mount option to enable branch below?
 	 */
-	if ((fname->dup.fa & FILE_ATTRIBUTE_REPARSE_POINT) &&
+	if (((fname->dup.fa & FILE_ATTRIBUTE_REPARSE_POINT) ||
+	     fname->dup.ea_size) &&
 	    ino != ni->mi.rno) {
 		struct inode *inode = ntfs_iget5(sbi->sb, &e->ref, NULL);
 		if (!IS_ERR_OR_NULL(inode)) {
@@ -617,4 +618,11 @@ const struct file_operations ntfs_dir_operations = {
 	.compat_ioctl   = ntfs_compat_ioctl,
 #endif
 };
+
+const struct file_operations ntfs_legacy_dir_operations = {
+	.llseek		= generic_file_llseek,
+	.read		= generic_read_dir,
+	.iterate_shared	= ntfs_readdir,
+	.open		= ntfs_file_open,
+};
 // clang-format on
diff --git a/fs/ntfs3/file.c b/fs/ntfs3/file.c
index 5418662c80..9ae202901f 100644
--- a/fs/ntfs3/file.c
+++ b/fs/ntfs3/file.c
@@ -299,10 +299,7 @@ static int ntfs_file_mmap(struct file *file, struct vm_area_struct *vma)
 		}
 
 		if (ni->i_valid < to) {
-			if (!inode_trylock(inode)) {
-				err = -EAGAIN;
-				goto out;
-			}
+			inode_lock(inode);
 			err = ntfs_extend_initialized_size(file, ni,
 							   ni->i_valid, to);
 			inode_unlock(inode);
@@ -578,6 +575,15 @@ static long ntfs_fallocate(struct file *file, int mode, loff_t vbo, loff_t len)
 		/* Check new size. */
 		u8 cluster_bits = sbi->cluster_bits;
 
+		/* Be sure file is non resident. */
+		if (is_resident(ni)) {
+			ni_lock(ni);
+			err = attr_force_nonresident(ni);
+			ni_unlock(ni);
+			if (err)
+				goto out;
+		}
+
 		/* generic/213: expected -ENOSPC instead of -EFBIG. */
 		if (!is_supported_holes) {
 			loff_t to_alloc = new_size - inode_get_bytes(inode);
@@ -1236,4 +1242,12 @@ const struct file_operations ntfs_file_operations = {
 	.fallocate	= ntfs_fallocate,
 	.release	= ntfs_file_release,
 };
+
+const struct file_operations ntfs_legacy_file_operations = {
+	.llseek		= generic_file_llseek,
+	.read_iter	= ntfs_file_read_iter,
+	.splice_read	= ntfs_file_splice_read,
+	.open		= ntfs_file_open,
+	.release	= ntfs_file_release,
+};
 // clang-format on
diff --git a/fs/ntfs3/frecord.c b/fs/ntfs3/frecord.c
index 7f27382e0c..ded451a84b 100644
--- a/fs/ntfs3/frecord.c
+++ b/fs/ntfs3/frecord.c
@@ -1501,7 +1501,7 @@ int ni_insert_nonresident(struct ntfs_inode *ni, enum ATTR_TYPE type,
 
 	if (is_ext) {
 		if (flags & ATTR_FLAG_COMPRESSED)
-			attr->nres.c_unit = COMPRESSION_UNIT;
+			attr->nres.c_unit = NTFS_LZNT_CUNIT;
 		attr->nres.total_size = attr->nres.alloc_size;
 	}
 
@@ -1897,6 +1897,47 @@ enum REPARSE_SIGN ni_parse_reparse(struct ntfs_inode *ni, struct ATTRIB *attr,
 }
 
 /*
+ * fiemap_fill_next_extent_k - a copy of fiemap_fill_next_extent
+ * but it accepts kernel address for fi_extents_start
+ */
+static int fiemap_fill_next_extent_k(struct fiemap_extent_info *fieinfo,
+				     u64 logical, u64 phys, u64 len, u32 flags)
+{
+	struct fiemap_extent extent;
+	struct fiemap_extent __user *dest = fieinfo->fi_extents_start;
+
+	/* only count the extents */
+	if (fieinfo->fi_extents_max == 0) {
+		fieinfo->fi_extents_mapped++;
+		return (flags & FIEMAP_EXTENT_LAST) ? 1 : 0;
+	}
+
+	if (fieinfo->fi_extents_mapped >= fieinfo->fi_extents_max)
+		return 1;
+
+	if (flags & FIEMAP_EXTENT_DELALLOC)
+		flags |= FIEMAP_EXTENT_UNKNOWN;
+	if (flags & FIEMAP_EXTENT_DATA_ENCRYPTED)
+		flags |= FIEMAP_EXTENT_ENCODED;
+	if (flags & (FIEMAP_EXTENT_DATA_TAIL | FIEMAP_EXTENT_DATA_INLINE))
+		flags |= FIEMAP_EXTENT_NOT_ALIGNED;
+
+	memset(&extent, 0, sizeof(extent));
+	extent.fe_logical = logical;
+	extent.fe_physical = phys;
+	extent.fe_length = len;
+	extent.fe_flags = flags;
+
+	dest += fieinfo->fi_extents_mapped;
+	memcpy(dest, &extent, sizeof(extent));
+
+	fieinfo->fi_extents_mapped++;
+	if (fieinfo->fi_extents_mapped == fieinfo->fi_extents_max)
+		return 1;
+	return (flags & FIEMAP_EXTENT_LAST) ? 1 : 0;
+}
+
+/*
  * ni_fiemap - Helper for file_fiemap().
  *
  * Assumed ni_lock.
@@ -1906,6 +1947,8 @@ int ni_fiemap(struct ntfs_inode *ni, struct fiemap_extent_info *fieinfo,
 	      __u64 vbo, __u64 len)
 {
 	int err = 0;
+	struct fiemap_extent __user *fe_u = fieinfo->fi_extents_start;
+	struct fiemap_extent *fe_k = NULL;
 	struct ntfs_sb_info *sbi = ni->mi.sbi;
 	u8 cluster_bits = sbi->cluster_bits;
 	struct runs_tree *run;
@@ -1953,6 +1996,18 @@ int ni_fiemap(struct ntfs_inode *ni, struct fiemap_extent_info *fieinfo,
 		goto out;
 	}
 
+	/*
+	 * To avoid lock problems replace pointer to user memory by pointer to kernel memory.
+	 */
+	fe_k = kmalloc_array(fieinfo->fi_extents_max,
+			     sizeof(struct fiemap_extent),
+			     GFP_NOFS | __GFP_ZERO);
+	if (!fe_k) {
+		err = -ENOMEM;
+		goto out;
+	}
+	fieinfo->fi_extents_start = fe_k;
+
 	end = vbo + len;
 	alloc_size = le64_to_cpu(attr->nres.alloc_size);
 	if (end > alloc_size)
@@ -2041,8 +2096,9 @@ int ni_fiemap(struct ntfs_inode *ni, struct fiemap_extent_info *fieinfo,
 			if (vbo + dlen >= end)
 				flags |= FIEMAP_EXTENT_LAST;
 
-			err = fiemap_fill_next_extent(fieinfo, vbo, lbo, dlen,
-						      flags);
+			err = fiemap_fill_next_extent_k(fieinfo, vbo, lbo, dlen,
+							flags);
+
 			if (err < 0)
 				break;
 			if (err == 1) {
@@ -2062,7 +2118,8 @@ int ni_fiemap(struct ntfs_inode *ni, struct fiemap_extent_info *fieinfo,
 		if (vbo + bytes >= end)
 			flags |= FIEMAP_EXTENT_LAST;
 
-		err = fiemap_fill_next_extent(fieinfo, vbo, lbo, bytes, flags);
+		err = fiemap_fill_next_extent_k(fieinfo, vbo, lbo, bytes,
+						flags);
 		if (err < 0)
 			break;
 		if (err == 1) {
@@ -2075,7 +2132,19 @@ int ni_fiemap(struct ntfs_inode *ni, struct fiemap_extent_info *fieinfo,
 
 	up_read(run_lock);
 
+	/*
+	 * Copy to user memory out of lock
+	 */
+	if (copy_to_user(fe_u, fe_k,
+			 fieinfo->fi_extents_max *
+				 sizeof(struct fiemap_extent))) {
+		err = -EFAULT;
+	}
+
 out:
+	/* Restore original pointer. */
+	fieinfo->fi_extents_start = fe_u;
+	kfree(fe_k);
 	return err;
 }
 
@@ -2636,7 +2705,7 @@ int ni_read_frame(struct ntfs_inode *ni, u64 frame_vbo, struct page **pages,
 		goto out1;
 	}
 
-	pages_disk = kzalloc(npages_disk * sizeof(struct page *), GFP_NOFS);
+	pages_disk = kcalloc(npages_disk, sizeof(*pages_disk), GFP_NOFS);
 	if (!pages_disk) {
 		err = -ENOMEM;
 		goto out2;
diff --git a/fs/ntfs3/fslog.c b/fs/ntfs3/fslog.c
index 855519713b..fcb3e49911 100644
--- a/fs/ntfs3/fslog.c
+++ b/fs/ntfs3/fslog.c
@@ -517,7 +517,7 @@ static inline bool is_rst_area_valid(const struct RESTART_HDR *rhdr)
 		seq_bits -= 1;
 	}
 
-	if (seq_bits != ra->seq_num_bits)
+	if (seq_bits != le32_to_cpu(ra->seq_num_bits))
 		return false;
 
 	/* The log page data offset and record header length must be quad-aligned. */
@@ -724,7 +724,8 @@ static bool check_rstbl(const struct RESTART_TABLE *rt, size_t bytes)
 
 	if (!rsize || rsize > bytes ||
 	    rsize + sizeof(struct RESTART_TABLE) > bytes || bytes < ts ||
-	    le16_to_cpu(rt->total) > ne || ff > ts || lf > ts ||
+	    le16_to_cpu(rt->total) > ne ||
+			ff > ts - sizeof(__le32) || lf > ts - sizeof(__le32) ||
 	    (ff && ff < sizeof(struct RESTART_TABLE)) ||
 	    (lf && lf < sizeof(struct RESTART_TABLE))) {
 		return false;
@@ -754,6 +755,9 @@ static bool check_rstbl(const struct RESTART_TABLE *rt, size_t bytes)
 			return false;
 
 		off = le32_to_cpu(*(__le32 *)Add2Ptr(rt, off));
+
+		if (off > ts - sizeof(__le32))
+			return false;
 	}
 
 	return true;
@@ -1184,7 +1188,8 @@ out:
 static int log_read_rst(struct ntfs_log *log, bool first,
 			struct restart_info *info)
 {
-	u32 skip, vbo;
+	u32 skip;
+	u64 vbo;
 	struct RESTART_HDR *r_page = NULL;
 
 	/* Determine which restart area we are looking for. */
@@ -2991,7 +2996,7 @@ static struct ATTRIB *attr_create_nonres_log(struct ntfs_sb_info *sbi,
 	if (is_ext) {
 		attr->name_off = SIZEOF_NONRESIDENT_EX_LE;
 		if (is_attr_compressed(attr))
-			attr->nres.c_unit = COMPRESSION_UNIT;
+			attr->nres.c_unit = NTFS_LZNT_CUNIT;
 
 		attr->nres.run_off =
 			cpu_to_le16(SIZEOF_NONRESIDENT_EX + name_size);
@@ -3721,6 +3726,8 @@ int log_replay(struct ntfs_inode *ni, bool *initialized)
 
 	u64 rec_lsn, checkpt_lsn = 0, rlsn = 0;
 	struct ATTR_NAME_ENTRY *attr_names = NULL;
+	u32 attr_names_bytes = 0;
+	u32 oatbl_bytes = 0;
 	struct RESTART_TABLE *dptbl = NULL;
 	struct RESTART_TABLE *trtbl = NULL;
 	const struct RESTART_TABLE *rt;
@@ -3735,6 +3742,7 @@ int log_replay(struct ntfs_inode *ni, bool *initialized)
 	struct NTFS_RESTART *rst = NULL;
 	struct lcb *lcb = NULL;
 	struct OPEN_ATTR_ENRTY *oe;
+	struct ATTR_NAME_ENTRY *ane;
 	struct TRANSACTION_ENTRY *tr;
 	struct DIR_PAGE_ENTRY *dp;
 	u32 i, bytes_per_attr_entry;
@@ -3914,6 +3922,9 @@ check_restart_area:
 		goto out;
 	}
 
+	log->page_mask = log->page_size - 1;
+	log->page_bits = blksize_bits(log->page_size);
+
 	/* If the file size has shrunk then we won't mount it. */
 	if (log->l_size < le64_to_cpu(ra2->l_size)) {
 		err = -EINVAL;
@@ -4313,17 +4324,40 @@ check_attr_table:
 	lcb = NULL;
 
 check_attribute_names2:
-	if (rst->attr_names_len && oatbl) {
-		struct ATTR_NAME_ENTRY *ane = attr_names;
-		while (ane->off) {
+	if (attr_names && oatbl) {
+		off = 0;
+		for (;;) {
+			/* Check we can use attribute name entry 'ane'. */
+			static_assert(sizeof(*ane) == 4);
+			if (off + sizeof(*ane) > attr_names_bytes) {
+				/* just ignore the rest. */
+				break;
+			}
+
+			ane = Add2Ptr(attr_names, off);
+			t16 = le16_to_cpu(ane->off);
+			if (!t16) {
+				/* this is the only valid exit. */
+				break;
+			}
+
+			/* Check we can use open attribute entry 'oe'. */
+			if (t16 + sizeof(*oe) > oatbl_bytes) {
+				/* just ignore the rest. */
+				break;
+			}
+
 			/* TODO: Clear table on exit! */
-			oe = Add2Ptr(oatbl, le16_to_cpu(ane->off));
+			oe = Add2Ptr(oatbl, t16);
 			t16 = le16_to_cpu(ane->name_bytes);
+			off += t16 + sizeof(*ane);
+			if (off > attr_names_bytes) {
+				/* just ignore the rest. */
+				break;
+			}
 			oe->name_len = t16 / sizeof(short);
 			oe->ptr = ane->name;
 			oe->is_attr_name = 2;
-			ane = Add2Ptr(ane,
-				      sizeof(struct ATTR_NAME_ENTRY) + t16);
 		}
 	}
 
diff --git a/fs/ntfs3/fsntfs.c b/fs/ntfs3/fsntfs.c
index ae2ef5c118..626d3f2c7e 100644
--- a/fs/ntfs3/fsntfs.c
+++ b/fs/ntfs3/fsntfs.c
@@ -522,7 +522,7 @@ static int ntfs_extend_mft(struct ntfs_sb_info *sbi)
 	ni->mi.dirty = true;
 
 	/* Step 2: Resize $MFT::BITMAP. */
-	new_bitmap_bytes = bitmap_size(new_mft_total);
+	new_bitmap_bytes = ntfs3_bitmap_size(new_mft_total);
 
 	err = attr_set_size(ni, ATTR_BITMAP, NULL, 0, &sbi->mft.bitmap.run,
 			    new_bitmap_bytes, &new_bitmap_bytes, true, NULL);
diff --git a/fs/ntfs3/index.c b/fs/ntfs3/index.c
index 14284f0ed4..9089c58a00 100644
--- a/fs/ntfs3/index.c
+++ b/fs/ntfs3/index.c
@@ -978,7 +978,7 @@ static struct indx_node *indx_new(struct ntfs_index *indx,
 		hdr->used =
 			cpu_to_le32(eo + sizeof(struct NTFS_DE) + sizeof(u64));
 		de_set_vbn_le(e, *sub_vbn);
-		hdr->flags = 1;
+		hdr->flags = NTFS_INDEX_HDR_HAS_SUBNODES;
 	} else {
 		e->size = cpu_to_le16(sizeof(struct NTFS_DE));
 		hdr->used = cpu_to_le32(eo + sizeof(struct NTFS_DE));
@@ -1456,8 +1456,8 @@ static int indx_create_allocate(struct ntfs_index *indx, struct ntfs_inode *ni,
 
 	alloc->nres.valid_size = alloc->nres.data_size = cpu_to_le64(data_size);
 
-	err = ni_insert_resident(ni, bitmap_size(1), ATTR_BITMAP, in->name,
-				 in->name_len, &bitmap, NULL, NULL);
+	err = ni_insert_resident(ni, ntfs3_bitmap_size(1), ATTR_BITMAP,
+				 in->name, in->name_len, &bitmap, NULL, NULL);
 	if (err)
 		goto out2;
 
@@ -1518,8 +1518,9 @@ static int indx_add_allocate(struct ntfs_index *indx, struct ntfs_inode *ni,
 	if (bmp) {
 		/* Increase bitmap. */
 		err = attr_set_size(ni, ATTR_BITMAP, in->name, in->name_len,
-				    &indx->bitmap_run, bitmap_size(bit + 1),
-				    NULL, true, NULL);
+				    &indx->bitmap_run,
+				    ntfs3_bitmap_size(bit + 1), NULL, true,
+				    NULL);
 		if (err)
 			goto out1;
 	}
@@ -1682,7 +1683,7 @@ static int indx_insert_into_root(struct ntfs_index *indx, struct ntfs_inode *ni,
 	e->size = cpu_to_le16(sizeof(struct NTFS_DE) + sizeof(u64));
 	e->flags = NTFS_IE_HAS_SUBNODES | NTFS_IE_LAST;
 
-	hdr->flags = 1;
+	hdr->flags = NTFS_INDEX_HDR_HAS_SUBNODES;
 	hdr->used = hdr->total =
 		cpu_to_le32(new_root_size - offsetof(struct INDEX_ROOT, ihdr));
 
@@ -2098,7 +2099,7 @@ static int indx_shrink(struct ntfs_index *indx, struct ntfs_inode *ni,
 	if (in->name == I30_NAME)
 		i_size_write(&ni->vfs_inode, new_data);
 
-	bpb = bitmap_size(bit);
+	bpb = ntfs3_bitmap_size(bit);
 	if (bpb * 8 == nbits)
 		return 0;
 
diff --git a/fs/ntfs3/inode.c b/fs/ntfs3/inode.c
index 05f169018c..9559d72f86 100644
--- a/fs/ntfs3/inode.c
+++ b/fs/ntfs3/inode.c
@@ -441,7 +441,10 @@ end_enum:
 		 * Usually a hard links to directories are disabled.
 		 */
 		inode->i_op = &ntfs_dir_inode_operations;
-		inode->i_fop = &ntfs_dir_operations;
+		if (is_legacy_ntfs(inode->i_sb))
+			inode->i_fop = &ntfs_legacy_dir_operations;
+		else
+			inode->i_fop = &ntfs_dir_operations;
 		ni->i_valid = 0;
 	} else if (S_ISLNK(mode)) {
 		ni->std_fa &= ~FILE_ATTRIBUTE_DIRECTORY;
@@ -451,7 +454,10 @@ end_enum:
 	} else if (S_ISREG(mode)) {
 		ni->std_fa &= ~FILE_ATTRIBUTE_DIRECTORY;
 		inode->i_op = &ntfs_file_inode_operations;
-		inode->i_fop = &ntfs_file_operations;
+		if (is_legacy_ntfs(inode->i_sb))
+			inode->i_fop = &ntfs_legacy_file_operations;
+		else
+			inode->i_fop = &ntfs_file_operations;
 		inode->i_mapping->a_ops = is_compressed(ni) ? &ntfs_aops_cmpr :
 							      &ntfs_aops;
 		if (ino != MFT_REC_MFT)
@@ -571,13 +577,18 @@ static noinline int ntfs_get_block_vbo(struct inode *inode, u64 vbo,
 	clear_buffer_uptodate(bh);
 
 	if (is_resident(ni)) {
-		ni_lock(ni);
-		err = attr_data_read_resident(ni, &folio->page);
-		ni_unlock(ni);
-
-		if (!err)
-			set_buffer_uptodate(bh);
+		bh->b_blocknr = RESIDENT_LCN;
 		bh->b_size = block_size;
+		if (!folio) {
+			err = 0;
+		} else {
+			ni_lock(ni);
+			err = attr_data_read_resident(ni, &folio->page);
+			ni_unlock(ni);
+
+			if (!err)
+				set_buffer_uptodate(bh);
+		}
 		return err;
 	}
 
@@ -1211,11 +1222,10 @@ out:
  *
  * NOTE: if fnd != NULL (ntfs_atomic_open) then @dir is locked
  */
-struct inode *ntfs_create_inode(struct mnt_idmap *idmap, struct inode *dir,
-				struct dentry *dentry,
-				const struct cpu_str *uni, umode_t mode,
-				dev_t dev, const char *symname, u32 size,
-				struct ntfs_fnd *fnd)
+int ntfs_create_inode(struct mnt_idmap *idmap, struct inode *dir,
+		      struct dentry *dentry, const struct cpu_str *uni,
+		      umode_t mode, dev_t dev, const char *symname, u32 size,
+		      struct ntfs_fnd *fnd)
 {
 	int err;
 	struct super_block *sb = dir->i_sb;
@@ -1240,6 +1250,9 @@ struct inode *ntfs_create_inode(struct mnt_idmap *idmap, struct inode *dir,
 	struct REPARSE_DATA_BUFFER *rp = NULL;
 	bool rp_inserted = false;
 
+	/* New file will be resident or non resident. */
+	const bool new_file_resident = 1;
+
 	if (!fnd)
 		ni_lock_dir(dir_ni);
 
@@ -1479,7 +1492,7 @@ struct inode *ntfs_create_inode(struct mnt_idmap *idmap, struct inode *dir,
 		attr->size = cpu_to_le32(SIZEOF_RESIDENT);
 		attr->name_off = SIZEOF_RESIDENT_LE;
 		attr->res.data_off = SIZEOF_RESIDENT_LE;
-	} else if (S_ISREG(mode)) {
+	} else if (!new_file_resident && S_ISREG(mode)) {
 		/*
 		 * Regular file. Create empty non resident data attribute.
 		 */
@@ -1495,7 +1508,7 @@ struct inode *ntfs_create_inode(struct mnt_idmap *idmap, struct inode *dir,
 			attr->size = cpu_to_le32(SIZEOF_NONRESIDENT_EX + 8);
 			attr->name_off = SIZEOF_NONRESIDENT_EX_LE;
 			attr->flags = ATTR_FLAG_COMPRESSED;
-			attr->nres.c_unit = COMPRESSION_UNIT;
+			attr->nres.c_unit = NTFS_LZNT_CUNIT;
 			asize = SIZEOF_NONRESIDENT_EX + 8;
 		} else {
 			attr->size = cpu_to_le32(SIZEOF_NONRESIDENT + 8);
@@ -1615,7 +1628,10 @@ struct inode *ntfs_create_inode(struct mnt_idmap *idmap, struct inode *dir,
 
 	if (S_ISDIR(mode)) {
 		inode->i_op = &ntfs_dir_inode_operations;
-		inode->i_fop = &ntfs_dir_operations;
+		if (is_legacy_ntfs(inode->i_sb))
+			inode->i_fop = &ntfs_legacy_dir_operations;
+		else
+			inode->i_fop = &ntfs_dir_operations;
 	} else if (S_ISLNK(mode)) {
 		inode->i_op = &ntfs_link_inode_operations;
 		inode->i_fop = NULL;
@@ -1624,7 +1640,10 @@ struct inode *ntfs_create_inode(struct mnt_idmap *idmap, struct inode *dir,
 		inode_nohighmem(inode);
 	} else if (S_ISREG(mode)) {
 		inode->i_op = &ntfs_file_inode_operations;
-		inode->i_fop = &ntfs_file_operations;
+		if (is_legacy_ntfs(inode->i_sb))
+			inode->i_fop = &ntfs_legacy_file_operations;
+		else
+			inode->i_fop = &ntfs_file_operations;
 		inode->i_mapping->a_ops = is_compressed(ni) ? &ntfs_aops_cmpr :
 							      &ntfs_aops;
 		init_rwsem(&ni->file.run_lock);
@@ -1649,7 +1668,9 @@ struct inode *ntfs_create_inode(struct mnt_idmap *idmap, struct inode *dir,
 	 * The packed size of extended attribute is stored in direntry too.
 	 * 'fname' here points to inside new_de.
 	 */
-	ntfs_save_wsl_perm(inode, &fname->dup.ea_size);
+	err = ntfs_save_wsl_perm(inode, &fname->dup.ea_size);
+	if (err)
+		goto out6;
 
 	/*
 	 * update ea_size in file_name attribute too.
@@ -1693,6 +1714,12 @@ struct inode *ntfs_create_inode(struct mnt_idmap *idmap, struct inode *dir,
 	goto out2;
 
 out6:
+	attr = ni_find_attr(ni, NULL, NULL, ATTR_EA, NULL, 0, NULL, NULL);
+	if (attr && attr->non_res) {
+		/* Delete ATTR_EA, if non-resident. */
+		attr_set_size(ni, ATTR_EA, NULL, 0, NULL, 0, NULL, false, NULL);
+	}
+
 	if (rp_inserted)
 		ntfs_remove_reparse(sbi, IO_REPARSE_TAG_SYMLINK, &new_de->ref);
 
@@ -1716,12 +1743,10 @@ out1:
 	if (!fnd)
 		ni_unlock(dir_ni);
 
-	if (err)
-		return ERR_PTR(err);
+	if (!err)
+		unlock_new_inode(inode);
 
-	unlock_new_inode(inode);
-
-	return inode;
+	return err;
 }
 
 int ntfs_link_inode(struct inode *inode, struct dentry *dentry)
@@ -2116,5 +2141,6 @@ const struct address_space_operations ntfs_aops = {
 const struct address_space_operations ntfs_aops_cmpr = {
 	.read_folio	= ntfs_read_folio,
 	.readahead	= ntfs_readahead,
+	.dirty_folio	= block_dirty_folio,
 };
 // clang-format on
diff --git a/fs/ntfs3/namei.c b/fs/ntfs3/namei.c
index cae41db0aa..71498421ce 100644
--- a/fs/ntfs3/namei.c
+++ b/fs/ntfs3/namei.c
@@ -107,12 +107,8 @@ static struct dentry *ntfs_lookup(struct inode *dir, struct dentry *dentry,
 static int ntfs_create(struct mnt_idmap *idmap, struct inode *dir,
 		       struct dentry *dentry, umode_t mode, bool excl)
 {
-	struct inode *inode;
-
-	inode = ntfs_create_inode(idmap, dir, dentry, NULL, S_IFREG | mode, 0,
-				  NULL, 0, NULL);
-
-	return IS_ERR(inode) ? PTR_ERR(inode) : 0;
+	return ntfs_create_inode(idmap, dir, dentry, NULL, S_IFREG | mode, 0,
+				 NULL, 0, NULL);
 }
 
 /*
@@ -123,12 +119,8 @@ static int ntfs_create(struct mnt_idmap *idmap, struct inode *dir,
 static int ntfs_mknod(struct mnt_idmap *idmap, struct inode *dir,
 		      struct dentry *dentry, umode_t mode, dev_t rdev)
 {
-	struct inode *inode;
-
-	inode = ntfs_create_inode(idmap, dir, dentry, NULL, mode, rdev, NULL, 0,
-				  NULL);
-
-	return IS_ERR(inode) ? PTR_ERR(inode) : 0;
+	return ntfs_create_inode(idmap, dir, dentry, NULL, mode, rdev, NULL, 0,
+				 NULL);
 }
 
 /*
@@ -200,15 +192,12 @@ static int ntfs_symlink(struct mnt_idmap *idmap, struct inode *dir,
 			struct dentry *dentry, const char *symname)
 {
 	u32 size = strlen(symname);
-	struct inode *inode;
 
 	if (unlikely(ntfs3_forced_shutdown(dir->i_sb)))
 		return -EIO;
 
-	inode = ntfs_create_inode(idmap, dir, dentry, NULL, S_IFLNK | 0777, 0,
-				  symname, size, NULL);
-
-	return IS_ERR(inode) ? PTR_ERR(inode) : 0;
+	return ntfs_create_inode(idmap, dir, dentry, NULL, S_IFLNK | 0777, 0,
+				 symname, size, NULL);
 }
 
 /*
@@ -217,12 +206,8 @@ static int ntfs_symlink(struct mnt_idmap *idmap, struct inode *dir,
 static int ntfs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
 		      struct dentry *dentry, umode_t mode)
 {
-	struct inode *inode;
-
-	inode = ntfs_create_inode(idmap, dir, dentry, NULL, S_IFDIR | mode, 0,
-				  NULL, 0, NULL);
-
-	return IS_ERR(inode) ? PTR_ERR(inode) : 0;
+	return ntfs_create_inode(idmap, dir, dentry, NULL, S_IFDIR | mode, 0,
+				 NULL, 0, NULL);
 }
 
 /*
@@ -358,95 +343,6 @@ out:
 	return err;
 }
 
-/*
- * ntfs_atomic_open
- *
- * inode_operations::atomic_open
- */
-static int ntfs_atomic_open(struct inode *dir, struct dentry *dentry,
-			    struct file *file, u32 flags, umode_t mode)
-{
-	int err;
-	struct inode *inode;
-	struct ntfs_fnd *fnd = NULL;
-	struct ntfs_inode *ni = ntfs_i(dir);
-	struct dentry *d = NULL;
-	struct cpu_str *uni = __getname();
-	bool locked = false;
-
-	if (!uni)
-		return -ENOMEM;
-
-	err = ntfs_nls_to_utf16(ni->mi.sbi, dentry->d_name.name,
-				dentry->d_name.len, uni, NTFS_NAME_LEN,
-				UTF16_HOST_ENDIAN);
-	if (err < 0)
-		goto out;
-
-#ifdef CONFIG_NTFS3_FS_POSIX_ACL
-	if (IS_POSIXACL(dir)) {
-		/*
-		 * Load in cache current acl to avoid ni_lock(dir):
-		 * ntfs_create_inode -> ntfs_init_acl -> posix_acl_create ->
-		 * ntfs_get_acl -> ntfs_get_acl_ex -> ni_lock
-		 */
-		struct posix_acl *p = get_inode_acl(dir, ACL_TYPE_DEFAULT);
-
-		if (IS_ERR(p)) {
-			err = PTR_ERR(p);
-			goto out;
-		}
-		posix_acl_release(p);
-	}
-#endif
-
-	if (d_in_lookup(dentry)) {
-		ni_lock_dir(ni);
-		locked = true;
-		fnd = fnd_get();
-		if (!fnd) {
-			err = -ENOMEM;
-			goto out1;
-		}
-
-		d = d_splice_alias(dir_search_u(dir, uni, fnd), dentry);
-		if (IS_ERR(d)) {
-			err = PTR_ERR(d);
-			d = NULL;
-			goto out2;
-		}
-
-		if (d)
-			dentry = d;
-	}
-
-	if (!(flags & O_CREAT) || d_really_is_positive(dentry)) {
-		err = finish_no_open(file, d);
-		goto out2;
-	}
-
-	file->f_mode |= FMODE_CREATED;
-
-	/*
-	 * fnd contains tree's path to insert to.
-	 * If fnd is not NULL then dir is locked.
-	 */
-	inode = ntfs_create_inode(mnt_idmap(file->f_path.mnt), dir, dentry, uni,
-				  mode, 0, NULL, 0, fnd);
-	err = IS_ERR(inode) ? PTR_ERR(inode) :
-			      finish_open(file, dentry, ntfs_file_open);
-	dput(d);
-
-out2:
-	fnd_put(fnd);
-out1:
-	if (locked)
-		ni_unlock(ni);
-out:
-	__putname(uni);
-	return err;
-}
-
 struct dentry *ntfs3_get_parent(struct dentry *child)
 {
 	struct inode *inode = d_inode(child);
@@ -612,7 +508,6 @@ const struct inode_operations ntfs_dir_inode_operations = {
 	.setattr	= ntfs3_setattr,
 	.getattr	= ntfs_getattr,
 	.listxattr	= ntfs_listxattr,
-	.atomic_open	= ntfs_atomic_open,
 	.fiemap		= ntfs_fiemap,
 };
 
diff --git a/fs/ntfs3/ntfs.h b/fs/ntfs3/ntfs.h
index 9c7478150a..e1889ad092 100644
--- a/fs/ntfs3/ntfs.h
+++ b/fs/ntfs3/ntfs.h
@@ -59,7 +59,7 @@ struct GUID {
 struct cpu_str {
 	u8 len;
 	u8 unused;
-	u16 name[10];
+	u16 name[];
 };
 
 struct le_str {
@@ -82,9 +82,6 @@ typedef u32 CLST;
 #define RESIDENT_LCN   ((CLST)-2)
 #define COMPRESSED_LCN ((CLST)-3)
 
-#define COMPRESSION_UNIT     4
-#define COMPRESS_MAX_CLUSTER 0x1000
-
 enum RECORD_NUM {
 	MFT_REC_MFT		= 0,
 	MFT_REC_MIRR		= 1,
@@ -696,14 +693,15 @@ static inline bool de_has_vcn_ex(const struct NTFS_DE *e)
 	      offsetof(struct ATTR_FILE_NAME, name) + \
 	      NTFS_NAME_LEN * sizeof(short), 8)
 
+#define NTFS_INDEX_HDR_HAS_SUBNODES cpu_to_le32(1)
+
 struct INDEX_HDR {
 	__le32 de_off;	// 0x00: The offset from the start of this structure
 			// to the first NTFS_DE.
 	__le32 used;	// 0x04: The size of this structure plus all
 			// entries (quad-word aligned).
 	__le32 total;	// 0x08: The allocated size of for this structure plus all entries.
-	u8 flags;	// 0x0C: 0x00 = Small directory, 0x01 = Large directory.
-	u8 res[3];
+	__le32 flags;	// 0x0C: 0x00 = Small directory, 0x01 = Large directory.
 
 	//
 	// de_off + used <= total
@@ -751,7 +749,7 @@ static inline struct NTFS_DE *hdr_next_de(const struct INDEX_HDR *hdr,
 
 static inline bool hdr_has_subnode(const struct INDEX_HDR *hdr)
 {
-	return hdr->flags & 1;
+	return hdr->flags & NTFS_INDEX_HDR_HAS_SUBNODES;
 }
 
 struct INDEX_BUFFER {
@@ -771,7 +769,7 @@ static inline bool ib_is_empty(const struct INDEX_BUFFER *ib)
 
 static inline bool ib_is_leaf(const struct INDEX_BUFFER *ib)
 {
-	return !(ib->ihdr.flags & 1);
+	return !(ib->ihdr.flags & NTFS_INDEX_HDR_HAS_SUBNODES);
 }
 
 /* Index root structure ( 0x90 ). */
diff --git a/fs/ntfs3/ntfs_fs.h b/fs/ntfs3/ntfs_fs.h
index 79356fd29a..f9ed6d2b06 100644
--- a/fs/ntfs3/ntfs_fs.h
+++ b/fs/ntfs3/ntfs_fs.h
@@ -452,6 +452,7 @@ int attr_allocate_frame(struct ntfs_inode *ni, CLST frame, size_t compr_size,
 int attr_collapse_range(struct ntfs_inode *ni, u64 vbo, u64 bytes);
 int attr_insert_range(struct ntfs_inode *ni, u64 vbo, u64 bytes);
 int attr_punch_hole(struct ntfs_inode *ni, u64 vbo, u64 bytes, u32 *frame_size);
+int attr_force_nonresident(struct ntfs_inode *ni);
 
 /* Functions from attrlist.c */
 void al_destroy(struct ntfs_inode *ni);
@@ -493,6 +494,7 @@ struct inode *dir_search_u(struct inode *dir, const struct cpu_str *uni,
 			   struct ntfs_fnd *fnd);
 bool dir_is_empty(struct inode *dir);
 extern const struct file_operations ntfs_dir_operations;
+extern const struct file_operations ntfs_legacy_dir_operations;
 
 /* Globals from file.c */
 int ntfs_getattr(struct mnt_idmap *idmap, const struct path *path,
@@ -507,6 +509,7 @@ long ntfs_compat_ioctl(struct file *filp, u32 cmd, unsigned long arg);
 extern const struct inode_operations ntfs_special_inode_operations;
 extern const struct inode_operations ntfs_file_inode_operations;
 extern const struct file_operations ntfs_file_operations;
+extern const struct file_operations ntfs_legacy_file_operations;
 
 /* Globals from frecord.c */
 void ni_remove_mi(struct ntfs_inode *ni, struct mft_inode *mi);
@@ -714,11 +717,10 @@ int ntfs_sync_inode(struct inode *inode);
 int ntfs_flush_inodes(struct super_block *sb, struct inode *i1,
 		      struct inode *i2);
 int inode_write_data(struct inode *inode, const void *data, size_t bytes);
-struct inode *ntfs_create_inode(struct mnt_idmap *idmap, struct inode *dir,
-				struct dentry *dentry,
-				const struct cpu_str *uni, umode_t mode,
-				dev_t dev, const char *symname, u32 size,
-				struct ntfs_fnd *fnd);
+int ntfs_create_inode(struct mnt_idmap *idmap, struct inode *dir,
+		      struct dentry *dentry, const struct cpu_str *uni,
+		      umode_t mode, dev_t dev, const char *symname, u32 size,
+		      struct ntfs_fnd *fnd);
 int ntfs_link_inode(struct inode *inode, struct dentry *dentry);
 int ntfs_unlink_inode(struct inode *dir, const struct dentry *dentry);
 void ntfs_evict_inode(struct inode *inode);
@@ -966,9 +968,9 @@ static inline bool run_is_empty(struct runs_tree *run)
 }
 
 /* NTFS uses quad aligned bitmaps. */
-static inline size_t bitmap_size(size_t bits)
+static inline size_t ntfs3_bitmap_size(size_t bits)
 {
-	return ALIGN((bits + 7) >> 3, 8);
+	return BITS_TO_U64(bits) * sizeof(u64);
 }
 
 #define _100ns2seconds 10000000
@@ -1154,4 +1156,6 @@ static inline void le64_sub_cpu(__le64 *var, u64 val)
 	*var = cpu_to_le64(le64_to_cpu(*var) - val);
 }
 
+bool is_legacy_ntfs(struct super_block *sb);
+
 #endif /* _LINUX_NTFS3_NTFS_FS_H */
diff --git a/fs/ntfs3/super.c b/fs/ntfs3/super.c
index 4643b06b15..02b6f51ce6 100644
--- a/fs/ntfs3/super.c
+++ b/fs/ntfs3/super.c
@@ -275,7 +275,7 @@ static const struct fs_parameter_spec ntfs_fs_parameters[] = {
 	fsparam_flag_no("acl",			Opt_acl),
 	fsparam_string("iocharset",		Opt_iocharset),
 	fsparam_flag_no("prealloc",		Opt_prealloc),
-	fsparam_flag_no("nocase",		Opt_nocase),
+	fsparam_flag_no("case",		Opt_nocase),
 	{}
 };
 // clang-format on
@@ -408,6 +408,12 @@ static int ntfs_fs_reconfigure(struct fs_context *fc)
 	struct ntfs_mount_options *new_opts = fc->fs_private;
 	int ro_rw;
 
+	/* If ntfs3 is used as legacy ntfs enforce read-only mode. */
+	if (is_legacy_ntfs(sb)) {
+		fc->sb_flags |= SB_RDONLY;
+		goto out;
+	}
+
 	ro_rw = sb_rdonly(sb) && !(fc->sb_flags & SB_RDONLY);
 	if (ro_rw && (sbi->flags & NTFS_FLAGS_NEED_REPLAY)) {
 		errorf(fc,
@@ -427,8 +433,6 @@ static int ntfs_fs_reconfigure(struct fs_context *fc)
 			fc,
 			"ntfs3: Cannot use different iocharset when remounting!");
 
-	sync_filesystem(sb);
-
 	if (ro_rw && (sbi->volume.flags & VOLUME_FLAG_DIRTY) &&
 	    !new_opts->force) {
 		errorf(fc,
@@ -436,6 +440,8 @@ static int ntfs_fs_reconfigure(struct fs_context *fc)
 		return -EINVAL;
 	}
 
+out:
+	sync_filesystem(sb);
 	swap(sbi->options, fc->fs_private);
 
 	return 0;
@@ -462,7 +468,7 @@ static int ntfs3_volinfo(struct seq_file *m, void *o)
 	struct super_block *sb = m->private;
 	struct ntfs_sb_info *sbi = sb->s_fs_info;
 
-	seq_printf(m, "ntfs%d.%d\n%u\n%zu\n\%zu\n%zu\n%s\n%s\n",
+	seq_printf(m, "ntfs%d.%d\n%u\n%zu\n%zu\n%zu\n%s\n%s\n",
 		   sbi->volume.major_ver, sbi->volume.minor_ver,
 		   sbi->cluster_size, sbi->used.bitmap.nbits,
 		   sbi->mft.bitmap.nbits,
@@ -1341,7 +1347,7 @@ static int ntfs_fill_super(struct super_block *sb, struct fs_context *fc)
 
 	/* Check bitmap boundary. */
 	tt = sbi->used.bitmap.nbits;
-	if (inode->i_size < bitmap_size(tt)) {
+	if (inode->i_size < ntfs3_bitmap_size(tt)) {
 		ntfs_err(sb, "$Bitmap is corrupted.");
 		err = -EINVAL;
 		goto put_inode_out;
@@ -1613,6 +1619,8 @@ load_root:
 	}
 #endif
 
+	if (is_legacy_ntfs(sb))
+		sb->s_flags |= SB_RDONLY;
 	return 0;
 
 put_inode_out:
@@ -1730,7 +1738,7 @@ static const struct fs_context_operations ntfs_context_ops = {
  * This will called when mount/remount. We will first initialize
  * options so that if remount we can use just that.
  */
-static int ntfs_init_fs_context(struct fs_context *fc)
+static int __ntfs_init_fs_context(struct fs_context *fc)
 {
 	struct ntfs_mount_options *opts;
 	struct ntfs_sb_info *sbi;
@@ -1778,6 +1786,11 @@ free_opts:
 	return -ENOMEM;
 }
 
+static int ntfs_init_fs_context(struct fs_context *fc)
+{
+	return __ntfs_init_fs_context(fc);
+}
+
 static void ntfs3_kill_sb(struct super_block *sb)
 {
 	struct ntfs_sb_info *sbi = sb->s_fs_info;
@@ -1798,6 +1811,50 @@ static struct file_system_type ntfs_fs_type = {
 	.kill_sb		= ntfs3_kill_sb,
 	.fs_flags		= FS_REQUIRES_DEV | FS_ALLOW_IDMAP,
 };
+
+#if IS_ENABLED(CONFIG_NTFS_FS)
+static int ntfs_legacy_init_fs_context(struct fs_context *fc)
+{
+	int ret;
+
+	ret = __ntfs_init_fs_context(fc);
+	/* If ntfs3 is used as legacy ntfs enforce read-only mode. */
+	fc->sb_flags |= SB_RDONLY;
+	return ret;
+}
+
+static struct file_system_type ntfs_legacy_fs_type = {
+	.owner			= THIS_MODULE,
+	.name			= "ntfs",
+	.init_fs_context	= ntfs_legacy_init_fs_context,
+	.parameters		= ntfs_fs_parameters,
+	.kill_sb		= ntfs3_kill_sb,
+	.fs_flags		= FS_REQUIRES_DEV | FS_ALLOW_IDMAP,
+};
+MODULE_ALIAS_FS("ntfs");
+
+static inline void register_as_ntfs_legacy(void)
+{
+	int err = register_filesystem(&ntfs_legacy_fs_type);
+	if (err)
+		pr_warn("ntfs3: Failed to register legacy ntfs filesystem driver: %d\n", err);
+}
+
+static inline void unregister_as_ntfs_legacy(void)
+{
+	unregister_filesystem(&ntfs_legacy_fs_type);
+}
+bool is_legacy_ntfs(struct super_block *sb)
+{
+	return sb->s_type == &ntfs_legacy_fs_type;
+}
+#else
+static inline void register_as_ntfs_legacy(void) {}
+static inline void unregister_as_ntfs_legacy(void) {}
+bool is_legacy_ntfs(struct super_block *sb) { return false; }
+#endif
+
+
 // clang-format on
 
 static int __init init_ntfs_fs(void)
@@ -1823,13 +1880,14 @@ static int __init init_ntfs_fs(void)
 
 	ntfs_inode_cachep = kmem_cache_create(
 		"ntfs_inode_cache", sizeof(struct ntfs_inode), 0,
-		(SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD | SLAB_ACCOUNT),
+		(SLAB_RECLAIM_ACCOUNT | SLAB_ACCOUNT),
 		init_once);
 	if (!ntfs_inode_cachep) {
 		err = -ENOMEM;
 		goto out1;
 	}
 
+	register_as_ntfs_legacy();
 	err = register_filesystem(&ntfs_fs_type);
 	if (err)
 		goto out;
@@ -1847,6 +1905,7 @@ static void __exit exit_ntfs_fs(void)
 	rcu_barrier();
 	kmem_cache_destroy(ntfs_inode_cachep);
 	unregister_filesystem(&ntfs_fs_type);
+	unregister_as_ntfs_legacy();
 	ntfs3_exit_bitmap();
 
 #ifdef CONFIG_PROC_FS
diff --git a/fs/ntfs3/xattr.c b/fs/ntfs3/xattr.c
index 53e7d1fa03..73785dece7 100644
--- a/fs/ntfs3/xattr.c
+++ b/fs/ntfs3/xattr.c
@@ -219,8 +219,11 @@ static ssize_t ntfs_list_ea(struct ntfs_inode *ni, char *buffer,
 		if (!ea->name_len)
 			break;
 
-		if (ea->name_len > ea_size)
+		if (ea->name_len > ea_size) {
+			ntfs_set_state(ni->mi.sbi, NTFS_DIRTY_ERROR);
+			err = -EINVAL; /* corrupted fs */
 			break;
+		}
 
 		if (buffer) {
 			/* Check if we can use field ea->name */
diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
index b82185075d..6be175a1ab 100644
--- a/fs/ocfs2/aops.c
+++ b/fs/ocfs2/aops.c
@@ -2283,8 +2283,6 @@ unlock:
 	ocfs2_inode_unlock(inode, 1);
 	brelse(di_bh);
 out:
-	if (ret < 0)
-		ret = -EIO;
 	return ret;
 }
 
@@ -2368,6 +2366,11 @@ static int ocfs2_dio_end_io_write(struct inode *inode,
 	}
 
 	list_for_each_entry(ue, &dwc->dw_zero_list, ue_node) {
+		ret = ocfs2_assure_trans_credits(handle, credits);
+		if (ret < 0) {
+			mlog_errno(ret);
+			break;
+		}
 		ret = ocfs2_mark_extent_written(inode, &et, handle,
 						ue->ue_cpos, 1,
 						ue->ue_phys,
diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c
index 4d7efefa98..1bde1281d5 100644
--- a/fs/ocfs2/cluster/heartbeat.c
+++ b/fs/ocfs2/cluster/heartbeat.c
@@ -213,7 +213,7 @@ struct o2hb_region {
 	unsigned int		hr_num_pages;
 
 	struct page             **hr_slot_data;
-	struct bdev_handle	*hr_bdev_handle;
+	struct file		*hr_bdev_file;
 	struct o2hb_disk_slot	*hr_slots;
 
 	/* live node map of this region */
@@ -263,7 +263,7 @@ struct o2hb_region {
 
 static inline struct block_device *reg_bdev(struct o2hb_region *reg)
 {
-	return reg->hr_bdev_handle ? reg->hr_bdev_handle->bdev : NULL;
+	return reg->hr_bdev_file ? file_bdev(reg->hr_bdev_file) : NULL;
 }
 
 struct o2hb_bio_wait_ctxt {
@@ -1509,8 +1509,8 @@ static void o2hb_region_release(struct config_item *item)
 		kfree(reg->hr_slot_data);
 	}
 
-	if (reg->hr_bdev_handle)
-		bdev_release(reg->hr_bdev_handle);
+	if (reg->hr_bdev_file)
+		fput(reg->hr_bdev_file);
 
 	kfree(reg->hr_slots);
 
@@ -1569,7 +1569,7 @@ static ssize_t o2hb_region_block_bytes_store(struct config_item *item,
 	unsigned long block_bytes;
 	unsigned int block_bits;
 
-	if (reg->hr_bdev_handle)
+	if (reg->hr_bdev_file)
 		return -EINVAL;
 
 	status = o2hb_read_block_input(reg, page, &block_bytes,
@@ -1598,7 +1598,7 @@ static ssize_t o2hb_region_start_block_store(struct config_item *item,
 	char *p = (char *)page;
 	ssize_t ret;
 
-	if (reg->hr_bdev_handle)
+	if (reg->hr_bdev_file)
 		return -EINVAL;
 
 	ret = kstrtoull(p, 0, &tmp);
@@ -1623,7 +1623,7 @@ static ssize_t o2hb_region_blocks_store(struct config_item *item,
 	unsigned long tmp;
 	char *p = (char *)page;
 
-	if (reg->hr_bdev_handle)
+	if (reg->hr_bdev_file)
 		return -EINVAL;
 
 	tmp = simple_strtoul(p, &p, 0);
@@ -1642,7 +1642,7 @@ static ssize_t o2hb_region_dev_show(struct config_item *item, char *page)
 {
 	unsigned int ret = 0;
 
-	if (to_o2hb_region(item)->hr_bdev_handle)
+	if (to_o2hb_region(item)->hr_bdev_file)
 		ret = sprintf(page, "%pg\n", reg_bdev(to_o2hb_region(item)));
 
 	return ret;
@@ -1753,7 +1753,7 @@ out:
 }
 
 /*
- * this is acting as commit; we set up all of hr_bdev_handle and hr_task or
+ * this is acting as commit; we set up all of hr_bdev_file and hr_task or
  * nothing
  */
 static ssize_t o2hb_region_dev_store(struct config_item *item,
@@ -1769,7 +1769,7 @@ static ssize_t o2hb_region_dev_store(struct config_item *item,
 	ssize_t ret = -EINVAL;
 	int live_threshold;
 
-	if (reg->hr_bdev_handle)
+	if (reg->hr_bdev_file)
 		goto out;
 
 	/* We can't heartbeat without having had our node number
@@ -1795,11 +1795,11 @@ static ssize_t o2hb_region_dev_store(struct config_item *item,
 	if (!S_ISBLK(f.file->f_mapping->host->i_mode))
 		goto out2;
 
-	reg->hr_bdev_handle = bdev_open_by_dev(f.file->f_mapping->host->i_rdev,
+	reg->hr_bdev_file = bdev_file_open_by_dev(f.file->f_mapping->host->i_rdev,
 			BLK_OPEN_WRITE | BLK_OPEN_READ, NULL, NULL);
-	if (IS_ERR(reg->hr_bdev_handle)) {
-		ret = PTR_ERR(reg->hr_bdev_handle);
-		reg->hr_bdev_handle = NULL;
+	if (IS_ERR(reg->hr_bdev_file)) {
+		ret = PTR_ERR(reg->hr_bdev_file);
+		reg->hr_bdev_file = NULL;
 		goto out2;
 	}
 
@@ -1903,8 +1903,8 @@ static ssize_t o2hb_region_dev_store(struct config_item *item,
 
 out3:
 	if (ret < 0) {
-		bdev_release(reg->hr_bdev_handle);
-		reg->hr_bdev_handle = NULL;
+		fput(reg->hr_bdev_file);
+		reg->hr_bdev_file = NULL;
 	}
 out2:
 	fdput(f);
diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c
index 960080753d..2b8fa3e782 100644
--- a/fs/ocfs2/cluster/tcp.c
+++ b/fs/ocfs2/cluster/tcp.c
@@ -1784,6 +1784,9 @@ static int o2net_accept_one(struct socket *sock, int *more)
 	struct o2nm_node *node = NULL;
 	struct o2nm_node *local_node = NULL;
 	struct o2net_sock_container *sc = NULL;
+	struct proto_accept_arg arg = {
+		.flags = O_NONBLOCK,
+	};
 	struct o2net_node *nn;
 	unsigned int nofs_flag;
 
@@ -1802,7 +1805,7 @@ static int o2net_accept_one(struct socket *sock, int *more)
 
 	new_sock->type = sock->type;
 	new_sock->ops = sock->ops;
-	ret = sock->ops->accept(sock, new_sock, O_NONBLOCK, false);
+	ret = sock->ops->accept(sock, new_sock, &arg);
 	if (ret < 0)
 		goto out;
 
diff --git a/fs/ocfs2/dir.c b/fs/ocfs2/dir.c
index d620d4c53c..f0beb173db 100644
--- a/fs/ocfs2/dir.c
+++ b/fs/ocfs2/dir.c
@@ -294,13 +294,16 @@ out:
  * bh passed here can be an inode block or a dir data block, depending
  * on the inode inline data flag.
  */
-static int ocfs2_check_dir_entry(struct inode * dir,
-				 struct ocfs2_dir_entry * de,
-				 struct buffer_head * bh,
+static int ocfs2_check_dir_entry(struct inode *dir,
+				 struct ocfs2_dir_entry *de,
+				 struct buffer_head *bh,
+				 char *buf,
+				 unsigned int size,
 				 unsigned long offset)
 {
 	const char *error_msg = NULL;
 	const int rlen = le16_to_cpu(de->rec_len);
+	const unsigned long next_offset = ((char *) de - buf) + rlen;
 
 	if (unlikely(rlen < OCFS2_DIR_REC_LEN(1)))
 		error_msg = "rec_len is smaller than minimal";
@@ -308,9 +311,11 @@ static int ocfs2_check_dir_entry(struct inode * dir,
 		error_msg = "rec_len % 4 != 0";
 	else if (unlikely(rlen < OCFS2_DIR_REC_LEN(de->name_len)))
 		error_msg = "rec_len is too small for name_len";
-	else if (unlikely(
-		 ((char *) de - bh->b_data) + rlen > dir->i_sb->s_blocksize))
-		error_msg = "directory entry across blocks";
+	else if (unlikely(next_offset > size))
+		error_msg = "directory entry overrun";
+	else if (unlikely(next_offset > size - OCFS2_DIR_REC_LEN(1)) &&
+		 next_offset != size)
+		error_msg = "directory entry too close to end";
 
 	if (unlikely(error_msg != NULL))
 		mlog(ML_ERROR, "bad entry in directory #%llu: %s - "
@@ -352,16 +357,17 @@ static inline int ocfs2_search_dirblock(struct buffer_head *bh,
 	de_buf = first_de;
 	dlimit = de_buf + bytes;
 
-	while (de_buf < dlimit) {
+	while (de_buf < dlimit - OCFS2_DIR_MEMBER_LEN) {
 		/* this code is executed quadratically often */
 		/* do minimal checking `by hand' */
 
 		de = (struct ocfs2_dir_entry *) de_buf;
 
-		if (de_buf + namelen <= dlimit &&
+		if (de->name + namelen <= dlimit &&
 		    ocfs2_match(namelen, name, de)) {
 			/* found a match - just to be sure, do a full check */
-			if (!ocfs2_check_dir_entry(dir, de, bh, offset)) {
+			if (!ocfs2_check_dir_entry(dir, de, bh, first_de,
+						   bytes, offset)) {
 				ret = -1;
 				goto bail;
 			}
@@ -1138,7 +1144,7 @@ static int __ocfs2_delete_entry(handle_t *handle, struct inode *dir,
 	pde = NULL;
 	de = (struct ocfs2_dir_entry *) first_de;
 	while (i < bytes) {
-		if (!ocfs2_check_dir_entry(dir, de, bh, i)) {
+		if (!ocfs2_check_dir_entry(dir, de, bh, first_de, bytes, i)) {
 			status = -EIO;
 			mlog_errno(status);
 			goto bail;
@@ -1635,7 +1641,8 @@ int __ocfs2_add_entry(handle_t *handle,
 		/* These checks should've already been passed by the
 		 * prepare function, but I guess we can leave them
 		 * here anyway. */
-		if (!ocfs2_check_dir_entry(dir, de, insert_bh, offset)) {
+		if (!ocfs2_check_dir_entry(dir, de, insert_bh, data_start,
+					   size, offset)) {
 			retval = -ENOENT;
 			goto bail;
 		}
@@ -1774,7 +1781,8 @@ static int ocfs2_dir_foreach_blk_id(struct inode *inode,
 		}
 
 		de = (struct ocfs2_dir_entry *) (data->id_data + ctx->pos);
-		if (!ocfs2_check_dir_entry(inode, de, di_bh, ctx->pos)) {
+		if (!ocfs2_check_dir_entry(inode, de, di_bh, (char *)data->id_data,
+					   i_size_read(inode), ctx->pos)) {
 			/* On error, skip the f_pos to the end. */
 			ctx->pos = i_size_read(inode);
 			break;
@@ -1867,7 +1875,8 @@ static int ocfs2_dir_foreach_blk_el(struct inode *inode,
 		while (ctx->pos < i_size_read(inode)
 		       && offset < sb->s_blocksize) {
 			de = (struct ocfs2_dir_entry *) (bh->b_data + offset);
-			if (!ocfs2_check_dir_entry(inode, de, bh, offset)) {
+			if (!ocfs2_check_dir_entry(inode, de, bh, bh->b_data,
+						   sb->s_blocksize, offset)) {
 				/* On error, skip the f_pos to the
 				   next block. */
 				ctx->pos = (ctx->pos | (sb->s_blocksize - 1)) + 1;
@@ -3339,7 +3348,7 @@ static int ocfs2_find_dir_space_id(struct inode *dir, struct buffer_head *di_bh,
 	struct super_block *sb = dir->i_sb;
 	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
 	struct ocfs2_dir_entry *de, *last_de = NULL;
-	char *de_buf, *limit;
+	char *first_de, *de_buf, *limit;
 	unsigned long offset = 0;
 	unsigned int rec_len, new_rec_len, free_space;
 
@@ -3352,14 +3361,16 @@ static int ocfs2_find_dir_space_id(struct inode *dir, struct buffer_head *di_bh,
 	else
 		free_space = dir->i_sb->s_blocksize - i_size_read(dir);
 
-	de_buf = di->id2.i_data.id_data;
+	first_de = di->id2.i_data.id_data;
+	de_buf = first_de;
 	limit = de_buf + i_size_read(dir);
 	rec_len = OCFS2_DIR_REC_LEN(namelen);
 
 	while (de_buf < limit) {
 		de = (struct ocfs2_dir_entry *)de_buf;
 
-		if (!ocfs2_check_dir_entry(dir, de, di_bh, offset)) {
+		if (!ocfs2_check_dir_entry(dir, de, di_bh, first_de,
+					   i_size_read(dir), offset)) {
 			ret = -ENOENT;
 			goto out;
 		}
@@ -3441,7 +3452,8 @@ static int ocfs2_find_dir_space_el(struct inode *dir, const char *name,
 			/* move to next block */
 			de = (struct ocfs2_dir_entry *) bh->b_data;
 		}
-		if (!ocfs2_check_dir_entry(dir, de, bh, offset)) {
+		if (!ocfs2_check_dir_entry(dir, de, bh, bh->b_data, blocksize,
+					   offset)) {
 			status = -ENOENT;
 			goto bail;
 		}
diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c
index 5c04dde999..2018501b22 100644
--- a/fs/ocfs2/dlm/dlmdomain.c
+++ b/fs/ocfs2/dlm/dlmdomain.c
@@ -1274,7 +1274,7 @@ static int dlm_query_nodeinfo_handler(struct o2net_msg *msg, u32 len,
 {
 	struct dlm_query_nodeinfo *qn;
 	struct dlm_ctxt *dlm = NULL;
-	int locked = 0, status = -EINVAL;
+	int status = -EINVAL;
 
 	qn = (struct dlm_query_nodeinfo *) msg->buf;
 
@@ -1290,12 +1290,11 @@ static int dlm_query_nodeinfo_handler(struct o2net_msg *msg, u32 len,
 	}
 
 	spin_lock(&dlm->spinlock);
-	locked = 1;
 	if (dlm->joining_node != qn->qn_nodenum) {
 		mlog(ML_ERROR, "Node %d queried nodes on domain %s but "
 		     "joining node is %d\n", qn->qn_nodenum, qn->qn_domain,
 		     dlm->joining_node);
-		goto bail;
+		goto unlock;
 	}
 
 	/* Support for node query was added in 1.1 */
@@ -1305,14 +1304,14 @@ static int dlm_query_nodeinfo_handler(struct o2net_msg *msg, u32 len,
 		     "but active dlm protocol is %d.%d\n", qn->qn_nodenum,
 		     qn->qn_domain, dlm->dlm_locking_proto.pv_major,
 		     dlm->dlm_locking_proto.pv_minor);
-		goto bail;
+		goto unlock;
 	}
 
 	status = dlm_match_nodes(dlm, qn);
 
+unlock:
+	spin_unlock(&dlm->spinlock);
 bail:
-	if (locked)
-		spin_unlock(&dlm->spinlock);
 	spin_unlock(&dlm_domain_lock);
 
 	return status;
@@ -1528,7 +1527,6 @@ static void dlm_send_join_asserts(struct dlm_ctxt *dlm,
 {
 	int status, node, live;
 
-	status = 0;
 	node = -1;
 	while ((node = find_next_bit(node_map, O2NM_MAX_NODES,
 				     node + 1)) < O2NM_MAX_NODES) {
diff --git a/fs/ocfs2/dlmfs/dlmfs.c b/fs/ocfs2/dlmfs/dlmfs.c
index 85215162c9..7fc0e920ed 100644
--- a/fs/ocfs2/dlmfs/dlmfs.c
+++ b/fs/ocfs2/dlmfs/dlmfs.c
@@ -578,7 +578,7 @@ static int __init init_dlmfs_fs(void)
 	dlmfs_inode_cache = kmem_cache_create("dlmfs_inode_cache",
 				sizeof(struct dlmfs_inode_private),
 				0, (SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
-					SLAB_MEM_SPREAD|SLAB_ACCOUNT),
+					SLAB_ACCOUNT),
 				dlmfs_init_once);
 	if (!dlmfs_inode_cache) {
 		status = -ENOMEM;
diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c
index 64a6ef6384..cb40cafbc0 100644
--- a/fs/ocfs2/dlmglue.c
+++ b/fs/ocfs2/dlmglue.c
@@ -1615,7 +1615,7 @@ update_holders:
 unlock:
 	lockres_clear_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING);
 
-	/* ocfs2_unblock_lock reques on seeing OCFS2_LOCK_UPCONVERT_FINISHING */
+	/* ocfs2_unblock_lock request on seeing OCFS2_LOCK_UPCONVERT_FINISHING */
 	kick_dc = (lockres->l_flags & OCFS2_LOCK_BLOCKED);
 
 	spin_unlock_irqrestore(&lockres->l_lock, flags);
diff --git a/fs/ocfs2/export.c b/fs/ocfs2/export.c
index b8b6a191b5..96b684763b 100644
--- a/fs/ocfs2/export.c
+++ b/fs/ocfs2/export.c
@@ -255,9 +255,9 @@ static struct dentry *ocfs2_fh_to_dentry(struct super_block *sb,
 	if (fh_len < 3 || fh_type > 2)
 		return NULL;
 
-	handle.ih_blkno = (u64)le32_to_cpu(fid->raw[0]) << 32;
-	handle.ih_blkno |= (u64)le32_to_cpu(fid->raw[1]);
-	handle.ih_generation = le32_to_cpu(fid->raw[2]);
+	handle.ih_blkno = (u64)le32_to_cpu((__force __le32)fid->raw[0]) << 32;
+	handle.ih_blkno |= (u64)le32_to_cpu((__force __le32)fid->raw[1]);
+	handle.ih_generation = le32_to_cpu((__force __le32)fid->raw[2]);
 	return ocfs2_get_dentry(sb, &handle);
 }
 
@@ -269,9 +269,9 @@ static struct dentry *ocfs2_fh_to_parent(struct super_block *sb,
 	if (fh_type != 2 || fh_len < 6)
 		return NULL;
 
-	parent.ih_blkno = (u64)le32_to_cpu(fid->raw[3]) << 32;
-	parent.ih_blkno |= (u64)le32_to_cpu(fid->raw[4]);
-	parent.ih_generation = le32_to_cpu(fid->raw[5]);
+	parent.ih_blkno = (u64)le32_to_cpu((__force __le32)fid->raw[3]) << 32;
+	parent.ih_blkno |= (u64)le32_to_cpu((__force __le32)fid->raw[4]);
+	parent.ih_generation = le32_to_cpu((__force __le32)fid->raw[5]);
 	return ocfs2_get_dentry(sb, &parent);
 }
 
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
index 8b6d150107..ccc57038a9 100644
--- a/fs/ocfs2/file.c
+++ b/fs/ocfs2/file.c
@@ -1936,6 +1936,8 @@ static int __ocfs2_change_file_space(struct file *file, struct inode *inode,
 
 	inode_lock(inode);
 
+	/* Wait all existing dio workers, newcomers will block on i_rwsem */
+	inode_dio_wait(inode);
 	/*
 	 * This prevents concurrent writes on other nodes
 	 */
@@ -2763,6 +2765,7 @@ const struct inode_operations ocfs2_file_iops = {
 const struct inode_operations ocfs2_special_file_iops = {
 	.setattr	= ocfs2_setattr,
 	.getattr	= ocfs2_getattr,
+	.listxattr	= ocfs2_listxattr,
 	.permission	= ocfs2_permission,
 	.get_inode_acl	= ocfs2_iop_get_acl,
 	.set_acl	= ocfs2_iop_set_acl,
diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c
index 999111bfc2..2cc5c99fe9 100644
--- a/fs/ocfs2/inode.c
+++ b/fs/ocfs2/inode.c
@@ -1621,6 +1621,7 @@ static struct super_block *ocfs2_inode_cache_get_super(struct ocfs2_caching_info
 }
 
 static void ocfs2_inode_cache_lock(struct ocfs2_caching_info *ci)
+__acquires(&oi->ip_lock)
 {
 	struct ocfs2_inode_info *oi = cache_info_to_inode(ci);
 
@@ -1628,6 +1629,7 @@ static void ocfs2_inode_cache_lock(struct ocfs2_caching_info *ci)
 }
 
 static void ocfs2_inode_cache_unlock(struct ocfs2_caching_info *ci)
+__releases(&oi->ip_lock)
 {
 	struct ocfs2_inode_info *oi = cache_info_to_inode(ci);
 
diff --git a/fs/ocfs2/ioctl.c b/fs/ocfs2/ioctl.c
index b1550ba73f..71beef7f8a 100644
--- a/fs/ocfs2/ioctl.c
+++ b/fs/ocfs2/ioctl.c
@@ -125,6 +125,7 @@ int ocfs2_fileattr_set(struct mnt_idmap *idmap,
 
 	ocfs2_inode->ip_attr = flags;
 	ocfs2_set_inode_flags(inode);
+	inode_set_ctime_current(inode);
 
 	status = ocfs2_mark_inode_dirty(handle, inode, bh);
 	if (status < 0)
diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c
index 604fea3a26..530fba34f6 100644
--- a/fs/ocfs2/journal.c
+++ b/fs/ocfs2/journal.c
@@ -446,6 +446,23 @@ bail:
 }
 
 /*
+ * Make sure handle has at least 'nblocks' credits available. If it does not
+ * have that many credits available, we will try to extend the handle to have
+ * enough credits. If that fails, we will restart transaction to have enough
+ * credits. Similar notes regarding data consistency and locking implications
+ * as for ocfs2_extend_trans() apply here.
+ */
+int ocfs2_assure_trans_credits(handle_t *handle, int nblocks)
+{
+	int old_nblks = jbd2_handle_buffer_credits(handle);
+
+	trace_ocfs2_assure_trans_credits(old_nblks);
+	if (old_nblks >= nblocks)
+		return 0;
+	return ocfs2_extend_trans(handle, nblocks - old_nblks);
+}
+
+/*
  * If we have fewer than thresh credits, extend by OCFS2_MAX_TRANS_DATA.
  * If that fails, restart the transaction & regain write access for the
  * buffer head which is used for metadata modifications.
@@ -479,12 +496,6 @@ bail:
 	return status;
 }
 
-
-struct ocfs2_triggers {
-	struct jbd2_buffer_trigger_type	ot_triggers;
-	int				ot_offset;
-};
-
 static inline struct ocfs2_triggers *to_ocfs2_trigger(struct jbd2_buffer_trigger_type *triggers)
 {
 	return container_of(triggers, struct ocfs2_triggers, ot_triggers);
@@ -548,85 +559,76 @@ static void ocfs2_db_frozen_trigger(struct jbd2_buffer_trigger_type *triggers,
 static void ocfs2_abort_trigger(struct jbd2_buffer_trigger_type *triggers,
 				struct buffer_head *bh)
 {
+	struct ocfs2_triggers *ot = to_ocfs2_trigger(triggers);
+
 	mlog(ML_ERROR,
 	     "ocfs2_abort_trigger called by JBD2.  bh = 0x%lx, "
 	     "bh->b_blocknr = %llu\n",
 	     (unsigned long)bh,
 	     (unsigned long long)bh->b_blocknr);
 
-	ocfs2_error(bh->b_assoc_map->host->i_sb,
+	ocfs2_error(ot->sb,
 		    "JBD2 has aborted our journal, ocfs2 cannot continue\n");
 }
 
-static struct ocfs2_triggers di_triggers = {
-	.ot_triggers = {
-		.t_frozen = ocfs2_frozen_trigger,
-		.t_abort = ocfs2_abort_trigger,
-	},
-	.ot_offset	= offsetof(struct ocfs2_dinode, i_check),
-};
-
-static struct ocfs2_triggers eb_triggers = {
-	.ot_triggers = {
-		.t_frozen = ocfs2_frozen_trigger,
-		.t_abort = ocfs2_abort_trigger,
-	},
-	.ot_offset	= offsetof(struct ocfs2_extent_block, h_check),
-};
-
-static struct ocfs2_triggers rb_triggers = {
-	.ot_triggers = {
-		.t_frozen = ocfs2_frozen_trigger,
-		.t_abort = ocfs2_abort_trigger,
-	},
-	.ot_offset	= offsetof(struct ocfs2_refcount_block, rf_check),
-};
-
-static struct ocfs2_triggers gd_triggers = {
-	.ot_triggers = {
-		.t_frozen = ocfs2_frozen_trigger,
-		.t_abort = ocfs2_abort_trigger,
-	},
-	.ot_offset	= offsetof(struct ocfs2_group_desc, bg_check),
-};
-
-static struct ocfs2_triggers db_triggers = {
-	.ot_triggers = {
-		.t_frozen = ocfs2_db_frozen_trigger,
-		.t_abort = ocfs2_abort_trigger,
-	},
-};
+static void ocfs2_setup_csum_triggers(struct super_block *sb,
+				      enum ocfs2_journal_trigger_type type,
+				      struct ocfs2_triggers *ot)
+{
+	BUG_ON(type >= OCFS2_JOURNAL_TRIGGER_COUNT);
 
-static struct ocfs2_triggers xb_triggers = {
-	.ot_triggers = {
-		.t_frozen = ocfs2_frozen_trigger,
-		.t_abort = ocfs2_abort_trigger,
-	},
-	.ot_offset	= offsetof(struct ocfs2_xattr_block, xb_check),
-};
+	switch (type) {
+	case OCFS2_JTR_DI:
+		ot->ot_triggers.t_frozen = ocfs2_frozen_trigger;
+		ot->ot_offset = offsetof(struct ocfs2_dinode, i_check);
+		break;
+	case OCFS2_JTR_EB:
+		ot->ot_triggers.t_frozen = ocfs2_frozen_trigger;
+		ot->ot_offset = offsetof(struct ocfs2_extent_block, h_check);
+		break;
+	case OCFS2_JTR_RB:
+		ot->ot_triggers.t_frozen = ocfs2_frozen_trigger;
+		ot->ot_offset = offsetof(struct ocfs2_refcount_block, rf_check);
+		break;
+	case OCFS2_JTR_GD:
+		ot->ot_triggers.t_frozen = ocfs2_frozen_trigger;
+		ot->ot_offset = offsetof(struct ocfs2_group_desc, bg_check);
+		break;
+	case OCFS2_JTR_DB:
+		ot->ot_triggers.t_frozen = ocfs2_db_frozen_trigger;
+		break;
+	case OCFS2_JTR_XB:
+		ot->ot_triggers.t_frozen = ocfs2_frozen_trigger;
+		ot->ot_offset = offsetof(struct ocfs2_xattr_block, xb_check);
+		break;
+	case OCFS2_JTR_DQ:
+		ot->ot_triggers.t_frozen = ocfs2_dq_frozen_trigger;
+		break;
+	case OCFS2_JTR_DR:
+		ot->ot_triggers.t_frozen = ocfs2_frozen_trigger;
+		ot->ot_offset = offsetof(struct ocfs2_dx_root_block, dr_check);
+		break;
+	case OCFS2_JTR_DL:
+		ot->ot_triggers.t_frozen = ocfs2_frozen_trigger;
+		ot->ot_offset = offsetof(struct ocfs2_dx_leaf, dl_check);
+		break;
+	case OCFS2_JTR_NONE:
+		/* To make compiler happy... */
+		return;
+	}
 
-static struct ocfs2_triggers dq_triggers = {
-	.ot_triggers = {
-		.t_frozen = ocfs2_dq_frozen_trigger,
-		.t_abort = ocfs2_abort_trigger,
-	},
-};
+	ot->ot_triggers.t_abort = ocfs2_abort_trigger;
+	ot->sb = sb;
+}
 
-static struct ocfs2_triggers dr_triggers = {
-	.ot_triggers = {
-		.t_frozen = ocfs2_frozen_trigger,
-		.t_abort = ocfs2_abort_trigger,
-	},
-	.ot_offset	= offsetof(struct ocfs2_dx_root_block, dr_check),
-};
+void ocfs2_initialize_journal_triggers(struct super_block *sb,
+				       struct ocfs2_triggers triggers[])
+{
+	enum ocfs2_journal_trigger_type type;
 
-static struct ocfs2_triggers dl_triggers = {
-	.ot_triggers = {
-		.t_frozen = ocfs2_frozen_trigger,
-		.t_abort = ocfs2_abort_trigger,
-	},
-	.ot_offset	= offsetof(struct ocfs2_dx_leaf, dl_check),
-};
+	for (type = OCFS2_JTR_DI; type < OCFS2_JOURNAL_TRIGGER_COUNT; type++)
+		ocfs2_setup_csum_triggers(sb, type, &triggers[type]);
+}
 
 static int __ocfs2_journal_access(handle_t *handle,
 				  struct ocfs2_caching_info *ci,
@@ -708,56 +710,91 @@ static int __ocfs2_journal_access(handle_t *handle,
 int ocfs2_journal_access_di(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type)
 {
-	return __ocfs2_journal_access(handle, ci, bh, &di_triggers, type);
+	struct ocfs2_super *osb = OCFS2_SB(ocfs2_metadata_cache_get_super(ci));
+
+	return __ocfs2_journal_access(handle, ci, bh,
+				      &osb->s_journal_triggers[OCFS2_JTR_DI],
+				      type);
 }
 
 int ocfs2_journal_access_eb(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type)
 {
-	return __ocfs2_journal_access(handle, ci, bh, &eb_triggers, type);
+	struct ocfs2_super *osb = OCFS2_SB(ocfs2_metadata_cache_get_super(ci));
+
+	return __ocfs2_journal_access(handle, ci, bh,
+				      &osb->s_journal_triggers[OCFS2_JTR_EB],
+				      type);
 }
 
 int ocfs2_journal_access_rb(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type)
 {
-	return __ocfs2_journal_access(handle, ci, bh, &rb_triggers,
+	struct ocfs2_super *osb = OCFS2_SB(ocfs2_metadata_cache_get_super(ci));
+
+	return __ocfs2_journal_access(handle, ci, bh,
+				      &osb->s_journal_triggers[OCFS2_JTR_RB],
 				      type);
 }
 
 int ocfs2_journal_access_gd(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type)
 {
-	return __ocfs2_journal_access(handle, ci, bh, &gd_triggers, type);
+	struct ocfs2_super *osb = OCFS2_SB(ocfs2_metadata_cache_get_super(ci));
+
+	return __ocfs2_journal_access(handle, ci, bh,
+				     &osb->s_journal_triggers[OCFS2_JTR_GD],
+				     type);
 }
 
 int ocfs2_journal_access_db(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type)
 {
-	return __ocfs2_journal_access(handle, ci, bh, &db_triggers, type);
+	struct ocfs2_super *osb = OCFS2_SB(ocfs2_metadata_cache_get_super(ci));
+
+	return __ocfs2_journal_access(handle, ci, bh,
+				     &osb->s_journal_triggers[OCFS2_JTR_DB],
+				     type);
 }
 
 int ocfs2_journal_access_xb(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type)
 {
-	return __ocfs2_journal_access(handle, ci, bh, &xb_triggers, type);
+	struct ocfs2_super *osb = OCFS2_SB(ocfs2_metadata_cache_get_super(ci));
+
+	return __ocfs2_journal_access(handle, ci, bh,
+				     &osb->s_journal_triggers[OCFS2_JTR_XB],
+				     type);
 }
 
 int ocfs2_journal_access_dq(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type)
 {
-	return __ocfs2_journal_access(handle, ci, bh, &dq_triggers, type);
+	struct ocfs2_super *osb = OCFS2_SB(ocfs2_metadata_cache_get_super(ci));
+
+	return __ocfs2_journal_access(handle, ci, bh,
+				     &osb->s_journal_triggers[OCFS2_JTR_DQ],
+				     type);
 }
 
 int ocfs2_journal_access_dr(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type)
 {
-	return __ocfs2_journal_access(handle, ci, bh, &dr_triggers, type);
+	struct ocfs2_super *osb = OCFS2_SB(ocfs2_metadata_cache_get_super(ci));
+
+	return __ocfs2_journal_access(handle, ci, bh,
+				     &osb->s_journal_triggers[OCFS2_JTR_DR],
+				     type);
 }
 
 int ocfs2_journal_access_dl(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type)
 {
-	return __ocfs2_journal_access(handle, ci, bh, &dl_triggers, type);
+	struct ocfs2_super *osb = OCFS2_SB(ocfs2_metadata_cache_get_super(ci));
+
+	return __ocfs2_journal_access(handle, ci, bh,
+				     &osb->s_journal_triggers[OCFS2_JTR_DL],
+				     type);
 }
 
 int ocfs2_journal_access(handle_t *handle, struct ocfs2_caching_info *ci,
@@ -778,13 +815,15 @@ void ocfs2_journal_dirty(handle_t *handle, struct buffer_head *bh)
 		if (!is_handle_aborted(handle)) {
 			journal_t *journal = handle->h_transaction->t_journal;
 
-			mlog(ML_ERROR, "jbd2_journal_dirty_metadata failed. "
-					"Aborting transaction and journal.\n");
+			mlog(ML_ERROR, "jbd2_journal_dirty_metadata failed: "
+			     "handle type %u started at line %u, credits %u/%u "
+			     "errcode %d. Aborting transaction and journal.\n",
+			     handle->h_type, handle->h_line_no,
+			     handle->h_requested_credits,
+			     jbd2_handle_buffer_credits(handle), status);
 			handle->h_err = status;
 			jbd2_journal_abort_handle(handle);
 			jbd2_journal_abort(journal, status);
-			ocfs2_abort(bh->b_assoc_map->host->i_sb,
-				    "Journal already aborted.\n");
 		}
 	}
 }
diff --git a/fs/ocfs2/journal.h b/fs/ocfs2/journal.h
index 41c9fe7e62..e3c3a35dc5 100644
--- a/fs/ocfs2/journal.h
+++ b/fs/ocfs2/journal.h
@@ -243,6 +243,8 @@ handle_t		    *ocfs2_start_trans(struct ocfs2_super *osb,
 int			     ocfs2_commit_trans(struct ocfs2_super *osb,
 						handle_t *handle);
 int			     ocfs2_extend_trans(handle_t *handle, int nblocks);
+int			     ocfs2_assure_trans_credits(handle_t *handle,
+						int nblocks);
 int			     ocfs2_allocate_extend_trans(handle_t *handle,
 						int thresh);
 
diff --git a/fs/ocfs2/localalloc.c b/fs/ocfs2/localalloc.c
index c803c10dd9..5df34561c5 100644
--- a/fs/ocfs2/localalloc.c
+++ b/fs/ocfs2/localalloc.c
@@ -212,14 +212,15 @@ static inline int ocfs2_la_state_enabled(struct ocfs2_super *osb)
 void ocfs2_local_alloc_seen_free_bits(struct ocfs2_super *osb,
 				      unsigned int num_clusters)
 {
-	spin_lock(&osb->osb_lock);
-	if (osb->local_alloc_state == OCFS2_LA_DISABLED ||
-	    osb->local_alloc_state == OCFS2_LA_THROTTLED)
-		if (num_clusters >= osb->local_alloc_default_bits) {
+	if (num_clusters >= osb->local_alloc_default_bits) {
+		spin_lock(&osb->osb_lock);
+		if (osb->local_alloc_state == OCFS2_LA_DISABLED ||
+		    osb->local_alloc_state == OCFS2_LA_THROTTLED) {
 			cancel_delayed_work(&osb->la_enable_wq);
 			osb->local_alloc_state = OCFS2_LA_ENABLED;
 		}
-	spin_unlock(&osb->osb_lock);
+		spin_unlock(&osb->osb_lock);
+	}
 }
 
 void ocfs2_la_enable_worker(struct work_struct *work)
@@ -335,7 +336,7 @@ int ocfs2_load_local_alloc(struct ocfs2_super *osb)
 		     "found = %u, set = %u, taken = %u, off = %u\n",
 		     num_used, le32_to_cpu(alloc->id1.bitmap1.i_used),
 		     le32_to_cpu(alloc->id1.bitmap1.i_total),
-		     OCFS2_LOCAL_ALLOC(alloc)->la_bm_off);
+		     le32_to_cpu(OCFS2_LOCAL_ALLOC(alloc)->la_bm_off));
 
 		status = -EINVAL;
 		goto bail;
@@ -863,14 +864,8 @@ static int ocfs2_local_alloc_find_clear_bits(struct ocfs2_super *osb,
 
 	numfound = bitoff = startoff = 0;
 	left = le32_to_cpu(alloc->id1.bitmap1.i_total);
-	while ((bitoff = ocfs2_find_next_zero_bit(bitmap, left, startoff)) != -1) {
-		if (bitoff == left) {
-			/* mlog(0, "bitoff (%d) == left", bitoff); */
-			break;
-		}
-		/* mlog(0, "Found a zero: bitoff = %d, startoff = %d, "
-		   "numfound = %d\n", bitoff, startoff, numfound);*/
-
+	while ((bitoff = ocfs2_find_next_zero_bit(bitmap, left, startoff)) <
+	       left) {
 		/* Ok, we found a zero bit... is it contig. or do we
 		 * start over?*/
 		if (bitoff == startoff) {
@@ -976,9 +971,9 @@ static int ocfs2_sync_local_to_main(struct ocfs2_super *osb,
 	start = count = 0;
 	left = le32_to_cpu(alloc->id1.bitmap1.i_total);
 
-	while ((bit_off = ocfs2_find_next_zero_bit(bitmap, left, start))
-	       != -1) {
-		if ((bit_off < left) && (bit_off == start)) {
+	while ((bit_off = ocfs2_find_next_zero_bit(bitmap, left, start)) <
+	       left) {
+		if (bit_off == start) {
 			count++;
 			start++;
 			continue;
@@ -1002,8 +997,7 @@ static int ocfs2_sync_local_to_main(struct ocfs2_super *osb,
 				goto bail;
 			}
 		}
-		if (bit_off >= left)
-			break;
+
 		count = 1;
 		start = bit_off + 1;
 	}
@@ -1220,7 +1214,7 @@ retry_enospc:
 			     OCFS2_LOCAL_ALLOC(alloc)->la_bitmap);
 
 	trace_ocfs2_local_alloc_new_window_result(
-		OCFS2_LOCAL_ALLOC(alloc)->la_bm_off,
+		le32_to_cpu(OCFS2_LOCAL_ALLOC(alloc)->la_bm_off),
 		le32_to_cpu(alloc->id1.bitmap1.i_total));
 
 bail:
diff --git a/fs/ocfs2/locks.c b/fs/ocfs2/locks.c
index f37174e79f..6de944818c 100644
--- a/fs/ocfs2/locks.c
+++ b/fs/ocfs2/locks.c
@@ -27,7 +27,7 @@ static int ocfs2_do_flock(struct file *file, struct inode *inode,
 	struct ocfs2_file_private *fp = file->private_data;
 	struct ocfs2_lock_res *lockres = &fp->fp_flock;
 
-	if (fl->fl_type == F_WRLCK)
+	if (lock_is_write(fl))
 		level = 1;
 	if (!IS_SETLKW(cmd))
 		trylock = 1;
@@ -53,8 +53,8 @@ static int ocfs2_do_flock(struct file *file, struct inode *inode,
 		 */
 
 		locks_init_lock(&request);
-		request.fl_type = F_UNLCK;
-		request.fl_flags = FL_FLOCK;
+		request.c.flc_type = F_UNLCK;
+		request.c.flc_flags = FL_FLOCK;
 		locks_lock_file_wait(file, &request);
 
 		ocfs2_file_unlock(file);
@@ -100,14 +100,14 @@ int ocfs2_flock(struct file *file, int cmd, struct file_lock *fl)
 	struct inode *inode = file->f_mapping->host;
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 
-	if (!(fl->fl_flags & FL_FLOCK))
+	if (!(fl->c.flc_flags & FL_FLOCK))
 		return -ENOLCK;
 
 	if ((osb->s_mount_opt & OCFS2_MOUNT_LOCALFLOCKS) ||
 	    ocfs2_mount_local(osb))
 		return locks_lock_file_wait(file, fl);
 
-	if (fl->fl_type == F_UNLCK)
+	if (lock_is_unlock(fl))
 		return ocfs2_do_funlock(file, cmd, fl);
 	else
 		return ocfs2_do_flock(file, inode, cmd, fl);
@@ -118,7 +118,7 @@ int ocfs2_lock(struct file *file, int cmd, struct file_lock *fl)
 	struct inode *inode = file->f_mapping->host;
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 
-	if (!(fl->fl_flags & FL_POSIX))
+	if (!(fl->c.flc_flags & FL_POSIX))
 		return -ENOLCK;
 
 	return ocfs2_plock(osb->cconn, OCFS2_I(inode)->ip_blkno, file, cmd, fl);
diff --git a/fs/ocfs2/move_extents.c b/fs/ocfs2/move_extents.c
index 1f9ed117e7..f9d6a4f9ca 100644
--- a/fs/ocfs2/move_extents.c
+++ b/fs/ocfs2/move_extents.c
@@ -685,7 +685,7 @@ static int ocfs2_move_extent(struct ocfs2_move_extents_context *context,
 	}
 
 	ret = ocfs2_block_group_set_bits(handle, gb_inode, gd, gd_bh,
-					 goal_bit, len);
+					 goal_bit, len, 0, 0);
 	if (ret) {
 		ocfs2_rollback_alloc_dinode_counts(gb_inode, gb_bh, len,
 					       le16_to_cpu(gd->bg_chain));
diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c
index 9221a33f91..4d1ea8703f 100644
--- a/fs/ocfs2/namei.c
+++ b/fs/ocfs2/namei.c
@@ -566,7 +566,7 @@ static int __ocfs2_mknod_locked(struct inode *dir,
 	fe->i_last_eb_blk = 0;
 	strcpy(fe->i_signature, OCFS2_INODE_SIGNATURE);
 	fe->i_flags |= cpu_to_le32(OCFS2_VALID_FL);
-	ktime_get_real_ts64(&ts);
+	ktime_get_coarse_real_ts64(&ts);
 	fe->i_atime = fe->i_ctime = fe->i_mtime =
 		cpu_to_le64(ts.tv_sec);
 	fe->i_mtime_nsec = fe->i_ctime_nsec = fe->i_atime_nsec =
@@ -797,6 +797,7 @@ static int ocfs2_link(struct dentry *old_dentry,
 	ocfs2_set_links_count(fe, inode->i_nlink);
 	fe->i_ctime = cpu_to_le64(inode_get_ctime_sec(inode));
 	fe->i_ctime_nsec = cpu_to_le32(inode_get_ctime_nsec(inode));
+	ocfs2_update_inode_fsync_trans(handle, inode, 0);
 	ocfs2_journal_dirty(handle, fe_bh);
 
 	err = ocfs2_add_entry(handle, dentry, inode,
@@ -993,6 +994,7 @@ static int ocfs2_unlink(struct inode *dir,
 		drop_nlink(inode);
 	drop_nlink(inode);
 	ocfs2_set_links_count(fe, inode->i_nlink);
+	ocfs2_update_inode_fsync_trans(handle, inode, 0);
 	ocfs2_journal_dirty(handle, fe_bh);
 
 	inode_set_mtime_to_ts(dir, inode_set_ctime_current(dir));
diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h
index a503c553ba..8fe826143d 100644
--- a/fs/ocfs2/ocfs2.h
+++ b/fs/ocfs2/ocfs2.h
@@ -284,6 +284,30 @@ enum ocfs2_mount_options
 #define OCFS2_OSB_ERROR_FS	0x0004
 #define OCFS2_DEFAULT_ATIME_QUANTUM	60
 
+struct ocfs2_triggers {
+	struct jbd2_buffer_trigger_type	ot_triggers;
+	int				ot_offset;
+	struct super_block		*sb;
+};
+
+enum ocfs2_journal_trigger_type {
+	OCFS2_JTR_DI,
+	OCFS2_JTR_EB,
+	OCFS2_JTR_RB,
+	OCFS2_JTR_GD,
+	OCFS2_JTR_DB,
+	OCFS2_JTR_XB,
+	OCFS2_JTR_DQ,
+	OCFS2_JTR_DR,
+	OCFS2_JTR_DL,
+	OCFS2_JTR_NONE  /* This must be the last entry */
+};
+
+#define OCFS2_JOURNAL_TRIGGER_COUNT OCFS2_JTR_NONE
+
+void ocfs2_initialize_journal_triggers(struct super_block *sb,
+				       struct ocfs2_triggers triggers[]);
+
 struct ocfs2_journal;
 struct ocfs2_slot_info;
 struct ocfs2_recovery_map;
@@ -351,6 +375,9 @@ struct ocfs2_super
 	struct ocfs2_journal *journal;
 	unsigned long osb_commit_interval;
 
+	/* Journal triggers for checksum */
+	struct ocfs2_triggers s_journal_triggers[OCFS2_JOURNAL_TRIGGER_COUNT];
+
 	struct delayed_work		la_enable_wq;
 
 	/*
diff --git a/fs/ocfs2/ocfs2_fs.h b/fs/ocfs2/ocfs2_fs.h
index 7aebdbf5cc..c93689b568 100644
--- a/fs/ocfs2/ocfs2_fs.h
+++ b/fs/ocfs2/ocfs2_fs.h
@@ -883,7 +883,8 @@ struct ocfs2_group_desc
 	__le16	bg_free_bits_count;     /* Free bits count */
 	__le16   bg_chain;               /* What chain I am in. */
 /*10*/	__le32   bg_generation;
-	__le32	bg_reserved1;
+	__le16   bg_contig_free_bits;   /* max contig free bits length */
+	__le16   bg_reserved1;
 	__le64   bg_next_group;          /* Next group in my list, in
 					   blocks */
 /*20*/	__le64   bg_parent_dinode;       /* dinode which owns me, in
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index 9898c11bdf..0511c69c9f 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -82,7 +82,7 @@ DECLARE_EVENT_CLASS(ocfs2__string,
 		__string(name,name)
 	),
 	TP_fast_assign(
-		__assign_str(name, name);
+		__assign_str(name);
 	),
 	TP_printk("%s", __get_str(name))
 );
@@ -1289,7 +1289,7 @@ DECLARE_EVENT_CLASS(ocfs2__file_ops,
 		__entry->dentry = dentry;
 		__entry->ino = ino;
 		__entry->d_len = d_len;
-		__assign_str(d_name, d_name);
+		__assign_str(d_name);
 		__entry->para = para;
 	),
 	TP_printk("%p %p %p %llu %llu %.*s", __entry->inode, __entry->file,
@@ -1425,7 +1425,7 @@ TRACE_EVENT(ocfs2_setattr,
 		__entry->dentry = dentry;
 		__entry->ino = ino;
 		__entry->d_len = d_len;
-		__assign_str(d_name, d_name);
+		__assign_str(d_name);
 		__entry->ia_valid = ia_valid;
 		__entry->ia_mode = ia_mode;
 		__entry->ia_uid = ia_uid;
@@ -1683,7 +1683,7 @@ TRACE_EVENT(ocfs2_parse_options,
 	),
 	TP_fast_assign(
 		__entry->is_remount = is_remount;
-		__assign_str(options, options);
+		__assign_str(options);
 	),
 	TP_printk("%d %s", __entry->is_remount, __get_str(options))
 );
@@ -1718,8 +1718,8 @@ TRACE_EVENT(ocfs2_initialize_super,
 		__field(int, cluster_bits)
 	),
 	TP_fast_assign(
-		__assign_str(label, label);
-		__assign_str(uuid_str, uuid_str);
+		__assign_str(label);
+		__assign_str(uuid_str);
 		__entry->root_dir = root_dir;
 		__entry->system_dir = system_dir;
 		__entry->cluster_bits = cluster_bits;
@@ -1746,7 +1746,7 @@ TRACE_EVENT(ocfs2_init_xattr_set_ctxt,
 		__field(int, credits)
 	),
 	TP_fast_assign(
-		__assign_str(name, name);
+		__assign_str(name);
 		__entry->meta = meta;
 		__entry->clusters = clusters;
 		__entry->credits = credits;
@@ -1770,7 +1770,7 @@ DECLARE_EVENT_CLASS(ocfs2__xattr_find,
 	),
 	TP_fast_assign(
 		__entry->ino = ino;
-		__assign_str(name, name);
+		__assign_str(name);
 		__entry->name_index = name_index;
 		__entry->hash = hash;
 		__entry->location = location;
@@ -2019,7 +2019,7 @@ TRACE_EVENT(ocfs2_sync_dquot_helper,
 		__entry->dq_id = dq_id;
 		__entry->dq_type = dq_type;
 		__entry->type = type;
-		__assign_str(s_id, s_id);
+		__assign_str(s_id);
 	),
 	TP_printk("%u %u %lu %s", __entry->dq_id, __entry->dq_type,
 		  __entry->type, __get_str(s_id))
@@ -2060,7 +2060,7 @@ TRACE_EVENT(ocfs2_dx_dir_search,
 	TP_fast_assign(
 		__entry->ino = ino;
 		__entry->namelen = namelen;
-		__assign_str(name, name);
+		__assign_str(name);
 		__entry->major_hash = major_hash;
 		__entry->minor_hash = minor_hash;
 		__entry->blkno = blkno;
@@ -2088,7 +2088,7 @@ TRACE_EVENT(ocfs2_find_files_on_disk,
 	),
 	TP_fast_assign(
 		__entry->namelen = namelen;
-		__assign_str(name, name);
+		__assign_str(name);
 		__entry->blkno = blkno;
 		__entry->dir = dir;
 	),
@@ -2107,7 +2107,7 @@ TRACE_EVENT(ocfs2_check_dir_for_entry,
 	TP_fast_assign(
 		__entry->dir = dir;
 		__entry->namelen = namelen;
-		__assign_str(name, name);
+		__assign_str(name);
 	),
 	TP_printk("%llu %.*s", __entry->dir,
 		  __entry->namelen, __get_str(name))
@@ -2135,7 +2135,7 @@ TRACE_EVENT(ocfs2_dx_dir_index_root_block,
 		__entry->major_hash = major_hash;
 		__entry->minor_hash = minor_hash;
 		__entry->namelen = namelen;
-		__assign_str(name, name);
+		__assign_str(name);
 		__entry->num_used = num_used;
 	),
 	TP_printk("%llu %x %x %.*s %u", __entry->dir,
@@ -2171,7 +2171,7 @@ DECLARE_EVENT_CLASS(ocfs2__dentry_ops,
 		__entry->dir = dir;
 		__entry->dentry = dentry;
 		__entry->name_len = name_len;
-		__assign_str(name, name);
+		__assign_str(name);
 		__entry->dir_blkno = dir_blkno;
 		__entry->extra = extra;
 	),
@@ -2217,7 +2217,7 @@ TRACE_EVENT(ocfs2_mknod,
 		__entry->dir = dir;
 		__entry->dentry = dentry;
 		__entry->name_len = name_len;
-		__assign_str(name, name);
+		__assign_str(name);
 		__entry->dir_blkno = dir_blkno;
 		__entry->dev = dev;
 		__entry->mode = mode;
@@ -2241,9 +2241,9 @@ TRACE_EVENT(ocfs2_link,
 	TP_fast_assign(
 		__entry->ino = ino;
 		__entry->old_len = old_len;
-		__assign_str(old_name, old_name);
+		__assign_str(old_name);
 		__entry->name_len = name_len;
-		__assign_str(name, name);
+		__assign_str(name);
 	),
 	TP_printk("%llu %.*s %.*s", __entry->ino,
 		  __entry->old_len, __get_str(old_name),
@@ -2279,9 +2279,9 @@ TRACE_EVENT(ocfs2_rename,
 		__entry->new_dir = new_dir;
 		__entry->new_dentry = new_dentry;
 		__entry->old_len = old_len;
-		__assign_str(old_name, old_name);
+		__assign_str(old_name);
 		__entry->new_len = new_len;
-		__assign_str(new_name, new_name);
+		__assign_str(new_name);
 	),
 	TP_printk("%p %p %p %p %.*s %.*s",
 		  __entry->old_dir, __entry->old_dentry,
@@ -2301,7 +2301,7 @@ TRACE_EVENT(ocfs2_rename_target_exists,
 	),
 	TP_fast_assign(
 		__entry->new_len = new_len;
-		__assign_str(new_name, new_name);
+		__assign_str(new_name);
 	),
 	TP_printk("%.*s", __entry->new_len, __get_str(new_name))
 );
@@ -2344,7 +2344,7 @@ TRACE_EVENT(ocfs2_symlink_begin,
 		__entry->dentry = dentry;
 		__entry->symname = symname;
 		__entry->len = len;
-		__assign_str(name, name);
+		__assign_str(name);
 	),
 	TP_printk("%p %p %s %.*s", __entry->dir, __entry->dentry,
 		  __entry->symname, __entry->len, __get_str(name))
@@ -2360,7 +2360,7 @@ TRACE_EVENT(ocfs2_blkno_stringify,
 	),
 	TP_fast_assign(
 		__entry->blkno = blkno;
-		__assign_str(name, name);
+		__assign_str(name);
 		__entry->namelen = namelen;
 	),
 	TP_printk("%llu %s %d", __entry->blkno, __get_str(name),
@@ -2381,7 +2381,7 @@ TRACE_EVENT(ocfs2_orphan_del,
 	),
 	TP_fast_assign(
 		__entry->dir = dir;
-		__assign_str(name, name);
+		__assign_str(name);
 		__entry->namelen = namelen;
 	),
 	TP_printk("%llu %s %d", __entry->dir, __get_str(name),
@@ -2403,7 +2403,7 @@ TRACE_EVENT(ocfs2_dentry_revalidate,
 	TP_fast_assign(
 		__entry->dentry = dentry;
 		__entry->len = len;
-		__assign_str(name, name);
+		__assign_str(name);
 	),
 	TP_printk("%p %.*s", __entry->dentry, __entry->len, __get_str(name))
 );
@@ -2420,7 +2420,7 @@ TRACE_EVENT(ocfs2_dentry_revalidate_negative,
 	),
 	TP_fast_assign(
 		__entry->len = len;
-		__assign_str(name, name);
+		__assign_str(name);
 		__entry->pgen = pgen;
 		__entry->gen = gen;
 	),
@@ -2445,7 +2445,7 @@ TRACE_EVENT(ocfs2_find_local_alias,
 	),
 	TP_fast_assign(
 		__entry->len = len;
-		__assign_str(name, name);
+		__assign_str(name);
 	),
 	TP_printk("%.*s", __entry->len, __get_str(name))
 );
@@ -2462,7 +2462,7 @@ TRACE_EVENT(ocfs2_dentry_attach_lock,
 	),
 	TP_fast_assign(
 		__entry->len = len;
-		__assign_str(name, name);
+		__assign_str(name);
 		__entry->parent = parent;
 		__entry->fsdata = fsdata;
 	),
@@ -2480,7 +2480,7 @@ TRACE_EVENT(ocfs2_dentry_attach_lock_found,
 		__field(unsigned long long, ino)
 	),
 	TP_fast_assign(
-		__assign_str(name, name);
+		__assign_str(name);
 		__entry->parent = parent;
 		__entry->ino = ino;
 	),
@@ -2527,7 +2527,7 @@ TRACE_EVENT(ocfs2_get_parent,
 	TP_fast_assign(
 		__entry->child = child;
 		__entry->len = len;
-		__assign_str(name, name);
+		__assign_str(name);
 		__entry->ino = ino;
 	),
 	TP_printk("%p %.*s %llu", __entry->child, __entry->len,
@@ -2551,7 +2551,7 @@ TRACE_EVENT(ocfs2_encode_fh_begin,
 	TP_fast_assign(
 		__entry->dentry = dentry;
 		__entry->name_len = name_len;
-		__assign_str(name, name);
+		__assign_str(name);
 		__entry->fh = fh;
 		__entry->len = len;
 		__entry->connectable = connectable;
@@ -2577,6 +2577,8 @@ DEFINE_OCFS2_ULL_UINT_EVENT(ocfs2_commit_cache_end);
 
 DEFINE_OCFS2_INT_INT_EVENT(ocfs2_extend_trans);
 
+DEFINE_OCFS2_INT_EVENT(ocfs2_assure_trans_credits);
+
 DEFINE_OCFS2_INT_EVENT(ocfs2_extend_trans_restart);
 
 DEFINE_OCFS2_INT_INT_EVENT(ocfs2_allocate_extend_trans);
diff --git a/fs/ocfs2/quota_global.c b/fs/ocfs2/quota_global.c
index dc9f76ab7e..0575c2d060 100644
--- a/fs/ocfs2/quota_global.c
+++ b/fs/ocfs2/quota_global.c
@@ -447,14 +447,17 @@ int ocfs2_global_write_info(struct super_block *sb, int type)
 	int err;
 	struct quota_info *dqopt = sb_dqopt(sb);
 	struct ocfs2_mem_dqinfo *info = dqopt->info[type].dqi_priv;
+	unsigned int memalloc;
 
 	down_write(&dqopt->dqio_sem);
+	memalloc = memalloc_nofs_save();
 	err = ocfs2_qinfo_lock(info, 1);
 	if (err < 0)
 		goto out_sem;
 	err = __ocfs2_global_write_info(sb, type);
 	ocfs2_qinfo_unlock(info, 1);
 out_sem:
+	memalloc_nofs_restore(memalloc);
 	up_write(&dqopt->dqio_sem);
 	return err;
 }
@@ -601,6 +604,7 @@ static int ocfs2_sync_dquot_helper(struct dquot *dquot, unsigned long type)
 	struct ocfs2_mem_dqinfo *oinfo = sb_dqinfo(sb, type)->dqi_priv;
 	struct ocfs2_super *osb = OCFS2_SB(sb);
 	int status = 0;
+	unsigned int memalloc;
 
 	trace_ocfs2_sync_dquot_helper(from_kqid(&init_user_ns, dquot->dq_id),
 				      dquot->dq_id.type,
@@ -618,6 +622,7 @@ static int ocfs2_sync_dquot_helper(struct dquot *dquot, unsigned long type)
 		goto out_ilock;
 	}
 	down_write(&sb_dqopt(sb)->dqio_sem);
+	memalloc = memalloc_nofs_save();
 	status = ocfs2_sync_dquot(dquot);
 	if (status < 0)
 		mlog_errno(status);
@@ -625,6 +630,7 @@ static int ocfs2_sync_dquot_helper(struct dquot *dquot, unsigned long type)
 	status = ocfs2_local_write_dquot(dquot);
 	if (status < 0)
 		mlog_errno(status);
+	memalloc_nofs_restore(memalloc);
 	up_write(&sb_dqopt(sb)->dqio_sem);
 	ocfs2_commit_trans(osb, handle);
 out_ilock:
@@ -662,6 +668,7 @@ static int ocfs2_write_dquot(struct dquot *dquot)
 	handle_t *handle;
 	struct ocfs2_super *osb = OCFS2_SB(dquot->dq_sb);
 	int status = 0;
+	unsigned int memalloc;
 
 	trace_ocfs2_write_dquot(from_kqid(&init_user_ns, dquot->dq_id),
 				dquot->dq_id.type);
@@ -673,7 +680,9 @@ static int ocfs2_write_dquot(struct dquot *dquot)
 		goto out;
 	}
 	down_write(&sb_dqopt(dquot->dq_sb)->dqio_sem);
+	memalloc = memalloc_nofs_save();
 	status = ocfs2_local_write_dquot(dquot);
+	memalloc_nofs_restore(memalloc);
 	up_write(&sb_dqopt(dquot->dq_sb)->dqio_sem);
 	ocfs2_commit_trans(osb, handle);
 out:
@@ -920,6 +929,7 @@ static int ocfs2_mark_dquot_dirty(struct dquot *dquot)
 	struct ocfs2_mem_dqinfo *oinfo = sb_dqinfo(sb, type)->dqi_priv;
 	handle_t *handle;
 	struct ocfs2_super *osb = OCFS2_SB(sb);
+	unsigned int memalloc;
 
 	trace_ocfs2_mark_dquot_dirty(from_kqid(&init_user_ns, dquot->dq_id),
 				     type);
@@ -946,6 +956,7 @@ static int ocfs2_mark_dquot_dirty(struct dquot *dquot)
 		goto out_ilock;
 	}
 	down_write(&sb_dqopt(sb)->dqio_sem);
+	memalloc = memalloc_nofs_save();
 	status = ocfs2_sync_dquot(dquot);
 	if (status < 0) {
 		mlog_errno(status);
@@ -954,6 +965,7 @@ static int ocfs2_mark_dquot_dirty(struct dquot *dquot)
 	/* Now write updated local dquot structure */
 	status = ocfs2_local_write_dquot(dquot);
 out_dlock:
+	memalloc_nofs_restore(memalloc);
 	up_write(&sb_dqopt(sb)->dqio_sem);
 	ocfs2_commit_trans(osb, handle);
 out_ilock:
diff --git a/fs/ocfs2/quota_local.c b/fs/ocfs2/quota_local.c
index e09842fc9d..8ce462c64c 100644
--- a/fs/ocfs2/quota_local.c
+++ b/fs/ocfs2/quota_local.c
@@ -470,6 +470,7 @@ static int ocfs2_recover_local_quota_file(struct inode *lqinode,
 	int bit, chunk;
 	struct ocfs2_recovery_chunk *rchunk, *next;
 	qsize_t spacechange, inodechange;
+	unsigned int memalloc;
 
 	trace_ocfs2_recover_local_quota_file((unsigned long)lqinode->i_ino, type);
 
@@ -521,6 +522,7 @@ static int ocfs2_recover_local_quota_file(struct inode *lqinode,
 				goto out_drop_lock;
 			}
 			down_write(&sb_dqopt(sb)->dqio_sem);
+			memalloc = memalloc_nofs_save();
 			spin_lock(&dquot->dq_dqb_lock);
 			/* Add usage from quota entry into quota changes
 			 * of our node. Auxiliary variables are important
@@ -553,6 +555,7 @@ static int ocfs2_recover_local_quota_file(struct inode *lqinode,
 			unlock_buffer(qbh);
 			ocfs2_journal_dirty(handle, qbh);
 out_commit:
+			memalloc_nofs_restore(memalloc);
 			up_write(&sb_dqopt(sb)->dqio_sem);
 			ocfs2_commit_trans(OCFS2_SB(sb), handle);
 out_drop_lock:
diff --git a/fs/ocfs2/refcounttree.c b/fs/ocfs2/refcounttree.c
index 3f80a56d0d..1f303b1adf 100644
--- a/fs/ocfs2/refcounttree.c
+++ b/fs/ocfs2/refcounttree.c
@@ -630,7 +630,7 @@ static int ocfs2_create_refcount_tree(struct inode *inode,
 	rb->rf_records.rl_count =
 			cpu_to_le16(ocfs2_refcount_recs_per_rb(osb->sb));
 	spin_lock(&osb->osb_lock);
-	rb->rf_generation = osb->s_next_generation++;
+	rb->rf_generation = cpu_to_le32(osb->s_next_generation++);
 	spin_unlock(&osb->osb_lock);
 
 	ocfs2_journal_dirty(handle, new_bh);
diff --git a/fs/ocfs2/reservations.c b/fs/ocfs2/reservations.c
index a9d1296d73..1fe61974d9 100644
--- a/fs/ocfs2/reservations.c
+++ b/fs/ocfs2/reservations.c
@@ -414,7 +414,7 @@ static int ocfs2_resmap_find_free_bits(struct ocfs2_reservation_map *resmap,
 
 	start = search_start;
 	while ((offset = ocfs2_find_next_zero_bit(bitmap, resmap->m_bitmap_len,
-						 start)) != -1) {
+					start)) < resmap->m_bitmap_len) {
 		/* Search reached end of the region */
 		if (offset >= (search_start + search_len))
 			break;
diff --git a/fs/ocfs2/resize.c b/fs/ocfs2/resize.c
index d65d43c618..c4a4016d38 100644
--- a/fs/ocfs2/resize.c
+++ b/fs/ocfs2/resize.c
@@ -91,6 +91,8 @@ static int ocfs2_update_last_group_and_inode(handle_t *handle,
 	u16 cl_bpc = le16_to_cpu(cl->cl_bpc);
 	u16 cl_cpg = le16_to_cpu(cl->cl_cpg);
 	u16 old_bg_clusters;
+	u16 contig_bits;
+	__le16 old_bg_contig_free_bits;
 
 	trace_ocfs2_update_last_group_and_inode(new_clusters,
 						first_new_cluster);
@@ -122,6 +124,11 @@ static int ocfs2_update_last_group_and_inode(handle_t *handle,
 		le16_add_cpu(&group->bg_free_bits_count, -1 * backups);
 	}
 
+	contig_bits = ocfs2_find_max_contig_free_bits(group->bg_bitmap,
+					le16_to_cpu(group->bg_bits), 0);
+	old_bg_contig_free_bits = group->bg_contig_free_bits;
+	group->bg_contig_free_bits = cpu_to_le16(contig_bits);
+
 	ocfs2_journal_dirty(handle, group_bh);
 
 	/* update the inode accordingly. */
@@ -160,6 +167,7 @@ out_rollback:
 		le16_add_cpu(&group->bg_free_bits_count, backups);
 		le16_add_cpu(&group->bg_bits, -1 * num_bits);
 		le16_add_cpu(&group->bg_free_bits_count, -1 * num_bits);
+		group->bg_contig_free_bits = old_bg_contig_free_bits;
 	}
 out:
 	if (ret)
diff --git a/fs/ocfs2/stack_user.c b/fs/ocfs2/stack_user.c
index 9b76ee66ae..c11406cd87 100644
--- a/fs/ocfs2/stack_user.c
+++ b/fs/ocfs2/stack_user.c
@@ -744,7 +744,7 @@ static int user_plock(struct ocfs2_cluster_connection *conn,
 		return dlm_posix_cancel(conn->cc_lockspace, ino, file, fl);
 	else if (IS_GETLK(cmd))
 		return dlm_posix_get(conn->cc_lockspace, ino, file, fl);
-	else if (fl->fl_type == F_UNLCK)
+	else if (lock_is_unlock(fl))
 		return dlm_posix_unlock(conn->cc_lockspace, ino, file, fl);
 	else
 		return dlm_posix_lock(conn->cc_lockspace, ino, file, cmd, fl);
diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c
index 166c8918c8..f7b483f0de 100644
--- a/fs/ocfs2/suballoc.c
+++ b/fs/ocfs2/suballoc.c
@@ -50,6 +50,10 @@ struct ocfs2_suballoc_result {
 	u64		sr_blkno;	/* The first allocated block */
 	unsigned int	sr_bit_offset;	/* The bit in the bg */
 	unsigned int	sr_bits;	/* How many bits we claimed */
+	unsigned int	sr_max_contig_bits; /* The length for contiguous
+					     * free bits, only available
+					     * for cluster group
+					     */
 };
 
 static u64 ocfs2_group_from_res(struct ocfs2_suballoc_result *res)
@@ -1272,6 +1276,26 @@ static int ocfs2_test_bg_bit_allocatable(struct buffer_head *bg_bh,
 	return ret;
 }
 
+u16 ocfs2_find_max_contig_free_bits(void *bitmap,
+			 u16 total_bits, u16 start)
+{
+	u16 offset, free_bits;
+	u16 contig_bits = 0;
+
+	while (start < total_bits) {
+		offset = ocfs2_find_next_zero_bit(bitmap, total_bits, start);
+		if (offset == total_bits)
+			break;
+
+		start = ocfs2_find_next_bit(bitmap, total_bits, offset);
+		free_bits = start - offset;
+		if (contig_bits < free_bits)
+			contig_bits = free_bits;
+	}
+
+	return contig_bits;
+}
+
 static int ocfs2_block_group_find_clear_bits(struct ocfs2_super *osb,
 					     struct buffer_head *bg_bh,
 					     unsigned int bits_wanted,
@@ -1280,6 +1304,7 @@ static int ocfs2_block_group_find_clear_bits(struct ocfs2_super *osb,
 {
 	void *bitmap;
 	u16 best_offset, best_size;
+	u16 prev_best_size = 0;
 	int offset, start, found, status = 0;
 	struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data;
 
@@ -1290,10 +1315,8 @@ static int ocfs2_block_group_find_clear_bits(struct ocfs2_super *osb,
 	found = start = best_offset = best_size = 0;
 	bitmap = bg->bg_bitmap;
 
-	while((offset = ocfs2_find_next_zero_bit(bitmap, total_bits, start)) != -1) {
-		if (offset == total_bits)
-			break;
-
+	while ((offset = ocfs2_find_next_zero_bit(bitmap, total_bits, start)) <
+	       total_bits) {
 		if (!ocfs2_test_bg_bit_allocatable(bg_bh, offset)) {
 			/* We found a zero, but we can't use it as it
 			 * hasn't been put to disk yet! */
@@ -1308,6 +1331,7 @@ static int ocfs2_block_group_find_clear_bits(struct ocfs2_super *osb,
 			/* got a zero after some ones */
 			found = 1;
 			start = offset + 1;
+			prev_best_size = best_size;
 		}
 		if (found > best_size) {
 			best_size = found;
@@ -1320,6 +1344,8 @@ static int ocfs2_block_group_find_clear_bits(struct ocfs2_super *osb,
 		}
 	}
 
+	/* best_size will be allocated, we save prev_best_size */
+	res->sr_max_contig_bits = prev_best_size;
 	if (best_size) {
 		res->sr_bit_offset = best_offset;
 		res->sr_bits = best_size;
@@ -1337,11 +1363,16 @@ int ocfs2_block_group_set_bits(handle_t *handle,
 					     struct ocfs2_group_desc *bg,
 					     struct buffer_head *group_bh,
 					     unsigned int bit_off,
-					     unsigned int num_bits)
+					     unsigned int num_bits,
+					     unsigned int max_contig_bits,
+					     int fastpath)
 {
 	int status;
 	void *bitmap = bg->bg_bitmap;
 	int journal_type = OCFS2_JOURNAL_ACCESS_WRITE;
+	unsigned int start = bit_off + num_bits;
+	u16 contig_bits;
+	struct ocfs2_super *osb = OCFS2_SB(alloc_inode->i_sb);
 
 	/* All callers get the descriptor via
 	 * ocfs2_read_group_descriptor().  Any corruption is a code bug. */
@@ -1373,6 +1404,29 @@ int ocfs2_block_group_set_bits(handle_t *handle,
 	while(num_bits--)
 		ocfs2_set_bit(bit_off++, bitmap);
 
+	/*
+	 * this is optimize path, caller set old contig value
+	 * in max_contig_bits to bypass finding action.
+	 */
+	if (fastpath) {
+		bg->bg_contig_free_bits = cpu_to_le16(max_contig_bits);
+	} else if (ocfs2_is_cluster_bitmap(alloc_inode)) {
+		/*
+		 * Usually, the block group bitmap allocates only 1 bit
+		 * at a time, while the cluster group allocates n bits
+		 * each time. Therefore, we only save the contig bits for
+		 * the cluster group.
+		 */
+		contig_bits = ocfs2_find_max_contig_free_bits(bitmap,
+				    le16_to_cpu(bg->bg_bits), start);
+		if (contig_bits > max_contig_bits)
+			max_contig_bits = contig_bits;
+		bg->bg_contig_free_bits = cpu_to_le16(max_contig_bits);
+		ocfs2_local_alloc_seen_free_bits(osb, max_contig_bits);
+	} else {
+		bg->bg_contig_free_bits = 0;
+	}
+
 	ocfs2_journal_dirty(handle, group_bh);
 
 bail:
@@ -1486,7 +1540,12 @@ static int ocfs2_cluster_group_search(struct inode *inode,
 
 	BUG_ON(!ocfs2_is_cluster_bitmap(inode));
 
-	if (gd->bg_free_bits_count) {
+	if (le16_to_cpu(gd->bg_contig_free_bits) &&
+	    le16_to_cpu(gd->bg_contig_free_bits) < bits_wanted)
+		return -ENOSPC;
+
+	/* ->bg_contig_free_bits may un-initialized, so compare again */
+	if (le16_to_cpu(gd->bg_free_bits_count) >= bits_wanted) {
 		max_bits = le16_to_cpu(gd->bg_bits);
 
 		/* Tail groups in cluster bitmaps which aren't cpg
@@ -1530,13 +1589,6 @@ static int ocfs2_cluster_group_search(struct inode *inode,
 		 * of bits. */
 		if (min_bits <= res->sr_bits)
 			search = 0; /* success */
-		else if (res->sr_bits) {
-			/*
-			 * Don't show bits which we'll be returning
-			 * for allocation to the local alloc bitmap.
-			 */
-			ocfs2_local_alloc_seen_free_bits(osb, res->sr_bits);
-		}
 	}
 
 	return search;
@@ -1555,7 +1607,7 @@ static int ocfs2_block_group_search(struct inode *inode,
 	BUG_ON(min_bits != 1);
 	BUG_ON(ocfs2_is_cluster_bitmap(inode));
 
-	if (bg->bg_free_bits_count) {
+	if (le16_to_cpu(bg->bg_free_bits_count) >= bits_wanted) {
 		ret = ocfs2_block_group_find_clear_bits(OCFS2_SB(inode->i_sb),
 							group_bh, bits_wanted,
 							le16_to_cpu(bg->bg_bits),
@@ -1715,7 +1767,8 @@ static int ocfs2_search_one_group(struct ocfs2_alloc_context *ac,
 	}
 
 	ret = ocfs2_block_group_set_bits(handle, alloc_inode, gd, group_bh,
-					 res->sr_bit_offset, res->sr_bits);
+					 res->sr_bit_offset, res->sr_bits,
+					 res->sr_max_contig_bits, 0);
 	if (ret < 0) {
 		ocfs2_rollback_alloc_dinode_counts(alloc_inode, ac->ac_bh,
 					       res->sr_bits,
@@ -1849,7 +1902,9 @@ static int ocfs2_search_chain(struct ocfs2_alloc_context *ac,
 					    bg,
 					    group_bh,
 					    res->sr_bit_offset,
-					    res->sr_bits);
+					    res->sr_bits,
+					    res->sr_max_contig_bits,
+					    0);
 	if (status < 0) {
 		ocfs2_rollback_alloc_dinode_counts(alloc_inode,
 					ac->ac_bh, res->sr_bits, chain);
@@ -1951,7 +2006,7 @@ static int ocfs2_claim_suballoc_bits(struct ocfs2_alloc_context *ac,
 	for (i = 0; i < le16_to_cpu(cl->cl_next_free_rec); i ++) {
 		if (i == victim)
 			continue;
-		if (!cl->cl_recs[i].c_free)
+		if (le32_to_cpu(cl->cl_recs[i].c_free) < bits_wanted)
 			continue;
 
 		ac->ac_chain = i;
@@ -2163,7 +2218,9 @@ int ocfs2_claim_new_inode_at_loc(handle_t *handle,
 					 bg,
 					 bg_bh,
 					 res->sr_bit_offset,
-					 res->sr_bits);
+					 res->sr_bits,
+					 res->sr_max_contig_bits,
+					 0);
 	if (ret < 0) {
 		ocfs2_rollback_alloc_dinode_counts(ac->ac_inode,
 					       ac->ac_bh, res->sr_bits, chain);
@@ -2382,11 +2439,13 @@ static int ocfs2_block_group_clear_bits(handle_t *handle,
 					struct buffer_head *group_bh,
 					unsigned int bit_off,
 					unsigned int num_bits,
+					unsigned int max_contig_bits,
 					void (*undo_fn)(unsigned int bit,
 							unsigned long *bmap))
 {
 	int status;
 	unsigned int tmp;
+	u16 contig_bits;
 	struct ocfs2_group_desc *undo_bg = NULL;
 	struct journal_head *jh;
 
@@ -2433,6 +2492,20 @@ static int ocfs2_block_group_clear_bits(handle_t *handle,
 				   num_bits);
 	}
 
+	/*
+	 * TODO: even 'num_bits == 1' (the worst case, release 1 cluster),
+	 * we still need to rescan whole bitmap.
+	 */
+	if (ocfs2_is_cluster_bitmap(alloc_inode)) {
+		contig_bits = ocfs2_find_max_contig_free_bits(bg->bg_bitmap,
+				    le16_to_cpu(bg->bg_bits), 0);
+		if (contig_bits > max_contig_bits)
+			max_contig_bits = contig_bits;
+		bg->bg_contig_free_bits = cpu_to_le16(max_contig_bits);
+	} else {
+		bg->bg_contig_free_bits = 0;
+	}
+
 	if (undo_fn)
 		spin_unlock(&jh->b_state_lock);
 
@@ -2459,6 +2532,7 @@ static int _ocfs2_free_suballoc_bits(handle_t *handle,
 	struct ocfs2_chain_list *cl = &fe->id2.i_chain;
 	struct buffer_head *group_bh = NULL;
 	struct ocfs2_group_desc *group;
+	__le16 old_bg_contig_free_bits = 0;
 
 	/* The alloc_bh comes from ocfs2_free_dinode() or
 	 * ocfs2_free_clusters().  The callers have all locked the
@@ -2483,9 +2557,11 @@ static int _ocfs2_free_suballoc_bits(handle_t *handle,
 
 	BUG_ON((count + start_bit) > le16_to_cpu(group->bg_bits));
 
+	if (ocfs2_is_cluster_bitmap(alloc_inode))
+		old_bg_contig_free_bits = group->bg_contig_free_bits;
 	status = ocfs2_block_group_clear_bits(handle, alloc_inode,
 					      group, group_bh,
-					      start_bit, count, undo_fn);
+					      start_bit, count, 0, undo_fn);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
@@ -2496,7 +2572,8 @@ static int _ocfs2_free_suballoc_bits(handle_t *handle,
 	if (status < 0) {
 		mlog_errno(status);
 		ocfs2_block_group_set_bits(handle, alloc_inode, group, group_bh,
-				start_bit, count);
+				start_bit, count,
+				le16_to_cpu(old_bg_contig_free_bits), 1);
 		goto bail;
 	}
 
diff --git a/fs/ocfs2/suballoc.h b/fs/ocfs2/suballoc.h
index 9c74eace3a..b481b83485 100644
--- a/fs/ocfs2/suballoc.h
+++ b/fs/ocfs2/suballoc.h
@@ -79,12 +79,16 @@ void ocfs2_rollback_alloc_dinode_counts(struct inode *inode,
 			 struct buffer_head *di_bh,
 			 u32 num_bits,
 			 u16 chain);
+u16 ocfs2_find_max_contig_free_bits(void *bitmap,
+			 u16 total_bits, u16 start);
 int ocfs2_block_group_set_bits(handle_t *handle,
 			 struct inode *alloc_inode,
 			 struct ocfs2_group_desc *bg,
 			 struct buffer_head *group_bh,
 			 unsigned int bit_off,
-			 unsigned int num_bits);
+			 unsigned int num_bits,
+			 unsigned int max_contig_bits,
+			 int fastpath);
 
 int ocfs2_claim_metadata(handle_t *handle,
 			 struct ocfs2_alloc_context *ac,
diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index 1259fe02cd..afee70125a 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -1075,9 +1075,11 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
 	debugfs_create_file("fs_state", S_IFREG|S_IRUSR, osb->osb_debug_root,
 			    osb, &ocfs2_osb_debug_fops);
 
-	if (ocfs2_meta_ecc(osb))
+	if (ocfs2_meta_ecc(osb)) {
+		ocfs2_initialize_journal_triggers(sb, osb->s_journal_triggers);
 		ocfs2_blockcheck_stats_debugfs_install( &osb->osb_ecc_stats,
 							osb->osb_debug_root);
+	}
 
 	status = ocfs2_mount_volume(sb);
 	if (status < 0)
@@ -1706,18 +1708,17 @@ static int ocfs2_initialize_mem_caches(void)
 				       sizeof(struct ocfs2_inode_info),
 				       0,
 				       (SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
-						SLAB_MEM_SPREAD|SLAB_ACCOUNT),
+						SLAB_ACCOUNT),
 				       ocfs2_inode_init_once);
 	ocfs2_dquot_cachep = kmem_cache_create("ocfs2_dquot_cache",
 					sizeof(struct ocfs2_dquot),
 					0,
-					(SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
-						SLAB_MEM_SPREAD),
+					SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT,
 					NULL);
 	ocfs2_qf_chunk_cachep = kmem_cache_create("ocfs2_qf_chunk_cache",
 					sizeof(struct ocfs2_quota_chunk),
 					0,
-					(SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD),
+					SLAB_RECLAIM_ACCOUNT,
 					NULL);
 	if (!ocfs2_inode_cachep || !ocfs2_dquot_cachep ||
 	    !ocfs2_qf_chunk_cachep) {
@@ -2027,8 +2028,8 @@ static int ocfs2_initialize_super(struct super_block *sb,
 	cbits = le32_to_cpu(di->id2.i_super.s_clustersize_bits);
 	bbits = le32_to_cpu(di->id2.i_super.s_blocksize_bits);
 	sb->s_maxbytes = ocfs2_max_file_offset(bbits, cbits);
-	memcpy(&sb->s_uuid, di->id2.i_super.s_uuid,
-	       sizeof(di->id2.i_super.s_uuid));
+	super_set_uuid(sb, di->id2.i_super.s_uuid,
+		       sizeof(di->id2.i_super.s_uuid));
 
 	osb->osb_dx_mask = (1 << (cbits - bbits)) - 1;
 
diff --git a/fs/open.c b/fs/open.c
index a84d21e55c..278b3edcda 100644
--- a/fs/open.c
+++ b/fs/open.c
@@ -29,7 +29,6 @@
 #include <linux/audit.h>
 #include <linux/falloc.h>
 #include <linux/fs_struct.h>
-#include <linux/ima.h>
 #include <linux/dnotify.h>
 #include <linux/compat.h>
 #include <linux/mnt_idmapping.h>
@@ -154,59 +153,62 @@ COMPAT_SYSCALL_DEFINE2(truncate, const char __user *, path, compat_off_t, length
 }
 #endif
 
-long do_sys_ftruncate(unsigned int fd, loff_t length, int small)
+long do_ftruncate(struct file *file, loff_t length, int small)
 {
 	struct inode *inode;
 	struct dentry *dentry;
-	struct fd f;
 	int error;
 
-	error = -EINVAL;
-	if (length < 0)
-		goto out;
-	error = -EBADF;
-	f = fdget(fd);
-	if (!f.file)
-		goto out;
-
 	/* explicitly opened as large or we are on 64-bit box */
-	if (f.file->f_flags & O_LARGEFILE)
+	if (file->f_flags & O_LARGEFILE)
 		small = 0;
 
-	dentry = f.file->f_path.dentry;
+	dentry = file->f_path.dentry;
 	inode = dentry->d_inode;
-	error = -EINVAL;
-	if (!S_ISREG(inode->i_mode) || !(f.file->f_mode & FMODE_WRITE))
-		goto out_putf;
+	if (!S_ISREG(inode->i_mode) || !(file->f_mode & FMODE_WRITE))
+		return -EINVAL;
 
-	error = -EINVAL;
 	/* Cannot ftruncate over 2^31 bytes without large file support */
 	if (small && length > MAX_NON_LFS)
-		goto out_putf;
+		return -EINVAL;
 
-	error = -EPERM;
 	/* Check IS_APPEND on real upper inode */
-	if (IS_APPEND(file_inode(f.file)))
-		goto out_putf;
+	if (IS_APPEND(file_inode(file)))
+		return -EPERM;
 	sb_start_write(inode->i_sb);
-	error = security_file_truncate(f.file);
+	error = security_file_truncate(file);
 	if (!error)
-		error = do_truncate(file_mnt_idmap(f.file), dentry, length,
-				    ATTR_MTIME | ATTR_CTIME, f.file);
+		error = do_truncate(file_mnt_idmap(file), dentry, length,
+				    ATTR_MTIME | ATTR_CTIME, file);
 	sb_end_write(inode->i_sb);
-out_putf:
+
+	return error;
+}
+
+long do_sys_ftruncate(unsigned int fd, loff_t length, int small)
+{
+	struct fd f;
+	int error;
+
+	if (length < 0)
+		return -EINVAL;
+	f = fdget(fd);
+	if (!f.file)
+		return -EBADF;
+
+	error = do_ftruncate(f.file, length, small);
+
 	fdput(f);
-out:
 	return error;
 }
 
-SYSCALL_DEFINE2(ftruncate, unsigned int, fd, unsigned long, length)
+SYSCALL_DEFINE2(ftruncate, unsigned int, fd, off_t, length)
 {
 	return do_sys_ftruncate(fd, length, 1);
 }
 
 #ifdef CONFIG_COMPAT
-COMPAT_SYSCALL_DEFINE2(ftruncate, unsigned int, fd, compat_ulong_t, length)
+COMPAT_SYSCALL_DEFINE2(ftruncate, unsigned int, fd, compat_off_t, length)
 {
 	return do_sys_ftruncate(fd, length, 1);
 }
@@ -900,10 +902,10 @@ cleanup_inode:
 }
 
 static int do_dentry_open(struct file *f,
-			  struct inode *inode,
 			  int (*open)(struct inode *, struct file *))
 {
 	static const struct file_operations empty_fops = {};
+	struct inode *inode = f->f_path.dentry->d_inode;
 	int error;
 
 	path_get(&f->f_path);
@@ -1002,11 +1004,6 @@ static int do_dentry_open(struct file *f,
 		}
 	}
 
-	/*
-	 * Once we return a file with FMODE_OPENED, __fput() will call
-	 * fsnotify_close(), so we need fsnotify_open() here for symmetry.
-	 */
-	fsnotify_open(f);
 	return 0;
 
 cleanup_all:
@@ -1045,7 +1042,7 @@ int finish_open(struct file *file, struct dentry *dentry,
 	BUG_ON(file->f_mode & FMODE_OPENED); /* once it's opened, it's opened */
 
 	file->f_path.dentry = dentry;
-	return do_dentry_open(file, d_backing_inode(dentry), open);
+	return do_dentry_open(file, open);
 }
 EXPORT_SYMBOL(finish_open);
 
@@ -1083,8 +1080,19 @@ EXPORT_SYMBOL(file_path);
  */
 int vfs_open(const struct path *path, struct file *file)
 {
+	int ret;
+
 	file->f_path = *path;
-	return do_dentry_open(file, d_backing_inode(path->dentry), NULL);
+	ret = do_dentry_open(file, NULL);
+	if (!ret) {
+		/*
+		 * Once we return a file with FMODE_OPENED, __fput() will call
+		 * fsnotify_close(), so we need fsnotify_open() here for
+		 * symmetry.
+		 */
+		fsnotify_open(file);
+	}
+	return ret;
 }
 
 struct file *dentry_open(const struct path *path, int flags,
@@ -1153,7 +1161,6 @@ EXPORT_SYMBOL(dentry_create);
  * kernel_file_open - open a file for kernel internal use
  * @path:	path of the file to open
  * @flags:	open flags
- * @inode:	the inode
  * @cred:	credentials for open
  *
  * Open a file for use by in-kernel consumers. The file is not accounted
@@ -1163,7 +1170,7 @@ EXPORT_SYMBOL(dentry_create);
  * Return: Opened file on success, an error pointer on failure.
  */
 struct file *kernel_file_open(const struct path *path, int flags,
-				struct inode *inode, const struct cred *cred)
+				const struct cred *cred)
 {
 	struct file *f;
 	int error;
@@ -1173,11 +1180,13 @@ struct file *kernel_file_open(const struct path *path, int flags,
 		return f;
 
 	f->f_path = *path;
-	error = do_dentry_open(f, inode, NULL);
+	error = do_dentry_open(f, NULL);
 	if (error) {
 		fput(f);
-		f = ERR_PTR(error);
+		return ERR_PTR(error);
 	}
+
+	fsnotify_open(f);
 	return f;
 }
 EXPORT_SYMBOL_GPL(kernel_file_open);
@@ -1364,7 +1373,7 @@ struct file *filp_open(const char *filename, int flags, umode_t mode)
 {
 	struct filename *name = getname_kernel(filename);
 	struct file *file = ERR_CAST(name);
-	
+
 	if (!IS_ERR(name)) {
 		file = file_open_name(name, flags, mode);
 		putname(name);
diff --git a/fs/openpromfs/inode.c b/fs/openpromfs/inode.c
index 5cf1809d47..a7b527ea50 100644
--- a/fs/openpromfs/inode.c
+++ b/fs/openpromfs/inode.c
@@ -446,7 +446,7 @@ static int __init init_openprom_fs(void)
 					    sizeof(struct op_inode_info),
 					    0,
 					    (SLAB_RECLAIM_ACCOUNT |
-					     SLAB_MEM_SPREAD | SLAB_ACCOUNT),
+					     SLAB_ACCOUNT),
 					    op_inode_init_once);
 	if (!op_inode_cachep)
 		return -ENOMEM;
diff --git a/fs/orangefs/dcache.c b/fs/orangefs/dcache.c
index 8bbe9486e3..395a00ed8a 100644
--- a/fs/orangefs/dcache.c
+++ b/fs/orangefs/dcache.c
@@ -33,9 +33,7 @@ static int orangefs_revalidate_lookup(struct dentry *dentry)
 
 	new_op->upcall.req.lookup.sym_follow = ORANGEFS_LOOKUP_LINK_NO_FOLLOW;
 	new_op->upcall.req.lookup.parent_refn = parent->refn;
-	strncpy(new_op->upcall.req.lookup.d_name,
-		dentry->d_name.name,
-		ORANGEFS_NAME_MAX - 1);
+	strscpy(new_op->upcall.req.lookup.d_name, dentry->d_name.name);
 
 	gossip_debug(GOSSIP_DCACHE_DEBUG,
 		     "%s:%s:%d interrupt flag [%d]\n",
diff --git a/fs/orangefs/namei.c b/fs/orangefs/namei.c
index c9dfd5c6a0..200558ec72 100644
--- a/fs/orangefs/namei.c
+++ b/fs/orangefs/namei.c
@@ -41,8 +41,7 @@ static int orangefs_create(struct mnt_idmap *idmap,
 	fill_default_sys_attrs(new_op->upcall.req.create.attributes,
 			       ORANGEFS_TYPE_METAFILE, mode);
 
-	strncpy(new_op->upcall.req.create.d_name,
-		dentry->d_name.name, ORANGEFS_NAME_MAX - 1);
+	strscpy(new_op->upcall.req.create.d_name, dentry->d_name.name);
 
 	ret = service_operation(new_op, __func__, get_interruptible_flag(dir));
 
@@ -137,8 +136,7 @@ static struct dentry *orangefs_lookup(struct inode *dir, struct dentry *dentry,
 		     &parent->refn.khandle);
 	new_op->upcall.req.lookup.parent_refn = parent->refn;
 
-	strncpy(new_op->upcall.req.lookup.d_name, dentry->d_name.name,
-		ORANGEFS_NAME_MAX - 1);
+	strscpy(new_op->upcall.req.lookup.d_name, dentry->d_name.name);
 
 	gossip_debug(GOSSIP_NAME_DEBUG,
 		     "%s: doing lookup on %s under %pU,%d\n",
@@ -192,8 +190,7 @@ static int orangefs_unlink(struct inode *dir, struct dentry *dentry)
 		return -ENOMEM;
 
 	new_op->upcall.req.remove.parent_refn = parent->refn;
-	strncpy(new_op->upcall.req.remove.d_name, dentry->d_name.name,
-		ORANGEFS_NAME_MAX - 1);
+	strscpy(new_op->upcall.req.remove.d_name, dentry->d_name.name);
 
 	ret = service_operation(new_op, "orangefs_unlink",
 				get_interruptible_flag(inode));
@@ -247,10 +244,8 @@ static int orangefs_symlink(struct mnt_idmap *idmap,
 			       ORANGEFS_TYPE_SYMLINK,
 			       mode);
 
-	strncpy(new_op->upcall.req.sym.entry_name,
-		dentry->d_name.name,
-		ORANGEFS_NAME_MAX - 1);
-	strncpy(new_op->upcall.req.sym.target, symname, ORANGEFS_NAME_MAX - 1);
+	strscpy(new_op->upcall.req.sym.entry_name, dentry->d_name.name);
+	strscpy(new_op->upcall.req.sym.target, symname);
 
 	ret = service_operation(new_op, __func__, get_interruptible_flag(dir));
 
@@ -324,8 +319,7 @@ static int orangefs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
 	fill_default_sys_attrs(new_op->upcall.req.mkdir.attributes,
 			      ORANGEFS_TYPE_DIRECTORY, mode);
 
-	strncpy(new_op->upcall.req.mkdir.d_name,
-		dentry->d_name.name, ORANGEFS_NAME_MAX - 1);
+	strscpy(new_op->upcall.req.mkdir.d_name, dentry->d_name.name);
 
 	ret = service_operation(new_op, __func__, get_interruptible_flag(dir));
 
@@ -405,12 +399,8 @@ static int orangefs_rename(struct mnt_idmap *idmap,
 	new_op->upcall.req.rename.old_parent_refn = ORANGEFS_I(old_dir)->refn;
 	new_op->upcall.req.rename.new_parent_refn = ORANGEFS_I(new_dir)->refn;
 
-	strncpy(new_op->upcall.req.rename.d_old_name,
-		old_dentry->d_name.name,
-		ORANGEFS_NAME_MAX - 1);
-	strncpy(new_op->upcall.req.rename.d_new_name,
-		new_dentry->d_name.name,
-		ORANGEFS_NAME_MAX - 1);
+	strscpy(new_op->upcall.req.rename.d_old_name, old_dentry->d_name.name);
+	strscpy(new_op->upcall.req.rename.d_new_name, new_dentry->d_name.name);
 
 	ret = service_operation(new_op,
 				"orangefs_rename",
diff --git a/fs/orangefs/orangefs-cache.c b/fs/orangefs/orangefs-cache.c
index 3b6982bf6b..e75e173a91 100644
--- a/fs/orangefs/orangefs-cache.c
+++ b/fs/orangefs/orangefs-cache.c
@@ -22,7 +22,7 @@ int op_cache_initialize(void)
 	op_cache = kmem_cache_create("orangefs_op_cache",
 				     sizeof(struct orangefs_kernel_op_s),
 				     0,
-				     ORANGEFS_CACHE_CREATE_FLAGS,
+				     0,
 				     NULL);
 
 	if (!op_cache) {
diff --git a/fs/orangefs/orangefs-kernel.h b/fs/orangefs/orangefs-kernel.h
index 926d9c0a42..e2df7eeadc 100644
--- a/fs/orangefs/orangefs-kernel.h
+++ b/fs/orangefs/orangefs-kernel.h
@@ -93,16 +93,6 @@ enum orangefs_vfs_op_states {
 	OP_VFS_STATE_GIVEN_UP = 16,
 };
 
-/*
- * orangefs kernel memory related flags
- */
-
-#if (defined CONFIG_DEBUG_SLAB)
-#define ORANGEFS_CACHE_CREATE_FLAGS SLAB_RED_ZONE
-#else
-#define ORANGEFS_CACHE_CREATE_FLAGS 0
-#endif
-
 extern const struct xattr_handler * const orangefs_xattr_handlers[];
 
 extern struct posix_acl *orangefs_get_acl(struct inode *inode, int type, bool rcu);
diff --git a/fs/orangefs/super.c b/fs/orangefs/super.c
index 4ca8ed410c..eba3e35719 100644
--- a/fs/orangefs/super.c
+++ b/fs/orangefs/super.c
@@ -201,7 +201,8 @@ static int orangefs_statfs(struct dentry *dentry, struct kstatfs *buf)
 		     (long)new_op->downcall.resp.statfs.files_avail);
 
 	buf->f_type = sb->s_magic;
-	memcpy(&buf->f_fsid, &ORANGEFS_SB(sb)->fs_id, sizeof(buf->f_fsid));
+	buf->f_fsid.val[0] = ORANGEFS_SB(sb)->fs_id;
+	buf->f_fsid.val[1] = ORANGEFS_SB(sb)->id;
 	buf->f_bsize = new_op->downcall.resp.statfs.block_size;
 	buf->f_namelen = ORANGEFS_NAME_MAX;
 
@@ -253,9 +254,8 @@ int orangefs_remount(struct orangefs_sb_info_s *orangefs_sb)
 	new_op = op_alloc(ORANGEFS_VFS_OP_FS_MOUNT);
 	if (!new_op)
 		return -ENOMEM;
-	strncpy(new_op->upcall.req.fs_mount.orangefs_config_server,
-		orangefs_sb->devname,
-		ORANGEFS_MAX_SERVER_ADDR_LEN);
+	strscpy(new_op->upcall.req.fs_mount.orangefs_config_server,
+		orangefs_sb->devname);
 
 	gossip_debug(GOSSIP_SUPER_DEBUG,
 		     "Attempting ORANGEFS Remount via host %s\n",
@@ -400,8 +400,7 @@ static int orangefs_unmount(int id, __s32 fs_id, const char *devname)
 		return -ENOMEM;
 	op->upcall.req.fs_umount.id = id;
 	op->upcall.req.fs_umount.fs_id = fs_id;
-	strncpy(op->upcall.req.fs_umount.orangefs_config_server,
-	    devname, ORANGEFS_MAX_SERVER_ADDR_LEN - 1);
+	strscpy(op->upcall.req.fs_umount.orangefs_config_server, devname);
 	r = service_operation(op, "orangefs_fs_umount", 0);
 	/* Not much to do about an error here. */
 	if (r)
@@ -494,9 +493,7 @@ struct dentry *orangefs_mount(struct file_system_type *fst,
 	if (!new_op)
 		return ERR_PTR(-ENOMEM);
 
-	strncpy(new_op->upcall.req.fs_mount.orangefs_config_server,
-		devname,
-		ORANGEFS_MAX_SERVER_ADDR_LEN - 1);
+	strscpy(new_op->upcall.req.fs_mount.orangefs_config_server, devname);
 
 	gossip_debug(GOSSIP_SUPER_DEBUG,
 		     "Attempting ORANGEFS Mount via host %s\n",
@@ -543,9 +540,8 @@ struct dentry *orangefs_mount(struct file_system_type *fst,
 	 * on successful mount, store the devname and data
 	 * used
 	 */
-	strncpy(ORANGEFS_SB(sb)->devname,
-		devname,
-		ORANGEFS_MAX_SERVER_ADDR_LEN - 1);
+	strscpy(ORANGEFS_SB(sb)->devname, devname);
+
 
 	/* mount_pending must be cleared */
 	ORANGEFS_SB(sb)->mount_pending = 0;
@@ -644,7 +640,7 @@ int orangefs_inode_cache_initialize(void)
 					"orangefs_inode_cache",
 					sizeof(struct orangefs_inode_s),
 					0,
-					ORANGEFS_CACHE_CREATE_FLAGS,
+					0,
 					offsetof(struct orangefs_inode_s,
 						link_target),
 					sizeof_field(struct orangefs_inode_s,
diff --git a/fs/overlayfs/copy_up.c b/fs/overlayfs/copy_up.c
index 0762575a1e..a5ef2005a2 100644
--- a/fs/overlayfs/copy_up.c
+++ b/fs/overlayfs/copy_up.c
@@ -114,7 +114,7 @@ int ovl_copy_xattr(struct super_block *sb, const struct path *oldpath, struct de
 		if (ovl_is_private_xattr(sb, name))
 			continue;
 
-		error = security_inode_copy_up_xattr(name);
+		error = security_inode_copy_up_xattr(old, name);
 		if (error < 0 && error != -EOPNOTSUPP)
 			break;
 		if (error == 1) {
diff --git a/fs/overlayfs/dir.c b/fs/overlayfs/dir.c
index 0f8b4a7192..ab65e98a1d 100644
--- a/fs/overlayfs/dir.c
+++ b/fs/overlayfs/dir.c
@@ -14,6 +14,7 @@
 #include <linux/posix_acl_xattr.h>
 #include <linux/atomic.h>
 #include <linux/ratelimit.h>
+#include <linux/backing-file.h>
 #include "overlayfs.h"
 
 static unsigned short ovl_redirect_max = 256;
@@ -260,14 +261,13 @@ static int ovl_set_opaque(struct dentry *dentry, struct dentry *upperdentry)
  * may not use to instantiate the new dentry.
  */
 static int ovl_instantiate(struct dentry *dentry, struct inode *inode,
-			   struct dentry *newdentry, bool hardlink)
+			   struct dentry *newdentry, bool hardlink, struct file *tmpfile)
 {
 	struct ovl_inode_params oip = {
 		.upperdentry = newdentry,
 		.newinode = inode,
 	};
 
-	ovl_dir_modified(dentry->d_parent, false);
 	ovl_dentry_set_upper_alias(dentry);
 	ovl_dentry_init_reval(dentry, newdentry, NULL);
 
@@ -295,6 +295,9 @@ static int ovl_instantiate(struct dentry *dentry, struct inode *inode,
 		inc_nlink(inode);
 	}
 
+	if (tmpfile)
+		d_mark_tmpfile(tmpfile, inode);
+
 	d_instantiate(dentry, inode);
 	if (inode != oip.newinode) {
 		pr_warn_ratelimited("newly created inode found in cache (%pd2)\n",
@@ -327,9 +330,6 @@ static int ovl_create_upper(struct dentry *dentry, struct inode *inode,
 	struct dentry *newdentry;
 	int err;
 
-	if (!attr->hardlink && !IS_POSIXACL(udir))
-		attr->mode &= ~current_umask();
-
 	inode_lock_nested(udir, I_MUTEX_PARENT);
 	newdentry = ovl_create_real(ofs, udir,
 				    ovl_lookup_upper(ofs, dentry->d_name.name,
@@ -345,7 +345,8 @@ static int ovl_create_upper(struct dentry *dentry, struct inode *inode,
 		ovl_set_opaque(dentry, newdentry);
 	}
 
-	err = ovl_instantiate(dentry, inode, newdentry, !!attr->hardlink);
+	ovl_dir_modified(dentry->d_parent, false);
+	err = ovl_instantiate(dentry, inode, newdentry, !!attr->hardlink, NULL);
 	if (err)
 		goto out_cleanup;
 out_unlock:
@@ -529,7 +530,8 @@ static int ovl_create_over_whiteout(struct dentry *dentry, struct inode *inode,
 		if (err)
 			goto out_cleanup;
 	}
-	err = ovl_instantiate(dentry, inode, newdentry, hardlink);
+	ovl_dir_modified(dentry->d_parent, false);
+	err = ovl_instantiate(dentry, inode, newdentry, hardlink, NULL);
 	if (err) {
 		ovl_cleanup(ofs, udir, newdentry);
 		dput(newdentry);
@@ -551,12 +553,35 @@ out_cleanup:
 	goto out_dput;
 }
 
+static int ovl_setup_cred_for_create(struct dentry *dentry, struct inode *inode,
+				     umode_t mode, const struct cred *old_cred)
+{
+	int err;
+	struct cred *override_cred;
+
+	override_cred = prepare_creds();
+	if (!override_cred)
+		return -ENOMEM;
+
+	override_cred->fsuid = inode->i_uid;
+	override_cred->fsgid = inode->i_gid;
+	err = security_dentry_create_files_as(dentry, mode, &dentry->d_name,
+					      old_cred, override_cred);
+	if (err) {
+		put_cred(override_cred);
+		return err;
+	}
+	put_cred(override_creds(override_cred));
+	put_cred(override_cred);
+
+	return 0;
+}
+
 static int ovl_create_or_link(struct dentry *dentry, struct inode *inode,
 			      struct ovl_cattr *attr, bool origin)
 {
 	int err;
 	const struct cred *old_cred;
-	struct cred *override_cred;
 	struct dentry *parent = dentry->d_parent;
 
 	old_cred = ovl_override_creds(dentry->d_sb);
@@ -572,10 +597,6 @@ static int ovl_create_or_link(struct dentry *dentry, struct inode *inode,
 	}
 
 	if (!attr->hardlink) {
-		err = -ENOMEM;
-		override_cred = prepare_creds();
-		if (!override_cred)
-			goto out_revert_creds;
 		/*
 		 * In the creation cases(create, mkdir, mknod, symlink),
 		 * ovl should transfer current's fs{u,g}id to underlying
@@ -589,17 +610,9 @@ static int ovl_create_or_link(struct dentry *dentry, struct inode *inode,
 		 * create a new inode, so just use the ovl mounter's
 		 * fs{u,g}id.
 		 */
-		override_cred->fsuid = inode->i_uid;
-		override_cred->fsgid = inode->i_gid;
-		err = security_dentry_create_files_as(dentry,
-				attr->mode, &dentry->d_name, old_cred,
-				override_cred);
-		if (err) {
-			put_cred(override_cred);
+		err = ovl_setup_cred_for_create(dentry, inode, attr->mode, old_cred);
+		if (err)
 			goto out_revert_creds;
-		}
-		put_cred(override_creds(override_cred));
-		put_cred(override_cred);
 	}
 
 	if (!ovl_dentry_is_whiteout(dentry))
@@ -1290,6 +1303,100 @@ out:
 	return err;
 }
 
+static int ovl_create_tmpfile(struct file *file, struct dentry *dentry,
+			      struct inode *inode, umode_t mode)
+{
+	const struct cred *old_cred;
+	struct path realparentpath;
+	struct file *realfile;
+	struct dentry *newdentry;
+	/* It's okay to set O_NOATIME, since the owner will be current fsuid */
+	int flags = file->f_flags | OVL_OPEN_FLAGS;
+	int err;
+
+	old_cred = ovl_override_creds(dentry->d_sb);
+	err = ovl_setup_cred_for_create(dentry, inode, mode, old_cred);
+	if (err)
+		goto out_revert_creds;
+
+	ovl_path_upper(dentry->d_parent, &realparentpath);
+	realfile = backing_tmpfile_open(&file->f_path, flags, &realparentpath,
+					mode, current_cred());
+	err = PTR_ERR_OR_ZERO(realfile);
+	pr_debug("tmpfile/open(%pd2, 0%o) = %i\n", realparentpath.dentry, mode, err);
+	if (err)
+		goto out_revert_creds;
+
+	/* ovl_instantiate() consumes the newdentry reference on success */
+	newdentry = dget(realfile->f_path.dentry);
+	err = ovl_instantiate(dentry, inode, newdentry, false, file);
+	if (!err) {
+		file->private_data = realfile;
+	} else {
+		dput(newdentry);
+		fput(realfile);
+	}
+out_revert_creds:
+	revert_creds(old_cred);
+	return err;
+}
+
+static int ovl_dummy_open(struct inode *inode, struct file *file)
+{
+	return 0;
+}
+
+static int ovl_tmpfile(struct mnt_idmap *idmap, struct inode *dir,
+		       struct file *file, umode_t mode)
+{
+	int err;
+	struct dentry *dentry = file->f_path.dentry;
+	struct inode *inode;
+
+	if (!OVL_FS(dentry->d_sb)->tmpfile)
+		return -EOPNOTSUPP;
+
+	err = ovl_copy_up(dentry->d_parent);
+	if (err)
+		return err;
+
+	err = ovl_want_write(dentry);
+	if (err)
+		return err;
+
+	err = -ENOMEM;
+	inode = ovl_new_inode(dentry->d_sb, mode, 0);
+	if (!inode)
+		goto drop_write;
+
+	inode_init_owner(&nop_mnt_idmap, inode, dir, mode);
+	err = ovl_create_tmpfile(file, dentry, inode, inode->i_mode);
+	if (err)
+		goto put_inode;
+
+	/*
+	 * Check if the preallocated inode was actually used.  Having something
+	 * else assigned to the dentry shouldn't happen as that would indicate
+	 * that the backing tmpfile "leaked" out of overlayfs.
+	 */
+	err = -EIO;
+	if (WARN_ON(inode != d_inode(dentry)))
+		goto put_realfile;
+
+	/* inode reference was transferred to dentry */
+	inode = NULL;
+	err = finish_open(file, dentry, ovl_dummy_open);
+put_realfile:
+	/* Without FMODE_OPENED ->release() won't be called on @file */
+	if (!(file->f_mode & FMODE_OPENED))
+		fput(file->private_data);
+put_inode:
+	iput(inode);
+drop_write:
+	ovl_drop_write(dentry);
+	return err;
+}
+
 const struct inode_operations ovl_dir_inode_operations = {
 	.lookup		= ovl_lookup,
 	.mkdir		= ovl_mkdir,
@@ -1310,4 +1417,5 @@ const struct inode_operations ovl_dir_inode_operations = {
 	.update_time	= ovl_update_time,
 	.fileattr_get	= ovl_fileattr_get,
 	.fileattr_set	= ovl_fileattr_set,
+	.tmpfile	= ovl_tmpfile,
 };
diff --git a/fs/overlayfs/export.c b/fs/overlayfs/export.c
index 063409069f..5868cb2229 100644
--- a/fs/overlayfs/export.c
+++ b/fs/overlayfs/export.c
@@ -181,6 +181,10 @@ static int ovl_check_encode_origin(struct dentry *dentry)
 	struct ovl_fs *ofs = OVL_FS(dentry->d_sb);
 	bool decodable = ofs->config.nfs_export;
 
+	/* No upper layer? */
+	if (!ovl_upper_mnt(ofs))
+		return 1;
+
 	/* Lower file handle for non-upper non-decodable */
 	if (!ovl_dentry_upper(dentry) && !decodable)
 		return 1;
@@ -209,7 +213,7 @@ static int ovl_check_encode_origin(struct dentry *dentry)
 	 * ovl_connect_layer() will try to make origin's layer "connected" by
 	 * copying up a "connectable" ancestor.
 	 */
-	if (d_is_dir(dentry) && ovl_upper_mnt(ofs) && decodable)
+	if (d_is_dir(dentry) && decodable)
 		return ovl_connect_layer(dentry);
 
 	/* Lower file handle for indexed and non-upper dir/non-dir */
diff --git a/fs/overlayfs/file.c b/fs/overlayfs/file.c
index 05536964d3..1a411cae57 100644
--- a/fs/overlayfs/file.c
+++ b/fs/overlayfs/file.c
@@ -24,9 +24,6 @@ static char ovl_whatisit(struct inode *inode, struct inode *realinode)
 		return 'm';
 }
 
-/* No atime modification on underlying */
-#define OVL_OPEN_FLAGS (O_NOATIME)
-
 static struct file *ovl_open_realfile(const struct file *file,
 				      const struct path *realpath)
 {
diff --git a/fs/overlayfs/inode.c b/fs/overlayfs/inode.c
index c63b31a460..35fd3e3e17 100644
--- a/fs/overlayfs/inode.c
+++ b/fs/overlayfs/inode.c
@@ -8,7 +8,6 @@
 #include <linux/slab.h>
 #include <linux/cred.h>
 #include <linux/xattr.h>
-#include <linux/posix_acl.h>
 #include <linux/ratelimit.h>
 #include <linux/fiemap.h>
 #include <linux/fileattr.h>
diff --git a/fs/overlayfs/overlayfs.h b/fs/overlayfs/overlayfs.h
index ee949f3e7c..0bfe35da4b 100644
--- a/fs/overlayfs/overlayfs.h
+++ b/fs/overlayfs/overlayfs.h
@@ -175,6 +175,9 @@ static inline int ovl_metadata_digest_size(const struct ovl_metacopy *metacopy)
 	return (int)metacopy->len - OVL_METACOPY_MIN_SIZE;
 }
 
+/* No atime modification on underlying */
+#define OVL_OPEN_FLAGS (O_NOATIME)
+
 extern const char *const ovl_xattr_table[][2];
 static inline const char *ovl_xattr(struct ovl_fs *ofs, enum ovl_xattr ox)
 {
diff --git a/fs/overlayfs/params.c b/fs/overlayfs/params.c
index 36dcc530ac..4860fcc461 100644
--- a/fs/overlayfs/params.c
+++ b/fs/overlayfs/params.c
@@ -139,10 +139,6 @@ static int ovl_verity_mode_def(void)
 	return OVL_VERITY_OFF;
 }
 
-#define fsparam_string_empty(NAME, OPT) \
-	__fsparam(fs_param_is_string, NAME, OPT, fs_param_can_be_empty, NULL)
-
-
 const struct fs_parameter_spec ovl_parameter_spec[] = {
 	fsparam_string_empty("lowerdir",    Opt_lowerdir),
 	fsparam_string("lowerdir+",         Opt_lowerdir_add),
diff --git a/fs/overlayfs/super.c b/fs/overlayfs/super.c
index 2eef6c70b2..06a231970c 100644
--- a/fs/overlayfs/super.c
+++ b/fs/overlayfs/super.c
@@ -28,41 +28,38 @@ MODULE_LICENSE("GPL");
 
 struct ovl_dir_cache;
 
-static struct dentry *ovl_d_real(struct dentry *dentry,
-				 const struct inode *inode)
+static struct dentry *ovl_d_real(struct dentry *dentry, enum d_real_type type)
 {
-	struct dentry *real = NULL, *lower;
+	struct dentry *upper, *lower;
 	int err;
 
-	/*
-	 * vfs is only expected to call d_real() with NULL from d_real_inode()
-	 * and with overlay inode from file_dentry() on an overlay file.
-	 *
-	 * TODO: remove @inode argument from d_real() API, remove code in this
-	 * function that deals with non-NULL @inode and remove d_real() call
-	 * from file_dentry().
-	 */
-	if (inode && d_inode(dentry) == inode)
-		return dentry;
-	else if (inode)
+	switch (type) {
+	case D_REAL_DATA:
+	case D_REAL_METADATA:
+		break;
+	default:
 		goto bug;
+	}
 
 	if (!d_is_reg(dentry)) {
 		/* d_real_inode() is only relevant for regular files */
 		return dentry;
 	}
 
-	real = ovl_dentry_upper(dentry);
-	if (real && (inode == d_inode(real)))
-		return real;
+	upper = ovl_dentry_upper(dentry);
+	if (upper && (type == D_REAL_METADATA ||
+		      ovl_has_upperdata(d_inode(dentry))))
+		return upper;
 
-	if (real && !inode && ovl_has_upperdata(d_inode(dentry)))
-		return real;
+	if (type == D_REAL_METADATA) {
+		lower = ovl_dentry_lower(dentry);
+		goto real_lower;
+	}
 
 	/*
-	 * Best effort lazy lookup of lowerdata for !inode case to return
+	 * Best effort lazy lookup of lowerdata for D_REAL_DATA case to return
 	 * the real lowerdata dentry.  The only current caller of d_real() with
-	 * NULL inode is d_real_inode() from trace_uprobe and this caller is
+	 * D_REAL_DATA is d_real_inode() from trace_uprobe and this caller is
 	 * likely going to be followed reading from the file, before placing
 	 * uprobes on offset within the file, so lowerdata should be available
 	 * when setting the uprobe.
@@ -73,18 +70,13 @@ static struct dentry *ovl_d_real(struct dentry *dentry,
 	lower = ovl_dentry_lowerdata(dentry);
 	if (!lower)
 		goto bug;
-	real = lower;
 
-	/* Handle recursion */
-	real = d_real(real, inode);
+real_lower:
+	/* Handle recursion into stacked lower fs */
+	return d_real(lower, type);
 
-	if (!inode || inode == d_inode(real))
-		return real;
 bug:
-	WARN(1, "%s(%pd4, %s:%lu): real dentry (%p/%lu) not found\n",
-	     __func__, dentry, inode ? inode->i_sb->s_id : "NULL",
-	     inode ? inode->i_ino : 0, real,
-	     real && d_inode(real) ? d_inode(real)->i_ino : 0);
+	WARN(1, "%s(%pd4, %d): real dentry not found\n", __func__, dentry, type);
 	return dentry;
 }
 
@@ -1468,7 +1460,7 @@ int ovl_fill_super(struct super_block *sb, struct fs_context *fc)
 	 * lead to unexpected results.
 	 */
 	sb->s_iflags |= SB_I_NOUMASK;
-	sb->s_iflags |= SB_I_EVM_UNSUPPORTED;
+	sb->s_iflags |= SB_I_EVM_HMAC_UNSUPPORTED;
 
 	err = -ENOMEM;
 	root_dentry = ovl_get_root(sb, ctx->upper.dentry, oe);
@@ -1511,7 +1503,7 @@ static int __init ovl_init(void)
 	ovl_inode_cachep = kmem_cache_create("ovl_inode",
 					     sizeof(struct ovl_inode), 0,
 					     (SLAB_RECLAIM_ACCOUNT|
-					      SLAB_MEM_SPREAD|SLAB_ACCOUNT),
+					      SLAB_ACCOUNT),
 					     ovl_inode_init_once);
 	if (ovl_inode_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/overlayfs/util.c b/fs/overlayfs/util.c
index a8e17f14d7..edc9216f6e 100644
--- a/fs/overlayfs/util.c
+++ b/fs/overlayfs/util.c
@@ -774,13 +774,14 @@ bool ovl_init_uuid_xattr(struct super_block *sb, struct ovl_fs *ofs,
 			 const struct path *upperpath)
 {
 	bool set = false;
+	uuid_t uuid;
 	int res;
 
 	/* Try to load existing persistent uuid */
-	res = ovl_path_getxattr(ofs, upperpath, OVL_XATTR_UUID, sb->s_uuid.b,
+	res = ovl_path_getxattr(ofs, upperpath, OVL_XATTR_UUID, uuid.b,
 				UUID_SIZE);
 	if (res == UUID_SIZE)
-		return true;
+		goto set_uuid;
 
 	if (res != -ENODATA)
 		goto fail;
@@ -808,17 +809,20 @@ bool ovl_init_uuid_xattr(struct super_block *sb, struct ovl_fs *ofs,
 	}
 
 	/* Generate overlay instance uuid */
-	uuid_gen(&sb->s_uuid);
+	uuid_gen(&uuid);
 
 	/* Try to store persistent uuid */
 	set = true;
-	res = ovl_setxattr(ofs, upperpath->dentry, OVL_XATTR_UUID, sb->s_uuid.b,
+	res = ovl_setxattr(ofs, upperpath->dentry, OVL_XATTR_UUID, uuid.b,
 			   UUID_SIZE);
-	if (res == 0)
-		return true;
+	if (res)
+		goto fail;
+
+set_uuid:
+	super_set_uuid(sb, uuid.b, sizeof(uuid));
+	return true;
 
 fail:
-	memset(sb->s_uuid.b, 0, UUID_SIZE);
 	ofs->config.uuid = OVL_UUID_NULL;
 	pr_warn("failed to %s uuid (%pd2, err=%i); falling back to uuid=null.\n",
 		set ? "set" : "get", upperpath->dentry, res);
@@ -1372,7 +1376,7 @@ int ovl_ensure_verity_loaded(struct path *datapath)
 		 * If this inode was not yet opened, the verity info hasn't been
 		 * loaded yet, so we need to do that here to force it into memory.
 		 */
-		filp = kernel_file_open(datapath, O_RDONLY, inode, current_cred());
+		filp = kernel_file_open(datapath, O_RDONLY, current_cred());
 		if (IS_ERR(filp))
 			return PTR_ERR(filp);
 		fput(filp);
diff --git a/fs/pidfs.c b/fs/pidfs.c
new file mode 100644
index 0000000000..dbb9d854d1
--- /dev/null
+++ b/fs/pidfs.c
@@ -0,0 +1,303 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <linux/anon_inodes.h>
+#include <linux/file.h>
+#include <linux/fs.h>
+#include <linux/magic.h>
+#include <linux/mount.h>
+#include <linux/pid.h>
+#include <linux/pidfs.h>
+#include <linux/pid_namespace.h>
+#include <linux/poll.h>
+#include <linux/proc_fs.h>
+#include <linux/proc_ns.h>
+#include <linux/pseudo_fs.h>
+#include <linux/seq_file.h>
+#include <uapi/linux/pidfd.h>
+
+#include "internal.h"
+
+#ifdef CONFIG_PROC_FS
+/**
+ * pidfd_show_fdinfo - print information about a pidfd
+ * @m: proc fdinfo file
+ * @f: file referencing a pidfd
+ *
+ * Pid:
+ * This function will print the pid that a given pidfd refers to in the
+ * pid namespace of the procfs instance.
+ * If the pid namespace of the process is not a descendant of the pid
+ * namespace of the procfs instance 0 will be shown as its pid. This is
+ * similar to calling getppid() on a process whose parent is outside of
+ * its pid namespace.
+ *
+ * NSpid:
+ * If pid namespaces are supported then this function will also print
+ * the pid of a given pidfd refers to for all descendant pid namespaces
+ * starting from the current pid namespace of the instance, i.e. the
+ * Pid field and the first entry in the NSpid field will be identical.
+ * If the pid namespace of the process is not a descendant of the pid
+ * namespace of the procfs instance 0 will be shown as its first NSpid
+ * entry and no others will be shown.
+ * Note that this differs from the Pid and NSpid fields in
+ * /proc/<pid>/status where Pid and NSpid are always shown relative to
+ * the  pid namespace of the procfs instance. The difference becomes
+ * obvious when sending around a pidfd between pid namespaces from a
+ * different branch of the tree, i.e. where no ancestral relation is
+ * present between the pid namespaces:
+ * - create two new pid namespaces ns1 and ns2 in the initial pid
+ *   namespace (also take care to create new mount namespaces in the
+ *   new pid namespace and mount procfs)
+ * - create a process with a pidfd in ns1
+ * - send pidfd from ns1 to ns2
+ * - read /proc/self/fdinfo/<pidfd> and observe that both Pid and NSpid
+ *   have exactly one entry, which is 0
+ */
+static void pidfd_show_fdinfo(struct seq_file *m, struct file *f)
+{
+	struct pid *pid = pidfd_pid(f);
+	struct pid_namespace *ns;
+	pid_t nr = -1;
+
+	if (likely(pid_has_task(pid, PIDTYPE_PID))) {
+		ns = proc_pid_ns(file_inode(m->file)->i_sb);
+		nr = pid_nr_ns(pid, ns);
+	}
+
+	seq_put_decimal_ll(m, "Pid:\t", nr);
+
+#ifdef CONFIG_PID_NS
+	seq_put_decimal_ll(m, "\nNSpid:\t", nr);
+	if (nr > 0) {
+		int i;
+
+		/* If nr is non-zero it means that 'pid' is valid and that
+		 * ns, i.e. the pid namespace associated with the procfs
+		 * instance, is in the pid namespace hierarchy of pid.
+		 * Start at one below the already printed level.
+		 */
+		for (i = ns->level + 1; i <= pid->level; i++)
+			seq_put_decimal_ll(m, "\t", pid->numbers[i].nr);
+	}
+#endif
+	seq_putc(m, '\n');
+}
+#endif
+
+/*
+ * Poll support for process exit notification.
+ */
+static __poll_t pidfd_poll(struct file *file, struct poll_table_struct *pts)
+{
+	struct pid *pid = pidfd_pid(file);
+	bool thread = file->f_flags & PIDFD_THREAD;
+	struct task_struct *task;
+	__poll_t poll_flags = 0;
+
+	poll_wait(file, &pid->wait_pidfd, pts);
+	/*
+	 * Depending on PIDFD_THREAD, inform pollers when the thread
+	 * or the whole thread-group exits.
+	 */
+	guard(rcu)();
+	task = pid_task(pid, PIDTYPE_PID);
+	if (!task)
+		poll_flags = EPOLLIN | EPOLLRDNORM | EPOLLHUP;
+	else if (task->exit_state && (thread || thread_group_empty(task)))
+		poll_flags = EPOLLIN | EPOLLRDNORM;
+
+	return poll_flags;
+}
+
+static const struct file_operations pidfs_file_operations = {
+	.poll		= pidfd_poll,
+#ifdef CONFIG_PROC_FS
+	.show_fdinfo	= pidfd_show_fdinfo,
+#endif
+};
+
+struct pid *pidfd_pid(const struct file *file)
+{
+	if (file->f_op != &pidfs_file_operations)
+		return ERR_PTR(-EBADF);
+	return file_inode(file)->i_private;
+}
+
+static struct vfsmount *pidfs_mnt __ro_after_init;
+
+#if BITS_PER_LONG == 32
+/*
+ * Provide a fallback mechanism for 32-bit systems so processes remain
+ * reliably comparable by inode number even on those systems.
+ */
+static DEFINE_IDA(pidfd_inum_ida);
+
+static int pidfs_inum(struct pid *pid, unsigned long *ino)
+{
+	int ret;
+
+	ret = ida_alloc_range(&pidfd_inum_ida, RESERVED_PIDS + 1,
+			      UINT_MAX, GFP_ATOMIC);
+	if (ret < 0)
+		return -ENOSPC;
+
+	*ino = ret;
+	return 0;
+}
+
+static inline void pidfs_free_inum(unsigned long ino)
+{
+	if (ino > 0)
+		ida_free(&pidfd_inum_ida, ino);
+}
+#else
+static inline int pidfs_inum(struct pid *pid, unsigned long *ino)
+{
+	*ino = pid->ino;
+	return 0;
+}
+#define pidfs_free_inum(ino) ((void)(ino))
+#endif
+
+/*
+ * The vfs falls back to simple_setattr() if i_op->setattr() isn't
+ * implemented. Let's reject it completely until we have a clean
+ * permission concept for pidfds.
+ */
+static int pidfs_setattr(struct mnt_idmap *idmap, struct dentry *dentry,
+			 struct iattr *attr)
+{
+	return -EOPNOTSUPP;
+}
+
+
+/*
+ * User space expects pidfs inodes to have no file type in st_mode.
+ *
+ * In particular, 'lsof' has this legacy logic:
+ *
+ *	type = s->st_mode & S_IFMT;
+ *	switch (type) {
+ *	  ...
+ *	case 0:
+ *		if (!strcmp(p, "anon_inode"))
+ *			Lf->ntype = Ntype = N_ANON_INODE;
+ *
+ * to detect our old anon_inode logic.
+ *
+ * Rather than mess with our internal sane inode data, just fix it
+ * up here in getattr() by masking off the format bits.
+ */
+static int pidfs_getattr(struct mnt_idmap *idmap, const struct path *path,
+			 struct kstat *stat, u32 request_mask,
+			 unsigned int query_flags)
+{
+	struct inode *inode = d_inode(path->dentry);
+
+	generic_fillattr(&nop_mnt_idmap, request_mask, inode, stat);
+	stat->mode &= ~S_IFMT;
+	return 0;
+}
+
+static const struct inode_operations pidfs_inode_operations = {
+	.getattr = pidfs_getattr,
+	.setattr = pidfs_setattr,
+};
+
+static void pidfs_evict_inode(struct inode *inode)
+{
+	struct pid *pid = inode->i_private;
+
+	clear_inode(inode);
+	put_pid(pid);
+	pidfs_free_inum(inode->i_ino);
+}
+
+static const struct super_operations pidfs_sops = {
+	.drop_inode	= generic_delete_inode,
+	.evict_inode	= pidfs_evict_inode,
+	.statfs		= simple_statfs,
+};
+
+/*
+ * 'lsof' has knowledge of out historical anon_inode use, and expects
+ * the pidfs dentry name to start with 'anon_inode'.
+ */
+static char *pidfs_dname(struct dentry *dentry, char *buffer, int buflen)
+{
+	return dynamic_dname(buffer, buflen, "anon_inode:[pidfd]");
+}
+
+static const struct dentry_operations pidfs_dentry_operations = {
+	.d_delete	= always_delete_dentry,
+	.d_dname	= pidfs_dname,
+	.d_prune	= stashed_dentry_prune,
+};
+
+static int pidfs_init_inode(struct inode *inode, void *data)
+{
+	inode->i_private = data;
+	inode->i_flags |= S_PRIVATE;
+	inode->i_mode |= S_IRWXU;
+	inode->i_op = &pidfs_inode_operations;
+	inode->i_fop = &pidfs_file_operations;
+	/*
+	 * Inode numbering for pidfs start at RESERVED_PIDS + 1. This
+	 * avoids collisions with the root inode which is 1 for pseudo
+	 * filesystems.
+	 */
+	return pidfs_inum(data, &inode->i_ino);
+}
+
+static void pidfs_put_data(void *data)
+{
+	struct pid *pid = data;
+	put_pid(pid);
+}
+
+static const struct stashed_operations pidfs_stashed_ops = {
+	.init_inode = pidfs_init_inode,
+	.put_data = pidfs_put_data,
+};
+
+static int pidfs_init_fs_context(struct fs_context *fc)
+{
+	struct pseudo_fs_context *ctx;
+
+	ctx = init_pseudo(fc, PID_FS_MAGIC);
+	if (!ctx)
+		return -ENOMEM;
+
+	ctx->ops = &pidfs_sops;
+	ctx->dops = &pidfs_dentry_operations;
+	fc->s_fs_info = (void *)&pidfs_stashed_ops;
+	return 0;
+}
+
+static struct file_system_type pidfs_type = {
+	.name			= "pidfs",
+	.init_fs_context	= pidfs_init_fs_context,
+	.kill_sb		= kill_anon_super,
+};
+
+struct file *pidfs_alloc_file(struct pid *pid, unsigned int flags)
+{
+
+	struct file *pidfd_file;
+	struct path path;
+	int ret;
+
+	ret = path_from_stashed(&pid->stashed, pidfs_mnt, get_pid(pid), &path);
+	if (ret < 0)
+		return ERR_PTR(ret);
+
+	pidfd_file = dentry_open(&path, flags, current_cred());
+	path_put(&path);
+	return pidfd_file;
+}
+
+void __init pidfs_init(void)
+{
+	pidfs_mnt = kern_mount(&pidfs_type);
+	if (IS_ERR(pidfs_mnt))
+		panic("Failed to mount pidfs pseudo filesystem");
+}
diff --git a/fs/pipe.c b/fs/pipe.c
index f1adbfe743..50c8a8596b 100644
--- a/fs/pipe.c
+++ b/fs/pipe.c
@@ -76,18 +76,20 @@ static unsigned long pipe_user_pages_soft = PIPE_DEF_BUFFERS * INR_OPEN_CUR;
  * -- Manfred Spraul <manfred@colorfullife.com> 2002-05-09
  */
 
-static void pipe_lock_nested(struct pipe_inode_info *pipe, int subclass)
+#define cmp_int(l, r)		((l > r) - (l < r))
+
+#ifdef CONFIG_PROVE_LOCKING
+static int pipe_lock_cmp_fn(const struct lockdep_map *a,
+			    const struct lockdep_map *b)
 {
-	if (pipe->files)
-		mutex_lock_nested(&pipe->mutex, subclass);
+	return cmp_int((unsigned long) a, (unsigned long) b);
 }
+#endif
 
 void pipe_lock(struct pipe_inode_info *pipe)
 {
-	/*
-	 * pipe_lock() nests non-pipe inode locks (for writing to a file)
-	 */
-	pipe_lock_nested(pipe, I_MUTEX_PARENT);
+	if (pipe->files)
+		mutex_lock(&pipe->mutex);
 }
 EXPORT_SYMBOL(pipe_lock);
 
@@ -98,28 +100,16 @@ void pipe_unlock(struct pipe_inode_info *pipe)
 }
 EXPORT_SYMBOL(pipe_unlock);
 
-static inline void __pipe_lock(struct pipe_inode_info *pipe)
-{
-	mutex_lock_nested(&pipe->mutex, I_MUTEX_PARENT);
-}
-
-static inline void __pipe_unlock(struct pipe_inode_info *pipe)
-{
-	mutex_unlock(&pipe->mutex);
-}
-
 void pipe_double_lock(struct pipe_inode_info *pipe1,
 		      struct pipe_inode_info *pipe2)
 {
 	BUG_ON(pipe1 == pipe2);
 
-	if (pipe1 < pipe2) {
-		pipe_lock_nested(pipe1, I_MUTEX_PARENT);
-		pipe_lock_nested(pipe2, I_MUTEX_CHILD);
-	} else {
-		pipe_lock_nested(pipe2, I_MUTEX_PARENT);
-		pipe_lock_nested(pipe1, I_MUTEX_CHILD);
-	}
+	if (pipe1 > pipe2)
+		swap(pipe1, pipe2);
+
+	pipe_lock(pipe1);
+	pipe_lock(pipe2);
 }
 
 static void anon_pipe_buf_release(struct pipe_inode_info *pipe,
@@ -271,7 +261,7 @@ pipe_read(struct kiocb *iocb, struct iov_iter *to)
 		return 0;
 
 	ret = 0;
-	__pipe_lock(pipe);
+	mutex_lock(&pipe->mutex);
 
 	/*
 	 * We only wake up writers if the pipe was full when we started
@@ -368,7 +358,7 @@ pipe_read(struct kiocb *iocb, struct iov_iter *to)
 			ret = -EAGAIN;
 			break;
 		}
-		__pipe_unlock(pipe);
+		mutex_unlock(&pipe->mutex);
 
 		/*
 		 * We only get here if we didn't actually read anything.
@@ -400,13 +390,13 @@ pipe_read(struct kiocb *iocb, struct iov_iter *to)
 		if (wait_event_interruptible_exclusive(pipe->rd_wait, pipe_readable(pipe)) < 0)
 			return -ERESTARTSYS;
 
-		__pipe_lock(pipe);
+		mutex_lock(&pipe->mutex);
 		was_full = pipe_full(pipe->head, pipe->tail, pipe->max_usage);
 		wake_next_reader = true;
 	}
 	if (pipe_empty(pipe->head, pipe->tail))
 		wake_next_reader = false;
-	__pipe_unlock(pipe);
+	mutex_unlock(&pipe->mutex);
 
 	if (was_full)
 		wake_up_interruptible_sync_poll(&pipe->wr_wait, EPOLLOUT | EPOLLWRNORM);
@@ -462,7 +452,7 @@ pipe_write(struct kiocb *iocb, struct iov_iter *from)
 	if (unlikely(total_len == 0))
 		return 0;
 
-	__pipe_lock(pipe);
+	mutex_lock(&pipe->mutex);
 
 	if (!pipe->readers) {
 		send_sig(SIGPIPE, current, 0);
@@ -582,19 +572,19 @@ pipe_write(struct kiocb *iocb, struct iov_iter *from)
 		 * after waiting we need to re-check whether the pipe
 		 * become empty while we dropped the lock.
 		 */
-		__pipe_unlock(pipe);
+		mutex_unlock(&pipe->mutex);
 		if (was_empty)
 			wake_up_interruptible_sync_poll(&pipe->rd_wait, EPOLLIN | EPOLLRDNORM);
 		kill_fasync(&pipe->fasync_readers, SIGIO, POLL_IN);
 		wait_event_interruptible_exclusive(pipe->wr_wait, pipe_writable(pipe));
-		__pipe_lock(pipe);
+		mutex_lock(&pipe->mutex);
 		was_empty = pipe_empty(pipe->head, pipe->tail);
 		wake_next_writer = true;
 	}
 out:
 	if (pipe_full(pipe->head, pipe->tail, pipe->max_usage))
 		wake_next_writer = false;
-	__pipe_unlock(pipe);
+	mutex_unlock(&pipe->mutex);
 
 	/*
 	 * If we do do a wakeup event, we do a 'sync' wakeup, because we
@@ -629,7 +619,7 @@ static long pipe_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 
 	switch (cmd) {
 	case FIONREAD:
-		__pipe_lock(pipe);
+		mutex_lock(&pipe->mutex);
 		count = 0;
 		head = pipe->head;
 		tail = pipe->tail;
@@ -639,16 +629,16 @@ static long pipe_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 			count += pipe->bufs[tail & mask].len;
 			tail++;
 		}
-		__pipe_unlock(pipe);
+		mutex_unlock(&pipe->mutex);
 
 		return put_user(count, (int __user *)arg);
 
 #ifdef CONFIG_WATCH_QUEUE
 	case IOC_WATCH_QUEUE_SET_SIZE: {
 		int ret;
-		__pipe_lock(pipe);
+		mutex_lock(&pipe->mutex);
 		ret = watch_queue_set_size(pipe, arg);
-		__pipe_unlock(pipe);
+		mutex_unlock(&pipe->mutex);
 		return ret;
 	}
 
@@ -734,7 +724,7 @@ pipe_release(struct inode *inode, struct file *file)
 {
 	struct pipe_inode_info *pipe = file->private_data;
 
-	__pipe_lock(pipe);
+	mutex_lock(&pipe->mutex);
 	if (file->f_mode & FMODE_READ)
 		pipe->readers--;
 	if (file->f_mode & FMODE_WRITE)
@@ -747,7 +737,7 @@ pipe_release(struct inode *inode, struct file *file)
 		kill_fasync(&pipe->fasync_readers, SIGIO, POLL_IN);
 		kill_fasync(&pipe->fasync_writers, SIGIO, POLL_OUT);
 	}
-	__pipe_unlock(pipe);
+	mutex_unlock(&pipe->mutex);
 
 	put_pipe_info(inode, pipe);
 	return 0;
@@ -759,7 +749,7 @@ pipe_fasync(int fd, struct file *filp, int on)
 	struct pipe_inode_info *pipe = filp->private_data;
 	int retval = 0;
 
-	__pipe_lock(pipe);
+	mutex_lock(&pipe->mutex);
 	if (filp->f_mode & FMODE_READ)
 		retval = fasync_helper(fd, filp, on, &pipe->fasync_readers);
 	if ((filp->f_mode & FMODE_WRITE) && retval >= 0) {
@@ -768,7 +758,7 @@ pipe_fasync(int fd, struct file *filp, int on)
 			/* this can happen only if on == T */
 			fasync_helper(-1, filp, 0, &pipe->fasync_readers);
 	}
-	__pipe_unlock(pipe);
+	mutex_unlock(&pipe->mutex);
 	return retval;
 }
 
@@ -834,6 +824,7 @@ struct pipe_inode_info *alloc_pipe_info(void)
 		pipe->nr_accounted = pipe_bufs;
 		pipe->user = user;
 		mutex_init(&pipe->mutex);
+		lock_set_cmp_fn(&pipe->mutex, pipe_lock_cmp_fn, NULL);
 		return pipe;
 	}
 
@@ -1144,7 +1135,7 @@ static int fifo_open(struct inode *inode, struct file *filp)
 	filp->private_data = pipe;
 	/* OK, we have a pipe and it's pinned down */
 
-	__pipe_lock(pipe);
+	mutex_lock(&pipe->mutex);
 
 	/* We can only do regular read/write on fifos */
 	stream_open(inode, filp);
@@ -1214,7 +1205,7 @@ static int fifo_open(struct inode *inode, struct file *filp)
 	}
 
 	/* Ok! */
-	__pipe_unlock(pipe);
+	mutex_unlock(&pipe->mutex);
 	return 0;
 
 err_rd:
@@ -1230,7 +1221,7 @@ err_wr:
 	goto err;
 
 err:
-	__pipe_unlock(pipe);
+	mutex_unlock(&pipe->mutex);
 
 	put_pipe_info(inode, pipe);
 	return ret;
@@ -1411,7 +1402,7 @@ long pipe_fcntl(struct file *file, unsigned int cmd, unsigned int arg)
 	if (!pipe)
 		return -EBADF;
 
-	__pipe_lock(pipe);
+	mutex_lock(&pipe->mutex);
 
 	switch (cmd) {
 	case F_SETPIPE_SZ:
@@ -1425,7 +1416,7 @@ long pipe_fcntl(struct file *file, unsigned int cmd, unsigned int arg)
 		break;
 	}
 
-	__pipe_unlock(pipe);
+	mutex_unlock(&pipe->mutex);
 	return ret;
 }
 
diff --git a/fs/posix_acl.c b/fs/posix_acl.c
index e1af20893e..3f87297dbf 100644
--- a/fs/posix_acl.c
+++ b/fs/posix_acl.c
@@ -26,7 +26,6 @@
 #include <linux/mnt_idmapping.h>
 #include <linux/iversion.h>
 #include <linux/security.h>
-#include <linux/evm.h>
 #include <linux/fsnotify.h>
 #include <linux/filelock.h>
 
@@ -786,12 +785,12 @@ struct posix_acl *posix_acl_from_xattr(struct user_namespace *userns,
 		return ERR_PTR(count);
 	if (count == 0)
 		return NULL;
-	
+
 	acl = posix_acl_alloc(count, GFP_NOFS);
 	if (!acl)
 		return ERR_PTR(-ENOMEM);
 	acl_e = acl->a_entries;
-	
+
 	for (end = entry + count; entry != end; acl_e++, entry++) {
 		acl_e->e_tag  = le16_to_cpu(entry->e_tag);
 		acl_e->e_perm = le16_to_cpu(entry->e_perm);
@@ -1137,7 +1136,7 @@ retry_deleg:
 		error = -EIO;
 	if (!error) {
 		fsnotify_xattr(dentry);
-		evm_inode_post_set_acl(dentry, acl_name, kacl);
+		security_inode_post_set_acl(dentry, acl_name, kacl);
 	}
 
 out_inode_unlock:
@@ -1245,7 +1244,7 @@ retry_deleg:
 		error = -EIO;
 	if (!error) {
 		fsnotify_xattr(dentry);
-		evm_inode_post_remove_acl(idmap, dentry, acl_name);
+		security_inode_post_remove_acl(idmap, dentry, acl_name);
 	}
 
 out_inode_unlock:
diff --git a/fs/proc/Kconfig b/fs/proc/Kconfig
index 32b1116ae1..d80a1431ef 100644
--- a/fs/proc/Kconfig
+++ b/fs/proc/Kconfig
@@ -32,7 +32,7 @@ config PROC_FS
 config PROC_KCORE
 	bool "/proc/kcore support" if !ARM
 	depends on PROC_FS && MMU
-	select CRASH_CORE
+	select VMCORE_INFO
 	help
 	  Provides a virtual ELF core file of the live kernel.  This can
 	  be read with gdb and other ELF tools.  No modifications can be
diff --git a/fs/proc/base.c b/fs/proc/base.c
index 18550c071d..72a1acd036 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -3214,7 +3214,7 @@ static int proc_pid_ksm_stat(struct seq_file *m, struct pid_namespace *ns,
 	mm = get_task_mm(task);
 	if (mm) {
 		seq_printf(m, "ksm_rmap_items %lu\n", mm->ksm_rmap_items);
-		seq_printf(m, "ksm_zero_pages %lu\n", mm->ksm_zero_pages);
+		seq_printf(m, "ksm_zero_pages %ld\n", mm_ksm_zero_pages(mm));
 		seq_printf(m, "ksm_merging_pages %lu\n", mm->ksm_merging_pages);
 		seq_printf(m, "ksm_process_profit %ld\n", ksm_process_profit(mm));
 		mmput(mm);
diff --git a/fs/proc/fd.c b/fs/proc/fd.c
index 6e72e5ad42..586bbc84ca 100644
--- a/fs/proc/fd.c
+++ b/fs/proc/fd.c
@@ -39,10 +39,8 @@ static int seq_show(struct seq_file *m, void *v)
 		spin_lock(&files->file_lock);
 		file = files_lookup_fd_locked(files, fd);
 		if (file) {
-			struct fdtable *fdt = files_fdtable(files);
-
 			f_flags = file->f_flags;
-			if (close_on_exec(fd, fdt))
+			if (close_on_exec(fd, files))
 				f_flags |= O_CLOEXEC;
 
 			get_file(file);
@@ -74,7 +72,18 @@ out:
 	return 0;
 }
 
-static int proc_fdinfo_access_allowed(struct inode *inode)
+static int seq_fdinfo_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, seq_show, inode);
+}
+
+/**
+ * Shared /proc/pid/fdinfo and /proc/pid/fdinfo/fd permission helper to ensure
+ * that the current task has PTRACE_MODE_READ in addition to the normal
+ * POSIX-like checks.
+ */
+static int proc_fdinfo_permission(struct mnt_idmap *idmap, struct inode *inode,
+				  int mask)
 {
 	bool allowed = false;
 	struct task_struct *task = get_proc_task(inode);
@@ -88,18 +97,13 @@ static int proc_fdinfo_access_allowed(struct inode *inode)
 	if (!allowed)
 		return -EACCES;
 
-	return 0;
+	return generic_permission(idmap, inode, mask);
 }
 
-static int seq_fdinfo_open(struct inode *inode, struct file *file)
-{
-	int ret = proc_fdinfo_access_allowed(inode);
-
-	if (ret)
-		return ret;
-
-	return single_open(file, seq_show, inode);
-}
+static const struct inode_operations proc_fdinfo_file_inode_operations = {
+	.permission	= proc_fdinfo_permission,
+	.setattr	= proc_setattr,
+};
 
 static const struct file_operations proc_fdinfo_file_operations = {
 	.open		= seq_fdinfo_open,
@@ -388,6 +392,8 @@ static struct dentry *proc_fdinfo_instantiate(struct dentry *dentry,
 	ei = PROC_I(inode);
 	ei->fd = data->fd;
 
+	inode->i_op = &proc_fdinfo_file_inode_operations;
+
 	inode->i_fop = &proc_fdinfo_file_operations;
 	tid_fd_update_inode(task, inode, 0);
 
@@ -407,23 +413,13 @@ static int proc_readfdinfo(struct file *file, struct dir_context *ctx)
 				  proc_fdinfo_instantiate);
 }
 
-static int proc_open_fdinfo(struct inode *inode, struct file *file)
-{
-	int ret = proc_fdinfo_access_allowed(inode);
-
-	if (ret)
-		return ret;
-
-	return 0;
-}
-
 const struct inode_operations proc_fdinfo_inode_operations = {
 	.lookup		= proc_lookupfdinfo,
+	.permission	= proc_fdinfo_permission,
 	.setattr	= proc_setattr,
 };
 
 const struct file_operations proc_fdinfo_operations = {
-	.open		= proc_open_fdinfo,
 	.read		= generic_read_dir,
 	.iterate_shared	= proc_readfdinfo,
 	.llseek		= generic_file_llseek,
diff --git a/fs/proc/inode.c b/fs/proc/inode.c
index 05350f3c28..d19434e2a5 100644
--- a/fs/proc/inode.c
+++ b/fs/proc/inode.c
@@ -92,7 +92,7 @@ void __init proc_init_kmemcache(void)
 	proc_inode_cachep = kmem_cache_create("proc_inode_cache",
 					     sizeof(struct proc_inode),
 					     0, (SLAB_RECLAIM_ACCOUNT|
-						SLAB_MEM_SPREAD|SLAB_ACCOUNT|
+						SLAB_ACCOUNT|
 						SLAB_PANIC),
 					     init_once);
 	pde_opener_cache =
@@ -451,15 +451,13 @@ pde_get_unmapped_area(struct proc_dir_entry *pde, struct file *file, unsigned lo
 			   unsigned long len, unsigned long pgoff,
 			   unsigned long flags)
 {
-	typeof_member(struct proc_ops, proc_get_unmapped_area) get_area;
+	if (pde->proc_ops->proc_get_unmapped_area)
+		return pde->proc_ops->proc_get_unmapped_area(file, orig_addr, len, pgoff, flags);
 
-	get_area = pde->proc_ops->proc_get_unmapped_area;
 #ifdef CONFIG_MMU
-	if (!get_area)
-		get_area = current->mm->get_unmapped_area;
+	return mm_get_unmapped_area(current->mm, file, orig_addr, len, pgoff, flags);
 #endif
-	if (get_area)
-		return get_area(file, orig_addr, len, pgoff, flags);
+
 	return orig_addr;
 }
 
diff --git a/fs/proc/kcore.c b/fs/proc/kcore.c
index 6422e569b0..8e08a9a1b7 100644
--- a/fs/proc/kcore.c
+++ b/fs/proc/kcore.c
@@ -10,7 +10,7 @@
  *	Safe accesses to vmalloc/direct-mapped discontiguous areas, Kanoj Sarcar <kanoj@sgi.com>
  */
 
-#include <linux/crash_core.h>
+#include <linux/vmcore_info.h>
 #include <linux/mm.h>
 #include <linux/proc_fs.h>
 #include <linux/kcore.h>
diff --git a/fs/proc/meminfo.c b/fs/proc/meminfo.c
index 45af9a989d..245171d916 100644
--- a/fs/proc/meminfo.c
+++ b/fs/proc/meminfo.c
@@ -89,8 +89,7 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
 	show_val_kb(m, "SwapTotal:      ", i.totalswap);
 	show_val_kb(m, "SwapFree:       ", i.freeswap);
 #ifdef CONFIG_ZSWAP
-	seq_printf(m,  "Zswap:          %8lu kB\n",
-		   (unsigned long)(zswap_pool_total_size >> 10));
+	show_val_kb(m, "Zswap:          ", zswap_total_pages());
 	seq_printf(m,  "Zswapped:       %8lu kB\n",
 		   (unsigned long)atomic_read(&zswap_stored_pages) <<
 		   (PAGE_SHIFT - 10));
diff --git a/fs/proc/page.c b/fs/proc/page.c
index 9223856c93..2fb64bdb64 100644
--- a/fs/proc/page.c
+++ b/fs/proc/page.c
@@ -107,10 +107,13 @@ static inline u64 kpf_copy_bit(u64 kflags, int ubit, int kbit)
 	return ((kflags >> kbit) & 1) << ubit;
 }
 
-u64 stable_page_flags(struct page *page)
+u64 stable_page_flags(const struct page *page)
 {
-	u64 k;
-	u64 u;
+	const struct folio *folio;
+	unsigned long k;
+	unsigned long mapping;
+	bool is_anon;
+	u64 u = 0;
 
 	/*
 	 * pseudo flag: KPF_NOPAGE
@@ -118,49 +121,47 @@ u64 stable_page_flags(struct page *page)
 	 */
 	if (!page)
 		return 1 << KPF_NOPAGE;
+	folio = page_folio(page);
 
-	k = page->flags;
-	u = 0;
+	k = folio->flags;
+	mapping = (unsigned long)folio->mapping;
+	is_anon = mapping & PAGE_MAPPING_ANON;
 
 	/*
 	 * pseudo flags for the well known (anonymous) memory mapped pages
 	 */
 	if (page_mapped(page))
 		u |= 1 << KPF_MMAP;
-	if (PageAnon(page))
+	if (is_anon) {
 		u |= 1 << KPF_ANON;
-	if (PageKsm(page))
-		u |= 1 << KPF_KSM;
+		if (mapping & PAGE_MAPPING_KSM)
+			u |= 1 << KPF_KSM;
+	}
 
 	/*
 	 * compound pages: export both head/tail info
 	 * they together define a compound page's start/end pos and order
 	 */
-	if (PageHead(page))
-		u |= 1 << KPF_COMPOUND_HEAD;
-	if (PageTail(page))
+	if (page == &folio->page)
+		u |= kpf_copy_bit(k, KPF_COMPOUND_HEAD, PG_head);
+	else
 		u |= 1 << KPF_COMPOUND_TAIL;
-	if (PageHuge(page))
+	if (folio_test_hugetlb(folio))
 		u |= 1 << KPF_HUGE;
 	/*
-	 * PageTransCompound can be true for non-huge compound pages (slab
-	 * pages or pages allocated by drivers with __GFP_COMP) because it
-	 * just checks PG_head/PG_tail, so we need to check PageLRU/PageAnon
+	 * We need to check PageLRU/PageAnon
 	 * to make sure a given page is a thp, not a non-huge compound page.
 	 */
-	else if (PageTransCompound(page)) {
-		struct page *head = compound_head(page);
-
-		if (PageLRU(head) || PageAnon(head))
+	else if (folio_test_large(folio)) {
+		if ((k & (1 << PG_lru)) || is_anon)
 			u |= 1 << KPF_THP;
-		else if (is_huge_zero_page(head)) {
+		else if (is_huge_zero_folio(folio)) {
 			u |= 1 << KPF_ZERO_PAGE;
 			u |= 1 << KPF_THP;
 		}
 	} else if (is_zero_pfn(page_to_pfn(page)))
 		u |= 1 << KPF_ZERO_PAGE;
 
-
 	/*
 	 * Caveats on high order pages: PG_buddy and PG_slab will only be set
 	 * on the head page.
@@ -174,16 +175,17 @@ u64 stable_page_flags(struct page *page)
 		u |= 1 << KPF_OFFLINE;
 	if (PageTable(page))
 		u |= 1 << KPF_PGTABLE;
+	if (folio_test_slab(folio))
+		u |= 1 << KPF_SLAB;
 
-	if (page_is_idle(page))
+#if defined(CONFIG_PAGE_IDLE_FLAG) && defined(CONFIG_64BIT)
+	u |= kpf_copy_bit(k, KPF_IDLE,          PG_idle);
+#else
+	if (folio_test_idle(folio))
 		u |= 1 << KPF_IDLE;
+#endif
 
 	u |= kpf_copy_bit(k, KPF_LOCKED,	PG_locked);
-
-	u |= kpf_copy_bit(k, KPF_SLAB,		PG_slab);
-	if (PageTail(page) && PageSlab(page))
-		u |= 1 << KPF_SLAB;
-
 	u |= kpf_copy_bit(k, KPF_ERROR,		PG_error);
 	u |= kpf_copy_bit(k, KPF_DIRTY,		PG_dirty);
 	u |= kpf_copy_bit(k, KPF_UPTODATE,	PG_uptodate);
@@ -194,7 +196,8 @@ u64 stable_page_flags(struct page *page)
 	u |= kpf_copy_bit(k, KPF_ACTIVE,	PG_active);
 	u |= kpf_copy_bit(k, KPF_RECLAIM,	PG_reclaim);
 
-	if (PageSwapCache(page))
+#define SWAPCACHE ((1 << PG_swapbacked) | (1 << PG_swapcache))
+	if ((k & SWAPCACHE) == SWAPCACHE)
 		u |= 1 << KPF_SWAPCACHE;
 	u |= kpf_copy_bit(k, KPF_SWAPBACKED,	PG_swapbacked);
 
@@ -202,7 +205,10 @@ u64 stable_page_flags(struct page *page)
 	u |= kpf_copy_bit(k, KPF_MLOCKED,	PG_mlocked);
 
 #ifdef CONFIG_MEMORY_FAILURE
-	u |= kpf_copy_bit(k, KPF_HWPOISON,	PG_hwpoison);
+	if (u & (1 << KPF_HUGE))
+		u |= kpf_copy_bit(k, KPF_HWPOISON,	PG_hwpoison);
+	else
+		u |= kpf_copy_bit(page->flags, KPF_HWPOISON,	PG_hwpoison);
 #endif
 
 #ifdef CONFIG_ARCH_USES_PG_UNCACHED
@@ -228,7 +234,6 @@ static ssize_t kpageflags_read(struct file *file, char __user *buf,
 {
 	const unsigned long max_dump_pfn = get_max_dump_pfn();
 	u64 __user *out = (u64 __user *)buf;
-	struct page *ppage;
 	unsigned long src = *ppos;
 	unsigned long pfn;
 	ssize_t ret = 0;
@@ -245,9 +250,9 @@ static ssize_t kpageflags_read(struct file *file, char __user *buf,
 		 * TODO: ZONE_DEVICE support requires to identify
 		 * memmaps that were actually initialized.
 		 */
-		ppage = pfn_to_online_page(pfn);
+		struct page *page = pfn_to_online_page(pfn);
 
-		if (put_user(stable_page_flags(ppage), out)) {
+		if (put_user(stable_page_flags(page), out)) {
 			ret = -EFAULT;
 			break;
 		}
diff --git a/fs/proc/proc_net.c b/fs/proc/proc_net.c
index 2ba31b6d68..52f0b75cbc 100644
--- a/fs/proc/proc_net.c
+++ b/fs/proc/proc_net.c
@@ -135,6 +135,7 @@ EXPORT_SYMBOL_GPL(proc_create_net_data);
  * @parent: The parent directory in which to create.
  * @ops: The seq_file ops with which to read the file.
  * @write: The write method with which to 'modify' the file.
+ * @state_size: The size of the per-file private state to allocate.
  * @data: Data for retrieval by pde_data().
  *
  * Create a network namespaced proc file in the @parent directory with the
diff --git a/fs/proc/proc_sysctl.c b/fs/proc/proc_sysctl.c
index 37cde0efee..dd7b462387 100644
--- a/fs/proc/proc_sysctl.c
+++ b/fs/proc/proc_sysctl.c
@@ -30,9 +30,7 @@ static const struct file_operations proc_sys_dir_file_operations;
 static const struct inode_operations proc_sys_dir_operations;
 
 /* Support for permanently empty directories */
-static struct ctl_table sysctl_mount_point[] = {
-	{.type = SYSCTL_TABLE_TYPE_PERMANENTLY_EMPTY }
-};
+static struct ctl_table sysctl_mount_point[] = { };
 
 /**
  * register_sysctl_mount_point() - registers a sysctl mount point
@@ -48,14 +46,12 @@ struct ctl_table_header *register_sysctl_mount_point(const char *path)
 }
 EXPORT_SYMBOL(register_sysctl_mount_point);
 
-#define sysctl_is_perm_empty_ctl_table(tptr)		\
-	(tptr[0].type == SYSCTL_TABLE_TYPE_PERMANENTLY_EMPTY)
 #define sysctl_is_perm_empty_ctl_header(hptr)		\
-	(sysctl_is_perm_empty_ctl_table(hptr->ctl_table))
+	(hptr->type == SYSCTL_TABLE_TYPE_PERMANENTLY_EMPTY)
 #define sysctl_set_perm_empty_ctl_header(hptr)		\
-	(hptr->ctl_table[0].type = SYSCTL_TABLE_TYPE_PERMANENTLY_EMPTY)
+	(hptr->type = SYSCTL_TABLE_TYPE_PERMANENTLY_EMPTY)
 #define sysctl_clear_perm_empty_ctl_header(hptr)	\
-	(hptr->ctl_table[0].type = SYSCTL_TABLE_TYPE_DEFAULT)
+	(hptr->type = SYSCTL_TABLE_TYPE_DEFAULT)
 
 void proc_sys_poll_notify(struct ctl_table_poll *poll)
 {
@@ -210,6 +206,8 @@ static void init_header(struct ctl_table_header *head,
 			node++;
 		}
 	}
+	if (table == sysctl_mount_point)
+		sysctl_set_perm_empty_ctl_header(head);
 }
 
 static void erase_header(struct ctl_table_header *head)
@@ -232,8 +230,7 @@ static int insert_header(struct ctl_dir *dir, struct ctl_table_header *header)
 		return -EROFS;
 
 	/* Am I creating a permanently empty directory? */
-	if (header->ctl_table_size > 0 &&
-	    sysctl_is_perm_empty_ctl_table(header->ctl_table)) {
+	if (sysctl_is_perm_empty_ctl_header(header)) {
 		if (!RB_EMPTY_ROOT(&dir->root))
 			return -EINVAL;
 		sysctl_set_perm_empty_ctl_header(dir_h);
@@ -479,12 +476,10 @@ static struct inode *proc_sys_make_inode(struct super_block *sb,
 			make_empty_dir_inode(inode);
 	}
 
+	inode->i_uid = GLOBAL_ROOT_UID;
+	inode->i_gid = GLOBAL_ROOT_GID;
 	if (root->set_ownership)
-		root->set_ownership(head, table, &inode->i_uid, &inode->i_gid);
-	else {
-		inode->i_uid = GLOBAL_ROOT_UID;
-		inode->i_gid = GLOBAL_ROOT_GID;
-	}
+		root->set_ownership(head, &inode->i_uid, &inode->i_gid);
 
 	return inode;
 }
@@ -1204,7 +1199,7 @@ static bool get_links(struct ctl_dir *dir,
 	struct ctl_table *entry, *link;
 
 	if (header->ctl_table_size == 0 ||
-	    sysctl_is_perm_empty_ctl_table(header->ctl_table))
+	    sysctl_is_perm_empty_ctl_header(header))
 		return true;
 
 	/* Are there links available for every entry in table? */
diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c
index 8433f076e9..a45f2da0ad 100644
--- a/fs/proc/task_mmu.c
+++ b/fs/proc/task_mmu.c
@@ -411,14 +411,14 @@ struct mem_size_stats {
 };
 
 static void smaps_page_accumulate(struct mem_size_stats *mss,
-		struct page *page, unsigned long size, unsigned long pss,
+		struct folio *folio, unsigned long size, unsigned long pss,
 		bool dirty, bool locked, bool private)
 {
 	mss->pss += pss;
 
-	if (PageAnon(page))
+	if (folio_test_anon(folio))
 		mss->pss_anon += pss;
-	else if (PageSwapBacked(page))
+	else if (folio_test_swapbacked(folio))
 		mss->pss_shmem += pss;
 	else
 		mss->pss_file += pss;
@@ -426,7 +426,7 @@ static void smaps_page_accumulate(struct mem_size_stats *mss,
 	if (locked)
 		mss->pss_locked += pss;
 
-	if (dirty || PageDirty(page)) {
+	if (dirty || folio_test_dirty(folio)) {
 		mss->pss_dirty += pss;
 		if (private)
 			mss->private_dirty += size;
@@ -444,6 +444,7 @@ static void smaps_account(struct mem_size_stats *mss, struct page *page,
 		bool compound, bool young, bool dirty, bool locked,
 		bool migration)
 {
+	struct folio *folio = page_folio(page);
 	int i, nr = compound ? compound_nr(page) : 1;
 	unsigned long size = nr * PAGE_SIZE;
 
@@ -451,27 +452,28 @@ static void smaps_account(struct mem_size_stats *mss, struct page *page,
 	 * First accumulate quantities that depend only on |size| and the type
 	 * of the compound page.
 	 */
-	if (PageAnon(page)) {
+	if (folio_test_anon(folio)) {
 		mss->anonymous += size;
-		if (!PageSwapBacked(page) && !dirty && !PageDirty(page))
+		if (!folio_test_swapbacked(folio) && !dirty &&
+		    !folio_test_dirty(folio))
 			mss->lazyfree += size;
 	}
 
-	if (PageKsm(page))
+	if (folio_test_ksm(folio))
 		mss->ksm += size;
 
 	mss->resident += size;
 	/* Accumulate the size in pages that have been accessed. */
-	if (young || page_is_young(page) || PageReferenced(page))
+	if (young || folio_test_young(folio) || folio_test_referenced(folio))
 		mss->referenced += size;
 
 	/*
 	 * Then accumulate quantities that may depend on sharing, or that may
 	 * differ page-by-page.
 	 *
-	 * page_count(page) == 1 guarantees the page is mapped exactly once.
+	 * refcount == 1 guarantees the page is mapped exactly once.
 	 * If any subpage of the compound page mapped with PTE it would elevate
-	 * page_count().
+	 * the refcount.
 	 *
 	 * The page_mapcount() is called to get a snapshot of the mapcount.
 	 * Without holding the page lock this snapshot can be slightly wrong as
@@ -480,9 +482,9 @@ static void smaps_account(struct mem_size_stats *mss, struct page *page,
 	 * especially for migration entries.  Treat regular migration entries
 	 * as mapcount == 1.
 	 */
-	if ((page_count(page) == 1) || migration) {
-		smaps_page_accumulate(mss, page, size, size << PSS_SHIFT, dirty,
-			locked, true);
+	if ((folio_ref_count(folio) == 1) || migration) {
+		smaps_page_accumulate(mss, folio, size, size << PSS_SHIFT,
+				dirty, locked, true);
 		return;
 	}
 	for (i = 0; i < nr; i++, page++) {
@@ -490,8 +492,8 @@ static void smaps_account(struct mem_size_stats *mss, struct page *page,
 		unsigned long pss = PAGE_SIZE << PSS_SHIFT;
 		if (mapcount >= 2)
 			pss /= mapcount;
-		smaps_page_accumulate(mss, page, PAGE_SIZE, pss, dirty, locked,
-				      mapcount < 2);
+		smaps_page_accumulate(mss, folio, PAGE_SIZE, pss,
+				dirty, locked, mapcount < 2);
 	}
 }
 
@@ -576,6 +578,7 @@ static void smaps_pmd_entry(pmd_t *pmd, unsigned long addr,
 	struct vm_area_struct *vma = walk->vma;
 	bool locked = !!(vma->vm_flags & VM_LOCKED);
 	struct page *page = NULL;
+	struct folio *folio;
 	bool migration = false;
 
 	if (pmd_present(*pmd)) {
@@ -590,11 +593,12 @@ static void smaps_pmd_entry(pmd_t *pmd, unsigned long addr,
 	}
 	if (IS_ERR_OR_NULL(page))
 		return;
-	if (PageAnon(page))
+	folio = page_folio(page);
+	if (folio_test_anon(folio))
 		mss->anonymous_thp += HPAGE_PMD_SIZE;
-	else if (PageSwapBacked(page))
+	else if (folio_test_swapbacked(folio))
 		mss->shmem_thp += HPAGE_PMD_SIZE;
-	else if (is_zone_device_page(page))
+	else if (folio_is_zone_device(folio))
 		/* pass */;
 	else
 		mss->file_thp += HPAGE_PMD_SIZE;
@@ -703,6 +707,9 @@ static void show_smap_vma_flags(struct seq_file *m, struct vm_area_struct *vma)
 #ifdef CONFIG_X86_USER_SHADOW_STACK
 		[ilog2(VM_SHADOW_STACK)] = "ss",
 #endif
+#ifdef CONFIG_64BIT
+		[ilog2(VM_SEALED)] = "sl",
+#endif
 	};
 	size_t i;
 
@@ -726,19 +733,20 @@ static int smaps_hugetlb_range(pte_t *pte, unsigned long hmask,
 {
 	struct mem_size_stats *mss = walk->private;
 	struct vm_area_struct *vma = walk->vma;
-	struct page *page = NULL;
-	pte_t ptent = ptep_get(pte);
+	pte_t ptent = huge_ptep_get(pte);
+	struct folio *folio = NULL;
 
 	if (pte_present(ptent)) {
-		page = vm_normal_page(vma, addr, ptent);
+		folio = page_folio(pte_page(ptent));
 	} else if (is_swap_pte(ptent)) {
 		swp_entry_t swpent = pte_to_swp_entry(ptent);
 
 		if (is_pfn_swap_entry(swpent))
-			page = pfn_swap_entry_to_page(swpent);
+			folio = pfn_swap_entry_folio(swpent);
 	}
-	if (page) {
-		if (page_mapcount(page) >= 2 || hugetlb_pmd_shared(pte))
+	if (folio) {
+		if (folio_likely_mapped_shared(folio) ||
+		    hugetlb_pmd_shared(pte))
 			mss->shared_hugetlb += huge_page_size(hstate_vma(vma));
 		else
 			mss->private_hugetlb += huge_page_size(hstate_vma(vma));
@@ -866,8 +874,8 @@ static int show_smap(struct seq_file *m, void *v)
 	__show_smap(m, &mss, false);
 
 	seq_printf(m, "THPeligible:    %8u\n",
-		   !!thp_vma_allowable_orders(vma, vma->vm_flags, true, false,
-					      true, THP_ORDERS_ALL));
+		   !!thp_vma_allowable_orders(vma, vma->vm_flags,
+			   TVA_SMAPS | TVA_ENFORCE_SYSFS, THP_ORDERS_ALL));
 
 	if (arch_pkeys_enabled())
 		seq_printf(m, "ProtectionKey:  %8u\n", vma_pkey(vma));
@@ -965,12 +973,17 @@ static int show_smaps_rollup(struct seq_file *m, void *v)
 				break;
 
 			/* Case 1 and 2 above */
-			if (vma->vm_start >= last_vma_end)
+			if (vma->vm_start >= last_vma_end) {
+				smap_gather_stats(vma, &mss, 0);
+				last_vma_end = vma->vm_end;
 				continue;
+			}
 
 			/* Case 4 above */
-			if (vma->vm_end > last_vma_end)
+			if (vma->vm_end > last_vma_end) {
 				smap_gather_stats(vma, &mss, last_vma_end);
+				last_vma_end = vma->vm_end;
+			}
 		}
 	} for_each_vma(vmi, vma);
 
@@ -1161,7 +1174,7 @@ static int clear_refs_pte_range(pmd_t *pmd, unsigned long addr,
 	struct vm_area_struct *vma = walk->vma;
 	pte_t *pte, ptent;
 	spinlock_t *ptl;
-	struct page *page;
+	struct folio *folio;
 
 	ptl = pmd_trans_huge_lock(pmd, vma);
 	if (ptl) {
@@ -1173,12 +1186,12 @@ static int clear_refs_pte_range(pmd_t *pmd, unsigned long addr,
 		if (!pmd_present(*pmd))
 			goto out;
 
-		page = pmd_page(*pmd);
+		folio = pmd_folio(*pmd);
 
 		/* Clear accessed and referenced bits. */
 		pmdp_test_and_clear_young(vma, addr, pmd);
-		test_and_clear_page_young(page);
-		ClearPageReferenced(page);
+		folio_test_clear_young(folio);
+		folio_clear_referenced(folio);
 out:
 		spin_unlock(ptl);
 		return 0;
@@ -1200,14 +1213,14 @@ out:
 		if (!pte_present(ptent))
 			continue;
 
-		page = vm_normal_page(vma, addr, ptent);
-		if (!page)
+		folio = vm_normal_folio(vma, addr, ptent);
+		if (!folio)
 			continue;
 
 		/* Clear accessed and referenced bits. */
 		ptep_test_and_clear_young(vma, addr, pte);
-		test_and_clear_page_young(page);
-		ClearPageReferenced(page);
+		folio_test_clear_young(folio);
+		folio_clear_referenced(folio);
 	}
 	pte_unmap_unlock(pte - 1, ptl);
 	cond_resched();
@@ -1352,8 +1365,7 @@ static inline pagemap_entry_t make_pme(u64 frame, u64 flags)
 	return (pagemap_entry_t) { .pme = (frame & PM_PFRAME_MASK) | flags };
 }
 
-static int add_to_pagemap(unsigned long addr, pagemap_entry_t *pme,
-			  struct pagemapread *pm)
+static int add_to_pagemap(pagemap_entry_t *pme, struct pagemapread *pm)
 {
 	pm->buffer[pm->pos++] = *pme;
 	if (pm->pos >= pm->len)
@@ -1380,7 +1392,7 @@ static int pagemap_pte_hole(unsigned long start, unsigned long end,
 			hole_end = end;
 
 		for (; addr < hole_end; addr += PAGE_SIZE) {
-			err = add_to_pagemap(addr, &pme, pm);
+			err = add_to_pagemap(&pme, pm);
 			if (err)
 				goto out;
 		}
@@ -1392,7 +1404,7 @@ static int pagemap_pte_hole(unsigned long start, unsigned long end,
 		if (vma->vm_flags & VM_SOFTDIRTY)
 			pme = make_pme(0, PM_SOFT_DIRTY);
 		for (; addr < min(end, vma->vm_end); addr += PAGE_SIZE) {
-			err = add_to_pagemap(addr, &pme, pm);
+			err = add_to_pagemap(&pme, pm);
 			if (err)
 				goto out;
 		}
@@ -1406,7 +1418,6 @@ static pagemap_entry_t pte_to_pagemap_entry(struct pagemapread *pm,
 {
 	u64 frame = 0, flags = 0;
 	struct page *page = NULL;
-	bool migration = false;
 
 	if (pte_present(pte)) {
 		if (pm->show_pfn)
@@ -1438,7 +1449,6 @@ static pagemap_entry_t pte_to_pagemap_entry(struct pagemapread *pm,
 			    (offset << MAX_SWAPFILES_SHIFT);
 		}
 		flags |= PM_SWAP;
-		migration = is_migration_entry(entry);
 		if (is_pfn_swap_entry(entry))
 			page = pfn_swap_entry_to_page(entry);
 		if (pte_marker_entry_uffd_wp(entry))
@@ -1447,7 +1457,7 @@ static pagemap_entry_t pte_to_pagemap_entry(struct pagemapread *pm,
 
 	if (page && !PageAnon(page))
 		flags |= PM_FILE;
-	if (page && !migration && page_mapcount(page) == 1)
+	if (page && (flags & PM_PRESENT) && page_mapcount(page) == 1)
 		flags |= PM_MMAP_EXCLUSIVE;
 	if (vma->vm_flags & VM_SOFTDIRTY)
 		flags |= PM_SOFT_DIRTY;
@@ -1464,10 +1474,10 @@ static int pagemap_pmd_range(pmd_t *pmdp, unsigned long addr, unsigned long end,
 	pte_t *pte, *orig_pte;
 	int err = 0;
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
-	bool migration = false;
 
 	ptl = pmd_trans_huge_lock(pmdp, vma);
 	if (ptl) {
+		unsigned int idx = (addr & ~PMD_MASK) >> PAGE_SHIFT;
 		u64 flags = 0, frame = 0;
 		pmd_t pmd = *pmdp;
 		struct page *page = NULL;
@@ -1484,8 +1494,7 @@ static int pagemap_pmd_range(pmd_t *pmdp, unsigned long addr, unsigned long end,
 			if (pmd_uffd_wp(pmd))
 				flags |= PM_UFFD_WP;
 			if (pm->show_pfn)
-				frame = pmd_pfn(pmd) +
-					((addr & ~PMD_MASK) >> PAGE_SHIFT);
+				frame = pmd_pfn(pmd) + idx;
 		}
 #ifdef CONFIG_ARCH_ENABLE_THP_MIGRATION
 		else if (is_swap_pmd(pmd)) {
@@ -1494,11 +1503,9 @@ static int pagemap_pmd_range(pmd_t *pmdp, unsigned long addr, unsigned long end,
 
 			if (pm->show_pfn) {
 				if (is_pfn_swap_entry(entry))
-					offset = swp_offset_pfn(entry);
+					offset = swp_offset_pfn(entry) + idx;
 				else
-					offset = swp_offset(entry);
-				offset = offset +
-					((addr & ~PMD_MASK) >> PAGE_SHIFT);
+					offset = swp_offset(entry) + idx;
 				frame = swp_type(entry) |
 					(offset << MAX_SWAPFILES_SHIFT);
 			}
@@ -1508,18 +1515,23 @@ static int pagemap_pmd_range(pmd_t *pmdp, unsigned long addr, unsigned long end,
 			if (pmd_swp_uffd_wp(pmd))
 				flags |= PM_UFFD_WP;
 			VM_BUG_ON(!is_pmd_migration_entry(pmd));
-			migration = is_migration_entry(entry);
 			page = pfn_swap_entry_to_page(entry);
 		}
 #endif
 
-		if (page && !migration && page_mapcount(page) == 1)
-			flags |= PM_MMAP_EXCLUSIVE;
+		if (page && !PageAnon(page))
+			flags |= PM_FILE;
+
+		for (; addr != end; addr += PAGE_SIZE, idx++) {
+			unsigned long cur_flags = flags;
+			pagemap_entry_t pme;
 
-		for (; addr != end; addr += PAGE_SIZE) {
-			pagemap_entry_t pme = make_pme(frame, flags);
+			if (page && (flags & PM_PRESENT) &&
+			    page_mapcount(page + idx) == 1)
+				cur_flags |= PM_MMAP_EXCLUSIVE;
 
-			err = add_to_pagemap(addr, &pme, pm);
+			pme = make_pme(frame, cur_flags);
+			err = add_to_pagemap(&pme, pm);
 			if (err)
 				break;
 			if (pm->show_pfn) {
@@ -1547,7 +1559,7 @@ static int pagemap_pmd_range(pmd_t *pmdp, unsigned long addr, unsigned long end,
 		pagemap_entry_t pme;
 
 		pme = pte_to_pagemap_entry(pm, vma, addr, ptep_get(pte));
-		err = add_to_pagemap(addr, &pme, pm);
+		err = add_to_pagemap(&pme, pm);
 		if (err)
 			break;
 	}
@@ -1575,12 +1587,13 @@ static int pagemap_hugetlb_range(pte_t *ptep, unsigned long hmask,
 
 	pte = huge_ptep_get(ptep);
 	if (pte_present(pte)) {
-		struct page *page = pte_page(pte);
+		struct folio *folio = page_folio(pte_page(pte));
 
-		if (!PageAnon(page))
+		if (!folio_test_anon(folio))
 			flags |= PM_FILE;
 
-		if (page_mapcount(page) == 1)
+		if (!folio_likely_mapped_shared(folio) &&
+		    !hugetlb_pmd_shared(ptep))
 			flags |= PM_MMAP_EXCLUSIVE;
 
 		if (huge_pte_uffd_wp(pte))
@@ -1597,7 +1610,7 @@ static int pagemap_hugetlb_range(pte_t *ptep, unsigned long hmask,
 	for (; addr != end; addr += PAGE_SIZE) {
 		pagemap_entry_t pme = make_pme(frame, flags);
 
-		err = add_to_pagemap(addr, &pme, pm);
+		err = add_to_pagemap(&pme, pm);
 		if (err)
 			return err;
 		if (pm->show_pfn && (flags & PM_PRESENT))
@@ -1807,7 +1820,7 @@ static unsigned long pagemap_page_category(struct pagemap_scan_private *p,
 		if (p->masks_of_interest & PAGE_IS_FILE) {
 			swp = pte_to_swp_entry(pte);
 			if (is_pfn_swap_entry(swp) &&
-			    !PageAnon(pfn_swap_entry_to_page(swp)))
+			    !folio_test_anon(pfn_swap_entry_folio(swp)))
 				categories |= PAGE_IS_FILE;
 		}
 		if (pte_swp_soft_dirty(pte))
@@ -1871,7 +1884,7 @@ static unsigned long pagemap_thp_category(struct pagemap_scan_private *p,
 		if (p->masks_of_interest & PAGE_IS_FILE) {
 			swp = pmd_to_swp_entry(pmd);
 			if (is_pfn_swap_entry(swp) &&
-			    !PageAnon(pfn_swap_entry_to_page(swp)))
+			    !folio_test_anon(pfn_swap_entry_folio(swp)))
 				categories |= PAGE_IS_FILE;
 		}
 	}
@@ -2552,28 +2565,29 @@ struct numa_maps_private {
 static void gather_stats(struct page *page, struct numa_maps *md, int pte_dirty,
 			unsigned long nr_pages)
 {
+	struct folio *folio = page_folio(page);
 	int count = page_mapcount(page);
 
 	md->pages += nr_pages;
-	if (pte_dirty || PageDirty(page))
+	if (pte_dirty || folio_test_dirty(folio))
 		md->dirty += nr_pages;
 
-	if (PageSwapCache(page))
+	if (folio_test_swapcache(folio))
 		md->swapcache += nr_pages;
 
-	if (PageActive(page) || PageUnevictable(page))
+	if (folio_test_active(folio) || folio_test_unevictable(folio))
 		md->active += nr_pages;
 
-	if (PageWriteback(page))
+	if (folio_test_writeback(folio))
 		md->writeback += nr_pages;
 
-	if (PageAnon(page))
+	if (folio_test_anon(folio))
 		md->anon += nr_pages;
 
 	if (count > md->mapcount_max)
 		md->mapcount_max = count;
 
-	md->node[page_to_nid(page)] += nr_pages;
+	md->node[folio_nid(folio)] += nr_pages;
 }
 
 static struct page *can_gather_numa_stats(pte_t pte, struct vm_area_struct *vma,
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index 1fb213f379..b52d85f8ad 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -383,6 +383,8 @@ static ssize_t __read_vmcore(struct iov_iter *iter, loff_t *fpos)
 		/* leave now if filled buffer already */
 		if (!iov_iter_count(iter))
 			return acc;
+
+		cond_resched();
 	}
 
 	list_for_each_entry(m, &vmcore_list, list) {
@@ -1370,9 +1372,8 @@ static void vmcoredd_write_header(void *buf, struct vmcoredd_data *data,
 	vdd_hdr->n_descsz = size + sizeof(vdd_hdr->dump_name);
 	vdd_hdr->n_type = NT_VMCOREDD;
 
-	strncpy((char *)vdd_hdr->name, VMCOREDD_NOTE_NAME,
-		sizeof(vdd_hdr->name));
-	memcpy(vdd_hdr->dump_name, data->dump_name, sizeof(vdd_hdr->dump_name));
+	strscpy_pad(vdd_hdr->name, VMCOREDD_NOTE_NAME);
+	strscpy_pad(vdd_hdr->dump_name, data->dump_name);
 }
 
 /**
diff --git a/fs/pstore/ram.c b/fs/pstore/ram.c
index 88b34fdbf7..b1a455f42e 100644
--- a/fs/pstore/ram.c
+++ b/fs/pstore/ram.c
@@ -893,6 +893,7 @@ static const struct of_device_id dt_match[] = {
 	{ .compatible = "ramoops" },
 	{}
 };
+MODULE_DEVICE_TABLE(of, dt_match);
 
 static struct platform_driver ramoops_driver = {
 	.probe		= ramoops_probe,
diff --git a/fs/pstore/zone.c b/fs/pstore/zone.c
index abca117725..694db61666 100644
--- a/fs/pstore/zone.c
+++ b/fs/pstore/zone.c
@@ -1217,7 +1217,6 @@ static struct pstore_zone **psz_init_zones(enum pstore_type_id type,
 		pr_err("allocate for zones %s failed\n", name);
 		return ERR_PTR(-ENOMEM);
 	}
-	memset(zones, 0, c * sizeof(*zones));
 
 	for (i = 0; i < c; i++) {
 		zone = psz_init_zone(type, off, record_size);
diff --git a/fs/qnx4/inode.c b/fs/qnx4/inode.c
index 6eb9bb369b..d79841e944 100644
--- a/fs/qnx4/inode.c
+++ b/fs/qnx4/inode.c
@@ -21,6 +21,7 @@
 #include <linux/buffer_head.h>
 #include <linux/writeback.h>
 #include <linux/statfs.h>
+#include <linux/fs_context.h>
 #include "qnx4.h"
 
 #define QNX4_VERSION  4
@@ -30,28 +31,33 @@ static const struct super_operations qnx4_sops;
 
 static struct inode *qnx4_alloc_inode(struct super_block *sb);
 static void qnx4_free_inode(struct inode *inode);
-static int qnx4_remount(struct super_block *sb, int *flags, char *data);
 static int qnx4_statfs(struct dentry *, struct kstatfs *);
+static int qnx4_get_tree(struct fs_context *fc);
 
 static const struct super_operations qnx4_sops =
 {
 	.alloc_inode	= qnx4_alloc_inode,
 	.free_inode	= qnx4_free_inode,
 	.statfs		= qnx4_statfs,
-	.remount_fs	= qnx4_remount,
 };
 
-static int qnx4_remount(struct super_block *sb, int *flags, char *data)
+static int qnx4_reconfigure(struct fs_context *fc)
 {
+	struct super_block *sb = fc->root->d_sb;
 	struct qnx4_sb_info *qs;
 
 	sync_filesystem(sb);
 	qs = qnx4_sb(sb);
 	qs->Version = QNX4_VERSION;
-	*flags |= SB_RDONLY;
+	fc->sb_flags |= SB_RDONLY;
 	return 0;
 }
 
+static const struct fs_context_operations qnx4_context_opts = {
+	.get_tree	= qnx4_get_tree,
+	.reconfigure	= qnx4_reconfigure,
+};
+
 static int qnx4_get_block( struct inode *inode, sector_t iblock, struct buffer_head *bh, int create )
 {
 	unsigned long phys;
@@ -183,12 +189,13 @@ static const char *qnx4_checkroot(struct super_block *sb,
 	return "bitmap file not found.";
 }
 
-static int qnx4_fill_super(struct super_block *s, void *data, int silent)
+static int qnx4_fill_super(struct super_block *s, struct fs_context *fc)
 {
 	struct buffer_head *bh;
 	struct inode *root;
 	const char *errmsg;
 	struct qnx4_sb_info *qs;
+	int silent = fc->sb_flags & SB_SILENT;
 
 	qs = kzalloc(sizeof(struct qnx4_sb_info), GFP_KERNEL);
 	if (!qs)
@@ -216,7 +223,7 @@ static int qnx4_fill_super(struct super_block *s, void *data, int silent)
 	errmsg = qnx4_checkroot(s, (struct qnx4_super_block *) bh->b_data);
 	brelse(bh);
 	if (errmsg != NULL) {
- 		if (!silent)
+		if (!silent)
 			printk(KERN_ERR "qnx4: %s\n", errmsg);
 		return -EINVAL;
 	}
@@ -235,6 +242,18 @@ static int qnx4_fill_super(struct super_block *s, void *data, int silent)
 	return 0;
 }
 
+static int qnx4_get_tree(struct fs_context *fc)
+{
+	return get_tree_bdev(fc, qnx4_fill_super);
+}
+
+static int qnx4_init_fs_context(struct fs_context *fc)
+{
+	fc->ops = &qnx4_context_opts;
+
+	return 0;
+}
+
 static void qnx4_kill_sb(struct super_block *sb)
 {
 	struct qnx4_sb_info *qs = qnx4_sb(sb);
@@ -359,7 +378,7 @@ static int init_inodecache(void)
 	qnx4_inode_cachep = kmem_cache_create("qnx4_inode_cache",
 					     sizeof(struct qnx4_inode_info),
 					     0, (SLAB_RECLAIM_ACCOUNT|
-						SLAB_MEM_SPREAD|SLAB_ACCOUNT),
+						SLAB_ACCOUNT),
 					     init_once);
 	if (qnx4_inode_cachep == NULL)
 		return -ENOMEM;
@@ -376,18 +395,12 @@ static void destroy_inodecache(void)
 	kmem_cache_destroy(qnx4_inode_cachep);
 }
 
-static struct dentry *qnx4_mount(struct file_system_type *fs_type,
-	int flags, const char *dev_name, void *data)
-{
-	return mount_bdev(fs_type, flags, dev_name, data, qnx4_fill_super);
-}
-
 static struct file_system_type qnx4_fs_type = {
-	.owner		= THIS_MODULE,
-	.name		= "qnx4",
-	.mount		= qnx4_mount,
-	.kill_sb	= qnx4_kill_sb,
-	.fs_flags	= FS_REQUIRES_DEV,
+	.owner			= THIS_MODULE,
+	.name			= "qnx4",
+	.kill_sb		= qnx4_kill_sb,
+	.fs_flags		= FS_REQUIRES_DEV,
+	.init_fs_context	= qnx4_init_fs_context,
 };
 MODULE_ALIAS_FS("qnx4");
 
diff --git a/fs/qnx6/inode.c b/fs/qnx6/inode.c
index a286c54571..d62fbef838 100644
--- a/fs/qnx6/inode.c
+++ b/fs/qnx6/inode.c
@@ -19,11 +19,11 @@
 #include <linux/buffer_head.h>
 #include <linux/writeback.h>
 #include <linux/statfs.h>
-#include <linux/parser.h>
 #include <linux/seq_file.h>
-#include <linux/mount.h>
 #include <linux/crc32.h>
 #include <linux/mpage.h>
+#include <linux/fs_parser.h>
+#include <linux/fs_context.h>
 #include "qnx6.h"
 
 static const struct super_operations qnx6_sops;
@@ -31,7 +31,7 @@ static const struct super_operations qnx6_sops;
 static void qnx6_put_super(struct super_block *sb);
 static struct inode *qnx6_alloc_inode(struct super_block *sb);
 static void qnx6_free_inode(struct inode *inode);
-static int qnx6_remount(struct super_block *sb, int *flags, char *data);
+static int qnx6_reconfigure(struct fs_context *fc);
 static int qnx6_statfs(struct dentry *dentry, struct kstatfs *buf);
 static int qnx6_show_options(struct seq_file *seq, struct dentry *root);
 
@@ -40,7 +40,6 @@ static const struct super_operations qnx6_sops = {
 	.free_inode	= qnx6_free_inode,
 	.put_super	= qnx6_put_super,
 	.statfs		= qnx6_statfs,
-	.remount_fs	= qnx6_remount,
 	.show_options	= qnx6_show_options,
 };
 
@@ -54,10 +53,12 @@ static int qnx6_show_options(struct seq_file *seq, struct dentry *root)
 	return 0;
 }
 
-static int qnx6_remount(struct super_block *sb, int *flags, char *data)
+static int qnx6_reconfigure(struct fs_context *fc)
 {
+	struct super_block *sb = fc->root->d_sb;
+
 	sync_filesystem(sb);
-	*flags |= SB_RDONLY;
+	fc->sb_flags |= SB_RDONLY;
 	return 0;
 }
 
@@ -218,39 +219,36 @@ void qnx6_superblock_debug(struct qnx6_super_block *sb, struct super_block *s)
 #endif
 
 enum {
-	Opt_mmifs,
-	Opt_err
+	Opt_mmifs
+};
+
+struct qnx6_context {
+	unsigned long s_mount_opts;
 };
 
-static const match_table_t tokens = {
-	{Opt_mmifs, "mmi_fs"},
-	{Opt_err, NULL}
+static const struct fs_parameter_spec qnx6_param_spec[] = {
+	fsparam_flag	("mmi_fs",	Opt_mmifs),
+	{}
 };
 
-static int qnx6_parse_options(char *options, struct super_block *sb)
+static int qnx6_parse_param(struct fs_context *fc, struct fs_parameter *param)
 {
-	char *p;
-	struct qnx6_sb_info *sbi = QNX6_SB(sb);
-	substring_t args[MAX_OPT_ARGS];
-
-	if (!options)
-		return 1;
-
-	while ((p = strsep(&options, ",")) != NULL) {
-		int token;
-		if (!*p)
-			continue;
-
-		token = match_token(p, tokens, args);
-		switch (token) {
-		case Opt_mmifs:
-			set_opt(sbi->s_mount_opt, MMI_FS);
-			break;
-		default:
-			return 0;
-		}
+	struct qnx6_context *ctx = fc->fs_private;
+	struct fs_parse_result result;
+	int opt;
+
+	opt = fs_parse(fc, qnx6_param_spec, param, &result);
+	if (opt < 0)
+		return opt;
+
+	switch (opt) {
+	case Opt_mmifs:
+		ctx->s_mount_opts |= QNX6_MOUNT_MMI_FS;
+		break;
+	default:
+		return -EINVAL;
 	}
-	return 1;
+	return 0;
 }
 
 static struct buffer_head *qnx6_check_first_superblock(struct super_block *s,
@@ -293,22 +291,25 @@ static struct buffer_head *qnx6_check_first_superblock(struct super_block *s,
 static struct inode *qnx6_private_inode(struct super_block *s,
 					struct qnx6_root_node *p);
 
-static int qnx6_fill_super(struct super_block *s, void *data, int silent)
+static int qnx6_fill_super(struct super_block *s, struct fs_context *fc)
 {
 	struct buffer_head *bh1 = NULL, *bh2 = NULL;
 	struct qnx6_super_block *sb1 = NULL, *sb2 = NULL;
 	struct qnx6_sb_info *sbi;
+	struct qnx6_context *ctx = fc->fs_private;
 	struct inode *root;
 	const char *errmsg;
 	struct qnx6_sb_info *qs;
 	int ret = -EINVAL;
 	u64 offset;
 	int bootblock_offset = QNX6_BOOTBLOCK_SIZE;
+	int silent = fc->sb_flags & SB_SILENT;
 
 	qs = kzalloc(sizeof(struct qnx6_sb_info), GFP_KERNEL);
 	if (!qs)
 		return -ENOMEM;
 	s->s_fs_info = qs;
+	qs->s_mount_opt = ctx->s_mount_opts;
 
 	/* Superblock always is 512 Byte long */
 	if (!sb_set_blocksize(s, QNX6_SUPERBLOCK_SIZE)) {
@@ -316,12 +317,7 @@ static int qnx6_fill_super(struct super_block *s, void *data, int silent)
 		goto outnobh;
 	}
 
-	/* parse the mount-options */
-	if (!qnx6_parse_options((char *) data, s)) {
-		pr_err("invalid mount options.\n");
-		goto outnobh;
-	}
-	if (test_opt(s, MMI_FS)) {
+	if (qs->s_mount_opt == QNX6_MOUNT_MMI_FS) {
 		sb1 = qnx6_mmi_fill_super(s, silent);
 		if (sb1)
 			goto mmi_success;
@@ -615,7 +611,7 @@ static int init_inodecache(void)
 	qnx6_inode_cachep = kmem_cache_create("qnx6_inode_cache",
 					     sizeof(struct qnx6_inode_info),
 					     0, (SLAB_RECLAIM_ACCOUNT|
-						SLAB_MEM_SPREAD|SLAB_ACCOUNT),
+						SLAB_ACCOUNT),
 					     init_once);
 	if (!qnx6_inode_cachep)
 		return -ENOMEM;
@@ -632,18 +628,43 @@ static void destroy_inodecache(void)
 	kmem_cache_destroy(qnx6_inode_cachep);
 }
 
-static struct dentry *qnx6_mount(struct file_system_type *fs_type,
-	int flags, const char *dev_name, void *data)
+static int qnx6_get_tree(struct fs_context *fc)
+{
+	return get_tree_bdev(fc, qnx6_fill_super);
+}
+
+static void qnx6_free_fc(struct fs_context *fc)
 {
-	return mount_bdev(fs_type, flags, dev_name, data, qnx6_fill_super);
+	kfree(fc->fs_private);
+}
+
+static const struct fs_context_operations qnx6_context_ops = {
+	.parse_param	= qnx6_parse_param,
+	.get_tree	= qnx6_get_tree,
+	.reconfigure	= qnx6_reconfigure,
+	.free		= qnx6_free_fc,
+};
+
+static int qnx6_init_fs_context(struct fs_context *fc)
+{
+	struct qnx6_context *ctx;
+
+	ctx = kzalloc(sizeof(struct qnx6_context), GFP_KERNEL);
+	if (!ctx)
+		return -ENOMEM;
+	fc->ops = &qnx6_context_ops;
+	fc->fs_private = ctx;
+
+	return 0;
 }
 
 static struct file_system_type qnx6_fs_type = {
-	.owner		= THIS_MODULE,
-	.name		= "qnx6",
-	.mount		= qnx6_mount,
-	.kill_sb	= kill_block_super,
-	.fs_flags	= FS_REQUIRES_DEV,
+	.owner			= THIS_MODULE,
+	.name			= "qnx6",
+	.kill_sb		= kill_block_super,
+	.fs_flags		= FS_REQUIRES_DEV,
+	.init_fs_context	= qnx6_init_fs_context,
+	.parameters		= qnx6_param_spec,
 };
 MODULE_ALIAS_FS("qnx6");
 
diff --git a/fs/quota/dquot.c b/fs/quota/dquot.c
index 51ff554dc8..627eb2f72e 100644
--- a/fs/quota/dquot.c
+++ b/fs/quota/dquot.c
@@ -410,7 +410,7 @@ static inline int mark_all_dquot_dirty(struct dquot __rcu * const *dquots)
 		if (dquot)
 			/* Even in case of error we have to continue */
 			ret = mark_dquot_dirty(dquot);
-		if (!err)
+		if (!err && ret < 0)
 			err = ret;
 	}
 	return err;
@@ -877,10 +877,7 @@ void dqput(struct dquot *dquot)
 	}
 
 	/* Need to release dquot? */
-#ifdef CONFIG_QUOTA_DEBUG
-	/* sanity check */
-	BUG_ON(!list_empty(&dquot->dq_free));
-#endif
+	WARN_ON_ONCE(!list_empty(&dquot->dq_free));
 	put_releasing_dquots(dquot);
 	atomic_dec(&dquot->dq_count);
 	spin_unlock(&dq_list_lock);
@@ -989,9 +986,8 @@ we_slept:
 	 * smp_mb__before_atomic() in dquot_acquire().
 	 */
 	smp_rmb();
-#ifdef CONFIG_QUOTA_DEBUG
-	BUG_ON(!dquot->dq_sb);	/* Has somebody invalidated entry under us? */
-#endif
+	/* Has somebody invalidated entry under us? */
+	WARN_ON_ONCE(hlist_unhashed(&dquot->dq_hash));
 out:
 	if (empty)
 		do_destroy_dquot(empty);
@@ -1741,7 +1737,7 @@ int __dquot_alloc_space(struct inode *inode, qsize_t number, int flags)
 
 	if (reserve)
 		goto out_flush_warn;
-	mark_all_dquot_dirty(dquots);
+	ret = mark_all_dquot_dirty(dquots);
 out_flush_warn:
 	srcu_read_unlock(&dquot_srcu, index);
 	flush_warnings(warn);
@@ -1790,7 +1786,7 @@ int dquot_alloc_inode(struct inode *inode)
 warn_put_all:
 	spin_unlock(&inode->i_lock);
 	if (ret == 0)
-		mark_all_dquot_dirty(dquots);
+		ret = mark_all_dquot_dirty(dquots);
 	srcu_read_unlock(&dquot_srcu, index);
 	flush_warnings(warn);
 	return ret;
@@ -1994,7 +1990,7 @@ int __dquot_transfer(struct inode *inode, struct dquot **transfer_to)
 	qsize_t inode_usage = 1;
 	struct dquot __rcu **dquots;
 	struct dquot *transfer_from[MAXQUOTAS] = {};
-	int cnt, index, ret = 0;
+	int cnt, index, ret = 0, err;
 	char is_valid[MAXQUOTAS] = {};
 	struct dquot_warn warn_to[MAXQUOTAS];
 	struct dquot_warn warn_from_inodes[MAXQUOTAS];
@@ -2091,8 +2087,12 @@ int __dquot_transfer(struct inode *inode, struct dquot **transfer_to)
 	 * mark_all_dquot_dirty().
 	 */
 	index = srcu_read_lock(&dquot_srcu);
-	mark_all_dquot_dirty((struct dquot __rcu **)transfer_from);
-	mark_all_dquot_dirty((struct dquot __rcu **)transfer_to);
+	err = mark_all_dquot_dirty((struct dquot __rcu **)transfer_from);
+	if (err < 0)
+		ret = err;
+	err = mark_all_dquot_dirty((struct dquot __rcu **)transfer_to);
+	if (err < 0)
+		ret = err;
 	srcu_read_unlock(&dquot_srcu, index);
 
 	flush_warnings(warn_to);
@@ -2102,7 +2102,7 @@ int __dquot_transfer(struct inode *inode, struct dquot **transfer_to)
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
 		if (is_valid[cnt])
 			transfer_to[cnt] = transfer_from[cnt];
-	return 0;
+	return ret;
 over_quota:
 	/* Back out changes we already did */
 	for (cnt--; cnt >= 0; cnt--) {
@@ -2415,7 +2415,8 @@ int dquot_load_quota_sb(struct super_block *sb, int type, int format_id,
 	lockdep_assert_held_write(&sb->s_umount);
 
 	/* Just unsuspend quotas? */
-	BUG_ON(flags & DQUOT_SUSPENDED);
+	if (WARN_ON_ONCE(flags & DQUOT_SUSPENDED))
+		return -EINVAL;
 
 	if (!fmt)
 		return -ESRCH;
@@ -2729,6 +2730,7 @@ static int do_set_dqblk(struct dquot *dquot, struct qc_dqblk *di)
 	struct mem_dqblk *dm = &dquot->dq_dqb;
 	int check_blim = 0, check_ilim = 0;
 	struct mem_dqinfo *dqi = &sb_dqopt(dquot->dq_sb)->info[dquot->dq_id.type];
+	int ret;
 
 	if (di->d_fieldmask & ~VFS_QC_MASK)
 		return -EINVAL;
@@ -2810,8 +2812,9 @@ static int do_set_dqblk(struct dquot *dquot, struct qc_dqblk *di)
 	else
 		set_bit(DQ_FAKE_B, &dquot->dq_flags);
 	spin_unlock(&dquot->dq_dqb_lock);
-	mark_dquot_dirty(dquot);
-
+	ret = mark_dquot_dirty(dquot);
+	if (ret < 0)
+		return ret;
 	return 0;
 }
 
@@ -3011,7 +3014,7 @@ static int __init dquot_init(void)
 	dquot_cachep = kmem_cache_create("dquot",
 			sizeof(struct dquot), sizeof(unsigned long) * 4,
 			(SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
-				SLAB_MEM_SPREAD|SLAB_PANIC),
+				SLAB_PANIC),
 			NULL);
 
 	order = 0;
@@ -3019,11 +3022,10 @@ static int __init dquot_init(void)
 	if (!dquot_hash)
 		panic("Cannot create dquot hash table");
 
-	for (i = 0; i < _DQST_DQSTAT_LAST; i++) {
-		ret = percpu_counter_init(&dqstats.counter[i], 0, GFP_KERNEL);
-		if (ret)
-			panic("Cannot create dquot stat counters");
-	}
+	ret = percpu_counter_init_many(dqstats.counter, 0, GFP_KERNEL,
+				       _DQST_DQSTAT_LAST);
+	if (ret)
+		panic("Cannot create dquot stat counters");
 
 	/* Find power-of-two hlist_heads which can fit into allocation */
 	nr_hash = (1UL << order) * PAGE_SIZE / sizeof(struct hlist_head);
diff --git a/fs/quota/quota_tree.c b/fs/quota/quota_tree.c
index 0f1493e0f6..afceef3ddf 100644
--- a/fs/quota/quota_tree.c
+++ b/fs/quota/quota_tree.c
@@ -21,6 +21,12 @@ MODULE_AUTHOR("Jan Kara");
 MODULE_DESCRIPTION("Quota trie support");
 MODULE_LICENSE("GPL");
 
+/*
+ * Maximum quota tree depth we support. Only to limit recursion when working
+ * with the tree.
+ */
+#define MAX_QTREE_DEPTH 6
+
 #define __QUOTA_QT_PARANOIA
 
 static int __get_index(struct qtree_mem_dqinfo *info, qid_t id, int depth)
@@ -108,7 +114,7 @@ static int check_dquot_block_header(struct qtree_mem_dqinfo *info,
 /* Remove empty block from list and return it */
 static int get_free_dqblk(struct qtree_mem_dqinfo *info)
 {
-	char *buf = kmalloc(info->dqi_usable_bs, GFP_NOFS);
+	char *buf = kmalloc(info->dqi_usable_bs, GFP_KERNEL);
 	struct qt_disk_dqdbheader *dh = (struct qt_disk_dqdbheader *)buf;
 	int ret, blk;
 
@@ -160,7 +166,7 @@ static int put_free_dqblk(struct qtree_mem_dqinfo *info, char *buf, uint blk)
 static int remove_free_dqentry(struct qtree_mem_dqinfo *info, char *buf,
 			       uint blk)
 {
-	char *tmpbuf = kmalloc(info->dqi_usable_bs, GFP_NOFS);
+	char *tmpbuf = kmalloc(info->dqi_usable_bs, GFP_KERNEL);
 	struct qt_disk_dqdbheader *dh = (struct qt_disk_dqdbheader *)buf;
 	uint nextblk = le32_to_cpu(dh->dqdh_next_free);
 	uint prevblk = le32_to_cpu(dh->dqdh_prev_free);
@@ -207,7 +213,7 @@ out_buf:
 static int insert_free_dqentry(struct qtree_mem_dqinfo *info, char *buf,
 			       uint blk)
 {
-	char *tmpbuf = kmalloc(info->dqi_usable_bs, GFP_NOFS);
+	char *tmpbuf = kmalloc(info->dqi_usable_bs, GFP_KERNEL);
 	struct qt_disk_dqdbheader *dh = (struct qt_disk_dqdbheader *)buf;
 	int err;
 
@@ -255,7 +261,7 @@ static uint find_free_dqentry(struct qtree_mem_dqinfo *info,
 {
 	uint blk, i;
 	struct qt_disk_dqdbheader *dh;
-	char *buf = kmalloc(info->dqi_usable_bs, GFP_NOFS);
+	char *buf = kmalloc(info->dqi_usable_bs, GFP_KERNEL);
 	char *ddquot;
 
 	*err = 0;
@@ -327,27 +333,36 @@ out_buf:
 
 /* Insert reference to structure into the trie */
 static int do_insert_tree(struct qtree_mem_dqinfo *info, struct dquot *dquot,
-			  uint *treeblk, int depth)
+			  uint *blks, int depth)
 {
-	char *buf = kmalloc(info->dqi_usable_bs, GFP_NOFS);
+	char *buf = kmalloc(info->dqi_usable_bs, GFP_KERNEL);
 	int ret = 0, newson = 0, newact = 0;
 	__le32 *ref;
 	uint newblk;
+	int i;
 
 	if (!buf)
 		return -ENOMEM;
-	if (!*treeblk) {
+	if (!blks[depth]) {
 		ret = get_free_dqblk(info);
 		if (ret < 0)
 			goto out_buf;
-		*treeblk = ret;
+		for (i = 0; i < depth; i++)
+			if (ret == blks[i]) {
+				quota_error(dquot->dq_sb,
+					"Free block already used in tree: block %u",
+					ret);
+				ret = -EIO;
+				goto out_buf;
+			}
+		blks[depth] = ret;
 		memset(buf, 0, info->dqi_usable_bs);
 		newact = 1;
 	} else {
-		ret = read_blk(info, *treeblk, buf);
+		ret = read_blk(info, blks[depth], buf);
 		if (ret < 0) {
 			quota_error(dquot->dq_sb, "Can't read tree quota "
-				    "block %u", *treeblk);
+				    "block %u", blks[depth]);
 			goto out_buf;
 		}
 	}
@@ -357,8 +372,20 @@ static int do_insert_tree(struct qtree_mem_dqinfo *info, struct dquot *dquot,
 			     info->dqi_blocks - 1);
 	if (ret)
 		goto out_buf;
-	if (!newblk)
+	if (!newblk) {
 		newson = 1;
+	} else {
+		for (i = 0; i <= depth; i++)
+			if (newblk == blks[i]) {
+				quota_error(dquot->dq_sb,
+					"Cycle in quota tree detected: block %u index %u",
+					blks[depth],
+					get_index(info, dquot->dq_id, depth));
+				ret = -EIO;
+				goto out_buf;
+			}
+	}
+	blks[depth + 1] = newblk;
 	if (depth == info->dqi_qtree_depth - 1) {
 #ifdef __QUOTA_QT_PARANOIA
 		if (newblk) {
@@ -370,16 +397,16 @@ static int do_insert_tree(struct qtree_mem_dqinfo *info, struct dquot *dquot,
 			goto out_buf;
 		}
 #endif
-		newblk = find_free_dqentry(info, dquot, &ret);
+		blks[depth + 1] = find_free_dqentry(info, dquot, &ret);
 	} else {
-		ret = do_insert_tree(info, dquot, &newblk, depth+1);
+		ret = do_insert_tree(info, dquot, blks, depth + 1);
 	}
 	if (newson && ret >= 0) {
 		ref[get_index(info, dquot->dq_id, depth)] =
-							cpu_to_le32(newblk);
-		ret = write_blk(info, *treeblk, buf);
+						cpu_to_le32(blks[depth + 1]);
+		ret = write_blk(info, blks[depth], buf);
 	} else if (newact && ret < 0) {
-		put_free_dqblk(info, buf, *treeblk);
+		put_free_dqblk(info, buf, blks[depth]);
 	}
 out_buf:
 	kfree(buf);
@@ -390,7 +417,7 @@ out_buf:
 static inline int dq_insert_tree(struct qtree_mem_dqinfo *info,
 				 struct dquot *dquot)
 {
-	int tmp = QT_TREEOFF;
+	uint blks[MAX_QTREE_DEPTH] = { QT_TREEOFF };
 
 #ifdef __QUOTA_QT_PARANOIA
 	if (info->dqi_blocks <= QT_TREEOFF) {
@@ -398,7 +425,11 @@ static inline int dq_insert_tree(struct qtree_mem_dqinfo *info,
 		return -EIO;
 	}
 #endif
-	return do_insert_tree(info, dquot, &tmp, 0);
+	if (info->dqi_qtree_depth >= MAX_QTREE_DEPTH) {
+		quota_error(dquot->dq_sb, "Quota tree depth too big!");
+		return -EIO;
+	}
+	return do_insert_tree(info, dquot, blks, 0);
 }
 
 /*
@@ -410,7 +441,7 @@ int qtree_write_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot)
 	int type = dquot->dq_id.type;
 	struct super_block *sb = dquot->dq_sb;
 	ssize_t ret;
-	char *ddquot = kmalloc(info->dqi_entry_size, GFP_NOFS);
+	char *ddquot = kmalloc(info->dqi_entry_size, GFP_KERNEL);
 
 	if (!ddquot)
 		return -ENOMEM;
@@ -449,7 +480,7 @@ static int free_dqentry(struct qtree_mem_dqinfo *info, struct dquot *dquot,
 			uint blk)
 {
 	struct qt_disk_dqdbheader *dh;
-	char *buf = kmalloc(info->dqi_usable_bs, GFP_NOFS);
+	char *buf = kmalloc(info->dqi_usable_bs, GFP_KERNEL);
 	int ret = 0;
 
 	if (!buf)
@@ -511,19 +542,20 @@ out_buf:
 
 /* Remove reference to dquot from tree */
 static int remove_tree(struct qtree_mem_dqinfo *info, struct dquot *dquot,
-		       uint *blk, int depth)
+		       uint *blks, int depth)
 {
-	char *buf = kmalloc(info->dqi_usable_bs, GFP_NOFS);
+	char *buf = kmalloc(info->dqi_usable_bs, GFP_KERNEL);
 	int ret = 0;
 	uint newblk;
 	__le32 *ref = (__le32 *)buf;
+	int i;
 
 	if (!buf)
 		return -ENOMEM;
-	ret = read_blk(info, *blk, buf);
+	ret = read_blk(info, blks[depth], buf);
 	if (ret < 0) {
 		quota_error(dquot->dq_sb, "Can't read quota data block %u",
-			    *blk);
+			    blks[depth]);
 		goto out_buf;
 	}
 	newblk = le32_to_cpu(ref[get_index(info, dquot->dq_id, depth)]);
@@ -532,29 +564,38 @@ static int remove_tree(struct qtree_mem_dqinfo *info, struct dquot *dquot,
 	if (ret)
 		goto out_buf;
 
+	for (i = 0; i <= depth; i++)
+		if (newblk == blks[i]) {
+			quota_error(dquot->dq_sb,
+				"Cycle in quota tree detected: block %u index %u",
+				blks[depth],
+				get_index(info, dquot->dq_id, depth));
+			ret = -EIO;
+			goto out_buf;
+		}
 	if (depth == info->dqi_qtree_depth - 1) {
 		ret = free_dqentry(info, dquot, newblk);
-		newblk = 0;
+		blks[depth + 1] = 0;
 	} else {
-		ret = remove_tree(info, dquot, &newblk, depth+1);
+		blks[depth + 1] = newblk;
+		ret = remove_tree(info, dquot, blks, depth + 1);
 	}
-	if (ret >= 0 && !newblk) {
-		int i;
+	if (ret >= 0 && !blks[depth + 1]) {
 		ref[get_index(info, dquot->dq_id, depth)] = cpu_to_le32(0);
 		/* Block got empty? */
 		for (i = 0; i < (info->dqi_usable_bs >> 2) && !ref[i]; i++)
 			;
 		/* Don't put the root block into the free block list */
 		if (i == (info->dqi_usable_bs >> 2)
-		    && *blk != QT_TREEOFF) {
-			put_free_dqblk(info, buf, *blk);
-			*blk = 0;
+		    && blks[depth] != QT_TREEOFF) {
+			put_free_dqblk(info, buf, blks[depth]);
+			blks[depth] = 0;
 		} else {
-			ret = write_blk(info, *blk, buf);
+			ret = write_blk(info, blks[depth], buf);
 			if (ret < 0)
 				quota_error(dquot->dq_sb,
 					    "Can't write quota tree block %u",
-					    *blk);
+					    blks[depth]);
 		}
 	}
 out_buf:
@@ -565,11 +606,15 @@ out_buf:
 /* Delete dquot from tree */
 int qtree_delete_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot)
 {
-	uint tmp = QT_TREEOFF;
+	uint blks[MAX_QTREE_DEPTH] = { QT_TREEOFF };
 
 	if (!dquot->dq_off)	/* Even not allocated? */
 		return 0;
-	return remove_tree(info, dquot, &tmp, 0);
+	if (info->dqi_qtree_depth >= MAX_QTREE_DEPTH) {
+		quota_error(dquot->dq_sb, "Quota tree depth too big!");
+		return -EIO;
+	}
+	return remove_tree(info, dquot, blks, 0);
 }
 EXPORT_SYMBOL(qtree_delete_dquot);
 
@@ -577,7 +622,7 @@ EXPORT_SYMBOL(qtree_delete_dquot);
 static loff_t find_block_dqentry(struct qtree_mem_dqinfo *info,
 				 struct dquot *dquot, uint blk)
 {
-	char *buf = kmalloc(info->dqi_usable_bs, GFP_NOFS);
+	char *buf = kmalloc(info->dqi_usable_bs, GFP_KERNEL);
 	loff_t ret = 0;
 	int i;
 	char *ddquot;
@@ -613,18 +658,20 @@ out_buf:
 
 /* Find entry for given id in the tree */
 static loff_t find_tree_dqentry(struct qtree_mem_dqinfo *info,
-				struct dquot *dquot, uint blk, int depth)
+				struct dquot *dquot, uint *blks, int depth)
 {
-	char *buf = kmalloc(info->dqi_usable_bs, GFP_NOFS);
+	char *buf = kmalloc(info->dqi_usable_bs, GFP_KERNEL);
 	loff_t ret = 0;
 	__le32 *ref = (__le32 *)buf;
+	uint blk;
+	int i;
 
 	if (!buf)
 		return -ENOMEM;
-	ret = read_blk(info, blk, buf);
+	ret = read_blk(info, blks[depth], buf);
 	if (ret < 0) {
 		quota_error(dquot->dq_sb, "Can't read quota tree block %u",
-			    blk);
+			    blks[depth]);
 		goto out_buf;
 	}
 	ret = 0;
@@ -636,8 +683,19 @@ static loff_t find_tree_dqentry(struct qtree_mem_dqinfo *info,
 	if (ret)
 		goto out_buf;
 
+	/* Check for cycles in the tree */
+	for (i = 0; i <= depth; i++)
+		if (blk == blks[i]) {
+			quota_error(dquot->dq_sb,
+				"Cycle in quota tree detected: block %u index %u",
+				blks[depth],
+				get_index(info, dquot->dq_id, depth));
+			ret = -EIO;
+			goto out_buf;
+		}
+	blks[depth + 1] = blk;
 	if (depth < info->dqi_qtree_depth - 1)
-		ret = find_tree_dqentry(info, dquot, blk, depth+1);
+		ret = find_tree_dqentry(info, dquot, blks, depth + 1);
 	else
 		ret = find_block_dqentry(info, dquot, blk);
 out_buf:
@@ -649,7 +707,13 @@ out_buf:
 static inline loff_t find_dqentry(struct qtree_mem_dqinfo *info,
 				  struct dquot *dquot)
 {
-	return find_tree_dqentry(info, dquot, QT_TREEOFF, 0);
+	uint blks[MAX_QTREE_DEPTH] = { QT_TREEOFF };
+
+	if (info->dqi_qtree_depth >= MAX_QTREE_DEPTH) {
+		quota_error(dquot->dq_sb, "Quota tree depth too big!");
+		return -EIO;
+	}
+	return find_tree_dqentry(info, dquot, blks, 0);
 }
 
 int qtree_read_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot)
@@ -684,7 +748,7 @@ int qtree_read_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot)
 		}
 		dquot->dq_off = offset;
 	}
-	ddquot = kmalloc(info->dqi_entry_size, GFP_NOFS);
+	ddquot = kmalloc(info->dqi_entry_size, GFP_KERNEL);
 	if (!ddquot)
 		return -ENOMEM;
 	ret = sb->s_op->quota_read(sb, type, ddquot, info->dqi_entry_size,
@@ -728,7 +792,7 @@ EXPORT_SYMBOL(qtree_release_dquot);
 static int find_next_id(struct qtree_mem_dqinfo *info, qid_t *id,
 			unsigned int blk, int depth)
 {
-	char *buf = kmalloc(info->dqi_usable_bs, GFP_NOFS);
+	char *buf = kmalloc(info->dqi_usable_bs, GFP_KERNEL);
 	__le32 *ref = (__le32 *)buf;
 	ssize_t ret;
 	unsigned int epb = info->dqi_usable_bs >> 2;
diff --git a/fs/quota/quota_v1.c b/fs/quota/quota_v1.c
index a0db3f195e..3f3e8acc05 100644
--- a/fs/quota/quota_v1.c
+++ b/fs/quota/quota_v1.c
@@ -160,9 +160,11 @@ static int v1_read_file_info(struct super_block *sb, int type)
 {
 	struct quota_info *dqopt = sb_dqopt(sb);
 	struct v1_disk_dqblk dqblk;
+	unsigned int memalloc;
 	int ret;
 
 	down_read(&dqopt->dqio_sem);
+	memalloc = memalloc_nofs_save();
 	ret = sb->s_op->quota_read(sb, type, (char *)&dqblk,
 				sizeof(struct v1_disk_dqblk), v1_dqoff(0));
 	if (ret != sizeof(struct v1_disk_dqblk)) {
@@ -179,6 +181,7 @@ static int v1_read_file_info(struct super_block *sb, int type)
 	dqopt->info[type].dqi_bgrace =
 			dqblk.dqb_btime ? dqblk.dqb_btime : MAX_DQ_TIME;
 out:
+	memalloc_nofs_restore(memalloc);
 	up_read(&dqopt->dqio_sem);
 	return ret;
 }
@@ -187,9 +190,11 @@ static int v1_write_file_info(struct super_block *sb, int type)
 {
 	struct quota_info *dqopt = sb_dqopt(sb);
 	struct v1_disk_dqblk dqblk;
+	unsigned int memalloc;
 	int ret;
 
 	down_write(&dqopt->dqio_sem);
+	memalloc = memalloc_nofs_save();
 	ret = sb->s_op->quota_read(sb, type, (char *)&dqblk,
 				sizeof(struct v1_disk_dqblk), v1_dqoff(0));
 	if (ret != sizeof(struct v1_disk_dqblk)) {
@@ -209,6 +214,7 @@ static int v1_write_file_info(struct super_block *sb, int type)
 	else if (ret >= 0)
 		ret = -EIO;
 out:
+	memalloc_nofs_restore(memalloc);
 	up_write(&dqopt->dqio_sem);
 	return ret;
 }
diff --git a/fs/quota/quota_v2.c b/fs/quota/quota_v2.c
index ae99e7b882..c48c233f3b 100644
--- a/fs/quota/quota_v2.c
+++ b/fs/quota/quota_v2.c
@@ -96,9 +96,11 @@ static int v2_read_file_info(struct super_block *sb, int type)
 	struct qtree_mem_dqinfo *qinfo;
 	ssize_t size;
 	unsigned int version;
+	unsigned int memalloc;
 	int ret;
 
 	down_read(&dqopt->dqio_sem);
+	memalloc = memalloc_nofs_save();
 	ret = v2_read_header(sb, type, &dqhead);
 	if (ret < 0)
 		goto out;
@@ -119,7 +121,7 @@ static int v2_read_file_info(struct super_block *sb, int type)
 			ret = -EIO;
 		goto out;
 	}
-	info->dqi_priv = kmalloc(sizeof(struct qtree_mem_dqinfo), GFP_NOFS);
+	info->dqi_priv = kmalloc(sizeof(struct qtree_mem_dqinfo), GFP_KERNEL);
 	if (!info->dqi_priv) {
 		ret = -ENOMEM;
 		goto out;
@@ -166,14 +168,17 @@ static int v2_read_file_info(struct super_block *sb, int type)
 		    i_size_read(sb_dqopt(sb)->files[type]));
 		goto out_free;
 	}
-	if (qinfo->dqi_free_blk >= qinfo->dqi_blocks) {
-		quota_error(sb, "Free block number too big (%u >= %u).",
-			    qinfo->dqi_free_blk, qinfo->dqi_blocks);
+	if (qinfo->dqi_free_blk && (qinfo->dqi_free_blk <= QT_TREEOFF ||
+	    qinfo->dqi_free_blk >= qinfo->dqi_blocks)) {
+		quota_error(sb, "Free block number %u out of range (%u, %u).",
+			    qinfo->dqi_free_blk, QT_TREEOFF, qinfo->dqi_blocks);
 		goto out_free;
 	}
-	if (qinfo->dqi_free_entry >= qinfo->dqi_blocks) {
-		quota_error(sb, "Block with free entry too big (%u >= %u).",
-			    qinfo->dqi_free_entry, qinfo->dqi_blocks);
+	if (qinfo->dqi_free_entry && (qinfo->dqi_free_entry <= QT_TREEOFF ||
+	    qinfo->dqi_free_entry >= qinfo->dqi_blocks)) {
+		quota_error(sb, "Block with free entry %u out of range (%u, %u).",
+			    qinfo->dqi_free_entry, QT_TREEOFF,
+			    qinfo->dqi_blocks);
 		goto out_free;
 	}
 	ret = 0;
@@ -183,6 +188,7 @@ out_free:
 		info->dqi_priv = NULL;
 	}
 out:
+	memalloc_nofs_restore(memalloc);
 	up_read(&dqopt->dqio_sem);
 	return ret;
 }
@@ -195,8 +201,10 @@ static int v2_write_file_info(struct super_block *sb, int type)
 	struct mem_dqinfo *info = &dqopt->info[type];
 	struct qtree_mem_dqinfo *qinfo = info->dqi_priv;
 	ssize_t size;
+	unsigned int memalloc;
 
 	down_write(&dqopt->dqio_sem);
+	memalloc = memalloc_nofs_save();
 	spin_lock(&dq_data_lock);
 	info->dqi_flags &= ~DQF_INFO_DIRTY;
 	dinfo.dqi_bgrace = cpu_to_le32(info->dqi_bgrace);
@@ -209,6 +217,7 @@ static int v2_write_file_info(struct super_block *sb, int type)
 	dinfo.dqi_free_entry = cpu_to_le32(qinfo->dqi_free_entry);
 	size = sb->s_op->quota_write(sb, type, (char *)&dinfo,
 	       sizeof(struct v2_disk_dqinfo), V2_DQINFOOFF);
+	memalloc_nofs_restore(memalloc);
 	up_write(&dqopt->dqio_sem);
 	if (size != sizeof(struct v2_disk_dqinfo)) {
 		quota_error(sb, "Can't write info structure");
@@ -328,11 +337,14 @@ static int v2_read_dquot(struct dquot *dquot)
 {
 	struct quota_info *dqopt = sb_dqopt(dquot->dq_sb);
 	int ret;
+	unsigned int memalloc;
 
 	down_read(&dqopt->dqio_sem);
+	memalloc = memalloc_nofs_save();
 	ret = qtree_read_dquot(
 			sb_dqinfo(dquot->dq_sb, dquot->dq_id.type)->dqi_priv,
 			dquot);
+	memalloc_nofs_restore(memalloc);
 	up_read(&dqopt->dqio_sem);
 	return ret;
 }
@@ -342,6 +354,7 @@ static int v2_write_dquot(struct dquot *dquot)
 	struct quota_info *dqopt = sb_dqopt(dquot->dq_sb);
 	int ret;
 	bool alloc = false;
+	unsigned int memalloc;
 
 	/*
 	 * If space for dquot is already allocated, we don't need any
@@ -355,9 +368,11 @@ static int v2_write_dquot(struct dquot *dquot)
 	} else {
 		down_read(&dqopt->dqio_sem);
 	}
+	memalloc = memalloc_nofs_save();
 	ret = qtree_write_dquot(
 			sb_dqinfo(dquot->dq_sb, dquot->dq_id.type)->dqi_priv,
 			dquot);
+	memalloc_nofs_restore(memalloc);
 	if (alloc)
 		up_write(&dqopt->dqio_sem);
 	else
@@ -368,10 +383,13 @@ static int v2_write_dquot(struct dquot *dquot)
 static int v2_release_dquot(struct dquot *dquot)
 {
 	struct quota_info *dqopt = sb_dqopt(dquot->dq_sb);
+	unsigned int memalloc;
 	int ret;
 
 	down_write(&dqopt->dqio_sem);
+	memalloc = memalloc_nofs_save();
 	ret = qtree_release_dquot(sb_dqinfo(dquot->dq_sb, dquot->dq_id.type)->dqi_priv, dquot);
+	memalloc_nofs_restore(memalloc);
 	up_write(&dqopt->dqio_sem);
 
 	return ret;
@@ -386,10 +404,13 @@ static int v2_free_file_info(struct super_block *sb, int type)
 static int v2_get_next_id(struct super_block *sb, struct kqid *qid)
 {
 	struct quota_info *dqopt = sb_dqopt(sb);
+	unsigned int memalloc;
 	int ret;
 
 	down_read(&dqopt->dqio_sem);
+	memalloc = memalloc_nofs_save();
 	ret = qtree_get_next_id(sb_dqinfo(sb, qid->type)->dqi_priv, qid);
+	memalloc_nofs_restore(memalloc);
 	up_read(&dqopt->dqio_sem);
 	return ret;
 }
diff --git a/fs/ramfs/file-mmu.c b/fs/ramfs/file-mmu.c
index c7a1aa3c88..b45c7edc32 100644
--- a/fs/ramfs/file-mmu.c
+++ b/fs/ramfs/file-mmu.c
@@ -35,7 +35,7 @@ static unsigned long ramfs_mmu_get_unmapped_area(struct file *file,
 		unsigned long addr, unsigned long len, unsigned long pgoff,
 		unsigned long flags)
 {
-	return current->mm->get_unmapped_area(file, addr, len, pgoff, flags);
+	return mm_get_unmapped_area(current->mm, file, addr, len, pgoff, flags);
 }
 
 const struct file_operations ramfs_file_operations = {
diff --git a/fs/ramfs/inode.c b/fs/ramfs/inode.c
index 4ac05a9e25..8006faaaf0 100644
--- a/fs/ramfs/inode.c
+++ b/fs/ramfs/inode.c
@@ -102,11 +102,20 @@ ramfs_mknod(struct mnt_idmap *idmap, struct inode *dir,
 	int error = -ENOSPC;
 
 	if (inode) {
+		error = security_inode_init_security(inode, dir,
+						     &dentry->d_name, NULL,
+						     NULL);
+		if (error) {
+			iput(inode);
+			goto out;
+		}
+
 		d_instantiate(dentry, inode);
 		dget(dentry);	/* Extra count - pin the dentry in core */
 		error = 0;
 		inode_set_mtime_to_ts(dir, inode_set_ctime_current(dir));
 	}
+out:
 	return error;
 }
 
@@ -134,6 +143,15 @@ static int ramfs_symlink(struct mnt_idmap *idmap, struct inode *dir,
 	inode = ramfs_get_inode(dir->i_sb, dir, S_IFLNK|S_IRWXUGO, 0);
 	if (inode) {
 		int l = strlen(symname)+1;
+
+		error = security_inode_init_security(inode, dir,
+						     &dentry->d_name, NULL,
+						     NULL);
+		if (error) {
+			iput(inode);
+			goto out;
+		}
+
 		error = page_symlink(inode, symname, l);
 		if (!error) {
 			d_instantiate(dentry, inode);
@@ -143,6 +161,7 @@ static int ramfs_symlink(struct mnt_idmap *idmap, struct inode *dir,
 		} else
 			iput(inode);
 	}
+out:
 	return error;
 }
 
@@ -150,12 +169,23 @@ static int ramfs_tmpfile(struct mnt_idmap *idmap,
 			 struct inode *dir, struct file *file, umode_t mode)
 {
 	struct inode *inode;
+	int error;
 
 	inode = ramfs_get_inode(dir->i_sb, dir, mode, 0);
 	if (!inode)
 		return -ENOSPC;
+
+	error = security_inode_init_security(inode, dir,
+					     &file_dentry(file)->d_name, NULL,
+					     NULL);
+	if (error) {
+		iput(inode);
+		goto out;
+	}
+
 	d_tmpfile(file, inode);
-	return finish_open_simple(file, 0);
+out:
+	return finish_open_simple(file, error);
 }
 
 static const struct inode_operations ramfs_dir_inode_operations = {
diff --git a/fs/read_write.c b/fs/read_write.c
index d4c036e82b..ef63393913 100644
--- a/fs/read_write.c
+++ b/fs/read_write.c
@@ -392,7 +392,7 @@ static ssize_t new_sync_read(struct file *filp, char __user *buf, size_t len, lo
 	kiocb.ki_pos = (ppos ? *ppos : 0);
 	iov_iter_ubuf(&iter, ITER_DEST, buf, len);
 
-	ret = call_read_iter(filp, &kiocb, &iter);
+	ret = filp->f_op->read_iter(&kiocb, &iter);
 	BUG_ON(ret == -EIOCBQUEUED);
 	if (ppos)
 		*ppos = kiocb.ki_pos;
@@ -494,7 +494,7 @@ static ssize_t new_sync_write(struct file *filp, const char __user *buf, size_t
 	kiocb.ki_pos = (ppos ? *ppos : 0);
 	iov_iter_ubuf(&iter, ITER_SOURCE, (void __user *)buf, len);
 
-	ret = call_write_iter(filp, &kiocb, &iter);
+	ret = filp->f_op->write_iter(&kiocb, &iter);
 	BUG_ON(ret == -EIOCBQUEUED);
 	if (ret > 0 && ppos)
 		*ppos = kiocb.ki_pos;
@@ -736,9 +736,9 @@ static ssize_t do_iter_readv_writev(struct file *filp, struct iov_iter *iter,
 	kiocb.ki_pos = (ppos ? *ppos : 0);
 
 	if (type == READ)
-		ret = call_read_iter(filp, &kiocb, iter);
+		ret = filp->f_op->read_iter(&kiocb, iter);
 	else
-		ret = call_write_iter(filp, &kiocb, iter);
+		ret = filp->f_op->write_iter(&kiocb, iter);
 	BUG_ON(ret == -EIOCBQUEUED);
 	if (ppos)
 		*ppos = kiocb.ki_pos;
@@ -799,7 +799,7 @@ ssize_t vfs_iocb_iter_read(struct file *file, struct kiocb *iocb,
 	if (ret < 0)
 		return ret;
 
-	ret = call_read_iter(file, iocb, iter);
+	ret = file->f_op->read_iter(iocb, iter);
 out:
 	if (ret >= 0)
 		fsnotify_access(file);
@@ -860,7 +860,7 @@ ssize_t vfs_iocb_iter_write(struct file *file, struct kiocb *iocb,
 		return ret;
 
 	kiocb_start_write(iocb);
-	ret = call_write_iter(file, iocb, iter);
+	ret = file->f_op->write_iter(iocb, iter);
 	if (ret != -EIOCBQUEUED)
 		kiocb_end_write(iocb);
 	if (ret > 0)
@@ -1667,6 +1667,7 @@ int generic_write_check_limits(struct file *file, loff_t pos, loff_t *count)
 
 	return 0;
 }
+EXPORT_SYMBOL_GPL(generic_write_check_limits);
 
 /* Like generic_write_checks(), but takes size of write instead of iter. */
 int generic_write_checks_count(struct kiocb *iocb, loff_t *count)
@@ -1685,7 +1686,7 @@ int generic_write_checks_count(struct kiocb *iocb, loff_t *count)
 
 	if ((iocb->ki_flags & IOCB_NOWAIT) &&
 	    !((iocb->ki_flags & IOCB_DIRECT) ||
-	      (file->f_mode & FMODE_BUF_WASYNC)))
+	      (file->f_op->fop_flags & FOP_BUFFER_WASYNC)))
 		return -EINVAL;
 
 	return generic_write_check_limits(iocb->ki_filp, iocb->ki_pos, count);
diff --git a/fs/reiserfs/README b/fs/reiserfs/README
index e2f7a264e3..11e9ecf24b 100644
--- a/fs/reiserfs/README
+++ b/fs/reiserfs/README
@@ -102,19 +102,9 @@ that start on a node aligned boundary (there are reasons to want to node
 align files), and he invented and implemented indirect items and
 unformatted nodes as the solution.
 
-Konstantin Shvachko, with the help of the Russian version of a VC,
-tried to put me in a position where I was forced into giving control
-of the project to him.  (Fortunately, as the person paying the money
-for all salaries from my dayjob I owned all copyrights, and you can't
-really force takeovers of sole proprietorships.)  This was something
-curious, because he never really understood the value of our project,
-why we should do what we do, or why innovation was possible in
-general, but he was sure that he ought to be controlling it.  Every
-innovation had to be forced past him while he was with us.  He added
-two years to the time required to complete reiserfs, and was a net
-loss for me.  Mikhail Gilula was a brilliant innovator who also left
-in a destructive way that erased the value of his contributions, and
-that he was shown much generosity just makes it more painful.
+Konstantin Shvachko was taking part in the early days.
+
+Mikhail Gilula was a brilliant innovator that has shown much generosity.
 
 Grigory Zaigralin was an extremely effective system administrator for
 our group.
diff --git a/fs/reiserfs/inode.c b/fs/reiserfs/inode.c
index 1d825459ee..c1daedc50f 100644
--- a/fs/reiserfs/inode.c
+++ b/fs/reiserfs/inode.c
@@ -2503,8 +2503,8 @@ out:
  * start/recovery path as __block_write_full_folio, along with special
  * code to handle reiserfs tails.
  */
-static int reiserfs_write_full_folio(struct folio *folio,
-				    struct writeback_control *wbc)
+static int reiserfs_write_folio(struct folio *folio,
+		struct writeback_control *wbc, void *data)
 {
 	struct inode *inode = folio->mapping->host;
 	unsigned long end_index = inode->i_size >> PAGE_SHIFT;
@@ -2721,12 +2721,11 @@ static int reiserfs_read_folio(struct file *f, struct folio *folio)
 	return block_read_full_folio(folio, reiserfs_get_block);
 }
 
-static int reiserfs_writepage(struct page *page, struct writeback_control *wbc)
+static int reiserfs_writepages(struct address_space *mapping,
+		struct writeback_control *wbc)
 {
-	struct folio *folio = page_folio(page);
-	struct inode *inode = folio->mapping->host;
-	reiserfs_wait_on_write_block(inode->i_sb);
-	return reiserfs_write_full_folio(folio, wbc);
+	reiserfs_wait_on_write_block(mapping->host->i_sb);
+	return write_cache_pages(mapping, wbc, reiserfs_write_folio, NULL);
 }
 
 static void reiserfs_truncate_failed_write(struct inode *inode)
@@ -3405,7 +3404,7 @@ out:
 }
 
 const struct address_space_operations reiserfs_address_space_operations = {
-	.writepage = reiserfs_writepage,
+	.writepages = reiserfs_writepages,
 	.read_folio = reiserfs_read_folio,
 	.readahead = reiserfs_readahead,
 	.release_folio = reiserfs_release_folio,
@@ -3415,4 +3414,5 @@ const struct address_space_operations reiserfs_address_space_operations = {
 	.bmap = reiserfs_aop_bmap,
 	.direct_IO = reiserfs_direct_IO,
 	.dirty_folio = reiserfs_dirty_folio,
+	.migrate_folio = buffer_migrate_folio,
 };
diff --git a/fs/reiserfs/item_ops.c b/fs/reiserfs/item_ops.c
index 3a5a752d96..5011c10287 100644
--- a/fs/reiserfs/item_ops.c
+++ b/fs/reiserfs/item_ops.c
@@ -389,16 +389,9 @@ static void direntry_print_item(struct item_head *ih, char *item)
 		name = item + deh_location(deh);
 		if (name[namelen - 1] == 0)
 			namelen = strlen(name);
-		namebuf[0] = '"';
-		if (namelen > sizeof(namebuf) - 3) {
-			strncpy(namebuf + 1, name, sizeof(namebuf) - 3);
-			namebuf[sizeof(namebuf) - 2] = '"';
-			namebuf[sizeof(namebuf) - 1] = 0;
-		} else {
-			memcpy(namebuf + 1, name, namelen);
-			namebuf[namelen + 1] = '"';
-			namebuf[namelen + 2] = 0;
-		}
+
+		scnprintf(namebuf, sizeof(namebuf), "\"%.*s\"",
+			  (int)sizeof(namebuf)-3, name);
 
 		printk("%d:  %-15s%-15d%-15d%-15lld%-15lld(%s)\n",
 		       i, namebuf,
diff --git a/fs/reiserfs/journal.c b/fs/reiserfs/journal.c
index 171c912af5..e477ee0ff3 100644
--- a/fs/reiserfs/journal.c
+++ b/fs/reiserfs/journal.c
@@ -2386,7 +2386,7 @@ static int journal_read(struct super_block *sb)
 
 	cur_dblock = SB_ONDISK_JOURNAL_1st_BLOCK(sb);
 	reiserfs_info(sb, "checking transaction log (%pg)\n",
-		      journal->j_bdev_handle->bdev);
+		      file_bdev(journal->j_bdev_file));
 	start = ktime_get_seconds();
 
 	/*
@@ -2447,7 +2447,7 @@ static int journal_read(struct super_block *sb)
 		 * device and journal device to be the same
 		 */
 		d_bh =
-		    reiserfs_breada(journal->j_bdev_handle->bdev, cur_dblock,
+		    reiserfs_breada(file_bdev(journal->j_bdev_file), cur_dblock,
 				    sb->s_blocksize,
 				    SB_ONDISK_JOURNAL_1st_BLOCK(sb) +
 				    SB_ONDISK_JOURNAL_SIZE(sb));
@@ -2588,9 +2588,9 @@ static void journal_list_init(struct super_block *sb)
 
 static void release_journal_dev(struct reiserfs_journal *journal)
 {
-	if (journal->j_bdev_handle) {
-		bdev_release(journal->j_bdev_handle);
-		journal->j_bdev_handle = NULL;
+	if (journal->j_bdev_file) {
+		bdev_fput(journal->j_bdev_file);
+		journal->j_bdev_file = NULL;
 	}
 }
 
@@ -2605,7 +2605,7 @@ static int journal_init_dev(struct super_block *super,
 
 	result = 0;
 
-	journal->j_bdev_handle = NULL;
+	journal->j_bdev_file = NULL;
 	jdev = SB_ONDISK_JOURNAL_DEVICE(super) ?
 	    new_decode_dev(SB_ONDISK_JOURNAL_DEVICE(super)) : super->s_dev;
 
@@ -2616,37 +2616,36 @@ static int journal_init_dev(struct super_block *super,
 	if ((!jdev_name || !jdev_name[0])) {
 		if (jdev == super->s_dev)
 			holder = NULL;
-		journal->j_bdev_handle = bdev_open_by_dev(jdev, blkdev_mode,
+		journal->j_bdev_file = bdev_file_open_by_dev(jdev, blkdev_mode,
 							  holder, NULL);
-		if (IS_ERR(journal->j_bdev_handle)) {
-			result = PTR_ERR(journal->j_bdev_handle);
-			journal->j_bdev_handle = NULL;
+		if (IS_ERR(journal->j_bdev_file)) {
+			result = PTR_ERR(journal->j_bdev_file);
+			journal->j_bdev_file = NULL;
 			reiserfs_warning(super, "sh-458",
 					 "cannot init journal device unknown-block(%u,%u): %i",
 					 MAJOR(jdev), MINOR(jdev), result);
 			return result;
 		} else if (jdev != super->s_dev)
-			set_blocksize(journal->j_bdev_handle->bdev,
-				      super->s_blocksize);
+			set_blocksize(journal->j_bdev_file, super->s_blocksize);
 
 		return 0;
 	}
 
-	journal->j_bdev_handle = bdev_open_by_path(jdev_name, blkdev_mode,
+	journal->j_bdev_file = bdev_file_open_by_path(jdev_name, blkdev_mode,
 						   holder, NULL);
-	if (IS_ERR(journal->j_bdev_handle)) {
-		result = PTR_ERR(journal->j_bdev_handle);
-		journal->j_bdev_handle = NULL;
+	if (IS_ERR(journal->j_bdev_file)) {
+		result = PTR_ERR(journal->j_bdev_file);
+		journal->j_bdev_file = NULL;
 		reiserfs_warning(super, "sh-457",
 				 "journal_init_dev: Cannot open '%s': %i",
 				 jdev_name, result);
 		return result;
 	}
 
-	set_blocksize(journal->j_bdev_handle->bdev, super->s_blocksize);
+	set_blocksize(journal->j_bdev_file, super->s_blocksize);
 	reiserfs_info(super,
 		      "journal_init_dev: journal device: %pg\n",
-		      journal->j_bdev_handle->bdev);
+		      file_bdev(journal->j_bdev_file));
 	return 0;
 }
 
@@ -2804,7 +2803,7 @@ int journal_init(struct super_block *sb, const char *j_dev_name,
 				 "journal header magic %x (device %pg) does "
 				 "not match to magic found in super block %x",
 				 jh->jh_journal.jp_journal_magic,
-				 journal->j_bdev_handle->bdev,
+				 file_bdev(journal->j_bdev_file),
 				 sb_jp_journal_magic(rs));
 		brelse(bhjh);
 		goto free_and_return;
@@ -2828,7 +2827,7 @@ int journal_init(struct super_block *sb, const char *j_dev_name,
 	reiserfs_info(sb, "journal params: device %pg, size %u, "
 		      "journal first block %u, max trans len %u, max batch %u, "
 		      "max commit age %u, max trans age %u\n",
-		      journal->j_bdev_handle->bdev,
+		      file_bdev(journal->j_bdev_file),
 		      SB_ONDISK_JOURNAL_SIZE(sb),
 		      SB_ONDISK_JOURNAL_1st_BLOCK(sb),
 		      journal->j_trans_max,
diff --git a/fs/reiserfs/procfs.c b/fs/reiserfs/procfs.c
index 83cb9402e0..5c68a4a52d 100644
--- a/fs/reiserfs/procfs.c
+++ b/fs/reiserfs/procfs.c
@@ -354,7 +354,7 @@ static int show_journal(struct seq_file *m, void *unused)
 		   "prepare: \t%12lu\n"
 		   "prepare_retry: \t%12lu\n",
 		   DJP(jp_journal_1st_block),
-		   SB_JOURNAL(sb)->j_bdev_handle->bdev,
+		   file_bdev(SB_JOURNAL(sb)->j_bdev_file),
 		   DJP(jp_journal_dev),
 		   DJP(jp_journal_size),
 		   DJP(jp_journal_trans_max),
diff --git a/fs/reiserfs/reiserfs.h b/fs/reiserfs/reiserfs.h
index b655491645..f0e1f29f20 100644
--- a/fs/reiserfs/reiserfs.h
+++ b/fs/reiserfs/reiserfs.h
@@ -299,7 +299,7 @@ struct reiserfs_journal {
 	/* oldest journal block.  start here for traverse */
 	struct reiserfs_journal_cnode *j_first;
 
-	struct bdev_handle *j_bdev_handle;
+	struct file *j_bdev_file;
 
 	/* first block on s_dev of reserved area journal */
 	int j_1st_reserved_block;
@@ -2810,10 +2810,10 @@ struct reiserfs_journal_header {
 
 /* We need these to make journal.c code more readable */
 #define journal_find_get_block(s, block) __find_get_block(\
-		SB_JOURNAL(s)->j_bdev_handle->bdev, block, s->s_blocksize)
-#define journal_getblk(s, block) __getblk(SB_JOURNAL(s)->j_bdev_handle->bdev,\
+		file_bdev(SB_JOURNAL(s)->j_bdev_file), block, s->s_blocksize)
+#define journal_getblk(s, block) __getblk(file_bdev(SB_JOURNAL(s)->j_bdev_file),\
 		block, s->s_blocksize)
-#define journal_bread(s, block) __bread(SB_JOURNAL(s)->j_bdev_handle->bdev,\
+#define journal_bread(s, block) __bread(file_bdev(SB_JOURNAL(s)->j_bdev_file),\
 		block, s->s_blocksize)
 
 enum reiserfs_bh_state_bits {
diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c
index 7b3d5aeb2a..ab76468da0 100644
--- a/fs/reiserfs/super.c
+++ b/fs/reiserfs/super.c
@@ -670,7 +670,6 @@ static int __init init_inodecache(void)
 						  sizeof(struct
 							 reiserfs_inode_info),
 						  0, (SLAB_RECLAIM_ACCOUNT|
-						      SLAB_MEM_SPREAD|
 						      SLAB_ACCOUNT),
 						  init_once);
 	if (reiserfs_inode_cachep == NULL)
diff --git a/fs/remap_range.c b/fs/remap_range.c
index de07f978ce..28246dfc84 100644
--- a/fs/remap_range.c
+++ b/fs/remap_range.c
@@ -99,8 +99,7 @@ static int generic_remap_checks(struct file *file_in, loff_t pos_in,
 	return 0;
 }
 
-static int remap_verify_area(struct file *file, loff_t pos, loff_t len,
-			     bool write)
+int remap_verify_area(struct file *file, loff_t pos, loff_t len, bool write)
 {
 	int mask = write ? MAY_WRITE : MAY_READ;
 	loff_t tmp;
@@ -118,6 +117,7 @@ static int remap_verify_area(struct file *file, loff_t pos, loff_t len,
 
 	return fsnotify_file_area_perm(file, mask, &pos, len);
 }
+EXPORT_SYMBOL_GPL(remap_verify_area);
 
 /*
  * Ensure that we don't remap a partial EOF block in the middle of something
diff --git a/fs/romfs/super.c b/fs/romfs/super.c
index 545ad44f96..2cbb924620 100644
--- a/fs/romfs/super.c
+++ b/fs/romfs/super.c
@@ -594,7 +594,7 @@ static void romfs_kill_sb(struct super_block *sb)
 #ifdef CONFIG_ROMFS_ON_BLOCK
 	if (sb->s_bdev) {
 		sync_blockdev(sb->s_bdev);
-		bdev_release(sb->s_bdev_handle);
+		bdev_fput(sb->s_bdev_file);
 	}
 #endif
 }
@@ -630,8 +630,8 @@ static int __init init_romfs_fs(void)
 	romfs_inode_cachep =
 		kmem_cache_create("romfs_i",
 				  sizeof(struct romfs_inode_info), 0,
-				  SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD |
-				  SLAB_ACCOUNT, romfs_i_init_once);
+				  SLAB_RECLAIM_ACCOUNT | SLAB_ACCOUNT,
+				  romfs_i_init_once);
 
 	if (!romfs_inode_cachep) {
 		pr_err("Failed to initialise inode cache\n");
diff --git a/fs/select.c b/fs/select.c
index d4d881d439..9515c3fa1a 100644
--- a/fs/select.c
+++ b/fs/select.c
@@ -839,7 +839,7 @@ SYSCALL_DEFINE1(old_select, struct sel_arg_struct __user *, arg)
 
 struct poll_list {
 	struct poll_list *next;
-	int len;
+	unsigned int len;
 	struct pollfd entries[];
 };
 
@@ -975,14 +975,15 @@ static int do_sys_poll(struct pollfd __user *ufds, unsigned int nfds,
 		struct timespec64 *end_time)
 {
 	struct poll_wqueues table;
-	int err = -EFAULT, fdcount, len;
+	int err = -EFAULT, fdcount;
 	/* Allocate small arguments on the stack to save memory and be
 	   faster - use long to make sure the buffer is aligned properly
 	   on 64 bit archs to avoid unaligned access */
 	long stack_pps[POLL_STACK_ALLOC/sizeof(long)];
 	struct poll_list *const head = (struct poll_list *)stack_pps;
  	struct poll_list *walk = head;
- 	unsigned long todo = nfds;
+	unsigned int todo = nfds;
+	unsigned int len;
 
 	if (nfds > rlimit(RLIMIT_NOFILE))
 		return -EINVAL;
@@ -998,9 +999,9 @@ static int do_sys_poll(struct pollfd __user *ufds, unsigned int nfds,
 					sizeof(struct pollfd) * walk->len))
 			goto out_fds;
 
-		todo -= walk->len;
-		if (!todo)
+		if (walk->len >= todo)
 			break;
+		todo -= walk->len;
 
 		len = min(todo, POLLFD_PER_PAGE);
 		walk = walk->next = kmalloc(struct_size(walk, entries, len),
@@ -1020,7 +1021,7 @@ static int do_sys_poll(struct pollfd __user *ufds, unsigned int nfds,
 
 	for (walk = head; walk; walk = walk->next) {
 		struct pollfd *fds = walk->entries;
-		int j;
+		unsigned int j;
 
 		for (j = walk->len; j; fds++, ufds++, j--)
 			unsafe_put_user(fds->revents, &ufds->revents, Efault);
diff --git a/fs/seq_file.c b/fs/seq_file.c
index f5fdaf3b15..e676c8b0cf 100644
--- a/fs/seq_file.c
+++ b/fs/seq_file.c
@@ -669,18 +669,11 @@ void seq_putc(struct seq_file *m, char c)
 }
 EXPORT_SYMBOL(seq_putc);
 
-void seq_puts(struct seq_file *m, const char *s)
+void __seq_puts(struct seq_file *m, const char *s)
 {
-	int len = strlen(s);
-
-	if (m->count + len >= m->size) {
-		seq_set_overflow(m);
-		return;
-	}
-	memcpy(m->buf + m->count, s, len);
-	m->count += len;
+	seq_write(m, s, strlen(s));
 }
-EXPORT_SYMBOL(seq_puts);
+EXPORT_SYMBOL(__seq_puts);
 
 /**
  * seq_put_decimal_ull_width - A helper routine for putting decimal numbers
diff --git a/fs/signalfd.c b/fs/signalfd.c
index e20d1484c6..ec7b2da247 100644
--- a/fs/signalfd.c
+++ b/fs/signalfd.c
@@ -68,8 +68,7 @@ static __poll_t signalfd_poll(struct file *file, poll_table *wait)
 /*
  * Copied from copy_siginfo_to_user() in kernel/signal.c
  */
-static int signalfd_copyinfo(struct signalfd_siginfo __user *uinfo,
-			     kernel_siginfo_t const *kinfo)
+static int signalfd_copyinfo(struct iov_iter *to, kernel_siginfo_t const *kinfo)
 {
 	struct signalfd_siginfo new;
 
@@ -146,10 +145,10 @@ static int signalfd_copyinfo(struct signalfd_siginfo __user *uinfo,
 		break;
 	}
 
-	if (copy_to_user(uinfo, &new, sizeof(struct signalfd_siginfo)))
+	if (!copy_to_iter_full(&new, sizeof(struct signalfd_siginfo), to))
 		return -EFAULT;
 
-	return sizeof(*uinfo);
+	return sizeof(struct signalfd_siginfo);
 }
 
 static ssize_t signalfd_dequeue(struct signalfd_ctx *ctx, kernel_siginfo_t *info,
@@ -199,28 +198,27 @@ static ssize_t signalfd_dequeue(struct signalfd_ctx *ctx, kernel_siginfo_t *info
  * error code. The "count" parameter must be at least the size of a
  * "struct signalfd_siginfo".
  */
-static ssize_t signalfd_read(struct file *file, char __user *buf, size_t count,
-			     loff_t *ppos)
+static ssize_t signalfd_read_iter(struct kiocb *iocb, struct iov_iter *to)
 {
+	struct file *file = iocb->ki_filp;
 	struct signalfd_ctx *ctx = file->private_data;
-	struct signalfd_siginfo __user *siginfo;
-	int nonblock = file->f_flags & O_NONBLOCK;
+	size_t count = iov_iter_count(to);
 	ssize_t ret, total = 0;
 	kernel_siginfo_t info;
+	bool nonblock;
 
 	count /= sizeof(struct signalfd_siginfo);
 	if (!count)
 		return -EINVAL;
 
-	siginfo = (struct signalfd_siginfo __user *) buf;
+	nonblock = file->f_flags & O_NONBLOCK || iocb->ki_flags & IOCB_NOWAIT;
 	do {
 		ret = signalfd_dequeue(ctx, &info, nonblock);
 		if (unlikely(ret <= 0))
 			break;
-		ret = signalfd_copyinfo(siginfo, &info);
+		ret = signalfd_copyinfo(to, &info);
 		if (ret < 0)
 			break;
-		siginfo++;
 		total += ret;
 		nonblock = 1;
 	} while (--count);
@@ -246,7 +244,7 @@ static const struct file_operations signalfd_fops = {
 #endif
 	.release	= signalfd_release,
 	.poll		= signalfd_poll,
-	.read		= signalfd_read,
+	.read_iter	= signalfd_read_iter,
 	.llseek		= noop_llseek,
 };
 
@@ -265,20 +263,30 @@ static int do_signalfd4(int ufd, sigset_t *mask, int flags)
 	signotset(mask);
 
 	if (ufd == -1) {
+		struct file *file;
+
 		ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
 		if (!ctx)
 			return -ENOMEM;
 
 		ctx->sigmask = *mask;
 
-		/*
-		 * When we call this, the initialization must be complete, since
-		 * anon_inode_getfd() will install the fd.
-		 */
-		ufd = anon_inode_getfd("[signalfd]", &signalfd_fops, ctx,
-				       O_RDWR | (flags & (O_CLOEXEC | O_NONBLOCK)));
-		if (ufd < 0)
+		ufd = get_unused_fd_flags(flags & O_CLOEXEC);
+		if (ufd < 0) {
+			kfree(ctx);
+			return ufd;
+		}
+
+		file = anon_inode_getfile("[signalfd]", &signalfd_fops, ctx,
+				       O_RDWR | (flags & O_NONBLOCK));
+		if (IS_ERR(file)) {
+			put_unused_fd(ufd);
 			kfree(ctx);
+			return PTR_ERR(file);
+		}
+		file->f_mode |= FMODE_NOWAIT;
+
+		fd_install(ufd, file);
 	} else {
 		struct fd f = fdget(ufd);
 		if (!f.file)
diff --git a/fs/smb/client/Kconfig b/fs/smb/client/Kconfig
index 2927bd174a..2517dc2423 100644
--- a/fs/smb/client/Kconfig
+++ b/fs/smb/client/Kconfig
@@ -2,6 +2,7 @@
 config CIFS
 	tristate "SMB3 and CIFS support (advanced network filesystem)"
 	depends on INET
+	select NETFS_SUPPORT
 	select NLS
 	select NLS_UCS2_UTILS
 	select CRYPTO
diff --git a/fs/smb/client/Makefile b/fs/smb/client/Makefile
index 0b07eb94c9..e11985f246 100644
--- a/fs/smb/client/Makefile
+++ b/fs/smb/client/Makefile
@@ -12,7 +12,7 @@ cifs-y := trace.o cifsfs.o cifs_debug.o connect.o dir.o file.o \
 	  smb2ops.o smb2maperror.o smb2transport.o \
 	  smb2misc.o smb2pdu.o smb2inode.o smb2file.o cifsacl.o fs_context.o \
 	  dns_resolve.o cifs_spnego_negtokeninit.asn1.o asn1.o \
-	  namespace.o
+	  namespace.o reparse.o
 
 $(obj)/asn1.o: $(obj)/cifs_spnego_negtokeninit.asn1.h
 
diff --git a/fs/smb/client/cifs_debug.c b/fs/smb/client/cifs_debug.c
index aa95fa95ca..4a20e92474 100644
--- a/fs/smb/client/cifs_debug.c
+++ b/fs/smb/client/cifs_debug.c
@@ -280,6 +280,24 @@ static int cifs_debug_files_proc_show(struct seq_file *m, void *v)
 	return 0;
 }
 
+static __always_inline const char *compression_alg_str(__le16 alg)
+{
+	switch (alg) {
+	case SMB3_COMPRESS_NONE:
+		return "NONE";
+	case SMB3_COMPRESS_LZNT1:
+		return "LZNT1";
+	case SMB3_COMPRESS_LZ77:
+		return "LZ77";
+	case SMB3_COMPRESS_LZ77_HUFF:
+		return "LZ77-Huffman";
+	case SMB3_COMPRESS_PATTERN:
+		return "Pattern_V1";
+	default:
+		return "invalid";
+	}
+}
+
 static int cifs_debug_data_proc_show(struct seq_file *m, void *v)
 {
 	struct mid_q_entry *mid_entry;
@@ -425,12 +443,6 @@ skip_rdma:
 			server->echo_credits,
 			server->oplock_credits,
 			server->dialect);
-		if (server->compress_algorithm == SMB3_COMPRESS_LZNT1)
-			seq_printf(m, " COMPRESS_LZNT1");
-		else if (server->compress_algorithm == SMB3_COMPRESS_LZ77)
-			seq_printf(m, " COMPRESS_LZ77");
-		else if (server->compress_algorithm == SMB3_COMPRESS_LZ77_HUFF)
-			seq_printf(m, " COMPRESS_LZ77_HUFF");
 		if (server->sign)
 			seq_printf(m, " signed");
 		if (server->posix_ext_supported)
@@ -462,6 +474,14 @@ skip_rdma:
 				   server->leaf_fullpath);
 		}
 
+		seq_puts(m, "\nCompression: ");
+		if (!server->compression.requested)
+			seq_puts(m, "disabled on mount");
+		else if (server->compression.enabled)
+			seq_printf(m, "enabled (%s)", compression_alg_str(server->compression.alg));
+		else
+			seq_puts(m, "disabled (not supported by this server)");
+
 		seq_printf(m, "\n\n\tSessions: ");
 		i = 0;
 		list_for_each_entry(ses, &server->smb_ses_list, smb_ses_list) {
@@ -1052,7 +1072,7 @@ static int cifs_security_flags_proc_open(struct inode *inode, struct file *file)
 static void
 cifs_security_flags_handle_must_flags(unsigned int *flags)
 {
-	unsigned int signflags = *flags & CIFSSEC_MUST_SIGN;
+	unsigned int signflags = *flags & (CIFSSEC_MUST_SIGN | CIFSSEC_MUST_SEAL);
 
 	if ((*flags & CIFSSEC_MUST_KRB5) == CIFSSEC_MUST_KRB5)
 		*flags = CIFSSEC_MUST_KRB5;
diff --git a/fs/smb/client/cifsfs.c b/fs/smb/client/cifsfs.c
index 81fe608158..2c4b357d85 100644
--- a/fs/smb/client/cifsfs.c
+++ b/fs/smb/client/cifsfs.c
@@ -134,7 +134,7 @@ module_param(enable_oplocks, bool, 0644);
 MODULE_PARM_DESC(enable_oplocks, "Enable or disable oplocks. Default: y/Y/1");
 
 module_param(enable_gcm_256, bool, 0644);
-MODULE_PARM_DESC(enable_gcm_256, "Enable requesting strongest (256 bit) GCM encryption. Default: n/N/0");
+MODULE_PARM_DESC(enable_gcm_256, "Enable requesting strongest (256 bit) GCM encryption. Default: y/Y/0");
 
 module_param(require_gcm_256, bool, 0644);
 MODULE_PARM_DESC(require_gcm_256, "Require strongest (256 bit) GCM encryption. Default: n/N/0");
@@ -151,10 +151,6 @@ MODULE_PARM_DESC(disable_legacy_dialects, "To improve security it may be "
 				  "vers=1.0 (CIFS/SMB1) and vers=2.0 are weaker"
 				  " and less secure. Default: n/N/0");
 
-extern mempool_t *cifs_sm_req_poolp;
-extern mempool_t *cifs_req_poolp;
-extern mempool_t *cifs_mid_poolp;
-
 struct workqueue_struct	*cifsiod_wq;
 struct workqueue_struct	*decrypt_wq;
 struct workqueue_struct	*fileinfo_put_wq;
@@ -375,9 +371,13 @@ static struct kmem_cache *cifs_inode_cachep;
 static struct kmem_cache *cifs_req_cachep;
 static struct kmem_cache *cifs_mid_cachep;
 static struct kmem_cache *cifs_sm_req_cachep;
+static struct kmem_cache *cifs_io_request_cachep;
+static struct kmem_cache *cifs_io_subrequest_cachep;
 mempool_t *cifs_sm_req_poolp;
 mempool_t *cifs_req_poolp;
 mempool_t *cifs_mid_poolp;
+mempool_t cifs_io_request_pool;
+mempool_t cifs_io_subrequest_pool;
 
 static struct inode *
 cifs_alloc_inode(struct super_block *sb)
@@ -431,6 +431,7 @@ cifs_free_inode(struct inode *inode)
 static void
 cifs_evict_inode(struct inode *inode)
 {
+	netfs_wait_for_outstanding_io(inode);
 	truncate_inode_pages_final(&inode->i_data);
 	if (inode->i_state & I_PINNING_NETFS_WB)
 		cifs_fscache_unuse_inode_cookie(inode, true);
@@ -675,6 +676,8 @@ cifs_show_options(struct seq_file *s, struct dentry *root)
 		seq_printf(s, ",backupgid=%u",
 			   from_kgid_munged(&init_user_ns,
 					    cifs_sb->ctx->backupgid));
+	seq_show_option(s, "reparse",
+			cifs_reparse_type_str(cifs_sb->ctx->reparse_type));
 
 	seq_printf(s, ",rsize=%u", cifs_sb->ctx->rsize);
 	seq_printf(s, ",wsize=%u", cifs_sb->ctx->wsize);
@@ -742,6 +745,8 @@ static void cifs_umount_begin(struct super_block *sb)
 
 	spin_lock(&cifs_tcp_ses_lock);
 	spin_lock(&tcon->tc_lock);
+	trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+			    netfs_trace_tcon_ref_see_umount);
 	if ((tcon->tc_count > 1) || (tcon->status == TID_EXITING)) {
 		/* we have other mounts to same share or we have
 		   already tried to umount this and woken up
@@ -986,61 +991,6 @@ out:
 	return root;
 }
 
-
-static ssize_t
-cifs_loose_read_iter(struct kiocb *iocb, struct iov_iter *iter)
-{
-	ssize_t rc;
-	struct inode *inode = file_inode(iocb->ki_filp);
-
-	if (iocb->ki_flags & IOCB_DIRECT)
-		return cifs_user_readv(iocb, iter);
-
-	rc = cifs_revalidate_mapping(inode);
-	if (rc)
-		return rc;
-
-	return generic_file_read_iter(iocb, iter);
-}
-
-static ssize_t cifs_file_write_iter(struct kiocb *iocb, struct iov_iter *from)
-{
-	struct inode *inode = file_inode(iocb->ki_filp);
-	struct cifsInodeInfo *cinode = CIFS_I(inode);
-	ssize_t written;
-	int rc;
-
-	if (iocb->ki_filp->f_flags & O_DIRECT) {
-		written = cifs_user_writev(iocb, from);
-		if (written > 0 && CIFS_CACHE_READ(cinode)) {
-			cifs_zap_mapping(inode);
-			cifs_dbg(FYI,
-				 "Set no oplock for inode=%p after a write operation\n",
-				 inode);
-			cinode->oplock = 0;
-		}
-		return written;
-	}
-
-	written = cifs_get_writer(cinode);
-	if (written)
-		return written;
-
-	written = generic_file_write_iter(iocb, from);
-
-	if (CIFS_CACHE_WRITE(CIFS_I(inode)))
-		goto out;
-
-	rc = filemap_fdatawrite(inode->i_mapping);
-	if (rc)
-		cifs_dbg(FYI, "cifs_file_write_iter: %d rc on %p inode\n",
-			 rc, inode);
-
-out:
-	cifs_put_writer(cinode);
-	return written;
-}
-
 static loff_t cifs_llseek(struct file *file, loff_t offset, int whence)
 {
 	struct cifsFileInfo *cfile = file->private_data;
@@ -1087,7 +1037,7 @@ static loff_t cifs_llseek(struct file *file, loff_t offset, int whence)
 }
 
 static int
-cifs_setlease(struct file *file, int arg, struct file_lock **lease, void **priv)
+cifs_setlease(struct file *file, int arg, struct file_lease **lease, void **priv)
 {
 	/*
 	 * Note that this is called by vfs setlease with i_lock held to
@@ -1096,9 +1046,6 @@ cifs_setlease(struct file *file, int arg, struct file_lock **lease, void **priv)
 	struct inode *inode = file_inode(file);
 	struct cifsFileInfo *cfile = file->private_data;
 
-	if (!(S_ISREG(inode->i_mode)))
-		return -EINVAL;
-
 	/* Check if file is oplocked if this is request for new lease */
 	if (arg == F_UNLCK ||
 	    ((arg == F_RDLCK) && CIFS_CACHE_READ(CIFS_I(inode))) ||
@@ -1280,7 +1227,7 @@ static loff_t cifs_remap_file_range(struct file *src_file, loff_t off,
 	struct cifsFileInfo *smb_file_src = src_file->private_data;
 	struct cifsFileInfo *smb_file_target = dst_file->private_data;
 	struct cifs_tcon *target_tcon, *src_tcon;
-	unsigned long long destend, fstart, fend, new_size;
+	unsigned long long destend, fstart, fend, old_size, new_size;
 	unsigned int xid;
 	int rc;
 
@@ -1345,6 +1292,9 @@ static loff_t cifs_remap_file_range(struct file *src_file, loff_t off,
 	rc = cifs_flush_folio(target_inode, destend, &fstart, &fend, false);
 	if (rc)
 		goto unlock;
+	if (fend > target_cifsi->netfs.zero_point)
+		target_cifsi->netfs.zero_point = fend + 1;
+	old_size = target_cifsi->netfs.remote_i_size;
 
 	/* Discard all the folios that overlap the destination region. */
 	cifs_dbg(FYI, "about to discard pages %llx-%llx\n", fstart, fend);
@@ -1357,12 +1307,13 @@ static loff_t cifs_remap_file_range(struct file *src_file, loff_t off,
 	if (target_tcon->ses->server->ops->duplicate_extents) {
 		rc = target_tcon->ses->server->ops->duplicate_extents(xid,
 			smb_file_src, smb_file_target, off, len, destoff);
-		if (rc == 0 && new_size > i_size_read(target_inode)) {
+		if (rc == 0 && new_size > old_size) {
 			truncate_setsize(target_inode, new_size);
-			netfs_resize_file(&target_cifsi->netfs, new_size, true);
 			fscache_resize_cookie(cifs_inode_cookie(target_inode),
 					      new_size);
 		}
+		if (rc == 0 && new_size > target_cifsi->netfs.zero_point)
+			target_cifsi->netfs.zero_point = new_size;
 	}
 
 	/* force revalidate of size and timestamps of target file now
@@ -1408,7 +1359,7 @@ ssize_t cifs_file_copychunk_range(unsigned int xid,
 	target_tcon = tlink_tcon(smb_file_target->tlink);
 
 	if (src_tcon->ses != target_tcon->ses) {
-		cifs_dbg(VFS, "source and target of copy not on same server\n");
+		cifs_dbg(FYI, "source and target of copy not on same server\n");
 		goto out;
 	}
 
@@ -1454,6 +1405,8 @@ ssize_t cifs_file_copychunk_range(unsigned int xid,
 	rc = cifs_flush_folio(target_inode, destend, &fstart, &fend, false);
 	if (rc)
 		goto unlock;
+	if (fend > target_cifsi->netfs.zero_point)
+		target_cifsi->netfs.zero_point = fend + 1;
 
 	/* Discard all the folios that overlap the destination region. */
 	truncate_inode_pages_range(&target_inode->i_data, fstart, fend);
@@ -1570,8 +1523,8 @@ const struct file_operations cifs_file_strict_ops = {
 };
 
 const struct file_operations cifs_file_direct_ops = {
-	.read_iter = cifs_direct_readv,
-	.write_iter = cifs_direct_writev,
+	.read_iter = netfs_unbuffered_read_iter,
+	.write_iter = netfs_file_write_iter,
 	.open = cifs_open,
 	.release = cifs_close,
 	.lock = cifs_lock,
@@ -1626,8 +1579,8 @@ const struct file_operations cifs_file_strict_nobrl_ops = {
 };
 
 const struct file_operations cifs_file_direct_nobrl_ops = {
-	.read_iter = cifs_direct_readv,
-	.write_iter = cifs_direct_writev,
+	.read_iter = netfs_unbuffered_read_iter,
+	.write_iter = netfs_file_write_iter,
 	.open = cifs_open,
 	.release = cifs_close,
 	.fsync = cifs_fsync,
@@ -1669,7 +1622,7 @@ cifs_init_inodecache(void)
 	cifs_inode_cachep = kmem_cache_create("cifs_inode_cache",
 					      sizeof(struct cifsInodeInfo),
 					      0, (SLAB_RECLAIM_ACCOUNT|
-						SLAB_MEM_SPREAD|SLAB_ACCOUNT),
+						SLAB_ACCOUNT),
 					      cifs_init_once);
 	if (cifs_inode_cachep == NULL)
 		return -ENOMEM;
@@ -1802,6 +1755,48 @@ static void destroy_mids(void)
 	kmem_cache_destroy(cifs_mid_cachep);
 }
 
+static int cifs_init_netfs(void)
+{
+	cifs_io_request_cachep =
+		kmem_cache_create("cifs_io_request",
+				  sizeof(struct cifs_io_request), 0,
+				  SLAB_HWCACHE_ALIGN, NULL);
+	if (!cifs_io_request_cachep)
+		goto nomem_req;
+
+	if (mempool_init_slab_pool(&cifs_io_request_pool, 100, cifs_io_request_cachep) < 0)
+		goto nomem_reqpool;
+
+	cifs_io_subrequest_cachep =
+		kmem_cache_create("cifs_io_subrequest",
+				  sizeof(struct cifs_io_subrequest), 0,
+				  SLAB_HWCACHE_ALIGN, NULL);
+	if (!cifs_io_subrequest_cachep)
+		goto nomem_subreq;
+
+	if (mempool_init_slab_pool(&cifs_io_subrequest_pool, 100, cifs_io_subrequest_cachep) < 0)
+		goto nomem_subreqpool;
+
+	return 0;
+
+nomem_subreqpool:
+	kmem_cache_destroy(cifs_io_subrequest_cachep);
+nomem_subreq:
+	mempool_destroy(&cifs_io_request_pool);
+nomem_reqpool:
+	kmem_cache_destroy(cifs_io_request_cachep);
+nomem_req:
+	return -ENOMEM;
+}
+
+static void cifs_destroy_netfs(void)
+{
+	mempool_exit(&cifs_io_subrequest_pool);
+	kmem_cache_destroy(cifs_io_subrequest_cachep);
+	mempool_exit(&cifs_io_request_pool);
+	kmem_cache_destroy(cifs_io_request_cachep);
+}
+
 static int __init
 init_cifs(void)
 {
@@ -1899,17 +1894,21 @@ init_cifs(void)
 					   WQ_FREEZABLE|WQ_MEM_RECLAIM, 0);
 	if (!serverclose_wq) {
 		rc = -ENOMEM;
-		goto out_destroy_serverclose_wq;
+		goto out_destroy_deferredclose_wq;
 	}
 
 	rc = cifs_init_inodecache();
 	if (rc)
-		goto out_destroy_deferredclose_wq;
+		goto out_destroy_serverclose_wq;
 
-	rc = init_mids();
+	rc = cifs_init_netfs();
 	if (rc)
 		goto out_destroy_inodecache;
 
+	rc = init_mids();
+	if (rc)
+		goto out_destroy_netfs;
+
 	rc = cifs_init_request_bufs();
 	if (rc)
 		goto out_destroy_mids;
@@ -1964,8 +1963,12 @@ out_destroy_request_bufs:
 	cifs_destroy_request_bufs();
 out_destroy_mids:
 	destroy_mids();
+out_destroy_netfs:
+	cifs_destroy_netfs();
 out_destroy_inodecache:
 	cifs_destroy_inodecache();
+out_destroy_serverclose_wq:
+	destroy_workqueue(serverclose_wq);
 out_destroy_deferredclose_wq:
 	destroy_workqueue(deferredclose_wq);
 out_destroy_cifsoplockd_wq:
@@ -1976,8 +1979,6 @@ out_destroy_decrypt_wq:
 	destroy_workqueue(decrypt_wq);
 out_destroy_cifsiod_wq:
 	destroy_workqueue(cifsiod_wq);
-out_destroy_serverclose_wq:
-	destroy_workqueue(serverclose_wq);
 out_clean_proc:
 	cifs_proc_clean();
 	return rc;
@@ -2002,6 +2003,7 @@ exit_cifs(void)
 #endif
 	cifs_destroy_request_bufs();
 	destroy_mids();
+	cifs_destroy_netfs();
 	cifs_destroy_inodecache();
 	destroy_workqueue(deferredclose_wq);
 	destroy_workqueue(cifsoplockd_wq);
diff --git a/fs/smb/client/cifsfs.h b/fs/smb/client/cifsfs.h
index 685f7d1139..62d5fee3e5 100644
--- a/fs/smb/client/cifsfs.h
+++ b/fs/smb/client/cifsfs.h
@@ -69,7 +69,6 @@ extern int cifs_revalidate_file_attr(struct file *filp);
 extern int cifs_revalidate_dentry_attr(struct dentry *);
 extern int cifs_revalidate_file(struct file *filp);
 extern int cifs_revalidate_dentry(struct dentry *);
-extern int cifs_invalidate_mapping(struct inode *inode);
 extern int cifs_revalidate_mapping(struct inode *inode);
 extern int cifs_zap_mapping(struct inode *inode);
 extern int cifs_getattr(struct mnt_idmap *, const struct path *,
@@ -85,6 +84,7 @@ extern const struct inode_operations cifs_namespace_inode_operations;
 
 
 /* Functions related to files and directories */
+extern const struct netfs_request_ops cifs_req_ops;
 extern const struct file_operations cifs_file_ops;
 extern const struct file_operations cifs_file_direct_ops; /* if directio mnt */
 extern const struct file_operations cifs_file_strict_ops; /* if strictio mnt */
@@ -94,12 +94,10 @@ extern const struct file_operations cifs_file_strict_nobrl_ops;
 extern int cifs_open(struct inode *inode, struct file *file);
 extern int cifs_close(struct inode *inode, struct file *file);
 extern int cifs_closedir(struct inode *inode, struct file *file);
-extern ssize_t cifs_user_readv(struct kiocb *iocb, struct iov_iter *to);
-extern ssize_t cifs_direct_readv(struct kiocb *iocb, struct iov_iter *to);
 extern ssize_t cifs_strict_readv(struct kiocb *iocb, struct iov_iter *to);
-extern ssize_t cifs_user_writev(struct kiocb *iocb, struct iov_iter *from);
-extern ssize_t cifs_direct_writev(struct kiocb *iocb, struct iov_iter *from);
 extern ssize_t cifs_strict_writev(struct kiocb *iocb, struct iov_iter *from);
+ssize_t cifs_file_write_iter(struct kiocb *iocb, struct iov_iter *from);
+ssize_t cifs_loose_read_iter(struct kiocb *iocb, struct iov_iter *iter);
 extern int cifs_flock(struct file *pfile, int cmd, struct file_lock *plock);
 extern int cifs_lock(struct file *, int, struct file_lock *);
 extern int cifs_fsync(struct file *, loff_t, loff_t, int);
@@ -110,9 +108,6 @@ extern int cifs_file_strict_mmap(struct file *file, struct vm_area_struct *vma);
 extern const struct file_operations cifs_dir_ops;
 extern int cifs_dir_open(struct inode *inode, struct file *file);
 extern int cifs_readdir(struct file *file, struct dir_context *ctx);
-extern void cifs_pages_written_back(struct inode *inode, loff_t start, unsigned int len);
-extern void cifs_pages_write_failed(struct inode *inode, loff_t start, unsigned int len);
-extern void cifs_pages_write_redirty(struct inode *inode, loff_t start, unsigned int len);
 
 /* Functions related to dir entries */
 extern const struct dentry_operations cifs_dentry_ops;
@@ -152,6 +147,6 @@ extern const struct export_operations cifs_export_ops;
 #endif /* CONFIG_CIFS_NFSD_EXPORT */
 
 /* when changing internal version - update following two lines at same time */
-#define SMB3_PRODUCT_BUILD 47
-#define CIFS_VERSION   "2.47"
+#define SMB3_PRODUCT_BUILD 49
+#define CIFS_VERSION   "2.49"
 #endif				/* _CIFSFS_H */
diff --git a/fs/smb/client/cifsglob.h b/fs/smb/client/cifsglob.h
index a393d505e8..d4bcc7da70 100644
--- a/fs/smb/client/cifsglob.h
+++ b/fs/smb/client/cifsglob.h
@@ -153,6 +153,24 @@ enum securityEnum {
 	Kerberos,		/* Kerberos via SPNEGO */
 };
 
+enum cifs_reparse_type {
+	CIFS_REPARSE_TYPE_NFS,
+	CIFS_REPARSE_TYPE_WSL,
+	CIFS_REPARSE_TYPE_DEFAULT = CIFS_REPARSE_TYPE_NFS,
+};
+
+static inline const char *cifs_reparse_type_str(enum cifs_reparse_type type)
+{
+	switch (type) {
+	case CIFS_REPARSE_TYPE_NFS:
+		return "nfs";
+	case CIFS_REPARSE_TYPE_WSL:
+		return "wsl";
+	default:
+		return "unknown";
+	}
+}
+
 struct session_key {
 	unsigned int len;
 	char *response;
@@ -208,6 +226,10 @@ struct cifs_open_info_data {
 			struct reparse_posix_data *posix;
 		};
 	} reparse;
+	struct {
+		__u8		eas[SMB2_WSL_MAX_QUERY_EA_RESP_SIZE];
+		unsigned int	eas_len;
+	} wsl;
 	char *symlink_target;
 	struct cifs_sid posix_owner;
 	struct cifs_sid posix_group;
@@ -217,19 +239,6 @@ struct cifs_open_info_data {
 	};
 };
 
-static inline bool cifs_open_data_reparse(struct cifs_open_info_data *data)
-{
-	struct smb2_file_all_info *fi = &data->fi;
-	u32 attrs = le32_to_cpu(fi->Attributes);
-	bool ret;
-
-	ret = data->reparse_point || (attrs & ATTR_REPARSE);
-	if (ret)
-		attrs |= ATTR_REPARSE;
-	fi->Attributes = cpu_to_le32(attrs);
-	return ret;
-}
-
 /*
  *****************************************************************
  * Except the CIFS PDUs themselves all the
@@ -259,8 +268,7 @@ struct dfs_info3_param;
 struct cifs_fattr;
 struct smb3_fs_context;
 struct cifs_fid;
-struct cifs_readdata;
-struct cifs_writedata;
+struct cifs_io_subrequest;
 struct cifs_io_parms;
 struct cifs_search_info;
 struct cifsInodeInfo;
@@ -374,7 +382,8 @@ struct smb_version_operations {
 			    struct cifs_open_info_data *data);
 	/* set size by path */
 	int (*set_path_size)(const unsigned int, struct cifs_tcon *,
-			     const char *, __u64, struct cifs_sb_info *, bool);
+			     const char *, __u64, struct cifs_sb_info *, bool,
+				 struct dentry *);
 	/* set size by file handle */
 	int (*set_file_size)(const unsigned int, struct cifs_tcon *,
 			     struct cifsFileInfo *, __u64, bool);
@@ -404,7 +413,7 @@ struct smb_version_operations {
 		     struct cifs_sb_info *);
 	/* unlink file */
 	int (*unlink)(const unsigned int, struct cifs_tcon *, const char *,
-		      struct cifs_sb_info *);
+		      struct cifs_sb_info *, struct dentry *);
 	/* open, rename and delete file */
 	int (*rename_pending_delete)(const char *, struct dentry *,
 				     const unsigned int);
@@ -440,10 +449,9 @@ struct smb_version_operations {
 	/* send a flush request to the server */
 	int (*flush)(const unsigned int, struct cifs_tcon *, struct cifs_fid *);
 	/* async read from the server */
-	int (*async_readv)(struct cifs_readdata *);
+	int (*async_readv)(struct cifs_io_subrequest *);
 	/* async write to the server */
-	int (*async_writev)(struct cifs_writedata *,
-			    void (*release)(struct kref *));
+	void (*async_writev)(struct cifs_io_subrequest *);
 	/* sync read from the server */
 	int (*sync_read)(const unsigned int, struct cifs_fid *,
 			 struct cifs_io_parms *, unsigned int *, char **,
@@ -538,8 +546,8 @@ struct smb_version_operations {
 	/* writepages retry size */
 	unsigned int (*wp_retry_size)(struct inode *);
 	/* get mtu credits */
-	int (*wait_mtu_credits)(struct TCP_Server_Info *, unsigned int,
-				unsigned int *, struct cifs_credits *);
+	int (*wait_mtu_credits)(struct TCP_Server_Info *, size_t,
+				size_t *, struct cifs_credits *);
 	/* adjust previously taken mtu credits to request size */
 	int (*adjust_credits)(struct TCP_Server_Info *server,
 			      struct cifs_credits *credits,
@@ -762,7 +770,11 @@ struct TCP_Server_Info {
 	unsigned int	max_write;
 	unsigned int	min_offload;
 	unsigned int	retrans;
-	__le16	compress_algorithm;
+	struct {
+		bool requested; /* "compress" mount option set*/
+		bool enabled; /* actually negotiated with server */
+		__le16 alg; /* preferred alg negotiated with server */
+	} compression;
 	__u16	signing_algorithm;
 	__le16	cipher_type;
 	 /* save initital negprot hash */
@@ -869,11 +881,12 @@ add_credits(struct TCP_Server_Info *server, const struct cifs_credits *credits,
 
 static inline void
 add_credits_and_wake_if(struct TCP_Server_Info *server,
-			const struct cifs_credits *credits, const int optype)
+			struct cifs_credits *credits, const int optype)
 {
 	if (credits->value) {
 		server->ops->add_credits(server, credits, optype);
 		wake_up(&server->request_q);
+		credits->value = 0;
 	}
 }
 
@@ -1176,6 +1189,7 @@ struct cifs_fattr {
  */
 struct cifs_tcon {
 	struct list_head tcon_list;
+	int debug_id;		/* Debugging for tracing */
 	int tc_count;
 	struct list_head rlist; /* reconnect list */
 	spinlock_t tc_lock;  /* protect anything here that is not protected */
@@ -1385,6 +1399,7 @@ struct cifs_open_parms {
 	umode_t mode;
 	bool reconnect:1;
 	bool replay:1; /* indicates that this open is for a replay */
+	struct kvec *ea_cctx;
 };
 
 struct cifs_fid {
@@ -1426,6 +1441,7 @@ struct cifsFileInfo {
 	bool invalidHandle:1;	/* file closed via session abend */
 	bool swapfile:1;
 	bool oplock_break_cancelled:1;
+	bool status_file_deleted:1; /* file has been deleted */
 	bool offload:1; /* offload final part of _put to a wq */
 	unsigned int oplock_epoch; /* epoch from the lease break */
 	__u32 oplock_level; /* oplock/lease level from the lease break */
@@ -1475,50 +1491,31 @@ struct cifs_aio_ctx {
 	bool			direct_io;
 };
 
-/* asynchronous read support */
-struct cifs_readdata {
-	struct kref			refcount;
-	struct list_head		list;
-	struct completion		done;
+struct cifs_io_request {
+	struct netfs_io_request		rreq;
 	struct cifsFileInfo		*cfile;
-	struct address_space		*mapping;
-	struct cifs_aio_ctx		*ctx;
-	__u64				offset;
-	ssize_t				got_bytes;
-	unsigned int			bytes;
-	pid_t				pid;
-	int				result;
-	struct work_struct		work;
-	struct iov_iter			iter;
-	struct kvec			iov[2];
 	struct TCP_Server_Info		*server;
-#ifdef CONFIG_CIFS_SMB_DIRECT
-	struct smbd_mr			*mr;
-#endif
-	struct cifs_credits		credits;
+	pid_t				pid;
 };
 
-/* asynchronous write support */
-struct cifs_writedata {
-	struct kref			refcount;
-	struct list_head		list;
-	struct completion		done;
-	enum writeback_sync_modes	sync_mode;
-	struct work_struct		work;
-	struct cifsFileInfo		*cfile;
-	struct cifs_aio_ctx		*ctx;
-	struct iov_iter			iter;
-	struct bio_vec			*bv;
-	__u64				offset;
-	pid_t				pid;
-	unsigned int			bytes;
+/* asynchronous read support */
+struct cifs_io_subrequest {
+	union {
+		struct netfs_io_subrequest subreq;
+		struct netfs_io_request *rreq;
+		struct cifs_io_request *req;
+	};
+	ssize_t				got_bytes;
+	unsigned int			xid;
 	int				result;
+	bool				have_xid;
+	bool				replay;
+	struct kvec			iov[2];
 	struct TCP_Server_Info		*server;
 #ifdef CONFIG_CIFS_SMB_DIRECT
 	struct smbd_mr			*mr;
 #endif
 	struct cifs_credits		credits;
-	bool				replay;
 };
 
 /*
@@ -1904,7 +1901,7 @@ static inline bool is_replayable_error(int error)
 #define   CIFSSEC_MAY_SIGN	0x00001
 #define   CIFSSEC_MAY_NTLMV2	0x00004
 #define   CIFSSEC_MAY_KRB5	0x00008
-#define   CIFSSEC_MAY_SEAL	0x00040 /* not supported yet */
+#define   CIFSSEC_MAY_SEAL	0x00040
 #define   CIFSSEC_MAY_NTLMSSP	0x00080 /* raw ntlmssp with ntlmv2 */
 
 #define   CIFSSEC_MUST_SIGN	0x01001
@@ -1914,15 +1911,15 @@ require use of the stronger protocol */
 #define   CIFSSEC_MUST_NTLMV2	0x04004
 #define   CIFSSEC_MUST_KRB5	0x08008
 #ifdef CONFIG_CIFS_UPCALL
-#define   CIFSSEC_MASK          0x8F08F /* flags supported if no weak allowed */
+#define   CIFSSEC_MASK          0xCF0CF /* flags supported if no weak allowed */
 #else
-#define	  CIFSSEC_MASK          0x87087 /* flags supported if no weak allowed */
+#define	  CIFSSEC_MASK          0xC70C7 /* flags supported if no weak allowed */
 #endif /* UPCALL */
-#define   CIFSSEC_MUST_SEAL	0x40040 /* not supported yet */
+#define   CIFSSEC_MUST_SEAL	0x40040
 #define   CIFSSEC_MUST_NTLMSSP	0x80080 /* raw ntlmssp with ntlmv2 */
 
-#define   CIFSSEC_DEF (CIFSSEC_MAY_SIGN | CIFSSEC_MAY_NTLMV2 | CIFSSEC_MAY_NTLMSSP)
-#define   CIFSSEC_MAX (CIFSSEC_MUST_NTLMV2)
+#define   CIFSSEC_DEF (CIFSSEC_MAY_SIGN | CIFSSEC_MAY_NTLMV2 | CIFSSEC_MAY_NTLMSSP | CIFSSEC_MAY_SEAL)
+#define   CIFSSEC_MAX (CIFSSEC_MAY_SIGN | CIFSSEC_MUST_KRB5 | CIFSSEC_MAY_SEAL)
 #define   CIFSSEC_AUTH_MASK (CIFSSEC_MAY_NTLMV2 | CIFSSEC_MAY_KRB5 | CIFSSEC_MAY_NTLMSSP)
 /*
  *****************************************************************
@@ -1999,6 +1996,7 @@ require use of the stronger protocol */
  *				->chans_need_reconnect
  *				->chans_in_reconnect
  * cifs_tcon->tc_lock		(anything that is not protected by another lock and can change)
+ * inode->i_rwsem, taken by fs/netfs/locking.c e.g. should be taken before cifsInodeInfo locks
  * cifsInodeInfo->open_file_lock	cifsInodeInfo->openFileList	cifs_alloc_inode
  * cifsInodeInfo->writers_lock	cifsInodeInfo->writers		cifsInodeInfo_alloc
  * cifsInodeInfo->lock_sem	cifsInodeInfo->llist		cifs_init_once
@@ -2095,7 +2093,11 @@ extern struct workqueue_struct *deferredclose_wq;
 extern struct workqueue_struct *serverclose_wq;
 extern __u32 cifs_lock_secret;
 
+extern mempool_t *cifs_sm_req_poolp;
+extern mempool_t *cifs_req_poolp;
 extern mempool_t *cifs_mid_poolp;
+extern mempool_t cifs_io_request_pool;
+extern mempool_t cifs_io_subrequest_pool;
 
 /* Operations for different SMB versions */
 #define SMB1_VERSION_STRING	"1.0"
@@ -2285,6 +2287,17 @@ static inline void cifs_sg_set_buf(struct sg_table *sgtable,
 	}
 }
 
+#define CIFS_OPARMS(_cifs_sb, _tcon, _path, _da, _cd, _co, _mode) \
+	((struct cifs_open_parms) { \
+		.tcon = _tcon, \
+		.path = _path, \
+		.desired_access = (_da), \
+		.disposition = (_cd), \
+		.create_options = cifs_create_options(_cifs_sb, (_co)), \
+		.mode = (_mode), \
+		.cifs_sb = _cifs_sb, \
+	})
+
 struct smb2_compound_vars {
 	struct cifs_open_parms oparms;
 	struct kvec rsp_iov[MAX_COMPOUND];
@@ -2296,6 +2309,7 @@ struct smb2_compound_vars {
 	struct kvec close_iov;
 	struct smb2_file_rename_info rename_info;
 	struct smb2_file_link_info link_info;
+	struct kvec ea_iov;
 };
 
 static inline bool cifs_ses_exiting(struct cifs_ses *ses)
diff --git a/fs/smb/client/cifspdu.h b/fs/smb/client/cifspdu.h
index c46d418c1c..a2072ab9e5 100644
--- a/fs/smb/client/cifspdu.h
+++ b/fs/smb/client/cifspdu.h
@@ -2574,7 +2574,7 @@ typedef struct {
 
 
 struct win_dev {
-	unsigned char type[8]; /* IntxCHR or IntxBLK or LnxFIFO*/
+	unsigned char type[8]; /* IntxCHR or IntxBLK or LnxFIFO or LnxSOCK */
 	__le64 major;
 	__le64 minor;
 } __attribute__((packed));
diff --git a/fs/smb/client/cifsproto.h b/fs/smb/client/cifsproto.h
index 996ca413dd..c15bb5ee7e 100644
--- a/fs/smb/client/cifsproto.h
+++ b/fs/smb/client/cifsproto.h
@@ -121,7 +121,7 @@ extern struct mid_q_entry *cifs_setup_async_request(struct TCP_Server_Info *,
 extern int cifs_check_receive(struct mid_q_entry *mid,
 			struct TCP_Server_Info *server, bool log_error);
 extern int cifs_wait_mtu_credits(struct TCP_Server_Info *server,
-				 unsigned int size, unsigned int *num,
+				 size_t size, size_t *num,
 				 struct cifs_credits *credits);
 extern int SendReceive2(const unsigned int /* xid */ , struct cifs_ses *,
 			struct kvec *, int /* nvec to send */,
@@ -148,6 +148,8 @@ extern bool is_size_safe_to_change(struct cifsInodeInfo *cifsInode, __u64 eof,
 				   bool from_readdir);
 extern void cifs_update_eof(struct cifsInodeInfo *cifsi, loff_t offset,
 			    unsigned int bytes_written);
+void cifs_write_subrequest_terminated(struct cifs_io_subrequest *wdata, ssize_t result,
+				      bool was_async);
 extern struct cifsFileInfo *find_writable_file(struct cifsInodeInfo *, int);
 extern int cifs_get_writable_file(struct cifsInodeInfo *cifs_inode,
 				  int flags,
@@ -210,10 +212,6 @@ extern struct inode *cifs_iget(struct super_block *sb,
 int cifs_get_inode_info(struct inode **inode, const char *full_path,
 			struct cifs_open_info_data *data, struct super_block *sb, int xid,
 			const struct cifs_fid *fid);
-bool cifs_reparse_point_to_fattr(struct cifs_sb_info *cifs_sb,
-				 struct cifs_fattr *fattr,
-				 struct cifs_open_info_data *data);
-
 extern int smb311_posix_get_inode_info(struct inode **inode,
 				       const char *full_path,
 				       struct cifs_open_info_data *data,
@@ -298,12 +296,16 @@ extern void cifs_close_all_deferred_files(struct cifs_tcon *cifs_tcon);
 
 extern void cifs_close_deferred_file_under_dentry(struct cifs_tcon *cifs_tcon,
 				const char *path);
+
+extern void cifs_mark_open_handles_for_deleted_file(struct inode *inode,
+				const char *path);
+
 extern struct TCP_Server_Info *
 cifs_get_tcp_session(struct smb3_fs_context *ctx,
 		     struct TCP_Server_Info *primary_server);
 extern void cifs_put_tcp_session(struct TCP_Server_Info *server,
 				 int from_reconnect);
-extern void cifs_put_tcon(struct cifs_tcon *tcon);
+extern void cifs_put_tcon(struct cifs_tcon *tcon, enum smb3_tcon_ref_trace trace);
 
 extern void cifs_release_automount_timer(void);
 
@@ -404,7 +406,8 @@ extern int CIFSSMBSetFileDisposition(const unsigned int xid,
 				     __u32 pid_of_opener);
 extern int CIFSSMBSetEOF(const unsigned int xid, struct cifs_tcon *tcon,
 			 const char *file_name, __u64 size,
-			 struct cifs_sb_info *cifs_sb, bool set_allocation);
+			 struct cifs_sb_info *cifs_sb, bool set_allocation,
+			 struct dentry *dentry);
 extern int CIFSSMBSetFileSize(const unsigned int xid, struct cifs_tcon *tcon,
 			      struct cifsFileInfo *cfile, __u64 size,
 			      bool set_allocation);
@@ -440,7 +443,8 @@ extern int CIFSPOSIXDelFile(const unsigned int xid, struct cifs_tcon *tcon,
 			const struct nls_table *nls_codepage,
 			int remap_special_chars);
 extern int CIFSSMBDelFile(const unsigned int xid, struct cifs_tcon *tcon,
-			  const char *name, struct cifs_sb_info *cifs_sb);
+			  const char *name, struct cifs_sb_info *cifs_sb,
+			  struct dentry *dentry);
 int CIFSSMBRename(const unsigned int xid, struct cifs_tcon *tcon,
 		  struct dentry *source_dentry,
 		  const char *from_name, const char *to_name,
@@ -528,8 +532,9 @@ extern int CIFSSMBLogoff(const unsigned int xid, struct cifs_ses *ses);
 
 extern struct cifs_ses *sesInfoAlloc(void);
 extern void sesInfoFree(struct cifs_ses *);
-extern struct cifs_tcon *tcon_info_alloc(bool dir_leases_enabled);
-extern void tconInfoFree(struct cifs_tcon *);
+extern struct cifs_tcon *tcon_info_alloc(bool dir_leases_enabled,
+					 enum smb3_tcon_ref_trace trace);
+extern void tconInfoFree(struct cifs_tcon *tcon, enum smb3_tcon_ref_trace trace);
 
 extern int cifs_sign_rqst(struct smb_rqst *rqst, struct TCP_Server_Info *server,
 		   __u32 *pexpected_response_sequence_number);
@@ -596,15 +601,11 @@ void __cifs_put_smb_ses(struct cifs_ses *ses);
 extern struct cifs_ses *
 cifs_get_smb_ses(struct TCP_Server_Info *server, struct smb3_fs_context *ctx);
 
-void cifs_readdata_release(struct kref *refcount);
-int cifs_async_readv(struct cifs_readdata *rdata);
+int cifs_async_readv(struct cifs_io_subrequest *rdata);
 int cifs_readv_receive(struct TCP_Server_Info *server, struct mid_q_entry *mid);
 
-int cifs_async_writev(struct cifs_writedata *wdata,
-		      void (*release)(struct kref *kref));
+void cifs_async_writev(struct cifs_io_subrequest *wdata);
 void cifs_writev_complete(struct work_struct *work);
-struct cifs_writedata *cifs_writedata_alloc(work_func_t complete);
-void cifs_writedata_release(struct kref *refcount);
 int cifs_query_mf_symlink(unsigned int xid, struct cifs_tcon *tcon,
 			  struct cifs_sb_info *cifs_sb,
 			  const unsigned char *path, char *pbuf,
@@ -719,8 +720,6 @@ static inline int cifs_create_options(struct cifs_sb_info *cifs_sb, int options)
 		return options;
 }
 
-struct super_block *cifs_get_tcon_super(struct cifs_tcon *tcon);
-void cifs_put_tcon_super(struct super_block *sb);
 int cifs_wait_for_server_reconnect(struct TCP_Server_Info *server, bool retry);
 
 /* Put references of @ses and its children */
diff --git a/fs/smb/client/cifssmb.c b/fs/smb/client/cifssmb.c
index 01e89070df..595c4b6737 100644
--- a/fs/smb/client/cifssmb.c
+++ b/fs/smb/client/cifssmb.c
@@ -24,6 +24,8 @@
 #include <linux/swap.h>
 #include <linux/task_io_accounting_ops.h>
 #include <linux/uaccess.h>
+#include <linux/netfs.h>
+#include <trace/events/netfs.h>
 #include "cifspdu.h"
 #include "cifsfs.h"
 #include "cifsglob.h"
@@ -738,7 +740,7 @@ PsxDelete:
 
 int
 CIFSSMBDelFile(const unsigned int xid, struct cifs_tcon *tcon, const char *name,
-	       struct cifs_sb_info *cifs_sb)
+	       struct cifs_sb_info *cifs_sb, struct dentry *dentry)
 {
 	DELETE_FILE_REQ *pSMB = NULL;
 	DELETE_FILE_RSP *pSMBr = NULL;
@@ -1262,18 +1264,17 @@ openRetry:
 static void
 cifs_readv_callback(struct mid_q_entry *mid)
 {
-	struct cifs_readdata *rdata = mid->callback_data;
-	struct cifs_tcon *tcon = tlink_tcon(rdata->cfile->tlink);
+	struct cifs_io_subrequest *rdata = mid->callback_data;
+	struct cifs_tcon *tcon = tlink_tcon(rdata->req->cfile->tlink);
 	struct TCP_Server_Info *server = tcon->ses->server;
 	struct smb_rqst rqst = { .rq_iov = rdata->iov,
 				 .rq_nvec = 2,
-				 .rq_iter_size = iov_iter_count(&rdata->iter),
-				 .rq_iter = rdata->iter };
+				 .rq_iter = rdata->subreq.io_iter };
 	struct cifs_credits credits = { .value = 1, .instance = 0 };
 
-	cifs_dbg(FYI, "%s: mid=%llu state=%d result=%d bytes=%u\n",
+	cifs_dbg(FYI, "%s: mid=%llu state=%d result=%d bytes=%zu\n",
 		 __func__, mid->mid, mid->mid_state, rdata->result,
-		 rdata->bytes);
+		 rdata->subreq.len);
 
 	switch (mid->mid_state) {
 	case MID_RESPONSE_RECEIVED:
@@ -1305,30 +1306,36 @@ cifs_readv_callback(struct mid_q_entry *mid)
 		rdata->result = -EIO;
 	}
 
-	queue_work(cifsiod_wq, &rdata->work);
+	if (rdata->result == 0 || rdata->result == -EAGAIN)
+		iov_iter_advance(&rdata->subreq.io_iter, rdata->got_bytes);
+	rdata->credits.value = 0;
+	netfs_subreq_terminated(&rdata->subreq,
+				(rdata->result == 0 || rdata->result == -EAGAIN) ?
+				rdata->got_bytes : rdata->result,
+				false);
 	release_mid(mid);
 	add_credits(server, &credits, 0);
 }
 
 /* cifs_async_readv - send an async write, and set up mid to handle result */
 int
-cifs_async_readv(struct cifs_readdata *rdata)
+cifs_async_readv(struct cifs_io_subrequest *rdata)
 {
 	int rc;
 	READ_REQ *smb = NULL;
 	int wct;
-	struct cifs_tcon *tcon = tlink_tcon(rdata->cfile->tlink);
+	struct cifs_tcon *tcon = tlink_tcon(rdata->req->cfile->tlink);
 	struct smb_rqst rqst = { .rq_iov = rdata->iov,
 				 .rq_nvec = 2 };
 
-	cifs_dbg(FYI, "%s: offset=%llu bytes=%u\n",
-		 __func__, rdata->offset, rdata->bytes);
+	cifs_dbg(FYI, "%s: offset=%llu bytes=%zu\n",
+		 __func__, rdata->subreq.start, rdata->subreq.len);
 
 	if (tcon->ses->capabilities & CAP_LARGE_FILES)
 		wct = 12;
 	else {
 		wct = 10; /* old style read */
-		if ((rdata->offset >> 32) > 0)  {
+		if ((rdata->subreq.start >> 32) > 0)  {
 			/* can not handle this big offset for old */
 			return -EIO;
 		}
@@ -1338,17 +1345,17 @@ cifs_async_readv(struct cifs_readdata *rdata)
 	if (rc)
 		return rc;
 
-	smb->hdr.Pid = cpu_to_le16((__u16)rdata->pid);
-	smb->hdr.PidHigh = cpu_to_le16((__u16)(rdata->pid >> 16));
+	smb->hdr.Pid = cpu_to_le16((__u16)rdata->req->pid);
+	smb->hdr.PidHigh = cpu_to_le16((__u16)(rdata->req->pid >> 16));
 
 	smb->AndXCommand = 0xFF;	/* none */
-	smb->Fid = rdata->cfile->fid.netfid;
-	smb->OffsetLow = cpu_to_le32(rdata->offset & 0xFFFFFFFF);
+	smb->Fid = rdata->req->cfile->fid.netfid;
+	smb->OffsetLow = cpu_to_le32(rdata->subreq.start & 0xFFFFFFFF);
 	if (wct == 12)
-		smb->OffsetHigh = cpu_to_le32(rdata->offset >> 32);
+		smb->OffsetHigh = cpu_to_le32(rdata->subreq.start >> 32);
 	smb->Remaining = 0;
-	smb->MaxCount = cpu_to_le16(rdata->bytes & 0xFFFF);
-	smb->MaxCountHigh = cpu_to_le32(rdata->bytes >> 16);
+	smb->MaxCount = cpu_to_le16(rdata->subreq.len & 0xFFFF);
+	smb->MaxCountHigh = cpu_to_le32(rdata->subreq.len >> 16);
 	if (wct == 12)
 		smb->ByteCount = 0;
 	else {
@@ -1364,15 +1371,11 @@ cifs_async_readv(struct cifs_readdata *rdata)
 	rdata->iov[1].iov_base = (char *)smb + 4;
 	rdata->iov[1].iov_len = get_rfc1002_length(smb);
 
-	kref_get(&rdata->refcount);
 	rc = cifs_call_async(tcon->ses->server, &rqst, cifs_readv_receive,
 			     cifs_readv_callback, NULL, rdata, 0, NULL);
 
 	if (rc == 0)
 		cifs_stats_inc(&tcon->stats.cifs_stats.num_reads);
-	else
-		kref_put(&rdata->refcount, cifs_readdata_release);
-
 	cifs_small_buf_release(smb);
 	return rc;
 }
@@ -1615,16 +1618,17 @@ CIFSSMBWrite(const unsigned int xid, struct cifs_io_parms *io_parms,
 static void
 cifs_writev_callback(struct mid_q_entry *mid)
 {
-	struct cifs_writedata *wdata = mid->callback_data;
-	struct cifs_tcon *tcon = tlink_tcon(wdata->cfile->tlink);
-	unsigned int written;
+	struct cifs_io_subrequest *wdata = mid->callback_data;
+	struct cifs_tcon *tcon = tlink_tcon(wdata->req->cfile->tlink);
 	WRITE_RSP *smb = (WRITE_RSP *)mid->resp_buf;
 	struct cifs_credits credits = { .value = 1, .instance = 0 };
+	ssize_t result;
+	size_t written;
 
 	switch (mid->mid_state) {
 	case MID_RESPONSE_RECEIVED:
-		wdata->result = cifs_check_receive(mid, tcon->ses->server, 0);
-		if (wdata->result != 0)
+		result = cifs_check_receive(mid, tcon->ses->server, 0);
+		if (result != 0)
 			break;
 
 		written = le16_to_cpu(smb->CountHigh);
@@ -1636,37 +1640,37 @@ cifs_writev_callback(struct mid_q_entry *mid)
 		 * client. OS/2 servers are known to set incorrect
 		 * CountHigh values.
 		 */
-		if (written > wdata->bytes)
+		if (written > wdata->subreq.len)
 			written &= 0xFFFF;
 
-		if (written < wdata->bytes)
-			wdata->result = -ENOSPC;
+		if (written < wdata->subreq.len)
+			result = -ENOSPC;
 		else
-			wdata->bytes = written;
+			result = written;
 		break;
 	case MID_REQUEST_SUBMITTED:
 	case MID_RETRY_NEEDED:
-		wdata->result = -EAGAIN;
+		result = -EAGAIN;
 		break;
 	default:
-		wdata->result = -EIO;
+		result = -EIO;
 		break;
 	}
 
-	queue_work(cifsiod_wq, &wdata->work);
+	wdata->credits.value = 0;
+	cifs_write_subrequest_terminated(wdata, result, true);
 	release_mid(mid);
 	add_credits(tcon->ses->server, &credits, 0);
 }
 
 /* cifs_async_writev - send an async write, and set up mid to handle result */
-int
-cifs_async_writev(struct cifs_writedata *wdata,
-		  void (*release)(struct kref *kref))
+void
+cifs_async_writev(struct cifs_io_subrequest *wdata)
 {
 	int rc = -EACCES;
 	WRITE_REQ *smb = NULL;
 	int wct;
-	struct cifs_tcon *tcon = tlink_tcon(wdata->cfile->tlink);
+	struct cifs_tcon *tcon = tlink_tcon(wdata->req->cfile->tlink);
 	struct kvec iov[2];
 	struct smb_rqst rqst = { };
 
@@ -1674,9 +1678,10 @@ cifs_async_writev(struct cifs_writedata *wdata,
 		wct = 14;
 	} else {
 		wct = 12;
-		if (wdata->offset >> 32 > 0) {
+		if (wdata->subreq.start >> 32 > 0) {
 			/* can not handle big offset for old srv */
-			return -EIO;
+			rc = -EIO;
+			goto out;
 		}
 	}
 
@@ -1684,14 +1689,14 @@ cifs_async_writev(struct cifs_writedata *wdata,
 	if (rc)
 		goto async_writev_out;
 
-	smb->hdr.Pid = cpu_to_le16((__u16)wdata->pid);
-	smb->hdr.PidHigh = cpu_to_le16((__u16)(wdata->pid >> 16));
+	smb->hdr.Pid = cpu_to_le16((__u16)wdata->req->pid);
+	smb->hdr.PidHigh = cpu_to_le16((__u16)(wdata->req->pid >> 16));
 
 	smb->AndXCommand = 0xFF;	/* none */
-	smb->Fid = wdata->cfile->fid.netfid;
-	smb->OffsetLow = cpu_to_le32(wdata->offset & 0xFFFFFFFF);
+	smb->Fid = wdata->req->cfile->fid.netfid;
+	smb->OffsetLow = cpu_to_le32(wdata->subreq.start & 0xFFFFFFFF);
 	if (wct == 14)
-		smb->OffsetHigh = cpu_to_le32(wdata->offset >> 32);
+		smb->OffsetHigh = cpu_to_le32(wdata->subreq.start >> 32);
 	smb->Reserved = 0xFFFFFFFF;
 	smb->WriteMode = 0;
 	smb->Remaining = 0;
@@ -1707,39 +1712,40 @@ cifs_async_writev(struct cifs_writedata *wdata,
 
 	rqst.rq_iov = iov;
 	rqst.rq_nvec = 2;
-	rqst.rq_iter = wdata->iter;
-	rqst.rq_iter_size = iov_iter_count(&wdata->iter);
+	rqst.rq_iter = wdata->subreq.io_iter;
+	rqst.rq_iter_size = iov_iter_count(&wdata->subreq.io_iter);
 
-	cifs_dbg(FYI, "async write at %llu %u bytes\n",
-		 wdata->offset, wdata->bytes);
+	cifs_dbg(FYI, "async write at %llu %zu bytes\n",
+		 wdata->subreq.start, wdata->subreq.len);
 
-	smb->DataLengthLow = cpu_to_le16(wdata->bytes & 0xFFFF);
-	smb->DataLengthHigh = cpu_to_le16(wdata->bytes >> 16);
+	smb->DataLengthLow = cpu_to_le16(wdata->subreq.len & 0xFFFF);
+	smb->DataLengthHigh = cpu_to_le16(wdata->subreq.len >> 16);
 
 	if (wct == 14) {
-		inc_rfc1001_len(&smb->hdr, wdata->bytes + 1);
-		put_bcc(wdata->bytes + 1, &smb->hdr);
+		inc_rfc1001_len(&smb->hdr, wdata->subreq.len + 1);
+		put_bcc(wdata->subreq.len + 1, &smb->hdr);
 	} else {
 		/* wct == 12 */
 		struct smb_com_writex_req *smbw =
 				(struct smb_com_writex_req *)smb;
-		inc_rfc1001_len(&smbw->hdr, wdata->bytes + 5);
-		put_bcc(wdata->bytes + 5, &smbw->hdr);
+		inc_rfc1001_len(&smbw->hdr, wdata->subreq.len + 5);
+		put_bcc(wdata->subreq.len + 5, &smbw->hdr);
 		iov[1].iov_len += 4; /* pad bigger by four bytes */
 	}
 
-	kref_get(&wdata->refcount);
 	rc = cifs_call_async(tcon->ses->server, &rqst, NULL,
 			     cifs_writev_callback, NULL, wdata, 0, NULL);
-
+	/* Can't touch wdata if rc == 0 */
 	if (rc == 0)
 		cifs_stats_inc(&tcon->stats.cifs_stats.num_writes);
-	else
-		kref_put(&wdata->refcount, release);
 
 async_writev_out:
 	cifs_small_buf_release(smb);
-	return rc;
+out:
+	if (rc) {
+		add_credits_and_wake_if(wdata->server, &wdata->credits, 0);
+		cifs_write_subrequest_terminated(wdata, rc, false);
+	}
 }
 
 int
@@ -2066,20 +2072,20 @@ CIFSSMBPosixLock(const unsigned int xid, struct cifs_tcon *tcon,
 		parm_data = (struct cifs_posix_lock *)
 			((char *)&pSMBr->hdr.Protocol + data_offset);
 		if (parm_data->lock_type == cpu_to_le16(CIFS_UNLCK))
-			pLockData->fl_type = F_UNLCK;
+			pLockData->c.flc_type = F_UNLCK;
 		else {
 			if (parm_data->lock_type ==
 					cpu_to_le16(CIFS_RDLCK))
-				pLockData->fl_type = F_RDLCK;
+				pLockData->c.flc_type = F_RDLCK;
 			else if (parm_data->lock_type ==
 					cpu_to_le16(CIFS_WRLCK))
-				pLockData->fl_type = F_WRLCK;
+				pLockData->c.flc_type = F_WRLCK;
 
 			pLockData->fl_start = le64_to_cpu(parm_data->start);
 			pLockData->fl_end = pLockData->fl_start +
 				(le64_to_cpu(parm_data->length) ?
 				 le64_to_cpu(parm_data->length) - 1 : 0);
-			pLockData->fl_pid = -le32_to_cpu(parm_data->pid);
+			pLockData->c.flc_pid = -le32_to_cpu(parm_data->pid);
 		}
 	}
 
@@ -4993,7 +4999,7 @@ QFSPosixRetry:
 int
 CIFSSMBSetEOF(const unsigned int xid, struct cifs_tcon *tcon,
 	      const char *file_name, __u64 size, struct cifs_sb_info *cifs_sb,
-	      bool set_allocation)
+	      bool set_allocation, struct dentry *dentry)
 {
 	struct smb_com_transaction2_spi_req *pSMB = NULL;
 	struct smb_com_transaction2_spi_rsp *pSMBr = NULL;
@@ -5854,10 +5860,8 @@ SetEARetry:
 	parm_data->list.EA_flags = 0;
 	/* we checked above that name len is less than 255 */
 	parm_data->list.name_len = (__u8)name_len;
-	/* EA names are always ASCII */
-	if (ea_name)
-		strncpy(parm_data->list.name, ea_name, name_len);
-	parm_data->list.name[name_len] = '\0';
+	/* EA names are always ASCII and NUL-terminated */
+	strscpy(parm_data->list.name, ea_name ?: "", name_len + 1);
 	parm_data->list.value_len = cpu_to_le16(ea_value_len);
 	/* caller ensures that ea_value_len is less than 64K but
 	we need to ensure that it fits within the smb */
diff --git a/fs/smb/client/connect.c b/fs/smb/client/connect.c
index e28f011f11..d2307162a2 100644
--- a/fs/smb/client/connect.c
+++ b/fs/smb/client/connect.c
@@ -52,9 +52,6 @@
 #include "fs_context.h"
 #include "cifs_swn.h"
 
-extern mempool_t *cifs_req_poolp;
-extern bool disable_legacy_dialects;
-
 /* FIXME: should these be tunable? */
 #define TLINK_ERROR_EXPIRE	(1 * HZ)
 #define TLINK_IDLE_EXPIRE	(600 * HZ)
@@ -1748,7 +1745,7 @@ cifs_get_tcp_session(struct smb3_fs_context *ctx,
 	tcp_ses->channel_sequence_num = 0; /* only tracked for primary channel */
 	tcp_ses->reconnect_instance = 1;
 	tcp_ses->lstrp = jiffies;
-	tcp_ses->compress_algorithm = cpu_to_le16(ctx->compression);
+	tcp_ses->compression.requested = ctx->compress;
 	spin_lock_init(&tcp_ses->req_lock);
 	spin_lock_init(&tcp_ses->srv_lock);
 	spin_lock_init(&tcp_ses->mid_lock);
@@ -1946,7 +1943,7 @@ cifs_setup_ipc(struct cifs_ses *ses, struct smb3_fs_context *ctx)
 	}
 
 	/* no need to setup directory caching on IPC share, so pass in false */
-	tcon = tcon_info_alloc(false);
+	tcon = tcon_info_alloc(false, netfs_trace_tcon_ref_new_ipc);
 	if (tcon == NULL)
 		return -ENOMEM;
 
@@ -1963,7 +1960,7 @@ cifs_setup_ipc(struct cifs_ses *ses, struct smb3_fs_context *ctx)
 
 	if (rc) {
 		cifs_server_dbg(VFS, "failed to connect to IPC (rc=%d)\n", rc);
-		tconInfoFree(tcon);
+		tconInfoFree(tcon, netfs_trace_tcon_ref_free_ipc_fail);
 		goto out;
 	}
 
@@ -2046,7 +2043,7 @@ void __cifs_put_smb_ses(struct cifs_ses *ses)
 	 * files on session close, as specified in MS-SMB2 3.3.5.6 Receiving an
 	 * SMB2 LOGOFF Request.
 	 */
-	tconInfoFree(tcon);
+	tconInfoFree(tcon, netfs_trace_tcon_ref_free_ipc);
 	if (do_logoff) {
 		xid = get_xid();
 		rc = server->ops->logoff(xid, ses);
@@ -2435,6 +2432,8 @@ cifs_find_tcon(struct cifs_ses *ses, struct smb3_fs_context *ctx)
 			continue;
 		}
 		++tcon->tc_count;
+		trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+				    netfs_trace_tcon_ref_get_find);
 		spin_unlock(&tcon->tc_lock);
 		spin_unlock(&cifs_tcp_ses_lock);
 		return tcon;
@@ -2444,7 +2443,7 @@ cifs_find_tcon(struct cifs_ses *ses, struct smb3_fs_context *ctx)
 }
 
 void
-cifs_put_tcon(struct cifs_tcon *tcon)
+cifs_put_tcon(struct cifs_tcon *tcon, enum smb3_tcon_ref_trace trace)
 {
 	unsigned int xid;
 	struct cifs_ses *ses;
@@ -2460,6 +2459,7 @@ cifs_put_tcon(struct cifs_tcon *tcon)
 	cifs_dbg(FYI, "%s: tc_count=%d\n", __func__, tcon->tc_count);
 	spin_lock(&cifs_tcp_ses_lock);
 	spin_lock(&tcon->tc_lock);
+	trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count - 1, trace);
 	if (--tcon->tc_count > 0) {
 		spin_unlock(&tcon->tc_lock);
 		spin_unlock(&cifs_tcp_ses_lock);
@@ -2496,7 +2496,7 @@ cifs_put_tcon(struct cifs_tcon *tcon)
 	_free_xid(xid);
 
 	cifs_fscache_release_super_cookie(tcon);
-	tconInfoFree(tcon);
+	tconInfoFree(tcon, netfs_trace_tcon_ref_free);
 	cifs_put_smb_ses(ses);
 }
 
@@ -2550,7 +2550,7 @@ cifs_get_tcon(struct cifs_ses *ses, struct smb3_fs_context *ctx)
 		nohandlecache = ctx->nohandlecache;
 	else
 		nohandlecache = true;
-	tcon = tcon_info_alloc(!nohandlecache);
+	tcon = tcon_info_alloc(!nohandlecache, netfs_trace_tcon_ref_new);
 	if (tcon == NULL) {
 		rc = -ENOMEM;
 		goto out_fail;
@@ -2614,6 +2614,13 @@ cifs_get_tcon(struct cifs_ses *ses, struct smb3_fs_context *ctx)
 			cifs_dbg(VFS, "Server does not support mounting with posix SMB3.11 extensions\n");
 			rc = -EOPNOTSUPP;
 			goto out_fail;
+		} else if (ses->server->vals->protocol_id == SMB10_PROT_ID)
+			if (cap_unix(ses))
+				cifs_dbg(FYI, "Unix Extensions requested on SMB1 mount\n");
+			else {
+				cifs_dbg(VFS, "SMB1 Unix Extensions not supported by server\n");
+				rc = -EOPNOTSUPP;
+				goto out_fail;
 		} else {
 			cifs_dbg(VFS,
 				"Check vers= mount option. SMB3.11 disabled but required for POSIX extensions\n");
@@ -2740,7 +2747,7 @@ cifs_get_tcon(struct cifs_ses *ses, struct smb3_fs_context *ctx)
 	return tcon;
 
 out_fail:
-	tconInfoFree(tcon);
+	tconInfoFree(tcon, netfs_trace_tcon_ref_free_fail);
 	return ERR_PTR(rc);
 }
 
@@ -2757,7 +2764,7 @@ cifs_put_tlink(struct tcon_link *tlink)
 	}
 
 	if (!IS_ERR(tlink_tcon(tlink)))
-		cifs_put_tcon(tlink_tcon(tlink));
+		cifs_put_tcon(tlink_tcon(tlink), netfs_trace_tcon_ref_put_tlink);
 	kfree(tlink);
 }
 
@@ -2805,6 +2812,8 @@ compare_mount_options(struct super_block *sb, struct cifs_mnt_data *mnt_data)
 		return 0;
 	if (old->ctx->closetimeo != new->ctx->closetimeo)
 		return 0;
+	if (old->ctx->reparse_type != new->ctx->reparse_type)
+		return 0;
 
 	return 1;
 }
@@ -3320,7 +3329,7 @@ void cifs_mount_put_conns(struct cifs_mount_ctx *mnt_ctx)
 	int rc = 0;
 
 	if (mnt_ctx->tcon)
-		cifs_put_tcon(mnt_ctx->tcon);
+		cifs_put_tcon(mnt_ctx->tcon, netfs_trace_tcon_ref_put_mnt_ctx);
 	else if (mnt_ctx->ses)
 		cifs_put_smb_ses(mnt_ctx->ses);
 	else if (mnt_ctx->server)
@@ -3684,6 +3693,7 @@ error:
 }
 #endif
 
+#ifdef CONFIG_CIFS_ALLOW_INSECURE_LEGACY
 /*
  * Issue a TREE_CONNECT request.
  */
@@ -3805,11 +3815,25 @@ CIFSTCon(const unsigned int xid, struct cifs_ses *ses,
 		else
 			tcon->Flags = 0;
 		cifs_dbg(FYI, "Tcon flags: 0x%x\n", tcon->Flags);
-	}
 
+		/*
+		 * reset_cifs_unix_caps calls QFSInfo which requires
+		 * need_reconnect to be false, but we would not need to call
+		 * reset_caps if this were not a reconnect case so must check
+		 * need_reconnect flag here.  The caller will also clear
+		 * need_reconnect when tcon was successful but needed to be
+		 * cleared earlier in the case of unix extensions reconnect
+		 */
+		if (tcon->need_reconnect && tcon->unix_ext) {
+			cifs_dbg(FYI, "resetting caps for %s\n", tcon->tree_name);
+			tcon->need_reconnect = false;
+			reset_cifs_unix_caps(xid, tcon, NULL, NULL);
+		}
+	}
 	cifs_buf_release(smb_buffer);
 	return rc;
 }
+#endif /* CONFIG_CIFS_ALLOW_INSECURE_LEGACY */
 
 static void delayed_free(struct rcu_head *p)
 {
diff --git a/fs/smb/client/dir.c b/fs/smb/client/dir.c
index 37897b919d..864b194dba 100644
--- a/fs/smb/client/dir.c
+++ b/fs/smb/client/dir.c
@@ -627,11 +627,18 @@ int cifs_mknod(struct mnt_idmap *idmap, struct inode *inode,
 		goto mknod_out;
 	}
 
+	trace_smb3_mknod_enter(xid, tcon->ses->Suid, tcon->tid, full_path);
+
 	rc = tcon->ses->server->ops->make_node(xid, inode, direntry, tcon,
 					       full_path, mode,
 					       device_number);
 
 mknod_out:
+	if (rc)
+		trace_smb3_mknod_err(xid,  tcon->ses->Suid, tcon->tid, rc);
+	else
+		trace_smb3_mknod_done(xid, tcon->ses->Suid, tcon->tid);
+
 	free_dentry_path(page);
 	free_xid(xid);
 	cifs_put_tlink(tlink);
diff --git a/fs/smb/client/file.c b/fs/smb/client/file.c
index af5c476db6..04ec1b9737 100644
--- a/fs/smb/client/file.c
+++ b/fs/smb/client/file.c
@@ -36,131 +36,284 @@
 #include "fs_context.h"
 #include "cifs_ioctl.h"
 #include "cached_dir.h"
+#include <trace/events/netfs.h>
+
+static int cifs_reopen_file(struct cifsFileInfo *cfile, bool can_flush);
 
 /*
- * Remove the dirty flags from a span of pages.
+ * Prepare a subrequest to upload to the server.  We need to allocate credits
+ * so that we know the maximum amount of data that we can include in it.
  */
-static void cifs_undirty_folios(struct inode *inode, loff_t start, unsigned int len)
+static void cifs_prepare_write(struct netfs_io_subrequest *subreq)
 {
-	struct address_space *mapping = inode->i_mapping;
-	struct folio *folio;
-	pgoff_t end;
+	struct cifs_io_subrequest *wdata =
+		container_of(subreq, struct cifs_io_subrequest, subreq);
+	struct cifs_io_request *req = wdata->req;
+	struct TCP_Server_Info *server;
+	struct cifsFileInfo *open_file = req->cfile;
+	size_t wsize = req->rreq.wsize;
+	int rc;
 
-	XA_STATE(xas, &mapping->i_pages, start / PAGE_SIZE);
+	if (!wdata->have_xid) {
+		wdata->xid = get_xid();
+		wdata->have_xid = true;
+	}
 
-	rcu_read_lock();
+	server = cifs_pick_channel(tlink_tcon(open_file->tlink)->ses);
+	wdata->server = server;
 
-	end = (start + len - 1) / PAGE_SIZE;
-	xas_for_each_marked(&xas, folio, end, PAGECACHE_TAG_DIRTY) {
-		if (xas_retry(&xas, folio))
-			continue;
-		xas_pause(&xas);
-		rcu_read_unlock();
-		folio_lock(folio);
-		folio_clear_dirty_for_io(folio);
-		folio_unlock(folio);
-		rcu_read_lock();
+retry:
+	if (open_file->invalidHandle) {
+		rc = cifs_reopen_file(open_file, false);
+		if (rc < 0) {
+			if (rc == -EAGAIN)
+				goto retry;
+			subreq->error = rc;
+			return netfs_prepare_write_failed(subreq);
+		}
+	}
+
+	rc = server->ops->wait_mtu_credits(server, wsize, &wdata->subreq.max_len,
+					   &wdata->credits);
+	if (rc < 0) {
+		subreq->error = rc;
+		return netfs_prepare_write_failed(subreq);
 	}
 
-	rcu_read_unlock();
+#ifdef CONFIG_CIFS_SMB_DIRECT
+	if (server->smbd_conn)
+		subreq->max_nr_segs = server->smbd_conn->max_frmr_depth;
+#endif
 }
 
 /*
- * Completion of write to server.
+ * Issue a subrequest to upload to the server.
  */
-void cifs_pages_written_back(struct inode *inode, loff_t start, unsigned int len)
+static void cifs_issue_write(struct netfs_io_subrequest *subreq)
 {
-	struct address_space *mapping = inode->i_mapping;
-	struct folio *folio;
-	pgoff_t end;
+	struct cifs_io_subrequest *wdata =
+		container_of(subreq, struct cifs_io_subrequest, subreq);
+	struct cifs_sb_info *sbi = CIFS_SB(subreq->rreq->inode->i_sb);
+	int rc;
 
-	XA_STATE(xas, &mapping->i_pages, start / PAGE_SIZE);
+	if (cifs_forced_shutdown(sbi)) {
+		rc = -EIO;
+		goto fail;
+	}
 
-	if (!len)
-		return;
+	rc = adjust_credits(wdata->server, &wdata->credits, wdata->subreq.len);
+	if (rc)
+		goto fail;
 
-	rcu_read_lock();
+	rc = -EAGAIN;
+	if (wdata->req->cfile->invalidHandle)
+		goto fail;
 
-	end = (start + len - 1) / PAGE_SIZE;
-	xas_for_each(&xas, folio, end) {
-		if (xas_retry(&xas, folio))
-			continue;
-		if (!folio_test_writeback(folio)) {
-			WARN_ONCE(1, "bad %x @%llx page %lx %lx\n",
-				  len, start, folio->index, end);
-			continue;
-		}
+	wdata->server->ops->async_writev(wdata);
+out:
+	return;
+
+fail:
+	if (rc == -EAGAIN)
+		trace_netfs_sreq(subreq, netfs_sreq_trace_retry);
+	else
+		trace_netfs_sreq(subreq, netfs_sreq_trace_fail);
+	add_credits_and_wake_if(wdata->server, &wdata->credits, 0);
+	cifs_write_subrequest_terminated(wdata, rc, false);
+	goto out;
+}
+
+static void cifs_netfs_invalidate_cache(struct netfs_io_request *wreq)
+{
+	cifs_invalidate_cache(wreq->inode, 0);
+}
 
-		folio_detach_private(folio);
-		folio_end_writeback(folio);
+/*
+ * Split the read up according to how many credits we can get for each piece.
+ * It's okay to sleep here if we need to wait for more credit to become
+ * available.
+ *
+ * We also choose the server and allocate an operation ID to be cleaned up
+ * later.
+ */
+static bool cifs_clamp_length(struct netfs_io_subrequest *subreq)
+{
+	struct netfs_io_request *rreq = subreq->rreq;
+	struct cifs_io_subrequest *rdata = container_of(subreq, struct cifs_io_subrequest, subreq);
+	struct cifs_io_request *req = container_of(subreq->rreq, struct cifs_io_request, rreq);
+	struct TCP_Server_Info *server = req->server;
+	struct cifs_sb_info *cifs_sb = CIFS_SB(rreq->inode->i_sb);
+	size_t rsize = 0;
+	int rc;
+
+	rdata->xid = get_xid();
+	rdata->have_xid = true;
+	rdata->server = server;
+
+	if (cifs_sb->ctx->rsize == 0)
+		cifs_sb->ctx->rsize =
+			server->ops->negotiate_rsize(tlink_tcon(req->cfile->tlink),
+						     cifs_sb->ctx);
+
+
+	rc = server->ops->wait_mtu_credits(server, cifs_sb->ctx->rsize, &rsize,
+					   &rdata->credits);
+	if (rc) {
+		subreq->error = rc;
+		return false;
 	}
 
-	rcu_read_unlock();
+	subreq->len = min_t(size_t, subreq->len, rsize);
+#ifdef CONFIG_CIFS_SMB_DIRECT
+	if (server->smbd_conn)
+		subreq->max_nr_segs = server->smbd_conn->max_frmr_depth;
+#endif
+	return true;
 }
 
 /*
- * Failure of write to server.
+ * Issue a read operation on behalf of the netfs helper functions.  We're asked
+ * to make a read of a certain size at a point in the file.  We are permitted
+ * to only read a portion of that, but as long as we read something, the netfs
+ * helper will call us again so that we can issue another read.
  */
-void cifs_pages_write_failed(struct inode *inode, loff_t start, unsigned int len)
+static void cifs_req_issue_read(struct netfs_io_subrequest *subreq)
 {
-	struct address_space *mapping = inode->i_mapping;
-	struct folio *folio;
-	pgoff_t end;
+	struct netfs_io_request *rreq = subreq->rreq;
+	struct cifs_io_subrequest *rdata = container_of(subreq, struct cifs_io_subrequest, subreq);
+	struct cifs_io_request *req = container_of(subreq->rreq, struct cifs_io_request, rreq);
+	int rc = 0;
+
+	cifs_dbg(FYI, "%s: op=%08x[%x] mapping=%p len=%zu/%zu\n",
+		 __func__, rreq->debug_id, subreq->debug_index, rreq->mapping,
+		 subreq->transferred, subreq->len);
+
+	if (req->cfile->invalidHandle) {
+		do {
+			rc = cifs_reopen_file(req->cfile, true);
+		} while (rc == -EAGAIN);
+		if (rc)
+			goto out;
+	}
+
+	__set_bit(NETFS_SREQ_CLEAR_TAIL, &subreq->flags);
+
+	rc = rdata->server->ops->async_readv(rdata);
+out:
+	if (rc)
+		netfs_subreq_terminated(subreq, rc, false);
+}
 
-	XA_STATE(xas, &mapping->i_pages, start / PAGE_SIZE);
+/*
+ * Writeback calls this when it finds a folio that needs uploading.  This isn't
+ * called if writeback only has copy-to-cache to deal with.
+ */
+static void cifs_begin_writeback(struct netfs_io_request *wreq)
+{
+	struct cifs_io_request *req = container_of(wreq, struct cifs_io_request, rreq);
+	int ret;
 
-	if (!len)
+	ret = cifs_get_writable_file(CIFS_I(wreq->inode), FIND_WR_ANY, &req->cfile);
+	if (ret) {
+		cifs_dbg(VFS, "No writable handle in writepages ret=%d\n", ret);
 		return;
+	}
 
-	rcu_read_lock();
+	wreq->io_streams[0].avail = true;
+}
 
-	end = (start + len - 1) / PAGE_SIZE;
-	xas_for_each(&xas, folio, end) {
-		if (xas_retry(&xas, folio))
-			continue;
-		if (!folio_test_writeback(folio)) {
-			WARN_ONCE(1, "bad %x @%llx page %lx %lx\n",
-				  len, start, folio->index, end);
-			continue;
-		}
+/*
+ * Initialise a request.
+ */
+static int cifs_init_request(struct netfs_io_request *rreq, struct file *file)
+{
+	struct cifs_io_request *req = container_of(rreq, struct cifs_io_request, rreq);
+	struct cifs_sb_info *cifs_sb = CIFS_SB(rreq->inode->i_sb);
+	struct cifsFileInfo *open_file = NULL;
 
-		folio_set_error(folio);
-		folio_end_writeback(folio);
+	rreq->rsize = cifs_sb->ctx->rsize;
+	rreq->wsize = cifs_sb->ctx->wsize;
+	req->pid = current->tgid; // Ummm...  This may be a workqueue
+
+	if (file) {
+		open_file = file->private_data;
+		rreq->netfs_priv = file->private_data;
+		req->cfile = cifsFileInfo_get(open_file);
+		req->server = cifs_pick_channel(tlink_tcon(req->cfile->tlink)->ses);
+		if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_RWPIDFORWARD)
+			req->pid = req->cfile->pid;
+	} else if (rreq->origin != NETFS_WRITEBACK) {
+		WARN_ON_ONCE(1);
+		return -EIO;
 	}
 
-	rcu_read_unlock();
+	return 0;
 }
 
 /*
- * Redirty pages after a temporary failure.
+ * Completion of a request operation.
  */
-void cifs_pages_write_redirty(struct inode *inode, loff_t start, unsigned int len)
+static void cifs_rreq_done(struct netfs_io_request *rreq)
 {
-	struct address_space *mapping = inode->i_mapping;
-	struct folio *folio;
-	pgoff_t end;
+	struct timespec64 atime, mtime;
+	struct inode *inode = rreq->inode;
 
-	XA_STATE(xas, &mapping->i_pages, start / PAGE_SIZE);
+	/* we do not want atime to be less than mtime, it broke some apps */
+	atime = inode_set_atime_to_ts(inode, current_time(inode));
+	mtime = inode_get_mtime(inode);
+	if (timespec64_compare(&atime, &mtime))
+		inode_set_atime_to_ts(inode, inode_get_mtime(inode));
+}
 
-	if (!len)
-		return;
+static void cifs_post_modify(struct inode *inode)
+{
+	/* Indication to update ctime and mtime as close is deferred */
+	set_bit(CIFS_INO_MODIFIED_ATTR, &CIFS_I(inode)->flags);
+}
+
+static void cifs_free_request(struct netfs_io_request *rreq)
+{
+	struct cifs_io_request *req = container_of(rreq, struct cifs_io_request, rreq);
 
-	rcu_read_lock();
+	if (req->cfile)
+		cifsFileInfo_put(req->cfile);
+}
 
-	end = (start + len - 1) / PAGE_SIZE;
-	xas_for_each(&xas, folio, end) {
-		if (!folio_test_writeback(folio)) {
-			WARN_ONCE(1, "bad %x @%llx page %lx %lx\n",
-				  len, start, folio->index, end);
-			continue;
-		}
+static void cifs_free_subrequest(struct netfs_io_subrequest *subreq)
+{
+	struct cifs_io_subrequest *rdata =
+		container_of(subreq, struct cifs_io_subrequest, subreq);
+	int rc = subreq->error;
 
-		filemap_dirty_folio(folio->mapping, folio);
-		folio_end_writeback(folio);
+	if (rdata->subreq.source == NETFS_DOWNLOAD_FROM_SERVER) {
+#ifdef CONFIG_CIFS_SMB_DIRECT
+		if (rdata->mr) {
+			smbd_deregister_mr(rdata->mr);
+			rdata->mr = NULL;
+		}
+#endif
 	}
 
-	rcu_read_unlock();
-}
+	add_credits_and_wake_if(rdata->server, &rdata->credits, 0);
+	if (rdata->have_xid)
+		free_xid(rdata->xid);
+}
+
+const struct netfs_request_ops cifs_req_ops = {
+	.request_pool		= &cifs_io_request_pool,
+	.subrequest_pool	= &cifs_io_subrequest_pool,
+	.init_request		= cifs_init_request,
+	.free_request		= cifs_free_request,
+	.free_subrequest	= cifs_free_subrequest,
+	.clamp_length		= cifs_clamp_length,
+	.issue_read		= cifs_req_issue_read,
+	.done			= cifs_rreq_done,
+	.post_modify		= cifs_post_modify,
+	.begin_writeback	= cifs_begin_writeback,
+	.prepare_write		= cifs_prepare_write,
+	.issue_write		= cifs_issue_write,
+	.invalidate_cache	= cifs_netfs_invalidate_cache,
+};
 
 /*
  * Mark as invalid, all open files on tree connections since they
@@ -1151,6 +1304,19 @@ void smb2_deferred_work_close(struct work_struct *work)
 	_cifsFileInfo_put(cfile, true, false);
 }
 
+static bool
+smb2_can_defer_close(struct inode *inode, struct cifs_deferred_close *dclose)
+{
+	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
+	struct cifsInodeInfo *cinode = CIFS_I(inode);
+
+	return (cifs_sb->ctx->closetimeo && cinode->lease_granted && dclose &&
+			(cinode->oplock == CIFS_CACHE_RHW_FLG ||
+			 cinode->oplock == CIFS_CACHE_RH_FLG) &&
+			!test_bit(CIFS_INO_CLOSE_ON_LOCK, &cinode->flags));
+
+}
+
 int cifs_close(struct inode *inode, struct file *file)
 {
 	struct cifsFileInfo *cfile;
@@ -1164,10 +1330,8 @@ int cifs_close(struct inode *inode, struct file *file)
 		cfile = file->private_data;
 		file->private_data = NULL;
 		dclose = kmalloc(sizeof(struct cifs_deferred_close), GFP_KERNEL);
-		if ((cifs_sb->ctx->closetimeo && cinode->oplock == CIFS_CACHE_RHW_FLG)
-		    && cinode->lease_granted &&
-		    !test_bit(CIFS_INO_CLOSE_ON_LOCK, &cinode->flags) &&
-		    dclose) {
+		if ((cfile->status_file_deleted == false) &&
+		    (smb2_can_defer_close(inode, dclose))) {
 			if (test_and_clear_bit(CIFS_INO_MODIFIED_ATTR, &cinode->flags)) {
 				inode_set_mtime_to_ts(inode,
 						      inode_set_ctime_current(inode));
@@ -1394,20 +1558,20 @@ cifs_lock_test(struct cifsFileInfo *cfile, __u64 offset, __u64 length,
 	down_read(&cinode->lock_sem);
 
 	exist = cifs_find_lock_conflict(cfile, offset, length, type,
-					flock->fl_flags, &conf_lock,
+					flock->c.flc_flags, &conf_lock,
 					CIFS_LOCK_OP);
 	if (exist) {
 		flock->fl_start = conf_lock->offset;
 		flock->fl_end = conf_lock->offset + conf_lock->length - 1;
-		flock->fl_pid = conf_lock->pid;
+		flock->c.flc_pid = conf_lock->pid;
 		if (conf_lock->type & server->vals->shared_lock_type)
-			flock->fl_type = F_RDLCK;
+			flock->c.flc_type = F_RDLCK;
 		else
-			flock->fl_type = F_WRLCK;
+			flock->c.flc_type = F_WRLCK;
 	} else if (!cinode->can_cache_brlcks)
 		rc = 1;
 	else
-		flock->fl_type = F_UNLCK;
+		flock->c.flc_type = F_UNLCK;
 
 	up_read(&cinode->lock_sem);
 	return rc;
@@ -1483,16 +1647,16 @@ cifs_posix_lock_test(struct file *file, struct file_lock *flock)
 {
 	int rc = 0;
 	struct cifsInodeInfo *cinode = CIFS_I(file_inode(file));
-	unsigned char saved_type = flock->fl_type;
+	unsigned char saved_type = flock->c.flc_type;
 
-	if ((flock->fl_flags & FL_POSIX) == 0)
+	if ((flock->c.flc_flags & FL_POSIX) == 0)
 		return 1;
 
 	down_read(&cinode->lock_sem);
 	posix_test_lock(file, flock);
 
-	if (flock->fl_type == F_UNLCK && !cinode->can_cache_brlcks) {
-		flock->fl_type = saved_type;
+	if (lock_is_unlock(flock) && !cinode->can_cache_brlcks) {
+		flock->c.flc_type = saved_type;
 		rc = 1;
 	}
 
@@ -1513,7 +1677,7 @@ cifs_posix_lock_set(struct file *file, struct file_lock *flock)
 	struct cifsInodeInfo *cinode = CIFS_I(file_inode(file));
 	int rc = FILE_LOCK_DEFERRED + 1;
 
-	if ((flock->fl_flags & FL_POSIX) == 0)
+	if ((flock->c.flc_flags & FL_POSIX) == 0)
 		return rc;
 
 	cifs_down_write(&cinode->lock_sem);
@@ -1663,7 +1827,9 @@ cifs_push_posix_locks(struct cifsFileInfo *cfile)
 
 	el = locks_to_send.next;
 	spin_lock(&flctx->flc_lock);
-	list_for_each_entry(flock, &flctx->flc_posix, fl_list) {
+	for_each_file_lock(flock, &flctx->flc_posix) {
+		unsigned char ftype = flock->c.flc_type;
+
 		if (el == &locks_to_send) {
 			/*
 			 * The list ended. We don't have enough allocated
@@ -1673,12 +1839,12 @@ cifs_push_posix_locks(struct cifsFileInfo *cfile)
 			break;
 		}
 		length = cifs_flock_len(flock);
-		if (flock->fl_type == F_RDLCK || flock->fl_type == F_SHLCK)
+		if (ftype == F_RDLCK || ftype == F_SHLCK)
 			type = CIFS_RDLCK;
 		else
 			type = CIFS_WRLCK;
 		lck = list_entry(el, struct lock_to_push, llist);
-		lck->pid = hash_lockowner(flock->fl_owner);
+		lck->pid = hash_lockowner(flock->c.flc_owner);
 		lck->netfid = cfile->fid.netfid;
 		lck->length = length;
 		lck->type = type;
@@ -1745,42 +1911,43 @@ static void
 cifs_read_flock(struct file_lock *flock, __u32 *type, int *lock, int *unlock,
 		bool *wait_flag, struct TCP_Server_Info *server)
 {
-	if (flock->fl_flags & FL_POSIX)
+	if (flock->c.flc_flags & FL_POSIX)
 		cifs_dbg(FYI, "Posix\n");
-	if (flock->fl_flags & FL_FLOCK)
+	if (flock->c.flc_flags & FL_FLOCK)
 		cifs_dbg(FYI, "Flock\n");
-	if (flock->fl_flags & FL_SLEEP) {
+	if (flock->c.flc_flags & FL_SLEEP) {
 		cifs_dbg(FYI, "Blocking lock\n");
 		*wait_flag = true;
 	}
-	if (flock->fl_flags & FL_ACCESS)
+	if (flock->c.flc_flags & FL_ACCESS)
 		cifs_dbg(FYI, "Process suspended by mandatory locking - not implemented yet\n");
-	if (flock->fl_flags & FL_LEASE)
+	if (flock->c.flc_flags & FL_LEASE)
 		cifs_dbg(FYI, "Lease on file - not implemented yet\n");
-	if (flock->fl_flags &
+	if (flock->c.flc_flags &
 	    (~(FL_POSIX | FL_FLOCK | FL_SLEEP |
 	       FL_ACCESS | FL_LEASE | FL_CLOSE | FL_OFDLCK)))
-		cifs_dbg(FYI, "Unknown lock flags 0x%x\n", flock->fl_flags);
+		cifs_dbg(FYI, "Unknown lock flags 0x%x\n",
+		         flock->c.flc_flags);
 
 	*type = server->vals->large_lock_type;
-	if (flock->fl_type == F_WRLCK) {
+	if (lock_is_write(flock)) {
 		cifs_dbg(FYI, "F_WRLCK\n");
 		*type |= server->vals->exclusive_lock_type;
 		*lock = 1;
-	} else if (flock->fl_type == F_UNLCK) {
+	} else if (lock_is_unlock(flock)) {
 		cifs_dbg(FYI, "F_UNLCK\n");
 		*type |= server->vals->unlock_lock_type;
 		*unlock = 1;
 		/* Check if unlock includes more than one lock range */
-	} else if (flock->fl_type == F_RDLCK) {
+	} else if (lock_is_read(flock)) {
 		cifs_dbg(FYI, "F_RDLCK\n");
 		*type |= server->vals->shared_lock_type;
 		*lock = 1;
-	} else if (flock->fl_type == F_EXLCK) {
+	} else if (flock->c.flc_type == F_EXLCK) {
 		cifs_dbg(FYI, "F_EXLCK\n");
 		*type |= server->vals->exclusive_lock_type;
 		*lock = 1;
-	} else if (flock->fl_type == F_SHLCK) {
+	} else if (flock->c.flc_type == F_SHLCK) {
 		cifs_dbg(FYI, "F_SHLCK\n");
 		*type |= server->vals->shared_lock_type;
 		*lock = 1;
@@ -1812,7 +1979,7 @@ cifs_getlk(struct file *file, struct file_lock *flock, __u32 type,
 		else
 			posix_lock_type = CIFS_WRLCK;
 		rc = CIFSSMBPosixLock(xid, tcon, netfid,
-				      hash_lockowner(flock->fl_owner),
+				      hash_lockowner(flock->c.flc_owner),
 				      flock->fl_start, length, flock,
 				      posix_lock_type, wait_flag);
 		return rc;
@@ -1829,7 +1996,7 @@ cifs_getlk(struct file *file, struct file_lock *flock, __u32 type,
 	if (rc == 0) {
 		rc = server->ops->mand_lock(xid, cfile, flock->fl_start, length,
 					    type, 0, 1, false);
-		flock->fl_type = F_UNLCK;
+		flock->c.flc_type = F_UNLCK;
 		if (rc != 0)
 			cifs_dbg(VFS, "Error unlocking previously locked range %d during test of lock\n",
 				 rc);
@@ -1837,7 +2004,7 @@ cifs_getlk(struct file *file, struct file_lock *flock, __u32 type,
 	}
 
 	if (type & server->vals->shared_lock_type) {
-		flock->fl_type = F_WRLCK;
+		flock->c.flc_type = F_WRLCK;
 		return 0;
 	}
 
@@ -1849,12 +2016,12 @@ cifs_getlk(struct file *file, struct file_lock *flock, __u32 type,
 	if (rc == 0) {
 		rc = server->ops->mand_lock(xid, cfile, flock->fl_start, length,
 			type | server->vals->shared_lock_type, 0, 1, false);
-		flock->fl_type = F_RDLCK;
+		flock->c.flc_type = F_RDLCK;
 		if (rc != 0)
 			cifs_dbg(VFS, "Error unlocking previously locked range %d during test of lock\n",
 				 rc);
 	} else
-		flock->fl_type = F_WRLCK;
+		flock->c.flc_type = F_WRLCK;
 
 	return 0;
 }
@@ -2022,7 +2189,7 @@ cifs_setlk(struct file *file, struct file_lock *flock, __u32 type,
 			posix_lock_type = CIFS_UNLCK;
 
 		rc = CIFSSMBPosixLock(xid, tcon, cfile->fid.netfid,
-				      hash_lockowner(flock->fl_owner),
+				      hash_lockowner(flock->c.flc_owner),
 				      flock->fl_start, length,
 				      NULL, posix_lock_type, wait_flag);
 		goto out;
@@ -2032,7 +2199,7 @@ cifs_setlk(struct file *file, struct file_lock *flock, __u32 type,
 		struct cifsLockInfo *lock;
 
 		lock = cifs_lock_init(flock->fl_start, length, type,
-				      flock->fl_flags);
+				      flock->c.flc_flags);
 		if (!lock)
 			return -ENOMEM;
 
@@ -2071,7 +2238,7 @@ cifs_setlk(struct file *file, struct file_lock *flock, __u32 type,
 		rc = server->ops->mand_unlock_range(cfile, flock, xid);
 
 out:
-	if ((flock->fl_flags & FL_POSIX) || (flock->fl_flags & FL_FLOCK)) {
+	if ((flock->c.flc_flags & FL_POSIX) || (flock->c.flc_flags & FL_FLOCK)) {
 		/*
 		 * If this is a request to remove all locks because we
 		 * are closing the file, it doesn't matter if the
@@ -2080,7 +2247,7 @@ out:
 		 */
 		if (rc) {
 			cifs_dbg(VFS, "%s failed rc=%d\n", __func__, rc);
-			if (!(flock->fl_flags & FL_CLOSE))
+			if (!(flock->c.flc_flags & FL_CLOSE))
 				return rc;
 		}
 		rc = locks_lock_file_wait(file, flock);
@@ -2101,7 +2268,7 @@ int cifs_flock(struct file *file, int cmd, struct file_lock *fl)
 
 	xid = get_xid();
 
-	if (!(fl->fl_flags & FL_FLOCK)) {
+	if (!(fl->c.flc_flags & FL_FLOCK)) {
 		rc = -ENOLCK;
 		free_xid(xid);
 		return rc;
@@ -2152,7 +2319,8 @@ int cifs_lock(struct file *file, int cmd, struct file_lock *flock)
 	xid = get_xid();
 
 	cifs_dbg(FYI, "%s: %pD2 cmd=0x%x type=0x%x flags=0x%x r=%lld:%lld\n", __func__, file, cmd,
-		 flock->fl_flags, flock->fl_type, (long long)flock->fl_start,
+		 flock->c.flc_flags, flock->c.flc_type,
+		 (long long)flock->fl_start,
 		 (long long)flock->fl_end);
 
 	cfile = (struct cifsFileInfo *)file->private_data;
@@ -2192,102 +2360,25 @@ int cifs_lock(struct file *file, int cmd, struct file_lock *flock)
 	return rc;
 }
 
-/*
- * update the file size (if needed) after a write. Should be called with
- * the inode->i_lock held
- */
-void
-cifs_update_eof(struct cifsInodeInfo *cifsi, loff_t offset,
-		      unsigned int bytes_written)
-{
-	loff_t end_of_write = offset + bytes_written;
-
-	if (end_of_write > cifsi->netfs.remote_i_size)
-		netfs_resize_file(&cifsi->netfs, end_of_write, true);
-}
-
-static ssize_t
-cifs_write(struct cifsFileInfo *open_file, __u32 pid, const char *write_data,
-	   size_t write_size, loff_t *offset)
+void cifs_write_subrequest_terminated(struct cifs_io_subrequest *wdata, ssize_t result,
+				      bool was_async)
 {
-	int rc = 0;
-	unsigned int bytes_written = 0;
-	unsigned int total_written;
-	struct cifs_tcon *tcon;
-	struct TCP_Server_Info *server;
-	unsigned int xid;
-	struct dentry *dentry = open_file->dentry;
-	struct cifsInodeInfo *cifsi = CIFS_I(d_inode(dentry));
-	struct cifs_io_parms io_parms = {0};
-
-	cifs_dbg(FYI, "write %zd bytes to offset %lld of %pd\n",
-		 write_size, *offset, dentry);
-
-	tcon = tlink_tcon(open_file->tlink);
-	server = tcon->ses->server;
-
-	if (!server->ops->sync_write)
-		return -ENOSYS;
+	struct netfs_io_request *wreq = wdata->rreq;
+	struct netfs_inode *ictx = netfs_inode(wreq->inode);
+	loff_t wrend;
 
-	xid = get_xid();
-
-	for (total_written = 0; write_size > total_written;
-	     total_written += bytes_written) {
-		rc = -EAGAIN;
-		while (rc == -EAGAIN) {
-			struct kvec iov[2];
-			unsigned int len;
-
-			if (open_file->invalidHandle) {
-				/* we could deadlock if we called
-				   filemap_fdatawait from here so tell
-				   reopen_file not to flush data to
-				   server now */
-				rc = cifs_reopen_file(open_file, false);
-				if (rc != 0)
-					break;
-			}
+	if (result > 0) {
+		wrend = wdata->subreq.start + wdata->subreq.transferred + result;
 
-			len = min(server->ops->wp_retry_size(d_inode(dentry)),
-				  (unsigned int)write_size - total_written);
-			/* iov[0] is reserved for smb header */
-			iov[1].iov_base = (char *)write_data + total_written;
-			iov[1].iov_len = len;
-			io_parms.pid = pid;
-			io_parms.tcon = tcon;
-			io_parms.offset = *offset;
-			io_parms.length = len;
-			rc = server->ops->sync_write(xid, &open_file->fid,
-					&io_parms, &bytes_written, iov, 1);
-		}
-		if (rc || (bytes_written == 0)) {
-			if (total_written)
-				break;
-			else {
-				free_xid(xid);
-				return rc;
-			}
-		} else {
-			spin_lock(&d_inode(dentry)->i_lock);
-			cifs_update_eof(cifsi, *offset, bytes_written);
-			spin_unlock(&d_inode(dentry)->i_lock);
-			*offset += bytes_written;
-		}
+		if (wrend > ictx->zero_point &&
+		    (wdata->rreq->origin == NETFS_UNBUFFERED_WRITE ||
+		     wdata->rreq->origin == NETFS_DIO_WRITE))
+			ictx->zero_point = wrend;
+		if (wrend > ictx->remote_i_size)
+			netfs_resize_file(ictx, wrend, true);
 	}
 
-	cifs_stats_bytes_written(tcon, total_written);
-
-	if (total_written > 0) {
-		spin_lock(&d_inode(dentry)->i_lock);
-		if (*offset > d_inode(dentry)->i_size) {
-			i_size_write(d_inode(dentry), *offset);
-			d_inode(dentry)->i_blocks = (512 - 1 + *offset) >> 9;
-		}
-		spin_unlock(&d_inode(dentry)->i_lock);
-	}
-	mark_inode_dirty_sync(d_inode(dentry));
-	free_xid(xid);
-	return total_written;
+	netfs_write_subrequest_terminated(&wdata->subreq, result, was_async);
 }
 
 struct cifsFileInfo *find_readable_file(struct cifsInodeInfo *cifs_inode,
@@ -2494,730 +2585,9 @@ cifs_get_readable_path(struct cifs_tcon *tcon, const char *name,
 	return -ENOENT;
 }
 
-void
-cifs_writedata_release(struct kref *refcount)
-{
-	struct cifs_writedata *wdata = container_of(refcount,
-					struct cifs_writedata, refcount);
-#ifdef CONFIG_CIFS_SMB_DIRECT
-	if (wdata->mr) {
-		smbd_deregister_mr(wdata->mr);
-		wdata->mr = NULL;
-	}
-#endif
-
-	if (wdata->cfile)
-		cifsFileInfo_put(wdata->cfile);
-
-	kfree(wdata);
-}
-
-/*
- * Write failed with a retryable error. Resend the write request. It's also
- * possible that the page was redirtied so re-clean the page.
- */
-static void
-cifs_writev_requeue(struct cifs_writedata *wdata)
-{
-	int rc = 0;
-	struct inode *inode = d_inode(wdata->cfile->dentry);
-	struct TCP_Server_Info *server;
-	unsigned int rest_len = wdata->bytes;
-	loff_t fpos = wdata->offset;
-
-	server = tlink_tcon(wdata->cfile->tlink)->ses->server;
-	do {
-		struct cifs_writedata *wdata2;
-		unsigned int wsize, cur_len;
-
-		wsize = server->ops->wp_retry_size(inode);
-		if (wsize < rest_len) {
-			if (wsize < PAGE_SIZE) {
-				rc = -EOPNOTSUPP;
-				break;
-			}
-			cur_len = min(round_down(wsize, PAGE_SIZE), rest_len);
-		} else {
-			cur_len = rest_len;
-		}
-
-		wdata2 = cifs_writedata_alloc(cifs_writev_complete);
-		if (!wdata2) {
-			rc = -ENOMEM;
-			break;
-		}
-
-		wdata2->sync_mode = wdata->sync_mode;
-		wdata2->offset	= fpos;
-		wdata2->bytes	= cur_len;
-		wdata2->iter	= wdata->iter;
-
-		iov_iter_advance(&wdata2->iter, fpos - wdata->offset);
-		iov_iter_truncate(&wdata2->iter, wdata2->bytes);
-
-		if (iov_iter_is_xarray(&wdata2->iter))
-			/* Check for pages having been redirtied and clean
-			 * them.  We can do this by walking the xarray.  If
-			 * it's not an xarray, then it's a DIO and we shouldn't
-			 * be mucking around with the page bits.
-			 */
-			cifs_undirty_folios(inode, fpos, cur_len);
-
-		rc = cifs_get_writable_file(CIFS_I(inode), FIND_WR_ANY,
-					    &wdata2->cfile);
-		if (!wdata2->cfile) {
-			cifs_dbg(VFS, "No writable handle to retry writepages rc=%d\n",
-				 rc);
-			if (!is_retryable_error(rc))
-				rc = -EBADF;
-		} else {
-			wdata2->pid = wdata2->cfile->pid;
-			rc = server->ops->async_writev(wdata2,
-						       cifs_writedata_release);
-		}
-
-		kref_put(&wdata2->refcount, cifs_writedata_release);
-		if (rc) {
-			if (is_retryable_error(rc))
-				continue;
-			fpos += cur_len;
-			rest_len -= cur_len;
-			break;
-		}
-
-		fpos += cur_len;
-		rest_len -= cur_len;
-	} while (rest_len > 0);
-
-	/* Clean up remaining pages from the original wdata */
-	if (iov_iter_is_xarray(&wdata->iter))
-		cifs_pages_write_failed(inode, fpos, rest_len);
-
-	if (rc != 0 && !is_retryable_error(rc))
-		mapping_set_error(inode->i_mapping, rc);
-	kref_put(&wdata->refcount, cifs_writedata_release);
-}
-
-void
-cifs_writev_complete(struct work_struct *work)
-{
-	struct cifs_writedata *wdata = container_of(work,
-						struct cifs_writedata, work);
-	struct inode *inode = d_inode(wdata->cfile->dentry);
-
-	if (wdata->result == 0) {
-		spin_lock(&inode->i_lock);
-		cifs_update_eof(CIFS_I(inode), wdata->offset, wdata->bytes);
-		spin_unlock(&inode->i_lock);
-		cifs_stats_bytes_written(tlink_tcon(wdata->cfile->tlink),
-					 wdata->bytes);
-	} else if (wdata->sync_mode == WB_SYNC_ALL && wdata->result == -EAGAIN)
-		return cifs_writev_requeue(wdata);
-
-	if (wdata->result == -EAGAIN)
-		cifs_pages_write_redirty(inode, wdata->offset, wdata->bytes);
-	else if (wdata->result < 0)
-		cifs_pages_write_failed(inode, wdata->offset, wdata->bytes);
-	else
-		cifs_pages_written_back(inode, wdata->offset, wdata->bytes);
-
-	if (wdata->result != -EAGAIN)
-		mapping_set_error(inode->i_mapping, wdata->result);
-	kref_put(&wdata->refcount, cifs_writedata_release);
-}
-
-struct cifs_writedata *cifs_writedata_alloc(work_func_t complete)
-{
-	struct cifs_writedata *wdata;
-
-	wdata = kzalloc(sizeof(*wdata), GFP_NOFS);
-	if (wdata != NULL) {
-		kref_init(&wdata->refcount);
-		INIT_LIST_HEAD(&wdata->list);
-		init_completion(&wdata->done);
-		INIT_WORK(&wdata->work, complete);
-	}
-	return wdata;
-}
-
-static int cifs_partialpagewrite(struct page *page, unsigned from, unsigned to)
-{
-	struct address_space *mapping = page->mapping;
-	loff_t offset = (loff_t)page->index << PAGE_SHIFT;
-	char *write_data;
-	int rc = -EFAULT;
-	int bytes_written = 0;
-	struct inode *inode;
-	struct cifsFileInfo *open_file;
-
-	if (!mapping || !mapping->host)
-		return -EFAULT;
-
-	inode = page->mapping->host;
-
-	offset += (loff_t)from;
-	write_data = kmap(page);
-	write_data += from;
-
-	if ((to > PAGE_SIZE) || (from > to)) {
-		kunmap(page);
-		return -EIO;
-	}
-
-	/* racing with truncate? */
-	if (offset > mapping->host->i_size) {
-		kunmap(page);
-		return 0; /* don't care */
-	}
-
-	/* check to make sure that we are not extending the file */
-	if (mapping->host->i_size - offset < (loff_t)to)
-		to = (unsigned)(mapping->host->i_size - offset);
-
-	rc = cifs_get_writable_file(CIFS_I(mapping->host), FIND_WR_ANY,
-				    &open_file);
-	if (!rc) {
-		bytes_written = cifs_write(open_file, open_file->pid,
-					   write_data, to - from, &offset);
-		cifsFileInfo_put(open_file);
-		/* Does mm or vfs already set times? */
-		simple_inode_init_ts(inode);
-		if ((bytes_written > 0) && (offset))
-			rc = 0;
-		else if (bytes_written < 0)
-			rc = bytes_written;
-		else
-			rc = -EFAULT;
-	} else {
-		cifs_dbg(FYI, "No writable handle for write page rc=%d\n", rc);
-		if (!is_retryable_error(rc))
-			rc = -EIO;
-	}
-
-	kunmap(page);
-	return rc;
-}
-
 /*
- * Extend the region to be written back to include subsequent contiguously
- * dirty pages if possible, but don't sleep while doing so.
+ * Flush data on a strict file.
  */
-static void cifs_extend_writeback(struct address_space *mapping,
-				  struct xa_state *xas,
-				  long *_count,
-				  loff_t start,
-				  int max_pages,
-				  loff_t max_len,
-				  size_t *_len)
-{
-	struct folio_batch batch;
-	struct folio *folio;
-	unsigned int nr_pages;
-	pgoff_t index = (start + *_len) / PAGE_SIZE;
-	size_t len;
-	bool stop = true;
-	unsigned int i;
-
-	folio_batch_init(&batch);
-
-	do {
-		/* Firstly, we gather up a batch of contiguous dirty pages
-		 * under the RCU read lock - but we can't clear the dirty flags
-		 * there if any of those pages are mapped.
-		 */
-		rcu_read_lock();
-
-		xas_for_each(xas, folio, ULONG_MAX) {
-			stop = true;
-			if (xas_retry(xas, folio))
-				continue;
-			if (xa_is_value(folio))
-				break;
-			if (folio->index != index) {
-				xas_reset(xas);
-				break;
-			}
-
-			if (!folio_try_get_rcu(folio)) {
-				xas_reset(xas);
-				continue;
-			}
-			nr_pages = folio_nr_pages(folio);
-			if (nr_pages > max_pages) {
-				xas_reset(xas);
-				break;
-			}
-
-			/* Has the page moved or been split? */
-			if (unlikely(folio != xas_reload(xas))) {
-				folio_put(folio);
-				xas_reset(xas);
-				break;
-			}
-
-			if (!folio_trylock(folio)) {
-				folio_put(folio);
-				xas_reset(xas);
-				break;
-			}
-			if (!folio_test_dirty(folio) ||
-			    folio_test_writeback(folio)) {
-				folio_unlock(folio);
-				folio_put(folio);
-				xas_reset(xas);
-				break;
-			}
-
-			max_pages -= nr_pages;
-			len = folio_size(folio);
-			stop = false;
-
-			index += nr_pages;
-			*_count -= nr_pages;
-			*_len += len;
-			if (max_pages <= 0 || *_len >= max_len || *_count <= 0)
-				stop = true;
-
-			if (!folio_batch_add(&batch, folio))
-				break;
-			if (stop)
-				break;
-		}
-
-		xas_pause(xas);
-		rcu_read_unlock();
-
-		/* Now, if we obtained any pages, we can shift them to being
-		 * writable and mark them for caching.
-		 */
-		if (!folio_batch_count(&batch))
-			break;
-
-		for (i = 0; i < folio_batch_count(&batch); i++) {
-			folio = batch.folios[i];
-			/* The folio should be locked, dirty and not undergoing
-			 * writeback from the loop above.
-			 */
-			if (!folio_clear_dirty_for_io(folio))
-				WARN_ON(1);
-			folio_start_writeback(folio);
-			folio_unlock(folio);
-		}
-
-		folio_batch_release(&batch);
-		cond_resched();
-	} while (!stop);
-}
-
-/*
- * Write back the locked page and any subsequent non-locked dirty pages.
- */
-static ssize_t cifs_write_back_from_locked_folio(struct address_space *mapping,
-						 struct writeback_control *wbc,
-						 struct xa_state *xas,
-						 struct folio *folio,
-						 unsigned long long start,
-						 unsigned long long end)
-{
-	struct inode *inode = mapping->host;
-	struct TCP_Server_Info *server;
-	struct cifs_writedata *wdata;
-	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
-	struct cifs_credits credits_on_stack;
-	struct cifs_credits *credits = &credits_on_stack;
-	struct cifsFileInfo *cfile = NULL;
-	unsigned long long i_size = i_size_read(inode), max_len;
-	unsigned int xid, wsize;
-	size_t len = folio_size(folio);
-	long count = wbc->nr_to_write;
-	int rc;
-
-	/* The folio should be locked, dirty and not undergoing writeback. */
-	if (!folio_clear_dirty_for_io(folio))
-		WARN_ON_ONCE(1);
-	folio_start_writeback(folio);
-
-	count -= folio_nr_pages(folio);
-
-	xid = get_xid();
-	server = cifs_pick_channel(cifs_sb_master_tcon(cifs_sb)->ses);
-
-	rc = cifs_get_writable_file(CIFS_I(inode), FIND_WR_ANY, &cfile);
-	if (rc) {
-		cifs_dbg(VFS, "No writable handle in writepages rc=%d\n", rc);
-		goto err_xid;
-	}
-
-	rc = server->ops->wait_mtu_credits(server, cifs_sb->ctx->wsize,
-					   &wsize, credits);
-	if (rc != 0)
-		goto err_close;
-
-	wdata = cifs_writedata_alloc(cifs_writev_complete);
-	if (!wdata) {
-		rc = -ENOMEM;
-		goto err_uncredit;
-	}
-
-	wdata->sync_mode = wbc->sync_mode;
-	wdata->offset = folio_pos(folio);
-	wdata->pid = cfile->pid;
-	wdata->credits = credits_on_stack;
-	wdata->cfile = cfile;
-	wdata->server = server;
-	cfile = NULL;
-
-	/* Find all consecutive lockable dirty pages that have contiguous
-	 * written regions, stopping when we find a page that is not
-	 * immediately lockable, is not dirty or is missing, or we reach the
-	 * end of the range.
-	 */
-	if (start < i_size) {
-		/* Trim the write to the EOF; the extra data is ignored.  Also
-		 * put an upper limit on the size of a single storedata op.
-		 */
-		max_len = wsize;
-		max_len = min_t(unsigned long long, max_len, end - start + 1);
-		max_len = min_t(unsigned long long, max_len, i_size - start);
-
-		if (len < max_len) {
-			int max_pages = INT_MAX;
-
-#ifdef CONFIG_CIFS_SMB_DIRECT
-			if (server->smbd_conn)
-				max_pages = server->smbd_conn->max_frmr_depth;
-#endif
-			max_pages -= folio_nr_pages(folio);
-
-			if (max_pages > 0)
-				cifs_extend_writeback(mapping, xas, &count, start,
-						      max_pages, max_len, &len);
-		}
-	}
-	len = min_t(unsigned long long, len, i_size - start);
-
-	/* We now have a contiguous set of dirty pages, each with writeback
-	 * set; the first page is still locked at this point, but all the rest
-	 * have been unlocked.
-	 */
-	folio_unlock(folio);
-	wdata->bytes = len;
-
-	if (start < i_size) {
-		iov_iter_xarray(&wdata->iter, ITER_SOURCE, &mapping->i_pages,
-				start, len);
-
-		rc = adjust_credits(wdata->server, &wdata->credits, wdata->bytes);
-		if (rc)
-			goto err_wdata;
-
-		if (wdata->cfile->invalidHandle)
-			rc = -EAGAIN;
-		else
-			rc = wdata->server->ops->async_writev(wdata,
-							      cifs_writedata_release);
-		if (rc >= 0) {
-			kref_put(&wdata->refcount, cifs_writedata_release);
-			goto err_close;
-		}
-	} else {
-		/* The dirty region was entirely beyond the EOF. */
-		cifs_pages_written_back(inode, start, len);
-		rc = 0;
-	}
-
-err_wdata:
-	kref_put(&wdata->refcount, cifs_writedata_release);
-err_uncredit:
-	add_credits_and_wake_if(server, credits, 0);
-err_close:
-	if (cfile)
-		cifsFileInfo_put(cfile);
-err_xid:
-	free_xid(xid);
-	if (rc == 0) {
-		wbc->nr_to_write = count;
-		rc = len;
-	} else if (is_retryable_error(rc)) {
-		cifs_pages_write_redirty(inode, start, len);
-	} else {
-		cifs_pages_write_failed(inode, start, len);
-		mapping_set_error(mapping, rc);
-	}
-	/* Indication to update ctime and mtime as close is deferred */
-	set_bit(CIFS_INO_MODIFIED_ATTR, &CIFS_I(inode)->flags);
-	return rc;
-}
-
-/*
- * write a region of pages back to the server
- */
-static ssize_t cifs_writepages_begin(struct address_space *mapping,
-				     struct writeback_control *wbc,
-				     struct xa_state *xas,
-				     unsigned long long *_start,
-				     unsigned long long end)
-{
-	struct folio *folio;
-	unsigned long long start = *_start;
-	ssize_t ret;
-	int skips = 0;
-
-search_again:
-	/* Find the first dirty page. */
-	rcu_read_lock();
-
-	for (;;) {
-		folio = xas_find_marked(xas, end / PAGE_SIZE, PAGECACHE_TAG_DIRTY);
-		if (xas_retry(xas, folio) || xa_is_value(folio))
-			continue;
-		if (!folio)
-			break;
-
-		if (!folio_try_get_rcu(folio)) {
-			xas_reset(xas);
-			continue;
-		}
-
-		if (unlikely(folio != xas_reload(xas))) {
-			folio_put(folio);
-			xas_reset(xas);
-			continue;
-		}
-
-		xas_pause(xas);
-		break;
-	}
-	rcu_read_unlock();
-	if (!folio)
-		return 0;
-
-	start = folio_pos(folio); /* May regress with THPs */
-
-	/* At this point we hold neither the i_pages lock nor the page lock:
-	 * the page may be truncated or invalidated (changing page->mapping to
-	 * NULL), or even swizzled back from swapper_space to tmpfs file
-	 * mapping
-	 */
-lock_again:
-	if (wbc->sync_mode != WB_SYNC_NONE) {
-		ret = folio_lock_killable(folio);
-		if (ret < 0)
-			return ret;
-	} else {
-		if (!folio_trylock(folio))
-			goto search_again;
-	}
-
-	if (folio->mapping != mapping ||
-	    !folio_test_dirty(folio)) {
-		start += folio_size(folio);
-		folio_unlock(folio);
-		goto search_again;
-	}
-
-	if (folio_test_writeback(folio) ||
-	    folio_test_fscache(folio)) {
-		folio_unlock(folio);
-		if (wbc->sync_mode != WB_SYNC_NONE) {
-			folio_wait_writeback(folio);
-#ifdef CONFIG_CIFS_FSCACHE
-			folio_wait_fscache(folio);
-#endif
-			goto lock_again;
-		}
-
-		start += folio_size(folio);
-		if (wbc->sync_mode == WB_SYNC_NONE) {
-			if (skips >= 5 || need_resched()) {
-				ret = 0;
-				goto out;
-			}
-			skips++;
-		}
-		goto search_again;
-	}
-
-	ret = cifs_write_back_from_locked_folio(mapping, wbc, xas, folio, start, end);
-out:
-	if (ret > 0)
-		*_start = start + ret;
-	return ret;
-}
-
-/*
- * Write a region of pages back to the server
- */
-static int cifs_writepages_region(struct address_space *mapping,
-				  struct writeback_control *wbc,
-				  unsigned long long *_start,
-				  unsigned long long end)
-{
-	ssize_t ret;
-
-	XA_STATE(xas, &mapping->i_pages, *_start / PAGE_SIZE);
-
-	do {
-		ret = cifs_writepages_begin(mapping, wbc, &xas, _start, end);
-		if (ret > 0 && wbc->nr_to_write > 0)
-			cond_resched();
-	} while (ret > 0 && wbc->nr_to_write > 0);
-
-	return ret > 0 ? 0 : ret;
-}
-
-/*
- * Write some of the pending data back to the server
- */
-static int cifs_writepages(struct address_space *mapping,
-			   struct writeback_control *wbc)
-{
-	loff_t start, end;
-	int ret;
-
-	/* We have to be careful as we can end up racing with setattr()
-	 * truncating the pagecache since the caller doesn't take a lock here
-	 * to prevent it.
-	 */
-
-	if (wbc->range_cyclic && mapping->writeback_index) {
-		start = mapping->writeback_index * PAGE_SIZE;
-		ret = cifs_writepages_region(mapping, wbc, &start, LLONG_MAX);
-		if (ret < 0)
-			goto out;
-
-		if (wbc->nr_to_write <= 0) {
-			mapping->writeback_index = start / PAGE_SIZE;
-			goto out;
-		}
-
-		start = 0;
-		end = mapping->writeback_index * PAGE_SIZE;
-		mapping->writeback_index = 0;
-		ret = cifs_writepages_region(mapping, wbc, &start, end);
-		if (ret == 0)
-			mapping->writeback_index = start / PAGE_SIZE;
-	} else if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX) {
-		start = 0;
-		ret = cifs_writepages_region(mapping, wbc, &start, LLONG_MAX);
-		if (wbc->nr_to_write > 0 && ret == 0)
-			mapping->writeback_index = start / PAGE_SIZE;
-	} else {
-		start = wbc->range_start;
-		ret = cifs_writepages_region(mapping, wbc, &start, wbc->range_end);
-	}
-
-out:
-	return ret;
-}
-
-static int
-cifs_writepage_locked(struct page *page, struct writeback_control *wbc)
-{
-	int rc;
-	unsigned int xid;
-
-	xid = get_xid();
-/* BB add check for wbc flags */
-	get_page(page);
-	if (!PageUptodate(page))
-		cifs_dbg(FYI, "ppw - page not up to date\n");
-
-	/*
-	 * Set the "writeback" flag, and clear "dirty" in the radix tree.
-	 *
-	 * A writepage() implementation always needs to do either this,
-	 * or re-dirty the page with "redirty_page_for_writepage()" in
-	 * the case of a failure.
-	 *
-	 * Just unlocking the page will cause the radix tree tag-bits
-	 * to fail to update with the state of the page correctly.
-	 */
-	set_page_writeback(page);
-retry_write:
-	rc = cifs_partialpagewrite(page, 0, PAGE_SIZE);
-	if (is_retryable_error(rc)) {
-		if (wbc->sync_mode == WB_SYNC_ALL && rc == -EAGAIN)
-			goto retry_write;
-		redirty_page_for_writepage(wbc, page);
-	} else if (rc != 0) {
-		SetPageError(page);
-		mapping_set_error(page->mapping, rc);
-	} else {
-		SetPageUptodate(page);
-	}
-	end_page_writeback(page);
-	put_page(page);
-	free_xid(xid);
-	return rc;
-}
-
-static int cifs_write_end(struct file *file, struct address_space *mapping,
-			loff_t pos, unsigned len, unsigned copied,
-			struct page *page, void *fsdata)
-{
-	int rc;
-	struct inode *inode = mapping->host;
-	struct cifsFileInfo *cfile = file->private_data;
-	struct cifs_sb_info *cifs_sb = CIFS_SB(cfile->dentry->d_sb);
-	struct folio *folio = page_folio(page);
-	__u32 pid;
-
-	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_RWPIDFORWARD)
-		pid = cfile->pid;
-	else
-		pid = current->tgid;
-
-	cifs_dbg(FYI, "write_end for page %p from pos %lld with %d bytes\n",
-		 page, pos, copied);
-
-	if (folio_test_checked(folio)) {
-		if (copied == len)
-			folio_mark_uptodate(folio);
-		folio_clear_checked(folio);
-	} else if (!folio_test_uptodate(folio) && copied == PAGE_SIZE)
-		folio_mark_uptodate(folio);
-
-	if (!folio_test_uptodate(folio)) {
-		char *page_data;
-		unsigned offset = pos & (PAGE_SIZE - 1);
-		unsigned int xid;
-
-		xid = get_xid();
-		/* this is probably better than directly calling
-		   partialpage_write since in this function the file handle is
-		   known which we might as well	leverage */
-		/* BB check if anything else missing out of ppw
-		   such as updating last write time */
-		page_data = kmap(page);
-		rc = cifs_write(cfile, pid, page_data + offset, copied, &pos);
-		/* if (rc < 0) should we set writebehind rc? */
-		kunmap(page);
-
-		free_xid(xid);
-	} else {
-		rc = copied;
-		pos += copied;
-		set_page_dirty(page);
-	}
-
-	if (rc > 0) {
-		spin_lock(&inode->i_lock);
-		if (pos > inode->i_size) {
-			i_size_write(inode, pos);
-			inode->i_blocks = (512 - 1 + pos) >> 9;
-		}
-		spin_unlock(&inode->i_lock);
-	}
-
-	unlock_page(page);
-	put_page(page);
-	/* Indication to update ctime and mtime as close is deferred */
-	set_bit(CIFS_INO_MODIFIED_ATTR, &CIFS_I(inode)->flags);
-
-	return rc;
-}
-
 int cifs_strict_fsync(struct file *file, loff_t start, loff_t end,
 		      int datasync)
 {
@@ -3272,6 +2642,9 @@ strict_fsync_exit:
 	return rc;
 }
 
+/*
+ * Flush data on a non-strict data.
+ */
 int cifs_fsync(struct file *file, loff_t start, loff_t end, int datasync)
 {
 	unsigned int xid;
@@ -3338,481 +2711,6 @@ int cifs_flush(struct file *file, fl_owner_t id)
 	return rc;
 }
 
-static void
-cifs_uncached_writedata_release(struct kref *refcount)
-{
-	struct cifs_writedata *wdata = container_of(refcount,
-					struct cifs_writedata, refcount);
-
-	kref_put(&wdata->ctx->refcount, cifs_aio_ctx_release);
-	cifs_writedata_release(refcount);
-}
-
-static void collect_uncached_write_data(struct cifs_aio_ctx *ctx);
-
-static void
-cifs_uncached_writev_complete(struct work_struct *work)
-{
-	struct cifs_writedata *wdata = container_of(work,
-					struct cifs_writedata, work);
-	struct inode *inode = d_inode(wdata->cfile->dentry);
-	struct cifsInodeInfo *cifsi = CIFS_I(inode);
-
-	spin_lock(&inode->i_lock);
-	cifs_update_eof(cifsi, wdata->offset, wdata->bytes);
-	if (cifsi->netfs.remote_i_size > inode->i_size)
-		i_size_write(inode, cifsi->netfs.remote_i_size);
-	spin_unlock(&inode->i_lock);
-
-	complete(&wdata->done);
-	collect_uncached_write_data(wdata->ctx);
-	/* the below call can possibly free the last ref to aio ctx */
-	kref_put(&wdata->refcount, cifs_uncached_writedata_release);
-}
-
-static int
-cifs_resend_wdata(struct cifs_writedata *wdata, struct list_head *wdata_list,
-	struct cifs_aio_ctx *ctx)
-{
-	unsigned int wsize;
-	struct cifs_credits credits;
-	int rc;
-	struct TCP_Server_Info *server = wdata->server;
-
-	do {
-		if (wdata->cfile->invalidHandle) {
-			rc = cifs_reopen_file(wdata->cfile, false);
-			if (rc == -EAGAIN)
-				continue;
-			else if (rc)
-				break;
-		}
-
-
-		/*
-		 * Wait for credits to resend this wdata.
-		 * Note: we are attempting to resend the whole wdata not in
-		 * segments
-		 */
-		do {
-			rc = server->ops->wait_mtu_credits(server, wdata->bytes,
-						&wsize, &credits);
-			if (rc)
-				goto fail;
-
-			if (wsize < wdata->bytes) {
-				add_credits_and_wake_if(server, &credits, 0);
-				msleep(1000);
-			}
-		} while (wsize < wdata->bytes);
-		wdata->credits = credits;
-
-		rc = adjust_credits(server, &wdata->credits, wdata->bytes);
-
-		if (!rc) {
-			if (wdata->cfile->invalidHandle)
-				rc = -EAGAIN;
-			else {
-				wdata->replay = true;
-#ifdef CONFIG_CIFS_SMB_DIRECT
-				if (wdata->mr) {
-					wdata->mr->need_invalidate = true;
-					smbd_deregister_mr(wdata->mr);
-					wdata->mr = NULL;
-				}
-#endif
-				rc = server->ops->async_writev(wdata,
-					cifs_uncached_writedata_release);
-			}
-		}
-
-		/* If the write was successfully sent, we are done */
-		if (!rc) {
-			list_add_tail(&wdata->list, wdata_list);
-			return 0;
-		}
-
-		/* Roll back credits and retry if needed */
-		add_credits_and_wake_if(server, &wdata->credits, 0);
-	} while (rc == -EAGAIN);
-
-fail:
-	kref_put(&wdata->refcount, cifs_uncached_writedata_release);
-	return rc;
-}
-
-/*
- * Select span of a bvec iterator we're going to use.  Limit it by both maximum
- * size and maximum number of segments.
- */
-static size_t cifs_limit_bvec_subset(const struct iov_iter *iter, size_t max_size,
-				     size_t max_segs, unsigned int *_nsegs)
-{
-	const struct bio_vec *bvecs = iter->bvec;
-	unsigned int nbv = iter->nr_segs, ix = 0, nsegs = 0;
-	size_t len, span = 0, n = iter->count;
-	size_t skip = iter->iov_offset;
-
-	if (WARN_ON(!iov_iter_is_bvec(iter)) || n == 0)
-		return 0;
-
-	while (n && ix < nbv && skip) {
-		len = bvecs[ix].bv_len;
-		if (skip < len)
-			break;
-		skip -= len;
-		n -= len;
-		ix++;
-	}
-
-	while (n && ix < nbv) {
-		len = min3(n, bvecs[ix].bv_len - skip, max_size);
-		span += len;
-		max_size -= len;
-		nsegs++;
-		ix++;
-		if (max_size == 0 || nsegs >= max_segs)
-			break;
-		skip = 0;
-		n -= len;
-	}
-
-	*_nsegs = nsegs;
-	return span;
-}
-
-static int
-cifs_write_from_iter(loff_t fpos, size_t len, struct iov_iter *from,
-		     struct cifsFileInfo *open_file,
-		     struct cifs_sb_info *cifs_sb, struct list_head *wdata_list,
-		     struct cifs_aio_ctx *ctx)
-{
-	int rc = 0;
-	size_t cur_len, max_len;
-	struct cifs_writedata *wdata;
-	pid_t pid;
-	struct TCP_Server_Info *server;
-	unsigned int xid, max_segs = INT_MAX;
-
-	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_RWPIDFORWARD)
-		pid = open_file->pid;
-	else
-		pid = current->tgid;
-
-	server = cifs_pick_channel(tlink_tcon(open_file->tlink)->ses);
-	xid = get_xid();
-
-#ifdef CONFIG_CIFS_SMB_DIRECT
-	if (server->smbd_conn)
-		max_segs = server->smbd_conn->max_frmr_depth;
-#endif
-
-	do {
-		struct cifs_credits credits_on_stack;
-		struct cifs_credits *credits = &credits_on_stack;
-		unsigned int wsize, nsegs = 0;
-
-		if (signal_pending(current)) {
-			rc = -EINTR;
-			break;
-		}
-
-		if (open_file->invalidHandle) {
-			rc = cifs_reopen_file(open_file, false);
-			if (rc == -EAGAIN)
-				continue;
-			else if (rc)
-				break;
-		}
-
-		rc = server->ops->wait_mtu_credits(server, cifs_sb->ctx->wsize,
-						   &wsize, credits);
-		if (rc)
-			break;
-
-		max_len = min_t(const size_t, len, wsize);
-		if (!max_len) {
-			rc = -EAGAIN;
-			add_credits_and_wake_if(server, credits, 0);
-			break;
-		}
-
-		cur_len = cifs_limit_bvec_subset(from, max_len, max_segs, &nsegs);
-		cifs_dbg(FYI, "write_from_iter len=%zx/%zx nsegs=%u/%lu/%u\n",
-			 cur_len, max_len, nsegs, from->nr_segs, max_segs);
-		if (cur_len == 0) {
-			rc = -EIO;
-			add_credits_and_wake_if(server, credits, 0);
-			break;
-		}
-
-		wdata = cifs_writedata_alloc(cifs_uncached_writev_complete);
-		if (!wdata) {
-			rc = -ENOMEM;
-			add_credits_and_wake_if(server, credits, 0);
-			break;
-		}
-
-		wdata->sync_mode = WB_SYNC_ALL;
-		wdata->offset	= (__u64)fpos;
-		wdata->cfile	= cifsFileInfo_get(open_file);
-		wdata->server	= server;
-		wdata->pid	= pid;
-		wdata->bytes	= cur_len;
-		wdata->credits	= credits_on_stack;
-		wdata->iter	= *from;
-		wdata->ctx	= ctx;
-		kref_get(&ctx->refcount);
-
-		iov_iter_truncate(&wdata->iter, cur_len);
-
-		rc = adjust_credits(server, &wdata->credits, wdata->bytes);
-
-		if (!rc) {
-			if (wdata->cfile->invalidHandle)
-				rc = -EAGAIN;
-			else
-				rc = server->ops->async_writev(wdata,
-					cifs_uncached_writedata_release);
-		}
-
-		if (rc) {
-			add_credits_and_wake_if(server, &wdata->credits, 0);
-			kref_put(&wdata->refcount,
-				 cifs_uncached_writedata_release);
-			if (rc == -EAGAIN)
-				continue;
-			break;
-		}
-
-		list_add_tail(&wdata->list, wdata_list);
-		iov_iter_advance(from, cur_len);
-		fpos += cur_len;
-		len -= cur_len;
-	} while (len > 0);
-
-	free_xid(xid);
-	return rc;
-}
-
-static void collect_uncached_write_data(struct cifs_aio_ctx *ctx)
-{
-	struct cifs_writedata *wdata, *tmp;
-	struct cifs_tcon *tcon;
-	struct cifs_sb_info *cifs_sb;
-	struct dentry *dentry = ctx->cfile->dentry;
-	ssize_t rc;
-
-	tcon = tlink_tcon(ctx->cfile->tlink);
-	cifs_sb = CIFS_SB(dentry->d_sb);
-
-	mutex_lock(&ctx->aio_mutex);
-
-	if (list_empty(&ctx->list)) {
-		mutex_unlock(&ctx->aio_mutex);
-		return;
-	}
-
-	rc = ctx->rc;
-	/*
-	 * Wait for and collect replies for any successful sends in order of
-	 * increasing offset. Once an error is hit, then return without waiting
-	 * for any more replies.
-	 */
-restart_loop:
-	list_for_each_entry_safe(wdata, tmp, &ctx->list, list) {
-		if (!rc) {
-			if (!try_wait_for_completion(&wdata->done)) {
-				mutex_unlock(&ctx->aio_mutex);
-				return;
-			}
-
-			if (wdata->result)
-				rc = wdata->result;
-			else
-				ctx->total_len += wdata->bytes;
-
-			/* resend call if it's a retryable error */
-			if (rc == -EAGAIN) {
-				struct list_head tmp_list;
-				struct iov_iter tmp_from = ctx->iter;
-
-				INIT_LIST_HEAD(&tmp_list);
-				list_del_init(&wdata->list);
-
-				if (ctx->direct_io)
-					rc = cifs_resend_wdata(
-						wdata, &tmp_list, ctx);
-				else {
-					iov_iter_advance(&tmp_from,
-						 wdata->offset - ctx->pos);
-
-					rc = cifs_write_from_iter(wdata->offset,
-						wdata->bytes, &tmp_from,
-						ctx->cfile, cifs_sb, &tmp_list,
-						ctx);
-
-					kref_put(&wdata->refcount,
-						cifs_uncached_writedata_release);
-				}
-
-				list_splice(&tmp_list, &ctx->list);
-				goto restart_loop;
-			}
-		}
-		list_del_init(&wdata->list);
-		kref_put(&wdata->refcount, cifs_uncached_writedata_release);
-	}
-
-	cifs_stats_bytes_written(tcon, ctx->total_len);
-	set_bit(CIFS_INO_INVALID_MAPPING, &CIFS_I(dentry->d_inode)->flags);
-
-	ctx->rc = (rc == 0) ? ctx->total_len : rc;
-
-	mutex_unlock(&ctx->aio_mutex);
-
-	if (ctx->iocb && ctx->iocb->ki_complete)
-		ctx->iocb->ki_complete(ctx->iocb, ctx->rc);
-	else
-		complete(&ctx->done);
-}
-
-static ssize_t __cifs_writev(
-	struct kiocb *iocb, struct iov_iter *from, bool direct)
-{
-	struct file *file = iocb->ki_filp;
-	ssize_t total_written = 0;
-	struct cifsFileInfo *cfile;
-	struct cifs_tcon *tcon;
-	struct cifs_sb_info *cifs_sb;
-	struct cifs_aio_ctx *ctx;
-	int rc;
-
-	rc = generic_write_checks(iocb, from);
-	if (rc <= 0)
-		return rc;
-
-	cifs_sb = CIFS_FILE_SB(file);
-	cfile = file->private_data;
-	tcon = tlink_tcon(cfile->tlink);
-
-	if (!tcon->ses->server->ops->async_writev)
-		return -ENOSYS;
-
-	ctx = cifs_aio_ctx_alloc();
-	if (!ctx)
-		return -ENOMEM;
-
-	ctx->cfile = cifsFileInfo_get(cfile);
-
-	if (!is_sync_kiocb(iocb))
-		ctx->iocb = iocb;
-
-	ctx->pos = iocb->ki_pos;
-	ctx->direct_io = direct;
-	ctx->nr_pinned_pages = 0;
-
-	if (user_backed_iter(from)) {
-		/*
-		 * Extract IOVEC/UBUF-type iterators to a BVEC-type iterator as
-		 * they contain references to the calling process's virtual
-		 * memory layout which won't be available in an async worker
-		 * thread.  This also takes a pin on every folio involved.
-		 */
-		rc = netfs_extract_user_iter(from, iov_iter_count(from),
-					     &ctx->iter, 0);
-		if (rc < 0) {
-			kref_put(&ctx->refcount, cifs_aio_ctx_release);
-			return rc;
-		}
-
-		ctx->nr_pinned_pages = rc;
-		ctx->bv = (void *)ctx->iter.bvec;
-		ctx->bv_need_unpin = iov_iter_extract_will_pin(from);
-	} else if ((iov_iter_is_bvec(from) || iov_iter_is_kvec(from)) &&
-		   !is_sync_kiocb(iocb)) {
-		/*
-		 * If the op is asynchronous, we need to copy the list attached
-		 * to a BVEC/KVEC-type iterator, but we assume that the storage
-		 * will be pinned by the caller; in any case, we may or may not
-		 * be able to pin the pages, so we don't try.
-		 */
-		ctx->bv = (void *)dup_iter(&ctx->iter, from, GFP_KERNEL);
-		if (!ctx->bv) {
-			kref_put(&ctx->refcount, cifs_aio_ctx_release);
-			return -ENOMEM;
-		}
-	} else {
-		/*
-		 * Otherwise, we just pass the iterator down as-is and rely on
-		 * the caller to make sure the pages referred to by the
-		 * iterator don't evaporate.
-		 */
-		ctx->iter = *from;
-	}
-
-	ctx->len = iov_iter_count(&ctx->iter);
-
-	/* grab a lock here due to read response handlers can access ctx */
-	mutex_lock(&ctx->aio_mutex);
-
-	rc = cifs_write_from_iter(iocb->ki_pos, ctx->len, &ctx->iter,
-				  cfile, cifs_sb, &ctx->list, ctx);
-
-	/*
-	 * If at least one write was successfully sent, then discard any rc
-	 * value from the later writes. If the other write succeeds, then
-	 * we'll end up returning whatever was written. If it fails, then
-	 * we'll get a new rc value from that.
-	 */
-	if (!list_empty(&ctx->list))
-		rc = 0;
-
-	mutex_unlock(&ctx->aio_mutex);
-
-	if (rc) {
-		kref_put(&ctx->refcount, cifs_aio_ctx_release);
-		return rc;
-	}
-
-	if (!is_sync_kiocb(iocb)) {
-		kref_put(&ctx->refcount, cifs_aio_ctx_release);
-		return -EIOCBQUEUED;
-	}
-
-	rc = wait_for_completion_killable(&ctx->done);
-	if (rc) {
-		mutex_lock(&ctx->aio_mutex);
-		ctx->rc = rc = -EINTR;
-		total_written = ctx->total_len;
-		mutex_unlock(&ctx->aio_mutex);
-	} else {
-		rc = ctx->rc;
-		total_written = ctx->total_len;
-	}
-
-	kref_put(&ctx->refcount, cifs_aio_ctx_release);
-
-	if (unlikely(!total_written))
-		return rc;
-
-	iocb->ki_pos += total_written;
-	return total_written;
-}
-
-ssize_t cifs_direct_writev(struct kiocb *iocb, struct iov_iter *from)
-{
-	struct file *file = iocb->ki_filp;
-
-	cifs_revalidate_mapping(file->f_inode);
-	return __cifs_writev(iocb, from, true);
-}
-
-ssize_t cifs_user_writev(struct kiocb *iocb, struct iov_iter *from)
-{
-	return __cifs_writev(iocb, from, false);
-}
-
 static ssize_t
 cifs_writev(struct kiocb *iocb, struct iov_iter *from)
 {
@@ -3823,7 +2721,10 @@ cifs_writev(struct kiocb *iocb, struct iov_iter *from)
 	struct TCP_Server_Info *server = tlink_tcon(cfile->tlink)->ses->server;
 	ssize_t rc;
 
-	inode_lock(inode);
+	rc = netfs_start_io_write(inode);
+	if (rc < 0)
+		return rc;
+
 	/*
 	 * We need to hold the sem to be sure nobody modifies lock list
 	 * with a brlock that prevents writing.
@@ -3837,13 +2738,12 @@ cifs_writev(struct kiocb *iocb, struct iov_iter *from)
 	if (!cifs_find_lock_conflict(cfile, iocb->ki_pos, iov_iter_count(from),
 				     server->vals->exclusive_lock_type, 0,
 				     NULL, CIFS_WRITE_OP))
-		rc = __generic_file_write_iter(iocb, from);
+		rc = netfs_buffered_write_iter_locked(iocb, from, NULL);
 	else
 		rc = -EACCES;
 out:
 	up_read(&cinode->lock_sem);
-	inode_unlock(inode);
-
+	netfs_end_io_write(inode);
 	if (rc > 0)
 		rc = generic_write_sync(iocb, rc);
 	return rc;
@@ -3866,9 +2766,9 @@ cifs_strict_writev(struct kiocb *iocb, struct iov_iter *from)
 
 	if (CIFS_CACHE_WRITE(cinode)) {
 		if (cap_unix(tcon->ses) &&
-		(CIFS_UNIX_FCNTL_CAP & le64_to_cpu(tcon->fsUnixInfo.Capability))
-		  && ((cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NOPOSIXBRL) == 0)) {
-			written = generic_file_write_iter(iocb, from);
+		    (CIFS_UNIX_FCNTL_CAP & le64_to_cpu(tcon->fsUnixInfo.Capability)) &&
+		    ((cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NOPOSIXBRL) == 0)) {
+			written = netfs_file_write_iter(iocb, from);
 			goto out;
 		}
 		written = cifs_writev(iocb, from);
@@ -3880,7 +2780,7 @@ cifs_strict_writev(struct kiocb *iocb, struct iov_iter *from)
 	 * affected pages because it may cause a error with mandatory locks on
 	 * these pages but not on the region from pos to ppos+len-1.
 	 */
-	written = cifs_user_writev(iocb, from);
+	written = netfs_file_write_iter(iocb, from);
 	if (CIFS_CACHE_READ(cinode)) {
 		/*
 		 * We have read level caching and we have just sent a write
@@ -3899,449 +2799,55 @@ out:
 	return written;
 }
 
-static struct cifs_readdata *cifs_readdata_alloc(work_func_t complete)
-{
-	struct cifs_readdata *rdata;
-
-	rdata = kzalloc(sizeof(*rdata), GFP_KERNEL);
-	if (rdata) {
-		kref_init(&rdata->refcount);
-		INIT_LIST_HEAD(&rdata->list);
-		init_completion(&rdata->done);
-		INIT_WORK(&rdata->work, complete);
-	}
-
-	return rdata;
-}
-
-void
-cifs_readdata_release(struct kref *refcount)
-{
-	struct cifs_readdata *rdata = container_of(refcount,
-					struct cifs_readdata, refcount);
-
-	if (rdata->ctx)
-		kref_put(&rdata->ctx->refcount, cifs_aio_ctx_release);
-#ifdef CONFIG_CIFS_SMB_DIRECT
-	if (rdata->mr) {
-		smbd_deregister_mr(rdata->mr);
-		rdata->mr = NULL;
-	}
-#endif
-	if (rdata->cfile)
-		cifsFileInfo_put(rdata->cfile);
-
-	kfree(rdata);
-}
-
-static void collect_uncached_read_data(struct cifs_aio_ctx *ctx);
-
-static void
-cifs_uncached_readv_complete(struct work_struct *work)
-{
-	struct cifs_readdata *rdata = container_of(work,
-						struct cifs_readdata, work);
-
-	complete(&rdata->done);
-	collect_uncached_read_data(rdata->ctx);
-	/* the below call can possibly free the last ref to aio ctx */
-	kref_put(&rdata->refcount, cifs_readdata_release);
-}
-
-static int cifs_resend_rdata(struct cifs_readdata *rdata,
-			struct list_head *rdata_list,
-			struct cifs_aio_ctx *ctx)
-{
-	unsigned int rsize;
-	struct cifs_credits credits;
-	int rc;
-	struct TCP_Server_Info *server;
-
-	/* XXX: should we pick a new channel here? */
-	server = rdata->server;
-
-	do {
-		if (rdata->cfile->invalidHandle) {
-			rc = cifs_reopen_file(rdata->cfile, true);
-			if (rc == -EAGAIN)
-				continue;
-			else if (rc)
-				break;
-		}
-
-		/*
-		 * Wait for credits to resend this rdata.
-		 * Note: we are attempting to resend the whole rdata not in
-		 * segments
-		 */
-		do {
-			rc = server->ops->wait_mtu_credits(server, rdata->bytes,
-						&rsize, &credits);
-
-			if (rc)
-				goto fail;
-
-			if (rsize < rdata->bytes) {
-				add_credits_and_wake_if(server, &credits, 0);
-				msleep(1000);
-			}
-		} while (rsize < rdata->bytes);
-		rdata->credits = credits;
-
-		rc = adjust_credits(server, &rdata->credits, rdata->bytes);
-		if (!rc) {
-			if (rdata->cfile->invalidHandle)
-				rc = -EAGAIN;
-			else {
-#ifdef CONFIG_CIFS_SMB_DIRECT
-				if (rdata->mr) {
-					rdata->mr->need_invalidate = true;
-					smbd_deregister_mr(rdata->mr);
-					rdata->mr = NULL;
-				}
-#endif
-				rc = server->ops->async_readv(rdata);
-			}
-		}
-
-		/* If the read was successfully sent, we are done */
-		if (!rc) {
-			/* Add to aio pending list */
-			list_add_tail(&rdata->list, rdata_list);
-			return 0;
-		}
-
-		/* Roll back credits and retry if needed */
-		add_credits_and_wake_if(server, &rdata->credits, 0);
-	} while (rc == -EAGAIN);
-
-fail:
-	kref_put(&rdata->refcount, cifs_readdata_release);
-	return rc;
-}
-
-static int
-cifs_send_async_read(loff_t fpos, size_t len, struct cifsFileInfo *open_file,
-		     struct cifs_sb_info *cifs_sb, struct list_head *rdata_list,
-		     struct cifs_aio_ctx *ctx)
+ssize_t cifs_loose_read_iter(struct kiocb *iocb, struct iov_iter *iter)
 {
-	struct cifs_readdata *rdata;
-	unsigned int rsize, nsegs, max_segs = INT_MAX;
-	struct cifs_credits credits_on_stack;
-	struct cifs_credits *credits = &credits_on_stack;
-	size_t cur_len, max_len;
-	int rc;
-	pid_t pid;
-	struct TCP_Server_Info *server;
-
-	server = cifs_pick_channel(tlink_tcon(open_file->tlink)->ses);
-
-#ifdef CONFIG_CIFS_SMB_DIRECT
-	if (server->smbd_conn)
-		max_segs = server->smbd_conn->max_frmr_depth;
-#endif
-
-	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_RWPIDFORWARD)
-		pid = open_file->pid;
-	else
-		pid = current->tgid;
-
-	do {
-		if (open_file->invalidHandle) {
-			rc = cifs_reopen_file(open_file, true);
-			if (rc == -EAGAIN)
-				continue;
-			else if (rc)
-				break;
-		}
-
-		if (cifs_sb->ctx->rsize == 0)
-			cifs_sb->ctx->rsize =
-				server->ops->negotiate_rsize(tlink_tcon(open_file->tlink),
-							     cifs_sb->ctx);
-
-		rc = server->ops->wait_mtu_credits(server, cifs_sb->ctx->rsize,
-						   &rsize, credits);
-		if (rc)
-			break;
-
-		max_len = min_t(size_t, len, rsize);
-
-		cur_len = cifs_limit_bvec_subset(&ctx->iter, max_len,
-						 max_segs, &nsegs);
-		cifs_dbg(FYI, "read-to-iter len=%zx/%zx nsegs=%u/%lu/%u\n",
-			 cur_len, max_len, nsegs, ctx->iter.nr_segs, max_segs);
-		if (cur_len == 0) {
-			rc = -EIO;
-			add_credits_and_wake_if(server, credits, 0);
-			break;
-		}
-
-		rdata = cifs_readdata_alloc(cifs_uncached_readv_complete);
-		if (!rdata) {
-			add_credits_and_wake_if(server, credits, 0);
-			rc = -ENOMEM;
-			break;
-		}
-
-		rdata->server	= server;
-		rdata->cfile	= cifsFileInfo_get(open_file);
-		rdata->offset	= fpos;
-		rdata->bytes	= cur_len;
-		rdata->pid	= pid;
-		rdata->credits	= credits_on_stack;
-		rdata->ctx	= ctx;
-		kref_get(&ctx->refcount);
-
-		rdata->iter	= ctx->iter;
-		iov_iter_truncate(&rdata->iter, cur_len);
-
-		rc = adjust_credits(server, &rdata->credits, rdata->bytes);
-
-		if (!rc) {
-			if (rdata->cfile->invalidHandle)
-				rc = -EAGAIN;
-			else
-				rc = server->ops->async_readv(rdata);
-		}
+	ssize_t rc;
+	struct inode *inode = file_inode(iocb->ki_filp);
 
-		if (rc) {
-			add_credits_and_wake_if(server, &rdata->credits, 0);
-			kref_put(&rdata->refcount, cifs_readdata_release);
-			if (rc == -EAGAIN)
-				continue;
-			break;
-		}
+	if (iocb->ki_flags & IOCB_DIRECT)
+		return netfs_unbuffered_read_iter(iocb, iter);
 
-		list_add_tail(&rdata->list, rdata_list);
-		iov_iter_advance(&ctx->iter, cur_len);
-		fpos += cur_len;
-		len -= cur_len;
-	} while (len > 0);
+	rc = cifs_revalidate_mapping(inode);
+	if (rc)
+		return rc;
 
-	return rc;
+	return netfs_file_read_iter(iocb, iter);
 }
 
-static void
-collect_uncached_read_data(struct cifs_aio_ctx *ctx)
+ssize_t cifs_file_write_iter(struct kiocb *iocb, struct iov_iter *from)
 {
-	struct cifs_readdata *rdata, *tmp;
-	struct cifs_sb_info *cifs_sb;
+	struct inode *inode = file_inode(iocb->ki_filp);
+	struct cifsInodeInfo *cinode = CIFS_I(inode);
+	ssize_t written;
 	int rc;
 
-	cifs_sb = CIFS_SB(ctx->cfile->dentry->d_sb);
-
-	mutex_lock(&ctx->aio_mutex);
-
-	if (list_empty(&ctx->list)) {
-		mutex_unlock(&ctx->aio_mutex);
-		return;
-	}
-
-	rc = ctx->rc;
-	/* the loop below should proceed in the order of increasing offsets */
-again:
-	list_for_each_entry_safe(rdata, tmp, &ctx->list, list) {
-		if (!rc) {
-			if (!try_wait_for_completion(&rdata->done)) {
-				mutex_unlock(&ctx->aio_mutex);
-				return;
-			}
-
-			if (rdata->result == -EAGAIN) {
-				/* resend call if it's a retryable error */
-				struct list_head tmp_list;
-				unsigned int got_bytes = rdata->got_bytes;
-
-				list_del_init(&rdata->list);
-				INIT_LIST_HEAD(&tmp_list);
-
-				if (ctx->direct_io) {
-					/*
-					 * Re-use rdata as this is a
-					 * direct I/O
-					 */
-					rc = cifs_resend_rdata(
-						rdata,
-						&tmp_list, ctx);
-				} else {
-					rc = cifs_send_async_read(
-						rdata->offset + got_bytes,
-						rdata->bytes - got_bytes,
-						rdata->cfile, cifs_sb,
-						&tmp_list, ctx);
-
-					kref_put(&rdata->refcount,
-						cifs_readdata_release);
-				}
-
-				list_splice(&tmp_list, &ctx->list);
-
-				goto again;
-			} else if (rdata->result)
-				rc = rdata->result;
-
-			/* if there was a short read -- discard anything left */
-			if (rdata->got_bytes && rdata->got_bytes < rdata->bytes)
-				rc = -ENODATA;
-
-			ctx->total_len += rdata->got_bytes;
-		}
-		list_del_init(&rdata->list);
-		kref_put(&rdata->refcount, cifs_readdata_release);
-	}
-
-	/* mask nodata case */
-	if (rc == -ENODATA)
-		rc = 0;
-
-	ctx->rc = (rc == 0) ? (ssize_t)ctx->total_len : rc;
-
-	mutex_unlock(&ctx->aio_mutex);
-
-	if (ctx->iocb && ctx->iocb->ki_complete)
-		ctx->iocb->ki_complete(ctx->iocb, ctx->rc);
-	else
-		complete(&ctx->done);
-}
-
-static ssize_t __cifs_readv(
-	struct kiocb *iocb, struct iov_iter *to, bool direct)
-{
-	size_t len;
-	struct file *file = iocb->ki_filp;
-	struct cifs_sb_info *cifs_sb;
-	struct cifsFileInfo *cfile;
-	struct cifs_tcon *tcon;
-	ssize_t rc, total_read = 0;
-	loff_t offset = iocb->ki_pos;
-	struct cifs_aio_ctx *ctx;
-
-	len = iov_iter_count(to);
-	if (!len)
-		return 0;
-
-	cifs_sb = CIFS_FILE_SB(file);
-	cfile = file->private_data;
-	tcon = tlink_tcon(cfile->tlink);
-
-	if (!tcon->ses->server->ops->async_readv)
-		return -ENOSYS;
-
-	if ((file->f_flags & O_ACCMODE) == O_WRONLY)
-		cifs_dbg(FYI, "attempting read on write only file instance\n");
-
-	ctx = cifs_aio_ctx_alloc();
-	if (!ctx)
-		return -ENOMEM;
-
-	ctx->pos	= offset;
-	ctx->direct_io	= direct;
-	ctx->len	= len;
-	ctx->cfile	= cifsFileInfo_get(cfile);
-	ctx->nr_pinned_pages = 0;
-
-	if (!is_sync_kiocb(iocb))
-		ctx->iocb = iocb;
-
-	if (user_backed_iter(to)) {
-		/*
-		 * Extract IOVEC/UBUF-type iterators to a BVEC-type iterator as
-		 * they contain references to the calling process's virtual
-		 * memory layout which won't be available in an async worker
-		 * thread.  This also takes a pin on every folio involved.
-		 */
-		rc = netfs_extract_user_iter(to, iov_iter_count(to),
-					     &ctx->iter, 0);
-		if (rc < 0) {
-			kref_put(&ctx->refcount, cifs_aio_ctx_release);
-			return rc;
-		}
-
-		ctx->nr_pinned_pages = rc;
-		ctx->bv = (void *)ctx->iter.bvec;
-		ctx->bv_need_unpin = iov_iter_extract_will_pin(to);
-		ctx->should_dirty = true;
-	} else if ((iov_iter_is_bvec(to) || iov_iter_is_kvec(to)) &&
-		   !is_sync_kiocb(iocb)) {
-		/*
-		 * If the op is asynchronous, we need to copy the list attached
-		 * to a BVEC/KVEC-type iterator, but we assume that the storage
-		 * will be retained by the caller; in any case, we may or may
-		 * not be able to pin the pages, so we don't try.
-		 */
-		ctx->bv = (void *)dup_iter(&ctx->iter, to, GFP_KERNEL);
-		if (!ctx->bv) {
-			kref_put(&ctx->refcount, cifs_aio_ctx_release);
-			return -ENOMEM;
-		}
-	} else {
-		/*
-		 * Otherwise, we just pass the iterator down as-is and rely on
-		 * the caller to make sure the pages referred to by the
-		 * iterator don't evaporate.
-		 */
-		ctx->iter = *to;
-	}
-
-	if (direct) {
-		rc = filemap_write_and_wait_range(file->f_inode->i_mapping,
-						  offset, offset + len - 1);
-		if (rc) {
-			kref_put(&ctx->refcount, cifs_aio_ctx_release);
-			return -EAGAIN;
+	if (iocb->ki_filp->f_flags & O_DIRECT) {
+		written = netfs_unbuffered_write_iter(iocb, from);
+		if (written > 0 && CIFS_CACHE_READ(cinode)) {
+			cifs_zap_mapping(inode);
+			cifs_dbg(FYI,
+				 "Set no oplock for inode=%p after a write operation\n",
+				 inode);
+			cinode->oplock = 0;
 		}
+		return written;
 	}
 
-	/* grab a lock here due to read response handlers can access ctx */
-	mutex_lock(&ctx->aio_mutex);
-
-	rc = cifs_send_async_read(offset, len, cfile, cifs_sb, &ctx->list, ctx);
-
-	/* if at least one read request send succeeded, then reset rc */
-	if (!list_empty(&ctx->list))
-		rc = 0;
-
-	mutex_unlock(&ctx->aio_mutex);
-
-	if (rc) {
-		kref_put(&ctx->refcount, cifs_aio_ctx_release);
-		return rc;
-	}
-
-	if (!is_sync_kiocb(iocb)) {
-		kref_put(&ctx->refcount, cifs_aio_ctx_release);
-		return -EIOCBQUEUED;
-	}
-
-	rc = wait_for_completion_killable(&ctx->done);
-	if (rc) {
-		mutex_lock(&ctx->aio_mutex);
-		ctx->rc = rc = -EINTR;
-		total_read = ctx->total_len;
-		mutex_unlock(&ctx->aio_mutex);
-	} else {
-		rc = ctx->rc;
-		total_read = ctx->total_len;
-	}
+	written = cifs_get_writer(cinode);
+	if (written)
+		return written;
 
-	kref_put(&ctx->refcount, cifs_aio_ctx_release);
+	written = netfs_file_write_iter(iocb, from);
 
-	if (total_read) {
-		iocb->ki_pos += total_read;
-		return total_read;
+	if (!CIFS_CACHE_WRITE(CIFS_I(inode))) {
+		rc = filemap_fdatawrite(inode->i_mapping);
+		if (rc)
+			cifs_dbg(FYI, "cifs_file_write_iter: %d rc on %p inode\n",
+				 rc, inode);
 	}
-	return rc;
-}
-
-ssize_t cifs_direct_readv(struct kiocb *iocb, struct iov_iter *to)
-{
-	return __cifs_readv(iocb, to, true);
-}
 
-ssize_t cifs_user_readv(struct kiocb *iocb, struct iov_iter *to)
-{
-	return __cifs_readv(iocb, to, false);
+	cifs_put_writer(cinode);
+	return written;
 }
 
 ssize_t
@@ -4364,140 +2870,54 @@ cifs_strict_readv(struct kiocb *iocb, struct iov_iter *to)
 	 * pos+len-1.
 	 */
 	if (!CIFS_CACHE_READ(cinode))
-		return cifs_user_readv(iocb, to);
+		return netfs_unbuffered_read_iter(iocb, to);
 
 	if (cap_unix(tcon->ses) &&
 	    (CIFS_UNIX_FCNTL_CAP & le64_to_cpu(tcon->fsUnixInfo.Capability)) &&
-	    ((cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NOPOSIXBRL) == 0))
-		return generic_file_read_iter(iocb, to);
+	    ((cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NOPOSIXBRL) == 0)) {
+		if (iocb->ki_flags & IOCB_DIRECT)
+			return netfs_unbuffered_read_iter(iocb, to);
+		return netfs_buffered_read_iter(iocb, to);
+	}
 
 	/*
 	 * We need to hold the sem to be sure nobody modifies lock list
 	 * with a brlock that prevents reading.
 	 */
-	down_read(&cinode->lock_sem);
-	if (!cifs_find_lock_conflict(cfile, iocb->ki_pos, iov_iter_count(to),
-				     tcon->ses->server->vals->shared_lock_type,
-				     0, NULL, CIFS_READ_OP))
-		rc = generic_file_read_iter(iocb, to);
-	up_read(&cinode->lock_sem);
-	return rc;
-}
-
-static ssize_t
-cifs_read(struct file *file, char *read_data, size_t read_size, loff_t *offset)
-{
-	int rc = -EACCES;
-	unsigned int bytes_read = 0;
-	unsigned int total_read;
-	unsigned int current_read_size;
-	unsigned int rsize;
-	struct cifs_sb_info *cifs_sb;
-	struct cifs_tcon *tcon;
-	struct TCP_Server_Info *server;
-	unsigned int xid;
-	char *cur_offset;
-	struct cifsFileInfo *open_file;
-	struct cifs_io_parms io_parms = {0};
-	int buf_type = CIFS_NO_BUFFER;
-	__u32 pid;
-
-	xid = get_xid();
-	cifs_sb = CIFS_FILE_SB(file);
-
-	/* FIXME: set up handlers for larger reads and/or convert to async */
-	rsize = min_t(unsigned int, cifs_sb->ctx->rsize, CIFSMaxBufSize);
-
-	if (file->private_data == NULL) {
-		rc = -EBADF;
-		free_xid(xid);
-		return rc;
-	}
-	open_file = file->private_data;
-	tcon = tlink_tcon(open_file->tlink);
-	server = cifs_pick_channel(tcon->ses);
-
-	if (!server->ops->sync_read) {
-		free_xid(xid);
-		return -ENOSYS;
-	}
-
-	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_RWPIDFORWARD)
-		pid = open_file->pid;
-	else
-		pid = current->tgid;
-
-	if ((file->f_flags & O_ACCMODE) == O_WRONLY)
-		cifs_dbg(FYI, "attempting read on write only file instance\n");
-
-	for (total_read = 0, cur_offset = read_data; read_size > total_read;
-	     total_read += bytes_read, cur_offset += bytes_read) {
-		do {
-			current_read_size = min_t(uint, read_size - total_read,
-						  rsize);
-			/*
-			 * For windows me and 9x we do not want to request more
-			 * than it negotiated since it will refuse the read
-			 * then.
-			 */
-			if (!(tcon->ses->capabilities &
-				tcon->ses->server->vals->cap_large_files)) {
-				current_read_size = min_t(uint,
-					current_read_size, CIFSMaxBufSize);
-			}
-			if (open_file->invalidHandle) {
-				rc = cifs_reopen_file(open_file, true);
-				if (rc != 0)
-					break;
-			}
-			io_parms.pid = pid;
-			io_parms.tcon = tcon;
-			io_parms.offset = *offset;
-			io_parms.length = current_read_size;
-			io_parms.server = server;
-			rc = server->ops->sync_read(xid, &open_file->fid, &io_parms,
-						    &bytes_read, &cur_offset,
-						    &buf_type);
-		} while (rc == -EAGAIN);
-
-		if (rc || (bytes_read == 0)) {
-			if (total_read) {
-				break;
-			} else {
-				free_xid(xid);
-				return rc;
-			}
-		} else {
-			cifs_stats_bytes_read(tcon, total_read);
-			*offset += bytes_read;
-		}
+	if (iocb->ki_flags & IOCB_DIRECT) {
+		rc = netfs_start_io_direct(inode);
+		if (rc < 0)
+			goto out;
+		rc = -EACCES;
+		down_read(&cinode->lock_sem);
+		if (!cifs_find_lock_conflict(
+			    cfile, iocb->ki_pos, iov_iter_count(to),
+			    tcon->ses->server->vals->shared_lock_type,
+			    0, NULL, CIFS_READ_OP))
+			rc = netfs_unbuffered_read_iter_locked(iocb, to);
+		up_read(&cinode->lock_sem);
+		netfs_end_io_direct(inode);
+	} else {
+		rc = netfs_start_io_read(inode);
+		if (rc < 0)
+			goto out;
+		rc = -EACCES;
+		down_read(&cinode->lock_sem);
+		if (!cifs_find_lock_conflict(
+			    cfile, iocb->ki_pos, iov_iter_count(to),
+			    tcon->ses->server->vals->shared_lock_type,
+			    0, NULL, CIFS_READ_OP))
+			rc = filemap_read(iocb, to, 0);
+		up_read(&cinode->lock_sem);
+		netfs_end_io_read(inode);
 	}
-	free_xid(xid);
-	return total_read;
+out:
+	return rc;
 }
 
-/*
- * If the page is mmap'ed into a process' page tables, then we need to make
- * sure that it doesn't change while being written back.
- */
 static vm_fault_t cifs_page_mkwrite(struct vm_fault *vmf)
 {
-	struct folio *folio = page_folio(vmf->page);
-
-	/* Wait for the folio to be written to the cache before we allow it to
-	 * be modified.  We then assume the entire folio will need writing back.
-	 */
-#ifdef CONFIG_CIFS_FSCACHE
-	if (folio_test_fscache(folio) &&
-	    folio_wait_fscache_killable(folio) < 0)
-		return VM_FAULT_RETRY;
-#endif
-
-	folio_wait_writeback(folio);
-
-	if (folio_lock_killable(folio) < 0)
-		return VM_FAULT_RETRY;
-	return VM_FAULT_LOCKED;
+	return netfs_page_mkwrite(vmf, NULL);
 }
 
 static const struct vm_operations_struct cifs_file_vm_ops = {
@@ -4543,290 +2963,6 @@ int cifs_file_mmap(struct file *file, struct vm_area_struct *vma)
 	return rc;
 }
 
-/*
- * Unlock a bunch of folios in the pagecache.
- */
-static void cifs_unlock_folios(struct address_space *mapping, pgoff_t first, pgoff_t last)
-{
-	struct folio *folio;
-	XA_STATE(xas, &mapping->i_pages, first);
-
-	rcu_read_lock();
-	xas_for_each(&xas, folio, last) {
-		folio_unlock(folio);
-	}
-	rcu_read_unlock();
-}
-
-static void cifs_readahead_complete(struct work_struct *work)
-{
-	struct cifs_readdata *rdata = container_of(work,
-						   struct cifs_readdata, work);
-	struct folio *folio;
-	pgoff_t last;
-	bool good = rdata->result == 0 || (rdata->result == -EAGAIN && rdata->got_bytes);
-
-	XA_STATE(xas, &rdata->mapping->i_pages, rdata->offset / PAGE_SIZE);
-
-	if (good)
-		cifs_readahead_to_fscache(rdata->mapping->host,
-					  rdata->offset, rdata->bytes);
-
-	if (iov_iter_count(&rdata->iter) > 0)
-		iov_iter_zero(iov_iter_count(&rdata->iter), &rdata->iter);
-
-	last = (rdata->offset + rdata->bytes - 1) / PAGE_SIZE;
-
-	rcu_read_lock();
-	xas_for_each(&xas, folio, last) {
-		if (good) {
-			flush_dcache_folio(folio);
-			folio_mark_uptodate(folio);
-		}
-		folio_unlock(folio);
-	}
-	rcu_read_unlock();
-
-	kref_put(&rdata->refcount, cifs_readdata_release);
-}
-
-static void cifs_readahead(struct readahead_control *ractl)
-{
-	struct cifsFileInfo *open_file = ractl->file->private_data;
-	struct cifs_sb_info *cifs_sb = CIFS_FILE_SB(ractl->file);
-	struct TCP_Server_Info *server;
-	unsigned int xid, nr_pages, cache_nr_pages = 0;
-	unsigned int ra_pages;
-	pgoff_t next_cached = ULONG_MAX, ra_index;
-	bool caching = fscache_cookie_enabled(cifs_inode_cookie(ractl->mapping->host)) &&
-		cifs_inode_cookie(ractl->mapping->host)->cache_priv;
-	bool check_cache = caching;
-	pid_t pid;
-	int rc = 0;
-
-	/* Note that readahead_count() lags behind our dequeuing of pages from
-	 * the ractl, wo we have to keep track for ourselves.
-	 */
-	ra_pages = readahead_count(ractl);
-	ra_index = readahead_index(ractl);
-
-	xid = get_xid();
-
-	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_RWPIDFORWARD)
-		pid = open_file->pid;
-	else
-		pid = current->tgid;
-
-	server = cifs_pick_channel(tlink_tcon(open_file->tlink)->ses);
-
-	cifs_dbg(FYI, "%s: file=%p mapping=%p num_pages=%u\n",
-		 __func__, ractl->file, ractl->mapping, ra_pages);
-
-	/*
-	 * Chop the readahead request up into rsize-sized read requests.
-	 */
-	while ((nr_pages = ra_pages)) {
-		unsigned int i, rsize;
-		struct cifs_readdata *rdata;
-		struct cifs_credits credits_on_stack;
-		struct cifs_credits *credits = &credits_on_stack;
-		struct folio *folio;
-		pgoff_t fsize;
-
-		/*
-		 * Find out if we have anything cached in the range of
-		 * interest, and if so, where the next chunk of cached data is.
-		 */
-		if (caching) {
-			if (check_cache) {
-				rc = cifs_fscache_query_occupancy(
-					ractl->mapping->host, ra_index, nr_pages,
-					&next_cached, &cache_nr_pages);
-				if (rc < 0)
-					caching = false;
-				check_cache = false;
-			}
-
-			if (ra_index == next_cached) {
-				/*
-				 * TODO: Send a whole batch of pages to be read
-				 * by the cache.
-				 */
-				folio = readahead_folio(ractl);
-				fsize = folio_nr_pages(folio);
-				ra_pages -= fsize;
-				ra_index += fsize;
-				if (cifs_readpage_from_fscache(ractl->mapping->host,
-							       &folio->page) < 0) {
-					/*
-					 * TODO: Deal with cache read failure
-					 * here, but for the moment, delegate
-					 * that to readpage.
-					 */
-					caching = false;
-				}
-				folio_unlock(folio);
-				next_cached += fsize;
-				cache_nr_pages -= fsize;
-				if (cache_nr_pages == 0)
-					check_cache = true;
-				continue;
-			}
-		}
-
-		if (open_file->invalidHandle) {
-			rc = cifs_reopen_file(open_file, true);
-			if (rc) {
-				if (rc == -EAGAIN)
-					continue;
-				break;
-			}
-		}
-
-		if (cifs_sb->ctx->rsize == 0)
-			cifs_sb->ctx->rsize =
-				server->ops->negotiate_rsize(tlink_tcon(open_file->tlink),
-							     cifs_sb->ctx);
-
-		rc = server->ops->wait_mtu_credits(server, cifs_sb->ctx->rsize,
-						   &rsize, credits);
-		if (rc)
-			break;
-		nr_pages = min_t(size_t, rsize / PAGE_SIZE, ra_pages);
-		if (next_cached != ULONG_MAX)
-			nr_pages = min_t(size_t, nr_pages, next_cached - ra_index);
-
-		/*
-		 * Give up immediately if rsize is too small to read an entire
-		 * page. The VFS will fall back to readpage. We should never
-		 * reach this point however since we set ra_pages to 0 when the
-		 * rsize is smaller than a cache page.
-		 */
-		if (unlikely(!nr_pages)) {
-			add_credits_and_wake_if(server, credits, 0);
-			break;
-		}
-
-		rdata = cifs_readdata_alloc(cifs_readahead_complete);
-		if (!rdata) {
-			/* best to give up if we're out of mem */
-			add_credits_and_wake_if(server, credits, 0);
-			break;
-		}
-
-		rdata->offset	= ra_index * PAGE_SIZE;
-		rdata->bytes	= nr_pages * PAGE_SIZE;
-		rdata->cfile	= cifsFileInfo_get(open_file);
-		rdata->server	= server;
-		rdata->mapping	= ractl->mapping;
-		rdata->pid	= pid;
-		rdata->credits	= credits_on_stack;
-
-		for (i = 0; i < nr_pages; i++) {
-			if (!readahead_folio(ractl))
-				WARN_ON(1);
-		}
-		ra_pages -= nr_pages;
-		ra_index += nr_pages;
-
-		iov_iter_xarray(&rdata->iter, ITER_DEST, &rdata->mapping->i_pages,
-				rdata->offset, rdata->bytes);
-
-		rc = adjust_credits(server, &rdata->credits, rdata->bytes);
-		if (!rc) {
-			if (rdata->cfile->invalidHandle)
-				rc = -EAGAIN;
-			else
-				rc = server->ops->async_readv(rdata);
-		}
-
-		if (rc) {
-			add_credits_and_wake_if(server, &rdata->credits, 0);
-			cifs_unlock_folios(rdata->mapping,
-					   rdata->offset / PAGE_SIZE,
-					   (rdata->offset + rdata->bytes - 1) / PAGE_SIZE);
-			/* Fallback to the readpage in error/reconnect cases */
-			kref_put(&rdata->refcount, cifs_readdata_release);
-			break;
-		}
-
-		kref_put(&rdata->refcount, cifs_readdata_release);
-	}
-
-	free_xid(xid);
-}
-
-/*
- * cifs_readpage_worker must be called with the page pinned
- */
-static int cifs_readpage_worker(struct file *file, struct page *page,
-	loff_t *poffset)
-{
-	struct inode *inode = file_inode(file);
-	struct timespec64 atime, mtime;
-	char *read_data;
-	int rc;
-
-	/* Is the page cached? */
-	rc = cifs_readpage_from_fscache(inode, page);
-	if (rc == 0)
-		goto read_complete;
-
-	read_data = kmap(page);
-	/* for reads over a certain size could initiate async read ahead */
-
-	rc = cifs_read(file, read_data, PAGE_SIZE, poffset);
-
-	if (rc < 0)
-		goto io_error;
-	else
-		cifs_dbg(FYI, "Bytes read %d\n", rc);
-
-	/* we do not want atime to be less than mtime, it broke some apps */
-	atime = inode_set_atime_to_ts(inode, current_time(inode));
-	mtime = inode_get_mtime(inode);
-	if (timespec64_compare(&atime, &mtime) < 0)
-		inode_set_atime_to_ts(inode, inode_get_mtime(inode));
-
-	if (PAGE_SIZE > rc)
-		memset(read_data + rc, 0, PAGE_SIZE - rc);
-
-	flush_dcache_page(page);
-	SetPageUptodate(page);
-	rc = 0;
-
-io_error:
-	kunmap(page);
-
-read_complete:
-	unlock_page(page);
-	return rc;
-}
-
-static int cifs_read_folio(struct file *file, struct folio *folio)
-{
-	struct page *page = &folio->page;
-	loff_t offset = page_file_offset(page);
-	int rc = -EACCES;
-	unsigned int xid;
-
-	xid = get_xid();
-
-	if (file->private_data == NULL) {
-		rc = -EBADF;
-		free_xid(xid);
-		return rc;
-	}
-
-	cifs_dbg(FYI, "read_folio %p at offset %d 0x%x\n",
-		 page, (int)offset, (int)offset);
-
-	rc = cifs_readpage_worker(file, page, &offset);
-
-	free_xid(xid);
-	return rc;
-}
-
 static int is_inode_writable(struct cifsInodeInfo *cifs_inode)
 {
 	struct cifsFileInfo *open_file;
@@ -4874,123 +3010,6 @@ bool is_size_safe_to_change(struct cifsInodeInfo *cifsInode, __u64 end_of_file,
 		return true;
 }
 
-static int cifs_write_begin(struct file *file, struct address_space *mapping,
-			loff_t pos, unsigned len,
-			struct page **pagep, void **fsdata)
-{
-	int oncethru = 0;
-	pgoff_t index = pos >> PAGE_SHIFT;
-	loff_t offset = pos & (PAGE_SIZE - 1);
-	loff_t page_start = pos & PAGE_MASK;
-	loff_t i_size;
-	struct page *page;
-	int rc = 0;
-
-	cifs_dbg(FYI, "write_begin from %lld len %d\n", (long long)pos, len);
-
-start:
-	page = grab_cache_page_write_begin(mapping, index);
-	if (!page) {
-		rc = -ENOMEM;
-		goto out;
-	}
-
-	if (PageUptodate(page))
-		goto out;
-
-	/*
-	 * If we write a full page it will be up to date, no need to read from
-	 * the server. If the write is short, we'll end up doing a sync write
-	 * instead.
-	 */
-	if (len == PAGE_SIZE)
-		goto out;
-
-	/*
-	 * optimize away the read when we have an oplock, and we're not
-	 * expecting to use any of the data we'd be reading in. That
-	 * is, when the page lies beyond the EOF, or straddles the EOF
-	 * and the write will cover all of the existing data.
-	 */
-	if (CIFS_CACHE_READ(CIFS_I(mapping->host))) {
-		i_size = i_size_read(mapping->host);
-		if (page_start >= i_size ||
-		    (offset == 0 && (pos + len) >= i_size)) {
-			zero_user_segments(page, 0, offset,
-					   offset + len,
-					   PAGE_SIZE);
-			/*
-			 * PageChecked means that the parts of the page
-			 * to which we're not writing are considered up
-			 * to date. Once the data is copied to the
-			 * page, it can be set uptodate.
-			 */
-			SetPageChecked(page);
-			goto out;
-		}
-	}
-
-	if ((file->f_flags & O_ACCMODE) != O_WRONLY && !oncethru) {
-		/*
-		 * might as well read a page, it is fast enough. If we get
-		 * an error, we don't need to return it. cifs_write_end will
-		 * do a sync write instead since PG_uptodate isn't set.
-		 */
-		cifs_readpage_worker(file, page, &page_start);
-		put_page(page);
-		oncethru = 1;
-		goto start;
-	} else {
-		/* we could try using another file handle if there is one -
-		   but how would we lock it to prevent close of that handle
-		   racing with this read? In any case
-		   this will be written out by write_end so is fine */
-	}
-out:
-	*pagep = page;
-	return rc;
-}
-
-static bool cifs_release_folio(struct folio *folio, gfp_t gfp)
-{
-	if (folio_test_private(folio))
-		return 0;
-	if (folio_test_fscache(folio)) {
-		if (current_is_kswapd() || !(gfp & __GFP_FS))
-			return false;
-		folio_wait_fscache(folio);
-	}
-	fscache_note_page_release(cifs_inode_cookie(folio->mapping->host));
-	return true;
-}
-
-static void cifs_invalidate_folio(struct folio *folio, size_t offset,
-				 size_t length)
-{
-	folio_wait_fscache(folio);
-}
-
-static int cifs_launder_folio(struct folio *folio)
-{
-	int rc = 0;
-	loff_t range_start = folio_pos(folio);
-	loff_t range_end = range_start + folio_size(folio);
-	struct writeback_control wbc = {
-		.sync_mode = WB_SYNC_ALL,
-		.nr_to_write = 0,
-		.range_start = range_start,
-		.range_end = range_end,
-	};
-
-	cifs_dbg(FYI, "Launder page: %lu\n", folio->index);
-
-	if (folio_clear_dirty_for_io(folio))
-		rc = cifs_writepage_locked(&folio->page, &wbc);
-
-	folio_wait_fscache(folio);
-	return rc;
-}
-
 void cifs_oplock_break(struct work_struct *work)
 {
 	struct cifsFileInfo *cfile = container_of(work, struct cifsFileInfo,
@@ -5080,25 +3099,6 @@ out:
 	cifs_done_oplock_break(cinode);
 }
 
-/*
- * The presence of cifs_direct_io() in the address space ops vector
- * allowes open() O_DIRECT flags which would have failed otherwise.
- *
- * In the non-cached mode (mount with cache=none), we shunt off direct read and write requests
- * so this method should never be called.
- *
- * Direct IO is not yet supported in the cached mode.
- */
-static ssize_t
-cifs_direct_io(struct kiocb *iocb, struct iov_iter *iter)
-{
-        /*
-         * FIXME
-         * Eventually need to support direct IO for non forcedirectio mounts
-         */
-        return -EINVAL;
-}
-
 static int cifs_swap_activate(struct swap_info_struct *sis,
 			      struct file *swap_file, sector_t *span)
 {
@@ -5159,24 +3159,42 @@ static void cifs_swap_deactivate(struct file *file)
 	/* do we need to unpin (or unlock) the file */
 }
 
+/**
+ * cifs_swap_rw - SMB3 address space operation for swap I/O
+ * @iocb: target I/O control block
+ * @iter: I/O buffer
+ *
+ * Perform IO to the swap-file.  This is much like direct IO.
+ */
+static int cifs_swap_rw(struct kiocb *iocb, struct iov_iter *iter)
+{
+	ssize_t ret;
+
+	if (iov_iter_rw(iter) == READ)
+		ret = netfs_unbuffered_read_iter_locked(iocb, iter);
+	else
+		ret = netfs_unbuffered_write_iter_locked(iocb, iter, NULL);
+	if (ret < 0)
+		return ret;
+	return 0;
+}
+
 const struct address_space_operations cifs_addr_ops = {
-	.read_folio = cifs_read_folio,
-	.readahead = cifs_readahead,
-	.writepages = cifs_writepages,
-	.write_begin = cifs_write_begin,
-	.write_end = cifs_write_end,
-	.dirty_folio = netfs_dirty_folio,
-	.release_folio = cifs_release_folio,
-	.direct_IO = cifs_direct_io,
-	.invalidate_folio = cifs_invalidate_folio,
-	.launder_folio = cifs_launder_folio,
-	.migrate_folio = filemap_migrate_folio,
+	.read_folio	= netfs_read_folio,
+	.readahead	= netfs_readahead,
+	.writepages	= netfs_writepages,
+	.dirty_folio	= netfs_dirty_folio,
+	.release_folio	= netfs_release_folio,
+	.direct_IO	= noop_direct_IO,
+	.invalidate_folio = netfs_invalidate_folio,
+	.migrate_folio	= filemap_migrate_folio,
 	/*
 	 * TODO: investigate and if useful we could add an is_dirty_writeback
 	 * helper if needed
 	 */
-	.swap_activate = cifs_swap_activate,
+	.swap_activate	= cifs_swap_activate,
 	.swap_deactivate = cifs_swap_deactivate,
+	.swap_rw = cifs_swap_rw,
 };
 
 /*
@@ -5185,13 +3203,10 @@ const struct address_space_operations cifs_addr_ops = {
  * to leave cifs_readahead out of the address space operations.
  */
 const struct address_space_operations cifs_addr_ops_smallbuf = {
-	.read_folio = cifs_read_folio,
-	.writepages = cifs_writepages,
-	.write_begin = cifs_write_begin,
-	.write_end = cifs_write_end,
-	.dirty_folio = netfs_dirty_folio,
-	.release_folio = cifs_release_folio,
-	.invalidate_folio = cifs_invalidate_folio,
-	.launder_folio = cifs_launder_folio,
-	.migrate_folio = filemap_migrate_folio,
+	.read_folio	= netfs_read_folio,
+	.writepages	= netfs_writepages,
+	.dirty_folio	= netfs_dirty_folio,
+	.release_folio	= netfs_release_folio,
+	.invalidate_folio = netfs_invalidate_folio,
+	.migrate_folio	= filemap_migrate_folio,
 };
diff --git a/fs/smb/client/fs_context.c b/fs/smb/client/fs_context.c
index f119035a82..3bbac925d0 100644
--- a/fs/smb/client/fs_context.c
+++ b/fs/smb/client/fs_context.c
@@ -175,6 +175,7 @@ const struct fs_parameter_spec smb3_fs_parameters[] = {
 	fsparam_string("vers", Opt_vers),
 	fsparam_string("sec", Opt_sec),
 	fsparam_string("cache", Opt_cache),
+	fsparam_string("reparse", Opt_reparse),
 
 	/* Arguments that should be ignored */
 	fsparam_flag("guest", Opt_ignore),
@@ -297,6 +298,35 @@ cifs_parse_cache_flavor(struct fs_context *fc, char *value, struct smb3_fs_conte
 	return 0;
 }
 
+static const match_table_t reparse_flavor_tokens = {
+	{ Opt_reparse_default,	"default" },
+	{ Opt_reparse_nfs,	"nfs" },
+	{ Opt_reparse_wsl,	"wsl" },
+	{ Opt_reparse_err,	NULL },
+};
+
+static int parse_reparse_flavor(struct fs_context *fc, char *value,
+				struct smb3_fs_context *ctx)
+{
+	substring_t args[MAX_OPT_ARGS];
+
+	switch (match_token(value, reparse_flavor_tokens, args)) {
+	case Opt_reparse_default:
+		ctx->reparse_type = CIFS_REPARSE_TYPE_DEFAULT;
+		break;
+	case Opt_reparse_nfs:
+		ctx->reparse_type = CIFS_REPARSE_TYPE_NFS;
+		break;
+	case Opt_reparse_wsl:
+		ctx->reparse_type = CIFS_REPARSE_TYPE_WSL;
+		break;
+	default:
+		cifs_errorf(fc, "bad reparse= option: %s\n", value);
+		return 1;
+	}
+	return 0;
+}
+
 #define DUP_CTX_STR(field)						\
 do {									\
 	if (ctx->field) {						\
@@ -948,7 +978,7 @@ static int smb3_fs_context_parse_param(struct fs_context *fc,
 
 	switch (opt) {
 	case Opt_compress:
-		ctx->compression = UNKNOWN_TYPE;
+		ctx->compress = true;
 		cifs_dbg(VFS,
 			"SMB3 compression support is experimental\n");
 		break;
@@ -1595,6 +1625,10 @@ static int smb3_fs_context_parse_param(struct fs_context *fc,
 	case Opt_rdma:
 		ctx->rdma = true;
 		break;
+	case Opt_reparse:
+		if (parse_reparse_flavor(fc, param->string, ctx))
+			goto cifs_parse_mount_err;
+		break;
 	}
 	/* case Opt_ignore: - is ignored as expected ... */
 
@@ -1683,6 +1717,7 @@ int smb3_init_fs_context(struct fs_context *fc)
 	ctx->backupgid_specified = false; /* no backup intent for a group */
 
 	ctx->retrans = 1;
+	ctx->reparse_type = CIFS_REPARSE_TYPE_DEFAULT;
 
 /*
  *	short int override_uid = -1;
diff --git a/fs/smb/client/fs_context.h b/fs/smb/client/fs_context.h
index 369a3fea1d..cf577ec0dd 100644
--- a/fs/smb/client/fs_context.h
+++ b/fs/smb/client/fs_context.h
@@ -41,6 +41,13 @@ enum {
 	Opt_cache_err
 };
 
+enum cifs_reparse_parm {
+	Opt_reparse_default,
+	Opt_reparse_nfs,
+	Opt_reparse_wsl,
+	Opt_reparse_err
+};
+
 enum cifs_sec_param {
 	Opt_sec_krb5,
 	Opt_sec_krb5i,
@@ -149,6 +156,7 @@ enum cifs_param {
 	Opt_vers,
 	Opt_sec,
 	Opt_cache,
+	Opt_reparse,
 
 	/* Mount options to be ignored */
 	Opt_ignore,
@@ -269,12 +277,13 @@ struct smb3_fs_context {
 	unsigned int max_credits; /* smb3 max_credits 10 < credits < 60000 */
 	unsigned int max_channels;
 	unsigned int max_cached_dirs;
-	__u16 compression; /* compression algorithm 0xFFFF default 0=disabled */
+	bool compress; /* enable SMB2 messages (READ/WRITE) de/compression */
 	bool rootfs:1; /* if it's a SMB root file system */
 	bool witness:1; /* use witness protocol */
 	char *leaf_fullpath;
 	struct cifs_ses *dfs_root_ses;
 	bool dfs_automount:1; /* set for dfs automount only */
+	enum cifs_reparse_type reparse_type;
 };
 
 extern const struct fs_parameter_spec smb3_fs_parameters[];
diff --git a/fs/smb/client/fscache.c b/fs/smb/client/fscache.c
index 113bde8f1e..01424a5cdb 100644
--- a/fs/smb/client/fscache.c
+++ b/fs/smb/client/fscache.c
@@ -94,6 +94,11 @@ int cifs_fscache_get_super_cookie(struct cifs_tcon *tcon)
 		}
 		pr_err("Cache volume key already in use (%s)\n", key);
 		vcookie = NULL;
+		trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+				    netfs_trace_tcon_ref_see_fscache_collision);
+	} else {
+		trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+				    netfs_trace_tcon_ref_see_fscache_okay);
 	}
 
 	tcon->fscache = vcookie;
@@ -115,6 +120,8 @@ void cifs_fscache_release_super_cookie(struct cifs_tcon *tcon)
 	cifs_fscache_fill_volume_coherency(tcon, &cd);
 	fscache_relinquish_volume(tcon->fscache, &cd, false);
 	tcon->fscache = NULL;
+	trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+			    netfs_trace_tcon_ref_see_fscache_relinq);
 }
 
 void cifs_fscache_get_inode_cookie(struct inode *inode)
@@ -163,112 +170,3 @@ void cifs_fscache_release_inode_cookie(struct inode *inode)
 		cifsi->netfs.cache = NULL;
 	}
 }
-
-/*
- * Fallback page reading interface.
- */
-static int fscache_fallback_read_page(struct inode *inode, struct page *page)
-{
-	struct netfs_cache_resources cres;
-	struct fscache_cookie *cookie = cifs_inode_cookie(inode);
-	struct iov_iter iter;
-	struct bio_vec bvec;
-	int ret;
-
-	memset(&cres, 0, sizeof(cres));
-	bvec_set_page(&bvec, page, PAGE_SIZE, 0);
-	iov_iter_bvec(&iter, ITER_DEST, &bvec, 1, PAGE_SIZE);
-
-	ret = fscache_begin_read_operation(&cres, cookie);
-	if (ret < 0)
-		return ret;
-
-	ret = fscache_read(&cres, page_offset(page), &iter, NETFS_READ_HOLE_FAIL,
-			   NULL, NULL);
-	fscache_end_operation(&cres);
-	return ret;
-}
-
-/*
- * Fallback page writing interface.
- */
-static int fscache_fallback_write_pages(struct inode *inode, loff_t start, size_t len,
-					bool no_space_allocated_yet)
-{
-	struct netfs_cache_resources cres;
-	struct fscache_cookie *cookie = cifs_inode_cookie(inode);
-	struct iov_iter iter;
-	int ret;
-
-	memset(&cres, 0, sizeof(cres));
-	iov_iter_xarray(&iter, ITER_SOURCE, &inode->i_mapping->i_pages, start, len);
-
-	ret = fscache_begin_write_operation(&cres, cookie);
-	if (ret < 0)
-		return ret;
-
-	ret = cres.ops->prepare_write(&cres, &start, &len, len, i_size_read(inode),
-				      no_space_allocated_yet);
-	if (ret == 0)
-		ret = fscache_write(&cres, start, &iter, NULL, NULL);
-	fscache_end_operation(&cres);
-	return ret;
-}
-
-/*
- * Retrieve a page from FS-Cache
- */
-int __cifs_readpage_from_fscache(struct inode *inode, struct page *page)
-{
-	int ret;
-
-	cifs_dbg(FYI, "%s: (fsc:%p, p:%p, i:0x%p\n",
-		 __func__, cifs_inode_cookie(inode), page, inode);
-
-	ret = fscache_fallback_read_page(inode, page);
-	if (ret < 0)
-		return ret;
-
-	/* Read completed synchronously */
-	SetPageUptodate(page);
-	return 0;
-}
-
-void __cifs_readahead_to_fscache(struct inode *inode, loff_t pos, size_t len)
-{
-	cifs_dbg(FYI, "%s: (fsc: %p, p: %llx, l: %zx, i: %p)\n",
-		 __func__, cifs_inode_cookie(inode), pos, len, inode);
-
-	fscache_fallback_write_pages(inode, pos, len, true);
-}
-
-/*
- * Query the cache occupancy.
- */
-int __cifs_fscache_query_occupancy(struct inode *inode,
-				   pgoff_t first, unsigned int nr_pages,
-				   pgoff_t *_data_first,
-				   unsigned int *_data_nr_pages)
-{
-	struct netfs_cache_resources cres;
-	struct fscache_cookie *cookie = cifs_inode_cookie(inode);
-	loff_t start, data_start;
-	size_t len, data_len;
-	int ret;
-
-	ret = fscache_begin_read_operation(&cres, cookie);
-	if (ret < 0)
-		return ret;
-
-	start = first * PAGE_SIZE;
-	len = nr_pages * PAGE_SIZE;
-	ret = cres.ops->query_occupancy(&cres, start, len, PAGE_SIZE,
-					&data_start, &data_len);
-	if (ret == 0) {
-		*_data_first = data_start / PAGE_SIZE;
-		*_data_nr_pages = len / PAGE_SIZE;
-	}
-
-	fscache_end_operation(&cres);
-	return ret;
-}
diff --git a/fs/smb/client/fscache.h b/fs/smb/client/fscache.h
index 1f2ea9f5cc..f06cb24f5f 100644
--- a/fs/smb/client/fscache.h
+++ b/fs/smb/client/fscache.h
@@ -74,41 +74,6 @@ static inline void cifs_invalidate_cache(struct inode *inode, unsigned int flags
 			   i_size_read(inode), flags);
 }
 
-extern int __cifs_fscache_query_occupancy(struct inode *inode,
-					  pgoff_t first, unsigned int nr_pages,
-					  pgoff_t *_data_first,
-					  unsigned int *_data_nr_pages);
-
-static inline int cifs_fscache_query_occupancy(struct inode *inode,
-					       pgoff_t first, unsigned int nr_pages,
-					       pgoff_t *_data_first,
-					       unsigned int *_data_nr_pages)
-{
-	if (!cifs_inode_cookie(inode))
-		return -ENOBUFS;
-	return __cifs_fscache_query_occupancy(inode, first, nr_pages,
-					      _data_first, _data_nr_pages);
-}
-
-extern int __cifs_readpage_from_fscache(struct inode *pinode, struct page *ppage);
-extern void __cifs_readahead_to_fscache(struct inode *pinode, loff_t pos, size_t len);
-
-
-static inline int cifs_readpage_from_fscache(struct inode *inode,
-					     struct page *page)
-{
-	if (cifs_inode_cookie(inode))
-		return __cifs_readpage_from_fscache(inode, page);
-	return -ENOBUFS;
-}
-
-static inline void cifs_readahead_to_fscache(struct inode *inode,
-					     loff_t pos, size_t len)
-{
-	if (cifs_inode_cookie(inode))
-		__cifs_readahead_to_fscache(inode, pos, len);
-}
-
 static inline bool cifs_fscache_enabled(struct inode *inode)
 {
 	return fscache_cookie_enabled(cifs_inode_cookie(inode));
@@ -131,25 +96,6 @@ static inline struct fscache_cookie *cifs_inode_cookie(struct inode *inode) { re
 static inline void cifs_invalidate_cache(struct inode *inode, unsigned int flags) {}
 static inline bool cifs_fscache_enabled(struct inode *inode) { return false; }
 
-static inline int cifs_fscache_query_occupancy(struct inode *inode,
-					       pgoff_t first, unsigned int nr_pages,
-					       pgoff_t *_data_first,
-					       unsigned int *_data_nr_pages)
-{
-	*_data_first = ULONG_MAX;
-	*_data_nr_pages = 0;
-	return -ENOBUFS;
-}
-
-static inline int
-cifs_readpage_from_fscache(struct inode *inode, struct page *page)
-{
-	return -ENOBUFS;
-}
-
-static inline
-void cifs_readahead_to_fscache(struct inode *inode, loff_t pos, size_t len) {}
-
 #endif /* CONFIG_CIFS_FSCACHE */
 
 #endif /* _CIFS_FSCACHE_H */
diff --git a/fs/smb/client/inode.c b/fs/smb/client/inode.c
index 4c3dec384f..dd0afa2373 100644
--- a/fs/smb/client/inode.c
+++ b/fs/smb/client/inode.c
@@ -26,15 +26,28 @@
 #include "fs_context.h"
 #include "cifs_ioctl.h"
 #include "cached_dir.h"
+#include "reparse.h"
+
+/*
+ * Set parameters for the netfs library
+ */
+static void cifs_set_netfs_context(struct inode *inode)
+{
+	struct cifsInodeInfo *cifs_i = CIFS_I(inode);
+
+	netfs_inode_init(&cifs_i->netfs, &cifs_req_ops, true);
+}
 
 static void cifs_set_ops(struct inode *inode)
 {
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
+	struct netfs_inode *ictx = netfs_inode(inode);
 
 	switch (inode->i_mode & S_IFMT) {
 	case S_IFREG:
 		inode->i_op = &cifs_file_inode_ops;
 		if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_DIRECT_IO) {
+			set_bit(NETFS_ICTX_UNBUFFERED, &ictx->flags);
 			if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_BRL)
 				inode->i_fop = &cifs_file_direct_nobrl_ops;
 			else
@@ -56,6 +69,7 @@ static void cifs_set_ops(struct inode *inode)
 			inode->i_data.a_ops = &cifs_addr_ops_smallbuf;
 		else
 			inode->i_data.a_ops = &cifs_addr_ops;
+		mapping_set_large_folios(inode->i_mapping);
 		break;
 	case S_IFDIR:
 		if (IS_AUTOMOUNT(inode)) {
@@ -220,8 +234,10 @@ cifs_fattr_to_inode(struct inode *inode, struct cifs_fattr *fattr,
 
 	if (fattr->cf_flags & CIFS_FATTR_JUNCTION)
 		inode->i_flags |= S_AUTOMOUNT;
-	if (inode->i_state & I_NEW)
+	if (inode->i_state & I_NEW) {
+		cifs_set_netfs_context(inode);
 		cifs_set_ops(inode);
+	}
 	return 0;
 }
 
@@ -400,7 +416,6 @@ cifs_get_file_info_unix(struct file *filp)
 		cifs_unix_basic_to_fattr(&fattr, &find_data, cifs_sb);
 	} else if (rc == -EREMOTE) {
 		cifs_create_junction_fattr(&fattr, inode->i_sb);
-		rc = 0;
 	} else
 		goto cifs_gfiunix_out;
 
@@ -591,6 +606,10 @@ cifs_sfu_type(struct cifs_fattr *fattr, const char *path,
 				mnr = le64_to_cpu(*(__le64 *)(pbuf+16));
 				fattr->cf_rdev = MKDEV(mjr, mnr);
 			}
+		} else if (memcmp("LnxSOCK", pbuf, 8) == 0) {
+			cifs_dbg(FYI, "Socket\n");
+			fattr->cf_mode |= S_IFSOCK;
+			fattr->cf_dtype = DT_SOCK;
 		} else if (memcmp("IntxLNK", pbuf, 7) == 0) {
 			cifs_dbg(FYI, "Symlink\n");
 			fattr->cf_mode |= S_IFLNK;
@@ -728,84 +747,6 @@ out_reparse:
 		fattr->cf_mode, fattr->cf_uniqueid, fattr->cf_nlink);
 }
 
-static inline dev_t nfs_mkdev(struct reparse_posix_data *buf)
-{
-	u64 v = le64_to_cpu(*(__le64 *)buf->DataBuffer);
-
-	return MKDEV(v >> 32, v & 0xffffffff);
-}
-
-bool cifs_reparse_point_to_fattr(struct cifs_sb_info *cifs_sb,
-				 struct cifs_fattr *fattr,
-				 struct cifs_open_info_data *data)
-{
-	struct reparse_posix_data *buf = data->reparse.posix;
-	u32 tag = data->reparse.tag;
-
-	if (tag == IO_REPARSE_TAG_NFS && buf) {
-		switch (le64_to_cpu(buf->InodeType)) {
-		case NFS_SPECFILE_CHR:
-			fattr->cf_mode |= S_IFCHR;
-			fattr->cf_dtype = DT_CHR;
-			fattr->cf_rdev = nfs_mkdev(buf);
-			break;
-		case NFS_SPECFILE_BLK:
-			fattr->cf_mode |= S_IFBLK;
-			fattr->cf_dtype = DT_BLK;
-			fattr->cf_rdev = nfs_mkdev(buf);
-			break;
-		case NFS_SPECFILE_FIFO:
-			fattr->cf_mode |= S_IFIFO;
-			fattr->cf_dtype = DT_FIFO;
-			break;
-		case NFS_SPECFILE_SOCK:
-			fattr->cf_mode |= S_IFSOCK;
-			fattr->cf_dtype = DT_SOCK;
-			break;
-		case NFS_SPECFILE_LNK:
-			fattr->cf_mode |= S_IFLNK;
-			fattr->cf_dtype = DT_LNK;
-			break;
-		default:
-			WARN_ON_ONCE(1);
-			return false;
-		}
-		return true;
-	}
-
-	switch (tag) {
-	case IO_REPARSE_TAG_LX_SYMLINK:
-		fattr->cf_mode |= S_IFLNK;
-		fattr->cf_dtype = DT_LNK;
-		break;
-	case IO_REPARSE_TAG_LX_FIFO:
-		fattr->cf_mode |= S_IFIFO;
-		fattr->cf_dtype = DT_FIFO;
-		break;
-	case IO_REPARSE_TAG_AF_UNIX:
-		fattr->cf_mode |= S_IFSOCK;
-		fattr->cf_dtype = DT_SOCK;
-		break;
-	case IO_REPARSE_TAG_LX_CHR:
-		fattr->cf_mode |= S_IFCHR;
-		fattr->cf_dtype = DT_CHR;
-		break;
-	case IO_REPARSE_TAG_LX_BLK:
-		fattr->cf_mode |= S_IFBLK;
-		fattr->cf_dtype = DT_BLK;
-		break;
-	case 0: /* SMB1 symlink */
-	case IO_REPARSE_TAG_SYMLINK:
-	case IO_REPARSE_TAG_NFS:
-		fattr->cf_mode |= S_IFLNK;
-		fattr->cf_dtype = DT_LNK;
-		break;
-	default:
-		return false;
-	}
-	return true;
-}
-
 static void cifs_open_info_to_fattr(struct cifs_fattr *fattr,
 				    struct cifs_open_info_data *data,
 				    struct super_block *sb)
@@ -836,6 +777,8 @@ static void cifs_open_info_to_fattr(struct cifs_fattr *fattr,
 	fattr->cf_bytes = le64_to_cpu(info->AllocationSize);
 	fattr->cf_createtime = le64_to_cpu(info->CreationTime);
 	fattr->cf_nlink = le32_to_cpu(info->NumberOfLinks);
+	fattr->cf_uid = cifs_sb->ctx->linux_uid;
+	fattr->cf_gid = cifs_sb->ctx->linux_gid;
 
 	fattr->cf_mode = cifs_sb->ctx->file_mode;
 	if (cifs_open_data_reparse(data) &&
@@ -878,9 +821,6 @@ out_reparse:
 		fattr->cf_symlink_target = data->symlink_target;
 		data->symlink_target = NULL;
 	}
-
-	fattr->cf_uid = cifs_sb->ctx->linux_uid;
-	fattr->cf_gid = cifs_sb->ctx->linux_gid;
 }
 
 static int
@@ -894,9 +834,14 @@ cifs_get_file_info(struct file *filp)
 	struct cifsFileInfo *cfile = filp->private_data;
 	struct cifs_tcon *tcon = tlink_tcon(cfile->tlink);
 	struct TCP_Server_Info *server = tcon->ses->server;
+	struct dentry *dentry = filp->f_path.dentry;
+	void *page = alloc_dentry_path();
+	const unsigned char *path;
 
-	if (!server->ops->query_file_info)
+	if (!server->ops->query_file_info) {
+		free_dentry_path(page);
 		return -ENOSYS;
+	}
 
 	xid = get_xid();
 	rc = server->ops->query_file_info(xid, tcon, cfile, &data);
@@ -908,11 +853,17 @@ cifs_get_file_info(struct file *filp)
 			data.symlink = true;
 			data.reparse.tag = IO_REPARSE_TAG_SYMLINK;
 		}
+		path = build_path_from_dentry(dentry, page);
+		if (IS_ERR(path)) {
+			rc = PTR_ERR(path);
+			goto cgfi_exit;
+		}
 		cifs_open_info_to_fattr(&fattr, &data, inode->i_sb);
+		if (fattr.cf_flags & CIFS_FATTR_DELETE_PENDING)
+			cifs_mark_open_handles_for_deleted_file(inode, path);
 		break;
 	case -EREMOTE:
 		cifs_create_junction_fattr(&fattr, inode->i_sb);
-		rc = 0;
 		break;
 	case -EOPNOTSUPP:
 	case -EINVAL:
@@ -938,6 +889,7 @@ cifs_get_file_info(struct file *filp)
 	rc = cifs_fattr_to_inode(inode, &fattr, false);
 cgfi_exit:
 	cifs_free_open_info(&data);
+	free_dentry_path(page);
 	free_xid(xid);
 	return rc;
 }
@@ -1090,13 +1042,26 @@ static int reparse_info_to_fattr(struct cifs_open_info_data *data,
 	}
 
 	rc = -EOPNOTSUPP;
-	switch ((data->reparse.tag = tag)) {
-	case 0: /* SMB1 symlink */
+	data->reparse.tag = tag;
+	if (!data->reparse.tag) {
 		if (server->ops->query_symlink) {
 			rc = server->ops->query_symlink(xid, tcon,
 							cifs_sb, full_path,
 							&data->symlink_target);
 		}
+		if (rc == -EOPNOTSUPP)
+			data->reparse.tag = IO_REPARSE_TAG_INTERNAL;
+	}
+
+	switch (data->reparse.tag) {
+	case 0: /* SMB1 symlink */
+		break;
+	case IO_REPARSE_TAG_INTERNAL:
+		rc = 0;
+		if (le32_to_cpu(data->fi.Attributes) & ATTR_DIRECTORY) {
+			cifs_create_junction_fattr(fattr, sb);
+			goto out;
+		}
 		break;
 	case IO_REPARSE_TAG_MOUNT_POINT:
 		cifs_create_junction_fattr(fattr, sb);
@@ -1172,6 +1137,9 @@ static int cifs_get_fattr(struct cifs_open_info_data *data,
 		} else {
 			cifs_open_info_to_fattr(fattr, data, sb);
 		}
+		if (!rc && *inode &&
+		    (fattr->cf_flags & CIFS_FATTR_DELETE_PENDING))
+			cifs_mark_open_handles_for_deleted_file(*inode, full_path);
 		break;
 	case -EREMOTE:
 		/* DFS link, no metadata available on this server */
@@ -1400,6 +1368,8 @@ int smb311_posix_get_inode_info(struct inode **inode,
 		goto out;
 
 	rc = update_inode_info(sb, &fattr, inode);
+	if (!rc && fattr.cf_flags & CIFS_FATTR_DELETE_PENDING)
+		cifs_mark_open_handles_for_deleted_file(*inode, full_path);
 out:
 	kfree(fattr.cf_symlink_target);
 	return rc;
@@ -1563,6 +1533,9 @@ iget_root:
 		goto out;
 	}
 
+	if (!rc && fattr.cf_flags & CIFS_FATTR_DELETE_PENDING)
+		cifs_mark_open_handles_for_deleted_file(inode, path);
+
 	if (rc && tcon->pipe) {
 		cifs_dbg(FYI, "ipc connection - fake read inode\n");
 		spin_lock(&inode->i_lock);
@@ -1849,20 +1822,24 @@ retry_std_delete:
 		goto psx_del_no_retry;
 	}
 
-	rc = server->ops->unlink(xid, tcon, full_path, cifs_sb);
+	rc = server->ops->unlink(xid, tcon, full_path, cifs_sb, dentry);
 
 psx_del_no_retry:
 	if (!rc) {
-		if (inode)
+		if (inode) {
+			cifs_mark_open_handles_for_deleted_file(inode, full_path);
 			cifs_drop_nlink(inode);
+		}
 	} else if (rc == -ENOENT) {
 		d_drop(dentry);
 	} else if (rc == -EBUSY) {
 		if (server->ops->rename_pending_delete) {
 			rc = server->ops->rename_pending_delete(full_path,
 								dentry, xid);
-			if (rc == 0)
+			if (rc == 0) {
+				cifs_mark_open_handles_for_deleted_file(inode, full_path);
 				cifs_drop_nlink(inode);
+			}
 		}
 	} else if ((rc == -EACCES) && (dosattr == 0) && inode) {
 		attrs = kzalloc(sizeof(*attrs), GFP_KERNEL);
@@ -2486,24 +2463,6 @@ cifs_dentry_needs_reval(struct dentry *dentry)
 	return false;
 }
 
-/*
- * Zap the cache. Called when invalid_mapping flag is set.
- */
-int
-cifs_invalidate_mapping(struct inode *inode)
-{
-	int rc = 0;
-
-	if (inode->i_mapping && inode->i_mapping->nrpages != 0) {
-		rc = invalidate_inode_pages2(inode->i_mapping);
-		if (rc)
-			cifs_dbg(VFS, "%s: invalidate inode %p failed with rc %d\n",
-				 __func__, inode, rc);
-	}
-
-	return rc;
-}
-
 /**
  * cifs_wait_bit_killable - helper for functions that are sleeping on bit locks
  *
@@ -2523,7 +2482,8 @@ int
 cifs_revalidate_mapping(struct inode *inode)
 {
 	int rc;
-	unsigned long *flags = &CIFS_I(inode)->flags;
+	struct cifsInodeInfo *cifs_inode = CIFS_I(inode);
+	unsigned long *flags = &cifs_inode->flags;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
 
 	/* swapfiles are not supposed to be shared */
@@ -2540,9 +2500,13 @@ cifs_revalidate_mapping(struct inode *inode)
 		if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_RW_CACHE)
 			goto skip_invalidate;
 
-		rc = cifs_invalidate_mapping(inode);
-		if (rc)
+		cifs_inode->netfs.zero_point = cifs_inode->netfs.remote_i_size;
+		rc = filemap_invalidate_inode(inode, true, 0, LLONG_MAX);
+		if (rc) {
+			cifs_dbg(VFS, "%s: invalidate inode %p failed with rc %d\n",
+				 __func__, inode, rc);
 			set_bit(CIFS_INO_INVALID_MAPPING, flags);
+		}
 	}
 
 skip_invalidate:
@@ -2800,7 +2764,7 @@ void cifs_setsize(struct inode *inode, loff_t offset)
 
 static int
 cifs_set_file_size(struct inode *inode, struct iattr *attrs,
-		   unsigned int xid, const char *full_path)
+		   unsigned int xid, const char *full_path, struct dentry *dentry)
 {
 	int rc;
 	struct cifsFileInfo *open_file;
@@ -2851,7 +2815,7 @@ cifs_set_file_size(struct inode *inode, struct iattr *attrs,
 	 */
 	if (server->ops->set_path_size)
 		rc = server->ops->set_path_size(xid, tcon, full_path,
-						attrs->ia_size, cifs_sb, false);
+						attrs->ia_size, cifs_sb, false, dentry);
 	else
 		rc = -ENOSYS;
 	cifs_dbg(FYI, "SetEOF by path (setattrs) rc = %d\n", rc);
@@ -2941,7 +2905,7 @@ cifs_setattr_unix(struct dentry *direntry, struct iattr *attrs)
 	rc = 0;
 
 	if (attrs->ia_valid & ATTR_SIZE) {
-		rc = cifs_set_file_size(inode, attrs, xid, full_path);
+		rc = cifs_set_file_size(inode, attrs, xid, full_path, direntry);
 		if (rc != 0)
 			goto out;
 	}
@@ -3108,7 +3072,7 @@ cifs_setattr_nounix(struct dentry *direntry, struct iattr *attrs)
 	}
 
 	if (attrs->ia_valid & ATTR_SIZE) {
-		rc = cifs_set_file_size(inode, attrs, xid, full_path);
+		rc = cifs_set_file_size(inode, attrs, xid, full_path, direntry);
 		if (rc != 0)
 			goto cifs_setattr_exit;
 	}
diff --git a/fs/smb/client/ioctl.c b/fs/smb/client/ioctl.c
index 682eabdd1d..855ac5a62e 100644
--- a/fs/smb/client/ioctl.c
+++ b/fs/smb/client/ioctl.c
@@ -349,6 +349,11 @@ long cifs_ioctl(struct file *filep, unsigned int command, unsigned long arg)
 	xid = get_xid();
 
 	cifs_dbg(FYI, "cifs ioctl 0x%x\n", command);
+	if (pSMBFile == NULL)
+		trace_smb3_ioctl(xid, 0, command);
+	else
+		trace_smb3_ioctl(xid, pSMBFile->fid.persistent_fid, command);
+
 	switch (command) {
 		case FS_IOC_GETFLAGS:
 			if (pSMBFile == NULL)
diff --git a/fs/smb/client/misc.c b/fs/smb/client/misc.c
index d56959d02e..65d4b72b4d 100644
--- a/fs/smb/client/misc.c
+++ b/fs/smb/client/misc.c
@@ -27,9 +27,6 @@
 #include "fs_context.h"
 #include "cached_dir.h"
 
-extern mempool_t *cifs_sm_req_poolp;
-extern mempool_t *cifs_req_poolp;
-
 /* The xid serves as a useful identifier for each incoming vfs request,
    in a similar way to the mid which is useful to track each sent smb,
    and CurrentXid can also provide a running counter (although it
@@ -114,9 +111,10 @@ sesInfoFree(struct cifs_ses *buf_to_free)
 }
 
 struct cifs_tcon *
-tcon_info_alloc(bool dir_leases_enabled)
+tcon_info_alloc(bool dir_leases_enabled, enum smb3_tcon_ref_trace trace)
 {
 	struct cifs_tcon *ret_buf;
+	static atomic_t tcon_debug_id;
 
 	ret_buf = kzalloc(sizeof(*ret_buf), GFP_KERNEL);
 	if (!ret_buf)
@@ -133,7 +131,8 @@ tcon_info_alloc(bool dir_leases_enabled)
 
 	atomic_inc(&tconInfoAllocCount);
 	ret_buf->status = TID_NEW;
-	++ret_buf->tc_count;
+	ret_buf->debug_id = atomic_inc_return(&tcon_debug_id);
+	ret_buf->tc_count = 1;
 	spin_lock_init(&ret_buf->tc_lock);
 	INIT_LIST_HEAD(&ret_buf->openFileList);
 	INIT_LIST_HEAD(&ret_buf->tcon_list);
@@ -145,17 +144,19 @@ tcon_info_alloc(bool dir_leases_enabled)
 #ifdef CONFIG_CIFS_FSCACHE
 	mutex_init(&ret_buf->fscache_lock);
 #endif
+	trace_smb3_tcon_ref(ret_buf->debug_id, ret_buf->tc_count, trace);
 
 	return ret_buf;
 }
 
 void
-tconInfoFree(struct cifs_tcon *tcon)
+tconInfoFree(struct cifs_tcon *tcon, enum smb3_tcon_ref_trace trace)
 {
 	if (tcon == NULL) {
 		cifs_dbg(FYI, "Null buffer passed to tconInfoFree\n");
 		return;
 	}
+	trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count, trace);
 	free_cached_dirs(tcon->cfids);
 	atomic_dec(&tconInfoAllocCount);
 	kfree(tcon->nativeFileSystem);
@@ -853,6 +854,40 @@ cifs_close_deferred_file_under_dentry(struct cifs_tcon *tcon, const char *path)
 	free_dentry_path(page);
 }
 
+/*
+ * If a dentry has been deleted, all corresponding open handles should know that
+ * so that we do not defer close them.
+ */
+void cifs_mark_open_handles_for_deleted_file(struct inode *inode,
+					     const char *path)
+{
+	struct cifsFileInfo *cfile;
+	void *page;
+	const char *full_path;
+	struct cifsInodeInfo *cinode = CIFS_I(inode);
+
+	page = alloc_dentry_path();
+	spin_lock(&cinode->open_file_lock);
+
+	/*
+	 * note: we need to construct path from dentry and compare only if the
+	 * inode has any hardlinks. When number of hardlinks is 1, we can just
+	 * mark all open handles since they are going to be from the same file.
+	 */
+	if (inode->i_nlink > 1) {
+		list_for_each_entry(cfile, &cinode->openFileList, flist) {
+			full_path = build_path_from_dentry(cfile->dentry, page);
+			if (!IS_ERR(full_path) && strcmp(full_path, path) == 0)
+				cfile->status_file_deleted = true;
+		}
+	} else {
+		list_for_each_entry(cfile, &cinode->openFileList, flist)
+			cfile->status_file_deleted = true;
+	}
+	spin_unlock(&cinode->open_file_lock);
+	free_dentry_path(page);
+}
+
 /* parses DFS referral V3 structure
  * caller is responsible for freeing target_nodes
  * returns:
@@ -1253,6 +1288,7 @@ int cifs_inval_name_dfs_link_error(const unsigned int xid,
 				   const char *full_path,
 				   bool *islink)
 {
+	struct TCP_Server_Info *server = tcon->ses->server;
 	struct cifs_ses *ses = tcon->ses;
 	size_t len;
 	char *path;
@@ -1269,12 +1305,12 @@ int cifs_inval_name_dfs_link_error(const unsigned int xid,
 	    !is_tcon_dfs(tcon))
 		return 0;
 
-	spin_lock(&tcon->tc_lock);
-	if (!tcon->origin_fullpath) {
-		spin_unlock(&tcon->tc_lock);
+	spin_lock(&server->srv_lock);
+	if (!server->leaf_fullpath) {
+		spin_unlock(&server->srv_lock);
 		return 0;
 	}
-	spin_unlock(&tcon->tc_lock);
+	spin_unlock(&server->srv_lock);
 
 	/*
 	 * Slow path - tcon is DFS and @full_path has prefix path, so attempt
diff --git a/fs/smb/client/readdir.c b/fs/smb/client/readdir.c
index 132ae7d884..ebe1cb30e1 100644
--- a/fs/smb/client/readdir.c
+++ b/fs/smb/client/readdir.c
@@ -22,6 +22,7 @@
 #include "smb2proto.h"
 #include "fs_context.h"
 #include "cached_dir.h"
+#include "reparse.h"
 
 /*
  * To be safe - for UCS to UTF-8 with strings loaded with the rare long
@@ -56,23 +57,6 @@ static inline void dump_cifs_file_struct(struct file *file, char *label)
 #endif /* DEBUG2 */
 
 /*
- * Match a reparse point inode if reparse tag and ctime haven't changed.
- *
- * Windows Server updates ctime of reparse points when their data have changed.
- * The server doesn't allow changing reparse tags from existing reparse points,
- * though it's worth checking.
- */
-static inline bool reparse_inode_match(struct inode *inode,
-				       struct cifs_fattr *fattr)
-{
-	struct timespec64 ctime = inode_get_ctime(inode);
-
-	return (CIFS_I(inode)->cifsAttrs & ATTR_REPARSE) &&
-		CIFS_I(inode)->reparse_tag == fattr->cf_cifstag &&
-		timespec64_equal(&ctime, &fattr->cf_ctime);
-}
-
-/*
  * Attempt to preload the dcache with the results from the FIND_FIRST/NEXT
  *
  * Find the dentry that matches "name". If there isn't one, create one. If it's
@@ -141,6 +125,8 @@ retry:
 					if (likely(reparse_inode_match(inode, fattr))) {
 						fattr->cf_mode = inode->i_mode;
 						fattr->cf_rdev = inode->i_rdev;
+						fattr->cf_uid = inode->i_uid;
+						fattr->cf_gid = inode->i_gid;
 						fattr->cf_eof = CIFS_I(inode)->netfs.remote_i_size;
 						fattr->cf_symlink_target = NULL;
 					} else {
diff --git a/fs/smb/client/reparse.c b/fs/smb/client/reparse.c
new file mode 100644
index 0000000000..689d8a506d
--- /dev/null
+++ b/fs/smb/client/reparse.c
@@ -0,0 +1,536 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) 2024 Paulo Alcantara <pc@manguebit.com>
+ */
+
+#include <linux/fs.h>
+#include <linux/stat.h>
+#include <linux/slab.h>
+#include "cifsglob.h"
+#include "smb2proto.h"
+#include "cifsproto.h"
+#include "cifs_unicode.h"
+#include "cifs_debug.h"
+#include "fs_context.h"
+#include "reparse.h"
+
+int smb2_create_reparse_symlink(const unsigned int xid, struct inode *inode,
+				struct dentry *dentry, struct cifs_tcon *tcon,
+				const char *full_path, const char *symname)
+{
+	struct reparse_symlink_data_buffer *buf = NULL;
+	struct cifs_open_info_data data;
+	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
+	struct inode *new;
+	struct kvec iov;
+	__le16 *path;
+	char *sym, sep = CIFS_DIR_SEP(cifs_sb);
+	u16 len, plen;
+	int rc = 0;
+
+	sym = kstrdup(symname, GFP_KERNEL);
+	if (!sym)
+		return -ENOMEM;
+
+	data = (struct cifs_open_info_data) {
+		.reparse_point = true,
+		.reparse = { .tag = IO_REPARSE_TAG_SYMLINK, },
+		.symlink_target = sym,
+	};
+
+	convert_delimiter(sym, sep);
+	path = cifs_convert_path_to_utf16(sym, cifs_sb);
+	if (!path) {
+		rc = -ENOMEM;
+		goto out;
+	}
+
+	plen = 2 * UniStrnlen((wchar_t *)path, PATH_MAX);
+	len = sizeof(*buf) + plen * 2;
+	buf = kzalloc(len, GFP_KERNEL);
+	if (!buf) {
+		rc = -ENOMEM;
+		goto out;
+	}
+
+	buf->ReparseTag = cpu_to_le32(IO_REPARSE_TAG_SYMLINK);
+	buf->ReparseDataLength = cpu_to_le16(len - sizeof(struct reparse_data_buffer));
+	buf->SubstituteNameOffset = cpu_to_le16(plen);
+	buf->SubstituteNameLength = cpu_to_le16(plen);
+	memcpy(&buf->PathBuffer[plen], path, plen);
+	buf->PrintNameOffset = 0;
+	buf->PrintNameLength = cpu_to_le16(plen);
+	memcpy(buf->PathBuffer, path, plen);
+	buf->Flags = cpu_to_le32(*symname != '/' ? SYMLINK_FLAG_RELATIVE : 0);
+	if (*sym != sep)
+		buf->Flags = cpu_to_le32(SYMLINK_FLAG_RELATIVE);
+
+	convert_delimiter(sym, '/');
+	iov.iov_base = buf;
+	iov.iov_len = len;
+	new = smb2_get_reparse_inode(&data, inode->i_sb, xid,
+				     tcon, full_path, &iov, NULL);
+	if (!IS_ERR(new))
+		d_instantiate(dentry, new);
+	else
+		rc = PTR_ERR(new);
+out:
+	kfree(path);
+	cifs_free_open_info(&data);
+	kfree(buf);
+	return rc;
+}
+
+static int nfs_set_reparse_buf(struct reparse_posix_data *buf,
+			       mode_t mode, dev_t dev,
+			       struct kvec *iov)
+{
+	u64 type;
+	u16 len, dlen;
+
+	len = sizeof(*buf);
+
+	switch ((type = reparse_mode_nfs_type(mode))) {
+	case NFS_SPECFILE_BLK:
+	case NFS_SPECFILE_CHR:
+		dlen = sizeof(__le64);
+		break;
+	case NFS_SPECFILE_FIFO:
+	case NFS_SPECFILE_SOCK:
+		dlen = 0;
+		break;
+	default:
+		return -EOPNOTSUPP;
+	}
+
+	buf->ReparseTag = cpu_to_le32(IO_REPARSE_TAG_NFS);
+	buf->Reserved = 0;
+	buf->InodeType = cpu_to_le64(type);
+	buf->ReparseDataLength = cpu_to_le16(len + dlen -
+					     sizeof(struct reparse_data_buffer));
+	*(__le64 *)buf->DataBuffer = cpu_to_le64(((u64)MAJOR(dev) << 32) |
+						 MINOR(dev));
+	iov->iov_base = buf;
+	iov->iov_len = len + dlen;
+	return 0;
+}
+
+static int mknod_nfs(unsigned int xid, struct inode *inode,
+		     struct dentry *dentry, struct cifs_tcon *tcon,
+		     const char *full_path, umode_t mode, dev_t dev)
+{
+	struct cifs_open_info_data data;
+	struct reparse_posix_data *p;
+	struct inode *new;
+	struct kvec iov;
+	__u8 buf[sizeof(*p) + sizeof(__le64)];
+	int rc;
+
+	p = (struct reparse_posix_data *)buf;
+	rc = nfs_set_reparse_buf(p, mode, dev, &iov);
+	if (rc)
+		return rc;
+
+	data = (struct cifs_open_info_data) {
+		.reparse_point = true,
+		.reparse = { .tag = IO_REPARSE_TAG_NFS, .posix = p, },
+	};
+
+	new = smb2_get_reparse_inode(&data, inode->i_sb, xid,
+				     tcon, full_path, &iov, NULL);
+	if (!IS_ERR(new))
+		d_instantiate(dentry, new);
+	else
+		rc = PTR_ERR(new);
+	cifs_free_open_info(&data);
+	return rc;
+}
+
+static int wsl_set_reparse_buf(struct reparse_data_buffer *buf,
+			       mode_t mode, struct kvec *iov)
+{
+	u32 tag;
+
+	switch ((tag = reparse_mode_wsl_tag(mode))) {
+	case IO_REPARSE_TAG_LX_BLK:
+	case IO_REPARSE_TAG_LX_CHR:
+	case IO_REPARSE_TAG_LX_FIFO:
+	case IO_REPARSE_TAG_AF_UNIX:
+		break;
+	default:
+		return -EOPNOTSUPP;
+	}
+
+	buf->ReparseTag = cpu_to_le32(tag);
+	buf->Reserved = 0;
+	buf->ReparseDataLength = 0;
+	iov->iov_base = buf;
+	iov->iov_len = sizeof(*buf);
+	return 0;
+}
+
+static struct smb2_create_ea_ctx *ea_create_context(u32 dlen, size_t *cc_len)
+{
+	struct smb2_create_ea_ctx *cc;
+
+	*cc_len = round_up(sizeof(*cc) + dlen, 8);
+	cc = kzalloc(*cc_len, GFP_KERNEL);
+	if (!cc)
+		return ERR_PTR(-ENOMEM);
+
+	cc->ctx.NameOffset = cpu_to_le16(offsetof(struct smb2_create_ea_ctx,
+						  name));
+	cc->ctx.NameLength = cpu_to_le16(4);
+	memcpy(cc->name, SMB2_CREATE_EA_BUFFER, strlen(SMB2_CREATE_EA_BUFFER));
+	cc->ctx.DataOffset = cpu_to_le16(offsetof(struct smb2_create_ea_ctx, ea));
+	cc->ctx.DataLength = cpu_to_le32(dlen);
+	return cc;
+}
+
+struct wsl_xattr {
+	const char	*name;
+	__le64		value;
+	u16		size;
+	u32		next;
+};
+
+static int wsl_set_xattrs(struct inode *inode, umode_t _mode,
+			  dev_t _dev, struct kvec *iov)
+{
+	struct smb2_file_full_ea_info *ea;
+	struct smb2_create_ea_ctx *cc;
+	struct smb3_fs_context *ctx = CIFS_SB(inode->i_sb)->ctx;
+	__le64 uid = cpu_to_le64(from_kuid(current_user_ns(), ctx->linux_uid));
+	__le64 gid = cpu_to_le64(from_kgid(current_user_ns(), ctx->linux_gid));
+	__le64 dev = cpu_to_le64(((u64)MINOR(_dev) << 32) | MAJOR(_dev));
+	__le64 mode = cpu_to_le64(_mode);
+	struct wsl_xattr xattrs[] = {
+		{ .name = SMB2_WSL_XATTR_UID,  .value = uid,  .size = SMB2_WSL_XATTR_UID_SIZE, },
+		{ .name = SMB2_WSL_XATTR_GID,  .value = gid,  .size = SMB2_WSL_XATTR_GID_SIZE, },
+		{ .name = SMB2_WSL_XATTR_MODE, .value = mode, .size = SMB2_WSL_XATTR_MODE_SIZE, },
+		{ .name = SMB2_WSL_XATTR_DEV,  .value = dev, .size = SMB2_WSL_XATTR_DEV_SIZE, },
+	};
+	size_t cc_len;
+	u32 dlen = 0, next = 0;
+	int i, num_xattrs;
+	u8 name_size = SMB2_WSL_XATTR_NAME_LEN + 1;
+
+	memset(iov, 0, sizeof(*iov));
+
+	/* Exclude $LXDEV xattr for sockets and fifos */
+	if (S_ISSOCK(_mode) || S_ISFIFO(_mode))
+		num_xattrs = ARRAY_SIZE(xattrs) - 1;
+	else
+		num_xattrs = ARRAY_SIZE(xattrs);
+
+	for (i = 0; i < num_xattrs; i++) {
+		xattrs[i].next = ALIGN(sizeof(*ea) + name_size +
+				       xattrs[i].size, 4);
+		dlen += xattrs[i].next;
+	}
+
+	cc = ea_create_context(dlen, &cc_len);
+	if (IS_ERR(cc))
+		return PTR_ERR(cc);
+
+	ea = &cc->ea;
+	for (i = 0; i < num_xattrs; i++) {
+		ea = (void *)((u8 *)ea + next);
+		next = xattrs[i].next;
+		ea->next_entry_offset = cpu_to_le32(next);
+
+		ea->ea_name_length = name_size - 1;
+		ea->ea_value_length = cpu_to_le16(xattrs[i].size);
+		memcpy(ea->ea_data, xattrs[i].name, name_size);
+		memcpy(&ea->ea_data[name_size],
+		       &xattrs[i].value, xattrs[i].size);
+	}
+	ea->next_entry_offset = 0;
+
+	iov->iov_base = cc;
+	iov->iov_len = cc_len;
+	return 0;
+}
+
+static int mknod_wsl(unsigned int xid, struct inode *inode,
+		     struct dentry *dentry, struct cifs_tcon *tcon,
+		     const char *full_path, umode_t mode, dev_t dev)
+{
+	struct cifs_open_info_data data;
+	struct reparse_data_buffer buf;
+	struct smb2_create_ea_ctx *cc;
+	struct inode *new;
+	unsigned int len;
+	struct kvec reparse_iov, xattr_iov;
+	int rc;
+
+	rc = wsl_set_reparse_buf(&buf, mode, &reparse_iov);
+	if (rc)
+		return rc;
+
+	rc = wsl_set_xattrs(inode, mode, dev, &xattr_iov);
+	if (rc)
+		return rc;
+
+	data = (struct cifs_open_info_data) {
+		.reparse_point = true,
+		.reparse = { .tag = le32_to_cpu(buf.ReparseTag), .buf = &buf, },
+	};
+
+	cc = xattr_iov.iov_base;
+	len = le32_to_cpu(cc->ctx.DataLength);
+	memcpy(data.wsl.eas, &cc->ea, len);
+	data.wsl.eas_len = len;
+
+	new = smb2_get_reparse_inode(&data, inode->i_sb,
+				     xid, tcon, full_path,
+				     &reparse_iov, &xattr_iov);
+	if (!IS_ERR(new))
+		d_instantiate(dentry, new);
+	else
+		rc = PTR_ERR(new);
+	cifs_free_open_info(&data);
+	kfree(xattr_iov.iov_base);
+	return rc;
+}
+
+int smb2_mknod_reparse(unsigned int xid, struct inode *inode,
+		       struct dentry *dentry, struct cifs_tcon *tcon,
+		       const char *full_path, umode_t mode, dev_t dev)
+{
+	struct smb3_fs_context *ctx = CIFS_SB(inode->i_sb)->ctx;
+	int rc = -EOPNOTSUPP;
+
+	switch (ctx->reparse_type) {
+	case CIFS_REPARSE_TYPE_NFS:
+		rc = mknod_nfs(xid, inode, dentry, tcon, full_path, mode, dev);
+		break;
+	case CIFS_REPARSE_TYPE_WSL:
+		rc = mknod_wsl(xid, inode, dentry, tcon, full_path, mode, dev);
+		break;
+	}
+	return rc;
+}
+
+/* See MS-FSCC 2.1.2.6 for the 'NFS' style reparse tags */
+static int parse_reparse_posix(struct reparse_posix_data *buf,
+			       struct cifs_sb_info *cifs_sb,
+			       struct cifs_open_info_data *data)
+{
+	unsigned int len;
+	u64 type;
+
+	switch ((type = le64_to_cpu(buf->InodeType))) {
+	case NFS_SPECFILE_LNK:
+		len = le16_to_cpu(buf->ReparseDataLength);
+		data->symlink_target = cifs_strndup_from_utf16(buf->DataBuffer,
+							       len, true,
+							       cifs_sb->local_nls);
+		if (!data->symlink_target)
+			return -ENOMEM;
+		convert_delimiter(data->symlink_target, '/');
+		cifs_dbg(FYI, "%s: target path: %s\n",
+			 __func__, data->symlink_target);
+		break;
+	case NFS_SPECFILE_CHR:
+	case NFS_SPECFILE_BLK:
+	case NFS_SPECFILE_FIFO:
+	case NFS_SPECFILE_SOCK:
+		break;
+	default:
+		cifs_dbg(VFS, "%s: unhandled inode type: 0x%llx\n",
+			 __func__, type);
+		return -EOPNOTSUPP;
+	}
+	return 0;
+}
+
+static int parse_reparse_symlink(struct reparse_symlink_data_buffer *sym,
+				 u32 plen, bool unicode,
+				 struct cifs_sb_info *cifs_sb,
+				 struct cifs_open_info_data *data)
+{
+	unsigned int len;
+	unsigned int offs;
+
+	/* We handle Symbolic Link reparse tag here. See: MS-FSCC 2.1.2.4 */
+
+	offs = le16_to_cpu(sym->SubstituteNameOffset);
+	len = le16_to_cpu(sym->SubstituteNameLength);
+	if (offs + 20 > plen || offs + len + 20 > plen) {
+		cifs_dbg(VFS, "srv returned malformed symlink buffer\n");
+		return -EIO;
+	}
+
+	data->symlink_target = cifs_strndup_from_utf16(sym->PathBuffer + offs,
+						       len, unicode,
+						       cifs_sb->local_nls);
+	if (!data->symlink_target)
+		return -ENOMEM;
+
+	convert_delimiter(data->symlink_target, '/');
+	cifs_dbg(FYI, "%s: target path: %s\n", __func__, data->symlink_target);
+
+	return 0;
+}
+
+int parse_reparse_point(struct reparse_data_buffer *buf,
+			u32 plen, struct cifs_sb_info *cifs_sb,
+			bool unicode, struct cifs_open_info_data *data)
+{
+	data->reparse.buf = buf;
+
+	/* See MS-FSCC 2.1.2 */
+	switch (le32_to_cpu(buf->ReparseTag)) {
+	case IO_REPARSE_TAG_NFS:
+		return parse_reparse_posix((struct reparse_posix_data *)buf,
+					   cifs_sb, data);
+	case IO_REPARSE_TAG_SYMLINK:
+		return parse_reparse_symlink(
+			(struct reparse_symlink_data_buffer *)buf,
+			plen, unicode, cifs_sb, data);
+	case IO_REPARSE_TAG_LX_SYMLINK:
+	case IO_REPARSE_TAG_AF_UNIX:
+	case IO_REPARSE_TAG_LX_FIFO:
+	case IO_REPARSE_TAG_LX_CHR:
+	case IO_REPARSE_TAG_LX_BLK:
+		return 0;
+	default:
+		cifs_dbg(VFS, "%s: unhandled reparse tag: 0x%08x\n",
+			 __func__, le32_to_cpu(buf->ReparseTag));
+		return -EOPNOTSUPP;
+	}
+}
+
+int smb2_parse_reparse_point(struct cifs_sb_info *cifs_sb,
+			     struct kvec *rsp_iov,
+			     struct cifs_open_info_data *data)
+{
+	struct reparse_data_buffer *buf;
+	struct smb2_ioctl_rsp *io = rsp_iov->iov_base;
+	u32 plen = le32_to_cpu(io->OutputCount);
+
+	buf = (struct reparse_data_buffer *)((u8 *)io +
+					     le32_to_cpu(io->OutputOffset));
+	return parse_reparse_point(buf, plen, cifs_sb, true, data);
+}
+
+static void wsl_to_fattr(struct cifs_open_info_data *data,
+			 struct cifs_sb_info *cifs_sb,
+			 u32 tag, struct cifs_fattr *fattr)
+{
+	struct smb2_file_full_ea_info *ea;
+	u32 next = 0;
+
+	switch (tag) {
+	case IO_REPARSE_TAG_LX_SYMLINK:
+		fattr->cf_mode |= S_IFLNK;
+		break;
+	case IO_REPARSE_TAG_LX_FIFO:
+		fattr->cf_mode |= S_IFIFO;
+		break;
+	case IO_REPARSE_TAG_AF_UNIX:
+		fattr->cf_mode |= S_IFSOCK;
+		break;
+	case IO_REPARSE_TAG_LX_CHR:
+		fattr->cf_mode |= S_IFCHR;
+		break;
+	case IO_REPARSE_TAG_LX_BLK:
+		fattr->cf_mode |= S_IFBLK;
+		break;
+	}
+
+	if (!data->wsl.eas_len)
+		goto out;
+
+	ea = (struct smb2_file_full_ea_info *)data->wsl.eas;
+	do {
+		const char *name;
+		void *v;
+		u8 nlen;
+
+		ea = (void *)((u8 *)ea + next);
+		next = le32_to_cpu(ea->next_entry_offset);
+		if (!le16_to_cpu(ea->ea_value_length))
+			continue;
+
+		name = ea->ea_data;
+		nlen = ea->ea_name_length;
+		v = (void *)((u8 *)ea->ea_data + ea->ea_name_length + 1);
+
+		if (!strncmp(name, SMB2_WSL_XATTR_UID, nlen))
+			fattr->cf_uid = wsl_make_kuid(cifs_sb, v);
+		else if (!strncmp(name, SMB2_WSL_XATTR_GID, nlen))
+			fattr->cf_gid = wsl_make_kgid(cifs_sb, v);
+		else if (!strncmp(name, SMB2_WSL_XATTR_MODE, nlen))
+			fattr->cf_mode = (umode_t)le32_to_cpu(*(__le32 *)v);
+		else if (!strncmp(name, SMB2_WSL_XATTR_DEV, nlen))
+			fattr->cf_rdev = wsl_mkdev(v);
+	} while (next);
+out:
+	fattr->cf_dtype = S_DT(fattr->cf_mode);
+}
+
+bool cifs_reparse_point_to_fattr(struct cifs_sb_info *cifs_sb,
+				 struct cifs_fattr *fattr,
+				 struct cifs_open_info_data *data)
+{
+	struct reparse_posix_data *buf = data->reparse.posix;
+	u32 tag = data->reparse.tag;
+
+	if (tag == IO_REPARSE_TAG_NFS && buf) {
+		switch (le64_to_cpu(buf->InodeType)) {
+		case NFS_SPECFILE_CHR:
+			fattr->cf_mode |= S_IFCHR;
+			fattr->cf_rdev = reparse_nfs_mkdev(buf);
+			break;
+		case NFS_SPECFILE_BLK:
+			fattr->cf_mode |= S_IFBLK;
+			fattr->cf_rdev = reparse_nfs_mkdev(buf);
+			break;
+		case NFS_SPECFILE_FIFO:
+			fattr->cf_mode |= S_IFIFO;
+			break;
+		case NFS_SPECFILE_SOCK:
+			fattr->cf_mode |= S_IFSOCK;
+			break;
+		case NFS_SPECFILE_LNK:
+			fattr->cf_mode |= S_IFLNK;
+			break;
+		default:
+			WARN_ON_ONCE(1);
+			return false;
+		}
+		goto out;
+	}
+
+	switch (tag) {
+	case IO_REPARSE_TAG_INTERNAL:
+		if (!(fattr->cf_cifsattrs & ATTR_DIRECTORY))
+			return false;
+		fallthrough;
+	case IO_REPARSE_TAG_DFS:
+	case IO_REPARSE_TAG_DFSR:
+	case IO_REPARSE_TAG_MOUNT_POINT:
+		/* See cifs_create_junction_fattr() */
+		fattr->cf_mode = S_IFDIR | 0711;
+		break;
+	case IO_REPARSE_TAG_LX_SYMLINK:
+	case IO_REPARSE_TAG_LX_FIFO:
+	case IO_REPARSE_TAG_AF_UNIX:
+	case IO_REPARSE_TAG_LX_CHR:
+	case IO_REPARSE_TAG_LX_BLK:
+		wsl_to_fattr(data, cifs_sb, tag, fattr);
+		break;
+	case 0: /* SMB1 symlink */
+	case IO_REPARSE_TAG_SYMLINK:
+	case IO_REPARSE_TAG_NFS:
+		fattr->cf_mode |= S_IFLNK;
+		break;
+	default:
+		return false;
+	}
+out:
+	fattr->cf_dtype = S_DT(fattr->cf_mode);
+	return true;
+}
diff --git a/fs/smb/client/reparse.h b/fs/smb/client/reparse.h
new file mode 100644
index 0000000000..2c0644bc4e
--- /dev/null
+++ b/fs/smb/client/reparse.h
@@ -0,0 +1,128 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (c) 2024 Paulo Alcantara <pc@manguebit.com>
+ */
+
+#ifndef _CIFS_REPARSE_H
+#define _CIFS_REPARSE_H
+
+#include <linux/fs.h>
+#include <linux/stat.h>
+#include <linux/uidgid.h>
+#include "fs_context.h"
+#include "cifsglob.h"
+
+/*
+ * Used only by cifs.ko to ignore reparse points from files when client or
+ * server doesn't support FSCTL_GET_REPARSE_POINT.
+ */
+#define IO_REPARSE_TAG_INTERNAL ((__u32)~0U)
+
+static inline dev_t reparse_nfs_mkdev(struct reparse_posix_data *buf)
+{
+	u64 v = le64_to_cpu(*(__le64 *)buf->DataBuffer);
+
+	return MKDEV(v >> 32, v & 0xffffffff);
+}
+
+static inline dev_t wsl_mkdev(void *ptr)
+{
+	u64 v = le64_to_cpu(*(__le64 *)ptr);
+
+	return MKDEV(v & 0xffffffff, v >> 32);
+}
+
+static inline kuid_t wsl_make_kuid(struct cifs_sb_info *cifs_sb,
+				   void *ptr)
+{
+	u32 uid = le32_to_cpu(*(__le32 *)ptr);
+
+	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_OVERR_UID)
+		return cifs_sb->ctx->linux_uid;
+	return make_kuid(current_user_ns(), uid);
+}
+
+static inline kgid_t wsl_make_kgid(struct cifs_sb_info *cifs_sb,
+				   void *ptr)
+{
+	u32 gid = le32_to_cpu(*(__le32 *)ptr);
+
+	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_OVERR_GID)
+		return cifs_sb->ctx->linux_gid;
+	return make_kgid(current_user_ns(), gid);
+}
+
+static inline u64 reparse_mode_nfs_type(mode_t mode)
+{
+	switch (mode & S_IFMT) {
+	case S_IFBLK: return NFS_SPECFILE_BLK;
+	case S_IFCHR: return NFS_SPECFILE_CHR;
+	case S_IFIFO: return NFS_SPECFILE_FIFO;
+	case S_IFSOCK: return NFS_SPECFILE_SOCK;
+	}
+	return 0;
+}
+
+static inline u32 reparse_mode_wsl_tag(mode_t mode)
+{
+	switch (mode & S_IFMT) {
+	case S_IFBLK: return IO_REPARSE_TAG_LX_BLK;
+	case S_IFCHR: return IO_REPARSE_TAG_LX_CHR;
+	case S_IFIFO: return IO_REPARSE_TAG_LX_FIFO;
+	case S_IFSOCK: return IO_REPARSE_TAG_AF_UNIX;
+	}
+	return 0;
+}
+
+/*
+ * Match a reparse point inode if reparse tag and ctime haven't changed.
+ *
+ * Windows Server updates ctime of reparse points when their data have changed.
+ * The server doesn't allow changing reparse tags from existing reparse points,
+ * though it's worth checking.
+ */
+static inline bool reparse_inode_match(struct inode *inode,
+				       struct cifs_fattr *fattr)
+{
+	struct cifsInodeInfo *cinode = CIFS_I(inode);
+	struct timespec64 ctime = inode_get_ctime(inode);
+
+	/*
+	 * Do not match reparse tags when client or server doesn't support
+	 * FSCTL_GET_REPARSE_POINT.  @fattr->cf_cifstag should contain correct
+	 * reparse tag from query dir response but the client won't be able to
+	 * read the reparse point data anyway.  This spares us a revalidation.
+	 */
+	if (cinode->reparse_tag != IO_REPARSE_TAG_INTERNAL &&
+	    cinode->reparse_tag != fattr->cf_cifstag)
+		return false;
+	return (cinode->cifsAttrs & ATTR_REPARSE) &&
+		timespec64_equal(&ctime, &fattr->cf_ctime);
+}
+
+static inline bool cifs_open_data_reparse(struct cifs_open_info_data *data)
+{
+	struct smb2_file_all_info *fi = &data->fi;
+	u32 attrs = le32_to_cpu(fi->Attributes);
+	bool ret;
+
+	ret = data->reparse_point || (attrs & ATTR_REPARSE);
+	if (ret)
+		attrs |= ATTR_REPARSE;
+	fi->Attributes = cpu_to_le32(attrs);
+	return ret;
+}
+
+bool cifs_reparse_point_to_fattr(struct cifs_sb_info *cifs_sb,
+				 struct cifs_fattr *fattr,
+				 struct cifs_open_info_data *data);
+int smb2_create_reparse_symlink(const unsigned int xid, struct inode *inode,
+				struct dentry *dentry, struct cifs_tcon *tcon,
+				const char *full_path, const char *symname);
+int smb2_mknod_reparse(unsigned int xid, struct inode *inode,
+		       struct dentry *dentry, struct cifs_tcon *tcon,
+		       const char *full_path, umode_t mode, dev_t dev);
+int smb2_parse_reparse_point(struct cifs_sb_info *cifs_sb, struct kvec *rsp_iov,
+			     struct cifs_open_info_data *data);
+
+#endif /* _CIFS_REPARSE_H */
diff --git a/fs/smb/client/smb2file.c b/fs/smb/client/smb2file.c
index e0ee96d69d..c23478ab1c 100644
--- a/fs/smb/client/smb2file.c
+++ b/fs/smb/client/smb2file.c
@@ -228,7 +228,7 @@ smb2_unlock_range(struct cifsFileInfo *cfile, struct file_lock *flock,
 			 * flock and OFD lock are associated with an open
 			 * file description, not the process.
 			 */
-			if (!(flock->fl_flags & (FL_FLOCK | FL_OFDLCK)))
+			if (!(flock->c.flc_flags & (FL_FLOCK | FL_OFDLCK)))
 				continue;
 		if (cinode->can_cache_brlcks) {
 			/*
diff --git a/fs/smb/client/smb2glob.h b/fs/smb/client/smb2glob.h
index a0c156996f..2466e61551 100644
--- a/fs/smb/client/smb2glob.h
+++ b/fs/smb/client/smb2glob.h
@@ -36,7 +36,8 @@ enum smb2_compound_ops {
 	SMB2_OP_RMDIR,
 	SMB2_OP_POSIX_QUERY_INFO,
 	SMB2_OP_SET_REPARSE,
-	SMB2_OP_GET_REPARSE
+	SMB2_OP_GET_REPARSE,
+	SMB2_OP_QUERY_WSL_EA,
 };
 
 /* Used when constructing chained read requests. */
diff --git a/fs/smb/client/smb2inode.c b/fs/smb/client/smb2inode.c
index 05818cd6d9..062b86a493 100644
--- a/fs/smb/client/smb2inode.c
+++ b/fs/smb/client/smb2inode.c
@@ -85,6 +85,82 @@ static int parse_posix_sids(struct cifs_open_info_data *data,
 	return 0;
 }
 
+struct wsl_query_ea {
+	__le32	next;
+	__u8	name_len;
+	__u8	name[SMB2_WSL_XATTR_NAME_LEN + 1];
+} __packed;
+
+#define NEXT_OFF cpu_to_le32(sizeof(struct wsl_query_ea))
+
+static const struct wsl_query_ea wsl_query_eas[] = {
+	{ .next = NEXT_OFF, .name_len = SMB2_WSL_XATTR_NAME_LEN, .name = SMB2_WSL_XATTR_UID, },
+	{ .next = NEXT_OFF, .name_len = SMB2_WSL_XATTR_NAME_LEN, .name = SMB2_WSL_XATTR_GID, },
+	{ .next = NEXT_OFF, .name_len = SMB2_WSL_XATTR_NAME_LEN, .name = SMB2_WSL_XATTR_MODE, },
+	{ .next = 0,        .name_len = SMB2_WSL_XATTR_NAME_LEN, .name = SMB2_WSL_XATTR_DEV, },
+};
+
+static int check_wsl_eas(struct kvec *rsp_iov)
+{
+	struct smb2_file_full_ea_info *ea;
+	struct smb2_query_info_rsp *rsp = rsp_iov->iov_base;
+	unsigned long addr;
+	u32 outlen, next;
+	u16 vlen;
+	u8 nlen;
+	u8 *end;
+
+	outlen = le32_to_cpu(rsp->OutputBufferLength);
+	if (outlen < SMB2_WSL_MIN_QUERY_EA_RESP_SIZE ||
+	    outlen > SMB2_WSL_MAX_QUERY_EA_RESP_SIZE)
+		return -EINVAL;
+
+	ea = (void *)((u8 *)rsp_iov->iov_base +
+		      le16_to_cpu(rsp->OutputBufferOffset));
+	end = (u8 *)rsp_iov->iov_base + rsp_iov->iov_len;
+	for (;;) {
+		if ((u8 *)ea > end - sizeof(*ea))
+			return -EINVAL;
+
+		nlen = ea->ea_name_length;
+		vlen = le16_to_cpu(ea->ea_value_length);
+		if (nlen != SMB2_WSL_XATTR_NAME_LEN ||
+		    (u8 *)ea + nlen + 1 + vlen > end)
+			return -EINVAL;
+
+		switch (vlen) {
+		case 4:
+			if (strncmp(ea->ea_data, SMB2_WSL_XATTR_UID, nlen) &&
+			    strncmp(ea->ea_data, SMB2_WSL_XATTR_GID, nlen) &&
+			    strncmp(ea->ea_data, SMB2_WSL_XATTR_MODE, nlen))
+				return -EINVAL;
+			break;
+		case 8:
+			if (strncmp(ea->ea_data, SMB2_WSL_XATTR_DEV, nlen))
+				return -EINVAL;
+			break;
+		case 0:
+			if (!strncmp(ea->ea_data, SMB2_WSL_XATTR_UID, nlen) ||
+			    !strncmp(ea->ea_data, SMB2_WSL_XATTR_GID, nlen) ||
+			    !strncmp(ea->ea_data, SMB2_WSL_XATTR_MODE, nlen) ||
+			    !strncmp(ea->ea_data, SMB2_WSL_XATTR_DEV, nlen))
+				break;
+			fallthrough;
+		default:
+			return -EINVAL;
+		}
+
+		next = le32_to_cpu(ea->next_entry_offset);
+		if (!next)
+			break;
+		if (!IS_ALIGNED(next, 4) ||
+		    check_add_overflow((unsigned long)ea, next, &addr))
+			return -EINVAL;
+		ea = (void *)addr;
+	}
+	return 0;
+}
+
 /*
  * note: If cfile is passed, the reference to it is dropped here.
  * So make sure that you do not reuse cfile after return from this func.
@@ -95,10 +171,9 @@ static int parse_posix_sids(struct cifs_open_info_data *data,
  */
 static int smb2_compound_op(const unsigned int xid, struct cifs_tcon *tcon,
 			    struct cifs_sb_info *cifs_sb, const char *full_path,
-			    __u32 desired_access, __u32 create_disposition,
-			    __u32 create_options, umode_t mode, struct kvec *in_iov,
+			    struct cifs_open_parms *oparms, struct kvec *in_iov,
 			    int *cmds, int num_cmds, struct cifsFileInfo *cfile,
-			    struct kvec *out_iov, int *out_buftype)
+			    struct kvec *out_iov, int *out_buftype, struct dentry *dentry)
 {
 
 	struct reparse_data_buffer *rbuf;
@@ -115,11 +190,12 @@ static int smb2_compound_op(const unsigned int xid, struct cifs_tcon *tcon,
 	int resp_buftype[MAX_COMPOUND];
 	struct smb2_query_info_rsp *qi_rsp = NULL;
 	struct cifs_open_info_data *idata;
+	struct inode *inode = NULL;
 	int flags = 0;
 	__u8 delete_pending[8] = {1, 0, 0, 0, 0, 0, 0, 0};
 	unsigned int size[2];
 	void *data[2];
-	int len;
+	unsigned int len;
 	int retries = 0, cur_sleep = 1;
 
 replay_again:
@@ -152,16 +228,28 @@ replay_again:
 		goto finished;
 	}
 
-	vars->oparms = (struct cifs_open_parms) {
-		.tcon = tcon,
-		.path = full_path,
-		.desired_access = desired_access,
-		.disposition = create_disposition,
-		.create_options = cifs_create_options(cifs_sb, create_options),
-		.fid = &fid,
-		.mode = mode,
-		.cifs_sb = cifs_sb,
-	};
+	/* if there is an existing lease, reuse it */
+
+	/*
+	 * note: files with hardlinks cause unexpected behaviour. As per MS-SMB2,
+	 * lease keys are associated with the filepath. We are maintaining lease keys
+	 * with the inode on the client. If the file has hardlinks, it is possible
+	 * that the lease for a file be reused for an operation on its hardlink or
+	 * vice versa.
+	 * As a workaround, send request using an existing lease key and if the server
+	 * returns STATUS_INVALID_PARAMETER, which maps to EINVAL, send the request
+	 * again without the lease.
+	 */
+	if (dentry) {
+		inode = d_inode(dentry);
+		if (CIFS_I(inode)->lease_granted && server->ops->get_lease_key) {
+			oplock = SMB2_OPLOCK_LEVEL_LEASE;
+			server->ops->get_lease_key(inode, &fid);
+		}
+	}
+
+	vars->oparms = *oparms;
+	vars->oparms.fid = &fid;
 
 	rqst[num_rqst].rq_iov = &vars->open_iov[0];
 	rqst[num_rqst].rq_nvec = SMB2_CREATE_IOV_SIZE;
@@ -202,14 +290,13 @@ replay_again:
 							  SMB2_O_INFO_FILE, 0,
 							  sizeof(struct smb2_file_all_info) +
 							  PATH_MAX * 2, 0, NULL);
-				if (!rc) {
-					smb2_set_next_command(tcon, &rqst[num_rqst]);
-					smb2_set_related(&rqst[num_rqst]);
-				}
 			}
-
-			if (rc)
+			if (!rc && (!cfile || num_rqst > 1)) {
+				smb2_set_next_command(tcon, &rqst[num_rqst]);
+				smb2_set_related(&rqst[num_rqst]);
+			} else if (rc) {
 				goto finished;
+			}
 			num_rqst++;
 			trace_smb3_query_info_compound_enter(xid, ses->Suid,
 							     tcon->tid, full_path);
@@ -239,14 +326,13 @@ replay_again:
 							  sizeof(struct smb311_posix_qinfo *) +
 							  (PATH_MAX * 2) +
 							  (sizeof(struct cifs_sid) * 2), 0, NULL);
-				if (!rc) {
-					smb2_set_next_command(tcon, &rqst[num_rqst]);
-					smb2_set_related(&rqst[num_rqst]);
-				}
 			}
-
-			if (rc)
+			if (!rc && (!cfile || num_rqst > 1)) {
+				smb2_set_next_command(tcon, &rqst[num_rqst]);
+				smb2_set_related(&rqst[num_rqst]);
+			} else if (rc) {
 				goto finished;
+			}
 			num_rqst++;
 			trace_smb3_posix_query_info_compound_enter(xid, ses->Suid,
 								   tcon->tid, full_path);
@@ -304,13 +390,13 @@ replay_again:
 							FILE_END_OF_FILE_INFORMATION,
 							SMB2_O_INFO_FILE, 0,
 							data, size);
-				if (!rc) {
-					smb2_set_next_command(tcon, &rqst[num_rqst]);
-					smb2_set_related(&rqst[num_rqst]);
-				}
 			}
-			if (rc)
+			if (!rc && (!cfile || num_rqst > 1)) {
+				smb2_set_next_command(tcon, &rqst[num_rqst]);
+				smb2_set_related(&rqst[num_rqst]);
+			} else if (rc) {
 				goto finished;
+			}
 			num_rqst++;
 			trace_smb3_set_eof_enter(xid, ses->Suid, tcon->tid, full_path);
 			break;
@@ -335,14 +421,13 @@ replay_again:
 							COMPOUND_FID, current->tgid,
 							FILE_BASIC_INFORMATION,
 							SMB2_O_INFO_FILE, 0, data, size);
-				if (!rc) {
-					smb2_set_next_command(tcon, &rqst[num_rqst]);
-					smb2_set_related(&rqst[num_rqst]);
-				}
 			}
-
-			if (rc)
+			if (!rc && (!cfile || num_rqst > 1)) {
+				smb2_set_next_command(tcon, &rqst[num_rqst]);
+				smb2_set_related(&rqst[num_rqst]);
+			} else if (rc) {
 				goto finished;
+			}
 			num_rqst++;
 			trace_smb3_set_info_compound_enter(xid, ses->Suid,
 							   tcon->tid, full_path);
@@ -376,13 +461,13 @@ replay_again:
 							COMPOUND_FID, COMPOUND_FID,
 							current->tgid, FILE_RENAME_INFORMATION,
 							SMB2_O_INFO_FILE, 0, data, size);
-				if (!rc) {
-					smb2_set_next_command(tcon, &rqst[num_rqst]);
-					smb2_set_related(&rqst[num_rqst]);
-				}
 			}
-			if (rc)
+			if (!rc && (!cfile || num_rqst > 1)) {
+				smb2_set_next_command(tcon, &rqst[num_rqst]);
+				smb2_set_related(&rqst[num_rqst]);
+			} else if (rc) {
 				goto finished;
+			}
 			num_rqst++;
 			trace_smb3_rename_enter(xid, ses->Suid, tcon->tid, full_path);
 			break;
@@ -417,15 +502,27 @@ replay_again:
 			rqst[num_rqst].rq_iov = vars->io_iov;
 			rqst[num_rqst].rq_nvec = ARRAY_SIZE(vars->io_iov);
 
-			rc = SMB2_ioctl_init(tcon, server, &rqst[num_rqst],
-					     COMPOUND_FID, COMPOUND_FID,
-					     FSCTL_SET_REPARSE_POINT,
-					     in_iov[i].iov_base,
-					     in_iov[i].iov_len, 0);
-			if (rc)
+			if (cfile) {
+				rc = SMB2_ioctl_init(tcon, server, &rqst[num_rqst],
+						     cfile->fid.persistent_fid,
+						     cfile->fid.volatile_fid,
+						     FSCTL_SET_REPARSE_POINT,
+						     in_iov[i].iov_base,
+						     in_iov[i].iov_len, 0);
+			} else {
+				rc = SMB2_ioctl_init(tcon, server, &rqst[num_rqst],
+						     COMPOUND_FID, COMPOUND_FID,
+						     FSCTL_SET_REPARSE_POINT,
+						     in_iov[i].iov_base,
+						     in_iov[i].iov_len, 0);
+			}
+			if (!rc && (!cfile || num_rqst > 1)) {
+				smb2_set_next_command(tcon, &rqst[num_rqst]);
+				smb2_set_related(&rqst[num_rqst]);
+			} else if (rc) {
 				goto finished;
-			smb2_set_next_command(tcon, &rqst[num_rqst]);
-			smb2_set_related(&rqst[num_rqst++]);
+			}
+			num_rqst++;
 			trace_smb3_set_reparse_compound_enter(xid, ses->Suid,
 							      tcon->tid, full_path);
 			break;
@@ -433,17 +530,61 @@ replay_again:
 			rqst[num_rqst].rq_iov = vars->io_iov;
 			rqst[num_rqst].rq_nvec = ARRAY_SIZE(vars->io_iov);
 
-			rc = SMB2_ioctl_init(tcon, server, &rqst[num_rqst],
-					     COMPOUND_FID, COMPOUND_FID,
-					     FSCTL_GET_REPARSE_POINT,
-					     NULL, 0, CIFSMaxBufSize);
-			if (rc)
+			if (cfile) {
+				rc = SMB2_ioctl_init(tcon, server, &rqst[num_rqst],
+						     cfile->fid.persistent_fid,
+						     cfile->fid.volatile_fid,
+						     FSCTL_GET_REPARSE_POINT,
+						     NULL, 0, CIFSMaxBufSize);
+			} else {
+				rc = SMB2_ioctl_init(tcon, server, &rqst[num_rqst],
+						     COMPOUND_FID, COMPOUND_FID,
+						     FSCTL_GET_REPARSE_POINT,
+						     NULL, 0, CIFSMaxBufSize);
+			}
+			if (!rc && (!cfile || num_rqst > 1)) {
+				smb2_set_next_command(tcon, &rqst[num_rqst]);
+				smb2_set_related(&rqst[num_rqst]);
+			} else if (rc) {
 				goto finished;
-			smb2_set_next_command(tcon, &rqst[num_rqst]);
-			smb2_set_related(&rqst[num_rqst++]);
+			}
+			num_rqst++;
 			trace_smb3_get_reparse_compound_enter(xid, ses->Suid,
 							      tcon->tid, full_path);
 			break;
+		case SMB2_OP_QUERY_WSL_EA:
+			rqst[num_rqst].rq_iov = &vars->ea_iov;
+			rqst[num_rqst].rq_nvec = 1;
+
+			if (cfile) {
+				rc = SMB2_query_info_init(tcon, server,
+							  &rqst[num_rqst],
+							  cfile->fid.persistent_fid,
+							  cfile->fid.volatile_fid,
+							  FILE_FULL_EA_INFORMATION,
+							  SMB2_O_INFO_FILE, 0,
+							  SMB2_WSL_MAX_QUERY_EA_RESP_SIZE,
+							  sizeof(wsl_query_eas),
+							  (void *)wsl_query_eas);
+			} else {
+				rc = SMB2_query_info_init(tcon, server,
+							  &rqst[num_rqst],
+							  COMPOUND_FID,
+							  COMPOUND_FID,
+							  FILE_FULL_EA_INFORMATION,
+							  SMB2_O_INFO_FILE, 0,
+							  SMB2_WSL_MAX_QUERY_EA_RESP_SIZE,
+							  sizeof(wsl_query_eas),
+							  (void *)wsl_query_eas);
+			}
+			if (!rc && (!cfile || num_rqst > 1)) {
+				smb2_set_next_command(tcon, &rqst[num_rqst]);
+				smb2_set_related(&rqst[num_rqst]);
+			} else if (rc) {
+				goto finished;
+			}
+			num_rqst++;
+			break;
 		default:
 			cifs_dbg(VFS, "Invalid command\n");
 			rc = -EINVAL;
@@ -551,8 +692,15 @@ finished:
 		case SMB2_OP_DELETE:
 			if (rc)
 				trace_smb3_delete_err(xid,  ses->Suid, tcon->tid, rc);
-			else
+			else {
+				/*
+				 * If dentry (hence, inode) is NULL, lease break is going to
+				 * take care of degrading leases on handles for deleted files.
+				 */
+				if (inode)
+					cifs_mark_open_handles_for_deleted_file(inode, full_path);
 				trace_smb3_delete_done(xid, ses->Suid, tcon->tid);
+			}
 			break;
 		case SMB2_OP_MKDIR:
 			if (rc)
@@ -626,11 +774,32 @@ finished:
 				memset(iov, 0, sizeof(*iov));
 				resp_buftype[i + 1] = CIFS_NO_BUFFER;
 			} else {
-				trace_smb3_set_reparse_compound_err(xid,  ses->Suid,
+				trace_smb3_set_reparse_compound_err(xid, ses->Suid,
 								    tcon->tid, rc);
 			}
 			SMB2_ioctl_free(&rqst[num_rqst++]);
 			break;
+		case SMB2_OP_QUERY_WSL_EA:
+			if (!rc) {
+				idata = in_iov[i].iov_base;
+				qi_rsp = rsp_iov[i + 1].iov_base;
+				data[0] = (u8 *)qi_rsp + le16_to_cpu(qi_rsp->OutputBufferOffset);
+				size[0] = le32_to_cpu(qi_rsp->OutputBufferLength);
+				rc = check_wsl_eas(&rsp_iov[i + 1]);
+				if (!rc) {
+					memcpy(idata->wsl.eas, data[0], size[0]);
+					idata->wsl.eas_len = size[0];
+				}
+			}
+			if (!rc) {
+				trace_smb3_query_wsl_ea_compound_done(xid, ses->Suid,
+								      tcon->tid);
+			} else {
+				trace_smb3_query_wsl_ea_compound_err(xid, ses->Suid,
+								     tcon->tid, rc);
+			}
+			SMB2_query_info_free(&rqst[num_rqst++]);
+			break;
 		}
 	}
 	SMB2_close_free(&rqst[num_rqst]);
@@ -693,15 +862,16 @@ int smb2_query_path_info(const unsigned int xid,
 			 const char *full_path,
 			 struct cifs_open_info_data *data)
 {
+	struct cifs_open_parms oparms;
 	__u32 create_options = 0;
 	struct cifsFileInfo *cfile;
 	struct cached_fid *cfid = NULL;
 	struct smb2_hdr *hdr;
-	struct kvec in_iov[2], out_iov[3] = {};
+	struct kvec in_iov[3], out_iov[3] = {};
 	int out_buftype[3] = {};
-	int cmds[2];
+	int cmds[3];
 	bool islink;
-	int i, num_cmds;
+	int i, num_cmds = 0;
 	int rc, rc2;
 
 	data->adjust_tz = false;
@@ -734,20 +904,22 @@ int smb2_query_path_info(const unsigned int xid,
 			close_cached_dir(cfid);
 			return rc;
 		}
-		cmds[0] = SMB2_OP_QUERY_INFO;
+		cmds[num_cmds++] = SMB2_OP_QUERY_INFO;
 	} else {
-		cmds[0] = SMB2_OP_POSIX_QUERY_INFO;
+		cmds[num_cmds++] = SMB2_OP_POSIX_QUERY_INFO;
 	}
 
 	in_iov[0].iov_base = data;
 	in_iov[0].iov_len = sizeof(*data);
 	in_iov[1] = in_iov[0];
+	in_iov[2] = in_iov[0];
 
 	cifs_get_readable_path(tcon, full_path, &cfile);
+	oparms = CIFS_OPARMS(cifs_sb, tcon, full_path, FILE_READ_ATTRIBUTES,
+			     FILE_OPEN, create_options, ACL_NO_MODE);
 	rc = smb2_compound_op(xid, tcon, cifs_sb, full_path,
-			      FILE_READ_ATTRIBUTES, FILE_OPEN,
-			      create_options, ACL_NO_MODE, in_iov,
-			      cmds, 1, cfile, out_iov, out_buftype);
+			      &oparms, in_iov, cmds, num_cmds,
+			      cfile, out_iov, out_buftype, NULL);
 	hdr = out_iov[0].iov_base;
 	/*
 	 * If first iov is unset, then SMB session was dropped or we've got a
@@ -758,6 +930,8 @@ int smb2_query_path_info(const unsigned int xid,
 
 	switch (rc) {
 	case 0:
+		rc = parse_create_response(data, cifs_sb, &out_iov[0]);
+		break;
 	case -EOPNOTSUPP:
 		/*
 		 * BB TODO: When support for special files added to Samba
@@ -767,19 +941,22 @@ int smb2_query_path_info(const unsigned int xid,
 		if (rc || !data->reparse_point)
 			goto out;
 
-		if (data->reparse.tag == IO_REPARSE_TAG_SYMLINK) {
-			/* symlink already parsed in create response */
-			num_cmds = 1;
-		} else {
-			cmds[1] = SMB2_OP_GET_REPARSE;
-			num_cmds = 2;
-		}
-		create_options |= OPEN_REPARSE_POINT;
+		cmds[num_cmds++] = SMB2_OP_QUERY_WSL_EA;
+		/*
+		 * Skip SMB2_OP_GET_REPARSE if symlink already parsed in create
+		 * response.
+		 */
+		if (data->reparse.tag != IO_REPARSE_TAG_SYMLINK)
+			cmds[num_cmds++] = SMB2_OP_GET_REPARSE;
+
+		oparms = CIFS_OPARMS(cifs_sb, tcon, full_path,
+				     FILE_READ_ATTRIBUTES | FILE_READ_EA,
+				     FILE_OPEN, create_options |
+				     OPEN_REPARSE_POINT, ACL_NO_MODE);
 		cifs_get_readable_path(tcon, full_path, &cfile);
 		rc = smb2_compound_op(xid, tcon, cifs_sb, full_path,
-				      FILE_READ_ATTRIBUTES, FILE_OPEN,
-				      create_options, ACL_NO_MODE, in_iov,
-				      cmds, num_cmds, cfile, NULL, NULL);
+				      &oparms, in_iov, cmds, num_cmds,
+				      cfile, NULL, NULL, NULL);
 		break;
 	case -EREMOTE:
 		break;
@@ -807,11 +984,14 @@ smb2_mkdir(const unsigned int xid, struct inode *parent_inode, umode_t mode,
 	   struct cifs_tcon *tcon, const char *name,
 	   struct cifs_sb_info *cifs_sb)
 {
-	return smb2_compound_op(xid, tcon, cifs_sb, name,
-				FILE_WRITE_ATTRIBUTES, FILE_CREATE,
-				CREATE_NOT_FILE, mode,
-				NULL, &(int){SMB2_OP_MKDIR}, 1,
-				NULL, NULL, NULL);
+	struct cifs_open_parms oparms;
+
+	oparms = CIFS_OPARMS(cifs_sb, tcon, name, FILE_WRITE_ATTRIBUTES,
+			     FILE_CREATE, CREATE_NOT_FILE, mode);
+	return smb2_compound_op(xid, tcon, cifs_sb,
+				name, &oparms, NULL,
+				&(int){SMB2_OP_MKDIR}, 1,
+				NULL, NULL, NULL, NULL);
 }
 
 void
@@ -819,6 +999,7 @@ smb2_mkdir_setinfo(struct inode *inode, const char *name,
 		   struct cifs_sb_info *cifs_sb, struct cifs_tcon *tcon,
 		   const unsigned int xid)
 {
+	struct cifs_open_parms oparms;
 	FILE_BASIC_INFO data = {};
 	struct cifsInodeInfo *cifs_i;
 	struct cifsFileInfo *cfile;
@@ -832,11 +1013,12 @@ smb2_mkdir_setinfo(struct inode *inode, const char *name,
 	dosattrs = cifs_i->cifsAttrs | ATTR_READONLY;
 	data.Attributes = cpu_to_le32(dosattrs);
 	cifs_get_writable_path(tcon, name, FIND_WR_ANY, &cfile);
+	oparms = CIFS_OPARMS(cifs_sb, tcon, name, FILE_WRITE_ATTRIBUTES,
+			     FILE_CREATE, CREATE_NOT_FILE, ACL_NO_MODE);
 	tmprc = smb2_compound_op(xid, tcon, cifs_sb, name,
-				 FILE_WRITE_ATTRIBUTES, FILE_CREATE,
-				 CREATE_NOT_FILE, ACL_NO_MODE, &in_iov,
+				 &oparms, &in_iov,
 				 &(int){SMB2_OP_SET_INFO}, 1,
-				 cfile, NULL, NULL);
+				 cfile, NULL, NULL, NULL);
 	if (tmprc == 0)
 		cifs_i->cifsAttrs = dosattrs;
 }
@@ -845,31 +1027,47 @@ int
 smb2_rmdir(const unsigned int xid, struct cifs_tcon *tcon, const char *name,
 	   struct cifs_sb_info *cifs_sb)
 {
+	struct cifs_open_parms oparms;
+
 	drop_cached_dir_by_name(xid, tcon, name, cifs_sb);
-	return smb2_compound_op(xid, tcon, cifs_sb, name,
-				DELETE, FILE_OPEN, CREATE_NOT_FILE,
-				ACL_NO_MODE, NULL,
+	oparms = CIFS_OPARMS(cifs_sb, tcon, name, DELETE,
+			     FILE_OPEN, CREATE_NOT_FILE, ACL_NO_MODE);
+	return smb2_compound_op(xid, tcon, cifs_sb,
+				name, &oparms, NULL,
 				&(int){SMB2_OP_RMDIR}, 1,
-				NULL, NULL, NULL);
+				NULL, NULL, NULL, NULL);
 }
 
 int
 smb2_unlink(const unsigned int xid, struct cifs_tcon *tcon, const char *name,
-	    struct cifs_sb_info *cifs_sb)
+	    struct cifs_sb_info *cifs_sb, struct dentry *dentry)
 {
-	return smb2_compound_op(xid, tcon, cifs_sb, name, DELETE, FILE_OPEN,
-				CREATE_DELETE_ON_CLOSE | OPEN_REPARSE_POINT,
-				ACL_NO_MODE, NULL,
-				&(int){SMB2_OP_DELETE}, 1,
-				NULL, NULL, NULL);
+	struct cifs_open_parms oparms;
+
+	oparms = CIFS_OPARMS(cifs_sb, tcon, name,
+			     DELETE, FILE_OPEN,
+			     CREATE_DELETE_ON_CLOSE | OPEN_REPARSE_POINT,
+			     ACL_NO_MODE);
+	int rc = smb2_compound_op(xid, tcon, cifs_sb, name, &oparms,
+				  NULL, &(int){SMB2_OP_DELETE}, 1,
+				  NULL, NULL, NULL, dentry);
+	if (rc == -EINVAL) {
+		cifs_dbg(FYI, "invalid lease key, resending request without lease");
+		rc = smb2_compound_op(xid, tcon, cifs_sb, name, &oparms,
+				      NULL, &(int){SMB2_OP_DELETE}, 1,
+				      NULL, NULL, NULL, NULL);
+	}
+	return rc;
 }
 
 static int smb2_set_path_attr(const unsigned int xid, struct cifs_tcon *tcon,
 			      const char *from_name, const char *to_name,
 			      struct cifs_sb_info *cifs_sb,
 			      __u32 create_options, __u32 access,
-			      int command, struct cifsFileInfo *cfile)
+			      int command, struct cifsFileInfo *cfile,
+				  struct dentry *dentry)
 {
+	struct cifs_open_parms oparms;
 	struct kvec in_iov;
 	__le16 *smb2_to_name = NULL;
 	int rc;
@@ -881,9 +1079,11 @@ static int smb2_set_path_attr(const unsigned int xid, struct cifs_tcon *tcon,
 	}
 	in_iov.iov_base = smb2_to_name;
 	in_iov.iov_len = 2 * UniStrnlen((wchar_t *)smb2_to_name, PATH_MAX);
-	rc = smb2_compound_op(xid, tcon, cifs_sb, from_name, access,
-			      FILE_OPEN, create_options, ACL_NO_MODE,
-			      &in_iov, &command, 1, cfile, NULL, NULL);
+	oparms = CIFS_OPARMS(cifs_sb, tcon, from_name, access, FILE_OPEN,
+			     create_options, ACL_NO_MODE);
+	rc = smb2_compound_op(xid, tcon, cifs_sb, from_name,
+			      &oparms, &in_iov, &command, 1,
+			      cfile, NULL, NULL, dentry);
 smb2_rename_path:
 	kfree(smb2_to_name);
 	return rc;
@@ -901,8 +1101,14 @@ int smb2_rename_path(const unsigned int xid,
 	drop_cached_dir_by_name(xid, tcon, from_name, cifs_sb);
 	cifs_get_writable_path(tcon, from_name, FIND_WR_WITH_DELETE, &cfile);
 
-	return smb2_set_path_attr(xid, tcon, from_name, to_name, cifs_sb,
-				  co, DELETE, SMB2_OP_RENAME, cfile);
+	int rc = smb2_set_path_attr(xid, tcon, from_name, to_name, cifs_sb,
+				  co, DELETE, SMB2_OP_RENAME, cfile, source_dentry);
+	if (rc == -EINVAL) {
+		cifs_dbg(FYI, "invalid lease key, resending request without lease");
+		rc = smb2_set_path_attr(xid, tcon, from_name, to_name, cifs_sb,
+				  co, DELETE, SMB2_OP_RENAME, cfile, NULL);
+	}
+	return rc;
 }
 
 int smb2_create_hardlink(const unsigned int xid,
@@ -915,32 +1121,46 @@ int smb2_create_hardlink(const unsigned int xid,
 
 	return smb2_set_path_attr(xid, tcon, from_name, to_name,
 				  cifs_sb, co, FILE_READ_ATTRIBUTES,
-				  SMB2_OP_HARDLINK, NULL);
+				  SMB2_OP_HARDLINK, NULL, NULL);
 }
 
 int
 smb2_set_path_size(const unsigned int xid, struct cifs_tcon *tcon,
 		   const char *full_path, __u64 size,
-		   struct cifs_sb_info *cifs_sb, bool set_alloc)
+		   struct cifs_sb_info *cifs_sb, bool set_alloc,
+		   struct dentry *dentry)
 {
+	struct cifs_open_parms oparms;
 	struct cifsFileInfo *cfile;
 	struct kvec in_iov;
 	__le64 eof = cpu_to_le64(size);
+	int rc;
 
 	in_iov.iov_base = &eof;
 	in_iov.iov_len = sizeof(eof);
 	cifs_get_writable_path(tcon, full_path, FIND_WR_ANY, &cfile);
-	return smb2_compound_op(xid, tcon, cifs_sb, full_path,
-				FILE_WRITE_DATA, FILE_OPEN,
-				0, ACL_NO_MODE, &in_iov,
-				&(int){SMB2_OP_SET_EOF}, 1,
-				cfile, NULL, NULL);
+
+	oparms = CIFS_OPARMS(cifs_sb, tcon, full_path, FILE_WRITE_DATA,
+			     FILE_OPEN, 0, ACL_NO_MODE);
+	rc = smb2_compound_op(xid, tcon, cifs_sb,
+			      full_path, &oparms, &in_iov,
+			      &(int){SMB2_OP_SET_EOF}, 1,
+			      cfile, NULL, NULL, dentry);
+	if (rc == -EINVAL) {
+		cifs_dbg(FYI, "invalid lease key, resending request without lease");
+		rc = smb2_compound_op(xid, tcon, cifs_sb,
+				      full_path, &oparms, &in_iov,
+				      &(int){SMB2_OP_SET_EOF}, 1,
+				      cfile, NULL, NULL, NULL);
+	}
+	return rc;
 }
 
 int
 smb2_set_file_info(struct inode *inode, const char *full_path,
 		   FILE_BASIC_INFO *buf, const unsigned int xid)
 {
+	struct cifs_open_parms oparms;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
 	struct tcon_link *tlink;
 	struct cifs_tcon *tcon;
@@ -959,11 +1179,12 @@ smb2_set_file_info(struct inode *inode, const char *full_path,
 	tcon = tlink_tcon(tlink);
 
 	cifs_get_writable_path(tcon, full_path, FIND_WR_ANY, &cfile);
-	rc = smb2_compound_op(xid, tcon, cifs_sb, full_path,
-			      FILE_WRITE_ATTRIBUTES, FILE_OPEN,
-			      0, ACL_NO_MODE, &in_iov,
+	oparms = CIFS_OPARMS(cifs_sb, tcon, full_path, FILE_WRITE_ATTRIBUTES,
+			     FILE_OPEN, 0, ACL_NO_MODE);
+	rc = smb2_compound_op(xid, tcon, cifs_sb,
+			      full_path, &oparms, &in_iov,
 			      &(int){SMB2_OP_SET_INFO}, 1,
-			      cfile, NULL, NULL);
+			      cfile, NULL, NULL, NULL);
 	cifs_put_tlink(tlink);
 	return rc;
 }
@@ -973,32 +1194,37 @@ struct inode *smb2_get_reparse_inode(struct cifs_open_info_data *data,
 				     const unsigned int xid,
 				     struct cifs_tcon *tcon,
 				     const char *full_path,
-				     struct kvec *iov)
+				     struct kvec *reparse_iov,
+				     struct kvec *xattr_iov)
 {
+	struct cifs_open_parms oparms;
 	struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
 	struct cifsFileInfo *cfile;
 	struct inode *new = NULL;
 	struct kvec in_iov[2];
 	int cmds[2];
-	int da, co, cd;
 	int rc;
 
-	da = SYNCHRONIZE | DELETE |
-		FILE_READ_ATTRIBUTES |
-		FILE_WRITE_ATTRIBUTES;
-	co = CREATE_NOT_DIR | OPEN_REPARSE_POINT;
-	cd = FILE_CREATE;
+	oparms = CIFS_OPARMS(cifs_sb, tcon, full_path,
+			     SYNCHRONIZE | DELETE |
+			     FILE_READ_ATTRIBUTES |
+			     FILE_WRITE_ATTRIBUTES,
+			     FILE_CREATE,
+			     CREATE_NOT_DIR | OPEN_REPARSE_POINT,
+			     ACL_NO_MODE);
+	if (xattr_iov)
+		oparms.ea_cctx = xattr_iov;
+
 	cmds[0] = SMB2_OP_SET_REPARSE;
-	in_iov[0] = *iov;
+	in_iov[0] = *reparse_iov;
 	in_iov[1].iov_base = data;
 	in_iov[1].iov_len = sizeof(*data);
 
 	if (tcon->posix_extensions) {
 		cmds[1] = SMB2_OP_POSIX_QUERY_INFO;
 		cifs_get_writable_path(tcon, full_path, FIND_WR_ANY, &cfile);
-		rc = smb2_compound_op(xid, tcon, cifs_sb, full_path,
-				      da, cd, co, ACL_NO_MODE, in_iov,
-				      cmds, 2, cfile, NULL, NULL);
+		rc = smb2_compound_op(xid, tcon, cifs_sb, full_path, &oparms,
+				      in_iov, cmds, 2, cfile, NULL, NULL, NULL);
 		if (!rc) {
 			rc = smb311_posix_get_inode_info(&new, full_path,
 							 data, sb, xid);
@@ -1006,9 +1232,8 @@ struct inode *smb2_get_reparse_inode(struct cifs_open_info_data *data,
 	} else {
 		cmds[1] = SMB2_OP_QUERY_INFO;
 		cifs_get_writable_path(tcon, full_path, FIND_WR_ANY, &cfile);
-		rc = smb2_compound_op(xid, tcon, cifs_sb, full_path,
-				      da, cd, co, ACL_NO_MODE, in_iov,
-				      cmds, 2, cfile, NULL, NULL);
+		rc = smb2_compound_op(xid, tcon, cifs_sb, full_path, &oparms,
+				      in_iov, cmds, 2, cfile, NULL, NULL, NULL);
 		if (!rc) {
 			rc = cifs_get_inode_info(&new, full_path,
 						 data, sb, xid, NULL);
@@ -1024,6 +1249,7 @@ int smb2_query_reparse_point(const unsigned int xid,
 			     u32 *tag, struct kvec *rsp,
 			     int *rsp_buftype)
 {
+	struct cifs_open_parms oparms;
 	struct cifs_open_info_data data = {};
 	struct cifsFileInfo *cfile;
 	struct kvec in_iov = { .iov_base = &data, .iov_len = sizeof(data), };
@@ -1032,11 +1258,12 @@ int smb2_query_reparse_point(const unsigned int xid,
 	cifs_dbg(FYI, "%s: path: %s\n", __func__, full_path);
 
 	cifs_get_readable_path(tcon, full_path, &cfile);
-	rc = smb2_compound_op(xid, tcon, cifs_sb, full_path,
-			      FILE_READ_ATTRIBUTES, FILE_OPEN,
-			      OPEN_REPARSE_POINT, ACL_NO_MODE, &in_iov,
+	oparms = CIFS_OPARMS(cifs_sb, tcon, full_path, FILE_READ_ATTRIBUTES,
+			     FILE_OPEN, OPEN_REPARSE_POINT, ACL_NO_MODE);
+	rc = smb2_compound_op(xid, tcon, cifs_sb,
+			      full_path, &oparms, &in_iov,
 			      &(int){SMB2_OP_GET_REPARSE}, 1,
-			      cfile, NULL, NULL);
+			      cfile, NULL, NULL, NULL);
 	if (rc)
 		goto out;
 
diff --git a/fs/smb/client/smb2misc.c b/fs/smb/client/smb2misc.c
index cc72be5a93..677ef6f99a 100644
--- a/fs/smb/client/smb2misc.c
+++ b/fs/smb/client/smb2misc.c
@@ -767,7 +767,7 @@ smb2_cancelled_close_fid(struct work_struct *work)
 	if (rc)
 		cifs_tcon_dbg(VFS, "Close cancelled mid failed rc:%d\n", rc);
 
-	cifs_put_tcon(tcon);
+	cifs_put_tcon(tcon, netfs_trace_tcon_ref_put_cancelled_close_fid);
 	kfree(cancelled);
 }
 
@@ -811,6 +811,8 @@ smb2_handle_cancelled_close(struct cifs_tcon *tcon, __u64 persistent_fid,
 	if (tcon->tc_count <= 0) {
 		struct TCP_Server_Info *server = NULL;
 
+		trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+				    netfs_trace_tcon_ref_see_cancelled_close);
 		WARN_ONCE(tcon->tc_count < 0, "tcon refcount is negative");
 		spin_unlock(&cifs_tcp_ses_lock);
 
@@ -823,12 +825,14 @@ smb2_handle_cancelled_close(struct cifs_tcon *tcon, __u64 persistent_fid,
 		return 0;
 	}
 	tcon->tc_count++;
+	trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+			    netfs_trace_tcon_ref_get_cancelled_close);
 	spin_unlock(&cifs_tcp_ses_lock);
 
 	rc = __smb2_handle_cancelled_cmd(tcon, SMB2_CLOSE_HE, 0,
 					 persistent_fid, volatile_fid);
 	if (rc)
-		cifs_put_tcon(tcon);
+		cifs_put_tcon(tcon, netfs_trace_tcon_ref_put_cancelled_close);
 
 	return rc;
 }
@@ -856,7 +860,7 @@ smb2_handle_cancelled_mid(struct mid_q_entry *mid, struct TCP_Server_Info *serve
 					 rsp->PersistentFileId,
 					 rsp->VolatileFileId);
 	if (rc)
-		cifs_put_tcon(tcon);
+		cifs_put_tcon(tcon, netfs_trace_tcon_ref_put_cancelled_mid);
 
 	return rc;
 }
diff --git a/fs/smb/client/smb2ops.c b/fs/smb/client/smb2ops.c
index ab77ecb2d4..c8e5365408 100644
--- a/fs/smb/client/smb2ops.c
+++ b/fs/smb/client/smb2ops.c
@@ -28,6 +28,7 @@
 #include "fscache.h"
 #include "fs_context.h"
 #include "cached_dir.h"
+#include "reparse.h"
 
 /* Change credits for different ops and return the total number of credits */
 static int
@@ -216,8 +217,8 @@ smb2_get_credits(struct mid_q_entry *mid)
 }
 
 static int
-smb2_wait_mtu_credits(struct TCP_Server_Info *server, unsigned int size,
-		      unsigned int *num, struct cifs_credits *credits)
+smb2_wait_mtu_credits(struct TCP_Server_Info *server, size_t size,
+		      size_t *num, struct cifs_credits *credits)
 {
 	int rc = 0;
 	unsigned int scredits, in_flight;
@@ -2027,6 +2028,7 @@ smb2_duplicate_extents(const unsigned int xid,
 		 * size will be queried on next revalidate, but it is important
 		 * to make sure that file's cached size is updated immediately
 		 */
+		netfs_resize_file(netfs_inode(inode), dest_off + len, true);
 		cifs_setsize(inode, dest_off + len);
 	}
 	rc = SMB2_ioctl(xid, tcon, trgtfile->fid.persistent_fid,
@@ -2914,8 +2916,11 @@ smb2_get_dfs_refer(const unsigned int xid, struct cifs_ses *ses,
 		tcon = list_first_entry_or_null(&ses->tcon_list,
 						struct cifs_tcon,
 						tcon_list);
-		if (tcon)
+		if (tcon) {
 			tcon->tc_count++;
+			trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+					    netfs_trace_tcon_ref_get_dfs_refer);
+		}
 		spin_unlock(&cifs_tcp_ses_lock);
 	}
 
@@ -2979,6 +2984,8 @@ smb2_get_dfs_refer(const unsigned int xid, struct cifs_ses *ses,
 		/* ipc tcons are not refcounted */
 		spin_lock(&cifs_tcp_ses_lock);
 		tcon->tc_count--;
+		trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+				    netfs_trace_tcon_ref_dec_dfs_refer);
 		/* tc_count can never go negative */
 		WARN_ON(tcon->tc_count < 0);
 		spin_unlock(&cifs_tcp_ses_lock);
@@ -2989,109 +2996,6 @@ smb2_get_dfs_refer(const unsigned int xid, struct cifs_ses *ses,
 	return rc;
 }
 
-/* See MS-FSCC 2.1.2.6 for the 'NFS' style reparse tags */
-static int parse_reparse_posix(struct reparse_posix_data *buf,
-			       struct cifs_sb_info *cifs_sb,
-			       struct cifs_open_info_data *data)
-{
-	unsigned int len;
-	u64 type;
-
-	switch ((type = le64_to_cpu(buf->InodeType))) {
-	case NFS_SPECFILE_LNK:
-		len = le16_to_cpu(buf->ReparseDataLength);
-		data->symlink_target = cifs_strndup_from_utf16(buf->DataBuffer,
-							       len, true,
-							       cifs_sb->local_nls);
-		if (!data->symlink_target)
-			return -ENOMEM;
-		convert_delimiter(data->symlink_target, '/');
-		cifs_dbg(FYI, "%s: target path: %s\n",
-			 __func__, data->symlink_target);
-		break;
-	case NFS_SPECFILE_CHR:
-	case NFS_SPECFILE_BLK:
-	case NFS_SPECFILE_FIFO:
-	case NFS_SPECFILE_SOCK:
-		break;
-	default:
-		cifs_dbg(VFS, "%s: unhandled inode type: 0x%llx\n",
-			 __func__, type);
-		return -EOPNOTSUPP;
-	}
-	return 0;
-}
-
-static int parse_reparse_symlink(struct reparse_symlink_data_buffer *sym,
-				 u32 plen, bool unicode,
-				 struct cifs_sb_info *cifs_sb,
-				 struct cifs_open_info_data *data)
-{
-	unsigned int len;
-	unsigned int offs;
-
-	/* We handle Symbolic Link reparse tag here. See: MS-FSCC 2.1.2.4 */
-
-	offs = le16_to_cpu(sym->SubstituteNameOffset);
-	len = le16_to_cpu(sym->SubstituteNameLength);
-	if (offs + 20 > plen || offs + len + 20 > plen) {
-		cifs_dbg(VFS, "srv returned malformed symlink buffer\n");
-		return -EIO;
-	}
-
-	data->symlink_target = cifs_strndup_from_utf16(sym->PathBuffer + offs,
-						       len, unicode,
-						       cifs_sb->local_nls);
-	if (!data->symlink_target)
-		return -ENOMEM;
-
-	convert_delimiter(data->symlink_target, '/');
-	cifs_dbg(FYI, "%s: target path: %s\n", __func__, data->symlink_target);
-
-	return 0;
-}
-
-int parse_reparse_point(struct reparse_data_buffer *buf,
-			u32 plen, struct cifs_sb_info *cifs_sb,
-			bool unicode, struct cifs_open_info_data *data)
-{
-	data->reparse.buf = buf;
-
-	/* See MS-FSCC 2.1.2 */
-	switch (le32_to_cpu(buf->ReparseTag)) {
-	case IO_REPARSE_TAG_NFS:
-		return parse_reparse_posix((struct reparse_posix_data *)buf,
-					   cifs_sb, data);
-	case IO_REPARSE_TAG_SYMLINK:
-		return parse_reparse_symlink(
-			(struct reparse_symlink_data_buffer *)buf,
-			plen, unicode, cifs_sb, data);
-	case IO_REPARSE_TAG_LX_SYMLINK:
-	case IO_REPARSE_TAG_AF_UNIX:
-	case IO_REPARSE_TAG_LX_FIFO:
-	case IO_REPARSE_TAG_LX_CHR:
-	case IO_REPARSE_TAG_LX_BLK:
-		return 0;
-	default:
-		cifs_dbg(VFS, "%s: unhandled reparse tag: 0x%08x\n",
-			 __func__, le32_to_cpu(buf->ReparseTag));
-		return -EOPNOTSUPP;
-	}
-}
-
-static int smb2_parse_reparse_point(struct cifs_sb_info *cifs_sb,
-				    struct kvec *rsp_iov,
-				    struct cifs_open_info_data *data)
-{
-	struct reparse_data_buffer *buf;
-	struct smb2_ioctl_rsp *io = rsp_iov->iov_base;
-	u32 plen = le32_to_cpu(io->OutputCount);
-
-	buf = (struct reparse_data_buffer *)((u8 *)io +
-					     le32_to_cpu(io->OutputOffset));
-	return parse_reparse_point(buf, plen, cifs_sb, true, data);
-}
-
 static struct cifs_ntsd *
 get_smb2_acl_by_fid(struct cifs_sb_info *cifs_sb,
 		    const struct cifs_fid *cifsfid, u32 *pacllen, u32 info)
@@ -3733,6 +3637,7 @@ static long smb3_insert_range(struct file *file, struct cifs_tcon *tcon,
 	rc = smb2_copychunk_range(xid, cfile, cfile, off, count, off + len);
 	if (rc < 0)
 		goto out_2;
+	cifsi->netfs.zero_point = new_eof;
 
 	rc = smb3_zero_data(file, tcon, off, len, xid);
 	if (rc < 0)
@@ -4018,7 +3923,7 @@ smb21_set_oplock_level(struct cifsInodeInfo *cinode, __u32 oplock,
 		strcat(message, "W");
 	}
 	if (!new_oplock)
-		strncpy(message, "None", sizeof(message));
+		strscpy(message, "None");
 
 	cinode->oplock = new_oplock;
 	cifs_dbg(FYI, "%s Lease granted on inode %p\n", message,
@@ -4587,7 +4492,7 @@ handle_read_data(struct TCP_Server_Info *server, struct mid_q_entry *mid,
 	unsigned int cur_off;
 	unsigned int cur_page_idx;
 	unsigned int pad_len;
-	struct cifs_readdata *rdata = mid->callback_data;
+	struct cifs_io_subrequest *rdata = mid->callback_data;
 	struct smb2_hdr *shdr = (struct smb2_hdr *)buf;
 	int length;
 	bool use_rdma_mr = false;
@@ -4689,7 +4594,7 @@ handle_read_data(struct TCP_Server_Info *server, struct mid_q_entry *mid,
 
 		/* Copy the data to the output I/O iterator. */
 		rdata->result = cifs_copy_pages_to_iter(pages, pages_len,
-							cur_off, &rdata->iter);
+							cur_off, &rdata->subreq.io_iter);
 		if (rdata->result != 0) {
 			if (is_offloaded)
 				mid->mid_state = MID_RESPONSE_MALFORMED;
@@ -4703,7 +4608,7 @@ handle_read_data(struct TCP_Server_Info *server, struct mid_q_entry *mid,
 		/* read response payload is in buf */
 		WARN_ONCE(pages && !xa_empty(pages),
 			  "read data can be either in buf or in pages");
-		length = copy_to_iter(buf + data_offset, data_len, &rdata->iter);
+		length = copy_to_iter(buf + data_offset, data_len, &rdata->subreq.io_iter);
 		if (length < 0)
 			return length;
 		rdata->got_bytes = data_len;
@@ -5066,214 +4971,87 @@ static int smb2_next_header(struct TCP_Server_Info *server, char *buf,
 	return 0;
 }
 
-int cifs_sfu_make_node(unsigned int xid, struct inode *inode,
-		       struct dentry *dentry, struct cifs_tcon *tcon,
-		       const char *full_path, umode_t mode, dev_t dev)
+static int __cifs_sfu_make_node(unsigned int xid, struct inode *inode,
+				struct dentry *dentry, struct cifs_tcon *tcon,
+				const char *full_path, umode_t mode, dev_t dev)
 {
-	struct cifs_open_info_data buf = {};
 	struct TCP_Server_Info *server = tcon->ses->server;
 	struct cifs_open_parms oparms;
 	struct cifs_io_parms io_parms = {};
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
 	struct cifs_fid fid;
 	unsigned int bytes_written;
-	struct win_dev *pdev;
+	struct win_dev pdev = {};
 	struct kvec iov[2];
 	__u32 oplock = server->oplocks ? REQ_OPLOCK : 0;
 	int rc;
 
-	if (!S_ISCHR(mode) && !S_ISBLK(mode) && !S_ISFIFO(mode))
+	switch (mode & S_IFMT) {
+	case S_IFCHR:
+		strscpy(pdev.type, "IntxCHR");
+		pdev.major = cpu_to_le64(MAJOR(dev));
+		pdev.minor = cpu_to_le64(MINOR(dev));
+		break;
+	case S_IFBLK:
+		strscpy(pdev.type, "IntxBLK");
+		pdev.major = cpu_to_le64(MAJOR(dev));
+		pdev.minor = cpu_to_le64(MINOR(dev));
+		break;
+	case S_IFSOCK:
+		strscpy(pdev.type, "LnxSOCK");
+		break;
+	case S_IFIFO:
+		strscpy(pdev.type, "LnxFIFO");
+		break;
+	default:
 		return -EPERM;
+	}
 
-	oparms = (struct cifs_open_parms) {
-		.tcon = tcon,
-		.cifs_sb = cifs_sb,
-		.desired_access = GENERIC_WRITE,
-		.create_options = cifs_create_options(cifs_sb, CREATE_NOT_DIR |
-						      CREATE_OPTION_SPECIAL),
-		.disposition = FILE_CREATE,
-		.path = full_path,
-		.fid = &fid,
-	};
+	oparms = CIFS_OPARMS(cifs_sb, tcon, full_path, GENERIC_WRITE,
+			     FILE_CREATE, CREATE_NOT_DIR |
+			     CREATE_OPTION_SPECIAL, ACL_NO_MODE);
+	oparms.fid = &fid;
 
-	rc = server->ops->open(xid, &oparms, &oplock, &buf);
+	rc = server->ops->open(xid, &oparms, &oplock, NULL);
 	if (rc)
 		return rc;
 
-	/*
-	 * BB Do not bother to decode buf since no local inode yet to put
-	 * timestamps in, but we can reuse it safely.
-	 */
-	pdev = (struct win_dev *)&buf.fi;
 	io_parms.pid = current->tgid;
 	io_parms.tcon = tcon;
-	io_parms.length = sizeof(*pdev);
-	iov[1].iov_base = pdev;
-	iov[1].iov_len = sizeof(*pdev);
-	if (S_ISCHR(mode)) {
-		memcpy(pdev->type, "IntxCHR", 8);
-		pdev->major = cpu_to_le64(MAJOR(dev));
-		pdev->minor = cpu_to_le64(MINOR(dev));
-	} else if (S_ISBLK(mode)) {
-		memcpy(pdev->type, "IntxBLK", 8);
-		pdev->major = cpu_to_le64(MAJOR(dev));
-		pdev->minor = cpu_to_le64(MINOR(dev));
-	} else if (S_ISFIFO(mode)) {
-		memcpy(pdev->type, "LnxFIFO", 8);
-	}
+	io_parms.length = sizeof(pdev);
+	iov[1].iov_base = &pdev;
+	iov[1].iov_len = sizeof(pdev);
 
 	rc = server->ops->sync_write(xid, &fid, &io_parms,
 				     &bytes_written, iov, 1);
 	server->ops->close(xid, tcon, &fid);
-	d_drop(dentry);
-	/* FIXME: add code here to set EAs */
-	cifs_free_open_info(&buf);
 	return rc;
 }
 
-static inline u64 mode_nfs_type(mode_t mode)
-{
-	switch (mode & S_IFMT) {
-	case S_IFBLK: return NFS_SPECFILE_BLK;
-	case S_IFCHR: return NFS_SPECFILE_CHR;
-	case S_IFIFO: return NFS_SPECFILE_FIFO;
-	case S_IFSOCK: return NFS_SPECFILE_SOCK;
-	}
-	return 0;
-}
-
-static int nfs_set_reparse_buf(struct reparse_posix_data *buf,
-			       mode_t mode, dev_t dev,
-			       struct kvec *iov)
-{
-	u64 type;
-	u16 len, dlen;
-
-	len = sizeof(*buf);
-
-	switch ((type = mode_nfs_type(mode))) {
-	case NFS_SPECFILE_BLK:
-	case NFS_SPECFILE_CHR:
-		dlen = sizeof(__le64);
-		break;
-	case NFS_SPECFILE_FIFO:
-	case NFS_SPECFILE_SOCK:
-		dlen = 0;
-		break;
-	default:
-		return -EOPNOTSUPP;
-	}
-
-	buf->ReparseTag = cpu_to_le32(IO_REPARSE_TAG_NFS);
-	buf->Reserved = 0;
-	buf->InodeType = cpu_to_le64(type);
-	buf->ReparseDataLength = cpu_to_le16(len + dlen -
-					     sizeof(struct reparse_data_buffer));
-	*(__le64 *)buf->DataBuffer = cpu_to_le64(((u64)MAJOR(dev) << 32) |
-						 MINOR(dev));
-	iov->iov_base = buf;
-	iov->iov_len = len + dlen;
-	return 0;
-}
-
-static int nfs_make_node(unsigned int xid, struct inode *inode,
-			 struct dentry *dentry, struct cifs_tcon *tcon,
-			 const char *full_path, umode_t mode, dev_t dev)
+int cifs_sfu_make_node(unsigned int xid, struct inode *inode,
+		       struct dentry *dentry, struct cifs_tcon *tcon,
+		       const char *full_path, umode_t mode, dev_t dev)
 {
-	struct cifs_open_info_data data;
-	struct reparse_posix_data *p;
-	struct inode *new;
-	struct kvec iov;
-	__u8 buf[sizeof(*p) + sizeof(__le64)];
+	struct inode *new = NULL;
 	int rc;
 
-	p = (struct reparse_posix_data *)buf;
-	rc = nfs_set_reparse_buf(p, mode, dev, &iov);
+	rc = __cifs_sfu_make_node(xid, inode, dentry, tcon,
+				  full_path, mode, dev);
 	if (rc)
 		return rc;
 
-	data = (struct cifs_open_info_data) {
-		.reparse_point = true,
-		.reparse = { .tag = IO_REPARSE_TAG_NFS, .posix = p, },
-	};
-
-	new = smb2_get_reparse_inode(&data, inode->i_sb, xid,
-				     tcon, full_path, &iov);
-	if (!IS_ERR(new))
-		d_instantiate(dentry, new);
-	else
-		rc = PTR_ERR(new);
-	cifs_free_open_info(&data);
-	return rc;
-}
-
-static int smb2_create_reparse_symlink(const unsigned int xid,
-				       struct inode *inode,
-				       struct dentry *dentry,
-				       struct cifs_tcon *tcon,
-				       const char *full_path,
-				       const char *symname)
-{
-	struct reparse_symlink_data_buffer *buf = NULL;
-	struct cifs_open_info_data data;
-	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
-	struct inode *new;
-	struct kvec iov;
-	__le16 *path;
-	char *sym, sep = CIFS_DIR_SEP(cifs_sb);
-	u16 len, plen;
-	int rc = 0;
-
-	sym = kstrdup(symname, GFP_KERNEL);
-	if (!sym)
-		return -ENOMEM;
-
-	data = (struct cifs_open_info_data) {
-		.reparse_point = true,
-		.reparse = { .tag = IO_REPARSE_TAG_SYMLINK, },
-		.symlink_target = sym,
-	};
-
-	convert_delimiter(sym, sep);
-	path = cifs_convert_path_to_utf16(sym, cifs_sb);
-	if (!path) {
-		rc = -ENOMEM;
-		goto out;
-	}
-
-	plen = 2 * UniStrnlen((wchar_t *)path, PATH_MAX);
-	len = sizeof(*buf) + plen * 2;
-	buf = kzalloc(len, GFP_KERNEL);
-	if (!buf) {
-		rc = -ENOMEM;
-		goto out;
+	if (tcon->posix_extensions) {
+		rc = smb311_posix_get_inode_info(&new, full_path, NULL,
+						 inode->i_sb, xid);
+	} else if (tcon->unix_ext) {
+		rc = cifs_get_inode_info_unix(&new, full_path,
+					      inode->i_sb, xid);
+	} else {
+		rc = cifs_get_inode_info(&new, full_path, NULL,
+					 inode->i_sb, xid, NULL);
 	}
-
-	buf->ReparseTag = cpu_to_le32(IO_REPARSE_TAG_SYMLINK);
-	buf->ReparseDataLength = cpu_to_le16(len - sizeof(struct reparse_data_buffer));
-	buf->SubstituteNameOffset = cpu_to_le16(plen);
-	buf->SubstituteNameLength = cpu_to_le16(plen);
-	memcpy(&buf->PathBuffer[plen], path, plen);
-	buf->PrintNameOffset = 0;
-	buf->PrintNameLength = cpu_to_le16(plen);
-	memcpy(buf->PathBuffer, path, plen);
-	buf->Flags = cpu_to_le32(*symname != '/' ? SYMLINK_FLAG_RELATIVE : 0);
-	if (*sym != sep)
-		buf->Flags = cpu_to_le32(SYMLINK_FLAG_RELATIVE);
-
-	convert_delimiter(sym, '/');
-	iov.iov_base = buf;
-	iov.iov_len = len;
-	new = smb2_get_reparse_inode(&data, inode->i_sb, xid,
-				     tcon, full_path, &iov);
-	if (!IS_ERR(new))
+	if (!rc)
 		d_instantiate(dentry, new);
-	else
-		rc = PTR_ERR(new);
-out:
-	kfree(path);
-	cifs_free_open_info(&data);
-	kfree(buf);
 	return rc;
 }
 
@@ -5294,8 +5072,8 @@ static int smb2_make_node(unsigned int xid, struct inode *inode,
 		rc = cifs_sfu_make_node(xid, inode, dentry, tcon,
 					full_path, mode, dev);
 	} else {
-		rc = nfs_make_node(xid, inode, dentry, tcon,
-				   full_path, mode, dev);
+		rc = smb2_mknod_reparse(xid, inode, dentry, tcon,
+					full_path, mode, dev);
 	}
 	return rc;
 }
diff --git a/fs/smb/client/smb2pdu.c b/fs/smb/client/smb2pdu.c
index 60793143e2..896147ba66 100644
--- a/fs/smb/client/smb2pdu.c
+++ b/fs/smb/client/smb2pdu.c
@@ -23,6 +23,8 @@
 #include <linux/uuid.h>
 #include <linux/pagemap.h>
 #include <linux/xattr.h>
+#include <linux/netfs.h>
+#include <trace/events/netfs.h>
 #include "cifsglob.h"
 #include "cifsacl.h"
 #include "cifsproto.h"
@@ -80,6 +82,9 @@ int smb3_encryption_required(const struct cifs_tcon *tcon)
 	if (tcon->seal &&
 	    (tcon->ses->server->capabilities & SMB2_GLOBAL_CAP_ENCRYPTION))
 		return 1;
+	if (((global_secflags & CIFSSEC_MUST_SEAL) == CIFSSEC_MUST_SEAL) &&
+	    (tcon->ses->server->capabilities & SMB2_GLOBAL_CAP_ENCRYPTION))
+		return 1;
 	return 0;
 }
 
@@ -743,7 +748,7 @@ assemble_neg_contexts(struct smb2_negotiate_req *req,
 	pneg_ctxt += sizeof(struct smb2_posix_neg_context);
 	neg_context_count++;
 
-	if (server->compress_algorithm) {
+	if (server->compression.requested) {
 		build_compression_ctxt((struct smb2_compression_capabilities_context *)
 				pneg_ctxt);
 		ctxt_len = ALIGN(sizeof(struct smb2_compression_capabilities_context), 8);
@@ -791,6 +796,9 @@ static void decode_compress_ctx(struct TCP_Server_Info *server,
 			 struct smb2_compression_capabilities_context *ctxt)
 {
 	unsigned int len = le16_to_cpu(ctxt->DataLength);
+	__le16 alg;
+
+	server->compression.enabled = false;
 
 	/*
 	 * Caller checked that DataLength remains within SMB boundary. We still
@@ -801,15 +809,22 @@ static void decode_compress_ctx(struct TCP_Server_Info *server,
 		pr_warn_once("server sent bad compression cntxt\n");
 		return;
 	}
+
 	if (le16_to_cpu(ctxt->CompressionAlgorithmCount) != 1) {
-		pr_warn_once("Invalid SMB3 compress algorithm count\n");
+		pr_warn_once("invalid SMB3 compress algorithm count\n");
 		return;
 	}
-	if (le16_to_cpu(ctxt->CompressionAlgorithms[0]) > 3) {
-		pr_warn_once("unknown compression algorithm\n");
+
+	alg = ctxt->CompressionAlgorithms[0];
+
+	/* 'NONE' (0) compressor type is never negotiated */
+	if (alg == 0 || le16_to_cpu(alg) > 3) {
+		pr_warn_once("invalid compression algorithm '%u'\n", alg);
 		return;
 	}
-	server->compress_algorithm = ctxt->CompressionAlgorithms[0];
+
+	server->compression.alg = alg;
+	server->compression.enabled = true;
 }
 
 static int decode_encrypt_ctx(struct TCP_Server_Info *server,
@@ -2732,6 +2747,17 @@ add_query_id_context(struct kvec *iov, unsigned int *num_iovec)
 	return 0;
 }
 
+static void add_ea_context(struct cifs_open_parms *oparms,
+			   struct kvec *rq_iov, unsigned int *num_iovs)
+{
+	struct kvec *iov = oparms->ea_cctx;
+
+	if (iov && iov->iov_base && iov->iov_len) {
+		rq_iov[(*num_iovs)++] = *iov;
+		memset(iov, 0, sizeof(*iov));
+	}
+}
+
 static int
 alloc_path_with_tree_prefix(__le16 **out_path, int *out_size, int *out_len,
 			    const char *treename, const __le16 *path)
@@ -3098,6 +3124,7 @@ SMB2_open_init(struct cifs_tcon *tcon, struct TCP_Server_Info *server,
 	}
 
 	add_query_id_context(iov, &n_iov);
+	add_ea_context(oparms, iov, &n_iov);
 
 	if (n_iov > 2) {
 		/*
@@ -4116,6 +4143,8 @@ void smb2_reconnect_server(struct work_struct *work)
 		list_for_each_entry(tcon, &ses->tcon_list, tcon_list) {
 			if (tcon->need_reconnect || tcon->need_reopen_files) {
 				tcon->tc_count++;
+				trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+						    netfs_trace_tcon_ref_get_reconnect_server);
 				list_add_tail(&tcon->rlist, &tmp_list);
 				tcon_selected = true;
 			}
@@ -4154,14 +4183,14 @@ void smb2_reconnect_server(struct work_struct *work)
 		if (tcon->ipc)
 			cifs_put_smb_ses(tcon->ses);
 		else
-			cifs_put_tcon(tcon);
+			cifs_put_tcon(tcon, netfs_trace_tcon_ref_put_reconnect_server);
 	}
 
 	if (!ses_exist)
 		goto done;
 
 	/* allocate a dummy tcon struct used for reconnect */
-	tcon = tcon_info_alloc(false);
+	tcon = tcon_info_alloc(false, netfs_trace_tcon_ref_new_reconnect_server);
 	if (!tcon) {
 		resched = true;
 		list_for_each_entry_safe(ses, ses2, &tmp_ses_list, rlist) {
@@ -4184,7 +4213,7 @@ void smb2_reconnect_server(struct work_struct *work)
 		list_del_init(&ses->rlist);
 		cifs_put_smb_ses(ses);
 	}
-	tconInfoFree(tcon);
+	tconInfoFree(tcon, netfs_trace_tcon_ref_free_reconnect_server);
 
 done:
 	cifs_dbg(FYI, "Reconnecting tcons and channels finished\n");
@@ -4367,7 +4396,7 @@ static inline bool smb3_use_rdma_offload(struct cifs_io_parms *io_parms)
  */
 static int
 smb2_new_read_req(void **buf, unsigned int *total_len,
-	struct cifs_io_parms *io_parms, struct cifs_readdata *rdata,
+	struct cifs_io_parms *io_parms, struct cifs_io_subrequest *rdata,
 	unsigned int remaining_bytes, int request_type)
 {
 	int rc = -EACCES;
@@ -4395,10 +4424,12 @@ smb2_new_read_req(void **buf, unsigned int *total_len,
 	req->Length = cpu_to_le32(io_parms->length);
 	req->Offset = cpu_to_le64(io_parms->offset);
 
-	trace_smb3_read_enter(0 /* xid */,
-			io_parms->persistent_fid,
-			io_parms->tcon->tid, io_parms->tcon->ses->Suid,
-			io_parms->offset, io_parms->length);
+	trace_smb3_read_enter(rdata ? rdata->rreq->debug_id : 0,
+			      rdata ? rdata->subreq.debug_index : 0,
+			      rdata ? rdata->xid : 0,
+			      io_parms->persistent_fid,
+			      io_parms->tcon->tid, io_parms->tcon->ses->Suid,
+			      io_parms->offset, io_parms->length);
 #ifdef CONFIG_CIFS_SMB_DIRECT
 	/*
 	 * If we want to do a RDMA write, fill in and append
@@ -4408,7 +4439,7 @@ smb2_new_read_req(void **buf, unsigned int *total_len,
 		struct smbd_buffer_descriptor_v1 *v1;
 		bool need_invalidate = server->dialect == SMB30_PROT_ID;
 
-		rdata->mr = smbd_register_mr(server->smbd_conn, &rdata->iter,
+		rdata->mr = smbd_register_mr(server->smbd_conn, &rdata->subreq.io_iter,
 					     true, need_invalidate);
 		if (!rdata->mr)
 			return -EAGAIN;
@@ -4456,11 +4487,21 @@ smb2_new_read_req(void **buf, unsigned int *total_len,
 	return rc;
 }
 
+static void smb2_readv_worker(struct work_struct *work)
+{
+	struct cifs_io_subrequest *rdata =
+		container_of(work, struct cifs_io_subrequest, subreq.work);
+
+	netfs_subreq_terminated(&rdata->subreq,
+				(rdata->result == 0 || rdata->result == -EAGAIN) ?
+				rdata->got_bytes : rdata->result, true);
+}
+
 static void
 smb2_readv_callback(struct mid_q_entry *mid)
 {
-	struct cifs_readdata *rdata = mid->callback_data;
-	struct cifs_tcon *tcon = tlink_tcon(rdata->cfile->tlink);
+	struct cifs_io_subrequest *rdata = mid->callback_data;
+	struct cifs_tcon *tcon = tlink_tcon(rdata->req->cfile->tlink);
 	struct TCP_Server_Info *server = rdata->server;
 	struct smb2_hdr *shdr =
 				(struct smb2_hdr *)rdata->iov[0].iov_base;
@@ -4468,17 +4509,17 @@ smb2_readv_callback(struct mid_q_entry *mid)
 	struct smb_rqst rqst = { .rq_iov = &rdata->iov[1], .rq_nvec = 1 };
 
 	if (rdata->got_bytes) {
-		rqst.rq_iter	  = rdata->iter;
-		rqst.rq_iter_size = iov_iter_count(&rdata->iter);
+		rqst.rq_iter	  = rdata->subreq.io_iter;
+		rqst.rq_iter_size = iov_iter_count(&rdata->subreq.io_iter);
 	}
 
 	WARN_ONCE(rdata->server != mid->server,
 		  "rdata server %p != mid server %p",
 		  rdata->server, mid->server);
 
-	cifs_dbg(FYI, "%s: mid=%llu state=%d result=%d bytes=%u\n",
+	cifs_dbg(FYI, "%s: mid=%llu state=%d result=%d bytes=%zu\n",
 		 __func__, mid->mid, mid->mid_state, rdata->result,
-		 rdata->bytes);
+		 rdata->subreq.len);
 
 	switch (mid->mid_state) {
 	case MID_RESPONSE_RECEIVED:
@@ -4488,7 +4529,6 @@ smb2_readv_callback(struct mid_q_entry *mid)
 		if (server->sign && !mid->decrypted) {
 			int rc;
 
-			iov_iter_revert(&rqst.rq_iter, rdata->got_bytes);
 			iov_iter_truncate(&rqst.rq_iter, rdata->got_bytes);
 			rc = smb2_verify_signature(&rqst, server);
 			if (rc)
@@ -4529,24 +4569,37 @@ smb2_readv_callback(struct mid_q_entry *mid)
 #endif
 	if (rdata->result && rdata->result != -ENODATA) {
 		cifs_stats_fail_inc(tcon, SMB2_READ_HE);
-		trace_smb3_read_err(0 /* xid */,
-				    rdata->cfile->fid.persistent_fid,
-				    tcon->tid, tcon->ses->Suid, rdata->offset,
-				    rdata->bytes, rdata->result);
+		trace_smb3_read_err(rdata->rreq->debug_id,
+				    rdata->subreq.debug_index,
+				    rdata->xid,
+				    rdata->req->cfile->fid.persistent_fid,
+				    tcon->tid, tcon->ses->Suid, rdata->subreq.start,
+				    rdata->subreq.len, rdata->result);
 	} else
-		trace_smb3_read_done(0 /* xid */,
-				     rdata->cfile->fid.persistent_fid,
+		trace_smb3_read_done(rdata->rreq->debug_id,
+				     rdata->subreq.debug_index,
+				     rdata->xid,
+				     rdata->req->cfile->fid.persistent_fid,
 				     tcon->tid, tcon->ses->Suid,
-				     rdata->offset, rdata->got_bytes);
+				     rdata->subreq.start, rdata->got_bytes);
 
-	queue_work(cifsiod_wq, &rdata->work);
+	if (rdata->result == -ENODATA) {
+		/* We may have got an EOF error because fallocate
+		 * failed to enlarge the file.
+		 */
+		if (rdata->subreq.start < rdata->subreq.rreq->i_size)
+			rdata->result = 0;
+	}
+	rdata->credits.value = 0;
+	INIT_WORK(&rdata->subreq.work, smb2_readv_worker);
+	queue_work(cifsiod_wq, &rdata->subreq.work);
 	release_mid(mid);
 	add_credits(server, &credits, 0);
 }
 
 /* smb2_async_readv - send an async read, and set up mid to handle result */
 int
-smb2_async_readv(struct cifs_readdata *rdata)
+smb2_async_readv(struct cifs_io_subrequest *rdata)
 {
 	int rc, flags = 0;
 	char *buf;
@@ -4555,23 +4608,23 @@ smb2_async_readv(struct cifs_readdata *rdata)
 	struct smb_rqst rqst = { .rq_iov = rdata->iov,
 				 .rq_nvec = 1 };
 	struct TCP_Server_Info *server;
-	struct cifs_tcon *tcon = tlink_tcon(rdata->cfile->tlink);
+	struct cifs_tcon *tcon = tlink_tcon(rdata->req->cfile->tlink);
 	unsigned int total_len;
 	int credit_request;
 
-	cifs_dbg(FYI, "%s: offset=%llu bytes=%u\n",
-		 __func__, rdata->offset, rdata->bytes);
+	cifs_dbg(FYI, "%s: offset=%llu bytes=%zu\n",
+		 __func__, rdata->subreq.start, rdata->subreq.len);
 
 	if (!rdata->server)
 		rdata->server = cifs_pick_channel(tcon->ses);
 
-	io_parms.tcon = tlink_tcon(rdata->cfile->tlink);
+	io_parms.tcon = tlink_tcon(rdata->req->cfile->tlink);
 	io_parms.server = server = rdata->server;
-	io_parms.offset = rdata->offset;
-	io_parms.length = rdata->bytes;
-	io_parms.persistent_fid = rdata->cfile->fid.persistent_fid;
-	io_parms.volatile_fid = rdata->cfile->fid.volatile_fid;
-	io_parms.pid = rdata->pid;
+	io_parms.offset = rdata->subreq.start;
+	io_parms.length = rdata->subreq.len;
+	io_parms.persistent_fid = rdata->req->cfile->fid.persistent_fid;
+	io_parms.volatile_fid = rdata->req->cfile->fid.volatile_fid;
+	io_parms.pid = rdata->req->pid;
 
 	rc = smb2_new_read_req(
 		(void **) &buf, &total_len, &io_parms, rdata, 0, 0);
@@ -4587,7 +4640,7 @@ smb2_async_readv(struct cifs_readdata *rdata)
 	shdr = (struct smb2_hdr *)buf;
 
 	if (rdata->credits.value > 0) {
-		shdr->CreditCharge = cpu_to_le16(DIV_ROUND_UP(rdata->bytes,
+		shdr->CreditCharge = cpu_to_le16(DIV_ROUND_UP(rdata->subreq.len,
 						SMB2_MAX_BUFFER_SIZE));
 		credit_request = le16_to_cpu(shdr->CreditCharge) + 8;
 		if (server->credits >= server->max_credits)
@@ -4597,22 +4650,22 @@ smb2_async_readv(struct cifs_readdata *rdata)
 				min_t(int, server->max_credits -
 						server->credits, credit_request));
 
-		rc = adjust_credits(server, &rdata->credits, rdata->bytes);
+		rc = adjust_credits(server, &rdata->credits, rdata->subreq.len);
 		if (rc)
 			goto async_readv_out;
 
 		flags |= CIFS_HAS_CREDITS;
 	}
 
-	kref_get(&rdata->refcount);
 	rc = cifs_call_async(server, &rqst,
 			     cifs_readv_receive, smb2_readv_callback,
 			     smb3_handle_read_data, rdata, flags,
 			     &rdata->credits);
 	if (rc) {
-		kref_put(&rdata->refcount, cifs_readdata_release);
 		cifs_stats_fail_inc(io_parms.tcon, SMB2_READ_HE);
-		trace_smb3_read_err(0 /* xid */, io_parms.persistent_fid,
+		trace_smb3_read_err(rdata->rreq->debug_id,
+				    rdata->subreq.debug_index,
+				    rdata->xid, io_parms.persistent_fid,
 				    io_parms.tcon->tid,
 				    io_parms.tcon->ses->Suid,
 				    io_parms.offset, io_parms.length, rc);
@@ -4663,22 +4716,23 @@ SMB2_read(const unsigned int xid, struct cifs_io_parms *io_parms,
 		if (rc != -ENODATA) {
 			cifs_stats_fail_inc(io_parms->tcon, SMB2_READ_HE);
 			cifs_dbg(VFS, "Send error in read = %d\n", rc);
-			trace_smb3_read_err(xid,
+			trace_smb3_read_err(0, 0, xid,
 					    req->PersistentFileId,
 					    io_parms->tcon->tid, ses->Suid,
 					    io_parms->offset, io_parms->length,
 					    rc);
 		} else
-			trace_smb3_read_done(xid, req->PersistentFileId, io_parms->tcon->tid,
+			trace_smb3_read_done(0, 0, xid,
+					     req->PersistentFileId, io_parms->tcon->tid,
 					     ses->Suid, io_parms->offset, 0);
 		free_rsp_buf(resp_buftype, rsp_iov.iov_base);
 		cifs_small_buf_release(req);
 		return rc == -ENODATA ? 0 : rc;
 	} else
-		trace_smb3_read_done(xid,
-				    req->PersistentFileId,
-				    io_parms->tcon->tid, ses->Suid,
-				    io_parms->offset, io_parms->length);
+		trace_smb3_read_done(0, 0, xid,
+				     req->PersistentFileId,
+				     io_parms->tcon->tid, ses->Suid,
+				     io_parms->offset, io_parms->length);
 
 	cifs_small_buf_release(req);
 
@@ -4711,12 +4765,13 @@ SMB2_read(const unsigned int xid, struct cifs_io_parms *io_parms,
 static void
 smb2_writev_callback(struct mid_q_entry *mid)
 {
-	struct cifs_writedata *wdata = mid->callback_data;
-	struct cifs_tcon *tcon = tlink_tcon(wdata->cfile->tlink);
+	struct cifs_io_subrequest *wdata = mid->callback_data;
+	struct cifs_tcon *tcon = tlink_tcon(wdata->req->cfile->tlink);
 	struct TCP_Server_Info *server = wdata->server;
-	unsigned int written;
 	struct smb2_write_rsp *rsp = (struct smb2_write_rsp *)mid->resp_buf;
 	struct cifs_credits credits = { .value = 0, .instance = 0 };
+	ssize_t result = 0;
+	size_t written;
 
 	WARN_ONCE(wdata->server != mid->server,
 		  "wdata server %p != mid server %p",
@@ -4726,8 +4781,8 @@ smb2_writev_callback(struct mid_q_entry *mid)
 	case MID_RESPONSE_RECEIVED:
 		credits.value = le16_to_cpu(rsp->hdr.CreditRequest);
 		credits.instance = server->reconnect_instance;
-		wdata->result = smb2_check_receive(mid, server, 0);
-		if (wdata->result != 0)
+		result = smb2_check_receive(mid, server, 0);
+		if (result != 0)
 			break;
 
 		written = le32_to_cpu(rsp->DataLength);
@@ -4737,24 +4792,24 @@ smb2_writev_callback(struct mid_q_entry *mid)
 		 * client. OS/2 servers are known to set incorrect
 		 * CountHigh values.
 		 */
-		if (written > wdata->bytes)
+		if (written > wdata->subreq.len)
 			written &= 0xFFFF;
 
-		if (written < wdata->bytes)
+		if (written < wdata->subreq.len)
 			wdata->result = -ENOSPC;
 		else
-			wdata->bytes = written;
+			wdata->subreq.len = written;
 		break;
 	case MID_REQUEST_SUBMITTED:
 	case MID_RETRY_NEEDED:
-		wdata->result = -EAGAIN;
+		result = -EAGAIN;
 		break;
 	case MID_RESPONSE_MALFORMED:
 		credits.value = le16_to_cpu(rsp->hdr.CreditRequest);
 		credits.instance = server->reconnect_instance;
 		fallthrough;
 	default:
-		wdata->result = -EIO;
+		result = -EIO;
 		break;
 	}
 #ifdef CONFIG_CIFS_SMB_DIRECT
@@ -4770,46 +4825,43 @@ smb2_writev_callback(struct mid_q_entry *mid)
 		wdata->mr = NULL;
 	}
 #endif
-	if (wdata->result) {
+	if (result) {
 		cifs_stats_fail_inc(tcon, SMB2_WRITE_HE);
-		trace_smb3_write_err(0 /* no xid */,
-				     wdata->cfile->fid.persistent_fid,
-				     tcon->tid, tcon->ses->Suid, wdata->offset,
-				     wdata->bytes, wdata->result);
+		trace_smb3_write_err(wdata->xid,
+				     wdata->req->cfile->fid.persistent_fid,
+				     tcon->tid, tcon->ses->Suid, wdata->subreq.start,
+				     wdata->subreq.len, wdata->result);
 		if (wdata->result == -ENOSPC)
 			pr_warn_once("Out of space writing to %s\n",
 				     tcon->tree_name);
 	} else
 		trace_smb3_write_done(0 /* no xid */,
-				      wdata->cfile->fid.persistent_fid,
+				      wdata->req->cfile->fid.persistent_fid,
 				      tcon->tid, tcon->ses->Suid,
-				      wdata->offset, wdata->bytes);
+				      wdata->subreq.start, wdata->subreq.len);
 
-	queue_work(cifsiod_wq, &wdata->work);
+	wdata->credits.value = 0;
+	cifs_write_subrequest_terminated(wdata, result ?: written, true);
 	release_mid(mid);
 	add_credits(server, &credits, 0);
 }
 
 /* smb2_async_writev - send an async write, and set up mid to handle result */
-int
-smb2_async_writev(struct cifs_writedata *wdata,
-		  void (*release)(struct kref *kref))
+void
+smb2_async_writev(struct cifs_io_subrequest *wdata)
 {
 	int rc = -EACCES, flags = 0;
 	struct smb2_write_req *req = NULL;
 	struct smb2_hdr *shdr;
-	struct cifs_tcon *tcon = tlink_tcon(wdata->cfile->tlink);
+	struct cifs_tcon *tcon = tlink_tcon(wdata->req->cfile->tlink);
 	struct TCP_Server_Info *server = wdata->server;
 	struct kvec iov[1];
 	struct smb_rqst rqst = { };
-	unsigned int total_len;
+	unsigned int total_len, xid = wdata->xid;
 	struct cifs_io_parms _io_parms;
 	struct cifs_io_parms *io_parms = NULL;
 	int credit_request;
 
-	if (!wdata->server || wdata->replay)
-		server = wdata->server = cifs_pick_channel(tcon->ses);
-
 	/*
 	 * in future we may get cifs_io_parms passed in from the caller,
 	 * but for now we construct it here...
@@ -4817,18 +4869,18 @@ smb2_async_writev(struct cifs_writedata *wdata,
 	_io_parms = (struct cifs_io_parms) {
 		.tcon = tcon,
 		.server = server,
-		.offset = wdata->offset,
-		.length = wdata->bytes,
-		.persistent_fid = wdata->cfile->fid.persistent_fid,
-		.volatile_fid = wdata->cfile->fid.volatile_fid,
-		.pid = wdata->pid,
+		.offset = wdata->subreq.start,
+		.length = wdata->subreq.len,
+		.persistent_fid = wdata->req->cfile->fid.persistent_fid,
+		.volatile_fid = wdata->req->cfile->fid.volatile_fid,
+		.pid = wdata->req->pid,
 	};
 	io_parms = &_io_parms;
 
 	rc = smb2_plain_req_init(SMB2_WRITE, tcon, server,
 				 (void **) &req, &total_len);
 	if (rc)
-		return rc;
+		goto out;
 
 	if (smb3_encryption_required(tcon))
 		flags |= CIFS_TRANSFORM_REQ;
@@ -4846,7 +4898,7 @@ smb2_async_writev(struct cifs_writedata *wdata,
 				offsetof(struct smb2_write_req, Buffer));
 	req->RemainingBytes = 0;
 
-	trace_smb3_write_enter(0 /* xid */,
+	trace_smb3_write_enter(wdata->xid,
 			       io_parms->persistent_fid,
 			       io_parms->tcon->tid,
 			       io_parms->tcon->ses->Suid,
@@ -4860,10 +4912,10 @@ smb2_async_writev(struct cifs_writedata *wdata,
 	 */
 	if (smb3_use_rdma_offload(io_parms)) {
 		struct smbd_buffer_descriptor_v1 *v1;
-		size_t data_size = iov_iter_count(&wdata->iter);
+		size_t data_size = iov_iter_count(&wdata->subreq.io_iter);
 		bool need_invalidate = server->dialect == SMB30_PROT_ID;
 
-		wdata->mr = smbd_register_mr(server->smbd_conn, &wdata->iter,
+		wdata->mr = smbd_register_mr(server->smbd_conn, &wdata->subreq.io_iter,
 					     false, need_invalidate);
 		if (!wdata->mr) {
 			rc = -EAGAIN;
@@ -4890,9 +4942,9 @@ smb2_async_writev(struct cifs_writedata *wdata,
 
 	rqst.rq_iov = iov;
 	rqst.rq_nvec = 1;
-	rqst.rq_iter = wdata->iter;
+	rqst.rq_iter = wdata->subreq.io_iter;
 	rqst.rq_iter_size = iov_iter_count(&rqst.rq_iter);
-	if (wdata->replay)
+	if (test_bit(NETFS_SREQ_RETRYING, &wdata->subreq.flags))
 		smb2_set_replay(server, &rqst);
 #ifdef CONFIG_CIFS_SMB_DIRECT
 	if (wdata->mr)
@@ -4910,7 +4962,7 @@ smb2_async_writev(struct cifs_writedata *wdata,
 #endif
 
 	if (wdata->credits.value > 0) {
-		shdr->CreditCharge = cpu_to_le16(DIV_ROUND_UP(wdata->bytes,
+		shdr->CreditCharge = cpu_to_le16(DIV_ROUND_UP(wdata->subreq.len,
 						    SMB2_MAX_BUFFER_SIZE));
 		credit_request = le16_to_cpu(shdr->CreditCharge) + 8;
 		if (server->credits >= server->max_credits)
@@ -4927,25 +4979,27 @@ smb2_async_writev(struct cifs_writedata *wdata,
 		flags |= CIFS_HAS_CREDITS;
 	}
 
-	kref_get(&wdata->refcount);
 	rc = cifs_call_async(server, &rqst, NULL, smb2_writev_callback, NULL,
 			     wdata, flags, &wdata->credits);
-
+	/* Can't touch wdata if rc == 0 */
 	if (rc) {
-		trace_smb3_write_err(0 /* no xid */,
+		trace_smb3_write_err(xid,
 				     io_parms->persistent_fid,
 				     io_parms->tcon->tid,
 				     io_parms->tcon->ses->Suid,
 				     io_parms->offset,
 				     io_parms->length,
 				     rc);
-		kref_put(&wdata->refcount, release);
 		cifs_stats_fail_inc(tcon, SMB2_WRITE_HE);
 	}
 
 async_writev_out:
 	cifs_small_buf_release(req);
-	return rc;
+out:
+	if (rc) {
+		add_credits_and_wake_if(wdata->server, &wdata->credits, 0);
+		cifs_write_subrequest_terminated(wdata, rc, true);
+	}
 }
 
 /*
diff --git a/fs/smb/client/smb2pdu.h b/fs/smb/client/smb2pdu.h
index b00f707bdd..5c458ab3b0 100644
--- a/fs/smb/client/smb2pdu.h
+++ b/fs/smb/client/smb2pdu.h
@@ -117,9 +117,10 @@ struct share_redirect_error_context_rsp {
  * [4] : posix context
  * [5] : time warp context
  * [6] : query id context
- * [7] : compound padding
+ * [7] : create ea context
+ * [8] : compound padding
  */
-#define SMB2_CREATE_IOV_SIZE 8
+#define SMB2_CREATE_IOV_SIZE 9
 
 /*
  * Maximum size of a SMB2_CREATE response is 64 (smb2 header) +
@@ -144,7 +145,7 @@ struct durable_context_v2 {
 } __packed;
 
 struct create_durable_v2 {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8   Name[8];
 	struct durable_context_v2 dcontext;
 } __packed;
@@ -166,7 +167,7 @@ struct durable_reconnect_context_v2_rsp {
 } __packed;
 
 struct create_durable_handle_reconnect_v2 {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8   Name[8];
 	struct durable_reconnect_context_v2 dcontext;
 	__u8   Pad[4];
@@ -174,7 +175,7 @@ struct create_durable_handle_reconnect_v2 {
 
 /* See MS-SMB2 2.2.13.2.5 */
 struct crt_twarp_ctxt {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8	Name[8];
 	__le64	Timestamp;
 
@@ -182,12 +183,12 @@ struct crt_twarp_ctxt {
 
 /* See MS-SMB2 2.2.13.2.9 */
 struct crt_query_id_ctxt {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8	Name[8];
 } __packed;
 
 struct crt_sd_ctxt {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8	Name[8];
 	struct smb3_sd sd;
 } __packed;
@@ -413,4 +414,35 @@ struct smb2_posix_info_parsed {
 	const u8 *name;
 };
 
+struct smb2_create_ea_ctx {
+	struct create_context_hdr ctx;
+	__u8 name[8];
+	struct smb2_file_full_ea_info ea;
+} __packed;
+
+#define SMB2_WSL_XATTR_UID		"$LXUID"
+#define SMB2_WSL_XATTR_GID		"$LXGID"
+#define SMB2_WSL_XATTR_MODE		"$LXMOD"
+#define SMB2_WSL_XATTR_DEV		"$LXDEV"
+#define SMB2_WSL_XATTR_NAME_LEN	6
+#define SMB2_WSL_NUM_XATTRS		4
+
+#define SMB2_WSL_XATTR_UID_SIZE	4
+#define SMB2_WSL_XATTR_GID_SIZE	4
+#define SMB2_WSL_XATTR_MODE_SIZE	4
+#define SMB2_WSL_XATTR_DEV_SIZE	8
+
+#define SMB2_WSL_MIN_QUERY_EA_RESP_SIZE \
+	(ALIGN((SMB2_WSL_NUM_XATTRS - 1) * \
+	       (SMB2_WSL_XATTR_NAME_LEN + 1 + \
+		sizeof(struct smb2_file_full_ea_info)), 4) + \
+	 SMB2_WSL_XATTR_NAME_LEN + 1 + sizeof(struct smb2_file_full_ea_info))
+
+#define SMB2_WSL_MAX_QUERY_EA_RESP_SIZE \
+	(ALIGN(SMB2_WSL_MIN_QUERY_EA_RESP_SIZE + \
+	       SMB2_WSL_XATTR_UID_SIZE + \
+	       SMB2_WSL_XATTR_GID_SIZE + \
+	       SMB2_WSL_XATTR_MODE_SIZE + \
+	       SMB2_WSL_XATTR_DEV_SIZE, 4))
+
 #endif				/* _SMB2PDU_H */
diff --git a/fs/smb/client/smb2proto.h b/fs/smb/client/smb2proto.h
index b3069911e9..b208232b12 100644
--- a/fs/smb/client/smb2proto.h
+++ b/fs/smb/client/smb2proto.h
@@ -61,7 +61,8 @@ struct inode *smb2_get_reparse_inode(struct cifs_open_info_data *data,
 				     const unsigned int xid,
 				     struct cifs_tcon *tcon,
 				     const char *full_path,
-				     struct kvec *iov);
+				     struct kvec *reparse_iov,
+				     struct kvec *xattr_iov);
 int smb2_query_reparse_point(const unsigned int xid,
 			     struct cifs_tcon *tcon,
 			     struct cifs_sb_info *cifs_sb,
@@ -75,7 +76,8 @@ int smb2_query_path_info(const unsigned int xid,
 			 struct cifs_open_info_data *data);
 extern int smb2_set_path_size(const unsigned int xid, struct cifs_tcon *tcon,
 			      const char *full_path, __u64 size,
-			      struct cifs_sb_info *cifs_sb, bool set_alloc);
+			      struct cifs_sb_info *cifs_sb, bool set_alloc,
+				  struct dentry *dentry);
 extern int smb2_set_file_info(struct inode *inode, const char *full_path,
 			      FILE_BASIC_INFO *buf, const unsigned int xid);
 extern int smb311_posix_mkdir(const unsigned int xid, struct inode *inode,
@@ -91,7 +93,8 @@ extern void smb2_mkdir_setinfo(struct inode *inode, const char *full_path,
 extern int smb2_rmdir(const unsigned int xid, struct cifs_tcon *tcon,
 		      const char *name, struct cifs_sb_info *cifs_sb);
 extern int smb2_unlink(const unsigned int xid, struct cifs_tcon *tcon,
-		       const char *name, struct cifs_sb_info *cifs_sb);
+		       const char *name, struct cifs_sb_info *cifs_sb,
+			   struct dentry *dentry);
 int smb2_rename_path(const unsigned int xid,
 		     struct cifs_tcon *tcon,
 		     struct dentry *source_dentry,
@@ -207,11 +210,10 @@ extern int SMB2_query_acl(const unsigned int xid, struct cifs_tcon *tcon,
 extern int SMB2_get_srv_num(const unsigned int xid, struct cifs_tcon *tcon,
 			    u64 persistent_fid, u64 volatile_fid,
 			    __le64 *uniqueid);
-extern int smb2_async_readv(struct cifs_readdata *rdata);
+extern int smb2_async_readv(struct cifs_io_subrequest *rdata);
 extern int SMB2_read(const unsigned int xid, struct cifs_io_parms *io_parms,
 		     unsigned int *nbytes, char **buf, int *buf_type);
-extern int smb2_async_writev(struct cifs_writedata *wdata,
-			     void (*release)(struct kref *kref));
+extern void smb2_async_writev(struct cifs_io_subrequest *wdata);
 extern int SMB2_write(const unsigned int xid, struct cifs_io_parms *io_parms,
 		      unsigned int *nbytes, struct kvec *iov, int n_vec);
 extern int SMB2_echo(struct TCP_Server_Info *server);
@@ -308,5 +310,11 @@ int smb311_posix_query_path_info(const unsigned int xid,
 int posix_info_parse(const void *beg, const void *end,
 		     struct smb2_posix_info_parsed *out);
 int posix_info_sid_size(const void *beg, const void *end);
+int smb2_create_reparse_symlink(const unsigned int xid, struct inode *inode,
+				struct dentry *dentry, struct cifs_tcon *tcon,
+				const char *full_path, const char *symname);
+int smb2_make_nfs_node(unsigned int xid, struct inode *inode,
+		       struct dentry *dentry, struct cifs_tcon *tcon,
+		       const char *full_path, umode_t mode, dev_t dev);
 
 #endif			/* _SMB2PROTO_H */
diff --git a/fs/smb/client/smb2transport.c b/fs/smb/client/smb2transport.c
index 5a3ca62d2f..1476c445ca 100644
--- a/fs/smb/client/smb2transport.c
+++ b/fs/smb/client/smb2transport.c
@@ -189,6 +189,8 @@ smb2_find_smb_sess_tcon_unlocked(struct cifs_ses *ses, __u32  tid)
 		if (tcon->tid != tid)
 			continue;
 		++tcon->tc_count;
+		trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+				    netfs_trace_tcon_ref_get_find_sess_tcon);
 		return tcon;
 	}
 
@@ -214,8 +216,8 @@ smb2_find_smb_tcon(struct TCP_Server_Info *server, __u64 ses_id, __u32  tid)
 	}
 	tcon = smb2_find_smb_sess_tcon_unlocked(ses, tid);
 	if (!tcon) {
-		cifs_put_smb_ses(ses);
 		spin_unlock(&cifs_tcp_ses_lock);
+		cifs_put_smb_ses(ses);
 		return NULL;
 	}
 	spin_unlock(&cifs_tcp_ses_lock);
@@ -659,7 +661,7 @@ smb2_sign_rqst(struct smb_rqst *rqst, struct TCP_Server_Info *server)
 	}
 	spin_unlock(&server->srv_lock);
 	if (!is_binding && !server->session_estab) {
-		strncpy(shdr->Signature, "BSRSPYL", 8);
+		strscpy(shdr->Signature, "BSRSPYL");
 		return 0;
 	}
 
diff --git a/fs/smb/client/trace.h b/fs/smb/client/trace.h
index 522fa387fc..36d47ce596 100644
--- a/fs/smb/client/trace.h
+++ b/fs/smb/client/trace.h
@@ -3,6 +3,9 @@
  *   Copyright (C) 2018, Microsoft Corporation.
  *
  *   Author(s): Steve French <stfrench@microsoft.com>
+ *
+ * Please use this 3-part article as a reference for writing new tracepoints:
+ * https://lwn.net/Articles/379903/
  */
 #undef TRACE_SYSTEM
 #define TRACE_SYSTEM cifs
@@ -15,12 +18,129 @@
 #include <linux/inet.h>
 
 /*
- * Please use this 3-part article as a reference for writing new tracepoints:
- * https://lwn.net/Articles/379903/
+ * Specify enums for tracing information.
  */
+#define smb3_tcon_ref_traces					      \
+	EM(netfs_trace_tcon_ref_dec_dfs_refer,		"DEC DfsRef") \
+	EM(netfs_trace_tcon_ref_free,			"FRE       ") \
+	EM(netfs_trace_tcon_ref_free_fail,		"FRE Fail  ") \
+	EM(netfs_trace_tcon_ref_free_ipc,		"FRE Ipc   ") \
+	EM(netfs_trace_tcon_ref_free_ipc_fail,		"FRE Ipc-F ") \
+	EM(netfs_trace_tcon_ref_free_reconnect_server,	"FRE Reconn") \
+	EM(netfs_trace_tcon_ref_get_cancelled_close,	"GET Cn-Cls") \
+	EM(netfs_trace_tcon_ref_get_dfs_refer,		"GET DfsRef") \
+	EM(netfs_trace_tcon_ref_get_find,		"GET Find  ") \
+	EM(netfs_trace_tcon_ref_get_find_sess_tcon,	"GET FndSes") \
+	EM(netfs_trace_tcon_ref_get_reconnect_server,	"GET Reconn") \
+	EM(netfs_trace_tcon_ref_new,			"NEW       ") \
+	EM(netfs_trace_tcon_ref_new_ipc,		"NEW Ipc   ") \
+	EM(netfs_trace_tcon_ref_new_reconnect_server,	"NEW Reconn") \
+	EM(netfs_trace_tcon_ref_put_cancelled_close,	"PUT Cn-Cls") \
+	EM(netfs_trace_tcon_ref_put_cancelled_close_fid, "PUT Cn-Fid") \
+	EM(netfs_trace_tcon_ref_put_cancelled_mid,	"PUT Cn-Mid") \
+	EM(netfs_trace_tcon_ref_put_mnt_ctx,		"PUT MntCtx") \
+	EM(netfs_trace_tcon_ref_put_reconnect_server,	"PUT Reconn") \
+	EM(netfs_trace_tcon_ref_put_tlink,		"PUT Tlink ") \
+	EM(netfs_trace_tcon_ref_see_cancelled_close,	"SEE Cn-Cls") \
+	EM(netfs_trace_tcon_ref_see_fscache_collision,	"SEE FV-CO!") \
+	EM(netfs_trace_tcon_ref_see_fscache_okay,	"SEE FV-Ok ") \
+	EM(netfs_trace_tcon_ref_see_fscache_relinq,	"SEE FV-Rlq") \
+	E_(netfs_trace_tcon_ref_see_umount,		"SEE Umount")
+
+#undef EM
+#undef E_
+
+/*
+ * Define those tracing enums.
+ */
+#ifndef __SMB3_DECLARE_TRACE_ENUMS_ONCE_ONLY
+#define __SMB3_DECLARE_TRACE_ENUMS_ONCE_ONLY
+
+#define EM(a, b) a,
+#define E_(a, b) a
+
+enum smb3_tcon_ref_trace { smb3_tcon_ref_traces } __mode(byte);
+
+#undef EM
+#undef E_
+#endif
+
+/*
+ * Export enum symbols via userspace.
+ */
+#define EM(a, b) TRACE_DEFINE_ENUM(a);
+#define E_(a, b) TRACE_DEFINE_ENUM(a);
+
+smb3_tcon_ref_traces;
+
+#undef EM
+#undef E_
+
+/*
+ * Now redefine the EM() and E_() macros to map the enums to the strings that
+ * will be printed in the output.
+ */
+#define EM(a, b)	{ a, b },
+#define E_(a, b)	{ a, b }
 
 /* For logging errors in read or write */
 DECLARE_EVENT_CLASS(smb3_rw_err_class,
+	TP_PROTO(unsigned int rreq_debug_id,
+		 unsigned int rreq_debug_index,
+		 unsigned int xid,
+		 __u64	fid,
+		 __u32	tid,
+		 __u64	sesid,
+		 __u64	offset,
+		 __u32	len,
+		 int	rc),
+	TP_ARGS(rreq_debug_id, rreq_debug_index,
+		xid, fid, tid, sesid, offset, len, rc),
+	TP_STRUCT__entry(
+		__field(unsigned int, rreq_debug_id)
+		__field(unsigned int, rreq_debug_index)
+		__field(unsigned int, xid)
+		__field(__u64, fid)
+		__field(__u32, tid)
+		__field(__u64, sesid)
+		__field(__u64, offset)
+		__field(__u32, len)
+		__field(int, rc)
+	),
+	TP_fast_assign(
+		__entry->rreq_debug_id = rreq_debug_id;
+		__entry->rreq_debug_index = rreq_debug_index;
+		__entry->xid = xid;
+		__entry->fid = fid;
+		__entry->tid = tid;
+		__entry->sesid = sesid;
+		__entry->offset = offset;
+		__entry->len = len;
+		__entry->rc = rc;
+	),
+	TP_printk("\tR=%08x[%x] xid=%u sid=0x%llx tid=0x%x fid=0x%llx offset=0x%llx len=0x%x rc=%d",
+		  __entry->rreq_debug_id, __entry->rreq_debug_index,
+		  __entry->xid, __entry->sesid, __entry->tid, __entry->fid,
+		  __entry->offset, __entry->len, __entry->rc)
+)
+
+#define DEFINE_SMB3_RW_ERR_EVENT(name)          \
+DEFINE_EVENT(smb3_rw_err_class, smb3_##name,    \
+	TP_PROTO(unsigned int rreq_debug_id,	\
+		 unsigned int rreq_debug_index,		\
+		 unsigned int xid,			\
+		 __u64	fid,				\
+		 __u32	tid,				\
+		 __u64	sesid,				\
+		 __u64	offset,				\
+		 __u32	len,				\
+		 int	rc),				\
+	TP_ARGS(rreq_debug_id, rreq_debug_index, xid, fid, tid, sesid, offset, len, rc))
+
+DEFINE_SMB3_RW_ERR_EVENT(read_err);
+
+/* For logging errors in other file I/O ops */
+DECLARE_EVENT_CLASS(smb3_other_err_class,
 	TP_PROTO(unsigned int xid,
 		__u64	fid,
 		__u32	tid,
@@ -52,8 +172,8 @@ DECLARE_EVENT_CLASS(smb3_rw_err_class,
 		__entry->offset, __entry->len, __entry->rc)
 )
 
-#define DEFINE_SMB3_RW_ERR_EVENT(name)          \
-DEFINE_EVENT(smb3_rw_err_class, smb3_##name,    \
+#define DEFINE_SMB3_OTHER_ERR_EVENT(name)	\
+DEFINE_EVENT(smb3_other_err_class, smb3_##name, \
 	TP_PROTO(unsigned int xid,		\
 		__u64	fid,			\
 		__u32	tid,			\
@@ -63,15 +183,67 @@ DEFINE_EVENT(smb3_rw_err_class, smb3_##name,    \
 		int	rc),			\
 	TP_ARGS(xid, fid, tid, sesid, offset, len, rc))
 
-DEFINE_SMB3_RW_ERR_EVENT(write_err);
-DEFINE_SMB3_RW_ERR_EVENT(read_err);
-DEFINE_SMB3_RW_ERR_EVENT(query_dir_err);
-DEFINE_SMB3_RW_ERR_EVENT(zero_err);
-DEFINE_SMB3_RW_ERR_EVENT(falloc_err);
+DEFINE_SMB3_OTHER_ERR_EVENT(write_err);
+DEFINE_SMB3_OTHER_ERR_EVENT(query_dir_err);
+DEFINE_SMB3_OTHER_ERR_EVENT(zero_err);
+DEFINE_SMB3_OTHER_ERR_EVENT(falloc_err);
 
 
 /* For logging successful read or write */
 DECLARE_EVENT_CLASS(smb3_rw_done_class,
+	TP_PROTO(unsigned int rreq_debug_id,
+		 unsigned int rreq_debug_index,
+		 unsigned int xid,
+		 __u64	fid,
+		 __u32	tid,
+		 __u64	sesid,
+		 __u64	offset,
+		 __u32	len),
+	TP_ARGS(rreq_debug_id, rreq_debug_index,
+		xid, fid, tid, sesid, offset, len),
+	TP_STRUCT__entry(
+		__field(unsigned int, rreq_debug_id)
+		__field(unsigned int, rreq_debug_index)
+		__field(unsigned int, xid)
+		__field(__u64, fid)
+		__field(__u32, tid)
+		__field(__u64, sesid)
+		__field(__u64, offset)
+		__field(__u32, len)
+	),
+	TP_fast_assign(
+		__entry->rreq_debug_id = rreq_debug_id;
+		__entry->rreq_debug_index = rreq_debug_index;
+		__entry->xid = xid;
+		__entry->fid = fid;
+		__entry->tid = tid;
+		__entry->sesid = sesid;
+		__entry->offset = offset;
+		__entry->len = len;
+	),
+	TP_printk("R=%08x[%x] xid=%u sid=0x%llx tid=0x%x fid=0x%llx offset=0x%llx len=0x%x",
+		  __entry->rreq_debug_id, __entry->rreq_debug_index,
+		  __entry->xid, __entry->sesid, __entry->tid, __entry->fid,
+		  __entry->offset, __entry->len)
+)
+
+#define DEFINE_SMB3_RW_DONE_EVENT(name)         \
+DEFINE_EVENT(smb3_rw_done_class, smb3_##name,   \
+	TP_PROTO(unsigned int rreq_debug_id,	\
+		 unsigned int rreq_debug_index,	\
+		 unsigned int xid,		\
+		__u64	fid,			\
+		__u32	tid,			\
+		__u64	sesid,			\
+		__u64	offset,			\
+		__u32	len),			\
+	TP_ARGS(rreq_debug_id, rreq_debug_index, xid, fid, tid, sesid, offset, len))
+
+DEFINE_SMB3_RW_DONE_EVENT(read_enter);
+DEFINE_SMB3_RW_DONE_EVENT(read_done);
+
+/* For logging successful other op */
+DECLARE_EVENT_CLASS(smb3_other_done_class,
 	TP_PROTO(unsigned int xid,
 		__u64	fid,
 		__u32	tid,
@@ -100,8 +272,8 @@ DECLARE_EVENT_CLASS(smb3_rw_done_class,
 		__entry->offset, __entry->len)
 )
 
-#define DEFINE_SMB3_RW_DONE_EVENT(name)         \
-DEFINE_EVENT(smb3_rw_done_class, smb3_##name,   \
+#define DEFINE_SMB3_OTHER_DONE_EVENT(name)         \
+DEFINE_EVENT(smb3_other_done_class, smb3_##name,   \
 	TP_PROTO(unsigned int xid,		\
 		__u64	fid,			\
 		__u32	tid,			\
@@ -110,16 +282,14 @@ DEFINE_EVENT(smb3_rw_done_class, smb3_##name,   \
 		__u32	len),			\
 	TP_ARGS(xid, fid, tid, sesid, offset, len))
 
-DEFINE_SMB3_RW_DONE_EVENT(write_enter);
-DEFINE_SMB3_RW_DONE_EVENT(read_enter);
-DEFINE_SMB3_RW_DONE_EVENT(query_dir_enter);
-DEFINE_SMB3_RW_DONE_EVENT(zero_enter);
-DEFINE_SMB3_RW_DONE_EVENT(falloc_enter);
-DEFINE_SMB3_RW_DONE_EVENT(write_done);
-DEFINE_SMB3_RW_DONE_EVENT(read_done);
-DEFINE_SMB3_RW_DONE_EVENT(query_dir_done);
-DEFINE_SMB3_RW_DONE_EVENT(zero_done);
-DEFINE_SMB3_RW_DONE_EVENT(falloc_done);
+DEFINE_SMB3_OTHER_DONE_EVENT(write_enter);
+DEFINE_SMB3_OTHER_DONE_EVENT(query_dir_enter);
+DEFINE_SMB3_OTHER_DONE_EVENT(zero_enter);
+DEFINE_SMB3_OTHER_DONE_EVENT(falloc_enter);
+DEFINE_SMB3_OTHER_DONE_EVENT(write_done);
+DEFINE_SMB3_OTHER_DONE_EVENT(query_dir_done);
+DEFINE_SMB3_OTHER_DONE_EVENT(zero_done);
+DEFINE_SMB3_OTHER_DONE_EVENT(falloc_done);
 
 /* For logging successful set EOF (truncate) */
 DECLARE_EVENT_CLASS(smb3_eof_class,
@@ -348,7 +518,7 @@ DECLARE_EVENT_CLASS(smb3_inf_compound_enter_class,
 		__entry->xid = xid;
 		__entry->tid = tid;
 		__entry->sesid = sesid;
-		__assign_str(path, full_path);
+		__assign_str(path);
 	),
 	TP_printk("xid=%u sid=0x%llx tid=0x%x path=%s",
 		__entry->xid, __entry->sesid, __entry->tid,
@@ -375,6 +545,7 @@ DEFINE_SMB3_INF_COMPOUND_ENTER_EVENT(get_reparse_compound_enter);
 DEFINE_SMB3_INF_COMPOUND_ENTER_EVENT(delete_enter);
 DEFINE_SMB3_INF_COMPOUND_ENTER_EVENT(mkdir_enter);
 DEFINE_SMB3_INF_COMPOUND_ENTER_EVENT(tdis_enter);
+DEFINE_SMB3_INF_COMPOUND_ENTER_EVENT(mknod_enter);
 
 DECLARE_EVENT_CLASS(smb3_inf_compound_done_class,
 	TP_PROTO(unsigned int xid,
@@ -411,10 +582,11 @@ DEFINE_SMB3_INF_COMPOUND_DONE_EVENT(set_eof_done);
 DEFINE_SMB3_INF_COMPOUND_DONE_EVENT(set_info_compound_done);
 DEFINE_SMB3_INF_COMPOUND_DONE_EVENT(set_reparse_compound_done);
 DEFINE_SMB3_INF_COMPOUND_DONE_EVENT(get_reparse_compound_done);
+DEFINE_SMB3_INF_COMPOUND_DONE_EVENT(query_wsl_ea_compound_done);
 DEFINE_SMB3_INF_COMPOUND_DONE_EVENT(delete_done);
 DEFINE_SMB3_INF_COMPOUND_DONE_EVENT(mkdir_done);
 DEFINE_SMB3_INF_COMPOUND_DONE_EVENT(tdis_done);
-
+DEFINE_SMB3_INF_COMPOUND_DONE_EVENT(mknod_done);
 
 DECLARE_EVENT_CLASS(smb3_inf_compound_err_class,
 	TP_PROTO(unsigned int xid,
@@ -456,9 +628,11 @@ DEFINE_SMB3_INF_COMPOUND_ERR_EVENT(set_eof_err);
 DEFINE_SMB3_INF_COMPOUND_ERR_EVENT(set_info_compound_err);
 DEFINE_SMB3_INF_COMPOUND_ERR_EVENT(set_reparse_compound_err);
 DEFINE_SMB3_INF_COMPOUND_ERR_EVENT(get_reparse_compound_err);
+DEFINE_SMB3_INF_COMPOUND_ERR_EVENT(query_wsl_ea_compound_err);
 DEFINE_SMB3_INF_COMPOUND_ERR_EVENT(mkdir_err);
 DEFINE_SMB3_INF_COMPOUND_ERR_EVENT(delete_err);
 DEFINE_SMB3_INF_COMPOUND_ERR_EVENT(tdis_err);
+DEFINE_SMB3_INF_COMPOUND_ERR_EVENT(mknod_err);
 
 /*
  * For logging SMB3 Status code and Command for responses which return errors
@@ -588,7 +762,7 @@ DECLARE_EVENT_CLASS(smb3_exit_err_class,
 	),
 	TP_fast_assign(
 		__entry->xid = xid;
-		__assign_str(func_name, func_name);
+		__assign_str(func_name);
 		__entry->rc = rc;
 	),
 	TP_printk("\t%s: xid=%u rc=%d",
@@ -641,7 +815,7 @@ DECLARE_EVENT_CLASS(smb3_enter_exit_class,
 	),
 	TP_fast_assign(
 		__entry->xid = xid;
-		__assign_str(func_name, func_name);
+		__assign_str(func_name);
 	),
 	TP_printk("\t%s: xid=%u",
 		__get_str(func_name), __entry->xid)
@@ -678,7 +852,7 @@ DECLARE_EVENT_CLASS(smb3_tcon_class,
 		__entry->xid = xid;
 		__entry->tid = tid;
 		__entry->sesid = sesid;
-		__assign_str(name, unc_name);
+		__assign_str(name);
 		__entry->rc = rc;
 	),
 	TP_printk("xid=%u sid=0x%llx tid=0x%x unc_name=%s rc=%d",
@@ -722,7 +896,7 @@ DECLARE_EVENT_CLASS(smb3_open_enter_class,
 		__entry->xid = xid;
 		__entry->tid = tid;
 		__entry->sesid = sesid;
-		__assign_str(path, full_path);
+		__assign_str(path);
 		__entry->create_options = create_options;
 		__entry->desired_access = desired_access;
 	),
@@ -924,7 +1098,7 @@ DECLARE_EVENT_CLASS(smb3_connect_class,
 		__entry->conn_id = conn_id;
 		pss = (struct sockaddr_storage *)__entry->dst_addr;
 		*pss = *dst_addr;
-		__assign_str(hostname, hostname);
+		__assign_str(hostname);
 	),
 	TP_printk("conn_id=0x%llx server=%s addr=%pISpsfc",
 		__entry->conn_id,
@@ -960,7 +1134,7 @@ DECLARE_EVENT_CLASS(smb3_connect_err_class,
 		__entry->rc = rc;
 		pss = (struct sockaddr_storage *)__entry->dst_addr;
 		*pss = *dst_addr;
-		__assign_str(hostname, hostname);
+		__assign_str(hostname);
 	),
 	TP_printk("rc=%d conn_id=0x%llx server=%s addr=%pISpsfc",
 		__entry->rc,
@@ -992,7 +1166,7 @@ DECLARE_EVENT_CLASS(smb3_reconnect_class,
 	TP_fast_assign(
 		__entry->currmid = currmid;
 		__entry->conn_id = conn_id;
-		__assign_str(hostname, hostname);
+		__assign_str(hostname);
 	),
 	TP_printk("conn_id=0x%llx server=%s current_mid=%llu",
 		__entry->conn_id,
@@ -1030,6 +1204,38 @@ DEFINE_EVENT(smb3_ses_class, smb3_##name,  \
 
 DEFINE_SMB3_SES_EVENT(ses_not_found);
 
+DECLARE_EVENT_CLASS(smb3_ioctl_class,
+	TP_PROTO(unsigned int xid,
+		__u64	fid,
+		unsigned int command),
+	TP_ARGS(xid, fid, command),
+	TP_STRUCT__entry(
+		__field(unsigned int, xid)
+		__field(__u64, fid)
+		__field(unsigned int, command)
+	),
+	TP_fast_assign(
+		__entry->xid = xid;
+		__entry->fid = fid;
+		__entry->command = command;
+	),
+	TP_printk("xid=%u fid=0x%llx ioctl cmd=0x%x",
+		__entry->xid, __entry->fid, __entry->command)
+)
+
+#define DEFINE_SMB3_IOCTL_EVENT(name)        \
+DEFINE_EVENT(smb3_ioctl_class, smb3_##name,  \
+	TP_PROTO(unsigned int xid,	     \
+		__u64 fid,		     \
+		unsigned int command),	     \
+	TP_ARGS(xid, fid, command))
+
+DEFINE_SMB3_IOCTL_EVENT(ioctl);
+
+
+
+
+
 DECLARE_EVENT_CLASS(smb3_credit_class,
 	TP_PROTO(__u64	currmid,
 		__u64 conn_id,
@@ -1049,7 +1255,7 @@ DECLARE_EVENT_CLASS(smb3_credit_class,
 	TP_fast_assign(
 		__entry->currmid = currmid;
 		__entry->conn_id = conn_id;
-		__assign_str(hostname, hostname);
+		__assign_str(hostname);
 		__entry->credits = credits;
 		__entry->credits_to_add = credits_to_add;
 		__entry->in_flight = in_flight;
@@ -1089,6 +1295,30 @@ DEFINE_SMB3_CREDIT_EVENT(waitff_credits);
 DEFINE_SMB3_CREDIT_EVENT(overflow_credits);
 DEFINE_SMB3_CREDIT_EVENT(set_credits);
 
+
+TRACE_EVENT(smb3_tcon_ref,
+	    TP_PROTO(unsigned int tcon_debug_id, int ref,
+		     enum smb3_tcon_ref_trace trace),
+	    TP_ARGS(tcon_debug_id, ref, trace),
+	    TP_STRUCT__entry(
+		    __field(unsigned int,		tcon)
+		    __field(int,			ref)
+		    __field(enum smb3_tcon_ref_trace,	trace)
+			     ),
+	    TP_fast_assign(
+		    __entry->tcon	= tcon_debug_id;
+		    __entry->ref	= ref;
+		    __entry->trace	= trace;
+			   ),
+	    TP_printk("TC=%08x %s r=%u",
+		      __entry->tcon,
+		      __print_symbolic(__entry->trace, smb3_tcon_ref_traces),
+		      __entry->ref)
+	    );
+
+
+#undef EM
+#undef E_
 #endif /* _CIFS_TRACE_H */
 
 #undef TRACE_INCLUDE_PATH
diff --git a/fs/smb/client/transport.c b/fs/smb/client/transport.c
index ddf1a3aafe..012b9bd069 100644
--- a/fs/smb/client/transport.c
+++ b/fs/smb/client/transport.c
@@ -691,8 +691,8 @@ wait_for_compound_request(struct TCP_Server_Info *server, int num,
 }
 
 int
-cifs_wait_mtu_credits(struct TCP_Server_Info *server, unsigned int size,
-		      unsigned int *num, struct cifs_credits *credits)
+cifs_wait_mtu_credits(struct TCP_Server_Info *server, size_t size,
+		      size_t *num, struct cifs_credits *credits)
 {
 	*num = size;
 	credits->value = 0;
@@ -1692,7 +1692,7 @@ __cifs_readv_discard(struct TCP_Server_Info *server, struct mid_q_entry *mid,
 static int
 cifs_readv_discard(struct TCP_Server_Info *server, struct mid_q_entry *mid)
 {
-	struct cifs_readdata *rdata = mid->callback_data;
+	struct cifs_io_subrequest *rdata = mid->callback_data;
 
 	return  __cifs_readv_discard(server, mid, rdata->result);
 }
@@ -1702,13 +1702,13 @@ cifs_readv_receive(struct TCP_Server_Info *server, struct mid_q_entry *mid)
 {
 	int length, len;
 	unsigned int data_offset, data_len;
-	struct cifs_readdata *rdata = mid->callback_data;
+	struct cifs_io_subrequest *rdata = mid->callback_data;
 	char *buf = server->smallbuf;
 	unsigned int buflen = server->pdu_size + HEADER_PREAMBLE_SIZE(server);
 	bool use_rdma_mr = false;
 
-	cifs_dbg(FYI, "%s: mid=%llu offset=%llu bytes=%u\n",
-		 __func__, mid->mid, rdata->offset, rdata->bytes);
+	cifs_dbg(FYI, "%s: mid=%llu offset=%llu bytes=%zu\n",
+		 __func__, mid->mid, rdata->subreq.start, rdata->subreq.len);
 
 	/*
 	 * read the rest of READ_RSP header (sans Data array), or whatever we
@@ -1813,8 +1813,11 @@ cifs_readv_receive(struct TCP_Server_Info *server, struct mid_q_entry *mid)
 		length = data_len; /* An RDMA read is already done. */
 	else
 #endif
-		length = cifs_read_iter_from_socket(server, &rdata->iter,
+	{
+		length = cifs_read_iter_from_socket(server, &rdata->subreq.io_iter,
 						    data_len);
+		iov_iter_revert(&rdata->subreq.io_iter, data_len);
+	}
 	if (length > 0)
 		rdata->got_bytes += length;
 	server->total_read += length;
diff --git a/fs/smb/common/cifs_arc4.c b/fs/smb/common/cifs_arc4.c
index 043e4cb839..df360ca478 100644
--- a/fs/smb/common/cifs_arc4.c
+++ b/fs/smb/common/cifs_arc4.c
@@ -10,6 +10,7 @@
 #include <linux/module.h>
 #include "arc4.h"
 
+MODULE_DESCRIPTION("ARC4 Cipher Algorithm");
 MODULE_LICENSE("GPL");
 
 int cifs_arc4_setkey(struct arc4_ctx *ctx, const u8 *in_key, unsigned int key_len)
diff --git a/fs/smb/common/cifs_md4.c b/fs/smb/common/cifs_md4.c
index 50f78cfc6c..7ee7f4dad9 100644
--- a/fs/smb/common/cifs_md4.c
+++ b/fs/smb/common/cifs_md4.c
@@ -24,6 +24,7 @@
 #include <asm/byteorder.h>
 #include "md4.h"
 
+MODULE_DESCRIPTION("MD4 Message Digest Algorithm (RFC1320)");
 MODULE_LICENSE("GPL");
 
 static inline u32 lshift(u32 x, unsigned int s)
diff --git a/fs/smb/common/smb2pdu.h b/fs/smb/common/smb2pdu.h
index a233a24352..c3ee42188d 100644
--- a/fs/smb/common/smb2pdu.h
+++ b/fs/smb/common/smb2pdu.h
@@ -208,38 +208,45 @@ struct smb2_transform_hdr {
 	__le64  SessionId;
 } __packed;
 
+/*
+ * These are simplified versions from the spec, as we don't need a fully fledged
+ * form of both unchained and chained structs.
+ *
+ * Moreover, even in chained compressed payloads, the initial compression header
+ * has the form of the unchained one -- i.e. it never has the
+ * OriginalPayloadSize field and ::Offset field always represent an offset
+ * (instead of a length, as it is in the chained header).
+ *
+ * See MS-SMB2 2.2.42 for more details.
+ */
+#define SMB2_COMPRESSION_FLAG_NONE	0x0000
+#define SMB2_COMPRESSION_FLAG_CHAINED	0x0001
 
-/* See MS-SMB2 2.2.42 */
-struct smb2_compression_transform_hdr_unchained {
-	__le32 ProtocolId;	/* 0xFC 'S' 'M' 'B' */
+struct smb2_compression_hdr {
+	__le32 ProtocolId; /* 0xFC 'S' 'M' 'B' */
 	__le32 OriginalCompressedSegmentSize;
 	__le16 CompressionAlgorithm;
 	__le16 Flags;
-	__le16 Length; /* if chained it is length, else offset */
+	__le32 Offset; /* this is the size of the uncompressed SMB2 header below */
+	/* uncompressed SMB2 header (READ or WRITE) goes here */
+	/* compressed data goes here */
 } __packed;
 
-/* See MS-SMB2 2.2.42.1 */
-#define SMB2_COMPRESSION_FLAG_NONE	0x0000
-#define SMB2_COMPRESSION_FLAG_CHAINED	0x0001
-
-struct compression_payload_header {
+/*
+ * ... OTOH, set compression payload header to always have OriginalPayloadSize
+ * as it's easier to pass the struct size minus sizeof(OriginalPayloadSize)
+ * than to juggle around the header/data memory.
+ */
+struct smb2_compression_payload_hdr {
 	__le16	CompressionAlgorithm;
 	__le16	Flags;
 	__le32	Length; /* length of compressed playload including field below if present */
-	/* __le32 OriginalPayloadSize; */ /* optional, present when LZNT1, LZ77, LZ77+Huffman */
+	__le32 OriginalPayloadSize; /* accounted when LZNT1, LZ77, LZ77+Huffman */
 } __packed;
 
-/* See MS-SMB2 2.2.42.2 */
-struct smb2_compression_transform_hdr_chained {
-	__le32 ProtocolId;	/* 0xFC 'S' 'M' 'B' */
-	__le32 OriginalCompressedSegmentSize;
-	/* struct compression_payload_header[] */
-} __packed;
-
-/* See MS-SMB2 2.2.42.2.2 */
-struct compression_pattern_payload_v1 {
-	__le16	Pattern;
-	__le16	Reserved1;
+struct smb2_compression_pattern_v1 {
+	__u8	Pattern;
+	__u8	Reserved1;
 	__le16	Reserved2;
 	__le32	Repetitions;
 } __packed;
@@ -273,15 +280,16 @@ struct smb3_blob_data {
 #define SE_GROUP_RESOURCE		0x20000000
 #define SE_GROUP_LOGON_ID		0xC0000000
 
-/* struct sid_attr_data is SidData array in BlobData format then le32 Attr */
-
 struct sid_array_data {
 	__le16 SidAttrCount;
 	/* SidAttrList - array of sid_attr_data structs */
 } __packed;
 
-struct luid_attr_data {
-
+/* struct sid_attr_data is SidData array in BlobData format then le32 Attr */
+struct sid_attr_data {
+	__le16 BlobSize;
+	__u8 BlobData[];
+	/* __le32 Attr */
 } __packed;
 
 /*
@@ -495,6 +503,7 @@ struct smb2_encryption_neg_context {
 #define SMB3_COMPRESS_LZ77_HUFF	cpu_to_le16(0x0003)
 /* Pattern scanning algorithm See MS-SMB2 3.1.4.4.1 */
 #define SMB3_COMPRESS_PATTERN	cpu_to_le16(0x0004) /* Pattern_V1 */
+#define SMB3_COMPRESS_LZ4	cpu_to_le16(0x0005)
 
 /* Compression Flags */
 #define SMB2_COMPRESSION_CAPABILITIES_FLAG_NONE		cpu_to_le32(0x00000000)
@@ -908,6 +917,40 @@ struct smb2_query_directory_rsp {
 	__u8   Buffer[];
 } __packed;
 
+/* DeviceType Flags */
+#define FILE_DEVICE_CD_ROM              0x00000002
+#define FILE_DEVICE_CD_ROM_FILE_SYSTEM  0x00000003
+#define FILE_DEVICE_DFS                 0x00000006
+#define FILE_DEVICE_DISK                0x00000007
+#define FILE_DEVICE_DISK_FILE_SYSTEM    0x00000008
+#define FILE_DEVICE_FILE_SYSTEM         0x00000009
+#define FILE_DEVICE_NAMED_PIPE          0x00000011
+#define FILE_DEVICE_NETWORK             0x00000012
+#define FILE_DEVICE_NETWORK_FILE_SYSTEM 0x00000014
+#define FILE_DEVICE_NULL                0x00000015
+#define FILE_DEVICE_PARALLEL_PORT       0x00000016
+#define FILE_DEVICE_PRINTER             0x00000018
+#define FILE_DEVICE_SERIAL_PORT         0x0000001b
+#define FILE_DEVICE_STREAMS             0x0000001e
+#define FILE_DEVICE_TAPE                0x0000001f
+#define FILE_DEVICE_TAPE_FILE_SYSTEM    0x00000020
+#define FILE_DEVICE_VIRTUAL_DISK        0x00000024
+#define FILE_DEVICE_NETWORK_REDIRECTOR  0x00000028
+
+/* Device Characteristics */
+#define FILE_REMOVABLE_MEDIA			0x00000001
+#define FILE_READ_ONLY_DEVICE			0x00000002
+#define FILE_FLOPPY_DISKETTE			0x00000004
+#define FILE_WRITE_ONCE_MEDIA			0x00000008
+#define FILE_REMOTE_DEVICE			0x00000010
+#define FILE_DEVICE_IS_MOUNTED			0x00000020
+#define FILE_VIRTUAL_VOLUME			0x00000040
+#define FILE_DEVICE_SECURE_OPEN			0x00000100
+#define FILE_CHARACTERISTIC_TS_DEVICE		0x00001000
+#define FILE_CHARACTERISTIC_WEBDAV_DEVICE	0x00002000
+#define FILE_PORTABLE_DEVICE			0x00004000
+#define FILE_DEVICE_ALLOW_APPCONTAINER_TRAVERSAL 0x00020000
+
 /*
  * Maximum number of iovs we need for a set-info request.
  * The largest one is rename/hardlink
@@ -1162,12 +1205,15 @@ struct smb2_server_client_notification {
 #define SMB2_CREATE_FLAG_REPARSEPOINT 0x01
 
 struct create_context {
-	__le32 Next;
-	__le16 NameOffset;
-	__le16 NameLength;
-	__le16 Reserved;
-	__le16 DataOffset;
-	__le32 DataLength;
+	/* New members must be added within the struct_group() macro below. */
+	__struct_group(create_context_hdr, hdr, __packed,
+		__le32 Next;
+		__le16 NameOffset;
+		__le16 NameLength;
+		__le16 Reserved;
+		__le16 DataOffset;
+		__le32 DataLength;
+	);
 	__u8 Buffer[];
 } __packed;
 
@@ -1213,7 +1259,7 @@ struct smb2_create_rsp {
 } __packed;
 
 struct create_posix {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8    Name[16];
 	__le32  Mode;
 	__u32   Reserved;
@@ -1221,7 +1267,7 @@ struct create_posix {
 
 /* See MS-SMB2 2.2.13.2.3 and MS-SMB2 2.2.13.2.4 */
 struct create_durable {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8   Name[8];
 	union {
 		__u8  Reserved[16];
@@ -1234,14 +1280,14 @@ struct create_durable {
 
 /* See MS-SMB2 2.2.13.2.5 */
 struct create_mxac_req {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8   Name[8];
 	__le64 Timestamp;
 } __packed;
 
 /* See MS-SMB2 2.2.14.2.5 */
 struct create_mxac_rsp {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8   Name[8];
 	__le32 QueryStatus;
 	__le32 MaximalAccess;
@@ -1277,13 +1323,13 @@ struct lease_context_v2 {
 } __packed;
 
 struct create_lease {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8   Name[8];
 	struct lease_context lcontext;
 } __packed;
 
 struct create_lease_v2 {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8   Name[8];
 	struct lease_context_v2 lcontext;
 	__u8   Pad[4];
@@ -1291,7 +1337,7 @@ struct create_lease_v2 {
 
 /* See MS-SMB2 2.2.14.2.9 */
 struct create_disk_id_rsp {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8   Name[8];
 	__le64 DiskFileId;
 	__le64 VolumeId;
@@ -1300,7 +1346,7 @@ struct create_disk_id_rsp {
 
 /* See MS-SMB2 2.2.13.2.13 */
 struct create_app_inst_id {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8 Name[16];
 	__le32 StructureSize; /* Must be 20 */
 	__u16 Reserved;
@@ -1309,7 +1355,7 @@ struct create_app_inst_id {
 
 /* See MS-SMB2 2.2.13.2.15 */
 struct create_app_inst_id_vers {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8 Name[16];
 	__le32 StructureSize; /* Must be 24 */
 	__u16 Reserved;
diff --git a/fs/smb/common/smbfsctl.h b/fs/smb/common/smbfsctl.h
index edd7fc2a79..a94d658b88 100644
--- a/fs/smb/common/smbfsctl.h
+++ b/fs/smb/common/smbfsctl.h
@@ -158,12 +158,6 @@
 #define IO_REPARSE_TAG_LX_CHR	     0x80000025
 #define IO_REPARSE_TAG_LX_BLK	     0x80000026
 
-#define IO_REPARSE_TAG_LX_SYMLINK_LE	cpu_to_le32(0xA000001D)
-#define IO_REPARSE_TAG_AF_UNIX_LE	cpu_to_le32(0x80000023)
-#define IO_REPARSE_TAG_LX_FIFO_LE	cpu_to_le32(0x80000024)
-#define IO_REPARSE_TAG_LX_CHR_LE	cpu_to_le32(0x80000025)
-#define IO_REPARSE_TAG_LX_BLK_LE	cpu_to_le32(0x80000026)
-
 /* fsctl flags */
 /* If Flags is set to this value, the request is an FSCTL not ioctl request */
 #define SMB2_0_IOCTL_IS_FSCTL		0x00000001
diff --git a/fs/smb/server/glob.h b/fs/smb/server/glob.h
index 5b8f3e0ebd..d528b20b37 100644
--- a/fs/smb/server/glob.h
+++ b/fs/smb/server/glob.h
@@ -12,8 +12,6 @@
 #include "unicode.h"
 #include "vfs_cache.h"
 
-#define KSMBD_VERSION	"3.4.2"
-
 extern int ksmbd_debug_types;
 
 #define KSMBD_DEBUG_SMB		BIT(0)
diff --git a/fs/smb/server/ksmbd_netlink.h b/fs/smb/server/ksmbd_netlink.h
index 4464a62228..f4e5519993 100644
--- a/fs/smb/server/ksmbd_netlink.h
+++ b/fs/smb/server/ksmbd_netlink.h
@@ -75,6 +75,7 @@ struct ksmbd_heartbeat {
 #define KSMBD_GLOBAL_FLAG_SMB2_ENCRYPTION	BIT(1)
 #define KSMBD_GLOBAL_FLAG_SMB3_MULTICHANNEL	BIT(2)
 #define KSMBD_GLOBAL_FLAG_SMB2_ENCRYPTION_OFF	BIT(3)
+#define KSMBD_GLOBAL_FLAG_DURABLE_HANDLE	BIT(4)
 
 /*
  * IPC request for ksmbd server startup
@@ -339,23 +340,24 @@ enum KSMBD_TREE_CONN_STATUS {
 /*
  * Share config flags.
  */
-#define KSMBD_SHARE_FLAG_INVALID		(0)
-#define KSMBD_SHARE_FLAG_AVAILABLE		BIT(0)
-#define KSMBD_SHARE_FLAG_BROWSEABLE		BIT(1)
-#define KSMBD_SHARE_FLAG_WRITEABLE		BIT(2)
-#define KSMBD_SHARE_FLAG_READONLY		BIT(3)
-#define KSMBD_SHARE_FLAG_GUEST_OK		BIT(4)
-#define KSMBD_SHARE_FLAG_GUEST_ONLY		BIT(5)
-#define KSMBD_SHARE_FLAG_STORE_DOS_ATTRS	BIT(6)
-#define KSMBD_SHARE_FLAG_OPLOCKS		BIT(7)
-#define KSMBD_SHARE_FLAG_PIPE			BIT(8)
-#define KSMBD_SHARE_FLAG_HIDE_DOT_FILES		BIT(9)
-#define KSMBD_SHARE_FLAG_INHERIT_OWNER		BIT(10)
-#define KSMBD_SHARE_FLAG_STREAMS		BIT(11)
-#define KSMBD_SHARE_FLAG_FOLLOW_SYMLINKS	BIT(12)
-#define KSMBD_SHARE_FLAG_ACL_XATTR		BIT(13)
-#define KSMBD_SHARE_FLAG_UPDATE			BIT(14)
-#define KSMBD_SHARE_FLAG_CROSSMNT		BIT(15)
+#define KSMBD_SHARE_FLAG_INVALID			(0)
+#define KSMBD_SHARE_FLAG_AVAILABLE			BIT(0)
+#define KSMBD_SHARE_FLAG_BROWSEABLE			BIT(1)
+#define KSMBD_SHARE_FLAG_WRITEABLE			BIT(2)
+#define KSMBD_SHARE_FLAG_READONLY			BIT(3)
+#define KSMBD_SHARE_FLAG_GUEST_OK			BIT(4)
+#define KSMBD_SHARE_FLAG_GUEST_ONLY			BIT(5)
+#define KSMBD_SHARE_FLAG_STORE_DOS_ATTRS		BIT(6)
+#define KSMBD_SHARE_FLAG_OPLOCKS			BIT(7)
+#define KSMBD_SHARE_FLAG_PIPE				BIT(8)
+#define KSMBD_SHARE_FLAG_HIDE_DOT_FILES			BIT(9)
+#define KSMBD_SHARE_FLAG_INHERIT_OWNER			BIT(10)
+#define KSMBD_SHARE_FLAG_STREAMS			BIT(11)
+#define KSMBD_SHARE_FLAG_FOLLOW_SYMLINKS		BIT(12)
+#define KSMBD_SHARE_FLAG_ACL_XATTR			BIT(13)
+#define KSMBD_SHARE_FLAG_UPDATE				BIT(14)
+#define KSMBD_SHARE_FLAG_CROSSMNT			BIT(15)
+#define KSMBD_SHARE_FLAG_CONTINUOUS_AVAILABILITY	BIT(16)
 
 /*
  * Tree connect request flags.
diff --git a/fs/smb/server/mgmt/user_session.c b/fs/smb/server/mgmt/user_session.c
index 15f68ee050..aec0a7a124 100644
--- a/fs/smb/server/mgmt/user_session.c
+++ b/fs/smb/server/mgmt/user_session.c
@@ -156,7 +156,7 @@ void ksmbd_session_destroy(struct ksmbd_session *sess)
 	kfree(sess);
 }
 
-static struct ksmbd_session *__session_lookup(unsigned long long id)
+struct ksmbd_session *__session_lookup(unsigned long long id)
 {
 	struct ksmbd_session *sess;
 
@@ -305,6 +305,32 @@ struct preauth_session *ksmbd_preauth_session_alloc(struct ksmbd_conn *conn,
 	return sess;
 }
 
+void destroy_previous_session(struct ksmbd_conn *conn,
+			      struct ksmbd_user *user, u64 id)
+{
+	struct ksmbd_session *prev_sess;
+	struct ksmbd_user *prev_user;
+
+	down_write(&sessions_table_lock);
+	down_write(&conn->session_lock);
+	prev_sess = __session_lookup(id);
+	if (!prev_sess || prev_sess->state == SMB2_SESSION_EXPIRED)
+		goto out;
+
+	prev_user = prev_sess->user;
+	if (!prev_user ||
+	    strcmp(user->name, prev_user->name) ||
+	    user->passkey_sz != prev_user->passkey_sz ||
+	    memcmp(user->passkey, prev_user->passkey, user->passkey_sz))
+		goto out;
+
+	ksmbd_destroy_file_table(&prev_sess->file_table);
+	prev_sess->state = SMB2_SESSION_EXPIRED;
+out:
+	up_write(&conn->session_lock);
+	up_write(&sessions_table_lock);
+}
+
 static bool ksmbd_preauth_session_id_match(struct preauth_session *sess,
 					   unsigned long long id)
 {
diff --git a/fs/smb/server/mgmt/user_session.h b/fs/smb/server/mgmt/user_session.h
index 63cb08fffd..dc9fded2cd 100644
--- a/fs/smb/server/mgmt/user_session.h
+++ b/fs/smb/server/mgmt/user_session.h
@@ -88,8 +88,11 @@ struct ksmbd_session *ksmbd_session_lookup(struct ksmbd_conn *conn,
 int ksmbd_session_register(struct ksmbd_conn *conn,
 			   struct ksmbd_session *sess);
 void ksmbd_sessions_deregister(struct ksmbd_conn *conn);
+struct ksmbd_session *__session_lookup(unsigned long long id);
 struct ksmbd_session *ksmbd_session_lookup_all(struct ksmbd_conn *conn,
 					       unsigned long long id);
+void destroy_previous_session(struct ksmbd_conn *conn,
+			      struct ksmbd_user *user, u64 id);
 struct preauth_session *ksmbd_preauth_session_alloc(struct ksmbd_conn *conn,
 						    u64 sess_id);
 struct preauth_session *ksmbd_preauth_session_lookup(struct ksmbd_conn *conn,
diff --git a/fs/smb/server/oplock.c b/fs/smb/server/oplock.c
index 2292ca6ff0..a8f52c4ebb 100644
--- a/fs/smb/server/oplock.c
+++ b/fs/smb/server/oplock.c
@@ -159,7 +159,8 @@ static struct oplock_info *opinfo_get_list(struct ksmbd_inode *ci)
 	opinfo = list_first_or_null_rcu(&ci->m_op_list, struct oplock_info,
 					op_entry);
 	if (opinfo) {
-		if (!atomic_inc_not_zero(&opinfo->refcount))
+		if (opinfo->conn == NULL ||
+		    !atomic_inc_not_zero(&opinfo->refcount))
 			opinfo = NULL;
 		else {
 			atomic_inc(&opinfo->conn->r_count);
@@ -206,9 +207,9 @@ static void opinfo_add(struct oplock_info *opinfo)
 {
 	struct ksmbd_inode *ci = opinfo->o_fp->f_ci;
 
-	write_lock(&ci->m_lock);
+	down_write(&ci->m_lock);
 	list_add_rcu(&opinfo->op_entry, &ci->m_op_list);
-	write_unlock(&ci->m_lock);
+	up_write(&ci->m_lock);
 }
 
 static void opinfo_del(struct oplock_info *opinfo)
@@ -220,9 +221,9 @@ static void opinfo_del(struct oplock_info *opinfo)
 		lease_del_list(opinfo);
 		write_unlock(&lease_list_lock);
 	}
-	write_lock(&ci->m_lock);
+	down_write(&ci->m_lock);
 	list_del_rcu(&opinfo->op_entry);
-	write_unlock(&ci->m_lock);
+	up_write(&ci->m_lock);
 }
 
 static unsigned long opinfo_count(struct ksmbd_file *fp)
@@ -525,21 +526,18 @@ static struct oplock_info *same_client_has_lease(struct ksmbd_inode *ci,
 	 * Compare lease key and client_guid to know request from same owner
 	 * of same client
 	 */
-	read_lock(&ci->m_lock);
+	down_read(&ci->m_lock);
 	list_for_each_entry(opinfo, &ci->m_op_list, op_entry) {
-		if (!opinfo->is_lease)
+		if (!opinfo->is_lease || !opinfo->conn)
 			continue;
-		read_unlock(&ci->m_lock);
 		lease = opinfo->o_lease;
 
 		ret = compare_guid_key(opinfo, client_guid, lctx->lease_key);
 		if (ret) {
 			m_opinfo = opinfo;
 			/* skip upgrading lease about breaking lease */
-			if (atomic_read(&opinfo->breaking_cnt)) {
-				read_lock(&ci->m_lock);
+			if (atomic_read(&opinfo->breaking_cnt))
 				continue;
-			}
 
 			/* upgrading lease */
 			if ((atomic_read(&ci->op_count) +
@@ -569,9 +567,8 @@ static struct oplock_info *same_client_has_lease(struct ksmbd_inode *ci,
 				lease_none_upgrade(opinfo, lctx->req_state);
 			}
 		}
-		read_lock(&ci->m_lock);
 	}
-	read_unlock(&ci->m_lock);
+	up_read(&ci->m_lock);
 
 	return m_opinfo;
 }
@@ -656,7 +653,7 @@ static void __smb2_oplock_break_noti(struct work_struct *wk)
 	struct smb2_hdr *rsp_hdr;
 	struct ksmbd_file *fp;
 
-	fp = ksmbd_lookup_durable_fd(br_info->fid);
+	fp = ksmbd_lookup_global_fd(br_info->fid);
 	if (!fp)
 		goto out;
 
@@ -1118,9 +1115,9 @@ void smb_send_parent_lease_break_noti(struct ksmbd_file *fp,
 	if (!p_ci)
 		return;
 
-	read_lock(&p_ci->m_lock);
+	down_read(&p_ci->m_lock);
 	list_for_each_entry(opinfo, &p_ci->m_op_list, op_entry) {
-		if (!opinfo->is_lease)
+		if (opinfo->conn == NULL || !opinfo->is_lease)
 			continue;
 
 		if (opinfo->o_lease->state != SMB2_OPLOCK_LEVEL_NONE &&
@@ -1136,13 +1133,11 @@ void smb_send_parent_lease_break_noti(struct ksmbd_file *fp,
 				continue;
 			}
 
-			read_unlock(&p_ci->m_lock);
 			oplock_break(opinfo, SMB2_OPLOCK_LEVEL_NONE);
 			opinfo_conn_put(opinfo);
-			read_lock(&p_ci->m_lock);
 		}
 	}
-	read_unlock(&p_ci->m_lock);
+	up_read(&p_ci->m_lock);
 
 	ksmbd_inode_put(p_ci);
 }
@@ -1156,16 +1151,16 @@ void smb_lazy_parent_lease_break_close(struct ksmbd_file *fp)
 	opinfo = rcu_dereference(fp->f_opinfo);
 	rcu_read_unlock();
 
-	if (!opinfo->is_lease || opinfo->o_lease->version != 2)
+	if (!opinfo || !opinfo->is_lease || opinfo->o_lease->version != 2)
 		return;
 
 	p_ci = ksmbd_inode_lookup_lock(fp->filp->f_path.dentry->d_parent);
 	if (!p_ci)
 		return;
 
-	read_lock(&p_ci->m_lock);
+	down_read(&p_ci->m_lock);
 	list_for_each_entry(opinfo, &p_ci->m_op_list, op_entry) {
-		if (!opinfo->is_lease)
+		if (opinfo->conn == NULL || !opinfo->is_lease)
 			continue;
 
 		if (opinfo->o_lease->state != SMB2_OPLOCK_LEVEL_NONE) {
@@ -1177,13 +1172,11 @@ void smb_lazy_parent_lease_break_close(struct ksmbd_file *fp)
 				atomic_dec(&opinfo->conn->r_count);
 				continue;
 			}
-			read_unlock(&p_ci->m_lock);
 			oplock_break(opinfo, SMB2_OPLOCK_LEVEL_NONE);
 			opinfo_conn_put(opinfo);
-			read_lock(&p_ci->m_lock);
 		}
 	}
-	read_unlock(&p_ci->m_lock);
+	up_read(&p_ci->m_lock);
 
 	ksmbd_inode_put(p_ci);
 }
@@ -1377,6 +1370,9 @@ void smb_break_all_levII_oplock(struct ksmbd_work *work, struct ksmbd_file *fp,
 
 	rcu_read_lock();
 	list_for_each_entry_rcu(brk_op, &ci->m_op_list, op_entry) {
+		if (brk_op->conn == NULL)
+			continue;
+
 		if (!atomic_inc_not_zero(&brk_op->refcount))
 			continue;
 
@@ -1513,11 +1509,10 @@ void create_lease_buf(u8 *rbuf, struct lease *lease)
 /**
  * parse_lease_state() - parse lease context containted in file open request
  * @open_req:	buffer containing smb2 file open(create) request
- * @is_dir:	whether leasing file is directory
  *
  * Return:  oplock state, -ENOENT if create lease context not found
  */
-struct lease_ctx_info *parse_lease_state(void *open_req, bool is_dir)
+struct lease_ctx_info *parse_lease_state(void *open_req)
 {
 	struct create_context *cc;
 	struct smb2_create_req *req = (struct smb2_create_req *)open_req;
@@ -1535,12 +1530,7 @@ struct lease_ctx_info *parse_lease_state(void *open_req, bool is_dir)
 		struct create_lease_v2 *lc = (struct create_lease_v2 *)cc;
 
 		memcpy(lreq->lease_key, lc->lcontext.LeaseKey, SMB2_LEASE_KEY_SIZE);
-		if (is_dir) {
-			lreq->req_state = lc->lcontext.LeaseState &
-				~SMB2_LEASE_WRITE_CACHING_LE;
-			lreq->is_dir = true;
-		} else
-			lreq->req_state = lc->lcontext.LeaseState;
+		lreq->req_state = lc->lcontext.LeaseState;
 		lreq->flags = lc->lcontext.LeaseFlags;
 		lreq->epoch = lc->lcontext.Epoch;
 		lreq->duration = lc->lcontext.LeaseDuration;
@@ -1664,6 +1654,8 @@ void create_durable_v2_rsp_buf(char *cc, struct ksmbd_file *fp)
 	buf->Name[3] = 'Q';
 
 	buf->Timeout = cpu_to_le32(fp->durable_timeout);
+	if (fp->is_persistent)
+		buf->Flags = cpu_to_le32(SMB2_DHANDLE_FLAG_PERSISTENT);
 }
 
 /**
@@ -1831,3 +1823,71 @@ out:
 	read_unlock(&lease_list_lock);
 	return ret_op;
 }
+
+int smb2_check_durable_oplock(struct ksmbd_conn *conn,
+			      struct ksmbd_share_config *share,
+			      struct ksmbd_file *fp,
+			      struct lease_ctx_info *lctx,
+			      char *name)
+{
+	struct oplock_info *opinfo = opinfo_get(fp);
+	int ret = 0;
+
+	if (!opinfo)
+		return 0;
+
+	if (opinfo->is_lease == false) {
+		if (lctx) {
+			pr_err("create context include lease\n");
+			ret = -EBADF;
+			goto out;
+		}
+
+		if (opinfo->level != SMB2_OPLOCK_LEVEL_BATCH) {
+			pr_err("oplock level is not equal to SMB2_OPLOCK_LEVEL_BATCH\n");
+			ret = -EBADF;
+		}
+
+		goto out;
+	}
+
+	if (memcmp(conn->ClientGUID, fp->client_guid,
+				SMB2_CLIENT_GUID_SIZE)) {
+		ksmbd_debug(SMB, "Client guid of fp is not equal to the one of connection\n");
+		ret = -EBADF;
+		goto out;
+	}
+
+	if (!lctx) {
+		ksmbd_debug(SMB, "create context does not include lease\n");
+		ret = -EBADF;
+		goto out;
+	}
+
+	if (memcmp(opinfo->o_lease->lease_key, lctx->lease_key,
+				SMB2_LEASE_KEY_SIZE)) {
+		ksmbd_debug(SMB,
+			    "lease key of fp does not match lease key in create context\n");
+		ret = -EBADF;
+		goto out;
+	}
+
+	if (!(opinfo->o_lease->state & SMB2_LEASE_HANDLE_CACHING_LE)) {
+		ksmbd_debug(SMB, "lease state does not contain SMB2_LEASE_HANDLE_CACHING\n");
+		ret = -EBADF;
+		goto out;
+	}
+
+	if (opinfo->o_lease->version != lctx->version) {
+		ksmbd_debug(SMB,
+			    "lease version of fp does not match the one in create context\n");
+		ret = -EBADF;
+		goto out;
+	}
+
+	if (!ksmbd_inode_pending_delete(fp))
+		ret = ksmbd_validate_name_reconnect(share, fp, name);
+out:
+	opinfo_put(opinfo);
+	return ret;
+}
diff --git a/fs/smb/server/oplock.h b/fs/smb/server/oplock.h
index 5b93ea9196..e9da63f25b 100644
--- a/fs/smb/server/oplock.h
+++ b/fs/smb/server/oplock.h
@@ -111,7 +111,7 @@ void opinfo_put(struct oplock_info *opinfo);
 
 /* Lease related functions */
 void create_lease_buf(u8 *rbuf, struct lease *lease);
-struct lease_ctx_info *parse_lease_state(void *open_req, bool is_dir);
+struct lease_ctx_info *parse_lease_state(void *open_req);
 __u8 smb2_map_lease_to_oplock(__le32 lease_state);
 int lease_read_to_write(struct oplock_info *opinfo);
 
@@ -130,4 +130,9 @@ void destroy_lease_table(struct ksmbd_conn *conn);
 void smb_send_parent_lease_break_noti(struct ksmbd_file *fp,
 				      struct lease_ctx_info *lctx);
 void smb_lazy_parent_lease_break_close(struct ksmbd_file *fp);
+int smb2_check_durable_oplock(struct ksmbd_conn *conn,
+			      struct ksmbd_share_config *share,
+			      struct ksmbd_file *fp,
+			      struct lease_ctx_info *lctx,
+			      char *name);
 #endif /* __KSMBD_OPLOCK_H */
diff --git a/fs/smb/server/server.c b/fs/smb/server/server.c
index 2bbc3c3316..c67fbc8d66 100644
--- a/fs/smb/server/server.c
+++ b/fs/smb/server/server.c
@@ -622,7 +622,6 @@ static void __exit ksmbd_server_exit(void)
 }
 
 MODULE_AUTHOR("Namjae Jeon <linkinjeon@kernel.org>");
-MODULE_VERSION(KSMBD_VERSION);
 MODULE_DESCRIPTION("Linux kernel CIFS/SMB SERVER");
 MODULE_LICENSE("GPL");
 MODULE_SOFTDEP("pre: ecb");
diff --git a/fs/smb/server/smb2ops.c b/fs/smb/server/smb2ops.c
index 8600f32c98..606aa3c518 100644
--- a/fs/smb/server/smb2ops.c
+++ b/fs/smb/server/smb2ops.c
@@ -261,6 +261,9 @@ void init_smb3_02_server(struct ksmbd_conn *conn)
 
 	if (server_conf.flags & KSMBD_GLOBAL_FLAG_SMB3_MULTICHANNEL)
 		conn->vals->capabilities |= SMB2_GLOBAL_CAP_MULTI_CHANNEL;
+
+	if (server_conf.flags & KSMBD_GLOBAL_FLAG_DURABLE_HANDLE)
+		conn->vals->capabilities |= SMB2_GLOBAL_CAP_PERSISTENT_HANDLES;
 }
 
 /**
@@ -283,6 +286,9 @@ int init_smb3_11_server(struct ksmbd_conn *conn)
 	if (server_conf.flags & KSMBD_GLOBAL_FLAG_SMB3_MULTICHANNEL)
 		conn->vals->capabilities |= SMB2_GLOBAL_CAP_MULTI_CHANNEL;
 
+	if (server_conf.flags & KSMBD_GLOBAL_FLAG_DURABLE_HANDLE)
+		conn->vals->capabilities |= SMB2_GLOBAL_CAP_PERSISTENT_HANDLES;
+
 	INIT_LIST_HEAD(&conn->preauth_sess_table);
 	return 0;
 }
diff --git a/fs/smb/server/smb2pdu.c b/fs/smb/server/smb2pdu.c
index 8ae0c4d5ab..840c71c66b 100644
--- a/fs/smb/server/smb2pdu.c
+++ b/fs/smb/server/smb2pdu.c
@@ -611,30 +611,6 @@ int smb2_check_user_session(struct ksmbd_work *work)
 	return -ENOENT;
 }
 
-static void destroy_previous_session(struct ksmbd_conn *conn,
-				     struct ksmbd_user *user, u64 id)
-{
-	struct ksmbd_session *prev_sess = ksmbd_session_lookup_slowpath(id);
-	struct ksmbd_user *prev_user;
-	struct channel *chann;
-	long index;
-
-	if (!prev_sess)
-		return;
-
-	prev_user = prev_sess->user;
-
-	if (!prev_user ||
-	    strcmp(user->name, prev_user->name) ||
-	    user->passkey_sz != prev_user->passkey_sz ||
-	    memcmp(user->passkey, prev_user->passkey, user->passkey_sz))
-		return;
-
-	prev_sess->state = SMB2_SESSION_EXPIRED;
-	xa_for_each(&prev_sess->ksmbd_chann_list, index, chann)
-		ksmbd_conn_set_exiting(chann->conn);
-}
-
 /**
  * smb2_get_name() - get filename string from on the wire smb format
  * @src:	source buffer
@@ -654,6 +630,12 @@ smb2_get_name(const char *src, const int maxlen, struct nls_table *local_nls)
 		return name;
 	}
 
+	if (*name == '\\') {
+		pr_err("not allow directory name included leading slash\n");
+		kfree(name);
+		return ERR_PTR(-EINVAL);
+	}
+
 	ksmbd_conv_path_to_unix(name);
 	ksmbd_strip_last_slash(name);
 	return name;
@@ -1950,7 +1932,7 @@ int smb2_tree_connect(struct ksmbd_work *work)
 	struct ksmbd_session *sess = work->sess;
 	char *treename = NULL, *name = NULL;
 	struct ksmbd_tree_conn_status status;
-	struct ksmbd_share_config *share;
+	struct ksmbd_share_config *share = NULL;
 	int rc = -EINVAL;
 
 	WORK_BUFFERS(work, req, rsp);
@@ -2012,7 +1994,12 @@ int smb2_tree_connect(struct ksmbd_work *work)
 	write_unlock(&sess->tree_conns_lock);
 	rsp->StructureSize = cpu_to_le16(16);
 out_err1:
-	rsp->Capabilities = 0;
+	if (server_conf.flags & KSMBD_GLOBAL_FLAG_DURABLE_HANDLE && share &&
+	    test_share_config_flag(share,
+				   KSMBD_SHARE_FLAG_CONTINUOUS_AVAILABILITY))
+		rsp->Capabilities = SMB2_SHARE_CAP_CONTINUOUS_AVAILABILITY;
+	else
+		rsp->Capabilities = 0;
 	rsp->Reserved = 0;
 	/* default manual caching */
 	rsp->ShareFlags = SMB2_SHAREFLAG_MANUAL_CACHING;
@@ -2064,15 +2051,22 @@ out_err1:
  * @access:		file access flags
  * @disposition:	file disposition flags
  * @may_flags:		set with MAY_ flags
+ * @is_dir:		is creating open flags for directory
  *
  * Return:      file open flags
  */
 static int smb2_create_open_flags(bool file_present, __le32 access,
 				  __le32 disposition,
-				  int *may_flags)
+				  int *may_flags,
+				  bool is_dir)
 {
 	int oflags = O_NONBLOCK | O_LARGEFILE;
 
+	if (is_dir) {
+		access &= ~FILE_WRITE_DESIRE_ACCESS_LE;
+		ksmbd_debug(SMB, "Discard write access to a directory\n");
+	}
+
 	if (access & FILE_READ_DESIRED_ACCESS_LE &&
 	    access & FILE_WRITE_DESIRE_ACCESS_LE) {
 		oflags |= O_RDWR;
@@ -2380,7 +2374,8 @@ static int smb2_set_ea(struct smb2_ea_info *eabuf, unsigned int buf_len,
 			if (rc > 0) {
 				rc = ksmbd_vfs_remove_xattr(idmap,
 							    path,
-							    attr_name);
+							    attr_name,
+							    get_write);
 
 				if (rc < 0) {
 					ksmbd_debug(SMB,
@@ -2395,7 +2390,7 @@ static int smb2_set_ea(struct smb2_ea_info *eabuf, unsigned int buf_len,
 		} else {
 			rc = ksmbd_vfs_setxattr(idmap, path, attr_name, value,
 						le16_to_cpu(eabuf->EaValueLength),
-						0, true);
+						0, get_write);
 			if (rc < 0) {
 				ksmbd_debug(SMB,
 					    "ksmbd_vfs_setxattr is failed(%d)\n",
@@ -2487,7 +2482,7 @@ static int smb2_remove_smb_xattrs(const struct path *path)
 		    !strncmp(&name[XATTR_USER_PREFIX_LEN], STREAM_PREFIX,
 			     STREAM_PREFIX_LEN)) {
 			err = ksmbd_vfs_remove_xattr(idmap, path,
-						     name);
+						     name, true);
 			if (err)
 				ksmbd_debug(SMB, "remove xattr failed : %s\n",
 					    name);
@@ -2646,6 +2641,165 @@ static void ksmbd_acls_fattr(struct smb_fattr *fattr,
 	}
 }
 
+enum {
+	DURABLE_RECONN_V2 = 1,
+	DURABLE_RECONN,
+	DURABLE_REQ_V2,
+	DURABLE_REQ,
+};
+
+struct durable_info {
+	struct ksmbd_file *fp;
+	unsigned short int type;
+	bool persistent;
+	bool reconnected;
+	unsigned int timeout;
+	char *CreateGuid;
+};
+
+static int parse_durable_handle_context(struct ksmbd_work *work,
+					struct smb2_create_req *req,
+					struct lease_ctx_info *lc,
+					struct durable_info *dh_info)
+{
+	struct ksmbd_conn *conn = work->conn;
+	struct create_context *context;
+	int dh_idx, err = 0;
+	u64 persistent_id = 0;
+	int req_op_level;
+	static const char * const durable_arr[] = {"DH2C", "DHnC", "DH2Q", "DHnQ"};
+
+	req_op_level = req->RequestedOplockLevel;
+	for (dh_idx = DURABLE_RECONN_V2; dh_idx <= ARRAY_SIZE(durable_arr);
+	     dh_idx++) {
+		context = smb2_find_context_vals(req, durable_arr[dh_idx - 1], 4);
+		if (IS_ERR(context)) {
+			err = PTR_ERR(context);
+			goto out;
+		}
+		if (!context)
+			continue;
+
+		switch (dh_idx) {
+		case DURABLE_RECONN_V2:
+		{
+			struct create_durable_reconn_v2_req *recon_v2;
+
+			if (dh_info->type == DURABLE_RECONN ||
+			    dh_info->type == DURABLE_REQ_V2) {
+				err = -EINVAL;
+				goto out;
+			}
+
+			recon_v2 = (struct create_durable_reconn_v2_req *)context;
+			persistent_id = recon_v2->Fid.PersistentFileId;
+			dh_info->fp = ksmbd_lookup_durable_fd(persistent_id);
+			if (!dh_info->fp) {
+				ksmbd_debug(SMB, "Failed to get durable handle state\n");
+				err = -EBADF;
+				goto out;
+			}
+
+			if (memcmp(dh_info->fp->create_guid, recon_v2->CreateGuid,
+				   SMB2_CREATE_GUID_SIZE)) {
+				err = -EBADF;
+				ksmbd_put_durable_fd(dh_info->fp);
+				goto out;
+			}
+
+			dh_info->type = dh_idx;
+			dh_info->reconnected = true;
+			ksmbd_debug(SMB,
+				"reconnect v2 Persistent-id from reconnect = %llu\n",
+					persistent_id);
+			break;
+		}
+		case DURABLE_RECONN:
+		{
+			struct create_durable_reconn_req *recon;
+
+			if (dh_info->type == DURABLE_RECONN_V2 ||
+			    dh_info->type == DURABLE_REQ_V2) {
+				err = -EINVAL;
+				goto out;
+			}
+
+			recon = (struct create_durable_reconn_req *)context;
+			persistent_id = recon->Data.Fid.PersistentFileId;
+			dh_info->fp = ksmbd_lookup_durable_fd(persistent_id);
+			if (!dh_info->fp) {
+				ksmbd_debug(SMB, "Failed to get durable handle state\n");
+				err = -EBADF;
+				goto out;
+			}
+
+			dh_info->type = dh_idx;
+			dh_info->reconnected = true;
+			ksmbd_debug(SMB, "reconnect Persistent-id from reconnect = %llu\n",
+				    persistent_id);
+			break;
+		}
+		case DURABLE_REQ_V2:
+		{
+			struct create_durable_req_v2 *durable_v2_blob;
+
+			if (dh_info->type == DURABLE_RECONN ||
+			    dh_info->type == DURABLE_RECONN_V2) {
+				err = -EINVAL;
+				goto out;
+			}
+
+			durable_v2_blob =
+				(struct create_durable_req_v2 *)context;
+			ksmbd_debug(SMB, "Request for durable v2 open\n");
+			dh_info->fp = ksmbd_lookup_fd_cguid(durable_v2_blob->CreateGuid);
+			if (dh_info->fp) {
+				if (!memcmp(conn->ClientGUID, dh_info->fp->client_guid,
+					    SMB2_CLIENT_GUID_SIZE)) {
+					if (!(req->hdr.Flags & SMB2_FLAGS_REPLAY_OPERATION)) {
+						err = -ENOEXEC;
+						goto out;
+					}
+
+					dh_info->fp->conn = conn;
+					dh_info->reconnected = true;
+					goto out;
+				}
+			}
+
+			if (((lc && (lc->req_state & SMB2_LEASE_HANDLE_CACHING_LE)) ||
+			     req_op_level == SMB2_OPLOCK_LEVEL_BATCH)) {
+				dh_info->CreateGuid =
+					durable_v2_blob->CreateGuid;
+				dh_info->persistent =
+					le32_to_cpu(durable_v2_blob->Flags);
+				dh_info->timeout =
+					le32_to_cpu(durable_v2_blob->Timeout);
+				dh_info->type = dh_idx;
+			}
+			break;
+		}
+		case DURABLE_REQ:
+			if (dh_info->type == DURABLE_RECONN)
+				goto out;
+			if (dh_info->type == DURABLE_RECONN_V2 ||
+			    dh_info->type == DURABLE_REQ_V2) {
+				err = -EINVAL;
+				goto out;
+			}
+
+			if (((lc && (lc->req_state & SMB2_LEASE_HANDLE_CACHING_LE)) ||
+			     req_op_level == SMB2_OPLOCK_LEVEL_BATCH)) {
+				ksmbd_debug(SMB, "Request for durable open\n");
+				dh_info->type = dh_idx;
+			}
+		}
+	}
+
+out:
+	return err;
+}
+
 /**
  * smb2_open() - handler for smb file open request
  * @work:	smb work containing request buffer
@@ -2669,6 +2823,7 @@ int smb2_open(struct ksmbd_work *work)
 	struct lease_ctx_info *lc = NULL;
 	struct create_ea_buf_req *ea_buf = NULL;
 	struct oplock_info *opinfo;
+	struct durable_info dh_info = {0};
 	__le32 *next_ptr = NULL;
 	int req_op_level = 0, open_flags = 0, may_flags = 0, file_info = 0;
 	int rc = 0;
@@ -2701,20 +2856,11 @@ int smb2_open(struct ksmbd_work *work)
 	}
 
 	if (req->NameLength) {
-		if ((req->CreateOptions & FILE_DIRECTORY_FILE_LE) &&
-		    *(char *)req->Buffer == '\\') {
-			pr_err("not allow directory name included leading slash\n");
-			rc = -EINVAL;
-			goto err_out2;
-		}
-
 		name = smb2_get_name((char *)req + le16_to_cpu(req->NameOffset),
 				     le16_to_cpu(req->NameLength),
 				     work->conn->local_nls);
 		if (IS_ERR(name)) {
 			rc = PTR_ERR(name);
-			if (rc != -ENOMEM)
-				rc = -ENOENT;
 			name = NULL;
 			goto err_out2;
 		}
@@ -2749,6 +2895,49 @@ int smb2_open(struct ksmbd_work *work)
 		}
 	}
 
+	req_op_level = req->RequestedOplockLevel;
+
+	if (server_conf.flags & KSMBD_GLOBAL_FLAG_DURABLE_HANDLE &&
+	    req->CreateContextsOffset) {
+		lc = parse_lease_state(req);
+		rc = parse_durable_handle_context(work, req, lc, &dh_info);
+		if (rc) {
+			ksmbd_debug(SMB, "error parsing durable handle context\n");
+			goto err_out2;
+		}
+
+		if (dh_info.reconnected == true) {
+			rc = smb2_check_durable_oplock(conn, share, dh_info.fp, lc, name);
+			if (rc) {
+				ksmbd_put_durable_fd(dh_info.fp);
+				goto err_out2;
+			}
+
+			rc = ksmbd_reopen_durable_fd(work, dh_info.fp);
+			if (rc) {
+				ksmbd_put_durable_fd(dh_info.fp);
+				goto err_out2;
+			}
+
+			if (ksmbd_override_fsids(work)) {
+				rc = -ENOMEM;
+				ksmbd_put_durable_fd(dh_info.fp);
+				goto err_out2;
+			}
+
+			fp = dh_info.fp;
+			file_info = FILE_OPENED;
+
+			rc = ksmbd_vfs_getattr(&fp->filp->f_path, &stat);
+			if (rc)
+				goto err_out2;
+
+			ksmbd_put_durable_fd(fp);
+			goto reconnected_fp;
+		}
+	} else if (req_op_level == SMB2_OPLOCK_LEVEL_LEASE)
+		lc = parse_lease_state(req);
+
 	if (le32_to_cpu(req->ImpersonationLevel) > le32_to_cpu(IL_DELEGATE)) {
 		pr_err("Invalid impersonationlevel : 0x%x\n",
 		       le32_to_cpu(req->ImpersonationLevel));
@@ -2985,7 +3174,9 @@ int smb2_open(struct ksmbd_work *work)
 
 	open_flags = smb2_create_open_flags(file_present, daccess,
 					    req->CreateDisposition,
-					    &may_flags);
+					    &may_flags,
+		req->CreateOptions & FILE_DIRECTORY_FILE_LE ||
+		(file_present && S_ISDIR(d_inode(path.dentry)->i_mode)));
 
 	if (!test_tree_conn_flag(tcon, KSMBD_TREE_CONN_FLAG_WRITABLE)) {
 		if (open_flags & (O_CREAT | O_TRUNC)) {
@@ -3192,9 +3383,9 @@ int smb2_open(struct ksmbd_work *work)
 	 * after daccess, saccess, attrib_only, and stream are
 	 * initialized.
 	 */
-	write_lock(&fp->f_ci->m_lock);
+	down_write(&fp->f_ci->m_lock);
 	list_add(&fp->node, &fp->f_ci->m_fp_list);
-	write_unlock(&fp->f_ci->m_lock);
+	up_write(&fp->f_ci->m_lock);
 
 	/* Check delete pending among previous fp before oplock break */
 	if (ksmbd_inode_pending_delete(fp)) {
@@ -3211,10 +3402,6 @@ int smb2_open(struct ksmbd_work *work)
 		need_truncate = 1;
 	}
 
-	req_op_level = req->RequestedOplockLevel;
-	if (req_op_level == SMB2_OPLOCK_LEVEL_LEASE)
-		lc = parse_lease_state(req, S_ISDIR(file_inode(filp)->i_mode));
-
 	share_ret = ksmbd_smb_check_shared_mode(fp->filp, fp);
 	if (!test_share_config_flag(work->tcon->share_conf, KSMBD_SHARE_FLAG_OPLOCKS) ||
 	    (req_op_level == SMB2_OPLOCK_LEVEL_LEASE &&
@@ -3225,6 +3412,11 @@ int smb2_open(struct ksmbd_work *work)
 		}
 	} else {
 		if (req_op_level == SMB2_OPLOCK_LEVEL_LEASE) {
+			if (S_ISDIR(file_inode(filp)->i_mode)) {
+				lc->req_state &= ~SMB2_LEASE_WRITE_CACHING_LE;
+				lc->is_dir = true;
+			}
+
 			/*
 			 * Compare parent lease using parent key. If there is no
 			 * a lease that has same parent key, Send lease break
@@ -3321,6 +3513,26 @@ int smb2_open(struct ksmbd_work *work)
 
 	memcpy(fp->client_guid, conn->ClientGUID, SMB2_CLIENT_GUID_SIZE);
 
+	if (dh_info.type == DURABLE_REQ_V2 || dh_info.type == DURABLE_REQ) {
+		if (dh_info.type == DURABLE_REQ_V2 && dh_info.persistent &&
+		    test_share_config_flag(work->tcon->share_conf,
+					   KSMBD_SHARE_FLAG_CONTINUOUS_AVAILABILITY))
+			fp->is_persistent = true;
+		else
+			fp->is_durable = true;
+
+		if (dh_info.type == DURABLE_REQ_V2) {
+			memcpy(fp->create_guid, dh_info.CreateGuid,
+					SMB2_CREATE_GUID_SIZE);
+			if (dh_info.timeout)
+				fp->durable_timeout = min(dh_info.timeout,
+						300000);
+			else
+				fp->durable_timeout = 60;
+		}
+	}
+
+reconnected_fp:
 	rsp->StructureSize = cpu_to_le16(89);
 	rcu_read_lock();
 	opinfo = rcu_dereference(fp->f_opinfo);
@@ -3407,6 +3619,33 @@ int smb2_open(struct ksmbd_work *work)
 		next_off = conn->vals->create_disk_id_size;
 	}
 
+	if (dh_info.type == DURABLE_REQ || dh_info.type == DURABLE_REQ_V2) {
+		struct create_context *durable_ccontext;
+
+		durable_ccontext = (struct create_context *)(rsp->Buffer +
+				le32_to_cpu(rsp->CreateContextsLength));
+		contxt_cnt++;
+		if (dh_info.type == DURABLE_REQ) {
+			create_durable_rsp_buf(rsp->Buffer +
+					le32_to_cpu(rsp->CreateContextsLength));
+			le32_add_cpu(&rsp->CreateContextsLength,
+					conn->vals->create_durable_size);
+			iov_len += conn->vals->create_durable_size;
+		} else {
+			create_durable_v2_rsp_buf(rsp->Buffer +
+					le32_to_cpu(rsp->CreateContextsLength),
+					fp);
+			le32_add_cpu(&rsp->CreateContextsLength,
+					conn->vals->create_durable_v2_size);
+			iov_len += conn->vals->create_durable_v2_size;
+		}
+
+		if (next_ptr)
+			*next_ptr = cpu_to_le32(next_off);
+		next_ptr = &durable_ccontext->Next;
+		next_off = conn->vals->create_durable_size;
+	}
+
 	if (posix_ctxt) {
 		contxt_cnt++;
 		create_posix_rsp_buf(rsp->Buffer +
@@ -5084,8 +5323,13 @@ static int smb2_get_info_filesystem(struct ksmbd_work *work,
 
 		info = (struct filesystem_device_info *)rsp->Buffer;
 
-		info->DeviceType = cpu_to_le32(stfs.f_type);
-		info->DeviceCharacteristics = cpu_to_le32(0x00000020);
+		info->DeviceType = cpu_to_le32(FILE_DEVICE_DISK);
+		info->DeviceCharacteristics =
+			cpu_to_le32(FILE_DEVICE_IS_MOUNTED);
+		if (!test_tree_conn_flag(work->tcon,
+					 KSMBD_TREE_CONN_FLAG_WRITABLE))
+			info->DeviceCharacteristics |=
+				cpu_to_le32(FILE_READ_ONLY_DEVICE);
 		rsp->OutputBufferLength = cpu_to_le32(8);
 		break;
 	}
@@ -6827,10 +7071,10 @@ struct file_lock *smb_flock_init(struct file *f)
 
 	locks_init_lock(fl);
 
-	fl->fl_owner = f;
-	fl->fl_pid = current->tgid;
-	fl->fl_file = f;
-	fl->fl_flags = FL_POSIX;
+	fl->c.flc_owner = f;
+	fl->c.flc_pid = current->tgid;
+	fl->c.flc_file = f;
+	fl->c.flc_flags = FL_POSIX;
 	fl->fl_ops = NULL;
 	fl->fl_lmops = NULL;
 
@@ -6847,30 +7091,30 @@ static int smb2_set_flock_flags(struct file_lock *flock, int flags)
 	case SMB2_LOCKFLAG_SHARED:
 		ksmbd_debug(SMB, "received shared request\n");
 		cmd = F_SETLKW;
-		flock->fl_type = F_RDLCK;
-		flock->fl_flags |= FL_SLEEP;
+		flock->c.flc_type = F_RDLCK;
+		flock->c.flc_flags |= FL_SLEEP;
 		break;
 	case SMB2_LOCKFLAG_EXCLUSIVE:
 		ksmbd_debug(SMB, "received exclusive request\n");
 		cmd = F_SETLKW;
-		flock->fl_type = F_WRLCK;
-		flock->fl_flags |= FL_SLEEP;
+		flock->c.flc_type = F_WRLCK;
+		flock->c.flc_flags |= FL_SLEEP;
 		break;
 	case SMB2_LOCKFLAG_SHARED | SMB2_LOCKFLAG_FAIL_IMMEDIATELY:
 		ksmbd_debug(SMB,
 			    "received shared & fail immediately request\n");
 		cmd = F_SETLK;
-		flock->fl_type = F_RDLCK;
+		flock->c.flc_type = F_RDLCK;
 		break;
 	case SMB2_LOCKFLAG_EXCLUSIVE | SMB2_LOCKFLAG_FAIL_IMMEDIATELY:
 		ksmbd_debug(SMB,
 			    "received exclusive & fail immediately request\n");
 		cmd = F_SETLK;
-		flock->fl_type = F_WRLCK;
+		flock->c.flc_type = F_WRLCK;
 		break;
 	case SMB2_LOCKFLAG_UNLOCK:
 		ksmbd_debug(SMB, "received unlock request\n");
-		flock->fl_type = F_UNLCK;
+		flock->c.flc_type = F_UNLCK;
 		cmd = F_SETLK;
 		break;
 	}
@@ -6908,13 +7152,13 @@ static void smb2_remove_blocked_lock(void **argv)
 	struct file_lock *flock = (struct file_lock *)argv[0];
 
 	ksmbd_vfs_posix_lock_unblock(flock);
-	wake_up(&flock->fl_wait);
+	locks_wake_up(flock);
 }
 
 static inline bool lock_defer_pending(struct file_lock *fl)
 {
 	/* check pending lock waiters */
-	return waitqueue_active(&fl->fl_wait);
+	return waitqueue_active(&fl->c.flc_wait);
 }
 
 /**
@@ -7005,8 +7249,8 @@ int smb2_lock(struct ksmbd_work *work)
 		list_for_each_entry(cmp_lock, &lock_list, llist) {
 			if (cmp_lock->fl->fl_start <= flock->fl_start &&
 			    cmp_lock->fl->fl_end >= flock->fl_end) {
-				if (cmp_lock->fl->fl_type != F_UNLCK &&
-				    flock->fl_type != F_UNLCK) {
+				if (cmp_lock->fl->c.flc_type != F_UNLCK &&
+				    flock->c.flc_type != F_UNLCK) {
 					pr_err("conflict two locks in one request\n");
 					err = -EINVAL;
 					locks_free_lock(flock);
@@ -7054,12 +7298,12 @@ int smb2_lock(struct ksmbd_work *work)
 		list_for_each_entry(conn, &conn_list, conns_list) {
 			spin_lock(&conn->llist_lock);
 			list_for_each_entry_safe(cmp_lock, tmp2, &conn->lock_list, clist) {
-				if (file_inode(cmp_lock->fl->fl_file) !=
-				    file_inode(smb_lock->fl->fl_file))
+				if (file_inode(cmp_lock->fl->c.flc_file) !=
+				    file_inode(smb_lock->fl->c.flc_file))
 					continue;
 
-				if (smb_lock->fl->fl_type == F_UNLCK) {
-					if (cmp_lock->fl->fl_file == smb_lock->fl->fl_file &&
+				if (lock_is_unlock(smb_lock->fl)) {
+					if (cmp_lock->fl->c.flc_file == smb_lock->fl->c.flc_file &&
 					    cmp_lock->start == smb_lock->start &&
 					    cmp_lock->end == smb_lock->end &&
 					    !lock_defer_pending(cmp_lock->fl)) {
@@ -7076,7 +7320,7 @@ int smb2_lock(struct ksmbd_work *work)
 					continue;
 				}
 
-				if (cmp_lock->fl->fl_file == smb_lock->fl->fl_file) {
+				if (cmp_lock->fl->c.flc_file == smb_lock->fl->c.flc_file) {
 					if (smb_lock->flags & SMB2_LOCKFLAG_SHARED)
 						continue;
 				} else {
@@ -7118,7 +7362,7 @@ int smb2_lock(struct ksmbd_work *work)
 		}
 		up_read(&conn_list_lock);
 out_check_cl:
-		if (smb_lock->fl->fl_type == F_UNLCK && nolock) {
+		if (lock_is_unlock(smb_lock->fl) && nolock) {
 			pr_err("Try to unlock nolocked range\n");
 			rsp->hdr.Status = STATUS_RANGE_NOT_LOCKED;
 			goto out;
@@ -7242,7 +7486,7 @@ out:
 		struct file_lock *rlock = NULL;
 
 		rlock = smb_flock_init(filp);
-		rlock->fl_type = F_UNLCK;
+		rlock->c.flc_type = F_UNLCK;
 		rlock->fl_start = smb_lock->start;
 		rlock->fl_end = smb_lock->end;
 
diff --git a/fs/smb/server/smb2pdu.h b/fs/smb/server/smb2pdu.h
index d12cfd3b09..643f5e1cfe 100644
--- a/fs/smb/server/smb2pdu.h
+++ b/fs/smb/server/smb2pdu.h
@@ -64,7 +64,7 @@ struct preauth_integrity_info {
 #define SMB2_SESSION_TIMEOUT		(10 * HZ)
 
 struct create_durable_req_v2 {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8   Name[8];
 	__le32 Timeout;
 	__le32 Flags;
@@ -72,8 +72,20 @@ struct create_durable_req_v2 {
 	__u8 CreateGuid[16];
 } __packed;
 
+struct create_durable_reconn_req {
+	struct create_context_hdr ccontext;
+	__u8   Name[8];
+	union {
+		__u8  Reserved[16];
+		struct {
+			__u64 PersistentFileId;
+			__u64 VolatileFileId;
+		} Fid;
+	} Data;
+} __packed;
+
 struct create_durable_reconn_v2_req {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8   Name[8];
 	struct {
 		__u64 PersistentFileId;
@@ -84,13 +96,13 @@ struct create_durable_reconn_v2_req {
 } __packed;
 
 struct create_alloc_size_req {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8   Name[8];
 	__le64 AllocationSize;
 } __packed;
 
 struct create_durable_rsp {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8   Name[8];
 	union {
 		__u8  Reserved[8];
@@ -98,8 +110,11 @@ struct create_durable_rsp {
 	} Data;
 } __packed;
 
+/* See MS-SMB2 2.2.13.2.11 */
+/* Flags */
+#define SMB2_DHANDLE_FLAG_PERSISTENT	0x00000002
 struct create_durable_v2_rsp {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8   Name[8];
 	__le32 Timeout;
 	__le32 Flags;
@@ -107,7 +122,7 @@ struct create_durable_v2_rsp {
 
 /* equivalent of the contents of SMB3.1.1 POSIX open context response */
 struct create_posix_rsp {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8    Name[16];
 	__le32 nlink;
 	__le32 reparse_tag;
@@ -366,13 +381,13 @@ struct smb2_ea_info {
 } __packed; /* level 15 Query */
 
 struct create_ea_buf_req {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8   Name[8];
 	struct smb2_ea_info ea;
 } __packed;
 
 struct create_sd_buf_req {
-	struct create_context ccontext;
+	struct create_context_hdr ccontext;
 	__u8   Name[8];
 	struct smb_ntsd ntsd;
 } __packed;
diff --git a/fs/smb/server/smb_common.c b/fs/smb/server/smb_common.c
index fcaf373cc0..474dadf6b7 100644
--- a/fs/smb/server/smb_common.c
+++ b/fs/smb/server/smb_common.c
@@ -646,7 +646,7 @@ int ksmbd_smb_check_shared_mode(struct file *filp, struct ksmbd_file *curr_fp)
 	 * Lookup fp in master fp list, and check desired access and
 	 * shared mode between previous open and current open.
 	 */
-	read_lock(&curr_fp->f_ci->m_lock);
+	down_read(&curr_fp->f_ci->m_lock);
 	list_for_each_entry(prev_fp, &curr_fp->f_ci->m_fp_list, node) {
 		if (file_inode(filp) != file_inode(prev_fp->filp))
 			continue;
@@ -722,7 +722,7 @@ int ksmbd_smb_check_shared_mode(struct file *filp, struct ksmbd_file *curr_fp)
 			break;
 		}
 	}
-	read_unlock(&curr_fp->f_ci->m_lock);
+	up_read(&curr_fp->f_ci->m_lock);
 
 	return rc;
 }
diff --git a/fs/smb/server/vfs.c b/fs/smb/server/vfs.c
index d7906efaa9..9e859ba010 100644
--- a/fs/smb/server/vfs.c
+++ b/fs/smb/server/vfs.c
@@ -337,18 +337,18 @@ static int check_lock_range(struct file *filp, loff_t start, loff_t end,
 		return 0;
 
 	spin_lock(&ctx->flc_lock);
-	list_for_each_entry(flock, &ctx->flc_posix, fl_list) {
+	for_each_file_lock(flock, &ctx->flc_posix) {
 		/* check conflict locks */
 		if (flock->fl_end >= start && end >= flock->fl_start) {
-			if (flock->fl_type == F_RDLCK) {
+			if (lock_is_read(flock)) {
 				if (type == WRITE) {
 					pr_err("not allow write by shared lock\n");
 					error = 1;
 					goto out;
 				}
-			} else if (flock->fl_type == F_WRLCK) {
+			} else if (lock_is_write(flock)) {
 				/* check owner in lock */
-				if (flock->fl_file != filp) {
+				if (flock->c.flc_file != filp) {
 					error = 1;
 					pr_err("not allow rw access by exclusive lock from other opens\n");
 					goto out;
@@ -1058,16 +1058,21 @@ int ksmbd_vfs_fqar_lseek(struct ksmbd_file *fp, loff_t start, loff_t length,
 }
 
 int ksmbd_vfs_remove_xattr(struct mnt_idmap *idmap,
-			   const struct path *path, char *attr_name)
+			   const struct path *path, char *attr_name,
+			   bool get_write)
 {
 	int err;
 
-	err = mnt_want_write(path->mnt);
-	if (err)
-		return err;
+	if (get_write == true) {
+		err = mnt_want_write(path->mnt);
+		if (err)
+			return err;
+	}
 
 	err = vfs_removexattr(idmap, path->dentry, attr_name);
-	mnt_drop_write(path->mnt);
+
+	if (get_write == true)
+		mnt_drop_write(path->mnt);
 
 	return err;
 }
@@ -1380,7 +1385,7 @@ int ksmbd_vfs_remove_sd_xattrs(struct mnt_idmap *idmap, const struct path *path)
 		ksmbd_debug(SMB, "%s, len %zd\n", name, strlen(name));
 
 		if (!strncmp(name, XATTR_NAME_SD, XATTR_NAME_SD_LEN)) {
-			err = ksmbd_vfs_remove_xattr(idmap, path, name);
+			err = ksmbd_vfs_remove_xattr(idmap, path, name, true);
 			if (err)
 				ksmbd_debug(SMB, "remove xattr failed : %s\n", name);
 		}
@@ -1850,13 +1855,13 @@ int ksmbd_vfs_copy_file_ranges(struct ksmbd_work *work,
 
 void ksmbd_vfs_posix_lock_wait(struct file_lock *flock)
 {
-	wait_event(flock->fl_wait, !flock->fl_blocker);
+	wait_event(flock->c.flc_wait, !flock->c.flc_blocker);
 }
 
 int ksmbd_vfs_posix_lock_wait_timeout(struct file_lock *flock, long timeout)
 {
-	return wait_event_interruptible_timeout(flock->fl_wait,
-						!flock->fl_blocker,
+	return wait_event_interruptible_timeout(flock->c.flc_wait,
+						!flock->c.flc_blocker,
 						timeout);
 }
 
diff --git a/fs/smb/server/vfs.h b/fs/smb/server/vfs.h
index cfe1c8092f..cb76f4b5ba 100644
--- a/fs/smb/server/vfs.h
+++ b/fs/smb/server/vfs.h
@@ -114,7 +114,8 @@ int ksmbd_vfs_setxattr(struct mnt_idmap *idmap,
 int ksmbd_vfs_xattr_stream_name(char *stream_name, char **xattr_stream_name,
 				size_t *xattr_stream_name_size, int s_type);
 int ksmbd_vfs_remove_xattr(struct mnt_idmap *idmap,
-			   const struct path *path, char *attr_name);
+			   const struct path *path, char *attr_name,
+			   bool get_write);
 int ksmbd_vfs_kern_path_locked(struct ksmbd_work *work, char *name,
 			       unsigned int flags, struct path *parent_path,
 			       struct path *path, bool caseless);
diff --git a/fs/smb/server/vfs_cache.c b/fs/smb/server/vfs_cache.c
index 4e82ff627d..8b2e37c871 100644
--- a/fs/smb/server/vfs_cache.c
+++ b/fs/smb/server/vfs_cache.c
@@ -165,7 +165,7 @@ static int ksmbd_inode_init(struct ksmbd_inode *ci, struct ksmbd_file *fp)
 	ci->m_fattr = 0;
 	INIT_LIST_HEAD(&ci->m_fp_list);
 	INIT_LIST_HEAD(&ci->m_op_list);
-	rwlock_init(&ci->m_lock);
+	init_rwsem(&ci->m_lock);
 	ci->m_de = fp->filp->f_path.dentry;
 	return 0;
 }
@@ -254,21 +254,22 @@ static void __ksmbd_inode_close(struct ksmbd_file *fp)
 		ci->m_flags &= ~S_DEL_ON_CLS_STREAM;
 		err = ksmbd_vfs_remove_xattr(file_mnt_idmap(filp),
 					     &filp->f_path,
-					     fp->stream.name);
+					     fp->stream.name,
+					     true);
 		if (err)
 			pr_err("remove xattr failed : %s\n",
 			       fp->stream.name);
 	}
 
 	if (atomic_dec_and_test(&ci->m_count)) {
-		write_lock(&ci->m_lock);
+		down_write(&ci->m_lock);
 		if (ci->m_flags & (S_DEL_ON_CLS | S_DEL_PENDING)) {
 			ci->m_flags &= ~(S_DEL_ON_CLS | S_DEL_PENDING);
-			write_unlock(&ci->m_lock);
+			up_write(&ci->m_lock);
 			ksmbd_vfs_unlink(filp);
-			write_lock(&ci->m_lock);
+			down_write(&ci->m_lock);
 		}
-		write_unlock(&ci->m_lock);
+		up_write(&ci->m_lock);
 
 		ksmbd_inode_free(ci);
 	}
@@ -289,9 +290,9 @@ static void __ksmbd_remove_fd(struct ksmbd_file_table *ft, struct ksmbd_file *fp
 	if (!has_file_id(fp->volatile_id))
 		return;
 
-	write_lock(&fp->f_ci->m_lock);
+	down_write(&fp->f_ci->m_lock);
 	list_del_init(&fp->node);
-	write_unlock(&fp->f_ci->m_lock);
+	up_write(&fp->f_ci->m_lock);
 
 	write_lock(&ft->lock);
 	idr_remove(ft->idr, fp->volatile_id);
@@ -305,7 +306,8 @@ static void __ksmbd_close_fd(struct ksmbd_file_table *ft, struct ksmbd_file *fp)
 
 	fd_limit_close();
 	__ksmbd_remove_durable_fd(fp);
-	__ksmbd_remove_fd(ft, fp);
+	if (ft)
+		__ksmbd_remove_fd(ft, fp);
 
 	close_id_del_oplock(fp);
 	filp = fp->filp;
@@ -465,11 +467,32 @@ struct ksmbd_file *ksmbd_lookup_fd_slow(struct ksmbd_work *work, u64 id,
 	return fp;
 }
 
-struct ksmbd_file *ksmbd_lookup_durable_fd(unsigned long long id)
+struct ksmbd_file *ksmbd_lookup_global_fd(unsigned long long id)
 {
 	return __ksmbd_lookup_fd(&global_ft, id);
 }
 
+struct ksmbd_file *ksmbd_lookup_durable_fd(unsigned long long id)
+{
+	struct ksmbd_file *fp;
+
+	fp = __ksmbd_lookup_fd(&global_ft, id);
+	if (fp && fp->conn) {
+		ksmbd_put_durable_fd(fp);
+		fp = NULL;
+	}
+
+	return fp;
+}
+
+void ksmbd_put_durable_fd(struct ksmbd_file *fp)
+{
+	if (!atomic_dec_and_test(&fp->refcount))
+		return;
+
+	__ksmbd_close_fd(NULL, fp);
+}
+
 struct ksmbd_file *ksmbd_lookup_fd_cguid(char *cguid)
 {
 	struct ksmbd_file	*fp = NULL;
@@ -501,17 +524,17 @@ struct ksmbd_file *ksmbd_lookup_fd_inode(struct dentry *dentry)
 	if (!ci)
 		return NULL;
 
-	read_lock(&ci->m_lock);
+	down_read(&ci->m_lock);
 	list_for_each_entry(lfp, &ci->m_fp_list, node) {
 		if (inode == file_inode(lfp->filp)) {
 			atomic_dec(&ci->m_count);
 			lfp = ksmbd_fp_get(lfp);
-			read_unlock(&ci->m_lock);
+			up_read(&ci->m_lock);
 			return lfp;
 		}
 	}
 	atomic_dec(&ci->m_count);
-	read_unlock(&ci->m_lock);
+	up_read(&ci->m_lock);
 	return NULL;
 }
 
@@ -639,6 +662,32 @@ __close_file_table_ids(struct ksmbd_file_table *ft,
 	return num;
 }
 
+static inline bool is_reconnectable(struct ksmbd_file *fp)
+{
+	struct oplock_info *opinfo = opinfo_get(fp);
+	bool reconn = false;
+
+	if (!opinfo)
+		return false;
+
+	if (opinfo->op_state != OPLOCK_STATE_NONE) {
+		opinfo_put(opinfo);
+		return false;
+	}
+
+	if (fp->is_resilient || fp->is_persistent)
+		reconn = true;
+	else if (fp->is_durable && opinfo->is_lease &&
+		 opinfo->o_lease->state & SMB2_LEASE_HANDLE_CACHING_LE)
+		reconn = true;
+
+	else if (fp->is_durable && opinfo->level == SMB2_OPLOCK_LEVEL_BATCH)
+		reconn = true;
+
+	opinfo_put(opinfo);
+	return reconn;
+}
+
 static bool tree_conn_fd_check(struct ksmbd_tree_connect *tcon,
 			       struct ksmbd_file *fp)
 {
@@ -648,7 +697,28 @@ static bool tree_conn_fd_check(struct ksmbd_tree_connect *tcon,
 static bool session_fd_check(struct ksmbd_tree_connect *tcon,
 			     struct ksmbd_file *fp)
 {
-	return false;
+	struct ksmbd_inode *ci;
+	struct oplock_info *op;
+	struct ksmbd_conn *conn;
+
+	if (!is_reconnectable(fp))
+		return false;
+
+	conn = fp->conn;
+	ci = fp->f_ci;
+	down_write(&ci->m_lock);
+	list_for_each_entry_rcu(op, &ci->m_op_list, op_entry) {
+		if (op->conn != conn)
+			continue;
+		op->conn = NULL;
+	}
+	up_write(&ci->m_lock);
+
+	fp->conn = NULL;
+	fp->tcon = NULL;
+	fp->volatile_id = KSMBD_NO_FID;
+
+	return true;
 }
 
 void ksmbd_close_tree_conn_fds(struct ksmbd_work *work)
@@ -687,6 +757,68 @@ void ksmbd_free_global_file_table(void)
 	ksmbd_destroy_file_table(&global_ft);
 }
 
+int ksmbd_validate_name_reconnect(struct ksmbd_share_config *share,
+				  struct ksmbd_file *fp, char *name)
+{
+	char *pathname, *ab_pathname;
+	int ret = 0;
+
+	pathname = kmalloc(PATH_MAX, GFP_KERNEL);
+	if (!pathname)
+		return -EACCES;
+
+	ab_pathname = d_path(&fp->filp->f_path, pathname, PATH_MAX);
+	if (IS_ERR(ab_pathname)) {
+		kfree(pathname);
+		return -EACCES;
+	}
+
+	if (name && strcmp(&ab_pathname[share->path_sz + 1], name)) {
+		ksmbd_debug(SMB, "invalid name reconnect %s\n", name);
+		ret = -EINVAL;
+	}
+
+	kfree(pathname);
+
+	return ret;
+}
+
+int ksmbd_reopen_durable_fd(struct ksmbd_work *work, struct ksmbd_file *fp)
+{
+	struct ksmbd_inode *ci;
+	struct oplock_info *op;
+
+	if (!fp->is_durable || fp->conn || fp->tcon) {
+		pr_err("Invalid durable fd [%p:%p]\n", fp->conn, fp->tcon);
+		return -EBADF;
+	}
+
+	if (has_file_id(fp->volatile_id)) {
+		pr_err("Still in use durable fd: %llu\n", fp->volatile_id);
+		return -EBADF;
+	}
+
+	fp->conn = work->conn;
+	fp->tcon = work->tcon;
+
+	ci = fp->f_ci;
+	down_write(&ci->m_lock);
+	list_for_each_entry_rcu(op, &ci->m_op_list, op_entry) {
+		if (op->conn)
+			continue;
+		op->conn = fp->conn;
+	}
+	up_write(&ci->m_lock);
+
+	__open_id(&work->sess->file_table, fp, OPEN_ID_TYPE_VOLATILE_ID);
+	if (!has_file_id(fp->volatile_id)) {
+		fp->conn = NULL;
+		fp->tcon = NULL;
+		return -EBADF;
+	}
+	return 0;
+}
+
 int ksmbd_init_file_table(struct ksmbd_file_table *ft)
 {
 	ft->idr = kzalloc(sizeof(struct idr), GFP_KERNEL);
diff --git a/fs/smb/server/vfs_cache.h b/fs/smb/server/vfs_cache.h
index a528f0cc77..5a225e7055 100644
--- a/fs/smb/server/vfs_cache.h
+++ b/fs/smb/server/vfs_cache.h
@@ -14,6 +14,7 @@
 #include <linux/workqueue.h>
 
 #include "vfs.h"
+#include "mgmt/share_config.h"
 
 /* Windows style file permissions for extended response */
 #define	FILE_GENERIC_ALL	0x1F01FF
@@ -46,7 +47,7 @@ struct stream {
 };
 
 struct ksmbd_inode {
-	rwlock_t			m_lock;
+	struct rw_semaphore		m_lock;
 	atomic_t			m_count;
 	atomic_t			op_count;
 	/* opinfo count for streams */
@@ -106,6 +107,9 @@ struct ksmbd_file {
 	int				dot_dotdot[2];
 	unsigned int			f_state;
 	bool				reserve_lease_break;
+	bool				is_durable;
+	bool				is_persistent;
+	bool				is_resilient;
 };
 
 static inline void set_ctx_actor(struct dir_context *ctx,
@@ -141,7 +145,9 @@ struct ksmbd_file *ksmbd_lookup_fd_slow(struct ksmbd_work *work, u64 id,
 void ksmbd_fd_put(struct ksmbd_work *work, struct ksmbd_file *fp);
 struct ksmbd_inode *ksmbd_inode_lookup_lock(struct dentry *d);
 void ksmbd_inode_put(struct ksmbd_inode *ci);
+struct ksmbd_file *ksmbd_lookup_global_fd(unsigned long long id);
 struct ksmbd_file *ksmbd_lookup_durable_fd(unsigned long long id);
+void ksmbd_put_durable_fd(struct ksmbd_file *fp);
 struct ksmbd_file *ksmbd_lookup_fd_cguid(char *cguid);
 struct ksmbd_file *ksmbd_lookup_fd_inode(struct dentry *dentry);
 unsigned int ksmbd_open_durable_fd(struct ksmbd_file *fp);
@@ -173,6 +179,9 @@ void ksmbd_set_inode_pending_delete(struct ksmbd_file *fp);
 void ksmbd_clear_inode_pending_delete(struct ksmbd_file *fp);
 void ksmbd_fd_set_delete_on_close(struct ksmbd_file *fp,
 				  int file_info);
+int ksmbd_reopen_durable_fd(struct ksmbd_work *work, struct ksmbd_file *fp);
+int ksmbd_validate_name_reconnect(struct ksmbd_share_config *share,
+				  struct ksmbd_file *fp, char *name);
 int ksmbd_init_file_cache(void);
 void ksmbd_exit_file_cache(void);
 #endif /* __VFS_CACHE_H__ */
diff --git a/fs/splice.c b/fs/splice.c
index 218e24b1ac..60aed8de21 100644
--- a/fs/splice.c
+++ b/fs/splice.c
@@ -362,7 +362,7 @@ ssize_t copy_splice_read(struct file *in, loff_t *ppos,
 	iov_iter_bvec(&to, ITER_DEST, bv, npages, len);
 	init_sync_kiocb(&kiocb, in);
 	kiocb.ki_pos = *ppos;
-	ret = call_read_iter(in, &kiocb, &to);
+	ret = in->f_op->read_iter(&kiocb, &to);
 
 	if (ret > 0) {
 		keep = DIV_ROUND_UP(ret, PAGE_SIZE);
@@ -740,7 +740,7 @@ iter_file_splice_write(struct pipe_inode_info *pipe, struct file *out,
 		iov_iter_bvec(&from, ITER_SOURCE, array, n, sd.total_len - left);
 		init_sync_kiocb(&kiocb, out);
 		kiocb.ki_pos = sd.pos;
-		ret = call_write_iter(out, &kiocb, &from);
+		ret = out->f_op->write_iter(&kiocb, &from);
 		sd.pos = kiocb.ki_pos;
 		if (ret <= 0)
 			break;
diff --git a/fs/squashfs/file.c b/fs/squashfs/file.c
index e8df643044..a8c1e7f9a6 100644
--- a/fs/squashfs/file.c
+++ b/fs/squashfs/file.c
@@ -375,8 +375,6 @@ void squashfs_fill_page(struct page *page, struct squashfs_cache_entry *buffer,
 	flush_dcache_page(page);
 	if (copied == avail)
 		SetPageUptodate(page);
-	else
-		SetPageError(page);
 }
 
 /* Copy data into page cache  */
@@ -471,7 +469,7 @@ static int squashfs_read_folio(struct file *file, struct folio *folio)
 
 		res = read_blocklist(inode, index, &block);
 		if (res < 0)
-			goto error_out;
+			goto out;
 
 		if (res == 0)
 			res = squashfs_readpage_sparse(page, expected);
@@ -483,8 +481,6 @@ static int squashfs_read_folio(struct file *file, struct folio *folio)
 	if (!res)
 		return 0;
 
-error_out:
-	SetPageError(page);
 out:
 	pageaddr = kmap_atomic(page);
 	memset(pageaddr, 0, PAGE_SIZE);
diff --git a/fs/squashfs/file_direct.c b/fs/squashfs/file_direct.c
index 763a3f7a75..2a689ce71d 100644
--- a/fs/squashfs/file_direct.c
+++ b/fs/squashfs/file_direct.c
@@ -106,14 +106,13 @@ int squashfs_readpage_block(struct page *target_page, u64 block, int bsize,
 	return 0;
 
 mark_errored:
-	/* Decompression failed, mark pages as errored.  Target_page is
+	/* Decompression failed.  Target_page is
 	 * dealt with by the caller
 	 */
 	for (i = 0; i < pages; i++) {
 		if (page[i] == NULL || page[i] == target_page)
 			continue;
 		flush_dcache_page(page[i]);
-		SetPageError(page[i]);
 		unlock_page(page[i]);
 		put_page(page[i]);
 	}
diff --git a/fs/squashfs/namei.c b/fs/squashfs/namei.c
index 11e4539b9e..65aae7e2a8 100644
--- a/fs/squashfs/namei.c
+++ b/fs/squashfs/namei.c
@@ -62,27 +62,21 @@
  */
 static int get_dir_index_using_name(struct super_block *sb,
 			u64 *next_block, int *next_offset, u64 index_start,
-			int index_offset, int i_count, const char *name,
-			int len)
+			int index_offset, int i_count, const char *name)
 {
 	struct squashfs_sb_info *msblk = sb->s_fs_info;
 	int i, length = 0, err;
 	unsigned int size;
 	struct squashfs_dir_index *index;
-	char *str;
 
 	TRACE("Entered get_dir_index_using_name, i_count %d\n", i_count);
 
-	index = kmalloc(sizeof(*index) + SQUASHFS_NAME_LEN * 2 + 2, GFP_KERNEL);
+	index = kmalloc(sizeof(*index) + SQUASHFS_NAME_LEN + 1, GFP_KERNEL);
 	if (index == NULL) {
 		ERROR("Failed to allocate squashfs_dir_index\n");
 		goto out;
 	}
 
-	str = &index->name[SQUASHFS_NAME_LEN + 1];
-	strncpy(str, name, len);
-	str[len] = '\0';
-
 	for (i = 0; i < i_count; i++) {
 		err = squashfs_read_metadata(sb, index, &index_start,
 					&index_offset, sizeof(*index));
@@ -101,7 +95,7 @@ static int get_dir_index_using_name(struct super_block *sb,
 
 		index->name[size] = '\0';
 
-		if (strcmp(index->name, str) > 0)
+		if (strcmp(index->name, name) > 0)
 			break;
 
 		length = le32_to_cpu(index->index);
@@ -153,7 +147,7 @@ static struct dentry *squashfs_lookup(struct inode *dir, struct dentry *dentry,
 	length = get_dir_index_using_name(dir->i_sb, &block, &offset,
 				squashfs_i(dir)->dir_idx_start,
 				squashfs_i(dir)->dir_idx_offset,
-				squashfs_i(dir)->dir_idx_cnt, name, len);
+				squashfs_i(dir)->dir_idx_cnt, name);
 
 	while (length < i_size_read(dir)) {
 		/*
diff --git a/fs/squashfs/symlink.c b/fs/squashfs/symlink.c
index 2bf977a52c..6ef735bd84 100644
--- a/fs/squashfs/symlink.c
+++ b/fs/squashfs/symlink.c
@@ -32,20 +32,19 @@
 
 static int squashfs_symlink_read_folio(struct file *file, struct folio *folio)
 {
-	struct page *page = &folio->page;
-	struct inode *inode = page->mapping->host;
+	struct inode *inode = folio->mapping->host;
 	struct super_block *sb = inode->i_sb;
 	struct squashfs_sb_info *msblk = sb->s_fs_info;
-	int index = page->index << PAGE_SHIFT;
+	int index = folio_pos(folio);
 	u64 block = squashfs_i(inode)->start;
 	int offset = squashfs_i(inode)->offset;
 	int length = min_t(int, i_size_read(inode) - index, PAGE_SIZE);
-	int bytes, copied;
+	int bytes, copied, error;
 	void *pageaddr;
 	struct squashfs_cache_entry *entry;
 
 	TRACE("Entered squashfs_symlink_readpage, page index %ld, start block "
-			"%llx, offset %x\n", page->index, block, offset);
+			"%llx, offset %x\n", folio->index, block, offset);
 
 	/*
 	 * Skip index bytes into symlink metadata.
@@ -57,14 +56,15 @@ static int squashfs_symlink_read_folio(struct file *file, struct folio *folio)
 			ERROR("Unable to read symlink [%llx:%x]\n",
 				squashfs_i(inode)->start,
 				squashfs_i(inode)->offset);
-			goto error_out;
+			error = bytes;
+			goto out;
 		}
 	}
 
 	/*
 	 * Read length bytes from symlink metadata.  Squashfs_read_metadata
 	 * is not used here because it can sleep and we want to use
-	 * kmap_atomic to map the page.  Instead call the underlying
+	 * kmap_local to map the folio.  Instead call the underlying
 	 * squashfs_cache_get routine.  As length bytes may overlap metadata
 	 * blocks, we may need to call squashfs_cache_get multiple times.
 	 */
@@ -75,29 +75,26 @@ static int squashfs_symlink_read_folio(struct file *file, struct folio *folio)
 				squashfs_i(inode)->start,
 				squashfs_i(inode)->offset);
 			squashfs_cache_put(entry);
-			goto error_out;
+			error = entry->error;
+			goto out;
 		}
 
-		pageaddr = kmap_atomic(page);
+		pageaddr = kmap_local_folio(folio, 0);
 		copied = squashfs_copy_data(pageaddr + bytes, entry, offset,
 								length - bytes);
 		if (copied == length - bytes)
 			memset(pageaddr + length, 0, PAGE_SIZE - length);
 		else
 			block = entry->next_index;
-		kunmap_atomic(pageaddr);
+		kunmap_local(pageaddr);
 		squashfs_cache_put(entry);
 	}
 
-	flush_dcache_page(page);
-	SetPageUptodate(page);
-	unlock_page(page);
-	return 0;
-
-error_out:
-	SetPageError(page);
-	unlock_page(page);
-	return 0;
+	flush_dcache_folio(folio);
+	error = 0;
+out:
+	folio_end_read(folio, error == 0);
+	return error;
 }
 
 
diff --git a/fs/stat.c b/fs/stat.c
index 77cdc69eb4..70bd3e888c 100644
--- a/fs/stat.c
+++ b/fs/stat.c
@@ -658,6 +658,7 @@ cp_statx(const struct kstat *stat, struct statx __user *buffer)
 	tmp.stx_mnt_id = stat->mnt_id;
 	tmp.stx_dio_mem_align = stat->dio_mem_align;
 	tmp.stx_dio_offset_align = stat->dio_offset_align;
+	tmp.stx_subvol = stat->subvol;
 
 	return copy_to_user(buffer, &tmp, sizeof(tmp)) ? -EFAULT : 0;
 }
diff --git a/fs/super.c b/fs/super.c
index d6efeba0d0..38d72a3cf6 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -274,6 +274,7 @@ static void destroy_super_work(struct work_struct *work)
 {
 	struct super_block *s = container_of(work, struct super_block,
 							destroy_work);
+	fsnotify_sb_free(s);
 	security_sb_free(s);
 	put_user_ns(s->s_user_ns);
 	kfree(s->s_subtype);
@@ -735,6 +736,17 @@ struct super_block *sget_fc(struct fs_context *fc,
 	struct user_namespace *user_ns = fc->global ? &init_user_ns : fc->user_ns;
 	int err;
 
+	/*
+	 * Never allow s_user_ns != &init_user_ns when FS_USERNS_MOUNT is
+	 * not set, as the filesystem is likely unprepared to handle it.
+	 * This can happen when fsconfig() is called from init_user_ns with
+	 * an fs_fd opened in another user namespace.
+	 */
+	if (user_ns != &init_user_ns && !(fc->fs_type->fs_flags & FS_USERNS_MOUNT)) {
+		errorfc(fc, "VFS: Mounting from non-initial user namespace is not allowed");
+		return ERR_PTR(-EPERM);
+	}
+
 retry:
 	spin_lock(&sb_lock);
 	if (test) {
@@ -1501,8 +1513,17 @@ static int fs_bdev_thaw(struct block_device *bdev)
 
 	lockdep_assert_held(&bdev->bd_fsfreeze_mutex);
 
+	/*
+	 * The block device may have been frozen before it was claimed by a
+	 * filesystem. Concurrently another process might try to mount that
+	 * frozen block device and has temporarily claimed the block device for
+	 * that purpose causing a concurrent fs_bdev_thaw() to end up here. The
+	 * mounter is already about to abort mounting because they still saw an
+	 * elevanted bdev->bd_fsfreeze_count so get_bdev_super() will return
+	 * NULL in that case.
+	 */
 	sb = get_bdev_super(bdev);
-	if (WARN_ON_ONCE(!sb))
+	if (!sb)
 		return -EINVAL;
 
 	if (sb->s_op->thaw_super)
@@ -1527,16 +1548,16 @@ int setup_bdev_super(struct super_block *sb, int sb_flags,
 		struct fs_context *fc)
 {
 	blk_mode_t mode = sb_open_mode(sb_flags);
-	struct bdev_handle *bdev_handle;
+	struct file *bdev_file;
 	struct block_device *bdev;
 
-	bdev_handle = bdev_open_by_dev(sb->s_dev, mode, sb, &fs_holder_ops);
-	if (IS_ERR(bdev_handle)) {
+	bdev_file = bdev_file_open_by_dev(sb->s_dev, mode, sb, &fs_holder_ops);
+	if (IS_ERR(bdev_file)) {
 		if (fc)
 			errorf(fc, "%s: Can't open blockdev", fc->source);
-		return PTR_ERR(bdev_handle);
+		return PTR_ERR(bdev_file);
 	}
-	bdev = bdev_handle->bdev;
+	bdev = file_bdev(bdev_file);
 
 	/*
 	 * This really should be in blkdev_get_by_dev, but right now can't due
@@ -1544,7 +1565,7 @@ int setup_bdev_super(struct super_block *sb, int sb_flags,
 	 * writable from userspace even for a read-only block device.
 	 */
 	if ((mode & BLK_OPEN_WRITE) && bdev_read_only(bdev)) {
-		bdev_release(bdev_handle);
+		bdev_fput(bdev_file);
 		return -EACCES;
 	}
 
@@ -1555,11 +1576,11 @@ int setup_bdev_super(struct super_block *sb, int sb_flags,
 	if (atomic_read(&bdev->bd_fsfreeze_count) > 0) {
 		if (fc)
 			warnf(fc, "%pg: Can't mount, blockdev is frozen", bdev);
-		bdev_release(bdev_handle);
+		bdev_fput(bdev_file);
 		return -EBUSY;
 	}
 	spin_lock(&sb_lock);
-	sb->s_bdev_handle = bdev_handle;
+	sb->s_bdev_file = bdev_file;
 	sb->s_bdev = bdev;
 	sb->s_bdi = bdi_get(bdev->bd_disk->bdi);
 	if (bdev_stable_writes(bdev))
@@ -1675,7 +1696,7 @@ void kill_block_super(struct super_block *sb)
 	generic_shutdown_super(sb);
 	if (bdev) {
 		sync_blockdev(bdev);
-		bdev_release(sb->s_bdev_handle);
+		bdev_fput(sb->s_bdev_file);
 	}
 }
 
diff --git a/fs/sysfs/file.c b/fs/sysfs/file.c
index 7cd64021d4..d1995e2d6c 100644
--- a/fs/sysfs/file.c
+++ b/fs/sysfs/file.c
@@ -785,3 +785,30 @@ int sysfs_emit_at(char *buf, int at, const char *fmt, ...)
 	return len;
 }
 EXPORT_SYMBOL_GPL(sysfs_emit_at);
+
+/**
+ *	sysfs_bin_attr_simple_read - read callback to simply copy from memory.
+ *	@file:	attribute file which is being read.
+ *	@kobj:	object to which the attribute belongs.
+ *	@attr:	attribute descriptor.
+ *	@buf:	destination buffer.
+ *	@off:	offset in bytes from which to read.
+ *	@count:	maximum number of bytes to read.
+ *
+ * Simple ->read() callback for bin_attributes backed by a buffer in memory.
+ * The @private and @size members in struct bin_attribute must be set to the
+ * buffer's location and size before the bin_attribute is created in sysfs.
+ *
+ * Bounds check for @off and @count is done in sysfs_kf_bin_read().
+ * Negative value check for @off is done in vfs_setpos() and default_llseek().
+ *
+ * Returns number of bytes written to @buf.
+ */
+ssize_t sysfs_bin_attr_simple_read(struct file *file, struct kobject *kobj,
+				   struct bin_attribute *attr, char *buf,
+				   loff_t off, size_t count)
+{
+	memcpy(buf, attr->private + off, count);
+	return count;
+}
+EXPORT_SYMBOL_GPL(sysfs_bin_attr_simple_read);
diff --git a/fs/sysfs/group.c b/fs/sysfs/group.c
index 1386764633..d22ad67a0f 100644
--- a/fs/sysfs/group.c
+++ b/fs/sysfs/group.c
@@ -31,6 +31,17 @@ static void remove_files(struct kernfs_node *parent,
 			kernfs_remove_by_name(parent, (*bin_attr)->attr.name);
 }
 
+static umode_t __first_visible(const struct attribute_group *grp, struct kobject *kobj)
+{
+	if (grp->attrs && grp->attrs[0] && grp->is_visible)
+		return grp->is_visible(kobj, grp->attrs[0], 0);
+
+	if (grp->bin_attrs && grp->bin_attrs[0] && grp->is_bin_visible)
+		return grp->is_bin_visible(kobj, grp->bin_attrs[0], 0);
+
+	return 0;
+}
+
 static int create_files(struct kernfs_node *parent, struct kobject *kobj,
 			kuid_t uid, kgid_t gid,
 			const struct attribute_group *grp, int update)
@@ -52,6 +63,7 @@ static int create_files(struct kernfs_node *parent, struct kobject *kobj,
 				kernfs_remove_by_name(parent, (*attr)->name);
 			if (grp->is_visible) {
 				mode = grp->is_visible(kobj, *attr, i);
+				mode &= ~SYSFS_GROUP_INVISIBLE;
 				if (!mode)
 					continue;
 			}
@@ -81,6 +93,7 @@ static int create_files(struct kernfs_node *parent, struct kobject *kobj,
 						(*bin_attr)->attr.name);
 			if (grp->is_bin_visible) {
 				mode = grp->is_bin_visible(kobj, *bin_attr, i);
+				mode &= ~SYSFS_GROUP_INVISIBLE;
 				if (!mode)
 					continue;
 			}
@@ -127,16 +140,31 @@ static int internal_create_group(struct kobject *kobj, int update,
 
 	kobject_get_ownership(kobj, &uid, &gid);
 	if (grp->name) {
+		umode_t mode = __first_visible(grp, kobj);
+
+		if (mode & SYSFS_GROUP_INVISIBLE)
+			mode = 0;
+		else
+			mode = S_IRWXU | S_IRUGO | S_IXUGO;
+
 		if (update) {
 			kn = kernfs_find_and_get(kobj->sd, grp->name);
 			if (!kn) {
-				pr_warn("Can't update unknown attr grp name: %s/%s\n",
-					kobj->name, grp->name);
-				return -EINVAL;
+				pr_debug("attr grp %s/%s not created yet\n",
+					 kobj->name, grp->name);
+				/* may have been invisible prior to this update */
+				update = 0;
+			} else if (!mode) {
+				sysfs_remove_group(kobj, grp);
+				kernfs_put(kn);
+				return 0;
 			}
-		} else {
-			kn = kernfs_create_dir_ns(kobj->sd, grp->name,
-						  S_IRWXU | S_IRUGO | S_IXUGO,
+		}
+
+		if (!update) {
+			if (!mode)
+				return 0;
+			kn = kernfs_create_dir_ns(kobj->sd, grp->name, mode,
 						  uid, gid, kobj, NULL);
 			if (IS_ERR(kn)) {
 				if (PTR_ERR(kn) == -EEXIST)
@@ -279,9 +307,8 @@ void sysfs_remove_group(struct kobject *kobj,
 	if (grp->name) {
 		kn = kernfs_find_and_get(parent, grp->name);
 		if (!kn) {
-			WARN(!kn, KERN_WARNING
-			     "sysfs group '%s' not found for kobject '%s'\n",
-			     grp->name, kobject_name(kobj));
+			pr_debug("sysfs group '%s' not found for kobject '%s'\n",
+				 grp->name, kobject_name(kobj));
 			return;
 		}
 	} else {
@@ -318,13 +345,13 @@ void sysfs_remove_groups(struct kobject *kobj,
 EXPORT_SYMBOL_GPL(sysfs_remove_groups);
 
 /**
- * sysfs_merge_group - merge files into a pre-existing attribute group.
+ * sysfs_merge_group - merge files into a pre-existing named attribute group.
  * @kobj:	The kobject containing the group.
  * @grp:	The files to create and the attribute group they belong to.
  *
- * This function returns an error if the group doesn't exist or any of the
- * files already exist in that group, in which case none of the new files
- * are created.
+ * This function returns an error if the group doesn't exist, the .name field is
+ * NULL or any of the files already exist in that group, in which case none of
+ * the new files are created.
  */
 int sysfs_merge_group(struct kobject *kobj,
 		       const struct attribute_group *grp)
@@ -356,7 +383,7 @@ int sysfs_merge_group(struct kobject *kobj,
 EXPORT_SYMBOL_GPL(sysfs_merge_group);
 
 /**
- * sysfs_unmerge_group - remove files from a pre-existing attribute group.
+ * sysfs_unmerge_group - remove files from a pre-existing named attribute group.
  * @kobj:	The kobject containing the group.
  * @grp:	The files to remove and the attribute group they belong to.
  */
diff --git a/fs/sysv/inode.c b/fs/sysv/inode.c
index 5a915b2e68..76bc2d5e75 100644
--- a/fs/sysv/inode.c
+++ b/fs/sysv/inode.c
@@ -336,7 +336,7 @@ int __init sysv_init_icache(void)
 {
 	sysv_inode_cachep = kmem_cache_create("sysv_inode_cache",
 			sizeof(struct sysv_inode_info), 0,
-			SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD|SLAB_ACCOUNT,
+			SLAB_RECLAIM_ACCOUNT|SLAB_ACCOUNT,
 			init_once);
 	if (!sysv_inode_cachep)
 		return -ENOMEM;
diff --git a/fs/timerfd.c b/fs/timerfd.c
index e9c96a0c79..4bf2f8bfec 100644
--- a/fs/timerfd.c
+++ b/fs/timerfd.c
@@ -262,17 +262,18 @@ static __poll_t timerfd_poll(struct file *file, poll_table *wait)
 	return events;
 }
 
-static ssize_t timerfd_read(struct file *file, char __user *buf, size_t count,
-			    loff_t *ppos)
+static ssize_t timerfd_read_iter(struct kiocb *iocb, struct iov_iter *to)
 {
+	struct file *file = iocb->ki_filp;
 	struct timerfd_ctx *ctx = file->private_data;
 	ssize_t res;
 	u64 ticks = 0;
 
-	if (count < sizeof(ticks))
+	if (iov_iter_count(to) < sizeof(ticks))
 		return -EINVAL;
+
 	spin_lock_irq(&ctx->wqh.lock);
-	if (file->f_flags & O_NONBLOCK)
+	if (file->f_flags & O_NONBLOCK || iocb->ki_flags & IOCB_NOWAIT)
 		res = -EAGAIN;
 	else
 		res = wait_event_interruptible_locked_irq(ctx->wqh, ctx->ticks);
@@ -312,8 +313,11 @@ static ssize_t timerfd_read(struct file *file, char __user *buf, size_t count,
 		ctx->ticks = 0;
 	}
 	spin_unlock_irq(&ctx->wqh.lock);
-	if (ticks)
-		res = put_user(ticks, (u64 __user *) buf) ? -EFAULT: sizeof(ticks);
+	if (ticks) {
+		res = copy_to_iter(&ticks, sizeof(ticks), to);
+		if (!res)
+			res = -EFAULT;
+	}
 	return res;
 }
 
@@ -384,7 +388,7 @@ static long timerfd_ioctl(struct file *file, unsigned int cmd, unsigned long arg
 static const struct file_operations timerfd_fops = {
 	.release	= timerfd_release,
 	.poll		= timerfd_poll,
-	.read		= timerfd_read,
+	.read_iter	= timerfd_read_iter,
 	.llseek		= noop_llseek,
 	.show_fdinfo	= timerfd_show,
 	.unlocked_ioctl	= timerfd_ioctl,
@@ -407,6 +411,7 @@ SYSCALL_DEFINE2(timerfd_create, int, clockid, int, flags)
 {
 	int ufd;
 	struct timerfd_ctx *ctx;
+	struct file *file;
 
 	/* Check the TFD_* constants for consistency.  */
 	BUILD_BUG_ON(TFD_CLOEXEC != O_CLOEXEC);
@@ -443,11 +448,22 @@ SYSCALL_DEFINE2(timerfd_create, int, clockid, int, flags)
 
 	ctx->moffs = ktime_mono_to_real(0);
 
-	ufd = anon_inode_getfd("[timerfd]", &timerfd_fops, ctx,
-			       O_RDWR | (flags & TFD_SHARED_FCNTL_FLAGS));
-	if (ufd < 0)
+	ufd = get_unused_fd_flags(flags & TFD_SHARED_FCNTL_FLAGS);
+	if (ufd < 0) {
+		kfree(ctx);
+		return ufd;
+	}
+
+	file = anon_inode_getfile("[timerfd]", &timerfd_fops, ctx,
+				    O_RDWR | (flags & TFD_SHARED_FCNTL_FLAGS));
+	if (IS_ERR(file)) {
+		put_unused_fd(ufd);
 		kfree(ctx);
+		return PTR_ERR(file);
+	}
 
+	file->f_mode |= FMODE_NOWAIT;
+	fd_install(ufd, file);
 	return ufd;
 }
 
diff --git a/fs/tracefs/event_inode.c b/fs/tracefs/event_inode.c
index 56d1741fe0..01e99e9845 100644
--- a/fs/tracefs/event_inode.c
+++ b/fs/tracefs/event_inode.c
@@ -35,11 +35,26 @@ static DEFINE_MUTEX(eventfs_mutex);
 /* Choose something "unique" ;-) */
 #define EVENTFS_FILE_INODE_INO		0x12c4e37
 
+struct eventfs_root_inode {
+	struct eventfs_inode		ei;
+	struct dentry			*events_dir;
+};
+
+static struct eventfs_root_inode *get_root_inode(struct eventfs_inode *ei)
+{
+	WARN_ON_ONCE(!ei->is_events);
+	return container_of(ei, struct eventfs_root_inode, ei);
+}
+
 /* Just try to make something consistent and unique */
 static int eventfs_dir_ino(struct eventfs_inode *ei)
 {
-	if (!ei->ino)
+	if (!ei->ino) {
 		ei->ino = get_next_ino();
+		/* Must not have the file inode number */
+		if (ei->ino == EVENTFS_FILE_INODE_INO)
+			ei->ino = get_next_ino();
+	}
 
 	return ei->ino;
 }
@@ -57,11 +72,25 @@ enum {
 	EVENTFS_SAVE_MODE	= BIT(16),
 	EVENTFS_SAVE_UID	= BIT(17),
 	EVENTFS_SAVE_GID	= BIT(18),
-	EVENTFS_TOPLEVEL	= BIT(19),
 };
 
 #define EVENTFS_MODE_MASK	(EVENTFS_SAVE_MODE - 1)
 
+static void free_ei_rcu(struct rcu_head *rcu)
+{
+	struct eventfs_inode *ei = container_of(rcu, struct eventfs_inode, rcu);
+	struct eventfs_root_inode *rei;
+
+	kfree(ei->entry_attrs);
+	kfree_const(ei->name);
+	if (ei->is_events) {
+		rei = get_root_inode(ei);
+		kfree(rei);
+	} else {
+		kfree(ei);
+	}
+}
+
 /*
  * eventfs_inode reference count management.
  *
@@ -73,12 +102,17 @@ enum {
 static void release_ei(struct kref *ref)
 {
 	struct eventfs_inode *ei = container_of(ref, struct eventfs_inode, kref);
+	const struct eventfs_entry *entry;
 
 	WARN_ON_ONCE(!ei->is_freed);
 
-	kfree(ei->entry_attrs);
-	kfree_const(ei->name);
-	kfree_rcu(ei, rcu);
+	for (int i = 0; i < ei->nr_entries; i++) {
+		entry = &ei->entries[i];
+		if (entry->release)
+			entry->release(entry->name, ei->data);
+	}
+
+	call_srcu(&eventfs_srcu, &ei->rcu, free_ei_rcu);
 }
 
 static inline void put_ei(struct eventfs_inode *ei)
@@ -95,6 +129,18 @@ static inline void free_ei(struct eventfs_inode *ei)
 	}
 }
 
+/*
+ * Called when creation of an ei fails, do not call release() functions.
+ */
+static inline void cleanup_ei(struct eventfs_inode *ei)
+{
+	if (ei) {
+		/* Set nr_entries to 0 to prevent release() function being called */
+		ei->nr_entries = 0;
+		free_ei(ei);
+	}
+}
+
 static inline struct eventfs_inode *get_ei(struct eventfs_inode *ei)
 {
 	if (ei)
@@ -164,7 +210,9 @@ static int eventfs_set_attr(struct mnt_idmap *idmap, struct dentry *dentry,
 	 * determined by the parent directory.
 	 */
 	if (dentry->d_inode->i_mode & S_IFDIR) {
-		update_attr(&ei->attr, iattr);
+		/* Just use the inode permissions for the events directory */
+		if (!ei->is_events)
+			update_attr(&ei->attr, iattr);
 
 	} else {
 		name = dentry->d_name.name;
@@ -182,63 +230,9 @@ static int eventfs_set_attr(struct mnt_idmap *idmap, struct dentry *dentry,
 	return ret;
 }
 
-static void update_top_events_attr(struct eventfs_inode *ei, struct super_block *sb)
-{
-	struct inode *root;
-
-	/* Only update if the "events" was on the top level */
-	if (!ei || !(ei->attr.mode & EVENTFS_TOPLEVEL))
-		return;
-
-	/* Get the tracefs root inode. */
-	root = d_inode(sb->s_root);
-	ei->attr.uid = root->i_uid;
-	ei->attr.gid = root->i_gid;
-}
-
-static void set_top_events_ownership(struct inode *inode)
-{
-	struct tracefs_inode *ti = get_tracefs(inode);
-	struct eventfs_inode *ei = ti->private;
-
-	/* The top events directory doesn't get automatically updated */
-	if (!ei || !ei->is_events || !(ei->attr.mode & EVENTFS_TOPLEVEL))
-		return;
-
-	update_top_events_attr(ei, inode->i_sb);
-
-	if (!(ei->attr.mode & EVENTFS_SAVE_UID))
-		inode->i_uid = ei->attr.uid;
-
-	if (!(ei->attr.mode & EVENTFS_SAVE_GID))
-		inode->i_gid = ei->attr.gid;
-}
-
-static int eventfs_get_attr(struct mnt_idmap *idmap,
-			    const struct path *path, struct kstat *stat,
-			    u32 request_mask, unsigned int flags)
-{
-	struct dentry *dentry = path->dentry;
-	struct inode *inode = d_backing_inode(dentry);
-
-	set_top_events_ownership(inode);
-
-	generic_fillattr(idmap, request_mask, inode, stat);
-	return 0;
-}
-
-static int eventfs_permission(struct mnt_idmap *idmap,
-			      struct inode *inode, int mask)
-{
-	set_top_events_ownership(inode);
-	return generic_permission(idmap, inode, mask);
-}
-
-static const struct inode_operations eventfs_root_dir_inode_operations = {
+static const struct inode_operations eventfs_dir_inode_operations = {
 	.lookup		= eventfs_root_lookup,
 	.setattr	= eventfs_set_attr,
-	.getattr	= eventfs_get_attr,
-	.permission	= eventfs_permission,
 };
 
 static const struct inode_operations eventfs_file_inode_operations = {
@@ -251,89 +245,114 @@ static const struct file_operations eventfs_file_operations = {
 	.llseek		= generic_file_llseek,
 };
 
-/*
- * On a remount of tracefs, if UID or GID options are set, then
- * the mount point inode permissions should be used.
- * Reset the saved permission flags appropriately.
- */
-void eventfs_remount(struct tracefs_inode *ti, bool update_uid, bool update_gid)
+static void eventfs_set_attrs(struct eventfs_inode *ei, bool update_uid, kuid_t uid,
+			      bool update_gid, kgid_t gid, int level)
 {
-	struct eventfs_inode *ei = ti->private;
+	struct eventfs_inode *ei_child;
 
-	if (!ei)
+	/* Update events/<system>/<event> */
+	if (WARN_ON_ONCE(level > 3))
 		return;
 
-	if (update_uid)
+	if (update_uid) {
 		ei->attr.mode &= ~EVENTFS_SAVE_UID;
+		ei->attr.uid = uid;
+	}
 
-	if (update_gid)
+	if (update_gid) {
 		ei->attr.mode &= ~EVENTFS_SAVE_GID;
+		ei->attr.gid = gid;
+	}
+
+	list_for_each_entry(ei_child, &ei->children, list) {
+		eventfs_set_attrs(ei_child, update_uid, uid, update_gid, gid, level + 1);
+	}
 
 	if (!ei->entry_attrs)
 		return;
 
 	for (int i = 0; i < ei->nr_entries; i++) {
-		if (update_uid)
+		if (update_uid) {
 			ei->entry_attrs[i].mode &= ~EVENTFS_SAVE_UID;
-		if (update_gid)
+			ei->entry_attrs[i].uid = uid;
+		}
+		if (update_gid) {
 			ei->entry_attrs[i].mode &= ~EVENTFS_SAVE_GID;
+			ei->entry_attrs[i].gid = gid;
+		}
 	}
+
 }
 
-/* Return the evenfs_inode of the "events" directory */
-static struct eventfs_inode *eventfs_find_events(struct dentry *dentry)
+/*
+ * On a remount of tracefs, if UID or GID options are set, then
+ * the mount point inode permissions should be used.
+ * Reset the saved permission flags appropriately.
+ */
+void eventfs_remount(struct tracefs_inode *ti, bool update_uid, bool update_gid)
 {
-	struct eventfs_inode *ei;
+	struct eventfs_inode *ei = ti->private;
 
-	do {
-		// The parent is stable because we do not do renames
-		dentry = dentry->d_parent;
-		// ... and directories always have d_fsdata
-		ei = dentry->d_fsdata;
+	/* Only the events directory does the updates */
+	if (!ei || !ei->is_events || ei->is_freed)
+		return;
 
-		/*
-		 * If the ei is being freed, the ownership of the children
-		 * doesn't matter.
-		 */
-		if (ei->is_freed) {
-			ei = NULL;
-			break;
-		}
-		// Walk upwards until you find the events inode
-	} while (!ei->is_events);
+	eventfs_set_attrs(ei, update_uid, ti->vfs_inode.i_uid,
+			  update_gid, ti->vfs_inode.i_gid, 0);
+}
 
-	update_top_events_attr(ei, dentry->d_sb);
+static void update_inode_attr(struct inode *inode, umode_t mode,
+			      struct eventfs_attr *attr, struct eventfs_root_inode *rei)
+{
+	if (attr && attr->mode & EVENTFS_SAVE_MODE)
+		inode->i_mode = attr->mode & EVENTFS_MODE_MASK;
+	else
+		inode->i_mode = mode;
 
-	return ei;
+	if (attr && attr->mode & EVENTFS_SAVE_UID)
+		inode->i_uid = attr->uid;
+	else
+		inode->i_uid = rei->ei.attr.uid;
+
+	if (attr && attr->mode & EVENTFS_SAVE_GID)
+		inode->i_gid = attr->gid;
+	else
+		inode->i_gid = rei->ei.attr.gid;
 }
 
-static void update_inode_attr(struct dentry *dentry, struct inode *inode,
-			      struct eventfs_attr *attr, umode_t mode)
+static struct inode *eventfs_get_inode(struct dentry *dentry, struct eventfs_attr *attr,
+				       umode_t mode,  struct eventfs_inode *ei)
 {
-	struct eventfs_inode *events_ei = eventfs_find_events(dentry);
+	struct eventfs_root_inode *rei;
+	struct eventfs_inode *pei;
+	struct tracefs_inode *ti;
+	struct inode *inode;
 
-	if (!events_ei)
-		return;
+	inode = tracefs_get_inode(dentry->d_sb);
+	if (!inode)
+		return NULL;
 
-	inode->i_mode = mode;
-	inode->i_uid = events_ei->attr.uid;
-	inode->i_gid = events_ei->attr.gid;
+	ti = get_tracefs(inode);
+	ti->private = ei;
+	ti->flags |= TRACEFS_EVENT_INODE;
 
-	if (!attr)
-		return;
+	/* Find the top dentry that holds the "events" directory */
+	do {
+		dentry = dentry->d_parent;
+		/* Directories always have d_fsdata */
+		pei = dentry->d_fsdata;
+	} while (!pei->is_events);
 
-	if (attr->mode & EVENTFS_SAVE_MODE)
-		inode->i_mode = attr->mode & EVENTFS_MODE_MASK;
+	rei = get_root_inode(pei);
 
-	if (attr->mode & EVENTFS_SAVE_UID)
-		inode->i_uid = attr->uid;
+	update_inode_attr(inode, mode, attr, rei);
 
-	if (attr->mode & EVENTFS_SAVE_GID)
-		inode->i_gid = attr->gid;
+	return inode;
 }
 
 /**
  * lookup_file - look up a file in the tracefs filesystem
+ * @parent_ei: Pointer to the eventfs_inode that represents parent of the file
  * @dentry: the dentry to look up
  * @mode: the permission that the file should have.
  * @attr: saved attributes changed by user
@@ -351,7 +370,6 @@ static struct dentry *lookup_file(struct eventfs_inode *parent_ei,
 				  void *data,
 				  const struct file_operations *fop)
 {
-	struct tracefs_inode *ti;
 	struct inode *inode;
 
 	if (!(mode & S_IFMT))
@@ -360,13 +378,11 @@ static struct dentry *lookup_file(struct eventfs_inode *parent_ei,
 	if (WARN_ON_ONCE(!S_ISREG(mode)))
 		return ERR_PTR(-EIO);
 
-	inode = tracefs_get_inode(dentry->d_sb);
+	/* Only directories have ti->private set to an ei, not files */
+	inode = eventfs_get_inode(dentry, attr, mode, NULL);
 	if (unlikely(!inode))
 		return ERR_PTR(-ENOMEM);
 
-	/* If the user updated the directory's attributes, use them */
-	update_inode_attr(dentry, inode, attr, mode);
-
 	inode->i_op = &eventfs_file_inode_operations;
 	inode->i_fop = fop;
 	inode->i_private = data;
@@ -374,9 +390,6 @@ static struct dentry *lookup_file(struct eventfs_inode *parent_ei,
 	/* All files will have the same inode number */
 	inode->i_ino = EVENTFS_FILE_INODE_INO;
 
-	ti = get_tracefs(inode);
-	ti->flags |= TRACEFS_EVENT_INODE;
-
 	// Files have their parent's ei as their fsdata
 	dentry->d_fsdata = get_ei(parent_ei);
 
@@ -387,6 +400,7 @@ static struct dentry *lookup_file(struct eventfs_inode *parent_ei,
 /**
  * lookup_dir_entry - look up a dir in the tracefs filesystem
  * @dentry: the directory to look up
+ * @pei: Pointer to the parent eventfs_inode if available
  * @ei: the eventfs_inode that represents the directory to create
  *
  * This function will look up a dentry for a directory represented by
@@ -395,47 +409,62 @@ static struct dentry *lookup_file(struct eventfs_inode *parent_ei,
 static struct dentry *lookup_dir_entry(struct dentry *dentry,
 	struct eventfs_inode *pei, struct eventfs_inode *ei)
 {
-	struct tracefs_inode *ti;
 	struct inode *inode;
+	umode_t mode = S_IFDIR | S_IRWXU | S_IRUGO | S_IXUGO;
 
-	inode = tracefs_get_inode(dentry->d_sb);
+	inode = eventfs_get_inode(dentry, &ei->attr, mode, ei);
 	if (unlikely(!inode))
 		return ERR_PTR(-ENOMEM);
 
-	/* If the user updated the directory's attributes, use them */
-	update_inode_attr(dentry, inode, &ei->attr,
-			  S_IFDIR | S_IRWXU | S_IRUGO | S_IXUGO);
-
-	inode->i_op = &eventfs_root_dir_inode_operations;
+	inode->i_op = &eventfs_dir_inode_operations;
 	inode->i_fop = &eventfs_file_operations;
 
 	/* All directories will have the same inode number */
 	inode->i_ino = eventfs_dir_ino(ei);
 
-	ti = get_tracefs(inode);
-	ti->flags |= TRACEFS_EVENT_INODE;
-	/* Only directories have ti->private set to an ei, not files */
-	ti->private = ei;
-
 	dentry->d_fsdata = get_ei(ei);
 
 	d_add(dentry, inode);
 	return NULL;
 }
 
+static inline struct eventfs_inode *init_ei(struct eventfs_inode *ei, const char *name)
+{
+	ei->name = kstrdup_const(name, GFP_KERNEL);
+	if (!ei->name)
+		return NULL;
+	kref_init(&ei->kref);
+	return ei;
+}
+
 static inline struct eventfs_inode *alloc_ei(const char *name)
 {
 	struct eventfs_inode *ei = kzalloc(sizeof(*ei), GFP_KERNEL);
+	struct eventfs_inode *result;
 
 	if (!ei)
 		return NULL;
 
-	ei->name = kstrdup_const(name, GFP_KERNEL);
-	if (!ei->name) {
+	result = init_ei(ei, name);
+	if (!result)
 		kfree(ei);
+
+	return result;
+}
+
+static inline struct eventfs_inode *alloc_root_ei(const char *name)
+{
+	struct eventfs_root_inode *rei = kzalloc(sizeof(*rei), GFP_KERNEL);
+	struct eventfs_inode *ei;
+
+	if (!rei)
 		return NULL;
-	}
-	kref_init(&ei->kref);
+
+	rei->ei.is_events = 1;
+	ei = init_ei(&rei->ei, name);
+	if (!ei)
+		kfree(rei);
+
 	return ei;
 }
 
@@ -452,16 +481,20 @@ void eventfs_d_release(struct dentry *dentry)
 
 /**
  * lookup_file_dentry - create a dentry for a file of an eventfs_inode
+ * @dentry: The parent dentry under which the new file's dentry will be created
  * @ei: the eventfs_inode that the file will be created under
  * @idx: the index into the entry_attrs[] of the @ei
- * @parent: The parent dentry of the created file.
- * @name: The name of the file to create
  * @mode: The mode of the file.
  * @data: The data to use to set the inode of the file with on open()
  * @fops: The fops of the file to be created.
  *
- * Create a dentry for a file of an eventfs_inode @ei and place it into the
- * address located at @e_dentry.
+ * This function creates a dentry for a file associated with an
+ * eventfs_inode @ei. It uses the entry attributes specified by @idx,
+ * if available. The file will have the specified @mode and its inode will be
+ * set up with @data upon open. The file operations will be set to @fops.
+ *
+ * Return: Returns a pointer to the newly created file's dentry or an error
+ * pointer.
  */
 static struct dentry *
 lookup_file_dentry(struct dentry *dentry,
@@ -498,7 +531,7 @@ static struct dentry *eventfs_root_lookup(struct inode *dir,
 	struct dentry *result = NULL;
 
 	ti = get_tracefs(dir);
-	if (!(ti->flags & TRACEFS_EVENT_INODE))
+	if (WARN_ON_ONCE(!(ti->flags & TRACEFS_EVENT_INODE)))
 		return ERR_PTR(-EIO);
 
 	mutex_lock(&eventfs_mutex);
@@ -510,7 +543,8 @@ static struct dentry *eventfs_root_lookup(struct inode *dir,
 	list_for_each_entry(ei_child, &ei->children, list) {
 		if (strcmp(ei_child->name, name) != 0)
 			continue;
-		if (ei_child->is_freed)
+		/* A child is freed and removed from the list at the same time */
+		if (WARN_ON_ONCE(ei_child->is_freed))
 			goto out;
 		result = lookup_dir_entry(dentry, ei, ei_child);
 		goto out;
@@ -701,8 +735,8 @@ struct eventfs_inode *eventfs_create_dir(const char *name, struct eventfs_inode
 
 	/* Was the parent freed? */
 	if (list_empty(&ei->list)) {
-		free_ei(ei);
-		ei = NULL;
+		cleanup_ei(ei);
+		ei = ERR_PTR(-EBUSY);
 	}
 	return ei;
 }
@@ -724,6 +758,7 @@ struct eventfs_inode *eventfs_create_events_dir(const char *name, struct dentry
 						int size, void *data)
 {
 	struct dentry *dentry = tracefs_start_creating(name, parent);
+	struct eventfs_root_inode *rei;
 	struct eventfs_inode *ei;
 	struct tracefs_inode *ti;
 	struct inode *inode;
@@ -736,7 +771,7 @@ struct eventfs_inode *eventfs_create_events_dir(const char *name, struct dentry
 	if (IS_ERR(dentry))
 		return ERR_CAST(dentry);
 
-	ei = alloc_ei(name);
+	ei = alloc_root_ei(name);
 	if (!ei)
 		goto fail;
 
@@ -745,10 +780,11 @@ struct eventfs_inode *eventfs_create_events_dir(const char *name, struct dentry
 		goto fail;
 
 	// Note: we have a ref to the dentry from tracefs_start_creating()
-	ei->events_dir = dentry;
+	rei = get_root_inode(ei);
+	rei->events_dir = dentry;
+
 	ei->entries = entries;
 	ei->nr_entries = size;
-	ei->is_events = 1;
 	ei->data = data;
 
 	/* Save the ownership of this directory */
@@ -756,14 +792,9 @@ struct eventfs_inode *eventfs_create_events_dir(const char *name, struct dentry
 	gid = d_inode(dentry->d_parent)->i_gid;
 
 	/*
-	 * If the events directory is of the top instance, then parent
-	 * is NULL. Set the attr.mode to reflect this and its permissions will
-	 * default to the tracefs root dentry.
+	 * The ei->attr will be used as the default values for the
+	 * files beneath this directory.
 	 */
-	if (!parent)
-		ei->attr.mode = EVENTFS_TOPLEVEL;
-
-	/* This is used as the default ownership of the files and directories */
 	ei->attr.uid = uid;
 	ei->attr.gid = gid;
 
@@ -771,13 +802,13 @@ struct eventfs_inode *eventfs_create_events_dir(const char *name, struct dentry
 	INIT_LIST_HEAD(&ei->list);
 
 	ti = get_tracefs(inode);
-	ti->flags |= TRACEFS_EVENT_INODE | TRACEFS_EVENT_TOP_INODE;
+	ti->flags |= TRACEFS_EVENT_INODE;
 	ti->private = ei;
 
 	inode->i_mode = S_IFDIR | S_IRWXU | S_IRUGO | S_IXUGO;
 	inode->i_uid = uid;
 	inode->i_gid = gid;
-	inode->i_op = &eventfs_root_dir_inode_operations;
+	inode->i_op = &eventfs_dir_inode_operations;
 	inode->i_fop = &eventfs_file_operations;
 
 	dentry->d_fsdata = get_ei(ei);
@@ -800,7 +831,7 @@ struct eventfs_inode *eventfs_create_events_dir(const char *name, struct dentry
 	return ei;
 
  fail:
-	free_ei(ei);
+	cleanup_ei(ei);
 	tracefs_failed_creating(dentry);
 	return ERR_PTR(-ENOMEM);
 }
@@ -859,13 +890,15 @@ void eventfs_remove_dir(struct eventfs_inode *ei)
  */
 void eventfs_remove_events_dir(struct eventfs_inode *ei)
 {
+	struct eventfs_root_inode *rei;
 	struct dentry *dentry;
 
-	dentry = ei->events_dir;
+	rei = get_root_inode(ei);
+	dentry = rei->events_dir;
 	if (!dentry)
 		return;
 
-	ei->events_dir = NULL;
+	rei->events_dir = NULL;
 	eventfs_remove_dir(ei);
 
 	/*
diff --git a/fs/tracefs/inode.c b/fs/tracefs/inode.c
index e30b74228e..507a61004e 100644
--- a/fs/tracefs/inode.c
+++ b/fs/tracefs/inode.c
@@ -11,14 +11,14 @@
 
 #include <linux/module.h>
 #include <linux/fs.h>
-#include <linux/mount.h>
+#include <linux/fs_context.h>
+#include <linux/fs_parser.h>
 #include <linux/kobject.h>
 #include <linux/namei.h>
 #include <linux/tracefs.h>
 #include <linux/fsnotify.h>
 #include <linux/security.h>
 #include <linux/seq_file.h>
-#include <linux/parser.h>
 #include <linux/magic.h>
 #include <linux/slab.h>
 #include "internal.h"
@@ -42,7 +42,7 @@ static struct inode *tracefs_alloc_inode(struct super_block *sb)
 	struct tracefs_inode *ti;
 	unsigned long flags;
 
-	ti = kmem_cache_alloc(tracefs_inode_cachep, GFP_KERNEL);
+	ti = alloc_inode_sb(sb, tracefs_inode_cachep, GFP_KERNEL);
 	if (!ti)
 		return NULL;
 
@@ -53,15 +53,14 @@ static struct inode *tracefs_alloc_inode(struct super_block *sb)
 	return &ti->vfs_inode;
 }
 
-static void tracefs_free_inode_rcu(struct rcu_head *rcu)
+static void tracefs_free_inode(struct inode *inode)
 {
-	struct tracefs_inode *ti;
+	struct tracefs_inode *ti = get_tracefs(inode);
 
-	ti = container_of(rcu, struct tracefs_inode, rcu);
 	kmem_cache_free(tracefs_inode_cachep, ti);
 }
 
-static void tracefs_free_inode(struct inode *inode)
+static void tracefs_destroy_inode(struct inode *inode)
 {
 	struct tracefs_inode *ti = get_tracefs(inode);
 	unsigned long flags;
@@ -69,8 +68,6 @@ static void tracefs_free_inode(struct inode *inode)
 	spin_lock_irqsave(&tracefs_inode_lock, flags);
 	list_del_rcu(&ti->list);
 	spin_unlock_irqrestore(&tracefs_inode_lock, flags);
-
-	call_rcu(&ti->rcu, tracefs_free_inode_rcu);
 }
 
 static ssize_t default_read_file(struct file *file, char __user *buf,
@@ -281,7 +278,7 @@ struct inode *tracefs_get_inode(struct super_block *sb)
 	return inode;
 }
 
-struct tracefs_mount_opts {
+struct tracefs_fs_info {
 	kuid_t uid;
 	kgid_t gid;
 	umode_t mode;
@@ -293,68 +290,51 @@ enum {
 	Opt_uid,
 	Opt_gid,
 	Opt_mode,
-	Opt_err
-};
-
-static const match_table_t tokens = {
-	{Opt_uid, "uid=%u"},
-	{Opt_gid, "gid=%u"},
-	{Opt_mode, "mode=%o"},
-	{Opt_err, NULL}
 };
 
-struct tracefs_fs_info {
-	struct tracefs_mount_opts mount_opts;
+static const struct fs_parameter_spec tracefs_param_specs[] = {
+	fsparam_u32	("gid",		Opt_gid),
+	fsparam_u32oct	("mode",	Opt_mode),
+	fsparam_u32	("uid",		Opt_uid),
+	{}
 };
 
-static int tracefs_parse_options(char *data, struct tracefs_mount_opts *opts)
+static int tracefs_parse_param(struct fs_context *fc, struct fs_parameter *param)
 {
-	substring_t args[MAX_OPT_ARGS];
-	int option;
-	int token;
+	struct tracefs_fs_info *opts = fc->s_fs_info;
+	struct fs_parse_result result;
 	kuid_t uid;
 	kgid_t gid;
-	char *p;
-
-	opts->opts = 0;
-	opts->mode = TRACEFS_DEFAULT_MODE;
-
-	while ((p = strsep(&data, ",")) != NULL) {
-		if (!*p)
-			continue;
-
-		token = match_token(p, tokens, args);
-		switch (token) {
-		case Opt_uid:
-			if (match_int(&args[0], &option))
-				return -EINVAL;
-			uid = make_kuid(current_user_ns(), option);
-			if (!uid_valid(uid))
-				return -EINVAL;
-			opts->uid = uid;
-			break;
-		case Opt_gid:
-			if (match_int(&args[0], &option))
-				return -EINVAL;
-			gid = make_kgid(current_user_ns(), option);
-			if (!gid_valid(gid))
-				return -EINVAL;
-			opts->gid = gid;
-			break;
-		case Opt_mode:
-			if (match_octal(&args[0], &option))
-				return -EINVAL;
-			opts->mode = option & S_IALLUGO;
-			break;
-		/*
-		 * We might like to report bad mount options here;
-		 * but traditionally tracefs has ignored all mount options
-		 */
-		}
-
-		opts->opts |= BIT(token);
+	int opt;
+
+	opt = fs_parse(fc, tracefs_param_specs, param, &result);
+	if (opt < 0)
+		return opt;
+
+	switch (opt) {
+	case Opt_uid:
+		uid = make_kuid(current_user_ns(), result.uint_32);
+		if (!uid_valid(uid))
+			return invalf(fc, "Unknown uid");
+		opts->uid = uid;
+		break;
+	case Opt_gid:
+		gid = make_kgid(current_user_ns(), result.uint_32);
+		if (!gid_valid(gid))
+			return invalf(fc, "Unknown gid");
+		opts->gid = gid;
+		break;
+	case Opt_mode:
+		opts->mode = result.uint_32 & S_IALLUGO;
+		break;
+	/*
+	 * We might like to report bad mount options here;
+	 * but traditionally tracefs has ignored all mount options
+	 */
 	}
 
+	opts->opts |= BIT(opt);
+
 	return 0;
 }
 
@@ -362,7 +342,6 @@ static int tracefs_apply_options(struct super_block *sb, bool remount)
 {
 	struct tracefs_fs_info *fsi = sb->s_fs_info;
 	struct inode *inode = d_inode(sb->s_root);
-	struct tracefs_mount_opts *opts = &fsi->mount_opts;
 	struct tracefs_inode *ti;
 	bool update_uid, update_gid;
 	umode_t tmp_mode;
@@ -372,31 +351,40 @@ static int tracefs_apply_options(struct super_block *sb, bool remount)
 	 * options.
 	 */
 
-	if (!remount || opts->opts & BIT(Opt_mode)) {
+	if (!remount || fsi->opts & BIT(Opt_mode)) {
 		tmp_mode = READ_ONCE(inode->i_mode) & ~S_IALLUGO;
-		tmp_mode |= opts->mode;
+		tmp_mode |= fsi->mode;
 		WRITE_ONCE(inode->i_mode, tmp_mode);
 	}
 
-	if (!remount || opts->opts & BIT(Opt_uid))
-		inode->i_uid = opts->uid;
+	if (!remount || fsi->opts & BIT(Opt_uid))
+		inode->i_uid = fsi->uid;
 
-	if (!remount || opts->opts & BIT(Opt_gid))
-		inode->i_gid = opts->gid;
+	if (!remount || fsi->opts & BIT(Opt_gid))
+		inode->i_gid = fsi->gid;
 
-	if (remount && (opts->opts & BIT(Opt_uid) || opts->opts & BIT(Opt_gid))) {
+	if (remount && (fsi->opts & BIT(Opt_uid) || fsi->opts & BIT(Opt_gid))) {
 
-		update_uid = opts->opts & BIT(Opt_uid);
-		update_gid = opts->opts & BIT(Opt_gid);
+		update_uid = fsi->opts & BIT(Opt_uid);
+		update_gid = fsi->opts & BIT(Opt_gid);
 
 		rcu_read_lock();
 		list_for_each_entry_rcu(ti, &tracefs_inodes, list) {
-			if (update_uid)
+			if (update_uid) {
 				ti->flags &= ~TRACEFS_UID_PERM_SET;
+				ti->vfs_inode.i_uid = fsi->uid;
+			}
 
-			if (update_gid)
+			if (update_gid) {
 				ti->flags &= ~TRACEFS_GID_PERM_SET;
-
+				ti->vfs_inode.i_gid = fsi->gid;
+			}
+
+			/*
+			 * Note, the above ti->vfs_inode updates are
+			 * used in eventfs_remount() so they must come
+			 * before calling it.
+			 */
 			if (ti->flags & TRACEFS_EVENT_INODE)
 				eventfs_remount(ti, update_uid, update_gid);
 		}
@@ -406,45 +394,57 @@ static int tracefs_apply_options(struct super_block *sb, bool remount)
 	return 0;
 }
 
-static int tracefs_remount(struct super_block *sb, int *flags, char *data)
+static int tracefs_reconfigure(struct fs_context *fc)
 {
-	int err;
-	struct tracefs_fs_info *fsi = sb->s_fs_info;
+	struct super_block *sb = fc->root->d_sb;
+	struct tracefs_fs_info *sb_opts = sb->s_fs_info;
+	struct tracefs_fs_info *new_opts = fc->s_fs_info;
 
 	sync_filesystem(sb);
-	err = tracefs_parse_options(data, &fsi->mount_opts);
-	if (err)
-		goto fail;
+	/* structure copy of new mount options to sb */
+	*sb_opts = *new_opts;
 
-	tracefs_apply_options(sb, true);
-
-fail:
-	return err;
+	return tracefs_apply_options(sb, true);
 }
 
 static int tracefs_show_options(struct seq_file *m, struct dentry *root)
 {
 	struct tracefs_fs_info *fsi = root->d_sb->s_fs_info;
-	struct tracefs_mount_opts *opts = &fsi->mount_opts;
 
-	if (!uid_eq(opts->uid, GLOBAL_ROOT_UID))
+	if (!uid_eq(fsi->uid, GLOBAL_ROOT_UID))
 		seq_printf(m, ",uid=%u",
-			   from_kuid_munged(&init_user_ns, opts->uid));
-	if (!gid_eq(opts->gid, GLOBAL_ROOT_GID))
+			   from_kuid_munged(&init_user_ns, fsi->uid));
+	if (!gid_eq(fsi->gid, GLOBAL_ROOT_GID))
 		seq_printf(m, ",gid=%u",
-			   from_kgid_munged(&init_user_ns, opts->gid));
-	if (opts->mode != TRACEFS_DEFAULT_MODE)
-		seq_printf(m, ",mode=%o", opts->mode);
+			   from_kgid_munged(&init_user_ns, fsi->gid));
+	if (fsi->mode != TRACEFS_DEFAULT_MODE)
+		seq_printf(m, ",mode=%o", fsi->mode);
 
 	return 0;
 }
 
+static int tracefs_drop_inode(struct inode *inode)
+{
+	struct tracefs_inode *ti = get_tracefs(inode);
+
+	/*
+	 * This inode is being freed and cannot be used for
+	 * eventfs. Clear the flag so that it doesn't call into
+	 * eventfs during the remount flag updates. The eventfs_inode
+	 * gets freed after an RCU cycle, so the content will still
+	 * be safe if the iteration is going on now.
+	 */
+	ti->flags &= ~TRACEFS_EVENT_INODE;
+
+	return 1;
+}
+
 static const struct super_operations tracefs_super_operations = {
 	.alloc_inode    = tracefs_alloc_inode,
 	.free_inode     = tracefs_free_inode,
-	.drop_inode     = generic_delete_inode,
+	.destroy_inode  = tracefs_destroy_inode,
+	.drop_inode     = tracefs_drop_inode,
 	.statfs		= simple_statfs,
-	.remount_fs	= tracefs_remount,
 	.show_options	= tracefs_show_options,
 };
 
@@ -469,46 +469,19 @@ static int tracefs_d_revalidate(struct dentry *dentry, unsigned int flags)
 	return !(ei && ei->is_freed);
 }
 
-static void tracefs_d_iput(struct dentry *dentry, struct inode *inode)
-{
-	struct tracefs_inode *ti = get_tracefs(inode);
-
-	/*
-	 * This inode is being freed and cannot be used for
-	 * eventfs. Clear the flag so that it doesn't call into
-	 * eventfs during the remount flag updates. The eventfs_inode
-	 * gets freed after an RCU cycle, so the content will still
-	 * be safe if the iteration is going on now.
-	 */
-	ti->flags &= ~TRACEFS_EVENT_INODE;
-}
-
 static const struct dentry_operations tracefs_dentry_operations = {
-	.d_iput = tracefs_d_iput,
 	.d_revalidate = tracefs_d_revalidate,
 	.d_release = tracefs_d_release,
 };
 
-static int trace_fill_super(struct super_block *sb, void *data, int silent)
+static int tracefs_fill_super(struct super_block *sb, struct fs_context *fc)
 {
 	static const struct tree_descr trace_files[] = {{""}};
-	struct tracefs_fs_info *fsi;
 	int err;
 
-	fsi = kzalloc(sizeof(struct tracefs_fs_info), GFP_KERNEL);
-	sb->s_fs_info = fsi;
-	if (!fsi) {
-		err = -ENOMEM;
-		goto fail;
-	}
-
-	err = tracefs_parse_options(data, &fsi->mount_opts);
-	if (err)
-		goto fail;
-
-	err  =  simple_fill_super(sb, TRACEFS_MAGIC, trace_files);
+	err = simple_fill_super(sb, TRACEFS_MAGIC, trace_files);
 	if (err)
-		goto fail;
+		return err;
 
 	sb->s_op = &tracefs_super_operations;
 	sb->s_d_op = &tracefs_dentry_operations;
@@ -516,24 +489,45 @@ static int trace_fill_super(struct super_block *sb, void *data, int silent)
 	tracefs_apply_options(sb, false);
 
 	return 0;
+}
 
-fail:
-	kfree(fsi);
-	sb->s_fs_info = NULL;
-	return err;
+static int tracefs_get_tree(struct fs_context *fc)
+{
+	return get_tree_single(fc, tracefs_fill_super);
 }
 
-static struct dentry *trace_mount(struct file_system_type *fs_type,
-			int flags, const char *dev_name,
-			void *data)
+static void tracefs_free_fc(struct fs_context *fc)
 {
-	return mount_single(fs_type, flags, data, trace_fill_super);
+	kfree(fc->s_fs_info);
+}
+
+static const struct fs_context_operations tracefs_context_ops = {
+	.free		= tracefs_free_fc,
+	.parse_param	= tracefs_parse_param,
+	.get_tree	= tracefs_get_tree,
+	.reconfigure	= tracefs_reconfigure,
+};
+
+static int tracefs_init_fs_context(struct fs_context *fc)
+{
+	struct tracefs_fs_info *fsi;
+
+	fsi = kzalloc(sizeof(struct tracefs_fs_info), GFP_KERNEL);
+	if (!fsi)
+		return -ENOMEM;
+
+	fsi->mode = TRACEFS_DEFAULT_MODE;
+
+	fc->s_fs_info = fsi;
+	fc->ops = &tracefs_context_ops;
+	return 0;
 }
 
 static struct file_system_type trace_fs_type = {
 	.owner =	THIS_MODULE,
 	.name =		"tracefs",
-	.mount =	trace_mount,
+	.init_fs_context = tracefs_init_fs_context,
+	.parameters	= tracefs_param_specs,
 	.kill_sb =	kill_litter_super,
 };
 MODULE_ALIAS_FS("tracefs");
@@ -817,7 +811,6 @@ static int __init tracefs_init(void)
 	tracefs_inode_cachep = kmem_cache_create("tracefs_inode_cache",
 						 sizeof(struct tracefs_inode),
 						 0, (SLAB_RECLAIM_ACCOUNT|
-						     SLAB_MEM_SPREAD|
 						     SLAB_ACCOUNT),
 						 init_once);
 	if (!tracefs_inode_cachep)
diff --git a/fs/tracefs/internal.h b/fs/tracefs/internal.h
index 824cbe8367..d83c2a25f2 100644
--- a/fs/tracefs/internal.h
+++ b/fs/tracefs/internal.h
@@ -4,17 +4,13 @@
 
 enum {
 	TRACEFS_EVENT_INODE		= BIT(1),
-	TRACEFS_EVENT_TOP_INODE		= BIT(2),
-	TRACEFS_GID_PERM_SET		= BIT(3),
-	TRACEFS_UID_PERM_SET		= BIT(4),
-	TRACEFS_INSTANCE_INODE		= BIT(5),
+	TRACEFS_GID_PERM_SET		= BIT(2),
+	TRACEFS_UID_PERM_SET		= BIT(3),
+	TRACEFS_INSTANCE_INODE		= BIT(4),
 };
 
 struct tracefs_inode {
-	union {
-		struct inode            vfs_inode;
-		struct rcu_head		rcu;
-	};
+	struct inode            vfs_inode;
 	/* The below gets initialized with memset_after(ti, 0, vfs_inode) */
 	struct list_head	list;
 	unsigned long           flags;
@@ -40,7 +36,6 @@ struct eventfs_attr {
  * @children:	link list into the child eventfs_inode
  * @entries:	the array of entries representing the files in the directory
  * @name:	the name of the directory to create
- * @events_dir: the dentry of the events directory
  * @entry_attrs: Saved mode and ownership of the @d_children
  * @data:	The private data to pass to the callbacks
  * @attr:	Saved mode and ownership of eventfs_inode itself
@@ -58,7 +53,6 @@ struct eventfs_inode {
 	struct list_head		children;
 	const struct eventfs_entry	*entries;
 	const char			*name;
-	struct dentry			*events_dir;
 	struct eventfs_attr		*entry_attrs;
 	void				*data;
 	struct eventfs_attr		attr;
diff --git a/fs/ubifs/debug.c b/fs/ubifs/debug.c
index d013c5b3f1..ac77ac1fd7 100644
--- a/fs/ubifs/debug.c
+++ b/fs/ubifs/debug.c
@@ -1742,17 +1742,22 @@ int dbg_check_idx_size(struct ubifs_info *c, long long idx_size)
 	err = dbg_walk_index(c, NULL, add_size, &calc);
 	if (err) {
 		ubifs_err(c, "error %d while walking the index", err);
-		return err;
+		goto out_err;
 	}
 
 	if (calc != idx_size) {
 		ubifs_err(c, "index size check failed: calculated size is %lld, should be %lld",
 			  calc, idx_size);
 		dump_stack();
-		return -EINVAL;
+		err = -EINVAL;
+		goto out_err;
 	}
 
 	return 0;
+
+out_err:
+	ubifs_destroy_tnc_tree(c);
+	return err;
 }
 
 /**
diff --git a/fs/ubifs/dir.c b/fs/ubifs/dir.c
index 6b3db00d9b..eac0fef801 100644
--- a/fs/ubifs/dir.c
+++ b/fs/ubifs/dir.c
@@ -205,7 +205,6 @@ static struct dentry *ubifs_lookup(struct inode *dir, struct dentry *dentry,
 	dbg_gen("'%pd' in dir ino %lu", dentry, dir->i_ino);
 
 	err = fscrypt_prepare_lookup(dir, dentry, &nm);
-	generic_set_encrypted_ci_d_ops(dentry);
 	if (err == -ENOENT)
 		return d_splice_alias(NULL, dentry);
 	if (err)
diff --git a/fs/ubifs/file.c b/fs/ubifs/file.c
index d0694b83dd..a1f4691993 100644
--- a/fs/ubifs/file.c
+++ b/fs/ubifs/file.c
@@ -96,36 +96,36 @@ dump:
 	return -EINVAL;
 }
 
-static int do_readpage(struct page *page)
+static int do_readpage(struct folio *folio)
 {
 	void *addr;
 	int err = 0, i;
 	unsigned int block, beyond;
-	struct ubifs_data_node *dn;
-	struct inode *inode = page->mapping->host;
+	struct ubifs_data_node *dn = NULL;
+	struct inode *inode = folio->mapping->host;
 	struct ubifs_info *c = inode->i_sb->s_fs_info;
 	loff_t i_size = i_size_read(inode);
 
 	dbg_gen("ino %lu, pg %lu, i_size %lld, flags %#lx",
-		inode->i_ino, page->index, i_size, page->flags);
-	ubifs_assert(c, !PageChecked(page));
-	ubifs_assert(c, !PagePrivate(page));
+		inode->i_ino, folio->index, i_size, folio->flags);
+	ubifs_assert(c, !folio_test_checked(folio));
+	ubifs_assert(c, !folio->private);
 
-	addr = kmap(page);
+	addr = kmap_local_folio(folio, 0);
 
-	block = page->index << UBIFS_BLOCKS_PER_PAGE_SHIFT;
+	block = folio->index << UBIFS_BLOCKS_PER_PAGE_SHIFT;
 	beyond = (i_size + UBIFS_BLOCK_SIZE - 1) >> UBIFS_BLOCK_SHIFT;
 	if (block >= beyond) {
 		/* Reading beyond inode */
-		SetPageChecked(page);
-		memset(addr, 0, PAGE_SIZE);
+		folio_set_checked(folio);
+		addr = folio_zero_tail(folio, 0, addr);
 		goto out;
 	}
 
 	dn = kmalloc(UBIFS_MAX_DATA_NODE_SZ, GFP_NOFS);
 	if (!dn) {
 		err = -ENOMEM;
-		goto error;
+		goto out;
 	}
 
 	i = 0;
@@ -150,39 +150,35 @@ static int do_readpage(struct page *page)
 					memset(addr + ilen, 0, dlen - ilen);
 			}
 		}
-		if (++i >= UBIFS_BLOCKS_PER_PAGE)
+		if (++i >= (UBIFS_BLOCKS_PER_PAGE << folio_order(folio)))
 			break;
 		block += 1;
 		addr += UBIFS_BLOCK_SIZE;
+		if (folio_test_highmem(folio) && (offset_in_page(addr) == 0)) {
+			kunmap_local(addr - UBIFS_BLOCK_SIZE);
+			addr = kmap_local_folio(folio, i * UBIFS_BLOCK_SIZE);
+		}
 	}
+
 	if (err) {
 		struct ubifs_info *c = inode->i_sb->s_fs_info;
 		if (err == -ENOENT) {
 			/* Not found, so it must be a hole */
-			SetPageChecked(page);
+			folio_set_checked(folio);
 			dbg_gen("hole");
-			goto out_free;
+			err = 0;
+		} else {
+			ubifs_err(c, "cannot read page %lu of inode %lu, error %d",
+				  folio->index, inode->i_ino, err);
 		}
-		ubifs_err(c, "cannot read page %lu of inode %lu, error %d",
-			  page->index, inode->i_ino, err);
-		goto error;
 	}
 
-out_free:
-	kfree(dn);
 out:
-	SetPageUptodate(page);
-	ClearPageError(page);
-	flush_dcache_page(page);
-	kunmap(page);
-	return 0;
-
-error:
 	kfree(dn);
-	ClearPageUptodate(page);
-	SetPageError(page);
-	flush_dcache_page(page);
-	kunmap(page);
+	if (!err)
+		folio_mark_uptodate(folio);
+	flush_dcache_folio(folio);
+	kunmap_local(addr);
 	return err;
 }
 
@@ -222,16 +218,16 @@ static int write_begin_slow(struct address_space *mapping,
 	pgoff_t index = pos >> PAGE_SHIFT;
 	struct ubifs_budget_req req = { .new_page = 1 };
 	int err, appending = !!(pos + len > inode->i_size);
-	struct page *page;
+	struct folio *folio;
 
 	dbg_gen("ino %lu, pos %llu, len %u, i_size %lld",
 		inode->i_ino, pos, len, inode->i_size);
 
 	/*
-	 * At the slow path we have to budget before locking the page, because
-	 * budgeting may force write-back, which would wait on locked pages and
-	 * deadlock if we had the page locked. At this point we do not know
-	 * anything about the page, so assume that this is a new page which is
+	 * At the slow path we have to budget before locking the folio, because
+	 * budgeting may force write-back, which would wait on locked folios and
+	 * deadlock if we had the folio locked. At this point we do not know
+	 * anything about the folio, so assume that this is a new folio which is
 	 * written to a hole. This corresponds to largest budget. Later the
 	 * budget will be amended if this is not true.
 	 */
@@ -243,42 +239,43 @@ static int write_begin_slow(struct address_space *mapping,
 	if (unlikely(err))
 		return err;
 
-	page = grab_cache_page_write_begin(mapping, index);
-	if (unlikely(!page)) {
+	folio = __filemap_get_folio(mapping, index, FGP_WRITEBEGIN,
+			mapping_gfp_mask(mapping));
+	if (IS_ERR(folio)) {
 		ubifs_release_budget(c, &req);
-		return -ENOMEM;
+		return PTR_ERR(folio);
 	}
 
-	if (!PageUptodate(page)) {
-		if (!(pos & ~PAGE_MASK) && len == PAGE_SIZE)
-			SetPageChecked(page);
+	if (!folio_test_uptodate(folio)) {
+		if (pos == folio_pos(folio) && len >= folio_size(folio))
+			folio_set_checked(folio);
 		else {
-			err = do_readpage(page);
+			err = do_readpage(folio);
 			if (err) {
-				unlock_page(page);
-				put_page(page);
+				folio_unlock(folio);
+				folio_put(folio);
 				ubifs_release_budget(c, &req);
 				return err;
 			}
 		}
 	}
 
-	if (PagePrivate(page))
+	if (folio->private)
 		/*
-		 * The page is dirty, which means it was budgeted twice:
+		 * The folio is dirty, which means it was budgeted twice:
 		 *   o first time the budget was allocated by the task which
-		 *     made the page dirty and set the PG_private flag;
+		 *     made the folio dirty and set the private field;
 		 *   o and then we budgeted for it for the second time at the
 		 *     very beginning of this function.
 		 *
-		 * So what we have to do is to release the page budget we
+		 * So what we have to do is to release the folio budget we
 		 * allocated.
 		 */
 		release_new_page_budget(c);
-	else if (!PageChecked(page))
+	else if (!folio_test_checked(folio))
 		/*
-		 * We are changing a page which already exists on the media.
-		 * This means that changing the page does not make the amount
+		 * We are changing a folio which already exists on the media.
+		 * This means that changing the folio does not make the amount
 		 * of indexing information larger, and this part of the budget
 		 * which we have already acquired may be released.
 		 */
@@ -301,14 +298,14 @@ static int write_begin_slow(struct address_space *mapping,
 			ubifs_release_dirty_inode_budget(c, ui);
 	}
 
-	*pagep = page;
+	*pagep = &folio->page;
 	return 0;
 }
 
 /**
  * allocate_budget - allocate budget for 'ubifs_write_begin()'.
  * @c: UBIFS file-system description object
- * @page: page to allocate budget for
+ * @folio: folio to allocate budget for
  * @ui: UBIFS inode object the page belongs to
  * @appending: non-zero if the page is appended
  *
@@ -319,15 +316,15 @@ static int write_begin_slow(struct address_space *mapping,
  *
  * Returns: %0 in case of success and %-ENOSPC in case of failure.
  */
-static int allocate_budget(struct ubifs_info *c, struct page *page,
+static int allocate_budget(struct ubifs_info *c, struct folio *folio,
 			   struct ubifs_inode *ui, int appending)
 {
 	struct ubifs_budget_req req = { .fast = 1 };
 
-	if (PagePrivate(page)) {
+	if (folio->private) {
 		if (!appending)
 			/*
-			 * The page is dirty and we are not appending, which
+			 * The folio is dirty and we are not appending, which
 			 * means no budget is needed at all.
 			 */
 			return 0;
@@ -351,11 +348,11 @@ static int allocate_budget(struct ubifs_info *c, struct page *page,
 		 */
 		req.dirtied_ino = 1;
 	} else {
-		if (PageChecked(page))
+		if (folio_test_checked(folio))
 			/*
 			 * The page corresponds to a hole and does not
 			 * exist on the media. So changing it makes
-			 * make the amount of indexing information
+			 * the amount of indexing information
 			 * larger, and we have to budget for a new
 			 * page.
 			 */
@@ -425,7 +422,7 @@ static int ubifs_write_begin(struct file *file, struct address_space *mapping,
 	pgoff_t index = pos >> PAGE_SHIFT;
 	int err, appending = !!(pos + len > inode->i_size);
 	int skipped_read = 0;
-	struct page *page;
+	struct folio *folio;
 
 	ubifs_assert(c, ubifs_inode(inode)->ui_size == inode->i_size);
 	ubifs_assert(c, !c->ro_media && !c->ro_mount);
@@ -434,13 +431,14 @@ static int ubifs_write_begin(struct file *file, struct address_space *mapping,
 		return -EROFS;
 
 	/* Try out the fast-path part first */
-	page = grab_cache_page_write_begin(mapping, index);
-	if (unlikely(!page))
-		return -ENOMEM;
+	folio = __filemap_get_folio(mapping, index, FGP_WRITEBEGIN,
+			mapping_gfp_mask(mapping));
+	if (IS_ERR(folio))
+		return PTR_ERR(folio);
 
-	if (!PageUptodate(page)) {
+	if (!folio_test_uptodate(folio)) {
 		/* The page is not loaded from the flash */
-		if (!(pos & ~PAGE_MASK) && len == PAGE_SIZE) {
+		if (pos == folio_pos(folio) && len >= folio_size(folio)) {
 			/*
 			 * We change whole page so no need to load it. But we
 			 * do not know whether this page exists on the media or
@@ -450,19 +448,19 @@ static int ubifs_write_begin(struct file *file, struct address_space *mapping,
 			 * media. Thus, we are setting the @PG_checked flag
 			 * here.
 			 */
-			SetPageChecked(page);
+			folio_set_checked(folio);
 			skipped_read = 1;
 		} else {
-			err = do_readpage(page);
+			err = do_readpage(folio);
 			if (err) {
-				unlock_page(page);
-				put_page(page);
+				folio_unlock(folio);
+				folio_put(folio);
 				return err;
 			}
 		}
 	}
 
-	err = allocate_budget(c, page, ui, appending);
+	err = allocate_budget(c, folio, ui, appending);
 	if (unlikely(err)) {
 		ubifs_assert(c, err == -ENOSPC);
 		/*
@@ -470,7 +468,7 @@ static int ubifs_write_begin(struct file *file, struct address_space *mapping,
 		 * write all of it, then it is not up to date.
 		 */
 		if (skipped_read)
-			ClearPageChecked(page);
+			folio_clear_checked(folio);
 		/*
 		 * Budgeting failed which means it would have to force
 		 * write-back but didn't, because we set the @fast flag in the
@@ -482,8 +480,8 @@ static int ubifs_write_begin(struct file *file, struct address_space *mapping,
 			ubifs_assert(c, mutex_is_locked(&ui->ui_mutex));
 			mutex_unlock(&ui->ui_mutex);
 		}
-		unlock_page(page);
-		put_page(page);
+		folio_unlock(folio);
+		folio_put(folio);
 
 		return write_begin_slow(mapping, pos, len, pagep);
 	}
@@ -494,22 +492,21 @@ static int ubifs_write_begin(struct file *file, struct address_space *mapping,
 	 * with @ui->ui_mutex locked if we are appending pages, and unlocked
 	 * otherwise. This is an optimization (slightly hacky though).
 	 */
-	*pagep = page;
+	*pagep = &folio->page;
 	return 0;
-
 }
 
 /**
  * cancel_budget - cancel budget.
  * @c: UBIFS file-system description object
- * @page: page to cancel budget for
+ * @folio: folio to cancel budget for
  * @ui: UBIFS inode object the page belongs to
  * @appending: non-zero if the page is appended
  *
  * This is a helper function for a page write operation. It unlocks the
  * @ui->ui_mutex in case of appending.
  */
-static void cancel_budget(struct ubifs_info *c, struct page *page,
+static void cancel_budget(struct ubifs_info *c, struct folio *folio,
 			  struct ubifs_inode *ui, int appending)
 {
 	if (appending) {
@@ -517,8 +514,8 @@ static void cancel_budget(struct ubifs_info *c, struct page *page,
 			ubifs_release_dirty_inode_budget(c, ui);
 		mutex_unlock(&ui->ui_mutex);
 	}
-	if (!PagePrivate(page)) {
-		if (PageChecked(page))
+	if (!folio->private) {
+		if (folio_test_checked(folio))
 			release_new_page_budget(c);
 		else
 			release_existing_page_budget(c);
@@ -529,6 +526,7 @@ static int ubifs_write_end(struct file *file, struct address_space *mapping,
 			   loff_t pos, unsigned len, unsigned copied,
 			   struct page *page, void *fsdata)
 {
+	struct folio *folio = page_folio(page);
 	struct inode *inode = mapping->host;
 	struct ubifs_inode *ui = ubifs_inode(inode);
 	struct ubifs_info *c = inode->i_sb->s_fs_info;
@@ -536,47 +534,47 @@ static int ubifs_write_end(struct file *file, struct address_space *mapping,
 	int appending = !!(end_pos > inode->i_size);
 
 	dbg_gen("ino %lu, pos %llu, pg %lu, len %u, copied %d, i_size %lld",
-		inode->i_ino, pos, page->index, len, copied, inode->i_size);
+		inode->i_ino, pos, folio->index, len, copied, inode->i_size);
 
-	if (unlikely(copied < len && len == PAGE_SIZE)) {
+	if (unlikely(copied < len && !folio_test_uptodate(folio))) {
 		/*
-		 * VFS copied less data to the page that it intended and
+		 * VFS copied less data to the folio than it intended and
 		 * declared in its '->write_begin()' call via the @len
-		 * argument. If the page was not up-to-date, and @len was
-		 * @PAGE_SIZE, the 'ubifs_write_begin()' function did
+		 * argument. If the folio was not up-to-date,
+		 * the 'ubifs_write_begin()' function did
 		 * not load it from the media (for optimization reasons). This
-		 * means that part of the page contains garbage. So read the
-		 * page now.
+		 * means that part of the folio contains garbage. So read the
+		 * folio now.
 		 */
 		dbg_gen("copied %d instead of %d, read page and repeat",
 			copied, len);
-		cancel_budget(c, page, ui, appending);
-		ClearPageChecked(page);
+		cancel_budget(c, folio, ui, appending);
+		folio_clear_checked(folio);
 
 		/*
 		 * Return 0 to force VFS to repeat the whole operation, or the
 		 * error code if 'do_readpage()' fails.
 		 */
-		copied = do_readpage(page);
+		copied = do_readpage(folio);
 		goto out;
 	}
 
-	if (len == PAGE_SIZE)
-		SetPageUptodate(page);
+	if (len == folio_size(folio))
+		folio_mark_uptodate(folio);
 
-	if (!PagePrivate(page)) {
-		attach_page_private(page, (void *)1);
+	if (!folio->private) {
+		folio_attach_private(folio, (void *)1);
 		atomic_long_inc(&c->dirty_pg_cnt);
-		__set_page_dirty_nobuffers(page);
+		filemap_dirty_folio(mapping, folio);
 	}
 
 	if (appending) {
 		i_size_write(inode, end_pos);
 		ui->ui_size = end_pos;
 		/*
-		 * Note, we do not set @I_DIRTY_PAGES (which means that the
-		 * inode has dirty pages), this has been done in
-		 * '__set_page_dirty_nobuffers()'.
+		 * We do not set @I_DIRTY_PAGES (which means that
+		 * the inode has dirty pages), this was done in
+		 * filemap_dirty_folio().
 		 */
 		__mark_inode_dirty(inode, I_DIRTY_DATASYNC);
 		ubifs_assert(c, mutex_is_locked(&ui->ui_mutex));
@@ -584,43 +582,43 @@ static int ubifs_write_end(struct file *file, struct address_space *mapping,
 	}
 
 out:
-	unlock_page(page);
-	put_page(page);
+	folio_unlock(folio);
+	folio_put(folio);
 	return copied;
 }
 
 /**
  * populate_page - copy data nodes into a page for bulk-read.
  * @c: UBIFS file-system description object
- * @page: page
+ * @folio: folio
  * @bu: bulk-read information
  * @n: next zbranch slot
  *
  * Returns: %0 on success and a negative error code on failure.
  */
-static int populate_page(struct ubifs_info *c, struct page *page,
+static int populate_page(struct ubifs_info *c, struct folio *folio,
 			 struct bu_info *bu, int *n)
 {
 	int i = 0, nn = *n, offs = bu->zbranch[0].offs, hole = 0, read = 0;
-	struct inode *inode = page->mapping->host;
+	struct inode *inode = folio->mapping->host;
 	loff_t i_size = i_size_read(inode);
 	unsigned int page_block;
 	void *addr, *zaddr;
 	pgoff_t end_index;
 
 	dbg_gen("ino %lu, pg %lu, i_size %lld, flags %#lx",
-		inode->i_ino, page->index, i_size, page->flags);
+		inode->i_ino, folio->index, i_size, folio->flags);
 
-	addr = zaddr = kmap(page);
+	addr = zaddr = kmap_local_folio(folio, 0);
 
 	end_index = (i_size - 1) >> PAGE_SHIFT;
-	if (!i_size || page->index > end_index) {
+	if (!i_size || folio->index > end_index) {
 		hole = 1;
-		memset(addr, 0, PAGE_SIZE);
+		addr = folio_zero_tail(folio, 0, addr);
 		goto out_hole;
 	}
 
-	page_block = page->index << UBIFS_BLOCKS_PER_PAGE_SHIFT;
+	page_block = folio->index << UBIFS_BLOCKS_PER_PAGE_SHIFT;
 	while (1) {
 		int err, len, out_len, dlen;
 
@@ -669,9 +667,13 @@ static int populate_page(struct ubifs_info *c, struct page *page,
 			break;
 		addr += UBIFS_BLOCK_SIZE;
 		page_block += 1;
+		if (folio_test_highmem(folio) && (offset_in_page(addr) == 0)) {
+			kunmap_local(addr - UBIFS_BLOCK_SIZE);
+			addr = kmap_local_folio(folio, i * UBIFS_BLOCK_SIZE);
+		}
 	}
 
-	if (end_index == page->index) {
+	if (end_index == folio->index) {
 		int len = i_size & (PAGE_SIZE - 1);
 
 		if (len && len < read)
@@ -680,22 +682,19 @@ static int populate_page(struct ubifs_info *c, struct page *page,
 
 out_hole:
 	if (hole) {
-		SetPageChecked(page);
+		folio_set_checked(folio);
 		dbg_gen("hole");
 	}
 
-	SetPageUptodate(page);
-	ClearPageError(page);
-	flush_dcache_page(page);
-	kunmap(page);
+	folio_mark_uptodate(folio);
+	flush_dcache_folio(folio);
+	kunmap_local(addr);
 	*n = nn;
 	return 0;
 
 out_err:
-	ClearPageUptodate(page);
-	SetPageError(page);
-	flush_dcache_page(page);
-	kunmap(page);
+	flush_dcache_folio(folio);
+	kunmap_local(addr);
 	ubifs_err(c, "bad data node (block %u, inode %lu)",
 		  page_block, inode->i_ino);
 	return -EINVAL;
@@ -705,15 +704,15 @@ out_err:
  * ubifs_do_bulk_read - do bulk-read.
  * @c: UBIFS file-system description object
  * @bu: bulk-read information
- * @page1: first page to read
+ * @folio1: first folio to read
  *
  * Returns: %1 if the bulk-read is done, otherwise %0 is returned.
  */
 static int ubifs_do_bulk_read(struct ubifs_info *c, struct bu_info *bu,
-			      struct page *page1)
+			      struct folio *folio1)
 {
-	pgoff_t offset = page1->index, end_index;
-	struct address_space *mapping = page1->mapping;
+	pgoff_t offset = folio1->index, end_index;
+	struct address_space *mapping = folio1->mapping;
 	struct inode *inode = mapping->host;
 	struct ubifs_inode *ui = ubifs_inode(inode);
 	int err, page_idx, page_cnt, ret = 0, n = 0;
@@ -763,11 +762,11 @@ static int ubifs_do_bulk_read(struct ubifs_info *c, struct bu_info *bu,
 			goto out_warn;
 	}
 
-	err = populate_page(c, page1, bu, &n);
+	err = populate_page(c, folio1, bu, &n);
 	if (err)
 		goto out_warn;
 
-	unlock_page(page1);
+	folio_unlock(folio1);
 	ret = 1;
 
 	isize = i_size_read(inode);
@@ -777,19 +776,19 @@ static int ubifs_do_bulk_read(struct ubifs_info *c, struct bu_info *bu,
 
 	for (page_idx = 1; page_idx < page_cnt; page_idx++) {
 		pgoff_t page_offset = offset + page_idx;
-		struct page *page;
+		struct folio *folio;
 
 		if (page_offset > end_index)
 			break;
-		page = pagecache_get_page(mapping, page_offset,
+		folio = __filemap_get_folio(mapping, page_offset,
 				 FGP_LOCK|FGP_ACCESSED|FGP_CREAT|FGP_NOWAIT,
 				 ra_gfp_mask);
-		if (!page)
+		if (IS_ERR(folio))
 			break;
-		if (!PageUptodate(page))
-			err = populate_page(c, page, bu, &n);
-		unlock_page(page);
-		put_page(page);
+		if (!folio_test_uptodate(folio))
+			err = populate_page(c, folio, bu, &n);
+		folio_unlock(folio);
+		folio_put(folio);
 		if (err)
 			break;
 	}
@@ -812,7 +811,7 @@ out_bu_off:
 
 /**
  * ubifs_bulk_read - determine whether to bulk-read and, if so, do it.
- * @page: page from which to start bulk-read.
+ * @folio: folio from which to start bulk-read.
  *
  * Some flash media are capable of reading sequentially at faster rates. UBIFS
  * bulk-read facility is designed to take advantage of that, by reading in one
@@ -821,12 +820,12 @@ out_bu_off:
  *
  * Returns: %1 if a bulk-read is done and %0 otherwise.
  */
-static int ubifs_bulk_read(struct page *page)
+static int ubifs_bulk_read(struct folio *folio)
 {
-	struct inode *inode = page->mapping->host;
+	struct inode *inode = folio->mapping->host;
 	struct ubifs_info *c = inode->i_sb->s_fs_info;
 	struct ubifs_inode *ui = ubifs_inode(inode);
-	pgoff_t index = page->index, last_page_read = ui->last_page_read;
+	pgoff_t index = folio->index, last_page_read = ui->last_page_read;
 	struct bu_info *bu;
 	int err = 0, allocated = 0;
 
@@ -874,8 +873,8 @@ static int ubifs_bulk_read(struct page *page)
 
 	bu->buf_len = c->max_bu_buf_len;
 	data_key_init(c, &bu->key, inode->i_ino,
-		      page->index << UBIFS_BLOCKS_PER_PAGE_SHIFT);
-	err = ubifs_do_bulk_read(c, bu, page);
+		      folio->index << UBIFS_BLOCKS_PER_PAGE_SHIFT);
+	err = ubifs_do_bulk_read(c, bu, folio);
 
 	if (!allocated)
 		mutex_unlock(&c->bu_mutex);
@@ -889,69 +888,71 @@ out_unlock:
 
 static int ubifs_read_folio(struct file *file, struct folio *folio)
 {
-	struct page *page = &folio->page;
-
-	if (ubifs_bulk_read(page))
+	if (ubifs_bulk_read(folio))
 		return 0;
-	do_readpage(page);
+	do_readpage(folio);
 	folio_unlock(folio);
 	return 0;
 }
 
-static int do_writepage(struct page *page, int len)
+static int do_writepage(struct folio *folio, size_t len)
 {
-	int err = 0, i, blen;
+	int err = 0, blen;
 	unsigned int block;
 	void *addr;
+	size_t offset = 0;
 	union ubifs_key key;
-	struct inode *inode = page->mapping->host;
+	struct inode *inode = folio->mapping->host;
 	struct ubifs_info *c = inode->i_sb->s_fs_info;
 
 #ifdef UBIFS_DEBUG
 	struct ubifs_inode *ui = ubifs_inode(inode);
 	spin_lock(&ui->ui_lock);
-	ubifs_assert(c, page->index <= ui->synced_i_size >> PAGE_SHIFT);
+	ubifs_assert(c, folio->index <= ui->synced_i_size >> PAGE_SHIFT);
 	spin_unlock(&ui->ui_lock);
 #endif
 
-	/* Update radix tree tags */
-	set_page_writeback(page);
+	folio_start_writeback(folio);
 
-	addr = kmap(page);
-	block = page->index << UBIFS_BLOCKS_PER_PAGE_SHIFT;
-	i = 0;
-	while (len) {
-		blen = min_t(int, len, UBIFS_BLOCK_SIZE);
+	addr = kmap_local_folio(folio, offset);
+	block = folio->index << UBIFS_BLOCKS_PER_PAGE_SHIFT;
+	for (;;) {
+		blen = min_t(size_t, len, UBIFS_BLOCK_SIZE);
 		data_key_init(c, &key, inode->i_ino, block);
 		err = ubifs_jnl_write_data(c, inode, &key, addr, blen);
 		if (err)
 			break;
-		if (++i >= UBIFS_BLOCKS_PER_PAGE)
+		len -= blen;
+		if (!len)
 			break;
 		block += 1;
 		addr += blen;
-		len -= blen;
+		if (folio_test_highmem(folio) && !offset_in_page(addr)) {
+			kunmap_local(addr - blen);
+			offset += PAGE_SIZE;
+			addr = kmap_local_folio(folio, offset);
+		}
 	}
+	kunmap_local(addr);
 	if (err) {
-		SetPageError(page);
-		ubifs_err(c, "cannot write page %lu of inode %lu, error %d",
-			  page->index, inode->i_ino, err);
+		mapping_set_error(folio->mapping, err);
+		ubifs_err(c, "cannot write folio %lu of inode %lu, error %d",
+			  folio->index, inode->i_ino, err);
 		ubifs_ro_mode(c, err);
 	}
 
-	ubifs_assert(c, PagePrivate(page));
-	if (PageChecked(page))
+	ubifs_assert(c, folio->private != NULL);
+	if (folio_test_checked(folio))
 		release_new_page_budget(c);
 	else
 		release_existing_page_budget(c);
 
 	atomic_long_dec(&c->dirty_pg_cnt);
-	detach_page_private(page);
-	ClearPageChecked(page);
+	folio_detach_private(folio);
+	folio_clear_checked(folio);
 
-	kunmap(page);
-	unlock_page(page);
-	end_page_writeback(page);
+	folio_unlock(folio);
+	folio_end_writeback(folio);
 	return err;
 }
 
@@ -1001,22 +1002,21 @@ static int do_writepage(struct page *page, int len)
  * on the page lock and it would not write the truncated inode node to the
  * journal before we have finished.
  */
-static int ubifs_writepage(struct page *page, struct writeback_control *wbc)
+static int ubifs_writepage(struct folio *folio, struct writeback_control *wbc,
+		void *data)
 {
-	struct inode *inode = page->mapping->host;
+	struct inode *inode = folio->mapping->host;
 	struct ubifs_info *c = inode->i_sb->s_fs_info;
 	struct ubifs_inode *ui = ubifs_inode(inode);
 	loff_t i_size =  i_size_read(inode), synced_i_size;
-	pgoff_t end_index = i_size >> PAGE_SHIFT;
-	int err, len = i_size & (PAGE_SIZE - 1);
-	void *kaddr;
+	int err, len = folio_size(folio);
 
 	dbg_gen("ino %lu, pg %lu, pg flags %#lx",
-		inode->i_ino, page->index, page->flags);
-	ubifs_assert(c, PagePrivate(page));
+		inode->i_ino, folio->index, folio->flags);
+	ubifs_assert(c, folio->private != NULL);
 
-	/* Is the page fully outside @i_size? (truncate in progress) */
-	if (page->index > end_index || (page->index == end_index && !len)) {
+	/* Is the folio fully outside @i_size? (truncate in progress) */
+	if (folio_pos(folio) >= i_size) {
 		err = 0;
 		goto out_unlock;
 	}
@@ -1025,9 +1025,9 @@ static int ubifs_writepage(struct page *page, struct writeback_control *wbc)
 	synced_i_size = ui->synced_i_size;
 	spin_unlock(&ui->ui_lock);
 
-	/* Is the page fully inside @i_size? */
-	if (page->index < end_index) {
-		if (page->index >= synced_i_size >> PAGE_SHIFT) {
+	/* Is the folio fully inside i_size? */
+	if (folio_pos(folio) + len <= i_size) {
+		if (folio_pos(folio) >= synced_i_size) {
 			err = inode->i_sb->s_op->write_inode(inode, NULL);
 			if (err)
 				goto out_redirty;
@@ -1040,20 +1040,18 @@ static int ubifs_writepage(struct page *page, struct writeback_control *wbc)
 			 * with this.
 			 */
 		}
-		return do_writepage(page, PAGE_SIZE);
+		return do_writepage(folio, len);
 	}
 
 	/*
-	 * The page straddles @i_size. It must be zeroed out on each and every
+	 * The folio straddles @i_size. It must be zeroed out on each and every
 	 * writepage invocation because it may be mmapped. "A file is mapped
 	 * in multiples of the page size. For a file that is not a multiple of
 	 * the page size, the remaining memory is zeroed when mapped, and
 	 * writes to that region are not written out to the file."
 	 */
-	kaddr = kmap_atomic(page);
-	memset(kaddr + len, 0, PAGE_SIZE - len);
-	flush_dcache_page(page);
-	kunmap_atomic(kaddr);
+	len = i_size - folio_pos(folio);
+	folio_zero_segment(folio, len, folio_size(folio));
 
 	if (i_size > synced_i_size) {
 		err = inode->i_sb->s_op->write_inode(inode, NULL);
@@ -1061,19 +1059,25 @@ static int ubifs_writepage(struct page *page, struct writeback_control *wbc)
 			goto out_redirty;
 	}
 
-	return do_writepage(page, len);
+	return do_writepage(folio, len);
 out_redirty:
 	/*
-	 * redirty_page_for_writepage() won't call ubifs_dirty_inode() because
+	 * folio_redirty_for_writepage() won't call ubifs_dirty_inode() because
 	 * it passes I_DIRTY_PAGES flag while calling __mark_inode_dirty(), so
 	 * there is no need to do space budget for dirty inode.
 	 */
-	redirty_page_for_writepage(wbc, page);
+	folio_redirty_for_writepage(wbc, folio);
 out_unlock:
-	unlock_page(page);
+	folio_unlock(folio);
 	return err;
 }
 
+static int ubifs_writepages(struct address_space *mapping,
+		struct writeback_control *wbc)
+{
+	return write_cache_pages(mapping, wbc, ubifs_writepage, NULL);
+}
+
 /**
  * do_attr_changes - change inode attributes.
  * @inode: inode to change attributes for
@@ -1150,11 +1154,11 @@ static int do_truncation(struct ubifs_info *c, struct inode *inode,
 
 	if (offset) {
 		pgoff_t index = new_size >> PAGE_SHIFT;
-		struct page *page;
+		struct folio *folio;
 
-		page = find_lock_page(inode->i_mapping, index);
-		if (page) {
-			if (PageDirty(page)) {
+		folio = filemap_lock_folio(inode->i_mapping, index);
+		if (!IS_ERR(folio)) {
+			if (folio_test_dirty(folio)) {
 				/*
 				 * 'ubifs_jnl_truncate()' will try to truncate
 				 * the last data node, but it contains
@@ -1163,14 +1167,14 @@ static int do_truncation(struct ubifs_info *c, struct inode *inode,
 				 * 'ubifs_jnl_truncate()' will see an already
 				 * truncated (and up to date) data node.
 				 */
-				ubifs_assert(c, PagePrivate(page));
+				ubifs_assert(c, folio->private != NULL);
 
-				clear_page_dirty_for_io(page);
+				folio_clear_dirty_for_io(folio);
 				if (UBIFS_BLOCKS_PER_PAGE_SHIFT)
-					offset = new_size &
-						 (PAGE_SIZE - 1);
-				err = do_writepage(page, offset);
-				put_page(page);
+					offset = offset_in_folio(folio,
+							new_size);
+				err = do_writepage(folio, offset);
+				folio_put(folio);
 				if (err)
 					goto out_budg;
 				/*
@@ -1183,8 +1187,8 @@ static int do_truncation(struct ubifs_info *c, struct inode *inode,
 				 * to 'ubifs_jnl_truncate()' to save it from
 				 * having to read it.
 				 */
-				unlock_page(page);
-				put_page(page);
+				folio_unlock(folio);
+				folio_put(folio);
 			}
 		}
 	}
@@ -1507,14 +1511,14 @@ static bool ubifs_release_folio(struct folio *folio, gfp_t unused_gfp_flags)
  */
 static vm_fault_t ubifs_vm_page_mkwrite(struct vm_fault *vmf)
 {
-	struct page *page = vmf->page;
+	struct folio *folio = page_folio(vmf->page);
 	struct inode *inode = file_inode(vmf->vma->vm_file);
 	struct ubifs_info *c = inode->i_sb->s_fs_info;
 	struct timespec64 now = current_time(inode);
 	struct ubifs_budget_req req = { .new_page = 1 };
 	int err, update_time;
 
-	dbg_gen("ino %lu, pg %lu, i_size %lld",	inode->i_ino, page->index,
+	dbg_gen("ino %lu, pg %lu, i_size %lld",	inode->i_ino, folio->index,
 		i_size_read(inode));
 	ubifs_assert(c, !c->ro_media && !c->ro_mount);
 
@@ -1522,17 +1526,17 @@ static vm_fault_t ubifs_vm_page_mkwrite(struct vm_fault *vmf)
 		return VM_FAULT_SIGBUS; /* -EROFS */
 
 	/*
-	 * We have not locked @page so far so we may budget for changing the
-	 * page. Note, we cannot do this after we locked the page, because
+	 * We have not locked @folio so far so we may budget for changing the
+	 * folio. Note, we cannot do this after we locked the folio, because
 	 * budgeting may cause write-back which would cause deadlock.
 	 *
-	 * At the moment we do not know whether the page is dirty or not, so we
-	 * assume that it is not and budget for a new page. We could look at
+	 * At the moment we do not know whether the folio is dirty or not, so we
+	 * assume that it is not and budget for a new folio. We could look at
 	 * the @PG_private flag and figure this out, but we may race with write
-	 * back and the page state may change by the time we lock it, so this
+	 * back and the folio state may change by the time we lock it, so this
 	 * would need additional care. We do not bother with this at the
 	 * moment, although it might be good idea to do. Instead, we allocate
-	 * budget for a new page and amend it later on if the page was in fact
+	 * budget for a new folio and amend it later on if the folio was in fact
 	 * dirty.
 	 *
 	 * The budgeting-related logic of this function is similar to what we
@@ -1555,21 +1559,21 @@ static vm_fault_t ubifs_vm_page_mkwrite(struct vm_fault *vmf)
 		return VM_FAULT_SIGBUS;
 	}
 
-	lock_page(page);
-	if (unlikely(page->mapping != inode->i_mapping ||
-		     page_offset(page) > i_size_read(inode))) {
-		/* Page got truncated out from underneath us */
+	folio_lock(folio);
+	if (unlikely(folio->mapping != inode->i_mapping ||
+		     folio_pos(folio) >= i_size_read(inode))) {
+		/* Folio got truncated out from underneath us */
 		goto sigbus;
 	}
 
-	if (PagePrivate(page))
+	if (folio->private)
 		release_new_page_budget(c);
 	else {
-		if (!PageChecked(page))
+		if (!folio_test_checked(folio))
 			ubifs_convert_page_budget(c);
-		attach_page_private(page, (void *)1);
+		folio_attach_private(folio, (void *)1);
 		atomic_long_inc(&c->dirty_pg_cnt);
-		__set_page_dirty_nobuffers(page);
+		filemap_dirty_folio(folio->mapping, folio);
 	}
 
 	if (update_time) {
@@ -1585,11 +1589,11 @@ static vm_fault_t ubifs_vm_page_mkwrite(struct vm_fault *vmf)
 			ubifs_release_dirty_inode_budget(c, ui);
 	}
 
-	wait_for_stable_page(page);
+	folio_wait_stable(folio);
 	return VM_FAULT_LOCKED;
 
 sigbus:
-	unlock_page(page);
+	folio_unlock(folio);
 	ubifs_release_budget(c, &req);
 	return VM_FAULT_SIGBUS;
 }
@@ -1643,7 +1647,7 @@ static int ubifs_symlink_getattr(struct mnt_idmap *idmap,
 
 const struct address_space_operations ubifs_file_address_operations = {
 	.read_folio     = ubifs_read_folio,
-	.writepage      = ubifs_writepage,
+	.writepages     = ubifs_writepages,
 	.write_begin    = ubifs_write_begin,
 	.write_end      = ubifs_write_end,
 	.invalidate_folio = ubifs_invalidate_folio,
diff --git a/fs/ubifs/find.c b/fs/ubifs/find.c
index 873e6e1c92..6ebf3c04ac 100644
--- a/fs/ubifs/find.c
+++ b/fs/ubifs/find.c
@@ -82,8 +82,9 @@ static int valuable(struct ubifs_info *c, const struct ubifs_lprops *lprops)
  */
 static int scan_for_dirty_cb(struct ubifs_info *c,
 			     const struct ubifs_lprops *lprops, int in_tree,
-			     struct scan_data *data)
+			     void *arg)
 {
+	struct scan_data *data = arg;
 	int ret = LPT_SCAN_CONTINUE;
 
 	/* Exclude LEBs that are currently in use */
@@ -166,8 +167,7 @@ static const struct ubifs_lprops *scan_for_dirty(struct ubifs_info *c,
 	data.pick_free = pick_free;
 	data.lnum = -1;
 	data.exclude_index = exclude_index;
-	err = ubifs_lpt_scan_nolock(c, -1, c->lscan_lnum,
-				    (ubifs_lpt_scan_callback)scan_for_dirty_cb,
+	err = ubifs_lpt_scan_nolock(c, -1, c->lscan_lnum, scan_for_dirty_cb,
 				    &data);
 	if (err)
 		return ERR_PTR(err);
@@ -349,8 +349,9 @@ out:
  */
 static int scan_for_free_cb(struct ubifs_info *c,
 			    const struct ubifs_lprops *lprops, int in_tree,
-			    struct scan_data *data)
+			    void *arg)
 {
+	struct scan_data *data = arg;
 	int ret = LPT_SCAN_CONTINUE;
 
 	/* Exclude LEBs that are currently in use */
@@ -446,7 +447,7 @@ const struct ubifs_lprops *do_find_free_space(struct ubifs_info *c,
 	data.pick_free = pick_free;
 	data.lnum = -1;
 	err = ubifs_lpt_scan_nolock(c, -1, c->lscan_lnum,
-				    (ubifs_lpt_scan_callback)scan_for_free_cb,
+				    scan_for_free_cb,
 				    &data);
 	if (err)
 		return ERR_PTR(err);
@@ -589,8 +590,9 @@ out:
  */
 static int scan_for_idx_cb(struct ubifs_info *c,
 			   const struct ubifs_lprops *lprops, int in_tree,
-			   struct scan_data *data)
+			   void *arg)
 {
+	struct scan_data *data = arg;
 	int ret = LPT_SCAN_CONTINUE;
 
 	/* Exclude LEBs that are currently in use */
@@ -625,8 +627,7 @@ static const struct ubifs_lprops *scan_for_leb_for_idx(struct ubifs_info *c)
 	int err;
 
 	data.lnum = -1;
-	err = ubifs_lpt_scan_nolock(c, -1, c->lscan_lnum,
-				    (ubifs_lpt_scan_callback)scan_for_idx_cb,
+	err = ubifs_lpt_scan_nolock(c, -1, c->lscan_lnum, scan_for_idx_cb,
 				    &data);
 	if (err)
 		return ERR_PTR(err);
@@ -726,11 +727,10 @@ out:
 	return err;
 }
 
-static int cmp_dirty_idx(const struct ubifs_lprops **a,
-			 const struct ubifs_lprops **b)
+static int cmp_dirty_idx(const void *a, const void *b)
 {
-	const struct ubifs_lprops *lpa = *a;
-	const struct ubifs_lprops *lpb = *b;
+	const struct ubifs_lprops *lpa = *(const struct ubifs_lprops **)a;
+	const struct ubifs_lprops *lpb = *(const struct ubifs_lprops **)b;
 
 	return lpa->dirty + lpa->free - lpb->dirty - lpb->free;
 }
@@ -754,7 +754,7 @@ int ubifs_save_dirty_idx_lnums(struct ubifs_info *c)
 	       sizeof(void *) * c->dirty_idx.cnt);
 	/* Sort it so that the dirtiest is now at the end */
 	sort(c->dirty_idx.arr, c->dirty_idx.cnt, sizeof(void *),
-	     (int (*)(const void *, const void *))cmp_dirty_idx, NULL);
+	     cmp_dirty_idx, NULL);
 	dbg_find("found %d dirty index LEBs", c->dirty_idx.cnt);
 	if (c->dirty_idx.cnt)
 		dbg_find("dirtiest index LEB is %d with dirty %d and free %d",
@@ -782,8 +782,9 @@ int ubifs_save_dirty_idx_lnums(struct ubifs_info *c)
  */
 static int scan_dirty_idx_cb(struct ubifs_info *c,
 			   const struct ubifs_lprops *lprops, int in_tree,
-			   struct scan_data *data)
+			   void *arg)
 {
+	struct scan_data *data = arg;
 	int ret = LPT_SCAN_CONTINUE;
 
 	/* Exclude LEBs that are currently in use */
@@ -842,8 +843,7 @@ static int find_dirty_idx_leb(struct ubifs_info *c)
 	if (c->pnodes_have >= c->pnode_cnt)
 		/* All pnodes are in memory, so skip scan */
 		return -ENOSPC;
-	err = ubifs_lpt_scan_nolock(c, -1, c->lscan_lnum,
-				    (ubifs_lpt_scan_callback)scan_dirty_idx_cb,
+	err = ubifs_lpt_scan_nolock(c, -1, c->lscan_lnum, scan_dirty_idx_cb,
 				    &data);
 	if (err)
 		return err;
diff --git a/fs/ubifs/journal.c b/fs/ubifs/journal.c
index f0a5538c84..74aee92433 100644
--- a/fs/ubifs/journal.c
+++ b/fs/ubifs/journal.c
@@ -293,6 +293,96 @@ static int write_head(struct ubifs_info *c, int jhead, void *buf, int len,
 }
 
 /**
+ * __queue_and_wait - queue a task and wait until the task is waked up.
+ * @c: UBIFS file-system description object
+ *
+ * This function adds current task in queue and waits until the task is waked
+ * up. This function should be called with @c->reserve_space_wq locked.
+ */
+static void __queue_and_wait(struct ubifs_info *c)
+{
+	DEFINE_WAIT(wait);
+
+	__add_wait_queue_entry_tail_exclusive(&c->reserve_space_wq, &wait);
+	set_current_state(TASK_UNINTERRUPTIBLE);
+	spin_unlock(&c->reserve_space_wq.lock);
+
+	schedule();
+	finish_wait(&c->reserve_space_wq, &wait);
+}
+
+/**
+ * wait_for_reservation - try queuing current task to wait until waked up.
+ * @c: UBIFS file-system description object
+ *
+ * This function queues current task to wait until waked up, if queuing is
+ * started(@c->need_wait_space is not %0). Returns %true if current task is
+ * added in queue, otherwise %false is returned.
+ */
+static bool wait_for_reservation(struct ubifs_info *c)
+{
+	if (likely(atomic_read(&c->need_wait_space) == 0))
+		/* Quick path to check whether queuing is started. */
+		return false;
+
+	spin_lock(&c->reserve_space_wq.lock);
+	if (atomic_read(&c->need_wait_space) == 0) {
+		/* Queuing is not started, don't queue current task. */
+		spin_unlock(&c->reserve_space_wq.lock);
+		return false;
+	}
+
+	__queue_and_wait(c);
+	return true;
+}
+
+/**
+ * wake_up_reservation - wake up first task in queue or stop queuing.
+ * @c: UBIFS file-system description object
+ *
+ * This function wakes up the first task in queue if it exists, or stops
+ * queuing if no tasks in queue.
+ */
+static void wake_up_reservation(struct ubifs_info *c)
+{
+	spin_lock(&c->reserve_space_wq.lock);
+	if (waitqueue_active(&c->reserve_space_wq))
+		wake_up_locked(&c->reserve_space_wq);
+	else
+		/*
+		 * Compared with wait_for_reservation(), set @c->need_wait_space
+		 * under the protection of wait queue lock, which can avoid that
+		 * @c->need_wait_space is set to 0 after new task queued.
+		 */
+		atomic_set(&c->need_wait_space, 0);
+	spin_unlock(&c->reserve_space_wq.lock);
+}
+
+/**
+ * wake_up_reservation - add current task in queue or start queuing.
+ * @c: UBIFS file-system description object
+ *
+ * This function starts queuing if queuing is not started, otherwise adds
+ * current task in queue.
+ */
+static void add_or_start_queue(struct ubifs_info *c)
+{
+	spin_lock(&c->reserve_space_wq.lock);
+	if (atomic_cmpxchg(&c->need_wait_space, 0, 1) == 0) {
+		/* Starts queuing, task can go on directly. */
+		spin_unlock(&c->reserve_space_wq.lock);
+		return;
+	}
+
+	/*
+	 * There are at least two tasks have retried more than 32 times
+	 * at certain point, first task has started queuing, just queue
+	 * the left tasks.
+	 */
+	__queue_and_wait(c);
+}
+
+/**
  * make_reservation - reserve journal space.
  * @c: UBIFS file-system description object
  * @jhead: journal head
@@ -311,33 +401,27 @@ static int write_head(struct ubifs_info *c, int jhead, void *buf, int len,
 static int make_reservation(struct ubifs_info *c, int jhead, int len)
 {
 	int err, cmt_retries = 0, nospc_retries = 0;
+	bool blocked = wait_for_reservation(c);
 
 again:
 	down_read(&c->commit_sem);
 	err = reserve_space(c, jhead, len);
-	if (!err)
+	if (!err) {
 		/* c->commit_sem will get released via finish_reservation(). */
-		return 0;
+		goto out_wake_up;
+	}
 	up_read(&c->commit_sem);
 
 	if (err == -ENOSPC) {
 		/*
 		 * GC could not make any progress. We should try to commit
-		 * once because it could make some dirty space and GC would
-		 * make progress, so make the error -EAGAIN so that the below
+		 * because it could make some dirty space and GC would make
+		 * progress, so make the error -EAGAIN so that the below
 		 * will commit and re-try.
 		 */
-		if (nospc_retries++ < 2) {
-			dbg_jnl("no space, retry");
-			err = -EAGAIN;
-		}
-
-		/*
-		 * This means that the budgeting is incorrect. We always have
-		 * to be able to write to the media, because all operations are
-		 * budgeted. Deletions are not budgeted, though, but we reserve
-		 * an extra LEB for them.
-		 */
+		nospc_retries++;
+		dbg_jnl("no space, retry");
+		err = -EAGAIN;
 	}
 
 	if (err != -EAGAIN)
@@ -349,15 +433,37 @@ again:
 	 */
 	if (cmt_retries > 128) {
 		/*
-		 * This should not happen unless the journal size limitations
-		 * are too tough.
+		 * This should not happen unless:
+		 * 1. The journal size limitations are too tough.
+		 * 2. The budgeting is incorrect. We always have to be able to
+		 *    write to the media, because all operations are budgeted.
+		 *    Deletions are not budgeted, though, but we reserve an
+		 *    extra LEB for them.
 		 */
-		ubifs_err(c, "stuck in space allocation");
+		ubifs_err(c, "stuck in space allocation, nospc_retries %d",
+			  nospc_retries);
 		err = -ENOSPC;
 		goto out;
-	} else if (cmt_retries > 32)
-		ubifs_warn(c, "too many space allocation re-tries (%d)",
-			   cmt_retries);
+	} else if (cmt_retries > 32) {
+		/*
+		 * It's almost impossible to happen, unless there are many tasks
+		 * making reservation concurrently and someone task has retried
+		 * gc + commit for many times, generated available space during
+		 * this period are grabbed by other tasks.
+		 * But if it happens, start queuing up all tasks that will make
+		 * space reservation, then there is only one task making space
+		 * reservation at any time, and it can always make success under
+		 * the premise of correct budgeting.
+		 */
+		ubifs_warn(c, "too many space allocation cmt_retries (%d) "
+			   "nospc_retries (%d), start queuing tasks",
+			   cmt_retries, nospc_retries);
+
+		if (!blocked) {
+			blocked = true;
+			add_or_start_queue(c);
+		}
+	}
 
 	dbg_jnl("-EAGAIN, commit and retry (retried %d times)",
 		cmt_retries);
@@ -365,7 +471,7 @@ again:
 
 	err = ubifs_run_commit(c);
 	if (err)
-		return err;
+		goto out_wake_up;
 	goto again;
 
 out:
@@ -380,6 +486,27 @@ out:
 		cmt_retries = dbg_check_lprops(c);
 		up_write(&c->commit_sem);
 	}
+out_wake_up:
+	if (blocked) {
+		/*
+		 * Only tasks that have ever started queuing or ever been queued
+		 * can wake up other queued tasks, which can make sure that
+		 * there is only one task waked up to make space reservation.
+		 * For example:
+		 *      task A          task B           task C
+		 *                 make_reservation  make_reservation
+		 * reserve_space // 0
+		 * wake_up_reservation
+		 *                  atomic_cmpxchg // 0, start queuing
+		 *                  reserve_space
+		 *                                    wait_for_reservation
+		 *                                     __queue_and_wait
+		 *                                      add_wait_queue
+		 *  if (blocked) // false
+		 *  // So that task C won't be waked up to race with task B
+		 */
+		wake_up_reservation(c);
+	}
 	return err;
 }
 
diff --git a/fs/ubifs/lprops.c b/fs/ubifs/lprops.c
index 6d6cd85c2b..a11c3dab7e 100644
--- a/fs/ubifs/lprops.c
+++ b/fs/ubifs/lprops.c
@@ -1014,8 +1014,9 @@ out:
  */
 static int scan_check_cb(struct ubifs_info *c,
 			 const struct ubifs_lprops *lp, int in_tree,
-			 struct ubifs_lp_stats *lst)
+			 void *arg)
 {
+	struct ubifs_lp_stats *lst = arg;
 	struct ubifs_scan_leb *sleb;
 	struct ubifs_scan_node *snod;
 	int cat, lnum = lp->lnum, is_idx = 0, used = 0, free, dirty, ret;
@@ -1269,8 +1270,7 @@ int dbg_check_lprops(struct ubifs_info *c)
 
 	memset(&lst, 0, sizeof(struct ubifs_lp_stats));
 	err = ubifs_lpt_scan_nolock(c, c->main_first, c->leb_cnt - 1,
-				    (ubifs_lpt_scan_callback)scan_check_cb,
-				    &lst);
+				    scan_check_cb, &lst);
 	if (err && err != -ENOSPC)
 		goto out;
 
diff --git a/fs/ubifs/lpt_commit.c b/fs/ubifs/lpt_commit.c
index c4d079328b..07351fdce7 100644
--- a/fs/ubifs/lpt_commit.c
+++ b/fs/ubifs/lpt_commit.c
@@ -1646,7 +1646,6 @@ static int dbg_check_ltab_lnum(struct ubifs_info *c, int lnum)
 		len -= node_len;
 	}
 
-	err = 0;
 out:
 	vfree(buf);
 	return err;
diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c
index 09e270d6ed..291583005d 100644
--- a/fs/ubifs/super.c
+++ b/fs/ubifs/super.c
@@ -2151,6 +2151,8 @@ static struct ubifs_info *alloc_ubifs_info(struct ubi_volume_desc *ubi)
 		mutex_init(&c->bu_mutex);
 		mutex_init(&c->write_reserve_mutex);
 		init_waitqueue_head(&c->cmt_wq);
+		init_waitqueue_head(&c->reserve_space_wq);
+		atomic_set(&c->need_wait_space, 0);
 		c->buds = RB_ROOT;
 		c->old_idx = RB_ROOT;
 		c->size_tree = RB_ROOT;
@@ -2239,13 +2241,14 @@ static int ubifs_fill_super(struct super_block *sb, void *data, int silent)
 		goto out_umount;
 	}
 
+	generic_set_sb_d_ops(sb);
 	sb->s_root = d_make_root(root);
 	if (!sb->s_root) {
 		err = -ENOMEM;
 		goto out_umount;
 	}
 
-	import_uuid(&sb->s_uuid, c->uuid);
+	super_set_uuid(sb, c->uuid, sizeof(c->uuid));
 
 	mutex_unlock(&c->umount_mutex);
 	return 0;
@@ -2433,8 +2436,8 @@ static int __init ubifs_init(void)
 
 	ubifs_inode_slab = kmem_cache_create("ubifs_inode_slab",
 				sizeof(struct ubifs_inode), 0,
-				SLAB_MEM_SPREAD | SLAB_RECLAIM_ACCOUNT |
-				SLAB_ACCOUNT, &inode_slab_ctor);
+				SLAB_RECLAIM_ACCOUNT | SLAB_ACCOUNT,
+				&inode_slab_ctor);
 	if (!ubifs_inode_slab)
 		return -ENOMEM;
 
diff --git a/fs/ubifs/tnc.c b/fs/ubifs/tnc.c
index f4728e65d1..45cacdcd47 100644
--- a/fs/ubifs/tnc.c
+++ b/fs/ubifs/tnc.c
@@ -3116,14 +3116,7 @@ static void tnc_destroy_cnext(struct ubifs_info *c)
 void ubifs_tnc_close(struct ubifs_info *c)
 {
 	tnc_destroy_cnext(c);
-	if (c->zroot.znode) {
-		long n, freed;
-
-		n = atomic_long_read(&c->clean_zn_cnt);
-		freed = ubifs_destroy_tnc_subtree(c, c->zroot.znode);
-		ubifs_assert(c, freed == n);
-		atomic_long_sub(n, &ubifs_clean_zn_cnt);
-	}
+	ubifs_destroy_tnc_tree(c);
 	kfree(c->gap_lebs);
 	kfree(c->ilebs);
 	destroy_old_idx(c);
diff --git a/fs/ubifs/tnc_misc.c b/fs/ubifs/tnc_misc.c
index 4d686e34e6..d3f8a6aa1f 100644
--- a/fs/ubifs/tnc_misc.c
+++ b/fs/ubifs/tnc_misc.c
@@ -251,6 +251,28 @@ long ubifs_destroy_tnc_subtree(const struct ubifs_info *c,
 }
 
 /**
+ * ubifs_destroy_tnc_tree - destroy all znodes connected to the TNC tree.
+ * @c: UBIFS file-system description object
+ *
+ * This function destroys the whole TNC tree and updates clean global znode
+ * count.
+ */
+void ubifs_destroy_tnc_tree(struct ubifs_info *c)
+{
+	long n, freed;
+
+	if (!c->zroot.znode)
+		return;
+
+	n = atomic_long_read(&c->clean_zn_cnt);
+	freed = ubifs_destroy_tnc_subtree(c, c->zroot.znode);
+	ubifs_assert(c, freed == n);
+	atomic_long_sub(n, &ubifs_clean_zn_cnt);
+
+	c->zroot.znode = NULL;
+}
+
+/**
  * read_znode - read an indexing node from flash and fill znode.
  * @c: UBIFS file-system description object
  * @zzbr: the zbranch describing the node to read
diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h
index 3916dc4f30..1f3ea879d9 100644
--- a/fs/ubifs/ubifs.h
+++ b/fs/ubifs/ubifs.h
@@ -1047,6 +1047,8 @@ struct ubifs_debug_info;
  * @bg_bud_bytes: number of bud bytes when background commit is initiated
  * @old_buds: buds to be released after commit ends
  * @max_bud_cnt: maximum number of buds
+ * @need_wait_space: Non %0 means space reservation tasks need to wait in queue
+ * @reserve_space_wq: wait queue to sleep on if @need_wait_space is not %0
  *
  * @commit_sem: synchronizes committer with other processes
  * @cmt_state: commit state
@@ -1305,6 +1307,8 @@ struct ubifs_info {
 	long long bg_bud_bytes;
 	struct list_head old_buds;
 	int max_bud_cnt;
+	atomic_t need_wait_space;
+	wait_queue_head_t reserve_space_wq;
 
 	struct rw_semaphore commit_sem;
 	int cmt_state;
@@ -1903,6 +1907,7 @@ struct ubifs_znode *ubifs_tnc_postorder_next(const struct ubifs_info *c,
 					     struct ubifs_znode *znode);
 long ubifs_destroy_tnc_subtree(const struct ubifs_info *c,
 			       struct ubifs_znode *zr);
+void ubifs_destroy_tnc_tree(struct ubifs_info *c);
 struct ubifs_znode *ubifs_load_znode(struct ubifs_info *c,
 				     struct ubifs_zbranch *zbr,
 				     struct ubifs_znode *parent, int iip);
diff --git a/fs/udf/balloc.c b/fs/udf/balloc.c
index ab3ffc3559..bb471ec364 100644
--- a/fs/udf/balloc.c
+++ b/fs/udf/balloc.c
@@ -18,6 +18,7 @@
 #include "udfdecl.h"
 
 #include <linux/bitops.h>
+#include <linux/overflow.h>
 
 #include "udf_i.h"
 #include "udf_sb.h"
@@ -64,8 +65,12 @@ static int read_block_bitmap(struct super_block *sb,
 	}
 
 	for (i = 0; i < count; i++)
-		if (udf_test_bit(i + off, bh->b_data))
+		if (udf_test_bit(i + off, bh->b_data)) {
+			bitmap->s_block_bitmap[bitmap_nr] =
+							ERR_PTR(-EFSCORRUPTED);
+			brelse(bh);
 			return -EFSCORRUPTED;
+		}
 	return 0;
 }
 
@@ -81,8 +86,15 @@ static int __load_block_bitmap(struct super_block *sb,
 			  block_group, nr_groups);
 	}
 
-	if (bitmap->s_block_bitmap[block_group])
+	if (bitmap->s_block_bitmap[block_group]) {
+		/*
+		 * The bitmap failed verification in the past. No point in
+		 * trying again.
+		 */
+		if (IS_ERR(bitmap->s_block_bitmap[block_group]))
+			return PTR_ERR(bitmap->s_block_bitmap[block_group]);
 		return block_group;
+	}
 
 	retval = read_block_bitmap(sb, bitmap, block_group, block_group);
 	if (retval < 0)
@@ -129,7 +141,6 @@ static void udf_bitmap_free_blocks(struct super_block *sb,
 {
 	struct udf_sb_info *sbi = UDF_SB(sb);
 	struct buffer_head *bh = NULL;
-	struct udf_part_map *partmap;
 	unsigned long block;
 	unsigned long block_group;
 	unsigned long bit;
@@ -138,19 +149,9 @@ static void udf_bitmap_free_blocks(struct super_block *sb,
 	unsigned long overflow;
 
 	mutex_lock(&sbi->s_alloc_mutex);
-	partmap = &sbi->s_partmaps[bloc->partitionReferenceNum];
-	if (bloc->logicalBlockNum + count < count ||
-	    (bloc->logicalBlockNum + count) > partmap->s_partition_len) {
-		udf_debug("%u < %d || %u + %u > %u\n",
-			  bloc->logicalBlockNum, 0,
-			  bloc->logicalBlockNum, count,
-			  partmap->s_partition_len);
-		goto error_return;
-	}
-
+	/* We make sure this cannot overflow when mounting the filesystem */
 	block = bloc->logicalBlockNum + offset +
 		(sizeof(struct spaceBitmapDesc) << 3);
-
 	do {
 		overflow = 0;
 		block_group = block >> (sb->s_blocksize_bits + 3);
@@ -380,7 +381,6 @@ static void udf_table_free_blocks(struct super_block *sb,
 				  uint32_t count)
 {
 	struct udf_sb_info *sbi = UDF_SB(sb);
-	struct udf_part_map *partmap;
 	uint32_t start, end;
 	uint32_t elen;
 	struct kernel_lb_addr eloc;
@@ -389,16 +389,6 @@ static void udf_table_free_blocks(struct super_block *sb,
 	struct udf_inode_info *iinfo;
 
 	mutex_lock(&sbi->s_alloc_mutex);
-	partmap = &sbi->s_partmaps[bloc->partitionReferenceNum];
-	if (bloc->logicalBlockNum + count < count ||
-	    (bloc->logicalBlockNum + count) > partmap->s_partition_len) {
-		udf_debug("%u < %d || %u + %u > %u\n",
-			  bloc->logicalBlockNum, 0,
-			  bloc->logicalBlockNum, count,
-			  partmap->s_partition_len);
-		goto error_return;
-	}
-
 	iinfo = UDF_I(table);
 	udf_add_free_space(sb, sbi->s_partition, count);
 
@@ -673,6 +663,17 @@ void udf_free_blocks(struct super_block *sb, struct inode *inode,
 {
 	uint16_t partition = bloc->partitionReferenceNum;
 	struct udf_part_map *map = &UDF_SB(sb)->s_partmaps[partition];
+	uint32_t blk;
+
+	if (check_add_overflow(bloc->logicalBlockNum, offset, &blk) ||
+	    check_add_overflow(blk, count, &blk) ||
+	    bloc->logicalBlockNum + count > map->s_partition_len) {
+		udf_debug("Invalid request to free blocks: (%d, %u), off %u, "
+			  "len %u, partition len %u\n",
+			  partition, bloc->logicalBlockNum, offset, count,
+			  map->s_partition_len);
+		return;
+	}
 
 	if (map->s_partition_flags & UDF_PART_FLAG_UNALLOC_BITMAP) {
 		udf_bitmap_free_blocks(sb, map->s_uspace.s_bitmap,
diff --git a/fs/udf/dir.c b/fs/udf/dir.c
index f6533f9385..f94f45fe2c 100644
--- a/fs/udf/dir.c
+++ b/fs/udf/dir.c
@@ -67,7 +67,7 @@ static int udf_readdir(struct file *file, struct dir_context *ctx)
 		pos_valid = true;
 	}
 
-	fname = kmalloc(UDF_NAME_LEN, GFP_NOFS);
+	fname = kmalloc(UDF_NAME_LEN, GFP_KERNEL);
 	if (!fname) {
 		ret = -ENOMEM;
 		goto out;
diff --git a/fs/udf/file.c b/fs/udf/file.c
index 0ceac4b593..3a4179de31 100644
--- a/fs/udf/file.c
+++ b/fs/udf/file.c
@@ -39,7 +39,7 @@ static vm_fault_t udf_page_mkwrite(struct vm_fault *vmf)
 	struct vm_area_struct *vma = vmf->vma;
 	struct inode *inode = file_inode(vma->vm_file);
 	struct address_space *mapping = inode->i_mapping;
-	struct page *page = vmf->page;
+	struct folio *folio = page_folio(vmf->page);
 	loff_t size;
 	unsigned int end;
 	vm_fault_t ret = VM_FAULT_LOCKED;
@@ -48,31 +48,31 @@ static vm_fault_t udf_page_mkwrite(struct vm_fault *vmf)
 	sb_start_pagefault(inode->i_sb);
 	file_update_time(vma->vm_file);
 	filemap_invalidate_lock_shared(mapping);
-	lock_page(page);
+	folio_lock(folio);
 	size = i_size_read(inode);
-	if (page->mapping != inode->i_mapping || page_offset(page) >= size) {
-		unlock_page(page);
+	if (folio->mapping != inode->i_mapping || folio_pos(folio) >= size) {
+		folio_unlock(folio);
 		ret = VM_FAULT_NOPAGE;
 		goto out_unlock;
 	}
 	/* Space is already allocated for in-ICB file */
 	if (UDF_I(inode)->i_alloc_type == ICBTAG_FLAG_AD_IN_ICB)
 		goto out_dirty;
-	if (page->index == size >> PAGE_SHIFT)
+	if (folio->index == size >> PAGE_SHIFT)
 		end = size & ~PAGE_MASK;
 	else
 		end = PAGE_SIZE;
-	err = __block_write_begin(page, 0, end, udf_get_block);
+	err = __block_write_begin(&folio->page, 0, end, udf_get_block);
 	if (err) {
-		unlock_page(page);
+		folio_unlock(folio);
 		ret = vmf_fs_error(err);
 		goto out_unlock;
 	}
 
-	block_commit_write(page, 0, end);
+	block_commit_write(&folio->page, 0, end);
 out_dirty:
-	set_page_dirty(page);
-	wait_for_stable_page(page);
+	folio_mark_dirty(folio);
+	folio_wait_stable(folio);
 out_unlock:
 	filemap_invalidate_unlock_shared(mapping);
 	sb_end_pagefault(inode->i_sb);
@@ -232,7 +232,9 @@ static int udf_setattr(struct mnt_idmap *idmap, struct dentry *dentry,
 
 	if ((attr->ia_valid & ATTR_SIZE) &&
 	    attr->ia_size != i_size_read(inode)) {
+		filemap_invalidate_lock(inode->i_mapping);
 		error = udf_setsize(inode, attr->ia_size);
+		filemap_invalidate_unlock(inode->i_mapping);
 		if (error)
 			return error;
 	}
diff --git a/fs/udf/inode.c b/fs/udf/inode.c
index d8493449d4..08767bd21e 100644
--- a/fs/udf/inode.c
+++ b/fs/udf/inode.c
@@ -208,19 +208,14 @@ static int udf_writepages(struct address_space *mapping,
 	return write_cache_pages(mapping, wbc, udf_adinicb_writepage, NULL);
 }
 
-static void udf_adinicb_readpage(struct page *page)
+static void udf_adinicb_read_folio(struct folio *folio)
 {
-	struct inode *inode = page->mapping->host;
-	char *kaddr;
+	struct inode *inode = folio->mapping->host;
 	struct udf_inode_info *iinfo = UDF_I(inode);
 	loff_t isize = i_size_read(inode);
 
-	kaddr = kmap_local_page(page);
-	memcpy(kaddr, iinfo->i_data + iinfo->i_lenEAttr, isize);
-	memset(kaddr + isize, 0, PAGE_SIZE - isize);
-	flush_dcache_page(page);
-	SetPageUptodate(page);
-	kunmap_local(kaddr);
+	folio_fill_tail(folio, 0, iinfo->i_data + iinfo->i_lenEAttr, isize);
+	folio_mark_uptodate(folio);
 }
 
 static int udf_read_folio(struct file *file, struct folio *folio)
@@ -228,7 +223,7 @@ static int udf_read_folio(struct file *file, struct folio *folio)
 	struct udf_inode_info *iinfo = UDF_I(file_inode(file));
 
 	if (iinfo->i_alloc_type == ICBTAG_FLAG_AD_IN_ICB) {
-		udf_adinicb_readpage(&folio->page);
+		udf_adinicb_read_folio(folio);
 		folio_unlock(folio);
 		return 0;
 	}
@@ -254,7 +249,7 @@ static int udf_write_begin(struct file *file, struct address_space *mapping,
 			   struct page **pagep, void **fsdata)
 {
 	struct udf_inode_info *iinfo = UDF_I(file_inode(file));
-	struct page *page;
+	struct folio *folio;
 	int ret;
 
 	if (iinfo->i_alloc_type != ICBTAG_FLAG_AD_IN_ICB) {
@@ -266,12 +261,13 @@ static int udf_write_begin(struct file *file, struct address_space *mapping,
 	}
 	if (WARN_ON_ONCE(pos >= PAGE_SIZE))
 		return -EIO;
-	page = grab_cache_page_write_begin(mapping, 0);
-	if (!page)
-		return -ENOMEM;
-	*pagep = page;
-	if (!PageUptodate(page))
-		udf_adinicb_readpage(page);
+	folio = __filemap_get_folio(mapping, 0, FGP_WRITEBEGIN,
+			mapping_gfp_mask(mapping));
+	if (IS_ERR(folio))
+		return PTR_ERR(folio);
+	*pagep = &folio->page;
+	if (!folio_test_uptodate(folio))
+		udf_adinicb_read_folio(folio);
 	return 0;
 }
 
@@ -280,17 +276,19 @@ static int udf_write_end(struct file *file, struct address_space *mapping,
 			 struct page *page, void *fsdata)
 {
 	struct inode *inode = file_inode(file);
+	struct folio *folio;
 	loff_t last_pos;
 
 	if (UDF_I(inode)->i_alloc_type != ICBTAG_FLAG_AD_IN_ICB)
 		return generic_write_end(file, mapping, pos, len, copied, page,
 					 fsdata);
+	folio = page_folio(page);
 	last_pos = pos + copied;
 	if (last_pos > inode->i_size)
 		i_size_write(inode, last_pos);
-	set_page_dirty(page);
-	unlock_page(page);
-	put_page(page);
+	folio_mark_dirty(folio);
+	folio_unlock(folio);
+	folio_put(folio);
 
 	return copied;
 }
@@ -341,7 +339,7 @@ const struct address_space_operations udf_aops = {
  */
 int udf_expand_file_adinicb(struct inode *inode)
 {
-	struct page *page;
+	struct folio *folio;
 	struct udf_inode_info *iinfo = UDF_I(inode);
 	int err;
 
@@ -357,12 +355,13 @@ int udf_expand_file_adinicb(struct inode *inode)
 		return 0;
 	}
 
-	page = find_or_create_page(inode->i_mapping, 0, GFP_NOFS);
-	if (!page)
-		return -ENOMEM;
+	folio = __filemap_get_folio(inode->i_mapping, 0,
+			FGP_LOCK | FGP_ACCESSED | FGP_CREAT, GFP_KERNEL);
+	if (IS_ERR(folio))
+		return PTR_ERR(folio);
 
-	if (!PageUptodate(page))
-		udf_adinicb_readpage(page);
+	if (!folio_test_uptodate(folio))
+		udf_adinicb_read_folio(folio);
 	down_write(&iinfo->i_data_sem);
 	memset(iinfo->i_data + iinfo->i_lenEAttr, 0x00,
 	       iinfo->i_lenAlloc);
@@ -371,22 +370,22 @@ int udf_expand_file_adinicb(struct inode *inode)
 		iinfo->i_alloc_type = ICBTAG_FLAG_AD_SHORT;
 	else
 		iinfo->i_alloc_type = ICBTAG_FLAG_AD_LONG;
-	set_page_dirty(page);
-	unlock_page(page);
+	folio_mark_dirty(folio);
+	folio_unlock(folio);
 	up_write(&iinfo->i_data_sem);
 	err = filemap_fdatawrite(inode->i_mapping);
 	if (err) {
 		/* Restore everything back so that we don't lose data... */
-		lock_page(page);
+		folio_lock(folio);
 		down_write(&iinfo->i_data_sem);
-		memcpy_to_page(page, 0, iinfo->i_data + iinfo->i_lenEAttr,
-			       inode->i_size);
-		unlock_page(page);
+		memcpy_from_folio(iinfo->i_data + iinfo->i_lenEAttr,
+				folio, 0, inode->i_size);
+		folio_unlock(folio);
 		iinfo->i_alloc_type = ICBTAG_FLAG_AD_IN_ICB;
 		iinfo->i_lenAlloc = inode->i_size;
 		up_write(&iinfo->i_data_sem);
 	}
-	put_page(page);
+	folio_put(folio);
 	mark_inode_dirty(inode);
 
 	return err;
@@ -1251,7 +1250,6 @@ int udf_setsize(struct inode *inode, loff_t newsize)
 	if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
 		return -EPERM;
 
-	filemap_invalidate_lock(inode->i_mapping);
 	iinfo = UDF_I(inode);
 	if (newsize > inode->i_size) {
 		if (iinfo->i_alloc_type == ICBTAG_FLAG_AD_IN_ICB) {
@@ -1264,11 +1262,11 @@ int udf_setsize(struct inode *inode, loff_t newsize)
 			}
 			err = udf_expand_file_adinicb(inode);
 			if (err)
-				goto out_unlock;
+				return err;
 		}
 		err = udf_extend_file(inode, newsize);
 		if (err)
-			goto out_unlock;
+			return err;
 set_size:
 		truncate_setsize(inode, newsize);
 	} else {
@@ -1286,14 +1284,14 @@ set_size:
 		err = block_truncate_page(inode->i_mapping, newsize,
 					  udf_get_block);
 		if (err)
-			goto out_unlock;
+			return err;
 		truncate_setsize(inode, newsize);
 		down_write(&iinfo->i_data_sem);
 		udf_clear_extent_cache(inode);
 		err = udf_truncate_extents(inode);
 		up_write(&iinfo->i_data_sem);
 		if (err)
-			goto out_unlock;
+			return err;
 	}
 update_time:
 	inode_set_mtime_to_ts(inode, inode_set_ctime_current(inode));
@@ -1301,8 +1299,6 @@ update_time:
 		udf_sync_inode(inode);
 	else
 		mark_inode_dirty(inode);
-out_unlock:
-	filemap_invalidate_unlock(inode->i_mapping);
 	return err;
 }
 
diff --git a/fs/udf/namei.c b/fs/udf/namei.c
index 1bb6ed9489..78a603129d 100644
--- a/fs/udf/namei.c
+++ b/fs/udf/namei.c
@@ -59,7 +59,7 @@ static int udf_fiiter_find_entry(struct inode *dir, const struct qstr *child,
 		child->name[0] == '.' && child->name[1] == '.';
 	int ret;
 
-	fname = kmalloc(UDF_NAME_LEN, GFP_NOFS);
+	fname = kmalloc(UDF_NAME_LEN, GFP_KERNEL);
 	if (!fname)
 		return -ENOMEM;
 
@@ -566,7 +566,7 @@ out:
 static int udf_symlink(struct mnt_idmap *idmap, struct inode *dir,
 		       struct dentry *dentry, const char *symname)
 {
-	struct inode *inode = udf_new_inode(dir, S_IFLNK | 0777);
+	struct inode *inode;
 	struct pathComponent *pc;
 	const char *compstart;
 	struct extent_position epos = {};
@@ -579,17 +579,20 @@ static int udf_symlink(struct mnt_idmap *idmap, struct inode *dir,
 	struct udf_inode_info *iinfo;
 	struct super_block *sb = dir->i_sb;
 
-	if (IS_ERR(inode))
-		return PTR_ERR(inode);
-
-	iinfo = UDF_I(inode);
-	down_write(&iinfo->i_data_sem);
-	name = kmalloc(UDF_NAME_LEN_CS0, GFP_NOFS);
+	name = kmalloc(UDF_NAME_LEN_CS0, GFP_KERNEL);
 	if (!name) {
 		err = -ENOMEM;
-		goto out_no_entry;
+		goto out;
+	}
+
+	inode = udf_new_inode(dir, S_IFLNK | 0777);
+	if (IS_ERR(inode)) {
+		err = PTR_ERR(inode);
+		goto out;
 	}
 
+	iinfo = UDF_I(inode);
+	down_write(&iinfo->i_data_sem);
 	inode->i_data.a_ops = &udf_symlink_aops;
 	inode->i_op = &udf_symlink_inode_operations;
 	inode_nohighmem(inode);
@@ -873,8 +876,6 @@ static int udf_rename(struct mnt_idmap *idmap, struct inode *old_dir,
 	if (has_diriter) {
 		diriter.fi.icb.extLocation =
 					cpu_to_lelb(UDF_I(new_dir)->i_location);
-		udf_update_tag((char *)&diriter.fi,
-			       udf_dir_entry_len(&diriter.fi));
 		udf_fiiter_write_fi(&diriter, NULL);
 		udf_fiiter_release(&diriter);
 	}
diff --git a/fs/udf/super.c b/fs/udf/super.c
index 928a04d9d9..92d4770539 100644
--- a/fs/udf/super.c
+++ b/fs/udf/super.c
@@ -40,20 +40,20 @@
 #include <linux/slab.h>
 #include <linux/kernel.h>
 #include <linux/module.h>
-#include <linux/parser.h>
 #include <linux/stat.h>
 #include <linux/cdrom.h>
 #include <linux/nls.h>
 #include <linux/vfs.h>
 #include <linux/vmalloc.h>
 #include <linux/errno.h>
-#include <linux/mount.h>
 #include <linux/seq_file.h>
 #include <linux/bitmap.h>
 #include <linux/crc-itu-t.h>
 #include <linux/log2.h>
 #include <asm/byteorder.h>
 #include <linux/iversion.h>
+#include <linux/fs_context.h>
+#include <linux/fs_parser.h>
 
 #include "udf_sb.h"
 #include "udf_i.h"
@@ -91,16 +91,20 @@ enum { UDF_MAX_LINKS = 0xffff };
 #define UDF_MAX_FILESIZE (1ULL << 42)
 
 /* These are the "meat" - everything else is stuffing */
-static int udf_fill_super(struct super_block *, void *, int);
+static int udf_fill_super(struct super_block *sb, struct fs_context *fc);
 static void udf_put_super(struct super_block *);
 static int udf_sync_fs(struct super_block *, int);
-static int udf_remount_fs(struct super_block *, int *, char *);
 static void udf_load_logicalvolint(struct super_block *, struct kernel_extent_ad);
 static void udf_open_lvid(struct super_block *);
 static void udf_close_lvid(struct super_block *);
 static unsigned int udf_count_free(struct super_block *);
 static int udf_statfs(struct dentry *, struct kstatfs *);
 static int udf_show_options(struct seq_file *, struct dentry *);
+static int udf_init_fs_context(struct fs_context *fc);
+static int udf_parse_param(struct fs_context *fc, struct fs_parameter *param);
+static int udf_reconfigure(struct fs_context *fc);
+static void udf_free_fc(struct fs_context *fc);
+static const struct fs_parameter_spec udf_param_spec[];
 
 struct logicalVolIntegrityDescImpUse *udf_sb_lvidiu(struct super_block *sb)
 {
@@ -119,18 +123,25 @@ struct logicalVolIntegrityDescImpUse *udf_sb_lvidiu(struct super_block *sb)
 }
 
 /* UDF filesystem type */
-static struct dentry *udf_mount(struct file_system_type *fs_type,
-		      int flags, const char *dev_name, void *data)
+static int udf_get_tree(struct fs_context *fc)
 {
-	return mount_bdev(fs_type, flags, dev_name, data, udf_fill_super);
+	return get_tree_bdev(fc, udf_fill_super);
 }
 
+static const struct fs_context_operations udf_context_ops = {
+	.parse_param	= udf_parse_param,
+	.get_tree	= udf_get_tree,
+	.reconfigure	= udf_reconfigure,
+	.free		= udf_free_fc,
+};
+
 static struct file_system_type udf_fstype = {
 	.owner		= THIS_MODULE,
 	.name		= "udf",
-	.mount		= udf_mount,
 	.kill_sb	= kill_block_super,
 	.fs_flags	= FS_REQUIRES_DEV,
+	.init_fs_context = udf_init_fs_context,
+	.parameters	= udf_param_spec,
 };
 MODULE_ALIAS_FS("udf");
 
@@ -177,7 +188,6 @@ static int __init init_inodecache(void)
 	udf_inode_cachep = kmem_cache_create("udf_inode_cache",
 					     sizeof(struct udf_inode_info),
 					     0, (SLAB_RECLAIM_ACCOUNT |
-						 SLAB_MEM_SPREAD |
 						 SLAB_ACCOUNT),
 					     init_once);
 	if (!udf_inode_cachep)
@@ -204,12 +214,10 @@ static const struct super_operations udf_sb_ops = {
 	.put_super	= udf_put_super,
 	.sync_fs	= udf_sync_fs,
 	.statfs		= udf_statfs,
-	.remount_fs	= udf_remount_fs,
 	.show_options	= udf_show_options,
 };
 
 struct udf_options {
-	unsigned char novrs;
 	unsigned int blocksize;
 	unsigned int session;
 	unsigned int lastblock;
@@ -223,6 +231,65 @@ struct udf_options {
 	struct nls_table *nls_map;
 };
 
+/*
+ * UDF has historically preserved prior mount options across
+ * a remount, so copy those here if remounting, otherwise set
+ * initial mount defaults.
+ */
+static void udf_init_options(struct fs_context *fc, struct udf_options *uopt)
+{
+	if (fc->purpose == FS_CONTEXT_FOR_RECONFIGURE) {
+		struct super_block *sb = fc->root->d_sb;
+		struct udf_sb_info *sbi = UDF_SB(sb);
+
+		uopt->flags = sbi->s_flags;
+		uopt->uid   = sbi->s_uid;
+		uopt->gid   = sbi->s_gid;
+		uopt->umask = sbi->s_umask;
+		uopt->fmode = sbi->s_fmode;
+		uopt->dmode = sbi->s_dmode;
+		uopt->nls_map = NULL;
+	} else {
+		uopt->flags = (1 << UDF_FLAG_USE_AD_IN_ICB) |
+			      (1 << UDF_FLAG_STRICT);
+		/*
+		 * By default we'll use overflow[ug]id when UDF
+		 * inode [ug]id == -1
+		 */
+		uopt->uid = make_kuid(current_user_ns(), overflowuid);
+		uopt->gid = make_kgid(current_user_ns(), overflowgid);
+		uopt->umask = 0;
+		uopt->fmode = UDF_INVALID_MODE;
+		uopt->dmode = UDF_INVALID_MODE;
+		uopt->nls_map = NULL;
+		uopt->session = 0xFFFFFFFF;
+	}
+}
+
+static int udf_init_fs_context(struct fs_context *fc)
+{
+	struct udf_options *uopt;
+
+	uopt = kzalloc(sizeof(*uopt), GFP_KERNEL);
+	if (!uopt)
+		return -ENOMEM;
+
+	udf_init_options(fc, uopt);
+
+	fc->fs_private = uopt;
+	fc->ops = &udf_context_ops;
+
+	return 0;
+}
+
+static void udf_free_fc(struct fs_context *fc)
+{
+	struct udf_options *uopt = fc->fs_private;
+
+	unload_nls(uopt->nls_map);
+	kfree(fc->fs_private);
+}
+
 static int __init init_udf_fs(void)
 {
 	int err;
@@ -269,7 +336,8 @@ static void udf_sb_free_bitmap(struct udf_bitmap *bitmap)
 	int nr_groups = bitmap->s_nr_groups;
 
 	for (i = 0; i < nr_groups; i++)
-		brelse(bitmap->s_block_bitmap[i]);
+		if (!IS_ERR_OR_NULL(bitmap->s_block_bitmap[i]))
+			brelse(bitmap->s_block_bitmap[i]);
 
 	kvfree(bitmap);
 }
@@ -358,7 +426,7 @@ static int udf_show_options(struct seq_file *seq, struct dentry *root)
 }
 
 /*
- * udf_parse_options
+ * udf_parse_param
  *
  * PURPOSE
  *	Parse mount options.
@@ -401,12 +469,12 @@ static int udf_show_options(struct seq_file *seq, struct dentry *root)
  *		yield highly unpredictable results.
  *
  * PRE-CONDITIONS
- *	options		Pointer to mount options string.
- *	uopts		Pointer to mount options variable.
+ *	fc		fs_context with pointer to mount options variable.
+ *	param		Pointer to fs_parameter being parsed.
  *
  * POST-CONDITIONS
- *	<return>	1	Mount options parsed okay.
- *	<return>	0	Error parsing mount options.
+ *	<return>	0	Mount options parsed okay.
+ *	<return>	errno	Error parsing mount options.
  *
  * HISTORY
  *	July 1, 1997 - Andrew E. Mileski
@@ -418,229 +486,193 @@ enum {
 	Opt_noadinicb, Opt_adinicb, Opt_shortad, Opt_longad,
 	Opt_gid, Opt_uid, Opt_umask, Opt_session, Opt_lastblock,
 	Opt_anchor, Opt_volume, Opt_partition, Opt_fileset,
-	Opt_rootdir, Opt_utf8, Opt_iocharset,
-	Opt_err, Opt_uforget, Opt_uignore, Opt_gforget, Opt_gignore,
-	Opt_fmode, Opt_dmode
-};
-
-static const match_table_t tokens = {
-	{Opt_novrs,	"novrs"},
-	{Opt_nostrict,	"nostrict"},
-	{Opt_bs,	"bs=%u"},
-	{Opt_unhide,	"unhide"},
-	{Opt_undelete,	"undelete"},
-	{Opt_noadinicb,	"noadinicb"},
-	{Opt_adinicb,	"adinicb"},
-	{Opt_shortad,	"shortad"},
-	{Opt_longad,	"longad"},
-	{Opt_uforget,	"uid=forget"},
-	{Opt_uignore,	"uid=ignore"},
-	{Opt_gforget,	"gid=forget"},
-	{Opt_gignore,	"gid=ignore"},
-	{Opt_gid,	"gid=%u"},
-	{Opt_uid,	"uid=%u"},
-	{Opt_umask,	"umask=%o"},
-	{Opt_session,	"session=%u"},
-	{Opt_lastblock,	"lastblock=%u"},
-	{Opt_anchor,	"anchor=%u"},
-	{Opt_volume,	"volume=%u"},
-	{Opt_partition,	"partition=%u"},
-	{Opt_fileset,	"fileset=%u"},
-	{Opt_rootdir,	"rootdir=%u"},
-	{Opt_utf8,	"utf8"},
-	{Opt_iocharset,	"iocharset=%s"},
-	{Opt_fmode,     "mode=%o"},
-	{Opt_dmode,     "dmode=%o"},
-	{Opt_err,	NULL}
+	Opt_rootdir, Opt_utf8, Opt_iocharset, Opt_err, Opt_fmode, Opt_dmode
 };
 
-static int udf_parse_options(char *options, struct udf_options *uopt,
-			     bool remount)
+static const struct fs_parameter_spec udf_param_spec[] = {
+	fsparam_flag	("novrs",		Opt_novrs),
+	fsparam_flag	("nostrict",		Opt_nostrict),
+	fsparam_u32	("bs",			Opt_bs),
+	fsparam_flag	("unhide",		Opt_unhide),
+	fsparam_flag	("undelete",		Opt_undelete),
+	fsparam_flag_no	("adinicb",		Opt_adinicb),
+	fsparam_flag	("shortad",		Opt_shortad),
+	fsparam_flag	("longad",		Opt_longad),
+	fsparam_string	("gid",			Opt_gid),
+	fsparam_string	("uid",			Opt_uid),
+	fsparam_u32	("umask",		Opt_umask),
+	fsparam_u32	("session",		Opt_session),
+	fsparam_u32	("lastblock",		Opt_lastblock),
+	fsparam_u32	("anchor",		Opt_anchor),
+	fsparam_u32	("volume",		Opt_volume),
+	fsparam_u32	("partition",		Opt_partition),
+	fsparam_u32	("fileset",		Opt_fileset),
+	fsparam_u32	("rootdir",		Opt_rootdir),
+	fsparam_flag	("utf8",		Opt_utf8),
+	fsparam_string	("iocharset",		Opt_iocharset),
+	fsparam_u32	("mode",		Opt_fmode),
+	fsparam_u32	("dmode",		Opt_dmode),
+	{}
+ };
+
+static int udf_parse_param(struct fs_context *fc, struct fs_parameter *param)
 {
-	char *p;
-	int option;
 	unsigned int uv;
-
-	uopt->novrs = 0;
-	uopt->session = 0xFFFFFFFF;
-	uopt->lastblock = 0;
-	uopt->anchor = 0;
-
-	if (!options)
-		return 1;
-
-	while ((p = strsep(&options, ",")) != NULL) {
-		substring_t args[MAX_OPT_ARGS];
-		int token;
-		unsigned n;
-		if (!*p)
-			continue;
-
-		token = match_token(p, tokens, args);
-		switch (token) {
-		case Opt_novrs:
-			uopt->novrs = 1;
-			break;
-		case Opt_bs:
-			if (match_int(&args[0], &option))
-				return 0;
-			n = option;
-			if (n != 512 && n != 1024 && n != 2048 && n != 4096)
-				return 0;
-			uopt->blocksize = n;
-			uopt->flags |= (1 << UDF_FLAG_BLOCKSIZE_SET);
-			break;
-		case Opt_unhide:
-			uopt->flags |= (1 << UDF_FLAG_UNHIDE);
-			break;
-		case Opt_undelete:
-			uopt->flags |= (1 << UDF_FLAG_UNDELETE);
-			break;
-		case Opt_noadinicb:
+	unsigned int n;
+	struct udf_options *uopt = fc->fs_private;
+	struct fs_parse_result result;
+	int token;
+	bool remount = (fc->purpose & FS_CONTEXT_FOR_RECONFIGURE);
+
+	token = fs_parse(fc, udf_param_spec, param, &result);
+	if (token < 0)
+		return token;
+
+	switch (token) {
+	case Opt_novrs:
+		uopt->flags |= (1 << UDF_FLAG_NOVRS);
+		break;
+	case Opt_bs:
+		n = result.uint_32;
+		if (n != 512 && n != 1024 && n != 2048 && n != 4096)
+			return -EINVAL;
+		uopt->blocksize = n;
+		uopt->flags |= (1 << UDF_FLAG_BLOCKSIZE_SET);
+		break;
+	case Opt_unhide:
+		uopt->flags |= (1 << UDF_FLAG_UNHIDE);
+		break;
+	case Opt_undelete:
+		uopt->flags |= (1 << UDF_FLAG_UNDELETE);
+		break;
+	case Opt_adinicb:
+		if (result.negated)
 			uopt->flags &= ~(1 << UDF_FLAG_USE_AD_IN_ICB);
-			break;
-		case Opt_adinicb:
+		else
 			uopt->flags |= (1 << UDF_FLAG_USE_AD_IN_ICB);
-			break;
-		case Opt_shortad:
-			uopt->flags |= (1 << UDF_FLAG_USE_SHORT_AD);
-			break;
-		case Opt_longad:
-			uopt->flags &= ~(1 << UDF_FLAG_USE_SHORT_AD);
-			break;
-		case Opt_gid:
-			if (match_uint(args, &uv))
-				return 0;
-			uopt->gid = make_kgid(current_user_ns(), uv);
-			if (!gid_valid(uopt->gid))
-				return 0;
+		break;
+	case Opt_shortad:
+		uopt->flags |= (1 << UDF_FLAG_USE_SHORT_AD);
+		break;
+	case Opt_longad:
+		uopt->flags &= ~(1 << UDF_FLAG_USE_SHORT_AD);
+		break;
+	case Opt_gid:
+		if (kstrtoint(param->string, 10, &uv) == 0) {
+			kgid_t gid = make_kgid(current_user_ns(), uv);
+			if (!gid_valid(gid))
+				return -EINVAL;
+			uopt->gid = gid;
 			uopt->flags |= (1 << UDF_FLAG_GID_SET);
-			break;
-		case Opt_uid:
-			if (match_uint(args, &uv))
-				return 0;
-			uopt->uid = make_kuid(current_user_ns(), uv);
-			if (!uid_valid(uopt->uid))
-				return 0;
+		} else if (!strcmp(param->string, "forget")) {
+			uopt->flags |= (1 << UDF_FLAG_GID_FORGET);
+		} else if (!strcmp(param->string, "ignore")) {
+			/* this option is superseded by gid=<number> */
+			;
+		} else {
+			return -EINVAL;
+		}
+		break;
+	case Opt_uid:
+		if (kstrtoint(param->string, 10, &uv) == 0) {
+			kuid_t uid = make_kuid(current_user_ns(), uv);
+			if (!uid_valid(uid))
+				return -EINVAL;
+			uopt->uid = uid;
 			uopt->flags |= (1 << UDF_FLAG_UID_SET);
-			break;
-		case Opt_umask:
-			if (match_octal(args, &option))
-				return 0;
-			uopt->umask = option;
-			break;
-		case Opt_nostrict:
-			uopt->flags &= ~(1 << UDF_FLAG_STRICT);
-			break;
-		case Opt_session:
-			if (match_int(args, &option))
-				return 0;
-			uopt->session = option;
-			if (!remount)
-				uopt->flags |= (1 << UDF_FLAG_SESSION_SET);
-			break;
-		case Opt_lastblock:
-			if (match_int(args, &option))
-				return 0;
-			uopt->lastblock = option;
-			if (!remount)
-				uopt->flags |= (1 << UDF_FLAG_LASTBLOCK_SET);
-			break;
-		case Opt_anchor:
-			if (match_int(args, &option))
-				return 0;
-			uopt->anchor = option;
-			break;
-		case Opt_volume:
-		case Opt_partition:
-		case Opt_fileset:
-		case Opt_rootdir:
-			/* Ignored (never implemented properly) */
-			break;
-		case Opt_utf8:
-			if (!remount) {
-				unload_nls(uopt->nls_map);
-				uopt->nls_map = NULL;
-			}
-			break;
-		case Opt_iocharset:
-			if (!remount) {
-				unload_nls(uopt->nls_map);
-				uopt->nls_map = NULL;
-			}
-			/* When nls_map is not loaded then UTF-8 is used */
-			if (!remount && strcmp(args[0].from, "utf8") != 0) {
-				uopt->nls_map = load_nls(args[0].from);
-				if (!uopt->nls_map) {
-					pr_err("iocharset %s not found\n",
-						args[0].from);
-					return 0;
-				}
-			}
-			break;
-		case Opt_uforget:
+		} else if (!strcmp(param->string, "forget")) {
 			uopt->flags |= (1 << UDF_FLAG_UID_FORGET);
-			break;
-		case Opt_uignore:
-		case Opt_gignore:
-			/* These options are superseeded by uid=<number> */
-			break;
-		case Opt_gforget:
-			uopt->flags |= (1 << UDF_FLAG_GID_FORGET);
-			break;
-		case Opt_fmode:
-			if (match_octal(args, &option))
-				return 0;
-			uopt->fmode = option & 0777;
-			break;
-		case Opt_dmode:
-			if (match_octal(args, &option))
-				return 0;
-			uopt->dmode = option & 0777;
-			break;
-		default:
-			pr_err("bad mount option \"%s\" or missing value\n", p);
-			return 0;
+		} else if (!strcmp(param->string, "ignore")) {
+			/* this option is superseded by uid=<number> */
+			;
+		} else {
+			return -EINVAL;
+		}
+		break;
+	case Opt_umask:
+		uopt->umask = result.uint_32;
+		break;
+	case Opt_nostrict:
+		uopt->flags &= ~(1 << UDF_FLAG_STRICT);
+		break;
+	case Opt_session:
+		uopt->session = result.uint_32;
+		if (!remount)
+			uopt->flags |= (1 << UDF_FLAG_SESSION_SET);
+		break;
+	case Opt_lastblock:
+		uopt->lastblock = result.uint_32;
+		if (!remount)
+			uopt->flags |= (1 << UDF_FLAG_LASTBLOCK_SET);
+		break;
+	case Opt_anchor:
+		uopt->anchor = result.uint_32;
+		break;
+	case Opt_volume:
+	case Opt_partition:
+	case Opt_fileset:
+	case Opt_rootdir:
+		/* Ignored (never implemented properly) */
+		break;
+	case Opt_utf8:
+		if (!remount) {
+			unload_nls(uopt->nls_map);
+			uopt->nls_map = NULL;
 		}
+		break;
+	case Opt_iocharset:
+		if (!remount) {
+			unload_nls(uopt->nls_map);
+			uopt->nls_map = NULL;
+		}
+		/* When nls_map is not loaded then UTF-8 is used */
+		if (!remount && strcmp(param->string, "utf8") != 0) {
+			uopt->nls_map = load_nls(param->string);
+			if (!uopt->nls_map) {
+				errorf(fc, "iocharset %s not found",
+					param->string);
+				return -EINVAL;
+			}
+		}
+		break;
+	case Opt_fmode:
+		uopt->fmode = result.uint_32 & 0777;
+		break;
+	case Opt_dmode:
+		uopt->dmode = result.uint_32 & 0777;
+		break;
+	default:
+		return -EINVAL;
 	}
-	return 1;
+	return 0;
 }
 
-static int udf_remount_fs(struct super_block *sb, int *flags, char *options)
+static int udf_reconfigure(struct fs_context *fc)
 {
-	struct udf_options uopt;
+	struct udf_options *uopt = fc->fs_private;
+	struct super_block *sb = fc->root->d_sb;
 	struct udf_sb_info *sbi = UDF_SB(sb);
+	int readonly = fc->sb_flags & SB_RDONLY;
 	int error = 0;
 
-	if (!(*flags & SB_RDONLY) && UDF_QUERY_FLAG(sb, UDF_FLAG_RW_INCOMPAT))
+	if (!readonly && UDF_QUERY_FLAG(sb, UDF_FLAG_RW_INCOMPAT))
 		return -EACCES;
 
 	sync_filesystem(sb);
 
-	uopt.flags = sbi->s_flags;
-	uopt.uid   = sbi->s_uid;
-	uopt.gid   = sbi->s_gid;
-	uopt.umask = sbi->s_umask;
-	uopt.fmode = sbi->s_fmode;
-	uopt.dmode = sbi->s_dmode;
-	uopt.nls_map = NULL;
-
-	if (!udf_parse_options(options, &uopt, true))
-		return -EINVAL;
-
 	write_lock(&sbi->s_cred_lock);
-	sbi->s_flags = uopt.flags;
-	sbi->s_uid   = uopt.uid;
-	sbi->s_gid   = uopt.gid;
-	sbi->s_umask = uopt.umask;
-	sbi->s_fmode = uopt.fmode;
-	sbi->s_dmode = uopt.dmode;
+	sbi->s_flags = uopt->flags;
+	sbi->s_uid   = uopt->uid;
+	sbi->s_gid   = uopt->gid;
+	sbi->s_umask = uopt->umask;
+	sbi->s_fmode = uopt->fmode;
+	sbi->s_dmode = uopt->dmode;
 	write_unlock(&sbi->s_cred_lock);
 
-	if ((bool)(*flags & SB_RDONLY) == sb_rdonly(sb))
+	if (readonly == sb_rdonly(sb))
 		goto out_unlock;
 
-	if (*flags & SB_RDONLY)
+	if (readonly)
 		udf_close_lvid(sb);
 	else
 		udf_open_lvid(sb);
@@ -864,7 +896,7 @@ static int udf_load_pvoldesc(struct super_block *sb, sector_t block)
 	int ret;
 	struct timestamp *ts;
 
-	outstr = kmalloc(128, GFP_NOFS);
+	outstr = kzalloc(128, GFP_KERNEL);
 	if (!outstr)
 		return -ENOMEM;
 
@@ -890,11 +922,11 @@ static int udf_load_pvoldesc(struct super_block *sb, sector_t block)
 
 	ret = udf_dstrCS0toChar(sb, outstr, 31, pvoldesc->volIdent, 32);
 	if (ret < 0) {
-		strcpy(UDF_SB(sb)->s_volume_ident, "InvalidName");
+		strscpy_pad(UDF_SB(sb)->s_volume_ident, "InvalidName");
 		pr_warn("incorrect volume identification, setting to "
 			"'InvalidName'\n");
 	} else {
-		strncpy(UDF_SB(sb)->s_volume_ident, outstr, ret);
+		strscpy_pad(UDF_SB(sb)->s_volume_ident, outstr);
 	}
 	udf_debug("volIdent[] = '%s'\n", UDF_SB(sb)->s_volume_ident);
 
@@ -1539,6 +1571,20 @@ out_bh:
 	return ret;
 }
 
+static bool udf_lvid_valid(struct super_block *sb,
+			   struct logicalVolIntegrityDesc *lvid)
+{
+	u32 parts, impuselen;
+
+	parts = le32_to_cpu(lvid->numOfPartitions);
+	impuselen = le32_to_cpu(lvid->lengthOfImpUse);
+	if (parts >= sb->s_blocksize || impuselen >= sb->s_blocksize ||
+	    sizeof(struct logicalVolIntegrityDesc) + impuselen +
+	    2 * parts * sizeof(u32) > sb->s_blocksize)
+		return false;
+	return true;
+}
+
 /*
  * Find the prevailing Logical Volume Integrity Descriptor.
  */
@@ -1549,7 +1595,6 @@ static void udf_load_logicalvolint(struct super_block *sb, struct kernel_extent_
 	struct udf_sb_info *sbi = UDF_SB(sb);
 	struct logicalVolIntegrityDesc *lvid;
 	int indirections = 0;
-	u32 parts, impuselen;
 
 	while (++indirections <= UDF_MAX_LVID_NESTING) {
 		final_bh = NULL;
@@ -1571,32 +1616,27 @@ static void udf_load_logicalvolint(struct super_block *sb, struct kernel_extent_
 		if (!final_bh)
 			return;
 
-		brelse(sbi->s_lvid_bh);
-		sbi->s_lvid_bh = final_bh;
-
 		lvid = (struct logicalVolIntegrityDesc *)final_bh->b_data;
+		if (udf_lvid_valid(sb, lvid)) {
+			brelse(sbi->s_lvid_bh);
+			sbi->s_lvid_bh = final_bh;
+		} else {
+			udf_warn(sb, "Corrupted LVID (parts=%u, impuselen=%u), "
+				 "ignoring.\n",
+				 le32_to_cpu(lvid->numOfPartitions),
+				 le32_to_cpu(lvid->lengthOfImpUse));
+		}
+
 		if (lvid->nextIntegrityExt.extLength == 0)
-			goto check;
+			return;
 
 		loc = leea_to_cpu(lvid->nextIntegrityExt);
 	}
 
 	udf_warn(sb, "Too many LVID indirections (max %u), ignoring.\n",
 		UDF_MAX_LVID_NESTING);
-out_err:
 	brelse(sbi->s_lvid_bh);
 	sbi->s_lvid_bh = NULL;
-	return;
-check:
-	parts = le32_to_cpu(lvid->numOfPartitions);
-	impuselen = le32_to_cpu(lvid->lengthOfImpUse);
-	if (parts >= sb->s_blocksize || impuselen >= sb->s_blocksize ||
-	    sizeof(struct logicalVolIntegrityDesc) + impuselen +
-	    2 * parts * sizeof(u32) > sb->s_blocksize) {
-		udf_warn(sb, "Corrupted LVID (parts=%u, impuselen=%u), "
-			 "ignoring.\n", parts, impuselen);
-		goto out_err;
-	}
 }
 
 /*
@@ -1946,7 +1986,7 @@ static int udf_load_vrs(struct super_block *sb, struct udf_options *uopt,
 		return -EINVAL;
 	}
 	sbi->s_last_block = uopt->lastblock;
-	if (!uopt->novrs) {
+	if (!UDF_QUERY_FLAG(sb, UDF_FLAG_NOVRS)) {
 		/* Check that it is NSR02 compliant */
 		nsr = udf_check_vsd(sb);
 		if (!nsr) {
@@ -2084,23 +2124,15 @@ u64 lvid_get_unique_id(struct super_block *sb)
 	return ret;
 }
 
-static int udf_fill_super(struct super_block *sb, void *options, int silent)
+static int udf_fill_super(struct super_block *sb, struct fs_context *fc)
 {
 	int ret = -EINVAL;
 	struct inode *inode = NULL;
-	struct udf_options uopt;
+	struct udf_options *uopt = fc->fs_private;
 	struct kernel_lb_addr rootdir, fileset;
 	struct udf_sb_info *sbi;
 	bool lvid_open = false;
-
-	uopt.flags = (1 << UDF_FLAG_USE_AD_IN_ICB) | (1 << UDF_FLAG_STRICT);
-	/* By default we'll use overflow[ug]id when UDF inode [ug]id == -1 */
-	uopt.uid = make_kuid(current_user_ns(), overflowuid);
-	uopt.gid = make_kgid(current_user_ns(), overflowgid);
-	uopt.umask = 0;
-	uopt.fmode = UDF_INVALID_MODE;
-	uopt.dmode = UDF_INVALID_MODE;
-	uopt.nls_map = NULL;
+	int silent = fc->sb_flags & SB_SILENT;
 
 	sbi = kzalloc(sizeof(*sbi), GFP_KERNEL);
 	if (!sbi)
@@ -2110,25 +2142,23 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent)
 
 	mutex_init(&sbi->s_alloc_mutex);
 
-	if (!udf_parse_options((char *)options, &uopt, false))
-		goto parse_options_failure;
-
 	fileset.logicalBlockNum = 0xFFFFFFFF;
 	fileset.partitionReferenceNum = 0xFFFF;
 
-	sbi->s_flags = uopt.flags;
-	sbi->s_uid = uopt.uid;
-	sbi->s_gid = uopt.gid;
-	sbi->s_umask = uopt.umask;
-	sbi->s_fmode = uopt.fmode;
-	sbi->s_dmode = uopt.dmode;
-	sbi->s_nls_map = uopt.nls_map;
+	sbi->s_flags = uopt->flags;
+	sbi->s_uid = uopt->uid;
+	sbi->s_gid = uopt->gid;
+	sbi->s_umask = uopt->umask;
+	sbi->s_fmode = uopt->fmode;
+	sbi->s_dmode = uopt->dmode;
+	sbi->s_nls_map = uopt->nls_map;
+	uopt->nls_map = NULL;
 	rwlock_init(&sbi->s_cred_lock);
 
-	if (uopt.session == 0xFFFFFFFF)
+	if (uopt->session == 0xFFFFFFFF)
 		sbi->s_session = udf_get_last_session(sb);
 	else
-		sbi->s_session = uopt.session;
+		sbi->s_session = uopt->session;
 
 	udf_debug("Multi-session=%d\n", sbi->s_session);
 
@@ -2139,16 +2169,16 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent)
 	sb->s_magic = UDF_SUPER_MAGIC;
 	sb->s_time_gran = 1000;
 
-	if (uopt.flags & (1 << UDF_FLAG_BLOCKSIZE_SET)) {
-		ret = udf_load_vrs(sb, &uopt, silent, &fileset);
+	if (uopt->flags & (1 << UDF_FLAG_BLOCKSIZE_SET)) {
+		ret = udf_load_vrs(sb, uopt, silent, &fileset);
 	} else {
-		uopt.blocksize = bdev_logical_block_size(sb->s_bdev);
-		while (uopt.blocksize <= 4096) {
-			ret = udf_load_vrs(sb, &uopt, silent, &fileset);
+		uopt->blocksize = bdev_logical_block_size(sb->s_bdev);
+		while (uopt->blocksize <= 4096) {
+			ret = udf_load_vrs(sb, uopt, silent, &fileset);
 			if (ret < 0) {
 				if (!silent && ret != -EACCES) {
 					pr_notice("Scanning with blocksize %u failed\n",
-						  uopt.blocksize);
+						  uopt->blocksize);
 				}
 				brelse(sbi->s_lvid_bh);
 				sbi->s_lvid_bh = NULL;
@@ -2161,7 +2191,7 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent)
 			} else
 				break;
 
-			uopt.blocksize <<= 1;
+			uopt->blocksize <<= 1;
 		}
 	}
 	if (ret < 0) {
@@ -2266,8 +2296,7 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent)
 
 error_out:
 	iput(sbi->s_vat_inode);
-parse_options_failure:
-	unload_nls(uopt.nls_map);
+	unload_nls(uopt->nls_map);
 	if (lvid_open)
 		udf_close_lvid(sb);
 	brelse(sbi->s_lvid_bh);
diff --git a/fs/udf/symlink.c b/fs/udf/symlink.c
index f7eaf7b145..fe03745d09 100644
--- a/fs/udf/symlink.c
+++ b/fs/udf/symlink.c
@@ -99,18 +99,17 @@ static int udf_pc_to_char(struct super_block *sb, unsigned char *from,
 
 static int udf_symlink_filler(struct file *file, struct folio *folio)
 {
-	struct page *page = &folio->page;
-	struct inode *inode = page->mapping->host;
+	struct inode *inode = folio->mapping->host;
 	struct buffer_head *bh = NULL;
 	unsigned char *symlink;
 	int err = 0;
-	unsigned char *p = page_address(page);
+	unsigned char *p = folio_address(folio);
 	struct udf_inode_info *iinfo = UDF_I(inode);
 
 	/* We don't support symlinks longer than one block */
 	if (inode->i_size > inode->i_sb->s_blocksize) {
 		err = -ENAMETOOLONG;
-		goto out_unlock;
+		goto out;
 	}
 
 	if (iinfo->i_alloc_type == ICBTAG_FLAG_AD_IN_ICB) {
@@ -120,24 +119,15 @@ static int udf_symlink_filler(struct file *file, struct folio *folio)
 		if (!bh) {
 			if (!err)
 				err = -EFSCORRUPTED;
-			goto out_err;
+			goto out;
 		}
 		symlink = bh->b_data;
 	}
 
 	err = udf_pc_to_char(inode->i_sb, symlink, inode->i_size, p, PAGE_SIZE);
 	brelse(bh);
-	if (err)
-		goto out_err;
-
-	SetPageUptodate(page);
-	unlock_page(page);
-	return 0;
-
-out_err:
-	SetPageError(page);
-out_unlock:
-	unlock_page(page);
+out:
+	folio_end_read(folio, err == 0);
 	return err;
 }
 
@@ -147,12 +137,12 @@ static int udf_symlink_getattr(struct mnt_idmap *idmap,
 {
 	struct dentry *dentry = path->dentry;
 	struct inode *inode = d_backing_inode(dentry);
-	struct page *page;
+	struct folio *folio;
 
 	generic_fillattr(&nop_mnt_idmap, request_mask, inode, stat);
-	page = read_mapping_page(inode->i_mapping, 0, NULL);
-	if (IS_ERR(page))
-		return PTR_ERR(page);
+	folio = read_mapping_folio(inode->i_mapping, 0, NULL);
+	if (IS_ERR(folio))
+		return PTR_ERR(folio);
 	/*
 	 * UDF uses non-trivial encoding of symlinks so i_size does not match
 	 * number of characters reported by readlink(2) which apparently some
@@ -162,8 +152,8 @@ static int udf_symlink_getattr(struct mnt_idmap *idmap,
 	 * let's report the length of string returned by readlink(2) for
 	 * st_size.
 	 */
-	stat->size = strlen(page_address(page));
-	put_page(page);
+	stat->size = strlen(folio_address(folio));
+	folio_put(folio);
 
 	return 0;
 }
diff --git a/fs/udf/udf_sb.h b/fs/udf/udf_sb.h
index f9a60bc1ab..08ec8756b9 100644
--- a/fs/udf/udf_sb.h
+++ b/fs/udf/udf_sb.h
@@ -23,6 +23,7 @@
 #define UDF_FLAG_STRICT			5
 #define UDF_FLAG_UNDELETE		6
 #define UDF_FLAG_UNHIDE			7
+#define UDF_FLAG_NOVRS			8
 #define UDF_FLAG_UID_FORGET     11    /* save -1 for uid to disk */
 #define UDF_FLAG_GID_FORGET     12
 #define UDF_FLAG_UID_SET	13
diff --git a/fs/udf/udftime.c b/fs/udf/udftime.c
index 758163af39..78ecc63360 100644
--- a/fs/udf/udftime.c
+++ b/fs/udf/udftime.c
@@ -46,13 +46,18 @@ udf_disk_stamp_to_time(struct timespec64 *dest, struct timestamp src)
 	dest->tv_sec = mktime64(year, src.month, src.day, src.hour, src.minute,
 			src.second);
 	dest->tv_sec -= offset * 60;
-	dest->tv_nsec = 1000 * (src.centiseconds * 10000 +
-			src.hundredsOfMicroseconds * 100 + src.microseconds);
+
 	/*
 	 * Sanitize nanosecond field since reportedly some filesystems are
 	 * recorded with bogus sub-second values.
 	 */
-	dest->tv_nsec %= NSEC_PER_SEC;
+	if (src.centiseconds < 100 && src.hundredsOfMicroseconds < 100 &&
+	    src.microseconds < 100) {
+		dest->tv_nsec = 1000 * (src.centiseconds * 10000 +
+			src.hundredsOfMicroseconds * 100 + src.microseconds);
+	} else {
+		dest->tv_nsec = 0;
+	}
 }
 
 void
diff --git a/fs/ufs/super.c b/fs/ufs/super.c
index a480810cd4..44666afc62 100644
--- a/fs/ufs/super.c
+++ b/fs/ufs/super.c
@@ -1470,8 +1470,7 @@ static int __init init_inodecache(void)
 {
 	ufs_inode_cachep = kmem_cache_create_usercopy("ufs_inode_cache",
 				sizeof(struct ufs_inode_info), 0,
-				(SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD|
-					SLAB_ACCOUNT),
+				(SLAB_RECLAIM_ACCOUNT | SLAB_ACCOUNT),
 				offsetof(struct ufs_inode_info, i_u1.i_symlink),
 				sizeof_field(struct ufs_inode_info,
 					i_u1.i_symlink),
diff --git a/fs/unicode/Makefile b/fs/unicode/Makefile
index 0e51c0025a..e309afe2b2 100644
--- a/fs/unicode/Makefile
+++ b/fs/unicode/Makefile
@@ -18,13 +18,13 @@ ifdef REGENERATE_UTF8DATA
 
 quiet_cmd_utf8data = GEN     $@
       cmd_utf8data = $< \
-		-a $(srctree)/$(src)/DerivedAge.txt \
-		-c $(srctree)/$(src)/DerivedCombiningClass.txt \
-		-p $(srctree)/$(src)/DerivedCoreProperties.txt \
-		-d $(srctree)/$(src)/UnicodeData.txt \
-		-f $(srctree)/$(src)/CaseFolding.txt \
-		-n $(srctree)/$(src)/NormalizationCorrections.txt \
-		-t $(srctree)/$(src)/NormalizationTest.txt \
+		-a $(src)/DerivedAge.txt \
+		-c $(src)/DerivedCombiningClass.txt \
+		-p $(src)/DerivedCoreProperties.txt \
+		-d $(src)/UnicodeData.txt \
+		-f $(src)/CaseFolding.txt \
+		-n $(src)/NormalizationCorrections.txt \
+		-t $(src)/NormalizationTest.txt \
 		-o $@
 
 $(obj)/utf8data.c: $(obj)/mkutf8data $(filter %.txt, $(cmd_utf8data)) FORCE
diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c
index 13ef4e1fc1..17e409ceaa 100644
--- a/fs/userfaultfd.c
+++ b/fs/userfaultfd.c
@@ -31,6 +31,7 @@
 #include <linux/hugetlb.h>
 #include <linux/swapops.h>
 #include <linux/miscdevice.h>
+#include <linux/uio.h>
 
 static int sysctl_unprivileged_userfaultfd __read_mostly;
 
@@ -50,45 +51,6 @@ static struct ctl_table vm_userfaultfd_table[] = {
 
 static struct kmem_cache *userfaultfd_ctx_cachep __ro_after_init;
 
-/*
- * Start with fault_pending_wqh and fault_wqh so they're more likely
- * to be in the same cacheline.
- *
- * Locking order:
- *	fd_wqh.lock
- *		fault_pending_wqh.lock
- *			fault_wqh.lock
- *		event_wqh.lock
- *
- * To avoid deadlocks, IRQs must be disabled when taking any of the above locks,
- * since fd_wqh.lock is taken by aio_poll() while it's holding a lock that's
- * also taken in IRQ context.
- */
-struct userfaultfd_ctx {
-	/* waitqueue head for the pending (i.e. not read) userfaults */
-	wait_queue_head_t fault_pending_wqh;
-	/* waitqueue head for the userfaults */
-	wait_queue_head_t fault_wqh;
-	/* waitqueue head for the pseudo fd to wakeup poll/read */
-	wait_queue_head_t fd_wqh;
-	/* waitqueue head for events */
-	wait_queue_head_t event_wqh;
-	/* a refile sequence protected by fault_pending_wqh lock */
-	seqcount_spinlock_t refile_seq;
-	/* pseudo fd refcounting */
-	refcount_t refcount;
-	/* userfaultfd syscall flags */
-	unsigned int flags;
-	/* features requested from the userspace */
-	unsigned int features;
-	/* released */
-	bool released;
-	/* memory mappings are changing because of non-cooperative event */
-	atomic_t mmap_changing;
-	/* mm with one ore more vmas attached to this userfaultfd_ctx */
-	struct mm_struct *mm;
-};
-
 struct userfaultfd_fork_ctx {
 	struct userfaultfd_ctx *orig;
 	struct userfaultfd_ctx *new;
@@ -321,7 +283,7 @@ static inline bool userfaultfd_huge_must_wait(struct userfaultfd_ctx *ctx,
 /*
  * Verify the pagetables are still not ok after having reigstered into
  * the fault_pending_wqh to avoid userland having to UFFDIO_WAKE any
- * userfault that has already been resolved, if userfaultfd_read and
+ * userfault that has already been resolved, if userfaultfd_read_iter and
  * UFFDIO_COPY|ZEROPAGE are being run simultaneously on two different
  * threads.
  */
@@ -696,7 +658,10 @@ int dup_userfaultfd(struct vm_area_struct *vma, struct list_head *fcs)
 	struct userfaultfd_fork_ctx *fctx;
 
 	octx = vma->vm_userfaultfd_ctx.ctx;
-	if (!octx || !(octx->features & UFFD_FEATURE_EVENT_FORK)) {
+	if (!octx)
+		return 0;
+
+	if (!(octx->features & UFFD_FEATURE_EVENT_FORK)) {
 		vma_start_write(vma);
 		vma->vm_userfaultfd_ctx = NULL_VM_UFFD_CTX;
 		userfaultfd_set_vm_flags(vma, vma->vm_flags & ~__VM_UFFD_FLAGS);
@@ -724,12 +689,15 @@ int dup_userfaultfd(struct vm_area_struct *vma, struct list_head *fcs)
 		ctx->flags = octx->flags;
 		ctx->features = octx->features;
 		ctx->released = false;
+		init_rwsem(&ctx->map_changing_lock);
 		atomic_set(&ctx->mmap_changing, 0);
 		ctx->mm = vma->vm_mm;
 		mmgrab(ctx->mm);
 
 		userfaultfd_ctx_get(octx);
+		down_write(&octx->map_changing_lock);
 		atomic_inc(&octx->mmap_changing);
+		up_write(&octx->map_changing_lock);
 		fctx->orig = octx;
 		fctx->new = ctx;
 		list_add_tail(&fctx->list, fcs);
@@ -776,7 +744,9 @@ void mremap_userfaultfd_prep(struct vm_area_struct *vma,
 	if (ctx->features & UFFD_FEATURE_EVENT_REMAP) {
 		vm_ctx->ctx = ctx;
 		userfaultfd_ctx_get(ctx);
+		down_write(&ctx->map_changing_lock);
 		atomic_inc(&ctx->mmap_changing);
+		up_write(&ctx->map_changing_lock);
 	} else {
 		/* Drop uffd context if remap feature not enabled */
 		vma_start_write(vma);
@@ -822,7 +792,9 @@ bool userfaultfd_remove(struct vm_area_struct *vma,
 		return true;
 
 	userfaultfd_ctx_get(ctx);
+	down_write(&ctx->map_changing_lock);
 	atomic_inc(&ctx->mmap_changing);
+	up_write(&ctx->map_changing_lock);
 	mmap_read_unlock(mm);
 
 	msg_init(&ewq.msg);
@@ -864,7 +836,9 @@ int userfaultfd_unmap_prep(struct vm_area_struct *vma, unsigned long start,
 		return -ENOMEM;
 
 	userfaultfd_ctx_get(ctx);
+	down_write(&ctx->map_changing_lock);
 	atomic_inc(&ctx->mmap_changing);
+	up_write(&ctx->map_changing_lock);
 	unmap_ctx->ctx = ctx;
 	unmap_ctx->start = start;
 	unmap_ctx->end = end;
@@ -1211,34 +1185,34 @@ static ssize_t userfaultfd_ctx_read(struct userfaultfd_ctx *ctx, int no_wait,
 	return ret;
 }
 
-static ssize_t userfaultfd_read(struct file *file, char __user *buf,
-				size_t count, loff_t *ppos)
+static ssize_t userfaultfd_read_iter(struct kiocb *iocb, struct iov_iter *to)
 {
+	struct file *file = iocb->ki_filp;
 	struct userfaultfd_ctx *ctx = file->private_data;
 	ssize_t _ret, ret = 0;
 	struct uffd_msg msg;
-	int no_wait = file->f_flags & O_NONBLOCK;
 	struct inode *inode = file_inode(file);
+	bool no_wait;
 
 	if (!userfaultfd_is_initialized(ctx))
 		return -EINVAL;
 
+	no_wait = file->f_flags & O_NONBLOCK || iocb->ki_flags & IOCB_NOWAIT;
 	for (;;) {
-		if (count < sizeof(msg))
+		if (iov_iter_count(to) < sizeof(msg))
 			return ret ? ret : -EINVAL;
 		_ret = userfaultfd_ctx_read(ctx, no_wait, &msg, inode);
 		if (_ret < 0)
 			return ret ? ret : _ret;
-		if (copy_to_user((__u64 __user *) buf, &msg, sizeof(msg)))
+		_ret = !copy_to_iter_full(&msg, sizeof(msg), to);
+		if (_ret)
 			return ret ? ret : -EFAULT;
 		ret += sizeof(msg);
-		buf += sizeof(msg);
-		count -= sizeof(msg);
 		/*
 		 * Allow to read more than one fault at time but only
 		 * block if waiting for the very first one.
 		 */
-		no_wait = O_NONBLOCK;
+		no_wait = true;
 	}
 }
 
@@ -1752,9 +1726,8 @@ static int userfaultfd_copy(struct userfaultfd_ctx *ctx,
 	if (uffdio_copy.mode & UFFDIO_COPY_MODE_WP)
 		flags |= MFILL_ATOMIC_WP;
 	if (mmget_not_zero(ctx->mm)) {
-		ret = mfill_atomic_copy(ctx->mm, uffdio_copy.dst, uffdio_copy.src,
-					uffdio_copy.len, &ctx->mmap_changing,
-					flags);
+		ret = mfill_atomic_copy(ctx, uffdio_copy.dst, uffdio_copy.src,
+					uffdio_copy.len, flags);
 		mmput(ctx->mm);
 	} else {
 		return -ESRCH;
@@ -1804,9 +1777,8 @@ static int userfaultfd_zeropage(struct userfaultfd_ctx *ctx,
 		goto out;
 
 	if (mmget_not_zero(ctx->mm)) {
-		ret = mfill_atomic_zeropage(ctx->mm, uffdio_zeropage.range.start,
-					   uffdio_zeropage.range.len,
-					   &ctx->mmap_changing);
+		ret = mfill_atomic_zeropage(ctx, uffdio_zeropage.range.start,
+					   uffdio_zeropage.range.len);
 		mmput(ctx->mm);
 	} else {
 		return -ESRCH;
@@ -1861,9 +1833,8 @@ static int userfaultfd_writeprotect(struct userfaultfd_ctx *ctx,
 		return -EINVAL;
 
 	if (mmget_not_zero(ctx->mm)) {
-		ret = mwriteprotect_range(ctx->mm, uffdio_wp.range.start,
-					  uffdio_wp.range.len, mode_wp,
-					  &ctx->mmap_changing);
+		ret = mwriteprotect_range(ctx, uffdio_wp.range.start,
+					  uffdio_wp.range.len, mode_wp);
 		mmput(ctx->mm);
 	} else {
 		return -ESRCH;
@@ -1913,9 +1884,8 @@ static int userfaultfd_continue(struct userfaultfd_ctx *ctx, unsigned long arg)
 		flags |= MFILL_ATOMIC_WP;
 
 	if (mmget_not_zero(ctx->mm)) {
-		ret = mfill_atomic_continue(ctx->mm, uffdio_continue.range.start,
-					    uffdio_continue.range.len,
-					    &ctx->mmap_changing, flags);
+		ret = mfill_atomic_continue(ctx, uffdio_continue.range.start,
+					    uffdio_continue.range.len, flags);
 		mmput(ctx->mm);
 	} else {
 		return -ESRCH;
@@ -1968,9 +1938,8 @@ static inline int userfaultfd_poison(struct userfaultfd_ctx *ctx, unsigned long
 		goto out;
 
 	if (mmget_not_zero(ctx->mm)) {
-		ret = mfill_atomic_poison(ctx->mm, uffdio_poison.range.start,
-					  uffdio_poison.range.len,
-					  &ctx->mmap_changing, 0);
+		ret = mfill_atomic_poison(ctx, uffdio_poison.range.start,
+					  uffdio_poison.range.len, 0);
 		mmput(ctx->mm);
 	} else {
 		return -ESRCH;
@@ -2044,16 +2013,8 @@ static int userfaultfd_move(struct userfaultfd_ctx *ctx,
 		return -EINVAL;
 
 	if (mmget_not_zero(mm)) {
-		mmap_read_lock(mm);
-
-		/* Re-check after taking mmap_lock */
-		if (likely(!atomic_read(&ctx->mmap_changing)))
-			ret = move_pages(ctx, mm, uffdio_move.dst, uffdio_move.src,
-					 uffdio_move.len, uffdio_move.mode);
-		else
-			ret = -EINVAL;
-
-		mmap_read_unlock(mm);
+		ret = move_pages(ctx, uffdio_move.dst, uffdio_move.src,
+				 uffdio_move.len, uffdio_move.mode);
 		mmput(mm);
 	} else {
 		return -ESRCH;
@@ -2096,7 +2057,7 @@ static int userfaultfd_api(struct userfaultfd_ctx *ctx,
 		goto out;
 	features = uffdio_api.features;
 	ret = -EINVAL;
-	if (uffdio_api.api != UFFD_API || (features & ~UFFD_API_FEATURES))
+	if (uffdio_api.api != UFFD_API)
 		goto err_out;
 	ret = -EPERM;
 	if ((features & UFFD_FEATURE_EVENT_FORK) && !capable(CAP_SYS_PTRACE))
@@ -2120,6 +2081,11 @@ static int userfaultfd_api(struct userfaultfd_ctx *ctx,
 	uffdio_api.features &= ~UFFD_FEATURE_WP_UNPOPULATED;
 	uffdio_api.features &= ~UFFD_FEATURE_WP_ASYNC;
 #endif
+
+	ret = -EINVAL;
+	if (features & ~uffdio_api.features)
+		goto err_out;
+
 	uffdio_api.ioctls = UFFD_API_IOCTLS;
 	ret = -EFAULT;
 	if (copy_to_user(buf, &uffdio_api, sizeof(uffdio_api)))
@@ -2219,7 +2185,7 @@ static const struct file_operations userfaultfd_fops = {
 #endif
 	.release	= userfaultfd_release,
 	.poll		= userfaultfd_poll,
-	.read		= userfaultfd_read,
+	.read_iter	= userfaultfd_read_iter,
 	.unlocked_ioctl = userfaultfd_ioctl,
 	.compat_ioctl	= compat_ptr_ioctl,
 	.llseek		= noop_llseek,
@@ -2239,6 +2205,7 @@ static void init_once_userfaultfd_ctx(void *mem)
 static int new_userfaultfd(int flags)
 {
 	struct userfaultfd_ctx *ctx;
+	struct file *file;
 	int fd;
 
 	BUG_ON(!current->mm);
@@ -2259,18 +2226,29 @@ static int new_userfaultfd(int flags)
 	ctx->flags = flags;
 	ctx->features = 0;
 	ctx->released = false;
+	init_rwsem(&ctx->map_changing_lock);
 	atomic_set(&ctx->mmap_changing, 0);
 	ctx->mm = current->mm;
-	/* prevent the mm struct to be freed */
-	mmgrab(ctx->mm);
+
+	fd = get_unused_fd_flags(flags & UFFD_SHARED_FCNTL_FLAGS);
+	if (fd < 0)
+		goto err_out;
 
 	/* Create a new inode so that the LSM can block the creation.  */
-	fd = anon_inode_create_getfd("[userfaultfd]", &userfaultfd_fops, ctx,
+	file = anon_inode_create_getfile("[userfaultfd]", &userfaultfd_fops, ctx,
 			O_RDONLY | (flags & UFFD_SHARED_FCNTL_FLAGS), NULL);
-	if (fd < 0) {
-		mmdrop(ctx->mm);
-		kmem_cache_free(userfaultfd_ctx_cachep, ctx);
+	if (IS_ERR(file)) {
+		put_unused_fd(fd);
+		fd = PTR_ERR(file);
+		goto err_out;
 	}
+	/* prevent the mm struct to be freed */
+	mmgrab(ctx->mm);
+	file->f_mode |= FMODE_NOWAIT;
+	fd_install(fd, file);
+	return fd;
+err_out:
+	kmem_cache_free(userfaultfd_ctx_cachep, ctx);
 	return fd;
 }
 
diff --git a/fs/vboxsf/super.c b/fs/vboxsf/super.c
index 9848af7821..ffb1d565da 100644
--- a/fs/vboxsf/super.c
+++ b/fs/vboxsf/super.c
@@ -155,7 +155,7 @@ static int vboxsf_fill_super(struct super_block *sb, struct fs_context *fc)
 		}
 	}
 
-	sbi->bdi_id = ida_simple_get(&vboxsf_bdi_ida, 0, 0, GFP_KERNEL);
+	sbi->bdi_id = ida_alloc(&vboxsf_bdi_ida, GFP_KERNEL);
 	if (sbi->bdi_id < 0) {
 		err = sbi->bdi_id;
 		goto fail_free;
@@ -221,7 +221,7 @@ fail_unmap:
 	vboxsf_unmap_folder(sbi->root);
 fail_free:
 	if (sbi->bdi_id >= 0)
-		ida_simple_remove(&vboxsf_bdi_ida, sbi->bdi_id);
+		ida_free(&vboxsf_bdi_ida, sbi->bdi_id);
 	if (sbi->nls)
 		unload_nls(sbi->nls);
 fail_destroy_idr:
@@ -269,7 +269,7 @@ static void vboxsf_put_super(struct super_block *sb)
 
 	vboxsf_unmap_folder(sbi->root);
 	if (sbi->bdi_id >= 0)
-		ida_simple_remove(&vboxsf_bdi_ida, sbi->bdi_id);
+		ida_free(&vboxsf_bdi_ida, sbi->bdi_id);
 	if (sbi->nls)
 		unload_nls(sbi->nls);
 
@@ -340,8 +340,7 @@ static int vboxsf_setup(void)
 	vboxsf_inode_cachep =
 		kmem_cache_create("vboxsf_inode_cache",
 				  sizeof(struct vboxsf_inode), 0,
-				  (SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD |
-				   SLAB_ACCOUNT),
+				  SLAB_RECLAIM_ACCOUNT | SLAB_ACCOUNT,
 				  vboxsf_inode_init_once);
 	if (!vboxsf_inode_cachep) {
 		err = -ENOMEM;
diff --git a/fs/vboxsf/utils.c b/fs/vboxsf/utils.c
index 72ac9320e6..9515bbf0b5 100644
--- a/fs/vboxsf/utils.c
+++ b/fs/vboxsf/utils.c
@@ -440,7 +440,6 @@ int vboxsf_nlscpy(struct vboxsf_sbi *sbi, char *name, size_t name_bound_len,
 {
 	const char *in;
 	char *out;
-	size_t out_len;
 	size_t out_bound_len;
 	size_t in_bound_len;
 
@@ -448,7 +447,6 @@ int vboxsf_nlscpy(struct vboxsf_sbi *sbi, char *name, size_t name_bound_len,
 	in_bound_len = utf8_len;
 
 	out = name;
-	out_len = 0;
 	/* Reserve space for terminating 0 */
 	out_bound_len = name_bound_len - 1;
 
@@ -469,7 +467,6 @@ int vboxsf_nlscpy(struct vboxsf_sbi *sbi, char *name, size_t name_bound_len,
 
 		out += nb;
 		out_bound_len -= nb;
-		out_len += nb;
 	}
 
 	*out = 0;
diff --git a/fs/verity/fsverity_private.h b/fs/verity/fsverity_private.h
index a6a6b27492..b3506f56e1 100644
--- a/fs/verity/fsverity_private.h
+++ b/fs/verity/fsverity_private.h
@@ -69,7 +69,6 @@ struct fsverity_info {
 	u8 file_digest[FS_VERITY_MAX_DIGEST_SIZE];
 	const struct inode *inode;
 	unsigned long *hash_block_verified;
-	spinlock_t hash_page_init_lock;
 };
 
 #define FS_VERITY_MAX_SIGNATURE_SIZE	(FS_VERITY_MAX_DESCRIPTOR_SIZE - \
diff --git a/fs/verity/init.c b/fs/verity/init.c
index cb2c9aac61..f440f0e61e 100644
--- a/fs/verity/init.c
+++ b/fs/verity/init.c
@@ -10,8 +10,6 @@
 #include <linux/ratelimit.h>
 
 #ifdef CONFIG_SYSCTL
-static struct ctl_table_header *fsverity_sysctl_header;
-
 static struct ctl_table fsverity_sysctl_table[] = {
 #ifdef CONFIG_FS_VERITY_BUILTIN_SIGNATURES
 	{
@@ -28,10 +26,7 @@ static struct ctl_table fsverity_sysctl_table[] = {
 
 static void __init fsverity_init_sysctl(void)
 {
-	fsverity_sysctl_header = register_sysctl("fs/verity",
-						 fsverity_sysctl_table);
-	if (!fsverity_sysctl_header)
-		panic("fsverity sysctl registration failed");
+	register_sysctl_init("fs/verity", fsverity_sysctl_table);
 }
 #else /* CONFIG_SYSCTL */
 static inline void fsverity_init_sysctl(void)
diff --git a/fs/verity/measure.c b/fs/verity/measure.c
index bf7a5f4ccc..3969d54158 100644
--- a/fs/verity/measure.c
+++ b/fs/verity/measure.c
@@ -159,9 +159,9 @@ __bpf_kfunc int bpf_get_fsverity_digest(struct file *file, struct bpf_dynptr_ker
 
 __bpf_kfunc_end_defs();
 
-BTF_SET8_START(fsverity_set_ids)
+BTF_KFUNCS_START(fsverity_set_ids)
 BTF_ID_FLAGS(func, bpf_get_fsverity_digest, KF_TRUSTED_ARGS)
-BTF_SET8_END(fsverity_set_ids)
+BTF_KFUNCS_END(fsverity_set_ids)
 
 static int bpf_get_fsverity_digest_filter(const struct bpf_prog *prog, u32 kfunc_id)
 {
diff --git a/fs/verity/open.c b/fs/verity/open.c
index 6c31a871b8..fdeb95eca3 100644
--- a/fs/verity/open.c
+++ b/fs/verity/open.c
@@ -239,7 +239,6 @@ struct fsverity_info *fsverity_create_info(const struct inode *inode,
 			err = -ENOMEM;
 			goto fail;
 		}
-		spin_lock_init(&vi->hash_page_init_lock);
 	}
 
 	return vi;
diff --git a/fs/verity/verify.c b/fs/verity/verify.c
index 904ccd7e8e..4fcad0825a 100644
--- a/fs/verity/verify.c
+++ b/fs/verity/verify.c
@@ -19,7 +19,6 @@ static struct workqueue_struct *fsverity_read_workqueue;
 static bool is_hash_block_verified(struct fsverity_info *vi, struct page *hpage,
 				   unsigned long hblock_idx)
 {
-	bool verified;
 	unsigned int blocks_per_page;
 	unsigned int i;
 
@@ -43,12 +42,20 @@ static bool is_hash_block_verified(struct fsverity_info *vi, struct page *hpage,
 	 * re-instantiated from the backing storage are re-verified.  To do
 	 * this, we use PG_checked again, but now it doesn't really mean
 	 * "checked".  Instead, now it just serves as an indicator for whether
-	 * the hash page is newly instantiated or not.
+	 * the hash page is newly instantiated or not.  If the page is new, as
+	 * indicated by PG_checked=0, we clear the bitmap bits for the page's
+	 * blocks since they are untrustworthy, then set PG_checked=1.
+	 * Otherwise we return the bitmap bit for the requested block.
 	 *
-	 * The first thread that sees PG_checked=0 must clear the corresponding
-	 * bitmap bits, then set PG_checked=1.  This requires a spinlock.  To
-	 * avoid having to take this spinlock in the common case of
-	 * PG_checked=1, we start with an opportunistic lockless read.
+	 * Multiple threads may execute this code concurrently on the same page.
+	 * This is safe because we use memory barriers to ensure that if a
+	 * thread sees PG_checked=1, then it also sees the associated bitmap
+	 * clearing to have occurred.  Also, all writes and their corresponding
+	 * reads are atomic, and all writes are safe to repeat in the event that
+	 * multiple threads get into the PG_checked=0 section.  (Clearing a
+	 * bitmap bit again at worst causes a hash block to be verified
+	 * redundantly.  That event should be very rare, so it's not worth using
+	 * a lock to avoid.  Setting PG_checked again has no effect.)
 	 */
 	if (PageChecked(hpage)) {
 		/*
@@ -58,24 +65,17 @@ static bool is_hash_block_verified(struct fsverity_info *vi, struct page *hpage,
 		smp_rmb();
 		return test_bit(hblock_idx, vi->hash_block_verified);
 	}
-	spin_lock(&vi->hash_page_init_lock);
-	if (PageChecked(hpage)) {
-		verified = test_bit(hblock_idx, vi->hash_block_verified);
-	} else {
-		blocks_per_page = vi->tree_params.blocks_per_page;
-		hblock_idx = round_down(hblock_idx, blocks_per_page);
-		for (i = 0; i < blocks_per_page; i++)
-			clear_bit(hblock_idx + i, vi->hash_block_verified);
-		/*
-		 * A write memory barrier is needed here to give RELEASE
-		 * semantics to the below SetPageChecked() operation.
-		 */
-		smp_wmb();
-		SetPageChecked(hpage);
-		verified = false;
-	}
-	spin_unlock(&vi->hash_page_init_lock);
-	return verified;
+	blocks_per_page = vi->tree_params.blocks_per_page;
+	hblock_idx = round_down(hblock_idx, blocks_per_page);
+	for (i = 0; i < blocks_per_page; i++)
+		clear_bit(hblock_idx + i, vi->hash_block_verified);
+	/*
+	 * A write memory barrier is needed here to give RELEASE semantics to
+	 * the below SetPageChecked() operation.
+	 */
+	smp_wmb();
+	SetPageChecked(hpage);
+	return false;
 }
 
 /*
diff --git a/fs/xattr.c b/fs/xattr.c
index 09d9276034..f8b643f91a 100644
--- a/fs/xattr.c
+++ b/fs/xattr.c
@@ -16,7 +16,6 @@
 #include <linux/mount.h>
 #include <linux/namei.h>
 #include <linux/security.h>
-#include <linux/evm.h>
 #include <linux/syscalls.h>
 #include <linux/export.h>
 #include <linux/fsnotify.h>
@@ -552,11 +551,11 @@ __vfs_removexattr_locked(struct mnt_idmap *idmap,
 		goto out;
 
 	error = __vfs_removexattr(idmap, dentry, name);
+	if (error)
+		return error;
 
-	if (!error) {
-		fsnotify_xattr(dentry);
-		evm_inode_post_removexattr(dentry, name);
-	}
+	fsnotify_xattr(dentry);
+	security_inode_post_removexattr(dentry, name);
 
 out:
 	return error;
diff --git a/fs/xfs/Kconfig b/fs/xfs/Kconfig
index 567fb37274..d41edd3038 100644
--- a/fs/xfs/Kconfig
+++ b/fs/xfs/Kconfig
@@ -124,12 +124,24 @@ config XFS_DRAIN_INTENTS
 	bool
 	select JUMP_LABEL if HAVE_ARCH_JUMP_LABEL
 
+config XFS_LIVE_HOOKS
+	bool
+	select JUMP_LABEL if HAVE_ARCH_JUMP_LABEL
+
+config XFS_MEMORY_BUFS
+	bool
+
+config XFS_BTREE_IN_MEM
+	bool
+
 config XFS_ONLINE_SCRUB
 	bool "XFS online metadata check support"
 	default n
 	depends on XFS_FS
 	depends on TMPFS && SHMEM
+	select XFS_LIVE_HOOKS
 	select XFS_DRAIN_INTENTS
+	select XFS_MEMORY_BUFS
 	help
 	  If you say Y here you will be able to check metadata on a
 	  mounted XFS filesystem.  This feature is intended to reduce
@@ -164,6 +176,7 @@ config XFS_ONLINE_REPAIR
 	bool "XFS online metadata repair support"
 	default n
 	depends on XFS_FS && XFS_ONLINE_SCRUB
+	select XFS_BTREE_IN_MEM
 	help
 	  If you say Y here you will be able to repair metadata on a
 	  mounted XFS filesystem.  This feature is intended to reduce
diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile
index fbe3cdc790..c50447548d 100644
--- a/fs/xfs/Makefile
+++ b/fs/xfs/Makefile
@@ -4,8 +4,8 @@
 # All Rights Reserved.
 #
 
-ccflags-y += -I $(srctree)/$(src)		# needed for trace events
-ccflags-y += -I $(srctree)/$(src)/libxfs
+ccflags-y += -I $(src)		# needed for trace events
+ccflags-y += -I $(src)/libxfs
 
 obj-$(CONFIG_XFS_FS)		+= xfs.o
 
@@ -34,6 +34,7 @@ xfs-y				+= $(addprefix libxfs/, \
 				   xfs_dir2_node.o \
 				   xfs_dir2_sf.o \
 				   xfs_dquot_buf.o \
+				   xfs_exchmaps.o \
 				   xfs_ialloc.o \
 				   xfs_ialloc_btree.o \
 				   xfs_iext_tree.o \
@@ -41,6 +42,7 @@ xfs-y				+= $(addprefix libxfs/, \
 				   xfs_inode_buf.o \
 				   xfs_log_rlimit.o \
 				   xfs_ag_resv.o \
+				   xfs_parent.o \
 				   xfs_rmap.o \
 				   xfs_rmap_btree.o \
 				   xfs_refcount.o \
@@ -49,6 +51,7 @@ xfs-y				+= $(addprefix libxfs/, \
 				   xfs_symlink_remote.o \
 				   xfs_trans_inode.o \
 				   xfs_trans_resv.o \
+				   xfs_trans_space.o \
 				   xfs_types.o \
 				   )
 # xfs_rtbitmap is shared with libxfs
@@ -67,6 +70,7 @@ xfs-y				+= xfs_aops.o \
 				   xfs_dir2_readdir.o \
 				   xfs_discard.o \
 				   xfs_error.o \
+				   xfs_exchrange.o \
 				   xfs_export.o \
 				   xfs_extent_busy.o \
 				   xfs_file.o \
@@ -74,6 +78,7 @@ xfs-y				+= xfs_aops.o \
 				   xfs_fsmap.o \
 				   xfs_fsops.o \
 				   xfs_globals.o \
+				   xfs_handle.o \
 				   xfs_health.o \
 				   xfs_icache.o \
 				   xfs_ioctl.o \
@@ -92,8 +97,7 @@ xfs-y				+= xfs_aops.o \
 				   xfs_symlink.o \
 				   xfs_sysfs.o \
 				   xfs_trans.o \
-				   xfs_xattr.o \
-				   kmem.o
+				   xfs_xattr.o
 
 # low-level transaction/log code
 xfs-y				+= xfs_log.o \
@@ -102,6 +106,7 @@ xfs-y				+= xfs_log.o \
 				   xfs_buf_item.o \
 				   xfs_buf_item_recover.o \
 				   xfs_dquot_item_recover.o \
+				   xfs_exchmaps_item.o \
 				   xfs_extfree_item.o \
 				   xfs_attr_item.o \
 				   xfs_icreate_item.o \
@@ -137,6 +142,9 @@ xfs-$(CONFIG_FS_DAX)		+= xfs_notify_failure.o
 endif
 
 xfs-$(CONFIG_XFS_DRAIN_INTENTS)	+= xfs_drain.o
+xfs-$(CONFIG_XFS_LIVE_HOOKS)	+= xfs_hooks.o
+xfs-$(CONFIG_XFS_MEMORY_BUFS)	+= xfs_buf_mem.o
+xfs-$(CONFIG_XFS_BTREE_IN_MEM)	+= libxfs/xfs_btree_mem.o
 
 # online scrub/repair
 ifeq ($(CONFIG_XFS_ONLINE_SCRUB),y)
@@ -155,10 +163,14 @@ xfs-y				+= $(addprefix scrub/, \
 				   common.o \
 				   dabtree.o \
 				   dir.o \
+				   dirtree.o \
 				   fscounters.o \
 				   health.o \
 				   ialloc.o \
 				   inode.o \
+				   iscan.o \
+				   listxattr.o \
+				   nlinks.o \
 				   parent.o \
 				   readdir.o \
 				   refcount.o \
@@ -166,6 +178,7 @@ xfs-y				+= $(addprefix scrub/, \
 				   scrub.o \
 				   symlink.o \
 				   xfarray.o \
+				   xfblob.o \
 				   xfile.o \
 				   )
 
@@ -179,6 +192,7 @@ xfs-$(CONFIG_XFS_RT)		+= $(addprefix scrub/, \
 xfs-$(CONFIG_XFS_QUOTA)		+= $(addprefix scrub/, \
 				   dqiterate.o \
 				   quota.o \
+				   quotacheck.o \
 				   )
 
 # online repair
@@ -186,22 +200,37 @@ ifeq ($(CONFIG_XFS_ONLINE_REPAIR),y)
 xfs-y				+= $(addprefix scrub/, \
 				   agheader_repair.o \
 				   alloc_repair.o \
+				   attr_repair.o \
 				   bmap_repair.o \
 				   cow_repair.o \
+				   dir_repair.o \
+				   dirtree_repair.o \
+				   findparent.o \
+				   fscounters_repair.o \
 				   ialloc_repair.o \
 				   inode_repair.o \
 				   newbt.o \
+				   nlinks_repair.o \
+				   orphanage.o \
+				   parent_repair.o \
+				   rcbag_btree.o \
+				   rcbag.o \
 				   reap.o \
 				   refcount_repair.o \
 				   repair.o \
+				   rmap_repair.o \
+				   symlink_repair.o \
+				   tempfile.o \
 				   )
 
 xfs-$(CONFIG_XFS_RT)		+= $(addprefix scrub/, \
 				   rtbitmap_repair.o \
+				   rtsummary_repair.o \
 				   )
 
 xfs-$(CONFIG_XFS_QUOTA)		+= $(addprefix scrub/, \
 				   quota_repair.o \
+				   quotacheck_repair.o \
 				   )
 endif
 endif
diff --git a/fs/xfs/kmem.c b/fs/xfs/kmem.c
deleted file mode 100644
index c557a030ac..0000000000
--- a/fs/xfs/kmem.c
+++ /dev/null
@@ -1,30 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0
-/*
- * Copyright (c) 2000-2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- */
-#include "xfs.h"
-#include "xfs_message.h"
-#include "xfs_trace.h"
-
-void *
-kmem_alloc(size_t size, xfs_km_flags_t flags)
-{
-	int	retries = 0;
-	gfp_t	lflags = kmem_flags_convert(flags);
-	void	*ptr;
-
-	trace_kmem_alloc(size, flags, _RET_IP_);
-
-	do {
-		ptr = kmalloc(size, lflags);
-		if (ptr || (flags & KM_MAYFAIL))
-			return ptr;
-		if (!(++retries % 100))
-			xfs_err(NULL,
-	"%s(%u) possible memory allocation deadlock size %u in %s (mode:0x%x)",
-				current->comm, current->pid,
-				(unsigned int)size, __func__, lflags);
-		memalloc_retry_wait(lflags);
-	} while (1);
-}
diff --git a/fs/xfs/kmem.h b/fs/xfs/kmem.h
deleted file mode 100644
index b987dc2c68..0000000000
--- a/fs/xfs/kmem.h
+++ /dev/null
@@ -1,83 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0 */
-/*
- * Copyright (c) 2000-2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- */
-#ifndef __XFS_SUPPORT_KMEM_H__
-#define __XFS_SUPPORT_KMEM_H__
-
-#include <linux/slab.h>
-#include <linux/sched.h>
-#include <linux/mm.h>
-#include <linux/vmalloc.h>
-
-/*
- * General memory allocation interfaces
- */
-
-typedef unsigned __bitwise xfs_km_flags_t;
-#define KM_NOFS		((__force xfs_km_flags_t)0x0004u)
-#define KM_MAYFAIL	((__force xfs_km_flags_t)0x0008u)
-#define KM_ZERO		((__force xfs_km_flags_t)0x0010u)
-#define KM_NOLOCKDEP	((__force xfs_km_flags_t)0x0020u)
-
-/*
- * We use a special process flag to avoid recursive callbacks into
- * the filesystem during transactions.  We will also issue our own
- * warnings, so we explicitly skip any generic ones (silly of us).
- */
-static inline gfp_t
-kmem_flags_convert(xfs_km_flags_t flags)
-{
-	gfp_t	lflags;
-
-	BUG_ON(flags & ~(KM_NOFS | KM_MAYFAIL | KM_ZERO | KM_NOLOCKDEP));
-
-	lflags = GFP_KERNEL | __GFP_NOWARN;
-	if (flags & KM_NOFS)
-		lflags &= ~__GFP_FS;
-
-	/*
-	 * Default page/slab allocator behavior is to retry for ever
-	 * for small allocations. We can override this behavior by using
-	 * __GFP_RETRY_MAYFAIL which will tell the allocator to retry as long
-	 * as it is feasible but rather fail than retry forever for all
-	 * request sizes.
-	 */
-	if (flags & KM_MAYFAIL)
-		lflags |= __GFP_RETRY_MAYFAIL;
-
-	if (flags & KM_ZERO)
-		lflags |= __GFP_ZERO;
-
-	if (flags & KM_NOLOCKDEP)
-		lflags |= __GFP_NOLOCKDEP;
-
-	return lflags;
-}
-
-extern void *kmem_alloc(size_t, xfs_km_flags_t);
-static inline void  kmem_free(const void *ptr)
-{
-	kvfree(ptr);
-}
-
-
-static inline void *
-kmem_zalloc(size_t size, xfs_km_flags_t flags)
-{
-	return kmem_alloc(size, flags | KM_ZERO);
-}
-
-/*
- * Zone interfaces
- */
-static inline struct page *
-kmem_to_page(void *addr)
-{
-	if (is_vmalloc_addr(addr))
-		return vmalloc_to_page(addr);
-	return virt_to_page(addr);
-}
-
-#endif /* __XFS_SUPPORT_KMEM_H__ */
diff --git a/fs/xfs/libxfs/xfs_ag.c b/fs/xfs/libxfs/xfs_ag.c
index 39d9525270..240e079cb3 100644
--- a/fs/xfs/libxfs/xfs_ag.c
+++ b/fs/xfs/libxfs/xfs_ag.c
@@ -194,7 +194,7 @@ xfs_initialize_perag_data(
 		pag = xfs_perag_get(mp, index);
 		error = xfs_alloc_read_agf(pag, NULL, 0, NULL);
 		if (!error)
-			error = xfs_ialloc_read_agi(pag, NULL, NULL);
+			error = xfs_ialloc_read_agi(pag, NULL, 0, NULL);
 		if (error) {
 			xfs_perag_put(pag);
 			return error;
@@ -217,6 +217,7 @@ xfs_initialize_perag_data(
 	 */
 	if (fdblocks > sbp->sb_dblocks || ifree > ialloc) {
 		xfs_alert(mp, "AGF corruption. Please run xfs_repair.");
+		xfs_fs_mark_sick(mp, XFS_SICK_FS_COUNTERS);
 		error = -EFSCORRUPTED;
 		goto out;
 	}
@@ -241,7 +242,7 @@ __xfs_free_perag(
 	struct xfs_perag *pag = container_of(head, struct xfs_perag, rcu_head);
 
 	ASSERT(!delayed_work_pending(&pag->pag_blockgc_work));
-	kmem_free(pag);
+	kfree(pag);
 }
 
 /*
@@ -263,7 +264,7 @@ xfs_free_perag(
 		xfs_defer_drain_free(&pag->pag_intents_drain);
 
 		cancel_delayed_work_sync(&pag->pag_blockgc_work);
-		xfs_buf_hash_destroy(pag);
+		xfs_buf_cache_destroy(&pag->pag_bcache);
 
 		/* drop the mount's active reference */
 		xfs_perag_rele(pag);
@@ -351,9 +352,9 @@ xfs_free_unused_perag_range(
 		spin_unlock(&mp->m_perag_lock);
 		if (!pag)
 			break;
-		xfs_buf_hash_destroy(pag);
+		xfs_buf_cache_destroy(&pag->pag_bcache);
 		xfs_defer_drain_free(&pag->pag_intents_drain);
-		kmem_free(pag);
+		kfree(pag);
 	}
 }
 
@@ -381,7 +382,7 @@ xfs_initialize_perag(
 			continue;
 		}
 
-		pag = kmem_zalloc(sizeof(*pag), KM_MAYFAIL);
+		pag = kzalloc(sizeof(*pag), GFP_KERNEL | __GFP_RETRY_MAYFAIL);
 		if (!pag) {
 			error = -ENOMEM;
 			goto out_unwind_new_pags;
@@ -389,7 +390,7 @@ xfs_initialize_perag(
 		pag->pag_agno = index;
 		pag->pag_mount = mp;
 
-		error = radix_tree_preload(GFP_NOFS);
+		error = radix_tree_preload(GFP_KERNEL | __GFP_RETRY_MAYFAIL);
 		if (error)
 			goto out_free_pag;
 
@@ -416,9 +417,10 @@ xfs_initialize_perag(
 		init_waitqueue_head(&pag->pag_active_wq);
 		pag->pagb_count = 0;
 		pag->pagb_tree = RB_ROOT;
+		xfs_hooks_init(&pag->pag_rmap_update_hooks);
 #endif /* __KERNEL__ */
 
-		error = xfs_buf_hash_init(pag);
+		error = xfs_buf_cache_init(&pag->pag_bcache);
 		if (error)
 			goto out_remove_pag;
 
@@ -453,7 +455,7 @@ out_remove_pag:
 	radix_tree_delete(&mp->m_perag_tree, index);
 	spin_unlock(&mp->m_perag_lock);
 out_free_pag:
-	kmem_free(pag);
+	kfree(pag);
 out_unwind_new_pags:
 	/* unwind any prior newly initialized pags */
 	xfs_free_unused_perag_range(mp, first_initialised, agcount);
@@ -491,7 +493,7 @@ xfs_btroot_init(
 	struct xfs_buf		*bp,
 	struct aghdr_init_data	*id)
 {
-	xfs_btree_init_block(mp, bp, id->type, 0, 0, id->agno);
+	xfs_btree_init_buf(mp, bp, id->bc_ops, 0, 0, id->agno);
 }
 
 /* Finish initializing a free space btree. */
@@ -549,7 +551,7 @@ xfs_freesp_init_recs(
 }
 
 /*
- * Alloc btree root block init functions
+ * bnobt/cntbt btree root block init functions
  */
 static void
 xfs_bnoroot_init(
@@ -557,17 +559,7 @@ xfs_bnoroot_init(
 	struct xfs_buf		*bp,
 	struct aghdr_init_data	*id)
 {
-	xfs_btree_init_block(mp, bp, XFS_BTNUM_BNO, 0, 0, id->agno);
-	xfs_freesp_init_recs(mp, bp, id);
-}
-
-static void
-xfs_cntroot_init(
-	struct xfs_mount	*mp,
-	struct xfs_buf		*bp,
-	struct aghdr_init_data	*id)
-{
-	xfs_btree_init_block(mp, bp, XFS_BTNUM_CNT, 0, 0, id->agno);
+	xfs_btree_init_buf(mp, bp, id->bc_ops, 0, 0, id->agno);
 	xfs_freesp_init_recs(mp, bp, id);
 }
 
@@ -583,7 +575,7 @@ xfs_rmaproot_init(
 	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
 	struct xfs_rmap_rec	*rrec;
 
-	xfs_btree_init_block(mp, bp, XFS_BTNUM_RMAP, 0, 4, id->agno);
+	xfs_btree_init_buf(mp, bp, id->bc_ops, 0, 4, id->agno);
 
 	/*
 	 * mark the AG header regions as static metadata The BNO
@@ -678,14 +670,13 @@ xfs_agfblock_init(
 	agf->agf_versionnum = cpu_to_be32(XFS_AGF_VERSION);
 	agf->agf_seqno = cpu_to_be32(id->agno);
 	agf->agf_length = cpu_to_be32(id->agsize);
-	agf->agf_roots[XFS_BTNUM_BNOi] = cpu_to_be32(XFS_BNO_BLOCK(mp));
-	agf->agf_roots[XFS_BTNUM_CNTi] = cpu_to_be32(XFS_CNT_BLOCK(mp));
-	agf->agf_levels[XFS_BTNUM_BNOi] = cpu_to_be32(1);
-	agf->agf_levels[XFS_BTNUM_CNTi] = cpu_to_be32(1);
+	agf->agf_bno_root = cpu_to_be32(XFS_BNO_BLOCK(mp));
+	agf->agf_cnt_root = cpu_to_be32(XFS_CNT_BLOCK(mp));
+	agf->agf_bno_level = cpu_to_be32(1);
+	agf->agf_cnt_level = cpu_to_be32(1);
 	if (xfs_has_rmapbt(mp)) {
-		agf->agf_roots[XFS_BTNUM_RMAPi] =
-					cpu_to_be32(XFS_RMAP_BLOCK(mp));
-		agf->agf_levels[XFS_BTNUM_RMAPi] = cpu_to_be32(1);
+		agf->agf_rmap_root = cpu_to_be32(XFS_RMAP_BLOCK(mp));
+		agf->agf_rmap_level = cpu_to_be32(1);
 		agf->agf_rmap_blocks = cpu_to_be32(1);
 	}
 
@@ -796,7 +787,7 @@ struct xfs_aghdr_grow_data {
 	size_t			numblks;
 	const struct xfs_buf_ops *ops;
 	aghdr_init_work_f	work;
-	xfs_btnum_t		type;
+	const struct xfs_btree_ops *bc_ops;
 	bool			need_init;
 };
 
@@ -850,13 +841,15 @@ xfs_ag_init_headers(
 		.numblks = BTOBB(mp->m_sb.sb_blocksize),
 		.ops = &xfs_bnobt_buf_ops,
 		.work = &xfs_bnoroot_init,
+		.bc_ops = &xfs_bnobt_ops,
 		.need_init = true
 	},
 	{ /* CNT root block */
 		.daddr = XFS_AGB_TO_DADDR(mp, id->agno, XFS_CNT_BLOCK(mp)),
 		.numblks = BTOBB(mp->m_sb.sb_blocksize),
 		.ops = &xfs_cntbt_buf_ops,
-		.work = &xfs_cntroot_init,
+		.work = &xfs_bnoroot_init,
+		.bc_ops = &xfs_cntbt_ops,
 		.need_init = true
 	},
 	{ /* INO root block */
@@ -864,7 +857,7 @@ xfs_ag_init_headers(
 		.numblks = BTOBB(mp->m_sb.sb_blocksize),
 		.ops = &xfs_inobt_buf_ops,
 		.work = &xfs_btroot_init,
-		.type = XFS_BTNUM_INO,
+		.bc_ops = &xfs_inobt_ops,
 		.need_init = true
 	},
 	{ /* FINO root block */
@@ -872,7 +865,7 @@ xfs_ag_init_headers(
 		.numblks = BTOBB(mp->m_sb.sb_blocksize),
 		.ops = &xfs_finobt_buf_ops,
 		.work = &xfs_btroot_init,
-		.type = XFS_BTNUM_FINO,
+		.bc_ops = &xfs_finobt_ops,
 		.need_init =  xfs_has_finobt(mp)
 	},
 	{ /* RMAP root block */
@@ -880,6 +873,7 @@ xfs_ag_init_headers(
 		.numblks = BTOBB(mp->m_sb.sb_blocksize),
 		.ops = &xfs_rmapbt_buf_ops,
 		.work = &xfs_rmaproot_init,
+		.bc_ops = &xfs_rmapbt_ops,
 		.need_init = xfs_has_rmapbt(mp)
 	},
 	{ /* REFC root block */
@@ -887,7 +881,7 @@ xfs_ag_init_headers(
 		.numblks = BTOBB(mp->m_sb.sb_blocksize),
 		.ops = &xfs_refcountbt_buf_ops,
 		.work = &xfs_btroot_init,
-		.type = XFS_BTNUM_REFC,
+		.bc_ops = &xfs_refcountbt_ops,
 		.need_init = xfs_has_reflink(mp)
 	},
 	{ /* NULL terminating block */
@@ -905,7 +899,7 @@ xfs_ag_init_headers(
 
 		id->daddr = dp->daddr;
 		id->numblks = dp->numblks;
-		id->type = dp->type;
+		id->bc_ops = dp->bc_ops;
 		error = xfs_ag_init_hdr(mp, id, dp->work, dp->ops);
 		if (error)
 			break;
@@ -937,7 +931,7 @@ xfs_ag_shrink_space(
 	int			error, err2;
 
 	ASSERT(pag->pag_agno == mp->m_sb.sb_agcount - 1);
-	error = xfs_ialloc_read_agi(pag, *tpp, &agibp);
+	error = xfs_ialloc_read_agi(pag, *tpp, 0, &agibp);
 	if (error)
 		return error;
 
@@ -950,8 +944,10 @@ xfs_ag_shrink_space(
 	agf = agfbp->b_addr;
 	aglen = be32_to_cpu(agi->agi_length);
 	/* some extra paranoid checks before we shrink the ag */
-	if (XFS_IS_CORRUPT(mp, agf->agf_length != agi->agi_length))
+	if (XFS_IS_CORRUPT(mp, agf->agf_length != agi->agi_length)) {
+		xfs_ag_mark_sick(pag, XFS_SICK_AG_AGF);
 		return -EFSCORRUPTED;
+	}
 	if (delta >= aglen)
 		return -EINVAL;
 
@@ -967,9 +963,7 @@ xfs_ag_shrink_space(
 	 * Disable perag reservations so it doesn't cause the allocation request
 	 * to fail. We'll reestablish reservation before we return.
 	 */
-	error = xfs_ag_resv_free(pag);
-	if (error)
-		return error;
+	xfs_ag_resv_free(pag);
 
 	/* internal log shouldn't also show up in the free space btrees */
 	error = xfs_alloc_vextent_exact_bno(&args,
@@ -979,14 +973,23 @@ xfs_ag_shrink_space(
 
 	if (error) {
 		/*
-		 * if extent allocation fails, need to roll the transaction to
+		 * If extent allocation fails, need to roll the transaction to
 		 * ensure that the AGFL fixup has been committed anyway.
+		 *
+		 * We need to hold the AGF across the roll to ensure nothing can
+		 * access the AG for allocation until the shrink is fully
+		 * cleaned up. And due to the resetting of the AG block
+		 * reservation space needing to lock the AGI, we also have to
+		 * hold that so we don't get AGI/AGF lock order inversions in
+		 * the error handling path.
 		 */
 		xfs_trans_bhold(*tpp, agfbp);
+		xfs_trans_bhold(*tpp, agibp);
 		err2 = xfs_trans_roll(tpp);
 		if (err2)
 			return err2;
 		xfs_trans_bjoin(*tpp, agfbp);
+		xfs_trans_bjoin(*tpp, agibp);
 		goto resv_init_out;
 	}
 
@@ -1057,7 +1060,7 @@ xfs_ag_extend_space(
 
 	ASSERT(pag->pag_agno == pag->pag_mount->m_sb.sb_agcount - 1);
 
-	error = xfs_ialloc_read_agi(pag, tp, &bp);
+	error = xfs_ialloc_read_agi(pag, tp, 0, &bp);
 	if (error)
 		return error;
 
@@ -1114,7 +1117,7 @@ xfs_ag_get_geometry(
 	int			error;
 
 	/* Lock the AG headers. */
-	error = xfs_ialloc_read_agi(pag, NULL, &agi_bp);
+	error = xfs_ialloc_read_agi(pag, NULL, 0, &agi_bp);
 	if (error)
 		return error;
 	error = xfs_alloc_read_agf(pag, NULL, 0, &agf_bp);
diff --git a/fs/xfs/libxfs/xfs_ag.h b/fs/xfs/libxfs/xfs_ag.h
index 4b343c4fac..35de09a251 100644
--- a/fs/xfs/libxfs/xfs_ag.h
+++ b/fs/xfs/libxfs/xfs_ag.h
@@ -36,8 +36,9 @@ struct xfs_perag {
 	atomic_t	pag_active_ref;	/* active reference count */
 	wait_queue_head_t pag_active_wq;/* woken active_ref falls to zero */
 	unsigned long	pag_opstate;
-	uint8_t		pagf_levels[XFS_BTNUM_AGF];
-					/* # of levels in bno & cnt btree */
+	uint8_t		pagf_bno_level;	/* # of levels in bno btree */
+	uint8_t		pagf_cnt_level;	/* # of levels in cnt btree */
+	uint8_t		pagf_rmap_level;/* # of levels in rmap btree */
 	uint32_t	pagf_flcount;	/* count of blocks in freelist */
 	xfs_extlen_t	pagf_freeblks;	/* total free blocks */
 	xfs_extlen_t	pagf_longest;	/* longest free space */
@@ -86,8 +87,10 @@ struct xfs_perag {
 	 * Alternate btree heights so that online repair won't trip the write
 	 * verifiers while rebuilding the AG btrees.
 	 */
-	uint8_t		pagf_repair_levels[XFS_BTNUM_AGF];
+	uint8_t		pagf_repair_bno_level;
+	uint8_t		pagf_repair_cnt_level;
 	uint8_t		pagf_repair_refcount_level;
+	uint8_t		pagf_repair_rmap_level;
 #endif
 
 	spinlock_t	pag_state_lock;
@@ -104,9 +107,7 @@ struct xfs_perag {
 	int		pag_ici_reclaimable;	/* reclaimable inodes */
 	unsigned long	pag_ici_reclaim_cursor;	/* reclaim restart point */
 
-	/* buffer cache index */
-	spinlock_t	pag_buf_lock;	/* lock for pag_buf_hash */
-	struct rhashtable pag_buf_hash;
+	struct xfs_buf_cache	pag_bcache;
 
 	/* background prealloc block trimming */
 	struct delayed_work	pag_blockgc_work;
@@ -119,6 +120,9 @@ struct xfs_perag {
 	 * inconsistencies.
 	 */
 	struct xfs_defer_drain	pag_intents_drain;
+
+	/* Hook to feed rmapbt updates to an active online repair. */
+	struct xfs_hooks	pag_rmap_update_hooks;
 #endif /* __KERNEL__ */
 };
 
@@ -331,7 +335,7 @@ struct aghdr_init_data {
 	/* per header data */
 	xfs_daddr_t		daddr;		/* header location */
 	size_t			numblks;	/* size of header */
-	xfs_btnum_t		type;		/* type of btree root block */
+	const struct xfs_btree_ops *bc_ops;	/* btree ops */
 };
 
 int xfs_ag_init_headers(struct xfs_mount *mp, struct aghdr_init_data *id);
diff --git a/fs/xfs/libxfs/xfs_ag_resv.c b/fs/xfs/libxfs/xfs_ag_resv.c
index da1057bd0e..216423df93 100644
--- a/fs/xfs/libxfs/xfs_ag_resv.c
+++ b/fs/xfs/libxfs/xfs_ag_resv.c
@@ -126,14 +126,13 @@ xfs_ag_resv_needed(
 }
 
 /* Clean out a reservation */
-static int
+static void
 __xfs_ag_resv_free(
 	struct xfs_perag		*pag,
 	enum xfs_ag_resv_type		type)
 {
 	struct xfs_ag_resv		*resv;
 	xfs_extlen_t			oldresv;
-	int				error;
 
 	trace_xfs_ag_resv_free(pag, type, 0);
 
@@ -149,30 +148,19 @@ __xfs_ag_resv_free(
 		oldresv = resv->ar_orig_reserved;
 	else
 		oldresv = resv->ar_reserved;
-	error = xfs_mod_fdblocks(pag->pag_mount, oldresv, true);
+	xfs_add_fdblocks(pag->pag_mount, oldresv);
 	resv->ar_reserved = 0;
 	resv->ar_asked = 0;
 	resv->ar_orig_reserved = 0;
-
-	if (error)
-		trace_xfs_ag_resv_free_error(pag->pag_mount, pag->pag_agno,
-				error, _RET_IP_);
-	return error;
 }
 
 /* Free a per-AG reservation. */
-int
+void
 xfs_ag_resv_free(
 	struct xfs_perag		*pag)
 {
-	int				error;
-	int				err2;
-
-	error = __xfs_ag_resv_free(pag, XFS_AG_RESV_RMAPBT);
-	err2 = __xfs_ag_resv_free(pag, XFS_AG_RESV_METADATA);
-	if (err2 && !error)
-		error = err2;
-	return error;
+	__xfs_ag_resv_free(pag, XFS_AG_RESV_RMAPBT);
+	__xfs_ag_resv_free(pag, XFS_AG_RESV_METADATA);
 }
 
 static int
@@ -216,7 +204,7 @@ __xfs_ag_resv_init(
 	if (XFS_TEST_ERROR(false, mp, XFS_ERRTAG_AG_RESV_FAIL))
 		error = -ENOSPC;
 	else
-		error = xfs_mod_fdblocks(mp, -(int64_t)hidden_space, true);
+		error = xfs_dec_fdblocks(mp, hidden_space, true);
 	if (error) {
 		trace_xfs_ag_resv_init_error(pag->pag_mount, pag->pag_agno,
 				error, _RET_IP_);
diff --git a/fs/xfs/libxfs/xfs_ag_resv.h b/fs/xfs/libxfs/xfs_ag_resv.h
index b74b210008..ff20ed93de 100644
--- a/fs/xfs/libxfs/xfs_ag_resv.h
+++ b/fs/xfs/libxfs/xfs_ag_resv.h
@@ -6,7 +6,7 @@
 #ifndef __XFS_AG_RESV_H__
 #define	__XFS_AG_RESV_H__
 
-int xfs_ag_resv_free(struct xfs_perag *pag);
+void xfs_ag_resv_free(struct xfs_perag *pag);
 int xfs_ag_resv_init(struct xfs_perag *pag, struct xfs_trans *tp);
 
 bool xfs_ag_resv_critical(struct xfs_perag *pag, enum xfs_ag_resv_type type);
diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c
index 3bd0a33fee..6c55a6e88e 100644
--- a/fs/xfs/libxfs/xfs_alloc.c
+++ b/fs/xfs/libxfs/xfs_alloc.c
@@ -26,6 +26,7 @@
 #include "xfs_ag.h"
 #include "xfs_ag_resv.h"
 #include "xfs_bmap.h"
+#include "xfs_health.h"
 
 struct kmem_cache	*xfs_extfree_item_cache;
 
@@ -78,7 +79,7 @@ xfs_prealloc_blocks(
 }
 
 /*
- * The number of blocks per AG that we withhold from xfs_mod_fdblocks to
+ * The number of blocks per AG that we withhold from xfs_dec_fdblocks to
  * guarantee that we can refill the AGFL prior to allocating space in a nearly
  * full AG.  Although the space described by the free space btrees, the
  * blocks used by the freesp btrees themselves, and the blocks owned by the
@@ -88,7 +89,7 @@ xfs_prealloc_blocks(
  * until the fs goes down, we subtract this many AG blocks from the incore
  * fdblocks to ensure user allocation does not overcommit the space the
  * filesystem needs for the AGFLs.  The rmap btree uses a per-AG reservation to
- * withhold space from xfs_mod_fdblocks, so we do not account for that here.
+ * withhold space from xfs_dec_fdblocks, so we do not account for that here.
  */
 #define XFS_ALLOCBT_AGFL_RESERVE	4
 
@@ -150,23 +151,38 @@ xfs_alloc_ag_max_usable(
 	return mp->m_sb.sb_agblocks - blocks;
 }
 
+
+static int
+xfs_alloc_lookup(
+	struct xfs_btree_cur	*cur,
+	xfs_lookup_t		dir,
+	xfs_agblock_t		bno,
+	xfs_extlen_t		len,
+	int			*stat)
+{
+	int			error;
+
+	cur->bc_rec.a.ar_startblock = bno;
+	cur->bc_rec.a.ar_blockcount = len;
+	error = xfs_btree_lookup(cur, dir, stat);
+	if (*stat == 1)
+		cur->bc_flags |= XFS_BTREE_ALLOCBT_ACTIVE;
+	else
+		cur->bc_flags &= ~XFS_BTREE_ALLOCBT_ACTIVE;
+	return error;
+}
+
 /*
  * Lookup the record equal to [bno, len] in the btree given by cur.
  */
-STATIC int				/* error */
+static inline int				/* error */
 xfs_alloc_lookup_eq(
 	struct xfs_btree_cur	*cur,	/* btree cursor */
 	xfs_agblock_t		bno,	/* starting block of extent */
 	xfs_extlen_t		len,	/* length of extent */
 	int			*stat)	/* success/failure */
 {
-	int			error;
-
-	cur->bc_rec.a.ar_startblock = bno;
-	cur->bc_rec.a.ar_blockcount = len;
-	error = xfs_btree_lookup(cur, XFS_LOOKUP_EQ, stat);
-	cur->bc_ag.abt.active = (*stat == 1);
-	return error;
+	return xfs_alloc_lookup(cur, XFS_LOOKUP_EQ, bno, len, stat);
 }
 
 /*
@@ -180,13 +196,7 @@ xfs_alloc_lookup_ge(
 	xfs_extlen_t		len,	/* length of extent */
 	int			*stat)	/* success/failure */
 {
-	int			error;
-
-	cur->bc_rec.a.ar_startblock = bno;
-	cur->bc_rec.a.ar_blockcount = len;
-	error = xfs_btree_lookup(cur, XFS_LOOKUP_GE, stat);
-	cur->bc_ag.abt.active = (*stat == 1);
-	return error;
+	return xfs_alloc_lookup(cur, XFS_LOOKUP_GE, bno, len, stat);
 }
 
 /*
@@ -200,19 +210,14 @@ xfs_alloc_lookup_le(
 	xfs_extlen_t		len,	/* length of extent */
 	int			*stat)	/* success/failure */
 {
-	int			error;
-	cur->bc_rec.a.ar_startblock = bno;
-	cur->bc_rec.a.ar_blockcount = len;
-	error = xfs_btree_lookup(cur, XFS_LOOKUP_LE, stat);
-	cur->bc_ag.abt.active = (*stat == 1);
-	return error;
+	return xfs_alloc_lookup(cur, XFS_LOOKUP_LE, bno, len, stat);
 }
 
 static inline bool
 xfs_alloc_cur_active(
 	struct xfs_btree_cur	*cur)
 {
-	return cur && cur->bc_ag.abt.active;
+	return cur && (cur->bc_flags & XFS_BTREE_ALLOCBT_ACTIVE);
 }
 
 /*
@@ -268,12 +273,12 @@ xfs_alloc_complain_bad_rec(
 	struct xfs_mount		*mp = cur->bc_mp;
 
 	xfs_warn(mp,
-		"%s Freespace BTree record corruption in AG %d detected at %pS!",
-		cur->bc_btnum == XFS_BTNUM_BNO ? "Block" : "Size",
-		cur->bc_ag.pag->pag_agno, fa);
+		"%sbt record corruption in AG %d detected at %pS!",
+		cur->bc_ops->name, cur->bc_ag.pag->pag_agno, fa);
 	xfs_warn(mp,
 		"start block 0x%x block count 0x%x", irec->ar_startblock,
 		irec->ar_blockcount);
+	xfs_btree_mark_sick(cur);
 	return -EFSCORRUPTED;
 }
 
@@ -497,14 +502,18 @@ xfs_alloc_fixup_trees(
 		if (XFS_IS_CORRUPT(mp,
 				   i != 1 ||
 				   nfbno1 != fbno ||
-				   nflen1 != flen))
+				   nflen1 != flen)) {
+			xfs_btree_mark_sick(cnt_cur);
 			return -EFSCORRUPTED;
+		}
 #endif
 	} else {
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, fbno, flen, &i)))
 			return error;
-		if (XFS_IS_CORRUPT(mp, i != 1))
+		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cnt_cur);
 			return -EFSCORRUPTED;
+		}
 	}
 	/*
 	 * Look up the record in the by-block tree if necessary.
@@ -516,14 +525,18 @@ xfs_alloc_fixup_trees(
 		if (XFS_IS_CORRUPT(mp,
 				   i != 1 ||
 				   nfbno1 != fbno ||
-				   nflen1 != flen))
+				   nflen1 != flen)) {
+			xfs_btree_mark_sick(bno_cur);
 			return -EFSCORRUPTED;
+		}
 #endif
 	} else {
 		if ((error = xfs_alloc_lookup_eq(bno_cur, fbno, flen, &i)))
 			return error;
-		if (XFS_IS_CORRUPT(mp, i != 1))
+		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(bno_cur);
 			return -EFSCORRUPTED;
+		}
 	}
 
 #ifdef DEBUG
@@ -536,8 +549,10 @@ xfs_alloc_fixup_trees(
 
 		if (XFS_IS_CORRUPT(mp,
 				   bnoblock->bb_numrecs !=
-				   cntblock->bb_numrecs))
+				   cntblock->bb_numrecs)) {
+			xfs_btree_mark_sick(bno_cur);
 			return -EFSCORRUPTED;
+		}
 	}
 #endif
 
@@ -567,30 +582,40 @@ xfs_alloc_fixup_trees(
 	 */
 	if ((error = xfs_btree_delete(cnt_cur, &i)))
 		return error;
-	if (XFS_IS_CORRUPT(mp, i != 1))
+	if (XFS_IS_CORRUPT(mp, i != 1)) {
+		xfs_btree_mark_sick(cnt_cur);
 		return -EFSCORRUPTED;
+	}
 	/*
 	 * Add new by-size btree entry(s).
 	 */
 	if (nfbno1 != NULLAGBLOCK) {
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, nfbno1, nflen1, &i)))
 			return error;
-		if (XFS_IS_CORRUPT(mp, i != 0))
+		if (XFS_IS_CORRUPT(mp, i != 0)) {
+			xfs_btree_mark_sick(cnt_cur);
 			return -EFSCORRUPTED;
+		}
 		if ((error = xfs_btree_insert(cnt_cur, &i)))
 			return error;
-		if (XFS_IS_CORRUPT(mp, i != 1))
+		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cnt_cur);
 			return -EFSCORRUPTED;
+		}
 	}
 	if (nfbno2 != NULLAGBLOCK) {
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, nfbno2, nflen2, &i)))
 			return error;
-		if (XFS_IS_CORRUPT(mp, i != 0))
+		if (XFS_IS_CORRUPT(mp, i != 0)) {
+			xfs_btree_mark_sick(cnt_cur);
 			return -EFSCORRUPTED;
+		}
 		if ((error = xfs_btree_insert(cnt_cur, &i)))
 			return error;
-		if (XFS_IS_CORRUPT(mp, i != 1))
+		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cnt_cur);
 			return -EFSCORRUPTED;
+		}
 	}
 	/*
 	 * Fix up the by-block btree entry(s).
@@ -601,8 +626,10 @@ xfs_alloc_fixup_trees(
 		 */
 		if ((error = xfs_btree_delete(bno_cur, &i)))
 			return error;
-		if (XFS_IS_CORRUPT(mp, i != 1))
+		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(bno_cur);
 			return -EFSCORRUPTED;
+		}
 	} else {
 		/*
 		 * Update the by-block entry to start later|be shorter.
@@ -616,12 +643,16 @@ xfs_alloc_fixup_trees(
 		 */
 		if ((error = xfs_alloc_lookup_eq(bno_cur, nfbno2, nflen2, &i)))
 			return error;
-		if (XFS_IS_CORRUPT(mp, i != 0))
+		if (XFS_IS_CORRUPT(mp, i != 0)) {
+			xfs_btree_mark_sick(bno_cur);
 			return -EFSCORRUPTED;
+		}
 		if ((error = xfs_btree_insert(bno_cur, &i)))
 			return error;
-		if (XFS_IS_CORRUPT(mp, i != 1))
+		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(bno_cur);
 			return -EFSCORRUPTED;
+		}
 	}
 	return 0;
 }
@@ -755,6 +786,8 @@ xfs_alloc_read_agfl(
 			mp, tp, mp->m_ddev_targp,
 			XFS_AG_DADDR(mp, pag->pag_agno, XFS_AGFL_DADDR(mp)),
 			XFS_FSS_TO_BB(mp, 1), 0, &bp, &xfs_agfl_buf_ops);
+	if (xfs_metadata_is_sick(error))
+		xfs_ag_mark_sick(pag, XFS_SICK_AG_AGFL);
 	if (error)
 		return error;
 	xfs_buf_set_ref(bp, XFS_AGFL_REF);
@@ -776,6 +809,7 @@ xfs_alloc_update_counters(
 	if (unlikely(be32_to_cpu(agf->agf_freeblks) >
 		     be32_to_cpu(agf->agf_length))) {
 		xfs_buf_mark_corrupt(agbp);
+		xfs_ag_mark_sick(agbp->b_pag, XFS_SICK_AG_AGF);
 		return -EFSCORRUPTED;
 	}
 
@@ -828,8 +862,8 @@ xfs_alloc_cur_setup(
 	 * attempt a small allocation.
 	 */
 	if (!acur->cnt)
-		acur->cnt = xfs_allocbt_init_cursor(args->mp, args->tp,
-					args->agbp, args->pag, XFS_BTNUM_CNT);
+		acur->cnt = xfs_cntbt_init_cursor(args->mp, args->tp,
+					args->agbp, args->pag);
 	error = xfs_alloc_lookup_ge(acur->cnt, 0, args->maxlen, &i);
 	if (error)
 		return error;
@@ -838,11 +872,11 @@ xfs_alloc_cur_setup(
 	 * Allocate the bnobt left and right search cursors.
 	 */
 	if (!acur->bnolt)
-		acur->bnolt = xfs_allocbt_init_cursor(args->mp, args->tp,
-					args->agbp, args->pag, XFS_BTNUM_BNO);
+		acur->bnolt = xfs_bnobt_init_cursor(args->mp, args->tp,
+					args->agbp, args->pag);
 	if (!acur->bnogt)
-		acur->bnogt = xfs_allocbt_init_cursor(args->mp, args->tp,
-					args->agbp, args->pag, XFS_BTNUM_BNO);
+		acur->bnogt = xfs_bnobt_init_cursor(args->mp, args->tp,
+					args->agbp, args->pag);
 	return i == 1 ? 0 : -ENOSPC;
 }
 
@@ -884,15 +918,17 @@ xfs_alloc_cur_check(
 	bool			busy;
 	unsigned		busy_gen = 0;
 	bool			deactivate = false;
-	bool			isbnobt = cur->bc_btnum == XFS_BTNUM_BNO;
+	bool			isbnobt = xfs_btree_is_bno(cur->bc_ops);
 
 	*new = 0;
 
 	error = xfs_alloc_get_rec(cur, &bno, &len, &i);
 	if (error)
 		return error;
-	if (XFS_IS_CORRUPT(args->mp, i != 1))
+	if (XFS_IS_CORRUPT(args->mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		return -EFSCORRUPTED;
+	}
 
 	/*
 	 * Check minlen and deactivate a cntbt cursor if out of acceptable size
@@ -958,9 +994,8 @@ xfs_alloc_cur_check(
 		deactivate = true;
 out:
 	if (deactivate)
-		cur->bc_ag.abt.active = false;
-	trace_xfs_alloc_cur_check(args->mp, cur->bc_btnum, bno, len, diff,
-				  *new);
+		cur->bc_flags &= ~XFS_BTREE_ALLOCBT_ACTIVE;
+	trace_xfs_alloc_cur_check(cur, bno, len, diff, *new);
 	return 0;
 }
 
@@ -973,13 +1008,12 @@ xfs_alloc_cur_finish(
 	struct xfs_alloc_arg	*args,
 	struct xfs_alloc_cur	*acur)
 {
-	struct xfs_agf __maybe_unused *agf = args->agbp->b_addr;
 	int			error;
 
 	ASSERT(acur->cnt && acur->bnolt);
 	ASSERT(acur->bno >= acur->rec_bno);
 	ASSERT(acur->bno + acur->len <= acur->rec_bno + acur->rec_len);
-	ASSERT(acur->rec_bno + acur->rec_len <= be32_to_cpu(agf->agf_length));
+	ASSERT(xfs_verify_agbext(args->pag, acur->rec_bno, acur->rec_len));
 
 	error = xfs_alloc_fixup_trees(acur->cnt, acur->bnolt, acur->rec_bno,
 				      acur->rec_len, acur->bno, acur->len, 0);
@@ -1098,6 +1132,7 @@ xfs_alloc_ag_vextent_small(
 		if (error)
 			goto error;
 		if (XFS_IS_CORRUPT(args->mp, i != 1)) {
+			xfs_btree_mark_sick(ccur);
 			error = -EFSCORRUPTED;
 			goto error;
 		}
@@ -1132,6 +1167,7 @@ xfs_alloc_ag_vextent_small(
 	*fbnop = args->agbno = fbno;
 	*flenp = args->len = 1;
 	if (XFS_IS_CORRUPT(args->mp, fbno >= be32_to_cpu(agf->agf_length))) {
+		xfs_btree_mark_sick(ccur);
 		error = -EFSCORRUPTED;
 		goto error;
 	}
@@ -1180,7 +1216,6 @@ STATIC int			/* error */
 xfs_alloc_ag_vextent_exact(
 	xfs_alloc_arg_t	*args)	/* allocation argument structure */
 {
-	struct xfs_agf __maybe_unused *agf = args->agbp->b_addr;
 	struct xfs_btree_cur *bno_cur;/* by block-number btree cursor */
 	struct xfs_btree_cur *cnt_cur;/* by count btree cursor */
 	int		error;
@@ -1197,8 +1232,8 @@ xfs_alloc_ag_vextent_exact(
 	/*
 	 * Allocate/initialize a cursor for the by-number freespace btree.
 	 */
-	bno_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp,
-					  args->pag, XFS_BTNUM_BNO);
+	bno_cur = xfs_bnobt_init_cursor(args->mp, args->tp, args->agbp,
+					  args->pag);
 
 	/*
 	 * Lookup bno and minlen in the btree (minlen is irrelevant, really).
@@ -1218,6 +1253,7 @@ xfs_alloc_ag_vextent_exact(
 	if (error)
 		goto error0;
 	if (XFS_IS_CORRUPT(args->mp, i != 1)) {
+		xfs_btree_mark_sick(bno_cur);
 		error = -EFSCORRUPTED;
 		goto error0;
 	}
@@ -1257,9 +1293,9 @@ xfs_alloc_ag_vextent_exact(
 	 * We are allocating agbno for args->len
 	 * Allocate/initialize a cursor for the by-size btree.
 	 */
-	cnt_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp,
-					args->pag, XFS_BTNUM_CNT);
-	ASSERT(args->agbno + args->len <= be32_to_cpu(agf->agf_length));
+	cnt_cur = xfs_cntbt_init_cursor(args->mp, args->tp, args->agbp,
+					args->pag);
+	ASSERT(xfs_verify_agbext(args->pag, args->agbno, args->len));
 	error = xfs_alloc_fixup_trees(cnt_cur, bno_cur, fbno, flen, args->agbno,
 				      args->len, XFSA_FIXUP_BNO_OK);
 	if (error) {
@@ -1330,7 +1366,7 @@ xfs_alloc_walk_iter(
 		if (error)
 			return error;
 		if (i == 0)
-			cur->bc_ag.abt.active = false;
+			cur->bc_flags &= ~XFS_BTREE_ALLOCBT_ACTIVE;
 
 		if (count > 0)
 			count--;
@@ -1444,7 +1480,7 @@ xfs_alloc_ag_vextent_locality(
 		if (error)
 			return error;
 		if (i) {
-			acur->cnt->bc_ag.abt.active = true;
+			acur->cnt->bc_flags |= XFS_BTREE_ALLOCBT_ACTIVE;
 			fbcur = acur->cnt;
 			fbinc = false;
 		}
@@ -1497,8 +1533,10 @@ xfs_alloc_ag_vextent_lastblock(
 			error = xfs_alloc_get_rec(acur->cnt, bno, len, &i);
 			if (error)
 				return error;
-			if (XFS_IS_CORRUPT(args->mp, i != 1))
+			if (XFS_IS_CORRUPT(args->mp, i != 1)) {
+				xfs_btree_mark_sick(acur->cnt);
 				return -EFSCORRUPTED;
+			}
 			if (*len >= args->minlen)
 				break;
 			error = xfs_btree_increment(acur->cnt, 0, &i);
@@ -1670,8 +1708,8 @@ restart:
 	/*
 	 * Allocate and initialize a cursor for the by-size btree.
 	 */
-	cnt_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp,
-					args->pag, XFS_BTNUM_CNT);
+	cnt_cur = xfs_cntbt_init_cursor(args->mp, args->tp, args->agbp,
+					args->pag);
 	bno_cur = NULL;
 
 	/*
@@ -1710,6 +1748,7 @@ restart:
 			if (error)
 				goto error0;
 			if (XFS_IS_CORRUPT(args->mp, i != 1)) {
+				xfs_btree_mark_sick(cnt_cur);
 				error = -EFSCORRUPTED;
 				goto error0;
 			}
@@ -1756,6 +1795,7 @@ restart:
 			   rlen != 0 &&
 			   (rlen > flen ||
 			    rbno + rlen > fbno + flen))) {
+		xfs_btree_mark_sick(cnt_cur);
 		error = -EFSCORRUPTED;
 		goto error0;
 	}
@@ -1778,6 +1818,7 @@ restart:
 					&i)))
 				goto error0;
 			if (XFS_IS_CORRUPT(args->mp, i != 1)) {
+				xfs_btree_mark_sick(cnt_cur);
 				error = -EFSCORRUPTED;
 				goto error0;
 			}
@@ -1790,6 +1831,7 @@ restart:
 					   rlen != 0 &&
 					   (rlen > flen ||
 					    rbno + rlen > fbno + flen))) {
+				xfs_btree_mark_sick(cnt_cur);
 				error = -EFSCORRUPTED;
 				goto error0;
 			}
@@ -1806,6 +1848,7 @@ restart:
 				&i)))
 			goto error0;
 		if (XFS_IS_CORRUPT(args->mp, i != 1)) {
+			xfs_btree_mark_sick(cnt_cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -1844,14 +1887,15 @@ restart:
 
 	rlen = args->len;
 	if (XFS_IS_CORRUPT(args->mp, rlen > flen)) {
+		xfs_btree_mark_sick(cnt_cur);
 		error = -EFSCORRUPTED;
 		goto error0;
 	}
 	/*
 	 * Allocate and initialize a cursor for the by-block tree.
 	 */
-	bno_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp,
-					args->pag, XFS_BTNUM_BNO);
+	bno_cur = xfs_bnobt_init_cursor(args->mp, args->tp, args->agbp,
+					args->pag);
 	if ((error = xfs_alloc_fixup_trees(cnt_cur, bno_cur, fbno, flen,
 			rbno, rlen, XFSA_FIXUP_CNT_OK)))
 		goto error0;
@@ -1863,6 +1907,7 @@ restart:
 	if (XFS_IS_CORRUPT(args->mp,
 			   args->agbno + args->len >
 			   be32_to_cpu(agf->agf_length))) {
+		xfs_ag_mark_sick(args->pag, XFS_SICK_AG_BNOBT);
 		error = -EFSCORRUPTED;
 		goto error0;
 	}
@@ -1924,7 +1969,7 @@ xfs_free_ag_extent(
 	/*
 	 * Allocate and initialize a cursor for the by-block btree.
 	 */
-	bno_cur = xfs_allocbt_init_cursor(mp, tp, agbp, pag, XFS_BTNUM_BNO);
+	bno_cur = xfs_bnobt_init_cursor(mp, tp, agbp, pag);
 	/*
 	 * Look for a neighboring block on the left (lower block numbers)
 	 * that is contiguous with this space.
@@ -1938,6 +1983,7 @@ xfs_free_ag_extent(
 		if ((error = xfs_alloc_get_rec(bno_cur, &ltbno, &ltlen, &i)))
 			goto error0;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(bno_cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -1953,6 +1999,7 @@ xfs_free_ag_extent(
 			 * Very bad.
 			 */
 			if (XFS_IS_CORRUPT(mp, ltbno + ltlen > bno)) {
+				xfs_btree_mark_sick(bno_cur);
 				error = -EFSCORRUPTED;
 				goto error0;
 			}
@@ -1971,6 +2018,7 @@ xfs_free_ag_extent(
 		if ((error = xfs_alloc_get_rec(bno_cur, &gtbno, &gtlen, &i)))
 			goto error0;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(bno_cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -1986,6 +2034,7 @@ xfs_free_ag_extent(
 			 * Very bad.
 			 */
 			if (XFS_IS_CORRUPT(mp, bno + len > gtbno)) {
+				xfs_btree_mark_sick(bno_cur);
 				error = -EFSCORRUPTED;
 				goto error0;
 			}
@@ -1994,7 +2043,7 @@ xfs_free_ag_extent(
 	/*
 	 * Now allocate and initialize a cursor for the by-size tree.
 	 */
-	cnt_cur = xfs_allocbt_init_cursor(mp, tp, agbp, pag, XFS_BTNUM_CNT);
+	cnt_cur = xfs_cntbt_init_cursor(mp, tp, agbp, pag);
 	/*
 	 * Have both left and right contiguous neighbors.
 	 * Merge all three into a single free block.
@@ -2006,12 +2055,14 @@ xfs_free_ag_extent(
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, ltbno, ltlen, &i)))
 			goto error0;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cnt_cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
 		if ((error = xfs_btree_delete(cnt_cur, &i)))
 			goto error0;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cnt_cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -2021,12 +2072,14 @@ xfs_free_ag_extent(
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, gtbno, gtlen, &i)))
 			goto error0;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cnt_cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
 		if ((error = xfs_btree_delete(cnt_cur, &i)))
 			goto error0;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cnt_cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -2036,6 +2089,7 @@ xfs_free_ag_extent(
 		if ((error = xfs_btree_delete(bno_cur, &i)))
 			goto error0;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(bno_cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -2045,6 +2099,7 @@ xfs_free_ag_extent(
 		if ((error = xfs_btree_decrement(bno_cur, 0, &i)))
 			goto error0;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(bno_cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -2064,6 +2119,7 @@ xfs_free_ag_extent(
 					   i != 1 ||
 					   xxbno != ltbno ||
 					   xxlen != ltlen)) {
+				xfs_btree_mark_sick(bno_cur);
 				error = -EFSCORRUPTED;
 				goto error0;
 			}
@@ -2088,12 +2144,14 @@ xfs_free_ag_extent(
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, ltbno, ltlen, &i)))
 			goto error0;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cnt_cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
 		if ((error = xfs_btree_delete(cnt_cur, &i)))
 			goto error0;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cnt_cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -2104,6 +2162,7 @@ xfs_free_ag_extent(
 		if ((error = xfs_btree_decrement(bno_cur, 0, &i)))
 			goto error0;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(bno_cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -2123,12 +2182,14 @@ xfs_free_ag_extent(
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, gtbno, gtlen, &i)))
 			goto error0;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cnt_cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
 		if ((error = xfs_btree_delete(cnt_cur, &i)))
 			goto error0;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cnt_cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -2151,6 +2212,7 @@ xfs_free_ag_extent(
 		if ((error = xfs_btree_insert(bno_cur, &i)))
 			goto error0;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(bno_cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -2163,12 +2225,14 @@ xfs_free_ag_extent(
 	if ((error = xfs_alloc_lookup_eq(cnt_cur, nbno, nlen, &i)))
 		goto error0;
 	if (XFS_IS_CORRUPT(mp, i != 0)) {
+		xfs_btree_mark_sick(cnt_cur);
 		error = -EFSCORRUPTED;
 		goto error0;
 	}
 	if ((error = xfs_btree_insert(cnt_cur, &i)))
 		goto error0;
 	if (XFS_IS_CORRUPT(mp, i != 1)) {
+		xfs_btree_mark_sick(cnt_cur);
 		error = -EFSCORRUPTED;
 		goto error0;
 	}
@@ -2267,8 +2331,9 @@ xfs_alloc_min_freelist(
 	struct xfs_perag	*pag)
 {
 	/* AG btrees have at least 1 level. */
-	static const uint8_t	fake_levels[XFS_BTNUM_AGF] = {1, 1, 1};
-	const uint8_t		*levels = pag ? pag->pagf_levels : fake_levels;
+	const unsigned int	bno_level = pag ? pag->pagf_bno_level : 1;
+	const unsigned int	cnt_level = pag ? pag->pagf_cnt_level : 1;
+	const unsigned int	rmap_level = pag ? pag->pagf_rmap_level : 1;
 	unsigned int		min_free;
 
 	ASSERT(mp->m_alloc_maxlevels > 0);
@@ -2295,16 +2360,12 @@ xfs_alloc_min_freelist(
 	 */
 
 	/* space needed by-bno freespace btree */
-	min_free = min_t(unsigned int, levels[XFS_BTNUM_BNOi] + 1,
-				       mp->m_alloc_maxlevels) * 2 - 2;
+	min_free = min(bno_level + 1, mp->m_alloc_maxlevels) * 2 - 2;
 	/* space needed by-size freespace btree */
-	min_free += min_t(unsigned int, levels[XFS_BTNUM_CNTi] + 1,
-				       mp->m_alloc_maxlevels) * 2 - 2;
+	min_free += min(cnt_level + 1, mp->m_alloc_maxlevels) * 2 - 2;
 	/* space needed reverse mapping used space btree */
 	if (xfs_has_rmapbt(mp))
-		min_free += min_t(unsigned int, levels[XFS_BTNUM_RMAPi] + 1,
-						mp->m_rmap_maxlevels) * 2 - 2;
-
+		min_free += min(rmap_level + 1, mp->m_rmap_maxlevels) * 2 - 2;
 	return min_free;
 }
 
@@ -2691,13 +2752,14 @@ xfs_exact_minlen_extent_available(
 	xfs_extlen_t		flen;
 	int			error = 0;
 
-	cnt_cur = xfs_allocbt_init_cursor(args->mp, args->tp, agbp,
-					args->pag, XFS_BTNUM_CNT);
+	cnt_cur = xfs_cntbt_init_cursor(args->mp, args->tp, agbp,
+					args->pag);
 	error = xfs_alloc_lookup_ge(cnt_cur, 0, args->minlen, stat);
 	if (error)
 		goto out;
 
 	if (*stat == 0) {
+		xfs_btree_mark_sick(cnt_cur);
 		error = -EFSCORRUPTED;
 		goto out;
 	}
@@ -2987,8 +3049,8 @@ xfs_alloc_log_agf(
 		offsetof(xfs_agf_t, agf_versionnum),
 		offsetof(xfs_agf_t, agf_seqno),
 		offsetof(xfs_agf_t, agf_length),
-		offsetof(xfs_agf_t, agf_roots[0]),
-		offsetof(xfs_agf_t, agf_levels[0]),
+		offsetof(xfs_agf_t, agf_bno_root),   /* also cnt/rmap root */
+		offsetof(xfs_agf_t, agf_bno_level),  /* also cnt/rmap levels */
 		offsetof(xfs_agf_t, agf_flfirst),
 		offsetof(xfs_agf_t, agf_fllast),
 		offsetof(xfs_agf_t, agf_flcount),
@@ -3167,12 +3229,10 @@ xfs_agf_verify(
 	    be32_to_cpu(agf->agf_freeblks) > agf_length)
 		return __this_address;
 
-	if (be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]) < 1 ||
-	    be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]) < 1 ||
-	    be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]) >
-						mp->m_alloc_maxlevels ||
-	    be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]) >
-						mp->m_alloc_maxlevels)
+	if (be32_to_cpu(agf->agf_bno_level) < 1 ||
+	    be32_to_cpu(agf->agf_cnt_level) < 1 ||
+	    be32_to_cpu(agf->agf_bno_level) > mp->m_alloc_maxlevels ||
+	    be32_to_cpu(agf->agf_cnt_level) > mp->m_alloc_maxlevels)
 		return __this_address;
 
 	if (xfs_has_lazysbcount(mp) &&
@@ -3183,9 +3243,8 @@ xfs_agf_verify(
 		if (be32_to_cpu(agf->agf_rmap_blocks) > agf_length)
 			return __this_address;
 
-		if (be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]) < 1 ||
-		    be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]) >
-							mp->m_rmap_maxlevels)
+		if (be32_to_cpu(agf->agf_rmap_level) < 1 ||
+		    be32_to_cpu(agf->agf_rmap_level) > mp->m_rmap_maxlevels)
 			return __this_address;
 	}
 
@@ -3268,6 +3327,8 @@ xfs_read_agf(
 	error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp,
 			XFS_AG_DADDR(mp, pag->pag_agno, XFS_AGF_DADDR(mp)),
 			XFS_FSS_TO_BB(mp, 1), flags, agfbpp, &xfs_agf_buf_ops);
+	if (xfs_metadata_is_sick(error))
+		xfs_ag_mark_sick(pag, XFS_SICK_AG_AGF);
 	if (error)
 		return error;
 
@@ -3309,12 +3370,9 @@ xfs_alloc_read_agf(
 		pag->pagf_btreeblks = be32_to_cpu(agf->agf_btreeblks);
 		pag->pagf_flcount = be32_to_cpu(agf->agf_flcount);
 		pag->pagf_longest = be32_to_cpu(agf->agf_longest);
-		pag->pagf_levels[XFS_BTNUM_BNOi] =
-			be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNOi]);
-		pag->pagf_levels[XFS_BTNUM_CNTi] =
-			be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNTi]);
-		pag->pagf_levels[XFS_BTNUM_RMAPi] =
-			be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAPi]);
+		pag->pagf_bno_level = be32_to_cpu(agf->agf_bno_level);
+		pag->pagf_cnt_level = be32_to_cpu(agf->agf_cnt_level);
+		pag->pagf_rmap_level = be32_to_cpu(agf->agf_rmap_level);
 		pag->pagf_refcount_level = be32_to_cpu(agf->agf_refcount_level);
 		if (xfs_agfl_needs_reset(pag->pag_mount, agf))
 			set_bit(XFS_AGSTATE_AGFL_NEEDS_RESET, &pag->pag_opstate);
@@ -3343,10 +3401,8 @@ xfs_alloc_read_agf(
 		ASSERT(pag->pagf_btreeblks == be32_to_cpu(agf->agf_btreeblks));
 		ASSERT(pag->pagf_flcount == be32_to_cpu(agf->agf_flcount));
 		ASSERT(pag->pagf_longest == be32_to_cpu(agf->agf_longest));
-		ASSERT(pag->pagf_levels[XFS_BTNUM_BNOi] ==
-		       be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNOi]));
-		ASSERT(pag->pagf_levels[XFS_BTNUM_CNTi] ==
-		       be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNTi]));
+		ASSERT(pag->pagf_bno_level == be32_to_cpu(agf->agf_bno_level));
+		ASSERT(pag->pagf_cnt_level == be32_to_cpu(agf->agf_cnt_level));
 	}
 #endif
 	if (agfbpp)
@@ -3895,17 +3951,23 @@ __xfs_free_extent(
 		return -EIO;
 
 	error = xfs_free_extent_fix_freelist(tp, pag, &agbp);
-	if (error)
+	if (error) {
+		if (xfs_metadata_is_sick(error))
+			xfs_ag_mark_sick(pag, XFS_SICK_AG_BNOBT);
 		return error;
+	}
+
 	agf = agbp->b_addr;
 
 	if (XFS_IS_CORRUPT(mp, agbno >= mp->m_sb.sb_agblocks)) {
+		xfs_ag_mark_sick(pag, XFS_SICK_AG_BNOBT);
 		error = -EFSCORRUPTED;
 		goto err_release;
 	}
 
 	/* validate the extent size is legal now we have the agf locked */
 	if (XFS_IS_CORRUPT(mp, agbno + len > be32_to_cpu(agf->agf_length))) {
+		xfs_ag_mark_sick(pag, XFS_SICK_AG_BNOBT);
 		error = -EFSCORRUPTED;
 		goto err_release;
 	}
@@ -3962,7 +4024,7 @@ xfs_alloc_query_range(
 	union xfs_btree_irec			high_brec = { .a = *high_rec };
 	struct xfs_alloc_query_range_info	query = { .priv = priv, .fn = fn };
 
-	ASSERT(cur->bc_btnum == XFS_BTNUM_BNO);
+	ASSERT(xfs_btree_is_bno(cur->bc_ops));
 	return xfs_btree_query_range(cur, &low_brec, &high_brec,
 			xfs_alloc_query_range_helper, &query);
 }
@@ -3976,7 +4038,7 @@ xfs_alloc_query_all(
 {
 	struct xfs_alloc_query_range_info	query;
 
-	ASSERT(cur->bc_btnum == XFS_BTNUM_BNO);
+	ASSERT(xfs_btree_is_bno(cur->bc_ops));
 	query.priv = priv;
 	query.fn = fn;
 	return xfs_btree_query_all(cur, xfs_alloc_query_range_helper, &query);
diff --git a/fs/xfs/libxfs/xfs_alloc_btree.c b/fs/xfs/libxfs/xfs_alloc_btree.c
index a7032bf0cd..6ef5ddd896 100644
--- a/fs/xfs/libxfs/xfs_alloc_btree.c
+++ b/fs/xfs/libxfs/xfs_alloc_btree.c
@@ -16,6 +16,7 @@
 #include "xfs_alloc.h"
 #include "xfs_extent_busy.h"
 #include "xfs_error.h"
+#include "xfs_health.h"
 #include "xfs_trace.h"
 #include "xfs_trans.h"
 #include "xfs_ag.h"
@@ -23,13 +24,22 @@
 static struct kmem_cache	*xfs_allocbt_cur_cache;
 
 STATIC struct xfs_btree_cur *
-xfs_allocbt_dup_cursor(
+xfs_bnobt_dup_cursor(
 	struct xfs_btree_cur	*cur)
 {
-	return xfs_allocbt_init_cursor(cur->bc_mp, cur->bc_tp,
-			cur->bc_ag.agbp, cur->bc_ag.pag, cur->bc_btnum);
+	return xfs_bnobt_init_cursor(cur->bc_mp, cur->bc_tp, cur->bc_ag.agbp,
+			cur->bc_ag.pag);
 }
 
+STATIC struct xfs_btree_cur *
+xfs_cntbt_dup_cursor(
+	struct xfs_btree_cur	*cur)
+{
+	return xfs_cntbt_init_cursor(cur->bc_mp, cur->bc_tp, cur->bc_ag.agbp,
+			cur->bc_ag.pag);
+}
+
+
 STATIC void
 xfs_allocbt_set_root(
 	struct xfs_btree_cur		*cur,
@@ -38,13 +48,18 @@ xfs_allocbt_set_root(
 {
 	struct xfs_buf		*agbp = cur->bc_ag.agbp;
 	struct xfs_agf		*agf = agbp->b_addr;
-	int			btnum = cur->bc_btnum;
 
 	ASSERT(ptr->s != 0);
 
-	agf->agf_roots[btnum] = ptr->s;
-	be32_add_cpu(&agf->agf_levels[btnum], inc);
-	cur->bc_ag.pag->pagf_levels[btnum] += inc;
+	if (xfs_btree_is_bno(cur->bc_ops)) {
+		agf->agf_bno_root = ptr->s;
+		be32_add_cpu(&agf->agf_bno_level, inc);
+		cur->bc_ag.pag->pagf_bno_level += inc;
+	} else {
+		agf->agf_cnt_root = ptr->s;
+		be32_add_cpu(&agf->agf_cnt_level, inc);
+		cur->bc_ag.pag->pagf_cnt_level += inc;
+	}
 
 	xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS);
 }
@@ -116,7 +131,7 @@ xfs_allocbt_update_lastrec(
 	__be32			len;
 	int			numrecs;
 
-	ASSERT(cur->bc_btnum == XFS_BTNUM_CNT);
+	ASSERT(!xfs_btree_is_bno(cur->bc_ops));
 
 	switch (reason) {
 	case LASTREC_UPDATE:
@@ -226,7 +241,10 @@ xfs_allocbt_init_ptr_from_cur(
 
 	ASSERT(cur->bc_ag.pag->pag_agno == be32_to_cpu(agf->agf_seqno));
 
-	ptr->s = agf->agf_roots[cur->bc_btnum];
+	if (xfs_btree_is_bno(cur->bc_ops))
+		ptr->s = agf->agf_bno_root;
+	else
+		ptr->s = agf->agf_cnt_root;
 }
 
 STATIC int64_t
@@ -299,13 +317,12 @@ xfs_allocbt_verify(
 	struct xfs_perag	*pag = bp->b_pag;
 	xfs_failaddr_t		fa;
 	unsigned int		level;
-	xfs_btnum_t		btnum = XFS_BTNUM_BNOi;
 
 	if (!xfs_verify_magic(bp, block->bb_magic))
 		return __this_address;
 
 	if (xfs_has_crc(mp)) {
-		fa = xfs_btree_sblock_v5hdr_verify(bp);
+		fa = xfs_btree_agblock_v5hdr_verify(bp);
 		if (fa)
 			return fa;
 	}
@@ -320,26 +337,32 @@ xfs_allocbt_verify(
 	 * against.
 	 */
 	level = be16_to_cpu(block->bb_level);
-	if (bp->b_ops->magic[0] == cpu_to_be32(XFS_ABTC_MAGIC))
-		btnum = XFS_BTNUM_CNTi;
 	if (pag && xfs_perag_initialised_agf(pag)) {
-		unsigned int	maxlevel = pag->pagf_levels[btnum];
+		unsigned int	maxlevel, repair_maxlevel = 0;
 
-#ifdef CONFIG_XFS_ONLINE_REPAIR
 		/*
 		 * Online repair could be rewriting the free space btrees, so
 		 * we'll validate against the larger of either tree while this
 		 * is going on.
 		 */
-		maxlevel = max_t(unsigned int, maxlevel,
-				 pag->pagf_repair_levels[btnum]);
+		if (bp->b_ops->magic[0] == cpu_to_be32(XFS_ABTC_MAGIC)) {
+			maxlevel = pag->pagf_cnt_level;
+#ifdef CONFIG_XFS_ONLINE_REPAIR
+			repair_maxlevel = pag->pagf_repair_cnt_level;
+#endif
+		} else {
+			maxlevel = pag->pagf_bno_level;
+#ifdef CONFIG_XFS_ONLINE_REPAIR
+			repair_maxlevel = pag->pagf_repair_bno_level;
 #endif
-		if (level >= maxlevel)
+		}
+
+		if (level >= max(maxlevel, repair_maxlevel))
 			return __this_address;
 	} else if (level >= mp->m_alloc_maxlevels)
 		return __this_address;
 
-	return xfs_btree_sblock_verify(bp, mp->m_alloc_mxr[level != 0]);
+	return xfs_btree_agblock_verify(bp, mp->m_alloc_mxr[level != 0]);
 }
 
 static void
@@ -348,7 +371,7 @@ xfs_allocbt_read_verify(
 {
 	xfs_failaddr_t	fa;
 
-	if (!xfs_btree_sblock_verify_crc(bp))
+	if (!xfs_btree_agblock_verify_crc(bp))
 		xfs_verifier_error(bp, -EFSBADCRC, __this_address);
 	else {
 		fa = xfs_allocbt_verify(bp);
@@ -372,7 +395,7 @@ xfs_allocbt_write_verify(
 		xfs_verifier_error(bp, -EFSCORRUPTED, fa);
 		return;
 	}
-	xfs_btree_sblock_calc_crc(bp);
+	xfs_btree_agblock_calc_crc(bp);
 
 }
 
@@ -454,11 +477,19 @@ xfs_allocbt_keys_contiguous(
 				 be32_to_cpu(key2->alloc.ar_startblock));
 }
 
-static const struct xfs_btree_ops xfs_bnobt_ops = {
+const struct xfs_btree_ops xfs_bnobt_ops = {
+	.name			= "bno",
+	.type			= XFS_BTREE_TYPE_AG,
+
 	.rec_len		= sizeof(xfs_alloc_rec_t),
 	.key_len		= sizeof(xfs_alloc_key_t),
+	.ptr_len		= XFS_BTREE_SHORT_PTR_LEN,
 
-	.dup_cursor		= xfs_allocbt_dup_cursor,
+	.lru_refs		= XFS_ALLOC_BTREE_REF,
+	.statoff		= XFS_STATS_CALC_INDEX(xs_abtb_2),
+	.sick_mask		= XFS_SICK_AG_BNOBT,
+
+	.dup_cursor		= xfs_bnobt_dup_cursor,
 	.set_root		= xfs_allocbt_set_root,
 	.alloc_block		= xfs_allocbt_alloc_block,
 	.free_block		= xfs_allocbt_free_block,
@@ -477,11 +508,20 @@ static const struct xfs_btree_ops xfs_bnobt_ops = {
 	.keys_contiguous	= xfs_allocbt_keys_contiguous,
 };
 
-static const struct xfs_btree_ops xfs_cntbt_ops = {
+const struct xfs_btree_ops xfs_cntbt_ops = {
+	.name			= "cnt",
+	.type			= XFS_BTREE_TYPE_AG,
+	.geom_flags		= XFS_BTGEO_LASTREC_UPDATE,
+
 	.rec_len		= sizeof(xfs_alloc_rec_t),
 	.key_len		= sizeof(xfs_alloc_key_t),
+	.ptr_len		= XFS_BTREE_SHORT_PTR_LEN,
+
+	.lru_refs		= XFS_ALLOC_BTREE_REF,
+	.statoff		= XFS_STATS_CALC_INDEX(xs_abtc_2),
+	.sick_mask		= XFS_SICK_AG_CNTBT,
 
-	.dup_cursor		= xfs_allocbt_dup_cursor,
+	.dup_cursor		= xfs_cntbt_dup_cursor,
 	.set_root		= xfs_allocbt_set_root,
 	.alloc_block		= xfs_allocbt_alloc_block,
 	.free_block		= xfs_allocbt_free_block,
@@ -500,76 +540,55 @@ static const struct xfs_btree_ops xfs_cntbt_ops = {
 	.keys_contiguous	= NULL, /* not needed right now */
 };
 
-/* Allocate most of a new allocation btree cursor. */
-STATIC struct xfs_btree_cur *
-xfs_allocbt_init_common(
+/*
+ * Allocate a new bnobt cursor.
+ *
+ * For staging cursors tp and agbp are NULL.
+ */
+struct xfs_btree_cur *
+xfs_bnobt_init_cursor(
 	struct xfs_mount	*mp,
 	struct xfs_trans	*tp,
-	struct xfs_perag	*pag,
-	xfs_btnum_t		btnum)
+	struct xfs_buf		*agbp,
+	struct xfs_perag	*pag)
 {
 	struct xfs_btree_cur	*cur;
 
-	ASSERT(btnum == XFS_BTNUM_BNO || btnum == XFS_BTNUM_CNT);
-
-	cur = xfs_btree_alloc_cursor(mp, tp, btnum, mp->m_alloc_maxlevels,
-			xfs_allocbt_cur_cache);
-	cur->bc_ag.abt.active = false;
-
-	if (btnum == XFS_BTNUM_CNT) {
-		cur->bc_ops = &xfs_cntbt_ops;
-		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_abtc_2);
-		cur->bc_flags = XFS_BTREE_LASTREC_UPDATE;
-	} else {
-		cur->bc_ops = &xfs_bnobt_ops;
-		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_abtb_2);
-	}
-
+	cur = xfs_btree_alloc_cursor(mp, tp, &xfs_bnobt_ops,
+			mp->m_alloc_maxlevels, xfs_allocbt_cur_cache);
 	cur->bc_ag.pag = xfs_perag_hold(pag);
+	cur->bc_ag.agbp = agbp;
+	if (agbp) {
+		struct xfs_agf		*agf = agbp->b_addr;
 
-	if (xfs_has_crc(mp))
-		cur->bc_flags |= XFS_BTREE_CRC_BLOCKS;
-
+		cur->bc_nlevels = be32_to_cpu(agf->agf_bno_level);
+	}
 	return cur;
 }
 
 /*
- * Allocate a new allocation btree cursor.
+ * Allocate a new cntbt cursor.
+ *
+ * For staging cursors tp and agbp are NULL.
  */
-struct xfs_btree_cur *			/* new alloc btree cursor */
-xfs_allocbt_init_cursor(
-	struct xfs_mount	*mp,		/* file system mount point */
-	struct xfs_trans	*tp,		/* transaction pointer */
-	struct xfs_buf		*agbp,		/* buffer for agf structure */
-	struct xfs_perag	*pag,
-	xfs_btnum_t		btnum)		/* btree identifier */
-{
-	struct xfs_agf		*agf = agbp->b_addr;
-	struct xfs_btree_cur	*cur;
-
-	cur = xfs_allocbt_init_common(mp, tp, pag, btnum);
-	if (btnum == XFS_BTNUM_CNT)
-		cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]);
-	else
-		cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]);
-
-	cur->bc_ag.agbp = agbp;
-
-	return cur;
-}
-
-/* Create a free space btree cursor with a fake root for staging. */
 struct xfs_btree_cur *
-xfs_allocbt_stage_cursor(
+xfs_cntbt_init_cursor(
 	struct xfs_mount	*mp,
-	struct xbtree_afakeroot	*afake,
-	struct xfs_perag	*pag,
-	xfs_btnum_t		btnum)
+	struct xfs_trans	*tp,
+	struct xfs_buf		*agbp,
+	struct xfs_perag	*pag)
 {
 	struct xfs_btree_cur	*cur;
 
-	cur = xfs_allocbt_init_common(mp, NULL, pag, btnum);
-	xfs_btree_stage_afakeroot(cur, afake);
+	cur = xfs_btree_alloc_cursor(mp, tp, &xfs_cntbt_ops,
+			mp->m_alloc_maxlevels, xfs_allocbt_cur_cache);
+	cur->bc_ag.pag = xfs_perag_hold(pag);
+	cur->bc_ag.agbp = agbp;
+	if (agbp) {
+		struct xfs_agf		*agf = agbp->b_addr;
+
+		cur->bc_nlevels = be32_to_cpu(agf->agf_cnt_level);
+	}
 	return cur;
 }
 
@@ -588,16 +607,16 @@ xfs_allocbt_commit_staged_btree(
 
 	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
 
-	agf->agf_roots[cur->bc_btnum] = cpu_to_be32(afake->af_root);
-	agf->agf_levels[cur->bc_btnum] = cpu_to_be32(afake->af_levels);
-	xfs_alloc_log_agf(tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS);
-
-	if (cur->bc_btnum == XFS_BTNUM_BNO) {
-		xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_bnobt_ops);
+	if (xfs_btree_is_bno(cur->bc_ops)) {
+		agf->agf_bno_root = cpu_to_be32(afake->af_root);
+		agf->agf_bno_level = cpu_to_be32(afake->af_levels);
 	} else {
-		cur->bc_flags |= XFS_BTREE_LASTREC_UPDATE;
-		xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_cntbt_ops);
+		agf->agf_cnt_root = cpu_to_be32(afake->af_root);
+		agf->agf_cnt_level = cpu_to_be32(afake->af_levels);
 	}
+	xfs_alloc_log_agf(tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS);
+
+	xfs_btree_commit_afakeroot(cur, tp, agbp);
 }
 
 /* Calculate number of records in an alloc btree block. */
diff --git a/fs/xfs/libxfs/xfs_alloc_btree.h b/fs/xfs/libxfs/xfs_alloc_btree.h
index 45df893ef6..155b47f231 100644
--- a/fs/xfs/libxfs/xfs_alloc_btree.h
+++ b/fs/xfs/libxfs/xfs_alloc_btree.h
@@ -47,12 +47,12 @@ struct xbtree_afakeroot;
 		 (maxrecs) * sizeof(xfs_alloc_key_t) + \
 		 ((index) - 1) * sizeof(xfs_alloc_ptr_t)))
 
-extern struct xfs_btree_cur *xfs_allocbt_init_cursor(struct xfs_mount *mp,
+struct xfs_btree_cur *xfs_bnobt_init_cursor(struct xfs_mount *mp,
 		struct xfs_trans *tp, struct xfs_buf *bp,
-		struct xfs_perag *pag, xfs_btnum_t btnum);
-struct xfs_btree_cur *xfs_allocbt_stage_cursor(struct xfs_mount *mp,
-		struct xbtree_afakeroot *afake, struct xfs_perag *pag,
-		xfs_btnum_t btnum);
+		struct xfs_perag *pag);
+struct xfs_btree_cur *xfs_cntbt_init_cursor(struct xfs_mount *mp,
+		struct xfs_trans *tp, struct xfs_buf *bp,
+		struct xfs_perag *pag);
 extern int xfs_allocbt_maxrecs(struct xfs_mount *, int, int);
 extern xfs_extlen_t xfs_allocbt_calc_size(struct xfs_mount *mp,
 		unsigned long long len);
diff --git a/fs/xfs/libxfs/xfs_attr.c b/fs/xfs/libxfs/xfs_attr.c
index e965a48e7d..f30bcc6410 100644
--- a/fs/xfs/libxfs/xfs_attr.c
+++ b/fs/xfs/libxfs/xfs_attr.c
@@ -26,6 +26,7 @@
 #include "xfs_trace.h"
 #include "xfs_attr_item.h"
 #include "xfs_xattr.h"
+#include "xfs_parent.h"
 
 struct kmem_cache		*xfs_attr_intent_cache;
 
@@ -87,6 +88,8 @@ xfs_attr_is_leaf(
 	struct xfs_iext_cursor	icur;
 	struct xfs_bmbt_irec	imap;
 
+	ASSERT(!xfs_need_iread_extents(ifp));
+
 	if (ifp->if_nextents != 1 || ifp->if_format != XFS_DINODE_FMT_EXTENTS)
 		return false;
 
@@ -224,11 +227,21 @@ int
 xfs_attr_get_ilocked(
 	struct xfs_da_args	*args)
 {
-	ASSERT(xfs_isilocked(args->dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL));
+	int			error;
+
+	xfs_assert_ilocked(args->dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL);
 
 	if (!xfs_inode_hasattr(args->dp))
 		return -ENOATTR;
 
+	/*
+	 * The incore attr fork iext tree must be loaded for xfs_attr_is_leaf
+	 * to work correctly.
+	 */
+	error = xfs_iread_extents(args->trans, args->dp, XFS_ATTR_FORK);
+	if (error)
+		return error;
+
 	if (args->dp->i_af.if_format == XFS_DINODE_FMT_LOCAL)
 		return xfs_attr_shortform_getvalue(args);
 	if (xfs_attr_is_leaf(args->dp))
@@ -264,9 +277,11 @@ xfs_attr_get(
 	if (xfs_is_shutdown(args->dp->i_mount))
 		return -EIO;
 
+	if (!args->owner)
+		args->owner = args->dp->i_ino;
 	args->geo = args->dp->i_mount->m_attr_geo;
 	args->whichfork = XFS_ATTR_FORK;
-	args->hashval = xfs_da_hashname(args->name, args->namelen);
+	xfs_attr_sethash(args);
 
 	/* Entirely possible to look up a name which doesn't exist */
 	args->op_flags = XFS_DA_OP_OKNOENT;
@@ -314,26 +329,20 @@ xfs_attr_calc_size(
 	return nblks;
 }
 
-/* Initialize transaction reservation for attr operations */
-void
-xfs_init_attr_trans(
-	struct xfs_da_args	*args,
-	struct xfs_trans_res	*tres,
-	unsigned int		*total)
+/* Initialize transaction reservation for an xattr set/replace/upsert */
+inline struct xfs_trans_res
+xfs_attr_set_resv(
+	const struct xfs_da_args	*args)
 {
-	struct xfs_mount	*mp = args->dp->i_mount;
-
-	if (args->value) {
-		tres->tr_logres = M_RES(mp)->tr_attrsetm.tr_logres +
-				 M_RES(mp)->tr_attrsetrt.tr_logres *
-				 args->total;
-		tres->tr_logcount = XFS_ATTRSET_LOG_COUNT;
-		tres->tr_logflags = XFS_TRANS_PERM_LOG_RES;
-		*total = args->total;
-	} else {
-		*tres = M_RES(mp)->tr_attrrm;
-		*total = XFS_ATTRRM_SPACE_RES(mp);
-	}
+	struct xfs_mount		*mp = args->dp->i_mount;
+	struct xfs_trans_res		ret = {
+		.tr_logres = M_RES(mp)->tr_attrsetm.tr_logres +
+			    M_RES(mp)->tr_attrsetrt.tr_logres * args->total,
+		.tr_logcount		= XFS_ATTRSET_LOG_COUNT,
+		.tr_logflags		= XFS_TRANS_PERM_LOG_RES,
+	};
+
+	return ret;
 }
 
 /*
@@ -363,7 +372,7 @@ xfs_attr_try_sf_addname(
 	 * Commit the shortform mods, and we're done.
 	 * NOTE: this is also the error path (EEXIST, etc).
 	 */
-	if (!error && !(args->op_flags & XFS_DA_OP_NOTIME))
+	if (!error)
 		xfs_trans_ichgtime(args->trans, dp, XFS_ICHGTIME_CHG);
 
 	if (xfs_has_wsync(dp->i_mount))
@@ -401,6 +410,50 @@ out:
 	return error;
 }
 
+/* Compute the hash value for a user/root/secure extended attribute */
+xfs_dahash_t
+xfs_attr_hashname(
+	const uint8_t		*name,
+	int			namelen)
+{
+	return xfs_da_hashname(name, namelen);
+}
+
+/* Compute the hash value for any extended attribute from any namespace. */
+xfs_dahash_t
+xfs_attr_hashval(
+	struct xfs_mount	*mp,
+	unsigned int		attr_flags,
+	const uint8_t		*name,
+	int			namelen,
+	const void		*value,
+	int			valuelen)
+{
+	ASSERT(xfs_attr_check_namespace(attr_flags));
+
+	if (attr_flags & XFS_ATTR_PARENT)
+		return xfs_parent_hashattr(mp, name, namelen, value, valuelen);
+
+	return xfs_attr_hashname(name, namelen);
+}
+
+/*
+ * PPTR_REPLACE operations require the caller to set the old and new names and
+ * values explicitly.  Update the canonical fields to the new name and value
+ * here now that the removal phase has finished.
+ */
+static void
+xfs_attr_update_pptr_replace_args(
+	struct xfs_da_args	*args)
+{
+	ASSERT(args->new_namelen > 0);
+	args->name = args->new_name;
+	args->namelen = args->new_namelen;
+	args->value = args->new_value;
+	args->valuelen = args->new_valuelen;
+	xfs_attr_sethash(args);
+}
+
 /*
  * Handle the state change on completion of a multi-state attr operation.
  *
@@ -418,14 +471,15 @@ xfs_attr_complete_op(
 	enum xfs_delattr_state	replace_state)
 {
 	struct xfs_da_args	*args = attr->xattri_da_args;
-	bool			do_replace = args->op_flags & XFS_DA_OP_REPLACE;
+
+	if (!(args->op_flags & XFS_DA_OP_REPLACE))
+		replace_state = XFS_DAS_DONE;
+	else if (xfs_attr_intent_op(attr) == XFS_ATTRI_OP_FLAGS_PPTR_REPLACE)
+		xfs_attr_update_pptr_replace_args(args);
 
 	args->op_flags &= ~XFS_DA_OP_REPLACE;
 	args->attr_filter &= ~XFS_ATTR_INCOMPLETE;
-	if (do_replace)
-		return replace_state;
-
-	return XFS_DAS_DONE;
+	return replace_state;
 }
 
 static int
@@ -647,8 +701,8 @@ xfs_attr_leaf_remove_attr(
 	int				forkoff;
 	int				error;
 
-	error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno,
-				   &bp);
+	error = xfs_attr3_leaf_read(args->trans, args->dp, args->owner,
+			args->blkno, &bp);
 	if (error)
 		return error;
 
@@ -679,7 +733,7 @@ xfs_attr_leaf_shrink(
 	if (!xfs_attr_is_leaf(dp))
 		return 0;
 
-	error = xfs_attr3_leaf_read(args->trans, args->dp, 0, &bp);
+	error = xfs_attr3_leaf_read(args->trans, args->dp, args->owner, 0, &bp);
 	if (error)
 		return error;
 
@@ -868,6 +922,11 @@ xfs_attr_lookup(
 		return -ENOATTR;
 	}
 
+	/* Prerequisite for xfs_attr_is_leaf */
+	error = xfs_iread_extents(args->trans, args->dp, XFS_ATTR_FORK);
+	if (error)
+		return error;
+
 	if (xfs_attr_is_leaf(dp)) {
 		error = xfs_attr_leaf_hasname(args, &bp);
 
@@ -883,73 +942,72 @@ xfs_attr_lookup(
 	return error;
 }
 
-static void
-xfs_attr_defer_add(
-	struct xfs_da_args	*args,
-	unsigned int		op_flags)
+int
+xfs_attr_add_fork(
+	struct xfs_inode	*ip,		/* incore inode pointer */
+	int			size,		/* space new attribute needs */
+	int			rsvd)		/* xact may use reserved blks */
 {
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_trans	*tp;		/* transaction pointer */
+	unsigned int		blks;		/* space reservation */
+	int			error;		/* error return value */
 
-	struct xfs_attr_intent	*new;
+	ASSERT(!XFS_NOT_DQATTACHED(mp, ip));
 
-	new = kmem_cache_zalloc(xfs_attr_intent_cache, GFP_NOFS | __GFP_NOFAIL);
-	new->xattri_op_flags = op_flags;
-	new->xattri_da_args = args;
+	blks = XFS_ADDAFORK_SPACE_RES(mp);
 
-	switch (op_flags) {
-	case XFS_ATTRI_OP_FLAGS_SET:
-		new->xattri_dela_state = xfs_attr_init_add_state(args);
-		break;
-	case XFS_ATTRI_OP_FLAGS_REPLACE:
-		new->xattri_dela_state = xfs_attr_init_replace_state(args);
-		break;
-	case XFS_ATTRI_OP_FLAGS_REMOVE:
-		new->xattri_dela_state = xfs_attr_init_remove_state(args);
-		break;
-	default:
-		ASSERT(0);
-	}
+	error = xfs_trans_alloc_inode(ip, &M_RES(mp)->tr_addafork, blks, 0,
+			rsvd, &tp);
+	if (error)
+		return error;
+
+	if (xfs_inode_has_attr_fork(ip))
+		goto trans_cancel;
+
+	error = xfs_bmap_add_attrfork(tp, ip, size, rsvd);
+	if (error)
+		goto trans_cancel;
 
-	xfs_defer_add(args->trans, &new->xattri_list, &xfs_attr_defer_type);
-	trace_xfs_attr_defer_add(new->xattri_dela_state, args->dp);
+	error = xfs_trans_commit(tp);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	return error;
+
+trans_cancel:
+	xfs_trans_cancel(tp);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	return error;
 }
 
 /*
- * Note: If args->value is NULL the attribute will be removed, just like the
- * Linux ->setattr API.
+ * Make a change to the xattr structure.
+ *
+ * The caller must have initialized @args, attached dquots, and must not hold
+ * any ILOCKs.  Reserved data blocks may be used if @rsvd is set.
+ *
+ * Returns -EEXIST for XFS_ATTRUPDATE_CREATE if the name already exists.
+ * Returns -ENOATTR for XFS_ATTRUPDATE_REMOVE if the name does not exist.
+ * Returns 0 on success, or a negative errno if something else went wrong.
  */
 int
 xfs_attr_set(
-	struct xfs_da_args	*args)
+	struct xfs_da_args	*args,
+	enum xfs_attr_update	op,
+	bool			rsvd)
 {
 	struct xfs_inode	*dp = args->dp;
 	struct xfs_mount	*mp = dp->i_mount;
 	struct xfs_trans_res	tres;
-	bool			rsvd = (args->attr_filter & XFS_ATTR_ROOT);
 	int			error, local;
 	int			rmt_blks = 0;
-	unsigned int		total;
+	unsigned int		total = 0;
 
-	if (xfs_is_shutdown(dp->i_mount))
-		return -EIO;
+	ASSERT(!args->trans);
 
-	error = xfs_qm_dqattach(dp);
-	if (error)
-		return error;
-
-	args->geo = mp->m_attr_geo;
-	args->whichfork = XFS_ATTR_FORK;
-	args->hashval = xfs_da_hashname(args->name, args->namelen);
-
-	/*
-	 * We have no control over the attribute names that userspace passes us
-	 * to remove, so we have to allow the name lookup prior to attribute
-	 * removal to fail as well.  Preserve the logged flag, since we need
-	 * to pass that through to the logging code.
-	 */
-	args->op_flags = XFS_DA_OP_OKNOENT |
-					(args->op_flags & XFS_DA_OP_LOGGED);
-
-	if (args->value) {
+	switch (op) {
+	case XFS_ATTRUPDATE_UPSERT:
+	case XFS_ATTRUPDATE_CREATE:
+	case XFS_ATTRUPDATE_REPLACE:
 		XFS_STATS_INC(mp, xs_attr_set);
 		args->total = xfs_attr_calc_size(args, &local);
 
@@ -962,33 +1020,36 @@ xfs_attr_set(
 				xfs_attr_sf_entsize_byname(args->namelen,
 						args->valuelen);
 
-			error = xfs_bmap_add_attrfork(dp, sf_size, rsvd);
+			error = xfs_attr_add_fork(dp, sf_size, rsvd);
 			if (error)
 				return error;
 		}
 
 		if (!local)
 			rmt_blks = xfs_attr3_rmt_blocks(mp, args->valuelen);
-	} else {
+
+		tres = xfs_attr_set_resv(args);
+		total = args->total;
+		break;
+	case XFS_ATTRUPDATE_REMOVE:
 		XFS_STATS_INC(mp, xs_attr_remove);
-		rmt_blks = xfs_attr3_rmt_blocks(mp, XFS_XATTR_SIZE_MAX);
+		rmt_blks = xfs_attr3_max_rmt_blocks(mp);
+		tres = M_RES(mp)->tr_attrrm;
+		total = XFS_ATTRRM_SPACE_RES(mp);
+		break;
 	}
 
 	/*
 	 * Root fork attributes can use reserved data blocks for this
 	 * operation if necessary
 	 */
-	xfs_init_attr_trans(args, &tres, &total);
 	error = xfs_trans_alloc_inode(dp, &tres, total, 0, rsvd, &args->trans);
 	if (error)
 		return error;
 
-	if (args->value || xfs_inode_hasattr(dp)) {
-		error = xfs_iext_count_may_overflow(dp, XFS_ATTR_FORK,
+	if (op != XFS_ATTRUPDATE_REMOVE || xfs_inode_hasattr(dp)) {
+		error = xfs_iext_count_extend(args->trans, dp, XFS_ATTR_FORK,
 				XFS_IEXT_ATTR_MANIP_CNT(rmt_blks));
-		if (error == -EFBIG)
-			error = xfs_iext_count_upgrade(args->trans, dp,
-					XFS_IEXT_ATTR_MANIP_CNT(rmt_blks));
 		if (error)
 			goto out_trans_cancel;
 	}
@@ -996,26 +1057,26 @@ xfs_attr_set(
 	error = xfs_attr_lookup(args);
 	switch (error) {
 	case -EEXIST:
-		if (!args->value) {
+		if (op == XFS_ATTRUPDATE_REMOVE) {
 			/* if no value, we are performing a remove operation */
-			xfs_attr_defer_add(args, XFS_ATTRI_OP_FLAGS_REMOVE);
+			xfs_attr_defer_add(args, XFS_ATTR_DEFER_REMOVE);
 			break;
 		}
 
 		/* Pure create fails if the attr already exists */
-		if (args->attr_flags & XATTR_CREATE)
+		if (op == XFS_ATTRUPDATE_CREATE)
 			goto out_trans_cancel;
-		xfs_attr_defer_add(args, XFS_ATTRI_OP_FLAGS_REPLACE);
+		xfs_attr_defer_add(args, XFS_ATTR_DEFER_REPLACE);
 		break;
 	case -ENOATTR:
 		/* Can't remove what isn't there. */
-		if (!args->value)
+		if (op == XFS_ATTRUPDATE_REMOVE)
 			goto out_trans_cancel;
 
 		/* Pure replace fails if no existing attr to replace. */
-		if (args->attr_flags & XATTR_REPLACE)
+		if (op == XFS_ATTRUPDATE_REPLACE)
 			goto out_trans_cancel;
-		xfs_attr_defer_add(args, XFS_ATTRI_OP_FLAGS_SET);
+		xfs_attr_defer_add(args, XFS_ATTR_DEFER_SET);
 		break;
 	default:
 		goto out_trans_cancel;
@@ -1028,8 +1089,7 @@ xfs_attr_set(
 	if (xfs_has_wsync(mp))
 		xfs_trans_set_sync(args->trans);
 
-	if (!(args->op_flags & XFS_DA_OP_NOTIME))
-		xfs_trans_ichgtime(args->trans, dp, XFS_ICHGTIME_CHG);
+	xfs_trans_ichgtime(args->trans, dp, XFS_ICHGTIME_CHG);
 
 	/*
 	 * Commit the last in the sequence of transactions.
@@ -1038,6 +1098,7 @@ xfs_attr_set(
 	error = xfs_trans_commit(args->trans);
 out_unlock:
 	xfs_iunlock(dp, XFS_ILOCK_EXCL);
+	args->trans = NULL;
 	return error;
 
 out_trans_cancel:
@@ -1050,7 +1111,7 @@ out_trans_cancel:
  * External routines when attribute list is inside the inode
  *========================================================================*/
 
-static inline int xfs_attr_sf_totsize(struct xfs_inode *dp)
+int xfs_attr_sf_totsize(struct xfs_inode *dp)
 {
 	struct xfs_attr_sf_hdr *sf = dp->i_af.if_data;
 
@@ -1153,7 +1214,7 @@ xfs_attr_leaf_try_add(
 	struct xfs_buf		*bp;
 	int			error;
 
-	error = xfs_attr3_leaf_read(args->trans, args->dp, 0, &bp);
+	error = xfs_attr3_leaf_read(args->trans, args->dp, args->owner, 0, &bp);
 	if (error)
 		return error;
 
@@ -1201,7 +1262,7 @@ xfs_attr_leaf_hasname(
 {
 	int                     error = 0;
 
-	error = xfs_attr3_leaf_read(args->trans, args->dp, 0, bp);
+	error = xfs_attr3_leaf_read(args->trans, args->dp, args->owner, 0, bp);
 	if (error)
 		return error;
 
@@ -1510,12 +1571,23 @@ out_release:
 	return error;
 }
 
+/* Enforce that there is at most one namespace bit per attr. */
+inline bool xfs_attr_check_namespace(unsigned int attr_flags)
+{
+	return hweight32(attr_flags & XFS_ATTR_NSP_ONDISK_MASK) < 2;
+}
+
 /* Returns true if the attribute entry name is valid. */
 bool
 xfs_attr_namecheck(
+	unsigned int	attr_flags,
 	const void	*name,
 	size_t		length)
 {
+	/* Only one namespace bit allowed. */
+	if (!xfs_attr_check_namespace(attr_flags))
+		return false;
+
 	/*
 	 * MAXNAMELEN includes the trailing null, but (name/length) leave it
 	 * out, so use >= for the length check.
@@ -1523,6 +1595,10 @@ xfs_attr_namecheck(
 	if (length >= MAXNAMELEN)
 		return false;
 
+	/* Parent pointers have their own validation. */
+	if (attr_flags & XFS_ATTR_PARENT)
+		return xfs_parent_namecheck(attr_flags, name, length);
+
 	/* There shouldn't be any nulls here */
 	return !memchr(name, 0, length);
 }
diff --git a/fs/xfs/libxfs/xfs_attr.h b/fs/xfs/libxfs/xfs_attr.h
index 81be9b3e40..0e51d0723f 100644
--- a/fs/xfs/libxfs/xfs_attr.h
+++ b/fs/xfs/libxfs/xfs_attr.h
@@ -47,8 +47,9 @@ struct xfs_attrlist_cursor_kern {
 
 
 /* void; state communicated via *context */
-typedef void (*put_listent_func_t)(struct xfs_attr_list_context *, int,
-			      unsigned char *, int, int);
+typedef void (*put_listent_func_t)(struct xfs_attr_list_context *context,
+		int flags, unsigned char *name, int namelen, void *value,
+		int valuelen);
 
 struct xfs_attr_list_context {
 	struct xfs_trans	*tp;
@@ -510,8 +511,8 @@ struct xfs_attr_intent {
 	struct xfs_da_args		*xattri_da_args;
 
 	/*
-	 * Shared buffer containing the attr name and value so that the logging
-	 * code can share large memory buffers between log items.
+	 * Shared buffer containing the attr name, new name, and value so that
+	 * the logging code can share large memory buffers between log items.
 	 */
 	struct xfs_attri_log_nameval	*xattri_nameval;
 
@@ -529,6 +530,11 @@ struct xfs_attr_intent {
 	struct xfs_bmbt_irec		xattri_map;
 };
 
+static inline unsigned int
+xfs_attr_intent_op(const struct xfs_attr_intent *attr)
+{
+	return attr->xattri_op_flags & XFS_ATTRI_OP_FLAGS_TYPE_MASK;
+}
 
 /*========================================================================
  * Function prototypes for the kernel.
@@ -544,13 +550,22 @@ int xfs_inode_hasattr(struct xfs_inode *ip);
 bool xfs_attr_is_leaf(struct xfs_inode *ip);
 int xfs_attr_get_ilocked(struct xfs_da_args *args);
 int xfs_attr_get(struct xfs_da_args *args);
-int xfs_attr_set(struct xfs_da_args *args);
+
+enum xfs_attr_update {
+	XFS_ATTRUPDATE_REMOVE,	/* remove attr */
+	XFS_ATTRUPDATE_UPSERT,	/* set value, replace any existing attr */
+	XFS_ATTRUPDATE_CREATE,	/* set value, fail if attr already exists */
+	XFS_ATTRUPDATE_REPLACE,	/* set value, fail if attr does not exist */
+};
+
+int xfs_attr_set(struct xfs_da_args *args, enum xfs_attr_update op, bool rsvd);
 int xfs_attr_set_iter(struct xfs_attr_intent *attr);
 int xfs_attr_remove_iter(struct xfs_attr_intent *attr);
-bool xfs_attr_namecheck(const void *name, size_t length);
+bool xfs_attr_check_namespace(unsigned int attr_flags);
+bool xfs_attr_namecheck(unsigned int attr_flags, const void *name,
+		size_t length);
 int xfs_attr_calc_size(struct xfs_da_args *args, int *local);
-void xfs_init_attr_trans(struct xfs_da_args *args, struct xfs_trans_res *tres,
-			 unsigned int *total);
+struct xfs_trans_res xfs_attr_set_resv(const struct xfs_da_args *args);
 
 /*
  * Check to see if the attr should be upgraded from non-existent or shortform to
@@ -590,7 +605,6 @@ xfs_attr_init_add_state(struct xfs_da_args *args)
 static inline enum xfs_delattr_state
 xfs_attr_init_remove_state(struct xfs_da_args *args)
 {
-	args->op_flags |= XFS_DA_OP_REMOVE;
 	if (xfs_attr_is_shortform(args->dp))
 		return XFS_DAS_SF_REMOVE;
 	if (xfs_attr_is_leaf(args->dp))
@@ -614,8 +628,25 @@ xfs_attr_init_replace_state(struct xfs_da_args *args)
 	return xfs_attr_init_add_state(args);
 }
 
+xfs_dahash_t xfs_attr_hashname(const uint8_t *name, int namelen);
+
+xfs_dahash_t xfs_attr_hashval(struct xfs_mount *mp, unsigned int attr_flags,
+		const uint8_t *name, int namelen, const void *value,
+		int valuelen);
+
+/* Set the hash value for any extended attribute from any namespace. */
+static inline void xfs_attr_sethash(struct xfs_da_args *args)
+{
+	args->hashval = xfs_attr_hashval(args->dp->i_mount, args->attr_filter,
+					 args->name, args->namelen,
+					 args->value, args->valuelen);
+}
+
 extern struct kmem_cache *xfs_attr_intent_cache;
 int __init xfs_attr_intent_init_cache(void);
 void xfs_attr_intent_destroy_cache(void);
 
+int xfs_attr_sf_totsize(struct xfs_inode *dp);
+int xfs_attr_add_fork(struct xfs_inode *ip, int size, int rsvd);
+
 #endif	/* __XFS_ATTR_H__ */
diff --git a/fs/xfs/libxfs/xfs_attr_leaf.c b/fs/xfs/libxfs/xfs_attr_leaf.c
index 6374bf1072..b9e98950eb 100644
--- a/fs/xfs/libxfs/xfs_attr_leaf.c
+++ b/fs/xfs/libxfs/xfs_attr_leaf.c
@@ -29,6 +29,7 @@
 #include "xfs_log.h"
 #include "xfs_ag.h"
 #include "xfs_errortag.h"
+#include "xfs_health.h"
 
 
 /*
@@ -387,6 +388,27 @@ xfs_attr3_leaf_verify(
 	return NULL;
 }
 
+xfs_failaddr_t
+xfs_attr3_leaf_header_check(
+	struct xfs_buf		*bp,
+	xfs_ino_t		owner)
+{
+	struct xfs_mount	*mp = bp->b_mount;
+
+	if (xfs_has_crc(mp)) {
+		struct xfs_attr3_leafblock *hdr3 = bp->b_addr;
+
+		if (hdr3->hdr.info.hdr.magic !=
+				cpu_to_be16(XFS_ATTR3_LEAF_MAGIC))
+			return __this_address;
+
+		if (be64_to_cpu(hdr3->hdr.info.owner) != owner)
+			return __this_address;
+	}
+
+	return NULL;
+}
+
 static void
 xfs_attr3_leaf_write_verify(
 	struct xfs_buf	*bp)
@@ -447,16 +469,30 @@ int
 xfs_attr3_leaf_read(
 	struct xfs_trans	*tp,
 	struct xfs_inode	*dp,
+	xfs_ino_t		owner,
 	xfs_dablk_t		bno,
 	struct xfs_buf		**bpp)
 {
+	xfs_failaddr_t		fa;
 	int			err;
 
 	err = xfs_da_read_buf(tp, dp, bno, 0, bpp, XFS_ATTR_FORK,
 			&xfs_attr3_leaf_buf_ops);
-	if (!err && tp && *bpp)
+	if (err || !(*bpp))
+		return err;
+
+	fa = xfs_attr3_leaf_header_check(*bpp, owner);
+	if (fa) {
+		__xfs_buf_mark_corrupt(*bpp, fa);
+		xfs_trans_brelse(tp, *bpp);
+		*bpp = NULL;
+		xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK);
+		return -EFSCORRUPTED;
+	}
+
+	if (tp)
 		xfs_trans_buf_set_type(tp, *bpp, XFS_BLFT_ATTR_LEAF_BUF);
-	return err;
+	return 0;
 }
 
 /*========================================================================
@@ -471,28 +507,57 @@ xfs_attr3_leaf_read(
  * INCOMPLETE flag will not be set in attr->attr_filter, but rather
  * XFS_DA_OP_RECOVERY will be set in args->op_flags.
  */
+static inline unsigned int xfs_attr_match_mask(const struct xfs_da_args *args)
+{
+	if (args->op_flags & XFS_DA_OP_RECOVERY)
+		return XFS_ATTR_NSP_ONDISK_MASK;
+	return XFS_ATTR_NSP_ONDISK_MASK | XFS_ATTR_INCOMPLETE;
+}
+
+static inline bool
+xfs_attr_parent_match(
+	const struct xfs_da_args	*args,
+	const void			*value,
+	unsigned int			valuelen)
+{
+	ASSERT(args->value != NULL);
+
+	/* Parent pointers do not use remote values */
+	if (!value)
+		return false;
+
+	/*
+	 * The only value we support is a parent rec.  However, we'll accept
+	 * any valuelen so that offline repair can delete ATTR_PARENT values
+	 * that are not parent pointers.
+	 */
+	if (valuelen != args->valuelen)
+		return false;
+
+	return memcmp(args->value, value, valuelen) == 0;
+}
+
 static bool
 xfs_attr_match(
 	struct xfs_da_args	*args,
-	uint8_t			namelen,
-	unsigned char		*name,
-	int			flags)
+	unsigned int		attr_flags,
+	const unsigned char	*name,
+	unsigned int		namelen,
+	const void		*value,
+	unsigned int		valuelen)
 {
+	unsigned int		mask = xfs_attr_match_mask(args);
 
 	if (args->namelen != namelen)
 		return false;
+	if ((args->attr_filter & mask) != (attr_flags & mask))
+		return false;
 	if (memcmp(args->name, name, namelen) != 0)
 		return false;
 
-	/* Recovery ignores the INCOMPLETE flag. */
-	if ((args->op_flags & XFS_DA_OP_RECOVERY) &&
-	    args->attr_filter == (flags & XFS_ATTR_NSP_ONDISK_MASK))
-		return true;
+	if (attr_flags & XFS_ATTR_PARENT)
+		return xfs_attr_parent_match(args, value, valuelen);
 
-	/* All remaining matches need to be filtered by INCOMPLETE state. */
-	if (args->attr_filter !=
-	    (flags & (XFS_ATTR_NSP_ONDISK_MASK | XFS_ATTR_INCOMPLETE)))
-		return false;
 	return true;
 }
 
@@ -503,6 +568,13 @@ xfs_attr_copy_value(
 	int			valuelen)
 {
 	/*
+	 * Parent pointer lookups require the caller to specify the name and
+	 * value, so don't copy anything.
+	 */
+	if (args->attr_filter & XFS_ATTR_PARENT)
+		return 0;
+
+	/*
 	 * No copy if all we have to do is get the length
 	 */
 	if (!args->valuelen) {
@@ -710,8 +782,9 @@ xfs_attr_sf_findname(
 	for (sfe = xfs_attr_sf_firstentry(sf);
 	     sfe < xfs_attr_sf_endptr(sf);
 	     sfe = xfs_attr_sf_nextentry(sfe)) {
-		if (xfs_attr_match(args, sfe->namelen, sfe->nameval,
-				sfe->flags))
+		if (xfs_attr_match(args, sfe->flags, sfe->nameval,
+				sfe->namelen, &sfe->nameval[sfe->namelen],
+				sfe->valuelen))
 			return sfe;
 	}
 
@@ -818,7 +891,8 @@ xfs_attr_sf_removename(
 	 */
 	if (totsize == sizeof(struct xfs_attr_sf_hdr) && xfs_has_attr2(mp) &&
 	    (dp->i_df.if_format != XFS_DINODE_FMT_BTREE) &&
-	    !(args->op_flags & (XFS_DA_OP_ADDNAME | XFS_DA_OP_REPLACE))) {
+	    !(args->op_flags & (XFS_DA_OP_ADDNAME | XFS_DA_OP_REPLACE)) &&
+	    !xfs_has_parent(mp)) {
 		xfs_attr_fork_remove(dp, args->trans);
 	} else {
 		xfs_idata_realloc(dp, -size, XFS_ATTR_FORK);
@@ -827,7 +901,8 @@ xfs_attr_sf_removename(
 		ASSERT(totsize > sizeof(struct xfs_attr_sf_hdr) ||
 				(args->op_flags & XFS_DA_OP_ADDNAME) ||
 				!xfs_has_attr2(mp) ||
-				dp->i_df.if_format == XFS_DINODE_FMT_BTREE);
+				dp->i_df.if_format == XFS_DINODE_FMT_BTREE ||
+				xfs_has_parent(mp));
 		xfs_trans_log_inode(args->trans, dp,
 					XFS_ILOG_CORE | XFS_ILOG_ADATA);
 	}
@@ -879,8 +954,7 @@ xfs_attr_shortform_to_leaf(
 
 	trace_xfs_attr_sf_to_leaf(args);
 
-	tmpbuffer = kmem_alloc(size, 0);
-	ASSERT(tmpbuffer != NULL);
+	tmpbuffer = kmalloc(size, GFP_KERNEL | __GFP_NOFAIL);
 	memcpy(tmpbuffer, ifp->if_data, size);
 	sf = (struct xfs_attr_sf_hdr *)tmpbuffer;
 
@@ -904,6 +978,7 @@ xfs_attr_shortform_to_leaf(
 	nargs.whichfork = XFS_ATTR_FORK;
 	nargs.trans = args->trans;
 	nargs.op_flags = XFS_DA_OP_OKNOENT;
+	nargs.owner = args->owner;
 
 	sfe = xfs_attr_sf_firstentry(sf);
 	for (i = 0; i < sf->count; i++) {
@@ -911,9 +986,13 @@ xfs_attr_shortform_to_leaf(
 		nargs.namelen = sfe->namelen;
 		nargs.value = &sfe->nameval[nargs.namelen];
 		nargs.valuelen = sfe->valuelen;
-		nargs.hashval = xfs_da_hashname(sfe->nameval,
-						sfe->namelen);
 		nargs.attr_filter = sfe->flags & XFS_ATTR_NSP_ONDISK_MASK;
+		if (!xfs_attr_check_namespace(sfe->flags)) {
+			xfs_da_mark_sick(args);
+			error = -EFSCORRUPTED;
+			goto out;
+		}
+		xfs_attr_sethash(&nargs);
 		error = xfs_attr3_leaf_lookup_int(bp, &nargs); /* set a->index */
 		ASSERT(error == -ENOATTR);
 		error = xfs_attr3_leaf_add(bp, &nargs);
@@ -924,7 +1003,7 @@ xfs_attr_shortform_to_leaf(
 	}
 	error = 0;
 out:
-	kmem_free(tmpbuffer);
+	kfree(tmpbuffer);
 	return error;
 }
 
@@ -1027,7 +1106,7 @@ xfs_attr_shortform_verify(
 		 * one namespace flag per xattr, so we can just count the
 		 * bits (i.e. hweight) here.
 		 */
-		if (hweight8(sfep->flags & XFS_ATTR_NSP_ONDISK_MASK) > 1)
+		if (!xfs_attr_check_namespace(sfep->flags))
 			return __this_address;
 
 		sfep = next_sfep;
@@ -1059,7 +1138,7 @@ xfs_attr3_leaf_to_shortform(
 
 	trace_xfs_attr_leaf_to_sf(args);
 
-	tmpbuffer = kmem_alloc(args->geo->blksize, 0);
+	tmpbuffer = kmalloc(args->geo->blksize, GFP_KERNEL | __GFP_NOFAIL);
 	if (!tmpbuffer)
 		return -ENOMEM;
 
@@ -1106,6 +1185,7 @@ xfs_attr3_leaf_to_shortform(
 	nargs.whichfork = XFS_ATTR_FORK;
 	nargs.trans = args->trans;
 	nargs.op_flags = XFS_DA_OP_OKNOENT;
+	nargs.owner = args->owner;
 
 	for (i = 0; i < ichdr.count; entry++, i++) {
 		if (entry->flags & XFS_ATTR_INCOMPLETE)
@@ -1125,7 +1205,7 @@ xfs_attr3_leaf_to_shortform(
 	error = 0;
 
 out:
-	kmem_free(tmpbuffer);
+	kfree(tmpbuffer);
 	return error;
 }
 
@@ -1158,7 +1238,7 @@ xfs_attr3_leaf_to_node(
 	error = xfs_da_grow_inode(args, &blkno);
 	if (error)
 		goto out;
-	error = xfs_attr3_leaf_read(args->trans, dp, 0, &bp1);
+	error = xfs_attr3_leaf_read(args->trans, dp, args->owner, 0, &bp1);
 	if (error)
 		goto out;
 
@@ -1237,7 +1317,7 @@ xfs_attr3_leaf_create(
 		ichdr.magic = XFS_ATTR3_LEAF_MAGIC;
 
 		hdr3->blkno = cpu_to_be64(xfs_buf_daddr(bp));
-		hdr3->owner = cpu_to_be64(dp->i_ino);
+		hdr3->owner = cpu_to_be64(args->owner);
 		uuid_copy(&hdr3->uuid, &mp->m_sb.sb_meta_uuid);
 
 		ichdr.freemap[0].base = sizeof(struct xfs_attr3_leaf_hdr);
@@ -1533,7 +1613,7 @@ xfs_attr3_leaf_compact(
 
 	trace_xfs_attr_leaf_compact(args);
 
-	tmpbuffer = kmem_alloc(args->geo->blksize, 0);
+	tmpbuffer = kmalloc(args->geo->blksize, GFP_KERNEL | __GFP_NOFAIL);
 	memcpy(tmpbuffer, bp->b_addr, args->geo->blksize);
 	memset(bp->b_addr, 0, args->geo->blksize);
 	leaf_src = (xfs_attr_leafblock_t *)tmpbuffer;
@@ -1571,7 +1651,7 @@ xfs_attr3_leaf_compact(
 	 */
 	xfs_trans_log_buf(trans, bp, 0, args->geo->blksize - 1);
 
-	kmem_free(tmpbuffer);
+	kfree(tmpbuffer);
 }
 
 /*
@@ -1993,7 +2073,7 @@ xfs_attr3_leaf_toosmall(
 		if (blkno == 0)
 			continue;
 		error = xfs_attr3_leaf_read(state->args->trans, state->args->dp,
-					blkno, &bp);
+					state->args->owner, blkno, &bp);
 		if (error)
 			return error;
 
@@ -2250,7 +2330,8 @@ xfs_attr3_leaf_unbalance(
 		struct xfs_attr_leafblock *tmp_leaf;
 		struct xfs_attr3_icleaf_hdr tmphdr;
 
-		tmp_leaf = kmem_zalloc(state->args->geo->blksize, 0);
+		tmp_leaf = kzalloc(state->args->geo->blksize,
+				GFP_KERNEL | __GFP_NOFAIL);
 
 		/*
 		 * Copy the header into the temp leaf so that all the stuff
@@ -2290,7 +2371,7 @@ xfs_attr3_leaf_unbalance(
 		}
 		memcpy(save_leaf, tmp_leaf, state->args->geo->blksize);
 		savehdr = tmphdr; /* struct copy */
-		kmem_free(tmp_leaf);
+		kfree(tmp_leaf);
 	}
 
 	xfs_attr3_leaf_hdr_to_disk(state->args->geo, save_leaf, &savehdr);
@@ -2343,6 +2424,7 @@ xfs_attr3_leaf_lookup_int(
 	entries = xfs_attr3_leaf_entryp(leaf);
 	if (ichdr.count >= args->geo->blksize / 8) {
 		xfs_buf_mark_corrupt(bp);
+		xfs_da_mark_sick(args);
 		return -EFSCORRUPTED;
 	}
 
@@ -2362,10 +2444,12 @@ xfs_attr3_leaf_lookup_int(
 	}
 	if (!(probe >= 0 && (!ichdr.count || probe < ichdr.count))) {
 		xfs_buf_mark_corrupt(bp);
+		xfs_da_mark_sick(args);
 		return -EFSCORRUPTED;
 	}
 	if (!(span <= 4 || be32_to_cpu(entry->hashval) == hashval)) {
 		xfs_buf_mark_corrupt(bp);
+		xfs_da_mark_sick(args);
 		return -EFSCORRUPTED;
 	}
 
@@ -2397,18 +2481,23 @@ xfs_attr3_leaf_lookup_int(
  */
 		if (entry->flags & XFS_ATTR_LOCAL) {
 			name_loc = xfs_attr3_leaf_name_local(leaf, probe);
-			if (!xfs_attr_match(args, name_loc->namelen,
-					name_loc->nameval, entry->flags))
+			if (!xfs_attr_match(args, entry->flags,
+					name_loc->nameval, name_loc->namelen,
+					&name_loc->nameval[name_loc->namelen],
+					be16_to_cpu(name_loc->valuelen)))
 				continue;
 			args->index = probe;
 			return -EEXIST;
 		} else {
+			unsigned int	valuelen;
+
 			name_rmt = xfs_attr3_leaf_name_remote(leaf, probe);
-			if (!xfs_attr_match(args, name_rmt->namelen,
-					name_rmt->name, entry->flags))
+			valuelen = be32_to_cpu(name_rmt->valuelen);
+			if (!xfs_attr_match(args, entry->flags, name_rmt->name,
+					name_rmt->namelen, NULL, valuelen))
 				continue;
 			args->index = probe;
-			args->rmtvaluelen = be32_to_cpu(name_rmt->valuelen);
+			args->rmtvaluelen = valuelen;
 			args->rmtblkno = be32_to_cpu(name_rmt->valueblk);
 			args->rmtblkcnt = xfs_attr3_rmt_blocks(
 							args->dp->i_mount,
@@ -2711,7 +2800,8 @@ xfs_attr3_leaf_clearflag(
 	/*
 	 * Set up the operation.
 	 */
-	error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno, &bp);
+	error = xfs_attr3_leaf_read(args->trans, args->dp, args->owner,
+			args->blkno, &bp);
 	if (error)
 		return error;
 
@@ -2775,7 +2865,8 @@ xfs_attr3_leaf_setflag(
 	/*
 	 * Set up the operation.
 	 */
-	error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno, &bp);
+	error = xfs_attr3_leaf_read(args->trans, args->dp, args->owner,
+			args->blkno, &bp);
 	if (error)
 		return error;
 
@@ -2834,7 +2925,8 @@ xfs_attr3_leaf_flipflags(
 	/*
 	 * Read the block containing the "old" attr
 	 */
-	error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno, &bp1);
+	error = xfs_attr3_leaf_read(args->trans, args->dp, args->owner,
+			args->blkno, &bp1);
 	if (error)
 		return error;
 
@@ -2842,8 +2934,8 @@ xfs_attr3_leaf_flipflags(
 	 * Read the block containing the "new" attr, if it is different
 	 */
 	if (args->blkno2 != args->blkno) {
-		error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno2,
-					   &bp2);
+		error = xfs_attr3_leaf_read(args->trans, args->dp, args->owner,
+				args->blkno2, &bp2);
 		if (error)
 			return error;
 	} else {
diff --git a/fs/xfs/libxfs/xfs_attr_leaf.h b/fs/xfs/libxfs/xfs_attr_leaf.h
index 9b9948639c..bac2195898 100644
--- a/fs/xfs/libxfs/xfs_attr_leaf.h
+++ b/fs/xfs/libxfs/xfs_attr_leaf.h
@@ -98,12 +98,14 @@ int	xfs_attr_leaf_order(struct xfs_buf *leaf1_bp,
 				   struct xfs_buf *leaf2_bp);
 int	xfs_attr_leaf_newentsize(struct xfs_da_args *args, int *local);
 int	xfs_attr3_leaf_read(struct xfs_trans *tp, struct xfs_inode *dp,
-			xfs_dablk_t bno, struct xfs_buf **bpp);
+			xfs_ino_t owner, xfs_dablk_t bno, struct xfs_buf **bpp);
 void	xfs_attr3_leaf_hdr_from_disk(struct xfs_da_geometry *geo,
 				     struct xfs_attr3_icleaf_hdr *to,
 				     struct xfs_attr_leafblock *from);
 void	xfs_attr3_leaf_hdr_to_disk(struct xfs_da_geometry *geo,
 				   struct xfs_attr_leafblock *to,
 				   struct xfs_attr3_icleaf_hdr *from);
+xfs_failaddr_t xfs_attr3_leaf_header_check(struct xfs_buf *bp,
+		xfs_ino_t owner);
 
 #endif	/* __XFS_ATTR_LEAF_H__ */
diff --git a/fs/xfs/libxfs/xfs_attr_remote.c b/fs/xfs/libxfs/xfs_attr_remote.c
index d440393b40..4c44ce1c8a 100644
--- a/fs/xfs/libxfs/xfs_attr_remote.c
+++ b/fs/xfs/libxfs/xfs_attr_remote.c
@@ -22,6 +22,7 @@
 #include "xfs_attr_remote.h"
 #include "xfs_trace.h"
 #include "xfs_error.h"
+#include "xfs_health.h"
 
 #define ATTR_RMTVALUE_MAPSIZE	1	/* # of map entries at once */
 
@@ -42,19 +43,32 @@
  * the logging system and therefore never have a log item.
  */
 
-/*
- * Each contiguous block has a header, so it is not just a simple attribute
- * length to FSB conversion.
- */
-int
+/* How many bytes can be stored in a remote value buffer? */
+inline unsigned int
+xfs_attr3_rmt_buf_space(
+	struct xfs_mount	*mp)
+{
+	unsigned int		blocksize = mp->m_attr_geo->blksize;
+
+	if (xfs_has_crc(mp))
+		return blocksize - sizeof(struct xfs_attr3_rmt_hdr);
+
+	return blocksize;
+}
+
+/* Compute number of fsblocks needed to store a remote attr value */
+unsigned int
 xfs_attr3_rmt_blocks(
-	struct xfs_mount *mp,
-	int		attrlen)
+	struct xfs_mount	*mp,
+	unsigned int		attrlen)
 {
-	if (xfs_has_crc(mp)) {
-		int buflen = XFS_ATTR3_RMT_BUF_SPACE(mp, mp->m_sb.sb_blocksize);
-		return (attrlen + buflen - 1) / buflen;
-	}
+	/*
+	 * Each contiguous block has a header, so it is not just a simple
+	 * attribute length to FSB conversion.
+	 */
+	if (xfs_has_crc(mp))
+		return howmany(attrlen, xfs_attr3_rmt_buf_space(mp));
+
 	return XFS_B_TO_FSB(mp, attrlen);
 }
 
@@ -91,7 +105,6 @@ xfs_attr3_rmt_verify(
 	struct xfs_mount	*mp,
 	struct xfs_buf		*bp,
 	void			*ptr,
-	int			fsbsize,
 	xfs_daddr_t		bno)
 {
 	struct xfs_attr3_rmt_hdr *rmt = ptr;
@@ -102,7 +115,7 @@ xfs_attr3_rmt_verify(
 		return __this_address;
 	if (be64_to_cpu(rmt->rm_blkno) != bno)
 		return __this_address;
-	if (be32_to_cpu(rmt->rm_bytes) > fsbsize - sizeof(*rmt))
+	if (be32_to_cpu(rmt->rm_bytes) > mp->m_attr_geo->blksize - sizeof(*rmt))
 		return __this_address;
 	if (be32_to_cpu(rmt->rm_offset) +
 				be32_to_cpu(rmt->rm_bytes) > XFS_XATTR_SIZE_MAX)
@@ -121,9 +134,9 @@ __xfs_attr3_rmt_read_verify(
 {
 	struct xfs_mount *mp = bp->b_mount;
 	char		*ptr;
-	int		len;
+	unsigned int	len;
 	xfs_daddr_t	bno;
-	int		blksize = mp->m_attr_geo->blksize;
+	unsigned int	blksize = mp->m_attr_geo->blksize;
 
 	/* no verification of non-crc buffers */
 	if (!xfs_has_crc(mp))
@@ -140,7 +153,7 @@ __xfs_attr3_rmt_read_verify(
 			*failaddr = __this_address;
 			return -EFSBADCRC;
 		}
-		*failaddr = xfs_attr3_rmt_verify(mp, bp, ptr, blksize, bno);
+		*failaddr = xfs_attr3_rmt_verify(mp, bp, ptr, bno);
 		if (*failaddr)
 			return -EFSCORRUPTED;
 		len -= blksize;
@@ -185,7 +198,7 @@ xfs_attr3_rmt_write_verify(
 {
 	struct xfs_mount *mp = bp->b_mount;
 	xfs_failaddr_t	fa;
-	int		blksize = mp->m_attr_geo->blksize;
+	unsigned int	blksize = mp->m_attr_geo->blksize;
 	char		*ptr;
 	int		len;
 	xfs_daddr_t	bno;
@@ -202,7 +215,7 @@ xfs_attr3_rmt_write_verify(
 	while (len > 0) {
 		struct xfs_attr3_rmt_hdr *rmt = (struct xfs_attr3_rmt_hdr *)ptr;
 
-		fa = xfs_attr3_rmt_verify(mp, bp, ptr, blksize, bno);
+		fa = xfs_attr3_rmt_verify(mp, bp, ptr, bno);
 		if (fa) {
 			xfs_verifier_error(bp, -EFSCORRUPTED, fa);
 			return;
@@ -276,32 +289,34 @@ xfs_attr3_rmt_hdr_set(
  */
 STATIC int
 xfs_attr_rmtval_copyout(
-	struct xfs_mount *mp,
-	struct xfs_buf	*bp,
-	xfs_ino_t	ino,
-	int		*offset,
-	int		*valuelen,
-	uint8_t		**dst)
+	struct xfs_mount	*mp,
+	struct xfs_buf		*bp,
+	struct xfs_inode	*dp,
+	xfs_ino_t		owner,
+	unsigned int		*offset,
+	unsigned int		*valuelen,
+	uint8_t			**dst)
 {
-	char		*src = bp->b_addr;
-	xfs_daddr_t	bno = xfs_buf_daddr(bp);
-	int		len = BBTOB(bp->b_length);
-	int		blksize = mp->m_attr_geo->blksize;
+	char			*src = bp->b_addr;
+	xfs_daddr_t		bno = xfs_buf_daddr(bp);
+	unsigned int		len = BBTOB(bp->b_length);
+	unsigned int		blksize = mp->m_attr_geo->blksize;
 
 	ASSERT(len >= blksize);
 
 	while (len > 0 && *valuelen > 0) {
-		int hdr_size = 0;
-		int byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, blksize);
+		unsigned int hdr_size = 0;
+		unsigned int byte_cnt = xfs_attr3_rmt_buf_space(mp);
 
 		byte_cnt = min(*valuelen, byte_cnt);
 
 		if (xfs_has_crc(mp)) {
-			if (xfs_attr3_rmt_hdr_ok(src, ino, *offset,
+			if (xfs_attr3_rmt_hdr_ok(src, owner, *offset,
 						  byte_cnt, bno)) {
 				xfs_alert(mp,
 "remote attribute header mismatch bno/off/len/owner (0x%llx/0x%x/Ox%x/0x%llx)",
-					bno, *offset, byte_cnt, ino);
+					bno, *offset, byte_cnt, owner);
+				xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK);
 				return -EFSCORRUPTED;
 			}
 			hdr_size = sizeof(struct xfs_attr3_rmt_hdr);
@@ -327,20 +342,20 @@ xfs_attr_rmtval_copyin(
 	struct xfs_mount *mp,
 	struct xfs_buf	*bp,
 	xfs_ino_t	ino,
-	int		*offset,
-	int		*valuelen,
+	unsigned int	*offset,
+	unsigned int	*valuelen,
 	uint8_t		**src)
 {
 	char		*dst = bp->b_addr;
 	xfs_daddr_t	bno = xfs_buf_daddr(bp);
-	int		len = BBTOB(bp->b_length);
-	int		blksize = mp->m_attr_geo->blksize;
+	unsigned int	len = BBTOB(bp->b_length);
+	unsigned int	blksize = mp->m_attr_geo->blksize;
 
 	ASSERT(len >= blksize);
 
 	while (len > 0 && *valuelen > 0) {
-		int hdr_size;
-		int byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, blksize);
+		unsigned int hdr_size;
+		unsigned int byte_cnt = xfs_attr3_rmt_buf_space(mp);
 
 		byte_cnt = min(*valuelen, byte_cnt);
 		hdr_size = xfs_attr3_rmt_hdr_set(mp, dst, ino, *offset,
@@ -386,12 +401,12 @@ xfs_attr_rmtval_get(
 	struct xfs_buf		*bp;
 	xfs_dablk_t		lblkno = args->rmtblkno;
 	uint8_t			*dst = args->value;
-	int			valuelen;
+	unsigned int		valuelen;
 	int			nmap;
 	int			error;
-	int			blkcnt = args->rmtblkcnt;
+	unsigned int		blkcnt = args->rmtblkcnt;
 	int			i;
-	int			offset = 0;
+	unsigned int		offset = 0;
 
 	trace_xfs_attr_rmtval_get(args);
 
@@ -418,12 +433,13 @@ xfs_attr_rmtval_get(
 			dblkcnt = XFS_FSB_TO_BB(mp, map[i].br_blockcount);
 			error = xfs_buf_read(mp->m_ddev_targp, dblkno, dblkcnt,
 					0, &bp, &xfs_attr3_rmt_buf_ops);
+			if (xfs_metadata_is_sick(error))
+				xfs_dirattr_mark_sick(args->dp, XFS_ATTR_FORK);
 			if (error)
 				return error;
 
-			error = xfs_attr_rmtval_copyout(mp, bp, args->dp->i_ino,
-							&offset, &valuelen,
-							&dst);
+			error = xfs_attr_rmtval_copyout(mp, bp, args->dp,
+					args->owner, &offset, &valuelen, &dst);
 			xfs_buf_relse(bp);
 			if (error)
 				return error;
@@ -448,7 +464,7 @@ xfs_attr_rmt_find_hole(
 	struct xfs_inode	*dp = args->dp;
 	struct xfs_mount	*mp = dp->i_mount;
 	int			error;
-	int			blkcnt;
+	unsigned int		blkcnt;
 	xfs_fileoff_t		lfileoff = 0;
 
 	/*
@@ -477,11 +493,11 @@ xfs_attr_rmtval_set_value(
 	struct xfs_bmbt_irec	map;
 	xfs_dablk_t		lblkno;
 	uint8_t			*src = args->value;
-	int			blkcnt;
-	int			valuelen;
+	unsigned int		blkcnt;
+	unsigned int		valuelen;
 	int			nmap;
 	int			error;
-	int			offset = 0;
+	unsigned int		offset = 0;
 
 	/*
 	 * Roll through the "value", copying the attribute value to the
@@ -517,8 +533,8 @@ xfs_attr_rmtval_set_value(
 			return error;
 		bp->b_ops = &xfs_attr3_rmt_buf_ops;
 
-		xfs_attr_rmtval_copyin(mp, bp, args->dp->i_ino, &offset,
-				       &valuelen, &src);
+		xfs_attr_rmtval_copyin(mp, bp, args->owner, &offset, &valuelen,
+				&src);
 
 		error = xfs_bwrite(bp);	/* GROT: NOTE: synchronous write */
 		xfs_buf_relse(bp);
@@ -545,11 +561,13 @@ xfs_attr_rmtval_stale(
 	struct xfs_buf		*bp;
 	int			error;
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 
 	if (XFS_IS_CORRUPT(mp, map->br_startblock == DELAYSTARTBLOCK) ||
-	    XFS_IS_CORRUPT(mp, map->br_startblock == HOLESTARTBLOCK))
+	    XFS_IS_CORRUPT(mp, map->br_startblock == HOLESTARTBLOCK)) {
+		xfs_bmap_mark_sick(ip, XFS_ATTR_FORK);
 		return -EFSCORRUPTED;
+	}
 
 	error = xfs_buf_incore(mp->m_ddev_targp,
 			XFS_FSB_TO_DADDR(mp, map->br_startblock),
@@ -619,7 +637,6 @@ xfs_attr_rmtval_set_blk(
 	if (error)
 		return error;
 
-	ASSERT(nmap == 1);
 	ASSERT((map->br_startblock != DELAYSTARTBLOCK) &&
 	       (map->br_startblock != HOLESTARTBLOCK));
 
@@ -639,7 +656,7 @@ xfs_attr_rmtval_invalidate(
 	struct xfs_da_args	*args)
 {
 	xfs_dablk_t		lblkno;
-	int			blkcnt;
+	unsigned int		blkcnt;
 	int			error;
 
 	/*
@@ -659,8 +676,10 @@ xfs_attr_rmtval_invalidate(
 				       blkcnt, &map, &nmap, XFS_BMAPI_ATTRFORK);
 		if (error)
 			return error;
-		if (XFS_IS_CORRUPT(args->dp->i_mount, nmap != 1))
+		if (XFS_IS_CORRUPT(args->dp->i_mount, nmap != 1)) {
+			xfs_bmap_mark_sick(args->dp, XFS_ATTR_FORK);
 			return -EFSCORRUPTED;
+		}
 		error = xfs_attr_rmtval_stale(args->dp, &map, XBF_TRYLOCK);
 		if (error)
 			return error;
diff --git a/fs/xfs/libxfs/xfs_attr_remote.h b/fs/xfs/libxfs/xfs_attr_remote.h
index d097ec6c4d..e3c6c7d774 100644
--- a/fs/xfs/libxfs/xfs_attr_remote.h
+++ b/fs/xfs/libxfs/xfs_attr_remote.h
@@ -6,7 +6,13 @@
 #ifndef __XFS_ATTR_REMOTE_H__
 #define	__XFS_ATTR_REMOTE_H__
 
-int xfs_attr3_rmt_blocks(struct xfs_mount *mp, int attrlen);
+unsigned int xfs_attr3_rmt_blocks(struct xfs_mount *mp, unsigned int attrlen);
+
+/* Number of rmt blocks needed to store the maximally sized attr value */
+static inline unsigned int xfs_attr3_max_rmt_blocks(struct xfs_mount *mp)
+{
+	return xfs_attr3_rmt_blocks(mp, XFS_XATTR_SIZE_MAX);
+}
 
 int xfs_attr_rmtval_get(struct xfs_da_args *args);
 int xfs_attr_rmtval_stale(struct xfs_inode *ip, struct xfs_bmbt_irec *map,
diff --git a/fs/xfs/libxfs/xfs_attr_sf.h b/fs/xfs/libxfs/xfs_attr_sf.h
index bc44222230..73bdc0e556 100644
--- a/fs/xfs/libxfs/xfs_attr_sf.h
+++ b/fs/xfs/libxfs/xfs_attr_sf.h
@@ -16,6 +16,7 @@ typedef struct xfs_attr_sf_sort {
 	uint8_t		flags;		/* flags bits (see xfs_attr_leaf.h) */
 	xfs_dahash_t	hash;		/* this entry's hash value */
 	unsigned char	*name;		/* name value, pointer into buffer */
+	void		*value;
 } xfs_attr_sf_sort_t;
 
 #define XFS_ATTR_SF_ENTSIZE_MAX			/* max space for name&value */ \
diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index f362345467..6af6f744fd 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -36,6 +36,9 @@
 #include "xfs_refcount.h"
 #include "xfs_icache.h"
 #include "xfs_iomap.h"
+#include "xfs_health.h"
+#include "xfs_bmap_item.h"
+#include "xfs_symlink_remote.h"
 
 struct kmem_cache		*xfs_bmap_intent_cache;
 
@@ -225,6 +228,28 @@ xfs_bmap_forkoff_reset(
 	}
 }
 
+static int
+xfs_bmap_read_buf(
+	struct xfs_mount	*mp,		/* file system mount point */
+	struct xfs_trans	*tp,		/* transaction pointer */
+	xfs_fsblock_t		fsbno,		/* file system block number */
+	struct xfs_buf		**bpp)		/* buffer for fsbno */
+{
+	struct xfs_buf		*bp;		/* return value */
+	int			error;
+
+	if (!xfs_verify_fsbno(mp, fsbno))
+		return -EFSCORRUPTED;
+	error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp,
+			XFS_FSB_TO_DADDR(mp, fsbno), mp->m_bsize, 0, &bp,
+			&xfs_bmbt_buf_ops);
+	if (!error) {
+		xfs_buf_set_ref(bp, XFS_BMAP_BTREE_REF);
+		*bpp = bp;
+	}
+	return error;
+}
+
 #ifdef DEBUG
 STATIC struct xfs_buf *
 xfs_bmap_get_bp(
@@ -364,9 +389,9 @@ xfs_bmap_check_leaf_extents(
 		bp = xfs_bmap_get_bp(cur, XFS_FSB_TO_DADDR(mp, bno));
 		if (!bp) {
 			bp_release = 1;
-			error = xfs_btree_read_bufl(mp, NULL, bno, &bp,
-						XFS_BMAP_BTREE_REF,
-						&xfs_bmbt_buf_ops);
+			error = xfs_bmap_read_buf(mp, NULL, bno, &bp);
+			if (xfs_metadata_is_sick(error))
+				xfs_btree_mark_sick(cur);
 			if (error)
 				goto error_norelse;
 		}
@@ -383,6 +408,7 @@ xfs_bmap_check_leaf_extents(
 		pp = XFS_BMBT_PTR_ADDR(mp, block, 1, mp->m_bmap_dmxr[1]);
 		bno = be64_to_cpu(*pp);
 		if (XFS_IS_CORRUPT(mp, !xfs_verify_fsbno(mp, bno))) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -450,9 +476,9 @@ xfs_bmap_check_leaf_extents(
 		bp = xfs_bmap_get_bp(cur, XFS_FSB_TO_DADDR(mp, bno));
 		if (!bp) {
 			bp_release = 1;
-			error = xfs_btree_read_bufl(mp, NULL, bno, &bp,
-						XFS_BMAP_BTREE_REF,
-						&xfs_bmbt_buf_ops);
+			error = xfs_bmap_read_buf(mp, NULL, bno, &bp);
+			if (xfs_metadata_is_sick(error))
+				xfs_btree_mark_sick(cur);
 			if (error)
 				goto error_norelse;
 		}
@@ -562,11 +588,14 @@ xfs_bmap_btree_to_extents(
 	pp = XFS_BMAP_BROOT_PTR_ADDR(mp, rblock, 1, ifp->if_broot_bytes);
 	cbno = be64_to_cpu(*pp);
 #ifdef DEBUG
-	if (XFS_IS_CORRUPT(cur->bc_mp, !xfs_btree_check_lptr(cur, cbno, 1)))
+	if (XFS_IS_CORRUPT(cur->bc_mp, !xfs_verify_fsbno(mp, cbno))) {
+		xfs_btree_mark_sick(cur);
 		return -EFSCORRUPTED;
+	}
 #endif
-	error = xfs_btree_read_bufl(mp, tp, cbno, &cbp, XFS_BMAP_BTREE_REF,
-				&xfs_bmbt_buf_ops);
+	error = xfs_bmap_read_buf(mp, tp, cbno, &cbp);
+	if (xfs_metadata_is_sick(error))
+		xfs_btree_mark_sick(cur);
 	if (error)
 		return error;
 	cblock = XFS_BUF_TO_BLOCK(cbp);
@@ -634,14 +663,13 @@ xfs_bmap_extents_to_btree(
 	 * Fill in the root.
 	 */
 	block = ifp->if_broot;
-	xfs_btree_init_block_int(mp, block, XFS_BUF_DADDR_NULL,
-				 XFS_BTNUM_BMAP, 1, 1, ip->i_ino,
-				 XFS_BTREE_LONG_PTRS);
+	xfs_bmbt_init_block(ip, block, NULL, 1, 1);
 	/*
 	 * Need a cursor.  Can't allocate until bb_level is filled in.
 	 */
 	cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
-	cur->bc_ino.flags = wasdel ? XFS_BTCUR_BMBT_WASDEL : 0;
+	if (wasdel)
+		cur->bc_flags |= XFS_BTREE_BMBT_WASDEL;
 	/*
 	 * Convert to a btree with two levels, one record in root.
 	 */
@@ -667,7 +695,7 @@ xfs_bmap_extents_to_btree(
 		goto out_root_realloc;
 	}
 
-	cur->bc_ino.allocated++;
+	cur->bc_bmap.allocated++;
 	ip->i_nblocks++;
 	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, 1L);
 	error = xfs_trans_get_buf(tp, mp->m_ddev_targp,
@@ -679,11 +707,8 @@ xfs_bmap_extents_to_btree(
 	/*
 	 * Fill in the child block.
 	 */
-	abp->b_ops = &xfs_bmbt_buf_ops;
 	ablock = XFS_BUF_TO_BLOCK(abp);
-	xfs_btree_init_block_int(mp, ablock, xfs_buf_daddr(abp),
-				XFS_BTNUM_BMAP, 0, 0, ip->i_ino,
-				XFS_BTREE_LONG_PTRS);
+	xfs_bmbt_init_block(ip, ablock, abp, 0, 0);
 
 	for_each_xfs_iext(ifp, &icur, &rec) {
 		if (isnullstartblock(rec.br_startblock))
@@ -754,7 +779,7 @@ xfs_bmap_local_to_extents_empty(
 }
 
 
-STATIC int				/* error */
+int					/* error */
 xfs_bmap_local_to_extents(
 	xfs_trans_t	*tp,		/* transaction pointer */
 	xfs_inode_t	*ip,		/* incore inode pointer */
@@ -764,7 +789,8 @@ xfs_bmap_local_to_extents(
 	void		(*init_fn)(struct xfs_trans *tp,
 				   struct xfs_buf *bp,
 				   struct xfs_inode *ip,
-				   struct xfs_ifork *ifp))
+				   struct xfs_ifork *ifp, void *priv),
+	void		*priv)
 {
 	int		error = 0;
 	int		flags;		/* logging flags returned */
@@ -825,7 +851,7 @@ xfs_bmap_local_to_extents(
 	 * log here. Note that init_fn must also set the buffer log item type
 	 * correctly.
 	 */
-	init_fn(tp, bp, ip, ifp);
+	init_fn(tp, bp, ip, ifp, priv);
 
 	/* account for the change in fork size */
 	xfs_idata_realloc(ip, -ifp->if_bytes, whichfork);
@@ -878,6 +904,7 @@ xfs_bmap_add_attrfork_btree(
 			goto error0;
 		/* must be at least one entry */
 		if (XFS_IS_CORRUPT(mp, stat != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -887,7 +914,7 @@ xfs_bmap_add_attrfork_btree(
 			xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
 			return -ENOSPC;
 		}
-		cur->bc_ino.allocated = 0;
+		cur->bc_bmap.allocated = 0;
 		xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
 	}
 	return 0;
@@ -915,7 +942,7 @@ xfs_bmap_add_attrfork_extents(
 	error = xfs_bmap_extents_to_btree(tp, ip, &cur, 0, flags,
 					  XFS_DATA_FORK);
 	if (cur) {
-		cur->bc_ino.allocated = 0;
+		cur->bc_bmap.allocated = 0;
 		xfs_btree_del_cursor(cur, error);
 	}
 	return error;
@@ -950,16 +977,18 @@ xfs_bmap_add_attrfork_local(
 		dargs.total = dargs.geo->fsbcount;
 		dargs.whichfork = XFS_DATA_FORK;
 		dargs.trans = tp;
+		dargs.owner = ip->i_ino;
 		return xfs_dir2_sf_to_block(&dargs);
 	}
 
 	if (S_ISLNK(VFS_I(ip)->i_mode))
 		return xfs_bmap_local_to_extents(tp, ip, 1, flags,
-						 XFS_DATA_FORK,
-						 xfs_symlink_local_to_remote);
+				XFS_DATA_FORK, xfs_symlink_local_to_remote,
+				NULL);
 
 	/* should only be called for types that support local format data */
 	ASSERT(0);
+	xfs_bmap_mark_sick(ip, XFS_ATTR_FORK);
 	return -EFSCORRUPTED;
 }
 
@@ -996,40 +1025,29 @@ xfs_bmap_set_attrforkoff(
 }
 
 /*
- * Convert inode from non-attributed to attributed.
- * Must not be in a transaction, ip must not be locked.
+ * Convert inode from non-attributed to attributed.  Caller must hold the
+ * ILOCK_EXCL and the file cannot have an attr fork.
  */
 int						/* error code */
 xfs_bmap_add_attrfork(
-	xfs_inode_t		*ip,		/* incore inode pointer */
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip,		/* incore inode pointer */
 	int			size,		/* space new attribute needs */
 	int			rsvd)		/* xact may use reserved blks */
 {
-	xfs_mount_t		*mp;		/* mount structure */
-	xfs_trans_t		*tp;		/* transaction pointer */
-	int			blks;		/* space reservation */
+	struct xfs_mount	*mp = tp->t_mountp;
 	int			version = 1;	/* superblock attr version */
 	int			logflags;	/* logging flags */
 	int			error;		/* error return value */
 
-	ASSERT(xfs_inode_has_attr_fork(ip) == 0);
-
-	mp = ip->i_mount;
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 	ASSERT(!XFS_NOT_DQATTACHED(mp, ip));
-
-	blks = XFS_ADDAFORK_SPACE_RES(mp);
-
-	error = xfs_trans_alloc_inode(ip, &M_RES(mp)->tr_addafork, blks, 0,
-			rsvd, &tp);
-	if (error)
-		return error;
-	if (xfs_inode_has_attr_fork(ip))
-		goto trans_cancel;
+	ASSERT(!xfs_inode_has_attr_fork(ip));
 
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 	error = xfs_bmap_set_attrforkoff(ip, size, &version);
 	if (error)
-		goto trans_cancel;
+		return error;
 
 	xfs_ifork_init_attr(ip, XFS_DINODE_FMT_EXTENTS, 0);
 	logflags = 0;
@@ -1050,7 +1068,7 @@ xfs_bmap_add_attrfork(
 	if (logflags)
 		xfs_trans_log_inode(tp, ip, logflags);
 	if (error)
-		goto trans_cancel;
+		return error;
 	if (!xfs_has_attr(mp) ||
 	   (!xfs_has_attr2(mp) && version == 2)) {
 		bool log_sb = false;
@@ -1069,14 +1087,7 @@ xfs_bmap_add_attrfork(
 			xfs_log_sb(tp);
 	}
 
-	error = xfs_trans_commit(tp);
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
-	return error;
-
-trans_cancel:
-	xfs_trans_cancel(tp);
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
-	return error;
+	return 0;
 }
 
 /*
@@ -1143,6 +1154,7 @@ xfs_iread_bmbt_block(
 				(unsigned long long)ip->i_ino);
 		xfs_inode_verifier_error(ip, -EFSCORRUPTED, __func__, block,
 				sizeof(*block), __this_address);
+		xfs_bmap_mark_sick(ip, whichfork);
 		return -EFSCORRUPTED;
 	}
 
@@ -1158,6 +1170,7 @@ xfs_iread_bmbt_block(
 			xfs_inode_verifier_error(ip, -EFSCORRUPTED,
 					"xfs_iread_extents(2)", frp,
 					sizeof(*frp), fa);
+			xfs_bmap_mark_sick(ip, whichfork);
 			return xfs_bmap_complain_bad_rec(ip, whichfork, fa,
 					&new);
 		}
@@ -1189,7 +1202,7 @@ xfs_iread_extents(
 	if (!xfs_need_iread_extents(ifp))
 		return 0;
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 
 	ir.loaded = 0;
 	xfs_iext_first(ifp, &ir.icur);
@@ -1201,6 +1214,7 @@ xfs_iread_extents(
 		goto out;
 
 	if (XFS_IS_CORRUPT(mp, ir.loaded != ifp->if_nextents)) {
+		xfs_bmap_mark_sick(ip, whichfork);
 		error = -EFSCORRUPTED;
 		goto out;
 	}
@@ -1213,6 +1227,8 @@ xfs_iread_extents(
 	smp_store_release(&ifp->if_needextents, 0);
 	return 0;
 out:
+	if (xfs_metadata_is_sick(error))
+		xfs_bmap_mark_sick(ip, whichfork);
 	xfs_iext_destroy(ifp);
 	return error;
 }
@@ -1292,6 +1308,7 @@ xfs_bmap_last_before(
 		break;
 	default:
 		ASSERT(0);
+		xfs_bmap_mark_sick(ip, whichfork);
 		return -EFSCORRUPTED;
 	}
 
@@ -1388,8 +1405,10 @@ xfs_bmap_last_offset(
 	if (ifp->if_format == XFS_DINODE_FMT_LOCAL)
 		return 0;
 
-	if (XFS_IS_CORRUPT(ip->i_mount, !xfs_ifork_has_extents(ifp)))
+	if (XFS_IS_CORRUPT(ip->i_mount, !xfs_ifork_has_extents(ifp))) {
+		xfs_bmap_mark_sick(ip, whichfork);
 		return -EFSCORRUPTED;
+	}
 
 	error = xfs_bmap_last_extent(NULL, ip, whichfork, &rec, &is_empty);
 	if (error || is_empty)
@@ -1429,8 +1448,7 @@ xfs_bmap_add_extent_delay_real(
 
 	ASSERT(whichfork != XFS_ATTR_FORK);
 	ASSERT(!isnullstartblock(new->br_startblock));
-	ASSERT(!bma->cur ||
-	       (bma->cur->bc_ino.flags & XFS_BTCUR_BMBT_WASDEL));
+	ASSERT(!bma->cur || (bma->cur->bc_flags & XFS_BTREE_BMBT_WASDEL));
 
 	XFS_STATS_INC(mp, xs_add_exlist);
 
@@ -1528,6 +1546,7 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(bma->cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -1535,6 +1554,7 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(bma->cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -1542,6 +1562,7 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(bma->cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -1549,6 +1570,7 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 		}
+		ASSERT(da_new <= da_old);
 		break;
 
 	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_LEFT_CONTIG:
@@ -1571,6 +1593,7 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(bma->cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -1578,6 +1601,7 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 		}
+		ASSERT(da_new <= da_old);
 		break;
 
 	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
@@ -1604,6 +1628,7 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(bma->cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -1611,6 +1636,7 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 		}
+		ASSERT(da_new <= da_old);
 		break;
 
 	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING:
@@ -1632,6 +1658,7 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 0)) {
+				xfs_btree_mark_sick(bma->cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -1639,10 +1666,12 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(bma->cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
 		}
+		ASSERT(da_new <= da_old);
 		break;
 
 	case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG:
@@ -1673,6 +1702,7 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(bma->cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -1680,6 +1710,7 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 		}
+		ASSERT(da_new <= da_old);
 		break;
 
 	case BMAP_LEFT_FILLING:
@@ -1698,6 +1729,7 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 0)) {
+				xfs_btree_mark_sick(bma->cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -1705,6 +1737,7 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(bma->cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -1721,7 +1754,7 @@ xfs_bmap_add_extent_delay_real(
 		temp = PREV.br_blockcount - new->br_blockcount;
 		da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(bma->ip, temp),
 			startblockval(PREV.br_startblock) -
-			(bma->cur ? bma->cur->bc_ino.allocated : 0));
+			(bma->cur ? bma->cur->bc_bmap.allocated : 0));
 
 		PREV.br_startoff = new_endoff;
 		PREV.br_blockcount = temp;
@@ -1749,6 +1782,7 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(bma->cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -1767,6 +1801,7 @@ xfs_bmap_add_extent_delay_real(
 		xfs_iext_update_extent(bma->ip, state, &bma->icur, &PREV);
 		xfs_iext_next(ifp, &bma->icur);
 		xfs_iext_update_extent(bma->ip, state, &bma->icur, &RIGHT);
+		ASSERT(da_new <= da_old);
 		break;
 
 	case BMAP_RIGHT_FILLING:
@@ -1785,6 +1820,7 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 0)) {
+				xfs_btree_mark_sick(bma->cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -1792,6 +1828,7 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(bma->cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -1808,12 +1845,13 @@ xfs_bmap_add_extent_delay_real(
 		temp = PREV.br_blockcount - new->br_blockcount;
 		da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(bma->ip, temp),
 			startblockval(PREV.br_startblock) -
-			(bma->cur ? bma->cur->bc_ino.allocated : 0));
+			(bma->cur ? bma->cur->bc_bmap.allocated : 0));
 
 		PREV.br_startblock = nullstartblock(da_new);
 		PREV.br_blockcount = temp;
 		xfs_iext_insert(bma->ip, &bma->icur, &PREV, state);
 		xfs_iext_next(ifp, &bma->icur);
+		ASSERT(da_new <= da_old);
 		break;
 
 	case 0:
@@ -1871,6 +1909,7 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 0)) {
+				xfs_btree_mark_sick(bma->cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -1878,6 +1917,7 @@ xfs_bmap_add_extent_delay_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(bma->cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -1926,19 +1966,18 @@ xfs_bmap_add_extent_delay_real(
 	}
 
 	if (da_new != da_old)
-		xfs_mod_delalloc(mp, (int64_t)da_new - da_old);
+		xfs_mod_delalloc(bma->ip, 0, (int64_t)da_new - da_old);
 
 	if (bma->cur) {
-		da_new += bma->cur->bc_ino.allocated;
-		bma->cur->bc_ino.allocated = 0;
+		da_new += bma->cur->bc_bmap.allocated;
+		bma->cur->bc_bmap.allocated = 0;
 	}
 
 	/* adjust for changes in reserved delayed indirect blocks */
-	if (da_new != da_old) {
-		ASSERT(state == 0 || da_new < da_old);
-		error = xfs_mod_fdblocks(mp, (int64_t)(da_old - da_new),
-				false);
-	}
+	if (da_new < da_old)
+		xfs_add_fdblocks(mp, da_old - da_new);
+	else if (da_new > da_old)
+		error = xfs_dec_fdblocks(mp, da_new - da_old, true);
 
 	xfs_bmap_check_leaf_extents(bma->cur, bma->ip, whichfork);
 done:
@@ -2074,30 +2113,35 @@ xfs_bmap_add_extent_unwritten_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
 			if ((error = xfs_btree_delete(cur, &i)))
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
 			if ((error = xfs_btree_decrement(cur, 0, &i)))
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
 			if ((error = xfs_btree_delete(cur, &i)))
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
 			if ((error = xfs_btree_decrement(cur, 0, &i)))
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2126,18 +2170,21 @@ xfs_bmap_add_extent_unwritten_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
 			if ((error = xfs_btree_delete(cur, &i)))
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
 			if ((error = xfs_btree_decrement(cur, 0, &i)))
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2169,18 +2216,21 @@ xfs_bmap_add_extent_unwritten_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
 			if ((error = xfs_btree_delete(cur, &i)))
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
 			if ((error = xfs_btree_decrement(cur, 0, &i)))
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2207,6 +2257,7 @@ xfs_bmap_add_extent_unwritten_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2240,6 +2291,7 @@ xfs_bmap_add_extent_unwritten_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2277,6 +2329,7 @@ xfs_bmap_add_extent_unwritten_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2287,6 +2340,7 @@ xfs_bmap_add_extent_unwritten_real(
 			if ((error = xfs_btree_insert(cur, &i)))
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2317,6 +2371,7 @@ xfs_bmap_add_extent_unwritten_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2353,6 +2408,7 @@ xfs_bmap_add_extent_unwritten_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2363,12 +2419,14 @@ xfs_bmap_add_extent_unwritten_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 0)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
 			if ((error = xfs_btree_insert(cur, &i)))
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2405,6 +2463,7 @@ xfs_bmap_add_extent_unwritten_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2417,6 +2476,7 @@ xfs_bmap_add_extent_unwritten_real(
 			if ((error = xfs_btree_insert(cur, &i)))
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2429,6 +2489,7 @@ xfs_bmap_add_extent_unwritten_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 0)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2436,6 +2497,7 @@ xfs_bmap_add_extent_unwritten_real(
 			if ((error = xfs_btree_insert(cur, &i)))
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2472,7 +2534,7 @@ xfs_bmap_add_extent_unwritten_real(
 
 	/* clear out the allocated field, done with it now in any case. */
 	if (cur) {
-		cur->bc_ino.allocated = 0;
+		cur->bc_bmap.allocated = 0;
 		*curp = cur;
 	}
 
@@ -2616,12 +2678,12 @@ xfs_bmap_add_extent_hole_delay(
 	}
 	if (oldlen != newlen) {
 		ASSERT(oldlen > newlen);
-		xfs_mod_fdblocks(ip->i_mount, (int64_t)(oldlen - newlen),
-				 false);
+		xfs_add_fdblocks(ip->i_mount, oldlen - newlen);
+
 		/*
 		 * Nothing to do for disk quota accounting here.
 		 */
-		xfs_mod_delalloc(ip->i_mount, (int64_t)newlen - oldlen);
+		xfs_mod_delalloc(ip, 0, (int64_t)newlen - oldlen);
 	}
 }
 
@@ -2651,7 +2713,7 @@ xfs_bmap_add_extent_hole_real(
 	struct xfs_bmbt_irec	old;
 
 	ASSERT(!isnullstartblock(new->br_startblock));
-	ASSERT(!cur || !(cur->bc_ino.flags & XFS_BTCUR_BMBT_WASDEL));
+	ASSERT(!cur || !(cur->bc_flags & XFS_BTREE_BMBT_WASDEL));
 
 	XFS_STATS_INC(mp, xs_add_exlist);
 
@@ -2721,6 +2783,7 @@ xfs_bmap_add_extent_hole_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2728,6 +2791,7 @@ xfs_bmap_add_extent_hole_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2735,6 +2799,7 @@ xfs_bmap_add_extent_hole_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2764,6 +2829,7 @@ xfs_bmap_add_extent_hole_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2794,6 +2860,7 @@ xfs_bmap_add_extent_hole_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2820,6 +2887,7 @@ xfs_bmap_add_extent_hole_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 0)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2827,6 +2895,7 @@ xfs_bmap_add_extent_hole_real(
 			if (error)
 				goto done;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto done;
 			}
@@ -2853,7 +2922,7 @@ xfs_bmap_add_extent_hole_real(
 
 	/* clear out the allocated field, done with it now in any case. */
 	if (cur)
-		cur->bc_ino.allocated = 0;
+		cur->bc_bmap.allocated = 0;
 
 	xfs_bmap_check_leaf_extents(cur, ip, whichfork);
 done:
@@ -3291,7 +3360,7 @@ xfs_bmap_alloc_account(
 		 * yet.
 		 */
 		if (ap->wasdel) {
-			xfs_mod_delalloc(ap->ip->i_mount, -(int64_t)ap->length);
+			xfs_mod_delalloc(ap->ip, -(int64_t)ap->length, 0);
 			return;
 		}
 
@@ -3315,7 +3384,7 @@ xfs_bmap_alloc_account(
 	xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE);
 	if (ap->wasdel) {
 		ap->ip->i_delayed_blks -= ap->length;
-		xfs_mod_delalloc(ap->ip->i_mount, -(int64_t)ap->length);
+		xfs_mod_delalloc(ap->ip, -(int64_t)ap->length, 0);
 		fld = isrt ? XFS_TRANS_DQ_DELRTBCOUNT : XFS_TRANS_DQ_DELBCOUNT;
 	} else {
 		fld = isrt ? XFS_TRANS_DQ_RTBCOUNT : XFS_TRANS_DQ_BCOUNT;
@@ -3898,14 +3967,18 @@ xfs_bmapi_read(
 
 	ASSERT(*nmap >= 1);
 	ASSERT(!(flags & ~(XFS_BMAPI_ATTRFORK | XFS_BMAPI_ENTIRE)));
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_SHARED|XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL);
 
-	if (WARN_ON_ONCE(!ifp))
+	if (WARN_ON_ONCE(!ifp)) {
+		xfs_bmap_mark_sick(ip, whichfork);
 		return -EFSCORRUPTED;
+	}
 
 	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)) ||
-	    XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT))
+	    XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) {
+		xfs_bmap_mark_sick(ip, whichfork);
 		return -EFSCORRUPTED;
+	}
 
 	if (xfs_is_shutdown(mp))
 		return -EIO;
@@ -3983,21 +4056,34 @@ xfs_bmapi_reserve_delalloc(
 	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, whichfork);
 	xfs_extlen_t		alen;
 	xfs_extlen_t		indlen;
+	uint64_t		fdblocks;
 	int			error;
-	xfs_fileoff_t		aoff = off;
+	xfs_fileoff_t		aoff;
+	bool			use_cowextszhint =
+					whichfork == XFS_COW_FORK && !prealloc;
 
+retry:
 	/*
 	 * Cap the alloc length. Keep track of prealloc so we know whether to
 	 * tag the inode before we return.
 	 */
+	aoff = off;
 	alen = XFS_FILBLKS_MIN(len + prealloc, XFS_MAX_BMBT_EXTLEN);
 	if (!eof)
 		alen = XFS_FILBLKS_MIN(alen, got->br_startoff - aoff);
 	if (prealloc && alen >= len)
 		prealloc = alen - len;
 
-	/* Figure out the extent size, adjust alen */
-	if (whichfork == XFS_COW_FORK) {
+	/*
+	 * If we're targetting the COW fork but aren't creating a speculative
+	 * posteof preallocation, try to expand the reservation to align with
+	 * the COW extent size hint if there's sufficient free space.
+	 *
+	 * Unlike the data fork, the CoW cancellation functions will free all
+	 * the reservations at inactivation, so we don't require that every
+	 * delalloc reservation have a dirty pagecache.
+	 */
+	if (use_cowextszhint) {
 		struct xfs_bmbt_irec	prev;
 		xfs_extlen_t		extsz = xfs_get_cowextsz_hint(ip);
 
@@ -4016,7 +4102,7 @@ xfs_bmapi_reserve_delalloc(
 	 */
 	error = xfs_quota_reserve_blkres(ip, alen);
 	if (error)
-		return error;
+		goto out;
 
 	/*
 	 * Split changing sb for alen and indlen since they could be coming
@@ -4025,17 +4111,21 @@ xfs_bmapi_reserve_delalloc(
 	indlen = (xfs_extlen_t)xfs_bmap_worst_indlen(ip, alen);
 	ASSERT(indlen > 0);
 
-	error = xfs_mod_fdblocks(mp, -((int64_t)alen), false);
-	if (error)
-		goto out_unreserve_quota;
+	fdblocks = indlen;
+	if (XFS_IS_REALTIME_INODE(ip)) {
+		error = xfs_dec_frextents(mp, xfs_rtb_to_rtx(mp, alen));
+		if (error)
+			goto out_unreserve_quota;
+	} else {
+		fdblocks += alen;
+	}
 
-	error = xfs_mod_fdblocks(mp, -((int64_t)indlen), false);
+	error = xfs_dec_fdblocks(mp, fdblocks, false);
 	if (error)
-		goto out_unreserve_blocks;
-
+		goto out_unreserve_frextents;
 
 	ip->i_delayed_blks += alen;
-	xfs_mod_delalloc(ip->i_mount, alen + indlen);
+	xfs_mod_delalloc(ip, alen, indlen);
 
 	got->br_startoff = aoff;
 	got->br_startblock = nullstartblock(indlen);
@@ -4056,11 +4146,23 @@ xfs_bmapi_reserve_delalloc(
 
 	return 0;
 
-out_unreserve_blocks:
-	xfs_mod_fdblocks(mp, alen, false);
+out_unreserve_frextents:
+	if (XFS_IS_REALTIME_INODE(ip))
+		xfs_add_frextents(mp, xfs_rtb_to_rtx(mp, alen));
 out_unreserve_quota:
 	if (XFS_IS_QUOTA_ON(mp))
 		xfs_quota_unreserve_blkres(ip, alen);
+out:
+	if (error == -ENOSPC || error == -EDQUOT) {
+		trace_xfs_delalloc_enospc(ip, off, len);
+
+		if (prealloc || use_cowextszhint) {
+			/* retry without any preallocation */
+			use_cowextszhint = false;
+			prealloc = 0;
+			goto retry;
+		}
+	}
 	return error;
 }
 
@@ -4108,26 +4210,10 @@ xfs_bmapi_allocate(
 	struct xfs_mount	*mp = bma->ip->i_mount;
 	int			whichfork = xfs_bmapi_whichfork(bma->flags);
 	struct xfs_ifork	*ifp = xfs_ifork_ptr(bma->ip, whichfork);
-	int			tmp_logflags = 0;
 	int			error;
 
 	ASSERT(bma->length > 0);
-
-	/*
-	 * For the wasdelay case, we could also just allocate the stuff asked
-	 * for in this bmap call but that wouldn't be as good.
-	 */
-	if (bma->wasdel) {
-		bma->length = (xfs_extlen_t)bma->got.br_blockcount;
-		bma->offset = bma->got.br_startoff;
-		if (!xfs_iext_peek_prev_extent(ifp, &bma->icur, &bma->prev))
-			bma->prev.br_startoff = NULLFILEOFF;
-	} else {
-		bma->length = XFS_FILBLKS_MIN(bma->length, XFS_MAX_BMBT_EXTLEN);
-		if (!bma->eof)
-			bma->length = XFS_FILBLKS_MIN(bma->length,
-					bma->got.br_startoff - bma->offset);
-	}
+	ASSERT(bma->length <= XFS_MAX_BMBT_EXTLEN);
 
 	if (bma->flags & XFS_BMAPI_CONTIG)
 		bma->minlen = bma->length;
@@ -4143,8 +4229,15 @@ xfs_bmapi_allocate(
 	} else {
 		error = xfs_bmap_alloc_userdata(bma);
 	}
-	if (error || bma->blkno == NULLFSBLOCK)
+	if (error)
 		return error;
+	if (bma->blkno == NULLFSBLOCK)
+		return -ENOSPC;
+
+	if (WARN_ON_ONCE(!xfs_valid_startblock(bma->ip, bma->blkno))) {
+		xfs_bmap_mark_sick(bma->ip, whichfork);
+		return -EFSCORRUPTED;
+	}
 
 	if (bma->flags & XFS_BMAPI_ZERO) {
 		error = xfs_zero_extent(bma->ip, bma->blkno, bma->length);
@@ -4160,9 +4253,8 @@ xfs_bmapi_allocate(
 	 */
 	bma->nallocs++;
 
-	if (bma->cur)
-		bma->cur->bc_ino.flags =
-			bma->wasdel ? XFS_BTCUR_BMBT_WASDEL : 0;
+	if (bma->cur && bma->wasdel)
+		bma->cur->bc_flags |= XFS_BTREE_BMBT_WASDEL;
 
 	bma->got.br_startoff = bma->offset;
 	bma->got.br_startblock = bma->blkno;
@@ -4178,8 +4270,6 @@ xfs_bmapi_allocate(
 		error = xfs_bmap_add_extent_hole_real(bma->tp, bma->ip,
 				whichfork, &bma->icur, &bma->cur, &bma->got,
 				&bma->logflags, bma->flags);
-
-	bma->logflags |= tmp_logflags;
 	if (error)
 		return error;
 
@@ -4324,6 +4414,15 @@ xfs_bmapi_finish(
  * extent state if necessary.  Details behaviour is controlled by the flags
  * parameter.  Only allocates blocks from a single allocation group, to avoid
  * locking problems.
+ *
+ * Returns 0 on success and places the extent mappings in mval.  nmaps is used
+ * as an input/output parameter where the caller specifies the maximum number
+ * of mappings that may be returned and xfs_bmapi_write passes back the number
+ * of mappings (including existing mappings) it found.
+ *
+ * Returns a negative error code on failure, including -ENOSPC when it could not
+ * allocate any blocks and -ENOSR when it did allocate blocks to convert a
+ * delalloc range, but those blocks were before the passed in range.
  */
 int
 xfs_bmapi_write(
@@ -4369,7 +4468,7 @@ xfs_bmapi_write(
 	ASSERT(tp != NULL);
 	ASSERT(len > 0);
 	ASSERT(ifp->if_format != XFS_DINODE_FMT_LOCAL);
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 	ASSERT(!(flags & XFS_BMAPI_REMAP));
 
 	/* zeroing is for currently only for data extents, not metadata */
@@ -4386,6 +4485,7 @@ xfs_bmapi_write(
 
 	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)) ||
 	    XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) {
+		xfs_bmap_mark_sick(ip, whichfork);
 		return -EFSCORRUPTED;
 	}
 
@@ -4441,20 +4541,33 @@ xfs_bmapi_write(
 			 * allocation length request (which can be 64 bits in
 			 * length) and the bma length request, which is
 			 * xfs_extlen_t and therefore 32 bits. Hence we have to
-			 * check for 32-bit overflows and handle them here.
+			 * be careful and do the min() using the larger type to
+			 * avoid overflows.
 			 */
-			if (len > (xfs_filblks_t)XFS_MAX_BMBT_EXTLEN)
-				bma.length = XFS_MAX_BMBT_EXTLEN;
-			else
-				bma.length = len;
+			bma.length = XFS_FILBLKS_MIN(len, XFS_MAX_BMBT_EXTLEN);
+
+			if (wasdelay) {
+				bma.length = XFS_FILBLKS_MIN(bma.length,
+					bma.got.br_blockcount -
+					(bno - bma.got.br_startoff));
+			} else {
+				if (!eof)
+					bma.length = XFS_FILBLKS_MIN(bma.length,
+						bma.got.br_startoff - bno);
+			}
 
-			ASSERT(len > 0);
 			ASSERT(bma.length > 0);
 			error = xfs_bmapi_allocate(&bma);
-			if (error)
+			if (error) {
+				/*
+				 * If we already allocated space in a previous
+				 * iteration return what we go so far when
+				 * running out of space.
+				 */
+				if (error == -ENOSPC && bma.nallocs)
+					break;
 				goto error0;
-			if (bma.blkno == NULLFSBLOCK)
-				break;
+			}
 
 			/*
 			 * If this is a CoW allocation, record the data in
@@ -4492,7 +4605,6 @@ xfs_bmapi_write(
 		if (!xfs_iext_next_extent(ifp, &bma.icur, &bma.got))
 			eof = true;
 	}
-	*nmap = n;
 
 	error = xfs_bmap_btree_to_extents(tp, ip, bma.cur, &bma.logflags,
 			whichfork);
@@ -4503,7 +4615,22 @@ xfs_bmapi_write(
 	       ifp->if_nextents > XFS_IFORK_MAXEXT(ip, whichfork));
 	xfs_bmapi_finish(&bma, whichfork, 0);
 	xfs_bmap_validate_ret(orig_bno, orig_len, orig_flags, orig_mval,
-		orig_nmap, *nmap);
+		orig_nmap, n);
+
+	/*
+	 * When converting delayed allocations, xfs_bmapi_allocate ignores
+	 * the passed in bno and always converts from the start of the found
+	 * delalloc extent.
+	 *
+	 * To avoid a successful return with *nmap set to 0, return the magic
+	 * -ENOSR error code for this particular case so that the caller can
+	 * handle it.
+	 */
+	if (!n) {
+		ASSERT(bma.nallocs >= *nmap);
+		return -ENOSR;
+	}
+	*nmap = n;
 	return 0;
 error0:
 	xfs_bmapi_finish(&bma, whichfork, error);
@@ -4516,8 +4643,8 @@ error0:
  * invocations to allocate the target offset if a large enough physical extent
  * is not available.
  */
-int
-xfs_bmapi_convert_delalloc(
+static int
+xfs_bmapi_convert_one_delalloc(
 	struct xfs_inode	*ip,
 	int			whichfork,
 	xfs_off_t		offset,
@@ -4547,11 +4674,8 @@ xfs_bmapi_convert_delalloc(
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
 	xfs_trans_ijoin(tp, ip, 0);
 
-	error = xfs_iext_count_may_overflow(ip, whichfork,
+	error = xfs_iext_count_extend(tp, ip, whichfork,
 			XFS_IEXT_ADD_NOSPLIT_CNT);
-	if (error == -EFBIG)
-		error = xfs_iext_count_upgrade(tp, ip,
-				XFS_IEXT_ADD_NOSPLIT_CNT);
 	if (error)
 		goto out_trans_cancel;
 
@@ -4574,19 +4698,25 @@ xfs_bmapi_convert_delalloc(
 	if (!isnullstartblock(bma.got.br_startblock)) {
 		xfs_bmbt_to_iomap(ip, iomap, &bma.got, 0, flags,
 				xfs_iomap_inode_sequence(ip, flags));
-		*seq = READ_ONCE(ifp->if_seq);
+		if (seq)
+			*seq = READ_ONCE(ifp->if_seq);
 		goto out_trans_cancel;
 	}
 
 	bma.tp = tp;
 	bma.ip = ip;
 	bma.wasdel = true;
-	bma.offset = bma.got.br_startoff;
-	bma.length = max_t(xfs_filblks_t, bma.got.br_blockcount,
-			XFS_MAX_BMBT_EXTLEN);
 	bma.minleft = xfs_bmapi_minleft(tp, ip, whichfork);
 
 	/*
+	 * Always allocate convert from the start of the delalloc extent even if
+	 * that is outside the passed in range to create large contiguous
+	 * extents on disk.
+	 */
+	bma.offset = bma.got.br_startoff;
+	bma.length = bma.got.br_blockcount;
+
+	/*
 	 * When we're converting the delalloc reservations backing dirty pages
 	 * in the page cache, we must be careful about how we create the new
 	 * extents:
@@ -4610,20 +4740,14 @@ xfs_bmapi_convert_delalloc(
 	if (error)
 		goto out_finish;
 
-	error = -ENOSPC;
-	if (WARN_ON_ONCE(bma.blkno == NULLFSBLOCK))
-		goto out_finish;
-	error = -EFSCORRUPTED;
-	if (WARN_ON_ONCE(!xfs_valid_startblock(ip, bma.got.br_startblock)))
-		goto out_finish;
-
 	XFS_STATS_ADD(mp, xs_xstrat_bytes, XFS_FSB_TO_B(mp, bma.length));
 	XFS_STATS_INC(mp, xs_xstrat_quick);
 
 	ASSERT(!isnullstartblock(bma.got.br_startblock));
 	xfs_bmbt_to_iomap(ip, iomap, &bma.got, 0, flags,
 				xfs_iomap_inode_sequence(ip, flags));
-	*seq = READ_ONCE(ifp->if_seq);
+	if (seq)
+		*seq = READ_ONCE(ifp->if_seq);
 
 	if (whichfork == XFS_COW_FORK)
 		xfs_refcount_alloc_cow_extent(tp, bma.blkno, bma.length);
@@ -4646,6 +4770,36 @@ out_trans_cancel:
 	return error;
 }
 
+/*
+ * Pass in a dellalloc extent and convert it to real extents, return the real
+ * extent that maps offset_fsb in iomap.
+ */
+int
+xfs_bmapi_convert_delalloc(
+	struct xfs_inode	*ip,
+	int			whichfork,
+	loff_t			offset,
+	struct iomap		*iomap,
+	unsigned int		*seq)
+{
+	int			error;
+
+	/*
+	 * Attempt to allocate whatever delalloc extent currently backs offset
+	 * and put the result into iomap.  Allocate in a loop because it may
+	 * take several attempts to allocate real blocks for a contiguous
+	 * delalloc extent if free space is sufficiently fragmented.
+	 */
+	do {
+		error = xfs_bmapi_convert_one_delalloc(ip, whichfork, offset,
+					iomap, seq);
+		if (error)
+			return error;
+	} while (iomap->offset + iomap->length <= offset);
+
+	return 0;
+}
+
 int
 xfs_bmapi_remap(
 	struct xfs_trans	*tp,
@@ -4666,7 +4820,7 @@ xfs_bmapi_remap(
 	ifp = xfs_ifork_ptr(ip, whichfork);
 	ASSERT(len > 0);
 	ASSERT(len <= (xfs_filblks_t)XFS_MAX_BMBT_EXTLEN);
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 	ASSERT(!(flags & ~(XFS_BMAPI_ATTRFORK | XFS_BMAPI_PREALLOC |
 			   XFS_BMAPI_NORMAP)));
 	ASSERT((flags & (XFS_BMAPI_ATTRFORK | XFS_BMAPI_PREALLOC)) !=
@@ -4674,6 +4828,7 @@ xfs_bmapi_remap(
 
 	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)) ||
 	    XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) {
+		xfs_bmap_mark_sick(ip, whichfork);
 		return -EFSCORRUPTED;
 	}
 
@@ -4693,10 +4848,8 @@ xfs_bmapi_remap(
 	ip->i_nblocks += len;
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 
-	if (ifp->if_format == XFS_DINODE_FMT_BTREE) {
+	if (ifp->if_format == XFS_DINODE_FMT_BTREE)
 		cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
-		cur->bc_ino.flags = 0;
-	}
 
 	got.br_startoff = bno;
 	got.br_startblock = startblock;
@@ -4738,32 +4891,18 @@ error0:
  * ores == 1). The number of stolen blocks is returned. The availability and
  * subsequent accounting of stolen blocks is the responsibility of the caller.
  */
-static xfs_filblks_t
+static void
 xfs_bmap_split_indlen(
 	xfs_filblks_t			ores,		/* original res. */
 	xfs_filblks_t			*indlen1,	/* ext1 worst indlen */
-	xfs_filblks_t			*indlen2,	/* ext2 worst indlen */
-	xfs_filblks_t			avail)		/* stealable blocks */
+	xfs_filblks_t			*indlen2)	/* ext2 worst indlen */
 {
 	xfs_filblks_t			len1 = *indlen1;
 	xfs_filblks_t			len2 = *indlen2;
 	xfs_filblks_t			nres = len1 + len2; /* new total res. */
-	xfs_filblks_t			stolen = 0;
 	xfs_filblks_t			resfactor;
 
 	/*
-	 * Steal as many blocks as we can to try and satisfy the worst case
-	 * indlen for both new extents.
-	 */
-	if (ores < nres && avail)
-		stolen = XFS_FILBLKS_MIN(nres - ores, avail);
-	ores += stolen;
-
-	 /* nothing else to do if we've satisfied the new reservation */
-	if (ores >= nres)
-		return stolen;
-
-	/*
 	 * We can't meet the total required reservation for the two extents.
 	 * Calculate the percent of the overall shortage between both extents
 	 * and apply this percentage to each of the requested indlen values.
@@ -4807,11 +4946,9 @@ xfs_bmap_split_indlen(
 
 	*indlen1 = len1;
 	*indlen2 = len2;
-
-	return stolen;
 }
 
-int
+void
 xfs_bmap_del_extent_delay(
 	struct xfs_inode	*ip,
 	int			whichfork,
@@ -4824,14 +4961,14 @@ xfs_bmap_del_extent_delay(
 	struct xfs_bmbt_irec	new;
 	int64_t			da_old, da_new, da_diff = 0;
 	xfs_fileoff_t		del_endoff, got_endoff;
-	xfs_filblks_t		got_indlen, new_indlen, stolen;
+	xfs_filblks_t		got_indlen, new_indlen, stolen = 0;
 	uint32_t		state = xfs_bmap_fork_to_state(whichfork);
-	int			error = 0;
+	uint64_t		fdblocks;
 	bool			isrt;
 
 	XFS_STATS_INC(mp, xs_del_exlist);
 
-	isrt = (whichfork == XFS_DATA_FORK) && XFS_IS_REALTIME_INODE(ip);
+	isrt = xfs_ifork_is_realtime(ip, whichfork);
 	del_endoff = del->br_startoff + del->br_blockcount;
 	got_endoff = got->br_startoff + got->br_blockcount;
 	da_old = startblockval(got->br_startblock);
@@ -4841,18 +4978,12 @@ xfs_bmap_del_extent_delay(
 	ASSERT(got->br_startoff <= del->br_startoff);
 	ASSERT(got_endoff >= del_endoff);
 
-	if (isrt)
-		xfs_mod_frextents(mp, xfs_rtb_to_rtx(mp, del->br_blockcount));
-
 	/*
 	 * Update the inode delalloc counter now and wait to update the
 	 * sb counters as we might have to borrow some blocks for the
 	 * indirect block accounting.
 	 */
-	ASSERT(!isrt);
-	error = xfs_quota_unreserve_blkres(ip, del->br_blockcount);
-	if (error)
-		return error;
+	xfs_quota_unreserve_blkres(ip, del->br_blockcount);
 	ip->i_delayed_blks -= del->br_blockcount;
 
 	if (got->br_startoff == del->br_startoff)
@@ -4906,8 +5037,24 @@ xfs_bmap_del_extent_delay(
 		new_indlen = xfs_bmap_worst_indlen(ip, new.br_blockcount);
 
 		WARN_ON_ONCE(!got_indlen || !new_indlen);
-		stolen = xfs_bmap_split_indlen(da_old, &got_indlen, &new_indlen,
-						       del->br_blockcount);
+		/*
+		 * Steal as many blocks as we can to try and satisfy the worst
+		 * case indlen for both new extents.
+		 *
+		 * However, we can't just steal reservations from the data
+		 * blocks if this is an RT inodes as the data and metadata
+		 * blocks come from different pools.  We'll have to live with
+		 * under-filled indirect reservation in this case.
+		 */
+		da_new = got_indlen + new_indlen;
+		if (da_new > da_old && !isrt) {
+			stolen = XFS_FILBLKS_MIN(da_new - da_old,
+						 del->br_blockcount);
+			da_old += stolen;
+		}
+		if (da_new > da_old)
+			xfs_bmap_split_indlen(da_old, &got_indlen, &new_indlen);
+		da_new = got_indlen + new_indlen;
 
 		got->br_startblock = nullstartblock((int)got_indlen);
 
@@ -4919,20 +5066,21 @@ xfs_bmap_del_extent_delay(
 		xfs_iext_next(ifp, icur);
 		xfs_iext_insert(ip, icur, &new, state);
 
-		da_new = got_indlen + new_indlen - stolen;
 		del->br_blockcount -= stolen;
 		break;
 	}
 
 	ASSERT(da_old >= da_new);
 	da_diff = da_old - da_new;
-	if (!isrt)
-		da_diff += del->br_blockcount;
-	if (da_diff) {
-		xfs_mod_fdblocks(mp, da_diff, false);
-		xfs_mod_delalloc(mp, -da_diff);
-	}
-	return error;
+	fdblocks = da_diff;
+
+	if (isrt)
+		xfs_add_frextents(mp, xfs_rtb_to_rtx(mp, del->br_blockcount));
+	else
+		fdblocks += del->br_blockcount;
+
+	xfs_add_fdblocks(mp, fdblocks);
+	xfs_mod_delalloc(ip, -(int64_t)del->br_blockcount, -da_diff);
 }
 
 void
@@ -5023,8 +5171,7 @@ xfs_bmap_del_extent_real(
 {
 	xfs_fsblock_t		del_endblock=0;	/* first block past del */
 	xfs_fileoff_t		del_endoff;	/* first offset past del */
-	int			do_fx;	/* free extent at end of routine */
-	int			error;	/* error return value */
+	int			error = 0;	/* error return value */
 	struct xfs_bmbt_irec	got;	/* current extent entry */
 	xfs_fileoff_t		got_endoff;	/* first offset past got */
 	int			i;	/* temp state */
@@ -5067,20 +5214,10 @@ xfs_bmap_del_extent_real(
 		return -ENOSPC;
 
 	*logflagsp = XFS_ILOG_CORE;
-	if (whichfork == XFS_DATA_FORK && XFS_IS_REALTIME_INODE(ip)) {
-		if (!(bflags & XFS_BMAPI_REMAP)) {
-			error = xfs_rtfree_blocks(tp, del->br_startblock,
-					del->br_blockcount);
-			if (error)
-				return error;
-		}
-
-		do_fx = 0;
+	if (xfs_ifork_is_realtime(ip, whichfork))
 		qfield = XFS_TRANS_DQ_RTBCOUNT;
-	} else {
-		do_fx = 1;
+	else
 		qfield = XFS_TRANS_DQ_BCOUNT;
-	}
 	nblks = del->br_blockcount;
 
 	del_endblock = del->br_startblock + del->br_blockcount;
@@ -5088,8 +5225,10 @@ xfs_bmap_del_extent_real(
 		error = xfs_bmbt_lookup_eq(cur, &got, &i);
 		if (error)
 			return error;
-		if (XFS_IS_CORRUPT(mp, i != 1))
+		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			return -EFSCORRUPTED;
+		}
 	}
 
 	if (got.br_startoff == del->br_startoff)
@@ -5113,8 +5252,10 @@ xfs_bmap_del_extent_real(
 		}
 		if ((error = xfs_btree_delete(cur, &i)))
 			return error;
-		if (XFS_IS_CORRUPT(mp, i != 1))
+		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			return -EFSCORRUPTED;
+		}
 		break;
 	case BMAP_LEFT_FILLING:
 		/*
@@ -5186,8 +5327,10 @@ xfs_bmap_del_extent_real(
 				error = xfs_bmbt_lookup_eq(cur, &got, &i);
 				if (error)
 					return error;
-				if (XFS_IS_CORRUPT(mp, i != 1))
+				if (XFS_IS_CORRUPT(mp, i != 1)) {
+					xfs_btree_mark_sick(cur);
 					return -EFSCORRUPTED;
+				}
 				/*
 				 * Update the btree record back
 				 * to the original value.
@@ -5203,8 +5346,10 @@ xfs_bmap_del_extent_real(
 				*logflagsp = 0;
 				return -ENOSPC;
 			}
-			if (XFS_IS_CORRUPT(mp, i != 1))
+			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				return -EFSCORRUPTED;
+			}
 		} else
 			*logflagsp |= xfs_ilog_fext(whichfork);
 
@@ -5220,18 +5365,29 @@ xfs_bmap_del_extent_real(
 	/*
 	 * If we need to, add to list of extents to delete.
 	 */
-	if (do_fx && !(bflags & XFS_BMAPI_REMAP)) {
+	if (!(bflags & XFS_BMAPI_REMAP)) {
 		if (xfs_is_reflink_inode(ip) && whichfork == XFS_DATA_FORK) {
 			xfs_refcount_decrease_extent(tp, del);
+		} else if (xfs_ifork_is_realtime(ip, whichfork)) {
+			/*
+			 * Ensure the bitmap and summary inodes are locked
+			 * and joined to the transaction before modifying them.
+			 */
+			if (!(tp->t_flags & XFS_TRANS_RTBITMAP_LOCKED)) {
+				tp->t_flags |= XFS_TRANS_RTBITMAP_LOCKED;
+				xfs_rtbitmap_lock(tp, mp);
+			}
+			error = xfs_rtfree_blocks(tp, del->br_startblock,
+					del->br_blockcount);
 		} else {
 			error = xfs_free_extent_later(tp, del->br_startblock,
 					del->br_blockcount, NULL,
 					XFS_AG_RESV_NONE,
 					((bflags & XFS_BMAPI_NODISCARD) ||
 					del->br_state == XFS_EXT_UNWRITTEN));
-			if (error)
-				return error;
 		}
+		if (error)
+			return error;
 	}
 
 	/*
@@ -5286,12 +5442,14 @@ __xfs_bunmapi(
 	whichfork = xfs_bmapi_whichfork(flags);
 	ASSERT(whichfork != XFS_COW_FORK);
 	ifp = xfs_ifork_ptr(ip, whichfork);
-	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)))
+	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp))) {
+		xfs_bmap_mark_sick(ip, whichfork);
 		return -EFSCORRUPTED;
+	}
 	if (xfs_is_shutdown(mp))
 		return -EIO;
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 	ASSERT(len > 0);
 	ASSERT(nexts >= 0);
 
@@ -5304,7 +5462,7 @@ __xfs_bunmapi(
 		return 0;
 	}
 	XFS_STATS_INC(mp, xs_blk_unmap);
-	isrt = (whichfork == XFS_DATA_FORK) && XFS_IS_REALTIME_INODE(ip);
+	isrt = xfs_ifork_is_realtime(ip, whichfork);
 	end = start + len;
 
 	if (!xfs_iext_lookup_extent_before(ip, ifp, &end, &icur, &got)) {
@@ -5317,20 +5475,9 @@ __xfs_bunmapi(
 	if (ifp->if_format == XFS_DINODE_FMT_BTREE) {
 		ASSERT(ifp->if_format == XFS_DINODE_FMT_BTREE);
 		cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
-		cur->bc_ino.flags = 0;
 	} else
 		cur = NULL;
 
-	if (isrt) {
-		/*
-		 * Synchronize by locking the bitmap inode.
-		 */
-		xfs_ilock(mp->m_rbmip, XFS_ILOCK_EXCL|XFS_ILOCK_RTBITMAP);
-		xfs_trans_ijoin(tp, mp->m_rbmip, XFS_ILOCK_EXCL);
-		xfs_ilock(mp->m_rsumip, XFS_ILOCK_EXCL|XFS_ILOCK_RTSUM);
-		xfs_trans_ijoin(tp, mp->m_rsumip, XFS_ILOCK_EXCL);
-	}
-
 	extno = 0;
 	while (end != (xfs_fileoff_t)-1 && end >= start &&
 	       (nexts == 0 || extno < nexts)) {
@@ -5367,7 +5514,7 @@ __xfs_bunmapi(
 		if (del.br_startoff + del.br_blockcount > end + 1)
 			del.br_blockcount = end + 1 - del.br_startoff;
 
-		if (!isrt)
+		if (!isrt || (flags & XFS_BMAPI_REMAP))
 			goto delete;
 
 		mod = xfs_rtb_to_rtxoff(mp,
@@ -5385,7 +5532,7 @@ __xfs_bunmapi(
 				 * This piece is unwritten, or we're not
 				 * using unwritten extents.  Skip over it.
 				 */
-				ASSERT(end >= mod);
+				ASSERT((flags & XFS_BMAPI_REMAP) || end >= mod);
 				end -= mod > del.br_blockcount ?
 					del.br_blockcount : mod;
 				if (end < got.br_startoff &&
@@ -5491,18 +5638,16 @@ __xfs_bunmapi(
 
 delete:
 		if (wasdel) {
-			error = xfs_bmap_del_extent_delay(ip, whichfork, &icur,
-					&got, &del);
+			xfs_bmap_del_extent_delay(ip, whichfork, &icur, &got, &del);
 		} else {
 			error = xfs_bmap_del_extent_real(ip, tp, &icur, cur,
 					&del, &tmp_logflags, whichfork,
 					flags);
 			logflags |= tmp_logflags;
+			if (error)
+				goto error0;
 		}
 
-		if (error)
-			goto error0;
-
 		end = del.br_startoff - 1;
 nodelete:
 		/*
@@ -5555,7 +5700,7 @@ error0:
 		xfs_trans_log_inode(tp, ip, logflags);
 	if (cur) {
 		if (!error)
-			cur->bc_ino.allocated = 0;
+			cur->bc_bmap.allocated = 0;
 		xfs_btree_del_cursor(cur, error);
 	}
 	return error;
@@ -5635,8 +5780,7 @@ xfs_bmse_merge(
 
 	blockcount = left->br_blockcount + got->br_blockcount;
 
-	ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL));
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
 	ASSERT(xfs_bmse_can_merge(left, got, shift));
 
 	new = *left;
@@ -5657,21 +5801,27 @@ xfs_bmse_merge(
 	error = xfs_bmbt_lookup_eq(cur, got, &i);
 	if (error)
 		return error;
-	if (XFS_IS_CORRUPT(mp, i != 1))
+	if (XFS_IS_CORRUPT(mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		return -EFSCORRUPTED;
+	}
 
 	error = xfs_btree_delete(cur, &i);
 	if (error)
 		return error;
-	if (XFS_IS_CORRUPT(mp, i != 1))
+	if (XFS_IS_CORRUPT(mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		return -EFSCORRUPTED;
+	}
 
 	/* lookup and update size of the previous extent */
 	error = xfs_bmbt_lookup_eq(cur, left, &i);
 	if (error)
 		return error;
-	if (XFS_IS_CORRUPT(mp, i != 1))
+	if (XFS_IS_CORRUPT(mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		return -EFSCORRUPTED;
+	}
 
 	error = xfs_bmbt_update(cur, &new);
 	if (error)
@@ -5719,8 +5869,10 @@ xfs_bmap_shift_update_extent(
 		error = xfs_bmbt_lookup_eq(cur, &prev, &i);
 		if (error)
 			return error;
-		if (XFS_IS_CORRUPT(mp, i != 1))
+		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			return -EFSCORRUPTED;
+		}
 
 		error = xfs_bmbt_update(cur, got);
 		if (error)
@@ -5758,28 +5910,28 @@ xfs_bmap_collapse_extents(
 
 	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)) ||
 	    XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) {
+		xfs_bmap_mark_sick(ip, whichfork);
 		return -EFSCORRUPTED;
 	}
 
 	if (xfs_is_shutdown(mp))
 		return -EIO;
 
-	ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
 
 	error = xfs_iread_extents(tp, ip, whichfork);
 	if (error)
 		return error;
 
-	if (ifp->if_format == XFS_DINODE_FMT_BTREE) {
+	if (ifp->if_format == XFS_DINODE_FMT_BTREE)
 		cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
-		cur->bc_ino.flags = 0;
-	}
 
 	if (!xfs_iext_lookup_extent(ip, ifp, *next_fsb, &icur, &got)) {
 		*done = true;
 		goto del_cursor;
 	}
 	if (XFS_IS_CORRUPT(mp, isnullstartblock(got.br_startblock))) {
+		xfs_bmap_mark_sick(ip, whichfork);
 		error = -EFSCORRUPTED;
 		goto del_cursor;
 	}
@@ -5837,7 +5989,7 @@ xfs_bmap_can_insert_extents(
 	int			is_empty;
 	int			error = 0;
 
-	ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL);
 
 	if (xfs_is_shutdown(ip->i_mount))
 		return -EIO;
@@ -5873,22 +6025,21 @@ xfs_bmap_insert_extents(
 
 	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)) ||
 	    XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) {
+		xfs_bmap_mark_sick(ip, whichfork);
 		return -EFSCORRUPTED;
 	}
 
 	if (xfs_is_shutdown(mp))
 		return -EIO;
 
-	ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
 
 	error = xfs_iread_extents(tp, ip, whichfork);
 	if (error)
 		return error;
 
-	if (ifp->if_format == XFS_DINODE_FMT_BTREE) {
+	if (ifp->if_format == XFS_DINODE_FMT_BTREE)
 		cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
-		cur->bc_ino.flags = 0;
-	}
 
 	if (*next_fsb == NULLFSBLOCK) {
 		xfs_iext_last(ifp, &icur);
@@ -5904,11 +6055,13 @@ xfs_bmap_insert_extents(
 		}
 	}
 	if (XFS_IS_CORRUPT(mp, isnullstartblock(got.br_startblock))) {
+		xfs_bmap_mark_sick(ip, whichfork);
 		error = -EFSCORRUPTED;
 		goto del_cursor;
 	}
 
 	if (XFS_IS_CORRUPT(mp, stop_fsb > got.br_startoff)) {
+		xfs_bmap_mark_sick(ip, whichfork);
 		error = -EFSCORRUPTED;
 		goto del_cursor;
 	}
@@ -5976,6 +6129,7 @@ xfs_bmap_split_extent(
 
 	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)) ||
 	    XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) {
+		xfs_bmap_mark_sick(ip, whichfork);
 		return -EFSCORRUPTED;
 	}
 
@@ -6002,11 +6156,11 @@ xfs_bmap_split_extent(
 
 	if (ifp->if_format == XFS_DINODE_FMT_BTREE) {
 		cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
-		cur->bc_ino.flags = 0;
 		error = xfs_bmbt_lookup_eq(cur, &got, &i);
 		if (error)
 			goto del_cursor;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto del_cursor;
 		}
@@ -6034,6 +6188,7 @@ xfs_bmap_split_extent(
 		if (error)
 			goto del_cursor;
 		if (XFS_IS_CORRUPT(mp, i != 0)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto del_cursor;
 		}
@@ -6041,6 +6196,7 @@ xfs_bmap_split_extent(
 		if (error)
 			goto del_cursor;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto del_cursor;
 		}
@@ -6060,7 +6216,7 @@ xfs_bmap_split_extent(
 
 del_cursor:
 	if (cur) {
-		cur->bc_ino.allocated = 0;
+		cur->bc_bmap.allocated = 0;
 		xfs_btree_del_cursor(cur, error);
 	}
 
@@ -6069,17 +6225,8 @@ del_cursor:
 	return error;
 }
 
-/* Deferred mapping is only for real extents in the data fork. */
-static bool
-xfs_bmap_is_update_needed(
-	struct xfs_bmbt_irec	*bmap)
-{
-	return  bmap->br_startblock != HOLESTARTBLOCK &&
-		bmap->br_startblock != DELAYSTARTBLOCK;
-}
-
 /* Record a bmap intent. */
-static int
+static inline void
 __xfs_bmap_add(
 	struct xfs_trans		*tp,
 	enum xfs_bmap_intent_type	type,
@@ -6089,25 +6236,19 @@ __xfs_bmap_add(
 {
 	struct xfs_bmap_intent		*bi;
 
-	trace_xfs_bmap_defer(tp->t_mountp,
-			XFS_FSB_TO_AGNO(tp->t_mountp, bmap->br_startblock),
-			type,
-			XFS_FSB_TO_AGBNO(tp->t_mountp, bmap->br_startblock),
-			ip->i_ino, whichfork,
-			bmap->br_startoff,
-			bmap->br_blockcount,
-			bmap->br_state);
+	if ((whichfork != XFS_DATA_FORK && whichfork != XFS_ATTR_FORK) ||
+	    bmap->br_startblock == HOLESTARTBLOCK ||
+	    bmap->br_startblock == DELAYSTARTBLOCK)
+		return;
 
-	bi = kmem_cache_alloc(xfs_bmap_intent_cache, GFP_NOFS | __GFP_NOFAIL);
+	bi = kmem_cache_alloc(xfs_bmap_intent_cache, GFP_KERNEL | __GFP_NOFAIL);
 	INIT_LIST_HEAD(&bi->bi_list);
 	bi->bi_type = type;
 	bi->bi_owner = ip;
 	bi->bi_whichfork = whichfork;
 	bi->bi_bmap = *bmap;
 
-	xfs_bmap_update_get_group(tp->t_mountp, bi);
-	xfs_defer_add(tp, &bi->bi_list, &xfs_bmap_update_defer_type);
-	return 0;
+	xfs_bmap_defer_add(tp, bi);
 }
 
 /* Map an extent into a file. */
@@ -6115,12 +6256,10 @@ void
 xfs_bmap_map_extent(
 	struct xfs_trans	*tp,
 	struct xfs_inode	*ip,
+	int			whichfork,
 	struct xfs_bmbt_irec	*PREV)
 {
-	if (!xfs_bmap_is_update_needed(PREV))
-		return;
-
-	__xfs_bmap_add(tp, XFS_BMAP_MAP, ip, XFS_DATA_FORK, PREV);
+	__xfs_bmap_add(tp, XFS_BMAP_MAP, ip, whichfork, PREV);
 }
 
 /* Unmap an extent out of a file. */
@@ -6128,12 +6267,10 @@ void
 xfs_bmap_unmap_extent(
 	struct xfs_trans	*tp,
 	struct xfs_inode	*ip,
+	int			whichfork,
 	struct xfs_bmbt_irec	*PREV)
 {
-	if (!xfs_bmap_is_update_needed(PREV))
-		return;
-
-	__xfs_bmap_add(tp, XFS_BMAP_UNMAP, ip, XFS_DATA_FORK, PREV);
+	__xfs_bmap_add(tp, XFS_BMAP_UNMAP, ip, whichfork, PREV);
 }
 
 /*
@@ -6147,36 +6284,35 @@ xfs_bmap_finish_one(
 {
 	struct xfs_bmbt_irec		*bmap = &bi->bi_bmap;
 	int				error = 0;
+	int				flags = 0;
 
-	ASSERT(tp->t_highest_agno == NULLAGNUMBER);
+	if (bi->bi_whichfork == XFS_ATTR_FORK)
+		flags |= XFS_BMAPI_ATTRFORK;
 
-	trace_xfs_bmap_deferred(tp->t_mountp,
-			XFS_FSB_TO_AGNO(tp->t_mountp, bmap->br_startblock),
-			bi->bi_type,
-			XFS_FSB_TO_AGBNO(tp->t_mountp, bmap->br_startblock),
-			bi->bi_owner->i_ino, bi->bi_whichfork,
-			bmap->br_startoff, bmap->br_blockcount,
-			bmap->br_state);
+	ASSERT(tp->t_highest_agno == NULLAGNUMBER);
 
-	if (WARN_ON_ONCE(bi->bi_whichfork != XFS_DATA_FORK))
-		return -EFSCORRUPTED;
+	trace_xfs_bmap_deferred(bi);
 
-	if (XFS_TEST_ERROR(false, tp->t_mountp,
-			XFS_ERRTAG_BMAP_FINISH_ONE))
+	if (XFS_TEST_ERROR(false, tp->t_mountp, XFS_ERRTAG_BMAP_FINISH_ONE))
 		return -EIO;
 
 	switch (bi->bi_type) {
 	case XFS_BMAP_MAP:
+		if (bi->bi_bmap.br_state == XFS_EXT_UNWRITTEN)
+			flags |= XFS_BMAPI_PREALLOC;
 		error = xfs_bmapi_remap(tp, bi->bi_owner, bmap->br_startoff,
-				bmap->br_blockcount, bmap->br_startblock, 0);
+				bmap->br_blockcount, bmap->br_startblock,
+				flags);
 		bmap->br_blockcount = 0;
 		break;
 	case XFS_BMAP_UNMAP:
 		error = __xfs_bunmapi(tp, bi->bi_owner, bmap->br_startoff,
-				&bmap->br_blockcount, XFS_BMAPI_REMAP, 1);
+				&bmap->br_blockcount, flags | XFS_BMAPI_REMAP,
+				1);
 		break;
 	default:
 		ASSERT(0);
+		xfs_bmap_mark_sick(bi->bi_owner, bi->bi_whichfork);
 		error = -EFSCORRUPTED;
 	}
 
@@ -6257,7 +6393,7 @@ xfs_bunmapi_range(
 	xfs_filblks_t		unmap_len = endoff - startoff + 1;
 	int			error = 0;
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 
 	while (unmap_len > 0) {
 		ASSERT((*tpp)->t_highest_agno == NULLAGNUMBER);
@@ -6270,7 +6406,51 @@ xfs_bunmapi_range(
 		error = xfs_defer_finish(tpp);
 		if (error)
 			goto out;
+		cond_resched();
 	}
 out:
 	return error;
 }
+
+struct xfs_bmap_query_range {
+	xfs_bmap_query_range_fn	fn;
+	void			*priv;
+};
+
+/* Format btree record and pass to our callback. */
+STATIC int
+xfs_bmap_query_range_helper(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_rec	*rec,
+	void				*priv)
+{
+	struct xfs_bmap_query_range	*query = priv;
+	struct xfs_bmbt_irec		irec;
+	xfs_failaddr_t			fa;
+
+	xfs_bmbt_disk_get_all(&rec->bmbt, &irec);
+	fa = xfs_bmap_validate_extent(cur->bc_ino.ip, cur->bc_ino.whichfork,
+			&irec);
+	if (fa) {
+		xfs_btree_mark_sick(cur);
+		return xfs_bmap_complain_bad_rec(cur->bc_ino.ip,
+				cur->bc_ino.whichfork, fa, &irec);
+	}
+
+	return query->fn(cur, &irec, query->priv);
+}
+
+/* Find all bmaps. */
+int
+xfs_bmap_query_all(
+	struct xfs_btree_cur		*cur,
+	xfs_bmap_query_range_fn		fn,
+	void				*priv)
+{
+	struct xfs_bmap_query_range	query = {
+		.priv			= priv,
+		.fn			= fn,
+	};
+
+	return xfs_btree_query_all(cur, xfs_bmap_query_range_helper, &query);
+}
diff --git a/fs/xfs/libxfs/xfs_bmap.h b/fs/xfs/libxfs/xfs_bmap.h
index f6b73f1bad..667b0c2b33 100644
--- a/fs/xfs/libxfs/xfs_bmap.h
+++ b/fs/xfs/libxfs/xfs_bmap.h
@@ -158,7 +158,7 @@ static inline bool xfs_bmap_is_real_extent(const struct xfs_bmbt_irec *irec)
  * Return true if the extent is a real, allocated extent, or false if it is  a
  * delayed allocation, and unwritten extent or a hole.
  */
-static inline bool xfs_bmap_is_written_extent(struct xfs_bmbt_irec *irec)
+static inline bool xfs_bmap_is_written_extent(const struct xfs_bmbt_irec *irec)
 {
 	return xfs_bmap_is_real_extent(irec) &&
 	       irec->br_state != XFS_EXT_UNWRITTEN;
@@ -176,9 +176,16 @@ int	xfs_bmap_longest_free_extent(struct xfs_perag *pag,
 void	xfs_trim_extent(struct xfs_bmbt_irec *irec, xfs_fileoff_t bno,
 		xfs_filblks_t len);
 unsigned int xfs_bmap_compute_attr_offset(struct xfs_mount *mp);
-int	xfs_bmap_add_attrfork(struct xfs_inode *ip, int size, int rsvd);
+int	xfs_bmap_add_attrfork(struct xfs_trans *tp, struct xfs_inode *ip,
+		int size, int rsvd);
 void	xfs_bmap_local_to_extents_empty(struct xfs_trans *tp,
 		struct xfs_inode *ip, int whichfork);
+int xfs_bmap_local_to_extents(struct xfs_trans *tp, struct xfs_inode *ip,
+		xfs_extlen_t total, int *logflagsp, int whichfork,
+		void (*init_fn)(struct xfs_trans *tp, struct xfs_buf *bp,
+				struct xfs_inode *ip, struct xfs_ifork *ifp,
+				void *priv),
+		void *priv);
 void	xfs_bmap_compute_maxlevels(struct xfs_mount *mp, int whichfork);
 int	xfs_bmap_first_unused(struct xfs_trans *tp, struct xfs_inode *ip,
 		xfs_extlen_t len, xfs_fileoff_t *unused, int whichfork);
@@ -195,7 +202,7 @@ int	xfs_bmapi_write(struct xfs_trans *tp, struct xfs_inode *ip,
 int	xfs_bunmapi(struct xfs_trans *tp, struct xfs_inode *ip,
 		xfs_fileoff_t bno, xfs_filblks_t len, uint32_t flags,
 		xfs_extnum_t nexts, int *done);
-int	xfs_bmap_del_extent_delay(struct xfs_inode *ip, int whichfork,
+void	xfs_bmap_del_extent_delay(struct xfs_inode *ip, int whichfork,
 		struct xfs_iext_cursor *cur, struct xfs_bmbt_irec *got,
 		struct xfs_bmbt_irec *del);
 void	xfs_bmap_del_extent_cow(struct xfs_inode *ip,
@@ -232,6 +239,10 @@ enum xfs_bmap_intent_type {
 	XFS_BMAP_UNMAP,
 };
 
+#define XFS_BMAP_INTENT_STRINGS \
+	{ XFS_BMAP_MAP,		"map" }, \
+	{ XFS_BMAP_UNMAP,	"unmap" }
+
 struct xfs_bmap_intent {
 	struct list_head			bi_list;
 	enum xfs_bmap_intent_type		bi_type;
@@ -241,14 +252,11 @@ struct xfs_bmap_intent {
 	struct xfs_bmbt_irec			bi_bmap;
 };
 
-void xfs_bmap_update_get_group(struct xfs_mount *mp,
-		struct xfs_bmap_intent *bi);
-
 int	xfs_bmap_finish_one(struct xfs_trans *tp, struct xfs_bmap_intent *bi);
 void	xfs_bmap_map_extent(struct xfs_trans *tp, struct xfs_inode *ip,
-		struct xfs_bmbt_irec *imap);
+		int whichfork, struct xfs_bmbt_irec *imap);
 void	xfs_bmap_unmap_extent(struct xfs_trans *tp, struct xfs_inode *ip,
-		struct xfs_bmbt_irec *imap);
+		int whichfork, struct xfs_bmbt_irec *imap);
 
 static inline uint32_t xfs_bmap_fork_to_state(int whichfork)
 {
@@ -280,4 +288,12 @@ extern struct kmem_cache	*xfs_bmap_intent_cache;
 int __init xfs_bmap_intent_init_cache(void);
 void xfs_bmap_intent_destroy_cache(void);
 
+typedef int (*xfs_bmap_query_range_fn)(
+	struct xfs_btree_cur	*cur,
+	struct xfs_bmbt_irec	*rec,
+	void			*priv);
+
+int xfs_bmap_query_all(struct xfs_btree_cur *cur, xfs_bmap_query_range_fn fn,
+		void *priv);
+
 #endif	/* __XFS_BMAP_H__ */
diff --git a/fs/xfs/libxfs/xfs_bmap_btree.c b/fs/xfs/libxfs/xfs_bmap_btree.c
index 71f2d50f78..f5d84dcb58 100644
--- a/fs/xfs/libxfs/xfs_bmap_btree.c
+++ b/fs/xfs/libxfs/xfs_bmap_btree.c
@@ -26,6 +26,22 @@
 
 static struct kmem_cache	*xfs_bmbt_cur_cache;
 
+void
+xfs_bmbt_init_block(
+	struct xfs_inode		*ip,
+	struct xfs_btree_block		*buf,
+	struct xfs_buf			*bp,
+	__u16				level,
+	__u16				numrecs)
+{
+	if (bp)
+		xfs_btree_init_buf(ip->i_mount, bp, &xfs_bmbt_ops, level,
+				numrecs, ip->i_ino);
+	else
+		xfs_btree_init_block(ip->i_mount, buf, &xfs_bmbt_ops, level,
+				numrecs, ip->i_ino);
+}
+
 /*
  * Convert on-disk form of btree root to in-memory form.
  */
@@ -44,9 +60,7 @@ xfs_bmdr_to_bmbt(
 	xfs_bmbt_key_t		*tkp;
 	__be64			*tpp;
 
-	xfs_btree_init_block_int(mp, rblock, XFS_BUF_DADDR_NULL,
-				 XFS_BTNUM_BMAP, 0, 0, ip->i_ino,
-				 XFS_BTREE_LONG_PTRS);
+	xfs_bmbt_init_block(ip, rblock, NULL, 0, 0);
 	rblock->bb_level = dblock->bb_level;
 	ASSERT(be16_to_cpu(rblock->bb_level) > 0);
 	rblock->bb_numrecs = dblock->bb_numrecs;
@@ -171,13 +185,8 @@ xfs_bmbt_dup_cursor(
 
 	new = xfs_bmbt_init_cursor(cur->bc_mp, cur->bc_tp,
 			cur->bc_ino.ip, cur->bc_ino.whichfork);
-
-	/*
-	 * Copy the firstblock, dfops, and flags values,
-	 * since init cursor doesn't get them.
-	 */
-	new->bc_ino.flags = cur->bc_ino.flags;
-
+	new->bc_flags |= (cur->bc_flags &
+		(XFS_BTREE_BMBT_INVALID_OWNER | XFS_BTREE_BMBT_WASDEL));
 	return new;
 }
 
@@ -189,10 +198,10 @@ xfs_bmbt_update_cursor(
 	ASSERT((dst->bc_tp->t_highest_agno != NULLAGNUMBER) ||
 	       (dst->bc_ino.ip->i_diflags & XFS_DIFLAG_REALTIME));
 
-	dst->bc_ino.allocated += src->bc_ino.allocated;
+	dst->bc_bmap.allocated += src->bc_bmap.allocated;
 	dst->bc_tp->t_highest_agno = src->bc_tp->t_highest_agno;
 
-	src->bc_ino.allocated = 0;
+	src->bc_bmap.allocated = 0;
 }
 
 STATIC int
@@ -211,7 +220,7 @@ xfs_bmbt_alloc_block(
 	xfs_rmap_ino_bmbt_owner(&args.oinfo, cur->bc_ino.ip->i_ino,
 			cur->bc_ino.whichfork);
 	args.minlen = args.maxlen = args.prod = 1;
-	args.wasdel = cur->bc_ino.flags & XFS_BTCUR_BMBT_WASDEL;
+	args.wasdel = cur->bc_flags & XFS_BTREE_BMBT_WASDEL;
 	if (!args.wasdel && args.tp->t_blk_res == 0)
 		return -ENOSPC;
 
@@ -247,7 +256,7 @@ xfs_bmbt_alloc_block(
 	}
 
 	ASSERT(args.len == 1);
-	cur->bc_ino.allocated++;
+	cur->bc_bmap.allocated++;
 	cur->bc_ino.ip->i_nblocks++;
 	xfs_trans_log_inode(args.tp, cur->bc_ino.ip, XFS_ILOG_CORE);
 	xfs_trans_mod_dquot_byino(args.tp, cur->bc_ino.ip,
@@ -360,14 +369,6 @@ xfs_bmbt_init_rec_from_cur(
 	xfs_bmbt_disk_set_all(&rec->bmbt, &cur->bc_rec.b);
 }
 
-STATIC void
-xfs_bmbt_init_ptr_from_cur(
-	struct xfs_btree_cur	*cur,
-	union xfs_btree_ptr	*ptr)
-{
-	ptr->l = 0;
-}
-
 STATIC int64_t
 xfs_bmbt_key_diff(
 	struct xfs_btree_cur		*cur,
@@ -419,7 +420,7 @@ xfs_bmbt_verify(
 		 * XXX: need a better way of verifying the owner here. Right now
 		 * just make sure there has been one set.
 		 */
-		fa = xfs_btree_lblock_v5hdr_verify(bp, XFS_RMAP_OWN_UNKNOWN);
+		fa = xfs_btree_fsblock_v5hdr_verify(bp, XFS_RMAP_OWN_UNKNOWN);
 		if (fa)
 			return fa;
 	}
@@ -435,7 +436,7 @@ xfs_bmbt_verify(
 	if (level > max(mp->m_bm_maxlevels[0], mp->m_bm_maxlevels[1]))
 		return __this_address;
 
-	return xfs_btree_lblock_verify(bp, mp->m_bmap_dmxr[level != 0]);
+	return xfs_btree_fsblock_verify(bp, mp->m_bmap_dmxr[level != 0]);
 }
 
 static void
@@ -444,7 +445,7 @@ xfs_bmbt_read_verify(
 {
 	xfs_failaddr_t	fa;
 
-	if (!xfs_btree_lblock_verify_crc(bp))
+	if (!xfs_btree_fsblock_verify_crc(bp))
 		xfs_verifier_error(bp, -EFSBADCRC, __this_address);
 	else {
 		fa = xfs_bmbt_verify(bp);
@@ -468,7 +469,7 @@ xfs_bmbt_write_verify(
 		xfs_verifier_error(bp, -EFSCORRUPTED, fa);
 		return;
 	}
-	xfs_btree_lblock_calc_crc(bp);
+	xfs_btree_fsblock_calc_crc(bp);
 }
 
 const struct xfs_buf_ops xfs_bmbt_buf_ops = {
@@ -515,9 +516,16 @@ xfs_bmbt_keys_contiguous(
 				 be64_to_cpu(key2->bmbt.br_startoff));
 }
 
-static const struct xfs_btree_ops xfs_bmbt_ops = {
+const struct xfs_btree_ops xfs_bmbt_ops = {
+	.name			= "bmap",
+	.type			= XFS_BTREE_TYPE_INODE,
+
 	.rec_len		= sizeof(xfs_bmbt_rec_t),
 	.key_len		= sizeof(xfs_bmbt_key_t),
+	.ptr_len		= XFS_BTREE_LONG_PTR_LEN,
+
+	.lru_refs		= XFS_BMAP_BTREE_REF,
+	.statoff		= XFS_STATS_CALC_INDEX(xs_bmbt_2),
 
 	.dup_cursor		= xfs_bmbt_dup_cursor,
 	.update_cursor		= xfs_bmbt_update_cursor,
@@ -529,7 +537,6 @@ static const struct xfs_btree_ops xfs_bmbt_ops = {
 	.init_key_from_rec	= xfs_bmbt_init_key_from_rec,
 	.init_high_key_from_rec	= xfs_bmbt_init_high_key_from_rec,
 	.init_rec_from_cur	= xfs_bmbt_init_rec_from_cur,
-	.init_ptr_from_cur	= xfs_bmbt_init_ptr_from_cur,
 	.key_diff		= xfs_bmbt_key_diff,
 	.diff_two_keys		= xfs_bmbt_diff_two_keys,
 	.buf_ops		= &xfs_bmbt_buf_ops,
@@ -538,35 +545,10 @@ static const struct xfs_btree_ops xfs_bmbt_ops = {
 	.keys_contiguous	= xfs_bmbt_keys_contiguous,
 };
 
-static struct xfs_btree_cur *
-xfs_bmbt_init_common(
-	struct xfs_mount	*mp,
-	struct xfs_trans	*tp,
-	struct xfs_inode	*ip,
-	int			whichfork)
-{
-	struct xfs_btree_cur	*cur;
-
-	ASSERT(whichfork != XFS_COW_FORK);
-
-	cur = xfs_btree_alloc_cursor(mp, tp, XFS_BTNUM_BMAP,
-			mp->m_bm_maxlevels[whichfork], xfs_bmbt_cur_cache);
-	cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_bmbt_2);
-
-	cur->bc_ops = &xfs_bmbt_ops;
-	cur->bc_flags = XFS_BTREE_LONG_PTRS | XFS_BTREE_ROOT_IN_INODE;
-	if (xfs_has_crc(mp))
-		cur->bc_flags |= XFS_BTREE_CRC_BLOCKS;
-
-	cur->bc_ino.ip = ip;
-	cur->bc_ino.allocated = 0;
-	cur->bc_ino.flags = 0;
-
-	return cur;
-}
-
 /*
- * Allocate a new bmap btree cursor.
+ * Create a new bmap btree cursor.
+ *
+ * For staging cursors -1 in passed in whichfork.
  */
 struct xfs_btree_cur *
 xfs_bmbt_init_cursor(
@@ -575,15 +557,34 @@ xfs_bmbt_init_cursor(
 	struct xfs_inode	*ip,
 	int			whichfork)
 {
-	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, whichfork);
 	struct xfs_btree_cur	*cur;
+	unsigned int		maxlevels;
 
-	cur = xfs_bmbt_init_common(mp, tp, ip, whichfork);
+	ASSERT(whichfork != XFS_COW_FORK);
 
-	cur->bc_nlevels = be16_to_cpu(ifp->if_broot->bb_level) + 1;
-	cur->bc_ino.forksize = xfs_inode_fork_size(ip, whichfork);
+	/*
+	 * The Data fork always has larger maxlevel, so use that for staging
+	 * cursors.
+	 */
+	switch (whichfork) {
+	case XFS_STAGING_FORK:
+		maxlevels = mp->m_bm_maxlevels[XFS_DATA_FORK];
+		break;
+	default:
+		maxlevels = mp->m_bm_maxlevels[whichfork];
+		break;
+	}
+	cur = xfs_btree_alloc_cursor(mp, tp, &xfs_bmbt_ops, maxlevels,
+			xfs_bmbt_cur_cache);
+	cur->bc_ino.ip = ip;
 	cur->bc_ino.whichfork = whichfork;
+	cur->bc_bmap.allocated = 0;
+	if (whichfork != XFS_STAGING_FORK) {
+		struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, whichfork);
 
+		cur->bc_nlevels = be16_to_cpu(ifp->if_broot->bb_level) + 1;
+		cur->bc_ino.forksize = xfs_inode_fork_size(ip, whichfork);
+	}
 	return cur;
 }
 
@@ -599,33 +600,6 @@ xfs_bmbt_block_maxrecs(
 }
 
 /*
- * Allocate a new bmap btree cursor for reloading an inode block mapping data
- * structure.  Note that callers can use the staged cursor to reload extents
- * format inode forks if they rebuild the iext tree and commit the staged
- * cursor immediately.
- */
-struct xfs_btree_cur *
-xfs_bmbt_stage_cursor(
-	struct xfs_mount	*mp,
-	struct xfs_inode	*ip,
-	struct xbtree_ifakeroot	*ifake)
-{
-	struct xfs_btree_cur	*cur;
-	struct xfs_btree_ops	*ops;
-
-	/* data fork always has larger maxheight */
-	cur = xfs_bmbt_init_common(mp, NULL, ip, XFS_DATA_FORK);
-	cur->bc_nlevels = ifake->if_levels;
-	cur->bc_ino.forksize = ifake->if_fork_size;
-
-	/* Don't let anyone think we're attached to the real fork yet. */
-	cur->bc_ino.whichfork = -1;
-	xfs_btree_stage_ifakeroot(cur, ifake, &ops);
-	ops->update_cursor = NULL;
-	return cur;
-}
-
-/*
  * Swap in the new inode fork root.  Once we pass this point the newly rebuilt
  * mappings are in place and we have to kill off any old btree blocks.
  */
@@ -665,7 +639,7 @@ xfs_bmbt_commit_staged_btree(
 		break;
 	}
 	xfs_trans_log_inode(tp, cur->bc_ino.ip, flags);
-	xfs_btree_commit_ifakeroot(cur, tp, whichfork, &xfs_bmbt_ops);
+	xfs_btree_commit_ifakeroot(cur, tp, whichfork);
 }
 
 /*
@@ -751,7 +725,7 @@ xfs_bmbt_change_owner(
 	ASSERT(xfs_ifork_ptr(ip, whichfork)->if_format == XFS_DINODE_FMT_BTREE);
 
 	cur = xfs_bmbt_init_cursor(ip->i_mount, tp, ip, whichfork);
-	cur->bc_ino.flags |= XFS_BTCUR_BMBT_INVALID_OWNER;
+	cur->bc_flags |= XFS_BTREE_BMBT_INVALID_OWNER;
 
 	error = xfs_btree_change_owner(cur, new_owner, buffer_list);
 	xfs_btree_del_cursor(cur, error);
diff --git a/fs/xfs/libxfs/xfs_bmap_btree.h b/fs/xfs/libxfs/xfs_bmap_btree.h
index 151b8491f6..de1b73f122 100644
--- a/fs/xfs/libxfs/xfs_bmap_btree.h
+++ b/fs/xfs/libxfs/xfs_bmap_btree.h
@@ -107,8 +107,6 @@ extern int xfs_bmbt_change_owner(struct xfs_trans *tp, struct xfs_inode *ip,
 
 extern struct xfs_btree_cur *xfs_bmbt_init_cursor(struct xfs_mount *,
 		struct xfs_trans *, struct xfs_inode *, int);
-struct xfs_btree_cur *xfs_bmbt_stage_cursor(struct xfs_mount *mp,
-		struct xfs_inode *ip, struct xbtree_ifakeroot *ifake);
 void xfs_bmbt_commit_staged_btree(struct xfs_btree_cur *cur,
 		struct xfs_trans *tp, int whichfork);
 
@@ -120,4 +118,7 @@ unsigned int xfs_bmbt_maxlevels_ondisk(void);
 int __init xfs_bmbt_init_cur_cache(void);
 void xfs_bmbt_destroy_cur_cache(void);
 
+void xfs_bmbt_init_block(struct xfs_inode *ip, struct xfs_btree_block *buf,
+		struct xfs_buf *bp, __u16 level, __u16 numrecs);
+
 #endif	/* __XFS_BMAP_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c
index ea8d3659df..d29547572a 100644
--- a/fs/xfs/libxfs/xfs_btree.c
+++ b/fs/xfs/libxfs/xfs_btree.c
@@ -27,28 +27,24 @@
 #include "xfs_bmap_btree.h"
 #include "xfs_rmap_btree.h"
 #include "xfs_refcount_btree.h"
+#include "xfs_health.h"
+#include "xfs_buf_mem.h"
+#include "xfs_btree_mem.h"
 
 /*
  * Btree magic numbers.
  */
-static const uint32_t xfs_magics[2][XFS_BTNUM_MAX] = {
-	{ XFS_ABTB_MAGIC, XFS_ABTC_MAGIC, 0, XFS_BMAP_MAGIC, XFS_IBT_MAGIC,
-	  XFS_FIBT_MAGIC, 0 },
-	{ XFS_ABTB_CRC_MAGIC, XFS_ABTC_CRC_MAGIC, XFS_RMAP_CRC_MAGIC,
-	  XFS_BMAP_CRC_MAGIC, XFS_IBT_CRC_MAGIC, XFS_FIBT_CRC_MAGIC,
-	  XFS_REFC_CRC_MAGIC }
-};
-
 uint32_t
 xfs_btree_magic(
-	int			crc,
-	xfs_btnum_t		btnum)
+	struct xfs_mount		*mp,
+	const struct xfs_btree_ops	*ops)
 {
-	uint32_t		magic = xfs_magics[crc][btnum];
+	int				idx = xfs_has_crc(mp) ? 1 : 0;
+	__be32				magic = ops->buf_ops->magic[idx];
 
 	/* Ensure we asked for crc for crc-only magics. */
 	ASSERT(magic != 0);
-	return magic;
+	return be32_to_cpu(magic);
 }
 
 /*
@@ -63,10 +59,8 @@ xfs_btree_magic(
  * bytes.
  */
 static inline xfs_failaddr_t
-xfs_btree_check_lblock_siblings(
+xfs_btree_check_fsblock_siblings(
 	struct xfs_mount	*mp,
-	struct xfs_btree_cur	*cur,
-	int			level,
 	xfs_fsblock_t		fsb,
 	__be64			dsibling)
 {
@@ -78,22 +72,33 @@ xfs_btree_check_lblock_siblings(
 	sibling = be64_to_cpu(dsibling);
 	if (sibling == fsb)
 		return __this_address;
-	if (level >= 0) {
-		if (!xfs_btree_check_lptr(cur, sibling, level + 1))
-			return __this_address;
-	} else {
-		if (!xfs_verify_fsbno(mp, sibling))
-			return __this_address;
-	}
+	if (!xfs_verify_fsbno(mp, sibling))
+		return __this_address;
+	return NULL;
+}
 
+static inline xfs_failaddr_t
+xfs_btree_check_memblock_siblings(
+	struct xfs_buftarg	*btp,
+	xfbno_t			bno,
+	__be64			dsibling)
+{
+	xfbno_t			sibling;
+
+	if (dsibling == cpu_to_be64(NULLFSBLOCK))
+		return NULL;
+
+	sibling = be64_to_cpu(dsibling);
+	if (sibling == bno)
+		return __this_address;
+	if (!xmbuf_verify_daddr(btp, xfbno_to_daddr(sibling)))
+		return __this_address;
 	return NULL;
 }
 
 static inline xfs_failaddr_t
-xfs_btree_check_sblock_siblings(
+xfs_btree_check_agblock_siblings(
 	struct xfs_perag	*pag,
-	struct xfs_btree_cur	*cur,
-	int			level,
 	xfs_agblock_t		agbno,
 	__be32			dsibling)
 {
@@ -105,34 +110,21 @@ xfs_btree_check_sblock_siblings(
 	sibling = be32_to_cpu(dsibling);
 	if (sibling == agbno)
 		return __this_address;
-	if (level >= 0) {
-		if (!xfs_btree_check_sptr(cur, sibling, level + 1))
-			return __this_address;
-	} else {
-		if (!xfs_verify_agbno(pag, sibling))
-			return __this_address;
-	}
+	if (!xfs_verify_agbno(pag, sibling))
+		return __this_address;
 	return NULL;
 }
 
-/*
- * Check a long btree block header.  Return the address of the failing check,
- * or NULL if everything is ok.
- */
-xfs_failaddr_t
-__xfs_btree_check_lblock(
+static xfs_failaddr_t
+__xfs_btree_check_lblock_hdr(
 	struct xfs_btree_cur	*cur,
 	struct xfs_btree_block	*block,
 	int			level,
 	struct xfs_buf		*bp)
 {
 	struct xfs_mount	*mp = cur->bc_mp;
-	xfs_btnum_t		btnum = cur->bc_btnum;
-	int			crc = xfs_has_crc(mp);
-	xfs_failaddr_t		fa;
-	xfs_fsblock_t		fsb = NULLFSBLOCK;
 
-	if (crc) {
+	if (xfs_has_crc(mp)) {
 		if (!uuid_equal(&block->bb_u.l.bb_uuid, &mp->m_sb.sb_meta_uuid))
 			return __this_address;
 		if (block->bb_u.l.bb_blkno !=
@@ -142,7 +134,7 @@ __xfs_btree_check_lblock(
 			return __this_address;
 	}
 
-	if (be32_to_cpu(block->bb_magic) != xfs_btree_magic(crc, btnum))
+	if (be32_to_cpu(block->bb_magic) != xfs_btree_magic(mp, cur->bc_ops))
 		return __this_address;
 	if (be16_to_cpu(block->bb_level) != level)
 		return __this_address;
@@ -150,44 +142,83 @@ __xfs_btree_check_lblock(
 	    cur->bc_ops->get_maxrecs(cur, level))
 		return __this_address;
 
-	if (bp)
-		fsb = XFS_DADDR_TO_FSB(mp, xfs_buf_daddr(bp));
+	return NULL;
+}
+
+/*
+ * Check a long btree block header.  Return the address of the failing check,
+ * or NULL if everything is ok.
+ */
+static xfs_failaddr_t
+__xfs_btree_check_fsblock(
+	struct xfs_btree_cur	*cur,
+	struct xfs_btree_block	*block,
+	int			level,
+	struct xfs_buf		*bp)
+{
+	struct xfs_mount	*mp = cur->bc_mp;
+	xfs_failaddr_t		fa;
+	xfs_fsblock_t		fsb;
+
+	fa = __xfs_btree_check_lblock_hdr(cur, block, level, bp);
+	if (fa)
+		return fa;
 
-	fa = xfs_btree_check_lblock_siblings(mp, cur, level, fsb,
+	/*
+	 * For inode-rooted btrees, the root block sits in the inode fork.  In
+	 * that case bp is NULL, and the block must not have any siblings.
+	 */
+	if (!bp) {
+		if (block->bb_u.l.bb_leftsib != cpu_to_be64(NULLFSBLOCK))
+			return __this_address;
+		if (block->bb_u.l.bb_rightsib != cpu_to_be64(NULLFSBLOCK))
+			return __this_address;
+		return NULL;
+	}
+
+	fsb = XFS_DADDR_TO_FSB(mp, xfs_buf_daddr(bp));
+	fa = xfs_btree_check_fsblock_siblings(mp, fsb,
 			block->bb_u.l.bb_leftsib);
 	if (!fa)
-		fa = xfs_btree_check_lblock_siblings(mp, cur, level, fsb,
+		fa = xfs_btree_check_fsblock_siblings(mp, fsb,
 				block->bb_u.l.bb_rightsib);
 	return fa;
 }
 
-/* Check a long btree block header. */
-static int
-xfs_btree_check_lblock(
+/*
+ * Check an in-memory btree block header.  Return the address of the failing
+ * check, or NULL if everything is ok.
+ */
+static xfs_failaddr_t
+__xfs_btree_check_memblock(
 	struct xfs_btree_cur	*cur,
 	struct xfs_btree_block	*block,
 	int			level,
 	struct xfs_buf		*bp)
 {
-	struct xfs_mount	*mp = cur->bc_mp;
+	struct xfs_buftarg	*btp = cur->bc_mem.xfbtree->target;
 	xfs_failaddr_t		fa;
+	xfbno_t			bno;
 
-	fa = __xfs_btree_check_lblock(cur, block, level, bp);
-	if (XFS_IS_CORRUPT(mp, fa != NULL) ||
-	    XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BTREE_CHECK_LBLOCK)) {
-		if (bp)
-			trace_xfs_btree_corrupt(bp, _RET_IP_);
-		return -EFSCORRUPTED;
-	}
-	return 0;
+	fa = __xfs_btree_check_lblock_hdr(cur, block, level, bp);
+	if (fa)
+		return fa;
+
+	bno = xfs_daddr_to_xfbno(xfs_buf_daddr(bp));
+	fa = xfs_btree_check_memblock_siblings(btp, bno,
+			block->bb_u.l.bb_leftsib);
+	if (!fa)
+		fa = xfs_btree_check_memblock_siblings(btp, bno,
+				block->bb_u.l.bb_rightsib);
+	return fa;
 }
 
 /*
  * Check a short btree block header.  Return the address of the failing check,
  * or NULL if everything is ok.
  */
-xfs_failaddr_t
-__xfs_btree_check_sblock(
+static xfs_failaddr_t
+__xfs_btree_check_agblock(
 	struct xfs_btree_cur	*cur,
 	struct xfs_btree_block	*block,
 	int			level,
@@ -195,20 +226,17 @@ __xfs_btree_check_sblock(
 {
 	struct xfs_mount	*mp = cur->bc_mp;
 	struct xfs_perag	*pag = cur->bc_ag.pag;
-	xfs_btnum_t		btnum = cur->bc_btnum;
-	int			crc = xfs_has_crc(mp);
 	xfs_failaddr_t		fa;
-	xfs_agblock_t		agbno = NULLAGBLOCK;
+	xfs_agblock_t		agbno;
 
-	if (crc) {
+	if (xfs_has_crc(mp)) {
 		if (!uuid_equal(&block->bb_u.s.bb_uuid, &mp->m_sb.sb_meta_uuid))
 			return __this_address;
-		if (block->bb_u.s.bb_blkno !=
-		    cpu_to_be64(bp ? xfs_buf_daddr(bp) : XFS_BUF_DADDR_NULL))
+		if (block->bb_u.s.bb_blkno != cpu_to_be64(xfs_buf_daddr(bp)))
 			return __this_address;
 	}
 
-	if (be32_to_cpu(block->bb_magic) != xfs_btree_magic(crc, btnum))
+	if (be32_to_cpu(block->bb_magic) != xfs_btree_magic(mp, cur->bc_ops))
 		return __this_address;
 	if (be16_to_cpu(block->bb_level) != level)
 		return __this_address;
@@ -216,36 +244,45 @@ __xfs_btree_check_sblock(
 	    cur->bc_ops->get_maxrecs(cur, level))
 		return __this_address;
 
-	if (bp)
-		agbno = xfs_daddr_to_agbno(mp, xfs_buf_daddr(bp));
-
-	fa = xfs_btree_check_sblock_siblings(pag, cur, level, agbno,
+	agbno = xfs_daddr_to_agbno(mp, xfs_buf_daddr(bp));
+	fa = xfs_btree_check_agblock_siblings(pag, agbno,
 			block->bb_u.s.bb_leftsib);
 	if (!fa)
-		fa = xfs_btree_check_sblock_siblings(pag, cur, level, agbno,
+		fa = xfs_btree_check_agblock_siblings(pag, agbno,
 				block->bb_u.s.bb_rightsib);
 	return fa;
 }
 
-/* Check a short btree block header. */
-STATIC int
-xfs_btree_check_sblock(
+/*
+ * Internal btree block check.
+ *
+ * Return NULL if the block is ok or the address of the failed check otherwise.
+ */
+xfs_failaddr_t
+__xfs_btree_check_block(
 	struct xfs_btree_cur	*cur,
 	struct xfs_btree_block	*block,
 	int			level,
 	struct xfs_buf		*bp)
 {
-	struct xfs_mount	*mp = cur->bc_mp;
-	xfs_failaddr_t		fa;
-
-	fa = __xfs_btree_check_sblock(cur, block, level, bp);
-	if (XFS_IS_CORRUPT(mp, fa != NULL) ||
-	    XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BTREE_CHECK_SBLOCK)) {
-		if (bp)
-			trace_xfs_btree_corrupt(bp, _RET_IP_);
-		return -EFSCORRUPTED;
+	switch (cur->bc_ops->type) {
+	case XFS_BTREE_TYPE_MEM:
+		return __xfs_btree_check_memblock(cur, block, level, bp);
+	case XFS_BTREE_TYPE_AG:
+		return __xfs_btree_check_agblock(cur, block, level, bp);
+	case XFS_BTREE_TYPE_INODE:
+		return __xfs_btree_check_fsblock(cur, block, level, bp);
+	default:
+		ASSERT(0);
+		return __this_address;
 	}
-	return 0;
+}
+
+static inline unsigned int xfs_btree_block_errtag(struct xfs_btree_cur *cur)
+{
+	if (cur->bc_ops->ptr_len == XFS_BTREE_SHORT_PTR_LEN)
+		return XFS_ERRTAG_BTREE_CHECK_SBLOCK;
+	return XFS_ERRTAG_BTREE_CHECK_LBLOCK;
 }
 
 /*
@@ -258,34 +295,49 @@ xfs_btree_check_block(
 	int			level,	/* level of the btree block */
 	struct xfs_buf		*bp)	/* buffer containing block, if any */
 {
-	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
-		return xfs_btree_check_lblock(cur, block, level, bp);
-	else
-		return xfs_btree_check_sblock(cur, block, level, bp);
-}
+	struct xfs_mount	*mp = cur->bc_mp;
+	xfs_failaddr_t		fa;
 
-/* Check that this long pointer is valid and points within the fs. */
-bool
-xfs_btree_check_lptr(
-	struct xfs_btree_cur	*cur,
-	xfs_fsblock_t		fsbno,
-	int			level)
-{
-	if (level <= 0)
-		return false;
-	return xfs_verify_fsbno(cur->bc_mp, fsbno);
+	fa = __xfs_btree_check_block(cur, block, level, bp);
+	if (XFS_IS_CORRUPT(mp, fa != NULL) ||
+	    XFS_TEST_ERROR(false, mp, xfs_btree_block_errtag(cur))) {
+		if (bp)
+			trace_xfs_btree_corrupt(bp, _RET_IP_);
+		xfs_btree_mark_sick(cur);
+		return -EFSCORRUPTED;
+	}
+	return 0;
 }
 
-/* Check that this short pointer is valid and points within the AG. */
-bool
-xfs_btree_check_sptr(
-	struct xfs_btree_cur	*cur,
-	xfs_agblock_t		agbno,
-	int			level)
+int
+__xfs_btree_check_ptr(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_ptr	*ptr,
+	int				index,
+	int				level)
 {
 	if (level <= 0)
-		return false;
-	return xfs_verify_agbno(cur->bc_ag.pag, agbno);
+		return -EFSCORRUPTED;
+
+	switch (cur->bc_ops->type) {
+	case XFS_BTREE_TYPE_MEM:
+		if (!xfbtree_verify_bno(cur->bc_mem.xfbtree,
+				be64_to_cpu((&ptr->l)[index])))
+			return -EFSCORRUPTED;
+		break;
+	case XFS_BTREE_TYPE_INODE:
+		if (!xfs_verify_fsbno(cur->bc_mp,
+				be64_to_cpu((&ptr->l)[index])))
+			return -EFSCORRUPTED;
+		break;
+	case XFS_BTREE_TYPE_AG:
+		if (!xfs_verify_agbno(cur->bc_ag.pag,
+				be32_to_cpu((&ptr->s)[index])))
+			return -EFSCORRUPTED;
+		break;
+	}
+
+	return 0;
 }
 
 /*
@@ -299,26 +351,35 @@ xfs_btree_check_ptr(
 	int				index,
 	int				level)
 {
-	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
-		if (xfs_btree_check_lptr(cur, be64_to_cpu((&ptr->l)[index]),
-				level))
-			return 0;
-		xfs_err(cur->bc_mp,
-"Inode %llu fork %d: Corrupt btree %d pointer at level %d index %d.",
+	int				error;
+
+	error = __xfs_btree_check_ptr(cur, ptr, index, level);
+	if (error) {
+		switch (cur->bc_ops->type) {
+		case XFS_BTREE_TYPE_MEM:
+			xfs_err(cur->bc_mp,
+"In-memory: Corrupt %sbt flags 0x%x pointer at level %d index %d fa %pS.",
+				cur->bc_ops->name, cur->bc_flags, level, index,
+				__this_address);
+			break;
+		case XFS_BTREE_TYPE_INODE:
+			xfs_err(cur->bc_mp,
+"Inode %llu fork %d: Corrupt %sbt pointer at level %d index %d.",
 				cur->bc_ino.ip->i_ino,
-				cur->bc_ino.whichfork, cur->bc_btnum,
+				cur->bc_ino.whichfork, cur->bc_ops->name,
 				level, index);
-	} else {
-		if (xfs_btree_check_sptr(cur, be32_to_cpu((&ptr->s)[index]),
-				level))
-			return 0;
-		xfs_err(cur->bc_mp,
-"AG %u: Corrupt btree %d pointer at level %d index %d.",
-				cur->bc_ag.pag->pag_agno, cur->bc_btnum,
+			break;
+		case XFS_BTREE_TYPE_AG:
+			xfs_err(cur->bc_mp,
+"AG %u: Corrupt %sbt pointer at level %d index %d.",
+				cur->bc_ag.pag->pag_agno, cur->bc_ops->name,
 				level, index);
+			break;
+		}
+		xfs_btree_mark_sick(cur);
 	}
 
-	return -EFSCORRUPTED;
+	return error;
 }
 
 #ifdef DEBUG
@@ -336,7 +397,7 @@ xfs_btree_check_ptr(
  * it to disk.
  */
 void
-xfs_btree_lblock_calc_crc(
+xfs_btree_fsblock_calc_crc(
 	struct xfs_buf		*bp)
 {
 	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
@@ -350,7 +411,7 @@ xfs_btree_lblock_calc_crc(
 }
 
 bool
-xfs_btree_lblock_verify_crc(
+xfs_btree_fsblock_verify_crc(
 	struct xfs_buf		*bp)
 {
 	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
@@ -374,7 +435,7 @@ xfs_btree_lblock_verify_crc(
  * it to disk.
  */
 void
-xfs_btree_sblock_calc_crc(
+xfs_btree_agblock_calc_crc(
 	struct xfs_buf		*bp)
 {
 	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
@@ -388,7 +449,7 @@ xfs_btree_sblock_calc_crc(
 }
 
 bool
-xfs_btree_sblock_verify_crc(
+xfs_btree_agblock_verify_crc(
 	struct xfs_buf		*bp)
 {
 	struct xfs_btree_block  *block = XFS_BUF_TO_BLOCK(bp);
@@ -410,6 +471,17 @@ xfs_btree_free_block(
 {
 	int			error;
 
+	trace_xfs_btree_free_block(cur, bp);
+
+	/*
+	 * Don't allow block freeing for a staging cursor, because staging
+	 * cursors do not support regular btree modifications.
+	 */
+	if (unlikely(cur->bc_flags & XFS_BTREE_STAGING)) {
+		ASSERT(0);
+		return -EFSCORRUPTED;
+	}
+
 	error = cur->bc_ops->free_block(cur, bp);
 	if (!error) {
 		xfs_trans_binval(cur->bc_tp, bp);
@@ -448,33 +520,70 @@ xfs_btree_del_cursor(
 	 * zero, then we should be shut down or on our way to shutdown due to
 	 * cancelling a dirty transaction on error.
 	 */
-	ASSERT(cur->bc_btnum != XFS_BTNUM_BMAP || cur->bc_ino.allocated == 0 ||
+	ASSERT(!xfs_btree_is_bmap(cur->bc_ops) || cur->bc_bmap.allocated == 0 ||
 	       xfs_is_shutdown(cur->bc_mp) || error != 0);
-	if (unlikely(cur->bc_flags & XFS_BTREE_STAGING))
-		kmem_free(cur->bc_ops);
-	if (!(cur->bc_flags & XFS_BTREE_LONG_PTRS) && cur->bc_ag.pag)
-		xfs_perag_put(cur->bc_ag.pag);
+
+	switch (cur->bc_ops->type) {
+	case XFS_BTREE_TYPE_AG:
+		if (cur->bc_ag.pag)
+			xfs_perag_put(cur->bc_ag.pag);
+		break;
+	case XFS_BTREE_TYPE_INODE:
+		/* nothing to do */
+		break;
+	case XFS_BTREE_TYPE_MEM:
+		if (cur->bc_mem.pag)
+			xfs_perag_put(cur->bc_mem.pag);
+		break;
+	}
+
 	kmem_cache_free(cur->bc_cache, cur);
 }
 
+/* Return the buffer target for this btree's buffer. */
+static inline struct xfs_buftarg *
+xfs_btree_buftarg(
+	struct xfs_btree_cur	*cur)
+{
+	if (cur->bc_ops->type == XFS_BTREE_TYPE_MEM)
+		return cur->bc_mem.xfbtree->target;
+	return cur->bc_mp->m_ddev_targp;
+}
+
+/* Return the block size (in units of 512b sectors) for this btree. */
+static inline unsigned int
+xfs_btree_bbsize(
+	struct xfs_btree_cur	*cur)
+{
+	if (cur->bc_ops->type == XFS_BTREE_TYPE_MEM)
+		return XFBNO_BBSIZE;
+	return cur->bc_mp->m_bsize;
+}
+
 /*
  * Duplicate the btree cursor.
  * Allocate a new one, copy the record, re-get the buffers.
  */
-int					/* error */
+int						/* error */
 xfs_btree_dup_cursor(
-	struct xfs_btree_cur *cur,		/* input cursor */
-	struct xfs_btree_cur **ncur)		/* output cursor */
+	struct xfs_btree_cur	*cur,		/* input cursor */
+	struct xfs_btree_cur	**ncur)		/* output cursor */
 {
-	struct xfs_buf	*bp;		/* btree block's buffer pointer */
-	int		error;		/* error return value */
-	int		i;		/* level number of btree block */
-	xfs_mount_t	*mp;		/* mount structure for filesystem */
-	struct xfs_btree_cur *new;		/* new cursor value */
-	xfs_trans_t	*tp;		/* transaction pointer, can be NULL */
+	struct xfs_mount	*mp = cur->bc_mp;
+	struct xfs_trans	*tp = cur->bc_tp;
+	struct xfs_buf		*bp;
+	struct xfs_btree_cur	*new;
+	int			error;
+	int			i;
 
-	tp = cur->bc_tp;
-	mp = cur->bc_mp;
+	/*
+	 * Don't allow staging cursors to be duplicated because they're supposed
+	 * to be kept private to a single thread.
+	 */
+	if (unlikely(cur->bc_flags & XFS_BTREE_STAGING)) {
+		ASSERT(0);
+		return -EFSCORRUPTED;
+	}
 
 	/*
 	 * Allocate a new cursor like the old one.
@@ -494,10 +603,13 @@ xfs_btree_dup_cursor(
 		new->bc_levels[i].ra = cur->bc_levels[i].ra;
 		bp = cur->bc_levels[i].bp;
 		if (bp) {
-			error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp,
-						   xfs_buf_daddr(bp), mp->m_bsize,
-						   0, &bp,
-						   cur->bc_ops->buf_ops);
+			error = xfs_trans_read_buf(mp, tp,
+					xfs_btree_buftarg(cur),
+					xfs_buf_daddr(bp),
+					xfs_btree_bbsize(cur), 0, &bp,
+					cur->bc_ops->buf_ops);
+			if (xfs_metadata_is_sick(error))
+				xfs_btree_mark_sick(new);
 			if (error) {
 				xfs_btree_del_cursor(new, error);
 				*ncur = NULL;
@@ -539,7 +651,7 @@ xfs_btree_dup_cursor(
  * record, key or pointer (xfs_btree_*_addr).  Note that all addressing
  * inside the btree block is done using indices starting at one, not zero!
  *
- * If XFS_BTREE_OVERLAPPING is set, then this btree supports keys containing
+ * If XFS_BTGEO_OVERLAPPING is set, then this btree supports keys containing
  * overlapping intervals.  In such a tree, records are still sorted lowest to
  * highest and indexed by the smallest key value that refers to the record.
  * However, nodes are different: each pointer has two associated keys -- one
@@ -589,26 +701,17 @@ xfs_btree_dup_cursor(
  */
 static inline size_t xfs_btree_block_len(struct xfs_btree_cur *cur)
 {
-	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
-		if (cur->bc_flags & XFS_BTREE_CRC_BLOCKS)
+	if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) {
+		if (xfs_has_crc(cur->bc_mp))
 			return XFS_BTREE_LBLOCK_CRC_LEN;
 		return XFS_BTREE_LBLOCK_LEN;
 	}
-	if (cur->bc_flags & XFS_BTREE_CRC_BLOCKS)
+	if (xfs_has_crc(cur->bc_mp))
 		return XFS_BTREE_SBLOCK_CRC_LEN;
 	return XFS_BTREE_SBLOCK_LEN;
 }
 
 /*
- * Return size of btree block pointers for this btree instance.
- */
-static inline size_t xfs_btree_ptr_len(struct xfs_btree_cur *cur)
-{
-	return (cur->bc_flags & XFS_BTREE_LONG_PTRS) ?
-		sizeof(__be64) : sizeof(__be32);
-}
-
-/*
  * Calculate offset of the n-th record in a btree block.
  */
 STATIC size_t
@@ -655,7 +758,7 @@ xfs_btree_ptr_offset(
 {
 	return xfs_btree_block_len(cur) +
 		cur->bc_ops->get_maxrecs(cur, level) * cur->bc_ops->key_len +
-		(n - 1) * xfs_btree_ptr_len(cur);
+		(n - 1) * cur->bc_ops->ptr_len;
 }
 
 /*
@@ -718,7 +821,7 @@ struct xfs_ifork *
 xfs_btree_ifork_ptr(
 	struct xfs_btree_cur	*cur)
 {
-	ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE);
+	ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_INODE);
 
 	if (cur->bc_flags & XFS_BTREE_STAGING)
 		return cur->bc_ino.ifake->if_fork;
@@ -750,8 +853,7 @@ xfs_btree_get_block(
 	int			level,	/* level in btree */
 	struct xfs_buf		**bpp)	/* buffer containing the block */
 {
-	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
-	    (level == cur->bc_nlevels - 1)) {
+	if (xfs_btree_at_iroot(cur, level)) {
 		*bpp = NULL;
 		return xfs_btree_get_iroot(cur);
 	}
@@ -856,95 +958,52 @@ xfs_btree_offsets(
 	}
 }
 
-/*
- * Get a buffer for the block, return it read in.
- * Long-form addressing.
- */
-int
-xfs_btree_read_bufl(
-	struct xfs_mount	*mp,		/* file system mount point */
-	struct xfs_trans	*tp,		/* transaction pointer */
-	xfs_fsblock_t		fsbno,		/* file system block number */
-	struct xfs_buf		**bpp,		/* buffer for fsbno */
-	int			refval,		/* ref count value for buffer */
-	const struct xfs_buf_ops *ops)
-{
-	struct xfs_buf		*bp;		/* return value */
-	xfs_daddr_t		d;		/* real disk block address */
-	int			error;
-
-	if (!xfs_verify_fsbno(mp, fsbno))
-		return -EFSCORRUPTED;
-	d = XFS_FSB_TO_DADDR(mp, fsbno);
-	error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, d,
-				   mp->m_bsize, 0, &bp, ops);
-	if (error)
-		return error;
-	if (bp)
-		xfs_buf_set_ref(bp, refval);
-	*bpp = bp;
-	return 0;
-}
-
-/*
- * Read-ahead the block, don't wait for it, don't return a buffer.
- * Long-form addressing.
- */
-/* ARGSUSED */
-void
-xfs_btree_reada_bufl(
-	struct xfs_mount	*mp,		/* file system mount point */
-	xfs_fsblock_t		fsbno,		/* file system block number */
-	xfs_extlen_t		count,		/* count of filesystem blocks */
-	const struct xfs_buf_ops *ops)
+STATIC int
+xfs_btree_readahead_fsblock(
+	struct xfs_btree_cur	*cur,
+	int			lr,
+	struct xfs_btree_block	*block)
 {
-	xfs_daddr_t		d;
+	struct xfs_mount	*mp = cur->bc_mp;
+	xfs_fsblock_t		left = be64_to_cpu(block->bb_u.l.bb_leftsib);
+	xfs_fsblock_t		right = be64_to_cpu(block->bb_u.l.bb_rightsib);
+	int			rval = 0;
 
-	ASSERT(fsbno != NULLFSBLOCK);
-	d = XFS_FSB_TO_DADDR(mp, fsbno);
-	xfs_buf_readahead(mp->m_ddev_targp, d, mp->m_bsize * count, ops);
-}
+	if ((lr & XFS_BTCUR_LEFTRA) && left != NULLFSBLOCK) {
+		xfs_buf_readahead(mp->m_ddev_targp, XFS_FSB_TO_DADDR(mp, left),
+				mp->m_bsize, cur->bc_ops->buf_ops);
+		rval++;
+	}
 
-/*
- * Read-ahead the block, don't wait for it, don't return a buffer.
- * Short-form addressing.
- */
-/* ARGSUSED */
-void
-xfs_btree_reada_bufs(
-	struct xfs_mount	*mp,		/* file system mount point */
-	xfs_agnumber_t		agno,		/* allocation group number */
-	xfs_agblock_t		agbno,		/* allocation group block number */
-	xfs_extlen_t		count,		/* count of filesystem blocks */
-	const struct xfs_buf_ops *ops)
-{
-	xfs_daddr_t		d;
+	if ((lr & XFS_BTCUR_RIGHTRA) && right != NULLFSBLOCK) {
+		xfs_buf_readahead(mp->m_ddev_targp, XFS_FSB_TO_DADDR(mp, right),
+				mp->m_bsize, cur->bc_ops->buf_ops);
+		rval++;
+	}
 
-	ASSERT(agno != NULLAGNUMBER);
-	ASSERT(agbno != NULLAGBLOCK);
-	d = XFS_AGB_TO_DADDR(mp, agno, agbno);
-	xfs_buf_readahead(mp->m_ddev_targp, d, mp->m_bsize * count, ops);
+	return rval;
 }
 
 STATIC int
-xfs_btree_readahead_lblock(
+xfs_btree_readahead_memblock(
 	struct xfs_btree_cur	*cur,
 	int			lr,
 	struct xfs_btree_block	*block)
 {
+	struct xfs_buftarg	*btp = cur->bc_mem.xfbtree->target;
+	xfbno_t			left = be64_to_cpu(block->bb_u.l.bb_leftsib);
+	xfbno_t			right = be64_to_cpu(block->bb_u.l.bb_rightsib);
 	int			rval = 0;
-	xfs_fsblock_t		left = be64_to_cpu(block->bb_u.l.bb_leftsib);
-	xfs_fsblock_t		right = be64_to_cpu(block->bb_u.l.bb_rightsib);
 
 	if ((lr & XFS_BTCUR_LEFTRA) && left != NULLFSBLOCK) {
-		xfs_btree_reada_bufl(cur->bc_mp, left, 1,
-				     cur->bc_ops->buf_ops);
+		xfs_buf_readahead(btp, xfbno_to_daddr(left), XFBNO_BBSIZE,
+				cur->bc_ops->buf_ops);
 		rval++;
 	}
 
 	if ((lr & XFS_BTCUR_RIGHTRA) && right != NULLFSBLOCK) {
-		xfs_btree_reada_bufl(cur->bc_mp, right, 1,
-				     cur->bc_ops->buf_ops);
+		xfs_buf_readahead(btp, xfbno_to_daddr(right), XFBNO_BBSIZE,
+				cur->bc_ops->buf_ops);
 		rval++;
 	}
 
@@ -952,25 +1011,28 @@ xfs_btree_readahead_lblock(
 }
 
 STATIC int
-xfs_btree_readahead_sblock(
+xfs_btree_readahead_agblock(
 	struct xfs_btree_cur	*cur,
 	int			lr,
-	struct xfs_btree_block *block)
+	struct xfs_btree_block	*block)
 {
-	int			rval = 0;
+	struct xfs_mount	*mp = cur->bc_mp;
+	xfs_agnumber_t		agno = cur->bc_ag.pag->pag_agno;
 	xfs_agblock_t		left = be32_to_cpu(block->bb_u.s.bb_leftsib);
 	xfs_agblock_t		right = be32_to_cpu(block->bb_u.s.bb_rightsib);
-
+	int			rval = 0;
 
 	if ((lr & XFS_BTCUR_LEFTRA) && left != NULLAGBLOCK) {
-		xfs_btree_reada_bufs(cur->bc_mp, cur->bc_ag.pag->pag_agno,
-				     left, 1, cur->bc_ops->buf_ops);
+		xfs_buf_readahead(mp->m_ddev_targp,
+				XFS_AGB_TO_DADDR(mp, agno, left),
+				mp->m_bsize, cur->bc_ops->buf_ops);
 		rval++;
 	}
 
 	if ((lr & XFS_BTCUR_RIGHTRA) && right != NULLAGBLOCK) {
-		xfs_btree_reada_bufs(cur->bc_mp, cur->bc_ag.pag->pag_agno,
-				     right, 1, cur->bc_ops->buf_ops);
+		xfs_buf_readahead(mp->m_ddev_targp,
+				XFS_AGB_TO_DADDR(mp, agno, right),
+				mp->m_bsize, cur->bc_ops->buf_ops);
 		rval++;
 	}
 
@@ -993,8 +1055,7 @@ xfs_btree_readahead(
 	 * No readahead needed if we are at the root level and the
 	 * btree root is stored in the inode.
 	 */
-	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
-	    (lev == cur->bc_nlevels - 1))
+	if (xfs_btree_at_iroot(cur, lev))
 		return 0;
 
 	if ((cur->bc_levels[lev].ra | lr) == cur->bc_levels[lev].ra)
@@ -1003,9 +1064,17 @@ xfs_btree_readahead(
 	cur->bc_levels[lev].ra |= lr;
 	block = XFS_BUF_TO_BLOCK(cur->bc_levels[lev].bp);
 
-	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
-		return xfs_btree_readahead_lblock(cur, lr, block);
-	return xfs_btree_readahead_sblock(cur, lr, block);
+	switch (cur->bc_ops->type) {
+	case XFS_BTREE_TYPE_AG:
+		return xfs_btree_readahead_agblock(cur, lr, block);
+	case XFS_BTREE_TYPE_INODE:
+		return xfs_btree_readahead_fsblock(cur, lr, block);
+	case XFS_BTREE_TYPE_MEM:
+		return xfs_btree_readahead_memblock(cur, lr, block);
+	default:
+		ASSERT(0);
+		return 0;
+	}
 }
 
 STATIC int
@@ -1014,23 +1083,24 @@ xfs_btree_ptr_to_daddr(
 	const union xfs_btree_ptr	*ptr,
 	xfs_daddr_t			*daddr)
 {
-	xfs_fsblock_t		fsbno;
-	xfs_agblock_t		agbno;
 	int			error;
 
 	error = xfs_btree_check_ptr(cur, ptr, 0, 1);
 	if (error)
 		return error;
 
-	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
-		fsbno = be64_to_cpu(ptr->l);
-		*daddr = XFS_FSB_TO_DADDR(cur->bc_mp, fsbno);
-	} else {
-		agbno = be32_to_cpu(ptr->s);
+	switch (cur->bc_ops->type) {
+	case XFS_BTREE_TYPE_AG:
 		*daddr = XFS_AGB_TO_DADDR(cur->bc_mp, cur->bc_ag.pag->pag_agno,
-				agbno);
+				be32_to_cpu(ptr->s));
+		break;
+	case XFS_BTREE_TYPE_INODE:
+		*daddr = XFS_FSB_TO_DADDR(cur->bc_mp, be64_to_cpu(ptr->l));
+		break;
+	case XFS_BTREE_TYPE_MEM:
+		*daddr = xfbno_to_daddr(be64_to_cpu(ptr->l));
+		break;
 	}
-
 	return 0;
 }
 
@@ -1050,8 +1120,9 @@ xfs_btree_readahead_ptr(
 
 	if (xfs_btree_ptr_to_daddr(cur, ptr, &daddr))
 		return;
-	xfs_buf_readahead(cur->bc_mp->m_ddev_targp, daddr,
-			  cur->bc_mp->m_bsize * count, cur->bc_ops->buf_ops);
+	xfs_buf_readahead(xfs_btree_buftarg(cur), daddr,
+			xfs_btree_bbsize(cur) * count,
+			cur->bc_ops->buf_ops);
 }
 
 /*
@@ -1072,7 +1143,7 @@ xfs_btree_setbuf(
 	cur->bc_levels[lev].ra = 0;
 
 	b = XFS_BUF_TO_BLOCK(bp);
-	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+	if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) {
 		if (b->bb_u.l.bb_leftsib == cpu_to_be64(NULLFSBLOCK))
 			cur->bc_levels[lev].ra |= XFS_BTCUR_LEFTRA;
 		if (b->bb_u.l.bb_rightsib == cpu_to_be64(NULLFSBLOCK))
@@ -1090,7 +1161,7 @@ xfs_btree_ptr_is_null(
 	struct xfs_btree_cur		*cur,
 	const union xfs_btree_ptr	*ptr)
 {
-	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+	if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN)
 		return ptr->l == cpu_to_be64(NULLFSBLOCK);
 	else
 		return ptr->s == cpu_to_be32(NULLAGBLOCK);
@@ -1101,12 +1172,23 @@ xfs_btree_set_ptr_null(
 	struct xfs_btree_cur	*cur,
 	union xfs_btree_ptr	*ptr)
 {
-	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+	if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN)
 		ptr->l = cpu_to_be64(NULLFSBLOCK);
 	else
 		ptr->s = cpu_to_be32(NULLAGBLOCK);
 }
 
+static inline bool
+xfs_btree_ptrs_equal(
+	struct xfs_btree_cur		*cur,
+	union xfs_btree_ptr		*ptr1,
+	union xfs_btree_ptr		*ptr2)
+{
+	if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN)
+		return ptr1->l == ptr2->l;
+	return ptr1->s == ptr2->s;
+}
+
 /*
  * Get/set/init sibling pointers
  */
@@ -1119,7 +1201,7 @@ xfs_btree_get_sibling(
 {
 	ASSERT(lr == XFS_BB_LEFTSIB || lr == XFS_BB_RIGHTSIB);
 
-	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+	if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) {
 		if (lr == XFS_BB_RIGHTSIB)
 			ptr->l = block->bb_u.l.bb_rightsib;
 		else
@@ -1141,7 +1223,7 @@ xfs_btree_set_sibling(
 {
 	ASSERT(lr == XFS_BB_LEFTSIB || lr == XFS_BB_RIGHTSIB);
 
-	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+	if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) {
 		if (lr == XFS_BB_RIGHTSIB)
 			block->bb_u.l.bb_rightsib = ptr->l;
 		else
@@ -1154,25 +1236,24 @@ xfs_btree_set_sibling(
 	}
 }
 
-void
-xfs_btree_init_block_int(
+static void
+__xfs_btree_init_block(
 	struct xfs_mount	*mp,
 	struct xfs_btree_block	*buf,
+	const struct xfs_btree_ops *ops,
 	xfs_daddr_t		blkno,
-	xfs_btnum_t		btnum,
 	__u16			level,
 	__u16			numrecs,
-	__u64			owner,
-	unsigned int		flags)
+	__u64			owner)
 {
-	int			crc = xfs_has_crc(mp);
-	__u32			magic = xfs_btree_magic(crc, btnum);
+	bool			crc = xfs_has_crc(mp);
+	__u32			magic = xfs_btree_magic(mp, ops);
 
 	buf->bb_magic = cpu_to_be32(magic);
 	buf->bb_level = cpu_to_be16(level);
 	buf->bb_numrecs = cpu_to_be16(numrecs);
 
-	if (flags & XFS_BTREE_LONG_PTRS) {
+	if (ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) {
 		buf->bb_u.l.bb_leftsib = cpu_to_be64(NULLFSBLOCK);
 		buf->bb_u.l.bb_rightsib = cpu_to_be64(NULLFSBLOCK);
 		if (crc) {
@@ -1183,14 +1264,12 @@ xfs_btree_init_block_int(
 			buf->bb_u.l.bb_lsn = 0;
 		}
 	} else {
-		/* owner is a 32 bit value on short blocks */
-		__u32 __owner = (__u32)owner;
-
 		buf->bb_u.s.bb_leftsib = cpu_to_be32(NULLAGBLOCK);
 		buf->bb_u.s.bb_rightsib = cpu_to_be32(NULLAGBLOCK);
 		if (crc) {
 			buf->bb_u.s.bb_blkno = cpu_to_be64(blkno);
-			buf->bb_u.s.bb_owner = cpu_to_be32(__owner);
+			/* owner is a 32 bit value on short blocks */
+			buf->bb_u.s.bb_owner = cpu_to_be32((__u32)owner);
 			uuid_copy(&buf->bb_u.s.bb_uuid, &mp->m_sb.sb_meta_uuid);
 			buf->bb_u.s.bb_lsn = 0;
 		}
@@ -1199,15 +1278,46 @@ xfs_btree_init_block_int(
 
 void
 xfs_btree_init_block(
-	struct xfs_mount *mp,
-	struct xfs_buf	*bp,
-	xfs_btnum_t	btnum,
-	__u16		level,
-	__u16		numrecs,
-	__u64		owner)
+	struct xfs_mount	*mp,
+	struct xfs_btree_block	*block,
+	const struct xfs_btree_ops *ops,
+	__u16			level,
+	__u16			numrecs,
+	__u64			owner)
+{
+	__xfs_btree_init_block(mp, block, ops, XFS_BUF_DADDR_NULL, level,
+			numrecs, owner);
+}
+
+void
+xfs_btree_init_buf(
+	struct xfs_mount		*mp,
+	struct xfs_buf			*bp,
+	const struct xfs_btree_ops	*ops,
+	__u16				level,
+	__u16				numrecs,
+	__u64				owner)
+{
+	__xfs_btree_init_block(mp, XFS_BUF_TO_BLOCK(bp), ops,
+			xfs_buf_daddr(bp), level, numrecs, owner);
+	bp->b_ops = ops->buf_ops;
+}
+
+static inline __u64
+xfs_btree_owner(
+	struct xfs_btree_cur    *cur)
 {
-	xfs_btree_init_block_int(mp, XFS_BUF_TO_BLOCK(bp), xfs_buf_daddr(bp),
-				 btnum, level, numrecs, owner, 0);
+	switch (cur->bc_ops->type) {
+	case XFS_BTREE_TYPE_MEM:
+		return cur->bc_mem.xfbtree->owner;
+	case XFS_BTREE_TYPE_INODE:
+		return cur->bc_ino.ip->i_ino;
+	case XFS_BTREE_TYPE_AG:
+		return cur->bc_ag.pag->pag_agno;
+	default:
+		ASSERT(0);
+		return 0;
+	}
 }
 
 void
@@ -1217,22 +1327,8 @@ xfs_btree_init_block_cur(
 	int			level,
 	int			numrecs)
 {
-	__u64			owner;
-
-	/*
-	 * we can pull the owner from the cursor right now as the different
-	 * owners align directly with the pointer size of the btree. This may
-	 * change in future, but is safe for current users of the generic btree
-	 * code.
-	 */
-	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
-		owner = cur->bc_ino.ip->i_ino;
-	else
-		owner = cur->bc_ag.pag->pag_agno;
-
-	xfs_btree_init_block_int(cur->bc_mp, XFS_BUF_TO_BLOCK(bp),
-				xfs_buf_daddr(bp), cur->bc_btnum, level,
-				numrecs, owner, cur->bc_flags);
+	xfs_btree_init_buf(cur->bc_mp, bp, cur->bc_ops, level, numrecs,
+			xfs_btree_owner(cur));
 }
 
 /*
@@ -1250,7 +1346,7 @@ xfs_btree_is_lastrec(
 
 	if (level > 0)
 		return 0;
-	if (!(cur->bc_flags & XFS_BTREE_LASTREC_UPDATE))
+	if (!(cur->bc_ops->geom_flags & XFS_BTGEO_LASTREC_UPDATE))
 		return 0;
 
 	xfs_btree_get_sibling(cur, block, &ptr, XFS_BB_RIGHTSIB);
@@ -1265,41 +1361,27 @@ xfs_btree_buf_to_ptr(
 	struct xfs_buf		*bp,
 	union xfs_btree_ptr	*ptr)
 {
-	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
-		ptr->l = cpu_to_be64(XFS_DADDR_TO_FSB(cur->bc_mp,
-					xfs_buf_daddr(bp)));
-	else {
+	switch (cur->bc_ops->type) {
+	case XFS_BTREE_TYPE_AG:
 		ptr->s = cpu_to_be32(xfs_daddr_to_agbno(cur->bc_mp,
 					xfs_buf_daddr(bp)));
+		break;
+	case XFS_BTREE_TYPE_INODE:
+		ptr->l = cpu_to_be64(XFS_DADDR_TO_FSB(cur->bc_mp,
+					xfs_buf_daddr(bp)));
+		break;
+	case XFS_BTREE_TYPE_MEM:
+		ptr->l = cpu_to_be64(xfs_daddr_to_xfbno(xfs_buf_daddr(bp)));
+		break;
 	}
 }
 
-STATIC void
+static inline void
 xfs_btree_set_refs(
 	struct xfs_btree_cur	*cur,
 	struct xfs_buf		*bp)
 {
-	switch (cur->bc_btnum) {
-	case XFS_BTNUM_BNO:
-	case XFS_BTNUM_CNT:
-		xfs_buf_set_ref(bp, XFS_ALLOC_BTREE_REF);
-		break;
-	case XFS_BTNUM_INO:
-	case XFS_BTNUM_FINO:
-		xfs_buf_set_ref(bp, XFS_INO_BTREE_REF);
-		break;
-	case XFS_BTNUM_BMAP:
-		xfs_buf_set_ref(bp, XFS_BMAP_BTREE_REF);
-		break;
-	case XFS_BTNUM_RMAP:
-		xfs_buf_set_ref(bp, XFS_RMAP_BTREE_REF);
-		break;
-	case XFS_BTNUM_REFC:
-		xfs_buf_set_ref(bp, XFS_REFC_BTREE_REF);
-		break;
-	default:
-		ASSERT(0);
-	}
+	xfs_buf_set_ref(bp, cur->bc_ops->lru_refs);
 }
 
 int
@@ -1309,15 +1391,14 @@ xfs_btree_get_buf_block(
 	struct xfs_btree_block		**block,
 	struct xfs_buf			**bpp)
 {
-	struct xfs_mount	*mp = cur->bc_mp;
-	xfs_daddr_t		d;
-	int			error;
+	xfs_daddr_t			d;
+	int				error;
 
 	error = xfs_btree_ptr_to_daddr(cur, ptr, &d);
 	if (error)
 		return error;
-	error = xfs_trans_get_buf(cur->bc_tp, mp->m_ddev_targp, d, mp->m_bsize,
-			0, bpp);
+	error = xfs_trans_get_buf(cur->bc_tp, xfs_btree_buftarg(cur), d,
+			xfs_btree_bbsize(cur), 0, bpp);
 	if (error)
 		return error;
 
@@ -1348,9 +1429,11 @@ xfs_btree_read_buf_block(
 	error = xfs_btree_ptr_to_daddr(cur, ptr, &d);
 	if (error)
 		return error;
-	error = xfs_trans_read_buf(mp, cur->bc_tp, mp->m_ddev_targp, d,
-				   mp->m_bsize, flags, bpp,
-				   cur->bc_ops->buf_ops);
+	error = xfs_trans_read_buf(mp, cur->bc_tp, xfs_btree_buftarg(cur), d,
+			xfs_btree_bbsize(cur), flags, bpp,
+			cur->bc_ops->buf_ops);
+	if (xfs_metadata_is_sick(error))
+		xfs_btree_mark_sick(cur);
 	if (error)
 		return error;
 
@@ -1398,7 +1481,7 @@ xfs_btree_copy_ptrs(
 	int			numptrs)
 {
 	ASSERT(numptrs >= 0);
-	memcpy(dst_ptr, src_ptr, numptrs * xfs_btree_ptr_len(cur));
+	memcpy(dst_ptr, src_ptr, numptrs * cur->bc_ops->ptr_len);
 }
 
 /*
@@ -1454,8 +1537,8 @@ xfs_btree_shift_ptrs(
 	ASSERT(numptrs >= 0);
 	ASSERT(dir == 1 || dir == -1);
 
-	dst_ptr = (char *)ptr + (dir * xfs_btree_ptr_len(cur));
-	memmove(dst_ptr, ptr, numptrs * xfs_btree_ptr_len(cur));
+	dst_ptr = (char *)ptr + (dir * cur->bc_ops->ptr_len);
+	memmove(dst_ptr, ptr, numptrs * cur->bc_ops->ptr_len);
 }
 
 /*
@@ -1566,7 +1649,7 @@ xfs_btree_log_block(
 	if (bp) {
 		int nbits;
 
-		if (cur->bc_flags & XFS_BTREE_CRC_BLOCKS) {
+		if (xfs_has_crc(cur->bc_mp)) {
 			/*
 			 * We don't log the CRC when updating a btree
 			 * block but instead recreate it during log
@@ -1581,7 +1664,7 @@ xfs_btree_log_block(
 			nbits = XFS_BB_NUM_BITS;
 		}
 		xfs_btree_offsets(fields,
-				  (cur->bc_flags & XFS_BTREE_LONG_PTRS) ?
+				  (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) ?
 					loffsets : soffsets,
 				  nbits, &first, &last);
 		xfs_trans_buf_set_type(cur->bc_tp, bp, XFS_BLFT_BTREE_BUF);
@@ -1658,9 +1741,10 @@ xfs_btree_increment(
 	 * confused or have the tree root in an inode.
 	 */
 	if (lev == cur->bc_nlevels) {
-		if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE)
+		if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE)
 			goto out0;
 		ASSERT(0);
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto error0;
 	}
@@ -1751,9 +1835,10 @@ xfs_btree_decrement(
 	 * or the root of the tree is in an inode.
 	 */
 	if (lev == cur->bc_nlevels) {
-		if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE)
+		if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE)
 			goto out0;
 		ASSERT(0);
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto error0;
 	}
@@ -1786,6 +1871,33 @@ error0:
 	return error;
 }
 
+/*
+ * Check the btree block owner now that we have the context to know who the
+ * real owner is.
+ */
+static inline xfs_failaddr_t
+xfs_btree_check_block_owner(
+	struct xfs_btree_cur	*cur,
+	struct xfs_btree_block	*block)
+{
+	__u64			owner;
+
+	if (!xfs_has_crc(cur->bc_mp) ||
+	    (cur->bc_flags & XFS_BTREE_BMBT_INVALID_OWNER))
+		return NULL;
+
+	owner = xfs_btree_owner(cur);
+	if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) {
+		if (be64_to_cpu(block->bb_u.l.bb_owner) != owner)
+			return __this_address;
+	} else {
+		if (be32_to_cpu(block->bb_u.s.bb_owner) != owner)
+			return __this_address;
+	}
+
+	return NULL;
+}
+
 int
 xfs_btree_lookup_get_block(
 	struct xfs_btree_cur		*cur,	/* btree cursor */
@@ -1798,8 +1910,7 @@ xfs_btree_lookup_get_block(
 	int			error = 0;
 
 	/* special case the root block if in an inode */
-	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
-	    (level == cur->bc_nlevels - 1)) {
+	if (xfs_btree_at_iroot(cur, level)) {
 		*blkp = xfs_btree_get_iroot(cur);
 		return 0;
 	}
@@ -1824,11 +1935,7 @@ xfs_btree_lookup_get_block(
 		return error;
 
 	/* Check the inode owner since the verifiers don't. */
-	if (xfs_has_crc(cur->bc_mp) &&
-	    !(cur->bc_ino.flags & XFS_BTCUR_BMBT_INVALID_OWNER) &&
-	    (cur->bc_flags & XFS_BTREE_LONG_PTRS) &&
-	    be64_to_cpu((*blkp)->bb_u.l.bb_owner) !=
-			cur->bc_ino.ip->i_ino)
+	if (xfs_btree_check_block_owner(cur, *blkp) != NULL)
 		goto out_bad;
 
 	/* Did we get the level we were looking for? */
@@ -1846,6 +1953,7 @@ out_bad:
 	*blkp = NULL;
 	xfs_buf_mark_corrupt(bp);
 	xfs_trans_brelse(cur->bc_tp, bp);
+	xfs_btree_mark_sick(cur);
 	return -EFSCORRUPTED;
 }
 
@@ -1872,6 +1980,27 @@ xfs_lookup_get_search_key(
 }
 
 /*
+ * Initialize a pointer to the root block.
+ */
+void
+xfs_btree_init_ptr_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr)
+{
+	if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE) {
+		/*
+		 * Inode-rooted btrees call xfs_btree_get_iroot to find the root
+		 * in xfs_btree_lookup_get_block and don't need a pointer here.
+		 */
+		ptr->l = 0;
+	} else if (cur->bc_flags & XFS_BTREE_STAGING) {
+		ptr->s = cpu_to_be32(cur->bc_ag.afake->af_root);
+	} else {
+		cur->bc_ops->init_ptr_from_cur(cur, ptr);
+	}
+}
+
+/*
  * Lookup the record.  The cursor is made to point to it, based on dir.
  * stat is set to 0 if can't find any such record, 1 for success.
  */
@@ -1892,14 +2021,16 @@ xfs_btree_lookup(
 	XFS_BTREE_STATS_INC(cur, lookup);
 
 	/* No such thing as a zero-level tree. */
-	if (XFS_IS_CORRUPT(cur->bc_mp, cur->bc_nlevels == 0))
+	if (XFS_IS_CORRUPT(cur->bc_mp, cur->bc_nlevels == 0)) {
+		xfs_btree_mark_sick(cur);
 		return -EFSCORRUPTED;
+	}
 
 	block = NULL;
 	keyno = 0;
 
 	/* initialise start pointer from cursor */
-	cur->bc_ops->init_ptr_from_cur(cur, &ptr);
+	xfs_btree_init_ptr_from_cur(cur, &ptr);
 	pp = &ptr;
 
 	/*
@@ -1936,6 +2067,7 @@ xfs_btree_lookup(
 							XFS_ERRLEVEL_LOW,
 							cur->bc_mp, block,
 							sizeof(*block));
+					xfs_btree_mark_sick(cur);
 					return -EFSCORRUPTED;
 				}
 
@@ -2012,8 +2144,10 @@ xfs_btree_lookup(
 			error = xfs_btree_increment(cur, 0, &i);
 			if (error)
 				goto error0;
-			if (XFS_IS_CORRUPT(cur->bc_mp, i != 1))
+			if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				return -EFSCORRUPTED;
+			}
 			*stat = 1;
 			return 0;
 		}
@@ -2040,7 +2174,7 @@ xfs_btree_high_key_from_key(
 	struct xfs_btree_cur	*cur,
 	union xfs_btree_key	*key)
 {
-	ASSERT(cur->bc_flags & XFS_BTREE_OVERLAPPING);
+	ASSERT(cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING);
 	return (union xfs_btree_key *)((char *)key +
 			(cur->bc_ops->key_len / 2));
 }
@@ -2061,7 +2195,7 @@ xfs_btree_get_leaf_keys(
 	rec = xfs_btree_rec_addr(cur, 1, block);
 	cur->bc_ops->init_key_from_rec(key, rec);
 
-	if (cur->bc_flags & XFS_BTREE_OVERLAPPING) {
+	if (cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING) {
 
 		cur->bc_ops->init_high_key_from_rec(&max_hkey, rec);
 		for (n = 2; n <= xfs_btree_get_numrecs(block); n++) {
@@ -2088,7 +2222,7 @@ xfs_btree_get_node_keys(
 	union xfs_btree_key	*high;
 	int			n;
 
-	if (cur->bc_flags & XFS_BTREE_OVERLAPPING) {
+	if (cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING) {
 		memcpy(key, xfs_btree_key_addr(cur, 1, block),
 				cur->bc_ops->key_len / 2);
 
@@ -2132,7 +2266,7 @@ xfs_btree_needs_key_update(
 	struct xfs_btree_cur	*cur,
 	int			ptr)
 {
-	return (cur->bc_flags & XFS_BTREE_OVERLAPPING) || ptr == 1;
+	return (cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING) || ptr == 1;
 }
 
 /*
@@ -2156,7 +2290,7 @@ __xfs_btree_updkeys(
 	struct xfs_buf		*bp;
 	int			ptr;
 
-	ASSERT(cur->bc_flags & XFS_BTREE_OVERLAPPING);
+	ASSERT(cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING);
 
 	/* Exit if there aren't any parent levels to update. */
 	if (level + 1 >= cur->bc_nlevels)
@@ -2225,7 +2359,7 @@ xfs_btree_update_keys(
 	ASSERT(level >= 0);
 
 	block = xfs_btree_get_block(cur, level, &bp);
-	if (cur->bc_flags & XFS_BTREE_OVERLAPPING)
+	if (cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING)
 		return __xfs_btree_updkeys(cur, level, block, bp, false);
 
 	/*
@@ -2332,8 +2466,7 @@ xfs_btree_lshift(
 	int			error;		/* error return value */
 	int			i;
 
-	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
-	    level == cur->bc_nlevels - 1)
+	if (xfs_btree_at_iroot(cur, level))
 		goto out0;
 
 	/* Set up variables for this block as "right". */
@@ -2460,12 +2593,13 @@ xfs_btree_lshift(
 	 * Using a temporary cursor, update the parent key values of the
 	 * block on the left.
 	 */
-	if (cur->bc_flags & XFS_BTREE_OVERLAPPING) {
+	if (cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING) {
 		error = xfs_btree_dup_cursor(cur, &tcur);
 		if (error)
 			goto error0;
 		i = xfs_btree_firstrec(tcur, level);
 		if (XFS_IS_CORRUPT(tcur->bc_mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -2527,8 +2661,7 @@ xfs_btree_rshift(
 	int			error;		/* error return value */
 	int			i;		/* loop counter */
 
-	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
-	    (level == cur->bc_nlevels - 1))
+	if (xfs_btree_at_iroot(cur, level))
 		goto out0;
 
 	/* Set up variables for this block as "left". */
@@ -2636,6 +2769,7 @@ xfs_btree_rshift(
 		goto error0;
 	i = xfs_btree_lastrec(tcur, level);
 	if (XFS_IS_CORRUPT(tcur->bc_mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto error0;
 	}
@@ -2645,7 +2779,7 @@ xfs_btree_rshift(
 		goto error1;
 
 	/* Update the parent high keys of the left block, if needed. */
-	if (cur->bc_flags & XFS_BTREE_OVERLAPPING) {
+	if (cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING) {
 		error = xfs_btree_update_keys(cur, level);
 		if (error)
 			goto error1;
@@ -2673,6 +2807,32 @@ error1:
 	return error;
 }
 
+static inline int
+xfs_btree_alloc_block(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_ptr	*hint_block,
+	union xfs_btree_ptr		*new_block,
+	int				*stat)
+{
+	int				error;
+
+	/*
+	 * Don't allow block allocation for a staging cursor, because staging
+	 * cursors do not support regular btree modifications.
+	 *
+	 * Bulk loading uses a separate callback to obtain new blocks from a
+	 * preallocated list, which prevents ENOSPC failures during loading.
+	 */
+	if (unlikely(cur->bc_flags & XFS_BTREE_STAGING)) {
+		ASSERT(0);
+		return -EFSCORRUPTED;
+	}
+
+	error = cur->bc_ops->alloc_block(cur, hint_block, new_block, stat);
+	trace_xfs_btree_alloc_block(cur, new_block, *stat, error);
+	return error;
+}
+
 /*
  * Split cur/level block in half.
  * Return new block number and the key to its first
@@ -2716,7 +2876,7 @@ __xfs_btree_split(
 	xfs_btree_buf_to_ptr(cur, lbp, &lptr);
 
 	/* Allocate the new block. If we can't do it, we're toast. Give up. */
-	error = cur->bc_ops->alloc_block(cur, &lptr, &rptr, stat);
+	error = xfs_btree_alloc_block(cur, &lptr, &rptr, stat);
 	if (error)
 		goto error0;
 	if (*stat == 0)
@@ -2823,7 +2983,7 @@ __xfs_btree_split(
 	}
 
 	/* Update the parent high keys of the left block, if needed. */
-	if (cur->bc_flags & XFS_BTREE_OVERLAPPING) {
+	if (cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING) {
 		error = xfs_btree_update_keys(cur, level);
 		if (error)
 			goto error0;
@@ -2941,7 +3101,7 @@ xfs_btree_split(
 	struct xfs_btree_split_args	args;
 	DECLARE_COMPLETION_ONSTACK(done);
 
-	if (cur->bc_btnum != XFS_BTNUM_BMAP ||
+	if (!xfs_btree_is_bmap(cur->bc_ops) ||
 	    cur->bc_tp->t_highest_agno == NULLAGNUMBER)
 		return __xfs_btree_split(cur, level, ptrp, key, curp, stat);
 
@@ -2963,7 +3123,6 @@ xfs_btree_split(
 #define xfs_btree_split	__xfs_btree_split
 #endif /* __KERNEL__ */
 
-
 /*
  * Copy the old inode root contents into a real block and make the
  * broot point to it.
@@ -2988,7 +3147,7 @@ xfs_btree_new_iroot(
 
 	XFS_BTREE_STATS_INC(cur, newroot);
 
-	ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE);
+	ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_INODE);
 
 	level = cur->bc_nlevels - 1;
 
@@ -2996,7 +3155,7 @@ xfs_btree_new_iroot(
 	pp = xfs_btree_ptr_addr(cur, 1, block);
 
 	/* Allocate the new block. If we can't do it, we're toast. Give up. */
-	error = cur->bc_ops->alloc_block(cur, pp, &nptr, stat);
+	error = xfs_btree_alloc_block(cur, pp, &nptr, stat);
 	if (error)
 		goto error0;
 	if (*stat == 0)
@@ -3014,9 +3173,9 @@ xfs_btree_new_iroot(
 	 * In that case have to also ensure the blkno remains correct
 	 */
 	memcpy(cblock, block, xfs_btree_block_len(cur));
-	if (cur->bc_flags & XFS_BTREE_CRC_BLOCKS) {
+	if (xfs_has_crc(cur->bc_mp)) {
 		__be64 bno = cpu_to_be64(xfs_buf_daddr(cbp));
-		if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+		if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN)
 			cblock->bb_u.l.bb_blkno = bno;
 		else
 			cblock->bb_u.s.bb_blkno = bno;
@@ -3069,6 +3228,21 @@ error0:
 	return error;
 }
 
+static void
+xfs_btree_set_root(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_ptr	*ptr,
+	int				inc)
+{
+	if (cur->bc_flags & XFS_BTREE_STAGING) {
+		/* Update the btree root information for a per-AG fake root. */
+		cur->bc_ag.afake->af_root = be32_to_cpu(ptr->s);
+		cur->bc_ag.afake->af_levels += inc;
+	} else {
+		cur->bc_ops->set_root(cur, ptr, inc);
+	}
+}
+
 /*
  * Allocate a new root block, fill it in.
  */
@@ -3093,10 +3267,10 @@ xfs_btree_new_root(
 	XFS_BTREE_STATS_INC(cur, newroot);
 
 	/* initialise our start point from the cursor */
-	cur->bc_ops->init_ptr_from_cur(cur, &rptr);
+	xfs_btree_init_ptr_from_cur(cur, &rptr);
 
 	/* Allocate the new block. If we can't do it, we're toast. Give up. */
-	error = cur->bc_ops->alloc_block(cur, &rptr, &lptr, stat);
+	error = xfs_btree_alloc_block(cur, &rptr, &lptr, stat);
 	if (error)
 		goto error0;
 	if (*stat == 0)
@@ -3109,7 +3283,7 @@ xfs_btree_new_root(
 		goto error0;
 
 	/* Set the root in the holding structure  increasing the level by 1. */
-	cur->bc_ops->set_root(cur, &lptr, 1);
+	xfs_btree_set_root(cur, &lptr, 1);
 
 	/*
 	 * At the previous root level there are now two blocks: the old root,
@@ -3213,8 +3387,7 @@ xfs_btree_make_block_unfull(
 {
 	int			error = 0;
 
-	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
-	    level == cur->bc_nlevels - 1) {
+	if (xfs_btree_at_iroot(cur, level)) {
 		struct xfs_inode *ip = cur->bc_ino.ip;
 
 		if (numrecs < cur->bc_ops->get_dmaxrecs(cur, level)) {
@@ -3299,8 +3472,8 @@ xfs_btree_insrec(
 	 * If we have an external root pointer, and we've made it to the
 	 * root level, allocate a new root block and we're done.
 	 */
-	if (!(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
-	    (level >= cur->bc_nlevels)) {
+	if (cur->bc_ops->type != XFS_BTREE_TYPE_INODE &&
+	    level >= cur->bc_nlevels) {
 		error = xfs_btree_new_root(cur, stat);
 		xfs_btree_set_ptr_null(cur, ptrp);
 
@@ -3524,6 +3697,7 @@ xfs_btree_insert(
 		}
 
 		if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -3537,7 +3711,8 @@ xfs_btree_insert(
 		if (pcur != cur &&
 		    (ncur || xfs_btree_ptr_is_null(cur, &nptr))) {
 			/* Save the state from the cursor before we trash it */
-			if (cur->bc_ops->update_cursor)
+			if (cur->bc_ops->update_cursor &&
+			    !(cur->bc_flags & XFS_BTREE_STAGING))
 				cur->bc_ops->update_cursor(pcur, cur);
 			cur->bc_nlevels = pcur->bc_nlevels;
 			xfs_btree_del_cursor(pcur, XFS_BTREE_NOERROR);
@@ -3586,7 +3761,7 @@ xfs_btree_kill_iroot(
 #endif
 	int			i;
 
-	ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE);
+	ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_INODE);
 	ASSERT(cur->bc_nlevels > 1);
 
 	/*
@@ -3680,7 +3855,7 @@ xfs_btree_kill_root(
 	 * Update the root pointer, decreasing the level by 1 and then
 	 * free the old root.
 	 */
-	cur->bc_ops->set_root(cur, newroot, -1);
+	xfs_btree_set_root(cur, newroot, -1);
 
 	error = xfs_btree_free_block(cur, bp);
 	if (error)
@@ -3822,27 +3997,25 @@ xfs_btree_delrec(
 	 * Try to get rid of the next level down.  If we can't then there's
 	 * nothing left to do.
 	 */
-	if (level == cur->bc_nlevels - 1) {
-		if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) {
-			xfs_iroot_realloc(cur->bc_ino.ip, -1,
-					  cur->bc_ino.whichfork);
+	if (xfs_btree_at_iroot(cur, level)) {
+		xfs_iroot_realloc(cur->bc_ino.ip, -1, cur->bc_ino.whichfork);
 
-			error = xfs_btree_kill_iroot(cur);
-			if (error)
-				goto error0;
+		error = xfs_btree_kill_iroot(cur);
+		if (error)
+			goto error0;
 
-			error = xfs_btree_dec_cursor(cur, level, stat);
-			if (error)
-				goto error0;
-			*stat = 1;
-			return 0;
-		}
+		error = xfs_btree_dec_cursor(cur, level, stat);
+		if (error)
+			goto error0;
+		*stat = 1;
+		return 0;
+	}
 
-		/*
-		 * If this is the root level, and there's only one entry left,
-		 * and it's NOT the leaf level, then we can get rid of this
-		 * level.
-		 */
+	/*
+	 * If this is the root level, and there's only one entry left, and it's
+	 * NOT the leaf level, then we can get rid of this level.
+	 */
+	if (level == cur->bc_nlevels - 1) {
 		if (numrecs == 1 && level > 0) {
 			union xfs_btree_ptr	*pp;
 			/*
@@ -3891,7 +4064,7 @@ xfs_btree_delrec(
 	xfs_btree_get_sibling(cur, block, &rptr, XFS_BB_RIGHTSIB);
 	xfs_btree_get_sibling(cur, block, &lptr, XFS_BB_LEFTSIB);
 
-	if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) {
+	if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE) {
 		/*
 		 * One child of root, need to get a chance to copy its contents
 		 * into the root and delete it. Can't go up to next level,
@@ -3931,6 +4104,7 @@ xfs_btree_delrec(
 		 */
 		i = xfs_btree_lastrec(tcur, level);
 		if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -3939,12 +4113,14 @@ xfs_btree_delrec(
 		if (error)
 			goto error0;
 		if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
 
 		i = xfs_btree_lastrec(tcur, level);
 		if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -3992,6 +4168,7 @@ xfs_btree_delrec(
 		if (!xfs_btree_ptr_is_null(cur, &lptr)) {
 			i = xfs_btree_firstrec(tcur, level);
 			if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto error0;
 			}
@@ -4000,6 +4177,7 @@ xfs_btree_delrec(
 			if (error)
 				goto error0;
 			if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto error0;
 			}
@@ -4017,6 +4195,7 @@ xfs_btree_delrec(
 		 */
 		i = xfs_btree_firstrec(tcur, level);
 		if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -4026,6 +4205,7 @@ xfs_btree_delrec(
 			goto error0;
 		i = xfs_btree_firstrec(tcur, level);
 		if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -4201,8 +4381,8 @@ xfs_btree_delrec(
 	 * If we joined with the right neighbor and there's a level above
 	 * us, increment the cursor at that level.
 	 */
-	else if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) ||
-		   (level + 1 < cur->bc_nlevels)) {
+	else if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE ||
+		 level + 1 < cur->bc_nlevels) {
 		error = xfs_btree_increment(cur, level + 1, &i);
 		if (error)
 			goto error0;
@@ -4270,7 +4450,7 @@ xfs_btree_delete(
 	 * If we combined blocks as part of deleting the record, delrec won't
 	 * have updated the parent high keys so we have to do that here.
 	 */
-	if (joined && (cur->bc_flags & XFS_BTREE_OVERLAPPING)) {
+	if (joined && (cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING)) {
 		error = xfs_btree_updkeys_force(cur, 0);
 		if (error)
 			goto error0;
@@ -4344,7 +4524,7 @@ xfs_btree_visit_block(
 {
 	struct xfs_btree_block		*block;
 	struct xfs_buf			*bp;
-	union xfs_btree_ptr		rptr;
+	union xfs_btree_ptr		rptr, bufptr;
 	int				error;
 
 	/* do right sibling readahead */
@@ -4367,15 +4547,12 @@ xfs_btree_visit_block(
 	 * return the same block without checking if the right sibling points
 	 * back to us and creates a cyclic reference in the btree.
 	 */
-	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
-		if (be64_to_cpu(rptr.l) == XFS_DADDR_TO_FSB(cur->bc_mp,
-							xfs_buf_daddr(bp)))
-			return -EFSCORRUPTED;
-	} else {
-		if (be32_to_cpu(rptr.s) == xfs_daddr_to_agbno(cur->bc_mp,
-							xfs_buf_daddr(bp)))
-			return -EFSCORRUPTED;
+	xfs_btree_buf_to_ptr(cur, bp, &bufptr);
+	if (xfs_btree_ptrs_equal(cur, &rptr, &bufptr)) {
+		xfs_btree_mark_sick(cur);
+		return -EFSCORRUPTED;
 	}
+
 	return xfs_btree_lookup_get_block(cur, level, &rptr, &block);
 }
 
@@ -4393,7 +4570,7 @@ xfs_btree_visit_blocks(
 	struct xfs_btree_block		*block = NULL;
 	int				error = 0;
 
-	cur->bc_ops->init_ptr_from_cur(cur, &lptr);
+	xfs_btree_init_ptr_from_cur(cur, &lptr);
 
 	/* for each level */
 	for (level = cur->bc_nlevels - 1; level >= 0; level--) {
@@ -4471,7 +4648,7 @@ xfs_btree_block_change_owner(
 
 	/* modify the owner */
 	block = xfs_btree_get_block(cur, level, &bp);
-	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+	if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) {
 		if (block->bb_u.l.bb_owner == cpu_to_be64(bbcoi->new_owner))
 			return 0;
 		block->bb_u.l.bb_owner = cpu_to_be64(bbcoi->new_owner);
@@ -4489,7 +4666,7 @@ xfs_btree_block_change_owner(
 	 * though, so everything is consistent in memory.
 	 */
 	if (!bp) {
-		ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE);
+		ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_INODE);
 		ASSERT(level == cur->bc_nlevels - 1);
 		return 0;
 	}
@@ -4523,7 +4700,7 @@ xfs_btree_change_owner(
 
 /* Verify the v5 fields of a long-format btree block. */
 xfs_failaddr_t
-xfs_btree_lblock_v5hdr_verify(
+xfs_btree_fsblock_v5hdr_verify(
 	struct xfs_buf		*bp,
 	uint64_t		owner)
 {
@@ -4544,7 +4721,7 @@ xfs_btree_lblock_v5hdr_verify(
 
 /* Verify a long-format btree block. */
 xfs_failaddr_t
-xfs_btree_lblock_verify(
+xfs_btree_fsblock_verify(
 	struct xfs_buf		*bp,
 	unsigned int		max_recs)
 {
@@ -4553,28 +4730,60 @@ xfs_btree_lblock_verify(
 	xfs_fsblock_t		fsb;
 	xfs_failaddr_t		fa;
 
+	ASSERT(!xfs_buftarg_is_mem(bp->b_target));
+
 	/* numrecs verification */
 	if (be16_to_cpu(block->bb_numrecs) > max_recs)
 		return __this_address;
 
 	/* sibling pointer verification */
 	fsb = XFS_DADDR_TO_FSB(mp, xfs_buf_daddr(bp));
-	fa = xfs_btree_check_lblock_siblings(mp, NULL, -1, fsb,
+	fa = xfs_btree_check_fsblock_siblings(mp, fsb,
 			block->bb_u.l.bb_leftsib);
 	if (!fa)
-		fa = xfs_btree_check_lblock_siblings(mp, NULL, -1, fsb,
+		fa = xfs_btree_check_fsblock_siblings(mp, fsb,
 				block->bb_u.l.bb_rightsib);
 	return fa;
 }
 
+/* Verify an in-memory btree block. */
+xfs_failaddr_t
+xfs_btree_memblock_verify(
+	struct xfs_buf		*bp,
+	unsigned int		max_recs)
+{
+	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
+	struct xfs_buftarg	*btp = bp->b_target;
+	xfs_failaddr_t		fa;
+	xfbno_t			bno;
+
+	ASSERT(xfs_buftarg_is_mem(bp->b_target));
+
+	/* numrecs verification */
+	if (be16_to_cpu(block->bb_numrecs) > max_recs)
+		return __this_address;
+
+	/* sibling pointer verification */
+	bno = xfs_daddr_to_xfbno(xfs_buf_daddr(bp));
+	fa = xfs_btree_check_memblock_siblings(btp, bno,
+			block->bb_u.l.bb_leftsib);
+	if (fa)
+		return fa;
+	fa = xfs_btree_check_memblock_siblings(btp, bno,
+			block->bb_u.l.bb_rightsib);
+	if (fa)
+		return fa;
+
+	return NULL;
+}
 /**
- * xfs_btree_sblock_v5hdr_verify() -- verify the v5 fields of a short-format
+ * xfs_btree_agblock_v5hdr_verify() -- verify the v5 fields of a short-format
  *				      btree block
  *
  * @bp: buffer containing the btree block
  */
 xfs_failaddr_t
-xfs_btree_sblock_v5hdr_verify(
+xfs_btree_agblock_v5hdr_verify(
 	struct xfs_buf		*bp)
 {
 	struct xfs_mount	*mp = bp->b_mount;
@@ -4593,13 +4802,13 @@ xfs_btree_sblock_v5hdr_verify(
 }
 
 /**
- * xfs_btree_sblock_verify() -- verify a short-format btree block
+ * xfs_btree_agblock_verify() -- verify a short-format btree block
  *
  * @bp: buffer containing the btree block
  * @max_recs: maximum records allowed in this btree node
  */
 xfs_failaddr_t
-xfs_btree_sblock_verify(
+xfs_btree_agblock_verify(
 	struct xfs_buf		*bp,
 	unsigned int		max_recs)
 {
@@ -4608,16 +4817,18 @@ xfs_btree_sblock_verify(
 	xfs_agblock_t		agbno;
 	xfs_failaddr_t		fa;
 
+	ASSERT(!xfs_buftarg_is_mem(bp->b_target));
+
 	/* numrecs verification */
 	if (be16_to_cpu(block->bb_numrecs) > max_recs)
 		return __this_address;
 
 	/* sibling pointer verification */
 	agbno = xfs_daddr_to_agbno(mp, xfs_buf_daddr(bp));
-	fa = xfs_btree_check_sblock_siblings(bp->b_pag, NULL, -1, agbno,
+	fa = xfs_btree_check_agblock_siblings(bp->b_pag, agbno,
 			block->bb_u.s.bb_leftsib);
 	if (!fa)
-		fa = xfs_btree_check_sblock_siblings(bp->b_pag, NULL, -1, agbno,
+		fa = xfs_btree_check_agblock_siblings(bp->b_pag, agbno,
 				block->bb_u.s.bb_rightsib);
 	return fa;
 }
@@ -4815,7 +5026,7 @@ xfs_btree_overlapped_query_range(
 
 	/* Load the root of the btree. */
 	level = cur->bc_nlevels - 1;
-	cur->bc_ops->init_ptr_from_cur(cur, &ptr);
+	xfs_btree_init_ptr_from_cur(cur, &ptr);
 	error = xfs_btree_lookup_get_block(cur, level, &ptr, &block);
 	if (error)
 		return error;
@@ -4966,7 +5177,7 @@ xfs_btree_query_range(
 	if (!xfs_btree_keycmp_le(cur, &low_key, &high_key))
 		return -EINVAL;
 
-	if (!(cur->bc_flags & XFS_BTREE_OVERLAPPING))
+	if (!(cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING))
 		return xfs_btree_simple_query_range(cur, &low_key,
 				&high_key, fn, priv);
 	return xfs_btree_overlapped_query_range(cur, &low_key, &high_key,
@@ -5020,7 +5231,7 @@ xfs_btree_diff_two_ptrs(
 	const union xfs_btree_ptr	*a,
 	const union xfs_btree_ptr	*b)
 {
-	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+	if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN)
 		return (int64_t)be64_to_cpu(a->l) - be64_to_cpu(b->l);
 	return (int64_t)be32_to_cpu(a->s) - be32_to_cpu(b->s);
 }
@@ -5074,7 +5285,7 @@ xfs_btree_has_records_helper(
 		key_contig = cur->bc_ops->keys_contiguous(cur, &info->high_key,
 					&rec_key, info->key_mask);
 		if (key_contig == XBTREE_KEY_OVERLAP &&
-				!(cur->bc_flags & XFS_BTREE_OVERLAPPING))
+				!(cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING))
 			return -EFSCORRUPTED;
 		if (key_contig == XBTREE_KEY_GAP)
 			return -ECANCELED;
@@ -5168,7 +5379,7 @@ xfs_btree_has_more_records(
 		return true;
 
 	/* There are more record blocks. */
-	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+	if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN)
 		return block->bb_u.l.bb_rightsib != cpu_to_be64(NULLFSBLOCK);
 	else
 		return block->bb_u.s.bb_rightsib != cpu_to_be32(NULLAGBLOCK);
@@ -5233,6 +5444,7 @@ xfs_btree_goto_left_edge(
 		return error;
 	if (stat != 0) {
 		ASSERT(0);
+		xfs_btree_mark_sick(cur);
 		return -EFSCORRUPTED;
 	}
 
diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h
index d906324e25..f93374278a 100644
--- a/fs/xfs/libxfs/xfs_btree.h
+++ b/fs/xfs/libxfs/xfs_btree.h
@@ -55,15 +55,8 @@ union xfs_btree_rec {
 #define	XFS_LOOKUP_LE	((xfs_lookup_t)XFS_LOOKUP_LEi)
 #define	XFS_LOOKUP_GE	((xfs_lookup_t)XFS_LOOKUP_GEi)
 
-#define	XFS_BTNUM_BNO	((xfs_btnum_t)XFS_BTNUM_BNOi)
-#define	XFS_BTNUM_CNT	((xfs_btnum_t)XFS_BTNUM_CNTi)
-#define	XFS_BTNUM_BMAP	((xfs_btnum_t)XFS_BTNUM_BMAPi)
-#define	XFS_BTNUM_INO	((xfs_btnum_t)XFS_BTNUM_INOi)
-#define	XFS_BTNUM_FINO	((xfs_btnum_t)XFS_BTNUM_FINOi)
-#define	XFS_BTNUM_RMAP	((xfs_btnum_t)XFS_BTNUM_RMAPi)
-#define	XFS_BTNUM_REFC	((xfs_btnum_t)XFS_BTNUM_REFCi)
-
-uint32_t xfs_btree_magic(int crc, xfs_btnum_t btnum);
+struct xfs_btree_ops;
+uint32_t xfs_btree_magic(struct xfs_mount *mp, const struct xfs_btree_ops *ops);
 
 /*
  * For logging record fields.
@@ -86,9 +79,11 @@ uint32_t xfs_btree_magic(int crc, xfs_btnum_t btnum);
  * Generic stats interface
  */
 #define XFS_BTREE_STATS_INC(cur, stat)	\
-	XFS_STATS_INC_OFF((cur)->bc_mp, (cur)->bc_statoff + __XBTS_ ## stat)
+	XFS_STATS_INC_OFF((cur)->bc_mp, \
+		(cur)->bc_ops->statoff + __XBTS_ ## stat)
 #define XFS_BTREE_STATS_ADD(cur, stat, val)	\
-	XFS_STATS_ADD_OFF((cur)->bc_mp, (cur)->bc_statoff + __XBTS_ ## stat, val)
+	XFS_STATS_ADD_OFF((cur)->bc_mp, \
+		(cur)->bc_ops->statoff + __XBTS_ ## stat, val)
 
 enum xbtree_key_contig {
 	XBTREE_KEY_GAP = 0,
@@ -111,10 +106,37 @@ static inline enum xbtree_key_contig xbtree_key_contig(uint64_t x, uint64_t y)
 	return XBTREE_KEY_OVERLAP;
 }
 
+#define XFS_BTREE_LONG_PTR_LEN		(sizeof(__be64))
+#define XFS_BTREE_SHORT_PTR_LEN		(sizeof(__be32))
+
+enum xfs_btree_type {
+	XFS_BTREE_TYPE_AG,
+	XFS_BTREE_TYPE_INODE,
+	XFS_BTREE_TYPE_MEM,
+};
+
 struct xfs_btree_ops {
-	/* size of the key and record structures */
-	size_t	key_len;
-	size_t	rec_len;
+	const char		*name;
+
+	/* Type of btree - AG-rooted or inode-rooted */
+	enum xfs_btree_type	type;
+
+	/* XFS_BTGEO_* flags that determine the geometry of the btree */
+	unsigned int		geom_flags;
+
+	/* size of the key, pointer, and record structures */
+	size_t			key_len;
+	size_t			ptr_len;
+	size_t			rec_len;
+
+	/* LRU refcount to set on each btree buffer created */
+	unsigned int		lru_refs;
+
+	/* offset of btree stats array */
+	unsigned int		statoff;
+
+	/* sick mask for health reporting (only for XFS_BTREE_TYPE_AG) */
+	unsigned int		sick_mask;
 
 	/* cursor operations */
 	struct xfs_btree_cur *(*dup_cursor)(struct xfs_btree_cur *);
@@ -199,6 +221,10 @@ struct xfs_btree_ops {
 			       const union xfs_btree_key *mask);
 };
 
+/* btree geometry flags */
+#define XFS_BTGEO_LASTREC_UPDATE	(1U << 0) /* track last rec externally */
+#define XFS_BTGEO_OVERLAPPING		(1U << 1) /* overlapping intervals */
+
 /*
  * Reasons for the update_lastrec method to be called.
  */
@@ -215,39 +241,6 @@ union xfs_btree_irec {
 	struct xfs_refcount_irec	rc;
 };
 
-/* Per-AG btree information. */
-struct xfs_btree_cur_ag {
-	struct xfs_perag		*pag;
-	union {
-		struct xfs_buf		*agbp;
-		struct xbtree_afakeroot	*afake;	/* for staging cursor */
-	};
-	union {
-		struct {
-			unsigned int	nr_ops;	/* # record updates */
-			unsigned int	shape_changes;	/* # of extent splits */
-		} refc;
-		struct {
-			bool		active;	/* allocation cursor state */
-		} abt;
-	};
-};
-
-/* Btree-in-inode cursor information */
-struct xfs_btree_cur_ino {
-	struct xfs_inode		*ip;
-	struct xbtree_ifakeroot		*ifake;	/* for staging cursor */
-	int				allocated;
-	short				forksize;
-	char				whichfork;
-	char				flags;
-/* We are converting a delalloc reservation */
-#define	XFS_BTCUR_BMBT_WASDEL		(1 << 0)
-
-/* For extent swap, ignore owner check in verifier */
-#define	XFS_BTCUR_BMBT_INVALID_OWNER	(1 << 1)
-};
-
 struct xfs_btree_level {
 	/* buffer pointer */
 	struct xfs_buf		*bp;
@@ -272,21 +265,38 @@ struct xfs_btree_cur
 	const struct xfs_btree_ops *bc_ops;
 	struct kmem_cache	*bc_cache; /* cursor cache */
 	unsigned int		bc_flags; /* btree features - below */
-	xfs_btnum_t		bc_btnum; /* identifies which btree type */
 	union xfs_btree_irec	bc_rec;	/* current insert/search record value */
 	uint8_t			bc_nlevels; /* number of levels in the tree */
 	uint8_t			bc_maxlevels; /* maximum levels for this btree type */
-	int			bc_statoff; /* offset of btree stats array */
 
-	/*
-	 * Short btree pointers need an agno to be able to turn the pointers
-	 * into physical addresses for IO, so the btree cursor switches between
-	 * bc_ino and bc_ag based on whether XFS_BTREE_LONG_PTRS is set for the
-	 * cursor.
-	 */
+	/* per-type information */
 	union {
-		struct xfs_btree_cur_ag	bc_ag;
-		struct xfs_btree_cur_ino bc_ino;
+		struct {
+			struct xfs_inode	*ip;
+			short			forksize;
+			char			whichfork;
+			struct xbtree_ifakeroot	*ifake;	/* for staging cursor */
+		} bc_ino;
+		struct {
+			struct xfs_perag	*pag;
+			struct xfs_buf		*agbp;
+			struct xbtree_afakeroot	*afake;	/* for staging cursor */
+		} bc_ag;
+		struct {
+			struct xfbtree		*xfbtree;
+			struct xfs_perag	*pag;
+		} bc_mem;
+	};
+
+	/* per-format private data */
+	union {
+		struct {
+			int		allocated;
+		} bc_bmap;	/* bmapbt */
+		struct {
+			unsigned int	nr_ops;		/* # record updates */
+			unsigned int	shape_changes;	/* # of extent splits */
+		} bc_refc;	/* refcountbt */
 	};
 
 	/* Must be at the end of the struct! */
@@ -304,18 +314,22 @@ xfs_btree_cur_sizeof(unsigned int nlevels)
 	return struct_size_t(struct xfs_btree_cur, bc_levels, nlevels);
 }
 
-/* cursor flags */
-#define XFS_BTREE_LONG_PTRS		(1<<0)	/* pointers are 64bits long */
-#define XFS_BTREE_ROOT_IN_INODE		(1<<1)	/* root may be variable size */
-#define XFS_BTREE_LASTREC_UPDATE	(1<<2)	/* track last rec externally */
-#define XFS_BTREE_CRC_BLOCKS		(1<<3)	/* uses extended btree blocks */
-#define XFS_BTREE_OVERLAPPING		(1<<4)	/* overlapping intervals */
+/* cursor state flags */
 /*
  * The root of this btree is a fakeroot structure so that we can stage a btree
  * rebuild without leaving it accessible via primary metadata.  The ops struct
  * is dynamically allocated and must be freed when the cursor is deleted.
  */
-#define XFS_BTREE_STAGING		(1<<5)
+#define XFS_BTREE_STAGING		(1U << 0)
+
+/* We are converting a delalloc reservation (only for bmbt btrees) */
+#define	XFS_BTREE_BMBT_WASDEL		(1U << 1)
+
+/* For extent swap, ignore owner check in verifier (only for bmbt btrees) */
+#define	XFS_BTREE_BMBT_INVALID_OWNER	(1U << 2)
+
+/* Cursor is active (only for allocbt btrees) */
+#define	XFS_BTREE_ALLOCBT_ACTIVE	(1U << 3)
 
 #define	XFS_BTREE_NOERROR	0
 #define	XFS_BTREE_ERROR		1
@@ -325,14 +339,10 @@ xfs_btree_cur_sizeof(unsigned int nlevels)
  */
 #define	XFS_BUF_TO_BLOCK(bp)	((struct xfs_btree_block *)((bp)->b_addr))
 
-/*
- * Internal long and short btree block checks.  They return NULL if the
- * block is ok or the address of the failed check otherwise.
- */
-xfs_failaddr_t __xfs_btree_check_lblock(struct xfs_btree_cur *cur,
-		struct xfs_btree_block *block, int level, struct xfs_buf *bp);
-xfs_failaddr_t __xfs_btree_check_sblock(struct xfs_btree_cur *cur,
+xfs_failaddr_t __xfs_btree_check_block(struct xfs_btree_cur *cur,
 		struct xfs_btree_block *block, int level, struct xfs_buf *bp);
+int __xfs_btree_check_ptr(struct xfs_btree_cur *cur,
+		const union xfs_btree_ptr *ptr, int index, int level);
 
 /*
  * Check that block header is ok.
@@ -345,24 +355,6 @@ xfs_btree_check_block(
 	struct xfs_buf		*bp);	/* buffer containing block, if any */
 
 /*
- * Check that (long) pointer is ok.
- */
-bool					/* error (0 or EFSCORRUPTED) */
-xfs_btree_check_lptr(
-	struct xfs_btree_cur	*cur,	/* btree cursor */
-	xfs_fsblock_t		fsbno,	/* btree block disk address */
-	int			level);	/* btree block level */
-
-/*
- * Check that (short) pointer is ok.
- */
-bool					/* error (0 or EFSCORRUPTED) */
-xfs_btree_check_sptr(
-	struct xfs_btree_cur	*cur,	/* btree cursor */
-	xfs_agblock_t		agbno,	/* btree block disk address */
-	int			level);	/* btree block level */
-
-/*
  * Delete the btree cursor.
  */
 void
@@ -392,63 +384,14 @@ xfs_btree_offsets(
 	int			*last);	/* output: last byte offset */
 
 /*
- * Get a buffer for the block, return it read in.
- * Long-form addressing.
- */
-int					/* error */
-xfs_btree_read_bufl(
-	struct xfs_mount	*mp,	/* file system mount point */
-	struct xfs_trans	*tp,	/* transaction pointer */
-	xfs_fsblock_t		fsbno,	/* file system block number */
-	struct xfs_buf		**bpp,	/* buffer for fsbno */
-	int			refval,	/* ref count value for buffer */
-	const struct xfs_buf_ops *ops);
-
-/*
- * Read-ahead the block, don't wait for it, don't return a buffer.
- * Long-form addressing.
- */
-void					/* error */
-xfs_btree_reada_bufl(
-	struct xfs_mount	*mp,	/* file system mount point */
-	xfs_fsblock_t		fsbno,	/* file system block number */
-	xfs_extlen_t		count,	/* count of filesystem blocks */
-	const struct xfs_buf_ops *ops);
-
-/*
- * Read-ahead the block, don't wait for it, don't return a buffer.
- * Short-form addressing.
- */
-void					/* error */
-xfs_btree_reada_bufs(
-	struct xfs_mount	*mp,	/* file system mount point */
-	xfs_agnumber_t		agno,	/* allocation group number */
-	xfs_agblock_t		agbno,	/* allocation group block number */
-	xfs_extlen_t		count,	/* count of filesystem blocks */
-	const struct xfs_buf_ops *ops);
-
-/*
  * Initialise a new btree block header
  */
-void
-xfs_btree_init_block(
-	struct xfs_mount *mp,
-	struct xfs_buf	*bp,
-	xfs_btnum_t	btnum,
-	__u16		level,
-	__u16		numrecs,
-	__u64		owner);
-
-void
-xfs_btree_init_block_int(
-	struct xfs_mount	*mp,
-	struct xfs_btree_block	*buf,
-	xfs_daddr_t		blkno,
-	xfs_btnum_t		btnum,
-	__u16			level,
-	__u16			numrecs,
-	__u64			owner,
-	unsigned int		flags);
+void xfs_btree_init_buf(struct xfs_mount *mp, struct xfs_buf *bp,
+		const struct xfs_btree_ops *ops, __u16 level, __u16 numrecs,
+		__u64 owner);
+void xfs_btree_init_block(struct xfs_mount *mp,
+		struct xfs_btree_block *buf, const struct xfs_btree_ops *ops,
+		__u16 level, __u16 numrecs, __u64 owner);
 
 /*
  * Common btree core entry points.
@@ -467,10 +410,10 @@ int xfs_btree_change_owner(struct xfs_btree_cur *cur, uint64_t new_owner,
 /*
  * btree block CRC helpers
  */
-void xfs_btree_lblock_calc_crc(struct xfs_buf *);
-bool xfs_btree_lblock_verify_crc(struct xfs_buf *);
-void xfs_btree_sblock_calc_crc(struct xfs_buf *);
-bool xfs_btree_sblock_verify_crc(struct xfs_buf *);
+void xfs_btree_fsblock_calc_crc(struct xfs_buf *);
+bool xfs_btree_fsblock_verify_crc(struct xfs_buf *);
+void xfs_btree_agblock_calc_crc(struct xfs_buf *);
+bool xfs_btree_agblock_verify_crc(struct xfs_buf *);
 
 /*
  * Internal btree helpers also used by xfs_bmap.c.
@@ -510,12 +453,14 @@ static inline int xfs_btree_get_level(const struct xfs_btree_block *block)
 #define	XFS_FILBLKS_MIN(a,b)	min_t(xfs_filblks_t, (a), (b))
 #define	XFS_FILBLKS_MAX(a,b)	max_t(xfs_filblks_t, (a), (b))
 
-xfs_failaddr_t xfs_btree_sblock_v5hdr_verify(struct xfs_buf *bp);
-xfs_failaddr_t xfs_btree_sblock_verify(struct xfs_buf *bp,
+xfs_failaddr_t xfs_btree_agblock_v5hdr_verify(struct xfs_buf *bp);
+xfs_failaddr_t xfs_btree_agblock_verify(struct xfs_buf *bp,
 		unsigned int max_recs);
-xfs_failaddr_t xfs_btree_lblock_v5hdr_verify(struct xfs_buf *bp,
+xfs_failaddr_t xfs_btree_fsblock_v5hdr_verify(struct xfs_buf *bp,
 		uint64_t owner);
-xfs_failaddr_t xfs_btree_lblock_verify(struct xfs_buf *bp,
+xfs_failaddr_t xfs_btree_fsblock_verify(struct xfs_buf *bp,
+		unsigned int max_recs);
+xfs_failaddr_t xfs_btree_memblock_verify(struct xfs_buf *bp,
 		unsigned int max_recs);
 
 unsigned int xfs_btree_compute_maxlevels(const unsigned int *limits,
@@ -690,7 +635,7 @@ xfs_btree_islastblock(
 
 	block = xfs_btree_get_block(cur, level, &bp);
 
-	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+	if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN)
 		return block->bb_u.l.bb_rightsib == cpu_to_be64(NULLFSBLOCK);
 	return block->bb_u.s.bb_rightsib == cpu_to_be32(NULLAGBLOCK);
 }
@@ -714,21 +659,28 @@ void xfs_btree_copy_ptrs(struct xfs_btree_cur *cur,
 void xfs_btree_copy_keys(struct xfs_btree_cur *cur,
 		union xfs_btree_key *dst_key,
 		const union xfs_btree_key *src_key, int numkeys);
+void xfs_btree_init_ptr_from_cur(struct xfs_btree_cur *cur,
+		union xfs_btree_ptr *ptr);
 
 static inline struct xfs_btree_cur *
 xfs_btree_alloc_cursor(
 	struct xfs_mount	*mp,
 	struct xfs_trans	*tp,
-	xfs_btnum_t		btnum,
+	const struct xfs_btree_ops *ops,
 	uint8_t			maxlevels,
 	struct kmem_cache	*cache)
 {
 	struct xfs_btree_cur	*cur;
 
-	cur = kmem_cache_zalloc(cache, GFP_NOFS | __GFP_NOFAIL);
+	ASSERT(ops->ptr_len == XFS_BTREE_LONG_PTR_LEN ||
+	       ops->ptr_len == XFS_BTREE_SHORT_PTR_LEN);
+
+	/* BMBT allocations can come through from non-transactional context. */
+	cur = kmem_cache_zalloc(cache,
+			GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL);
+	cur->bc_ops = ops;
 	cur->bc_tp = tp;
 	cur->bc_mp = mp;
-	cur->bc_btnum = btnum;
 	cur->bc_maxlevels = maxlevels;
 	cur->bc_cache = cache;
 
@@ -740,4 +692,14 @@ void xfs_btree_destroy_cur_caches(void);
 
 int xfs_btree_goto_left_edge(struct xfs_btree_cur *cur);
 
+/* Does this level of the cursor point to the inode root (and not a block)? */
+static inline bool
+xfs_btree_at_iroot(
+	const struct xfs_btree_cur	*cur,
+	int				level)
+{
+	return cur->bc_ops->type == XFS_BTREE_TYPE_INODE &&
+	       level == cur->bc_nlevels - 1;
+}
+
 #endif	/* __XFS_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_btree_mem.c b/fs/xfs/libxfs/xfs_btree_mem.c
new file mode 100644
index 0000000000..036061fe32
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_btree_mem.c
@@ -0,0 +1,347 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (c) 2021-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_trans.h"
+#include "xfs_btree.h"
+#include "xfs_error.h"
+#include "xfs_buf_mem.h"
+#include "xfs_btree_mem.h"
+#include "xfs_ag.h"
+#include "xfs_buf_item.h"
+#include "xfs_trace.h"
+
+/* Set the root of an in-memory btree. */
+void
+xfbtree_set_root(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_ptr	*ptr,
+	int				inc)
+{
+	ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_MEM);
+
+	cur->bc_mem.xfbtree->root = *ptr;
+	cur->bc_mem.xfbtree->nlevels += inc;
+}
+
+/* Initialize a pointer from the in-memory btree header. */
+void
+xfbtree_init_ptr_from_cur(
+	struct xfs_btree_cur		*cur,
+	union xfs_btree_ptr		*ptr)
+{
+	ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_MEM);
+
+	*ptr = cur->bc_mem.xfbtree->root;
+}
+
+/* Duplicate an in-memory btree cursor. */
+struct xfs_btree_cur *
+xfbtree_dup_cursor(
+	struct xfs_btree_cur		*cur)
+{
+	struct xfs_btree_cur		*ncur;
+
+	ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_MEM);
+
+	ncur = xfs_btree_alloc_cursor(cur->bc_mp, cur->bc_tp, cur->bc_ops,
+			cur->bc_maxlevels, cur->bc_cache);
+	ncur->bc_flags = cur->bc_flags;
+	ncur->bc_nlevels = cur->bc_nlevels;
+	ncur->bc_mem.xfbtree = cur->bc_mem.xfbtree;
+
+	if (cur->bc_mem.pag)
+		ncur->bc_mem.pag = xfs_perag_hold(cur->bc_mem.pag);
+
+	return ncur;
+}
+
+/* Close the btree xfile and release all resources. */
+void
+xfbtree_destroy(
+	struct xfbtree		*xfbt)
+{
+	xfs_buftarg_drain(xfbt->target);
+}
+
+/* Compute the number of bytes available for records. */
+static inline unsigned int
+xfbtree_rec_bytes(
+	struct xfs_mount		*mp,
+	const struct xfs_btree_ops	*ops)
+{
+	return XMBUF_BLOCKSIZE - XFS_BTREE_LBLOCK_CRC_LEN;
+}
+
+/* Initialize an empty leaf block as the btree root. */
+STATIC int
+xfbtree_init_leaf_block(
+	struct xfs_mount		*mp,
+	struct xfbtree			*xfbt,
+	const struct xfs_btree_ops	*ops)
+{
+	struct xfs_buf			*bp;
+	xfbno_t				bno = xfbt->highest_bno++;
+	int				error;
+
+	error = xfs_buf_get(xfbt->target, xfbno_to_daddr(bno), XFBNO_BBSIZE,
+			&bp);
+	if (error)
+		return error;
+
+	trace_xfbtree_create_root_buf(xfbt, bp);
+
+	bp->b_ops = ops->buf_ops;
+	xfs_btree_init_buf(mp, bp, ops, 0, 0, xfbt->owner);
+	xfs_buf_relse(bp);
+
+	xfbt->root.l = cpu_to_be64(bno);
+	return 0;
+}
+
+/*
+ * Create an in-memory btree root that can be used with the given xmbuf.
+ * Callers must set xfbt->owner.
+ */
+int
+xfbtree_init(
+	struct xfs_mount		*mp,
+	struct xfbtree			*xfbt,
+	struct xfs_buftarg		*btp,
+	const struct xfs_btree_ops	*ops)
+{
+	unsigned int			blocklen = xfbtree_rec_bytes(mp, ops);
+	unsigned int			keyptr_len;
+	int				error;
+
+	/* Requires a long-format CRC-format btree */
+	if (!xfs_has_crc(mp)) {
+		ASSERT(xfs_has_crc(mp));
+		return -EINVAL;
+	}
+	if (ops->ptr_len != XFS_BTREE_LONG_PTR_LEN) {
+		ASSERT(ops->ptr_len == XFS_BTREE_LONG_PTR_LEN);
+		return -EINVAL;
+	}
+
+	memset(xfbt, 0, sizeof(*xfbt));
+	xfbt->target = btp;
+
+	/* Set up min/maxrecs for this btree. */
+	keyptr_len = ops->key_len + sizeof(__be64);
+	xfbt->maxrecs[0] = blocklen / ops->rec_len;
+	xfbt->maxrecs[1] = blocklen / keyptr_len;
+	xfbt->minrecs[0] = xfbt->maxrecs[0] / 2;
+	xfbt->minrecs[1] = xfbt->maxrecs[1] / 2;
+	xfbt->highest_bno = 0;
+	xfbt->nlevels = 1;
+
+	/* Initialize the empty btree. */
+	error = xfbtree_init_leaf_block(mp, xfbt, ops);
+	if (error)
+		goto err_freesp;
+
+	trace_xfbtree_init(mp, xfbt, ops);
+
+	return 0;
+
+err_freesp:
+	xfs_buftarg_drain(xfbt->target);
+	return error;
+}
+
+/* Allocate a block to our in-memory btree. */
+int
+xfbtree_alloc_block(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_ptr	*start,
+	union xfs_btree_ptr		*new,
+	int				*stat)
+{
+	struct xfbtree			*xfbt = cur->bc_mem.xfbtree;
+	xfbno_t				bno = xfbt->highest_bno++;
+
+	ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_MEM);
+
+	trace_xfbtree_alloc_block(xfbt, cur, bno);
+
+	/* Fail if the block address exceeds the maximum for the buftarg. */
+	if (!xfbtree_verify_bno(xfbt, bno)) {
+		ASSERT(xfbtree_verify_bno(xfbt, bno));
+		*stat = 0;
+		return 0;
+	}
+
+	new->l = cpu_to_be64(bno);
+	*stat = 1;
+	return 0;
+}
+
+/* Free a block from our in-memory btree. */
+int
+xfbtree_free_block(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp)
+{
+	struct xfbtree		*xfbt = cur->bc_mem.xfbtree;
+	xfs_daddr_t		daddr = xfs_buf_daddr(bp);
+	xfbno_t			bno = xfs_daddr_to_xfbno(daddr);
+
+	ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_MEM);
+
+	trace_xfbtree_free_block(xfbt, cur, bno);
+
+	if (bno + 1 == xfbt->highest_bno)
+		xfbt->highest_bno--;
+
+	return 0;
+}
+
+/* Return the minimum number of records for a btree block. */
+int
+xfbtree_get_minrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
+{
+	struct xfbtree		*xfbt = cur->bc_mem.xfbtree;
+
+	return xfbt->minrecs[level != 0];
+}
+
+/* Return the maximum number of records for a btree block. */
+int
+xfbtree_get_maxrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
+{
+	struct xfbtree		*xfbt = cur->bc_mem.xfbtree;
+
+	return xfbt->maxrecs[level != 0];
+}
+
+/* If this log item is a buffer item that came from the xfbtree, return it. */
+static inline struct xfs_buf *
+xfbtree_buf_match(
+	struct xfbtree			*xfbt,
+	const struct xfs_log_item	*lip)
+{
+	const struct xfs_buf_log_item	*bli;
+	struct xfs_buf			*bp;
+
+	if (lip->li_type != XFS_LI_BUF)
+		return NULL;
+
+	bli = container_of(lip, struct xfs_buf_log_item, bli_item);
+	bp = bli->bli_buf;
+	if (bp->b_target != xfbt->target)
+		return NULL;
+
+	return bp;
+}
+
+/*
+ * Commit changes to the incore btree immediately by writing all dirty xfbtree
+ * buffers to the backing xfile.  This detaches all xfbtree buffers from the
+ * transaction, even on failure.  The buffer locks are dropped between the
+ * delwri queue and submit, so the caller must synchronize btree access.
+ *
+ * Normally we'd let the buffers commit with the transaction and get written to
+ * the xfile via the log, but online repair stages ephemeral btrees in memory
+ * and uses the btree_staging functions to write new btrees to disk atomically.
+ * The in-memory btree (and its backing store) are discarded at the end of the
+ * repair phase, which means that xfbtree buffers cannot commit with the rest
+ * of a transaction.
+ *
+ * In other words, online repair only needs the transaction to collect buffer
+ * pointers and to avoid buffer deadlocks, not to guarantee consistency of
+ * updates.
+ */
+int
+xfbtree_trans_commit(
+	struct xfbtree		*xfbt,
+	struct xfs_trans	*tp)
+{
+	struct xfs_log_item	*lip, *n;
+	bool			tp_dirty = false;
+	int			error = 0;
+
+	/*
+	 * For each xfbtree buffer attached to the transaction, write the dirty
+	 * buffers to the xfile and release them.
+	 */
+	list_for_each_entry_safe(lip, n, &tp->t_items, li_trans) {
+		struct xfs_buf	*bp = xfbtree_buf_match(xfbt, lip);
+
+		if (!bp) {
+			if (test_bit(XFS_LI_DIRTY, &lip->li_flags))
+				tp_dirty |= true;
+			continue;
+		}
+
+		trace_xfbtree_trans_commit_buf(xfbt, bp);
+
+		xmbuf_trans_bdetach(tp, bp);
+
+		/*
+		 * If the buffer fails verification, note the failure but
+		 * continue walking the transaction items so that we remove all
+		 * ephemeral btree buffers.
+		 */
+		if (!error)
+			error = xmbuf_finalize(bp);
+
+		xfs_buf_relse(bp);
+	}
+
+	/*
+	 * Reset the transaction's dirty flag to reflect the dirty state of the
+	 * log items that are still attached.
+	 */
+	tp->t_flags = (tp->t_flags & ~XFS_TRANS_DIRTY) |
+			(tp_dirty ? XFS_TRANS_DIRTY : 0);
+
+	return error;
+}
+
+/*
+ * Cancel changes to the incore btree by detaching all the xfbtree buffers.
+ * Changes are not undone, so callers must not access the btree ever again.
+ */
+void
+xfbtree_trans_cancel(
+	struct xfbtree		*xfbt,
+	struct xfs_trans	*tp)
+{
+	struct xfs_log_item	*lip, *n;
+	bool			tp_dirty = false;
+
+	list_for_each_entry_safe(lip, n, &tp->t_items, li_trans) {
+		struct xfs_buf	*bp = xfbtree_buf_match(xfbt, lip);
+
+		if (!bp) {
+			if (test_bit(XFS_LI_DIRTY, &lip->li_flags))
+				tp_dirty |= true;
+			continue;
+		}
+
+		trace_xfbtree_trans_cancel_buf(xfbt, bp);
+
+		xmbuf_trans_bdetach(tp, bp);
+		xfs_buf_relse(bp);
+	}
+
+	/*
+	 * Reset the transaction's dirty flag to reflect the dirty state of the
+	 * log items that are still attached.
+	 */
+	tp->t_flags = (tp->t_flags & ~XFS_TRANS_DIRTY) |
+			(tp_dirty ? XFS_TRANS_DIRTY : 0);
+}
diff --git a/fs/xfs/libxfs/xfs_btree_mem.h b/fs/xfs/libxfs/xfs_btree_mem.h
new file mode 100644
index 0000000000..1c3825786e
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_btree_mem.h
@@ -0,0 +1,75 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (c) 2021-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_BTREE_MEM_H__
+#define __XFS_BTREE_MEM_H__
+
+typedef uint64_t xfbno_t;
+
+#define XFBNO_BLOCKSIZE			(XMBUF_BLOCKSIZE)
+#define XFBNO_BBSHIFT			(XMBUF_BLOCKSHIFT - BBSHIFT)
+#define XFBNO_BBSIZE			(XFBNO_BLOCKSIZE >> BBSHIFT)
+
+static inline xfs_daddr_t xfbno_to_daddr(xfbno_t blkno)
+{
+	return blkno << XFBNO_BBSHIFT;
+}
+
+static inline xfbno_t xfs_daddr_to_xfbno(xfs_daddr_t daddr)
+{
+	return daddr >> XFBNO_BBSHIFT;
+}
+
+struct xfbtree {
+	/* buffer cache target for this in-memory btree */
+	struct xfs_buftarg		*target;
+
+	/* Highest block number that has been written to. */
+	xfbno_t				highest_bno;
+
+	/* Owner of this btree. */
+	unsigned long long		owner;
+
+	/* Btree header */
+	union xfs_btree_ptr		root;
+	unsigned int			nlevels;
+
+	/* Minimum and maximum records per block. */
+	unsigned int			maxrecs[2];
+	unsigned int			minrecs[2];
+};
+
+#ifdef CONFIG_XFS_BTREE_IN_MEM
+static inline bool xfbtree_verify_bno(struct xfbtree *xfbt, xfbno_t bno)
+{
+	return xmbuf_verify_daddr(xfbt->target, xfbno_to_daddr(bno));
+}
+
+void xfbtree_set_root(struct xfs_btree_cur *cur,
+		const union xfs_btree_ptr *ptr, int inc);
+void xfbtree_init_ptr_from_cur(struct xfs_btree_cur *cur,
+		union xfs_btree_ptr *ptr);
+struct xfs_btree_cur *xfbtree_dup_cursor(struct xfs_btree_cur *cur);
+
+int xfbtree_get_minrecs(struct xfs_btree_cur *cur, int level);
+int xfbtree_get_maxrecs(struct xfs_btree_cur *cur, int level);
+
+int xfbtree_alloc_block(struct xfs_btree_cur *cur,
+		const union xfs_btree_ptr *start, union xfs_btree_ptr *ptr,
+		int *stat);
+int xfbtree_free_block(struct xfs_btree_cur *cur, struct xfs_buf *bp);
+
+/* Callers must set xfbt->target and xfbt->owner before calling this */
+int xfbtree_init(struct xfs_mount *mp, struct xfbtree *xfbt,
+		struct xfs_buftarg *btp, const struct xfs_btree_ops *ops);
+void xfbtree_destroy(struct xfbtree *xfbt);
+
+int xfbtree_trans_commit(struct xfbtree *xfbt, struct xfs_trans *tp);
+void xfbtree_trans_cancel(struct xfbtree *xfbt, struct xfs_trans *tp);
+#else
+# define xfbtree_verify_bno(...)	(false)
+#endif /* CONFIG_XFS_BTREE_IN_MEM */
+
+#endif /* __XFS_BTREE_MEM_H__ */
diff --git a/fs/xfs/libxfs/xfs_btree_staging.c b/fs/xfs/libxfs/xfs_btree_staging.c
index e276eba87c..6949297031 100644
--- a/fs/xfs/libxfs/xfs_btree_staging.c
+++ b/fs/xfs/libxfs/xfs_btree_staging.c
@@ -39,63 +39,6 @@
  */
 
 /*
- * Don't allow staging cursors to be duplicated because they're supposed to be
- * kept private to a single thread.
- */
-STATIC struct xfs_btree_cur *
-xfs_btree_fakeroot_dup_cursor(
-	struct xfs_btree_cur	*cur)
-{
-	ASSERT(0);
-	return NULL;
-}
-
-/*
- * Don't allow block allocation for a staging cursor, because staging cursors
- * do not support regular btree modifications.
- *
- * Bulk loading uses a separate callback to obtain new blocks from a
- * preallocated list, which prevents ENOSPC failures during loading.
- */
-STATIC int
-xfs_btree_fakeroot_alloc_block(
-	struct xfs_btree_cur		*cur,
-	const union xfs_btree_ptr	*start_bno,
-	union xfs_btree_ptr		*new_bno,
-	int				*stat)
-{
-	ASSERT(0);
-	return -EFSCORRUPTED;
-}
-
-/*
- * Don't allow block freeing for a staging cursor, because staging cursors
- * do not support regular btree modifications.
- */
-STATIC int
-xfs_btree_fakeroot_free_block(
-	struct xfs_btree_cur	*cur,
-	struct xfs_buf		*bp)
-{
-	ASSERT(0);
-	return -EFSCORRUPTED;
-}
-
-/* Initialize a pointer to the root block from the fakeroot. */
-STATIC void
-xfs_btree_fakeroot_init_ptr_from_cur(
-	struct xfs_btree_cur	*cur,
-	union xfs_btree_ptr	*ptr)
-{
-	struct xbtree_afakeroot	*afake;
-
-	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
-
-	afake = cur->bc_ag.afake;
-	ptr->s = cpu_to_be32(afake->af_root);
-}
-
-/*
  * Bulk Loading for AG Btrees
  * ==========================
  *
@@ -109,47 +52,20 @@ xfs_btree_fakeroot_init_ptr_from_cur(
  * cursor into a regular btree cursor.
  */
 
-/* Update the btree root information for a per-AG fake root. */
-STATIC void
-xfs_btree_afakeroot_set_root(
-	struct xfs_btree_cur		*cur,
-	const union xfs_btree_ptr	*ptr,
-	int				inc)
-{
-	struct xbtree_afakeroot	*afake = cur->bc_ag.afake;
-
-	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
-	afake->af_root = be32_to_cpu(ptr->s);
-	afake->af_levels += inc;
-}
-
 /*
  * Initialize a AG-rooted btree cursor with the given AG btree fake root.
- * The btree cursor's bc_ops will be overridden as needed to make the staging
- * functionality work.
  */
 void
 xfs_btree_stage_afakeroot(
 	struct xfs_btree_cur		*cur,
 	struct xbtree_afakeroot		*afake)
 {
-	struct xfs_btree_ops		*nops;
-
 	ASSERT(!(cur->bc_flags & XFS_BTREE_STAGING));
-	ASSERT(!(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE));
+	ASSERT(cur->bc_ops->type != XFS_BTREE_TYPE_INODE);
 	ASSERT(cur->bc_tp == NULL);
 
-	nops = kmem_alloc(sizeof(struct xfs_btree_ops), KM_NOFS);
-	memcpy(nops, cur->bc_ops, sizeof(struct xfs_btree_ops));
-	nops->alloc_block = xfs_btree_fakeroot_alloc_block;
-	nops->free_block = xfs_btree_fakeroot_free_block;
-	nops->init_ptr_from_cur = xfs_btree_fakeroot_init_ptr_from_cur;
-	nops->set_root = xfs_btree_afakeroot_set_root;
-	nops->dup_cursor = xfs_btree_fakeroot_dup_cursor;
-
 	cur->bc_ag.afake = afake;
 	cur->bc_nlevels = afake->af_levels;
-	cur->bc_ops = nops;
 	cur->bc_flags |= XFS_BTREE_STAGING;
 }
 
@@ -163,17 +79,15 @@ void
 xfs_btree_commit_afakeroot(
 	struct xfs_btree_cur		*cur,
 	struct xfs_trans		*tp,
-	struct xfs_buf			*agbp,
-	const struct xfs_btree_ops	*ops)
+	struct xfs_buf			*agbp)
 {
 	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
 	ASSERT(cur->bc_tp == NULL);
 
 	trace_xfs_btree_commit_afakeroot(cur);
 
-	kmem_free((void *)cur->bc_ops);
+	cur->bc_ag.afake = NULL;
 	cur->bc_ag.agbp = agbp;
-	cur->bc_ops = ops;
 	cur->bc_flags &= ~XFS_BTREE_STAGING;
 	cur->bc_tp = tp;
 }
@@ -211,29 +125,16 @@ xfs_btree_commit_afakeroot(
 void
 xfs_btree_stage_ifakeroot(
 	struct xfs_btree_cur		*cur,
-	struct xbtree_ifakeroot		*ifake,
-	struct xfs_btree_ops		**new_ops)
+	struct xbtree_ifakeroot		*ifake)
 {
-	struct xfs_btree_ops		*nops;
-
 	ASSERT(!(cur->bc_flags & XFS_BTREE_STAGING));
-	ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE);
+	ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_INODE);
 	ASSERT(cur->bc_tp == NULL);
 
-	nops = kmem_alloc(sizeof(struct xfs_btree_ops), KM_NOFS);
-	memcpy(nops, cur->bc_ops, sizeof(struct xfs_btree_ops));
-	nops->alloc_block = xfs_btree_fakeroot_alloc_block;
-	nops->free_block = xfs_btree_fakeroot_free_block;
-	nops->init_ptr_from_cur = xfs_btree_fakeroot_init_ptr_from_cur;
-	nops->dup_cursor = xfs_btree_fakeroot_dup_cursor;
-
 	cur->bc_ino.ifake = ifake;
 	cur->bc_nlevels = ifake->if_levels;
-	cur->bc_ops = nops;
+	cur->bc_ino.forksize = ifake->if_fork_size;
 	cur->bc_flags |= XFS_BTREE_STAGING;
-
-	if (new_ops)
-		*new_ops = nops;
 }
 
 /*
@@ -246,18 +147,15 @@ void
 xfs_btree_commit_ifakeroot(
 	struct xfs_btree_cur		*cur,
 	struct xfs_trans		*tp,
-	int				whichfork,
-	const struct xfs_btree_ops	*ops)
+	int				whichfork)
 {
 	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
 	ASSERT(cur->bc_tp == NULL);
 
 	trace_xfs_btree_commit_ifakeroot(cur);
 
-	kmem_free((void *)cur->bc_ops);
 	cur->bc_ino.ifake = NULL;
 	cur->bc_ino.whichfork = whichfork;
-	cur->bc_ops = ops;
 	cur->bc_flags &= ~XFS_BTREE_STAGING;
 	cur->bc_tp = tp;
 }
@@ -397,8 +295,7 @@ xfs_btree_bload_prep_block(
 	struct xfs_btree_block		*new_block;
 	int				ret;
 
-	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
-	    level == cur->bc_nlevels - 1) {
+	if (xfs_btree_at_iroot(cur, level)) {
 		struct xfs_ifork	*ifp = xfs_btree_ifork_ptr(cur);
 		size_t			new_size;
 
@@ -406,14 +303,12 @@ xfs_btree_bload_prep_block(
 
 		/* Allocate a new incore btree root block. */
 		new_size = bbl->iroot_size(cur, level, nr_this_block, priv);
-		ifp->if_broot = kmem_zalloc(new_size, 0);
+		ifp->if_broot = kzalloc(new_size, GFP_KERNEL | __GFP_NOFAIL);
 		ifp->if_broot_bytes = (int)new_size;
 
 		/* Initialize it and send it out. */
-		xfs_btree_init_block_int(cur->bc_mp, ifp->if_broot,
-				XFS_BUF_DADDR_NULL, cur->bc_btnum, level,
-				nr_this_block, cur->bc_ino.ip->i_ino,
-				cur->bc_flags);
+		xfs_btree_init_block(cur->bc_mp, ifp->if_broot, cur->bc_ops,
+				level, nr_this_block, cur->bc_ino.ip->i_ino);
 
 		*bpp = NULL;
 		*blockp = ifp->if_broot;
@@ -704,7 +599,7 @@ xfs_btree_bload_compute_geometry(
 		xfs_btree_bload_level_geometry(cur, bbl, level, nr_this_level,
 				&avg_per_block, &level_blocks, &dontcare64);
 
-		if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) {
+		if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE) {
 			/*
 			 * If all the items we want to store at this level
 			 * would fit in the inode root block, then we have our
@@ -763,7 +658,7 @@ xfs_btree_bload_compute_geometry(
 		return -EOVERFLOW;
 
 	bbl->btree_height = cur->bc_nlevels;
-	if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE)
+	if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE)
 		bbl->nr_blocks = nr_blocks - 1;
 	else
 		bbl->nr_blocks = nr_blocks;
@@ -890,7 +785,7 @@ xfs_btree_bload(
 	}
 
 	/* Initialize the new root. */
-	if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) {
+	if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE) {
 		ASSERT(xfs_btree_ptr_is_null(cur, &ptr));
 		cur->bc_ino.ifake->if_levels = cur->bc_nlevels;
 		cur->bc_ino.ifake->if_blocks = total_blocks - 1;
diff --git a/fs/xfs/libxfs/xfs_btree_staging.h b/fs/xfs/libxfs/xfs_btree_staging.h
index 055ea43b1e..0c9c2ffb12 100644
--- a/fs/xfs/libxfs/xfs_btree_staging.h
+++ b/fs/xfs/libxfs/xfs_btree_staging.h
@@ -22,7 +22,7 @@ struct xbtree_afakeroot {
 void xfs_btree_stage_afakeroot(struct xfs_btree_cur *cur,
 		struct xbtree_afakeroot *afake);
 void xfs_btree_commit_afakeroot(struct xfs_btree_cur *cur, struct xfs_trans *tp,
-		struct xfs_buf *agbp, const struct xfs_btree_ops *ops);
+		struct xfs_buf *agbp);
 
 /* Fake root for an inode-rooted btree. */
 struct xbtree_ifakeroot {
@@ -41,10 +41,9 @@ struct xbtree_ifakeroot {
 
 /* Cursor interactions with fake roots for inode-rooted btrees. */
 void xfs_btree_stage_ifakeroot(struct xfs_btree_cur *cur,
-		struct xbtree_ifakeroot *ifake,
-		struct xfs_btree_ops **new_ops);
+		struct xbtree_ifakeroot *ifake);
 void xfs_btree_commit_ifakeroot(struct xfs_btree_cur *cur, struct xfs_trans *tp,
-		int whichfork, const struct xfs_btree_ops *ops);
+		int whichfork);
 
 /* Bulk loading of staged btrees. */
 typedef int (*xfs_btree_bload_get_records_fn)(struct xfs_btree_cur *cur,
@@ -76,8 +75,7 @@ struct xfs_btree_bload {
 
 	/*
 	 * This function should return the size of the in-core btree root
-	 * block.  It is only necessary for XFS_BTREE_ROOT_IN_INODE btree
-	 * types.
+	 * block.  It is only necessary for XFS_BTREE_TYPE_INODE btrees.
 	 */
 	xfs_btree_bload_iroot_size_fn	iroot_size;
 
diff --git a/fs/xfs/libxfs/xfs_da_btree.c b/fs/xfs/libxfs/xfs_da_btree.c
index 5457188bb4..16a529a887 100644
--- a/fs/xfs/libxfs/xfs_da_btree.c
+++ b/fs/xfs/libxfs/xfs_da_btree.c
@@ -23,6 +23,7 @@
 #include "xfs_buf_item.h"
 #include "xfs_log.h"
 #include "xfs_errortag.h"
+#include "xfs_health.h"
 
 /*
  * xfs_da_btree.c
@@ -85,7 +86,8 @@ xfs_da_state_alloc(
 {
 	struct xfs_da_state	*state;
 
-	state = kmem_cache_zalloc(xfs_da_state_cache, GFP_NOFS | __GFP_NOFAIL);
+	state = kmem_cache_zalloc(xfs_da_state_cache,
+			GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL);
 	state->args = args;
 	state->mp = args->dp->i_mount;
 	return state;
@@ -250,6 +252,51 @@ xfs_da3_node_verify(
 	return NULL;
 }
 
+xfs_failaddr_t
+xfs_da3_node_header_check(
+	struct xfs_buf		*bp,
+	xfs_ino_t		owner)
+{
+	struct xfs_mount	*mp = bp->b_mount;
+
+	if (xfs_has_crc(mp)) {
+		struct xfs_da3_blkinfo *hdr3 = bp->b_addr;
+
+		if (hdr3->hdr.magic != cpu_to_be16(XFS_DA3_NODE_MAGIC))
+			return __this_address;
+
+		if (be64_to_cpu(hdr3->owner) != owner)
+			return __this_address;
+	}
+
+	return NULL;
+}
+
+xfs_failaddr_t
+xfs_da3_header_check(
+	struct xfs_buf		*bp,
+	xfs_ino_t		owner)
+{
+	struct xfs_mount	*mp = bp->b_mount;
+	struct xfs_da_blkinfo	*hdr = bp->b_addr;
+
+	if (!xfs_has_crc(mp))
+		return NULL;
+
+	switch (hdr->magic) {
+	case cpu_to_be16(XFS_ATTR3_LEAF_MAGIC):
+		return xfs_attr3_leaf_header_check(bp, owner);
+	case cpu_to_be16(XFS_DA3_NODE_MAGIC):
+		return xfs_da3_node_header_check(bp, owner);
+	case cpu_to_be16(XFS_DIR3_LEAF1_MAGIC):
+	case cpu_to_be16(XFS_DIR3_LEAFN_MAGIC):
+		return xfs_dir3_leaf_header_check(bp, owner);
+	}
+
+	ASSERT(0);
+	return NULL;
+}
+
 static void
 xfs_da3_node_write_verify(
 	struct xfs_buf	*bp)
@@ -352,6 +399,8 @@ const struct xfs_buf_ops xfs_da3_node_buf_ops = {
 static int
 xfs_da3_node_set_type(
 	struct xfs_trans	*tp,
+	struct xfs_inode	*dp,
+	int			whichfork,
 	struct xfs_buf		*bp)
 {
 	struct xfs_da_blkinfo	*info = bp->b_addr;
@@ -373,6 +422,7 @@ xfs_da3_node_set_type(
 		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, tp->t_mountp,
 				info, sizeof(*info));
 		xfs_trans_brelse(tp, bp);
+		xfs_dirattr_mark_sick(dp, whichfork);
 		return -EFSCORRUPTED;
 	}
 }
@@ -391,7 +441,7 @@ xfs_da3_node_read(
 			&xfs_da3_node_buf_ops);
 	if (error || !*bpp || !tp)
 		return error;
-	return xfs_da3_node_set_type(tp, *bpp);
+	return xfs_da3_node_set_type(tp, dp, whichfork, *bpp);
 }
 
 int
@@ -408,6 +458,8 @@ xfs_da3_node_read_mapped(
 	error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, mappedbno,
 			XFS_FSB_TO_BB(mp, xfs_dabuf_nfsb(mp, whichfork)), 0,
 			bpp, &xfs_da3_node_buf_ops);
+	if (xfs_metadata_is_sick(error))
+		xfs_dirattr_mark_sick(dp, whichfork);
 	if (error || !*bpp)
 		return error;
 
@@ -418,7 +470,7 @@ xfs_da3_node_read_mapped(
 
 	if (!tp)
 		return 0;
-	return xfs_da3_node_set_type(tp, *bpp);
+	return xfs_da3_node_set_type(tp, dp, whichfork, *bpp);
 }
 
 /*
@@ -479,7 +531,7 @@ xfs_da3_node_create(
 		memset(hdr3, 0, sizeof(struct xfs_da3_node_hdr));
 		ichdr.magic = XFS_DA3_NODE_MAGIC;
 		hdr3->info.blkno = cpu_to_be64(xfs_buf_daddr(bp));
-		hdr3->info.owner = cpu_to_be64(args->dp->i_ino);
+		hdr3->info.owner = cpu_to_be64(args->owner);
 		uuid_copy(&hdr3->info.uuid, &mp->m_sb.sb_meta_uuid);
 	} else {
 		ichdr.magic = XFS_DA_NODE_MAGIC;
@@ -631,6 +683,7 @@ xfs_da3_split(
 	if (node->hdr.info.forw) {
 		if (be32_to_cpu(node->hdr.info.forw) != addblk->blkno) {
 			xfs_buf_mark_corrupt(oldblk->bp);
+			xfs_da_mark_sick(state->args);
 			error = -EFSCORRUPTED;
 			goto out;
 		}
@@ -644,6 +697,7 @@ xfs_da3_split(
 	if (node->hdr.info.back) {
 		if (be32_to_cpu(node->hdr.info.back) != addblk->blkno) {
 			xfs_buf_mark_corrupt(oldblk->bp);
+			xfs_da_mark_sick(state->args);
 			error = -EFSCORRUPTED;
 			goto out;
 		}
@@ -1190,6 +1244,7 @@ xfs_da3_root_join(
 	struct xfs_da3_icnode_hdr oldroothdr;
 	int			error;
 	struct xfs_inode	*dp = state->args->dp;
+	xfs_failaddr_t		fa;
 
 	trace_xfs_da_root_join(state->args);
 
@@ -1216,6 +1271,13 @@ xfs_da3_root_join(
 	error = xfs_da3_node_read(args->trans, dp, child, &bp, args->whichfork);
 	if (error)
 		return error;
+	fa = xfs_da3_header_check(bp, args->owner);
+	if (fa) {
+		__xfs_buf_mark_corrupt(bp, fa);
+		xfs_trans_brelse(args->trans, bp);
+		xfs_da_mark_sick(args);
+		return -EFSCORRUPTED;
+	}
 	xfs_da_blkinfo_onlychild_validate(bp->b_addr, oldroothdr.level);
 
 	/*
@@ -1250,6 +1312,7 @@ xfs_da3_node_toosmall(
 	struct xfs_da_blkinfo	*info;
 	xfs_dablk_t		blkno;
 	struct xfs_buf		*bp;
+	xfs_failaddr_t		fa;
 	struct xfs_da3_icnode_hdr nodehdr;
 	int			count;
 	int			forward;
@@ -1324,6 +1387,13 @@ xfs_da3_node_toosmall(
 				state->args->whichfork);
 		if (error)
 			return error;
+		fa = xfs_da3_node_header_check(bp, state->args->owner);
+		if (fa) {
+			__xfs_buf_mark_corrupt(bp, fa);
+			xfs_trans_brelse(state->args->trans, bp);
+			xfs_da_mark_sick(state->args);
+			return -EFSCORRUPTED;
+		}
 
 		node = bp->b_addr;
 		xfs_da3_node_hdr_from_disk(dp->i_mount, &thdr, node);
@@ -1582,6 +1652,7 @@ xfs_da3_node_lookup_int(
 	struct xfs_da_node_entry *btree;
 	struct xfs_da3_icnode_hdr nodehdr;
 	struct xfs_da_args	*args;
+	xfs_failaddr_t		fa;
 	xfs_dablk_t		blkno;
 	xfs_dahash_t		hashval;
 	xfs_dahash_t		btreehashval;
@@ -1620,6 +1691,12 @@ xfs_da3_node_lookup_int(
 
 		if (magic == XFS_ATTR_LEAF_MAGIC ||
 		    magic == XFS_ATTR3_LEAF_MAGIC) {
+			fa = xfs_attr3_leaf_header_check(blk->bp, args->owner);
+			if (fa) {
+				__xfs_buf_mark_corrupt(blk->bp, fa);
+				xfs_da_mark_sick(args);
+				return -EFSCORRUPTED;
+			}
 			blk->magic = XFS_ATTR_LEAF_MAGIC;
 			blk->hashval = xfs_attr_leaf_lasthash(blk->bp, NULL);
 			break;
@@ -1627,6 +1704,12 @@ xfs_da3_node_lookup_int(
 
 		if (magic == XFS_DIR2_LEAFN_MAGIC ||
 		    magic == XFS_DIR3_LEAFN_MAGIC) {
+			fa = xfs_dir3_leaf_header_check(blk->bp, args->owner);
+			if (fa) {
+				__xfs_buf_mark_corrupt(blk->bp, fa);
+				xfs_da_mark_sick(args);
+				return -EFSCORRUPTED;
+			}
 			blk->magic = XFS_DIR2_LEAFN_MAGIC;
 			blk->hashval = xfs_dir2_leaf_lasthash(args->dp,
 							      blk->bp, NULL);
@@ -1635,6 +1718,14 @@ xfs_da3_node_lookup_int(
 
 		if (magic != XFS_DA_NODE_MAGIC && magic != XFS_DA3_NODE_MAGIC) {
 			xfs_buf_mark_corrupt(blk->bp);
+			xfs_da_mark_sick(args);
+			return -EFSCORRUPTED;
+		}
+
+		fa = xfs_da3_node_header_check(blk->bp, args->owner);
+		if (fa) {
+			__xfs_buf_mark_corrupt(blk->bp, fa);
+			xfs_da_mark_sick(args);
 			return -EFSCORRUPTED;
 		}
 
@@ -1650,6 +1741,7 @@ xfs_da3_node_lookup_int(
 		/* Tree taller than we can handle; bail out! */
 		if (nodehdr.level >= XFS_DA_NODE_MAXDEPTH) {
 			xfs_buf_mark_corrupt(blk->bp);
+			xfs_da_mark_sick(args);
 			return -EFSCORRUPTED;
 		}
 
@@ -1658,6 +1750,7 @@ xfs_da3_node_lookup_int(
 			expected_level = nodehdr.level - 1;
 		else if (expected_level != nodehdr.level) {
 			xfs_buf_mark_corrupt(blk->bp);
+			xfs_da_mark_sick(args);
 			return -EFSCORRUPTED;
 		} else
 			expected_level--;
@@ -1709,12 +1802,16 @@ xfs_da3_node_lookup_int(
 		}
 
 		/* We can't point back to the root. */
-		if (XFS_IS_CORRUPT(dp->i_mount, blkno == args->geo->leafblk))
+		if (XFS_IS_CORRUPT(dp->i_mount, blkno == args->geo->leafblk)) {
+			xfs_da_mark_sick(args);
 			return -EFSCORRUPTED;
+		}
 	}
 
-	if (XFS_IS_CORRUPT(dp->i_mount, expected_level != 0))
+	if (XFS_IS_CORRUPT(dp->i_mount, expected_level != 0)) {
+		xfs_da_mark_sick(args);
 		return -EFSCORRUPTED;
+	}
 
 	/*
 	 * A leaf block that ends in the hashval that we are interested in
@@ -1732,6 +1829,7 @@ xfs_da3_node_lookup_int(
 			args->blkno = blk->blkno;
 		} else {
 			ASSERT(0);
+			xfs_da_mark_sick(args);
 			return -EFSCORRUPTED;
 		}
 		if (((retval == -ENOENT) || (retval == -ENOATTR)) &&
@@ -1803,6 +1901,7 @@ xfs_da3_blk_link(
 	struct xfs_da_blkinfo	*tmp_info;
 	struct xfs_da_args	*args;
 	struct xfs_buf		*bp;
+	xfs_failaddr_t		fa;
 	int			before = 0;
 	int			error;
 	struct xfs_inode	*dp = state->args->dp;
@@ -1846,6 +1945,13 @@ xfs_da3_blk_link(
 						&bp, args->whichfork);
 			if (error)
 				return error;
+			fa = xfs_da3_header_check(bp, args->owner);
+			if (fa) {
+				__xfs_buf_mark_corrupt(bp, fa);
+				xfs_trans_brelse(args->trans, bp);
+				xfs_da_mark_sick(args);
+				return -EFSCORRUPTED;
+			}
 			ASSERT(bp != NULL);
 			tmp_info = bp->b_addr;
 			ASSERT(tmp_info->magic == old_info->magic);
@@ -1867,6 +1973,13 @@ xfs_da3_blk_link(
 						&bp, args->whichfork);
 			if (error)
 				return error;
+			fa = xfs_da3_header_check(bp, args->owner);
+			if (fa) {
+				__xfs_buf_mark_corrupt(bp, fa);
+				xfs_trans_brelse(args->trans, bp);
+				xfs_da_mark_sick(args);
+				return -EFSCORRUPTED;
+			}
 			ASSERT(bp != NULL);
 			tmp_info = bp->b_addr;
 			ASSERT(tmp_info->magic == old_info->magic);
@@ -1896,6 +2009,7 @@ xfs_da3_blk_unlink(
 	struct xfs_da_blkinfo	*tmp_info;
 	struct xfs_da_args	*args;
 	struct xfs_buf		*bp;
+	xfs_failaddr_t		fa;
 	int			error;
 
 	/*
@@ -1926,6 +2040,13 @@ xfs_da3_blk_unlink(
 						&bp, args->whichfork);
 			if (error)
 				return error;
+			fa = xfs_da3_header_check(bp, args->owner);
+			if (fa) {
+				__xfs_buf_mark_corrupt(bp, fa);
+				xfs_trans_brelse(args->trans, bp);
+				xfs_da_mark_sick(args);
+				return -EFSCORRUPTED;
+			}
 			ASSERT(bp != NULL);
 			tmp_info = bp->b_addr;
 			ASSERT(tmp_info->magic == save_info->magic);
@@ -1943,6 +2064,13 @@ xfs_da3_blk_unlink(
 						&bp, args->whichfork);
 			if (error)
 				return error;
+			fa = xfs_da3_header_check(bp, args->owner);
+			if (fa) {
+				__xfs_buf_mark_corrupt(bp, fa);
+				xfs_trans_brelse(args->trans, bp);
+				xfs_da_mark_sick(args);
+				return -EFSCORRUPTED;
+			}
 			ASSERT(bp != NULL);
 			tmp_info = bp->b_addr;
 			ASSERT(tmp_info->magic == save_info->magic);
@@ -1979,6 +2107,7 @@ xfs_da3_path_shift(
 	struct xfs_da_node_entry *btree;
 	struct xfs_da3_icnode_hdr nodehdr;
 	struct xfs_buf		*bp;
+	xfs_failaddr_t		fa;
 	xfs_dablk_t		blkno = 0;
 	int			level;
 	int			error;
@@ -2057,6 +2186,12 @@ xfs_da3_path_shift(
 		switch (be16_to_cpu(info->magic)) {
 		case XFS_DA_NODE_MAGIC:
 		case XFS_DA3_NODE_MAGIC:
+			fa = xfs_da3_node_header_check(blk->bp, args->owner);
+			if (fa) {
+				__xfs_buf_mark_corrupt(blk->bp, fa);
+				xfs_da_mark_sick(args);
+				return -EFSCORRUPTED;
+			}
 			blk->magic = XFS_DA_NODE_MAGIC;
 			xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr,
 						   bp->b_addr);
@@ -2070,6 +2205,12 @@ xfs_da3_path_shift(
 			break;
 		case XFS_ATTR_LEAF_MAGIC:
 		case XFS_ATTR3_LEAF_MAGIC:
+			fa = xfs_attr3_leaf_header_check(blk->bp, args->owner);
+			if (fa) {
+				__xfs_buf_mark_corrupt(blk->bp, fa);
+				xfs_da_mark_sick(args);
+				return -EFSCORRUPTED;
+			}
 			blk->magic = XFS_ATTR_LEAF_MAGIC;
 			ASSERT(level == path->active-1);
 			blk->index = 0;
@@ -2077,6 +2218,12 @@ xfs_da3_path_shift(
 			break;
 		case XFS_DIR2_LEAFN_MAGIC:
 		case XFS_DIR3_LEAFN_MAGIC:
+			fa = xfs_dir3_leaf_header_check(blk->bp, args->owner);
+			if (fa) {
+				__xfs_buf_mark_corrupt(blk->bp, fa);
+				xfs_da_mark_sick(args);
+				return -EFSCORRUPTED;
+			}
 			blk->magic = XFS_DIR2_LEAFN_MAGIC;
 			ASSERT(level == path->active-1);
 			blk->index = 0;
@@ -2150,8 +2297,8 @@ xfs_da_grow_inode_int(
 	struct xfs_inode	*dp = args->dp;
 	int			w = args->whichfork;
 	xfs_rfsblock_t		nblks = dp->i_nblocks;
-	struct xfs_bmbt_irec	map, *mapp;
-	int			nmap, error, got, i, mapi;
+	struct xfs_bmbt_irec	map, *mapp = &map;
+	int			nmap, error, got, i, mapi = 1;
 
 	/*
 	 * Find a spot in the file space to put the new block.
@@ -2167,14 +2314,7 @@ xfs_da_grow_inode_int(
 	error = xfs_bmapi_write(tp, dp, *bno, count,
 			xfs_bmapi_aflag(w)|XFS_BMAPI_METADATA|XFS_BMAPI_CONTIG,
 			args->total, &map, &nmap);
-	if (error)
-		return error;
-
-	ASSERT(nmap <= 1);
-	if (nmap == 1) {
-		mapp = &map;
-		mapi = 1;
-	} else if (nmap == 0 && count > 1) {
+	if (error == -ENOSPC && count > 1) {
 		xfs_fileoff_t		b;
 		int			c;
 
@@ -2182,7 +2322,8 @@ xfs_da_grow_inode_int(
 		 * If we didn't get it and the block might work if fragmented,
 		 * try without the CONTIG flag.  Loop until we get it all.
 		 */
-		mapp = kmem_alloc(sizeof(*mapp) * count, 0);
+		mapp = kmalloc(sizeof(*mapp) * count,
+				GFP_KERNEL | __GFP_NOFAIL);
 		for (b = *bno, mapi = 0; b < *bno + count; ) {
 			c = (int)(*bno + count - b);
 			nmap = min(XFS_BMAP_MAX_NMAP, c);
@@ -2191,16 +2332,13 @@ xfs_da_grow_inode_int(
 					args->total, &mapp[mapi], &nmap);
 			if (error)
 				goto out_free_map;
-			if (nmap < 1)
-				break;
 			mapi += nmap;
 			b = mapp[mapi - 1].br_startoff +
 			    mapp[mapi - 1].br_blockcount;
 		}
-	} else {
-		mapi = 0;
-		mapp = NULL;
 	}
+	if (error)
+		goto out_free_map;
 
 	/*
 	 * Count the blocks we got, make sure it matches the total.
@@ -2219,7 +2357,7 @@ xfs_da_grow_inode_int(
 
 out_free_map:
 	if (mapp != &map)
-		kmem_free(mapp);
+		kfree(mapp);
 	return error;
 }
 
@@ -2272,6 +2410,7 @@ xfs_da3_swap_lastblock(
 	struct xfs_buf		*last_buf;
 	struct xfs_buf		*sib_buf;
 	struct xfs_buf		*par_buf;
+	xfs_failaddr_t		fa;
 	xfs_dahash_t		dead_hash;
 	xfs_fileoff_t		lastoff;
 	xfs_dablk_t		dead_blkno;
@@ -2297,8 +2436,10 @@ xfs_da3_swap_lastblock(
 	error = xfs_bmap_last_before(tp, dp, &lastoff, w);
 	if (error)
 		return error;
-	if (XFS_IS_CORRUPT(mp, lastoff == 0))
+	if (XFS_IS_CORRUPT(mp, lastoff == 0)) {
+		xfs_da_mark_sick(args);
 		return -EFSCORRUPTED;
+	}
 	/*
 	 * Read the last block in the btree space.
 	 */
@@ -2306,6 +2447,14 @@ xfs_da3_swap_lastblock(
 	error = xfs_da3_node_read(tp, dp, last_blkno, &last_buf, w);
 	if (error)
 		return error;
+	fa = xfs_da3_header_check(last_buf, args->owner);
+	if (fa) {
+		__xfs_buf_mark_corrupt(last_buf, fa);
+		xfs_trans_brelse(tp, last_buf);
+		xfs_da_mark_sick(args);
+		return -EFSCORRUPTED;
+	}
+
 	/*
 	 * Copy the last block into the dead buffer and log it.
 	 */
@@ -2344,10 +2493,18 @@ xfs_da3_swap_lastblock(
 		error = xfs_da3_node_read(tp, dp, sib_blkno, &sib_buf, w);
 		if (error)
 			goto done;
+		fa = xfs_da3_header_check(sib_buf, args->owner);
+		if (fa) {
+			__xfs_buf_mark_corrupt(sib_buf, fa);
+			xfs_da_mark_sick(args);
+			error = -EFSCORRUPTED;
+			goto done;
+		}
 		sib_info = sib_buf->b_addr;
 		if (XFS_IS_CORRUPT(mp,
 				   be32_to_cpu(sib_info->forw) != last_blkno ||
 				   sib_info->magic != dead_info->magic)) {
+			xfs_da_mark_sick(args);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -2364,10 +2521,18 @@ xfs_da3_swap_lastblock(
 		error = xfs_da3_node_read(tp, dp, sib_blkno, &sib_buf, w);
 		if (error)
 			goto done;
+		fa = xfs_da3_header_check(sib_buf, args->owner);
+		if (fa) {
+			__xfs_buf_mark_corrupt(sib_buf, fa);
+			xfs_da_mark_sick(args);
+			error = -EFSCORRUPTED;
+			goto done;
+		}
 		sib_info = sib_buf->b_addr;
 		if (XFS_IS_CORRUPT(mp,
 				   be32_to_cpu(sib_info->back) != last_blkno ||
 				   sib_info->magic != dead_info->magic)) {
+			xfs_da_mark_sick(args);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -2386,10 +2551,18 @@ xfs_da3_swap_lastblock(
 		error = xfs_da3_node_read(tp, dp, par_blkno, &par_buf, w);
 		if (error)
 			goto done;
+		fa = xfs_da3_node_header_check(par_buf, args->owner);
+		if (fa) {
+			__xfs_buf_mark_corrupt(par_buf, fa);
+			xfs_da_mark_sick(args);
+			error = -EFSCORRUPTED;
+			goto done;
+		}
 		par_node = par_buf->b_addr;
 		xfs_da3_node_hdr_from_disk(dp->i_mount, &par_hdr, par_node);
 		if (XFS_IS_CORRUPT(mp,
 				   level >= 0 && level != par_hdr.level + 1)) {
+			xfs_da_mark_sick(args);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -2401,6 +2574,7 @@ xfs_da3_swap_lastblock(
 		     entno++)
 			continue;
 		if (XFS_IS_CORRUPT(mp, entno == par_hdr.count)) {
+			xfs_da_mark_sick(args);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -2426,15 +2600,24 @@ xfs_da3_swap_lastblock(
 		xfs_trans_brelse(tp, par_buf);
 		par_buf = NULL;
 		if (XFS_IS_CORRUPT(mp, par_blkno == 0)) {
+			xfs_da_mark_sick(args);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
 		error = xfs_da3_node_read(tp, dp, par_blkno, &par_buf, w);
 		if (error)
 			goto done;
+		fa = xfs_da3_node_header_check(par_buf, args->owner);
+		if (fa) {
+			__xfs_buf_mark_corrupt(par_buf, fa);
+			xfs_da_mark_sick(args);
+			error = -EFSCORRUPTED;
+			goto done;
+		}
 		par_node = par_buf->b_addr;
 		xfs_da3_node_hdr_from_disk(dp->i_mount, &par_hdr, par_node);
 		if (XFS_IS_CORRUPT(mp, par_hdr.level != level)) {
+			xfs_da_mark_sick(args);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -2518,7 +2701,8 @@ xfs_dabuf_map(
 	int			error = 0, nirecs, i;
 
 	if (nfsb > 1)
-		irecs = kmem_zalloc(sizeof(irec) * nfsb, KM_NOFS);
+		irecs = kzalloc(sizeof(irec) * nfsb,
+				GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL);
 
 	nirecs = nfsb;
 	error = xfs_bmapi_read(dp, bno, nfsb, irecs, &nirecs,
@@ -2531,7 +2715,8 @@ xfs_dabuf_map(
 	 * larger one that needs to be free by the caller.
 	 */
 	if (nirecs > 1) {
-		map = kmem_zalloc(nirecs * sizeof(struct xfs_buf_map), KM_NOFS);
+		map = kzalloc(nirecs * sizeof(struct xfs_buf_map),
+				GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL);
 		if (!map) {
 			error = -ENOMEM;
 			goto out_free_irecs;
@@ -2557,12 +2742,13 @@ xfs_dabuf_map(
 	*nmaps = nirecs;
 out_free_irecs:
 	if (irecs != &irec)
-		kmem_free(irecs);
+		kfree(irecs);
 	return error;
 
 invalid_mapping:
 	/* Caller ok with no mapping. */
 	if (XFS_IS_CORRUPT(mp, !(flags & XFS_DABUF_MAP_HOLE_OK))) {
+		xfs_dirattr_mark_sick(dp, whichfork);
 		error = -EFSCORRUPTED;
 		if (xfs_error_level >= XFS_ERRLEVEL_LOW) {
 			xfs_alert(mp, "%s: bno %u inode %llu",
@@ -2613,7 +2799,7 @@ xfs_da_get_buf(
 
 out_free:
 	if (mapp != &map)
-		kmem_free(mapp);
+		kfree(mapp);
 
 	return error;
 }
@@ -2644,6 +2830,8 @@ xfs_da_read_buf(
 
 	error = xfs_trans_read_buf_map(mp, tp, mp->m_ddev_targp, mapp, nmap, 0,
 			&bp, ops);
+	if (xfs_metadata_is_sick(error))
+		xfs_dirattr_mark_sick(dp, whichfork);
 	if (error)
 		goto out_free;
 
@@ -2654,7 +2842,7 @@ xfs_da_read_buf(
 	*bpp = bp;
 out_free:
 	if (mapp != &map)
-		kmem_free(mapp);
+		kfree(mapp);
 
 	return error;
 }
@@ -2685,7 +2873,7 @@ xfs_da_reada_buf(
 
 out_free:
 	if (mapp != &map)
-		kmem_free(mapp);
+		kfree(mapp);
 
 	return error;
 }
diff --git a/fs/xfs/libxfs/xfs_da_btree.h b/fs/xfs/libxfs/xfs_da_btree.h
index 706baf36e1..354d5d6504 100644
--- a/fs/xfs/libxfs/xfs_da_btree.h
+++ b/fs/xfs/libxfs/xfs_da_btree.h
@@ -54,17 +54,24 @@ enum xfs_dacmp {
  */
 typedef struct xfs_da_args {
 	struct xfs_da_geometry *geo;	/* da block geometry */
-	const uint8_t		*name;		/* string (maybe not NULL terminated) */
-	int		namelen;	/* length of string (maybe no NULL) */
-	uint8_t		filetype;	/* filetype of inode for directories */
+	const uint8_t	*name;		/* string (maybe not NULL terminated) */
+	const uint8_t	*new_name;	/* new attr name */
 	void		*value;		/* set of bytes (maybe contain NULLs) */
-	int		valuelen;	/* length of value */
-	unsigned int	attr_filter;	/* XFS_ATTR_{ROOT,SECURE,INCOMPLETE} */
-	unsigned int	attr_flags;	/* XATTR_{CREATE,REPLACE} */
-	xfs_dahash_t	hashval;	/* hash value of name */
-	xfs_ino_t	inumber;	/* input/output inode number */
+	void		*new_value;	/* new xattr value (may contain NULLs) */
 	struct xfs_inode *dp;		/* directory inode to manipulate */
 	struct xfs_trans *trans;	/* current trans (changes over time) */
+
+	xfs_ino_t	inumber;	/* input/output inode number */
+	xfs_ino_t	owner;		/* inode that owns the dir/attr data */
+
+	int		valuelen;	/* length of value */
+	int		new_valuelen;	/* length of new_value */
+	uint8_t		filetype;	/* filetype of inode for directories */
+	uint8_t		op_flags;	/* operation flags */
+	uint8_t		attr_filter;	/* XFS_ATTR_{ROOT,SECURE,INCOMPLETE} */
+	short		namelen;	/* length of string (maybe no NULL) */
+	short		new_namelen;	/* length of new attr name */
+	xfs_dahash_t	hashval;	/* hash value of name */
 	xfs_extlen_t	total;		/* total blocks needed, for 1st bmap */
 	int		whichfork;	/* data or attribute fork */
 	xfs_dablk_t	blkno;		/* blkno of attr leaf of interest */
@@ -77,7 +84,6 @@ typedef struct xfs_da_args {
 	xfs_dablk_t	rmtblkno2;	/* remote attr value starting blkno */
 	int		rmtblkcnt2;	/* remote attr value block count */
 	int		rmtvaluelen2;	/* remote attr value length in bytes */
-	uint32_t	op_flags;	/* operation flags */
 	enum xfs_dacmp	cmpresult;	/* name compare result for lookups */
 } xfs_da_args_t;
 
@@ -89,10 +95,8 @@ typedef struct xfs_da_args {
 #define XFS_DA_OP_ADDNAME	(1u << 2) /* this is an add operation */
 #define XFS_DA_OP_OKNOENT	(1u << 3) /* lookup op, ENOENT ok, else die */
 #define XFS_DA_OP_CILOOKUP	(1u << 4) /* lookup returns CI name if found */
-#define XFS_DA_OP_NOTIME	(1u << 5) /* don't update inode timestamps */
-#define XFS_DA_OP_REMOVE	(1u << 6) /* this is a remove operation */
-#define XFS_DA_OP_RECOVERY	(1u << 7) /* Log recovery operation */
-#define XFS_DA_OP_LOGGED	(1u << 8) /* Use intent items to track op */
+#define XFS_DA_OP_RECOVERY	(1u << 5) /* Log recovery operation */
+#define XFS_DA_OP_LOGGED	(1u << 6) /* Use intent items to track op */
 
 #define XFS_DA_OP_FLAGS \
 	{ XFS_DA_OP_JUSTCHECK,	"JUSTCHECK" }, \
@@ -100,8 +104,6 @@ typedef struct xfs_da_args {
 	{ XFS_DA_OP_ADDNAME,	"ADDNAME" }, \
 	{ XFS_DA_OP_OKNOENT,	"OKNOENT" }, \
 	{ XFS_DA_OP_CILOOKUP,	"CILOOKUP" }, \
-	{ XFS_DA_OP_NOTIME,	"NOTIME" }, \
-	{ XFS_DA_OP_REMOVE,	"REMOVE" }, \
 	{ XFS_DA_OP_RECOVERY,	"RECOVERY" }, \
 	{ XFS_DA_OP_LOGGED,	"LOGGED" }
 
@@ -235,6 +237,8 @@ void	xfs_da3_node_hdr_from_disk(struct xfs_mount *mp,
 		struct xfs_da3_icnode_hdr *to, struct xfs_da_intnode *from);
 void	xfs_da3_node_hdr_to_disk(struct xfs_mount *mp,
 		struct xfs_da_intnode *to, struct xfs_da3_icnode_hdr *from);
+xfs_failaddr_t xfs_da3_header_check(struct xfs_buf *bp, xfs_ino_t owner);
+xfs_failaddr_t xfs_da3_node_header_check(struct xfs_buf *bp, xfs_ino_t owner);
 
 extern struct kmem_cache	*xfs_da_state_cache;
 
diff --git a/fs/xfs/libxfs/xfs_da_format.h b/fs/xfs/libxfs/xfs_da_format.h
index 24f9d1461f..86de99e2f7 100644
--- a/fs/xfs/libxfs/xfs_da_format.h
+++ b/fs/xfs/libxfs/xfs_da_format.h
@@ -159,6 +159,17 @@ struct xfs_da3_intnode {
 
 #define XFS_DIR3_FT_MAX			9
 
+#define XFS_DIR3_FTYPE_STR \
+	{ XFS_DIR3_FT_UNKNOWN,	"unknown" }, \
+	{ XFS_DIR3_FT_REG_FILE,	"file" }, \
+	{ XFS_DIR3_FT_DIR,	"directory" }, \
+	{ XFS_DIR3_FT_CHRDEV,	"char" }, \
+	{ XFS_DIR3_FT_BLKDEV,	"block" }, \
+	{ XFS_DIR3_FT_FIFO,	"fifo" }, \
+	{ XFS_DIR3_FT_SOCK,	"sock" }, \
+	{ XFS_DIR3_FT_SYMLINK,	"symlink" }, \
+	{ XFS_DIR3_FT_WHT,	"whiteout" }
+
 /*
  * Byte offset in data block and shortform entry.
  */
@@ -703,12 +714,30 @@ struct xfs_attr3_leafblock {
 #define	XFS_ATTR_LOCAL_BIT	0	/* attr is stored locally */
 #define	XFS_ATTR_ROOT_BIT	1	/* limit access to trusted attrs */
 #define	XFS_ATTR_SECURE_BIT	2	/* limit access to secure attrs */
+#define	XFS_ATTR_PARENT_BIT	3	/* parent pointer attrs */
 #define	XFS_ATTR_INCOMPLETE_BIT	7	/* attr in middle of create/delete */
 #define XFS_ATTR_LOCAL		(1u << XFS_ATTR_LOCAL_BIT)
 #define XFS_ATTR_ROOT		(1u << XFS_ATTR_ROOT_BIT)
 #define XFS_ATTR_SECURE		(1u << XFS_ATTR_SECURE_BIT)
+#define XFS_ATTR_PARENT		(1u << XFS_ATTR_PARENT_BIT)
 #define XFS_ATTR_INCOMPLETE	(1u << XFS_ATTR_INCOMPLETE_BIT)
-#define XFS_ATTR_NSP_ONDISK_MASK	(XFS_ATTR_ROOT | XFS_ATTR_SECURE)
+
+#define XFS_ATTR_NSP_ONDISK_MASK	(XFS_ATTR_ROOT | \
+					 XFS_ATTR_SECURE | \
+					 XFS_ATTR_PARENT)
+
+/* Private attr namespaces not exposed to userspace */
+#define XFS_ATTR_PRIVATE_NSP_MASK	(XFS_ATTR_PARENT)
+
+#define XFS_ATTR_ONDISK_MASK	(XFS_ATTR_NSP_ONDISK_MASK | \
+				 XFS_ATTR_LOCAL | \
+				 XFS_ATTR_INCOMPLETE)
+
+#define XFS_ATTR_NAMESPACE_STR \
+	{ XFS_ATTR_LOCAL,	"local" }, \
+	{ XFS_ATTR_ROOT,	"root" }, \
+	{ XFS_ATTR_SECURE,	"secure" }, \
+	{ XFS_ATTR_PARENT,	"parent" }
 
 /*
  * Alignment for namelist and valuelist entries (since they are mixed
@@ -851,9 +880,7 @@ struct xfs_attr3_rmt_hdr {
 
 #define XFS_ATTR3_RMT_CRC_OFF	offsetof(struct xfs_attr3_rmt_hdr, rm_crc)
 
-#define XFS_ATTR3_RMT_BUF_SPACE(mp, bufsize)	\
-	((bufsize) - (xfs_has_crc((mp)) ? \
-			sizeof(struct xfs_attr3_rmt_hdr) : 0))
+unsigned int xfs_attr3_rmt_buf_space(struct xfs_mount *mp);
 
 /* Number of bytes in a directory block. */
 static inline unsigned int xfs_dir2_dirblock_bytes(struct xfs_sb *sbp)
@@ -864,4 +891,17 @@ static inline unsigned int xfs_dir2_dirblock_bytes(struct xfs_sb *sbp)
 xfs_failaddr_t xfs_da3_blkinfo_verify(struct xfs_buf *bp,
 				      struct xfs_da3_blkinfo *hdr3);
 
+/*
+ * Parent pointer attribute format definition
+ *
+ * The xattr name contains the dirent name.
+ * The xattr value encodes the parent inode number and generation to ease
+ * opening parents by handle.
+ * The xattr hashval is xfs_dir2_namehash() ^ p_ino
+ */
+struct xfs_parent_rec {
+	__be64	p_ino;
+	__be32	p_gen;
+} __packed;
+
 #endif /* __XFS_DA_FORMAT_H__ */
diff --git a/fs/xfs/libxfs/xfs_defer.c b/fs/xfs/libxfs/xfs_defer.c
index 66a17910d0..4a078e07e1 100644
--- a/fs/xfs/libxfs/xfs_defer.c
+++ b/fs/xfs/libxfs/xfs_defer.c
@@ -27,6 +27,7 @@
 #include "xfs_da_btree.h"
 #include "xfs_attr.h"
 #include "xfs_trans_priv.h"
+#include "xfs_exchmaps.h"
 
 static struct kmem_cache	*xfs_defer_pending_cache;
 
@@ -819,16 +820,16 @@ xfs_defer_can_append(
 /* Create a new pending item at the end of the transaction list. */
 static inline struct xfs_defer_pending *
 xfs_defer_alloc(
-	struct xfs_trans		*tp,
+	struct list_head		*dfops,
 	const struct xfs_defer_op_type	*ops)
 {
 	struct xfs_defer_pending	*dfp;
 
 	dfp = kmem_cache_zalloc(xfs_defer_pending_cache,
-			GFP_NOFS | __GFP_NOFAIL);
+			GFP_KERNEL | __GFP_NOFAIL);
 	dfp->dfp_ops = ops;
 	INIT_LIST_HEAD(&dfp->dfp_work);
-	list_add_tail(&dfp->dfp_list, &tp->t_dfops);
+	list_add_tail(&dfp->dfp_list, dfops);
 
 	return dfp;
 }
@@ -846,7 +847,7 @@ xfs_defer_add(
 
 	dfp = xfs_defer_find_last(tp, ops);
 	if (!dfp || !xfs_defer_can_append(dfp, ops))
-		dfp = xfs_defer_alloc(tp, ops);
+		dfp = xfs_defer_alloc(&tp->t_dfops, ops);
 
 	xfs_defer_add_item(dfp, li);
 	trace_xfs_defer_add_item(tp->t_mountp, dfp, li);
@@ -870,7 +871,7 @@ xfs_defer_add_barrier(
 	if (dfp)
 		return;
 
-	xfs_defer_alloc(tp, &xfs_barrier_defer_type);
+	xfs_defer_alloc(&tp->t_dfops, &xfs_barrier_defer_type);
 
 	trace_xfs_defer_add_item(tp->t_mountp, dfp, NULL);
 }
@@ -885,14 +886,9 @@ xfs_defer_start_recovery(
 	struct list_head		*r_dfops,
 	const struct xfs_defer_op_type	*ops)
 {
-	struct xfs_defer_pending	*dfp;
+	struct xfs_defer_pending	*dfp = xfs_defer_alloc(r_dfops, ops);
 
-	dfp = kmem_cache_zalloc(xfs_defer_pending_cache,
-			GFP_NOFS | __GFP_NOFAIL);
-	dfp->dfp_ops = ops;
 	dfp->dfp_intent = lip;
-	INIT_LIST_HEAD(&dfp->dfp_work);
-	list_add_tail(&dfp->dfp_list, r_dfops);
 }
 
 /*
@@ -979,7 +975,7 @@ xfs_defer_ops_capture(
 		return ERR_PTR(error);
 
 	/* Create an object to capture the defer ops. */
-	dfc = kmem_zalloc(sizeof(*dfc), KM_NOFS);
+	dfc = kzalloc(sizeof(*dfc), GFP_KERNEL | __GFP_NOFAIL);
 	INIT_LIST_HEAD(&dfc->dfc_list);
 	INIT_LIST_HEAD(&dfc->dfc_dfops);
 
@@ -1011,7 +1007,7 @@ xfs_defer_ops_capture(
 	 * transaction.
 	 */
 	for (i = 0; i < dfc->dfc_held.dr_inos; i++) {
-		ASSERT(xfs_isilocked(dfc->dfc_held.dr_ip[i], XFS_ILOCK_EXCL));
+		xfs_assert_ilocked(dfc->dfc_held.dr_ip[i], XFS_ILOCK_EXCL);
 		ihold(VFS_I(dfc->dfc_held.dr_ip[i]));
 	}
 
@@ -1038,7 +1034,7 @@ xfs_defer_ops_capture_abort(
 	for (i = 0; i < dfc->dfc_held.dr_inos; i++)
 		xfs_irele(dfc->dfc_held.dr_ip[i]);
 
-	kmem_free(dfc);
+	kfree(dfc);
 }
 
 /*
@@ -1096,7 +1092,11 @@ xfs_defer_ops_continue(
 	ASSERT(!(tp->t_flags & XFS_TRANS_DIRTY));
 
 	/* Lock the captured resources to the new transaction. */
-	if (dfc->dfc_held.dr_inos == 2)
+	if (dfc->dfc_held.dr_inos > 2) {
+		xfs_sort_inodes(dfc->dfc_held.dr_ip, dfc->dfc_held.dr_inos);
+		xfs_lock_inodes(dfc->dfc_held.dr_ip, dfc->dfc_held.dr_inos,
+				XFS_ILOCK_EXCL);
+	} else if (dfc->dfc_held.dr_inos == 2)
 		xfs_lock_two_inodes(dfc->dfc_held.dr_ip[0], XFS_ILOCK_EXCL,
 				    dfc->dfc_held.dr_ip[1], XFS_ILOCK_EXCL);
 	else if (dfc->dfc_held.dr_inos == 1)
@@ -1114,7 +1114,7 @@ xfs_defer_ops_continue(
 	list_splice_init(&dfc->dfc_dfops, &tp->t_dfops);
 	tp->t_flags |= dfc->dfc_tpflags;
 
-	kmem_free(dfc);
+	kfree(dfc);
 }
 
 /* Release the resources captured and continued during recovery. */
@@ -1181,6 +1181,10 @@ xfs_defer_init_item_caches(void)
 	error = xfs_attr_intent_init_cache();
 	if (error)
 		goto err;
+	error = xfs_exchmaps_intent_init_cache();
+	if (error)
+		goto err;
+
 	return 0;
 err:
 	xfs_defer_destroy_item_caches();
@@ -1191,6 +1195,7 @@ err:
 void
 xfs_defer_destroy_item_caches(void)
 {
+	xfs_exchmaps_intent_destroy_cache();
 	xfs_attr_intent_destroy_cache();
 	xfs_extfree_intent_destroy_cache();
 	xfs_bmap_intent_destroy_cache();
diff --git a/fs/xfs/libxfs/xfs_defer.h b/fs/xfs/libxfs/xfs_defer.h
index 18a9fb92dd..8b338031e4 100644
--- a/fs/xfs/libxfs/xfs_defer.h
+++ b/fs/xfs/libxfs/xfs_defer.h
@@ -72,12 +72,18 @@ extern const struct xfs_defer_op_type xfs_rmap_update_defer_type;
 extern const struct xfs_defer_op_type xfs_extent_free_defer_type;
 extern const struct xfs_defer_op_type xfs_agfl_free_defer_type;
 extern const struct xfs_defer_op_type xfs_attr_defer_type;
-
+extern const struct xfs_defer_op_type xfs_exchmaps_defer_type;
 
 /*
  * Deferred operation item relogging limits.
  */
-#define XFS_DEFER_OPS_NR_INODES	2	/* join up to two inodes */
+
+/*
+ * Rename w/ parent pointers can require up to 5 inodes with deferred ops to
+ * be joined to the transaction: src_dp, target_dp, src_ip, target_ip, and wip.
+ * These inodes are locked in sorted order by their inode numbers
+ */
+#define XFS_DEFER_OPS_NR_INODES	5
 #define XFS_DEFER_OPS_NR_BUFS	2	/* join up to two buffers */
 
 /* Resources that must be held across a transaction roll. */
diff --git a/fs/xfs/libxfs/xfs_dir2.c b/fs/xfs/libxfs/xfs_dir2.c
index a766732815..457f9a38f8 100644
--- a/fs/xfs/libxfs/xfs_dir2.c
+++ b/fs/xfs/libxfs/xfs_dir2.c
@@ -18,6 +18,7 @@
 #include "xfs_errortag.h"
 #include "xfs_error.h"
 #include "xfs_trace.h"
+#include "xfs_health.h"
 
 const struct xfs_name xfs_name_dotdot = {
 	.name	= (const unsigned char *)"..",
@@ -25,6 +26,12 @@ const struct xfs_name xfs_name_dotdot = {
 	.type	= XFS_DIR3_FT_DIR,
 };
 
+const struct xfs_name xfs_name_dot = {
+	.name	= (const unsigned char *)".",
+	.len	= 1,
+	.type	= XFS_DIR3_FT_DIR,
+};
+
 /*
  * Convert inode mode to directory entry filetype
  */
@@ -104,13 +111,13 @@ xfs_da_mount(
 	ASSERT(mp->m_sb.sb_versionnum & XFS_SB_VERSION_DIRV2BIT);
 	ASSERT(xfs_dir2_dirblock_bytes(&mp->m_sb) <= XFS_MAX_BLOCKSIZE);
 
-	mp->m_dir_geo = kmem_zalloc(sizeof(struct xfs_da_geometry),
-				    KM_MAYFAIL);
-	mp->m_attr_geo = kmem_zalloc(sizeof(struct xfs_da_geometry),
-				     KM_MAYFAIL);
+	mp->m_dir_geo = kzalloc(sizeof(struct xfs_da_geometry),
+				GFP_KERNEL | __GFP_RETRY_MAYFAIL);
+	mp->m_attr_geo = kzalloc(sizeof(struct xfs_da_geometry),
+				GFP_KERNEL | __GFP_RETRY_MAYFAIL);
 	if (!mp->m_dir_geo || !mp->m_attr_geo) {
-		kmem_free(mp->m_dir_geo);
-		kmem_free(mp->m_attr_geo);
+		kfree(mp->m_dir_geo);
+		kfree(mp->m_attr_geo);
 		return -ENOMEM;
 	}
 
@@ -178,8 +185,8 @@ void
 xfs_da_unmount(
 	struct xfs_mount	*mp)
 {
-	kmem_free(mp->m_dir_geo);
-	kmem_free(mp->m_attr_geo);
+	kfree(mp->m_dir_geo);
+	kfree(mp->m_attr_geo);
 }
 
 /*
@@ -236,18 +243,75 @@ xfs_dir_init(
 	if (error)
 		return error;
 
-	args = kmem_zalloc(sizeof(*args), KM_NOFS);
+	args = kzalloc(sizeof(*args), GFP_KERNEL | __GFP_NOFAIL);
 	if (!args)
 		return -ENOMEM;
 
 	args->geo = dp->i_mount->m_dir_geo;
 	args->dp = dp;
 	args->trans = tp;
+	args->owner = dp->i_ino;
 	error = xfs_dir2_sf_create(args, pdp->i_ino);
-	kmem_free(args);
+	kfree(args);
 	return error;
 }
 
+enum xfs_dir2_fmt
+xfs_dir2_format(
+	struct xfs_da_args	*args,
+	int			*error)
+{
+	struct xfs_inode	*dp = args->dp;
+	struct xfs_mount	*mp = dp->i_mount;
+	struct xfs_da_geometry	*geo = mp->m_dir_geo;
+	xfs_fileoff_t		eof;
+
+	xfs_assert_ilocked(dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL);
+
+	*error = 0;
+	if (dp->i_df.if_format == XFS_DINODE_FMT_LOCAL)
+		return XFS_DIR2_FMT_SF;
+
+	*error = xfs_bmap_last_offset(dp, &eof, XFS_DATA_FORK);
+	if (*error)
+		return XFS_DIR2_FMT_ERROR;
+
+	if (eof == XFS_B_TO_FSB(mp, geo->blksize)) {
+		if (XFS_IS_CORRUPT(mp, dp->i_disk_size != geo->blksize)) {
+			xfs_da_mark_sick(args);
+			*error = -EFSCORRUPTED;
+			return XFS_DIR2_FMT_ERROR;
+		}
+		return XFS_DIR2_FMT_BLOCK;
+	}
+	if (eof == geo->leafblk + geo->fsbcount)
+		return XFS_DIR2_FMT_LEAF;
+	return XFS_DIR2_FMT_NODE;
+}
+
+int
+xfs_dir_createname_args(
+	struct xfs_da_args	*args)
+{
+	int			error;
+
+	if (!args->inumber)
+		args->op_flags |= XFS_DA_OP_JUSTCHECK;
+
+	switch (xfs_dir2_format(args, &error)) {
+	case XFS_DIR2_FMT_SF:
+		return xfs_dir2_sf_addname(args);
+	case XFS_DIR2_FMT_BLOCK:
+		return xfs_dir2_block_addname(args);
+	case XFS_DIR2_FMT_LEAF:
+		return xfs_dir2_leaf_addname(args);
+	case XFS_DIR2_FMT_NODE:
+		return xfs_dir2_node_addname(args);
+	default:
+		return error;
+	}
+}
+
 /*
  * Enter a name in a directory, or check for available space.
  * If inum is 0, only the available space test is performed.
@@ -262,7 +326,6 @@ xfs_dir_createname(
 {
 	struct xfs_da_args	*args;
 	int			rval;
-	bool			v;
 
 	ASSERT(S_ISDIR(VFS_I(dp)->i_mode));
 
@@ -273,7 +336,7 @@ xfs_dir_createname(
 		XFS_STATS_INC(dp->i_mount, xs_dir_create);
 	}
 
-	args = kmem_zalloc(sizeof(*args), KM_NOFS);
+	args = kzalloc(sizeof(*args), GFP_KERNEL | __GFP_NOFAIL);
 	if (!args)
 		return -ENOMEM;
 
@@ -288,32 +351,10 @@ xfs_dir_createname(
 	args->whichfork = XFS_DATA_FORK;
 	args->trans = tp;
 	args->op_flags = XFS_DA_OP_ADDNAME | XFS_DA_OP_OKNOENT;
-	if (!inum)
-		args->op_flags |= XFS_DA_OP_JUSTCHECK;
+	args->owner = dp->i_ino;
 
-	if (dp->i_df.if_format == XFS_DINODE_FMT_LOCAL) {
-		rval = xfs_dir2_sf_addname(args);
-		goto out_free;
-	}
-
-	rval = xfs_dir2_isblock(args, &v);
-	if (rval)
-		goto out_free;
-	if (v) {
-		rval = xfs_dir2_block_addname(args);
-		goto out_free;
-	}
-
-	rval = xfs_dir2_isleaf(args, &v);
-	if (rval)
-		goto out_free;
-	if (v)
-		rval = xfs_dir2_leaf_addname(args);
-	else
-		rval = xfs_dir2_node_addname(args);
-
-out_free:
-	kmem_free(args);
+	rval = xfs_dir_createname_args(args);
+	kfree(args);
 	return rval;
 }
 
@@ -333,7 +374,8 @@ xfs_dir_cilookup_result(
 					!(args->op_flags & XFS_DA_OP_CILOOKUP))
 		return -EEXIST;
 
-	args->value = kmem_alloc(len, KM_NOFS | KM_MAYFAIL);
+	args->value = kmalloc(len,
+			GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_RETRY_MAYFAIL);
 	if (!args->value)
 		return -ENOMEM;
 
@@ -342,6 +384,34 @@ xfs_dir_cilookup_result(
 	return -EEXIST;
 }
 
+int
+xfs_dir_lookup_args(
+	struct xfs_da_args	*args)
+{
+	int			error;
+
+	switch (xfs_dir2_format(args, &error)) {
+	case XFS_DIR2_FMT_SF:
+		error = xfs_dir2_sf_lookup(args);
+		break;
+	case XFS_DIR2_FMT_BLOCK:
+		error = xfs_dir2_block_lookup(args);
+		break;
+	case XFS_DIR2_FMT_LEAF:
+		error = xfs_dir2_leaf_lookup(args);
+		break;
+	case XFS_DIR2_FMT_NODE:
+		error = xfs_dir2_node_lookup(args);
+		break;
+	default:
+		break;
+	}
+
+	if (error != -EEXIST)
+		return error;
+	return 0;
+}
+
 /*
  * Lookup a name in a directory, give back the inode number.
  * If ci_name is not NULL, returns the actual name in ci_name if it differs
@@ -358,21 +428,13 @@ xfs_dir_lookup(
 {
 	struct xfs_da_args	*args;
 	int			rval;
-	bool			v;
 	int			lock_mode;
 
 	ASSERT(S_ISDIR(VFS_I(dp)->i_mode));
 	XFS_STATS_INC(dp->i_mount, xs_dir_lookup);
 
-	/*
-	 * We need to use KM_NOFS here so that lockdep will not throw false
-	 * positive deadlock warnings on a non-transactional lookup path. It is
-	 * safe to recurse into inode recalim in that case, but lockdep can't
-	 * easily be taught about it. Hence KM_NOFS avoids having to add more
-	 * lockdep Doing this avoids having to add a bunch of lockdep class
-	 * annotations into the reclaim path for the ilock.
-	 */
-	args = kmem_zalloc(sizeof(*args), KM_NOFS);
+	args = kzalloc(sizeof(*args),
+			GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL);
 	args->geo = dp->i_mount->m_dir_geo;
 	args->name = name->name;
 	args->namelen = name->len;
@@ -382,34 +444,12 @@ xfs_dir_lookup(
 	args->whichfork = XFS_DATA_FORK;
 	args->trans = tp;
 	args->op_flags = XFS_DA_OP_OKNOENT;
+	args->owner = dp->i_ino;
 	if (ci_name)
 		args->op_flags |= XFS_DA_OP_CILOOKUP;
 
 	lock_mode = xfs_ilock_data_map_shared(dp);
-	if (dp->i_df.if_format == XFS_DINODE_FMT_LOCAL) {
-		rval = xfs_dir2_sf_lookup(args);
-		goto out_check_rval;
-	}
-
-	rval = xfs_dir2_isblock(args, &v);
-	if (rval)
-		goto out_free;
-	if (v) {
-		rval = xfs_dir2_block_lookup(args);
-		goto out_check_rval;
-	}
-
-	rval = xfs_dir2_isleaf(args, &v);
-	if (rval)
-		goto out_free;
-	if (v)
-		rval = xfs_dir2_leaf_lookup(args);
-	else
-		rval = xfs_dir2_node_lookup(args);
-
-out_check_rval:
-	if (rval == -EEXIST)
-		rval = 0;
+	rval = xfs_dir_lookup_args(args);
 	if (!rval) {
 		*inum = args->inumber;
 		if (ci_name) {
@@ -417,12 +457,31 @@ out_check_rval:
 			ci_name->len = args->valuelen;
 		}
 	}
-out_free:
 	xfs_iunlock(dp, lock_mode);
-	kmem_free(args);
+	kfree(args);
 	return rval;
 }
 
+int
+xfs_dir_removename_args(
+	struct xfs_da_args	*args)
+{
+	int			error;
+
+	switch (xfs_dir2_format(args, &error)) {
+	case XFS_DIR2_FMT_SF:
+		return xfs_dir2_sf_removename(args);
+	case XFS_DIR2_FMT_BLOCK:
+		return xfs_dir2_block_removename(args);
+	case XFS_DIR2_FMT_LEAF:
+		return xfs_dir2_leaf_removename(args);
+	case XFS_DIR2_FMT_NODE:
+		return xfs_dir2_node_removename(args);
+	default:
+		return error;
+	}
+}
+
 /*
  * Remove an entry from a directory.
  */
@@ -430,18 +489,17 @@ int
 xfs_dir_removename(
 	struct xfs_trans	*tp,
 	struct xfs_inode	*dp,
-	struct xfs_name		*name,
+	const struct xfs_name	*name,
 	xfs_ino_t		ino,
 	xfs_extlen_t		total)		/* bmap's total block count */
 {
 	struct xfs_da_args	*args;
 	int			rval;
-	bool			v;
 
 	ASSERT(S_ISDIR(VFS_I(dp)->i_mode));
 	XFS_STATS_INC(dp->i_mount, xs_dir_remove);
 
-	args = kmem_zalloc(sizeof(*args), KM_NOFS);
+	args = kzalloc(sizeof(*args), GFP_KERNEL | __GFP_NOFAIL);
 	if (!args)
 		return -ENOMEM;
 
@@ -455,30 +513,30 @@ xfs_dir_removename(
 	args->total = total;
 	args->whichfork = XFS_DATA_FORK;
 	args->trans = tp;
+	args->owner = dp->i_ino;
+	rval = xfs_dir_removename_args(args);
+	kfree(args);
+	return rval;
+}
 
-	if (dp->i_df.if_format == XFS_DINODE_FMT_LOCAL) {
-		rval = xfs_dir2_sf_removename(args);
-		goto out_free;
-	}
+int
+xfs_dir_replace_args(
+	struct xfs_da_args	*args)
+{
+	int			error;
 
-	rval = xfs_dir2_isblock(args, &v);
-	if (rval)
-		goto out_free;
-	if (v) {
-		rval = xfs_dir2_block_removename(args);
-		goto out_free;
+	switch (xfs_dir2_format(args, &error)) {
+	case XFS_DIR2_FMT_SF:
+		return xfs_dir2_sf_replace(args);
+	case XFS_DIR2_FMT_BLOCK:
+		return xfs_dir2_block_replace(args);
+	case XFS_DIR2_FMT_LEAF:
+		return xfs_dir2_leaf_replace(args);
+	case XFS_DIR2_FMT_NODE:
+		return xfs_dir2_node_replace(args);
+	default:
+		return error;
 	}
-
-	rval = xfs_dir2_isleaf(args, &v);
-	if (rval)
-		goto out_free;
-	if (v)
-		rval = xfs_dir2_leaf_removename(args);
-	else
-		rval = xfs_dir2_node_removename(args);
-out_free:
-	kmem_free(args);
-	return rval;
 }
 
 /*
@@ -494,7 +552,6 @@ xfs_dir_replace(
 {
 	struct xfs_da_args	*args;
 	int			rval;
-	bool			v;
 
 	ASSERT(S_ISDIR(VFS_I(dp)->i_mode));
 
@@ -502,7 +559,7 @@ xfs_dir_replace(
 	if (rval)
 		return rval;
 
-	args = kmem_zalloc(sizeof(*args), KM_NOFS);
+	args = kzalloc(sizeof(*args), GFP_KERNEL | __GFP_NOFAIL);
 	if (!args)
 		return -ENOMEM;
 
@@ -516,29 +573,9 @@ xfs_dir_replace(
 	args->total = total;
 	args->whichfork = XFS_DATA_FORK;
 	args->trans = tp;
-
-	if (dp->i_df.if_format == XFS_DINODE_FMT_LOCAL) {
-		rval = xfs_dir2_sf_replace(args);
-		goto out_free;
-	}
-
-	rval = xfs_dir2_isblock(args, &v);
-	if (rval)
-		goto out_free;
-	if (v) {
-		rval = xfs_dir2_block_replace(args);
-		goto out_free;
-	}
-
-	rval = xfs_dir2_isleaf(args, &v);
-	if (rval)
-		goto out_free;
-	if (v)
-		rval = xfs_dir2_leaf_replace(args);
-	else
-		rval = xfs_dir2_node_replace(args);
-out_free:
-	kmem_free(args);
+	args->owner = dp->i_ino;
+	rval = xfs_dir_replace_args(args);
+	kfree(args);
 	return rval;
 }
 
@@ -606,55 +643,6 @@ xfs_dir2_grow_inode(
 }
 
 /*
- * See if the directory is a single-block form directory.
- */
-int
-xfs_dir2_isblock(
-	struct xfs_da_args	*args,
-	bool			*isblock)
-{
-	struct xfs_mount	*mp = args->dp->i_mount;
-	xfs_fileoff_t		eof;
-	int			error;
-
-	error = xfs_bmap_last_offset(args->dp, &eof, XFS_DATA_FORK);
-	if (error)
-		return error;
-
-	*isblock = false;
-	if (XFS_FSB_TO_B(mp, eof) != args->geo->blksize)
-		return 0;
-
-	*isblock = true;
-	if (XFS_IS_CORRUPT(mp, args->dp->i_disk_size != args->geo->blksize))
-		return -EFSCORRUPTED;
-	return 0;
-}
-
-/*
- * See if the directory is a single-leaf form directory.
- */
-int
-xfs_dir2_isleaf(
-	struct xfs_da_args	*args,
-	bool			*isleaf)
-{
-	xfs_fileoff_t		eof;
-	int			error;
-
-	error = xfs_bmap_last_offset(args->dp, &eof, XFS_DATA_FORK);
-	if (error)
-		return error;
-
-	*isleaf = false;
-	if (eof != args->geo->leafblk + args->geo->fsbcount)
-		return 0;
-
-	*isleaf = true;
-	return 0;
-}
-
-/*
  * Remove the given block from the directory.
  * This routine is used for data and free blocks, leaf/node are done
  * by xfs_da_shrink_inode.
diff --git a/fs/xfs/libxfs/xfs_dir2.h b/fs/xfs/libxfs/xfs_dir2.h
index 19af22a16c..6dbe6e9ecb 100644
--- a/fs/xfs/libxfs/xfs_dir2.h
+++ b/fs/xfs/libxfs/xfs_dir2.h
@@ -22,6 +22,29 @@ struct xfs_dir3_icfree_hdr;
 struct xfs_dir3_icleaf_hdr;
 
 extern const struct xfs_name	xfs_name_dotdot;
+extern const struct xfs_name	xfs_name_dot;
+
+static inline bool
+xfs_dir2_samename(
+	const struct xfs_name	*n1,
+	const struct xfs_name	*n2)
+{
+	if (n1 == n2)
+		return true;
+	if (n1->len != n2->len)
+		return false;
+	return !memcmp(n1->name, n2->name, n1->len);
+}
+
+enum xfs_dir2_fmt {
+	XFS_DIR2_FMT_SF,
+	XFS_DIR2_FMT_BLOCK,
+	XFS_DIR2_FMT_LEAF,
+	XFS_DIR2_FMT_NODE,
+	XFS_DIR2_FMT_ERROR,
+};
+
+enum xfs_dir2_fmt xfs_dir2_format(struct xfs_da_args *args, int *error);
 
 /*
  * Convert inode mode to directory entry filetype
@@ -45,7 +68,7 @@ extern int xfs_dir_lookup(struct xfs_trans *tp, struct xfs_inode *dp,
 				const struct xfs_name *name, xfs_ino_t *inum,
 				struct xfs_name *ci_name);
 extern int xfs_dir_removename(struct xfs_trans *tp, struct xfs_inode *dp,
-				struct xfs_name *name, xfs_ino_t ino,
+				const struct xfs_name *name, xfs_ino_t ino,
 				xfs_extlen_t tot);
 extern int xfs_dir_replace(struct xfs_trans *tp, struct xfs_inode *dp,
 				const struct xfs_name *name, xfs_ino_t inum,
@@ -53,6 +76,11 @@ extern int xfs_dir_replace(struct xfs_trans *tp, struct xfs_inode *dp,
 extern int xfs_dir_canenter(struct xfs_trans *tp, struct xfs_inode *dp,
 				struct xfs_name *name);
 
+int xfs_dir_lookup_args(struct xfs_da_args *args);
+int xfs_dir_createname_args(struct xfs_da_args *args);
+int xfs_dir_removename_args(struct xfs_da_args *args);
+int xfs_dir_replace_args(struct xfs_da_args *args);
+
 /*
  * Direct call from the bmap code, bypassing the generic directory layer.
  */
@@ -61,8 +89,6 @@ extern int xfs_dir2_sf_to_block(struct xfs_da_args *args);
 /*
  * Interface routines used by userspace utilities
  */
-extern int xfs_dir2_isblock(struct xfs_da_args *args, bool *isblock);
-extern int xfs_dir2_isleaf(struct xfs_da_args *args, bool *isleaf);
 extern int xfs_dir2_shrink_inode(struct xfs_da_args *args, xfs_dir2_db_t db,
 				struct xfs_buf *bp);
 
@@ -88,6 +114,10 @@ extern struct xfs_dir2_data_free *xfs_dir2_data_freefind(
 
 extern int xfs_dir_ino_validate(struct xfs_mount *mp, xfs_ino_t ino);
 
+xfs_failaddr_t xfs_dir3_leaf_header_check(struct xfs_buf *bp, xfs_ino_t owner);
+xfs_failaddr_t xfs_dir3_data_header_check(struct xfs_buf *bp, xfs_ino_t owner);
+xfs_failaddr_t xfs_dir3_block_header_check(struct xfs_buf *bp, xfs_ino_t owner);
+
 extern const struct xfs_buf_ops xfs_dir3_block_buf_ops;
 extern const struct xfs_buf_ops xfs_dir3_leafn_buf_ops;
 extern const struct xfs_buf_ops xfs_dir3_leaf1_buf_ops;
diff --git a/fs/xfs/libxfs/xfs_dir2_block.c b/fs/xfs/libxfs/xfs_dir2_block.c
index 3c256d4cc4..0f93ed1a4a 100644
--- a/fs/xfs/libxfs/xfs_dir2_block.c
+++ b/fs/xfs/libxfs/xfs_dir2_block.c
@@ -20,6 +20,7 @@
 #include "xfs_error.h"
 #include "xfs_trace.h"
 #include "xfs_log.h"
+#include "xfs_health.h"
 
 /*
  * Local function prototypes.
@@ -114,17 +115,20 @@ const struct xfs_buf_ops xfs_dir3_block_buf_ops = {
 	.verify_struct = xfs_dir3_block_verify,
 };
 
-static xfs_failaddr_t
+xfs_failaddr_t
 xfs_dir3_block_header_check(
-	struct xfs_inode	*dp,
-	struct xfs_buf		*bp)
+	struct xfs_buf		*bp,
+	xfs_ino_t		owner)
 {
-	struct xfs_mount	*mp = dp->i_mount;
+	struct xfs_mount	*mp = bp->b_mount;
 
 	if (xfs_has_crc(mp)) {
 		struct xfs_dir3_blk_hdr *hdr3 = bp->b_addr;
 
-		if (be64_to_cpu(hdr3->owner) != dp->i_ino)
+		if (hdr3->magic != cpu_to_be32(XFS_DIR3_BLOCK_MAGIC))
+			return __this_address;
+
+		if (be64_to_cpu(hdr3->owner) != owner)
 			return __this_address;
 	}
 
@@ -135,6 +139,7 @@ int
 xfs_dir3_block_read(
 	struct xfs_trans	*tp,
 	struct xfs_inode	*dp,
+	xfs_ino_t		owner,
 	struct xfs_buf		**bpp)
 {
 	struct xfs_mount	*mp = dp->i_mount;
@@ -147,11 +152,12 @@ xfs_dir3_block_read(
 		return err;
 
 	/* Check things that we can't do in the verifier. */
-	fa = xfs_dir3_block_header_check(dp, *bpp);
+	fa = xfs_dir3_block_header_check(*bpp, owner);
 	if (fa) {
 		__xfs_buf_mark_corrupt(*bpp, fa);
 		xfs_trans_brelse(tp, *bpp);
 		*bpp = NULL;
+		xfs_dirattr_mark_sick(dp, XFS_DATA_FORK);
 		return -EFSCORRUPTED;
 	}
 
@@ -161,12 +167,13 @@ xfs_dir3_block_read(
 
 static void
 xfs_dir3_block_init(
-	struct xfs_mount	*mp,
-	struct xfs_trans	*tp,
-	struct xfs_buf		*bp,
-	struct xfs_inode	*dp)
+	struct xfs_da_args	*args,
+	struct xfs_buf		*bp)
 {
-	struct xfs_dir3_blk_hdr *hdr3 = bp->b_addr;
+	struct xfs_trans	*tp = args->trans;
+	struct xfs_inode	*dp = args->dp;
+	struct xfs_mount	*mp = dp->i_mount;
+	struct xfs_dir3_blk_hdr	*hdr3 = bp->b_addr;
 
 	bp->b_ops = &xfs_dir3_block_buf_ops;
 	xfs_trans_buf_set_type(tp, bp, XFS_BLFT_DIR_BLOCK_BUF);
@@ -175,7 +182,7 @@ xfs_dir3_block_init(
 		memset(hdr3, 0, sizeof(*hdr3));
 		hdr3->magic = cpu_to_be32(XFS_DIR3_BLOCK_MAGIC);
 		hdr3->blkno = cpu_to_be64(xfs_buf_daddr(bp));
-		hdr3->owner = cpu_to_be64(dp->i_ino);
+		hdr3->owner = cpu_to_be64(args->owner);
 		uuid_copy(&hdr3->uuid, &mp->m_sb.sb_meta_uuid);
 		return;
 
@@ -380,7 +387,7 @@ xfs_dir2_block_addname(
 	tp = args->trans;
 
 	/* Read the (one and only) directory block into bp. */
-	error = xfs_dir3_block_read(tp, dp, &bp);
+	error = xfs_dir3_block_read(tp, dp, args->owner, &bp);
 	if (error)
 		return error;
 
@@ -695,7 +702,7 @@ xfs_dir2_block_lookup_int(
 	dp = args->dp;
 	tp = args->trans;
 
-	error = xfs_dir3_block_read(tp, dp, &bp);
+	error = xfs_dir3_block_read(tp, dp, args->owner, &bp);
 	if (error)
 		return error;
 
@@ -979,7 +986,8 @@ xfs_dir2_leaf_to_block(
 	 * Read the data block if we don't already have it, give up if it fails.
 	 */
 	if (!dbp) {
-		error = xfs_dir3_data_read(tp, dp, args->geo->datablk, 0, &dbp);
+		error = xfs_dir3_data_read(tp, dp, args->owner,
+				args->geo->datablk, 0, &dbp);
 		if (error)
 			return error;
 	}
@@ -1007,7 +1015,7 @@ xfs_dir2_leaf_to_block(
 	/*
 	 * Start converting it to block form.
 	 */
-	xfs_dir3_block_init(mp, tp, dbp, dp);
+	xfs_dir3_block_init(args, dbp);
 
 	needlog = 1;
 	needscan = 0;
@@ -1108,7 +1116,7 @@ xfs_dir2_sf_to_block(
 	 * Copy the directory into a temporary buffer.
 	 * Then pitch the incore inode data so we can make extents.
 	 */
-	sfp = kmem_alloc(ifp->if_bytes, 0);
+	sfp = kmalloc(ifp->if_bytes, GFP_KERNEL | __GFP_NOFAIL);
 	memcpy(sfp, oldsfp, ifp->if_bytes);
 
 	xfs_idata_realloc(dp, -ifp->if_bytes, XFS_DATA_FORK);
@@ -1127,7 +1135,7 @@ xfs_dir2_sf_to_block(
 	error = xfs_dir3_data_init(args, blkno, &bp);
 	if (error)
 		goto out_free;
-	xfs_dir3_block_init(mp, tp, bp, dp);
+	xfs_dir3_block_init(args, bp);
 	hdr = bp->b_addr;
 
 	/*
@@ -1167,7 +1175,7 @@ xfs_dir2_sf_to_block(
 	 * Create entry for .
 	 */
 	dep = bp->b_addr + offset;
-	dep->inumber = cpu_to_be64(dp->i_ino);
+	dep->inumber = cpu_to_be64(args->owner);
 	dep->namelen = 1;
 	dep->name[0] = '.';
 	xfs_dir2_data_put_ftype(mp, dep, XFS_DIR3_FT_DIR);
@@ -1253,7 +1261,7 @@ xfs_dir2_sf_to_block(
 			sfep = xfs_dir2_sf_nextentry(mp, sfp, sfep);
 	}
 	/* Done with the temporary buffer */
-	kmem_free(sfp);
+	kfree(sfp);
 	/*
 	 * Sort the leaf entries by hash value.
 	 */
@@ -1268,6 +1276,6 @@ xfs_dir2_sf_to_block(
 	xfs_dir3_data_check(dp, bp);
 	return 0;
 out_free:
-	kmem_free(sfp);
+	kfree(sfp);
 	return error;
 }
diff --git a/fs/xfs/libxfs/xfs_dir2_data.c b/fs/xfs/libxfs/xfs_dir2_data.c
index dbcf58979a..ea0b9628df 100644
--- a/fs/xfs/libxfs/xfs_dir2_data.c
+++ b/fs/xfs/libxfs/xfs_dir2_data.c
@@ -18,6 +18,7 @@
 #include "xfs_trans.h"
 #include "xfs_buf_item.h"
 #include "xfs_log.h"
+#include "xfs_health.h"
 
 static xfs_failaddr_t xfs_dir2_data_freefind_verify(
 		struct xfs_dir2_data_hdr *hdr, struct xfs_dir2_data_free *bf,
@@ -394,17 +395,20 @@ static const struct xfs_buf_ops xfs_dir3_data_reada_buf_ops = {
 	.verify_write = xfs_dir3_data_write_verify,
 };
 
-static xfs_failaddr_t
+xfs_failaddr_t
 xfs_dir3_data_header_check(
-	struct xfs_inode	*dp,
-	struct xfs_buf		*bp)
+	struct xfs_buf		*bp,
+	xfs_ino_t		owner)
 {
-	struct xfs_mount	*mp = dp->i_mount;
+	struct xfs_mount	*mp = bp->b_mount;
 
 	if (xfs_has_crc(mp)) {
 		struct xfs_dir3_data_hdr *hdr3 = bp->b_addr;
 
-		if (be64_to_cpu(hdr3->hdr.owner) != dp->i_ino)
+		if (hdr3->hdr.magic != cpu_to_be32(XFS_DIR3_DATA_MAGIC))
+			return __this_address;
+
+		if (be64_to_cpu(hdr3->hdr.owner) != owner)
 			return __this_address;
 	}
 
@@ -415,6 +419,7 @@ int
 xfs_dir3_data_read(
 	struct xfs_trans	*tp,
 	struct xfs_inode	*dp,
+	xfs_ino_t		owner,
 	xfs_dablk_t		bno,
 	unsigned int		flags,
 	struct xfs_buf		**bpp)
@@ -428,11 +433,12 @@ xfs_dir3_data_read(
 		return err;
 
 	/* Check things that we can't do in the verifier. */
-	fa = xfs_dir3_data_header_check(dp, *bpp);
+	fa = xfs_dir3_data_header_check(*bpp, owner);
 	if (fa) {
 		__xfs_buf_mark_corrupt(*bpp, fa);
 		xfs_trans_brelse(tp, *bpp);
 		*bpp = NULL;
+		xfs_dirattr_mark_sick(dp, XFS_DATA_FORK);
 		return -EFSCORRUPTED;
 	}
 
@@ -723,7 +729,7 @@ xfs_dir3_data_init(
 		memset(hdr3, 0, sizeof(*hdr3));
 		hdr3->magic = cpu_to_be32(XFS_DIR3_DATA_MAGIC);
 		hdr3->blkno = cpu_to_be64(xfs_buf_daddr(bp));
-		hdr3->owner = cpu_to_be64(dp->i_ino);
+		hdr3->owner = cpu_to_be64(args->owner);
 		uuid_copy(&hdr3->uuid, &mp->m_sb.sb_meta_uuid);
 
 	} else
@@ -1198,6 +1204,7 @@ xfs_dir2_data_use_free(
 corrupt:
 	xfs_corruption_error(__func__, XFS_ERRLEVEL_LOW, args->dp->i_mount,
 			hdr, sizeof(*hdr), __FILE__, __LINE__, fa);
+	xfs_da_mark_sick(args);
 	return -EFSCORRUPTED;
 }
 
diff --git a/fs/xfs/libxfs/xfs_dir2_leaf.c b/fs/xfs/libxfs/xfs_dir2_leaf.c
index cb9e950a91..71c2f22a3f 100644
--- a/fs/xfs/libxfs/xfs_dir2_leaf.c
+++ b/fs/xfs/libxfs/xfs_dir2_leaf.c
@@ -19,6 +19,7 @@
 #include "xfs_trace.h"
 #include "xfs_trans.h"
 #include "xfs_buf_item.h"
+#include "xfs_health.h"
 
 /*
  * Local function declarations.
@@ -207,6 +208,29 @@ xfs_dir3_leaf_verify(
 	return xfs_dir3_leaf_check_int(mp, &leafhdr, bp->b_addr, true);
 }
 
+xfs_failaddr_t
+xfs_dir3_leaf_header_check(
+	struct xfs_buf		*bp,
+	xfs_ino_t		owner)
+{
+	struct xfs_mount	*mp = bp->b_mount;
+
+	if (xfs_has_crc(mp)) {
+		struct xfs_dir3_leaf *hdr3 = bp->b_addr;
+
+		if (hdr3->hdr.info.hdr.magic !=
+					cpu_to_be16(XFS_DIR3_LEAF1_MAGIC) &&
+		    hdr3->hdr.info.hdr.magic !=
+					cpu_to_be16(XFS_DIR3_LEAFN_MAGIC))
+			return __this_address;
+
+		if (be64_to_cpu(hdr3->hdr.info.owner) != owner)
+			return __this_address;
+	}
+
+	return NULL;
+}
+
 static void
 xfs_dir3_leaf_read_verify(
 	struct xfs_buf  *bp)
@@ -270,32 +294,60 @@ int
 xfs_dir3_leaf_read(
 	struct xfs_trans	*tp,
 	struct xfs_inode	*dp,
+	xfs_ino_t		owner,
 	xfs_dablk_t		fbno,
 	struct xfs_buf		**bpp)
 {
+	xfs_failaddr_t		fa;
 	int			err;
 
 	err = xfs_da_read_buf(tp, dp, fbno, 0, bpp, XFS_DATA_FORK,
 			&xfs_dir3_leaf1_buf_ops);
-	if (!err && tp && *bpp)
+	if (err || !(*bpp))
+		return err;
+
+	fa = xfs_dir3_leaf_header_check(*bpp, owner);
+	if (fa) {
+		__xfs_buf_mark_corrupt(*bpp, fa);
+		xfs_trans_brelse(tp, *bpp);
+		*bpp = NULL;
+		xfs_dirattr_mark_sick(dp, XFS_DATA_FORK);
+		return -EFSCORRUPTED;
+	}
+
+	if (tp)
 		xfs_trans_buf_set_type(tp, *bpp, XFS_BLFT_DIR_LEAF1_BUF);
-	return err;
+	return 0;
 }
 
 int
 xfs_dir3_leafn_read(
 	struct xfs_trans	*tp,
 	struct xfs_inode	*dp,
+	xfs_ino_t		owner,
 	xfs_dablk_t		fbno,
 	struct xfs_buf		**bpp)
 {
+	xfs_failaddr_t		fa;
 	int			err;
 
 	err = xfs_da_read_buf(tp, dp, fbno, 0, bpp, XFS_DATA_FORK,
 			&xfs_dir3_leafn_buf_ops);
-	if (!err && tp && *bpp)
+	if (err || !(*bpp))
+		return err;
+
+	fa = xfs_dir3_leaf_header_check(*bpp, owner);
+	if (fa) {
+		__xfs_buf_mark_corrupt(*bpp, fa);
+		xfs_trans_brelse(tp, *bpp);
+		*bpp = NULL;
+		xfs_dirattr_mark_sick(dp, XFS_DATA_FORK);
+		return -EFSCORRUPTED;
+	}
+
+	if (tp)
 		xfs_trans_buf_set_type(tp, *bpp, XFS_BLFT_DIR_LEAFN_BUF);
-	return err;
+	return 0;
 }
 
 /*
@@ -303,12 +355,12 @@ xfs_dir3_leafn_read(
  */
 static void
 xfs_dir3_leaf_init(
-	struct xfs_mount	*mp,
-	struct xfs_trans	*tp,
+	struct xfs_da_args	*args,
 	struct xfs_buf		*bp,
-	xfs_ino_t		owner,
 	uint16_t		type)
 {
+	struct xfs_mount	*mp = args->dp->i_mount;
+	struct xfs_trans	*tp = args->trans;
 	struct xfs_dir2_leaf	*leaf = bp->b_addr;
 
 	ASSERT(type == XFS_DIR2_LEAF1_MAGIC || type == XFS_DIR2_LEAFN_MAGIC);
@@ -322,7 +374,7 @@ xfs_dir3_leaf_init(
 					 ? cpu_to_be16(XFS_DIR3_LEAF1_MAGIC)
 					 : cpu_to_be16(XFS_DIR3_LEAFN_MAGIC);
 		leaf3->info.blkno = cpu_to_be64(xfs_buf_daddr(bp));
-		leaf3->info.owner = cpu_to_be64(owner);
+		leaf3->info.owner = cpu_to_be64(args->owner);
 		uuid_copy(&leaf3->info.uuid, &mp->m_sb.sb_meta_uuid);
 	} else {
 		memset(leaf, 0, sizeof(*leaf));
@@ -355,7 +407,6 @@ xfs_dir3_leaf_get_buf(
 {
 	struct xfs_inode	*dp = args->dp;
 	struct xfs_trans	*tp = args->trans;
-	struct xfs_mount	*mp = dp->i_mount;
 	struct xfs_buf		*bp;
 	int			error;
 
@@ -368,7 +419,7 @@ xfs_dir3_leaf_get_buf(
 	if (error)
 		return error;
 
-	xfs_dir3_leaf_init(mp, tp, bp, dp->i_ino, magic);
+	xfs_dir3_leaf_init(args, bp, magic);
 	xfs_dir3_leaf_log_header(args, bp);
 	if (magic == XFS_DIR2_LEAF1_MAGIC)
 		xfs_dir3_leaf_log_tail(args, bp);
@@ -646,7 +697,8 @@ xfs_dir2_leaf_addname(
 
 	trace_xfs_dir2_leaf_addname(args);
 
-	error = xfs_dir3_leaf_read(tp, dp, args->geo->leafblk, &lbp);
+	error = xfs_dir3_leaf_read(tp, dp, args->owner, args->geo->leafblk,
+			&lbp);
 	if (error)
 		return error;
 
@@ -833,9 +885,9 @@ xfs_dir2_leaf_addname(
 		 * Already had space in some data block.
 		 * Just read that one in.
 		 */
-		error = xfs_dir3_data_read(tp, dp,
-				   xfs_dir2_db_to_da(args->geo, use_block),
-				   0, &dbp);
+		error = xfs_dir3_data_read(tp, dp, args->owner,
+				xfs_dir2_db_to_da(args->geo, use_block), 0,
+				&dbp);
 		if (error) {
 			xfs_trans_brelse(tp, lbp);
 			return error;
@@ -1237,7 +1289,8 @@ xfs_dir2_leaf_lookup_int(
 	tp = args->trans;
 	mp = dp->i_mount;
 
-	error = xfs_dir3_leaf_read(tp, dp, args->geo->leafblk, &lbp);
+	error = xfs_dir3_leaf_read(tp, dp, args->owner, args->geo->leafblk,
+			&lbp);
 	if (error)
 		return error;
 
@@ -1275,9 +1328,9 @@ xfs_dir2_leaf_lookup_int(
 		if (newdb != curdb) {
 			if (dbp)
 				xfs_trans_brelse(tp, dbp);
-			error = xfs_dir3_data_read(tp, dp,
-					   xfs_dir2_db_to_da(args->geo, newdb),
-					   0, &dbp);
+			error = xfs_dir3_data_read(tp, dp, args->owner,
+					xfs_dir2_db_to_da(args->geo, newdb), 0,
+					&dbp);
 			if (error) {
 				xfs_trans_brelse(tp, lbp);
 				return error;
@@ -1317,9 +1370,9 @@ xfs_dir2_leaf_lookup_int(
 		ASSERT(cidb != -1);
 		if (cidb != curdb) {
 			xfs_trans_brelse(tp, dbp);
-			error = xfs_dir3_data_read(tp, dp,
-					   xfs_dir2_db_to_da(args->geo, cidb),
-					   0, &dbp);
+			error = xfs_dir3_data_read(tp, dp, args->owner,
+					xfs_dir2_db_to_da(args->geo, cidb), 0,
+					&dbp);
 			if (error) {
 				xfs_trans_brelse(tp, lbp);
 				return error;
@@ -1393,8 +1446,10 @@ xfs_dir2_leaf_removename(
 	bestsp = xfs_dir2_leaf_bests_p(ltp);
 	if (be16_to_cpu(bestsp[db]) != oldbest) {
 		xfs_buf_mark_corrupt(lbp);
+		xfs_da_mark_sick(args);
 		return -EFSCORRUPTED;
 	}
+
 	/*
 	 * Mark the former data entry unused.
 	 */
@@ -1611,7 +1666,8 @@ xfs_dir2_leaf_trim_data(
 	/*
 	 * Read the offending data block.  We need its buffer.
 	 */
-	error = xfs_dir3_data_read(tp, dp, xfs_dir2_db_to_da(geo, db), 0, &dbp);
+	error = xfs_dir3_data_read(tp, dp, args->owner,
+			xfs_dir2_db_to_da(geo, db), 0, &dbp);
 	if (error)
 		return error;
 
@@ -1750,7 +1806,8 @@ xfs_dir2_node_to_leaf(
 	/*
 	 * Read the freespace block.
 	 */
-	error = xfs_dir2_free_read(tp, dp,  args->geo->freeblk, &fbp);
+	error = xfs_dir2_free_read(tp, dp, args->owner, args->geo->freeblk,
+			&fbp);
 	if (error)
 		return error;
 	xfs_dir2_free_hdr_from_disk(mp, &freehdr, fbp->b_addr);
diff --git a/fs/xfs/libxfs/xfs_dir2_node.c b/fs/xfs/libxfs/xfs_dir2_node.c
index 7a03aeb9f4..fe8d4fa131 100644
--- a/fs/xfs/libxfs/xfs_dir2_node.c
+++ b/fs/xfs/libxfs/xfs_dir2_node.c
@@ -20,6 +20,7 @@
 #include "xfs_trans.h"
 #include "xfs_buf_item.h"
 #include "xfs_log.h"
+#include "xfs_health.h"
 
 /*
  * Function declarations.
@@ -174,11 +175,11 @@ const struct xfs_buf_ops xfs_dir3_free_buf_ops = {
 /* Everything ok in the free block header? */
 static xfs_failaddr_t
 xfs_dir3_free_header_check(
-	struct xfs_inode	*dp,
-	xfs_dablk_t		fbno,
-	struct xfs_buf		*bp)
+	struct xfs_buf		*bp,
+	xfs_ino_t		owner,
+	xfs_dablk_t		fbno)
 {
-	struct xfs_mount	*mp = dp->i_mount;
+	struct xfs_mount	*mp = bp->b_mount;
 	int			maxbests = mp->m_dir_geo->free_max_bests;
 	unsigned int		firstdb;
 
@@ -194,7 +195,7 @@ xfs_dir3_free_header_check(
 			return __this_address;
 		if (be32_to_cpu(hdr3->nvalid) < be32_to_cpu(hdr3->nused))
 			return __this_address;
-		if (be64_to_cpu(hdr3->hdr.owner) != dp->i_ino)
+		if (be64_to_cpu(hdr3->hdr.owner) != owner)
 			return __this_address;
 	} else {
 		struct xfs_dir2_free_hdr *hdr = bp->b_addr;
@@ -213,6 +214,7 @@ static int
 __xfs_dir3_free_read(
 	struct xfs_trans	*tp,
 	struct xfs_inode	*dp,
+	xfs_ino_t		owner,
 	xfs_dablk_t		fbno,
 	unsigned int		flags,
 	struct xfs_buf		**bpp)
@@ -226,11 +228,12 @@ __xfs_dir3_free_read(
 		return err;
 
 	/* Check things that we can't do in the verifier. */
-	fa = xfs_dir3_free_header_check(dp, fbno, *bpp);
+	fa = xfs_dir3_free_header_check(*bpp, owner, fbno);
 	if (fa) {
 		__xfs_buf_mark_corrupt(*bpp, fa);
 		xfs_trans_brelse(tp, *bpp);
 		*bpp = NULL;
+		xfs_dirattr_mark_sick(dp, XFS_DATA_FORK);
 		return -EFSCORRUPTED;
 	}
 
@@ -297,20 +300,23 @@ int
 xfs_dir2_free_read(
 	struct xfs_trans	*tp,
 	struct xfs_inode	*dp,
+	xfs_ino_t		owner,
 	xfs_dablk_t		fbno,
 	struct xfs_buf		**bpp)
 {
-	return __xfs_dir3_free_read(tp, dp, fbno, 0, bpp);
+	return __xfs_dir3_free_read(tp, dp, owner, fbno, 0, bpp);
 }
 
 static int
 xfs_dir2_free_try_read(
 	struct xfs_trans	*tp,
 	struct xfs_inode	*dp,
+	xfs_ino_t		owner,
 	xfs_dablk_t		fbno,
 	struct xfs_buf		**bpp)
 {
-	return __xfs_dir3_free_read(tp, dp, fbno, XFS_DABUF_MAP_HOLE_OK, bpp);
+	return __xfs_dir3_free_read(tp, dp, owner, fbno, XFS_DABUF_MAP_HOLE_OK,
+			bpp);
 }
 
 static int
@@ -347,7 +353,7 @@ xfs_dir3_free_get_buf(
 		hdr.magic = XFS_DIR3_FREE_MAGIC;
 
 		hdr3->hdr.blkno = cpu_to_be64(xfs_buf_daddr(bp));
-		hdr3->hdr.owner = cpu_to_be64(dp->i_ino);
+		hdr3->hdr.owner = cpu_to_be64(args->owner);
 		uuid_copy(&hdr3->hdr.uuid, &mp->m_sb.sb_meta_uuid);
 	} else
 		hdr.magic = XFS_DIR2_FREE_MAGIC;
@@ -443,6 +449,7 @@ xfs_dir2_leaf_to_node(
 	if (be32_to_cpu(ltp->bestcount) >
 				(uint)dp->i_disk_size / args->geo->blksize) {
 		xfs_buf_mark_corrupt(lbp);
+		xfs_da_mark_sick(args);
 		return -EFSCORRUPTED;
 	}
 
@@ -517,6 +524,7 @@ xfs_dir2_leafn_add(
 	 */
 	if (index < 0) {
 		xfs_buf_mark_corrupt(bp);
+		xfs_da_mark_sick(args);
 		return -EFSCORRUPTED;
 	}
 
@@ -713,7 +721,7 @@ xfs_dir2_leafn_lookup_for_addname(
 				if (curbp)
 					xfs_trans_brelse(tp, curbp);
 
-				error = xfs_dir2_free_read(tp, dp,
+				error = xfs_dir2_free_read(tp, dp, args->owner,
 						xfs_dir2_db_to_da(args->geo,
 								  newfdb),
 						&curbp);
@@ -736,6 +744,7 @@ xfs_dir2_leafn_lookup_for_addname(
 					   cpu_to_be16(NULLDATAOFF))) {
 				if (curfdb != newfdb)
 					xfs_trans_brelse(tp, curbp);
+				xfs_da_mark_sick(args);
 				return -EFSCORRUPTED;
 			}
 			curfdb = newfdb;
@@ -804,6 +813,7 @@ xfs_dir2_leafn_lookup_for_entry(
 	xfs_dir3_leaf_check(dp, bp);
 	if (leafhdr.count <= 0) {
 		xfs_buf_mark_corrupt(bp);
+		xfs_da_mark_sick(args);
 		return -EFSCORRUPTED;
 	}
 
@@ -857,7 +867,7 @@ xfs_dir2_leafn_lookup_for_entry(
 				ASSERT(state->extravalid);
 				curbp = state->extrablk.bp;
 			} else {
-				error = xfs_dir3_data_read(tp, dp,
+				error = xfs_dir3_data_read(tp, dp, args->owner,
 						xfs_dir2_db_to_da(args->geo,
 								  newdb),
 						0, &curbp);
@@ -1350,8 +1360,8 @@ xfs_dir2_leafn_remove(
 		 * read in the free block.
 		 */
 		fdb = xfs_dir2_db_to_fdb(geo, db);
-		error = xfs_dir2_free_read(tp, dp, xfs_dir2_db_to_da(geo, fdb),
-					   &fbp);
+		error = xfs_dir2_free_read(tp, dp, args->owner,
+				xfs_dir2_db_to_da(geo, fdb), &fbp);
 		if (error)
 			return error;
 		free = fbp->b_addr;
@@ -1556,7 +1566,8 @@ xfs_dir2_leafn_toosmall(
 		/*
 		 * Read the sibling leaf block.
 		 */
-		error = xfs_dir3_leafn_read(state->args->trans, dp, blkno, &bp);
+		error = xfs_dir3_leafn_read(state->args->trans, dp,
+				state->args->owner, blkno, &bp);
 		if (error)
 			return error;
 
@@ -1709,7 +1720,7 @@ xfs_dir2_node_add_datablk(
 	 * that was just allocated.
 	 */
 	fbno = xfs_dir2_db_to_fdb(args->geo, *dbno);
-	error = xfs_dir2_free_try_read(tp, dp,
+	error = xfs_dir2_free_try_read(tp, dp, args->owner,
 			       xfs_dir2_db_to_da(args->geo, fbno), &fbp);
 	if (error)
 		return error;
@@ -1739,6 +1750,7 @@ xfs_dir2_node_add_datablk(
 			} else {
 				xfs_alert(mp, " ... fblk is NULL");
 			}
+			xfs_da_mark_sick(args);
 			return -EFSCORRUPTED;
 		}
 
@@ -1855,7 +1867,7 @@ xfs_dir2_node_find_freeblk(
 		 * so this might not succeed.  This should be really rare, so
 		 * there's no reason to avoid it.
 		 */
-		error = xfs_dir2_free_try_read(tp, dp,
+		error = xfs_dir2_free_try_read(tp, dp, args->owner,
 				xfs_dir2_db_to_da(args->geo, fbno),
 				&fbp);
 		if (error)
@@ -1941,9 +1953,8 @@ xfs_dir2_node_addname_int(
 						  &freehdr, &findex);
 	} else {
 		/* Read the data block in. */
-		error = xfs_dir3_data_read(tp, dp,
-					   xfs_dir2_db_to_da(args->geo, dbno),
-					   0, &dbp);
+		error = xfs_dir3_data_read(tp, dp, args->owner,
+				xfs_dir2_db_to_da(args->geo, dbno), 0, &dbp);
 	}
 	if (error)
 		return error;
@@ -2295,7 +2306,7 @@ xfs_dir2_node_trim_free(
 	/*
 	 * Read the freespace block.
 	 */
-	error = xfs_dir2_free_try_read(tp, dp, fo, &bp);
+	error = xfs_dir2_free_try_read(tp, dp, args->owner, fo, &bp);
 	if (error)
 		return error;
 	/*
diff --git a/fs/xfs/libxfs/xfs_dir2_priv.h b/fs/xfs/libxfs/xfs_dir2_priv.h
index 1db2e60ba8..3befb32509 100644
--- a/fs/xfs/libxfs/xfs_dir2_priv.h
+++ b/fs/xfs/libxfs/xfs_dir2_priv.h
@@ -50,8 +50,8 @@ extern int xfs_dir_cilookup_result(struct xfs_da_args *args,
 
 
 /* xfs_dir2_block.c */
-extern int xfs_dir3_block_read(struct xfs_trans *tp, struct xfs_inode *dp,
-			       struct xfs_buf **bpp);
+int xfs_dir3_block_read(struct xfs_trans *tp, struct xfs_inode *dp,
+		xfs_ino_t owner, struct xfs_buf **bpp);
 extern int xfs_dir2_block_addname(struct xfs_da_args *args);
 extern int xfs_dir2_block_lookup(struct xfs_da_args *args);
 extern int xfs_dir2_block_removename(struct xfs_da_args *args);
@@ -78,7 +78,8 @@ extern void xfs_dir3_data_check(struct xfs_inode *dp, struct xfs_buf *bp);
 extern xfs_failaddr_t __xfs_dir3_data_check(struct xfs_inode *dp,
 		struct xfs_buf *bp);
 int xfs_dir3_data_read(struct xfs_trans *tp, struct xfs_inode *dp,
-		xfs_dablk_t bno, unsigned int flags, struct xfs_buf **bpp);
+		xfs_ino_t owner, xfs_dablk_t bno, unsigned int flags,
+		struct xfs_buf **bpp);
 int xfs_dir3_data_readahead(struct xfs_inode *dp, xfs_dablk_t bno,
 		unsigned int flags);
 
@@ -95,9 +96,9 @@ void xfs_dir2_leaf_hdr_from_disk(struct xfs_mount *mp,
 void xfs_dir2_leaf_hdr_to_disk(struct xfs_mount *mp, struct xfs_dir2_leaf *to,
 		struct xfs_dir3_icleaf_hdr *from);
 int xfs_dir3_leaf_read(struct xfs_trans *tp, struct xfs_inode *dp,
-		xfs_dablk_t fbno, struct xfs_buf **bpp);
+		xfs_ino_t owner, xfs_dablk_t fbno, struct xfs_buf **bpp);
 int xfs_dir3_leafn_read(struct xfs_trans *tp, struct xfs_inode *dp,
-		xfs_dablk_t fbno, struct xfs_buf **bpp);
+		xfs_ino_t owner, xfs_dablk_t fbno, struct xfs_buf **bpp);
 extern int xfs_dir2_block_to_leaf(struct xfs_da_args *args,
 		struct xfs_buf *dbp);
 extern int xfs_dir2_leaf_addname(struct xfs_da_args *args);
@@ -154,8 +155,8 @@ extern int xfs_dir2_node_removename(struct xfs_da_args *args);
 extern int xfs_dir2_node_replace(struct xfs_da_args *args);
 extern int xfs_dir2_node_trim_free(struct xfs_da_args *args, xfs_fileoff_t fo,
 		int *rvalp);
-extern int xfs_dir2_free_read(struct xfs_trans *tp, struct xfs_inode *dp,
-		xfs_dablk_t fbno, struct xfs_buf **bpp);
+int xfs_dir2_free_read(struct xfs_trans *tp, struct xfs_inode *dp,
+		xfs_ino_t owner, xfs_dablk_t fbno, struct xfs_buf **bpp);
 
 /* xfs_dir2_sf.c */
 xfs_ino_t xfs_dir2_sf_get_ino(struct xfs_mount *mp, struct xfs_dir2_sf_hdr *hdr,
diff --git a/fs/xfs/libxfs/xfs_dir2_sf.c b/fs/xfs/libxfs/xfs_dir2_sf.c
index e1f83fc7b6..17a20384c8 100644
--- a/fs/xfs/libxfs/xfs_dir2_sf.c
+++ b/fs/xfs/libxfs/xfs_dir2_sf.c
@@ -276,7 +276,7 @@ xfs_dir2_block_to_sf(
 	 * format the data into.  Once we have formatted the data, we can free
 	 * the block and copy the formatted data into the inode literal area.
 	 */
-	sfp = kmem_alloc(mp->m_sb.sb_inodesize, 0);
+	sfp = kmalloc(mp->m_sb.sb_inodesize, GFP_KERNEL | __GFP_NOFAIL);
 	memcpy(sfp, sfhp, xfs_dir2_sf_hdr_size(sfhp->i8count));
 
 	/*
@@ -350,7 +350,7 @@ xfs_dir2_block_to_sf(
 	xfs_dir2_sf_check(args);
 out:
 	xfs_trans_log_inode(args->trans, dp, logflags);
-	kmem_free(sfp);
+	kfree(sfp);
 	return error;
 }
 
@@ -524,7 +524,7 @@ xfs_dir2_sf_addname_hard(
 	 * Copy the old directory to the stack buffer.
 	 */
 	old_isize = (int)dp->i_disk_size;
-	buf = kmem_alloc(old_isize, 0);
+	buf = kmalloc(old_isize, GFP_KERNEL | __GFP_NOFAIL);
 	oldsfp = (xfs_dir2_sf_hdr_t *)buf;
 	memcpy(oldsfp, dp->i_df.if_data, old_isize);
 	/*
@@ -576,7 +576,7 @@ xfs_dir2_sf_addname_hard(
 		sfep = xfs_dir2_sf_nextentry(mp, sfp, sfep);
 		memcpy(sfep, oldsfep, old_isize - nbytes);
 	}
-	kmem_free(buf);
+	kfree(buf);
 	dp->i_disk_size = new_isize;
 	xfs_dir2_sf_check(args);
 }
@@ -1151,7 +1151,7 @@ xfs_dir2_sf_toino4(
 	 * Don't want xfs_idata_realloc copying the data here.
 	 */
 	oldsize = dp->i_df.if_bytes;
-	buf = kmem_alloc(oldsize, 0);
+	buf = kmalloc(oldsize, GFP_KERNEL | __GFP_NOFAIL);
 	ASSERT(oldsfp->i8count == 1);
 	memcpy(buf, oldsfp, oldsize);
 	/*
@@ -1190,7 +1190,7 @@ xfs_dir2_sf_toino4(
 	/*
 	 * Clean up the inode.
 	 */
-	kmem_free(buf);
+	kfree(buf);
 	dp->i_disk_size = newsize;
 	xfs_trans_log_inode(args->trans, dp, XFS_ILOG_CORE | XFS_ILOG_DDATA);
 }
@@ -1223,7 +1223,7 @@ xfs_dir2_sf_toino8(
 	 * Don't want xfs_idata_realloc copying the data here.
 	 */
 	oldsize = dp->i_df.if_bytes;
-	buf = kmem_alloc(oldsize, 0);
+	buf = kmalloc(oldsize, GFP_KERNEL | __GFP_NOFAIL);
 	ASSERT(oldsfp->i8count == 0);
 	memcpy(buf, oldsfp, oldsize);
 	/*
@@ -1262,7 +1262,7 @@ xfs_dir2_sf_toino8(
 	/*
 	 * Clean up the inode.
 	 */
-	kmem_free(buf);
+	kfree(buf);
 	dp->i_disk_size = newsize;
 	xfs_trans_log_inode(args->trans, dp, XFS_ILOG_CORE | XFS_ILOG_DDATA);
 }
diff --git a/fs/xfs/libxfs/xfs_errortag.h b/fs/xfs/libxfs/xfs_errortag.h
index 01a9e86b30..7002d7676a 100644
--- a/fs/xfs/libxfs/xfs_errortag.h
+++ b/fs/xfs/libxfs/xfs_errortag.h
@@ -63,7 +63,8 @@
 #define XFS_ERRTAG_ATTR_LEAF_TO_NODE			41
 #define XFS_ERRTAG_WB_DELAY_MS				42
 #define XFS_ERRTAG_WRITE_DELAY_MS			43
-#define XFS_ERRTAG_MAX					44
+#define XFS_ERRTAG_EXCHMAPS_FINISH_ONE			44
+#define XFS_ERRTAG_MAX					45
 
 /*
  * Random factors for above tags, 1 means always, 2 means 1/2 time, etc.
@@ -111,5 +112,6 @@
 #define XFS_RANDOM_ATTR_LEAF_TO_NODE			1
 #define XFS_RANDOM_WB_DELAY_MS				3000
 #define XFS_RANDOM_WRITE_DELAY_MS			3000
+#define XFS_RANDOM_EXCHMAPS_FINISH_ONE			1
 
 #endif /* __XFS_ERRORTAG_H_ */
diff --git a/fs/xfs/libxfs/xfs_exchmaps.c b/fs/xfs/libxfs/xfs_exchmaps.c
new file mode 100644
index 0000000000..2021396651
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_exchmaps.c
@@ -0,0 +1,1235 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2020-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_inode.h"
+#include "xfs_trans.h"
+#include "xfs_bmap.h"
+#include "xfs_icache.h"
+#include "xfs_quota.h"
+#include "xfs_exchmaps.h"
+#include "xfs_trace.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_trans_space.h"
+#include "xfs_error.h"
+#include "xfs_errortag.h"
+#include "xfs_health.h"
+#include "xfs_exchmaps_item.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
+#include "xfs_attr_leaf.h"
+#include "xfs_attr.h"
+#include "xfs_dir2_priv.h"
+#include "xfs_dir2.h"
+#include "xfs_symlink_remote.h"
+
+struct kmem_cache	*xfs_exchmaps_intent_cache;
+
+/* bmbt mappings adjacent to a pair of records. */
+struct xfs_exchmaps_adjacent {
+	struct xfs_bmbt_irec		left1;
+	struct xfs_bmbt_irec		right1;
+	struct xfs_bmbt_irec		left2;
+	struct xfs_bmbt_irec		right2;
+};
+
+#define ADJACENT_INIT { \
+	.left1  = { .br_startblock = HOLESTARTBLOCK }, \
+	.right1 = { .br_startblock = HOLESTARTBLOCK }, \
+	.left2  = { .br_startblock = HOLESTARTBLOCK }, \
+	.right2 = { .br_startblock = HOLESTARTBLOCK }, \
+}
+
+/* Information to reset reflink flag / CoW fork state after an exchange. */
+
+/*
+ * If the reflink flag is set on either inode, make sure it has an incore CoW
+ * fork, since all reflink inodes must have them.  If there's a CoW fork and it
+ * has mappings in it, make sure the inodes are tagged appropriately so that
+ * speculative preallocations can be GC'd if we run low of space.
+ */
+static inline void
+xfs_exchmaps_ensure_cowfork(
+	struct xfs_inode	*ip)
+{
+	struct xfs_ifork	*cfork;
+
+	if (xfs_is_reflink_inode(ip))
+		xfs_ifork_init_cow(ip);
+
+	cfork = xfs_ifork_ptr(ip, XFS_COW_FORK);
+	if (!cfork)
+		return;
+	if (cfork->if_bytes > 0)
+		xfs_inode_set_cowblocks_tag(ip);
+	else
+		xfs_inode_clear_cowblocks_tag(ip);
+}
+
+/*
+ * Adjust the on-disk inode size upwards if needed so that we never add
+ * mappings into the file past EOF.  This is crucial so that log recovery won't
+ * get confused by the sudden appearance of post-eof mappings.
+ */
+STATIC void
+xfs_exchmaps_update_size(
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip,
+	struct xfs_bmbt_irec	*imap,
+	xfs_fsize_t		new_isize)
+{
+	struct xfs_mount	*mp = tp->t_mountp;
+	xfs_fsize_t		len;
+
+	if (new_isize < 0)
+		return;
+
+	len = min(XFS_FSB_TO_B(mp, imap->br_startoff + imap->br_blockcount),
+		  new_isize);
+
+	if (len <= ip->i_disk_size)
+		return;
+
+	trace_xfs_exchmaps_update_inode_size(ip, len);
+
+	ip->i_disk_size = len;
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+}
+
+/* Advance the incore state tracking after exchanging a mapping. */
+static inline void
+xmi_advance(
+	struct xfs_exchmaps_intent	*xmi,
+	const struct xfs_bmbt_irec	*irec)
+{
+	xmi->xmi_startoff1 += irec->br_blockcount;
+	xmi->xmi_startoff2 += irec->br_blockcount;
+	xmi->xmi_blockcount -= irec->br_blockcount;
+}
+
+/* Do we still have more mappings to exchange? */
+static inline bool
+xmi_has_more_exchange_work(const struct xfs_exchmaps_intent *xmi)
+{
+	return xmi->xmi_blockcount > 0;
+}
+
+/* Do we have post-operation cleanups to perform? */
+static inline bool
+xmi_has_postop_work(const struct xfs_exchmaps_intent *xmi)
+{
+	return xmi->xmi_flags & (XFS_EXCHMAPS_CLEAR_INO1_REFLINK |
+				 XFS_EXCHMAPS_CLEAR_INO2_REFLINK |
+				 __XFS_EXCHMAPS_INO2_SHORTFORM);
+}
+
+/* Check all mappings to make sure we can actually exchange them. */
+int
+xfs_exchmaps_check_forks(
+	struct xfs_mount		*mp,
+	const struct xfs_exchmaps_req	*req)
+{
+	struct xfs_ifork		*ifp1, *ifp2;
+	int				whichfork = xfs_exchmaps_reqfork(req);
+
+	/* No fork? */
+	ifp1 = xfs_ifork_ptr(req->ip1, whichfork);
+	ifp2 = xfs_ifork_ptr(req->ip2, whichfork);
+	if (!ifp1 || !ifp2)
+		return -EINVAL;
+
+	/* We don't know how to exchange local format forks. */
+	if (ifp1->if_format == XFS_DINODE_FMT_LOCAL ||
+	    ifp2->if_format == XFS_DINODE_FMT_LOCAL)
+		return -EINVAL;
+
+	return 0;
+}
+
+#ifdef CONFIG_XFS_QUOTA
+/* Log the actual updates to the quota accounting. */
+static inline void
+xfs_exchmaps_update_quota(
+	struct xfs_trans		*tp,
+	struct xfs_exchmaps_intent	*xmi,
+	struct xfs_bmbt_irec		*irec1,
+	struct xfs_bmbt_irec		*irec2)
+{
+	int64_t				ip1_delta = 0, ip2_delta = 0;
+	unsigned int			qflag;
+
+	qflag = XFS_IS_REALTIME_INODE(xmi->xmi_ip1) ? XFS_TRANS_DQ_RTBCOUNT :
+						      XFS_TRANS_DQ_BCOUNT;
+
+	if (xfs_bmap_is_real_extent(irec1)) {
+		ip1_delta -= irec1->br_blockcount;
+		ip2_delta += irec1->br_blockcount;
+	}
+
+	if (xfs_bmap_is_real_extent(irec2)) {
+		ip1_delta += irec2->br_blockcount;
+		ip2_delta -= irec2->br_blockcount;
+	}
+
+	xfs_trans_mod_dquot_byino(tp, xmi->xmi_ip1, qflag, ip1_delta);
+	xfs_trans_mod_dquot_byino(tp, xmi->xmi_ip2, qflag, ip2_delta);
+}
+#else
+# define xfs_exchmaps_update_quota(tp, xmi, irec1, irec2)	((void)0)
+#endif
+
+/* Decide if we want to skip this mapping from file1. */
+static inline bool
+xfs_exchmaps_can_skip_mapping(
+	struct xfs_exchmaps_intent	*xmi,
+	struct xfs_bmbt_irec		*irec)
+{
+	struct xfs_mount		*mp = xmi->xmi_ip1->i_mount;
+
+	/* Do not skip this mapping if the caller did not tell us to. */
+	if (!(xmi->xmi_flags & XFS_EXCHMAPS_INO1_WRITTEN))
+		return false;
+
+	/* Do not skip mapped, written mappings. */
+	if (xfs_bmap_is_written_extent(irec))
+		return false;
+
+	/*
+	 * The mapping is unwritten or a hole.  It cannot be a delalloc
+	 * reservation because we already excluded those.  It cannot be an
+	 * unwritten extent with dirty page cache because we flushed the page
+	 * cache.  For files where the allocation unit is 1FSB (files on the
+	 * data dev, rt files if the extent size is 1FSB), we can safely
+	 * skip this mapping.
+	 */
+	if (!xfs_inode_has_bigrtalloc(xmi->xmi_ip1))
+		return true;
+
+	/*
+	 * For a realtime file with a multi-fsb allocation unit, the decision
+	 * is trickier because we can only swap full allocation units.
+	 * Unwritten mappings can appear in the middle of an rtx if the rtx is
+	 * partially written, but they can also appear for preallocations.
+	 *
+	 * If the mapping is a hole, skip it entirely.  Holes should align with
+	 * rtx boundaries.
+	 */
+	if (!xfs_bmap_is_real_extent(irec))
+		return true;
+
+	/*
+	 * All mappings below this point are unwritten.
+	 *
+	 * - If the beginning is not aligned to an rtx, trim the end of the
+	 *   mapping so that it does not cross an rtx boundary, and swap it.
+	 *
+	 * - If both ends are aligned to an rtx, skip the entire mapping.
+	 */
+	if (!isaligned_64(irec->br_startoff, mp->m_sb.sb_rextsize)) {
+		xfs_fileoff_t	new_end;
+
+		new_end = roundup_64(irec->br_startoff, mp->m_sb.sb_rextsize);
+		irec->br_blockcount = min(irec->br_blockcount,
+					  new_end - irec->br_startoff);
+		return false;
+	}
+	if (isaligned_64(irec->br_blockcount, mp->m_sb.sb_rextsize))
+		return true;
+
+	/*
+	 * All mappings below this point are unwritten, start on an rtx
+	 * boundary, and do not end on an rtx boundary.
+	 *
+	 * - If the mapping is longer than one rtx, trim the end of the mapping
+	 *   down to an rtx boundary and skip it.
+	 *
+	 * - The mapping is shorter than one rtx.  Swap it.
+	 */
+	if (irec->br_blockcount > mp->m_sb.sb_rextsize) {
+		xfs_fileoff_t	new_end;
+
+		new_end = rounddown_64(irec->br_startoff + irec->br_blockcount,
+				mp->m_sb.sb_rextsize);
+		irec->br_blockcount = new_end - irec->br_startoff;
+		return true;
+	}
+
+	return false;
+}
+
+/*
+ * Walk forward through the file ranges in @xmi until we find two different
+ * mappings to exchange.  If there is work to do, return the mappings;
+ * otherwise we've reached the end of the range and xmi_blockcount will be
+ * zero.
+ *
+ * If the walk skips over a pair of mappings to the same storage, save them as
+ * the left records in @adj (if provided) so that the simulation phase can
+ * avoid an extra lookup.
+  */
+static int
+xfs_exchmaps_find_mappings(
+	struct xfs_exchmaps_intent	*xmi,
+	struct xfs_bmbt_irec		*irec1,
+	struct xfs_bmbt_irec		*irec2,
+	struct xfs_exchmaps_adjacent	*adj)
+{
+	int				nimaps;
+	int				bmap_flags;
+	int				error;
+
+	bmap_flags = xfs_bmapi_aflag(xfs_exchmaps_whichfork(xmi));
+
+	for (; xmi_has_more_exchange_work(xmi); xmi_advance(xmi, irec1)) {
+		/* Read mapping from the first file */
+		nimaps = 1;
+		error = xfs_bmapi_read(xmi->xmi_ip1, xmi->xmi_startoff1,
+				xmi->xmi_blockcount, irec1, &nimaps,
+				bmap_flags);
+		if (error)
+			return error;
+		if (nimaps != 1 ||
+		    irec1->br_startblock == DELAYSTARTBLOCK ||
+		    irec1->br_startoff != xmi->xmi_startoff1) {
+			/*
+			 * We should never get no mapping or a delalloc mapping
+			 * or something that doesn't match what we asked for,
+			 * since the caller flushed both inodes and we hold the
+			 * ILOCKs for both inodes.
+			 */
+			ASSERT(0);
+			return -EINVAL;
+		}
+
+		if (xfs_exchmaps_can_skip_mapping(xmi, irec1)) {
+			trace_xfs_exchmaps_mapping1_skip(xmi->xmi_ip1, irec1);
+			continue;
+		}
+
+		/* Read mapping from the second file */
+		nimaps = 1;
+		error = xfs_bmapi_read(xmi->xmi_ip2, xmi->xmi_startoff2,
+				irec1->br_blockcount, irec2, &nimaps,
+				bmap_flags);
+		if (error)
+			return error;
+		if (nimaps != 1 ||
+		    irec2->br_startblock == DELAYSTARTBLOCK ||
+		    irec2->br_startoff != xmi->xmi_startoff2) {
+			/*
+			 * We should never get no mapping or a delalloc mapping
+			 * or something that doesn't match what we asked for,
+			 * since the caller flushed both inodes and we hold the
+			 * ILOCKs for both inodes.
+			 */
+			ASSERT(0);
+			return -EINVAL;
+		}
+
+		/*
+		 * We can only exchange as many blocks as the smaller of the
+		 * two mapping maps.
+		 */
+		irec1->br_blockcount = min(irec1->br_blockcount,
+					   irec2->br_blockcount);
+
+		trace_xfs_exchmaps_mapping1(xmi->xmi_ip1, irec1);
+		trace_xfs_exchmaps_mapping2(xmi->xmi_ip2, irec2);
+
+		/* We found something to exchange, so return it. */
+		if (irec1->br_startblock != irec2->br_startblock)
+			return 0;
+
+		/*
+		 * Two mappings pointing to the same physical block must not
+		 * have different states; that's filesystem corruption.  Move
+		 * on to the next mapping if they're both holes or both point
+		 * to the same physical space extent.
+		 */
+		if (irec1->br_state != irec2->br_state) {
+			xfs_bmap_mark_sick(xmi->xmi_ip1,
+					xfs_exchmaps_whichfork(xmi));
+			xfs_bmap_mark_sick(xmi->xmi_ip2,
+					xfs_exchmaps_whichfork(xmi));
+			return -EFSCORRUPTED;
+		}
+
+		/*
+		 * Save the mappings if we're estimating work and skipping
+		 * these identical mappings.
+		 */
+		if (adj) {
+			memcpy(&adj->left1, irec1, sizeof(*irec1));
+			memcpy(&adj->left2, irec2, sizeof(*irec2));
+		}
+	}
+
+	return 0;
+}
+
+/* Exchange these two mappings. */
+static void
+xfs_exchmaps_one_step(
+	struct xfs_trans		*tp,
+	struct xfs_exchmaps_intent	*xmi,
+	struct xfs_bmbt_irec		*irec1,
+	struct xfs_bmbt_irec		*irec2)
+{
+	int				whichfork = xfs_exchmaps_whichfork(xmi);
+
+	xfs_exchmaps_update_quota(tp, xmi, irec1, irec2);
+
+	/* Remove both mappings. */
+	xfs_bmap_unmap_extent(tp, xmi->xmi_ip1, whichfork, irec1);
+	xfs_bmap_unmap_extent(tp, xmi->xmi_ip2, whichfork, irec2);
+
+	/*
+	 * Re-add both mappings.  We exchange the file offsets between the two
+	 * maps and add the opposite map, which has the effect of filling the
+	 * logical offsets we just unmapped, but with with the physical mapping
+	 * information exchanged.
+	 */
+	swap(irec1->br_startoff, irec2->br_startoff);
+	xfs_bmap_map_extent(tp, xmi->xmi_ip1, whichfork, irec2);
+	xfs_bmap_map_extent(tp, xmi->xmi_ip2, whichfork, irec1);
+
+	/* Make sure we're not adding mappings past EOF. */
+	if (whichfork == XFS_DATA_FORK) {
+		xfs_exchmaps_update_size(tp, xmi->xmi_ip1, irec2,
+				xmi->xmi_isize1);
+		xfs_exchmaps_update_size(tp, xmi->xmi_ip2, irec1,
+				xmi->xmi_isize2);
+	}
+
+	/*
+	 * Advance our cursor and exit.   The caller (either defer ops or log
+	 * recovery) will log the XMD item, and if *blockcount is nonzero, it
+	 * will log a new XMI item for the remainder and call us back.
+	 */
+	xmi_advance(xmi, irec1);
+}
+
+/* Convert inode2's leaf attr fork back to shortform, if possible.. */
+STATIC int
+xfs_exchmaps_attr_to_sf(
+	struct xfs_trans		*tp,
+	struct xfs_exchmaps_intent	*xmi)
+{
+	struct xfs_da_args	args = {
+		.dp		= xmi->xmi_ip2,
+		.geo		= tp->t_mountp->m_attr_geo,
+		.whichfork	= XFS_ATTR_FORK,
+		.trans		= tp,
+		.owner		= xmi->xmi_ip2->i_ino,
+	};
+	struct xfs_buf		*bp;
+	int			forkoff;
+	int			error;
+
+	if (!xfs_attr_is_leaf(xmi->xmi_ip2))
+		return 0;
+
+	error = xfs_attr3_leaf_read(tp, xmi->xmi_ip2, xmi->xmi_ip2->i_ino, 0,
+			&bp);
+	if (error)
+		return error;
+
+	forkoff = xfs_attr_shortform_allfit(bp, xmi->xmi_ip2);
+	if (forkoff == 0)
+		return 0;
+
+	return xfs_attr3_leaf_to_shortform(bp, &args, forkoff);
+}
+
+/* Convert inode2's block dir fork back to shortform, if possible.. */
+STATIC int
+xfs_exchmaps_dir_to_sf(
+	struct xfs_trans		*tp,
+	struct xfs_exchmaps_intent	*xmi)
+{
+	struct xfs_da_args	args = {
+		.dp		= xmi->xmi_ip2,
+		.geo		= tp->t_mountp->m_dir_geo,
+		.whichfork	= XFS_DATA_FORK,
+		.trans		= tp,
+		.owner		= xmi->xmi_ip2->i_ino,
+	};
+	struct xfs_dir2_sf_hdr	sfh;
+	struct xfs_buf		*bp;
+	int			size;
+	int			error = 0;
+
+	if (xfs_dir2_format(&args, &error) != XFS_DIR2_FMT_BLOCK)
+		return error;
+
+	error = xfs_dir3_block_read(tp, xmi->xmi_ip2, xmi->xmi_ip2->i_ino, &bp);
+	if (error)
+		return error;
+
+	size = xfs_dir2_block_sfsize(xmi->xmi_ip2, bp->b_addr, &sfh);
+	if (size > xfs_inode_data_fork_size(xmi->xmi_ip2))
+		return 0;
+
+	return xfs_dir2_block_to_sf(&args, bp, size, &sfh);
+}
+
+/* Convert inode2's remote symlink target back to shortform, if possible. */
+STATIC int
+xfs_exchmaps_link_to_sf(
+	struct xfs_trans		*tp,
+	struct xfs_exchmaps_intent	*xmi)
+{
+	struct xfs_inode		*ip = xmi->xmi_ip2;
+	struct xfs_ifork		*ifp = xfs_ifork_ptr(ip, XFS_DATA_FORK);
+	char				*buf;
+	int				error;
+
+	if (ifp->if_format == XFS_DINODE_FMT_LOCAL ||
+	    ip->i_disk_size > xfs_inode_data_fork_size(ip))
+		return 0;
+
+	/* Read the current symlink target into a buffer. */
+	buf = kmalloc(ip->i_disk_size + 1,
+			GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL);
+	if (!buf) {
+		ASSERT(0);
+		return -ENOMEM;
+	}
+
+	error = xfs_symlink_remote_read(ip, buf);
+	if (error)
+		goto free;
+
+	/* Remove the blocks. */
+	error = xfs_symlink_remote_truncate(tp, ip);
+	if (error)
+		goto free;
+
+	/* Convert fork to local format and log our changes. */
+	xfs_idestroy_fork(ifp);
+	ifp->if_bytes = 0;
+	ifp->if_format = XFS_DINODE_FMT_LOCAL;
+	xfs_init_local_fork(ip, XFS_DATA_FORK, buf, ip->i_disk_size);
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_DDATA | XFS_ILOG_CORE);
+free:
+	kfree(buf);
+	return error;
+}
+
+/* Clear the reflink flag after an exchange. */
+static inline void
+xfs_exchmaps_clear_reflink(
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip)
+{
+	trace_xfs_reflink_unset_inode_flag(ip);
+
+	ip->i_diflags2 &= ~XFS_DIFLAG2_REFLINK;
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+}
+
+/* Finish whatever work might come after an exchange operation. */
+static int
+xfs_exchmaps_do_postop_work(
+	struct xfs_trans		*tp,
+	struct xfs_exchmaps_intent	*xmi)
+{
+	if (xmi->xmi_flags & __XFS_EXCHMAPS_INO2_SHORTFORM) {
+		int			error = 0;
+
+		if (xmi->xmi_flags & XFS_EXCHMAPS_ATTR_FORK)
+			error = xfs_exchmaps_attr_to_sf(tp, xmi);
+		else if (S_ISDIR(VFS_I(xmi->xmi_ip2)->i_mode))
+			error = xfs_exchmaps_dir_to_sf(tp, xmi);
+		else if (S_ISLNK(VFS_I(xmi->xmi_ip2)->i_mode))
+			error = xfs_exchmaps_link_to_sf(tp, xmi);
+		xmi->xmi_flags &= ~__XFS_EXCHMAPS_INO2_SHORTFORM;
+		if (error)
+			return error;
+	}
+
+	if (xmi->xmi_flags & XFS_EXCHMAPS_CLEAR_INO1_REFLINK) {
+		xfs_exchmaps_clear_reflink(tp, xmi->xmi_ip1);
+		xmi->xmi_flags &= ~XFS_EXCHMAPS_CLEAR_INO1_REFLINK;
+	}
+
+	if (xmi->xmi_flags & XFS_EXCHMAPS_CLEAR_INO2_REFLINK) {
+		xfs_exchmaps_clear_reflink(tp, xmi->xmi_ip2);
+		xmi->xmi_flags &= ~XFS_EXCHMAPS_CLEAR_INO2_REFLINK;
+	}
+
+	return 0;
+}
+
+/* Finish one step in a mapping exchange operation, possibly relogging. */
+int
+xfs_exchmaps_finish_one(
+	struct xfs_trans		*tp,
+	struct xfs_exchmaps_intent	*xmi)
+{
+	struct xfs_bmbt_irec		irec1, irec2;
+	int				error;
+
+	if (xmi_has_more_exchange_work(xmi)) {
+		/*
+		 * If the operation state says that some range of the files
+		 * have not yet been exchanged, look for mappings in that range
+		 * to exchange.  If we find some mappings, exchange them.
+		 */
+		error = xfs_exchmaps_find_mappings(xmi, &irec1, &irec2, NULL);
+		if (error)
+			return error;
+
+		if (xmi_has_more_exchange_work(xmi))
+			xfs_exchmaps_one_step(tp, xmi, &irec1, &irec2);
+
+		/*
+		 * If the caller asked us to exchange the file sizes after the
+		 * exchange and either we just exchanged the last mappings in
+		 * the range or we didn't find anything to exchange, update the
+		 * ondisk file sizes.
+		 */
+		if ((xmi->xmi_flags & XFS_EXCHMAPS_SET_SIZES) &&
+		    !xmi_has_more_exchange_work(xmi)) {
+			xmi->xmi_ip1->i_disk_size = xmi->xmi_isize1;
+			xmi->xmi_ip2->i_disk_size = xmi->xmi_isize2;
+
+			xfs_trans_log_inode(tp, xmi->xmi_ip1, XFS_ILOG_CORE);
+			xfs_trans_log_inode(tp, xmi->xmi_ip2, XFS_ILOG_CORE);
+		}
+	} else if (xmi_has_postop_work(xmi)) {
+		/*
+		 * Now that we're finished with the exchange operation,
+		 * complete the post-op cleanup work.
+		 */
+		error = xfs_exchmaps_do_postop_work(tp, xmi);
+		if (error)
+			return error;
+	}
+
+	if (XFS_TEST_ERROR(false, tp->t_mountp, XFS_ERRTAG_EXCHMAPS_FINISH_ONE))
+		return -EIO;
+
+	/* If we still have work to do, ask for a new transaction. */
+	if (xmi_has_more_exchange_work(xmi) || xmi_has_postop_work(xmi)) {
+		trace_xfs_exchmaps_defer(tp->t_mountp, xmi);
+		return -EAGAIN;
+	}
+
+	/*
+	 * If we reach here, we've finished all the exchange work and the post
+	 * operation work.  The last thing we need to do before returning to
+	 * the caller is to make sure that COW forks are set up correctly.
+	 */
+	if (!(xmi->xmi_flags & XFS_EXCHMAPS_ATTR_FORK)) {
+		xfs_exchmaps_ensure_cowfork(xmi->xmi_ip1);
+		xfs_exchmaps_ensure_cowfork(xmi->xmi_ip2);
+	}
+
+	return 0;
+}
+
+/*
+ * Compute the amount of bmbt blocks we should reserve for each file.  In the
+ * worst case, each exchange will fill a hole with a new mapping, which could
+ * result in a btree split every time we add a new leaf block.
+ */
+static inline uint64_t
+xfs_exchmaps_bmbt_blocks(
+	struct xfs_mount		*mp,
+	const struct xfs_exchmaps_req	*req)
+{
+	return howmany_64(req->nr_exchanges,
+					XFS_MAX_CONTIG_BMAPS_PER_BLOCK(mp)) *
+			XFS_EXTENTADD_SPACE_RES(mp, xfs_exchmaps_reqfork(req));
+}
+
+/* Compute the space we should reserve for the rmap btree expansions. */
+static inline uint64_t
+xfs_exchmaps_rmapbt_blocks(
+	struct xfs_mount		*mp,
+	const struct xfs_exchmaps_req	*req)
+{
+	if (!xfs_has_rmapbt(mp))
+		return 0;
+	if (XFS_IS_REALTIME_INODE(req->ip1))
+		return 0;
+
+	return howmany_64(req->nr_exchanges,
+					XFS_MAX_CONTIG_RMAPS_PER_BLOCK(mp)) *
+			XFS_RMAPADD_SPACE_RES(mp);
+}
+
+/* Estimate the bmbt and rmapbt overhead required to exchange mappings. */
+int
+xfs_exchmaps_estimate_overhead(
+	struct xfs_exchmaps_req		*req)
+{
+	struct xfs_mount		*mp = req->ip1->i_mount;
+	xfs_filblks_t			bmbt_blocks;
+	xfs_filblks_t			rmapbt_blocks;
+	xfs_filblks_t			resblks = req->resblks;
+
+	/*
+	 * Compute the number of bmbt and rmapbt blocks we might need to handle
+	 * the estimated number of exchanges.
+	 */
+	bmbt_blocks = xfs_exchmaps_bmbt_blocks(mp, req);
+	rmapbt_blocks = xfs_exchmaps_rmapbt_blocks(mp, req);
+
+	trace_xfs_exchmaps_overhead(mp, bmbt_blocks, rmapbt_blocks);
+
+	/* Make sure the change in file block count doesn't overflow. */
+	if (check_add_overflow(req->ip1_bcount, bmbt_blocks, &req->ip1_bcount))
+		return -EFBIG;
+	if (check_add_overflow(req->ip2_bcount, bmbt_blocks, &req->ip2_bcount))
+		return -EFBIG;
+
+	/*
+	 * Add together the number of blocks we need to handle btree growth,
+	 * then add it to the number of blocks we need to reserve to this
+	 * transaction.
+	 */
+	if (check_add_overflow(resblks, bmbt_blocks, &resblks))
+		return -ENOSPC;
+	if (check_add_overflow(resblks, bmbt_blocks, &resblks))
+		return -ENOSPC;
+	if (check_add_overflow(resblks, rmapbt_blocks, &resblks))
+		return -ENOSPC;
+	if (check_add_overflow(resblks, rmapbt_blocks, &resblks))
+		return -ENOSPC;
+
+	/* Can't actually reserve more than UINT_MAX blocks. */
+	if (req->resblks > UINT_MAX)
+		return -ENOSPC;
+
+	req->resblks = resblks;
+	trace_xfs_exchmaps_final_estimate(req);
+	return 0;
+}
+
+/* Decide if we can merge two real mappings. */
+static inline bool
+xmi_can_merge(
+	const struct xfs_bmbt_irec	*b1,
+	const struct xfs_bmbt_irec	*b2)
+{
+	/* Don't merge holes. */
+	if (b1->br_startblock == HOLESTARTBLOCK ||
+	    b2->br_startblock == HOLESTARTBLOCK)
+		return false;
+
+	/* We don't merge holes. */
+	if (!xfs_bmap_is_real_extent(b1) || !xfs_bmap_is_real_extent(b2))
+		return false;
+
+	if (b1->br_startoff   + b1->br_blockcount == b2->br_startoff &&
+	    b1->br_startblock + b1->br_blockcount == b2->br_startblock &&
+	    b1->br_state			  == b2->br_state &&
+	    b1->br_blockcount + b2->br_blockcount <= XFS_MAX_BMBT_EXTLEN)
+		return true;
+
+	return false;
+}
+
+/*
+ * Decide if we can merge three mappings.  Caller must ensure all three
+ * mappings must not be holes or delalloc reservations.
+ */
+static inline bool
+xmi_can_merge_all(
+	const struct xfs_bmbt_irec	*l,
+	const struct xfs_bmbt_irec	*m,
+	const struct xfs_bmbt_irec	*r)
+{
+	xfs_filblks_t			new_len;
+
+	new_len = l->br_blockcount + m->br_blockcount + r->br_blockcount;
+	return new_len <= XFS_MAX_BMBT_EXTLEN;
+}
+
+#define CLEFT_CONTIG	0x01
+#define CRIGHT_CONTIG	0x02
+#define CHOLE		0x04
+#define CBOTH_CONTIG	(CLEFT_CONTIG | CRIGHT_CONTIG)
+
+#define NLEFT_CONTIG	0x10
+#define NRIGHT_CONTIG	0x20
+#define NHOLE		0x40
+#define NBOTH_CONTIG	(NLEFT_CONTIG | NRIGHT_CONTIG)
+
+/* Estimate the effect of a single exchange on mapping count. */
+static inline int
+xmi_delta_nextents_step(
+	struct xfs_mount		*mp,
+	const struct xfs_bmbt_irec	*left,
+	const struct xfs_bmbt_irec	*curr,
+	const struct xfs_bmbt_irec	*new,
+	const struct xfs_bmbt_irec	*right)
+{
+	bool				lhole, rhole, chole, nhole;
+	unsigned int			state = 0;
+	int				ret = 0;
+
+	lhole = left->br_startblock == HOLESTARTBLOCK;
+	rhole = right->br_startblock == HOLESTARTBLOCK;
+	chole = curr->br_startblock == HOLESTARTBLOCK;
+	nhole = new->br_startblock == HOLESTARTBLOCK;
+
+	if (chole)
+		state |= CHOLE;
+	if (!lhole && !chole && xmi_can_merge(left, curr))
+		state |= CLEFT_CONTIG;
+	if (!rhole && !chole && xmi_can_merge(curr, right))
+		state |= CRIGHT_CONTIG;
+	if ((state & CBOTH_CONTIG) == CBOTH_CONTIG &&
+	    !xmi_can_merge_all(left, curr, right))
+		state &= ~CRIGHT_CONTIG;
+
+	if (nhole)
+		state |= NHOLE;
+	if (!lhole && !nhole && xmi_can_merge(left, new))
+		state |= NLEFT_CONTIG;
+	if (!rhole && !nhole && xmi_can_merge(new, right))
+		state |= NRIGHT_CONTIG;
+	if ((state & NBOTH_CONTIG) == NBOTH_CONTIG &&
+	    !xmi_can_merge_all(left, new, right))
+		state &= ~NRIGHT_CONTIG;
+
+	switch (state & (CLEFT_CONTIG | CRIGHT_CONTIG | CHOLE)) {
+	case CLEFT_CONTIG | CRIGHT_CONTIG:
+		/*
+		 * left/curr/right are the same mapping, so deleting curr
+		 * causes 2 new mappings to be created.
+		 */
+		ret += 2;
+		break;
+	case 0:
+		/*
+		 * curr is not contiguous with any mapping, so we remove curr
+		 * completely
+		 */
+		ret--;
+		break;
+	case CHOLE:
+		/* hole, do nothing */
+		break;
+	case CLEFT_CONTIG:
+	case CRIGHT_CONTIG:
+		/* trim either left or right, no change */
+		break;
+	}
+
+	switch (state & (NLEFT_CONTIG | NRIGHT_CONTIG | NHOLE)) {
+	case NLEFT_CONTIG | NRIGHT_CONTIG:
+		/*
+		 * left/curr/right will become the same mapping, so adding
+		 * curr causes the deletion of right.
+		 */
+		ret--;
+		break;
+	case 0:
+		/* new is not contiguous with any mapping */
+		ret++;
+		break;
+	case NHOLE:
+		/* hole, do nothing. */
+		break;
+	case NLEFT_CONTIG:
+	case NRIGHT_CONTIG:
+		/* new is absorbed into left or right, no change */
+		break;
+	}
+
+	trace_xfs_exchmaps_delta_nextents_step(mp, left, curr, new, right, ret,
+			state);
+	return ret;
+}
+
+/* Make sure we don't overflow the extent (mapping) counters. */
+static inline int
+xmi_ensure_delta_nextents(
+	struct xfs_exchmaps_req	*req,
+	struct xfs_inode	*ip,
+	int64_t			delta)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	int			whichfork = xfs_exchmaps_reqfork(req);
+	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, whichfork);
+	uint64_t		new_nextents;
+	xfs_extnum_t		max_nextents;
+
+	if (delta < 0)
+		return 0;
+
+	/*
+	 * It's always an error if the delta causes integer overflow.  delta
+	 * needs an explicit cast here to avoid warnings about implicit casts
+	 * coded into the overflow check.
+	 */
+	if (check_add_overflow(ifp->if_nextents, (uint64_t)delta,
+				&new_nextents))
+		return -EFBIG;
+
+	if (XFS_TEST_ERROR(false, mp, XFS_ERRTAG_REDUCE_MAX_IEXTENTS) &&
+	    new_nextents > 10)
+		return -EFBIG;
+
+	/*
+	 * We always promote both inodes to have large extent counts if the
+	 * superblock feature is enabled, so we only need to check against the
+	 * theoretical maximum.
+	 */
+	max_nextents = xfs_iext_max_nextents(xfs_has_large_extent_counts(mp),
+					     whichfork);
+	if (new_nextents > max_nextents)
+		return -EFBIG;
+
+	return 0;
+}
+
+/* Find the next mapping after irec. */
+static inline int
+xmi_next(
+	struct xfs_inode		*ip,
+	int				bmap_flags,
+	const struct xfs_bmbt_irec	*irec,
+	struct xfs_bmbt_irec		*nrec)
+{
+	xfs_fileoff_t			off;
+	xfs_filblks_t			blockcount;
+	int				nimaps = 1;
+	int				error;
+
+	off = irec->br_startoff + irec->br_blockcount;
+	blockcount = XFS_MAX_FILEOFF - off;
+	error = xfs_bmapi_read(ip, off, blockcount, nrec, &nimaps, bmap_flags);
+	if (error)
+		return error;
+	if (nrec->br_startblock == DELAYSTARTBLOCK ||
+	    nrec->br_startoff != off) {
+		/*
+		 * If we don't get the mapping we want, return a zero-length
+		 * mapping, which our estimator function will pretend is a hole.
+		 * We shouldn't get delalloc reservations.
+		 */
+		nrec->br_startblock = HOLESTARTBLOCK;
+	}
+
+	return 0;
+}
+
+int __init
+xfs_exchmaps_intent_init_cache(void)
+{
+	xfs_exchmaps_intent_cache = kmem_cache_create("xfs_exchmaps_intent",
+			sizeof(struct xfs_exchmaps_intent),
+			0, 0, NULL);
+
+	return xfs_exchmaps_intent_cache != NULL ? 0 : -ENOMEM;
+}
+
+void
+xfs_exchmaps_intent_destroy_cache(void)
+{
+	kmem_cache_destroy(xfs_exchmaps_intent_cache);
+	xfs_exchmaps_intent_cache = NULL;
+}
+
+/*
+ * Decide if we will exchange the reflink flags between the two files after the
+ * exchange.  The only time we want to do this is if we're exchanging all
+ * mappings under EOF and the inode reflink flags have different states.
+ */
+static inline bool
+xmi_can_exchange_reflink_flags(
+	const struct xfs_exchmaps_req	*req,
+	unsigned int			reflink_state)
+{
+	struct xfs_mount		*mp = req->ip1->i_mount;
+
+	if (hweight32(reflink_state) != 1)
+		return false;
+	if (req->startoff1 != 0 || req->startoff2 != 0)
+		return false;
+	if (req->blockcount != XFS_B_TO_FSB(mp, req->ip1->i_disk_size))
+		return false;
+	if (req->blockcount != XFS_B_TO_FSB(mp, req->ip2->i_disk_size))
+		return false;
+	return true;
+}
+
+
+/* Allocate and initialize a new incore intent item from a request. */
+struct xfs_exchmaps_intent *
+xfs_exchmaps_init_intent(
+	const struct xfs_exchmaps_req	*req)
+{
+	struct xfs_exchmaps_intent	*xmi;
+	unsigned int			rs = 0;
+
+	xmi = kmem_cache_zalloc(xfs_exchmaps_intent_cache,
+			GFP_NOFS | __GFP_NOFAIL);
+	INIT_LIST_HEAD(&xmi->xmi_list);
+	xmi->xmi_ip1 = req->ip1;
+	xmi->xmi_ip2 = req->ip2;
+	xmi->xmi_startoff1 = req->startoff1;
+	xmi->xmi_startoff2 = req->startoff2;
+	xmi->xmi_blockcount = req->blockcount;
+	xmi->xmi_isize1 = xmi->xmi_isize2 = -1;
+	xmi->xmi_flags = req->flags & XFS_EXCHMAPS_PARAMS;
+
+	if (xfs_exchmaps_whichfork(xmi) == XFS_ATTR_FORK) {
+		xmi->xmi_flags |= __XFS_EXCHMAPS_INO2_SHORTFORM;
+		return xmi;
+	}
+
+	if (req->flags & XFS_EXCHMAPS_SET_SIZES) {
+		xmi->xmi_flags |= XFS_EXCHMAPS_SET_SIZES;
+		xmi->xmi_isize1 = req->ip2->i_disk_size;
+		xmi->xmi_isize2 = req->ip1->i_disk_size;
+	}
+
+	/* Record the state of each inode's reflink flag before the op. */
+	if (xfs_is_reflink_inode(req->ip1))
+		rs |= 1;
+	if (xfs_is_reflink_inode(req->ip2))
+		rs |= 2;
+
+	/*
+	 * Figure out if we're clearing the reflink flags (which effectively
+	 * exchanges them) after the operation.
+	 */
+	if (xmi_can_exchange_reflink_flags(req, rs)) {
+		if (rs & 1)
+			xmi->xmi_flags |= XFS_EXCHMAPS_CLEAR_INO1_REFLINK;
+		if (rs & 2)
+			xmi->xmi_flags |= XFS_EXCHMAPS_CLEAR_INO2_REFLINK;
+	}
+
+	if (S_ISDIR(VFS_I(xmi->xmi_ip2)->i_mode) ||
+	    S_ISLNK(VFS_I(xmi->xmi_ip2)->i_mode))
+		xmi->xmi_flags |= __XFS_EXCHMAPS_INO2_SHORTFORM;
+
+	return xmi;
+}
+
+/*
+ * Estimate the number of exchange operations and the number of file blocks
+ * in each file that will be affected by the exchange operation.
+ */
+int
+xfs_exchmaps_estimate(
+	struct xfs_exchmaps_req		*req)
+{
+	struct xfs_exchmaps_intent	*xmi;
+	struct xfs_bmbt_irec		irec1, irec2;
+	struct xfs_exchmaps_adjacent	adj = ADJACENT_INIT;
+	xfs_filblks_t			ip1_blocks = 0, ip2_blocks = 0;
+	int64_t				d_nexts1, d_nexts2;
+	int				bmap_flags;
+	int				error;
+
+	ASSERT(!(req->flags & ~XFS_EXCHMAPS_PARAMS));
+
+	bmap_flags = xfs_bmapi_aflag(xfs_exchmaps_reqfork(req));
+	xmi = xfs_exchmaps_init_intent(req);
+
+	/*
+	 * To guard against the possibility of overflowing the extent counters,
+	 * we have to estimate an upper bound on the potential increase in that
+	 * counter.  We can split the mapping at each end of the range, and for
+	 * each step of the exchange we can split the mapping that we're
+	 * working on if the mappings do not align.
+	 */
+	d_nexts1 = d_nexts2 = 3;
+
+	while (xmi_has_more_exchange_work(xmi)) {
+		/*
+		 * Walk through the file ranges until we find something to
+		 * exchange.  Because we're simulating the exchange, pass in
+		 * adj to capture skipped mappings for correct estimation of
+		 * bmbt record merges.
+		 */
+		error = xfs_exchmaps_find_mappings(xmi, &irec1, &irec2, &adj);
+		if (error)
+			goto out_free;
+		if (!xmi_has_more_exchange_work(xmi))
+			break;
+
+		/* Update accounting. */
+		if (xfs_bmap_is_real_extent(&irec1))
+			ip1_blocks += irec1.br_blockcount;
+		if (xfs_bmap_is_real_extent(&irec2))
+			ip2_blocks += irec2.br_blockcount;
+		req->nr_exchanges++;
+
+		/* Read the next mappings from both files. */
+		error = xmi_next(req->ip1, bmap_flags, &irec1, &adj.right1);
+		if (error)
+			goto out_free;
+
+		error = xmi_next(req->ip2, bmap_flags, &irec2, &adj.right2);
+		if (error)
+			goto out_free;
+
+		/* Update extent count deltas. */
+		d_nexts1 += xmi_delta_nextents_step(req->ip1->i_mount,
+				&adj.left1, &irec1, &irec2, &adj.right1);
+
+		d_nexts2 += xmi_delta_nextents_step(req->ip1->i_mount,
+				&adj.left2, &irec2, &irec1, &adj.right2);
+
+		/* Now pretend we exchanged the mappings. */
+		if (xmi_can_merge(&adj.left2, &irec1))
+			adj.left2.br_blockcount += irec1.br_blockcount;
+		else
+			memcpy(&adj.left2, &irec1, sizeof(irec1));
+
+		if (xmi_can_merge(&adj.left1, &irec2))
+			adj.left1.br_blockcount += irec2.br_blockcount;
+		else
+			memcpy(&adj.left1, &irec2, sizeof(irec2));
+
+		xmi_advance(xmi, &irec1);
+	}
+
+	/* Account for the blocks that are being exchanged. */
+	if (XFS_IS_REALTIME_INODE(req->ip1) &&
+	    xfs_exchmaps_reqfork(req) == XFS_DATA_FORK) {
+		req->ip1_rtbcount = ip1_blocks;
+		req->ip2_rtbcount = ip2_blocks;
+	} else {
+		req->ip1_bcount = ip1_blocks;
+		req->ip2_bcount = ip2_blocks;
+	}
+
+	/*
+	 * Make sure that both forks have enough slack left in their extent
+	 * counters that the exchange operation will not overflow.
+	 */
+	trace_xfs_exchmaps_delta_nextents(req, d_nexts1, d_nexts2);
+	if (req->ip1 == req->ip2) {
+		error = xmi_ensure_delta_nextents(req, req->ip1,
+				d_nexts1 + d_nexts2);
+	} else {
+		error = xmi_ensure_delta_nextents(req, req->ip1, d_nexts1);
+		if (error)
+			goto out_free;
+		error = xmi_ensure_delta_nextents(req, req->ip2, d_nexts2);
+	}
+	if (error)
+		goto out_free;
+
+	trace_xfs_exchmaps_initial_estimate(req);
+	error = xfs_exchmaps_estimate_overhead(req);
+out_free:
+	kmem_cache_free(xfs_exchmaps_intent_cache, xmi);
+	return error;
+}
+
+/* Set the reflink flag before an operation. */
+static inline void
+xfs_exchmaps_set_reflink(
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip)
+{
+	trace_xfs_reflink_set_inode_flag(ip);
+
+	ip->i_diflags2 |= XFS_DIFLAG2_REFLINK;
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+}
+
+/*
+ * If either file has shared blocks and we're exchanging data forks, we must
+ * flag the other file as having shared blocks so that we get the shared-block
+ * rmap functions if we need to fix up the rmaps.
+ */
+void
+xfs_exchmaps_ensure_reflink(
+	struct xfs_trans			*tp,
+	const struct xfs_exchmaps_intent	*xmi)
+{
+	unsigned int				rs = 0;
+
+	if (xfs_is_reflink_inode(xmi->xmi_ip1))
+		rs |= 1;
+	if (xfs_is_reflink_inode(xmi->xmi_ip2))
+		rs |= 2;
+
+	if ((rs & 1) && !xfs_is_reflink_inode(xmi->xmi_ip2))
+		xfs_exchmaps_set_reflink(tp, xmi->xmi_ip2);
+
+	if ((rs & 2) && !xfs_is_reflink_inode(xmi->xmi_ip1))
+		xfs_exchmaps_set_reflink(tp, xmi->xmi_ip1);
+}
+
+/* Set the large extent count flag before an operation if needed. */
+static inline void
+xfs_exchmaps_ensure_large_extent_counts(
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip)
+{
+	if (xfs_inode_has_large_extent_counts(ip))
+		return;
+
+	ip->i_diflags2 |= XFS_DIFLAG2_NREXT64;
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+}
+
+/* Widen the extent counter fields of both inodes if necessary. */
+void
+xfs_exchmaps_upgrade_extent_counts(
+	struct xfs_trans			*tp,
+	const struct xfs_exchmaps_intent	*xmi)
+{
+	if (!xfs_has_large_extent_counts(tp->t_mountp))
+		return;
+
+	xfs_exchmaps_ensure_large_extent_counts(tp, xmi->xmi_ip1);
+	xfs_exchmaps_ensure_large_extent_counts(tp, xmi->xmi_ip2);
+}
+
+/*
+ * Schedule an exchange a range of mappings from one inode to another.
+ *
+ * The use of file mapping exchange log intent items ensures the operation can
+ * be resumed even if the system goes down.  The caller must commit the
+ * transaction to start the work.
+ *
+ * The caller must ensure the inodes must be joined to the transaction and
+ * ILOCKd; they will still be joined to the transaction at exit.
+ */
+void
+xfs_exchange_mappings(
+	struct xfs_trans		*tp,
+	const struct xfs_exchmaps_req	*req)
+{
+	struct xfs_exchmaps_intent	*xmi;
+
+	BUILD_BUG_ON(XFS_EXCHMAPS_INTERNAL_FLAGS & XFS_EXCHMAPS_LOGGED_FLAGS);
+
+	xfs_assert_ilocked(req->ip1, XFS_ILOCK_EXCL);
+	xfs_assert_ilocked(req->ip2, XFS_ILOCK_EXCL);
+	ASSERT(!(req->flags & ~XFS_EXCHMAPS_LOGGED_FLAGS));
+	if (req->flags & XFS_EXCHMAPS_SET_SIZES)
+		ASSERT(!(req->flags & XFS_EXCHMAPS_ATTR_FORK));
+	ASSERT(xfs_has_exchange_range(tp->t_mountp));
+
+	if (req->blockcount == 0)
+		return;
+
+	xmi = xfs_exchmaps_init_intent(req);
+	xfs_exchmaps_defer_add(tp, xmi);
+	xfs_exchmaps_ensure_reflink(tp, xmi);
+	xfs_exchmaps_upgrade_extent_counts(tp, xmi);
+}
diff --git a/fs/xfs/libxfs/xfs_exchmaps.h b/fs/xfs/libxfs/xfs_exchmaps.h
new file mode 100644
index 0000000000..fa822dff20
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_exchmaps.h
@@ -0,0 +1,124 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (c) 2020-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_EXCHMAPS_H__
+#define __XFS_EXCHMAPS_H__
+
+/* In-core deferred operation info about a file mapping exchange request. */
+struct xfs_exchmaps_intent {
+	/* List of other incore deferred work. */
+	struct list_head	xmi_list;
+
+	/* Inodes participating in the operation. */
+	struct xfs_inode	*xmi_ip1;
+	struct xfs_inode	*xmi_ip2;
+
+	/* File offset range information. */
+	xfs_fileoff_t		xmi_startoff1;
+	xfs_fileoff_t		xmi_startoff2;
+	xfs_filblks_t		xmi_blockcount;
+
+	/* Set these file sizes after the operation, unless negative. */
+	xfs_fsize_t		xmi_isize1;
+	xfs_fsize_t		xmi_isize2;
+
+	uint64_t		xmi_flags;	/* XFS_EXCHMAPS_* flags */
+};
+
+/* Try to convert inode2 from block to short format at the end, if possible. */
+#define __XFS_EXCHMAPS_INO2_SHORTFORM	(1ULL << 63)
+
+#define XFS_EXCHMAPS_INTERNAL_FLAGS	(__XFS_EXCHMAPS_INO2_SHORTFORM)
+
+/* flags that can be passed to xfs_exchmaps_{estimate,mappings} */
+#define XFS_EXCHMAPS_PARAMS		(XFS_EXCHMAPS_ATTR_FORK | \
+					 XFS_EXCHMAPS_SET_SIZES | \
+					 XFS_EXCHMAPS_INO1_WRITTEN)
+
+static inline int
+xfs_exchmaps_whichfork(const struct xfs_exchmaps_intent *xmi)
+{
+	if (xmi->xmi_flags & XFS_EXCHMAPS_ATTR_FORK)
+		return XFS_ATTR_FORK;
+	return XFS_DATA_FORK;
+}
+
+/* Parameters for a mapping exchange request. */
+struct xfs_exchmaps_req {
+	/* Inodes participating in the operation. */
+	struct xfs_inode	*ip1;
+	struct xfs_inode	*ip2;
+
+	/* File offset range information. */
+	xfs_fileoff_t		startoff1;
+	xfs_fileoff_t		startoff2;
+	xfs_filblks_t		blockcount;
+
+	/* XFS_EXCHMAPS_* operation flags */
+	uint64_t		flags;
+
+	/*
+	 * Fields below this line are filled out by xfs_exchmaps_estimate;
+	 * callers should initialize this part of the struct to zero.
+	 */
+
+	/*
+	 * Data device blocks to be moved out of ip1, and free space needed to
+	 * handle the bmbt changes.
+	 */
+	xfs_filblks_t		ip1_bcount;
+
+	/*
+	 * Data device blocks to be moved out of ip2, and free space needed to
+	 * handle the bmbt changes.
+	 */
+	xfs_filblks_t		ip2_bcount;
+
+	/* rt blocks to be moved out of ip1. */
+	xfs_filblks_t		ip1_rtbcount;
+
+	/* rt blocks to be moved out of ip2. */
+	xfs_filblks_t		ip2_rtbcount;
+
+	/* Free space needed to handle the bmbt changes */
+	unsigned long long	resblks;
+
+	/* Number of exchanges needed to complete the operation */
+	unsigned long long	nr_exchanges;
+};
+
+static inline int
+xfs_exchmaps_reqfork(const struct xfs_exchmaps_req *req)
+{
+	if (req->flags & XFS_EXCHMAPS_ATTR_FORK)
+		return XFS_ATTR_FORK;
+	return XFS_DATA_FORK;
+}
+
+int xfs_exchmaps_estimate_overhead(struct xfs_exchmaps_req *req);
+int xfs_exchmaps_estimate(struct xfs_exchmaps_req *req);
+
+extern struct kmem_cache	*xfs_exchmaps_intent_cache;
+
+int __init xfs_exchmaps_intent_init_cache(void);
+void xfs_exchmaps_intent_destroy_cache(void);
+
+struct xfs_exchmaps_intent *xfs_exchmaps_init_intent(
+		const struct xfs_exchmaps_req *req);
+void xfs_exchmaps_ensure_reflink(struct xfs_trans *tp,
+		const struct xfs_exchmaps_intent *xmi);
+void xfs_exchmaps_upgrade_extent_counts(struct xfs_trans *tp,
+		const struct xfs_exchmaps_intent *xmi);
+
+int xfs_exchmaps_finish_one(struct xfs_trans *tp,
+		struct xfs_exchmaps_intent *xmi);
+
+int xfs_exchmaps_check_forks(struct xfs_mount *mp,
+		const struct xfs_exchmaps_req *req);
+
+void xfs_exchange_mappings(struct xfs_trans *tp,
+		const struct xfs_exchmaps_req *req);
+
+#endif /* __XFS_EXCHMAPS_H__ */
diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index 382ab1e71c..61f51becff 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -367,19 +367,23 @@ xfs_sb_has_ro_compat_feature(
 	return (sbp->sb_features_ro_compat & feature) != 0;
 }
 
-#define XFS_SB_FEAT_INCOMPAT_FTYPE	(1 << 0)	/* filetype in dirent */
-#define XFS_SB_FEAT_INCOMPAT_SPINODES	(1 << 1)	/* sparse inode chunks */
-#define XFS_SB_FEAT_INCOMPAT_META_UUID	(1 << 2)	/* metadata UUID */
-#define XFS_SB_FEAT_INCOMPAT_BIGTIME	(1 << 3)	/* large timestamps */
-#define XFS_SB_FEAT_INCOMPAT_NEEDSREPAIR (1 << 4)	/* needs xfs_repair */
-#define XFS_SB_FEAT_INCOMPAT_NREXT64	(1 << 5)	/* large extent counters */
+#define XFS_SB_FEAT_INCOMPAT_FTYPE	(1 << 0)  /* filetype in dirent */
+#define XFS_SB_FEAT_INCOMPAT_SPINODES	(1 << 1)  /* sparse inode chunks */
+#define XFS_SB_FEAT_INCOMPAT_META_UUID	(1 << 2)  /* metadata UUID */
+#define XFS_SB_FEAT_INCOMPAT_BIGTIME	(1 << 3)  /* large timestamps */
+#define XFS_SB_FEAT_INCOMPAT_NEEDSREPAIR (1 << 4) /* needs xfs_repair */
+#define XFS_SB_FEAT_INCOMPAT_NREXT64	(1 << 5)  /* large extent counters */
+#define XFS_SB_FEAT_INCOMPAT_EXCHRANGE	(1 << 6)  /* exchangerange supported */
+#define XFS_SB_FEAT_INCOMPAT_PARENT	(1 << 7)  /* parent pointers */
 #define XFS_SB_FEAT_INCOMPAT_ALL \
-		(XFS_SB_FEAT_INCOMPAT_FTYPE|	\
-		 XFS_SB_FEAT_INCOMPAT_SPINODES|	\
-		 XFS_SB_FEAT_INCOMPAT_META_UUID| \
-		 XFS_SB_FEAT_INCOMPAT_BIGTIME| \
-		 XFS_SB_FEAT_INCOMPAT_NEEDSREPAIR| \
-		 XFS_SB_FEAT_INCOMPAT_NREXT64)
+		(XFS_SB_FEAT_INCOMPAT_FTYPE | \
+		 XFS_SB_FEAT_INCOMPAT_SPINODES | \
+		 XFS_SB_FEAT_INCOMPAT_META_UUID | \
+		 XFS_SB_FEAT_INCOMPAT_BIGTIME | \
+		 XFS_SB_FEAT_INCOMPAT_NEEDSREPAIR | \
+		 XFS_SB_FEAT_INCOMPAT_NREXT64 | \
+		 XFS_SB_FEAT_INCOMPAT_EXCHRANGE | \
+		 XFS_SB_FEAT_INCOMPAT_PARENT)
 
 #define XFS_SB_FEAT_INCOMPAT_UNKNOWN	~XFS_SB_FEAT_INCOMPAT_ALL
 static inline bool
@@ -477,15 +481,9 @@ xfs_is_quota_inode(struct xfs_sb *sbp, xfs_ino_t ino)
 #define	XFS_AGI_GOOD_VERSION(v)	((v) == XFS_AGI_VERSION)
 
 /*
- * Btree number 0 is bno, 1 is cnt, 2 is rmap. This value gives the size of the
- * arrays below.
- */
-#define	XFS_BTNUM_AGF	((int)XFS_BTNUM_RMAPi + 1)
-
-/*
- * The second word of agf_levels in the first a.g. overlaps the EFS
- * superblock's magic number.  Since the magic numbers valid for EFS
- * are > 64k, our value cannot be confused for an EFS superblock's.
+ * agf_cnt_level in the first AGF overlaps the EFS superblock's magic number.
+ * Since the magic numbers valid for EFS are > 64k, our value cannot be confused
+ * for an EFS superblock.
  */
 
 typedef struct xfs_agf {
@@ -499,8 +497,13 @@ typedef struct xfs_agf {
 	/*
 	 * Freespace and rmap information
 	 */
-	__be32		agf_roots[XFS_BTNUM_AGF];	/* root blocks */
-	__be32		agf_levels[XFS_BTNUM_AGF];	/* btree levels */
+	__be32		agf_bno_root;	/* bnobt root block */
+	__be32		agf_cnt_root;	/* cntbt root block */
+	__be32		agf_rmap_root;	/* rmapbt root block */
+
+	__be32		agf_bno_level;	/* bnobt btree levels */
+	__be32		agf_cnt_level;	/* cntbt btree levels */
+	__be32		agf_rmap_level;	/* rmapbt btree levels */
 
 	__be32		agf_flfirst;	/* first freelist block's index */
 	__be32		agf_fllast;	/* last freelist block's index */
@@ -899,6 +902,12 @@ static inline uint xfs_dinode_size(int version)
 #define	XFS_MAXLINK		((1U << 31) - 1U)
 
 /*
+ * Any file that hits the maximum ondisk link count should be pinned to avoid
+ * a use-after-free situation.
+ */
+#define	XFS_NLINK_PINNED	(~0U)
+
+/*
  * Values for di_format
  *
  * This enum is used in string mapping in xfs_trace.h; please keep the
diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h
index 6360073865..454b63ef72 100644
--- a/fs/xfs/libxfs/xfs_fs.h
+++ b/fs/xfs/libxfs/xfs_fs.h
@@ -195,6 +195,8 @@ struct xfs_fsop_geom {
 #define XFS_FSOP_GEOM_SICK_PQUOTA	(1 << 3)  /* project quota */
 #define XFS_FSOP_GEOM_SICK_RT_BITMAP	(1 << 4)  /* realtime bitmap */
 #define XFS_FSOP_GEOM_SICK_RT_SUMMARY	(1 << 5)  /* realtime summary */
+#define XFS_FSOP_GEOM_SICK_QUOTACHECK	(1 << 6)  /* quota counts */
+#define XFS_FSOP_GEOM_SICK_NLINKS	(1 << 7)  /* inode link counts */
 
 /* Output for XFS_FS_COUNTS */
 typedef struct xfs_fsop_counts {
@@ -237,6 +239,8 @@ typedef struct xfs_fsop_resblks {
 #define XFS_FSOP_GEOM_FLAGS_BIGTIME	(1 << 21) /* 64-bit nsec timestamps */
 #define XFS_FSOP_GEOM_FLAGS_INOBTCNT	(1 << 22) /* inobt btree counter */
 #define XFS_FSOP_GEOM_FLAGS_NREXT64	(1 << 23) /* large extent counters */
+#define XFS_FSOP_GEOM_FLAGS_EXCHANGE_RANGE (1 << 24) /* exchange range */
+#define XFS_FSOP_GEOM_FLAGS_PARENT	(1 << 25) /* linux parent pointers */
 
 /*
  * Minimum and maximum sizes need for growth checks.
@@ -292,6 +296,7 @@ struct xfs_ag_geometry {
 #define XFS_AG_GEOM_SICK_FINOBT	(1 << 7)  /* free inode index */
 #define XFS_AG_GEOM_SICK_RMAPBT	(1 << 8)  /* reverse mappings */
 #define XFS_AG_GEOM_SICK_REFCNTBT (1 << 9)  /* reference counts */
+#define XFS_AG_GEOM_SICK_INODES	(1 << 10) /* bad inodes were seen */
 
 /*
  * Structures for XFS_IOC_FSGROWFSDATA, XFS_IOC_FSGROWFSLOG & XFS_IOC_FSGROWFSRT
@@ -406,6 +411,7 @@ struct xfs_bulkstat {
 #define XFS_BS_SICK_XATTR	(1 << 5)  /* extended attributes */
 #define XFS_BS_SICK_SYMLINK	(1 << 6)  /* symbolic link remote target */
 #define XFS_BS_SICK_PARENT	(1 << 7)  /* parent pointers */
+#define XFS_BS_SICK_DIRTREE	(1 << 8)  /* directory tree structure */
 
 /*
  * Project quota id helpers (previously projid was 16bit only
@@ -629,7 +635,9 @@ typedef struct xfs_fsop_attrmulti_handlereq {
 /*
  * per machine unique filesystem identifier types.
  */
-typedef struct { __u32 val[2]; } xfs_fsid_t; /* file system id type */
+typedef struct xfs_fsid {
+	__u32	val[2];			/* file system id type */
+} xfs_fsid_t;
 
 typedef struct xfs_fid {
 	__u16	fid_len;		/* length of remainder	*/
@@ -709,9 +717,22 @@ struct xfs_scrub_metadata {
 #define XFS_SCRUB_TYPE_GQUOTA	22	/* group quotas */
 #define XFS_SCRUB_TYPE_PQUOTA	23	/* project quotas */
 #define XFS_SCRUB_TYPE_FSCOUNTERS 24	/* fs summary counters */
+#define XFS_SCRUB_TYPE_QUOTACHECK 25	/* quota counters */
+#define XFS_SCRUB_TYPE_NLINKS	26	/* inode link counts */
+#define XFS_SCRUB_TYPE_HEALTHY	27	/* everything checked out ok */
+#define XFS_SCRUB_TYPE_DIRTREE	28	/* directory tree structure */
 
 /* Number of scrub subcommands. */
-#define XFS_SCRUB_TYPE_NR	25
+#define XFS_SCRUB_TYPE_NR	29
+
+/*
+ * This special type code only applies to the vectored scrub implementation.
+ *
+ * If any of the previous scrub vectors recorded runtime errors or have
+ * sv_flags bits set that match the OFLAG bits in the barrier vector's
+ * sv_flags, set the barrier's sv_ret to -ECANCELED and return to userspace.
+ */
+#define XFS_SCRUB_TYPE_BARRIER	(0xFFFFFFFF)
 
 /* i: Repair this metadata. */
 #define XFS_SCRUB_IFLAG_REPAIR		(1u << 0)
@@ -757,6 +778,29 @@ struct xfs_scrub_metadata {
 				 XFS_SCRUB_OFLAG_NO_REPAIR_NEEDED)
 #define XFS_SCRUB_FLAGS_ALL	(XFS_SCRUB_FLAGS_IN | XFS_SCRUB_FLAGS_OUT)
 
+/* Vectored scrub calls to reduce the number of kernel transitions. */
+
+struct xfs_scrub_vec {
+	__u32 sv_type;		/* XFS_SCRUB_TYPE_* */
+	__u32 sv_flags;		/* XFS_SCRUB_FLAGS_* */
+	__s32 sv_ret;		/* 0 or a negative error code */
+	__u32 sv_reserved;	/* must be zero */
+};
+
+/* Vectored metadata scrub control structure. */
+struct xfs_scrub_vec_head {
+	__u64 svh_ino;		/* inode number. */
+	__u32 svh_gen;		/* inode generation. */
+	__u32 svh_agno;		/* ag number. */
+	__u32 svh_flags;	/* XFS_SCRUB_VEC_FLAGS_* */
+	__u16 svh_rest_us;	/* wait this much time between vector items */
+	__u16 svh_nr;		/* number of svh_vectors */
+	__u64 svh_reserved;	/* must be zero */
+	__u64 svh_vectors;	/* pointer to buffer of xfs_scrub_vec */
+};
+
+#define XFS_SCRUB_VEC_FLAGS_ALL		(0)
+
 /*
  * ioctl limits
  */
@@ -766,6 +810,118 @@ struct xfs_scrub_metadata {
 #  define XFS_XATTR_LIST_MAX 65536
 #endif
 
+/*
+ * Exchange part of file1 with part of the file that this ioctl that is being
+ * called against (which we'll call file2).  Filesystems must be able to
+ * restart and complete the operation even after the system goes down.
+ */
+struct xfs_exchange_range {
+	__s32		file1_fd;
+	__u32		pad;		/* must be zeroes */
+	__u64		file1_offset;	/* file1 offset, bytes */
+	__u64		file2_offset;	/* file2 offset, bytes */
+	__u64		length;		/* bytes to exchange */
+
+	__u64		flags;		/* see XFS_EXCHANGE_RANGE_* below */
+};
+
+/*
+ * Exchange file data all the way to the ends of both files, and then exchange
+ * the file sizes.  This flag can be used to replace a file's contents with a
+ * different amount of data.  length will be ignored.
+ */
+#define XFS_EXCHANGE_RANGE_TO_EOF	(1ULL << 0)
+
+/* Flush all changes in file data and file metadata to disk before returning. */
+#define XFS_EXCHANGE_RANGE_DSYNC	(1ULL << 1)
+
+/* Dry run; do all the parameter verification but do not change anything. */
+#define XFS_EXCHANGE_RANGE_DRY_RUN	(1ULL << 2)
+
+/*
+ * Exchange only the parts of the two files where the file allocation units
+ * mapped to file1's range have been written to.  This can accelerate
+ * scatter-gather atomic writes with a temp file if all writes are aligned to
+ * the file allocation unit.
+ */
+#define XFS_EXCHANGE_RANGE_FILE1_WRITTEN (1ULL << 3)
+
+#define XFS_EXCHANGE_RANGE_ALL_FLAGS	(XFS_EXCHANGE_RANGE_TO_EOF | \
+					 XFS_EXCHANGE_RANGE_DSYNC | \
+					 XFS_EXCHANGE_RANGE_DRY_RUN | \
+					 XFS_EXCHANGE_RANGE_FILE1_WRITTEN)
+
+/* Iterating parent pointers of files. */
+
+/* target was the root directory */
+#define XFS_GETPARENTS_OFLAG_ROOT	(1U << 0)
+
+/* Cursor is done iterating pptrs */
+#define XFS_GETPARENTS_OFLAG_DONE	(1U << 1)
+
+#define XFS_GETPARENTS_OFLAGS_ALL	(XFS_GETPARENTS_OFLAG_ROOT | \
+					 XFS_GETPARENTS_OFLAG_DONE)
+
+#define XFS_GETPARENTS_IFLAGS_ALL	(0)
+
+struct xfs_getparents_rec {
+	struct xfs_handle	gpr_parent; /* Handle to parent */
+	__u32			gpr_reclen; /* Length of entire record */
+	__u32			gpr_reserved; /* zero */
+	char			gpr_name[]; /* Null-terminated filename */
+};
+
+/* Iterate through this file's directory parent pointers */
+struct xfs_getparents {
+	/*
+	 * Structure to track progress in iterating the parent pointers.
+	 * Must be initialized to zeroes before the first ioctl call, and
+	 * not touched by callers after that.
+	 */
+	struct xfs_attrlist_cursor	gp_cursor;
+
+	/* Input flags: XFS_GETPARENTS_IFLAG* */
+	__u16				gp_iflags;
+
+	/* Output flags: XFS_GETPARENTS_OFLAG* */
+	__u16				gp_oflags;
+
+	/* Size of the gp_buffer in bytes */
+	__u32				gp_bufsize;
+
+	/* Must be set to zero */
+	__u64				gp_reserved;
+
+	/* Pointer to a buffer in which to place xfs_getparents_rec */
+	__u64				gp_buffer;
+};
+
+static inline struct xfs_getparents_rec *
+xfs_getparents_first_rec(struct xfs_getparents *gp)
+{
+	return (struct xfs_getparents_rec *)(uintptr_t)gp->gp_buffer;
+}
+
+static inline struct xfs_getparents_rec *
+xfs_getparents_next_rec(struct xfs_getparents *gp,
+			struct xfs_getparents_rec *gpr)
+{
+	void *next = ((void *)gpr + gpr->gpr_reclen);
+	void *end = (void *)(uintptr_t)(gp->gp_buffer + gp->gp_bufsize);
+
+	if (next >= end)
+		return NULL;
+
+	return next;
+}
+
+/* Iterate through this file handle's directory parent pointers. */
+struct xfs_getparents_by_handle {
+	/* Handle to file whose parents we want. */
+	struct xfs_handle		gph_handle;
+
+	struct xfs_getparents		gph_request;
+};
 
 /*
  * ioctl commands that are used by Linux filesystems
@@ -802,6 +958,9 @@ struct xfs_scrub_metadata {
 /*	XFS_IOC_GETFSMAP ------ hoisted 59         */
 #define XFS_IOC_SCRUB_METADATA	_IOWR('X', 60, struct xfs_scrub_metadata)
 #define XFS_IOC_AG_GEOMETRY	_IOWR('X', 61, struct xfs_ag_geometry)
+#define XFS_IOC_GETPARENTS	_IOWR('X', 62, struct xfs_getparents)
+#define XFS_IOC_GETPARENTS_BY_HANDLE _IOWR('X', 63, struct xfs_getparents_by_handle)
+#define XFS_IOC_SCRUBV_METADATA	_IOWR('X', 64, struct xfs_scrub_vec_head)
 
 /*
  * ioctl commands that replace IRIX syssgi()'s
@@ -837,6 +996,7 @@ struct xfs_scrub_metadata {
 #define XFS_IOC_FSGEOMETRY	     _IOR ('X', 126, struct xfs_fsop_geom)
 #define XFS_IOC_BULKSTAT	     _IOR ('X', 127, struct xfs_bulkstat_req)
 #define XFS_IOC_INUMBERS	     _IOR ('X', 128, struct xfs_inumbers_req)
+#define XFS_IOC_EXCHANGE_RANGE	     _IOW ('X', 129, struct xfs_exchange_range)
 /*	XFS_IOC_GETFSUUID ---------- deprecated 140	 */
 
 
diff --git a/fs/xfs/libxfs/xfs_health.h b/fs/xfs/libxfs/xfs_health.h
index 6296993ff8..b0edb4288e 100644
--- a/fs/xfs/libxfs/xfs_health.h
+++ b/fs/xfs/libxfs/xfs_health.h
@@ -26,21 +26,40 @@
  * and the "sick" field tells us if that piece was found to need repairs.
  * Therefore we can conclude that for a given sick flag value:
  *
- *  - checked && sick  => metadata needs repair
- *  - checked && !sick => metadata is ok
- *  - !checked         => has not been examined since mount
+ *  - checked && sick   => metadata needs repair
+ *  - checked && !sick  => metadata is ok
+ *  - !checked && sick  => errors have been observed during normal operation,
+ *                         but the metadata has not been checked thoroughly
+ *  - !checked && !sick => has not been examined since mount
+ *
+ * Evidence of health problems can be sorted into three basic categories:
+ *
+ * a) Primary evidence, which signals that something is defective within the
+ *    general grouping of metadata.
+ *
+ * b) Secondary evidence, which are side effects of primary problem but are
+ *    not themselves problems.  These can be forgotten when the primary
+ *    health problems are addressed.
+ *
+ * c) Indirect evidence, which points to something being wrong in another
+ *    group, but we had to release resources and this is all that's left of
+ *    that state.
  */
 
 struct xfs_mount;
 struct xfs_perag;
 struct xfs_inode;
 struct xfs_fsop_geom;
+struct xfs_btree_cur;
+struct xfs_da_args;
 
 /* Observable health issues for metadata spanning the entire filesystem. */
 #define XFS_SICK_FS_COUNTERS	(1 << 0)  /* summary counters */
 #define XFS_SICK_FS_UQUOTA	(1 << 1)  /* user quota */
 #define XFS_SICK_FS_GQUOTA	(1 << 2)  /* group quota */
 #define XFS_SICK_FS_PQUOTA	(1 << 3)  /* project quota */
+#define XFS_SICK_FS_QUOTACHECK	(1 << 4)  /* quota counts */
+#define XFS_SICK_FS_NLINKS	(1 << 5)  /* inode link counts */
 
 /* Observable health issues for realtime volume metadata. */
 #define XFS_SICK_RT_BITMAP	(1 << 0)  /* realtime bitmap */
@@ -57,6 +76,7 @@ struct xfs_fsop_geom;
 #define XFS_SICK_AG_FINOBT	(1 << 7)  /* free inode index */
 #define XFS_SICK_AG_RMAPBT	(1 << 8)  /* reverse mappings */
 #define XFS_SICK_AG_REFCNTBT	(1 << 9)  /* reference counts */
+#define XFS_SICK_AG_INODES	(1 << 10) /* inactivated bad inodes */
 
 /* Observable health issues for inode metadata. */
 #define XFS_SICK_INO_CORE	(1 << 0)  /* inode core */
@@ -73,11 +93,17 @@ struct xfs_fsop_geom;
 #define XFS_SICK_INO_DIR_ZAPPED		(1 << 10) /* directory erased */
 #define XFS_SICK_INO_SYMLINK_ZAPPED	(1 << 11) /* symlink erased */
 
+/* Don't propagate sick status to ag health summary during inactivation */
+#define XFS_SICK_INO_FORGET	(1 << 12)
+#define XFS_SICK_INO_DIRTREE	(1 << 13)  /* directory tree structure */
+
 /* Primary evidence of health problems in a given group. */
 #define XFS_SICK_FS_PRIMARY	(XFS_SICK_FS_COUNTERS | \
 				 XFS_SICK_FS_UQUOTA | \
 				 XFS_SICK_FS_GQUOTA | \
-				 XFS_SICK_FS_PQUOTA)
+				 XFS_SICK_FS_PQUOTA | \
+				 XFS_SICK_FS_QUOTACHECK | \
+				 XFS_SICK_FS_NLINKS)
 
 #define XFS_SICK_RT_PRIMARY	(XFS_SICK_RT_BITMAP | \
 				 XFS_SICK_RT_SUMMARY)
@@ -100,36 +126,94 @@ struct xfs_fsop_geom;
 				 XFS_SICK_INO_DIR | \
 				 XFS_SICK_INO_XATTR | \
 				 XFS_SICK_INO_SYMLINK | \
-				 XFS_SICK_INO_PARENT)
+				 XFS_SICK_INO_PARENT | \
+				 XFS_SICK_INO_DIRTREE)
 
 #define XFS_SICK_INO_ZAPPED	(XFS_SICK_INO_BMBTD_ZAPPED | \
 				 XFS_SICK_INO_BMBTA_ZAPPED | \
 				 XFS_SICK_INO_DIR_ZAPPED | \
 				 XFS_SICK_INO_SYMLINK_ZAPPED)
 
-/* These functions must be provided by the xfs implementation. */
+/* Secondary state related to (but not primary evidence of) health problems. */
+#define XFS_SICK_FS_SECONDARY	(0)
+#define XFS_SICK_RT_SECONDARY	(0)
+#define XFS_SICK_AG_SECONDARY	(0)
+#define XFS_SICK_INO_SECONDARY	(XFS_SICK_INO_FORGET)
+
+/* Evidence of health problems elsewhere. */
+#define XFS_SICK_FS_INDIRECT	(0)
+#define XFS_SICK_RT_INDIRECT	(0)
+#define XFS_SICK_AG_INDIRECT	(XFS_SICK_AG_INODES)
+#define XFS_SICK_INO_INDIRECT	(0)
+
+/* All health masks. */
+#define XFS_SICK_FS_ALL	(XFS_SICK_FS_PRIMARY | \
+				 XFS_SICK_FS_SECONDARY | \
+				 XFS_SICK_FS_INDIRECT)
+
+#define XFS_SICK_RT_ALL	(XFS_SICK_RT_PRIMARY | \
+				 XFS_SICK_RT_SECONDARY | \
+				 XFS_SICK_RT_INDIRECT)
+
+#define XFS_SICK_AG_ALL	(XFS_SICK_AG_PRIMARY | \
+				 XFS_SICK_AG_SECONDARY | \
+				 XFS_SICK_AG_INDIRECT)
+
+#define XFS_SICK_INO_ALL	(XFS_SICK_INO_PRIMARY | \
+				 XFS_SICK_INO_SECONDARY | \
+				 XFS_SICK_INO_INDIRECT | \
+				 XFS_SICK_INO_ZAPPED)
+
+/*
+ * These functions must be provided by the xfs implementation.  Function
+ * behavior with respect to the first argument should be as follows:
+ *
+ * xfs_*_mark_sick:        Set the sick flags and do not set checked flags.
+ *                         Runtime code should call this upon encountering
+ *                         a corruption.
+ *
+ * xfs_*_mark_corrupt:     Set the sick and checked flags simultaneously.
+ *                         Fsck tools should call this when corruption is
+ *                         found.
+ *
+ * xfs_*_mark_healthy:     Clear the sick flags and set the checked flags.
+ *                         Fsck tools should call this after correcting errors.
+ *
+ * xfs_*_measure_sickness: Return the sick and check status in the provided
+ *                         out parameters.
+ */
 
 void xfs_fs_mark_sick(struct xfs_mount *mp, unsigned int mask);
+void xfs_fs_mark_corrupt(struct xfs_mount *mp, unsigned int mask);
 void xfs_fs_mark_healthy(struct xfs_mount *mp, unsigned int mask);
 void xfs_fs_measure_sickness(struct xfs_mount *mp, unsigned int *sick,
 		unsigned int *checked);
 
 void xfs_rt_mark_sick(struct xfs_mount *mp, unsigned int mask);
+void xfs_rt_mark_corrupt(struct xfs_mount *mp, unsigned int mask);
 void xfs_rt_mark_healthy(struct xfs_mount *mp, unsigned int mask);
 void xfs_rt_measure_sickness(struct xfs_mount *mp, unsigned int *sick,
 		unsigned int *checked);
 
+void xfs_agno_mark_sick(struct xfs_mount *mp, xfs_agnumber_t agno,
+		unsigned int mask);
 void xfs_ag_mark_sick(struct xfs_perag *pag, unsigned int mask);
+void xfs_ag_mark_corrupt(struct xfs_perag *pag, unsigned int mask);
 void xfs_ag_mark_healthy(struct xfs_perag *pag, unsigned int mask);
 void xfs_ag_measure_sickness(struct xfs_perag *pag, unsigned int *sick,
 		unsigned int *checked);
 
 void xfs_inode_mark_sick(struct xfs_inode *ip, unsigned int mask);
+void xfs_inode_mark_corrupt(struct xfs_inode *ip, unsigned int mask);
 void xfs_inode_mark_healthy(struct xfs_inode *ip, unsigned int mask);
 void xfs_inode_measure_sickness(struct xfs_inode *ip, unsigned int *sick,
 		unsigned int *checked);
 
 void xfs_health_unmount(struct xfs_mount *mp);
+void xfs_bmap_mark_sick(struct xfs_inode *ip, int whichfork);
+void xfs_btree_mark_sick(struct xfs_btree_cur *cur);
+void xfs_dirattr_mark_sick(struct xfs_inode *ip, int whichfork);
+void xfs_da_mark_sick(struct xfs_da_args *args);
 
 /* Now some helpers. */
 
@@ -197,4 +281,7 @@ void xfs_fsop_geom_health(struct xfs_mount *mp, struct xfs_fsop_geom *geo);
 void xfs_ag_geom_health(struct xfs_perag *pag, struct xfs_ag_geometry *ageo);
 void xfs_bulkstat_health(struct xfs_inode *ip, struct xfs_bulkstat *bs);
 
+#define xfs_metadata_is_sick(error) \
+	(unlikely((error) == -EFSCORRUPTED || (error) == -EFSBADCRC))
+
 #endif	/* __XFS_HEALTH_H__ */
diff --git a/fs/xfs/libxfs/xfs_ialloc.c b/fs/xfs/libxfs/xfs_ialloc.c
index 2361a22035..14c81f227c 100644
--- a/fs/xfs/libxfs/xfs_ialloc.c
+++ b/fs/xfs/libxfs/xfs_ialloc.c
@@ -27,6 +27,7 @@
 #include "xfs_log.h"
 #include "xfs_rmap.h"
 #include "xfs_ag.h"
+#include "xfs_health.h"
 
 /*
  * Lookup a record by ino in the btree given by cur.
@@ -140,13 +141,13 @@ xfs_inobt_complain_bad_rec(
 	struct xfs_mount		*mp = cur->bc_mp;
 
 	xfs_warn(mp,
-		"%s Inode BTree record corruption in AG %d detected at %pS!",
-		cur->bc_btnum == XFS_BTNUM_INO ? "Used" : "Free",
-		cur->bc_ag.pag->pag_agno, fa);
+		"%sbt record corruption in AG %d detected at %pS!",
+		cur->bc_ops->name, cur->bc_ag.pag->pag_agno, fa);
 	xfs_warn(mp,
 "start inode 0x%x, count 0x%x, free 0x%x freemask 0x%llx, holemask 0x%x",
 		irec->ir_startino, irec->ir_count, irec->ir_freecount,
 		irec->ir_free, irec->ir_holemask);
+	xfs_btree_mark_sick(cur);
 	return -EFSCORRUPTED;
 }
 
@@ -205,14 +206,17 @@ xfs_inobt_insert(
 	struct xfs_buf		*agbp,
 	xfs_agino_t		newino,
 	xfs_agino_t		newlen,
-	xfs_btnum_t		btnum)
+	bool			is_finobt)
 {
 	struct xfs_btree_cur	*cur;
 	xfs_agino_t		thisino;
 	int			i;
 	int			error;
 
-	cur = xfs_inobt_init_cursor(pag, tp, agbp, btnum);
+	if (is_finobt)
+		cur = xfs_finobt_init_cursor(pag, tp, agbp);
+	else
+		cur = xfs_inobt_init_cursor(pag, tp, agbp);
 
 	for (thisino = newino;
 	     thisino < newino + newlen;
@@ -528,16 +532,14 @@ __xfs_inobt_rec_merge(
 }
 
 /*
- * Insert a new sparse inode chunk into the associated inode btree. The inode
- * record for the sparse chunk is pre-aligned to a startino that should match
- * any pre-existing sparse inode record in the tree. This allows sparse chunks
- * to fill over time.
+ * Insert a new sparse inode chunk into the associated inode allocation btree.
+ * The inode record for the sparse chunk is pre-aligned to a startino that
+ * should match any pre-existing sparse inode record in the tree. This allows
+ * sparse chunks to fill over time.
  *
- * This function supports two modes of handling preexisting records depending on
- * the merge flag. If merge is true, the provided record is merged with the
+ * If no preexisting record exists, the provided record is inserted.
+ * If there is a preexisting record, the provided record is merged with the
  * existing record and updated in place. The merged record is returned in nrec.
- * If merge is false, an existing record is replaced with the provided record.
- * If no preexisting record exists, the provided record is always inserted.
  *
  * It is considered corruption if a merge is requested and not possible. Given
  * the sparse inode alignment constraints, this should never happen.
@@ -547,9 +549,7 @@ xfs_inobt_insert_sprec(
 	struct xfs_perag		*pag,
 	struct xfs_trans		*tp,
 	struct xfs_buf			*agbp,
-	int				btnum,
-	struct xfs_inobt_rec_incore	*nrec,	/* in/out: new/merged rec. */
-	bool				merge)	/* merge or replace */
+	struct xfs_inobt_rec_incore	*nrec)	/* in/out: new/merged rec. */
 {
 	struct xfs_mount		*mp = pag->pag_mount;
 	struct xfs_btree_cur		*cur;
@@ -557,7 +557,7 @@ xfs_inobt_insert_sprec(
 	int				i;
 	struct xfs_inobt_rec_incore	rec;
 
-	cur = xfs_inobt_init_cursor(pag, tp, agbp, btnum);
+	cur = xfs_inobt_init_cursor(pag, tp, agbp);
 
 	/* the new record is pre-aligned so we know where to look */
 	error = xfs_inobt_lookup(cur, nrec->ir_startino, XFS_LOOKUP_EQ, &i);
@@ -571,6 +571,7 @@ xfs_inobt_insert_sprec(
 		if (error)
 			goto error;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto error;
 		}
@@ -579,45 +580,45 @@ xfs_inobt_insert_sprec(
 	}
 
 	/*
-	 * A record exists at this startino. Merge or replace the record
-	 * depending on what we've been asked to do.
+	 * A record exists at this startino.  Merge the records.
 	 */
-	if (merge) {
-		error = xfs_inobt_get_rec(cur, &rec, &i);
-		if (error)
-			goto error;
-		if (XFS_IS_CORRUPT(mp, i != 1)) {
-			error = -EFSCORRUPTED;
-			goto error;
-		}
-		if (XFS_IS_CORRUPT(mp, rec.ir_startino != nrec->ir_startino)) {
-			error = -EFSCORRUPTED;
-			goto error;
-		}
+	error = xfs_inobt_get_rec(cur, &rec, &i);
+	if (error)
+		goto error;
+	if (XFS_IS_CORRUPT(mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
+		error = -EFSCORRUPTED;
+		goto error;
+	}
+	if (XFS_IS_CORRUPT(mp, rec.ir_startino != nrec->ir_startino)) {
+		xfs_btree_mark_sick(cur);
+		error = -EFSCORRUPTED;
+		goto error;
+	}
 
-		/*
-		 * This should never fail. If we have coexisting records that
-		 * cannot merge, something is seriously wrong.
-		 */
-		if (XFS_IS_CORRUPT(mp, !__xfs_inobt_can_merge(nrec, &rec))) {
-			error = -EFSCORRUPTED;
-			goto error;
-		}
+	/*
+	 * This should never fail. If we have coexisting records that
+	 * cannot merge, something is seriously wrong.
+	 */
+	if (XFS_IS_CORRUPT(mp, !__xfs_inobt_can_merge(nrec, &rec))) {
+		xfs_btree_mark_sick(cur);
+		error = -EFSCORRUPTED;
+		goto error;
+	}
 
-		trace_xfs_irec_merge_pre(mp, pag->pag_agno, rec.ir_startino,
-					 rec.ir_holemask, nrec->ir_startino,
-					 nrec->ir_holemask);
+	trace_xfs_irec_merge_pre(mp, pag->pag_agno, rec.ir_startino,
+				 rec.ir_holemask, nrec->ir_startino,
+				 nrec->ir_holemask);
 
-		/* merge to nrec to output the updated record */
-		__xfs_inobt_rec_merge(nrec, &rec);
+	/* merge to nrec to output the updated record */
+	__xfs_inobt_rec_merge(nrec, &rec);
 
-		trace_xfs_irec_merge_post(mp, pag->pag_agno, nrec->ir_startino,
-					  nrec->ir_holemask);
+	trace_xfs_irec_merge_post(mp, pag->pag_agno, nrec->ir_startino,
+				  nrec->ir_holemask);
 
-		error = xfs_inobt_rec_check_count(mp, nrec);
-		if (error)
-			goto error;
-	}
+	error = xfs_inobt_rec_check_count(mp, nrec);
+	if (error)
+		goto error;
 
 	error = xfs_inobt_update(cur, nrec);
 	if (error)
@@ -632,6 +633,59 @@ error:
 }
 
 /*
+ * Insert a new sparse inode chunk into the free inode btree. The inode
+ * record for the sparse chunk is pre-aligned to a startino that should match
+ * any pre-existing sparse inode record in the tree. This allows sparse chunks
+ * to fill over time.
+ *
+ * The new record is always inserted, overwriting a pre-existing record if
+ * there is one.
+ */
+STATIC int
+xfs_finobt_insert_sprec(
+	struct xfs_perag		*pag,
+	struct xfs_trans		*tp,
+	struct xfs_buf			*agbp,
+	struct xfs_inobt_rec_incore	*nrec)	/* in/out: new rec. */
+{
+	struct xfs_mount		*mp = pag->pag_mount;
+	struct xfs_btree_cur		*cur;
+	int				error;
+	int				i;
+
+	cur = xfs_finobt_init_cursor(pag, tp, agbp);
+
+	/* the new record is pre-aligned so we know where to look */
+	error = xfs_inobt_lookup(cur, nrec->ir_startino, XFS_LOOKUP_EQ, &i);
+	if (error)
+		goto error;
+	/* if nothing there, insert a new record and return */
+	if (i == 0) {
+		error = xfs_inobt_insert_rec(cur, nrec->ir_holemask,
+					     nrec->ir_count, nrec->ir_freecount,
+					     nrec->ir_free, &i);
+		if (error)
+			goto error;
+		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
+			error = -EFSCORRUPTED;
+			goto error;
+		}
+	} else {
+		error = xfs_inobt_update(cur, nrec);
+		if (error)
+			goto error;
+	}
+
+	xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
+	return 0;
+error:
+	xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
+	return error;
+}
+
+
+/*
  * Allocate new inodes in the allocation group specified by agbp.  Returns 0 if
  * inodes were allocated in this AG; -EAGAIN if there was no space in this AG so
  * the caller knows it can try another AG, a hard -ENOSPC when over the maximum
@@ -857,8 +911,7 @@ sparse_alloc:
 		 * if necessary. If a merge does occur, rec is updated to the
 		 * merged record.
 		 */
-		error = xfs_inobt_insert_sprec(pag, tp, agbp,
-				XFS_BTNUM_INO, &rec, true);
+		error = xfs_inobt_insert_sprec(pag, tp, agbp, &rec);
 		if (error == -EFSCORRUPTED) {
 			xfs_alert(args.mp,
 	"invalid sparse inode record: ino 0x%llx holemask 0x%x count %u",
@@ -882,21 +935,19 @@ sparse_alloc:
 		 * existing record with this one.
 		 */
 		if (xfs_has_finobt(args.mp)) {
-			error = xfs_inobt_insert_sprec(pag, tp, agbp,
-				       XFS_BTNUM_FINO, &rec, false);
+			error = xfs_finobt_insert_sprec(pag, tp, agbp, &rec);
 			if (error)
 				return error;
 		}
 	} else {
 		/* full chunk - insert new records to both btrees */
-		error = xfs_inobt_insert(pag, tp, agbp, newino, newlen,
-					 XFS_BTNUM_INO);
+		error = xfs_inobt_insert(pag, tp, agbp, newino, newlen, false);
 		if (error)
 			return error;
 
 		if (xfs_has_finobt(args.mp)) {
 			error = xfs_inobt_insert(pag, tp, agbp, newino,
-						 newlen, XFS_BTNUM_FINO);
+						 newlen, true);
 			if (error)
 				return error;
 		}
@@ -949,8 +1000,10 @@ xfs_ialloc_next_rec(
 		error = xfs_inobt_get_rec(cur, rec, &i);
 		if (error)
 			return error;
-		if (XFS_IS_CORRUPT(cur->bc_mp, i != 1))
+		if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			return -EFSCORRUPTED;
+		}
 	}
 
 	return 0;
@@ -974,8 +1027,10 @@ xfs_ialloc_get_rec(
 		error = xfs_inobt_get_rec(cur, rec, &i);
 		if (error)
 			return error;
-		if (XFS_IS_CORRUPT(cur->bc_mp, i != 1))
+		if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			return -EFSCORRUPTED;
+		}
 	}
 
 	return 0;
@@ -1003,6 +1058,33 @@ xfs_inobt_first_free_inode(
 }
 
 /*
+ * If this AG has corrupt inodes, check if allocating this inode would fail
+ * with corruption errors.  Returns 0 if we're clear, or EAGAIN to try again
+ * somewhere else.
+ */
+static int
+xfs_dialloc_check_ino(
+	struct xfs_perag	*pag,
+	struct xfs_trans	*tp,
+	xfs_ino_t		ino)
+{
+	struct xfs_imap		imap;
+	struct xfs_buf		*bp;
+	int			error;
+
+	error = xfs_imap(pag, tp, ino, &imap, 0);
+	if (error)
+		return -EAGAIN;
+
+	error = xfs_imap_to_bp(pag->pag_mount, tp, &imap, &bp);
+	if (error)
+		return -EAGAIN;
+
+	xfs_trans_brelse(tp, bp);
+	return 0;
+}
+
+/*
  * Allocate an inode using the inobt-only algorithm.
  */
 STATIC int
@@ -1030,7 +1112,7 @@ xfs_dialloc_ag_inobt(
 	ASSERT(pag->pagi_freecount > 0);
 
  restart_pagno:
-	cur = xfs_inobt_init_cursor(pag, tp, agbp, XFS_BTNUM_INO);
+	cur = xfs_inobt_init_cursor(pag, tp, agbp);
 	/*
 	 * If pagino is 0 (this is the root inode allocation) use newino.
 	 * This must work because we've just allocated some.
@@ -1053,6 +1135,7 @@ xfs_dialloc_ag_inobt(
 		if (error)
 			goto error0;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -1061,6 +1144,7 @@ xfs_dialloc_ag_inobt(
 		if (error)
 			goto error0;
 		if (XFS_IS_CORRUPT(mp, j != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -1219,6 +1303,7 @@ xfs_dialloc_ag_inobt(
 	if (error)
 		goto error0;
 	if (XFS_IS_CORRUPT(mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto error0;
 	}
@@ -1228,6 +1313,7 @@ xfs_dialloc_ag_inobt(
 		if (error)
 			goto error0;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -1237,6 +1323,7 @@ xfs_dialloc_ag_inobt(
 		if (error)
 			goto error0;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto error0;
 		}
@@ -1249,6 +1336,13 @@ alloc_inode:
 	ASSERT((XFS_AGINO_TO_OFFSET(mp, rec.ir_startino) %
 				   XFS_INODES_PER_CHUNK) == 0);
 	ino = XFS_AGINO_TO_INO(mp, pag->pag_agno, rec.ir_startino + offset);
+
+	if (xfs_ag_has_sickness(pag, XFS_SICK_AG_INODES)) {
+		error = xfs_dialloc_check_ino(pag, tp, ino);
+		if (error)
+			goto error0;
+	}
+
 	rec.ir_free &= ~XFS_INOBT_MASK(offset);
 	rec.ir_freecount--;
 	error = xfs_inobt_update(cur, &rec);
@@ -1297,8 +1391,10 @@ xfs_dialloc_ag_finobt_near(
 		error = xfs_inobt_get_rec(lcur, rec, &i);
 		if (error)
 			return error;
-		if (XFS_IS_CORRUPT(lcur->bc_mp, i != 1))
+		if (XFS_IS_CORRUPT(lcur->bc_mp, i != 1)) {
+			xfs_btree_mark_sick(lcur);
 			return -EFSCORRUPTED;
+		}
 
 		/*
 		 * See if we've landed in the parent inode record. The finobt
@@ -1322,12 +1418,14 @@ xfs_dialloc_ag_finobt_near(
 		if (error)
 			goto error_rcur;
 		if (XFS_IS_CORRUPT(lcur->bc_mp, j != 1)) {
+			xfs_btree_mark_sick(lcur);
 			error = -EFSCORRUPTED;
 			goto error_rcur;
 		}
 	}
 
 	if (XFS_IS_CORRUPT(lcur->bc_mp, i != 1 && j != 1)) {
+		xfs_btree_mark_sick(lcur);
 		error = -EFSCORRUPTED;
 		goto error_rcur;
 	}
@@ -1383,8 +1481,10 @@ xfs_dialloc_ag_finobt_newino(
 			error = xfs_inobt_get_rec(cur, rec, &i);
 			if (error)
 				return error;
-			if (XFS_IS_CORRUPT(cur->bc_mp, i != 1))
+			if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				return -EFSCORRUPTED;
+			}
 			return 0;
 		}
 	}
@@ -1395,14 +1495,18 @@ xfs_dialloc_ag_finobt_newino(
 	error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &i);
 	if (error)
 		return error;
-	if (XFS_IS_CORRUPT(cur->bc_mp, i != 1))
+	if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		return -EFSCORRUPTED;
+	}
 
 	error = xfs_inobt_get_rec(cur, rec, &i);
 	if (error)
 		return error;
-	if (XFS_IS_CORRUPT(cur->bc_mp, i != 1))
+	if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		return -EFSCORRUPTED;
+	}
 
 	return 0;
 }
@@ -1424,14 +1528,18 @@ xfs_dialloc_ag_update_inobt(
 	error = xfs_inobt_lookup(cur, frec->ir_startino, XFS_LOOKUP_EQ, &i);
 	if (error)
 		return error;
-	if (XFS_IS_CORRUPT(cur->bc_mp, i != 1))
+	if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		return -EFSCORRUPTED;
+	}
 
 	error = xfs_inobt_get_rec(cur, &rec, &i);
 	if (error)
 		return error;
-	if (XFS_IS_CORRUPT(cur->bc_mp, i != 1))
+	if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		return -EFSCORRUPTED;
+	}
 	ASSERT((XFS_AGINO_TO_OFFSET(cur->bc_mp, rec.ir_startino) %
 				   XFS_INODES_PER_CHUNK) == 0);
 
@@ -1440,8 +1548,10 @@ xfs_dialloc_ag_update_inobt(
 
 	if (XFS_IS_CORRUPT(cur->bc_mp,
 			   rec.ir_free != frec->ir_free ||
-			   rec.ir_freecount != frec->ir_freecount))
+			   rec.ir_freecount != frec->ir_freecount)) {
+		xfs_btree_mark_sick(cur);
 		return -EFSCORRUPTED;
+	}
 
 	return xfs_inobt_update(cur, &rec);
 }
@@ -1483,7 +1593,7 @@ xfs_dialloc_ag(
 	if (!pagino)
 		pagino = be32_to_cpu(agi->agi_newino);
 
-	cur = xfs_inobt_init_cursor(pag, tp, agbp, XFS_BTNUM_FINO);
+	cur = xfs_finobt_init_cursor(pag, tp, agbp);
 
 	error = xfs_check_agi_freecount(cur);
 	if (error)
@@ -1508,6 +1618,12 @@ xfs_dialloc_ag(
 				   XFS_INODES_PER_CHUNK) == 0);
 	ino = XFS_AGINO_TO_INO(mp, pag->pag_agno, rec.ir_startino + offset);
 
+	if (xfs_ag_has_sickness(pag, XFS_SICK_AG_INODES)) {
+		error = xfs_dialloc_check_ino(pag, tp, ino);
+		if (error)
+			goto error_cur;
+	}
+
 	/*
 	 * Modify or remove the finobt record.
 	 */
@@ -1526,7 +1642,7 @@ xfs_dialloc_ag(
 	 * the original freecount. If all is well, make the equivalent update to
 	 * the inobt using the finobt record and offset information.
 	 */
-	icur = xfs_inobt_init_cursor(pag, tp, agbp, XFS_BTNUM_INO);
+	icur = xfs_inobt_init_cursor(pag, tp, agbp);
 
 	error = xfs_check_agi_freecount(icur);
 	if (error)
@@ -1623,7 +1739,7 @@ xfs_dialloc_good_ag(
 		return false;
 
 	if (!xfs_perag_initialised_agi(pag)) {
-		error = xfs_ialloc_read_agi(pag, tp, NULL);
+		error = xfs_ialloc_read_agi(pag, tp, 0, NULL);
 		if (error)
 			return false;
 	}
@@ -1692,7 +1808,7 @@ xfs_dialloc_try_ag(
 	 * Then read in the AGI buffer and recheck with the AGI buffer
 	 * lock held.
 	 */
-	error = xfs_ialloc_read_agi(pag, *tpp, &agbp);
+	error = xfs_ialloc_read_agi(pag, *tpp, 0, &agbp);
 	if (error)
 		return error;
 
@@ -1943,7 +2059,7 @@ xfs_difree_inobt(
 	/*
 	 * Initialize the cursor.
 	 */
-	cur = xfs_inobt_init_cursor(pag, tp, agbp, XFS_BTNUM_INO);
+	cur = xfs_inobt_init_cursor(pag, tp, agbp);
 
 	error = xfs_check_agi_freecount(cur);
 	if (error)
@@ -1958,6 +2074,7 @@ xfs_difree_inobt(
 		goto error0;
 	}
 	if (XFS_IS_CORRUPT(mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto error0;
 	}
@@ -1968,6 +2085,7 @@ xfs_difree_inobt(
 		goto error0;
 	}
 	if (XFS_IS_CORRUPT(mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto error0;
 	}
@@ -2068,7 +2186,7 @@ xfs_difree_finobt(
 	int				error;
 	int				i;
 
-	cur = xfs_inobt_init_cursor(pag, tp, agbp, XFS_BTNUM_FINO);
+	cur = xfs_finobt_init_cursor(pag, tp, agbp);
 
 	error = xfs_inobt_lookup(cur, ibtrec->ir_startino, XFS_LOOKUP_EQ, &i);
 	if (error)
@@ -2080,6 +2198,7 @@ xfs_difree_finobt(
 		 * something is out of sync.
 		 */
 		if (XFS_IS_CORRUPT(mp, ibtrec->ir_freecount != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto error;
 		}
@@ -2106,6 +2225,7 @@ xfs_difree_finobt(
 	if (error)
 		goto error;
 	if (XFS_IS_CORRUPT(mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto error;
 	}
@@ -2116,6 +2236,7 @@ xfs_difree_finobt(
 	if (XFS_IS_CORRUPT(mp,
 			   rec.ir_free != ibtrec->ir_free ||
 			   rec.ir_freecount != ibtrec->ir_freecount)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto error;
 	}
@@ -2205,7 +2326,7 @@ xfs_difree(
 	/*
 	 * Get the allocation group header.
 	 */
-	error = xfs_ialloc_read_agi(pag, tp, &agbp);
+	error = xfs_ialloc_read_agi(pag, tp, 0, &agbp);
 	if (error) {
 		xfs_warn(mp, "%s: xfs_ialloc_read_agi() returned error %d.",
 			__func__, error);
@@ -2251,7 +2372,7 @@ xfs_imap_lookup(
 	int			error;
 	int			i;
 
-	error = xfs_ialloc_read_agi(pag, tp, &agbp);
+	error = xfs_ialloc_read_agi(pag, tp, 0, &agbp);
 	if (error) {
 		xfs_alert(mp,
 			"%s: xfs_ialloc_read_agi() returned error %d, agno %d",
@@ -2265,7 +2386,7 @@ xfs_imap_lookup(
 	 * we have a record, we need to ensure it contains the inode number
 	 * we are looking up.
 	 */
-	cur = xfs_inobt_init_cursor(pag, tp, agbp, XFS_BTNUM_INO);
+	cur = xfs_inobt_init_cursor(pag, tp, agbp);
 	error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, &i);
 	if (!error) {
 		if (i)
@@ -2594,6 +2715,7 @@ int
 xfs_read_agi(
 	struct xfs_perag	*pag,
 	struct xfs_trans	*tp,
+	xfs_buf_flags_t		flags,
 	struct xfs_buf		**agibpp)
 {
 	struct xfs_mount	*mp = pag->pag_mount;
@@ -2603,7 +2725,9 @@ xfs_read_agi(
 
 	error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp,
 			XFS_AG_DADDR(mp, pag->pag_agno, XFS_AGI_DADDR(mp)),
-			XFS_FSS_TO_BB(mp, 1), 0, agibpp, &xfs_agi_buf_ops);
+			XFS_FSS_TO_BB(mp, 1), flags, agibpp, &xfs_agi_buf_ops);
+	if (xfs_metadata_is_sick(error))
+		xfs_ag_mark_sick(pag, XFS_SICK_AG_AGI);
 	if (error)
 		return error;
 	if (tp)
@@ -2621,6 +2745,7 @@ int
 xfs_ialloc_read_agi(
 	struct xfs_perag	*pag,
 	struct xfs_trans	*tp,
+	int			flags,
 	struct xfs_buf		**agibpp)
 {
 	struct xfs_buf		*agibp;
@@ -2629,7 +2754,9 @@ xfs_ialloc_read_agi(
 
 	trace_xfs_ialloc_read_agi(pag->pag_mount, pag->pag_agno);
 
-	error = xfs_read_agi(pag, tp, &agibp);
+	error = xfs_read_agi(pag, tp,
+			(flags & XFS_IALLOC_FLAG_TRYLOCK) ? XBF_TRYLOCK : 0,
+			&agibp);
 	if (error)
 		return error;
 
@@ -2765,7 +2892,7 @@ xfs_ialloc_count_inodes(
 	struct xfs_ialloc_count_inodes	ci = {0};
 	int				error;
 
-	ASSERT(cur->bc_btnum == XFS_BTNUM_INO);
+	ASSERT(xfs_btree_is_ino(cur->bc_ops));
 	error = xfs_btree_query_all(cur, xfs_ialloc_count_inodes_rec, &ci);
 	if (error)
 		return error;
@@ -2982,7 +3109,7 @@ xfs_ialloc_check_shrink(
 	if (!xfs_has_sparseinodes(pag->pag_mount))
 		return 0;
 
-	cur = xfs_inobt_init_cursor(pag, tp, agibp, XFS_BTNUM_INO);
+	cur = xfs_inobt_init_cursor(pag, tp, agibp);
 
 	/* Look up the inobt record that would correspond to the new EOFS. */
 	agino = XFS_AGB_TO_AGINO(pag->pag_mount, new_length);
@@ -2995,6 +3122,7 @@ xfs_ialloc_check_shrink(
 		goto out;
 
 	if (!has) {
+		xfs_ag_mark_sick(pag, XFS_SICK_AG_INOBT);
 		error = -EFSCORRUPTED;
 		goto out;
 	}
diff --git a/fs/xfs/libxfs/xfs_ialloc.h b/fs/xfs/libxfs/xfs_ialloc.h
index f1412183bb..b549627e3a 100644
--- a/fs/xfs/libxfs/xfs_ialloc.h
+++ b/fs/xfs/libxfs/xfs_ialloc.h
@@ -63,10 +63,11 @@ xfs_ialloc_log_agi(
 	struct xfs_buf	*bp,		/* allocation group header buffer */
 	uint32_t	fields);	/* bitmask of fields to log */
 
-int xfs_read_agi(struct xfs_perag *pag, struct xfs_trans *tp,
+int xfs_read_agi(struct xfs_perag *pag, struct xfs_trans *tp, xfs_buf_flags_t flags,
 		struct xfs_buf **agibpp);
 int xfs_ialloc_read_agi(struct xfs_perag *pag, struct xfs_trans *tp,
-		struct xfs_buf **agibpp);
+		int flags, struct xfs_buf **agibpp);
+#define	XFS_IALLOC_FLAG_TRYLOCK	(1U << 0)  /* use trylock for buffer locking */
 
 /*
  * Lookup a record by ino in the btree given by cur.
diff --git a/fs/xfs/libxfs/xfs_ialloc_btree.c b/fs/xfs/libxfs/xfs_ialloc_btree.c
index 42a5e1f227..42e9fd47f6 100644
--- a/fs/xfs/libxfs/xfs_ialloc_btree.c
+++ b/fs/xfs/libxfs/xfs_ialloc_btree.c
@@ -17,6 +17,7 @@
 #include "xfs_ialloc_btree.h"
 #include "xfs_alloc.h"
 #include "xfs_error.h"
+#include "xfs_health.h"
 #include "xfs_trace.h"
 #include "xfs_trans.h"
 #include "xfs_rmap.h"
@@ -37,7 +38,15 @@ xfs_inobt_dup_cursor(
 	struct xfs_btree_cur	*cur)
 {
 	return xfs_inobt_init_cursor(cur->bc_ag.pag, cur->bc_tp,
-			cur->bc_ag.agbp, cur->bc_btnum);
+			cur->bc_ag.agbp);
+}
+
+STATIC struct xfs_btree_cur *
+xfs_finobt_dup_cursor(
+	struct xfs_btree_cur	*cur)
+{
+	return xfs_finobt_init_cursor(cur->bc_ag.pag, cur->bc_tp,
+			cur->bc_ag.agbp);
 }
 
 STATIC void
@@ -81,9 +90,9 @@ xfs_inobt_mod_blockcount(
 	if (!xfs_has_inobtcounts(cur->bc_mp))
 		return;
 
-	if (cur->bc_btnum == XFS_BTNUM_FINO)
+	if (xfs_btree_is_fino(cur->bc_ops))
 		be32_add_cpu(&agi->agi_fblocks, howmuch);
-	else if (cur->bc_btnum == XFS_BTNUM_INO)
+	else
 		be32_add_cpu(&agi->agi_iblocks, howmuch);
 	xfs_ialloc_log_agi(cur->bc_tp, agbp, XFS_AGI_IBLOCKS);
 }
@@ -300,7 +309,7 @@ xfs_inobt_verify(
 	 * xfs_perag_initialised_agi(pag)) if we ever do.
 	 */
 	if (xfs_has_crc(mp)) {
-		fa = xfs_btree_sblock_v5hdr_verify(bp);
+		fa = xfs_btree_agblock_v5hdr_verify(bp);
 		if (fa)
 			return fa;
 	}
@@ -310,7 +319,7 @@ xfs_inobt_verify(
 	if (level >= M_IGEO(mp)->inobt_maxlevels)
 		return __this_address;
 
-	return xfs_btree_sblock_verify(bp,
+	return xfs_btree_agblock_verify(bp,
 			M_IGEO(mp)->inobt_mxr[level != 0]);
 }
 
@@ -320,7 +329,7 @@ xfs_inobt_read_verify(
 {
 	xfs_failaddr_t	fa;
 
-	if (!xfs_btree_sblock_verify_crc(bp))
+	if (!xfs_btree_agblock_verify_crc(bp))
 		xfs_verifier_error(bp, -EFSBADCRC, __this_address);
 	else {
 		fa = xfs_inobt_verify(bp);
@@ -344,7 +353,7 @@ xfs_inobt_write_verify(
 		xfs_verifier_error(bp, -EFSCORRUPTED, fa);
 		return;
 	}
-	xfs_btree_sblock_calc_crc(bp);
+	xfs_btree_agblock_calc_crc(bp);
 
 }
 
@@ -398,9 +407,17 @@ xfs_inobt_keys_contiguous(
 				 be32_to_cpu(key2->inobt.ir_startino));
 }
 
-static const struct xfs_btree_ops xfs_inobt_ops = {
+const struct xfs_btree_ops xfs_inobt_ops = {
+	.name			= "ino",
+	.type			= XFS_BTREE_TYPE_AG,
+
 	.rec_len		= sizeof(xfs_inobt_rec_t),
 	.key_len		= sizeof(xfs_inobt_key_t),
+	.ptr_len		= XFS_BTREE_SHORT_PTR_LEN,
+
+	.lru_refs		= XFS_INO_BTREE_REF,
+	.statoff		= XFS_STATS_CALC_INDEX(xs_ibt_2),
+	.sick_mask		= XFS_SICK_AG_INOBT,
 
 	.dup_cursor		= xfs_inobt_dup_cursor,
 	.set_root		= xfs_inobt_set_root,
@@ -420,11 +437,19 @@ static const struct xfs_btree_ops xfs_inobt_ops = {
 	.keys_contiguous	= xfs_inobt_keys_contiguous,
 };
 
-static const struct xfs_btree_ops xfs_finobt_ops = {
+const struct xfs_btree_ops xfs_finobt_ops = {
+	.name			= "fino",
+	.type			= XFS_BTREE_TYPE_AG,
+
 	.rec_len		= sizeof(xfs_inobt_rec_t),
 	.key_len		= sizeof(xfs_inobt_key_t),
+	.ptr_len		= XFS_BTREE_SHORT_PTR_LEN,
 
-	.dup_cursor		= xfs_inobt_dup_cursor,
+	.lru_refs		= XFS_INO_BTREE_REF,
+	.statoff		= XFS_STATS_CALC_INDEX(xs_fibt_2),
+	.sick_mask		= XFS_SICK_AG_FINOBT,
+
+	.dup_cursor		= xfs_finobt_dup_cursor,
 	.set_root		= xfs_finobt_set_root,
 	.alloc_block		= xfs_finobt_alloc_block,
 	.free_block		= xfs_finobt_free_block,
@@ -443,65 +468,54 @@ static const struct xfs_btree_ops xfs_finobt_ops = {
 };
 
 /*
- * Initialize a new inode btree cursor.
+ * Create an inode btree cursor.
+ *
+ * For staging cursors tp and agbp are NULL.
  */
-static struct xfs_btree_cur *
-xfs_inobt_init_common(
+struct xfs_btree_cur *
+xfs_inobt_init_cursor(
 	struct xfs_perag	*pag,
-	struct xfs_trans	*tp,		/* transaction pointer */
-	xfs_btnum_t		btnum)		/* ialloc or free ino btree */
+	struct xfs_trans	*tp,
+	struct xfs_buf		*agbp)
 {
 	struct xfs_mount	*mp = pag->pag_mount;
 	struct xfs_btree_cur	*cur;
 
-	cur = xfs_btree_alloc_cursor(mp, tp, btnum,
+	cur = xfs_btree_alloc_cursor(mp, tp, &xfs_inobt_ops,
 			M_IGEO(mp)->inobt_maxlevels, xfs_inobt_cur_cache);
-	if (btnum == XFS_BTNUM_INO) {
-		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_ibt_2);
-		cur->bc_ops = &xfs_inobt_ops;
-	} else {
-		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_fibt_2);
-		cur->bc_ops = &xfs_finobt_ops;
-	}
-
-	if (xfs_has_crc(mp))
-		cur->bc_flags |= XFS_BTREE_CRC_BLOCKS;
-
 	cur->bc_ag.pag = xfs_perag_hold(pag);
+	cur->bc_ag.agbp = agbp;
+	if (agbp) {
+		struct xfs_agi		*agi = agbp->b_addr;
+
+		cur->bc_nlevels = be32_to_cpu(agi->agi_level);
+	}
 	return cur;
 }
 
-/* Create an inode btree cursor. */
+/*
+ * Create a free inode btree cursor.
+ *
+ * For staging cursors tp and agbp are NULL.
+ */
 struct xfs_btree_cur *
-xfs_inobt_init_cursor(
+xfs_finobt_init_cursor(
 	struct xfs_perag	*pag,
 	struct xfs_trans	*tp,
-	struct xfs_buf		*agbp,
-	xfs_btnum_t		btnum)
+	struct xfs_buf		*agbp)
 {
+	struct xfs_mount	*mp = pag->pag_mount;
 	struct xfs_btree_cur	*cur;
-	struct xfs_agi		*agi = agbp->b_addr;
 
-	cur = xfs_inobt_init_common(pag, tp, btnum);
-	if (btnum == XFS_BTNUM_INO)
-		cur->bc_nlevels = be32_to_cpu(agi->agi_level);
-	else
-		cur->bc_nlevels = be32_to_cpu(agi->agi_free_level);
+	cur = xfs_btree_alloc_cursor(mp, tp, &xfs_finobt_ops,
+			M_IGEO(mp)->inobt_maxlevels, xfs_inobt_cur_cache);
+	cur->bc_ag.pag = xfs_perag_hold(pag);
 	cur->bc_ag.agbp = agbp;
-	return cur;
-}
+	if (agbp) {
+		struct xfs_agi		*agi = agbp->b_addr;
 
-/* Create an inode btree cursor with a fake root for staging. */
-struct xfs_btree_cur *
-xfs_inobt_stage_cursor(
-	struct xfs_perag	*pag,
-	struct xbtree_afakeroot	*afake,
-	xfs_btnum_t		btnum)
-{
-	struct xfs_btree_cur	*cur;
-
-	cur = xfs_inobt_init_common(pag, NULL, btnum);
-	xfs_btree_stage_afakeroot(cur, afake);
+		cur->bc_nlevels = be32_to_cpu(agi->agi_free_level);
+	}
 	return cur;
 }
 
@@ -521,7 +535,7 @@ xfs_inobt_commit_staged_btree(
 
 	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
 
-	if (cur->bc_btnum == XFS_BTNUM_INO) {
+	if (xfs_btree_is_ino(cur->bc_ops)) {
 		fields = XFS_AGI_ROOT | XFS_AGI_LEVEL;
 		agi->agi_root = cpu_to_be32(afake->af_root);
 		agi->agi_level = cpu_to_be32(afake->af_levels);
@@ -530,7 +544,7 @@ xfs_inobt_commit_staged_btree(
 			fields |= XFS_AGI_IBLOCKS;
 		}
 		xfs_ialloc_log_agi(tp, agbp, fields);
-		xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_inobt_ops);
+		xfs_btree_commit_afakeroot(cur, tp, agbp);
 	} else {
 		fields = XFS_AGI_FREE_ROOT | XFS_AGI_FREE_LEVEL;
 		agi->agi_free_root = cpu_to_be32(afake->af_root);
@@ -540,7 +554,7 @@ xfs_inobt_commit_staged_btree(
 			fields |= XFS_AGI_IBLOCKS;
 		}
 		xfs_ialloc_log_agi(tp, agbp, fields);
-		xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_finobt_ops);
+		xfs_btree_commit_afakeroot(cur, tp, agbp);
 	}
 }
 
@@ -721,45 +735,21 @@ xfs_inobt_max_size(
 					XFS_INODES_PER_CHUNK);
 }
 
-/* Read AGI and create inobt cursor. */
-int
-xfs_inobt_cur(
-	struct xfs_perag	*pag,
-	struct xfs_trans	*tp,
-	xfs_btnum_t		which,
-	struct xfs_btree_cur	**curpp,
-	struct xfs_buf		**agi_bpp)
-{
-	struct xfs_btree_cur	*cur;
-	int			error;
-
-	ASSERT(*agi_bpp == NULL);
-	ASSERT(*curpp == NULL);
-
-	error = xfs_ialloc_read_agi(pag, tp, agi_bpp);
-	if (error)
-		return error;
-
-	cur = xfs_inobt_init_cursor(pag, tp, *agi_bpp, which);
-	*curpp = cur;
-	return 0;
-}
-
 static int
-xfs_inobt_count_blocks(
+xfs_finobt_count_blocks(
 	struct xfs_perag	*pag,
 	struct xfs_trans	*tp,
-	xfs_btnum_t		btnum,
 	xfs_extlen_t		*tree_blocks)
 {
 	struct xfs_buf		*agbp = NULL;
-	struct xfs_btree_cur	*cur = NULL;
+	struct xfs_btree_cur	*cur;
 	int			error;
 
-	error = xfs_inobt_cur(pag, tp, btnum, &cur, &agbp);
+	error = xfs_ialloc_read_agi(pag, tp, 0, &agbp);
 	if (error)
 		return error;
 
+	cur = xfs_inobt_init_cursor(pag, tp, agbp);
 	error = xfs_btree_count_blocks(cur, tree_blocks);
 	xfs_btree_del_cursor(cur, error);
 	xfs_trans_brelse(tp, agbp);
@@ -778,7 +768,7 @@ xfs_finobt_read_blocks(
 	struct xfs_agi		*agi;
 	int			error;
 
-	error = xfs_ialloc_read_agi(pag, tp, &agbp);
+	error = xfs_ialloc_read_agi(pag, tp, 0, &agbp);
 	if (error)
 		return error;
 
@@ -807,8 +797,7 @@ xfs_finobt_calc_reserves(
 	if (xfs_has_inobtcounts(pag->pag_mount))
 		error = xfs_finobt_read_blocks(pag, tp, &tree_len);
 	else
-		error = xfs_inobt_count_blocks(pag, tp, XFS_BTNUM_FINO,
-				&tree_len);
+		error = xfs_finobt_count_blocks(pag, tp, &tree_len);
 	if (error)
 		return error;
 
diff --git a/fs/xfs/libxfs/xfs_ialloc_btree.h b/fs/xfs/libxfs/xfs_ialloc_btree.h
index 3262c3fe5e..6472ec1ecb 100644
--- a/fs/xfs/libxfs/xfs_ialloc_btree.h
+++ b/fs/xfs/libxfs/xfs_ialloc_btree.h
@@ -46,10 +46,10 @@ struct xfs_perag;
 		 (maxrecs) * sizeof(xfs_inobt_key_t) + \
 		 ((index) - 1) * sizeof(xfs_inobt_ptr_t)))
 
-extern struct xfs_btree_cur *xfs_inobt_init_cursor(struct xfs_perag *pag,
-		struct xfs_trans *tp, struct xfs_buf *agbp, xfs_btnum_t btnum);
-struct xfs_btree_cur *xfs_inobt_stage_cursor(struct xfs_perag *pag,
-		struct xbtree_afakeroot *afake, xfs_btnum_t btnum);
+struct xfs_btree_cur *xfs_inobt_init_cursor(struct xfs_perag *pag,
+		struct xfs_trans *tp, struct xfs_buf *agbp);
+struct xfs_btree_cur *xfs_finobt_init_cursor(struct xfs_perag *pag,
+		struct xfs_trans *tp, struct xfs_buf *agbp);
 extern int xfs_inobt_maxrecs(struct xfs_mount *, int, int);
 
 /* ir_holemask to inode allocation bitmap conversion */
@@ -66,9 +66,6 @@ int xfs_finobt_calc_reserves(struct xfs_perag *perag, struct xfs_trans *tp,
 		xfs_extlen_t *ask, xfs_extlen_t *used);
 extern xfs_extlen_t xfs_iallocbt_calc_size(struct xfs_mount *mp,
 		unsigned long long len);
-int xfs_inobt_cur(struct xfs_perag *pag, struct xfs_trans *tp,
-		xfs_btnum_t btnum, struct xfs_btree_cur **curpp,
-		struct xfs_buf **agi_bpp);
 
 void xfs_inobt_commit_staged_btree(struct xfs_btree_cur *cur,
 		struct xfs_trans *tp, struct xfs_buf *agbp);
diff --git a/fs/xfs/libxfs/xfs_iext_tree.c b/fs/xfs/libxfs/xfs_iext_tree.c
index f4e6b200cd..8796f2b3e5 100644
--- a/fs/xfs/libxfs/xfs_iext_tree.c
+++ b/fs/xfs/libxfs/xfs_iext_tree.c
@@ -394,11 +394,18 @@ xfs_iext_leaf_key(
 	return leaf->recs[n].lo & XFS_IEXT_STARTOFF_MASK;
 }
 
+static inline void *
+xfs_iext_alloc_node(
+	int	size)
+{
+	return kzalloc(size, GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL);
+}
+
 static void
 xfs_iext_grow(
 	struct xfs_ifork	*ifp)
 {
-	struct xfs_iext_node	*node = kmem_zalloc(NODE_SIZE, KM_NOFS);
+	struct xfs_iext_node	*node = xfs_iext_alloc_node(NODE_SIZE);
 	int			i;
 
 	if (ifp->if_height == 1) {
@@ -454,7 +461,7 @@ xfs_iext_split_node(
 	int			*nr_entries)
 {
 	struct xfs_iext_node	*node = *nodep;
-	struct xfs_iext_node	*new = kmem_zalloc(NODE_SIZE, KM_NOFS);
+	struct xfs_iext_node	*new = xfs_iext_alloc_node(NODE_SIZE);
 	const int		nr_move = KEYS_PER_NODE / 2;
 	int			nr_keep = nr_move + (KEYS_PER_NODE & 1);
 	int			i = 0;
@@ -542,7 +549,7 @@ xfs_iext_split_leaf(
 	int			*nr_entries)
 {
 	struct xfs_iext_leaf	*leaf = cur->leaf;
-	struct xfs_iext_leaf	*new = kmem_zalloc(NODE_SIZE, KM_NOFS);
+	struct xfs_iext_leaf	*new = xfs_iext_alloc_node(NODE_SIZE);
 	const int		nr_move = RECS_PER_LEAF / 2;
 	int			nr_keep = nr_move + (RECS_PER_LEAF & 1);
 	int			i;
@@ -583,7 +590,7 @@ xfs_iext_alloc_root(
 {
 	ASSERT(ifp->if_bytes == 0);
 
-	ifp->if_data = kmem_zalloc(sizeof(struct xfs_iext_rec), KM_NOFS);
+	ifp->if_data = xfs_iext_alloc_node(sizeof(struct xfs_iext_rec));
 	ifp->if_height = 1;
 
 	/* now that we have a node step into it */
@@ -603,7 +610,8 @@ xfs_iext_realloc_root(
 	if (new_size / sizeof(struct xfs_iext_rec) == RECS_PER_LEAF)
 		new_size = NODE_SIZE;
 
-	new = krealloc(ifp->if_data, new_size, GFP_NOFS | __GFP_NOFAIL);
+	new = krealloc(ifp->if_data, new_size,
+			GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL);
 	memset(new + ifp->if_bytes, 0, new_size - ifp->if_bytes);
 	ifp->if_data = new;
 	cur->leaf = new;
@@ -743,7 +751,7 @@ xfs_iext_remove_node(
 again:
 	ASSERT(node->ptrs[pos]);
 	ASSERT(node->ptrs[pos] == victim);
-	kmem_free(victim);
+	kfree(victim);
 
 	nr_entries = xfs_iext_node_nr_entries(node, pos) - 1;
 	offset = node->keys[0];
@@ -789,7 +797,7 @@ again:
 		ASSERT(node == ifp->if_data);
 		ifp->if_data = node->ptrs[0];
 		ifp->if_height--;
-		kmem_free(node);
+		kfree(node);
 	}
 }
 
@@ -863,7 +871,7 @@ xfs_iext_free_last_leaf(
 	struct xfs_ifork	*ifp)
 {
 	ifp->if_height--;
-	kmem_free(ifp->if_data);
+	kfree(ifp->if_data);
 	ifp->if_data = NULL;
 }
 
@@ -1044,7 +1052,7 @@ xfs_iext_destroy_node(
 		}
 	}
 
-	kmem_free(node);
+	kfree(node);
 }
 
 void
diff --git a/fs/xfs/libxfs/xfs_inode_buf.c b/fs/xfs/libxfs/xfs_inode_buf.c
index 137a65bda9..513b50da62 100644
--- a/fs/xfs/libxfs/xfs_inode_buf.c
+++ b/fs/xfs/libxfs/xfs_inode_buf.c
@@ -18,6 +18,7 @@
 #include "xfs_trans.h"
 #include "xfs_ialloc.h"
 #include "xfs_dir2.h"
+#include "xfs_health.h"
 
 #include <linux/iversion.h>
 
@@ -132,9 +133,14 @@ xfs_imap_to_bp(
 	struct xfs_imap		*imap,
 	struct xfs_buf		**bpp)
 {
-	return xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, imap->im_blkno,
-				   imap->im_len, XBF_UNMAPPED, bpp,
-				   &xfs_inode_buf_ops);
+	int			error;
+
+	error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, imap->im_blkno,
+			imap->im_len, XBF_UNMAPPED, bpp, &xfs_inode_buf_ops);
+	if (xfs_metadata_is_sick(error))
+		xfs_agno_mark_sick(mp, xfs_daddr_to_agno(mp, imap->im_blkno),
+				XFS_SICK_AG_INODES);
+	return error;
 }
 
 static inline struct timespec64 xfs_inode_decode_bigtime(uint64_t ts)
@@ -368,17 +374,40 @@ xfs_dinode_verify_fork(
 	/*
 	 * For fork types that can contain local data, check that the fork
 	 * format matches the size of local data contained within the fork.
-	 *
-	 * For all types, check that when the size says the should be in extent
-	 * or btree format, the inode isn't claiming it is in local format.
 	 */
 	if (whichfork == XFS_DATA_FORK) {
-		if (S_ISDIR(mode) || S_ISLNK(mode)) {
+		/*
+		 * A directory small enough to fit in the inode must be stored
+		 * in local format.  The directory sf <-> extents conversion
+		 * code updates the directory size accordingly.  Directories
+		 * being truncated have zero size and are not subject to this
+		 * check.
+		 */
+		if (S_ISDIR(mode)) {
+			if (dip->di_size &&
+			    be64_to_cpu(dip->di_size) <= fork_size &&
+			    fork_format != XFS_DINODE_FMT_LOCAL)
+				return __this_address;
+		}
+
+		/*
+		 * A symlink with a target small enough to fit in the inode can
+		 * be stored in extents format if xattrs were added (thus
+		 * converting the data fork from shortform to remote format)
+		 * and then removed.
+		 */
+		if (S_ISLNK(mode)) {
 			if (be64_to_cpu(dip->di_size) <= fork_size &&
+			    fork_format != XFS_DINODE_FMT_EXTENTS &&
 			    fork_format != XFS_DINODE_FMT_LOCAL)
 				return __this_address;
 		}
 
+		/*
+		 * For all types, check that when the size says the fork should
+		 * be in extent or btree format, the inode isn't claiming to be
+		 * in local format.
+		 */
 		if (be64_to_cpu(dip->di_size) > fork_size &&
 		    fork_format == XFS_DINODE_FMT_LOCAL)
 			return __this_address;
@@ -485,6 +514,14 @@ xfs_dinode_verify(
 			return __this_address;
 	}
 
+	if (dip->di_version > 1) {
+		if (dip->di_onlink)
+			return __this_address;
+	} else {
+		if (dip->di_nlink)
+			return __this_address;
+	}
+
 	/* don't allow invalid i_size */
 	di_size = be64_to_cpu(dip->di_size);
 	if (di_size & (1ULL << 63))
@@ -494,9 +531,19 @@ xfs_dinode_verify(
 	if (mode && xfs_mode_to_ftype(mode) == XFS_DIR3_FT_UNKNOWN)
 		return __this_address;
 
-	/* No zero-length symlinks/dirs. */
-	if ((S_ISLNK(mode) || S_ISDIR(mode)) && di_size == 0)
-		return __this_address;
+	/*
+	 * No zero-length symlinks/dirs unless they're unlinked and hence being
+	 * inactivated.
+	 */
+	if ((S_ISLNK(mode) || S_ISDIR(mode)) && di_size == 0) {
+		if (dip->di_version > 1) {
+			if (dip->di_nlink)
+				return __this_address;
+		} else {
+			if (dip->di_onlink)
+				return __this_address;
+		}
+	}
 
 	fa = xfs_dinode_verify_nrext64(mp, dip);
 	if (fa)
diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c
index f4569e18a8..9d11ae0159 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.c
+++ b/fs/xfs/libxfs/xfs_inode_fork.c
@@ -25,6 +25,8 @@
 #include "xfs_attr_leaf.h"
 #include "xfs_types.h"
 #include "xfs_errortag.h"
+#include "xfs_health.h"
+#include "xfs_symlink_remote.h"
 
 struct kmem_cache *xfs_ifork_cache;
 
@@ -50,7 +52,8 @@ xfs_init_local_fork(
 		mem_size++;
 
 	if (size) {
-		char *new_data = kmem_alloc(mem_size, KM_NOFS);
+		char *new_data = kmalloc(mem_size,
+				GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL);
 
 		memcpy(new_data, data, size);
 		if (zero_terminate)
@@ -77,7 +80,7 @@ xfs_iformat_local(
 	/*
 	 * If the size is unreasonable, then something
 	 * is wrong and we just bail out rather than crash in
-	 * kmem_alloc() or memcpy() below.
+	 * kmalloc() or memcpy() below.
 	 */
 	if (unlikely(size > XFS_DFORK_SIZE(dip, ip->i_mount, whichfork))) {
 		xfs_warn(ip->i_mount,
@@ -87,6 +90,7 @@ xfs_iformat_local(
 		xfs_inode_verifier_error(ip, -EFSCORRUPTED,
 				"xfs_iformat_local", dip, sizeof(*dip),
 				__this_address);
+		xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE);
 		return -EFSCORRUPTED;
 	}
 
@@ -116,7 +120,7 @@ xfs_iformat_extents(
 
 	/*
 	 * If the number of extents is unreasonable, then something is wrong and
-	 * we just bail out rather than crash in kmem_alloc() or memcpy() below.
+	 * we just bail out rather than crash in kmalloc() or memcpy() below.
 	 */
 	if (unlikely(size < 0 || size > XFS_DFORK_SIZE(dip, mp, whichfork))) {
 		xfs_warn(ip->i_mount, "corrupt inode %llu ((a)extents = %llu).",
@@ -124,6 +128,7 @@ xfs_iformat_extents(
 		xfs_inode_verifier_error(ip, -EFSCORRUPTED,
 				"xfs_iformat_extents(1)", dip, sizeof(*dip),
 				__this_address);
+		xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE);
 		return -EFSCORRUPTED;
 	}
 
@@ -143,6 +148,7 @@ xfs_iformat_extents(
 				xfs_inode_verifier_error(ip, -EFSCORRUPTED,
 						"xfs_iformat_extents(2)",
 						dp, sizeof(*dp), fa);
+				xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE);
 				return xfs_bmap_complain_bad_rec(ip, whichfork,
 						fa, &new);
 			}
@@ -201,11 +207,13 @@ xfs_iformat_btree(
 		xfs_inode_verifier_error(ip, -EFSCORRUPTED,
 				"xfs_iformat_btree", dfp, size,
 				__this_address);
+		xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE);
 		return -EFSCORRUPTED;
 	}
 
 	ifp->if_broot_bytes = size;
-	ifp->if_broot = kmem_alloc(size, KM_NOFS);
+	ifp->if_broot = kmalloc(size,
+				GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL);
 	ASSERT(ifp->if_broot != NULL);
 	/*
 	 * Copy and convert from the on-disk structure
@@ -265,12 +273,14 @@ xfs_iformat_data_fork(
 		default:
 			xfs_inode_verifier_error(ip, -EFSCORRUPTED, __func__,
 					dip, sizeof(*dip), __this_address);
+			xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE);
 			return -EFSCORRUPTED;
 		}
 		break;
 	default:
 		xfs_inode_verifier_error(ip, -EFSCORRUPTED, __func__, dip,
 				sizeof(*dip), __this_address);
+		xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE);
 		return -EFSCORRUPTED;
 	}
 }
@@ -342,6 +352,7 @@ xfs_iformat_attr_fork(
 	default:
 		xfs_inode_verifier_error(ip, error, __func__, dip,
 				sizeof(*dip), __this_address);
+		xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE);
 		error = -EFSCORRUPTED;
 		break;
 	}
@@ -399,7 +410,8 @@ xfs_iroot_realloc(
 		 */
 		if (ifp->if_broot_bytes == 0) {
 			new_size = XFS_BMAP_BROOT_SPACE_CALC(mp, rec_diff);
-			ifp->if_broot = kmem_alloc(new_size, KM_NOFS);
+			ifp->if_broot = kmalloc(new_size,
+						GFP_KERNEL | __GFP_NOFAIL);
 			ifp->if_broot_bytes = (int)new_size;
 			return;
 		}
@@ -414,7 +426,7 @@ xfs_iroot_realloc(
 		new_max = cur_max + rec_diff;
 		new_size = XFS_BMAP_BROOT_SPACE_CALC(mp, new_max);
 		ifp->if_broot = krealloc(ifp->if_broot, new_size,
-					 GFP_NOFS | __GFP_NOFAIL);
+					 GFP_KERNEL | __GFP_NOFAIL);
 		op = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, ifp->if_broot, 1,
 						     ifp->if_broot_bytes);
 		np = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, ifp->if_broot, 1,
@@ -440,7 +452,7 @@ xfs_iroot_realloc(
 	else
 		new_size = 0;
 	if (new_size > 0) {
-		new_broot = kmem_alloc(new_size, KM_NOFS);
+		new_broot = kmalloc(new_size, GFP_KERNEL | __GFP_NOFAIL);
 		/*
 		 * First copy over the btree block header.
 		 */
@@ -470,7 +482,7 @@ xfs_iroot_realloc(
 						     (int)new_size);
 		memcpy(np, op, new_max * (uint)sizeof(xfs_fsblock_t));
 	}
-	kmem_free(ifp->if_broot);
+	kfree(ifp->if_broot);
 	ifp->if_broot = new_broot;
 	ifp->if_broot_bytes = (int)new_size;
 	if (ifp->if_broot)
@@ -488,7 +500,7 @@ xfs_iroot_realloc(
  *
  * If the amount of space needed has decreased below the size of the
  * inline buffer, then switch to using the inline buffer.  Otherwise,
- * use kmem_realloc() or kmem_alloc() to adjust the size of the buffer
+ * use krealloc() or kmalloc() to adjust the size of the buffer
  * to what is needed.
  *
  * ip -- the inode whose if_data area is changing
@@ -509,7 +521,7 @@ xfs_idata_realloc(
 
 	if (byte_diff) {
 		ifp->if_data = krealloc(ifp->if_data, new_size,
-					GFP_NOFS | __GFP_NOFAIL);
+					GFP_KERNEL | __GFP_NOFAIL);
 		if (new_size == 0)
 			ifp->if_data = NULL;
 		ifp->if_bytes = new_size;
@@ -524,13 +536,13 @@ xfs_idestroy_fork(
 	struct xfs_ifork	*ifp)
 {
 	if (ifp->if_broot != NULL) {
-		kmem_free(ifp->if_broot);
+		kfree(ifp->if_broot);
 		ifp->if_broot = NULL;
 	}
 
 	switch (ifp->if_format) {
 	case XFS_DINODE_FMT_LOCAL:
-		kmem_free(ifp->if_data);
+		kfree(ifp->if_data);
 		ifp->if_data = NULL;
 		break;
 	case XFS_DINODE_FMT_EXTENTS:
@@ -562,7 +574,7 @@ xfs_iextents_copy(
 	struct xfs_bmbt_irec	rec;
 	int64_t			copied = 0;
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL | XFS_ILOCK_SHARED));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL | XFS_ILOCK_SHARED);
 	ASSERT(ifp->if_bytes > 0);
 
 	for_each_xfs_iext(ifp, &icur, &rec) {
@@ -689,7 +701,7 @@ xfs_ifork_init_cow(
 		return;
 
 	ip->i_cowfp = kmem_cache_zalloc(xfs_ifork_cache,
-				       GFP_NOFS | __GFP_NOFAIL);
+				GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL);
 	ip->i_cowfp->if_format = XFS_DINODE_FMT_EXTENTS;
 }
 
@@ -753,52 +765,54 @@ xfs_ifork_verify_local_attr(
 	return 0;
 }
 
+/*
+ * Check if the inode fork supports adding nr_to_add more extents.
+ *
+ * If it doesn't but we can upgrade it to large extent counters, do the upgrade.
+ * If we can't upgrade or are already using big counters but still can't fit the
+ * additional extents, return -EFBIG.
+ */
 int
-xfs_iext_count_may_overflow(
+xfs_iext_count_extend(
+	struct xfs_trans	*tp,
 	struct xfs_inode	*ip,
 	int			whichfork,
-	int			nr_to_add)
+	uint			nr_to_add)
 {
+	struct xfs_mount	*mp = ip->i_mount;
+	bool			has_large =
+		xfs_inode_has_large_extent_counts(ip);
 	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, whichfork);
-	uint64_t		max_exts;
 	uint64_t		nr_exts;
 
+	ASSERT(nr_to_add <= XFS_MAX_EXTCNT_UPGRADE_NR);
+
 	if (whichfork == XFS_COW_FORK)
 		return 0;
 
-	max_exts = xfs_iext_max_nextents(xfs_inode_has_large_extent_counts(ip),
-				whichfork);
-
-	if (XFS_TEST_ERROR(false, ip->i_mount, XFS_ERRTAG_REDUCE_MAX_IEXTENTS))
-		max_exts = 10;
-
+	/* no point in upgrading if if_nextents overflows */
 	nr_exts = ifp->if_nextents + nr_to_add;
-	if (nr_exts < ifp->if_nextents || nr_exts > max_exts)
+	if (nr_exts < ifp->if_nextents)
+		return -EFBIG;
+
+	if (XFS_TEST_ERROR(false, mp, XFS_ERRTAG_REDUCE_MAX_IEXTENTS) &&
+	    nr_exts > 10)
 		return -EFBIG;
 
+	if (nr_exts > xfs_iext_max_nextents(has_large, whichfork)) {
+		if (has_large || !xfs_has_large_extent_counts(mp))
+			return -EFBIG;
+		ip->i_diflags2 |= XFS_DIFLAG2_NREXT64;
+		xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+	}
 	return 0;
 }
 
-/*
- * Upgrade this inode's extent counter fields to be able to handle a potential
- * increase in the extent count by nr_to_add.  Normally this is the same
- * quantity that caused xfs_iext_count_may_overflow() to return -EFBIG.
- */
-int
-xfs_iext_count_upgrade(
-	struct xfs_trans	*tp,
+/* Decide if a file mapping is on the realtime device or not. */
+bool
+xfs_ifork_is_realtime(
 	struct xfs_inode	*ip,
-	uint			nr_to_add)
+	int			whichfork)
 {
-	ASSERT(nr_to_add <= XFS_MAX_EXTCNT_UPGRADE_NR);
-
-	if (!xfs_has_large_extent_counts(ip->i_mount) ||
-	    xfs_inode_has_large_extent_counts(ip) ||
-	    XFS_TEST_ERROR(false, ip->i_mount, XFS_ERRTAG_REDUCE_MAX_IEXTENTS))
-		return -EFBIG;
-
-	ip->i_diflags2 |= XFS_DIFLAG2_NREXT64;
-	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
-
-	return 0;
+	return XFS_IS_REALTIME_INODE(ip) && whichfork != XFS_ATTR_FORK;
 }
diff --git a/fs/xfs/libxfs/xfs_inode_fork.h b/fs/xfs/libxfs/xfs_inode_fork.h
index 96303249d2..2373d12fd4 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.h
+++ b/fs/xfs/libxfs/xfs_inode_fork.h
@@ -256,10 +256,9 @@ extern void xfs_ifork_init_cow(struct xfs_inode *ip);
 
 int xfs_ifork_verify_local_data(struct xfs_inode *ip);
 int xfs_ifork_verify_local_attr(struct xfs_inode *ip);
-int xfs_iext_count_may_overflow(struct xfs_inode *ip, int whichfork,
-		int nr_to_add);
-int xfs_iext_count_upgrade(struct xfs_trans *tp, struct xfs_inode *ip,
-		uint nr_to_add);
+int xfs_iext_count_extend(struct xfs_trans *tp, struct xfs_inode *ip,
+		int whichfork, uint nr_to_add);
+bool xfs_ifork_is_realtime(struct xfs_inode *ip, int whichfork);
 
 /* returns true if the fork has extents but they are not read in yet. */
 static inline bool xfs_need_iread_extents(const struct xfs_ifork *ifp)
diff --git a/fs/xfs/libxfs/xfs_log_format.h b/fs/xfs/libxfs/xfs_log_format.h
index 269573c828..3e6682ed65 100644
--- a/fs/xfs/libxfs/xfs_log_format.h
+++ b/fs/xfs/libxfs/xfs_log_format.h
@@ -115,10 +115,13 @@ struct xfs_unmount_log_format {
 #define XLOG_REG_TYPE_BUD_FORMAT	26
 #define XLOG_REG_TYPE_ATTRI_FORMAT	27
 #define XLOG_REG_TYPE_ATTRD_FORMAT	28
-#define XLOG_REG_TYPE_ATTR_NAME	29
+#define XLOG_REG_TYPE_ATTR_NAME		29
 #define XLOG_REG_TYPE_ATTR_VALUE	30
-#define XLOG_REG_TYPE_MAX		30
-
+#define XLOG_REG_TYPE_XMI_FORMAT	31
+#define XLOG_REG_TYPE_XMD_FORMAT	32
+#define XLOG_REG_TYPE_ATTR_NEWNAME	33
+#define XLOG_REG_TYPE_ATTR_NEWVALUE	34
+#define XLOG_REG_TYPE_MAX		34
 
 /*
  * Flags to log operation header
@@ -243,6 +246,8 @@ typedef struct xfs_trans_header {
 #define	XFS_LI_BUD		0x1245
 #define	XFS_LI_ATTRI		0x1246  /* attr set/remove intent*/
 #define	XFS_LI_ATTRD		0x1247  /* attr set/remove done */
+#define	XFS_LI_XMI		0x1248  /* mapping exchange intent */
+#define	XFS_LI_XMD		0x1249  /* mapping exchange done */
 
 #define XFS_LI_TYPE_DESC \
 	{ XFS_LI_EFI,		"XFS_LI_EFI" }, \
@@ -260,7 +265,9 @@ typedef struct xfs_trans_header {
 	{ XFS_LI_BUI,		"XFS_LI_BUI" }, \
 	{ XFS_LI_BUD,		"XFS_LI_BUD" }, \
 	{ XFS_LI_ATTRI,		"XFS_LI_ATTRI" }, \
-	{ XFS_LI_ATTRD,		"XFS_LI_ATTRD" }
+	{ XFS_LI_ATTRD,		"XFS_LI_ATTRD" }, \
+	{ XFS_LI_XMI,		"XFS_LI_XMI" }, \
+	{ XFS_LI_XMD,		"XFS_LI_XMD" }
 
 /*
  * Inode Log Item Format definitions.
@@ -838,10 +845,12 @@ struct xfs_cud_log_format {
 
 #define XFS_BMAP_EXTENT_ATTR_FORK	(1U << 31)
 #define XFS_BMAP_EXTENT_UNWRITTEN	(1U << 30)
+#define XFS_BMAP_EXTENT_REALTIME	(1U << 29)
 
 #define XFS_BMAP_EXTENT_FLAGS		(XFS_BMAP_EXTENT_TYPE_MASK | \
 					 XFS_BMAP_EXTENT_ATTR_FORK | \
-					 XFS_BMAP_EXTENT_UNWRITTEN)
+					 XFS_BMAP_EXTENT_UNWRITTEN | \
+					 XFS_BMAP_EXTENT_REALTIME)
 
 /*
  * This is the structure used to lay out an bui log item in the
@@ -877,6 +886,61 @@ struct xfs_bud_log_format {
 };
 
 /*
+ * XMI/XMD (file mapping exchange) log format definitions
+ */
+
+/* This is the structure used to lay out an mapping exchange log item. */
+struct xfs_xmi_log_format {
+	uint16_t		xmi_type;	/* xmi log item type */
+	uint16_t		xmi_size;	/* size of this item */
+	uint32_t		__pad;		/* must be zero */
+	uint64_t		xmi_id;		/* xmi identifier */
+
+	uint64_t		xmi_inode1;	/* inumber of first file */
+	uint64_t		xmi_inode2;	/* inumber of second file */
+	uint32_t		xmi_igen1;	/* generation of first file */
+	uint32_t		xmi_igen2;	/* generation of second file */
+	uint64_t		xmi_startoff1;	/* block offset into file1 */
+	uint64_t		xmi_startoff2;	/* block offset into file2 */
+	uint64_t		xmi_blockcount;	/* number of blocks */
+	uint64_t		xmi_flags;	/* XFS_EXCHMAPS_* */
+	uint64_t		xmi_isize1;	/* intended file1 size */
+	uint64_t		xmi_isize2;	/* intended file2 size */
+};
+
+/* Exchange mappings between extended attribute forks instead of data forks. */
+#define XFS_EXCHMAPS_ATTR_FORK		(1ULL << 0)
+
+/* Set the file sizes when finished. */
+#define XFS_EXCHMAPS_SET_SIZES		(1ULL << 1)
+
+/*
+ * Exchange the mappings of the two files only if the file allocation units
+ * mapped to file1's range have been written.
+ */
+#define XFS_EXCHMAPS_INO1_WRITTEN	(1ULL << 2)
+
+/* Clear the reflink flag from inode1 after the operation. */
+#define XFS_EXCHMAPS_CLEAR_INO1_REFLINK	(1ULL << 3)
+
+/* Clear the reflink flag from inode2 after the operation. */
+#define XFS_EXCHMAPS_CLEAR_INO2_REFLINK	(1ULL << 4)
+
+#define XFS_EXCHMAPS_LOGGED_FLAGS	(XFS_EXCHMAPS_ATTR_FORK | \
+					 XFS_EXCHMAPS_SET_SIZES | \
+					 XFS_EXCHMAPS_INO1_WRITTEN | \
+					 XFS_EXCHMAPS_CLEAR_INO1_REFLINK | \
+					 XFS_EXCHMAPS_CLEAR_INO2_REFLINK)
+
+/* This is the structure used to lay out an mapping exchange done log item. */
+struct xfs_xmd_log_format {
+	uint16_t		xmd_type;	/* xmd log item type */
+	uint16_t		xmd_size;	/* size of this item */
+	uint32_t		__pad;
+	uint64_t		xmd_xmi_id;	/* id of corresponding xmi */
+};
+
+/*
  * Dquot Log format definitions.
  *
  * The first two fields must be the type and size fitting into
@@ -964,6 +1028,9 @@ struct xfs_icreate_log {
 #define XFS_ATTRI_OP_FLAGS_SET		1	/* Set the attribute */
 #define XFS_ATTRI_OP_FLAGS_REMOVE	2	/* Remove the attribute */
 #define XFS_ATTRI_OP_FLAGS_REPLACE	3	/* Replace the attribute */
+#define XFS_ATTRI_OP_FLAGS_PPTR_SET	4	/* Set parent pointer */
+#define XFS_ATTRI_OP_FLAGS_PPTR_REMOVE	5	/* Remove parent pointer */
+#define XFS_ATTRI_OP_FLAGS_PPTR_REPLACE	6	/* Replace parent pointer */
 #define XFS_ATTRI_OP_FLAGS_TYPE_MASK	0xFF	/* Flags type mask */
 
 /*
@@ -972,6 +1039,7 @@ struct xfs_icreate_log {
  */
 #define XFS_ATTRI_FILTER_MASK		(XFS_ATTR_ROOT | \
 					 XFS_ATTR_SECURE | \
+					 XFS_ATTR_PARENT | \
 					 XFS_ATTR_INCOMPLETE)
 
 /*
@@ -981,11 +1049,22 @@ struct xfs_icreate_log {
 struct xfs_attri_log_format {
 	uint16_t	alfi_type;	/* attri log item type */
 	uint16_t	alfi_size;	/* size of this item */
-	uint32_t	__pad;		/* pad to 64 bit aligned */
+	uint32_t	alfi_igen;	/* generation of alfi_ino for pptr ops */
 	uint64_t	alfi_id;	/* attri identifier */
 	uint64_t	alfi_ino;	/* the inode for this attr operation */
 	uint32_t	alfi_op_flags;	/* marks the op as a set or remove */
-	uint32_t	alfi_name_len;	/* attr name length */
+	union {
+		uint32_t	alfi_name_len;	/* attr name length */
+		struct {
+			/*
+			 * For PPTR_REPLACE, these are the lengths of the old
+			 * and new attr names.  The new and old values must
+			 * have the same length.
+			 */
+			uint16_t	alfi_old_name_len;
+			uint16_t	alfi_new_name_len;
+		};
+	};
 	uint32_t	alfi_value_len;	/* attr value length */
 	uint32_t	alfi_attr_filter;/* attr filter flags */
 };
diff --git a/fs/xfs/libxfs/xfs_log_recover.h b/fs/xfs/libxfs/xfs_log_recover.h
index 9fe7a9564b..521d327e4c 100644
--- a/fs/xfs/libxfs/xfs_log_recover.h
+++ b/fs/xfs/libxfs/xfs_log_recover.h
@@ -75,6 +75,8 @@ extern const struct xlog_recover_item_ops xlog_cui_item_ops;
 extern const struct xlog_recover_item_ops xlog_cud_item_ops;
 extern const struct xlog_recover_item_ops xlog_attri_item_ops;
 extern const struct xlog_recover_item_ops xlog_attrd_item_ops;
+extern const struct xlog_recover_item_ops xlog_xmi_item_ops;
+extern const struct xlog_recover_item_ops xlog_xmd_item_ops;
 
 /*
  * Macros, structures, prototypes for internal log manager use.
@@ -121,6 +123,8 @@ bool xlog_is_buffer_cancelled(struct xlog *log, xfs_daddr_t blkno, uint len);
 
 int xlog_recover_iget(struct xfs_mount *mp, xfs_ino_t ino,
 		struct xfs_inode **ipp);
+int xlog_recover_iget_handle(struct xfs_mount *mp, xfs_ino_t ino, uint32_t gen,
+		struct xfs_inode **ipp);
 void xlog_recover_release_intent(struct xlog *log, unsigned short intent_type,
 		uint64_t intent_id);
 int xlog_alloc_buf_cancel_table(struct xlog *log);
diff --git a/fs/xfs/libxfs/xfs_log_rlimit.c b/fs/xfs/libxfs/xfs_log_rlimit.c
index 9975b93a74..d3bd6a86c8 100644
--- a/fs/xfs/libxfs/xfs_log_rlimit.c
+++ b/fs/xfs/libxfs/xfs_log_rlimit.c
@@ -17,6 +17,34 @@
 #include "xfs_trace.h"
 
 /*
+ * Shortly after enabling the large extents count feature in 2023, longstanding
+ * bugs were found in the code that computes the minimum log size.  Luckily,
+ * the bugs resulted in over-estimates of that size, so there's no impact to
+ * existing users.  However, we don't want to reduce the minimum log size
+ * because that can create the situation where a newer mkfs writes a new
+ * filesystem that an older kernel won't mount.
+ *
+ * Several years prior, we also discovered that the transaction reservations
+ * for rmap and reflink operations were unnecessarily large.  That was fixed,
+ * but the minimum log size computation was left alone to avoid the
+ * compatibility problems noted above.  Fix that too.
+ *
+ * Therefore, we only may correct the computation starting with filesystem
+ * features that didn't exist in 2023.  In other words, only turn this on if
+ * the filesystem has parent pointers.
+ *
+ * This function can be called before the XFS_HAS_* flags have been set up,
+ * (e.g. mkfs) so we must check the ondisk superblock.
+ */
+static inline bool
+xfs_want_minlogsize_fixes(
+	struct xfs_sb	*sb)
+{
+	return xfs_sb_is_v5(sb) &&
+	       xfs_sb_has_incompat_feature(sb, XFS_SB_FEAT_INCOMPAT_PARENT);
+}
+
+/*
  * Calculate the maximum length in bytes that would be required for a local
  * attribute value as large attributes out of line are not logged.
  */
@@ -31,6 +59,15 @@ xfs_log_calc_max_attrsetm_res(
 	       MAXNAMELEN - 1;
 	nblks = XFS_DAENTER_SPACE_RES(mp, XFS_ATTR_FORK);
 	nblks += XFS_B_TO_FSB(mp, size);
+
+	/*
+	 * If the feature set is new enough, correct a unit conversion error in
+	 * the xattr transaction reservation code that resulted in oversized
+	 * minimum log size computations.
+	 */
+	if (xfs_want_minlogsize_fixes(&mp->m_sb))
+		size = XFS_B_TO_FSB(mp, size);
+
 	nblks += XFS_NEXTENTADD_SPACE_RES(mp, size, XFS_ATTR_FORK);
 
 	return  M_RES(mp)->tr_attrsetm.tr_logres +
@@ -49,6 +86,15 @@ xfs_log_calc_trans_resv_for_minlogblocks(
 	unsigned int		rmap_maxlevels = mp->m_rmap_maxlevels;
 
 	/*
+	 * If the feature set is new enough, drop the oversized minimum log
+	 * size computation introduced by the original reflink code.
+	 */
+	if (xfs_want_minlogsize_fixes(&mp->m_sb)) {
+		xfs_trans_resv_calc(mp, resv);
+		return;
+	}
+
+	/*
 	 * In the early days of rmap+reflink, we always set the rmap maxlevels
 	 * to 9 even if the AG was small enough that it would never grow to
 	 * that height.  Transaction reservation sizes influence the minimum
diff --git a/fs/xfs/libxfs/xfs_ondisk.h b/fs/xfs/libxfs/xfs_ondisk.h
index 81885a6a02..e8cdd77d03 100644
--- a/fs/xfs/libxfs/xfs_ondisk.h
+++ b/fs/xfs/libxfs/xfs_ondisk.h
@@ -119,6 +119,7 @@ xfs_check_ondisk_structs(void)
 	XFS_CHECK_OFFSET(xfs_dir2_sf_entry_t, offset,		1);
 	XFS_CHECK_OFFSET(xfs_dir2_sf_entry_t, name,		3);
 	XFS_CHECK_STRUCT_SIZE(xfs_dir2_sf_hdr_t,		10);
+	XFS_CHECK_STRUCT_SIZE(struct xfs_parent_rec,		12);
 
 	/* log structures */
 	XFS_CHECK_STRUCT_SIZE(struct xfs_buf_log_format,	88);
@@ -155,6 +156,11 @@ xfs_check_ondisk_structs(void)
 	XFS_CHECK_OFFSET(struct xfs_efi_log_format_32, efi_extents,	16);
 	XFS_CHECK_OFFSET(struct xfs_efi_log_format_64, efi_extents,	16);
 
+	/* parent pointer ioctls */
+	XFS_CHECK_STRUCT_SIZE(struct xfs_getparents_rec,	32);
+	XFS_CHECK_STRUCT_SIZE(struct xfs_getparents,		40);
+	XFS_CHECK_STRUCT_SIZE(struct xfs_getparents_by_handle,	64);
+
 	/*
 	 * The v5 superblock format extended several v4 header structures with
 	 * additional data. While new fields are only accessible on v5
diff --git a/fs/xfs/libxfs/xfs_parent.c b/fs/xfs/libxfs/xfs_parent.c
new file mode 100644
index 0000000000..69366c44a7
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_parent.c
@@ -0,0 +1,379 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2022-2024 Oracle.
+ * All rights reserved.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_format.h"
+#include "xfs_da_format.h"
+#include "xfs_log_format.h"
+#include "xfs_shared.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_inode.h"
+#include "xfs_error.h"
+#include "xfs_trace.h"
+#include "xfs_trans.h"
+#include "xfs_da_btree.h"
+#include "xfs_attr.h"
+#include "xfs_dir2.h"
+#include "xfs_dir2_priv.h"
+#include "xfs_attr_sf.h"
+#include "xfs_bmap.h"
+#include "xfs_defer.h"
+#include "xfs_log.h"
+#include "xfs_xattr.h"
+#include "xfs_parent.h"
+#include "xfs_trans_space.h"
+#include "xfs_attr_item.h"
+#include "xfs_health.h"
+
+struct kmem_cache		*xfs_parent_args_cache;
+
+/*
+ * Parent pointer attribute handling.
+ *
+ * Because the attribute name is a filename component, it will never be longer
+ * than 255 bytes and must not contain nulls or slashes.  These are roughly the
+ * same constraints that apply to attribute names.
+ *
+ * The attribute value must always be a struct xfs_parent_rec.  This means the
+ * attribute will never be in remote format because 12 bytes is nowhere near
+ * xfs_attr_leaf_entsize_local_max() (~75% of block size).
+ *
+ * Creating a new parent attribute will always create a new attribute - there
+ * should never, ever be an existing attribute in the tree for a new inode.
+ * ENOSPC behavior is problematic - creating the inode without the parent
+ * pointer is effectively a corruption, so we allow parent attribute creation
+ * to dip into the reserve block pool to avoid unexpected ENOSPC errors from
+ * occurring.
+ */
+
+/* Return true if parent pointer attr name is valid. */
+bool
+xfs_parent_namecheck(
+	unsigned int			attr_flags,
+	const void			*name,
+	size_t				length)
+{
+	/*
+	 * Parent pointers always use logged operations, so there should never
+	 * be incomplete xattrs.
+	 */
+	if (attr_flags & XFS_ATTR_INCOMPLETE)
+		return false;
+
+	return xfs_dir2_namecheck(name, length);
+}
+
+/* Return true if parent pointer attr value is valid. */
+bool
+xfs_parent_valuecheck(
+	struct xfs_mount		*mp,
+	const void			*value,
+	size_t				valuelen)
+{
+	const struct xfs_parent_rec	*rec = value;
+
+	if (!xfs_has_parent(mp))
+		return false;
+
+	/* The xattr value must be a parent record. */
+	if (valuelen != sizeof(struct xfs_parent_rec))
+		return false;
+
+	/* The parent record must be local. */
+	if (value == NULL)
+		return false;
+
+	/* The parent inumber must be valid. */
+	if (!xfs_verify_dir_ino(mp, be64_to_cpu(rec->p_ino)))
+		return false;
+
+	return true;
+}
+
+/* Compute the attribute name hash for a parent pointer. */
+xfs_dahash_t
+xfs_parent_hashval(
+	struct xfs_mount		*mp,
+	const uint8_t			*name,
+	int				namelen,
+	xfs_ino_t			parent_ino)
+{
+	struct xfs_name			xname = {
+		.name			= name,
+		.len			= namelen,
+	};
+
+	/*
+	 * Use the same dirent name hash as would be used on the directory, but
+	 * mix in the parent inode number to avoid collisions on hardlinked
+	 * files with identical names but different parents.
+	 */
+	return xfs_dir2_hashname(mp, &xname) ^
+		upper_32_bits(parent_ino) ^ lower_32_bits(parent_ino);
+}
+
+/* Compute the attribute name hash from the xattr components. */
+xfs_dahash_t
+xfs_parent_hashattr(
+	struct xfs_mount		*mp,
+	const uint8_t			*name,
+	int				namelen,
+	const void			*value,
+	int				valuelen)
+{
+	const struct xfs_parent_rec	*rec = value;
+
+	/* Requires a local attr value in xfs_parent_rec format */
+	if (valuelen != sizeof(struct xfs_parent_rec)) {
+		ASSERT(valuelen == sizeof(struct xfs_parent_rec));
+		return 0;
+	}
+
+	if (!value) {
+		ASSERT(value != NULL);
+		return 0;
+	}
+
+	return xfs_parent_hashval(mp, name, namelen, be64_to_cpu(rec->p_ino));
+}
+
+/*
+ * Initialize the parent pointer arguments structure.  Caller must have zeroed
+ * the contents of @args.  @tp is only required for updates.
+ */
+static void
+xfs_parent_da_args_init(
+	struct xfs_da_args	*args,
+	struct xfs_trans	*tp,
+	struct xfs_parent_rec	*rec,
+	struct xfs_inode	*child,
+	xfs_ino_t		owner,
+	const struct xfs_name	*parent_name)
+{
+	args->geo = child->i_mount->m_attr_geo;
+	args->whichfork = XFS_ATTR_FORK;
+	args->attr_filter = XFS_ATTR_PARENT;
+	args->op_flags = XFS_DA_OP_LOGGED | XFS_DA_OP_OKNOENT;
+	args->trans = tp;
+	args->dp = child;
+	args->owner = owner;
+	args->name = parent_name->name;
+	args->namelen = parent_name->len;
+	args->value = rec;
+	args->valuelen = sizeof(struct xfs_parent_rec);
+	xfs_attr_sethash(args);
+}
+
+/* Make sure the incore state is ready for a parent pointer query/update. */
+static inline int
+xfs_parent_iread_extents(
+	struct xfs_trans	*tp,
+	struct xfs_inode	*child)
+{
+	/* Parent pointers require that the attr fork must exist. */
+	if (XFS_IS_CORRUPT(child->i_mount, !xfs_inode_has_attr_fork(child))) {
+		xfs_inode_mark_sick(child, XFS_SICK_INO_PARENT);
+		return -EFSCORRUPTED;
+	}
+
+	return xfs_iread_extents(tp, child, XFS_ATTR_FORK);
+}
+
+/* Add a parent pointer to reflect a dirent addition. */
+int
+xfs_parent_addname(
+	struct xfs_trans	*tp,
+	struct xfs_parent_args	*ppargs,
+	struct xfs_inode	*dp,
+	const struct xfs_name	*parent_name,
+	struct xfs_inode	*child)
+{
+	int			error;
+
+	error = xfs_parent_iread_extents(tp, child);
+	if (error)
+		return error;
+
+	xfs_inode_to_parent_rec(&ppargs->rec, dp);
+	xfs_parent_da_args_init(&ppargs->args, tp, &ppargs->rec, child,
+			child->i_ino, parent_name);
+	xfs_attr_defer_add(&ppargs->args, XFS_ATTR_DEFER_SET);
+	return 0;
+}
+
+/* Remove a parent pointer to reflect a dirent removal. */
+int
+xfs_parent_removename(
+	struct xfs_trans	*tp,
+	struct xfs_parent_args	*ppargs,
+	struct xfs_inode	*dp,
+	const struct xfs_name	*parent_name,
+	struct xfs_inode	*child)
+{
+	int			error;
+
+	error = xfs_parent_iread_extents(tp, child);
+	if (error)
+		return error;
+
+	xfs_inode_to_parent_rec(&ppargs->rec, dp);
+	xfs_parent_da_args_init(&ppargs->args, tp, &ppargs->rec, child,
+			child->i_ino, parent_name);
+	xfs_attr_defer_add(&ppargs->args, XFS_ATTR_DEFER_REMOVE);
+	return 0;
+}
+
+/* Replace one parent pointer with another to reflect a rename. */
+int
+xfs_parent_replacename(
+	struct xfs_trans	*tp,
+	struct xfs_parent_args	*ppargs,
+	struct xfs_inode	*old_dp,
+	const struct xfs_name	*old_name,
+	struct xfs_inode	*new_dp,
+	const struct xfs_name	*new_name,
+	struct xfs_inode	*child)
+{
+	int			error;
+
+	error = xfs_parent_iread_extents(tp, child);
+	if (error)
+		return error;
+
+	xfs_inode_to_parent_rec(&ppargs->rec, old_dp);
+	xfs_parent_da_args_init(&ppargs->args, tp, &ppargs->rec, child,
+			child->i_ino, old_name);
+
+	xfs_inode_to_parent_rec(&ppargs->new_rec, new_dp);
+	ppargs->args.new_name = new_name->name;
+	ppargs->args.new_namelen = new_name->len;
+	ppargs->args.new_value = &ppargs->new_rec;
+	ppargs->args.new_valuelen = sizeof(struct xfs_parent_rec);
+	xfs_attr_defer_add(&ppargs->args, XFS_ATTR_DEFER_REPLACE);
+	return 0;
+}
+
+/*
+ * Extract parent pointer information from any parent pointer xattr into
+ * @parent_ino/gen.  The last two parameters can be NULL pointers.
+ *
+ * Returns 0 if this is not a parent pointer xattr at all; or -EFSCORRUPTED for
+ * garbage.
+ */
+int
+xfs_parent_from_attr(
+	struct xfs_mount	*mp,
+	unsigned int		attr_flags,
+	const unsigned char	*name,
+	unsigned int		namelen,
+	const void		*value,
+	unsigned int		valuelen,
+	xfs_ino_t		*parent_ino,
+	uint32_t		*parent_gen)
+{
+	const struct xfs_parent_rec	*rec = value;
+
+	ASSERT(attr_flags & XFS_ATTR_PARENT);
+
+	if (!xfs_parent_namecheck(attr_flags, name, namelen))
+		return -EFSCORRUPTED;
+	if (!xfs_parent_valuecheck(mp, value, valuelen))
+		return -EFSCORRUPTED;
+
+	if (parent_ino)
+		*parent_ino = be64_to_cpu(rec->p_ino);
+	if (parent_gen)
+		*parent_gen = be32_to_cpu(rec->p_gen);
+	return 0;
+}
+
+/*
+ * Look up a parent pointer record (@parent_name -> @pptr) of @ip.
+ *
+ * Caller must hold at least ILOCK_SHARED.  The scratchpad need not be
+ * initialized.
+ *
+ * Returns 0 if the pointer is found, -ENOATTR if there is no match, or a
+ * negative errno.
+ */
+int
+xfs_parent_lookup(
+	struct xfs_trans		*tp,
+	struct xfs_inode		*ip,
+	const struct xfs_name		*parent_name,
+	struct xfs_parent_rec		*pptr,
+	struct xfs_da_args		*scratch)
+{
+	memset(scratch, 0, sizeof(struct xfs_da_args));
+	xfs_parent_da_args_init(scratch, tp, pptr, ip, ip->i_ino, parent_name);
+	return xfs_attr_get_ilocked(scratch);
+}
+
+/* Sanity-check a parent pointer before we try to perform repairs. */
+static inline bool
+xfs_parent_sanity_check(
+	struct xfs_mount		*mp,
+	const struct xfs_name		*parent_name,
+	const struct xfs_parent_rec	*pptr)
+{
+	if (!xfs_parent_namecheck(XFS_ATTR_PARENT, parent_name->name,
+				parent_name->len))
+		return false;
+
+	if (!xfs_parent_valuecheck(mp, pptr, sizeof(*pptr)))
+		return false;
+
+	return true;
+}
+
+
+/*
+ * Attach the parent pointer (@parent_name -> @pptr) to @ip immediately.
+ * Caller must not have a transaction or hold the ILOCK.  This is for
+ * specialized repair functions only.  The scratchpad need not be initialized.
+ */
+int
+xfs_parent_set(
+	struct xfs_inode	*ip,
+	xfs_ino_t		owner,
+	const struct xfs_name	*parent_name,
+	struct xfs_parent_rec	*pptr,
+	struct xfs_da_args	*scratch)
+{
+	if (!xfs_parent_sanity_check(ip->i_mount, parent_name, pptr)) {
+		ASSERT(0);
+		return -EFSCORRUPTED;
+	}
+
+	memset(scratch, 0, sizeof(struct xfs_da_args));
+	xfs_parent_da_args_init(scratch, NULL, pptr, ip, owner, parent_name);
+	return xfs_attr_set(scratch, XFS_ATTRUPDATE_CREATE, false);
+}
+
+/*
+ * Remove the parent pointer (@parent_name -> @pptr) from @ip immediately.
+ * Caller must not have a transaction or hold the ILOCK.  This is for
+ * specialized repair functions only.  The scratchpad need not be initialized.
+ */
+int
+xfs_parent_unset(
+	struct xfs_inode		*ip,
+	xfs_ino_t			owner,
+	const struct xfs_name		*parent_name,
+	struct xfs_parent_rec		*pptr,
+	struct xfs_da_args		*scratch)
+{
+	if (!xfs_parent_sanity_check(ip->i_mount, parent_name, pptr)) {
+		ASSERT(0);
+		return -EFSCORRUPTED;
+	}
+
+	memset(scratch, 0, sizeof(struct xfs_da_args));
+	xfs_parent_da_args_init(scratch, NULL, pptr, ip, owner, parent_name);
+	return xfs_attr_set(scratch, XFS_ATTRUPDATE_REMOVE, false);
+}
diff --git a/fs/xfs/libxfs/xfs_parent.h b/fs/xfs/libxfs/xfs_parent.h
new file mode 100644
index 0000000000..b8036527cd
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_parent.h
@@ -0,0 +1,110 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2022-2024 Oracle.
+ * All Rights Reserved.
+ */
+#ifndef	__XFS_PARENT_H__
+#define	__XFS_PARENT_H__
+
+/* Metadata validators */
+bool xfs_parent_namecheck(unsigned int attr_flags, const void *name,
+		size_t length);
+bool xfs_parent_valuecheck(struct xfs_mount *mp, const void *value,
+		size_t valuelen);
+
+xfs_dahash_t xfs_parent_hashval(struct xfs_mount *mp, const uint8_t *name,
+		int namelen, xfs_ino_t parent_ino);
+xfs_dahash_t xfs_parent_hashattr(struct xfs_mount *mp, const uint8_t *name,
+		int namelen, const void *value, int valuelen);
+
+/* Initializes a xfs_parent_rec to be stored as an attribute name. */
+static inline void
+xfs_parent_rec_init(
+	struct xfs_parent_rec	*rec,
+	xfs_ino_t		ino,
+	uint32_t		gen)
+{
+	rec->p_ino = cpu_to_be64(ino);
+	rec->p_gen = cpu_to_be32(gen);
+}
+
+/* Initializes a xfs_parent_rec to be stored as an attribute name. */
+static inline void
+xfs_inode_to_parent_rec(
+	struct xfs_parent_rec	*rec,
+	const struct xfs_inode	*dp)
+{
+	xfs_parent_rec_init(rec, dp->i_ino, VFS_IC(dp)->i_generation);
+}
+
+extern struct kmem_cache	*xfs_parent_args_cache;
+
+/*
+ * Parent pointer information needed to pass around the deferred xattr update
+ * machinery.
+ */
+struct xfs_parent_args {
+	struct xfs_parent_rec	rec;
+	struct xfs_parent_rec	new_rec;
+	struct xfs_da_args	args;
+};
+
+/*
+ * Start a parent pointer update by allocating the context object we need to
+ * perform a parent pointer update.
+ */
+static inline int
+xfs_parent_start(
+	struct xfs_mount	*mp,
+	struct xfs_parent_args	**ppargsp)
+{
+	if (!xfs_has_parent(mp)) {
+		*ppargsp = NULL;
+		return 0;
+	}
+
+	*ppargsp = kmem_cache_zalloc(xfs_parent_args_cache, GFP_KERNEL);
+	if (!*ppargsp)
+		return -ENOMEM;
+	return 0;
+}
+
+/* Finish a parent pointer update by freeing the context object. */
+static inline void
+xfs_parent_finish(
+	struct xfs_mount	*mp,
+	struct xfs_parent_args	*ppargs)
+{
+	if (ppargs)
+		kmem_cache_free(xfs_parent_args_cache, ppargs);
+}
+
+int xfs_parent_addname(struct xfs_trans *tp, struct xfs_parent_args *ppargs,
+		struct xfs_inode *dp, const struct xfs_name *parent_name,
+		struct xfs_inode *child);
+int xfs_parent_removename(struct xfs_trans *tp, struct xfs_parent_args *ppargs,
+		struct xfs_inode *dp, const struct xfs_name *parent_name,
+		struct xfs_inode *child);
+int xfs_parent_replacename(struct xfs_trans *tp,
+		struct xfs_parent_args *ppargs,
+		struct xfs_inode *old_dp, const struct xfs_name *old_name,
+		struct xfs_inode *new_dp, const struct xfs_name *new_name,
+		struct xfs_inode *child);
+
+int xfs_parent_from_attr(struct xfs_mount *mp, unsigned int attr_flags,
+		const unsigned char *name, unsigned int namelen,
+		const void *value, unsigned int valuelen,
+		xfs_ino_t *parent_ino, uint32_t *parent_gen);
+
+/* Repair functions */
+int xfs_parent_lookup(struct xfs_trans *tp, struct xfs_inode *ip,
+		const struct xfs_name *name, struct xfs_parent_rec *pptr,
+		struct xfs_da_args *scratch);
+int xfs_parent_set(struct xfs_inode *ip, xfs_ino_t owner,
+		const struct xfs_name *name, struct xfs_parent_rec *pptr,
+		struct xfs_da_args *scratch);
+int xfs_parent_unset(struct xfs_inode *ip, xfs_ino_t owner,
+		const struct xfs_name *name, struct xfs_parent_rec *pptr,
+		struct xfs_da_args *scratch);
+
+#endif /* __XFS_PARENT_H__ */
diff --git a/fs/xfs/libxfs/xfs_refcount.c b/fs/xfs/libxfs/xfs_refcount.c
index 6709a7f8ba..511c912d51 100644
--- a/fs/xfs/libxfs/xfs_refcount.c
+++ b/fs/xfs/libxfs/xfs_refcount.c
@@ -23,6 +23,7 @@
 #include "xfs_refcount.h"
 #include "xfs_rmap.h"
 #include "xfs_ag.h"
+#include "xfs_health.h"
 
 struct kmem_cache	*xfs_refcount_intent_cache;
 
@@ -156,6 +157,7 @@ xfs_refcount_complain_bad_rec(
 	xfs_warn(mp,
 		"Start block 0x%x, block count 0x%x, references 0x%x",
 		irec->rc_startblock, irec->rc_blockcount, irec->rc_refcount);
+	xfs_btree_mark_sick(cur);
 	return -EFSCORRUPTED;
 }
 
@@ -238,6 +240,7 @@ xfs_refcount_insert(
 	if (error)
 		goto out_error;
 	if (XFS_IS_CORRUPT(cur->bc_mp, *i != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
@@ -268,12 +271,14 @@ xfs_refcount_delete(
 	if (error)
 		goto out_error;
 	if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
 	trace_xfs_refcount_delete(cur->bc_mp, cur->bc_ag.pag->pag_agno, &irec);
 	error = xfs_btree_delete(cur, i);
 	if (XFS_IS_CORRUPT(cur->bc_mp, *i != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
@@ -398,6 +403,7 @@ xfs_refcount_split_extent(
 	if (error)
 		goto out_error;
 	if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
@@ -425,6 +431,7 @@ xfs_refcount_split_extent(
 	if (error)
 		goto out_error;
 	if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
@@ -470,6 +477,7 @@ xfs_refcount_merge_center_extents(
 	if (error)
 		goto out_error;
 	if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
@@ -478,6 +486,7 @@ xfs_refcount_merge_center_extents(
 	if (error)
 		goto out_error;
 	if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
@@ -487,6 +496,7 @@ xfs_refcount_merge_center_extents(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -498,6 +508,7 @@ xfs_refcount_merge_center_extents(
 	if (error)
 		goto out_error;
 	if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
@@ -542,6 +553,7 @@ xfs_refcount_merge_left_extent(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -550,6 +562,7 @@ xfs_refcount_merge_left_extent(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -561,6 +574,7 @@ xfs_refcount_merge_left_extent(
 	if (error)
 		goto out_error;
 	if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
@@ -608,6 +622,7 @@ xfs_refcount_merge_right_extent(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -616,6 +631,7 @@ xfs_refcount_merge_right_extent(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -627,6 +643,7 @@ xfs_refcount_merge_right_extent(
 	if (error)
 		goto out_error;
 	if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
@@ -674,6 +691,7 @@ xfs_refcount_find_left_extents(
 	if (error)
 		goto out_error;
 	if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
@@ -693,6 +711,7 @@ xfs_refcount_find_left_extents(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -767,6 +786,7 @@ xfs_refcount_find_right_extents(
 	if (error)
 		goto out_error;
 	if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
@@ -786,6 +806,7 @@ xfs_refcount_find_right_extents(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -1056,7 +1077,7 @@ xfs_refcount_still_have_space(
 	 * to handle each of the shape changes to the refcount btree.
 	 */
 	overhead = xfs_allocfree_block_count(cur->bc_mp,
-				cur->bc_ag.refc.shape_changes);
+				cur->bc_refc.shape_changes);
 	overhead += cur->bc_mp->m_refc_maxlevels;
 	overhead *= cur->bc_mp->m_sb.sb_blocksize;
 
@@ -1064,17 +1085,17 @@ xfs_refcount_still_have_space(
 	 * Only allow 2 refcount extent updates per transaction if the
 	 * refcount continue update "error" has been injected.
 	 */
-	if (cur->bc_ag.refc.nr_ops > 2 &&
+	if (cur->bc_refc.nr_ops > 2 &&
 	    XFS_TEST_ERROR(false, cur->bc_mp,
 			XFS_ERRTAG_REFCOUNT_CONTINUE_UPDATE))
 		return false;
 
-	if (cur->bc_ag.refc.nr_ops == 0)
+	if (cur->bc_refc.nr_ops == 0)
 		return true;
 	else if (overhead > cur->bc_tp->t_log_res)
 		return false;
-	return  cur->bc_tp->t_log_res - overhead >
-		cur->bc_ag.refc.nr_ops * XFS_REFCOUNT_ITEM_OVERHEAD;
+	return cur->bc_tp->t_log_res - overhead >
+		cur->bc_refc.nr_ops * XFS_REFCOUNT_ITEM_OVERHEAD;
 }
 
 /*
@@ -1134,7 +1155,7 @@ xfs_refcount_adjust_extents(
 			 * Either cover the hole (increment) or
 			 * delete the range (decrement).
 			 */
-			cur->bc_ag.refc.nr_ops++;
+			cur->bc_refc.nr_ops++;
 			if (tmp.rc_refcount) {
 				error = xfs_refcount_insert(cur, &tmp,
 						&found_tmp);
@@ -1142,6 +1163,7 @@ xfs_refcount_adjust_extents(
 					goto out_error;
 				if (XFS_IS_CORRUPT(cur->bc_mp,
 						   found_tmp != 1)) {
+					xfs_btree_mark_sick(cur);
 					error = -EFSCORRUPTED;
 					goto out_error;
 				}
@@ -1180,6 +1202,7 @@ xfs_refcount_adjust_extents(
 		 */
 		if (XFS_IS_CORRUPT(cur->bc_mp, ext.rc_blockcount == 0) ||
 		    XFS_IS_CORRUPT(cur->bc_mp, ext.rc_blockcount > *aglen)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -1193,7 +1216,7 @@ xfs_refcount_adjust_extents(
 		ext.rc_refcount += adj;
 		trace_xfs_refcount_modify_extent(cur->bc_mp,
 				cur->bc_ag.pag->pag_agno, &ext);
-		cur->bc_ag.refc.nr_ops++;
+		cur->bc_refc.nr_ops++;
 		if (ext.rc_refcount > 1) {
 			error = xfs_refcount_update(cur, &ext);
 			if (error)
@@ -1203,6 +1226,7 @@ xfs_refcount_adjust_extents(
 			if (error)
 				goto out_error;
 			if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto out_error;
 			}
@@ -1281,7 +1305,7 @@ xfs_refcount_adjust(
 	if (shape_changed)
 		shape_changes++;
 	if (shape_changes)
-		cur->bc_ag.refc.shape_changes++;
+		cur->bc_refc.shape_changes++;
 
 	/* Now that we've taken care of the ends, adjust the middle extents */
 	error = xfs_refcount_adjust_extents(cur, agbno, aglen, adj);
@@ -1327,8 +1351,10 @@ xfs_refcount_continue_op(
 	struct xfs_perag		*pag = cur->bc_ag.pag;
 
 	if (XFS_IS_CORRUPT(mp, !xfs_verify_agbext(pag, new_agbno,
-					ri->ri_blockcount)))
+					ri->ri_blockcount))) {
+		xfs_btree_mark_sick(cur);
 		return -EFSCORRUPTED;
+	}
 
 	ri->ri_startblock = XFS_AGB_TO_FSB(mp, pag->pag_agno, new_agbno);
 
@@ -1374,8 +1400,8 @@ xfs_refcount_finish_one(
 	 */
 	rcur = *pcur;
 	if (rcur != NULL && rcur->bc_ag.pag != ri->ri_pag) {
-		nr_ops = rcur->bc_ag.refc.nr_ops;
-		shape_changes = rcur->bc_ag.refc.shape_changes;
+		nr_ops = rcur->bc_refc.nr_ops;
+		shape_changes = rcur->bc_refc.shape_changes;
 		xfs_refcount_finish_one_cleanup(tp, rcur, 0);
 		rcur = NULL;
 		*pcur = NULL;
@@ -1387,8 +1413,8 @@ xfs_refcount_finish_one(
 			return error;
 
 		rcur = xfs_refcountbt_init_cursor(mp, tp, agbp, ri->ri_pag);
-		rcur->bc_ag.refc.nr_ops = nr_ops;
-		rcur->bc_ag.refc.shape_changes = shape_changes;
+		rcur->bc_refc.nr_ops = nr_ops;
+		rcur->bc_refc.shape_changes = shape_changes;
 	}
 	*pcur = rcur;
 
@@ -1449,7 +1475,7 @@ __xfs_refcount_add(
 			blockcount);
 
 	ri = kmem_cache_alloc(xfs_refcount_intent_cache,
-			GFP_NOFS | __GFP_NOFAIL);
+			GFP_KERNEL | __GFP_NOFAIL);
 	INIT_LIST_HEAD(&ri->ri_list);
 	ri->ri_type = type;
 	ri->ri_startblock = startblock;
@@ -1535,6 +1561,7 @@ xfs_refcount_find_shared(
 	if (error)
 		goto out_error;
 	if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
@@ -1552,6 +1579,7 @@ xfs_refcount_find_shared(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -1585,6 +1613,7 @@ xfs_refcount_find_shared(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -1682,6 +1711,7 @@ xfs_refcount_adjust_cow_extents(
 		goto out_error;
 	if (XFS_IS_CORRUPT(cur->bc_mp, found_rec &&
 				ext.rc_domain != XFS_REFC_DOMAIN_COW)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
@@ -1697,6 +1727,7 @@ xfs_refcount_adjust_cow_extents(
 		/* Adding a CoW reservation, there should be nothing here. */
 		if (XFS_IS_CORRUPT(cur->bc_mp,
 				   agbno + aglen > ext.rc_startblock)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -1714,6 +1745,7 @@ xfs_refcount_adjust_cow_extents(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(cur->bc_mp, found_tmp != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -1721,14 +1753,17 @@ xfs_refcount_adjust_cow_extents(
 	case XFS_REFCOUNT_ADJUST_COW_FREE:
 		/* Removing a CoW reservation, there should be one extent. */
 		if (XFS_IS_CORRUPT(cur->bc_mp, ext.rc_startblock != agbno)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
 		if (XFS_IS_CORRUPT(cur->bc_mp, ext.rc_blockcount != aglen)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
 		if (XFS_IS_CORRUPT(cur->bc_mp, ext.rc_refcount != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -1740,6 +1775,7 @@ xfs_refcount_adjust_cow_extents(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -1889,8 +1925,10 @@ xfs_refcount_recover_extent(
 	struct xfs_refcount_recovery	*rr;
 
 	if (XFS_IS_CORRUPT(cur->bc_mp,
-			   be32_to_cpu(rec->refc.rc_refcount) != 1))
+			   be32_to_cpu(rec->refc.rc_refcount) != 1)) {
+		xfs_btree_mark_sick(cur);
 		return -EFSCORRUPTED;
+	}
 
 	rr = kmalloc(sizeof(struct xfs_refcount_recovery),
 			GFP_KERNEL | __GFP_NOFAIL);
@@ -1900,6 +1938,7 @@ xfs_refcount_recover_extent(
 	if (xfs_refcount_check_irec(cur->bc_ag.pag, &rr->rr_rrec) != NULL ||
 	    XFS_IS_CORRUPT(cur->bc_mp,
 			   rr->rr_rrec.rc_domain != XFS_REFC_DOMAIN_COW)) {
+		xfs_btree_mark_sick(cur);
 		kfree(rr);
 		return -EFSCORRUPTED;
 	}
diff --git a/fs/xfs/libxfs/xfs_refcount_btree.c b/fs/xfs/libxfs/xfs_refcount_btree.c
index 0d80bd9914..ca59f6c89f 100644
--- a/fs/xfs/libxfs/xfs_refcount_btree.c
+++ b/fs/xfs/libxfs/xfs_refcount_btree.c
@@ -16,6 +16,7 @@
 #include "xfs_refcount.h"
 #include "xfs_alloc.h"
 #include "xfs_error.h"
+#include "xfs_health.h"
 #include "xfs_trace.h"
 #include "xfs_trans.h"
 #include "xfs_bit.h"
@@ -77,8 +78,6 @@ xfs_refcountbt_alloc_block(
 					xfs_refc_block(args.mp)));
 	if (error)
 		goto out_error;
-	trace_xfs_refcountbt_alloc_block(cur->bc_mp, cur->bc_ag.pag->pag_agno,
-			args.agbno, 1);
 	if (args.fsbno == NULLFSBLOCK) {
 		*stat = 0;
 		return 0;
@@ -107,8 +106,6 @@ xfs_refcountbt_free_block(
 	struct xfs_agf		*agf = agbp->b_addr;
 	xfs_fsblock_t		fsbno = XFS_DADDR_TO_FSB(mp, xfs_buf_daddr(bp));
 
-	trace_xfs_refcountbt_free_block(cur->bc_mp, cur->bc_ag.pag->pag_agno,
-			XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno), 1);
 	be32_add_cpu(&agf->agf_refcount_blocks, -1);
 	xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_REFCOUNT_BLOCKS);
 	return xfs_free_extent_later(cur->bc_tp, fsbno, 1,
@@ -220,7 +217,7 @@ xfs_refcountbt_verify(
 
 	if (!xfs_has_reflink(mp))
 		return __this_address;
-	fa = xfs_btree_sblock_v5hdr_verify(bp);
+	fa = xfs_btree_agblock_v5hdr_verify(bp);
 	if (fa)
 		return fa;
 
@@ -242,7 +239,7 @@ xfs_refcountbt_verify(
 	} else if (level >= mp->m_refc_maxlevels)
 		return __this_address;
 
-	return xfs_btree_sblock_verify(bp, mp->m_refc_mxr[level != 0]);
+	return xfs_btree_agblock_verify(bp, mp->m_refc_mxr[level != 0]);
 }
 
 STATIC void
@@ -251,7 +248,7 @@ xfs_refcountbt_read_verify(
 {
 	xfs_failaddr_t	fa;
 
-	if (!xfs_btree_sblock_verify_crc(bp))
+	if (!xfs_btree_agblock_verify_crc(bp))
 		xfs_verifier_error(bp, -EFSBADCRC, __this_address);
 	else {
 		fa = xfs_refcountbt_verify(bp);
@@ -275,7 +272,7 @@ xfs_refcountbt_write_verify(
 		xfs_verifier_error(bp, -EFSCORRUPTED, fa);
 		return;
 	}
-	xfs_btree_sblock_calc_crc(bp);
+	xfs_btree_agblock_calc_crc(bp);
 
 }
 
@@ -321,9 +318,17 @@ xfs_refcountbt_keys_contiguous(
 				 be32_to_cpu(key2->refc.rc_startblock));
 }
 
-static const struct xfs_btree_ops xfs_refcountbt_ops = {
+const struct xfs_btree_ops xfs_refcountbt_ops = {
+	.name			= "refcount",
+	.type			= XFS_BTREE_TYPE_AG,
+
 	.rec_len		= sizeof(struct xfs_refcount_rec),
 	.key_len		= sizeof(struct xfs_refcount_key),
+	.ptr_len		= XFS_BTREE_SHORT_PTR_LEN,
+
+	.lru_refs		= XFS_REFC_BTREE_REF,
+	.statoff		= XFS_STATS_CALC_INDEX(xs_refcbt_2),
+	.sick_mask		= XFS_SICK_AG_REFCNTBT,
 
 	.dup_cursor		= xfs_refcountbt_dup_cursor,
 	.set_root		= xfs_refcountbt_set_root,
@@ -344,59 +349,32 @@ static const struct xfs_btree_ops xfs_refcountbt_ops = {
 };
 
 /*
- * Initialize a new refcount btree cursor.
+ * Create a new refcount btree cursor.
+ *
+ * For staging cursors tp and agbp are NULL.
  */
-static struct xfs_btree_cur *
-xfs_refcountbt_init_common(
+struct xfs_btree_cur *
+xfs_refcountbt_init_cursor(
 	struct xfs_mount	*mp,
 	struct xfs_trans	*tp,
+	struct xfs_buf		*agbp,
 	struct xfs_perag	*pag)
 {
 	struct xfs_btree_cur	*cur;
 
 	ASSERT(pag->pag_agno < mp->m_sb.sb_agcount);
 
-	cur = xfs_btree_alloc_cursor(mp, tp, XFS_BTNUM_REFC,
+	cur = xfs_btree_alloc_cursor(mp, tp, &xfs_refcountbt_ops,
 			mp->m_refc_maxlevels, xfs_refcountbt_cur_cache);
-	cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_refcbt_2);
-
-	cur->bc_flags |= XFS_BTREE_CRC_BLOCKS;
-
 	cur->bc_ag.pag = xfs_perag_hold(pag);
-	cur->bc_ag.refc.nr_ops = 0;
-	cur->bc_ag.refc.shape_changes = 0;
-	cur->bc_ops = &xfs_refcountbt_ops;
-	return cur;
-}
-
-/* Create a btree cursor. */
-struct xfs_btree_cur *
-xfs_refcountbt_init_cursor(
-	struct xfs_mount	*mp,
-	struct xfs_trans	*tp,
-	struct xfs_buf		*agbp,
-	struct xfs_perag	*pag)
-{
-	struct xfs_agf		*agf = agbp->b_addr;
-	struct xfs_btree_cur	*cur;
-
-	cur = xfs_refcountbt_init_common(mp, tp, pag);
-	cur->bc_nlevels = be32_to_cpu(agf->agf_refcount_level);
+	cur->bc_refc.nr_ops = 0;
+	cur->bc_refc.shape_changes = 0;
 	cur->bc_ag.agbp = agbp;
-	return cur;
-}
+	if (agbp) {
+		struct xfs_agf		*agf = agbp->b_addr;
 
-/* Create a btree cursor with a fake root for staging. */
-struct xfs_btree_cur *
-xfs_refcountbt_stage_cursor(
-	struct xfs_mount	*mp,
-	struct xbtree_afakeroot	*afake,
-	struct xfs_perag	*pag)
-{
-	struct xfs_btree_cur	*cur;
-
-	cur = xfs_refcountbt_init_common(mp, NULL, pag);
-	xfs_btree_stage_afakeroot(cur, afake);
+		cur->bc_nlevels = be32_to_cpu(agf->agf_refcount_level);
+	}
 	return cur;
 }
 
@@ -421,7 +399,7 @@ xfs_refcountbt_commit_staged_btree(
 	xfs_alloc_log_agf(tp, agbp, XFS_AGF_REFCOUNT_BLOCKS |
 				    XFS_AGF_REFCOUNT_ROOT |
 				    XFS_AGF_REFCOUNT_LEVEL);
-	xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_refcountbt_ops);
+	xfs_btree_commit_afakeroot(cur, tp, agbp);
 }
 
 /* Calculate number of records in a refcount btree block. */
diff --git a/fs/xfs/libxfs/xfs_refcount_btree.h b/fs/xfs/libxfs/xfs_refcount_btree.h
index d66b37259b..1e0ab25f6c 100644
--- a/fs/xfs/libxfs/xfs_refcount_btree.h
+++ b/fs/xfs/libxfs/xfs_refcount_btree.h
@@ -48,8 +48,6 @@ struct xbtree_afakeroot;
 extern struct xfs_btree_cur *xfs_refcountbt_init_cursor(struct xfs_mount *mp,
 		struct xfs_trans *tp, struct xfs_buf *agbp,
 		struct xfs_perag *pag);
-struct xfs_btree_cur *xfs_refcountbt_stage_cursor(struct xfs_mount *mp,
-		struct xbtree_afakeroot *afake, struct xfs_perag *pag);
 extern int xfs_refcountbt_maxrecs(int blocklen, bool leaf);
 extern void xfs_refcountbt_compute_maxlevels(struct xfs_mount *mp);
 
diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c
index 76bf7f48cb..ef16f6f9ce 100644
--- a/fs/xfs/libxfs/xfs_rmap.c
+++ b/fs/xfs/libxfs/xfs_rmap.c
@@ -23,6 +23,7 @@
 #include "xfs_error.h"
 #include "xfs_inode.h"
 #include "xfs_ag.h"
+#include "xfs_health.h"
 
 struct kmem_cache	*xfs_rmap_intent_cache;
 
@@ -56,8 +57,10 @@ xfs_rmap_lookup_le(
 	error = xfs_rmap_get_rec(cur, irec, &get_stat);
 	if (error)
 		return error;
-	if (!get_stat)
+	if (!get_stat) {
+		xfs_btree_mark_sick(cur);
 		return -EFSCORRUPTED;
+	}
 
 	return 0;
 }
@@ -132,6 +135,7 @@ xfs_rmap_insert(
 	if (error)
 		goto done;
 	if (XFS_IS_CORRUPT(rcur->bc_mp, i != 0)) {
+		xfs_btree_mark_sick(rcur);
 		error = -EFSCORRUPTED;
 		goto done;
 	}
@@ -145,6 +149,7 @@ xfs_rmap_insert(
 	if (error)
 		goto done;
 	if (XFS_IS_CORRUPT(rcur->bc_mp, i != 1)) {
+		xfs_btree_mark_sick(rcur);
 		error = -EFSCORRUPTED;
 		goto done;
 	}
@@ -174,6 +179,7 @@ xfs_rmap_delete(
 	if (error)
 		goto done;
 	if (XFS_IS_CORRUPT(rcur->bc_mp, i != 1)) {
+		xfs_btree_mark_sick(rcur);
 		error = -EFSCORRUPTED;
 		goto done;
 	}
@@ -182,6 +188,7 @@ xfs_rmap_delete(
 	if (error)
 		goto done;
 	if (XFS_IS_CORRUPT(rcur->bc_mp, i != 1)) {
+		xfs_btree_mark_sick(rcur);
 		error = -EFSCORRUPTED;
 		goto done;
 	}
@@ -208,10 +215,10 @@ xfs_rmap_btrec_to_irec(
 /* Simple checks for rmap records. */
 xfs_failaddr_t
 xfs_rmap_check_irec(
-	struct xfs_btree_cur		*cur,
+	struct xfs_perag		*pag,
 	const struct xfs_rmap_irec	*irec)
 {
-	struct xfs_mount		*mp = cur->bc_mp;
+	struct xfs_mount		*mp = pag->pag_mount;
 	bool				is_inode;
 	bool				is_unwritten;
 	bool				is_bmbt;
@@ -226,8 +233,8 @@ xfs_rmap_check_irec(
 			return __this_address;
 	} else {
 		/* check for valid extent range, including overflow */
-		if (!xfs_verify_agbext(cur->bc_ag.pag, irec->rm_startblock,
-						       irec->rm_blockcount))
+		if (!xfs_verify_agbext(pag, irec->rm_startblock,
+					    irec->rm_blockcount))
 			return __this_address;
 	}
 
@@ -262,6 +269,16 @@ xfs_rmap_check_irec(
 	return NULL;
 }
 
+static inline xfs_failaddr_t
+xfs_rmap_check_btrec(
+	struct xfs_btree_cur		*cur,
+	const struct xfs_rmap_irec	*irec)
+{
+	if (xfs_btree_is_mem_rmap(cur->bc_ops))
+		return xfs_rmap_check_irec(cur->bc_mem.pag, irec);
+	return xfs_rmap_check_irec(cur->bc_ag.pag, irec);
+}
+
 static inline int
 xfs_rmap_complain_bad_rec(
 	struct xfs_btree_cur		*cur,
@@ -270,13 +287,18 @@ xfs_rmap_complain_bad_rec(
 {
 	struct xfs_mount		*mp = cur->bc_mp;
 
-	xfs_warn(mp,
-		"Reverse Mapping BTree record corruption in AG %d detected at %pS!",
-		cur->bc_ag.pag->pag_agno, fa);
+	if (xfs_btree_is_mem_rmap(cur->bc_ops))
+		xfs_warn(mp,
+ "In-Memory Reverse Mapping BTree record corruption detected at %pS!", fa);
+	else
+		xfs_warn(mp,
+ "Reverse Mapping BTree record corruption in AG %d detected at %pS!",
+			cur->bc_ag.pag->pag_agno, fa);
 	xfs_warn(mp,
 		"Owner 0x%llx, flags 0x%x, start block 0x%x block count 0x%x",
 		irec->rm_owner, irec->rm_flags, irec->rm_startblock,
 		irec->rm_blockcount);
+	xfs_btree_mark_sick(cur);
 	return -EFSCORRUPTED;
 }
 
@@ -299,7 +321,7 @@ xfs_rmap_get_rec(
 
 	fa = xfs_rmap_btrec_to_irec(rec, irec);
 	if (!fa)
-		fa = xfs_rmap_check_irec(cur, irec);
+		fa = xfs_rmap_check_btrec(cur, irec);
 	if (fa)
 		return xfs_rmap_complain_bad_rec(cur, fa, irec);
 
@@ -512,7 +534,7 @@ xfs_rmap_lookup_le_range(
  */
 static int
 xfs_rmap_free_check_owner(
-	struct xfs_mount	*mp,
+	struct xfs_btree_cur	*cur,
 	uint64_t		ltoff,
 	struct xfs_rmap_irec	*rec,
 	xfs_filblks_t		len,
@@ -520,6 +542,7 @@ xfs_rmap_free_check_owner(
 	uint64_t		offset,
 	unsigned int		flags)
 {
+	struct xfs_mount	*mp = cur->bc_mp;
 	int			error = 0;
 
 	if (owner == XFS_RMAP_OWN_UNKNOWN)
@@ -529,12 +552,14 @@ xfs_rmap_free_check_owner(
 	if (XFS_IS_CORRUPT(mp,
 			   (flags & XFS_RMAP_UNWRITTEN) !=
 			   (rec->rm_flags & XFS_RMAP_UNWRITTEN))) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out;
 	}
 
 	/* Make sure the owner matches what we expect to find in the tree. */
 	if (XFS_IS_CORRUPT(mp, owner != rec->rm_owner)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out;
 	}
@@ -546,16 +571,19 @@ xfs_rmap_free_check_owner(
 	if (flags & XFS_RMAP_BMBT_BLOCK) {
 		if (XFS_IS_CORRUPT(mp,
 				   !(rec->rm_flags & XFS_RMAP_BMBT_BLOCK))) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out;
 		}
 	} else {
 		if (XFS_IS_CORRUPT(mp, rec->rm_offset > offset)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out;
 		}
 		if (XFS_IS_CORRUPT(mp,
 				   offset + len > ltoff + rec->rm_blockcount)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out;
 		}
@@ -618,6 +646,7 @@ xfs_rmap_unmap(
 	if (error)
 		goto out_error;
 	if (XFS_IS_CORRUPT(mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
@@ -639,6 +668,7 @@ xfs_rmap_unmap(
 		if (XFS_IS_CORRUPT(mp,
 				   bno <
 				   ltrec.rm_startblock + ltrec.rm_blockcount)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -665,6 +695,7 @@ xfs_rmap_unmap(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -677,12 +708,13 @@ xfs_rmap_unmap(
 			   ltrec.rm_startblock > bno ||
 			   ltrec.rm_startblock + ltrec.rm_blockcount <
 			   bno + len)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
 
 	/* Check owner information. */
-	error = xfs_rmap_free_check_owner(mp, ltoff, &ltrec, len, owner,
+	error = xfs_rmap_free_check_owner(cur, ltoff, &ltrec, len, owner,
 			offset, flags);
 	if (error)
 		goto out_error;
@@ -697,6 +729,7 @@ xfs_rmap_unmap(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -788,6 +821,86 @@ out_error:
 	return error;
 }
 
+#ifdef CONFIG_XFS_LIVE_HOOKS
+/*
+ * Use a static key here to reduce the overhead of rmapbt live updates.  If
+ * the compiler supports jump labels, the static branch will be replaced by a
+ * nop sled when there are no hook users.  Online fsck is currently the only
+ * caller, so this is a reasonable tradeoff.
+ *
+ * Note: Patching the kernel code requires taking the cpu hotplug lock.  Other
+ * parts of the kernel allocate memory with that lock held, which means that
+ * XFS callers cannot hold any locks that might be used by memory reclaim or
+ * writeback when calling the static_branch_{inc,dec} functions.
+ */
+DEFINE_STATIC_XFS_HOOK_SWITCH(xfs_rmap_hooks_switch);
+
+void
+xfs_rmap_hook_disable(void)
+{
+	xfs_hooks_switch_off(&xfs_rmap_hooks_switch);
+}
+
+void
+xfs_rmap_hook_enable(void)
+{
+	xfs_hooks_switch_on(&xfs_rmap_hooks_switch);
+}
+
+/* Call downstream hooks for a reverse mapping update. */
+static inline void
+xfs_rmap_update_hook(
+	struct xfs_trans		*tp,
+	struct xfs_perag		*pag,
+	enum xfs_rmap_intent_type	op,
+	xfs_agblock_t			startblock,
+	xfs_extlen_t			blockcount,
+	bool				unwritten,
+	const struct xfs_owner_info	*oinfo)
+{
+	if (xfs_hooks_switched_on(&xfs_rmap_hooks_switch)) {
+		struct xfs_rmap_update_params	p = {
+			.startblock	= startblock,
+			.blockcount	= blockcount,
+			.unwritten	= unwritten,
+			.oinfo		= *oinfo, /* struct copy */
+		};
+
+		if (pag)
+			xfs_hooks_call(&pag->pag_rmap_update_hooks, op, &p);
+	}
+}
+
+/* Call the specified function during a reverse mapping update. */
+int
+xfs_rmap_hook_add(
+	struct xfs_perag	*pag,
+	struct xfs_rmap_hook	*hook)
+{
+	return xfs_hooks_add(&pag->pag_rmap_update_hooks, &hook->rmap_hook);
+}
+
+/* Stop calling the specified function during a reverse mapping update. */
+void
+xfs_rmap_hook_del(
+	struct xfs_perag	*pag,
+	struct xfs_rmap_hook	*hook)
+{
+	xfs_hooks_del(&pag->pag_rmap_update_hooks, &hook->rmap_hook);
+}
+
+/* Configure rmap update hook functions. */
+void
+xfs_rmap_hook_setup(
+	struct xfs_rmap_hook	*hook,
+	notifier_fn_t		mod_fn)
+{
+	xfs_hook_setup(&hook->rmap_hook, mod_fn);
+}
+#else
+# define xfs_rmap_update_hook(t, p, o, s, b, u, oi)	do { } while (0)
+#endif /* CONFIG_XFS_LIVE_HOOKS */
+
 /*
  * Remove a reference to an extent in the rmap btree.
  */
@@ -808,7 +921,7 @@ xfs_rmap_free(
 		return 0;
 
 	cur = xfs_rmapbt_init_cursor(mp, tp, agbp, pag);
-
+	xfs_rmap_update_hook(tp, pag, XFS_RMAP_UNMAP, bno, len, false, oinfo);
 	error = xfs_rmap_unmap(cur, bno, len, false, oinfo);
 
 	xfs_btree_del_cursor(cur, error);
@@ -900,6 +1013,7 @@ xfs_rmap_map(
 	if (XFS_IS_CORRUPT(mp,
 			   have_lt != 0 &&
 			   ltrec.rm_startblock + ltrec.rm_blockcount > bno)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
@@ -917,10 +1031,12 @@ xfs_rmap_map(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(mp, have_gt != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
 		if (XFS_IS_CORRUPT(mp, bno + len > gtrec.rm_startblock)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -974,6 +1090,7 @@ xfs_rmap_map(
 			if (error)
 				goto out_error;
 			if (XFS_IS_CORRUPT(mp, i != 1)) {
+				xfs_btree_mark_sick(cur);
 				error = -EFSCORRUPTED;
 				goto out_error;
 			}
@@ -1021,6 +1138,7 @@ xfs_rmap_map(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -1055,6 +1173,7 @@ xfs_rmap_alloc(
 		return 0;
 
 	cur = xfs_rmapbt_init_cursor(mp, tp, agbp, pag);
+	xfs_rmap_update_hook(tp, pag, XFS_RMAP_MAP, bno, len, false, oinfo);
 	error = xfs_rmap_map(cur, bno, len, false, oinfo);
 
 	xfs_btree_del_cursor(cur, error);
@@ -1116,6 +1235,7 @@ xfs_rmap_convert(
 	if (error)
 		goto done;
 	if (XFS_IS_CORRUPT(mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto done;
 	}
@@ -1153,12 +1273,14 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
 		if (XFS_IS_CORRUPT(mp,
 				   LEFT.rm_startblock + LEFT.rm_blockcount >
 				   bno)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1181,6 +1303,7 @@ xfs_rmap_convert(
 	if (error)
 		goto done;
 	if (XFS_IS_CORRUPT(mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto done;
 	}
@@ -1193,10 +1316,12 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
 		if (XFS_IS_CORRUPT(mp, bno + len > RIGHT.rm_startblock)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1227,6 +1352,7 @@ xfs_rmap_convert(
 	if (error)
 		goto done;
 	if (XFS_IS_CORRUPT(mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto done;
 	}
@@ -1246,6 +1372,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1257,6 +1384,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1264,6 +1392,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1275,6 +1404,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1282,6 +1412,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1305,6 +1436,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1312,6 +1444,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1331,6 +1464,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1342,6 +1476,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1349,6 +1484,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1419,6 +1555,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1461,6 +1598,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 0)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1476,6 +1614,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1509,6 +1648,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1522,6 +1662,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 0)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1534,6 +1675,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1606,6 +1748,7 @@ xfs_rmap_convert_shared(
 	if (error)
 		goto done;
 	if (XFS_IS_CORRUPT(mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto done;
 	}
@@ -1634,6 +1777,7 @@ xfs_rmap_convert_shared(
 		if (XFS_IS_CORRUPT(mp,
 				   LEFT.rm_startblock + LEFT.rm_blockcount >
 				   bno)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1652,10 +1796,12 @@ xfs_rmap_convert_shared(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
 		if (XFS_IS_CORRUPT(mp, bno + len > RIGHT.rm_startblock)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1706,6 +1852,7 @@ xfs_rmap_convert_shared(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1732,6 +1879,7 @@ xfs_rmap_convert_shared(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1758,6 +1906,7 @@ xfs_rmap_convert_shared(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1781,6 +1930,7 @@ xfs_rmap_convert_shared(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1816,6 +1966,7 @@ xfs_rmap_convert_shared(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1861,6 +2012,7 @@ xfs_rmap_convert_shared(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1896,6 +2048,7 @@ xfs_rmap_convert_shared(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -1934,6 +2087,7 @@ xfs_rmap_convert_shared(
 		if (error)
 			goto done;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto done;
 		}
@@ -2023,6 +2177,7 @@ xfs_rmap_unmap_shared(
 	if (error)
 		goto out_error;
 	if (XFS_IS_CORRUPT(mp, i != 1)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
@@ -2033,12 +2188,14 @@ xfs_rmap_unmap_shared(
 			   ltrec.rm_startblock > bno ||
 			   ltrec.rm_startblock + ltrec.rm_blockcount <
 			   bno + len)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
 
 	/* Make sure the owner matches what we expect to find in the tree. */
 	if (XFS_IS_CORRUPT(mp, owner != ltrec.rm_owner)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
@@ -2047,16 +2204,19 @@ xfs_rmap_unmap_shared(
 	if (XFS_IS_CORRUPT(mp,
 			   (flags & XFS_RMAP_UNWRITTEN) !=
 			   (ltrec.rm_flags & XFS_RMAP_UNWRITTEN))) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
 
 	/* Check the offset. */
 	if (XFS_IS_CORRUPT(mp, ltrec.rm_offset > offset)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
 	if (XFS_IS_CORRUPT(mp, offset > ltoff + ltrec.rm_blockcount)) {
+		xfs_btree_mark_sick(cur);
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
@@ -2113,6 +2273,7 @@ xfs_rmap_unmap_shared(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -2142,6 +2303,7 @@ xfs_rmap_unmap_shared(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -2221,6 +2383,7 @@ xfs_rmap_map_shared(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(mp, have_gt != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -2273,6 +2436,7 @@ xfs_rmap_map_shared(
 		if (error)
 			goto out_error;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out_error;
 		}
@@ -2335,15 +2499,12 @@ xfs_rmap_map_raw(
 {
 	struct xfs_owner_info	oinfo;
 
-	oinfo.oi_owner = rmap->rm_owner;
-	oinfo.oi_offset = rmap->rm_offset;
-	oinfo.oi_flags = 0;
-	if (rmap->rm_flags & XFS_RMAP_ATTR_FORK)
-		oinfo.oi_flags |= XFS_OWNER_INFO_ATTR_FORK;
-	if (rmap->rm_flags & XFS_RMAP_BMBT_BLOCK)
-		oinfo.oi_flags |= XFS_OWNER_INFO_BMBT_BLOCK;
+	xfs_owner_info_pack(&oinfo, rmap->rm_owner, rmap->rm_offset,
+			rmap->rm_flags);
 
-	if (rmap->rm_flags || XFS_RMAP_NON_INODE_OWNER(rmap->rm_owner))
+	if ((rmap->rm_flags & (XFS_RMAP_ATTR_FORK | XFS_RMAP_BMBT_BLOCK |
+			       XFS_RMAP_UNWRITTEN)) ||
+	    XFS_RMAP_NON_INODE_OWNER(rmap->rm_owner))
 		return xfs_rmap_map(cur, rmap->rm_startblock,
 				rmap->rm_blockcount,
 				rmap->rm_flags & XFS_RMAP_UNWRITTEN,
@@ -2373,7 +2534,7 @@ xfs_rmap_query_range_helper(
 
 	fa = xfs_rmap_btrec_to_irec(rec, &irec);
 	if (!fa)
-		fa = xfs_rmap_check_irec(cur, &irec);
+		fa = xfs_rmap_check_btrec(cur, &irec);
 	if (fa)
 		return xfs_rmap_complain_bad_rec(cur, fa, &irec);
 
@@ -2428,6 +2589,38 @@ xfs_rmap_finish_one_cleanup(
 		xfs_trans_brelse(tp, agbp);
 }
 
+/* Commit an rmap operation into the ondisk tree. */
+int
+__xfs_rmap_finish_intent(
+	struct xfs_btree_cur		*rcur,
+	enum xfs_rmap_intent_type	op,
+	xfs_agblock_t			bno,
+	xfs_extlen_t			len,
+	const struct xfs_owner_info	*oinfo,
+	bool				unwritten)
+{
+	switch (op) {
+	case XFS_RMAP_ALLOC:
+	case XFS_RMAP_MAP:
+		return xfs_rmap_map(rcur, bno, len, unwritten, oinfo);
+	case XFS_RMAP_MAP_SHARED:
+		return xfs_rmap_map_shared(rcur, bno, len, unwritten, oinfo);
+	case XFS_RMAP_FREE:
+	case XFS_RMAP_UNMAP:
+		return xfs_rmap_unmap(rcur, bno, len, unwritten, oinfo);
+	case XFS_RMAP_UNMAP_SHARED:
+		return xfs_rmap_unmap_shared(rcur, bno, len, unwritten, oinfo);
+	case XFS_RMAP_CONVERT:
+		return xfs_rmap_convert(rcur, bno, len, !unwritten, oinfo);
+	case XFS_RMAP_CONVERT_SHARED:
+		return xfs_rmap_convert_shared(rcur, bno, len, !unwritten,
+				oinfo);
+	default:
+		ASSERT(0);
+		return -EFSCORRUPTED;
+	}
+}
+
 /*
  * Process one of the deferred rmap operations.  We pass back the
  * btree cursor to maintain our lock on the rmapbt between calls.
@@ -2476,10 +2669,14 @@ xfs_rmap_finish_one(
 		 * allocate blocks.
 		 */
 		error = xfs_free_extent_fix_freelist(tp, ri->ri_pag, &agbp);
-		if (error)
+		if (error) {
+			xfs_ag_mark_sick(ri->ri_pag, XFS_SICK_AG_AGFL);
 			return error;
-		if (XFS_IS_CORRUPT(tp->t_mountp, !agbp))
+		}
+		if (XFS_IS_CORRUPT(tp->t_mountp, !agbp)) {
+			xfs_ag_mark_sick(ri->ri_pag, XFS_SICK_AG_AGFL);
 			return -EFSCORRUPTED;
+		}
 
 		rcur = xfs_rmapbt_init_cursor(mp, tp, agbp, ri->ri_pag);
 	}
@@ -2490,39 +2687,14 @@ xfs_rmap_finish_one(
 	unwritten = ri->ri_bmap.br_state == XFS_EXT_UNWRITTEN;
 	bno = XFS_FSB_TO_AGBNO(rcur->bc_mp, ri->ri_bmap.br_startblock);
 
-	switch (ri->ri_type) {
-	case XFS_RMAP_ALLOC:
-	case XFS_RMAP_MAP:
-		error = xfs_rmap_map(rcur, bno, ri->ri_bmap.br_blockcount,
-				unwritten, &oinfo);
-		break;
-	case XFS_RMAP_MAP_SHARED:
-		error = xfs_rmap_map_shared(rcur, bno,
-				ri->ri_bmap.br_blockcount, unwritten, &oinfo);
-		break;
-	case XFS_RMAP_FREE:
-	case XFS_RMAP_UNMAP:
-		error = xfs_rmap_unmap(rcur, bno, ri->ri_bmap.br_blockcount,
-				unwritten, &oinfo);
-		break;
-	case XFS_RMAP_UNMAP_SHARED:
-		error = xfs_rmap_unmap_shared(rcur, bno,
-				ri->ri_bmap.br_blockcount, unwritten, &oinfo);
-		break;
-	case XFS_RMAP_CONVERT:
-		error = xfs_rmap_convert(rcur, bno, ri->ri_bmap.br_blockcount,
-				!unwritten, &oinfo);
-		break;
-	case XFS_RMAP_CONVERT_SHARED:
-		error = xfs_rmap_convert_shared(rcur, bno,
-				ri->ri_bmap.br_blockcount, !unwritten, &oinfo);
-		break;
-	default:
-		ASSERT(0);
-		error = -EFSCORRUPTED;
-	}
+	error = __xfs_rmap_finish_intent(rcur, ri->ri_type, bno,
+			ri->ri_bmap.br_blockcount, &oinfo, unwritten);
+	if (error)
+		return error;
 
-	return error;
+	xfs_rmap_update_hook(tp, ri->ri_pag, ri->ri_type, bno,
+			ri->ri_bmap.br_blockcount, unwritten, &oinfo);
+	return 0;
 }
 
 /*
@@ -2559,7 +2731,7 @@ __xfs_rmap_add(
 			bmap->br_blockcount,
 			bmap->br_state);
 
-	ri = kmem_cache_alloc(xfs_rmap_intent_cache, GFP_NOFS | __GFP_NOFAIL);
+	ri = kmem_cache_alloc(xfs_rmap_intent_cache, GFP_KERNEL | __GFP_NOFAIL);
 	INIT_LIST_HEAD(&ri->ri_list);
 	ri->ri_type = type;
 	ri->ri_owner = owner;
diff --git a/fs/xfs/libxfs/xfs_rmap.h b/fs/xfs/libxfs/xfs_rmap.h
index 3c98d9d50a..9d01fe6894 100644
--- a/fs/xfs/libxfs/xfs_rmap.h
+++ b/fs/xfs/libxfs/xfs_rmap.h
@@ -186,6 +186,10 @@ void xfs_rmap_finish_one_cleanup(struct xfs_trans *tp,
 		struct xfs_btree_cur *rcur, int error);
 int xfs_rmap_finish_one(struct xfs_trans *tp, struct xfs_rmap_intent *ri,
 		struct xfs_btree_cur **pcur);
+int __xfs_rmap_finish_intent(struct xfs_btree_cur *rcur,
+		enum xfs_rmap_intent_type op, xfs_agblock_t bno,
+		xfs_extlen_t len, const struct xfs_owner_info *oinfo,
+		bool unwritten);
 
 int xfs_rmap_lookup_le_range(struct xfs_btree_cur *cur, xfs_agblock_t bno,
 		uint64_t owner, uint64_t offset, unsigned int flags,
@@ -195,7 +199,7 @@ int xfs_rmap_compare(const struct xfs_rmap_irec *a,
 union xfs_btree_rec;
 xfs_failaddr_t xfs_rmap_btrec_to_irec(const union xfs_btree_rec *rec,
 		struct xfs_rmap_irec *irec);
-xfs_failaddr_t xfs_rmap_check_irec(struct xfs_btree_cur *cur,
+xfs_failaddr_t xfs_rmap_check_irec(struct xfs_perag *pag,
 		const struct xfs_rmap_irec *irec);
 
 int xfs_rmap_has_records(struct xfs_btree_cur *cur, xfs_agblock_t bno,
@@ -235,4 +239,29 @@ extern struct kmem_cache	*xfs_rmap_intent_cache;
 int __init xfs_rmap_intent_init_cache(void);
 void xfs_rmap_intent_destroy_cache(void);
 
+/*
+ * Parameters for tracking reverse mapping changes.  The hook function arg
+ * parameter is enum xfs_rmap_intent_type, and the rest is below.
+ */
+struct xfs_rmap_update_params {
+	xfs_agblock_t			startblock;
+	xfs_extlen_t			blockcount;
+	struct xfs_owner_info		oinfo;
+	bool				unwritten;
+};
+
+#ifdef CONFIG_XFS_LIVE_HOOKS
+
+struct xfs_rmap_hook {
+	struct xfs_hook			rmap_hook;
+};
+
+void xfs_rmap_hook_disable(void);
+void xfs_rmap_hook_enable(void);
+
+int xfs_rmap_hook_add(struct xfs_perag *pag, struct xfs_rmap_hook *hook);
+void xfs_rmap_hook_del(struct xfs_perag *pag, struct xfs_rmap_hook *hook);
+void xfs_rmap_hook_setup(struct xfs_rmap_hook *hook, notifier_fn_t mod_fn);
+#endif
+
 #endif	/* __XFS_RMAP_H__ */
diff --git a/fs/xfs/libxfs/xfs_rmap_btree.c b/fs/xfs/libxfs/xfs_rmap_btree.c
index 6c81b20e97..9e759efa81 100644
--- a/fs/xfs/libxfs/xfs_rmap_btree.c
+++ b/fs/xfs/libxfs/xfs_rmap_btree.c
@@ -16,11 +16,14 @@
 #include "xfs_btree_staging.h"
 #include "xfs_rmap.h"
 #include "xfs_rmap_btree.h"
+#include "xfs_health.h"
 #include "xfs_trace.h"
 #include "xfs_error.h"
 #include "xfs_extent_busy.h"
 #include "xfs_ag.h"
 #include "xfs_ag_resv.h"
+#include "xfs_buf_mem.h"
+#include "xfs_btree_mem.h"
 
 static struct kmem_cache	*xfs_rmapbt_cur_cache;
 
@@ -65,13 +68,12 @@ xfs_rmapbt_set_root(
 {
 	struct xfs_buf		*agbp = cur->bc_ag.agbp;
 	struct xfs_agf		*agf = agbp->b_addr;
-	int			btnum = cur->bc_btnum;
 
 	ASSERT(ptr->s != 0);
 
-	agf->agf_roots[btnum] = ptr->s;
-	be32_add_cpu(&agf->agf_levels[btnum], inc);
-	cur->bc_ag.pag->pagf_levels[btnum] += inc;
+	agf->agf_rmap_root = ptr->s;
+	be32_add_cpu(&agf->agf_rmap_level, inc);
+	cur->bc_ag.pag->pagf_rmap_level += inc;
 
 	xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS);
 }
@@ -94,8 +96,6 @@ xfs_rmapbt_alloc_block(
 				       &bno, 1);
 	if (error)
 		return error;
-
-	trace_xfs_rmapbt_alloc_block(cur->bc_mp, pag->pag_agno, bno, 1);
 	if (bno == NULLAGBLOCK) {
 		*stat = 0;
 		return 0;
@@ -125,8 +125,6 @@ xfs_rmapbt_free_block(
 	int			error;
 
 	bno = xfs_daddr_to_agbno(cur->bc_mp, xfs_buf_daddr(bp));
-	trace_xfs_rmapbt_free_block(cur->bc_mp, pag->pag_agno,
-			bno, 1);
 	be32_add_cpu(&agf->agf_rmap_blocks, -1);
 	xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_RMAP_BLOCKS);
 	error = xfs_alloc_put_freelist(pag, cur->bc_tp, agbp, NULL, bno, 1);
@@ -226,7 +224,7 @@ xfs_rmapbt_init_ptr_from_cur(
 
 	ASSERT(cur->bc_ag.pag->pag_agno == be32_to_cpu(agf->agf_seqno));
 
-	ptr->s = agf->agf_roots[cur->bc_btnum];
+	ptr->s = agf->agf_rmap_root;
 }
 
 /*
@@ -340,18 +338,29 @@ xfs_rmapbt_verify(
 
 	if (!xfs_has_rmapbt(mp))
 		return __this_address;
-	fa = xfs_btree_sblock_v5hdr_verify(bp);
+	fa = xfs_btree_agblock_v5hdr_verify(bp);
 	if (fa)
 		return fa;
 
 	level = be16_to_cpu(block->bb_level);
 	if (pag && xfs_perag_initialised_agf(pag)) {
-		if (level >= pag->pagf_levels[XFS_BTNUM_RMAPi])
+		unsigned int	maxlevel = pag->pagf_rmap_level;
+
+#ifdef CONFIG_XFS_ONLINE_REPAIR
+		/*
+		 * Online repair could be rewriting the free space btrees, so
+		 * we'll validate against the larger of either tree while this
+		 * is going on.
+		 */
+		maxlevel = max_t(unsigned int, maxlevel,
+				pag->pagf_repair_rmap_level);
+#endif
+		if (level >= maxlevel)
 			return __this_address;
 	} else if (level >= mp->m_rmap_maxlevels)
 		return __this_address;
 
-	return xfs_btree_sblock_verify(bp, mp->m_rmap_mxr[level != 0]);
+	return xfs_btree_agblock_verify(bp, mp->m_rmap_mxr[level != 0]);
 }
 
 static void
@@ -360,7 +369,7 @@ xfs_rmapbt_read_verify(
 {
 	xfs_failaddr_t	fa;
 
-	if (!xfs_btree_sblock_verify_crc(bp))
+	if (!xfs_btree_agblock_verify_crc(bp))
 		xfs_verifier_error(bp, -EFSBADCRC, __this_address);
 	else {
 		fa = xfs_rmapbt_verify(bp);
@@ -384,7 +393,7 @@ xfs_rmapbt_write_verify(
 		xfs_verifier_error(bp, -EFSCORRUPTED, fa);
 		return;
 	}
-	xfs_btree_sblock_calc_crc(bp);
+	xfs_btree_agblock_calc_crc(bp);
 
 }
 
@@ -476,9 +485,19 @@ xfs_rmapbt_keys_contiguous(
 				 be32_to_cpu(key2->rmap.rm_startblock));
 }
 
-static const struct xfs_btree_ops xfs_rmapbt_ops = {
+const struct xfs_btree_ops xfs_rmapbt_ops = {
+	.name			= "rmap",
+	.type			= XFS_BTREE_TYPE_AG,
+	.geom_flags		= XFS_BTGEO_OVERLAPPING,
+
 	.rec_len		= sizeof(struct xfs_rmap_rec),
+	/* Overlapping btree; 2 keys per pointer. */
 	.key_len		= 2 * sizeof(struct xfs_rmap_key),
+	.ptr_len		= XFS_BTREE_SHORT_PTR_LEN,
+
+	.lru_refs		= XFS_RMAP_BTREE_REF,
+	.statoff		= XFS_STATS_CALC_INDEX(xs_rmap_2),
+	.sick_mask		= XFS_SICK_AG_RMAPBT,
 
 	.dup_cursor		= xfs_rmapbt_dup_cursor,
 	.set_root		= xfs_rmapbt_set_root,
@@ -498,55 +517,176 @@ static const struct xfs_btree_ops xfs_rmapbt_ops = {
 	.keys_contiguous	= xfs_rmapbt_keys_contiguous,
 };
 
-static struct xfs_btree_cur *
-xfs_rmapbt_init_common(
+/*
+ * Create a new reverse mapping btree cursor.
+ *
+ * For staging cursors tp and agbp are NULL.
+ */
+struct xfs_btree_cur *
+xfs_rmapbt_init_cursor(
 	struct xfs_mount	*mp,
 	struct xfs_trans	*tp,
+	struct xfs_buf		*agbp,
 	struct xfs_perag	*pag)
 {
 	struct xfs_btree_cur	*cur;
 
-	/* Overlapping btree; 2 keys per pointer. */
-	cur = xfs_btree_alloc_cursor(mp, tp, XFS_BTNUM_RMAP,
+	cur = xfs_btree_alloc_cursor(mp, tp, &xfs_rmapbt_ops,
 			mp->m_rmap_maxlevels, xfs_rmapbt_cur_cache);
-	cur->bc_flags = XFS_BTREE_CRC_BLOCKS | XFS_BTREE_OVERLAPPING;
-	cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_rmap_2);
-	cur->bc_ops = &xfs_rmapbt_ops;
-
 	cur->bc_ag.pag = xfs_perag_hold(pag);
+	cur->bc_ag.agbp = agbp;
+	if (agbp) {
+		struct xfs_agf		*agf = agbp->b_addr;
+
+		cur->bc_nlevels = be32_to_cpu(agf->agf_rmap_level);
+	}
 	return cur;
 }
 
-/* Create a new reverse mapping btree cursor. */
+#ifdef CONFIG_XFS_BTREE_IN_MEM
+static inline unsigned int
+xfs_rmapbt_mem_block_maxrecs(
+	unsigned int		blocklen,
+	bool			leaf)
+{
+	if (leaf)
+		return blocklen / sizeof(struct xfs_rmap_rec);
+	return blocklen /
+		(2 * sizeof(struct xfs_rmap_key) + sizeof(__be64));
+}
+
+/*
+ * Validate an in-memory rmap btree block.  Callers are allowed to generate an
+ * in-memory btree even if the ondisk feature is not enabled.
+ */
+static xfs_failaddr_t
+xfs_rmapbt_mem_verify(
+	struct xfs_buf		*bp)
+{
+	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
+	xfs_failaddr_t		fa;
+	unsigned int		level;
+	unsigned int		maxrecs;
+
+	if (!xfs_verify_magic(bp, block->bb_magic))
+		return __this_address;
+
+	fa = xfs_btree_fsblock_v5hdr_verify(bp, XFS_RMAP_OWN_UNKNOWN);
+	if (fa)
+		return fa;
+
+	level = be16_to_cpu(block->bb_level);
+	if (level >= xfs_rmapbt_maxlevels_ondisk())
+		return __this_address;
+
+	maxrecs = xfs_rmapbt_mem_block_maxrecs(
+			XFBNO_BLOCKSIZE - XFS_BTREE_LBLOCK_CRC_LEN, level == 0);
+	return xfs_btree_memblock_verify(bp, maxrecs);
+}
+
+static void
+xfs_rmapbt_mem_rw_verify(
+	struct xfs_buf	*bp)
+{
+	xfs_failaddr_t	fa = xfs_rmapbt_mem_verify(bp);
+
+	if (fa)
+		xfs_verifier_error(bp, -EFSCORRUPTED, fa);
+}
+
+/* skip crc checks on in-memory btrees to save time */
+static const struct xfs_buf_ops xfs_rmapbt_mem_buf_ops = {
+	.name			= "xfs_rmapbt_mem",
+	.magic			= { 0, cpu_to_be32(XFS_RMAP_CRC_MAGIC) },
+	.verify_read		= xfs_rmapbt_mem_rw_verify,
+	.verify_write		= xfs_rmapbt_mem_rw_verify,
+	.verify_struct		= xfs_rmapbt_mem_verify,
+};
+
+const struct xfs_btree_ops xfs_rmapbt_mem_ops = {
+	.name			= "mem_rmap",
+	.type			= XFS_BTREE_TYPE_MEM,
+	.geom_flags		= XFS_BTGEO_OVERLAPPING,
+
+	.rec_len		= sizeof(struct xfs_rmap_rec),
+	/* Overlapping btree; 2 keys per pointer. */
+	.key_len		= 2 * sizeof(struct xfs_rmap_key),
+	.ptr_len		= XFS_BTREE_LONG_PTR_LEN,
+
+	.lru_refs		= XFS_RMAP_BTREE_REF,
+	.statoff		= XFS_STATS_CALC_INDEX(xs_rmap_mem_2),
+
+	.dup_cursor		= xfbtree_dup_cursor,
+	.set_root		= xfbtree_set_root,
+	.alloc_block		= xfbtree_alloc_block,
+	.free_block		= xfbtree_free_block,
+	.get_minrecs		= xfbtree_get_minrecs,
+	.get_maxrecs		= xfbtree_get_maxrecs,
+	.init_key_from_rec	= xfs_rmapbt_init_key_from_rec,
+	.init_high_key_from_rec	= xfs_rmapbt_init_high_key_from_rec,
+	.init_rec_from_cur	= xfs_rmapbt_init_rec_from_cur,
+	.init_ptr_from_cur	= xfbtree_init_ptr_from_cur,
+	.key_diff		= xfs_rmapbt_key_diff,
+	.buf_ops		= &xfs_rmapbt_mem_buf_ops,
+	.diff_two_keys		= xfs_rmapbt_diff_two_keys,
+	.keys_inorder		= xfs_rmapbt_keys_inorder,
+	.recs_inorder		= xfs_rmapbt_recs_inorder,
+	.keys_contiguous	= xfs_rmapbt_keys_contiguous,
+};
+
+/* Create a cursor for an in-memory btree. */
 struct xfs_btree_cur *
-xfs_rmapbt_init_cursor(
-	struct xfs_mount	*mp,
+xfs_rmapbt_mem_cursor(
+	struct xfs_perag	*pag,
 	struct xfs_trans	*tp,
-	struct xfs_buf		*agbp,
-	struct xfs_perag	*pag)
+	struct xfbtree		*xfbt)
 {
-	struct xfs_agf		*agf = agbp->b_addr;
 	struct xfs_btree_cur	*cur;
+	struct xfs_mount	*mp = pag->pag_mount;
 
-	cur = xfs_rmapbt_init_common(mp, tp, pag);
-	cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]);
-	cur->bc_ag.agbp = agbp;
+	cur = xfs_btree_alloc_cursor(mp, tp, &xfs_rmapbt_mem_ops,
+			xfs_rmapbt_maxlevels_ondisk(), xfs_rmapbt_cur_cache);
+	cur->bc_mem.xfbtree = xfbt;
+	cur->bc_nlevels = xfbt->nlevels;
+
+	cur->bc_mem.pag = xfs_perag_hold(pag);
 	return cur;
 }
 
-/* Create a new reverse mapping btree cursor with a fake root for staging. */
-struct xfs_btree_cur *
-xfs_rmapbt_stage_cursor(
+/* Create an in-memory rmap btree. */
+int
+xfs_rmapbt_mem_init(
 	struct xfs_mount	*mp,
-	struct xbtree_afakeroot	*afake,
-	struct xfs_perag	*pag)
+	struct xfbtree		*xfbt,
+	struct xfs_buftarg	*btp,
+	xfs_agnumber_t		agno)
 {
-	struct xfs_btree_cur	*cur;
+	xfbt->owner = agno;
+	return xfbtree_init(mp, xfbt, btp, &xfs_rmapbt_mem_ops);
+}
 
-	cur = xfs_rmapbt_init_common(mp, NULL, pag);
-	xfs_btree_stage_afakeroot(cur, afake);
-	return cur;
+/* Compute the max possible height for reverse mapping btrees in memory. */
+static unsigned int
+xfs_rmapbt_mem_maxlevels(void)
+{
+	unsigned int		minrecs[2];
+	unsigned int		blocklen;
+
+	blocklen = XFBNO_BLOCKSIZE - XFS_BTREE_LBLOCK_CRC_LEN;
+
+	minrecs[0] = xfs_rmapbt_mem_block_maxrecs(blocklen, true) / 2;
+	minrecs[1] = xfs_rmapbt_mem_block_maxrecs(blocklen, false) / 2;
+
+	/*
+	 * How tall can an in-memory rmap btree become if we filled the entire
+	 * AG with rmap records?
+	 */
+	return xfs_btree_compute_maxlevels(minrecs,
+			XFS_MAX_AG_BYTES / sizeof(struct xfs_rmap_rec));
 }
+#else
+# define xfs_rmapbt_mem_maxlevels()	(0)
+#endif /* CONFIG_XFS_BTREE_IN_MEM */
 
 /*
  * Install a new reverse mapping btree root.  Caller is responsible for
@@ -563,12 +703,12 @@ xfs_rmapbt_commit_staged_btree(
 
 	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
 
-	agf->agf_roots[cur->bc_btnum] = cpu_to_be32(afake->af_root);
-	agf->agf_levels[cur->bc_btnum] = cpu_to_be32(afake->af_levels);
+	agf->agf_rmap_root = cpu_to_be32(afake->af_root);
+	agf->agf_rmap_level = cpu_to_be32(afake->af_levels);
 	agf->agf_rmap_blocks = cpu_to_be32(afake->af_blocks);
 	xfs_alloc_log_agf(tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS |
 				    XFS_AGF_RMAP_BLOCKS);
-	xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_rmapbt_ops);
+	xfs_btree_commit_afakeroot(cur, tp, agbp);
 }
 
 /* Calculate number of records in a reverse mapping btree block. */
@@ -618,7 +758,8 @@ xfs_rmapbt_maxlevels_ondisk(void)
 	 * like if it consumes almost all the blocks in the AG due to maximal
 	 * sharing factor.
 	 */
-	return xfs_btree_space_to_height(minrecs, XFS_MAX_CRC_AG_BLOCKS);
+	return max(xfs_btree_space_to_height(minrecs, XFS_MAX_CRC_AG_BLOCKS),
+		   xfs_rmapbt_mem_maxlevels());
 }
 
 /* Compute the maximum height of an rmap btree. */
diff --git a/fs/xfs/libxfs/xfs_rmap_btree.h b/fs/xfs/libxfs/xfs_rmap_btree.h
index 3244715dd1..eb90d89e80 100644
--- a/fs/xfs/libxfs/xfs_rmap_btree.h
+++ b/fs/xfs/libxfs/xfs_rmap_btree.h
@@ -10,6 +10,7 @@ struct xfs_buf;
 struct xfs_btree_cur;
 struct xfs_mount;
 struct xbtree_afakeroot;
+struct xfbtree;
 
 /* rmaps only exist on crc enabled filesystems */
 #define XFS_RMAP_BLOCK_LEN	XFS_BTREE_SBLOCK_CRC_LEN
@@ -44,8 +45,6 @@ struct xbtree_afakeroot;
 struct xfs_btree_cur *xfs_rmapbt_init_cursor(struct xfs_mount *mp,
 				struct xfs_trans *tp, struct xfs_buf *bp,
 				struct xfs_perag *pag);
-struct xfs_btree_cur *xfs_rmapbt_stage_cursor(struct xfs_mount *mp,
-		struct xbtree_afakeroot *afake, struct xfs_perag *pag);
 void xfs_rmapbt_commit_staged_btree(struct xfs_btree_cur *cur,
 		struct xfs_trans *tp, struct xfs_buf *agbp);
 int xfs_rmapbt_maxrecs(int blocklen, int leaf);
@@ -64,4 +63,9 @@ unsigned int xfs_rmapbt_maxlevels_ondisk(void);
 int __init xfs_rmapbt_init_cur_cache(void);
 void xfs_rmapbt_destroy_cur_cache(void);
 
+struct xfs_btree_cur *xfs_rmapbt_mem_cursor(struct xfs_perag *pag,
+		struct xfs_trans *tp, struct xfbtree *xfbtree);
+int xfs_rmapbt_mem_init(struct xfs_mount *mp, struct xfbtree *xfbtree,
+		struct xfs_buftarg *btp, xfs_agnumber_t agno);
+
 #endif /* __XFS_RMAP_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_rtbitmap.c b/fs/xfs/libxfs/xfs_rtbitmap.c
index e31663cb7b..386b672c50 100644
--- a/fs/xfs/libxfs/xfs_rtbitmap.c
+++ b/fs/xfs/libxfs/xfs_rtbitmap.c
@@ -17,6 +17,7 @@
 #include "xfs_rtalloc.h"
 #include "xfs_error.h"
 #include "xfs_rtbitmap.h"
+#include "xfs_health.h"
 
 /*
  * Realtime allocator bitmap functions shared with userspace.
@@ -115,13 +116,19 @@ xfs_rtbuf_get(
 	if (error)
 		return error;
 
-	if (XFS_IS_CORRUPT(mp, nmap == 0 || !xfs_bmap_is_written_extent(&map)))
+	if (XFS_IS_CORRUPT(mp, nmap == 0 || !xfs_bmap_is_written_extent(&map))) {
+		xfs_rt_mark_sick(mp, issum ? XFS_SICK_RT_SUMMARY :
+					     XFS_SICK_RT_BITMAP);
 		return -EFSCORRUPTED;
+	}
 
 	ASSERT(map.br_startblock != NULLFSBLOCK);
 	error = xfs_trans_read_buf(mp, args->tp, mp->m_ddev_targp,
 				   XFS_FSB_TO_DADDR(mp, map.br_startblock),
 				   mp->m_bsize, 0, &bp, &xfs_rtbuf_ops);
+	if (xfs_metadata_is_sick(error))
+		xfs_rt_mark_sick(mp, issum ? XFS_SICK_RT_SUMMARY :
+					     XFS_SICK_RT_BITMAP);
 	if (error)
 		return error;
 
@@ -934,7 +941,7 @@ xfs_rtfree_extent(
 	struct timespec64	atime;
 
 	ASSERT(mp->m_rbmip->i_itemp != NULL);
-	ASSERT(xfs_isilocked(mp->m_rbmip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(mp->m_rbmip, XFS_ILOCK_EXCL);
 
 	error = xfs_rtcheck_alloc_range(&args, start, len);
 	if (error)
@@ -1161,3 +1168,60 @@ xfs_rtsummary_wordcount(
 	blocks = xfs_rtsummary_blockcount(mp, rsumlevels, rbmblocks);
 	return XFS_FSB_TO_B(mp, blocks) >> XFS_WORDLOG;
 }
+
+/*
+ * Lock both realtime free space metadata inodes for a freespace update.  If a
+ * transaction is given, the inodes will be joined to the transaction and the
+ * ILOCKs will be released on transaction commit.
+ */
+void
+xfs_rtbitmap_lock(
+	struct xfs_trans	*tp,
+	struct xfs_mount	*mp)
+{
+	xfs_ilock(mp->m_rbmip, XFS_ILOCK_EXCL | XFS_ILOCK_RTBITMAP);
+	if (tp)
+		xfs_trans_ijoin(tp, mp->m_rbmip, XFS_ILOCK_EXCL);
+
+	xfs_ilock(mp->m_rsumip, XFS_ILOCK_EXCL | XFS_ILOCK_RTSUM);
+	if (tp)
+		xfs_trans_ijoin(tp, mp->m_rsumip, XFS_ILOCK_EXCL);
+}
+
+/* Unlock both realtime free space metadata inodes after a freespace update. */
+void
+xfs_rtbitmap_unlock(
+	struct xfs_mount	*mp)
+{
+	xfs_iunlock(mp->m_rsumip, XFS_ILOCK_EXCL | XFS_ILOCK_RTSUM);
+	xfs_iunlock(mp->m_rbmip, XFS_ILOCK_EXCL | XFS_ILOCK_RTBITMAP);
+}
+
+/*
+ * Lock the realtime free space metadata inodes for a freespace scan.  Callers
+ * must walk metadata blocks in order of increasing file offset.
+ */
+void
+xfs_rtbitmap_lock_shared(
+	struct xfs_mount	*mp,
+	unsigned int		rbmlock_flags)
+{
+	if (rbmlock_flags & XFS_RBMLOCK_BITMAP)
+		xfs_ilock(mp->m_rbmip, XFS_ILOCK_SHARED | XFS_ILOCK_RTBITMAP);
+
+	if (rbmlock_flags & XFS_RBMLOCK_SUMMARY)
+		xfs_ilock(mp->m_rsumip, XFS_ILOCK_SHARED | XFS_ILOCK_RTSUM);
+}
+
+/* Unlock the realtime free space metadata inodes after a freespace scan. */
+void
+xfs_rtbitmap_unlock_shared(
+	struct xfs_mount	*mp,
+	unsigned int		rbmlock_flags)
+{
+	if (rbmlock_flags & XFS_RBMLOCK_SUMMARY)
+		xfs_iunlock(mp->m_rsumip, XFS_ILOCK_SHARED | XFS_ILOCK_RTSUM);
+
+	if (rbmlock_flags & XFS_RBMLOCK_BITMAP)
+		xfs_iunlock(mp->m_rbmip, XFS_ILOCK_SHARED | XFS_ILOCK_RTBITMAP);
+}
diff --git a/fs/xfs/libxfs/xfs_rtbitmap.h b/fs/xfs/libxfs/xfs_rtbitmap.h
index 152a66750a..6186585f2c 100644
--- a/fs/xfs/libxfs/xfs_rtbitmap.h
+++ b/fs/xfs/libxfs/xfs_rtbitmap.h
@@ -360,6 +360,19 @@ xfs_filblks_t xfs_rtsummary_blockcount(struct xfs_mount *mp,
 		unsigned int rsumlevels, xfs_extlen_t rbmblocks);
 unsigned long long xfs_rtsummary_wordcount(struct xfs_mount *mp,
 		unsigned int rsumlevels, xfs_extlen_t rbmblocks);
+
+void xfs_rtbitmap_lock(struct xfs_trans *tp, struct xfs_mount *mp);
+void xfs_rtbitmap_unlock(struct xfs_mount *mp);
+
+/* Lock the rt bitmap inode in shared mode */
+#define XFS_RBMLOCK_BITMAP	(1U << 0)
+/* Lock the rt summary inode in shared mode */
+#define XFS_RBMLOCK_SUMMARY	(1U << 1)
+
+void xfs_rtbitmap_lock_shared(struct xfs_mount *mp,
+		unsigned int rbmlock_flags);
+void xfs_rtbitmap_unlock_shared(struct xfs_mount *mp,
+		unsigned int rbmlock_flags);
 #else /* CONFIG_XFS_RT */
 # define xfs_rtfree_extent(t,b,l)			(-ENOSYS)
 # define xfs_rtfree_blocks(t,rb,rl)			(-ENOSYS)
@@ -378,6 +391,10 @@ xfs_rtbitmap_blockcount(struct xfs_mount *mp, xfs_rtbxlen_t rtextents)
 # define xfs_rtbitmap_wordcount(mp, r)			(0)
 # define xfs_rtsummary_blockcount(mp, l, b)		(0)
 # define xfs_rtsummary_wordcount(mp, l, b)		(0)
+# define xfs_rtbitmap_lock(tp, mp)		do { } while (0)
+# define xfs_rtbitmap_unlock(mp)		do { } while (0)
+# define xfs_rtbitmap_lock_shared(mp, lf)	do { } while (0)
+# define xfs_rtbitmap_unlock_shared(mp, lf)	do { } while (0)
 #endif /* CONFIG_XFS_RT */
 
 #endif /* __XFS_RTBITMAP_H__ */
diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c
index 5bb6e2bd6d..6b56f0f6d4 100644
--- a/fs/xfs/libxfs/xfs_sb.c
+++ b/fs/xfs/libxfs/xfs_sb.c
@@ -26,6 +26,7 @@
 #include "xfs_health.h"
 #include "xfs_ag.h"
 #include "xfs_rtbitmap.h"
+#include "xfs_exchrange.h"
 
 /*
  * Physical superblock buffer manipulations. Shared with libxfs in userspace.
@@ -175,6 +176,10 @@ xfs_sb_version_to_features(
 		features |= XFS_FEAT_NEEDSREPAIR;
 	if (sbp->sb_features_incompat & XFS_SB_FEAT_INCOMPAT_NREXT64)
 		features |= XFS_FEAT_NREXT64;
+	if (sbp->sb_features_incompat & XFS_SB_FEAT_INCOMPAT_EXCHRANGE)
+		features |= XFS_FEAT_EXCHANGE_RANGE;
+	if (sbp->sb_features_incompat & XFS_SB_FEAT_INCOMPAT_PARENT)
+		features |= XFS_FEAT_PARENT;
 
 	return features;
 }
@@ -530,7 +535,8 @@ xfs_validate_sb_common(
 	}
 
 	if (!xfs_validate_stripe_geometry(mp, XFS_FSB_TO_B(mp, sbp->sb_unit),
-			XFS_FSB_TO_B(mp, sbp->sb_width), 0, false))
+			XFS_FSB_TO_B(mp, sbp->sb_width), 0,
+			xfs_buf_daddr(bp) == XFS_SB_DADDR, false))
 		return -EFSCORRUPTED;
 
 	/*
@@ -1032,11 +1038,12 @@ xfs_log_sb(
 	 * and hence we don't need have to update it here.
 	 */
 	if (xfs_has_lazysbcount(mp)) {
-		mp->m_sb.sb_icount = percpu_counter_sum(&mp->m_icount);
+		mp->m_sb.sb_icount = percpu_counter_sum_positive(&mp->m_icount);
 		mp->m_sb.sb_ifree = min_t(uint64_t,
-				percpu_counter_sum(&mp->m_ifree),
+				percpu_counter_sum_positive(&mp->m_ifree),
 				mp->m_sb.sb_icount);
-		mp->m_sb.sb_fdblocks = percpu_counter_sum(&mp->m_fdblocks);
+		mp->m_sb.sb_fdblocks =
+				percpu_counter_sum_positive(&mp->m_fdblocks);
 	}
 
 	xfs_sb_to_disk(bp->b_addr, &mp->m_sb);
@@ -1250,6 +1257,8 @@ xfs_fs_geometry(
 		geo->flags |= XFS_FSOP_GEOM_FLAGS_BIGTIME;
 	if (xfs_has_inobtcounts(mp))
 		geo->flags |= XFS_FSOP_GEOM_FLAGS_INOBTCNT;
+	if (xfs_has_parent(mp))
+		geo->flags |= XFS_FSOP_GEOM_FLAGS_PARENT;
 	if (xfs_has_sector(mp)) {
 		geo->flags |= XFS_FSOP_GEOM_FLAGS_SECTOR;
 		geo->logsectsize = sbp->sb_logsectsize;
@@ -1258,6 +1267,8 @@ xfs_fs_geometry(
 	}
 	if (xfs_has_large_extent_counts(mp))
 		geo->flags |= XFS_FSOP_GEOM_FLAGS_NREXT64;
+	if (xfs_has_exchange_range(mp))
+		geo->flags |= XFS_FSOP_GEOM_FLAGS_EXCHANGE_RANGE;
 	geo->rtsectsize = sbp->sb_blocksize;
 	geo->dirblocksize = xfs_dir2_dirblock_bytes(sbp);
 
@@ -1290,6 +1301,8 @@ xfs_sb_read_secondary(
 	error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp,
 			XFS_AG_DADDR(mp, agno, XFS_SB_BLOCK(mp)),
 			XFS_FSS_TO_BB(mp, 1), 0, &bp, &xfs_sb_buf_ops);
+	if (xfs_metadata_is_sick(error))
+		xfs_agno_mark_sick(mp, agno, XFS_SICK_AG_SB);
 	if (error)
 		return error;
 	xfs_buf_set_ref(bp, XFS_SSB_REF);
@@ -1321,8 +1334,10 @@ xfs_sb_get_secondary(
 }
 
 /*
- * sunit, swidth, sectorsize(optional with 0) should be all in bytes,
- * so users won't be confused by values in error messages.
+ * sunit, swidth, sectorsize(optional with 0) should be all in bytes, so users
+ * won't be confused by values in error messages.  This function returns false
+ * if the stripe geometry is invalid and the caller is unable to repair the
+ * stripe configuration later in the mount process.
  */
 bool
 xfs_validate_stripe_geometry(
@@ -1330,20 +1345,21 @@ xfs_validate_stripe_geometry(
 	__s64			sunit,
 	__s64			swidth,
 	int			sectorsize,
+	bool			may_repair,
 	bool			silent)
 {
 	if (swidth > INT_MAX) {
 		if (!silent)
 			xfs_notice(mp,
 "stripe width (%lld) is too large", swidth);
-		return false;
+		goto check_override;
 	}
 
 	if (sunit > swidth) {
 		if (!silent)
 			xfs_notice(mp,
 "stripe unit (%lld) is larger than the stripe width (%lld)", sunit, swidth);
-		return false;
+		goto check_override;
 	}
 
 	if (sectorsize && (int)sunit % sectorsize) {
@@ -1351,21 +1367,21 @@ xfs_validate_stripe_geometry(
 			xfs_notice(mp,
 "stripe unit (%lld) must be a multiple of the sector size (%d)",
 				   sunit, sectorsize);
-		return false;
+		goto check_override;
 	}
 
 	if (sunit && !swidth) {
 		if (!silent)
 			xfs_notice(mp,
 "invalid stripe unit (%lld) and stripe width of 0", sunit);
-		return false;
+		goto check_override;
 	}
 
 	if (!sunit && swidth) {
 		if (!silent)
 			xfs_notice(mp,
 "invalid stripe width (%lld) and stripe unit of 0", swidth);
-		return false;
+		goto check_override;
 	}
 
 	if (sunit && (int)swidth % (int)sunit) {
@@ -1373,9 +1389,27 @@ xfs_validate_stripe_geometry(
 			xfs_notice(mp,
 "stripe width (%lld) must be a multiple of the stripe unit (%lld)",
 				   swidth, sunit);
-		return false;
+		goto check_override;
 	}
 	return true;
+
+check_override:
+	if (!may_repair)
+		return false;
+	/*
+	 * During mount, mp->m_dalign will not be set unless the sunit mount
+	 * option was set. If it was set, ignore the bad stripe alignment values
+	 * and allow the validation and overwrite later in the mount process to
+	 * attempt to overwrite the bad stripe alignment values with the values
+	 * supplied by mount options.
+	 */
+	if (!mp->m_dalign)
+		return false;
+	if (!silent)
+		xfs_notice(mp,
+"Will try to correct with specified mount options sunit (%d) and swidth (%d)",
+			BBTOB(mp->m_dalign), BBTOB(mp->m_swidth));
+	return true;
 }
 
 /*
diff --git a/fs/xfs/libxfs/xfs_sb.h b/fs/xfs/libxfs/xfs_sb.h
index 2e8e8d63d4..37b1ed1bc2 100644
--- a/fs/xfs/libxfs/xfs_sb.h
+++ b/fs/xfs/libxfs/xfs_sb.h
@@ -35,8 +35,9 @@ extern int	xfs_sb_get_secondary(struct xfs_mount *mp,
 				struct xfs_trans *tp, xfs_agnumber_t agno,
 				struct xfs_buf **bpp);
 
-extern bool	xfs_validate_stripe_geometry(struct xfs_mount *mp,
-		__s64 sunit, __s64 swidth, int sectorsize, bool silent);
+bool	xfs_validate_stripe_geometry(struct xfs_mount *mp,
+		__s64 sunit, __s64 swidth, int sectorsize, bool may_repair,
+		bool silent);
 
 uint8_t xfs_compute_rextslog(xfs_rtbxlen_t rtextents);
 
diff --git a/fs/xfs/libxfs/xfs_shared.h b/fs/xfs/libxfs/xfs_shared.h
index 4220d3584c..34f104ed37 100644
--- a/fs/xfs/libxfs/xfs_shared.h
+++ b/fs/xfs/libxfs/xfs_shared.h
@@ -43,6 +43,60 @@ extern const struct xfs_buf_ops xfs_sb_buf_ops;
 extern const struct xfs_buf_ops xfs_sb_quiet_buf_ops;
 extern const struct xfs_buf_ops xfs_symlink_buf_ops;
 
+/* btree ops */
+extern const struct xfs_btree_ops xfs_bnobt_ops;
+extern const struct xfs_btree_ops xfs_cntbt_ops;
+extern const struct xfs_btree_ops xfs_inobt_ops;
+extern const struct xfs_btree_ops xfs_finobt_ops;
+extern const struct xfs_btree_ops xfs_bmbt_ops;
+extern const struct xfs_btree_ops xfs_refcountbt_ops;
+extern const struct xfs_btree_ops xfs_rmapbt_ops;
+extern const struct xfs_btree_ops xfs_rmapbt_mem_ops;
+
+static inline bool xfs_btree_is_bno(const struct xfs_btree_ops *ops)
+{
+	return ops == &xfs_bnobt_ops;
+}
+
+static inline bool xfs_btree_is_cnt(const struct xfs_btree_ops *ops)
+{
+	return ops == &xfs_cntbt_ops;
+}
+
+static inline bool xfs_btree_is_bmap(const struct xfs_btree_ops *ops)
+{
+	return ops == &xfs_bmbt_ops;
+}
+
+static inline bool xfs_btree_is_ino(const struct xfs_btree_ops *ops)
+{
+	return ops == &xfs_inobt_ops;
+}
+
+static inline bool xfs_btree_is_fino(const struct xfs_btree_ops *ops)
+{
+	return ops == &xfs_finobt_ops;
+}
+
+static inline bool xfs_btree_is_refcount(const struct xfs_btree_ops *ops)
+{
+	return ops == &xfs_refcountbt_ops;
+}
+
+static inline bool xfs_btree_is_rmap(const struct xfs_btree_ops *ops)
+{
+	return ops == &xfs_rmapbt_ops;
+}
+
+#ifdef CONFIG_XFS_BTREE_IN_MEM
+static inline bool xfs_btree_is_mem_rmap(const struct xfs_btree_ops *ops)
+{
+	return ops == &xfs_rmapbt_mem_ops;
+}
+#else
+# define xfs_btree_is_mem_rmap(...)	(false)
+#endif
+
 /* log size calculation functions */
 int	xfs_log_calc_unit_res(struct xfs_mount *mp, int unit_bytes);
 int	xfs_log_calc_minimum_size(struct xfs_mount *);
@@ -70,7 +124,6 @@ void	xfs_log_get_max_trans_res(struct xfs_mount *mp,
 #define XFS_TRANS_RES_FDBLKS		(1u << 6)
 /* Transaction contains an intent done log item */
 #define XFS_TRANS_HAS_INTENT_DONE	(1u << 7)
-
 /*
  * LOWMODE is used by the allocator to activate the lowspace algorithm - when
  * free space is running low the extent allocator may choose to allocate an
@@ -82,7 +135,10 @@ void	xfs_log_get_max_trans_res(struct xfs_mount *mp,
  * for free space from AG 0. If the correct transaction reservations have been
  * made then this algorithm will eventually find all the space it needs.
  */
-#define XFS_TRANS_LOWMODE	0x100	/* allocate in low space mode */
+#define XFS_TRANS_LOWMODE		(1u << 8)
+
+/* Transaction has locked the rtbitmap and rtsum inodes */
+#define XFS_TRANS_RTBITMAP_LOCKED	(1u << 9)
 
 /*
  * Field values for xfs_trans_mod_sb.
@@ -128,19 +184,6 @@ void	xfs_log_get_max_trans_res(struct xfs_mount *mp,
 #define	XFS_ICHGTIME_CHG	0x2	/* inode field change timestamp */
 #define	XFS_ICHGTIME_CREATE	0x4	/* inode create timestamp */
 
-
-/*
- * Symlink decoding/encoding functions
- */
-int xfs_symlink_blocks(struct xfs_mount *mp, int pathlen);
-int xfs_symlink_hdr_set(struct xfs_mount *mp, xfs_ino_t ino, uint32_t offset,
-			uint32_t size, struct xfs_buf *bp);
-bool xfs_symlink_hdr_ok(xfs_ino_t ino, uint32_t offset,
-			uint32_t size, struct xfs_buf *bp);
-void xfs_symlink_local_to_remote(struct xfs_trans *tp, struct xfs_buf *bp,
-				 struct xfs_inode *ip, struct xfs_ifork *ifp);
-xfs_failaddr_t xfs_symlink_shortform_verify(void *sfp, int64_t size);
-
 /* Computed inode geometry for the filesystem. */
 struct xfs_ino_geometry {
 	/* Maximum inode count in this filesystem. */
diff --git a/fs/xfs/libxfs/xfs_symlink_remote.c b/fs/xfs/libxfs/xfs_symlink_remote.c
index 160aa20aa4..f228127a88 100644
--- a/fs/xfs/libxfs/xfs_symlink_remote.c
+++ b/fs/xfs/libxfs/xfs_symlink_remote.c
@@ -16,7 +16,10 @@
 #include "xfs_trans.h"
 #include "xfs_buf_item.h"
 #include "xfs_log.h"
-
+#include "xfs_symlink_remote.h"
+#include "xfs_bit.h"
+#include "xfs_bmap.h"
+#include "xfs_health.h"
 
 /*
  * Each contiguous block has a header, so it is not just a simple pathlen
@@ -166,7 +169,8 @@ xfs_symlink_local_to_remote(
 	struct xfs_trans	*tp,
 	struct xfs_buf		*bp,
 	struct xfs_inode	*ip,
-	struct xfs_ifork	*ifp)
+	struct xfs_ifork	*ifp,
+	void			*priv)
 {
 	struct xfs_mount	*mp = ip->i_mount;
 	char			*buf;
@@ -227,3 +231,200 @@ xfs_symlink_shortform_verify(
 		return __this_address;
 	return NULL;
 }
+
+/* Read a remote symlink target into the buffer. */
+int
+xfs_symlink_remote_read(
+	struct xfs_inode	*ip,
+	char			*link)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_bmbt_irec	mval[XFS_SYMLINK_MAPS];
+	struct xfs_buf		*bp;
+	xfs_daddr_t		d;
+	char			*cur_chunk;
+	int			pathlen = ip->i_disk_size;
+	int			nmaps = XFS_SYMLINK_MAPS;
+	int			byte_cnt;
+	int			n;
+	int			error = 0;
+	int			fsblocks = 0;
+	int			offset;
+
+	xfs_assert_ilocked(ip, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL);
+
+	fsblocks = xfs_symlink_blocks(mp, pathlen);
+	error = xfs_bmapi_read(ip, 0, fsblocks, mval, &nmaps, 0);
+	if (error)
+		goto out;
+
+	offset = 0;
+	for (n = 0; n < nmaps; n++) {
+		d = XFS_FSB_TO_DADDR(mp, mval[n].br_startblock);
+		byte_cnt = XFS_FSB_TO_B(mp, mval[n].br_blockcount);
+
+		error = xfs_buf_read(mp->m_ddev_targp, d, BTOBB(byte_cnt), 0,
+				&bp, &xfs_symlink_buf_ops);
+		if (xfs_metadata_is_sick(error))
+			xfs_inode_mark_sick(ip, XFS_SICK_INO_SYMLINK);
+		if (error)
+			return error;
+		byte_cnt = XFS_SYMLINK_BUF_SPACE(mp, byte_cnt);
+		if (pathlen < byte_cnt)
+			byte_cnt = pathlen;
+
+		cur_chunk = bp->b_addr;
+		if (xfs_has_crc(mp)) {
+			if (!xfs_symlink_hdr_ok(ip->i_ino, offset,
+							byte_cnt, bp)) {
+				xfs_inode_mark_sick(ip, XFS_SICK_INO_SYMLINK);
+				error = -EFSCORRUPTED;
+				xfs_alert(mp,
+"symlink header does not match required off/len/owner (0x%x/0x%x,0x%llx)",
+					offset, byte_cnt, ip->i_ino);
+				xfs_buf_relse(bp);
+				goto out;
+
+			}
+
+			cur_chunk += sizeof(struct xfs_dsymlink_hdr);
+		}
+
+		memcpy(link + offset, cur_chunk, byte_cnt);
+
+		pathlen -= byte_cnt;
+		offset += byte_cnt;
+
+		xfs_buf_relse(bp);
+	}
+	ASSERT(pathlen == 0);
+
+	link[ip->i_disk_size] = '\0';
+	error = 0;
+
+ out:
+	return error;
+}
+
+/* Write the symlink target into the inode. */
+int
+xfs_symlink_write_target(
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip,
+	xfs_ino_t		owner,
+	const char		*target_path,
+	int			pathlen,
+	xfs_fsblock_t		fs_blocks,
+	uint			resblks)
+{
+	struct xfs_bmbt_irec	mval[XFS_SYMLINK_MAPS];
+	struct xfs_mount	*mp = tp->t_mountp;
+	const char		*cur_chunk;
+	struct xfs_buf		*bp;
+	xfs_daddr_t		d;
+	int			byte_cnt;
+	int			nmaps;
+	int			offset = 0;
+	int			n;
+	int			error;
+
+	/*
+	 * If the symlink will fit into the inode, write it inline.
+	 */
+	if (pathlen <= xfs_inode_data_fork_size(ip)) {
+		xfs_init_local_fork(ip, XFS_DATA_FORK, target_path, pathlen);
+
+		ip->i_disk_size = pathlen;
+		ip->i_df.if_format = XFS_DINODE_FMT_LOCAL;
+		xfs_trans_log_inode(tp, ip, XFS_ILOG_DDATA | XFS_ILOG_CORE);
+		return 0;
+	}
+
+	nmaps = XFS_SYMLINK_MAPS;
+	error = xfs_bmapi_write(tp, ip, 0, fs_blocks, XFS_BMAPI_METADATA,
+			resblks, mval, &nmaps);
+	if (error)
+		return error;
+
+	ip->i_disk_size = pathlen;
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+
+	cur_chunk = target_path;
+	offset = 0;
+	for (n = 0; n < nmaps; n++) {
+		char	*buf;
+
+		d = XFS_FSB_TO_DADDR(mp, mval[n].br_startblock);
+		byte_cnt = XFS_FSB_TO_B(mp, mval[n].br_blockcount);
+		error = xfs_trans_get_buf(tp, mp->m_ddev_targp, d,
+				BTOBB(byte_cnt), 0, &bp);
+		if (error)
+			return error;
+		bp->b_ops = &xfs_symlink_buf_ops;
+
+		byte_cnt = XFS_SYMLINK_BUF_SPACE(mp, byte_cnt);
+		byte_cnt = min(byte_cnt, pathlen);
+
+		buf = bp->b_addr;
+		buf += xfs_symlink_hdr_set(mp, owner, offset, byte_cnt, bp);
+
+		memcpy(buf, cur_chunk, byte_cnt);
+
+		cur_chunk += byte_cnt;
+		pathlen -= byte_cnt;
+		offset += byte_cnt;
+
+		xfs_trans_buf_set_type(tp, bp, XFS_BLFT_SYMLINK_BUF);
+		xfs_trans_log_buf(tp, bp, 0, (buf + byte_cnt - 1) -
+						(char *)bp->b_addr);
+	}
+	ASSERT(pathlen == 0);
+	return 0;
+}
+
+/* Remove all the blocks from a symlink and invalidate buffers. */
+int
+xfs_symlink_remote_truncate(
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip)
+{
+	struct xfs_bmbt_irec	mval[XFS_SYMLINK_MAPS];
+	struct xfs_mount	*mp = tp->t_mountp;
+	struct xfs_buf		*bp;
+	int			nmaps = XFS_SYMLINK_MAPS;
+	int			done = 0;
+	int			i;
+	int			error;
+
+	/* Read mappings and invalidate buffers. */
+	error = xfs_bmapi_read(ip, 0, XFS_MAX_FILEOFF, mval, &nmaps, 0);
+	if (error)
+		return error;
+
+	for (i = 0; i < nmaps; i++) {
+		if (!xfs_bmap_is_real_extent(&mval[i]))
+			break;
+
+		error = xfs_trans_get_buf(tp, mp->m_ddev_targp,
+				XFS_FSB_TO_DADDR(mp, mval[i].br_startblock),
+				XFS_FSB_TO_BB(mp, mval[i].br_blockcount), 0,
+				&bp);
+		if (error)
+			return error;
+
+		xfs_trans_binval(tp, bp);
+	}
+
+	/* Unmap the remote blocks. */
+	error = xfs_bunmapi(tp, ip, 0, XFS_MAX_FILEOFF, 0, nmaps, &done);
+	if (error)
+		return error;
+	if (!done) {
+		ASSERT(done);
+		xfs_inode_mark_sick(ip, XFS_SICK_INO_SYMLINK);
+		return -EFSCORRUPTED;
+	}
+
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+	return 0;
+}
diff --git a/fs/xfs/libxfs/xfs_symlink_remote.h b/fs/xfs/libxfs/xfs_symlink_remote.h
new file mode 100644
index 0000000000..c1672fe1f1
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_symlink_remote.h
@@ -0,0 +1,28 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * Copyright (c) 2013 Red Hat, Inc.
+ * All Rights Reserved.
+ */
+#ifndef __XFS_SYMLINK_REMOTE_H
+#define __XFS_SYMLINK_REMOTE_H
+
+/*
+ * Symlink decoding/encoding functions
+ */
+int xfs_symlink_blocks(struct xfs_mount *mp, int pathlen);
+int xfs_symlink_hdr_set(struct xfs_mount *mp, xfs_ino_t ino, uint32_t offset,
+			uint32_t size, struct xfs_buf *bp);
+bool xfs_symlink_hdr_ok(xfs_ino_t ino, uint32_t offset,
+			uint32_t size, struct xfs_buf *bp);
+void xfs_symlink_local_to_remote(struct xfs_trans *tp, struct xfs_buf *bp,
+				 struct xfs_inode *ip, struct xfs_ifork *ifp,
+				 void *priv);
+xfs_failaddr_t xfs_symlink_shortform_verify(void *sfp, int64_t size);
+int xfs_symlink_remote_read(struct xfs_inode *ip, char *link);
+int xfs_symlink_write_target(struct xfs_trans *tp, struct xfs_inode *ip,
+		xfs_ino_t owner, const char *target_path, int pathlen,
+		xfs_fsblock_t fs_blocks, uint resblks);
+int xfs_symlink_remote_truncate(struct xfs_trans *tp, struct xfs_inode *ip);
+
+#endif /* __XFS_SYMLINK_REMOTE_H */
diff --git a/fs/xfs/libxfs/xfs_trans_inode.c b/fs/xfs/libxfs/xfs_trans_inode.c
index 70e97ea6ee..69fc5b9813 100644
--- a/fs/xfs/libxfs/xfs_trans_inode.c
+++ b/fs/xfs/libxfs/xfs_trans_inode.c
@@ -31,7 +31,7 @@ xfs_trans_ijoin(
 {
 	struct xfs_inode_log_item *iip;
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 	if (ip->i_itemp == NULL)
 		xfs_inode_item_init(ip, ip->i_mount);
 	iip = ip->i_itemp;
@@ -60,7 +60,7 @@ xfs_trans_ichgtime(
 	struct timespec64	tv;
 
 	ASSERT(tp);
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 
 	tv = current_time(inode);
 
@@ -90,7 +90,7 @@ xfs_trans_log_inode(
 	struct inode		*inode = VFS_I(ip);
 
 	ASSERT(iip);
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 	ASSERT(!xfs_iflags_test(ip, XFS_ISTALE));
 
 	tp->t_flags |= XFS_TRANS_DIRTY;
diff --git a/fs/xfs/libxfs/xfs_trans_resv.c b/fs/xfs/libxfs/xfs_trans_resv.c
index 6cd45e8c11..6dbe6e7251 100644
--- a/fs/xfs/libxfs/xfs_trans_resv.c
+++ b/fs/xfs/libxfs/xfs_trans_resv.c
@@ -20,6 +20,9 @@
 #include "xfs_qm.h"
 #include "xfs_trans_space.h"
 #include "xfs_rtbitmap.h"
+#include "xfs_attr_item.h"
+#include "xfs_log.h"
+#include "xfs_da_format.h"
 
 #define _ALLOC	true
 #define _FREE	false
@@ -422,29 +425,110 @@ xfs_calc_itruncate_reservation_minlogsize(
 	return xfs_calc_itruncate_reservation(mp, true);
 }
 
+static inline unsigned int xfs_calc_pptr_link_overhead(void)
+{
+	return sizeof(struct xfs_attri_log_format) +
+			xlog_calc_iovec_len(sizeof(struct xfs_parent_rec)) +
+			xlog_calc_iovec_len(MAXNAMELEN - 1);
+}
+static inline unsigned int xfs_calc_pptr_unlink_overhead(void)
+{
+	return sizeof(struct xfs_attri_log_format) +
+			xlog_calc_iovec_len(sizeof(struct xfs_parent_rec)) +
+			xlog_calc_iovec_len(MAXNAMELEN - 1);
+}
+static inline unsigned int xfs_calc_pptr_replace_overhead(void)
+{
+	return sizeof(struct xfs_attri_log_format) +
+			xlog_calc_iovec_len(sizeof(struct xfs_parent_rec)) +
+			xlog_calc_iovec_len(MAXNAMELEN - 1) +
+			xlog_calc_iovec_len(sizeof(struct xfs_parent_rec)) +
+			xlog_calc_iovec_len(MAXNAMELEN - 1);
+}
+
 /*
  * In renaming a files we can modify:
  *    the five inodes involved: 5 * inode size
  *    the two directory btrees: 2 * (max depth + v2) * dir block size
  *    the two directory bmap btrees: 2 * max depth * block size
  * And the bmap_finish transaction can free dir and bmap blocks (two sets
- *	of bmap blocks) giving:
+ *	of bmap blocks) giving (t2):
  *    the agf for the ags in which the blocks live: 3 * sector size
  *    the agfl for the ags in which the blocks live: 3 * sector size
  *    the superblock for the free block count: sector size
  *    the allocation btrees: 3 exts * 2 trees * (2 * max depth - 1) * block size
+ * If parent pointers are enabled (t3), then each transaction in the chain
+ *    must be capable of setting or removing the extended attribute
+ *    containing the parent information.  It must also be able to handle
+ *    the three xattr intent items that track the progress of the parent
+ *    pointer update.
  */
 STATIC uint
 xfs_calc_rename_reservation(
 	struct xfs_mount	*mp)
 {
-	return XFS_DQUOT_LOGRES(mp) +
-		max((xfs_calc_inode_res(mp, 5) +
-		     xfs_calc_buf_res(2 * XFS_DIROP_LOG_COUNT(mp),
-				      XFS_FSB_TO_B(mp, 1))),
-		    (xfs_calc_buf_res(7, mp->m_sb.sb_sectsize) +
-		     xfs_calc_buf_res(xfs_allocfree_block_count(mp, 3),
-				      XFS_FSB_TO_B(mp, 1))));
+	unsigned int		overhead = XFS_DQUOT_LOGRES(mp);
+	struct xfs_trans_resv	*resp = M_RES(mp);
+	unsigned int		t1, t2, t3 = 0;
+
+	t1 = xfs_calc_inode_res(mp, 5) +
+	     xfs_calc_buf_res(2 * XFS_DIROP_LOG_COUNT(mp),
+			XFS_FSB_TO_B(mp, 1));
+
+	t2 = xfs_calc_buf_res(7, mp->m_sb.sb_sectsize) +
+	     xfs_calc_buf_res(xfs_allocfree_block_count(mp, 3),
+			XFS_FSB_TO_B(mp, 1));
+
+	if (xfs_has_parent(mp)) {
+		unsigned int	rename_overhead, exchange_overhead;
+
+		t3 = max(resp->tr_attrsetm.tr_logres,
+			 resp->tr_attrrm.tr_logres);
+
+		/*
+		 * For a standard rename, the three xattr intent log items
+		 * are (1) replacing the pptr for the source file; (2)
+		 * removing the pptr on the dest file; and (3) adding a
+		 * pptr for the whiteout file in the src dir.
+		 *
+		 * For an RENAME_EXCHANGE, there are two xattr intent
+		 * items to replace the pptr for both src and dest
+		 * files.  Link counts don't change and there is no
+		 * whiteout.
+		 *
+		 * In the worst case we can end up relogging all log
+		 * intent items to allow the log tail to move ahead, so
+		 * they become overhead added to each transaction in a
+		 * processing chain.
+		 */
+		rename_overhead = xfs_calc_pptr_replace_overhead() +
+				  xfs_calc_pptr_unlink_overhead() +
+				  xfs_calc_pptr_link_overhead();
+		exchange_overhead = 2 * xfs_calc_pptr_replace_overhead();
+
+		overhead += max(rename_overhead, exchange_overhead);
+	}
+
+	return overhead + max3(t1, t2, t3);
+}
+
+static inline unsigned int
+xfs_rename_log_count(
+	struct xfs_mount	*mp,
+	struct xfs_trans_resv	*resp)
+{
+	/* One for the rename, one more for freeing blocks */
+	unsigned int		ret = XFS_RENAME_LOG_COUNT;
+
+	/*
+	 * Pre-reserve enough log reservation to handle the transaction
+	 * rolling needed to remove or add one parent pointer.
+	 */
+	if (xfs_has_parent(mp))
+		ret += max(resp->tr_attrsetm.tr_logcount,
+			   resp->tr_attrrm.tr_logcount);
+
+	return ret;
 }
 
 /*
@@ -461,6 +545,23 @@ xfs_calc_iunlink_remove_reservation(
 	       2 * M_IGEO(mp)->inode_cluster_size;
 }
 
+static inline unsigned int
+xfs_link_log_count(
+	struct xfs_mount	*mp,
+	struct xfs_trans_resv	*resp)
+{
+	unsigned int		ret = XFS_LINK_LOG_COUNT;
+
+	/*
+	 * Pre-reserve enough log reservation to handle the transaction
+	 * rolling needed to add one parent pointer.
+	 */
+	if (xfs_has_parent(mp))
+		ret += resp->tr_attrsetm.tr_logcount;
+
+	return ret;
+}
+
 /*
  * For creating a link to an inode:
  *    the parent directory inode: inode size
@@ -477,14 +578,23 @@ STATIC uint
 xfs_calc_link_reservation(
 	struct xfs_mount	*mp)
 {
-	return XFS_DQUOT_LOGRES(mp) +
-		xfs_calc_iunlink_remove_reservation(mp) +
-		max((xfs_calc_inode_res(mp, 2) +
-		     xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp),
-				      XFS_FSB_TO_B(mp, 1))),
-		    (xfs_calc_buf_res(3, mp->m_sb.sb_sectsize) +
-		     xfs_calc_buf_res(xfs_allocfree_block_count(mp, 1),
-				      XFS_FSB_TO_B(mp, 1))));
+	unsigned int		overhead = XFS_DQUOT_LOGRES(mp);
+	struct xfs_trans_resv	*resp = M_RES(mp);
+	unsigned int		t1, t2, t3 = 0;
+
+	overhead += xfs_calc_iunlink_remove_reservation(mp);
+	t1 = xfs_calc_inode_res(mp, 2) +
+	     xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp), XFS_FSB_TO_B(mp, 1));
+	t2 = xfs_calc_buf_res(3, mp->m_sb.sb_sectsize) +
+	     xfs_calc_buf_res(xfs_allocfree_block_count(mp, 1),
+			      XFS_FSB_TO_B(mp, 1));
+
+	if (xfs_has_parent(mp)) {
+		t3 = resp->tr_attrsetm.tr_logres;
+		overhead += xfs_calc_pptr_link_overhead();
+	}
+
+	return overhead + max3(t1, t2, t3);
 }
 
 /*
@@ -499,6 +609,23 @@ xfs_calc_iunlink_add_reservation(xfs_mount_t *mp)
 			M_IGEO(mp)->inode_cluster_size;
 }
 
+static inline unsigned int
+xfs_remove_log_count(
+	struct xfs_mount	*mp,
+	struct xfs_trans_resv	*resp)
+{
+	unsigned int		ret = XFS_REMOVE_LOG_COUNT;
+
+	/*
+	 * Pre-reserve enough log reservation to handle the transaction
+	 * rolling needed to add one parent pointer.
+	 */
+	if (xfs_has_parent(mp))
+		ret += resp->tr_attrrm.tr_logcount;
+
+	return ret;
+}
+
 /*
  * For removing a directory entry we can modify:
  *    the parent directory inode: inode size
@@ -515,14 +642,24 @@ STATIC uint
 xfs_calc_remove_reservation(
 	struct xfs_mount	*mp)
 {
-	return XFS_DQUOT_LOGRES(mp) +
-		xfs_calc_iunlink_add_reservation(mp) +
-		max((xfs_calc_inode_res(mp, 2) +
-		     xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp),
-				      XFS_FSB_TO_B(mp, 1))),
-		    (xfs_calc_buf_res(4, mp->m_sb.sb_sectsize) +
-		     xfs_calc_buf_res(xfs_allocfree_block_count(mp, 2),
-				      XFS_FSB_TO_B(mp, 1))));
+	unsigned int            overhead = XFS_DQUOT_LOGRES(mp);
+	struct xfs_trans_resv   *resp = M_RES(mp);
+	unsigned int            t1, t2, t3 = 0;
+
+	overhead += xfs_calc_iunlink_add_reservation(mp);
+
+	t1 = xfs_calc_inode_res(mp, 2) +
+	     xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp), XFS_FSB_TO_B(mp, 1));
+	t2 = xfs_calc_buf_res(4, mp->m_sb.sb_sectsize) +
+	     xfs_calc_buf_res(xfs_allocfree_block_count(mp, 2),
+			      XFS_FSB_TO_B(mp, 1));
+
+	if (xfs_has_parent(mp)) {
+		t3 = resp->tr_attrrm.tr_logres;
+		overhead += xfs_calc_pptr_unlink_overhead();
+	}
+
+	return overhead + max3(t1, t2, t3);
 }
 
 /*
@@ -571,12 +708,40 @@ xfs_calc_icreate_resv_alloc(
 		xfs_calc_finobt_res(mp);
 }
 
+static inline unsigned int
+xfs_icreate_log_count(
+	struct xfs_mount	*mp,
+	struct xfs_trans_resv	*resp)
+{
+	unsigned int		ret = XFS_CREATE_LOG_COUNT;
+
+	/*
+	 * Pre-reserve enough log reservation to handle the transaction
+	 * rolling needed to add one parent pointer.
+	 */
+	if (xfs_has_parent(mp))
+		ret += resp->tr_attrsetm.tr_logcount;
+
+	return ret;
+}
+
 STATIC uint
-xfs_calc_icreate_reservation(xfs_mount_t *mp)
+xfs_calc_icreate_reservation(
+	struct xfs_mount	*mp)
 {
-	return XFS_DQUOT_LOGRES(mp) +
-		max(xfs_calc_icreate_resv_alloc(mp),
-		    xfs_calc_create_resv_modify(mp));
+	struct xfs_trans_resv	*resp = M_RES(mp);
+	unsigned int		overhead = XFS_DQUOT_LOGRES(mp);
+	unsigned int		t1, t2, t3 = 0;
+
+	t1 = xfs_calc_icreate_resv_alloc(mp);
+	t2 = xfs_calc_create_resv_modify(mp);
+
+	if (xfs_has_parent(mp)) {
+		t3 = resp->tr_attrsetm.tr_logres;
+		overhead += xfs_calc_pptr_link_overhead();
+	}
+
+	return overhead + max3(t1, t2, t3);
 }
 
 STATIC uint
@@ -589,6 +754,23 @@ xfs_calc_create_tmpfile_reservation(
 	return res + xfs_calc_iunlink_add_reservation(mp);
 }
 
+static inline unsigned int
+xfs_mkdir_log_count(
+	struct xfs_mount	*mp,
+	struct xfs_trans_resv	*resp)
+{
+	unsigned int		ret = XFS_MKDIR_LOG_COUNT;
+
+	/*
+	 * Pre-reserve enough log reservation to handle the transaction
+	 * rolling needed to add one parent pointer.
+	 */
+	if (xfs_has_parent(mp))
+		ret += resp->tr_attrsetm.tr_logcount;
+
+	return ret;
+}
+
 /*
  * Making a new directory is the same as creating a new file.
  */
@@ -599,6 +781,22 @@ xfs_calc_mkdir_reservation(
 	return xfs_calc_icreate_reservation(mp);
 }
 
+static inline unsigned int
+xfs_symlink_log_count(
+	struct xfs_mount	*mp,
+	struct xfs_trans_resv	*resp)
+{
+	unsigned int		ret = XFS_SYMLINK_LOG_COUNT;
+
+	/*
+	 * Pre-reserve enough log reservation to handle the transaction
+	 * rolling needed to add one parent pointer.
+	 */
+	if (xfs_has_parent(mp))
+		ret += resp->tr_attrsetm.tr_logcount;
+
+	return ret;
+}
 
 /*
  * Making a new symplink is the same as creating a new file, but
@@ -911,54 +1109,76 @@ xfs_calc_sb_reservation(
 	return xfs_calc_buf_res(1, mp->m_sb.sb_sectsize);
 }
 
-void
-xfs_trans_resv_calc(
+/*
+ * Namespace reservations.
+ *
+ * These get tricky when parent pointers are enabled as we have attribute
+ * modifications occurring from within these transactions. Rather than confuse
+ * each of these reservation calculations with the conditional attribute
+ * reservations, add them here in a clear and concise manner. This requires that
+ * the attribute reservations have already been calculated.
+ *
+ * Note that we only include the static attribute reservation here; the runtime
+ * reservation will have to be modified by the size of the attributes being
+ * added/removed/modified. See the comments on the attribute reservation
+ * calculations for more details.
+ */
+STATIC void
+xfs_calc_namespace_reservations(
 	struct xfs_mount	*mp,
 	struct xfs_trans_resv	*resp)
 {
-	int			logcount_adj = 0;
-
-	/*
-	 * The following transactions are logged in physical format and
-	 * require a permanent reservation on space.
-	 */
-	resp->tr_write.tr_logres = xfs_calc_write_reservation(mp, false);
-	resp->tr_write.tr_logcount = XFS_WRITE_LOG_COUNT;
-	resp->tr_write.tr_logflags |= XFS_TRANS_PERM_LOG_RES;
-
-	resp->tr_itruncate.tr_logres = xfs_calc_itruncate_reservation(mp, false);
-	resp->tr_itruncate.tr_logcount = XFS_ITRUNCATE_LOG_COUNT;
-	resp->tr_itruncate.tr_logflags |= XFS_TRANS_PERM_LOG_RES;
+	ASSERT(resp->tr_attrsetm.tr_logres > 0);
 
 	resp->tr_rename.tr_logres = xfs_calc_rename_reservation(mp);
-	resp->tr_rename.tr_logcount = XFS_RENAME_LOG_COUNT;
+	resp->tr_rename.tr_logcount = xfs_rename_log_count(mp, resp);
 	resp->tr_rename.tr_logflags |= XFS_TRANS_PERM_LOG_RES;
 
 	resp->tr_link.tr_logres = xfs_calc_link_reservation(mp);
-	resp->tr_link.tr_logcount = XFS_LINK_LOG_COUNT;
+	resp->tr_link.tr_logcount = xfs_link_log_count(mp, resp);
 	resp->tr_link.tr_logflags |= XFS_TRANS_PERM_LOG_RES;
 
 	resp->tr_remove.tr_logres = xfs_calc_remove_reservation(mp);
-	resp->tr_remove.tr_logcount = XFS_REMOVE_LOG_COUNT;
+	resp->tr_remove.tr_logcount = xfs_remove_log_count(mp, resp);
 	resp->tr_remove.tr_logflags |= XFS_TRANS_PERM_LOG_RES;
 
 	resp->tr_symlink.tr_logres = xfs_calc_symlink_reservation(mp);
-	resp->tr_symlink.tr_logcount = XFS_SYMLINK_LOG_COUNT;
+	resp->tr_symlink.tr_logcount = xfs_symlink_log_count(mp, resp);
 	resp->tr_symlink.tr_logflags |= XFS_TRANS_PERM_LOG_RES;
 
 	resp->tr_create.tr_logres = xfs_calc_icreate_reservation(mp);
-	resp->tr_create.tr_logcount = XFS_CREATE_LOG_COUNT;
+	resp->tr_create.tr_logcount = xfs_icreate_log_count(mp, resp);
 	resp->tr_create.tr_logflags |= XFS_TRANS_PERM_LOG_RES;
 
+	resp->tr_mkdir.tr_logres = xfs_calc_mkdir_reservation(mp);
+	resp->tr_mkdir.tr_logcount = xfs_mkdir_log_count(mp, resp);
+	resp->tr_mkdir.tr_logflags |= XFS_TRANS_PERM_LOG_RES;
+}
+
+void
+xfs_trans_resv_calc(
+	struct xfs_mount	*mp,
+	struct xfs_trans_resv	*resp)
+{
+	int			logcount_adj = 0;
+
+	/*
+	 * The following transactions are logged in physical format and
+	 * require a permanent reservation on space.
+	 */
+	resp->tr_write.tr_logres = xfs_calc_write_reservation(mp, false);
+	resp->tr_write.tr_logcount = XFS_WRITE_LOG_COUNT;
+	resp->tr_write.tr_logflags |= XFS_TRANS_PERM_LOG_RES;
+
+	resp->tr_itruncate.tr_logres = xfs_calc_itruncate_reservation(mp, false);
+	resp->tr_itruncate.tr_logcount = XFS_ITRUNCATE_LOG_COUNT;
+	resp->tr_itruncate.tr_logflags |= XFS_TRANS_PERM_LOG_RES;
+
 	resp->tr_create_tmpfile.tr_logres =
 			xfs_calc_create_tmpfile_reservation(mp);
 	resp->tr_create_tmpfile.tr_logcount = XFS_CREATE_TMPFILE_LOG_COUNT;
 	resp->tr_create_tmpfile.tr_logflags |= XFS_TRANS_PERM_LOG_RES;
 
-	resp->tr_mkdir.tr_logres = xfs_calc_mkdir_reservation(mp);
-	resp->tr_mkdir.tr_logcount = XFS_MKDIR_LOG_COUNT;
-	resp->tr_mkdir.tr_logflags |= XFS_TRANS_PERM_LOG_RES;
-
 	resp->tr_ifree.tr_logres = xfs_calc_ifree_reservation(mp);
 	resp->tr_ifree.tr_logcount = XFS_INACTIVE_LOG_COUNT;
 	resp->tr_ifree.tr_logflags |= XFS_TRANS_PERM_LOG_RES;
@@ -988,6 +1208,8 @@ xfs_trans_resv_calc(
 	resp->tr_qm_dqalloc.tr_logcount = XFS_WRITE_LOG_COUNT;
 	resp->tr_qm_dqalloc.tr_logflags |= XFS_TRANS_PERM_LOG_RES;
 
+	xfs_calc_namespace_reservations(mp, resp);
+
 	/*
 	 * The following transactions are logged in logical format with
 	 * a default log count.
diff --git a/fs/xfs/libxfs/xfs_trans_space.c b/fs/xfs/libxfs/xfs_trans_space.c
new file mode 100644
index 0000000000..b9dc3752f7
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_trans_space.c
@@ -0,0 +1,121 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2000,2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_da_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_da_btree.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_trans_space.h"
+
+/* Calculate the disk space required to add a parent pointer. */
+unsigned int
+xfs_parent_calc_space_res(
+	struct xfs_mount	*mp,
+	unsigned int		namelen)
+{
+	/*
+	 * Parent pointers are always the first attr in an attr tree, and never
+	 * larger than a block
+	 */
+	return XFS_DAENTER_SPACE_RES(mp, XFS_ATTR_FORK) +
+	       XFS_NEXTENTADD_SPACE_RES(mp, namelen, XFS_ATTR_FORK);
+}
+
+unsigned int
+xfs_create_space_res(
+	struct xfs_mount	*mp,
+	unsigned int		namelen)
+{
+	unsigned int		ret;
+
+	ret = XFS_IALLOC_SPACE_RES(mp) + XFS_DIRENTER_SPACE_RES(mp, namelen);
+	if (xfs_has_parent(mp))
+		ret += xfs_parent_calc_space_res(mp, namelen);
+
+	return ret;
+}
+
+unsigned int
+xfs_mkdir_space_res(
+	struct xfs_mount	*mp,
+	unsigned int		namelen)
+{
+	return xfs_create_space_res(mp, namelen);
+}
+
+unsigned int
+xfs_link_space_res(
+	struct xfs_mount	*mp,
+	unsigned int		namelen)
+{
+	unsigned int		ret;
+
+	ret = XFS_DIRENTER_SPACE_RES(mp, namelen);
+	if (xfs_has_parent(mp))
+		ret += xfs_parent_calc_space_res(mp, namelen);
+
+	return ret;
+}
+
+unsigned int
+xfs_symlink_space_res(
+	struct xfs_mount	*mp,
+	unsigned int		namelen,
+	unsigned int		fsblocks)
+{
+	unsigned int		ret;
+
+	ret = XFS_IALLOC_SPACE_RES(mp) + XFS_DIRENTER_SPACE_RES(mp, namelen) +
+			fsblocks;
+
+	if (xfs_has_parent(mp))
+		ret += xfs_parent_calc_space_res(mp, namelen);
+
+	return ret;
+}
+
+unsigned int
+xfs_remove_space_res(
+	struct xfs_mount	*mp,
+	unsigned int		namelen)
+{
+	unsigned int		ret = XFS_DIRREMOVE_SPACE_RES(mp);
+
+	if (xfs_has_parent(mp))
+		ret += xfs_parent_calc_space_res(mp, namelen);
+
+	return ret;
+}
+
+unsigned int
+xfs_rename_space_res(
+	struct xfs_mount	*mp,
+	unsigned int		src_namelen,
+	bool			target_exists,
+	unsigned int		target_namelen,
+	bool			has_whiteout)
+{
+	unsigned int		ret;
+
+	ret = XFS_DIRREMOVE_SPACE_RES(mp) +
+			XFS_DIRENTER_SPACE_RES(mp, target_namelen);
+
+	if (xfs_has_parent(mp)) {
+		if (has_whiteout)
+			ret += xfs_parent_calc_space_res(mp, src_namelen);
+		ret += 2 * xfs_parent_calc_space_res(mp, target_namelen);
+	}
+
+	if (target_exists)
+		ret += xfs_parent_calc_space_res(mp, target_namelen);
+
+	return ret;
+}
diff --git a/fs/xfs/libxfs/xfs_trans_space.h b/fs/xfs/libxfs/xfs_trans_space.h
index 87b31c69a7..1155ff2d37 100644
--- a/fs/xfs/libxfs/xfs_trans_space.h
+++ b/fs/xfs/libxfs/xfs_trans_space.h
@@ -10,6 +10,10 @@
  * Components of space reservations.
  */
 
+/* Worst case number of bmaps that can be held in a block. */
+#define XFS_MAX_CONTIG_BMAPS_PER_BLOCK(mp)    \
+		(((mp)->m_bmap_dmxr[0]) - ((mp)->m_bmap_dmnr[0]))
+
 /* Worst case number of rmaps that can be held in a block. */
 #define XFS_MAX_CONTIG_RMAPS_PER_BLOCK(mp)    \
 		(((mp)->m_rmap_mxr[0]) - ((mp)->m_rmap_mnr[0]))
@@ -76,31 +80,32 @@
 /* This macro is not used - see inline code in xfs_attr_set */
 #define	XFS_ATTRSET_SPACE_RES(mp, v)	\
 	(XFS_DAENTER_SPACE_RES(mp, XFS_ATTR_FORK) + XFS_B_TO_FSB(mp, v))
-#define	XFS_CREATE_SPACE_RES(mp,nl)	\
-	(XFS_IALLOC_SPACE_RES(mp) + XFS_DIRENTER_SPACE_RES(mp,nl))
 #define	XFS_DIOSTRAT_SPACE_RES(mp, v)	\
 	(XFS_EXTENTADD_SPACE_RES(mp, XFS_DATA_FORK) + (v))
 #define	XFS_GROWFS_SPACE_RES(mp)	\
 	(2 * (mp)->m_alloc_maxlevels)
 #define	XFS_GROWFSRT_SPACE_RES(mp,b)	\
 	((b) + XFS_EXTENTADD_SPACE_RES(mp, XFS_DATA_FORK))
-#define	XFS_LINK_SPACE_RES(mp,nl)	\
-	XFS_DIRENTER_SPACE_RES(mp,nl)
-#define	XFS_MKDIR_SPACE_RES(mp,nl)	\
-	(XFS_IALLOC_SPACE_RES(mp) + XFS_DIRENTER_SPACE_RES(mp,nl))
 #define	XFS_QM_DQALLOC_SPACE_RES(mp)	\
 	(XFS_EXTENTADD_SPACE_RES(mp, XFS_DATA_FORK) + \
 	 XFS_DQUOT_CLUSTER_SIZE_FSB)
 #define	XFS_QM_QINOCREATE_SPACE_RES(mp)	\
 	XFS_IALLOC_SPACE_RES(mp)
-#define	XFS_REMOVE_SPACE_RES(mp)	\
-	XFS_DIRREMOVE_SPACE_RES(mp)
-#define	XFS_RENAME_SPACE_RES(mp,nl)	\
-	(XFS_DIRREMOVE_SPACE_RES(mp) + XFS_DIRENTER_SPACE_RES(mp,nl))
-#define	XFS_SYMLINK_SPACE_RES(mp,nl,b)	\
-	(XFS_IALLOC_SPACE_RES(mp) + XFS_DIRENTER_SPACE_RES(mp,nl) + (b))
 #define XFS_IFREE_SPACE_RES(mp)		\
 	(xfs_has_finobt(mp) ? M_IGEO(mp)->inobt_maxlevels : 0)
 
+unsigned int xfs_parent_calc_space_res(struct xfs_mount *mp,
+		unsigned int namelen);
+
+unsigned int xfs_create_space_res(struct xfs_mount *mp, unsigned int namelen);
+unsigned int xfs_mkdir_space_res(struct xfs_mount *mp, unsigned int namelen);
+unsigned int xfs_link_space_res(struct xfs_mount *mp, unsigned int namelen);
+unsigned int xfs_symlink_space_res(struct xfs_mount *mp, unsigned int namelen,
+		unsigned int fsblocks);
+unsigned int xfs_remove_space_res(struct xfs_mount *mp, unsigned int namelen);
+
+unsigned int xfs_rename_space_res(struct xfs_mount *mp,
+		unsigned int src_namelen, bool target_exists,
+		unsigned int target_namelen, bool has_whiteout);
 
 #endif	/* __XFS_TRANS_SPACE_H__ */
diff --git a/fs/xfs/libxfs/xfs_types.h b/fs/xfs/libxfs/xfs_types.h
index 62e02d5380..76eb9e3288 100644
--- a/fs/xfs/libxfs/xfs_types.h
+++ b/fs/xfs/libxfs/xfs_types.h
@@ -80,11 +80,13 @@ typedef void *		xfs_failaddr_t;
 /*
  * Inode fork identifiers.
  */
-#define	XFS_DATA_FORK	0
-#define	XFS_ATTR_FORK	1
-#define	XFS_COW_FORK	2
+#define XFS_STAGING_FORK	(-1)	/* fake fork for staging a btree */
+#define	XFS_DATA_FORK		(0)
+#define	XFS_ATTR_FORK		(1)
+#define	XFS_COW_FORK		(2)
 
 #define XFS_WHICHFORK_STRINGS \
+	{ XFS_STAGING_FORK, 	"staging" }, \
 	{ XFS_DATA_FORK, 	"data" }, \
 	{ XFS_ATTR_FORK,	"attr" }, \
 	{ XFS_COW_FORK,		"cow" }
@@ -114,24 +116,6 @@ typedef enum {
 	{ XFS_LOOKUP_LEi,	"le" }, \
 	{ XFS_LOOKUP_GEi,	"ge" }
 
-/*
- * This enum is used in string mapping in xfs_trace.h and scrub/trace.h;
- * please keep the TRACE_DEFINE_ENUMs for it up to date.
- */
-typedef enum {
-	XFS_BTNUM_BNOi, XFS_BTNUM_CNTi, XFS_BTNUM_RMAPi, XFS_BTNUM_BMAPi,
-	XFS_BTNUM_INOi, XFS_BTNUM_FINOi, XFS_BTNUM_REFCi, XFS_BTNUM_MAX
-} xfs_btnum_t;
-
-#define XFS_BTNUM_STRINGS \
-	{ XFS_BTNUM_BNOi,	"bnobt" }, \
-	{ XFS_BTNUM_CNTi,	"cntbt" }, \
-	{ XFS_BTNUM_RMAPi,	"rmapbt" }, \
-	{ XFS_BTNUM_BMAPi,	"bmbt" }, \
-	{ XFS_BTNUM_INOi,	"inobt" }, \
-	{ XFS_BTNUM_FINOi,	"finobt" }, \
-	{ XFS_BTNUM_REFCi,	"refcbt" }
-
 struct xfs_name {
 	const unsigned char	*name;
 	int			len;
diff --git a/fs/xfs/mrlock.h b/fs/xfs/mrlock.h
deleted file mode 100644
index 79155eec34..0000000000
--- a/fs/xfs/mrlock.h
+++ /dev/null
@@ -1,78 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0
-/*
- * Copyright (c) 2000-2006 Silicon Graphics, Inc.
- * All Rights Reserved.
- */
-#ifndef __XFS_SUPPORT_MRLOCK_H__
-#define __XFS_SUPPORT_MRLOCK_H__
-
-#include <linux/rwsem.h>
-
-typedef struct {
-	struct rw_semaphore	mr_lock;
-#if defined(DEBUG) || defined(XFS_WARN)
-	int			mr_writer;
-#endif
-} mrlock_t;
-
-#if defined(DEBUG) || defined(XFS_WARN)
-#define mrinit(mrp, name)	\
-	do { (mrp)->mr_writer = 0; init_rwsem(&(mrp)->mr_lock); } while (0)
-#else
-#define mrinit(mrp, name)	\
-	do { init_rwsem(&(mrp)->mr_lock); } while (0)
-#endif
-
-#define mrlock_init(mrp, t,n,s)	mrinit(mrp, n)
-#define mrfree(mrp)		do { } while (0)
-
-static inline void mraccess_nested(mrlock_t *mrp, int subclass)
-{
-	down_read_nested(&mrp->mr_lock, subclass);
-}
-
-static inline void mrupdate_nested(mrlock_t *mrp, int subclass)
-{
-	down_write_nested(&mrp->mr_lock, subclass);
-#if defined(DEBUG) || defined(XFS_WARN)
-	mrp->mr_writer = 1;
-#endif
-}
-
-static inline int mrtryaccess(mrlock_t *mrp)
-{
-	return down_read_trylock(&mrp->mr_lock);
-}
-
-static inline int mrtryupdate(mrlock_t *mrp)
-{
-	if (!down_write_trylock(&mrp->mr_lock))
-		return 0;
-#if defined(DEBUG) || defined(XFS_WARN)
-	mrp->mr_writer = 1;
-#endif
-	return 1;
-}
-
-static inline void mrunlock_excl(mrlock_t *mrp)
-{
-#if defined(DEBUG) || defined(XFS_WARN)
-	mrp->mr_writer = 0;
-#endif
-	up_write(&mrp->mr_lock);
-}
-
-static inline void mrunlock_shared(mrlock_t *mrp)
-{
-	up_read(&mrp->mr_lock);
-}
-
-static inline void mrdemote(mrlock_t *mrp)
-{
-#if defined(DEBUG) || defined(XFS_WARN)
-	mrp->mr_writer = 0;
-#endif
-	downgrade_write(&mrp->mr_lock);
-}
-
-#endif /* __XFS_SUPPORT_MRLOCK_H__ */
diff --git a/fs/xfs/scrub/agb_bitmap.h b/fs/xfs/scrub/agb_bitmap.h
index ed08f76ff4..e488e1f4f6 100644
--- a/fs/xfs/scrub/agb_bitmap.h
+++ b/fs/xfs/scrub/agb_bitmap.h
@@ -65,4 +65,9 @@ int xagb_bitmap_set_btblocks(struct xagb_bitmap *bitmap,
 int xagb_bitmap_set_btcur_path(struct xagb_bitmap *bitmap,
 		struct xfs_btree_cur *cur);
 
+static inline uint32_t xagb_bitmap_count_set_regions(struct xagb_bitmap *b)
+{
+	return xbitmap32_count_set_regions(&b->agbitmap);
+}
+
 #endif	/* __XFS_SCRUB_AGB_BITMAP_H__ */
diff --git a/fs/xfs/scrub/agheader.c b/fs/xfs/scrub/agheader.c
index 6c6e5eba42..f8e5b67128 100644
--- a/fs/xfs/scrub/agheader.c
+++ b/fs/xfs/scrub/agheader.c
@@ -15,6 +15,7 @@
 #include "xfs_ialloc.h"
 #include "xfs_rmap.h"
 #include "xfs_ag.h"
+#include "xfs_inode.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 
@@ -165,8 +166,7 @@ xchk_superblock(
 		xchk_block_set_corrupt(sc, bp);
 
 	/* Check sb_versionnum bits that are set at mkfs time. */
-	vernum_mask = cpu_to_be16(~XFS_SB_VERSION_OKBITS |
-				  XFS_SB_VERSION_NUMBITS |
+	vernum_mask = cpu_to_be16(XFS_SB_VERSION_NUMBITS |
 				  XFS_SB_VERSION_ALIGNBIT |
 				  XFS_SB_VERSION_DALIGNBIT |
 				  XFS_SB_VERSION_SHAREDBIT |
@@ -556,28 +556,28 @@ xchk_agf(
 		xchk_block_set_corrupt(sc, sc->sa.agf_bp);
 
 	/* Check the AGF btree roots and levels */
-	agbno = be32_to_cpu(agf->agf_roots[XFS_BTNUM_BNO]);
+	agbno = be32_to_cpu(agf->agf_bno_root);
 	if (!xfs_verify_agbno(pag, agbno))
 		xchk_block_set_corrupt(sc, sc->sa.agf_bp);
 
-	agbno = be32_to_cpu(agf->agf_roots[XFS_BTNUM_CNT]);
+	agbno = be32_to_cpu(agf->agf_cnt_root);
 	if (!xfs_verify_agbno(pag, agbno))
 		xchk_block_set_corrupt(sc, sc->sa.agf_bp);
 
-	level = be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]);
+	level = be32_to_cpu(agf->agf_bno_level);
 	if (level <= 0 || level > mp->m_alloc_maxlevels)
 		xchk_block_set_corrupt(sc, sc->sa.agf_bp);
 
-	level = be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]);
+	level = be32_to_cpu(agf->agf_cnt_level);
 	if (level <= 0 || level > mp->m_alloc_maxlevels)
 		xchk_block_set_corrupt(sc, sc->sa.agf_bp);
 
 	if (xfs_has_rmapbt(mp)) {
-		agbno = be32_to_cpu(agf->agf_roots[XFS_BTNUM_RMAP]);
+		agbno = be32_to_cpu(agf->agf_rmap_root);
 		if (!xfs_verify_agbno(pag, agbno))
 			xchk_block_set_corrupt(sc, sc->sa.agf_bp);
 
-		level = be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]);
+		level = be32_to_cpu(agf->agf_rmap_level);
 		if (level <= 0 || level > mp->m_rmap_maxlevels)
 			xchk_block_set_corrupt(sc, sc->sa.agf_bp);
 	}
@@ -865,6 +865,43 @@ xchk_agi_xref(
 	/* scrub teardown will take care of sc->sa for us */
 }
 
+/*
+ * Check the unlinked buckets for links to bad inodes.  We hold the AGI, so
+ * there cannot be any threads updating unlinked list pointers in this AG.
+ */
+STATIC void
+xchk_iunlink(
+	struct xfs_scrub	*sc,
+	struct xfs_agi		*agi)
+{
+	unsigned int		i;
+	struct xfs_inode	*ip;
+
+	for (i = 0; i < XFS_AGI_UNLINKED_BUCKETS; i++) {
+		xfs_agino_t	agino = be32_to_cpu(agi->agi_unlinked[i]);
+
+		while (agino != NULLAGINO) {
+			if (agino % XFS_AGI_UNLINKED_BUCKETS != i) {
+				xchk_block_set_corrupt(sc, sc->sa.agi_bp);
+				return;
+			}
+
+			ip = xfs_iunlink_lookup(sc->sa.pag, agino);
+			if (!ip) {
+				xchk_block_set_corrupt(sc, sc->sa.agi_bp);
+				return;
+			}
+
+			if (!xfs_inode_on_unlinked_list(ip)) {
+				xchk_block_set_corrupt(sc, sc->sa.agi_bp);
+				return;
+			}
+
+			agino = ip->i_next_unlinked;
+		}
+	}
+}
+
 /* Scrub the AGI. */
 int
 xchk_agi(
@@ -949,6 +986,8 @@ xchk_agi(
 	if (pag->pagi_freecount != be32_to_cpu(agi->agi_freecount))
 		xchk_block_set_corrupt(sc, sc->sa.agi_bp);
 
+	xchk_iunlink(sc, agi);
+
 	xchk_agi_xref(sc);
 out:
 	return error;
diff --git a/fs/xfs/scrub/agheader_repair.c b/fs/xfs/scrub/agheader_repair.c
index 26bd1ff68f..0dbc484b18 100644
--- a/fs/xfs/scrub/agheader_repair.c
+++ b/fs/xfs/scrub/agheader_repair.c
@@ -21,13 +21,18 @@
 #include "xfs_rmap_btree.h"
 #include "xfs_refcount_btree.h"
 #include "xfs_ag.h"
+#include "xfs_inode.h"
+#include "xfs_iunlink_item.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
 #include "scrub/repair.h"
 #include "scrub/bitmap.h"
 #include "scrub/agb_bitmap.h"
+#include "scrub/agino_bitmap.h"
 #include "scrub/reap.h"
+#include "scrub/xfile.h"
+#include "scrub/xfarray.h"
 
 /* Superblock */
 
@@ -174,8 +179,7 @@ xrep_agf_find_btrees(
 	 * We relied on the rmapbt to reconstruct the AGF.  If we get a
 	 * different root then something's seriously wrong.
 	 */
-	if (fab[XREP_AGF_RMAPBT].root !=
-	    be32_to_cpu(old_agf->agf_roots[XFS_BTNUM_RMAPi]))
+	if (fab[XREP_AGF_RMAPBT].root != be32_to_cpu(old_agf->agf_rmap_root))
 		return -EFSCORRUPTED;
 
 	/* We must find the refcountbt root if that feature is enabled. */
@@ -224,20 +228,14 @@ xrep_agf_set_roots(
 	struct xfs_agf			*agf,
 	struct xrep_find_ag_btree	*fab)
 {
-	agf->agf_roots[XFS_BTNUM_BNOi] =
-			cpu_to_be32(fab[XREP_AGF_BNOBT].root);
-	agf->agf_levels[XFS_BTNUM_BNOi] =
-			cpu_to_be32(fab[XREP_AGF_BNOBT].height);
+	agf->agf_bno_root = cpu_to_be32(fab[XREP_AGF_BNOBT].root);
+	agf->agf_bno_level = cpu_to_be32(fab[XREP_AGF_BNOBT].height);
 
-	agf->agf_roots[XFS_BTNUM_CNTi] =
-			cpu_to_be32(fab[XREP_AGF_CNTBT].root);
-	agf->agf_levels[XFS_BTNUM_CNTi] =
-			cpu_to_be32(fab[XREP_AGF_CNTBT].height);
+	agf->agf_cnt_root = cpu_to_be32(fab[XREP_AGF_CNTBT].root);
+	agf->agf_cnt_level = cpu_to_be32(fab[XREP_AGF_CNTBT].height);
 
-	agf->agf_roots[XFS_BTNUM_RMAPi] =
-			cpu_to_be32(fab[XREP_AGF_RMAPBT].root);
-	agf->agf_levels[XFS_BTNUM_RMAPi] =
-			cpu_to_be32(fab[XREP_AGF_RMAPBT].height);
+	agf->agf_rmap_root = cpu_to_be32(fab[XREP_AGF_RMAPBT].root);
+	agf->agf_rmap_level = cpu_to_be32(fab[XREP_AGF_RMAPBT].height);
 
 	if (xfs_has_reflink(sc->mp)) {
 		agf->agf_refcount_root =
@@ -262,8 +260,7 @@ xrep_agf_calc_from_btrees(
 	int			error;
 
 	/* Update the AGF counters from the bnobt. */
-	cur = xfs_allocbt_init_cursor(mp, sc->tp, agf_bp,
-			sc->sa.pag, XFS_BTNUM_BNO);
+	cur = xfs_bnobt_init_cursor(mp, sc->tp, agf_bp, sc->sa.pag);
 	error = xfs_alloc_query_all(cur, xrep_agf_walk_allocbt, &raa);
 	if (error)
 		goto err;
@@ -276,8 +273,7 @@ xrep_agf_calc_from_btrees(
 	agf->agf_longest = cpu_to_be32(raa.longest);
 
 	/* Update the AGF counters from the cntbt. */
-	cur = xfs_allocbt_init_cursor(mp, sc->tp, agf_bp,
-			sc->sa.pag, XFS_BTNUM_CNT);
+	cur = xfs_cntbt_init_cursor(mp, sc->tp, agf_bp, sc->sa.pag);
 	error = xfs_btree_count_blocks(cur, &blocks);
 	if (error)
 		goto err;
@@ -333,12 +329,9 @@ xrep_agf_commit_new(
 	pag->pagf_btreeblks = be32_to_cpu(agf->agf_btreeblks);
 	pag->pagf_freeblks = be32_to_cpu(agf->agf_freeblks);
 	pag->pagf_longest = be32_to_cpu(agf->agf_longest);
-	pag->pagf_levels[XFS_BTNUM_BNOi] =
-			be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNOi]);
-	pag->pagf_levels[XFS_BTNUM_CNTi] =
-			be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNTi]);
-	pag->pagf_levels[XFS_BTNUM_RMAPi] =
-			be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAPi]);
+	pag->pagf_bno_level = be32_to_cpu(agf->agf_bno_level);
+	pag->pagf_cnt_level = be32_to_cpu(agf->agf_cnt_level);
+	pag->pagf_rmap_level = be32_to_cpu(agf->agf_rmap_level);
 	pag->pagf_refcount_level = be32_to_cpu(agf->agf_refcount_level);
 	set_bit(XFS_AGSTATE_AGF_INIT, &pag->pag_opstate);
 
@@ -559,16 +552,14 @@ xrep_agfl_collect_blocks(
 		goto out_bmp;
 
 	/* Find all blocks currently being used by the bnobt. */
-	cur = xfs_allocbt_init_cursor(mp, sc->tp, agf_bp,
-			sc->sa.pag, XFS_BTNUM_BNO);
+	cur = xfs_bnobt_init_cursor(mp, sc->tp, agf_bp, sc->sa.pag);
 	error = xagb_bitmap_set_btblocks(&ra.agmetablocks, cur);
 	xfs_btree_del_cursor(cur, error);
 	if (error)
 		goto out_bmp;
 
 	/* Find all blocks currently being used by the cntbt. */
-	cur = xfs_allocbt_init_cursor(mp, sc->tp, agf_bp,
-			sc->sa.pag, XFS_BTNUM_CNT);
+	cur = xfs_cntbt_init_cursor(mp, sc->tp, agf_bp, sc->sa.pag);
 	error = xagb_bitmap_set_btblocks(&ra.agmetablocks, cur);
 	xfs_btree_del_cursor(cur, error);
 	if (error)
@@ -810,15 +801,57 @@ enum {
 	XREP_AGI_MAX
 };
 
+#define XREP_AGI_LOOKUP_BATCH		32
+
+struct xrep_agi {
+	struct xfs_scrub		*sc;
+
+	/* AGI buffer, tracked separately */
+	struct xfs_buf			*agi_bp;
+
+	/* context for finding btree roots */
+	struct xrep_find_ag_btree	fab[XREP_AGI_MAX];
+
+	/* old AGI contents in case we have to revert */
+	struct xfs_agi			old_agi;
+
+	/* bitmap of which inodes are unlinked */
+	struct xagino_bitmap		iunlink_bmp;
+
+	/* heads of the unlinked inode bucket lists */
+	xfs_agino_t			iunlink_heads[XFS_AGI_UNLINKED_BUCKETS];
+
+	/* scratchpad for batched lookups of the radix tree */
+	struct xfs_inode		*lookup_batch[XREP_AGI_LOOKUP_BATCH];
+
+	/* Map of ino -> next_ino for unlinked inode processing. */
+	struct xfarray			*iunlink_next;
+
+	/* Map of ino -> prev_ino for unlinked inode processing. */
+	struct xfarray			*iunlink_prev;
+};
+
+static void
+xrep_agi_buf_cleanup(
+	void		*buf)
+{
+	struct xrep_agi	*ragi = buf;
+
+	xfarray_destroy(ragi->iunlink_prev);
+	xfarray_destroy(ragi->iunlink_next);
+	xagino_bitmap_destroy(&ragi->iunlink_bmp);
+}
+
 /*
  * Given the inode btree roots described by *fab, find the roots, check them
  * for sanity, and pass the root data back out via *fab.
  */
 STATIC int
 xrep_agi_find_btrees(
-	struct xfs_scrub		*sc,
-	struct xrep_find_ag_btree	*fab)
+	struct xrep_agi			*ragi)
 {
+	struct xfs_scrub		*sc = ragi->sc;
+	struct xrep_find_ag_btree	*fab = ragi->fab;
 	struct xfs_buf			*agf_bp;
 	struct xfs_mount		*mp = sc->mp;
 	int				error;
@@ -851,10 +884,11 @@ xrep_agi_find_btrees(
  */
 STATIC void
 xrep_agi_init_header(
-	struct xfs_scrub	*sc,
-	struct xfs_buf		*agi_bp,
-	struct xfs_agi		*old_agi)
+	struct xrep_agi		*ragi)
 {
+	struct xfs_scrub	*sc = ragi->sc;
+	struct xfs_buf		*agi_bp = ragi->agi_bp;
+	struct xfs_agi		*old_agi = &ragi->old_agi;
 	struct xfs_agi		*agi = agi_bp->b_addr;
 	struct xfs_perag	*pag = sc->sa.pag;
 	struct xfs_mount	*mp = sc->mp;
@@ -870,10 +904,6 @@ xrep_agi_init_header(
 	if (xfs_has_crc(mp))
 		uuid_copy(&agi->agi_uuid, &mp->m_sb.sb_meta_uuid);
 
-	/* We don't know how to fix the unlinked list yet. */
-	memcpy(&agi->agi_unlinked, &old_agi->agi_unlinked,
-			sizeof(agi->agi_unlinked));
-
 	/* Mark the incore AGF data stale until we're done fixing things. */
 	ASSERT(xfs_perag_initialised_agi(pag));
 	clear_bit(XFS_AGSTATE_AGI_INIT, &pag->pag_opstate);
@@ -882,10 +912,12 @@ xrep_agi_init_header(
 /* Set btree root information in an AGI. */
 STATIC void
 xrep_agi_set_roots(
-	struct xfs_scrub		*sc,
-	struct xfs_agi			*agi,
-	struct xrep_find_ag_btree	*fab)
+	struct xrep_agi			*ragi)
 {
+	struct xfs_scrub		*sc = ragi->sc;
+	struct xfs_agi			*agi = ragi->agi_bp->b_addr;
+	struct xrep_find_ag_btree	*fab = ragi->fab;
+
 	agi->agi_root = cpu_to_be32(fab[XREP_AGI_INOBT].root);
 	agi->agi_level = cpu_to_be32(fab[XREP_AGI_INOBT].height);
 
@@ -898,9 +930,10 @@ xrep_agi_set_roots(
 /* Update the AGI counters. */
 STATIC int
 xrep_agi_calc_from_btrees(
-	struct xfs_scrub	*sc,
-	struct xfs_buf		*agi_bp)
+	struct xrep_agi		*ragi)
 {
+	struct xfs_scrub	*sc = ragi->sc;
+	struct xfs_buf		*agi_bp = ragi->agi_bp;
 	struct xfs_btree_cur	*cur;
 	struct xfs_agi		*agi = agi_bp->b_addr;
 	struct xfs_mount	*mp = sc->mp;
@@ -908,7 +941,7 @@ xrep_agi_calc_from_btrees(
 	xfs_agino_t		freecount;
 	int			error;
 
-	cur = xfs_inobt_init_cursor(sc->sa.pag, sc->tp, agi_bp, XFS_BTNUM_INO);
+	cur = xfs_inobt_init_cursor(sc->sa.pag, sc->tp, agi_bp);
 	error = xfs_ialloc_count_inodes(cur, &count, &freecount);
 	if (error)
 		goto err;
@@ -928,8 +961,7 @@ xrep_agi_calc_from_btrees(
 	if (xfs_has_finobt(mp) && xfs_has_inobtcounts(mp)) {
 		xfs_agblock_t	blocks;
 
-		cur = xfs_inobt_init_cursor(sc->sa.pag, sc->tp, agi_bp,
-				XFS_BTNUM_FINO);
+		cur = xfs_finobt_init_cursor(sc->sa.pag, sc->tp, agi_bp);
 		error = xfs_btree_count_blocks(cur, &blocks);
 		if (error)
 			goto err;
@@ -943,12 +975,721 @@ err:
 	return error;
 }
 
+/*
+ * Record a forwards unlinked chain pointer from agino -> next_agino in our
+ * staging information.
+ */
+static inline int
+xrep_iunlink_store_next(
+	struct xrep_agi		*ragi,
+	xfs_agino_t		agino,
+	xfs_agino_t		next_agino)
+{
+	ASSERT(next_agino != 0);
+
+	return xfarray_store(ragi->iunlink_next, agino, &next_agino);
+}
+
+/*
+ * Record a backwards unlinked chain pointer from prev_ino <- agino in our
+ * staging information.
+ */
+static inline int
+xrep_iunlink_store_prev(
+	struct xrep_agi		*ragi,
+	xfs_agino_t		agino,
+	xfs_agino_t		prev_agino)
+{
+	ASSERT(prev_agino != 0);
+
+	return xfarray_store(ragi->iunlink_prev, agino, &prev_agino);
+}
+
+/*
+ * Given an @agino, look up the next inode in the iunlink bucket.  Returns
+ * NULLAGINO if we're at the end of the chain, 0 if @agino is not in memory
+ * like it should be, or a per-AG inode number.
+ */
+static inline xfs_agino_t
+xrep_iunlink_next(
+	struct xfs_scrub	*sc,
+	xfs_agino_t		agino)
+{
+	struct xfs_inode	*ip;
+
+	ip = xfs_iunlink_lookup(sc->sa.pag, agino);
+	if (!ip)
+		return 0;
+
+	return ip->i_next_unlinked;
+}
+
+/*
+ * Load the inode @agino into memory, set its i_prev_unlinked, and drop the
+ * inode so it can be inactivated.  Returns NULLAGINO if we're at the end of
+ * the chain or if we should stop walking the chain due to corruption; or a
+ * per-AG inode number.
+ */
+STATIC xfs_agino_t
+xrep_iunlink_reload_next(
+	struct xrep_agi		*ragi,
+	xfs_agino_t		prev_agino,
+	xfs_agino_t		agino)
+{
+	struct xfs_scrub	*sc = ragi->sc;
+	struct xfs_inode	*ip;
+	xfs_ino_t		ino;
+	xfs_agino_t		ret = NULLAGINO;
+	int			error;
+
+	ino = XFS_AGINO_TO_INO(sc->mp, sc->sa.pag->pag_agno, agino);
+	error = xchk_iget(ragi->sc, ino, &ip);
+	if (error)
+		return ret;
+
+	trace_xrep_iunlink_reload_next(ip, prev_agino);
+
+	/* If this is a linked inode, stop processing the chain. */
+	if (VFS_I(ip)->i_nlink != 0) {
+		xrep_iunlink_store_next(ragi, agino, NULLAGINO);
+		goto rele;
+	}
+
+	ip->i_prev_unlinked = prev_agino;
+	ret = ip->i_next_unlinked;
+
+	/*
+	 * Drop the inode reference that we just took.  We hold the AGI, so
+	 * this inode cannot move off the unlinked list and hence cannot be
+	 * reclaimed.
+	 */
+rele:
+	xchk_irele(sc, ip);
+	return ret;
+}
+
+/*
+ * Walk an AGI unlinked bucket's list to load incore any unlinked inodes that
+ * still existed at mount time.  This can happen if iunlink processing fails
+ * during log recovery.
+ */
+STATIC int
+xrep_iunlink_walk_ondisk_bucket(
+	struct xrep_agi		*ragi,
+	unsigned int		bucket)
+{
+	struct xfs_scrub	*sc = ragi->sc;
+	struct xfs_agi		*agi = sc->sa.agi_bp->b_addr;
+	xfs_agino_t		prev_agino = NULLAGINO;
+	xfs_agino_t		next_agino;
+	int			error = 0;
+
+	next_agino = be32_to_cpu(agi->agi_unlinked[bucket]);
+	while (next_agino != NULLAGINO) {
+		xfs_agino_t	agino = next_agino;
+
+		if (xchk_should_terminate(ragi->sc, &error))
+			return error;
+
+		trace_xrep_iunlink_walk_ondisk_bucket(sc->sa.pag, bucket,
+				prev_agino, agino);
+
+		if (bucket != agino % XFS_AGI_UNLINKED_BUCKETS)
+			break;
+
+		next_agino = xrep_iunlink_next(sc, agino);
+		if (!next_agino)
+			next_agino = xrep_iunlink_reload_next(ragi, prev_agino,
+					agino);
+
+		prev_agino = agino;
+	}
+
+	return 0;
+}
+
+/* Decide if this is an unlinked inode in this AG. */
+STATIC bool
+xrep_iunlink_igrab(
+	struct xfs_perag	*pag,
+	struct xfs_inode	*ip)
+{
+	struct xfs_mount	*mp = pag->pag_mount;
+
+	if (XFS_INO_TO_AGNO(mp, ip->i_ino) != pag->pag_agno)
+		return false;
+
+	if (!xfs_inode_on_unlinked_list(ip))
+		return false;
+
+	return true;
+}
+
+/*
+ * Mark the given inode in the lookup batch in our unlinked inode bitmap, and
+ * remember if this inode is the start of the unlinked chain.
+ */
+STATIC int
+xrep_iunlink_visit(
+	struct xrep_agi		*ragi,
+	unsigned int		batch_idx)
+{
+	struct xfs_mount	*mp = ragi->sc->mp;
+	struct xfs_inode	*ip = ragi->lookup_batch[batch_idx];
+	xfs_agino_t		agino;
+	unsigned int		bucket;
+	int			error;
+
+	ASSERT(XFS_INO_TO_AGNO(mp, ip->i_ino) == ragi->sc->sa.pag->pag_agno);
+	ASSERT(xfs_inode_on_unlinked_list(ip));
+
+	agino = XFS_INO_TO_AGINO(mp, ip->i_ino);
+	bucket = agino % XFS_AGI_UNLINKED_BUCKETS;
+
+	trace_xrep_iunlink_visit(ragi->sc->sa.pag, bucket,
+			ragi->iunlink_heads[bucket], ip);
+
+	error = xagino_bitmap_set(&ragi->iunlink_bmp, agino, 1);
+	if (error)
+		return error;
+
+	if (ip->i_prev_unlinked == NULLAGINO) {
+		if (ragi->iunlink_heads[bucket] == NULLAGINO)
+			ragi->iunlink_heads[bucket] = agino;
+	}
+
+	return 0;
+}
+
+/*
+ * Find all incore unlinked inodes so that we can rebuild the unlinked buckets.
+ * We hold the AGI so there should not be any modifications to the unlinked
+ * list.
+ */
+STATIC int
+xrep_iunlink_mark_incore(
+	struct xrep_agi		*ragi)
+{
+	struct xfs_perag	*pag = ragi->sc->sa.pag;
+	struct xfs_mount	*mp = pag->pag_mount;
+	uint32_t		first_index = 0;
+	bool			done = false;
+	unsigned int		nr_found = 0;
+
+	do {
+		unsigned int	i;
+		int		error = 0;
+
+		if (xchk_should_terminate(ragi->sc, &error))
+			return error;
+
+		rcu_read_lock();
+
+		nr_found = radix_tree_gang_lookup(&pag->pag_ici_root,
+				(void **)&ragi->lookup_batch, first_index,
+				XREP_AGI_LOOKUP_BATCH);
+		if (!nr_found) {
+			rcu_read_unlock();
+			return 0;
+		}
+
+		for (i = 0; i < nr_found; i++) {
+			struct xfs_inode *ip = ragi->lookup_batch[i];
+
+			if (done || !xrep_iunlink_igrab(pag, ip))
+				ragi->lookup_batch[i] = NULL;
+
+			/*
+			 * Update the index for the next lookup. Catch
+			 * overflows into the next AG range which can occur if
+			 * we have inodes in the last block of the AG and we
+			 * are currently pointing to the last inode.
+			 *
+			 * Because we may see inodes that are from the wrong AG
+			 * due to RCU freeing and reallocation, only update the
+			 * index if it lies in this AG. It was a race that lead
+			 * us to see this inode, so another lookup from the
+			 * same index will not find it again.
+			 */
+			if (XFS_INO_TO_AGNO(mp, ip->i_ino) != pag->pag_agno)
+				continue;
+			first_index = XFS_INO_TO_AGINO(mp, ip->i_ino + 1);
+			if (first_index < XFS_INO_TO_AGINO(mp, ip->i_ino))
+				done = true;
+		}
+
+		/* unlock now we've grabbed the inodes. */
+		rcu_read_unlock();
+
+		for (i = 0; i < nr_found; i++) {
+			if (!ragi->lookup_batch[i])
+				continue;
+			error = xrep_iunlink_visit(ragi, i);
+			if (error)
+				return error;
+		}
+	} while (!done);
+
+	return 0;
+}
+
+/* Mark all the unlinked ondisk inodes in this inobt record in iunlink_bmp. */
+STATIC int
+xrep_iunlink_mark_ondisk_rec(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_rec	*rec,
+	void				*priv)
+{
+	struct xfs_inobt_rec_incore	irec;
+	struct xrep_agi			*ragi = priv;
+	struct xfs_scrub		*sc = ragi->sc;
+	struct xfs_mount		*mp = cur->bc_mp;
+	xfs_agino_t			agino;
+	unsigned int			i;
+	int				error = 0;
+
+	xfs_inobt_btrec_to_irec(mp, rec, &irec);
+
+	for (i = 0, agino = irec.ir_startino;
+	     i < XFS_INODES_PER_CHUNK;
+	     i++, agino++) {
+		struct xfs_inode	*ip;
+		unsigned int		len = 1;
+
+		/* Skip free inodes */
+		if (XFS_INOBT_MASK(i) & irec.ir_free)
+			continue;
+		/* Skip inodes we've seen before */
+		if (xagino_bitmap_test(&ragi->iunlink_bmp, agino, &len))
+			continue;
+
+		/*
+		 * Skip incore inodes; these were already picked up by
+		 * the _mark_incore step.
+		 */
+		rcu_read_lock();
+		ip = radix_tree_lookup(&sc->sa.pag->pag_ici_root, agino);
+		rcu_read_unlock();
+		if (ip)
+			continue;
+
+		/*
+		 * Try to look up this inode.  If we can't get it, just move
+		 * on because we haven't actually scrubbed the inobt or the
+		 * inodes yet.
+		 */
+		error = xchk_iget(ragi->sc,
+				XFS_AGINO_TO_INO(mp, sc->sa.pag->pag_agno,
+						 agino),
+				&ip);
+		if (error)
+			continue;
+
+		trace_xrep_iunlink_reload_ondisk(ip);
+
+		if (VFS_I(ip)->i_nlink == 0)
+			error = xagino_bitmap_set(&ragi->iunlink_bmp, agino, 1);
+		xchk_irele(sc, ip);
+		if (error)
+			break;
+	}
+
+	return error;
+}
+
+/*
+ * Find ondisk inodes that are unlinked and not in cache, and mark them in
+ * iunlink_bmp.   We haven't checked the inobt yet, so we don't error out if
+ * the btree is corrupt.
+ */
+STATIC void
+xrep_iunlink_mark_ondisk(
+	struct xrep_agi		*ragi)
+{
+	struct xfs_scrub	*sc = ragi->sc;
+	struct xfs_buf		*agi_bp = ragi->agi_bp;
+	struct xfs_btree_cur	*cur;
+	int			error;
+
+	cur = xfs_inobt_init_cursor(sc->sa.pag, sc->tp, agi_bp);
+	error = xfs_btree_query_all(cur, xrep_iunlink_mark_ondisk_rec, ragi);
+	xfs_btree_del_cursor(cur, error);
+}
+
+/*
+ * Walk an iunlink bucket's inode list.  For each inode that should be on this
+ * chain, clear its entry in in iunlink_bmp because it's ok and we don't need
+ * to touch it further.
+ */
+STATIC int
+xrep_iunlink_resolve_bucket(
+	struct xrep_agi		*ragi,
+	unsigned int		bucket)
+{
+	struct xfs_scrub	*sc = ragi->sc;
+	struct xfs_inode	*ip;
+	xfs_agino_t		prev_agino = NULLAGINO;
+	xfs_agino_t		next_agino = ragi->iunlink_heads[bucket];
+	int			error = 0;
+
+	while (next_agino != NULLAGINO) {
+		if (xchk_should_terminate(ragi->sc, &error))
+			return error;
+
+		/* Find the next inode in the chain. */
+		ip = xfs_iunlink_lookup(sc->sa.pag, next_agino);
+		if (!ip) {
+			/* Inode not incore?  Terminate the chain. */
+			trace_xrep_iunlink_resolve_uncached(sc->sa.pag,
+					bucket, prev_agino, next_agino);
+
+			next_agino = NULLAGINO;
+			break;
+		}
+
+		if (next_agino % XFS_AGI_UNLINKED_BUCKETS != bucket) {
+			/*
+			 * Inode is in the wrong bucket.  Advance the list,
+			 * but pretend we didn't see this inode.
+			 */
+			trace_xrep_iunlink_resolve_wronglist(sc->sa.pag,
+					bucket, prev_agino, next_agino);
+
+			next_agino = ip->i_next_unlinked;
+			continue;
+		}
+
+		if (!xfs_inode_on_unlinked_list(ip)) {
+			/*
+			 * Incore inode doesn't think this inode is on an
+			 * unlinked list.  This is probably because we reloaded
+			 * it from disk.  Advance the list, but pretend we
+			 * didn't see this inode; we'll fix that later.
+			 */
+			trace_xrep_iunlink_resolve_nolist(sc->sa.pag,
+					bucket, prev_agino, next_agino);
+			next_agino = ip->i_next_unlinked;
+			continue;
+		}
+
+		trace_xrep_iunlink_resolve_ok(sc->sa.pag, bucket, prev_agino,
+				next_agino);
+
+		/*
+		 * Otherwise, this inode's unlinked pointers are ok.  Clear it
+		 * from the unlinked bitmap since we're done with it, and make
+		 * sure the chain is still correct.
+		 */
+		error = xagino_bitmap_clear(&ragi->iunlink_bmp, next_agino, 1);
+		if (error)
+			return error;
+
+		/* Remember the previous inode's next pointer. */
+		if (prev_agino != NULLAGINO) {
+			error = xrep_iunlink_store_next(ragi, prev_agino,
+					next_agino);
+			if (error)
+				return error;
+		}
+
+		/* Remember this inode's previous pointer. */
+		error = xrep_iunlink_store_prev(ragi, next_agino, prev_agino);
+		if (error)
+			return error;
+
+		/* Advance the list and remember this inode. */
+		prev_agino = next_agino;
+		next_agino = ip->i_next_unlinked;
+	}
+
+	/* Update the previous inode's next pointer. */
+	if (prev_agino != NULLAGINO) {
+		error = xrep_iunlink_store_next(ragi, prev_agino, next_agino);
+		if (error)
+			return error;
+	}
+
+	return 0;
+}
+
+/* Reinsert this unlinked inode into the head of the staged bucket list. */
+STATIC int
+xrep_iunlink_add_to_bucket(
+	struct xrep_agi		*ragi,
+	xfs_agino_t		agino)
+{
+	xfs_agino_t		current_head;
+	unsigned int		bucket;
+	int			error;
+
+	bucket = agino % XFS_AGI_UNLINKED_BUCKETS;
+
+	/* Point this inode at the current head of the bucket list. */
+	current_head = ragi->iunlink_heads[bucket];
+
+	trace_xrep_iunlink_add_to_bucket(ragi->sc->sa.pag, bucket, agino,
+			current_head);
+
+	error = xrep_iunlink_store_next(ragi, agino, current_head);
+	if (error)
+		return error;
+
+	/* Remember the head inode's previous pointer. */
+	if (current_head != NULLAGINO) {
+		error = xrep_iunlink_store_prev(ragi, current_head, agino);
+		if (error)
+			return error;
+	}
+
+	ragi->iunlink_heads[bucket] = agino;
+	return 0;
+}
+
+/* Reinsert unlinked inodes into the staged iunlink buckets. */
+STATIC int
+xrep_iunlink_add_lost_inodes(
+	uint32_t		start,
+	uint32_t		len,
+	void			*priv)
+{
+	struct xrep_agi		*ragi = priv;
+	int			error;
+
+	for (; len > 0; start++, len--) {
+		error = xrep_iunlink_add_to_bucket(ragi, start);
+		if (error)
+			return error;
+	}
+
+	return 0;
+}
+
+/*
+ * Figure out the iunlink bucket values and find inodes that need to be
+ * reinserted into the list.
+ */
+STATIC int
+xrep_iunlink_rebuild_buckets(
+	struct xrep_agi		*ragi)
+{
+	unsigned int		i;
+	int			error;
+
+	/*
+	 * Walk the ondisk AGI unlinked list to find inodes that are on the
+	 * list but aren't in memory.  This can happen if a past log recovery
+	 * tried to clear the iunlinked list but failed.  Our scan rebuilds the
+	 * unlinked list using incore inodes, so we must load and link them
+	 * properly.
+	 */
+	for (i = 0; i < XFS_AGI_UNLINKED_BUCKETS; i++) {
+		error = xrep_iunlink_walk_ondisk_bucket(ragi, i);
+		if (error)
+			return error;
+	}
+
+	/*
+	 * Record all the incore unlinked inodes in iunlink_bmp that we didn't
+	 * find by walking the ondisk iunlink buckets.  This shouldn't happen,
+	 * but we can't risk forgetting an inode somewhere.
+	 */
+	error = xrep_iunlink_mark_incore(ragi);
+	if (error)
+		return error;
+
+	/*
+	 * If there are ondisk inodes that are unlinked and are not been loaded
+	 * into cache, record them in iunlink_bmp.
+	 */
+	xrep_iunlink_mark_ondisk(ragi);
+
+	/*
+	 * Walk each iunlink bucket to (re)construct as much of the incore list
+	 * as would be correct.  For each inode that survives this step, mark
+	 * it clear in iunlink_bmp; we're done with those inodes.
+	 */
+	for (i = 0; i < XFS_AGI_UNLINKED_BUCKETS; i++) {
+		error = xrep_iunlink_resolve_bucket(ragi, i);
+		if (error)
+			return error;
+	}
+
+	/*
+	 * Any unlinked inodes that we didn't find through the bucket list
+	 * walk (or was ignored by the walk) must be inserted into the bucket
+	 * list.  Stage this in memory for now.
+	 */
+	return xagino_bitmap_walk(&ragi->iunlink_bmp,
+			xrep_iunlink_add_lost_inodes, ragi);
+}
+
+/* Update i_next_iunlinked for the inode @agino. */
+STATIC int
+xrep_iunlink_relink_next(
+	struct xrep_agi		*ragi,
+	xfarray_idx_t		idx,
+	xfs_agino_t		next_agino)
+{
+	struct xfs_scrub	*sc = ragi->sc;
+	struct xfs_perag	*pag = sc->sa.pag;
+	struct xfs_inode	*ip;
+	xfarray_idx_t		agino = idx - 1;
+	bool			want_rele = false;
+	int			error = 0;
+
+	ip = xfs_iunlink_lookup(pag, agino);
+	if (!ip) {
+		xfs_ino_t	ino;
+		xfs_agino_t	prev_agino;
+
+		/*
+		 * No inode exists in cache.  Load it off the disk so that we
+		 * can reinsert it into the incore unlinked list.
+		 */
+		ino = XFS_AGINO_TO_INO(sc->mp, pag->pag_agno, agino);
+		error = xchk_iget(sc, ino, &ip);
+		if (error)
+			return -EFSCORRUPTED;
+
+		want_rele = true;
+
+		/* Set the backward pointer since this just came off disk. */
+		error = xfarray_load(ragi->iunlink_prev, agino, &prev_agino);
+		if (error)
+			goto out_rele;
+
+		trace_xrep_iunlink_relink_prev(ip, prev_agino);
+		ip->i_prev_unlinked = prev_agino;
+	}
+
+	/* Update the forward pointer. */
+	if (ip->i_next_unlinked != next_agino) {
+		error = xfs_iunlink_log_inode(sc->tp, ip, pag, next_agino);
+		if (error)
+			goto out_rele;
+
+		trace_xrep_iunlink_relink_next(ip, next_agino);
+		ip->i_next_unlinked = next_agino;
+	}
+
+out_rele:
+	/*
+	 * The iunlink lookup doesn't igrab because we hold the AGI buffer lock
+	 * and the inode cannot be reclaimed.  However, if we used iget to load
+	 * a missing inode, we must irele it here.
+	 */
+	if (want_rele)
+		xchk_irele(sc, ip);
+	return error;
+}
+
+/* Update i_prev_iunlinked for the inode @agino. */
+STATIC int
+xrep_iunlink_relink_prev(
+	struct xrep_agi		*ragi,
+	xfarray_idx_t		idx,
+	xfs_agino_t		prev_agino)
+{
+	struct xfs_scrub	*sc = ragi->sc;
+	struct xfs_perag	*pag = sc->sa.pag;
+	struct xfs_inode	*ip;
+	xfarray_idx_t		agino = idx - 1;
+	bool			want_rele = false;
+	int			error = 0;
+
+	ASSERT(prev_agino != 0);
+
+	ip = xfs_iunlink_lookup(pag, agino);
+	if (!ip) {
+		xfs_ino_t	ino;
+		xfs_agino_t	next_agino;
+
+		/*
+		 * No inode exists in cache.  Load it off the disk so that we
+		 * can reinsert it into the incore unlinked list.
+		 */
+		ino = XFS_AGINO_TO_INO(sc->mp, pag->pag_agno, agino);
+		error = xchk_iget(sc, ino, &ip);
+		if (error)
+			return -EFSCORRUPTED;
+
+		want_rele = true;
+
+		/* Set the forward pointer since this just came off disk. */
+		error = xfarray_load(ragi->iunlink_prev, agino, &next_agino);
+		if (error)
+			goto out_rele;
+
+		error = xfs_iunlink_log_inode(sc->tp, ip, pag, next_agino);
+		if (error)
+			goto out_rele;
+
+		trace_xrep_iunlink_relink_next(ip, next_agino);
+		ip->i_next_unlinked = next_agino;
+	}
+
+	/* Update the backward pointer. */
+	if (ip->i_prev_unlinked != prev_agino) {
+		trace_xrep_iunlink_relink_prev(ip, prev_agino);
+		ip->i_prev_unlinked = prev_agino;
+	}
+
+out_rele:
+	/*
+	 * The iunlink lookup doesn't igrab because we hold the AGI buffer lock
+	 * and the inode cannot be reclaimed.  However, if we used iget to load
+	 * a missing inode, we must irele it here.
+	 */
+	if (want_rele)
+		xchk_irele(sc, ip);
+	return error;
+}
+
+/* Log all the iunlink updates we need to finish regenerating the AGI. */
+STATIC int
+xrep_iunlink_commit(
+	struct xrep_agi		*ragi)
+{
+	struct xfs_agi		*agi = ragi->agi_bp->b_addr;
+	xfarray_idx_t		idx = XFARRAY_CURSOR_INIT;
+	xfs_agino_t		agino;
+	unsigned int		i;
+	int			error;
+
+	/* Fix all the forward links */
+	while ((error = xfarray_iter(ragi->iunlink_next, &idx, &agino)) == 1) {
+		error = xrep_iunlink_relink_next(ragi, idx, agino);
+		if (error)
+			return error;
+	}
+
+	/* Fix all the back links */
+	idx = XFARRAY_CURSOR_INIT;
+	while ((error = xfarray_iter(ragi->iunlink_prev, &idx, &agino)) == 1) {
+		error = xrep_iunlink_relink_prev(ragi, idx, agino);
+		if (error)
+			return error;
+	}
+
+	/* Copy the staged iunlink buckets to the new AGI. */
+	for (i = 0; i < XFS_AGI_UNLINKED_BUCKETS; i++) {
+		trace_xrep_iunlink_commit_bucket(ragi->sc->sa.pag, i,
+				be32_to_cpu(ragi->old_agi.agi_unlinked[i]),
+				ragi->iunlink_heads[i]);
+
+		agi->agi_unlinked[i] = cpu_to_be32(ragi->iunlink_heads[i]);
+	}
+
+	return 0;
+}
+
 /* Trigger reinitialization of the in-core data. */
 STATIC int
 xrep_agi_commit_new(
-	struct xfs_scrub	*sc,
-	struct xfs_buf		*agi_bp)
+	struct xrep_agi		*ragi)
 {
+	struct xfs_scrub	*sc = ragi->sc;
+	struct xfs_buf		*agi_bp = ragi->agi_bp;
 	struct xfs_perag	*pag;
 	struct xfs_agi		*agi = agi_bp->b_addr;
 
@@ -971,33 +1712,58 @@ xrep_agi_commit_new(
 /* Repair the AGI. */
 int
 xrep_agi(
-	struct xfs_scrub		*sc)
+	struct xfs_scrub	*sc)
 {
-	struct xrep_find_ag_btree	fab[XREP_AGI_MAX] = {
-		[XREP_AGI_INOBT] = {
-			.rmap_owner = XFS_RMAP_OWN_INOBT,
-			.buf_ops = &xfs_inobt_buf_ops,
-			.maxlevels = M_IGEO(sc->mp)->inobt_maxlevels,
-		},
-		[XREP_AGI_FINOBT] = {
-			.rmap_owner = XFS_RMAP_OWN_INOBT,
-			.buf_ops = &xfs_finobt_buf_ops,
-			.maxlevels = M_IGEO(sc->mp)->inobt_maxlevels,
-		},
-		[XREP_AGI_END] = {
-			.buf_ops = NULL
-		},
-	};
-	struct xfs_agi			old_agi;
-	struct xfs_mount		*mp = sc->mp;
-	struct xfs_buf			*agi_bp;
-	struct xfs_agi			*agi;
-	int				error;
+	struct xrep_agi		*ragi;
+	struct xfs_mount	*mp = sc->mp;
+	char			*descr;
+	unsigned int		i;
+	int			error;
 
 	/* We require the rmapbt to rebuild anything. */
 	if (!xfs_has_rmapbt(mp))
 		return -EOPNOTSUPP;
 
+	sc->buf = kzalloc(sizeof(struct xrep_agi), XCHK_GFP_FLAGS);
+	if (!sc->buf)
+		return -ENOMEM;
+	ragi = sc->buf;
+	ragi->sc = sc;
+
+	ragi->fab[XREP_AGI_INOBT] = (struct xrep_find_ag_btree){
+		.rmap_owner	= XFS_RMAP_OWN_INOBT,
+		.buf_ops	= &xfs_inobt_buf_ops,
+		.maxlevels	= M_IGEO(sc->mp)->inobt_maxlevels,
+	};
+	ragi->fab[XREP_AGI_FINOBT] = (struct xrep_find_ag_btree){
+		.rmap_owner	= XFS_RMAP_OWN_INOBT,
+		.buf_ops	= &xfs_finobt_buf_ops,
+		.maxlevels	= M_IGEO(sc->mp)->inobt_maxlevels,
+	};
+	ragi->fab[XREP_AGI_END] = (struct xrep_find_ag_btree){
+		.buf_ops	= NULL,
+	};
+
+	for (i = 0; i < XFS_AGI_UNLINKED_BUCKETS; i++)
+		ragi->iunlink_heads[i] = NULLAGINO;
+
+	xagino_bitmap_init(&ragi->iunlink_bmp);
+	sc->buf_cleanup = xrep_agi_buf_cleanup;
+
+	descr = xchk_xfile_ag_descr(sc, "iunlinked next pointers");
+	error = xfarray_create(descr, 0, sizeof(xfs_agino_t),
+			&ragi->iunlink_next);
+	kfree(descr);
+	if (error)
+		return error;
+
+	descr = xchk_xfile_ag_descr(sc, "iunlinked prev pointers");
+	error = xfarray_create(descr, 0, sizeof(xfs_agino_t),
+			&ragi->iunlink_prev);
+	kfree(descr);
+	if (error)
+		return error;
+
 	/*
 	 * Make sure we have the AGI buffer, as scrub might have decided it
 	 * was corrupt after xfs_ialloc_read_agi failed with -EFSCORRUPTED.
@@ -1005,14 +1771,17 @@ xrep_agi(
 	error = xfs_trans_read_buf(mp, sc->tp, mp->m_ddev_targp,
 			XFS_AG_DADDR(mp, sc->sa.pag->pag_agno,
 						XFS_AGI_DADDR(mp)),
-			XFS_FSS_TO_BB(mp, 1), 0, &agi_bp, NULL);
+			XFS_FSS_TO_BB(mp, 1), 0, &ragi->agi_bp, NULL);
 	if (error)
 		return error;
-	agi_bp->b_ops = &xfs_agi_buf_ops;
-	agi = agi_bp->b_addr;
+	ragi->agi_bp->b_ops = &xfs_agi_buf_ops;
 
 	/* Find the AGI btree roots. */
-	error = xrep_agi_find_btrees(sc, fab);
+	error = xrep_agi_find_btrees(ragi);
+	if (error)
+		return error;
+
+	error = xrep_iunlink_rebuild_buckets(ragi);
 	if (error)
 		return error;
 
@@ -1021,18 +1790,21 @@ xrep_agi(
 		return error;
 
 	/* Start rewriting the header and implant the btrees we found. */
-	xrep_agi_init_header(sc, agi_bp, &old_agi);
-	xrep_agi_set_roots(sc, agi, fab);
-	error = xrep_agi_calc_from_btrees(sc, agi_bp);
+	xrep_agi_init_header(ragi);
+	xrep_agi_set_roots(ragi);
+	error = xrep_agi_calc_from_btrees(ragi);
+	if (error)
+		goto out_revert;
+	error = xrep_iunlink_commit(ragi);
 	if (error)
 		goto out_revert;
 
 	/* Reinitialize in-core state. */
-	return xrep_agi_commit_new(sc, agi_bp);
+	return xrep_agi_commit_new(ragi);
 
 out_revert:
 	/* Mark the incore AGI state stale and revert the AGI. */
 	clear_bit(XFS_AGSTATE_AGI_INIT, &sc->sa.pag->pag_opstate);
-	memcpy(agi, &old_agi, sizeof(old_agi));
+	memcpy(ragi->agi_bp->b_addr, &ragi->old_agi, sizeof(struct xfs_agi));
 	return error;
 }
diff --git a/fs/xfs/scrub/agino_bitmap.h b/fs/xfs/scrub/agino_bitmap.h
new file mode 100644
index 0000000000..56d7db5f16
--- /dev/null
+++ b/fs/xfs/scrub/agino_bitmap.h
@@ -0,0 +1,49 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2018-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_AGINO_BITMAP_H__
+#define __XFS_SCRUB_AGINO_BITMAP_H__
+
+/* Bitmaps, but for type-checked for xfs_agino_t */
+
+struct xagino_bitmap {
+	struct xbitmap32	aginobitmap;
+};
+
+static inline void xagino_bitmap_init(struct xagino_bitmap *bitmap)
+{
+	xbitmap32_init(&bitmap->aginobitmap);
+}
+
+static inline void xagino_bitmap_destroy(struct xagino_bitmap *bitmap)
+{
+	xbitmap32_destroy(&bitmap->aginobitmap);
+}
+
+static inline int xagino_bitmap_clear(struct xagino_bitmap *bitmap,
+		xfs_agino_t agino, unsigned int len)
+{
+	return xbitmap32_clear(&bitmap->aginobitmap, agino, len);
+}
+
+static inline int xagino_bitmap_set(struct xagino_bitmap *bitmap,
+		xfs_agino_t agino, unsigned int len)
+{
+	return xbitmap32_set(&bitmap->aginobitmap, agino, len);
+}
+
+static inline bool xagino_bitmap_test(struct xagino_bitmap *bitmap,
+		xfs_agino_t agino, unsigned int *len)
+{
+	return xbitmap32_test(&bitmap->aginobitmap, agino, len);
+}
+
+static inline int xagino_bitmap_walk(struct xagino_bitmap *bitmap,
+		xbitmap32_walk_fn fn, void *priv)
+{
+	return xbitmap32_walk(&bitmap->aginobitmap, fn, priv);
+}
+
+#endif	/* __XFS_SCRUB_AGINO_BITMAP_H__ */
diff --git a/fs/xfs/scrub/alloc_repair.c b/fs/xfs/scrub/alloc_repair.c
index 45edda0968..30295898cc 100644
--- a/fs/xfs/scrub/alloc_repair.c
+++ b/fs/xfs/scrub/alloc_repair.c
@@ -687,8 +687,8 @@ xrep_abt_reset_counters(
 	 * height values before re-initializing the perag info from the updated
 	 * AGF to capture all the new values.
 	 */
-	pag->pagf_repair_levels[XFS_BTNUM_BNOi] = pag->pagf_levels[XFS_BTNUM_BNOi];
-	pag->pagf_repair_levels[XFS_BTNUM_CNTi] = pag->pagf_levels[XFS_BTNUM_CNTi];
+	pag->pagf_repair_bno_level = pag->pagf_bno_level;
+	pag->pagf_repair_cnt_level = pag->pagf_cnt_level;
 
 	/* Reinitialize with the values we just logged. */
 	return xrep_reinit_pagf(sc);
@@ -735,10 +735,11 @@ xrep_abt_build_new_trees(
 	ra->new_cntbt.bload.claim_block = xrep_abt_claim_block;
 
 	/* Allocate cursors for the staged btrees. */
-	bno_cur = xfs_allocbt_stage_cursor(sc->mp, &ra->new_bnobt.afake,
-			pag, XFS_BTNUM_BNO);
-	cnt_cur = xfs_allocbt_stage_cursor(sc->mp, &ra->new_cntbt.afake,
-			pag, XFS_BTNUM_CNT);
+	bno_cur = xfs_bnobt_init_cursor(sc->mp, NULL, NULL, pag);
+	xfs_btree_stage_afakeroot(bno_cur, &ra->new_bnobt.afake);
+
+	cnt_cur = xfs_cntbt_init_cursor(sc->mp, NULL, NULL, pag);
+	xfs_btree_stage_afakeroot(cnt_cur, &ra->new_cntbt.afake);
 
 	/* Last chance to abort before we start committing fixes. */
 	if (xchk_should_terminate(sc, &error))
@@ -765,10 +766,8 @@ xrep_abt_build_new_trees(
 	 * height so that we don't trip the verifiers when writing the new
 	 * btree blocks to disk.
 	 */
-	pag->pagf_repair_levels[XFS_BTNUM_BNOi] =
-					ra->new_bnobt.bload.btree_height;
-	pag->pagf_repair_levels[XFS_BTNUM_CNTi] =
-					ra->new_cntbt.bload.btree_height;
+	pag->pagf_repair_bno_level = ra->new_bnobt.bload.btree_height;
+	pag->pagf_repair_cnt_level = ra->new_cntbt.bload.btree_height;
 
 	/* Load the free space by length tree. */
 	ra->array_cur = XFARRAY_CURSOR_INIT;
@@ -779,7 +778,7 @@ xrep_abt_build_new_trees(
 
 	error = xrep_bnobt_sort_records(ra);
 	if (error)
-		return error;
+		goto err_levels;
 
 	/* Load the free space by block number tree. */
 	ra->array_cur = XFARRAY_CURSOR_INIT;
@@ -807,8 +806,8 @@ xrep_abt_build_new_trees(
 	return xrep_roll_ag_trans(sc);
 
 err_levels:
-	pag->pagf_repair_levels[XFS_BTNUM_BNOi] = 0;
-	pag->pagf_repair_levels[XFS_BTNUM_CNTi] = 0;
+	pag->pagf_repair_bno_level = 0;
+	pag->pagf_repair_cnt_level = 0;
 err_cur:
 	xfs_btree_del_cursor(cnt_cur, error);
 	xfs_btree_del_cursor(bno_cur, error);
@@ -838,8 +837,8 @@ xrep_abt_remove_old_trees(
 	 * Now that we've zapped all the old allocbt blocks we can turn off
 	 * the alternate height mechanism.
 	 */
-	pag->pagf_repair_levels[XFS_BTNUM_BNOi] = 0;
-	pag->pagf_repair_levels[XFS_BTNUM_CNTi] = 0;
+	pag->pagf_repair_bno_level = 0;
+	pag->pagf_repair_cnt_level = 0;
 	return 0;
 }
 
diff --git a/fs/xfs/scrub/attr.c b/fs/xfs/scrub/attr.c
index 83c7feb387..708334f9b2 100644
--- a/fs/xfs/scrub/attr.c
+++ b/fs/xfs/scrub/attr.c
@@ -10,16 +10,20 @@
 #include "xfs_trans_resv.h"
 #include "xfs_mount.h"
 #include "xfs_log_format.h"
+#include "xfs_trans.h"
 #include "xfs_inode.h"
 #include "xfs_da_format.h"
 #include "xfs_da_btree.h"
 #include "xfs_attr.h"
 #include "xfs_attr_leaf.h"
 #include "xfs_attr_sf.h"
+#include "xfs_parent.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/dabtree.h"
 #include "scrub/attr.h"
+#include "scrub/listxattr.h"
+#include "scrub/repair.h"
 
 /* Free the buffers linked from the xattr buffer. */
 static void
@@ -35,6 +39,8 @@ xchk_xattr_buf_cleanup(
 	kvfree(ab->value);
 	ab->value = NULL;
 	ab->value_sz = 0;
+	kvfree(ab->name);
+	ab->name = NULL;
 }
 
 /*
@@ -65,7 +71,7 @@ xchk_xattr_want_freemap(
  * reallocating the buffer if necessary.  Buffer contents are not preserved
  * across a reallocation.
  */
-static int
+int
 xchk_setup_xattr_buf(
 	struct xfs_scrub	*sc,
 	size_t			value_size)
@@ -95,6 +101,12 @@ xchk_setup_xattr_buf(
 			return -ENOMEM;
 	}
 
+	if (xchk_could_repair(sc)) {
+		ab->name = kvmalloc(XATTR_NAME_MAX + 1, XCHK_GFP_FLAGS);
+		if (!ab->name)
+			return -ENOMEM;
+	}
+
 resize_value:
 	if (ab->value_sz >= value_size)
 		return 0;
@@ -121,6 +133,12 @@ xchk_setup_xattr(
 {
 	int			error;
 
+	if (xchk_could_repair(sc)) {
+		error = xrep_setup_xattr(sc);
+		if (error)
+			return error;
+	}
+
 	/*
 	 * We failed to get memory while checking attrs, so this time try to
 	 * get all the memory we're ever going to need.  Allocate the buffer
@@ -137,106 +155,105 @@ xchk_setup_xattr(
 
 /* Extended Attributes */
 
-struct xchk_xattr {
-	struct xfs_attr_list_context	context;
-	struct xfs_scrub		*sc;
-};
-
 /*
  * Check that an extended attribute key can be looked up by hash.
  *
- * We use the XFS attribute list iterator (i.e. xfs_attr_list_ilocked)
- * to call this function for every attribute key in an inode.  Once
- * we're here, we load the attribute value to see if any errors happen,
- * or if we get more or less data than we expected.
+ * We use the extended attribute walk helper to call this function for every
+ * attribute key in an inode.  Once we're here, we load the attribute value to
+ * see if any errors happen, or if we get more or less data than we expected.
  */
-static void
-xchk_xattr_listent(
-	struct xfs_attr_list_context	*context,
-	int				flags,
-	unsigned char			*name,
-	int				namelen,
-	int				valuelen)
+static int
+xchk_xattr_actor(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*ip,
+	unsigned int		attr_flags,
+	const unsigned char	*name,
+	unsigned int		namelen,
+	const void		*value,
+	unsigned int		valuelen,
+	void			*priv)
 {
 	struct xfs_da_args		args = {
-		.op_flags		= XFS_DA_OP_NOTIME,
-		.attr_filter		= flags & XFS_ATTR_NSP_ONDISK_MASK,
-		.geo			= context->dp->i_mount->m_attr_geo,
+		.attr_filter		= attr_flags & XFS_ATTR_NSP_ONDISK_MASK,
+		.geo			= sc->mp->m_attr_geo,
 		.whichfork		= XFS_ATTR_FORK,
-		.dp			= context->dp,
+		.dp			= ip,
 		.name			= name,
 		.namelen		= namelen,
-		.hashval		= xfs_da_hashname(name, namelen),
-		.trans			= context->tp,
+		.trans			= sc->tp,
 		.valuelen		= valuelen,
+		.owner			= ip->i_ino,
 	};
 	struct xchk_xattr_buf		*ab;
-	struct xchk_xattr		*sx;
 	int				error = 0;
 
-	sx = container_of(context, struct xchk_xattr, context);
-	ab = sx->sc->buf;
+	ab = sc->buf;
 
-	if (xchk_should_terminate(sx->sc, &error)) {
-		context->seen_enough = error;
-		return;
+	if (xchk_should_terminate(sc, &error))
+		return error;
+
+	if (attr_flags & ~XFS_ATTR_ONDISK_MASK) {
+		xchk_fblock_set_corrupt(sc, XFS_ATTR_FORK, args.blkno);
+		return -ECANCELED;
 	}
 
-	if (flags & XFS_ATTR_INCOMPLETE) {
+	if (attr_flags & XFS_ATTR_INCOMPLETE) {
 		/* Incomplete attr key, just mark the inode for preening. */
-		xchk_ino_set_preen(sx->sc, context->dp->i_ino);
-		return;
+		xchk_ino_set_preen(sc, ip->i_ino);
+		return 0;
 	}
 
-	/* Only one namespace bit allowed. */
-	if (hweight32(flags & XFS_ATTR_NSP_ONDISK_MASK) > 1) {
-		xchk_fblock_set_corrupt(sx->sc, XFS_ATTR_FORK, args.blkno);
-		goto fail_xref;
+	/* Does this name make sense? */
+	if (!xfs_attr_namecheck(attr_flags, name, namelen)) {
+		xchk_fblock_set_corrupt(sc, XFS_ATTR_FORK, args.blkno);
+		return -ECANCELED;
 	}
 
-	/* Does this name make sense? */
-	if (!xfs_attr_namecheck(name, namelen)) {
-		xchk_fblock_set_corrupt(sx->sc, XFS_ATTR_FORK, args.blkno);
-		goto fail_xref;
+	/* Check parent pointer record. */
+	if ((attr_flags & XFS_ATTR_PARENT) &&
+	    !xfs_parent_valuecheck(sc->mp, value, valuelen)) {
+		xchk_fblock_set_corrupt(sc, XFS_ATTR_FORK, args.blkno);
+		return -ECANCELED;
 	}
 
 	/*
-	 * Local xattr values are stored in the attr leaf block, so we don't
-	 * need to retrieve the value from a remote block to detect corruption
-	 * problems.
+	 * Try to allocate enough memory to extract the attr value.  If that
+	 * doesn't work, return -EDEADLOCK as a signal to try again with a
+	 * maximally sized buffer.
 	 */
-	if (flags & XFS_ATTR_LOCAL)
-		goto fail_xref;
+	error = xchk_setup_xattr_buf(sc, valuelen);
+	if (error == -ENOMEM)
+		error = -EDEADLOCK;
+	if (error)
+		return error;
 
 	/*
-	 * Try to allocate enough memory to extrat the attr value.  If that
-	 * doesn't work, we overload the seen_enough variable to convey
-	 * the error message back to the main scrub function.
+	 * Parent pointers are matched on attr name and value, so we must
+	 * supply the xfs_parent_rec here when confirming that the dabtree
+	 * indexing works correctly.
 	 */
-	error = xchk_setup_xattr_buf(sx->sc, valuelen);
-	if (error == -ENOMEM)
-		error = -EDEADLOCK;
-	if (error) {
-		context->seen_enough = error;
-		return;
-	}
+	if (attr_flags & XFS_ATTR_PARENT)
+		memcpy(ab->value, value, valuelen);
 
 	args.value = ab->value;
 
+	/*
+	 * Get the attr value to ensure that lookup can find this attribute
+	 * through the dabtree indexing and that remote value retrieval also
+	 * works correctly.
+	 */
+	xfs_attr_sethash(&args);
 	error = xfs_attr_get_ilocked(&args);
 	/* ENODATA means the hash lookup failed and the attr is bad */
 	if (error == -ENODATA)
 		error = -EFSCORRUPTED;
-	if (!xchk_fblock_process_error(sx->sc, XFS_ATTR_FORK, args.blkno,
+	if (!xchk_fblock_process_error(sc, XFS_ATTR_FORK, args.blkno,
 			&error))
-		goto fail_xref;
+		return error;
 	if (args.valuelen != valuelen)
-		xchk_fblock_set_corrupt(sx->sc, XFS_ATTR_FORK,
-					     args.blkno);
-fail_xref:
-	if (sx->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
-		context->seen_enough = 1;
-	return;
+		xchk_fblock_set_corrupt(sc, XFS_ATTR_FORK, args.blkno);
+
+	return 0;
 }
 
 /*
@@ -246,7 +263,7 @@ fail_xref:
  * Within a char, the lowest bit of the char represents the byte with
  * the smallest address
  */
-STATIC bool
+bool
 xchk_xattr_set_map(
 	struct xfs_scrub	*sc,
 	unsigned long		*map,
@@ -403,6 +420,17 @@ xchk_xattr_block(
 	xfs_attr3_leaf_hdr_from_disk(mp->m_attr_geo, &leafhdr, leaf);
 	hdrsize = xfs_attr3_leaf_hdr_size(leaf);
 
+	/*
+	 * Empty xattr leaf blocks mapped at block 0 are probably a byproduct
+	 * of a race between setxattr and a log shutdown.  Anywhere else in the
+	 * attr fork is a corruption.
+	 */
+	if (leafhdr.count == 0) {
+		if (blk->blkno == 0)
+			xchk_da_set_preen(ds, level);
+		else
+			xchk_da_set_corrupt(ds, level);
+	}
 	if (leafhdr.usedbytes > mp->m_attr_geo->blksize)
 		xchk_da_set_corrupt(ds, level);
 	if (leafhdr.firstused > mp->m_attr_geo->blksize)
@@ -411,6 +439,8 @@ xchk_xattr_block(
 		xchk_da_set_corrupt(ds, level);
 	if (!xchk_xattr_set_map(ds->sc, ab->usedmap, 0, hdrsize))
 		xchk_da_set_corrupt(ds, level);
+	if (leafhdr.holes)
+		xchk_da_set_preen(ds, level);
 
 	if (ds->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
 		goto out;
@@ -463,7 +493,6 @@ xchk_xattr_rec(
 	xfs_dahash_t			hash;
 	int				nameidx;
 	int				hdrsize;
-	unsigned int			badflags;
 	int				error;
 
 	ASSERT(blk->magic == XFS_ATTR_LEAF_MAGIC);
@@ -493,10 +522,15 @@ xchk_xattr_rec(
 
 	/* Retrieve the entry and check it. */
 	hash = be32_to_cpu(ent->hashval);
-	badflags = ~(XFS_ATTR_LOCAL | XFS_ATTR_ROOT | XFS_ATTR_SECURE |
-			XFS_ATTR_INCOMPLETE);
-	if ((ent->flags & badflags) != 0)
+	if (ent->flags & ~XFS_ATTR_ONDISK_MASK) {
+		xchk_da_set_corrupt(ds, level);
+		return 0;
+	}
+	if (!xfs_attr_check_namespace(ent->flags)) {
 		xchk_da_set_corrupt(ds, level);
+		return 0;
+	}
+
 	if (ent->flags & XFS_ATTR_LOCAL) {
 		lentry = (struct xfs_attr_leaf_name_local *)
 				(((char *)bp->b_addr) + nameidx);
@@ -504,7 +538,10 @@ xchk_xattr_rec(
 			xchk_da_set_corrupt(ds, level);
 			goto out;
 		}
-		calc_hash = xfs_da_hashname(lentry->nameval, lentry->namelen);
+		calc_hash = xfs_attr_hashval(mp, ent->flags, lentry->nameval,
+					     lentry->namelen,
+					     lentry->nameval + lentry->namelen,
+					     be16_to_cpu(lentry->valuelen));
 	} else {
 		rentry = (struct xfs_attr_leaf_name_remote *)
 				(((char *)bp->b_addr) + nameidx);
@@ -512,7 +549,13 @@ xchk_xattr_rec(
 			xchk_da_set_corrupt(ds, level);
 			goto out;
 		}
-		calc_hash = xfs_da_hashname(rentry->name, rentry->namelen);
+		if (ent->flags & XFS_ATTR_PARENT) {
+			xchk_da_set_corrupt(ds, level);
+			goto out;
+		}
+		calc_hash = xfs_attr_hashval(mp, ent->flags, rentry->name,
+					     rentry->namelen, NULL,
+					     be32_to_cpu(rentry->valuelen));
 	}
 	if (calc_hash != hash)
 		xchk_da_set_corrupt(ds, level);
@@ -556,6 +599,15 @@ xchk_xattr_check_sf(
 			break;
 		}
 
+		/*
+		 * Shortform entries do not set LOCAL or INCOMPLETE, so the
+		 * only valid flag bits here are for namespaces.
+		 */
+		if (sfe->flags & ~XFS_ATTR_NSP_ONDISK_MASK) {
+			xchk_fblock_set_corrupt(sc, XFS_ATTR_FORK, 0);
+			break;
+		}
+
 		if (!xchk_xattr_set_map(sc, ab->usedmap,
 				(char *)sfe - (char *)sf,
 				sizeof(struct xfs_attr_sf_entry))) {
@@ -588,16 +640,6 @@ int
 xchk_xattr(
 	struct xfs_scrub		*sc)
 {
-	struct xchk_xattr		sx = {
-		.sc			= sc,
-		.context		= {
-			.dp		= sc->ip,
-			.tp		= sc->tp,
-			.resynch	= 1,
-			.put_listent	= xchk_xattr_listent,
-			.allow_incomplete = true,
-		},
-	};
 	xfs_dablk_t			last_checked = -1U;
 	int				error = 0;
 
@@ -626,12 +668,6 @@ xchk_xattr(
 	/*
 	 * Look up every xattr in this file by name and hash.
 	 *
-	 * Use the backend implementation of xfs_attr_list to call
-	 * xchk_xattr_listent on every attribute key in this inode.
-	 * In other words, we use the same iterator/callback mechanism
-	 * that listattr uses to scrub extended attributes, though in our
-	 * _listent function, we check the value of the attribute.
-	 *
 	 * The VFS only locks i_rwsem when modifying attrs, so keep all
 	 * three locks held because that's the only way to ensure we're
 	 * the only thread poking into the da btree.  We traverse the da
@@ -639,13 +675,9 @@ xchk_xattr(
 	 * iteration, which doesn't really follow the usual buffer
 	 * locking order.
 	 */
-	error = xfs_attr_list_ilocked(&sx.context);
+	error = xchk_xattr_walk(sc, sc->ip, xchk_xattr_actor, NULL, NULL);
 	if (!xchk_fblock_process_error(sc, XFS_ATTR_FORK, 0, &error))
 		return error;
 
-	/* Did our listent function try to return any errors? */
-	if (sx.context.seen_enough < 0)
-		return sx.context.seen_enough;
-
 	return 0;
 }
diff --git a/fs/xfs/scrub/attr.h b/fs/xfs/scrub/attr.h
index 48fd9402c4..7db58af566 100644
--- a/fs/xfs/scrub/attr.h
+++ b/fs/xfs/scrub/attr.h
@@ -16,9 +16,16 @@ struct xchk_xattr_buf {
 	/* Bitmap of free space in xattr leaf blocks. */
 	unsigned long		*freemap;
 
+	/* Memory buffer used to hold salvaged xattr names. */
+	unsigned char		*name;
+
 	/* Memory buffer used to extract xattr values. */
 	void			*value;
 	size_t			value_sz;
 };
 
+bool xchk_xattr_set_map(struct xfs_scrub *sc, unsigned long *map,
+		unsigned int start, unsigned int len);
+int xchk_setup_xattr_buf(struct xfs_scrub *sc, size_t value_size);
+
 #endif	/* __XFS_SCRUB_ATTR_H__ */
diff --git a/fs/xfs/scrub/attr_repair.c b/fs/xfs/scrub/attr_repair.c
new file mode 100644
index 0000000000..c7eb94069c
--- /dev/null
+++ b/fs/xfs/scrub/attr_repair.c
@@ -0,0 +1,1663 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2018-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
+#include "xfs_dir2.h"
+#include "xfs_attr.h"
+#include "xfs_attr_leaf.h"
+#include "xfs_attr_sf.h"
+#include "xfs_attr_remote.h"
+#include "xfs_bmap.h"
+#include "xfs_bmap_util.h"
+#include "xfs_exchmaps.h"
+#include "xfs_exchrange.h"
+#include "xfs_acl.h"
+#include "xfs_parent.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+#include "scrub/repair.h"
+#include "scrub/tempfile.h"
+#include "scrub/tempexch.h"
+#include "scrub/xfile.h"
+#include "scrub/xfarray.h"
+#include "scrub/xfblob.h"
+#include "scrub/attr.h"
+#include "scrub/reap.h"
+#include "scrub/attr_repair.h"
+
+/*
+ * Extended Attribute Repair
+ * =========================
+ *
+ * We repair extended attributes by reading the attr leaf blocks looking for
+ * attributes entries that look salvageable (name passes verifiers, value can
+ * be retrieved, etc).  Each extended attribute worth salvaging is stashed in
+ * memory, and the stashed entries are periodically replayed into a temporary
+ * file to constrain memory use.  Batching the construction of the temporary
+ * extended attribute structure in this fashion reduces lock cycling of the
+ * file being repaired and the temporary file.
+ *
+ * When salvaging completes, the remaining stashed attributes are replayed to
+ * the temporary file.  An atomic file contents exchange is used to commit the
+ * new xattr blocks to the file being repaired.  This will disrupt attrmulti
+ * cursors.
+ */
+
+struct xrep_xattr_key {
+	/* Cookie for retrieval of the xattr name. */
+	xfblob_cookie		name_cookie;
+
+	/* Cookie for retrieval of the xattr value. */
+	xfblob_cookie		value_cookie;
+
+	/* XFS_ATTR_* flags */
+	int			flags;
+
+	/* Length of the value and name. */
+	uint32_t		valuelen;
+	uint16_t		namelen;
+};
+
+/*
+ * Stash up to 8 pages of attrs in xattr_records/xattr_blobs before we write
+ * them to the temp file.
+ */
+#define XREP_XATTR_MAX_STASH_BYTES	(PAGE_SIZE * 8)
+
+struct xrep_xattr {
+	struct xfs_scrub	*sc;
+
+	/* Information for exchanging attr fork mappings at the end. */
+	struct xrep_tempexch	tx;
+
+	/* xattr keys */
+	struct xfarray		*xattr_records;
+
+	/* xattr values */
+	struct xfblob		*xattr_blobs;
+
+	/* Number of attributes that we are salvaging. */
+	unsigned long long	attrs_found;
+
+	/* Can we flush stashed attrs to the tempfile? */
+	bool			can_flush;
+
+	/* Did the live update fail, and hence the repair is now out of date? */
+	bool			live_update_aborted;
+
+	/* Lock protecting parent pointer updates */
+	struct mutex		lock;
+
+	/* Fixed-size array of xrep_xattr_pptr structures. */
+	struct xfarray		*pptr_recs;
+
+	/* Blobs containing parent pointer names. */
+	struct xfblob		*pptr_names;
+
+	/* Hook to capture parent pointer updates. */
+	struct xfs_dir_hook	dhook;
+
+	/* Scratch buffer for capturing parent pointers. */
+	struct xfs_da_args	pptr_args;
+
+	/* Name buffer */
+	struct xfs_name		xname;
+	char			namebuf[MAXNAMELEN];
+};
+
+/* Create a parent pointer in the tempfile. */
+#define XREP_XATTR_PPTR_ADD	(1)
+
+/* Remove a parent pointer from the tempfile. */
+#define XREP_XATTR_PPTR_REMOVE	(2)
+
+/* A stashed parent pointer update. */
+struct xrep_xattr_pptr {
+	/* Cookie for retrieval of the pptr name. */
+	xfblob_cookie		name_cookie;
+
+	/* Parent pointer record. */
+	struct xfs_parent_rec	pptr_rec;
+
+	/* Length of the pptr name. */
+	uint8_t			namelen;
+
+	/* XREP_XATTR_PPTR_{ADD,REMOVE} */
+	uint8_t			action;
+};
+
+/* Set up to recreate the extended attributes. */
+int
+xrep_setup_xattr(
+	struct xfs_scrub	*sc)
+{
+	if (xfs_has_parent(sc->mp))
+		xchk_fsgates_enable(sc, XCHK_FSGATES_DIRENTS);
+
+	return xrep_tempfile_create(sc, S_IFREG);
+}
+
+/*
+ * Decide if we want to salvage this attribute.  We don't bother with
+ * incomplete or oversized keys or values.  The @value parameter can be null
+ * for remote attrs.
+ */
+STATIC int
+xrep_xattr_want_salvage(
+	struct xrep_xattr	*rx,
+	unsigned int		attr_flags,
+	const void		*name,
+	int			namelen,
+	const void		*value,
+	int			valuelen)
+{
+	if (attr_flags & XFS_ATTR_INCOMPLETE)
+		return false;
+	if (namelen > XATTR_NAME_MAX || namelen <= 0)
+		return false;
+	if (!xfs_attr_namecheck(attr_flags, name, namelen))
+		return false;
+	if (valuelen > XATTR_SIZE_MAX || valuelen < 0)
+		return false;
+	if (attr_flags & XFS_ATTR_PARENT)
+		return xfs_parent_valuecheck(rx->sc->mp, value, valuelen);
+
+	return true;
+}
+
+/* Allocate an in-core record to hold xattrs while we rebuild the xattr data. */
+STATIC int
+xrep_xattr_salvage_key(
+	struct xrep_xattr	*rx,
+	int			flags,
+	unsigned char		*name,
+	int			namelen,
+	unsigned char		*value,
+	int			valuelen)
+{
+	struct xrep_xattr_key	key = {
+		.valuelen	= valuelen,
+		.flags		= flags & XFS_ATTR_NSP_ONDISK_MASK,
+	};
+	unsigned int		i = 0;
+	int			error = 0;
+
+	if (xchk_should_terminate(rx->sc, &error))
+		return error;
+
+	/*
+	 * Truncate the name to the first character that would trip namecheck.
+	 * If we no longer have a name after that, ignore this attribute.
+	 */
+	if (flags & XFS_ATTR_PARENT) {
+		key.namelen = namelen;
+
+		trace_xrep_xattr_salvage_pptr(rx->sc->ip, flags, name,
+				key.namelen, value, valuelen);
+	} else {
+		while (i < namelen && name[i] != 0)
+			i++;
+		if (i == 0)
+			return 0;
+		key.namelen = i;
+
+		trace_xrep_xattr_salvage_rec(rx->sc->ip, flags, name,
+				key.namelen, valuelen);
+	}
+
+	error = xfblob_store(rx->xattr_blobs, &key.name_cookie, name,
+			key.namelen);
+	if (error)
+		return error;
+
+	error = xfblob_store(rx->xattr_blobs, &key.value_cookie, value,
+			key.valuelen);
+	if (error)
+		return error;
+
+	error = xfarray_append(rx->xattr_records, &key);
+	if (error)
+		return error;
+
+	rx->attrs_found++;
+	return 0;
+}
+
+/*
+ * Record a shortform extended attribute key & value for later reinsertion
+ * into the inode.
+ */
+STATIC int
+xrep_xattr_salvage_sf_attr(
+	struct xrep_xattr		*rx,
+	struct xfs_attr_sf_hdr		*hdr,
+	struct xfs_attr_sf_entry	*sfe)
+{
+	struct xfs_scrub		*sc = rx->sc;
+	struct xchk_xattr_buf		*ab = sc->buf;
+	unsigned char			*name = sfe->nameval;
+	unsigned char			*value = &sfe->nameval[sfe->namelen];
+
+	if (!xchk_xattr_set_map(sc, ab->usedmap, (char *)name - (char *)hdr,
+			sfe->namelen))
+		return 0;
+
+	if (!xchk_xattr_set_map(sc, ab->usedmap, (char *)value - (char *)hdr,
+			sfe->valuelen))
+		return 0;
+
+	if (!xrep_xattr_want_salvage(rx, sfe->flags, sfe->nameval,
+			sfe->namelen, value, sfe->valuelen))
+		return 0;
+
+	return xrep_xattr_salvage_key(rx, sfe->flags, sfe->nameval,
+			sfe->namelen, value, sfe->valuelen);
+}
+
+/*
+ * Record a local format extended attribute key & value for later reinsertion
+ * into the inode.
+ */
+STATIC int
+xrep_xattr_salvage_local_attr(
+	struct xrep_xattr		*rx,
+	struct xfs_attr_leaf_entry	*ent,
+	unsigned int			nameidx,
+	const char			*buf_end,
+	struct xfs_attr_leaf_name_local	*lentry)
+{
+	struct xchk_xattr_buf		*ab = rx->sc->buf;
+	unsigned char			*value;
+	unsigned int			valuelen;
+	unsigned int			namesize;
+
+	/*
+	 * Decode the leaf local entry format.  If something seems wrong, we
+	 * junk the attribute.
+	 */
+	value = &lentry->nameval[lentry->namelen];
+	valuelen = be16_to_cpu(lentry->valuelen);
+	namesize = xfs_attr_leaf_entsize_local(lentry->namelen, valuelen);
+	if ((char *)lentry + namesize > buf_end)
+		return 0;
+	if (!xrep_xattr_want_salvage(rx, ent->flags, lentry->nameval,
+			lentry->namelen, value, valuelen))
+		return 0;
+	if (!xchk_xattr_set_map(rx->sc, ab->usedmap, nameidx, namesize))
+		return 0;
+
+	/* Try to save this attribute. */
+	return xrep_xattr_salvage_key(rx, ent->flags, lentry->nameval,
+			lentry->namelen, value, valuelen);
+}
+
+/*
+ * Record a remote format extended attribute key & value for later reinsertion
+ * into the inode.
+ */
+STATIC int
+xrep_xattr_salvage_remote_attr(
+	struct xrep_xattr		*rx,
+	struct xfs_attr_leaf_entry	*ent,
+	unsigned int			nameidx,
+	const char			*buf_end,
+	struct xfs_attr_leaf_name_remote *rentry,
+	unsigned int			ent_idx,
+	struct xfs_buf			*leaf_bp)
+{
+	struct xchk_xattr_buf		*ab = rx->sc->buf;
+	struct xfs_da_args		args = {
+		.trans			= rx->sc->tp,
+		.dp			= rx->sc->ip,
+		.index			= ent_idx,
+		.geo			= rx->sc->mp->m_attr_geo,
+		.owner			= rx->sc->ip->i_ino,
+		.attr_filter		= ent->flags & XFS_ATTR_NSP_ONDISK_MASK,
+		.namelen		= rentry->namelen,
+		.name			= rentry->name,
+		.value			= ab->value,
+		.valuelen		= be32_to_cpu(rentry->valuelen),
+	};
+	unsigned int			namesize;
+	int				error;
+
+	/*
+	 * Decode the leaf remote entry format.  If something seems wrong, we
+	 * junk the attribute.  Note that we should never find a zero-length
+	 * remote attribute value.
+	 */
+	namesize = xfs_attr_leaf_entsize_remote(rentry->namelen);
+	if ((char *)rentry + namesize > buf_end)
+		return 0;
+	if (args.valuelen == 0 ||
+	    !xrep_xattr_want_salvage(rx, ent->flags, rentry->name,
+			rentry->namelen, NULL, args.valuelen))
+		return 0;
+	if (!xchk_xattr_set_map(rx->sc, ab->usedmap, nameidx, namesize))
+		return 0;
+
+	/*
+	 * Enlarge the buffer (if needed) to hold the value that we're trying
+	 * to salvage from the old extended attribute data.
+	 */
+	error = xchk_setup_xattr_buf(rx->sc, args.valuelen);
+	if (error == -ENOMEM)
+		error = -EDEADLOCK;
+	if (error)
+		return error;
+
+	/* Look up the remote value and stash it for reconstruction. */
+	error = xfs_attr3_leaf_getvalue(leaf_bp, &args);
+	if (error || args.rmtblkno == 0)
+		goto err_free;
+
+	error = xfs_attr_rmtval_get(&args);
+	if (error)
+		goto err_free;
+
+	/* Try to save this attribute. */
+	error = xrep_xattr_salvage_key(rx, ent->flags, rentry->name,
+			rentry->namelen, ab->value, args.valuelen);
+err_free:
+	/* remote value was garbage, junk it */
+	if (error == -EFSBADCRC || error == -EFSCORRUPTED)
+		error = 0;
+	return error;
+}
+
+/* Extract every xattr key that we can from this attr fork block. */
+STATIC int
+xrep_xattr_recover_leaf(
+	struct xrep_xattr		*rx,
+	struct xfs_buf			*bp)
+{
+	struct xfs_attr3_icleaf_hdr	leafhdr;
+	struct xfs_scrub		*sc = rx->sc;
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_attr_leafblock	*leaf;
+	struct xfs_attr_leaf_name_local	*lentry;
+	struct xfs_attr_leaf_name_remote *rentry;
+	struct xfs_attr_leaf_entry	*ent;
+	struct xfs_attr_leaf_entry	*entries;
+	struct xchk_xattr_buf		*ab = rx->sc->buf;
+	char				*buf_end;
+	size_t				off;
+	unsigned int			nameidx;
+	unsigned int			hdrsize;
+	int				i;
+	int				error = 0;
+
+	bitmap_zero(ab->usedmap, mp->m_attr_geo->blksize);
+
+	/* Check the leaf header */
+	leaf = bp->b_addr;
+	xfs_attr3_leaf_hdr_from_disk(mp->m_attr_geo, &leafhdr, leaf);
+	hdrsize = xfs_attr3_leaf_hdr_size(leaf);
+	xchk_xattr_set_map(sc, ab->usedmap, 0, hdrsize);
+	entries = xfs_attr3_leaf_entryp(leaf);
+
+	buf_end = (char *)bp->b_addr + mp->m_attr_geo->blksize;
+	for (i = 0, ent = entries; i < leafhdr.count; ent++, i++) {
+		if (xchk_should_terminate(sc, &error))
+			return error;
+
+		/* Skip key if it conflicts with something else? */
+		off = (char *)ent - (char *)leaf;
+		if (!xchk_xattr_set_map(sc, ab->usedmap, off,
+				sizeof(xfs_attr_leaf_entry_t)))
+			continue;
+
+		/* Check the name information. */
+		nameidx = be16_to_cpu(ent->nameidx);
+		if (nameidx < leafhdr.firstused ||
+		    nameidx >= mp->m_attr_geo->blksize)
+			continue;
+
+		if (ent->flags & XFS_ATTR_LOCAL) {
+			lentry = xfs_attr3_leaf_name_local(leaf, i);
+			error = xrep_xattr_salvage_local_attr(rx, ent, nameidx,
+					buf_end, lentry);
+		} else {
+			rentry = xfs_attr3_leaf_name_remote(leaf, i);
+			error = xrep_xattr_salvage_remote_attr(rx, ent, nameidx,
+					buf_end, rentry, i, bp);
+		}
+		if (error)
+			return error;
+	}
+
+	return 0;
+}
+
+/* Try to recover shortform attrs. */
+STATIC int
+xrep_xattr_recover_sf(
+	struct xrep_xattr		*rx)
+{
+	struct xfs_scrub		*sc = rx->sc;
+	struct xchk_xattr_buf		*ab = sc->buf;
+	struct xfs_attr_sf_hdr		*hdr;
+	struct xfs_attr_sf_entry	*sfe;
+	struct xfs_attr_sf_entry	*next;
+	struct xfs_ifork		*ifp;
+	unsigned char			*end;
+	int				i;
+	int				error = 0;
+
+	ifp = xfs_ifork_ptr(rx->sc->ip, XFS_ATTR_FORK);
+	hdr = ifp->if_data;
+
+	bitmap_zero(ab->usedmap, ifp->if_bytes);
+	end = (unsigned char *)ifp->if_data + ifp->if_bytes;
+	xchk_xattr_set_map(sc, ab->usedmap, 0, sizeof(*hdr));
+
+	sfe = xfs_attr_sf_firstentry(hdr);
+	if ((unsigned char *)sfe > end)
+		return 0;
+
+	for (i = 0; i < hdr->count; i++) {
+		if (xchk_should_terminate(sc, &error))
+			return error;
+
+		next = xfs_attr_sf_nextentry(sfe);
+		if ((unsigned char *)next > end)
+			break;
+
+		if (xchk_xattr_set_map(sc, ab->usedmap,
+				(char *)sfe - (char *)hdr,
+				sizeof(struct xfs_attr_sf_entry))) {
+			/*
+			 * No conflicts with the sf entry; let's save this
+			 * attribute.
+			 */
+			error = xrep_xattr_salvage_sf_attr(rx, hdr, sfe);
+			if (error)
+				return error;
+		}
+
+		sfe = next;
+	}
+
+	return 0;
+}
+
+/*
+ * Try to return a buffer of xattr data for a given physical extent.
+ *
+ * Because the buffer cache get function complains if it finds a buffer
+ * matching the block number but not matching the length, we must be careful to
+ * look for incore buffers (up to the maximum length of a remote value) that
+ * could be hiding anywhere in the physical range.  If we find an incore
+ * buffer, we can pass that to the caller.  Optionally, read a single block and
+ * pass that back.
+ *
+ * Note the subtlety that remote attr value blocks for which there is no incore
+ * buffer will be passed to the callback one block at a time.  These buffers
+ * will not have any ops attached and must be staled to prevent aliasing with
+ * multiblock buffers once we drop the ILOCK.
+ */
+STATIC int
+xrep_xattr_find_buf(
+	struct xfs_mount	*mp,
+	xfs_fsblock_t		fsbno,
+	xfs_extlen_t		max_len,
+	bool			can_read,
+	struct xfs_buf		**bpp)
+{
+	struct xrep_bufscan	scan = {
+		.daddr		= XFS_FSB_TO_DADDR(mp, fsbno),
+		.max_sectors	= xrep_bufscan_max_sectors(mp, max_len),
+		.daddr_step	= XFS_FSB_TO_BB(mp, 1),
+	};
+	struct xfs_buf		*bp;
+
+	while ((bp = xrep_bufscan_advance(mp, &scan)) != NULL) {
+		*bpp = bp;
+		return 0;
+	}
+
+	if (!can_read) {
+		*bpp = NULL;
+		return 0;
+	}
+
+	return xfs_buf_read(mp->m_ddev_targp, scan.daddr, XFS_FSB_TO_BB(mp, 1),
+			XBF_TRYLOCK, bpp, NULL);
+}
+
+/*
+ * Deal with a buffer that we found during our walk of the attr fork.
+ *
+ * Attribute leaf and node blocks are simple -- they're a single block, so we
+ * can walk them one at a time and we never have to worry about discontiguous
+ * multiblock buffers like we do for directories.
+ *
+ * Unfortunately, remote attr blocks add a lot of complexity here.  Each disk
+ * block is totally self contained, in the sense that the v5 header provides no
+ * indication that there could be more data in the next block.  The incore
+ * buffers can span multiple blocks, though they never cross extent records.
+ * However, they don't necessarily start or end on an extent record boundary.
+ * Therefore, we need a special buffer find function to walk the buffer cache
+ * for us.
+ *
+ * The caller must hold the ILOCK on the file being repaired.  We use
+ * XBF_TRYLOCK here to skip any locked buffer on the assumption that we don't
+ * own the block and don't want to hang the system on a potentially garbage
+ * buffer.
+ */
+STATIC int
+xrep_xattr_recover_block(
+	struct xrep_xattr	*rx,
+	xfs_dablk_t		dabno,
+	xfs_fsblock_t		fsbno,
+	xfs_extlen_t		max_len,
+	xfs_extlen_t		*actual_len)
+{
+	struct xfs_da_blkinfo	*info;
+	struct xfs_buf		*bp;
+	int			error;
+
+	error = xrep_xattr_find_buf(rx->sc->mp, fsbno, max_len, true, &bp);
+	if (error)
+		return error;
+	info = bp->b_addr;
+	*actual_len = XFS_BB_TO_FSB(rx->sc->mp, bp->b_length);
+
+	trace_xrep_xattr_recover_leafblock(rx->sc->ip, dabno,
+			be16_to_cpu(info->magic));
+
+	/*
+	 * If the buffer has the right magic number for an attr leaf block and
+	 * passes a structure check (we don't care about checksums), salvage
+	 * as much as we can from the block. */
+	if (info->magic == cpu_to_be16(XFS_ATTR3_LEAF_MAGIC) &&
+	    xrep_buf_verify_struct(bp, &xfs_attr3_leaf_buf_ops) &&
+	    xfs_attr3_leaf_header_check(bp, rx->sc->ip->i_ino) == NULL)
+		error = xrep_xattr_recover_leaf(rx, bp);
+
+	/*
+	 * If the buffer didn't already have buffer ops set, it was read in by
+	 * the _find_buf function and could very well be /part/ of a multiblock
+	 * remote block.  Mark it stale so that it doesn't hang around in
+	 * memory to cause problems.
+	 */
+	if (bp->b_ops == NULL)
+		xfs_buf_stale(bp);
+
+	xfs_buf_relse(bp);
+	return error;
+}
+
+/* Insert one xattr key/value. */
+STATIC int
+xrep_xattr_insert_rec(
+	struct xrep_xattr		*rx,
+	const struct xrep_xattr_key	*key)
+{
+	struct xfs_da_args		args = {
+		.dp			= rx->sc->tempip,
+		.attr_filter		= key->flags,
+		.namelen		= key->namelen,
+		.valuelen		= key->valuelen,
+		.owner			= rx->sc->ip->i_ino,
+		.geo			= rx->sc->mp->m_attr_geo,
+		.whichfork		= XFS_ATTR_FORK,
+		.op_flags		= XFS_DA_OP_OKNOENT,
+	};
+	struct xchk_xattr_buf		*ab = rx->sc->buf;
+	int				error;
+
+	/*
+	 * Grab pointers to the scrub buffer so that we can use them to insert
+	 * attrs into the temp file.
+	 */
+	args.name = ab->name;
+	args.value = ab->value;
+
+	/*
+	 * The attribute name is stored near the end of the in-core buffer,
+	 * though we reserve one more byte to ensure null termination.
+	 */
+	ab->name[XATTR_NAME_MAX] = 0;
+
+	error = xfblob_load(rx->xattr_blobs, key->name_cookie, ab->name,
+			key->namelen);
+	if (error)
+		return error;
+
+	error = xfblob_free(rx->xattr_blobs, key->name_cookie);
+	if (error)
+		return error;
+
+	error = xfblob_load(rx->xattr_blobs, key->value_cookie, args.value,
+			key->valuelen);
+	if (error)
+		return error;
+
+	error = xfblob_free(rx->xattr_blobs, key->value_cookie);
+	if (error)
+		return error;
+
+	ab->name[key->namelen] = 0;
+
+	if (key->flags & XFS_ATTR_PARENT) {
+		trace_xrep_xattr_insert_pptr(rx->sc->tempip, key->flags,
+				ab->name, key->namelen, ab->value,
+				key->valuelen);
+		args.op_flags |= XFS_DA_OP_LOGGED;
+	} else {
+		trace_xrep_xattr_insert_rec(rx->sc->tempip, key->flags,
+				ab->name, key->namelen, key->valuelen);
+	}
+
+	/*
+	 * xfs_attr_set creates and commits its own transaction.  If the attr
+	 * already exists, we'll just drop it during the rebuild.
+	 */
+	xfs_attr_sethash(&args);
+	error = xfs_attr_set(&args, XFS_ATTRUPDATE_CREATE, false);
+	if (error == -EEXIST)
+		error = 0;
+
+	return error;
+}
+
+/*
+ * Periodically flush salvaged attributes to the temporary file.  This is done
+ * to reduce the memory requirements of the xattr rebuild because files can
+ * contain millions of attributes.
+ */
+STATIC int
+xrep_xattr_flush_stashed(
+	struct xrep_xattr	*rx)
+{
+	xfarray_idx_t		array_cur;
+	int			error;
+
+	/*
+	 * Entering this function, the scrub context has a reference to the
+	 * inode being repaired, the temporary file, and a scrub transaction
+	 * that we use during xattr salvaging to avoid livelocking if there
+	 * are cycles in the xattr structures.  We hold ILOCK_EXCL on both
+	 * the inode being repaired, though it is not ijoined to the scrub
+	 * transaction.
+	 *
+	 * To constrain kernel memory use, we occasionally flush salvaged
+	 * xattrs from the xfarray and xfblob structures into the temporary
+	 * file in preparation for exchanging the xattr structures at the end.
+	 * Updating the temporary file requires a transaction, so we commit the
+	 * scrub transaction and drop the two ILOCKs so that xfs_attr_set can
+	 * allocate whatever transaction it wants.
+	 *
+	 * We still hold IOLOCK_EXCL on the inode being repaired, which
+	 * prevents anyone from modifying the damaged xattr data while we
+	 * repair it.
+	 */
+	error = xrep_trans_commit(rx->sc);
+	if (error)
+		return error;
+	xchk_iunlock(rx->sc, XFS_ILOCK_EXCL);
+
+	/*
+	 * Take the IOLOCK of the temporary file while we modify xattrs.  This
+	 * isn't strictly required because the temporary file is never revealed
+	 * to userspace, but we follow the same locking rules.  We still hold
+	 * sc->ip's IOLOCK.
+	 */
+	error = xrep_tempfile_iolock_polled(rx->sc);
+	if (error)
+		return error;
+
+	/* Add all the salvaged attrs to the temporary file. */
+	foreach_xfarray_idx(rx->xattr_records, array_cur) {
+		struct xrep_xattr_key	key;
+
+		error = xfarray_load(rx->xattr_records, array_cur, &key);
+		if (error)
+			return error;
+
+		error = xrep_xattr_insert_rec(rx, &key);
+		if (error)
+			return error;
+	}
+
+	/* Empty out both arrays now that we've added the entries. */
+	xfarray_truncate(rx->xattr_records);
+	xfblob_truncate(rx->xattr_blobs);
+
+	xrep_tempfile_iounlock(rx->sc);
+
+	/* Recreate the salvage transaction and relock the inode. */
+	error = xchk_trans_alloc(rx->sc, 0);
+	if (error)
+		return error;
+	xchk_ilock(rx->sc, XFS_ILOCK_EXCL);
+	return 0;
+}
+
+/* Decide if we've stashed too much xattr data in memory. */
+static inline bool
+xrep_xattr_want_flush_stashed(
+	struct xrep_xattr	*rx)
+{
+	unsigned long long	bytes;
+
+	if (!rx->can_flush)
+		return false;
+
+	bytes = xfarray_bytes(rx->xattr_records) +
+		xfblob_bytes(rx->xattr_blobs);
+	return bytes > XREP_XATTR_MAX_STASH_BYTES;
+}
+
+/*
+ * Did we observe rename changing parent pointer xattrs while we were flushing
+ * salvaged attrs?
+ */
+static inline bool
+xrep_xattr_saw_pptr_conflict(
+	struct xrep_xattr	*rx)
+{
+	bool			ret;
+
+	ASSERT(rx->can_flush);
+
+	if (!xfs_has_parent(rx->sc->mp))
+		return false;
+
+	xfs_assert_ilocked(rx->sc->ip, XFS_ILOCK_EXCL);
+
+	mutex_lock(&rx->lock);
+	ret = xfarray_bytes(rx->pptr_recs) > 0;
+	mutex_unlock(&rx->lock);
+
+	return ret;
+}
+
+/*
+ * Reset the entire repair state back to initial conditions, now that we've
+ * detected a parent pointer update to the attr structure while we were
+ * flushing salvaged attrs.  See the locking notes in dir_repair.c for more
+ * information on why this is all necessary.
+ */
+STATIC int
+xrep_xattr_full_reset(
+	struct xrep_xattr	*rx)
+{
+	struct xfs_scrub	*sc = rx->sc;
+	struct xfs_attr_sf_hdr	*hdr;
+	struct xfs_ifork	*ifp = &sc->tempip->i_af;
+	int			error;
+
+	trace_xrep_xattr_full_reset(sc->ip, sc->tempip);
+
+	/* The temporary file's data fork had better not be in btree format. */
+	if (sc->tempip->i_df.if_format == XFS_DINODE_FMT_BTREE) {
+		ASSERT(0);
+		return -EIO;
+	}
+
+	/*
+	 * We begin in transaction context with sc->ip ILOCKed but not joined
+	 * to the transaction.  To reset to the initial state, we must hold
+	 * sc->ip's ILOCK to prevent rename from updating parent pointer
+	 * information and the tempfile's ILOCK to clear its contents.
+	 */
+	xchk_iunlock(rx->sc, XFS_ILOCK_EXCL);
+	xrep_tempfile_ilock_both(sc);
+	xfs_trans_ijoin(sc->tp, sc->ip, 0);
+	xfs_trans_ijoin(sc->tp, sc->tempip, 0);
+
+	/*
+	 * Free all the blocks of the attr fork of the temp file, and reset
+	 * it back to local format.
+	 */
+	if (xfs_ifork_has_extents(&sc->tempip->i_af)) {
+		error = xrep_reap_ifork(sc, sc->tempip, XFS_ATTR_FORK);
+		if (error)
+			return error;
+
+		ASSERT(ifp->if_bytes == 0);
+		ifp->if_format = XFS_DINODE_FMT_LOCAL;
+		xfs_idata_realloc(sc->tempip, sizeof(*hdr), XFS_ATTR_FORK);
+	}
+
+	/* Reinitialize the attr fork to an empty shortform structure. */
+	hdr = ifp->if_data;
+	memset(hdr, 0, sizeof(*hdr));
+	hdr->totsize = cpu_to_be16(sizeof(*hdr));
+	xfs_trans_log_inode(sc->tp, sc->tempip, XFS_ILOG_CORE | XFS_ILOG_ADATA);
+
+	/*
+	 * Roll this transaction to commit our reset ondisk.  The tempfile
+	 * should no longer be joined to the transaction, so we drop its ILOCK.
+	 * This should leave us in transaction context with sc->ip ILOCKed but
+	 * not joined to the transaction.
+	 */
+	error = xrep_roll_trans(sc);
+	if (error)
+		return error;
+	xrep_tempfile_iunlock(sc);
+
+	/*
+	 * Erase any accumulated parent pointer updates now that we've erased
+	 * the tempfile's attr fork.  We're resetting the entire repair state
+	 * back to where we were initially, except now we won't flush salvaged
+	 * xattrs until the very end.
+	 */
+	mutex_lock(&rx->lock);
+	xfarray_truncate(rx->pptr_recs);
+	xfblob_truncate(rx->pptr_names);
+	mutex_unlock(&rx->lock);
+
+	rx->can_flush = false;
+	rx->attrs_found = 0;
+
+	ASSERT(xfarray_bytes(rx->xattr_records) == 0);
+	ASSERT(xfblob_bytes(rx->xattr_blobs) == 0);
+	return 0;
+}
+
+/* Extract as many attribute keys and values as we can. */
+STATIC int
+xrep_xattr_recover(
+	struct xrep_xattr	*rx)
+{
+	struct xfs_bmbt_irec	got;
+	struct xfs_scrub	*sc = rx->sc;
+	struct xfs_da_geometry	*geo = sc->mp->m_attr_geo;
+	xfs_fileoff_t		offset;
+	xfs_extlen_t		len;
+	xfs_dablk_t		dabno;
+	int			nmap;
+	int			error;
+
+restart:
+	/*
+	 * Iterate each xattr leaf block in the attr fork to scan them for any
+	 * attributes that we might salvage.
+	 */
+	for (offset = 0;
+	     offset < XFS_MAX_FILEOFF;
+	     offset = got.br_startoff + got.br_blockcount) {
+		nmap = 1;
+		error = xfs_bmapi_read(sc->ip, offset, XFS_MAX_FILEOFF - offset,
+				&got, &nmap, XFS_BMAPI_ATTRFORK);
+		if (error)
+			return error;
+		if (nmap != 1)
+			return -EFSCORRUPTED;
+		if (!xfs_bmap_is_written_extent(&got))
+			continue;
+
+		for (dabno = round_up(got.br_startoff, geo->fsbcount);
+		     dabno < got.br_startoff + got.br_blockcount;
+		     dabno += len) {
+			xfs_fileoff_t	curr_offset = dabno - got.br_startoff;
+			xfs_extlen_t	maxlen;
+
+			if (xchk_should_terminate(rx->sc, &error))
+				return error;
+
+			maxlen = min_t(xfs_filblks_t, INT_MAX,
+					got.br_blockcount - curr_offset);
+			error = xrep_xattr_recover_block(rx, dabno,
+					curr_offset + got.br_startblock,
+					maxlen, &len);
+			if (error)
+				return error;
+
+			if (xrep_xattr_want_flush_stashed(rx)) {
+				error = xrep_xattr_flush_stashed(rx);
+				if (error)
+					return error;
+
+				if (xrep_xattr_saw_pptr_conflict(rx)) {
+					error = xrep_xattr_full_reset(rx);
+					if (error)
+						return error;
+
+					goto restart;
+				}
+			}
+		}
+	}
+
+	return 0;
+}
+
+/*
+ * Reset the extended attribute fork to a state where we can start re-adding
+ * the salvaged attributes.
+ */
+STATIC int
+xrep_xattr_fork_remove(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*ip)
+{
+	struct xfs_attr_sf_hdr	*hdr;
+	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, XFS_ATTR_FORK);
+
+	/*
+	 * If the data fork is in btree format, we can't change di_forkoff
+	 * because we could run afoul of the rule that the data fork isn't
+	 * supposed to be in btree format if there's enough space in the fork
+	 * that it could have used extents format.  Instead, reinitialize the
+	 * attr fork to have a shortform structure with zero attributes.
+	 */
+	if (ip->i_df.if_format == XFS_DINODE_FMT_BTREE) {
+		ifp->if_format = XFS_DINODE_FMT_LOCAL;
+		hdr = xfs_idata_realloc(ip, (int)sizeof(*hdr) - ifp->if_bytes,
+				XFS_ATTR_FORK);
+		hdr->count = 0;
+		hdr->totsize = cpu_to_be16(sizeof(*hdr));
+		xfs_trans_log_inode(sc->tp, ip,
+				XFS_ILOG_CORE | XFS_ILOG_ADATA);
+		return 0;
+	}
+
+	/* If we still have attr fork extents, something's wrong. */
+	if (ifp->if_nextents != 0) {
+		struct xfs_iext_cursor	icur;
+		struct xfs_bmbt_irec	irec;
+		unsigned int		i = 0;
+
+		xfs_emerg(sc->mp,
+	"inode 0x%llx attr fork still has %llu attr extents, format %d?!",
+				ip->i_ino, ifp->if_nextents, ifp->if_format);
+		for_each_xfs_iext(ifp, &icur, &irec) {
+			xfs_err(sc->mp,
+	"[%u]: startoff %llu startblock %llu blockcount %llu state %u",
+					i++, irec.br_startoff,
+					irec.br_startblock, irec.br_blockcount,
+					irec.br_state);
+		}
+		ASSERT(0);
+		return -EFSCORRUPTED;
+	}
+
+	xfs_attr_fork_remove(ip, sc->tp);
+	return 0;
+}
+
+/*
+ * Free all the attribute fork blocks of the file being repaired and delete the
+ * fork.  The caller must ILOCK the scrub file and join it to the transaction.
+ * This function returns with the inode joined to a clean transaction.
+ */
+int
+xrep_xattr_reset_fork(
+	struct xfs_scrub	*sc)
+{
+	int			error;
+
+	trace_xrep_xattr_reset_fork(sc->ip, sc->ip);
+
+	/* Unmap all the attr blocks. */
+	if (xfs_ifork_has_extents(&sc->ip->i_af)) {
+		error = xrep_reap_ifork(sc, sc->ip, XFS_ATTR_FORK);
+		if (error)
+			return error;
+	}
+
+	error = xrep_xattr_fork_remove(sc, sc->ip);
+	if (error)
+		return error;
+
+	return xfs_trans_roll_inode(&sc->tp, sc->ip);
+}
+
+/*
+ * Free all the attribute fork blocks of the temporary file and delete the attr
+ * fork.  The caller must ILOCK the tempfile and join it to the transaction.
+ * This function returns with the inode joined to a clean scrub transaction.
+ */
+int
+xrep_xattr_reset_tempfile_fork(
+	struct xfs_scrub	*sc)
+{
+	int			error;
+
+	trace_xrep_xattr_reset_fork(sc->ip, sc->tempip);
+
+	/*
+	 * Wipe out the attr fork of the temp file so that regular inode
+	 * inactivation won't trip over the corrupt attr fork.
+	 */
+	if (xfs_ifork_has_extents(&sc->tempip->i_af)) {
+		error = xrep_reap_ifork(sc, sc->tempip, XFS_ATTR_FORK);
+		if (error)
+			return error;
+	}
+
+	return xrep_xattr_fork_remove(sc, sc->tempip);
+}
+
+/*
+ * Find all the extended attributes for this inode by scraping them out of the
+ * attribute key blocks by hand, and flushing them into the temp file.
+ * When we're done, free the staging memory before exchanging the xattr
+ * structures to reduce memory usage.
+ */
+STATIC int
+xrep_xattr_salvage_attributes(
+	struct xrep_xattr	*rx)
+{
+	struct xfs_inode	*ip = rx->sc->ip;
+	int			error;
+
+	/* Short format xattrs are easy! */
+	if (rx->sc->ip->i_af.if_format == XFS_DINODE_FMT_LOCAL) {
+		error = xrep_xattr_recover_sf(rx);
+		if (error)
+			return error;
+
+		return xrep_xattr_flush_stashed(rx);
+	}
+
+	/*
+	 * For non-inline xattr structures, the salvage function scans the
+	 * buffer cache looking for potential attr leaf blocks.  The scan
+	 * requires the ability to lock any buffer found and runs independently
+	 * of any transaction <-> buffer item <-> buffer linkage.  Therefore,
+	 * roll the transaction to ensure there are no buffers joined.  We hold
+	 * the ILOCK independently of the transaction.
+	 */
+	error = xfs_trans_roll(&rx->sc->tp);
+	if (error)
+		return error;
+
+	error = xfs_iread_extents(rx->sc->tp, ip, XFS_ATTR_FORK);
+	if (error)
+		return error;
+
+	error = xrep_xattr_recover(rx);
+	if (error)
+		return error;
+
+	return xrep_xattr_flush_stashed(rx);
+}
+
+/*
+ * Add this stashed incore parent pointer to the temporary file.  The caller
+ * must hold the tempdir's IOLOCK, must not hold any ILOCKs, and must not be in
+ * transaction context.
+ */
+STATIC int
+xrep_xattr_replay_pptr_update(
+	struct xrep_xattr		*rx,
+	const struct xfs_name		*xname,
+	struct xrep_xattr_pptr		*pptr)
+{
+	struct xfs_scrub		*sc = rx->sc;
+	int				error;
+
+	switch (pptr->action) {
+	case XREP_XATTR_PPTR_ADD:
+		/* Create parent pointer. */
+		trace_xrep_xattr_replay_parentadd(sc->tempip, xname,
+				&pptr->pptr_rec);
+
+		error = xfs_parent_set(sc->tempip, sc->ip->i_ino, xname,
+				&pptr->pptr_rec, &rx->pptr_args);
+		ASSERT(error != -EEXIST);
+		return error;
+	case XREP_XATTR_PPTR_REMOVE:
+		/* Remove parent pointer. */
+		trace_xrep_xattr_replay_parentremove(sc->tempip, xname,
+				&pptr->pptr_rec);
+
+		error = xfs_parent_unset(sc->tempip, sc->ip->i_ino, xname,
+				&pptr->pptr_rec, &rx->pptr_args);
+		ASSERT(error != -ENOATTR);
+		return error;
+	}
+
+	ASSERT(0);
+	return -EIO;
+}
+
+/*
+ * Flush stashed parent pointer updates that have been recorded by the scanner.
+ * This is done to reduce the memory requirements of the xattr rebuild, since
+ * files can have a lot of hardlinks and the fs can be busy.
+ *
+ * Caller must not hold transactions or ILOCKs.  Caller must hold the tempfile
+ * IOLOCK.
+ */
+STATIC int
+xrep_xattr_replay_pptr_updates(
+	struct xrep_xattr	*rx)
+{
+	xfarray_idx_t		array_cur;
+	int			error;
+
+	mutex_lock(&rx->lock);
+	foreach_xfarray_idx(rx->pptr_recs, array_cur) {
+		struct xrep_xattr_pptr	pptr;
+
+		error = xfarray_load(rx->pptr_recs, array_cur, &pptr);
+		if (error)
+			goto out_unlock;
+
+		error = xfblob_loadname(rx->pptr_names, pptr.name_cookie,
+				&rx->xname, pptr.namelen);
+		if (error)
+			goto out_unlock;
+		mutex_unlock(&rx->lock);
+
+		error = xrep_xattr_replay_pptr_update(rx, &rx->xname, &pptr);
+		if (error)
+			return error;
+
+		mutex_lock(&rx->lock);
+	}
+
+	/* Empty out both arrays now that we've added the entries. */
+	xfarray_truncate(rx->pptr_recs);
+	xfblob_truncate(rx->pptr_names);
+	mutex_unlock(&rx->lock);
+	return 0;
+out_unlock:
+	mutex_unlock(&rx->lock);
+	return error;
+}
+
+/*
+ * Remember that we want to create a parent pointer in the tempfile.  These
+ * stashed actions will be replayed later.
+ */
+STATIC int
+xrep_xattr_stash_parentadd(
+	struct xrep_xattr	*rx,
+	const struct xfs_name	*name,
+	const struct xfs_inode	*dp)
+{
+	struct xrep_xattr_pptr	pptr = {
+		.action		= XREP_XATTR_PPTR_ADD,
+		.namelen	= name->len,
+	};
+	int			error;
+
+	trace_xrep_xattr_stash_parentadd(rx->sc->tempip, dp, name);
+
+	xfs_inode_to_parent_rec(&pptr.pptr_rec, dp);
+	error = xfblob_storename(rx->pptr_names, &pptr.name_cookie, name);
+	if (error)
+		return error;
+
+	return xfarray_append(rx->pptr_recs, &pptr);
+}
+
+/*
+ * Remember that we want to remove a parent pointer from the tempfile.  These
+ * stashed actions will be replayed later.
+ */
+STATIC int
+xrep_xattr_stash_parentremove(
+	struct xrep_xattr	*rx,
+	const struct xfs_name	*name,
+	const struct xfs_inode	*dp)
+{
+	struct xrep_xattr_pptr	pptr = {
+		.action		= XREP_XATTR_PPTR_REMOVE,
+		.namelen	= name->len,
+	};
+	int			error;
+
+	trace_xrep_xattr_stash_parentremove(rx->sc->tempip, dp, name);
+
+	xfs_inode_to_parent_rec(&pptr.pptr_rec, dp);
+	error = xfblob_storename(rx->pptr_names, &pptr.name_cookie, name);
+	if (error)
+		return error;
+
+	return xfarray_append(rx->pptr_recs, &pptr);
+}
+
+/*
+ * Capture dirent updates being made by other threads.  We will have to replay
+ * the parent pointer updates before exchanging attr forks.
+ */
+STATIC int
+xrep_xattr_live_dirent_update(
+	struct notifier_block		*nb,
+	unsigned long			action,
+	void				*data)
+{
+	struct xfs_dir_update_params	*p = data;
+	struct xrep_xattr		*rx;
+	struct xfs_scrub		*sc;
+	int				error;
+
+	rx = container_of(nb, struct xrep_xattr, dhook.dirent_hook.nb);
+	sc = rx->sc;
+
+	/*
+	 * This thread updated a dirent that points to the file that we're
+	 * repairing, so stash the update for replay against the temporary
+	 * file.
+	 */
+	if (p->ip->i_ino != sc->ip->i_ino)
+		return NOTIFY_DONE;
+
+	mutex_lock(&rx->lock);
+	if (p->delta > 0)
+		error = xrep_xattr_stash_parentadd(rx, p->name, p->dp);
+	else
+		error = xrep_xattr_stash_parentremove(rx, p->name, p->dp);
+	if (error)
+		rx->live_update_aborted = true;
+	mutex_unlock(&rx->lock);
+	return NOTIFY_DONE;
+}
+
+/*
+ * Prepare both inodes' attribute forks for an exchange.  Promote the tempfile
+ * from short format to leaf format, and if the file being repaired has a short
+ * format attr fork, turn it into an empty extent list.
+ */
+STATIC int
+xrep_xattr_swap_prep(
+	struct xfs_scrub	*sc,
+	bool			temp_local,
+	bool			ip_local)
+{
+	int			error;
+
+	/*
+	 * If the tempfile's attributes are in shortform format, convert that
+	 * to a single leaf extent so that we can use the atomic mapping
+	 * exchange.
+	 */
+	if (temp_local) {
+		struct xfs_da_args	args = {
+			.dp		= sc->tempip,
+			.geo		= sc->mp->m_attr_geo,
+			.whichfork	= XFS_ATTR_FORK,
+			.trans		= sc->tp,
+			.total		= 1,
+			.owner		= sc->ip->i_ino,
+		};
+
+		error = xfs_attr_shortform_to_leaf(&args);
+		if (error)
+			return error;
+
+		/*
+		 * Roll the deferred log items to get us back to a clean
+		 * transaction.
+		 */
+		error = xfs_defer_finish(&sc->tp);
+		if (error)
+			return error;
+	}
+
+	/*
+	 * If the file being repaired had a shortform attribute fork, convert
+	 * that to an empty extent list in preparation for the atomic mapping
+	 * exchange.
+	 */
+	if (ip_local) {
+		struct xfs_ifork	*ifp;
+
+		ifp = xfs_ifork_ptr(sc->ip, XFS_ATTR_FORK);
+
+		xfs_idestroy_fork(ifp);
+		ifp->if_format = XFS_DINODE_FMT_EXTENTS;
+		ifp->if_nextents = 0;
+		ifp->if_bytes = 0;
+		ifp->if_data = NULL;
+		ifp->if_height = 0;
+
+		xfs_trans_log_inode(sc->tp, sc->ip,
+				XFS_ILOG_CORE | XFS_ILOG_ADATA);
+	}
+
+	return 0;
+}
+
+/* Exchange the temporary file's attribute fork with the one being repaired. */
+int
+xrep_xattr_swap(
+	struct xfs_scrub	*sc,
+	struct xrep_tempexch	*tx)
+{
+	bool			ip_local, temp_local;
+	int			error = 0;
+
+	ip_local = sc->ip->i_af.if_format == XFS_DINODE_FMT_LOCAL;
+	temp_local = sc->tempip->i_af.if_format == XFS_DINODE_FMT_LOCAL;
+
+	/*
+	 * If the both files have a local format attr fork and the rebuilt
+	 * xattr data would fit in the repaired file's attr fork, just copy
+	 * the contents from the tempfile and declare ourselves done.
+	 */
+	if (ip_local && temp_local) {
+		int	forkoff;
+		int	newsize;
+
+		newsize = xfs_attr_sf_totsize(sc->tempip);
+		forkoff = xfs_attr_shortform_bytesfit(sc->ip, newsize);
+		if (forkoff > 0) {
+			sc->ip->i_forkoff = forkoff;
+			xrep_tempfile_copyout_local(sc, XFS_ATTR_FORK);
+			return 0;
+		}
+	}
+
+	/* Otherwise, make sure both attr forks are in block-mapping mode. */
+	error = xrep_xattr_swap_prep(sc, temp_local, ip_local);
+	if (error)
+		return error;
+
+	return xrep_tempexch_contents(sc, tx);
+}
+
+/*
+ * Finish replaying stashed parent pointer updates, allocate a transaction for
+ * exchanging extent mappings, and take the ILOCKs of both files before we
+ * commit the new extended attribute structure.
+ */
+STATIC int
+xrep_xattr_finalize_tempfile(
+	struct xrep_xattr	*rx)
+{
+	struct xfs_scrub	*sc = rx->sc;
+	int			error;
+
+	if (!xfs_has_parent(sc->mp))
+		return xrep_tempexch_trans_alloc(sc, XFS_ATTR_FORK, &rx->tx);
+
+	/*
+	 * Repair relies on the ILOCK to quiesce all possible xattr updates.
+	 * Replay all queued parent pointer updates into the tempfile before
+	 * exchanging the contents, even if that means dropping the ILOCKs and
+	 * the transaction.
+	 */
+	do {
+		error = xrep_xattr_replay_pptr_updates(rx);
+		if (error)
+			return error;
+
+		error = xrep_tempexch_trans_alloc(sc, XFS_ATTR_FORK, &rx->tx);
+		if (error)
+			return error;
+
+		if (xfarray_length(rx->pptr_recs) == 0)
+			break;
+
+		xchk_trans_cancel(sc);
+		xrep_tempfile_iunlock_both(sc);
+	} while (!xchk_should_terminate(sc, &error));
+	return error;
+}
+
+/*
+ * Exchange the new extended attribute data (which we created in the tempfile)
+ * with the file being repaired.
+ */
+STATIC int
+xrep_xattr_rebuild_tree(
+	struct xrep_xattr	*rx)
+{
+	struct xfs_scrub	*sc = rx->sc;
+	int			error;
+
+	/*
+	 * If we didn't find any attributes to salvage, repair the file by
+	 * zapping its attr fork.
+	 */
+	if (rx->attrs_found == 0) {
+		xfs_trans_ijoin(sc->tp, sc->ip, 0);
+		error = xrep_xattr_reset_fork(sc);
+		if (error)
+			return error;
+
+		goto forget_acls;
+	}
+
+	trace_xrep_xattr_rebuild_tree(sc->ip, sc->tempip);
+
+	/*
+	 * Commit the repair transaction and drop the ILOCKs so that we can use
+	 * the atomic file content exchange helper functions to compute the
+	 * correct resource reservations.
+	 *
+	 * We still hold IOLOCK_EXCL (aka i_rwsem) which will prevent xattr
+	 * modifications, but there's nothing to prevent userspace from reading
+	 * the attributes until we're ready for the exchange operation.  Reads
+	 * will return -EIO without shutting down the fs, so we're ok with
+	 * that.
+	 */
+	error = xrep_trans_commit(sc);
+	if (error)
+		return error;
+
+	xchk_iunlock(sc, XFS_ILOCK_EXCL);
+
+	/*
+	 * Take the IOLOCK on the temporary file so that we can run xattr
+	 * operations with the same locks held as we would for a normal file.
+	 * We still hold sc->ip's IOLOCK.
+	 */
+	error = xrep_tempfile_iolock_polled(rx->sc);
+	if (error)
+		return error;
+
+	/*
+	 * Allocate transaction, lock inodes, and make sure that we've replayed
+	 * all the stashed parent pointer updates to the temp file.  After this
+	 * point, we're ready to exchange attr fork mappings.
+	 */
+	error = xrep_xattr_finalize_tempfile(rx);
+	if (error)
+		return error;
+
+	/*
+	 * Exchange the blocks mapped by the tempfile's attr fork with the file
+	 * being repaired.  The old attr blocks will then be attached to the
+	 * tempfile, so reap its attr fork.
+	 */
+	error = xrep_xattr_swap(sc, &rx->tx);
+	if (error)
+		return error;
+
+	error = xrep_xattr_reset_tempfile_fork(sc);
+	if (error)
+		return error;
+
+	/*
+	 * Roll to get a transaction without any inodes joined to it.  Then we
+	 * can drop the tempfile's ILOCK and IOLOCK before doing more work on
+	 * the scrub target file.
+	 */
+	error = xfs_trans_roll(&sc->tp);
+	if (error)
+		return error;
+
+	xrep_tempfile_iunlock(sc);
+	xrep_tempfile_iounlock(sc);
+
+forget_acls:
+	/* Invalidate cached ACLs now that we've reloaded all the xattrs. */
+	xfs_forget_acl(VFS_I(sc->ip), SGI_ACL_FILE);
+	xfs_forget_acl(VFS_I(sc->ip), SGI_ACL_DEFAULT);
+	return 0;
+}
+
+/* Tear down all the incore scan stuff we created. */
+STATIC void
+xrep_xattr_teardown(
+	struct xrep_xattr	*rx)
+{
+	if (xfs_has_parent(rx->sc->mp))
+		xfs_dir_hook_del(rx->sc->mp, &rx->dhook);
+	if (rx->pptr_names)
+		xfblob_destroy(rx->pptr_names);
+	if (rx->pptr_recs)
+		xfarray_destroy(rx->pptr_recs);
+	xfblob_destroy(rx->xattr_blobs);
+	xfarray_destroy(rx->xattr_records);
+	mutex_destroy(&rx->lock);
+	kfree(rx);
+}
+
+/* Set up the filesystem scan so we can regenerate extended attributes. */
+STATIC int
+xrep_xattr_setup_scan(
+	struct xfs_scrub	*sc,
+	struct xrep_xattr	**rxp)
+{
+	struct xrep_xattr	*rx;
+	char			*descr;
+	int			max_len;
+	int			error;
+
+	rx = kzalloc(sizeof(struct xrep_xattr), XCHK_GFP_FLAGS);
+	if (!rx)
+		return -ENOMEM;
+	rx->sc = sc;
+	rx->can_flush = true;
+	rx->xname.name = rx->namebuf;
+
+	mutex_init(&rx->lock);
+
+	/*
+	 * Allocate enough memory to handle loading local attr values from the
+	 * xfblob data while flushing stashed attrs to the temporary file.
+	 * We only realloc the buffer when salvaging remote attr values.
+	 */
+	max_len = xfs_attr_leaf_entsize_local_max(sc->mp->m_attr_geo->blksize);
+	error = xchk_setup_xattr_buf(rx->sc, max_len);
+	if (error == -ENOMEM)
+		error = -EDEADLOCK;
+	if (error)
+		goto out_rx;
+
+	/* Set up some staging for salvaged attribute keys and values */
+	descr = xchk_xfile_ino_descr(sc, "xattr keys");
+	error = xfarray_create(descr, 0, sizeof(struct xrep_xattr_key),
+			&rx->xattr_records);
+	kfree(descr);
+	if (error)
+		goto out_rx;
+
+	descr = xchk_xfile_ino_descr(sc, "xattr names");
+	error = xfblob_create(descr, &rx->xattr_blobs);
+	kfree(descr);
+	if (error)
+		goto out_keys;
+
+	if (xfs_has_parent(sc->mp)) {
+		ASSERT(sc->flags & XCHK_FSGATES_DIRENTS);
+
+		descr = xchk_xfile_ino_descr(sc,
+				"xattr retained parent pointer entries");
+		error = xfarray_create(descr, 0,
+				sizeof(struct xrep_xattr_pptr),
+				&rx->pptr_recs);
+		kfree(descr);
+		if (error)
+			goto out_values;
+
+		descr = xchk_xfile_ino_descr(sc,
+				"xattr retained parent pointer names");
+		error = xfblob_create(descr, &rx->pptr_names);
+		kfree(descr);
+		if (error)
+			goto out_pprecs;
+
+		xfs_dir_hook_setup(&rx->dhook, xrep_xattr_live_dirent_update);
+		error = xfs_dir_hook_add(sc->mp, &rx->dhook);
+		if (error)
+			goto out_ppnames;
+	}
+
+	*rxp = rx;
+	return 0;
+out_ppnames:
+	xfblob_destroy(rx->pptr_names);
+out_pprecs:
+	xfarray_destroy(rx->pptr_recs);
+out_values:
+	xfblob_destroy(rx->xattr_blobs);
+out_keys:
+	xfarray_destroy(rx->xattr_records);
+out_rx:
+	mutex_destroy(&rx->lock);
+	kfree(rx);
+	return error;
+}
+
+/*
+ * Repair the extended attribute metadata.
+ *
+ * XXX: Remote attribute value buffers encompass the entire (up to 64k) buffer.
+ * The buffer cache in XFS can't handle aliased multiblock buffers, so this
+ * might misbehave if the attr fork is crosslinked with other filesystem
+ * metadata.
+ */
+int
+xrep_xattr(
+	struct xfs_scrub	*sc)
+{
+	struct xrep_xattr	*rx = NULL;
+	int			error;
+
+	if (!xfs_inode_hasattr(sc->ip))
+		return -ENOENT;
+
+	/* The rmapbt is required to reap the old attr fork. */
+	if (!xfs_has_rmapbt(sc->mp))
+		return -EOPNOTSUPP;
+	/* We require atomic file exchange range to rebuild anything. */
+	if (!xfs_has_exchange_range(sc->mp))
+		return -EOPNOTSUPP;
+
+	error = xrep_xattr_setup_scan(sc, &rx);
+	if (error)
+		return error;
+
+	ASSERT(sc->ilock_flags & XFS_ILOCK_EXCL);
+
+	error = xrep_xattr_salvage_attributes(rx);
+	if (error)
+		goto out_scan;
+
+	if (rx->live_update_aborted) {
+		error = -EIO;
+		goto out_scan;
+	}
+
+	/* Last chance to abort before we start committing fixes. */
+	if (xchk_should_terminate(sc, &error))
+		goto out_scan;
+
+	error = xrep_xattr_rebuild_tree(rx);
+	if (error)
+		goto out_scan;
+
+out_scan:
+	xrep_xattr_teardown(rx);
+	return error;
+}
diff --git a/fs/xfs/scrub/attr_repair.h b/fs/xfs/scrub/attr_repair.h
new file mode 100644
index 0000000000..979729bd4a
--- /dev/null
+++ b/fs/xfs/scrub/attr_repair.h
@@ -0,0 +1,15 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2018-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_ATTR_REPAIR_H__
+#define __XFS_SCRUB_ATTR_REPAIR_H__
+
+struct xrep_tempexch;
+
+int xrep_xattr_swap(struct xfs_scrub *sc, struct xrep_tempexch *tx);
+int xrep_xattr_reset_fork(struct xfs_scrub *sc);
+int xrep_xattr_reset_tempfile_fork(struct xfs_scrub *sc);
+
+#endif /* __XFS_SCRUB_ATTR_REPAIR_H__ */
diff --git a/fs/xfs/scrub/bitmap.c b/fs/xfs/scrub/bitmap.c
index 1449bb5262..7ba35a7a79 100644
--- a/fs/xfs/scrub/bitmap.c
+++ b/fs/xfs/scrub/bitmap.c
@@ -40,22 +40,23 @@ struct xbitmap64_node {
  * These functions are defined by the INTERVAL_TREE_DEFINE macro, but we'll
  * forward-declare them anyway for clarity.
  */
-static inline void
+static inline __maybe_unused void
 xbitmap64_tree_insert(struct xbitmap64_node *node, struct rb_root_cached *root);
 
-static inline void
+static inline __maybe_unused void
 xbitmap64_tree_remove(struct xbitmap64_node *node, struct rb_root_cached *root);
 
-static inline struct xbitmap64_node *
+static inline __maybe_unused struct xbitmap64_node *
 xbitmap64_tree_iter_first(struct rb_root_cached *root, uint64_t start,
 			uint64_t last);
 
-static inline struct xbitmap64_node *
+static inline __maybe_unused struct xbitmap64_node *
 xbitmap64_tree_iter_next(struct xbitmap64_node *node, uint64_t start,
 		       uint64_t last);
 
 INTERVAL_TREE_DEFINE(struct xbitmap64_node, bn_rbnode, uint64_t,
-		__bn_subtree_last, START, LAST, static inline, xbitmap64_tree)
+		__bn_subtree_last, START, LAST, static inline __maybe_unused,
+		xbitmap64_tree)
 
 /* Iterate each interval of a bitmap.  Do not change the bitmap. */
 #define for_each_xbitmap64_extent(bn, bitmap) \
@@ -314,22 +315,23 @@ struct xbitmap32_node {
  * These functions are defined by the INTERVAL_TREE_DEFINE macro, but we'll
  * forward-declare them anyway for clarity.
  */
-static inline void
+static inline __maybe_unused void
 xbitmap32_tree_insert(struct xbitmap32_node *node, struct rb_root_cached *root);
 
-static inline void
+static inline __maybe_unused void
 xbitmap32_tree_remove(struct xbitmap32_node *node, struct rb_root_cached *root);
 
-static inline struct xbitmap32_node *
+static inline __maybe_unused struct xbitmap32_node *
 xbitmap32_tree_iter_first(struct rb_root_cached *root, uint32_t start,
 			  uint32_t last);
 
-static inline struct xbitmap32_node *
+static inline __maybe_unused struct xbitmap32_node *
 xbitmap32_tree_iter_next(struct xbitmap32_node *node, uint32_t start,
 			 uint32_t last);
 
 INTERVAL_TREE_DEFINE(struct xbitmap32_node, bn_rbnode, uint32_t,
-		__bn_subtree_last, START, LAST, static inline, xbitmap32_tree)
+		__bn_subtree_last, START, LAST, static inline __maybe_unused,
+		xbitmap32_tree)
 
 /* Iterate each interval of a bitmap.  Do not change the bitmap. */
 #define for_each_xbitmap32_extent(bn, bitmap) \
@@ -566,3 +568,17 @@ xbitmap32_test(
 	*len = bn->bn_start - start;
 	return false;
 }
+
+/* Count the number of set regions in this bitmap. */
+uint32_t
+xbitmap32_count_set_regions(
+	struct xbitmap32	*bitmap)
+{
+	struct xbitmap32_node	*bn;
+	uint32_t		nr = 0;
+
+	for_each_xbitmap32_extent(bn, bitmap)
+		nr++;
+
+	return nr;
+}
diff --git a/fs/xfs/scrub/bitmap.h b/fs/xfs/scrub/bitmap.h
index 2df8911606..710c1ac5e3 100644
--- a/fs/xfs/scrub/bitmap.h
+++ b/fs/xfs/scrub/bitmap.h
@@ -62,4 +62,6 @@ int xbitmap32_walk(struct xbitmap32 *bitmap, xbitmap32_walk_fn fn,
 bool xbitmap32_empty(struct xbitmap32 *bitmap);
 bool xbitmap32_test(struct xbitmap32 *bitmap, uint32_t start, uint32_t *len);
 
+uint32_t xbitmap32_count_set_regions(struct xbitmap32 *bitmap);
+
 #endif	/* __XFS_SCRUB_BITMAP_H__ */
diff --git a/fs/xfs/scrub/bmap.c b/fs/xfs/scrub/bmap.c
index b169cddde6..24a15bf784 100644
--- a/fs/xfs/scrub/bmap.c
+++ b/fs/xfs/scrub/bmap.c
@@ -924,7 +924,7 @@ xchk_bmap(
 	if (!ifp)
 		return -ENOENT;
 
-	info.is_rt = whichfork == XFS_DATA_FORK && XFS_IS_REALTIME_INODE(ip);
+	info.is_rt = xfs_ifork_is_realtime(ip, whichfork);
 	info.whichfork = whichfork;
 	info.is_shared = whichfork == XFS_DATA_FORK && xfs_is_reflink_inode(ip);
 	info.sc = sc;
diff --git a/fs/xfs/scrub/bmap_repair.c b/fs/xfs/scrub/bmap_repair.c
index a4bb89fdd5..1e656fab5e 100644
--- a/fs/xfs/scrub/bmap_repair.c
+++ b/fs/xfs/scrub/bmap_repair.c
@@ -639,7 +639,13 @@ xrep_bmap_build_new_fork(
 	rb->new_bmapbt.bload.get_records = xrep_bmap_get_records;
 	rb->new_bmapbt.bload.claim_block = xrep_bmap_claim_block;
 	rb->new_bmapbt.bload.iroot_size = xrep_bmap_iroot_size;
-	bmap_cur = xfs_bmbt_stage_cursor(sc->mp, sc->ip, ifake);
+
+	/*
+	 * Allocate a new bmap btree cursor for reloading an inode block mapping
+	 * data structure.
+	 */
+	bmap_cur = xfs_bmbt_init_cursor(sc->mp, NULL, sc->ip, XFS_STAGING_FORK);
+	xfs_btree_stage_ifakeroot(bmap_cur, ifake);
 
 	/*
 	 * Figure out the size and format of the new fork, then fill it with
diff --git a/fs/xfs/scrub/btree.c b/fs/xfs/scrub/btree.c
index 1935b9ce18..fe678a0438 100644
--- a/fs/xfs/scrub/btree.c
+++ b/fs/xfs/scrub/btree.c
@@ -47,7 +47,7 @@ __xchk_btree_process_error(
 		*error = 0;
 		fallthrough;
 	default:
-		if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE)
+		if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE)
 			trace_xchk_ifork_btree_op_error(sc, cur, level,
 					*error, ret_ip);
 		else
@@ -91,7 +91,7 @@ __xchk_btree_set_corrupt(
 {
 	sc->sm->sm_flags |= errflag;
 
-	if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE)
+	if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE)
 		trace_xchk_ifork_btree_error(sc, cur, level,
 				ret_ip);
 	else
@@ -168,7 +168,7 @@ xchk_btree_rec(
 	if (xfs_btree_keycmp_lt(cur, &key, keyp))
 		xchk_btree_set_corrupt(bs->sc, cur, 1);
 
-	if (!(cur->bc_flags & XFS_BTREE_OVERLAPPING))
+	if (!(cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING))
 		return;
 
 	/* Is high_key(rec) no larger than the parent high key? */
@@ -215,7 +215,7 @@ xchk_btree_key(
 	if (xfs_btree_keycmp_lt(cur, key, keyp))
 		xchk_btree_set_corrupt(bs->sc, cur, level);
 
-	if (!(cur->bc_flags & XFS_BTREE_OVERLAPPING))
+	if (!(cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING))
 		return;
 
 	/* Is this block's high key no larger than the parent high key? */
@@ -236,22 +236,18 @@ xchk_btree_ptr_ok(
 	int			level,
 	union xfs_btree_ptr	*ptr)
 {
-	bool			res;
-
 	/* A btree rooted in an inode has no block pointer to the root. */
-	if ((bs->cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	if (bs->cur->bc_ops->type == XFS_BTREE_TYPE_INODE &&
 	    level == bs->cur->bc_nlevels)
 		return true;
 
 	/* Otherwise, check the pointers. */
-	if (bs->cur->bc_flags & XFS_BTREE_LONG_PTRS)
-		res = xfs_btree_check_lptr(bs->cur, be64_to_cpu(ptr->l), level);
-	else
-		res = xfs_btree_check_sptr(bs->cur, be32_to_cpu(ptr->s), level);
-	if (!res)
+	if (__xfs_btree_check_ptr(bs->cur, ptr, 0, level)) {
 		xchk_btree_set_corrupt(bs->sc, bs->cur, level);
+		return false;
+	}
 
-	return res;
+	return true;
 }
 
 /* Check that a btree block's sibling matches what we expect it. */
@@ -374,18 +370,21 @@ xchk_btree_check_block_owner(
 {
 	xfs_agnumber_t		agno;
 	xfs_agblock_t		agbno;
-	xfs_btnum_t		btnum;
 	bool			init_sa;
 	int			error = 0;
 
 	if (!bs->cur)
 		return 0;
 
-	btnum = bs->cur->bc_btnum;
 	agno = xfs_daddr_to_agno(bs->cur->bc_mp, daddr);
 	agbno = xfs_daddr_to_agbno(bs->cur->bc_mp, daddr);
 
-	init_sa = bs->cur->bc_flags & XFS_BTREE_LONG_PTRS;
+	/*
+	 * If the btree being examined is not itself a per-AG btree, initialize
+	 * sc->sa so that we can check for the presence of an ownership record
+	 * in the rmap btree for the AG containing the block.
+	 */
+	init_sa = bs->cur->bc_ops->type != XFS_BTREE_TYPE_AG;
 	if (init_sa) {
 		error = xchk_ag_init_existing(bs->sc, agno, &bs->sc->sa);
 		if (!xchk_btree_xref_process_error(bs->sc, bs->cur,
@@ -399,11 +398,11 @@ xchk_btree_check_block_owner(
 	 * have to nullify it (to shut down further block owner checks) if
 	 * self-xref encounters problems.
 	 */
-	if (!bs->sc->sa.bno_cur && btnum == XFS_BTNUM_BNO)
+	if (!bs->sc->sa.bno_cur && xfs_btree_is_bno(bs->cur->bc_ops))
 		bs->cur = NULL;
 
 	xchk_xref_is_only_owned_by(bs->sc, agbno, 1, bs->oinfo);
-	if (!bs->sc->sa.rmap_cur && btnum == XFS_BTNUM_RMAP)
+	if (!bs->sc->sa.rmap_cur && xfs_btree_is_rmap(bs->cur->bc_ops))
 		bs->cur = NULL;
 
 out_free:
@@ -429,7 +428,7 @@ xchk_btree_check_owner(
 	 * up.
 	 */
 	if (bp == NULL) {
-		if (!(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE))
+		if (cur->bc_ops->type != XFS_BTREE_TYPE_INODE)
 			xchk_btree_set_corrupt(bs->sc, bs->cur, level);
 		return 0;
 	}
@@ -442,7 +441,7 @@ xchk_btree_check_owner(
 	 * duplicate cursors.  Therefore, save the buffer daddr for
 	 * later scanning.
 	 */
-	if (cur->bc_btnum == XFS_BTNUM_BNO || cur->bc_btnum == XFS_BTNUM_RMAP) {
+	if (xfs_btree_is_bno(cur->bc_ops) || xfs_btree_is_rmap(cur->bc_ops)) {
 		struct check_owner	*co;
 
 		co = kmalloc(sizeof(struct check_owner), XCHK_GFP_FLAGS);
@@ -475,7 +474,7 @@ xchk_btree_check_iroot_minrecs(
 	 * existing filesystems, so instead we disable the check for data fork
 	 * bmap btrees when there's an attr fork.
 	 */
-	if (bs->cur->bc_btnum == XFS_BTNUM_BMAP &&
+	if (xfs_btree_is_bmap(bs->cur->bc_ops) &&
 	    bs->cur->bc_ino.whichfork == XFS_DATA_FORK &&
 	    xfs_inode_has_attr_fork(bs->sc->ip))
 		return false;
@@ -508,7 +507,7 @@ xchk_btree_check_minrecs(
 	 * child block might be less than the standard minrecs, but that's ok
 	 * provided that there's only one direct child of the root.
 	 */
-	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE &&
 	    level == cur->bc_nlevels - 2) {
 		struct xfs_btree_block	*root_block;
 		struct xfs_buf		*root_bp;
@@ -562,7 +561,7 @@ xchk_btree_block_check_keys(
 		return;
 	}
 
-	if (!(cur->bc_flags & XFS_BTREE_OVERLAPPING))
+	if (!(cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING))
 		return;
 
 	/* Make sure the high key of this block matches the parent. */
@@ -585,7 +584,6 @@ xchk_btree_get_block(
 	struct xfs_btree_block	**pblock,
 	struct xfs_buf		**pbp)
 {
-	xfs_failaddr_t		failed_at;
 	int			error;
 
 	*pblock = NULL;
@@ -597,13 +595,7 @@ xchk_btree_get_block(
 		return error;
 
 	xfs_btree_get_block(bs->cur, level, pbp);
-	if (bs->cur->bc_flags & XFS_BTREE_LONG_PTRS)
-		failed_at = __xfs_btree_check_lblock(bs->cur, *pblock,
-				level, *pbp);
-	else
-		failed_at = __xfs_btree_check_sblock(bs->cur, *pblock,
-				 level, *pbp);
-	if (failed_at) {
+	if (__xfs_btree_check_block(bs->cur, *pblock, level, *pbp)) {
 		xchk_btree_set_corrupt(bs->sc, bs->cur, level);
 		return 0;
 	}
@@ -664,7 +656,7 @@ xchk_btree_block_keys(
 	if (xfs_btree_keycmp_ne(cur, &block_keys, parent_keys))
 		xchk_btree_set_corrupt(bs->sc, cur, 1);
 
-	if (!(cur->bc_flags & XFS_BTREE_OVERLAPPING))
+	if (!(cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING))
 		return;
 
 	/* Get high keys */
@@ -728,7 +720,7 @@ xchk_btree(
 	 * error codes for us.
 	 */
 	level = cur->bc_nlevels - 1;
-	cur->bc_ops->init_ptr_from_cur(cur, &ptr);
+	xfs_btree_init_ptr_from_cur(cur, &ptr);
 	if (!xchk_btree_ptr_ok(bs, cur->bc_nlevels, &ptr))
 		goto out;
 	error = xchk_btree_get_block(bs, level, &ptr, &block, &bp);
diff --git a/fs/xfs/scrub/common.c b/fs/xfs/scrub/common.c
index 81f2b96bb5..1ad8ec63a7 100644
--- a/fs/xfs/scrub/common.c
+++ b/fs/xfs/scrub/common.c
@@ -29,6 +29,10 @@
 #include "xfs_attr.h"
 #include "xfs_reflink.h"
 #include "xfs_ag.h"
+#include "xfs_error.h"
+#include "xfs_quota.h"
+#include "xfs_exchmaps.h"
+#include "xfs_rtbitmap.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
@@ -82,6 +86,15 @@ __xchk_process_error(
 				sc->ip ? sc->ip : XFS_I(file_inode(sc->file)),
 				sc->sm, *error);
 		break;
+	case -ECANCELED:
+		/*
+		 * ECANCELED here means that the caller set one of the scrub
+		 * outcome flags (corrupt, xfail, xcorrupt) and wants to exit
+		 * quickly.  Set error to zero and do not continue.
+		 */
+		trace_xchk_op_error(sc, agno, bno, *error, ret_ip);
+		*error = 0;
+		break;
 	case -EFSBADCRC:
 	case -EFSCORRUPTED:
 		/* Note the badness but don't abort. */
@@ -89,8 +102,7 @@ __xchk_process_error(
 		*error = 0;
 		fallthrough;
 	default:
-		trace_xchk_op_error(sc, agno, bno, *error,
-				ret_ip);
+		trace_xchk_op_error(sc, agno, bno, *error, ret_ip);
 		break;
 	}
 	return false;
@@ -136,6 +148,16 @@ __xchk_fblock_process_error(
 		/* Used to restart an op with deadlock avoidance. */
 		trace_xchk_deadlock_retry(sc->ip, sc->sm, *error);
 		break;
+	case -ECANCELED:
+		/*
+		 * ECANCELED here means that the caller set one of the scrub
+		 * outcome flags (corrupt, xfail, xcorrupt) and wants to exit
+		 * quickly.  Set error to zero and do not continue.
+		 */
+		trace_xchk_file_op_error(sc, whichfork, offset, *error,
+				ret_ip);
+		*error = 0;
+		break;
 	case -EFSBADCRC:
 	case -EFSCORRUPTED:
 		/* Note the badness but don't abort. */
@@ -227,6 +249,19 @@ xchk_block_set_corrupt(
 	trace_xchk_block_error(sc, xfs_buf_daddr(bp), __return_address);
 }
 
+#ifdef CONFIG_XFS_QUOTA
+/* Record a corrupt quota counter. */
+void
+xchk_qcheck_set_corrupt(
+	struct xfs_scrub	*sc,
+	unsigned int		dqtype,
+	xfs_dqid_t		id)
+{
+	sc->sm->sm_flags |= XFS_SCRUB_OFLAG_CORRUPT;
+	trace_xchk_qcheck_error(sc, dqtype, id, __return_address);
+}
+#endif
+
 /* Record a corruption while cross-referencing. */
 void
 xchk_block_xref_set_corrupt(
@@ -412,7 +447,7 @@ xchk_perag_read_headers(
 {
 	int			error;
 
-	error = xfs_ialloc_read_agi(sa->pag, sc->tp, &sa->agi_bp);
+	error = xfs_ialloc_read_agi(sa->pag, sc->tp, 0, &sa->agi_bp);
 	if (error && want_ag_read_header_failure(sc, XFS_SCRUB_TYPE_AGI))
 		return error;
 
@@ -427,7 +462,7 @@ xchk_perag_read_headers(
  * Grab the AG headers for the attached perag structure and wait for pending
  * intents to drain.
  */
-static int
+int
 xchk_perag_drain_and_lock(
 	struct xfs_scrub	*sc)
 {
@@ -555,46 +590,50 @@ xchk_ag_btcur_init(
 {
 	struct xfs_mount	*mp = sc->mp;
 
-	if (sa->agf_bp &&
-	    xchk_ag_btree_healthy_enough(sc, sa->pag, XFS_BTNUM_BNO)) {
+	if (sa->agf_bp) {
 		/* Set up a bnobt cursor for cross-referencing. */
-		sa->bno_cur = xfs_allocbt_init_cursor(mp, sc->tp, sa->agf_bp,
-				sa->pag, XFS_BTNUM_BNO);
-	}
+		sa->bno_cur = xfs_bnobt_init_cursor(mp, sc->tp, sa->agf_bp,
+				sa->pag);
+		xchk_ag_btree_del_cursor_if_sick(sc, &sa->bno_cur,
+				XFS_SCRUB_TYPE_BNOBT);
 
-	if (sa->agf_bp &&
-	    xchk_ag_btree_healthy_enough(sc, sa->pag, XFS_BTNUM_CNT)) {
 		/* Set up a cntbt cursor for cross-referencing. */
-		sa->cnt_cur = xfs_allocbt_init_cursor(mp, sc->tp, sa->agf_bp,
-				sa->pag, XFS_BTNUM_CNT);
-	}
-
-	/* Set up a inobt cursor for cross-referencing. */
-	if (sa->agi_bp &&
-	    xchk_ag_btree_healthy_enough(sc, sa->pag, XFS_BTNUM_INO)) {
-		sa->ino_cur = xfs_inobt_init_cursor(sa->pag, sc->tp, sa->agi_bp,
-				XFS_BTNUM_INO);
-	}
-
-	/* Set up a finobt cursor for cross-referencing. */
-	if (sa->agi_bp && xfs_has_finobt(mp) &&
-	    xchk_ag_btree_healthy_enough(sc, sa->pag, XFS_BTNUM_FINO)) {
-		sa->fino_cur = xfs_inobt_init_cursor(sa->pag, sc->tp, sa->agi_bp,
-				XFS_BTNUM_FINO);
-	}
-
-	/* Set up a rmapbt cursor for cross-referencing. */
-	if (sa->agf_bp && xfs_has_rmapbt(mp) &&
-	    xchk_ag_btree_healthy_enough(sc, sa->pag, XFS_BTNUM_RMAP)) {
-		sa->rmap_cur = xfs_rmapbt_init_cursor(mp, sc->tp, sa->agf_bp,
+		sa->cnt_cur = xfs_cntbt_init_cursor(mp, sc->tp, sa->agf_bp,
 				sa->pag);
+		xchk_ag_btree_del_cursor_if_sick(sc, &sa->cnt_cur,
+				XFS_SCRUB_TYPE_CNTBT);
+
+		/* Set up a rmapbt cursor for cross-referencing. */
+		if (xfs_has_rmapbt(mp)) {
+			sa->rmap_cur = xfs_rmapbt_init_cursor(mp, sc->tp,
+					sa->agf_bp, sa->pag);
+			xchk_ag_btree_del_cursor_if_sick(sc, &sa->rmap_cur,
+					XFS_SCRUB_TYPE_RMAPBT);
+		}
+
+		/* Set up a refcountbt cursor for cross-referencing. */
+		if (xfs_has_reflink(mp)) {
+			sa->refc_cur = xfs_refcountbt_init_cursor(mp, sc->tp,
+					sa->agf_bp, sa->pag);
+			xchk_ag_btree_del_cursor_if_sick(sc, &sa->refc_cur,
+					XFS_SCRUB_TYPE_REFCNTBT);
+		}
 	}
 
-	/* Set up a refcountbt cursor for cross-referencing. */
-	if (sa->agf_bp && xfs_has_reflink(mp) &&
-	    xchk_ag_btree_healthy_enough(sc, sa->pag, XFS_BTNUM_REFC)) {
-		sa->refc_cur = xfs_refcountbt_init_cursor(mp, sc->tp,
-				sa->agf_bp, sa->pag);
+	if (sa->agi_bp) {
+		/* Set up a inobt cursor for cross-referencing. */
+		sa->ino_cur = xfs_inobt_init_cursor(sa->pag, sc->tp,
+				sa->agi_bp);
+		xchk_ag_btree_del_cursor_if_sick(sc, &sa->ino_cur,
+				XFS_SCRUB_TYPE_INOBT);
+
+		/* Set up a finobt cursor for cross-referencing. */
+		if (xfs_has_finobt(mp)) {
+			sa->fino_cur = xfs_finobt_init_cursor(sa->pag, sc->tp,
+					sa->agi_bp);
+			xchk_ag_btree_del_cursor_if_sick(sc, &sa->fino_cur,
+					XFS_SCRUB_TYPE_FINOBT);
+		}
 	}
 }
 
@@ -653,6 +692,13 @@ xchk_trans_cancel(
 	sc->tp = NULL;
 }
 
+int
+xchk_trans_alloc_empty(
+	struct xfs_scrub	*sc)
+{
+	return xfs_trans_alloc_empty(sc->mp, &sc->tp);
+}
+
 /*
  * Grab an empty transaction so that we can re-grab locked buffers if
  * one of our btrees turns out to be cyclic.
@@ -672,7 +718,7 @@ xchk_trans_alloc(
 		return xfs_trans_alloc(sc->mp, &M_RES(sc->mp)->tr_itruncate,
 				resblks, 0, 0, &sc->tp);
 
-	return xfs_trans_alloc_empty(sc->mp, &sc->tp);
+	return xchk_trans_alloc_empty(sc);
 }
 
 /* Set us up with a transaction and an empty context. */
@@ -737,7 +783,7 @@ xchk_iget(
 {
 	ASSERT(sc->tp != NULL);
 
-	return xfs_iget(sc->mp, sc->tp, inum, XFS_IGET_UNTRUSTED, 0, ipp);
+	return xfs_iget(sc->mp, sc->tp, inum, XCHK_IGET_FLAGS, 0, ipp);
 }
 
 /*
@@ -783,13 +829,13 @@ again:
 	 * in the iget cache miss path.
 	 */
 	pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, inum));
-	error = xfs_ialloc_read_agi(pag, tp, agi_bpp);
+	error = xfs_ialloc_read_agi(pag, tp, 0, agi_bpp);
 	xfs_perag_put(pag);
 	if (error)
 		return error;
 
-	error = xfs_iget(mp, tp, inum,
-			XFS_IGET_NORETRY | XFS_IGET_UNTRUSTED, 0, ipp);
+	error = xfs_iget(mp, tp, inum, XFS_IGET_NORETRY | XCHK_IGET_FLAGS, 0,
+			ipp);
 	if (error == -EAGAIN) {
 		/*
 		 * The inode may be in core but temporarily unavailable and may
@@ -1000,9 +1046,7 @@ xchk_irele(
 	struct xfs_scrub	*sc,
 	struct xfs_inode	*ip)
 {
-	if (current->journal_info != NULL) {
-		ASSERT(current->journal_info == sc->tp);
-
+	if (sc->tp) {
 		/*
 		 * If we are in a transaction, we /cannot/ drop the inode
 		 * ourselves, because the VFS will trigger writeback, which
@@ -1018,12 +1062,6 @@ xchk_irele(
 		spin_lock(&VFS_I(ip)->i_lock);
 		VFS_I(ip)->i_state &= ~I_DONTCACHE;
 		spin_unlock(&VFS_I(ip)->i_lock);
-	} else if (atomic_read(&VFS_I(ip)->i_count) == 1) {
-		/*
-		 * If this is the last reference to the inode and the caller
-		 * permits it, set DONTCACHE to avoid thrashing.
-		 */
-		d_mark_dontcache(VFS_I(ip));
 	}
 
 	xfs_irele(ip);
@@ -1160,27 +1198,12 @@ xchk_metadata_inode_subtype(
 	struct xfs_scrub	*sc,
 	unsigned int		scrub_type)
 {
-	__u32			smtype = sc->sm->sm_type;
-	unsigned int		sick_mask = sc->sick_mask;
+	struct xfs_scrub_subord	*sub;
 	int			error;
 
-	sc->sm->sm_type = scrub_type;
-
-	switch (scrub_type) {
-	case XFS_SCRUB_TYPE_INODE:
-		error = xchk_inode(sc);
-		break;
-	case XFS_SCRUB_TYPE_BMBTD:
-		error = xchk_bmap_data(sc);
-		break;
-	default:
-		ASSERT(0);
-		error = -EFSCORRUPTED;
-		break;
-	}
-
-	sc->sick_mask = sick_mask;
-	sc->sm->sm_type = smtype;
+	sub = xchk_scrub_create_subord(sc, scrub_type);
+	error = sub->sc.ops->scrub(&sub->sc);
+	xchk_scrub_free_subord(sub);
 	return error;
 }
 
@@ -1259,6 +1282,15 @@ xchk_fsgates_enable(
 	if (scrub_fsgates & XCHK_FSGATES_DRAIN)
 		xfs_drain_wait_enable();
 
+	if (scrub_fsgates & XCHK_FSGATES_QUOTA)
+		xfs_dqtrx_hook_enable();
+
+	if (scrub_fsgates & XCHK_FSGATES_DIRENTS)
+		xfs_dir_hook_enable();
+
+	if (scrub_fsgates & XCHK_FSGATES_RMAP)
+		xfs_rmap_hook_enable();
+
 	sc->flags |= scrub_fsgates;
 }
 
diff --git a/fs/xfs/scrub/common.h b/fs/xfs/scrub/common.h
index da09580b45..3d5f1f6b4b 100644
--- a/fs/xfs/scrub/common.h
+++ b/fs/xfs/scrub/common.h
@@ -6,32 +6,8 @@
 #ifndef __XFS_SCRUB_COMMON_H__
 #define __XFS_SCRUB_COMMON_H__
 
-/*
- * We /could/ terminate a scrub/repair operation early.  If we're not
- * in a good place to continue (fatal signal, etc.) then bail out.
- * Note that we're careful not to make any judgements about *error.
- */
-static inline bool
-xchk_should_terminate(
-	struct xfs_scrub	*sc,
-	int			*error)
-{
-	/*
-	 * If preemption is disabled, we need to yield to the scheduler every
-	 * few seconds so that we don't run afoul of the soft lockup watchdog
-	 * or RCU stall detector.
-	 */
-	cond_resched();
-
-	if (fatal_signal_pending(current)) {
-		if (*error == 0)
-			*error = -EINTR;
-		return true;
-	}
-	return false;
-}
-
 int xchk_trans_alloc(struct xfs_scrub *sc, uint resblks);
+int xchk_trans_alloc_empty(struct xfs_scrub *sc);
 void xchk_trans_cancel(struct xfs_scrub *sc);
 
 bool xchk_process_error(struct xfs_scrub *sc, xfs_agnumber_t agno,
@@ -54,6 +30,10 @@ void xchk_block_set_corrupt(struct xfs_scrub *sc,
 void xchk_ino_set_corrupt(struct xfs_scrub *sc, xfs_ino_t ino);
 void xchk_fblock_set_corrupt(struct xfs_scrub *sc, int whichfork,
 		xfs_fileoff_t offset);
+#ifdef CONFIG_XFS_QUOTA
+void xchk_qcheck_set_corrupt(struct xfs_scrub *sc, unsigned int dqtype,
+		xfs_dqid_t id);
+#endif
 
 void xchk_block_xref_set_corrupt(struct xfs_scrub *sc,
 		struct xfs_buf *bp);
@@ -87,6 +67,7 @@ int xchk_setup_directory(struct xfs_scrub *sc);
 int xchk_setup_xattr(struct xfs_scrub *sc);
 int xchk_setup_symlink(struct xfs_scrub *sc);
 int xchk_setup_parent(struct xfs_scrub *sc);
+int xchk_setup_dirtree(struct xfs_scrub *sc);
 #ifdef CONFIG_XFS_RT
 int xchk_setup_rtbitmap(struct xfs_scrub *sc);
 int xchk_setup_rtsummary(struct xfs_scrub *sc);
@@ -105,6 +86,7 @@ xchk_setup_rtsummary(struct xfs_scrub *sc)
 #ifdef CONFIG_XFS_QUOTA
 int xchk_ino_dqattach(struct xfs_scrub *sc);
 int xchk_setup_quota(struct xfs_scrub *sc);
+int xchk_setup_quotacheck(struct xfs_scrub *sc);
 #else
 static inline int
 xchk_ino_dqattach(struct xfs_scrub *sc)
@@ -116,12 +98,19 @@ xchk_setup_quota(struct xfs_scrub *sc)
 {
 	return -ENOENT;
 }
+static inline int
+xchk_setup_quotacheck(struct xfs_scrub *sc)
+{
+	return -ENOENT;
+}
 #endif
 int xchk_setup_fscounters(struct xfs_scrub *sc);
+int xchk_setup_nlinks(struct xfs_scrub *sc);
 
 void xchk_ag_free(struct xfs_scrub *sc, struct xchk_ag *sa);
 int xchk_ag_init(struct xfs_scrub *sc, xfs_agnumber_t agno,
 		struct xchk_ag *sa);
+int xchk_perag_drain_and_lock(struct xfs_scrub *sc);
 
 /*
  * Grab all AG resources, treating the inability to grab the perag structure as
@@ -199,6 +188,7 @@ static inline bool xchk_skip_xref(struct xfs_scrub_metadata *sm)
 }
 
 bool xchk_dir_looks_zapped(struct xfs_inode *dp);
+bool xchk_pptr_looks_zapped(struct xfs_inode *ip);
 
 #ifdef CONFIG_XFS_ONLINE_REPAIR
 /* Decide if a repair is required. */
diff --git a/fs/xfs/scrub/cow_repair.c b/fs/xfs/scrub/cow_repair.c
index 1e82c727af..4de3f0f40f 100644
--- a/fs/xfs/scrub/cow_repair.c
+++ b/fs/xfs/scrub/cow_repair.c
@@ -609,6 +609,6 @@ xrep_bmap_cow(
 out_bitmap:
 	xfsb_bitmap_destroy(&xc->old_cowfork_fsblocks);
 	xoff_bitmap_destroy(&xc->bad_fileoffs);
-	kmem_free(xc);
+	kfree(xc);
 	return error;
 }
diff --git a/fs/xfs/scrub/dab_bitmap.h b/fs/xfs/scrub/dab_bitmap.h
new file mode 100644
index 0000000000..0c6e3aad43
--- /dev/null
+++ b/fs/xfs/scrub/dab_bitmap.h
@@ -0,0 +1,37 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2022-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_DAB_BITMAP_H__
+#define __XFS_SCRUB_DAB_BITMAP_H__
+
+/* Bitmaps, but for type-checked for xfs_dablk_t */
+
+struct xdab_bitmap {
+	struct xbitmap32	dabitmap;
+};
+
+static inline void xdab_bitmap_init(struct xdab_bitmap *bitmap)
+{
+	xbitmap32_init(&bitmap->dabitmap);
+}
+
+static inline void xdab_bitmap_destroy(struct xdab_bitmap *bitmap)
+{
+	xbitmap32_destroy(&bitmap->dabitmap);
+}
+
+static inline int xdab_bitmap_set(struct xdab_bitmap *bitmap,
+		xfs_dablk_t dabno, xfs_extlen_t len)
+{
+	return xbitmap32_set(&bitmap->dabitmap, dabno, len);
+}
+
+static inline bool xdab_bitmap_test(struct xdab_bitmap *bitmap,
+		xfs_dablk_t dabno, xfs_extlen_t *len)
+{
+	return xbitmap32_test(&bitmap->dabitmap, dabno, len);
+}
+
+#endif	/* __XFS_SCRUB_DAB_BITMAP_H__ */
diff --git a/fs/xfs/scrub/dabtree.c b/fs/xfs/scrub/dabtree.c
index 82b150d3b8..056de4819f 100644
--- a/fs/xfs/scrub/dabtree.c
+++ b/fs/xfs/scrub/dabtree.c
@@ -78,6 +78,22 @@ xchk_da_set_corrupt(
 			__return_address);
 }
 
+/* Flag a da btree node in need of optimization. */
+void
+xchk_da_set_preen(
+	struct xchk_da_btree	*ds,
+	int			level)
+{
+	struct xfs_scrub	*sc = ds->sc;
+
+	sc->sm->sm_flags |= XFS_SCRUB_OFLAG_PREEN;
+	trace_xchk_fblock_preen(sc, ds->dargs.whichfork,
+			xfs_dir2_da_to_db(ds->dargs.geo,
+				ds->state->path.blk[level].blkno),
+			__return_address);
+}
+
+/* Find an entry at a certain level in a da btree. */
 static struct xfs_da_node_entry *
 xchk_da_btree_node_entry(
 	struct xchk_da_btree		*ds,
@@ -320,6 +336,7 @@ xchk_da_btree_block(
 	struct xfs_da3_blkinfo		*hdr3;
 	struct xfs_da_args		*dargs = &ds->dargs;
 	struct xfs_inode		*ip = ds->dargs.dp;
+	xfs_failaddr_t			fa;
 	xfs_ino_t			owner;
 	int				*pmaxrecs;
 	struct xfs_da3_icnode_hdr	nodehdr;
@@ -442,6 +459,12 @@ xchk_da_btree_block(
 		goto out_freebp;
 	}
 
+	fa = xfs_da3_header_check(blk->bp, dargs->owner);
+	if (fa) {
+		xchk_da_set_corrupt(ds, level);
+		goto out_freebp;
+	}
+
 	/*
 	 * If we've been handed a block that is below the dabtree root, does
 	 * its hashval match what the parent block expected to see?
@@ -494,6 +517,7 @@ xchk_da_btree(
 	ds->dargs.whichfork = whichfork;
 	ds->dargs.trans = sc->tp;
 	ds->dargs.op_flags = XFS_DA_OP_OKNOENT;
+	ds->dargs.owner = sc->ip->i_ino;
 	ds->state = xfs_da_state_alloc(&ds->dargs);
 	ds->sc = sc;
 	ds->private = private;
diff --git a/fs/xfs/scrub/dabtree.h b/fs/xfs/scrub/dabtree.h
index 4f8c2138a1..de291e3b77 100644
--- a/fs/xfs/scrub/dabtree.h
+++ b/fs/xfs/scrub/dabtree.h
@@ -35,6 +35,9 @@ bool xchk_da_process_error(struct xchk_da_btree *ds, int level, int *error);
 
 /* Check for da btree corruption. */
 void xchk_da_set_corrupt(struct xchk_da_btree *ds, int level);
+void xchk_da_set_preen(struct xchk_da_btree *ds, int level);
+
+void xchk_da_set_preen(struct xchk_da_btree *ds, int level);
 
 int xchk_da_btree_hash(struct xchk_da_btree *ds, int level, __be32 *hashp);
 int xchk_da_btree(struct xfs_scrub *sc, int whichfork,
diff --git a/fs/xfs/scrub/dir.c b/fs/xfs/scrub/dir.c
index d86ab51af9..bf9199e8df 100644
--- a/fs/xfs/scrub/dir.c
+++ b/fs/xfs/scrub/dir.c
@@ -16,22 +16,70 @@
 #include "xfs_dir2.h"
 #include "xfs_dir2_priv.h"
 #include "xfs_health.h"
+#include "xfs_attr.h"
+#include "xfs_parent.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/dabtree.h"
 #include "scrub/readdir.h"
 #include "scrub/health.h"
+#include "scrub/repair.h"
+#include "scrub/trace.h"
+#include "scrub/xfile.h"
+#include "scrub/xfarray.h"
+#include "scrub/xfblob.h"
 
 /* Set us up to scrub directories. */
 int
 xchk_setup_directory(
 	struct xfs_scrub	*sc)
 {
+	int			error;
+
+	if (xchk_could_repair(sc)) {
+		error = xrep_setup_directory(sc);
+		if (error)
+			return error;
+	}
+
 	return xchk_setup_inode_contents(sc, 0);
 }
 
 /* Directories */
 
+/* Deferred directory entry that we saved for later. */
+struct xchk_dirent {
+	/* Cookie for retrieval of the dirent name. */
+	xfblob_cookie		name_cookie;
+
+	/* Child inode number. */
+	xfs_ino_t		ino;
+
+	/* Length of the pptr name. */
+	uint8_t			namelen;
+};
+
+struct xchk_dir {
+	struct xfs_scrub	*sc;
+
+	/* information for parent pointer validation. */
+	struct xfs_parent_rec	pptr_rec;
+	struct xfs_da_args	pptr_args;
+
+	/* Fixed-size array of xchk_dirent structures. */
+	struct xfarray		*dir_entries;
+
+	/* Blobs containing dirent names. */
+	struct xfblob		*dir_names;
+
+	/* If we've cycled the ILOCK, we must revalidate deferred dirents. */
+	bool			need_revalidate;
+
+	/* Name buffer for dirent revalidation. */
+	struct xfs_name		xname;
+	uint8_t			namebuf[MAXNAMELEN];
+};
+
 /* Scrub a directory entry. */
 
 /* Check that an inode's mode matches a given XFS_DIR3_FT_* type. */
@@ -55,6 +103,108 @@ xchk_dir_check_ftype(
 }
 
 /*
+ * Try to lock a child file for checking parent pointers.  Returns the inode
+ * flags for the locks we now hold, or zero if we failed.
+ */
+STATIC unsigned int
+xchk_dir_lock_child(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*ip)
+{
+	if (!xfs_ilock_nowait(ip, XFS_IOLOCK_SHARED))
+		return 0;
+
+	if (!xfs_ilock_nowait(ip, XFS_ILOCK_SHARED)) {
+		xfs_iunlock(ip, XFS_IOLOCK_SHARED);
+		return 0;
+	}
+
+	if (!xfs_inode_has_attr_fork(ip) || !xfs_need_iread_extents(&ip->i_af))
+		return XFS_IOLOCK_SHARED | XFS_ILOCK_SHARED;
+
+	xfs_iunlock(ip, XFS_ILOCK_SHARED);
+
+	if (!xfs_ilock_nowait(ip, XFS_ILOCK_EXCL)) {
+		xfs_iunlock(ip, XFS_IOLOCK_SHARED);
+		return 0;
+	}
+
+	return XFS_IOLOCK_SHARED | XFS_ILOCK_EXCL;
+}
+
+/* Check the backwards link (parent pointer) associated with this dirent. */
+STATIC int
+xchk_dir_parent_pointer(
+	struct xchk_dir		*sd,
+	const struct xfs_name	*name,
+	struct xfs_inode	*ip)
+{
+	struct xfs_scrub	*sc = sd->sc;
+	int			error;
+
+	xfs_inode_to_parent_rec(&sd->pptr_rec, sc->ip);
+	error = xfs_parent_lookup(sc->tp, ip, name, &sd->pptr_rec,
+			&sd->pptr_args);
+	if (error == -ENOATTR)
+		xchk_fblock_xref_set_corrupt(sc, XFS_DATA_FORK, 0);
+
+	return 0;
+}
+
+/* Look for a parent pointer matching this dirent, if the child isn't busy. */
+STATIC int
+xchk_dir_check_pptr_fast(
+	struct xchk_dir		*sd,
+	xfs_dir2_dataptr_t	dapos,
+	const struct xfs_name	*name,
+	struct xfs_inode	*ip)
+{
+	struct xfs_scrub	*sc = sd->sc;
+	unsigned int		lockmode;
+	int			error;
+
+	/* dot and dotdot entries do not have parent pointers */
+	if (xfs_dir2_samename(name, &xfs_name_dot) ||
+	    xfs_dir2_samename(name, &xfs_name_dotdot))
+		return 0;
+
+	/* No self-referential non-dot or dotdot dirents. */
+	if (ip == sc->ip) {
+		xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
+		return -ECANCELED;
+	}
+
+	/* Try to lock the inode. */
+	lockmode = xchk_dir_lock_child(sc, ip);
+	if (!lockmode) {
+		struct xchk_dirent	save_de = {
+			.namelen	= name->len,
+			.ino		= ip->i_ino,
+		};
+
+		/* Couldn't lock the inode, so save the dirent for later. */
+		trace_xchk_dir_defer(sc->ip, name, ip->i_ino);
+
+		error = xfblob_storename(sd->dir_names, &save_de.name_cookie,
+				name);
+		if (!xchk_fblock_xref_process_error(sc, XFS_DATA_FORK, 0,
+					&error))
+			return error;
+
+		error = xfarray_append(sd->dir_entries, &save_de);
+		if (!xchk_fblock_xref_process_error(sc, XFS_DATA_FORK, 0,
+					&error))
+			return error;
+
+		return 0;
+	}
+
+	error = xchk_dir_parent_pointer(sd, name, ip);
+	xfs_iunlock(ip, lockmode);
+	return error;
+}
+
+/*
  * Scrub a single directory entry.
  *
  * Check the inode number to make sure it's sane, then we check that we can
@@ -71,6 +221,7 @@ xchk_dir_actor(
 {
 	struct xfs_mount	*mp = dp->i_mount;
 	struct xfs_inode	*ip;
+	struct xchk_dir		*sd = priv;
 	xfs_ino_t		lookup_ino;
 	xfs_dablk_t		offset;
 	int			error = 0;
@@ -93,11 +244,11 @@ xchk_dir_actor(
 		return -ECANCELED;
 	}
 
-	if (!strncmp(".", name->name, name->len)) {
+	if (xfs_dir2_samename(name, &xfs_name_dot)) {
 		/* If this is "." then check that the inum matches the dir. */
 		if (ino != dp->i_ino)
 			xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, offset);
-	} else if (!strncmp("..", name->name, name->len)) {
+	} else if (xfs_dir2_samename(name, &xfs_name_dotdot)) {
 		/*
 		 * If this is ".." in the root inode, check that the inum
 		 * matches this dir.
@@ -137,6 +288,14 @@ xchk_dir_actor(
 		goto out;
 
 	xchk_dir_check_ftype(sc, offset, ip, name->type);
+
+	if (xfs_has_parent(mp)) {
+		error = xchk_dir_check_pptr_fast(sd, dapos, name, ip);
+		if (error)
+			goto out_rele;
+	}
+
+out_rele:
 	xchk_irele(sc, ip);
 out:
 	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
@@ -196,8 +355,8 @@ xchk_dir_rec(
 		xchk_da_set_corrupt(ds, level);
 		goto out;
 	}
-	error = xfs_dir3_data_read(ds->dargs.trans, dp, rec_bno,
-			XFS_DABUF_MAP_HOLE_OK, &bp);
+	error = xfs_dir3_data_read(ds->dargs.trans, dp, ds->dargs.owner,
+			rec_bno, XFS_DABUF_MAP_HOLE_OK, &bp);
 	if (!xchk_fblock_process_error(ds->sc, XFS_DATA_FORK, rec_bno,
 			&error))
 		goto out;
@@ -315,10 +474,11 @@ xchk_directory_data_bestfree(
 		/* dir block format */
 		if (lblk != XFS_B_TO_FSBT(mp, XFS_DIR2_DATA_OFFSET))
 			xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
-		error = xfs_dir3_block_read(sc->tp, sc->ip, &bp);
+		error = xfs_dir3_block_read(sc->tp, sc->ip, sc->ip->i_ino, &bp);
 	} else {
 		/* dir data format */
-		error = xfs_dir3_data_read(sc->tp, sc->ip, lblk, 0, &bp);
+		error = xfs_dir3_data_read(sc->tp, sc->ip, sc->ip->i_ino, lblk,
+				0, &bp);
 	}
 	if (!xchk_fblock_process_error(sc, XFS_DATA_FORK, lblk, &error))
 		goto out;
@@ -470,7 +630,7 @@ xchk_directory_leaf1_bestfree(
 	int				error;
 
 	/* Read the free space block. */
-	error = xfs_dir3_leaf_read(sc->tp, sc->ip, lblk, &bp);
+	error = xfs_dir3_leaf_read(sc->tp, sc->ip, sc->ip->i_ino, lblk, &bp);
 	if (!xchk_fblock_process_error(sc, XFS_DATA_FORK, lblk, &error))
 		return error;
 	xchk_buffer_recheck(sc, bp);
@@ -531,10 +691,9 @@ xchk_directory_leaf1_bestfree(
 	/* Check all the bestfree entries. */
 	for (i = 0; i < bestcount; i++, bestp++) {
 		best = be16_to_cpu(*bestp);
-		error = xfs_dir3_data_read(sc->tp, sc->ip,
+		error = xfs_dir3_data_read(sc->tp, sc->ip, args->owner,
 				xfs_dir2_db_to_da(args->geo, i),
-				XFS_DABUF_MAP_HOLE_OK,
-				&dbp);
+				XFS_DABUF_MAP_HOLE_OK, &dbp);
 		if (!xchk_fblock_process_error(sc, XFS_DATA_FORK, lblk,
 				&error))
 			break;
@@ -577,7 +736,7 @@ xchk_directory_free_bestfree(
 	int				error;
 
 	/* Read the free space block */
-	error = xfs_dir2_free_read(sc->tp, sc->ip, lblk, &bp);
+	error = xfs_dir2_free_read(sc->tp, sc->ip, sc->ip->i_ino, lblk, &bp);
 	if (!xchk_fblock_process_error(sc, XFS_DATA_FORK, lblk, &error))
 		return error;
 	xchk_buffer_recheck(sc, bp);
@@ -597,7 +756,7 @@ xchk_directory_free_bestfree(
 			stale++;
 			continue;
 		}
-		error = xfs_dir3_data_read(sc->tp, sc->ip,
+		error = xfs_dir3_data_read(sc->tp, sc->ip, args->owner,
 				(freehdr.firstdb + i) * args->geo->fsbcount,
 				0, &dbp);
 		if (!xchk_fblock_process_error(sc, XFS_DATA_FORK, lblk,
@@ -621,10 +780,11 @@ xchk_directory_blocks(
 {
 	struct xfs_bmbt_irec	got;
 	struct xfs_da_args	args = {
-		.dp		= sc ->ip,
+		.dp		= sc->ip,
 		.whichfork	= XFS_DATA_FORK,
 		.geo		= sc->mp->m_dir_geo,
 		.trans		= sc->tp,
+		.owner		= sc->ip->i_ino,
 	};
 	struct xfs_ifork	*ifp = xfs_ifork_ptr(sc->ip, XFS_DATA_FORK);
 	struct xfs_mount	*mp = sc->mp;
@@ -648,7 +808,8 @@ xchk_directory_blocks(
 	free_lblk = XFS_B_TO_FSB(mp, XFS_DIR2_FREE_OFFSET);
 
 	/* Is this a block dir? */
-	error = xfs_dir2_isblock(&args, &is_block);
+	if (xfs_dir2_format(&args, &error) == XFS_DIR2_FMT_BLOCK)
+		is_block = true;
 	if (!xchk_fblock_process_error(sc, XFS_DATA_FORK, lblk, &error))
 		goto out;
 
@@ -752,11 +913,148 @@ out:
 	return error;
 }
 
+/*
+ * Revalidate a dirent that we collected in the past but couldn't check because
+ * of lock contention.  Returns 0 if the dirent is still valid, -ENOENT if it
+ * has gone away on us, or a negative errno.
+ */
+STATIC int
+xchk_dir_revalidate_dirent(
+	struct xchk_dir		*sd,
+	const struct xfs_name	*xname,
+	xfs_ino_t		ino)
+{
+	struct xfs_scrub	*sc = sd->sc;
+	xfs_ino_t		child_ino;
+	int			error;
+
+	/*
+	 * Look up the directory entry.  If we get -ENOENT, the directory entry
+	 * went away and there's nothing to revalidate.  Return any other
+	 * error.
+	 */
+	error = xchk_dir_lookup(sc, sc->ip, xname, &child_ino);
+	if (error)
+		return error;
+
+	/* The inode number changed, nothing to revalidate. */
+	if (ino != child_ino)
+		return -ENOENT;
+
+	return 0;
+}
+
+/*
+ * Check a directory entry's parent pointers the slow way, which means we cycle
+ * locks a bunch and put up with revalidation until we get it done.
+ */
+STATIC int
+xchk_dir_slow_dirent(
+	struct xchk_dir		*sd,
+	struct xchk_dirent	*dirent,
+	const struct xfs_name	*xname)
+{
+	struct xfs_scrub	*sc = sd->sc;
+	struct xfs_inode	*ip;
+	unsigned int		lockmode;
+	int			error;
+
+	/* Check that the deferred dirent still exists. */
+	if (sd->need_revalidate) {
+		error = xchk_dir_revalidate_dirent(sd, xname, dirent->ino);
+		if (error == -ENOENT)
+			return 0;
+		if (!xchk_fblock_xref_process_error(sc, XFS_DATA_FORK, 0,
+					&error))
+			return error;
+	}
+
+	error = xchk_iget(sc, dirent->ino, &ip);
+	if (error == -EINVAL || error == -ENOENT) {
+		xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
+		return 0;
+	}
+	if (!xchk_fblock_xref_process_error(sc, XFS_DATA_FORK, 0, &error))
+		return error;
+
+	/*
+	 * If we can grab both IOLOCK and ILOCK of the alleged child, we can
+	 * proceed with the validation.
+	 */
+	lockmode = xchk_dir_lock_child(sc, ip);
+	if (lockmode) {
+		trace_xchk_dir_slowpath(sc->ip, xname, ip->i_ino);
+		goto check_pptr;
+	}
+
+	/*
+	 * We couldn't lock the child file.  Drop all the locks and try to
+	 * get them again, one at a time.
+	 */
+	xchk_iunlock(sc, sc->ilock_flags);
+	sd->need_revalidate = true;
+
+	trace_xchk_dir_ultraslowpath(sc->ip, xname, ip->i_ino);
+
+	error = xchk_dir_trylock_for_pptrs(sc, ip, &lockmode);
+	if (error)
+		goto out_rele;
+
+	/* Revalidate, since we just cycled the locks. */
+	error = xchk_dir_revalidate_dirent(sd, xname, dirent->ino);
+	if (error == -ENOENT) {
+		error = 0;
+		goto out_unlock;
+	}
+	if (!xchk_fblock_xref_process_error(sc, XFS_DATA_FORK, 0, &error))
+		goto out_unlock;
+
+check_pptr:
+	error = xchk_dir_parent_pointer(sd, xname, ip);
+out_unlock:
+	xfs_iunlock(ip, lockmode);
+out_rele:
+	xchk_irele(sc, ip);
+	return error;
+}
+
+/* Check all the dirents that we deferred the first time around. */
+STATIC int
+xchk_dir_finish_slow_dirents(
+	struct xchk_dir		*sd)
+{
+	xfarray_idx_t		array_cur;
+	int			error;
+
+	foreach_xfarray_idx(sd->dir_entries, array_cur) {
+		struct xchk_dirent	dirent;
+
+		if (sd->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+			return 0;
+
+		error = xfarray_load(sd->dir_entries, array_cur, &dirent);
+		if (error)
+			return error;
+
+		error = xfblob_loadname(sd->dir_names, dirent.name_cookie,
+				&sd->xname, dirent.namelen);
+		if (error)
+			return error;
+
+		error = xchk_dir_slow_dirent(sd, &dirent, &sd->xname);
+		if (error)
+			return error;
+	}
+
+	return 0;
+}
+
 /* Scrub a whole directory. */
 int
 xchk_directory(
 	struct xfs_scrub	*sc)
 {
+	struct xchk_dir		*sd;
 	int			error;
 
 	if (!S_ISDIR(VFS_I(sc->ip)->i_mode))
@@ -789,9 +1087,60 @@ xchk_directory(
 	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
 		return 0;
 
+	sd = kvzalloc(sizeof(struct xchk_dir), XCHK_GFP_FLAGS);
+	if (!sd)
+		return -ENOMEM;
+	sd->sc = sc;
+	sd->xname.name = sd->namebuf;
+
+	if (xfs_has_parent(sc->mp)) {
+		char		*descr;
+
+		/*
+		 * Set up some staging memory for dirents that we can't check
+		 * due to locking contention.
+		 */
+		descr = xchk_xfile_ino_descr(sc, "slow directory entries");
+		error = xfarray_create(descr, 0, sizeof(struct xchk_dirent),
+				&sd->dir_entries);
+		kfree(descr);
+		if (error)
+			goto out_sd;
+
+		descr = xchk_xfile_ino_descr(sc, "slow directory entry names");
+		error = xfblob_create(descr, &sd->dir_names);
+		kfree(descr);
+		if (error)
+			goto out_entries;
+	}
+
 	/* Look up every name in this directory by hash. */
-	error = xchk_dir_walk(sc, sc->ip, xchk_dir_actor, NULL);
-	if (error && error != -ECANCELED)
+	error = xchk_dir_walk(sc, sc->ip, xchk_dir_actor, sd);
+	if (error == -ECANCELED)
+		error = 0;
+	if (error)
+		goto out_names;
+
+	if (xfs_has_parent(sc->mp)) {
+		error = xchk_dir_finish_slow_dirents(sd);
+		if (error == -ETIMEDOUT) {
+			/* Couldn't grab a lock, scrub was marked incomplete */
+			error = 0;
+			goto out_names;
+		}
+		if (error)
+			goto out_names;
+	}
+
+out_names:
+	if (sd->dir_names)
+		xfblob_destroy(sd->dir_names);
+out_entries:
+	if (sd->dir_entries)
+		xfarray_destroy(sd->dir_entries);
+out_sd:
+	kvfree(sd);
+	if (error)
 		return error;
 
 	/* If the dir is clean, it is clearly not zapped. */
diff --git a/fs/xfs/scrub/dir_repair.c b/fs/xfs/scrub/dir_repair.c
new file mode 100644
index 0000000000..64679fe084
--- /dev/null
+++ b/fs/xfs/scrub/dir_repair.c
@@ -0,0 +1,1958 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2020-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_icache.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
+#include "xfs_dir2.h"
+#include "xfs_dir2_priv.h"
+#include "xfs_bmap.h"
+#include "xfs_quota.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_trans_space.h"
+#include "xfs_bmap_util.h"
+#include "xfs_exchmaps.h"
+#include "xfs_exchrange.h"
+#include "xfs_ag.h"
+#include "xfs_parent.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+#include "scrub/repair.h"
+#include "scrub/tempfile.h"
+#include "scrub/tempexch.h"
+#include "scrub/xfile.h"
+#include "scrub/xfarray.h"
+#include "scrub/xfblob.h"
+#include "scrub/iscan.h"
+#include "scrub/readdir.h"
+#include "scrub/reap.h"
+#include "scrub/findparent.h"
+#include "scrub/orphanage.h"
+#include "scrub/listxattr.h"
+
+/*
+ * Directory Repair
+ * ================
+ *
+ * We repair directories by reading the directory data blocks looking for
+ * directory entries that look salvageable (name passes verifiers, entry points
+ * to a valid allocated inode, etc).  Each entry worth salvaging is stashed in
+ * memory, and the stashed entries are periodically replayed into a temporary
+ * directory to constrain memory use.  Batching the construction of the
+ * temporary directory in this fashion reduces lock cycling of the directory
+ * being repaired and the temporary directory, and will later become important
+ * for parent pointer scanning.
+ *
+ * If parent pointers are enabled on this filesystem, we instead reconstruct
+ * the directory by visiting each parent pointer of each file in the filesystem
+ * and translating the relevant parent pointer records into dirents.  In this
+ * case, it is advantageous to stash all directory entries created from parent
+ * pointers for a single child file before replaying them into the temporary
+ * directory.  To save memory, the live filesystem scan reuses the findparent
+ * fields.  Directory repair chooses either parent pointer scanning or
+ * directory entry salvaging, but not both.
+ *
+ * Directory entries added to the temporary directory do not elevate the link
+ * counts of the inodes found.  When salvaging completes, the remaining stashed
+ * entries are replayed to the temporary directory.  An atomic mapping exchange
+ * is used to commit the new directory blocks to the directory being repaired.
+ * This will disrupt readdir cursors.
+ *
+ * Locking Issues
+ * --------------
+ *
+ * If /a, /a/b, and /c are all directories, the VFS does not take i_rwsem on
+ * /a/b for a "mv /a/b /c/" operation.  This means that only b's ILOCK protects
+ * b's dotdot update.  This is in contrast to every other dotdot update (link,
+ * remove, mkdir).  If the repair code drops the ILOCK, it must either
+ * revalidate the dotdot entry or use dirent hooks to capture updates from
+ * other threads.
+ */
+
+/* Create a dirent in the tempdir. */
+#define XREP_DIRENT_ADD		(1)
+
+/* Remove a dirent from the tempdir. */
+#define XREP_DIRENT_REMOVE	(2)
+
+/* Directory entry to be restored in the new directory. */
+struct xrep_dirent {
+	/* Cookie for retrieval of the dirent name. */
+	xfblob_cookie		name_cookie;
+
+	/* Target inode number. */
+	xfs_ino_t		ino;
+
+	/* Length of the dirent name. */
+	uint8_t			namelen;
+
+	/* File type of the dirent. */
+	uint8_t			ftype;
+
+	/* XREP_DIRENT_{ADD,REMOVE} */
+	uint8_t			action;
+};
+
+/*
+ * Stash up to 8 pages of recovered dirent data in dir_entries and dir_names
+ * before we write them to the temp dir.
+ */
+#define XREP_DIR_MAX_STASH_BYTES	(PAGE_SIZE * 8)
+
+struct xrep_dir {
+	struct xfs_scrub	*sc;
+
+	/* Fixed-size array of xrep_dirent structures. */
+	struct xfarray		*dir_entries;
+
+	/* Blobs containing directory entry names. */
+	struct xfblob		*dir_names;
+
+	/* Information for exchanging data forks at the end. */
+	struct xrep_tempexch	tx;
+
+	/* Preallocated args struct for performing dir operations */
+	struct xfs_da_args	args;
+
+	/*
+	 * Information used to scan the filesystem to find the inumber of the
+	 * dotdot entry for this directory.  For directory salvaging when
+	 * parent pointers are not enabled, we use the findparent_* functions
+	 * on this object and access only the parent_ino field directly.
+	 *
+	 * When parent pointers are enabled, however, the pptr scanner uses the
+	 * iscan, hooks, lock, and parent_ino fields of this object directly.
+	 * @pscan.lock coordinates access to dir_entries, dir_names,
+	 * parent_ino, subdirs, dirents, and args.  This reduces the memory
+	 * requirements of this structure.
+	 */
+	struct xrep_parent_scan_info pscan;
+
+	/*
+	 * Context information for attaching this directory to the lost+found
+	 * if this directory does not have a parent.
+	 */
+	struct xrep_adoption	adoption;
+
+	/* How many subdirectories did we find? */
+	uint64_t		subdirs;
+
+	/* How many dirents did we find? */
+	unsigned int		dirents;
+
+	/* Should we move this directory to the orphanage? */
+	bool			needs_adoption;
+
+	/* Directory entry name, plus the trailing null. */
+	struct xfs_name		xname;
+	unsigned char		namebuf[MAXNAMELEN];
+};
+
+/* Tear down all the incore stuff we created. */
+static void
+xrep_dir_teardown(
+	struct xfs_scrub	*sc)
+{
+	struct xrep_dir		*rd = sc->buf;
+
+	xrep_findparent_scan_teardown(&rd->pscan);
+	xfblob_destroy(rd->dir_names);
+	xfarray_destroy(rd->dir_entries);
+}
+
+/* Set up for a directory repair. */
+int
+xrep_setup_directory(
+	struct xfs_scrub	*sc)
+{
+	struct xrep_dir		*rd;
+	int			error;
+
+	xchk_fsgates_enable(sc, XCHK_FSGATES_DIRENTS);
+
+	error = xrep_orphanage_try_create(sc);
+	if (error)
+		return error;
+
+	error = xrep_tempfile_create(sc, S_IFDIR);
+	if (error)
+		return error;
+
+	rd = kvzalloc(sizeof(struct xrep_dir), XCHK_GFP_FLAGS);
+	if (!rd)
+		return -ENOMEM;
+	rd->sc = sc;
+	rd->xname.name = rd->namebuf;
+	sc->buf = rd;
+
+	return 0;
+}
+
+/*
+ * Look up the dotdot entry and confirm that it's really the parent.
+ * Returns NULLFSINO if we don't know what to do.
+ */
+static inline xfs_ino_t
+xrep_dir_lookup_parent(
+	struct xrep_dir		*rd)
+{
+	struct xfs_scrub	*sc = rd->sc;
+	xfs_ino_t		ino;
+	int			error;
+
+	error = xfs_dir_lookup(sc->tp, sc->ip, &xfs_name_dotdot, &ino, NULL);
+	if (error)
+		return NULLFSINO;
+	if (!xfs_verify_dir_ino(sc->mp, ino))
+		return NULLFSINO;
+
+	error = xrep_findparent_confirm(sc, &ino);
+	if (error)
+		return NULLFSINO;
+
+	return ino;
+}
+
+/*
+ * Look up '..' in the dentry cache and confirm that it's really the parent.
+ * Returns NULLFSINO if the dcache misses or if the hit is implausible.
+ */
+static inline xfs_ino_t
+xrep_dir_dcache_parent(
+	struct xrep_dir		*rd)
+{
+	struct xfs_scrub	*sc = rd->sc;
+	xfs_ino_t		parent_ino;
+	int			error;
+
+	parent_ino = xrep_findparent_from_dcache(sc);
+	if (parent_ino == NULLFSINO)
+		return parent_ino;
+
+	error = xrep_findparent_confirm(sc, &parent_ino);
+	if (error)
+		return NULLFSINO;
+
+	return parent_ino;
+}
+
+/* Try to find the parent of the directory being repaired. */
+STATIC int
+xrep_dir_find_parent(
+	struct xrep_dir		*rd)
+{
+	xfs_ino_t		ino;
+
+	ino = xrep_findparent_self_reference(rd->sc);
+	if (ino != NULLFSINO) {
+		xrep_findparent_scan_finish_early(&rd->pscan, ino);
+		return 0;
+	}
+
+	ino = xrep_dir_dcache_parent(rd);
+	if (ino != NULLFSINO) {
+		xrep_findparent_scan_finish_early(&rd->pscan, ino);
+		return 0;
+	}
+
+	ino = xrep_dir_lookup_parent(rd);
+	if (ino != NULLFSINO) {
+		xrep_findparent_scan_finish_early(&rd->pscan, ino);
+		return 0;
+	}
+
+	/*
+	 * A full filesystem scan is the last resort.  On a busy filesystem,
+	 * the scan can fail with -EBUSY if we cannot grab IOLOCKs.  That means
+	 * that we don't know what who the parent is, so we should return to
+	 * userspace.
+	 */
+	return xrep_findparent_scan(&rd->pscan);
+}
+
+/*
+ * Decide if we want to salvage this entry.  We don't bother with oversized
+ * names or the dot entry.
+ */
+STATIC int
+xrep_dir_want_salvage(
+	struct xrep_dir		*rd,
+	const char		*name,
+	int			namelen,
+	xfs_ino_t		ino)
+{
+	struct xfs_mount	*mp = rd->sc->mp;
+
+	/* No pointers to ourselves or to garbage. */
+	if (ino == rd->sc->ip->i_ino)
+		return false;
+	if (!xfs_verify_dir_ino(mp, ino))
+		return false;
+
+	/* No weird looking names or dot entries. */
+	if (namelen >= MAXNAMELEN || namelen <= 0)
+		return false;
+	if (namelen == 1 && name[0] == '.')
+		return false;
+	if (!xfs_dir2_namecheck(name, namelen))
+		return false;
+
+	return true;
+}
+
+/*
+ * Remember that we want to create a dirent in the tempdir.  These stashed
+ * actions will be replayed later.
+ */
+STATIC int
+xrep_dir_stash_createname(
+	struct xrep_dir		*rd,
+	const struct xfs_name	*name,
+	xfs_ino_t		ino)
+{
+	struct xrep_dirent	dirent = {
+		.action		= XREP_DIRENT_ADD,
+		.ino		= ino,
+		.namelen	= name->len,
+		.ftype		= name->type,
+	};
+	int			error;
+
+	trace_xrep_dir_stash_createname(rd->sc->tempip, name, ino);
+
+	error = xfblob_storename(rd->dir_names, &dirent.name_cookie, name);
+	if (error)
+		return error;
+
+	return xfarray_append(rd->dir_entries, &dirent);
+}
+
+/*
+ * Remember that we want to remove a dirent from the tempdir.  These stashed
+ * actions will be replayed later.
+ */
+STATIC int
+xrep_dir_stash_removename(
+	struct xrep_dir		*rd,
+	const struct xfs_name	*name,
+	xfs_ino_t		ino)
+{
+	struct xrep_dirent	dirent = {
+		.action		= XREP_DIRENT_REMOVE,
+		.ino		= ino,
+		.namelen	= name->len,
+		.ftype		= name->type,
+	};
+	int			error;
+
+	trace_xrep_dir_stash_removename(rd->sc->tempip, name, ino);
+
+	error = xfblob_storename(rd->dir_names, &dirent.name_cookie, name);
+	if (error)
+		return error;
+
+	return xfarray_append(rd->dir_entries, &dirent);
+}
+
+/* Allocate an in-core record to hold entries while we rebuild the dir data. */
+STATIC int
+xrep_dir_salvage_entry(
+	struct xrep_dir		*rd,
+	unsigned char		*name,
+	unsigned int		namelen,
+	xfs_ino_t		ino)
+{
+	struct xfs_name		xname = {
+		.name		= name,
+	};
+	struct xfs_scrub	*sc = rd->sc;
+	struct xfs_inode	*ip;
+	unsigned int		i = 0;
+	int			error = 0;
+
+	if (xchk_should_terminate(sc, &error))
+		return error;
+
+	/*
+	 * Truncate the name to the first character that would trip namecheck.
+	 * If we no longer have a name after that, ignore this entry.
+	 */
+	while (i < namelen && name[i] != 0 && name[i] != '/')
+		i++;
+	if (i == 0)
+		return 0;
+	xname.len = i;
+
+	/* Ignore '..' entries; we already picked the new parent. */
+	if (xname.len == 2 && name[0] == '.' && name[1] == '.') {
+		trace_xrep_dir_salvaged_parent(sc->ip, ino);
+		return 0;
+	}
+
+	trace_xrep_dir_salvage_entry(sc->ip, &xname, ino);
+
+	/*
+	 * Compute the ftype or dump the entry if we can't.  We don't lock the
+	 * inode because inodes can't change type while we have a reference.
+	 */
+	error = xchk_iget(sc, ino, &ip);
+	if (error)
+		return 0;
+
+	xname.type = xfs_mode_to_ftype(VFS_I(ip)->i_mode);
+	xchk_irele(sc, ip);
+
+	return xrep_dir_stash_createname(rd, &xname, ino);
+}
+
+/* Record a shortform directory entry for later reinsertion. */
+STATIC int
+xrep_dir_salvage_sf_entry(
+	struct xrep_dir			*rd,
+	struct xfs_dir2_sf_hdr		*sfp,
+	struct xfs_dir2_sf_entry	*sfep)
+{
+	xfs_ino_t			ino;
+
+	ino = xfs_dir2_sf_get_ino(rd->sc->mp, sfp, sfep);
+	if (!xrep_dir_want_salvage(rd, sfep->name, sfep->namelen, ino))
+		return 0;
+
+	return xrep_dir_salvage_entry(rd, sfep->name, sfep->namelen, ino);
+}
+
+/* Record a regular directory entry for later reinsertion. */
+STATIC int
+xrep_dir_salvage_data_entry(
+	struct xrep_dir			*rd,
+	struct xfs_dir2_data_entry	*dep)
+{
+	xfs_ino_t			ino;
+
+	ino = be64_to_cpu(dep->inumber);
+	if (!xrep_dir_want_salvage(rd, dep->name, dep->namelen, ino))
+		return 0;
+
+	return xrep_dir_salvage_entry(rd, dep->name, dep->namelen, ino);
+}
+
+/* Try to recover block/data format directory entries. */
+STATIC int
+xrep_dir_recover_data(
+	struct xrep_dir		*rd,
+	struct xfs_buf		*bp)
+{
+	struct xfs_da_geometry	*geo = rd->sc->mp->m_dir_geo;
+	unsigned int		offset;
+	unsigned int		end;
+	int			error = 0;
+
+	/*
+	 * Loop over the data portion of the block.
+	 * Each object is a real entry (dep) or an unused one (dup).
+	 */
+	offset = geo->data_entry_offset;
+	end = min_t(unsigned int, BBTOB(bp->b_length),
+			xfs_dir3_data_end_offset(geo, bp->b_addr));
+
+	while (offset < end) {
+		struct xfs_dir2_data_unused	*dup = bp->b_addr + offset;
+		struct xfs_dir2_data_entry	*dep = bp->b_addr + offset;
+
+		if (xchk_should_terminate(rd->sc, &error))
+			return error;
+
+		/* Skip unused entries. */
+		if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
+			offset += be16_to_cpu(dup->length);
+			continue;
+		}
+
+		/* Don't walk off the end of the block. */
+		offset += xfs_dir2_data_entsize(rd->sc->mp, dep->namelen);
+		if (offset > end)
+			break;
+
+		/* Ok, let's save this entry. */
+		error = xrep_dir_salvage_data_entry(rd, dep);
+		if (error)
+			return error;
+
+	}
+
+	return 0;
+}
+
+/* Try to recover shortform directory entries. */
+STATIC int
+xrep_dir_recover_sf(
+	struct xrep_dir			*rd)
+{
+	struct xfs_dir2_sf_hdr		*hdr;
+	struct xfs_dir2_sf_entry	*sfep;
+	struct xfs_dir2_sf_entry	*next;
+	struct xfs_ifork		*ifp;
+	xfs_ino_t			ino;
+	unsigned char			*end;
+	int				error = 0;
+
+	ifp = xfs_ifork_ptr(rd->sc->ip, XFS_DATA_FORK);
+	hdr = ifp->if_data;
+	end = (unsigned char *)ifp->if_data + ifp->if_bytes;
+
+	ino = xfs_dir2_sf_get_parent_ino(hdr);
+	trace_xrep_dir_salvaged_parent(rd->sc->ip, ino);
+
+	sfep = xfs_dir2_sf_firstentry(hdr);
+	while ((unsigned char *)sfep < end) {
+		if (xchk_should_terminate(rd->sc, &error))
+			return error;
+
+		next = xfs_dir2_sf_nextentry(rd->sc->mp, hdr, sfep);
+		if ((unsigned char *)next > end)
+			break;
+
+		/* Ok, let's save this entry. */
+		error = xrep_dir_salvage_sf_entry(rd, hdr, sfep);
+		if (error)
+			return error;
+
+		sfep = next;
+	}
+
+	return 0;
+}
+
+/*
+ * Try to figure out the format of this directory from the data fork mappings
+ * and the directory size.  If we can be reasonably sure of format, we can be
+ * more aggressive in salvaging directory entries.  On return, @magic_guess
+ * will be set to DIR3_BLOCK_MAGIC if we think this is a "block format"
+ * directory; DIR3_DATA_MAGIC if we think this is a "data format" directory,
+ * and 0 if we can't tell.
+ */
+STATIC void
+xrep_dir_guess_format(
+	struct xrep_dir		*rd,
+	__be32			*magic_guess)
+{
+	struct xfs_inode	*dp = rd->sc->ip;
+	struct xfs_mount	*mp = rd->sc->mp;
+	struct xfs_da_geometry	*geo = mp->m_dir_geo;
+	xfs_fileoff_t		last;
+	int			error;
+
+	ASSERT(xfs_has_crc(mp));
+
+	*magic_guess = 0;
+
+	/*
+	 * If there's a single directory block and the directory size is
+	 * exactly one block, this has to be a single block format directory.
+	 */
+	error = xfs_bmap_last_offset(dp, &last, XFS_DATA_FORK);
+	if (!error && XFS_FSB_TO_B(mp, last) == geo->blksize &&
+	    dp->i_disk_size == geo->blksize) {
+		*magic_guess = cpu_to_be32(XFS_DIR3_BLOCK_MAGIC);
+		return;
+	}
+
+	/*
+	 * If the last extent before the leaf offset matches the directory
+	 * size and the directory size is larger than 1 block, this is a
+	 * data format directory.
+	 */
+	last = geo->leafblk;
+	error = xfs_bmap_last_before(rd->sc->tp, dp, &last, XFS_DATA_FORK);
+	if (!error &&
+	    XFS_FSB_TO_B(mp, last) > geo->blksize &&
+	    XFS_FSB_TO_B(mp, last) == dp->i_disk_size) {
+		*magic_guess = cpu_to_be32(XFS_DIR3_DATA_MAGIC);
+		return;
+	}
+}
+
+/* Recover directory entries from a specific directory block. */
+STATIC int
+xrep_dir_recover_dirblock(
+	struct xrep_dir		*rd,
+	__be32			magic_guess,
+	xfs_dablk_t		dabno)
+{
+	struct xfs_dir2_data_hdr *hdr;
+	struct xfs_buf		*bp;
+	__be32			oldmagic;
+	int			error;
+
+	/*
+	 * Try to read buffer.  We invalidate them in the next step so we don't
+	 * bother to set a buffer type or ops.
+	 */
+	error = xfs_da_read_buf(rd->sc->tp, rd->sc->ip, dabno,
+			XFS_DABUF_MAP_HOLE_OK, &bp, XFS_DATA_FORK, NULL);
+	if (error || !bp)
+		return error;
+
+	hdr = bp->b_addr;
+	oldmagic = hdr->magic;
+
+	trace_xrep_dir_recover_dirblock(rd->sc->ip, dabno,
+			be32_to_cpu(hdr->magic), be32_to_cpu(magic_guess));
+
+	/*
+	 * If we're sure of the block's format, proceed with the salvage
+	 * operation using the specified magic number.
+	 */
+	if (magic_guess) {
+		hdr->magic = magic_guess;
+		goto recover;
+	}
+
+	/*
+	 * If we couldn't guess what type of directory this is, then we will
+	 * only salvage entries from directory blocks that match the magic
+	 * number and pass verifiers.
+	 */
+	switch (hdr->magic) {
+	case cpu_to_be32(XFS_DIR2_BLOCK_MAGIC):
+	case cpu_to_be32(XFS_DIR3_BLOCK_MAGIC):
+		if (!xrep_buf_verify_struct(bp, &xfs_dir3_block_buf_ops))
+			goto out;
+		if (xfs_dir3_block_header_check(bp, rd->sc->ip->i_ino) != NULL)
+			goto out;
+		break;
+	case cpu_to_be32(XFS_DIR2_DATA_MAGIC):
+	case cpu_to_be32(XFS_DIR3_DATA_MAGIC):
+		if (!xrep_buf_verify_struct(bp, &xfs_dir3_data_buf_ops))
+			goto out;
+		if (xfs_dir3_data_header_check(bp, rd->sc->ip->i_ino) != NULL)
+			goto out;
+		break;
+	default:
+		goto out;
+	}
+
+recover:
+	error = xrep_dir_recover_data(rd, bp);
+
+out:
+	hdr->magic = oldmagic;
+	xfs_trans_brelse(rd->sc->tp, bp);
+	return error;
+}
+
+static inline void
+xrep_dir_init_args(
+	struct xrep_dir		*rd,
+	struct xfs_inode	*dp,
+	const struct xfs_name	*name)
+{
+	memset(&rd->args, 0, sizeof(struct xfs_da_args));
+	rd->args.geo = rd->sc->mp->m_dir_geo;
+	rd->args.whichfork = XFS_DATA_FORK;
+	rd->args.owner = rd->sc->ip->i_ino;
+	rd->args.trans = rd->sc->tp;
+	rd->args.dp = dp;
+	if (!name)
+		return;
+	rd->args.name = name->name;
+	rd->args.namelen = name->len;
+	rd->args.filetype = name->type;
+	rd->args.hashval = xfs_dir2_hashname(rd->sc->mp, name);
+}
+
+/* Replay a stashed createname into the temporary directory. */
+STATIC int
+xrep_dir_replay_createname(
+	struct xrep_dir		*rd,
+	const struct xfs_name	*name,
+	xfs_ino_t		inum,
+	xfs_extlen_t		total)
+{
+	struct xfs_scrub	*sc = rd->sc;
+	struct xfs_inode	*dp = rd->sc->tempip;
+	int			error;
+
+	ASSERT(S_ISDIR(VFS_I(dp)->i_mode));
+
+	error = xfs_dir_ino_validate(sc->mp, inum);
+	if (error)
+		return error;
+
+	trace_xrep_dir_replay_createname(dp, name, inum);
+
+	xrep_dir_init_args(rd, dp, name);
+	rd->args.inumber = inum;
+	rd->args.total = total;
+	rd->args.op_flags = XFS_DA_OP_ADDNAME | XFS_DA_OP_OKNOENT;
+	return xfs_dir_createname_args(&rd->args);
+}
+
+/* Replay a stashed removename onto the temporary directory. */
+STATIC int
+xrep_dir_replay_removename(
+	struct xrep_dir		*rd,
+	const struct xfs_name	*name,
+	xfs_extlen_t		total)
+{
+	struct xfs_inode	*dp = rd->args.dp;
+
+	ASSERT(S_ISDIR(VFS_I(dp)->i_mode));
+
+	xrep_dir_init_args(rd, dp, name);
+	rd->args.op_flags = 0;
+	rd->args.total = total;
+
+	trace_xrep_dir_replay_removename(dp, name, 0);
+	return xfs_dir_removename_args(&rd->args);
+}
+
+/*
+ * Add this stashed incore directory entry to the temporary directory.
+ * The caller must hold the tempdir's IOLOCK, must not hold any ILOCKs, and
+ * must not be in transaction context.
+ */
+STATIC int
+xrep_dir_replay_update(
+	struct xrep_dir			*rd,
+	const struct xfs_name		*xname,
+	const struct xrep_dirent	*dirent)
+{
+	struct xfs_mount		*mp = rd->sc->mp;
+#ifdef DEBUG
+	xfs_ino_t			ino;
+#endif
+	uint				resblks;
+	int				error;
+
+	resblks = xfs_link_space_res(mp, xname->len);
+	error = xchk_trans_alloc(rd->sc, resblks);
+	if (error)
+		return error;
+
+	/* Lock the temporary directory and join it to the transaction */
+	xrep_tempfile_ilock(rd->sc);
+	xfs_trans_ijoin(rd->sc->tp, rd->sc->tempip, 0);
+
+	switch (dirent->action) {
+	case XREP_DIRENT_ADD:
+		/*
+		 * Create a replacement dirent in the temporary directory.
+		 * Note that _createname doesn't check for existing entries.
+		 * There shouldn't be any in the temporary dir, but we'll
+		 * verify this in debug mode.
+		 */
+#ifdef DEBUG
+		error = xchk_dir_lookup(rd->sc, rd->sc->tempip, xname, &ino);
+		if (error != -ENOENT) {
+			ASSERT(error != -ENOENT);
+			goto out_cancel;
+		}
+#endif
+
+		error = xrep_dir_replay_createname(rd, xname, dirent->ino,
+				resblks);
+		if (error)
+			goto out_cancel;
+
+		if (xname->type == XFS_DIR3_FT_DIR)
+			rd->subdirs++;
+		rd->dirents++;
+		break;
+	case XREP_DIRENT_REMOVE:
+		/*
+		 * Remove a dirent from the temporary directory.  Note that
+		 * _removename doesn't check the inode target of the exist
+		 * entry.  There should be a perfect match in the temporary
+		 * dir, but we'll verify this in debug mode.
+		 */
+#ifdef DEBUG
+		error = xchk_dir_lookup(rd->sc, rd->sc->tempip, xname, &ino);
+		if (error) {
+			ASSERT(error != 0);
+			goto out_cancel;
+		}
+		if (ino != dirent->ino) {
+			ASSERT(ino == dirent->ino);
+			error = -EIO;
+			goto out_cancel;
+		}
+#endif
+
+		error = xrep_dir_replay_removename(rd, xname, resblks);
+		if (error)
+			goto out_cancel;
+
+		if (xname->type == XFS_DIR3_FT_DIR)
+			rd->subdirs--;
+		rd->dirents--;
+		break;
+	default:
+		ASSERT(0);
+		error = -EIO;
+		goto out_cancel;
+	}
+
+	/* Commit and unlock. */
+	error = xrep_trans_commit(rd->sc);
+	if (error)
+		return error;
+
+	xrep_tempfile_iunlock(rd->sc);
+	return 0;
+out_cancel:
+	xchk_trans_cancel(rd->sc);
+	xrep_tempfile_iunlock(rd->sc);
+	return error;
+}
+
+/*
+ * Flush stashed incore dirent updates that have been recorded by the scanner.
+ * This is done to reduce the memory requirements of the directory rebuild,
+ * since directories can contain up to 32GB of directory data.
+ *
+ * Caller must not hold transactions or ILOCKs.  Caller must hold the tempdir
+ * IOLOCK.
+ */
+STATIC int
+xrep_dir_replay_updates(
+	struct xrep_dir		*rd)
+{
+	xfarray_idx_t		array_cur;
+	int			error;
+
+	/* Add all the salvaged dirents to the temporary directory. */
+	mutex_lock(&rd->pscan.lock);
+	foreach_xfarray_idx(rd->dir_entries, array_cur) {
+		struct xrep_dirent	dirent;
+
+		error = xfarray_load(rd->dir_entries, array_cur, &dirent);
+		if (error)
+			goto out_unlock;
+
+		error = xfblob_loadname(rd->dir_names, dirent.name_cookie,
+				&rd->xname, dirent.namelen);
+		if (error)
+			goto out_unlock;
+		rd->xname.type = dirent.ftype;
+		mutex_unlock(&rd->pscan.lock);
+
+		error = xrep_dir_replay_update(rd, &rd->xname, &dirent);
+		if (error)
+			return error;
+		mutex_lock(&rd->pscan.lock);
+	}
+
+	/* Empty out both arrays now that we've added the entries. */
+	xfarray_truncate(rd->dir_entries);
+	xfblob_truncate(rd->dir_names);
+	mutex_unlock(&rd->pscan.lock);
+	return 0;
+out_unlock:
+	mutex_unlock(&rd->pscan.lock);
+	return error;
+}
+
+/*
+ * Periodically flush stashed directory entries to the temporary dir.  This
+ * is done to reduce the memory requirements of the directory rebuild, since
+ * directories can contain up to 32GB of directory data.
+ */
+STATIC int
+xrep_dir_flush_stashed(
+	struct xrep_dir		*rd)
+{
+	int			error;
+
+	/*
+	 * Entering this function, the scrub context has a reference to the
+	 * inode being repaired, the temporary file, and a scrub transaction
+	 * that we use during dirent salvaging to avoid livelocking if there
+	 * are cycles in the directory structures.  We hold ILOCK_EXCL on both
+	 * the inode being repaired and the temporary file, though they are
+	 * not ijoined to the scrub transaction.
+	 *
+	 * To constrain kernel memory use, we occasionally write salvaged
+	 * dirents from the xfarray and xfblob structures into the temporary
+	 * directory in preparation for exchanging the directory structures at
+	 * the end.  Updating the temporary file requires a transaction, so we
+	 * commit the scrub transaction and drop the two ILOCKs so that
+	 * we can allocate whatever transaction we want.
+	 *
+	 * We still hold IOLOCK_EXCL on the inode being repaired, which
+	 * prevents anyone from accessing the damaged directory data while we
+	 * repair it.
+	 */
+	error = xrep_trans_commit(rd->sc);
+	if (error)
+		return error;
+	xchk_iunlock(rd->sc, XFS_ILOCK_EXCL);
+
+	/*
+	 * Take the IOLOCK of the temporary file while we modify dirents.  This
+	 * isn't strictly required because the temporary file is never revealed
+	 * to userspace, but we follow the same locking rules.  We still hold
+	 * sc->ip's IOLOCK.
+	 */
+	error = xrep_tempfile_iolock_polled(rd->sc);
+	if (error)
+		return error;
+
+	/* Write to the tempdir all the updates that we've stashed. */
+	error = xrep_dir_replay_updates(rd);
+	xrep_tempfile_iounlock(rd->sc);
+	if (error)
+		return error;
+
+	/*
+	 * Recreate the salvage transaction and relock the dir we're salvaging.
+	 */
+	error = xchk_trans_alloc(rd->sc, 0);
+	if (error)
+		return error;
+	xchk_ilock(rd->sc, XFS_ILOCK_EXCL);
+	return 0;
+}
+
+/* Decide if we've stashed too much dirent data in memory. */
+static inline bool
+xrep_dir_want_flush_stashed(
+	struct xrep_dir		*rd)
+{
+	unsigned long long	bytes;
+
+	bytes = xfarray_bytes(rd->dir_entries) + xfblob_bytes(rd->dir_names);
+	return bytes > XREP_DIR_MAX_STASH_BYTES;
+}
+
+/* Extract as many directory entries as we can. */
+STATIC int
+xrep_dir_recover(
+	struct xrep_dir		*rd)
+{
+	struct xfs_bmbt_irec	got;
+	struct xfs_scrub	*sc = rd->sc;
+	struct xfs_da_geometry	*geo = sc->mp->m_dir_geo;
+	xfs_fileoff_t		offset;
+	xfs_dablk_t		dabno;
+	__be32			magic_guess;
+	int			nmap;
+	int			error;
+
+	xrep_dir_guess_format(rd, &magic_guess);
+
+	/* Iterate each directory data block in the data fork. */
+	for (offset = 0;
+	     offset < geo->leafblk;
+	     offset = got.br_startoff + got.br_blockcount) {
+		nmap = 1;
+		error = xfs_bmapi_read(sc->ip, offset, geo->leafblk - offset,
+				&got, &nmap, 0);
+		if (error)
+			return error;
+		if (nmap != 1)
+			return -EFSCORRUPTED;
+		if (!xfs_bmap_is_written_extent(&got))
+			continue;
+
+		for (dabno = round_up(got.br_startoff, geo->fsbcount);
+		     dabno < got.br_startoff + got.br_blockcount;
+		     dabno += geo->fsbcount) {
+			if (xchk_should_terminate(rd->sc, &error))
+				return error;
+
+			error = xrep_dir_recover_dirblock(rd,
+					magic_guess, dabno);
+			if (error)
+				return error;
+
+			/* Flush dirents to constrain memory usage. */
+			if (xrep_dir_want_flush_stashed(rd)) {
+				error = xrep_dir_flush_stashed(rd);
+				if (error)
+					return error;
+			}
+		}
+	}
+
+	return 0;
+}
+
+/*
+ * Find all the directory entries for this inode by scraping them out of the
+ * directory leaf blocks by hand, and flushing them into the temp dir.
+ */
+STATIC int
+xrep_dir_find_entries(
+	struct xrep_dir		*rd)
+{
+	struct xfs_inode	*dp = rd->sc->ip;
+	int			error;
+
+	/*
+	 * Salvage directory entries from the old directory, and write them to
+	 * the temporary directory.
+	 */
+	if (dp->i_df.if_format == XFS_DINODE_FMT_LOCAL) {
+		error = xrep_dir_recover_sf(rd);
+	} else {
+		error = xfs_iread_extents(rd->sc->tp, dp, XFS_DATA_FORK);
+		if (error)
+			return error;
+
+		error = xrep_dir_recover(rd);
+	}
+	if (error)
+		return error;
+
+	return xrep_dir_flush_stashed(rd);
+}
+
+/* Scan all files in the filesystem for dirents. */
+STATIC int
+xrep_dir_salvage_entries(
+	struct xrep_dir		*rd)
+{
+	struct xfs_scrub	*sc = rd->sc;
+	int			error;
+
+	/*
+	 * Drop the ILOCK on this directory so that we can scan for this
+	 * directory's parent.  Figure out who is going to be the parent of
+	 * this directory, then retake the ILOCK so that we can salvage
+	 * directory entries.
+	 */
+	xchk_iunlock(sc, XFS_ILOCK_EXCL);
+	error = xrep_dir_find_parent(rd);
+	xchk_ilock(sc, XFS_ILOCK_EXCL);
+	if (error)
+		return error;
+
+	/*
+	 * Collect directory entries by parsing raw leaf blocks to salvage
+	 * whatever we can.  When we're done, free the staging memory before
+	 * exchanging the directories to reduce memory usage.
+	 */
+	error = xrep_dir_find_entries(rd);
+	if (error)
+		return error;
+
+	/*
+	 * Cancel the repair transaction and drop the ILOCK so that we can
+	 * (later) use the atomic mapping exchange functions to compute the
+	 * correct block reservations and re-lock the inodes.
+	 *
+	 * We still hold IOLOCK_EXCL (aka i_rwsem) which will prevent directory
+	 * modifications, but there's nothing to prevent userspace from reading
+	 * the directory until we're ready for the exchange operation.  Reads
+	 * will return -EIO without shutting down the fs, so we're ok with
+	 * that.
+	 *
+	 * The VFS can change dotdot on us, but the findparent scan will keep
+	 * our incore parent inode up to date.  See the note on locking issues
+	 * for more details.
+	 */
+	error = xrep_trans_commit(sc);
+	if (error)
+		return error;
+
+	xchk_iunlock(sc, XFS_ILOCK_EXCL);
+	return 0;
+}
+
+
+/*
+ * Examine a parent pointer of a file.  If it leads us back to the directory
+ * that we're rebuilding, create an incore dirent from the parent pointer and
+ * stash it.
+ */
+STATIC int
+xrep_dir_scan_pptr(
+	struct xfs_scrub		*sc,
+	struct xfs_inode		*ip,
+	unsigned int			attr_flags,
+	const unsigned char		*name,
+	unsigned int			namelen,
+	const void			*value,
+	unsigned int			valuelen,
+	void				*priv)
+{
+	struct xfs_name			xname = {
+		.name			= name,
+		.len			= namelen,
+		.type			= xfs_mode_to_ftype(VFS_I(ip)->i_mode),
+	};
+	xfs_ino_t			parent_ino;
+	uint32_t			parent_gen;
+	struct xrep_dir			*rd = priv;
+	int				error;
+
+	if (!(attr_flags & XFS_ATTR_PARENT))
+		return 0;
+
+	/*
+	 * Ignore parent pointers that point back to a different dir, list the
+	 * wrong generation number, or are invalid.
+	 */
+	error = xfs_parent_from_attr(sc->mp, attr_flags, name, namelen, value,
+			valuelen, &parent_ino, &parent_gen);
+	if (error)
+		return error;
+
+	if (parent_ino != sc->ip->i_ino ||
+	    parent_gen != VFS_I(sc->ip)->i_generation)
+		return 0;
+
+	mutex_lock(&rd->pscan.lock);
+	error = xrep_dir_stash_createname(rd, &xname, ip->i_ino);
+	mutex_unlock(&rd->pscan.lock);
+	return error;
+}
+
+/*
+ * If this child dirent points to the directory being repaired, remember that
+ * fact so that we can reset the dotdot entry if necessary.
+ */
+STATIC int
+xrep_dir_scan_dirent(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*dp,
+	xfs_dir2_dataptr_t	dapos,
+	const struct xfs_name	*name,
+	xfs_ino_t		ino,
+	void			*priv)
+{
+	struct xrep_dir		*rd = priv;
+
+	/* Dirent doesn't point to this directory. */
+	if (ino != rd->sc->ip->i_ino)
+		return 0;
+
+	/* Ignore garbage inum. */
+	if (!xfs_verify_dir_ino(rd->sc->mp, ino))
+		return 0;
+
+	/* No weird looking names. */
+	if (name->len >= MAXNAMELEN || name->len <= 0)
+		return 0;
+
+	/* Don't pick up dot or dotdot entries; we only want child dirents. */
+	if (xfs_dir2_samename(name, &xfs_name_dotdot) ||
+	    xfs_dir2_samename(name, &xfs_name_dot))
+		return 0;
+
+	trace_xrep_dir_stash_createname(sc->tempip, &xfs_name_dotdot,
+			dp->i_ino);
+
+	xrep_findparent_scan_found(&rd->pscan, dp->i_ino);
+	return 0;
+}
+
+/*
+ * Decide if we want to look for child dirents or parent pointers in this file.
+ * Skip the dir being repaired and any files being used to stage repairs.
+ */
+static inline bool
+xrep_dir_want_scan(
+	struct xrep_dir		*rd,
+	const struct xfs_inode	*ip)
+{
+	return ip != rd->sc->ip && !xrep_is_tempfile(ip);
+}
+
+/*
+ * Take ILOCK on a file that we want to scan.
+ *
+ * Select ILOCK_EXCL if the file is a directory with an unloaded data bmbt or
+ * has an unloaded attr bmbt.  Otherwise, take ILOCK_SHARED.
+ */
+static inline unsigned int
+xrep_dir_scan_ilock(
+	struct xrep_dir		*rd,
+	struct xfs_inode	*ip)
+{
+	uint			lock_mode = XFS_ILOCK_SHARED;
+
+	/* Need to take the shared ILOCK to advance the iscan cursor. */
+	if (!xrep_dir_want_scan(rd, ip))
+		goto lock;
+
+	if (S_ISDIR(VFS_I(ip)->i_mode) && xfs_need_iread_extents(&ip->i_df)) {
+		lock_mode = XFS_ILOCK_EXCL;
+		goto lock;
+	}
+
+	if (xfs_inode_has_attr_fork(ip) && xfs_need_iread_extents(&ip->i_af))
+		lock_mode = XFS_ILOCK_EXCL;
+
+lock:
+	xfs_ilock(ip, lock_mode);
+	return lock_mode;
+}
+
+/*
+ * Scan this file for relevant child dirents or parent pointers that point to
+ * the directory we're rebuilding.
+ */
+STATIC int
+xrep_dir_scan_file(
+	struct xrep_dir		*rd,
+	struct xfs_inode	*ip)
+{
+	unsigned int		lock_mode;
+	int			error = 0;
+
+	lock_mode = xrep_dir_scan_ilock(rd, ip);
+
+	if (!xrep_dir_want_scan(rd, ip))
+		goto scan_done;
+
+	/*
+	 * If the extended attributes look as though they has been zapped by
+	 * the inode record repair code, we cannot scan for parent pointers.
+	 */
+	if (xchk_pptr_looks_zapped(ip)) {
+		error = -EBUSY;
+		goto scan_done;
+	}
+
+	error = xchk_xattr_walk(rd->sc, ip, xrep_dir_scan_pptr, NULL, rd);
+	if (error)
+		goto scan_done;
+
+	if (S_ISDIR(VFS_I(ip)->i_mode)) {
+		/*
+		 * If the directory looks as though it has been zapped by the
+		 * inode record repair code, we cannot scan for child dirents.
+		 */
+		if (xchk_dir_looks_zapped(ip)) {
+			error = -EBUSY;
+			goto scan_done;
+		}
+
+		error = xchk_dir_walk(rd->sc, ip, xrep_dir_scan_dirent, rd);
+		if (error)
+			goto scan_done;
+	}
+
+scan_done:
+	xchk_iscan_mark_visited(&rd->pscan.iscan, ip);
+	xfs_iunlock(ip, lock_mode);
+	return error;
+}
+
+/*
+ * Scan all files in the filesystem for parent pointers that we can turn into
+ * replacement dirents, and a dirent that we can use to set the dotdot pointer.
+ */
+STATIC int
+xrep_dir_scan_dirtree(
+	struct xrep_dir		*rd)
+{
+	struct xfs_scrub	*sc = rd->sc;
+	struct xfs_inode	*ip;
+	int			error;
+
+	/* Roots of directory trees are their own parents. */
+	if (sc->ip == sc->mp->m_rootip)
+		xrep_findparent_scan_found(&rd->pscan, sc->ip->i_ino);
+
+	/*
+	 * Filesystem scans are time consuming.  Drop the directory ILOCK and
+	 * all other resources for the duration of the scan and hope for the
+	 * best.  The live update hooks will keep our scan information up to
+	 * date even though we've dropped the locks.
+	 */
+	xchk_trans_cancel(sc);
+	if (sc->ilock_flags & (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL))
+		xchk_iunlock(sc, sc->ilock_flags & (XFS_ILOCK_SHARED |
+						    XFS_ILOCK_EXCL));
+	error = xchk_trans_alloc_empty(sc);
+	if (error)
+		return error;
+
+	while ((error = xchk_iscan_iter(&rd->pscan.iscan, &ip)) == 1) {
+		bool		flush;
+
+		error = xrep_dir_scan_file(rd, ip);
+		xchk_irele(sc, ip);
+		if (error)
+			break;
+
+		/* Flush stashed dirent updates to constrain memory usage. */
+		mutex_lock(&rd->pscan.lock);
+		flush = xrep_dir_want_flush_stashed(rd);
+		mutex_unlock(&rd->pscan.lock);
+		if (flush) {
+			xchk_trans_cancel(sc);
+
+			error = xrep_tempfile_iolock_polled(sc);
+			if (error)
+				break;
+
+			error = xrep_dir_replay_updates(rd);
+			xrep_tempfile_iounlock(sc);
+			if (error)
+				break;
+
+			error = xchk_trans_alloc_empty(sc);
+			if (error)
+				break;
+		}
+
+		if (xchk_should_terminate(sc, &error))
+			break;
+	}
+	xchk_iscan_iter_finish(&rd->pscan.iscan);
+	if (error) {
+		/*
+		 * If we couldn't grab an inode that was busy with a state
+		 * change, change the error code so that we exit to userspace
+		 * as quickly as possible.
+		 */
+		if (error == -EBUSY)
+			return -ECANCELED;
+		return error;
+	}
+
+	/*
+	 * Cancel the empty transaction so that we can (later) use the atomic
+	 * file mapping exchange functions to lock files and commit the new
+	 * directory.
+	 */
+	xchk_trans_cancel(rd->sc);
+	return 0;
+}
+
+/*
+ * Capture dirent updates being made by other threads which are relevant to the
+ * directory being repaired.
+ */
+STATIC int
+xrep_dir_live_update(
+	struct notifier_block		*nb,
+	unsigned long			action,
+	void				*data)
+{
+	struct xfs_dir_update_params	*p = data;
+	struct xrep_dir			*rd;
+	struct xfs_scrub		*sc;
+	int				error = 0;
+
+	rd = container_of(nb, struct xrep_dir, pscan.dhook.dirent_hook.nb);
+	sc = rd->sc;
+
+	/*
+	 * This thread updated a child dirent in the directory that we're
+	 * rebuilding.  Stash the update for replay against the temporary
+	 * directory.
+	 */
+	if (p->dp->i_ino == sc->ip->i_ino &&
+	    xchk_iscan_want_live_update(&rd->pscan.iscan, p->ip->i_ino)) {
+		mutex_lock(&rd->pscan.lock);
+		if (p->delta > 0)
+			error = xrep_dir_stash_createname(rd, p->name,
+					p->ip->i_ino);
+		else
+			error = xrep_dir_stash_removename(rd, p->name,
+					p->ip->i_ino);
+		mutex_unlock(&rd->pscan.lock);
+		if (error)
+			goto out_abort;
+	}
+
+	/*
+	 * This thread updated another directory's child dirent that points to
+	 * the directory that we're rebuilding, so remember the new dotdot
+	 * target.
+	 */
+	if (p->ip->i_ino == sc->ip->i_ino &&
+	    xchk_iscan_want_live_update(&rd->pscan.iscan, p->dp->i_ino)) {
+		if (p->delta > 0) {
+			trace_xrep_dir_stash_createname(sc->tempip,
+					&xfs_name_dotdot,
+					p->dp->i_ino);
+
+			xrep_findparent_scan_found(&rd->pscan, p->dp->i_ino);
+		} else {
+			trace_xrep_dir_stash_removename(sc->tempip,
+					&xfs_name_dotdot,
+					rd->pscan.parent_ino);
+
+			xrep_findparent_scan_found(&rd->pscan, NULLFSINO);
+		}
+	}
+
+	return NOTIFY_DONE;
+out_abort:
+	xchk_iscan_abort(&rd->pscan.iscan);
+	return NOTIFY_DONE;
+}
+
+/*
+ * Free all the directory blocks and reset the data fork.  The caller must
+ * join the inode to the transaction.  This function returns with the inode
+ * joined to a clean scrub transaction.
+ */
+STATIC int
+xrep_dir_reset_fork(
+	struct xrep_dir		*rd,
+	xfs_ino_t		parent_ino)
+{
+	struct xfs_scrub	*sc = rd->sc;
+	struct xfs_ifork	*ifp = xfs_ifork_ptr(sc->tempip, XFS_DATA_FORK);
+	int			error;
+
+	/* Unmap all the directory buffers. */
+	if (xfs_ifork_has_extents(ifp)) {
+		error = xrep_reap_ifork(sc, sc->tempip, XFS_DATA_FORK);
+		if (error)
+			return error;
+	}
+
+	trace_xrep_dir_reset_fork(sc->tempip, parent_ino);
+
+	/* Reset the data fork to an empty data fork. */
+	xfs_idestroy_fork(ifp);
+	ifp->if_bytes = 0;
+	sc->tempip->i_disk_size = 0;
+
+	/* Reinitialize the short form directory. */
+	xrep_dir_init_args(rd, sc->tempip, NULL);
+	return xfs_dir2_sf_create(&rd->args, parent_ino);
+}
+
+/*
+ * Prepare both inodes' directory forks for exchanging mappings.  Promote the
+ * tempfile from short format to leaf format, and if the file being repaired
+ * has a short format data fork, turn it into an empty extent list.
+ */
+STATIC int
+xrep_dir_swap_prep(
+	struct xfs_scrub	*sc,
+	bool			temp_local,
+	bool			ip_local)
+{
+	int			error;
+
+	/*
+	 * If the tempfile's directory is in shortform format, convert that to
+	 * a single leaf extent so that we can use the atomic mapping exchange.
+	 */
+	if (temp_local) {
+		struct xfs_da_args	args = {
+			.dp		= sc->tempip,
+			.geo		= sc->mp->m_dir_geo,
+			.whichfork	= XFS_DATA_FORK,
+			.trans		= sc->tp,
+			.total		= 1,
+			.owner		= sc->ip->i_ino,
+		};
+
+		error = xfs_dir2_sf_to_block(&args);
+		if (error)
+			return error;
+
+		/*
+		 * Roll the deferred log items to get us back to a clean
+		 * transaction.
+		 */
+		error = xfs_defer_finish(&sc->tp);
+		if (error)
+			return error;
+	}
+
+	/*
+	 * If the file being repaired had a shortform data fork, convert that
+	 * to an empty extent list in preparation for the atomic mapping
+	 * exchange.
+	 */
+	if (ip_local) {
+		struct xfs_ifork	*ifp;
+
+		ifp = xfs_ifork_ptr(sc->ip, XFS_DATA_FORK);
+		xfs_idestroy_fork(ifp);
+		ifp->if_format = XFS_DINODE_FMT_EXTENTS;
+		ifp->if_nextents = 0;
+		ifp->if_bytes = 0;
+		ifp->if_data = NULL;
+		ifp->if_height = 0;
+
+		xfs_trans_log_inode(sc->tp, sc->ip,
+				XFS_ILOG_CORE | XFS_ILOG_DDATA);
+	}
+
+	return 0;
+}
+
+/*
+ * Replace the inode number of a directory entry.
+ */
+static int
+xrep_dir_replace(
+	struct xrep_dir		*rd,
+	struct xfs_inode	*dp,
+	const struct xfs_name	*name,
+	xfs_ino_t		inum,
+	xfs_extlen_t		total)
+{
+	struct xfs_scrub	*sc = rd->sc;
+	int			error;
+
+	ASSERT(S_ISDIR(VFS_I(dp)->i_mode));
+
+	error = xfs_dir_ino_validate(sc->mp, inum);
+	if (error)
+		return error;
+
+	xrep_dir_init_args(rd, dp, name);
+	rd->args.inumber = inum;
+	rd->args.total = total;
+	return xfs_dir_replace_args(&rd->args);
+}
+
+/*
+ * Reset the link count of this directory and adjust the unlinked list pointers
+ * as needed.
+ */
+STATIC int
+xrep_dir_set_nlink(
+	struct xrep_dir		*rd)
+{
+	struct xfs_scrub	*sc = rd->sc;
+	struct xfs_inode	*dp = sc->ip;
+	struct xfs_perag	*pag;
+	unsigned int		new_nlink = min_t(unsigned long long,
+						  rd->subdirs + 2,
+						  XFS_NLINK_PINNED);
+	int			error;
+
+	/*
+	 * The directory is not on the incore unlinked list, which means that
+	 * it needs to be reachable via the directory tree.  Update the nlink
+	 * with our observed link count.  If the directory has no parent, it
+	 * will be moved to the orphanage.
+	 */
+	if (!xfs_inode_on_unlinked_list(dp))
+		goto reset_nlink;
+
+	/*
+	 * The directory is on the unlinked list and we did not find any
+	 * dirents.  Set the link count to zero and let the directory
+	 * inactivate when the last reference drops.
+	 */
+	if (rd->dirents == 0) {
+		rd->needs_adoption = false;
+		new_nlink = 0;
+		goto reset_nlink;
+	}
+
+	/*
+	 * The directory is on the unlinked list and we found dirents.  This
+	 * directory needs to be reachable via the directory tree.  Remove the
+	 * dir from the unlinked list and update nlink with the observed link
+	 * count.  If the directory has no parent, it will be moved to the
+	 * orphanage.
+	 */
+	pag = xfs_perag_get(sc->mp, XFS_INO_TO_AGNO(sc->mp, dp->i_ino));
+	if (!pag) {
+		ASSERT(0);
+		return -EFSCORRUPTED;
+	}
+
+	error = xfs_iunlink_remove(sc->tp, pag, dp);
+	xfs_perag_put(pag);
+	if (error)
+		return error;
+
+reset_nlink:
+	if (VFS_I(dp)->i_nlink != new_nlink)
+		set_nlink(VFS_I(dp), new_nlink);
+	return 0;
+}
+
+/*
+ * Finish replaying stashed dirent updates, allocate a transaction for
+ * exchanging data fork mappings, and take the ILOCKs of both directories
+ * before we commit the new directory structure.
+ */
+STATIC int
+xrep_dir_finalize_tempdir(
+	struct xrep_dir		*rd)
+{
+	struct xfs_scrub	*sc = rd->sc;
+	int			error;
+
+	if (!xfs_has_parent(sc->mp))
+		return xrep_tempexch_trans_alloc(sc, XFS_DATA_FORK, &rd->tx);
+
+	/*
+	 * Repair relies on the ILOCK to quiesce all possible dirent updates.
+	 * Replay all queued dirent updates into the tempdir before exchanging
+	 * the contents, even if that means dropping the ILOCKs and the
+	 * transaction.
+	 */
+	do {
+		error = xrep_dir_replay_updates(rd);
+		if (error)
+			return error;
+
+		error = xrep_tempexch_trans_alloc(sc, XFS_DATA_FORK, &rd->tx);
+		if (error)
+			return error;
+
+		if (xfarray_length(rd->dir_entries) == 0)
+			break;
+
+		xchk_trans_cancel(sc);
+		xrep_tempfile_iunlock_both(sc);
+	} while (!xchk_should_terminate(sc, &error));
+	return error;
+}
+
+/* Exchange the temporary directory's data fork with the one being repaired. */
+STATIC int
+xrep_dir_swap(
+	struct xrep_dir		*rd)
+{
+	struct xfs_scrub	*sc = rd->sc;
+	bool			ip_local, temp_local;
+	int			error = 0;
+
+	/*
+	 * If we never found the parent for this directory, temporarily assign
+	 * the root dir as the parent; we'll move this to the orphanage after
+	 * exchanging the dir contents.  We hold the ILOCK of the dir being
+	 * repaired, so we're not worried about racy updates of dotdot.
+	 */
+	ASSERT(sc->ilock_flags & XFS_ILOCK_EXCL);
+	if (rd->pscan.parent_ino == NULLFSINO) {
+		rd->needs_adoption = true;
+		rd->pscan.parent_ino = rd->sc->mp->m_sb.sb_rootino;
+	}
+
+	/*
+	 * Reset the temporary directory's '..' entry to point to the parent
+	 * that we found.  The temporary directory was created with the root
+	 * directory as the parent, so we can skip this if repairing a
+	 * subdirectory of the root.
+	 *
+	 * It's also possible that this replacement could also expand a sf
+	 * tempdir into block format.
+	 */
+	if (rd->pscan.parent_ino != sc->mp->m_rootip->i_ino) {
+		error = xrep_dir_replace(rd, rd->sc->tempip, &xfs_name_dotdot,
+				rd->pscan.parent_ino, rd->tx.req.resblks);
+		if (error)
+			return error;
+	}
+
+	/*
+	 * Changing the dot and dotdot entries could have changed the shape of
+	 * the directory, so we recompute these.
+	 */
+	ip_local = sc->ip->i_df.if_format == XFS_DINODE_FMT_LOCAL;
+	temp_local = sc->tempip->i_df.if_format == XFS_DINODE_FMT_LOCAL;
+
+	/*
+	 * If the both files have a local format data fork and the rebuilt
+	 * directory data would fit in the repaired file's data fork, copy
+	 * the contents from the tempfile and update the directory link count.
+	 * We're done now.
+	 */
+	if (ip_local && temp_local &&
+	    sc->tempip->i_disk_size <= xfs_inode_data_fork_size(sc->ip)) {
+		xrep_tempfile_copyout_local(sc, XFS_DATA_FORK);
+		return xrep_dir_set_nlink(rd);
+	}
+
+	/*
+	 * Clean the transaction before we start working on exchanging
+	 * directory contents.
+	 */
+	error = xrep_tempfile_roll_trans(rd->sc);
+	if (error)
+		return error;
+
+	/* Otherwise, make sure both data forks are in block-mapping mode. */
+	error = xrep_dir_swap_prep(sc, temp_local, ip_local);
+	if (error)
+		return error;
+
+	/*
+	 * Set nlink of the directory in the same transaction sequence that
+	 * (atomically) commits the new directory data.
+	 */
+	error = xrep_dir_set_nlink(rd);
+	if (error)
+		return error;
+
+	return xrep_tempexch_contents(sc, &rd->tx);
+}
+
+/*
+ * Exchange the new directory contents (which we created in the tempfile) with
+ * the directory being repaired.
+ */
+STATIC int
+xrep_dir_rebuild_tree(
+	struct xrep_dir		*rd)
+{
+	struct xfs_scrub	*sc = rd->sc;
+	int			error;
+
+	trace_xrep_dir_rebuild_tree(sc->ip, rd->pscan.parent_ino);
+
+	/*
+	 * Take the IOLOCK on the temporary file so that we can run dir
+	 * operations with the same locks held as we would for a normal file.
+	 * We still hold sc->ip's IOLOCK.
+	 */
+	error = xrep_tempfile_iolock_polled(rd->sc);
+	if (error)
+		return error;
+
+	/*
+	 * Allocate transaction, lock inodes, and make sure that we've replayed
+	 * all the stashed dirent updates to the tempdir.  After this point,
+	 * we're ready to exchange data fork mappings.
+	 */
+	error = xrep_dir_finalize_tempdir(rd);
+	if (error)
+		return error;
+
+	if (xchk_iscan_aborted(&rd->pscan.iscan))
+		return -ECANCELED;
+
+	/*
+	 * Exchange the tempdir's data fork with the file being repaired.  This
+	 * recreates the transaction and re-takes the ILOCK in the scrub
+	 * context.
+	 */
+	error = xrep_dir_swap(rd);
+	if (error)
+		return error;
+
+	/*
+	 * Release the old directory blocks and reset the data fork of the temp
+	 * directory to an empty shortform directory because inactivation does
+	 * nothing for directories.
+	 */
+	error = xrep_dir_reset_fork(rd, sc->mp->m_rootip->i_ino);
+	if (error)
+		return error;
+
+	/*
+	 * Roll to get a transaction without any inodes joined to it.  Then we
+	 * can drop the tempfile's ILOCK and IOLOCK before doing more work on
+	 * the scrub target directory.
+	 */
+	error = xfs_trans_roll(&sc->tp);
+	if (error)
+		return error;
+
+	xrep_tempfile_iunlock(sc);
+	xrep_tempfile_iounlock(sc);
+	return 0;
+}
+
+/* Set up the filesystem scan so we can regenerate directory entries. */
+STATIC int
+xrep_dir_setup_scan(
+	struct xrep_dir		*rd)
+{
+	struct xfs_scrub	*sc = rd->sc;
+	char			*descr;
+	int			error;
+
+	/* Set up some staging memory for salvaging dirents. */
+	descr = xchk_xfile_ino_descr(sc, "directory entries");
+	error = xfarray_create(descr, 0, sizeof(struct xrep_dirent),
+			&rd->dir_entries);
+	kfree(descr);
+	if (error)
+		return error;
+
+	descr = xchk_xfile_ino_descr(sc, "directory entry names");
+	error = xfblob_create(descr, &rd->dir_names);
+	kfree(descr);
+	if (error)
+		goto out_xfarray;
+
+	if (xfs_has_parent(sc->mp))
+		error = __xrep_findparent_scan_start(sc, &rd->pscan,
+				xrep_dir_live_update);
+	else
+		error = xrep_findparent_scan_start(sc, &rd->pscan);
+	if (error)
+		goto out_xfblob;
+
+	return 0;
+
+out_xfblob:
+	xfblob_destroy(rd->dir_names);
+	rd->dir_names = NULL;
+out_xfarray:
+	xfarray_destroy(rd->dir_entries);
+	rd->dir_entries = NULL;
+	return error;
+}
+
+/*
+ * Move the current file to the orphanage.
+ *
+ * Caller must hold IOLOCK_EXCL on @sc->ip, and no other inode locks.  Upon
+ * successful return, the scrub transaction will have enough extra reservation
+ * to make the move; it will hold IOLOCK_EXCL and ILOCK_EXCL of @sc->ip and the
+ * orphanage; and both inodes will be ijoined.
+ */
+STATIC int
+xrep_dir_move_to_orphanage(
+	struct xrep_dir		*rd)
+{
+	struct xfs_scrub	*sc = rd->sc;
+	xfs_ino_t		orig_parent, new_parent;
+	int			error;
+
+	/*
+	 * We are about to drop the ILOCK on sc->ip to lock the orphanage and
+	 * prepare for the adoption.  Therefore, look up the old dotdot entry
+	 * for sc->ip so that we can compare it after we re-lock sc->ip.
+	 */
+	error = xchk_dir_lookup(sc, sc->ip, &xfs_name_dotdot, &orig_parent);
+	if (error)
+		return error;
+
+	/*
+	 * Drop the ILOCK on the scrub target and commit the transaction.
+	 * Adoption computes its own resource requirements and gathers the
+	 * necessary components.
+	 */
+	error = xrep_trans_commit(sc);
+	if (error)
+		return error;
+	xchk_iunlock(sc, XFS_ILOCK_EXCL);
+
+	/* If we can take the orphanage's iolock then we're ready to move. */
+	if (!xrep_orphanage_ilock_nowait(sc, XFS_IOLOCK_EXCL)) {
+		xchk_iunlock(sc, sc->ilock_flags);
+		error = xrep_orphanage_iolock_two(sc);
+		if (error)
+			return error;
+	}
+
+	/* Grab transaction and ILOCK the two files. */
+	error = xrep_adoption_trans_alloc(sc, &rd->adoption);
+	if (error)
+		return error;
+
+	error = xrep_adoption_compute_name(&rd->adoption, &rd->xname);
+	if (error)
+		return error;
+
+	/*
+	 * Now that we've reacquired the ILOCK on sc->ip, look up the dotdot
+	 * entry again.  If the parent changed or the child was unlinked while
+	 * the child directory was unlocked, we don't need to move the child to
+	 * the orphanage after all.
+	 */
+	error = xchk_dir_lookup(sc, sc->ip, &xfs_name_dotdot, &new_parent);
+	if (error)
+		return error;
+
+	/*
+	 * Attach to the orphanage if we still have a linked directory and it
+	 * hasn't been moved.
+	 */
+	if (orig_parent == new_parent && VFS_I(sc->ip)->i_nlink > 0) {
+		error = xrep_adoption_move(&rd->adoption);
+		if (error)
+			return error;
+	}
+
+	/*
+	 * Launder the scrub transaction so we can drop the orphanage ILOCK
+	 * and IOLOCK.  Return holding the scrub target's ILOCK and IOLOCK.
+	 */
+	error = xrep_adoption_trans_roll(&rd->adoption);
+	if (error)
+		return error;
+
+	xrep_orphanage_iunlock(sc, XFS_ILOCK_EXCL);
+	xrep_orphanage_iunlock(sc, XFS_IOLOCK_EXCL);
+	return 0;
+}
+
+/*
+ * Repair the directory metadata.
+ *
+ * XXX: Directory entry buffers can be multiple fsblocks in size.  The buffer
+ * cache in XFS can't handle aliased multiblock buffers, so this might
+ * misbehave if the directory blocks are crosslinked with other filesystem
+ * metadata.
+ *
+ * XXX: Is it necessary to check the dcache for this directory to make sure
+ * that we always recreate every cached entry?
+ */
+int
+xrep_directory(
+	struct xfs_scrub	*sc)
+{
+	struct xrep_dir		*rd = sc->buf;
+	int			error;
+
+	/* The rmapbt is required to reap the old data fork. */
+	if (!xfs_has_rmapbt(sc->mp))
+		return -EOPNOTSUPP;
+	/* We require atomic file exchange range to rebuild anything. */
+	if (!xfs_has_exchange_range(sc->mp))
+		return -EOPNOTSUPP;
+
+	error = xrep_dir_setup_scan(rd);
+	if (error)
+		return error;
+
+	if (xfs_has_parent(sc->mp))
+		error = xrep_dir_scan_dirtree(rd);
+	else
+		error = xrep_dir_salvage_entries(rd);
+	if (error)
+		goto out_teardown;
+
+	/* Last chance to abort before we start committing fixes. */
+	if (xchk_should_terminate(sc, &error))
+		goto out_teardown;
+
+	error = xrep_dir_rebuild_tree(rd);
+	if (error)
+		goto out_teardown;
+
+	if (rd->needs_adoption) {
+		if (!xrep_orphanage_can_adopt(rd->sc))
+			error = -EFSCORRUPTED;
+		else
+			error = xrep_dir_move_to_orphanage(rd);
+		if (error)
+			goto out_teardown;
+	}
+
+out_teardown:
+	xrep_dir_teardown(sc);
+	return error;
+}
diff --git a/fs/xfs/scrub/dirtree.c b/fs/xfs/scrub/dirtree.c
new file mode 100644
index 0000000000..bde58fb561
--- /dev/null
+++ b/fs/xfs/scrub/dirtree.c
@@ -0,0 +1,985 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2023-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_inode.h"
+#include "xfs_icache.h"
+#include "xfs_dir2.h"
+#include "xfs_dir2_priv.h"
+#include "xfs_attr.h"
+#include "xfs_parent.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/bitmap.h"
+#include "scrub/ino_bitmap.h"
+#include "scrub/xfile.h"
+#include "scrub/xfarray.h"
+#include "scrub/xfblob.h"
+#include "scrub/listxattr.h"
+#include "scrub/trace.h"
+#include "scrub/repair.h"
+#include "scrub/orphanage.h"
+#include "scrub/dirtree.h"
+
+/*
+ * Directory Tree Structure Validation
+ * ===================================
+ *
+ * Validating the tree qualities of the directory tree structure can be
+ * difficult.  If the tree is frozen, running a depth (or breadth) first search
+ * and marking a bitmap suffices to determine if there is a cycle.  XORing the
+ * mark bitmap with the inode bitmap afterwards tells us if there are
+ * disconnected cycles.  If the tree is not frozen, directory updates can move
+ * subtrees across the scanner wavefront, which complicates the design greatly.
+ *
+ * Directory parent pointers change that by enabling an incremental approach to
+ * validation of the tree structure.  Instead of using one thread to scan the
+ * entire filesystem, we instead can have multiple threads walking individual
+ * subdirectories upwards to the root.  In a perfect world, the IOLOCK would
+ * suffice to stabilize two directories in a parent -> child relationship.
+ * Unfortunately, the VFS does not take the IOLOCK when moving a child
+ * subdirectory, so we instead synchronize on ILOCK and use dirent update hooks
+ * to detect a race.  If a race occurs in a path, we restart the scan.
+ *
+ * If the walk terminates without reaching the root, we know the path is
+ * disconnected and ought to be attached to the lost and found.  If on the walk
+ * we find the same subdir that we're scanning, we know this is a cycle and
+ * should delete an incoming edge.  If we find multiple paths to the root, we
+ * know to delete an incoming edge.
+ *
+ * There are two big hitches with this approach: first, all file link counts
+ * must be correct to prevent other writers from doing the wrong thing with the
+ * directory tree structure.  Second, because we're walking upwards in a tree
+ * of arbitrary depth, we cannot hold all the ILOCKs.  Instead, we will use a
+ * directory update hook to invalidate the scan results if one of the paths
+ * we've scanned has changed.
+ */
+
+/* Clean up the dirtree checking resources. */
+STATIC void
+xchk_dirtree_buf_cleanup(
+	void			*buf)
+{
+	struct xchk_dirtree	*dl = buf;
+	struct xchk_dirpath	*path, *n;
+
+	if (dl->scan_ino != NULLFSINO)
+		xfs_dir_hook_del(dl->sc->mp, &dl->dhook);
+
+	xchk_dirtree_for_each_path_safe(dl, path, n) {
+		list_del_init(&path->list);
+		xino_bitmap_destroy(&path->seen_inodes);
+		kfree(path);
+	}
+
+	xfblob_destroy(dl->path_names);
+	xfarray_destroy(dl->path_steps);
+	mutex_destroy(&dl->lock);
+}
+
+/* Set us up to look for directory loops. */
+int
+xchk_setup_dirtree(
+	struct xfs_scrub	*sc)
+{
+	struct xchk_dirtree	*dl;
+	char			*descr;
+	int			error;
+
+	xchk_fsgates_enable(sc, XCHK_FSGATES_DIRENTS);
+
+	if (xchk_could_repair(sc)) {
+		error = xrep_setup_dirtree(sc);
+		if (error)
+			return error;
+	}
+
+	dl = kvzalloc(sizeof(struct xchk_dirtree), XCHK_GFP_FLAGS);
+	if (!dl)
+		return -ENOMEM;
+	dl->sc = sc;
+	dl->xname.name = dl->namebuf;
+	dl->hook_xname.name = dl->hook_namebuf;
+	INIT_LIST_HEAD(&dl->path_list);
+	dl->root_ino = NULLFSINO;
+	dl->scan_ino = NULLFSINO;
+	dl->parent_ino = NULLFSINO;
+
+	mutex_init(&dl->lock);
+
+	descr = xchk_xfile_ino_descr(sc, "dirtree path steps");
+	error = xfarray_create(descr, 0, sizeof(struct xchk_dirpath_step),
+			&dl->path_steps);
+	kfree(descr);
+	if (error)
+		goto out_dl;
+
+	descr = xchk_xfile_ino_descr(sc, "dirtree path names");
+	error = xfblob_create(descr, &dl->path_names);
+	kfree(descr);
+	if (error)
+		goto out_steps;
+
+	error = xchk_setup_inode_contents(sc, 0);
+	if (error)
+		goto out_names;
+
+	sc->buf = dl;
+	sc->buf_cleanup = xchk_dirtree_buf_cleanup;
+	return 0;
+
+out_names:
+	xfblob_destroy(dl->path_names);
+out_steps:
+	xfarray_destroy(dl->path_steps);
+out_dl:
+	mutex_destroy(&dl->lock);
+	kvfree(dl);
+	return error;
+}
+
+/*
+ * Add the parent pointer described by @dl->pptr to the given path as a new
+ * step.  Returns -ELNRNG if the path is too deep.
+ */
+int
+xchk_dirpath_append(
+	struct xchk_dirtree		*dl,
+	struct xfs_inode		*ip,
+	struct xchk_dirpath		*path,
+	const struct xfs_name		*name,
+	const struct xfs_parent_rec	*pptr)
+{
+	struct xchk_dirpath_step	step = {
+		.pptr_rec		= *pptr, /* struct copy */
+		.name_len		= name->len,
+	};
+	int				error;
+
+	/*
+	 * If this path is more than 2 billion steps long, this directory tree
+	 * is too far gone to fix.
+	 */
+	if (path->nr_steps >= XFS_MAXLINK)
+		return -ELNRNG;
+
+	error = xfblob_storename(dl->path_names, &step.name_cookie, name);
+	if (error)
+		return error;
+
+	error = xino_bitmap_set(&path->seen_inodes, ip->i_ino);
+	if (error)
+		return error;
+
+	error = xfarray_append(dl->path_steps, &step);
+	if (error)
+		return error;
+
+	path->nr_steps++;
+	return 0;
+}
+
+/*
+ * Create an xchk_path for each parent pointer of the directory that we're
+ * scanning.  For each path created, we will eventually try to walk towards the
+ * root with the goal of deleting all parents except for one that leads to the
+ * root.
+ *
+ * Returns -EFSCORRUPTED to signal that the inode being scanned has a corrupt
+ * parent pointer and hence there's no point in continuing; or -ENOSR if there
+ * are too many parent pointers for this directory.
+ */
+STATIC int
+xchk_dirtree_create_path(
+	struct xfs_scrub		*sc,
+	struct xfs_inode		*ip,
+	unsigned int			attr_flags,
+	const unsigned char		*name,
+	unsigned int			namelen,
+	const void			*value,
+	unsigned int			valuelen,
+	void				*priv)
+{
+	struct xfs_name			xname = {
+		.name			= name,
+		.len			= namelen,
+	};
+	struct xchk_dirtree		*dl = priv;
+	struct xchk_dirpath		*path;
+	const struct xfs_parent_rec	*rec = value;
+	int				error;
+
+	if (!(attr_flags & XFS_ATTR_PARENT))
+		return 0;
+
+	error = xfs_parent_from_attr(sc->mp, attr_flags, name, namelen, value,
+			valuelen, NULL, NULL);
+	if (error)
+		return error;
+
+	/*
+	 * If there are more than 2 billion actual parent pointers for this
+	 * subdirectory, this fs is too far gone to fix.
+	 */
+	if (dl->nr_paths >= XFS_MAXLINK)
+		return -ENOSR;
+
+	trace_xchk_dirtree_create_path(sc, ip, dl->nr_paths, &xname, rec);
+
+	/*
+	 * Create a new xchk_path structure to remember this parent pointer
+	 * and record the first name step.
+	 */
+	path = kmalloc(sizeof(struct xchk_dirpath), XCHK_GFP_FLAGS);
+	if (!path)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&path->list);
+	xino_bitmap_init(&path->seen_inodes);
+	path->nr_steps = 0;
+	path->outcome = XCHK_DIRPATH_SCANNING;
+
+	error = xchk_dirpath_append(dl, sc->ip, path, &xname, rec);
+	if (error)
+		goto out_path;
+
+	path->first_step = xfarray_length(dl->path_steps) - 1;
+	path->second_step = XFARRAY_NULLIDX;
+	path->path_nr = dl->nr_paths;
+
+	list_add_tail(&path->list, &dl->path_list);
+	dl->nr_paths++;
+	return 0;
+out_path:
+	kfree(path);
+	return error;
+}
+
+/*
+ * Validate that the first step of this path still has a corresponding
+ * parent pointer in @sc->ip.  We probably dropped @sc->ip's ILOCK while
+ * walking towards the roots, which is why this is necessary.
+ *
+ * This function has a side effect of loading the first parent pointer of this
+ * path into the parent pointer scratch pad.  This prepares us to walk up the
+ * directory tree towards the root.  Returns -ESTALE if the scan data is now
+ * out of date.
+ */
+STATIC int
+xchk_dirpath_revalidate(
+	struct xchk_dirtree		*dl,
+	struct xchk_dirpath		*path)
+{
+	struct xfs_scrub		*sc = dl->sc;
+	int				error;
+
+	/*
+	 * Look up the parent pointer that corresponds to the start of this
+	 * path.  If the parent pointer has disappeared on us, dump all the
+	 * scan results and try again.
+	 */
+	error = xfs_parent_lookup(sc->tp, sc->ip, &dl->xname, &dl->pptr_rec,
+			&dl->pptr_args);
+	if (error == -ENOATTR) {
+		trace_xchk_dirpath_disappeared(dl->sc, sc->ip, path->path_nr,
+				path->first_step, &dl->xname, &dl->pptr_rec);
+		dl->stale = true;
+		return -ESTALE;
+	}
+
+	return error;
+}
+
+/*
+ * Walk the parent pointers of a directory at the end of a path and record
+ * the parent that we find in @dl->xname/pptr_rec.
+ */
+STATIC int
+xchk_dirpath_find_next_step(
+	struct xfs_scrub		*sc,
+	struct xfs_inode		*ip,
+	unsigned int			attr_flags,
+	const unsigned char		*name,
+	unsigned int			namelen,
+	const void			*value,
+	unsigned int			valuelen,
+	void				*priv)
+{
+	struct xchk_dirtree		*dl = priv;
+	const struct xfs_parent_rec	*rec = value;
+	int				error;
+
+	if (!(attr_flags & XFS_ATTR_PARENT))
+		return 0;
+
+	error = xfs_parent_from_attr(sc->mp, attr_flags, name, namelen, value,
+			valuelen, NULL, NULL);
+	if (error)
+		return error;
+
+	/*
+	 * If we've already set @dl->pptr_rec, then this directory has multiple
+	 * parents.  Signal this back to the caller via -EMLINK.
+	 */
+	if (dl->parents_found > 0)
+		return -EMLINK;
+
+	dl->parents_found++;
+	memcpy(dl->namebuf, name, namelen);
+	dl->xname.len = namelen;
+	dl->pptr_rec = *rec; /* struct copy */
+	return 0;
+}
+
+/* Set and log the outcome of a path walk. */
+static inline void
+xchk_dirpath_set_outcome(
+	struct xchk_dirtree		*dl,
+	struct xchk_dirpath		*path,
+	enum xchk_dirpath_outcome	outcome)
+{
+	trace_xchk_dirpath_set_outcome(dl->sc, path->path_nr, path->nr_steps,
+			outcome);
+
+	path->outcome = outcome;
+}
+
+/*
+ * Scan the directory at the end of this path for its parent directory link.
+ * If we find one, extend the path.  Returns -ESTALE if the scan data out of
+ * date.  Returns -EFSCORRUPTED if the parent pointer is bad; or -ELNRNG if
+ * the path got too deep.
+ */
+STATIC int
+xchk_dirpath_step_up(
+	struct xchk_dirtree	*dl,
+	struct xchk_dirpath	*path)
+{
+	struct xfs_scrub	*sc = dl->sc;
+	struct xfs_inode	*dp;
+	xfs_ino_t		parent_ino = be64_to_cpu(dl->pptr_rec.p_ino);
+	unsigned int		lock_mode;
+	int			error;
+
+	/* Grab and lock the parent directory. */
+	error = xchk_iget(sc, parent_ino, &dp);
+	if (error)
+		return error;
+
+	lock_mode = xfs_ilock_attr_map_shared(dp);
+	mutex_lock(&dl->lock);
+
+	if (dl->stale) {
+		error = -ESTALE;
+		goto out_scanlock;
+	}
+
+	/* We've reached the root directory; the path is ok. */
+	if (parent_ino == dl->root_ino) {
+		xchk_dirpath_set_outcome(dl, path, XCHK_DIRPATH_OK);
+		error = 0;
+		goto out_scanlock;
+	}
+
+	/*
+	 * The inode being scanned is its own distant ancestor!  Get rid of
+	 * this path.
+	 */
+	if (parent_ino == sc->ip->i_ino) {
+		xchk_dirpath_set_outcome(dl, path, XCHK_DIRPATH_DELETE);
+		error = 0;
+		goto out_scanlock;
+	}
+
+	/*
+	 * We've seen this inode before during the path walk.  There's a loop
+	 * above us in the directory tree.  This probably means that we cannot
+	 * continue, but let's keep walking paths to get a full picture.
+	 */
+	if (xino_bitmap_test(&path->seen_inodes, parent_ino)) {
+		xchk_dirpath_set_outcome(dl, path, XCHK_DIRPATH_LOOP);
+		error = 0;
+		goto out_scanlock;
+	}
+
+	/* The handle encoded in the parent pointer must match. */
+	if (VFS_I(dp)->i_generation != be32_to_cpu(dl->pptr_rec.p_gen)) {
+		trace_xchk_dirpath_badgen(dl->sc, dp, path->path_nr,
+				path->nr_steps, &dl->xname, &dl->pptr_rec);
+		error = -EFSCORRUPTED;
+		goto out_scanlock;
+	}
+
+	/* Parent pointer must point up to a directory. */
+	if (!S_ISDIR(VFS_I(dp)->i_mode)) {
+		trace_xchk_dirpath_nondir_parent(dl->sc, dp, path->path_nr,
+				path->nr_steps, &dl->xname, &dl->pptr_rec);
+		error = -EFSCORRUPTED;
+		goto out_scanlock;
+	}
+
+	/* Parent cannot be an unlinked directory. */
+	if (VFS_I(dp)->i_nlink == 0) {
+		trace_xchk_dirpath_unlinked_parent(dl->sc, dp, path->path_nr,
+				path->nr_steps, &dl->xname, &dl->pptr_rec);
+		error = -EFSCORRUPTED;
+		goto out_scanlock;
+	}
+
+	/*
+	 * If the extended attributes look as though they has been zapped by
+	 * the inode record repair code, we cannot scan for parent pointers.
+	 */
+	if (xchk_pptr_looks_zapped(dp)) {
+		error = -EBUSY;
+		xchk_set_incomplete(sc);
+		goto out_scanlock;
+	}
+
+	/*
+	 * Walk the parent pointers of @dp to find the parent of this directory
+	 * to find the next step in our walk.  If we find that @dp has exactly
+	 * one parent, the parent pointer information will be stored in
+	 * @dl->pptr_rec.  This prepares us for the next step of the walk.
+	 */
+	mutex_unlock(&dl->lock);
+	dl->parents_found = 0;
+	error = xchk_xattr_walk(sc, dp, xchk_dirpath_find_next_step, NULL, dl);
+	mutex_lock(&dl->lock);
+	if (error == -EFSCORRUPTED || error == -EMLINK ||
+	    (!error && dl->parents_found == 0)) {
+		/*
+		 * Further up the directory tree from @sc->ip, we found a
+		 * corrupt parent pointer, multiple parent pointers while
+		 * finding this directory's parent, or zero parents despite
+		 * having a nonzero link count.  Keep looking for other paths.
+		 */
+		xchk_dirpath_set_outcome(dl, path, XCHK_DIRPATH_CORRUPT);
+		error = 0;
+		goto out_scanlock;
+	}
+	if (error)
+		goto out_scanlock;
+
+	if (dl->stale) {
+		error = -ESTALE;
+		goto out_scanlock;
+	}
+
+	trace_xchk_dirpath_found_next_step(sc, dp, path->path_nr,
+			path->nr_steps, &dl->xname, &dl->pptr_rec);
+
+	/* Append to the path steps */
+	error = xchk_dirpath_append(dl, dp, path, &dl->xname, &dl->pptr_rec);
+	if (error)
+		goto out_scanlock;
+
+	if (path->second_step == XFARRAY_NULLIDX)
+		path->second_step = xfarray_length(dl->path_steps) - 1;
+
+out_scanlock:
+	mutex_unlock(&dl->lock);
+	xfs_iunlock(dp, lock_mode);
+	xchk_irele(sc, dp);
+	return error;
+}
+
+/*
+ * Walk the directory tree upwards towards what is hopefully the root
+ * directory, recording path steps as we go.  The current path components are
+ * stored in dl->pptr_rec and dl->xname.
+ *
+ * Returns -ESTALE if the scan data are out of date.  Returns -EFSCORRUPTED
+ * only if the direct parent pointer of @sc->ip associated with this path is
+ * corrupt.
+ */
+STATIC int
+xchk_dirpath_walk_upwards(
+	struct xchk_dirtree	*dl,
+	struct xchk_dirpath	*path)
+{
+	struct xfs_scrub	*sc = dl->sc;
+	int			error;
+
+	ASSERT(sc->ilock_flags & XFS_ILOCK_EXCL);
+
+	/* Reload the start of this path and make sure it's still there. */
+	error = xchk_dirpath_revalidate(dl, path);
+	if (error)
+		return error;
+
+	trace_xchk_dirpath_walk_upwards(sc, sc->ip, path->path_nr, &dl->xname,
+			&dl->pptr_rec);
+
+	/*
+	 * The inode being scanned is its own direct ancestor!
+	 * Get rid of this path.
+	 */
+	if (be64_to_cpu(dl->pptr_rec.p_ino) == sc->ip->i_ino) {
+		xchk_dirpath_set_outcome(dl, path, XCHK_DIRPATH_DELETE);
+		return 0;
+	}
+
+	/*
+	 * Drop ILOCK_EXCL on the inode being scanned.  We still hold
+	 * IOLOCK_EXCL on it, so it cannot move around or be renamed.
+	 *
+	 * Beyond this point we're walking up the directory tree, which means
+	 * that we can acquire and drop the ILOCK on an alias of sc->ip.  The
+	 * ILOCK state is no longer tracked in the scrub context.  Hence we
+	 * must drop @sc->ip's ILOCK during the walk.
+	 */
+	mutex_unlock(&dl->lock);
+	xchk_iunlock(sc, XFS_ILOCK_EXCL);
+
+	/*
+	 * Take the first step in the walk towards the root by checking the
+	 * start of this path, which is a direct parent pointer of @sc->ip.
+	 * If we see any kind of error here (including corruptions), the parent
+	 * pointer of @sc->ip is corrupt.  Stop the whole scan.
+	 */
+	error = xchk_dirpath_step_up(dl, path);
+	if (error) {
+		xchk_ilock(sc, XFS_ILOCK_EXCL);
+		mutex_lock(&dl->lock);
+		return error;
+	}
+
+	/*
+	 * Take steps upward from the second step in this path towards the
+	 * root.  If we hit corruption errors here, there's a problem
+	 * *somewhere* in the path, but we don't need to stop scanning.
+	 */
+	while (!error && path->outcome == XCHK_DIRPATH_SCANNING)
+		error = xchk_dirpath_step_up(dl, path);
+
+	/* Retake the locks we had, mark paths, etc. */
+	xchk_ilock(sc, XFS_ILOCK_EXCL);
+	mutex_lock(&dl->lock);
+	if (error == -EFSCORRUPTED) {
+		xchk_dirpath_set_outcome(dl, path, XCHK_DIRPATH_CORRUPT);
+		error = 0;
+	}
+	if (!error && dl->stale)
+		return -ESTALE;
+	return error;
+}
+
+/*
+ * Decide if this path step has been touched by this live update.  Returns
+ * 1 for yes, 0 for no, or a negative errno.
+ */
+STATIC int
+xchk_dirpath_step_is_stale(
+	struct xchk_dirtree		*dl,
+	struct xchk_dirpath		*path,
+	unsigned int			step_nr,
+	xfarray_idx_t			step_idx,
+	struct xfs_dir_update_params	*p,
+	xfs_ino_t			*cursor)
+{
+	struct xchk_dirpath_step	step;
+	xfs_ino_t			child_ino = *cursor;
+	int				error;
+
+	error = xfarray_load(dl->path_steps, step_idx, &step);
+	if (error)
+		return error;
+	*cursor = be64_to_cpu(step.pptr_rec.p_ino);
+
+	/*
+	 * If the parent and child being updated are not the ones mentioned in
+	 * this path step, the scan data is still ok.
+	 */
+	if (p->ip->i_ino != child_ino || p->dp->i_ino != *cursor)
+		return 0;
+
+	/*
+	 * If the dirent name lengths or byte sequences are different, the scan
+	 * data is still ok.
+	 */
+	if (p->name->len != step.name_len)
+		return 0;
+
+	error = xfblob_loadname(dl->path_names, step.name_cookie,
+			&dl->hook_xname, step.name_len);
+	if (error)
+		return error;
+
+	if (memcmp(dl->hook_xname.name, p->name->name, p->name->len) != 0)
+		return 0;
+
+	/*
+	 * If the update comes from the repair code itself, walk the state
+	 * machine forward.
+	 */
+	if (p->ip->i_ino == dl->scan_ino &&
+	    path->outcome == XREP_DIRPATH_ADOPTING) {
+		xchk_dirpath_set_outcome(dl, path, XREP_DIRPATH_ADOPTED);
+		return 0;
+	}
+
+	if (p->ip->i_ino == dl->scan_ino &&
+	    path->outcome == XREP_DIRPATH_DELETING) {
+		xchk_dirpath_set_outcome(dl, path, XREP_DIRPATH_DELETED);
+		return 0;
+	}
+
+	/* Exact match, scan data is out of date. */
+	trace_xchk_dirpath_changed(dl->sc, path->path_nr, step_nr, p->dp,
+			p->ip, p->name);
+	return 1;
+}
+
+/*
+ * Decide if this path has been touched by this live update.  Returns 1 for
+ * yes, 0 for no, or a negative errno.
+ */
+STATIC int
+xchk_dirpath_is_stale(
+	struct xchk_dirtree		*dl,
+	struct xchk_dirpath		*path,
+	struct xfs_dir_update_params	*p)
+{
+	xfs_ino_t			cursor = dl->scan_ino;
+	xfarray_idx_t			idx = path->first_step;
+	unsigned int			i;
+	int				ret;
+
+	/*
+	 * The child being updated has not been seen by this path at all; this
+	 * path cannot be stale.
+	 */
+	if (!xino_bitmap_test(&path->seen_inodes, p->ip->i_ino))
+		return 0;
+
+	ret = xchk_dirpath_step_is_stale(dl, path, 0, idx, p, &cursor);
+	if (ret != 0)
+		return ret;
+
+	for (i = 1, idx = path->second_step; i < path->nr_steps; i++, idx++) {
+		ret = xchk_dirpath_step_is_stale(dl, path, i, idx, p, &cursor);
+		if (ret != 0)
+			return ret;
+	}
+
+	return 0;
+}
+
+/*
+ * Decide if a directory update from the regular filesystem touches any of the
+ * paths we've scanned, and invalidate the scan data if true.
+ */
+STATIC int
+xchk_dirtree_live_update(
+	struct notifier_block		*nb,
+	unsigned long			action,
+	void				*data)
+{
+	struct xfs_dir_update_params	*p = data;
+	struct xchk_dirtree		*dl;
+	struct xchk_dirpath		*path;
+	int				ret;
+
+	dl = container_of(nb, struct xchk_dirtree, dhook.dirent_hook.nb);
+
+	trace_xchk_dirtree_live_update(dl->sc, p->dp, action, p->ip, p->delta,
+			p->name);
+
+	mutex_lock(&dl->lock);
+
+	if (dl->stale || dl->aborted)
+		goto out_unlock;
+
+	xchk_dirtree_for_each_path(dl, path) {
+		ret = xchk_dirpath_is_stale(dl, path, p);
+		if (ret < 0) {
+			dl->aborted = true;
+			break;
+		}
+		if (ret == 1) {
+			dl->stale = true;
+			break;
+		}
+	}
+
+out_unlock:
+	mutex_unlock(&dl->lock);
+	return NOTIFY_DONE;
+}
+
+/* Delete all the collected path information. */
+STATIC void
+xchk_dirtree_reset(
+	void			*buf)
+{
+	struct xchk_dirtree	*dl = buf;
+	struct xchk_dirpath	*path, *n;
+
+	ASSERT(dl->sc->ilock_flags & XFS_ILOCK_EXCL);
+
+	xchk_dirtree_for_each_path_safe(dl, path, n) {
+		list_del_init(&path->list);
+		xino_bitmap_destroy(&path->seen_inodes);
+		kfree(path);
+	}
+	dl->nr_paths = 0;
+
+	xfarray_truncate(dl->path_steps);
+	xfblob_truncate(dl->path_names);
+
+	dl->stale = false;
+}
+
+/*
+ * Load the name/pptr from the first step in this path into @dl->pptr_rec and
+ * @dl->xname.
+ */
+STATIC int
+xchk_dirtree_load_path(
+	struct xchk_dirtree		*dl,
+	struct xchk_dirpath		*path)
+{
+	struct xchk_dirpath_step	step;
+	int				error;
+
+	error = xfarray_load(dl->path_steps, path->first_step, &step);
+	if (error)
+		return error;
+
+	error = xfblob_loadname(dl->path_names, step.name_cookie, &dl->xname,
+			step.name_len);
+	if (error)
+		return error;
+
+	dl->pptr_rec = step.pptr_rec; /* struct copy */
+	return 0;
+}
+
+/*
+ * For each parent pointer of this subdir, trace a path upwards towards the
+ * root directory and record what we find.  Returns 0 for success;
+ * -EFSCORRUPTED if walking the parent pointers of @sc->ip failed, -ELNRNG if a
+ * path was too deep; -ENOSR if there were too many parent pointers; or
+ * a negative errno.
+ */
+int
+xchk_dirtree_find_paths_to_root(
+	struct xchk_dirtree	*dl)
+{
+	struct xfs_scrub	*sc = dl->sc;
+	struct xchk_dirpath	*path;
+	int			error = 0;
+
+	do {
+		if (xchk_should_terminate(sc, &error))
+			return error;
+
+		xchk_dirtree_reset(dl);
+
+		/*
+		 * If the extended attributes look as though they has been
+		 * zapped by the inode record repair code, we cannot scan for
+		 * parent pointers.
+		 */
+		if (xchk_pptr_looks_zapped(sc->ip)) {
+			xchk_set_incomplete(sc);
+			return -EBUSY;
+		}
+
+		/*
+		 * Create path walk contexts for each parent of the directory
+		 * that is being scanned.  Directories are supposed to have
+		 * only one parent, but this is how we detect multiple parents.
+		 */
+		error = xchk_xattr_walk(sc, sc->ip, xchk_dirtree_create_path,
+				NULL, dl);
+		if (error)
+			return error;
+
+		xchk_dirtree_for_each_path(dl, path) {
+			/* Load path components into dl->pptr/xname */
+			error = xchk_dirtree_load_path(dl, path);
+			if (error)
+				return error;
+
+			/*
+			 * Try to walk up each path to the root.  This enables
+			 * us to find directory loops in ancestors, and the
+			 * like.
+			 */
+			error = xchk_dirpath_walk_upwards(dl, path);
+			if (error == -EFSCORRUPTED) {
+				/*
+				 * A parent pointer of @sc->ip is bad, don't
+				 * bother continuing.
+				 */
+				break;
+			}
+			if (error == -ESTALE) {
+				/* This had better be an invalidation. */
+				ASSERT(dl->stale);
+				break;
+			}
+			if (error)
+				return error;
+			if (dl->aborted)
+				return 0;
+		}
+	} while (dl->stale);
+
+	return error;
+}
+
+/*
+ * Figure out what to do with the paths we tried to find.  Do not call this
+ * if the scan results are stale.
+ */
+void
+xchk_dirtree_evaluate(
+	struct xchk_dirtree		*dl,
+	struct xchk_dirtree_outcomes	*oc)
+{
+	struct xchk_dirpath		*path;
+
+	ASSERT(!dl->stale);
+
+	/* Scan the paths we have to decide what to do. */
+	memset(oc, 0, sizeof(struct xchk_dirtree_outcomes));
+	xchk_dirtree_for_each_path(dl, path) {
+		trace_xchk_dirpath_evaluate_path(dl->sc, path->path_nr,
+				path->nr_steps, path->outcome);
+
+		switch (path->outcome) {
+		case XCHK_DIRPATH_SCANNING:
+			/* shouldn't get here */
+			ASSERT(0);
+			break;
+		case XCHK_DIRPATH_DELETE:
+			/* This one is already going away. */
+			oc->bad++;
+			break;
+		case XCHK_DIRPATH_CORRUPT:
+		case XCHK_DIRPATH_LOOP:
+			/* Couldn't find the end of this path. */
+			oc->suspect++;
+			break;
+		case XCHK_DIRPATH_STALE:
+			/* shouldn't get here either */
+			ASSERT(0);
+			break;
+		case XCHK_DIRPATH_OK:
+			/* This path got all the way to the root. */
+			oc->good++;
+			break;
+		case XREP_DIRPATH_DELETING:
+		case XREP_DIRPATH_DELETED:
+		case XREP_DIRPATH_ADOPTING:
+		case XREP_DIRPATH_ADOPTED:
+			/* These should not be in progress! */
+			ASSERT(0);
+			break;
+		}
+	}
+
+	trace_xchk_dirtree_evaluate(dl, oc);
+}
+
+/* Look for directory loops. */
+int
+xchk_dirtree(
+	struct xfs_scrub		*sc)
+{
+	struct xchk_dirtree_outcomes	oc;
+	struct xchk_dirtree		*dl = sc->buf;
+	int				error;
+
+	/*
+	 * Nondirectories do not point downwards to other files, so they cannot
+	 * cause a cycle in the directory tree.
+	 */
+	if (!S_ISDIR(VFS_I(sc->ip)->i_mode))
+		return -ENOENT;
+
+	ASSERT(xfs_has_parent(sc->mp));
+
+	/*
+	 * Find the root of the directory tree.  Remember which directory to
+	 * scan, because the hook doesn't detach until after sc->ip gets
+	 * released during teardown.
+	 */
+	dl->root_ino = sc->mp->m_rootip->i_ino;
+	dl->scan_ino = sc->ip->i_ino;
+
+	trace_xchk_dirtree_start(sc->ip, sc->sm, 0);
+
+	/*
+	 * Hook into the directory entry code so that we can capture updates to
+	 * paths that we have already scanned.  The scanner thread takes each
+	 * directory's ILOCK, which means that any in-progress directory update
+	 * will finish before we can scan the directory.
+	 */
+	ASSERT(sc->flags & XCHK_FSGATES_DIRENTS);
+	xfs_dir_hook_setup(&dl->dhook, xchk_dirtree_live_update);
+	error = xfs_dir_hook_add(sc->mp, &dl->dhook);
+	if (error)
+		goto out;
+
+	mutex_lock(&dl->lock);
+
+	/* Trace each parent pointer's path to the root. */
+	error = xchk_dirtree_find_paths_to_root(dl);
+	if (error == -EFSCORRUPTED || error == -ELNRNG || error == -ENOSR) {
+		/*
+		 * Don't bother walking the paths if the xattr structure or the
+		 * parent pointers are corrupt; this scan cannot be completed
+		 * without full information.
+		 */
+		xchk_ino_xref_set_corrupt(sc, sc->ip->i_ino);
+		error = 0;
+		goto out_scanlock;
+	}
+	if (error == -EBUSY) {
+		/*
+		 * We couldn't scan some directory's parent pointers because
+		 * the attr fork looked like it had been zapped.  The
+		 * scan was marked incomplete, so no further error code
+		 * is necessary.
+		 */
+		error = 0;
+		goto out_scanlock;
+	}
+	if (error)
+		goto out_scanlock;
+	if (dl->aborted) {
+		xchk_set_incomplete(sc);
+		goto out_scanlock;
+	}
+
+	/* Assess what we found in our path evaluation. */
+	xchk_dirtree_evaluate(dl, &oc);
+	if (xchk_dirtree_parentless(dl)) {
+		if (oc.good || oc.bad || oc.suspect)
+			xchk_ino_set_corrupt(sc, sc->ip->i_ino);
+	} else {
+		if (oc.bad || oc.good + oc.suspect != 1)
+			xchk_ino_set_corrupt(sc, sc->ip->i_ino);
+		if (oc.suspect)
+			xchk_ino_xref_set_corrupt(sc, sc->ip->i_ino);
+	}
+
+out_scanlock:
+	mutex_unlock(&dl->lock);
+out:
+	trace_xchk_dirtree_done(sc->ip, sc->sm, error);
+	return error;
+}
diff --git a/fs/xfs/scrub/dirtree.h b/fs/xfs/scrub/dirtree.h
new file mode 100644
index 0000000000..1e1686365c
--- /dev/null
+++ b/fs/xfs/scrub/dirtree.h
@@ -0,0 +1,178 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (c) 2023-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_DIRTREE_H__
+#define __XFS_SCRUB_DIRTREE_H__
+
+/*
+ * Each of these represents one parent pointer path step in a chain going
+ * up towards the directory tree root.  These are stored inside an xfarray.
+ */
+struct xchk_dirpath_step {
+	/* Directory entry name associated with this parent link. */
+	xfblob_cookie		name_cookie;
+	unsigned int		name_len;
+
+	/* Handle of the parent directory. */
+	struct xfs_parent_rec	pptr_rec;
+};
+
+enum xchk_dirpath_outcome {
+	XCHK_DIRPATH_SCANNING = 0,	/* still being put together */
+	XCHK_DIRPATH_DELETE,		/* delete this path */
+	XCHK_DIRPATH_CORRUPT,		/* corruption detected in path */
+	XCHK_DIRPATH_LOOP,		/* cycle detected further up */
+	XCHK_DIRPATH_STALE,		/* path is stale */
+	XCHK_DIRPATH_OK,		/* path reaches the root */
+
+	XREP_DIRPATH_DELETING,		/* path is being deleted */
+	XREP_DIRPATH_DELETED,		/* path has been deleted */
+	XREP_DIRPATH_ADOPTING,		/* path is being adopted */
+	XREP_DIRPATH_ADOPTED,		/* path has been adopted */
+};
+
+/*
+ * Each of these represents one parent pointer path out of the directory being
+ * scanned.  These exist in-core, and hopefully there aren't more than a
+ * handful of them.
+ */
+struct xchk_dirpath {
+	struct list_head	list;
+
+	/* Index of the first step in this path. */
+	xfarray_idx_t		first_step;
+
+	/* Index of the second step in this path. */
+	xfarray_idx_t		second_step;
+
+	/* Inodes seen while walking this path. */
+	struct xino_bitmap	seen_inodes;
+
+	/* Number of steps in this path. */
+	unsigned int		nr_steps;
+
+	/* Which path is this? */
+	unsigned int		path_nr;
+
+	/* What did we conclude from following this path? */
+	enum xchk_dirpath_outcome outcome;
+};
+
+struct xchk_dirtree_outcomes {
+	/* Number of XCHK_DIRPATH_DELETE */
+	unsigned int		bad;
+
+	/* Number of XCHK_DIRPATH_CORRUPT or XCHK_DIRPATH_LOOP */
+	unsigned int		suspect;
+
+	/* Number of XCHK_DIRPATH_OK */
+	unsigned int		good;
+
+	/* Directory needs to be added to lost+found */
+	bool			needs_adoption;
+};
+
+struct xchk_dirtree {
+	struct xfs_scrub	*sc;
+
+	/* Root inode that we're looking for. */
+	xfs_ino_t		root_ino;
+
+	/*
+	 * This is the inode that we're scanning.  The live update hook can
+	 * continue to be called after xchk_teardown drops sc->ip but before
+	 * it calls buf_cleanup, so we keep a copy.
+	 */
+	xfs_ino_t		scan_ino;
+
+	/*
+	 * If we start deleting redundant paths to this subdirectory, this is
+	 * the inode number of the surviving parent and the dotdot entry will
+	 * be set to this value.  If the value is NULLFSINO, then use @root_ino
+	 * as a stand-in until the orphanage can adopt the subdirectory.
+	 */
+	xfs_ino_t		parent_ino;
+
+	/* Scratch buffer for scanning pptr xattrs */
+	struct xfs_parent_rec	pptr_rec;
+	struct xfs_da_args	pptr_args;
+
+	/* Name buffer */
+	struct xfs_name		xname;
+	char			namebuf[MAXNAMELEN];
+
+	/* Information for reparenting this directory. */
+	struct xrep_adoption	adoption;
+
+	/*
+	 * Hook into directory updates so that we can receive live updates
+	 * from other writer threads.
+	 */
+	struct xfs_dir_hook	dhook;
+
+	/* Parent pointer update arguments. */
+	struct xfs_parent_args	ppargs;
+
+	/* lock for everything below here */
+	struct mutex		lock;
+
+	/* buffer for the live update functions to use for dirent names */
+	struct xfs_name		hook_xname;
+	unsigned char		hook_namebuf[MAXNAMELEN];
+
+	/*
+	 * All path steps observed during this scan.  Each of the path
+	 * steps for a particular pathwalk are recorded in sequential
+	 * order in the xfarray.  A pathwalk ends either with a step
+	 * pointing to the root directory (success) or pointing to NULLFSINO
+	 * (loop detected, empty dir detected, etc).
+	 */
+	struct xfarray		*path_steps;
+
+	/* All names observed during this scan. */
+	struct xfblob		*path_names;
+
+	/* All paths being tracked by this scanner. */
+	struct list_head	path_list;
+
+	/* Number of paths in path_list. */
+	unsigned int		nr_paths;
+
+	/* Number of parents found by a pptr scan. */
+	unsigned int		parents_found;
+
+	/* Have the path data been invalidated by a concurrent update? */
+	bool			stale:1;
+
+	/* Has the scan been aborted? */
+	bool			aborted:1;
+};
+
+#define xchk_dirtree_for_each_path_safe(dl, path, n) \
+	list_for_each_entry_safe((path), (n), &(dl)->path_list, list)
+
+#define xchk_dirtree_for_each_path(dl, path) \
+	list_for_each_entry((path), &(dl)->path_list, list)
+
+static inline bool
+xchk_dirtree_parentless(const struct xchk_dirtree *dl)
+{
+	struct xfs_scrub	*sc = dl->sc;
+
+	if (sc->ip == sc->mp->m_rootip)
+		return true;
+	if (VFS_I(sc->ip)->i_nlink == 0)
+		return true;
+	return false;
+}
+
+int xchk_dirtree_find_paths_to_root(struct xchk_dirtree *dl);
+int xchk_dirpath_append(struct xchk_dirtree *dl, struct xfs_inode *ip,
+		struct xchk_dirpath *path, const struct xfs_name *name,
+		const struct xfs_parent_rec *pptr);
+void xchk_dirtree_evaluate(struct xchk_dirtree *dl,
+		struct xchk_dirtree_outcomes *oc);
+
+#endif /* __XFS_SCRUB_DIRTREE_H__ */
diff --git a/fs/xfs/scrub/dirtree_repair.c b/fs/xfs/scrub/dirtree_repair.c
new file mode 100644
index 0000000000..5c04e70ba9
--- /dev/null
+++ b/fs/xfs/scrub/dirtree_repair.c
@@ -0,0 +1,821 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2023-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_trans_space.h"
+#include "xfs_mount.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_inode.h"
+#include "xfs_icache.h"
+#include "xfs_dir2.h"
+#include "xfs_dir2_priv.h"
+#include "xfs_attr.h"
+#include "xfs_parent.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/bitmap.h"
+#include "scrub/ino_bitmap.h"
+#include "scrub/xfile.h"
+#include "scrub/xfarray.h"
+#include "scrub/xfblob.h"
+#include "scrub/listxattr.h"
+#include "scrub/trace.h"
+#include "scrub/repair.h"
+#include "scrub/orphanage.h"
+#include "scrub/dirtree.h"
+#include "scrub/readdir.h"
+
+/*
+ * Directory Tree Structure Repairs
+ * ================================
+ *
+ * If we decide that the directory being scanned is participating in a
+ * directory loop, the only change we can make is to remove directory entries
+ * pointing down to @sc->ip.  If that leaves it with no parents, the directory
+ * should be adopted by the orphanage.
+ */
+
+/* Set up to repair directory loops. */
+int
+xrep_setup_dirtree(
+	struct xfs_scrub	*sc)
+{
+	return xrep_orphanage_try_create(sc);
+}
+
+/* Change the outcome of this path. */
+static inline void
+xrep_dirpath_set_outcome(
+	struct xchk_dirtree		*dl,
+	struct xchk_dirpath		*path,
+	enum xchk_dirpath_outcome	outcome)
+{
+	trace_xrep_dirpath_set_outcome(dl->sc, path->path_nr, path->nr_steps,
+			outcome);
+
+	path->outcome = outcome;
+}
+
+/* Delete all paths. */
+STATIC void
+xrep_dirtree_delete_all_paths(
+	struct xchk_dirtree		*dl,
+	struct xchk_dirtree_outcomes	*oc)
+{
+	struct xchk_dirpath		*path;
+
+	xchk_dirtree_for_each_path(dl, path) {
+		switch (path->outcome) {
+		case XCHK_DIRPATH_CORRUPT:
+		case XCHK_DIRPATH_LOOP:
+			oc->suspect--;
+			oc->bad++;
+			xrep_dirpath_set_outcome(dl, path, XCHK_DIRPATH_DELETE);
+			break;
+		case XCHK_DIRPATH_OK:
+			oc->good--;
+			oc->bad++;
+			xrep_dirpath_set_outcome(dl, path, XCHK_DIRPATH_DELETE);
+			break;
+		default:
+			break;
+		}
+	}
+
+	ASSERT(oc->suspect == 0);
+	ASSERT(oc->good == 0);
+}
+
+/* Since this is the surviving path, set the dotdot entry to this value. */
+STATIC void
+xrep_dirpath_retain_parent(
+	struct xchk_dirtree		*dl,
+	struct xchk_dirpath		*path)
+{
+	struct xchk_dirpath_step	step;
+	int				error;
+
+	error = xfarray_load(dl->path_steps, path->first_step, &step);
+	if (error)
+		return;
+
+	dl->parent_ino = be64_to_cpu(step.pptr_rec.p_ino);
+}
+
+/* Find the one surviving path so we know how to set dotdot. */
+STATIC void
+xrep_dirtree_find_surviving_path(
+	struct xchk_dirtree		*dl,
+	struct xchk_dirtree_outcomes	*oc)
+{
+	struct xchk_dirpath		*path;
+	bool				foundit = false;
+
+	xchk_dirtree_for_each_path(dl, path) {
+		switch (path->outcome) {
+		case XCHK_DIRPATH_CORRUPT:
+		case XCHK_DIRPATH_LOOP:
+		case XCHK_DIRPATH_OK:
+			if (!foundit) {
+				xrep_dirpath_retain_parent(dl, path);
+				foundit = true;
+				continue;
+			}
+			ASSERT(foundit == false);
+			break;
+		default:
+			break;
+		}
+	}
+
+	ASSERT(oc->suspect + oc->good == 1);
+}
+
+/* Delete all paths except for the one good one. */
+STATIC void
+xrep_dirtree_keep_one_good_path(
+	struct xchk_dirtree		*dl,
+	struct xchk_dirtree_outcomes	*oc)
+{
+	struct xchk_dirpath		*path;
+	bool				foundit = false;
+
+	xchk_dirtree_for_each_path(dl, path) {
+		switch (path->outcome) {
+		case XCHK_DIRPATH_CORRUPT:
+		case XCHK_DIRPATH_LOOP:
+			oc->suspect--;
+			oc->bad++;
+			xrep_dirpath_set_outcome(dl, path, XCHK_DIRPATH_DELETE);
+			break;
+		case XCHK_DIRPATH_OK:
+			if (!foundit) {
+				xrep_dirpath_retain_parent(dl, path);
+				foundit = true;
+				continue;
+			}
+			oc->good--;
+			oc->bad++;
+			xrep_dirpath_set_outcome(dl, path, XCHK_DIRPATH_DELETE);
+			break;
+		default:
+			break;
+		}
+	}
+
+	ASSERT(oc->suspect == 0);
+	ASSERT(oc->good < 2);
+}
+
+/* Delete all paths except for one suspect one. */
+STATIC void
+xrep_dirtree_keep_one_suspect_path(
+	struct xchk_dirtree		*dl,
+	struct xchk_dirtree_outcomes	*oc)
+{
+	struct xchk_dirpath		*path;
+	bool				foundit = false;
+
+	xchk_dirtree_for_each_path(dl, path) {
+		switch (path->outcome) {
+		case XCHK_DIRPATH_CORRUPT:
+		case XCHK_DIRPATH_LOOP:
+			if (!foundit) {
+				xrep_dirpath_retain_parent(dl, path);
+				foundit = true;
+				continue;
+			}
+			oc->suspect--;
+			oc->bad++;
+			xrep_dirpath_set_outcome(dl, path, XCHK_DIRPATH_DELETE);
+			break;
+		case XCHK_DIRPATH_OK:
+			ASSERT(0);
+			break;
+		default:
+			break;
+		}
+	}
+
+	ASSERT(oc->suspect == 1);
+	ASSERT(oc->good == 0);
+}
+
+/*
+ * Figure out what to do with the paths we tried to find.  Returns -EDEADLOCK
+ * if the scan results have become stale.
+ */
+STATIC void
+xrep_dirtree_decide_fate(
+	struct xchk_dirtree		*dl,
+	struct xchk_dirtree_outcomes	*oc)
+{
+	xchk_dirtree_evaluate(dl, oc);
+
+	/* Parentless directories should not have any paths at all. */
+	if (xchk_dirtree_parentless(dl)) {
+		xrep_dirtree_delete_all_paths(dl, oc);
+		return;
+	}
+
+	/* One path is exactly the number of paths we want. */
+	if (oc->good + oc->suspect == 1) {
+		xrep_dirtree_find_surviving_path(dl, oc);
+		return;
+	}
+
+	/* Zero paths means we should reattach the subdir to the orphanage. */
+	if (oc->good + oc->suspect == 0) {
+		if (dl->sc->orphanage)
+			oc->needs_adoption = true;
+		return;
+	}
+
+	/*
+	 * Otherwise, this subdirectory has too many parents.  If there's at
+	 * least one good path, keep it and delete the others.
+	 */
+	if (oc->good > 0) {
+		xrep_dirtree_keep_one_good_path(dl, oc);
+		return;
+	}
+
+	/*
+	 * There are no good paths and there are too many suspect paths.
+	 * Keep the first suspect path and delete the rest.
+	 */
+	xrep_dirtree_keep_one_suspect_path(dl, oc);
+}
+
+/*
+ * Load the first step of this path into @step and @dl->xname/pptr
+ * for later repair work.
+ */
+STATIC int
+xrep_dirtree_prep_path(
+	struct xchk_dirtree		*dl,
+	struct xchk_dirpath		*path,
+	struct xchk_dirpath_step	*step)
+{
+	int				error;
+
+	error = xfarray_load(dl->path_steps, path->first_step, step);
+	if (error)
+		return error;
+
+	error = xfblob_loadname(dl->path_names, step->name_cookie, &dl->xname,
+			step->name_len);
+	if (error)
+		return error;
+
+	dl->pptr_rec = step->pptr_rec; /* struct copy */
+	return 0;
+}
+
+/* Delete the VFS dentry for a removed child. */
+STATIC int
+xrep_dirtree_purge_dentry(
+	struct xchk_dirtree	*dl,
+	struct xfs_inode	*dp,
+	const struct xfs_name	*name)
+{
+	struct qstr		qname = QSTR_INIT(name->name, name->len);
+	struct dentry		*parent_dentry, *child_dentry;
+	int			error = 0;
+
+	/*
+	 * Find the dentry for the parent directory.  If there isn't one, we're
+	 * done.  Caller already holds i_rwsem for parent and child.
+	 */
+	parent_dentry = d_find_alias(VFS_I(dp));
+	if (!parent_dentry)
+		return 0;
+
+	/* The VFS thinks the parent is a directory, right? */
+	if (!d_is_dir(parent_dentry)) {
+		ASSERT(d_is_dir(parent_dentry));
+		error = -EFSCORRUPTED;
+		goto out_dput_parent;
+	}
+
+	/*
+	 * Try to find the dirent pointing to the child.  If there isn't one,
+	 * we're done.
+	 */
+	qname.hash = full_name_hash(parent_dentry, name->name, name->len);
+	child_dentry = d_lookup(parent_dentry, &qname);
+	if (!child_dentry) {
+		error = 0;
+		goto out_dput_parent;
+	}
+
+	trace_xrep_dirtree_delete_child(dp->i_mount, child_dentry);
+
+	/* Child is not a directory?  We're screwed. */
+	if (!d_is_dir(child_dentry)) {
+		ASSERT(d_is_dir(child_dentry));
+		error = -EFSCORRUPTED;
+		goto out_dput_child;
+	}
+
+	/* Replace the child dentry with a negative one. */
+	d_delete(child_dentry);
+
+out_dput_child:
+	dput(child_dentry);
+out_dput_parent:
+	dput(parent_dentry);
+	return error;
+}
+
+/*
+ * Prepare to delete a link by taking the IOLOCK of the parent and the child
+ * (scrub target).  Caller must hold IOLOCK_EXCL on @sc->ip.  Returns 0 if we
+ * took both locks, or a negative errno if we couldn't lock the parent in time.
+ */
+static inline int
+xrep_dirtree_unlink_iolock(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*dp)
+{
+	int			error;
+
+	ASSERT(sc->ilock_flags & XFS_IOLOCK_EXCL);
+
+	if (xfs_ilock_nowait(dp, XFS_IOLOCK_EXCL))
+		return 0;
+
+	xchk_iunlock(sc, XFS_IOLOCK_EXCL);
+	do {
+		xfs_ilock(dp, XFS_IOLOCK_EXCL);
+		if (xchk_ilock_nowait(sc, XFS_IOLOCK_EXCL))
+			break;
+		xfs_iunlock(dp, XFS_IOLOCK_EXCL);
+
+		if (xchk_should_terminate(sc, &error)) {
+			xchk_ilock(sc, XFS_IOLOCK_EXCL);
+			return error;
+		}
+
+		delay(1);
+	} while (1);
+
+	return 0;
+}
+
+/*
+ * Remove a link from the directory tree and update the dcache.  Returns
+ * -ESTALE if the scan data are now out of date.
+ */
+STATIC int
+xrep_dirtree_unlink(
+	struct xchk_dirtree		*dl,
+	struct xfs_inode		*dp,
+	struct xchk_dirpath		*path,
+	struct xchk_dirpath_step	*step)
+{
+	struct xfs_scrub		*sc = dl->sc;
+	struct xfs_mount		*mp = sc->mp;
+	xfs_ino_t			dotdot_ino;
+	xfs_ino_t			parent_ino = dl->parent_ino;
+	unsigned int			resblks;
+	int				dontcare;
+	int				error;
+
+	/* Take IOLOCK_EXCL of the parent and child. */
+	error = xrep_dirtree_unlink_iolock(sc, dp);
+	if (error)
+		return error;
+
+	/*
+	 * Create the transaction that we need to sever the path.  Ignore
+	 * EDQUOT and ENOSPC being returned via nospace_error because the
+	 * directory code can handle a reservationless update.
+	 */
+	resblks = xfs_remove_space_res(mp, step->name_len);
+	error = xfs_trans_alloc_dir(dp, &M_RES(mp)->tr_remove, sc->ip,
+			&resblks, &sc->tp, &dontcare);
+	if (error)
+		goto out_iolock;
+
+	/*
+	 * Cancel if someone invalidate the paths while we were trying to get
+	 * the ILOCK.
+	 */
+	mutex_lock(&dl->lock);
+	if (dl->stale) {
+		mutex_unlock(&dl->lock);
+		error = -ESTALE;
+		goto out_trans_cancel;
+	}
+	xrep_dirpath_set_outcome(dl, path, XREP_DIRPATH_DELETING);
+	mutex_unlock(&dl->lock);
+
+	trace_xrep_dirtree_delete_path(dl->sc, sc->ip, path->path_nr,
+			&dl->xname, &dl->pptr_rec);
+
+	/*
+	 * Decide if we need to reset the dotdot entry.  Rules:
+	 *
+	 * - If there's a surviving parent, we want dotdot to point there.
+	 * - If we don't have any surviving parents, then point dotdot at the
+	 *   root dir.
+	 * - If dotdot is already set to the value we want, pass in NULLFSINO
+	 *   for no change necessary.
+	 *
+	 * Do this /before/ we dirty anything, in case the dotdot lookup
+	 * fails.
+	 */
+	error = xchk_dir_lookup(sc, sc->ip, &xfs_name_dotdot, &dotdot_ino);
+	if (error)
+		goto out_trans_cancel;
+	if (parent_ino == NULLFSINO)
+		parent_ino = dl->root_ino;
+	if (dotdot_ino == parent_ino)
+		parent_ino = NULLFSINO;
+
+	/* Drop the link from sc->ip's dotdot entry.  */
+	error = xfs_droplink(sc->tp, dp);
+	if (error)
+		goto out_trans_cancel;
+
+	/* Reset the dotdot entry to a surviving parent. */
+	if (parent_ino != NULLFSINO) {
+		error = xfs_dir_replace(sc->tp, sc->ip, &xfs_name_dotdot,
+				parent_ino, 0);
+		if (error)
+			goto out_trans_cancel;
+	}
+
+	/* Drop the link from dp to sc->ip. */
+	error = xfs_droplink(sc->tp, sc->ip);
+	if (error)
+		goto out_trans_cancel;
+
+	error = xfs_dir_removename(sc->tp, dp, &dl->xname, sc->ip->i_ino,
+			resblks);
+	if (error) {
+		ASSERT(error != -ENOENT);
+		goto out_trans_cancel;
+	}
+
+	if (xfs_has_parent(sc->mp)) {
+		error = xfs_parent_removename(sc->tp, &dl->ppargs, dp,
+				&dl->xname, sc->ip);
+		if (error)
+			goto out_trans_cancel;
+	}
+
+	/*
+	 * Notify dirent hooks that we removed the bad link, invalidate the
+	 * dcache, and commit the repair.
+	 */
+	xfs_dir_update_hook(dp, sc->ip, -1, &dl->xname);
+	error = xrep_dirtree_purge_dentry(dl, dp, &dl->xname);
+	if (error)
+		goto out_trans_cancel;
+
+	error = xrep_trans_commit(sc);
+	goto out_ilock;
+
+out_trans_cancel:
+	xchk_trans_cancel(sc);
+out_ilock:
+	xfs_iunlock(sc->ip, XFS_ILOCK_EXCL);
+	xfs_iunlock(dp, XFS_ILOCK_EXCL);
+out_iolock:
+	xfs_iunlock(dp, XFS_IOLOCK_EXCL);
+	return error;
+}
+
+/*
+ * Delete a directory entry that points to this directory.  Returns -ESTALE
+ * if the scan data are now out of date.
+ */
+STATIC int
+xrep_dirtree_delete_path(
+	struct xchk_dirtree		*dl,
+	struct xchk_dirpath		*path)
+{
+	struct xchk_dirpath_step	step;
+	struct xfs_scrub		*sc = dl->sc;
+	struct xfs_inode		*dp;
+	int				error;
+
+	/*
+	 * Load the parent pointer and directory inode for this path, then
+	 * drop the scan lock, the ILOCK, and the transaction so that
+	 * _delete_path can reserve the proper transaction.  This sets up
+	 * @dl->xname for the deletion.
+	 */
+	error = xrep_dirtree_prep_path(dl, path, &step);
+	if (error)
+		return error;
+
+	error = xchk_iget(sc, be64_to_cpu(step.pptr_rec.p_ino), &dp);
+	if (error)
+		return error;
+
+	mutex_unlock(&dl->lock);
+	xchk_trans_cancel(sc);
+	xchk_iunlock(sc, XFS_ILOCK_EXCL);
+
+	/* Delete the directory link and release the parent. */
+	error = xrep_dirtree_unlink(dl, dp, path, &step);
+	xchk_irele(sc, dp);
+
+	/*
+	 * Retake all the resources we had at the beginning even if the repair
+	 * failed or the scan data are now stale.  This keeps things simple for
+	 * the caller.
+	 */
+	xchk_trans_alloc_empty(sc);
+	xchk_ilock(sc, XFS_ILOCK_EXCL);
+	mutex_lock(&dl->lock);
+
+	if (!error && dl->stale)
+		error = -ESTALE;
+	return error;
+}
+
+/* Add a new path to represent our in-progress adoption. */
+STATIC int
+xrep_dirtree_create_adoption_path(
+	struct xchk_dirtree		*dl)
+{
+	struct xfs_scrub		*sc = dl->sc;
+	struct xchk_dirpath		*path;
+	int				error;
+
+	/*
+	 * We should have capped the number of paths at XFS_MAXLINK-1 in the
+	 * scanner.
+	 */
+	if (dl->nr_paths > XFS_MAXLINK) {
+		ASSERT(dl->nr_paths <= XFS_MAXLINK);
+		return -EFSCORRUPTED;
+	}
+
+	/*
+	 * Create a new xchk_path structure to remember this parent pointer
+	 * and record the first name step.
+	 */
+	path = kmalloc(sizeof(struct xchk_dirpath), XCHK_GFP_FLAGS);
+	if (!path)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&path->list);
+	xino_bitmap_init(&path->seen_inodes);
+	path->nr_steps = 0;
+	path->outcome = XREP_DIRPATH_ADOPTING;
+
+	/*
+	 * Record the new link that we just created in the orphanage.  Because
+	 * adoption is the last repair that we perform, we don't bother filling
+	 * in the path all the way back to the root.
+	 */
+	xfs_inode_to_parent_rec(&dl->pptr_rec, sc->orphanage);
+
+	error = xino_bitmap_set(&path->seen_inodes, sc->orphanage->i_ino);
+	if (error)
+		goto out_path;
+
+	trace_xrep_dirtree_create_adoption(sc, sc->ip, dl->nr_paths,
+			&dl->xname, &dl->pptr_rec);
+
+	error = xchk_dirpath_append(dl, sc->ip, path, &dl->xname,
+			&dl->pptr_rec);
+	if (error)
+		goto out_path;
+
+	path->first_step = xfarray_length(dl->path_steps) - 1;
+	path->second_step = XFARRAY_NULLIDX;
+	path->path_nr = dl->nr_paths;
+
+	list_add_tail(&path->list, &dl->path_list);
+	dl->nr_paths++;
+	return 0;
+
+out_path:
+	kfree(path);
+	return error;
+}
+
+/*
+ * Prepare to move a file to the orphanage by taking the IOLOCK of the
+ * orphanage and the child (scrub target).  Caller must hold IOLOCK_EXCL on
+ * @sc->ip.  Returns 0 if we took both locks, or a negative errno if we
+ * couldn't lock the orphanage in time.
+ */
+static inline int
+xrep_dirtree_adopt_iolock(
+	struct xfs_scrub	*sc)
+{
+	int			error;
+
+	ASSERT(sc->ilock_flags & XFS_IOLOCK_EXCL);
+
+	if (xrep_orphanage_ilock_nowait(sc, XFS_IOLOCK_EXCL))
+		return 0;
+
+	xchk_iunlock(sc, XFS_IOLOCK_EXCL);
+	do {
+		xrep_orphanage_ilock(sc, XFS_IOLOCK_EXCL);
+		if (xchk_ilock_nowait(sc, XFS_IOLOCK_EXCL))
+			break;
+		xrep_orphanage_iunlock(sc, XFS_IOLOCK_EXCL);
+
+		if (xchk_should_terminate(sc, &error)) {
+			xchk_ilock(sc, XFS_IOLOCK_EXCL);
+			return error;
+		}
+
+		delay(1);
+	} while (1);
+
+	return 0;
+}
+
+/*
+ * Reattach this orphaned directory to the orphanage.  Do not call this with
+ * any resources held.  Returns -ESTALE if the scan data have become out of
+ * date.
+ */
+STATIC int
+xrep_dirtree_adopt(
+	struct xchk_dirtree		*dl)
+{
+	struct xfs_scrub		*sc = dl->sc;
+	int				error;
+
+	/* Take the IOLOCK of the orphanage and the scrub target. */
+	error = xrep_dirtree_adopt_iolock(sc);
+	if (error)
+		return error;
+
+	/*
+	 * Set up for an adoption.  The directory tree fixer runs after the
+	 * link counts have been corrected.  Therefore, we must bump the
+	 * child's link count since there will be no further opportunity to fix
+	 * errors.
+	 */
+	error = xrep_adoption_trans_alloc(sc, &dl->adoption);
+	if (error)
+		goto out_iolock;
+	dl->adoption.bump_child_nlink = true;
+
+	/* Figure out what name we're going to use here. */
+	error = xrep_adoption_compute_name(&dl->adoption, &dl->xname);
+	if (error)
+		goto out_trans;
+
+	/*
+	 * Now that we have a proposed name for the orphanage entry, create
+	 * a faux path so that the live update hook will see it.
+	 */
+	mutex_lock(&dl->lock);
+	if (dl->stale) {
+		mutex_unlock(&dl->lock);
+		error = -ESTALE;
+		goto out_trans;
+	}
+	error = xrep_dirtree_create_adoption_path(dl);
+	mutex_unlock(&dl->lock);
+	if (error)
+		goto out_trans;
+
+	/* Reparent the directory. */
+	error = xrep_adoption_move(&dl->adoption);
+	if (error)
+		goto out_trans;
+
+	/*
+	 * Commit the name and release all inode locks except for the scrub
+	 * target's IOLOCK.
+	 */
+	error = xrep_trans_commit(sc);
+	goto out_ilock;
+
+out_trans:
+	xchk_trans_cancel(sc);
+out_ilock:
+	xchk_iunlock(sc, XFS_ILOCK_EXCL);
+	xrep_orphanage_iunlock(sc, XFS_ILOCK_EXCL);
+out_iolock:
+	xrep_orphanage_iunlock(sc, XFS_IOLOCK_EXCL);
+	return error;
+}
+
+/*
+ * This newly orphaned directory needs to be adopted by the orphanage.
+ * Make this happen.
+ */
+STATIC int
+xrep_dirtree_move_to_orphanage(
+	struct xchk_dirtree		*dl)
+{
+	struct xfs_scrub		*sc = dl->sc;
+	int				error;
+
+	/*
+	 * Start by dropping all the resources that we hold so that we can grab
+	 * all the resources that we need for the adoption.
+	 */
+	mutex_unlock(&dl->lock);
+	xchk_trans_cancel(sc);
+	xchk_iunlock(sc, XFS_ILOCK_EXCL);
+
+	/* Perform the adoption. */
+	error = xrep_dirtree_adopt(dl);
+
+	/*
+	 * Retake all the resources we had at the beginning even if the repair
+	 * failed or the scan data are now stale.  This keeps things simple for
+	 * the caller.
+	 */
+	xchk_trans_alloc_empty(sc);
+	xchk_ilock(sc, XFS_ILOCK_EXCL);
+	mutex_lock(&dl->lock);
+
+	if (!error && dl->stale)
+		error = -ESTALE;
+	return error;
+}
+
+/*
+ * Try to fix all the problems.  Returns -ESTALE if the scan data have become
+ * out of date.
+ */
+STATIC int
+xrep_dirtree_fix_problems(
+	struct xchk_dirtree		*dl,
+	struct xchk_dirtree_outcomes	*oc)
+{
+	struct xchk_dirpath		*path;
+	int				error;
+
+	/* Delete all the paths we don't want. */
+	xchk_dirtree_for_each_path(dl, path) {
+		if (path->outcome != XCHK_DIRPATH_DELETE)
+			continue;
+
+		error = xrep_dirtree_delete_path(dl, path);
+		if (error)
+			return error;
+	}
+
+	/* Reparent this directory to the orphanage. */
+	if (oc->needs_adoption) {
+		if (xrep_orphanage_can_adopt(dl->sc))
+			return xrep_dirtree_move_to_orphanage(dl);
+		return -EFSCORRUPTED;
+	}
+
+	return 0;
+}
+
+/* Fix directory loops involving this directory. */
+int
+xrep_dirtree(
+	struct xfs_scrub		*sc)
+{
+	struct xchk_dirtree		*dl = sc->buf;
+	struct xchk_dirtree_outcomes	oc;
+	int				error;
+
+	/*
+	 * Prepare to fix the directory tree by retaking the scan lock.  The
+	 * order of resource acquisition is still IOLOCK -> transaction ->
+	 * ILOCK -> scan lock.
+	 */
+	mutex_lock(&dl->lock);
+	do {
+		/*
+		 * Decide what we're going to do, then do it.  An -ESTALE
+		 * return here means the scan results are invalid and we have
+		 * to walk again.
+		 */
+		if (!dl->stale) {
+			xrep_dirtree_decide_fate(dl, &oc);
+
+			trace_xrep_dirtree_decided_fate(dl, &oc);
+
+			error = xrep_dirtree_fix_problems(dl, &oc);
+			if (!error || error != -ESTALE)
+				break;
+		}
+		error = xchk_dirtree_find_paths_to_root(dl);
+		if (error == -ELNRNG || error == -ENOSR)
+			error = -EFSCORRUPTED;
+	} while (!error);
+	mutex_unlock(&dl->lock);
+
+	return error;
+}
diff --git a/fs/xfs/scrub/findparent.c b/fs/xfs/scrub/findparent.c
new file mode 100644
index 0000000000..01766041ba
--- /dev/null
+++ b/fs/xfs/scrub/findparent.c
@@ -0,0 +1,454 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2020-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_icache.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
+#include "xfs_dir2.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_dir2_priv.h"
+#include "xfs_trans_space.h"
+#include "xfs_health.h"
+#include "xfs_exchmaps.h"
+#include "xfs_parent.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+#include "scrub/repair.h"
+#include "scrub/iscan.h"
+#include "scrub/findparent.h"
+#include "scrub/readdir.h"
+#include "scrub/tempfile.h"
+#include "scrub/listxattr.h"
+
+/*
+ * Finding the Parent of a Directory
+ * =================================
+ *
+ * Directories have parent pointers, in the sense that each directory contains
+ * a dotdot entry that points to the single allowed parent.  The brute force
+ * way to find the parent of a given directory is to scan every directory in
+ * the filesystem looking for a child dirent that references this directory.
+ *
+ * This module wraps the process of scanning the directory tree.  It requires
+ * that @sc->ip is the directory whose parent we want to find, and that the
+ * caller hold only the IOLOCK on that directory.  The scan itself needs to
+ * take the ILOCK of each directory visited.
+ *
+ * Because we cannot hold @sc->ip's ILOCK during a scan of the whole fs, it is
+ * necessary to use dirent hook to update the parent scan results.  Callers
+ * must not read the scan results without re-taking @sc->ip's ILOCK.
+ *
+ * There are a few shortcuts that we can take to avoid scanning the entire
+ * filesystem, such as noticing directory tree roots and querying the dentry
+ * cache for parent information.
+ */
+
+struct xrep_findparent_info {
+	/* The directory currently being scanned. */
+	struct xfs_inode	*dp;
+
+	/*
+	 * Scrub context.  We're looking for a @dp containing a directory
+	 * entry pointing to sc->ip->i_ino.
+	 */
+	struct xfs_scrub	*sc;
+
+	/* Optional scan information for a xrep_findparent_scan call. */
+	struct xrep_parent_scan_info *parent_scan;
+
+	/*
+	 * Parent that we've found for sc->ip.  If we're scanning the entire
+	 * directory tree, we need this to ensure that we only find /one/
+	 * parent directory.
+	 */
+	xfs_ino_t		found_parent;
+
+	/*
+	 * This is set to true if @found_parent was not observed directly from
+	 * the directory scan but by noticing a change in dotdot entries after
+	 * cycling the sc->ip IOLOCK.
+	 */
+	bool			parent_tentative;
+};
+
+/*
+ * If this directory entry points to the scrub target inode, then the directory
+ * we're scanning is the parent of the scrub target inode.
+ */
+STATIC int
+xrep_findparent_dirent(
+	struct xfs_scrub		*sc,
+	struct xfs_inode		*dp,
+	xfs_dir2_dataptr_t		dapos,
+	const struct xfs_name		*name,
+	xfs_ino_t			ino,
+	void				*priv)
+{
+	struct xrep_findparent_info	*fpi = priv;
+	int				error = 0;
+
+	if (xchk_should_terminate(fpi->sc, &error))
+		return error;
+
+	if (ino != fpi->sc->ip->i_ino)
+		return 0;
+
+	/* Ignore garbage directory entry names. */
+	if (name->len == 0 || !xfs_dir2_namecheck(name->name, name->len))
+		return -EFSCORRUPTED;
+
+	/*
+	 * Ignore dotdot and dot entries -- we're looking for parent -> child
+	 * links only.
+	 */
+	if (name->name[0] == '.' && (name->len == 1 ||
+				     (name->len == 2 && name->name[1] == '.')))
+		return 0;
+
+	/* Uhoh, more than one parent for a dir? */
+	if (fpi->found_parent != NULLFSINO &&
+	    !(fpi->parent_tentative && fpi->found_parent == fpi->dp->i_ino)) {
+		trace_xrep_findparent_dirent(fpi->sc->ip, 0);
+		return -EFSCORRUPTED;
+	}
+
+	/* We found a potential parent; remember this. */
+	trace_xrep_findparent_dirent(fpi->sc->ip, fpi->dp->i_ino);
+	fpi->found_parent = fpi->dp->i_ino;
+	fpi->parent_tentative = false;
+
+	if (fpi->parent_scan)
+		xrep_findparent_scan_found(fpi->parent_scan, fpi->dp->i_ino);
+
+	return 0;
+}
+
+/*
+ * If this is a directory, walk the dirents looking for any that point to the
+ * scrub target inode.
+ */
+STATIC int
+xrep_findparent_walk_directory(
+	struct xrep_findparent_info	*fpi)
+{
+	struct xfs_scrub		*sc = fpi->sc;
+	struct xfs_inode		*dp = fpi->dp;
+	unsigned int			lock_mode;
+	int				error = 0;
+
+	/*
+	 * The inode being scanned cannot be its own parent, nor can any
+	 * temporary directory we created to stage this repair.
+	 */
+	if (dp == sc->ip || dp == sc->tempip)
+		return 0;
+
+	/*
+	 * Similarly, temporary files created to stage a repair cannot be the
+	 * parent of this inode.
+	 */
+	if (xrep_is_tempfile(dp))
+		return 0;
+
+	/*
+	 * Scan the directory to see if there it contains an entry pointing to
+	 * the directory that we are repairing.
+	 */
+	lock_mode = xfs_ilock_data_map_shared(dp);
+
+	/*
+	 * If this directory is known to be sick, we cannot scan it reliably
+	 * and must abort.
+	 */
+	if (xfs_inode_has_sickness(dp, XFS_SICK_INO_CORE |
+				       XFS_SICK_INO_BMBTD |
+				       XFS_SICK_INO_DIR)) {
+		error = -EFSCORRUPTED;
+		goto out_unlock;
+	}
+
+	/*
+	 * We cannot complete our parent pointer scan if a directory looks as
+	 * though it has been zapped by the inode record repair code.
+	 */
+	if (xchk_dir_looks_zapped(dp)) {
+		error = -EBUSY;
+		goto out_unlock;
+	}
+
+	error = xchk_dir_walk(sc, dp, xrep_findparent_dirent, fpi);
+	if (error)
+		goto out_unlock;
+
+out_unlock:
+	xfs_iunlock(dp, lock_mode);
+	return error;
+}
+
+/*
+ * Update this directory's dotdot pointer based on ongoing dirent updates.
+ */
+STATIC int
+xrep_findparent_live_update(
+	struct notifier_block		*nb,
+	unsigned long			action,
+	void				*data)
+{
+	struct xfs_dir_update_params	*p = data;
+	struct xrep_parent_scan_info	*pscan;
+	struct xfs_scrub		*sc;
+
+	pscan = container_of(nb, struct xrep_parent_scan_info,
+			dhook.dirent_hook.nb);
+	sc = pscan->sc;
+
+	/*
+	 * If @p->ip is the subdirectory that we're interested in and we've
+	 * already scanned @p->dp, update the dotdot target inumber to the
+	 * parent inode.
+	 */
+	if (p->ip->i_ino == sc->ip->i_ino &&
+	    xchk_iscan_want_live_update(&pscan->iscan, p->dp->i_ino)) {
+		if (p->delta > 0) {
+			xrep_findparent_scan_found(pscan, p->dp->i_ino);
+		} else {
+			xrep_findparent_scan_found(pscan, NULLFSINO);
+		}
+	}
+
+	return NOTIFY_DONE;
+}
+
+/*
+ * Set up a scan to find the parent of a directory.  The provided dirent hook
+ * will be called when there is a dotdot update for the inode being repaired.
+ */
+int
+__xrep_findparent_scan_start(
+	struct xfs_scrub		*sc,
+	struct xrep_parent_scan_info	*pscan,
+	notifier_fn_t			custom_fn)
+{
+	int				error;
+
+	if (!(sc->flags & XCHK_FSGATES_DIRENTS)) {
+		ASSERT(sc->flags & XCHK_FSGATES_DIRENTS);
+		return -EINVAL;
+	}
+
+	pscan->sc = sc;
+	pscan->parent_ino = NULLFSINO;
+
+	mutex_init(&pscan->lock);
+
+	xchk_iscan_start(sc, 30000, 100, &pscan->iscan);
+
+	/*
+	 * Hook into the dirent update code.  The hook only operates on inodes
+	 * that were already scanned, and the scanner thread takes each inode's
+	 * ILOCK, which means that any in-progress inode updates will finish
+	 * before we can scan the inode.
+	 */
+	if (custom_fn)
+		xfs_dir_hook_setup(&pscan->dhook, custom_fn);
+	else
+		xfs_dir_hook_setup(&pscan->dhook, xrep_findparent_live_update);
+	error = xfs_dir_hook_add(sc->mp, &pscan->dhook);
+	if (error)
+		goto out_iscan;
+
+	return 0;
+out_iscan:
+	xchk_iscan_teardown(&pscan->iscan);
+	mutex_destroy(&pscan->lock);
+	return error;
+}
+
+/*
+ * Scan the entire filesystem looking for a parent inode for the inode being
+ * scrubbed.  @sc->ip must not be the root of a directory tree.  Callers must
+ * not hold a dirty transaction or any lock that would interfere with taking
+ * an ILOCK.
+ *
+ * Returns 0 with @pscan->parent_ino set to the parent that we found.
+ * Returns 0 with @pscan->parent_ino set to NULLFSINO if we found no parents.
+ * Returns the usual negative errno if something else happened.
+ */
+int
+xrep_findparent_scan(
+	struct xrep_parent_scan_info	*pscan)
+{
+	struct xrep_findparent_info	fpi = {
+		.sc			= pscan->sc,
+		.found_parent		= NULLFSINO,
+		.parent_scan		= pscan,
+	};
+	struct xfs_scrub		*sc = pscan->sc;
+	int				ret;
+
+	ASSERT(S_ISDIR(VFS_IC(sc->ip)->i_mode));
+
+	while ((ret = xchk_iscan_iter(&pscan->iscan, &fpi.dp)) == 1) {
+		if (S_ISDIR(VFS_I(fpi.dp)->i_mode))
+			ret = xrep_findparent_walk_directory(&fpi);
+		else
+			ret = 0;
+		xchk_iscan_mark_visited(&pscan->iscan, fpi.dp);
+		xchk_irele(sc, fpi.dp);
+		if (ret)
+			break;
+
+		if (xchk_should_terminate(sc, &ret))
+			break;
+	}
+	xchk_iscan_iter_finish(&pscan->iscan);
+
+	return ret;
+}
+
+/* Tear down a parent scan. */
+void
+xrep_findparent_scan_teardown(
+	struct xrep_parent_scan_info	*pscan)
+{
+	xfs_dir_hook_del(pscan->sc->mp, &pscan->dhook);
+	xchk_iscan_teardown(&pscan->iscan);
+	mutex_destroy(&pscan->lock);
+}
+
+/* Finish a parent scan early. */
+void
+xrep_findparent_scan_finish_early(
+	struct xrep_parent_scan_info	*pscan,
+	xfs_ino_t			ino)
+{
+	xrep_findparent_scan_found(pscan, ino);
+	xchk_iscan_finish_early(&pscan->iscan);
+}
+
+/*
+ * Confirm that the directory @parent_ino actually contains a directory entry
+ * pointing to the child @sc->ip->ino.  This function returns one of several
+ * ways:
+ *
+ * Returns 0 with @parent_ino unchanged if the parent was confirmed.
+ * Returns 0 with @parent_ino set to NULLFSINO if the parent was not valid.
+ * Returns the usual negative errno if something else happened.
+ */
+int
+xrep_findparent_confirm(
+	struct xfs_scrub	*sc,
+	xfs_ino_t		*parent_ino)
+{
+	struct xrep_findparent_info fpi = {
+		.sc		= sc,
+		.found_parent	= NULLFSINO,
+	};
+	int			error;
+
+	/*
+	 * The root directory always points to itself.  Unlinked dirs can point
+	 * anywhere, so we point them at the root dir too.
+	 */
+	if (sc->ip == sc->mp->m_rootip || VFS_I(sc->ip)->i_nlink == 0) {
+		*parent_ino = sc->mp->m_sb.sb_rootino;
+		return 0;
+	}
+
+	/* Reject garbage parent inode numbers and self-referential parents. */
+	if (*parent_ino == NULLFSINO)
+	       return 0;
+	if (!xfs_verify_dir_ino(sc->mp, *parent_ino) ||
+	    *parent_ino == sc->ip->i_ino) {
+		*parent_ino = NULLFSINO;
+		return 0;
+	}
+
+	error = xchk_iget(sc, *parent_ino, &fpi.dp);
+	if (error)
+		return error;
+
+	if (!S_ISDIR(VFS_I(fpi.dp)->i_mode)) {
+		*parent_ino = NULLFSINO;
+		goto out_rele;
+	}
+
+	error = xrep_findparent_walk_directory(&fpi);
+	if (error)
+		goto out_rele;
+
+	*parent_ino = fpi.found_parent;
+out_rele:
+	xchk_irele(sc, fpi.dp);
+	return error;
+}
+
+/*
+ * If we're the root of a directory tree, we are our own parent.  If we're an
+ * unlinked directory, the parent /won't/ have a link to us.  Set the parent
+ * directory to the root for both cases.  Returns NULLFSINO if we don't know
+ * what to do.
+ */
+xfs_ino_t
+xrep_findparent_self_reference(
+	struct xfs_scrub	*sc)
+{
+	if (sc->ip->i_ino == sc->mp->m_sb.sb_rootino)
+		return sc->mp->m_sb.sb_rootino;
+
+	if (VFS_I(sc->ip)->i_nlink == 0)
+		return sc->mp->m_sb.sb_rootino;
+
+	return NULLFSINO;
+}
+
+/* Check the dentry cache to see if knows of a parent for the scrub target. */
+xfs_ino_t
+xrep_findparent_from_dcache(
+	struct xfs_scrub	*sc)
+{
+	struct inode		*pip = NULL;
+	struct dentry		*dentry, *parent;
+	xfs_ino_t		ret = NULLFSINO;
+
+	dentry = d_find_alias(VFS_I(sc->ip));
+	if (!dentry)
+		goto out;
+
+	parent = dget_parent(dentry);
+	if (!parent)
+		goto out_dput;
+
+	ASSERT(parent->d_sb == sc->ip->i_mount->m_super);
+
+	pip = igrab(d_inode(parent));
+	dput(parent);
+
+	if (S_ISDIR(pip->i_mode)) {
+		trace_xrep_findparent_from_dcache(sc->ip, XFS_I(pip)->i_ino);
+		ret = XFS_I(pip)->i_ino;
+	}
+
+	xchk_irele(sc, XFS_I(pip));
+
+out_dput:
+	dput(dentry);
+out:
+	return ret;
+}
diff --git a/fs/xfs/scrub/findparent.h b/fs/xfs/scrub/findparent.h
new file mode 100644
index 0000000000..d998c7a881
--- /dev/null
+++ b/fs/xfs/scrub/findparent.h
@@ -0,0 +1,56 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (c) 2020-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_FINDPARENT_H__
+#define __XFS_SCRUB_FINDPARENT_H__
+
+struct xrep_parent_scan_info {
+	struct xfs_scrub	*sc;
+
+	/* Inode scan cursor. */
+	struct xchk_iscan	iscan;
+
+	/* Hook to capture directory entry updates. */
+	struct xfs_dir_hook	dhook;
+
+	/* Lock protecting parent_ino. */
+	struct mutex		lock;
+
+	/* Parent inode that we've found. */
+	xfs_ino_t		parent_ino;
+
+	bool			lookup_parent;
+};
+
+int __xrep_findparent_scan_start(struct xfs_scrub *sc,
+		struct xrep_parent_scan_info *pscan,
+		notifier_fn_t custom_fn);
+static inline int xrep_findparent_scan_start(struct xfs_scrub *sc,
+		struct xrep_parent_scan_info *pscan)
+{
+	return __xrep_findparent_scan_start(sc, pscan, NULL);
+}
+int xrep_findparent_scan(struct xrep_parent_scan_info *pscan);
+void xrep_findparent_scan_teardown(struct xrep_parent_scan_info *pscan);
+
+static inline void
+xrep_findparent_scan_found(
+	struct xrep_parent_scan_info	*pscan,
+	xfs_ino_t			ino)
+{
+	mutex_lock(&pscan->lock);
+	pscan->parent_ino = ino;
+	mutex_unlock(&pscan->lock);
+}
+
+void xrep_findparent_scan_finish_early(struct xrep_parent_scan_info *pscan,
+		xfs_ino_t ino);
+
+int xrep_findparent_confirm(struct xfs_scrub *sc, xfs_ino_t *parent_ino);
+
+xfs_ino_t xrep_findparent_self_reference(struct xfs_scrub *sc);
+xfs_ino_t xrep_findparent_from_dcache(struct xfs_scrub *sc);
+
+#endif /* __XFS_SCRUB_FINDPARENT_H__ */
diff --git a/fs/xfs/scrub/fscounters.c b/fs/xfs/scrub/fscounters.c
index 5799e9a94f..1d3e983469 100644
--- a/fs/xfs/scrub/fscounters.c
+++ b/fs/xfs/scrub/fscounters.c
@@ -22,6 +22,7 @@
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
+#include "scrub/fscounters.h"
 
 /*
  * FS Summary Counters
@@ -48,17 +49,6 @@
  * our tolerance for mismatch between expected and actual counter values.
  */
 
-struct xchk_fscounters {
-	struct xfs_scrub	*sc;
-	uint64_t		icount;
-	uint64_t		ifree;
-	uint64_t		fdblocks;
-	uint64_t		frextents;
-	unsigned long long	icount_min;
-	unsigned long long	icount_max;
-	bool			frozen;
-};
-
 /*
  * Since the expected value computation is lockless but only browses incore
  * values, the percpu counters should be fairly close to each other.  However,
@@ -95,7 +85,7 @@ xchk_fscount_warmup(
 			continue;
 
 		/* Lock both AG headers. */
-		error = xfs_ialloc_read_agi(pag, sc->tp, &agi_bp);
+		error = xfs_ialloc_read_agi(pag, sc->tp, 0, &agi_bp);
 		if (error)
 			break;
 		error = xfs_alloc_read_agf(pag, sc->tp, 0, &agf_bp);
@@ -235,14 +225,19 @@ xchk_setup_fscounters(
 	 * Pause all writer activity in the filesystem while we're scrubbing to
 	 * reduce the likelihood of background perturbations to the counters
 	 * throwing off our calculations.
+	 *
+	 * If we're repairing, we need to prevent any other thread from
+	 * changing the global fs summary counters while we're repairing them.
+	 * This requires the fs to be frozen, which will disable background
+	 * reclaim and purge all inactive inodes.
 	 */
-	if (sc->flags & XCHK_TRY_HARDER) {
+	if ((sc->flags & XCHK_TRY_HARDER) || xchk_could_repair(sc)) {
 		error = xchk_fscounters_freeze(sc);
 		if (error)
 			return error;
 	}
 
-	return xfs_trans_alloc_empty(sc->mp, &sc->tp);
+	return xchk_trans_alloc_empty(sc);
 }
 
 /*
@@ -254,7 +249,9 @@ xchk_setup_fscounters(
  * set the INCOMPLETE flag even when a negative errno is returned.  This care
  * must be taken with certain errno values (i.e. EFSBADCRC, EFSCORRUPTED,
  * ECANCELED) that are absorbed into a scrub state flag update by
- * xchk_*_process_error.
+ * xchk_*_process_error.  Scrub and repair share the same incore data
+ * structures, so the INCOMPLETE flag is critical to prevent a repair based on
+ * insufficient information.
  */
 
 /* Count free space btree blocks manually for pre-lazysbcount filesystems. */
@@ -415,10 +412,11 @@ xchk_fscount_count_frextents(
 	int			error;
 
 	fsc->frextents = 0;
+	fsc->frextents_delayed = 0;
 	if (!xfs_has_realtime(mp))
 		return 0;
 
-	xfs_ilock(sc->mp->m_rbmip, XFS_ILOCK_SHARED | XFS_ILOCK_RTBITMAP);
+	xfs_rtbitmap_lock_shared(sc->mp, XFS_RBMLOCK_BITMAP);
 	error = xfs_rtalloc_query_all(sc->mp, sc->tp,
 			xchk_fscount_add_frextent, fsc);
 	if (error) {
@@ -426,8 +424,10 @@ xchk_fscount_count_frextents(
 		goto out_unlock;
 	}
 
+	fsc->frextents_delayed = percpu_counter_sum(&mp->m_delalloc_rtextents);
+
 out_unlock:
-	xfs_iunlock(sc->mp->m_rbmip, XFS_ILOCK_SHARED | XFS_ILOCK_RTBITMAP);
+	xfs_rtbitmap_unlock_shared(sc->mp, XFS_RBMLOCK_BITMAP);
 	return error;
 }
 #else
@@ -437,6 +437,7 @@ xchk_fscount_count_frextents(
 	struct xchk_fscounters	*fsc)
 {
 	fsc->frextents = 0;
+	fsc->frextents_delayed = 0;
 	return 0;
 }
 #endif /* CONFIG_XFS_RT */
@@ -482,6 +483,10 @@ xchk_fscount_within_range(
 	if (curr_value == expected)
 		return true;
 
+	/* We require exact matches when repair is running. */
+	if (sc->sm->sm_flags & XFS_SCRUB_IFLAG_REPAIR)
+		return false;
+
 	min_value = min(old_value, curr_value);
 	max_value = max(old_value, curr_value);
 
@@ -516,7 +521,7 @@ xchk_fscounters(
 
 	/*
 	 * If the filesystem is not frozen, the counter summation calls above
-	 * can race with xfs_mod_freecounter, which subtracts a requested space
+	 * can race with xfs_dec_freecounter, which subtracts a requested space
 	 * reservation from the counter and undoes the subtraction if that made
 	 * the counter go negative.  Therefore, it's possible to see negative
 	 * values here, and we should only flag that as a corruption if we
@@ -592,7 +597,7 @@ xchk_fscounters(
 	}
 
 	if (!xchk_fscount_within_range(sc, frextents, &mp->m_frextents,
-			fsc->frextents)) {
+			fsc->frextents - fsc->frextents_delayed)) {
 		if (fsc->frozen)
 			xchk_set_corrupt(sc);
 		else
diff --git a/fs/xfs/scrub/fscounters.h b/fs/xfs/scrub/fscounters.h
new file mode 100644
index 0000000000..bcf56e1c36
--- /dev/null
+++ b/fs/xfs/scrub/fscounters.h
@@ -0,0 +1,21 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (c) 2021-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_FSCOUNTERS_H__
+#define __XFS_SCRUB_FSCOUNTERS_H__
+
+struct xchk_fscounters {
+	struct xfs_scrub	*sc;
+	uint64_t		icount;
+	uint64_t		ifree;
+	uint64_t		fdblocks;
+	uint64_t		frextents;
+	uint64_t		frextents_delayed;
+	unsigned long long	icount_min;
+	unsigned long long	icount_max;
+	bool			frozen;
+};
+
+#endif /* __XFS_SCRUB_FSCOUNTERS_H__ */
diff --git a/fs/xfs/scrub/fscounters_repair.c b/fs/xfs/scrub/fscounters_repair.c
new file mode 100644
index 0000000000..469bf645db
--- /dev/null
+++ b/fs/xfs/scrub/fscounters_repair.c
@@ -0,0 +1,82 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2018-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_alloc.h"
+#include "xfs_ialloc.h"
+#include "xfs_rmap.h"
+#include "xfs_health.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+#include "scrub/repair.h"
+#include "scrub/fscounters.h"
+
+/*
+ * FS Summary Counters
+ * ===================
+ *
+ * We correct errors in the filesystem summary counters by setting them to the
+ * values computed during the obligatory scrub phase.  However, we must be
+ * careful not to allow any other thread to change the counters while we're
+ * computing and setting new values.  To achieve this, we freeze the
+ * filesystem for the whole operation if the REPAIR flag is set.  The checking
+ * function is stricter when we've frozen the fs.
+ */
+
+/*
+ * Reset the superblock counters.  Caller is responsible for freezing the
+ * filesystem during the calculation and reset phases.
+ */
+int
+xrep_fscounters(
+	struct xfs_scrub	*sc)
+{
+	struct xfs_mount	*mp = sc->mp;
+	struct xchk_fscounters	*fsc = sc->buf;
+
+	/*
+	 * Reinitialize the in-core counters from what we computed.  We froze
+	 * the filesystem, so there shouldn't be anyone else trying to modify
+	 * these counters.
+	 */
+	if (!fsc->frozen) {
+		ASSERT(fsc->frozen);
+		return -EFSCORRUPTED;
+	}
+
+	trace_xrep_reset_counters(mp, fsc);
+
+	percpu_counter_set(&mp->m_icount, fsc->icount);
+	percpu_counter_set(&mp->m_ifree, fsc->ifree);
+	percpu_counter_set(&mp->m_fdblocks, fsc->fdblocks);
+
+	/*
+	 * Online repair is only supported on v5 file systems, which require
+	 * lazy sb counters and thus no update of sb_fdblocks here.  But as of
+	 * now we don't support lazy counting sb_frextents yet, and thus need
+	 * to also update it directly here.  And for that we need to keep
+	 * track of the delalloc reservations separately, as they are are
+	 * subtracted from m_frextents, but not included in sb_frextents.
+	 */
+	percpu_counter_set(&mp->m_frextents,
+		fsc->frextents - fsc->frextents_delayed);
+	mp->m_sb.sb_frextents = fsc->frextents;
+
+	return 0;
+}
diff --git a/fs/xfs/scrub/health.c b/fs/xfs/scrub/health.c
index 531006910c..b712a8bd34 100644
--- a/fs/xfs/scrub/health.c
+++ b/fs/xfs/scrub/health.c
@@ -14,6 +14,7 @@
 #include "xfs_health.h"
 #include "scrub/scrub.h"
 #include "scrub/health.h"
+#include "scrub/common.h"
 
 /*
  * Scrub and In-Core Filesystem Health Assessments
@@ -105,6 +106,9 @@ static const struct xchk_health_map type_to_health_flag[XFS_SCRUB_TYPE_NR] = {
 	[XFS_SCRUB_TYPE_GQUOTA]		= { XHG_FS,  XFS_SICK_FS_GQUOTA },
 	[XFS_SCRUB_TYPE_PQUOTA]		= { XHG_FS,  XFS_SICK_FS_PQUOTA },
 	[XFS_SCRUB_TYPE_FSCOUNTERS]	= { XHG_FS,  XFS_SICK_FS_COUNTERS },
+	[XFS_SCRUB_TYPE_QUOTACHECK]	= { XHG_FS,  XFS_SICK_FS_QUOTACHECK },
+	[XFS_SCRUB_TYPE_NLINKS]		= { XHG_FS,  XFS_SICK_FS_NLINKS },
+	[XFS_SCRUB_TYPE_DIRTREE]	= { XHG_INO, XFS_SICK_INO_DIRTREE },
 };
 
 /* Return the health status mask for this scrub type. */
@@ -148,6 +152,24 @@ xchk_file_looks_zapped(
 }
 
 /*
+ * Scrub gave the filesystem a clean bill of health, so clear all the indirect
+ * markers of past problems (at least for the fs and ags) so that we can be
+ * healthy again.
+ */
+STATIC void
+xchk_mark_all_healthy(
+	struct xfs_mount	*mp)
+{
+	struct xfs_perag	*pag;
+	xfs_agnumber_t		agno;
+
+	xfs_fs_mark_healthy(mp, XFS_SICK_FS_INDIRECT);
+	xfs_rt_mark_healthy(mp, XFS_SICK_RT_INDIRECT);
+	for_each_perag(mp, agno, pag)
+		xfs_ag_mark_healthy(pag, XFS_SICK_AG_INDIRECT);
+}
+
+/*
  * Update filesystem health assessments based on what we found and did.
  *
  * If the scrubber finds errors, we mark sick whatever's mentioned in
@@ -164,6 +186,18 @@ xchk_update_health(
 	struct xfs_perag	*pag;
 	bool			bad;
 
+	/*
+	 * The HEALTHY scrub type is a request from userspace to clear all the
+	 * indirect flags after a clean scan of the entire filesystem.  As such
+	 * there's no sick flag defined for it, so we branch here ahead of the
+	 * mask check.
+	 */
+	if (sc->sm->sm_type == XFS_SCRUB_TYPE_HEALTHY &&
+	    !(sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)) {
+		xchk_mark_all_healthy(sc->mp);
+		return;
+	}
+
 	if (!sc->sick_mask)
 		return;
 
@@ -173,7 +207,7 @@ xchk_update_health(
 	case XHG_AG:
 		pag = xfs_perag_get(sc->mp, sc->sm->sm_agno);
 		if (bad)
-			xfs_ag_mark_sick(pag, sc->sick_mask);
+			xfs_ag_mark_corrupt(pag, sc->sick_mask);
 		else
 			xfs_ag_mark_healthy(pag, sc->sick_mask);
 		xfs_perag_put(pag);
@@ -181,20 +215,30 @@ xchk_update_health(
 	case XHG_INO:
 		if (!sc->ip)
 			return;
-		if (bad)
-			xfs_inode_mark_sick(sc->ip, sc->sick_mask);
-		else
+		if (bad) {
+			unsigned int	mask = sc->sick_mask;
+
+			/*
+			 * If we're coming in for repairs then we don't want
+			 * sickness flags to propagate to the incore health
+			 * status if the inode gets inactivated before we can
+			 * fix it.
+			 */
+			if (sc->sm->sm_flags & XFS_SCRUB_IFLAG_REPAIR)
+				mask |= XFS_SICK_INO_FORGET;
+			xfs_inode_mark_corrupt(sc->ip, mask);
+		} else
 			xfs_inode_mark_healthy(sc->ip, sc->sick_mask);
 		break;
 	case XHG_FS:
 		if (bad)
-			xfs_fs_mark_sick(sc->mp, sc->sick_mask);
+			xfs_fs_mark_corrupt(sc->mp, sc->sick_mask);
 		else
 			xfs_fs_mark_healthy(sc->mp, sc->sick_mask);
 		break;
 	case XHG_RT:
 		if (bad)
-			xfs_rt_mark_sick(sc->mp, sc->sick_mask);
+			xfs_rt_mark_corrupt(sc->mp, sc->sick_mask);
 		else
 			xfs_rt_mark_healthy(sc->mp, sc->sick_mask);
 		break;
@@ -205,13 +249,13 @@ xchk_update_health(
 }
 
 /* Is the given per-AG btree healthy enough for scanning? */
-bool
-xchk_ag_btree_healthy_enough(
+void
+xchk_ag_btree_del_cursor_if_sick(
 	struct xfs_scrub	*sc,
-	struct xfs_perag	*pag,
-	xfs_btnum_t		btnum)
+	struct xfs_btree_cur	**curp,
+	unsigned int		sm_type)
 {
-	unsigned int		mask = 0;
+	unsigned int		mask = (*curp)->bc_ops->sick_mask;
 
 	/*
 	 * We always want the cursor if it's the same type as whatever we're
@@ -220,41 +264,8 @@ xchk_ag_btree_healthy_enough(
 	 * Otherwise, we're only interested in the btree for cross-referencing.
 	 * If we know the btree is bad then don't bother, just set XFAIL.
 	 */
-	switch (btnum) {
-	case XFS_BTNUM_BNO:
-		if (sc->sm->sm_type == XFS_SCRUB_TYPE_BNOBT)
-			return true;
-		mask = XFS_SICK_AG_BNOBT;
-		break;
-	case XFS_BTNUM_CNT:
-		if (sc->sm->sm_type == XFS_SCRUB_TYPE_CNTBT)
-			return true;
-		mask = XFS_SICK_AG_CNTBT;
-		break;
-	case XFS_BTNUM_INO:
-		if (sc->sm->sm_type == XFS_SCRUB_TYPE_INOBT)
-			return true;
-		mask = XFS_SICK_AG_INOBT;
-		break;
-	case XFS_BTNUM_FINO:
-		if (sc->sm->sm_type == XFS_SCRUB_TYPE_FINOBT)
-			return true;
-		mask = XFS_SICK_AG_FINOBT;
-		break;
-	case XFS_BTNUM_RMAP:
-		if (sc->sm->sm_type == XFS_SCRUB_TYPE_RMAPBT)
-			return true;
-		mask = XFS_SICK_AG_RMAPBT;
-		break;
-	case XFS_BTNUM_REFC:
-		if (sc->sm->sm_type == XFS_SCRUB_TYPE_REFCNTBT)
-			return true;
-		mask = XFS_SICK_AG_REFCNTBT;
-		break;
-	default:
-		ASSERT(0);
-		return true;
-	}
+	if (sc->sm->sm_type == sm_type)
+		return;
 
 	/*
 	 * If we just repaired some AG metadata, sc->sick_mask will reflect all
@@ -266,10 +277,42 @@ xchk_ag_btree_healthy_enough(
 	    type_to_health_flag[sc->sm->sm_type].group == XHG_AG)
 		mask &= ~sc->sick_mask;
 
-	if (xfs_ag_has_sickness(pag, mask)) {
+	if (xfs_ag_has_sickness((*curp)->bc_ag.pag, mask)) {
 		sc->sm->sm_flags |= XFS_SCRUB_OFLAG_XFAIL;
-		return false;
+		xfs_btree_del_cursor(*curp, XFS_BTREE_NOERROR);
+		*curp = NULL;
+	}
+}
+
+/*
+ * Quick scan to double-check that there isn't any evidence of lingering
+ * primary health problems.  If we're still clear, then the health update will
+ * take care of clearing the indirect evidence.
+ */
+int
+xchk_health_record(
+	struct xfs_scrub	*sc)
+{
+	struct xfs_mount	*mp = sc->mp;
+	struct xfs_perag	*pag;
+	xfs_agnumber_t		agno;
+
+	unsigned int		sick;
+	unsigned int		checked;
+
+	xfs_fs_measure_sickness(mp, &sick, &checked);
+	if (sick & XFS_SICK_FS_PRIMARY)
+		xchk_set_corrupt(sc);
+
+	xfs_rt_measure_sickness(mp, &sick, &checked);
+	if (sick & XFS_SICK_RT_PRIMARY)
+		xchk_set_corrupt(sc);
+
+	for_each_perag(mp, agno, pag) {
+		xfs_ag_measure_sickness(pag, &sick, &checked);
+		if (sick & XFS_SICK_AG_PRIMARY)
+			xchk_set_corrupt(sc);
 	}
 
-	return true;
+	return 0;
 }
diff --git a/fs/xfs/scrub/health.h b/fs/xfs/scrub/health.h
index a731b24673..63fc426eb5 100644
--- a/fs/xfs/scrub/health.h
+++ b/fs/xfs/scrub/health.h
@@ -8,9 +8,10 @@
 
 unsigned int xchk_health_mask_for_scrub_type(__u32 scrub_type);
 void xchk_update_health(struct xfs_scrub *sc);
-bool xchk_ag_btree_healthy_enough(struct xfs_scrub *sc, struct xfs_perag *pag,
-		xfs_btnum_t btnum);
+void xchk_ag_btree_del_cursor_if_sick(struct xfs_scrub *sc,
+		struct xfs_btree_cur **curp, unsigned int sm_type);
 void xchk_mark_healthy_if_clean(struct xfs_scrub *sc, unsigned int mask);
 bool xchk_file_looks_zapped(struct xfs_scrub *sc, unsigned int mask);
+int xchk_health_record(struct xfs_scrub *sc);
 
 #endif /* __XFS_SCRUB_HEALTH_H__ */
diff --git a/fs/xfs/scrub/ialloc.c b/fs/xfs/scrub/ialloc.c
index a720fc6226..750d7b0cd2 100644
--- a/fs/xfs/scrub/ialloc.c
+++ b/fs/xfs/scrub/ialloc.c
@@ -76,7 +76,7 @@ xchk_inobt_xref_finobt(
 	int			has_record;
 	int			error;
 
-	ASSERT(cur->bc_btnum == XFS_BTNUM_FINO);
+	ASSERT(xfs_btree_is_fino(cur->bc_ops));
 
 	error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, &has_record);
 	if (error)
@@ -179,7 +179,7 @@ xchk_finobt_xref_inobt(
 	int			has_record;
 	int			error;
 
-	ASSERT(cur->bc_btnum == XFS_BTNUM_INO);
+	ASSERT(xfs_btree_is_ino(cur->bc_ops));
 
 	error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, &has_record);
 	if (error)
@@ -514,7 +514,7 @@ xchk_iallocbt_rec_alignment(
 	 * Otherwise, we expect that the finobt record is aligned to the
 	 * cluster alignment as told by the superblock.
 	 */
-	if (bs->cur->bc_btnum == XFS_BTNUM_FINO) {
+	if (xfs_btree_is_fino(bs->cur->bc_ops)) {
 		unsigned int	imask;
 
 		imask = min_t(unsigned int, XFS_INODES_PER_CHUNK,
@@ -649,8 +649,7 @@ out:
  */
 STATIC void
 xchk_iallocbt_xref_rmap_btreeblks(
-	struct xfs_scrub	*sc,
-	int			which)
+	struct xfs_scrub	*sc)
 {
 	xfs_filblks_t		blocks;
 	xfs_extlen_t		inobt_blocks = 0;
@@ -688,7 +687,6 @@ xchk_iallocbt_xref_rmap_btreeblks(
 STATIC void
 xchk_iallocbt_xref_rmap_inodes(
 	struct xfs_scrub	*sc,
-	int			which,
 	unsigned long long	inodes)
 {
 	xfs_filblks_t		blocks;
@@ -719,17 +717,14 @@ xchk_iallocbt(
 		.next_startino	= NULLAGINO,
 		.next_cluster_ino = NULLAGINO,
 	};
-	xfs_btnum_t		which;
 	int			error;
 
 	switch (sc->sm->sm_type) {
 	case XFS_SCRUB_TYPE_INOBT:
 		cur = sc->sa.ino_cur;
-		which = XFS_BTNUM_INO;
 		break;
 	case XFS_SCRUB_TYPE_FINOBT:
 		cur = sc->sa.fino_cur;
-		which = XFS_BTNUM_FINO;
 		break;
 	default:
 		ASSERT(0);
@@ -741,7 +736,7 @@ xchk_iallocbt(
 	if (error)
 		return error;
 
-	xchk_iallocbt_xref_rmap_btreeblks(sc, which);
+	xchk_iallocbt_xref_rmap_btreeblks(sc);
 
 	/*
 	 * If we're scrubbing the inode btree, inode_blocks is the number of
@@ -750,9 +745,8 @@ xchk_iallocbt(
 	 * knows about.  We can't do this for the finobt since it only points
 	 * to inode chunks with free inodes.
 	 */
-	if (which == XFS_BTNUM_INO)
-		xchk_iallocbt_xref_rmap_inodes(sc, which, iabt.inodes);
-
+	if (sc->sm->sm_type == XFS_SCRUB_TYPE_INOBT)
+		xchk_iallocbt_xref_rmap_inodes(sc, iabt.inodes);
 	return error;
 }
 
diff --git a/fs/xfs/scrub/ialloc_repair.c b/fs/xfs/scrub/ialloc_repair.c
index b3f7182dd2..a00ec7ae17 100644
--- a/fs/xfs/scrub/ialloc_repair.c
+++ b/fs/xfs/scrub/ialloc_repair.c
@@ -369,7 +369,7 @@ xrep_ibt_check_inode_ext(
 	 * On a sparse inode fs, this cluster could be part of a sparse chunk.
 	 * Sparse clusters must be aligned to sparse chunk alignment.
 	 */
-	if (xfs_has_sparseinodes(mp) &&
+	if (xfs_has_sparseinodes(mp) && mp->m_sb.sb_spino_align &&
 	    (!IS_ALIGNED(agbno, mp->m_sb.sb_spino_align) ||
 	     !IS_ALIGNED(agbno + len, mp->m_sb.sb_spino_align)))
 		return -EFSCORRUPTED;
@@ -663,8 +663,8 @@ xrep_ibt_build_new_trees(
 	ri->new_inobt.bload.claim_block = xrep_ibt_claim_block;
 	ri->new_inobt.bload.get_records = xrep_ibt_get_records;
 
-	ino_cur = xfs_inobt_stage_cursor(sc->sa.pag, &ri->new_inobt.afake,
-			XFS_BTNUM_INO);
+	ino_cur = xfs_inobt_init_cursor(sc->sa.pag, NULL, NULL);
+	xfs_btree_stage_afakeroot(ino_cur, &ri->new_inobt.afake);
 	error = xfs_btree_bload_compute_geometry(ino_cur, &ri->new_inobt.bload,
 			xfarray_length(ri->inode_records));
 	if (error)
@@ -684,8 +684,8 @@ xrep_ibt_build_new_trees(
 		ri->new_finobt.bload.claim_block = xrep_fibt_claim_block;
 		ri->new_finobt.bload.get_records = xrep_fibt_get_records;
 
-		fino_cur = xfs_inobt_stage_cursor(sc->sa.pag,
-				&ri->new_finobt.afake, XFS_BTNUM_FINO);
+		fino_cur = xfs_finobt_init_cursor(sc->sa.pag, NULL, NULL);
+		xfs_btree_stage_afakeroot(fino_cur, &ri->new_finobt.afake);
 		error = xfs_btree_bload_compute_geometry(fino_cur,
 				&ri->new_finobt.bload, ri->finobt_recs);
 		if (error)
diff --git a/fs/xfs/scrub/ino_bitmap.h b/fs/xfs/scrub/ino_bitmap.h
new file mode 100644
index 0000000000..1300833679
--- /dev/null
+++ b/fs/xfs/scrub/ino_bitmap.h
@@ -0,0 +1,37 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2023-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_INO_BITMAP_H__
+#define __XFS_SCRUB_INO_BITMAP_H__
+
+/* Bitmaps, but for type-checked for xfs_ino_t */
+
+struct xino_bitmap {
+	struct xbitmap64	inobitmap;
+};
+
+static inline void xino_bitmap_init(struct xino_bitmap *bitmap)
+{
+	xbitmap64_init(&bitmap->inobitmap);
+}
+
+static inline void xino_bitmap_destroy(struct xino_bitmap *bitmap)
+{
+	xbitmap64_destroy(&bitmap->inobitmap);
+}
+
+static inline int xino_bitmap_set(struct xino_bitmap *bitmap, xfs_ino_t ino)
+{
+	return xbitmap64_set(&bitmap->inobitmap, ino, 1);
+}
+
+static inline int xino_bitmap_test(struct xino_bitmap *bitmap, xfs_ino_t ino)
+{
+	uint64_t	len = 1;
+
+	return xbitmap64_test(&bitmap->inobitmap, ino, &len);
+}
+
+#endif	/* __XFS_SCRUB_INO_BITMAP_H__ */
diff --git a/fs/xfs/scrub/inode.c b/fs/xfs/scrub/inode.c
index 6e2fe2d625..d32716fb2f 100644
--- a/fs/xfs/scrub/inode.c
+++ b/fs/xfs/scrub/inode.c
@@ -739,6 +739,23 @@ xchk_inode_check_reflink_iflag(
 		xchk_ino_set_corrupt(sc, ino);
 }
 
+/*
+ * If this inode has zero link count, it must be on the unlinked list.  If
+ * it has nonzero link count, it must not be on the unlinked list.
+ */
+STATIC void
+xchk_inode_check_unlinked(
+	struct xfs_scrub	*sc)
+{
+	if (VFS_I(sc->ip)->i_nlink == 0) {
+		if (!xfs_inode_on_unlinked_list(sc->ip))
+			xchk_ino_set_corrupt(sc, sc->ip->i_ino);
+	} else {
+		if (xfs_inode_on_unlinked_list(sc->ip))
+			xchk_ino_set_corrupt(sc, sc->ip->i_ino);
+	}
+}
+
 /* Scrub an inode. */
 int
 xchk_inode(
@@ -771,6 +788,8 @@ xchk_inode(
 	if (S_ISREG(VFS_I(sc->ip)->i_mode))
 		xchk_inode_check_reflink_iflag(sc, sc->ip->i_ino);
 
+	xchk_inode_check_unlinked(sc);
+
 	xchk_inode_xref(sc, sc->ip->i_ino, &di);
 out:
 	return error;
diff --git a/fs/xfs/scrub/inode_repair.c b/fs/xfs/scrub/inode_repair.c
index 0ca62d59f8..daf9f1ee7c 100644
--- a/fs/xfs/scrub/inode_repair.c
+++ b/fs/xfs/scrub/inode_repair.c
@@ -37,12 +37,16 @@
 #include "xfs_attr_leaf.h"
 #include "xfs_log_priv.h"
 #include "xfs_health.h"
+#include "xfs_symlink_remote.h"
 #include "scrub/xfs_scrub.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/btree.h"
 #include "scrub/trace.h"
 #include "scrub/repair.h"
+#include "scrub/iscan.h"
+#include "scrub/readdir.h"
+#include "scrub/tempfile.h"
 
 /*
  * Inode Record Repair
@@ -126,6 +130,10 @@ struct xrep_inode {
 
 	/* Must we remove all access from this file? */
 	bool			zap_acls;
+
+	/* Inode scanner to see if we can find the ftype from dirents */
+	struct xchk_iscan	ftype_iscan;
+	uint8_t			alleged_ftype;
 };
 
 /*
@@ -227,26 +235,296 @@ xrep_dinode_header(
 	dip->di_gen = cpu_to_be32(sc->sm->sm_gen);
 }
 
-/* Turn di_mode into /something/ recognizable. */
-STATIC void
+/*
+ * If this directory entry points to the scrub target inode, then the directory
+ * we're scanning is the parent of the scrub target inode.
+ */
+STATIC int
+xrep_dinode_findmode_dirent(
+	struct xfs_scrub		*sc,
+	struct xfs_inode		*dp,
+	xfs_dir2_dataptr_t		dapos,
+	const struct xfs_name		*name,
+	xfs_ino_t			ino,
+	void				*priv)
+{
+	struct xrep_inode		*ri = priv;
+	int				error = 0;
+
+	if (xchk_should_terminate(ri->sc, &error))
+		return error;
+
+	if (ino != sc->sm->sm_ino)
+		return 0;
+
+	/* Ignore garbage directory entry names. */
+	if (name->len == 0 || !xfs_dir2_namecheck(name->name, name->len))
+		return -EFSCORRUPTED;
+
+	/* Don't pick up dot or dotdot entries; we only want child dirents. */
+	if (xfs_dir2_samename(name, &xfs_name_dotdot) ||
+	    xfs_dir2_samename(name, &xfs_name_dot))
+		return 0;
+
+	/*
+	 * Uhoh, more than one parent for this inode and they don't agree on
+	 * the file type?
+	 */
+	if (ri->alleged_ftype != XFS_DIR3_FT_UNKNOWN &&
+	    ri->alleged_ftype != name->type) {
+		trace_xrep_dinode_findmode_dirent_inval(ri->sc, dp, name->type,
+				ri->alleged_ftype);
+		return -EFSCORRUPTED;
+	}
+
+	/* We found a potential parent; remember the ftype. */
+	trace_xrep_dinode_findmode_dirent(ri->sc, dp, name->type);
+	ri->alleged_ftype = name->type;
+	return 0;
+}
+
+/* Try to lock a directory, or wait a jiffy. */
+static inline int
+xrep_dinode_ilock_nowait(
+	struct xfs_inode	*dp,
+	unsigned int		lock_mode)
+{
+	if (xfs_ilock_nowait(dp, lock_mode))
+		return true;
+
+	schedule_timeout_killable(1);
+	return false;
+}
+
+/*
+ * Try to lock a directory to look for ftype hints.  Since we already hold the
+ * AGI buffer, we cannot block waiting for the ILOCK because rename can take
+ * the ILOCK and then try to lock AGIs.
+ */
+STATIC int
+xrep_dinode_trylock_directory(
+	struct xrep_inode	*ri,
+	struct xfs_inode	*dp,
+	unsigned int		*lock_modep)
+{
+	unsigned long		deadline = jiffies + msecs_to_jiffies(30000);
+	unsigned int		lock_mode;
+	int			error = 0;
+
+	do {
+		if (xchk_should_terminate(ri->sc, &error))
+			return error;
+
+		if (xfs_need_iread_extents(&dp->i_df))
+			lock_mode = XFS_ILOCK_EXCL;
+		else
+			lock_mode = XFS_ILOCK_SHARED;
+
+		if (xrep_dinode_ilock_nowait(dp, lock_mode)) {
+			*lock_modep = lock_mode;
+			return 0;
+		}
+	} while (!time_is_before_jiffies(deadline));
+	return -EBUSY;
+}
+
+/*
+ * If this is a directory, walk the dirents looking for any that point to the
+ * scrub target inode.
+ */
+STATIC int
+xrep_dinode_findmode_walk_directory(
+	struct xrep_inode	*ri,
+	struct xfs_inode	*dp)
+{
+	struct xfs_scrub	*sc = ri->sc;
+	unsigned int		lock_mode;
+	int			error = 0;
+
+	/* Ignore temporary repair directories. */
+	if (xrep_is_tempfile(dp))
+		return 0;
+
+	/*
+	 * Scan the directory to see if there it contains an entry pointing to
+	 * the directory that we are repairing.
+	 */
+	error = xrep_dinode_trylock_directory(ri, dp, &lock_mode);
+	if (error)
+		return error;
+
+	/*
+	 * If this directory is known to be sick, we cannot scan it reliably
+	 * and must abort.
+	 */
+	if (xfs_inode_has_sickness(dp, XFS_SICK_INO_CORE |
+				       XFS_SICK_INO_BMBTD |
+				       XFS_SICK_INO_DIR)) {
+		error = -EFSCORRUPTED;
+		goto out_unlock;
+	}
+
+	/*
+	 * We cannot complete our parent pointer scan if a directory looks as
+	 * though it has been zapped by the inode record repair code.
+	 */
+	if (xchk_dir_looks_zapped(dp)) {
+		error = -EBUSY;
+		goto out_unlock;
+	}
+
+	error = xchk_dir_walk(sc, dp, xrep_dinode_findmode_dirent, ri);
+	if (error)
+		goto out_unlock;
+
+out_unlock:
+	xfs_iunlock(dp, lock_mode);
+	return error;
+}
+
+/*
+ * Try to find the mode of the inode being repaired by looking for directories
+ * that point down to this file.
+ */
+STATIC int
+xrep_dinode_find_mode(
+	struct xrep_inode	*ri,
+	uint16_t		*mode)
+{
+	struct xfs_scrub	*sc = ri->sc;
+	struct xfs_inode	*dp;
+	int			error;
+
+	/* No ftype means we have no other metadata to consult. */
+	if (!xfs_has_ftype(sc->mp)) {
+		*mode = S_IFREG;
+		return 0;
+	}
+
+	/*
+	 * Scan all directories for parents that might point down to this
+	 * inode.  Skip the inode being repaired during the scan since it
+	 * cannot be its own parent.  Note that we still hold the AGI locked
+	 * so there's a real possibility that _iscan_iter can return EBUSY.
+	 */
+	xchk_iscan_start(sc, 5000, 100, &ri->ftype_iscan);
+	xchk_iscan_set_agi_trylock(&ri->ftype_iscan);
+	ri->ftype_iscan.skip_ino = sc->sm->sm_ino;
+	ri->alleged_ftype = XFS_DIR3_FT_UNKNOWN;
+	while ((error = xchk_iscan_iter(&ri->ftype_iscan, &dp)) == 1) {
+		if (S_ISDIR(VFS_I(dp)->i_mode))
+			error = xrep_dinode_findmode_walk_directory(ri, dp);
+		xchk_iscan_mark_visited(&ri->ftype_iscan, dp);
+		xchk_irele(sc, dp);
+		if (error < 0)
+			break;
+		if (xchk_should_terminate(sc, &error))
+			break;
+	}
+	xchk_iscan_iter_finish(&ri->ftype_iscan);
+	xchk_iscan_teardown(&ri->ftype_iscan);
+
+	if (error == -EBUSY) {
+		if (ri->alleged_ftype != XFS_DIR3_FT_UNKNOWN) {
+			/*
+			 * If we got an EBUSY after finding at least one
+			 * dirent, that means the scan found an inode on the
+			 * inactivation list and could not open it.  Accept the
+			 * alleged ftype and install a new mode below.
+			 */
+			error = 0;
+		} else if (!(sc->flags & XCHK_TRY_HARDER)) {
+			/*
+			 * Otherwise, retry the operation one time to see if
+			 * the reason for the delay is an inode from the same
+			 * cluster buffer waiting on the inactivation list.
+			 */
+			error = -EDEADLOCK;
+		}
+	}
+	if (error)
+		return error;
+
+	/*
+	 * Convert the discovered ftype into the file mode.  If all else fails,
+	 * return S_IFREG.
+	 */
+	switch (ri->alleged_ftype) {
+	case XFS_DIR3_FT_DIR:
+		*mode = S_IFDIR;
+		break;
+	case XFS_DIR3_FT_WHT:
+	case XFS_DIR3_FT_CHRDEV:
+		*mode = S_IFCHR;
+		break;
+	case XFS_DIR3_FT_BLKDEV:
+		*mode = S_IFBLK;
+		break;
+	case XFS_DIR3_FT_FIFO:
+		*mode = S_IFIFO;
+		break;
+	case XFS_DIR3_FT_SOCK:
+		*mode = S_IFSOCK;
+		break;
+	case XFS_DIR3_FT_SYMLINK:
+		*mode = S_IFLNK;
+		break;
+	default:
+		*mode = S_IFREG;
+		break;
+	}
+	return 0;
+}
+
+/* Turn di_mode into /something/ recognizable.  Returns true if we succeed. */
+STATIC int
 xrep_dinode_mode(
 	struct xrep_inode	*ri,
 	struct xfs_dinode	*dip)
 {
 	struct xfs_scrub	*sc = ri->sc;
 	uint16_t		mode = be16_to_cpu(dip->di_mode);
+	int			error;
 
 	trace_xrep_dinode_mode(sc, dip);
 
 	if (mode == 0 || xfs_mode_to_ftype(mode) != XFS_DIR3_FT_UNKNOWN)
-		return;
+		return 0;
+
+	/* Try to fix the mode.  If we cannot, then leave everything alone. */
+	error = xrep_dinode_find_mode(ri, &mode);
+	switch (error) {
+	case -EINTR:
+	case -EBUSY:
+	case -EDEADLOCK:
+		/* temporary failure or fatal signal */
+		return error;
+	case 0:
+		/* found mode */
+		break;
+	default:
+		/* some other error, assume S_IFREG */
+		mode = S_IFREG;
+		break;
+	}
 
 	/* bad mode, so we set it to a file that only root can read */
-	mode = S_IFREG;
 	dip->di_mode = cpu_to_be16(mode);
 	dip->di_uid = 0;
 	dip->di_gid = 0;
 	ri->zap_acls = true;
+	return 0;
+}
+
+/* Fix unused link count fields having nonzero values. */
+STATIC void
+xrep_dinode_nlinks(
+	struct xfs_dinode	*dip)
+{
+	if (dip->di_version > 1)
+		dip->di_onlink = 0;
+	else
+		dip->di_nlink = 0;
 }
 
 /* Fix any conflicting flags that the verifiers complain about. */
@@ -1107,12 +1385,16 @@ xrep_dinode_core(
 	/* Fix everything the verifier will complain about. */
 	dip = xfs_buf_offset(bp, ri->imap.im_boffset);
 	xrep_dinode_header(sc, dip);
-	xrep_dinode_mode(ri, dip);
+	iget_error = xrep_dinode_mode(ri, dip);
+	if (iget_error)
+		goto write;
+	xrep_dinode_nlinks(dip);
 	xrep_dinode_flags(sc, dip, ri->rt_extents > 0);
 	xrep_dinode_size(ri, dip);
 	xrep_dinode_extsize_hints(sc, dip);
 	xrep_dinode_zap_forks(ri, dip);
 
+write:
 	/* Write out the inode. */
 	trace_xrep_dinode_fixed(sc, dip);
 	xfs_dinode_calc_crc(sc->mp, dip);
@@ -1128,7 +1410,8 @@ xrep_dinode_core(
 	 * accessing the inode.  If iget fails, we still need to commit the
 	 * changes.
 	 */
-	iget_error = xchk_iget(sc, ino, &sc->ip);
+	if (!iget_error)
+		iget_error = xchk_iget(sc, ino, &sc->ip);
 	if (!iget_error)
 		xchk_ilock(sc, XFS_IOLOCK_EXCL);
 
@@ -1453,6 +1736,44 @@ xrep_inode_extsize(
 	}
 }
 
+/* Ensure this file has an attr fork if it needs to hold a parent pointer. */
+STATIC int
+xrep_inode_pptr(
+	struct xfs_scrub	*sc)
+{
+	struct xfs_mount	*mp = sc->mp;
+	struct xfs_inode	*ip = sc->ip;
+	struct inode		*inode = VFS_I(ip);
+
+	if (!xfs_has_parent(mp))
+		return 0;
+
+	/*
+	 * Unlinked inodes that cannot be added to the directory tree will not
+	 * have a parent pointer.
+	 */
+	if (inode->i_nlink == 0 && !(inode->i_state & I_LINKABLE))
+		return 0;
+
+	/* The root directory doesn't have a parent pointer. */
+	if (ip == mp->m_rootip)
+		return 0;
+
+	/*
+	 * Metadata inodes are rooted in the superblock and do not have any
+	 * parents.
+	 */
+	if (xfs_is_metadata_inode(ip))
+		return 0;
+
+	/* Inode already has an attr fork; no further work possible here. */
+	if (xfs_inode_has_attr_fork(ip))
+		return 0;
+
+	return xfs_bmap_add_attrfork(sc->tp, ip,
+			sizeof(struct xfs_attr_sf_hdr), true);
+}
+
 /* Fix any irregularities in an inode that the verifiers don't catch. */
 STATIC int
 xrep_inode_problems(
@@ -1463,6 +1784,9 @@ xrep_inode_problems(
 	error = xrep_inode_blockcounts(sc);
 	if (error)
 		return error;
+	error = xrep_inode_pptr(sc);
+	if (error)
+		return error;
 	xrep_inode_timestamps(sc->ip);
 	xrep_inode_flags(sc);
 	xrep_inode_ids(sc);
@@ -1479,6 +1803,46 @@ xrep_inode_problems(
 	return xrep_roll_trans(sc);
 }
 
+/*
+ * Make sure this inode's unlinked list pointers are consistent with its
+ * link count.
+ */
+STATIC int
+xrep_inode_unlinked(
+	struct xfs_scrub	*sc)
+{
+	unsigned int		nlink = VFS_I(sc->ip)->i_nlink;
+	int			error;
+
+	/*
+	 * If this inode is linked from the directory tree and on the unlinked
+	 * list, remove it from the unlinked list.
+	 */
+	if (nlink > 0 && xfs_inode_on_unlinked_list(sc->ip)) {
+		struct xfs_perag	*pag;
+		int			error;
+
+		pag = xfs_perag_get(sc->mp,
+				XFS_INO_TO_AGNO(sc->mp, sc->ip->i_ino));
+		error = xfs_iunlink_remove(sc->tp, pag, sc->ip);
+		xfs_perag_put(pag);
+		if (error)
+			return error;
+	}
+
+	/*
+	 * If this inode is not linked from the directory tree yet not on the
+	 * unlinked list, put it on the unlinked list.
+	 */
+	if (nlink == 0 && !xfs_inode_on_unlinked_list(sc->ip)) {
+		error = xfs_iunlink(sc->tp, sc->ip);
+		if (error)
+			return error;
+	}
+
+	return 0;
+}
+
 /* Repair an inode's fields. */
 int
 xrep_inode(
@@ -1496,6 +1860,13 @@ xrep_inode(
 		ASSERT(ri != NULL);
 
 		error = xrep_dinode_problems(ri);
+		if (error == -EBUSY) {
+			/*
+			 * Directory scan to recover inode mode encountered a
+			 * busy inode, so we did not continue repairing things.
+			 */
+			return 0;
+		}
 		if (error)
 			return error;
 
@@ -1521,5 +1892,10 @@ xrep_inode(
 			return error;
 	}
 
+	/* Reconnect incore unlinked list */
+	error = xrep_inode_unlinked(sc);
+	if (error)
+		return error;
+
 	return xrep_defer_finish(sc);
 }
diff --git a/fs/xfs/scrub/iscan.c b/fs/xfs/scrub/iscan.c
new file mode 100644
index 0000000000..cf9d983667
--- /dev/null
+++ b/fs/xfs/scrub/iscan.c
@@ -0,0 +1,826 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2021-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_inode.h"
+#include "xfs_btree.h"
+#include "xfs_ialloc.h"
+#include "xfs_ialloc_btree.h"
+#include "xfs_ag.h"
+#include "xfs_error.h"
+#include "xfs_bit.h"
+#include "xfs_icache.h"
+#include "scrub/scrub.h"
+#include "scrub/iscan.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+
+/*
+ * Live File Scan
+ * ==============
+ *
+ * Live file scans walk every inode in a live filesystem.  This is more or
+ * less like a regular iwalk, except that when we're advancing the scan cursor,
+ * we must ensure that inodes cannot be added or deleted anywhere between the
+ * old cursor value and the new cursor value.  If we're advancing the cursor
+ * by one inode, the caller must hold that inode; if we're finding the next
+ * inode to scan, we must grab the AGI and hold it until we've updated the
+ * scan cursor.
+ *
+ * Callers are expected to use this code to scan all files in the filesystem to
+ * construct a new metadata index of some kind.  The scan races against other
+ * live updates, which means there must be a provision to update the new index
+ * when updates are made to inodes that already been scanned.  The iscan lock
+ * can be used in live update hook code to stop the scan and protect this data
+ * structure.
+ *
+ * To keep the new index up to date with other metadata updates being made to
+ * the live filesystem, it is assumed that the caller will add hooks as needed
+ * to be notified when a metadata update occurs.  The inode scanner must tell
+ * the hook code when an inode has been visited with xchk_iscan_mark_visit.
+ * Hook functions can use xchk_iscan_want_live_update to decide if the
+ * scanner's observations must be updated.
+ */
+
+/*
+ * If the inobt record @rec covers @iscan->skip_ino, mark the inode free so
+ * that the scan ignores that inode.
+ */
+STATIC void
+xchk_iscan_mask_skipino(
+	struct xchk_iscan	*iscan,
+	struct xfs_perag	*pag,
+	struct xfs_inobt_rec_incore	*rec,
+	xfs_agino_t		lastrecino)
+{
+	struct xfs_scrub	*sc = iscan->sc;
+	struct xfs_mount	*mp = sc->mp;
+	xfs_agnumber_t		skip_agno = XFS_INO_TO_AGNO(mp, iscan->skip_ino);
+	xfs_agnumber_t		skip_agino = XFS_INO_TO_AGINO(mp, iscan->skip_ino);
+
+	if (pag->pag_agno != skip_agno)
+		return;
+	if (skip_agino < rec->ir_startino)
+		return;
+	if (skip_agino > lastrecino)
+		return;
+
+	rec->ir_free |= xfs_inobt_maskn(skip_agino - rec->ir_startino, 1);
+}
+
+/*
+ * Set *cursor to the next allocated inode after whatever it's set to now.
+ * If there are no more inodes in this AG, cursor is set to NULLAGINO.
+ */
+STATIC int
+xchk_iscan_find_next(
+	struct xchk_iscan	*iscan,
+	struct xfs_buf		*agi_bp,
+	struct xfs_perag	*pag,
+	xfs_inofree_t		*allocmaskp,
+	xfs_agino_t		*cursor,
+	uint8_t			*nr_inodesp)
+{
+	struct xfs_scrub	*sc = iscan->sc;
+	struct xfs_inobt_rec_incore	rec;
+	struct xfs_btree_cur	*cur;
+	struct xfs_mount	*mp = sc->mp;
+	struct xfs_trans	*tp = sc->tp;
+	xfs_agnumber_t		agno = pag->pag_agno;
+	xfs_agino_t		lastino = NULLAGINO;
+	xfs_agino_t		first, last;
+	xfs_agino_t		agino = *cursor;
+	int			has_rec;
+	int			error;
+
+	/* If the cursor is beyond the end of this AG, move to the next one. */
+	xfs_agino_range(mp, agno, &first, &last);
+	if (agino > last) {
+		*cursor = NULLAGINO;
+		return 0;
+	}
+
+	/*
+	 * Look up the inode chunk for the current cursor position.  If there
+	 * is no chunk here, we want the next one.
+	 */
+	cur = xfs_inobt_init_cursor(pag, tp, agi_bp);
+	error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, &has_rec);
+	if (!error && !has_rec)
+		error = xfs_btree_increment(cur, 0, &has_rec);
+	for (; !error; error = xfs_btree_increment(cur, 0, &has_rec)) {
+		xfs_inofree_t	allocmask;
+
+		/*
+		 * If we've run out of inobt records in this AG, move the
+		 * cursor on to the next AG and exit.  The caller can try
+		 * again with the next AG.
+		 */
+		if (!has_rec) {
+			*cursor = NULLAGINO;
+			break;
+		}
+
+		error = xfs_inobt_get_rec(cur, &rec, &has_rec);
+		if (error)
+			break;
+		if (!has_rec) {
+			error = -EFSCORRUPTED;
+			break;
+		}
+
+		/* Make sure that we always move forward. */
+		if (lastino != NULLAGINO &&
+		    XFS_IS_CORRUPT(mp, lastino >= rec.ir_startino)) {
+			error = -EFSCORRUPTED;
+			break;
+		}
+		lastino = rec.ir_startino + XFS_INODES_PER_CHUNK - 1;
+
+		/*
+		 * If this record only covers inodes that come before the
+		 * cursor, advance to the next record.
+		 */
+		if (rec.ir_startino + XFS_INODES_PER_CHUNK <= agino)
+			continue;
+
+		if (iscan->skip_ino)
+			xchk_iscan_mask_skipino(iscan, pag, &rec, lastino);
+
+		/*
+		 * If the incoming lookup put us in the middle of an inobt
+		 * record, mark it and the previous inodes "free" so that the
+		 * search for allocated inodes will start at the cursor.
+		 * We don't care about ir_freecount here.
+		 */
+		if (agino >= rec.ir_startino)
+			rec.ir_free |= xfs_inobt_maskn(0,
+						agino + 1 - rec.ir_startino);
+
+		/*
+		 * If there are allocated inodes in this chunk, find them
+		 * and update the scan cursor.
+		 */
+		allocmask = ~rec.ir_free;
+		if (hweight64(allocmask) > 0) {
+			int	next = xfs_lowbit64(allocmask);
+
+			ASSERT(next >= 0);
+			*cursor = rec.ir_startino + next;
+			*allocmaskp = allocmask >> next;
+			*nr_inodesp = XFS_INODES_PER_CHUNK - next;
+			break;
+		}
+	}
+
+	xfs_btree_del_cursor(cur, error);
+	return error;
+}
+
+/*
+ * Advance both the scan and the visited cursors.
+ *
+ * The inumber address space for a given filesystem is sparse, which means that
+ * the scan cursor can jump a long ways in a single iter() call.  There are no
+ * inodes in these sparse areas, so we must move the visited cursor forward at
+ * the same time so that the scan user can receive live updates for inodes that
+ * may get created once we release the AGI buffer.
+ */
+static inline void
+xchk_iscan_move_cursor(
+	struct xchk_iscan	*iscan,
+	xfs_agnumber_t		agno,
+	xfs_agino_t		agino)
+{
+	struct xfs_scrub	*sc = iscan->sc;
+	struct xfs_mount	*mp = sc->mp;
+	xfs_ino_t		cursor, visited;
+
+	BUILD_BUG_ON(XFS_MAXINUMBER == NULLFSINO);
+
+	/*
+	 * Special-case ino == 0 here so that we never set visited_ino to
+	 * NULLFSINO when wrapping around EOFS, for that will let through all
+	 * live updates.
+	 */
+	cursor = XFS_AGINO_TO_INO(mp, agno, agino);
+	if (cursor == 0)
+		visited = XFS_MAXINUMBER;
+	else
+		visited = cursor - 1;
+
+	mutex_lock(&iscan->lock);
+	iscan->cursor_ino = cursor;
+	iscan->__visited_ino = visited;
+	trace_xchk_iscan_move_cursor(iscan);
+	mutex_unlock(&iscan->lock);
+}
+
+/*
+ * Prepare to return agno/agino to the iscan caller by moving the lastino
+ * cursor to the previous inode.  Do this while we still hold the AGI so that
+ * no other threads can create or delete inodes in this AG.
+ */
+static inline void
+xchk_iscan_finish(
+	struct xchk_iscan	*iscan)
+{
+	mutex_lock(&iscan->lock);
+	iscan->cursor_ino = NULLFSINO;
+
+	/* All live updates will be applied from now on */
+	iscan->__visited_ino = NULLFSINO;
+
+	mutex_unlock(&iscan->lock);
+}
+
+/* Mark an inode scan finished before we actually scan anything. */
+void
+xchk_iscan_finish_early(
+	struct xchk_iscan	*iscan)
+{
+	ASSERT(iscan->cursor_ino == iscan->scan_start_ino);
+	ASSERT(iscan->__visited_ino == iscan->scan_start_ino);
+
+	xchk_iscan_finish(iscan);
+}
+
+/*
+ * Grab the AGI to advance the inode scan.  Returns 0 if *agi_bpp is now set,
+ * -ECANCELED if the live scan aborted, -EBUSY if the AGI could not be grabbed,
+ * or the usual negative errno.
+ */
+STATIC int
+xchk_iscan_read_agi(
+	struct xchk_iscan	*iscan,
+	struct xfs_perag	*pag,
+	struct xfs_buf		**agi_bpp)
+{
+	struct xfs_scrub	*sc = iscan->sc;
+	unsigned long		relax;
+	int			ret;
+
+	if (!xchk_iscan_agi_needs_trylock(iscan))
+		return xfs_ialloc_read_agi(pag, sc->tp, 0, agi_bpp);
+
+	relax = msecs_to_jiffies(iscan->iget_retry_delay);
+	do {
+		ret = xfs_ialloc_read_agi(pag, sc->tp, XFS_IALLOC_FLAG_TRYLOCK,
+				agi_bpp);
+		if (ret != -EAGAIN)
+			return ret;
+		if (!iscan->iget_timeout ||
+		    time_is_before_jiffies(iscan->__iget_deadline))
+			return -EBUSY;
+
+		trace_xchk_iscan_agi_retry_wait(iscan);
+	} while (!schedule_timeout_killable(relax) &&
+		 !xchk_iscan_aborted(iscan));
+	return -ECANCELED;
+}
+
+/*
+ * Advance ino to the next inode that the inobt thinks is allocated, being
+ * careful to jump to the next AG if we've reached the right end of this AG's
+ * inode btree.  Advancing ino effectively means that we've pushed the inode
+ * scan forward, so set the iscan cursor to (ino - 1) so that our live update
+ * predicates will track inode allocations in that part of the inode number
+ * key space once we release the AGI buffer.
+ *
+ * Returns 1 if there's a new inode to examine, 0 if we've run out of inodes,
+ * -ECANCELED if the live scan aborted, or the usual negative errno.
+ */
+STATIC int
+xchk_iscan_advance(
+	struct xchk_iscan	*iscan,
+	struct xfs_perag	**pagp,
+	struct xfs_buf		**agi_bpp,
+	xfs_inofree_t		*allocmaskp,
+	uint8_t			*nr_inodesp)
+{
+	struct xfs_scrub	*sc = iscan->sc;
+	struct xfs_mount	*mp = sc->mp;
+	struct xfs_buf		*agi_bp;
+	struct xfs_perag	*pag;
+	xfs_agnumber_t		agno;
+	xfs_agino_t		agino;
+	int			ret;
+
+	ASSERT(iscan->cursor_ino >= iscan->__visited_ino);
+
+	do {
+		if (xchk_iscan_aborted(iscan))
+			return -ECANCELED;
+
+		agno = XFS_INO_TO_AGNO(mp, iscan->cursor_ino);
+		pag = xfs_perag_get(mp, agno);
+		if (!pag)
+			return -ECANCELED;
+
+		ret = xchk_iscan_read_agi(iscan, pag, &agi_bp);
+		if (ret)
+			goto out_pag;
+
+		agino = XFS_INO_TO_AGINO(mp, iscan->cursor_ino);
+		ret = xchk_iscan_find_next(iscan, agi_bp, pag, allocmaskp,
+				&agino, nr_inodesp);
+		if (ret)
+			goto out_buf;
+
+		if (agino != NULLAGINO) {
+			/*
+			 * Found the next inode in this AG, so return it along
+			 * with the AGI buffer and the perag structure to
+			 * ensure it cannot go away.
+			 */
+			xchk_iscan_move_cursor(iscan, agno, agino);
+			*agi_bpp = agi_bp;
+			*pagp = pag;
+			return 1;
+		}
+
+		/*
+		 * Did not find any more inodes in this AG, move on to the next
+		 * AG.
+		 */
+		agno = (agno + 1) % mp->m_sb.sb_agcount;
+		xchk_iscan_move_cursor(iscan, agno, 0);
+		xfs_trans_brelse(sc->tp, agi_bp);
+		xfs_perag_put(pag);
+
+		trace_xchk_iscan_advance_ag(iscan);
+	} while (iscan->cursor_ino != iscan->scan_start_ino);
+
+	xchk_iscan_finish(iscan);
+	return 0;
+
+out_buf:
+	xfs_trans_brelse(sc->tp, agi_bp);
+out_pag:
+	xfs_perag_put(pag);
+	return ret;
+}
+
+/*
+ * Grabbing the inode failed, so we need to back up the scan and ask the caller
+ * to try to _advance the scan again.  Returns -EBUSY if we've run out of retry
+ * opportunities, -ECANCELED if the process has a fatal signal pending, or
+ * -EAGAIN if we should try again.
+ */
+STATIC int
+xchk_iscan_iget_retry(
+	struct xchk_iscan	*iscan,
+	bool			wait)
+{
+	ASSERT(iscan->cursor_ino == iscan->__visited_ino + 1);
+
+	if (!iscan->iget_timeout ||
+	    time_is_before_jiffies(iscan->__iget_deadline))
+		return -EBUSY;
+
+	if (wait) {
+		unsigned long	relax;
+
+		/*
+		 * Sleep for a period of time to let the rest of the system
+		 * catch up.  If we return early, someone sent a kill signal to
+		 * the calling process.
+		 */
+		relax = msecs_to_jiffies(iscan->iget_retry_delay);
+		trace_xchk_iscan_iget_retry_wait(iscan);
+
+		if (schedule_timeout_killable(relax) ||
+		    xchk_iscan_aborted(iscan))
+			return -ECANCELED;
+	}
+
+	iscan->cursor_ino--;
+	return -EAGAIN;
+}
+
+/*
+ * For an inode scan, we hold the AGI and want to try to grab a batch of
+ * inodes.  Holding the AGI prevents inodegc from clearing freed inodes,
+ * so we must use noretry here.  For every inode after the first one in the
+ * batch, we don't want to wait, so we use retry there too.  Finally, use
+ * dontcache to avoid polluting the cache.
+ */
+#define ISCAN_IGET_FLAGS	(XFS_IGET_NORETRY | XFS_IGET_DONTCACHE)
+
+/*
+ * Grab an inode as part of an inode scan.  While scanning this inode, the
+ * caller must ensure that no other threads can modify the inode until a call
+ * to xchk_iscan_visit succeeds.
+ *
+ * Returns the number of incore inodes grabbed; -EAGAIN if the caller should
+ * call again xchk_iscan_advance; -EBUSY if we couldn't grab an inode;
+ * -ECANCELED if there's a fatal signal pending; or some other negative errno.
+ */
+STATIC int
+xchk_iscan_iget(
+	struct xchk_iscan	*iscan,
+	struct xfs_perag	*pag,
+	struct xfs_buf		*agi_bp,
+	xfs_inofree_t		allocmask,
+	uint8_t			nr_inodes)
+{
+	struct xfs_scrub	*sc = iscan->sc;
+	struct xfs_mount	*mp = sc->mp;
+	xfs_ino_t		ino = iscan->cursor_ino;
+	unsigned int		idx = 0;
+	unsigned int		i;
+	int			error;
+
+	ASSERT(iscan->__inodes[0] == NULL);
+
+	/* Fill the first slot in the inode array. */
+	error = xfs_iget(sc->mp, sc->tp, ino, ISCAN_IGET_FLAGS, 0,
+			&iscan->__inodes[idx]);
+
+	trace_xchk_iscan_iget(iscan, error);
+
+	if (error == -ENOENT || error == -EAGAIN) {
+		xfs_trans_brelse(sc->tp, agi_bp);
+		xfs_perag_put(pag);
+
+		/*
+		 * It's possible that this inode has lost all of its links but
+		 * hasn't yet been inactivated.  If we don't have a transaction
+		 * or it's not writable, flush the inodegc workers and wait.
+		 * If we have a non-empty transaction, we must not block on
+		 * inodegc, which allocates its own transactions.
+		 */
+		if (sc->tp && !(sc->tp->t_flags & XFS_TRANS_NO_WRITECOUNT))
+			xfs_inodegc_push(mp);
+		else
+			xfs_inodegc_flush(mp);
+		return xchk_iscan_iget_retry(iscan, true);
+	}
+
+	if (error == -EINVAL) {
+		xfs_trans_brelse(sc->tp, agi_bp);
+		xfs_perag_put(pag);
+
+		/*
+		 * We thought the inode was allocated, but the inode btree
+		 * lookup failed, which means that it was freed since the last
+		 * time we advanced the cursor.  Back up and try again.  This
+		 * should never happen since still hold the AGI buffer from the
+		 * inobt check, but we need to be careful about infinite loops.
+		 */
+		return xchk_iscan_iget_retry(iscan, false);
+	}
+
+	if (error) {
+		xfs_trans_brelse(sc->tp, agi_bp);
+		xfs_perag_put(pag);
+		return error;
+	}
+	idx++;
+	ino++;
+	allocmask >>= 1;
+
+	/*
+	 * Now that we've filled the first slot in __inodes, try to fill the
+	 * rest of the batch with consecutively ordered inodes.  to reduce the
+	 * number of _iter calls.  Make a bitmap of unallocated inodes from the
+	 * zeroes in the inuse bitmap; these inodes will not be scanned, but
+	 * the _want_live_update predicate will pass through all live updates.
+	 *
+	 * If we can't iget an allocated inode, stop and return what we have.
+	 */
+	mutex_lock(&iscan->lock);
+	iscan->__batch_ino = ino - 1;
+	iscan->__skipped_inomask = 0;
+	mutex_unlock(&iscan->lock);
+
+	for (i = 1; i < nr_inodes; i++, ino++, allocmask >>= 1) {
+		if (!(allocmask & 1)) {
+			ASSERT(!(iscan->__skipped_inomask & (1ULL << i)));
+
+			mutex_lock(&iscan->lock);
+			iscan->cursor_ino = ino;
+			iscan->__skipped_inomask |= (1ULL << i);
+			mutex_unlock(&iscan->lock);
+			continue;
+		}
+
+		ASSERT(iscan->__inodes[idx] == NULL);
+
+		error = xfs_iget(sc->mp, sc->tp, ino, ISCAN_IGET_FLAGS, 0,
+				&iscan->__inodes[idx]);
+		if (error)
+			break;
+
+		mutex_lock(&iscan->lock);
+		iscan->cursor_ino = ino;
+		mutex_unlock(&iscan->lock);
+		idx++;
+	}
+
+	trace_xchk_iscan_iget_batch(sc->mp, iscan, nr_inodes, idx);
+	xfs_trans_brelse(sc->tp, agi_bp);
+	xfs_perag_put(pag);
+	return idx;
+}
+
+/*
+ * Advance the visit cursor to reflect skipped inodes beyond whatever we
+ * scanned.
+ */
+STATIC void
+xchk_iscan_finish_batch(
+	struct xchk_iscan	*iscan)
+{
+	xfs_ino_t		highest_skipped;
+
+	mutex_lock(&iscan->lock);
+
+	if (iscan->__batch_ino != NULLFSINO) {
+		highest_skipped = iscan->__batch_ino +
+					xfs_highbit64(iscan->__skipped_inomask);
+		iscan->__visited_ino = max(iscan->__visited_ino,
+					   highest_skipped);
+
+		trace_xchk_iscan_skip(iscan);
+	}
+
+	iscan->__batch_ino = NULLFSINO;
+	iscan->__skipped_inomask = 0;
+
+	mutex_unlock(&iscan->lock);
+}
+
+/*
+ * Advance the inode scan cursor to the next allocated inode and return up to
+ * 64 consecutive allocated inodes starting with the cursor position.
+ */
+STATIC int
+xchk_iscan_iter_batch(
+	struct xchk_iscan	*iscan)
+{
+	struct xfs_scrub	*sc = iscan->sc;
+	int			ret;
+
+	xchk_iscan_finish_batch(iscan);
+
+	if (iscan->iget_timeout)
+		iscan->__iget_deadline = jiffies +
+					 msecs_to_jiffies(iscan->iget_timeout);
+
+	do {
+		struct xfs_buf	*agi_bp = NULL;
+		struct xfs_perag *pag = NULL;
+		xfs_inofree_t	allocmask = 0;
+		uint8_t		nr_inodes = 0;
+
+		ret = xchk_iscan_advance(iscan, &pag, &agi_bp, &allocmask,
+				&nr_inodes);
+		if (ret != 1)
+			return ret;
+
+		if (xchk_iscan_aborted(iscan)) {
+			xfs_trans_brelse(sc->tp, agi_bp);
+			xfs_perag_put(pag);
+			ret = -ECANCELED;
+			break;
+		}
+
+		ret = xchk_iscan_iget(iscan, pag, agi_bp, allocmask, nr_inodes);
+	} while (ret == -EAGAIN);
+
+	return ret;
+}
+
+/*
+ * Advance the inode scan cursor to the next allocated inode and return the
+ * incore inode structure associated with it.
+ *
+ * Returns 1 if there's a new inode to examine, 0 if we've run out of inodes,
+ * -ECANCELED if the live scan aborted, -EBUSY if the incore inode could not be
+ * grabbed, or the usual negative errno.
+ *
+ * If the function returns -EBUSY and the caller can handle skipping an inode,
+ * it may call this function again to continue the scan with the next allocated
+ * inode.
+ */
+int
+xchk_iscan_iter(
+	struct xchk_iscan	*iscan,
+	struct xfs_inode	**ipp)
+{
+	unsigned int		i;
+	int			error;
+
+	/* Find a cached inode, or go get another batch. */
+	for (i = 0; i < XFS_INODES_PER_CHUNK; i++) {
+		if (iscan->__inodes[i])
+			goto foundit;
+	}
+
+	error = xchk_iscan_iter_batch(iscan);
+	if (error <= 0)
+		return error;
+
+	ASSERT(iscan->__inodes[0] != NULL);
+	i = 0;
+
+foundit:
+	/* Give the caller our reference. */
+	*ipp = iscan->__inodes[i];
+	iscan->__inodes[i] = NULL;
+	return 1;
+}
+
+/* Clean up an xfs_iscan_iter call by dropping any inodes that we still hold. */
+void
+xchk_iscan_iter_finish(
+	struct xchk_iscan	*iscan)
+{
+	struct xfs_scrub	*sc = iscan->sc;
+	unsigned int		i;
+
+	for (i = 0; i < XFS_INODES_PER_CHUNK; i++) {
+		if (iscan->__inodes[i]) {
+			xchk_irele(sc, iscan->__inodes[i]);
+			iscan->__inodes[i] = NULL;
+		}
+	}
+}
+
+/* Mark this inode scan finished and release resources. */
+void
+xchk_iscan_teardown(
+	struct xchk_iscan	*iscan)
+{
+	xchk_iscan_iter_finish(iscan);
+	xchk_iscan_finish(iscan);
+	mutex_destroy(&iscan->lock);
+}
+
+/* Pick an AG from which to start a scan. */
+static inline xfs_ino_t
+xchk_iscan_rotor(
+	struct xfs_mount	*mp)
+{
+	static atomic_t		agi_rotor;
+	unsigned int		r = atomic_inc_return(&agi_rotor) - 1;
+
+	/*
+	 * Rotoring *backwards* through the AGs, so we add one here before
+	 * subtracting from the agcount to arrive at an AG number.
+	 */
+	r = (r % mp->m_sb.sb_agcount) + 1;
+
+	return XFS_AGINO_TO_INO(mp, mp->m_sb.sb_agcount - r, 0);
+}
+
+/*
+ * Set ourselves up to start an inode scan.  If the @iget_timeout and
+ * @iget_retry_delay parameters are set, the scan will try to iget each inode
+ * for @iget_timeout milliseconds.  If an iget call indicates that the inode is
+ * waiting to be inactivated, the CPU will relax for @iget_retry_delay
+ * milliseconds after pushing the inactivation workers.
+ */
+void
+xchk_iscan_start(
+	struct xfs_scrub	*sc,
+	unsigned int		iget_timeout,
+	unsigned int		iget_retry_delay,
+	struct xchk_iscan	*iscan)
+{
+	xfs_ino_t		start_ino;
+
+	start_ino = xchk_iscan_rotor(sc->mp);
+
+	iscan->__batch_ino = NULLFSINO;
+	iscan->__skipped_inomask = 0;
+
+	iscan->sc = sc;
+	clear_bit(XCHK_ISCAN_OPSTATE_ABORTED, &iscan->__opstate);
+	iscan->iget_timeout = iget_timeout;
+	iscan->iget_retry_delay = iget_retry_delay;
+	iscan->__visited_ino = start_ino;
+	iscan->cursor_ino = start_ino;
+	iscan->scan_start_ino = start_ino;
+	mutex_init(&iscan->lock);
+	memset(iscan->__inodes, 0, sizeof(iscan->__inodes));
+
+	trace_xchk_iscan_start(iscan, start_ino);
+}
+
+/*
+ * Mark this inode as having been visited.  Callers must hold a sufficiently
+ * exclusive lock on the inode to prevent concurrent modifications.
+ */
+void
+xchk_iscan_mark_visited(
+	struct xchk_iscan	*iscan,
+	struct xfs_inode	*ip)
+{
+	mutex_lock(&iscan->lock);
+	iscan->__visited_ino = ip->i_ino;
+	trace_xchk_iscan_visit(iscan);
+	mutex_unlock(&iscan->lock);
+}
+
+/*
+ * Did we skip this inode because it wasn't allocated when we loaded the batch?
+ * If so, it is newly allocated and will not be scanned.  All live updates to
+ * this inode must be passed to the caller to maintain scan correctness.
+ */
+static inline bool
+xchk_iscan_skipped(
+	const struct xchk_iscan	*iscan,
+	xfs_ino_t		ino)
+{
+	if (iscan->__batch_ino == NULLFSINO)
+		return false;
+	if (ino < iscan->__batch_ino)
+		return false;
+	if (ino >= iscan->__batch_ino + XFS_INODES_PER_CHUNK)
+		return false;
+
+	return iscan->__skipped_inomask & (1ULL << (ino - iscan->__batch_ino));
+}
+
+/*
+ * Do we need a live update for this inode?  This is true if the scanner thread
+ * has visited this inode and the scan hasn't been aborted due to errors.
+ * Callers must hold a sufficiently exclusive lock on the inode to prevent
+ * scanners from reading any inode metadata.
+ */
+bool
+xchk_iscan_want_live_update(
+	struct xchk_iscan	*iscan,
+	xfs_ino_t		ino)
+{
+	bool			ret = false;
+
+	if (xchk_iscan_aborted(iscan))
+		return false;
+
+	mutex_lock(&iscan->lock);
+
+	trace_xchk_iscan_want_live_update(iscan, ino);
+
+	/* Scan is finished, caller should receive all updates. */
+	if (iscan->__visited_ino == NULLFSINO) {
+		ret = true;
+		goto unlock;
+	}
+
+	/*
+	 * No inodes have been visited yet, so the visited cursor points at the
+	 * start of the scan range.  The caller should not receive any updates.
+	 */
+	if (iscan->scan_start_ino == iscan->__visited_ino) {
+		ret = false;
+		goto unlock;
+	}
+
+	/*
+	 * This inode was not allocated at the time of the iscan batch.
+	 * The caller should receive all updates.
+	 */
+	if (xchk_iscan_skipped(iscan, ino)) {
+		ret = true;
+		goto unlock;
+	}
+
+	/*
+	 * The visited cursor hasn't yet wrapped around the end of the FS.  If
+	 * @ino is inside the starred range, the caller should receive updates:
+	 *
+	 * 0 ------------ S ************ V ------------ EOFS
+	 */
+	if (iscan->scan_start_ino <= iscan->__visited_ino) {
+		if (ino >= iscan->scan_start_ino &&
+		    ino <= iscan->__visited_ino)
+			ret = true;
+
+		goto unlock;
+	}
+
+	/*
+	 * The visited cursor wrapped around the end of the FS.  If @ino is
+	 * inside the starred range, the caller should receive updates:
+	 *
+	 * 0 ************ V ------------ S ************ EOFS
+	 */
+	if (ino >= iscan->scan_start_ino || ino <= iscan->__visited_ino)
+		ret = true;
+
+unlock:
+	mutex_unlock(&iscan->lock);
+	return ret;
+}
diff --git a/fs/xfs/scrub/iscan.h b/fs/xfs/scrub/iscan.h
new file mode 100644
index 0000000000..f9f47fa01a
--- /dev/null
+++ b/fs/xfs/scrub/iscan.h
@@ -0,0 +1,100 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (c) 2021-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_ISCAN_H__
+#define __XFS_SCRUB_ISCAN_H__
+
+struct xchk_iscan {
+	struct xfs_scrub	*sc;
+
+	/* Lock to protect the scan cursor. */
+	struct mutex		lock;
+
+	/*
+	 * This is the first inode in the inumber address space that we
+	 * examined.  When the scan wraps around back to here, the scan is
+	 * finished.
+	 */
+	xfs_ino_t		scan_start_ino;
+
+	/* This is the inode that will be examined next. */
+	xfs_ino_t		cursor_ino;
+
+	/* If nonzero and non-NULL, skip this inode when scanning. */
+	xfs_ino_t		skip_ino;
+
+	/*
+	 * This is the last inode that we've successfully scanned, either
+	 * because the caller scanned it, or we moved the cursor past an empty
+	 * part of the inode address space.  Scan callers should only use the
+	 * xchk_iscan_visit function to modify this.
+	 */
+	xfs_ino_t		__visited_ino;
+
+	/* Operational state of the livescan. */
+	unsigned long		__opstate;
+
+	/* Give up on iterating @cursor_ino if we can't iget it by this time. */
+	unsigned long		__iget_deadline;
+
+	/* Amount of time (in ms) that we will try to iget an inode. */
+	unsigned int		iget_timeout;
+
+	/* Wait this many ms to retry an iget. */
+	unsigned int		iget_retry_delay;
+
+	/*
+	 * The scan grabs batches of inodes and stashes them here before
+	 * handing them out with _iter.  Unallocated inodes are set in the
+	 * mask so that all updates to that inode are selected for live
+	 * update propagation.
+	 */
+	xfs_ino_t		__batch_ino;
+	xfs_inofree_t		__skipped_inomask;
+	struct xfs_inode	*__inodes[XFS_INODES_PER_CHUNK];
+};
+
+/* Set if the scan has been aborted due to some event in the fs. */
+#define XCHK_ISCAN_OPSTATE_ABORTED	(1)
+
+/* Use trylock to acquire the AGI */
+#define XCHK_ISCAN_OPSTATE_TRYLOCK_AGI	(2)
+
+static inline bool
+xchk_iscan_aborted(const struct xchk_iscan *iscan)
+{
+	return test_bit(XCHK_ISCAN_OPSTATE_ABORTED, &iscan->__opstate);
+}
+
+static inline void
+xchk_iscan_abort(struct xchk_iscan *iscan)
+{
+	set_bit(XCHK_ISCAN_OPSTATE_ABORTED, &iscan->__opstate);
+}
+
+static inline bool
+xchk_iscan_agi_needs_trylock(const struct xchk_iscan *iscan)
+{
+	return test_bit(XCHK_ISCAN_OPSTATE_TRYLOCK_AGI, &iscan->__opstate);
+}
+
+static inline void
+xchk_iscan_set_agi_trylock(struct xchk_iscan *iscan)
+{
+	set_bit(XCHK_ISCAN_OPSTATE_TRYLOCK_AGI, &iscan->__opstate);
+}
+
+void xchk_iscan_start(struct xfs_scrub *sc, unsigned int iget_timeout,
+		unsigned int iget_retry_delay, struct xchk_iscan *iscan);
+void xchk_iscan_finish_early(struct xchk_iscan *iscan);
+void xchk_iscan_teardown(struct xchk_iscan *iscan);
+
+int xchk_iscan_iter(struct xchk_iscan *iscan, struct xfs_inode **ipp);
+void xchk_iscan_iter_finish(struct xchk_iscan *iscan);
+
+void xchk_iscan_mark_visited(struct xchk_iscan *iscan, struct xfs_inode *ip);
+bool xchk_iscan_want_live_update(struct xchk_iscan *iscan, xfs_ino_t ino);
+
+#endif /* __XFS_SCRUB_ISCAN_H__ */
diff --git a/fs/xfs/scrub/listxattr.c b/fs/xfs/scrub/listxattr.c
new file mode 100644
index 0000000000..256ff7700c
--- /dev/null
+++ b/fs/xfs/scrub/listxattr.c
@@ -0,0 +1,320 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2022-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_inode.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
+#include "xfs_attr.h"
+#include "xfs_attr_leaf.h"
+#include "xfs_attr_sf.h"
+#include "xfs_trans.h"
+#include "scrub/scrub.h"
+#include "scrub/bitmap.h"
+#include "scrub/dab_bitmap.h"
+#include "scrub/listxattr.h"
+
+/* Call a function for every entry in a shortform xattr structure. */
+STATIC int
+xchk_xattr_walk_sf(
+	struct xfs_scrub		*sc,
+	struct xfs_inode		*ip,
+	xchk_xattr_fn			attr_fn,
+	void				*priv)
+{
+	struct xfs_attr_sf_hdr		*hdr = ip->i_af.if_data;
+	struct xfs_attr_sf_entry	*sfe;
+	unsigned int			i;
+	int				error;
+
+	sfe = xfs_attr_sf_firstentry(hdr);
+	for (i = 0; i < hdr->count; i++) {
+		error = attr_fn(sc, ip, sfe->flags, sfe->nameval, sfe->namelen,
+				&sfe->nameval[sfe->namelen], sfe->valuelen,
+				priv);
+		if (error)
+			return error;
+
+		sfe = xfs_attr_sf_nextentry(sfe);
+	}
+
+	return 0;
+}
+
+/* Call a function for every entry in this xattr leaf block. */
+STATIC int
+xchk_xattr_walk_leaf_entries(
+	struct xfs_scrub		*sc,
+	struct xfs_inode		*ip,
+	xchk_xattr_fn			attr_fn,
+	struct xfs_buf			*bp,
+	void				*priv)
+{
+	struct xfs_attr3_icleaf_hdr	ichdr;
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_attr_leafblock	*leaf = bp->b_addr;
+	struct xfs_attr_leaf_entry	*entry;
+	unsigned int			i;
+	int				error;
+
+	xfs_attr3_leaf_hdr_from_disk(mp->m_attr_geo, &ichdr, leaf);
+	entry = xfs_attr3_leaf_entryp(leaf);
+
+	for (i = 0; i < ichdr.count; entry++, i++) {
+		void			*value;
+		unsigned char		*name;
+		unsigned int		namelen, valuelen;
+
+		if (entry->flags & XFS_ATTR_LOCAL) {
+			struct xfs_attr_leaf_name_local		*name_loc;
+
+			name_loc = xfs_attr3_leaf_name_local(leaf, i);
+			name = name_loc->nameval;
+			namelen = name_loc->namelen;
+			value = &name_loc->nameval[name_loc->namelen];
+			valuelen = be16_to_cpu(name_loc->valuelen);
+		} else {
+			struct xfs_attr_leaf_name_remote	*name_rmt;
+
+			name_rmt = xfs_attr3_leaf_name_remote(leaf, i);
+			name = name_rmt->name;
+			namelen = name_rmt->namelen;
+			value = NULL;
+			valuelen = be32_to_cpu(name_rmt->valuelen);
+		}
+
+		error = attr_fn(sc, ip, entry->flags, name, namelen, value,
+				valuelen, priv);
+		if (error)
+			return error;
+
+	}
+
+	return 0;
+}
+
+/*
+ * Call a function for every entry in a leaf-format xattr structure.  Avoid
+ * memory allocations for the loop detector since there's only one block.
+ */
+STATIC int
+xchk_xattr_walk_leaf(
+	struct xfs_scrub		*sc,
+	struct xfs_inode		*ip,
+	xchk_xattr_fn			attr_fn,
+	void				*priv)
+{
+	struct xfs_buf			*leaf_bp;
+	int				error;
+
+	error = xfs_attr3_leaf_read(sc->tp, ip, ip->i_ino, 0, &leaf_bp);
+	if (error)
+		return error;
+
+	error = xchk_xattr_walk_leaf_entries(sc, ip, attr_fn, leaf_bp, priv);
+	xfs_trans_brelse(sc->tp, leaf_bp);
+	return error;
+}
+
+/* Find the leftmost leaf in the xattr dabtree. */
+STATIC int
+xchk_xattr_find_leftmost_leaf(
+	struct xfs_scrub		*sc,
+	struct xfs_inode		*ip,
+	struct xdab_bitmap		*seen_dablks,
+	struct xfs_buf			**leaf_bpp)
+{
+	struct xfs_da3_icnode_hdr	nodehdr;
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_trans		*tp = sc->tp;
+	struct xfs_da_intnode		*node;
+	struct xfs_da_node_entry	*btree;
+	struct xfs_buf			*bp;
+	xfs_failaddr_t			fa;
+	xfs_dablk_t			blkno = 0;
+	unsigned int			expected_level = 0;
+	int				error;
+
+	for (;;) {
+		xfs_extlen_t		len = 1;
+		uint16_t		magic;
+
+		/* Make sure we haven't seen this new block already. */
+		if (xdab_bitmap_test(seen_dablks, blkno, &len))
+			return -EFSCORRUPTED;
+
+		error = xfs_da3_node_read(tp, ip, blkno, &bp, XFS_ATTR_FORK);
+		if (error)
+			return error;
+
+		node = bp->b_addr;
+		magic = be16_to_cpu(node->hdr.info.magic);
+		if (magic == XFS_ATTR_LEAF_MAGIC ||
+		    magic == XFS_ATTR3_LEAF_MAGIC)
+			break;
+
+		error = -EFSCORRUPTED;
+		if (magic != XFS_DA_NODE_MAGIC &&
+		    magic != XFS_DA3_NODE_MAGIC)
+			goto out_buf;
+
+		fa = xfs_da3_node_header_check(bp, ip->i_ino);
+		if (fa)
+			goto out_buf;
+
+		xfs_da3_node_hdr_from_disk(mp, &nodehdr, node);
+
+		if (nodehdr.count == 0 || nodehdr.level >= XFS_DA_NODE_MAXDEPTH)
+			goto out_buf;
+
+		/* Check the level from the root node. */
+		if (blkno == 0)
+			expected_level = nodehdr.level - 1;
+		else if (expected_level != nodehdr.level)
+			goto out_buf;
+		else
+			expected_level--;
+
+		/* Remember that we've seen this node. */
+		error = xdab_bitmap_set(seen_dablks, blkno, 1);
+		if (error)
+			goto out_buf;
+
+		/* Find the next level towards the leaves of the dabtree. */
+		btree = nodehdr.btree;
+		blkno = be32_to_cpu(btree->before);
+		xfs_trans_brelse(tp, bp);
+	}
+
+	error = -EFSCORRUPTED;
+	fa = xfs_attr3_leaf_header_check(bp, ip->i_ino);
+	if (fa)
+		goto out_buf;
+
+	if (expected_level != 0)
+		goto out_buf;
+
+	/* Remember that we've seen this leaf. */
+	error = xdab_bitmap_set(seen_dablks, blkno, 1);
+	if (error)
+		goto out_buf;
+
+	*leaf_bpp = bp;
+	return 0;
+
+out_buf:
+	xfs_trans_brelse(tp, bp);
+	return error;
+}
+
+/* Call a function for every entry in a node-format xattr structure. */
+STATIC int
+xchk_xattr_walk_node(
+	struct xfs_scrub		*sc,
+	struct xfs_inode		*ip,
+	xchk_xattr_fn			attr_fn,
+	xchk_xattrleaf_fn		leaf_fn,
+	void				*priv)
+{
+	struct xfs_attr3_icleaf_hdr	leafhdr;
+	struct xdab_bitmap		seen_dablks;
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_attr_leafblock	*leaf;
+	struct xfs_buf			*leaf_bp;
+	int				error;
+
+	xdab_bitmap_init(&seen_dablks);
+
+	error = xchk_xattr_find_leftmost_leaf(sc, ip, &seen_dablks, &leaf_bp);
+	if (error)
+		goto out_bitmap;
+
+	for (;;) {
+		xfs_extlen_t	len;
+
+		error = xchk_xattr_walk_leaf_entries(sc, ip, attr_fn, leaf_bp,
+				priv);
+		if (error)
+			goto out_leaf;
+
+		/* Find the right sibling of this leaf block. */
+		leaf = leaf_bp->b_addr;
+		xfs_attr3_leaf_hdr_from_disk(mp->m_attr_geo, &leafhdr, leaf);
+		if (leafhdr.forw == 0)
+			goto out_leaf;
+
+		xfs_trans_brelse(sc->tp, leaf_bp);
+
+		if (leaf_fn) {
+			error = leaf_fn(sc, priv);
+			if (error)
+				goto out_bitmap;
+		}
+
+		/* Make sure we haven't seen this new leaf already. */
+		len = 1;
+		if (xdab_bitmap_test(&seen_dablks, leafhdr.forw, &len)) {
+			error = -EFSCORRUPTED;
+			goto out_bitmap;
+		}
+
+		error = xfs_attr3_leaf_read(sc->tp, ip, ip->i_ino,
+				leafhdr.forw, &leaf_bp);
+		if (error)
+			goto out_bitmap;
+
+		/* Remember that we've seen this new leaf. */
+		error = xdab_bitmap_set(&seen_dablks, leafhdr.forw, 1);
+		if (error)
+			goto out_leaf;
+	}
+
+out_leaf:
+	xfs_trans_brelse(sc->tp, leaf_bp);
+out_bitmap:
+	xdab_bitmap_destroy(&seen_dablks);
+	return error;
+}
+
+/*
+ * Call a function for every extended attribute in a file.
+ *
+ * Callers must hold the ILOCK.  No validation or cursor restarts allowed.
+ * Returns -EFSCORRUPTED on any problem, including loops in the dabtree.
+ */
+int
+xchk_xattr_walk(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*ip,
+	xchk_xattr_fn		attr_fn,
+	xchk_xattrleaf_fn	leaf_fn,
+	void			*priv)
+{
+	int			error;
+
+	xfs_assert_ilocked(ip, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL);
+
+	if (!xfs_inode_hasattr(ip))
+		return 0;
+
+	if (ip->i_af.if_format == XFS_DINODE_FMT_LOCAL)
+		return xchk_xattr_walk_sf(sc, ip, attr_fn, priv);
+
+	/* attr functions require that the attr fork is loaded */
+	error = xfs_iread_extents(sc->tp, ip, XFS_ATTR_FORK);
+	if (error)
+		return error;
+
+	if (xfs_attr_is_leaf(ip))
+		return xchk_xattr_walk_leaf(sc, ip, attr_fn, priv);
+
+	return xchk_xattr_walk_node(sc, ip, attr_fn, leaf_fn, priv);
+}
diff --git a/fs/xfs/scrub/listxattr.h b/fs/xfs/scrub/listxattr.h
new file mode 100644
index 0000000000..703cfb7b14
--- /dev/null
+++ b/fs/xfs/scrub/listxattr.h
@@ -0,0 +1,19 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (c) 2022-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_LISTXATTR_H__
+#define __XFS_SCRUB_LISTXATTR_H__
+
+typedef int (*xchk_xattr_fn)(struct xfs_scrub *sc, struct xfs_inode *ip,
+		unsigned int attr_flags, const unsigned char *name,
+		unsigned int namelen, const void *value, unsigned int valuelen,
+		void *priv);
+
+typedef int (*xchk_xattrleaf_fn)(struct xfs_scrub *sc, void *priv);
+
+int xchk_xattr_walk(struct xfs_scrub *sc, struct xfs_inode *ip,
+		xchk_xattr_fn attr_fn, xchk_xattrleaf_fn leaf_fn, void *priv);
+
+#endif /* __XFS_SCRUB_LISTXATTR_H__ */
diff --git a/fs/xfs/scrub/newbt.c b/fs/xfs/scrub/newbt.c
index bb6d980b4f..4a0271123d 100644
--- a/fs/xfs/scrub/newbt.c
+++ b/fs/xfs/scrub/newbt.c
@@ -239,7 +239,11 @@ xrep_newbt_alloc_ag_blocks(
 
 		xrep_newbt_validate_ag_alloc_hint(xnr);
 
-		error = xfs_alloc_vextent_near_bno(&args, xnr->alloc_hint);
+		if (xnr->alloc_vextent)
+			error = xnr->alloc_vextent(sc, &args, xnr->alloc_hint);
+		else
+			error = xfs_alloc_vextent_near_bno(&args,
+					xnr->alloc_hint);
 		if (error)
 			return error;
 		if (args.fsbno == NULLFSBLOCK)
@@ -309,7 +313,11 @@ xrep_newbt_alloc_file_blocks(
 
 		xrep_newbt_validate_file_alloc_hint(xnr);
 
-		error = xfs_alloc_vextent_start_ag(&args, xnr->alloc_hint);
+		if (xnr->alloc_vextent)
+			error = xnr->alloc_vextent(sc, &args, xnr->alloc_hint);
+		else
+			error = xfs_alloc_vextent_start_ag(&args,
+					xnr->alloc_hint);
 		if (error)
 			return error;
 		if (args.fsbno == NULLFSBLOCK)
@@ -535,7 +543,7 @@ xrep_newbt_claim_block(
 	trace_xrep_newbt_claim_block(mp, resv->pag->pag_agno, agbno, 1,
 			xnr->oinfo.oi_owner);
 
-	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+	if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN)
 		ptr->l = cpu_to_be64(XFS_AGB_TO_FSB(mp, resv->pag->pag_agno,
 								agbno));
 	else
diff --git a/fs/xfs/scrub/newbt.h b/fs/xfs/scrub/newbt.h
index 89f8e3970b..3d804d31af 100644
--- a/fs/xfs/scrub/newbt.h
+++ b/fs/xfs/scrub/newbt.h
@@ -6,6 +6,8 @@
 #ifndef __XFS_SCRUB_NEWBT_H__
 #define __XFS_SCRUB_NEWBT_H__
 
+struct xfs_alloc_arg;
+
 struct xrep_newbt_resv {
 	/* Link to list of extents that we've reserved. */
 	struct list_head	list;
@@ -28,6 +30,11 @@ struct xrep_newbt_resv {
 struct xrep_newbt {
 	struct xfs_scrub	*sc;
 
+	/* Custom allocation function, or NULL for xfs_alloc_vextent */
+	int			(*alloc_vextent)(struct xfs_scrub *sc,
+						 struct xfs_alloc_arg *args,
+						 xfs_fsblock_t alloc_hint);
+
 	/* List of extents that we've reserved. */
 	struct list_head	resv_list;
 
diff --git a/fs/xfs/scrub/nlinks.c b/fs/xfs/scrub/nlinks.c
new file mode 100644
index 0000000000..80aee30886
--- /dev/null
+++ b/fs/xfs/scrub/nlinks.c
@@ -0,0 +1,1049 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2021-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_inode.h"
+#include "xfs_icache.h"
+#include "xfs_iwalk.h"
+#include "xfs_ialloc.h"
+#include "xfs_dir2.h"
+#include "xfs_dir2_priv.h"
+#include "xfs_ag.h"
+#include "xfs_parent.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/repair.h"
+#include "scrub/xfile.h"
+#include "scrub/xfarray.h"
+#include "scrub/iscan.h"
+#include "scrub/orphanage.h"
+#include "scrub/nlinks.h"
+#include "scrub/trace.h"
+#include "scrub/readdir.h"
+#include "scrub/tempfile.h"
+#include "scrub/listxattr.h"
+
+/*
+ * Live Inode Link Count Checking
+ * ==============================
+ *
+ * Inode link counts are "summary" metadata, in the sense that they are
+ * computed as the number of directory entries referencing each file on the
+ * filesystem.  Therefore, we compute the correct link counts by creating a
+ * shadow link count structure and walking every inode.
+ */
+
+/* Set us up to scrub inode link counts. */
+int
+xchk_setup_nlinks(
+	struct xfs_scrub	*sc)
+{
+	struct xchk_nlink_ctrs	*xnc;
+	int			error;
+
+	xchk_fsgates_enable(sc, XCHK_FSGATES_DIRENTS);
+
+	if (xchk_could_repair(sc)) {
+		error = xrep_setup_nlinks(sc);
+		if (error)
+			return error;
+	}
+
+	xnc = kvzalloc(sizeof(struct xchk_nlink_ctrs), XCHK_GFP_FLAGS);
+	if (!xnc)
+		return -ENOMEM;
+	xnc->xname.name = xnc->namebuf;
+	xnc->sc = sc;
+	sc->buf = xnc;
+
+	return xchk_setup_fs(sc);
+}
+
+/*
+ * Part 1: Collecting file link counts.  For each file, we create a shadow link
+ * counting structure, then walk the entire directory tree, incrementing parent
+ * and child link counts for each directory entry seen.
+ *
+ * To avoid false corruption reports in part 2, any failure in this part must
+ * set the INCOMPLETE flag even when a negative errno is returned.  This care
+ * must be taken with certain errno values (i.e. EFSBADCRC, EFSCORRUPTED,
+ * ECANCELED) that are absorbed into a scrub state flag update by
+ * xchk_*_process_error.  Scrub and repair share the same incore data
+ * structures, so the INCOMPLETE flag is critical to prevent a repair based on
+ * insufficient information.
+ *
+ * Because we are scanning a live filesystem, it's possible that another thread
+ * will try to update the link counts for an inode that we've already scanned.
+ * This will cause our counts to be incorrect.  Therefore, we hook all
+ * directory entry updates because that is when link count updates occur.  By
+ * shadowing transaction updates in this manner, live nlink check can ensure by
+ * locking the inode and the shadow structure that its own copies are not out
+ * of date.  Because the hook code runs in a different process context from the
+ * scrub code and the scrub state flags are not accessed atomically, failures
+ * in the hook code must abort the iscan and the scrubber must notice the
+ * aborted scan and set the incomplete flag.
+ *
+ * Note that we use jump labels and srcu notifier hooks to minimize the
+ * overhead when live nlinks is /not/ running.  Locking order for nlink
+ * observations is inode ILOCK -> iscan_lock/xchk_nlink_ctrs lock.
+ */
+
+/*
+ * Add a delta to an nlink counter, clamping the value to U32_MAX.  Because
+ * XFS_MAXLINK < U32_MAX, the checking code will produce the correct results
+ * even if we lose some precision.
+ */
+static inline void
+careful_add(
+	xfs_nlink_t	*nlinkp,
+	int		delta)
+{
+	uint64_t	new_value = (uint64_t)(*nlinkp) + delta;
+
+	BUILD_BUG_ON(XFS_MAXLINK > U32_MAX);
+	*nlinkp = min_t(uint64_t, new_value, U32_MAX);
+}
+
+/* Update incore link count information.  Caller must hold the nlinks lock. */
+STATIC int
+xchk_nlinks_update_incore(
+	struct xchk_nlink_ctrs	*xnc,
+	xfs_ino_t		ino,
+	int			parents_delta,
+	int			backrefs_delta,
+	int			children_delta)
+{
+	struct xchk_nlink	nl;
+	int			error;
+
+	if (!xnc->nlinks)
+		return 0;
+
+	error = xfarray_load_sparse(xnc->nlinks, ino, &nl);
+	if (error)
+		return error;
+
+	trace_xchk_nlinks_update_incore(xnc->sc->mp, ino, &nl, parents_delta,
+			backrefs_delta, children_delta);
+
+	careful_add(&nl.parents, parents_delta);
+	careful_add(&nl.backrefs, backrefs_delta);
+	careful_add(&nl.children, children_delta);
+
+	nl.flags |= XCHK_NLINK_WRITTEN;
+	error = xfarray_store(xnc->nlinks, ino, &nl);
+	if (error == -EFBIG) {
+		/*
+		 * EFBIG means we tried to store data at too high a byte offset
+		 * in the sparse array.  IOWs, we cannot complete the check and
+		 * must notify userspace that the check was incomplete.
+		 */
+		error = -ECANCELED;
+	}
+	return error;
+}
+
+/*
+ * Apply a link count change from the regular filesystem into our shadow link
+ * count structure based on a directory update in progress.
+ */
+STATIC int
+xchk_nlinks_live_update(
+	struct notifier_block		*nb,
+	unsigned long			action,
+	void				*data)
+{
+	struct xfs_dir_update_params	*p = data;
+	struct xchk_nlink_ctrs		*xnc;
+	int				error;
+
+	xnc = container_of(nb, struct xchk_nlink_ctrs, dhook.dirent_hook.nb);
+
+	/*
+	 * Ignore temporary directories being used to stage dir repairs, since
+	 * we don't bump the link counts of the children.
+	 */
+	if (xrep_is_tempfile(p->dp))
+		return NOTIFY_DONE;
+
+	trace_xchk_nlinks_live_update(xnc->sc->mp, p->dp, action, p->ip->i_ino,
+			p->delta, p->name->name, p->name->len);
+
+	/*
+	 * If we've already scanned @dp, update the number of parents that link
+	 * to @ip.  If @ip is a subdirectory, update the number of child links
+	 * going out of @dp.
+	 */
+	if (xchk_iscan_want_live_update(&xnc->collect_iscan, p->dp->i_ino)) {
+		mutex_lock(&xnc->lock);
+		error = xchk_nlinks_update_incore(xnc, p->ip->i_ino, p->delta,
+				0, 0);
+		if (!error && S_ISDIR(VFS_IC(p->ip)->i_mode))
+			error = xchk_nlinks_update_incore(xnc, p->dp->i_ino, 0,
+					0, p->delta);
+		mutex_unlock(&xnc->lock);
+		if (error)
+			goto out_abort;
+	}
+
+	/*
+	 * If @ip is a subdirectory and we've already scanned it, update the
+	 * number of backrefs pointing to @dp.
+	 */
+	if (S_ISDIR(VFS_IC(p->ip)->i_mode) &&
+	    xchk_iscan_want_live_update(&xnc->collect_iscan, p->ip->i_ino)) {
+		mutex_lock(&xnc->lock);
+		error = xchk_nlinks_update_incore(xnc, p->dp->i_ino, 0,
+				p->delta, 0);
+		mutex_unlock(&xnc->lock);
+		if (error)
+			goto out_abort;
+	}
+
+	return NOTIFY_DONE;
+
+out_abort:
+	xchk_iscan_abort(&xnc->collect_iscan);
+	return NOTIFY_DONE;
+}
+
+/* Bump the observed link count for the inode referenced by this entry. */
+STATIC int
+xchk_nlinks_collect_dirent(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*dp,
+	xfs_dir2_dataptr_t	dapos,
+	const struct xfs_name	*name,
+	xfs_ino_t		ino,
+	void			*priv)
+{
+	struct xchk_nlink_ctrs	*xnc = priv;
+	bool			dot = false, dotdot = false;
+	int			error;
+
+	/* Does this name make sense? */
+	if (name->len == 0 || !xfs_dir2_namecheck(name->name, name->len)) {
+		error = -ECANCELED;
+		goto out_abort;
+	}
+
+	if (name->len == 1 && name->name[0] == '.')
+		dot = true;
+	else if (name->len == 2 && name->name[0] == '.' &&
+				   name->name[1] == '.')
+		dotdot = true;
+
+	/* Don't accept a '.' entry that points somewhere else. */
+	if (dot && ino != dp->i_ino) {
+		error = -ECANCELED;
+		goto out_abort;
+	}
+
+	/* Don't accept an invalid inode number. */
+	if (!xfs_verify_dir_ino(sc->mp, ino)) {
+		error = -ECANCELED;
+		goto out_abort;
+	}
+
+	/* Update the shadow link counts if we haven't already failed. */
+
+	if (xchk_iscan_aborted(&xnc->collect_iscan)) {
+		error = -ECANCELED;
+		goto out_incomplete;
+	}
+
+	trace_xchk_nlinks_collect_dirent(sc->mp, dp, ino, name);
+
+	mutex_lock(&xnc->lock);
+
+	/*
+	 * If this is a dotdot entry, it is a back link from dp to ino.  How
+	 * we handle this depends on whether or not dp is the root directory.
+	 *
+	 * The root directory is its own parent, so we pretend the dotdot entry
+	 * establishes the "parent" of the root directory.  Increment the
+	 * number of parents of the root directory.
+	 *
+	 * Otherwise, increment the number of backrefs pointing back to ino.
+	 *
+	 * If the filesystem has parent pointers, we walk the pptrs to
+	 * determine the backref count.
+	 */
+	if (dotdot) {
+		if (dp == sc->mp->m_rootip)
+			error = xchk_nlinks_update_incore(xnc, ino, 1, 0, 0);
+		else if (!xfs_has_parent(sc->mp))
+			error = xchk_nlinks_update_incore(xnc, ino, 0, 1, 0);
+		else
+			error = 0;
+		if (error)
+			goto out_unlock;
+	}
+
+	/*
+	 * If this dirent is a forward link from dp to ino, increment the
+	 * number of parents linking into ino.
+	 */
+	if (!dot && !dotdot) {
+		error = xchk_nlinks_update_incore(xnc, ino, 1, 0, 0);
+		if (error)
+			goto out_unlock;
+	}
+
+	/*
+	 * If this dirent is a forward link to a subdirectory, increment the
+	 * number of child links of dp.
+	 */
+	if (!dot && !dotdot && name->type == XFS_DIR3_FT_DIR) {
+		error = xchk_nlinks_update_incore(xnc, dp->i_ino, 0, 0, 1);
+		if (error)
+			goto out_unlock;
+	}
+
+	mutex_unlock(&xnc->lock);
+	return 0;
+
+out_unlock:
+	mutex_unlock(&xnc->lock);
+out_abort:
+	xchk_iscan_abort(&xnc->collect_iscan);
+out_incomplete:
+	xchk_set_incomplete(sc);
+	return error;
+}
+
+/* Bump the backref count for the inode referenced by this parent pointer. */
+STATIC int
+xchk_nlinks_collect_pptr(
+	struct xfs_scrub		*sc,
+	struct xfs_inode		*ip,
+	unsigned int			attr_flags,
+	const unsigned char		*name,
+	unsigned int			namelen,
+	const void			*value,
+	unsigned int			valuelen,
+	void				*priv)
+{
+	struct xfs_name			xname = {
+		.name			= name,
+		.len			= namelen,
+	};
+	struct xchk_nlink_ctrs		*xnc = priv;
+	const struct xfs_parent_rec	*pptr_rec = value;
+	xfs_ino_t			parent_ino;
+	int				error;
+
+	/* Update the shadow link counts if we haven't already failed. */
+
+	if (xchk_iscan_aborted(&xnc->collect_iscan)) {
+		error = -ECANCELED;
+		goto out_incomplete;
+	}
+
+	if (!(attr_flags & XFS_ATTR_PARENT))
+		return 0;
+
+	error = xfs_parent_from_attr(sc->mp, attr_flags, name, namelen, value,
+			valuelen, &parent_ino, NULL);
+	if (error)
+		return error;
+
+	trace_xchk_nlinks_collect_pptr(sc->mp, ip, &xname, pptr_rec);
+
+	mutex_lock(&xnc->lock);
+
+	error = xchk_nlinks_update_incore(xnc, parent_ino, 0, 1, 0);
+	if (error)
+		goto out_unlock;
+
+	mutex_unlock(&xnc->lock);
+	return 0;
+
+out_unlock:
+	mutex_unlock(&xnc->lock);
+	xchk_iscan_abort(&xnc->collect_iscan);
+out_incomplete:
+	xchk_set_incomplete(sc);
+	return error;
+}
+
+/* Walk a directory to bump the observed link counts of the children. */
+STATIC int
+xchk_nlinks_collect_dir(
+	struct xchk_nlink_ctrs	*xnc,
+	struct xfs_inode	*dp)
+{
+	struct xfs_scrub	*sc = xnc->sc;
+	unsigned int		lock_mode;
+	int			error = 0;
+
+	/*
+	 * Ignore temporary directories being used to stage dir repairs, since
+	 * we don't bump the link counts of the children.
+	 */
+	if (xrep_is_tempfile(dp))
+		return 0;
+
+	/* Prevent anyone from changing this directory while we walk it. */
+	xfs_ilock(dp, XFS_IOLOCK_SHARED);
+	lock_mode = xfs_ilock_data_map_shared(dp);
+
+	/*
+	 * The dotdot entry of an unlinked directory still points to the last
+	 * parent, but the parent no longer links to this directory.  Skip the
+	 * directory to avoid overcounting.
+	 */
+	if (VFS_I(dp)->i_nlink == 0)
+		goto out_unlock;
+
+	/*
+	 * We cannot count file links if the directory looks as though it has
+	 * been zapped by the inode record repair code.
+	 */
+	if (xchk_dir_looks_zapped(dp)) {
+		error = -EBUSY;
+		goto out_abort;
+	}
+
+	error = xchk_dir_walk(sc, dp, xchk_nlinks_collect_dirent, xnc);
+	if (error == -ECANCELED) {
+		error = 0;
+		goto out_unlock;
+	}
+	if (error)
+		goto out_abort;
+
+	/* Walk the parent pointers to get real backref counts. */
+	if (xfs_has_parent(sc->mp)) {
+		/*
+		 * If the extended attributes look as though they has been
+		 * zapped by the inode record repair code, we cannot scan for
+		 * parent pointers.
+		 */
+		if (xchk_pptr_looks_zapped(dp)) {
+			error = -EBUSY;
+			goto out_unlock;
+		}
+
+		error = xchk_xattr_walk(sc, dp, xchk_nlinks_collect_pptr, NULL,
+				xnc);
+		if (error == -ECANCELED) {
+			error = 0;
+			goto out_unlock;
+		}
+		if (error)
+			goto out_abort;
+	}
+
+	xchk_iscan_mark_visited(&xnc->collect_iscan, dp);
+	goto out_unlock;
+
+out_abort:
+	xchk_set_incomplete(sc);
+	xchk_iscan_abort(&xnc->collect_iscan);
+out_unlock:
+	xfs_iunlock(dp, lock_mode);
+	xfs_iunlock(dp, XFS_IOLOCK_SHARED);
+	return error;
+}
+
+/* If this looks like a valid pointer, count it. */
+static inline int
+xchk_nlinks_collect_metafile(
+	struct xchk_nlink_ctrs	*xnc,
+	xfs_ino_t		ino)
+{
+	if (!xfs_verify_ino(xnc->sc->mp, ino))
+		return 0;
+
+	trace_xchk_nlinks_collect_metafile(xnc->sc->mp, ino);
+	return xchk_nlinks_update_incore(xnc, ino, 1, 0, 0);
+}
+
+/* Bump the link counts of metadata files rooted in the superblock. */
+STATIC int
+xchk_nlinks_collect_metafiles(
+	struct xchk_nlink_ctrs	*xnc)
+{
+	struct xfs_mount	*mp = xnc->sc->mp;
+	int			error = -ECANCELED;
+
+
+	if (xchk_iscan_aborted(&xnc->collect_iscan))
+		goto out_incomplete;
+
+	mutex_lock(&xnc->lock);
+	error = xchk_nlinks_collect_metafile(xnc, mp->m_sb.sb_rbmino);
+	if (error)
+		goto out_abort;
+
+	error = xchk_nlinks_collect_metafile(xnc, mp->m_sb.sb_rsumino);
+	if (error)
+		goto out_abort;
+
+	error = xchk_nlinks_collect_metafile(xnc, mp->m_sb.sb_uquotino);
+	if (error)
+		goto out_abort;
+
+	error = xchk_nlinks_collect_metafile(xnc, mp->m_sb.sb_gquotino);
+	if (error)
+		goto out_abort;
+
+	error = xchk_nlinks_collect_metafile(xnc, mp->m_sb.sb_pquotino);
+	if (error)
+		goto out_abort;
+	mutex_unlock(&xnc->lock);
+
+	return 0;
+
+out_abort:
+	mutex_unlock(&xnc->lock);
+	xchk_iscan_abort(&xnc->collect_iscan);
+out_incomplete:
+	xchk_set_incomplete(xnc->sc);
+	return error;
+}
+
+/* Advance the collection scan cursor for this non-directory file. */
+static inline int
+xchk_nlinks_collect_file(
+	struct xchk_nlink_ctrs	*xnc,
+	struct xfs_inode	*ip)
+{
+	xfs_ilock(ip, XFS_IOLOCK_SHARED);
+	xchk_iscan_mark_visited(&xnc->collect_iscan, ip);
+	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
+	return 0;
+}
+
+/* Walk all directories and count inode links. */
+STATIC int
+xchk_nlinks_collect(
+	struct xchk_nlink_ctrs	*xnc)
+{
+	struct xfs_scrub	*sc = xnc->sc;
+	struct xfs_inode	*ip;
+	int			error;
+
+	/* Count the rt and quota files that are rooted in the superblock. */
+	error = xchk_nlinks_collect_metafiles(xnc);
+	if (error)
+		return error;
+
+	/*
+	 * Set up for a potentially lengthy filesystem scan by reducing our
+	 * transaction resource usage for the duration.  Specifically:
+	 *
+	 * Cancel the transaction to release the log grant space while we scan
+	 * the filesystem.
+	 *
+	 * Create a new empty transaction to eliminate the possibility of the
+	 * inode scan deadlocking on cyclical metadata.
+	 *
+	 * We pass the empty transaction to the file scanning function to avoid
+	 * repeatedly cycling empty transactions.  This can be done even though
+	 * we take the IOLOCK to quiesce the file because empty transactions
+	 * do not take sb_internal.
+	 */
+	xchk_trans_cancel(sc);
+	error = xchk_trans_alloc_empty(sc);
+	if (error)
+		return error;
+
+	while ((error = xchk_iscan_iter(&xnc->collect_iscan, &ip)) == 1) {
+		if (S_ISDIR(VFS_I(ip)->i_mode))
+			error = xchk_nlinks_collect_dir(xnc, ip);
+		else
+			error = xchk_nlinks_collect_file(xnc, ip);
+		xchk_irele(sc, ip);
+		if (error)
+			break;
+
+		if (xchk_should_terminate(sc, &error))
+			break;
+	}
+	xchk_iscan_iter_finish(&xnc->collect_iscan);
+	if (error) {
+		xchk_set_incomplete(sc);
+		/*
+		 * If we couldn't grab an inode that was busy with a state
+		 * change, change the error code so that we exit to userspace
+		 * as quickly as possible.
+		 */
+		if (error == -EBUSY)
+			return -ECANCELED;
+		return error;
+	}
+
+	/*
+	 * Switch out for a real transaction in preparation for building a new
+	 * tree.
+	 */
+	xchk_trans_cancel(sc);
+	return xchk_setup_fs(sc);
+}
+
+/*
+ * Part 2: Comparing file link counters.  Walk each inode and compare the link
+ * counts against our shadow information; and then walk each shadow link count
+ * structure (that wasn't covered in the first part), comparing it against the
+ * file.
+ */
+
+/* Read the observed link count for comparison with the actual inode. */
+STATIC int
+xchk_nlinks_comparison_read(
+	struct xchk_nlink_ctrs	*xnc,
+	xfs_ino_t		ino,
+	struct xchk_nlink	*obs)
+{
+	struct xchk_nlink	nl;
+	int			error;
+
+	error = xfarray_load_sparse(xnc->nlinks, ino, &nl);
+	if (error)
+		return error;
+
+	nl.flags |= (XCHK_NLINK_COMPARE_SCANNED | XCHK_NLINK_WRITTEN);
+
+	error = xfarray_store(xnc->nlinks, ino, &nl);
+	if (error == -EFBIG) {
+		/*
+		 * EFBIG means we tried to store data at too high a byte offset
+		 * in the sparse array.  IOWs, we cannot complete the check and
+		 * must notify userspace that the check was incomplete.  This
+		 * shouldn't really happen outside of the collection phase.
+		 */
+		xchk_set_incomplete(xnc->sc);
+		return -ECANCELED;
+	}
+	if (error)
+		return error;
+
+	/* Copy the counters, but do not expose the internal state. */
+	obs->parents = nl.parents;
+	obs->backrefs = nl.backrefs;
+	obs->children = nl.children;
+	obs->flags = 0;
+	return 0;
+}
+
+/* Check our link count against an inode. */
+STATIC int
+xchk_nlinks_compare_inode(
+	struct xchk_nlink_ctrs	*xnc,
+	struct xfs_inode	*ip)
+{
+	struct xchk_nlink	obs;
+	struct xfs_scrub	*sc = xnc->sc;
+	uint64_t		total_links;
+	unsigned int		actual_nlink;
+	int			error;
+
+	/*
+	 * Ignore temporary files being used to stage repairs, since we assume
+	 * they're correct for non-directories, and the directory repair code
+	 * doesn't bump the link counts for the children.
+	 */
+	if (xrep_is_tempfile(ip))
+		return 0;
+
+	xfs_ilock(ip, XFS_ILOCK_SHARED);
+	mutex_lock(&xnc->lock);
+
+	if (xchk_iscan_aborted(&xnc->collect_iscan)) {
+		xchk_set_incomplete(xnc->sc);
+		error = -ECANCELED;
+		goto out_scanlock;
+	}
+
+	error = xchk_nlinks_comparison_read(xnc, ip->i_ino, &obs);
+	if (error)
+		goto out_scanlock;
+
+	/*
+	 * If we don't have ftype to get an accurate count of the subdirectory
+	 * entries in this directory, take advantage of the fact that on a
+	 * consistent ftype=0 filesystem, the number of subdirectory
+	 * backreferences (dotdot entries) pointing towards this directory
+	 * should be equal to the number of subdirectory entries in the
+	 * directory.
+	 */
+	if (!xfs_has_ftype(sc->mp) && S_ISDIR(VFS_I(ip)->i_mode))
+		obs.children = obs.backrefs;
+
+	total_links = xchk_nlink_total(ip, &obs);
+	actual_nlink = VFS_I(ip)->i_nlink;
+
+	trace_xchk_nlinks_compare_inode(sc->mp, ip, &obs);
+
+	/*
+	 * If we found so many parents that we'd overflow i_nlink, we must flag
+	 * this as a corruption.  The VFS won't let users increase the link
+	 * count, but it will let them decrease it.
+	 */
+	if (total_links > XFS_NLINK_PINNED) {
+		xchk_ino_set_corrupt(sc, ip->i_ino);
+		goto out_corrupt;
+	} else if (total_links > XFS_MAXLINK) {
+		xchk_ino_set_warning(sc, ip->i_ino);
+	}
+
+	/* Link counts should match. */
+	if (total_links != actual_nlink) {
+		xchk_ino_set_corrupt(sc, ip->i_ino);
+		goto out_corrupt;
+	}
+
+	if (S_ISDIR(VFS_I(ip)->i_mode) && actual_nlink > 0) {
+		/*
+		 * The collection phase ignores directories with zero link
+		 * count, so we ignore them here too.
+		 *
+		 * The number of subdirectory backreferences (dotdot entries)
+		 * pointing towards this directory should be equal to the
+		 * number of subdirectory entries in the directory.
+		 */
+		if (obs.children != obs.backrefs)
+			xchk_ino_xref_set_corrupt(sc, ip->i_ino);
+	} else {
+		/*
+		 * Non-directories and unlinked directories should not have
+		 * back references.
+		 */
+		if (obs.backrefs != 0) {
+			xchk_ino_set_corrupt(sc, ip->i_ino);
+			goto out_corrupt;
+		}
+
+		/*
+		 * Non-directories and unlinked directories should not have
+		 * children.
+		 */
+		if (obs.children != 0) {
+			xchk_ino_set_corrupt(sc, ip->i_ino);
+			goto out_corrupt;
+		}
+	}
+
+	if (ip == sc->mp->m_rootip) {
+		/*
+		 * For the root of a directory tree, both the '.' and '..'
+		 * entries should point to the root directory.  The dotdot
+		 * entry is counted as a parent of the root /and/ a backref of
+		 * the root directory.
+		 */
+		if (obs.parents != 1) {
+			xchk_ino_set_corrupt(sc, ip->i_ino);
+			goto out_corrupt;
+		}
+	} else if (actual_nlink > 0) {
+		/*
+		 * Linked files that are not the root directory should have at
+		 * least one parent.
+		 */
+		if (obs.parents == 0) {
+			xchk_ino_set_corrupt(sc, ip->i_ino);
+			goto out_corrupt;
+		}
+	}
+
+out_corrupt:
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		error = -ECANCELED;
+out_scanlock:
+	mutex_unlock(&xnc->lock);
+	xfs_iunlock(ip, XFS_ILOCK_SHARED);
+	return error;
+}
+
+/*
+ * Check our link count against an inode that wasn't checked previously.  This
+ * is intended to catch directories with dangling links, though we could be
+ * racing with inode allocation in other threads.
+ */
+STATIC int
+xchk_nlinks_compare_inum(
+	struct xchk_nlink_ctrs	*xnc,
+	xfs_ino_t		ino)
+{
+	struct xchk_nlink	obs;
+	struct xfs_mount	*mp = xnc->sc->mp;
+	struct xfs_trans	*tp = xnc->sc->tp;
+	struct xfs_buf		*agi_bp;
+	struct xfs_inode	*ip;
+	int			error;
+
+	/*
+	 * The first iget failed, so try again with the variant that returns
+	 * either an incore inode or the AGI buffer.  If the function returns
+	 * EINVAL/ENOENT, it should have passed us the AGI buffer so that we
+	 * can guarantee that the inode won't be allocated while we check for
+	 * a zero link count in the observed link count data.
+	 */
+	error = xchk_iget_agi(xnc->sc, ino, &agi_bp, &ip);
+	if (!error) {
+		/* Actually got an inode, so use the inode compare. */
+		error = xchk_nlinks_compare_inode(xnc, ip);
+		xchk_irele(xnc->sc, ip);
+		return error;
+	}
+	if (error == -ENOENT || error == -EINVAL) {
+		/* No inode was found.  Check for zero link count below. */
+		error = 0;
+	}
+	if (error)
+		goto out_agi;
+
+	/* Ensure that we have protected against inode allocation/freeing. */
+	if (agi_bp == NULL) {
+		ASSERT(agi_bp != NULL);
+		xchk_set_incomplete(xnc->sc);
+		return -ECANCELED;
+	}
+
+	if (xchk_iscan_aborted(&xnc->collect_iscan)) {
+		xchk_set_incomplete(xnc->sc);
+		error = -ECANCELED;
+		goto out_agi;
+	}
+
+	mutex_lock(&xnc->lock);
+	error = xchk_nlinks_comparison_read(xnc, ino, &obs);
+	if (error)
+		goto out_scanlock;
+
+	trace_xchk_nlinks_check_zero(mp, ino, &obs);
+
+	/*
+	 * If we can't grab the inode, the link count had better be zero.  We
+	 * still hold the AGI to prevent inode allocation/freeing.
+	 */
+	if (xchk_nlink_total(NULL, &obs) != 0) {
+		xchk_ino_set_corrupt(xnc->sc, ino);
+		error = -ECANCELED;
+	}
+
+out_scanlock:
+	mutex_unlock(&xnc->lock);
+out_agi:
+	if (agi_bp)
+		xfs_trans_brelse(tp, agi_bp);
+	return error;
+}
+
+/*
+ * Try to visit every inode in the filesystem to compare the link count.  Move
+ * on if we can't grab an inode, since we'll revisit unchecked nlink records in
+ * the second part.
+ */
+static int
+xchk_nlinks_compare_iter(
+	struct xchk_nlink_ctrs	*xnc,
+	struct xfs_inode	**ipp)
+{
+	int			error;
+
+	do {
+		error = xchk_iscan_iter(&xnc->compare_iscan, ipp);
+	} while (error == -EBUSY);
+
+	return error;
+}
+
+/* Compare the link counts we observed against the live information. */
+STATIC int
+xchk_nlinks_compare(
+	struct xchk_nlink_ctrs	*xnc)
+{
+	struct xchk_nlink	nl;
+	struct xfs_scrub	*sc = xnc->sc;
+	struct xfs_inode	*ip;
+	xfarray_idx_t		cur = XFARRAY_CURSOR_INIT;
+	int			error;
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return 0;
+
+	/*
+	 * Create a new empty transaction so that we can advance the iscan
+	 * cursor without deadlocking if the inobt has a cycle and push on the
+	 * inactivation workqueue.
+	 */
+	xchk_trans_cancel(sc);
+	error = xchk_trans_alloc_empty(sc);
+	if (error)
+		return error;
+
+	/*
+	 * Use the inobt to walk all allocated inodes to compare the link
+	 * counts.  Inodes skipped by _compare_iter will be tried again in the
+	 * next phase of the scan.
+	 */
+	xchk_iscan_start(sc, 0, 0, &xnc->compare_iscan);
+	while ((error = xchk_nlinks_compare_iter(xnc, &ip)) == 1) {
+		error = xchk_nlinks_compare_inode(xnc, ip);
+		xchk_iscan_mark_visited(&xnc->compare_iscan, ip);
+		xchk_irele(sc, ip);
+		if (error)
+			break;
+
+		if (xchk_should_terminate(sc, &error))
+			break;
+	}
+	xchk_iscan_iter_finish(&xnc->compare_iscan);
+	xchk_iscan_teardown(&xnc->compare_iscan);
+	if (error)
+		return error;
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return 0;
+
+	/*
+	 * Walk all the non-null nlink observations that weren't checked in the
+	 * previous step.
+	 */
+	mutex_lock(&xnc->lock);
+	while ((error = xfarray_iter(xnc->nlinks, &cur, &nl)) == 1) {
+		xfs_ino_t	ino = cur - 1;
+
+		if (nl.flags & XCHK_NLINK_COMPARE_SCANNED)
+			continue;
+
+		mutex_unlock(&xnc->lock);
+
+		error = xchk_nlinks_compare_inum(xnc, ino);
+		if (error)
+			return error;
+
+		if (xchk_should_terminate(xnc->sc, &error))
+			return error;
+
+		mutex_lock(&xnc->lock);
+	}
+	mutex_unlock(&xnc->lock);
+
+	return error;
+}
+
+/* Tear down everything associated with a nlinks check. */
+static void
+xchk_nlinks_teardown_scan(
+	void			*priv)
+{
+	struct xchk_nlink_ctrs	*xnc = priv;
+
+	/* Discourage any hook functions that might be running. */
+	xchk_iscan_abort(&xnc->collect_iscan);
+
+	xfs_dir_hook_del(xnc->sc->mp, &xnc->dhook);
+
+	xfarray_destroy(xnc->nlinks);
+	xnc->nlinks = NULL;
+
+	xchk_iscan_teardown(&xnc->collect_iscan);
+	mutex_destroy(&xnc->lock);
+	xnc->sc = NULL;
+}
+
+/*
+ * Scan all inodes in the entire filesystem to generate link count data.  If
+ * the scan is successful, the counts will be left alive for a repair.  If any
+ * error occurs, we'll tear everything down.
+ */
+STATIC int
+xchk_nlinks_setup_scan(
+	struct xfs_scrub	*sc,
+	struct xchk_nlink_ctrs	*xnc)
+{
+	struct xfs_mount	*mp = sc->mp;
+	char			*descr;
+	unsigned long long	max_inos;
+	xfs_agnumber_t		last_agno = mp->m_sb.sb_agcount - 1;
+	xfs_agino_t		first_agino, last_agino;
+	int			error;
+
+	mutex_init(&xnc->lock);
+
+	/* Retry iget every tenth of a second for up to 30 seconds. */
+	xchk_iscan_start(sc, 30000, 100, &xnc->collect_iscan);
+
+	/*
+	 * Set up enough space to store an nlink record for the highest
+	 * possible inode number in this system.
+	 */
+	xfs_agino_range(mp, last_agno, &first_agino, &last_agino);
+	max_inos = XFS_AGINO_TO_INO(mp, last_agno, last_agino) + 1;
+	descr = xchk_xfile_descr(sc, "file link counts");
+	error = xfarray_create(descr, min(XFS_MAXINUMBER + 1, max_inos),
+			sizeof(struct xchk_nlink), &xnc->nlinks);
+	kfree(descr);
+	if (error)
+		goto out_teardown;
+
+	/*
+	 * Hook into the directory entry code so that we can capture updates to
+	 * file link counts.  The hook only triggers for inodes that were
+	 * already scanned, and the scanner thread takes each inode's ILOCK,
+	 * which means that any in-progress inode updates will finish before we
+	 * can scan the inode.
+	 */
+	ASSERT(sc->flags & XCHK_FSGATES_DIRENTS);
+	xfs_dir_hook_setup(&xnc->dhook, xchk_nlinks_live_update);
+	error = xfs_dir_hook_add(mp, &xnc->dhook);
+	if (error)
+		goto out_teardown;
+
+	/* Use deferred cleanup to pass the inode link count data to repair. */
+	sc->buf_cleanup = xchk_nlinks_teardown_scan;
+	return 0;
+
+out_teardown:
+	xchk_nlinks_teardown_scan(xnc);
+	return error;
+}
+
+/* Scrub the link count of all inodes on the filesystem. */
+int
+xchk_nlinks(
+	struct xfs_scrub	*sc)
+{
+	struct xchk_nlink_ctrs	*xnc = sc->buf;
+	int			error = 0;
+
+	/* Set ourselves up to check link counts on the live filesystem. */
+	error = xchk_nlinks_setup_scan(sc, xnc);
+	if (error)
+		return error;
+
+	/* Walk all inodes, picking up link count information. */
+	error = xchk_nlinks_collect(xnc);
+	if (!xchk_xref_process_error(sc, 0, 0, &error))
+		return error;
+
+	/* Fail fast if we're not playing with a full dataset. */
+	if (xchk_iscan_aborted(&xnc->collect_iscan))
+		xchk_set_incomplete(sc);
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_INCOMPLETE)
+		return 0;
+
+	/* Compare link counts. */
+	error = xchk_nlinks_compare(xnc);
+	if (!xchk_xref_process_error(sc, 0, 0, &error))
+		return error;
+
+	/* Check one last time for an incomplete dataset. */
+	if (xchk_iscan_aborted(&xnc->collect_iscan))
+		xchk_set_incomplete(sc);
+
+	return 0;
+}
diff --git a/fs/xfs/scrub/nlinks.h b/fs/xfs/scrub/nlinks.h
new file mode 100644
index 0000000000..b820712bfd
--- /dev/null
+++ b/fs/xfs/scrub/nlinks.h
@@ -0,0 +1,109 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (c) 2021-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_NLINKS_H__
+#define __XFS_SCRUB_NLINKS_H__
+
+/* Live link count control structure. */
+struct xchk_nlink_ctrs {
+	struct xfs_scrub	*sc;
+
+	/* Shadow link count data and its mutex. */
+	struct xfarray		*nlinks;
+	struct mutex		lock;
+
+	/*
+	 * The collection step uses a separate iscan context from the compare
+	 * step because the collection iscan coordinates live updates to the
+	 * observation data while this scanner is running.  The compare iscan
+	 * is secondary and can be reinitialized as needed.
+	 */
+	struct xchk_iscan	collect_iscan;
+	struct xchk_iscan	compare_iscan;
+
+	/*
+	 * Hook into directory updates so that we can receive live updates
+	 * from other writer threads.
+	 */
+	struct xfs_dir_hook	dhook;
+
+	/* Orphanage reparenting request. */
+	struct xrep_adoption	adoption;
+
+	/* Directory entry name, plus the trailing null. */
+	struct xfs_name		xname;
+	char			namebuf[MAXNAMELEN];
+};
+
+/*
+ * In-core link counts for a given inode in the filesystem.
+ *
+ * For an empty rootdir, the directory entries and the field to which they are
+ * accounted are as follows:
+ *
+ * Root directory:
+ *
+ * . points to self		(root.child)
+ * .. points to self		(root.parent)
+ * f1 points to a child file	(f1.parent)
+ * d1 points to a child dir	(d1.parent, root.child)
+ *
+ * Subdirectory d1:
+ *
+ * . points to self		(d1.child)
+ * .. points to root dir	(root.backref)
+ * f2 points to child file	(f2.parent)
+ * f3 points to root.f1		(f1.parent)
+ *
+ * root.nlink == 3 (root.dot, root.dotdot, root.d1)
+ * d1.nlink == 2 (root.d1, d1.dot)
+ * f1.nlink == 2 (root.f1, d1.f3)
+ * f2.nlink == 1 (d1.f2)
+ */
+struct xchk_nlink {
+	/* Count of forward links from parent directories to this file. */
+	xfs_nlink_t		parents;
+
+	/*
+	 * Count of back links to this parent directory from child
+	 * subdirectories.
+	 */
+	xfs_nlink_t		backrefs;
+
+	/*
+	 * Count of forward links from this directory to all child files and
+	 * the number of dot entries.  Should be zero for non-directories.
+	 */
+	xfs_nlink_t		children;
+
+	/* Record state flags */
+	unsigned int		flags;
+};
+
+/*
+ * This incore link count has been written at least once.  We never want to
+ * store an xchk_nlink that looks uninitialized.
+ */
+#define XCHK_NLINK_WRITTEN		(1U << 0)
+
+/* Already checked this link count record. */
+#define XCHK_NLINK_COMPARE_SCANNED	(1U << 1)
+
+/* Already made a repair with this link count record. */
+#define XREP_NLINK_DIRTY		(1U << 2)
+
+/* Compute total link count, using large enough variables to detect overflow. */
+static inline uint64_t
+xchk_nlink_total(struct xfs_inode *ip, const struct xchk_nlink *live)
+{
+	uint64_t	ret = live->parents;
+
+	/* Add one link count for the dot entry of any linked directory. */
+	if (ip && S_ISDIR(VFS_I(ip)->i_mode) && VFS_I(ip)->i_nlink)
+		ret++;
+	return ret + live->children;
+}
+
+#endif /* __XFS_SCRUB_NLINKS_H__ */
diff --git a/fs/xfs/scrub/nlinks_repair.c b/fs/xfs/scrub/nlinks_repair.c
new file mode 100644
index 0000000000..b3e707f47b
--- /dev/null
+++ b/fs/xfs/scrub/nlinks_repair.c
@@ -0,0 +1,353 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2021-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_inode.h"
+#include "xfs_icache.h"
+#include "xfs_bmap_util.h"
+#include "xfs_iwalk.h"
+#include "xfs_ialloc.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_dir2.h"
+#include "xfs_parent.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/repair.h"
+#include "scrub/xfile.h"
+#include "scrub/xfarray.h"
+#include "scrub/iscan.h"
+#include "scrub/orphanage.h"
+#include "scrub/nlinks.h"
+#include "scrub/trace.h"
+#include "scrub/tempfile.h"
+
+/*
+ * Live Inode Link Count Repair
+ * ============================
+ *
+ * Use the live inode link count information that we collected to replace the
+ * nlink values of the incore inodes.  A scrub->repair cycle should have left
+ * the live data and hooks active, so this is safe so long as we make sure the
+ * inode is locked.
+ */
+
+/* Set up to repair inode link counts. */
+int
+xrep_setup_nlinks(
+	struct xfs_scrub	*sc)
+{
+	return xrep_orphanage_try_create(sc);
+}
+
+/*
+ * Inodes that aren't the root directory or the orphanage, have a nonzero link
+ * count, and no observed parents should be moved to the orphanage.
+ */
+static inline bool
+xrep_nlinks_is_orphaned(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*ip,
+	unsigned int		actual_nlink,
+	const struct xchk_nlink	*obs)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+
+	if (obs->parents != 0)
+		return false;
+	if (ip == mp->m_rootip || ip == sc->orphanage)
+		return false;
+	return actual_nlink != 0;
+}
+
+/* Remove an inode from the unlinked list. */
+STATIC int
+xrep_nlinks_iunlink_remove(
+	struct xfs_scrub	*sc)
+{
+	struct xfs_perag	*pag;
+	int			error;
+
+	pag = xfs_perag_get(sc->mp, XFS_INO_TO_AGNO(sc->mp, sc->ip->i_ino));
+	error = xfs_iunlink_remove(sc->tp, pag, sc->ip);
+	xfs_perag_put(pag);
+	return error;
+}
+
+/*
+ * Correct the link count of the given inode.  Because we have to grab locks
+ * and resources in a certain order, it's possible that this will be a no-op.
+ */
+STATIC int
+xrep_nlinks_repair_inode(
+	struct xchk_nlink_ctrs	*xnc)
+{
+	struct xchk_nlink	obs;
+	struct xfs_scrub	*sc = xnc->sc;
+	struct xfs_mount	*mp = sc->mp;
+	struct xfs_inode	*ip = sc->ip;
+	uint64_t		total_links;
+	uint64_t		actual_nlink;
+	bool			orphanage_available = false;
+	bool			dirty = false;
+	int			error;
+
+	/*
+	 * Ignore temporary files being used to stage repairs, since we assume
+	 * they're correct for non-directories, and the directory repair code
+	 * doesn't bump the link counts for the children.
+	 */
+	if (xrep_is_tempfile(ip))
+		return 0;
+
+	/*
+	 * If the filesystem has an orphanage attached to the scrub context,
+	 * prepare for a link count repair that could involve @ip being adopted
+	 * by the lost+found.
+	 */
+	if (xrep_orphanage_can_adopt(sc)) {
+		error = xrep_orphanage_iolock_two(sc);
+		if (error)
+			return error;
+
+		error = xrep_adoption_trans_alloc(sc, &xnc->adoption);
+		if (error) {
+			xchk_iunlock(sc, XFS_IOLOCK_EXCL);
+			xrep_orphanage_iunlock(sc, XFS_IOLOCK_EXCL);
+		} else {
+			orphanage_available = true;
+		}
+	}
+
+	/*
+	 * Either there is no orphanage or we couldn't allocate resources for
+	 * that kind of update.  Let's try again with only the resources we
+	 * need for a simple link count update, since that's much more common.
+	 */
+	if (!orphanage_available) {
+		xchk_ilock(sc, XFS_IOLOCK_EXCL);
+
+		error = xfs_trans_alloc(mp, &M_RES(mp)->tr_link, 0, 0, 0,
+				&sc->tp);
+		if (error) {
+			xchk_iunlock(sc, XFS_IOLOCK_EXCL);
+			return error;
+		}
+
+		xchk_ilock(sc, XFS_ILOCK_EXCL);
+		xfs_trans_ijoin(sc->tp, ip, 0);
+	}
+
+	mutex_lock(&xnc->lock);
+
+	if (xchk_iscan_aborted(&xnc->collect_iscan)) {
+		error = -ECANCELED;
+		goto out_scanlock;
+	}
+
+	error = xfarray_load_sparse(xnc->nlinks, ip->i_ino, &obs);
+	if (error)
+		goto out_scanlock;
+
+	/*
+	 * We're done accessing the shared scan data, so we can drop the lock.
+	 * We still hold @ip's ILOCK, so its link count cannot change.
+	 */
+	mutex_unlock(&xnc->lock);
+
+	total_links = xchk_nlink_total(ip, &obs);
+	actual_nlink = VFS_I(ip)->i_nlink;
+
+	/*
+	 * Non-directories cannot have directories pointing up to them.
+	 *
+	 * We previously set error to zero, but set it again because one static
+	 * checker author fears that programmers will fail to maintain this
+	 * invariant and built their tool to flag this as a security risk.  A
+	 * different tool author made their bot complain about the redundant
+	 * store.  This is a never-ending and stupid battle; both tools missed
+	 * *actual bugs* elsewhere; and I no longer care.
+	 */
+	if (!S_ISDIR(VFS_I(ip)->i_mode) && obs.children != 0) {
+		trace_xrep_nlinks_unfixable_inode(mp, ip, &obs);
+		error = 0;
+		goto out_trans;
+	}
+
+	/*
+	 * Decide if we're going to move this file to the orphanage, and fix
+	 * up the incore link counts if we are.
+	 */
+	if (orphanage_available &&
+	    xrep_nlinks_is_orphaned(sc, ip, actual_nlink, &obs)) {
+		/* Figure out what name we're going to use here. */
+		error = xrep_adoption_compute_name(&xnc->adoption, &xnc->xname);
+		if (error)
+			goto out_trans;
+
+		/*
+		 * Reattach this file to the directory tree by moving it to
+		 * the orphanage per the adoption parameters that we already
+		 * computed.
+		 */
+		error = xrep_adoption_move(&xnc->adoption);
+		if (error)
+			goto out_trans;
+
+		/*
+		 * Re-read the link counts since the reparenting will have
+		 * updated our scan info.
+		 */
+		mutex_lock(&xnc->lock);
+		error = xfarray_load_sparse(xnc->nlinks, ip->i_ino, &obs);
+		mutex_unlock(&xnc->lock);
+		if (error)
+			goto out_trans;
+
+		total_links = xchk_nlink_total(ip, &obs);
+		actual_nlink = VFS_I(ip)->i_nlink;
+		dirty = true;
+	}
+
+	/*
+	 * If this inode is linked from the directory tree and on the unlinked
+	 * list, remove it from the unlinked list.
+	 */
+	if (total_links > 0 && xfs_inode_on_unlinked_list(ip)) {
+		error = xrep_nlinks_iunlink_remove(sc);
+		if (error)
+			goto out_trans;
+		dirty = true;
+	}
+
+	/*
+	 * If this inode is not linked from the directory tree yet not on the
+	 * unlinked list, put it on the unlinked list.
+	 */
+	if (total_links == 0 && !xfs_inode_on_unlinked_list(ip)) {
+		error = xfs_iunlink(sc->tp, ip);
+		if (error)
+			goto out_trans;
+		dirty = true;
+	}
+
+	/* Commit the new link count if it changed. */
+	if (total_links != actual_nlink) {
+		trace_xrep_nlinks_update_inode(mp, ip, &obs);
+
+		set_nlink(VFS_I(ip), min_t(unsigned long long, total_links,
+					   XFS_NLINK_PINNED));
+		dirty = true;
+	}
+
+	if (!dirty) {
+		error = 0;
+		goto out_trans;
+	}
+
+	xfs_trans_log_inode(sc->tp, ip, XFS_ILOG_CORE);
+
+	error = xrep_trans_commit(sc);
+	goto out_unlock;
+
+out_scanlock:
+	mutex_unlock(&xnc->lock);
+out_trans:
+	xchk_trans_cancel(sc);
+out_unlock:
+	xchk_iunlock(sc, XFS_ILOCK_EXCL);
+	if (orphanage_available) {
+		xrep_orphanage_iunlock(sc, XFS_ILOCK_EXCL);
+		xrep_orphanage_iunlock(sc, XFS_IOLOCK_EXCL);
+	}
+	xchk_iunlock(sc, XFS_IOLOCK_EXCL);
+	return error;
+}
+
+/*
+ * Try to visit every inode in the filesystem for repairs.  Move on if we can't
+ * grab an inode, since we're still making forward progress.
+ */
+static int
+xrep_nlinks_iter(
+	struct xchk_nlink_ctrs	*xnc,
+	struct xfs_inode	**ipp)
+{
+	int			error;
+
+	do {
+		error = xchk_iscan_iter(&xnc->compare_iscan, ipp);
+	} while (error == -EBUSY);
+
+	return error;
+}
+
+/* Commit the new inode link counters. */
+int
+xrep_nlinks(
+	struct xfs_scrub	*sc)
+{
+	struct xchk_nlink_ctrs	*xnc = sc->buf;
+	int			error;
+
+	/*
+	 * We need ftype for an accurate count of the number of child
+	 * subdirectory links.  Child subdirectories with a back link (dotdot
+	 * entry) but no forward link are moved to the orphanage, so we cannot
+	 * repair the link count of the parent directory based on the back link
+	 * count alone.  Filesystems without ftype support are rare (old V4) so
+	 * we just skip out here.
+	 */
+	if (!xfs_has_ftype(sc->mp))
+		return -EOPNOTSUPP;
+
+	/*
+	 * Use the inobt to walk all allocated inodes to compare and fix the
+	 * link counts.  Retry iget every tenth of a second for up to 30
+	 * seconds -- even if repair misses a few inodes, we still try to fix
+	 * as many of them as we can.
+	 */
+	xchk_iscan_start(sc, 30000, 100, &xnc->compare_iscan);
+	ASSERT(sc->ip == NULL);
+
+	while ((error = xrep_nlinks_iter(xnc, &sc->ip)) == 1) {
+		/*
+		 * Commit the scrub transaction so that we can create repair
+		 * transactions with the correct reservations.
+		 */
+		xchk_trans_cancel(sc);
+
+		error = xrep_nlinks_repair_inode(xnc);
+		xchk_iscan_mark_visited(&xnc->compare_iscan, sc->ip);
+		xchk_irele(sc, sc->ip);
+		sc->ip = NULL;
+		if (error)
+			break;
+
+		if (xchk_should_terminate(sc, &error))
+			break;
+
+		/*
+		 * Create a new empty transaction so that we can advance the
+		 * iscan cursor without deadlocking if the inobt has a cycle.
+		 * We can only push the inactivation workqueues with an empty
+		 * transaction.
+		 */
+		error = xchk_trans_alloc_empty(sc);
+		if (error)
+			break;
+	}
+	xchk_iscan_iter_finish(&xnc->compare_iscan);
+	xchk_iscan_teardown(&xnc->compare_iscan);
+
+	return error;
+}
diff --git a/fs/xfs/scrub/orphanage.c b/fs/xfs/scrub/orphanage.c
new file mode 100644
index 0000000000..7148d8362d
--- /dev/null
+++ b/fs/xfs/scrub/orphanage.c
@@ -0,0 +1,627 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2021-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_inode.h"
+#include "xfs_ialloc.h"
+#include "xfs_quota.h"
+#include "xfs_trans_space.h"
+#include "xfs_dir2.h"
+#include "xfs_icache.h"
+#include "xfs_bmap.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_parent.h"
+#include "xfs_attr_sf.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/repair.h"
+#include "scrub/trace.h"
+#include "scrub/orphanage.h"
+#include "scrub/readdir.h"
+
+#include <linux/namei.h>
+
+/*
+ * The Orphanage
+ * =============
+ *
+ * If the directory tree is damaged, children of that directory become
+ * inaccessible via that file path.  If a child has no other parents, the file
+ * is said to be orphaned.  xfs_repair fixes this situation by creating a
+ * orphanage directory (specifically, /lost+found) and creating a directory
+ * entry pointing to the orphaned file.
+ *
+ * Online repair follows this tactic by creating a root-owned /lost+found
+ * directory if one does not exist.  If an orphan is found, it will move that
+ * files into orphanage.
+ */
+
+/* Make the orphanage owned by root. */
+STATIC int
+xrep_chown_orphanage(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*dp)
+{
+	struct xfs_trans	*tp;
+	struct xfs_mount	*mp = sc->mp;
+	struct xfs_dquot	*udqp = NULL, *gdqp = NULL, *pdqp = NULL;
+	struct xfs_dquot	*oldu = NULL, *oldg = NULL, *oldp = NULL;
+	struct inode		*inode = VFS_I(dp);
+	int			error;
+
+	error = xfs_qm_vop_dqalloc(dp, GLOBAL_ROOT_UID, GLOBAL_ROOT_GID, 0,
+			XFS_QMOPT_QUOTALL, &udqp, &gdqp, &pdqp);
+	if (error)
+		return error;
+
+	error = xfs_trans_alloc_ichange(dp, udqp, gdqp, pdqp, true, &tp);
+	if (error)
+		goto out_dqrele;
+
+	/*
+	 * Always clear setuid/setgid/sticky on the orphanage since we don't
+	 * normally want that functionality on this directory and xfs_repair
+	 * doesn't create it this way either.  Leave the other access bits
+	 * unchanged.
+	 */
+	inode->i_mode &= ~(S_ISUID | S_ISGID | S_ISVTX);
+
+	/*
+	 * Change the ownerships and register quota modifications
+	 * in the transaction.
+	 */
+	if (!uid_eq(inode->i_uid, GLOBAL_ROOT_UID)) {
+		if (XFS_IS_UQUOTA_ON(mp))
+			oldu = xfs_qm_vop_chown(tp, dp, &dp->i_udquot, udqp);
+		inode->i_uid = GLOBAL_ROOT_UID;
+	}
+	if (!gid_eq(inode->i_gid, GLOBAL_ROOT_GID)) {
+		if (XFS_IS_GQUOTA_ON(mp))
+			oldg = xfs_qm_vop_chown(tp, dp, &dp->i_gdquot, gdqp);
+		inode->i_gid = GLOBAL_ROOT_GID;
+	}
+	if (dp->i_projid != 0) {
+		if (XFS_IS_PQUOTA_ON(mp))
+			oldp = xfs_qm_vop_chown(tp, dp, &dp->i_pdquot, pdqp);
+		dp->i_projid = 0;
+	}
+
+	dp->i_diflags &= ~(XFS_DIFLAG_REALTIME | XFS_DIFLAG_RTINHERIT);
+	xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
+
+	XFS_STATS_INC(mp, xs_ig_attrchg);
+
+	if (xfs_has_wsync(mp))
+		xfs_trans_set_sync(tp);
+	error = xfs_trans_commit(tp);
+
+	xfs_qm_dqrele(oldu);
+	xfs_qm_dqrele(oldg);
+	xfs_qm_dqrele(oldp);
+
+out_dqrele:
+	xfs_qm_dqrele(udqp);
+	xfs_qm_dqrele(gdqp);
+	xfs_qm_dqrele(pdqp);
+	return error;
+}
+
+#define ORPHANAGE	"lost+found"
+
+/* Create the orphanage directory, and set sc->orphanage to it. */
+int
+xrep_orphanage_create(
+	struct xfs_scrub	*sc)
+{
+	struct xfs_mount	*mp = sc->mp;
+	struct dentry		*root_dentry, *orphanage_dentry;
+	struct inode		*root_inode = VFS_I(sc->mp->m_rootip);
+	struct inode		*orphanage_inode;
+	int			error;
+
+	if (xfs_is_shutdown(mp))
+		return -EIO;
+	if (xfs_is_readonly(mp)) {
+		sc->orphanage = NULL;
+		return 0;
+	}
+
+	ASSERT(sc->tp == NULL);
+	ASSERT(sc->orphanage == NULL);
+
+	/* Find the dentry for the root directory... */
+	root_dentry = d_find_alias(root_inode);
+	if (!root_dentry) {
+		error = -EFSCORRUPTED;
+		goto out;
+	}
+
+	/* ...which is a directory, right? */
+	if (!d_is_dir(root_dentry)) {
+		error = -EFSCORRUPTED;
+		goto out_dput_root;
+	}
+
+	/* Try to find the orphanage directory. */
+	inode_lock_nested(root_inode, I_MUTEX_PARENT);
+	orphanage_dentry = lookup_one_len(ORPHANAGE, root_dentry,
+			strlen(ORPHANAGE));
+	if (IS_ERR(orphanage_dentry)) {
+		error = PTR_ERR(orphanage_dentry);
+		goto out_unlock_root;
+	}
+
+	/*
+	 * Nothing found?  Call mkdir to create the orphanage.  Create the
+	 * directory without other-user access because we're live and someone
+	 * could have been relying partly on minimal access to a parent
+	 * directory to control access to a file we put in here.
+	 */
+	if (d_really_is_negative(orphanage_dentry)) {
+		error = vfs_mkdir(&nop_mnt_idmap, root_inode, orphanage_dentry,
+				0750);
+		if (error)
+			goto out_dput_orphanage;
+	}
+
+	/* Not a directory? Bail out. */
+	if (!d_is_dir(orphanage_dentry)) {
+		error = -ENOTDIR;
+		goto out_dput_orphanage;
+	}
+
+	/*
+	 * Grab a reference to the orphanage.  This /should/ succeed since
+	 * we hold the root directory locked and therefore nobody can delete
+	 * the orphanage.
+	 */
+	orphanage_inode = igrab(d_inode(orphanage_dentry));
+	if (!orphanage_inode) {
+		error = -ENOENT;
+		goto out_dput_orphanage;
+	}
+
+	/* Make sure the orphanage is owned by root. */
+	error = xrep_chown_orphanage(sc, XFS_I(orphanage_inode));
+	if (error)
+		goto out_dput_orphanage;
+
+	/* Stash the reference for later and bail out. */
+	sc->orphanage = XFS_I(orphanage_inode);
+	sc->orphanage_ilock_flags = 0;
+
+out_dput_orphanage:
+	dput(orphanage_dentry);
+out_unlock_root:
+	inode_unlock(VFS_I(sc->mp->m_rootip));
+out_dput_root:
+	dput(root_dentry);
+out:
+	return error;
+}
+
+void
+xrep_orphanage_ilock(
+	struct xfs_scrub	*sc,
+	unsigned int		ilock_flags)
+{
+	sc->orphanage_ilock_flags |= ilock_flags;
+	xfs_ilock(sc->orphanage, ilock_flags);
+}
+
+bool
+xrep_orphanage_ilock_nowait(
+	struct xfs_scrub	*sc,
+	unsigned int		ilock_flags)
+{
+	if (xfs_ilock_nowait(sc->orphanage, ilock_flags)) {
+		sc->orphanage_ilock_flags |= ilock_flags;
+		return true;
+	}
+
+	return false;
+}
+
+void
+xrep_orphanage_iunlock(
+	struct xfs_scrub	*sc,
+	unsigned int		ilock_flags)
+{
+	xfs_iunlock(sc->orphanage, ilock_flags);
+	sc->orphanage_ilock_flags &= ~ilock_flags;
+}
+
+/* Grab the IOLOCK of the orphanage and sc->ip. */
+int
+xrep_orphanage_iolock_two(
+	struct xfs_scrub	*sc)
+{
+	int			error = 0;
+
+	while (true) {
+		if (xchk_should_terminate(sc, &error))
+			return error;
+
+		/*
+		 * Normal XFS takes the IOLOCK before grabbing a transaction.
+		 * Scrub holds a transaction, which means that we can't block
+		 * on either IOLOCK.
+		 */
+		if (xrep_orphanage_ilock_nowait(sc, XFS_IOLOCK_EXCL)) {
+			if (xchk_ilock_nowait(sc, XFS_IOLOCK_EXCL))
+				break;
+			xrep_orphanage_iunlock(sc, XFS_IOLOCK_EXCL);
+		}
+		delay(1);
+	}
+
+	return 0;
+}
+
+/* Release the orphanage. */
+void
+xrep_orphanage_rele(
+	struct xfs_scrub	*sc)
+{
+	if (!sc->orphanage)
+		return;
+
+	if (sc->orphanage_ilock_flags)
+		xfs_iunlock(sc->orphanage, sc->orphanage_ilock_flags);
+
+	xchk_irele(sc, sc->orphanage);
+	sc->orphanage = NULL;
+}
+
+/* Adoption moves a file into /lost+found */
+
+/* Can the orphanage adopt @sc->ip? */
+bool
+xrep_orphanage_can_adopt(
+	struct xfs_scrub	*sc)
+{
+	ASSERT(sc->ip != NULL);
+
+	if (!sc->orphanage)
+		return false;
+	if (sc->ip == sc->orphanage)
+		return false;
+	if (xfs_internal_inum(sc->mp, sc->ip->i_ino))
+		return false;
+	return true;
+}
+
+/*
+ * Create a new transaction to send a child to the orphanage.
+ *
+ * Allocate a new transaction with sufficient disk space to handle the
+ * adoption, take ILOCK_EXCL of the orphanage and sc->ip, joins them to the
+ * transaction, and reserve quota to reparent the latter.  Caller must hold the
+ * IOLOCK of the orphanage and sc->ip.
+ */
+int
+xrep_adoption_trans_alloc(
+	struct xfs_scrub	*sc,
+	struct xrep_adoption	*adopt)
+{
+	struct xfs_mount	*mp = sc->mp;
+	unsigned int		child_blkres = 0;
+	int			error;
+
+	ASSERT(sc->tp == NULL);
+	ASSERT(sc->ip != NULL);
+	ASSERT(sc->orphanage != NULL);
+	ASSERT(sc->ilock_flags & XFS_IOLOCK_EXCL);
+	ASSERT(sc->orphanage_ilock_flags & XFS_IOLOCK_EXCL);
+	ASSERT(!(sc->ilock_flags & (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)));
+	ASSERT(!(sc->orphanage_ilock_flags &
+				(XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)));
+
+	/* Compute the worst case space reservation that we need. */
+	adopt->sc = sc;
+	adopt->orphanage_blkres = xfs_link_space_res(mp, MAXNAMELEN);
+	if (S_ISDIR(VFS_I(sc->ip)->i_mode))
+		child_blkres = xfs_rename_space_res(mp, 0, false,
+						    xfs_name_dotdot.len, false);
+	if (xfs_has_parent(mp))
+		child_blkres += XFS_ADDAFORK_SPACE_RES(mp);
+	adopt->child_blkres = child_blkres;
+
+	/*
+	 * Allocate a transaction to link the child into the parent, along with
+	 * enough disk space to handle expansion of both the orphanage and the
+	 * dotdot entry of a child directory.
+	 */
+	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_link,
+			adopt->orphanage_blkres + adopt->child_blkres, 0, 0,
+			&sc->tp);
+	if (error)
+		return error;
+
+	xfs_lock_two_inodes(sc->orphanage, XFS_ILOCK_EXCL,
+			    sc->ip, XFS_ILOCK_EXCL);
+	sc->ilock_flags |= XFS_ILOCK_EXCL;
+	sc->orphanage_ilock_flags |= XFS_ILOCK_EXCL;
+
+	xfs_trans_ijoin(sc->tp, sc->orphanage, 0);
+	xfs_trans_ijoin(sc->tp, sc->ip, 0);
+
+	/*
+	 * Reserve enough quota in the orphan directory to add the new name.
+	 * Normally the orphanage should have user/group/project ids of zero
+	 * and hence is not subject to quota enforcement, but we're allowed to
+	 * exceed quota to reattach disconnected parts of the directory tree.
+	 */
+	error = xfs_trans_reserve_quota_nblks(sc->tp, sc->orphanage,
+			adopt->orphanage_blkres, 0, true);
+	if (error)
+		goto out_cancel;
+
+	/*
+	 * Reserve enough quota in the child directory to change dotdot.
+	 * Here we're also allowed to exceed file quota to repair inconsistent
+	 * metadata.
+	 */
+	if (adopt->child_blkres) {
+		error = xfs_trans_reserve_quota_nblks(sc->tp, sc->ip,
+				adopt->child_blkres, 0, true);
+		if (error)
+			goto out_cancel;
+	}
+
+	return 0;
+out_cancel:
+	xchk_trans_cancel(sc);
+	xrep_orphanage_iunlock(sc, XFS_ILOCK_EXCL);
+	xchk_iunlock(sc, XFS_ILOCK_EXCL);
+	return error;
+}
+
+/*
+ * Compute the xfs_name for the directory entry that we're adding to the
+ * orphanage.  Caller must hold ILOCKs of sc->ip and the orphanage and must not
+ * reuse namebuf until the adoption completes or is dissolved.
+ */
+int
+xrep_adoption_compute_name(
+	struct xrep_adoption	*adopt,
+	struct xfs_name		*xname)
+{
+	struct xfs_scrub	*sc = adopt->sc;
+	char			*namebuf = (void *)xname->name;
+	xfs_ino_t		ino;
+	unsigned int		incr = 0;
+	int			error = 0;
+
+	adopt->xname = xname;
+	xname->len = snprintf(namebuf, MAXNAMELEN, "%llu", sc->ip->i_ino);
+	xname->type = xfs_mode_to_ftype(VFS_I(sc->ip)->i_mode);
+
+	/* Make sure the filename is unique in the lost+found. */
+	error = xchk_dir_lookup(sc, sc->orphanage, xname, &ino);
+	while (error == 0 && incr < 10000) {
+		xname->len = snprintf(namebuf, MAXNAMELEN, "%llu.%u",
+				sc->ip->i_ino, ++incr);
+		error = xchk_dir_lookup(sc, sc->orphanage, xname, &ino);
+	}
+	if (error == 0) {
+		/* We already have 10,000 entries in the orphanage? */
+		return -EFSCORRUPTED;
+	}
+
+	if (error != -ENOENT)
+		return error;
+	return 0;
+}
+
+/*
+ * Make sure the dcache does not have a positive dentry for the name we've
+ * chosen.  The caller should have checked with the ondisk directory, so any
+ * discrepancy is a sign that something is seriously wrong.
+ */
+static int
+xrep_adoption_check_dcache(
+	struct xrep_adoption	*adopt)
+{
+	struct qstr		qname = QSTR_INIT(adopt->xname->name,
+						  adopt->xname->len);
+	struct xfs_scrub	*sc = adopt->sc;
+	struct dentry		*d_orphanage, *d_child;
+	int			error = 0;
+
+	d_orphanage = d_find_alias(VFS_I(sc->orphanage));
+	if (!d_orphanage)
+		return 0;
+
+	d_child = d_hash_and_lookup(d_orphanage, &qname);
+	if (d_child) {
+		trace_xrep_adoption_check_child(sc->mp, d_child);
+
+		if (d_is_positive(d_child)) {
+			ASSERT(d_is_negative(d_child));
+			error = -EFSCORRUPTED;
+		}
+
+		dput(d_child);
+	}
+
+	dput(d_orphanage);
+	return error;
+}
+
+/*
+ * Invalidate all dentries for the name that was added to the orphanage
+ * directory, and all dentries pointing to the child inode that was moved.
+ *
+ * There should not be any positive entries for the name, since we've
+ * maintained our lock on the orphanage directory.
+ */
+static void
+xrep_adoption_zap_dcache(
+	struct xrep_adoption	*adopt)
+{
+	struct qstr		qname = QSTR_INIT(adopt->xname->name,
+						  adopt->xname->len);
+	struct xfs_scrub	*sc = adopt->sc;
+	struct dentry		*d_orphanage, *d_child;
+
+	/* Invalidate all dentries for the adoption name */
+	d_orphanage = d_find_alias(VFS_I(sc->orphanage));
+	if (!d_orphanage)
+		return;
+
+	d_child = d_hash_and_lookup(d_orphanage, &qname);
+	while (d_child != NULL) {
+		trace_xrep_adoption_invalidate_child(sc->mp, d_child);
+
+		ASSERT(d_is_negative(d_child));
+		d_invalidate(d_child);
+		dput(d_child);
+		d_child = d_lookup(d_orphanage, &qname);
+	}
+
+	dput(d_orphanage);
+
+	/* Invalidate all the dentries pointing down to this file. */
+	while ((d_child = d_find_alias(VFS_I(sc->ip))) != NULL) {
+		trace_xrep_adoption_invalidate_child(sc->mp, d_child);
+
+		d_invalidate(d_child);
+		dput(d_child);
+	}
+}
+
+/*
+ * If we have to add an attr fork ahead of a parent pointer update, how much
+ * space should we ask for?
+ */
+static inline int
+xrep_adoption_attr_sizeof(
+	const struct xrep_adoption	*adopt)
+{
+	return sizeof(struct xfs_attr_sf_hdr) +
+		xfs_attr_sf_entsize_byname(sizeof(struct xfs_parent_rec),
+					   adopt->xname->len);
+}
+
+/*
+ * Move the current file to the orphanage under the computed name.
+ *
+ * Returns with a dirty transaction so that the caller can handle any other
+ * work, such as fixing up unlinked lists or resetting link counts.
+ */
+int
+xrep_adoption_move(
+	struct xrep_adoption	*adopt)
+{
+	struct xfs_scrub	*sc = adopt->sc;
+	bool			isdir = S_ISDIR(VFS_I(sc->ip)->i_mode);
+	int			error;
+
+	trace_xrep_adoption_reparent(sc->orphanage, adopt->xname,
+			sc->ip->i_ino);
+
+	error = xrep_adoption_check_dcache(adopt);
+	if (error)
+		return error;
+
+	/*
+	 * If this filesystem has parent pointers, ensure that the file being
+	 * moved to the orphanage has an attribute fork.  This is required
+	 * because the parent pointer code does not itself add attr forks.
+	 */
+	if (!xfs_inode_has_attr_fork(sc->ip) && xfs_has_parent(sc->mp)) {
+		int sf_size = xrep_adoption_attr_sizeof(adopt);
+
+		error = xfs_bmap_add_attrfork(sc->tp, sc->ip, sf_size, true);
+		if (error)
+			return error;
+	}
+
+	/* Create the new name in the orphanage. */
+	error = xfs_dir_createname(sc->tp, sc->orphanage, adopt->xname,
+			sc->ip->i_ino, adopt->orphanage_blkres);
+	if (error)
+		return error;
+
+	/*
+	 * Bump the link count of the orphanage if we just added a
+	 * subdirectory, and update its timestamps.
+	 */
+	xfs_trans_ichgtime(sc->tp, sc->orphanage,
+			XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
+	if (isdir)
+		xfs_bumplink(sc->tp, sc->orphanage);
+	xfs_trans_log_inode(sc->tp, sc->orphanage, XFS_ILOG_CORE);
+
+	/* Bump the link count of the child. */
+	if (adopt->bump_child_nlink) {
+		xfs_bumplink(sc->tp, sc->ip);
+		xfs_trans_log_inode(sc->tp, sc->ip, XFS_ILOG_CORE);
+	}
+
+	/* Replace the dotdot entry if the child is a subdirectory. */
+	if (isdir) {
+		error = xfs_dir_replace(sc->tp, sc->ip, &xfs_name_dotdot,
+				sc->orphanage->i_ino, adopt->child_blkres);
+		if (error)
+			return error;
+	}
+
+	/* Add a parent pointer from the file back to the lost+found. */
+	if (xfs_has_parent(sc->mp)) {
+		error = xfs_parent_addname(sc->tp, &adopt->ppargs,
+				sc->orphanage, adopt->xname, sc->ip);
+		if (error)
+			return error;
+	}
+
+	/*
+	 * Notify dirent hooks that we moved the file to /lost+found, and
+	 * finish all the deferred work so that we know the adoption is fully
+	 * recorded in the log.
+	 */
+	xfs_dir_update_hook(sc->orphanage, sc->ip, 1, adopt->xname);
+
+	/* Remove negative dentries from the lost+found's dcache */
+	xrep_adoption_zap_dcache(adopt);
+	return 0;
+}
+
+/*
+ * Roll to a clean scrub transaction so that we can release the orphanage,
+ * even if xrep_adoption_move was not called.
+ *
+ * Commits all the work and deferred ops attached to an adoption request and
+ * rolls to a clean scrub transaction.  On success, returns 0 with the scrub
+ * context holding a clean transaction with no inodes joined.  On failure,
+ * returns negative errno with no scrub transaction.  All inode locks are
+ * still held after this function returns.
+ */
+int
+xrep_adoption_trans_roll(
+	struct xrep_adoption	*adopt)
+{
+	struct xfs_scrub	*sc = adopt->sc;
+	int			error;
+
+	trace_xrep_adoption_trans_roll(sc->orphanage, sc->ip,
+			!!(sc->tp->t_flags & XFS_TRANS_DIRTY));
+
+	/* Finish all the deferred ops to commit all repairs. */
+	error = xrep_defer_finish(sc);
+	if (error)
+		return error;
+
+	/* Roll the transaction once more to detach the inodes. */
+	return xfs_trans_roll(&sc->tp);
+}
diff --git a/fs/xfs/scrub/orphanage.h b/fs/xfs/scrub/orphanage.h
new file mode 100644
index 0000000000..7c7a2e7d81
--- /dev/null
+++ b/fs/xfs/scrub/orphanage.h
@@ -0,0 +1,86 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2021-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_ORPHANAGE_H__
+#define __XFS_SCRUB_ORPHANAGE_H__
+
+#ifdef CONFIG_XFS_ONLINE_REPAIR
+int xrep_orphanage_create(struct xfs_scrub *sc);
+
+/*
+ * If we're doing a repair, ensure that the orphanage exists and attach it to
+ * the scrub context.
+ */
+static inline int
+xrep_orphanage_try_create(
+	struct xfs_scrub	*sc)
+{
+	int			error;
+
+	ASSERT(sc->sm->sm_flags & XFS_SCRUB_IFLAG_REPAIR);
+
+	error = xrep_orphanage_create(sc);
+	switch (error) {
+	case 0:
+	case -ENOENT:
+	case -ENOTDIR:
+	case -ENOSPC:
+		/*
+		 * If the orphanage can't be found or isn't a directory, we'll
+		 * keep going, but we won't be able to attach the file to the
+		 * orphanage if we can't find the parent.
+		 */
+		return 0;
+	}
+
+	return error;
+}
+
+int xrep_orphanage_iolock_two(struct xfs_scrub *sc);
+
+void xrep_orphanage_ilock(struct xfs_scrub *sc, unsigned int ilock_flags);
+bool xrep_orphanage_ilock_nowait(struct xfs_scrub *sc,
+		unsigned int ilock_flags);
+void xrep_orphanage_iunlock(struct xfs_scrub *sc, unsigned int ilock_flags);
+
+void xrep_orphanage_rele(struct xfs_scrub *sc);
+
+/* Information about a request to add a file to the orphanage. */
+struct xrep_adoption {
+	struct xfs_scrub	*sc;
+
+	/* Name used for the adoption. */
+	struct xfs_name		*xname;
+
+	/* Parent pointer context tracking */
+	struct xfs_parent_args	ppargs;
+
+	/* Block reservations for orphanage and child (if directory). */
+	unsigned int		orphanage_blkres;
+	unsigned int		child_blkres;
+
+	/*
+	 * Does the caller want us to bump the child link count?  This is not
+	 * needed when reattaching files that have become disconnected but have
+	 * nlink > 1.  It is necessary when changing the directory tree
+	 * structure.
+	 */
+	bool			bump_child_nlink:1;
+};
+
+bool xrep_orphanage_can_adopt(struct xfs_scrub *sc);
+
+int xrep_adoption_trans_alloc(struct xfs_scrub *sc,
+		struct xrep_adoption *adopt);
+int xrep_adoption_compute_name(struct xrep_adoption *adopt,
+		struct xfs_name *xname);
+int xrep_adoption_move(struct xrep_adoption *adopt);
+int xrep_adoption_trans_roll(struct xrep_adoption *adopt);
+#else
+struct xrep_adoption { /* empty */ };
+# define xrep_orphanage_rele(sc)	((void)0)
+#endif /* CONFIG_XFS_ONLINE_REPAIR */
+
+#endif /* __XFS_SCRUB_ORPHANAGE_H__ */
diff --git a/fs/xfs/scrub/parent.c b/fs/xfs/scrub/parent.c
index 7db8736721..733c410a22 100644
--- a/fs/xfs/scrub/parent.c
+++ b/fs/xfs/scrub/parent.c
@@ -10,19 +10,37 @@
 #include "xfs_trans_resv.h"
 #include "xfs_mount.h"
 #include "xfs_log_format.h"
+#include "xfs_trans.h"
 #include "xfs_inode.h"
 #include "xfs_icache.h"
 #include "xfs_dir2.h"
 #include "xfs_dir2_priv.h"
+#include "xfs_attr.h"
+#include "xfs_parent.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/readdir.h"
+#include "scrub/tempfile.h"
+#include "scrub/repair.h"
+#include "scrub/listxattr.h"
+#include "scrub/xfile.h"
+#include "scrub/xfarray.h"
+#include "scrub/xfblob.h"
+#include "scrub/trace.h"
 
 /* Set us up to scrub parents. */
 int
 xchk_setup_parent(
 	struct xfs_scrub	*sc)
 {
+	int			error;
+
+	if (xchk_could_repair(sc)) {
+		error = xrep_setup_parent(sc);
+		if (error)
+			return error;
+	}
+
 	return xchk_setup_inode_contents(sc, 0);
 }
 
@@ -143,7 +161,8 @@ xchk_parent_validate(
 	}
 	if (!xchk_fblock_xref_process_error(sc, XFS_DATA_FORK, 0, &error))
 		return error;
-	if (dp == sc->ip || !S_ISDIR(VFS_I(dp)->i_mode)) {
+	if (dp == sc->ip || xrep_is_tempfile(dp) ||
+	    !S_ISDIR(VFS_I(dp)->i_mode)) {
 		xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
 		goto out_rele;
 	}
@@ -185,6 +204,621 @@ out_rele:
 	return error;
 }
 
+/*
+ * Checking of Parent Pointers
+ * ===========================
+ *
+ * On filesystems with directory parent pointers, we check the referential
+ * integrity by visiting each parent pointer of a child file and checking that
+ * the directory referenced by the pointer actually has a dirent pointing
+ * forward to the child file.
+ */
+
+/* Deferred parent pointer entry that we saved for later. */
+struct xchk_pptr {
+	/* Cookie for retrieval of the pptr name. */
+	xfblob_cookie		name_cookie;
+
+	/* Parent pointer record. */
+	struct xfs_parent_rec	pptr_rec;
+
+	/* Length of the pptr name. */
+	uint8_t			namelen;
+};
+
+struct xchk_pptrs {
+	struct xfs_scrub	*sc;
+
+	/* How many parent pointers did we find at the end? */
+	unsigned long long	pptrs_found;
+
+	/* Parent of this directory. */
+	xfs_ino_t		parent_ino;
+
+	/* Fixed-size array of xchk_pptr structures. */
+	struct xfarray		*pptr_entries;
+
+	/* Blobs containing parent pointer names. */
+	struct xfblob		*pptr_names;
+
+	/* Scratch buffer for scanning pptr xattrs */
+	struct xfs_da_args	pptr_args;
+
+	/* If we've cycled the ILOCK, we must revalidate all deferred pptrs. */
+	bool			need_revalidate;
+
+	/* Name buffer */
+	struct xfs_name		xname;
+	char			namebuf[MAXNAMELEN];
+};
+
+/* Does this parent pointer match the dotdot entry? */
+STATIC int
+xchk_parent_scan_dotdot(
+	struct xfs_scrub		*sc,
+	struct xfs_inode		*ip,
+	unsigned int			attr_flags,
+	const unsigned char		*name,
+	unsigned int			namelen,
+	const void			*value,
+	unsigned int			valuelen,
+	void				*priv)
+{
+	struct xchk_pptrs		*pp = priv;
+	xfs_ino_t			parent_ino;
+	int				error;
+
+	if (!(attr_flags & XFS_ATTR_PARENT))
+		return 0;
+
+	error = xfs_parent_from_attr(sc->mp, attr_flags, name, namelen, value,
+			valuelen, &parent_ino, NULL);
+	if (error)
+		return error;
+
+	if (pp->parent_ino == parent_ino)
+		return -ECANCELED;
+
+	return 0;
+}
+
+/* Look up the dotdot entry so that we can check it as we walk the pptrs. */
+STATIC int
+xchk_parent_pptr_and_dotdot(
+	struct xchk_pptrs	*pp)
+{
+	struct xfs_scrub	*sc = pp->sc;
+	int			error;
+
+	/* Look up '..' */
+	error = xchk_dir_lookup(sc, sc->ip, &xfs_name_dotdot, &pp->parent_ino);
+	if (!xchk_fblock_process_error(sc, XFS_DATA_FORK, 0, &error))
+		return error;
+	if (!xfs_verify_dir_ino(sc->mp, pp->parent_ino)) {
+		xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
+		return 0;
+	}
+
+	/* Is this the root dir?  Then '..' must point to itself. */
+	if (sc->ip == sc->mp->m_rootip) {
+		if (sc->ip->i_ino != pp->parent_ino)
+			xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
+		return 0;
+	}
+
+	/*
+	 * If this is now an unlinked directory, the dotdot value is
+	 * meaningless as long as it points to a valid inode.
+	 */
+	if (VFS_I(sc->ip)->i_nlink == 0)
+		return 0;
+
+	if (pp->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return 0;
+
+	/* Otherwise, walk the pptrs again, and check. */
+	error = xchk_xattr_walk(sc, sc->ip, xchk_parent_scan_dotdot, NULL, pp);
+	if (error == -ECANCELED) {
+		/* Found a parent pointer that matches dotdot. */
+		return 0;
+	}
+	if (!error || error == -EFSCORRUPTED) {
+		/* Found a broken parent pointer or no match. */
+		xchk_fblock_set_corrupt(sc, XFS_ATTR_FORK, 0);
+		return 0;
+	}
+	return error;
+}
+
+/*
+ * Try to lock a parent directory for checking dirents.  Returns the inode
+ * flags for the locks we now hold, or zero if we failed.
+ */
+STATIC unsigned int
+xchk_parent_lock_dir(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*dp)
+{
+	if (!xfs_ilock_nowait(dp, XFS_IOLOCK_SHARED))
+		return 0;
+
+	if (!xfs_ilock_nowait(dp, XFS_ILOCK_SHARED)) {
+		xfs_iunlock(dp, XFS_IOLOCK_SHARED);
+		return 0;
+	}
+
+	if (!xfs_need_iread_extents(&dp->i_df))
+		return XFS_IOLOCK_SHARED | XFS_ILOCK_SHARED;
+
+	xfs_iunlock(dp, XFS_ILOCK_SHARED);
+
+	if (!xfs_ilock_nowait(dp, XFS_ILOCK_EXCL)) {
+		xfs_iunlock(dp, XFS_IOLOCK_SHARED);
+		return 0;
+	}
+
+	return XFS_IOLOCK_SHARED | XFS_ILOCK_EXCL;
+}
+
+/* Check the forward link (dirent) associated with this parent pointer. */
+STATIC int
+xchk_parent_dirent(
+	struct xchk_pptrs	*pp,
+	const struct xfs_name	*xname,
+	struct xfs_inode	*dp)
+{
+	struct xfs_scrub	*sc = pp->sc;
+	xfs_ino_t		child_ino;
+	int			error;
+
+	/*
+	 * Use the name attached to this parent pointer to look up the
+	 * directory entry in the alleged parent.
+	 */
+	error = xchk_dir_lookup(sc, dp, xname, &child_ino);
+	if (error == -ENOENT) {
+		xchk_fblock_xref_set_corrupt(sc, XFS_ATTR_FORK, 0);
+		return 0;
+	}
+	if (!xchk_fblock_xref_process_error(sc, XFS_ATTR_FORK, 0, &error))
+		return error;
+
+	/* Does the inode number match? */
+	if (child_ino != sc->ip->i_ino) {
+		xchk_fblock_xref_set_corrupt(sc, XFS_ATTR_FORK, 0);
+		return 0;
+	}
+
+	return 0;
+}
+
+/* Try to grab a parent directory. */
+STATIC int
+xchk_parent_iget(
+	struct xchk_pptrs	*pp,
+	const struct xfs_parent_rec	*pptr,
+	struct xfs_inode	**dpp)
+{
+	struct xfs_scrub	*sc = pp->sc;
+	struct xfs_inode	*ip;
+	xfs_ino_t		parent_ino = be64_to_cpu(pptr->p_ino);
+	int			error;
+
+	/* Validate inode number. */
+	error = xfs_dir_ino_validate(sc->mp, parent_ino);
+	if (error) {
+		xchk_fblock_set_corrupt(sc, XFS_ATTR_FORK, 0);
+		return -ECANCELED;
+	}
+
+	error = xchk_iget(sc, parent_ino, &ip);
+	if (error == -EINVAL || error == -ENOENT) {
+		xchk_fblock_set_corrupt(sc, XFS_ATTR_FORK, 0);
+		return -ECANCELED;
+	}
+	if (!xchk_fblock_xref_process_error(sc, XFS_ATTR_FORK, 0, &error))
+		return error;
+
+	/* The parent must be a directory. */
+	if (!S_ISDIR(VFS_I(ip)->i_mode)) {
+		xchk_fblock_xref_set_corrupt(sc, XFS_ATTR_FORK, 0);
+		goto out_rele;
+	}
+
+	/* Validate generation number. */
+	if (VFS_I(ip)->i_generation != be32_to_cpu(pptr->p_gen)) {
+		xchk_fblock_xref_set_corrupt(sc, XFS_ATTR_FORK, 0);
+		goto out_rele;
+	}
+
+	*dpp = ip;
+	return 0;
+out_rele:
+	xchk_irele(sc, ip);
+	return 0;
+}
+
+/*
+ * Walk an xattr of a file.  If this xattr is a parent pointer, follow it up
+ * to a parent directory and check that the parent has a dirent pointing back
+ * to us.
+ */
+STATIC int
+xchk_parent_scan_attr(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*ip,
+	unsigned int		attr_flags,
+	const unsigned char	*name,
+	unsigned int		namelen,
+	const void		*value,
+	unsigned int		valuelen,
+	void			*priv)
+{
+	struct xfs_name		xname = {
+		.name		= name,
+		.len		= namelen,
+	};
+	struct xchk_pptrs	*pp = priv;
+	struct xfs_inode	*dp = NULL;
+	const struct xfs_parent_rec *pptr_rec = value;
+	xfs_ino_t		parent_ino;
+	unsigned int		lockmode;
+	int			error;
+
+	if (!(attr_flags & XFS_ATTR_PARENT))
+		return 0;
+
+	error = xfs_parent_from_attr(sc->mp, attr_flags, name, namelen, value,
+			valuelen, &parent_ino, NULL);
+	if (error) {
+		xchk_fblock_set_corrupt(sc, XFS_ATTR_FORK, 0);
+		return error;
+	}
+
+	/* No self-referential parent pointers. */
+	if (parent_ino == sc->ip->i_ino) {
+		xchk_fblock_set_corrupt(sc, XFS_ATTR_FORK, 0);
+		return -ECANCELED;
+	}
+
+	pp->pptrs_found++;
+
+	error = xchk_parent_iget(pp, pptr_rec, &dp);
+	if (error)
+		return error;
+	if (!dp)
+		return 0;
+
+	/* Try to lock the inode. */
+	lockmode = xchk_parent_lock_dir(sc, dp);
+	if (!lockmode) {
+		struct xchk_pptr	save_pp = {
+			.pptr_rec	= *pptr_rec, /* struct copy */
+			.namelen	= namelen,
+		};
+
+		/* Couldn't lock the inode, so save the pptr for later. */
+		trace_xchk_parent_defer(sc->ip, &xname, dp->i_ino);
+
+		error = xfblob_storename(pp->pptr_names, &save_pp.name_cookie,
+				&xname);
+		if (!xchk_fblock_xref_process_error(sc, XFS_ATTR_FORK, 0,
+					&error))
+			goto out_rele;
+
+		error = xfarray_append(pp->pptr_entries, &save_pp);
+		if (!xchk_fblock_xref_process_error(sc, XFS_ATTR_FORK, 0,
+					&error))
+			goto out_rele;
+
+		goto out_rele;
+	}
+
+	error = xchk_parent_dirent(pp, &xname, dp);
+	if (error)
+		goto out_unlock;
+
+out_unlock:
+	xfs_iunlock(dp, lockmode);
+out_rele:
+	xchk_irele(sc, dp);
+	return error;
+}
+
+/*
+ * Revalidate a parent pointer that we collected in the past but couldn't check
+ * because of lock contention.  Returns 0 if the parent pointer is still valid,
+ * -ENOENT if it has gone away on us, or a negative errno.
+ */
+STATIC int
+xchk_parent_revalidate_pptr(
+	struct xchk_pptrs		*pp,
+	const struct xfs_name		*xname,
+	struct xfs_parent_rec		*pptr)
+{
+	struct xfs_scrub		*sc = pp->sc;
+	int				error;
+
+	error = xfs_parent_lookup(sc->tp, sc->ip, xname, pptr, &pp->pptr_args);
+	if (error == -ENOATTR) {
+		/* Parent pointer went away, nothing to revalidate. */
+		return -ENOENT;
+	}
+
+	return error;
+}
+
+/*
+ * Check a parent pointer the slow way, which means we cycle locks a bunch
+ * and put up with revalidation until we get it done.
+ */
+STATIC int
+xchk_parent_slow_pptr(
+	struct xchk_pptrs	*pp,
+	const struct xfs_name	*xname,
+	struct xfs_parent_rec	*pptr)
+{
+	struct xfs_scrub	*sc = pp->sc;
+	struct xfs_inode	*dp = NULL;
+	unsigned int		lockmode;
+	int			error;
+
+	/* Check that the deferred parent pointer still exists. */
+	if (pp->need_revalidate) {
+		error = xchk_parent_revalidate_pptr(pp, xname, pptr);
+		if (error == -ENOENT)
+			return 0;
+		if (!xchk_fblock_xref_process_error(sc, XFS_ATTR_FORK, 0,
+					&error))
+			return error;
+	}
+
+	error = xchk_parent_iget(pp, pptr, &dp);
+	if (error)
+		return error;
+	if (!dp)
+		return 0;
+
+	/*
+	 * If we can grab both IOLOCK and ILOCK of the alleged parent, we
+	 * can proceed with the validation.
+	 */
+	lockmode = xchk_parent_lock_dir(sc, dp);
+	if (lockmode) {
+		trace_xchk_parent_slowpath(sc->ip, xname, dp->i_ino);
+		goto check_dirent;
+	}
+
+	/*
+	 * We couldn't lock the parent dir.  Drop all the locks and try to
+	 * get them again, one at a time.
+	 */
+	xchk_iunlock(sc, sc->ilock_flags);
+	pp->need_revalidate = true;
+
+	trace_xchk_parent_ultraslowpath(sc->ip, xname, dp->i_ino);
+
+	error = xchk_dir_trylock_for_pptrs(sc, dp, &lockmode);
+	if (error)
+		goto out_rele;
+
+	/* Revalidate the parent pointer now that we cycled locks. */
+	error = xchk_parent_revalidate_pptr(pp, xname, pptr);
+	if (error == -ENOENT) {
+		error = 0;
+		goto out_unlock;
+	}
+	if (!xchk_fblock_xref_process_error(sc, XFS_ATTR_FORK, 0, &error))
+		goto out_unlock;
+
+check_dirent:
+	error = xchk_parent_dirent(pp, xname, dp);
+out_unlock:
+	xfs_iunlock(dp, lockmode);
+out_rele:
+	xchk_irele(sc, dp);
+	return error;
+}
+
+/* Check all the parent pointers that we deferred the first time around. */
+STATIC int
+xchk_parent_finish_slow_pptrs(
+	struct xchk_pptrs	*pp)
+{
+	xfarray_idx_t		array_cur;
+	int			error;
+
+	foreach_xfarray_idx(pp->pptr_entries, array_cur) {
+		struct xchk_pptr	pptr;
+
+		if (pp->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+			return 0;
+
+		error = xfarray_load(pp->pptr_entries, array_cur, &pptr);
+		if (error)
+			return error;
+
+		error = xfblob_loadname(pp->pptr_names, pptr.name_cookie,
+				&pp->xname, pptr.namelen);
+		if (error)
+			return error;
+
+		error = xchk_parent_slow_pptr(pp, &pp->xname, &pptr.pptr_rec);
+		if (error)
+			return error;
+	}
+
+	/* Empty out both xfiles now that we've checked everything. */
+	xfarray_truncate(pp->pptr_entries);
+	xfblob_truncate(pp->pptr_names);
+	return 0;
+}
+
+/* Count the number of parent pointers. */
+STATIC int
+xchk_parent_count_pptr(
+	struct xfs_scrub		*sc,
+	struct xfs_inode		*ip,
+	unsigned int			attr_flags,
+	const unsigned char		*name,
+	unsigned int			namelen,
+	const void			*value,
+	unsigned int			valuelen,
+	void				*priv)
+{
+	struct xchk_pptrs		*pp = priv;
+	int				error;
+
+	if (!(attr_flags & XFS_ATTR_PARENT))
+		return 0;
+
+	error = xfs_parent_from_attr(sc->mp, attr_flags, name, namelen, value,
+			valuelen, NULL, NULL);
+	if (error)
+		return error;
+
+	pp->pptrs_found++;
+	return 0;
+}
+
+/*
+ * Compare the number of parent pointers to the link count.  For
+ * non-directories these should be the same.  For unlinked directories the
+ * count should be zero; for linked directories, it should be nonzero.
+ */
+STATIC int
+xchk_parent_count_pptrs(
+	struct xchk_pptrs	*pp)
+{
+	struct xfs_scrub	*sc = pp->sc;
+	int			error;
+
+	/*
+	 * If we cycled the ILOCK while cross-checking parent pointers with
+	 * dirents, then we need to recalculate the number of parent pointers.
+	 */
+	if (pp->need_revalidate) {
+		pp->pptrs_found = 0;
+		error = xchk_xattr_walk(sc, sc->ip, xchk_parent_count_pptr,
+				NULL, pp);
+		if (error == -EFSCORRUPTED) {
+			/* Found a bad parent pointer */
+			xchk_fblock_set_corrupt(sc, XFS_ATTR_FORK, 0);
+			return 0;
+		}
+		if (error)
+			return error;
+	}
+
+	if (S_ISDIR(VFS_I(sc->ip)->i_mode)) {
+		if (sc->ip == sc->mp->m_rootip)
+			pp->pptrs_found++;
+
+		if (VFS_I(sc->ip)->i_nlink == 0 && pp->pptrs_found > 0)
+			xchk_ino_set_corrupt(sc, sc->ip->i_ino);
+		else if (VFS_I(sc->ip)->i_nlink > 0 &&
+			 pp->pptrs_found == 0)
+			xchk_ino_set_corrupt(sc, sc->ip->i_ino);
+	} else {
+		if (VFS_I(sc->ip)->i_nlink != pp->pptrs_found)
+			xchk_ino_set_corrupt(sc, sc->ip->i_ino);
+	}
+
+	return 0;
+}
+
+/* Check parent pointers of a file. */
+STATIC int
+xchk_parent_pptr(
+	struct xfs_scrub	*sc)
+{
+	struct xchk_pptrs	*pp;
+	char			*descr;
+	int			error;
+
+	pp = kvzalloc(sizeof(struct xchk_pptrs), XCHK_GFP_FLAGS);
+	if (!pp)
+		return -ENOMEM;
+	pp->sc = sc;
+	pp->xname.name = pp->namebuf;
+
+	/*
+	 * Set up some staging memory for parent pointers that we can't check
+	 * due to locking contention.
+	 */
+	descr = xchk_xfile_ino_descr(sc, "slow parent pointer entries");
+	error = xfarray_create(descr, 0, sizeof(struct xchk_pptr),
+			&pp->pptr_entries);
+	kfree(descr);
+	if (error)
+		goto out_pp;
+
+	descr = xchk_xfile_ino_descr(sc, "slow parent pointer names");
+	error = xfblob_create(descr, &pp->pptr_names);
+	kfree(descr);
+	if (error)
+		goto out_entries;
+
+	error = xchk_xattr_walk(sc, sc->ip, xchk_parent_scan_attr, NULL, pp);
+	if (error == -ECANCELED) {
+		error = 0;
+		goto out_names;
+	}
+	if (error)
+		goto out_names;
+
+	error = xchk_parent_finish_slow_pptrs(pp);
+	if (error == -ETIMEDOUT) {
+		/* Couldn't grab a lock, scrub was marked incomplete */
+		error = 0;
+		goto out_names;
+	}
+	if (error)
+		goto out_names;
+
+	if (pp->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		goto out_names;
+
+	/*
+	 * For subdirectories, make sure the dotdot entry references the same
+	 * inode as the parent pointers.
+	 *
+	 * If we're scanning a /consistent/ directory, there should only be
+	 * one parent pointer, and it should point to the same directory as
+	 * the dotdot entry.
+	 *
+	 * However, a corrupt directory tree might feature a subdirectory with
+	 * multiple parents.  The directory loop scanner is responsible for
+	 * correcting that kind of problem, so for now we only validate that
+	 * the dotdot entry matches /one/ of the parents.
+	 */
+	if (S_ISDIR(VFS_I(sc->ip)->i_mode)) {
+		error = xchk_parent_pptr_and_dotdot(pp);
+		if (error)
+			goto out_names;
+	}
+
+	if (pp->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		goto out_pp;
+
+	/*
+	 * Complain if the number of parent pointers doesn't match the link
+	 * count.  This could be a sign of missing parent pointers (or an
+	 * incorrect link count).
+	 */
+	error = xchk_parent_count_pptrs(pp);
+	if (error)
+		goto out_names;
+
+out_names:
+	xfblob_destroy(pp->pptr_names);
+out_entries:
+	xfarray_destroy(pp->pptr_entries);
+out_pp:
+	kvfree(pp);
+	return error;
+}
+
 /* Scrub a parent pointer. */
 int
 xchk_parent(
@@ -194,6 +828,9 @@ xchk_parent(
 	xfs_ino_t		parent_ino;
 	int			error = 0;
 
+	if (xfs_has_parent(mp))
+		return xchk_parent_pptr(sc);
+
 	/*
 	 * If we're a directory, check that the '..' link points up to
 	 * a directory that has one entry pointing to us.
@@ -237,3 +874,64 @@ xchk_parent(
 
 	return error;
 }
+
+/*
+ * Decide if this file's extended attributes (and therefore its parent
+ * pointers) have been zapped to satisfy the inode and ifork verifiers.
+ * Checking and repairing should be postponed until the extended attribute
+ * structure is fixed.
+ */
+bool
+xchk_pptr_looks_zapped(
+	struct xfs_inode	*ip)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	struct inode		*inode = VFS_I(ip);
+
+	ASSERT(xfs_has_parent(mp));
+
+	/*
+	 * Temporary files that cannot be linked into the directory tree do not
+	 * have attr forks because they cannot ever have parents.
+	 */
+	if (inode->i_nlink == 0 && !(inode->i_state & I_LINKABLE))
+		return false;
+
+	/*
+	 * Directory tree roots do not have parents, so the expected outcome
+	 * of a parent pointer scan is always the empty set.  It's safe to scan
+	 * them even if the attr fork was zapped.
+	 */
+	if (ip == mp->m_rootip)
+		return false;
+
+	/*
+	 * Metadata inodes are all rooted in the superblock and do not have
+	 * any parents.  Hence the attr fork will not be initialized, but
+	 * there are no parent pointers that might have been zapped.
+	 */
+	if (xfs_is_metadata_inode(ip))
+		return false;
+
+	/*
+	 * Linked and linkable non-rootdir files should always have an
+	 * attribute fork because that is where parent pointers are
+	 * stored.  If the fork is absent, something is amiss.
+	 */
+	if (!xfs_inode_has_attr_fork(ip))
+		return true;
+
+	/* Repair zapped this file's attr fork a short time ago */
+	if (xfs_ifork_zapped(ip, XFS_ATTR_FORK))
+		return true;
+
+	/*
+	 * If the dinode repair found a bad attr fork, it will reset the fork
+	 * to extents format with zero records and wait for the bmapbta
+	 * scrubber to reconstruct the block mappings.  The extended attribute
+	 * structure always contain some content when parent pointers are
+	 * enabled, so this is a clear sign of a zapped attr fork.
+	 */
+	return ip->i_af.if_format == XFS_DINODE_FMT_EXTENTS &&
+	       ip->i_af.if_nextents == 0;
+}
diff --git a/fs/xfs/scrub/parent_repair.c b/fs/xfs/scrub/parent_repair.c
new file mode 100644
index 0000000000..7b42b7f65a
--- /dev/null
+++ b/fs/xfs/scrub/parent_repair.c
@@ -0,0 +1,1612 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2020-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_icache.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
+#include "xfs_dir2.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_dir2_priv.h"
+#include "xfs_trans_space.h"
+#include "xfs_health.h"
+#include "xfs_exchmaps.h"
+#include "xfs_parent.h"
+#include "xfs_attr.h"
+#include "xfs_bmap.h"
+#include "xfs_ag.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+#include "scrub/repair.h"
+#include "scrub/iscan.h"
+#include "scrub/findparent.h"
+#include "scrub/readdir.h"
+#include "scrub/tempfile.h"
+#include "scrub/tempexch.h"
+#include "scrub/orphanage.h"
+#include "scrub/xfile.h"
+#include "scrub/xfarray.h"
+#include "scrub/xfblob.h"
+#include "scrub/attr_repair.h"
+#include "scrub/listxattr.h"
+
+/*
+ * Repairing The Directory Parent Pointer
+ * ======================================
+ *
+ * Currently, only directories support parent pointers (in the form of '..'
+ * entries), so we simply scan the filesystem and update the '..' entry.
+ *
+ * Note that because the only parent pointer is the dotdot entry, we won't
+ * touch an unhealthy directory, since the directory repair code is perfectly
+ * capable of rebuilding a directory with the proper parent inode.
+ *
+ * See the section on locking issues in dir_repair.c for more information about
+ * conflicts with the VFS.  The findparent code wll keep our incore parent
+ * inode up to date.
+ *
+ * If parent pointers are enabled, we instead reconstruct the parent pointer
+ * information by visiting every directory entry of every directory in the
+ * system and translating the relevant dirents into parent pointers.  In this
+ * case, it is advantageous to stash all parent pointers created from dirents
+ * from a single parent file before replaying them into the temporary file.  To
+ * save memory, the live filesystem scan reuses the findparent object.  Parent
+ * pointer repair chooses either directory scanning or findparent, but not
+ * both.
+ *
+ * When salvaging completes, the remaining stashed entries are replayed to the
+ * temporary file.  All non-parent pointer extended attributes are copied to
+ * the temporary file's extended attributes.  An atomic file mapping exchange
+ * is used to commit the new xattr blocks to the file being repaired.  This
+ * will disrupt attrmulti cursors.
+ */
+
+/* Create a parent pointer in the tempfile. */
+#define XREP_PPTR_ADD		(1)
+
+/* Remove a parent pointer from the tempfile. */
+#define XREP_PPTR_REMOVE	(2)
+
+/* A stashed parent pointer update. */
+struct xrep_pptr {
+	/* Cookie for retrieval of the pptr name. */
+	xfblob_cookie		name_cookie;
+
+	/* Parent pointer record. */
+	struct xfs_parent_rec	pptr_rec;
+
+	/* Length of the pptr name. */
+	uint8_t			namelen;
+
+	/* XREP_PPTR_{ADD,REMOVE} */
+	uint8_t			action;
+};
+
+/*
+ * Stash up to 8 pages of recovered parent pointers in pptr_recs and
+ * pptr_names before we write them to the temp file.
+ */
+#define XREP_PARENT_MAX_STASH_BYTES	(PAGE_SIZE * 8)
+
+struct xrep_parent {
+	struct xfs_scrub	*sc;
+
+	/* Fixed-size array of xrep_pptr structures. */
+	struct xfarray		*pptr_recs;
+
+	/* Blobs containing parent pointer names. */
+	struct xfblob		*pptr_names;
+
+	/* xattr keys */
+	struct xfarray		*xattr_records;
+
+	/* xattr values */
+	struct xfblob		*xattr_blobs;
+
+	/* Scratch buffers for saving extended attributes */
+	unsigned char		*xattr_name;
+	void			*xattr_value;
+	unsigned int		xattr_value_sz;
+
+	/*
+	 * Information used to exchange the attr fork mappings, if the fs
+	 * supports parent pointers.
+	 */
+	struct xrep_tempexch	tx;
+
+	/*
+	 * Information used to scan the filesystem to find the inumber of the
+	 * dotdot entry for this directory.  On filesystems without parent
+	 * pointers, we use the findparent_* functions on this object and
+	 * access only the parent_ino field directly.
+	 *
+	 * When parent pointers are enabled, the directory entry scanner uses
+	 * the iscan, hooks, and lock fields of this object directly.
+	 * @pscan.lock coordinates access to pptr_recs, pptr_names, pptr, and
+	 * pptr_scratch.  This reduces the memory requirements of this
+	 * structure.
+	 *
+	 * The lock also controls access to xattr_records and xattr_blobs(?)
+	 */
+	struct xrep_parent_scan_info pscan;
+
+	/* Orphanage reparenting request. */
+	struct xrep_adoption	adoption;
+
+	/* Directory entry name, plus the trailing null. */
+	struct xfs_name		xname;
+	unsigned char		namebuf[MAXNAMELEN];
+
+	/* Scratch buffer for scanning pptr xattrs */
+	struct xfs_da_args	pptr_args;
+
+	/* Have we seen any live updates of parent pointers recently? */
+	bool			saw_pptr_updates;
+
+	/* Number of parents we found after all other repairs */
+	unsigned long long	parents;
+};
+
+struct xrep_parent_xattr {
+	/* Cookie for retrieval of the xattr name. */
+	xfblob_cookie		name_cookie;
+
+	/* Cookie for retrieval of the xattr value. */
+	xfblob_cookie		value_cookie;
+
+	/* XFS_ATTR_* flags */
+	int			flags;
+
+	/* Length of the value and name. */
+	uint32_t		valuelen;
+	uint16_t		namelen;
+};
+
+/*
+ * Stash up to 8 pages of attrs in xattr_records/xattr_blobs before we write
+ * them to the temp file.
+ */
+#define XREP_PARENT_XATTR_MAX_STASH_BYTES	(PAGE_SIZE * 8)
+
+/* Tear down all the incore stuff we created. */
+static void
+xrep_parent_teardown(
+	struct xrep_parent	*rp)
+{
+	xrep_findparent_scan_teardown(&rp->pscan);
+	kvfree(rp->xattr_name);
+	rp->xattr_name = NULL;
+	kvfree(rp->xattr_value);
+	rp->xattr_value = NULL;
+	if (rp->xattr_blobs)
+		xfblob_destroy(rp->xattr_blobs);
+	rp->xattr_blobs = NULL;
+	if (rp->xattr_records)
+		xfarray_destroy(rp->xattr_records);
+	rp->xattr_records = NULL;
+	if (rp->pptr_names)
+		xfblob_destroy(rp->pptr_names);
+	rp->pptr_names = NULL;
+	if (rp->pptr_recs)
+		xfarray_destroy(rp->pptr_recs);
+	rp->pptr_recs = NULL;
+}
+
+/* Set up for a parent repair. */
+int
+xrep_setup_parent(
+	struct xfs_scrub	*sc)
+{
+	struct xrep_parent	*rp;
+	int			error;
+
+	xchk_fsgates_enable(sc, XCHK_FSGATES_DIRENTS);
+
+	rp = kvzalloc(sizeof(struct xrep_parent), XCHK_GFP_FLAGS);
+	if (!rp)
+		return -ENOMEM;
+	rp->sc = sc;
+	rp->xname.name = rp->namebuf;
+	sc->buf = rp;
+
+	error = xrep_tempfile_create(sc, S_IFREG);
+	if (error)
+		return error;
+
+	return xrep_orphanage_try_create(sc);
+}
+
+/*
+ * Scan all files in the filesystem for a child dirent that we can turn into
+ * the dotdot entry for this directory.
+ */
+STATIC int
+xrep_parent_find_dotdot(
+	struct xrep_parent	*rp)
+{
+	struct xfs_scrub	*sc = rp->sc;
+	xfs_ino_t		ino;
+	unsigned int		sick, checked;
+	int			error;
+
+	/*
+	 * Avoid sick directories.  There shouldn't be anyone else clearing the
+	 * directory's sick status.
+	 */
+	xfs_inode_measure_sickness(sc->ip, &sick, &checked);
+	if (sick & XFS_SICK_INO_DIR)
+		return -EFSCORRUPTED;
+
+	ino = xrep_findparent_self_reference(sc);
+	if (ino != NULLFSINO) {
+		xrep_findparent_scan_finish_early(&rp->pscan, ino);
+		return 0;
+	}
+
+	/*
+	 * Drop the ILOCK on this directory so that we can scan for the dotdot
+	 * entry.  Figure out who is going to be the parent of this directory,
+	 * then retake the ILOCK so that we can salvage directory entries.
+	 */
+	xchk_iunlock(sc, XFS_ILOCK_EXCL);
+
+	/* Does the VFS dcache have an answer for us? */
+	ino = xrep_findparent_from_dcache(sc);
+	if (ino != NULLFSINO) {
+		error = xrep_findparent_confirm(sc, &ino);
+		if (!error && ino != NULLFSINO) {
+			xrep_findparent_scan_finish_early(&rp->pscan, ino);
+			goto out_relock;
+		}
+	}
+
+	/* Scan the entire filesystem for a parent. */
+	error = xrep_findparent_scan(&rp->pscan);
+out_relock:
+	xchk_ilock(sc, XFS_ILOCK_EXCL);
+
+	return error;
+}
+
+/*
+ * Add this stashed incore parent pointer to the temporary file.
+ * The caller must hold the tempdir's IOLOCK, must not hold any ILOCKs, and
+ * must not be in transaction context.
+ */
+STATIC int
+xrep_parent_replay_update(
+	struct xrep_parent	*rp,
+	const struct xfs_name	*xname,
+	struct xrep_pptr	*pptr)
+{
+	struct xfs_scrub	*sc = rp->sc;
+
+	switch (pptr->action) {
+	case XREP_PPTR_ADD:
+		/* Create parent pointer. */
+		trace_xrep_parent_replay_parentadd(sc->tempip, xname,
+				&pptr->pptr_rec);
+
+		return xfs_parent_set(sc->tempip, sc->ip->i_ino, xname,
+				&pptr->pptr_rec, &rp->pptr_args);
+	case XREP_PPTR_REMOVE:
+		/* Remove parent pointer. */
+		trace_xrep_parent_replay_parentremove(sc->tempip, xname,
+				&pptr->pptr_rec);
+
+		return xfs_parent_unset(sc->tempip, sc->ip->i_ino, xname,
+				&pptr->pptr_rec, &rp->pptr_args);
+	}
+
+	ASSERT(0);
+	return -EIO;
+}
+
+/*
+ * Flush stashed parent pointer updates that have been recorded by the scanner.
+ * This is done to reduce the memory requirements of the parent pointer
+ * rebuild, since files can have a lot of hardlinks and the fs can be busy.
+ *
+ * Caller must not hold transactions or ILOCKs.  Caller must hold the tempfile
+ * IOLOCK.
+ */
+STATIC int
+xrep_parent_replay_updates(
+	struct xrep_parent	*rp)
+{
+	xfarray_idx_t		array_cur;
+	int			error;
+
+	mutex_lock(&rp->pscan.lock);
+	foreach_xfarray_idx(rp->pptr_recs, array_cur) {
+		struct xrep_pptr	pptr;
+
+		error = xfarray_load(rp->pptr_recs, array_cur, &pptr);
+		if (error)
+			goto out_unlock;
+
+		error = xfblob_loadname(rp->pptr_names, pptr.name_cookie,
+				&rp->xname, pptr.namelen);
+		if (error)
+			goto out_unlock;
+		rp->xname.len = pptr.namelen;
+		mutex_unlock(&rp->pscan.lock);
+
+		error = xrep_parent_replay_update(rp, &rp->xname, &pptr);
+		if (error)
+			return error;
+
+		mutex_lock(&rp->pscan.lock);
+	}
+
+	/* Empty out both arrays now that we've added the entries. */
+	xfarray_truncate(rp->pptr_recs);
+	xfblob_truncate(rp->pptr_names);
+	mutex_unlock(&rp->pscan.lock);
+	return 0;
+out_unlock:
+	mutex_unlock(&rp->pscan.lock);
+	return error;
+}
+
+/*
+ * Remember that we want to create a parent pointer in the tempfile.  These
+ * stashed actions will be replayed later.
+ */
+STATIC int
+xrep_parent_stash_parentadd(
+	struct xrep_parent	*rp,
+	const struct xfs_name	*name,
+	const struct xfs_inode	*dp)
+{
+	struct xrep_pptr	pptr = {
+		.action		= XREP_PPTR_ADD,
+		.namelen	= name->len,
+	};
+	int			error;
+
+	trace_xrep_parent_stash_parentadd(rp->sc->tempip, dp, name);
+
+	xfs_inode_to_parent_rec(&pptr.pptr_rec, dp);
+	error = xfblob_storename(rp->pptr_names, &pptr.name_cookie, name);
+	if (error)
+		return error;
+
+	return xfarray_append(rp->pptr_recs, &pptr);
+}
+
+/*
+ * Remember that we want to remove a parent pointer from the tempfile.  These
+ * stashed actions will be replayed later.
+ */
+STATIC int
+xrep_parent_stash_parentremove(
+	struct xrep_parent	*rp,
+	const struct xfs_name	*name,
+	const struct xfs_inode	*dp)
+{
+	struct xrep_pptr	pptr = {
+		.action		= XREP_PPTR_REMOVE,
+		.namelen	= name->len,
+	};
+	int			error;
+
+	trace_xrep_parent_stash_parentremove(rp->sc->tempip, dp, name);
+
+	xfs_inode_to_parent_rec(&pptr.pptr_rec, dp);
+	error = xfblob_storename(rp->pptr_names, &pptr.name_cookie, name);
+	if (error)
+		return error;
+
+	return xfarray_append(rp->pptr_recs, &pptr);
+}
+
+/*
+ * Examine an entry of a directory.  If this dirent leads us back to the file
+ * whose parent pointers we're rebuilding, add a pptr to the temporary
+ * directory.
+ */
+STATIC int
+xrep_parent_scan_dirent(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*dp,
+	xfs_dir2_dataptr_t	dapos,
+	const struct xfs_name	*name,
+	xfs_ino_t		ino,
+	void			*priv)
+{
+	struct xrep_parent	*rp = priv;
+	int			error;
+
+	/* Dirent doesn't point to this directory. */
+	if (ino != rp->sc->ip->i_ino)
+		return 0;
+
+	/* No weird looking names. */
+	if (name->len == 0 || !xfs_dir2_namecheck(name->name, name->len))
+		return -EFSCORRUPTED;
+
+	/* No mismatching ftypes. */
+	if (name->type != xfs_mode_to_ftype(VFS_I(sc->ip)->i_mode))
+		return -EFSCORRUPTED;
+
+	/* Don't pick up dot or dotdot entries; we only want child dirents. */
+	if (xfs_dir2_samename(name, &xfs_name_dotdot) ||
+	    xfs_dir2_samename(name, &xfs_name_dot))
+		return 0;
+
+	/*
+	 * Transform this dirent into a parent pointer and queue it for later
+	 * addition to the temporary file.
+	 */
+	mutex_lock(&rp->pscan.lock);
+	error = xrep_parent_stash_parentadd(rp, name, dp);
+	mutex_unlock(&rp->pscan.lock);
+	return error;
+}
+
+/*
+ * Decide if we want to look for dirents in this directory.  Skip the file
+ * being repaired and any files being used to stage repairs.
+ */
+static inline bool
+xrep_parent_want_scan(
+	struct xrep_parent	*rp,
+	const struct xfs_inode	*ip)
+{
+	return ip != rp->sc->ip && !xrep_is_tempfile(ip);
+}
+
+/*
+ * Take ILOCK on a file that we want to scan.
+ *
+ * Select ILOCK_EXCL if the file is a directory with an unloaded data bmbt.
+ * Otherwise, take ILOCK_SHARED.
+ */
+static inline unsigned int
+xrep_parent_scan_ilock(
+	struct xrep_parent	*rp,
+	struct xfs_inode	*ip)
+{
+	uint			lock_mode = XFS_ILOCK_SHARED;
+
+	/* Still need to take the shared ILOCK to advance the iscan cursor. */
+	if (!xrep_parent_want_scan(rp, ip))
+		goto lock;
+
+	if (S_ISDIR(VFS_I(ip)->i_mode) && xfs_need_iread_extents(&ip->i_df)) {
+		lock_mode = XFS_ILOCK_EXCL;
+		goto lock;
+	}
+
+lock:
+	xfs_ilock(ip, lock_mode);
+	return lock_mode;
+}
+
+/*
+ * Scan this file for relevant child dirents that point to the file whose
+ * parent pointers we're rebuilding.
+ */
+STATIC int
+xrep_parent_scan_file(
+	struct xrep_parent	*rp,
+	struct xfs_inode	*ip)
+{
+	unsigned int		lock_mode;
+	int			error = 0;
+
+	lock_mode = xrep_parent_scan_ilock(rp, ip);
+
+	if (!xrep_parent_want_scan(rp, ip))
+		goto scan_done;
+
+	if (S_ISDIR(VFS_I(ip)->i_mode)) {
+		/*
+		 * If the directory looks as though it has been zapped by the
+		 * inode record repair code, we cannot scan for child dirents.
+		 */
+		if (xchk_dir_looks_zapped(ip)) {
+			error = -EBUSY;
+			goto scan_done;
+		}
+
+		error = xchk_dir_walk(rp->sc, ip, xrep_parent_scan_dirent, rp);
+		if (error)
+			goto scan_done;
+	}
+
+scan_done:
+	xchk_iscan_mark_visited(&rp->pscan.iscan, ip);
+	xfs_iunlock(ip, lock_mode);
+	return error;
+}
+
+/* Decide if we've stashed too much pptr data in memory. */
+static inline bool
+xrep_parent_want_flush_stashed(
+	struct xrep_parent	*rp)
+{
+	unsigned long long	bytes;
+
+	bytes = xfarray_bytes(rp->pptr_recs) + xfblob_bytes(rp->pptr_names);
+	return bytes > XREP_PARENT_MAX_STASH_BYTES;
+}
+
+/*
+ * Scan all directories in the filesystem to look for dirents that we can turn
+ * into parent pointers.
+ */
+STATIC int
+xrep_parent_scan_dirtree(
+	struct xrep_parent	*rp)
+{
+	struct xfs_scrub	*sc = rp->sc;
+	struct xfs_inode	*ip;
+	int			error;
+
+	/*
+	 * Filesystem scans are time consuming.  Drop the file ILOCK and all
+	 * other resources for the duration of the scan and hope for the best.
+	 * The live update hooks will keep our scan information up to date.
+	 */
+	xchk_trans_cancel(sc);
+	if (sc->ilock_flags & (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL))
+		xchk_iunlock(sc, sc->ilock_flags & (XFS_ILOCK_SHARED |
+						    XFS_ILOCK_EXCL));
+	error = xchk_trans_alloc_empty(sc);
+	if (error)
+		return error;
+
+	while ((error = xchk_iscan_iter(&rp->pscan.iscan, &ip)) == 1) {
+		bool		flush;
+
+		error = xrep_parent_scan_file(rp, ip);
+		xchk_irele(sc, ip);
+		if (error)
+			break;
+
+		/* Flush stashed pptr updates to constrain memory usage. */
+		mutex_lock(&rp->pscan.lock);
+		flush = xrep_parent_want_flush_stashed(rp);
+		mutex_unlock(&rp->pscan.lock);
+		if (flush) {
+			xchk_trans_cancel(sc);
+
+			error = xrep_tempfile_iolock_polled(sc);
+			if (error)
+				break;
+
+			error = xrep_parent_replay_updates(rp);
+			xrep_tempfile_iounlock(sc);
+			if (error)
+				break;
+
+			error = xchk_trans_alloc_empty(sc);
+			if (error)
+				break;
+		}
+
+		if (xchk_should_terminate(sc, &error))
+			break;
+	}
+	xchk_iscan_iter_finish(&rp->pscan.iscan);
+	if (error) {
+		/*
+		 * If we couldn't grab an inode that was busy with a state
+		 * change, change the error code so that we exit to userspace
+		 * as quickly as possible.
+		 */
+		if (error == -EBUSY)
+			return -ECANCELED;
+		return error;
+	}
+
+	/*
+	 * Retake sc->ip's ILOCK now that we're done flushing stashed parent
+	 * pointers.  We end this function with an empty transaction and the
+	 * ILOCK.
+	 */
+	xchk_ilock(rp->sc, XFS_ILOCK_EXCL);
+	return 0;
+}
+
+/*
+ * Capture dirent updates being made by other threads which are relevant to the
+ * file being repaired.
+ */
+STATIC int
+xrep_parent_live_update(
+	struct notifier_block		*nb,
+	unsigned long			action,
+	void				*data)
+{
+	struct xfs_dir_update_params	*p = data;
+	struct xrep_parent		*rp;
+	struct xfs_scrub		*sc;
+	int				error;
+
+	rp = container_of(nb, struct xrep_parent, pscan.dhook.dirent_hook.nb);
+	sc = rp->sc;
+
+	/*
+	 * This thread updated a dirent that points to the file that we're
+	 * repairing, so stash the update for replay against the temporary
+	 * file.
+	 */
+	if (p->ip->i_ino == sc->ip->i_ino &&
+	    xchk_iscan_want_live_update(&rp->pscan.iscan, p->dp->i_ino)) {
+		mutex_lock(&rp->pscan.lock);
+		if (p->delta > 0)
+			error = xrep_parent_stash_parentadd(rp, p->name, p->dp);
+		else
+			error = xrep_parent_stash_parentremove(rp, p->name,
+					p->dp);
+		if (!error)
+			rp->saw_pptr_updates = true;
+		mutex_unlock(&rp->pscan.lock);
+		if (error)
+			goto out_abort;
+	}
+
+	return NOTIFY_DONE;
+out_abort:
+	xchk_iscan_abort(&rp->pscan.iscan);
+	return NOTIFY_DONE;
+}
+
+/* Reset a directory's dotdot entry, if needed. */
+STATIC int
+xrep_parent_reset_dotdot(
+	struct xrep_parent	*rp)
+{
+	struct xfs_scrub	*sc = rp->sc;
+	xfs_ino_t		ino;
+	unsigned int		spaceres;
+	int			error = 0;
+
+	ASSERT(sc->ilock_flags & XFS_ILOCK_EXCL);
+
+	error = xchk_dir_lookup(sc, sc->ip, &xfs_name_dotdot, &ino);
+	if (error || ino == rp->pscan.parent_ino)
+		return error;
+
+	xfs_trans_ijoin(sc->tp, sc->ip, 0);
+
+	trace_xrep_parent_reset_dotdot(sc->ip, rp->pscan.parent_ino);
+
+	/*
+	 * Reserve more space just in case we have to expand the dir.  We're
+	 * allowed to exceed quota to repair inconsistent metadata.
+	 */
+	spaceres = xfs_rename_space_res(sc->mp, 0, false, xfs_name_dotdot.len,
+			false);
+	error = xfs_trans_reserve_more_inode(sc->tp, sc->ip, spaceres, 0,
+			true);
+	if (error)
+		return error;
+
+	error = xfs_dir_replace(sc->tp, sc->ip, &xfs_name_dotdot,
+			rp->pscan.parent_ino, spaceres);
+	if (error)
+		return error;
+
+	/*
+	 * Roll transaction to detach the inode from the transaction but retain
+	 * ILOCK_EXCL.
+	 */
+	return xfs_trans_roll(&sc->tp);
+}
+
+/* Pass back the parent inumber if this a parent pointer */
+STATIC int
+xrep_parent_lookup_pptr(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*ip,
+	unsigned int		attr_flags,
+	const unsigned char	*name,
+	unsigned int		namelen,
+	const void		*value,
+	unsigned int		valuelen,
+	void			*priv)
+{
+	xfs_ino_t		*inop = priv;
+	xfs_ino_t		parent_ino;
+	int			error;
+
+	if (!(attr_flags & XFS_ATTR_PARENT))
+		return 0;
+
+	error = xfs_parent_from_attr(sc->mp, attr_flags, name, namelen, value,
+			valuelen, &parent_ino, NULL);
+	if (error)
+		return error;
+
+	*inop = parent_ino;
+	return -ECANCELED;
+}
+
+/*
+ * Find the first parent of the scrub target by walking parent pointers for
+ * the purpose of deciding if we're going to move it to the orphanage.
+ * We don't care if the attr fork is zapped.
+ */
+STATIC int
+xrep_parent_lookup_pptrs(
+	struct xfs_scrub	*sc,
+	xfs_ino_t		*inop)
+{
+	int			error;
+
+	*inop = NULLFSINO;
+
+	error = xchk_xattr_walk(sc, sc->ip, xrep_parent_lookup_pptr, NULL,
+			inop);
+	if (error && error != -ECANCELED)
+		return error;
+	return 0;
+}
+
+/*
+ * Move the current file to the orphanage.
+ *
+ * Caller must hold IOLOCK_EXCL on @sc->ip, and no other inode locks.  Upon
+ * successful return, the scrub transaction will have enough extra reservation
+ * to make the move; it will hold IOLOCK_EXCL and ILOCK_EXCL of @sc->ip and the
+ * orphanage; and both inodes will be ijoined.
+ */
+STATIC int
+xrep_parent_move_to_orphanage(
+	struct xrep_parent	*rp)
+{
+	struct xfs_scrub	*sc = rp->sc;
+	xfs_ino_t		orig_parent, new_parent;
+	int			error;
+
+	if (S_ISDIR(VFS_I(sc->ip)->i_mode)) {
+		/*
+		 * We are about to drop the ILOCK on sc->ip to lock the
+		 * orphanage and prepare for the adoption.  Therefore, look up
+		 * the old dotdot entry for sc->ip so that we can compare it
+		 * after we re-lock sc->ip.
+		 */
+		error = xchk_dir_lookup(sc, sc->ip, &xfs_name_dotdot,
+				&orig_parent);
+		if (error)
+			return error;
+	} else {
+		/*
+		 * We haven't dropped the ILOCK since we committed the new
+		 * xattr structure (and hence the new parent pointer records),
+		 * which means that the file cannot have been moved in the
+		 * directory tree, and there are no parents.
+		 */
+		orig_parent = NULLFSINO;
+	}
+
+	/*
+	 * Drop the ILOCK on the scrub target and commit the transaction.
+	 * Adoption computes its own resource requirements and gathers the
+	 * necessary components.
+	 */
+	error = xrep_trans_commit(sc);
+	if (error)
+		return error;
+	xchk_iunlock(sc, XFS_ILOCK_EXCL);
+
+	/* If we can take the orphanage's iolock then we're ready to move. */
+	if (!xrep_orphanage_ilock_nowait(sc, XFS_IOLOCK_EXCL)) {
+		xchk_iunlock(sc, sc->ilock_flags);
+		error = xrep_orphanage_iolock_two(sc);
+		if (error)
+			return error;
+	}
+
+	/* Grab transaction and ILOCK the two files. */
+	error = xrep_adoption_trans_alloc(sc, &rp->adoption);
+	if (error)
+		return error;
+
+	error = xrep_adoption_compute_name(&rp->adoption, &rp->xname);
+	if (error)
+		return error;
+
+	/*
+	 * Now that we've reacquired the ILOCK on sc->ip, look up the dotdot
+	 * entry again.  If the parent changed or the child was unlinked while
+	 * the child directory was unlocked, we don't need to move the child to
+	 * the orphanage after all.  For a non-directory, we have to scan for
+	 * the first parent pointer to see if one has been added.
+	 */
+	if (S_ISDIR(VFS_I(sc->ip)->i_mode))
+		error = xchk_dir_lookup(sc, sc->ip, &xfs_name_dotdot,
+				&new_parent);
+	else
+		error = xrep_parent_lookup_pptrs(sc, &new_parent);
+	if (error)
+		return error;
+
+	/*
+	 * Attach to the orphanage if we still have a linked directory and it
+	 * hasn't been moved.
+	 */
+	if (orig_parent == new_parent && VFS_I(sc->ip)->i_nlink > 0) {
+		error = xrep_adoption_move(&rp->adoption);
+		if (error)
+			return error;
+	}
+
+	/*
+	 * Launder the scrub transaction so we can drop the orphanage ILOCK
+	 * and IOLOCK.  Return holding the scrub target's ILOCK and IOLOCK.
+	 */
+	error = xrep_adoption_trans_roll(&rp->adoption);
+	if (error)
+		return error;
+
+	xrep_orphanage_iunlock(sc, XFS_ILOCK_EXCL);
+	xrep_orphanage_iunlock(sc, XFS_IOLOCK_EXCL);
+	return 0;
+}
+
+/* Ensure that the xattr value buffer is large enough. */
+STATIC int
+xrep_parent_alloc_xattr_value(
+	struct xrep_parent	*rp,
+	size_t			bufsize)
+{
+	void			*new_val;
+
+	if (rp->xattr_value_sz >= bufsize)
+		return 0;
+
+	if (rp->xattr_value) {
+		kvfree(rp->xattr_value);
+		rp->xattr_value = NULL;
+		rp->xattr_value_sz = 0;
+	}
+
+	new_val = kvmalloc(bufsize, XCHK_GFP_FLAGS);
+	if (!new_val)
+		return -ENOMEM;
+
+	rp->xattr_value = new_val;
+	rp->xattr_value_sz = bufsize;
+	return 0;
+}
+
+/* Retrieve the (remote) value of a non-pptr xattr. */
+STATIC int
+xrep_parent_fetch_xattr_remote(
+	struct xrep_parent	*rp,
+	struct xfs_inode	*ip,
+	unsigned int		attr_flags,
+	const unsigned char	*name,
+	unsigned int		namelen,
+	unsigned int		valuelen)
+{
+	struct xfs_scrub	*sc = rp->sc;
+	struct xfs_da_args	args = {
+		.attr_filter	= attr_flags & XFS_ATTR_NSP_ONDISK_MASK,
+		.geo		= sc->mp->m_attr_geo,
+		.whichfork	= XFS_ATTR_FORK,
+		.dp		= ip,
+		.name		= name,
+		.namelen	= namelen,
+		.trans		= sc->tp,
+		.valuelen	= valuelen,
+		.owner		= ip->i_ino,
+	};
+	int			error;
+
+	/*
+	 * If we need a larger value buffer, try to allocate one.  If that
+	 * fails, return with -EDEADLOCK to try harder.
+	 */
+	error = xrep_parent_alloc_xattr_value(rp, valuelen);
+	if (error == -ENOMEM)
+		return -EDEADLOCK;
+	if (error)
+		return error;
+
+	args.value = rp->xattr_value;
+	xfs_attr_sethash(&args);
+	return xfs_attr_get_ilocked(&args);
+}
+
+/* Stash non-pptr attributes for later replay into the temporary file. */
+STATIC int
+xrep_parent_stash_xattr(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*ip,
+	unsigned int		attr_flags,
+	const unsigned char	*name,
+	unsigned int		namelen,
+	const void		*value,
+	unsigned int		valuelen,
+	void			*priv)
+{
+	struct xrep_parent_xattr key = {
+		.valuelen	= valuelen,
+		.namelen	= namelen,
+		.flags		= attr_flags & XFS_ATTR_NSP_ONDISK_MASK,
+	};
+	struct xrep_parent	*rp = priv;
+	int			error;
+
+	if (attr_flags & (XFS_ATTR_INCOMPLETE | XFS_ATTR_PARENT))
+		return 0;
+
+	if (!value) {
+		error = xrep_parent_fetch_xattr_remote(rp, ip, attr_flags,
+				name, namelen, valuelen);
+		if (error)
+			return error;
+
+		value = rp->xattr_value;
+	}
+
+	trace_xrep_parent_stash_xattr(rp->sc->tempip, key.flags, (void *)name,
+			key.namelen, key.valuelen);
+
+	error = xfblob_store(rp->xattr_blobs, &key.name_cookie, name,
+			key.namelen);
+	if (error)
+		return error;
+
+	error = xfblob_store(rp->xattr_blobs, &key.value_cookie, value,
+			key.valuelen);
+	if (error)
+		return error;
+
+	return xfarray_append(rp->xattr_records, &key);
+}
+
+/* Insert one xattr key/value. */
+STATIC int
+xrep_parent_insert_xattr(
+	struct xrep_parent		*rp,
+	const struct xrep_parent_xattr	*key)
+{
+	struct xfs_da_args		args = {
+		.dp			= rp->sc->tempip,
+		.attr_filter		= key->flags,
+		.namelen		= key->namelen,
+		.valuelen		= key->valuelen,
+		.owner			= rp->sc->ip->i_ino,
+		.geo			= rp->sc->mp->m_attr_geo,
+		.whichfork		= XFS_ATTR_FORK,
+		.op_flags		= XFS_DA_OP_OKNOENT,
+	};
+	int				error;
+
+	ASSERT(!(key->flags & XFS_ATTR_PARENT));
+
+	/*
+	 * Grab pointers to the scrub buffer so that we can use them to insert
+	 * attrs into the temp file.
+	 */
+	args.name = rp->xattr_name;
+	args.value = rp->xattr_value;
+
+	/*
+	 * The attribute name is stored near the end of the in-core buffer,
+	 * though we reserve one more byte to ensure null termination.
+	 */
+	rp->xattr_name[XATTR_NAME_MAX] = 0;
+
+	error = xfblob_load(rp->xattr_blobs, key->name_cookie, rp->xattr_name,
+			key->namelen);
+	if (error)
+		return error;
+
+	error = xfblob_free(rp->xattr_blobs, key->name_cookie);
+	if (error)
+		return error;
+
+	error = xfblob_load(rp->xattr_blobs, key->value_cookie, args.value,
+			key->valuelen);
+	if (error)
+		return error;
+
+	error = xfblob_free(rp->xattr_blobs, key->value_cookie);
+	if (error)
+		return error;
+
+	rp->xattr_name[key->namelen] = 0;
+
+	trace_xrep_parent_insert_xattr(rp->sc->tempip, key->flags,
+			rp->xattr_name, key->namelen, key->valuelen);
+
+	xfs_attr_sethash(&args);
+	return xfs_attr_set(&args, XFS_ATTRUPDATE_UPSERT, false);
+}
+
+/*
+ * Periodically flush salvaged attributes to the temporary file.  This is done
+ * to reduce the memory requirements of the xattr rebuild because files can
+ * contain millions of attributes.
+ */
+STATIC int
+xrep_parent_flush_xattrs(
+	struct xrep_parent	*rp)
+{
+	xfarray_idx_t		array_cur;
+	int			error;
+
+	/*
+	 * Entering this function, the scrub context has a reference to the
+	 * inode being repaired, the temporary file, and the empty scrub
+	 * transaction that we created for the xattr scan.  We hold ILOCK_EXCL
+	 * on the inode being repaired.
+	 *
+	 * To constrain kernel memory use, we occasionally flush salvaged
+	 * xattrs from the xfarray and xfblob structures into the temporary
+	 * file in preparation for exchanging the xattr structures at the end.
+	 * Updating the temporary file requires a transaction, so we commit the
+	 * scrub transaction and drop the ILOCK so that xfs_attr_set can
+	 * allocate whatever transaction it wants.
+	 *
+	 * We still hold IOLOCK_EXCL on the inode being repaired, which
+	 * prevents anyone from adding xattrs (or parent pointers) while we're
+	 * flushing.
+	 */
+	xchk_trans_cancel(rp->sc);
+	xchk_iunlock(rp->sc, XFS_ILOCK_EXCL);
+
+	/*
+	 * Take the IOLOCK of the temporary file while we modify xattrs.  This
+	 * isn't strictly required because the temporary file is never revealed
+	 * to userspace, but we follow the same locking rules.  We still hold
+	 * sc->ip's IOLOCK.
+	 */
+	error = xrep_tempfile_iolock_polled(rp->sc);
+	if (error)
+		return error;
+
+	/* Add all the salvaged attrs to the temporary file. */
+	foreach_xfarray_idx(rp->xattr_records, array_cur) {
+		struct xrep_parent_xattr	key;
+
+		error = xfarray_load(rp->xattr_records, array_cur, &key);
+		if (error)
+			return error;
+
+		error = xrep_parent_insert_xattr(rp, &key);
+		if (error)
+			return error;
+	}
+
+	/* Empty out both arrays now that we've added the entries. */
+	xfarray_truncate(rp->xattr_records);
+	xfblob_truncate(rp->xattr_blobs);
+
+	xrep_tempfile_iounlock(rp->sc);
+
+	/* Recreate the empty transaction and relock the inode. */
+	error = xchk_trans_alloc_empty(rp->sc);
+	if (error)
+		return error;
+	xchk_ilock(rp->sc, XFS_ILOCK_EXCL);
+	return 0;
+}
+
+/* Decide if we've stashed too much xattr data in memory. */
+static inline bool
+xrep_parent_want_flush_xattrs(
+	struct xrep_parent	*rp)
+{
+	unsigned long long	bytes;
+
+	bytes = xfarray_bytes(rp->xattr_records) +
+		xfblob_bytes(rp->xattr_blobs);
+	return bytes > XREP_PARENT_XATTR_MAX_STASH_BYTES;
+}
+
+/* Flush staged attributes to the temporary file if we're over the limit. */
+STATIC int
+xrep_parent_try_flush_xattrs(
+	struct xfs_scrub	*sc,
+	void			*priv)
+{
+	struct xrep_parent	*rp = priv;
+	int			error;
+
+	if (!xrep_parent_want_flush_xattrs(rp))
+		return 0;
+
+	error = xrep_parent_flush_xattrs(rp);
+	if (error)
+		return error;
+
+	/*
+	 * If there were any parent pointer updates to the xattr structure
+	 * while we dropped the ILOCK, the xattr structure is now stale.
+	 * Signal to the attr copy process that we need to start over, but
+	 * this time without opportunistic attr flushing.
+	 *
+	 * This is unlikely to happen, so we're ok with restarting the copy.
+	 */
+	mutex_lock(&rp->pscan.lock);
+	if (rp->saw_pptr_updates)
+		error = -ESTALE;
+	mutex_unlock(&rp->pscan.lock);
+	return error;
+}
+
+/* Copy all the non-pptr extended attributes into the temporary file. */
+STATIC int
+xrep_parent_copy_xattrs(
+	struct xrep_parent	*rp)
+{
+	struct xfs_scrub	*sc = rp->sc;
+	int			error;
+
+	/*
+	 * Clear the pptr updates flag.  We hold sc->ip ILOCKed, so there
+	 * can't be any parent pointer updates in progress.
+	 */
+	mutex_lock(&rp->pscan.lock);
+	rp->saw_pptr_updates = false;
+	mutex_unlock(&rp->pscan.lock);
+
+	/* Copy xattrs, stopping periodically to flush the incore buffers. */
+	error = xchk_xattr_walk(sc, sc->ip, xrep_parent_stash_xattr,
+			xrep_parent_try_flush_xattrs, rp);
+	if (error && error != -ESTALE)
+		return error;
+
+	if (error == -ESTALE) {
+		/*
+		 * The xattr copy collided with a parent pointer update.
+		 * Restart the copy, but this time hold the ILOCK all the way
+		 * to the end to lock out any directory parent pointer updates.
+		 */
+		error = xchk_xattr_walk(sc, sc->ip, xrep_parent_stash_xattr,
+				NULL, rp);
+		if (error)
+			return error;
+	}
+
+	/* Flush any remaining stashed xattrs to the temporary file. */
+	if (xfarray_bytes(rp->xattr_records) == 0)
+		return 0;
+
+	return xrep_parent_flush_xattrs(rp);
+}
+
+/*
+ * Ensure that @sc->ip and @sc->tempip both have attribute forks before we head
+ * into the attr fork exchange transaction.  All files on a filesystem with
+ * parent pointers must have an attr fork because the parent pointer code does
+ * not itself add attribute forks.
+ *
+ * Note: Unlinkable unlinked files don't need one, but the overhead of having
+ * an unnecessary attr fork is not justified by the additional code complexity
+ * that would be needed to track that state correctly.
+ */
+STATIC int
+xrep_parent_ensure_attr_fork(
+	struct xrep_parent	*rp)
+{
+	struct xfs_scrub	*sc = rp->sc;
+	int			error;
+
+	error = xfs_attr_add_fork(sc->tempip,
+			sizeof(struct xfs_attr_sf_hdr), 1);
+	if (error)
+		return error;
+	return xfs_attr_add_fork(sc->ip, sizeof(struct xfs_attr_sf_hdr), 1);
+}
+
+/*
+ * Finish replaying stashed parent pointer updates, allocate a transaction for
+ * exchanging extent mappings, and take the ILOCKs of both files before we
+ * commit the new attribute structure.
+ */
+STATIC int
+xrep_parent_finalize_tempfile(
+	struct xrep_parent	*rp)
+{
+	struct xfs_scrub	*sc = rp->sc;
+	int			error;
+
+	/*
+	 * Repair relies on the ILOCK to quiesce all possible xattr updates.
+	 * Replay all queued parent pointer updates into the tempfile before
+	 * exchanging the contents, even if that means dropping the ILOCKs and
+	 * the transaction.
+	 */
+	do {
+		error = xrep_parent_replay_updates(rp);
+		if (error)
+			return error;
+
+		error = xrep_parent_ensure_attr_fork(rp);
+		if (error)
+			return error;
+
+		error = xrep_tempexch_trans_alloc(sc, XFS_ATTR_FORK, &rp->tx);
+		if (error)
+			return error;
+
+		if (xfarray_length(rp->pptr_recs) == 0)
+			break;
+
+		xchk_trans_cancel(sc);
+		xrep_tempfile_iunlock_both(sc);
+	} while (!xchk_should_terminate(sc, &error));
+	return error;
+}
+
+/*
+ * Replay all the stashed parent pointers into the temporary file, copy all
+ * the non-pptr xattrs from the file being repaired into the temporary file,
+ * and exchange the attr fork contents atomically.
+ */
+STATIC int
+xrep_parent_rebuild_pptrs(
+	struct xrep_parent	*rp)
+{
+	struct xfs_scrub	*sc = rp->sc;
+	xfs_ino_t		parent_ino = NULLFSINO;
+	int			error;
+
+	/*
+	 * Copy non-ppttr xattrs from the file being repaired into the
+	 * temporary file's xattr structure.  We hold sc->ip's IOLOCK, which
+	 * prevents setxattr/removexattr calls from occurring, but renames
+	 * update the parent pointers without holding IOLOCK.  If we detect
+	 * stale attr structures, we restart the scan but only flush at the
+	 * end.
+	 */
+	error = xrep_parent_copy_xattrs(rp);
+	if (error)
+		return error;
+
+	/*
+	 * Cancel the empty transaction that we used to walk and copy attrs,
+	 * and drop the ILOCK so that we can take the IOLOCK on the temporary
+	 * file.  We still hold sc->ip's IOLOCK.
+	 */
+	xchk_trans_cancel(sc);
+	xchk_iunlock(sc, XFS_ILOCK_EXCL);
+
+	error = xrep_tempfile_iolock_polled(sc);
+	if (error)
+		return error;
+
+	/*
+	 * Allocate transaction, lock inodes, and make sure that we've replayed
+	 * all the stashed pptr updates to the tempdir.  After this point,
+	 * we're ready to exchange the attr fork mappings.
+	 */
+	error = xrep_parent_finalize_tempfile(rp);
+	if (error)
+		return error;
+
+	/* Last chance to abort before we start committing pptr fixes. */
+	if (xchk_should_terminate(sc, &error))
+		return error;
+
+	if (xchk_iscan_aborted(&rp->pscan.iscan))
+		return -ECANCELED;
+
+	/*
+	 * Exchange the attr fork contents and junk the old attr fork contents,
+	 * which are now in the tempfile.
+	 */
+	error = xrep_xattr_swap(sc, &rp->tx);
+	if (error)
+		return error;
+	error = xrep_xattr_reset_tempfile_fork(sc);
+	if (error)
+		return error;
+
+	/*
+	 * Roll to get a transaction without any inodes joined to it.  Then we
+	 * can drop the tempfile's ILOCK and IOLOCK before doing more work on
+	 * the scrub target file.
+	 */
+	error = xfs_trans_roll(&sc->tp);
+	if (error)
+		return error;
+	xrep_tempfile_iunlock(sc);
+	xrep_tempfile_iounlock(sc);
+
+	/*
+	 * We've committed the new parent pointers.  Find at least one parent
+	 * so that we can decide if we're moving this file to the orphanage.
+	 * For this purpose, root directories are their own parents.
+	 */
+	if (sc->ip == sc->mp->m_rootip) {
+		xrep_findparent_scan_found(&rp->pscan, sc->ip->i_ino);
+	} else {
+		error = xrep_parent_lookup_pptrs(sc, &parent_ino);
+		if (error)
+			return error;
+		if (parent_ino != NULLFSINO)
+			xrep_findparent_scan_found(&rp->pscan, parent_ino);
+	}
+	return 0;
+}
+
+/*
+ * Commit the new parent pointer structure (currently only the dotdot entry) to
+ * the file that we're repairing.
+ */
+STATIC int
+xrep_parent_rebuild_tree(
+	struct xrep_parent	*rp)
+{
+	int			error;
+
+	if (xfs_has_parent(rp->sc->mp)) {
+		error = xrep_parent_rebuild_pptrs(rp);
+		if (error)
+			return error;
+	}
+
+	if (rp->pscan.parent_ino == NULLFSINO) {
+		if (xrep_orphanage_can_adopt(rp->sc))
+			return xrep_parent_move_to_orphanage(rp);
+		return -EFSCORRUPTED;
+	}
+
+	if (S_ISDIR(VFS_I(rp->sc->ip)->i_mode))
+		return xrep_parent_reset_dotdot(rp);
+
+	return 0;
+}
+
+/* Count the number of parent pointers. */
+STATIC int
+xrep_parent_count_pptr(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*ip,
+	unsigned int		attr_flags,
+	const unsigned char	*name,
+	unsigned int		namelen,
+	const void		*value,
+	unsigned int		valuelen,
+	void			*priv)
+{
+	struct xrep_parent	*rp = priv;
+	int			error;
+
+	if (!(attr_flags & XFS_ATTR_PARENT))
+		return 0;
+
+	error = xfs_parent_from_attr(sc->mp, attr_flags, name, namelen, value,
+			valuelen, NULL, NULL);
+	if (error)
+		return error;
+
+	rp->parents++;
+	return 0;
+}
+
+/*
+ * After all parent pointer rebuilding and adoption activity completes, reset
+ * the link count of this nondirectory, having scanned the fs to rebuild all
+ * parent pointers.
+ */
+STATIC int
+xrep_parent_set_nondir_nlink(
+	struct xrep_parent	*rp)
+{
+	struct xfs_scrub	*sc = rp->sc;
+	struct xfs_inode	*ip = sc->ip;
+	struct xfs_perag	*pag;
+	bool			joined = false;
+	int			error;
+
+	/* Count parent pointers so we can reset the file link count. */
+	rp->parents = 0;
+	error = xchk_xattr_walk(sc, ip, xrep_parent_count_pptr, NULL, rp);
+	if (error)
+		return error;
+
+	if (rp->parents > 0 && xfs_inode_on_unlinked_list(ip)) {
+		xfs_trans_ijoin(sc->tp, sc->ip, 0);
+		joined = true;
+
+		/*
+		 * The file is on the unlinked list but we found parents.
+		 * Remove the file from the unlinked list.
+		 */
+		pag = xfs_perag_get(sc->mp, XFS_INO_TO_AGNO(sc->mp, ip->i_ino));
+		if (!pag) {
+			ASSERT(0);
+			return -EFSCORRUPTED;
+		}
+
+		error = xfs_iunlink_remove(sc->tp, pag, ip);
+		xfs_perag_put(pag);
+		if (error)
+			return error;
+	} else if (rp->parents == 0 && !xfs_inode_on_unlinked_list(ip)) {
+		xfs_trans_ijoin(sc->tp, sc->ip, 0);
+		joined = true;
+
+		/*
+		 * The file is not on the unlinked list but we found no
+		 * parents.  Add the file to the unlinked list.
+		 */
+		error = xfs_iunlink(sc->tp, ip);
+		if (error)
+			return error;
+	}
+
+	/* Set the correct link count. */
+	if (VFS_I(ip)->i_nlink != rp->parents) {
+		if (!joined) {
+			xfs_trans_ijoin(sc->tp, sc->ip, 0);
+			joined = true;
+		}
+
+		set_nlink(VFS_I(ip), min_t(unsigned long long, rp->parents,
+					   XFS_NLINK_PINNED));
+	}
+
+	/* Log the inode to keep it moving forward if we dirtied anything. */
+	if (joined)
+		xfs_trans_log_inode(sc->tp, ip, XFS_ILOG_CORE);
+	return 0;
+}
+
+/* Set up the filesystem scan so we can look for parents. */
+STATIC int
+xrep_parent_setup_scan(
+	struct xrep_parent	*rp)
+{
+	struct xfs_scrub	*sc = rp->sc;
+	char			*descr;
+	struct xfs_da_geometry	*geo = sc->mp->m_attr_geo;
+	int			max_len;
+	int			error;
+
+	if (!xfs_has_parent(sc->mp))
+		return xrep_findparent_scan_start(sc, &rp->pscan);
+
+	/* Buffers for copying non-pptr attrs to the tempfile */
+	rp->xattr_name = kvmalloc(XATTR_NAME_MAX + 1, XCHK_GFP_FLAGS);
+	if (!rp->xattr_name)
+		return -ENOMEM;
+
+	/*
+	 * Allocate enough memory to handle loading local attr values from the
+	 * xfblob data while flushing stashed attrs to the temporary file.
+	 * We only realloc the buffer when salvaging remote attr values, so
+	 * TRY_HARDER means we allocate the maximal attr value size.
+	 */
+	if (sc->flags & XCHK_TRY_HARDER)
+		max_len = XATTR_SIZE_MAX;
+	else
+		max_len = xfs_attr_leaf_entsize_local_max(geo->blksize);
+	error = xrep_parent_alloc_xattr_value(rp, max_len);
+	if (error)
+		goto out_xattr_name;
+
+	/* Set up some staging memory for logging parent pointer updates. */
+	descr = xchk_xfile_ino_descr(sc, "parent pointer entries");
+	error = xfarray_create(descr, 0, sizeof(struct xrep_pptr),
+			&rp->pptr_recs);
+	kfree(descr);
+	if (error)
+		goto out_xattr_value;
+
+	descr = xchk_xfile_ino_descr(sc, "parent pointer names");
+	error = xfblob_create(descr, &rp->pptr_names);
+	kfree(descr);
+	if (error)
+		goto out_recs;
+
+	/* Set up some storage for copying attrs before the mapping exchange */
+	descr = xchk_xfile_ino_descr(sc,
+				"parent pointer retained xattr entries");
+	error = xfarray_create(descr, 0, sizeof(struct xrep_parent_xattr),
+			&rp->xattr_records);
+	kfree(descr);
+	if (error)
+		goto out_names;
+
+	descr = xchk_xfile_ino_descr(sc,
+				"parent pointer retained xattr values");
+	error = xfblob_create(descr, &rp->xattr_blobs);
+	kfree(descr);
+	if (error)
+		goto out_attr_keys;
+
+	error = __xrep_findparent_scan_start(sc, &rp->pscan,
+			xrep_parent_live_update);
+	if (error)
+		goto out_attr_values;
+
+	return 0;
+
+out_attr_values:
+	xfblob_destroy(rp->xattr_blobs);
+	rp->xattr_blobs = NULL;
+out_attr_keys:
+	xfarray_destroy(rp->xattr_records);
+	rp->xattr_records = NULL;
+out_names:
+	xfblob_destroy(rp->pptr_names);
+	rp->pptr_names = NULL;
+out_recs:
+	xfarray_destroy(rp->pptr_recs);
+	rp->pptr_recs = NULL;
+out_xattr_value:
+	kvfree(rp->xattr_value);
+	rp->xattr_value = NULL;
+out_xattr_name:
+	kvfree(rp->xattr_name);
+	rp->xattr_name = NULL;
+	return error;
+}
+
+int
+xrep_parent(
+	struct xfs_scrub	*sc)
+{
+	struct xrep_parent	*rp = sc->buf;
+	int			error;
+
+	/*
+	 * When the parent pointers feature is enabled, repairs are committed
+	 * by atomically committing a new xattr structure and reaping the old
+	 * attr fork.  Reaping requires rmap and exchange-range to be enabled.
+	 */
+	if (xfs_has_parent(sc->mp)) {
+		if (!xfs_has_rmapbt(sc->mp))
+			return -EOPNOTSUPP;
+		if (!xfs_has_exchange_range(sc->mp))
+			return -EOPNOTSUPP;
+	}
+
+	error = xrep_parent_setup_scan(rp);
+	if (error)
+		return error;
+
+	if (xfs_has_parent(sc->mp))
+		error = xrep_parent_scan_dirtree(rp);
+	else
+		error = xrep_parent_find_dotdot(rp);
+	if (error)
+		goto out_teardown;
+
+	/* Last chance to abort before we start committing dotdot fixes. */
+	if (xchk_should_terminate(sc, &error))
+		goto out_teardown;
+
+	error = xrep_parent_rebuild_tree(rp);
+	if (error)
+		goto out_teardown;
+	if (xfs_has_parent(sc->mp) && !S_ISDIR(VFS_I(sc->ip)->i_mode)) {
+		error = xrep_parent_set_nondir_nlink(rp);
+		if (error)
+			goto out_teardown;
+	}
+
+	error = xrep_defer_finish(sc);
+
+out_teardown:
+	xrep_parent_teardown(rp);
+	return error;
+}
diff --git a/fs/xfs/scrub/quota_repair.c b/fs/xfs/scrub/quota_repair.c
index 0bab4c30cb..90cd1512bb 100644
--- a/fs/xfs/scrub/quota_repair.c
+++ b/fs/xfs/scrub/quota_repair.c
@@ -77,8 +77,6 @@ xrep_quota_item_fill_bmap_hole(
 			irec, &nmaps);
 	if (error)
 		return error;
-	if (nmaps != 1)
-		return -ENOSPC;
 
 	dq->q_blkno = XFS_FSB_TO_DADDR(mp, irec->br_startblock);
 
@@ -444,10 +442,6 @@ xrep_quota_data_fork(
 					XFS_BMAPI_CONVERT, 0, &nrec, &nmap);
 			if (error)
 				goto out;
-			if (nmap != 1) {
-				error = -ENOSPC;
-				goto out;
-			}
 			ASSERT(nrec.br_startoff == irec.br_startoff);
 			ASSERT(nrec.br_blockcount == irec.br_blockcount);
 
diff --git a/fs/xfs/scrub/quotacheck.c b/fs/xfs/scrub/quotacheck.c
new file mode 100644
index 0000000000..c77eb2de8d
--- /dev/null
+++ b/fs/xfs/scrub/quotacheck.c
@@ -0,0 +1,867 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2020-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_inode.h"
+#include "xfs_quota.h"
+#include "xfs_qm.h"
+#include "xfs_icache.h"
+#include "xfs_bmap_util.h"
+#include "xfs_ialloc.h"
+#include "xfs_ag.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/repair.h"
+#include "scrub/xfile.h"
+#include "scrub/xfarray.h"
+#include "scrub/iscan.h"
+#include "scrub/quota.h"
+#include "scrub/quotacheck.h"
+#include "scrub/trace.h"
+
+/*
+ * Live Quotacheck
+ * ===============
+ *
+ * Quota counters are "summary" metadata, in the sense that they are computed
+ * as the summation of the block usage counts for every file on the filesystem.
+ * Therefore, we compute the correct icount, bcount, and rtbcount values by
+ * creating a shadow quota counter structure and walking every inode.
+ */
+
+/* Track the quota deltas for a dquot in a transaction. */
+struct xqcheck_dqtrx {
+	xfs_dqtype_t		q_type;
+	xfs_dqid_t		q_id;
+
+	int64_t			icount_delta;
+
+	int64_t			bcount_delta;
+	int64_t			delbcnt_delta;
+
+	int64_t			rtbcount_delta;
+	int64_t			delrtb_delta;
+};
+
+#define XQCHECK_MAX_NR_DQTRXS	(XFS_QM_TRANS_DQTYPES * XFS_QM_TRANS_MAXDQS)
+
+/*
+ * Track the quota deltas for all dquots attached to a transaction if the
+ * quota deltas are being applied to an inode that we already scanned.
+ */
+struct xqcheck_dqacct {
+	struct rhash_head	hash;
+	uintptr_t		tx_id;
+	struct xqcheck_dqtrx	dqtrx[XQCHECK_MAX_NR_DQTRXS];
+	unsigned int		refcount;
+};
+
+/* Free a shadow dquot accounting structure. */
+static void
+xqcheck_dqacct_free(
+	void			*ptr,
+	void			*arg)
+{
+	struct xqcheck_dqacct	*dqa = ptr;
+
+	kfree(dqa);
+}
+
+/* Set us up to scrub quota counters. */
+int
+xchk_setup_quotacheck(
+	struct xfs_scrub	*sc)
+{
+	if (!XFS_IS_QUOTA_ON(sc->mp))
+		return -ENOENT;
+
+	xchk_fsgates_enable(sc, XCHK_FSGATES_QUOTA);
+
+	sc->buf = kzalloc(sizeof(struct xqcheck), XCHK_GFP_FLAGS);
+	if (!sc->buf)
+		return -ENOMEM;
+
+	return xchk_setup_fs(sc);
+}
+
+/*
+ * Part 1: Collecting dquot resource usage counts.  For each xfs_dquot attached
+ * to each inode, we create a shadow dquot, and compute the inode count and add
+ * the data/rt block usage from what we see.
+ *
+ * To avoid false corruption reports in part 2, any failure in this part must
+ * set the INCOMPLETE flag even when a negative errno is returned.  This care
+ * must be taken with certain errno values (i.e. EFSBADCRC, EFSCORRUPTED,
+ * ECANCELED) that are absorbed into a scrub state flag update by
+ * xchk_*_process_error.  Scrub and repair share the same incore data
+ * structures, so the INCOMPLETE flag is critical to prevent a repair based on
+ * insufficient information.
+ *
+ * Because we are scanning a live filesystem, it's possible that another thread
+ * will try to update the quota counters for an inode that we've already
+ * scanned.  This will cause our counts to be incorrect.  Therefore, we hook
+ * the live transaction code in two places: (1) when the callers update the
+ * per-transaction dqtrx structure to log quota counter updates; and (2) when
+ * transaction commit actually logs those updates to the incore dquot.  By
+ * shadowing transaction updates in this manner, live quotacheck can ensure
+ * by locking the dquot and the shadow structure that its own copies are not
+ * out of date.  Because the hook code runs in a different process context from
+ * the scrub code and the scrub state flags are not accessed atomically,
+ * failures in the hook code must abort the iscan and the scrubber must notice
+ * the aborted scan and set the incomplete flag.
+ *
+ * Note that we use srcu notifier hooks to minimize the overhead when live
+ * quotacheck is /not/ running.
+ */
+
+/* Update an incore dquot counter information from a live update. */
+static int
+xqcheck_update_incore_counts(
+	struct xqcheck		*xqc,
+	struct xfarray		*counts,
+	xfs_dqid_t		id,
+	int64_t			inodes,
+	int64_t			nblks,
+	int64_t			rtblks)
+{
+	struct xqcheck_dquot	xcdq;
+	int			error;
+
+	error = xfarray_load_sparse(counts, id, &xcdq);
+	if (error)
+		return error;
+
+	xcdq.flags |= XQCHECK_DQUOT_WRITTEN;
+	xcdq.icount += inodes;
+	xcdq.bcount += nblks;
+	xcdq.rtbcount += rtblks;
+
+	error = xfarray_store(counts, id, &xcdq);
+	if (error == -EFBIG) {
+		/*
+		 * EFBIG means we tried to store data at too high a byte offset
+		 * in the sparse array.  IOWs, we cannot complete the check and
+		 * must notify userspace that the check was incomplete.
+		 */
+		error = -ECANCELED;
+	}
+	return error;
+}
+
+/* Decide if this is the shadow dquot accounting structure for a transaction. */
+static int
+xqcheck_dqacct_obj_cmpfn(
+	struct rhashtable_compare_arg	*arg,
+	const void			*obj)
+{
+	const uintptr_t			*tx_idp = arg->key;
+	const struct xqcheck_dqacct	*dqa = obj;
+
+	if (dqa->tx_id != *tx_idp)
+		return 1;
+	return 0;
+}
+
+static const struct rhashtable_params xqcheck_dqacct_hash_params = {
+	.min_size		= 32,
+	.key_len		= sizeof(uintptr_t),
+	.key_offset		= offsetof(struct xqcheck_dqacct, tx_id),
+	.head_offset		= offsetof(struct xqcheck_dqacct, hash),
+	.automatic_shrinking	= true,
+	.obj_cmpfn		= xqcheck_dqacct_obj_cmpfn,
+};
+
+/* Find a shadow dqtrx slot for the given dquot. */
+STATIC struct xqcheck_dqtrx *
+xqcheck_get_dqtrx(
+	struct xqcheck_dqacct	*dqa,
+	xfs_dqtype_t		q_type,
+	xfs_dqid_t		q_id)
+{
+	int			i;
+
+	for (i = 0; i < XQCHECK_MAX_NR_DQTRXS; i++) {
+		if (dqa->dqtrx[i].q_type == 0 ||
+		    (dqa->dqtrx[i].q_type == q_type &&
+		     dqa->dqtrx[i].q_id == q_id))
+			return &dqa->dqtrx[i];
+	}
+
+	return NULL;
+}
+
+/*
+ * Create and fill out a quota delta tracking structure to shadow the updates
+ * going on in the regular quota code.
+ */
+static int
+xqcheck_mod_live_ino_dqtrx(
+	struct notifier_block		*nb,
+	unsigned long			action,
+	void				*data)
+{
+	struct xfs_mod_ino_dqtrx_params *p = data;
+	struct xqcheck			*xqc;
+	struct xqcheck_dqacct		*dqa;
+	struct xqcheck_dqtrx		*dqtrx;
+	int				error;
+
+	xqc = container_of(nb, struct xqcheck, qhook.mod_hook.nb);
+
+	/* Skip quota reservation fields. */
+	switch (action) {
+	case XFS_TRANS_DQ_BCOUNT:
+	case XFS_TRANS_DQ_DELBCOUNT:
+	case XFS_TRANS_DQ_ICOUNT:
+	case XFS_TRANS_DQ_RTBCOUNT:
+	case XFS_TRANS_DQ_DELRTBCOUNT:
+		break;
+	default:
+		return NOTIFY_DONE;
+	}
+
+	/* Ignore dqtrx updates for quota types we don't care about. */
+	switch (p->q_type) {
+	case XFS_DQTYPE_USER:
+		if (!xqc->ucounts)
+			return NOTIFY_DONE;
+		break;
+	case XFS_DQTYPE_GROUP:
+		if (!xqc->gcounts)
+			return NOTIFY_DONE;
+		break;
+	case XFS_DQTYPE_PROJ:
+		if (!xqc->pcounts)
+			return NOTIFY_DONE;
+		break;
+	default:
+		return NOTIFY_DONE;
+	}
+
+	/* Skip inodes that haven't been scanned yet. */
+	if (!xchk_iscan_want_live_update(&xqc->iscan, p->ino))
+		return NOTIFY_DONE;
+
+	/* Make a shadow quota accounting tracker for this transaction. */
+	mutex_lock(&xqc->lock);
+	dqa = rhashtable_lookup_fast(&xqc->shadow_dquot_acct, &p->tx_id,
+			xqcheck_dqacct_hash_params);
+	if (!dqa) {
+		dqa = kzalloc(sizeof(struct xqcheck_dqacct), XCHK_GFP_FLAGS);
+		if (!dqa)
+			goto out_abort;
+
+		dqa->tx_id = p->tx_id;
+		error = rhashtable_insert_fast(&xqc->shadow_dquot_acct,
+				&dqa->hash, xqcheck_dqacct_hash_params);
+		if (error)
+			goto out_abort;
+	}
+
+	/* Find the shadow dqtrx (or an empty slot) here. */
+	dqtrx = xqcheck_get_dqtrx(dqa, p->q_type, p->q_id);
+	if (!dqtrx)
+		goto out_abort;
+	if (dqtrx->q_type == 0) {
+		dqtrx->q_type = p->q_type;
+		dqtrx->q_id = p->q_id;
+		dqa->refcount++;
+	}
+
+	/* Update counter */
+	switch (action) {
+	case XFS_TRANS_DQ_BCOUNT:
+		dqtrx->bcount_delta += p->delta;
+		break;
+	case XFS_TRANS_DQ_DELBCOUNT:
+		dqtrx->delbcnt_delta += p->delta;
+		break;
+	case XFS_TRANS_DQ_ICOUNT:
+		dqtrx->icount_delta += p->delta;
+		break;
+	case XFS_TRANS_DQ_RTBCOUNT:
+		dqtrx->rtbcount_delta += p->delta;
+		break;
+	case XFS_TRANS_DQ_DELRTBCOUNT:
+		dqtrx->delrtb_delta += p->delta;
+		break;
+	}
+
+	mutex_unlock(&xqc->lock);
+	return NOTIFY_DONE;
+
+out_abort:
+	xchk_iscan_abort(&xqc->iscan);
+	mutex_unlock(&xqc->lock);
+	return NOTIFY_DONE;
+}
+
+/*
+ * Apply the transaction quota deltas to our shadow quota accounting info when
+ * the regular quota code are doing the same.
+ */
+static int
+xqcheck_apply_live_dqtrx(
+	struct notifier_block		*nb,
+	unsigned long			action,
+	void				*data)
+{
+	struct xfs_apply_dqtrx_params	*p = data;
+	struct xqcheck			*xqc;
+	struct xqcheck_dqacct		*dqa;
+	struct xqcheck_dqtrx		*dqtrx;
+	struct xfarray			*counts;
+	int				error;
+
+	xqc = container_of(nb, struct xqcheck, qhook.apply_hook.nb);
+
+	/* Map the dquot type to an incore counter object. */
+	switch (p->q_type) {
+	case XFS_DQTYPE_USER:
+		counts = xqc->ucounts;
+		break;
+	case XFS_DQTYPE_GROUP:
+		counts = xqc->gcounts;
+		break;
+	case XFS_DQTYPE_PROJ:
+		counts = xqc->pcounts;
+		break;
+	default:
+		return NOTIFY_DONE;
+	}
+
+	if (xchk_iscan_aborted(&xqc->iscan) || counts == NULL)
+		return NOTIFY_DONE;
+
+	/*
+	 * Find the shadow dqtrx for this transaction and dquot, if any deltas
+	 * need to be applied here.  If not, we're finished early.
+	 */
+	mutex_lock(&xqc->lock);
+	dqa = rhashtable_lookup_fast(&xqc->shadow_dquot_acct, &p->tx_id,
+			xqcheck_dqacct_hash_params);
+	if (!dqa)
+		goto out_unlock;
+	dqtrx = xqcheck_get_dqtrx(dqa, p->q_type, p->q_id);
+	if (!dqtrx || dqtrx->q_type == 0)
+		goto out_unlock;
+
+	/* Update our shadow dquot if we're committing. */
+	if (action == XFS_APPLY_DQTRX_COMMIT) {
+		error = xqcheck_update_incore_counts(xqc, counts, p->q_id,
+				dqtrx->icount_delta,
+				dqtrx->bcount_delta + dqtrx->delbcnt_delta,
+				dqtrx->rtbcount_delta + dqtrx->delrtb_delta);
+		if (error)
+			goto out_abort;
+	}
+
+	/* Free the shadow accounting structure if that was the last user. */
+	dqa->refcount--;
+	if (dqa->refcount == 0) {
+		error = rhashtable_remove_fast(&xqc->shadow_dquot_acct,
+				&dqa->hash, xqcheck_dqacct_hash_params);
+		if (error)
+			goto out_abort;
+		xqcheck_dqacct_free(dqa, NULL);
+	}
+
+	mutex_unlock(&xqc->lock);
+	return NOTIFY_DONE;
+
+out_abort:
+	xchk_iscan_abort(&xqc->iscan);
+out_unlock:
+	mutex_unlock(&xqc->lock);
+	return NOTIFY_DONE;
+}
+
+/* Record this inode's quota usage in our shadow quota counter data. */
+STATIC int
+xqcheck_collect_inode(
+	struct xqcheck		*xqc,
+	struct xfs_inode	*ip)
+{
+	struct xfs_trans	*tp = xqc->sc->tp;
+	xfs_filblks_t		nblks, rtblks;
+	uint			ilock_flags = 0;
+	xfs_dqid_t		id;
+	bool			isreg = S_ISREG(VFS_I(ip)->i_mode);
+	int			error = 0;
+
+	if (xfs_is_quota_inode(&tp->t_mountp->m_sb, ip->i_ino)) {
+		/*
+		 * Quota files are never counted towards quota, so we do not
+		 * need to take the lock.
+		 */
+		xchk_iscan_mark_visited(&xqc->iscan, ip);
+		return 0;
+	}
+
+	/* Figure out the data / rt device block counts. */
+	xfs_ilock(ip, XFS_IOLOCK_SHARED);
+	if (isreg)
+		xfs_ilock(ip, XFS_MMAPLOCK_SHARED);
+	if (XFS_IS_REALTIME_INODE(ip)) {
+		/*
+		 * Read in the data fork for rt files so that _count_blocks
+		 * can count the number of blocks allocated from the rt volume.
+		 * Inodes do not track that separately.
+		 */
+		ilock_flags = xfs_ilock_data_map_shared(ip);
+		error = xfs_iread_extents(tp, ip, XFS_DATA_FORK);
+		if (error)
+			goto out_abort;
+	} else {
+		ilock_flags = XFS_ILOCK_SHARED;
+		xfs_ilock(ip, XFS_ILOCK_SHARED);
+	}
+	xfs_inode_count_blocks(tp, ip, &nblks, &rtblks);
+
+	if (xchk_iscan_aborted(&xqc->iscan)) {
+		error = -ECANCELED;
+		goto out_incomplete;
+	}
+
+	/* Update the shadow dquot counters. */
+	mutex_lock(&xqc->lock);
+	if (xqc->ucounts) {
+		id = xfs_qm_id_for_quotatype(ip, XFS_DQTYPE_USER);
+		error = xqcheck_update_incore_counts(xqc, xqc->ucounts, id, 1,
+				nblks, rtblks);
+		if (error)
+			goto out_mutex;
+	}
+
+	if (xqc->gcounts) {
+		id = xfs_qm_id_for_quotatype(ip, XFS_DQTYPE_GROUP);
+		error = xqcheck_update_incore_counts(xqc, xqc->gcounts, id, 1,
+				nblks, rtblks);
+		if (error)
+			goto out_mutex;
+	}
+
+	if (xqc->pcounts) {
+		id = xfs_qm_id_for_quotatype(ip, XFS_DQTYPE_PROJ);
+		error = xqcheck_update_incore_counts(xqc, xqc->pcounts, id, 1,
+				nblks, rtblks);
+		if (error)
+			goto out_mutex;
+	}
+	mutex_unlock(&xqc->lock);
+
+	xchk_iscan_mark_visited(&xqc->iscan, ip);
+	goto out_ilock;
+
+out_mutex:
+	mutex_unlock(&xqc->lock);
+out_abort:
+	xchk_iscan_abort(&xqc->iscan);
+out_incomplete:
+	xchk_set_incomplete(xqc->sc);
+out_ilock:
+	xfs_iunlock(ip, ilock_flags);
+	if (isreg)
+		xfs_iunlock(ip, XFS_MMAPLOCK_SHARED);
+	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
+	return error;
+}
+
+/* Walk all the allocated inodes and run a quota scan on them. */
+STATIC int
+xqcheck_collect_counts(
+	struct xqcheck		*xqc)
+{
+	struct xfs_scrub	*sc = xqc->sc;
+	struct xfs_inode	*ip;
+	int			error;
+
+	/*
+	 * Set up for a potentially lengthy filesystem scan by reducing our
+	 * transaction resource usage for the duration.  Specifically:
+	 *
+	 * Cancel the transaction to release the log grant space while we scan
+	 * the filesystem.
+	 *
+	 * Create a new empty transaction to eliminate the possibility of the
+	 * inode scan deadlocking on cyclical metadata.
+	 *
+	 * We pass the empty transaction to the file scanning function to avoid
+	 * repeatedly cycling empty transactions.  This can be done without
+	 * risk of deadlock between sb_internal and the IOLOCK (we take the
+	 * IOLOCK to quiesce the file before scanning) because empty
+	 * transactions do not take sb_internal.
+	 */
+	xchk_trans_cancel(sc);
+	error = xchk_trans_alloc_empty(sc);
+	if (error)
+		return error;
+
+	while ((error = xchk_iscan_iter(&xqc->iscan, &ip)) == 1) {
+		error = xqcheck_collect_inode(xqc, ip);
+		xchk_irele(sc, ip);
+		if (error)
+			break;
+
+		if (xchk_should_terminate(sc, &error))
+			break;
+	}
+	xchk_iscan_iter_finish(&xqc->iscan);
+	if (error) {
+		xchk_set_incomplete(sc);
+		/*
+		 * If we couldn't grab an inode that was busy with a state
+		 * change, change the error code so that we exit to userspace
+		 * as quickly as possible.
+		 */
+		if (error == -EBUSY)
+			return -ECANCELED;
+		return error;
+	}
+
+	/*
+	 * Switch out for a real transaction in preparation for building a new
+	 * tree.
+	 */
+	xchk_trans_cancel(sc);
+	return xchk_setup_fs(sc);
+}
+
+/*
+ * Part 2: Comparing dquot resource counters.  Walk each xfs_dquot, comparing
+ * the resource usage counters against our shadow dquots; and then walk each
+ * shadow dquot (that wasn't covered in the first part), comparing it against
+ * the xfs_dquot.
+ */
+
+/*
+ * Check the dquot data against what we observed.  Caller must hold the dquot
+ * lock.
+ */
+STATIC int
+xqcheck_compare_dquot(
+	struct xqcheck		*xqc,
+	xfs_dqtype_t		dqtype,
+	struct xfs_dquot	*dq)
+{
+	struct xqcheck_dquot	xcdq;
+	struct xfarray		*counts = xqcheck_counters_for(xqc, dqtype);
+	int			error;
+
+	if (xchk_iscan_aborted(&xqc->iscan)) {
+		xchk_set_incomplete(xqc->sc);
+		return -ECANCELED;
+	}
+
+	mutex_lock(&xqc->lock);
+	error = xfarray_load_sparse(counts, dq->q_id, &xcdq);
+	if (error)
+		goto out_unlock;
+
+	if (xcdq.icount != dq->q_ino.count)
+		xchk_qcheck_set_corrupt(xqc->sc, dqtype, dq->q_id);
+
+	if (xcdq.bcount != dq->q_blk.count)
+		xchk_qcheck_set_corrupt(xqc->sc, dqtype, dq->q_id);
+
+	if (xcdq.rtbcount != dq->q_rtb.count)
+		xchk_qcheck_set_corrupt(xqc->sc, dqtype, dq->q_id);
+
+	xcdq.flags |= (XQCHECK_DQUOT_COMPARE_SCANNED | XQCHECK_DQUOT_WRITTEN);
+	error = xfarray_store(counts, dq->q_id, &xcdq);
+	if (error == -EFBIG) {
+		/*
+		 * EFBIG means we tried to store data at too high a byte offset
+		 * in the sparse array.  IOWs, we cannot complete the check and
+		 * must notify userspace that the check was incomplete.  This
+		 * should never happen outside of the collection phase.
+		 */
+		xchk_set_incomplete(xqc->sc);
+		error = -ECANCELED;
+	}
+	mutex_unlock(&xqc->lock);
+	if (error)
+		return error;
+
+	if (xqc->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return -ECANCELED;
+
+	return 0;
+
+out_unlock:
+	mutex_unlock(&xqc->lock);
+	return error;
+}
+
+/*
+ * Walk all the observed dquots, and make sure there's a matching incore
+ * dquot and that its counts match ours.
+ */
+STATIC int
+xqcheck_walk_observations(
+	struct xqcheck		*xqc,
+	xfs_dqtype_t		dqtype)
+{
+	struct xqcheck_dquot	xcdq;
+	struct xfs_dquot	*dq;
+	struct xfarray		*counts = xqcheck_counters_for(xqc, dqtype);
+	xfarray_idx_t		cur = XFARRAY_CURSOR_INIT;
+	int			error;
+
+	mutex_lock(&xqc->lock);
+	while ((error = xfarray_iter(counts, &cur, &xcdq)) == 1) {
+		xfs_dqid_t	id = cur - 1;
+
+		if (xcdq.flags & XQCHECK_DQUOT_COMPARE_SCANNED)
+			continue;
+
+		mutex_unlock(&xqc->lock);
+
+		error = xfs_qm_dqget(xqc->sc->mp, id, dqtype, false, &dq);
+		if (error == -ENOENT) {
+			xchk_qcheck_set_corrupt(xqc->sc, dqtype, id);
+			return 0;
+		}
+		if (error)
+			return error;
+
+		error = xqcheck_compare_dquot(xqc, dqtype, dq);
+		xfs_qm_dqput(dq);
+		if (error)
+			return error;
+
+		if (xchk_should_terminate(xqc->sc, &error))
+			return error;
+
+		mutex_lock(&xqc->lock);
+	}
+	mutex_unlock(&xqc->lock);
+
+	return error;
+}
+
+/* Compare the quota counters we observed against the live dquots. */
+STATIC int
+xqcheck_compare_dqtype(
+	struct xqcheck		*xqc,
+	xfs_dqtype_t		dqtype)
+{
+	struct xchk_dqiter	cursor = { };
+	struct xfs_scrub	*sc = xqc->sc;
+	struct xfs_dquot	*dq;
+	int			error;
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return 0;
+
+	/* If the quota CHKD flag is cleared, we need to repair this quota. */
+	if (!(xfs_quota_chkd_flag(dqtype) & sc->mp->m_qflags)) {
+		xchk_qcheck_set_corrupt(xqc->sc, dqtype, 0);
+		return 0;
+	}
+
+	/* Compare what we observed against the actual dquots. */
+	xchk_dqiter_init(&cursor, sc, dqtype);
+	while ((error = xchk_dquot_iter(&cursor, &dq)) == 1) {
+		error = xqcheck_compare_dquot(xqc, dqtype, dq);
+		xfs_qm_dqput(dq);
+		if (error)
+			break;
+	}
+	if (error)
+		return error;
+
+	/* Walk all the observed dquots and compare to the incore ones. */
+	return xqcheck_walk_observations(xqc, dqtype);
+}
+
+/* Tear down everything associated with a quotacheck. */
+static void
+xqcheck_teardown_scan(
+	void			*priv)
+{
+	struct xqcheck		*xqc = priv;
+	struct xfs_quotainfo	*qi = xqc->sc->mp->m_quotainfo;
+
+	/* Discourage any hook functions that might be running. */
+	xchk_iscan_abort(&xqc->iscan);
+
+	/*
+	 * As noted above, the apply hook is responsible for cleaning up the
+	 * shadow dquot accounting data when a transaction completes.  The mod
+	 * hook must be removed before the apply hook so that we don't
+	 * mistakenly leave an active shadow account for the mod hook to get
+	 * its hands on.  No hooks should be running after these functions
+	 * return.
+	 */
+	xfs_dqtrx_hook_del(qi, &xqc->qhook);
+
+	if (xqc->shadow_dquot_acct.key_len) {
+		rhashtable_free_and_destroy(&xqc->shadow_dquot_acct,
+				xqcheck_dqacct_free, NULL);
+		xqc->shadow_dquot_acct.key_len = 0;
+	}
+
+	if (xqc->pcounts) {
+		xfarray_destroy(xqc->pcounts);
+		xqc->pcounts = NULL;
+	}
+
+	if (xqc->gcounts) {
+		xfarray_destroy(xqc->gcounts);
+		xqc->gcounts = NULL;
+	}
+
+	if (xqc->ucounts) {
+		xfarray_destroy(xqc->ucounts);
+		xqc->ucounts = NULL;
+	}
+
+	xchk_iscan_teardown(&xqc->iscan);
+	mutex_destroy(&xqc->lock);
+	xqc->sc = NULL;
+}
+
+/*
+ * Scan all inodes in the entire filesystem to generate quota counter data.
+ * If the scan is successful, the quota data will be left alive for a repair.
+ * If any error occurs, we'll tear everything down.
+ */
+STATIC int
+xqcheck_setup_scan(
+	struct xfs_scrub	*sc,
+	struct xqcheck		*xqc)
+{
+	char			*descr;
+	struct xfs_quotainfo	*qi = sc->mp->m_quotainfo;
+	unsigned long long	max_dquots = XFS_DQ_ID_MAX + 1ULL;
+	int			error;
+
+	ASSERT(xqc->sc == NULL);
+	xqc->sc = sc;
+
+	mutex_init(&xqc->lock);
+
+	/* Retry iget every tenth of a second for up to 30 seconds. */
+	xchk_iscan_start(sc, 30000, 100, &xqc->iscan);
+
+	error = -ENOMEM;
+	if (xfs_this_quota_on(sc->mp, XFS_DQTYPE_USER)) {
+		descr = xchk_xfile_descr(sc, "user dquot records");
+		error = xfarray_create(descr, max_dquots,
+				sizeof(struct xqcheck_dquot), &xqc->ucounts);
+		kfree(descr);
+		if (error)
+			goto out_teardown;
+	}
+
+	if (xfs_this_quota_on(sc->mp, XFS_DQTYPE_GROUP)) {
+		descr = xchk_xfile_descr(sc, "group dquot records");
+		error = xfarray_create(descr, max_dquots,
+				sizeof(struct xqcheck_dquot), &xqc->gcounts);
+		kfree(descr);
+		if (error)
+			goto out_teardown;
+	}
+
+	if (xfs_this_quota_on(sc->mp, XFS_DQTYPE_PROJ)) {
+		descr = xchk_xfile_descr(sc, "project dquot records");
+		error = xfarray_create(descr, max_dquots,
+				sizeof(struct xqcheck_dquot), &xqc->pcounts);
+		kfree(descr);
+		if (error)
+			goto out_teardown;
+	}
+
+	/*
+	 * Set up hash table to map transactions to our internal shadow dqtrx
+	 * structures.
+	 */
+	error = rhashtable_init(&xqc->shadow_dquot_acct,
+			&xqcheck_dqacct_hash_params);
+	if (error)
+		goto out_teardown;
+
+	/*
+	 * Hook into the quota code.  The hook only triggers for inodes that
+	 * were already scanned, and the scanner thread takes each inode's
+	 * ILOCK, which means that any in-progress inode updates will finish
+	 * before we can scan the inode.
+	 *
+	 * The apply hook (which removes the shadow dquot accounting struct)
+	 * must be installed before the mod hook so that we never fail to catch
+	 * the end of a quota update sequence and leave stale shadow data.
+	 */
+	ASSERT(sc->flags & XCHK_FSGATES_QUOTA);
+	xfs_dqtrx_hook_setup(&xqc->qhook, xqcheck_mod_live_ino_dqtrx,
+			xqcheck_apply_live_dqtrx);
+
+	error = xfs_dqtrx_hook_add(qi, &xqc->qhook);
+	if (error)
+		goto out_teardown;
+
+	/* Use deferred cleanup to pass the quota count data to repair. */
+	sc->buf_cleanup = xqcheck_teardown_scan;
+	return 0;
+
+out_teardown:
+	xqcheck_teardown_scan(xqc);
+	return error;
+}
+
+/* Scrub all counters for a given quota type. */
+int
+xchk_quotacheck(
+	struct xfs_scrub	*sc)
+{
+	struct xqcheck		*xqc = sc->buf;
+	int			error = 0;
+
+	/* Check quota counters on the live filesystem. */
+	error = xqcheck_setup_scan(sc, xqc);
+	if (error)
+		return error;
+
+	/* Walk all inodes, picking up quota information. */
+	error = xqcheck_collect_counts(xqc);
+	if (!xchk_xref_process_error(sc, 0, 0, &error))
+		return error;
+
+	/* Fail fast if we're not playing with a full dataset. */
+	if (xchk_iscan_aborted(&xqc->iscan))
+		xchk_set_incomplete(sc);
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_INCOMPLETE)
+		return 0;
+
+	/* Compare quota counters. */
+	if (xqc->ucounts) {
+		error = xqcheck_compare_dqtype(xqc, XFS_DQTYPE_USER);
+		if (!xchk_xref_process_error(sc, 0, 0, &error))
+			return error;
+	}
+	if (xqc->gcounts) {
+		error = xqcheck_compare_dqtype(xqc, XFS_DQTYPE_GROUP);
+		if (!xchk_xref_process_error(sc, 0, 0, &error))
+			return error;
+	}
+	if (xqc->pcounts) {
+		error = xqcheck_compare_dqtype(xqc, XFS_DQTYPE_PROJ);
+		if (!xchk_xref_process_error(sc, 0, 0, &error))
+			return error;
+	}
+
+	/* Check one last time for an incomplete dataset. */
+	if (xchk_iscan_aborted(&xqc->iscan))
+		xchk_set_incomplete(sc);
+
+	return 0;
+}
diff --git a/fs/xfs/scrub/quotacheck.h b/fs/xfs/scrub/quotacheck.h
new file mode 100644
index 0000000000..4ea5f249c9
--- /dev/null
+++ b/fs/xfs/scrub/quotacheck.h
@@ -0,0 +1,76 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (c) 2020-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_QUOTACHECK_H__
+#define __XFS_SCRUB_QUOTACHECK_H__
+
+/* Quota counters for live quotacheck. */
+struct xqcheck_dquot {
+	/* block usage count */
+	int64_t			bcount;
+
+	/* inode usage count */
+	int64_t			icount;
+
+	/* realtime block usage count */
+	int64_t			rtbcount;
+
+	/* Record state */
+	unsigned int		flags;
+};
+
+/*
+ * This incore dquot record has been written at least once.  We never want to
+ * store an xqcheck_dquot that looks uninitialized.
+ */
+#define XQCHECK_DQUOT_WRITTEN		(1U << 0)
+
+/* Already checked this dquot. */
+#define XQCHECK_DQUOT_COMPARE_SCANNED	(1U << 1)
+
+/* Already repaired this dquot. */
+#define XQCHECK_DQUOT_REPAIR_SCANNED	(1U << 2)
+
+/* Live quotacheck control structure. */
+struct xqcheck {
+	struct xfs_scrub	*sc;
+
+	/* Shadow dquot counter data. */
+	struct xfarray		*ucounts;
+	struct xfarray		*gcounts;
+	struct xfarray		*pcounts;
+
+	/* Lock protecting quotacheck count observations */
+	struct mutex		lock;
+
+	struct xchk_iscan	iscan;
+
+	/* Hooks into the quota code. */
+	struct xfs_dqtrx_hook	qhook;
+
+	/* Shadow quota delta tracking structure. */
+	struct rhashtable	shadow_dquot_acct;
+};
+
+/* Return the incore counter array for a given quota type. */
+static inline struct xfarray *
+xqcheck_counters_for(
+	struct xqcheck		*xqc,
+	xfs_dqtype_t		dqtype)
+{
+	switch (dqtype) {
+	case XFS_DQTYPE_USER:
+		return xqc->ucounts;
+	case XFS_DQTYPE_GROUP:
+		return xqc->gcounts;
+	case XFS_DQTYPE_PROJ:
+		return xqc->pcounts;
+	}
+
+	ASSERT(0);
+	return NULL;
+}
+
+#endif /* __XFS_SCRUB_QUOTACHECK_H__ */
diff --git a/fs/xfs/scrub/quotacheck_repair.c b/fs/xfs/scrub/quotacheck_repair.c
new file mode 100644
index 0000000000..dd8554c755
--- /dev/null
+++ b/fs/xfs/scrub/quotacheck_repair.c
@@ -0,0 +1,261 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2020-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_inode.h"
+#include "xfs_quota.h"
+#include "xfs_qm.h"
+#include "xfs_icache.h"
+#include "xfs_bmap_util.h"
+#include "xfs_iwalk.h"
+#include "xfs_ialloc.h"
+#include "xfs_sb.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/repair.h"
+#include "scrub/xfile.h"
+#include "scrub/xfarray.h"
+#include "scrub/iscan.h"
+#include "scrub/quota.h"
+#include "scrub/quotacheck.h"
+#include "scrub/trace.h"
+
+/*
+ * Live Quotacheck Repair
+ * ======================
+ *
+ * Use the live quota counter information that we collected to replace the
+ * counter values in the incore dquots.  A scrub->repair cycle should have left
+ * the live data and hooks active, so this is safe so long as we make sure the
+ * dquot is locked.
+ */
+
+/* Commit new counters to a dquot. */
+static int
+xqcheck_commit_dquot(
+	struct xqcheck		*xqc,
+	xfs_dqtype_t		dqtype,
+	struct xfs_dquot	*dq)
+{
+	struct xqcheck_dquot	xcdq;
+	struct xfarray		*counts = xqcheck_counters_for(xqc, dqtype);
+	int64_t			delta;
+	bool			dirty = false;
+	int			error = 0;
+
+	/* Unlock the dquot just long enough to allocate a transaction. */
+	xfs_dqunlock(dq);
+	error = xchk_trans_alloc(xqc->sc, 0);
+	xfs_dqlock(dq);
+	if (error)
+		return error;
+
+	xfs_trans_dqjoin(xqc->sc->tp, dq);
+
+	if (xchk_iscan_aborted(&xqc->iscan)) {
+		error = -ECANCELED;
+		goto out_cancel;
+	}
+
+	mutex_lock(&xqc->lock);
+	error = xfarray_load_sparse(counts, dq->q_id, &xcdq);
+	if (error)
+		goto out_unlock;
+
+	/* Adjust counters as needed. */
+	delta = (int64_t)xcdq.icount - dq->q_ino.count;
+	if (delta) {
+		dq->q_ino.reserved += delta;
+		dq->q_ino.count += delta;
+		dirty = true;
+	}
+
+	delta = (int64_t)xcdq.bcount - dq->q_blk.count;
+	if (delta) {
+		dq->q_blk.reserved += delta;
+		dq->q_blk.count += delta;
+		dirty = true;
+	}
+
+	delta = (int64_t)xcdq.rtbcount - dq->q_rtb.count;
+	if (delta) {
+		dq->q_rtb.reserved += delta;
+		dq->q_rtb.count += delta;
+		dirty = true;
+	}
+
+	xcdq.flags |= (XQCHECK_DQUOT_REPAIR_SCANNED | XQCHECK_DQUOT_WRITTEN);
+	error = xfarray_store(counts, dq->q_id, &xcdq);
+	if (error == -EFBIG) {
+		/*
+		 * EFBIG means we tried to store data at too high a byte offset
+		 * in the sparse array.  IOWs, we cannot complete the repair
+		 * and must cancel the whole operation.  This should never
+		 * happen, but we need to catch it anyway.
+		 */
+		error = -ECANCELED;
+	}
+	mutex_unlock(&xqc->lock);
+	if (error || !dirty)
+		goto out_cancel;
+
+	trace_xrep_quotacheck_dquot(xqc->sc->mp, dq->q_type, dq->q_id);
+
+	/* Commit the dirty dquot to disk. */
+	dq->q_flags |= XFS_DQFLAG_DIRTY;
+	if (dq->q_id)
+		xfs_qm_adjust_dqtimers(dq);
+	xfs_trans_log_dquot(xqc->sc->tp, dq);
+
+	/*
+	 * Transaction commit unlocks the dquot, so we must re-lock it so that
+	 * the caller can put the reference (which apparently requires a locked
+	 * dquot).
+	 */
+	error = xrep_trans_commit(xqc->sc);
+	xfs_dqlock(dq);
+	return error;
+
+out_unlock:
+	mutex_unlock(&xqc->lock);
+out_cancel:
+	xchk_trans_cancel(xqc->sc);
+
+	/* Re-lock the dquot so the caller can put the reference. */
+	xfs_dqlock(dq);
+	return error;
+}
+
+/* Commit new quota counters for a particular quota type. */
+STATIC int
+xqcheck_commit_dqtype(
+	struct xqcheck		*xqc,
+	unsigned int		dqtype)
+{
+	struct xchk_dqiter	cursor = { };
+	struct xqcheck_dquot	xcdq;
+	struct xfs_scrub	*sc = xqc->sc;
+	struct xfs_mount	*mp = sc->mp;
+	struct xfarray		*counts = xqcheck_counters_for(xqc, dqtype);
+	struct xfs_dquot	*dq;
+	xfarray_idx_t		cur = XFARRAY_CURSOR_INIT;
+	int			error;
+
+	/*
+	 * Update the counters of every dquot that the quota file knows about.
+	 */
+	xchk_dqiter_init(&cursor, sc, dqtype);
+	while ((error = xchk_dquot_iter(&cursor, &dq)) == 1) {
+		error = xqcheck_commit_dquot(xqc, dqtype, dq);
+		xfs_qm_dqput(dq);
+		if (error)
+			break;
+	}
+	if (error)
+		return error;
+
+	/*
+	 * Make a second pass to deal with the dquots that we know about but
+	 * the quota file previously did not know about.
+	 */
+	mutex_lock(&xqc->lock);
+	while ((error = xfarray_iter(counts, &cur, &xcdq)) == 1) {
+		xfs_dqid_t	id = cur - 1;
+
+		if (xcdq.flags & XQCHECK_DQUOT_REPAIR_SCANNED)
+			continue;
+
+		mutex_unlock(&xqc->lock);
+
+		/*
+		 * Grab the dquot, allowing for dquot block allocation in a
+		 * separate transaction.  We committed the scrub transaction
+		 * in a previous step, so we will not be creating nested
+		 * transactions here.
+		 */
+		error = xfs_qm_dqget(mp, id, dqtype, true, &dq);
+		if (error)
+			return error;
+
+		error = xqcheck_commit_dquot(xqc, dqtype, dq);
+		xfs_qm_dqput(dq);
+		if (error)
+			return error;
+
+		mutex_lock(&xqc->lock);
+	}
+	mutex_unlock(&xqc->lock);
+
+	return error;
+}
+
+/* Figure out quota CHKD flags for the running quota types. */
+static inline unsigned int
+xqcheck_chkd_flags(
+	struct xfs_mount	*mp)
+{
+	unsigned int		ret = 0;
+
+	if (XFS_IS_UQUOTA_ON(mp))
+		ret |= XFS_UQUOTA_CHKD;
+	if (XFS_IS_GQUOTA_ON(mp))
+		ret |= XFS_GQUOTA_CHKD;
+	if (XFS_IS_PQUOTA_ON(mp))
+		ret |= XFS_PQUOTA_CHKD;
+	return ret;
+}
+
+/* Commit the new dquot counters. */
+int
+xrep_quotacheck(
+	struct xfs_scrub	*sc)
+{
+	struct xqcheck		*xqc = sc->buf;
+	unsigned int		qflags = xqcheck_chkd_flags(sc->mp);
+	int			error;
+
+	/*
+	 * Clear the CHKD flag for the running quota types and commit the scrub
+	 * transaction so that we can allocate new quota block mappings if we
+	 * have to.  If we crash after this point, the sb still has the CHKD
+	 * flags cleared, so mount quotacheck will fix all of this up.
+	 */
+	xrep_update_qflags(sc, qflags, 0);
+	error = xrep_trans_commit(sc);
+	if (error)
+		return error;
+
+	/* Commit the new counters to the dquots. */
+	if (xqc->ucounts) {
+		error = xqcheck_commit_dqtype(xqc, XFS_DQTYPE_USER);
+		if (error)
+			return error;
+	}
+	if (xqc->gcounts) {
+		error = xqcheck_commit_dqtype(xqc, XFS_DQTYPE_GROUP);
+		if (error)
+			return error;
+	}
+	if (xqc->pcounts) {
+		error = xqcheck_commit_dqtype(xqc, XFS_DQTYPE_PROJ);
+		if (error)
+			return error;
+	}
+
+	/* Set the CHKD flags now that we've fixed quota counts. */
+	error = xchk_trans_alloc(sc, 0);
+	if (error)
+		return error;
+
+	xrep_update_qflags(sc, 0, qflags);
+	return xrep_trans_commit(sc);
+}
diff --git a/fs/xfs/scrub/rcbag.c b/fs/xfs/scrub/rcbag.c
new file mode 100644
index 0000000000..e1e52bc207
--- /dev/null
+++ b/fs/xfs/scrub/rcbag.c
@@ -0,0 +1,307 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2022-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_buf_mem.h"
+#include "xfs_btree_mem.h"
+#include "xfs_error.h"
+#include "scrub/scrub.h"
+#include "scrub/rcbag_btree.h"
+#include "scrub/rcbag.h"
+#include "scrub/trace.h"
+
+struct rcbag {
+	struct xfs_mount	*mp;
+	struct xfbtree		xfbtree;
+	uint64_t		nr_items;
+};
+
+int
+rcbag_init(
+	struct xfs_mount	*mp,
+	struct xfs_buftarg	*btp,
+	struct rcbag		**bagp)
+{
+	struct rcbag		*bag;
+	int			error;
+
+	bag = kzalloc(sizeof(struct rcbag), XCHK_GFP_FLAGS);
+	if (!bag)
+		return -ENOMEM;
+
+	bag->nr_items = 0;
+	bag->mp = mp;
+
+	error = rcbagbt_mem_init(mp, &bag->xfbtree, btp);
+	if (error)
+		goto out_bag;
+
+	*bagp = bag;
+	return 0;
+
+out_bag:
+	kfree(bag);
+	return error;
+}
+
+void
+rcbag_free(
+	struct rcbag		**bagp)
+{
+	struct rcbag		*bag = *bagp;
+
+	xfbtree_destroy(&bag->xfbtree);
+	kfree(bag);
+	*bagp = NULL;
+}
+
+/* Track an rmap in the refcount bag. */
+int
+rcbag_add(
+	struct rcbag			*bag,
+	struct xfs_trans		*tp,
+	const struct xfs_rmap_irec	*rmap)
+{
+	struct rcbag_rec		bagrec;
+	struct xfs_mount		*mp = bag->mp;
+	struct xfs_btree_cur		*cur;
+	int				has;
+	int				error;
+
+	cur = rcbagbt_mem_cursor(mp, tp, &bag->xfbtree);
+	error = rcbagbt_lookup_eq(cur, rmap, &has);
+	if (error)
+		goto out_cur;
+
+	if (has) {
+		error = rcbagbt_get_rec(cur, &bagrec, &has);
+		if (error)
+			goto out_cur;
+		if (!has) {
+			error = -EFSCORRUPTED;
+			goto out_cur;
+		}
+
+		bagrec.rbg_refcount++;
+		error = rcbagbt_update(cur, &bagrec);
+		if (error)
+			goto out_cur;
+	} else {
+		bagrec.rbg_startblock = rmap->rm_startblock;
+		bagrec.rbg_blockcount = rmap->rm_blockcount;
+		bagrec.rbg_refcount = 1;
+
+		error = rcbagbt_insert(cur, &bagrec, &has);
+		if (error)
+			goto out_cur;
+		if (!has) {
+			error = -EFSCORRUPTED;
+			goto out_cur;
+		}
+	}
+
+	xfs_btree_del_cursor(cur, 0);
+
+	error = xfbtree_trans_commit(&bag->xfbtree, tp);
+	if (error)
+		return error;
+
+	bag->nr_items++;
+	return 0;
+
+out_cur:
+	xfs_btree_del_cursor(cur, error);
+	xfbtree_trans_cancel(&bag->xfbtree, tp);
+	return error;
+}
+
+/* Return the number of records in the bag. */
+uint64_t
+rcbag_count(
+	const struct rcbag	*rcbag)
+{
+	return rcbag->nr_items;
+}
+
+static inline uint32_t rcbag_rec_next_bno(const struct rcbag_rec *r)
+{
+	return r->rbg_startblock + r->rbg_blockcount;
+}
+
+/*
+ * Find the next block where the refcount changes, given the next rmap we
+ * looked at and the ones we're already tracking.
+ */
+int
+rcbag_next_edge(
+	struct rcbag			*bag,
+	struct xfs_trans		*tp,
+	const struct xfs_rmap_irec	*next_rmap,
+	bool				next_valid,
+	uint32_t			*next_bnop)
+{
+	struct rcbag_rec		bagrec;
+	struct xfs_mount		*mp = bag->mp;
+	struct xfs_btree_cur		*cur;
+	uint32_t			next_bno = NULLAGBLOCK;
+	int				has;
+	int				error;
+
+	if (next_valid)
+		next_bno = next_rmap->rm_startblock;
+
+	cur = rcbagbt_mem_cursor(mp, tp, &bag->xfbtree);
+	error = xfs_btree_goto_left_edge(cur);
+	if (error)
+		goto out_cur;
+
+	while (true) {
+		error = xfs_btree_increment(cur, 0, &has);
+		if (error)
+			goto out_cur;
+		if (!has)
+			break;
+
+		error = rcbagbt_get_rec(cur, &bagrec, &has);
+		if (error)
+			goto out_cur;
+		if (!has) {
+			error = -EFSCORRUPTED;
+			goto out_cur;
+		}
+
+		next_bno = min(next_bno, rcbag_rec_next_bno(&bagrec));
+	}
+
+	/*
+	 * We should have found /something/ because either next_rrm is the next
+	 * interesting rmap to look at after emitting this refcount extent, or
+	 * there are other rmaps in rmap_bag contributing to the current
+	 * sharing count.  But if something is seriously wrong, bail out.
+	 */
+	if (next_bno == NULLAGBLOCK) {
+		error = -EFSCORRUPTED;
+		goto out_cur;
+	}
+
+	xfs_btree_del_cursor(cur, 0);
+
+	*next_bnop = next_bno;
+	return 0;
+
+out_cur:
+	xfs_btree_del_cursor(cur, error);
+	return error;
+}
+
+/* Pop all refcount bag records that end at next_bno */
+int
+rcbag_remove_ending_at(
+	struct rcbag		*bag,
+	struct xfs_trans	*tp,
+	uint32_t		next_bno)
+{
+	struct rcbag_rec	bagrec;
+	struct xfs_mount	*mp = bag->mp;
+	struct xfs_btree_cur	*cur;
+	int			has;
+	int			error;
+
+	/* go to the right edge of the tree */
+	cur = rcbagbt_mem_cursor(mp, tp, &bag->xfbtree);
+	memset(&cur->bc_rec, 0xFF, sizeof(cur->bc_rec));
+	error = xfs_btree_lookup(cur, XFS_LOOKUP_GE, &has);
+	if (error)
+		goto out_cur;
+
+	while (true) {
+		error = xfs_btree_decrement(cur, 0, &has);
+		if (error)
+			goto out_cur;
+		if (!has)
+			break;
+
+		error = rcbagbt_get_rec(cur, &bagrec, &has);
+		if (error)
+			goto out_cur;
+		if (!has) {
+			error = -EFSCORRUPTED;
+			goto out_cur;
+		}
+
+		if (rcbag_rec_next_bno(&bagrec) != next_bno)
+			continue;
+
+		error = xfs_btree_delete(cur, &has);
+		if (error)
+			goto out_cur;
+		if (!has) {
+			error = -EFSCORRUPTED;
+			goto out_cur;
+		}
+
+		bag->nr_items -= bagrec.rbg_refcount;
+	}
+
+	xfs_btree_del_cursor(cur, 0);
+	return xfbtree_trans_commit(&bag->xfbtree, tp);
+out_cur:
+	xfs_btree_del_cursor(cur, error);
+	xfbtree_trans_cancel(&bag->xfbtree, tp);
+	return error;
+}
+
+/* Dump the rcbag. */
+void
+rcbag_dump(
+	struct rcbag			*bag,
+	struct xfs_trans		*tp)
+{
+	struct rcbag_rec		bagrec;
+	struct xfs_mount		*mp = bag->mp;
+	struct xfs_btree_cur		*cur;
+	unsigned long long		nr = 0;
+	int				has;
+	int				error;
+
+	cur = rcbagbt_mem_cursor(mp, tp, &bag->xfbtree);
+	error = xfs_btree_goto_left_edge(cur);
+	if (error)
+		goto out_cur;
+
+	while (true) {
+		error = xfs_btree_increment(cur, 0, &has);
+		if (error)
+			goto out_cur;
+		if (!has)
+			break;
+
+		error = rcbagbt_get_rec(cur, &bagrec, &has);
+		if (error)
+			goto out_cur;
+		if (!has) {
+			error = -EFSCORRUPTED;
+			goto out_cur;
+		}
+
+		xfs_err(bag->mp, "[%llu]: bno 0x%x fsbcount 0x%x refcount 0x%llx\n",
+				nr++,
+				(unsigned int)bagrec.rbg_startblock,
+				(unsigned int)bagrec.rbg_blockcount,
+				(unsigned long long)bagrec.rbg_refcount);
+	}
+
+out_cur:
+	xfs_btree_del_cursor(cur, error);
+}
diff --git a/fs/xfs/scrub/rcbag.h b/fs/xfs/scrub/rcbag.h
new file mode 100644
index 0000000000..e29ef788ba
--- /dev/null
+++ b/fs/xfs/scrub/rcbag.h
@@ -0,0 +1,28 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2022-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_RCBAG_H__
+#define __XFS_SCRUB_RCBAG_H__
+
+struct xfs_mount;
+struct rcbag;
+struct xfs_buftarg;
+
+int rcbag_init(struct xfs_mount *mp, struct xfs_buftarg *btp,
+		struct rcbag **bagp);
+void rcbag_free(struct rcbag **bagp);
+int rcbag_add(struct rcbag *bag, struct xfs_trans *tp,
+		const struct xfs_rmap_irec *rmap);
+uint64_t rcbag_count(const struct rcbag *bag);
+
+int rcbag_next_edge(struct rcbag *bag, struct xfs_trans *tp,
+		const struct xfs_rmap_irec *next_rmap, bool next_valid,
+		uint32_t *next_bnop);
+int rcbag_remove_ending_at(struct rcbag *bag, struct xfs_trans *tp,
+		uint32_t next_bno);
+
+void rcbag_dump(struct rcbag *bag, struct xfs_trans *tp);
+
+#endif /* __XFS_SCRUB_RCBAG_H__ */
diff --git a/fs/xfs/scrub/rcbag_btree.c b/fs/xfs/scrub/rcbag_btree.c
new file mode 100644
index 0000000000..709356dc62
--- /dev/null
+++ b/fs/xfs/scrub/rcbag_btree.c
@@ -0,0 +1,370 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2022-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_buf_mem.h"
+#include "xfs_btree_mem.h"
+#include "xfs_error.h"
+#include "scrub/rcbag_btree.h"
+#include "scrub/trace.h"
+
+static struct kmem_cache	*rcbagbt_cur_cache;
+
+STATIC void
+rcbagbt_init_key_from_rec(
+	union xfs_btree_key		*key,
+	const union xfs_btree_rec	*rec)
+{
+	struct rcbag_key	*bag_key = (struct rcbag_key *)key;
+	const struct rcbag_rec	*bag_rec = (const struct rcbag_rec *)rec;
+
+	BUILD_BUG_ON(sizeof(struct rcbag_key) > sizeof(union xfs_btree_key));
+	BUILD_BUG_ON(sizeof(struct rcbag_rec) > sizeof(union xfs_btree_rec));
+
+	bag_key->rbg_startblock = bag_rec->rbg_startblock;
+	bag_key->rbg_blockcount = bag_rec->rbg_blockcount;
+}
+
+STATIC void
+rcbagbt_init_rec_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*rec)
+{
+	struct rcbag_rec	*bag_rec = (struct rcbag_rec *)rec;
+	struct rcbag_rec	*bag_irec = (struct rcbag_rec *)&cur->bc_rec;
+
+	bag_rec->rbg_startblock = bag_irec->rbg_startblock;
+	bag_rec->rbg_blockcount = bag_irec->rbg_blockcount;
+	bag_rec->rbg_refcount = bag_irec->rbg_refcount;
+}
+
+STATIC int64_t
+rcbagbt_key_diff(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_key	*key)
+{
+	struct rcbag_rec		*rec = (struct rcbag_rec *)&cur->bc_rec;
+	const struct rcbag_key		*kp = (const struct rcbag_key *)key;
+
+	if (kp->rbg_startblock > rec->rbg_startblock)
+		return 1;
+	if (kp->rbg_startblock < rec->rbg_startblock)
+		return -1;
+
+	if (kp->rbg_blockcount > rec->rbg_blockcount)
+		return 1;
+	if (kp->rbg_blockcount < rec->rbg_blockcount)
+		return -1;
+
+	return 0;
+}
+
+STATIC int64_t
+rcbagbt_diff_two_keys(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_key	*k1,
+	const union xfs_btree_key	*k2,
+	const union xfs_btree_key	*mask)
+{
+	const struct rcbag_key		*kp1 = (const struct rcbag_key *)k1;
+	const struct rcbag_key		*kp2 = (const struct rcbag_key *)k2;
+
+	ASSERT(mask == NULL);
+
+	if (kp1->rbg_startblock > kp2->rbg_startblock)
+		return 1;
+	if (kp1->rbg_startblock < kp2->rbg_startblock)
+		return -1;
+
+	if (kp1->rbg_blockcount > kp2->rbg_blockcount)
+		return 1;
+	if (kp1->rbg_blockcount < kp2->rbg_blockcount)
+		return -1;
+
+	return 0;
+}
+
+STATIC int
+rcbagbt_keys_inorder(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_key	*k1,
+	const union xfs_btree_key	*k2)
+{
+	const struct rcbag_key		*kp1 = (const struct rcbag_key *)k1;
+	const struct rcbag_key		*kp2 = (const struct rcbag_key *)k2;
+
+	if (kp1->rbg_startblock > kp2->rbg_startblock)
+		return 0;
+	if (kp1->rbg_startblock < kp2->rbg_startblock)
+		return 1;
+
+	if (kp1->rbg_blockcount > kp2->rbg_blockcount)
+		return 0;
+	if (kp1->rbg_blockcount < kp2->rbg_blockcount)
+		return 1;
+
+	return 0;
+}
+
+STATIC int
+rcbagbt_recs_inorder(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_rec	*r1,
+	const union xfs_btree_rec	*r2)
+{
+	const struct rcbag_rec		*rp1 = (const struct rcbag_rec *)r1;
+	const struct rcbag_rec		*rp2 = (const struct rcbag_rec *)r2;
+
+	if (rp1->rbg_startblock > rp2->rbg_startblock)
+		return 0;
+	if (rp1->rbg_startblock < rp2->rbg_startblock)
+		return 1;
+
+	if (rp1->rbg_blockcount > rp2->rbg_blockcount)
+		return 0;
+	if (rp1->rbg_blockcount < rp2->rbg_blockcount)
+		return 1;
+
+	return 0;
+}
+
+static xfs_failaddr_t
+rcbagbt_verify(
+	struct xfs_buf		*bp)
+{
+	struct xfs_mount	*mp = bp->b_mount;
+	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
+	xfs_failaddr_t		fa;
+	unsigned int		level;
+	unsigned int		maxrecs;
+
+	if (!xfs_verify_magic(bp, block->bb_magic))
+		return __this_address;
+
+	fa = xfs_btree_fsblock_v5hdr_verify(bp, XFS_RMAP_OWN_UNKNOWN);
+	if (fa)
+		return fa;
+
+	level = be16_to_cpu(block->bb_level);
+	if (level >= rcbagbt_maxlevels_possible())
+		return __this_address;
+
+	maxrecs = rcbagbt_maxrecs(mp, XFBNO_BLOCKSIZE, level == 0);
+	return xfs_btree_memblock_verify(bp, maxrecs);
+}
+
+static void
+rcbagbt_rw_verify(
+	struct xfs_buf	*bp)
+{
+	xfs_failaddr_t	fa = rcbagbt_verify(bp);
+
+	if (fa)
+		xfs_verifier_error(bp, -EFSCORRUPTED, fa);
+}
+
+/* skip crc checks on in-memory btrees to save time */
+static const struct xfs_buf_ops rcbagbt_mem_buf_ops = {
+	.name			= "rcbagbt_mem",
+	.magic			= { 0, cpu_to_be32(RCBAG_MAGIC) },
+	.verify_read		= rcbagbt_rw_verify,
+	.verify_write		= rcbagbt_rw_verify,
+	.verify_struct		= rcbagbt_verify,
+};
+
+static const struct xfs_btree_ops rcbagbt_mem_ops = {
+	.name			= "rcbag",
+	.type			= XFS_BTREE_TYPE_MEM,
+
+	.rec_len		= sizeof(struct rcbag_rec),
+	.key_len		= sizeof(struct rcbag_key),
+	.ptr_len		= XFS_BTREE_LONG_PTR_LEN,
+
+	.lru_refs		= 1,
+	.statoff		= XFS_STATS_CALC_INDEX(xs_rcbag_2),
+
+	.dup_cursor		= xfbtree_dup_cursor,
+	.set_root		= xfbtree_set_root,
+	.alloc_block		= xfbtree_alloc_block,
+	.free_block		= xfbtree_free_block,
+	.get_minrecs		= xfbtree_get_minrecs,
+	.get_maxrecs		= xfbtree_get_maxrecs,
+	.init_key_from_rec	= rcbagbt_init_key_from_rec,
+	.init_rec_from_cur	= rcbagbt_init_rec_from_cur,
+	.init_ptr_from_cur	= xfbtree_init_ptr_from_cur,
+	.key_diff		= rcbagbt_key_diff,
+	.buf_ops		= &rcbagbt_mem_buf_ops,
+	.diff_two_keys		= rcbagbt_diff_two_keys,
+	.keys_inorder		= rcbagbt_keys_inorder,
+	.recs_inorder		= rcbagbt_recs_inorder,
+};
+
+/* Create a cursor for an in-memory btree. */
+struct xfs_btree_cur *
+rcbagbt_mem_cursor(
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	struct xfbtree		*xfbtree)
+{
+	struct xfs_btree_cur	*cur;
+
+	cur = xfs_btree_alloc_cursor(mp, tp, &rcbagbt_mem_ops,
+			rcbagbt_maxlevels_possible(), rcbagbt_cur_cache);
+
+	cur->bc_mem.xfbtree = xfbtree;
+	cur->bc_nlevels = xfbtree->nlevels;
+	return cur;
+}
+
+/* Create an in-memory refcount bag btree. */
+int
+rcbagbt_mem_init(
+	struct xfs_mount	*mp,
+	struct xfbtree		*xfbt,
+	struct xfs_buftarg	*btp)
+{
+	xfbt->owner = 0;
+	return xfbtree_init(mp, xfbt, btp, &rcbagbt_mem_ops);
+}
+
+/* Calculate number of records in a refcount bag btree block. */
+static inline unsigned int
+rcbagbt_block_maxrecs(
+	unsigned int		blocklen,
+	bool			leaf)
+{
+	if (leaf)
+		return blocklen / sizeof(struct rcbag_rec);
+	return blocklen /
+		(sizeof(struct rcbag_key) + sizeof(rcbag_ptr_t));
+}
+
+/*
+ * Calculate number of records in an refcount bag btree block.
+ */
+unsigned int
+rcbagbt_maxrecs(
+	struct xfs_mount	*mp,
+	unsigned int		blocklen,
+	bool			leaf)
+{
+	blocklen -= RCBAG_BLOCK_LEN;
+	return rcbagbt_block_maxrecs(blocklen, leaf);
+}
+
+/* Compute the max possible height for refcount bag btrees. */
+unsigned int
+rcbagbt_maxlevels_possible(void)
+{
+	unsigned int		minrecs[2];
+	unsigned int		blocklen;
+
+	blocklen = XFBNO_BLOCKSIZE - XFS_BTREE_LBLOCK_CRC_LEN;
+
+	minrecs[0] = rcbagbt_block_maxrecs(blocklen, true) / 2;
+	minrecs[1] = rcbagbt_block_maxrecs(blocklen, false) / 2;
+
+	return xfs_btree_space_to_height(minrecs, ULLONG_MAX);
+}
+
+/* Calculate the refcount bag btree size for some records. */
+unsigned long long
+rcbagbt_calc_size(
+	unsigned long long	nr_records)
+{
+	unsigned int		minrecs[2];
+	unsigned int		blocklen;
+
+	blocklen = XFBNO_BLOCKSIZE - XFS_BTREE_LBLOCK_CRC_LEN;
+
+	minrecs[0] = rcbagbt_block_maxrecs(blocklen, true) / 2;
+	minrecs[1] = rcbagbt_block_maxrecs(blocklen, false) / 2;
+
+	return xfs_btree_calc_size(minrecs, nr_records);
+}
+
+int __init
+rcbagbt_init_cur_cache(void)
+{
+	rcbagbt_cur_cache = kmem_cache_create("xfs_rcbagbt_cur",
+			xfs_btree_cur_sizeof(rcbagbt_maxlevels_possible()),
+			0, 0, NULL);
+
+	if (!rcbagbt_cur_cache)
+		return -ENOMEM;
+	return 0;
+}
+
+void
+rcbagbt_destroy_cur_cache(void)
+{
+	kmem_cache_destroy(rcbagbt_cur_cache);
+	rcbagbt_cur_cache = NULL;
+}
+
+/* Look up the refcount bag record corresponding to this reverse mapping. */
+int
+rcbagbt_lookup_eq(
+	struct xfs_btree_cur		*cur,
+	const struct xfs_rmap_irec	*rmap,
+	int				*success)
+{
+	struct rcbag_rec		*rec = (struct rcbag_rec *)&cur->bc_rec;
+
+	rec->rbg_startblock = rmap->rm_startblock;
+	rec->rbg_blockcount = rmap->rm_blockcount;
+
+	return xfs_btree_lookup(cur, XFS_LOOKUP_EQ, success);
+}
+
+/* Get the data from the pointed-to record. */
+int
+rcbagbt_get_rec(
+	struct xfs_btree_cur	*cur,
+	struct rcbag_rec	*rec,
+	int			*has)
+{
+	union xfs_btree_rec	*btrec;
+	int			error;
+
+	error = xfs_btree_get_rec(cur, &btrec, has);
+	if (error || !(*has))
+		return error;
+
+	memcpy(rec, btrec, sizeof(struct rcbag_rec));
+	return 0;
+}
+
+/* Update the record referred to by cur to the value given. */
+int
+rcbagbt_update(
+	struct xfs_btree_cur	*cur,
+	const struct rcbag_rec	*rec)
+{
+	union xfs_btree_rec	btrec;
+
+	memcpy(&btrec, rec, sizeof(struct rcbag_rec));
+	return xfs_btree_update(cur, &btrec);
+}
+
+/* Update the record referred to by cur to the value given. */
+int
+rcbagbt_insert(
+	struct xfs_btree_cur	*cur,
+	const struct rcbag_rec	*rec,
+	int			*success)
+{
+	struct rcbag_rec	*btrec = (struct rcbag_rec *)&cur->bc_rec;
+
+	memcpy(btrec, rec, sizeof(struct rcbag_rec));
+	return xfs_btree_insert(cur, success);
+}
diff --git a/fs/xfs/scrub/rcbag_btree.h b/fs/xfs/scrub/rcbag_btree.h
new file mode 100644
index 0000000000..03cadb0325
--- /dev/null
+++ b/fs/xfs/scrub/rcbag_btree.h
@@ -0,0 +1,81 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2022-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_RCBAG_BTREE_H__
+#define __XFS_SCRUB_RCBAG_BTREE_H__
+
+#ifdef CONFIG_XFS_BTREE_IN_MEM
+
+struct xfs_buf;
+struct xfs_btree_cur;
+struct xfs_mount;
+
+#define RCBAG_MAGIC	0x74826671	/* 'JRBG' */
+
+struct rcbag_key {
+	uint32_t	rbg_startblock;
+	uint32_t	rbg_blockcount;
+};
+
+struct rcbag_rec {
+	uint32_t	rbg_startblock;
+	uint32_t	rbg_blockcount;
+	uint64_t	rbg_refcount;
+};
+
+typedef __be64 rcbag_ptr_t;
+
+/* reflinks only exist on crc enabled filesystems */
+#define RCBAG_BLOCK_LEN	XFS_BTREE_LBLOCK_CRC_LEN
+
+/*
+ * Record, key, and pointer address macros for btree blocks.
+ *
+ * (note that some of these may appear unused, but they are used in userspace)
+ */
+#define RCBAG_REC_ADDR(block, index) \
+	((struct rcbag_rec *) \
+		((char *)(block) + RCBAG_BLOCK_LEN + \
+		 (((index) - 1) * sizeof(struct rcbag_rec))))
+
+#define RCBAG_KEY_ADDR(block, index) \
+	((struct rcbag_key *) \
+		((char *)(block) + RCBAG_BLOCK_LEN + \
+		 ((index) - 1) * sizeof(struct rcbag_key)))
+
+#define RCBAG_PTR_ADDR(block, index, maxrecs) \
+	((rcbag_ptr_t *) \
+		((char *)(block) + RCBAG_BLOCK_LEN + \
+		 (maxrecs) * sizeof(struct rcbag_key) + \
+		 ((index) - 1) * sizeof(rcbag_ptr_t)))
+
+unsigned int rcbagbt_maxrecs(struct xfs_mount *mp, unsigned int blocklen,
+		bool leaf);
+
+unsigned long long rcbagbt_calc_size(unsigned long long nr_records);
+
+unsigned int rcbagbt_maxlevels_possible(void);
+
+int __init rcbagbt_init_cur_cache(void);
+void rcbagbt_destroy_cur_cache(void);
+
+struct xfs_btree_cur *rcbagbt_mem_cursor(struct xfs_mount *mp,
+		struct xfs_trans *tp, struct xfbtree *xfbtree);
+int rcbagbt_mem_init(struct xfs_mount *mp, struct xfbtree *xfbtree,
+		struct xfs_buftarg *btp);
+
+int rcbagbt_lookup_eq(struct xfs_btree_cur *cur,
+		const struct xfs_rmap_irec *rmap, int *success);
+int rcbagbt_get_rec(struct xfs_btree_cur *cur, struct rcbag_rec *rec, int *has);
+int rcbagbt_update(struct xfs_btree_cur *cur, const struct rcbag_rec *rec);
+int rcbagbt_insert(struct xfs_btree_cur *cur, const struct rcbag_rec *rec,
+		int *success);
+
+#else
+# define rcbagbt_init_cur_cache()		0
+# define rcbagbt_destroy_cur_cache()		((void)0)
+#endif /* CONFIG_XFS_BTREE_IN_MEM */
+
+#endif /* __XFS_SCRUB_RCBAG_BTREE_H__ */
diff --git a/fs/xfs/scrub/readdir.c b/fs/xfs/scrub/readdir.c
index 16462332c8..01c9a2dc0f 100644
--- a/fs/xfs/scrub/readdir.c
+++ b/fs/xfs/scrub/readdir.c
@@ -18,6 +18,7 @@
 #include "xfs_trans.h"
 #include "xfs_error.h"
 #include "scrub/scrub.h"
+#include "scrub/common.h"
 #include "scrub/readdir.h"
 
 /* Call a function for every entry in a shortform directory. */
@@ -99,7 +100,7 @@ xchk_dir_walk_block(
 	unsigned int		off, next_off, end;
 	int			error;
 
-	error = xfs_dir3_block_read(sc->tp, dp, &bp);
+	error = xfs_dir3_block_read(sc->tp, dp, dp->i_ino, &bp);
 	if (error)
 		return error;
 
@@ -175,7 +176,7 @@ xchk_read_leaf_dir_buf(
 	if (new_off > *curoff)
 		*curoff = new_off;
 
-	return xfs_dir3_data_read(tp, dp, map.br_startoff, 0, bpp);
+	return xfs_dir3_data_read(tp, dp, dp->i_ino, map.br_startoff, 0, bpp);
 }
 
 /* Call a function for every entry in a leaf directory. */
@@ -273,32 +274,27 @@ xchk_dir_walk(
 		.dp		= dp,
 		.geo		= dp->i_mount->m_dir_geo,
 		.trans		= sc->tp,
+		.owner		= dp->i_ino,
 	};
-	bool			isblock;
 	int			error;
 
 	if (xfs_is_shutdown(dp->i_mount))
 		return -EIO;
 
 	ASSERT(S_ISDIR(VFS_I(dp)->i_mode));
-	ASSERT(xfs_isilocked(dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL);
 
-	if (dp->i_df.if_format == XFS_DINODE_FMT_LOCAL)
+	switch (xfs_dir2_format(&args, &error)) {
+	case XFS_DIR2_FMT_SF:
 		return xchk_dir_walk_sf(sc, dp, dirent_fn, priv);
-
-	/* dir2 functions require that the data fork is loaded */
-	error = xfs_iread_extents(sc->tp, dp, XFS_DATA_FORK);
-	if (error)
-		return error;
-
-	error = xfs_dir2_isblock(&args, &isblock);
-	if (error)
-		return error;
-
-	if (isblock)
+	case XFS_DIR2_FMT_BLOCK:
 		return xchk_dir_walk_block(sc, dp, dirent_fn, priv);
-
-	return xchk_dir_walk_leaf(sc, dp, dirent_fn, priv);
+	case XFS_DIR2_FMT_LEAF:
+	case XFS_DIR2_FMT_NODE:
+		return xchk_dir_walk_leaf(sc, dp, dirent_fn, priv);
+	default:
+		return error;
+	}
 }
 
 /*
@@ -324,50 +320,102 @@ xchk_dir_lookup(
 		.hashval	= xfs_dir2_hashname(dp->i_mount, name),
 		.whichfork	= XFS_DATA_FORK,
 		.op_flags	= XFS_DA_OP_OKNOENT,
+		.owner		= dp->i_ino,
 	};
-	bool			isblock, isleaf;
 	int			error;
 
 	if (xfs_is_shutdown(dp->i_mount))
 		return -EIO;
 
+	/*
+	 * A temporary directory's block headers are written with the owner
+	 * set to sc->ip, so we must switch the owner here for the lookup.
+	 */
+	if (dp == sc->tempip)
+		args.owner = sc->ip->i_ino;
+
 	ASSERT(S_ISDIR(VFS_I(dp)->i_mode));
-	ASSERT(xfs_isilocked(dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL);
 
-	if (dp->i_df.if_format == XFS_DINODE_FMT_LOCAL) {
-		error = xfs_dir2_sf_lookup(&args);
-		goto out_check_rval;
-	}
+	error = xfs_dir_lookup_args(&args);
+	if (!error)
+		*ino = args.inumber;
+	return error;
+}
 
-	/* dir2 functions require that the data fork is loaded */
-	error = xfs_iread_extents(sc->tp, dp, XFS_DATA_FORK);
-	if (error)
-		return error;
+/*
+ * Try to grab the IOLOCK and ILOCK of sc->ip and ip, returning @ip's lock
+ * state.  The caller may have a transaction, so we must use trylock for both
+ * IOLOCKs.
+ */
+static inline unsigned int
+xchk_dir_trylock_both(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*ip)
+{
+	if (!xchk_ilock_nowait(sc, XFS_IOLOCK_EXCL))
+		return 0;
 
-	error = xfs_dir2_isblock(&args, &isblock);
-	if (error)
-		return error;
+	if (!xfs_ilock_nowait(ip, XFS_IOLOCK_SHARED))
+		goto parent_iolock;
 
-	if (isblock) {
-		error = xfs_dir2_block_lookup(&args);
-		goto out_check_rval;
-	}
+	xchk_ilock(sc, XFS_ILOCK_EXCL);
+	if (!xfs_ilock_nowait(ip, XFS_ILOCK_EXCL))
+		goto parent_ilock;
 
-	error = xfs_dir2_isleaf(&args, &isleaf);
-	if (error)
-		return error;
+	return XFS_IOLOCK_SHARED | XFS_ILOCK_EXCL;
+
+parent_ilock:
+	xchk_iunlock(sc, XFS_ILOCK_EXCL);
+	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
+parent_iolock:
+	xchk_iunlock(sc, XFS_IOLOCK_EXCL);
+	return 0;
+}
+
+/*
+ * Try for a limited time to grab the IOLOCK and ILOCK of both the scrub target
+ * (@sc->ip) and the inode at the other end (@ip) of a directory or parent
+ * pointer link so that we can check that link.
+ *
+ * We do not know ahead of time that the directory tree is /not/ corrupt, so we
+ * cannot use the "lock two inode" functions because we do not know that there
+ * is not a racing thread trying to take the locks in opposite order.  First
+ * take IOLOCK_EXCL of the scrub target, and then try to take IOLOCK_SHARED
+ * of @ip to synchronize with the VFS.  Next, take ILOCK_EXCL of the scrub
+ * target and @ip to synchronize with XFS.
+ *
+ * If the trylocks succeed, *lockmode will be set to the locks held for @ip;
+ * @sc->ilock_flags will be set for the locks held for @sc->ip; and zero will
+ * be returned.  If not, returns -EDEADLOCK to try again; or -ETIMEDOUT if
+ * XCHK_TRY_HARDER was set.  Returns -EINTR if the process has been killed.
+ */
+int
+xchk_dir_trylock_for_pptrs(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*ip,
+	unsigned int		*lockmode)
+{
+	unsigned int		nr;
+	int			error = 0;
+
+	ASSERT(sc->ilock_flags == 0);
+
+	for (nr = 0; nr < HZ; nr++) {
+		*lockmode = xchk_dir_trylock_both(sc, ip);
+		if (*lockmode)
+			return 0;
 
-	if (isleaf) {
-		error = xfs_dir2_leaf_lookup(&args);
-		goto out_check_rval;
+		if (xchk_should_terminate(sc, &error))
+			return error;
+
+		delay(1);
 	}
 
-	error = xfs_dir2_node_lookup(&args);
+	if (sc->flags & XCHK_TRY_HARDER) {
+		xchk_set_incomplete(sc);
+		return -ETIMEDOUT;
+	}
 
-out_check_rval:
-	if (error == -EEXIST)
-		error = 0;
-	if (!error)
-		*ino = args.inumber;
-	return error;
+	return -EDEADLOCK;
 }
diff --git a/fs/xfs/scrub/readdir.h b/fs/xfs/scrub/readdir.h
index 55787f4df1..da501877a6 100644
--- a/fs/xfs/scrub/readdir.h
+++ b/fs/xfs/scrub/readdir.h
@@ -16,4 +16,7 @@ int xchk_dir_walk(struct xfs_scrub *sc, struct xfs_inode *dp,
 int xchk_dir_lookup(struct xfs_scrub *sc, struct xfs_inode *dp,
 		const struct xfs_name *name, xfs_ino_t *ino);
 
+int xchk_dir_trylock_for_pptrs(struct xfs_scrub *sc, struct xfs_inode *ip,
+		unsigned int *lockmode);
+
 #endif /* __XFS_SCRUB_READDIR_H__ */
diff --git a/fs/xfs/scrub/reap.c b/fs/xfs/scrub/reap.c
index f99eca7998..be283153c2 100644
--- a/fs/xfs/scrub/reap.c
+++ b/fs/xfs/scrub/reap.c
@@ -114,7 +114,7 @@ xreap_put_freelist(
 	int			error;
 
 	/* Make sure there's space on the freelist. */
-	error = xrep_fix_freelist(sc, true);
+	error = xrep_fix_freelist(sc, 0);
 	if (error)
 		return error;
 
@@ -211,6 +211,48 @@ static inline void xreap_defer_finish_reset(struct xreap_state *rs)
 	rs->force_roll = false;
 }
 
+/*
+ * Compute the maximum length of a buffer cache scan (in units of sectors),
+ * given a quantity of fs blocks.
+ */
+xfs_daddr_t
+xrep_bufscan_max_sectors(
+	struct xfs_mount	*mp,
+	xfs_extlen_t		fsblocks)
+{
+	int			max_fsbs;
+
+	/* Remote xattr values are the largest buffers that we support. */
+	max_fsbs = xfs_attr3_max_rmt_blocks(mp);
+
+	return XFS_FSB_TO_BB(mp, min_t(xfs_extlen_t, fsblocks, max_fsbs));
+}
+
+/*
+ * Return an incore buffer from a sector scan, or NULL if there are no buffers
+ * left to return.
+ */
+struct xfs_buf *
+xrep_bufscan_advance(
+	struct xfs_mount	*mp,
+	struct xrep_bufscan	*scan)
+{
+	scan->__sector_count += scan->daddr_step;
+	while (scan->__sector_count <= scan->max_sectors) {
+		struct xfs_buf	*bp = NULL;
+		int		error;
+
+		error = xfs_buf_incore(mp->m_ddev_targp, scan->daddr,
+				scan->__sector_count, XBF_LIVESCAN, &bp);
+		if (!error)
+			return bp;
+
+		scan->__sector_count += scan->daddr_step;
+	}
+
+	return NULL;
+}
+
 /* Try to invalidate the incore buffers for an extent that we're freeing. */
 STATIC void
 xreap_agextent_binval(
@@ -241,28 +283,15 @@ xreap_agextent_binval(
 	 * of any plausible size.
 	 */
 	while (bno < agbno_next) {
-		xfs_agblock_t	fsbcount;
-		xfs_agblock_t	max_fsbs;
-
-		/*
-		 * Max buffer size is the max remote xattr buffer size, which
-		 * is one fs block larger than 64k.
-		 */
-		max_fsbs = min_t(xfs_agblock_t, agbno_next - bno,
-				xfs_attr3_rmt_blocks(mp, XFS_XATTR_SIZE_MAX));
-
-		for (fsbcount = 1; fsbcount <= max_fsbs; fsbcount++) {
-			struct xfs_buf	*bp = NULL;
-			xfs_daddr_t	daddr;
-			int		error;
-
-			daddr = XFS_AGB_TO_DADDR(mp, agno, bno);
-			error = xfs_buf_incore(mp->m_ddev_targp, daddr,
-					XFS_FSB_TO_BB(mp, fsbcount),
-					XBF_LIVESCAN, &bp);
-			if (error)
-				continue;
-
+		struct xrep_bufscan	scan = {
+			.daddr		= XFS_AGB_TO_DADDR(mp, agno, bno),
+			.max_sectors	= xrep_bufscan_max_sectors(mp,
+							agbno_next - bno),
+			.daddr_step	= XFS_FSB_TO_BB(mp, 1),
+		};
+		struct xfs_buf	*bp;
+
+		while ((bp = xrep_bufscan_advance(mp, &scan)) != NULL) {
 			xfs_trans_bjoin(sc->tp, bp);
 			xfs_trans_binval(sc->tp, bp);
 			rs->invalidated++;
@@ -646,3 +675,375 @@ xrep_reap_fsblocks(
 
 	return 0;
 }
+
+/*
+ * Metadata files are not supposed to share blocks with anything else.
+ * If blocks are shared, we remove the reverse mapping (thus reducing the
+ * crosslink factor); if blocks are not shared, we also need to free them.
+ *
+ * This first step determines the longest subset of the passed-in imap
+ * (starting at its beginning) that is either crosslinked or not crosslinked.
+ * The blockcount will be adjust down as needed.
+ */
+STATIC int
+xreap_bmapi_select(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*ip,
+	int			whichfork,
+	struct xfs_bmbt_irec	*imap,
+	bool			*crosslinked)
+{
+	struct xfs_owner_info	oinfo;
+	struct xfs_btree_cur	*cur;
+	xfs_filblks_t		len = 1;
+	xfs_agblock_t		bno;
+	xfs_agblock_t		agbno;
+	xfs_agblock_t		agbno_next;
+	int			error;
+
+	agbno = XFS_FSB_TO_AGBNO(sc->mp, imap->br_startblock);
+	agbno_next = agbno + imap->br_blockcount;
+
+	cur = xfs_rmapbt_init_cursor(sc->mp, sc->tp, sc->sa.agf_bp,
+			sc->sa.pag);
+
+	xfs_rmap_ino_owner(&oinfo, ip->i_ino, whichfork, imap->br_startoff);
+	error = xfs_rmap_has_other_keys(cur, agbno, 1, &oinfo, crosslinked);
+	if (error)
+		goto out_cur;
+
+	bno = agbno + 1;
+	while (bno < agbno_next) {
+		bool		also_crosslinked;
+
+		oinfo.oi_offset++;
+		error = xfs_rmap_has_other_keys(cur, bno, 1, &oinfo,
+				&also_crosslinked);
+		if (error)
+			goto out_cur;
+
+		if (also_crosslinked != *crosslinked)
+			break;
+
+		len++;
+		bno++;
+	}
+
+	imap->br_blockcount = len;
+	trace_xreap_bmapi_select(sc->sa.pag, agbno, len, *crosslinked);
+out_cur:
+	xfs_btree_del_cursor(cur, error);
+	return error;
+}
+
+/*
+ * Decide if this buffer can be joined to a transaction.  This is true for most
+ * buffers, but there are two cases that we want to catch: large remote xattr
+ * value buffers are not logged and can overflow the buffer log item dirty
+ * bitmap size; and oversized cached buffers if things have really gone
+ * haywire.
+ */
+static inline bool
+xreap_buf_loggable(
+	const struct xfs_buf	*bp)
+{
+	int			i;
+
+	for (i = 0; i < bp->b_map_count; i++) {
+		int		chunks;
+		int		map_size;
+
+		chunks = DIV_ROUND_UP(BBTOB(bp->b_maps[i].bm_len),
+				XFS_BLF_CHUNK);
+		map_size = DIV_ROUND_UP(chunks, NBWORD);
+		if (map_size > XFS_BLF_DATAMAP_SIZE)
+			return false;
+	}
+
+	return true;
+}
+
+/*
+ * Invalidate any buffers for this file mapping.  The @imap blockcount may be
+ * adjusted downward if we need to roll the transaction.
+ */
+STATIC int
+xreap_bmapi_binval(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*ip,
+	int			whichfork,
+	struct xfs_bmbt_irec	*imap)
+{
+	struct xfs_mount	*mp = sc->mp;
+	struct xfs_perag	*pag = sc->sa.pag;
+	int			bmap_flags = xfs_bmapi_aflag(whichfork);
+	xfs_fileoff_t		off;
+	xfs_fileoff_t		max_off;
+	xfs_extlen_t		scan_blocks;
+	xfs_agnumber_t		agno = sc->sa.pag->pag_agno;
+	xfs_agblock_t		bno;
+	xfs_agblock_t		agbno;
+	xfs_agblock_t		agbno_next;
+	unsigned int		invalidated = 0;
+	int			error;
+
+	/*
+	 * Avoid invalidating AG headers and post-EOFS blocks because we never
+	 * own those.
+	 */
+	agbno = bno = XFS_FSB_TO_AGBNO(sc->mp, imap->br_startblock);
+	agbno_next = agbno + imap->br_blockcount;
+	if (!xfs_verify_agbno(pag, agbno) ||
+	    !xfs_verify_agbno(pag, agbno_next - 1))
+		return 0;
+
+	/*
+	 * Buffers for file blocks can span multiple contiguous mappings.  This
+	 * means that for each block in the mapping, there could exist an
+	 * xfs_buf indexed by that block with any length up to the maximum
+	 * buffer size (remote xattr values) or to the next hole in the fork.
+	 * To set up our binval scan, first we need to figure out the location
+	 * of the next hole.
+	 */
+	off = imap->br_startoff + imap->br_blockcount;
+	max_off = off + xfs_attr3_max_rmt_blocks(mp);
+	while (off < max_off) {
+		struct xfs_bmbt_irec	hmap;
+		int			nhmaps = 1;
+
+		error = xfs_bmapi_read(ip, off, max_off - off, &hmap,
+				&nhmaps, bmap_flags);
+		if (error)
+			return error;
+		if (nhmaps != 1 || hmap.br_startblock == DELAYSTARTBLOCK) {
+			ASSERT(0);
+			return -EFSCORRUPTED;
+		}
+
+		if (!xfs_bmap_is_real_extent(&hmap))
+			break;
+
+		off = hmap.br_startoff + hmap.br_blockcount;
+	}
+	scan_blocks = off - imap->br_startoff;
+
+	trace_xreap_bmapi_binval_scan(sc, imap, scan_blocks);
+
+	/*
+	 * If there are incore buffers for these blocks, invalidate them.  If
+	 * we can't (try)lock the buffer we assume it's owned by someone else
+	 * and leave it alone.  The buffer cache cannot detect aliasing, so
+	 * employ nested loops to detect incore buffers of any plausible size.
+	 */
+	while (bno < agbno_next) {
+		struct xrep_bufscan	scan = {
+			.daddr		= XFS_AGB_TO_DADDR(mp, agno, bno),
+			.max_sectors	= xrep_bufscan_max_sectors(mp,
+								scan_blocks),
+			.daddr_step	= XFS_FSB_TO_BB(mp, 1),
+		};
+		struct xfs_buf		*bp;
+
+		while ((bp = xrep_bufscan_advance(mp, &scan)) != NULL) {
+			if (xreap_buf_loggable(bp)) {
+				xfs_trans_bjoin(sc->tp, bp);
+				xfs_trans_binval(sc->tp, bp);
+			} else {
+				xfs_buf_stale(bp);
+				xfs_buf_relse(bp);
+			}
+			invalidated++;
+
+			/*
+			 * Stop invalidating if we've hit the limit; we should
+			 * still have enough reservation left to free however
+			 * much of the mapping we've seen so far.
+			 */
+			if (invalidated > XREAP_MAX_BINVAL) {
+				imap->br_blockcount = agbno_next - bno;
+				goto out;
+			}
+		}
+
+		bno++;
+		scan_blocks--;
+	}
+
+out:
+	trace_xreap_bmapi_binval(sc->sa.pag, agbno, imap->br_blockcount);
+	return 0;
+}
+
+/*
+ * Dispose of as much of the beginning of this file fork mapping as possible.
+ * The number of blocks disposed of is returned in @imap->br_blockcount.
+ */
+STATIC int
+xrep_reap_bmapi_iter(
+	struct xfs_scrub		*sc,
+	struct xfs_inode		*ip,
+	int				whichfork,
+	struct xfs_bmbt_irec		*imap,
+	bool				crosslinked)
+{
+	int				error;
+
+	if (crosslinked) {
+		/*
+		 * If there are other rmappings, this block is cross linked and
+		 * must not be freed.  Remove the reverse mapping, leave the
+		 * buffer cache in its possibly confused state, and move on.
+		 * We don't want to risk discarding valid data buffers from
+		 * anybody else who thinks they own the block, even though that
+		 * runs the risk of stale buffer warnings in the future.
+		 */
+		trace_xreap_dispose_unmap_extent(sc->sa.pag,
+				XFS_FSB_TO_AGBNO(sc->mp, imap->br_startblock),
+				imap->br_blockcount);
+
+		/*
+		 * Schedule removal of the mapping from the fork.  We use
+		 * deferred log intents in this function to control the exact
+		 * sequence of metadata updates.
+		 */
+		xfs_bmap_unmap_extent(sc->tp, ip, whichfork, imap);
+		xfs_trans_mod_dquot_byino(sc->tp, ip, XFS_TRANS_DQ_BCOUNT,
+				-(int64_t)imap->br_blockcount);
+		xfs_rmap_unmap_extent(sc->tp, ip, whichfork, imap);
+		return 0;
+	}
+
+	/*
+	 * If the block is not crosslinked, we can invalidate all the incore
+	 * buffers for the extent, and then free the extent.  This is a bit of
+	 * a mess since we don't detect discontiguous buffers that are indexed
+	 * by a block starting before the first block of the extent but overlap
+	 * anyway.
+	 */
+	trace_xreap_dispose_free_extent(sc->sa.pag,
+			XFS_FSB_TO_AGBNO(sc->mp, imap->br_startblock),
+			imap->br_blockcount);
+
+	/*
+	 * Invalidate as many buffers as we can, starting at the beginning of
+	 * this mapping.  If this function sets blockcount to zero, the
+	 * transaction is full of logged buffer invalidations, so we need to
+	 * return early so that we can roll and retry.
+	 */
+	error = xreap_bmapi_binval(sc, ip, whichfork, imap);
+	if (error || imap->br_blockcount == 0)
+		return error;
+
+	/*
+	 * Schedule removal of the mapping from the fork.  We use deferred log
+	 * intents in this function to control the exact sequence of metadata
+	 * updates.
+	 */
+	xfs_bmap_unmap_extent(sc->tp, ip, whichfork, imap);
+	xfs_trans_mod_dquot_byino(sc->tp, ip, XFS_TRANS_DQ_BCOUNT,
+			-(int64_t)imap->br_blockcount);
+	return xfs_free_extent_later(sc->tp, imap->br_startblock,
+			imap->br_blockcount, NULL, XFS_AG_RESV_NONE, true);
+}
+
+/*
+ * Dispose of as much of this file extent as we can.  Upon successful return,
+ * the imap will reflect the mapping that was removed from the fork.
+ */
+STATIC int
+xreap_ifork_extent(
+	struct xfs_scrub		*sc,
+	struct xfs_inode		*ip,
+	int				whichfork,
+	struct xfs_bmbt_irec		*imap)
+{
+	xfs_agnumber_t			agno;
+	bool				crosslinked;
+	int				error;
+
+	ASSERT(sc->sa.pag == NULL);
+
+	trace_xreap_ifork_extent(sc, ip, whichfork, imap);
+
+	agno = XFS_FSB_TO_AGNO(sc->mp, imap->br_startblock);
+	sc->sa.pag = xfs_perag_get(sc->mp, agno);
+	if (!sc->sa.pag)
+		return -EFSCORRUPTED;
+
+	error = xfs_alloc_read_agf(sc->sa.pag, sc->tp, 0, &sc->sa.agf_bp);
+	if (error)
+		goto out_pag;
+
+	/*
+	 * Decide the fate of the blocks at the beginning of the mapping, then
+	 * update the mapping to use it with the unmap calls.
+	 */
+	error = xreap_bmapi_select(sc, ip, whichfork, imap, &crosslinked);
+	if (error)
+		goto out_agf;
+
+	error = xrep_reap_bmapi_iter(sc, ip, whichfork, imap, crosslinked);
+	if (error)
+		goto out_agf;
+
+out_agf:
+	xfs_trans_brelse(sc->tp, sc->sa.agf_bp);
+	sc->sa.agf_bp = NULL;
+out_pag:
+	xfs_perag_put(sc->sa.pag);
+	sc->sa.pag = NULL;
+	return error;
+}
+
+/*
+ * Dispose of each block mapped to the given fork of the given file.  Callers
+ * must hold ILOCK_EXCL, and ip can only be sc->ip or sc->tempip.  The fork
+ * must not have any delalloc reservations.
+ */
+int
+xrep_reap_ifork(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*ip,
+	int			whichfork)
+{
+	xfs_fileoff_t		off = 0;
+	int			bmap_flags = xfs_bmapi_aflag(whichfork);
+	int			error;
+
+	ASSERT(xfs_has_rmapbt(sc->mp));
+	ASSERT(ip == sc->ip || ip == sc->tempip);
+	ASSERT(whichfork == XFS_ATTR_FORK || !XFS_IS_REALTIME_INODE(ip));
+
+	while (off < XFS_MAX_FILEOFF) {
+		struct xfs_bmbt_irec	imap;
+		int			nimaps = 1;
+
+		/* Read the next extent, skip past holes and delalloc. */
+		error = xfs_bmapi_read(ip, off, XFS_MAX_FILEOFF - off, &imap,
+				&nimaps, bmap_flags);
+		if (error)
+			return error;
+		if (nimaps != 1 || imap.br_startblock == DELAYSTARTBLOCK) {
+			ASSERT(0);
+			return -EFSCORRUPTED;
+		}
+
+		/*
+		 * If this is a real space mapping, reap as much of it as we
+		 * can in a single transaction.
+		 */
+		if (xfs_bmap_is_real_extent(&imap)) {
+			error = xreap_ifork_extent(sc, ip, whichfork, &imap);
+			if (error)
+				return error;
+
+			error = xfs_defer_finish(&sc->tp);
+			if (error)
+				return error;
+		}
+
+		off = imap.br_startoff + imap.br_blockcount;
+	}
+
+	return 0;
+}
diff --git a/fs/xfs/scrub/reap.h b/fs/xfs/scrub/reap.h
index 0b69f16dd9..3f2f1775e2 100644
--- a/fs/xfs/scrub/reap.h
+++ b/fs/xfs/scrub/reap.h
@@ -13,5 +13,26 @@ int xrep_reap_agblocks(struct xfs_scrub *sc, struct xagb_bitmap *bitmap,
 		const struct xfs_owner_info *oinfo, enum xfs_ag_resv_type type);
 int xrep_reap_fsblocks(struct xfs_scrub *sc, struct xfsb_bitmap *bitmap,
 		const struct xfs_owner_info *oinfo);
+int xrep_reap_ifork(struct xfs_scrub *sc, struct xfs_inode *ip, int whichfork);
+
+/* Buffer cache scan context. */
+struct xrep_bufscan {
+	/* Disk address for the buffers we want to scan. */
+	xfs_daddr_t		daddr;
+
+	/* Maximum number of sectors to scan. */
+	xfs_daddr_t		max_sectors;
+
+	/* Each round, increment the search length by this number of sectors. */
+	xfs_daddr_t		daddr_step;
+
+	/* Internal scan state; initialize to zero. */
+	xfs_daddr_t		__sector_count;
+};
+
+xfs_daddr_t xrep_bufscan_max_sectors(struct xfs_mount *mp,
+		xfs_extlen_t fsblocks);
+struct xfs_buf *xrep_bufscan_advance(struct xfs_mount *mp,
+		struct xrep_bufscan *scan);
 
 #endif /* __XFS_SCRUB_REAP_H__ */
diff --git a/fs/xfs/scrub/refcount.c b/fs/xfs/scrub/refcount.c
index bf22f245bb..d0c7d4a29c 100644
--- a/fs/xfs/scrub/refcount.c
+++ b/fs/xfs/scrub/refcount.c
@@ -7,8 +7,10 @@
 #include "xfs_fs.h"
 #include "xfs_shared.h"
 #include "xfs_format.h"
+#include "xfs_log_format.h"
 #include "xfs_trans_resv.h"
 #include "xfs_mount.h"
+#include "xfs_trans.h"
 #include "xfs_ag.h"
 #include "xfs_btree.h"
 #include "xfs_rmap.h"
@@ -17,6 +19,7 @@
 #include "scrub/common.h"
 #include "scrub/btree.h"
 #include "scrub/trace.h"
+#include "scrub/repair.h"
 
 /*
  * Set us up to scrub reference count btrees.
@@ -27,6 +30,15 @@ xchk_setup_ag_refcountbt(
 {
 	if (xchk_need_intent_drain(sc))
 		xchk_fsgates_enable(sc, XCHK_FSGATES_DRAIN);
+
+	if (xchk_could_repair(sc)) {
+		int		error;
+
+		error = xrep_setup_ag_refcountbt(sc);
+		if (error)
+			return error;
+	}
+
 	return xchk_setup_ag_btree(sc, false);
 }
 
diff --git a/fs/xfs/scrub/refcount_repair.c b/fs/xfs/scrub/refcount_repair.c
index f38fccc42a..a00d7ce7ae 100644
--- a/fs/xfs/scrub/refcount_repair.c
+++ b/fs/xfs/scrub/refcount_repair.c
@@ -25,6 +25,7 @@
 #include "xfs_refcount_btree.h"
 #include "xfs_error.h"
 #include "xfs_ag.h"
+#include "xfs_health.h"
 #include "scrub/xfs_scrub.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
@@ -37,6 +38,7 @@
 #include "scrub/xfarray.h"
 #include "scrub/newbt.h"
 #include "scrub/reap.h"
+#include "scrub/rcbag.h"
 
 /*
  * Rebuilding the Reference Count Btree
@@ -97,12 +99,6 @@
  * insert all the records.
  */
 
-/* The only parts of the rmap that we care about for computing refcounts. */
-struct xrep_refc_rmap {
-	xfs_agblock_t		startblock;
-	xfs_extlen_t		blockcount;
-} __packed;
-
 struct xrep_refc {
 	/* refcount extents */
 	struct xfarray		*refcount_records;
@@ -122,6 +118,20 @@ struct xrep_refc {
 	xfs_extlen_t		btblocks;
 };
 
+/* Set us up to repair refcount btrees. */
+int
+xrep_setup_ag_refcountbt(
+	struct xfs_scrub	*sc)
+{
+	char			*descr;
+	int			error;
+
+	descr = xchk_xfile_ag_descr(sc, "rmap record bag");
+	error = xrep_setup_xfbtree(sc, descr);
+	kfree(descr);
+	return error;
+}
+
 /* Check for any obvious conflicts with this shared/CoW staging extent. */
 STATIC int
 xrep_refc_check_ext(
@@ -223,10 +233,9 @@ xrep_refc_rmap_shareable(
 STATIC int
 xrep_refc_walk_rmaps(
 	struct xrep_refc	*rr,
-	struct xrep_refc_rmap	*rrm,
+	struct xfs_rmap_irec	*rmap,
 	bool			*have_rec)
 {
-	struct xfs_rmap_irec	rmap;
 	struct xfs_btree_cur	*cur = rr->sc->sa.rmap_cur;
 	struct xfs_mount	*mp = cur->bc_mp;
 	int			have_gt;
@@ -250,29 +259,30 @@ xrep_refc_walk_rmaps(
 		if (!have_gt)
 			return 0;
 
-		error = xfs_rmap_get_rec(cur, &rmap, &have_gt);
+		error = xfs_rmap_get_rec(cur, rmap, &have_gt);
 		if (error)
 			return error;
-		if (XFS_IS_CORRUPT(mp, !have_gt))
+		if (XFS_IS_CORRUPT(mp, !have_gt)) {
+			xfs_btree_mark_sick(cur);
 			return -EFSCORRUPTED;
+		}
 
-		if (rmap.rm_owner == XFS_RMAP_OWN_COW) {
-			error = xrep_refc_stash_cow(rr, rmap.rm_startblock,
-					rmap.rm_blockcount);
+		if (rmap->rm_owner == XFS_RMAP_OWN_COW) {
+			error = xrep_refc_stash_cow(rr, rmap->rm_startblock,
+					rmap->rm_blockcount);
 			if (error)
 				return error;
-		} else if (rmap.rm_owner == XFS_RMAP_OWN_REFC) {
+		} else if (rmap->rm_owner == XFS_RMAP_OWN_REFC) {
 			/* refcountbt block, dump it when we're done. */
-			rr->btblocks += rmap.rm_blockcount;
+			rr->btblocks += rmap->rm_blockcount;
 			error = xagb_bitmap_set(&rr->old_refcountbt_blocks,
-					rmap.rm_startblock, rmap.rm_blockcount);
+					rmap->rm_startblock,
+					rmap->rm_blockcount);
 			if (error)
 				return error;
 		}
-	} while (!xrep_refc_rmap_shareable(mp, &rmap));
+	} while (!xrep_refc_rmap_shareable(mp, rmap));
 
-	rrm->startblock = rmap.rm_startblock;
-	rrm->blockcount = rmap.rm_blockcount;
 	*have_rec = true;
 	return 0;
 }
@@ -354,45 +364,6 @@ xrep_refc_sort_records(
 	return error;
 }
 
-#define RRM_NEXT(r)	((r).startblock + (r).blockcount)
-/*
- * Find the next block where the refcount changes, given the next rmap we
- * looked at and the ones we're already tracking.
- */
-static inline int
-xrep_refc_next_edge(
-	struct xfarray		*rmap_bag,
-	struct xrep_refc_rmap	*next_rrm,
-	bool			next_valid,
-	xfs_agblock_t		*nbnop)
-{
-	struct xrep_refc_rmap	rrm;
-	xfarray_idx_t		array_cur = XFARRAY_CURSOR_INIT;
-	xfs_agblock_t		nbno = NULLAGBLOCK;
-	int			error;
-
-	if (next_valid)
-		nbno = next_rrm->startblock;
-
-	while ((error = xfarray_iter(rmap_bag, &array_cur, &rrm)) == 1)
-		nbno = min_t(xfs_agblock_t, nbno, RRM_NEXT(rrm));
-
-	if (error)
-		return error;
-
-	/*
-	 * We should have found /something/ because either next_rrm is the next
-	 * interesting rmap to look at after emitting this refcount extent, or
-	 * there are other rmaps in rmap_bag contributing to the current
-	 * sharing count.  But if something is seriously wrong, bail out.
-	 */
-	if (nbno == NULLAGBLOCK)
-		return -EFSCORRUPTED;
-
-	*nbnop = nbno;
-	return 0;
-}
-
 /*
  * Walk forward through the rmap btree to collect all rmaps starting at
  * @bno in @rmap_bag.  These represent the file(s) that share ownership of
@@ -402,22 +373,21 @@ xrep_refc_next_edge(
 static int
 xrep_refc_push_rmaps_at(
 	struct xrep_refc	*rr,
-	struct xfarray		*rmap_bag,
+	struct rcbag		*rcstack,
 	xfs_agblock_t		bno,
-	struct xrep_refc_rmap	*rrm,
-	bool			*have,
-	uint64_t		*stack_sz)
+	struct xfs_rmap_irec	*rmap,
+	bool			*have)
 {
 	struct xfs_scrub	*sc = rr->sc;
 	int			have_gt;
 	int			error;
 
-	while (*have && rrm->startblock == bno) {
-		error = xfarray_store_anywhere(rmap_bag, rrm);
+	while (*have && rmap->rm_startblock == bno) {
+		error = rcbag_add(rcstack, rr->sc->tp, rmap);
 		if (error)
 			return error;
-		(*stack_sz)++;
-		error = xrep_refc_walk_rmaps(rr, rrm, have);
+
+		error = xrep_refc_walk_rmaps(rr, rmap, have);
 		if (error)
 			return error;
 	}
@@ -425,8 +395,10 @@ xrep_refc_push_rmaps_at(
 	error = xfs_btree_decrement(sc->sa.rmap_cur, 0, &have_gt);
 	if (error)
 		return error;
-	if (XFS_IS_CORRUPT(sc->mp, !have_gt))
+	if (XFS_IS_CORRUPT(sc->mp, !have_gt)) {
+		xfs_btree_mark_sick(sc->sa.rmap_cur);
 		return -EFSCORRUPTED;
+	}
 
 	return 0;
 }
@@ -436,12 +408,9 @@ STATIC int
 xrep_refc_find_refcounts(
 	struct xrep_refc	*rr)
 {
-	struct xrep_refc_rmap	rrm;
 	struct xfs_scrub	*sc = rr->sc;
-	struct xfarray		*rmap_bag;
-	char			*descr;
-	uint64_t		old_stack_sz;
-	uint64_t		stack_sz = 0;
+	struct rcbag		*rcstack;
+	uint64_t		old_stack_height;
 	xfs_agblock_t		sbno;
 	xfs_agblock_t		cbno;
 	xfs_agblock_t		nbno;
@@ -451,14 +420,11 @@ xrep_refc_find_refcounts(
 	xrep_ag_btcur_init(sc, &sc->sa);
 
 	/*
-	 * Set up a sparse array to store all the rmap records that we're
-	 * tracking to generate a reference count record.  If this exceeds
+	 * Set up a bag to store all the rmap records that we're tracking to
+	 * generate a reference count record.  If the size of the bag exceeds
 	 * MAXREFCOUNT, we clamp rc_refcount.
 	 */
-	descr = xchk_xfile_ag_descr(sc, "rmap record bag");
-	error = xfarray_create(descr, 0, sizeof(struct xrep_refc_rmap),
-			&rmap_bag);
-	kfree(descr);
+	error = rcbag_init(sc->mp, sc->xmbtp, &rcstack);
 	if (error)
 		goto out_cur;
 
@@ -469,62 +435,54 @@ xrep_refc_find_refcounts(
 
 	/* Process reverse mappings into refcount data. */
 	while (xfs_btree_has_more_records(sc->sa.rmap_cur)) {
+		struct xfs_rmap_irec	rmap;
+
 		/* Push all rmaps with pblk == sbno onto the stack */
-		error = xrep_refc_walk_rmaps(rr, &rrm, &have);
+		error = xrep_refc_walk_rmaps(rr, &rmap, &have);
 		if (error)
 			goto out_bag;
 		if (!have)
 			break;
-		sbno = cbno = rrm.startblock;
-		error = xrep_refc_push_rmaps_at(rr, rmap_bag, sbno,
-					&rrm, &have, &stack_sz);
+		sbno = cbno = rmap.rm_startblock;
+		error = xrep_refc_push_rmaps_at(rr, rcstack, sbno, &rmap,
+				&have);
 		if (error)
 			goto out_bag;
 
 		/* Set nbno to the bno of the next refcount change */
-		error = xrep_refc_next_edge(rmap_bag, &rrm, have, &nbno);
+		error = rcbag_next_edge(rcstack, sc->tp, &rmap, have, &nbno);
 		if (error)
 			goto out_bag;
 
 		ASSERT(nbno > sbno);
-		old_stack_sz = stack_sz;
+		old_stack_height = rcbag_count(rcstack);
 
 		/* While stack isn't empty... */
-		while (stack_sz) {
-			xfarray_idx_t	array_cur = XFARRAY_CURSOR_INIT;
-
+		while (rcbag_count(rcstack) > 0) {
 			/* Pop all rmaps that end at nbno */
-			while ((error = xfarray_iter(rmap_bag, &array_cur,
-								&rrm)) == 1) {
-				if (RRM_NEXT(rrm) != nbno)
-					continue;
-				error = xfarray_unset(rmap_bag, array_cur - 1);
-				if (error)
-					goto out_bag;
-				stack_sz--;
-			}
+			error = rcbag_remove_ending_at(rcstack, sc->tp, nbno);
 			if (error)
 				goto out_bag;
 
 			/* Push array items that start at nbno */
-			error = xrep_refc_walk_rmaps(rr, &rrm, &have);
+			error = xrep_refc_walk_rmaps(rr, &rmap, &have);
 			if (error)
 				goto out_bag;
 			if (have) {
-				error = xrep_refc_push_rmaps_at(rr, rmap_bag,
-						nbno, &rrm, &have, &stack_sz);
+				error = xrep_refc_push_rmaps_at(rr, rcstack,
+						nbno, &rmap, &have);
 				if (error)
 					goto out_bag;
 			}
 
 			/* Emit refcount if necessary */
 			ASSERT(nbno > cbno);
-			if (stack_sz != old_stack_sz) {
-				if (old_stack_sz > 1) {
+			if (rcbag_count(rcstack) != old_stack_height) {
+				if (old_stack_height > 1) {
 					error = xrep_refc_stash(rr,
 							XFS_REFC_DOMAIN_SHARED,
 							cbno, nbno - cbno,
-							old_stack_sz);
+							old_stack_height);
 					if (error)
 						goto out_bag;
 				}
@@ -532,13 +490,13 @@ xrep_refc_find_refcounts(
 			}
 
 			/* Stack empty, go find the next rmap */
-			if (stack_sz == 0)
+			if (rcbag_count(rcstack) == 0)
 				break;
-			old_stack_sz = stack_sz;
+			old_stack_height = rcbag_count(rcstack);
 			sbno = nbno;
 
 			/* Set nbno to the bno of the next refcount change */
-			error = xrep_refc_next_edge(rmap_bag, &rrm, have,
+			error = rcbag_next_edge(rcstack, sc->tp, &rmap, have,
 					&nbno);
 			if (error)
 				goto out_bag;
@@ -547,14 +505,13 @@ xrep_refc_find_refcounts(
 		}
 	}
 
-	ASSERT(stack_sz == 0);
+	ASSERT(rcbag_count(rcstack) == 0);
 out_bag:
-	xfarray_destroy(rmap_bag);
+	rcbag_free(&rcstack);
 out_cur:
 	xchk_ag_btcur_free(&sc->sa);
 	return error;
 }
-#undef RRM_NEXT
 
 /* Retrieve refcountbt data for bulk load. */
 STATIC int
@@ -653,8 +610,8 @@ xrep_refc_build_new_tree(
 	rr->new_btree.bload.claim_block = xrep_refc_claim_block;
 
 	/* Compute how many blocks we'll need. */
-	refc_cur = xfs_refcountbt_stage_cursor(sc->mp, &rr->new_btree.afake,
-			pag);
+	refc_cur = xfs_refcountbt_init_cursor(sc->mp, NULL, NULL, pag);
+	xfs_btree_stage_afakeroot(refc_cur, &rr->new_btree.afake);
 	error = xfs_btree_bload_compute_geometry(refc_cur,
 			&rr->new_btree.bload,
 			xfarray_length(rr->refcount_records));
diff --git a/fs/xfs/scrub/repair.c b/fs/xfs/scrub/repair.c
index 745d5b8f40..67478294f1 100644
--- a/fs/xfs/scrub/repair.c
+++ b/fs/xfs/scrub/repair.c
@@ -30,12 +30,20 @@
 #include "xfs_errortag.h"
 #include "xfs_error.h"
 #include "xfs_reflink.h"
+#include "xfs_health.h"
+#include "xfs_buf_mem.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
+#include "xfs_attr.h"
+#include "xfs_dir2.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
 #include "scrub/repair.h"
 #include "scrub/bitmap.h"
 #include "scrub/stats.h"
+#include "scrub/xfile.h"
+#include "scrub/attr_repair.h"
 
 /*
  * Attempt to repair some metadata, if the metadata is corrupt and userspace
@@ -287,7 +295,7 @@ xrep_calc_ag_resblks(
 		icount = pag->pagi_count;
 	} else {
 		/* Try to get the actual counters from disk. */
-		error = xfs_ialloc_read_agi(pag, NULL, &bp);
+		error = xfs_ialloc_read_agi(pag, NULL, 0, &bp);
 		if (!error) {
 			icount = pag->pagi_count;
 			xfs_buf_relse(bp);
@@ -400,7 +408,7 @@ xrep_calc_ag_resblks(
 int
 xrep_fix_freelist(
 	struct xfs_scrub	*sc,
-	bool			can_shrink)
+	int			alloc_flags)
 {
 	struct xfs_alloc_arg	args = {0};
 
@@ -410,8 +418,7 @@ xrep_fix_freelist(
 	args.alignment = 1;
 	args.pag = sc->sa.pag;
 
-	return xfs_alloc_fix_freelist(&args,
-			can_shrink ? 0 : XFS_ALLOC_FLAG_NOSHRINK);
+	return xfs_alloc_fix_freelist(&args, alloc_flags);
 }
 
 /*
@@ -687,6 +694,44 @@ xrep_find_ag_btree_roots(
 }
 
 #ifdef CONFIG_XFS_QUOTA
+/* Update some quota flags in the superblock. */
+void
+xrep_update_qflags(
+	struct xfs_scrub	*sc,
+	unsigned int		clear_flags,
+	unsigned int		set_flags)
+{
+	struct xfs_mount	*mp = sc->mp;
+	struct xfs_buf		*bp;
+
+	mutex_lock(&mp->m_quotainfo->qi_quotaofflock);
+	if ((mp->m_qflags & clear_flags) == 0 &&
+	    (mp->m_qflags & set_flags) == set_flags)
+		goto no_update;
+
+	mp->m_qflags &= ~clear_flags;
+	mp->m_qflags |= set_flags;
+
+	spin_lock(&mp->m_sb_lock);
+	mp->m_sb.sb_qflags &= ~clear_flags;
+	mp->m_sb.sb_qflags |= set_flags;
+	spin_unlock(&mp->m_sb_lock);
+
+	/*
+	 * Update the quota flags in the ondisk superblock without touching
+	 * the summary counters.  We have not quiesced inode chunk allocation,
+	 * so we cannot coordinate with updates to the icount and ifree percpu
+	 * counters.
+	 */
+	bp = xfs_trans_getsb(sc->tp);
+	xfs_sb_to_disk(bp->b_addr, &mp->m_sb);
+	xfs_trans_buf_set_type(sc->tp, bp, XFS_BLFT_SB_BUF);
+	xfs_trans_log_buf(sc->tp, bp, 0, sizeof(struct xfs_dsb) - 1);
+
+no_update:
+	mutex_unlock(&mp->m_quotainfo->qi_quotaofflock);
+}
+
 /* Force a quotacheck the next time we mount. */
 void
 xrep_force_quotacheck(
@@ -699,13 +744,7 @@ xrep_force_quotacheck(
 	if (!(flag & sc->mp->m_qflags))
 		return;
 
-	mutex_lock(&sc->mp->m_quotainfo->qi_quotaofflock);
-	sc->mp->m_qflags &= ~flag;
-	spin_lock(&sc->mp->m_sb_lock);
-	sc->mp->m_sb.sb_qflags &= ~flag;
-	spin_unlock(&sc->mp->m_sb_lock);
-	xfs_log_sb(sc->tp);
-	mutex_unlock(&sc->mp->m_quotainfo->qi_quotaofflock);
+	xrep_update_qflags(sc, flag, 0);
 }
 
 /*
@@ -799,20 +838,20 @@ xrep_ag_btcur_init(
 	/* Set up a bnobt cursor for cross-referencing. */
 	if (sc->sm->sm_type != XFS_SCRUB_TYPE_BNOBT &&
 	    sc->sm->sm_type != XFS_SCRUB_TYPE_CNTBT) {
-		sa->bno_cur = xfs_allocbt_init_cursor(mp, sc->tp, sa->agf_bp,
-				sc->sa.pag, XFS_BTNUM_BNO);
-		sa->cnt_cur = xfs_allocbt_init_cursor(mp, sc->tp, sa->agf_bp,
-				sc->sa.pag, XFS_BTNUM_CNT);
+		sa->bno_cur = xfs_bnobt_init_cursor(mp, sc->tp, sa->agf_bp,
+				sc->sa.pag);
+		sa->cnt_cur = xfs_cntbt_init_cursor(mp, sc->tp, sa->agf_bp,
+				sc->sa.pag);
 	}
 
 	/* Set up a inobt cursor for cross-referencing. */
 	if (sc->sm->sm_type != XFS_SCRUB_TYPE_INOBT &&
 	    sc->sm->sm_type != XFS_SCRUB_TYPE_FINOBT) {
 		sa->ino_cur = xfs_inobt_init_cursor(sc->sa.pag, sc->tp,
-				sa->agi_bp, XFS_BTNUM_INO);
+				sa->agi_bp);
 		if (xfs_has_finobt(mp))
-			sa->fino_cur = xfs_inobt_init_cursor(sc->sa.pag,
-					sc->tp, sa->agi_bp, XFS_BTNUM_FINO);
+			sa->fino_cur = xfs_finobt_init_cursor(sc->sa.pag,
+					sc->tp, sa->agi_bp);
 	}
 
 	/* Set up a rmapbt cursor for cross-referencing. */
@@ -874,7 +913,7 @@ xrep_reinit_pagi(
 	ASSERT(xfs_perag_initialised_agi(pag));
 
 	clear_bit(XFS_AGSTATE_AGI_INIT, &pag->pag_opstate);
-	error = xfs_ialloc_read_agi(pag, sc->tp, &bp);
+	error = xfs_ialloc_read_agi(pag, sc->tp, 0, &bp);
 	if (error)
 		return error;
 
@@ -900,7 +939,7 @@ xrep_ag_init(
 
 	ASSERT(!sa->pag);
 
-	error = xfs_ialloc_read_agi(pag, sc->tp, &sa->agi_bp);
+	error = xfs_ialloc_read_agi(pag, sc->tp, 0, &sa->agi_bp);
 	if (error)
 		return error;
 
@@ -929,9 +968,7 @@ xrep_reset_perag_resv(
 	ASSERT(sc->tp);
 
 	sc->flags &= ~XREP_RESET_PERAG_RESV;
-	error = xfs_ag_resv_free(sc->sa.pag);
-	if (error)
-		goto out;
+	xfs_ag_resv_free(sc->sa.pag);
 	error = xfs_ag_resv_init(sc->sa.pag, sc->tp);
 	if (error == -ENOSPC) {
 		xfs_err(sc->mp,
@@ -940,7 +977,6 @@ xrep_reset_perag_resv(
 		error = 0;
 	}
 
-out:
 	return error;
 }
 
@@ -970,55 +1006,27 @@ xrep_metadata_inode_subtype(
 	struct xfs_scrub	*sc,
 	unsigned int		scrub_type)
 {
-	__u32			smtype = sc->sm->sm_type;
-	__u32			smflags = sc->sm->sm_flags;
-	unsigned int		sick_mask = sc->sick_mask;
+	struct xfs_scrub_subord	*sub;
 	int			error;
 
 	/*
-	 * Let's see if the inode needs repair.  We're going to open-code calls
-	 * to the scrub and repair functions so that we can hang on to the
+	 * Let's see if the inode needs repair.  Use a subordinate scrub context
+	 * to call the scrub and repair functions so that we can hang on to the
 	 * resources that we already acquired instead of using the standard
 	 * setup/teardown routines.
 	 */
-	sc->sm->sm_flags &= ~XFS_SCRUB_FLAGS_OUT;
-	sc->sm->sm_type = scrub_type;
-
-	switch (scrub_type) {
-	case XFS_SCRUB_TYPE_INODE:
-		error = xchk_inode(sc);
-		break;
-	case XFS_SCRUB_TYPE_BMBTD:
-		error = xchk_bmap_data(sc);
-		break;
-	case XFS_SCRUB_TYPE_BMBTA:
-		error = xchk_bmap_attr(sc);
-		break;
-	default:
-		ASSERT(0);
-		error = -EFSCORRUPTED;
-	}
+	sub = xchk_scrub_create_subord(sc, scrub_type);
+	error = sub->sc.ops->scrub(&sub->sc);
 	if (error)
 		goto out;
-
-	if (!xrep_will_attempt(sc))
+	if (!xrep_will_attempt(&sub->sc))
 		goto out;
 
 	/*
 	 * Repair some part of the inode.  This will potentially join the inode
 	 * to the transaction.
 	 */
-	switch (scrub_type) {
-	case XFS_SCRUB_TYPE_INODE:
-		error = xrep_inode(sc);
-		break;
-	case XFS_SCRUB_TYPE_BMBTD:
-		error = xrep_bmap(sc, XFS_DATA_FORK, false);
-		break;
-	case XFS_SCRUB_TYPE_BMBTA:
-		error = xrep_bmap(sc, XFS_ATTR_FORK, false);
-		break;
-	}
+	error = sub->sc.ops->repair(&sub->sc);
 	if (error)
 		goto out;
 
@@ -1027,10 +1035,10 @@ xrep_metadata_inode_subtype(
 	 * that the inode will not be joined to the transaction when we exit
 	 * the function.
 	 */
-	error = xfs_defer_finish(&sc->tp);
+	error = xfs_defer_finish(&sub->sc.tp);
 	if (error)
 		goto out;
-	error = xfs_trans_roll(&sc->tp);
+	error = xfs_trans_roll(&sub->sc.tp);
 	if (error)
 		goto out;
 
@@ -1038,31 +1046,18 @@ xrep_metadata_inode_subtype(
 	 * Clear the corruption flags and re-check the metadata that we just
 	 * repaired.
 	 */
-	sc->sm->sm_flags &= ~XFS_SCRUB_FLAGS_OUT;
-
-	switch (scrub_type) {
-	case XFS_SCRUB_TYPE_INODE:
-		error = xchk_inode(sc);
-		break;
-	case XFS_SCRUB_TYPE_BMBTD:
-		error = xchk_bmap_data(sc);
-		break;
-	case XFS_SCRUB_TYPE_BMBTA:
-		error = xchk_bmap_attr(sc);
-		break;
-	}
+	sub->sc.sm->sm_flags &= ~XFS_SCRUB_FLAGS_OUT;
+	error = sub->sc.ops->scrub(&sub->sc);
 	if (error)
 		goto out;
 
 	/* If corruption persists, the repair has failed. */
-	if (xchk_needs_repair(sc->sm)) {
+	if (xchk_needs_repair(sub->sc.sm)) {
 		error = -EFSCORRUPTED;
 		goto out;
 	}
 out:
-	sc->sick_mask = sick_mask;
-	sc->sm->sm_type = smtype;
-	sc->sm->sm_flags = smflags;
+	xchk_scrub_free_subord(sub);
 	return error;
 }
 
@@ -1102,6 +1097,17 @@ xrep_metadata_inode_forks(
 			return error;
 	}
 
+	/* Clear the attr forks since metadata shouldn't have that. */
+	if (xfs_inode_hasattr(sc->ip)) {
+		if (!dirty) {
+			dirty = true;
+			xfs_trans_ijoin(sc->tp, sc->ip, 0);
+		}
+		error = xrep_xattr_reset_fork(sc);
+		if (error)
+			return error;
+	}
+
 	/*
 	 * If we modified the inode, roll the transaction but don't rejoin the
 	 * inode to the new transaction because xrep_bmap_data can do that.
@@ -1115,3 +1121,86 @@ xrep_metadata_inode_forks(
 
 	return 0;
 }
+
+/*
+ * Set up an in-memory buffer cache so that we can use the xfbtree.  Allocating
+ * a shmem file might take loks, so we cannot be in transaction context.  Park
+ * our resources in the scrub context and let the teardown function take care
+ * of them at the right time.
+ */
+int
+xrep_setup_xfbtree(
+	struct xfs_scrub	*sc,
+	const char		*descr)
+{
+	ASSERT(sc->tp == NULL);
+
+	return xmbuf_alloc(sc->mp, descr, &sc->xmbtp);
+}
+
+/*
+ * Create a dummy transaction for use in a live update hook function.  This
+ * function MUST NOT be called from regular repair code because the current
+ * process' transaction is saved via the cookie.
+ */
+int
+xrep_trans_alloc_hook_dummy(
+	struct xfs_mount	*mp,
+	void			**cookiep,
+	struct xfs_trans	**tpp)
+{
+	int			error;
+
+	*cookiep = current->journal_info;
+	current->journal_info = NULL;
+
+	error = xfs_trans_alloc_empty(mp, tpp);
+	if (!error)
+		return 0;
+
+	current->journal_info = *cookiep;
+	*cookiep = NULL;
+	return error;
+}
+
+/* Cancel a dummy transaction used by a live update hook function. */
+void
+xrep_trans_cancel_hook_dummy(
+	void			**cookiep,
+	struct xfs_trans	*tp)
+{
+	xfs_trans_cancel(tp);
+	current->journal_info = *cookiep;
+	*cookiep = NULL;
+}
+
+/*
+ * See if this buffer can pass the given ->verify_struct() function.
+ *
+ * If the buffer already has ops attached and they're not the ones that were
+ * passed in, we reject the buffer.  Otherwise, we perform the structure test
+ * (note that we do not check CRCs) and return the outcome of the test.  The
+ * buffer ops and error state are left unchanged.
+ */
+bool
+xrep_buf_verify_struct(
+	struct xfs_buf			*bp,
+	const struct xfs_buf_ops	*ops)
+{
+	const struct xfs_buf_ops	*old_ops = bp->b_ops;
+	xfs_failaddr_t			fa;
+	int				old_error;
+
+	if (old_ops) {
+		if (old_ops != ops)
+			return false;
+	}
+
+	old_error = bp->b_error;
+	bp->b_ops = ops;
+	fa = bp->b_ops->verify_struct(bp);
+	bp->b_ops = old_ops;
+	bp->b_error = old_error;
+
+	return fa == NULL;
+}
diff --git a/fs/xfs/scrub/repair.h b/fs/xfs/scrub/repair.h
index 17114327e6..0e0dc2bf98 100644
--- a/fs/xfs/scrub/repair.h
+++ b/fs/xfs/scrub/repair.h
@@ -51,7 +51,7 @@ struct xbitmap;
 struct xagb_bitmap;
 struct xfsb_bitmap;
 
-int xrep_fix_freelist(struct xfs_scrub *sc, bool can_shrink);
+int xrep_fix_freelist(struct xfs_scrub *sc, int alloc_flags);
 
 struct xrep_find_ag_btree {
 	/* in: rmap owner of the btree we're looking for */
@@ -72,6 +72,8 @@ int xrep_find_ag_btree_roots(struct xfs_scrub *sc, struct xfs_buf *agf_bp,
 		struct xrep_find_ag_btree *btree_info, struct xfs_buf *agfl_bp);
 
 #ifdef CONFIG_XFS_QUOTA
+void xrep_update_qflags(struct xfs_scrub *sc, unsigned int clear_flags,
+		unsigned int set_flags);
 void xrep_force_quotacheck(struct xfs_scrub *sc, xfs_dqtype_t type);
 int xrep_ino_dqattach(struct xfs_scrub *sc);
 #else
@@ -79,11 +81,21 @@ int xrep_ino_dqattach(struct xfs_scrub *sc);
 # define xrep_ino_dqattach(sc)			(0)
 #endif /* CONFIG_XFS_QUOTA */
 
+int xrep_setup_xfbtree(struct xfs_scrub *sc, const char *descr);
+
 int xrep_ino_ensure_extent_count(struct xfs_scrub *sc, int whichfork,
 		xfs_extnum_t nextents);
 int xrep_reset_perag_resv(struct xfs_scrub *sc);
 int xrep_bmap(struct xfs_scrub *sc, int whichfork, bool allow_unwritten);
 int xrep_metadata_inode_forks(struct xfs_scrub *sc);
+int xrep_setup_ag_rmapbt(struct xfs_scrub *sc);
+int xrep_setup_ag_refcountbt(struct xfs_scrub *sc);
+int xrep_setup_xattr(struct xfs_scrub *sc);
+int xrep_setup_directory(struct xfs_scrub *sc);
+int xrep_setup_parent(struct xfs_scrub *sc);
+int xrep_setup_nlinks(struct xfs_scrub *sc);
+int xrep_setup_symlink(struct xfs_scrub *sc, unsigned int *resblks);
+int xrep_setup_dirtree(struct xfs_scrub *sc);
 
 /* Repair setup functions */
 int xrep_setup_ag_allocbt(struct xfs_scrub *sc);
@@ -109,27 +121,45 @@ int xrep_agfl(struct xfs_scrub *sc);
 int xrep_agi(struct xfs_scrub *sc);
 int xrep_allocbt(struct xfs_scrub *sc);
 int xrep_iallocbt(struct xfs_scrub *sc);
+int xrep_rmapbt(struct xfs_scrub *sc);
 int xrep_refcountbt(struct xfs_scrub *sc);
 int xrep_inode(struct xfs_scrub *sc);
 int xrep_bmap_data(struct xfs_scrub *sc);
 int xrep_bmap_attr(struct xfs_scrub *sc);
 int xrep_bmap_cow(struct xfs_scrub *sc);
+int xrep_nlinks(struct xfs_scrub *sc);
+int xrep_fscounters(struct xfs_scrub *sc);
+int xrep_xattr(struct xfs_scrub *sc);
+int xrep_directory(struct xfs_scrub *sc);
+int xrep_parent(struct xfs_scrub *sc);
+int xrep_symlink(struct xfs_scrub *sc);
+int xrep_dirtree(struct xfs_scrub *sc);
 
 #ifdef CONFIG_XFS_RT
 int xrep_rtbitmap(struct xfs_scrub *sc);
+int xrep_rtsummary(struct xfs_scrub *sc);
 #else
 # define xrep_rtbitmap			xrep_notsupported
+# define xrep_rtsummary			xrep_notsupported
 #endif /* CONFIG_XFS_RT */
 
 #ifdef CONFIG_XFS_QUOTA
 int xrep_quota(struct xfs_scrub *sc);
+int xrep_quotacheck(struct xfs_scrub *sc);
 #else
 # define xrep_quota			xrep_notsupported
+# define xrep_quotacheck		xrep_notsupported
 #endif /* CONFIG_XFS_QUOTA */
 
 int xrep_reinit_pagf(struct xfs_scrub *sc);
 int xrep_reinit_pagi(struct xfs_scrub *sc);
 
+int xrep_trans_alloc_hook_dummy(struct xfs_mount *mp, void **cookiep,
+		struct xfs_trans **tpp);
+void xrep_trans_cancel_hook_dummy(void **cookiep, struct xfs_trans *tp);
+
+bool xrep_buf_verify_struct(struct xfs_buf *bp, const struct xfs_buf_ops *ops);
+
 #else
 
 #define xrep_ino_dqattach(sc)	(0)
@@ -171,9 +201,21 @@ xrep_setup_nothing(
 	return 0;
 }
 #define xrep_setup_ag_allocbt		xrep_setup_nothing
+#define xrep_setup_ag_rmapbt		xrep_setup_nothing
+#define xrep_setup_ag_refcountbt	xrep_setup_nothing
+#define xrep_setup_xattr		xrep_setup_nothing
+#define xrep_setup_directory		xrep_setup_nothing
+#define xrep_setup_parent		xrep_setup_nothing
+#define xrep_setup_nlinks		xrep_setup_nothing
+#define xrep_setup_dirtree		xrep_setup_nothing
 
 #define xrep_setup_inode(sc, imap)	((void)0)
 
+static inline int xrep_setup_symlink(struct xfs_scrub *sc, unsigned int *x)
+{
+	return 0;
+}
+
 #define xrep_revalidate_allocbt		(NULL)
 #define xrep_revalidate_iallocbt	(NULL)
 
@@ -184,6 +226,7 @@ xrep_setup_nothing(
 #define xrep_agi			xrep_notsupported
 #define xrep_allocbt			xrep_notsupported
 #define xrep_iallocbt			xrep_notsupported
+#define xrep_rmapbt			xrep_notsupported
 #define xrep_refcountbt			xrep_notsupported
 #define xrep_inode			xrep_notsupported
 #define xrep_bmap_data			xrep_notsupported
@@ -191,6 +234,15 @@ xrep_setup_nothing(
 #define xrep_bmap_cow			xrep_notsupported
 #define xrep_rtbitmap			xrep_notsupported
 #define xrep_quota			xrep_notsupported
+#define xrep_quotacheck			xrep_notsupported
+#define xrep_nlinks			xrep_notsupported
+#define xrep_fscounters			xrep_notsupported
+#define xrep_rtsummary			xrep_notsupported
+#define xrep_xattr			xrep_notsupported
+#define xrep_directory			xrep_notsupported
+#define xrep_parent			xrep_notsupported
+#define xrep_symlink			xrep_notsupported
+#define xrep_dirtree			xrep_notsupported
 
 #endif /* CONFIG_XFS_ONLINE_REPAIR */
 
diff --git a/fs/xfs/scrub/rmap.c b/fs/xfs/scrub/rmap.c
index c99d1714f2..ba5bbc3fb7 100644
--- a/fs/xfs/scrub/rmap.c
+++ b/fs/xfs/scrub/rmap.c
@@ -25,6 +25,7 @@
 #include "scrub/btree.h"
 #include "scrub/bitmap.h"
 #include "scrub/agb_bitmap.h"
+#include "scrub/repair.h"
 
 /*
  * Set us up to scrub reverse mapping btrees.
@@ -36,6 +37,14 @@ xchk_setup_ag_rmapbt(
 	if (xchk_need_intent_drain(sc))
 		xchk_fsgates_enable(sc, XCHK_FSGATES_DRAIN);
 
+	if (xchk_could_repair(sc)) {
+		int		error;
+
+		error = xrep_setup_ag_rmapbt(sc);
+		if (error)
+			return error;
+	}
+
 	return xchk_setup_ag_btree(sc, false);
 }
 
@@ -349,7 +358,7 @@ xchk_rmapbt_rec(
 	struct xfs_rmap_irec	irec;
 
 	if (xfs_rmap_btrec_to_irec(rec, &irec) != NULL ||
-	    xfs_rmap_check_irec(bs->cur, &irec) != NULL) {
+	    xfs_rmap_check_irec(bs->cur->bc_ag.pag, &irec) != NULL) {
 		xchk_btree_set_corrupt(bs->sc, bs->cur, 0);
 		return 0;
 	}
@@ -412,8 +421,8 @@ xchk_rmapbt_walk_ag_metadata(
 	/* OWN_AG: bnobt, cntbt, rmapbt, and AGFL */
 	cur = sc->sa.bno_cur;
 	if (!cur)
-		cur = xfs_allocbt_init_cursor(sc->mp, sc->tp, sc->sa.agf_bp,
-				sc->sa.pag, XFS_BTNUM_BNO);
+		cur = xfs_bnobt_init_cursor(sc->mp, sc->tp, sc->sa.agf_bp,
+				sc->sa.pag);
 	error = xagb_bitmap_set_btblocks(&cr->ag_owned, cur);
 	if (cur != sc->sa.bno_cur)
 		xfs_btree_del_cursor(cur, error);
@@ -422,8 +431,8 @@ xchk_rmapbt_walk_ag_metadata(
 
 	cur = sc->sa.cnt_cur;
 	if (!cur)
-		cur = xfs_allocbt_init_cursor(sc->mp, sc->tp, sc->sa.agf_bp,
-				sc->sa.pag, XFS_BTNUM_CNT);
+		cur = xfs_cntbt_init_cursor(sc->mp, sc->tp, sc->sa.agf_bp,
+				sc->sa.pag);
 	error = xagb_bitmap_set_btblocks(&cr->ag_owned, cur);
 	if (cur != sc->sa.cnt_cur)
 		xfs_btree_del_cursor(cur, error);
@@ -447,8 +456,7 @@ xchk_rmapbt_walk_ag_metadata(
 	/* OWN_INOBT: inobt, finobt */
 	cur = sc->sa.ino_cur;
 	if (!cur)
-		cur = xfs_inobt_init_cursor(sc->sa.pag, sc->tp, sc->sa.agi_bp,
-				XFS_BTNUM_INO);
+		cur = xfs_inobt_init_cursor(sc->sa.pag, sc->tp, sc->sa.agi_bp);
 	error = xagb_bitmap_set_btblocks(&cr->inobt_owned, cur);
 	if (cur != sc->sa.ino_cur)
 		xfs_btree_del_cursor(cur, error);
@@ -458,8 +466,8 @@ xchk_rmapbt_walk_ag_metadata(
 	if (xfs_has_finobt(sc->mp)) {
 		cur = sc->sa.fino_cur;
 		if (!cur)
-			cur = xfs_inobt_init_cursor(sc->sa.pag, sc->tp,
-					sc->sa.agi_bp, XFS_BTNUM_FINO);
+			cur = xfs_finobt_init_cursor(sc->sa.pag, sc->tp,
+					sc->sa.agi_bp);
 		error = xagb_bitmap_set_btblocks(&cr->inobt_owned, cur);
 		if (cur != sc->sa.fino_cur)
 			xfs_btree_del_cursor(cur, error);
diff --git a/fs/xfs/scrub/rmap_repair.c b/fs/xfs/scrub/rmap_repair.c
new file mode 100644
index 0000000000..e8080eba37
--- /dev/null
+++ b/fs/xfs/scrub/rmap_repair.c
@@ -0,0 +1,1675 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2018-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_btree_staging.h"
+#include "xfs_buf_mem.h"
+#include "xfs_btree_mem.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_alloc.h"
+#include "xfs_alloc_btree.h"
+#include "xfs_ialloc.h"
+#include "xfs_ialloc_btree.h"
+#include "xfs_rmap.h"
+#include "xfs_rmap_btree.h"
+#include "xfs_inode.h"
+#include "xfs_icache.h"
+#include "xfs_bmap.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_refcount.h"
+#include "xfs_refcount_btree.h"
+#include "xfs_ag.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/btree.h"
+#include "scrub/trace.h"
+#include "scrub/repair.h"
+#include "scrub/bitmap.h"
+#include "scrub/agb_bitmap.h"
+#include "scrub/xfile.h"
+#include "scrub/xfarray.h"
+#include "scrub/iscan.h"
+#include "scrub/newbt.h"
+#include "scrub/reap.h"
+
+/*
+ * Reverse Mapping Btree Repair
+ * ============================
+ *
+ * This is the most involved of all the AG space btree rebuilds.  Everywhere
+ * else in XFS we lock inodes and then AG data structures, but generating the
+ * list of rmap records requires that we be able to scan both block mapping
+ * btrees of every inode in the filesystem to see if it owns any extents in
+ * this AG.  We can't tolerate any inode updates while we do this, so we
+ * freeze the filesystem to lock everyone else out, and grant ourselves
+ * special privileges to run transactions with regular background reclamation
+ * turned off.
+ *
+ * We also have to be very careful not to allow inode reclaim to start a
+ * transaction because all transactions (other than our own) will block.
+ * Deferred inode inactivation helps us out there.
+ *
+ * I) Reverse mappings for all non-space metadata and file data are collected
+ * according to the following algorithm:
+ *
+ * 1. For each fork of each inode:
+ * 1.1. Create a bitmap BMBIT to track bmbt blocks if necessary.
+ * 1.2. If the incore extent map isn't loaded, walk the bmbt to accumulate
+ *      bmaps into rmap records (see 1.1.4).  Set bits in BMBIT for each btree
+ *      block.
+ * 1.3. If the incore extent map is loaded but the fork is in btree format,
+ *      just visit the bmbt blocks to set the corresponding BMBIT areas.
+ * 1.4. From the incore extent map, accumulate each bmap that falls into our
+ *      target AG.  Remember, multiple bmap records can map to a single rmap
+ *      record, so we cannot simply emit rmap records 1:1.
+ * 1.5. Emit rmap records for each extent in BMBIT and free it.
+ * 2. Create bitmaps INOBIT and ICHUNKBIT.
+ * 3. For each record in the inobt, set the corresponding areas in ICHUNKBIT,
+ *    and set bits in INOBIT for each btree block.  If the inobt has no records
+ *    at all, we must be careful to record its root in INOBIT.
+ * 4. For each block in the finobt, set the corresponding INOBIT area.
+ * 5. Emit rmap records for each extent in INOBIT and ICHUNKBIT and free them.
+ * 6. Create bitmaps REFCBIT and COWBIT.
+ * 7. For each CoW staging extent in the refcountbt, set the corresponding
+ *    areas in COWBIT.
+ * 8. For each block in the refcountbt, set the corresponding REFCBIT area.
+ * 9. Emit rmap records for each extent in REFCBIT and COWBIT and free them.
+ * A. Emit rmap for the AG headers.
+ * B. Emit rmap for the log, if there is one.
+ *
+ * II) The rmapbt shape and space metadata rmaps are computed as follows:
+ *
+ * 1. Count the rmaps collected in the previous step. (= NR)
+ * 2. Estimate the number of rmapbt blocks needed to store NR records. (= RMB)
+ * 3. Reserve RMB blocks through the newbt using the allocator in normap mode.
+ * 4. Create bitmap AGBIT.
+ * 5. For each reservation in the newbt, set the corresponding areas in AGBIT.
+ * 6. For each block in the AGFL, bnobt, and cntbt, set the bits in AGBIT.
+ * 7. Count the extents in AGBIT. (= AGNR)
+ * 8. Estimate the number of rmapbt blocks needed for NR + AGNR rmaps. (= RMB')
+ * 9. If RMB' >= RMB, reserve RMB' - RMB more newbt blocks, set RMB = RMB',
+ *    and clear AGBIT.  Go to step 5.
+ * A. Emit rmaps for each extent in AGBIT.
+ *
+ * III) The rmapbt is constructed and set in place as follows:
+ *
+ * 1. Sort the rmap records.
+ * 2. Bulk load the rmaps.
+ *
+ * IV) Reap the old btree blocks.
+ *
+ * 1. Create a bitmap OLDRMBIT.
+ * 2. For each gap in the new rmapbt, set the corresponding areas of OLDRMBIT.
+ * 3. For each extent in the bnobt, clear the corresponding parts of OLDRMBIT.
+ * 4. Reap the extents corresponding to the set areas in OLDRMBIT.  These are
+ *    the parts of the AG that the rmap didn't find during its scan of the
+ *    primary metadata and aren't known to be in the free space, which implies
+ *    that they were the old rmapbt blocks.
+ * 5. Commit.
+ *
+ * We use the 'xrep_rmap' prefix for all the rmap functions.
+ */
+
+/* Context for collecting rmaps */
+struct xrep_rmap {
+	/* new rmapbt information */
+	struct xrep_newbt	new_btree;
+
+	/* lock for the xfbtree and xfile */
+	struct mutex		lock;
+
+	/* rmap records generated from primary metadata */
+	struct xfbtree		rmap_btree;
+
+	struct xfs_scrub	*sc;
+
+	/* in-memory btree cursor for the xfs_btree_bload iteration */
+	struct xfs_btree_cur	*mcur;
+
+	/* Hooks into rmap update code. */
+	struct xfs_rmap_hook	rhook;
+
+	/* inode scan cursor */
+	struct xchk_iscan	iscan;
+
+	/* Number of non-freespace records found. */
+	unsigned long long	nr_records;
+
+	/* bnobt/cntbt contribution to btreeblks */
+	xfs_agblock_t		freesp_btblocks;
+
+	/* old agf_rmap_blocks counter */
+	unsigned int		old_rmapbt_fsbcount;
+};
+
+/* Set us up to repair reverse mapping btrees. */
+int
+xrep_setup_ag_rmapbt(
+	struct xfs_scrub	*sc)
+{
+	struct xrep_rmap	*rr;
+	char			*descr;
+	int			error;
+
+	xchk_fsgates_enable(sc, XCHK_FSGATES_RMAP);
+
+	descr = xchk_xfile_ag_descr(sc, "reverse mapping records");
+	error = xrep_setup_xfbtree(sc, descr);
+	kfree(descr);
+	if (error)
+		return error;
+
+	rr = kzalloc(sizeof(struct xrep_rmap), XCHK_GFP_FLAGS);
+	if (!rr)
+		return -ENOMEM;
+
+	rr->sc = sc;
+	sc->buf = rr;
+	return 0;
+}
+
+/* Make sure there's nothing funny about this mapping. */
+STATIC int
+xrep_rmap_check_mapping(
+	struct xfs_scrub	*sc,
+	const struct xfs_rmap_irec *rec)
+{
+	enum xbtree_recpacking	outcome;
+	int			error;
+
+	if (xfs_rmap_check_irec(sc->sa.pag, rec) != NULL)
+		return -EFSCORRUPTED;
+
+	/* Make sure this isn't free space. */
+	error = xfs_alloc_has_records(sc->sa.bno_cur, rec->rm_startblock,
+			rec->rm_blockcount, &outcome);
+	if (error)
+		return error;
+	if (outcome != XBTREE_RECPACKING_EMPTY)
+		return -EFSCORRUPTED;
+
+	return 0;
+}
+
+/* Store a reverse-mapping record. */
+static inline int
+xrep_rmap_stash(
+	struct xrep_rmap	*rr,
+	xfs_agblock_t		startblock,
+	xfs_extlen_t		blockcount,
+	uint64_t		owner,
+	uint64_t		offset,
+	unsigned int		flags)
+{
+	struct xfs_rmap_irec	rmap = {
+		.rm_startblock	= startblock,
+		.rm_blockcount	= blockcount,
+		.rm_owner	= owner,
+		.rm_offset	= offset,
+		.rm_flags	= flags,
+	};
+	struct xfs_scrub	*sc = rr->sc;
+	struct xfs_btree_cur	*mcur;
+	int			error = 0;
+
+	if (xchk_should_terminate(sc, &error))
+		return error;
+
+	if (xchk_iscan_aborted(&rr->iscan))
+		return -EFSCORRUPTED;
+
+	trace_xrep_rmap_found(sc->mp, sc->sa.pag->pag_agno, &rmap);
+
+	mutex_lock(&rr->lock);
+	mcur = xfs_rmapbt_mem_cursor(sc->sa.pag, sc->tp, &rr->rmap_btree);
+	error = xfs_rmap_map_raw(mcur, &rmap);
+	xfs_btree_del_cursor(mcur, error);
+	if (error)
+		goto out_cancel;
+
+	error = xfbtree_trans_commit(&rr->rmap_btree, sc->tp);
+	if (error)
+		goto out_abort;
+
+	mutex_unlock(&rr->lock);
+	return 0;
+
+out_cancel:
+	xfbtree_trans_cancel(&rr->rmap_btree, sc->tp);
+out_abort:
+	xchk_iscan_abort(&rr->iscan);
+	mutex_unlock(&rr->lock);
+	return error;
+}
+
+struct xrep_rmap_stash_run {
+	struct xrep_rmap	*rr;
+	uint64_t		owner;
+	unsigned int		rmap_flags;
+};
+
+static int
+xrep_rmap_stash_run(
+	uint32_t			start,
+	uint32_t			len,
+	void				*priv)
+{
+	struct xrep_rmap_stash_run	*rsr = priv;
+	struct xrep_rmap		*rr = rsr->rr;
+
+	return xrep_rmap_stash(rr, start, len, rsr->owner, 0, rsr->rmap_flags);
+}
+
+/*
+ * Emit rmaps for every extent of bits set in the bitmap.  Caller must ensure
+ * that the ranges are in units of FS blocks.
+ */
+STATIC int
+xrep_rmap_stash_bitmap(
+	struct xrep_rmap		*rr,
+	struct xagb_bitmap		*bitmap,
+	const struct xfs_owner_info	*oinfo)
+{
+	struct xrep_rmap_stash_run	rsr = {
+		.rr			= rr,
+		.owner			= oinfo->oi_owner,
+		.rmap_flags		= 0,
+	};
+
+	if (oinfo->oi_flags & XFS_OWNER_INFO_ATTR_FORK)
+		rsr.rmap_flags |= XFS_RMAP_ATTR_FORK;
+	if (oinfo->oi_flags & XFS_OWNER_INFO_BMBT_BLOCK)
+		rsr.rmap_flags |= XFS_RMAP_BMBT_BLOCK;
+
+	return xagb_bitmap_walk(bitmap, xrep_rmap_stash_run, &rsr);
+}
+
+/* Section (I): Finding all file and bmbt extents. */
+
+/* Context for accumulating rmaps for an inode fork. */
+struct xrep_rmap_ifork {
+	/*
+	 * Accumulate rmap data here to turn multiple adjacent bmaps into a
+	 * single rmap.
+	 */
+	struct xfs_rmap_irec	accum;
+
+	/* Bitmap of bmbt blocks in this AG. */
+	struct xagb_bitmap	bmbt_blocks;
+
+	struct xrep_rmap	*rr;
+
+	/* Which inode fork? */
+	int			whichfork;
+};
+
+/* Stash an rmap that we accumulated while walking an inode fork. */
+STATIC int
+xrep_rmap_stash_accumulated(
+	struct xrep_rmap_ifork	*rf)
+{
+	if (rf->accum.rm_blockcount == 0)
+		return 0;
+
+	return xrep_rmap_stash(rf->rr, rf->accum.rm_startblock,
+			rf->accum.rm_blockcount, rf->accum.rm_owner,
+			rf->accum.rm_offset, rf->accum.rm_flags);
+}
+
+/* Accumulate a bmbt record. */
+STATIC int
+xrep_rmap_visit_bmbt(
+	struct xfs_btree_cur	*cur,
+	struct xfs_bmbt_irec	*rec,
+	void			*priv)
+{
+	struct xrep_rmap_ifork	*rf = priv;
+	struct xfs_mount	*mp = rf->rr->sc->mp;
+	struct xfs_rmap_irec	*accum = &rf->accum;
+	xfs_agblock_t		agbno;
+	unsigned int		rmap_flags = 0;
+	int			error;
+
+	if (XFS_FSB_TO_AGNO(mp, rec->br_startblock) !=
+			rf->rr->sc->sa.pag->pag_agno)
+		return 0;
+
+	agbno = XFS_FSB_TO_AGBNO(mp, rec->br_startblock);
+	if (rf->whichfork == XFS_ATTR_FORK)
+		rmap_flags |= XFS_RMAP_ATTR_FORK;
+	if (rec->br_state == XFS_EXT_UNWRITTEN)
+		rmap_flags |= XFS_RMAP_UNWRITTEN;
+
+	/* If this bmap is adjacent to the previous one, just add it. */
+	if (accum->rm_blockcount > 0 &&
+	    rec->br_startoff == accum->rm_offset + accum->rm_blockcount &&
+	    agbno == accum->rm_startblock + accum->rm_blockcount &&
+	    rmap_flags == accum->rm_flags) {
+		accum->rm_blockcount += rec->br_blockcount;
+		return 0;
+	}
+
+	/* Otherwise stash the old rmap and start accumulating a new one. */
+	error = xrep_rmap_stash_accumulated(rf);
+	if (error)
+		return error;
+
+	accum->rm_startblock = agbno;
+	accum->rm_blockcount = rec->br_blockcount;
+	accum->rm_offset = rec->br_startoff;
+	accum->rm_flags = rmap_flags;
+	return 0;
+}
+
+/* Add a btree block to the bitmap. */
+STATIC int
+xrep_rmap_visit_iroot_btree_block(
+	struct xfs_btree_cur	*cur,
+	int			level,
+	void			*priv)
+{
+	struct xrep_rmap_ifork	*rf = priv;
+	struct xfs_buf		*bp;
+	xfs_fsblock_t		fsbno;
+	xfs_agblock_t		agbno;
+
+	xfs_btree_get_block(cur, level, &bp);
+	if (!bp)
+		return 0;
+
+	fsbno = XFS_DADDR_TO_FSB(cur->bc_mp, xfs_buf_daddr(bp));
+	if (XFS_FSB_TO_AGNO(cur->bc_mp, fsbno) != rf->rr->sc->sa.pag->pag_agno)
+		return 0;
+
+	agbno = XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno);
+	return xagb_bitmap_set(&rf->bmbt_blocks, agbno, 1);
+}
+
+/*
+ * Iterate a metadata btree rooted in an inode to collect rmap records for
+ * anything in this fork that matches the AG.
+ */
+STATIC int
+xrep_rmap_scan_iroot_btree(
+	struct xrep_rmap_ifork	*rf,
+	struct xfs_btree_cur	*cur)
+{
+	struct xfs_owner_info	oinfo;
+	struct xrep_rmap	*rr = rf->rr;
+	int			error;
+
+	xagb_bitmap_init(&rf->bmbt_blocks);
+
+	/* Record all the blocks in the btree itself. */
+	error = xfs_btree_visit_blocks(cur, xrep_rmap_visit_iroot_btree_block,
+			XFS_BTREE_VISIT_ALL, rf);
+	if (error)
+		goto out;
+
+	/* Emit rmaps for the btree blocks. */
+	xfs_rmap_ino_bmbt_owner(&oinfo, rf->accum.rm_owner, rf->whichfork);
+	error = xrep_rmap_stash_bitmap(rr, &rf->bmbt_blocks, &oinfo);
+	if (error)
+		goto out;
+
+	/* Stash any remaining accumulated rmaps. */
+	error = xrep_rmap_stash_accumulated(rf);
+out:
+	xagb_bitmap_destroy(&rf->bmbt_blocks);
+	return error;
+}
+
+/*
+ * Iterate the block mapping btree to collect rmap records for anything in this
+ * fork that matches the AG.  Sets @mappings_done to true if we've scanned the
+ * block mappings in this fork.
+ */
+STATIC int
+xrep_rmap_scan_bmbt(
+	struct xrep_rmap_ifork	*rf,
+	struct xfs_inode	*ip,
+	bool			*mappings_done)
+{
+	struct xrep_rmap	*rr = rf->rr;
+	struct xfs_btree_cur	*cur;
+	struct xfs_ifork	*ifp;
+	int			error;
+
+	*mappings_done = false;
+	ifp = xfs_ifork_ptr(ip, rf->whichfork);
+	cur = xfs_bmbt_init_cursor(rr->sc->mp, rr->sc->tp, ip, rf->whichfork);
+
+	if (!xfs_ifork_is_realtime(ip, rf->whichfork) &&
+	    xfs_need_iread_extents(ifp)) {
+		/*
+		 * If the incore extent cache isn't loaded, scan the bmbt for
+		 * mapping records.  This avoids loading the incore extent
+		 * tree, which will increase memory pressure at a time when
+		 * we're trying to run as quickly as we possibly can.  Ignore
+		 * realtime extents.
+		 */
+		error = xfs_bmap_query_all(cur, xrep_rmap_visit_bmbt, rf);
+		if (error)
+			goto out_cur;
+
+		*mappings_done = true;
+	}
+
+	/* Scan for the bmbt blocks, which always live on the data device. */
+	error = xrep_rmap_scan_iroot_btree(rf, cur);
+out_cur:
+	xfs_btree_del_cursor(cur, error);
+	return error;
+}
+
+/*
+ * Iterate the in-core extent cache to collect rmap records for anything in
+ * this fork that matches the AG.
+ */
+STATIC int
+xrep_rmap_scan_iext(
+	struct xrep_rmap_ifork	*rf,
+	struct xfs_ifork	*ifp)
+{
+	struct xfs_bmbt_irec	rec;
+	struct xfs_iext_cursor	icur;
+	int			error;
+
+	for_each_xfs_iext(ifp, &icur, &rec) {
+		if (isnullstartblock(rec.br_startblock))
+			continue;
+		error = xrep_rmap_visit_bmbt(NULL, &rec, rf);
+		if (error)
+			return error;
+	}
+
+	return xrep_rmap_stash_accumulated(rf);
+}
+
+/* Find all the extents from a given AG in an inode fork. */
+STATIC int
+xrep_rmap_scan_ifork(
+	struct xrep_rmap	*rr,
+	struct xfs_inode	*ip,
+	int			whichfork)
+{
+	struct xrep_rmap_ifork	rf = {
+		.accum		= { .rm_owner = ip->i_ino, },
+		.rr		= rr,
+		.whichfork	= whichfork,
+	};
+	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, whichfork);
+	int			error = 0;
+
+	if (!ifp)
+		return 0;
+
+	if (ifp->if_format == XFS_DINODE_FMT_BTREE) {
+		bool		mappings_done;
+
+		/*
+		 * Scan the bmap btree for data device mappings.  This includes
+		 * the btree blocks themselves, even if this is a realtime
+		 * file.
+		 */
+		error = xrep_rmap_scan_bmbt(&rf, ip, &mappings_done);
+		if (error || mappings_done)
+			return error;
+	} else if (ifp->if_format != XFS_DINODE_FMT_EXTENTS) {
+		return 0;
+	}
+
+	/* Scan incore extent cache if this isn't a realtime file. */
+	if (xfs_ifork_is_realtime(ip, whichfork))
+		return 0;
+
+	return xrep_rmap_scan_iext(&rf, ifp);
+}
+
+/*
+ * Take ILOCK on a file that we want to scan.
+ *
+ * Select ILOCK_EXCL if the file has an unloaded data bmbt or has an unloaded
+ * attr bmbt.  Otherwise, take ILOCK_SHARED.
+ */
+static inline unsigned int
+xrep_rmap_scan_ilock(
+	struct xfs_inode	*ip)
+{
+	uint			lock_mode = XFS_ILOCK_SHARED;
+
+	if (xfs_need_iread_extents(&ip->i_df)) {
+		lock_mode = XFS_ILOCK_EXCL;
+		goto lock;
+	}
+
+	if (xfs_inode_has_attr_fork(ip) && xfs_need_iread_extents(&ip->i_af))
+		lock_mode = XFS_ILOCK_EXCL;
+
+lock:
+	xfs_ilock(ip, lock_mode);
+	return lock_mode;
+}
+
+/* Record reverse mappings for a file. */
+STATIC int
+xrep_rmap_scan_inode(
+	struct xrep_rmap	*rr,
+	struct xfs_inode	*ip)
+{
+	unsigned int		lock_mode = xrep_rmap_scan_ilock(ip);
+	int			error;
+
+	/* Check the data fork. */
+	error = xrep_rmap_scan_ifork(rr, ip, XFS_DATA_FORK);
+	if (error)
+		goto out_unlock;
+
+	/* Check the attr fork. */
+	error = xrep_rmap_scan_ifork(rr, ip, XFS_ATTR_FORK);
+	if (error)
+		goto out_unlock;
+
+	/* COW fork extents are "owned" by the refcount btree. */
+
+	xchk_iscan_mark_visited(&rr->iscan, ip);
+out_unlock:
+	xfs_iunlock(ip, lock_mode);
+	return error;
+}
+
+/* Section (I): Find all AG metadata extents except for free space metadata. */
+
+struct xrep_rmap_inodes {
+	struct xrep_rmap	*rr;
+	struct xagb_bitmap	inobt_blocks;	/* INOBIT */
+	struct xagb_bitmap	ichunk_blocks;	/* ICHUNKBIT */
+};
+
+/* Record inode btree rmaps. */
+STATIC int
+xrep_rmap_walk_inobt(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_rec	*rec,
+	void				*priv)
+{
+	struct xfs_inobt_rec_incore	irec;
+	struct xrep_rmap_inodes		*ri = priv;
+	struct xfs_mount		*mp = cur->bc_mp;
+	xfs_agblock_t			agbno;
+	xfs_extlen_t			aglen;
+	xfs_agino_t			agino;
+	xfs_agino_t			iperhole;
+	unsigned int			i;
+	int				error;
+
+	/* Record the inobt blocks. */
+	error = xagb_bitmap_set_btcur_path(&ri->inobt_blocks, cur);
+	if (error)
+		return error;
+
+	xfs_inobt_btrec_to_irec(mp, rec, &irec);
+	if (xfs_inobt_check_irec(cur->bc_ag.pag, &irec) != NULL)
+		return -EFSCORRUPTED;
+
+	agino = irec.ir_startino;
+
+	/* Record a non-sparse inode chunk. */
+	if (!xfs_inobt_issparse(irec.ir_holemask)) {
+		agbno = XFS_AGINO_TO_AGBNO(mp, agino);
+		aglen = max_t(xfs_extlen_t, 1,
+				XFS_INODES_PER_CHUNK / mp->m_sb.sb_inopblock);
+
+		return xagb_bitmap_set(&ri->ichunk_blocks, agbno, aglen);
+	}
+
+	/* Iterate each chunk. */
+	iperhole = max_t(xfs_agino_t, mp->m_sb.sb_inopblock,
+			XFS_INODES_PER_HOLEMASK_BIT);
+	aglen = iperhole / mp->m_sb.sb_inopblock;
+	for (i = 0, agino = irec.ir_startino;
+	     i < XFS_INOBT_HOLEMASK_BITS;
+	     i += iperhole / XFS_INODES_PER_HOLEMASK_BIT, agino += iperhole) {
+		/* Skip holes. */
+		if (irec.ir_holemask & (1 << i))
+			continue;
+
+		/* Record the inode chunk otherwise. */
+		agbno = XFS_AGINO_TO_AGBNO(mp, agino);
+		error = xagb_bitmap_set(&ri->ichunk_blocks, agbno, aglen);
+		if (error)
+			return error;
+	}
+
+	return 0;
+}
+
+/* Collect rmaps for the blocks containing inode btrees and the inode chunks. */
+STATIC int
+xrep_rmap_find_inode_rmaps(
+	struct xrep_rmap	*rr)
+{
+	struct xrep_rmap_inodes	ri = {
+		.rr		= rr,
+	};
+	struct xfs_scrub	*sc = rr->sc;
+	int			error;
+
+	xagb_bitmap_init(&ri.inobt_blocks);
+	xagb_bitmap_init(&ri.ichunk_blocks);
+
+	/*
+	 * Iterate every record in the inobt so we can capture all the inode
+	 * chunks and the blocks in the inobt itself.
+	 */
+	error = xfs_btree_query_all(sc->sa.ino_cur, xrep_rmap_walk_inobt, &ri);
+	if (error)
+		goto out_bitmap;
+
+	/*
+	 * Note that if there are zero records in the inobt then query_all does
+	 * nothing and we have to account the empty inobt root manually.
+	 */
+	if (xagb_bitmap_empty(&ri.ichunk_blocks)) {
+		struct xfs_agi	*agi = sc->sa.agi_bp->b_addr;
+
+		error = xagb_bitmap_set(&ri.inobt_blocks,
+				be32_to_cpu(agi->agi_root), 1);
+		if (error)
+			goto out_bitmap;
+	}
+
+	/* Scan the finobt too. */
+	if (xfs_has_finobt(sc->mp)) {
+		error = xagb_bitmap_set_btblocks(&ri.inobt_blocks,
+				sc->sa.fino_cur);
+		if (error)
+			goto out_bitmap;
+	}
+
+	/* Generate rmaps for everything. */
+	error = xrep_rmap_stash_bitmap(rr, &ri.inobt_blocks,
+			&XFS_RMAP_OINFO_INOBT);
+	if (error)
+		goto out_bitmap;
+	error = xrep_rmap_stash_bitmap(rr, &ri.ichunk_blocks,
+			&XFS_RMAP_OINFO_INODES);
+
+out_bitmap:
+	xagb_bitmap_destroy(&ri.inobt_blocks);
+	xagb_bitmap_destroy(&ri.ichunk_blocks);
+	return error;
+}
+
+/* Record a CoW staging extent. */
+STATIC int
+xrep_rmap_walk_cowblocks(
+	struct xfs_btree_cur		*cur,
+	const struct xfs_refcount_irec	*irec,
+	void				*priv)
+{
+	struct xagb_bitmap		*bitmap = priv;
+
+	if (!xfs_refcount_check_domain(irec) ||
+	    irec->rc_domain != XFS_REFC_DOMAIN_COW)
+		return -EFSCORRUPTED;
+
+	return xagb_bitmap_set(bitmap, irec->rc_startblock, irec->rc_blockcount);
+}
+
+/*
+ * Collect rmaps for the blocks containing the refcount btree, and all CoW
+ * staging extents.
+ */
+STATIC int
+xrep_rmap_find_refcount_rmaps(
+	struct xrep_rmap	*rr)
+{
+	struct xagb_bitmap	refcountbt_blocks;	/* REFCBIT */
+	struct xagb_bitmap	cow_blocks;		/* COWBIT */
+	struct xfs_refcount_irec low = {
+		.rc_startblock	= 0,
+		.rc_domain	= XFS_REFC_DOMAIN_COW,
+	};
+	struct xfs_refcount_irec high = {
+		.rc_startblock	= -1U,
+		.rc_domain	= XFS_REFC_DOMAIN_COW,
+	};
+	struct xfs_scrub	*sc = rr->sc;
+	int			error;
+
+	if (!xfs_has_reflink(sc->mp))
+		return 0;
+
+	xagb_bitmap_init(&refcountbt_blocks);
+	xagb_bitmap_init(&cow_blocks);
+
+	/* refcountbt */
+	error = xagb_bitmap_set_btblocks(&refcountbt_blocks, sc->sa.refc_cur);
+	if (error)
+		goto out_bitmap;
+
+	/* Collect rmaps for CoW staging extents. */
+	error = xfs_refcount_query_range(sc->sa.refc_cur, &low, &high,
+			xrep_rmap_walk_cowblocks, &cow_blocks);
+	if (error)
+		goto out_bitmap;
+
+	/* Generate rmaps for everything. */
+	error = xrep_rmap_stash_bitmap(rr, &cow_blocks, &XFS_RMAP_OINFO_COW);
+	if (error)
+		goto out_bitmap;
+	error = xrep_rmap_stash_bitmap(rr, &refcountbt_blocks,
+			&XFS_RMAP_OINFO_REFC);
+
+out_bitmap:
+	xagb_bitmap_destroy(&cow_blocks);
+	xagb_bitmap_destroy(&refcountbt_blocks);
+	return error;
+}
+
+/* Generate rmaps for the AG headers (AGI/AGF/AGFL) */
+STATIC int
+xrep_rmap_find_agheader_rmaps(
+	struct xrep_rmap	*rr)
+{
+	struct xfs_scrub	*sc = rr->sc;
+
+	/* Create a record for the AG sb->agfl. */
+	return xrep_rmap_stash(rr, XFS_SB_BLOCK(sc->mp),
+			XFS_AGFL_BLOCK(sc->mp) - XFS_SB_BLOCK(sc->mp) + 1,
+			XFS_RMAP_OWN_FS, 0, 0);
+}
+
+/* Generate rmaps for the log, if it's in this AG. */
+STATIC int
+xrep_rmap_find_log_rmaps(
+	struct xrep_rmap	*rr)
+{
+	struct xfs_scrub	*sc = rr->sc;
+
+	if (!xfs_ag_contains_log(sc->mp, sc->sa.pag->pag_agno))
+		return 0;
+
+	return xrep_rmap_stash(rr,
+			XFS_FSB_TO_AGBNO(sc->mp, sc->mp->m_sb.sb_logstart),
+			sc->mp->m_sb.sb_logblocks, XFS_RMAP_OWN_LOG, 0, 0);
+}
+
+/* Check and count all the records that we gathered. */
+STATIC int
+xrep_rmap_check_record(
+	struct xfs_btree_cur		*cur,
+	const struct xfs_rmap_irec	*rec,
+	void				*priv)
+{
+	struct xrep_rmap		*rr = priv;
+	int				error;
+
+	error = xrep_rmap_check_mapping(rr->sc, rec);
+	if (error)
+		return error;
+
+	rr->nr_records++;
+	return 0;
+}
+
+/*
+ * Generate all the reverse-mappings for this AG, a list of the old rmapbt
+ * blocks, and the new btreeblks count.  Figure out if we have enough free
+ * space to reconstruct the inode btrees.  The caller must clean up the lists
+ * if anything goes wrong.  This implements section (I) above.
+ */
+STATIC int
+xrep_rmap_find_rmaps(
+	struct xrep_rmap	*rr)
+{
+	struct xfs_scrub	*sc = rr->sc;
+	struct xchk_ag		*sa = &sc->sa;
+	struct xfs_inode	*ip;
+	struct xfs_btree_cur	*mcur;
+	int			error;
+
+	/* Find all the per-AG metadata. */
+	xrep_ag_btcur_init(sc, &sc->sa);
+
+	error = xrep_rmap_find_inode_rmaps(rr);
+	if (error)
+		goto end_agscan;
+
+	error = xrep_rmap_find_refcount_rmaps(rr);
+	if (error)
+		goto end_agscan;
+
+	error = xrep_rmap_find_agheader_rmaps(rr);
+	if (error)
+		goto end_agscan;
+
+	error = xrep_rmap_find_log_rmaps(rr);
+end_agscan:
+	xchk_ag_btcur_free(&sc->sa);
+	if (error)
+		return error;
+
+	/*
+	 * Set up for a potentially lengthy filesystem scan by reducing our
+	 * transaction resource usage for the duration.  Specifically:
+	 *
+	 * Unlock the AG header buffers and cancel the transaction to release
+	 * the log grant space while we scan the filesystem.
+	 *
+	 * Create a new empty transaction to eliminate the possibility of the
+	 * inode scan deadlocking on cyclical metadata.
+	 *
+	 * We pass the empty transaction to the file scanning function to avoid
+	 * repeatedly cycling empty transactions.  This can be done even though
+	 * we take the IOLOCK to quiesce the file because empty transactions
+	 * do not take sb_internal.
+	 */
+	sa->agf_bp = NULL;
+	sa->agi_bp = NULL;
+	xchk_trans_cancel(sc);
+	error = xchk_trans_alloc_empty(sc);
+	if (error)
+		return error;
+
+	/* Iterate all AGs for inodes rmaps. */
+	while ((error = xchk_iscan_iter(&rr->iscan, &ip)) == 1) {
+		error = xrep_rmap_scan_inode(rr, ip);
+		xchk_irele(sc, ip);
+		if (error)
+			break;
+
+		if (xchk_should_terminate(sc, &error))
+			break;
+	}
+	xchk_iscan_iter_finish(&rr->iscan);
+	if (error)
+		return error;
+
+	/*
+	 * Switch out for a real transaction and lock the AG headers in
+	 * preparation for building a new tree.
+	 */
+	xchk_trans_cancel(sc);
+	error = xchk_setup_fs(sc);
+	if (error)
+		return error;
+	error = xchk_perag_drain_and_lock(sc);
+	if (error)
+		return error;
+
+	/*
+	 * If a hook failed to update the in-memory btree, we lack the data to
+	 * continue the repair.
+	 */
+	if (xchk_iscan_aborted(&rr->iscan))
+		return -EFSCORRUPTED;
+
+	/*
+	 * Now that we have everything locked again, we need to count the
+	 * number of rmap records stashed in the btree.  This should reflect
+	 * all actively-owned space in the filesystem.  At the same time, check
+	 * all our records before we start building a new btree, which requires
+	 * a bnobt cursor.
+	 */
+	mcur = xfs_rmapbt_mem_cursor(rr->sc->sa.pag, NULL, &rr->rmap_btree);
+	sc->sa.bno_cur = xfs_bnobt_init_cursor(sc->mp, sc->tp, sc->sa.agf_bp,
+			sc->sa.pag);
+
+	rr->nr_records = 0;
+	error = xfs_rmap_query_all(mcur, xrep_rmap_check_record, rr);
+
+	xfs_btree_del_cursor(sc->sa.bno_cur, error);
+	sc->sa.bno_cur = NULL;
+	xfs_btree_del_cursor(mcur, error);
+
+	return error;
+}
+
+/* Section (II): Reserving space for new rmapbt and setting free space bitmap */
+
+struct xrep_rmap_agfl {
+	struct xagb_bitmap	*bitmap;
+	xfs_agnumber_t		agno;
+};
+
+/* Add an AGFL block to the rmap list. */
+STATIC int
+xrep_rmap_walk_agfl(
+	struct xfs_mount	*mp,
+	xfs_agblock_t		agbno,
+	void			*priv)
+{
+	struct xrep_rmap_agfl	*ra = priv;
+
+	return xagb_bitmap_set(ra->bitmap, agbno, 1);
+}
+
+/*
+ * Run one round of reserving space for the new rmapbt and recomputing the
+ * number of blocks needed to store the previously observed rmapbt records and
+ * the ones we'll create for the free space metadata.  When we don't need more
+ * blocks, return a bitmap of OWN_AG extents in @freesp_blocks and set @done to
+ * true.
+ */
+STATIC int
+xrep_rmap_try_reserve(
+	struct xrep_rmap	*rr,
+	struct xfs_btree_cur	*rmap_cur,
+	struct xagb_bitmap	*freesp_blocks,
+	uint64_t		*blocks_reserved,
+	bool			*done)
+{
+	struct xrep_rmap_agfl	ra = {
+		.bitmap		= freesp_blocks,
+		.agno		= rr->sc->sa.pag->pag_agno,
+	};
+	struct xfs_scrub	*sc = rr->sc;
+	struct xrep_newbt_resv	*resv, *n;
+	struct xfs_agf		*agf = sc->sa.agf_bp->b_addr;
+	struct xfs_buf		*agfl_bp;
+	uint64_t		nr_blocks;	/* RMB */
+	uint64_t		freesp_records;
+	int			error;
+
+	/*
+	 * We're going to recompute new_btree.bload.nr_blocks at the end of
+	 * this function to reflect however many btree blocks we need to store
+	 * all the rmap records (including the ones that reflect the changes we
+	 * made to support the new rmapbt blocks), so we save the old value
+	 * here so we can decide if we've reserved enough blocks.
+	 */
+	nr_blocks = rr->new_btree.bload.nr_blocks;
+
+	/*
+	 * Make sure we've reserved enough space for the new btree.  This can
+	 * change the shape of the free space btrees, which can cause secondary
+	 * interactions with the rmap records because all three space btrees
+	 * have the same rmap owner.  We'll account for all that below.
+	 */
+	error = xrep_newbt_alloc_blocks(&rr->new_btree,
+			nr_blocks - *blocks_reserved);
+	if (error)
+		return error;
+
+	*blocks_reserved = rr->new_btree.bload.nr_blocks;
+
+	/* Clear everything in the bitmap. */
+	xagb_bitmap_destroy(freesp_blocks);
+
+	/* Set all the bnobt blocks in the bitmap. */
+	sc->sa.bno_cur = xfs_bnobt_init_cursor(sc->mp, sc->tp, sc->sa.agf_bp,
+			sc->sa.pag);
+	error = xagb_bitmap_set_btblocks(freesp_blocks, sc->sa.bno_cur);
+	xfs_btree_del_cursor(sc->sa.bno_cur, error);
+	sc->sa.bno_cur = NULL;
+	if (error)
+		return error;
+
+	/* Set all the cntbt blocks in the bitmap. */
+	sc->sa.cnt_cur = xfs_cntbt_init_cursor(sc->mp, sc->tp, sc->sa.agf_bp,
+			sc->sa.pag);
+	error = xagb_bitmap_set_btblocks(freesp_blocks, sc->sa.cnt_cur);
+	xfs_btree_del_cursor(sc->sa.cnt_cur, error);
+	sc->sa.cnt_cur = NULL;
+	if (error)
+		return error;
+
+	/* Record our new btreeblks value. */
+	rr->freesp_btblocks = xagb_bitmap_hweight(freesp_blocks) - 2;
+
+	/* Set all the new rmapbt blocks in the bitmap. */
+	list_for_each_entry_safe(resv, n, &rr->new_btree.resv_list, list) {
+		error = xagb_bitmap_set(freesp_blocks, resv->agbno, resv->len);
+		if (error)
+			return error;
+	}
+
+	/* Set all the AGFL blocks in the bitmap. */
+	error = xfs_alloc_read_agfl(sc->sa.pag, sc->tp, &agfl_bp);
+	if (error)
+		return error;
+
+	error = xfs_agfl_walk(sc->mp, agf, agfl_bp, xrep_rmap_walk_agfl, &ra);
+	if (error)
+		return error;
+
+	/* Count the extents in the bitmap. */
+	freesp_records = xagb_bitmap_count_set_regions(freesp_blocks);
+
+	/* Compute how many blocks we'll need for all the rmaps. */
+	error = xfs_btree_bload_compute_geometry(rmap_cur,
+			&rr->new_btree.bload, rr->nr_records + freesp_records);
+	if (error)
+		return error;
+
+	/* We're done when we don't need more blocks. */
+	*done = nr_blocks >= rr->new_btree.bload.nr_blocks;
+	return 0;
+}
+
+/*
+ * Iteratively reserve space for rmap btree while recording OWN_AG rmaps for
+ * the free space metadata.  This implements section (II) above.
+ */
+STATIC int
+xrep_rmap_reserve_space(
+	struct xrep_rmap	*rr,
+	struct xfs_btree_cur	*rmap_cur)
+{
+	struct xagb_bitmap	freesp_blocks;	/* AGBIT */
+	uint64_t		blocks_reserved = 0;
+	bool			done = false;
+	int			error;
+
+	/* Compute how many blocks we'll need for the rmaps collected so far. */
+	error = xfs_btree_bload_compute_geometry(rmap_cur,
+			&rr->new_btree.bload, rr->nr_records);
+	if (error)
+		return error;
+
+	/* Last chance to abort before we start committing fixes. */
+	if (xchk_should_terminate(rr->sc, &error))
+		return error;
+
+	xagb_bitmap_init(&freesp_blocks);
+
+	/*
+	 * Iteratively reserve space for the new rmapbt and recompute the
+	 * number of blocks needed to store the previously observed rmapbt
+	 * records and the ones we'll create for the free space metadata.
+	 * Finish when we don't need more blocks.
+	 */
+	do {
+		error = xrep_rmap_try_reserve(rr, rmap_cur, &freesp_blocks,
+				&blocks_reserved, &done);
+		if (error)
+			goto out_bitmap;
+	} while (!done);
+
+	/* Emit rmaps for everything in the free space bitmap. */
+	xrep_ag_btcur_init(rr->sc, &rr->sc->sa);
+	error = xrep_rmap_stash_bitmap(rr, &freesp_blocks, &XFS_RMAP_OINFO_AG);
+	xchk_ag_btcur_free(&rr->sc->sa);
+
+out_bitmap:
+	xagb_bitmap_destroy(&freesp_blocks);
+	return error;
+}
+
+/* Section (III): Building the new rmap btree. */
+
+/* Update the AGF counters. */
+STATIC int
+xrep_rmap_reset_counters(
+	struct xrep_rmap	*rr)
+{
+	struct xfs_scrub	*sc = rr->sc;
+	struct xfs_perag	*pag = sc->sa.pag;
+	struct xfs_agf		*agf = sc->sa.agf_bp->b_addr;
+	xfs_agblock_t		rmap_btblocks;
+
+	/*
+	 * The AGF header contains extra information related to the reverse
+	 * mapping btree, so we must update those fields here.
+	 */
+	rmap_btblocks = rr->new_btree.afake.af_blocks - 1;
+	agf->agf_btreeblks = cpu_to_be32(rr->freesp_btblocks + rmap_btblocks);
+	xfs_alloc_log_agf(sc->tp, sc->sa.agf_bp, XFS_AGF_BTREEBLKS);
+
+	/*
+	 * After we commit the new btree to disk, it is possible that the
+	 * process to reap the old btree blocks will race with the AIL trying
+	 * to checkpoint the old btree blocks into the filesystem.  If the new
+	 * tree is shorter than the old one, the rmapbt write verifier will
+	 * fail and the AIL will shut down the filesystem.
+	 *
+	 * To avoid this, save the old incore btree height values as the alt
+	 * height values before re-initializing the perag info from the updated
+	 * AGF to capture all the new values.
+	 */
+	pag->pagf_repair_rmap_level = pag->pagf_rmap_level;
+
+	/* Reinitialize with the values we just logged. */
+	return xrep_reinit_pagf(sc);
+}
+
+/* Retrieve rmapbt data for bulk load. */
+STATIC int
+xrep_rmap_get_records(
+	struct xfs_btree_cur	*cur,
+	unsigned int		idx,
+	struct xfs_btree_block	*block,
+	unsigned int		nr_wanted,
+	void			*priv)
+{
+	struct xrep_rmap	*rr = priv;
+	union xfs_btree_rec	*block_rec;
+	unsigned int		loaded;
+	int			error;
+
+	for (loaded = 0; loaded < nr_wanted; loaded++, idx++) {
+		int		stat = 0;
+
+		error = xfs_btree_increment(rr->mcur, 0, &stat);
+		if (error)
+			return error;
+		if (!stat)
+			return -EFSCORRUPTED;
+
+		error = xfs_rmap_get_rec(rr->mcur, &cur->bc_rec.r, &stat);
+		if (error)
+			return error;
+		if (!stat)
+			return -EFSCORRUPTED;
+
+		block_rec = xfs_btree_rec_addr(cur, idx, block);
+		cur->bc_ops->init_rec_from_cur(cur, block_rec);
+	}
+
+	return loaded;
+}
+
+/* Feed one of the new btree blocks to the bulk loader. */
+STATIC int
+xrep_rmap_claim_block(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr,
+	void			*priv)
+{
+	struct xrep_rmap        *rr = priv;
+
+	return xrep_newbt_claim_block(cur, &rr->new_btree, ptr);
+}
+
+/* Custom allocation function for new rmap btrees. */
+STATIC int
+xrep_rmap_alloc_vextent(
+	struct xfs_scrub	*sc,
+	struct xfs_alloc_arg	*args,
+	xfs_fsblock_t		alloc_hint)
+{
+	int			error;
+
+	/*
+	 * We don't want an rmap update on the allocation, since we iteratively
+	 * compute the OWN_AG records /after/ allocating blocks for the records
+	 * that we already know we need to store.  Therefore, fix the freelist
+	 * with the NORMAP flag set so that we don't also try to create an rmap
+	 * for new AGFL blocks.
+	 */
+	error = xrep_fix_freelist(sc, XFS_ALLOC_FLAG_NORMAP);
+	if (error)
+		return error;
+
+	/*
+	 * If xrep_fix_freelist fixed the freelist by moving blocks from the
+	 * free space btrees or by removing blocks from the AGFL and queueing
+	 * an EFI to free the block, the transaction will be dirty.  This
+	 * second case is of interest to us.
+	 *
+	 * Later on, we will need to compare gaps in the new recordset against
+	 * the block usage of all OWN_AG owners in order to free the old
+	 * btree's blocks, which means that we can't have EFIs for former AGFL
+	 * blocks attached to the repair transaction when we commit the new
+	 * btree.
+	 *
+	 * xrep_newbt_alloc_blocks guarantees this for us by calling
+	 * xrep_defer_finish to commit anything that fix_freelist may have
+	 * added to the transaction.
+	 */
+	return xfs_alloc_vextent_near_bno(args, alloc_hint);
+}
+
+
+/* Count the records in this btree. */
+STATIC int
+xrep_rmap_count_records(
+	struct xfs_btree_cur	*cur,
+	unsigned long long	*nr)
+{
+	int			running = 1;
+	int			error;
+
+	*nr = 0;
+
+	error = xfs_btree_goto_left_edge(cur);
+	if (error)
+		return error;
+
+	while (running && !(error = xfs_btree_increment(cur, 0, &running))) {
+		if (running)
+			(*nr)++;
+	}
+
+	return error;
+}
+/*
+ * Use the collected rmap information to stage a new rmap btree.  If this is
+ * successful we'll return with the new btree root information logged to the
+ * repair transaction but not yet committed.  This implements section (III)
+ * above.
+ */
+STATIC int
+xrep_rmap_build_new_tree(
+	struct xrep_rmap	*rr)
+{
+	struct xfs_scrub	*sc = rr->sc;
+	struct xfs_perag	*pag = sc->sa.pag;
+	struct xfs_agf		*agf = sc->sa.agf_bp->b_addr;
+	struct xfs_btree_cur	*rmap_cur;
+	xfs_fsblock_t		fsbno;
+	int			error;
+
+	/*
+	 * Preserve the old rmapbt block count so that we can adjust the
+	 * per-AG rmapbt reservation after we commit the new btree root and
+	 * want to dispose of the old btree blocks.
+	 */
+	rr->old_rmapbt_fsbcount = be32_to_cpu(agf->agf_rmap_blocks);
+
+	/*
+	 * Prepare to construct the new btree by reserving disk space for the
+	 * new btree and setting up all the accounting information we'll need
+	 * to root the new btree while it's under construction and before we
+	 * attach it to the AG header.  The new blocks are accounted to the
+	 * rmapbt per-AG reservation, which we will adjust further after
+	 * committing the new btree.
+	 */
+	fsbno = XFS_AGB_TO_FSB(sc->mp, pag->pag_agno, XFS_RMAP_BLOCK(sc->mp));
+	xrep_newbt_init_ag(&rr->new_btree, sc, &XFS_RMAP_OINFO_SKIP_UPDATE,
+			fsbno, XFS_AG_RESV_RMAPBT);
+	rr->new_btree.bload.get_records = xrep_rmap_get_records;
+	rr->new_btree.bload.claim_block = xrep_rmap_claim_block;
+	rr->new_btree.alloc_vextent = xrep_rmap_alloc_vextent;
+	rmap_cur = xfs_rmapbt_init_cursor(sc->mp, NULL, NULL, pag);
+	xfs_btree_stage_afakeroot(rmap_cur, &rr->new_btree.afake);
+
+	/*
+	 * Initialize @rr->new_btree, reserve space for the new rmapbt,
+	 * and compute OWN_AG rmaps.
+	 */
+	error = xrep_rmap_reserve_space(rr, rmap_cur);
+	if (error)
+		goto err_cur;
+
+	/*
+	 * Count the rmapbt records again, because the space reservation
+	 * for the rmapbt itself probably added more records to the btree.
+	 */
+	rr->mcur = xfs_rmapbt_mem_cursor(rr->sc->sa.pag, NULL,
+			&rr->rmap_btree);
+
+	error = xrep_rmap_count_records(rr->mcur, &rr->nr_records);
+	if (error)
+		goto err_mcur;
+
+	/*
+	 * Due to btree slack factors, it's possible for a new btree to be one
+	 * level taller than the old btree.  Update the incore btree height so
+	 * that we don't trip the verifiers when writing the new btree blocks
+	 * to disk.
+	 */
+	pag->pagf_repair_rmap_level = rr->new_btree.bload.btree_height;
+
+	/*
+	 * Move the cursor to the left edge of the tree so that the first
+	 * increment in ->get_records positions us at the first record.
+	 */
+	error = xfs_btree_goto_left_edge(rr->mcur);
+	if (error)
+		goto err_level;
+
+	/* Add all observed rmap records. */
+	error = xfs_btree_bload(rmap_cur, &rr->new_btree.bload, rr);
+	if (error)
+		goto err_level;
+
+	/*
+	 * Install the new btree in the AG header.  After this point the old
+	 * btree is no longer accessible and the new tree is live.
+	 */
+	xfs_rmapbt_commit_staged_btree(rmap_cur, sc->tp, sc->sa.agf_bp);
+	xfs_btree_del_cursor(rmap_cur, 0);
+	xfs_btree_del_cursor(rr->mcur, 0);
+	rr->mcur = NULL;
+
+	/*
+	 * Now that we've written the new btree to disk, we don't need to keep
+	 * updating the in-memory btree.  Abort the scan to stop live updates.
+	 */
+	xchk_iscan_abort(&rr->iscan);
+
+	/*
+	 * The newly committed rmap recordset includes mappings for the blocks
+	 * that we reserved to build the new btree.  If there is excess space
+	 * reservation to be freed, the corresponding rmap records must also be
+	 * removed.
+	 */
+	rr->new_btree.oinfo = XFS_RMAP_OINFO_AG;
+
+	/* Reset the AGF counters now that we've changed the btree shape. */
+	error = xrep_rmap_reset_counters(rr);
+	if (error)
+		goto err_newbt;
+
+	/* Dispose of any unused blocks and the accounting information. */
+	error = xrep_newbt_commit(&rr->new_btree);
+	if (error)
+		return error;
+
+	return xrep_roll_ag_trans(sc);
+
+err_level:
+	pag->pagf_repair_rmap_level = 0;
+err_mcur:
+	xfs_btree_del_cursor(rr->mcur, error);
+err_cur:
+	xfs_btree_del_cursor(rmap_cur, error);
+err_newbt:
+	xrep_newbt_cancel(&rr->new_btree);
+	return error;
+}
+
+/* Section (IV): Reaping the old btree. */
+
+struct xrep_rmap_find_gaps {
+	struct xagb_bitmap	rmap_gaps;
+	xfs_agblock_t		next_agbno;
+};
+
+/* Subtract each free extent in the bnobt from the rmap gaps. */
+STATIC int
+xrep_rmap_find_freesp(
+	struct xfs_btree_cur		*cur,
+	const struct xfs_alloc_rec_incore *rec,
+	void				*priv)
+{
+	struct xrep_rmap_find_gaps	*rfg = priv;
+
+	return xagb_bitmap_clear(&rfg->rmap_gaps, rec->ar_startblock,
+			rec->ar_blockcount);
+}
+
+/* Record the free space we find, as part of cleaning out the btree. */
+STATIC int
+xrep_rmap_find_gaps(
+	struct xfs_btree_cur		*cur,
+	const struct xfs_rmap_irec	*rec,
+	void				*priv)
+{
+	struct xrep_rmap_find_gaps	*rfg = priv;
+	int				error;
+
+	if (rec->rm_startblock > rfg->next_agbno) {
+		error = xagb_bitmap_set(&rfg->rmap_gaps, rfg->next_agbno,
+				rec->rm_startblock - rfg->next_agbno);
+		if (error)
+			return error;
+	}
+
+	rfg->next_agbno = max_t(xfs_agblock_t, rfg->next_agbno,
+				rec->rm_startblock + rec->rm_blockcount);
+	return 0;
+}
+
+/*
+ * Reap the old rmapbt blocks.  Now that the rmapbt is fully rebuilt, we make
+ * a list of gaps in the rmap records and a list of the extents mentioned in
+ * the bnobt.  Any block that's in the new rmapbt gap list but not mentioned
+ * in the bnobt is a block from the old rmapbt and can be removed.
+ */
+STATIC int
+xrep_rmap_remove_old_tree(
+	struct xrep_rmap	*rr)
+{
+	struct xrep_rmap_find_gaps rfg = {
+		.next_agbno	= 0,
+	};
+	struct xfs_scrub	*sc = rr->sc;
+	struct xfs_agf		*agf = sc->sa.agf_bp->b_addr;
+	struct xfs_perag	*pag = sc->sa.pag;
+	struct xfs_btree_cur	*mcur;
+	xfs_agblock_t		agend;
+	int			error;
+
+	xagb_bitmap_init(&rfg.rmap_gaps);
+
+	/* Compute free space from the new rmapbt. */
+	mcur = xfs_rmapbt_mem_cursor(rr->sc->sa.pag, NULL, &rr->rmap_btree);
+
+	error = xfs_rmap_query_all(mcur, xrep_rmap_find_gaps, &rfg);
+	xfs_btree_del_cursor(mcur, error);
+	if (error)
+		goto out_bitmap;
+
+	/* Insert a record for space between the last rmap and EOAG. */
+	agend = be32_to_cpu(agf->agf_length);
+	if (rfg.next_agbno < agend) {
+		error = xagb_bitmap_set(&rfg.rmap_gaps, rfg.next_agbno,
+				agend - rfg.next_agbno);
+		if (error)
+			goto out_bitmap;
+	}
+
+	/* Compute free space from the existing bnobt. */
+	sc->sa.bno_cur = xfs_bnobt_init_cursor(sc->mp, sc->tp, sc->sa.agf_bp,
+			sc->sa.pag);
+	error = xfs_alloc_query_all(sc->sa.bno_cur, xrep_rmap_find_freesp,
+			&rfg);
+	xfs_btree_del_cursor(sc->sa.bno_cur, error);
+	sc->sa.bno_cur = NULL;
+	if (error)
+		goto out_bitmap;
+
+	/*
+	 * Free the "free" blocks that the new rmapbt knows about but the bnobt
+	 * doesn't--these are the old rmapbt blocks.  Credit the old rmapbt
+	 * block usage count back to the per-AG rmapbt reservation (and not
+	 * fdblocks, since the rmap btree lives in free space) to keep the
+	 * reservation and free space accounting correct.
+	 */
+	error = xrep_reap_agblocks(sc, &rfg.rmap_gaps,
+			&XFS_RMAP_OINFO_ANY_OWNER, XFS_AG_RESV_RMAPBT);
+	if (error)
+		goto out_bitmap;
+
+	/*
+	 * Now that we've zapped all the old rmapbt blocks we can turn off
+	 * the alternate height mechanism and reset the per-AG space
+	 * reservation.
+	 */
+	pag->pagf_repair_rmap_level = 0;
+	sc->flags |= XREP_RESET_PERAG_RESV;
+out_bitmap:
+	xagb_bitmap_destroy(&rfg.rmap_gaps);
+	return error;
+}
+
+static inline bool
+xrep_rmapbt_want_live_update(
+	struct xchk_iscan		*iscan,
+	const struct xfs_owner_info	*oi)
+{
+	if (xchk_iscan_aborted(iscan))
+		return false;
+
+	/*
+	 * Before unlocking the AG header to perform the inode scan, we
+	 * recorded reverse mappings for all AG metadata except for the OWN_AG
+	 * metadata.  IOWs, the in-memory btree knows about the AG headers, the
+	 * two inode btrees, the CoW staging extents, and the refcount btrees.
+	 * For these types of metadata, we need to record the live updates in
+	 * the in-memory rmap btree.
+	 *
+	 * However, we do not scan the free space btrees or the AGFL until we
+	 * have re-locked the AGF and are ready to reserve space for the new
+	 * rmap btree, so we do not want live updates for OWN_AG metadata.
+	 */
+	if (XFS_RMAP_NON_INODE_OWNER(oi->oi_owner))
+		return oi->oi_owner != XFS_RMAP_OWN_AG;
+
+	/* Ignore updates to files that the scanner hasn't visited yet. */
+	return xchk_iscan_want_live_update(iscan, oi->oi_owner);
+}
+
+/*
+ * Apply a rmapbt update from the regular filesystem into our shadow btree.
+ * We're running from the thread that owns the AGF buffer and is generating
+ * the update, so we must be careful about which parts of the struct xrep_rmap
+ * that we change.
+ */
+static int
+xrep_rmapbt_live_update(
+	struct notifier_block		*nb,
+	unsigned long			action,
+	void				*data)
+{
+	struct xfs_rmap_update_params	*p = data;
+	struct xrep_rmap		*rr;
+	struct xfs_mount		*mp;
+	struct xfs_btree_cur		*mcur;
+	struct xfs_trans		*tp;
+	void				*txcookie;
+	int				error;
+
+	rr = container_of(nb, struct xrep_rmap, rhook.rmap_hook.nb);
+	mp = rr->sc->mp;
+
+	if (!xrep_rmapbt_want_live_update(&rr->iscan, &p->oinfo))
+		goto out_unlock;
+
+	trace_xrep_rmap_live_update(mp, rr->sc->sa.pag->pag_agno, action, p);
+
+	error = xrep_trans_alloc_hook_dummy(mp, &txcookie, &tp);
+	if (error)
+		goto out_abort;
+
+	mutex_lock(&rr->lock);
+	mcur = xfs_rmapbt_mem_cursor(rr->sc->sa.pag, tp, &rr->rmap_btree);
+	error = __xfs_rmap_finish_intent(mcur, action, p->startblock,
+			p->blockcount, &p->oinfo, p->unwritten);
+	xfs_btree_del_cursor(mcur, error);
+	if (error)
+		goto out_cancel;
+
+	error = xfbtree_trans_commit(&rr->rmap_btree, tp);
+	if (error)
+		goto out_cancel;
+
+	xrep_trans_cancel_hook_dummy(&txcookie, tp);
+	mutex_unlock(&rr->lock);
+	return NOTIFY_DONE;
+
+out_cancel:
+	xfbtree_trans_cancel(&rr->rmap_btree, tp);
+	xrep_trans_cancel_hook_dummy(&txcookie, tp);
+out_abort:
+	mutex_unlock(&rr->lock);
+	xchk_iscan_abort(&rr->iscan);
+out_unlock:
+	return NOTIFY_DONE;
+}
+
+/* Set up the filesystem scan components. */
+STATIC int
+xrep_rmap_setup_scan(
+	struct xrep_rmap	*rr)
+{
+	struct xfs_scrub	*sc = rr->sc;
+	int			error;
+
+	mutex_init(&rr->lock);
+
+	/* Set up in-memory rmap btree */
+	error = xfs_rmapbt_mem_init(sc->mp, &rr->rmap_btree, sc->xmbtp,
+			sc->sa.pag->pag_agno);
+	if (error)
+		goto out_mutex;
+
+	/* Retry iget every tenth of a second for up to 30 seconds. */
+	xchk_iscan_start(sc, 30000, 100, &rr->iscan);
+
+	/*
+	 * Hook into live rmap operations so that we can update our in-memory
+	 * btree to reflect live changes on the filesystem.  Since we drop the
+	 * AGF buffer to scan all the inodes, we need this piece to avoid
+	 * installing a stale btree.
+	 */
+	ASSERT(sc->flags & XCHK_FSGATES_RMAP);
+	xfs_rmap_hook_setup(&rr->rhook, xrep_rmapbt_live_update);
+	error = xfs_rmap_hook_add(sc->sa.pag, &rr->rhook);
+	if (error)
+		goto out_iscan;
+	return 0;
+
+out_iscan:
+	xchk_iscan_teardown(&rr->iscan);
+	xfbtree_destroy(&rr->rmap_btree);
+out_mutex:
+	mutex_destroy(&rr->lock);
+	return error;
+}
+
+/* Tear down scan components. */
+STATIC void
+xrep_rmap_teardown(
+	struct xrep_rmap	*rr)
+{
+	struct xfs_scrub	*sc = rr->sc;
+
+	xchk_iscan_abort(&rr->iscan);
+	xfs_rmap_hook_del(sc->sa.pag, &rr->rhook);
+	xchk_iscan_teardown(&rr->iscan);
+	xfbtree_destroy(&rr->rmap_btree);
+	mutex_destroy(&rr->lock);
+}
+
+/* Repair the rmap btree for some AG. */
+int
+xrep_rmapbt(
+	struct xfs_scrub	*sc)
+{
+	struct xrep_rmap	*rr = sc->buf;
+	int			error;
+
+	error = xrep_rmap_setup_scan(rr);
+	if (error)
+		return error;
+
+	/*
+	 * Collect rmaps for everything in this AG that isn't space metadata.
+	 * These rmaps won't change even as we try to allocate blocks.
+	 */
+	error = xrep_rmap_find_rmaps(rr);
+	if (error)
+		goto out_records;
+
+	/* Rebuild the rmap information. */
+	error = xrep_rmap_build_new_tree(rr);
+	if (error)
+		goto out_records;
+
+	/* Kill the old tree. */
+	error = xrep_rmap_remove_old_tree(rr);
+	if (error)
+		goto out_records;
+
+out_records:
+	xrep_rmap_teardown(rr);
+	return error;
+}
diff --git a/fs/xfs/scrub/rtbitmap_repair.c b/fs/xfs/scrub/rtbitmap_repair.c
index 46f5d5f605..0fef98e9f8 100644
--- a/fs/xfs/scrub/rtbitmap_repair.c
+++ b/fs/xfs/scrub/rtbitmap_repair.c
@@ -108,8 +108,6 @@ xrep_rtbitmap_data_mappings(
 				0, &map, &nmaps);
 		if (error)
 			return error;
-		if (nmaps != 1)
-			return -EFSCORRUPTED;
 
 		/* Commit new extent and all deferred work. */
 		error = xrep_defer_finish(sc);
diff --git a/fs/xfs/scrub/rtsummary.c b/fs/xfs/scrub/rtsummary.c
index b1ff4f3332..3fee603f52 100644
--- a/fs/xfs/scrub/rtsummary.c
+++ b/fs/xfs/scrub/rtsummary.c
@@ -17,10 +17,14 @@
 #include "xfs_bit.h"
 #include "xfs_bmap.h"
 #include "xfs_sb.h"
+#include "xfs_exchmaps.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
 #include "scrub/xfile.h"
+#include "scrub/repair.h"
+#include "scrub/tempexch.h"
+#include "scrub/rtsummary.h"
 
 /*
  * Realtime Summary
@@ -32,18 +36,6 @@
  * (potentially large) amount of data in pageable memory.
  */
 
-struct xchk_rtsummary {
-	struct xfs_rtalloc_args	args;
-
-	uint64_t		rextents;
-	uint64_t		rbmblocks;
-	uint64_t		rsumsize;
-	unsigned int		rsumlevels;
-
-	/* Memory buffer for the summary comparison. */
-	union xfs_suminfo_raw	words[];
-};
-
 /* Set us up to check the rtsummary file. */
 int
 xchk_setup_rtsummary(
@@ -60,6 +52,12 @@ xchk_setup_rtsummary(
 		return -ENOMEM;
 	sc->buf = rts;
 
+	if (xchk_could_repair(sc)) {
+		error = xrep_setup_rtsummary(sc, rts);
+		if (error)
+			return error;
+	}
+
 	/*
 	 * Create an xfile to construct a new rtsummary file.  The xfile allows
 	 * us to avoid pinning kernel memory for this purpose.
@@ -70,7 +68,7 @@ xchk_setup_rtsummary(
 	if (error)
 		return error;
 
-	error = xchk_trans_alloc(sc, 0);
+	error = xchk_trans_alloc(sc, rts->resblks);
 	if (error)
 		return error;
 
@@ -119,7 +117,7 @@ xfsum_load(
 	xfs_rtsumoff_t		sumoff,
 	union xfs_suminfo_raw	*rawinfo)
 {
-	return xfile_obj_load(sc->xfile, rawinfo,
+	return xfile_load(sc->xfile, rawinfo,
 			sizeof(union xfs_suminfo_raw),
 			sumoff << XFS_WORDLOG);
 }
@@ -130,19 +128,19 @@ xfsum_store(
 	xfs_rtsumoff_t		sumoff,
 	const union xfs_suminfo_raw rawinfo)
 {
-	return xfile_obj_store(sc->xfile, &rawinfo,
+	return xfile_store(sc->xfile, &rawinfo,
 			sizeof(union xfs_suminfo_raw),
 			sumoff << XFS_WORDLOG);
 }
 
-static inline int
+inline int
 xfsum_copyout(
 	struct xfs_scrub	*sc,
 	xfs_rtsumoff_t		sumoff,
 	union xfs_suminfo_raw	*rawinfo,
 	unsigned int		nr_words)
 {
-	return xfile_obj_load(sc->xfile, rawinfo, nr_words << XFS_WORDLOG,
+	return xfile_load(sc->xfile, rawinfo, nr_words << XFS_WORDLOG,
 			sumoff << XFS_WORDLOG);
 }
 
@@ -362,7 +360,12 @@ xchk_rtsummary(
 	error = xchk_rtsum_compare(sc);
 
 out_rbm:
-	/* Unlock the rtbitmap since we're done with it. */
+	/*
+	 * Unlock the rtbitmap since we're done with it.  All other writers of
+	 * the rt free space metadata grab the bitmap and summary ILOCKs in
+	 * that order, so we're still protected against allocation activities
+	 * even if we continue on to the repair function.
+	 */
 	xfs_iunlock(mp->m_rbmip, XFS_ILOCK_SHARED | XFS_ILOCK_RTBITMAP);
 	return error;
 }
diff --git a/fs/xfs/scrub/rtsummary.h b/fs/xfs/scrub/rtsummary.h
new file mode 100644
index 0000000000..e1d50304d8
--- /dev/null
+++ b/fs/xfs/scrub/rtsummary.h
@@ -0,0 +1,37 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2020-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_RTSUMMARY_H__
+#define __XFS_SCRUB_RTSUMMARY_H__
+
+struct xchk_rtsummary {
+#ifdef CONFIG_XFS_ONLINE_REPAIR
+	struct xrep_tempexch	tempexch;
+#endif
+	struct xfs_rtalloc_args	args;
+
+	uint64_t		rextents;
+	uint64_t		rbmblocks;
+	uint64_t		rsumsize;
+	unsigned int		rsumlevels;
+	unsigned int		resblks;
+
+	/* suminfo position of xfile as we write buffers to disk. */
+	xfs_rtsumoff_t		prep_wordoff;
+
+	/* Memory buffer for the summary comparison. */
+	union xfs_suminfo_raw	words[];
+};
+
+int xfsum_copyout(struct xfs_scrub *sc, xfs_rtsumoff_t sumoff,
+		union xfs_suminfo_raw *rawinfo, unsigned int nr_words);
+
+#ifdef CONFIG_XFS_ONLINE_REPAIR
+int xrep_setup_rtsummary(struct xfs_scrub *sc, struct xchk_rtsummary *rts);
+#else
+# define xrep_setup_rtsummary(sc, rts)	(0)
+#endif /* CONFIG_XFS_ONLINE_REPAIR */
+
+#endif /* __XFS_SCRUB_RTSUMMARY_H__ */
diff --git a/fs/xfs/scrub/rtsummary_repair.c b/fs/xfs/scrub/rtsummary_repair.c
new file mode 100644
index 0000000000..d9e971c4c7
--- /dev/null
+++ b/fs/xfs/scrub/rtsummary_repair.c
@@ -0,0 +1,175 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2020-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_btree.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_rtalloc.h"
+#include "xfs_inode.h"
+#include "xfs_bit.h"
+#include "xfs_bmap.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_exchmaps.h"
+#include "xfs_rtbitmap.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+#include "scrub/repair.h"
+#include "scrub/tempfile.h"
+#include "scrub/tempexch.h"
+#include "scrub/reap.h"
+#include "scrub/xfile.h"
+#include "scrub/rtsummary.h"
+
+/* Set us up to repair the rtsummary file. */
+int
+xrep_setup_rtsummary(
+	struct xfs_scrub	*sc,
+	struct xchk_rtsummary	*rts)
+{
+	struct xfs_mount	*mp = sc->mp;
+	unsigned long long	blocks;
+	int			error;
+
+	error = xrep_tempfile_create(sc, S_IFREG);
+	if (error)
+		return error;
+
+	/*
+	 * If we're doing a repair, we reserve enough blocks to write out a
+	 * completely new summary file, plus twice as many blocks as we would
+	 * need if we can only allocate one block per data fork mapping.  This
+	 * should cover the preallocation of the temporary file and exchanging
+	 * the extent mappings.
+	 *
+	 * We cannot use xfs_exchmaps_estimate because we have not yet
+	 * constructed the replacement rtsummary and therefore do not know how
+	 * many extents it will use.  By the time we do, we will have a dirty
+	 * transaction (which we cannot drop because we cannot drop the
+	 * rtsummary ILOCK) and cannot ask for more reservation.
+	 */
+	blocks = XFS_B_TO_FSB(mp, mp->m_rsumsize);
+	blocks += xfs_bmbt_calc_size(mp, blocks) * 2;
+	if (blocks > UINT_MAX)
+		return -EOPNOTSUPP;
+
+	rts->resblks += blocks;
+	return 0;
+}
+
+static int
+xrep_rtsummary_prep_buf(
+	struct xfs_scrub	*sc,
+	struct xfs_buf		*bp,
+	void			*data)
+{
+	struct xchk_rtsummary	*rts = data;
+	struct xfs_mount	*mp = sc->mp;
+	union xfs_suminfo_raw	*ondisk;
+	int			error;
+
+	rts->args.mp = sc->mp;
+	rts->args.tp = sc->tp;
+	rts->args.sumbp = bp;
+	ondisk = xfs_rsumblock_infoptr(&rts->args, 0);
+	rts->args.sumbp = NULL;
+
+	bp->b_ops = &xfs_rtbuf_ops;
+
+	error = xfsum_copyout(sc, rts->prep_wordoff, ondisk, mp->m_blockwsize);
+	if (error)
+		return error;
+
+	rts->prep_wordoff += mp->m_blockwsize;
+	xfs_trans_buf_set_type(sc->tp, bp, XFS_BLFT_RTSUMMARY_BUF);
+	return 0;
+}
+
+/* Repair the realtime summary. */
+int
+xrep_rtsummary(
+	struct xfs_scrub	*sc)
+{
+	struct xchk_rtsummary	*rts = sc->buf;
+	struct xfs_mount	*mp = sc->mp;
+	xfs_filblks_t		rsumblocks;
+	int			error;
+
+	/* We require the rmapbt to rebuild anything. */
+	if (!xfs_has_rmapbt(mp))
+		return -EOPNOTSUPP;
+	/* We require atomic file exchange range to rebuild anything. */
+	if (!xfs_has_exchange_range(mp))
+		return -EOPNOTSUPP;
+
+	/* Walk away if we disagree on the size of the rt bitmap. */
+	if (rts->rbmblocks != mp->m_sb.sb_rbmblocks)
+		return 0;
+
+	/* Make sure any problems with the fork are fixed. */
+	error = xrep_metadata_inode_forks(sc);
+	if (error)
+		return error;
+
+	/*
+	 * Try to take ILOCK_EXCL of the temporary file.  We had better be the
+	 * only ones holding onto this inode, but we can't block while holding
+	 * the rtsummary file's ILOCK_EXCL.
+	 */
+	while (!xrep_tempfile_ilock_nowait(sc)) {
+		if (xchk_should_terminate(sc, &error))
+			return error;
+		delay(1);
+	}
+
+	/* Make sure we have space allocated for the entire summary file. */
+	rsumblocks = XFS_B_TO_FSB(mp, rts->rsumsize);
+	xfs_trans_ijoin(sc->tp, sc->ip, 0);
+	xfs_trans_ijoin(sc->tp, sc->tempip, 0);
+	error = xrep_tempfile_prealloc(sc, 0, rsumblocks);
+	if (error)
+		return error;
+
+	/* Last chance to abort before we start committing fixes. */
+	if (xchk_should_terminate(sc, &error))
+		return error;
+
+	/* Copy the rtsummary file that we generated. */
+	error = xrep_tempfile_copyin(sc, 0, rsumblocks,
+			xrep_rtsummary_prep_buf, rts);
+	if (error)
+		return error;
+	error = xrep_tempfile_set_isize(sc, rts->rsumsize);
+	if (error)
+		return error;
+
+	/*
+	 * Now exchange the contents.  Nothing in repair uses the temporary
+	 * buffer, so we can reuse it for the tempfile exchrange information.
+	 */
+	error = xrep_tempexch_trans_reserve(sc, XFS_DATA_FORK, &rts->tempexch);
+	if (error)
+		return error;
+
+	error = xrep_tempexch_contents(sc, &rts->tempexch);
+	if (error)
+		return error;
+
+	/* Reset incore state and blow out the summary cache. */
+	if (mp->m_rsum_cache)
+		memset(mp->m_rsum_cache, 0xFF, mp->m_sb.sb_rbmblocks);
+
+	mp->m_rsumlevels = rts->rsumlevels;
+	mp->m_rsumsize = rts->rsumsize;
+
+	/* Free the old rtsummary blocks if they're not in use. */
+	return xrep_reap_ifork(sc, sc->tempip, XFS_DATA_FORK);
+}
diff --git a/fs/xfs/scrub/scrub.c b/fs/xfs/scrub/scrub.c
index caf324c2b9..4cbcf7a86d 100644
--- a/fs/xfs/scrub/scrub.c
+++ b/fs/xfs/scrub/scrub.c
@@ -15,6 +15,13 @@
 #include "xfs_quota.h"
 #include "xfs_qm.h"
 #include "xfs_scrub.h"
+#include "xfs_buf_mem.h"
+#include "xfs_rmap.h"
+#include "xfs_exchrange.h"
+#include "xfs_exchmaps.h"
+#include "xfs_dir2.h"
+#include "xfs_parent.h"
+#include "xfs_icache.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
@@ -22,6 +29,8 @@
 #include "scrub/health.h"
 #include "scrub/stats.h"
 #include "scrub/xfile.h"
+#include "scrub/tempfile.h"
+#include "scrub/orphanage.h"
 
 /*
  * Online Scrub and Repair
@@ -157,9 +166,51 @@ xchk_fsgates_disable(
 	if (sc->flags & XCHK_FSGATES_DRAIN)
 		xfs_drain_wait_disable();
 
+	if (sc->flags & XCHK_FSGATES_QUOTA)
+		xfs_dqtrx_hook_disable();
+
+	if (sc->flags & XCHK_FSGATES_DIRENTS)
+		xfs_dir_hook_disable();
+
+	if (sc->flags & XCHK_FSGATES_RMAP)
+		xfs_rmap_hook_disable();
+
 	sc->flags &= ~XCHK_FSGATES_ALL;
 }
 
+/* Free the resources associated with a scrub subtype. */
+void
+xchk_scrub_free_subord(
+	struct xfs_scrub_subord	*sub)
+{
+	struct xfs_scrub	*sc = sub->parent_sc;
+
+	ASSERT(sc->ip == sub->sc.ip);
+	ASSERT(sc->orphanage == sub->sc.orphanage);
+	ASSERT(sc->tempip == sub->sc.tempip);
+
+	sc->sm->sm_type = sub->old_smtype;
+	sc->sm->sm_flags = sub->old_smflags |
+				(sc->sm->sm_flags & XFS_SCRUB_FLAGS_OUT);
+	sc->tp = sub->sc.tp;
+
+	if (sub->sc.buf) {
+		if (sub->sc.buf_cleanup)
+			sub->sc.buf_cleanup(sub->sc.buf);
+		kvfree(sub->sc.buf);
+	}
+	if (sub->sc.xmbtp)
+		xmbuf_free(sub->sc.xmbtp);
+	if (sub->sc.xfile)
+		xfile_destroy(sub->sc.xfile);
+
+	sc->ilock_flags = sub->sc.ilock_flags;
+	sc->orphanage_ilock_flags = sub->sc.orphanage_ilock_flags;
+	sc->temp_ilock_flags = sub->sc.temp_ilock_flags;
+
+	kfree(sub);
+}
+
 /* Free all the resources and finish the transactions. */
 STATIC int
 xchk_teardown(
@@ -184,6 +235,10 @@ xchk_teardown(
 		sc->flags &= ~XCHK_HAVE_FREEZE_PROT;
 		mnt_drop_write_file(sc->file);
 	}
+	if (sc->xmbtp) {
+		xmbuf_free(sc->xmbtp);
+		sc->xmbtp = NULL;
+	}
 	if (sc->xfile) {
 		xfile_destroy(sc->xfile);
 		sc->xfile = NULL;
@@ -196,6 +251,8 @@ xchk_teardown(
 		sc->buf = NULL;
 	}
 
+	xrep_tempfile_rele(sc);
+	xrep_orphanage_rele(sc);
 	xchk_fsgates_disable(sc);
 	return error;
 }
@@ -267,7 +324,7 @@ static const struct xchk_meta_ops meta_scrub_ops[] = {
 		.setup	= xchk_setup_ag_rmapbt,
 		.scrub	= xchk_rmapbt,
 		.has	= xfs_has_rmapbt,
-		.repair	= xrep_notsupported,
+		.repair	= xrep_rmapbt,
 	},
 	[XFS_SCRUB_TYPE_REFCNTBT] = {	/* refcountbt */
 		.type	= ST_PERAG,
@@ -304,25 +361,25 @@ static const struct xchk_meta_ops meta_scrub_ops[] = {
 		.type	= ST_INODE,
 		.setup	= xchk_setup_directory,
 		.scrub	= xchk_directory,
-		.repair	= xrep_notsupported,
+		.repair	= xrep_directory,
 	},
 	[XFS_SCRUB_TYPE_XATTR] = {	/* extended attributes */
 		.type	= ST_INODE,
 		.setup	= xchk_setup_xattr,
 		.scrub	= xchk_xattr,
-		.repair	= xrep_notsupported,
+		.repair	= xrep_xattr,
 	},
 	[XFS_SCRUB_TYPE_SYMLINK] = {	/* symbolic link */
 		.type	= ST_INODE,
 		.setup	= xchk_setup_symlink,
 		.scrub	= xchk_symlink,
-		.repair	= xrep_notsupported,
+		.repair	= xrep_symlink,
 	},
 	[XFS_SCRUB_TYPE_PARENT] = {	/* parent pointers */
 		.type	= ST_INODE,
 		.setup	= xchk_setup_parent,
 		.scrub	= xchk_parent,
-		.repair	= xrep_notsupported,
+		.repair	= xrep_parent,
 	},
 	[XFS_SCRUB_TYPE_RTBITMAP] = {	/* realtime bitmap */
 		.type	= ST_FS,
@@ -334,7 +391,7 @@ static const struct xchk_meta_ops meta_scrub_ops[] = {
 		.type	= ST_FS,
 		.setup	= xchk_setup_rtsummary,
 		.scrub	= xchk_rtsummary,
-		.repair	= xrep_notsupported,
+		.repair	= xrep_rtsummary,
 	},
 	[XFS_SCRUB_TYPE_UQUOTA] = {	/* user quota */
 		.type	= ST_FS,
@@ -358,7 +415,32 @@ static const struct xchk_meta_ops meta_scrub_ops[] = {
 		.type	= ST_FS,
 		.setup	= xchk_setup_fscounters,
 		.scrub	= xchk_fscounters,
-		.repair	= xrep_notsupported,
+		.repair	= xrep_fscounters,
+	},
+	[XFS_SCRUB_TYPE_QUOTACHECK] = {	/* quota counters */
+		.type	= ST_FS,
+		.setup	= xchk_setup_quotacheck,
+		.scrub	= xchk_quotacheck,
+		.repair	= xrep_quotacheck,
+	},
+	[XFS_SCRUB_TYPE_NLINKS] = {	/* inode link counts */
+		.type	= ST_FS,
+		.setup	= xchk_setup_nlinks,
+		.scrub	= xchk_nlinks,
+		.repair	= xrep_nlinks,
+	},
+	[XFS_SCRUB_TYPE_HEALTHY] = {	/* fs healthy; clean all reminders */
+		.type	= ST_FS,
+		.setup	= xchk_setup_fs,
+		.scrub	= xchk_health_record,
+		.repair = xrep_notsupported,
+	},
+	[XFS_SCRUB_TYPE_DIRTREE] = {	/* directory tree structure */
+		.type	= ST_INODE,
+		.setup	= xchk_setup_dirtree,
+		.scrub	= xchk_dirtree,
+		.has	= xfs_has_parent,
+		.repair	= xrep_dirtree,
 	},
 };
 
@@ -464,8 +546,38 @@ static inline void xchk_postmortem(struct xfs_scrub *sc)
 }
 #endif /* CONFIG_XFS_ONLINE_REPAIR */
 
+/*
+ * Create a new scrub context from an existing one, but with a different scrub
+ * type.
+ */
+struct xfs_scrub_subord *
+xchk_scrub_create_subord(
+	struct xfs_scrub	*sc,
+	unsigned int		subtype)
+{
+	struct xfs_scrub_subord	*sub;
+
+	sub = kzalloc(sizeof(*sub), XCHK_GFP_FLAGS);
+	if (!sub)
+		return ERR_PTR(-ENOMEM);
+
+	sub->old_smtype = sc->sm->sm_type;
+	sub->old_smflags = sc->sm->sm_flags;
+	sub->parent_sc = sc;
+	memcpy(&sub->sc, sc, sizeof(struct xfs_scrub));
+	sub->sc.ops = &meta_scrub_ops[subtype];
+	sub->sc.sm->sm_type = subtype;
+	sub->sc.sm->sm_flags &= ~XFS_SCRUB_FLAGS_OUT;
+	sub->sc.buf = NULL;
+	sub->sc.buf_cleanup = NULL;
+	sub->sc.xfile = NULL;
+	sub->sc.xmbtp = NULL;
+
+	return sub;
+}
+
 /* Dispatch metadata scrubbing. */
-int
+STATIC int
 xfs_scrub_metadata(
 	struct file			*file,
 	struct xfs_scrub_metadata	*sm)
@@ -507,6 +619,7 @@ xfs_scrub_metadata(
 	sc->sm = sm;
 	sc->ops = &meta_scrub_ops[sm->sm_type];
 	sc->sick_mask = xchk_health_mask_for_scrub_type(sm->sm_type);
+	sc->relax = INIT_XCHK_RELAX;
 retry_op:
 	/*
 	 * When repairs are allowed, prevent freezing or readonly remount while
@@ -610,3 +723,221 @@ try_harder:
 	run.retries++;
 	goto retry_op;
 }
+
+/* Scrub one aspect of one piece of metadata. */
+int
+xfs_ioc_scrub_metadata(
+	struct file			*file,
+	void				__user *arg)
+{
+	struct xfs_scrub_metadata	scrub;
+	int				error;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	if (copy_from_user(&scrub, arg, sizeof(scrub)))
+		return -EFAULT;
+
+	error = xfs_scrub_metadata(file, &scrub);
+	if (error)
+		return error;
+
+	if (copy_to_user(arg, &scrub, sizeof(scrub)))
+		return -EFAULT;
+
+	return 0;
+}
+
+/* Decide if there have been any scrub failures up to this point. */
+static inline int
+xfs_scrubv_check_barrier(
+	struct xfs_mount		*mp,
+	const struct xfs_scrub_vec	*vectors,
+	const struct xfs_scrub_vec	*stop_vec)
+{
+	const struct xfs_scrub_vec	*v;
+	__u32				failmask;
+
+	failmask = stop_vec->sv_flags & XFS_SCRUB_FLAGS_OUT;
+
+	for (v = vectors; v < stop_vec; v++) {
+		if (v->sv_type == XFS_SCRUB_TYPE_BARRIER)
+			continue;
+
+		/*
+		 * Runtime errors count as a previous failure, except the ones
+		 * used to ask userspace to retry.
+		 */
+		switch (v->sv_ret) {
+		case -EBUSY:
+		case -ENOENT:
+		case -EUSERS:
+		case 0:
+			break;
+		default:
+			return -ECANCELED;
+		}
+
+		/*
+		 * If any of the out-flags on the scrub vector match the mask
+		 * that was set on the barrier vector, that's a previous fail.
+		 */
+		if (v->sv_flags & failmask)
+			return -ECANCELED;
+	}
+
+	return 0;
+}
+
+/*
+ * If the caller provided us with a nonzero inode number that isn't the ioctl
+ * file, try to grab a reference to it to eliminate all further untrusted inode
+ * lookups.  If we can't get the inode, let each scrub function try again.
+ */
+STATIC struct xfs_inode *
+xchk_scrubv_open_by_handle(
+	struct xfs_mount		*mp,
+	const struct xfs_scrub_vec_head	*head)
+{
+	struct xfs_trans		*tp;
+	struct xfs_inode		*ip;
+	int				error;
+
+	error = xfs_trans_alloc_empty(mp, &tp);
+	if (error)
+		return NULL;
+
+	error = xfs_iget(mp, tp, head->svh_ino, XCHK_IGET_FLAGS, 0, &ip);
+	xfs_trans_cancel(tp);
+	if (error)
+		return NULL;
+
+	if (VFS_I(ip)->i_generation != head->svh_gen) {
+		xfs_irele(ip);
+		return NULL;
+	}
+
+	return ip;
+}
+
+/* Vectored scrub implementation to reduce ioctl calls. */
+int
+xfs_ioc_scrubv_metadata(
+	struct file			*file,
+	void				__user *arg)
+{
+	struct xfs_scrub_vec_head	head;
+	struct xfs_scrub_vec_head	__user *uhead = arg;
+	struct xfs_scrub_vec		*vectors;
+	struct xfs_scrub_vec		__user *uvectors;
+	struct xfs_inode		*ip_in = XFS_I(file_inode(file));
+	struct xfs_mount		*mp = ip_in->i_mount;
+	struct xfs_inode		*handle_ip = NULL;
+	struct xfs_scrub_vec		*v;
+	size_t				vec_bytes;
+	unsigned int			i;
+	int				error = 0;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	if (copy_from_user(&head, uhead, sizeof(head)))
+		return -EFAULT;
+
+	if (head.svh_reserved)
+		return -EINVAL;
+	if (head.svh_flags & ~XFS_SCRUB_VEC_FLAGS_ALL)
+		return -EINVAL;
+	if (head.svh_nr == 0)
+		return 0;
+
+	vec_bytes = array_size(head.svh_nr, sizeof(struct xfs_scrub_vec));
+	if (vec_bytes > PAGE_SIZE)
+		return -ENOMEM;
+
+	uvectors = u64_to_user_ptr(head.svh_vectors);
+	vectors = memdup_user(uvectors, vec_bytes);
+	if (IS_ERR(vectors))
+		return PTR_ERR(vectors);
+
+	trace_xchk_scrubv_start(ip_in, &head);
+
+	for (i = 0, v = vectors; i < head.svh_nr; i++, v++) {
+		if (v->sv_reserved) {
+			error = -EINVAL;
+			goto out_free;
+		}
+
+		if (v->sv_type == XFS_SCRUB_TYPE_BARRIER &&
+		    (v->sv_flags & ~XFS_SCRUB_FLAGS_OUT)) {
+			error = -EINVAL;
+			goto out_free;
+		}
+
+		trace_xchk_scrubv_item(mp, &head, i, v);
+	}
+
+	/*
+	 * If the caller wants us to do a scrub-by-handle and the file used to
+	 * call the ioctl is not the same file, load the incore inode and pin
+	 * it across all the scrubv actions to avoid repeated UNTRUSTED
+	 * lookups.  The reference is not passed to deeper layers of scrub
+	 * because each scrubber gets to decide its own strategy and return
+	 * values for getting an inode.
+	 */
+	if (head.svh_ino && head.svh_ino != ip_in->i_ino)
+		handle_ip = xchk_scrubv_open_by_handle(mp, &head);
+
+	/* Run all the scrubbers. */
+	for (i = 0, v = vectors; i < head.svh_nr; i++, v++) {
+		struct xfs_scrub_metadata	sm = {
+			.sm_type		= v->sv_type,
+			.sm_flags		= v->sv_flags,
+			.sm_ino			= head.svh_ino,
+			.sm_gen			= head.svh_gen,
+			.sm_agno		= head.svh_agno,
+		};
+
+		if (v->sv_type == XFS_SCRUB_TYPE_BARRIER) {
+			v->sv_ret = xfs_scrubv_check_barrier(mp, vectors, v);
+			if (v->sv_ret) {
+				trace_xchk_scrubv_barrier_fail(mp, &head, i, v);
+				break;
+			}
+
+			continue;
+		}
+
+		v->sv_ret = xfs_scrub_metadata(file, &sm);
+		v->sv_flags = sm.sm_flags;
+
+		trace_xchk_scrubv_outcome(mp, &head, i, v);
+
+		if (head.svh_rest_us) {
+			ktime_t		expires;
+
+			expires = ktime_add_ns(ktime_get(),
+					head.svh_rest_us * 1000);
+			set_current_state(TASK_KILLABLE);
+			schedule_hrtimeout(&expires, HRTIMER_MODE_ABS);
+		}
+
+		if (fatal_signal_pending(current)) {
+			error = -EINTR;
+			goto out_free;
+		}
+	}
+
+	if (copy_to_user(uvectors, vectors, vec_bytes) ||
+	    copy_to_user(uhead, &head, sizeof(head))) {
+		error = -EFAULT;
+		goto out_free;
+	}
+
+out_free:
+	if (handle_ip)
+		xfs_irele(handle_ip);
+	kfree(vectors);
+	return error;
+}
diff --git a/fs/xfs/scrub/scrub.h b/fs/xfs/scrub/scrub.h
index 7fc50654c4..1bc33f010d 100644
--- a/fs/xfs/scrub/scrub.h
+++ b/fs/xfs/scrub/scrub.h
@@ -8,6 +8,49 @@
 
 struct xfs_scrub;
 
+struct xchk_relax {
+	unsigned long	next_resched;
+	unsigned int	resched_nr;
+	bool		interruptible;
+};
+
+/* Yield to the scheduler at most 10x per second. */
+#define XCHK_RELAX_NEXT		(jiffies + (HZ / 10))
+
+#define INIT_XCHK_RELAX	\
+	(struct xchk_relax){ \
+		.next_resched	= XCHK_RELAX_NEXT, \
+		.resched_nr	= 0, \
+		.interruptible	= true, \
+	}
+
+/*
+ * Relax during a scrub operation and exit if there's a fatal signal pending.
+ *
+ * If preemption is disabled, we need to yield to the scheduler every now and
+ * then so that we don't run afoul of the soft lockup watchdog or RCU stall
+ * detector.  cond_resched calls are somewhat expensive (~5ns) so we want to
+ * ratelimit this to 10x per second.  Amortize the cost of the other checks by
+ * only doing it once every 100 calls.
+ */
+static inline int xchk_maybe_relax(struct xchk_relax *widget)
+{
+	/* Amortize the cost of scheduling and checking signals. */
+	if (likely(++widget->resched_nr < 100))
+		return 0;
+	widget->resched_nr = 0;
+
+	if (unlikely(widget->next_resched <= jiffies)) {
+		cond_resched();
+		widget->next_resched = XCHK_RELAX_NEXT;
+	}
+
+	if (widget->interruptible && fatal_signal_pending(current))
+		return -EINTR;
+
+	return 0;
+}
+
 /*
  * Standard flags for allocating memory within scrub.  NOFS context is
  * configured by the process allocation scope.  Scrub and repair must be able
@@ -17,6 +60,13 @@ struct xfs_scrub;
 #define XCHK_GFP_FLAGS	((__force gfp_t)(GFP_KERNEL | __GFP_NOWARN | \
 					 __GFP_RETRY_MAYFAIL))
 
+/*
+ * For opening files by handle for fsck operations, we don't trust the inumber
+ * or the allocation state; therefore, perform an untrusted lookup.  We don't
+ * want these inodes to pollute the cache, so mark them for immediate removal.
+ */
+#define XCHK_IGET_FLAGS	(XFS_IGET_UNTRUSTED | XFS_IGET_DONTCACHE)
+
 /* Type info and names for the scrub types. */
 enum xchk_type {
 	ST_NONE = 1,	/* disabled */
@@ -99,9 +149,20 @@ struct xfs_scrub {
 	/* xfile used by the scrubbers; freed at teardown. */
 	struct xfile			*xfile;
 
+	/* buffer target for in-memory btrees; also freed at teardown. */
+	struct xfs_buftarg		*xmbtp;
+
 	/* Lock flags for @ip. */
 	uint				ilock_flags;
 
+	/* The orphanage, for stashing files that have lost their parent. */
+	uint				orphanage_ilock_flags;
+	struct xfs_inode		*orphanage;
+
+	/* A temporary file on this filesystem, for staging new metadata. */
+	struct xfs_inode		*tempip;
+	uint				temp_ilock_flags;
+
 	/* See the XCHK/XREP state flags below. */
 	unsigned int			flags;
 
@@ -112,6 +173,9 @@ struct xfs_scrub {
 	 */
 	unsigned int			sick_mask;
 
+	/* next time we want to cond_resched() */
+	struct xchk_relax		relax;
+
 	/* State tracking for single-AG operations. */
 	struct xchk_ag			sa;
 };
@@ -121,6 +185,9 @@ struct xfs_scrub {
 #define XCHK_HAVE_FREEZE_PROT	(1U << 1)  /* do we have freeze protection? */
 #define XCHK_FSGATES_DRAIN	(1U << 2)  /* defer ops draining enabled */
 #define XCHK_NEED_DRAIN		(1U << 3)  /* scrub needs to drain defer ops */
+#define XCHK_FSGATES_QUOTA	(1U << 4)  /* quota live update enabled */
+#define XCHK_FSGATES_DIRENTS	(1U << 5)  /* directory live update enabled */
+#define XCHK_FSGATES_RMAP	(1U << 6)  /* rmapbt live update enabled */
 #define XREP_RESET_PERAG_RESV	(1U << 30) /* must reset AG space reservation */
 #define XREP_ALREADY_FIXED	(1U << 31) /* checking our repair work */
 
@@ -130,7 +197,39 @@ struct xfs_scrub {
  * features are gated off via dynamic code patching, which is why the state
  * must be enabled during scrub setup and can only be torn down afterwards.
  */
-#define XCHK_FSGATES_ALL	(XCHK_FSGATES_DRAIN)
+#define XCHK_FSGATES_ALL	(XCHK_FSGATES_DRAIN | \
+				 XCHK_FSGATES_QUOTA | \
+				 XCHK_FSGATES_DIRENTS | \
+				 XCHK_FSGATES_RMAP)
+
+struct xfs_scrub_subord {
+	struct xfs_scrub	sc;
+	struct xfs_scrub	*parent_sc;
+	unsigned int		old_smtype;
+	unsigned int		old_smflags;
+};
+
+struct xfs_scrub_subord *xchk_scrub_create_subord(struct xfs_scrub *sc,
+		unsigned int subtype);
+void xchk_scrub_free_subord(struct xfs_scrub_subord *sub);
+
+/*
+ * We /could/ terminate a scrub/repair operation early.  If we're not
+ * in a good place to continue (fatal signal, etc.) then bail out.
+ * Note that we're careful not to make any judgements about *error.
+ */
+static inline bool
+xchk_should_terminate(
+	struct xfs_scrub	*sc,
+	int			*error)
+{
+	if (xchk_maybe_relax(&sc->relax)) {
+		if (*error == 0)
+			*error = -EINTR;
+		return true;
+	}
+	return false;
+}
 
 /* Metadata scrubbers */
 int xchk_tester(struct xfs_scrub *sc);
@@ -150,6 +249,7 @@ int xchk_directory(struct xfs_scrub *sc);
 int xchk_xattr(struct xfs_scrub *sc);
 int xchk_symlink(struct xfs_scrub *sc);
 int xchk_parent(struct xfs_scrub *sc);
+int xchk_dirtree(struct xfs_scrub *sc);
 #ifdef CONFIG_XFS_RT
 int xchk_rtbitmap(struct xfs_scrub *sc);
 int xchk_rtsummary(struct xfs_scrub *sc);
@@ -167,14 +267,21 @@ xchk_rtsummary(struct xfs_scrub *sc)
 #endif
 #ifdef CONFIG_XFS_QUOTA
 int xchk_quota(struct xfs_scrub *sc);
+int xchk_quotacheck(struct xfs_scrub *sc);
 #else
 static inline int
 xchk_quota(struct xfs_scrub *sc)
 {
 	return -ENOENT;
 }
+static inline int
+xchk_quotacheck(struct xfs_scrub *sc)
+{
+	return -ENOENT;
+}
 #endif
 int xchk_fscounters(struct xfs_scrub *sc);
+int xchk_nlinks(struct xfs_scrub *sc);
 
 /* cross-referencing helpers */
 void xchk_xref_is_used_space(struct xfs_scrub *sc, xfs_agblock_t agbno,
diff --git a/fs/xfs/scrub/stats.c b/fs/xfs/scrub/stats.c
index cd91db4a55..7996c23354 100644
--- a/fs/xfs/scrub/stats.c
+++ b/fs/xfs/scrub/stats.c
@@ -77,6 +77,9 @@ static const char *name_map[XFS_SCRUB_TYPE_NR] = {
 	[XFS_SCRUB_TYPE_GQUOTA]		= "grpquota",
 	[XFS_SCRUB_TYPE_PQUOTA]		= "prjquota",
 	[XFS_SCRUB_TYPE_FSCOUNTERS]	= "fscounters",
+	[XFS_SCRUB_TYPE_QUOTACHECK]	= "quotacheck",
+	[XFS_SCRUB_TYPE_NLINKS]		= "nlinks",
+	[XFS_SCRUB_TYPE_DIRTREE]	= "dirtree",
 };
 
 /* Format the scrub stats into a text buffer, similar to pcp style. */
@@ -329,9 +332,9 @@ xchk_stats_register(
 	if (!cs->cs_debugfs)
 		return;
 
-	debugfs_create_file("stats", 0644, cs->cs_debugfs, cs,
+	debugfs_create_file("stats", 0444, cs->cs_debugfs, cs,
 			&scrub_stats_fops);
-	debugfs_create_file("clear_stats", 0400, cs->cs_debugfs, cs,
+	debugfs_create_file("clear_stats", 0200, cs->cs_debugfs, cs,
 			&clear_scrub_stats_fops);
 }
 
diff --git a/fs/xfs/scrub/symlink.c b/fs/xfs/scrub/symlink.c
index ddff86713d..c848bcc07c 100644
--- a/fs/xfs/scrub/symlink.c
+++ b/fs/xfs/scrub/symlink.c
@@ -10,24 +10,36 @@
 #include "xfs_trans_resv.h"
 #include "xfs_mount.h"
 #include "xfs_log_format.h"
+#include "xfs_trans.h"
 #include "xfs_inode.h"
 #include "xfs_symlink.h"
 #include "xfs_health.h"
+#include "xfs_symlink_remote.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/health.h"
+#include "scrub/repair.h"
 
 /* Set us up to scrub a symbolic link. */
 int
 xchk_setup_symlink(
 	struct xfs_scrub	*sc)
 {
+	unsigned int		resblks = 0;
+	int			error;
+
 	/* Allocate the buffer without the inode lock held. */
 	sc->buf = kvzalloc(XFS_SYMLINK_MAXLEN + 1, XCHK_GFP_FLAGS);
 	if (!sc->buf)
 		return -ENOMEM;
 
-	return xchk_setup_inode_contents(sc, 0);
+	if (xchk_could_repair(sc)) {
+		error = xrep_setup_symlink(sc, &resblks);
+		if (error)
+			return error;
+	}
+
+	return xchk_setup_inode_contents(sc, resblks);
 }
 
 /* Symbolic links. */
@@ -67,7 +79,7 @@ xchk_symlink(
 	}
 
 	/* Remote symlink; must read the contents. */
-	error = xfs_readlink_bmap_ilocked(sc->ip, sc->buf);
+	error = xfs_symlink_remote_read(sc->ip, sc->buf);
 	if (!xchk_fblock_process_error(sc, XFS_DATA_FORK, 0, &error))
 		return error;
 	if (strnlen(sc->buf, XFS_SYMLINK_MAXLEN) < len)
diff --git a/fs/xfs/scrub/symlink_repair.c b/fs/xfs/scrub/symlink_repair.c
new file mode 100644
index 0000000000..d015a86ef4
--- /dev/null
+++ b/fs/xfs/scrub/symlink_repair.c
@@ -0,0 +1,509 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2018-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_inode_fork.h"
+#include "xfs_symlink.h"
+#include "xfs_bmap.h"
+#include "xfs_quota.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_trans_space.h"
+#include "xfs_symlink_remote.h"
+#include "xfs_exchmaps.h"
+#include "xfs_exchrange.h"
+#include "xfs_health.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+#include "scrub/repair.h"
+#include "scrub/tempfile.h"
+#include "scrub/tempexch.h"
+#include "scrub/reap.h"
+
+/*
+ * Symbolic Link Repair
+ * ====================
+ *
+ * We repair symbolic links by reading whatever target data we can find, up to
+ * the first NULL byte.  If the recovered target strlen matches i_size, then
+ * we rewrite the target.  In all other cases, we replace the target with an
+ * overly long string that cannot possibly resolve.  The new target is written
+ * into a private hidden temporary file, and then a file contents exchange
+ * commits the new symlink target to the file being repaired.
+ */
+
+/* Set us up to repair the symlink file. */
+int
+xrep_setup_symlink(
+	struct xfs_scrub	*sc,
+	unsigned int		*resblks)
+{
+	struct xfs_mount	*mp = sc->mp;
+	unsigned long long	blocks;
+	int			error;
+
+	error = xrep_tempfile_create(sc, S_IFLNK);
+	if (error)
+		return error;
+
+	/*
+	 * If we're doing a repair, we reserve enough blocks to write out a
+	 * completely new symlink file, plus twice as many blocks as we would
+	 * need if we can only allocate one block per data fork mapping.  This
+	 * should cover the preallocation of the temporary file and exchanging
+	 * the extent mappings.
+	 *
+	 * We cannot use xfs_exchmaps_estimate because we have not yet
+	 * constructed the replacement symlink and therefore do not know how
+	 * many extents it will use.  By the time we do, we will have a dirty
+	 * transaction (which we cannot drop because we cannot drop the
+	 * symlink ILOCK) and cannot ask for more reservation.
+	 */
+	blocks = xfs_symlink_blocks(sc->mp, XFS_SYMLINK_MAXLEN);
+	blocks += xfs_bmbt_calc_size(mp, blocks) * 2;
+	if (blocks > UINT_MAX)
+		return -EOPNOTSUPP;
+
+	*resblks += blocks;
+	return 0;
+}
+
+/*
+ * Try to salvage the pathname from remote blocks.  Returns the number of bytes
+ * salvaged or a negative errno.
+ */
+STATIC ssize_t
+xrep_symlink_salvage_remote(
+	struct xfs_scrub	*sc)
+{
+	struct xfs_bmbt_irec	mval[XFS_SYMLINK_MAPS];
+	struct xfs_inode	*ip = sc->ip;
+	struct xfs_buf		*bp;
+	char			*target_buf = sc->buf;
+	xfs_failaddr_t		fa;
+	xfs_filblks_t		fsblocks;
+	xfs_daddr_t		d;
+	loff_t			len;
+	loff_t			offset = 0;
+	unsigned int		byte_cnt;
+	bool			magic_ok;
+	bool			hdr_ok;
+	int			n;
+	int			nmaps = XFS_SYMLINK_MAPS;
+	int			error;
+
+	/* We'll only read until the buffer is full. */
+	len = min_t(loff_t, ip->i_disk_size, XFS_SYMLINK_MAXLEN);
+	fsblocks = xfs_symlink_blocks(sc->mp, len);
+	error = xfs_bmapi_read(ip, 0, fsblocks, mval, &nmaps, 0);
+	if (error)
+		return error;
+
+	for (n = 0; n < nmaps; n++) {
+		struct xfs_dsymlink_hdr	*dsl;
+
+		d = XFS_FSB_TO_DADDR(sc->mp, mval[n].br_startblock);
+
+		/* Read the rmt block.  We'll run the verifiers manually. */
+		error = xfs_trans_read_buf(sc->mp, sc->tp, sc->mp->m_ddev_targp,
+				d, XFS_FSB_TO_BB(sc->mp, mval[n].br_blockcount),
+				0, &bp, NULL);
+		if (error)
+			return error;
+		bp->b_ops = &xfs_symlink_buf_ops;
+
+		/* How many bytes do we expect to get out of this buffer? */
+		byte_cnt = XFS_FSB_TO_B(sc->mp, mval[n].br_blockcount);
+		byte_cnt = XFS_SYMLINK_BUF_SPACE(sc->mp, byte_cnt);
+		byte_cnt = min_t(unsigned int, byte_cnt, len);
+
+		/*
+		 * See if the verifiers accept this block.  We're willing to
+		 * salvage if the if the offset/byte/ino are ok and either the
+		 * verifier passed or the magic is ok.  Anything else and we
+		 * stop dead in our tracks.
+		 */
+		fa = bp->b_ops->verify_struct(bp);
+		dsl = bp->b_addr;
+		magic_ok = dsl->sl_magic == cpu_to_be32(XFS_SYMLINK_MAGIC);
+		hdr_ok = xfs_symlink_hdr_ok(ip->i_ino, offset, byte_cnt, bp);
+		if (!hdr_ok || (fa != NULL && !magic_ok))
+			break;
+
+		memcpy(target_buf + offset, dsl + 1, byte_cnt);
+
+		len -= byte_cnt;
+		offset += byte_cnt;
+	}
+	return offset;
+}
+
+/*
+ * Try to salvage an inline symlink's contents.  Returns the number of bytes
+ * salvaged or a negative errno.
+ */
+STATIC ssize_t
+xrep_symlink_salvage_inline(
+	struct xfs_scrub	*sc)
+{
+	struct xfs_inode	*ip = sc->ip;
+	char			*target_buf = sc->buf;
+	char			*old_target;
+	struct xfs_ifork	*ifp;
+	unsigned int		nr;
+
+	ifp = xfs_ifork_ptr(ip, XFS_DATA_FORK);
+	if (!ifp->if_data)
+		return 0;
+
+	/*
+	 * If inode repair zapped the link target, pretend that we didn't find
+	 * any bytes at all so that we can replace the (now totally lost) link
+	 * target with a warning message.
+	 */
+	old_target = ifp->if_data;
+	if (xfs_inode_has_sickness(sc->ip, XFS_SICK_INO_SYMLINK_ZAPPED) &&
+	    sc->ip->i_disk_size == 1 && old_target[0] == '?')
+		return 0;
+
+	nr = min(XFS_SYMLINK_MAXLEN, xfs_inode_data_fork_size(ip));
+	strncpy(target_buf, ifp->if_data, nr);
+	return nr;
+}
+
+#define DUMMY_TARGET \
+	"The target of this symbolic link could not be recovered at all and " \
+	"has been replaced with this explanatory message.  To avoid " \
+	"accidentally pointing to an existing file path, this message is " \
+	"longer than the maximum supported file name length.  That is an " \
+	"acceptable length for a symlink target on XFS but will produce " \
+	"File Name Too Long errors if resolved."
+
+/* Salvage whatever we can of the target. */
+STATIC int
+xrep_symlink_salvage(
+	struct xfs_scrub	*sc)
+{
+	char			*target_buf = sc->buf;
+	ssize_t			buflen = 0;
+
+	BUILD_BUG_ON(sizeof(DUMMY_TARGET) - 1 <= NAME_MAX);
+
+	/*
+	 * Salvage the target if there weren't any corruption problems observed
+	 * while scanning it.
+	 */
+	if (!(sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)) {
+		if (sc->ip->i_df.if_format == XFS_DINODE_FMT_LOCAL)
+			buflen = xrep_symlink_salvage_inline(sc);
+		else
+			buflen = xrep_symlink_salvage_remote(sc);
+		if (buflen < 0)
+			return buflen;
+
+		/*
+		 * NULL-terminate the buffer because the ondisk target does not
+		 * do that for us.  If salvage didn't find the exact amount of
+		 * data that we expected to find, don't salvage anything.
+		 */
+		target_buf[buflen] = 0;
+		if (strlen(target_buf) != sc->ip->i_disk_size)
+			buflen = 0;
+	}
+
+	/*
+	 * Change an empty target into a dummy target and clear the symlink
+	 * target zapped flag.
+	 */
+	if (buflen == 0) {
+		sc->sick_mask |= XFS_SICK_INO_SYMLINK_ZAPPED;
+		sprintf(target_buf, DUMMY_TARGET);
+	}
+
+	trace_xrep_symlink_salvage_target(sc->ip, target_buf,
+			strlen(target_buf));
+	return 0;
+}
+
+STATIC void
+xrep_symlink_local_to_remote(
+	struct xfs_trans	*tp,
+	struct xfs_buf		*bp,
+	struct xfs_inode	*ip,
+	struct xfs_ifork	*ifp,
+	void			*priv)
+{
+	struct xfs_scrub	*sc = priv;
+	struct xfs_dsymlink_hdr	*dsl = bp->b_addr;
+
+	xfs_symlink_local_to_remote(tp, bp, ip, ifp, NULL);
+
+	if (!xfs_has_crc(sc->mp))
+		return;
+
+	dsl->sl_owner = cpu_to_be64(sc->ip->i_ino);
+	xfs_trans_log_buf(tp, bp, 0,
+			  sizeof(struct xfs_dsymlink_hdr) + ifp->if_bytes - 1);
+}
+
+/*
+ * Prepare both links' data forks for an exchange.  Promote the tempfile from
+ * local format to extents format, and if the file being repaired has a short
+ * format data fork, turn it into an empty extent list.
+ */
+STATIC int
+xrep_symlink_swap_prep(
+	struct xfs_scrub	*sc,
+	bool			temp_local,
+	bool			ip_local)
+{
+	int			error;
+
+	/*
+	 * If the temp link is in shortform format, convert that to a remote
+	 * target so that we can use the atomic mapping exchange.
+	 */
+	if (temp_local) {
+		int		logflags = XFS_ILOG_CORE;
+
+		error = xfs_bmap_local_to_extents(sc->tp, sc->tempip, 1,
+				&logflags, XFS_DATA_FORK,
+				xrep_symlink_local_to_remote,
+				sc);
+		if (error)
+			return error;
+
+		xfs_trans_log_inode(sc->tp, sc->ip, 0);
+
+		error = xfs_defer_finish(&sc->tp);
+		if (error)
+			return error;
+	}
+
+	/*
+	 * If the file being repaired had a shortform data fork, convert that
+	 * to an empty extent list in preparation for the atomic mapping
+	 * exchange.
+	 */
+	if (ip_local) {
+		struct xfs_ifork	*ifp;
+
+		ifp = xfs_ifork_ptr(sc->ip, XFS_DATA_FORK);
+		xfs_idestroy_fork(ifp);
+		ifp->if_format = XFS_DINODE_FMT_EXTENTS;
+		ifp->if_nextents = 0;
+		ifp->if_bytes = 0;
+		ifp->if_data = NULL;
+		ifp->if_height = 0;
+
+		xfs_trans_log_inode(sc->tp, sc->ip,
+				XFS_ILOG_CORE | XFS_ILOG_DDATA);
+	}
+
+	return 0;
+}
+
+/* Exchange the temporary symlink's data fork with the one being repaired. */
+STATIC int
+xrep_symlink_swap(
+	struct xfs_scrub	*sc)
+{
+	struct xrep_tempexch	*tx = sc->buf;
+	bool			ip_local, temp_local;
+	int			error;
+
+	ip_local = sc->ip->i_df.if_format == XFS_DINODE_FMT_LOCAL;
+	temp_local = sc->tempip->i_df.if_format == XFS_DINODE_FMT_LOCAL;
+
+	/*
+	 * If the both links have a local format data fork and the rebuilt
+	 * remote data would fit in the repaired file's data fork, copy the
+	 * contents from the tempfile and declare ourselves done.
+	 */
+	if (ip_local && temp_local &&
+	    sc->tempip->i_disk_size <= xfs_inode_data_fork_size(sc->ip)) {
+		xrep_tempfile_copyout_local(sc, XFS_DATA_FORK);
+		return 0;
+	}
+
+	/* Otherwise, make sure both data forks are in block-mapping mode. */
+	error = xrep_symlink_swap_prep(sc, temp_local, ip_local);
+	if (error)
+		return error;
+
+	return xrep_tempexch_contents(sc, tx);
+}
+
+/*
+ * Free all the remote blocks and reset the data fork.  The caller must join
+ * the inode to the transaction.  This function returns with the inode joined
+ * to a clean scrub transaction.
+ */
+STATIC int
+xrep_symlink_reset_fork(
+	struct xfs_scrub	*sc)
+{
+	struct xfs_ifork	*ifp = xfs_ifork_ptr(sc->tempip, XFS_DATA_FORK);
+	int			error;
+
+	/* Unmap all the remote target buffers. */
+	if (xfs_ifork_has_extents(ifp)) {
+		error = xrep_reap_ifork(sc, sc->tempip, XFS_DATA_FORK);
+		if (error)
+			return error;
+	}
+
+	trace_xrep_symlink_reset_fork(sc->tempip);
+
+	/* Reset the temp symlink target to dummy content. */
+	xfs_idestroy_fork(ifp);
+	return xfs_symlink_write_target(sc->tp, sc->tempip, sc->tempip->i_ino,
+			"?", 1, 0, 0);
+}
+
+/*
+ * Reinitialize a link target.  Caller must ensure the inode is joined to
+ * the transaction.
+ */
+STATIC int
+xrep_symlink_rebuild(
+	struct xfs_scrub	*sc)
+{
+	struct xrep_tempexch	*tx;
+	char			*target_buf = sc->buf;
+	xfs_fsblock_t		fs_blocks;
+	unsigned int		target_len;
+	unsigned int		resblks;
+	int			error;
+
+	/* How many blocks do we need? */
+	target_len = strlen(target_buf);
+	ASSERT(target_len != 0);
+	if (target_len == 0 || target_len > XFS_SYMLINK_MAXLEN)
+		return -EFSCORRUPTED;
+
+	trace_xrep_symlink_rebuild(sc->ip);
+
+	/*
+	 * In preparation to write the new symlink target to the temporary
+	 * file, drop the ILOCK of the file being repaired (it shouldn't be
+	 * joined) and take the ILOCK of the temporary file.
+	 *
+	 * The VFS does not take the IOLOCK while reading a symlink (and new
+	 * symlinks are hidden with INEW until they've been written) so it's
+	 * possible that a readlink() could see the old corrupted contents
+	 * while we're doing this.
+	 */
+	xchk_iunlock(sc, XFS_ILOCK_EXCL);
+	xrep_tempfile_ilock(sc);
+	xfs_trans_ijoin(sc->tp, sc->tempip, 0);
+
+	/*
+	 * Reserve resources to reinitialize the target.  We're allowed to
+	 * exceed file quota to repair inconsistent metadata, though this is
+	 * unlikely.
+	 */
+	fs_blocks = xfs_symlink_blocks(sc->mp, target_len);
+	resblks = xfs_symlink_space_res(sc->mp, target_len, fs_blocks);
+	error = xfs_trans_reserve_quota_nblks(sc->tp, sc->tempip, resblks, 0,
+			true);
+	if (error)
+		return error;
+
+	/* Erase the dummy target set up by the tempfile initialization. */
+	xfs_idestroy_fork(&sc->tempip->i_df);
+	sc->tempip->i_df.if_bytes = 0;
+	sc->tempip->i_df.if_format = XFS_DINODE_FMT_EXTENTS;
+
+	/* Write the salvaged target to the temporary link. */
+	error = xfs_symlink_write_target(sc->tp, sc->tempip, sc->ip->i_ino,
+			target_buf, target_len, fs_blocks, resblks);
+	if (error)
+		return error;
+
+	/*
+	 * Commit the repair transaction so that we can use the atomic mapping
+	 * exchange functions to compute the correct block reservations and
+	 * re-lock the inodes.
+	 */
+	target_buf = NULL;
+	error = xrep_trans_commit(sc);
+	if (error)
+		return error;
+
+	/* Last chance to abort before we start committing fixes. */
+	if (xchk_should_terminate(sc, &error))
+		return error;
+
+	xrep_tempfile_iunlock(sc);
+
+	/*
+	 * We're done with the temporary buffer, so we can reuse it for the
+	 * tempfile contents exchange information.
+	 */
+	tx = sc->buf;
+	error = xrep_tempexch_trans_alloc(sc, XFS_DATA_FORK, tx);
+	if (error)
+		return error;
+
+	/*
+	 * Exchange the temp link's data fork with the file being repaired.
+	 * This recreates the transaction and takes the ILOCKs of the file
+	 * being repaired and the temporary file.
+	 */
+	error = xrep_symlink_swap(sc);
+	if (error)
+		return error;
+
+	/*
+	 * Release the old symlink blocks and reset the data fork of the temp
+	 * link to an empty shortform link.  This is the last repair action we
+	 * perform on the symlink, so we don't need to clean the transaction.
+	 */
+	return xrep_symlink_reset_fork(sc);
+}
+
+/* Repair a symbolic link. */
+int
+xrep_symlink(
+	struct xfs_scrub	*sc)
+{
+	int			error;
+
+	/* The rmapbt is required to reap the old data fork. */
+	if (!xfs_has_rmapbt(sc->mp))
+		return -EOPNOTSUPP;
+	/* We require atomic file exchange range to rebuild anything. */
+	if (!xfs_has_exchange_range(sc->mp))
+		return -EOPNOTSUPP;
+
+	ASSERT(sc->ilock_flags & XFS_ILOCK_EXCL);
+
+	error = xrep_symlink_salvage(sc);
+	if (error)
+		return error;
+
+	/* Now reset the target. */
+	error = xrep_symlink_rebuild(sc);
+	if (error)
+		return error;
+
+	return xrep_trans_commit(sc);
+}
diff --git a/fs/xfs/scrub/tempexch.h b/fs/xfs/scrub/tempexch.h
new file mode 100644
index 0000000000..995ba187c5
--- /dev/null
+++ b/fs/xfs/scrub/tempexch.h
@@ -0,0 +1,22 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2022-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_TEMPEXCH_H__
+#define __XFS_SCRUB_TEMPEXCH_H__
+
+#ifdef CONFIG_XFS_ONLINE_REPAIR
+struct xrep_tempexch {
+	struct xfs_exchmaps_req	req;
+};
+
+int xrep_tempexch_trans_reserve(struct xfs_scrub *sc, int whichfork,
+		struct xrep_tempexch *ti);
+int xrep_tempexch_trans_alloc(struct xfs_scrub *sc, int whichfork,
+		struct xrep_tempexch *ti);
+
+int xrep_tempexch_contents(struct xfs_scrub *sc, struct xrep_tempexch *ti);
+#endif /* CONFIG_XFS_ONLINE_REPAIR */
+
+#endif /* __XFS_SCRUB_TEMPEXCH_H__ */
diff --git a/fs/xfs/scrub/tempfile.c b/fs/xfs/scrub/tempfile.c
new file mode 100644
index 0000000000..b747b625c5
--- /dev/null
+++ b/fs/xfs/scrub/tempfile.c
@@ -0,0 +1,851 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2021-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_inode.h"
+#include "xfs_ialloc.h"
+#include "xfs_quota.h"
+#include "xfs_bmap.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_trans_space.h"
+#include "xfs_dir2.h"
+#include "xfs_exchrange.h"
+#include "xfs_exchmaps.h"
+#include "xfs_defer.h"
+#include "xfs_symlink_remote.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/repair.h"
+#include "scrub/trace.h"
+#include "scrub/tempfile.h"
+#include "scrub/tempexch.h"
+#include "scrub/xfile.h"
+
+/*
+ * Create a temporary file for reconstructing metadata, with the intention of
+ * atomically exchanging the temporary file's contents with the file that's
+ * being repaired.
+ */
+int
+xrep_tempfile_create(
+	struct xfs_scrub	*sc,
+	uint16_t		mode)
+{
+	struct xfs_mount	*mp = sc->mp;
+	struct xfs_trans	*tp = NULL;
+	struct xfs_dquot	*udqp = NULL;
+	struct xfs_dquot	*gdqp = NULL;
+	struct xfs_dquot	*pdqp = NULL;
+	struct xfs_trans_res	*tres;
+	struct xfs_inode	*dp = mp->m_rootip;
+	xfs_ino_t		ino;
+	unsigned int		resblks;
+	bool			is_dir = S_ISDIR(mode);
+	int			error;
+
+	if (xfs_is_shutdown(mp))
+		return -EIO;
+	if (xfs_is_readonly(mp))
+		return -EROFS;
+
+	ASSERT(sc->tp == NULL);
+	ASSERT(sc->tempip == NULL);
+
+	/*
+	 * Make sure that we have allocated dquot(s) on disk.  The temporary
+	 * inode should be completely root owned so that we don't fail due to
+	 * quota limits.
+	 */
+	error = xfs_qm_vop_dqalloc(dp, GLOBAL_ROOT_UID, GLOBAL_ROOT_GID, 0,
+			XFS_QMOPT_QUOTALL, &udqp, &gdqp, &pdqp);
+	if (error)
+		return error;
+
+	if (is_dir) {
+		resblks = xfs_mkdir_space_res(mp, 0);
+		tres = &M_RES(mp)->tr_mkdir;
+	} else {
+		resblks = XFS_IALLOC_SPACE_RES(mp);
+		tres = &M_RES(mp)->tr_create_tmpfile;
+	}
+
+	error = xfs_trans_alloc_icreate(mp, tres, udqp, gdqp, pdqp, resblks,
+			&tp);
+	if (error)
+		goto out_release_dquots;
+
+	/* Allocate inode, set up directory. */
+	error = xfs_dialloc(&tp, dp->i_ino, mode, &ino);
+	if (error)
+		goto out_trans_cancel;
+	error = xfs_init_new_inode(&nop_mnt_idmap, tp, dp, ino, mode, 0, 0,
+			0, false, &sc->tempip);
+	if (error)
+		goto out_trans_cancel;
+
+	/* Change the ownership of the inode to root. */
+	VFS_I(sc->tempip)->i_uid = GLOBAL_ROOT_UID;
+	VFS_I(sc->tempip)->i_gid = GLOBAL_ROOT_GID;
+	sc->tempip->i_diflags &= ~(XFS_DIFLAG_REALTIME | XFS_DIFLAG_RTINHERIT);
+	xfs_trans_log_inode(tp, sc->tempip, XFS_ILOG_CORE);
+
+	/*
+	 * Mark our temporary file as private so that LSMs and the ACL code
+	 * don't try to add their own metadata or reason about these files.
+	 * The file should never be exposed to userspace.
+	 */
+	VFS_I(sc->tempip)->i_flags |= S_PRIVATE;
+	VFS_I(sc->tempip)->i_opflags &= ~IOP_XATTR;
+
+	if (is_dir) {
+		error = xfs_dir_init(tp, sc->tempip, dp);
+		if (error)
+			goto out_trans_cancel;
+	} else if (S_ISLNK(VFS_I(sc->tempip)->i_mode)) {
+		/*
+		 * Initialize the temporary symlink with a meaningless target
+		 * that won't trip the verifiers.  Repair must rewrite the
+		 * target with meaningful content before swapping with the file
+		 * being repaired.  A single-byte target will not write a
+		 * remote target block, so the owner is irrelevant.
+		 */
+		error = xfs_symlink_write_target(tp, sc->tempip,
+				sc->tempip->i_ino, ".", 1, 0, 0);
+		if (error)
+			goto out_trans_cancel;
+	}
+
+	/*
+	 * Attach the dquot(s) to the inodes and modify them incore.
+	 * These ids of the inode couldn't have changed since the new
+	 * inode has been locked ever since it was created.
+	 */
+	xfs_qm_vop_create_dqattach(tp, sc->tempip, udqp, gdqp, pdqp);
+
+	/*
+	 * Put our temp file on the unlinked list so it's purged automatically.
+	 * All file-based metadata being reconstructed using this file must be
+	 * atomically exchanged with the original file because the contents
+	 * here will be purged when the inode is dropped or log recovery cleans
+	 * out the unlinked list.
+	 */
+	error = xfs_iunlink(tp, sc->tempip);
+	if (error)
+		goto out_trans_cancel;
+
+	error = xfs_trans_commit(tp);
+	if (error)
+		goto out_release_inode;
+
+	trace_xrep_tempfile_create(sc);
+
+	xfs_qm_dqrele(udqp);
+	xfs_qm_dqrele(gdqp);
+	xfs_qm_dqrele(pdqp);
+
+	/* Finish setting up the incore / vfs context. */
+	xfs_iunlock(sc->tempip, XFS_ILOCK_EXCL);
+	xfs_setup_iops(sc->tempip);
+	xfs_finish_inode_setup(sc->tempip);
+
+	sc->temp_ilock_flags = 0;
+	return error;
+
+out_trans_cancel:
+	xfs_trans_cancel(tp);
+out_release_inode:
+	/*
+	 * Wait until after the current transaction is aborted to finish the
+	 * setup of the inode and release the inode.  This prevents recursive
+	 * transactions and deadlocks from xfs_inactive.
+	 */
+	if (sc->tempip) {
+		xfs_iunlock(sc->tempip, XFS_ILOCK_EXCL);
+		xfs_finish_inode_setup(sc->tempip);
+		xchk_irele(sc, sc->tempip);
+	}
+out_release_dquots:
+	xfs_qm_dqrele(udqp);
+	xfs_qm_dqrele(gdqp);
+	xfs_qm_dqrele(pdqp);
+
+	return error;
+}
+
+/* Take IOLOCK_EXCL on the temporary file, maybe. */
+bool
+xrep_tempfile_iolock_nowait(
+	struct xfs_scrub	*sc)
+{
+	if (xfs_ilock_nowait(sc->tempip, XFS_IOLOCK_EXCL)) {
+		sc->temp_ilock_flags |= XFS_IOLOCK_EXCL;
+		return true;
+	}
+
+	return false;
+}
+
+/*
+ * Take the temporary file's IOLOCK while holding a different inode's IOLOCK.
+ * In theory nobody else should hold the tempfile's IOLOCK, but we use trylock
+ * to avoid deadlocks and lockdep complaints.
+ */
+int
+xrep_tempfile_iolock_polled(
+	struct xfs_scrub	*sc)
+{
+	int			error = 0;
+
+	while (!xrep_tempfile_iolock_nowait(sc)) {
+		if (xchk_should_terminate(sc, &error))
+			return error;
+		delay(1);
+	}
+
+	return 0;
+}
+
+/* Release IOLOCK_EXCL on the temporary file. */
+void
+xrep_tempfile_iounlock(
+	struct xfs_scrub	*sc)
+{
+	xfs_iunlock(sc->tempip, XFS_IOLOCK_EXCL);
+	sc->temp_ilock_flags &= ~XFS_IOLOCK_EXCL;
+}
+
+/* Prepare the temporary file for metadata updates by grabbing ILOCK_EXCL. */
+void
+xrep_tempfile_ilock(
+	struct xfs_scrub	*sc)
+{
+	sc->temp_ilock_flags |= XFS_ILOCK_EXCL;
+	xfs_ilock(sc->tempip, XFS_ILOCK_EXCL);
+}
+
+/* Try to grab ILOCK_EXCL on the temporary file. */
+bool
+xrep_tempfile_ilock_nowait(
+	struct xfs_scrub	*sc)
+{
+	if (xfs_ilock_nowait(sc->tempip, XFS_ILOCK_EXCL)) {
+		sc->temp_ilock_flags |= XFS_ILOCK_EXCL;
+		return true;
+	}
+
+	return false;
+}
+
+/* Unlock ILOCK_EXCL on the temporary file after an update. */
+void
+xrep_tempfile_iunlock(
+	struct xfs_scrub	*sc)
+{
+	xfs_iunlock(sc->tempip, XFS_ILOCK_EXCL);
+	sc->temp_ilock_flags &= ~XFS_ILOCK_EXCL;
+}
+
+/*
+ * Begin the process of making changes to both the file being scrubbed and
+ * the temporary file by taking ILOCK_EXCL on both.
+ */
+void
+xrep_tempfile_ilock_both(
+	struct xfs_scrub	*sc)
+{
+	xfs_lock_two_inodes(sc->ip, XFS_ILOCK_EXCL, sc->tempip, XFS_ILOCK_EXCL);
+	sc->ilock_flags |= XFS_ILOCK_EXCL;
+	sc->temp_ilock_flags |= XFS_ILOCK_EXCL;
+}
+
+/* Unlock ILOCK_EXCL on both files. */
+void
+xrep_tempfile_iunlock_both(
+	struct xfs_scrub	*sc)
+{
+	xrep_tempfile_iunlock(sc);
+	xchk_iunlock(sc, XFS_ILOCK_EXCL);
+}
+
+/* Release the temporary file. */
+void
+xrep_tempfile_rele(
+	struct xfs_scrub	*sc)
+{
+	if (!sc->tempip)
+		return;
+
+	if (sc->temp_ilock_flags) {
+		xfs_iunlock(sc->tempip, sc->temp_ilock_flags);
+		sc->temp_ilock_flags = 0;
+	}
+
+	xchk_irele(sc, sc->tempip);
+	sc->tempip = NULL;
+}
+
+/*
+ * Make sure that the given range of the data fork of the temporary file is
+ * mapped to written blocks.  The caller must ensure that both inodes are
+ * joined to the transaction.
+ */
+int
+xrep_tempfile_prealloc(
+	struct xfs_scrub	*sc,
+	xfs_fileoff_t		off,
+	xfs_filblks_t		len)
+{
+	struct xfs_bmbt_irec	map;
+	xfs_fileoff_t		end = off + len;
+	int			error;
+
+	ASSERT(sc->tempip != NULL);
+	ASSERT(!XFS_NOT_DQATTACHED(sc->mp, sc->tempip));
+
+	for (; off < end; off = map.br_startoff + map.br_blockcount) {
+		int		nmaps = 1;
+
+		/*
+		 * If we have a real extent mapping this block then we're
+		 * in ok shape.
+		 */
+		error = xfs_bmapi_read(sc->tempip, off, end - off, &map, &nmaps,
+				XFS_DATA_FORK);
+		if (error)
+			return error;
+		if (nmaps == 0) {
+			ASSERT(nmaps != 0);
+			return -EFSCORRUPTED;
+		}
+
+		if (xfs_bmap_is_written_extent(&map))
+			continue;
+
+		/*
+		 * If we find a delalloc reservation then something is very
+		 * very wrong.  Bail out.
+		 */
+		if (map.br_startblock == DELAYSTARTBLOCK)
+			return -EFSCORRUPTED;
+
+		/*
+		 * Make sure this block has a real zeroed extent allocated to
+		 * it.
+		 */
+		nmaps = 1;
+		error = xfs_bmapi_write(sc->tp, sc->tempip, off, end - off,
+				XFS_BMAPI_CONVERT | XFS_BMAPI_ZERO, 0, &map,
+				&nmaps);
+		if (error)
+			return error;
+		if (nmaps != 1)
+			return -EFSCORRUPTED;
+
+		trace_xrep_tempfile_prealloc(sc, XFS_DATA_FORK, &map);
+
+		/* Commit new extent and all deferred work. */
+		error = xfs_defer_finish(&sc->tp);
+		if (error)
+			return error;
+	}
+
+	return 0;
+}
+
+/*
+ * Write data to each block of a file.  The given range of the tempfile's data
+ * fork must already be populated with written extents.
+ */
+int
+xrep_tempfile_copyin(
+	struct xfs_scrub	*sc,
+	xfs_fileoff_t		off,
+	xfs_filblks_t		len,
+	xrep_tempfile_copyin_fn	prep_fn,
+	void			*data)
+{
+	LIST_HEAD(buffers_list);
+	struct xfs_mount	*mp = sc->mp;
+	struct xfs_buf		*bp;
+	xfs_fileoff_t		flush_mask;
+	xfs_fileoff_t		end = off + len;
+	loff_t			pos = XFS_FSB_TO_B(mp, off);
+	int			error = 0;
+
+	ASSERT(S_ISREG(VFS_I(sc->tempip)->i_mode));
+
+	/* Flush buffers to disk every 512K */
+	flush_mask = XFS_B_TO_FSBT(mp, (1U << 19)) - 1;
+
+	for (; off < end; off++, pos += mp->m_sb.sb_blocksize) {
+		struct xfs_bmbt_irec	map;
+		int			nmaps = 1;
+
+		/* Read block mapping for this file block. */
+		error = xfs_bmapi_read(sc->tempip, off, 1, &map, &nmaps, 0);
+		if (error)
+			goto out_err;
+		if (nmaps == 0 || !xfs_bmap_is_written_extent(&map)) {
+			error = -EFSCORRUPTED;
+			goto out_err;
+		}
+
+		/* Get the metadata buffer for this offset in the file. */
+		error = xfs_trans_get_buf(sc->tp, mp->m_ddev_targp,
+				XFS_FSB_TO_DADDR(mp, map.br_startblock),
+				mp->m_bsize, 0, &bp);
+		if (error)
+			goto out_err;
+
+		trace_xrep_tempfile_copyin(sc, XFS_DATA_FORK, &map);
+
+		/* Read in a block's worth of data from the xfile. */
+		error = prep_fn(sc, bp, data);
+		if (error) {
+			xfs_trans_brelse(sc->tp, bp);
+			goto out_err;
+		}
+
+		/* Queue buffer, and flush if we have too much dirty data. */
+		xfs_buf_delwri_queue_here(bp, &buffers_list);
+		xfs_trans_brelse(sc->tp, bp);
+
+		if (!(off & flush_mask)) {
+			error = xfs_buf_delwri_submit(&buffers_list);
+			if (error)
+				goto out_err;
+		}
+	}
+
+	/*
+	 * Write the new blocks to disk.  If the ordered list isn't empty after
+	 * that, then something went wrong and we have to fail.  This should
+	 * never happen, but we'll check anyway.
+	 */
+	error = xfs_buf_delwri_submit(&buffers_list);
+	if (error)
+		goto out_err;
+
+	if (!list_empty(&buffers_list)) {
+		ASSERT(list_empty(&buffers_list));
+		error = -EIO;
+		goto out_err;
+	}
+
+	return 0;
+
+out_err:
+	xfs_buf_delwri_cancel(&buffers_list);
+	return error;
+}
+
+/*
+ * Set the temporary file's size.  Caller must join the tempfile to the scrub
+ * transaction and is responsible for adjusting block mappings as needed.
+ */
+int
+xrep_tempfile_set_isize(
+	struct xfs_scrub	*sc,
+	unsigned long long	isize)
+{
+	if (sc->tempip->i_disk_size == isize)
+		return 0;
+
+	sc->tempip->i_disk_size = isize;
+	i_size_write(VFS_I(sc->tempip), isize);
+	return xrep_tempfile_roll_trans(sc);
+}
+
+/*
+ * Roll a repair transaction involving the temporary file.  Caller must join
+ * both the temporary file and the file being scrubbed to the transaction.
+ * This function return with both inodes joined to a new scrub transaction,
+ * or the usual negative errno.
+ */
+int
+xrep_tempfile_roll_trans(
+	struct xfs_scrub	*sc)
+{
+	int			error;
+
+	xfs_trans_log_inode(sc->tp, sc->tempip, XFS_ILOG_CORE);
+	error = xrep_roll_trans(sc);
+	if (error)
+		return error;
+
+	xfs_trans_ijoin(sc->tp, sc->tempip, 0);
+	return 0;
+}
+
+/*
+ * Fill out the mapping exchange request in preparation for atomically
+ * committing the contents of a metadata file that we've rebuilt in the temp
+ * file.
+ */
+STATIC int
+xrep_tempexch_prep_request(
+	struct xfs_scrub	*sc,
+	int			whichfork,
+	struct xrep_tempexch	*tx)
+{
+	struct xfs_exchmaps_req	*req = &tx->req;
+
+	memset(tx, 0, sizeof(struct xrep_tempexch));
+
+	/* COW forks don't exist on disk. */
+	if (whichfork == XFS_COW_FORK) {
+		ASSERT(0);
+		return -EINVAL;
+	}
+
+	/* Both files should have the relevant forks. */
+	if (!xfs_ifork_ptr(sc->ip, whichfork) ||
+	    !xfs_ifork_ptr(sc->tempip, whichfork)) {
+		ASSERT(xfs_ifork_ptr(sc->ip, whichfork) != NULL);
+		ASSERT(xfs_ifork_ptr(sc->tempip, whichfork) != NULL);
+		return -EINVAL;
+	}
+
+	/* Exchange all mappings in both forks. */
+	req->ip1 = sc->tempip;
+	req->ip2 = sc->ip;
+	req->startoff1 = 0;
+	req->startoff2 = 0;
+	switch (whichfork) {
+	case XFS_ATTR_FORK:
+		req->flags |= XFS_EXCHMAPS_ATTR_FORK;
+		break;
+	case XFS_DATA_FORK:
+		/* Always exchange sizes when exchanging data fork mappings. */
+		req->flags |= XFS_EXCHMAPS_SET_SIZES;
+		break;
+	}
+	req->blockcount = XFS_MAX_FILEOFF;
+
+	return 0;
+}
+
+/*
+ * Fill out the mapping exchange resource estimation structures in preparation
+ * for exchanging the contents of a metadata file that we've rebuilt in the
+ * temp file.  Caller must hold IOLOCK_EXCL but not ILOCK_EXCL on both files.
+ */
+STATIC int
+xrep_tempexch_estimate(
+	struct xfs_scrub	*sc,
+	struct xrep_tempexch	*tx)
+{
+	struct xfs_exchmaps_req	*req = &tx->req;
+	struct xfs_ifork	*ifp;
+	struct xfs_ifork	*tifp;
+	int			whichfork = xfs_exchmaps_reqfork(req);
+	int			state = 0;
+
+	/*
+	 * The exchmaps code only knows how to exchange file fork space
+	 * mappings.  Any fork data in local format must be promoted to a
+	 * single block before the exchange can take place.
+	 */
+	ifp = xfs_ifork_ptr(sc->ip, whichfork);
+	if (ifp->if_format == XFS_DINODE_FMT_LOCAL)
+		state |= 1;
+
+	tifp = xfs_ifork_ptr(sc->tempip, whichfork);
+	if (tifp->if_format == XFS_DINODE_FMT_LOCAL)
+		state |= 2;
+
+	switch (state) {
+	case 0:
+		/* Both files have mapped extents; use the regular estimate. */
+		return xfs_exchrange_estimate(req);
+	case 1:
+		/*
+		 * The file being repaired is in local format, but the temp
+		 * file has mapped extents.  To perform the exchange, the file
+		 * being repaired must have its shorform data converted to an
+		 * ondisk block so that the forks will be in extents format.
+		 * We need one resblk for the conversion; the number of
+		 * exchanges is (worst case) the temporary file's extent count
+		 * plus the block we converted.
+		 */
+		req->ip1_bcount = sc->tempip->i_nblocks;
+		req->ip2_bcount = 1;
+		req->nr_exchanges = 1 + tifp->if_nextents;
+		req->resblks = 1;
+		break;
+	case 2:
+		/*
+		 * The temporary file is in local format, but the file being
+		 * repaired has mapped extents.  To perform the exchange, the
+		 * temp file must have its shortform data converted to an
+		 * ondisk block, and the fork changed to extents format.  We
+		 * need one resblk for the conversion; the number of exchanges
+		 * is (worst case) the extent count of the file being repaired
+		 * plus the block we converted.
+		 */
+		req->ip1_bcount = 1;
+		req->ip2_bcount = sc->ip->i_nblocks;
+		req->nr_exchanges = 1 + ifp->if_nextents;
+		req->resblks = 1;
+		break;
+	case 3:
+		/*
+		 * Both forks are in local format.  To perform the exchange,
+		 * both files must have their shortform data converted to
+		 * fsblocks, and both forks must be converted to extents
+		 * format.  We need two resblks for the two conversions, and
+		 * the number of exchanges is 1 since there's only one block at
+		 * fileoff 0.  Presumably, the caller could not exchange the
+		 * two inode fork areas directly.
+		 */
+		req->ip1_bcount = 1;
+		req->ip2_bcount = 1;
+		req->nr_exchanges = 1;
+		req->resblks = 2;
+		break;
+	}
+
+	return xfs_exchmaps_estimate_overhead(req);
+}
+
+/*
+ * Obtain a quota reservation to make sure we don't hit EDQUOT.  We can skip
+ * this if quota enforcement is disabled or if both inodes' dquots are the
+ * same.  The qretry structure must be initialized to zeroes before the first
+ * call to this function.
+ */
+STATIC int
+xrep_tempexch_reserve_quota(
+	struct xfs_scrub		*sc,
+	const struct xrep_tempexch	*tx)
+{
+	struct xfs_trans		*tp = sc->tp;
+	const struct xfs_exchmaps_req	*req = &tx->req;
+	int64_t				ddelta, rdelta;
+	int				error;
+
+	/*
+	 * Don't bother with a quota reservation if we're not enforcing them
+	 * or the two inodes have the same dquots.
+	 */
+	if (!XFS_IS_QUOTA_ON(tp->t_mountp) || req->ip1 == req->ip2 ||
+	    (req->ip1->i_udquot == req->ip2->i_udquot &&
+	     req->ip1->i_gdquot == req->ip2->i_gdquot &&
+	     req->ip1->i_pdquot == req->ip2->i_pdquot))
+		return 0;
+
+	/*
+	 * Quota reservation for each file comes from two sources.  First, we
+	 * need to account for any net gain in mapped blocks during the
+	 * exchange.  Second, we need reservation for the gross gain in mapped
+	 * blocks so that we don't trip over any quota block reservation
+	 * assertions.  We must reserve the gross gain because the quota code
+	 * subtracts from bcount the number of blocks that we unmap; it does
+	 * not add that quantity back to the quota block reservation.
+	 */
+	ddelta = max_t(int64_t, 0, req->ip2_bcount - req->ip1_bcount);
+	rdelta = max_t(int64_t, 0, req->ip2_rtbcount - req->ip1_rtbcount);
+	error = xfs_trans_reserve_quota_nblks(tp, req->ip1,
+			ddelta + req->ip1_bcount, rdelta + req->ip1_rtbcount,
+			true);
+	if (error)
+		return error;
+
+	ddelta = max_t(int64_t, 0, req->ip1_bcount - req->ip2_bcount);
+	rdelta = max_t(int64_t, 0, req->ip1_rtbcount - req->ip2_rtbcount);
+	return xfs_trans_reserve_quota_nblks(tp, req->ip2,
+			ddelta + req->ip2_bcount, rdelta + req->ip2_rtbcount,
+			true);
+}
+
+/*
+ * Prepare an existing transaction for an atomic file contents exchange.
+ *
+ * This function fills out the mapping exchange request and resource estimation
+ * structures in preparation for exchanging the contents of a metadata file
+ * that has been rebuilt in the temp file.  Next, it reserves space and quota
+ * for the transaction.
+ *
+ * The caller must hold ILOCK_EXCL of the scrub target file and the temporary
+ * file.  The caller must join both inodes to the transaction with no unlock
+ * flags, and is responsible for dropping both ILOCKs when appropriate.  Only
+ * use this when those ILOCKs cannot be dropped.
+ */
+int
+xrep_tempexch_trans_reserve(
+	struct xfs_scrub	*sc,
+	int			whichfork,
+	struct xrep_tempexch	*tx)
+{
+	int			error;
+
+	ASSERT(sc->tp != NULL);
+	xfs_assert_ilocked(sc->ip, XFS_ILOCK_EXCL);
+	xfs_assert_ilocked(sc->tempip, XFS_ILOCK_EXCL);
+
+	error = xrep_tempexch_prep_request(sc, whichfork, tx);
+	if (error)
+		return error;
+
+	error = xfs_exchmaps_estimate(&tx->req);
+	if (error)
+		return error;
+
+	error = xfs_trans_reserve_more(sc->tp, tx->req.resblks, 0);
+	if (error)
+		return error;
+
+	return xrep_tempexch_reserve_quota(sc, tx);
+}
+
+/*
+ * Create a new transaction for a file contents exchange.
+ *
+ * This function fills out the mapping excahange request and resource
+ * estimation structures in preparation for exchanging the contents of a
+ * metadata file that has been rebuilt in the temp file.  Next, it reserves
+ * space, takes ILOCK_EXCL of both inodes, joins them to the transaction and
+ * reserves quota for the transaction.
+ *
+ * The caller is responsible for dropping both ILOCKs when appropriate.
+ */
+int
+xrep_tempexch_trans_alloc(
+	struct xfs_scrub	*sc,
+	int			whichfork,
+	struct xrep_tempexch	*tx)
+{
+	unsigned int		flags = 0;
+	int			error;
+
+	ASSERT(sc->tp == NULL);
+	ASSERT(xfs_has_exchange_range(sc->mp));
+
+	error = xrep_tempexch_prep_request(sc, whichfork, tx);
+	if (error)
+		return error;
+
+	error = xrep_tempexch_estimate(sc, tx);
+	if (error)
+		return error;
+
+	if (xfs_has_lazysbcount(sc->mp))
+		flags |= XFS_TRANS_RES_FDBLKS;
+
+	error = xfs_trans_alloc(sc->mp, &M_RES(sc->mp)->tr_itruncate,
+			tx->req.resblks, 0, flags, &sc->tp);
+	if (error)
+		return error;
+
+	sc->temp_ilock_flags |= XFS_ILOCK_EXCL;
+	sc->ilock_flags |= XFS_ILOCK_EXCL;
+	xfs_exchrange_ilock(sc->tp, sc->ip, sc->tempip);
+
+	return xrep_tempexch_reserve_quota(sc, tx);
+}
+
+/*
+ * Exchange file mappings (and hence file contents) between the file being
+ * repaired and the temporary file.  Returns with both inodes locked and joined
+ * to a clean scrub transaction.
+ */
+int
+xrep_tempexch_contents(
+	struct xfs_scrub	*sc,
+	struct xrep_tempexch	*tx)
+{
+	int			error;
+
+	ASSERT(xfs_has_exchange_range(sc->mp));
+
+	xfs_exchange_mappings(sc->tp, &tx->req);
+	error = xfs_defer_finish(&sc->tp);
+	if (error)
+		return error;
+
+	/*
+	 * If we exchanged the ondisk sizes of two metadata files, we must
+	 * exchanged the incore sizes as well.
+	 */
+	if (tx->req.flags & XFS_EXCHMAPS_SET_SIZES) {
+		loff_t	temp;
+
+		temp = i_size_read(VFS_I(sc->ip));
+		i_size_write(VFS_I(sc->ip), i_size_read(VFS_I(sc->tempip)));
+		i_size_write(VFS_I(sc->tempip), temp);
+	}
+
+	return 0;
+}
+
+/*
+ * Write local format data from one of the temporary file's forks into the same
+ * fork of file being repaired, and exchange the file sizes, if appropriate.
+ * Caller must ensure that the file being repaired has enough fork space to
+ * hold all the bytes.
+ */
+void
+xrep_tempfile_copyout_local(
+	struct xfs_scrub	*sc,
+	int			whichfork)
+{
+	struct xfs_ifork	*temp_ifp;
+	struct xfs_ifork	*ifp;
+	unsigned int		ilog_flags = XFS_ILOG_CORE;
+
+	temp_ifp = xfs_ifork_ptr(sc->tempip, whichfork);
+	ifp = xfs_ifork_ptr(sc->ip, whichfork);
+
+	ASSERT(temp_ifp != NULL);
+	ASSERT(ifp != NULL);
+	ASSERT(temp_ifp->if_format == XFS_DINODE_FMT_LOCAL);
+	ASSERT(ifp->if_format == XFS_DINODE_FMT_LOCAL);
+
+	switch (whichfork) {
+	case XFS_DATA_FORK:
+		ASSERT(sc->tempip->i_disk_size <=
+					xfs_inode_data_fork_size(sc->ip));
+		break;
+	case XFS_ATTR_FORK:
+		ASSERT(sc->tempip->i_forkoff >= sc->ip->i_forkoff);
+		break;
+	default:
+		ASSERT(0);
+		return;
+	}
+
+	/* Recreate @sc->ip's incore fork (ifp) with data from temp_ifp. */
+	xfs_idestroy_fork(ifp);
+	xfs_init_local_fork(sc->ip, whichfork, temp_ifp->if_data,
+			temp_ifp->if_bytes);
+
+	if (whichfork == XFS_DATA_FORK) {
+		i_size_write(VFS_I(sc->ip), i_size_read(VFS_I(sc->tempip)));
+		sc->ip->i_disk_size = sc->tempip->i_disk_size;
+	}
+
+	ilog_flags |= xfs_ilog_fdata(whichfork);
+	xfs_trans_log_inode(sc->tp, sc->ip, ilog_flags);
+}
+
+/* Decide if a given XFS inode is a temporary file for a repair. */
+bool
+xrep_is_tempfile(
+	const struct xfs_inode	*ip)
+{
+	const struct inode	*inode = &ip->i_vnode;
+
+	if (IS_PRIVATE(inode) && !(inode->i_opflags & IOP_XATTR))
+		return true;
+
+	return false;
+}
diff --git a/fs/xfs/scrub/tempfile.h b/fs/xfs/scrub/tempfile.h
new file mode 100644
index 0000000000..e51399f595
--- /dev/null
+++ b/fs/xfs/scrub/tempfile.h
@@ -0,0 +1,48 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2021-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_TEMPFILE_H__
+#define __XFS_SCRUB_TEMPFILE_H__
+
+#ifdef CONFIG_XFS_ONLINE_REPAIR
+int xrep_tempfile_create(struct xfs_scrub *sc, uint16_t mode);
+void xrep_tempfile_rele(struct xfs_scrub *sc);
+
+bool xrep_tempfile_iolock_nowait(struct xfs_scrub *sc);
+int xrep_tempfile_iolock_polled(struct xfs_scrub *sc);
+void xrep_tempfile_iounlock(struct xfs_scrub *sc);
+
+void xrep_tempfile_ilock(struct xfs_scrub *sc);
+bool xrep_tempfile_ilock_nowait(struct xfs_scrub *sc);
+void xrep_tempfile_iunlock(struct xfs_scrub *sc);
+void xrep_tempfile_iunlock_both(struct xfs_scrub *sc);
+void xrep_tempfile_ilock_both(struct xfs_scrub *sc);
+
+int xrep_tempfile_prealloc(struct xfs_scrub *sc, xfs_fileoff_t off,
+		xfs_filblks_t len);
+
+enum xfs_blft;
+
+typedef int (*xrep_tempfile_copyin_fn)(struct xfs_scrub *sc,
+		struct xfs_buf *bp, void *data);
+
+int xrep_tempfile_copyin(struct xfs_scrub *sc, xfs_fileoff_t off,
+		xfs_filblks_t len, xrep_tempfile_copyin_fn fn, void *data);
+
+int xrep_tempfile_set_isize(struct xfs_scrub *sc, unsigned long long isize);
+
+int xrep_tempfile_roll_trans(struct xfs_scrub *sc);
+void xrep_tempfile_copyout_local(struct xfs_scrub *sc, int whichfork);
+bool xrep_is_tempfile(const struct xfs_inode *ip);
+#else
+static inline void xrep_tempfile_iolock_both(struct xfs_scrub *sc)
+{
+	xchk_ilock(sc, XFS_IOLOCK_EXCL);
+}
+# define xrep_is_tempfile(ip)		(false)
+# define xrep_tempfile_rele(sc)
+#endif /* CONFIG_XFS_ONLINE_REPAIR */
+
+#endif /* __XFS_SCRUB_TEMPFILE_H__ */
diff --git a/fs/xfs/scrub/trace.c b/fs/xfs/scrub/trace.c
index d0e24ffaf7..4470ad0533 100644
--- a/fs/xfs/scrub/trace.c
+++ b/fs/xfs/scrub/trace.c
@@ -16,10 +16,22 @@
 #include "xfs_rtbitmap.h"
 #include "xfs_quota.h"
 #include "xfs_quota_defs.h"
+#include "xfs_da_format.h"
+#include "xfs_dir2.h"
+#include "xfs_rmap.h"
+#include "xfs_parent.h"
 #include "scrub/scrub.h"
 #include "scrub/xfile.h"
 #include "scrub/xfarray.h"
 #include "scrub/quota.h"
+#include "scrub/iscan.h"
+#include "scrub/orphanage.h"
+#include "scrub/nlinks.h"
+#include "scrub/fscounters.h"
+#include "scrub/bitmap.h"
+#include "scrub/ino_bitmap.h"
+#include "scrub/xfblob.h"
+#include "scrub/dirtree.h"
 
 /* Figure out which block the btree cursor was pointing to. */
 static inline xfs_fsblock_t
@@ -32,7 +44,7 @@ xchk_btree_cur_fsbno(
 				xfs_buf_daddr(cur->bc_levels[level].bp));
 
 	if (level == cur->bc_nlevels - 1 &&
-	    (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE))
+	    cur->bc_ops->type == XFS_BTREE_TYPE_INODE)
 		return XFS_INO_TO_FSB(cur->bc_mp, cur->bc_ino.ip->i_ino);
 
 	return NULLFSBLOCK;
diff --git a/fs/xfs/scrub/trace.h b/fs/xfs/scrub/trace.h
index 6bbb4e8639..92ef4cdc48 100644
--- a/fs/xfs/scrub/trace.h
+++ b/fs/xfs/scrub/trace.h
@@ -15,11 +15,21 @@
 
 #include <linux/tracepoint.h>
 #include "xfs_bit.h"
+#include "xfs_quota_defs.h"
 
+struct xfs_scrub;
 struct xfile;
 struct xfarray;
 struct xfarray_sortinfo;
 struct xchk_dqiter;
+struct xchk_iscan;
+struct xchk_nlink;
+struct xchk_fscounters;
+struct xfs_rmap_update_params;
+struct xfs_parent_rec;
+enum xchk_dirpath_outcome;
+struct xchk_dirtree;
+struct xchk_dirtree_outcomes;
 
 /*
  * ftrace's __print_symbolic requires that all enum values be wrapped in the
@@ -27,14 +37,6 @@ struct xchk_dqiter;
  * ring buffer.  Somehow this was only worth mentioning in the ftrace sample
  * code.
  */
-TRACE_DEFINE_ENUM(XFS_BTNUM_BNOi);
-TRACE_DEFINE_ENUM(XFS_BTNUM_CNTi);
-TRACE_DEFINE_ENUM(XFS_BTNUM_BMAPi);
-TRACE_DEFINE_ENUM(XFS_BTNUM_INOi);
-TRACE_DEFINE_ENUM(XFS_BTNUM_FINOi);
-TRACE_DEFINE_ENUM(XFS_BTNUM_RMAPi);
-TRACE_DEFINE_ENUM(XFS_BTNUM_REFCi);
-
 TRACE_DEFINE_ENUM(XFS_REFC_DOMAIN_SHARED);
 TRACE_DEFINE_ENUM(XFS_REFC_DOMAIN_COW);
 
@@ -63,6 +65,11 @@ TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_UQUOTA);
 TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_GQUOTA);
 TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_PQUOTA);
 TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_FSCOUNTERS);
+TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_QUOTACHECK);
+TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_NLINKS);
+TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_HEALTHY);
+TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_DIRTREE);
+TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_BARRIER);
 
 #define XFS_SCRUB_TYPE_STRINGS \
 	{ XFS_SCRUB_TYPE_PROBE,		"probe" }, \
@@ -89,7 +96,12 @@ TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_FSCOUNTERS);
 	{ XFS_SCRUB_TYPE_UQUOTA,	"usrquota" }, \
 	{ XFS_SCRUB_TYPE_GQUOTA,	"grpquota" }, \
 	{ XFS_SCRUB_TYPE_PQUOTA,	"prjquota" }, \
-	{ XFS_SCRUB_TYPE_FSCOUNTERS,	"fscounters" }
+	{ XFS_SCRUB_TYPE_FSCOUNTERS,	"fscounters" }, \
+	{ XFS_SCRUB_TYPE_QUOTACHECK,	"quotacheck" }, \
+	{ XFS_SCRUB_TYPE_NLINKS,	"nlinks" }, \
+	{ XFS_SCRUB_TYPE_HEALTHY,	"healthy" }, \
+	{ XFS_SCRUB_TYPE_DIRTREE,	"dirtree" }, \
+	{ XFS_SCRUB_TYPE_BARRIER,	"barrier" }
 
 #define XFS_SCRUB_FLAG_STRINGS \
 	{ XFS_SCRUB_IFLAG_REPAIR,		"repair" }, \
@@ -107,9 +119,21 @@ TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_FSCOUNTERS);
 	{ XCHK_HAVE_FREEZE_PROT,		"nofreeze" }, \
 	{ XCHK_FSGATES_DRAIN,			"fsgates_drain" }, \
 	{ XCHK_NEED_DRAIN,			"need_drain" }, \
+	{ XCHK_FSGATES_QUOTA,			"fsgates_quota" }, \
+	{ XCHK_FSGATES_DIRENTS,			"fsgates_dirents" }, \
+	{ XCHK_FSGATES_RMAP,			"fsgates_rmap" }, \
 	{ XREP_RESET_PERAG_RESV,		"reset_perag_resv" }, \
 	{ XREP_ALREADY_FIXED,			"already_fixed" }
 
+TRACE_DEFINE_ENUM(XFS_RMAP_MAP);
+TRACE_DEFINE_ENUM(XFS_RMAP_MAP_SHARED);
+TRACE_DEFINE_ENUM(XFS_RMAP_UNMAP);
+TRACE_DEFINE_ENUM(XFS_RMAP_UNMAP_SHARED);
+TRACE_DEFINE_ENUM(XFS_RMAP_CONVERT);
+TRACE_DEFINE_ENUM(XFS_RMAP_CONVERT_SHARED);
+TRACE_DEFINE_ENUM(XFS_RMAP_ALLOC);
+TRACE_DEFINE_ENUM(XFS_RMAP_FREE);
+
 DECLARE_EVENT_CLASS(xchk_class,
 	TP_PROTO(struct xfs_inode *ip, struct xfs_scrub_metadata *sm,
 		 int error),
@@ -153,6 +177,8 @@ DEFINE_EVENT(xchk_class, name, \
 DEFINE_SCRUB_EVENT(xchk_start);
 DEFINE_SCRUB_EVENT(xchk_done);
 DEFINE_SCRUB_EVENT(xchk_deadlock_retry);
+DEFINE_SCRUB_EVENT(xchk_dirtree_start);
+DEFINE_SCRUB_EVENT(xchk_dirtree_done);
 DEFINE_SCRUB_EVENT(xrep_attempt);
 DEFINE_SCRUB_EVENT(xrep_done);
 
@@ -183,6 +209,81 @@ DEFINE_EVENT(xchk_fsgate_class, name, \
 DEFINE_SCRUB_FSHOOK_EVENT(xchk_fsgates_enable);
 DEFINE_SCRUB_FSHOOK_EVENT(xchk_fsgates_disable);
 
+DECLARE_EVENT_CLASS(xchk_vector_head_class,
+	TP_PROTO(struct xfs_inode *ip, struct xfs_scrub_vec_head *vhead),
+	TP_ARGS(ip, vhead),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_ino_t, inum)
+		__field(unsigned int, gen)
+		__field(unsigned int, flags)
+		__field(unsigned short, rest_us)
+		__field(unsigned short, nr_vecs)
+	),
+	TP_fast_assign(
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->agno = vhead->svh_agno;
+		__entry->inum = vhead->svh_ino;
+		__entry->gen = vhead->svh_gen;
+		__entry->flags = vhead->svh_flags;
+		__entry->rest_us = vhead->svh_rest_us;
+		__entry->nr_vecs = vhead->svh_nr;
+	),
+	TP_printk("dev %d:%d ino 0x%llx agno 0x%x inum 0x%llx gen 0x%x flags 0x%x rest_us %u nr_vecs %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->agno,
+		  __entry->inum,
+		  __entry->gen,
+		  __entry->flags,
+		  __entry->rest_us,
+		  __entry->nr_vecs)
+)
+#define DEFINE_SCRUBV_HEAD_EVENT(name) \
+DEFINE_EVENT(xchk_vector_head_class, name, \
+	TP_PROTO(struct xfs_inode *ip, struct xfs_scrub_vec_head *vhead), \
+	TP_ARGS(ip, vhead))
+
+DEFINE_SCRUBV_HEAD_EVENT(xchk_scrubv_start);
+
+DECLARE_EVENT_CLASS(xchk_vector_class,
+	TP_PROTO(struct xfs_mount *mp, struct xfs_scrub_vec_head *vhead,
+		 unsigned int vec_nr, struct xfs_scrub_vec *v),
+	TP_ARGS(mp, vhead, vec_nr, v),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(unsigned int, vec_nr)
+		__field(unsigned int, vec_type)
+		__field(unsigned int, vec_flags)
+		__field(int, vec_ret)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->vec_nr = vec_nr;
+		__entry->vec_type = v->sv_type;
+		__entry->vec_flags = v->sv_flags;
+		__entry->vec_ret = v->sv_ret;
+	),
+	TP_printk("dev %d:%d vec[%u] type %s flags %s ret %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->vec_nr,
+		  __print_symbolic(__entry->vec_type, XFS_SCRUB_TYPE_STRINGS),
+		  __print_flags(__entry->vec_flags, "|", XFS_SCRUB_FLAG_STRINGS),
+		  __entry->vec_ret)
+)
+#define DEFINE_SCRUBV_EVENT(name) \
+DEFINE_EVENT(xchk_vector_class, name, \
+	TP_PROTO(struct xfs_mount *mp, struct xfs_scrub_vec_head *vhead, \
+		 unsigned int vec_nr, struct xfs_scrub_vec *v), \
+	TP_ARGS(mp, vhead, vec_nr, v))
+
+DEFINE_SCRUBV_EVENT(xchk_scrubv_barrier_fail);
+DEFINE_SCRUBV_EVENT(xchk_scrubv_item);
+DEFINE_SCRUBV_EVENT(xchk_scrubv_outcome);
+
 TRACE_EVENT(xchk_op_error,
 	TP_PROTO(struct xfs_scrub *sc, xfs_agnumber_t agno,
 		 xfs_agblock_t bno, int error, void *ret_ip),
@@ -348,6 +449,7 @@ DEFINE_EVENT(xchk_fblock_error_class, name, \
 
 DEFINE_SCRUB_FBLOCK_ERROR_EVENT(xchk_fblock_error);
 DEFINE_SCRUB_FBLOCK_ERROR_EVENT(xchk_fblock_warning);
+DEFINE_SCRUB_FBLOCK_ERROR_EVENT(xchk_fblock_preen);
 
 #ifdef CONFIG_XFS_QUOTA
 DECLARE_EVENT_CLASS(xchk_dqiter_class,
@@ -395,6 +497,29 @@ DEFINE_SCRUB_DQITER_EVENT(xchk_dquot_iter_revalidate_bmap);
 DEFINE_SCRUB_DQITER_EVENT(xchk_dquot_iter_advance_bmap);
 DEFINE_SCRUB_DQITER_EVENT(xchk_dquot_iter_advance_incore);
 DEFINE_SCRUB_DQITER_EVENT(xchk_dquot_iter);
+
+TRACE_EVENT(xchk_qcheck_error,
+	TP_PROTO(struct xfs_scrub *sc, xfs_dqtype_t dqtype, xfs_dqid_t id,
+		 void *ret_ip),
+	TP_ARGS(sc, dqtype, id, ret_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_dqtype_t, dqtype)
+		__field(xfs_dqid_t, id)
+		__field(void *, ret_ip)
+	),
+	TP_fast_assign(
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->dqtype = dqtype;
+		__entry->id = id;
+		__entry->ret_ip = ret_ip;
+	),
+	TP_printk("dev %d:%d dquot type %s id 0x%x ret_ip %pS",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->dqtype, XFS_DQTYPE_STRINGS),
+		  __entry->id,
+		  __entry->ret_ip)
+);
 #endif /* CONFIG_XFS_QUOTA */
 
 TRACE_EVENT(xchk_incomplete,
@@ -423,7 +548,7 @@ TRACE_EVENT(xchk_btree_op_error,
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
 		__field(unsigned int, type)
-		__field(xfs_btnum_t, btnum)
+		__string(name, cur->bc_ops->name)
 		__field(int, level)
 		__field(xfs_agnumber_t, agno)
 		__field(xfs_agblock_t, bno)
@@ -436,7 +561,7 @@ TRACE_EVENT(xchk_btree_op_error,
 
 		__entry->dev = sc->mp->m_super->s_dev;
 		__entry->type = sc->sm->sm_type;
-		__entry->btnum = cur->bc_btnum;
+		__assign_str(name);
 		__entry->level = level;
 		__entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, fsbno);
 		__entry->bno = XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno);
@@ -444,10 +569,10 @@ TRACE_EVENT(xchk_btree_op_error,
 		__entry->error = error;
 		__entry->ret_ip = ret_ip;
 	),
-	TP_printk("dev %d:%d type %s btree %s level %d ptr %d agno 0x%x agbno 0x%x error %d ret_ip %pS",
+	TP_printk("dev %d:%d type %s %sbt level %d ptr %d agno 0x%x agbno 0x%x error %d ret_ip %pS",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __print_symbolic(__entry->type, XFS_SCRUB_TYPE_STRINGS),
-		  __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS),
+		  __get_str(name),
 		  __entry->level,
 		  __entry->ptr,
 		  __entry->agno,
@@ -465,7 +590,7 @@ TRACE_EVENT(xchk_ifork_btree_op_error,
 		__field(xfs_ino_t, ino)
 		__field(int, whichfork)
 		__field(unsigned int, type)
-		__field(xfs_btnum_t, btnum)
+		__string(name, cur->bc_ops->name)
 		__field(int, level)
 		__field(int, ptr)
 		__field(xfs_agnumber_t, agno)
@@ -479,7 +604,7 @@ TRACE_EVENT(xchk_ifork_btree_op_error,
 		__entry->ino = sc->ip->i_ino;
 		__entry->whichfork = cur->bc_ino.whichfork;
 		__entry->type = sc->sm->sm_type;
-		__entry->btnum = cur->bc_btnum;
+		__assign_str(name);
 		__entry->level = level;
 		__entry->ptr = cur->bc_levels[level].ptr;
 		__entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, fsbno);
@@ -487,12 +612,12 @@ TRACE_EVENT(xchk_ifork_btree_op_error,
 		__entry->error = error;
 		__entry->ret_ip = ret_ip;
 	),
-	TP_printk("dev %d:%d ino 0x%llx fork %s type %s btree %s level %d ptr %d agno 0x%x agbno 0x%x error %d ret_ip %pS",
+	TP_printk("dev %d:%d ino 0x%llx fork %s type %s %sbt level %d ptr %d agno 0x%x agbno 0x%x error %d ret_ip %pS",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __entry->ino,
 		  __print_symbolic(__entry->whichfork, XFS_WHICHFORK_STRINGS),
 		  __print_symbolic(__entry->type, XFS_SCRUB_TYPE_STRINGS),
-		  __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS),
+		  __get_str(name),
 		  __entry->level,
 		  __entry->ptr,
 		  __entry->agno,
@@ -508,7 +633,7 @@ TRACE_EVENT(xchk_btree_error,
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
 		__field(unsigned int, type)
-		__field(xfs_btnum_t, btnum)
+		__string(name, cur->bc_ops->name)
 		__field(int, level)
 		__field(xfs_agnumber_t, agno)
 		__field(xfs_agblock_t, bno)
@@ -519,17 +644,17 @@ TRACE_EVENT(xchk_btree_error,
 		xfs_fsblock_t fsbno = xchk_btree_cur_fsbno(cur, level);
 		__entry->dev = sc->mp->m_super->s_dev;
 		__entry->type = sc->sm->sm_type;
-		__entry->btnum = cur->bc_btnum;
+		__assign_str(name);
 		__entry->level = level;
 		__entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, fsbno);
 		__entry->bno = XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno);
 		__entry->ptr = cur->bc_levels[level].ptr;
 		__entry->ret_ip = ret_ip;
 	),
-	TP_printk("dev %d:%d type %s btree %s level %d ptr %d agno 0x%x agbno 0x%x ret_ip %pS",
+	TP_printk("dev %d:%d type %s %sbt level %d ptr %d agno 0x%x agbno 0x%x ret_ip %pS",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __print_symbolic(__entry->type, XFS_SCRUB_TYPE_STRINGS),
-		  __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS),
+		  __get_str(name),
 		  __entry->level,
 		  __entry->ptr,
 		  __entry->agno,
@@ -546,7 +671,7 @@ TRACE_EVENT(xchk_ifork_btree_error,
 		__field(xfs_ino_t, ino)
 		__field(int, whichfork)
 		__field(unsigned int, type)
-		__field(xfs_btnum_t, btnum)
+		__string(name, cur->bc_ops->name)
 		__field(int, level)
 		__field(xfs_agnumber_t, agno)
 		__field(xfs_agblock_t, bno)
@@ -559,19 +684,19 @@ TRACE_EVENT(xchk_ifork_btree_error,
 		__entry->ino = sc->ip->i_ino;
 		__entry->whichfork = cur->bc_ino.whichfork;
 		__entry->type = sc->sm->sm_type;
-		__entry->btnum = cur->bc_btnum;
+		__assign_str(name);
 		__entry->level = level;
 		__entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, fsbno);
 		__entry->bno = XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno);
 		__entry->ptr = cur->bc_levels[level].ptr;
 		__entry->ret_ip = ret_ip;
 	),
-	TP_printk("dev %d:%d ino 0x%llx fork %s type %s btree %s level %d ptr %d agno 0x%x agbno 0x%x ret_ip %pS",
+	TP_printk("dev %d:%d ino 0x%llx fork %s type %s %sbt level %d ptr %d agno 0x%x agbno 0x%x ret_ip %pS",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __entry->ino,
 		  __print_symbolic(__entry->whichfork, XFS_WHICHFORK_STRINGS),
 		  __print_symbolic(__entry->type, XFS_SCRUB_TYPE_STRINGS),
-		  __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS),
+		  __get_str(name),
 		  __entry->level,
 		  __entry->ptr,
 		  __entry->agno,
@@ -586,7 +711,7 @@ DECLARE_EVENT_CLASS(xchk_sbtree_class,
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
 		__field(int, type)
-		__field(xfs_btnum_t, btnum)
+		__string(name, cur->bc_ops->name)
 		__field(xfs_agnumber_t, agno)
 		__field(xfs_agblock_t, bno)
 		__field(int, level)
@@ -598,17 +723,17 @@ DECLARE_EVENT_CLASS(xchk_sbtree_class,
 
 		__entry->dev = sc->mp->m_super->s_dev;
 		__entry->type = sc->sm->sm_type;
-		__entry->btnum = cur->bc_btnum;
+		__assign_str(name);
 		__entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, fsbno);
 		__entry->bno = XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno);
 		__entry->level = level;
 		__entry->nlevels = cur->bc_nlevels;
 		__entry->ptr = cur->bc_levels[level].ptr;
 	),
-	TP_printk("dev %d:%d type %s btree %s agno 0x%x agbno 0x%x level %d nlevels %d ptr %d",
+	TP_printk("dev %d:%d type %s %sbt agno 0x%x agbno 0x%x level %d nlevels %d ptr %d",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __print_symbolic(__entry->type, XFS_SCRUB_TYPE_STRINGS),
-		  __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS),
+		  __get_str(name),
 		  __entry->agno,
 		  __entry->bno,
 		  __entry->level,
@@ -861,18 +986,11 @@ TRACE_EVENT(xfile_destroy,
 		__field(loff_t, size)
 	),
 	TP_fast_assign(
-		struct xfile_stat	statbuf;
-		int			ret;
+		struct inode		*inode = file_inode(xf->file);
 
-		ret = xfile_stat(xf, &statbuf);
-		if (!ret) {
-			__entry->bytes = statbuf.bytes;
-			__entry->size = statbuf.size;
-		} else {
-			__entry->bytes = -1;
-			__entry->size = -1;
-		}
-		__entry->ino = file_inode(xf->file)->i_ino;
+		__entry->ino = inode->i_ino;
+		__entry->bytes = inode->i_blocks << SECTOR_SHIFT;
+		__entry->size = i_size_read(inode);
 	),
 	TP_printk("xfino 0x%lx mem_bytes 0x%llx isize 0x%llx",
 		  __entry->ino,
@@ -891,19 +1009,12 @@ DECLARE_EVENT_CLASS(xfile_class,
 		__field(unsigned long long, bytecount)
 	),
 	TP_fast_assign(
-		struct xfile_stat	statbuf;
-		int			ret;
+		struct inode		*inode = file_inode(xf->file);
 
-		ret = xfile_stat(xf, &statbuf);
-		if (!ret) {
-			__entry->bytes_used = statbuf.bytes;
-			__entry->size = statbuf.size;
-		} else {
-			__entry->bytes_used = -1;
-			__entry->size = -1;
-		}
-		__entry->ino = file_inode(xf->file)->i_ino;
+		__entry->ino = inode->i_ino;
+		__entry->bytes_used = inode->i_blocks << SECTOR_SHIFT;
 		__entry->pos = pos;
+		__entry->size = i_size_read(inode);
 		__entry->bytecount = bytecount;
 	),
 	TP_printk("xfino 0x%lx mem_bytes 0x%llx pos 0x%llx bytecount 0x%llx isize 0x%llx",
@@ -917,11 +1028,12 @@ DECLARE_EVENT_CLASS(xfile_class,
 DEFINE_EVENT(xfile_class, name, \
 	TP_PROTO(struct xfile *xf, loff_t pos, unsigned long long bytecount), \
 	TP_ARGS(xf, pos, bytecount))
-DEFINE_XFILE_EVENT(xfile_pread);
-DEFINE_XFILE_EVENT(xfile_pwrite);
+DEFINE_XFILE_EVENT(xfile_load);
+DEFINE_XFILE_EVENT(xfile_store);
 DEFINE_XFILE_EVENT(xfile_seek_data);
-DEFINE_XFILE_EVENT(xfile_get_page);
-DEFINE_XFILE_EVENT(xfile_put_page);
+DEFINE_XFILE_EVENT(xfile_get_folio);
+DEFINE_XFILE_EVENT(xfile_put_folio);
+DEFINE_XFILE_EVENT(xfile_discard);
 
 TRACE_EVENT(xfarray_create,
 	TP_PROTO(struct xfarray *xfa, unsigned long long required_capacity),
@@ -968,7 +1080,7 @@ TRACE_EVENT(xfarray_isort,
 		  __entry->hi - __entry->lo)
 );
 
-TRACE_EVENT(xfarray_pagesort,
+TRACE_EVENT(xfarray_foliosort,
 	TP_PROTO(struct xfarray_sortinfo *si, uint64_t lo, uint64_t hi),
 	TP_ARGS(si, lo, hi),
 	TP_STRUCT__entry(
@@ -1039,6 +1151,47 @@ TRACE_EVENT(xfarray_sort,
 		  __entry->bytes)
 );
 
+TRACE_EVENT(xfarray_sort_scan,
+	TP_PROTO(struct xfarray_sortinfo *si, unsigned long long idx),
+	TP_ARGS(si, idx),
+	TP_STRUCT__entry(
+		__field(unsigned long, ino)
+		__field(unsigned long long, nr)
+		__field(size_t, obj_size)
+		__field(unsigned long long, idx)
+		__field(unsigned long long, folio_pos)
+		__field(unsigned long, folio_bytes)
+		__field(unsigned long long, first_idx)
+		__field(unsigned long long, last_idx)
+	),
+	TP_fast_assign(
+		__entry->nr = si->array->nr;
+		__entry->obj_size = si->array->obj_size;
+		__entry->ino = file_inode(si->array->xfile->file)->i_ino;
+		__entry->idx = idx;
+		if (si->folio) {
+			__entry->folio_pos = folio_pos(si->folio);
+			__entry->folio_bytes = folio_size(si->folio);
+			__entry->first_idx = si->first_folio_idx;
+			__entry->last_idx = si->last_folio_idx;
+		} else {
+			__entry->folio_pos = 0;
+			__entry->folio_bytes = 0;
+			__entry->first_idx = 0;
+			__entry->last_idx = 0;
+		}
+	),
+	TP_printk("xfino 0x%lx nr %llu objsz %zu idx %llu folio_pos 0x%llx folio_bytes 0x%lx first_idx %llu last_idx %llu",
+		  __entry->ino,
+		  __entry->nr,
+		  __entry->obj_size,
+		  __entry->idx,
+		  __entry->folio_pos,
+		  __entry->folio_bytes,
+		  __entry->first_idx,
+		  __entry->last_idx)
+);
+
 TRACE_EVENT(xfarray_sort_stats,
 	TP_PROTO(struct xfarray_sortinfo *si, int error),
 	TP_ARGS(si, error),
@@ -1119,6 +1272,650 @@ TRACE_EVENT(xchk_rtsum_record_free,
 );
 #endif /* CONFIG_XFS_RT */
 
+DECLARE_EVENT_CLASS(xchk_iscan_class,
+	TP_PROTO(struct xchk_iscan *iscan),
+	TP_ARGS(iscan),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, cursor)
+		__field(xfs_ino_t, visited)
+	),
+	TP_fast_assign(
+		__entry->dev = iscan->sc->mp->m_super->s_dev;
+		__entry->cursor = iscan->cursor_ino;
+		__entry->visited = iscan->__visited_ino;
+	),
+	TP_printk("dev %d:%d iscan cursor 0x%llx visited 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->cursor,
+		  __entry->visited)
+)
+#define DEFINE_ISCAN_EVENT(name) \
+DEFINE_EVENT(xchk_iscan_class, name, \
+	TP_PROTO(struct xchk_iscan *iscan), \
+	TP_ARGS(iscan))
+DEFINE_ISCAN_EVENT(xchk_iscan_move_cursor);
+DEFINE_ISCAN_EVENT(xchk_iscan_visit);
+DEFINE_ISCAN_EVENT(xchk_iscan_skip);
+DEFINE_ISCAN_EVENT(xchk_iscan_advance_ag);
+
+DECLARE_EVENT_CLASS(xchk_iscan_ino_class,
+	TP_PROTO(struct xchk_iscan *iscan, xfs_ino_t ino),
+	TP_ARGS(iscan, ino),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, startino)
+		__field(xfs_ino_t, cursor)
+		__field(xfs_ino_t, visited)
+		__field(xfs_ino_t, ino)
+	),
+	TP_fast_assign(
+		__entry->dev = iscan->sc->mp->m_super->s_dev;
+		__entry->startino = iscan->scan_start_ino;
+		__entry->cursor = iscan->cursor_ino;
+		__entry->visited = iscan->__visited_ino;
+		__entry->ino = ino;
+	),
+	TP_printk("dev %d:%d iscan start 0x%llx cursor 0x%llx visited 0x%llx ino 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->startino,
+		  __entry->cursor,
+		  __entry->visited,
+		  __entry->ino)
+)
+#define DEFINE_ISCAN_INO_EVENT(name) \
+DEFINE_EVENT(xchk_iscan_ino_class, name, \
+	TP_PROTO(struct xchk_iscan *iscan, xfs_ino_t ino), \
+	TP_ARGS(iscan, ino))
+DEFINE_ISCAN_INO_EVENT(xchk_iscan_want_live_update);
+DEFINE_ISCAN_INO_EVENT(xchk_iscan_start);
+
+TRACE_EVENT(xchk_iscan_iget,
+	TP_PROTO(struct xchk_iscan *iscan, int error),
+	TP_ARGS(iscan, error),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, cursor)
+		__field(xfs_ino_t, visited)
+		__field(int, error)
+	),
+	TP_fast_assign(
+		__entry->dev = iscan->sc->mp->m_super->s_dev;
+		__entry->cursor = iscan->cursor_ino;
+		__entry->visited = iscan->__visited_ino;
+		__entry->error = error;
+	),
+	TP_printk("dev %d:%d iscan cursor 0x%llx visited 0x%llx error %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->cursor,
+		  __entry->visited,
+		  __entry->error)
+);
+
+TRACE_EVENT(xchk_iscan_iget_batch,
+	TP_PROTO(struct xfs_mount *mp, struct xchk_iscan *iscan,
+		 unsigned int nr, unsigned int avail),
+	TP_ARGS(mp, iscan, nr, avail),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, cursor)
+		__field(xfs_ino_t, visited)
+		__field(unsigned int, nr)
+		__field(unsigned int, avail)
+		__field(unsigned int, unavail)
+		__field(xfs_ino_t, batch_ino)
+		__field(unsigned long long, skipmask)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->cursor = iscan->cursor_ino;
+		__entry->visited = iscan->__visited_ino;
+		__entry->nr = nr;
+		__entry->avail = avail;
+		__entry->unavail = hweight64(iscan->__skipped_inomask);
+		__entry->batch_ino = iscan->__batch_ino;
+		__entry->skipmask = iscan->__skipped_inomask;
+	),
+	TP_printk("dev %d:%d iscan cursor 0x%llx visited 0x%llx batchino 0x%llx skipmask 0x%llx nr %u avail %u unavail %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->cursor,
+		  __entry->visited,
+		  __entry->batch_ino,
+		  __entry->skipmask,
+		  __entry->nr,
+		  __entry->avail,
+		  __entry->unavail)
+);
+
+DECLARE_EVENT_CLASS(xchk_iscan_retry_wait_class,
+	TP_PROTO(struct xchk_iscan *iscan),
+	TP_ARGS(iscan),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, cursor)
+		__field(xfs_ino_t, visited)
+		__field(unsigned int, retry_delay)
+		__field(unsigned long, remaining)
+		__field(unsigned int, iget_timeout)
+	),
+	TP_fast_assign(
+		__entry->dev = iscan->sc->mp->m_super->s_dev;
+		__entry->cursor = iscan->cursor_ino;
+		__entry->visited = iscan->__visited_ino;
+		__entry->retry_delay = iscan->iget_retry_delay;
+		__entry->remaining = jiffies_to_msecs(iscan->__iget_deadline - jiffies);
+		__entry->iget_timeout = iscan->iget_timeout;
+	),
+	TP_printk("dev %d:%d iscan cursor 0x%llx visited 0x%llx remaining %lu timeout %u delay %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->cursor,
+		  __entry->visited,
+		  __entry->remaining,
+		  __entry->iget_timeout,
+		  __entry->retry_delay)
+)
+#define DEFINE_ISCAN_RETRY_WAIT_EVENT(name) \
+DEFINE_EVENT(xchk_iscan_retry_wait_class, name, \
+	TP_PROTO(struct xchk_iscan *iscan), \
+	TP_ARGS(iscan))
+DEFINE_ISCAN_RETRY_WAIT_EVENT(xchk_iscan_iget_retry_wait);
+DEFINE_ISCAN_RETRY_WAIT_EVENT(xchk_iscan_agi_retry_wait);
+
+TRACE_EVENT(xchk_nlinks_collect_dirent,
+	TP_PROTO(struct xfs_mount *mp, struct xfs_inode *dp,
+		 xfs_ino_t ino, const struct xfs_name *name),
+	TP_ARGS(mp, dp, ino, name),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, dir)
+		__field(xfs_ino_t, ino)
+		__field(unsigned int, namelen)
+		__dynamic_array(char, name, name->len)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->dir = dp->i_ino;
+		__entry->ino = ino;
+		__entry->namelen = name->len;
+		memcpy(__get_str(name), name->name, name->len);
+	),
+	TP_printk("dev %d:%d dir 0x%llx -> ino 0x%llx name '%.*s'",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->dir,
+		  __entry->ino,
+		  __entry->namelen,
+		  __get_str(name))
+);
+
+TRACE_EVENT(xchk_nlinks_collect_pptr,
+	TP_PROTO(struct xfs_mount *mp, struct xfs_inode *dp,
+		 const struct xfs_name *name,
+		 const struct xfs_parent_rec *pptr),
+	TP_ARGS(mp, dp, name, pptr),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, dir)
+		__field(xfs_ino_t, ino)
+		__field(unsigned int, namelen)
+		__dynamic_array(char, name, name->len)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->dir = dp->i_ino;
+		__entry->ino = be64_to_cpu(pptr->p_ino);
+		__entry->namelen = name->len;
+		memcpy(__get_str(name), name->name, name->len);
+	),
+	TP_printk("dev %d:%d dir 0x%llx -> ino 0x%llx name '%.*s'",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->dir,
+		  __entry->ino,
+		  __entry->namelen,
+		  __get_str(name))
+);
+
+TRACE_EVENT(xchk_nlinks_collect_metafile,
+	TP_PROTO(struct xfs_mount *mp, xfs_ino_t ino),
+	TP_ARGS(mp, ino),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->ino = ino;
+	),
+	TP_printk("dev %d:%d ino 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino)
+);
+
+TRACE_EVENT(xchk_nlinks_live_update,
+	TP_PROTO(struct xfs_mount *mp, const struct xfs_inode *dp,
+		 int action, xfs_ino_t ino, int delta,
+		 const char *name, unsigned int namelen),
+	TP_ARGS(mp, dp, action, ino, delta, name, namelen),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, dir)
+		__field(int, action)
+		__field(xfs_ino_t, ino)
+		__field(int, delta)
+		__field(unsigned int, namelen)
+		__dynamic_array(char, name, namelen)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->dir = dp ? dp->i_ino : NULLFSINO;
+		__entry->action = action;
+		__entry->ino = ino;
+		__entry->delta = delta;
+		__entry->namelen = namelen;
+		memcpy(__get_str(name), name, namelen);
+	),
+	TP_printk("dev %d:%d dir 0x%llx ino 0x%llx nlink_delta %d name '%.*s'",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->dir,
+		  __entry->ino,
+		  __entry->delta,
+		  __entry->namelen,
+		  __get_str(name))
+);
+
+TRACE_EVENT(xchk_nlinks_check_zero,
+	TP_PROTO(struct xfs_mount *mp, xfs_ino_t ino,
+		 const struct xchk_nlink *live),
+	TP_ARGS(mp, ino, live),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_nlink_t, parents)
+		__field(xfs_nlink_t, backrefs)
+		__field(xfs_nlink_t, children)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->ino = ino;
+		__entry->parents = live->parents;
+		__entry->backrefs = live->backrefs;
+		__entry->children = live->children;
+	),
+	TP_printk("dev %d:%d ino 0x%llx parents %u backrefs %u children %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->parents,
+		  __entry->backrefs,
+		  __entry->children)
+);
+
+TRACE_EVENT(xchk_nlinks_update_incore,
+	TP_PROTO(struct xfs_mount *mp, xfs_ino_t ino,
+		 const struct xchk_nlink *live, int parents_delta,
+		 int backrefs_delta, int children_delta),
+	TP_ARGS(mp, ino, live, parents_delta, backrefs_delta, children_delta),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_nlink_t, parents)
+		__field(xfs_nlink_t, backrefs)
+		__field(xfs_nlink_t, children)
+		__field(int, parents_delta)
+		__field(int, backrefs_delta)
+		__field(int, children_delta)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->ino = ino;
+		__entry->parents = live->parents;
+		__entry->backrefs = live->backrefs;
+		__entry->children = live->children;
+		__entry->parents_delta = parents_delta;
+		__entry->backrefs_delta = backrefs_delta;
+		__entry->children_delta = children_delta;
+	),
+	TP_printk("dev %d:%d ino 0x%llx parents %d:%u backrefs %d:%u children %d:%u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->parents_delta,
+		  __entry->parents,
+		  __entry->backrefs_delta,
+		  __entry->backrefs,
+		  __entry->children_delta,
+		  __entry->children)
+);
+
+DECLARE_EVENT_CLASS(xchk_nlinks_diff_class,
+	TP_PROTO(struct xfs_mount *mp, struct xfs_inode *ip,
+		 const struct xchk_nlink *live),
+	TP_ARGS(mp, ip, live),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(uint8_t, ftype)
+		__field(xfs_nlink_t, nlink)
+		__field(xfs_nlink_t, parents)
+		__field(xfs_nlink_t, backrefs)
+		__field(xfs_nlink_t, children)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->ftype = xfs_mode_to_ftype(VFS_I(ip)->i_mode);
+		__entry->nlink = VFS_I(ip)->i_nlink;
+		__entry->parents = live->parents;
+		__entry->backrefs = live->backrefs;
+		__entry->children = live->children;
+	),
+	TP_printk("dev %d:%d ino 0x%llx ftype %s nlink %u parents %u backrefs %u children %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __print_symbolic(__entry->ftype, XFS_DIR3_FTYPE_STR),
+		  __entry->nlink,
+		  __entry->parents,
+		  __entry->backrefs,
+		  __entry->children)
+);
+#define DEFINE_SCRUB_NLINKS_DIFF_EVENT(name) \
+DEFINE_EVENT(xchk_nlinks_diff_class, name, \
+	TP_PROTO(struct xfs_mount *mp, struct xfs_inode *ip, \
+		 const struct xchk_nlink *live), \
+	TP_ARGS(mp, ip, live))
+DEFINE_SCRUB_NLINKS_DIFF_EVENT(xchk_nlinks_compare_inode);
+
+DECLARE_EVENT_CLASS(xchk_pptr_class,
+	TP_PROTO(struct xfs_inode *ip, const struct xfs_name *name,
+		 xfs_ino_t far_ino),
+	TP_ARGS(ip, name, far_ino),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(unsigned int, namelen)
+		__dynamic_array(char, name, name->len)
+		__field(xfs_ino_t, far_ino)
+	),
+	TP_fast_assign(
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->namelen = name->len;
+		memcpy(__get_str(name), name, name->len);
+		__entry->far_ino = far_ino;
+	),
+	TP_printk("dev %d:%d ino 0x%llx name '%.*s' far_ino 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->namelen,
+		  __get_str(name),
+		  __entry->far_ino)
+)
+#define DEFINE_XCHK_PPTR_EVENT(name) \
+DEFINE_EVENT(xchk_pptr_class, name, \
+	TP_PROTO(struct xfs_inode *ip, const struct xfs_name *name, \
+		 xfs_ino_t far_ino), \
+	TP_ARGS(ip, name, far_ino))
+DEFINE_XCHK_PPTR_EVENT(xchk_dir_defer);
+DEFINE_XCHK_PPTR_EVENT(xchk_dir_slowpath);
+DEFINE_XCHK_PPTR_EVENT(xchk_dir_ultraslowpath);
+DEFINE_XCHK_PPTR_EVENT(xchk_parent_defer);
+DEFINE_XCHK_PPTR_EVENT(xchk_parent_slowpath);
+DEFINE_XCHK_PPTR_EVENT(xchk_parent_ultraslowpath);
+
+DECLARE_EVENT_CLASS(xchk_dirtree_class,
+	TP_PROTO(struct xfs_scrub *sc, struct xfs_inode *ip,
+		 unsigned int path_nr, const struct xfs_name *name,
+		 const struct xfs_parent_rec *pptr),
+	TP_ARGS(sc, ip, path_nr, name, pptr),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(unsigned int, path_nr)
+		__field(xfs_ino_t, child_ino)
+		__field(unsigned int, child_gen)
+		__field(xfs_ino_t, parent_ino)
+		__field(unsigned int, parent_gen)
+		__field(unsigned int, namelen)
+		__dynamic_array(char, name, name->len)
+	),
+	TP_fast_assign(
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->path_nr = path_nr;
+		__entry->child_ino = ip->i_ino;
+		__entry->child_gen = VFS_I(ip)->i_generation;
+		__entry->parent_ino = be64_to_cpu(pptr->p_ino);
+		__entry->parent_gen = be32_to_cpu(pptr->p_gen);
+		__entry->namelen = name->len;
+		memcpy(__get_str(name), name->name, name->len);
+	),
+	TP_printk("dev %d:%d path %u child_ino 0x%llx child_gen 0x%x parent_ino 0x%llx parent_gen 0x%x name '%.*s'",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->path_nr,
+		  __entry->child_ino,
+		  __entry->child_gen,
+		  __entry->parent_ino,
+		  __entry->parent_gen,
+		  __entry->namelen,
+		  __get_str(name))
+);
+#define DEFINE_XCHK_DIRTREE_EVENT(name) \
+DEFINE_EVENT(xchk_dirtree_class, name, \
+	TP_PROTO(struct xfs_scrub *sc, struct xfs_inode *ip, \
+		 unsigned int path_nr, const struct xfs_name *name, \
+		 const struct xfs_parent_rec *pptr), \
+	TP_ARGS(sc, ip, path_nr, name, pptr))
+DEFINE_XCHK_DIRTREE_EVENT(xchk_dirtree_create_path);
+DEFINE_XCHK_DIRTREE_EVENT(xchk_dirpath_walk_upwards);
+
+DECLARE_EVENT_CLASS(xchk_dirpath_class,
+	TP_PROTO(struct xfs_scrub *sc, struct xfs_inode *ip,
+		 unsigned int path_nr, unsigned int step_nr,
+		 const struct xfs_name *name,
+		 const struct xfs_parent_rec *pptr),
+	TP_ARGS(sc, ip, path_nr, step_nr, name, pptr),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(unsigned int, path_nr)
+		__field(unsigned int, step_nr)
+		__field(xfs_ino_t, child_ino)
+		__field(unsigned int, child_gen)
+		__field(xfs_ino_t, parent_ino)
+		__field(unsigned int, parent_gen)
+		__field(unsigned int, namelen)
+		__dynamic_array(char, name, name->len)
+	),
+	TP_fast_assign(
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->path_nr = path_nr;
+		__entry->step_nr = step_nr;
+		__entry->child_ino = ip->i_ino;
+		__entry->child_gen = VFS_I(ip)->i_generation;
+		__entry->parent_ino = be64_to_cpu(pptr->p_ino);
+		__entry->parent_gen = be32_to_cpu(pptr->p_gen);
+		__entry->namelen = name->len;
+		memcpy(__get_str(name), name->name, name->len);
+	),
+	TP_printk("dev %d:%d path %u step %u child_ino 0x%llx child_gen 0x%x parent_ino 0x%llx parent_gen 0x%x name '%.*s'",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->path_nr,
+		  __entry->step_nr,
+		  __entry->child_ino,
+		  __entry->child_gen,
+		  __entry->parent_ino,
+		  __entry->parent_gen,
+		  __entry->namelen,
+		  __get_str(name))
+);
+#define DEFINE_XCHK_DIRPATH_EVENT(name) \
+DEFINE_EVENT(xchk_dirpath_class, name, \
+	TP_PROTO(struct xfs_scrub *sc, struct xfs_inode *ip, \
+		 unsigned int path_nr, unsigned int step_nr, \
+		 const struct xfs_name *name, \
+		 const struct xfs_parent_rec *pptr), \
+	TP_ARGS(sc, ip, path_nr, step_nr, name, pptr))
+DEFINE_XCHK_DIRPATH_EVENT(xchk_dirpath_disappeared);
+DEFINE_XCHK_DIRPATH_EVENT(xchk_dirpath_badgen);
+DEFINE_XCHK_DIRPATH_EVENT(xchk_dirpath_nondir_parent);
+DEFINE_XCHK_DIRPATH_EVENT(xchk_dirpath_unlinked_parent);
+DEFINE_XCHK_DIRPATH_EVENT(xchk_dirpath_found_next_step);
+
+TRACE_DEFINE_ENUM(XCHK_DIRPATH_SCANNING);
+TRACE_DEFINE_ENUM(XCHK_DIRPATH_DELETE);
+TRACE_DEFINE_ENUM(XCHK_DIRPATH_CORRUPT);
+TRACE_DEFINE_ENUM(XCHK_DIRPATH_LOOP);
+TRACE_DEFINE_ENUM(XCHK_DIRPATH_STALE);
+TRACE_DEFINE_ENUM(XCHK_DIRPATH_OK);
+TRACE_DEFINE_ENUM(XREP_DIRPATH_DELETING);
+TRACE_DEFINE_ENUM(XREP_DIRPATH_DELETED);
+TRACE_DEFINE_ENUM(XREP_DIRPATH_ADOPTING);
+TRACE_DEFINE_ENUM(XREP_DIRPATH_ADOPTED);
+
+#define XCHK_DIRPATH_OUTCOME_STRINGS \
+	{ XCHK_DIRPATH_SCANNING,	"scanning" }, \
+	{ XCHK_DIRPATH_DELETE,		"delete" }, \
+	{ XCHK_DIRPATH_CORRUPT,		"corrupt" }, \
+	{ XCHK_DIRPATH_LOOP,		"loop" }, \
+	{ XCHK_DIRPATH_STALE,		"stale" }, \
+	{ XCHK_DIRPATH_OK,		"ok" }, \
+	{ XREP_DIRPATH_DELETING,	"deleting" }, \
+	{ XREP_DIRPATH_DELETED,		"deleted" }, \
+	{ XREP_DIRPATH_ADOPTING,	"adopting" }, \
+	{ XREP_DIRPATH_ADOPTED,		"adopted" }
+
+DECLARE_EVENT_CLASS(xchk_dirpath_outcome_class,
+	TP_PROTO(struct xfs_scrub *sc, unsigned long long path_nr,
+		 unsigned int nr_steps, \
+		 unsigned int outcome),
+	TP_ARGS(sc, path_nr, nr_steps, outcome),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(unsigned long long, path_nr)
+		__field(unsigned int, nr_steps)
+		__field(unsigned int, outcome)
+	),
+	TP_fast_assign(
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->path_nr = path_nr;
+		__entry->nr_steps = nr_steps;
+		__entry->outcome = outcome;
+	),
+	TP_printk("dev %d:%d path %llu steps %u outcome %s",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->path_nr,
+		  __entry->nr_steps,
+		  __print_symbolic(__entry->outcome, XCHK_DIRPATH_OUTCOME_STRINGS))
+);
+#define DEFINE_XCHK_DIRPATH_OUTCOME_EVENT(name) \
+DEFINE_EVENT(xchk_dirpath_outcome_class, name, \
+	TP_PROTO(struct xfs_scrub *sc, unsigned long long path_nr, \
+		 unsigned int nr_steps, \
+		 unsigned int outcome), \
+	TP_ARGS(sc, path_nr, nr_steps, outcome))
+DEFINE_XCHK_DIRPATH_OUTCOME_EVENT(xchk_dirpath_set_outcome);
+DEFINE_XCHK_DIRPATH_OUTCOME_EVENT(xchk_dirpath_evaluate_path);
+
+DECLARE_EVENT_CLASS(xchk_dirtree_evaluate_class,
+	TP_PROTO(const struct xchk_dirtree *dl,
+		 const struct xchk_dirtree_outcomes *oc),
+	TP_ARGS(dl, oc),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_ino_t, rootino)
+		__field(unsigned int, nr_paths)
+		__field(unsigned int, bad)
+		__field(unsigned int, suspect)
+		__field(unsigned int, good)
+		__field(bool, needs_adoption)
+	),
+	TP_fast_assign(
+		__entry->dev = dl->sc->mp->m_super->s_dev;
+		__entry->ino = dl->sc->ip->i_ino;
+		__entry->rootino = dl->root_ino;
+		__entry->nr_paths = dl->nr_paths;
+		__entry->bad = oc->bad;
+		__entry->suspect = oc->suspect;
+		__entry->good = oc->good;
+		__entry->needs_adoption = oc->needs_adoption ? 1 : 0;
+	),
+	TP_printk("dev %d:%d ino 0x%llx rootino 0x%llx nr_paths %u bad %u suspect %u good %u adopt? %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->rootino,
+		  __entry->nr_paths,
+		  __entry->bad,
+		  __entry->suspect,
+		  __entry->good,
+		  __entry->needs_adoption)
+);
+#define DEFINE_XCHK_DIRTREE_EVALUATE_EVENT(name) \
+DEFINE_EVENT(xchk_dirtree_evaluate_class, name, \
+	TP_PROTO(const struct xchk_dirtree *dl, \
+		 const struct xchk_dirtree_outcomes *oc), \
+	TP_ARGS(dl, oc))
+DEFINE_XCHK_DIRTREE_EVALUATE_EVENT(xchk_dirtree_evaluate);
+
+TRACE_EVENT(xchk_dirpath_changed,
+	TP_PROTO(struct xfs_scrub *sc, unsigned int path_nr,
+		 unsigned int step_nr, const struct xfs_inode *dp,
+		 const struct xfs_inode *ip, const struct xfs_name *xname),
+	TP_ARGS(sc, path_nr, step_nr, dp, ip, xname),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(unsigned int, path_nr)
+		__field(unsigned int, step_nr)
+		__field(xfs_ino_t, child_ino)
+		__field(xfs_ino_t, parent_ino)
+		__field(unsigned int, namelen)
+		__dynamic_array(char, name, xname->len)
+	),
+	TP_fast_assign(
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->path_nr = path_nr;
+		__entry->step_nr = step_nr;
+		__entry->child_ino = ip->i_ino;
+		__entry->parent_ino = dp->i_ino;
+		__entry->namelen = xname->len;
+		memcpy(__get_str(name), xname->name, xname->len);
+	),
+	TP_printk("dev %d:%d path %u step %u child_ino 0x%llx parent_ino 0x%llx name '%.*s'",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->path_nr,
+		  __entry->step_nr,
+		  __entry->child_ino,
+		  __entry->parent_ino,
+		  __entry->namelen,
+		  __get_str(name))
+);
+
+TRACE_EVENT(xchk_dirtree_live_update,
+	TP_PROTO(struct xfs_scrub *sc, const struct xfs_inode *dp,
+		 int action, const struct xfs_inode *ip, int delta,
+		 const struct xfs_name *xname),
+	TP_ARGS(sc, dp, action, ip, delta, xname),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, parent_ino)
+		__field(int, action)
+		__field(xfs_ino_t, child_ino)
+		__field(int, delta)
+		__field(unsigned int, namelen)
+		__dynamic_array(char, name, xname->len)
+	),
+	TP_fast_assign(
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->parent_ino = dp->i_ino;
+		__entry->action = action;
+		__entry->child_ino = ip->i_ino;
+		__entry->delta = delta;
+		__entry->namelen = xname->len;
+		memcpy(__get_str(name), xname->name, xname->len);
+	),
+	TP_printk("dev %d:%d parent_ino 0x%llx child_ino 0x%llx nlink_delta %d name '%.*s'",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->parent_ino,
+		  __entry->child_ino,
+		  __entry->delta,
+		  __entry->namelen,
+		  __get_str(name))
+);
+
 /* repair tracepoints */
 #if IS_ENABLED(CONFIG_XFS_ONLINE_REPAIR)
 
@@ -1150,6 +1947,7 @@ DEFINE_EVENT(xrep_extent_class, name, \
 DEFINE_REPAIR_EXTENT_EVENT(xreap_dispose_unmap_extent);
 DEFINE_REPAIR_EXTENT_EVENT(xreap_dispose_free_extent);
 DEFINE_REPAIR_EXTENT_EVENT(xreap_agextent_binval);
+DEFINE_REPAIR_EXTENT_EVENT(xreap_bmapi_binval);
 DEFINE_REPAIR_EXTENT_EVENT(xrep_agfl_insert);
 
 DECLARE_EVENT_CLASS(xrep_reap_find_class,
@@ -1183,6 +1981,7 @@ DEFINE_EVENT(xrep_reap_find_class, name, \
 		 bool crosslinked), \
 	TP_ARGS(pag, agbno, len, crosslinked))
 DEFINE_REPAIR_REAP_FIND_EVENT(xreap_agextent_select);
+DEFINE_REPAIR_REAP_FIND_EVENT(xreap_bmapi_select);
 
 DECLARE_EVENT_CLASS(xrep_rmap_class,
 	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
@@ -1223,7 +2022,6 @@ DEFINE_EVENT(xrep_rmap_class, name, \
 		 uint64_t owner, uint64_t offset, unsigned int flags), \
 	TP_ARGS(mp, agno, agbno, len, owner, offset, flags))
 DEFINE_REPAIR_RMAP_EVENT(xrep_ibt_walk_rmap);
-DEFINE_REPAIR_RMAP_EVENT(xrep_rmap_extent_fn);
 DEFINE_REPAIR_RMAP_EVENT(xrep_bmap_walk_rmap);
 
 TRACE_EVENT(xrep_abt_found,
@@ -1341,6 +2139,38 @@ TRACE_EVENT(xrep_bmap_found,
 		  __entry->state)
 );
 
+TRACE_EVENT(xrep_rmap_found,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 const struct xfs_rmap_irec *rec),
+	TP_ARGS(mp, agno, rec),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, agbno)
+		__field(xfs_extlen_t, len)
+		__field(uint64_t, owner)
+		__field(uint64_t, offset)
+		__field(unsigned int, flags)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->agbno = rec->rm_startblock;
+		__entry->len = rec->rm_blockcount;
+		__entry->owner = rec->rm_owner;
+		__entry->offset = rec->rm_offset;
+		__entry->flags = rec->rm_flags;
+	),
+	TP_printk("dev %d:%d agno 0x%x agbno 0x%x fsbcount 0x%x owner 0x%llx fileoff 0x%llx flags 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->agbno,
+		  __entry->len,
+		  __entry->owner,
+		  __entry->offset,
+		  __entry->flags)
+);
+
 TRACE_EVENT(xrep_findroot_block,
 	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, xfs_agblock_t agbno,
 		 uint32_t magic, uint16_t level),
@@ -1425,16 +2255,28 @@ TRACE_EVENT(xrep_calc_ag_resblks_btsize,
 		  __entry->refcbt_sz)
 )
 TRACE_EVENT(xrep_reset_counters,
-	TP_PROTO(struct xfs_mount *mp),
-	TP_ARGS(mp),
+	TP_PROTO(struct xfs_mount *mp, struct xchk_fscounters *fsc),
+	TP_ARGS(mp, fsc),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
+		__field(uint64_t, icount)
+		__field(uint64_t, ifree)
+		__field(uint64_t, fdblocks)
+		__field(uint64_t, frextents)
 	),
 	TP_fast_assign(
 		__entry->dev = mp->m_super->s_dev;
+		__entry->icount = fsc->icount;
+		__entry->ifree = fsc->ifree;
+		__entry->fdblocks = fsc->fdblocks;
+		__entry->frextents = fsc->frextents;
 	),
-	TP_printk("dev %d:%d",
-		  MAJOR(__entry->dev), MINOR(__entry->dev))
+	TP_printk("dev %d:%d icount %llu ifree %llu fdblocks %llu frextents %llu",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->icount,
+		  __entry->ifree,
+		  __entry->fdblocks,
+		  __entry->frextents)
 )
 
 DECLARE_EVENT_CLASS(xrep_newbt_extent_class,
@@ -1645,6 +2487,55 @@ TRACE_EVENT(xrep_dinode_count_rmaps,
 		  __entry->attr_extents)
 );
 
+TRACE_EVENT(xrep_dinode_findmode_dirent,
+	TP_PROTO(struct xfs_scrub *sc, struct xfs_inode *dp,
+		 unsigned int ftype),
+	TP_ARGS(sc, dp, ftype),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_ino_t, parent_ino)
+		__field(unsigned int, ftype)
+	),
+	TP_fast_assign(
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->ino = sc->sm->sm_ino;
+		__entry->parent_ino = dp->i_ino;
+		__entry->ftype = ftype;
+	),
+	TP_printk("dev %d:%d ino 0x%llx parent_ino 0x%llx ftype '%s'",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->parent_ino,
+		  __print_symbolic(__entry->ftype, XFS_DIR3_FTYPE_STR))
+);
+
+TRACE_EVENT(xrep_dinode_findmode_dirent_inval,
+	TP_PROTO(struct xfs_scrub *sc, struct xfs_inode *dp,
+		 unsigned int ftype, unsigned int found_ftype),
+	TP_ARGS(sc, dp, ftype, found_ftype),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_ino_t, parent_ino)
+		__field(unsigned int, ftype)
+		__field(unsigned int, found_ftype)
+	),
+	TP_fast_assign(
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->ino = sc->sm->sm_ino;
+		__entry->parent_ino = dp->i_ino;
+		__entry->ftype = ftype;
+		__entry->found_ftype = found_ftype;
+	),
+	TP_printk("dev %d:%d ino 0x%llx parent_ino 0x%llx ftype '%s' found_ftype '%s'",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->parent_ino,
+		  __print_symbolic(__entry->ftype, XFS_DIR3_FTYPE_STR),
+		  __print_symbolic(__entry->found_ftype, XFS_DIR3_FTYPE_STR))
+);
+
 TRACE_EVENT(xrep_cow_mark_file_range,
 	TP_PROTO(struct xfs_inode *ip, xfs_fsblock_t startblock,
 		 xfs_fileoff_t startoff, xfs_filblks_t blockcount),
@@ -1756,8 +2647,933 @@ DEFINE_EVENT(xrep_dquot_class, name, \
 DEFINE_XREP_DQUOT_EVENT(xrep_dquot_item);
 DEFINE_XREP_DQUOT_EVENT(xrep_disk_dquot);
 DEFINE_XREP_DQUOT_EVENT(xrep_dquot_item_fill_bmap_hole);
+DEFINE_XREP_DQUOT_EVENT(xrep_quotacheck_dquot);
 #endif /* CONFIG_XFS_QUOTA */
 
+DEFINE_SCRUB_NLINKS_DIFF_EVENT(xrep_nlinks_update_inode);
+DEFINE_SCRUB_NLINKS_DIFF_EVENT(xrep_nlinks_unfixable_inode);
+
+TRACE_EVENT(xrep_rmap_live_update,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, unsigned int op,
+		 const struct xfs_rmap_update_params *p),
+	TP_ARGS(mp, agno, op, p),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(unsigned int, op)
+		__field(xfs_agblock_t, agbno)
+		__field(xfs_extlen_t, len)
+		__field(uint64_t, owner)
+		__field(uint64_t, offset)
+		__field(unsigned int, flags)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->op = op;
+		__entry->agbno = p->startblock;
+		__entry->len = p->blockcount;
+		xfs_owner_info_unpack(&p->oinfo, &__entry->owner,
+				&__entry->offset, &__entry->flags);
+		if (p->unwritten)
+			__entry->flags |= XFS_RMAP_UNWRITTEN;
+	),
+	TP_printk("dev %d:%d agno 0x%x op %d agbno 0x%x fsbcount 0x%x owner 0x%llx fileoff 0x%llx flags 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->op,
+		  __entry->agbno,
+		  __entry->len,
+		  __entry->owner,
+		  __entry->offset,
+		  __entry->flags)
+);
+
+TRACE_EVENT(xrep_tempfile_create,
+	TP_PROTO(struct xfs_scrub *sc),
+	TP_ARGS(sc),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(unsigned int, type)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_ino_t, inum)
+		__field(unsigned int, gen)
+		__field(unsigned int, flags)
+		__field(xfs_ino_t, temp_inum)
+	),
+	TP_fast_assign(
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->ino = sc->file ? XFS_I(file_inode(sc->file))->i_ino : 0;
+		__entry->type = sc->sm->sm_type;
+		__entry->agno = sc->sm->sm_agno;
+		__entry->inum = sc->sm->sm_ino;
+		__entry->gen = sc->sm->sm_gen;
+		__entry->flags = sc->sm->sm_flags;
+		__entry->temp_inum = sc->tempip->i_ino;
+	),
+	TP_printk("dev %d:%d ino 0x%llx type %s inum 0x%llx gen 0x%x flags 0x%x temp_inum 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __print_symbolic(__entry->type, XFS_SCRUB_TYPE_STRINGS),
+		  __entry->inum,
+		  __entry->gen,
+		  __entry->flags,
+		  __entry->temp_inum)
+);
+
+DECLARE_EVENT_CLASS(xrep_tempfile_class,
+	TP_PROTO(struct xfs_scrub *sc, int whichfork,
+		 struct xfs_bmbt_irec *irec),
+	TP_ARGS(sc, whichfork, irec),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(int, whichfork)
+		__field(xfs_fileoff_t, lblk)
+		__field(xfs_filblks_t, len)
+		__field(xfs_fsblock_t, pblk)
+		__field(int, state)
+	),
+	TP_fast_assign(
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->ino = sc->tempip->i_ino;
+		__entry->whichfork = whichfork;
+		__entry->lblk = irec->br_startoff;
+		__entry->len = irec->br_blockcount;
+		__entry->pblk = irec->br_startblock;
+		__entry->state = irec->br_state;
+	),
+	TP_printk("dev %d:%d ino 0x%llx whichfork %s fileoff 0x%llx fsbcount 0x%llx startblock 0x%llx state %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __print_symbolic(__entry->whichfork, XFS_WHICHFORK_STRINGS),
+		  __entry->lblk,
+		  __entry->len,
+		  __entry->pblk,
+		  __entry->state)
+);
+#define DEFINE_XREP_TEMPFILE_EVENT(name) \
+DEFINE_EVENT(xrep_tempfile_class, name, \
+	TP_PROTO(struct xfs_scrub *sc, int whichfork, \
+		 struct xfs_bmbt_irec *irec), \
+	TP_ARGS(sc, whichfork, irec))
+DEFINE_XREP_TEMPFILE_EVENT(xrep_tempfile_prealloc);
+DEFINE_XREP_TEMPFILE_EVENT(xrep_tempfile_copyin);
+
+TRACE_EVENT(xreap_ifork_extent,
+	TP_PROTO(struct xfs_scrub *sc, struct xfs_inode *ip, int whichfork,
+		 const struct xfs_bmbt_irec *irec),
+	TP_ARGS(sc, ip, whichfork, irec),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(int, whichfork)
+		__field(xfs_fileoff_t, fileoff)
+		__field(xfs_filblks_t, len)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, agbno)
+		__field(int, state)
+	),
+	TP_fast_assign(
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->whichfork = whichfork;
+		__entry->fileoff = irec->br_startoff;
+		__entry->len = irec->br_blockcount;
+		__entry->agno = XFS_FSB_TO_AGNO(sc->mp, irec->br_startblock);
+		__entry->agbno = XFS_FSB_TO_AGBNO(sc->mp, irec->br_startblock);
+		__entry->state = irec->br_state;
+	),
+	TP_printk("dev %d:%d ip 0x%llx whichfork %s agno 0x%x agbno 0x%x fileoff 0x%llx fsbcount 0x%llx state 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __print_symbolic(__entry->whichfork, XFS_WHICHFORK_STRINGS),
+		  __entry->agno,
+		  __entry->agbno,
+		  __entry->fileoff,
+		  __entry->len,
+		  __entry->state)
+);
+
+TRACE_EVENT(xreap_bmapi_binval_scan,
+	TP_PROTO(struct xfs_scrub *sc, const struct xfs_bmbt_irec *irec,
+		 xfs_extlen_t scan_blocks),
+	TP_ARGS(sc, irec, scan_blocks),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_filblks_t, len)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, agbno)
+		__field(xfs_extlen_t, scan_blocks)
+	),
+	TP_fast_assign(
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->len = irec->br_blockcount;
+		__entry->agno = XFS_FSB_TO_AGNO(sc->mp, irec->br_startblock);
+		__entry->agbno = XFS_FSB_TO_AGBNO(sc->mp, irec->br_startblock);
+		__entry->scan_blocks = scan_blocks;
+	),
+	TP_printk("dev %d:%d agno 0x%x agbno 0x%x fsbcount 0x%llx scan_blocks 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->agbno,
+		  __entry->len,
+		  __entry->scan_blocks)
+);
+
+TRACE_EVENT(xrep_xattr_recover_leafblock,
+	TP_PROTO(struct xfs_inode *ip, xfs_dablk_t dabno, uint16_t magic),
+	TP_ARGS(ip, dabno, magic),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_dablk_t, dabno)
+		__field(uint16_t, magic)
+	),
+	TP_fast_assign(
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->dabno = dabno;
+		__entry->magic = magic;
+	),
+	TP_printk("dev %d:%d ino 0x%llx dablk 0x%x magic 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->dabno,
+		  __entry->magic)
+);
+
+DECLARE_EVENT_CLASS(xrep_xattr_salvage_class,
+	TP_PROTO(struct xfs_inode *ip, unsigned int flags, char *name,
+		 unsigned int namelen, unsigned int valuelen),
+	TP_ARGS(ip, flags, name, namelen, valuelen),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(unsigned int, flags)
+		__field(unsigned int, namelen)
+		__dynamic_array(char, name, namelen)
+		__field(unsigned int, valuelen)
+	),
+	TP_fast_assign(
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->flags = flags;
+		__entry->namelen = namelen;
+		memcpy(__get_str(name), name, namelen);
+		__entry->valuelen = valuelen;
+	),
+	TP_printk("dev %d:%d ino 0x%llx flags %s name '%.*s' valuelen 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		   __print_flags(__entry->flags, "|", XFS_ATTR_NAMESPACE_STR),
+		  __entry->namelen,
+		  __get_str(name),
+		  __entry->valuelen)
+);
+#define DEFINE_XREP_XATTR_SALVAGE_EVENT(name) \
+DEFINE_EVENT(xrep_xattr_salvage_class, name, \
+	TP_PROTO(struct xfs_inode *ip, unsigned int flags, char *name, \
+		 unsigned int namelen, unsigned int valuelen), \
+	TP_ARGS(ip, flags, name, namelen, valuelen))
+DEFINE_XREP_XATTR_SALVAGE_EVENT(xrep_xattr_salvage_rec);
+DEFINE_XREP_XATTR_SALVAGE_EVENT(xrep_xattr_insert_rec);
+DEFINE_XREP_XATTR_SALVAGE_EVENT(xrep_parent_stash_xattr);
+DEFINE_XREP_XATTR_SALVAGE_EVENT(xrep_parent_insert_xattr);
+
+DECLARE_EVENT_CLASS(xrep_pptr_salvage_class,
+	TP_PROTO(struct xfs_inode *ip, unsigned int flags, const void *name,
+		 unsigned int namelen, const void *value, unsigned int valuelen),
+	TP_ARGS(ip, flags, name, namelen, value, valuelen),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_ino_t, parent_ino)
+		__field(unsigned int, parent_gen)
+		__field(unsigned int, namelen)
+		__dynamic_array(char, name, namelen)
+	),
+	TP_fast_assign(
+		const struct xfs_parent_rec	*rec = value;
+
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->parent_ino = be64_to_cpu(rec->p_ino);
+		__entry->parent_gen = be32_to_cpu(rec->p_gen);
+		__entry->namelen = namelen;
+		memcpy(__get_str(name), name, namelen);
+	),
+	TP_printk("dev %d:%d ino 0x%llx parent_ino 0x%llx parent_gen 0x%x name '%.*s'",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->parent_ino,
+		  __entry->parent_gen,
+		  __entry->namelen,
+		  __get_str(name))
+)
+#define DEFINE_XREP_PPTR_SALVAGE_EVENT(name) \
+DEFINE_EVENT(xrep_pptr_salvage_class, name, \
+	TP_PROTO(struct xfs_inode *ip, unsigned int flags, const void *name, \
+		 unsigned int namelen, const void *value, unsigned int valuelen), \
+	TP_ARGS(ip, flags, name, namelen, value, valuelen))
+DEFINE_XREP_PPTR_SALVAGE_EVENT(xrep_xattr_salvage_pptr);
+DEFINE_XREP_PPTR_SALVAGE_EVENT(xrep_xattr_insert_pptr);
+
+TRACE_EVENT(xrep_xattr_class,
+	TP_PROTO(struct xfs_inode *ip, struct xfs_inode *arg_ip),
+	TP_ARGS(ip, arg_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_ino_t, src_ino)
+	),
+	TP_fast_assign(
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->src_ino = arg_ip->i_ino;
+	),
+	TP_printk("dev %d:%d ino 0x%llx src 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->src_ino)
+)
+#define DEFINE_XREP_XATTR_EVENT(name) \
+DEFINE_EVENT(xrep_xattr_class, name, \
+	TP_PROTO(struct xfs_inode *ip, struct xfs_inode *arg_ip), \
+	TP_ARGS(ip, arg_ip))
+DEFINE_XREP_XATTR_EVENT(xrep_xattr_rebuild_tree);
+DEFINE_XREP_XATTR_EVENT(xrep_xattr_reset_fork);
+DEFINE_XREP_XATTR_EVENT(xrep_xattr_full_reset);
+
+DECLARE_EVENT_CLASS(xrep_xattr_pptr_scan_class,
+	TP_PROTO(struct xfs_inode *ip, const struct xfs_inode *dp,
+		 const struct xfs_name *name),
+	TP_ARGS(ip, dp, name),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_ino_t, parent_ino)
+		__field(unsigned int, parent_gen)
+		__field(unsigned int, namelen)
+		__dynamic_array(char, name, name->len)
+	),
+	TP_fast_assign(
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->parent_ino = dp->i_ino;
+		__entry->parent_gen = VFS_IC(dp)->i_generation;
+		__entry->namelen = name->len;
+		memcpy(__get_str(name), name->name, name->len);
+	),
+	TP_printk("dev %d:%d ino 0x%llx parent_ino 0x%llx parent_gen 0x%x name '%.*s'",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->parent_ino,
+		  __entry->parent_gen,
+		  __entry->namelen,
+		  __get_str(name))
+)
+#define DEFINE_XREP_XATTR_PPTR_SCAN_EVENT(name) \
+DEFINE_EVENT(xrep_xattr_pptr_scan_class, name, \
+	TP_PROTO(struct xfs_inode *ip, const struct xfs_inode *dp, \
+		 const struct xfs_name *name), \
+	TP_ARGS(ip, dp, name))
+DEFINE_XREP_XATTR_PPTR_SCAN_EVENT(xrep_xattr_stash_parentadd);
+DEFINE_XREP_XATTR_PPTR_SCAN_EVENT(xrep_xattr_stash_parentremove);
+
+TRACE_EVENT(xrep_dir_recover_dirblock,
+	TP_PROTO(struct xfs_inode *dp, xfs_dablk_t dabno, uint32_t magic,
+		 uint32_t magic_guess),
+	TP_ARGS(dp, dabno, magic, magic_guess),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, dir_ino)
+		__field(xfs_dablk_t, dabno)
+		__field(uint32_t, magic)
+		__field(uint32_t, magic_guess)
+	),
+	TP_fast_assign(
+		__entry->dev = dp->i_mount->m_super->s_dev;
+		__entry->dir_ino = dp->i_ino;
+		__entry->dabno = dabno;
+		__entry->magic = magic;
+		__entry->magic_guess = magic_guess;
+	),
+	TP_printk("dev %d:%d dir 0x%llx dablk 0x%x magic 0x%x magic_guess 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->dir_ino,
+		  __entry->dabno,
+		  __entry->magic,
+		  __entry->magic_guess)
+);
+
+DECLARE_EVENT_CLASS(xrep_dir_class,
+	TP_PROTO(struct xfs_inode *dp, xfs_ino_t parent_ino),
+	TP_ARGS(dp, parent_ino),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, dir_ino)
+		__field(xfs_ino_t, parent_ino)
+	),
+	TP_fast_assign(
+		__entry->dev = dp->i_mount->m_super->s_dev;
+		__entry->dir_ino = dp->i_ino;
+		__entry->parent_ino = parent_ino;
+	),
+	TP_printk("dev %d:%d dir 0x%llx parent 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->dir_ino,
+		  __entry->parent_ino)
+)
+#define DEFINE_XREP_DIR_EVENT(name) \
+DEFINE_EVENT(xrep_dir_class, name, \
+	TP_PROTO(struct xfs_inode *dp, xfs_ino_t parent_ino), \
+	TP_ARGS(dp, parent_ino))
+DEFINE_XREP_DIR_EVENT(xrep_dir_rebuild_tree);
+DEFINE_XREP_DIR_EVENT(xrep_dir_reset_fork);
+DEFINE_XREP_DIR_EVENT(xrep_parent_reset_dotdot);
+
+DECLARE_EVENT_CLASS(xrep_dirent_class,
+	TP_PROTO(struct xfs_inode *dp, const struct xfs_name *name,
+		 xfs_ino_t ino),
+	TP_ARGS(dp, name, ino),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, dir_ino)
+		__field(unsigned int, namelen)
+		__dynamic_array(char, name, name->len)
+		__field(xfs_ino_t, ino)
+		__field(uint8_t, ftype)
+	),
+	TP_fast_assign(
+		__entry->dev = dp->i_mount->m_super->s_dev;
+		__entry->dir_ino = dp->i_ino;
+		__entry->namelen = name->len;
+		memcpy(__get_str(name), name->name, name->len);
+		__entry->ino = ino;
+		__entry->ftype = name->type;
+	),
+	TP_printk("dev %d:%d dir 0x%llx ftype %s name '%.*s' ino 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->dir_ino,
+		  __print_symbolic(__entry->ftype, XFS_DIR3_FTYPE_STR),
+		  __entry->namelen,
+		  __get_str(name),
+		  __entry->ino)
+)
+#define DEFINE_XREP_DIRENT_EVENT(name) \
+DEFINE_EVENT(xrep_dirent_class, name, \
+	TP_PROTO(struct xfs_inode *dp, const struct xfs_name *name, \
+		 xfs_ino_t ino), \
+	TP_ARGS(dp, name, ino))
+DEFINE_XREP_DIRENT_EVENT(xrep_dir_salvage_entry);
+DEFINE_XREP_DIRENT_EVENT(xrep_dir_stash_createname);
+DEFINE_XREP_DIRENT_EVENT(xrep_dir_replay_createname);
+DEFINE_XREP_DIRENT_EVENT(xrep_adoption_reparent);
+DEFINE_XREP_DIRENT_EVENT(xrep_dir_stash_removename);
+DEFINE_XREP_DIRENT_EVENT(xrep_dir_replay_removename);
+
+DECLARE_EVENT_CLASS(xrep_adoption_class,
+	TP_PROTO(struct xfs_inode *dp, struct xfs_inode *ip, bool moved),
+	TP_ARGS(dp, ip, moved),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, dir_ino)
+		__field(xfs_ino_t, child_ino)
+		__field(bool, moved)
+	),
+	TP_fast_assign(
+		__entry->dev = dp->i_mount->m_super->s_dev;
+		__entry->dir_ino = dp->i_ino;
+		__entry->child_ino = ip->i_ino;
+		__entry->moved = moved;
+	),
+	TP_printk("dev %d:%d dir 0x%llx child 0x%llx moved? %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->dir_ino,
+		  __entry->child_ino,
+		  __entry->moved)
+);
+#define DEFINE_XREP_ADOPTION_EVENT(name) \
+DEFINE_EVENT(xrep_adoption_class, name, \
+	TP_PROTO(struct xfs_inode *dp, struct xfs_inode *ip, bool moved), \
+	TP_ARGS(dp, ip, moved))
+DEFINE_XREP_ADOPTION_EVENT(xrep_adoption_trans_roll);
+
+DECLARE_EVENT_CLASS(xrep_parent_salvage_class,
+	TP_PROTO(struct xfs_inode *dp, xfs_ino_t ino),
+	TP_ARGS(dp, ino),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, dir_ino)
+		__field(xfs_ino_t, ino)
+	),
+	TP_fast_assign(
+		__entry->dev = dp->i_mount->m_super->s_dev;
+		__entry->dir_ino = dp->i_ino;
+		__entry->ino = ino;
+	),
+	TP_printk("dev %d:%d dir 0x%llx parent 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->dir_ino,
+		  __entry->ino)
+)
+#define DEFINE_XREP_PARENT_SALVAGE_EVENT(name) \
+DEFINE_EVENT(xrep_parent_salvage_class, name, \
+	TP_PROTO(struct xfs_inode *dp, xfs_ino_t ino), \
+	TP_ARGS(dp, ino))
+DEFINE_XREP_PARENT_SALVAGE_EVENT(xrep_dir_salvaged_parent);
+DEFINE_XREP_PARENT_SALVAGE_EVENT(xrep_findparent_dirent);
+DEFINE_XREP_PARENT_SALVAGE_EVENT(xrep_findparent_from_dcache);
+
+DECLARE_EVENT_CLASS(xrep_pptr_class,
+	TP_PROTO(struct xfs_inode *ip, const struct xfs_name *name,
+		 const struct xfs_parent_rec *pptr),
+	TP_ARGS(ip, name, pptr),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_ino_t, parent_ino)
+		__field(unsigned int, parent_gen)
+		__field(unsigned int, namelen)
+		__dynamic_array(char, name, name->len)
+	),
+	TP_fast_assign(
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->parent_ino = be64_to_cpu(pptr->p_ino);
+		__entry->parent_gen = be32_to_cpu(pptr->p_gen);
+		__entry->namelen = name->len;
+		memcpy(__get_str(name), name->name, name->len);
+	),
+	TP_printk("dev %d:%d ino 0x%llx parent_ino 0x%llx parent_gen 0x%x name '%.*s'",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->parent_ino,
+		  __entry->parent_gen,
+		  __entry->namelen,
+		  __get_str(name))
+)
+#define DEFINE_XREP_PPTR_EVENT(name) \
+DEFINE_EVENT(xrep_pptr_class, name, \
+	TP_PROTO(struct xfs_inode *ip, const struct xfs_name *name, \
+		 const struct xfs_parent_rec *pptr), \
+	TP_ARGS(ip, name, pptr))
+DEFINE_XREP_PPTR_EVENT(xrep_xattr_replay_parentadd);
+DEFINE_XREP_PPTR_EVENT(xrep_xattr_replay_parentremove);
+DEFINE_XREP_PPTR_EVENT(xrep_parent_replay_parentadd);
+DEFINE_XREP_PPTR_EVENT(xrep_parent_replay_parentremove);
+
+DECLARE_EVENT_CLASS(xrep_pptr_scan_class,
+	TP_PROTO(struct xfs_inode *ip, const struct xfs_inode *dp,
+		 const struct xfs_name *name),
+	TP_ARGS(ip, dp, name),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_ino_t, parent_ino)
+		__field(unsigned int, parent_gen)
+		__field(unsigned int, namelen)
+		__dynamic_array(char, name, name->len)
+	),
+	TP_fast_assign(
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->parent_ino = dp->i_ino;
+		__entry->parent_gen = VFS_IC(dp)->i_generation;
+		__entry->namelen = name->len;
+		memcpy(__get_str(name), name->name, name->len);
+	),
+	TP_printk("dev %d:%d ino 0x%llx parent_ino 0x%llx parent_gen 0x%x name '%.*s'",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->parent_ino,
+		  __entry->parent_gen,
+		  __entry->namelen,
+		  __get_str(name))
+)
+#define DEFINE_XREP_PPTR_SCAN_EVENT(name) \
+DEFINE_EVENT(xrep_pptr_scan_class, name, \
+	TP_PROTO(struct xfs_inode *ip, const struct xfs_inode *dp, \
+		 const struct xfs_name *name), \
+	TP_ARGS(ip, dp, name))
+DEFINE_XREP_PPTR_SCAN_EVENT(xrep_parent_stash_parentadd);
+DEFINE_XREP_PPTR_SCAN_EVENT(xrep_parent_stash_parentremove);
+
+TRACE_EVENT(xrep_nlinks_set_record,
+	TP_PROTO(struct xfs_mount *mp, xfs_ino_t ino,
+		 const struct xchk_nlink *obs),
+	TP_ARGS(mp, ino, obs),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_nlink_t, parents)
+		__field(xfs_nlink_t, backrefs)
+		__field(xfs_nlink_t, children)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->ino = ino;
+		__entry->parents = obs->parents;
+		__entry->backrefs = obs->backrefs;
+		__entry->children = obs->children;
+	),
+	TP_printk("dev %d:%d ino 0x%llx parents %u backrefs %u children %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->parents,
+		  __entry->backrefs,
+		  __entry->children)
+);
+
+DECLARE_EVENT_CLASS(xrep_dentry_class,
+	TP_PROTO(struct xfs_mount *mp, const struct dentry *dentry),
+	TP_ARGS(mp, dentry),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(unsigned int, flags)
+		__field(unsigned long, ino)
+		__field(bool, positive)
+		__field(unsigned long, parent_ino)
+		__field(unsigned int, namelen)
+		__dynamic_array(char, name, dentry->d_name.len)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->flags = dentry->d_flags;
+		__entry->positive = d_is_positive(dentry);
+		if (dentry->d_parent && d_inode(dentry->d_parent))
+			__entry->parent_ino = d_inode(dentry->d_parent)->i_ino;
+		else
+			__entry->parent_ino = -1UL;
+		__entry->ino = d_inode(dentry) ? d_inode(dentry)->i_ino : 0;
+		__entry->namelen = dentry->d_name.len;
+		memcpy(__get_str(name), dentry->d_name.name, dentry->d_name.len);
+	),
+	TP_printk("dev %d:%d flags 0x%x positive? %d parent_ino 0x%lx ino 0x%lx name '%.*s'",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->flags,
+		  __entry->positive,
+		  __entry->parent_ino,
+		  __entry->ino,
+		  __entry->namelen,
+		  __get_str(name))
+);
+#define DEFINE_REPAIR_DENTRY_EVENT(name) \
+DEFINE_EVENT(xrep_dentry_class, name, \
+	TP_PROTO(struct xfs_mount *mp, const struct dentry *dentry), \
+	TP_ARGS(mp, dentry))
+DEFINE_REPAIR_DENTRY_EVENT(xrep_adoption_check_child);
+DEFINE_REPAIR_DENTRY_EVENT(xrep_adoption_invalidate_child);
+DEFINE_REPAIR_DENTRY_EVENT(xrep_dirtree_delete_child);
+
+TRACE_EVENT(xrep_symlink_salvage_target,
+	TP_PROTO(struct xfs_inode *ip, char *target, unsigned int targetlen),
+	TP_ARGS(ip, target, targetlen),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(unsigned int, targetlen)
+		__dynamic_array(char, target, targetlen + 1)
+	),
+	TP_fast_assign(
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->targetlen = targetlen;
+		memcpy(__get_str(target), target, targetlen);
+		__get_str(target)[targetlen] = 0;
+	),
+	TP_printk("dev %d:%d ip 0x%llx target '%.*s'",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->targetlen,
+		  __get_str(target))
+);
+
+DECLARE_EVENT_CLASS(xrep_symlink_class,
+	TP_PROTO(struct xfs_inode *ip),
+	TP_ARGS(ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+	),
+	TP_fast_assign(
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		__entry->ino = ip->i_ino;
+	),
+	TP_printk("dev %d:%d ip 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino)
+);
+
+#define DEFINE_XREP_SYMLINK_EVENT(name) \
+DEFINE_EVENT(xrep_symlink_class, name, \
+	TP_PROTO(struct xfs_inode *ip), \
+	TP_ARGS(ip))
+DEFINE_XREP_SYMLINK_EVENT(xrep_symlink_rebuild);
+DEFINE_XREP_SYMLINK_EVENT(xrep_symlink_reset_fork);
+
+TRACE_EVENT(xrep_iunlink_visit,
+	TP_PROTO(struct xfs_perag *pag, unsigned int bucket,
+		 xfs_agino_t bucket_agino, struct xfs_inode *ip),
+	TP_ARGS(pag, bucket, bucket_agino, ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agino_t, agino)
+		__field(unsigned int, bucket)
+		__field(xfs_agino_t, bucket_agino)
+		__field(xfs_agino_t, prev_agino)
+		__field(xfs_agino_t, next_agino)
+	),
+	TP_fast_assign(
+		__entry->dev = pag->pag_mount->m_super->s_dev;
+		__entry->agno = pag->pag_agno;
+		__entry->agino = XFS_INO_TO_AGINO(pag->pag_mount, ip->i_ino);
+		__entry->bucket = bucket;
+		__entry->bucket_agino = bucket_agino;
+		__entry->prev_agino = ip->i_prev_unlinked;
+		__entry->next_agino = ip->i_next_unlinked;
+	),
+	TP_printk("dev %d:%d agno 0x%x bucket %u agino 0x%x bucket_agino 0x%x prev_agino 0x%x next_agino 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->bucket,
+		  __entry->agino,
+		  __entry->bucket_agino,
+		  __entry->prev_agino,
+		  __entry->next_agino)
+);
+
+TRACE_EVENT(xrep_iunlink_reload_next,
+	TP_PROTO(struct xfs_inode *ip, xfs_agino_t prev_agino),
+	TP_ARGS(ip, prev_agino),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agino_t, agino)
+		__field(xfs_agino_t, old_prev_agino)
+		__field(xfs_agino_t, prev_agino)
+		__field(xfs_agino_t, next_agino)
+		__field(unsigned int, nlink)
+	),
+	TP_fast_assign(
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		__entry->agno = XFS_INO_TO_AGNO(ip->i_mount, ip->i_ino);
+		__entry->agino = XFS_INO_TO_AGINO(ip->i_mount, ip->i_ino);
+		__entry->old_prev_agino = ip->i_prev_unlinked;
+		__entry->prev_agino = prev_agino;
+		__entry->next_agino = ip->i_next_unlinked;
+		__entry->nlink = VFS_I(ip)->i_nlink;
+	),
+	TP_printk("dev %d:%d agno 0x%x bucket %u agino 0x%x nlink %u old_prev_agino %u prev_agino 0x%x next_agino 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->agino % XFS_AGI_UNLINKED_BUCKETS,
+		  __entry->agino,
+		  __entry->nlink,
+		  __entry->old_prev_agino,
+		  __entry->prev_agino,
+		  __entry->next_agino)
+);
+
+TRACE_EVENT(xrep_iunlink_reload_ondisk,
+	TP_PROTO(struct xfs_inode *ip),
+	TP_ARGS(ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agino_t, agino)
+		__field(unsigned int, nlink)
+		__field(xfs_agino_t, next_agino)
+	),
+	TP_fast_assign(
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		__entry->agno = XFS_INO_TO_AGNO(ip->i_mount, ip->i_ino);
+		__entry->agino = XFS_INO_TO_AGINO(ip->i_mount, ip->i_ino);
+		__entry->nlink = VFS_I(ip)->i_nlink;
+		__entry->next_agino = ip->i_next_unlinked;
+	),
+	TP_printk("dev %d:%d agno 0x%x bucket %u agino 0x%x nlink %u next_agino 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->agino % XFS_AGI_UNLINKED_BUCKETS,
+		  __entry->agino,
+		  __entry->nlink,
+		  __entry->next_agino)
+);
+
+TRACE_EVENT(xrep_iunlink_walk_ondisk_bucket,
+	TP_PROTO(struct xfs_perag *pag, unsigned int bucket,
+		 xfs_agino_t prev_agino, xfs_agino_t next_agino),
+	TP_ARGS(pag, bucket, prev_agino, next_agino),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(unsigned int, bucket)
+		__field(xfs_agino_t, prev_agino)
+		__field(xfs_agino_t, next_agino)
+	),
+	TP_fast_assign(
+		__entry->dev = pag->pag_mount->m_super->s_dev;
+		__entry->agno = pag->pag_agno;
+		__entry->bucket = bucket;
+		__entry->prev_agino = prev_agino;
+		__entry->next_agino = next_agino;
+	),
+	TP_printk("dev %d:%d agno 0x%x bucket %u prev_agino 0x%x next_agino 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->bucket,
+		  __entry->prev_agino,
+		  __entry->next_agino)
+);
+
+DECLARE_EVENT_CLASS(xrep_iunlink_resolve_class,
+	TP_PROTO(struct xfs_perag *pag, unsigned int bucket,
+		 xfs_agino_t prev_agino, xfs_agino_t next_agino),
+	TP_ARGS(pag, bucket, prev_agino, next_agino),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(unsigned int, bucket)
+		__field(xfs_agino_t, prev_agino)
+		__field(xfs_agino_t, next_agino)
+	),
+	TP_fast_assign(
+		__entry->dev = pag->pag_mount->m_super->s_dev;
+		__entry->agno = pag->pag_agno;
+		__entry->bucket = bucket;
+		__entry->prev_agino = prev_agino;
+		__entry->next_agino = next_agino;
+	),
+	TP_printk("dev %d:%d agno 0x%x bucket %u prev_agino 0x%x next_agino 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->bucket,
+		  __entry->prev_agino,
+		  __entry->next_agino)
+);
+#define DEFINE_REPAIR_IUNLINK_RESOLVE_EVENT(name) \
+DEFINE_EVENT(xrep_iunlink_resolve_class, name, \
+	TP_PROTO(struct xfs_perag *pag, unsigned int bucket, \
+		 xfs_agino_t prev_agino, xfs_agino_t next_agino), \
+	TP_ARGS(pag, bucket, prev_agino, next_agino))
+DEFINE_REPAIR_IUNLINK_RESOLVE_EVENT(xrep_iunlink_resolve_uncached);
+DEFINE_REPAIR_IUNLINK_RESOLVE_EVENT(xrep_iunlink_resolve_wronglist);
+DEFINE_REPAIR_IUNLINK_RESOLVE_EVENT(xrep_iunlink_resolve_nolist);
+DEFINE_REPAIR_IUNLINK_RESOLVE_EVENT(xrep_iunlink_resolve_ok);
+
+TRACE_EVENT(xrep_iunlink_relink_next,
+	TP_PROTO(struct xfs_inode *ip, xfs_agino_t next_agino),
+	TP_ARGS(ip, next_agino),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agino_t, agino)
+		__field(xfs_agino_t, next_agino)
+		__field(xfs_agino_t, new_next_agino)
+	),
+	TP_fast_assign(
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		__entry->agno = XFS_INO_TO_AGNO(ip->i_mount, ip->i_ino);
+		__entry->agino = XFS_INO_TO_AGINO(ip->i_mount, ip->i_ino);
+		__entry->next_agino = ip->i_next_unlinked;
+		__entry->new_next_agino = next_agino;
+	),
+	TP_printk("dev %d:%d agno 0x%x bucket %u agino 0x%x next_agino 0x%x -> 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->agino % XFS_AGI_UNLINKED_BUCKETS,
+		  __entry->agino,
+		  __entry->next_agino,
+		  __entry->new_next_agino)
+);
+
+TRACE_EVENT(xrep_iunlink_relink_prev,
+	TP_PROTO(struct xfs_inode *ip, xfs_agino_t prev_agino),
+	TP_ARGS(ip, prev_agino),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agino_t, agino)
+		__field(xfs_agino_t, prev_agino)
+		__field(xfs_agino_t, new_prev_agino)
+	),
+	TP_fast_assign(
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		__entry->agno = XFS_INO_TO_AGNO(ip->i_mount, ip->i_ino);
+		__entry->agino = XFS_INO_TO_AGINO(ip->i_mount, ip->i_ino);
+		__entry->prev_agino = ip->i_prev_unlinked;
+		__entry->new_prev_agino = prev_agino;
+	),
+	TP_printk("dev %d:%d agno 0x%x bucket %u agino 0x%x prev_agino 0x%x -> 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->agino % XFS_AGI_UNLINKED_BUCKETS,
+		  __entry->agino,
+		  __entry->prev_agino,
+		  __entry->new_prev_agino)
+);
+
+TRACE_EVENT(xrep_iunlink_add_to_bucket,
+	TP_PROTO(struct xfs_perag *pag, unsigned int bucket,
+		 xfs_agino_t agino, xfs_agino_t curr_head),
+	TP_ARGS(pag, bucket, agino, curr_head),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(unsigned int, bucket)
+		__field(xfs_agino_t, agino)
+		__field(xfs_agino_t, next_agino)
+	),
+	TP_fast_assign(
+		__entry->dev = pag->pag_mount->m_super->s_dev;
+		__entry->agno = pag->pag_agno;
+		__entry->bucket = bucket;
+		__entry->agino = agino;
+		__entry->next_agino = curr_head;
+	),
+	TP_printk("dev %d:%d agno 0x%x bucket %u agino 0x%x next_agino 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->bucket,
+		  __entry->agino,
+		  __entry->next_agino)
+);
+
+TRACE_EVENT(xrep_iunlink_commit_bucket,
+	TP_PROTO(struct xfs_perag *pag, unsigned int bucket,
+		 xfs_agino_t old_agino, xfs_agino_t agino),
+	TP_ARGS(pag, bucket, old_agino, agino),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(unsigned int, bucket)
+		__field(xfs_agino_t, old_agino)
+		__field(xfs_agino_t, agino)
+	),
+	TP_fast_assign(
+		__entry->dev = pag->pag_mount->m_super->s_dev;
+		__entry->agno = pag->pag_agno;
+		__entry->bucket = bucket;
+		__entry->old_agino = old_agino;
+		__entry->agino = agino;
+	),
+	TP_printk("dev %d:%d agno 0x%x bucket %u agino 0x%x -> 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->bucket,
+		  __entry->old_agino,
+		  __entry->agino)
+);
+
+DEFINE_XCHK_DIRPATH_OUTCOME_EVENT(xrep_dirpath_set_outcome);
+DEFINE_XCHK_DIRTREE_EVENT(xrep_dirtree_delete_path);
+DEFINE_XCHK_DIRTREE_EVENT(xrep_dirtree_create_adoption);
+DEFINE_XCHK_DIRTREE_EVALUATE_EVENT(xrep_dirtree_decided_fate);
+
 #endif /* IS_ENABLED(CONFIG_XFS_ONLINE_REPAIR) */
 
 #endif /* _TRACE_XFS_SCRUB_TRACE_H */
diff --git a/fs/xfs/scrub/xfarray.c b/fs/xfs/scrub/xfarray.c
index f0f532c10a..cdd13ed9c5 100644
--- a/fs/xfs/scrub/xfarray.c
+++ b/fs/xfs/scrub/xfarray.c
@@ -7,16 +7,16 @@
 #include "xfs_fs.h"
 #include "xfs_shared.h"
 #include "xfs_format.h"
+#include "scrub/scrub.h"
 #include "scrub/xfile.h"
 #include "scrub/xfarray.h"
-#include "scrub/scrub.h"
 #include "scrub/trace.h"
 
 /*
  * Large Arrays of Fixed-Size Records
  * ==================================
  *
- * This memory array uses an xfile (which itself is a memfd "file") to store
+ * This memory array uses an xfile (which itself is a shmem file) to store
  * large numbers of fixed-size records in memory that can be paged out.  This
  * puts less stress on the memory reclaim algorithms during an online repair
  * because we don't have to pin so much memory.  However, array access is less
@@ -136,7 +136,7 @@ xfarray_load(
 	if (idx >= array->nr)
 		return -ENODATA;
 
-	return xfile_obj_load(array->xfile, ptr, array->obj_size,
+	return xfile_load(array->xfile, ptr, array->obj_size,
 			xfarray_pos(array, idx));
 }
 
@@ -152,7 +152,7 @@ xfarray_is_unset(
 	if (array->unset_slots == 0)
 		return false;
 
-	error = xfile_obj_load(array->xfile, temp, array->obj_size, pos);
+	error = xfile_load(array->xfile, temp, array->obj_size, pos);
 	if (!error && xfarray_element_is_null(array, temp))
 		return true;
 
@@ -184,7 +184,7 @@ xfarray_unset(
 		return 0;
 
 	memset(temp, 0, array->obj_size);
-	error = xfile_obj_store(array->xfile, temp, array->obj_size, pos);
+	error = xfile_store(array->xfile, temp, array->obj_size, pos);
 	if (error)
 		return error;
 
@@ -209,7 +209,7 @@ xfarray_store(
 
 	ASSERT(!xfarray_element_is_null(array, ptr));
 
-	ret = xfile_obj_store(array->xfile, ptr, array->obj_size,
+	ret = xfile_store(array->xfile, ptr, array->obj_size,
 			xfarray_pos(array, idx));
 	if (ret)
 		return ret;
@@ -245,12 +245,12 @@ xfarray_store_anywhere(
 	for (pos = 0;
 	     pos < endpos && array->unset_slots > 0;
 	     pos += array->obj_size) {
-		error = xfile_obj_load(array->xfile, temp, array->obj_size,
+		error = xfile_load(array->xfile, temp, array->obj_size,
 				pos);
 		if (error || !xfarray_element_is_null(array, temp))
 			continue;
 
-		error = xfile_obj_store(array->xfile, ptr, array->obj_size,
+		error = xfile_store(array->xfile, ptr, array->obj_size,
 				pos);
 		if (error)
 			return error;
@@ -486,6 +486,9 @@ xfarray_sortinfo_alloc(
 
 	xfarray_sortinfo_lo(si)[0] = 0;
 	xfarray_sortinfo_hi(si)[0] = array->nr - 1;
+	si->relax = INIT_XCHK_RELAX;
+	if (flags & XFARRAY_SORT_KILLABLE)
+		si->relax.interruptible = false;
 
 	trace_xfarray_sort(si, nr_bytes);
 	*infop = si;
@@ -503,10 +506,7 @@ xfarray_sort_terminated(
 	 * few seconds so that we don't run afoul of the soft lockup watchdog
 	 * or RCU stall detector.
 	 */
-	cond_resched();
-
-	if ((si->flags & XFARRAY_SORT_KILLABLE) &&
-	    fatal_signal_pending(current)) {
+	if (xchk_maybe_relax(&si->relax)) {
 		if (*error == 0)
 			*error = -EINTR;
 		return true;
@@ -552,7 +552,7 @@ xfarray_isort(
 	trace_xfarray_isort(si, lo, hi);
 
 	xfarray_sort_bump_loads(si);
-	error = xfile_obj_load(si->array->xfile, scratch, len, lo_pos);
+	error = xfile_load(si->array->xfile, scratch, len, lo_pos);
 	if (error)
 		return error;
 
@@ -560,88 +560,45 @@ xfarray_isort(
 	sort(scratch, hi - lo + 1, si->array->obj_size, si->cmp_fn, NULL);
 
 	xfarray_sort_bump_stores(si);
-	return xfile_obj_store(si->array->xfile, scratch, len, lo_pos);
-}
-
-/* Grab a page for sorting records. */
-static inline int
-xfarray_sort_get_page(
-	struct xfarray_sortinfo	*si,
-	loff_t			pos,
-	uint64_t		len)
-{
-	int			error;
-
-	error = xfile_get_page(si->array->xfile, pos, len, &si->xfpage);
-	if (error)
-		return error;
-
-	/*
-	 * xfile pages must never be mapped into userspace, so we skip the
-	 * dcache flush when mapping the page.
-	 */
-	si->page_kaddr = kmap_local_page(si->xfpage.page);
-	return 0;
-}
-
-/* Release a page we grabbed for sorting records. */
-static inline int
-xfarray_sort_put_page(
-	struct xfarray_sortinfo	*si)
-{
-	if (!si->page_kaddr)
-		return 0;
-
-	kunmap_local(si->page_kaddr);
-	si->page_kaddr = NULL;
-
-	return xfile_put_page(si->array->xfile, &si->xfpage);
+	return xfile_store(si->array->xfile, scratch, len, lo_pos);
 }
 
-/* Decide if these records are eligible for in-page sorting. */
-static inline bool
-xfarray_want_pagesort(
-	struct xfarray_sortinfo	*si,
-	xfarray_idx_t		lo,
-	xfarray_idx_t		hi)
-{
-	pgoff_t			lo_page;
-	pgoff_t			hi_page;
-	loff_t			end_pos;
-
-	/* We can only map one page at a time. */
-	lo_page = xfarray_pos(si->array, lo) >> PAGE_SHIFT;
-	end_pos = xfarray_pos(si->array, hi) + si->array->obj_size - 1;
-	hi_page = end_pos >> PAGE_SHIFT;
-
-	return lo_page == hi_page;
-}
-
-/* Sort a bunch of records that all live in the same memory page. */
+/*
+ * Sort the records from lo to hi (inclusive) if they are all backed by the
+ * same memory folio.  Returns 1 if it sorted, 0 if it did not, or a negative
+ * errno.
+ */
 STATIC int
-xfarray_pagesort(
+xfarray_foliosort(
 	struct xfarray_sortinfo	*si,
 	xfarray_idx_t		lo,
 	xfarray_idx_t		hi)
 {
+	struct folio		*folio;
 	void			*startp;
 	loff_t			lo_pos = xfarray_pos(si->array, lo);
-	uint64_t		len = xfarray_pos(si->array, hi - lo);
-	int			error = 0;
+	uint64_t		len = xfarray_pos(si->array, hi - lo + 1);
 
-	trace_xfarray_pagesort(si, lo, hi);
+	/* No single folio could back this many records. */
+	if (len > XFILE_MAX_FOLIO_SIZE)
+		return 0;
 
 	xfarray_sort_bump_loads(si);
-	error = xfarray_sort_get_page(si, lo_pos, len);
-	if (error)
-		return error;
+	folio = xfile_get_folio(si->array->xfile, lo_pos, len, XFILE_ALLOC);
+	if (IS_ERR(folio))
+		return PTR_ERR(folio);
+	if (!folio)
+		return 0;
+
+	trace_xfarray_foliosort(si, lo, hi);
 
 	xfarray_sort_bump_heapsorts(si);
-	startp = si->page_kaddr + offset_in_page(lo_pos);
+	startp = folio_address(folio) + offset_in_folio(folio, lo_pos);
 	sort(startp, hi - lo + 1, si->array->obj_size, si->cmp_fn, NULL);
 
 	xfarray_sort_bump_stores(si);
-	return xfarray_sort_put_page(si);
+	xfile_put_folio(si->array->xfile, folio);
+	return 1;
 }
 
 /* Return a pointer to the xfarray pivot record within the sortinfo struct. */
@@ -829,63 +786,80 @@ xfarray_qsort_push(
 	return 0;
 }
 
+static inline void
+xfarray_sort_scan_done(
+	struct xfarray_sortinfo	*si)
+{
+	if (si->folio)
+		xfile_put_folio(si->array->xfile, si->folio);
+	si->folio = NULL;
+}
+
 /*
- * Load an element from the array into the first scratchpad and cache the page,
- * if possible.
+ * Cache the folio backing the start of the given array element.  If the array
+ * element is contained entirely within the folio, return a pointer to the
+ * cached folio.  Otherwise, load the element into the scratchpad and return a
+ * pointer to the scratchpad.
  */
 static inline int
-xfarray_sort_load_cached(
+xfarray_sort_scan(
 	struct xfarray_sortinfo	*si,
 	xfarray_idx_t		idx,
-	void			*ptr)
+	void			**ptrp)
 {
 	loff_t			idx_pos = xfarray_pos(si->array, idx);
-	pgoff_t			startpage;
-	pgoff_t			endpage;
 	int			error = 0;
 
-	/*
-	 * If this load would split a page, release the cached page, if any,
-	 * and perform a traditional read.
-	 */
-	startpage = idx_pos >> PAGE_SHIFT;
-	endpage = (idx_pos + si->array->obj_size - 1) >> PAGE_SHIFT;
-	if (startpage != endpage) {
-		error = xfarray_sort_put_page(si);
-		if (error)
-			return error;
+	if (xfarray_sort_terminated(si, &error))
+		return error;
 
-		if (xfarray_sort_terminated(si, &error))
-			return error;
+	trace_xfarray_sort_scan(si, idx);
 
-		return xfile_obj_load(si->array->xfile, ptr,
-				si->array->obj_size, idx_pos);
-	}
+	/* If the cached folio doesn't cover this index, release it. */
+	if (si->folio &&
+	    (idx < si->first_folio_idx || idx > si->last_folio_idx))
+		xfarray_sort_scan_done(si);
 
-	/* If the cached page is not the one we want, release it. */
-	if (xfile_page_cached(&si->xfpage) &&
-	    xfile_page_index(&si->xfpage) != startpage) {
-		error = xfarray_sort_put_page(si);
-		if (error)
-			return error;
+	/* Grab the first folio that backs this array element. */
+	if (!si->folio) {
+		struct folio	*folio;
+		loff_t		next_pos;
+
+		folio = xfile_get_folio(si->array->xfile, idx_pos,
+				si->array->obj_size, XFILE_ALLOC);
+		if (IS_ERR(folio))
+			return PTR_ERR(folio);
+		si->folio = folio;
+
+		si->first_folio_idx = xfarray_idx(si->array,
+				folio_pos(si->folio) + si->array->obj_size - 1);
+
+		next_pos = folio_pos(si->folio) + folio_size(si->folio);
+		si->last_folio_idx = xfarray_idx(si->array, next_pos - 1);
+		if (xfarray_pos(si->array, si->last_folio_idx + 1) > next_pos)
+			si->last_folio_idx--;
+
+		trace_xfarray_sort_scan(si, idx);
 	}
 
 	/*
-	 * If we don't have a cached page (and we know the load is contained
-	 * in a single page) then grab it.
+	 * If this folio still doesn't cover the desired element, it must cross
+	 * a folio boundary.  Read into the scratchpad and we're done.
 	 */
-	if (!xfile_page_cached(&si->xfpage)) {
-		if (xfarray_sort_terminated(si, &error))
-			return error;
+	if (idx < si->first_folio_idx || idx > si->last_folio_idx) {
+		void		*temp = xfarray_scratch(si->array);
 
-		error = xfarray_sort_get_page(si, startpage << PAGE_SHIFT,
-				PAGE_SIZE);
+		error = xfile_load(si->array->xfile, temp, si->array->obj_size,
+				idx_pos);
 		if (error)
 			return error;
+
+		*ptrp = temp;
+		return 0;
 	}
 
-	memcpy(ptr, si->page_kaddr + offset_in_page(idx_pos),
-			si->array->obj_size);
+	/* Otherwise return a pointer to the array element in the folio. */
+	*ptrp = folio_address(si->folio) + offset_in_folio(si->folio, idx_pos);
 	return 0;
 }
 
@@ -952,6 +926,8 @@ xfarray_sort(
 	pivot = xfarray_sortinfo_pivot(si);
 
 	while (si->stack_depth >= 0) {
+		int		ret;
+
 		lo = si_lo[si->stack_depth];
 		hi = si_hi[si->stack_depth];
 
@@ -964,13 +940,13 @@ xfarray_sort(
 		}
 
 		/*
-		 * If directly mapping the page and sorting can solve our
+		 * If directly mapping the folio and sorting can solve our
 		 * problems, we're done.
 		 */
-		if (xfarray_want_pagesort(si, lo, hi)) {
-			error = xfarray_pagesort(si, lo, hi);
-			if (error)
-				goto out_free;
+		ret = xfarray_foliosort(si, lo, hi);
+		if (ret < 0)
+			goto out_free;
+		if (ret == 1) {
 			si->stack_depth--;
 			continue;
 		}
@@ -995,25 +971,24 @@ xfarray_sort(
 		 * than the pivot is on the right side of the range.
 		 */
 		while (lo < hi) {
+			void	*p;
+
 			/*
 			 * Decrement hi until it finds an a[hi] less than the
 			 * pivot value.
 			 */
-			error = xfarray_sort_load_cached(si, hi, scratch);
+			error = xfarray_sort_scan(si, hi, &p);
 			if (error)
 				goto out_free;
-			while (xfarray_sort_cmp(si, scratch, pivot) >= 0 &&
-								lo < hi) {
+			while (xfarray_sort_cmp(si, p, pivot) >= 0 && lo < hi) {
 				hi--;
-				error = xfarray_sort_load_cached(si, hi,
-						scratch);
+				error = xfarray_sort_scan(si, hi, &p);
 				if (error)
 					goto out_free;
 			}
-			error = xfarray_sort_put_page(si);
-			if (error)
-				goto out_free;
-
+			if (p != scratch)
+				memcpy(scratch, p, si->array->obj_size);
+			xfarray_sort_scan_done(si);
 			if (xfarray_sort_terminated(si, &error))
 				goto out_free;
 
@@ -1028,21 +1003,18 @@ xfarray_sort(
 			 * Increment lo until it finds an a[lo] greater than
 			 * the pivot value.
 			 */
-			error = xfarray_sort_load_cached(si, lo, scratch);
+			error = xfarray_sort_scan(si, lo, &p);
 			if (error)
 				goto out_free;
-			while (xfarray_sort_cmp(si, scratch, pivot) <= 0 &&
-								lo < hi) {
+			while (xfarray_sort_cmp(si, p, pivot) <= 0 && lo < hi) {
 				lo++;
-				error = xfarray_sort_load_cached(si, lo,
-						scratch);
+				error = xfarray_sort_scan(si, lo, &p);
 				if (error)
 					goto out_free;
 			}
-			error = xfarray_sort_put_page(si);
-			if (error)
-				goto out_free;
-
+			if (p != scratch)
+				memcpy(scratch, p, si->array->obj_size);
+			xfarray_sort_scan_done(si);
 			if (xfarray_sort_terminated(si, &error))
 				goto out_free;
 
@@ -1078,6 +1050,24 @@ xfarray_sort(
 
 out_free:
 	trace_xfarray_sort_stats(si, error);
+	xfarray_sort_scan_done(si);
 	kvfree(si);
 	return error;
 }
+
+/* How many bytes is this array consuming? */
+unsigned long long
+xfarray_bytes(
+	struct xfarray		*array)
+{
+	return xfile_bytes(array->xfile);
+}
+
+/* Empty the entire array. */
+void
+xfarray_truncate(
+	struct xfarray	*array)
+{
+	xfile_discard(array->xfile, 0, MAX_LFS_FILESIZE);
+	array->nr = 0;
+}
diff --git a/fs/xfs/scrub/xfarray.h b/fs/xfs/scrub/xfarray.h
index 62b9c506fd..5eeeeed13a 100644
--- a/fs/xfs/scrub/xfarray.h
+++ b/fs/xfs/scrub/xfarray.h
@@ -8,6 +8,7 @@
 
 /* xfile array index type, along with cursor initialization */
 typedef uint64_t		xfarray_idx_t;
+#define XFARRAY_NULLIDX		((__force xfarray_idx_t)-1ULL)
 #define XFARRAY_CURSOR_INIT	((__force xfarray_idx_t)0)
 
 /* Iterate each index of an xfile array. */
@@ -44,6 +45,27 @@ int xfarray_unset(struct xfarray *array, xfarray_idx_t idx);
 int xfarray_store(struct xfarray *array, xfarray_idx_t idx, const void *ptr);
 int xfarray_store_anywhere(struct xfarray *array, const void *ptr);
 bool xfarray_element_is_null(struct xfarray *array, const void *ptr);
+void xfarray_truncate(struct xfarray *array);
+unsigned long long xfarray_bytes(struct xfarray *array);
+
+/*
+ * Load an array element, but zero the buffer if there's no data because we
+ * haven't stored to that array element yet.
+ */
+static inline int
+xfarray_load_sparse(
+	struct xfarray	*array,
+	uint64_t	idx,
+	void		*rec)
+{
+	int		error = xfarray_load(array, idx, rec);
+
+	if (error == -ENODATA) {
+		memset(rec, 0, array->obj_size);
+		return 0;
+	}
+	return error;
+}
 
 /* Append an element to the array. */
 static inline int xfarray_append(struct xfarray *array, const void *ptr)
@@ -105,9 +127,17 @@ struct xfarray_sortinfo {
 	/* XFARRAY_SORT_* flags; see below. */
 	unsigned int		flags;
 
-	/* Cache a page here for faster access. */
-	struct xfile_page	xfpage;
-	void			*page_kaddr;
+	/* next time we want to cond_resched() */
+	struct xchk_relax	relax;
+
+	/* Cache a folio here for faster scanning for pivots */
+	struct folio		*folio;
+
+	/* First array index in folio that is completely readable */
+	xfarray_idx_t		first_folio_idx;
+
+	/* Last array index in folio that is completely readable */
+	xfarray_idx_t		last_folio_idx;
 
 #ifdef DEBUG
 	/* Performance statistics. */
diff --git a/fs/xfs/scrub/xfblob.c b/fs/xfs/scrub/xfblob.c
new file mode 100644
index 0000000000..6ef2a9637f
--- /dev/null
+++ b/fs/xfs/scrub/xfblob.c
@@ -0,0 +1,168 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2021-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "scrub/scrub.h"
+#include "scrub/xfile.h"
+#include "scrub/xfarray.h"
+#include "scrub/xfblob.h"
+
+/*
+ * XFS Blob Storage
+ * ================
+ * Stores and retrieves blobs using an xfile.  Objects are appended to the file
+ * and the offset is returned as a magic cookie for retrieval.
+ */
+
+#define XB_KEY_MAGIC	0xABAADDAD
+struct xb_key {
+	uint32_t		xb_magic;  /* XB_KEY_MAGIC */
+	uint32_t		xb_size;   /* size of the blob, in bytes */
+	loff_t			xb_offset; /* byte offset of this key */
+	/* blob comes after here */
+} __packed;
+
+/* Initialize a blob storage object. */
+int
+xfblob_create(
+	const char		*description,
+	struct xfblob		**blobp)
+{
+	struct xfblob		*blob;
+	struct xfile		*xfile;
+	int			error;
+
+	error = xfile_create(description, 0, &xfile);
+	if (error)
+		return error;
+
+	blob = kmalloc(sizeof(struct xfblob), XCHK_GFP_FLAGS);
+	if (!blob) {
+		error = -ENOMEM;
+		goto out_xfile;
+	}
+
+	blob->xfile = xfile;
+	blob->last_offset = PAGE_SIZE;
+
+	*blobp = blob;
+	return 0;
+
+out_xfile:
+	xfile_destroy(xfile);
+	return error;
+}
+
+/* Destroy a blob storage object. */
+void
+xfblob_destroy(
+	struct xfblob	*blob)
+{
+	xfile_destroy(blob->xfile);
+	kfree(blob);
+}
+
+/* Retrieve a blob. */
+int
+xfblob_load(
+	struct xfblob	*blob,
+	xfblob_cookie	cookie,
+	void		*ptr,
+	uint32_t	size)
+{
+	struct xb_key	key;
+	int		error;
+
+	error = xfile_load(blob->xfile, &key, sizeof(key), cookie);
+	if (error)
+		return error;
+
+	if (key.xb_magic != XB_KEY_MAGIC || key.xb_offset != cookie) {
+		ASSERT(0);
+		return -ENODATA;
+	}
+	if (size < key.xb_size) {
+		ASSERT(0);
+		return -EFBIG;
+	}
+
+	return xfile_load(blob->xfile, ptr, key.xb_size,
+			cookie + sizeof(key));
+}
+
+/* Store a blob. */
+int
+xfblob_store(
+	struct xfblob	*blob,
+	xfblob_cookie	*cookie,
+	const void	*ptr,
+	uint32_t	size)
+{
+	struct xb_key	key = {
+		.xb_offset = blob->last_offset,
+		.xb_magic = XB_KEY_MAGIC,
+		.xb_size = size,
+	};
+	loff_t		pos = blob->last_offset;
+	int		error;
+
+	error = xfile_store(blob->xfile, &key, sizeof(key), pos);
+	if (error)
+		return error;
+
+	pos += sizeof(key);
+	error = xfile_store(blob->xfile, ptr, size, pos);
+	if (error)
+		goto out_err;
+
+	*cookie = blob->last_offset;
+	blob->last_offset += sizeof(key) + size;
+	return 0;
+out_err:
+	xfile_discard(blob->xfile, blob->last_offset, sizeof(key));
+	return error;
+}
+
+/* Free a blob. */
+int
+xfblob_free(
+	struct xfblob	*blob,
+	xfblob_cookie	cookie)
+{
+	struct xb_key	key;
+	int		error;
+
+	error = xfile_load(blob->xfile, &key, sizeof(key), cookie);
+	if (error)
+		return error;
+
+	if (key.xb_magic != XB_KEY_MAGIC || key.xb_offset != cookie) {
+		ASSERT(0);
+		return -ENODATA;
+	}
+
+	xfile_discard(blob->xfile, cookie, sizeof(key) + key.xb_size);
+	return 0;
+}
+
+/* How many bytes is this blob storage object consuming? */
+unsigned long long
+xfblob_bytes(
+	struct xfblob		*blob)
+{
+	return xfile_bytes(blob->xfile);
+}
+
+/* Drop all the blobs. */
+void
+xfblob_truncate(
+	struct xfblob	*blob)
+{
+	xfile_discard(blob->xfile, PAGE_SIZE, MAX_LFS_FILESIZE - PAGE_SIZE);
+	blob->last_offset = PAGE_SIZE;
+}
diff --git a/fs/xfs/scrub/xfblob.h b/fs/xfs/scrub/xfblob.h
new file mode 100644
index 0000000000..ae78322613
--- /dev/null
+++ b/fs/xfs/scrub/xfblob.h
@@ -0,0 +1,50 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (c) 2021-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_XFBLOB_H__
+#define __XFS_SCRUB_XFBLOB_H__
+
+struct xfblob {
+	struct xfile	*xfile;
+	loff_t		last_offset;
+};
+
+typedef loff_t		xfblob_cookie;
+
+int xfblob_create(const char *descr, struct xfblob **blobp);
+void xfblob_destroy(struct xfblob *blob);
+int xfblob_load(struct xfblob *blob, xfblob_cookie cookie, void *ptr,
+		uint32_t size);
+int xfblob_store(struct xfblob *blob, xfblob_cookie *cookie, const void *ptr,
+		uint32_t size);
+int xfblob_free(struct xfblob *blob, xfblob_cookie cookie);
+unsigned long long xfblob_bytes(struct xfblob *blob);
+void xfblob_truncate(struct xfblob *blob);
+
+static inline int
+xfblob_storename(
+	struct xfblob		*blob,
+	xfblob_cookie		*cookie,
+	const struct xfs_name	*xname)
+{
+	return xfblob_store(blob, cookie, xname->name, xname->len);
+}
+
+static inline int
+xfblob_loadname(
+	struct xfblob		*blob,
+	xfblob_cookie		cookie,
+	struct xfs_name		*xname,
+	uint32_t		size)
+{
+	int ret = xfblob_load(blob, cookie, (void *)xname->name, size);
+	if (ret)
+		return ret;
+
+	xname->len = size;
+	return 0;
+}
+
+#endif /* __XFS_SCRUB_XFBLOB_H__ */
diff --git a/fs/xfs/scrub/xfile.c b/fs/xfs/scrub/xfile.c
index 090c3ead43..d848222f80 100644
--- a/fs/xfs/scrub/xfile.c
+++ b/fs/xfs/scrub/xfile.c
@@ -10,9 +10,9 @@
 #include "xfs_log_format.h"
 #include "xfs_trans_resv.h"
 #include "xfs_mount.h"
+#include "scrub/scrub.h"
 #include "scrub/xfile.h"
 #include "scrub/xfarray.h"
-#include "scrub/scrub.h"
 #include "scrub/trace.h"
 #include <linux/shmem_fs.h>
 
@@ -34,13 +34,6 @@
  * xfiles assume that the caller will handle all required concurrency
  * management; standard vfs locks (freezer and inode) are not taken.  Reads
  * and writes are satisfied directly from the page cache.
- *
- * NOTE: The current shmemfs implementation has a quirk that in-kernel reads
- * of a hole cause a page to be mapped into the file.  If you are going to
- * create a sparse xfile, please be careful about reading from uninitialized
- * parts of the file.  These pages are !Uptodate and will eventually be
- * reclaimed if not written, but in the short term this boosts memory
- * consumption.
  */
 
 /*
@@ -62,38 +55,27 @@ xfile_create(
 {
 	struct inode		*inode;
 	struct xfile		*xf;
-	int			error = -ENOMEM;
+	int			error;
 
 	xf = kmalloc(sizeof(struct xfile), XCHK_GFP_FLAGS);
 	if (!xf)
 		return -ENOMEM;
 
-	xf->file = shmem_file_setup(description, isize, 0);
-	if (!xf->file)
-		goto out_xfile;
+	xf->file = shmem_kernel_file_setup(description, isize, VM_NORESERVE);
 	if (IS_ERR(xf->file)) {
 		error = PTR_ERR(xf->file);
 		goto out_xfile;
 	}
 
-	/*
-	 * We want a large sparse file that we can pread, pwrite, and seek.
-	 * xfile users are responsible for keeping the xfile hidden away from
-	 * all other callers, so we skip timestamp updates and security checks.
-	 * Make the inode only accessible by root, just in case the xfile ever
-	 * escapes.
-	 */
-	xf->file->f_mode |= FMODE_PREAD | FMODE_PWRITE | FMODE_NOCMTIME |
-			    FMODE_LSEEK;
-	xf->file->f_flags |= O_RDWR | O_LARGEFILE | O_NOATIME;
 	inode = file_inode(xf->file);
-	inode->i_flags |= S_PRIVATE | S_NOCMTIME | S_NOATIME;
-	inode->i_mode &= ~0177;
-	inode->i_uid = GLOBAL_ROOT_UID;
-	inode->i_gid = GLOBAL_ROOT_GID;
-
 	lockdep_set_class(&inode->i_rwsem, &xfile_i_mutex_key);
 
+	/*
+	 * We don't want to bother with kmapping data during repair, so don't
+	 * allow highmem pages to back this mapping.
+	 */
+	mapping_set_gfp_mask(inode->i_mapping, GFP_KERNEL);
+
 	trace_xfile_create(xf);
 
 	*xfilep = xf;
@@ -118,164 +100,128 @@ xfile_destroy(
 }
 
 /*
- * Read a memory object directly from the xfile's page cache.  Unlike regular
- * pread, we return -E2BIG and -EFBIG for reads that are too large or at too
- * high an offset, instead of truncating the read.  Otherwise, we return
- * bytes read or an error code, like regular pread.
+ * Load an object.  Since we're treating this file as "memory", any error or
+ * short IO is treated as a failure to allocate memory.
  */
-ssize_t
-xfile_pread(
+int
+xfile_load(
 	struct xfile		*xf,
 	void			*buf,
 	size_t			count,
 	loff_t			pos)
 {
 	struct inode		*inode = file_inode(xf->file);
-	struct address_space	*mapping = inode->i_mapping;
-	struct page		*page = NULL;
-	ssize_t			read = 0;
 	unsigned int		pflags;
-	int			error = 0;
 
 	if (count > MAX_RW_COUNT)
-		return -E2BIG;
+		return -ENOMEM;
 	if (inode->i_sb->s_maxbytes - pos < count)
-		return -EFBIG;
+		return -ENOMEM;
 
-	trace_xfile_pread(xf, pos, count);
+	trace_xfile_load(xf, pos, count);
 
 	pflags = memalloc_nofs_save();
 	while (count > 0) {
-		void		*p, *kaddr;
+		struct folio	*folio;
 		unsigned int	len;
+		unsigned int	offset;
 
-		len = min_t(ssize_t, count, PAGE_SIZE - offset_in_page(pos));
-
-		/*
-		 * In-kernel reads of a shmem file cause it to allocate a page
-		 * if the mapping shows a hole.  Therefore, if we hit ENOMEM
-		 * we can continue by zeroing the caller's buffer.
-		 */
-		page = shmem_read_mapping_page_gfp(mapping, pos >> PAGE_SHIFT,
-				__GFP_NOWARN);
-		if (IS_ERR(page)) {
-			error = PTR_ERR(page);
-			if (error != -ENOMEM)
-				break;
-
-			memset(buf, 0, len);
-			goto advance;
-		}
-
-		if (PageUptodate(page)) {
+		if (shmem_get_folio(inode, pos >> PAGE_SHIFT, &folio,
+				SGP_READ) < 0)
+			break;
+		if (!folio) {
 			/*
-			 * xfile pages must never be mapped into userspace, so
-			 * we skip the dcache flush.
+			 * No data stored at this offset, just zero the output
+			 * buffer until the next page boundary.
 			 */
-			kaddr = kmap_local_page(page);
-			p = kaddr + offset_in_page(pos);
-			memcpy(buf, p, len);
-			kunmap_local(kaddr);
-		} else {
+			len = min_t(ssize_t, count,
+				PAGE_SIZE - offset_in_page(pos));
 			memset(buf, 0, len);
-		}
-		put_page(page);
+		} else {
+			if (filemap_check_wb_err(inode->i_mapping, 0)) {
+				folio_unlock(folio);
+				folio_put(folio);
+				break;
+			}
 
-advance:
+			offset = offset_in_folio(folio, pos);
+			len = min_t(ssize_t, count, folio_size(folio) - offset);
+			memcpy(buf, folio_address(folio) + offset, len);
+
+			folio_unlock(folio);
+			folio_put(folio);
+		}
 		count -= len;
 		pos += len;
 		buf += len;
-		read += len;
 	}
 	memalloc_nofs_restore(pflags);
 
-	if (read > 0)
-		return read;
-	return error;
+	if (count)
+		return -ENOMEM;
+	return 0;
 }
 
 /*
- * Write a memory object directly to the xfile's page cache.  Unlike regular
- * pwrite, we return -E2BIG and -EFBIG for writes that are too large or at too
- * high an offset, instead of truncating the write.  Otherwise, we return
- * bytes written or an error code, like regular pwrite.
+ * Store an object.  Since we're treating this file as "memory", any error or
+ * short IO is treated as a failure to allocate memory.
  */
-ssize_t
-xfile_pwrite(
+int
+xfile_store(
 	struct xfile		*xf,
 	const void		*buf,
 	size_t			count,
 	loff_t			pos)
 {
 	struct inode		*inode = file_inode(xf->file);
-	struct address_space	*mapping = inode->i_mapping;
-	const struct address_space_operations *aops = mapping->a_ops;
-	struct page		*page = NULL;
-	ssize_t			written = 0;
 	unsigned int		pflags;
-	int			error = 0;
 
 	if (count > MAX_RW_COUNT)
-		return -E2BIG;
+		return -ENOMEM;
 	if (inode->i_sb->s_maxbytes - pos < count)
-		return -EFBIG;
+		return -ENOMEM;
 
-	trace_xfile_pwrite(xf, pos, count);
+	trace_xfile_store(xf, pos, count);
+
+	/*
+	 * Increase the file size first so that shmem_get_folio(..., SGP_CACHE),
+	 * actually allocates a folio instead of erroring out.
+	 */
+	if (pos + count > i_size_read(inode))
+		i_size_write(inode, pos + count);
 
 	pflags = memalloc_nofs_save();
 	while (count > 0) {
-		void		*fsdata = NULL;
-		void		*p, *kaddr;
+		struct folio	*folio;
 		unsigned int	len;
-		int		ret;
-
-		len = min_t(ssize_t, count, PAGE_SIZE - offset_in_page(pos));
-
-		/*
-		 * We call write_begin directly here to avoid all the freezer
-		 * protection lock-taking that happens in the normal path.
-		 * shmem doesn't support fs freeze, but lockdep doesn't know
-		 * that and will trip over that.
-		 */
-		error = aops->write_begin(NULL, mapping, pos, len, &page,
-				&fsdata);
-		if (error)
-			break;
+		unsigned int	offset;
 
-		/*
-		 * xfile pages must never be mapped into userspace, so we skip
-		 * the dcache flush.  If the page is not uptodate, zero it
-		 * before writing data.
-		 */
-		kaddr = kmap_local_page(page);
-		if (!PageUptodate(page)) {
-			memset(kaddr, 0, PAGE_SIZE);
-			SetPageUptodate(page);
-		}
-		p = kaddr + offset_in_page(pos);
-		memcpy(p, buf, len);
-		kunmap_local(kaddr);
-
-		ret = aops->write_end(NULL, mapping, pos, len, len, page,
-				fsdata);
-		if (ret < 0) {
-			error = ret;
+		if (shmem_get_folio(inode, pos >> PAGE_SHIFT, &folio,
+				SGP_CACHE) < 0)
+			break;
+		if (filemap_check_wb_err(inode->i_mapping, 0)) {
+			folio_unlock(folio);
+			folio_put(folio);
 			break;
 		}
 
-		written += ret;
-		if (ret != len)
-			break;
+		offset = offset_in_folio(folio, pos);
+		len = min_t(ssize_t, count, folio_size(folio) - offset);
+		memcpy(folio_address(folio) + offset, buf, len);
+
+		folio_mark_dirty(folio);
+		folio_unlock(folio);
+		folio_put(folio);
 
-		count -= ret;
-		pos += ret;
-		buf += ret;
+		count -= len;
+		pos += len;
+		buf += len;
 	}
 	memalloc_nofs_restore(pflags);
 
-	if (written > 0)
-		return written;
-	return error;
+	if (count)
+		return -ENOMEM;
+	return 0;
 }
 
 /* Find the next written area in the xfile data for a given offset. */
@@ -291,129 +237,88 @@ xfile_seek_data(
 	return ret;
 }
 
-/* Query stat information for an xfile. */
-int
-xfile_stat(
-	struct xfile		*xf,
-	struct xfile_stat	*statbuf)
-{
-	struct kstat		ks;
-	int			error;
-
-	error = vfs_getattr_nosec(&xf->file->f_path, &ks,
-			STATX_SIZE | STATX_BLOCKS, AT_STATX_DONT_SYNC);
-	if (error)
-		return error;
-
-	statbuf->size = ks.size;
-	statbuf->bytes = ks.blocks << SECTOR_SHIFT;
-	return 0;
-}
-
 /*
- * Grab the (locked) page for a memory object.  The object cannot span a page
- * boundary.  Returns 0 (and a locked page) if successful, -ENOTBLK if we
- * cannot grab the page, or the usual negative errno.
+ * Grab the (locked) folio for a memory object.  The object cannot span a folio
+ * boundary.  Returns the locked folio if successful, NULL if there was no
+ * folio or it didn't cover the range requested, or an ERR_PTR on failure.
  */
-int
-xfile_get_page(
+struct folio *
+xfile_get_folio(
 	struct xfile		*xf,
 	loff_t			pos,
-	unsigned int		len,
-	struct xfile_page	*xfpage)
+	size_t			len,
+	unsigned int		flags)
 {
 	struct inode		*inode = file_inode(xf->file);
-	struct address_space	*mapping = inode->i_mapping;
-	const struct address_space_operations *aops = mapping->a_ops;
-	struct page		*page = NULL;
-	void			*fsdata = NULL;
-	loff_t			key = round_down(pos, PAGE_SIZE);
+	struct folio		*folio = NULL;
 	unsigned int		pflags;
 	int			error;
 
 	if (inode->i_sb->s_maxbytes - pos < len)
-		return -ENOMEM;
-	if (len > PAGE_SIZE - offset_in_page(pos))
-		return -ENOTBLK;
+		return ERR_PTR(-ENOMEM);
 
-	trace_xfile_get_page(xf, pos, len);
-
-	pflags = memalloc_nofs_save();
+	trace_xfile_get_folio(xf, pos, len);
 
 	/*
-	 * We call write_begin directly here to avoid all the freezer
-	 * protection lock-taking that happens in the normal path.  shmem
-	 * doesn't support fs freeze, but lockdep doesn't know that and will
-	 * trip over that.
+	 * Increase the file size first so that shmem_get_folio(..., SGP_CACHE),
+	 * actually allocates a folio instead of erroring out.
 	 */
-	error = aops->write_begin(NULL, mapping, key, PAGE_SIZE, &page,
-			&fsdata);
+	if ((flags & XFILE_ALLOC) && pos + len > i_size_read(inode))
+		i_size_write(inode, pos + len);
+
+	pflags = memalloc_nofs_save();
+	error = shmem_get_folio(inode, pos >> PAGE_SHIFT, &folio,
+			(flags & XFILE_ALLOC) ? SGP_CACHE : SGP_READ);
+	memalloc_nofs_restore(pflags);
 	if (error)
-		goto out_pflags;
+		return ERR_PTR(error);
 
-	/* We got the page, so make sure we push out EOF. */
-	if (i_size_read(inode) < pos + len)
-		i_size_write(inode, pos + len);
+	if (!folio)
+		return NULL;
 
-	/*
-	 * If the page isn't up to date, fill it with zeroes before we hand it
-	 * to the caller and make sure the backing store will hold on to them.
-	 */
-	if (!PageUptodate(page)) {
-		void	*kaddr;
+	if (len > folio_size(folio) - offset_in_folio(folio, pos)) {
+		folio_unlock(folio);
+		folio_put(folio);
+		return NULL;
+	}
 
-		kaddr = kmap_local_page(page);
-		memset(kaddr, 0, PAGE_SIZE);
-		kunmap_local(kaddr);
-		SetPageUptodate(page);
+	if (filemap_check_wb_err(inode->i_mapping, 0)) {
+		folio_unlock(folio);
+		folio_put(folio);
+		return ERR_PTR(-EIO);
 	}
 
 	/*
-	 * Mark each page dirty so that the contents are written to some
-	 * backing store when we drop this buffer, and take an extra reference
-	 * to prevent the xfile page from being swapped or removed from the
-	 * page cache by reclaim if the caller unlocks the page.
+	 * Mark the folio dirty so that it won't be reclaimed once we drop the
+	 * (potentially last) reference in xfile_put_folio.
 	 */
-	set_page_dirty(page);
-	get_page(page);
-
-	xfpage->page = page;
-	xfpage->fsdata = fsdata;
-	xfpage->pos = key;
-out_pflags:
-	memalloc_nofs_restore(pflags);
-	return error;
+	if (flags & XFILE_ALLOC)
+		folio_set_dirty(folio);
+	return folio;
 }
 
 /*
- * Release the (locked) page for a memory object.  Returns 0 or a negative
- * errno.
+ * Release the (locked) folio for a memory object.
  */
-int
-xfile_put_page(
+void
+xfile_put_folio(
 	struct xfile		*xf,
-	struct xfile_page	*xfpage)
+	struct folio		*folio)
 {
-	struct inode		*inode = file_inode(xf->file);
-	struct address_space	*mapping = inode->i_mapping;
-	const struct address_space_operations *aops = mapping->a_ops;
-	unsigned int		pflags;
-	int			ret;
-
-	trace_xfile_put_page(xf, xfpage->pos, PAGE_SIZE);
+	trace_xfile_put_folio(xf, folio_pos(folio), folio_size(folio));
 
-	/* Give back the reference that we took in xfile_get_page. */
-	put_page(xfpage->page);
+	folio_unlock(folio);
+	folio_put(folio);
+}
 
-	pflags = memalloc_nofs_save();
-	ret = aops->write_end(NULL, mapping, xfpage->pos, PAGE_SIZE, PAGE_SIZE,
-			xfpage->page, xfpage->fsdata);
-	memalloc_nofs_restore(pflags);
-	memset(xfpage, 0, sizeof(struct xfile_page));
+/* Discard the page cache that's backing a range of the xfile. */
+void
+xfile_discard(
+	struct xfile		*xf,
+	loff_t			pos,
+	u64			count)
+{
+	trace_xfile_discard(xf, pos, count);
 
-	if (ret < 0)
-		return ret;
-	if (ret != PAGE_SIZE)
-		return -EIO;
-	return 0;
+	shmem_truncate_range(file_inode(xf->file), pos, pos + count - 1);
 }
diff --git a/fs/xfs/scrub/xfile.h b/fs/xfs/scrub/xfile.h
index d56643b0f4..cc2cc1714c 100644
--- a/fs/xfs/scrub/xfile.h
+++ b/fs/xfs/scrub/xfile.h
@@ -6,22 +6,6 @@
 #ifndef __XFS_SCRUB_XFILE_H__
 #define __XFS_SCRUB_XFILE_H__
 
-struct xfile_page {
-	struct page		*page;
-	void			*fsdata;
-	loff_t			pos;
-};
-
-static inline bool xfile_page_cached(const struct xfile_page *xfpage)
-{
-	return xfpage->page != NULL;
-}
-
-static inline pgoff_t xfile_page_index(const struct xfile_page *xfpage)
-{
-	return xfpage->page->index;
-}
-
 struct xfile {
 	struct file		*file;
 };
@@ -29,49 +13,23 @@ struct xfile {
 int xfile_create(const char *description, loff_t isize, struct xfile **xfilep);
 void xfile_destroy(struct xfile *xf);
 
-ssize_t xfile_pread(struct xfile *xf, void *buf, size_t count, loff_t pos);
-ssize_t xfile_pwrite(struct xfile *xf, const void *buf, size_t count,
+int xfile_load(struct xfile *xf, void *buf, size_t count, loff_t pos);
+int xfile_store(struct xfile *xf, const void *buf, size_t count,
 		loff_t pos);
 
-/*
- * Load an object.  Since we're treating this file as "memory", any error or
- * short IO is treated as a failure to allocate memory.
- */
-static inline int
-xfile_obj_load(struct xfile *xf, void *buf, size_t count, loff_t pos)
-{
-	ssize_t	ret = xfile_pread(xf, buf, count, pos);
-
-	if (ret < 0 || ret != count)
-		return -ENOMEM;
-	return 0;
-}
-
-/*
- * Store an object.  Since we're treating this file as "memory", any error or
- * short IO is treated as a failure to allocate memory.
- */
-static inline int
-xfile_obj_store(struct xfile *xf, const void *buf, size_t count, loff_t pos)
-{
-	ssize_t	ret = xfile_pwrite(xf, buf, count, pos);
-
-	if (ret < 0 || ret != count)
-		return -ENOMEM;
-	return 0;
-}
-
+void xfile_discard(struct xfile *xf, loff_t pos, u64 count);
 loff_t xfile_seek_data(struct xfile *xf, loff_t pos);
 
-struct xfile_stat {
-	loff_t			size;
-	unsigned long long	bytes;
-};
+#define XFILE_MAX_FOLIO_SIZE	(PAGE_SIZE << MAX_PAGECACHE_ORDER)
 
-int xfile_stat(struct xfile *xf, struct xfile_stat *statbuf);
+#define XFILE_ALLOC		(1 << 0) /* allocate folio if not present */
+struct folio *xfile_get_folio(struct xfile *xf, loff_t offset, size_t len,
+		unsigned int flags);
+void xfile_put_folio(struct xfile *xf, struct folio *folio);
 
-int xfile_get_page(struct xfile *xf, loff_t offset, unsigned int len,
-		struct xfile_page *xbuf);
-int xfile_put_page(struct xfile *xf, struct xfile_page *xbuf);
+static inline unsigned long long xfile_bytes(struct xfile *xf)
+{
+	return file_inode(xf->file)->i_blocks << SECTOR_SHIFT;
+}
 
 #endif /* __XFS_SCRUB_XFILE_H__ */
diff --git a/fs/xfs/scrub/xfs_scrub.h b/fs/xfs/scrub/xfs_scrub.h
index a39befa743..f17173b83e 100644
--- a/fs/xfs/scrub/xfs_scrub.h
+++ b/fs/xfs/scrub/xfs_scrub.h
@@ -7,9 +7,11 @@
 #define __XFS_SCRUB_H__
 
 #ifndef CONFIG_XFS_ONLINE_SCRUB
-# define xfs_scrub_metadata(file, sm)	(-ENOTTY)
+# define xfs_ioc_scrub_metadata(f, a)	(-ENOTTY)
+# define xfs_ioc_scrubv_metadata(f, a)	(-ENOTTY)
 #else
-int xfs_scrub_metadata(struct file *file, struct xfs_scrub_metadata *sm);
+int xfs_ioc_scrub_metadata(struct file *file, void __user *arg);
+int xfs_ioc_scrubv_metadata(struct file *file, void __user *arg);
 #endif /* CONFIG_XFS_ONLINE_SCRUB */
 
 #endif	/* __XFS_SCRUB_H__ */
diff --git a/fs/xfs/xfs_acl.c b/fs/xfs/xfs_acl.c
index 6b84030181..c7c3dcfa27 100644
--- a/fs/xfs/xfs_acl.c
+++ b/fs/xfs/xfs_acl.c
@@ -167,7 +167,7 @@ xfs_get_acl(struct inode *inode, int type, bool rcu)
 		acl = ERR_PTR(error);
 	}
 
-	kmem_free(args.value);
+	kvfree(args.value);
 	return acl;
 }
 
@@ -201,16 +201,17 @@ __xfs_set_acl(struct inode *inode, struct posix_acl *acl, int type)
 		if (!args.value)
 			return -ENOMEM;
 		xfs_acl_to_disk(args.value, acl);
+		error = xfs_attr_change(&args, XFS_ATTRUPDATE_UPSERT);
+		kvfree(args.value);
+	} else {
+		error = xfs_attr_change(&args, XFS_ATTRUPDATE_REMOVE);
+		/*
+		 * If the attribute didn't exist to start with that's fine.
+		 */
+		if (error == -ENOATTR)
+			error = 0;
 	}
 
-	error = xfs_attr_change(&args);
-	kmem_free(args.value);
-
-	/*
-	 * If the attribute didn't exist to start with that's fine.
-	 */
-	if (!acl && error == -ENOATTR)
-		error = 0;
 	if (!error)
 		set_cached_acl(inode, type, acl);
 	return error;
diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c
index 813f85156b..6dead20338 100644
--- a/fs/xfs/xfs_aops.c
+++ b/fs/xfs/xfs_aops.c
@@ -112,7 +112,7 @@ xfs_end_ioend(
 	 * longer dirty. If we don't remove delalloc blocks here, they become
 	 * stale and can corrupt free space accounting on unmount.
 	 */
-	error = blk_status_to_errno(ioend->io_bio->bi_status);
+	error = blk_status_to_errno(ioend->io_bio.bi_status);
 	if (unlikely(error)) {
 		if (ioend->io_flags & IOMAP_F_SHARED) {
 			xfs_reflink_cancel_cow_range(ip, offset, size, true);
@@ -179,7 +179,7 @@ STATIC void
 xfs_end_bio(
 	struct bio		*bio)
 {
-	struct iomap_ioend	*ioend = bio->bi_private;
+	struct iomap_ioend	*ioend = iomap_ioend_from_bio(bio);
 	struct xfs_inode	*ip = XFS_I(ioend->io_inode);
 	unsigned long		flags;
 
@@ -233,50 +233,12 @@ xfs_imap_valid(
 	return true;
 }
 
-/*
- * Pass in a dellalloc extent and convert it to real extents, return the real
- * extent that maps offset_fsb in wpc->iomap.
- *
- * The current page is held locked so nothing could have removed the block
- * backing offset_fsb, although it could have moved from the COW to the data
- * fork by another thread.
- */
-static int
-xfs_convert_blocks(
-	struct iomap_writepage_ctx *wpc,
-	struct xfs_inode	*ip,
-	int			whichfork,
-	loff_t			offset)
-{
-	int			error;
-	unsigned		*seq;
-
-	if (whichfork == XFS_COW_FORK)
-		seq = &XFS_WPC(wpc)->cow_seq;
-	else
-		seq = &XFS_WPC(wpc)->data_seq;
-
-	/*
-	 * Attempt to allocate whatever delalloc extent currently backs offset
-	 * and put the result into wpc->iomap.  Allocate in a loop because it
-	 * may take several attempts to allocate real blocks for a contiguous
-	 * delalloc extent if free space is sufficiently fragmented.
-	 */
-	do {
-		error = xfs_bmapi_convert_delalloc(ip, whichfork, offset,
-				&wpc->iomap, seq);
-		if (error)
-			return error;
-	} while (wpc->iomap.offset + wpc->iomap.length <= offset);
-
-	return 0;
-}
-
 static int
 xfs_map_blocks(
 	struct iomap_writepage_ctx *wpc,
 	struct inode		*inode,
-	loff_t			offset)
+	loff_t			offset,
+	unsigned int		len)
 {
 	struct xfs_inode	*ip = XFS_I(inode);
 	struct xfs_mount	*mp = ip->i_mount;
@@ -289,6 +251,7 @@ xfs_map_blocks(
 	struct xfs_iext_cursor	icur;
 	int			retries = 0;
 	int			error = 0;
+	unsigned int		*seq;
 
 	if (xfs_is_shutdown(mp))
 		return -EIO;
@@ -386,7 +349,19 @@ retry:
 	trace_xfs_map_blocks_found(ip, offset, count, whichfork, &imap);
 	return 0;
 allocate_blocks:
-	error = xfs_convert_blocks(wpc, ip, whichfork, offset);
+	/*
+	 * Convert a dellalloc extent to a real one. The current page is held
+	 * locked so nothing could have removed the block backing offset_fsb,
+	 * although it could have moved from the COW to the data fork by another
+	 * thread.
+	 */
+	if (whichfork == XFS_COW_FORK)
+		seq = &XFS_WPC(wpc)->cow_seq;
+	else
+		seq = &XFS_WPC(wpc)->data_seq;
+
+	error = xfs_bmapi_convert_delalloc(ip, whichfork, offset,
+				&wpc->iomap, seq);
 	if (error) {
 		/*
 		 * If we failed to find the extent in the COW fork we might have
@@ -444,7 +419,7 @@ xfs_prepare_ioend(
 	/* send ioends that might require a transaction to the completion wq */
 	if (xfs_ioend_is_append(ioend) || ioend->io_type == IOMAP_UNWRITTEN ||
 	    (ioend->io_flags & IOMAP_F_SHARED))
-		ioend->io_bio->bi_end_io = xfs_end_bio;
+		ioend->io_bio.bi_end_io = xfs_end_bio;
 	return status;
 }
 
@@ -468,7 +443,6 @@ xfs_discard_folio(
 {
 	struct xfs_inode	*ip = XFS_I(folio->mapping->host);
 	struct xfs_mount	*mp = ip->i_mount;
-	int			error;
 
 	if (xfs_is_shutdown(mp))
 		return;
@@ -482,11 +456,8 @@ xfs_discard_folio(
 	 * byte of the next folio. Hence the end offset is only dependent on the
 	 * folio itself and not the start offset that is passed in.
 	 */
-	error = xfs_bmap_punch_delalloc_range(ip, pos,
+	xfs_bmap_punch_delalloc_range(ip, pos,
 				folio_pos(folio) + folio_size(folio));
-
-	if (error && !xfs_is_shutdown(mp))
-		xfs_alert(mp, "page discard unable to remove delalloc mapping.");
 }
 
 static const struct iomap_writeback_ops xfs_writeback_ops = {
@@ -502,13 +473,6 @@ xfs_vm_writepages(
 {
 	struct xfs_writepage_ctx wpc = { };
 
-	/*
-	 * Writing back data in a transaction context can result in recursive
-	 * transactions. This is bad, so issue a warning and get out of here.
-	 */
-	if (WARN_ON_ONCE(current->journal_info))
-		return 0;
-
 	xfs_iflags_clear(XFS_I(mapping->host), XFS_ITRUNCATED);
 	return iomap_writepages(mapping, wbc, &wpc.ctx, &xfs_writeback_ops);
 }
diff --git a/fs/xfs/xfs_attr_inactive.c b/fs/xfs/xfs_attr_inactive.c
index 89c7a9f4f9..24fb12986a 100644
--- a/fs/xfs/xfs_attr_inactive.c
+++ b/fs/xfs/xfs_attr_inactive.c
@@ -23,6 +23,7 @@
 #include "xfs_quota.h"
 #include "xfs_dir2.h"
 #include "xfs_error.h"
+#include "xfs_health.h"
 
 /*
  * Invalidate any incore buffers associated with this remote attribute value
@@ -147,6 +148,7 @@ xfs_attr3_node_inactive(
 	if (level > XFS_DA_NODE_MAXDEPTH) {
 		xfs_buf_mark_corrupt(bp);
 		xfs_trans_brelse(*trans, bp);	/* no locks for later trans */
+		xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK);
 		return -EFSCORRUPTED;
 	}
 
@@ -197,6 +199,7 @@ xfs_attr3_node_inactive(
 		default:
 			xfs_buf_mark_corrupt(child_bp);
 			xfs_trans_brelse(*trans, child_bp);
+			xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK);
 			error = -EFSCORRUPTED;
 			break;
 		}
@@ -286,6 +289,7 @@ xfs_attr3_root_inactive(
 		error = xfs_attr3_leaf_inactive(trans, dp, bp);
 		break;
 	default:
+		xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK);
 		error = -EFSCORRUPTED;
 		xfs_buf_mark_corrupt(bp);
 		xfs_trans_brelse(*trans, bp);
diff --git a/fs/xfs/xfs_attr_item.c b/fs/xfs/xfs_attr_item.c
index 9e02111bd8..f683b7a932 100644
--- a/fs/xfs/xfs_attr_item.c
+++ b/fs/xfs/xfs_attr_item.c
@@ -27,6 +27,7 @@
 #include "xfs_error.h"
 #include "xfs_log_priv.h"
 #include "xfs_log_recover.h"
+#include "xfs_parent.h"
 
 struct kmem_cache		*xfs_attri_cache;
 struct kmem_cache		*xfs_attrd_cache;
@@ -73,8 +74,12 @@ static inline struct xfs_attri_log_nameval *
 xfs_attri_log_nameval_alloc(
 	const void			*name,
 	unsigned int			name_len,
+	const void			*new_name,
+	unsigned int			new_name_len,
 	const void			*value,
-	unsigned int			value_len)
+	unsigned int			value_len,
+	const void			*new_value,
+	unsigned int			new_value_len)
 {
 	struct xfs_attri_log_nameval	*nv;
 
@@ -83,15 +88,26 @@ xfs_attri_log_nameval_alloc(
 	 * this. But kvmalloc() utterly sucks, so we use our own version.
 	 */
 	nv = xlog_kvmalloc(sizeof(struct xfs_attri_log_nameval) +
-					name_len + value_len);
+					name_len + new_name_len + value_len +
+					new_value_len);
 
 	nv->name.i_addr = nv + 1;
 	nv->name.i_len = name_len;
 	nv->name.i_type = XLOG_REG_TYPE_ATTR_NAME;
 	memcpy(nv->name.i_addr, name, name_len);
 
+	if (new_name_len) {
+		nv->new_name.i_addr = nv->name.i_addr + name_len;
+		nv->new_name.i_len = new_name_len;
+		memcpy(nv->new_name.i_addr, new_name, new_name_len);
+	} else {
+		nv->new_name.i_addr = NULL;
+		nv->new_name.i_len = 0;
+	}
+	nv->new_name.i_type = XLOG_REG_TYPE_ATTR_NEWNAME;
+
 	if (value_len) {
-		nv->value.i_addr = nv->name.i_addr + name_len;
+		nv->value.i_addr = nv->name.i_addr + name_len + new_name_len;
 		nv->value.i_len = value_len;
 		memcpy(nv->value.i_addr, value, value_len);
 	} else {
@@ -100,6 +116,17 @@ xfs_attri_log_nameval_alloc(
 	}
 	nv->value.i_type = XLOG_REG_TYPE_ATTR_VALUE;
 
+	if (new_value_len) {
+		nv->new_value.i_addr = nv->name.i_addr + name_len +
+						new_name_len + value_len;
+		nv->new_value.i_len = new_value_len;
+		memcpy(nv->new_value.i_addr, new_value, new_value_len);
+	} else {
+		nv->new_value.i_addr = NULL;
+		nv->new_value.i_len = 0;
+	}
+	nv->new_value.i_type = XLOG_REG_TYPE_ATTR_NEWVALUE;
+
 	refcount_set(&nv->refcount, 1);
 	return nv;
 }
@@ -108,7 +135,7 @@ STATIC void
 xfs_attri_item_free(
 	struct xfs_attri_log_item	*attrip)
 {
-	kmem_free(attrip->attri_item.li_lv_shadow);
+	kvfree(attrip->attri_item.li_lv_shadow);
 	xfs_attri_log_nameval_put(attrip->attri_nameval);
 	kmem_cache_free(xfs_attri_cache, attrip);
 }
@@ -145,11 +172,20 @@ xfs_attri_item_size(
 	*nbytes += sizeof(struct xfs_attri_log_format) +
 			xlog_calc_iovec_len(nv->name.i_len);
 
-	if (!nv->value.i_len)
-		return;
+	if (nv->new_name.i_len) {
+		*nvecs += 1;
+		*nbytes += xlog_calc_iovec_len(nv->new_name.i_len);
+	}
 
-	*nvecs += 1;
-	*nbytes += xlog_calc_iovec_len(nv->value.i_len);
+	if (nv->value.i_len) {
+		*nvecs += 1;
+		*nbytes += xlog_calc_iovec_len(nv->value.i_len);
+	}
+
+	if (nv->new_value.i_len) {
+		*nvecs += 1;
+		*nbytes += xlog_calc_iovec_len(nv->new_value.i_len);
+	}
 }
 
 /*
@@ -179,15 +215,28 @@ xfs_attri_item_format(
 	ASSERT(nv->name.i_len > 0);
 	attrip->attri_format.alfi_size++;
 
+	if (nv->new_name.i_len > 0)
+		attrip->attri_format.alfi_size++;
+
 	if (nv->value.i_len > 0)
 		attrip->attri_format.alfi_size++;
 
+	if (nv->new_value.i_len > 0)
+		attrip->attri_format.alfi_size++;
+
 	xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_ATTRI_FORMAT,
 			&attrip->attri_format,
 			sizeof(struct xfs_attri_log_format));
 	xlog_copy_from_iovec(lv, &vecp, &nv->name);
+
+	if (nv->new_name.i_len > 0)
+		xlog_copy_from_iovec(lv, &vecp, &nv->new_name);
+
 	if (nv->value.i_len > 0)
 		xlog_copy_from_iovec(lv, &vecp, &nv->value);
+
+	if (nv->new_value.i_len > 0)
+		xlog_copy_from_iovec(lv, &vecp, &nv->new_value);
 }
 
 /*
@@ -226,7 +275,7 @@ xfs_attri_init(
 {
 	struct xfs_attri_log_item	*attrip;
 
-	attrip = kmem_cache_zalloc(xfs_attri_cache, GFP_NOFS | __GFP_NOFAIL);
+	attrip = kmem_cache_zalloc(xfs_attri_cache, GFP_KERNEL | __GFP_NOFAIL);
 
 	/*
 	 * Grab an extra reference to the name/value buffer for this log item.
@@ -251,7 +300,7 @@ static inline struct xfs_attrd_log_item *ATTRD_ITEM(struct xfs_log_item *lip)
 STATIC void
 xfs_attrd_item_free(struct xfs_attrd_log_item *attrdp)
 {
-	kmem_free(attrdp->attrd_item.li_lv_shadow);
+	kvfree(attrdp->attrd_item.li_lv_shadow);
 	kmem_cache_free(xfs_attrd_cache, attrdp);
 }
 
@@ -308,6 +357,12 @@ xfs_attrd_item_intent(
 	return &ATTRD_ITEM(lip)->attrd_attrip->attri_item;
 }
 
+static inline unsigned int
+xfs_attr_log_item_op(const struct xfs_attri_log_format *attrp)
+{
+	return attrp->alfi_op_flags & XFS_ATTRI_OP_FLAGS_TYPE_MASK;
+}
+
 /* Log an attr to the intent item. */
 STATIC void
 xfs_attr_log_item(
@@ -316,6 +371,8 @@ xfs_attr_log_item(
 	const struct xfs_attr_intent	*attr)
 {
 	struct xfs_attri_log_format	*attrp;
+	struct xfs_attri_log_nameval	*nv = attr->xattri_nameval;
+	struct xfs_da_args		*args = attr->xattri_da_args;
 
 	/*
 	 * At this point the xfs_attr_intent has been constructed, and we've
@@ -323,13 +380,30 @@ xfs_attr_log_item(
 	 * structure with fields from this xfs_attr_intent
 	 */
 	attrp = &attrip->attri_format;
-	attrp->alfi_ino = attr->xattri_da_args->dp->i_ino;
+	attrp->alfi_ino = args->dp->i_ino;
 	ASSERT(!(attr->xattri_op_flags & ~XFS_ATTRI_OP_FLAGS_TYPE_MASK));
 	attrp->alfi_op_flags = attr->xattri_op_flags;
-	attrp->alfi_value_len = attr->xattri_nameval->value.i_len;
-	attrp->alfi_name_len = attr->xattri_nameval->name.i_len;
-	ASSERT(!(attr->xattri_da_args->attr_filter & ~XFS_ATTRI_FILTER_MASK));
-	attrp->alfi_attr_filter = attr->xattri_da_args->attr_filter;
+	attrp->alfi_value_len = nv->value.i_len;
+
+	switch (xfs_attr_log_item_op(attrp)) {
+	case XFS_ATTRI_OP_FLAGS_PPTR_REPLACE:
+		ASSERT(nv->value.i_len == nv->new_value.i_len);
+
+		attrp->alfi_igen = VFS_I(args->dp)->i_generation;
+		attrp->alfi_old_name_len = nv->name.i_len;
+		attrp->alfi_new_name_len = nv->new_name.i_len;
+		break;
+	case XFS_ATTRI_OP_FLAGS_PPTR_REMOVE:
+	case XFS_ATTRI_OP_FLAGS_PPTR_SET:
+		attrp->alfi_igen = VFS_I(args->dp)->i_generation;
+		fallthrough;
+	default:
+		attrp->alfi_name_len = nv->name.i_len;
+		break;
+	}
+
+	ASSERT(!(args->attr_filter & ~XFS_ATTRI_FILTER_MASK));
+	attrp->alfi_attr_filter = args->attr_filter;
 }
 
 /* Get an ATTRI. */
@@ -368,8 +442,11 @@ xfs_attr_create_intent(
 		 * Transfer our reference to the name/value buffer to the
 		 * deferred work state structure.
 		 */
-		attr->xattri_nameval = xfs_attri_log_nameval_alloc(args->name,
-				args->namelen, args->value, args->valuelen);
+		attr->xattri_nameval = xfs_attri_log_nameval_alloc(
+				args->name, args->namelen,
+				args->new_name, args->new_namelen,
+				args->value, args->valuelen,
+				args->new_value, args->new_valuelen);
 	}
 
 	attrip = xfs_attri_init(mp, attr->xattri_nameval);
@@ -386,11 +463,16 @@ xfs_attr_free_item(
 		xfs_da_state_free(attr->xattri_da_state);
 	xfs_attri_log_nameval_put(attr->xattri_nameval);
 	if (attr->xattri_da_args->op_flags & XFS_DA_OP_RECOVERY)
-		kmem_free(attr);
+		kfree(attr);
 	else
 		kmem_cache_free(xfs_attr_intent_cache, attr);
 }
 
+static inline struct xfs_attr_intent *attri_entry(const struct list_head *e)
+{
+	return list_entry(e, struct xfs_attr_intent, xattri_list);
+}
+
 /* Process an attr. */
 STATIC int
 xfs_attr_finish_item(
@@ -399,11 +481,10 @@ xfs_attr_finish_item(
 	struct list_head		*item,
 	struct xfs_btree_cur		**state)
 {
-	struct xfs_attr_intent		*attr;
+	struct xfs_attr_intent		*attr = attri_entry(item);
 	struct xfs_da_args		*args;
 	int				error;
 
-	attr = container_of(item, struct xfs_attr_intent, xattri_list);
 	args = attr->xattri_da_args;
 
 	/* Reset trans after EAGAIN cycle since the transaction is new */
@@ -443,9 +524,8 @@ STATIC void
 xfs_attr_cancel_item(
 	struct list_head		*item)
 {
-	struct xfs_attr_intent		*attr;
+	struct xfs_attr_intent		*attr = attri_entry(item);
 
-	attr = container_of(item, struct xfs_attr_intent, xattri_list);
 	xfs_attr_free_item(attr);
 }
 
@@ -457,17 +537,19 @@ xfs_attri_item_match(
 	return ATTRI_ITEM(lip)->attri_format.alfi_id == intent_id;
 }
 
+static inline bool
+xfs_attri_validate_namelen(unsigned int namelen)
+{
+	return namelen > 0 && namelen <= XATTR_NAME_MAX;
+}
+
 /* Is this recovered ATTRI format ok? */
 static inline bool
 xfs_attri_validate(
 	struct xfs_mount		*mp,
 	struct xfs_attri_log_format	*attrp)
 {
-	unsigned int			op = attrp->alfi_op_flags &
-					     XFS_ATTRI_OP_FLAGS_TYPE_MASK;
-
-	if (attrp->__pad != 0)
-		return false;
+	unsigned int			op = xfs_attr_log_item_op(attrp);
 
 	if (attrp->alfi_op_flags & ~XFS_ATTRI_OP_FLAGS_TYPE_MASK)
 		return false;
@@ -475,24 +557,75 @@ xfs_attri_validate(
 	if (attrp->alfi_attr_filter & ~XFS_ATTRI_FILTER_MASK)
 		return false;
 
-	/* alfi_op_flags should be either a set or remove */
+	if (!xfs_attr_check_namespace(attrp->alfi_attr_filter &
+				      XFS_ATTR_NSP_ONDISK_MASK))
+		return false;
+
 	switch (op) {
+	case XFS_ATTRI_OP_FLAGS_PPTR_SET:
+	case XFS_ATTRI_OP_FLAGS_PPTR_REMOVE:
+		if (!xfs_has_parent(mp))
+			return false;
+		if (attrp->alfi_value_len != sizeof(struct xfs_parent_rec))
+			return false;
+		if (!xfs_attri_validate_namelen(attrp->alfi_name_len))
+			return false;
+		if (!(attrp->alfi_attr_filter & XFS_ATTR_PARENT))
+			return false;
+		break;
 	case XFS_ATTRI_OP_FLAGS_SET:
 	case XFS_ATTRI_OP_FLAGS_REPLACE:
+		if (!xfs_is_using_logged_xattrs(mp))
+			return false;
+		if (attrp->alfi_value_len > XATTR_SIZE_MAX)
+			return false;
+		if (!xfs_attri_validate_namelen(attrp->alfi_name_len))
+			return false;
+		break;
 	case XFS_ATTRI_OP_FLAGS_REMOVE:
+		if (!xfs_is_using_logged_xattrs(mp))
+			return false;
+		if (attrp->alfi_value_len != 0)
+			return false;
+		if (!xfs_attri_validate_namelen(attrp->alfi_name_len))
+			return false;
+		break;
+	case XFS_ATTRI_OP_FLAGS_PPTR_REPLACE:
+		if (!xfs_has_parent(mp))
+			return false;
+		if (!xfs_attri_validate_namelen(attrp->alfi_old_name_len))
+			return false;
+		if (!xfs_attri_validate_namelen(attrp->alfi_new_name_len))
+			return false;
+		if (attrp->alfi_value_len != sizeof(struct xfs_parent_rec))
+			return false;
+		if (!(attrp->alfi_attr_filter & XFS_ATTR_PARENT))
+			return false;
 		break;
 	default:
 		return false;
 	}
 
-	if (attrp->alfi_value_len > XATTR_SIZE_MAX)
-		return false;
+	return xfs_verify_ino(mp, attrp->alfi_ino);
+}
 
-	if ((attrp->alfi_name_len > XATTR_NAME_MAX) ||
-	    (attrp->alfi_name_len == 0))
-		return false;
+static int
+xfs_attri_iread_extents(
+	struct xfs_inode		*ip)
+{
+	struct xfs_trans		*tp;
+	int				error;
 
-	return xfs_verify_ino(mp, attrp->alfi_ino);
+	error = xfs_trans_alloc_empty(ip->i_mount, &tp);
+	if (error)
+		return error;
+
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	error = xfs_iread_extents(tp, ip, XFS_ATTR_FORK);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	xfs_trans_cancel(tp);
+
+	return error;
 }
 
 static inline struct xfs_attr_intent *
@@ -505,20 +638,46 @@ xfs_attri_recover_work(
 {
 	struct xfs_attr_intent		*attr;
 	struct xfs_da_args		*args;
+	struct xfs_inode		*ip;
 	int				local;
 	int				error;
 
-	error = xlog_recover_iget(mp,  attrp->alfi_ino, ipp);
-	if (error)
-		return ERR_PTR(error);
+	/*
+	 * Parent pointer attr items record the generation but regular logged
+	 * xattrs do not; select the right iget function.
+	 */
+	switch (xfs_attr_log_item_op(attrp)) {
+	case XFS_ATTRI_OP_FLAGS_PPTR_SET:
+	case XFS_ATTRI_OP_FLAGS_PPTR_REPLACE:
+	case XFS_ATTRI_OP_FLAGS_PPTR_REMOVE:
+		error = xlog_recover_iget_handle(mp, attrp->alfi_ino,
+				attrp->alfi_igen, &ip);
+		break;
+	default:
+		error = xlog_recover_iget(mp, attrp->alfi_ino, &ip);
+		break;
+	}
+	if (error) {
+		xfs_irele(ip);
+		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, attrp,
+				sizeof(*attrp));
+		return ERR_PTR(-EFSCORRUPTED);
+	}
 
-	attr = kmem_zalloc(sizeof(struct xfs_attr_intent) +
-			   sizeof(struct xfs_da_args), KM_NOFS);
+	if (xfs_inode_has_attr_fork(ip)) {
+		error = xfs_attri_iread_extents(ip);
+		if (error) {
+			xfs_irele(ip);
+			return ERR_PTR(error);
+		}
+	}
+
+	attr = kzalloc(sizeof(struct xfs_attr_intent) +
+			sizeof(struct xfs_da_args), GFP_KERNEL | __GFP_NOFAIL);
 	args = (struct xfs_da_args *)(attr + 1);
 
 	attr->xattri_da_args = args;
-	attr->xattri_op_flags = attrp->alfi_op_flags &
-						XFS_ATTRI_OP_FLAGS_TYPE_MASK;
+	attr->xattri_op_flags = xfs_attr_log_item_op(attrp);
 
 	/*
 	 * We're reconstructing the deferred work state structure from the
@@ -528,35 +687,42 @@ xfs_attri_recover_work(
 	attr->xattri_nameval = xfs_attri_log_nameval_get(nv);
 	ASSERT(attr->xattri_nameval);
 
-	args->dp = *ipp;
+	args->dp = ip;
 	args->geo = mp->m_attr_geo;
 	args->whichfork = XFS_ATTR_FORK;
 	args->name = nv->name.i_addr;
 	args->namelen = nv->name.i_len;
-	args->hashval = xfs_da_hashname(args->name, args->namelen);
+	args->new_name = nv->new_name.i_addr;
+	args->new_namelen = nv->new_name.i_len;
+	args->value = nv->value.i_addr;
+	args->valuelen = nv->value.i_len;
+	args->new_value = nv->new_value.i_addr;
+	args->new_valuelen = nv->new_value.i_len;
 	args->attr_filter = attrp->alfi_attr_filter & XFS_ATTRI_FILTER_MASK;
 	args->op_flags = XFS_DA_OP_RECOVERY | XFS_DA_OP_OKNOENT |
 			 XFS_DA_OP_LOGGED;
+	args->owner = args->dp->i_ino;
+	xfs_attr_sethash(args);
 
-	ASSERT(xfs_sb_version_haslogxattrs(&mp->m_sb));
-
-	switch (attr->xattri_op_flags) {
+	switch (xfs_attr_intent_op(attr)) {
+	case XFS_ATTRI_OP_FLAGS_PPTR_SET:
+	case XFS_ATTRI_OP_FLAGS_PPTR_REPLACE:
 	case XFS_ATTRI_OP_FLAGS_SET:
 	case XFS_ATTRI_OP_FLAGS_REPLACE:
-		args->value = nv->value.i_addr;
-		args->valuelen = nv->value.i_len;
 		args->total = xfs_attr_calc_size(args, &local);
 		if (xfs_inode_hasattr(args->dp))
 			attr->xattri_dela_state = xfs_attr_init_replace_state(args);
 		else
 			attr->xattri_dela_state = xfs_attr_init_add_state(args);
 		break;
+	case XFS_ATTRI_OP_FLAGS_PPTR_REMOVE:
 	case XFS_ATTRI_OP_FLAGS_REMOVE:
 		attr->xattri_dela_state = xfs_attr_init_remove_state(args);
 		break;
 	}
 
 	xfs_defer_add_item(dfp, &attr->xattri_list);
+	*ipp = ip;
 	return attr;
 }
 
@@ -580,7 +746,7 @@ xfs_attr_recover_work(
 	struct xfs_attri_log_format	*attrp;
 	struct xfs_attri_log_nameval	*nv = attrip->attri_nameval;
 	int				error;
-	int				total;
+	unsigned int			total = 0;
 
 	/*
 	 * First check the validity of the attr described by the ATTRI.  If any
@@ -588,7 +754,8 @@ xfs_attr_recover_work(
 	 */
 	attrp = &attrip->attri_format;
 	if (!xfs_attri_validate(mp, attrp) ||
-	    !xfs_attr_namecheck(nv->name.i_addr, nv->name.i_len))
+	    !xfs_attr_namecheck(attrp->alfi_attr_filter, nv->name.i_addr,
+				nv->name.i_len))
 		return -EFSCORRUPTED;
 
 	attr = xfs_attri_recover_work(mp, dfp, attrp, &ip, nv);
@@ -596,7 +763,20 @@ xfs_attr_recover_work(
 		return PTR_ERR(attr);
 	args = attr->xattri_da_args;
 
-	xfs_init_attr_trans(args, &resv, &total);
+	switch (xfs_attr_intent_op(attr)) {
+	case XFS_ATTRI_OP_FLAGS_PPTR_SET:
+	case XFS_ATTRI_OP_FLAGS_PPTR_REPLACE:
+	case XFS_ATTRI_OP_FLAGS_SET:
+	case XFS_ATTRI_OP_FLAGS_REPLACE:
+		resv = xfs_attr_set_resv(args);
+		total = args->total;
+		break;
+	case XFS_ATTRI_OP_FLAGS_PPTR_REMOVE:
+	case XFS_ATTRI_OP_FLAGS_REMOVE:
+		resv = M_RES(mp)->tr_attrrm;
+		total = XFS_ATTRRM_SPACE_RES(mp);
+		break;
+	}
 	resv = xlog_recover_resv(&resv);
 	error = xfs_trans_alloc(mp, &resv, total, 0, XFS_TRANS_RESERVE, &tp);
 	if (error)
@@ -611,16 +791,17 @@ xfs_attr_recover_work(
 		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
 				&attrip->attri_format,
 				sizeof(attrip->attri_format));
-	if (error) {
-		xfs_trans_cancel(tp);
-		goto out_unlock;
-	}
+	if (error)
+		goto out_cancel;
 
 	error = xfs_defer_ops_capture_and_commit(tp, capture_list);
 out_unlock:
 	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 	xfs_irele(ip);
 	return error;
+out_cancel:
+	xfs_trans_cancel(tp);
+	goto out_unlock;
 }
 
 /* Re-log an intent item to push the log tail forward. */
@@ -646,9 +827,20 @@ xfs_attr_relog_intent(
 	new_attrp = &new_attrip->attri_format;
 
 	new_attrp->alfi_ino = old_attrp->alfi_ino;
+	new_attrp->alfi_igen = old_attrp->alfi_igen;
 	new_attrp->alfi_op_flags = old_attrp->alfi_op_flags;
 	new_attrp->alfi_value_len = old_attrp->alfi_value_len;
-	new_attrp->alfi_name_len = old_attrp->alfi_name_len;
+
+	switch (xfs_attr_log_item_op(old_attrp)) {
+	case XFS_ATTRI_OP_FLAGS_PPTR_REPLACE:
+		new_attrp->alfi_new_name_len = old_attrp->alfi_new_name_len;
+		new_attrp->alfi_old_name_len = old_attrp->alfi_old_name_len;
+		break;
+	default:
+		new_attrp->alfi_name_len = old_attrp->alfi_name_len;
+		break;
+	}
+
 	new_attrp->alfi_attr_filter = old_attrp->alfi_attr_filter;
 
 	return &new_attrip->attri_item;
@@ -666,7 +858,7 @@ xfs_attr_create_done(
 
 	attrip = ATTRI_ITEM(intent);
 
-	attrdp = kmem_cache_zalloc(xfs_attrd_cache, GFP_NOFS | __GFP_NOFAIL);
+	attrdp = kmem_cache_zalloc(xfs_attrd_cache, GFP_KERNEL | __GFP_NOFAIL);
 
 	xfs_log_item_init(tp->t_mountp, &attrdp->attrd_item, XFS_LI_ATTRD,
 			  &xfs_attrd_item_ops);
@@ -676,6 +868,75 @@ xfs_attr_create_done(
 	return &attrdp->attrd_item;
 }
 
+void
+xfs_attr_defer_add(
+	struct xfs_da_args	*args,
+	enum xfs_attr_defer_op	op)
+{
+	struct xfs_attr_intent	*new;
+	unsigned int		log_op = 0;
+	bool			is_pptr = args->attr_filter & XFS_ATTR_PARENT;
+
+	if (is_pptr) {
+		ASSERT(xfs_has_parent(args->dp->i_mount));
+		ASSERT((args->attr_filter & ~XFS_ATTR_PARENT) == 0);
+		ASSERT(args->op_flags & XFS_DA_OP_LOGGED);
+		ASSERT(args->valuelen == sizeof(struct xfs_parent_rec));
+	}
+
+	new = kmem_cache_zalloc(xfs_attr_intent_cache,
+			GFP_NOFS | __GFP_NOFAIL);
+	new->xattri_da_args = args;
+
+	/* Compute log operation from the higher level op and namespace. */
+	switch (op) {
+	case XFS_ATTR_DEFER_SET:
+		if (is_pptr)
+			log_op = XFS_ATTRI_OP_FLAGS_PPTR_SET;
+		else
+			log_op = XFS_ATTRI_OP_FLAGS_SET;
+		break;
+	case XFS_ATTR_DEFER_REPLACE:
+		if (is_pptr)
+			log_op = XFS_ATTRI_OP_FLAGS_PPTR_REPLACE;
+		else
+			log_op = XFS_ATTRI_OP_FLAGS_REPLACE;
+		break;
+	case XFS_ATTR_DEFER_REMOVE:
+		if (is_pptr)
+			log_op = XFS_ATTRI_OP_FLAGS_PPTR_REMOVE;
+		else
+			log_op = XFS_ATTRI_OP_FLAGS_REMOVE;
+		break;
+	default:
+		ASSERT(0);
+		break;
+	}
+	new->xattri_op_flags = log_op;
+
+	/* Set up initial attr operation state. */
+	switch (log_op) {
+	case XFS_ATTRI_OP_FLAGS_PPTR_SET:
+	case XFS_ATTRI_OP_FLAGS_SET:
+		new->xattri_dela_state = xfs_attr_init_add_state(args);
+		break;
+	case XFS_ATTRI_OP_FLAGS_PPTR_REPLACE:
+		ASSERT(args->new_valuelen == args->valuelen);
+		new->xattri_dela_state = xfs_attr_init_replace_state(args);
+		break;
+	case XFS_ATTRI_OP_FLAGS_REPLACE:
+		new->xattri_dela_state = xfs_attr_init_replace_state(args);
+		break;
+	case XFS_ATTRI_OP_FLAGS_PPTR_REMOVE:
+	case XFS_ATTRI_OP_FLAGS_REMOVE:
+		new->xattri_dela_state = xfs_attr_init_remove_state(args);
+		break;
+	}
+
+	xfs_defer_add(args->trans, &new->xattri_list, &xfs_attr_defer_type);
+	trace_xfs_attr_defer_add(new->xattri_dela_state, args->dp);
+}
+
 const struct xfs_defer_op_type xfs_attr_defer_type = {
 	.name		= "attr",
 	.max_items	= 1,
@@ -688,6 +949,56 @@ const struct xfs_defer_op_type xfs_attr_defer_type = {
 	.relog_intent	= xfs_attr_relog_intent,
 };
 
+static inline void *
+xfs_attri_validate_name_iovec(
+	struct xfs_mount		*mp,
+	struct xfs_attri_log_format     *attri_formatp,
+	const struct xfs_log_iovec	*iovec,
+	unsigned int			name_len)
+{
+	if (iovec->i_len != xlog_calc_iovec_len(name_len)) {
+		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+				attri_formatp, sizeof(*attri_formatp));
+		return NULL;
+	}
+
+	if (!xfs_attr_namecheck(attri_formatp->alfi_attr_filter, iovec->i_addr,
+				name_len)) {
+		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+				attri_formatp, sizeof(*attri_formatp));
+		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+				iovec->i_addr, iovec->i_len);
+		return NULL;
+	}
+
+	return iovec->i_addr;
+}
+
+static inline void *
+xfs_attri_validate_value_iovec(
+	struct xfs_mount		*mp,
+	struct xfs_attri_log_format     *attri_formatp,
+	const struct xfs_log_iovec	*iovec,
+	unsigned int			value_len)
+{
+	if (iovec->i_len != xlog_calc_iovec_len(value_len)) {
+		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+				attri_formatp, sizeof(*attri_formatp));
+		return NULL;
+	}
+
+	if ((attri_formatp->alfi_attr_filter & XFS_ATTR_PARENT) &&
+	    !xfs_parent_valuecheck(mp, iovec->i_addr, value_len)) {
+		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+				attri_formatp, sizeof(*attri_formatp));
+		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+				iovec->i_addr, iovec->i_len);
+		return NULL;
+	}
+
+	return iovec->i_addr;
+}
+
 STATIC int
 xlog_recover_attri_commit_pass2(
 	struct xlog                     *log,
@@ -699,51 +1010,177 @@ xlog_recover_attri_commit_pass2(
 	struct xfs_attri_log_item       *attrip;
 	struct xfs_attri_log_format     *attri_formatp;
 	struct xfs_attri_log_nameval	*nv;
-	const void			*attr_value = NULL;
 	const void			*attr_name;
+	const void			*attr_value = NULL;
+	const void			*attr_new_name = NULL;
+	const void			*attr_new_value = NULL;
 	size_t				len;
-
-	attri_formatp = item->ri_buf[0].i_addr;
-	attr_name = item->ri_buf[1].i_addr;
+	unsigned int			name_len = 0;
+	unsigned int			value_len = 0;
+	unsigned int			new_name_len = 0;
+	unsigned int			new_value_len = 0;
+	unsigned int			op, i = 0;
 
 	/* Validate xfs_attri_log_format before the large memory allocation */
 	len = sizeof(struct xfs_attri_log_format);
-	if (item->ri_buf[0].i_len != len) {
+	if (item->ri_buf[i].i_len != len) {
 		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
 				item->ri_buf[0].i_addr, item->ri_buf[0].i_len);
 		return -EFSCORRUPTED;
 	}
 
+	attri_formatp = item->ri_buf[i].i_addr;
 	if (!xfs_attri_validate(mp, attri_formatp)) {
 		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
-				item->ri_buf[0].i_addr, item->ri_buf[0].i_len);
+				attri_formatp, len);
 		return -EFSCORRUPTED;
 	}
 
-	/* Validate the attr name */
-	if (item->ri_buf[1].i_len !=
-			xlog_calc_iovec_len(attri_formatp->alfi_name_len)) {
+	/* Check the number of log iovecs makes sense for the op code. */
+	op = xfs_attr_log_item_op(attri_formatp);
+	switch (op) {
+	case XFS_ATTRI_OP_FLAGS_PPTR_REMOVE:
+	case XFS_ATTRI_OP_FLAGS_PPTR_SET:
+		/* Log item, attr name, attr value */
+		if (item->ri_total != 3) {
+			XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+					     attri_formatp, len);
+			return -EFSCORRUPTED;
+		}
+		name_len = attri_formatp->alfi_name_len;
+		value_len = attri_formatp->alfi_value_len;
+		break;
+	case XFS_ATTRI_OP_FLAGS_SET:
+	case XFS_ATTRI_OP_FLAGS_REPLACE:
+		/* Log item, attr name, attr value */
+		if (item->ri_total != 3) {
+			XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+					     attri_formatp, len);
+			return -EFSCORRUPTED;
+		}
+		name_len = attri_formatp->alfi_name_len;
+		value_len = attri_formatp->alfi_value_len;
+		break;
+	case XFS_ATTRI_OP_FLAGS_REMOVE:
+		/* Log item, attr name */
+		if (item->ri_total != 2) {
+			XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+					     attri_formatp, len);
+			return -EFSCORRUPTED;
+		}
+		name_len = attri_formatp->alfi_name_len;
+		break;
+	case XFS_ATTRI_OP_FLAGS_PPTR_REPLACE:
+		/*
+		 * Log item, attr name, new attr name, attr value, new attr
+		 * value
+		 */
+		if (item->ri_total != 5) {
+			XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+					     attri_formatp, len);
+			return -EFSCORRUPTED;
+		}
+		name_len = attri_formatp->alfi_old_name_len;
+		new_name_len = attri_formatp->alfi_new_name_len;
+		new_value_len = value_len = attri_formatp->alfi_value_len;
+		break;
+	default:
 		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
-				item->ri_buf[0].i_addr, item->ri_buf[0].i_len);
+				     attri_formatp, len);
 		return -EFSCORRUPTED;
 	}
+	i++;
 
-	if (!xfs_attr_namecheck(attr_name, attri_formatp->alfi_name_len)) {
-		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
-				item->ri_buf[1].i_addr, item->ri_buf[1].i_len);
+	/* Validate the attr name */
+	attr_name = xfs_attri_validate_name_iovec(mp, attri_formatp,
+			&item->ri_buf[i], name_len);
+	if (!attr_name)
 		return -EFSCORRUPTED;
+	i++;
+
+	/* Validate the new attr name */
+	if (new_name_len > 0) {
+		attr_new_name = xfs_attri_validate_name_iovec(mp,
+					attri_formatp, &item->ri_buf[i],
+					new_name_len);
+		if (!attr_new_name)
+			return -EFSCORRUPTED;
+		i++;
 	}
 
 	/* Validate the attr value, if present */
-	if (attri_formatp->alfi_value_len != 0) {
-		if (item->ri_buf[2].i_len != xlog_calc_iovec_len(attri_formatp->alfi_value_len)) {
+	if (value_len != 0) {
+		attr_value = xfs_attri_validate_value_iovec(mp, attri_formatp,
+				&item->ri_buf[i], value_len);
+		if (!attr_value)
+			return -EFSCORRUPTED;
+		i++;
+	}
+
+	/* Validate the new attr value, if present */
+	if (new_value_len != 0) {
+		attr_new_value = xfs_attri_validate_value_iovec(mp,
+					attri_formatp, &item->ri_buf[i],
+					new_value_len);
+		if (!attr_new_value)
+			return -EFSCORRUPTED;
+		i++;
+	}
+
+	/*
+	 * Make sure we got the correct number of buffers for the operation
+	 * that we just loaded.
+	 */
+	if (i != item->ri_total) {
+		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+				attri_formatp, len);
+		return -EFSCORRUPTED;
+	}
+
+	switch (op) {
+	case XFS_ATTRI_OP_FLAGS_REMOVE:
+		/* Regular remove operations operate only on names. */
+		if (attr_value != NULL || value_len != 0) {
 			XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
-					item->ri_buf[0].i_addr,
-					item->ri_buf[0].i_len);
+					     attri_formatp, len);
 			return -EFSCORRUPTED;
 		}
-
-		attr_value = item->ri_buf[2].i_addr;
+		fallthrough;
+	case XFS_ATTRI_OP_FLAGS_PPTR_REMOVE:
+	case XFS_ATTRI_OP_FLAGS_PPTR_SET:
+	case XFS_ATTRI_OP_FLAGS_SET:
+	case XFS_ATTRI_OP_FLAGS_REPLACE:
+		/*
+		 * Regular xattr set/remove/replace operations require a name
+		 * and do not take a newname.  Values are optional for set and
+		 * replace.
+		 *
+		 * Name-value set/remove operations must have a name, do not
+		 * take a newname, and can take a value.
+		 */
+		if (attr_name == NULL || name_len == 0) {
+			XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+					     attri_formatp, len);
+			return -EFSCORRUPTED;
+		}
+		break;
+	case XFS_ATTRI_OP_FLAGS_PPTR_REPLACE:
+		/*
+		 * Name-value replace operations require the caller to
+		 * specify the old and new names and values explicitly.
+		 * Values are optional.
+		 */
+		if (attr_name == NULL || name_len == 0) {
+			XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+					     attri_formatp, len);
+			return -EFSCORRUPTED;
+		}
+		if (attr_new_name == NULL || new_name_len == 0) {
+			XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+					     attri_formatp, len);
+			return -EFSCORRUPTED;
+		}
+		break;
 	}
 
 	/*
@@ -751,9 +1188,10 @@ xlog_recover_attri_commit_pass2(
 	 * name/value buffer to the recovered incore log item and drop our
 	 * reference.
 	 */
-	nv = xfs_attri_log_nameval_alloc(attr_name,
-			attri_formatp->alfi_name_len, attr_value,
-			attri_formatp->alfi_value_len);
+	nv = xfs_attri_log_nameval_alloc(attr_name, name_len,
+			attr_new_name, new_name_len,
+			attr_value, value_len,
+			attr_new_value, new_value_len);
 
 	attrip = xfs_attri_init(mp, nv);
 	memcpy(&attrip->attri_format, attri_formatp, len);
diff --git a/fs/xfs/xfs_attr_item.h b/fs/xfs/xfs_attr_item.h
index 3280a79302..e74128cbb7 100644
--- a/fs/xfs/xfs_attr_item.h
+++ b/fs/xfs/xfs_attr_item.h
@@ -13,7 +13,9 @@ struct kmem_zone;
 
 struct xfs_attri_log_nameval {
 	struct xfs_log_iovec	name;
+	struct xfs_log_iovec	new_name;	/* PPTR_REPLACE only */
 	struct xfs_log_iovec	value;
+	struct xfs_log_iovec	new_value;	/* PPTR_REPLACE only */
 	refcount_t		refcount;
 
 	/* name and value follow the end of this struct */
@@ -51,4 +53,12 @@ struct xfs_attrd_log_item {
 extern struct kmem_cache	*xfs_attri_cache;
 extern struct kmem_cache	*xfs_attrd_cache;
 
+enum xfs_attr_defer_op {
+	XFS_ATTR_DEFER_SET,
+	XFS_ATTR_DEFER_REMOVE,
+	XFS_ATTR_DEFER_REPLACE,
+};
+
+void xfs_attr_defer_add(struct xfs_da_args *args, enum xfs_attr_defer_op op);
+
 #endif	/* __XFS_ATTR_ITEM_H__ */
diff --git a/fs/xfs/xfs_attr_list.c b/fs/xfs/xfs_attr_list.c
index e368ad671e..5c947e5ce8 100644
--- a/fs/xfs/xfs_attr_list.c
+++ b/fs/xfs/xfs_attr_list.c
@@ -22,6 +22,7 @@
 #include "xfs_error.h"
 #include "xfs_trace.h"
 #include "xfs_dir2.h"
+#include "xfs_health.h"
 
 STATIC int
 xfs_attr_shortform_compare(const void *a, const void *b)
@@ -81,13 +82,17 @@ xfs_attr_shortform_list(
 	     (dp->i_af.if_bytes + sf->count * 16) < context->bufsize)) {
 		for (i = 0, sfe = xfs_attr_sf_firstentry(sf); i < sf->count; i++) {
 			if (XFS_IS_CORRUPT(context->dp->i_mount,
-					   !xfs_attr_namecheck(sfe->nameval,
-							       sfe->namelen)))
+					   !xfs_attr_namecheck(sfe->flags,
+							       sfe->nameval,
+							       sfe->namelen))) {
+				xfs_dirattr_mark_sick(context->dp, XFS_ATTR_FORK);
 				return -EFSCORRUPTED;
+			}
 			context->put_listent(context,
 					     sfe->flags,
 					     sfe->nameval,
 					     (int)sfe->namelen,
+					     &sfe->nameval[sfe->namelen],
 					     (int)sfe->valuelen);
 			/*
 			 * Either search callback finished early or
@@ -109,7 +114,7 @@ xfs_attr_shortform_list(
 	 * It didn't all fit, so we have to sort everything on hashval.
 	 */
 	sbsize = sf->count * sizeof(*sbuf);
-	sbp = sbuf = kmem_alloc(sbsize, KM_NOFS);
+	sbp = sbuf = kmalloc(sbsize, GFP_KERNEL | __GFP_NOFAIL);
 
 	/*
 	 * Scan the attribute list for the rest of the entries, storing
@@ -119,22 +124,28 @@ xfs_attr_shortform_list(
 	for (i = 0, sfe = xfs_attr_sf_firstentry(sf); i < sf->count; i++) {
 		if (unlikely(
 		    ((char *)sfe < (char *)sf) ||
-		    ((char *)sfe >= ((char *)sf + dp->i_af.if_bytes)))) {
+		    ((char *)sfe >= ((char *)sf + dp->i_af.if_bytes)) ||
+		    !xfs_attr_check_namespace(sfe->flags))) {
 			XFS_CORRUPTION_ERROR("xfs_attr_shortform_list",
 					     XFS_ERRLEVEL_LOW,
 					     context->dp->i_mount, sfe,
 					     sizeof(*sfe));
-			kmem_free(sbuf);
+			kfree(sbuf);
+			xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK);
 			return -EFSCORRUPTED;
 		}
 
 		sbp->entno = i;
-		sbp->hash = xfs_da_hashname(sfe->nameval, sfe->namelen);
 		sbp->name = sfe->nameval;
 		sbp->namelen = sfe->namelen;
 		/* These are bytes, and both on-disk, don't endian-flip */
+		sbp->value = &sfe->nameval[sfe->namelen],
 		sbp->valuelen = sfe->valuelen;
 		sbp->flags = sfe->flags;
+		sbp->hash = xfs_attr_hashval(dp->i_mount, sfe->flags,
+					     sfe->nameval, sfe->namelen,
+					     sfe->nameval + sfe->namelen,
+					     sfe->valuelen);
 		sfe = xfs_attr_sf_nextentry(sfe);
 		sbp++;
 		nsbuf++;
@@ -173,8 +184,9 @@ xfs_attr_shortform_list(
 			cursor->offset = 0;
 		}
 		if (XFS_IS_CORRUPT(context->dp->i_mount,
-				   !xfs_attr_namecheck(sbp->name,
+				   !xfs_attr_namecheck(sbp->flags, sbp->name,
 						       sbp->namelen))) {
+			xfs_dirattr_mark_sick(context->dp, XFS_ATTR_FORK);
 			error = -EFSCORRUPTED;
 			goto out;
 		}
@@ -182,13 +194,14 @@ xfs_attr_shortform_list(
 				     sbp->flags,
 				     sbp->name,
 				     sbp->namelen,
+				     sbp->value,
 				     sbp->valuelen);
 		if (context->seen_enough)
 			break;
 		cursor->offset++;
 	}
 out:
-	kmem_free(sbuf);
+	kfree(sbuf);
 	return error;
 }
 
@@ -209,6 +222,7 @@ xfs_attr_node_list_lookup(
 	struct xfs_mount		*mp = dp->i_mount;
 	struct xfs_trans		*tp = context->tp;
 	struct xfs_buf			*bp;
+	xfs_failaddr_t			fa;
 	int				i;
 	int				error = 0;
 	unsigned int			expected_level = 0;
@@ -233,6 +247,10 @@ xfs_attr_node_list_lookup(
 			goto out_corruptbuf;
 		}
 
+		fa = xfs_da3_node_header_check(bp, dp->i_ino);
+		if (fa)
+			goto out_corruptbuf;
+
 		xfs_da3_node_hdr_from_disk(mp, &nodehdr, node);
 
 		/* Tree taller than we can handle; bail out! */
@@ -262,8 +280,16 @@ xfs_attr_node_list_lookup(
 			return 0;
 
 		/* We can't point back to the root. */
-		if (XFS_IS_CORRUPT(mp, cursor->blkno == 0))
+		if (XFS_IS_CORRUPT(mp, cursor->blkno == 0)) {
+			xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK);
 			return -EFSCORRUPTED;
+		}
+	}
+
+	fa = xfs_attr3_leaf_header_check(bp, dp->i_ino);
+	if (fa) {
+		__xfs_buf_mark_corrupt(bp, fa);
+		goto out_releasebuf;
 	}
 
 	if (expected_level != 0)
@@ -274,7 +300,9 @@ xfs_attr_node_list_lookup(
 
 out_corruptbuf:
 	xfs_buf_mark_corrupt(bp);
+out_releasebuf:
 	xfs_trans_brelse(tp, bp);
+	xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK);
 	return -EFSCORRUPTED;
 }
 
@@ -289,6 +317,7 @@ xfs_attr_node_list(
 	struct xfs_buf			*bp;
 	struct xfs_inode		*dp = context->dp;
 	struct xfs_mount		*mp = dp->i_mount;
+	xfs_failaddr_t			fa;
 	int				error = 0;
 
 	trace_xfs_attr_node_list(context);
@@ -302,44 +331,60 @@ xfs_attr_node_list(
 	 */
 	bp = NULL;
 	if (cursor->blkno > 0) {
+		struct xfs_attr_leaf_entry *entries;
+
 		error = xfs_da3_node_read(context->tp, dp, cursor->blkno, &bp,
 				XFS_ATTR_FORK);
-		if ((error != 0) && (error != -EFSCORRUPTED))
+		if (xfs_metadata_is_sick(error))
+			xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK);
+		if (error != 0 && error != -EFSCORRUPTED)
 			return error;
-		if (bp) {
-			struct xfs_attr_leaf_entry *entries;
+		if (!bp)
+			goto need_lookup;
 
-			node = bp->b_addr;
-			switch (be16_to_cpu(node->hdr.info.magic)) {
-			case XFS_DA_NODE_MAGIC:
-			case XFS_DA3_NODE_MAGIC:
-				trace_xfs_attr_list_wrong_blk(context);
+		node = bp->b_addr;
+		switch (be16_to_cpu(node->hdr.info.magic)) {
+		case XFS_DA_NODE_MAGIC:
+		case XFS_DA3_NODE_MAGIC:
+			trace_xfs_attr_list_wrong_blk(context);
+			fa = xfs_da3_node_header_check(bp, dp->i_ino);
+			if (fa) {
+				__xfs_buf_mark_corrupt(bp, fa);
+				xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK);
+			}
+			xfs_trans_brelse(context->tp, bp);
+			bp = NULL;
+			break;
+		case XFS_ATTR_LEAF_MAGIC:
+		case XFS_ATTR3_LEAF_MAGIC:
+			leaf = bp->b_addr;
+			fa = xfs_attr3_leaf_header_check(bp, dp->i_ino);
+			if (fa) {
+				__xfs_buf_mark_corrupt(bp, fa);
 				xfs_trans_brelse(context->tp, bp);
+				xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK);
 				bp = NULL;
 				break;
-			case XFS_ATTR_LEAF_MAGIC:
-			case XFS_ATTR3_LEAF_MAGIC:
-				leaf = bp->b_addr;
-				xfs_attr3_leaf_hdr_from_disk(mp->m_attr_geo,
-							     &leafhdr, leaf);
-				entries = xfs_attr3_leaf_entryp(leaf);
-				if (cursor->hashval > be32_to_cpu(
-						entries[leafhdr.count - 1].hashval)) {
-					trace_xfs_attr_list_wrong_blk(context);
-					xfs_trans_brelse(context->tp, bp);
-					bp = NULL;
-				} else if (cursor->hashval <= be32_to_cpu(
-						entries[0].hashval)) {
-					trace_xfs_attr_list_wrong_blk(context);
-					xfs_trans_brelse(context->tp, bp);
-					bp = NULL;
-				}
-				break;
-			default:
+			}
+			xfs_attr3_leaf_hdr_from_disk(mp->m_attr_geo,
+						     &leafhdr, leaf);
+			entries = xfs_attr3_leaf_entryp(leaf);
+			if (cursor->hashval > be32_to_cpu(
+					entries[leafhdr.count - 1].hashval)) {
+				trace_xfs_attr_list_wrong_blk(context);
+				xfs_trans_brelse(context->tp, bp);
+				bp = NULL;
+			} else if (cursor->hashval <= be32_to_cpu(
+					entries[0].hashval)) {
 				trace_xfs_attr_list_wrong_blk(context);
 				xfs_trans_brelse(context->tp, bp);
 				bp = NULL;
 			}
+			break;
+		default:
+			trace_xfs_attr_list_wrong_blk(context);
+			xfs_trans_brelse(context->tp, bp);
+			bp = NULL;
 		}
 	}
 
@@ -349,6 +394,7 @@ xfs_attr_node_list(
 	 * Note that start of node block is same as start of leaf block.
 	 */
 	if (bp == NULL) {
+need_lookup:
 		error = xfs_attr_node_list_lookup(context, cursor, &bp);
 		if (error || !bp)
 			return error;
@@ -370,8 +416,8 @@ xfs_attr_node_list(
 			break;
 		cursor->blkno = leafhdr.forw;
 		xfs_trans_brelse(context->tp, bp);
-		error = xfs_attr3_leaf_read(context->tp, dp, cursor->blkno,
-					    &bp);
+		error = xfs_attr3_leaf_read(context->tp, dp, dp->i_ino,
+				cursor->blkno, &bp);
 		if (error)
 			return error;
 	}
@@ -436,6 +482,7 @@ xfs_attr3_leaf_list_int(
 	 */
 	for (; i < ichdr.count; entry++, i++) {
 		char *name;
+		void *value;
 		int namelen, valuelen;
 
 		if (be32_to_cpu(entry->hashval) != cursor->hashval) {
@@ -453,6 +500,7 @@ xfs_attr3_leaf_list_int(
 			name_loc = xfs_attr3_leaf_name_local(leaf, i);
 			name = name_loc->nameval;
 			namelen = name_loc->namelen;
+			value = &name_loc->nameval[name_loc->namelen];
 			valuelen = be16_to_cpu(name_loc->valuelen);
 		} else {
 			xfs_attr_leaf_name_remote_t *name_rmt;
@@ -460,14 +508,18 @@ xfs_attr3_leaf_list_int(
 			name_rmt = xfs_attr3_leaf_name_remote(leaf, i);
 			name = name_rmt->name;
 			namelen = name_rmt->namelen;
+			value = NULL;
 			valuelen = be32_to_cpu(name_rmt->valuelen);
 		}
 
 		if (XFS_IS_CORRUPT(context->dp->i_mount,
-				   !xfs_attr_namecheck(name, namelen)))
+				   !xfs_attr_namecheck(entry->flags, name,
+						       namelen))) {
+			xfs_dirattr_mark_sick(context->dp, XFS_ATTR_FORK);
 			return -EFSCORRUPTED;
+		}
 		context->put_listent(context, entry->flags,
-					      name, namelen, valuelen);
+					      name, namelen, value, valuelen);
 		if (context->seen_enough)
 			break;
 		cursor->offset++;
@@ -489,7 +541,8 @@ xfs_attr_leaf_list(
 	trace_xfs_attr_leaf_list(context);
 
 	context->cursor.blkno = 0;
-	error = xfs_attr3_leaf_read(context->tp, context->dp, 0, &bp);
+	error = xfs_attr3_leaf_read(context->tp, context->dp,
+			context->dp->i_ino, 0, &bp);
 	if (error)
 		return error;
 
@@ -503,8 +556,9 @@ xfs_attr_list_ilocked(
 	struct xfs_attr_list_context	*context)
 {
 	struct xfs_inode		*dp = context->dp;
+	int				error;
 
-	ASSERT(xfs_isilocked(dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL);
 
 	/*
 	 * Decide on what work routines to call based on the inode size.
@@ -513,6 +567,12 @@ xfs_attr_list_ilocked(
 		return 0;
 	if (dp->i_af.if_format == XFS_DINODE_FMT_LOCAL)
 		return xfs_attr_shortform_list(context);
+
+	/* Prerequisite for xfs_attr_is_leaf */
+	error = xfs_iread_extents(NULL, dp, XFS_ATTR_FORK);
+	if (error)
+		return error;
+
 	if (xfs_attr_is_leaf(dp))
 		return xfs_attr_leaf_list(context);
 	return xfs_attr_node_list(context);
diff --git a/fs/xfs/xfs_bmap_item.c b/fs/xfs/xfs_bmap_item.c
index 52fb8a148b..a19d62e78a 100644
--- a/fs/xfs/xfs_bmap_item.c
+++ b/fs/xfs/xfs_bmap_item.c
@@ -25,6 +25,7 @@
 #include "xfs_log_priv.h"
 #include "xfs_log_recover.h"
 #include "xfs_ag.h"
+#include "xfs_trace.h"
 
 struct kmem_cache	*xfs_bui_cache;
 struct kmem_cache	*xfs_bud_cache;
@@ -40,7 +41,7 @@ STATIC void
 xfs_bui_item_free(
 	struct xfs_bui_log_item	*buip)
 {
-	kmem_free(buip->bui_item.li_lv_shadow);
+	kvfree(buip->bui_item.li_lv_shadow);
 	kmem_cache_free(xfs_bui_cache, buip);
 }
 
@@ -201,7 +202,7 @@ xfs_bud_item_release(
 	struct xfs_bud_log_item	*budp = BUD_ITEM(lip);
 
 	xfs_bui_release(budp->bud_buip);
-	kmem_free(budp->bud_item.li_lv_shadow);
+	kvfree(budp->bud_item.li_lv_shadow);
 	kmem_cache_free(xfs_bud_cache, budp);
 }
 
@@ -221,6 +222,11 @@ static const struct xfs_item_ops xfs_bud_item_ops = {
 	.iop_intent	= xfs_bud_item_intent,
 };
 
+static inline struct xfs_bmap_intent *bi_entry(const struct list_head *e)
+{
+	return list_entry(e, struct xfs_bmap_intent, bi_list);
+}
+
 /* Sort bmap intents by inode. */
 static int
 xfs_bmap_update_diff_items(
@@ -228,37 +234,12 @@ xfs_bmap_update_diff_items(
 	const struct list_head		*a,
 	const struct list_head		*b)
 {
-	struct xfs_bmap_intent		*ba;
-	struct xfs_bmap_intent		*bb;
+	struct xfs_bmap_intent		*ba = bi_entry(a);
+	struct xfs_bmap_intent		*bb = bi_entry(b);
 
-	ba = container_of(a, struct xfs_bmap_intent, bi_list);
-	bb = container_of(b, struct xfs_bmap_intent, bi_list);
 	return ba->bi_owner->i_ino - bb->bi_owner->i_ino;
 }
 
-/* Set the map extent flags for this mapping. */
-static void
-xfs_trans_set_bmap_flags(
-	struct xfs_map_extent		*map,
-	enum xfs_bmap_intent_type	type,
-	int				whichfork,
-	xfs_exntst_t			state)
-{
-	map->me_flags = 0;
-	switch (type) {
-	case XFS_BMAP_MAP:
-	case XFS_BMAP_UNMAP:
-		map->me_flags = type;
-		break;
-	default:
-		ASSERT(0);
-	}
-	if (state == XFS_EXT_UNWRITTEN)
-		map->me_flags |= XFS_BMAP_EXTENT_UNWRITTEN;
-	if (whichfork == XFS_ATTR_FORK)
-		map->me_flags |= XFS_BMAP_EXTENT_ATTR_FORK;
-}
-
 /* Log bmap updates in the intent item. */
 STATIC void
 xfs_bmap_update_log_item(
@@ -281,8 +262,21 @@ xfs_bmap_update_log_item(
 	map->me_startblock = bi->bi_bmap.br_startblock;
 	map->me_startoff = bi->bi_bmap.br_startoff;
 	map->me_len = bi->bi_bmap.br_blockcount;
-	xfs_trans_set_bmap_flags(map, bi->bi_type, bi->bi_whichfork,
-			bi->bi_bmap.br_state);
+
+	switch (bi->bi_type) {
+	case XFS_BMAP_MAP:
+	case XFS_BMAP_UNMAP:
+		map->me_flags = bi->bi_type;
+		break;
+	default:
+		ASSERT(0);
+	}
+	if (bi->bi_bmap.br_state == XFS_EXT_UNWRITTEN)
+		map->me_flags |= XFS_BMAP_EXTENT_UNWRITTEN;
+	if (bi->bi_whichfork == XFS_ATTR_FORK)
+		map->me_flags |= XFS_BMAP_EXTENT_ATTR_FORK;
+	if (xfs_ifork_is_realtime(bi->bi_owner, bi->bi_whichfork))
+		map->me_flags |= XFS_BMAP_EXTENT_REALTIME;
 }
 
 static struct xfs_log_item *
@@ -325,13 +319,16 @@ xfs_bmap_update_create_done(
 }
 
 /* Take a passive ref to the AG containing the space we're mapping. */
-void
+static inline void
 xfs_bmap_update_get_group(
 	struct xfs_mount	*mp,
 	struct xfs_bmap_intent	*bi)
 {
 	xfs_agnumber_t		agno;
 
+	if (xfs_ifork_is_realtime(bi->bi_owner, bi->bi_whichfork))
+		return;
+
 	agno = XFS_FSB_TO_AGNO(mp, bi->bi_bmap.br_startblock);
 
 	/*
@@ -344,14 +341,40 @@ xfs_bmap_update_get_group(
 	bi->bi_pag = xfs_perag_intent_get(mp, agno);
 }
 
+/* Add this deferred BUI to the transaction. */
+void
+xfs_bmap_defer_add(
+	struct xfs_trans	*tp,
+	struct xfs_bmap_intent	*bi)
+{
+	trace_xfs_bmap_defer(bi);
+
+	xfs_bmap_update_get_group(tp->t_mountp, bi);
+	xfs_defer_add(tp, &bi->bi_list, &xfs_bmap_update_defer_type);
+}
+
 /* Release a passive AG ref after finishing mapping work. */
 static inline void
 xfs_bmap_update_put_group(
 	struct xfs_bmap_intent	*bi)
 {
+	if (xfs_ifork_is_realtime(bi->bi_owner, bi->bi_whichfork))
+		return;
+
 	xfs_perag_intent_put(bi->bi_pag);
 }
 
+/* Cancel a deferred bmap update. */
+STATIC void
+xfs_bmap_update_cancel_item(
+	struct list_head		*item)
+{
+	struct xfs_bmap_intent		*bi = bi_entry(item);
+
+	xfs_bmap_update_put_group(bi);
+	kmem_cache_free(xfs_bmap_intent_cache, bi);
+}
+
 /* Process a deferred bmap update. */
 STATIC int
 xfs_bmap_update_finish_item(
@@ -360,19 +383,16 @@ xfs_bmap_update_finish_item(
 	struct list_head		*item,
 	struct xfs_btree_cur		**state)
 {
-	struct xfs_bmap_intent		*bi;
+	struct xfs_bmap_intent		*bi = bi_entry(item);
 	int				error;
 
-	bi = container_of(item, struct xfs_bmap_intent, bi_list);
-
 	error = xfs_bmap_finish_one(tp, bi);
 	if (!error && bi->bi_bmap.br_blockcount > 0) {
 		ASSERT(bi->bi_type == XFS_BMAP_UNMAP);
 		return -EAGAIN;
 	}
 
-	xfs_bmap_update_put_group(bi);
-	kmem_cache_free(xfs_bmap_intent_cache, bi);
+	xfs_bmap_update_cancel_item(item);
 	return error;
 }
 
@@ -384,19 +404,6 @@ xfs_bmap_update_abort_intent(
 	xfs_bui_release(BUI_ITEM(intent));
 }
 
-/* Cancel a deferred bmap update. */
-STATIC void
-xfs_bmap_update_cancel_item(
-	struct list_head		*item)
-{
-	struct xfs_bmap_intent		*bi;
-
-	bi = container_of(item, struct xfs_bmap_intent, bi_list);
-
-	xfs_bmap_update_put_group(bi);
-	kmem_cache_free(xfs_bmap_intent_cache, bi);
-}
-
 /* Is this recovered BUI ok? */
 static inline bool
 xfs_bui_validate(
@@ -428,6 +435,9 @@ xfs_bui_validate(
 	if (!xfs_verify_fileext(mp, map->me_startoff, map->me_len))
 		return false;
 
+	if (map->me_flags & XFS_BMAP_EXTENT_REALTIME)
+		return xfs_verify_rtbext(mp, map->me_startblock, map->me_len);
+
 	return xfs_verify_fsbext(mp, map->me_startblock, map->me_len);
 }
 
@@ -445,7 +455,8 @@ xfs_bui_recover_work(
 	if (error)
 		return ERR_PTR(error);
 
-	bi = kmem_cache_zalloc(xfs_bmap_intent_cache, GFP_NOFS | __GFP_NOFAIL);
+	bi = kmem_cache_zalloc(xfs_bmap_intent_cache,
+			GFP_KERNEL | __GFP_NOFAIL);
 	bi->bi_whichfork = (map->me_flags & XFS_BMAP_EXTENT_ATTR_FORK) ?
 			XFS_ATTR_FORK : XFS_DATA_FORK;
 	bi->bi_type = map->me_flags & XFS_BMAP_EXTENT_TYPE_MASK;
@@ -502,14 +513,18 @@ xfs_bmap_recover_work(
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
 	xfs_trans_ijoin(tp, ip, 0);
 
+	if (!!(map->me_flags & XFS_BMAP_EXTENT_REALTIME) !=
+	    xfs_ifork_is_realtime(ip, work->bi_whichfork)) {
+		error = -EFSCORRUPTED;
+		goto err_cancel;
+	}
+
 	if (work->bi_type == XFS_BMAP_MAP)
 		iext_delta = XFS_IEXT_ADD_NOSPLIT_CNT;
 	else
 		iext_delta = XFS_IEXT_PUNCH_HOLE_CNT;
 
-	error = xfs_iext_count_may_overflow(ip, work->bi_whichfork, iext_delta);
-	if (error == -EFBIG)
-		error = xfs_iext_count_upgrade(tp, ip, iext_delta);
+	error = xfs_iext_count_extend(tp, ip, work->bi_whichfork, iext_delta);
 	if (error)
 		goto err_cancel;
 
diff --git a/fs/xfs/xfs_bmap_item.h b/fs/xfs/xfs_bmap_item.h
index 3fafd3881a..6fee6a5083 100644
--- a/fs/xfs/xfs_bmap_item.h
+++ b/fs/xfs/xfs_bmap_item.h
@@ -68,4 +68,8 @@ struct xfs_bud_log_item {
 extern struct kmem_cache	*xfs_bui_cache;
 extern struct kmem_cache	*xfs_bud_cache;
 
+struct xfs_bmap_intent;
+
+void xfs_bmap_defer_add(struct xfs_trans *tp, struct xfs_bmap_intent *bi);
+
 #endif	/* __XFS_BMAP_ITEM_H__ */
diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c
index c2531c2890..a4d9fbc21b 100644
--- a/fs/xfs/xfs_bmap_util.c
+++ b/fs/xfs/xfs_bmap_util.c
@@ -66,7 +66,7 @@ xfs_zero_extent(
 	return blkdev_issue_zeroout(target->bt_bdev,
 		block << (mp->m_super->s_blocksize_bits - 9),
 		count_fsb << (mp->m_super->s_blocksize_bits - 9),
-		GFP_NOFS, 0);
+		GFP_KERNEL, 0);
 }
 
 /*
@@ -440,7 +440,7 @@ out_unlock_iolock:
  * if the ranges only partially overlap them, so it is up to the caller to
  * ensure that partial blocks are not passed in.
  */
-int
+void
 xfs_bmap_punch_delalloc_range(
 	struct xfs_inode	*ip,
 	xfs_off_t		start_byte,
@@ -452,7 +452,6 @@ xfs_bmap_punch_delalloc_range(
 	xfs_fileoff_t		end_fsb = XFS_B_TO_FSB(mp, end_byte);
 	struct xfs_bmbt_irec	got, del;
 	struct xfs_iext_cursor	icur;
-	int			error = 0;
 
 	ASSERT(!xfs_need_iread_extents(ifp));
 
@@ -476,26 +475,22 @@ xfs_bmap_punch_delalloc_range(
 			continue;
 		}
 
-		error = xfs_bmap_del_extent_delay(ip, XFS_DATA_FORK, &icur,
-						  &got, &del);
-		if (error || !xfs_iext_get_extent(ifp, &icur, &got))
+		xfs_bmap_del_extent_delay(ip, XFS_DATA_FORK, &icur, &got, &del);
+		if (!xfs_iext_get_extent(ifp, &icur, &got))
 			break;
 	}
 
 out_unlock:
 	xfs_iunlock(ip, XFS_ILOCK_EXCL);
-	return error;
 }
 
 /*
  * Test whether it is appropriate to check an inode for and free post EOF
- * blocks. The 'force' parameter determines whether we should also consider
- * regular files that are marked preallocated or append-only.
+ * blocks.
  */
 bool
 xfs_can_free_eofblocks(
-	struct xfs_inode	*ip,
-	bool			force)
+	struct xfs_inode	*ip)
 {
 	struct xfs_bmbt_irec	imap;
 	struct xfs_mount	*mp = ip->i_mount;
@@ -508,8 +503,8 @@ xfs_can_free_eofblocks(
 	 * Caller must either hold the exclusive io lock; or be inactivating
 	 * the inode, which guarantees there are no other users of the inode.
 	 */
-	ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL) ||
-	       (VFS_I(ip)->i_state & I_FREEING));
+	if (!(VFS_I(ip)->i_state & I_FREEING))
+		xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL);
 
 	/* prealloc/delalloc exists only on regular files */
 	if (!S_ISREG(VFS_I(ip)->i_mode))
@@ -529,11 +524,11 @@ xfs_can_free_eofblocks(
 		return false;
 
 	/*
-	 * Do not free real preallocated or append-only files unless the file
-	 * has delalloc blocks and we are forced to remove them.
+	 * Only free real extents for inodes with persistent preallocations or
+	 * the append-only flag.
 	 */
 	if (ip->i_diflags & (XFS_DIFLAG_PREALLOC | XFS_DIFLAG_APPEND))
-		if (!force || ip->i_delayed_blks == 0)
+		if (ip->i_delayed_blks == 0)
 			return false;
 
 	/*
@@ -542,7 +537,7 @@ xfs_can_free_eofblocks(
 	 * forever.
 	 */
 	end_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)XFS_ISIZE(ip));
-	if (XFS_IS_REALTIME_INODE(ip) && mp->m_sb.sb_rextsize > 1)
+	if (xfs_inode_has_bigrtalloc(ip))
 		end_fsb = xfs_rtb_roundup_rtx(mp, end_fsb);
 	last_fsb = XFS_B_TO_FSB(mp, mp->m_super->s_maxbytes);
 	if (last_fsb <= end_fsb)
@@ -587,6 +582,22 @@ xfs_free_eofblocks(
 	/* Wait on dio to ensure i_size has settled. */
 	inode_dio_wait(VFS_I(ip));
 
+	/*
+	 * For preallocated files only free delayed allocations.
+	 *
+	 * Note that this means we also leave speculative preallocations in
+	 * place for preallocated files.
+	 */
+	if (ip->i_diflags & (XFS_DIFLAG_PREALLOC | XFS_DIFLAG_APPEND)) {
+		if (ip->i_delayed_blks) {
+			xfs_bmap_punch_delalloc_range(ip,
+				round_up(XFS_ISIZE(ip), mp->m_sb.sb_blocksize),
+				LLONG_MAX);
+		}
+		xfs_inode_clear_eofblocks_tag(ip);
+		return 0;
+	}
+
 	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, &tp);
 	if (error) {
 		ASSERT(xfs_is_shutdown(mp));
@@ -713,41 +724,37 @@ xfs_alloc_file_space(
 		if (error)
 			break;
 
-		error = xfs_iext_count_may_overflow(ip, XFS_DATA_FORK,
+		error = xfs_iext_count_extend(tp, ip, XFS_DATA_FORK,
 				XFS_IEXT_ADD_NOSPLIT_CNT);
-		if (error == -EFBIG)
-			error = xfs_iext_count_upgrade(tp, ip,
-					XFS_IEXT_ADD_NOSPLIT_CNT);
-		if (error)
-			goto error;
-
-		error = xfs_bmapi_write(tp, ip, startoffset_fsb,
-				allocatesize_fsb, XFS_BMAPI_PREALLOC, 0, imapp,
-				&nimaps);
 		if (error)
 			goto error;
 
-		ip->i_diflags |= XFS_DIFLAG_PREALLOC;
-		xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
-
-		error = xfs_trans_commit(tp);
-		xfs_iunlock(ip, XFS_ILOCK_EXCL);
-		if (error)
-			break;
-
 		/*
 		 * If the allocator cannot find a single free extent large
 		 * enough to cover the start block of the requested range,
-		 * xfs_bmapi_write will return 0 but leave *nimaps set to 0.
+		 * xfs_bmapi_write will return -ENOSR.
 		 *
 		 * In that case we simply need to keep looping with the same
 		 * startoffset_fsb so that one of the following allocations
 		 * will eventually reach the requested range.
 		 */
-		if (nimaps) {
+		error = xfs_bmapi_write(tp, ip, startoffset_fsb,
+				allocatesize_fsb, XFS_BMAPI_PREALLOC, 0, imapp,
+				&nimaps);
+		if (error) {
+			if (error != -ENOSR)
+				goto error;
+			error = 0;
+		} else {
 			startoffset_fsb += imapp->br_blockcount;
 			allocatesize_fsb -= imapp->br_blockcount;
 		}
+
+		ip->i_diflags |= XFS_DIFLAG_PREALLOC;
+		xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+
+		error = xfs_trans_commit(tp);
+		xfs_iunlock(ip, XFS_ILOCK_EXCL);
 	}
 
 	return error;
@@ -775,10 +782,8 @@ xfs_unmap_extent(
 	if (error)
 		return error;
 
-	error = xfs_iext_count_may_overflow(ip, XFS_DATA_FORK,
+	error = xfs_iext_count_extend(tp, ip, XFS_DATA_FORK,
 			XFS_IEXT_PUNCH_HOLE_CNT);
-	if (error == -EFBIG)
-		error = xfs_iext_count_upgrade(tp, ip, XFS_IEXT_PUNCH_HOLE_CNT);
 	if (error)
 		goto out_trans_cancel;
 
@@ -843,7 +848,7 @@ xfs_free_file_space(
 	endoffset_fsb = XFS_B_TO_FSBT(mp, offset + len);
 
 	/* We can only free complete realtime extents. */
-	if (XFS_IS_REALTIME_INODE(ip) && mp->m_sb.sb_rextsize > 1) {
+	if (xfs_inode_has_bigrtalloc(ip)) {
 		startoffset_fsb = xfs_rtb_roundup_rtx(mp, startoffset_fsb);
 		endoffset_fsb = xfs_rtb_rounddown_rtx(mp, endoffset_fsb);
 	}
@@ -900,7 +905,7 @@ xfs_prepare_shift(
 	 * Trim eofblocks to avoid shifting uninitialized post-eof preallocation
 	 * into the accessible region of the file.
 	 */
-	if (xfs_can_free_eofblocks(ip, true)) {
+	if (xfs_can_free_eofblocks(ip)) {
 		error = xfs_free_eofblocks(ip);
 		if (error)
 			return error;
@@ -965,8 +970,7 @@ xfs_collapse_file_space(
 	xfs_fileoff_t		shift_fsb = XFS_B_TO_FSB(mp, len);
 	bool			done = false;
 
-	ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL));
-	ASSERT(xfs_isilocked(ip, XFS_MMAPLOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL);
 
 	trace_xfs_collapse_file_space(ip);
 
@@ -1035,8 +1039,7 @@ xfs_insert_file_space(
 	xfs_fileoff_t		shift_fsb = XFS_B_TO_FSB(mp, len);
 	bool			done = false;
 
-	ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL));
-	ASSERT(xfs_isilocked(ip, XFS_MMAPLOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL);
 
 	trace_xfs_insert_file_space(ip);
 
@@ -1056,10 +1059,8 @@ xfs_insert_file_space(
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
 	xfs_trans_ijoin(tp, ip, 0);
 
-	error = xfs_iext_count_may_overflow(ip, XFS_DATA_FORK,
+	error = xfs_iext_count_extend(tp, ip, XFS_DATA_FORK,
 			XFS_IEXT_PUNCH_HOLE_CNT);
-	if (error == -EFBIG)
-		error = xfs_iext_count_upgrade(tp, ip, XFS_IEXT_PUNCH_HOLE_CNT);
 	if (error)
 		goto out_trans_cancel;
 
@@ -1285,38 +1286,32 @@ xfs_swap_extent_rmap(
 			trace_xfs_swap_extent_rmap_remap_piece(tip, &uirec);
 
 			if (xfs_bmap_is_real_extent(&uirec)) {
-				error = xfs_iext_count_may_overflow(ip,
+				error = xfs_iext_count_extend(tp, ip,
 						XFS_DATA_FORK,
 						XFS_IEXT_SWAP_RMAP_CNT);
-				if (error == -EFBIG)
-					error = xfs_iext_count_upgrade(tp, ip,
-							XFS_IEXT_SWAP_RMAP_CNT);
 				if (error)
 					goto out;
 			}
 
 			if (xfs_bmap_is_real_extent(&irec)) {
-				error = xfs_iext_count_may_overflow(tip,
+				error = xfs_iext_count_extend(tp, tip,
 						XFS_DATA_FORK,
 						XFS_IEXT_SWAP_RMAP_CNT);
-				if (error == -EFBIG)
-					error = xfs_iext_count_upgrade(tp, ip,
-							XFS_IEXT_SWAP_RMAP_CNT);
 				if (error)
 					goto out;
 			}
 
 			/* Remove the mapping from the donor file. */
-			xfs_bmap_unmap_extent(tp, tip, &uirec);
+			xfs_bmap_unmap_extent(tp, tip, XFS_DATA_FORK, &uirec);
 
 			/* Remove the mapping from the source file. */
-			xfs_bmap_unmap_extent(tp, ip, &irec);
+			xfs_bmap_unmap_extent(tp, ip, XFS_DATA_FORK, &irec);
 
 			/* Map the donor file's blocks into the source file. */
-			xfs_bmap_map_extent(tp, ip, &uirec);
+			xfs_bmap_map_extent(tp, ip, XFS_DATA_FORK, &uirec);
 
 			/* Map the source file's blocks into the donor file. */
-			xfs_bmap_map_extent(tp, tip, &irec);
+			xfs_bmap_map_extent(tp, tip, XFS_DATA_FORK, &irec);
 
 			error = xfs_defer_finish(tpp);
 			tp = *tpp;
diff --git a/fs/xfs/xfs_bmap_util.h b/fs/xfs/xfs_bmap_util.h
index 77ecbb753e..eb0895bfb9 100644
--- a/fs/xfs/xfs_bmap_util.h
+++ b/fs/xfs/xfs_bmap_util.h
@@ -30,7 +30,7 @@ xfs_bmap_rtalloc(struct xfs_bmalloca *ap)
 }
 #endif /* CONFIG_XFS_RT */
 
-int	xfs_bmap_punch_delalloc_range(struct xfs_inode *ip,
+void	xfs_bmap_punch_delalloc_range(struct xfs_inode *ip,
 		xfs_off_t start_byte, xfs_off_t end_byte);
 
 struct kgetbmap {
@@ -63,7 +63,7 @@ int	xfs_insert_file_space(struct xfs_inode *, xfs_off_t offset,
 				xfs_off_t len);
 
 /* EOF block manipulation functions */
-bool	xfs_can_free_eofblocks(struct xfs_inode *ip, bool force);
+bool	xfs_can_free_eofblocks(struct xfs_inode *ip);
 int	xfs_free_eofblocks(struct xfs_inode *ip);
 
 int	xfs_swap_extents(struct xfs_inode *ip, struct xfs_inode *tip,
diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c
index 8e5bd50d29..aa4dbda7b5 100644
--- a/fs/xfs/xfs_buf.c
+++ b/fs/xfs/xfs_buf.c
@@ -21,6 +21,7 @@
 #include "xfs_errortag.h"
 #include "xfs_error.h"
 #include "xfs_ag.h"
+#include "xfs_buf_mem.h"
 
 struct kmem_cache *xfs_buf_cache;
 
@@ -60,6 +61,11 @@ xfs_buf_submit(
 	return __xfs_buf_submit(bp, !(bp->b_flags & XBF_ASYNC));
 }
 
+static inline bool xfs_buf_is_uncached(struct xfs_buf *bp)
+{
+	return bp->b_rhash_key == XFS_BUF_DADDR_NULL;
+}
+
 static inline int
 xfs_buf_is_vmapped(
 	struct xfs_buf	*bp)
@@ -189,8 +195,8 @@ xfs_buf_get_maps(
 		return 0;
 	}
 
-	bp->b_maps = kmem_zalloc(map_count * sizeof(struct xfs_buf_map),
-				KM_NOFS);
+	bp->b_maps = kzalloc(map_count * sizeof(struct xfs_buf_map),
+			GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL);
 	if (!bp->b_maps)
 		return -ENOMEM;
 	return 0;
@@ -204,7 +210,7 @@ xfs_buf_free_maps(
 	struct xfs_buf	*bp)
 {
 	if (bp->b_maps != &bp->__b_map) {
-		kmem_free(bp->b_maps);
+		kfree(bp->b_maps);
 		bp->b_maps = NULL;
 	}
 }
@@ -222,7 +228,8 @@ _xfs_buf_alloc(
 	int			i;
 
 	*bpp = NULL;
-	bp = kmem_cache_zalloc(xfs_buf_cache, GFP_NOFS | __GFP_NOFAIL);
+	bp = kmem_cache_zalloc(xfs_buf_cache,
+			GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL);
 
 	/*
 	 * We don't want certain flags to appear in b_flags unless they are
@@ -289,7 +296,7 @@ xfs_buf_free_pages(
 	mm_account_reclaimed_pages(bp->b_page_count);
 
 	if (bp->b_pages != bp->b_page_array)
-		kmem_free(bp->b_pages);
+		kfree(bp->b_pages);
 	bp->b_pages = NULL;
 	bp->b_flags &= ~_XBF_PAGES;
 }
@@ -312,10 +319,12 @@ xfs_buf_free(
 
 	ASSERT(list_empty(&bp->b_lru));
 
-	if (bp->b_flags & _XBF_PAGES)
+	if (xfs_buftarg_is_mem(bp->b_target))
+		xmbuf_unmap_page(bp);
+	else if (bp->b_flags & _XBF_PAGES)
 		xfs_buf_free_pages(bp);
 	else if (bp->b_flags & _XBF_KMEM)
-		kmem_free(bp->b_addr);
+		kfree(bp->b_addr);
 
 	call_rcu(&bp->b_rcu, xfs_buf_free_callback);
 }
@@ -325,21 +334,21 @@ xfs_buf_alloc_kmem(
 	struct xfs_buf	*bp,
 	xfs_buf_flags_t	flags)
 {
-	xfs_km_flags_t	kmflag_mask = KM_NOFS;
+	gfp_t		gfp_mask = GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL;
 	size_t		size = BBTOB(bp->b_length);
 
 	/* Assure zeroed buffer for non-read cases. */
 	if (!(flags & XBF_READ))
-		kmflag_mask |= KM_ZERO;
+		gfp_mask |= __GFP_ZERO;
 
-	bp->b_addr = kmem_alloc(size, kmflag_mask);
+	bp->b_addr = kmalloc(size, gfp_mask);
 	if (!bp->b_addr)
 		return -ENOMEM;
 
 	if (((unsigned long)(bp->b_addr + size - 1) & PAGE_MASK) !=
 	    ((unsigned long)bp->b_addr & PAGE_MASK)) {
 		/* b_addr spans two pages - use alloc_page instead */
-		kmem_free(bp->b_addr);
+		kfree(bp->b_addr);
 		bp->b_addr = NULL;
 		return -ENOMEM;
 	}
@@ -356,13 +365,11 @@ xfs_buf_alloc_pages(
 	struct xfs_buf	*bp,
 	xfs_buf_flags_t	flags)
 {
-	gfp_t		gfp_mask = __GFP_NOWARN;
+	gfp_t		gfp_mask = GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOWARN;
 	long		filled = 0;
 
 	if (flags & XBF_READ_AHEAD)
 		gfp_mask |= __GFP_NORETRY;
-	else
-		gfp_mask |= GFP_NOFS;
 
 	/* Make sure that we have a page list */
 	bp->b_page_count = DIV_ROUND_UP(BBTOB(bp->b_length), PAGE_SIZE);
@@ -429,11 +436,18 @@ _xfs_buf_map_pages(
 
 		/*
 		 * vm_map_ram() will allocate auxiliary structures (e.g.
-		 * pagetables) with GFP_KERNEL, yet we are likely to be under
-		 * GFP_NOFS context here. Hence we need to tell memory reclaim
-		 * that we are in such a context via PF_MEMALLOC_NOFS to prevent
-		 * memory reclaim re-entering the filesystem here and
-		 * potentially deadlocking.
+		 * pagetables) with GFP_KERNEL, yet we often under a scoped nofs
+		 * context here. Mixing GFP_KERNEL with GFP_NOFS allocations
+		 * from the same call site that can be run from both above and
+		 * below memory reclaim causes lockdep false positives. Hence we
+		 * always need to force this allocation to nofs context because
+		 * we can't pass __GFP_NOLOCKDEP down to auxillary structures to
+		 * prevent false positive lockdep reports.
+		 *
+		 * XXX(dgc): I think dquot reclaim is the only place we can get
+		 * to this function from memory reclaim context now. If we fix
+		 * that like we've fixed inode reclaim to avoid writeback from
+		 * reclaim, this nofs wrapping can go away.
 		 */
 		nofs_flag = memalloc_nofs_save();
 		do {
@@ -480,6 +494,9 @@ _xfs_buf_obj_cmp(
 		 * it stale has not yet committed. i.e. we are
 		 * reallocating a busy extent. Skip this buffer and
 		 * continue searching for an exact match.
+		 *
+		 * Note: If we're scanning for incore buffers to stale, don't
+		 * complain if we find non-stale buffers.
 		 */
 		if (!(map->bm_flags & XBM_LIVESCAN))
 			ASSERT(bp->b_flags & XBF_STALE);
@@ -499,18 +516,18 @@ static const struct rhashtable_params xfs_buf_hash_params = {
 };
 
 int
-xfs_buf_hash_init(
-	struct xfs_perag	*pag)
+xfs_buf_cache_init(
+	struct xfs_buf_cache	*bch)
 {
-	spin_lock_init(&pag->pag_buf_lock);
-	return rhashtable_init(&pag->pag_buf_hash, &xfs_buf_hash_params);
+	spin_lock_init(&bch->bc_lock);
+	return rhashtable_init(&bch->bc_hash, &xfs_buf_hash_params);
 }
 
 void
-xfs_buf_hash_destroy(
-	struct xfs_perag	*pag)
+xfs_buf_cache_destroy(
+	struct xfs_buf_cache	*bch)
 {
-	rhashtable_destroy(&pag->pag_buf_hash);
+	rhashtable_destroy(&bch->bc_hash);
 }
 
 static int
@@ -573,7 +590,7 @@ xfs_buf_find_lock(
 
 static inline int
 xfs_buf_lookup(
-	struct xfs_perag	*pag,
+	struct xfs_buf_cache	*bch,
 	struct xfs_buf_map	*map,
 	xfs_buf_flags_t		flags,
 	struct xfs_buf		**bpp)
@@ -582,7 +599,7 @@ xfs_buf_lookup(
 	int			error;
 
 	rcu_read_lock();
-	bp = rhashtable_lookup(&pag->pag_buf_hash, map, xfs_buf_hash_params);
+	bp = rhashtable_lookup(&bch->bc_hash, map, xfs_buf_hash_params);
 	if (!bp || !atomic_inc_not_zero(&bp->b_hold)) {
 		rcu_read_unlock();
 		return -ENOENT;
@@ -607,6 +624,7 @@ xfs_buf_lookup(
 static int
 xfs_buf_find_insert(
 	struct xfs_buftarg	*btp,
+	struct xfs_buf_cache	*bch,
 	struct xfs_perag	*pag,
 	struct xfs_buf_map	*cmap,
 	struct xfs_buf_map	*map,
@@ -622,31 +640,33 @@ xfs_buf_find_insert(
 	if (error)
 		goto out_drop_pag;
 
-	/*
-	 * For buffers that fit entirely within a single page, first attempt to
-	 * allocate the memory from the heap to minimise memory usage. If we
-	 * can't get heap memory for these small buffers, we fall back to using
-	 * the page allocator.
-	 */
-	if (BBTOB(new_bp->b_length) >= PAGE_SIZE ||
-	    xfs_buf_alloc_kmem(new_bp, flags) < 0) {
+	if (xfs_buftarg_is_mem(new_bp->b_target)) {
+		error = xmbuf_map_page(new_bp);
+	} else if (BBTOB(new_bp->b_length) >= PAGE_SIZE ||
+		   xfs_buf_alloc_kmem(new_bp, flags) < 0) {
+		/*
+		 * For buffers that fit entirely within a single page, first
+		 * attempt to allocate the memory from the heap to minimise
+		 * memory usage. If we can't get heap memory for these small
+		 * buffers, we fall back to using the page allocator.
+		 */
 		error = xfs_buf_alloc_pages(new_bp, flags);
-		if (error)
-			goto out_free_buf;
 	}
+	if (error)
+		goto out_free_buf;
 
-	spin_lock(&pag->pag_buf_lock);
-	bp = rhashtable_lookup_get_insert_fast(&pag->pag_buf_hash,
+	spin_lock(&bch->bc_lock);
+	bp = rhashtable_lookup_get_insert_fast(&bch->bc_hash,
 			&new_bp->b_rhash_head, xfs_buf_hash_params);
 	if (IS_ERR(bp)) {
 		error = PTR_ERR(bp);
-		spin_unlock(&pag->pag_buf_lock);
+		spin_unlock(&bch->bc_lock);
 		goto out_free_buf;
 	}
 	if (bp) {
 		/* found an existing buffer */
 		atomic_inc(&bp->b_hold);
-		spin_unlock(&pag->pag_buf_lock);
+		spin_unlock(&bch->bc_lock);
 		error = xfs_buf_find_lock(bp, flags);
 		if (error)
 			xfs_buf_rele(bp);
@@ -657,17 +677,40 @@ xfs_buf_find_insert(
 
 	/* The new buffer keeps the perag reference until it is freed. */
 	new_bp->b_pag = pag;
-	spin_unlock(&pag->pag_buf_lock);
+	spin_unlock(&bch->bc_lock);
 	*bpp = new_bp;
 	return 0;
 
 out_free_buf:
 	xfs_buf_free(new_bp);
 out_drop_pag:
-	xfs_perag_put(pag);
+	if (pag)
+		xfs_perag_put(pag);
 	return error;
 }
 
+static inline struct xfs_perag *
+xfs_buftarg_get_pag(
+	struct xfs_buftarg		*btp,
+	const struct xfs_buf_map	*map)
+{
+	struct xfs_mount		*mp = btp->bt_mount;
+
+	if (xfs_buftarg_is_mem(btp))
+		return NULL;
+	return xfs_perag_get(mp, xfs_daddr_to_agno(mp, map->bm_bn));
+}
+
+static inline struct xfs_buf_cache *
+xfs_buftarg_buf_cache(
+	struct xfs_buftarg		*btp,
+	struct xfs_perag		*pag)
+{
+	if (pag)
+		return &pag->pag_bcache;
+	return btp->bt_cache;
+}
+
 /*
  * Assembles a buffer covering the specified range. The code is optimised for
  * cache hits, as metadata intensive workloads will see 3 orders of magnitude
@@ -681,6 +724,7 @@ xfs_buf_get_map(
 	xfs_buf_flags_t		flags,
 	struct xfs_buf		**bpp)
 {
+	struct xfs_buf_cache	*bch;
 	struct xfs_perag	*pag;
 	struct xfs_buf		*bp = NULL;
 	struct xfs_buf_map	cmap = { .bm_bn = map[0].bm_bn };
@@ -696,10 +740,10 @@ xfs_buf_get_map(
 	if (error)
 		return error;
 
-	pag = xfs_perag_get(btp->bt_mount,
-			    xfs_daddr_to_agno(btp->bt_mount, cmap.bm_bn));
+	pag = xfs_buftarg_get_pag(btp, &cmap);
+	bch = xfs_buftarg_buf_cache(btp, pag);
 
-	error = xfs_buf_lookup(pag, &cmap, flags, &bp);
+	error = xfs_buf_lookup(bch, &cmap, flags, &bp);
 	if (error && error != -ENOENT)
 		goto out_put_perag;
 
@@ -711,13 +755,14 @@ xfs_buf_get_map(
 			goto out_put_perag;
 
 		/* xfs_buf_find_insert() consumes the perag reference. */
-		error = xfs_buf_find_insert(btp, pag, &cmap, map, nmaps,
+		error = xfs_buf_find_insert(btp, bch, pag, &cmap, map, nmaps,
 				flags, &bp);
 		if (error)
 			return error;
 	} else {
 		XFS_STATS_INC(btp->bt_mount, xb_get_locked);
-		xfs_perag_put(pag);
+		if (pag)
+			xfs_perag_put(pag);
 	}
 
 	/* We do not hold a perag reference anymore. */
@@ -745,7 +790,8 @@ xfs_buf_get_map(
 	return 0;
 
 out_put_perag:
-	xfs_perag_put(pag);
+	if (pag)
+		xfs_perag_put(pag);
 	return error;
 }
 
@@ -892,6 +938,13 @@ xfs_buf_readahead_map(
 {
 	struct xfs_buf		*bp;
 
+	/*
+	 * Currently we don't have a good means or justification for performing
+	 * xmbuf_map_page asynchronously, so we don't do readahead.
+	 */
+	if (xfs_buftarg_is_mem(target))
+		return;
+
 	xfs_buf_read_map(target, map, nmaps,
 		     XBF_TRYLOCK | XBF_ASYNC | XBF_READ_AHEAD, &bp, ops,
 		     __this_address);
@@ -957,7 +1010,10 @@ xfs_buf_get_uncached(
 	if (error)
 		return error;
 
-	error = xfs_buf_alloc_pages(bp, flags);
+	if (xfs_buftarg_is_mem(bp->b_target))
+		error = xmbuf_map_page(bp);
+	else
+		error = xfs_buf_alloc_pages(bp, flags);
 	if (error)
 		goto fail_free_buf;
 
@@ -990,29 +1046,29 @@ xfs_buf_hold(
 	atomic_inc(&bp->b_hold);
 }
 
-/*
- * Release a hold on the specified buffer. If the hold count is 1, the buffer is
- * placed on LRU or freed (depending on b_lru_ref).
- */
-void
-xfs_buf_rele(
+static void
+xfs_buf_rele_uncached(
+	struct xfs_buf		*bp)
+{
+	ASSERT(list_empty(&bp->b_lru));
+	if (atomic_dec_and_test(&bp->b_hold)) {
+		xfs_buf_ioacct_dec(bp);
+		xfs_buf_free(bp);
+	}
+}
+
+static void
+xfs_buf_rele_cached(
 	struct xfs_buf		*bp)
 {
+	struct xfs_buftarg	*btp = bp->b_target;
 	struct xfs_perag	*pag = bp->b_pag;
+	struct xfs_buf_cache	*bch = xfs_buftarg_buf_cache(btp, pag);
 	bool			release;
 	bool			freebuf = false;
 
 	trace_xfs_buf_rele(bp, _RET_IP_);
 
-	if (!pag) {
-		ASSERT(list_empty(&bp->b_lru));
-		if (atomic_dec_and_test(&bp->b_hold)) {
-			xfs_buf_ioacct_dec(bp);
-			xfs_buf_free(bp);
-		}
-		return;
-	}
-
 	ASSERT(atomic_read(&bp->b_hold) > 0);
 
 	/*
@@ -1026,7 +1082,7 @@ xfs_buf_rele(
 	 * leading to a use-after-free scenario.
 	 */
 	spin_lock(&bp->b_lock);
-	release = atomic_dec_and_lock(&bp->b_hold, &pag->pag_buf_lock);
+	release = atomic_dec_and_lock(&bp->b_hold, &bch->bc_lock);
 	if (!release) {
 		/*
 		 * Drop the in-flight state if the buffer is already on the LRU
@@ -1047,11 +1103,11 @@ xfs_buf_rele(
 		 * buffer for the LRU and clear the (now stale) dispose list
 		 * state flag
 		 */
-		if (list_lru_add_obj(&bp->b_target->bt_lru, &bp->b_lru)) {
+		if (list_lru_add_obj(&btp->bt_lru, &bp->b_lru)) {
 			bp->b_state &= ~XFS_BSTATE_DISPOSE;
 			atomic_inc(&bp->b_hold);
 		}
-		spin_unlock(&pag->pag_buf_lock);
+		spin_unlock(&bch->bc_lock);
 	} else {
 		/*
 		 * most of the time buffers will already be removed from the
@@ -1060,16 +1116,17 @@ xfs_buf_rele(
 		 * was on was the disposal list
 		 */
 		if (!(bp->b_state & XFS_BSTATE_DISPOSE)) {
-			list_lru_del_obj(&bp->b_target->bt_lru, &bp->b_lru);
+			list_lru_del_obj(&btp->bt_lru, &bp->b_lru);
 		} else {
 			ASSERT(list_empty(&bp->b_lru));
 		}
 
 		ASSERT(!(bp->b_flags & _XBF_DELWRI_Q));
-		rhashtable_remove_fast(&pag->pag_buf_hash, &bp->b_rhash_head,
-				       xfs_buf_hash_params);
-		spin_unlock(&pag->pag_buf_lock);
-		xfs_perag_put(pag);
+		rhashtable_remove_fast(&bch->bc_hash, &bp->b_rhash_head,
+				xfs_buf_hash_params);
+		spin_unlock(&bch->bc_lock);
+		if (pag)
+			xfs_perag_put(pag);
 		freebuf = true;
 	}
 
@@ -1080,6 +1137,19 @@ out_unlock:
 		xfs_buf_free(bp);
 }
 
+/*
+ * Release a hold on the specified buffer.
+ */
+void
+xfs_buf_rele(
+	struct xfs_buf		*bp)
+{
+	trace_xfs_buf_rele(bp, _RET_IP_);
+	if (xfs_buf_is_uncached(bp))
+		xfs_buf_rele_uncached(bp);
+	else
+		xfs_buf_rele_cached(bp);
+}
 
 /*
  *	Lock a buffer object, if it is not already locked.
@@ -1585,6 +1655,12 @@ _xfs_buf_ioapply(
 	/* we only use the buffer cache for meta-data */
 	op |= REQ_META;
 
+	/* in-memory targets are directly mapped, no IO required. */
+	if (xfs_buftarg_is_mem(bp->b_target)) {
+		xfs_buf_ioend(bp);
+		return;
+	}
+
 	/*
 	 * Walk all the vectors issuing IO on them. Set up the initial offset
 	 * into the buffer and the desired IO size before we start -
@@ -1940,112 +2016,119 @@ xfs_buftarg_shrink_count(
 }
 
 void
-xfs_free_buftarg(
+xfs_destroy_buftarg(
 	struct xfs_buftarg	*btp)
 {
 	shrinker_free(btp->bt_shrinker);
 	ASSERT(percpu_counter_sum(&btp->bt_io_count) == 0);
 	percpu_counter_destroy(&btp->bt_io_count);
 	list_lru_destroy(&btp->bt_lru);
+}
 
+void
+xfs_free_buftarg(
+	struct xfs_buftarg	*btp)
+{
+	xfs_destroy_buftarg(btp);
 	fs_put_dax(btp->bt_daxdev, btp->bt_mount);
 	/* the main block device is closed by kill_block_super */
 	if (btp->bt_bdev != btp->bt_mount->m_super->s_bdev)
-		bdev_release(btp->bt_bdev_handle);
-
-	kmem_free(btp);
+		bdev_fput(btp->bt_bdev_file);
+	kfree(btp);
 }
 
 int
 xfs_setsize_buftarg(
-	xfs_buftarg_t		*btp,
+	struct xfs_buftarg	*btp,
 	unsigned int		sectorsize)
 {
 	/* Set up metadata sector size info */
 	btp->bt_meta_sectorsize = sectorsize;
 	btp->bt_meta_sectormask = sectorsize - 1;
 
-	if (set_blocksize(btp->bt_bdev, sectorsize)) {
+	if (set_blocksize(btp->bt_bdev_file, sectorsize)) {
 		xfs_warn(btp->bt_mount,
 			"Cannot set_blocksize to %u on device %pg",
 			sectorsize, btp->bt_bdev);
 		return -EINVAL;
 	}
 
-	/* Set up device logical sector size mask */
-	btp->bt_logical_sectorsize = bdev_logical_block_size(btp->bt_bdev);
-	btp->bt_logical_sectormask = bdev_logical_block_size(btp->bt_bdev) - 1;
-
 	return 0;
 }
 
-/*
- * When allocating the initial buffer target we have not yet
- * read in the superblock, so don't know what sized sectors
- * are being used at this early stage.  Play safe.
- */
-STATIC int
-xfs_setsize_buftarg_early(
-	xfs_buftarg_t		*btp)
+int
+xfs_init_buftarg(
+	struct xfs_buftarg		*btp,
+	size_t				logical_sectorsize,
+	const char			*descr)
 {
-	return xfs_setsize_buftarg(btp, bdev_logical_block_size(btp->bt_bdev));
+	/* Set up device logical sector size mask */
+	btp->bt_logical_sectorsize = logical_sectorsize;
+	btp->bt_logical_sectormask = logical_sectorsize - 1;
+
+	/*
+	 * Buffer IO error rate limiting. Limit it to no more than 10 messages
+	 * per 30 seconds so as to not spam logs too much on repeated errors.
+	 */
+	ratelimit_state_init(&btp->bt_ioerror_rl, 30 * HZ,
+			     DEFAULT_RATELIMIT_BURST);
+
+	if (list_lru_init(&btp->bt_lru))
+		return -ENOMEM;
+	if (percpu_counter_init(&btp->bt_io_count, 0, GFP_KERNEL))
+		goto out_destroy_lru;
+
+	btp->bt_shrinker =
+		shrinker_alloc(SHRINKER_NUMA_AWARE, "xfs-buf:%s", descr);
+	if (!btp->bt_shrinker)
+		goto out_destroy_io_count;
+	btp->bt_shrinker->count_objects = xfs_buftarg_shrink_count;
+	btp->bt_shrinker->scan_objects = xfs_buftarg_shrink_scan;
+	btp->bt_shrinker->private_data = btp;
+	shrinker_register(btp->bt_shrinker);
+	return 0;
+
+out_destroy_io_count:
+	percpu_counter_destroy(&btp->bt_io_count);
+out_destroy_lru:
+	list_lru_destroy(&btp->bt_lru);
+	return -ENOMEM;
 }
 
 struct xfs_buftarg *
 xfs_alloc_buftarg(
 	struct xfs_mount	*mp,
-	struct bdev_handle	*bdev_handle)
+	struct file		*bdev_file)
 {
-	xfs_buftarg_t		*btp;
+	struct xfs_buftarg	*btp;
 	const struct dax_holder_operations *ops = NULL;
 
 #if defined(CONFIG_FS_DAX) && defined(CONFIG_MEMORY_FAILURE)
 	ops = &xfs_dax_holder_operations;
 #endif
-	btp = kmem_zalloc(sizeof(*btp), KM_NOFS);
+	btp = kzalloc(sizeof(*btp), GFP_KERNEL | __GFP_NOFAIL);
 
 	btp->bt_mount = mp;
-	btp->bt_bdev_handle = bdev_handle;
-	btp->bt_dev = bdev_handle->bdev->bd_dev;
-	btp->bt_bdev = bdev_handle->bdev;
+	btp->bt_bdev_file = bdev_file;
+	btp->bt_bdev = file_bdev(bdev_file);
+	btp->bt_dev = btp->bt_bdev->bd_dev;
 	btp->bt_daxdev = fs_dax_get_by_bdev(btp->bt_bdev, &btp->bt_dax_part_off,
 					    mp, ops);
 
 	/*
-	 * Buffer IO error rate limiting. Limit it to no more than 10 messages
-	 * per 30 seconds so as to not spam logs too much on repeated errors.
+	 * When allocating the buftargs we have not yet read the super block and
+	 * thus don't know the file system sector size yet.
 	 */
-	ratelimit_state_init(&btp->bt_ioerror_rl, 30 * HZ,
-			     DEFAULT_RATELIMIT_BURST);
-
-	if (xfs_setsize_buftarg_early(btp))
+	if (xfs_setsize_buftarg(btp, bdev_logical_block_size(btp->bt_bdev)))
 		goto error_free;
-
-	if (list_lru_init(&btp->bt_lru))
+	if (xfs_init_buftarg(btp, bdev_logical_block_size(btp->bt_bdev),
+			mp->m_super->s_id))
 		goto error_free;
 
-	if (percpu_counter_init(&btp->bt_io_count, 0, GFP_KERNEL))
-		goto error_lru;
-
-	btp->bt_shrinker = shrinker_alloc(SHRINKER_NUMA_AWARE, "xfs-buf:%s",
-					  mp->m_super->s_id);
-	if (!btp->bt_shrinker)
-		goto error_pcpu;
-
-	btp->bt_shrinker->count_objects = xfs_buftarg_shrink_count;
-	btp->bt_shrinker->scan_objects = xfs_buftarg_shrink_scan;
-	btp->bt_shrinker->private_data = btp;
-
-	shrinker_register(btp->bt_shrinker);
-
 	return btp;
 
-error_pcpu:
-	percpu_counter_destroy(&btp->bt_io_count);
-error_lru:
-	list_lru_destroy(&btp->bt_lru);
 error_free:
-	kmem_free(btp);
+	kfree(btp);
 	return NULL;
 }
 
diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h
index b470de08a4..b158064450 100644
--- a/fs/xfs/xfs_buf.h
+++ b/fs/xfs/xfs_buf.h
@@ -83,6 +83,14 @@ typedef unsigned int xfs_buf_flags_t;
 #define XFS_BSTATE_DISPOSE	 (1 << 0)	/* buffer being discarded */
 #define XFS_BSTATE_IN_FLIGHT	 (1 << 1)	/* I/O in flight */
 
+struct xfs_buf_cache {
+	spinlock_t		bc_lock;
+	struct rhashtable	bc_hash;
+};
+
+int xfs_buf_cache_init(struct xfs_buf_cache *bch);
+void xfs_buf_cache_destroy(struct xfs_buf_cache *bch);
+
 /*
  * The xfs_buftarg contains 2 notions of "sector size" -
  *
@@ -96,11 +104,12 @@ typedef unsigned int xfs_buf_flags_t;
  * The latter is derived from the underlying device, and controls direct IO
  * alignment constraints.
  */
-typedef struct xfs_buftarg {
+struct xfs_buftarg {
 	dev_t			bt_dev;
-	struct bdev_handle	*bt_bdev_handle;
+	struct file		*bt_bdev_file;
 	struct block_device	*bt_bdev;
 	struct dax_device	*bt_daxdev;
+	struct file		*bt_file;
 	u64			bt_dax_part_off;
 	struct xfs_mount	*bt_mount;
 	unsigned int		bt_meta_sectorsize;
@@ -114,7 +123,10 @@ typedef struct xfs_buftarg {
 
 	struct percpu_counter	bt_io_count;
 	struct ratelimit_state	bt_ioerror_rl;
-} xfs_buftarg_t;
+
+	/* built-in cache, if we're not using the perag one */
+	struct xfs_buf_cache	bt_cache[];
+};
 
 #define XB_PAGES	2
 
@@ -366,7 +378,7 @@ xfs_buf_update_cksum(struct xfs_buf *bp, unsigned long cksum_offset)
  *	Handling of buftargs.
  */
 struct xfs_buftarg *xfs_alloc_buftarg(struct xfs_mount *mp,
-		struct bdev_handle *bdev_handle);
+		struct file *bdev_file);
 extern void xfs_free_buftarg(struct xfs_buftarg *);
 extern void xfs_buftarg_wait(struct xfs_buftarg *);
 extern void xfs_buftarg_drain(struct xfs_buftarg *);
@@ -379,4 +391,9 @@ int xfs_buf_reverify(struct xfs_buf *bp, const struct xfs_buf_ops *ops);
 bool xfs_verify_magic(struct xfs_buf *bp, __be32 dmagic);
 bool xfs_verify_magic16(struct xfs_buf *bp, __be16 dmagic);
 
+/* for xfs_buf_mem.c only: */
+int xfs_init_buftarg(struct xfs_buftarg *btp, size_t logical_sectorsize,
+		const char *descr);
+void xfs_destroy_buftarg(struct xfs_buftarg *btp);
+
 #endif	/* __XFS_BUF_H__ */
diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c
index 023d4e0385..4303184234 100644
--- a/fs/xfs/xfs_buf_item.c
+++ b/fs/xfs/xfs_buf_item.c
@@ -805,8 +805,8 @@ xfs_buf_item_get_format(
 		return;
 	}
 
-	bip->bli_formats = kmem_zalloc(count * sizeof(struct xfs_buf_log_format),
-				0);
+	bip->bli_formats = kzalloc(count * sizeof(struct xfs_buf_log_format),
+				GFP_KERNEL | __GFP_NOFAIL);
 }
 
 STATIC void
@@ -814,7 +814,7 @@ xfs_buf_item_free_format(
 	struct xfs_buf_log_item	*bip)
 {
 	if (bip->bli_formats != &bip->__bli_format) {
-		kmem_free(bip->bli_formats);
+		kfree(bip->bli_formats);
 		bip->bli_formats = NULL;
 	}
 }
@@ -1044,7 +1044,7 @@ xfs_buf_item_free(
 	struct xfs_buf_log_item	*bip)
 {
 	xfs_buf_item_free_format(bip);
-	kmem_free(bip->bli_item.li_lv_shadow);
+	kvfree(bip->bli_item.li_lv_shadow);
 	kmem_cache_free(xfs_buf_item_cache, bip);
 }
 
diff --git a/fs/xfs/xfs_buf_item_recover.c b/fs/xfs/xfs_buf_item_recover.c
index 43167f543a..09e893cf56 100644
--- a/fs/xfs/xfs_buf_item_recover.c
+++ b/fs/xfs/xfs_buf_item_recover.c
@@ -85,7 +85,7 @@ xlog_add_buffer_cancelled(
 		return false;
 	}
 
-	bcp = kmem_alloc(sizeof(struct xfs_buf_cancel), 0);
+	bcp = kmalloc(sizeof(struct xfs_buf_cancel), GFP_KERNEL | __GFP_NOFAIL);
 	bcp->bc_blkno = blkno;
 	bcp->bc_len = len;
 	bcp->bc_refcount = 1;
@@ -129,7 +129,7 @@ xlog_put_buffer_cancelled(
 
 	if (--bcp->bc_refcount == 0) {
 		list_del(&bcp->bc_list);
-		kmem_free(bcp);
+		kfree(bcp);
 	}
 	return true;
 }
@@ -1062,10 +1062,10 @@ xlog_free_buf_cancel_table(
 				&log->l_buf_cancel_table[i],
 				struct xfs_buf_cancel, bc_list))) {
 			list_del(&bc->bc_list);
-			kmem_free(bc);
+			kfree(bc);
 		}
 	}
 
-	kmem_free(log->l_buf_cancel_table);
+	kfree(log->l_buf_cancel_table);
 	log->l_buf_cancel_table = NULL;
 }
diff --git a/fs/xfs/xfs_buf_mem.c b/fs/xfs/xfs_buf_mem.c
new file mode 100644
index 0000000000..9bb2d24de7
--- /dev/null
+++ b/fs/xfs/xfs_buf_mem.c
@@ -0,0 +1,270 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2023-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_buf.h"
+#include "xfs_buf_mem.h"
+#include "xfs_trace.h"
+#include <linux/shmem_fs.h>
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_buf_item.h"
+#include "xfs_error.h"
+
+/*
+ * Buffer Cache for In-Memory Files
+ * ================================
+ *
+ * Online fsck wants to create ephemeral ordered recordsets.  The existing
+ * btree infrastructure can do this, but we need the buffer cache to target
+ * memory instead of block devices.
+ *
+ * When CONFIG_TMPFS=y, shmemfs is enough of a filesystem to meet those
+ * requirements.  Therefore, the xmbuf mechanism uses an unlinked shmem file to
+ * store our staging data.  This file is not installed in the file descriptor
+ * table so that user programs cannot access the data, which means that the
+ * xmbuf must be freed with xmbuf_destroy.
+ *
+ * xmbufs assume that the caller will handle all required concurrency
+ * management; standard vfs locks (freezer and inode) are not taken.  Reads
+ * and writes are satisfied directly from the page cache.
+ *
+ * The only supported block size is PAGE_SIZE, and we cannot use highmem.
+ */
+
+/*
+ * shmem files used to back an in-memory buffer cache must not be exposed to
+ * userspace.  Upper layers must coordinate access to the one handle returned
+ * by the constructor, so establish a separate lock class for xmbufs to avoid
+ * confusing lockdep.
+ */
+static struct lock_class_key xmbuf_i_mutex_key;
+
+/*
+ * Allocate a buffer cache target for a memory-backed file and set up the
+ * buffer target.
+ */
+int
+xmbuf_alloc(
+	struct xfs_mount	*mp,
+	const char		*descr,
+	struct xfs_buftarg	**btpp)
+{
+	struct file		*file;
+	struct inode		*inode;
+	struct xfs_buftarg	*btp;
+	int			error;
+
+	btp = kzalloc(struct_size(btp, bt_cache, 1), GFP_KERNEL);
+	if (!btp)
+		return -ENOMEM;
+
+	file = shmem_kernel_file_setup(descr, 0, 0);
+	if (IS_ERR(file)) {
+		error = PTR_ERR(file);
+		goto out_free_btp;
+	}
+	inode = file_inode(file);
+
+	/* private file, private locking */
+	lockdep_set_class(&inode->i_rwsem, &xmbuf_i_mutex_key);
+
+	/*
+	 * We don't want to bother with kmapping data during repair, so don't
+	 * allow highmem pages to back this mapping.
+	 */
+	mapping_set_gfp_mask(inode->i_mapping, GFP_KERNEL);
+
+	/* ensure all writes are below EOF to avoid pagecache zeroing */
+	i_size_write(inode, inode->i_sb->s_maxbytes);
+
+	error = xfs_buf_cache_init(btp->bt_cache);
+	if (error)
+		goto out_file;
+
+	/* Initialize buffer target */
+	btp->bt_mount = mp;
+	btp->bt_dev = (dev_t)-1U;
+	btp->bt_bdev = NULL; /* in-memory buftargs have no bdev */
+	btp->bt_file = file;
+	btp->bt_meta_sectorsize = XMBUF_BLOCKSIZE;
+	btp->bt_meta_sectormask = XMBUF_BLOCKSIZE - 1;
+
+	error = xfs_init_buftarg(btp, XMBUF_BLOCKSIZE, descr);
+	if (error)
+		goto out_bcache;
+
+	trace_xmbuf_create(btp);
+
+	*btpp = btp;
+	return 0;
+
+out_bcache:
+	xfs_buf_cache_destroy(btp->bt_cache);
+out_file:
+	fput(file);
+out_free_btp:
+	kfree(btp);
+	return error;
+}
+
+/* Free a buffer cache target for a memory-backed buffer cache. */
+void
+xmbuf_free(
+	struct xfs_buftarg	*btp)
+{
+	ASSERT(xfs_buftarg_is_mem(btp));
+	ASSERT(percpu_counter_sum(&btp->bt_io_count) == 0);
+
+	trace_xmbuf_free(btp);
+
+	xfs_destroy_buftarg(btp);
+	xfs_buf_cache_destroy(btp->bt_cache);
+	fput(btp->bt_file);
+	kfree(btp);
+}
+
+/* Directly map a shmem page into the buffer cache. */
+int
+xmbuf_map_page(
+	struct xfs_buf		*bp)
+{
+	struct inode		*inode = file_inode(bp->b_target->bt_file);
+	struct folio		*folio = NULL;
+	struct page		*page;
+	loff_t                  pos = BBTOB(xfs_buf_daddr(bp));
+	int			error;
+
+	ASSERT(xfs_buftarg_is_mem(bp->b_target));
+
+	if (bp->b_map_count != 1)
+		return -ENOMEM;
+	if (BBTOB(bp->b_length) != XMBUF_BLOCKSIZE)
+		return -ENOMEM;
+	if (offset_in_page(pos) != 0) {
+		ASSERT(offset_in_page(pos));
+		return -ENOMEM;
+	}
+
+	error = shmem_get_folio(inode, pos >> PAGE_SHIFT, &folio, SGP_CACHE);
+	if (error)
+		return error;
+
+	if (filemap_check_wb_err(inode->i_mapping, 0)) {
+		folio_unlock(folio);
+		folio_put(folio);
+		return -EIO;
+	}
+
+	page = folio_file_page(folio, pos >> PAGE_SHIFT);
+
+	/*
+	 * Mark the page dirty so that it won't be reclaimed once we drop the
+	 * (potentially last) reference in xmbuf_unmap_page.
+	 */
+	set_page_dirty(page);
+	unlock_page(page);
+
+	bp->b_addr = page_address(page);
+	bp->b_pages = bp->b_page_array;
+	bp->b_pages[0] = page;
+	bp->b_page_count = 1;
+	return 0;
+}
+
+/* Unmap a shmem page that was mapped into the buffer cache. */
+void
+xmbuf_unmap_page(
+	struct xfs_buf		*bp)
+{
+	struct page		*page = bp->b_pages[0];
+
+	ASSERT(xfs_buftarg_is_mem(bp->b_target));
+
+	put_page(page);
+
+	bp->b_addr = NULL;
+	bp->b_pages[0] = NULL;
+	bp->b_pages = NULL;
+	bp->b_page_count = 0;
+}
+
+/* Is this a valid daddr within the buftarg? */
+bool
+xmbuf_verify_daddr(
+	struct xfs_buftarg	*btp,
+	xfs_daddr_t		daddr)
+{
+	struct inode		*inode = file_inode(btp->bt_file);
+
+	ASSERT(xfs_buftarg_is_mem(btp));
+
+	return daddr < (inode->i_sb->s_maxbytes >> BBSHIFT);
+}
+
+/* Discard the page backing this buffer. */
+static void
+xmbuf_stale(
+	struct xfs_buf		*bp)
+{
+	struct inode		*inode = file_inode(bp->b_target->bt_file);
+	loff_t			pos;
+
+	ASSERT(xfs_buftarg_is_mem(bp->b_target));
+
+	pos = BBTOB(xfs_buf_daddr(bp));
+	shmem_truncate_range(inode, pos, pos + BBTOB(bp->b_length) - 1);
+}
+
+/*
+ * Finalize a buffer -- discard the backing page if it's stale, or run the
+ * write verifier to detect problems.
+ */
+int
+xmbuf_finalize(
+	struct xfs_buf		*bp)
+{
+	xfs_failaddr_t		fa;
+	int			error = 0;
+
+	if (bp->b_flags & XBF_STALE) {
+		xmbuf_stale(bp);
+		return 0;
+	}
+
+	/*
+	 * Although this btree is ephemeral, validate the buffer structure so
+	 * that we can detect memory corruption errors and software bugs.
+	 */
+	fa = bp->b_ops->verify_struct(bp);
+	if (fa) {
+		error = -EFSCORRUPTED;
+		xfs_verifier_error(bp, error, fa);
+	}
+
+	return error;
+}
+
+/*
+ * Detach this xmbuf buffer from the transaction by any means necessary.
+ * All buffers are direct-mapped, so they do not need bwrite.
+ */
+void
+xmbuf_trans_bdetach(
+	struct xfs_trans	*tp,
+	struct xfs_buf		*bp)
+{
+	struct xfs_buf_log_item	*bli = bp->b_log_item;
+
+	ASSERT(bli != NULL);
+
+	bli->bli_flags &= ~(XFS_BLI_DIRTY | XFS_BLI_ORDERED |
+			    XFS_BLI_LOGGED | XFS_BLI_STALE);
+	clear_bit(XFS_LI_DIRTY, &bli->bli_item.li_flags);
+
+	while (bp->b_log_item != NULL)
+		xfs_trans_bdetach(tp, bp);
+}
diff --git a/fs/xfs/xfs_buf_mem.h b/fs/xfs/xfs_buf_mem.h
new file mode 100644
index 0000000000..eed4a7b632
--- /dev/null
+++ b/fs/xfs/xfs_buf_mem.h
@@ -0,0 +1,34 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2023-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_BUF_MEM_H__
+#define __XFS_BUF_MEM_H__
+
+#define XMBUF_BLOCKSIZE			(PAGE_SIZE)
+#define XMBUF_BLOCKSHIFT		(PAGE_SHIFT)
+
+#ifdef CONFIG_XFS_MEMORY_BUFS
+static inline bool xfs_buftarg_is_mem(const struct xfs_buftarg *btp)
+{
+	return btp->bt_bdev == NULL;
+}
+
+int xmbuf_alloc(struct xfs_mount *mp, const char *descr,
+		struct xfs_buftarg **btpp);
+void xmbuf_free(struct xfs_buftarg *btp);
+
+int xmbuf_map_page(struct xfs_buf *bp);
+void xmbuf_unmap_page(struct xfs_buf *bp);
+bool xmbuf_verify_daddr(struct xfs_buftarg *btp, xfs_daddr_t daddr);
+void xmbuf_trans_bdetach(struct xfs_trans *tp, struct xfs_buf *bp);
+int xmbuf_finalize(struct xfs_buf *bp);
+#else
+# define xfs_buftarg_is_mem(...)	(false)
+# define xmbuf_map_page(...)		(-ENOMEM)
+# define xmbuf_unmap_page(...)		((void)0)
+# define xmbuf_verify_daddr(...)	(false)
+#endif /* CONFIG_XFS_MEMORY_BUFS */
+
+#endif /* __XFS_BUF_MEM_H__ */
diff --git a/fs/xfs/xfs_dir2_readdir.c b/fs/xfs/xfs_dir2_readdir.c
index cc6dc56f45..06ac5a7de6 100644
--- a/fs/xfs/xfs_dir2_readdir.c
+++ b/fs/xfs/xfs_dir2_readdir.c
@@ -118,8 +118,10 @@ xfs_dir2_sf_getdents(
 		ctx->pos = off & 0x7fffffff;
 		if (XFS_IS_CORRUPT(dp->i_mount,
 				   !xfs_dir2_namecheck(sfep->name,
-						       sfep->namelen)))
+						       sfep->namelen))) {
+			xfs_dirattr_mark_sick(dp, XFS_DATA_FORK);
 			return -EFSCORRUPTED;
+		}
 		if (!dir_emit(ctx, (char *)sfep->name, sfep->namelen, ino,
 			    xfs_dir3_get_dtype(mp, filetype)))
 			return 0;
@@ -155,7 +157,7 @@ xfs_dir2_block_getdents(
 	if (xfs_dir2_dataptr_to_db(geo, ctx->pos) > geo->datablk)
 		return 0;
 
-	error = xfs_dir3_block_read(args->trans, dp, &bp);
+	error = xfs_dir3_block_read(args->trans, dp, args->owner, &bp);
 	if (error)
 		return error;
 
@@ -211,6 +213,7 @@ xfs_dir2_block_getdents(
 		if (XFS_IS_CORRUPT(dp->i_mount,
 				   !xfs_dir2_namecheck(dep->name,
 						       dep->namelen))) {
+			xfs_dirattr_mark_sick(dp, XFS_DATA_FORK);
 			error = -EFSCORRUPTED;
 			goto out_rele;
 		}
@@ -279,7 +282,8 @@ xfs_dir2_leaf_readbuf(
 	new_off = xfs_dir2_da_to_byte(geo, map.br_startoff);
 	if (new_off > *cur_off)
 		*cur_off = new_off;
-	error = xfs_dir3_data_read(args->trans, dp, map.br_startoff, 0, &bp);
+	error = xfs_dir3_data_read(args->trans, dp, args->owner,
+			map.br_startoff, 0, &bp);
 	if (error)
 		goto out;
 
@@ -465,6 +469,7 @@ xfs_dir2_leaf_getdents(
 		if (XFS_IS_CORRUPT(dp->i_mount,
 				   !xfs_dir2_namecheck(dep->name,
 						       dep->namelen))) {
+			xfs_dirattr_mark_sick(dp, XFS_DATA_FORK);
 			error = -EFSCORRUPTED;
 			break;
 		}
@@ -511,7 +516,6 @@ xfs_readdir(
 {
 	struct xfs_da_args	args = { NULL };
 	unsigned int		lock_mode;
-	bool			isblock;
 	int			error;
 
 	trace_xfs_readdir(dp);
@@ -522,29 +526,30 @@ xfs_readdir(
 		return -EIO;
 
 	ASSERT(S_ISDIR(VFS_I(dp)->i_mode));
-	ASSERT(xfs_isilocked(dp, XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL));
+	xfs_assert_ilocked(dp, XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL);
 	XFS_STATS_INC(dp->i_mount, xs_dir_getdents);
 
 	args.dp = dp;
 	args.geo = dp->i_mount->m_dir_geo;
 	args.trans = tp;
+	args.owner = dp->i_ino;
 
 	if (dp->i_df.if_format == XFS_DINODE_FMT_LOCAL)
 		return xfs_dir2_sf_getdents(&args, ctx);
 
 	lock_mode = xfs_ilock_data_map_shared(dp);
-	error = xfs_dir2_isblock(&args, &isblock);
-	if (error)
-		goto out_unlock;
-
-	if (isblock) {
+	switch (xfs_dir2_format(&args, &error)) {
+	case XFS_DIR2_FMT_BLOCK:
 		error = xfs_dir2_block_getdents(&args, ctx, &lock_mode);
-		goto out_unlock;
+		break;
+	case XFS_DIR2_FMT_LEAF:
+	case XFS_DIR2_FMT_NODE:
+		error = xfs_dir2_leaf_getdents(&args, ctx, bufsize, &lock_mode);
+		break;
+	default:
+		break;
 	}
 
-	error = xfs_dir2_leaf_getdents(&args, ctx, bufsize, &lock_mode);
-
-out_unlock:
 	if (lock_mode)
 		xfs_iunlock(dp, lock_mode);
 	return error;
diff --git a/fs/xfs/xfs_discard.c b/fs/xfs/xfs_discard.c
index d5787991bb..25fe3b932b 100644
--- a/fs/xfs/xfs_discard.c
+++ b/fs/xfs/xfs_discard.c
@@ -8,6 +8,7 @@
 #include "xfs_format.h"
 #include "xfs_log_format.h"
 #include "xfs_trans_resv.h"
+#include "xfs_trans.h"
 #include "xfs_mount.h"
 #include "xfs_btree.h"
 #include "xfs_alloc_btree.h"
@@ -18,6 +19,7 @@
 #include "xfs_trace.h"
 #include "xfs_log.h"
 #include "xfs_ag.h"
+#include "xfs_health.h"
 
 /*
  * Notes on an efficient, low latency fstrim algorithm
@@ -79,7 +81,7 @@ xfs_discard_endio_work(
 		container_of(work, struct xfs_busy_extents, endio_work);
 
 	xfs_extent_busy_clear(extents->mount, &extents->extent_list, false);
-	kmem_free(extents->owner);
+	kfree(extents->owner);
 }
 
 /*
@@ -120,7 +122,7 @@ xfs_discard_extents(
 		error = __blkdev_issue_discard(mp->m_ddev_targp->bt_bdev,
 				XFS_AGB_TO_DADDR(mp, busyp->agno, busyp->bno),
 				XFS_FSB_TO_BB(mp, busyp->length),
-				GFP_NOFS, &bio);
+				GFP_KERNEL, &bio);
 		if (error && error != -EOPNOTSUPP) {
 			xfs_info(mp,
 	 "discard failed for extent [0x%llx,%u], error %d",
@@ -143,18 +145,23 @@ xfs_discard_extents(
 	return error;
 }
 
+struct xfs_trim_cur {
+	xfs_agblock_t	start;
+	xfs_extlen_t	count;
+	xfs_agblock_t	end;
+	xfs_extlen_t	minlen;
+	bool		by_bno;
+};
 
 static int
 xfs_trim_gather_extents(
 	struct xfs_perag	*pag,
-	xfs_daddr_t		start,
-	xfs_daddr_t		end,
-	xfs_daddr_t		minlen,
-	struct xfs_alloc_rec_incore *tcur,
+	struct xfs_trim_cur	*tcur,
 	struct xfs_busy_extents	*extents,
 	uint64_t		*blocks_trimmed)
 {
 	struct xfs_mount	*mp = pag->pag_mount;
+	struct xfs_trans	*tp;
 	struct xfs_btree_cur	*cur;
 	struct xfs_buf		*agbp;
 	int			error;
@@ -168,25 +175,34 @@ xfs_trim_gather_extents(
 	 */
 	xfs_log_force(mp, XFS_LOG_SYNC);
 
-	error = xfs_alloc_read_agf(pag, NULL, 0, &agbp);
+	error = xfs_trans_alloc_empty(mp, &tp);
 	if (error)
 		return error;
 
-	cur = xfs_allocbt_init_cursor(mp, NULL, agbp, pag, XFS_BTNUM_CNT);
-
-	/*
-	 * Look up the extent length requested in the AGF and start with it.
-	 */
-	if (tcur->ar_startblock == NULLAGBLOCK)
-		error = xfs_alloc_lookup_ge(cur, 0, tcur->ar_blockcount, &i);
-	else
-		error = xfs_alloc_lookup_le(cur, tcur->ar_startblock,
-				tcur->ar_blockcount, &i);
+	error = xfs_alloc_read_agf(pag, tp, 0, &agbp);
+	if (error)
+		goto out_trans_cancel;
+
+	if (tcur->by_bno) {
+		/* sub-AG discard request always starts at tcur->start */
+		cur = xfs_bnobt_init_cursor(mp, tp, agbp, pag);
+		error = xfs_alloc_lookup_le(cur, tcur->start, 0, &i);
+		if (!error && !i)
+			error = xfs_alloc_lookup_ge(cur, tcur->start, 0, &i);
+	} else if (tcur->start == 0) {
+		/* first time through a by-len starts with max length */
+		cur = xfs_cntbt_init_cursor(mp, tp, agbp, pag);
+		error = xfs_alloc_lookup_ge(cur, 0, tcur->count, &i);
+	} else {
+		/* nth time through a by-len starts where we left off */
+		cur = xfs_cntbt_init_cursor(mp, tp, agbp, pag);
+		error = xfs_alloc_lookup_le(cur, tcur->start, tcur->count, &i);
+	}
 	if (error)
 		goto out_del_cursor;
 	if (i == 0) {
 		/* nothing of that length left in the AG, we are done */
-		tcur->ar_blockcount = 0;
+		tcur->count = 0;
 		goto out_del_cursor;
 	}
 
@@ -197,13 +213,12 @@ xfs_trim_gather_extents(
 	while (i) {
 		xfs_agblock_t	fbno;
 		xfs_extlen_t	flen;
-		xfs_daddr_t	dbno;
-		xfs_extlen_t	dlen;
 
 		error = xfs_alloc_get_rec(cur, &fbno, &flen, &i);
 		if (error)
 			break;
 		if (XFS_IS_CORRUPT(mp, i != 1)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			break;
 		}
@@ -213,38 +228,46 @@ xfs_trim_gather_extents(
 			 * Update the cursor to point at this extent so we
 			 * restart the next batch from this extent.
 			 */
-			tcur->ar_startblock = fbno;
-			tcur->ar_blockcount = flen;
-			break;
-		}
-
-		/*
-		 * use daddr format for all range/len calculations as that is
-		 * the format the range/len variables are supplied in by
-		 * userspace.
-		 */
-		dbno = XFS_AGB_TO_DADDR(mp, pag->pag_agno, fbno);
-		dlen = XFS_FSB_TO_BB(mp, flen);
-
-		/*
-		 * Too small?  Give up.
-		 */
-		if (dlen < minlen) {
-			trace_xfs_discard_toosmall(mp, pag->pag_agno, fbno, flen);
-			tcur->ar_blockcount = 0;
+			tcur->start = fbno;
+			tcur->count = flen;
 			break;
 		}
 
 		/*
 		 * If the extent is entirely outside of the range we are
-		 * supposed to discard skip it.  Do not bother to trim
-		 * down partially overlapping ranges for now.
+		 * supposed to skip it.  Do not bother to trim down partially
+		 * overlapping ranges for now.
 		 */
-		if (dbno + dlen < start || dbno > end) {
+		if (fbno + flen < tcur->start) {
+			trace_xfs_discard_exclude(mp, pag->pag_agno, fbno, flen);
+			goto next_extent;
+		}
+		if (fbno > tcur->end) {
 			trace_xfs_discard_exclude(mp, pag->pag_agno, fbno, flen);
+			if (tcur->by_bno) {
+				tcur->count = 0;
+				break;
+			}
 			goto next_extent;
 		}
 
+		/* Trim the extent returned to the range we want. */
+		if (fbno < tcur->start) {
+			flen -= tcur->start - fbno;
+			fbno = tcur->start;
+		}
+		if (fbno + flen > tcur->end + 1)
+			flen = tcur->end - fbno + 1;
+
+		/* Too small?  Give up. */
+		if (flen < tcur->minlen) {
+			trace_xfs_discard_toosmall(mp, pag->pag_agno, fbno, flen);
+			if (tcur->by_bno)
+				goto next_extent;
+			tcur->count = 0;
+			break;
+		}
+
 		/*
 		 * If any blocks in the range are still busy, skip the
 		 * discard and try again the next time.
@@ -258,7 +281,10 @@ xfs_trim_gather_extents(
 				&extents->extent_list);
 		*blocks_trimmed += flen;
 next_extent:
-		error = xfs_btree_decrement(cur, 0, &i);
+		if (tcur->by_bno)
+			error = xfs_btree_increment(cur, 0, &i);
+		else
+			error = xfs_btree_decrement(cur, 0, &i);
 		if (error)
 			break;
 
@@ -268,7 +294,7 @@ next_extent:
 		 * is no more extents to search.
 		 */
 		if (i == 0)
-			tcur->ar_blockcount = 0;
+			tcur->count = 0;
 	}
 
 	/*
@@ -279,7 +305,8 @@ next_extent:
 		xfs_extent_busy_clear(mp, &extents->extent_list, false);
 out_del_cursor:
 	xfs_btree_del_cursor(cur, error);
-	xfs_buf_relse(agbp);
+out_trans_cancel:
+	xfs_trans_cancel(tp);
 	return error;
 }
 
@@ -297,17 +324,22 @@ xfs_trim_should_stop(void)
 static int
 xfs_trim_extents(
 	struct xfs_perag	*pag,
-	xfs_daddr_t		start,
-	xfs_daddr_t		end,
-	xfs_daddr_t		minlen,
+	xfs_agblock_t		start,
+	xfs_agblock_t		end,
+	xfs_extlen_t		minlen,
 	uint64_t		*blocks_trimmed)
 {
-	struct xfs_alloc_rec_incore tcur = {
-		.ar_blockcount = pag->pagf_longest,
-		.ar_startblock = NULLAGBLOCK,
+	struct xfs_trim_cur	tcur = {
+		.start		= start,
+		.count		= pag->pagf_longest,
+		.end		= end,
+		.minlen		= minlen,
 	};
 	int			error = 0;
 
+	if (start != 0 || end != pag->block_count)
+		tcur.by_bno = true;
+
 	do {
 		struct xfs_busy_extents	*extents;
 
@@ -321,8 +353,8 @@ xfs_trim_extents(
 		extents->owner = extents;
 		INIT_LIST_HEAD(&extents->extent_list);
 
-		error = xfs_trim_gather_extents(pag, start, end, minlen,
-				&tcur, extents, blocks_trimmed);
+		error = xfs_trim_gather_extents(pag, &tcur, extents,
+				blocks_trimmed);
 		if (error) {
 			kfree(extents);
 			break;
@@ -345,7 +377,7 @@ xfs_trim_extents(
 		if (xfs_trim_should_stop())
 			break;
 
-	} while (tcur.ar_blockcount != 0);
+	} while (tcur.count != 0);
 
 	return error;
 
@@ -369,8 +401,10 @@ xfs_ioc_trim(
 	unsigned int		granularity =
 		bdev_discard_granularity(mp->m_ddev_targp->bt_bdev);
 	struct fstrim_range	range;
-	xfs_daddr_t		start, end, minlen;
-	xfs_agnumber_t		agno;
+	xfs_daddr_t		start, end;
+	xfs_extlen_t		minlen;
+	xfs_agnumber_t		start_agno, end_agno;
+	xfs_agblock_t		start_agbno, end_agbno;
 	uint64_t		blocks_trimmed = 0;
 	int			error, last_error = 0;
 
@@ -390,7 +424,8 @@ xfs_ioc_trim(
 		return -EFAULT;
 
 	range.minlen = max_t(u64, granularity, range.minlen);
-	minlen = BTOBB(range.minlen);
+	minlen = XFS_B_TO_FSB(mp, range.minlen);
+
 	/*
 	 * Truncating down the len isn't actually quite correct, but using
 	 * BBTOB would mean we trivially get overflows for values
@@ -404,15 +439,21 @@ xfs_ioc_trim(
 		return -EINVAL;
 
 	start = BTOBB(range.start);
-	end = start + BTOBBT(range.len) - 1;
+	end = min_t(xfs_daddr_t, start + BTOBBT(range.len),
+		    XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks)) - 1;
+
+	start_agno = xfs_daddr_to_agno(mp, start);
+	start_agbno = xfs_daddr_to_agbno(mp, start);
+	end_agno = xfs_daddr_to_agno(mp, end);
+	end_agbno = xfs_daddr_to_agbno(mp, end);
 
-	if (end > XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks) - 1)
-		end = XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks) - 1;
+	for_each_perag_range(mp, start_agno, end_agno, pag) {
+		xfs_agblock_t	agend = pag->block_count;
 
-	agno = xfs_daddr_to_agno(mp, start);
-	for_each_perag_range(mp, agno, xfs_daddr_to_agno(mp, end), pag) {
-		error = xfs_trim_extents(pag, start, end, minlen,
-					  &blocks_trimmed);
+		if (start_agno == end_agno)
+			agend = end_agbno;
+		error = xfs_trim_extents(pag, start_agbno, agend, minlen,
+				&blocks_trimmed);
 		if (error)
 			last_error = error;
 
@@ -420,6 +461,7 @@ xfs_ioc_trim(
 			xfs_perag_rele(pag);
 			break;
 		}
+		start_agbno = 0;
 	}
 
 	if (last_error)
diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c
index b4f20d9c8f..c1b211c260 100644
--- a/fs/xfs/xfs_dquot.c
+++ b/fs/xfs/xfs_dquot.c
@@ -24,6 +24,7 @@
 #include "xfs_log.h"
 #include "xfs_bmap_btree.h"
 #include "xfs_error.h"
+#include "xfs_health.h"
 
 /*
  * Lock order:
@@ -44,6 +45,29 @@ static struct kmem_cache	*xfs_dquot_cache;
 static struct lock_class_key xfs_dquot_group_class;
 static struct lock_class_key xfs_dquot_project_class;
 
+/* Record observations of quota corruption with the health tracking system. */
+static void
+xfs_dquot_mark_sick(
+	struct xfs_dquot	*dqp)
+{
+	struct xfs_mount	*mp = dqp->q_mount;
+
+	switch (dqp->q_type) {
+	case XFS_DQTYPE_USER:
+		xfs_fs_mark_sick(mp, XFS_SICK_FS_UQUOTA);
+		break;
+	case XFS_DQTYPE_GROUP:
+		xfs_fs_mark_sick(mp, XFS_SICK_FS_GQUOTA);
+		break;
+	case XFS_DQTYPE_PROJ:
+		xfs_fs_mark_sick(mp, XFS_SICK_FS_PQUOTA);
+		break;
+	default:
+		ASSERT(0);
+		break;
+	}
+}
+
 /*
  * This is called to free all the memory associated with a dquot
  */
@@ -53,7 +77,7 @@ xfs_qm_dqdestroy(
 {
 	ASSERT(list_empty(&dqp->q_lru));
 
-	kmem_free(dqp->q_logitem.qli_item.li_lv_shadow);
+	kvfree(dqp->q_logitem.qli_item.li_lv_shadow);
 	mutex_destroy(&dqp->q_qlock);
 
 	XFS_STATS_DEC(dqp->q_mount, xs_qm_dquot);
@@ -317,11 +341,8 @@ xfs_dquot_disk_alloc(
 		goto err_cancel;
 	}
 
-	error = xfs_iext_count_may_overflow(quotip, XFS_DATA_FORK,
+	error = xfs_iext_count_extend(tp, quotip, XFS_DATA_FORK,
 			XFS_IEXT_ADD_NOSPLIT_CNT);
-	if (error == -EFBIG)
-		error = xfs_iext_count_upgrade(tp, quotip,
-				XFS_IEXT_ADD_NOSPLIT_CNT);
 	if (error)
 		goto err_cancel;
 
@@ -333,7 +354,6 @@ xfs_dquot_disk_alloc(
 		goto err_cancel;
 
 	ASSERT(map.br_blockcount == XFS_DQUOT_CLUSTER_SIZE_FSB);
-	ASSERT(nmaps == 1);
 	ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
 	       (map.br_startblock != HOLESTARTBLOCK));
 
@@ -451,6 +471,8 @@ xfs_dquot_disk_read(
 	error = xfs_trans_read_buf(mp, NULL, mp->m_ddev_targp, dqp->q_blkno,
 			mp->m_quotainfo->qi_dqchunklen, 0, &bp,
 			&xfs_dquot_buf_ops);
+	if (xfs_metadata_is_sick(error))
+		xfs_dquot_mark_sick(dqp);
 	if (error) {
 		ASSERT(bp == NULL);
 		return error;
@@ -574,6 +596,7 @@ xfs_dquot_from_disk(
 			  "Metadata corruption detected at %pS, quota %u",
 			  __this_address, dqp->q_id);
 		xfs_alert(bp->b_mount, "Unmount and run xfs_repair");
+		xfs_dquot_mark_sick(dqp);
 		return -EFSCORRUPTED;
 	}
 
@@ -784,6 +807,12 @@ restart:
  * caller should throw away the dquot and start over.  Otherwise, the dquot
  * is returned locked (and held by the cache) as if there had been a cache
  * hit.
+ *
+ * The insert needs to be done under memalloc_nofs context because the radix
+ * tree can do memory allocation during insert. The qi->qi_tree_lock is taken in
+ * memory reclaim when freeing unused dquots, so we cannot have the radix tree
+ * node allocation recursing into filesystem reclaim whilst we hold the
+ * qi_tree_lock.
  */
 static int
 xfs_qm_dqget_cache_insert(
@@ -793,25 +822,27 @@ xfs_qm_dqget_cache_insert(
 	xfs_dqid_t		id,
 	struct xfs_dquot	*dqp)
 {
+	unsigned int		nofs_flags;
 	int			error;
 
+	nofs_flags = memalloc_nofs_save();
 	mutex_lock(&qi->qi_tree_lock);
 	error = radix_tree_insert(tree, id, dqp);
 	if (unlikely(error)) {
 		/* Duplicate found!  Caller must try again. */
-		mutex_unlock(&qi->qi_tree_lock);
 		trace_xfs_dqget_dup(dqp);
-		return error;
+		goto out_unlock;
 	}
 
 	/* Return a locked dquot to the caller, with a reference taken. */
 	xfs_dqlock(dqp);
 	dqp->q_nrefs = 1;
-
 	qi->qi_dquots++;
-	mutex_unlock(&qi->qi_tree_lock);
 
-	return 0;
+out_unlock:
+	mutex_unlock(&qi->qi_tree_lock);
+	memalloc_nofs_restore(nofs_flags);
+	return error;
 }
 
 /* Check our input parameters. */
@@ -950,7 +981,7 @@ xfs_qm_dqget_inode(
 	if (error)
 		return error;
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 	ASSERT(xfs_inode_dquot(ip, type) == NULL);
 
 	id = xfs_qm_id_for_quotatype(ip, type);
@@ -1007,7 +1038,7 @@ restart:
 	}
 
 dqret:
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 	trace_xfs_dqget_miss(dqp);
 	*O_dqpp = dqp;
 	return 0;
@@ -1238,6 +1269,8 @@ xfs_qm_dqflush(
 				   &bp, &xfs_dquot_buf_ops);
 	if (error == -EAGAIN)
 		goto out_unlock;
+	if (xfs_metadata_is_sick(error))
+		xfs_dquot_mark_sick(dqp);
 	if (error)
 		goto out_abort;
 
@@ -1246,6 +1279,7 @@ xfs_qm_dqflush(
 		xfs_alert(mp, "corrupt dquot ID 0x%x in memory at %pS",
 				dqp->q_id, fa);
 		xfs_buf_relse(bp);
+		xfs_dquot_mark_sick(dqp);
 		error = -EFSCORRUPTED;
 		goto out_abort;
 	}
@@ -1333,6 +1367,47 @@ xfs_dqlock2(
 	}
 }
 
+static int
+xfs_dqtrx_cmp(
+	const void		*a,
+	const void		*b)
+{
+	const struct xfs_dqtrx	*qa = a;
+	const struct xfs_dqtrx	*qb = b;
+
+	if (qa->qt_dquot->q_id > qb->qt_dquot->q_id)
+		return 1;
+	if (qa->qt_dquot->q_id < qb->qt_dquot->q_id)
+		return -1;
+	return 0;
+}
+
+void
+xfs_dqlockn(
+	struct xfs_dqtrx	*q)
+{
+	unsigned int		i;
+
+	BUILD_BUG_ON(XFS_QM_TRANS_MAXDQS > MAX_LOCKDEP_SUBCLASSES);
+
+	/* Sort in order of dquot id, do not allow duplicates */
+	for (i = 0; i < XFS_QM_TRANS_MAXDQS && q[i].qt_dquot != NULL; i++) {
+		unsigned int	j;
+
+		for (j = 0; j < i; j++)
+			ASSERT(q[i].qt_dquot != q[j].qt_dquot);
+	}
+	if (i == 0)
+		return;
+
+	sort(q, i, sizeof(struct xfs_dqtrx), xfs_dqtrx_cmp, NULL);
+
+	mutex_lock(&q[0].qt_dquot->q_qlock);
+	for (i = 1; i < XFS_QM_TRANS_MAXDQS && q[i].qt_dquot != NULL; i++)
+		mutex_lock_nested(&q[i].qt_dquot->q_qlock,
+				XFS_QLOCK_NESTED + i - 1);
+}
+
 int __init
 xfs_qm_init(void)
 {
diff --git a/fs/xfs/xfs_dquot.h b/fs/xfs/xfs_dquot.h
index 956272d9b3..677bb2dc9a 100644
--- a/fs/xfs/xfs_dquot.h
+++ b/fs/xfs/xfs_dquot.h
@@ -223,6 +223,7 @@ int		xfs_qm_dqget_uncached(struct xfs_mount *mp,
 void		xfs_qm_dqput(struct xfs_dquot *dqp);
 
 void		xfs_dqlock2(struct xfs_dquot *, struct xfs_dquot *);
+void		xfs_dqlockn(struct xfs_dqtrx *q);
 
 void		xfs_dquot_set_prealloc_limits(struct xfs_dquot *);
 
diff --git a/fs/xfs/xfs_error.c b/fs/xfs/xfs_error.c
index b2cbbba3e1..78cdc5064a 100644
--- a/fs/xfs/xfs_error.c
+++ b/fs/xfs/xfs_error.c
@@ -62,6 +62,7 @@ static unsigned int xfs_errortag_random_default[] = {
 	XFS_RANDOM_ATTR_LEAF_TO_NODE,
 	XFS_RANDOM_WB_DELAY_MS,
 	XFS_RANDOM_WRITE_DELAY_MS,
+	XFS_RANDOM_EXCHMAPS_FINISH_ONE,
 };
 
 struct xfs_errortag_attr {
@@ -179,6 +180,7 @@ XFS_ERRORTAG_ATTR_RW(da_leaf_split,	XFS_ERRTAG_DA_LEAF_SPLIT);
 XFS_ERRORTAG_ATTR_RW(attr_leaf_to_node,	XFS_ERRTAG_ATTR_LEAF_TO_NODE);
 XFS_ERRORTAG_ATTR_RW(wb_delay_ms,	XFS_ERRTAG_WB_DELAY_MS);
 XFS_ERRORTAG_ATTR_RW(write_delay_ms,	XFS_ERRTAG_WRITE_DELAY_MS);
+XFS_ERRORTAG_ATTR_RW(exchmaps_finish_one, XFS_ERRTAG_EXCHMAPS_FINISH_ONE);
 
 static struct attribute *xfs_errortag_attrs[] = {
 	XFS_ERRORTAG_ATTR_LIST(noerror),
@@ -224,6 +226,7 @@ static struct attribute *xfs_errortag_attrs[] = {
 	XFS_ERRORTAG_ATTR_LIST(attr_leaf_to_node),
 	XFS_ERRORTAG_ATTR_LIST(wb_delay_ms),
 	XFS_ERRORTAG_ATTR_LIST(write_delay_ms),
+	XFS_ERRORTAG_ATTR_LIST(exchmaps_finish_one),
 	NULL,
 };
 ATTRIBUTE_GROUPS(xfs_errortag);
@@ -240,15 +243,15 @@ xfs_errortag_init(
 {
 	int ret;
 
-	mp->m_errortag = kmem_zalloc(sizeof(unsigned int) * XFS_ERRTAG_MAX,
-			KM_MAYFAIL);
+	mp->m_errortag = kzalloc(sizeof(unsigned int) * XFS_ERRTAG_MAX,
+				GFP_KERNEL | __GFP_RETRY_MAYFAIL);
 	if (!mp->m_errortag)
 		return -ENOMEM;
 
 	ret = xfs_sysfs_init(&mp->m_errortag_kobj, &xfs_errortag_ktype,
 				&mp->m_kobj, "errortag");
 	if (ret)
-		kmem_free(mp->m_errortag);
+		kfree(mp->m_errortag);
 	return ret;
 }
 
@@ -257,7 +260,7 @@ xfs_errortag_del(
 	struct xfs_mount	*mp)
 {
 	xfs_sysfs_del(&mp->m_errortag_kobj);
-	kmem_free(mp->m_errortag);
+	kfree(mp->m_errortag);
 }
 
 static bool
diff --git a/fs/xfs/xfs_exchmaps_item.c b/fs/xfs/xfs_exchmaps_item.c
new file mode 100644
index 0000000000..264a121c5e
--- /dev/null
+++ b/fs/xfs/xfs_exchmaps_item.c
@@ -0,0 +1,614 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2020-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_bit.h"
+#include "xfs_shared.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_inode.h"
+#include "xfs_trans.h"
+#include "xfs_trans_priv.h"
+#include "xfs_exchmaps_item.h"
+#include "xfs_exchmaps.h"
+#include "xfs_log.h"
+#include "xfs_bmap.h"
+#include "xfs_icache.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_trans_space.h"
+#include "xfs_error.h"
+#include "xfs_log_priv.h"
+#include "xfs_log_recover.h"
+#include "xfs_exchrange.h"
+#include "xfs_trace.h"
+
+struct kmem_cache	*xfs_xmi_cache;
+struct kmem_cache	*xfs_xmd_cache;
+
+static const struct xfs_item_ops xfs_xmi_item_ops;
+
+static inline struct xfs_xmi_log_item *XMI_ITEM(struct xfs_log_item *lip)
+{
+	return container_of(lip, struct xfs_xmi_log_item, xmi_item);
+}
+
+STATIC void
+xfs_xmi_item_free(
+	struct xfs_xmi_log_item	*xmi_lip)
+{
+	kvfree(xmi_lip->xmi_item.li_lv_shadow);
+	kmem_cache_free(xfs_xmi_cache, xmi_lip);
+}
+
+/*
+ * Freeing the XMI requires that we remove it from the AIL if it has already
+ * been placed there. However, the XMI may not yet have been placed in the AIL
+ * when called by xfs_xmi_release() from XMD processing due to the ordering of
+ * committed vs unpin operations in bulk insert operations. Hence the reference
+ * count to ensure only the last caller frees the XMI.
+ */
+STATIC void
+xfs_xmi_release(
+	struct xfs_xmi_log_item	*xmi_lip)
+{
+	ASSERT(atomic_read(&xmi_lip->xmi_refcount) > 0);
+	if (atomic_dec_and_test(&xmi_lip->xmi_refcount)) {
+		xfs_trans_ail_delete(&xmi_lip->xmi_item, 0);
+		xfs_xmi_item_free(xmi_lip);
+	}
+}
+
+
+STATIC void
+xfs_xmi_item_size(
+	struct xfs_log_item	*lip,
+	int			*nvecs,
+	int			*nbytes)
+{
+	*nvecs += 1;
+	*nbytes += sizeof(struct xfs_xmi_log_format);
+}
+
+/*
+ * This is called to fill in the vector of log iovecs for the given xmi log
+ * item. We use only 1 iovec, and we point that at the xmi_log_format structure
+ * embedded in the xmi item.
+ */
+STATIC void
+xfs_xmi_item_format(
+	struct xfs_log_item	*lip,
+	struct xfs_log_vec	*lv)
+{
+	struct xfs_xmi_log_item	*xmi_lip = XMI_ITEM(lip);
+	struct xfs_log_iovec	*vecp = NULL;
+
+	xmi_lip->xmi_format.xmi_type = XFS_LI_XMI;
+	xmi_lip->xmi_format.xmi_size = 1;
+
+	xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_XMI_FORMAT,
+			&xmi_lip->xmi_format,
+			sizeof(struct xfs_xmi_log_format));
+}
+
+/*
+ * The unpin operation is the last place an XMI is manipulated in the log. It
+ * is either inserted in the AIL or aborted in the event of a log I/O error. In
+ * either case, the XMI transaction has been successfully committed to make it
+ * this far. Therefore, we expect whoever committed the XMI to either construct
+ * and commit the XMD or drop the XMD's reference in the event of error. Simply
+ * drop the log's XMI reference now that the log is done with it.
+ */
+STATIC void
+xfs_xmi_item_unpin(
+	struct xfs_log_item	*lip,
+	int			remove)
+{
+	struct xfs_xmi_log_item	*xmi_lip = XMI_ITEM(lip);
+
+	xfs_xmi_release(xmi_lip);
+}
+
+/*
+ * The XMI has been either committed or aborted if the transaction has been
+ * cancelled. If the transaction was cancelled, an XMD isn't going to be
+ * constructed and thus we free the XMI here directly.
+ */
+STATIC void
+xfs_xmi_item_release(
+	struct xfs_log_item	*lip)
+{
+	xfs_xmi_release(XMI_ITEM(lip));
+}
+
+/* Allocate and initialize an xmi item. */
+STATIC struct xfs_xmi_log_item *
+xfs_xmi_init(
+	struct xfs_mount	*mp)
+
+{
+	struct xfs_xmi_log_item	*xmi_lip;
+
+	xmi_lip = kmem_cache_zalloc(xfs_xmi_cache, GFP_KERNEL | __GFP_NOFAIL);
+
+	xfs_log_item_init(mp, &xmi_lip->xmi_item, XFS_LI_XMI, &xfs_xmi_item_ops);
+	xmi_lip->xmi_format.xmi_id = (uintptr_t)(void *)xmi_lip;
+	atomic_set(&xmi_lip->xmi_refcount, 2);
+
+	return xmi_lip;
+}
+
+static inline struct xfs_xmd_log_item *XMD_ITEM(struct xfs_log_item *lip)
+{
+	return container_of(lip, struct xfs_xmd_log_item, xmd_item);
+}
+
+STATIC void
+xfs_xmd_item_size(
+	struct xfs_log_item	*lip,
+	int			*nvecs,
+	int			*nbytes)
+{
+	*nvecs += 1;
+	*nbytes += sizeof(struct xfs_xmd_log_format);
+}
+
+/*
+ * This is called to fill in the vector of log iovecs for the given xmd log
+ * item. We use only 1 iovec, and we point that at the xmd_log_format structure
+ * embedded in the xmd item.
+ */
+STATIC void
+xfs_xmd_item_format(
+	struct xfs_log_item	*lip,
+	struct xfs_log_vec	*lv)
+{
+	struct xfs_xmd_log_item	*xmd_lip = XMD_ITEM(lip);
+	struct xfs_log_iovec	*vecp = NULL;
+
+	xmd_lip->xmd_format.xmd_type = XFS_LI_XMD;
+	xmd_lip->xmd_format.xmd_size = 1;
+
+	xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_XMD_FORMAT, &xmd_lip->xmd_format,
+			sizeof(struct xfs_xmd_log_format));
+}
+
+/*
+ * The XMD is either committed or aborted if the transaction is cancelled. If
+ * the transaction is cancelled, drop our reference to the XMI and free the
+ * XMD.
+ */
+STATIC void
+xfs_xmd_item_release(
+	struct xfs_log_item	*lip)
+{
+	struct xfs_xmd_log_item	*xmd_lip = XMD_ITEM(lip);
+
+	xfs_xmi_release(xmd_lip->xmd_intent_log_item);
+	kvfree(xmd_lip->xmd_item.li_lv_shadow);
+	kmem_cache_free(xfs_xmd_cache, xmd_lip);
+}
+
+static struct xfs_log_item *
+xfs_xmd_item_intent(
+	struct xfs_log_item	*lip)
+{
+	return &XMD_ITEM(lip)->xmd_intent_log_item->xmi_item;
+}
+
+static const struct xfs_item_ops xfs_xmd_item_ops = {
+	.flags		= XFS_ITEM_RELEASE_WHEN_COMMITTED |
+			  XFS_ITEM_INTENT_DONE,
+	.iop_size	= xfs_xmd_item_size,
+	.iop_format	= xfs_xmd_item_format,
+	.iop_release	= xfs_xmd_item_release,
+	.iop_intent	= xfs_xmd_item_intent,
+};
+
+/* Log file mapping exchange information in the intent item. */
+STATIC struct xfs_log_item *
+xfs_exchmaps_create_intent(
+	struct xfs_trans		*tp,
+	struct list_head		*items,
+	unsigned int			count,
+	bool				sort)
+{
+	struct xfs_xmi_log_item		*xmi_lip;
+	struct xfs_exchmaps_intent	*xmi;
+	struct xfs_xmi_log_format	*xlf;
+
+	ASSERT(count == 1);
+
+	xmi = list_first_entry_or_null(items, struct xfs_exchmaps_intent,
+			xmi_list);
+
+	xmi_lip = xfs_xmi_init(tp->t_mountp);
+	xlf = &xmi_lip->xmi_format;
+
+	xlf->xmi_inode1 = xmi->xmi_ip1->i_ino;
+	xlf->xmi_igen1 = VFS_I(xmi->xmi_ip1)->i_generation;
+	xlf->xmi_inode2 = xmi->xmi_ip2->i_ino;
+	xlf->xmi_igen2 = VFS_I(xmi->xmi_ip2)->i_generation;
+	xlf->xmi_startoff1 = xmi->xmi_startoff1;
+	xlf->xmi_startoff2 = xmi->xmi_startoff2;
+	xlf->xmi_blockcount = xmi->xmi_blockcount;
+	xlf->xmi_isize1 = xmi->xmi_isize1;
+	xlf->xmi_isize2 = xmi->xmi_isize2;
+	xlf->xmi_flags = xmi->xmi_flags & XFS_EXCHMAPS_LOGGED_FLAGS;
+
+	return &xmi_lip->xmi_item;
+}
+
+STATIC struct xfs_log_item *
+xfs_exchmaps_create_done(
+	struct xfs_trans		*tp,
+	struct xfs_log_item		*intent,
+	unsigned int			count)
+{
+	struct xfs_xmi_log_item		*xmi_lip = XMI_ITEM(intent);
+	struct xfs_xmd_log_item		*xmd_lip;
+
+	xmd_lip = kmem_cache_zalloc(xfs_xmd_cache, GFP_KERNEL | __GFP_NOFAIL);
+	xfs_log_item_init(tp->t_mountp, &xmd_lip->xmd_item, XFS_LI_XMD,
+			  &xfs_xmd_item_ops);
+	xmd_lip->xmd_intent_log_item = xmi_lip;
+	xmd_lip->xmd_format.xmd_xmi_id = xmi_lip->xmi_format.xmi_id;
+
+	return &xmd_lip->xmd_item;
+}
+
+/* Add this deferred XMI to the transaction. */
+void
+xfs_exchmaps_defer_add(
+	struct xfs_trans		*tp,
+	struct xfs_exchmaps_intent	*xmi)
+{
+	trace_xfs_exchmaps_defer(tp->t_mountp, xmi);
+
+	xfs_defer_add(tp, &xmi->xmi_list, &xfs_exchmaps_defer_type);
+}
+
+static inline struct xfs_exchmaps_intent *xmi_entry(const struct list_head *e)
+{
+	return list_entry(e, struct xfs_exchmaps_intent, xmi_list);
+}
+
+/* Cancel a deferred file mapping exchange. */
+STATIC void
+xfs_exchmaps_cancel_item(
+	struct list_head		*item)
+{
+	struct xfs_exchmaps_intent	*xmi = xmi_entry(item);
+
+	kmem_cache_free(xfs_exchmaps_intent_cache, xmi);
+}
+
+/* Process a deferred file mapping exchange. */
+STATIC int
+xfs_exchmaps_finish_item(
+	struct xfs_trans		*tp,
+	struct xfs_log_item		*done,
+	struct list_head		*item,
+	struct xfs_btree_cur		**state)
+{
+	struct xfs_exchmaps_intent	*xmi = xmi_entry(item);
+	int				error;
+
+	/*
+	 * Exchange one more mappings between two files.  If there's still more
+	 * work to do, we want to requeue ourselves after all other pending
+	 * deferred operations have finished.  This includes all of the dfops
+	 * that we queued directly as well as any new ones created in the
+	 * process of finishing the others.  Doing so prevents us from queuing
+	 * a large number of XMI log items in kernel memory, which in turn
+	 * prevents us from pinning the tail of the log (while logging those
+	 * new XMI items) until the first XMI items can be processed.
+	 */
+	error = xfs_exchmaps_finish_one(tp, xmi);
+	if (error != -EAGAIN)
+		xfs_exchmaps_cancel_item(item);
+	return error;
+}
+
+/* Abort all pending XMIs. */
+STATIC void
+xfs_exchmaps_abort_intent(
+	struct xfs_log_item		*intent)
+{
+	xfs_xmi_release(XMI_ITEM(intent));
+}
+
+/* Is this recovered XMI ok? */
+static inline bool
+xfs_xmi_validate(
+	struct xfs_mount		*mp,
+	struct xfs_xmi_log_item		*xmi_lip)
+{
+	struct xfs_xmi_log_format	*xlf = &xmi_lip->xmi_format;
+
+	if (!xfs_has_exchange_range(mp))
+		return false;
+
+	if (xmi_lip->xmi_format.__pad != 0)
+		return false;
+
+	if (xlf->xmi_flags & ~XFS_EXCHMAPS_LOGGED_FLAGS)
+		return false;
+
+	if (!xfs_verify_ino(mp, xlf->xmi_inode1) ||
+	    !xfs_verify_ino(mp, xlf->xmi_inode2))
+		return false;
+
+	if (!xfs_verify_fileext(mp, xlf->xmi_startoff1, xlf->xmi_blockcount))
+		return false;
+
+	return xfs_verify_fileext(mp, xlf->xmi_startoff2, xlf->xmi_blockcount);
+}
+
+/*
+ * Use the recovered log state to create a new request, estimate resource
+ * requirements, and create a new incore intent state.
+ */
+STATIC struct xfs_exchmaps_intent *
+xfs_xmi_item_recover_intent(
+	struct xfs_mount		*mp,
+	struct xfs_defer_pending	*dfp,
+	const struct xfs_xmi_log_format	*xlf,
+	struct xfs_exchmaps_req		*req,
+	struct xfs_inode		**ipp1,
+	struct xfs_inode		**ipp2)
+{
+	struct xfs_inode		*ip1, *ip2;
+	struct xfs_exchmaps_intent	*xmi;
+	int				error;
+
+	/*
+	 * Grab both inodes and set IRECOVERY to prevent trimming of post-eof
+	 * mappings and freeing of unlinked inodes until we're totally done
+	 * processing files.  The ondisk format of this new log item contains
+	 * file handle information, which is why recovery for other items do
+	 * not check the inode generation number.
+	 */
+	error = xlog_recover_iget_handle(mp, xlf->xmi_inode1, xlf->xmi_igen1,
+			&ip1);
+	if (error) {
+		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, xlf,
+				sizeof(*xlf));
+		return ERR_PTR(error);
+	}
+
+	error = xlog_recover_iget_handle(mp, xlf->xmi_inode2, xlf->xmi_igen2,
+			&ip2);
+	if (error) {
+		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, xlf,
+				sizeof(*xlf));
+		goto err_rele1;
+	}
+
+	req->ip1 = ip1;
+	req->ip2 = ip2;
+	req->startoff1 = xlf->xmi_startoff1;
+	req->startoff2 = xlf->xmi_startoff2;
+	req->blockcount = xlf->xmi_blockcount;
+	req->flags = xlf->xmi_flags & XFS_EXCHMAPS_PARAMS;
+
+	xfs_exchrange_ilock(NULL, ip1, ip2);
+	error = xfs_exchmaps_estimate(req);
+	xfs_exchrange_iunlock(ip1, ip2);
+	if (error)
+		goto err_rele2;
+
+	*ipp1 = ip1;
+	*ipp2 = ip2;
+	xmi = xfs_exchmaps_init_intent(req);
+	xfs_defer_add_item(dfp, &xmi->xmi_list);
+	return xmi;
+
+err_rele2:
+	xfs_irele(ip2);
+err_rele1:
+	xfs_irele(ip1);
+	req->ip2 = req->ip1 = NULL;
+	return ERR_PTR(error);
+}
+
+/* Process a file mapping exchange item that was recovered from the log. */
+STATIC int
+xfs_exchmaps_recover_work(
+	struct xfs_defer_pending	*dfp,
+	struct list_head		*capture_list)
+{
+	struct xfs_exchmaps_req		req = { .flags = 0 };
+	struct xfs_trans_res		resv;
+	struct xfs_exchmaps_intent	*xmi;
+	struct xfs_log_item		*lip = dfp->dfp_intent;
+	struct xfs_xmi_log_item		*xmi_lip = XMI_ITEM(lip);
+	struct xfs_mount		*mp = lip->li_log->l_mp;
+	struct xfs_trans		*tp;
+	struct xfs_inode		*ip1, *ip2;
+	int				error = 0;
+
+	if (!xfs_xmi_validate(mp, xmi_lip)) {
+		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+				&xmi_lip->xmi_format,
+				sizeof(xmi_lip->xmi_format));
+		return -EFSCORRUPTED;
+	}
+
+	xmi = xfs_xmi_item_recover_intent(mp, dfp, &xmi_lip->xmi_format, &req,
+			&ip1, &ip2);
+	if (IS_ERR(xmi))
+		return PTR_ERR(xmi);
+
+	trace_xfs_exchmaps_recover(mp, xmi);
+
+	resv = xlog_recover_resv(&M_RES(mp)->tr_write);
+	error = xfs_trans_alloc(mp, &resv, req.resblks, 0, 0, &tp);
+	if (error)
+		goto err_rele;
+
+	xfs_exchrange_ilock(tp, ip1, ip2);
+
+	xfs_exchmaps_ensure_reflink(tp, xmi);
+	xfs_exchmaps_upgrade_extent_counts(tp, xmi);
+	error = xlog_recover_finish_intent(tp, dfp);
+	if (error == -EFSCORRUPTED)
+		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+				&xmi_lip->xmi_format,
+				sizeof(xmi_lip->xmi_format));
+	if (error)
+		goto err_cancel;
+
+	/*
+	 * Commit transaction, which frees the transaction and saves the inodes
+	 * for later replay activities.
+	 */
+	error = xfs_defer_ops_capture_and_commit(tp, capture_list);
+	goto err_unlock;
+
+err_cancel:
+	xfs_trans_cancel(tp);
+err_unlock:
+	xfs_exchrange_iunlock(ip1, ip2);
+err_rele:
+	xfs_irele(ip2);
+	xfs_irele(ip1);
+	return error;
+}
+
+/* Relog an intent item to push the log tail forward. */
+static struct xfs_log_item *
+xfs_exchmaps_relog_intent(
+	struct xfs_trans		*tp,
+	struct xfs_log_item		*intent,
+	struct xfs_log_item		*done_item)
+{
+	struct xfs_xmi_log_item		*xmi_lip;
+	struct xfs_xmi_log_format	*old_xlf, *new_xlf;
+
+	old_xlf = &XMI_ITEM(intent)->xmi_format;
+
+	xmi_lip = xfs_xmi_init(tp->t_mountp);
+	new_xlf = &xmi_lip->xmi_format;
+
+	new_xlf->xmi_inode1	= old_xlf->xmi_inode1;
+	new_xlf->xmi_inode2	= old_xlf->xmi_inode2;
+	new_xlf->xmi_igen1	= old_xlf->xmi_igen1;
+	new_xlf->xmi_igen2	= old_xlf->xmi_igen2;
+	new_xlf->xmi_startoff1	= old_xlf->xmi_startoff1;
+	new_xlf->xmi_startoff2	= old_xlf->xmi_startoff2;
+	new_xlf->xmi_blockcount	= old_xlf->xmi_blockcount;
+	new_xlf->xmi_flags	= old_xlf->xmi_flags;
+	new_xlf->xmi_isize1	= old_xlf->xmi_isize1;
+	new_xlf->xmi_isize2	= old_xlf->xmi_isize2;
+
+	return &xmi_lip->xmi_item;
+}
+
+const struct xfs_defer_op_type xfs_exchmaps_defer_type = {
+	.name		= "exchmaps",
+	.max_items	= 1,
+	.create_intent	= xfs_exchmaps_create_intent,
+	.abort_intent	= xfs_exchmaps_abort_intent,
+	.create_done	= xfs_exchmaps_create_done,
+	.finish_item	= xfs_exchmaps_finish_item,
+	.cancel_item	= xfs_exchmaps_cancel_item,
+	.recover_work	= xfs_exchmaps_recover_work,
+	.relog_intent	= xfs_exchmaps_relog_intent,
+};
+
+STATIC bool
+xfs_xmi_item_match(
+	struct xfs_log_item	*lip,
+	uint64_t		intent_id)
+{
+	return XMI_ITEM(lip)->xmi_format.xmi_id == intent_id;
+}
+
+static const struct xfs_item_ops xfs_xmi_item_ops = {
+	.flags		= XFS_ITEM_INTENT,
+	.iop_size	= xfs_xmi_item_size,
+	.iop_format	= xfs_xmi_item_format,
+	.iop_unpin	= xfs_xmi_item_unpin,
+	.iop_release	= xfs_xmi_item_release,
+	.iop_match	= xfs_xmi_item_match,
+};
+
+/*
+ * This routine is called to create an in-core file mapping exchange item from
+ * the xmi format structure which was logged on disk.  It allocates an in-core
+ * xmi, copies the exchange information from the format structure into it, and
+ * adds the xmi to the AIL with the given LSN.
+ */
+STATIC int
+xlog_recover_xmi_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			lsn)
+{
+	struct xfs_mount		*mp = log->l_mp;
+	struct xfs_xmi_log_item		*xmi_lip;
+	struct xfs_xmi_log_format	*xmi_formatp;
+	size_t				len;
+
+	len = sizeof(struct xfs_xmi_log_format);
+	if (item->ri_buf[0].i_len != len) {
+		XFS_ERROR_REPORT(__func__, XFS_ERRLEVEL_LOW, log->l_mp);
+		return -EFSCORRUPTED;
+	}
+
+	xmi_formatp = item->ri_buf[0].i_addr;
+	if (xmi_formatp->__pad != 0) {
+		XFS_ERROR_REPORT(__func__, XFS_ERRLEVEL_LOW, log->l_mp);
+		return -EFSCORRUPTED;
+	}
+
+	xmi_lip = xfs_xmi_init(mp);
+	memcpy(&xmi_lip->xmi_format, xmi_formatp, len);
+
+	xlog_recover_intent_item(log, &xmi_lip->xmi_item, lsn,
+			&xfs_exchmaps_defer_type);
+	return 0;
+}
+
+const struct xlog_recover_item_ops xlog_xmi_item_ops = {
+	.item_type		= XFS_LI_XMI,
+	.commit_pass2		= xlog_recover_xmi_commit_pass2,
+};
+
+/*
+ * This routine is called when an XMD format structure is found in a committed
+ * transaction in the log. Its purpose is to cancel the corresponding XMI if it
+ * was still in the log. To do this it searches the AIL for the XMI with an id
+ * equal to that in the XMD format structure. If we find it we drop the XMD
+ * reference, which removes the XMI from the AIL and frees it.
+ */
+STATIC int
+xlog_recover_xmd_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			lsn)
+{
+	struct xfs_xmd_log_format	*xmd_formatp;
+
+	xmd_formatp = item->ri_buf[0].i_addr;
+	if (item->ri_buf[0].i_len != sizeof(struct xfs_xmd_log_format)) {
+		XFS_ERROR_REPORT(__func__, XFS_ERRLEVEL_LOW, log->l_mp);
+		return -EFSCORRUPTED;
+	}
+
+	xlog_recover_release_intent(log, XFS_LI_XMI, xmd_formatp->xmd_xmi_id);
+	return 0;
+}
+
+const struct xlog_recover_item_ops xlog_xmd_item_ops = {
+	.item_type		= XFS_LI_XMD,
+	.commit_pass2		= xlog_recover_xmd_commit_pass2,
+};
diff --git a/fs/xfs/xfs_exchmaps_item.h b/fs/xfs/xfs_exchmaps_item.h
new file mode 100644
index 0000000000..efa368d25d
--- /dev/null
+++ b/fs/xfs/xfs_exchmaps_item.h
@@ -0,0 +1,64 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (c) 2020-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef	__XFS_EXCHMAPS_ITEM_H__
+#define	__XFS_EXCHMAPS_ITEM_H__
+
+/*
+ * The file mapping exchange intent item helps us exchange multiple file
+ * mappings between two inode forks.  It does this by tracking the range of
+ * file block offsets that still need to be exchanged, and relogs as progress
+ * happens.
+ *
+ * *I items should be recorded in the *first* of a series of rolled
+ * transactions, and the *D items should be recorded in the same transaction
+ * that records the associated bmbt updates.
+ *
+ * Should the system crash after the commit of the first transaction but
+ * before the commit of the final transaction in a series, log recovery will
+ * use the redo information recorded by the intent items to replay the
+ * rest of the mapping exchanges.
+ */
+
+/* kernel only XMI/XMD definitions */
+
+struct xfs_mount;
+struct kmem_cache;
+
+/*
+ * This is the incore file mapping exchange intent log item.  It is used to log
+ * the fact that we are exchanging mappings between two files.  It is used in
+ * conjunction with the incore file mapping exchange done log item described
+ * below.
+ *
+ * These log items follow the same rules as struct xfs_efi_log_item; see the
+ * comments about that structure (in xfs_extfree_item.h) for more details.
+ */
+struct xfs_xmi_log_item {
+	struct xfs_log_item		xmi_item;
+	atomic_t			xmi_refcount;
+	struct xfs_xmi_log_format	xmi_format;
+};
+
+/*
+ * This is the incore file mapping exchange done log item.  It is used to log
+ * the fact that an exchange mentioned in an earlier xmi item have been
+ * performed.
+ */
+struct xfs_xmd_log_item {
+	struct xfs_log_item		xmd_item;
+	struct xfs_xmi_log_item		*xmd_intent_log_item;
+	struct xfs_xmd_log_format	xmd_format;
+};
+
+extern struct kmem_cache	*xfs_xmi_cache;
+extern struct kmem_cache	*xfs_xmd_cache;
+
+struct xfs_exchmaps_intent;
+
+void xfs_exchmaps_defer_add(struct xfs_trans *tp,
+		struct xfs_exchmaps_intent *xmi);
+
+#endif	/* __XFS_EXCHMAPS_ITEM_H__ */
diff --git a/fs/xfs/xfs_exchrange.c b/fs/xfs/xfs_exchrange.c
new file mode 100644
index 0000000000..c8a655c92c
--- /dev/null
+++ b/fs/xfs/xfs_exchrange.c
@@ -0,0 +1,804 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2020-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_inode.h"
+#include "xfs_trans.h"
+#include "xfs_quota.h"
+#include "xfs_bmap_util.h"
+#include "xfs_reflink.h"
+#include "xfs_trace.h"
+#include "xfs_exchrange.h"
+#include "xfs_exchmaps.h"
+#include "xfs_sb.h"
+#include "xfs_icache.h"
+#include "xfs_log.h"
+#include "xfs_rtbitmap.h"
+#include <linux/fsnotify.h>
+
+/* Lock (and optionally join) two inodes for a file range exchange. */
+void
+xfs_exchrange_ilock(
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip1,
+	struct xfs_inode	*ip2)
+{
+	if (ip1 != ip2)
+		xfs_lock_two_inodes(ip1, XFS_ILOCK_EXCL,
+				    ip2, XFS_ILOCK_EXCL);
+	else
+		xfs_ilock(ip1, XFS_ILOCK_EXCL);
+	if (tp) {
+		xfs_trans_ijoin(tp, ip1, 0);
+		if (ip2 != ip1)
+			xfs_trans_ijoin(tp, ip2, 0);
+	}
+
+}
+
+/* Unlock two inodes after a file range exchange operation. */
+void
+xfs_exchrange_iunlock(
+	struct xfs_inode	*ip1,
+	struct xfs_inode	*ip2)
+{
+	if (ip2 != ip1)
+		xfs_iunlock(ip2, XFS_ILOCK_EXCL);
+	xfs_iunlock(ip1, XFS_ILOCK_EXCL);
+}
+
+/*
+ * Estimate the resource requirements to exchange file contents between the two
+ * files.  The caller is required to hold the IOLOCK and the MMAPLOCK and to
+ * have flushed both inodes' pagecache and active direct-ios.
+ */
+int
+xfs_exchrange_estimate(
+	struct xfs_exchmaps_req	*req)
+{
+	int			error;
+
+	xfs_exchrange_ilock(NULL, req->ip1, req->ip2);
+	error = xfs_exchmaps_estimate(req);
+	xfs_exchrange_iunlock(req->ip1, req->ip2);
+	return error;
+}
+
+#define QRETRY_IP1	(0x1)
+#define QRETRY_IP2	(0x2)
+
+/*
+ * Obtain a quota reservation to make sure we don't hit EDQUOT.  We can skip
+ * this if quota enforcement is disabled or if both inodes' dquots are the
+ * same.  The qretry structure must be initialized to zeroes before the first
+ * call to this function.
+ */
+STATIC int
+xfs_exchrange_reserve_quota(
+	struct xfs_trans		*tp,
+	const struct xfs_exchmaps_req	*req,
+	unsigned int			*qretry)
+{
+	int64_t				ddelta, rdelta;
+	int				ip1_error = 0;
+	int				error;
+
+	/*
+	 * Don't bother with a quota reservation if we're not enforcing them
+	 * or the two inodes have the same dquots.
+	 */
+	if (!XFS_IS_QUOTA_ON(tp->t_mountp) || req->ip1 == req->ip2 ||
+	    (req->ip1->i_udquot == req->ip2->i_udquot &&
+	     req->ip1->i_gdquot == req->ip2->i_gdquot &&
+	     req->ip1->i_pdquot == req->ip2->i_pdquot))
+		return 0;
+
+	*qretry = 0;
+
+	/*
+	 * For each file, compute the net gain in the number of regular blocks
+	 * that will be mapped into that file and reserve that much quota.  The
+	 * quota counts must be able to absorb at least that much space.
+	 */
+	ddelta = req->ip2_bcount - req->ip1_bcount;
+	rdelta = req->ip2_rtbcount - req->ip1_rtbcount;
+	if (ddelta > 0 || rdelta > 0) {
+		error = xfs_trans_reserve_quota_nblks(tp, req->ip1,
+				ddelta > 0 ? ddelta : 0,
+				rdelta > 0 ? rdelta : 0,
+				false);
+		if (error == -EDQUOT || error == -ENOSPC) {
+			/*
+			 * Save this error and see what happens if we try to
+			 * reserve quota for ip2.  Then report both.
+			 */
+			*qretry |= QRETRY_IP1;
+			ip1_error = error;
+			error = 0;
+		}
+		if (error)
+			return error;
+	}
+	if (ddelta < 0 || rdelta < 0) {
+		error = xfs_trans_reserve_quota_nblks(tp, req->ip2,
+				ddelta < 0 ? -ddelta : 0,
+				rdelta < 0 ? -rdelta : 0,
+				false);
+		if (error == -EDQUOT || error == -ENOSPC)
+			*qretry |= QRETRY_IP2;
+		if (error)
+			return error;
+	}
+	if (ip1_error)
+		return ip1_error;
+
+	/*
+	 * For each file, forcibly reserve the gross gain in mapped blocks so
+	 * that we don't trip over any quota block reservation assertions.
+	 * We must reserve the gross gain because the quota code subtracts from
+	 * bcount the number of blocks that we unmap; it does not add that
+	 * quantity back to the quota block reservation.
+	 */
+	error = xfs_trans_reserve_quota_nblks(tp, req->ip1, req->ip1_bcount,
+			req->ip1_rtbcount, true);
+	if (error)
+		return error;
+
+	return xfs_trans_reserve_quota_nblks(tp, req->ip2, req->ip2_bcount,
+			req->ip2_rtbcount, true);
+}
+
+/* Exchange the mappings (and hence the contents) of two files' forks. */
+STATIC int
+xfs_exchrange_mappings(
+	const struct xfs_exchrange	*fxr,
+	struct xfs_inode		*ip1,
+	struct xfs_inode		*ip2)
+{
+	struct xfs_mount		*mp = ip1->i_mount;
+	struct xfs_exchmaps_req		req = {
+		.ip1			= ip1,
+		.ip2			= ip2,
+		.startoff1		= XFS_B_TO_FSBT(mp, fxr->file1_offset),
+		.startoff2		= XFS_B_TO_FSBT(mp, fxr->file2_offset),
+		.blockcount		= XFS_B_TO_FSB(mp, fxr->length),
+	};
+	struct xfs_trans		*tp;
+	unsigned int			qretry;
+	bool				retried = false;
+	int				error;
+
+	trace_xfs_exchrange_mappings(fxr, ip1, ip2);
+
+	if (fxr->flags & XFS_EXCHANGE_RANGE_TO_EOF)
+		req.flags |= XFS_EXCHMAPS_SET_SIZES;
+	if (fxr->flags & XFS_EXCHANGE_RANGE_FILE1_WRITTEN)
+		req.flags |= XFS_EXCHMAPS_INO1_WRITTEN;
+
+	/*
+	 * Round the request length up to the nearest file allocation unit.
+	 * The prep function already checked that the request offsets and
+	 * length in @fxr are safe to round up.
+	 */
+	if (xfs_inode_has_bigrtalloc(ip2))
+		req.blockcount = xfs_rtb_roundup_rtx(mp, req.blockcount);
+
+	error = xfs_exchrange_estimate(&req);
+	if (error)
+		return error;
+
+retry:
+	/* Allocate the transaction, lock the inodes, and join them. */
+	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, req.resblks, 0,
+			XFS_TRANS_RES_FDBLKS, &tp);
+	if (error)
+		return error;
+
+	xfs_exchrange_ilock(tp, ip1, ip2);
+
+	trace_xfs_exchrange_before(ip2, 2);
+	trace_xfs_exchrange_before(ip1, 1);
+
+	error = xfs_exchmaps_check_forks(mp, &req);
+	if (error)
+		goto out_trans_cancel;
+
+	/*
+	 * Reserve ourselves some quota if any of them are in enforcing mode.
+	 * In theory we only need enough to satisfy the change in the number
+	 * of blocks between the two ranges being remapped.
+	 */
+	error = xfs_exchrange_reserve_quota(tp, &req, &qretry);
+	if ((error == -EDQUOT || error == -ENOSPC) && !retried) {
+		xfs_trans_cancel(tp);
+		xfs_exchrange_iunlock(ip1, ip2);
+		if (qretry & QRETRY_IP1)
+			xfs_blockgc_free_quota(ip1, 0);
+		if (qretry & QRETRY_IP2)
+			xfs_blockgc_free_quota(ip2, 0);
+		retried = true;
+		goto retry;
+	}
+	if (error)
+		goto out_trans_cancel;
+
+	/* If we got this far on a dry run, all parameters are ok. */
+	if (fxr->flags & XFS_EXCHANGE_RANGE_DRY_RUN)
+		goto out_trans_cancel;
+
+	/* Update the mtime and ctime of both files. */
+	if (fxr->flags & __XFS_EXCHANGE_RANGE_UPD_CMTIME1)
+		xfs_trans_ichgtime(tp, ip1, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
+	if (fxr->flags & __XFS_EXCHANGE_RANGE_UPD_CMTIME2)
+		xfs_trans_ichgtime(tp, ip2, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
+
+	xfs_exchange_mappings(tp, &req);
+
+	/*
+	 * Force the log to persist metadata updates if the caller or the
+	 * administrator requires this.  The generic prep function already
+	 * flushed the relevant parts of the page cache.
+	 */
+	if (xfs_has_wsync(mp) || (fxr->flags & XFS_EXCHANGE_RANGE_DSYNC))
+		xfs_trans_set_sync(tp);
+
+	error = xfs_trans_commit(tp);
+
+	trace_xfs_exchrange_after(ip2, 2);
+	trace_xfs_exchrange_after(ip1, 1);
+
+	if (error)
+		goto out_unlock;
+
+	/*
+	 * If the caller wanted us to exchange the contents of two complete
+	 * files of unequal length, exchange the incore sizes now.  This should
+	 * be safe because we flushed both files' page caches, exchanged all
+	 * the mappings, and updated the ondisk sizes.
+	 */
+	if (fxr->flags & XFS_EXCHANGE_RANGE_TO_EOF) {
+		loff_t	temp;
+
+		temp = i_size_read(VFS_I(ip2));
+		i_size_write(VFS_I(ip2), i_size_read(VFS_I(ip1)));
+		i_size_write(VFS_I(ip1), temp);
+	}
+
+out_unlock:
+	xfs_exchrange_iunlock(ip1, ip2);
+	return error;
+
+out_trans_cancel:
+	xfs_trans_cancel(tp);
+	goto out_unlock;
+}
+
+/*
+ * Generic code for exchanging ranges of two files via XFS_IOC_EXCHANGE_RANGE.
+ * This part deals with struct file objects and byte ranges and does not deal
+ * with XFS-specific data structures such as xfs_inodes and block ranges.  This
+ * separation may some day facilitate porting to another filesystem.
+ *
+ * The goal is to exchange fxr.length bytes starting at fxr.file1_offset in
+ * file1 with the same number of bytes starting at fxr.file2_offset in file2.
+ * Implementations must call xfs_exchange_range_prep to prepare the two
+ * files prior to taking locks; and they must update the inode change and mod
+ * times of both files as part of the metadata update.  The timestamp update
+ * and freshness checks must be done atomically as part of the data exchange
+ * operation to ensure correctness of the freshness check.
+ * xfs_exchange_range_finish must be called after the operation completes
+ * successfully but before locks are dropped.
+ */
+
+/* Verify that we have security clearance to perform this operation. */
+static int
+xfs_exchange_range_verify_area(
+	struct xfs_exchrange	*fxr)
+{
+	int			ret;
+
+	ret = remap_verify_area(fxr->file1, fxr->file1_offset, fxr->length,
+			true);
+	if (ret)
+		return ret;
+
+	return remap_verify_area(fxr->file2, fxr->file2_offset, fxr->length,
+			true);
+}
+
+/*
+ * Performs necessary checks before doing a range exchange, having stabilized
+ * mutable inode attributes via i_rwsem.
+ */
+static inline int
+xfs_exchange_range_checks(
+	struct xfs_exchrange	*fxr,
+	unsigned int		alloc_unit)
+{
+	struct inode		*inode1 = file_inode(fxr->file1);
+	struct inode		*inode2 = file_inode(fxr->file2);
+	uint64_t		allocmask = alloc_unit - 1;
+	int64_t			test_len;
+	uint64_t		blen;
+	loff_t			size1, size2, tmp;
+	int			error;
+
+	/* Don't touch certain kinds of inodes */
+	if (IS_IMMUTABLE(inode1) || IS_IMMUTABLE(inode2))
+		return -EPERM;
+	if (IS_SWAPFILE(inode1) || IS_SWAPFILE(inode2))
+		return -ETXTBSY;
+
+	size1 = i_size_read(inode1);
+	size2 = i_size_read(inode2);
+
+	/* Ranges cannot start after EOF. */
+	if (fxr->file1_offset > size1 || fxr->file2_offset > size2)
+		return -EINVAL;
+
+	/*
+	 * If the caller said to exchange to EOF, we set the length of the
+	 * request large enough to cover everything to the end of both files.
+	 */
+	if (fxr->flags & XFS_EXCHANGE_RANGE_TO_EOF) {
+		fxr->length = max_t(int64_t, size1 - fxr->file1_offset,
+					     size2 - fxr->file2_offset);
+
+		error = xfs_exchange_range_verify_area(fxr);
+		if (error)
+			return error;
+	}
+
+	/*
+	 * The start of both ranges must be aligned to the file allocation
+	 * unit.
+	 */
+	if (!IS_ALIGNED(fxr->file1_offset, alloc_unit) ||
+	    !IS_ALIGNED(fxr->file2_offset, alloc_unit))
+		return -EINVAL;
+
+	/* Ensure offsets don't wrap. */
+	if (check_add_overflow(fxr->file1_offset, fxr->length, &tmp) ||
+	    check_add_overflow(fxr->file2_offset, fxr->length, &tmp))
+		return -EINVAL;
+
+	/*
+	 * We require both ranges to end within EOF, unless we're exchanging
+	 * to EOF.
+	 */
+	if (!(fxr->flags & XFS_EXCHANGE_RANGE_TO_EOF) &&
+	    (fxr->file1_offset + fxr->length > size1 ||
+	     fxr->file2_offset + fxr->length > size2))
+		return -EINVAL;
+
+	/*
+	 * Make sure we don't hit any file size limits.  If we hit any size
+	 * limits such that test_length was adjusted, we abort the whole
+	 * operation.
+	 */
+	test_len = fxr->length;
+	error = generic_write_check_limits(fxr->file2, fxr->file2_offset,
+			&test_len);
+	if (error)
+		return error;
+	error = generic_write_check_limits(fxr->file1, fxr->file1_offset,
+			&test_len);
+	if (error)
+		return error;
+	if (test_len != fxr->length)
+		return -EINVAL;
+
+	/*
+	 * If the user wanted us to exchange up to the infile's EOF, round up
+	 * to the next allocation unit boundary for this check.  Do the same
+	 * for the outfile.
+	 *
+	 * Otherwise, reject the range length if it's not aligned to an
+	 * allocation unit.
+	 */
+	if (fxr->file1_offset + fxr->length == size1)
+		blen = ALIGN(size1, alloc_unit) - fxr->file1_offset;
+	else if (fxr->file2_offset + fxr->length == size2)
+		blen = ALIGN(size2, alloc_unit) - fxr->file2_offset;
+	else if (!IS_ALIGNED(fxr->length, alloc_unit))
+		return -EINVAL;
+	else
+		blen = fxr->length;
+
+	/* Don't allow overlapped exchanges within the same file. */
+	if (inode1 == inode2 &&
+	    fxr->file2_offset + blen > fxr->file1_offset &&
+	    fxr->file1_offset + blen > fxr->file2_offset)
+		return -EINVAL;
+
+	/*
+	 * Ensure that we don't exchange a partial EOF block into the middle of
+	 * another file.
+	 */
+	if ((fxr->length & allocmask) == 0)
+		return 0;
+
+	blen = fxr->length;
+	if (fxr->file2_offset + blen < size2)
+		blen &= ~allocmask;
+
+	if (fxr->file1_offset + blen < size1)
+		blen &= ~allocmask;
+
+	return blen == fxr->length ? 0 : -EINVAL;
+}
+
+/*
+ * Check that the two inodes are eligible for range exchanges, the ranges make
+ * sense, and then flush all dirty data.  Caller must ensure that the inodes
+ * have been locked against any other modifications.
+ */
+static inline int
+xfs_exchange_range_prep(
+	struct xfs_exchrange	*fxr,
+	unsigned int		alloc_unit)
+{
+	struct inode		*inode1 = file_inode(fxr->file1);
+	struct inode		*inode2 = file_inode(fxr->file2);
+	bool			same_inode = (inode1 == inode2);
+	int			error;
+
+	/* Check that we don't violate system file offset limits. */
+	error = xfs_exchange_range_checks(fxr, alloc_unit);
+	if (error || fxr->length == 0)
+		return error;
+
+	/* Wait for the completion of any pending IOs on both files */
+	inode_dio_wait(inode1);
+	if (!same_inode)
+		inode_dio_wait(inode2);
+
+	error = filemap_write_and_wait_range(inode1->i_mapping,
+			fxr->file1_offset,
+			fxr->file1_offset + fxr->length - 1);
+	if (error)
+		return error;
+
+	error = filemap_write_and_wait_range(inode2->i_mapping,
+			fxr->file2_offset,
+			fxr->file2_offset + fxr->length - 1);
+	if (error)
+		return error;
+
+	/*
+	 * If the files or inodes involved require synchronous writes, amend
+	 * the request to force the filesystem to flush all data and metadata
+	 * to disk after the operation completes.
+	 */
+	if (((fxr->file1->f_flags | fxr->file2->f_flags) & O_SYNC) ||
+	    IS_SYNC(inode1) || IS_SYNC(inode2))
+		fxr->flags |= XFS_EXCHANGE_RANGE_DSYNC;
+
+	return 0;
+}
+
+/*
+ * Finish a range exchange operation, if it was successful.  Caller must ensure
+ * that the inodes are still locked against any other modifications.
+ */
+static inline int
+xfs_exchange_range_finish(
+	struct xfs_exchrange	*fxr)
+{
+	int			error;
+
+	error = file_remove_privs(fxr->file1);
+	if (error)
+		return error;
+	if (file_inode(fxr->file1) == file_inode(fxr->file2))
+		return 0;
+
+	return file_remove_privs(fxr->file2);
+}
+
+/*
+ * Check the alignment of an exchange request when the allocation unit size
+ * isn't a power of two.  The generic file-level helpers use (fast)
+ * bitmask-based alignment checks, but here we have to use slow long division.
+ */
+static int
+xfs_exchrange_check_rtalign(
+	const struct xfs_exchrange	*fxr,
+	struct xfs_inode		*ip1,
+	struct xfs_inode		*ip2,
+	unsigned int			alloc_unit)
+{
+	uint64_t			length = fxr->length;
+	uint64_t			blen;
+	loff_t				size1, size2;
+
+	size1 = i_size_read(VFS_I(ip1));
+	size2 = i_size_read(VFS_I(ip2));
+
+	/* The start of both ranges must be aligned to a rt extent. */
+	if (!isaligned_64(fxr->file1_offset, alloc_unit) ||
+	    !isaligned_64(fxr->file2_offset, alloc_unit))
+		return -EINVAL;
+
+	if (fxr->flags & XFS_EXCHANGE_RANGE_TO_EOF)
+		length = max_t(int64_t, size1 - fxr->file1_offset,
+					size2 - fxr->file2_offset);
+
+	/*
+	 * If the user wanted us to exchange up to the infile's EOF, round up
+	 * to the next rt extent boundary for this check.  Do the same for the
+	 * outfile.
+	 *
+	 * Otherwise, reject the range length if it's not rt extent aligned.
+	 * We already confirmed the starting offsets' rt extent block
+	 * alignment.
+	 */
+	if (fxr->file1_offset + length == size1)
+		blen = roundup_64(size1, alloc_unit) - fxr->file1_offset;
+	else if (fxr->file2_offset + length == size2)
+		blen = roundup_64(size2, alloc_unit) - fxr->file2_offset;
+	else if (!isaligned_64(length, alloc_unit))
+		return -EINVAL;
+	else
+		blen = length;
+
+	/* Don't allow overlapped exchanges within the same file. */
+	if (ip1 == ip2 &&
+	    fxr->file2_offset + blen > fxr->file1_offset &&
+	    fxr->file1_offset + blen > fxr->file2_offset)
+		return -EINVAL;
+
+	/*
+	 * Ensure that we don't exchange a partial EOF rt extent into the
+	 * middle of another file.
+	 */
+	if (isaligned_64(length, alloc_unit))
+		return 0;
+
+	blen = length;
+	if (fxr->file2_offset + length < size2)
+		blen = rounddown_64(blen, alloc_unit);
+
+	if (fxr->file1_offset + blen < size1)
+		blen = rounddown_64(blen, alloc_unit);
+
+	return blen == length ? 0 : -EINVAL;
+}
+
+/* Prepare two files to have their data exchanged. */
+STATIC int
+xfs_exchrange_prep(
+	struct xfs_exchrange	*fxr,
+	struct xfs_inode	*ip1,
+	struct xfs_inode	*ip2)
+{
+	struct xfs_mount	*mp = ip2->i_mount;
+	unsigned int		alloc_unit = xfs_inode_alloc_unitsize(ip2);
+	int			error;
+
+	trace_xfs_exchrange_prep(fxr, ip1, ip2);
+
+	/* Verify both files are either real-time or non-realtime */
+	if (XFS_IS_REALTIME_INODE(ip1) != XFS_IS_REALTIME_INODE(ip2))
+		return -EINVAL;
+
+	/* Check non-power of two alignment issues, if necessary. */
+	if (!is_power_of_2(alloc_unit)) {
+		error = xfs_exchrange_check_rtalign(fxr, ip1, ip2, alloc_unit);
+		if (error)
+			return error;
+
+		/*
+		 * Do the generic file-level checks with the regular block
+		 * alignment.
+		 */
+		alloc_unit = mp->m_sb.sb_blocksize;
+	}
+
+	error = xfs_exchange_range_prep(fxr, alloc_unit);
+	if (error || fxr->length == 0)
+		return error;
+
+	/* Attach dquots to both inodes before changing block maps. */
+	error = xfs_qm_dqattach(ip2);
+	if (error)
+		return error;
+	error = xfs_qm_dqattach(ip1);
+	if (error)
+		return error;
+
+	trace_xfs_exchrange_flush(fxr, ip1, ip2);
+
+	/* Flush the relevant ranges of both files. */
+	error = xfs_flush_unmap_range(ip2, fxr->file2_offset, fxr->length);
+	if (error)
+		return error;
+	error = xfs_flush_unmap_range(ip1, fxr->file1_offset, fxr->length);
+	if (error)
+		return error;
+
+	/*
+	 * Cancel CoW fork preallocations for the ranges of both files.  The
+	 * prep function should have flushed all the dirty data, so the only
+	 * CoW mappings remaining should be speculative.
+	 */
+	if (xfs_inode_has_cow_data(ip1)) {
+		error = xfs_reflink_cancel_cow_range(ip1, fxr->file1_offset,
+				fxr->length, true);
+		if (error)
+			return error;
+	}
+
+	if (xfs_inode_has_cow_data(ip2)) {
+		error = xfs_reflink_cancel_cow_range(ip2, fxr->file2_offset,
+				fxr->length, true);
+		if (error)
+			return error;
+	}
+
+	return 0;
+}
+
+/*
+ * Exchange contents of files.  This is the binding between the generic
+ * file-level concepts and the XFS inode-specific implementation.
+ */
+STATIC int
+xfs_exchrange_contents(
+	struct xfs_exchrange	*fxr)
+{
+	struct inode		*inode1 = file_inode(fxr->file1);
+	struct inode		*inode2 = file_inode(fxr->file2);
+	struct xfs_inode	*ip1 = XFS_I(inode1);
+	struct xfs_inode	*ip2 = XFS_I(inode2);
+	struct xfs_mount	*mp = ip1->i_mount;
+	int			error;
+
+	if (!xfs_has_exchange_range(mp))
+		return -EOPNOTSUPP;
+
+	if (fxr->flags & ~(XFS_EXCHANGE_RANGE_ALL_FLAGS |
+			   XFS_EXCHANGE_RANGE_PRIV_FLAGS))
+		return -EINVAL;
+
+	if (xfs_is_shutdown(mp))
+		return -EIO;
+
+	/* Lock both files against IO */
+	error = xfs_ilock2_io_mmap(ip1, ip2);
+	if (error)
+		goto out_err;
+
+	/* Prepare and then exchange file contents. */
+	error = xfs_exchrange_prep(fxr, ip1, ip2);
+	if (error)
+		goto out_unlock;
+
+	error = xfs_exchrange_mappings(fxr, ip1, ip2);
+	if (error)
+		goto out_unlock;
+
+	/*
+	 * Finish the exchange by removing special file privileges like any
+	 * other file write would do.  This may involve turning on support for
+	 * logged xattrs if either file has security capabilities.
+	 */
+	error = xfs_exchange_range_finish(fxr);
+	if (error)
+		goto out_unlock;
+
+out_unlock:
+	xfs_iunlock2_io_mmap(ip1, ip2);
+out_err:
+	if (error)
+		trace_xfs_exchrange_error(ip2, error, _RET_IP_);
+	return error;
+}
+
+/* Exchange parts of two files. */
+static int
+xfs_exchange_range(
+	struct xfs_exchrange	*fxr)
+{
+	struct inode		*inode1 = file_inode(fxr->file1);
+	struct inode		*inode2 = file_inode(fxr->file2);
+	int			ret;
+
+	BUILD_BUG_ON(XFS_EXCHANGE_RANGE_ALL_FLAGS &
+		     XFS_EXCHANGE_RANGE_PRIV_FLAGS);
+
+	/* Both files must be on the same mount/filesystem. */
+	if (fxr->file1->f_path.mnt != fxr->file2->f_path.mnt)
+		return -EXDEV;
+
+	if (fxr->flags & ~XFS_EXCHANGE_RANGE_ALL_FLAGS)
+		return -EINVAL;
+
+	/* Userspace requests only honored for regular files. */
+	if (S_ISDIR(inode1->i_mode) || S_ISDIR(inode2->i_mode))
+		return -EISDIR;
+	if (!S_ISREG(inode1->i_mode) || !S_ISREG(inode2->i_mode))
+		return -EINVAL;
+
+	/* Both files must be opened for read and write. */
+	if (!(fxr->file1->f_mode & FMODE_READ) ||
+	    !(fxr->file1->f_mode & FMODE_WRITE) ||
+	    !(fxr->file2->f_mode & FMODE_READ) ||
+	    !(fxr->file2->f_mode & FMODE_WRITE))
+		return -EBADF;
+
+	/* Neither file can be opened append-only. */
+	if ((fxr->file1->f_flags & O_APPEND) ||
+	    (fxr->file2->f_flags & O_APPEND))
+		return -EBADF;
+
+	/*
+	 * If we're not exchanging to EOF, we can check the areas before
+	 * stabilizing both files' i_size.
+	 */
+	if (!(fxr->flags & XFS_EXCHANGE_RANGE_TO_EOF)) {
+		ret = xfs_exchange_range_verify_area(fxr);
+		if (ret)
+			return ret;
+	}
+
+	/* Update cmtime if the fd/inode don't forbid it. */
+	if (!(fxr->file1->f_mode & FMODE_NOCMTIME) && !IS_NOCMTIME(inode1))
+		fxr->flags |= __XFS_EXCHANGE_RANGE_UPD_CMTIME1;
+	if (!(fxr->file2->f_mode & FMODE_NOCMTIME) && !IS_NOCMTIME(inode2))
+		fxr->flags |= __XFS_EXCHANGE_RANGE_UPD_CMTIME2;
+
+	file_start_write(fxr->file2);
+	ret = xfs_exchrange_contents(fxr);
+	file_end_write(fxr->file2);
+	if (ret)
+		return ret;
+
+	fsnotify_modify(fxr->file1);
+	if (fxr->file2 != fxr->file1)
+		fsnotify_modify(fxr->file2);
+	return 0;
+}
+
+/* Collect exchange-range arguments from userspace. */
+long
+xfs_ioc_exchange_range(
+	struct file			*file,
+	struct xfs_exchange_range __user *argp)
+{
+	struct xfs_exchrange		fxr = {
+		.file2			= file,
+	};
+	struct xfs_exchange_range	args;
+	struct fd			file1;
+	int				error;
+
+	if (copy_from_user(&args, argp, sizeof(args)))
+		return -EFAULT;
+	if (memchr_inv(&args.pad, 0, sizeof(args.pad)))
+		return -EINVAL;
+	if (args.flags & ~XFS_EXCHANGE_RANGE_ALL_FLAGS)
+		return -EINVAL;
+
+	fxr.file1_offset	= args.file1_offset;
+	fxr.file2_offset	= args.file2_offset;
+	fxr.length		= args.length;
+	fxr.flags		= args.flags;
+
+	file1 = fdget(args.file1_fd);
+	if (!file1.file)
+		return -EBADF;
+	fxr.file1 = file1.file;
+
+	error = xfs_exchange_range(&fxr);
+	fdput(file1);
+	return error;
+}
diff --git a/fs/xfs/xfs_exchrange.h b/fs/xfs/xfs_exchrange.h
new file mode 100644
index 0000000000..039abcca54
--- /dev/null
+++ b/fs/xfs/xfs_exchrange.h
@@ -0,0 +1,38 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (c) 2020-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_EXCHRANGE_H__
+#define __XFS_EXCHRANGE_H__
+
+/* Update the mtime/cmtime of file1 and file2 */
+#define __XFS_EXCHANGE_RANGE_UPD_CMTIME1	(1ULL << 63)
+#define __XFS_EXCHANGE_RANGE_UPD_CMTIME2	(1ULL << 62)
+
+#define XFS_EXCHANGE_RANGE_PRIV_FLAGS	(__XFS_EXCHANGE_RANGE_UPD_CMTIME1 | \
+					 __XFS_EXCHANGE_RANGE_UPD_CMTIME2)
+
+struct xfs_exchrange {
+	struct file		*file1;
+	struct file		*file2;
+
+	loff_t			file1_offset;
+	loff_t			file2_offset;
+	u64			length;
+
+	u64			flags;	/* XFS_EXCHANGE_RANGE flags */
+};
+
+long xfs_ioc_exchange_range(struct file *file,
+		struct xfs_exchange_range __user *argp);
+
+struct xfs_exchmaps_req;
+
+void xfs_exchrange_ilock(struct xfs_trans *tp, struct xfs_inode *ip1,
+		struct xfs_inode *ip2);
+void xfs_exchrange_iunlock(struct xfs_inode *ip1, struct xfs_inode *ip2);
+
+int xfs_exchrange_estimate(struct xfs_exchmaps_req *req);
+
+#endif /* __XFS_EXCHRANGE_H__ */
diff --git a/fs/xfs/xfs_export.c b/fs/xfs/xfs_export.c
index 7cd09c3a82..201489d3de 100644
--- a/fs/xfs/xfs_export.c
+++ b/fs/xfs/xfs_export.c
@@ -102,7 +102,7 @@ xfs_fs_encode_fh(
 	return fileid_type;
 }
 
-STATIC struct inode *
+struct inode *
 xfs_nfs_get_inode(
 	struct super_block	*sb,
 	u64			ino,
@@ -160,7 +160,7 @@ xfs_nfs_get_inode(
 		}
 	}
 
-	if (VFS_I(ip)->i_generation != generation) {
+	if (VFS_I(ip)->i_generation != generation || IS_PRIVATE(VFS_I(ip))) {
 		xfs_irele(ip);
 		return ERR_PTR(-ESTALE);
 	}
diff --git a/fs/xfs/xfs_export.h b/fs/xfs/xfs_export.h
index 64471a3ddb..3cd85e8901 100644
--- a/fs/xfs/xfs_export.h
+++ b/fs/xfs/xfs_export.h
@@ -57,4 +57,6 @@ struct xfs_fid64 {
 /* This flag goes on the wire.  Don't play with it. */
 #define XFS_FILEID_TYPE_64FLAG	0x80	/* NFS fileid has 64bit inodes */
 
+struct inode *xfs_nfs_get_inode(struct super_block *sb, u64 ino, u32 gen);
+
 #endif	/* __XFS_EXPORT_H__ */
diff --git a/fs/xfs/xfs_extent_busy.c b/fs/xfs/xfs_extent_busy.c
index 2ccde32c9a..a73e7c73b6 100644
--- a/fs/xfs/xfs_extent_busy.c
+++ b/fs/xfs/xfs_extent_busy.c
@@ -32,7 +32,8 @@ xfs_extent_busy_insert_list(
 	struct rb_node		**rbp;
 	struct rb_node		*parent = NULL;
 
-	new = kmem_zalloc(sizeof(struct xfs_extent_busy), 0);
+	new = kzalloc(sizeof(struct xfs_extent_busy),
+			GFP_KERNEL | __GFP_NOFAIL);
 	new->agno = pag->pag_agno;
 	new->bno = bno;
 	new->length = len;
@@ -517,35 +518,26 @@ fail:
 	goto out;
 }
 
-STATIC void
+static bool
 xfs_extent_busy_clear_one(
-	struct xfs_mount	*mp,
 	struct xfs_perag	*pag,
-	struct xfs_extent_busy	*busyp)
+	struct xfs_extent_busy	*busyp,
+	bool			do_discard)
 {
 	if (busyp->length) {
-		trace_xfs_extent_busy_clear(mp, busyp->agno, busyp->bno,
-						busyp->length);
+		if (do_discard &&
+		    !(busyp->flags & XFS_EXTENT_BUSY_SKIP_DISCARD)) {
+			busyp->flags = XFS_EXTENT_BUSY_DISCARDED;
+			return false;
+		}
+		trace_xfs_extent_busy_clear(pag->pag_mount, busyp->agno,
+				busyp->bno, busyp->length);
 		rb_erase(&busyp->rb_node, &pag->pagb_tree);
 	}
 
 	list_del_init(&busyp->list);
-	kmem_free(busyp);
-}
-
-static void
-xfs_extent_busy_put_pag(
-	struct xfs_perag	*pag,
-	bool			wakeup)
-		__releases(pag->pagb_lock)
-{
-	if (wakeup) {
-		pag->pagb_gen++;
-		wake_up_all(&pag->pagb_wait);
-	}
-
-	spin_unlock(&pag->pagb_lock);
-	xfs_perag_put(pag);
+	kfree(busyp);
+	return true;
 }
 
 /*
@@ -559,32 +551,33 @@ xfs_extent_busy_clear(
 	struct list_head	*list,
 	bool			do_discard)
 {
-	struct xfs_extent_busy	*busyp, *n;
-	struct xfs_perag	*pag = NULL;
-	xfs_agnumber_t		agno = NULLAGNUMBER;
-	bool			wakeup = false;
-
-	list_for_each_entry_safe(busyp, n, list, list) {
-		if (busyp->agno != agno) {
-			if (pag)
-				xfs_extent_busy_put_pag(pag, wakeup);
-			agno = busyp->agno;
-			pag = xfs_perag_get(mp, agno);
-			spin_lock(&pag->pagb_lock);
-			wakeup = false;
-		}
+	struct xfs_extent_busy	*busyp, *next;
 
-		if (do_discard && busyp->length &&
-		    !(busyp->flags & XFS_EXTENT_BUSY_SKIP_DISCARD)) {
-			busyp->flags = XFS_EXTENT_BUSY_DISCARDED;
-		} else {
-			xfs_extent_busy_clear_one(mp, pag, busyp);
-			wakeup = true;
-		}
-	}
+	busyp = list_first_entry_or_null(list, typeof(*busyp), list);
+	if (!busyp)
+		return;
 
-	if (pag)
-		xfs_extent_busy_put_pag(pag, wakeup);
+	do {
+		bool			wakeup = false;
+		struct xfs_perag	*pag;
+
+		pag = xfs_perag_get(mp, busyp->agno);
+		spin_lock(&pag->pagb_lock);
+		do {
+			next = list_next_entry(busyp, list);
+			if (xfs_extent_busy_clear_one(pag, busyp, do_discard))
+				wakeup = true;
+			busyp = next;
+		} while (!list_entry_is_head(busyp, list, list) &&
+			 busyp->agno == pag->pag_agno);
+
+		if (wakeup) {
+			pag->pagb_gen++;
+			wake_up_all(&pag->pagb_wait);
+		}
+		spin_unlock(&pag->pagb_lock);
+		xfs_perag_put(pag);
+	} while (!list_entry_is_head(busyp, list, list));
 }
 
 /*
diff --git a/fs/xfs/xfs_extfree_item.c b/fs/xfs/xfs_extfree_item.c
index 1d1185fca6..8c382f0923 100644
--- a/fs/xfs/xfs_extfree_item.c
+++ b/fs/xfs/xfs_extfree_item.c
@@ -40,9 +40,9 @@ STATIC void
 xfs_efi_item_free(
 	struct xfs_efi_log_item	*efip)
 {
-	kmem_free(efip->efi_item.li_lv_shadow);
+	kvfree(efip->efi_item.li_lv_shadow);
 	if (efip->efi_format.efi_nextents > XFS_EFI_MAX_FAST_EXTENTS)
-		kmem_free(efip);
+		kfree(efip);
 	else
 		kmem_cache_free(xfs_efi_cache, efip);
 }
@@ -229,9 +229,9 @@ static inline struct xfs_efd_log_item *EFD_ITEM(struct xfs_log_item *lip)
 STATIC void
 xfs_efd_item_free(struct xfs_efd_log_item *efdp)
 {
-	kmem_free(efdp->efd_item.li_lv_shadow);
+	kvfree(efdp->efd_item.li_lv_shadow);
 	if (efdp->efd_format.efd_nextents > XFS_EFD_MAX_FAST_EXTENTS)
-		kmem_free(efdp);
+		kfree(efdp);
 	else
 		kmem_cache_free(xfs_efd_cache, efdp);
 }
diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
index e33e5e13b9..b240ea5241 100644
--- a/fs/xfs/xfs_file.c
+++ b/fs/xfs/xfs_file.c
@@ -24,6 +24,7 @@
 #include "xfs_pnfs.h"
 #include "xfs_iomap.h"
 #include "xfs_reflink.h"
+#include "xfs_file.h"
 
 #include <linux/dax.h>
 #include <linux/falloc.h>
@@ -38,33 +39,19 @@ static const struct vm_operations_struct xfs_file_vm_ops;
  * Decide if the given file range is aligned to the size of the fundamental
  * allocation unit for the file.
  */
-static bool
+bool
 xfs_is_falloc_aligned(
 	struct xfs_inode	*ip,
 	loff_t			pos,
 	long long int		len)
 {
-	struct xfs_mount	*mp = ip->i_mount;
-	uint64_t		mask;
-
-	if (XFS_IS_REALTIME_INODE(ip)) {
-		if (!is_power_of_2(mp->m_sb.sb_rextsize)) {
-			u64	rextbytes;
-			u32	mod;
-
-			rextbytes = XFS_FSB_TO_B(mp, mp->m_sb.sb_rextsize);
-			div_u64_rem(pos, rextbytes, &mod);
-			if (mod)
-				return false;
-			div_u64_rem(len, rextbytes, &mod);
-			return mod == 0;
-		}
-		mask = XFS_FSB_TO_B(mp, mp->m_sb.sb_rextsize) - 1;
-	} else {
-		mask = mp->m_sb.sb_blocksize - 1;
-	}
+	unsigned int		alloc_unit = xfs_inode_alloc_unitsize(ip);
 
-	return !((pos | len) & mask);
+	if (!is_power_of_2(alloc_unit))
+		return isaligned_64(pos, alloc_unit) &&
+		       isaligned_64(len, alloc_unit);
+
+	return !((pos | len) & (alloc_unit - 1));
 }
 
 /*
@@ -861,67 +848,6 @@ xfs_file_write_iter(
 	return xfs_file_buffered_write(iocb, from);
 }
 
-static void
-xfs_wait_dax_page(
-	struct inode		*inode)
-{
-	struct xfs_inode        *ip = XFS_I(inode);
-
-	xfs_iunlock(ip, XFS_MMAPLOCK_EXCL);
-	schedule();
-	xfs_ilock(ip, XFS_MMAPLOCK_EXCL);
-}
-
-int
-xfs_break_dax_layouts(
-	struct inode		*inode,
-	bool			*retry)
-{
-	struct page		*page;
-
-	ASSERT(xfs_isilocked(XFS_I(inode), XFS_MMAPLOCK_EXCL));
-
-	page = dax_layout_busy_page(inode->i_mapping);
-	if (!page)
-		return 0;
-
-	*retry = true;
-	return ___wait_var_event(&page->_refcount,
-			atomic_read(&page->_refcount) == 1, TASK_INTERRUPTIBLE,
-			0, 0, xfs_wait_dax_page(inode));
-}
-
-int
-xfs_break_layouts(
-	struct inode		*inode,
-	uint			*iolock,
-	enum layout_break_reason reason)
-{
-	bool			retry;
-	int			error;
-
-	ASSERT(xfs_isilocked(XFS_I(inode), XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL));
-
-	do {
-		retry = false;
-		switch (reason) {
-		case BREAK_UNMAP:
-			error = xfs_break_dax_layouts(inode, &retry);
-			if (error || retry)
-				break;
-			fallthrough;
-		case BREAK_WRITE:
-			error = xfs_break_leased_layouts(inode, iolock, &retry);
-			break;
-		default:
-			WARN_ON_ONCE(1);
-			error = -EINVAL;
-		}
-	} while (error == 0 && retry);
-
-	return error;
-}
-
 /* Does this file, inode, or mount want synchronous writes? */
 static inline bool xfs_file_sync_writes(struct file *filp)
 {
@@ -1230,8 +1156,7 @@ xfs_file_open(
 {
 	if (xfs_is_shutdown(XFS_M(inode->i_sb)))
 		return -EIO;
-	file->f_mode |= FMODE_NOWAIT | FMODE_BUF_RASYNC | FMODE_BUF_WASYNC |
-			FMODE_DIO_PARALLEL_WRITE | FMODE_CAN_ODIRECT;
+	file->f_mode |= FMODE_NOWAIT | FMODE_CAN_ODIRECT;
 	return generic_file_open(inode, file);
 }
 
@@ -1244,7 +1169,9 @@ xfs_dir_open(
 	unsigned int	mode;
 	int		error;
 
-	error = xfs_file_open(inode, file);
+	if (xfs_is_shutdown(ip->i_mount))
+		return -EIO;
+	error = generic_file_open(inode, file);
 	if (error)
 		return error;
 
@@ -1490,7 +1417,6 @@ const struct file_operations xfs_file_operations = {
 	.compat_ioctl	= xfs_file_compat_ioctl,
 #endif
 	.mmap		= xfs_file_mmap,
-	.mmap_supported_flags = MAP_SYNC,
 	.open		= xfs_file_open,
 	.release	= xfs_file_release,
 	.fsync		= xfs_file_fsync,
@@ -1498,6 +1424,8 @@ const struct file_operations xfs_file_operations = {
 	.fallocate	= xfs_file_fallocate,
 	.fadvise	= xfs_file_fadvise,
 	.remap_file_range = xfs_file_remap_range,
+	.fop_flags	= FOP_MMAP_SYNC | FOP_BUFFER_RASYNC |
+			  FOP_BUFFER_WASYNC | FOP_DIO_PARALLEL_WRITE,
 };
 
 const struct file_operations xfs_dir_file_operations = {
diff --git a/fs/xfs/xfs_file.h b/fs/xfs/xfs_file.h
new file mode 100644
index 0000000000..2ad91f755c
--- /dev/null
+++ b/fs/xfs/xfs_file.h
@@ -0,0 +1,15 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ */
+#ifndef __XFS_FILE_H__
+#define __XFS_FILE_H__
+
+extern const struct file_operations xfs_file_operations;
+extern const struct file_operations xfs_dir_file_operations;
+
+bool xfs_is_falloc_aligned(struct xfs_inode *ip, loff_t pos,
+		long long int len);
+
+#endif /* __XFS_FILE_H__ */
diff --git a/fs/xfs/xfs_filestream.c b/fs/xfs/xfs_filestream.c
index 2fc98d3137..e3aaa05555 100644
--- a/fs/xfs/xfs_filestream.c
+++ b/fs/xfs/xfs_filestream.c
@@ -44,7 +44,7 @@ xfs_fstrm_free_func(
 	atomic_dec(&pag->pagf_fstrms);
 	xfs_perag_rele(pag);
 
-	kmem_free(item);
+	kfree(item);
 }
 
 /*
@@ -313,7 +313,7 @@ xfs_filestream_create_association(
 	 * we return a referenced AG, the allocation can still go ahead just
 	 * fine.
 	 */
-	item = kmem_alloc(sizeof(*item), KM_MAYFAIL);
+	item = kmalloc(sizeof(*item), GFP_KERNEL | __GFP_RETRY_MAYFAIL);
 	if (!item)
 		goto out_put_fstrms;
 
@@ -326,7 +326,7 @@ xfs_filestream_create_association(
 
 out_free_item:
 	xfs_perag_rele(item->pag);
-	kmem_free(item);
+	kfree(item);
 out_put_fstrms:
 	atomic_dec(&args->pag->pagf_fstrms);
 	return 0;
diff --git a/fs/xfs/xfs_fsmap.c b/fs/xfs/xfs_fsmap.c
index 5a72217f5f..85dbb46452 100644
--- a/fs/xfs/xfs_fsmap.c
+++ b/fs/xfs/xfs_fsmap.c
@@ -533,7 +533,7 @@ xfs_getfsmap_rtdev_rtbitmap(
 	trace_xfs_fsmap_low_key_linear(mp, info->dev, start_rtb);
 	trace_xfs_fsmap_high_key_linear(mp, info->dev, end_rtb);
 
-	xfs_ilock(mp->m_rbmip, XFS_ILOCK_SHARED | XFS_ILOCK_RTBITMAP);
+	xfs_rtbitmap_lock_shared(mp, XFS_RBMLOCK_BITMAP);
 
 	/*
 	 * Set up query parameters to return free rtextents covering the range
@@ -557,7 +557,7 @@ xfs_getfsmap_rtdev_rtbitmap(
 	if (error)
 		goto err;
 err:
-	xfs_iunlock(mp->m_rbmip, XFS_ILOCK_SHARED | XFS_ILOCK_RTBITMAP);
+	xfs_rtbitmap_unlock_shared(mp, XFS_RBMLOCK_BITMAP);
 	return error;
 }
 #endif /* CONFIG_XFS_RT */
@@ -763,8 +763,8 @@ xfs_getfsmap_datadev_bnobt_query(
 		return xfs_getfsmap_datadev_bnobt_helper(*curpp, &key[1], info);
 
 	/* Allocate cursor for this AG and query_range it. */
-	*curpp = xfs_allocbt_init_cursor(tp->t_mountp, tp, info->agf_bp,
-			info->pag, XFS_BTNUM_BNO);
+	*curpp = xfs_bnobt_init_cursor(tp->t_mountp, tp, info->agf_bp,
+			info->pag);
 	key->ar_startblock = info->low.rm_startblock;
 	key[1].ar_startblock = info->high.rm_startblock;
 	return xfs_alloc_query_range(*curpp, key, &key[1],
diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c
index 83f708f62e..c211ea2b63 100644
--- a/fs/xfs/xfs_fsops.c
+++ b/fs/xfs/xfs_fsops.c
@@ -213,10 +213,8 @@ xfs_growfs_data_private(
 			struct xfs_perag	*pag;
 
 			pag = xfs_perag_get(mp, id.agno);
-			error = xfs_ag_resv_free(pag);
+			xfs_ag_resv_free(pag);
 			xfs_perag_put(pag);
-			if (error)
-				return error;
 		}
 		/*
 		 * Reserve AG metadata blocks. ENOSPC here does not mean there
@@ -385,14 +383,14 @@ xfs_reserve_blocks(
 	 */
 	if (mp->m_resblks > request) {
 		lcounter = mp->m_resblks_avail - request;
-		if (lcounter  > 0) {		/* release unused blocks */
+		if (lcounter > 0) {		/* release unused blocks */
 			fdblks_delta = lcounter;
 			mp->m_resblks_avail -= lcounter;
 		}
 		mp->m_resblks = request;
 		if (fdblks_delta) {
 			spin_unlock(&mp->m_sb_lock);
-			error = xfs_mod_fdblocks(mp, fdblks_delta, 0);
+			xfs_add_fdblocks(mp, fdblks_delta);
 			spin_lock(&mp->m_sb_lock);
 		}
 
@@ -428,9 +426,9 @@ xfs_reserve_blocks(
 		 */
 		fdblks_delta = min(free, delta);
 		spin_unlock(&mp->m_sb_lock);
-		error = xfs_mod_fdblocks(mp, -fdblks_delta, 0);
+		error = xfs_dec_fdblocks(mp, fdblks_delta, 0);
 		if (!error)
-			xfs_mod_fdblocks(mp, fdblks_delta, 0);
+			xfs_add_fdblocks(mp, fdblks_delta);
 		spin_lock(&mp->m_sb_lock);
 	}
 out:
@@ -556,24 +554,13 @@ xfs_fs_reserve_ag_blocks(
 /*
  * Free space reserved for per-AG metadata.
  */
-int
+void
 xfs_fs_unreserve_ag_blocks(
 	struct xfs_mount	*mp)
 {
 	xfs_agnumber_t		agno;
 	struct xfs_perag	*pag;
-	int			error = 0;
-	int			err2;
 
-	for_each_perag(mp, agno, pag) {
-		err2 = xfs_ag_resv_free(pag);
-		if (err2 && !error)
-			error = err2;
-	}
-
-	if (error)
-		xfs_warn(mp,
-	"Error %d freeing per-AG metadata reserve pool.", error);
-
-	return error;
+	for_each_perag(mp, agno, pag)
+		xfs_ag_resv_free(pag);
 }
diff --git a/fs/xfs/xfs_fsops.h b/fs/xfs/xfs_fsops.h
index 44457b0a05..3e2f73bcf8 100644
--- a/fs/xfs/xfs_fsops.h
+++ b/fs/xfs/xfs_fsops.h
@@ -12,6 +12,6 @@ int xfs_reserve_blocks(struct xfs_mount *mp, uint64_t request);
 int xfs_fs_goingdown(struct xfs_mount *mp, uint32_t inflags);
 
 int xfs_fs_reserve_ag_blocks(struct xfs_mount *mp);
-int xfs_fs_unreserve_ag_blocks(struct xfs_mount *mp);
+void xfs_fs_unreserve_ag_blocks(struct xfs_mount *mp);
 
 #endif	/* __XFS_FSOPS_H__ */
diff --git a/fs/xfs/xfs_handle.c b/fs/xfs/xfs_handle.c
new file mode 100644
index 0000000000..a3f16e9b6f
--- /dev/null
+++ b/fs/xfs/xfs_handle.c
@@ -0,0 +1,947 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * Copyright (c) 2022-2024 Oracle.
+ * All rights reserved.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_shared.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_inode.h"
+#include "xfs_error.h"
+#include "xfs_trace.h"
+#include "xfs_trans.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
+#include "xfs_attr.h"
+#include "xfs_ioctl.h"
+#include "xfs_parent.h"
+#include "xfs_da_btree.h"
+#include "xfs_handle.h"
+#include "xfs_health.h"
+#include "xfs_icache.h"
+#include "xfs_export.h"
+#include "xfs_xattr.h"
+#include "xfs_acl.h"
+
+#include <linux/namei.h>
+
+static inline size_t
+xfs_filehandle_fid_len(void)
+{
+	struct xfs_handle	*handle = NULL;
+
+	return sizeof(struct xfs_fid) - sizeof(handle->ha_fid.fid_len);
+}
+
+static inline size_t
+xfs_filehandle_init(
+	struct xfs_mount	*mp,
+	xfs_ino_t		ino,
+	uint32_t		gen,
+	struct xfs_handle	*handle)
+{
+	memcpy(&handle->ha_fsid, mp->m_fixedfsid, sizeof(struct xfs_fsid));
+
+	handle->ha_fid.fid_len = xfs_filehandle_fid_len();
+	handle->ha_fid.fid_pad = 0;
+	handle->ha_fid.fid_gen = gen;
+	handle->ha_fid.fid_ino = ino;
+
+	return sizeof(struct xfs_handle);
+}
+
+static inline size_t
+xfs_fshandle_init(
+	struct xfs_mount	*mp,
+	struct xfs_handle	*handle)
+{
+	memcpy(&handle->ha_fsid, mp->m_fixedfsid, sizeof(struct xfs_fsid));
+	memset(&handle->ha_fid, 0, sizeof(handle->ha_fid));
+
+	return sizeof(struct xfs_fsid);
+}
+
+/*
+ * xfs_find_handle maps from userspace xfs_fsop_handlereq structure to
+ * a file or fs handle.
+ *
+ * XFS_IOC_PATH_TO_FSHANDLE
+ *    returns fs handle for a mount point or path within that mount point
+ * XFS_IOC_FD_TO_HANDLE
+ *    returns full handle for a FD opened in user space
+ * XFS_IOC_PATH_TO_HANDLE
+ *    returns full handle for a path
+ */
+int
+xfs_find_handle(
+	unsigned int		cmd,
+	xfs_fsop_handlereq_t	*hreq)
+{
+	int			hsize;
+	xfs_handle_t		handle;
+	struct inode		*inode;
+	struct fd		f = {NULL};
+	struct path		path;
+	int			error;
+	struct xfs_inode	*ip;
+
+	if (cmd == XFS_IOC_FD_TO_HANDLE) {
+		f = fdget(hreq->fd);
+		if (!f.file)
+			return -EBADF;
+		inode = file_inode(f.file);
+	} else {
+		error = user_path_at(AT_FDCWD, hreq->path, 0, &path);
+		if (error)
+			return error;
+		inode = d_inode(path.dentry);
+	}
+	ip = XFS_I(inode);
+
+	/*
+	 * We can only generate handles for inodes residing on a XFS filesystem,
+	 * and only for regular files, directories or symbolic links.
+	 */
+	error = -EINVAL;
+	if (inode->i_sb->s_magic != XFS_SB_MAGIC)
+		goto out_put;
+
+	error = -EBADF;
+	if (!S_ISREG(inode->i_mode) &&
+	    !S_ISDIR(inode->i_mode) &&
+	    !S_ISLNK(inode->i_mode))
+		goto out_put;
+
+
+	memcpy(&handle.ha_fsid, ip->i_mount->m_fixedfsid, sizeof(xfs_fsid_t));
+
+	if (cmd == XFS_IOC_PATH_TO_FSHANDLE)
+		hsize = xfs_fshandle_init(ip->i_mount, &handle);
+	else
+		hsize = xfs_filehandle_init(ip->i_mount, ip->i_ino,
+				inode->i_generation, &handle);
+
+	error = -EFAULT;
+	if (copy_to_user(hreq->ohandle, &handle, hsize) ||
+	    copy_to_user(hreq->ohandlen, &hsize, sizeof(__s32)))
+		goto out_put;
+
+	error = 0;
+
+ out_put:
+	if (cmd == XFS_IOC_FD_TO_HANDLE)
+		fdput(f);
+	else
+		path_put(&path);
+	return error;
+}
+
+/*
+ * No need to do permission checks on the various pathname components
+ * as the handle operations are privileged.
+ */
+STATIC int
+xfs_handle_acceptable(
+	void			*context,
+	struct dentry		*dentry)
+{
+	return 1;
+}
+
+/* Convert handle already copied to kernel space into a dentry. */
+static struct dentry *
+xfs_khandle_to_dentry(
+	struct file		*file,
+	struct xfs_handle	*handle)
+{
+	struct xfs_fid64        fid = {
+		.ino		= handle->ha_fid.fid_ino,
+		.gen		= handle->ha_fid.fid_gen,
+	};
+
+	/*
+	 * Only allow handle opens under a directory.
+	 */
+	if (!S_ISDIR(file_inode(file)->i_mode))
+		return ERR_PTR(-ENOTDIR);
+
+	if (handle->ha_fid.fid_len != xfs_filehandle_fid_len())
+		return ERR_PTR(-EINVAL);
+
+	return exportfs_decode_fh(file->f_path.mnt, (struct fid *)&fid, 3,
+			FILEID_INO32_GEN | XFS_FILEID_TYPE_64FLAG,
+			xfs_handle_acceptable, NULL);
+}
+
+/* Convert handle already copied to kernel space into an xfs_inode. */
+static struct xfs_inode *
+xfs_khandle_to_inode(
+	struct file		*file,
+	struct xfs_handle	*handle)
+{
+	struct xfs_inode	*ip = XFS_I(file_inode(file));
+	struct xfs_mount	*mp = ip->i_mount;
+	struct inode		*inode;
+
+	if (!S_ISDIR(VFS_I(ip)->i_mode))
+		return ERR_PTR(-ENOTDIR);
+
+	if (handle->ha_fid.fid_len != xfs_filehandle_fid_len())
+		return ERR_PTR(-EINVAL);
+
+	inode = xfs_nfs_get_inode(mp->m_super, handle->ha_fid.fid_ino,
+			handle->ha_fid.fid_gen);
+	if (IS_ERR(inode))
+		return ERR_CAST(inode);
+
+	return XFS_I(inode);
+}
+
+/*
+ * Convert userspace handle data into a dentry.
+ */
+struct dentry *
+xfs_handle_to_dentry(
+	struct file		*parfilp,
+	void __user		*uhandle,
+	u32			hlen)
+{
+	xfs_handle_t		handle;
+
+	if (hlen != sizeof(xfs_handle_t))
+		return ERR_PTR(-EINVAL);
+	if (copy_from_user(&handle, uhandle, hlen))
+		return ERR_PTR(-EFAULT);
+
+	return xfs_khandle_to_dentry(parfilp, &handle);
+}
+
+STATIC struct dentry *
+xfs_handlereq_to_dentry(
+	struct file		*parfilp,
+	xfs_fsop_handlereq_t	*hreq)
+{
+	return xfs_handle_to_dentry(parfilp, hreq->ihandle, hreq->ihandlen);
+}
+
+int
+xfs_open_by_handle(
+	struct file		*parfilp,
+	xfs_fsop_handlereq_t	*hreq)
+{
+	const struct cred	*cred = current_cred();
+	int			error;
+	int			fd;
+	int			permflag;
+	struct file		*filp;
+	struct inode		*inode;
+	struct dentry		*dentry;
+	fmode_t			fmode;
+	struct path		path;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	dentry = xfs_handlereq_to_dentry(parfilp, hreq);
+	if (IS_ERR(dentry))
+		return PTR_ERR(dentry);
+	inode = d_inode(dentry);
+
+	/* Restrict xfs_open_by_handle to directories & regular files. */
+	if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode))) {
+		error = -EPERM;
+		goto out_dput;
+	}
+
+#if BITS_PER_LONG != 32
+	hreq->oflags |= O_LARGEFILE;
+#endif
+
+	permflag = hreq->oflags;
+	fmode = OPEN_FMODE(permflag);
+	if ((!(permflag & O_APPEND) || (permflag & O_TRUNC)) &&
+	    (fmode & FMODE_WRITE) && IS_APPEND(inode)) {
+		error = -EPERM;
+		goto out_dput;
+	}
+
+	if ((fmode & FMODE_WRITE) && IS_IMMUTABLE(inode)) {
+		error = -EPERM;
+		goto out_dput;
+	}
+
+	/* Can't write directories. */
+	if (S_ISDIR(inode->i_mode) && (fmode & FMODE_WRITE)) {
+		error = -EISDIR;
+		goto out_dput;
+	}
+
+	fd = get_unused_fd_flags(0);
+	if (fd < 0) {
+		error = fd;
+		goto out_dput;
+	}
+
+	path.mnt = parfilp->f_path.mnt;
+	path.dentry = dentry;
+	filp = dentry_open(&path, hreq->oflags, cred);
+	dput(dentry);
+	if (IS_ERR(filp)) {
+		put_unused_fd(fd);
+		return PTR_ERR(filp);
+	}
+
+	if (S_ISREG(inode->i_mode)) {
+		filp->f_flags |= O_NOATIME;
+		filp->f_mode |= FMODE_NOCMTIME;
+	}
+
+	fd_install(fd, filp);
+	return fd;
+
+ out_dput:
+	dput(dentry);
+	return error;
+}
+
+int
+xfs_readlink_by_handle(
+	struct file		*parfilp,
+	xfs_fsop_handlereq_t	*hreq)
+{
+	struct dentry		*dentry;
+	__u32			olen;
+	int			error;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	dentry = xfs_handlereq_to_dentry(parfilp, hreq);
+	if (IS_ERR(dentry))
+		return PTR_ERR(dentry);
+
+	/* Restrict this handle operation to symlinks only. */
+	if (!d_is_symlink(dentry)) {
+		error = -EINVAL;
+		goto out_dput;
+	}
+
+	if (copy_from_user(&olen, hreq->ohandlen, sizeof(__u32))) {
+		error = -EFAULT;
+		goto out_dput;
+	}
+
+	error = vfs_readlink(dentry, hreq->ohandle, olen);
+
+ out_dput:
+	dput(dentry);
+	return error;
+}
+
+/*
+ * Format an attribute and copy it out to the user's buffer.
+ * Take care to check values and protect against them changing later,
+ * we may be reading them directly out of a user buffer.
+ */
+static void
+xfs_ioc_attr_put_listent(
+	struct xfs_attr_list_context *context,
+	int			flags,
+	unsigned char		*name,
+	int			namelen,
+	void			*value,
+	int			valuelen)
+{
+	struct xfs_attrlist	*alist = context->buffer;
+	struct xfs_attrlist_ent	*aep;
+	int			arraytop;
+
+	ASSERT(!context->seen_enough);
+	ASSERT(context->count >= 0);
+	ASSERT(context->count < (ATTR_MAX_VALUELEN/8));
+	ASSERT(context->firstu >= sizeof(*alist));
+	ASSERT(context->firstu <= context->bufsize);
+
+	/*
+	 * Only list entries in the right namespace.
+	 */
+	if (context->attr_filter != (flags & XFS_ATTR_NSP_ONDISK_MASK))
+		return;
+
+	arraytop = sizeof(*alist) +
+			context->count * sizeof(alist->al_offset[0]);
+
+	/* decrement by the actual bytes used by the attr */
+	context->firstu -= round_up(offsetof(struct xfs_attrlist_ent, a_name) +
+			namelen + 1, sizeof(uint32_t));
+	if (context->firstu < arraytop) {
+		trace_xfs_attr_list_full(context);
+		alist->al_more = 1;
+		context->seen_enough = 1;
+		return;
+	}
+
+	aep = context->buffer + context->firstu;
+	aep->a_valuelen = valuelen;
+	memcpy(aep->a_name, name, namelen);
+	aep->a_name[namelen] = 0;
+	alist->al_offset[context->count++] = context->firstu;
+	alist->al_count = context->count;
+	trace_xfs_attr_list_add(context);
+}
+
+static unsigned int
+xfs_attr_filter(
+	u32			ioc_flags)
+{
+	if (ioc_flags & XFS_IOC_ATTR_ROOT)
+		return XFS_ATTR_ROOT;
+	if (ioc_flags & XFS_IOC_ATTR_SECURE)
+		return XFS_ATTR_SECURE;
+	return 0;
+}
+
+static inline enum xfs_attr_update
+xfs_xattr_flags(
+	u32			ioc_flags,
+	void			*value)
+{
+	if (!value)
+		return XFS_ATTRUPDATE_REMOVE;
+	if (ioc_flags & XFS_IOC_ATTR_CREATE)
+		return XFS_ATTRUPDATE_CREATE;
+	if (ioc_flags & XFS_IOC_ATTR_REPLACE)
+		return XFS_ATTRUPDATE_REPLACE;
+	return XFS_ATTRUPDATE_UPSERT;
+}
+
+int
+xfs_ioc_attr_list(
+	struct xfs_inode		*dp,
+	void __user			*ubuf,
+	size_t				bufsize,
+	int				flags,
+	struct xfs_attrlist_cursor __user *ucursor)
+{
+	struct xfs_attr_list_context	context = { };
+	struct xfs_attrlist		*alist;
+	void				*buffer;
+	int				error;
+
+	if (bufsize < sizeof(struct xfs_attrlist) ||
+	    bufsize > XFS_XATTR_LIST_MAX)
+		return -EINVAL;
+
+	/*
+	 * Reject flags, only allow namespaces.
+	 */
+	if (flags & ~(XFS_IOC_ATTR_ROOT | XFS_IOC_ATTR_SECURE))
+		return -EINVAL;
+	if (flags == (XFS_IOC_ATTR_ROOT | XFS_IOC_ATTR_SECURE))
+		return -EINVAL;
+
+	/*
+	 * Validate the cursor.
+	 */
+	if (copy_from_user(&context.cursor, ucursor, sizeof(context.cursor)))
+		return -EFAULT;
+	if (context.cursor.pad1 || context.cursor.pad2)
+		return -EINVAL;
+	if (!context.cursor.initted &&
+	    (context.cursor.hashval || context.cursor.blkno ||
+	     context.cursor.offset))
+		return -EINVAL;
+
+	buffer = kvzalloc(bufsize, GFP_KERNEL);
+	if (!buffer)
+		return -ENOMEM;
+
+	/*
+	 * Initialize the output buffer.
+	 */
+	context.dp = dp;
+	context.resynch = 1;
+	context.attr_filter = xfs_attr_filter(flags);
+	context.buffer = buffer;
+	context.bufsize = round_down(bufsize, sizeof(uint32_t));
+	context.firstu = context.bufsize;
+	context.put_listent = xfs_ioc_attr_put_listent;
+
+	alist = context.buffer;
+	alist->al_count = 0;
+	alist->al_more = 0;
+	alist->al_offset[0] = context.bufsize;
+
+	error = xfs_attr_list(&context);
+	if (error)
+		goto out_free;
+
+	if (copy_to_user(ubuf, buffer, bufsize) ||
+	    copy_to_user(ucursor, &context.cursor, sizeof(context.cursor)))
+		error = -EFAULT;
+out_free:
+	kvfree(buffer);
+	return error;
+}
+
+int
+xfs_attrlist_by_handle(
+	struct file		*parfilp,
+	struct xfs_fsop_attrlist_handlereq __user *p)
+{
+	struct xfs_fsop_attrlist_handlereq al_hreq;
+	struct dentry		*dentry;
+	int			error = -ENOMEM;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+	if (copy_from_user(&al_hreq, p, sizeof(al_hreq)))
+		return -EFAULT;
+
+	dentry = xfs_handlereq_to_dentry(parfilp, &al_hreq.hreq);
+	if (IS_ERR(dentry))
+		return PTR_ERR(dentry);
+
+	error = xfs_ioc_attr_list(XFS_I(d_inode(dentry)), al_hreq.buffer,
+				  al_hreq.buflen, al_hreq.flags, &p->pos);
+	dput(dentry);
+	return error;
+}
+
+static int
+xfs_attrmulti_attr_get(
+	struct inode		*inode,
+	unsigned char		*name,
+	unsigned char		__user *ubuf,
+	uint32_t		*len,
+	uint32_t		flags)
+{
+	struct xfs_da_args	args = {
+		.dp		= XFS_I(inode),
+		.attr_filter	= xfs_attr_filter(flags),
+		.name		= name,
+		.namelen	= strlen(name),
+		.valuelen	= *len,
+	};
+	int			error;
+
+	if (*len > XFS_XATTR_SIZE_MAX)
+		return -EINVAL;
+
+	error = xfs_attr_get(&args);
+	if (error)
+		goto out_kfree;
+
+	*len = args.valuelen;
+	if (copy_to_user(ubuf, args.value, args.valuelen))
+		error = -EFAULT;
+
+out_kfree:
+	kvfree(args.value);
+	return error;
+}
+
+static int
+xfs_attrmulti_attr_set(
+	struct inode		*inode,
+	unsigned char		*name,
+	const unsigned char	__user *ubuf,
+	uint32_t		len,
+	uint32_t		flags)
+{
+	struct xfs_da_args	args = {
+		.dp		= XFS_I(inode),
+		.attr_filter	= xfs_attr_filter(flags),
+		.name		= name,
+		.namelen	= strlen(name),
+	};
+	int			error;
+
+	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
+		return -EPERM;
+
+	if (ubuf) {
+		if (len > XFS_XATTR_SIZE_MAX)
+			return -EINVAL;
+		args.value = memdup_user(ubuf, len);
+		if (IS_ERR(args.value))
+			return PTR_ERR(args.value);
+		args.valuelen = len;
+	}
+
+	error = xfs_attr_change(&args, xfs_xattr_flags(flags, args.value));
+	if (!error && (flags & XFS_IOC_ATTR_ROOT))
+		xfs_forget_acl(inode, name);
+	kfree(args.value);
+	return error;
+}
+
+int
+xfs_ioc_attrmulti_one(
+	struct file		*parfilp,
+	struct inode		*inode,
+	uint32_t		opcode,
+	void __user		*uname,
+	void __user		*value,
+	uint32_t		*len,
+	uint32_t		flags)
+{
+	unsigned char		*name;
+	int			error;
+
+	if ((flags & XFS_IOC_ATTR_ROOT) && (flags & XFS_IOC_ATTR_SECURE))
+		return -EINVAL;
+
+	name = strndup_user(uname, MAXNAMELEN);
+	if (IS_ERR(name))
+		return PTR_ERR(name);
+
+	switch (opcode) {
+	case ATTR_OP_GET:
+		error = xfs_attrmulti_attr_get(inode, name, value, len, flags);
+		break;
+	case ATTR_OP_REMOVE:
+		value = NULL;
+		*len = 0;
+		fallthrough;
+	case ATTR_OP_SET:
+		error = mnt_want_write_file(parfilp);
+		if (error)
+			break;
+		error = xfs_attrmulti_attr_set(inode, name, value, *len, flags);
+		mnt_drop_write_file(parfilp);
+		break;
+	default:
+		error = -EINVAL;
+		break;
+	}
+
+	kfree(name);
+	return error;
+}
+
+int
+xfs_attrmulti_by_handle(
+	struct file		*parfilp,
+	void			__user *arg)
+{
+	int			error;
+	xfs_attr_multiop_t	*ops;
+	xfs_fsop_attrmulti_handlereq_t am_hreq;
+	struct dentry		*dentry;
+	unsigned int		i, size;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+	if (copy_from_user(&am_hreq, arg, sizeof(xfs_fsop_attrmulti_handlereq_t)))
+		return -EFAULT;
+
+	/* overflow check */
+	if (am_hreq.opcount >= INT_MAX / sizeof(xfs_attr_multiop_t))
+		return -E2BIG;
+
+	dentry = xfs_handlereq_to_dentry(parfilp, &am_hreq.hreq);
+	if (IS_ERR(dentry))
+		return PTR_ERR(dentry);
+
+	error = -E2BIG;
+	size = am_hreq.opcount * sizeof(xfs_attr_multiop_t);
+	if (!size || size > 16 * PAGE_SIZE)
+		goto out_dput;
+
+	ops = memdup_user(am_hreq.ops, size);
+	if (IS_ERR(ops)) {
+		error = PTR_ERR(ops);
+		goto out_dput;
+	}
+
+	error = 0;
+	for (i = 0; i < am_hreq.opcount; i++) {
+		ops[i].am_error = xfs_ioc_attrmulti_one(parfilp,
+				d_inode(dentry), ops[i].am_opcode,
+				ops[i].am_attrname, ops[i].am_attrvalue,
+				&ops[i].am_length, ops[i].am_flags);
+	}
+
+	if (copy_to_user(am_hreq.ops, ops, size))
+		error = -EFAULT;
+
+	kfree(ops);
+ out_dput:
+	dput(dentry);
+	return error;
+}
+
+struct xfs_getparents_ctx {
+	struct xfs_attr_list_context	context;
+	struct xfs_getparents_by_handle	gph;
+
+	/* File to target */
+	struct xfs_inode		*ip;
+
+	/* Internal buffer where we format records */
+	void				*krecords;
+
+	/* Last record filled out */
+	struct xfs_getparents_rec	*lastrec;
+
+	unsigned int			count;
+};
+
+static inline unsigned int
+xfs_getparents_rec_sizeof(
+	unsigned int		namelen)
+{
+	return round_up(sizeof(struct xfs_getparents_rec) + namelen + 1,
+			sizeof(uint64_t));
+}
+
+static void
+xfs_getparents_put_listent(
+	struct xfs_attr_list_context	*context,
+	int				flags,
+	unsigned char			*name,
+	int				namelen,
+	void				*value,
+	int				valuelen)
+{
+	struct xfs_getparents_ctx	*gpx =
+		container_of(context, struct xfs_getparents_ctx, context);
+	struct xfs_inode		*ip = context->dp;
+	struct xfs_mount		*mp = ip->i_mount;
+	struct xfs_getparents		*gp = &gpx->gph.gph_request;
+	struct xfs_getparents_rec	*gpr = gpx->krecords + context->firstu;
+	unsigned short			reclen =
+		xfs_getparents_rec_sizeof(namelen);
+	xfs_ino_t			ino;
+	uint32_t			gen;
+	int				error;
+
+	if (!(flags & XFS_ATTR_PARENT))
+		return;
+
+	error = xfs_parent_from_attr(mp, flags, name, namelen, value, valuelen,
+			&ino, &gen);
+	if (error) {
+		xfs_inode_mark_sick(ip, XFS_SICK_INO_PARENT);
+		context->seen_enough = -EFSCORRUPTED;
+		return;
+	}
+
+	/*
+	 * We found a parent pointer, but we've filled up the buffer.  Signal
+	 * to the caller that we did /not/ reach the end of the parent pointer
+	 * recordset.
+	 */
+	if (context->firstu > context->bufsize - reclen) {
+		context->seen_enough = 1;
+		return;
+	}
+
+	/* Format the parent pointer directly into the caller buffer. */
+	gpr->gpr_reclen = reclen;
+	xfs_filehandle_init(mp, ino, gen, &gpr->gpr_parent);
+	memcpy(gpr->gpr_name, name, namelen);
+	gpr->gpr_name[namelen] = 0;
+
+	trace_xfs_getparents_put_listent(ip, gp, context, gpr);
+
+	context->firstu += reclen;
+	gpx->count++;
+	gpx->lastrec = gpr;
+}
+
+/* Expand the last record to fill the rest of the caller's buffer. */
+static inline void
+xfs_getparents_expand_lastrec(
+	struct xfs_getparents_ctx	*gpx)
+{
+	struct xfs_getparents		*gp = &gpx->gph.gph_request;
+	struct xfs_getparents_rec	*gpr = gpx->lastrec;
+
+	if (!gpx->lastrec)
+		gpr = gpx->krecords;
+
+	gpr->gpr_reclen = gp->gp_bufsize - ((void *)gpr - gpx->krecords);
+
+	trace_xfs_getparents_expand_lastrec(gpx->ip, gp, &gpx->context, gpr);
+}
+
+/* Retrieve the parent pointers for a given inode. */
+STATIC int
+xfs_getparents(
+	struct xfs_getparents_ctx	*gpx)
+{
+	struct xfs_getparents		*gp = &gpx->gph.gph_request;
+	struct xfs_inode		*ip = gpx->ip;
+	struct xfs_mount		*mp = ip->i_mount;
+	size_t				bufsize;
+	int				error;
+
+	/* Check size of buffer requested by user */
+	if (gp->gp_bufsize > XFS_XATTR_LIST_MAX)
+		return -ENOMEM;
+	if (gp->gp_bufsize < xfs_getparents_rec_sizeof(1))
+		return -EINVAL;
+
+	if (gp->gp_iflags & ~XFS_GETPARENTS_IFLAGS_ALL)
+		return -EINVAL;
+	if (gp->gp_reserved)
+		return -EINVAL;
+
+	bufsize = round_down(gp->gp_bufsize, sizeof(uint64_t));
+	gpx->krecords = kvzalloc(bufsize, GFP_KERNEL);
+	if (!gpx->krecords) {
+		bufsize = min(bufsize, PAGE_SIZE);
+		gpx->krecords = kvzalloc(bufsize, GFP_KERNEL);
+		if (!gpx->krecords)
+			return -ENOMEM;
+	}
+
+	gpx->context.dp = ip;
+	gpx->context.resynch = 1;
+	gpx->context.put_listent = xfs_getparents_put_listent;
+	gpx->context.bufsize = bufsize;
+	/* firstu is used to track the bytes filled in the buffer */
+	gpx->context.firstu = 0;
+
+	/* Copy the cursor provided by caller */
+	memcpy(&gpx->context.cursor, &gp->gp_cursor,
+			sizeof(struct xfs_attrlist_cursor));
+	gpx->count = 0;
+	gp->gp_oflags = 0;
+
+	trace_xfs_getparents_begin(ip, gp, &gpx->context.cursor);
+
+	error = xfs_attr_list(&gpx->context);
+	if (error)
+		goto out_free_buf;
+	if (gpx->context.seen_enough < 0) {
+		error = gpx->context.seen_enough;
+		goto out_free_buf;
+	}
+	xfs_getparents_expand_lastrec(gpx);
+
+	/* Update the caller with the current cursor position */
+	memcpy(&gp->gp_cursor, &gpx->context.cursor,
+			sizeof(struct xfs_attrlist_cursor));
+
+	/* Is this the root directory? */
+	if (ip->i_ino == mp->m_sb.sb_rootino)
+		gp->gp_oflags |= XFS_GETPARENTS_OFLAG_ROOT;
+
+	if (gpx->context.seen_enough == 0) {
+		/*
+		 * If we did not run out of buffer space, then we reached the
+		 * end of the pptr recordset, so set the DONE flag.
+		 */
+		gp->gp_oflags |= XFS_GETPARENTS_OFLAG_DONE;
+	} else if (gpx->count == 0) {
+		/*
+		 * If we ran out of buffer space before copying any parent
+		 * pointers at all, the caller's buffer was too short.  Tell
+		 * userspace that, erm, the message is too long.
+		 */
+		error = -EMSGSIZE;
+		goto out_free_buf;
+	}
+
+	trace_xfs_getparents_end(ip, gp, &gpx->context.cursor);
+
+	ASSERT(gpx->context.firstu <= gpx->gph.gph_request.gp_bufsize);
+
+	/* Copy the records to userspace. */
+	if (copy_to_user(u64_to_user_ptr(gpx->gph.gph_request.gp_buffer),
+				gpx->krecords, gpx->context.firstu))
+		error = -EFAULT;
+
+out_free_buf:
+	kvfree(gpx->krecords);
+	gpx->krecords = NULL;
+	return error;
+}
+
+/* Retrieve the parents of this file and pass them back to userspace. */
+int
+xfs_ioc_getparents(
+	struct file			*file,
+	struct xfs_getparents __user	*ureq)
+{
+	struct xfs_getparents_ctx	gpx = {
+		.ip			= XFS_I(file_inode(file)),
+	};
+	struct xfs_getparents		*kreq = &gpx.gph.gph_request;
+	struct xfs_mount		*mp = gpx.ip->i_mount;
+	int				error;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+	if (!xfs_has_parent(mp))
+		return -EOPNOTSUPP;
+	if (copy_from_user(kreq, ureq, sizeof(*kreq)))
+		return -EFAULT;
+
+	error = xfs_getparents(&gpx);
+	if (error)
+		return error;
+
+	if (copy_to_user(ureq, kreq, sizeof(*kreq)))
+		return -EFAULT;
+
+	return 0;
+}
+
+/* Retrieve the parents of this file handle and pass them back to userspace. */
+int
+xfs_ioc_getparents_by_handle(
+	struct file			*file,
+	struct xfs_getparents_by_handle __user	*ureq)
+{
+	struct xfs_getparents_ctx	gpx = { };
+	struct xfs_inode		*ip = XFS_I(file_inode(file));
+	struct xfs_mount		*mp = ip->i_mount;
+	struct xfs_getparents_by_handle	*kreq = &gpx.gph;
+	struct xfs_handle		*handle = &kreq->gph_handle;
+	int				error;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+	if (!xfs_has_parent(mp))
+		return -EOPNOTSUPP;
+	if (copy_from_user(kreq, ureq, sizeof(*kreq)))
+		return -EFAULT;
+
+	/*
+	 * We don't use exportfs_decode_fh because it does too much work here.
+	 * If the handle refers to a directory, the exportfs code will walk
+	 * upwards through the directory tree to connect the dentries to the
+	 * root directory dentry.  For GETPARENTS we don't care about that
+	 * because we're not actually going to open a file descriptor; we only
+	 * want to open an inode and read its parent pointers.
+	 *
+	 * Note that xfs_scrub uses GETPARENTS to log that it will try to fix a
+	 * corrupted file's metadata.  For this usecase we would really rather
+	 * userspace single-step the path reconstruction to avoid loops or
+	 * other strange things if the directory tree is corrupt.
+	 */
+	gpx.ip = xfs_khandle_to_inode(file, handle);
+	if (IS_ERR(gpx.ip))
+		return PTR_ERR(gpx.ip);
+
+	error = xfs_getparents(&gpx);
+	if (error)
+		goto out_rele;
+
+	if (copy_to_user(ureq, kreq, sizeof(*kreq)))
+		error = -EFAULT;
+
+out_rele:
+	xfs_irele(gpx.ip);
+	return error;
+}
diff --git a/fs/xfs/xfs_handle.h b/fs/xfs/xfs_handle.h
new file mode 100644
index 0000000000..6799a86d85
--- /dev/null
+++ b/fs/xfs/xfs_handle.h
@@ -0,0 +1,33 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * Copyright (c) 2022-2024 Oracle.
+ * All rights reserved.
+ */
+#ifndef	__XFS_HANDLE_H__
+#define	__XFS_HANDLE_H__
+
+int xfs_attrlist_by_handle(struct file *parfilp,
+		struct xfs_fsop_attrlist_handlereq __user *p);
+int xfs_attrmulti_by_handle(struct file *parfilp, void __user *arg);
+
+int xfs_find_handle(unsigned int cmd, struct xfs_fsop_handlereq *hreq);
+int xfs_open_by_handle(struct file *parfilp, struct xfs_fsop_handlereq *hreq);
+int xfs_readlink_by_handle(struct file *parfilp,
+		struct xfs_fsop_handlereq *hreq);
+
+int xfs_ioc_attrmulti_one(struct file *parfilp, struct inode *inode,
+		uint32_t opcode, void __user *uname, void __user *value,
+		uint32_t *len, uint32_t flags);
+int xfs_ioc_attr_list(struct xfs_inode *dp, void __user *ubuf,
+		      size_t bufsize, int flags,
+		      struct xfs_attrlist_cursor __user *ucursor);
+
+struct dentry *xfs_handle_to_dentry(struct file *parfilp, void __user *uhandle,
+		u32 hlen);
+
+int xfs_ioc_getparents(struct file *file, struct xfs_getparents __user *arg);
+int xfs_ioc_getparents_by_handle(struct file *file,
+		struct xfs_getparents_by_handle __user *arg);
+
+#endif	/* __XFS_HANDLE_H__ */
diff --git a/fs/xfs/xfs_health.c b/fs/xfs/xfs_health.c
index 9a57afee93..10f116d093 100644
--- a/fs/xfs/xfs_health.c
+++ b/fs/xfs/xfs_health.c
@@ -14,6 +14,10 @@
 #include "xfs_trace.h"
 #include "xfs_health.h"
 #include "xfs_ag.h"
+#include "xfs_btree.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
+#include "xfs_quota_defs.h"
 
 /*
  * Warn about metadata corruption that we detected but haven't fixed, and
@@ -93,11 +97,25 @@ xfs_fs_mark_sick(
 	struct xfs_mount	*mp,
 	unsigned int		mask)
 {
-	ASSERT(!(mask & ~XFS_SICK_FS_PRIMARY));
+	ASSERT(!(mask & ~XFS_SICK_FS_ALL));
 	trace_xfs_fs_mark_sick(mp, mask);
 
 	spin_lock(&mp->m_sb_lock);
 	mp->m_fs_sick |= mask;
+	spin_unlock(&mp->m_sb_lock);
+}
+
+/* Mark per-fs metadata as having been checked and found unhealthy by fsck. */
+void
+xfs_fs_mark_corrupt(
+	struct xfs_mount	*mp,
+	unsigned int		mask)
+{
+	ASSERT(!(mask & ~XFS_SICK_FS_ALL));
+	trace_xfs_fs_mark_corrupt(mp, mask);
+
+	spin_lock(&mp->m_sb_lock);
+	mp->m_fs_sick |= mask;
 	mp->m_fs_checked |= mask;
 	spin_unlock(&mp->m_sb_lock);
 }
@@ -108,11 +126,13 @@ xfs_fs_mark_healthy(
 	struct xfs_mount	*mp,
 	unsigned int		mask)
 {
-	ASSERT(!(mask & ~XFS_SICK_FS_PRIMARY));
+	ASSERT(!(mask & ~XFS_SICK_FS_ALL));
 	trace_xfs_fs_mark_healthy(mp, mask);
 
 	spin_lock(&mp->m_sb_lock);
 	mp->m_fs_sick &= ~mask;
+	if (!(mp->m_fs_sick & XFS_SICK_FS_PRIMARY))
+		mp->m_fs_sick &= ~XFS_SICK_FS_SECONDARY;
 	mp->m_fs_checked |= mask;
 	spin_unlock(&mp->m_sb_lock);
 }
@@ -136,11 +156,25 @@ xfs_rt_mark_sick(
 	struct xfs_mount	*mp,
 	unsigned int		mask)
 {
-	ASSERT(!(mask & ~XFS_SICK_RT_PRIMARY));
+	ASSERT(!(mask & ~XFS_SICK_RT_ALL));
 	trace_xfs_rt_mark_sick(mp, mask);
 
 	spin_lock(&mp->m_sb_lock);
 	mp->m_rt_sick |= mask;
+	spin_unlock(&mp->m_sb_lock);
+}
+
+/* Mark realtime metadata as having been checked and found unhealthy by fsck. */
+void
+xfs_rt_mark_corrupt(
+	struct xfs_mount	*mp,
+	unsigned int		mask)
+{
+	ASSERT(!(mask & ~XFS_SICK_RT_ALL));
+	trace_xfs_rt_mark_corrupt(mp, mask);
+
+	spin_lock(&mp->m_sb_lock);
+	mp->m_rt_sick |= mask;
 	mp->m_rt_checked |= mask;
 	spin_unlock(&mp->m_sb_lock);
 }
@@ -151,11 +185,13 @@ xfs_rt_mark_healthy(
 	struct xfs_mount	*mp,
 	unsigned int		mask)
 {
-	ASSERT(!(mask & ~XFS_SICK_RT_PRIMARY));
+	ASSERT(!(mask & ~XFS_SICK_RT_ALL));
 	trace_xfs_rt_mark_healthy(mp, mask);
 
 	spin_lock(&mp->m_sb_lock);
 	mp->m_rt_sick &= ~mask;
+	if (!(mp->m_rt_sick & XFS_SICK_RT_PRIMARY))
+		mp->m_rt_sick &= ~XFS_SICK_RT_SECONDARY;
 	mp->m_rt_checked |= mask;
 	spin_unlock(&mp->m_sb_lock);
 }
@@ -173,17 +209,48 @@ xfs_rt_measure_sickness(
 	spin_unlock(&mp->m_sb_lock);
 }
 
+/* Mark unhealthy per-ag metadata given a raw AG number. */
+void
+xfs_agno_mark_sick(
+	struct xfs_mount	*mp,
+	xfs_agnumber_t		agno,
+	unsigned int		mask)
+{
+	struct xfs_perag	*pag = xfs_perag_get(mp, agno);
+
+	/* per-ag structure not set up yet? */
+	if (!pag)
+		return;
+
+	xfs_ag_mark_sick(pag, mask);
+	xfs_perag_put(pag);
+}
+
 /* Mark unhealthy per-ag metadata. */
 void
 xfs_ag_mark_sick(
 	struct xfs_perag	*pag,
 	unsigned int		mask)
 {
-	ASSERT(!(mask & ~XFS_SICK_AG_PRIMARY));
+	ASSERT(!(mask & ~XFS_SICK_AG_ALL));
 	trace_xfs_ag_mark_sick(pag->pag_mount, pag->pag_agno, mask);
 
 	spin_lock(&pag->pag_state_lock);
 	pag->pag_sick |= mask;
+	spin_unlock(&pag->pag_state_lock);
+}
+
+/* Mark per-ag metadata as having been checked and found unhealthy by fsck. */
+void
+xfs_ag_mark_corrupt(
+	struct xfs_perag	*pag,
+	unsigned int		mask)
+{
+	ASSERT(!(mask & ~XFS_SICK_AG_ALL));
+	trace_xfs_ag_mark_corrupt(pag->pag_mount, pag->pag_agno, mask);
+
+	spin_lock(&pag->pag_state_lock);
+	pag->pag_sick |= mask;
 	pag->pag_checked |= mask;
 	spin_unlock(&pag->pag_state_lock);
 }
@@ -194,11 +261,13 @@ xfs_ag_mark_healthy(
 	struct xfs_perag	*pag,
 	unsigned int		mask)
 {
-	ASSERT(!(mask & ~XFS_SICK_AG_PRIMARY));
+	ASSERT(!(mask & ~XFS_SICK_AG_ALL));
 	trace_xfs_ag_mark_healthy(pag->pag_mount, pag->pag_agno, mask);
 
 	spin_lock(&pag->pag_state_lock);
 	pag->pag_sick &= ~mask;
+	if (!(pag->pag_sick & XFS_SICK_AG_PRIMARY))
+		pag->pag_sick &= ~XFS_SICK_AG_SECONDARY;
 	pag->pag_checked |= mask;
 	spin_unlock(&pag->pag_state_lock);
 }
@@ -222,11 +291,34 @@ xfs_inode_mark_sick(
 	struct xfs_inode	*ip,
 	unsigned int		mask)
 {
-	ASSERT(!(mask & ~(XFS_SICK_INO_PRIMARY | XFS_SICK_INO_ZAPPED)));
+	ASSERT(!(mask & ~XFS_SICK_INO_ALL));
 	trace_xfs_inode_mark_sick(ip, mask);
 
 	spin_lock(&ip->i_flags_lock);
 	ip->i_sick |= mask;
+	spin_unlock(&ip->i_flags_lock);
+
+	/*
+	 * Keep this inode around so we don't lose the sickness report.  Scrub
+	 * grabs inodes with DONTCACHE assuming that most inode are ok, which
+	 * is not the case here.
+	 */
+	spin_lock(&VFS_I(ip)->i_lock);
+	VFS_I(ip)->i_state &= ~I_DONTCACHE;
+	spin_unlock(&VFS_I(ip)->i_lock);
+}
+
+/* Mark inode metadata as having been checked and found unhealthy by fsck. */
+void
+xfs_inode_mark_corrupt(
+	struct xfs_inode	*ip,
+	unsigned int		mask)
+{
+	ASSERT(!(mask & ~XFS_SICK_INO_ALL));
+	trace_xfs_inode_mark_corrupt(ip, mask);
+
+	spin_lock(&ip->i_flags_lock);
+	ip->i_sick |= mask;
 	ip->i_checked |= mask;
 	spin_unlock(&ip->i_flags_lock);
 
@@ -246,11 +338,13 @@ xfs_inode_mark_healthy(
 	struct xfs_inode	*ip,
 	unsigned int		mask)
 {
-	ASSERT(!(mask & ~(XFS_SICK_INO_PRIMARY | XFS_SICK_INO_ZAPPED)));
+	ASSERT(!(mask & ~XFS_SICK_INO_ALL));
 	trace_xfs_inode_mark_healthy(ip, mask);
 
 	spin_lock(&ip->i_flags_lock);
 	ip->i_sick &= ~mask;
+	if (!(ip->i_sick & XFS_SICK_INO_PRIMARY))
+		ip->i_sick &= ~XFS_SICK_INO_SECONDARY;
 	ip->i_checked |= mask;
 	spin_unlock(&ip->i_flags_lock);
 }
@@ -280,6 +374,8 @@ static const struct ioctl_sick_map fs_map[] = {
 	{ XFS_SICK_FS_UQUOTA,	XFS_FSOP_GEOM_SICK_UQUOTA },
 	{ XFS_SICK_FS_GQUOTA,	XFS_FSOP_GEOM_SICK_GQUOTA },
 	{ XFS_SICK_FS_PQUOTA,	XFS_FSOP_GEOM_SICK_PQUOTA },
+	{ XFS_SICK_FS_QUOTACHECK, XFS_FSOP_GEOM_SICK_QUOTACHECK },
+	{ XFS_SICK_FS_NLINKS,	XFS_FSOP_GEOM_SICK_NLINKS },
 	{ 0, 0 },
 };
 
@@ -335,6 +431,7 @@ static const struct ioctl_sick_map ag_map[] = {
 	{ XFS_SICK_AG_FINOBT,	XFS_AG_GEOM_SICK_FINOBT },
 	{ XFS_SICK_AG_RMAPBT,	XFS_AG_GEOM_SICK_RMAPBT },
 	{ XFS_SICK_AG_REFCNTBT,	XFS_AG_GEOM_SICK_REFCNTBT },
+	{ XFS_SICK_AG_INODES,	XFS_AG_GEOM_SICK_INODES },
 	{ 0, 0 },
 };
 
@@ -373,6 +470,7 @@ static const struct ioctl_sick_map ino_map[] = {
 	{ XFS_SICK_INO_BMBTA_ZAPPED,	XFS_BS_SICK_BMBTA },
 	{ XFS_SICK_INO_DIR_ZAPPED,	XFS_BS_SICK_DIR },
 	{ XFS_SICK_INO_SYMLINK_ZAPPED,	XFS_BS_SICK_SYMLINK },
+	{ XFS_SICK_INO_DIRTREE,	XFS_BS_SICK_DIRTREE },
 	{ 0, 0 },
 };
 
@@ -397,3 +495,92 @@ xfs_bulkstat_health(
 			bs->bs_sick |= m->ioctl_mask;
 	}
 }
+
+/* Mark a block mapping sick. */
+void
+xfs_bmap_mark_sick(
+	struct xfs_inode	*ip,
+	int			whichfork)
+{
+	unsigned int		mask;
+
+	switch (whichfork) {
+	case XFS_DATA_FORK:
+		mask = XFS_SICK_INO_BMBTD;
+		break;
+	case XFS_ATTR_FORK:
+		mask = XFS_SICK_INO_BMBTA;
+		break;
+	case XFS_COW_FORK:
+		mask = XFS_SICK_INO_BMBTC;
+		break;
+	default:
+		ASSERT(0);
+		return;
+	}
+
+	xfs_inode_mark_sick(ip, mask);
+}
+
+/* Record observations of btree corruption with the health tracking system. */
+void
+xfs_btree_mark_sick(
+	struct xfs_btree_cur		*cur)
+{
+	switch (cur->bc_ops->type) {
+	case XFS_BTREE_TYPE_MEM:
+		/* no health state tracking for ephemeral btrees */
+		return;
+	case XFS_BTREE_TYPE_AG:
+		ASSERT(cur->bc_ops->sick_mask);
+		xfs_ag_mark_sick(cur->bc_ag.pag, cur->bc_ops->sick_mask);
+		return;
+	case XFS_BTREE_TYPE_INODE:
+		if (xfs_btree_is_bmap(cur->bc_ops)) {
+			xfs_bmap_mark_sick(cur->bc_ino.ip,
+					   cur->bc_ino.whichfork);
+			return;
+		}
+		fallthrough;
+	default:
+		ASSERT(0);
+		return;
+	}
+}
+
+/*
+ * Record observations of dir/attr btree corruption with the health tracking
+ * system.
+ */
+void
+xfs_dirattr_mark_sick(
+	struct xfs_inode	*ip,
+	int			whichfork)
+{
+	unsigned int		mask;
+
+	switch (whichfork) {
+	case XFS_DATA_FORK:
+		mask = XFS_SICK_INO_DIR;
+		break;
+	case XFS_ATTR_FORK:
+		mask = XFS_SICK_INO_XATTR;
+		break;
+	default:
+		ASSERT(0);
+		return;
+	}
+
+	xfs_inode_mark_sick(ip, mask);
+}
+
+/*
+ * Record observations of dir/attr btree corruption with the health tracking
+ * system.
+ */
+void
+xfs_da_mark_sick(
+	struct xfs_da_args	*args)
+{
+	xfs_dirattr_mark_sick(args->dp, args->whichfork);
+}
diff --git a/fs/xfs/xfs_hooks.c b/fs/xfs/xfs_hooks.c
new file mode 100644
index 0000000000..a58d1de2d3
--- /dev/null
+++ b/fs/xfs/xfs_hooks.c
@@ -0,0 +1,52 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2022-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_ag.h"
+#include "xfs_trace.h"
+
+/* Initialize a notifier chain. */
+void
+xfs_hooks_init(
+	struct xfs_hooks	*chain)
+{
+	BLOCKING_INIT_NOTIFIER_HEAD(&chain->head);
+}
+
+/* Make it so a function gets called whenever we hit a certain hook point. */
+int
+xfs_hooks_add(
+	struct xfs_hooks	*chain,
+	struct xfs_hook		*hook)
+{
+	ASSERT(hook->nb.notifier_call != NULL);
+	BUILD_BUG_ON(offsetof(struct xfs_hook, nb) != 0);
+
+	return blocking_notifier_chain_register(&chain->head, &hook->nb);
+}
+
+/* Remove a previously installed hook. */
+void
+xfs_hooks_del(
+	struct xfs_hooks	*chain,
+	struct xfs_hook		*hook)
+{
+	blocking_notifier_chain_unregister(&chain->head, &hook->nb);
+}
+
+/* Call a hook.  Returns the NOTIFY_* value returned by the last hook. */
+int
+xfs_hooks_call(
+	struct xfs_hooks	*chain,
+	unsigned long		val,
+	void			*priv)
+{
+	return blocking_notifier_call_chain(&chain->head, val, priv);
+}
diff --git a/fs/xfs/xfs_hooks.h b/fs/xfs/xfs_hooks.h
new file mode 100644
index 0000000000..60b8a58315
--- /dev/null
+++ b/fs/xfs/xfs_hooks.h
@@ -0,0 +1,65 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2022-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef XFS_HOOKS_H_
+#define XFS_HOOKS_H_
+
+#ifdef CONFIG_XFS_LIVE_HOOKS
+struct xfs_hooks {
+	struct blocking_notifier_head	head;
+};
+
+/*
+ * If jump labels are enabled in Kconfig, the static key uses nop sleds and
+ * code patching to eliminate the overhead of taking the rwsem in
+ * blocking_notifier_call_chain when there are no hooks configured.  If not,
+ * the static key per-call overhead is an atomic read.  Most arches that can
+ * handle XFS also support jump labels.
+ *
+ * Note: Patching the kernel code requires taking the cpu hotplug lock.  Other
+ * parts of the kernel allocate memory with that lock held, which means that
+ * XFS callers cannot hold any locks that might be used by memory reclaim or
+ * writeback when calling the static_branch_{inc,dec} functions.
+ */
+# define DEFINE_STATIC_XFS_HOOK_SWITCH(name) \
+	static DEFINE_STATIC_KEY_FALSE(name)
+# define xfs_hooks_switch_on(name)	static_branch_inc(name)
+# define xfs_hooks_switch_off(name)	static_branch_dec(name)
+# define xfs_hooks_switched_on(name)	static_branch_unlikely(name)
+
+struct xfs_hook {
+	/* This must come at the start of the structure. */
+	struct notifier_block		nb;
+};
+
+typedef	int (*xfs_hook_fn_t)(struct xfs_hook *hook, unsigned long action,
+		void *data);
+
+void xfs_hooks_init(struct xfs_hooks *chain);
+int xfs_hooks_add(struct xfs_hooks *chain, struct xfs_hook *hook);
+void xfs_hooks_del(struct xfs_hooks *chain, struct xfs_hook *hook);
+int xfs_hooks_call(struct xfs_hooks *chain, unsigned long action,
+		void *priv);
+
+static inline void xfs_hook_setup(struct xfs_hook *hook, notifier_fn_t fn)
+{
+	hook->nb.notifier_call = fn;
+	hook->nb.priority = 0;
+}
+
+#else
+
+struct xfs_hooks { /* empty */ };
+
+# define DEFINE_STATIC_XFS_HOOK_SWITCH(name)
+# define xfs_hooks_switch_on(name)		((void)0)
+# define xfs_hooks_switch_off(name)		((void)0)
+# define xfs_hooks_switched_on(name)		(false)
+
+# define xfs_hooks_init(chain)			((void)0)
+# define xfs_hooks_call(chain, val, priv)	(NOTIFY_DONE)
+#endif
+
+#endif /* XFS_HOOKS_H_ */
diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c
index dba514a2c8..9967334ea9 100644
--- a/fs/xfs/xfs_icache.c
+++ b/fs/xfs/xfs_icache.c
@@ -24,6 +24,7 @@
 #include "xfs_ialloc.h"
 #include "xfs_ag.h"
 #include "xfs_log_priv.h"
+#include "xfs_health.h"
 
 #include <linux/iversion.h>
 
@@ -415,6 +416,9 @@ xfs_iget_check_free_state(
 			xfs_warn(ip->i_mount,
 "Corruption detected! Free inode 0x%llx not marked free! (mode 0x%x)",
 				ip->i_ino, VFS_I(ip)->i_mode);
+			xfs_agno_mark_sick(ip->i_mount,
+					XFS_INO_TO_AGNO(ip->i_mount, ip->i_ino),
+					XFS_SICK_AG_INOBT);
 			return -EFSCORRUPTED;
 		}
 
@@ -422,6 +426,9 @@ xfs_iget_check_free_state(
 			xfs_warn(ip->i_mount,
 "Corruption detected! Free inode 0x%llx has blocks allocated!",
 				ip->i_ino);
+			xfs_agno_mark_sick(ip->i_mount,
+					XFS_INO_TO_AGNO(ip->i_mount, ip->i_ino),
+					XFS_SICK_AG_INOBT);
 			return -EFSCORRUPTED;
 		}
 		return 0;
@@ -606,7 +613,6 @@ xfs_iget_cache_miss(
 	struct xfs_inode	*ip;
 	int			error;
 	xfs_agino_t		agino = XFS_INO_TO_AGINO(mp, ino);
-	int			iflags;
 
 	ip = xfs_inode_alloc(mp, ino);
 	if (!ip)
@@ -640,6 +646,8 @@ xfs_iget_cache_miss(
 				xfs_buf_offset(bp, ip->i_imap.im_boffset));
 		if (!error)
 			xfs_buf_set_ref(bp, XFS_INO_REF);
+		else
+			xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE);
 		xfs_trans_brelse(tp, bp);
 
 		if (error)
@@ -659,10 +667,9 @@ xfs_iget_cache_miss(
 	/*
 	 * Preload the radix tree so we can insert safely under the
 	 * write spinlock. Note that we cannot sleep inside the preload
-	 * region. Since we can be called from transaction context, don't
-	 * recurse into the file system.
+	 * region.
 	 */
-	if (radix_tree_preload(GFP_NOFS)) {
+	if (radix_tree_preload(GFP_KERNEL | __GFP_NOLOCKDEP)) {
 		error = -EAGAIN;
 		goto out_destroy;
 	}
@@ -685,13 +692,12 @@ xfs_iget_cache_miss(
 	 * memory barrier that ensures this detection works correctly at lookup
 	 * time.
 	 */
-	iflags = XFS_INEW;
 	if (flags & XFS_IGET_DONTCACHE)
 		d_mark_dontcache(VFS_I(ip));
 	ip->i_udquot = NULL;
 	ip->i_gdquot = NULL;
 	ip->i_pdquot = NULL;
-	xfs_iflags_set(ip, iflags);
+	xfs_iflags_set(ip, XFS_INEW);
 
 	/* insert the new inode */
 	spin_lock(&pag->pag_ici_lock);
@@ -1149,7 +1155,7 @@ xfs_inode_free_eofblocks(
 	}
 	*lockflags |= XFS_IOLOCK_EXCL;
 
-	if (xfs_can_free_eofblocks(ip, false))
+	if (xfs_can_free_eofblocks(ip))
 		return xfs_free_eofblocks(ip);
 
 	/* inode could be preallocated or append-only */
@@ -2031,8 +2037,10 @@ xfs_inodegc_want_queue_work(
  *  - Memory shrinkers queued the inactivation worker and it hasn't finished.
  *  - The queue depth exceeds the maximum allowable percpu backlog.
  *
- * Note: If the current thread is running a transaction, we don't ever want to
- * wait for other transactions because that could introduce a deadlock.
+ * Note: If we are in a NOFS context here (e.g. current thread is running a
+ * transaction) the we don't want to block here as inodegc progress may require
+ * filesystem resources we hold to make progress and that could result in a
+ * deadlock. Hence we skip out of here if we are in a scoped NOFS context.
  */
 static inline bool
 xfs_inodegc_want_flush_work(
@@ -2040,7 +2048,7 @@ xfs_inodegc_want_flush_work(
 	unsigned int		items,
 	unsigned int		shrinker_hits)
 {
-	if (current->journal_info)
+	if (current->flags & PF_MEMALLOC_NOFS)
 		return false;
 
 	if (shrinker_hits > 0)
diff --git a/fs/xfs/xfs_icreate_item.c b/fs/xfs/xfs_icreate_item.c
index b05314d481..4345db5017 100644
--- a/fs/xfs/xfs_icreate_item.c
+++ b/fs/xfs/xfs_icreate_item.c
@@ -63,7 +63,7 @@ STATIC void
 xfs_icreate_item_release(
 	struct xfs_log_item	*lip)
 {
-	kmem_free(ICR_ITEM(lip)->ic_item.li_lv_shadow);
+	kvfree(ICR_ITEM(lip)->ic_item.li_lv_shadow);
 	kmem_cache_free(xfs_icreate_cache, ICR_ITEM(lip));
 }
 
diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index 1fd94958aa..a4e3cd8971 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -16,6 +16,7 @@
 #include "xfs_inode.h"
 #include "xfs_dir2.h"
 #include "xfs_attr.h"
+#include "xfs_bit.h"
 #include "xfs_trans_space.h"
 #include "xfs_trans.h"
 #include "xfs_buf_item.h"
@@ -38,13 +39,13 @@
 #include "xfs_ag.h"
 #include "xfs_log_priv.h"
 #include "xfs_health.h"
+#include "xfs_pnfs.h"
+#include "xfs_parent.h"
+#include "xfs_xattr.h"
+#include "xfs_sb.h"
 
 struct kmem_cache *xfs_inode_cache;
 
-STATIC int xfs_iunlink(struct xfs_trans *, struct xfs_inode *);
-STATIC int xfs_iunlink_remove(struct xfs_trans *tp, struct xfs_perag *pag,
-	struct xfs_inode *);
-
 /*
  * helper function to extract extent size hint from inode
  */
@@ -60,7 +61,8 @@ xfs_get_extsz_hint(
 		return 0;
 	if ((ip->i_diflags & XFS_DIFLAG_EXTSIZE) && ip->i_extsize)
 		return ip->i_extsize;
-	if (XFS_IS_REALTIME_INODE(ip))
+	if (XFS_IS_REALTIME_INODE(ip) &&
+	    ip->i_mount->m_sb.sb_rextsize > 1)
 		return ip->i_mount->m_sb.sb_rextsize;
 	return 0;
 }
@@ -203,9 +205,9 @@ xfs_ilock(
 	}
 
 	if (lock_flags & XFS_ILOCK_EXCL)
-		mrupdate_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags));
+		down_write_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags));
 	else if (lock_flags & XFS_ILOCK_SHARED)
-		mraccess_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags));
+		down_read_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags));
 }
 
 /*
@@ -246,10 +248,10 @@ xfs_ilock_nowait(
 	}
 
 	if (lock_flags & XFS_ILOCK_EXCL) {
-		if (!mrtryupdate(&ip->i_lock))
+		if (!down_write_trylock(&ip->i_lock))
 			goto out_undo_mmaplock;
 	} else if (lock_flags & XFS_ILOCK_SHARED) {
-		if (!mrtryaccess(&ip->i_lock))
+		if (!down_read_trylock(&ip->i_lock))
 			goto out_undo_mmaplock;
 	}
 	return 1;
@@ -298,9 +300,9 @@ xfs_iunlock(
 		up_read(&VFS_I(ip)->i_mapping->invalidate_lock);
 
 	if (lock_flags & XFS_ILOCK_EXCL)
-		mrunlock_excl(&ip->i_lock);
+		up_write(&ip->i_lock);
 	else if (lock_flags & XFS_ILOCK_SHARED)
-		mrunlock_shared(&ip->i_lock);
+		up_read(&ip->i_lock);
 
 	trace_xfs_iunlock(ip, lock_flags, _RET_IP_);
 }
@@ -319,7 +321,7 @@ xfs_ilock_demote(
 		~(XFS_IOLOCK_EXCL|XFS_MMAPLOCK_EXCL|XFS_ILOCK_EXCL)) == 0);
 
 	if (lock_flags & XFS_ILOCK_EXCL)
-		mrdemote(&ip->i_lock);
+		downgrade_write(&ip->i_lock);
 	if (lock_flags & XFS_MMAPLOCK_EXCL)
 		downgrade_write(&VFS_I(ip)->i_mapping->invalidate_lock);
 	if (lock_flags & XFS_IOLOCK_EXCL)
@@ -328,52 +330,30 @@ xfs_ilock_demote(
 	trace_xfs_ilock_demote(ip, lock_flags, _RET_IP_);
 }
 
-#if defined(DEBUG) || defined(XFS_WARN)
-static inline bool
-__xfs_rwsem_islocked(
-	struct rw_semaphore	*rwsem,
-	bool			shared)
-{
-	if (!debug_locks)
-		return rwsem_is_locked(rwsem);
-
-	if (!shared)
-		return lockdep_is_held_type(rwsem, 0);
-
-	/*
-	 * We are checking that the lock is held at least in shared
-	 * mode but don't care that it might be held exclusively
-	 * (i.e. shared | excl). Hence we check if the lock is held
-	 * in any mode rather than an explicit shared mode.
-	 */
-	return lockdep_is_held_type(rwsem, -1);
-}
-
-bool
-xfs_isilocked(
+void
+xfs_assert_ilocked(
 	struct xfs_inode	*ip,
 	uint			lock_flags)
 {
-	if (lock_flags & (XFS_ILOCK_EXCL|XFS_ILOCK_SHARED)) {
-		if (!(lock_flags & XFS_ILOCK_SHARED))
-			return !!ip->i_lock.mr_writer;
-		return rwsem_is_locked(&ip->i_lock.mr_lock);
-	}
-
-	if (lock_flags & (XFS_MMAPLOCK_EXCL|XFS_MMAPLOCK_SHARED)) {
-		return __xfs_rwsem_islocked(&VFS_I(ip)->i_mapping->invalidate_lock,
-				(lock_flags & XFS_MMAPLOCK_SHARED));
-	}
-
-	if (lock_flags & (XFS_IOLOCK_EXCL | XFS_IOLOCK_SHARED)) {
-		return __xfs_rwsem_islocked(&VFS_I(ip)->i_rwsem,
-				(lock_flags & XFS_IOLOCK_SHARED));
-	}
-
-	ASSERT(0);
-	return false;
+	/*
+	 * Sometimes we assert the ILOCK is held exclusively, but we're in
+	 * a workqueue, so lockdep doesn't know we're the owner.
+	 */
+	if (lock_flags & XFS_ILOCK_SHARED)
+		rwsem_assert_held(&ip->i_lock);
+	else if (lock_flags & XFS_ILOCK_EXCL)
+		rwsem_assert_held_write_nolockdep(&ip->i_lock);
+
+	if (lock_flags & XFS_MMAPLOCK_SHARED)
+		rwsem_assert_held(&VFS_I(ip)->i_mapping->invalidate_lock);
+	else if (lock_flags & XFS_MMAPLOCK_EXCL)
+		rwsem_assert_held_write(&VFS_I(ip)->i_mapping->invalidate_lock);
+
+	if (lock_flags & XFS_IOLOCK_SHARED)
+		rwsem_assert_held(&VFS_I(ip)->i_rwsem);
+	else if (lock_flags & XFS_IOLOCK_EXCL)
+		rwsem_assert_held_write(&VFS_I(ip)->i_rwsem);
 }
-#endif
 
 /*
  * xfs_lockdep_subclass_ok() is only used in an ASSERT, so is only called when
@@ -442,7 +422,7 @@ xfs_lock_inumorder(
  * lock more than one at a time, lockdep will report false positives saying we
  * have violated locking orders.
  */
-static void
+void
 xfs_lock_inodes(
 	struct xfs_inode	**ips,
 	int			inodes,
@@ -671,7 +651,7 @@ xfs_lookup(
 
 out_free_name:
 	if (ci_name)
-		kmem_free(ci_name->name);
+		kfree(ci_name->name);
 out_unlock:
 	*ipp = NULL;
 	return error;
@@ -771,6 +751,8 @@ xfs_inode_inherit_flags2(
 /*
  * Initialise a newly allocated inode and return the in-core inode to the
  * caller locked exclusively.
+ *
+ * Caller is responsible for unlocking the inode manually upon return
  */
 int
 xfs_init_new_inode(
@@ -802,6 +784,8 @@ xfs_init_new_inode(
 	 */
 	if ((pip && ino == pip->i_ino) || !xfs_verify_dir_ino(mp, ino)) {
 		xfs_alert(mp, "Allocated a known in-use inode 0x%llx!", ino);
+		xfs_agno_mark_sick(mp, XFS_INO_TO_AGNO(mp, ino),
+				XFS_SICK_AG_INOBT);
 		return -EFSCORRUPTED;
 	}
 
@@ -887,15 +871,22 @@ xfs_init_new_inode(
 	 * this saves us from needing to run a separate transaction to set the
 	 * fork offset in the immediate future.
 	 */
-	if (init_xattrs && xfs_has_attr(mp)) {
+	if (init_xattrs) {
 		ip->i_forkoff = xfs_default_attroffset(ip) >> 3;
 		xfs_ifork_init_attr(ip, XFS_DINODE_FMT_EXTENTS, 0);
+
+		if (!xfs_has_attr(mp)) {
+			spin_lock(&mp->m_sb_lock);
+			xfs_add_attr(mp);
+			spin_unlock(&mp->m_sb_lock);
+			xfs_log_sb(tp);
+		}
 	}
 
 	/*
 	 * Log the new values stuffed into the inode.
 	 */
-	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
+	xfs_trans_ijoin(tp, ip, 0);
 	xfs_trans_log_inode(tp, ip, flags);
 
 	/* now that we have an i_mode we can setup the inode structure */
@@ -910,24 +901,27 @@ xfs_init_new_inode(
  * link count to go to zero, move the inode to AGI unlinked list so that it can
  * be freed when the last active reference goes away via xfs_inactive().
  */
-static int			/* error */
+int
 xfs_droplink(
-	xfs_trans_t *tp,
-	xfs_inode_t *ip)
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip)
 {
-	if (VFS_I(ip)->i_nlink == 0) {
-		xfs_alert(ip->i_mount,
-			  "%s: Attempt to drop inode (%llu) with nlink zero.",
-			  __func__, ip->i_ino);
-		return -EFSCORRUPTED;
-	}
+	struct inode		*inode = VFS_I(ip);
 
 	xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG);
 
-	drop_nlink(VFS_I(ip));
+	if (inode->i_nlink == 0) {
+		xfs_info_ratelimited(tp->t_mountp,
+ "Inode 0x%llx link count dropped below zero.  Pinning link count.",
+				ip->i_ino);
+		set_nlink(inode, XFS_NLINK_PINNED);
+	}
+	if (inode->i_nlink != XFS_NLINK_PINNED)
+		drop_nlink(inode);
+
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 
-	if (VFS_I(ip)->i_nlink)
+	if (inode->i_nlink)
 		return 0;
 
 	return xfs_iunlink(tp, ip);
@@ -936,21 +930,104 @@ xfs_droplink(
 /*
  * Increment the link count on an inode & log the change.
  */
-static void
+void
 xfs_bumplink(
-	xfs_trans_t *tp,
-	xfs_inode_t *ip)
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip)
 {
+	struct inode		*inode = VFS_I(ip);
+
 	xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG);
 
-	inc_nlink(VFS_I(ip));
+	if (inode->i_nlink == XFS_NLINK_PINNED - 1)
+		xfs_info_ratelimited(tp->t_mountp,
+ "Inode 0x%llx link count exceeded maximum.  Pinning link count.",
+				ip->i_ino);
+	if (inode->i_nlink != XFS_NLINK_PINNED)
+		inc_nlink(inode);
+
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 }
 
+#ifdef CONFIG_XFS_LIVE_HOOKS
+/*
+ * Use a static key here to reduce the overhead of directory live update hooks.
+ * If the compiler supports jump labels, the static branch will be replaced by
+ * a nop sled when there are no hook users.  Online fsck is currently the only
+ * caller, so this is a reasonable tradeoff.
+ *
+ * Note: Patching the kernel code requires taking the cpu hotplug lock.  Other
+ * parts of the kernel allocate memory with that lock held, which means that
+ * XFS callers cannot hold any locks that might be used by memory reclaim or
+ * writeback when calling the static_branch_{inc,dec} functions.
+ */
+DEFINE_STATIC_XFS_HOOK_SWITCH(xfs_dir_hooks_switch);
+
+void
+xfs_dir_hook_disable(void)
+{
+	xfs_hooks_switch_off(&xfs_dir_hooks_switch);
+}
+
+void
+xfs_dir_hook_enable(void)
+{
+	xfs_hooks_switch_on(&xfs_dir_hooks_switch);
+}
+
+/* Call hooks for a directory update relating to a child dirent update. */
+inline void
+xfs_dir_update_hook(
+	struct xfs_inode		*dp,
+	struct xfs_inode		*ip,
+	int				delta,
+	const struct xfs_name		*name)
+{
+	if (xfs_hooks_switched_on(&xfs_dir_hooks_switch)) {
+		struct xfs_dir_update_params	p = {
+			.dp		= dp,
+			.ip		= ip,
+			.delta		= delta,
+			.name		= name,
+		};
+		struct xfs_mount	*mp = ip->i_mount;
+
+		xfs_hooks_call(&mp->m_dir_update_hooks, 0, &p);
+	}
+}
+
+/* Call the specified function during a directory update. */
+int
+xfs_dir_hook_add(
+	struct xfs_mount	*mp,
+	struct xfs_dir_hook	*hook)
+{
+	return xfs_hooks_add(&mp->m_dir_update_hooks, &hook->dirent_hook);
+}
+
+/* Stop calling the specified function during a directory update. */
+void
+xfs_dir_hook_del(
+	struct xfs_mount	*mp,
+	struct xfs_dir_hook	*hook)
+{
+	xfs_hooks_del(&mp->m_dir_update_hooks, &hook->dirent_hook);
+}
+
+/* Configure directory update hook functions. */
+void
+xfs_dir_hook_setup(
+	struct xfs_dir_hook	*hook,
+	notifier_fn_t		mod_fn)
+{
+	xfs_hook_setup(&hook->dirent_hook, mod_fn);
+}
+#endif /* CONFIG_XFS_LIVE_HOOKS */
+
 int
 xfs_create(
 	struct mnt_idmap	*idmap,
-	xfs_inode_t		*dp,
+	struct xfs_inode	*dp,
 	struct xfs_name		*name,
 	umode_t			mode,
 	dev_t			rdev,
@@ -962,7 +1039,7 @@ xfs_create(
 	struct xfs_inode	*ip = NULL;
 	struct xfs_trans	*tp = NULL;
 	int			error;
-	bool                    unlock_dp_on_error = false;
+	bool			unlock_dp_on_error = false;
 	prid_t			prid;
 	struct xfs_dquot	*udqp = NULL;
 	struct xfs_dquot	*gdqp = NULL;
@@ -970,6 +1047,7 @@ xfs_create(
 	struct xfs_trans_res	*tres;
 	uint			resblks;
 	xfs_ino_t		ino;
+	struct xfs_parent_args	*ppargs;
 
 	trace_xfs_create(dp, name);
 
@@ -991,13 +1069,17 @@ xfs_create(
 		return error;
 
 	if (is_dir) {
-		resblks = XFS_MKDIR_SPACE_RES(mp, name->len);
+		resblks = xfs_mkdir_space_res(mp, name->len);
 		tres = &M_RES(mp)->tr_mkdir;
 	} else {
-		resblks = XFS_CREATE_SPACE_RES(mp, name->len);
+		resblks = xfs_create_space_res(mp, name->len);
 		tres = &M_RES(mp)->tr_create;
 	}
 
+	error = xfs_parent_start(mp, &ppargs);
+	if (error)
+		goto out_release_dquots;
+
 	/*
 	 * Initially assume that the file does not exist and
 	 * reserve the resources for that case.  If that is not
@@ -1013,7 +1095,7 @@ xfs_create(
 				resblks, &tp);
 	}
 	if (error)
-		goto out_release_dquots;
+		goto out_parent;
 
 	xfs_ilock(dp, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT);
 	unlock_dp_on_error = true;
@@ -1037,8 +1119,7 @@ xfs_create(
 	 * the transaction cancel unlocking dp so don't do it explicitly in the
 	 * error path.
 	 */
-	xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
-	unlock_dp_on_error = false;
+	xfs_trans_ijoin(tp, dp, 0);
 
 	error = xfs_dir_createname(tp, dp, name, ip->i_ino,
 					resblks - XFS_IALLOC_SPACE_RES(mp));
@@ -1058,6 +1139,22 @@ xfs_create(
 	}
 
 	/*
+	 * If we have parent pointers, we need to add the attribute containing
+	 * the parent information now.
+	 */
+	if (ppargs) {
+		error = xfs_parent_addname(tp, ppargs, dp, name, ip);
+		if (error)
+			goto out_trans_cancel;
+	}
+
+	/*
+	 * Create ip with a reference from dp, and add '.' and '..' references
+	 * if it's a directory.
+	 */
+	xfs_dir_update_hook(dp, ip, 1, name);
+
+	/*
 	 * If this is a synchronous mount, make sure that the
 	 * create transaction goes to disk before returning to
 	 * the user.
@@ -1081,6 +1178,9 @@ xfs_create(
 	xfs_qm_dqrele(pdqp);
 
 	*ipp = ip;
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	xfs_iunlock(dp, XFS_ILOCK_EXCL);
+	xfs_parent_finish(mp, ppargs);
 	return 0;
 
  out_trans_cancel:
@@ -1092,9 +1192,12 @@ xfs_create(
 	 * transactions and deadlocks from xfs_inactive.
 	 */
 	if (ip) {
+		xfs_iunlock(ip, XFS_ILOCK_EXCL);
 		xfs_finish_inode_setup(ip);
 		xfs_irele(ip);
 	}
+ out_parent:
+	xfs_parent_finish(mp, ppargs);
  out_release_dquots:
 	xfs_qm_dqrele(udqp);
 	xfs_qm_dqrele(gdqp);
@@ -1110,6 +1213,7 @@ xfs_create_tmpfile(
 	struct mnt_idmap	*idmap,
 	struct xfs_inode	*dp,
 	umode_t			mode,
+	bool			init_xattrs,
 	struct xfs_inode	**ipp)
 {
 	struct xfs_mount	*mp = dp->i_mount;
@@ -1150,7 +1254,7 @@ xfs_create_tmpfile(
 	error = xfs_dialloc(&tp, dp->i_ino, mode, &ino);
 	if (!error)
 		error = xfs_init_new_inode(idmap, tp, dp, ino, mode,
-				0, 0, prid, false, &ip);
+				0, 0, prid, init_xattrs, &ip);
 	if (error)
 		goto out_trans_cancel;
 
@@ -1177,6 +1281,7 @@ xfs_create_tmpfile(
 	xfs_qm_dqrele(pdqp);
 
 	*ipp = ip;
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 	return 0;
 
  out_trans_cancel:
@@ -1188,6 +1293,7 @@ xfs_create_tmpfile(
 	 * transactions and deadlocks from xfs_inactive.
 	 */
 	if (ip) {
+		xfs_iunlock(ip, XFS_ILOCK_EXCL);
 		xfs_finish_inode_setup(ip);
 		xfs_irele(ip);
 	}
@@ -1201,14 +1307,15 @@ xfs_create_tmpfile(
 
 int
 xfs_link(
-	xfs_inode_t		*tdp,
-	xfs_inode_t		*sip,
+	struct xfs_inode	*tdp,
+	struct xfs_inode	*sip,
 	struct xfs_name		*target_name)
 {
-	xfs_mount_t		*mp = tdp->i_mount;
-	xfs_trans_t		*tp;
+	struct xfs_mount	*mp = tdp->i_mount;
+	struct xfs_trans	*tp;
 	int			error, nospace_error = 0;
 	int			resblks;
+	struct xfs_parent_args	*ppargs;
 
 	trace_xfs_link(tdp, target_name);
 
@@ -1227,11 +1334,25 @@ xfs_link(
 	if (error)
 		goto std_return;
 
-	resblks = XFS_LINK_SPACE_RES(mp, target_name->len);
+	error = xfs_parent_start(mp, &ppargs);
+	if (error)
+		goto std_return;
+
+	resblks = xfs_link_space_res(mp, target_name->len);
 	error = xfs_trans_alloc_dir(tdp, &M_RES(mp)->tr_link, sip, &resblks,
 			&tp, &nospace_error);
 	if (error)
-		goto std_return;
+		goto out_parent;
+
+	/*
+	 * We don't allow reservationless or quotaless hardlinking when parent
+	 * pointers are enabled because we can't back out if the xattrs must
+	 * grow.
+	 */
+	if (ppargs && nospace_error) {
+		error = nospace_error;
+		goto error_return;
+	}
 
 	/*
 	 * If we are using project inheritance, we only allow hard link
@@ -1240,8 +1361,19 @@ xfs_link(
 	 */
 	if (unlikely((tdp->i_diflags & XFS_DIFLAG_PROJINHERIT) &&
 		     tdp->i_projid != sip->i_projid)) {
-		error = -EXDEV;
-		goto error_return;
+		/*
+		 * Project quota setup skips special files which can
+		 * leave inodes in a PROJINHERIT directory without a
+		 * project ID set. We need to allow links to be made
+		 * to these "project-less" inodes because userspace
+		 * expects them to succeed after project ID setup,
+		 * but everything else should be rejected.
+		 */
+		if (!special_file(VFS_I(sip)->i_mode) ||
+		    sip->i_projid != 0) {
+			error = -EXDEV;
+			goto error_return;
+		}
 	}
 
 	if (!resblks) {
@@ -1273,6 +1405,20 @@ xfs_link(
 	xfs_bumplink(tp, sip);
 
 	/*
+	 * If we have parent pointers, we now need to add the parent record to
+	 * the attribute fork of the inode. If this is the initial parent
+	 * attribute, we need to create it correctly, otherwise we can just add
+	 * the parent to the inode.
+	 */
+	if (ppargs) {
+		error = xfs_parent_addname(tp, ppargs, tdp, target_name, sip);
+		if (error)
+			goto error_return;
+	}
+
+	xfs_dir_update_hook(tdp, sip, 1, target_name);
+
+	/*
 	 * If this is a synchronous mount, make sure that the
 	 * link transaction goes to disk before returning to
 	 * the user.
@@ -1280,10 +1426,18 @@ xfs_link(
 	if (xfs_has_wsync(mp) || xfs_has_dirsync(mp))
 		xfs_trans_set_sync(tp);
 
-	return xfs_trans_commit(tp);
+	error = xfs_trans_commit(tp);
+	xfs_iunlock(tdp, XFS_ILOCK_EXCL);
+	xfs_iunlock(sip, XFS_ILOCK_EXCL);
+	xfs_parent_finish(mp, ppargs);
+	return error;
 
  error_return:
 	xfs_trans_cancel(tp);
+	xfs_iunlock(tdp, XFS_ILOCK_EXCL);
+	xfs_iunlock(sip, XFS_ILOCK_EXCL);
+ out_parent:
+	xfs_parent_finish(mp, ppargs);
  std_return:
 	if (error == -ENOSPC && nospace_error)
 		error = nospace_error;
@@ -1342,9 +1496,9 @@ xfs_itruncate_extents_flags(
 	xfs_fileoff_t		first_unmap_block;
 	int			error = 0;
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-	ASSERT(!atomic_read(&VFS_I(ip)->i_count) ||
-	       xfs_isilocked(ip, XFS_IOLOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
+	if (atomic_read(&VFS_I(ip)->i_count))
+		xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL);
 	ASSERT(new_size <= XFS_ISIZE(ip));
 	ASSERT(tp->t_flags & XFS_TRANS_PERM_LOG_RES);
 	ASSERT(ip->i_itemp != NULL);
@@ -1449,7 +1603,7 @@ xfs_release(
 	if (!xfs_ilock_nowait(ip, XFS_IOLOCK_EXCL))
 		return 0;
 
-	if (xfs_can_free_eofblocks(ip, false)) {
+	if (xfs_can_free_eofblocks(ip)) {
 		/*
 		 * Check if the inode is being opened, written and closed
 		 * frequently and we have delayed allocation blocks outstanding
@@ -1482,6 +1636,51 @@ out_unlock:
 }
 
 /*
+ * Mark all the buffers attached to this directory stale.  In theory we should
+ * never be freeing a directory with any blocks at all, but this covers the
+ * case where we've recovered a directory swap with a "temporary" directory
+ * created by online repair and now need to dump it.
+ */
+STATIC void
+xfs_inactive_dir(
+	struct xfs_inode	*dp)
+{
+	struct xfs_iext_cursor	icur;
+	struct xfs_bmbt_irec	got;
+	struct xfs_mount	*mp = dp->i_mount;
+	struct xfs_da_geometry	*geo = mp->m_dir_geo;
+	struct xfs_ifork	*ifp = xfs_ifork_ptr(dp, XFS_DATA_FORK);
+	xfs_fileoff_t		off;
+
+	/*
+	 * Invalidate each directory block.  All directory blocks are of
+	 * fsbcount length and alignment, so we only need to walk those same
+	 * offsets.  We hold the only reference to this inode, so we must wait
+	 * for the buffer locks.
+	 */
+	for_each_xfs_iext(ifp, &icur, &got) {
+		for (off = round_up(got.br_startoff, geo->fsbcount);
+		     off < got.br_startoff + got.br_blockcount;
+		     off += geo->fsbcount) {
+			struct xfs_buf	*bp = NULL;
+			xfs_fsblock_t	fsbno;
+			int		error;
+
+			fsbno = (off - got.br_startoff) + got.br_startblock;
+			error = xfs_buf_incore(mp->m_ddev_targp,
+					XFS_FSB_TO_DADDR(mp, fsbno),
+					XFS_FSB_TO_BB(mp, geo->fsbcount),
+					XBF_LIVESCAN, &bp);
+			if (error)
+				continue;
+
+			xfs_buf_stale(bp);
+			xfs_buf_relse(bp);
+		}
+	}
+}
+
+/*
  * xfs_inactive_truncate
  *
  * Called to perform a truncate when an inode becomes unlinked.
@@ -1596,7 +1795,7 @@ xfs_inactive_ifree(
 	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
 
 	error = xfs_ifree(tp, ip);
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 	if (error) {
 		/*
 		 * If we fail to free the inode, shut down.  The cancel
@@ -1665,15 +1864,46 @@ xfs_inode_needs_inactive(
 
 	/*
 	 * This file isn't being freed, so check if there are post-eof blocks
-	 * to free.  @force is true because we are evicting an inode from the
-	 * cache.  Post-eof blocks must be freed, lest we end up with broken
-	 * free space accounting.
+	 * to free.
 	 *
 	 * Note: don't bother with iolock here since lockdep complains about
 	 * acquiring it in reclaim context. We have the only reference to the
 	 * inode at this point anyways.
 	 */
-	return xfs_can_free_eofblocks(ip, true);
+	return xfs_can_free_eofblocks(ip);
+}
+
+/*
+ * Save health status somewhere, if we're dumping an inode with uncorrected
+ * errors and online repair isn't running.
+ */
+static inline void
+xfs_inactive_health(
+	struct xfs_inode	*ip)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_perag	*pag;
+	unsigned int		sick;
+	unsigned int		checked;
+
+	xfs_inode_measure_sickness(ip, &sick, &checked);
+	if (!sick)
+		return;
+
+	trace_xfs_inode_unfixed_corruption(ip, sick);
+
+	if (sick & XFS_SICK_INO_FORGET)
+		return;
+
+	pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino));
+	if (!pag) {
+		/* There had better still be a perag structure! */
+		ASSERT(0);
+		return;
+	}
+
+	xfs_ag_mark_sick(pag, XFS_SICK_AG_INODES);
+	xfs_perag_put(pag);
 }
 
 /*
@@ -1704,6 +1934,8 @@ xfs_inactive(
 	mp = ip->i_mount;
 	ASSERT(!xfs_iflags_test(ip, XFS_IRECOVERY));
 
+	xfs_inactive_health(ip);
+
 	/*
 	 * If this is a read-only mount, don't do this (would generate I/O)
 	 * unless we're in log recovery and cleaning the iunlinked list.
@@ -1721,15 +1953,11 @@ xfs_inactive(
 
 	if (VFS_I(ip)->i_nlink != 0) {
 		/*
-		 * force is true because we are evicting an inode from the
-		 * cache. Post-eof blocks must be freed, lest we end up with
-		 * broken free space accounting.
-		 *
 		 * Note: don't bother with iolock here since lockdep complains
 		 * about acquiring it in reclaim context. We have the only
 		 * reference to the inode at this point anyways.
 		 */
-		if (xfs_can_free_eofblocks(ip, true))
+		if (xfs_can_free_eofblocks(ip))
 			error = xfs_free_eofblocks(ip);
 
 		goto out;
@@ -1756,6 +1984,11 @@ xfs_inactive(
 			goto out;
 	}
 
+	if (S_ISDIR(VFS_I(ip)->i_mode) && ip->i_df.if_nextents > 0) {
+		xfs_inactive_dir(ip);
+		truncate = 1;
+	}
+
 	if (S_ISLNK(VFS_I(ip)->i_mode))
 		error = xfs_inactive_symlink(ip);
 	else if (truncate)
@@ -1829,7 +2062,7 @@ out:
  * only unlinked, referenced inodes can be on the unlinked inode list.  If we
  * don't find the inode in cache, then let the caller handle the situation.
  */
-static struct xfs_inode *
+struct xfs_inode *
 xfs_iunlink_lookup(
 	struct xfs_perag	*pag,
 	xfs_agino_t		agino)
@@ -1910,6 +2143,7 @@ xfs_iunlink_update_bucket(
 	 */
 	if (old_value == new_agino) {
 		xfs_buf_mark_corrupt(agibp);
+		xfs_ag_mark_sick(pag, XFS_SICK_AG_AGI);
 		return -EFSCORRUPTED;
 	}
 
@@ -1959,11 +2193,14 @@ xfs_iunlink_reload_next(
 	 */
 	ino = XFS_AGINO_TO_INO(mp, pag->pag_agno, next_agino);
 	error = xfs_iget(mp, tp, ino, XFS_IGET_UNTRUSTED, 0, &next_ip);
-	if (error)
+	if (error) {
+		xfs_ag_mark_sick(pag, XFS_SICK_AG_AGI);
 		return error;
+	}
 
 	/* If this is not an unlinked inode, something is very wrong. */
 	if (VFS_I(next_ip)->i_nlink != 0) {
+		xfs_ag_mark_sick(pag, XFS_SICK_AG_AGI);
 		error = -EFSCORRUPTED;
 		goto rele;
 	}
@@ -2001,6 +2238,7 @@ xfs_iunlink_insert_inode(
 	if (next_agino == agino ||
 	    !xfs_verify_agino_or_null(pag, next_agino)) {
 		xfs_buf_mark_corrupt(agibp);
+		xfs_ag_mark_sick(pag, XFS_SICK_AG_AGI);
 		return -EFSCORRUPTED;
 	}
 
@@ -2037,7 +2275,7 @@ xfs_iunlink_insert_inode(
  * We place the on-disk inode on a list in the AGI.  It will be pulled from this
  * list when the inode is freed.
  */
-STATIC int
+int
 xfs_iunlink(
 	struct xfs_trans	*tp,
 	struct xfs_inode	*ip)
@@ -2054,7 +2292,7 @@ xfs_iunlink(
 	pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino));
 
 	/* Get the agi buffer first.  It ensures lock ordering on the list. */
-	error = xfs_read_agi(pag, tp, &agibp);
+	error = xfs_read_agi(pag, tp, 0, &agibp);
 	if (error)
 		goto out;
 
@@ -2088,6 +2326,7 @@ xfs_iunlink_remove_inode(
 	if (!xfs_verify_agino(pag, head_agino)) {
 		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
 				agi, sizeof(*agi));
+		xfs_ag_mark_sick(pag, XFS_SICK_AG_AGI);
 		return -EFSCORRUPTED;
 	}
 
@@ -2116,8 +2355,10 @@ xfs_iunlink_remove_inode(
 		struct xfs_inode	*prev_ip;
 
 		prev_ip = xfs_iunlink_lookup(pag, ip->i_prev_unlinked);
-		if (!prev_ip)
+		if (!prev_ip) {
+			xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE);
 			return -EFSCORRUPTED;
+		}
 
 		error = xfs_iunlink_log_inode(tp, prev_ip, pag,
 				ip->i_next_unlinked);
@@ -2136,7 +2377,7 @@ xfs_iunlink_remove_inode(
 /*
  * Pull the on-disk inode from the AGI unlinked list.
  */
-STATIC int
+int
 xfs_iunlink_remove(
 	struct xfs_trans	*tp,
 	struct xfs_perag	*pag,
@@ -2148,7 +2389,7 @@ xfs_iunlink_remove(
 	trace_xfs_iunlink_remove(ip);
 
 	/* Get the agi buffer first.  It ensures lock ordering on the list. */
-	error = xfs_read_agi(pag, tp, &agibp);
+	error = xfs_read_agi(pag, tp, 0, &agibp);
 	if (error)
 		return error;
 
@@ -2309,11 +2550,26 @@ xfs_ifree_cluster(
 		 * This buffer may not have been correctly initialised as we
 		 * didn't read it from disk. That's not important because we are
 		 * only using to mark the buffer as stale in the log, and to
-		 * attach stale cached inodes on it. That means it will never be
-		 * dispatched for IO. If it is, we want to know about it, and we
-		 * want it to fail. We can acheive this by adding a write
-		 * verifier to the buffer.
+		 * attach stale cached inodes on it.
+		 *
+		 * For the inode that triggered the cluster freeing, this
+		 * attachment may occur in xfs_inode_item_precommit() after we
+		 * have marked this buffer stale.  If this buffer was not in
+		 * memory before xfs_ifree_cluster() started, it will not be
+		 * marked XBF_DONE and this will cause problems later in
+		 * xfs_inode_item_precommit() when we trip over a (stale, !done)
+		 * buffer to attached to the transaction.
+		 *
+		 * Hence we have to mark the buffer as XFS_DONE here. This is
+		 * safe because we are also marking the buffer as XBF_STALE and
+		 * XFS_BLI_STALE. That means it will never be dispatched for
+		 * IO and it won't be unlocked until the cluster freeing has
+		 * been committed to the journal and the buffer unpinned. If it
+		 * is written, we want to know about it, and we want it to
+		 * fail. We can acheive this by adding a write verifier to the
+		 * buffer.
 		 */
+		bp->b_flags |= XBF_DONE;
 		bp->b_ops = &xfs_inode_buf_ops;
 
 		/*
@@ -2350,7 +2606,7 @@ xfs_ifree(
 	struct xfs_inode_log_item *iip = ip->i_itemp;
 	int			error;
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 	ASSERT(VFS_I(ip)->i_nlink == 0);
 	ASSERT(ip->i_df.if_nextents == 0);
 	ASSERT(ip->i_disk_size == 0 || !S_ISREG(VFS_I(ip)->i_mode));
@@ -2378,7 +2634,7 @@ xfs_ifree(
 	 * already been freed by xfs_attr_inactive.
 	 */
 	if (ip->i_df.if_format == XFS_DINODE_FMT_LOCAL) {
-		kmem_free(ip->i_df.if_data);
+		kfree(ip->i_df.if_data);
 		ip->i_df.if_data = NULL;
 		ip->i_df.if_bytes = 0;
 	}
@@ -2419,7 +2675,7 @@ static void
 xfs_iunpin(
 	struct xfs_inode	*ip)
 {
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL | XFS_ILOCK_SHARED);
 
 	trace_xfs_inode_unpin_nowait(ip, _RET_IP_);
 
@@ -2482,16 +2738,17 @@ xfs_iunpin_wait(
  */
 int
 xfs_remove(
-	xfs_inode_t             *dp,
+	struct xfs_inode	*dp,
 	struct xfs_name		*name,
-	xfs_inode_t		*ip)
+	struct xfs_inode	*ip)
 {
-	xfs_mount_t		*mp = dp->i_mount;
-	xfs_trans_t             *tp = NULL;
+	struct xfs_mount	*mp = dp->i_mount;
+	struct xfs_trans	*tp = NULL;
 	int			is_dir = S_ISDIR(VFS_I(ip)->i_mode);
 	int			dontcare;
 	int                     error = 0;
 	uint			resblks;
+	struct xfs_parent_args	*ppargs;
 
 	trace_xfs_remove(dp, name);
 
@@ -2508,6 +2765,10 @@ xfs_remove(
 	if (error)
 		goto std_return;
 
+	error = xfs_parent_start(mp, &ppargs);
+	if (error)
+		goto std_return;
+
 	/*
 	 * We try to get the real space reservation first, allowing for
 	 * directory btree deletion(s) implying possible bmap insert(s).  If we
@@ -2519,12 +2780,12 @@ xfs_remove(
 	 * the directory code can handle a reservationless update and we don't
 	 * want to prevent a user from trying to free space by deleting things.
 	 */
-	resblks = XFS_REMOVE_SPACE_RES(mp);
+	resblks = xfs_remove_space_res(mp, name->len);
 	error = xfs_trans_alloc_dir(dp, &M_RES(mp)->tr_remove, ip, &resblks,
 			&tp, &dontcare);
 	if (error) {
 		ASSERT(error != -ENOSPC);
-		goto std_return;
+		goto out_parent;
 	}
 
 	/*
@@ -2584,6 +2845,19 @@ xfs_remove(
 		goto out_trans_cancel;
 	}
 
+	/* Remove parent pointer. */
+	if (ppargs) {
+		error = xfs_parent_removename(tp, ppargs, dp, name, ip);
+		if (error)
+			goto out_trans_cancel;
+	}
+
+	/*
+	 * Drop the link from dp to ip, and if ip was a directory, remove the
+	 * '.' and '..' references since we freed the directory.
+	 */
+	xfs_dir_update_hook(dp, ip, -1, name);
+
 	/*
 	 * If this is a synchronous mount, make sure that the
 	 * remove transaction goes to disk before returning to
@@ -2594,19 +2868,42 @@ xfs_remove(
 
 	error = xfs_trans_commit(tp);
 	if (error)
-		goto std_return;
+		goto out_unlock;
 
 	if (is_dir && xfs_inode_is_filestream(ip))
 		xfs_filestream_deassociate(ip);
 
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	xfs_iunlock(dp, XFS_ILOCK_EXCL);
+	xfs_parent_finish(mp, ppargs);
 	return 0;
 
  out_trans_cancel:
 	xfs_trans_cancel(tp);
+ out_unlock:
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	xfs_iunlock(dp, XFS_ILOCK_EXCL);
+ out_parent:
+	xfs_parent_finish(mp, ppargs);
  std_return:
 	return error;
 }
 
+static inline void
+xfs_iunlock_rename(
+	struct xfs_inode	**i_tab,
+	int			num_inodes)
+{
+	int			i;
+
+	for (i = num_inodes - 1; i >= 0; i--) {
+		/* Skip duplicate inodes if src and target dps are the same */
+		if (!i_tab[i] || (i > 0 && i_tab[i] == i_tab[i - 1]))
+			continue;
+		xfs_iunlock(i_tab[i], XFS_ILOCK_EXCL);
+	}
+}
+
 /*
  * Enter all inodes for a rename transaction into a sorted array.
  */
@@ -2621,7 +2918,7 @@ xfs_sort_for_rename(
 	struct xfs_inode	**i_tab,/* out: sorted array of inodes */
 	int			*num_inodes)  /* in/out: inodes in array */
 {
-	int			i, j;
+	int			i;
 
 	ASSERT(*num_inodes == __XFS_SORT_INODES);
 	memset(i_tab, 0, *num_inodes * sizeof(struct xfs_inode *));
@@ -2643,17 +2940,26 @@ xfs_sort_for_rename(
 		i_tab[i++] = wip;
 	*num_inodes = i;
 
+	xfs_sort_inodes(i_tab, *num_inodes);
+}
+
+void
+xfs_sort_inodes(
+	struct xfs_inode	**i_tab,
+	unsigned int		num_inodes)
+{
+	int			i, j;
+
+	ASSERT(num_inodes <= __XFS_SORT_INODES);
+
 	/*
 	 * Sort the elements via bubble sort.  (Remember, there are at
 	 * most 5 elements to sort, so this is adequate.)
 	 */
-	for (i = 0; i < *num_inodes; i++) {
-		for (j = 1; j < *num_inodes; j++) {
-			if (i_tab[j]->i_ino < i_tab[j-1]->i_ino) {
-				struct xfs_inode *temp = i_tab[j];
-				i_tab[j] = i_tab[j-1];
-				i_tab[j-1] = temp;
-			}
+	for (i = 0; i < num_inodes; i++) {
+		for (j = 1; j < num_inodes; j++) {
+			if (i_tab[j]->i_ino < i_tab[j-1]->i_ino)
+				swap(i_tab[j], i_tab[j - 1]);
 		}
 	}
 }
@@ -2683,15 +2989,17 @@ xfs_cross_rename(
 	struct xfs_inode	*dp1,
 	struct xfs_name		*name1,
 	struct xfs_inode	*ip1,
+	struct xfs_parent_args	*ip1_ppargs,
 	struct xfs_inode	*dp2,
 	struct xfs_name		*name2,
 	struct xfs_inode	*ip2,
+	struct xfs_parent_args	*ip2_ppargs,
 	int			spaceres)
 {
-	int		error = 0;
-	int		ip1_flags = 0;
-	int		ip2_flags = 0;
-	int		dp2_flags = 0;
+	int			error = 0;
+	int			ip1_flags = 0;
+	int			ip2_flags = 0;
+	int			dp2_flags = 0;
 
 	/* Swap inode number for dirent in first parent */
 	error = xfs_dir_replace(tp, dp1, name1, ip2->i_ino, spaceres);
@@ -2760,6 +3068,21 @@ xfs_cross_rename(
 		}
 	}
 
+	/* Schedule parent pointer replacements */
+	if (ip1_ppargs) {
+		error = xfs_parent_replacename(tp, ip1_ppargs, dp1, name1, dp2,
+				name2, ip1);
+		if (error)
+			goto out_trans_abort;
+	}
+
+	if (ip2_ppargs) {
+		error = xfs_parent_replacename(tp, ip2_ppargs, dp2, name2, dp1,
+				name1, ip2);
+		if (error)
+			goto out_trans_abort;
+	}
+
 	if (ip1_flags) {
 		xfs_trans_ichgtime(tp, ip1, ip1_flags);
 		xfs_trans_log_inode(tp, ip1, XFS_ILOG_CORE);
@@ -2774,6 +3097,20 @@ xfs_cross_rename(
 	}
 	xfs_trans_ichgtime(tp, dp1, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
 	xfs_trans_log_inode(tp, dp1, XFS_ILOG_CORE);
+
+	/*
+	 * Inform our hook clients that we've finished an exchange operation as
+	 * follows: removed the source and target files from their directories;
+	 * added the target to the source directory; and added the source to
+	 * the target directory.  All inodes are locked, so it's ok to model a
+	 * rename this way so long as we say we deleted entries before we add
+	 * new ones.
+	 */
+	xfs_dir_update_hook(dp1, ip1, -1, name1);
+	xfs_dir_update_hook(dp2, ip2, -1, name2);
+	xfs_dir_update_hook(dp1, ip2, 1, name1);
+	xfs_dir_update_hook(dp2, ip1, 1, name2);
+
 	return xfs_finish_rename(tp);
 
 out_trans_abort:
@@ -2801,7 +3138,7 @@ xfs_rename_alloc_whiteout(
 	int			error;
 
 	error = xfs_create_tmpfile(idmap, dp, S_IFCHR | WHITEOUT_MODE,
-				   &tmpfile);
+			xfs_has_parent(dp->i_mount), &tmpfile);
 	if (error)
 		return error;
 
@@ -2845,6 +3182,9 @@ xfs_rename(
 	struct xfs_trans	*tp;
 	struct xfs_inode	*wip = NULL;		/* whiteout inode */
 	struct xfs_inode	*inodes[__XFS_SORT_INODES];
+	struct xfs_parent_args	*src_ppargs = NULL;
+	struct xfs_parent_args	*tgt_ppargs = NULL;
+	struct xfs_parent_args	*wip_ppargs = NULL;
 	int			i;
 	int			num_inodes = __XFS_SORT_INODES;
 	bool			new_parent = (src_dp != target_dp);
@@ -2876,9 +3216,26 @@ xfs_rename(
 	xfs_sort_for_rename(src_dp, target_dp, src_ip, target_ip, wip,
 				inodes, &num_inodes);
 
+	error = xfs_parent_start(mp, &src_ppargs);
+	if (error)
+		goto out_release_wip;
+
+	if (wip) {
+		error = xfs_parent_start(mp, &wip_ppargs);
+		if (error)
+			goto out_src_ppargs;
+	}
+
+	if (target_ip) {
+		error = xfs_parent_start(mp, &tgt_ppargs);
+		if (error)
+			goto out_wip_ppargs;
+	}
+
 retry:
 	nospace_error = 0;
-	spaceres = XFS_RENAME_SPACE_RES(mp, target_name->len);
+	spaceres = xfs_rename_space_res(mp, src_name->len, target_ip != NULL,
+			target_name->len, wip != NULL);
 	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_rename, spaceres, 0, 0, &tp);
 	if (error == -ENOSPC) {
 		nospace_error = error;
@@ -2887,14 +3244,26 @@ retry:
 				&tp);
 	}
 	if (error)
-		goto out_release_wip;
+		goto out_tgt_ppargs;
+
+	/*
+	 * We don't allow reservationless renaming when parent pointers are
+	 * enabled because we can't back out if the xattrs must grow.
+	 */
+	if (src_ppargs && nospace_error) {
+		error = nospace_error;
+		xfs_trans_cancel(tp);
+		goto out_tgt_ppargs;
+	}
 
 	/*
 	 * Attach the dquots to the inodes
 	 */
 	error = xfs_qm_vop_rename_dqattach(inodes);
-	if (error)
-		goto out_trans_cancel;
+	if (error) {
+		xfs_trans_cancel(tp);
+		goto out_tgt_ppargs;
+	}
 
 	/*
 	 * Lock all the participating inodes. Depending upon whether
@@ -2905,18 +3274,16 @@ retry:
 	xfs_lock_inodes(inodes, num_inodes, XFS_ILOCK_EXCL);
 
 	/*
-	 * Join all the inodes to the transaction. From this point on,
-	 * we can rely on either trans_commit or trans_cancel to unlock
-	 * them.
+	 * Join all the inodes to the transaction.
 	 */
-	xfs_trans_ijoin(tp, src_dp, XFS_ILOCK_EXCL);
+	xfs_trans_ijoin(tp, src_dp, 0);
 	if (new_parent)
-		xfs_trans_ijoin(tp, target_dp, XFS_ILOCK_EXCL);
-	xfs_trans_ijoin(tp, src_ip, XFS_ILOCK_EXCL);
+		xfs_trans_ijoin(tp, target_dp, 0);
+	xfs_trans_ijoin(tp, src_ip, 0);
 	if (target_ip)
-		xfs_trans_ijoin(tp, target_ip, XFS_ILOCK_EXCL);
+		xfs_trans_ijoin(tp, target_ip, 0);
 	if (wip)
-		xfs_trans_ijoin(tp, wip, XFS_ILOCK_EXCL);
+		xfs_trans_ijoin(tp, wip, 0);
 
 	/*
 	 * If we are using project inheritance, we only allow renames
@@ -2930,10 +3297,13 @@ retry:
 	}
 
 	/* RENAME_EXCHANGE is unique from here on. */
-	if (flags & RENAME_EXCHANGE)
-		return xfs_cross_rename(tp, src_dp, src_name, src_ip,
-					target_dp, target_name, target_ip,
-					spaceres);
+	if (flags & RENAME_EXCHANGE) {
+		error = xfs_cross_rename(tp, src_dp, src_name, src_ip,
+				src_ppargs, target_dp, target_name, target_ip,
+				tgt_ppargs, spaceres);
+		nospace_error = 0;
+		goto out_unlock;
+	}
 
 	/*
 	 * Try to reserve quota to handle an expansion of the target directory.
@@ -2947,6 +3317,7 @@ retry:
 		if (error == -EDQUOT || error == -ENOSPC) {
 			if (!retried) {
 				xfs_trans_cancel(tp);
+				xfs_iunlock_rename(inodes, num_inodes);
 				xfs_blockgc_free_quota(target_dp, 0);
 				retried = true;
 				goto retry;
@@ -2961,6 +3332,15 @@ retry:
 	}
 
 	/*
+	 * We don't allow quotaless renaming when parent pointers are enabled
+	 * because we can't back out if the xattrs must grow.
+	 */
+	if (src_ppargs && nospace_error) {
+		error = nospace_error;
+		goto out_trans_cancel;
+	}
+
+	/*
 	 * Check for expected errors before we dirty the transaction
 	 * so we can return an error without a transaction abort.
 	 */
@@ -3006,7 +3386,7 @@ retry:
 
 			pag = xfs_perag_get(mp,
 					XFS_INO_TO_AGNO(mp, inodes[i]->i_ino));
-			error = xfs_read_agi(pag, tp, &bp);
+			error = xfs_read_agi(pag, tp, 0, &bp);
 			xfs_perag_put(pag);
 			if (error)
 				goto out_trans_cancel;
@@ -3152,18 +3532,62 @@ retry:
 	if (error)
 		goto out_trans_cancel;
 
+	/* Schedule parent pointer updates. */
+	if (wip_ppargs) {
+		error = xfs_parent_addname(tp, wip_ppargs, src_dp, src_name,
+				wip);
+		if (error)
+			goto out_trans_cancel;
+	}
+
+	if (src_ppargs) {
+		error = xfs_parent_replacename(tp, src_ppargs, src_dp,
+				src_name, target_dp, target_name, src_ip);
+		if (error)
+			goto out_trans_cancel;
+	}
+
+	if (tgt_ppargs) {
+		error = xfs_parent_removename(tp, tgt_ppargs, target_dp,
+				target_name, target_ip);
+		if (error)
+			goto out_trans_cancel;
+	}
+
 	xfs_trans_ichgtime(tp, src_dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
 	xfs_trans_log_inode(tp, src_dp, XFS_ILOG_CORE);
 	if (new_parent)
 		xfs_trans_log_inode(tp, target_dp, XFS_ILOG_CORE);
 
-	error = xfs_finish_rename(tp);
+	/*
+	 * Inform our hook clients that we've finished a rename operation as
+	 * follows: removed the source and target files from their directories;
+	 * that we've added the source to the target directory; and finally
+	 * that we've added the whiteout, if there was one.  All inodes are
+	 * locked, so it's ok to model a rename this way so long as we say we
+	 * deleted entries before we add new ones.
+	 */
+	if (target_ip)
+		xfs_dir_update_hook(target_dp, target_ip, -1, target_name);
+	xfs_dir_update_hook(src_dp, src_ip, -1, src_name);
+	xfs_dir_update_hook(target_dp, src_ip, 1, target_name);
 	if (wip)
-		xfs_irele(wip);
-	return error;
+		xfs_dir_update_hook(src_dp, wip, 1, src_name);
+
+	error = xfs_finish_rename(tp);
+	nospace_error = 0;
+	goto out_unlock;
 
 out_trans_cancel:
 	xfs_trans_cancel(tp);
+out_unlock:
+	xfs_iunlock_rename(inodes, num_inodes);
+out_tgt_ppargs:
+	xfs_parent_finish(mp, tgt_ppargs);
+out_wip_ppargs:
+	xfs_parent_finish(mp, wip_ppargs);
+out_src_ppargs:
+	xfs_parent_finish(mp, src_ppargs);
 out_release_wip:
 	if (wip)
 		xfs_irele(wip);
@@ -3182,7 +3606,7 @@ xfs_iflush(
 	struct xfs_mount	*mp = ip->i_mount;
 	int			error;
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL | XFS_ILOCK_SHARED);
 	ASSERT(xfs_iflags_test(ip, XFS_IFLUSHING));
 	ASSERT(ip->i_df.if_format != XFS_DINODE_FMT_BTREE ||
 	       ip->i_df.if_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK));
@@ -3317,6 +3741,8 @@ flush_out:
 
 	/* generate the checksum. */
 	xfs_dinode_calc_crc(mp, dip);
+	if (error)
+		xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE);
 	return error;
 }
 
@@ -3661,7 +4087,7 @@ xfs_inode_reload_unlinked_bucket(
 
 	/* Grab the first inode in the list */
 	pag = xfs_perag_get(mp, agno);
-	error = xfs_ialloc_read_agi(pag, tp, &agibp);
+	error = xfs_ialloc_read_agi(pag, tp, 0, &agibp);
 	xfs_perag_put(pag);
 	if (error)
 		return error;
@@ -3777,3 +4203,93 @@ xfs_ifork_zapped(
 		return false;
 	}
 }
+
+/* Compute the number of data and realtime blocks used by a file. */
+void
+xfs_inode_count_blocks(
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip,
+	xfs_filblks_t		*dblocks,
+	xfs_filblks_t		*rblocks)
+{
+	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, XFS_DATA_FORK);
+
+	*rblocks = 0;
+	if (XFS_IS_REALTIME_INODE(ip))
+		xfs_bmap_count_leaves(ifp, rblocks);
+	*dblocks = ip->i_nblocks - *rblocks;
+}
+
+static void
+xfs_wait_dax_page(
+	struct inode		*inode)
+{
+	struct xfs_inode        *ip = XFS_I(inode);
+
+	xfs_iunlock(ip, XFS_MMAPLOCK_EXCL);
+	schedule();
+	xfs_ilock(ip, XFS_MMAPLOCK_EXCL);
+}
+
+int
+xfs_break_dax_layouts(
+	struct inode		*inode,
+	bool			*retry)
+{
+	struct page		*page;
+
+	xfs_assert_ilocked(XFS_I(inode), XFS_MMAPLOCK_EXCL);
+
+	page = dax_layout_busy_page(inode->i_mapping);
+	if (!page)
+		return 0;
+
+	*retry = true;
+	return ___wait_var_event(&page->_refcount,
+			atomic_read(&page->_refcount) == 1, TASK_INTERRUPTIBLE,
+			0, 0, xfs_wait_dax_page(inode));
+}
+
+int
+xfs_break_layouts(
+	struct inode		*inode,
+	uint			*iolock,
+	enum layout_break_reason reason)
+{
+	bool			retry;
+	int			error;
+
+	xfs_assert_ilocked(XFS_I(inode), XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL);
+
+	do {
+		retry = false;
+		switch (reason) {
+		case BREAK_UNMAP:
+			error = xfs_break_dax_layouts(inode, &retry);
+			if (error || retry)
+				break;
+			fallthrough;
+		case BREAK_WRITE:
+			error = xfs_break_leased_layouts(inode, iolock, &retry);
+			break;
+		default:
+			WARN_ON_ONCE(1);
+			error = -EINVAL;
+		}
+	} while (error == 0 && retry);
+
+	return error;
+}
+
+/* Returns the size of fundamental allocation unit for a file, in bytes. */
+unsigned int
+xfs_inode_alloc_unitsize(
+	struct xfs_inode	*ip)
+{
+	unsigned int		blocks = 1;
+
+	if (XFS_IS_REALTIME_INODE(ip))
+		blocks = ip->i_mount->m_sb.sb_rextsize;
+
+	return XFS_FSB_TO_B(ip->i_mount, blocks);
+}
diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
index 97f63bacd4..292b90b5f2 100644
--- a/fs/xfs/xfs_inode.h
+++ b/fs/xfs/xfs_inode.h
@@ -39,7 +39,7 @@ typedef struct xfs_inode {
 
 	/* Transaction and locking information. */
 	struct xfs_inode_log_item *i_itemp;	/* logging information */
-	mrlock_t		i_lock;		/* inode lock */
+	struct rw_semaphore	i_lock;		/* inode lock */
 	atomic_t		i_pincount;	/* inode pin count */
 	struct llist_node	i_gclist;	/* deferred inactivation list */
 
@@ -171,6 +171,12 @@ static inline struct inode *VFS_I(struct xfs_inode *ip)
 	return &ip->i_vnode;
 }
 
+/* convert from const xfs inode to const vfs inode */
+static inline const struct inode *VFS_IC(const struct xfs_inode *ip)
+{
+	return &ip->i_vnode;
+}
+
 /*
  * For regular files we only update the on-disk filesize when actually
  * writing data back to disk.  Until then only the copy in the VFS inode
@@ -201,13 +207,13 @@ xfs_new_eof(struct xfs_inode *ip, xfs_fsize_t new_size)
  * i_flags helper functions
  */
 static inline void
-__xfs_iflags_set(xfs_inode_t *ip, unsigned short flags)
+__xfs_iflags_set(xfs_inode_t *ip, unsigned long flags)
 {
 	ip->i_flags |= flags;
 }
 
 static inline void
-xfs_iflags_set(xfs_inode_t *ip, unsigned short flags)
+xfs_iflags_set(xfs_inode_t *ip, unsigned long flags)
 {
 	spin_lock(&ip->i_flags_lock);
 	__xfs_iflags_set(ip, flags);
@@ -215,7 +221,7 @@ xfs_iflags_set(xfs_inode_t *ip, unsigned short flags)
 }
 
 static inline void
-xfs_iflags_clear(xfs_inode_t *ip, unsigned short flags)
+xfs_iflags_clear(xfs_inode_t *ip, unsigned long flags)
 {
 	spin_lock(&ip->i_flags_lock);
 	ip->i_flags &= ~flags;
@@ -223,13 +229,13 @@ xfs_iflags_clear(xfs_inode_t *ip, unsigned short flags)
 }
 
 static inline int
-__xfs_iflags_test(xfs_inode_t *ip, unsigned short flags)
+__xfs_iflags_test(xfs_inode_t *ip, unsigned long flags)
 {
 	return (ip->i_flags & flags);
 }
 
 static inline int
-xfs_iflags_test(xfs_inode_t *ip, unsigned short flags)
+xfs_iflags_test(xfs_inode_t *ip, unsigned long flags)
 {
 	int ret;
 	spin_lock(&ip->i_flags_lock);
@@ -239,7 +245,7 @@ xfs_iflags_test(xfs_inode_t *ip, unsigned short flags)
 }
 
 static inline int
-xfs_iflags_test_and_clear(xfs_inode_t *ip, unsigned short flags)
+xfs_iflags_test_and_clear(xfs_inode_t *ip, unsigned long flags)
 {
 	int ret;
 
@@ -252,7 +258,7 @@ xfs_iflags_test_and_clear(xfs_inode_t *ip, unsigned short flags)
 }
 
 static inline int
-xfs_iflags_test_and_set(xfs_inode_t *ip, unsigned short flags)
+xfs_iflags_test_and_set(xfs_inode_t *ip, unsigned long flags)
 {
 	int ret;
 
@@ -306,6 +312,15 @@ static inline bool xfs_inode_has_large_extent_counts(struct xfs_inode *ip)
 }
 
 /*
+ * Decide if this file is a realtime file whose data allocation unit is larger
+ * than a single filesystem block.
+ */
+static inline bool xfs_inode_has_bigrtalloc(struct xfs_inode *ip)
+{
+	return XFS_IS_REALTIME_INODE(ip) && ip->i_mount->m_sb.sb_rextsize > 1;
+}
+
+/*
  * Return the buftarg used for data allocations on a given inode.
  */
 #define xfs_inode_buftarg(ip) \
@@ -507,7 +522,7 @@ int		xfs_create(struct mnt_idmap *idmap,
 			   umode_t mode, dev_t rdev, bool need_xattr,
 			   struct xfs_inode **ipp);
 int		xfs_create_tmpfile(struct mnt_idmap *idmap,
-			   struct xfs_inode *dp, umode_t mode,
+			   struct xfs_inode *dp, umode_t mode, bool init_xattrs,
 			   struct xfs_inode **ipp);
 int		xfs_remove(struct xfs_inode *dp, struct xfs_name *name,
 			   struct xfs_inode *ip);
@@ -523,7 +538,7 @@ void		xfs_ilock(xfs_inode_t *, uint);
 int		xfs_ilock_nowait(xfs_inode_t *, uint);
 void		xfs_iunlock(xfs_inode_t *, uint);
 void		xfs_ilock_demote(xfs_inode_t *, uint);
-bool		xfs_isilocked(struct xfs_inode *, uint);
+void		xfs_assert_ilocked(struct xfs_inode *, uint);
 uint		xfs_ilock_data_map_shared(struct xfs_inode *);
 uint		xfs_ilock_attr_map_shared(struct xfs_inode *);
 
@@ -559,16 +574,10 @@ xfs_itruncate_extents(
 	return xfs_itruncate_extents_flags(tpp, ip, whichfork, new_size, 0);
 }
 
-/* from xfs_file.c */
 int	xfs_break_dax_layouts(struct inode *inode, bool *retry);
 int	xfs_break_layouts(struct inode *inode, uint *iolock,
 		enum layout_break_reason reason);
 
-/* from xfs_iops.c */
-extern void xfs_setup_inode(struct xfs_inode *ip);
-extern void xfs_setup_iops(struct xfs_inode *ip);
-extern void xfs_diflags_to_iflags(struct xfs_inode *ip, bool init);
-
 static inline void xfs_update_stable_writes(struct xfs_inode *ip)
 {
 	if (bdev_stable_writes(xfs_inode_buftarg(ip)->bt_bdev))
@@ -607,11 +616,20 @@ extern struct kmem_cache	*xfs_inode_cache;
 
 bool xfs_inode_needs_inactive(struct xfs_inode *ip);
 
+int xfs_iunlink(struct xfs_trans *tp, struct xfs_inode *ip);
+int xfs_iunlink_remove(struct xfs_trans *tp, struct xfs_perag *pag,
+		struct xfs_inode *ip);
+struct xfs_inode *xfs_iunlink_lookup(struct xfs_perag *pag, xfs_agino_t agino);
+
 void xfs_end_io(struct work_struct *work);
 
 int xfs_ilock2_io_mmap(struct xfs_inode *ip1, struct xfs_inode *ip2);
 void xfs_iunlock2_io_mmap(struct xfs_inode *ip1, struct xfs_inode *ip2);
 void xfs_iunlock2_remapping(struct xfs_inode *ip1, struct xfs_inode *ip2);
+int xfs_droplink(struct xfs_trans *tp, struct xfs_inode *ip);
+void xfs_bumplink(struct xfs_trans *tp, struct xfs_inode *ip);
+void xfs_lock_inodes(struct xfs_inode **ips, int inodes, uint lock_mode);
+void xfs_sort_inodes(struct xfs_inode **i_tab, unsigned int num_inodes);
 
 static inline bool
 xfs_inode_unlinked_incomplete(
@@ -623,5 +641,33 @@ int xfs_inode_reload_unlinked_bucket(struct xfs_trans *tp, struct xfs_inode *ip)
 int xfs_inode_reload_unlinked(struct xfs_inode *ip);
 
 bool xfs_ifork_zapped(const struct xfs_inode *ip, int whichfork);
+void xfs_inode_count_blocks(struct xfs_trans *tp, struct xfs_inode *ip,
+		xfs_filblks_t *dblocks, xfs_filblks_t *rblocks);
+unsigned int xfs_inode_alloc_unitsize(struct xfs_inode *ip);
+
+struct xfs_dir_update_params {
+	const struct xfs_inode	*dp;
+	const struct xfs_inode	*ip;
+	const struct xfs_name	*name;
+	int			delta;
+};
+
+#ifdef CONFIG_XFS_LIVE_HOOKS
+void xfs_dir_update_hook(struct xfs_inode *dp, struct xfs_inode *ip,
+		int delta, const struct xfs_name *name);
+
+struct xfs_dir_hook {
+	struct xfs_hook		dirent_hook;
+};
+
+void xfs_dir_hook_disable(void);
+void xfs_dir_hook_enable(void);
+
+int xfs_dir_hook_add(struct xfs_mount *mp, struct xfs_dir_hook *hook);
+void xfs_dir_hook_del(struct xfs_mount *mp, struct xfs_dir_hook *hook);
+void xfs_dir_hook_setup(struct xfs_dir_hook *hook, notifier_fn_t mod_fn);
+#else
+# define xfs_dir_update_hook(dp, ip, delta, name)	((void)0)
+#endif /* CONFIG_XFS_LIVE_HOOKS */
 
 #endif	/* __XFS_INODE_H__ */
diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c
index 0aee97ba0b..f28d653300 100644
--- a/fs/xfs/xfs_inode_item.c
+++ b/fs/xfs/xfs_inode_item.c
@@ -650,7 +650,7 @@ xfs_inode_item_pin(
 {
 	struct xfs_inode	*ip = INODE_ITEM(lip)->ili_inode;
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 	ASSERT(lip->li_buf);
 
 	trace_xfs_inode_pin(ip, _RET_IP_);
@@ -756,7 +756,7 @@ xfs_inode_item_release(
 	unsigned short		lock_flags;
 
 	ASSERT(ip->i_itemp != NULL);
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 
 	lock_flags = iip->ili_lock_flags;
 	iip->ili_lock_flags = 0;
@@ -856,7 +856,7 @@ xfs_inode_item_destroy(
 	ASSERT(iip->ili_item.li_buf == NULL);
 
 	ip->i_itemp = NULL;
-	kmem_free(iip->ili_item.li_lv_shadow);
+	kvfree(iip->ili_item.li_lv_shadow);
 	kmem_cache_free(xfs_ili_cache, iip);
 }
 
diff --git a/fs/xfs/xfs_inode_item_recover.c b/fs/xfs/xfs_inode_item_recover.c
index 144198a6b2..dbdab4ce7c 100644
--- a/fs/xfs/xfs_inode_item_recover.c
+++ b/fs/xfs/xfs_inode_item_recover.c
@@ -291,7 +291,8 @@ xlog_recover_inode_commit_pass2(
 	if (item->ri_buf[0].i_len == sizeof(struct xfs_inode_log_format)) {
 		in_f = item->ri_buf[0].i_addr;
 	} else {
-		in_f = kmem_alloc(sizeof(struct xfs_inode_log_format), 0);
+		in_f = kmalloc(sizeof(struct xfs_inode_log_format),
+				GFP_KERNEL | __GFP_NOFAIL);
 		need_free = 1;
 		error = xfs_inode_item_format_convert(&item->ri_buf[0], in_f);
 		if (error)
@@ -553,7 +554,7 @@ out_release:
 	xfs_buf_relse(bp);
 error:
 	if (need_free)
-		kmem_free(in_f);
+		kfree(in_f);
 	return error;
 }
 
diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c
index f02b6e558a..f0117188f3 100644
--- a/fs/xfs/xfs_ioctl.c
+++ b/fs/xfs/xfs_ioctl.c
@@ -23,11 +23,9 @@
 #include "xfs_fsops.h"
 #include "xfs_discard.h"
 #include "xfs_quota.h"
-#include "xfs_export.h"
 #include "xfs_trace.h"
 #include "xfs_icache.h"
 #include "xfs_trans.h"
-#include "xfs_acl.h"
 #include "xfs_btree.h"
 #include <linux/fsmap.h>
 #include "xfs_fsmap.h"
@@ -39,596 +37,13 @@
 #include "xfs_ioctl.h"
 #include "xfs_xattr.h"
 #include "xfs_rtbitmap.h"
+#include "xfs_file.h"
+#include "xfs_exchrange.h"
+#include "xfs_handle.h"
 
 #include <linux/mount.h>
-#include <linux/namei.h>
 #include <linux/fileattr.h>
 
-/*
- * xfs_find_handle maps from userspace xfs_fsop_handlereq structure to
- * a file or fs handle.
- *
- * XFS_IOC_PATH_TO_FSHANDLE
- *    returns fs handle for a mount point or path within that mount point
- * XFS_IOC_FD_TO_HANDLE
- *    returns full handle for a FD opened in user space
- * XFS_IOC_PATH_TO_HANDLE
- *    returns full handle for a path
- */
-int
-xfs_find_handle(
-	unsigned int		cmd,
-	xfs_fsop_handlereq_t	*hreq)
-{
-	int			hsize;
-	xfs_handle_t		handle;
-	struct inode		*inode;
-	struct fd		f = {NULL};
-	struct path		path;
-	int			error;
-	struct xfs_inode	*ip;
-
-	if (cmd == XFS_IOC_FD_TO_HANDLE) {
-		f = fdget(hreq->fd);
-		if (!f.file)
-			return -EBADF;
-		inode = file_inode(f.file);
-	} else {
-		error = user_path_at(AT_FDCWD, hreq->path, 0, &path);
-		if (error)
-			return error;
-		inode = d_inode(path.dentry);
-	}
-	ip = XFS_I(inode);
-
-	/*
-	 * We can only generate handles for inodes residing on a XFS filesystem,
-	 * and only for regular files, directories or symbolic links.
-	 */
-	error = -EINVAL;
-	if (inode->i_sb->s_magic != XFS_SB_MAGIC)
-		goto out_put;
-
-	error = -EBADF;
-	if (!S_ISREG(inode->i_mode) &&
-	    !S_ISDIR(inode->i_mode) &&
-	    !S_ISLNK(inode->i_mode))
-		goto out_put;
-
-
-	memcpy(&handle.ha_fsid, ip->i_mount->m_fixedfsid, sizeof(xfs_fsid_t));
-
-	if (cmd == XFS_IOC_PATH_TO_FSHANDLE) {
-		/*
-		 * This handle only contains an fsid, zero the rest.
-		 */
-		memset(&handle.ha_fid, 0, sizeof(handle.ha_fid));
-		hsize = sizeof(xfs_fsid_t);
-	} else {
-		handle.ha_fid.fid_len = sizeof(xfs_fid_t) -
-					sizeof(handle.ha_fid.fid_len);
-		handle.ha_fid.fid_pad = 0;
-		handle.ha_fid.fid_gen = inode->i_generation;
-		handle.ha_fid.fid_ino = ip->i_ino;
-		hsize = sizeof(xfs_handle_t);
-	}
-
-	error = -EFAULT;
-	if (copy_to_user(hreq->ohandle, &handle, hsize) ||
-	    copy_to_user(hreq->ohandlen, &hsize, sizeof(__s32)))
-		goto out_put;
-
-	error = 0;
-
- out_put:
-	if (cmd == XFS_IOC_FD_TO_HANDLE)
-		fdput(f);
-	else
-		path_put(&path);
-	return error;
-}
-
-/*
- * No need to do permission checks on the various pathname components
- * as the handle operations are privileged.
- */
-STATIC int
-xfs_handle_acceptable(
-	void			*context,
-	struct dentry		*dentry)
-{
-	return 1;
-}
-
-/*
- * Convert userspace handle data into a dentry.
- */
-struct dentry *
-xfs_handle_to_dentry(
-	struct file		*parfilp,
-	void __user		*uhandle,
-	u32			hlen)
-{
-	xfs_handle_t		handle;
-	struct xfs_fid64	fid;
-
-	/*
-	 * Only allow handle opens under a directory.
-	 */
-	if (!S_ISDIR(file_inode(parfilp)->i_mode))
-		return ERR_PTR(-ENOTDIR);
-
-	if (hlen != sizeof(xfs_handle_t))
-		return ERR_PTR(-EINVAL);
-	if (copy_from_user(&handle, uhandle, hlen))
-		return ERR_PTR(-EFAULT);
-	if (handle.ha_fid.fid_len !=
-	    sizeof(handle.ha_fid) - sizeof(handle.ha_fid.fid_len))
-		return ERR_PTR(-EINVAL);
-
-	memset(&fid, 0, sizeof(struct fid));
-	fid.ino = handle.ha_fid.fid_ino;
-	fid.gen = handle.ha_fid.fid_gen;
-
-	return exportfs_decode_fh(parfilp->f_path.mnt, (struct fid *)&fid, 3,
-			FILEID_INO32_GEN | XFS_FILEID_TYPE_64FLAG,
-			xfs_handle_acceptable, NULL);
-}
-
-STATIC struct dentry *
-xfs_handlereq_to_dentry(
-	struct file		*parfilp,
-	xfs_fsop_handlereq_t	*hreq)
-{
-	return xfs_handle_to_dentry(parfilp, hreq->ihandle, hreq->ihandlen);
-}
-
-int
-xfs_open_by_handle(
-	struct file		*parfilp,
-	xfs_fsop_handlereq_t	*hreq)
-{
-	const struct cred	*cred = current_cred();
-	int			error;
-	int			fd;
-	int			permflag;
-	struct file		*filp;
-	struct inode		*inode;
-	struct dentry		*dentry;
-	fmode_t			fmode;
-	struct path		path;
-
-	if (!capable(CAP_SYS_ADMIN))
-		return -EPERM;
-
-	dentry = xfs_handlereq_to_dentry(parfilp, hreq);
-	if (IS_ERR(dentry))
-		return PTR_ERR(dentry);
-	inode = d_inode(dentry);
-
-	/* Restrict xfs_open_by_handle to directories & regular files. */
-	if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode))) {
-		error = -EPERM;
-		goto out_dput;
-	}
-
-#if BITS_PER_LONG != 32
-	hreq->oflags |= O_LARGEFILE;
-#endif
-
-	permflag = hreq->oflags;
-	fmode = OPEN_FMODE(permflag);
-	if ((!(permflag & O_APPEND) || (permflag & O_TRUNC)) &&
-	    (fmode & FMODE_WRITE) && IS_APPEND(inode)) {
-		error = -EPERM;
-		goto out_dput;
-	}
-
-	if ((fmode & FMODE_WRITE) && IS_IMMUTABLE(inode)) {
-		error = -EPERM;
-		goto out_dput;
-	}
-
-	/* Can't write directories. */
-	if (S_ISDIR(inode->i_mode) && (fmode & FMODE_WRITE)) {
-		error = -EISDIR;
-		goto out_dput;
-	}
-
-	fd = get_unused_fd_flags(0);
-	if (fd < 0) {
-		error = fd;
-		goto out_dput;
-	}
-
-	path.mnt = parfilp->f_path.mnt;
-	path.dentry = dentry;
-	filp = dentry_open(&path, hreq->oflags, cred);
-	dput(dentry);
-	if (IS_ERR(filp)) {
-		put_unused_fd(fd);
-		return PTR_ERR(filp);
-	}
-
-	if (S_ISREG(inode->i_mode)) {
-		filp->f_flags |= O_NOATIME;
-		filp->f_mode |= FMODE_NOCMTIME;
-	}
-
-	fd_install(fd, filp);
-	return fd;
-
- out_dput:
-	dput(dentry);
-	return error;
-}
-
-int
-xfs_readlink_by_handle(
-	struct file		*parfilp,
-	xfs_fsop_handlereq_t	*hreq)
-{
-	struct dentry		*dentry;
-	__u32			olen;
-	int			error;
-
-	if (!capable(CAP_SYS_ADMIN))
-		return -EPERM;
-
-	dentry = xfs_handlereq_to_dentry(parfilp, hreq);
-	if (IS_ERR(dentry))
-		return PTR_ERR(dentry);
-
-	/* Restrict this handle operation to symlinks only. */
-	if (!d_is_symlink(dentry)) {
-		error = -EINVAL;
-		goto out_dput;
-	}
-
-	if (copy_from_user(&olen, hreq->ohandlen, sizeof(__u32))) {
-		error = -EFAULT;
-		goto out_dput;
-	}
-
-	error = vfs_readlink(dentry, hreq->ohandle, olen);
-
- out_dput:
-	dput(dentry);
-	return error;
-}
-
-/*
- * Format an attribute and copy it out to the user's buffer.
- * Take care to check values and protect against them changing later,
- * we may be reading them directly out of a user buffer.
- */
-static void
-xfs_ioc_attr_put_listent(
-	struct xfs_attr_list_context *context,
-	int			flags,
-	unsigned char		*name,
-	int			namelen,
-	int			valuelen)
-{
-	struct xfs_attrlist	*alist = context->buffer;
-	struct xfs_attrlist_ent	*aep;
-	int			arraytop;
-
-	ASSERT(!context->seen_enough);
-	ASSERT(context->count >= 0);
-	ASSERT(context->count < (ATTR_MAX_VALUELEN/8));
-	ASSERT(context->firstu >= sizeof(*alist));
-	ASSERT(context->firstu <= context->bufsize);
-
-	/*
-	 * Only list entries in the right namespace.
-	 */
-	if (context->attr_filter != (flags & XFS_ATTR_NSP_ONDISK_MASK))
-		return;
-
-	arraytop = sizeof(*alist) +
-			context->count * sizeof(alist->al_offset[0]);
-
-	/* decrement by the actual bytes used by the attr */
-	context->firstu -= round_up(offsetof(struct xfs_attrlist_ent, a_name) +
-			namelen + 1, sizeof(uint32_t));
-	if (context->firstu < arraytop) {
-		trace_xfs_attr_list_full(context);
-		alist->al_more = 1;
-		context->seen_enough = 1;
-		return;
-	}
-
-	aep = context->buffer + context->firstu;
-	aep->a_valuelen = valuelen;
-	memcpy(aep->a_name, name, namelen);
-	aep->a_name[namelen] = 0;
-	alist->al_offset[context->count++] = context->firstu;
-	alist->al_count = context->count;
-	trace_xfs_attr_list_add(context);
-}
-
-static unsigned int
-xfs_attr_filter(
-	u32			ioc_flags)
-{
-	if (ioc_flags & XFS_IOC_ATTR_ROOT)
-		return XFS_ATTR_ROOT;
-	if (ioc_flags & XFS_IOC_ATTR_SECURE)
-		return XFS_ATTR_SECURE;
-	return 0;
-}
-
-static unsigned int
-xfs_attr_flags(
-	u32			ioc_flags)
-{
-	if (ioc_flags & XFS_IOC_ATTR_CREATE)
-		return XATTR_CREATE;
-	if (ioc_flags & XFS_IOC_ATTR_REPLACE)
-		return XATTR_REPLACE;
-	return 0;
-}
-
-int
-xfs_ioc_attr_list(
-	struct xfs_inode		*dp,
-	void __user			*ubuf,
-	size_t				bufsize,
-	int				flags,
-	struct xfs_attrlist_cursor __user *ucursor)
-{
-	struct xfs_attr_list_context	context = { };
-	struct xfs_attrlist		*alist;
-	void				*buffer;
-	int				error;
-
-	if (bufsize < sizeof(struct xfs_attrlist) ||
-	    bufsize > XFS_XATTR_LIST_MAX)
-		return -EINVAL;
-
-	/*
-	 * Reject flags, only allow namespaces.
-	 */
-	if (flags & ~(XFS_IOC_ATTR_ROOT | XFS_IOC_ATTR_SECURE))
-		return -EINVAL;
-	if (flags == (XFS_IOC_ATTR_ROOT | XFS_IOC_ATTR_SECURE))
-		return -EINVAL;
-
-	/*
-	 * Validate the cursor.
-	 */
-	if (copy_from_user(&context.cursor, ucursor, sizeof(context.cursor)))
-		return -EFAULT;
-	if (context.cursor.pad1 || context.cursor.pad2)
-		return -EINVAL;
-	if (!context.cursor.initted &&
-	    (context.cursor.hashval || context.cursor.blkno ||
-	     context.cursor.offset))
-		return -EINVAL;
-
-	buffer = kvzalloc(bufsize, GFP_KERNEL);
-	if (!buffer)
-		return -ENOMEM;
-
-	/*
-	 * Initialize the output buffer.
-	 */
-	context.dp = dp;
-	context.resynch = 1;
-	context.attr_filter = xfs_attr_filter(flags);
-	context.buffer = buffer;
-	context.bufsize = round_down(bufsize, sizeof(uint32_t));
-	context.firstu = context.bufsize;
-	context.put_listent = xfs_ioc_attr_put_listent;
-
-	alist = context.buffer;
-	alist->al_count = 0;
-	alist->al_more = 0;
-	alist->al_offset[0] = context.bufsize;
-
-	error = xfs_attr_list(&context);
-	if (error)
-		goto out_free;
-
-	if (copy_to_user(ubuf, buffer, bufsize) ||
-	    copy_to_user(ucursor, &context.cursor, sizeof(context.cursor)))
-		error = -EFAULT;
-out_free:
-	kmem_free(buffer);
-	return error;
-}
-
-STATIC int
-xfs_attrlist_by_handle(
-	struct file		*parfilp,
-	struct xfs_fsop_attrlist_handlereq __user *p)
-{
-	struct xfs_fsop_attrlist_handlereq al_hreq;
-	struct dentry		*dentry;
-	int			error = -ENOMEM;
-
-	if (!capable(CAP_SYS_ADMIN))
-		return -EPERM;
-	if (copy_from_user(&al_hreq, p, sizeof(al_hreq)))
-		return -EFAULT;
-
-	dentry = xfs_handlereq_to_dentry(parfilp, &al_hreq.hreq);
-	if (IS_ERR(dentry))
-		return PTR_ERR(dentry);
-
-	error = xfs_ioc_attr_list(XFS_I(d_inode(dentry)), al_hreq.buffer,
-				  al_hreq.buflen, al_hreq.flags, &p->pos);
-	dput(dentry);
-	return error;
-}
-
-static int
-xfs_attrmulti_attr_get(
-	struct inode		*inode,
-	unsigned char		*name,
-	unsigned char		__user *ubuf,
-	uint32_t		*len,
-	uint32_t		flags)
-{
-	struct xfs_da_args	args = {
-		.dp		= XFS_I(inode),
-		.attr_filter	= xfs_attr_filter(flags),
-		.attr_flags	= xfs_attr_flags(flags),
-		.name		= name,
-		.namelen	= strlen(name),
-		.valuelen	= *len,
-	};
-	int			error;
-
-	if (*len > XFS_XATTR_SIZE_MAX)
-		return -EINVAL;
-
-	error = xfs_attr_get(&args);
-	if (error)
-		goto out_kfree;
-
-	*len = args.valuelen;
-	if (copy_to_user(ubuf, args.value, args.valuelen))
-		error = -EFAULT;
-
-out_kfree:
-	kmem_free(args.value);
-	return error;
-}
-
-static int
-xfs_attrmulti_attr_set(
-	struct inode		*inode,
-	unsigned char		*name,
-	const unsigned char	__user *ubuf,
-	uint32_t		len,
-	uint32_t		flags)
-{
-	struct xfs_da_args	args = {
-		.dp		= XFS_I(inode),
-		.attr_filter	= xfs_attr_filter(flags),
-		.attr_flags	= xfs_attr_flags(flags),
-		.name		= name,
-		.namelen	= strlen(name),
-	};
-	int			error;
-
-	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
-		return -EPERM;
-
-	if (ubuf) {
-		if (len > XFS_XATTR_SIZE_MAX)
-			return -EINVAL;
-		args.value = memdup_user(ubuf, len);
-		if (IS_ERR(args.value))
-			return PTR_ERR(args.value);
-		args.valuelen = len;
-	}
-
-	error = xfs_attr_change(&args);
-	if (!error && (flags & XFS_IOC_ATTR_ROOT))
-		xfs_forget_acl(inode, name);
-	kfree(args.value);
-	return error;
-}
-
-int
-xfs_ioc_attrmulti_one(
-	struct file		*parfilp,
-	struct inode		*inode,
-	uint32_t		opcode,
-	void __user		*uname,
-	void __user		*value,
-	uint32_t		*len,
-	uint32_t		flags)
-{
-	unsigned char		*name;
-	int			error;
-
-	if ((flags & XFS_IOC_ATTR_ROOT) && (flags & XFS_IOC_ATTR_SECURE))
-		return -EINVAL;
-
-	name = strndup_user(uname, MAXNAMELEN);
-	if (IS_ERR(name))
-		return PTR_ERR(name);
-
-	switch (opcode) {
-	case ATTR_OP_GET:
-		error = xfs_attrmulti_attr_get(inode, name, value, len, flags);
-		break;
-	case ATTR_OP_REMOVE:
-		value = NULL;
-		*len = 0;
-		fallthrough;
-	case ATTR_OP_SET:
-		error = mnt_want_write_file(parfilp);
-		if (error)
-			break;
-		error = xfs_attrmulti_attr_set(inode, name, value, *len, flags);
-		mnt_drop_write_file(parfilp);
-		break;
-	default:
-		error = -EINVAL;
-		break;
-	}
-
-	kfree(name);
-	return error;
-}
-
-STATIC int
-xfs_attrmulti_by_handle(
-	struct file		*parfilp,
-	void			__user *arg)
-{
-	int			error;
-	xfs_attr_multiop_t	*ops;
-	xfs_fsop_attrmulti_handlereq_t am_hreq;
-	struct dentry		*dentry;
-	unsigned int		i, size;
-
-	if (!capable(CAP_SYS_ADMIN))
-		return -EPERM;
-	if (copy_from_user(&am_hreq, arg, sizeof(xfs_fsop_attrmulti_handlereq_t)))
-		return -EFAULT;
-
-	/* overflow check */
-	if (am_hreq.opcount >= INT_MAX / sizeof(xfs_attr_multiop_t))
-		return -E2BIG;
-
-	dentry = xfs_handlereq_to_dentry(parfilp, &am_hreq.hreq);
-	if (IS_ERR(dentry))
-		return PTR_ERR(dentry);
-
-	error = -E2BIG;
-	size = am_hreq.opcount * sizeof(xfs_attr_multiop_t);
-	if (!size || size > 16 * PAGE_SIZE)
-		goto out_dput;
-
-	ops = memdup_user(am_hreq.ops, size);
-	if (IS_ERR(ops)) {
-		error = PTR_ERR(ops);
-		goto out_dput;
-	}
-
-	error = 0;
-	for (i = 0; i < am_hreq.opcount; i++) {
-		ops[i].am_error = xfs_ioc_attrmulti_one(parfilp,
-				d_inode(dentry), ops[i].am_opcode,
-				ops[i].am_attrname, ops[i].am_attrvalue,
-				&ops[i].am_length, ops[i].am_flags);
-	}
-
-	if (copy_to_user(am_hreq.ops, ops, size))
-		error = -EFAULT;
-
-	kfree(ops);
- out_dput:
-	dput(dentry);
-	return error;
-}
-
 /* Return 0 on success or positive error */
 int
 xfs_fsbulkstat_one_fmt(
@@ -1506,7 +921,7 @@ xfs_ioc_getbmap(
 
 	error = 0;
 out_free_buf:
-	kmem_free(buf);
+	kvfree(buf);
 	return error;
 }
 
@@ -1636,34 +1051,10 @@ xfs_ioc_getfsmap(
 	}
 
 out_free:
-	kmem_free(recs);
+	kvfree(recs);
 	return error;
 }
 
-STATIC int
-xfs_ioc_scrub_metadata(
-	struct file			*file,
-	void				__user *arg)
-{
-	struct xfs_scrub_metadata	scrub;
-	int				error;
-
-	if (!capable(CAP_SYS_ADMIN))
-		return -EPERM;
-
-	if (copy_from_user(&scrub, arg, sizeof(scrub)))
-		return -EFAULT;
-
-	error = xfs_scrub_metadata(file, &scrub);
-	if (error)
-		return error;
-
-	if (copy_to_user(arg, &scrub, sizeof(scrub)))
-		return -EFAULT;
-
-	return 0;
-}
-
 int
 xfs_ioc_swapext(
 	xfs_swapext_t	*sxp)
@@ -2010,7 +1401,10 @@ xfs_file_ioctl(
 
 	case XFS_IOC_FSGETXATTRA:
 		return xfs_ioc_fsgetxattra(ip, arg);
-
+	case XFS_IOC_GETPARENTS:
+		return xfs_ioc_getparents(filp, arg);
+	case XFS_IOC_GETPARENTS_BY_HANDLE:
+		return xfs_ioc_getparents_by_handle(filp, arg);
 	case XFS_IOC_GETBMAP:
 	case XFS_IOC_GETBMAPA:
 	case XFS_IOC_GETBMAPX:
@@ -2019,6 +1413,8 @@ xfs_file_ioctl(
 	case FS_IOC_GETFSMAP:
 		return xfs_ioc_getfsmap(ip, arg);
 
+	case XFS_IOC_SCRUBV_METADATA:
+		return xfs_ioc_scrubv_metadata(filp, arg);
 	case XFS_IOC_SCRUB_METADATA:
 		return xfs_ioc_scrub_metadata(filp, arg);
 
@@ -2169,6 +1565,9 @@ xfs_file_ioctl(
 		return error;
 	}
 
+	case XFS_IOC_EXCHANGE_RANGE:
+		return xfs_ioc_exchange_range(filp, arg);
+
 	default:
 		return -ENOTTY;
 	}
diff --git a/fs/xfs/xfs_ioctl.h b/fs/xfs/xfs_ioctl.h
index 38be600b5e..12124946f3 100644
--- a/fs/xfs/xfs_ioctl.h
+++ b/fs/xfs/xfs_ioctl.h
@@ -15,34 +15,6 @@ xfs_ioc_swapext(
 	xfs_swapext_t	*sxp);
 
 extern int
-xfs_find_handle(
-	unsigned int		cmd,
-	xfs_fsop_handlereq_t	*hreq);
-
-extern int
-xfs_open_by_handle(
-	struct file		*parfilp,
-	xfs_fsop_handlereq_t	*hreq);
-
-extern int
-xfs_readlink_by_handle(
-	struct file		*parfilp,
-	xfs_fsop_handlereq_t	*hreq);
-
-int xfs_ioc_attrmulti_one(struct file *parfilp, struct inode *inode,
-		uint32_t opcode, void __user *uname, void __user *value,
-		uint32_t *len, uint32_t flags);
-int xfs_ioc_attr_list(struct xfs_inode *dp, void __user *ubuf,
-		      size_t bufsize, int flags,
-		      struct xfs_attrlist_cursor __user *ucursor);
-
-extern struct dentry *
-xfs_handle_to_dentry(
-	struct file		*parfilp,
-	void __user		*uhandle,
-	u32			hlen);
-
-extern int
 xfs_fileattr_get(
 	struct dentry		*dentry,
 	struct fileattr		*fa);
diff --git a/fs/xfs/xfs_ioctl32.c b/fs/xfs/xfs_ioctl32.c
index ee35eea1ec..b64785dc43 100644
--- a/fs/xfs/xfs_ioctl32.c
+++ b/fs/xfs/xfs_ioctl32.c
@@ -24,6 +24,7 @@
 #include "xfs_ioctl32.h"
 #include "xfs_trace.h"
 #include "xfs_sb.h"
+#include "xfs_handle.h"
 
 #define  _NATIVE_IOC(cmd, type) \
 	  _IOC(_IOC_DIR(cmd), _IOC_TYPE(cmd), _IOC_NR(cmd), sizeof(type))
diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
index 18c8f168b1..414903885a 100644
--- a/fs/xfs/xfs_iomap.c
+++ b/fs/xfs/xfs_iomap.c
@@ -27,6 +27,8 @@
 #include "xfs_dquot_item.h"
 #include "xfs_dquot.h"
 #include "xfs_reflink.h"
+#include "xfs_health.h"
+#include "xfs_rtbitmap.h"
 
 #define XFS_ALLOC_ALIGN(mp, off) \
 	(((off) >> mp->m_allocsize_log) << mp->m_allocsize_log)
@@ -45,6 +47,7 @@ xfs_alert_fsblock_zero(
 		(unsigned long long)imap->br_startoff,
 		(unsigned long long)imap->br_blockcount,
 		imap->br_state);
+	xfs_bmap_mark_sick(ip, XFS_DATA_FORK);
 	return -EFSCORRUPTED;
 }
 
@@ -99,8 +102,10 @@ xfs_bmbt_to_iomap(
 	struct xfs_mount	*mp = ip->i_mount;
 	struct xfs_buftarg	*target = xfs_inode_buftarg(ip);
 
-	if (unlikely(!xfs_valid_startblock(ip, imap->br_startblock)))
+	if (unlikely(!xfs_valid_startblock(ip, imap->br_startblock))) {
+		xfs_bmap_mark_sick(ip, XFS_DATA_FORK);
 		return xfs_alert_fsblock_zero(ip, imap);
+	}
 
 	if (imap->br_startblock == HOLESTARTBLOCK) {
 		iomap->addr = IOMAP_NULL_ADDR;
@@ -294,9 +299,7 @@ xfs_iomap_write_direct(
 	if (error)
 		return error;
 
-	error = xfs_iext_count_may_overflow(ip, XFS_DATA_FORK, nr_exts);
-	if (error == -EFBIG)
-		error = xfs_iext_count_upgrade(tp, ip, nr_exts);
+	error = xfs_iext_count_extend(tp, ip, XFS_DATA_FORK, nr_exts);
 	if (error)
 		goto out_trans_cancel;
 
@@ -317,16 +320,10 @@ xfs_iomap_write_direct(
 	if (error)
 		goto out_unlock;
 
-	/*
-	 * Copy any maps to caller's array and return any error.
-	 */
-	if (nimaps == 0) {
-		error = -ENOSPC;
-		goto out_unlock;
-	}
-
-	if (unlikely(!xfs_valid_startblock(ip, imap->br_startblock)))
+	if (unlikely(!xfs_valid_startblock(ip, imap->br_startblock))) {
+		xfs_bmap_mark_sick(ip, XFS_DATA_FORK);
 		error = xfs_alert_fsblock_zero(ip, imap);
+	}
 
 out_unlock:
 	*seq = xfs_iomap_inode_sequence(ip, 0);
@@ -398,6 +395,29 @@ xfs_quota_calc_throttle(
 	}
 }
 
+static int64_t
+xfs_iomap_freesp(
+	struct percpu_counter	*counter,
+	uint64_t		low_space[XFS_LOWSP_MAX],
+	int			*shift)
+{
+	int64_t			freesp;
+
+	freesp = percpu_counter_read_positive(counter);
+	if (freesp < low_space[XFS_LOWSP_5_PCNT]) {
+		*shift = 2;
+		if (freesp < low_space[XFS_LOWSP_4_PCNT])
+			(*shift)++;
+		if (freesp < low_space[XFS_LOWSP_3_PCNT])
+			(*shift)++;
+		if (freesp < low_space[XFS_LOWSP_2_PCNT])
+			(*shift)++;
+		if (freesp < low_space[XFS_LOWSP_1_PCNT])
+			(*shift)++;
+	}
+	return freesp;
+}
+
 /*
  * If we don't have a user specified preallocation size, dynamically increase
  * the preallocation size as the size of the file grows.  Cap the maximum size
@@ -480,18 +500,13 @@ xfs_iomap_prealloc_size(
 	alloc_blocks = XFS_FILEOFF_MIN(roundup_pow_of_two(XFS_MAX_BMBT_EXTLEN),
 				       alloc_blocks);
 
-	freesp = percpu_counter_read_positive(&mp->m_fdblocks);
-	if (freesp < mp->m_low_space[XFS_LOWSP_5_PCNT]) {
-		shift = 2;
-		if (freesp < mp->m_low_space[XFS_LOWSP_4_PCNT])
-			shift++;
-		if (freesp < mp->m_low_space[XFS_LOWSP_3_PCNT])
-			shift++;
-		if (freesp < mp->m_low_space[XFS_LOWSP_2_PCNT])
-			shift++;
-		if (freesp < mp->m_low_space[XFS_LOWSP_1_PCNT])
-			shift++;
-	}
+	if (unlikely(XFS_IS_REALTIME_INODE(ip)))
+		freesp = xfs_rtx_to_rtb(mp,
+			xfs_iomap_freesp(&mp->m_frextents,
+					mp->m_low_rtexts, &shift));
+	else
+		freesp = xfs_iomap_freesp(&mp->m_fdblocks, mp->m_low_space,
+				&shift);
 
 	/*
 	 * Check each quota to cap the prealloc size, provide a shift value to
@@ -600,11 +615,8 @@ xfs_iomap_write_unwritten(
 		if (error)
 			return error;
 
-		error = xfs_iext_count_may_overflow(ip, XFS_DATA_FORK,
+		error = xfs_iext_count_extend(tp, ip, XFS_DATA_FORK,
 				XFS_IEXT_WRITE_UNWRITTEN_CNT);
-		if (error == -EFBIG)
-			error = xfs_iext_count_upgrade(tp, ip,
-					XFS_IEXT_WRITE_UNWRITTEN_CNT);
 		if (error)
 			goto error_on_bmapi_transaction;
 
@@ -639,8 +651,10 @@ xfs_iomap_write_unwritten(
 		if (error)
 			return error;
 
-		if (unlikely(!xfs_valid_startblock(ip, imap.br_startblock)))
+		if (unlikely(!xfs_valid_startblock(ip, imap.br_startblock))) {
+			xfs_bmap_mark_sick(ip, XFS_DATA_FORK);
 			return xfs_alert_fsblock_zero(ip, &imap);
+		}
 
 		if ((numblks_fsb = imap.br_blockcount) == 0) {
 			/*
@@ -974,8 +988,6 @@ xfs_buffered_write_iomap_begin(
 		return xfs_direct_write_iomap_begin(inode, offset, count,
 				flags, iomap, srcmap);
 
-	ASSERT(!XFS_IS_REALTIME_INODE(ip));
-
 	error = xfs_qm_dqattach(ip);
 	if (error)
 		return error;
@@ -986,6 +998,7 @@ xfs_buffered_write_iomap_begin(
 
 	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(&ip->i_df)) ||
 	    XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) {
+		xfs_bmap_mark_sick(ip, XFS_DATA_FORK);
 		error = -EFSCORRUPTED;
 		goto out_unlock;
 	}
@@ -1014,6 +1027,24 @@ xfs_buffered_write_iomap_begin(
 	}
 
 	/*
+	 * For zeroing, trim a delalloc extent that extends beyond the EOF
+	 * block.  If it starts beyond the EOF block, convert it to an
+	 * unwritten extent.
+	 */
+	if ((flags & IOMAP_ZERO) && imap.br_startoff <= offset_fsb &&
+	    isnullstartblock(imap.br_startblock)) {
+		xfs_fileoff_t eof_fsb = XFS_B_TO_FSB(mp, XFS_ISIZE(ip));
+
+		if (offset_fsb >= eof_fsb)
+			goto convert_delay;
+		if (end_fsb > eof_fsb) {
+			end_fsb = eof_fsb;
+			xfs_trim_extent(&imap, offset_fsb,
+					end_fsb - offset_fsb);
+		}
+	}
+
+	/*
 	 * Search the COW fork extent list even if we did not find a data fork
 	 * extent.  This serves two purposes: first this implements the
 	 * speculative preallocation using cowextsize, so that we also unshare
@@ -1117,47 +1148,48 @@ xfs_buffered_write_iomap_begin(
 		}
 	}
 
-retry:
-	error = xfs_bmapi_reserve_delalloc(ip, allocfork, offset_fsb,
-			end_fsb - offset_fsb, prealloc_blocks,
-			allocfork == XFS_DATA_FORK ? &imap : &cmap,
-			allocfork == XFS_DATA_FORK ? &icur : &ccur,
-			allocfork == XFS_DATA_FORK ? eof : cow_eof);
-	switch (error) {
-	case 0:
-		break;
-	case -ENOSPC:
-	case -EDQUOT:
-		/* retry without any preallocation */
-		trace_xfs_delalloc_enospc(ip, offset, count);
-		if (prealloc_blocks) {
-			prealloc_blocks = 0;
-			goto retry;
-		}
-		fallthrough;
-	default:
-		goto out_unlock;
-	}
-
 	if (allocfork == XFS_COW_FORK) {
+		error = xfs_bmapi_reserve_delalloc(ip, allocfork, offset_fsb,
+				end_fsb - offset_fsb, prealloc_blocks, &cmap,
+				&ccur, cow_eof);
+		if (error)
+			goto out_unlock;
+
 		trace_xfs_iomap_alloc(ip, offset, count, allocfork, &cmap);
 		goto found_cow;
 	}
 
+	error = xfs_bmapi_reserve_delalloc(ip, allocfork, offset_fsb,
+			end_fsb - offset_fsb, prealloc_blocks, &imap, &icur,
+			eof);
+	if (error)
+		goto out_unlock;
+
 	/*
 	 * Flag newly allocated delalloc blocks with IOMAP_F_NEW so we punch
 	 * them out if the write happens to fail.
 	 */
 	seq = xfs_iomap_inode_sequence(ip, IOMAP_F_NEW);
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	xfs_iunlock(ip, lockmode);
 	trace_xfs_iomap_alloc(ip, offset, count, allocfork, &imap);
 	return xfs_bmbt_to_iomap(ip, iomap, &imap, flags, IOMAP_F_NEW, seq);
 
 found_imap:
 	seq = xfs_iomap_inode_sequence(ip, 0);
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	xfs_iunlock(ip, lockmode);
 	return xfs_bmbt_to_iomap(ip, iomap, &imap, flags, 0, seq);
 
+convert_delay:
+	xfs_iunlock(ip, lockmode);
+	truncate_pagecache(inode, offset);
+	error = xfs_bmapi_convert_delalloc(ip, XFS_DATA_FORK, offset,
+					   iomap, NULL);
+	if (error)
+		return error;
+
+	trace_xfs_iomap_alloc(ip, offset, count, XFS_DATA_FORK, &imap);
+	return 0;
+
 found_cow:
 	seq = xfs_iomap_inode_sequence(ip, 0);
 	if (imap.br_startoff <= offset_fsb) {
@@ -1165,17 +1197,17 @@ found_cow:
 		if (error)
 			goto out_unlock;
 		seq = xfs_iomap_inode_sequence(ip, IOMAP_F_SHARED);
-		xfs_iunlock(ip, XFS_ILOCK_EXCL);
+		xfs_iunlock(ip, lockmode);
 		return xfs_bmbt_to_iomap(ip, iomap, &cmap, flags,
 					 IOMAP_F_SHARED, seq);
 	}
 
 	xfs_trim_extent(&cmap, offset_fsb, imap.br_startoff - offset_fsb);
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	xfs_iunlock(ip, lockmode);
 	return xfs_bmbt_to_iomap(ip, iomap, &cmap, flags, 0, seq);
 
 out_unlock:
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	xfs_iunlock(ip, lockmode);
 	return error;
 }
 
@@ -1185,8 +1217,8 @@ xfs_buffered_write_delalloc_punch(
 	loff_t			offset,
 	loff_t			length)
 {
-	return xfs_bmap_punch_delalloc_range(XFS_I(inode), offset,
-			offset + length);
+	xfs_bmap_punch_delalloc_range(XFS_I(inode), offset, offset + length);
+	return 0;
 }
 
 static int
@@ -1323,7 +1355,7 @@ xfs_seek_iomap_begin(
 	if (cow_fsb != NULLFILEOFF && cow_fsb <= offset_fsb) {
 		if (data_fsb < cow_fsb + cmap.br_blockcount)
 			end_fsb = min(end_fsb, data_fsb);
-		xfs_trim_extent(&cmap, offset_fsb, end_fsb);
+		xfs_trim_extent(&cmap, offset_fsb, end_fsb - offset_fsb);
 		seq = xfs_iomap_inode_sequence(ip, IOMAP_F_SHARED);
 		error = xfs_bmbt_to_iomap(ip, iomap, &cmap, flags,
 				IOMAP_F_SHARED, seq);
@@ -1348,7 +1380,7 @@ xfs_seek_iomap_begin(
 	imap.br_state = XFS_EXT_NORM;
 done:
 	seq = xfs_iomap_inode_sequence(ip, 0);
-	xfs_trim_extent(&imap, offset_fsb, end_fsb);
+	xfs_trim_extent(&imap, offset_fsb, end_fsb - offset_fsb);
 	error = xfs_bmbt_to_iomap(ip, iomap, &imap, flags, 0, seq);
 out_unlock:
 	xfs_iunlock(ip, lockmode);
diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c
index a0d77f5f51..ff222827e5 100644
--- a/fs/xfs/xfs_iops.c
+++ b/fs/xfs/xfs_iops.c
@@ -25,6 +25,7 @@
 #include "xfs_error.h"
 #include "xfs_ioctl.h"
 #include "xfs_xattr.h"
+#include "xfs_file.h"
 
 #include <linux/posix_acl.h>
 #include <linux/security.h>
@@ -62,7 +63,7 @@ xfs_initxattrs(
 			.value		= xattr->value,
 			.valuelen	= xattr->value_len,
 		};
-		error = xfs_attr_change(&args);
+		error = xfs_attr_change(&args, XFS_ATTRUPDATE_UPSERT);
 		if (error < 0)
 			break;
 	}
@@ -156,6 +157,8 @@ xfs_create_need_xattr(
 	if (dir->i_sb->s_security)
 		return true;
 #endif
+	if (xfs_has_parent(XFS_I(dir)->i_mount))
+		return true;
 	return false;
 }
 
@@ -200,7 +203,18 @@ xfs_generic_create(
 				xfs_create_need_xattr(dir, default_acl, acl),
 				&ip);
 	} else {
-		error = xfs_create_tmpfile(idmap, XFS_I(dir), mode, &ip);
+		bool	init_xattrs = false;
+
+		/*
+		 * If this temporary file will be linkable, set up the file
+		 * with an attr fork to receive a parent pointer.
+		 */
+		if (!(tmpfile->f_flags & O_EXCL) &&
+		    xfs_has_parent(XFS_I(dir)->i_mount))
+			init_xattrs = true;
+
+		error = xfs_create_tmpfile(idmap, XFS_I(dir), mode,
+				init_xattrs, &ip);
 	}
 	if (unlikely(error))
 		goto out_free_acl;
@@ -346,7 +360,7 @@ xfs_vn_ci_lookup(
 	dname.name = ci_name.name;
 	dname.len = ci_name.len;
 	dentry = d_add_ci(dentry, VFS_I(ip), &dname);
-	kmem_free(ci_name.name);
+	kfree(ci_name.name);
 	return dentry;
 }
 
@@ -364,6 +378,9 @@ xfs_vn_link(
 	if (unlikely(error))
 		return error;
 
+	if (IS_PRIVATE(inode))
+		return -EPERM;
+
 	error = xfs_link(XFS_I(dir), XFS_I(inode), &name);
 	if (unlikely(error))
 		return error;
@@ -521,7 +538,7 @@ xfs_stat_blksize(
 	 * always return the realtime extent size.
 	 */
 	if (XFS_IS_REALTIME_INODE(ip))
-		return XFS_FSB_TO_B(mp, xfs_get_extsz_hint(ip));
+		return XFS_FSB_TO_B(mp, xfs_get_extsz_hint(ip) ? : 1);
 
 	/*
 	 * Allow large block sizes to be reported to userspace programs if the
@@ -796,8 +813,7 @@ xfs_setattr_size(
 	uint			lock_flags = 0;
 	bool			did_zeroing = false;
 
-	ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL));
-	ASSERT(xfs_isilocked(ip, XFS_MMAPLOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL);
 	ASSERT(S_ISREG(inode->i_mode));
 	ASSERT((iattr->ia_valid & (ATTR_UID|ATTR_GID|ATTR_ATIME|ATTR_ATIME_SET|
 		ATTR_MTIME_SET|ATTR_TIMES_SET)) == 0);
@@ -1285,9 +1301,9 @@ xfs_setup_inode(
 		 */
 		lockdep_set_class(&inode->i_rwsem,
 				  &inode->i_sb->s_type->i_mutex_dir_key);
-		lockdep_set_class(&ip->i_lock.mr_lock, &xfs_dir_ilock_class);
+		lockdep_set_class(&ip->i_lock, &xfs_dir_ilock_class);
 	} else {
-		lockdep_set_class(&ip->i_lock.mr_lock, &xfs_nondir_ilock_class);
+		lockdep_set_class(&ip->i_lock, &xfs_nondir_ilock_class);
 	}
 
 	/*
diff --git a/fs/xfs/xfs_iops.h b/fs/xfs/xfs_iops.h
index 7f84a0843b..3c1a2605ff 100644
--- a/fs/xfs/xfs_iops.h
+++ b/fs/xfs/xfs_iops.h
@@ -8,9 +8,6 @@
 
 struct xfs_inode;
 
-extern const struct file_operations xfs_file_operations;
-extern const struct file_operations xfs_dir_file_operations;
-
 extern ssize_t xfs_vn_listxattr(struct dentry *, char *data, size_t size);
 
 int xfs_vn_setattr_size(struct mnt_idmap *idmap,
@@ -19,4 +16,8 @@ int xfs_vn_setattr_size(struct mnt_idmap *idmap,
 int xfs_inode_init_security(struct inode *inode, struct inode *dir,
 		const struct qstr *qstr);
 
+extern void xfs_setup_inode(struct xfs_inode *ip);
+extern void xfs_setup_iops(struct xfs_inode *ip);
+extern void xfs_diflags_to_iflags(struct xfs_inode *ip, bool init);
+
 #endif /* __XFS_IOPS_H__ */
diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c
index 14462614fc..c0757ab994 100644
--- a/fs/xfs/xfs_itable.c
+++ b/fs/xfs/xfs_itable.c
@@ -97,6 +97,14 @@ xfs_bulkstat_one_int(
 	vfsuid = i_uid_into_vfsuid(idmap, inode);
 	vfsgid = i_gid_into_vfsgid(idmap, inode);
 
+	/* If this is a private inode, don't leak its details to userspace. */
+	if (IS_PRIVATE(inode)) {
+		xfs_iunlock(ip, XFS_ILOCK_SHARED);
+		xfs_irele(ip);
+		error = -EINVAL;
+		goto out_advance;
+	}
+
 	/* xfs_iget returns the following without needing
 	 * further change.
 	 */
@@ -197,8 +205,8 @@ xfs_bulkstat_one(
 
 	ASSERT(breq->icount == 1);
 
-	bc.buf = kmem_zalloc(sizeof(struct xfs_bulkstat),
-			KM_MAYFAIL);
+	bc.buf = kzalloc(sizeof(struct xfs_bulkstat),
+			GFP_KERNEL | __GFP_RETRY_MAYFAIL);
 	if (!bc.buf)
 		return -ENOMEM;
 
@@ -214,7 +222,7 @@ xfs_bulkstat_one(
 			breq->startino, &bc);
 	xfs_trans_cancel(tp);
 out:
-	kmem_free(bc.buf);
+	kfree(bc.buf);
 
 	/*
 	 * If we reported one inode to userspace then we abort because we hit
@@ -289,8 +297,8 @@ xfs_bulkstat(
 	if (xfs_bulkstat_already_done(breq->mp, breq->startino))
 		return 0;
 
-	bc.buf = kmem_zalloc(sizeof(struct xfs_bulkstat),
-			KM_MAYFAIL);
+	bc.buf = kzalloc(sizeof(struct xfs_bulkstat),
+			GFP_KERNEL | __GFP_RETRY_MAYFAIL);
 	if (!bc.buf)
 		return -ENOMEM;
 
@@ -309,7 +317,7 @@ xfs_bulkstat(
 			xfs_bulkstat_iwalk, breq->icount, &bc);
 	xfs_trans_cancel(tp);
 out:
-	kmem_free(bc.buf);
+	kfree(bc.buf);
 
 	/*
 	 * We found some inodes, so clear the error status and return them.
diff --git a/fs/xfs/xfs_iwalk.c b/fs/xfs/xfs_iwalk.c
index b3275e8d47..86f14ec7c3 100644
--- a/fs/xfs/xfs_iwalk.c
+++ b/fs/xfs/xfs_iwalk.c
@@ -22,6 +22,7 @@
 #include "xfs_trans.h"
 #include "xfs_pwork.h"
 #include "xfs_ag.h"
+#include "xfs_bit.h"
 
 /*
  * Walking Inodes in the Filesystem
@@ -99,6 +100,7 @@ xfs_iwalk_ichunk_ra(
 	struct xfs_inobt_rec_incore	*irec)
 {
 	struct xfs_ino_geometry		*igeo = M_IGEO(mp);
+	xfs_agnumber_t			agno = pag->pag_agno;
 	xfs_agblock_t			agbno;
 	struct blk_plug			plug;
 	int				i;	/* inode chunk index */
@@ -111,8 +113,9 @@ xfs_iwalk_ichunk_ra(
 
 		imask = xfs_inobt_maskn(i, igeo->inodes_per_cluster);
 		if (imask & ~irec->ir_free) {
-			xfs_btree_reada_bufs(mp, pag->pag_agno, agbno,
-					igeo->blocks_per_cluster,
+			xfs_buf_readahead(mp->m_ddev_targp,
+					XFS_AGB_TO_DADDR(mp, agno, agbno),
+					igeo->blocks_per_cluster * mp->m_bsize,
 					&xfs_inode_buf_ops);
 		}
 		agbno += igeo->blocks_per_cluster;
@@ -131,21 +134,11 @@ xfs_iwalk_adjust_start(
 	struct xfs_inobt_rec_incore	*irec)	/* btree record */
 {
 	int				idx;	/* index into inode chunk */
-	int				i;
 
 	idx = agino - irec->ir_startino;
 
-	/*
-	 * We got a right chunk with some left inodes allocated at it.  Grab
-	 * the chunk record.  Mark all the uninteresting inodes free because
-	 * they're before our start point.
-	 */
-	for (i = 0; i < idx; i++) {
-		if (XFS_INOBT_MASK(i) & ~irec->ir_free)
-			irec->ir_freecount++;
-	}
-
 	irec->ir_free |= xfs_inobt_maskn(0, idx);
+	irec->ir_freecount = hweight64(irec->ir_free);
 }
 
 /* Allocate memory for a walk. */
@@ -160,7 +153,7 @@ xfs_iwalk_alloc(
 
 	/* Allocate a prefetch buffer for inobt records. */
 	size = iwag->sz_recs * sizeof(struct xfs_inobt_rec_incore);
-	iwag->recs = kmem_alloc(size, KM_MAYFAIL);
+	iwag->recs = kmalloc(size, GFP_KERNEL | __GFP_RETRY_MAYFAIL);
 	if (iwag->recs == NULL)
 		return -ENOMEM;
 
@@ -172,7 +165,7 @@ STATIC void
 xfs_iwalk_free(
 	struct xfs_iwalk_ag	*iwag)
 {
-	kmem_free(iwag->recs);
+	kfree(iwag->recs);
 	iwag->recs = NULL;
 }
 
@@ -275,9 +268,10 @@ xfs_iwalk_ag_start(
 
 	/* Set up a fresh cursor and empty the inobt cache. */
 	iwag->nr_recs = 0;
-	error = xfs_inobt_cur(pag, tp, XFS_BTNUM_INO, curpp, agi_bpp);
+	error = xfs_ialloc_read_agi(pag, tp, 0, agi_bpp);
 	if (error)
 		return error;
+	*curpp = xfs_inobt_init_cursor(pag, tp, *agi_bpp);
 
 	/* Starting at the beginning of the AG?  That's easy! */
 	if (agino == 0)
@@ -306,8 +300,10 @@ xfs_iwalk_ag_start(
 	error = xfs_inobt_get_rec(*curpp, irec, has_more);
 	if (error)
 		return error;
-	if (XFS_IS_CORRUPT(mp, *has_more != 1))
+	if (XFS_IS_CORRUPT(mp, *has_more != 1)) {
+		xfs_btree_mark_sick(*curpp);
 		return -EFSCORRUPTED;
+	}
 
 	iwag->lastino = XFS_AGINO_TO_INO(mp, pag->pag_agno,
 				irec->ir_startino + XFS_INODES_PER_CHUNK - 1);
@@ -355,7 +351,6 @@ xfs_iwalk_run_callbacks(
 	int				*has_more)
 {
 	struct xfs_mount		*mp = iwag->mp;
-	struct xfs_inobt_rec_incore	*irec;
 	xfs_agino_t			next_agino;
 	int				error;
 
@@ -365,8 +360,8 @@ xfs_iwalk_run_callbacks(
 
 	/* Delete cursor but remember the last record we cached... */
 	xfs_iwalk_del_inobt(iwag->tp, curpp, agi_bpp, 0);
-	irec = &iwag->recs[iwag->nr_recs - 1];
-	ASSERT(next_agino >= irec->ir_startino + XFS_INODES_PER_CHUNK);
+	ASSERT(next_agino >= iwag->recs[iwag->nr_recs - 1].ir_startino +
+			XFS_INODES_PER_CHUNK);
 
 	if (iwag->drop_trans) {
 		xfs_trans_cancel(iwag->tp);
@@ -390,11 +385,10 @@ xfs_iwalk_run_callbacks(
 	}
 
 	/* ...and recreate the cursor just past where we left off. */
-	error = xfs_inobt_cur(iwag->pag, iwag->tp, XFS_BTNUM_INO, curpp,
-			agi_bpp);
+	error = xfs_ialloc_read_agi(iwag->pag, iwag->tp, 0, agi_bpp);
 	if (error)
 		return error;
-
+	*curpp = xfs_inobt_init_cursor(iwag->pag, iwag->tp, *agi_bpp);
 	return xfs_inobt_lookup(*curpp, next_agino, XFS_LOOKUP_GE, has_more);
 }
 
@@ -434,6 +428,7 @@ xfs_iwalk_ag(
 		rec_fsino = XFS_AGINO_TO_INO(mp, pag->pag_agno, irec->ir_startino);
 		if (iwag->lastino != NULLFSINO &&
 		    XFS_IS_CORRUPT(mp, iwag->lastino >= rec_fsino)) {
+			xfs_btree_mark_sick(cur);
 			error = -EFSCORRUPTED;
 			goto out;
 		}
@@ -627,7 +622,7 @@ xfs_iwalk_ag_work(
 	xfs_iwalk_free(iwag);
 out:
 	xfs_perag_put(iwag->pag);
-	kmem_free(iwag);
+	kfree(iwag);
 	return error;
 }
 
@@ -663,7 +658,8 @@ xfs_iwalk_threaded(
 		if (xfs_pwork_ctl_want_abort(&pctl))
 			break;
 
-		iwag = kmem_zalloc(sizeof(struct xfs_iwalk_ag), 0);
+		iwag = kzalloc(sizeof(struct xfs_iwalk_ag),
+				GFP_KERNEL | __GFP_NOFAIL);
 		iwag->mp = mp;
 
 		/*
diff --git a/fs/xfs/xfs_linux.h b/fs/xfs/xfs_linux.h
index d7873e0360..ac35532812 100644
--- a/fs/xfs/xfs_linux.h
+++ b/fs/xfs/xfs_linux.h
@@ -21,15 +21,13 @@ typedef __u32			xfs_nlink_t;
 
 #include "xfs_types.h"
 
-#include "kmem.h"
-#include "mrlock.h"
-
 #include <linux/semaphore.h>
 #include <linux/mm.h>
 #include <linux/sched/mm.h>
 #include <linux/kernel.h>
 #include <linux/blkdev.h>
 #include <linux/slab.h>
+#include <linux/vmalloc.h>
 #include <linux/crc32c.h>
 #include <linux/module.h>
 #include <linux/mutex.h>
@@ -51,6 +49,7 @@ typedef __u32			xfs_nlink_t;
 #include <linux/notifier.h>
 #include <linux/delay.h>
 #include <linux/log2.h>
+#include <linux/rwsem.h>
 #include <linux/spinlock.h>
 #include <linux/random.h>
 #include <linux/ctype.h>
@@ -82,6 +81,7 @@ typedef __u32			xfs_nlink_t;
 #include "xfs_buf.h"
 #include "xfs_message.h"
 #include "xfs_drain.h"
+#include "xfs_hooks.h"
 
 #ifdef __BIG_ENDIAN
 #define XFS_NATIVE_HOST 1
@@ -198,6 +198,11 @@ static inline uint64_t howmany_64(uint64_t x, uint32_t y)
 	return x;
 }
 
+static inline bool isaligned_64(uint64_t x, uint32_t y)
+{
+	return do_div(x, y) == 0;
+}
+
 /* If @b is a power of 2, return log2(b).  Else return -1. */
 static inline int8_t log2_if_power2(unsigned long b)
 {
@@ -269,4 +274,15 @@ int xfs_rw_bdev(struct block_device *bdev, sector_t sector, unsigned int count,
 # define PTR_FMT "%p"
 #endif
 
+/*
+ * Helper for IO routines to grab backing pages from allocated kernel memory.
+ */
+static inline struct page *
+kmem_to_page(void *addr)
+{
+	if (is_vmalloc_addr(addr))
+		return vmalloc_to_page(addr);
+	return virt_to_page(addr);
+}
+
 #endif /* __XFS_LINUX__ */
diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c
index a1650fc813..416c154949 100644
--- a/fs/xfs/xfs_log.c
+++ b/fs/xfs/xfs_log.c
@@ -633,14 +633,14 @@ xlog_state_release_iclog(
  */
 int
 xfs_log_mount(
-	xfs_mount_t	*mp,
-	xfs_buftarg_t	*log_target,
-	xfs_daddr_t	blk_offset,
-	int		num_bblks)
+	xfs_mount_t		*mp,
+	struct xfs_buftarg	*log_target,
+	xfs_daddr_t		blk_offset,
+	int			num_bblks)
 {
-	struct xlog	*log;
-	int		error = 0;
-	int		min_logfsbs;
+	struct xlog		*log;
+	int			error = 0;
+	int			min_logfsbs;
 
 	if (!xfs_has_norecovery(mp)) {
 		xfs_notice(mp, "Mounting V%d Filesystem %pU",
@@ -1448,7 +1448,7 @@ xfs_log_work_queue(
  * Clear the log incompat flags if we have the opportunity.
  *
  * This only happens if we're about to log the second dummy transaction as part
- * of covering the log and we can get the log incompat feature usage lock.
+ * of covering the log.
  */
 static inline void
 xlog_clear_incompat(
@@ -1463,11 +1463,7 @@ xlog_clear_incompat(
 	if (log->l_covered_state != XLOG_STATE_COVER_DONE2)
 		return;
 
-	if (!down_write_trylock(&log->l_incompat_users))
-		return;
-
 	xfs_clear_incompat_log_features(mp);
-	up_write(&log->l_incompat_users);
 }
 
 /*
@@ -1528,7 +1524,7 @@ xlog_alloc_log(
 	int			error = -ENOMEM;
 	uint			log2_size = 0;
 
-	log = kmem_zalloc(sizeof(struct xlog), KM_MAYFAIL);
+	log = kzalloc(sizeof(struct xlog), GFP_KERNEL | __GFP_RETRY_MAYFAIL);
 	if (!log) {
 		xfs_warn(mp, "Log allocation failed: No memory!");
 		goto out;
@@ -1585,8 +1581,6 @@ xlog_alloc_log(
 	}
 	log->l_sectBBsize = 1 << log2_size;
 
-	init_rwsem(&log->l_incompat_users);
-
 	xlog_get_iclog_buffer_size(mp, log);
 
 	spin_lock_init(&log->l_icloglock);
@@ -1605,7 +1599,8 @@ xlog_alloc_log(
 		size_t bvec_size = howmany(log->l_iclog_size, PAGE_SIZE) *
 				sizeof(struct bio_vec);
 
-		iclog = kmem_zalloc(sizeof(*iclog) + bvec_size, KM_MAYFAIL);
+		iclog = kzalloc(sizeof(*iclog) + bvec_size,
+				GFP_KERNEL | __GFP_RETRY_MAYFAIL);
 		if (!iclog)
 			goto out_free_iclog;
 
@@ -1661,13 +1656,13 @@ out_destroy_workqueue:
 out_free_iclog:
 	for (iclog = log->l_iclog; iclog; iclog = prev_iclog) {
 		prev_iclog = iclog->ic_next;
-		kmem_free(iclog->ic_data);
-		kmem_free(iclog);
+		kvfree(iclog->ic_data);
+		kfree(iclog);
 		if (prev_iclog == log->l_iclog)
 			break;
 	}
 out_free_log:
-	kmem_free(log);
+	kfree(log);
 out:
 	return ERR_PTR(error);
 }	/* xlog_alloc_log */
@@ -2118,14 +2113,14 @@ xlog_dealloc_log(
 	iclog = log->l_iclog;
 	for (i = 0; i < log->l_iclog_bufs; i++) {
 		next_iclog = iclog->ic_next;
-		kmem_free(iclog->ic_data);
-		kmem_free(iclog);
+		kvfree(iclog->ic_data);
+		kfree(iclog);
 		iclog = next_iclog;
 	}
 
 	log->l_mp->m_log = NULL;
 	destroy_workqueue(log->l_ioend_workqueue);
-	kmem_free(log);
+	kfree(log);
 }
 
 /*
@@ -3517,7 +3512,8 @@ xlog_ticket_alloc(
 	struct xlog_ticket	*tic;
 	int			unit_res;
 
-	tic = kmem_cache_zalloc(xfs_log_ticket_cache, GFP_NOFS | __GFP_NOFAIL);
+	tic = kmem_cache_zalloc(xfs_log_ticket_cache,
+			GFP_KERNEL | __GFP_NOFAIL);
 
 	unit_res = xlog_calc_unit_res(log, unit_bytes, &tic->t_iclog_hdrs);
 
@@ -3869,23 +3865,3 @@ xfs_log_check_lsn(
 
 	return valid;
 }
-
-/*
- * Notify the log that we're about to start using a feature that is protected
- * by a log incompat feature flag.  This will prevent log covering from
- * clearing those flags.
- */
-void
-xlog_use_incompat_feat(
-	struct xlog		*log)
-{
-	down_read(&log->l_incompat_users);
-}
-
-/* Notify the log that we've finished using log incompat features. */
-void
-xlog_drop_incompat_feat(
-	struct xlog		*log)
-{
-	up_read(&log->l_incompat_users);
-}
diff --git a/fs/xfs/xfs_log.h b/fs/xfs/xfs_log.h
index 2728886c29..d69acf8811 100644
--- a/fs/xfs/xfs_log.h
+++ b/fs/xfs/xfs_log.h
@@ -159,8 +159,6 @@ bool	xfs_log_check_lsn(struct xfs_mount *, xfs_lsn_t);
 xfs_lsn_t xlog_grant_push_threshold(struct xlog *log, int need_bytes);
 bool	  xlog_force_shutdown(struct xlog *log, uint32_t shutdown_flags);
 
-void xlog_use_incompat_feat(struct xlog *log);
-void xlog_drop_incompat_feat(struct xlog *log);
 int xfs_attr_use_log_assist(struct xfs_mount *mp);
 
 #endif	/* __XFS_LOG_H__ */
diff --git a/fs/xfs/xfs_log_cil.c b/fs/xfs/xfs_log_cil.c
index 67a99d9470..f51cbc6405 100644
--- a/fs/xfs/xfs_log_cil.c
+++ b/fs/xfs/xfs_log_cil.c
@@ -100,7 +100,7 @@ xlog_cil_ctx_alloc(void)
 {
 	struct xfs_cil_ctx	*ctx;
 
-	ctx = kmem_zalloc(sizeof(*ctx), KM_NOFS);
+	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL | __GFP_NOFAIL);
 	INIT_LIST_HEAD(&ctx->committing);
 	INIT_LIST_HEAD(&ctx->busy_extents.extent_list);
 	INIT_LIST_HEAD(&ctx->log_items);
@@ -339,7 +339,7 @@ xlog_cil_alloc_shadow_bufs(
 			 * the buffer, only the log vector header and the iovec
 			 * storage.
 			 */
-			kmem_free(lip->li_lv_shadow);
+			kvfree(lip->li_lv_shadow);
 			lv = xlog_kvmalloc(buf_size);
 
 			memset(lv, 0, xlog_cil_iovec_space(niovecs));
@@ -703,7 +703,7 @@ xlog_cil_free_logvec(
 	while (!list_empty(lv_chain)) {
 		lv = list_first_entry(lv_chain, struct xfs_log_vec, lv_list);
 		list_del_init(&lv->lv_list);
-		kmem_free(lv);
+		kvfree(lv);
 	}
 }
 
@@ -753,7 +753,7 @@ xlog_cil_committed(
 		return;
 	}
 
-	kmem_free(ctx);
+	kfree(ctx);
 }
 
 void
@@ -1116,11 +1116,18 @@ xlog_cil_cleanup_whiteouts(
  * same sequence twice.  If we get a race between multiple pushes for the same
  * sequence they will block on the first one and then abort, hence avoiding
  * needless pushes.
+ *
+ * This runs from a workqueue so it does not inherent any specific memory
+ * allocation context. However, we do not want to block on memory reclaim
+ * recursing back into the filesystem because this push may have been triggered
+ * by memory reclaim itself. Hence we really need to run under full GFP_NOFS
+ * contraints here.
  */
 static void
 xlog_cil_push_work(
 	struct work_struct	*work)
 {
+	unsigned int		nofs_flags = memalloc_nofs_save();
 	struct xfs_cil_ctx	*ctx =
 		container_of(work, struct xfs_cil_ctx, push_work);
 	struct xfs_cil		*cil = ctx->cil;
@@ -1334,12 +1341,14 @@ xlog_cil_push_work(
 	spin_unlock(&log->l_icloglock);
 	xlog_cil_cleanup_whiteouts(&whiteouts);
 	xfs_log_ticket_ungrant(log, ticket);
+	memalloc_nofs_restore(nofs_flags);
 	return;
 
 out_skip:
 	up_write(&cil->xc_ctx_lock);
 	xfs_log_ticket_put(new_ctx->ticket);
-	kmem_free(new_ctx);
+	kfree(new_ctx);
+	memalloc_nofs_restore(nofs_flags);
 	return;
 
 out_abort_free_ticket:
@@ -1348,6 +1357,7 @@ out_abort_free_ticket:
 	if (!ctx->commit_iclog) {
 		xfs_log_ticket_ungrant(log, ctx->ticket);
 		xlog_cil_committed(ctx);
+		memalloc_nofs_restore(nofs_flags);
 		return;
 	}
 	spin_lock(&log->l_icloglock);
@@ -1356,6 +1366,7 @@ out_abort_free_ticket:
 	/* Not safe to reference ctx now! */
 	spin_unlock(&log->l_icloglock);
 	xfs_log_ticket_ungrant(log, ticket);
+	memalloc_nofs_restore(nofs_flags);
 }
 
 /*
@@ -1367,7 +1378,7 @@ out_abort_free_ticket:
  */
 static void
 xlog_cil_push_background(
-	struct xlog	*log) __releases(cil->xc_ctx_lock)
+	struct xlog	*log)
 {
 	struct xfs_cil	*cil = log->l_cilp;
 	int		space_used = atomic_read(&cil->xc_ctx->space_used);
@@ -1533,7 +1544,7 @@ xlog_cil_process_intents(
 		set_bit(XFS_LI_WHITEOUT, &ilip->li_flags);
 		trace_xfs_cil_whiteout_mark(ilip);
 		len += ilip->li_lv->lv_bytes;
-		kmem_free(ilip->li_lv);
+		kvfree(ilip->li_lv);
 		ilip->li_lv = NULL;
 
 		xfs_trans_del_item(lip);
@@ -1747,7 +1758,7 @@ xlog_cil_init(
 	struct xlog_cil_pcp	*cilpcp;
 	int			cpu;
 
-	cil = kmem_zalloc(sizeof(*cil), KM_MAYFAIL);
+	cil = kzalloc(sizeof(*cil), GFP_KERNEL | __GFP_RETRY_MAYFAIL);
 	if (!cil)
 		return -ENOMEM;
 	/*
@@ -1786,7 +1797,7 @@ xlog_cil_init(
 out_destroy_wq:
 	destroy_workqueue(cil->xc_push_wq);
 out_destroy_cil:
-	kmem_free(cil);
+	kfree(cil);
 	return -ENOMEM;
 }
 
@@ -1799,12 +1810,12 @@ xlog_cil_destroy(
 	if (cil->xc_ctx) {
 		if (cil->xc_ctx->ticket)
 			xfs_log_ticket_put(cil->xc_ctx->ticket);
-		kmem_free(cil->xc_ctx);
+		kfree(cil->xc_ctx);
 	}
 
 	ASSERT(test_bit(XLOG_CIL_EMPTY, &cil->xc_flags));
 	free_percpu(cil->xc_pcp);
 	destroy_workqueue(cil->xc_push_wq);
-	kmem_free(cil);
+	kfree(cil);
 }
 
diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h
index e30c06ec20..40e22ec0fb 100644
--- a/fs/xfs/xfs_log_priv.h
+++ b/fs/xfs/xfs_log_priv.h
@@ -450,9 +450,6 @@ struct xlog {
 	xfs_lsn_t		l_recovery_lsn;
 
 	uint32_t		l_iclog_roundoff;/* padding roundoff */
-
-	/* Users of log incompat features should take a read lock. */
-	struct rw_semaphore	l_incompat_users;
 };
 
 /*
@@ -623,7 +620,8 @@ xlog_wait(
 	remove_wait_queue(wq, &wait);
 }
 
-int xlog_wait_on_iclog(struct xlog_in_core *iclog);
+int xlog_wait_on_iclog(struct xlog_in_core *iclog)
+		__releases(iclog->ic_log->l_icloglock);
 
 /*
  * The LSN is valid so long as it is behind the current LSN. If it isn't, this
@@ -683,7 +681,7 @@ xlog_valid_lsn(
  * flags to control the kmalloc() behaviour within kvmalloc(). Hence kmalloc()
  * will do direct reclaim and compaction in the slow path, both of which are
  * horrendously expensive. We just want kmalloc to fail fast and fall back to
- * vmalloc if it can't get somethign straight away from the free lists or
+ * vmalloc if it can't get something straight away from the free lists or
  * buddy allocator. Hence we have to open code kvmalloc outselves here.
  *
  * This assumes that the caller uses memalloc_nofs_save task context here, so
diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 1251c81e55..4fe627991e 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -361,7 +361,7 @@ xlog_find_verify_cycle(
 	*new_blk = -1;
 
 out:
-	kmem_free(buffer);
+	kvfree(buffer);
 	return error;
 }
 
@@ -477,7 +477,7 @@ xlog_find_verify_log_record(
 		*last_blk = i;
 
 out:
-	kmem_free(buffer);
+	kvfree(buffer);
 	return error;
 }
 
@@ -731,7 +731,7 @@ validate_head:
 			goto out_free_buffer;
 	}
 
-	kmem_free(buffer);
+	kvfree(buffer);
 	if (head_blk == log_bbnum)
 		*return_head_blk = 0;
 	else
@@ -745,7 +745,7 @@ validate_head:
 	return 0;
 
 out_free_buffer:
-	kmem_free(buffer);
+	kvfree(buffer);
 	if (error)
 		xfs_warn(log->l_mp, "failed to find log head");
 	return error;
@@ -999,7 +999,7 @@ xlog_verify_tail(
 		"Tail block (0x%llx) overwrite detected. Updated to 0x%llx",
 			 orig_tail, *tail_blk);
 out:
-	kmem_free(buffer);
+	kvfree(buffer);
 	return error;
 }
 
@@ -1046,7 +1046,7 @@ xlog_verify_head(
 	error = xlog_rseek_logrec_hdr(log, *head_blk, *tail_blk,
 				      XLOG_MAX_ICLOGS, tmp_buffer,
 				      &tmp_rhead_blk, &tmp_rhead, &tmp_wrapped);
-	kmem_free(tmp_buffer);
+	kvfree(tmp_buffer);
 	if (error < 0)
 		return error;
 
@@ -1365,7 +1365,7 @@ xlog_find_tail(
 		error = xlog_clear_stale_blocks(log, tail_lsn);
 
 done:
-	kmem_free(buffer);
+	kvfree(buffer);
 
 	if (error)
 		xfs_warn(log->l_mp, "failed to locate log tail");
@@ -1399,6 +1399,7 @@ xlog_find_zeroed(
 	xfs_daddr_t	new_blk, last_blk, start_blk;
 	xfs_daddr_t     num_scan_bblks;
 	int	        error, log_bbnum = log->l_logBBsize;
+	int		ret = 1;
 
 	*blk_no = 0;
 
@@ -1413,8 +1414,7 @@ xlog_find_zeroed(
 	first_cycle = xlog_get_cycle(offset);
 	if (first_cycle == 0) {		/* completely zeroed log */
 		*blk_no = 0;
-		kmem_free(buffer);
-		return 1;
+		goto out_free_buffer;
 	}
 
 	/* check partially zeroed log */
@@ -1424,8 +1424,8 @@ xlog_find_zeroed(
 
 	last_cycle = xlog_get_cycle(offset);
 	if (last_cycle != 0) {		/* log completely written to */
-		kmem_free(buffer);
-		return 0;
+		ret = 0;
+		goto out_free_buffer;
 	}
 
 	/* we have a partially zeroed log */
@@ -1471,10 +1471,10 @@ xlog_find_zeroed(
 
 	*blk_no = last_blk;
 out_free_buffer:
-	kmem_free(buffer);
+	kvfree(buffer);
 	if (error)
 		return error;
-	return 1;
+	return ret;
 }
 
 /*
@@ -1583,7 +1583,7 @@ xlog_write_log_records(
 	}
 
 out_free_buffer:
-	kmem_free(buffer);
+	kvfree(buffer);
 	return error;
 }
 
@@ -1767,6 +1767,37 @@ xlog_recover_iget(
 	return 0;
 }
 
+/*
+ * Get an inode so that we can recover a log operation.
+ *
+ * Log intent items that target inodes effectively contain a file handle.
+ * Check that the generation number matches the intent item like we do for
+ * other file handles.  Log intent items defined after this validation weakness
+ * was identified must use this function.
+ */
+int
+xlog_recover_iget_handle(
+	struct xfs_mount	*mp,
+	xfs_ino_t		ino,
+	uint32_t		gen,
+	struct xfs_inode	**ipp)
+{
+	struct xfs_inode	*ip;
+	int			error;
+
+	error = xlog_recover_iget(mp, ino, &ip);
+	if (error)
+		return error;
+
+	if (VFS_I(ip)->i_generation != gen) {
+		xfs_irele(ip);
+		return -EFSCORRUPTED;
+	}
+
+	*ipp = ip;
+	return 0;
+}
+
 /******************************************************************************
  *
  *		Log recover routines
@@ -1789,6 +1820,8 @@ static const struct xlog_recover_item_ops *xlog_recover_item_ops[] = {
 	&xlog_bud_item_ops,
 	&xlog_attri_item_ops,
 	&xlog_attrd_item_ops,
+	&xlog_xmi_item_ops,
+	&xlog_xmd_item_ops,
 };
 
 static const struct xlog_recover_item_ops *
@@ -2057,7 +2090,8 @@ xlog_recover_add_item(
 {
 	struct xlog_recover_item *item;
 
-	item = kmem_zalloc(sizeof(struct xlog_recover_item), 0);
+	item = kzalloc(sizeof(struct xlog_recover_item),
+			GFP_KERNEL | __GFP_NOFAIL);
 	INIT_LIST_HEAD(&item->ri_list);
 	list_add_tail(&item->ri_list, head);
 }
@@ -2160,7 +2194,7 @@ xlog_recover_add_to_trans(
 		return 0;
 	}
 
-	ptr = kmem_alloc(len, 0);
+	ptr = xlog_kvmalloc(len);
 	memcpy(ptr, dp, len);
 	in_f = (struct xfs_inode_log_format *)ptr;
 
@@ -2182,14 +2216,13 @@ xlog_recover_add_to_trans(
 		"bad number of regions (%d) in inode log format",
 				  in_f->ilf_size);
 			ASSERT(0);
-			kmem_free(ptr);
+			kvfree(ptr);
 			return -EFSCORRUPTED;
 		}
 
 		item->ri_total = in_f->ilf_size;
-		item->ri_buf =
-			kmem_zalloc(item->ri_total * sizeof(xfs_log_iovec_t),
-				    0);
+		item->ri_buf = kzalloc(item->ri_total * sizeof(xfs_log_iovec_t),
+				GFP_KERNEL | __GFP_NOFAIL);
 	}
 
 	if (item->ri_total <= item->ri_cnt) {
@@ -2197,7 +2230,7 @@ xlog_recover_add_to_trans(
 	"log item region count (%d) overflowed size (%d)",
 				item->ri_cnt, item->ri_total);
 		ASSERT(0);
-		kmem_free(ptr);
+		kvfree(ptr);
 		return -EFSCORRUPTED;
 	}
 
@@ -2227,13 +2260,13 @@ xlog_recover_free_trans(
 		/* Free the regions in the item. */
 		list_del(&item->ri_list);
 		for (i = 0; i < item->ri_cnt; i++)
-			kmem_free(item->ri_buf[i].i_addr);
+			kvfree(item->ri_buf[i].i_addr);
 		/* Free the item itself */
-		kmem_free(item->ri_buf);
-		kmem_free(item);
+		kfree(item->ri_buf);
+		kfree(item);
 	}
 	/* Free the transaction recover structure */
-	kmem_free(trans);
+	kfree(trans);
 }
 
 /*
@@ -2332,7 +2365,7 @@ xlog_recover_ophdr_to_trans(
 	 * This is a new transaction so allocate a new recovery container to
 	 * hold the recovery ops that will follow.
 	 */
-	trans = kmem_zalloc(sizeof(struct xlog_recover), 0);
+	trans = kzalloc(sizeof(struct xlog_recover), GFP_KERNEL | __GFP_NOFAIL);
 	trans->r_log_tid = tid;
 	trans->r_lsn = be64_to_cpu(rhead->h_lsn);
 	INIT_LIST_HEAD(&trans->r_itemq);
@@ -2656,7 +2689,7 @@ xlog_recover_clear_agi_bucket(
 	if (error)
 		goto out_error;
 
-	error = xfs_read_agi(pag, tp, &agibp);
+	error = xfs_read_agi(pag, tp, 0, &agibp);
 	if (error)
 		goto out_abort;
 
@@ -2772,7 +2805,7 @@ xlog_recover_iunlink_ag(
 	int			bucket;
 	int			error;
 
-	error = xfs_read_agi(pag, NULL, &agibp);
+	error = xfs_read_agi(pag, NULL, 0, &agibp);
 	if (error) {
 		/*
 		 * AGI is b0rked. Don't process it.
@@ -2966,7 +2999,7 @@ xlog_do_recovery_pass(
 	int			error = 0, h_size, h_len;
 	int			error2 = 0;
 	int			bblks, split_bblks;
-	int			hblks, split_hblks, wrapped_hblks;
+	int			hblks = 1, split_hblks, wrapped_hblks;
 	int			i;
 	struct hlist_head	rhash[XLOG_RHASH_SIZE];
 	LIST_HEAD		(buffer_list);
@@ -2977,6 +3010,10 @@ xlog_do_recovery_pass(
 	for (i = 0; i < XLOG_RHASH_SIZE; i++)
 		INIT_HLIST_HEAD(&rhash[i]);
 
+	hbp = xlog_alloc_buffer(log, hblks);
+	if (!hbp)
+		return -ENOMEM;
+
 	/*
 	 * Read the header of the tail block and get the iclog buffer size from
 	 * h_size.  Use this to tell how many sectors make up the log header.
@@ -2987,10 +3024,6 @@ xlog_do_recovery_pass(
 		 * iclog header and extract the header size from it.  Get a
 		 * new hbp that is the correct size.
 		 */
-		hbp = xlog_alloc_buffer(log, 1);
-		if (!hbp)
-			return -ENOMEM;
-
 		error = xlog_bread(log, tail_blk, 1, hbp, &offset);
 		if (error)
 			goto bread_err1;
@@ -3022,23 +3055,30 @@ xlog_do_recovery_pass(
 		if (error)
 			goto bread_err1;
 
-		hblks = xlog_logrec_hblks(log, rhead);
-		if (hblks != 1) {
-			kmem_free(hbp);
-			hbp = xlog_alloc_buffer(log, hblks);
+		/*
+		 * This open codes xlog_logrec_hblks so that we can reuse the
+		 * fixed up h_size value calculated above.  Without that we'd
+		 * still allocate the buffer based on the incorrect on-disk
+		 * size.
+		 */
+		if (h_size > XLOG_HEADER_CYCLE_SIZE &&
+		    (rhead->h_version & cpu_to_be32(XLOG_VERSION_2))) {
+			hblks = DIV_ROUND_UP(h_size, XLOG_HEADER_CYCLE_SIZE);
+			if (hblks > 1) {
+				kvfree(hbp);
+				hbp = xlog_alloc_buffer(log, hblks);
+				if (!hbp)
+					return -ENOMEM;
+			}
 		}
 	} else {
 		ASSERT(log->l_sectBBsize == 1);
-		hblks = 1;
-		hbp = xlog_alloc_buffer(log, 1);
 		h_size = XLOG_BIG_RECORD_BSIZE;
 	}
 
-	if (!hbp)
-		return -ENOMEM;
 	dbp = xlog_alloc_buffer(log, BTOBB(h_size));
 	if (!dbp) {
-		kmem_free(hbp);
+		kvfree(hbp);
 		return -ENOMEM;
 	}
 
@@ -3199,16 +3239,33 @@ xlog_do_recovery_pass(
 	}
 
  bread_err2:
-	kmem_free(dbp);
+	kvfree(dbp);
  bread_err1:
-	kmem_free(hbp);
+	kvfree(hbp);
 
 	/*
-	 * Submit buffers that have been added from the last record processed,
-	 * regardless of error status.
+	 * Submit buffers that have been dirtied by the last record recovered.
 	 */
-	if (!list_empty(&buffer_list))
+	if (!list_empty(&buffer_list)) {
+		if (error) {
+			/*
+			 * If there has been an item recovery error then we
+			 * cannot allow partial checkpoint writeback to
+			 * occur.  We might have multiple checkpoints with the
+			 * same start LSN in this buffer list, and partial
+			 * writeback of a checkpoint in this situation can
+			 * prevent future recovery of all the changes in the
+			 * checkpoints at this start LSN.
+			 *
+			 * Note: Shutting down the filesystem will result in the
+			 * delwri submission marking all the buffers stale,
+			 * completing them and cleaning up _XBF_LOGRECOVERY
+			 * state without doing any IO.
+			 */
+			xlog_force_shutdown(log, SHUTDOWN_LOG_IO_ERROR);
+		}
 		error2 = xfs_buf_delwri_submit(&buffer_list);
+	}
 
 	if (error && first_bad)
 		*first_bad = rhead_blk;
@@ -3443,12 +3500,19 @@ xlog_recover(
  * part of recovery so that the root and real-time bitmap inodes can be read in
  * from disk in between the two stages.  This is necessary so that we can free
  * space in the real-time portion of the file system.
+ *
+ * We run this whole process under GFP_NOFS allocation context. We do a
+ * combination of non-transactional and transactional work, yet we really don't
+ * want to recurse into the filesystem from direct reclaim during any of this
+ * processing. This allows all the recovery code run here not to care about the
+ * memory allocation context it is running in.
  */
 int
 xlog_recover_finish(
 	struct xlog	*log)
 {
-	int	error;
+	unsigned int	nofs_flags = memalloc_nofs_save();
+	int		error;
 
 	error = xlog_recover_process_intents(log);
 	if (error) {
@@ -3462,7 +3526,7 @@ xlog_recover_finish(
 		xlog_recover_cancel_intents(log);
 		xfs_alert(log->l_mp, "Failed to recover intents");
 		xlog_force_shutdown(log, SHUTDOWN_LOG_IO_ERROR);
-		return error;
+		goto out_error;
 	}
 
 	/*
@@ -3472,21 +3536,6 @@ xlog_recover_finish(
 	 */
 	xfs_log_force(log->l_mp, XFS_LOG_SYNC);
 
-	/*
-	 * Now that we've recovered the log and all the intents, we can clear
-	 * the log incompat feature bits in the superblock because there's no
-	 * longer anything to protect.  We rely on the AIL push to write out the
-	 * updated superblock after everything else.
-	 */
-	if (xfs_clear_incompat_log_features(log->l_mp)) {
-		error = xfs_sync_sb(log->l_mp, false);
-		if (error < 0) {
-			xfs_alert(log->l_mp,
-	"Failed to clear log incompat features on recovery");
-			return error;
-		}
-	}
-
 	xlog_recover_process_iunlinks(log);
 
 	/*
@@ -3508,9 +3557,13 @@ xlog_recover_finish(
 		 * and AIL.
 		 */
 		xlog_force_shutdown(log, SHUTDOWN_LOG_IO_ERROR);
+		error = 0;
+		goto out_error;
 	}
 
-	return 0;
+out_error:
+	memalloc_nofs_restore(nofs_flags);
+	return error;
 }
 
 void
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
index aabb25dc3e..09eef1721e 100644
--- a/fs/xfs/xfs_mount.c
+++ b/fs/xfs/xfs_mount.c
@@ -34,6 +34,7 @@
 #include "xfs_health.h"
 #include "xfs_trace.h"
 #include "xfs_ag.h"
+#include "xfs_rtbitmap.h"
 #include "scrub/stats.h"
 
 static DEFINE_MUTEX(xfs_uuid_table_mutex);
@@ -45,7 +46,7 @@ xfs_uuid_table_free(void)
 {
 	if (xfs_uuid_table_size == 0)
 		return;
-	kmem_free(xfs_uuid_table);
+	kfree(xfs_uuid_table);
 	xfs_uuid_table = NULL;
 	xfs_uuid_table_size = 0;
 }
@@ -62,7 +63,7 @@ xfs_uuid_mount(
 	int			hole, i;
 
 	/* Publish UUID in struct super_block */
-	uuid_copy(&mp->m_super->s_uuid, uuid);
+	super_set_uuid(mp->m_super, uuid->b, sizeof(*uuid));
 
 	if (xfs_has_nouuid(mp))
 		return 0;
@@ -230,6 +231,13 @@ reread:
 	mp->m_features |= xfs_sb_version_to_features(sbp);
 	xfs_reinit_percpu_counters(mp);
 
+	/*
+	 * If logged xattrs are enabled after log recovery finishes, then set
+	 * the opstate so that log recovery will work properly.
+	 */
+	if (xfs_sb_version_haslogxattrs(&mp->m_sb))
+		xfs_set_using_logged_xattrs(mp);
+
 	/* no need to be quiet anymore, so reset the buf ops */
 	bp->b_ops = &xfs_sb_buf_ops;
 
@@ -706,6 +714,8 @@ xfs_mountfs(
 	/* enable fail_at_unmount as default */
 	mp->m_fail_unmount = true;
 
+	super_set_sysfs_name_id(mp->m_super);
+
 	error = xfs_sysfs_init(&mp->m_kobj, &xfs_mp_ktype,
 			       NULL, mp->m_super->s_id);
 	if (error)
@@ -826,6 +836,15 @@ xfs_mountfs(
 		goto out_inodegc_shrinker;
 	}
 
+	/*
+	 * If logged xattrs are still enabled after log recovery finishes, then
+	 * they'll be available until unmount.  Otherwise, turn them off.
+	 */
+	if (xfs_sb_version_haslogxattrs(&mp->m_sb))
+		xfs_set_using_logged_xattrs(mp);
+	else
+		xfs_clear_using_logged_xattrs(mp);
+
 	/* Enable background inode inactivation workers. */
 	xfs_inodegc_start(mp);
 	xfs_blockgc_start(mp);
@@ -1093,6 +1112,11 @@ xfs_unmountfs(
 				"Freespace may not be correct on next mount.");
 	xfs_unmount_check(mp);
 
+	/*
+	 * Indicate that it's ok to clear log incompat bits before cleaning
+	 * the log and writing the unmount record.
+	 */
+	xfs_set_done_with_log_incompat(mp);
 	xfs_log_unmount(mp);
 	xfs_da_unmount(mp);
 	xfs_uuid_unmount(mp);
@@ -1129,16 +1153,44 @@ xfs_fs_writable(
 	return true;
 }
 
-/* Adjust m_fdblocks or m_frextents. */
+void
+xfs_add_freecounter(
+	struct xfs_mount	*mp,
+	struct percpu_counter	*counter,
+	uint64_t		delta)
+{
+	bool			has_resv_pool = (counter == &mp->m_fdblocks);
+	uint64_t		res_used;
+
+	/*
+	 * If the reserve pool is depleted, put blocks back into it first.
+	 * Most of the time the pool is full.
+	 */
+	if (!has_resv_pool || mp->m_resblks == mp->m_resblks_avail) {
+		percpu_counter_add(counter, delta);
+		return;
+	}
+
+	spin_lock(&mp->m_sb_lock);
+	res_used = mp->m_resblks - mp->m_resblks_avail;
+	if (res_used > delta) {
+		mp->m_resblks_avail += delta;
+	} else {
+		delta -= res_used;
+		mp->m_resblks_avail = mp->m_resblks;
+		percpu_counter_add(counter, delta);
+	}
+	spin_unlock(&mp->m_sb_lock);
+}
+
 int
-xfs_mod_freecounter(
+xfs_dec_freecounter(
 	struct xfs_mount	*mp,
 	struct percpu_counter	*counter,
-	int64_t			delta,
+	uint64_t		delta,
 	bool			rsvd)
 {
 	int64_t			lcounter;
-	long long		res_used;
 	uint64_t		set_aside = 0;
 	s32			batch;
 	bool			has_resv_pool;
@@ -1148,31 +1200,6 @@ xfs_mod_freecounter(
 	if (rsvd)
 		ASSERT(has_resv_pool);
 
-	if (delta > 0) {
-		/*
-		 * If the reserve pool is depleted, put blocks back into it
-		 * first. Most of the time the pool is full.
-		 */
-		if (likely(!has_resv_pool ||
-			   mp->m_resblks == mp->m_resblks_avail)) {
-			percpu_counter_add(counter, delta);
-			return 0;
-		}
-
-		spin_lock(&mp->m_sb_lock);
-		res_used = (long long)(mp->m_resblks - mp->m_resblks_avail);
-
-		if (res_used > delta) {
-			mp->m_resblks_avail += delta;
-		} else {
-			delta -= res_used;
-			mp->m_resblks_avail = mp->m_resblks;
-			percpu_counter_add(counter, delta);
-		}
-		spin_unlock(&mp->m_sb_lock);
-		return 0;
-	}
-
 	/*
 	 * Taking blocks away, need to be more accurate the closer we
 	 * are to zero.
@@ -1200,7 +1227,7 @@ xfs_mod_freecounter(
 	 */
 	if (has_resv_pool)
 		set_aside = xfs_fdblocks_unavailable(mp);
-	percpu_counter_add_batch(counter, delta, batch);
+	percpu_counter_add_batch(counter, -((int64_t)delta), batch);
 	if (__percpu_counter_compare(counter, set_aside,
 				     XFS_FDBLOCKS_BATCH) >= 0) {
 		/* we had space! */
@@ -1212,11 +1239,11 @@ xfs_mod_freecounter(
 	 * that took us to ENOSPC.
 	 */
 	spin_lock(&mp->m_sb_lock);
-	percpu_counter_add(counter, -delta);
+	percpu_counter_add(counter, delta);
 	if (!has_resv_pool || !rsvd)
 		goto fdblocks_enospc;
 
-	lcounter = (long long)mp->m_resblks_avail + delta;
+	lcounter = (long long)mp->m_resblks_avail - delta;
 	if (lcounter >= 0) {
 		mp->m_resblks_avail = lcounter;
 		spin_unlock(&mp->m_sb_lock);
@@ -1362,7 +1389,8 @@ xfs_clear_incompat_log_features(
 	if (!xfs_has_crc(mp) ||
 	    !xfs_sb_has_incompat_log_feature(&mp->m_sb,
 				XFS_SB_FEAT_INCOMPAT_LOG_ALL) ||
-	    xfs_is_shutdown(mp))
+	    xfs_is_shutdown(mp) ||
+	    !xfs_is_done_with_log_incompat(mp))
 		return false;
 
 	/*
@@ -1397,9 +1425,20 @@ xfs_clear_incompat_log_features(
 #define XFS_DELALLOC_BATCH	(4096)
 void
 xfs_mod_delalloc(
-	struct xfs_mount	*mp,
-	int64_t			delta)
+	struct xfs_inode	*ip,
+	int64_t			data_delta,
+	int64_t			ind_delta)
 {
-	percpu_counter_add_batch(&mp->m_delalloc_blks, delta,
+	struct xfs_mount	*mp = ip->i_mount;
+
+	if (XFS_IS_REALTIME_INODE(ip)) {
+		percpu_counter_add_batch(&mp->m_delalloc_rtextents,
+				xfs_rtb_to_rtx(mp, data_delta),
+				XFS_DELALLOC_BATCH);
+		if (!ind_delta)
+			return;
+		data_delta = 0;
+	}
+	percpu_counter_add_batch(&mp->m_delalloc_blks, data_delta + ind_delta,
 			XFS_DELALLOC_BATCH);
 }
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index 503fe3c7ed..d0567dfbc0 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -94,9 +94,9 @@ typedef struct xfs_mount {
 	struct xfs_inode	*m_rsumip;	/* pointer to summary inode */
 	struct xfs_inode	*m_rootip;	/* pointer to root directory */
 	struct xfs_quotainfo	*m_quotainfo;	/* disk quota information */
-	xfs_buftarg_t		*m_ddev_targp;	/* saves taking the address */
-	xfs_buftarg_t		*m_logdev_targp;/* ptr to log device */
-	xfs_buftarg_t		*m_rtdev_targp;	/* ptr to rt device */
+	struct xfs_buftarg	*m_ddev_targp;	/* data device */
+	struct xfs_buftarg	*m_logdev_targp;/* log device */
+	struct xfs_buftarg	*m_rtdev_targp;	/* rt device */
 	void __percpu		*m_inodegc;	/* percpu inodegc structures */
 
 	/*
@@ -195,6 +195,12 @@ typedef struct xfs_mount {
 	 * extents or anything related to the rt device.
 	 */
 	struct percpu_counter	m_delalloc_blks;
+
+	/*
+	 * RT version of the above.
+	 */
+	struct percpu_counter	m_delalloc_rtextents;
+
 	/*
 	 * Global count of allocation btree blocks in use across all AGs. Only
 	 * used when perag reservation is enabled. Helps prevent block
@@ -252,6 +258,9 @@ typedef struct xfs_mount {
 
 	/* cpus that have inodes queued for inactivation */
 	struct cpumask		m_inodegc_cpumask;
+
+	/* Hook to feed dirent updates to an active online repair. */
+	struct xfs_hooks	m_dir_update_hooks;
 } xfs_mount_t;
 
 #define M_IGEO(mp)		(&(mp)->m_ino_geo)
@@ -289,6 +298,7 @@ typedef struct xfs_mount {
 #define XFS_FEAT_BIGTIME	(1ULL << 24)	/* large timestamps */
 #define XFS_FEAT_NEEDSREPAIR	(1ULL << 25)	/* needs xfs_repair */
 #define XFS_FEAT_NREXT64	(1ULL << 26)	/* large extent counters */
+#define XFS_FEAT_EXCHANGE_RANGE	(1ULL << 27)	/* exchange range */
 
 /* Mount features */
 #define XFS_FEAT_NOATTR2	(1ULL << 48)	/* disable attr2 creation */
@@ -328,19 +338,10 @@ static inline void xfs_add_ ## name (struct xfs_mount *mp) \
 __XFS_ADD_FEAT(attr, ATTR)
 __XFS_HAS_FEAT(nlink, NLINK)
 __XFS_ADD_FEAT(quota, QUOTA)
-__XFS_HAS_FEAT(align, ALIGN)
 __XFS_HAS_FEAT(dalign, DALIGN)
-__XFS_HAS_FEAT(logv2, LOGV2)
 __XFS_HAS_FEAT(sector, SECTOR)
-__XFS_HAS_FEAT(extflg, EXTFLG)
 __XFS_HAS_FEAT(asciici, ASCIICI)
-__XFS_HAS_FEAT(lazysbcount, LAZYSBCOUNT)
-__XFS_ADD_FEAT(attr2, ATTR2)
 __XFS_HAS_FEAT(parent, PARENT)
-__XFS_ADD_FEAT(projid32, PROJID32)
-__XFS_HAS_FEAT(crc, CRC)
-__XFS_HAS_FEAT(v3inodes, V3INODES)
-__XFS_HAS_FEAT(pquotino, PQUOTINO)
 __XFS_HAS_FEAT(ftype, FTYPE)
 __XFS_HAS_FEAT(finobt, FINOBT)
 __XFS_HAS_FEAT(rmapbt, RMAPBT)
@@ -352,6 +353,38 @@ __XFS_HAS_FEAT(inobtcounts, INOBTCNT)
 __XFS_HAS_FEAT(bigtime, BIGTIME)
 __XFS_HAS_FEAT(needsrepair, NEEDSREPAIR)
 __XFS_HAS_FEAT(large_extent_counts, NREXT64)
+__XFS_HAS_FEAT(exchange_range, EXCHANGE_RANGE)
+
+/*
+ * Some features are always on for v5 file systems, allow the compiler to
+ * eliminiate dead code when building without v4 support.
+ */
+#define __XFS_HAS_V4_FEAT(name, NAME) \
+static inline bool xfs_has_ ## name (struct xfs_mount *mp) \
+{ \
+	return !IS_ENABLED(CONFIG_XFS_SUPPORT_V4) || \
+		(mp->m_features & XFS_FEAT_ ## NAME); \
+}
+
+#define __XFS_ADD_V4_FEAT(name, NAME) \
+	__XFS_HAS_V4_FEAT(name, NAME); \
+static inline void xfs_add_ ## name (struct xfs_mount *mp) \
+{ \
+	if (IS_ENABLED(CONFIG_XFS_SUPPORT_V4)) { \
+		mp->m_features |= XFS_FEAT_ ## NAME; \
+		xfs_sb_version_add ## name(&mp->m_sb); \
+	} \
+}
+
+__XFS_HAS_V4_FEAT(align, ALIGN)
+__XFS_HAS_V4_FEAT(logv2, LOGV2)
+__XFS_HAS_V4_FEAT(extflg, EXTFLG)
+__XFS_HAS_V4_FEAT(lazysbcount, LAZYSBCOUNT)
+__XFS_ADD_V4_FEAT(attr2, ATTR2)
+__XFS_ADD_V4_FEAT(projid32, PROJID32)
+__XFS_HAS_V4_FEAT(v3inodes, V3INODES)
+__XFS_HAS_V4_FEAT(crc, CRC)
+__XFS_HAS_V4_FEAT(pquotino, PQUOTINO)
 
 /*
  * Mount features
@@ -409,6 +442,10 @@ __XFS_HAS_FEAT(nouuid, NOUUID)
 #define XFS_OPSTATE_WARNED_LARP		9
 /* Mount time quotacheck is running */
 #define XFS_OPSTATE_QUOTACHECK_RUNNING	10
+/* Do we want to clear log incompat flags? */
+#define XFS_OPSTATE_UNSET_LOG_INCOMPAT	11
+/* Filesystem can use logged extended attributes */
+#define XFS_OPSTATE_USE_LARP		12
 
 #define __XFS_IS_OPSTATE(name, NAME) \
 static inline bool xfs_is_ ## name (struct xfs_mount *mp) \
@@ -436,6 +473,8 @@ __XFS_IS_OPSTATE(quotacheck_running, QUOTACHECK_RUNNING)
 #else
 # define xfs_is_quotacheck_running(mp)	(false)
 #endif
+__XFS_IS_OPSTATE(done_with_log_incompat, UNSET_LOG_INCOMPAT)
+__XFS_IS_OPSTATE(using_logged_xattrs, USE_LARP)
 
 static inline bool
 xfs_should_warn(struct xfs_mount *mp, long nr)
@@ -454,7 +493,9 @@ xfs_should_warn(struct xfs_mount *mp, long nr)
 	{ (1UL << XFS_OPSTATE_WARNED_SCRUB),		"wscrub" }, \
 	{ (1UL << XFS_OPSTATE_WARNED_SHRINK),		"wshrink" }, \
 	{ (1UL << XFS_OPSTATE_WARNED_LARP),		"wlarp" }, \
-	{ (1UL << XFS_OPSTATE_QUOTACHECK_RUNNING),	"quotacheck" }
+	{ (1UL << XFS_OPSTATE_QUOTACHECK_RUNNING),	"quotacheck" }, \
+	{ (1UL << XFS_OPSTATE_UNSET_LOG_INCOMPAT),	"unset_log_incompat" }, \
+	{ (1UL << XFS_OPSTATE_USE_LARP),		"logged_xattrs" }
 
 /*
  * Max and min values for mount-option defined I/O
@@ -502,9 +543,6 @@ xfs_daddr_to_agbno(struct xfs_mount *mp, xfs_daddr_t d)
 	return (xfs_agblock_t) do_div(ld, mp->m_sb.sb_agblocks);
 }
 
-int xfs_buf_hash_init(struct xfs_perag *pag);
-void xfs_buf_hash_destroy(struct xfs_perag *pag);
-
 extern void	xfs_uuid_table_free(void);
 extern uint64_t xfs_default_resblks(xfs_mount_t *mp);
 extern int	xfs_mountfs(xfs_mount_t *mp);
@@ -534,19 +572,30 @@ xfs_fdblocks_unavailable(
 	return mp->m_alloc_set_aside + atomic64_read(&mp->m_allocbt_blks);
 }
 
-int xfs_mod_freecounter(struct xfs_mount *mp, struct percpu_counter *counter,
-		int64_t delta, bool rsvd);
+int xfs_dec_freecounter(struct xfs_mount *mp, struct percpu_counter *counter,
+		uint64_t delta, bool rsvd);
+void xfs_add_freecounter(struct xfs_mount *mp, struct percpu_counter *counter,
+		uint64_t delta);
+
+static inline int xfs_dec_fdblocks(struct xfs_mount *mp, uint64_t delta,
+		bool reserved)
+{
+	return xfs_dec_freecounter(mp, &mp->m_fdblocks, delta, reserved);
+}
+
+static inline void xfs_add_fdblocks(struct xfs_mount *mp, uint64_t delta)
+{
+	xfs_add_freecounter(mp, &mp->m_fdblocks, delta);
+}
 
-static inline int
-xfs_mod_fdblocks(struct xfs_mount *mp, int64_t delta, bool reserved)
+static inline int xfs_dec_frextents(struct xfs_mount *mp, uint64_t delta)
 {
-	return xfs_mod_freecounter(mp, &mp->m_fdblocks, delta, reserved);
+	return xfs_dec_freecounter(mp, &mp->m_frextents, delta, false);
 }
 
-static inline int
-xfs_mod_frextents(struct xfs_mount *mp, int64_t delta)
+static inline void xfs_add_frextents(struct xfs_mount *mp, uint64_t delta)
 {
-	return xfs_mod_freecounter(mp, &mp->m_frextents, delta, false);
+	xfs_add_freecounter(mp, &mp->m_frextents, delta);
 }
 
 extern int	xfs_readsb(xfs_mount_t *, int);
@@ -566,6 +615,7 @@ struct xfs_error_cfg * xfs_error_get_cfg(struct xfs_mount *mp,
 void xfs_force_summary_recalc(struct xfs_mount *mp);
 int xfs_add_incompat_log_feature(struct xfs_mount *mp, uint32_t feature);
 bool xfs_clear_incompat_log_features(struct xfs_mount *mp);
-void xfs_mod_delalloc(struct xfs_mount *mp, int64_t delta);
+void xfs_mod_delalloc(struct xfs_inode *ip, int64_t data_delta,
+		int64_t ind_delta);
 
 #endif	/* __XFS_MOUNT_H__ */
diff --git a/fs/xfs/xfs_mru_cache.c b/fs/xfs/xfs_mru_cache.c
index f85e3b07ab..7443debaff 100644
--- a/fs/xfs/xfs_mru_cache.c
+++ b/fs/xfs/xfs_mru_cache.c
@@ -333,13 +333,14 @@ xfs_mru_cache_create(
 	if (!(grp_time = msecs_to_jiffies(lifetime_ms) / grp_count))
 		return -EINVAL;
 
-	if (!(mru = kmem_zalloc(sizeof(*mru), 0)))
+	mru = kzalloc(sizeof(*mru), GFP_KERNEL | __GFP_NOFAIL);
+	if (!mru)
 		return -ENOMEM;
 
 	/* An extra list is needed to avoid reaping up to a grp_time early. */
 	mru->grp_count = grp_count + 1;
-	mru->lists = kmem_zalloc(mru->grp_count * sizeof(*mru->lists), 0);
-
+	mru->lists = kzalloc(mru->grp_count * sizeof(*mru->lists),
+				GFP_KERNEL | __GFP_NOFAIL);
 	if (!mru->lists) {
 		err = -ENOMEM;
 		goto exit;
@@ -364,9 +365,9 @@ xfs_mru_cache_create(
 
 exit:
 	if (err && mru && mru->lists)
-		kmem_free(mru->lists);
+		kfree(mru->lists);
 	if (err && mru)
-		kmem_free(mru);
+		kfree(mru);
 
 	return err;
 }
@@ -406,8 +407,8 @@ xfs_mru_cache_destroy(
 
 	xfs_mru_cache_flush(mru);
 
-	kmem_free(mru->lists);
-	kmem_free(mru);
+	kfree(mru->lists);
+	kfree(mru);
 }
 
 /*
@@ -427,7 +428,7 @@ xfs_mru_cache_insert(
 	if (!mru || !mru->lists)
 		return -EINVAL;
 
-	if (radix_tree_preload(GFP_NOFS))
+	if (radix_tree_preload(GFP_KERNEL))
 		return -ENOMEM;
 
 	INIT_LIST_HEAD(&elem->list_node);
diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c
index 67d0a8564f..47120b745c 100644
--- a/fs/xfs/xfs_qm.c
+++ b/fs/xfs/xfs_qm.c
@@ -26,6 +26,7 @@
 #include "xfs_ag.h"
 #include "xfs_ialloc.h"
 #include "xfs_log_priv.h"
+#include "xfs_health.h"
 
 /*
  * The global quota manager. There is only one of these for the entire
@@ -254,7 +255,7 @@ xfs_qm_dqattach_one(
 	struct xfs_dquot	*dqp;
 	int			error;
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 	error = 0;
 
 	/*
@@ -322,7 +323,7 @@ xfs_qm_dqattach_locked(
 	if (!xfs_qm_need_dqattach(ip))
 		return 0;
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 
 	if (XFS_IS_UQUOTA_ON(mp) && !ip->i_udquot) {
 		error = xfs_qm_dqattach_one(ip, XFS_DQTYPE_USER,
@@ -353,7 +354,7 @@ done:
 	 * Don't worry about the dquots that we may have attached before any
 	 * error - they'll get detached later if it has not already been done.
 	 */
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 	return error;
 }
 
@@ -628,7 +629,8 @@ xfs_qm_init_quotainfo(
 
 	ASSERT(XFS_IS_QUOTA_ON(mp));
 
-	qinf = mp->m_quotainfo = kmem_zalloc(sizeof(struct xfs_quotainfo), 0);
+	qinf = mp->m_quotainfo = kzalloc(sizeof(struct xfs_quotainfo),
+					GFP_KERNEL | __GFP_NOFAIL);
 
 	error = list_lru_init(&qinf->qi_lru);
 	if (error)
@@ -642,9 +644,9 @@ xfs_qm_init_quotainfo(
 	if (error)
 		goto out_free_lru;
 
-	INIT_RADIX_TREE(&qinf->qi_uquota_tree, GFP_NOFS);
-	INIT_RADIX_TREE(&qinf->qi_gquota_tree, GFP_NOFS);
-	INIT_RADIX_TREE(&qinf->qi_pquota_tree, GFP_NOFS);
+	INIT_RADIX_TREE(&qinf->qi_uquota_tree, GFP_KERNEL);
+	INIT_RADIX_TREE(&qinf->qi_gquota_tree, GFP_KERNEL);
+	INIT_RADIX_TREE(&qinf->qi_pquota_tree, GFP_KERNEL);
 	mutex_init(&qinf->qi_tree_lock);
 
 	/* mutex used to serialize quotaoffs */
@@ -691,6 +693,9 @@ xfs_qm_init_quotainfo(
 
 	shrinker_register(qinf->qi_shrinker);
 
+	xfs_hooks_init(&qinf->qi_mod_ino_dqtrx_hooks);
+	xfs_hooks_init(&qinf->qi_apply_dqtrx_hooks);
+
 	return 0;
 
 out_free_inos:
@@ -700,7 +705,7 @@ out_free_inos:
 out_free_lru:
 	list_lru_destroy(&qinf->qi_lru);
 out_free_qinf:
-	kmem_free(qinf);
+	kfree(qinf);
 	mp->m_quotainfo = NULL;
 	return error;
 }
@@ -724,7 +729,7 @@ xfs_qm_destroy_quotainfo(
 	xfs_qm_destroy_quotainos(qi);
 	mutex_destroy(&qi->qi_tree_lock);
 	mutex_destroy(&qi->qi_quotaofflock);
-	kmem_free(qi);
+	kfree(qi);
 	mp->m_quotainfo = NULL;
 }
 
@@ -758,14 +763,18 @@ xfs_qm_qino_alloc(
 			     (mp->m_sb.sb_gquotino != NULLFSINO)) {
 			ino = mp->m_sb.sb_gquotino;
 			if (XFS_IS_CORRUPT(mp,
-					   mp->m_sb.sb_pquotino != NULLFSINO))
+					   mp->m_sb.sb_pquotino != NULLFSINO)) {
+				xfs_fs_mark_sick(mp, XFS_SICK_FS_PQUOTA);
 				return -EFSCORRUPTED;
+			}
 		} else if ((flags & XFS_QMOPT_GQUOTA) &&
 			     (mp->m_sb.sb_pquotino != NULLFSINO)) {
 			ino = mp->m_sb.sb_pquotino;
 			if (XFS_IS_CORRUPT(mp,
-					   mp->m_sb.sb_gquotino != NULLFSINO))
+					   mp->m_sb.sb_gquotino != NULLFSINO)) {
+				xfs_fs_mark_sick(mp, XFS_SICK_FS_GQUOTA);
 				return -EFSCORRUPTED;
+			}
 		}
 		if (ino != NULLFSINO) {
 			error = xfs_iget(mp, NULL, ino, 0, 0, ipp);
@@ -827,8 +836,10 @@ xfs_qm_qino_alloc(
 		ASSERT(xfs_is_shutdown(mp));
 		xfs_alert(mp, "%s failed (error %d)!", __func__, error);
 	}
-	if (need_alloc)
+	if (need_alloc) {
+		xfs_iunlock(*ipp, XFS_ILOCK_EXCL);
 		xfs_finish_inode_setup(*ipp);
+	}
 	return error;
 }
 
@@ -996,7 +1007,8 @@ xfs_qm_reset_dqcounts_buf(
 	if (qip->i_nblocks == 0)
 		return 0;
 
-	map = kmem_alloc(XFS_DQITER_MAP_SIZE * sizeof(*map), 0);
+	map = kmalloc(XFS_DQITER_MAP_SIZE * sizeof(*map),
+			GFP_KERNEL | __GFP_NOFAIL);
 
 	lblkno = 0;
 	maxlblkcnt = XFS_B_TO_FSB(mp, mp->m_super->s_maxbytes);
@@ -1058,7 +1070,7 @@ xfs_qm_reset_dqcounts_buf(
 	} while (nmaps > 0);
 
 out:
-	kmem_free(map);
+	kfree(map);
 	return error;
 }
 
@@ -1406,8 +1418,12 @@ error_return:
 			xfs_warn(mp,
 				"Quotacheck: Failed to reset quota flags.");
 		}
-	} else
+		xfs_fs_mark_sick(mp, XFS_SICK_FS_QUOTACHECK);
+	} else {
 		xfs_notice(mp, "Quotacheck: Done.");
+		xfs_fs_mark_healthy(mp, XFS_SICK_FS_QUOTACHECK);
+	}
+
 	return error;
 
 error_purge:
@@ -1809,7 +1825,7 @@ xfs_qm_vop_chown(
 				 XFS_TRANS_DQ_RTBCOUNT : XFS_TRANS_DQ_BCOUNT;
 
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 	ASSERT(XFS_IS_QUOTA_ON(ip->i_mount));
 
 	/* old dquot */
@@ -1817,12 +1833,12 @@ xfs_qm_vop_chown(
 	ASSERT(prevdq);
 	ASSERT(prevdq != newdq);
 
-	xfs_trans_mod_dquot(tp, prevdq, bfield, -(ip->i_nblocks));
-	xfs_trans_mod_dquot(tp, prevdq, XFS_TRANS_DQ_ICOUNT, -1);
+	xfs_trans_mod_ino_dquot(tp, ip, prevdq, bfield, -(ip->i_nblocks));
+	xfs_trans_mod_ino_dquot(tp, ip, prevdq, XFS_TRANS_DQ_ICOUNT, -1);
 
 	/* the sparkling new dquot */
-	xfs_trans_mod_dquot(tp, newdq, bfield, ip->i_nblocks);
-	xfs_trans_mod_dquot(tp, newdq, XFS_TRANS_DQ_ICOUNT, 1);
+	xfs_trans_mod_ino_dquot(tp, ip, newdq, bfield, ip->i_nblocks);
+	xfs_trans_mod_ino_dquot(tp, ip, newdq, XFS_TRANS_DQ_ICOUNT, 1);
 
 	/*
 	 * Back when we made quota reservations for the chown, we reserved the
@@ -1897,29 +1913,28 @@ xfs_qm_vop_create_dqattach(
 	if (!XFS_IS_QUOTA_ON(mp))
 		return;
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 
 	if (udqp && XFS_IS_UQUOTA_ON(mp)) {
 		ASSERT(ip->i_udquot == NULL);
 		ASSERT(i_uid_read(VFS_I(ip)) == udqp->q_id);
 
 		ip->i_udquot = xfs_qm_dqhold(udqp);
-		xfs_trans_mod_dquot(tp, udqp, XFS_TRANS_DQ_ICOUNT, 1);
 	}
 	if (gdqp && XFS_IS_GQUOTA_ON(mp)) {
 		ASSERT(ip->i_gdquot == NULL);
 		ASSERT(i_gid_read(VFS_I(ip)) == gdqp->q_id);
 
 		ip->i_gdquot = xfs_qm_dqhold(gdqp);
-		xfs_trans_mod_dquot(tp, gdqp, XFS_TRANS_DQ_ICOUNT, 1);
 	}
 	if (pdqp && XFS_IS_PQUOTA_ON(mp)) {
 		ASSERT(ip->i_pdquot == NULL);
 		ASSERT(ip->i_projid == pdqp->q_id);
 
 		ip->i_pdquot = xfs_qm_dqhold(pdqp);
-		xfs_trans_mod_dquot(tp, pdqp, XFS_TRANS_DQ_ICOUNT, 1);
 	}
+
+	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_ICOUNT, 1);
 }
 
 /* Decide if this inode's dquot is near an enforcement boundary. */
diff --git a/fs/xfs/xfs_qm.h b/fs/xfs/xfs_qm.h
index d5c9fc4ba5..6e09dfcd13 100644
--- a/fs/xfs/xfs_qm.h
+++ b/fs/xfs/xfs_qm.h
@@ -68,6 +68,10 @@ struct xfs_quotainfo {
 	/* Minimum and maximum quota expiration timestamp values. */
 	time64_t		qi_expiry_min;
 	time64_t		qi_expiry_max;
+
+	/* Hook to feed quota counter updates to an active online repair. */
+	struct xfs_hooks	qi_mod_ino_dqtrx_hooks;
+	struct xfs_hooks	qi_apply_dqtrx_hooks;
 };
 
 static inline struct radix_tree_root *
@@ -104,6 +108,18 @@ xfs_quota_inode(struct xfs_mount *mp, xfs_dqtype_t type)
 	return NULL;
 }
 
+/*
+ * Parameters for tracking dqtrx changes on behalf of an inode.  The hook
+ * function arg parameter is the field being updated.
+ */
+struct xfs_mod_ino_dqtrx_params {
+	uintptr_t		tx_id;
+	xfs_ino_t		ino;
+	xfs_dqtype_t		q_type;
+	xfs_dqid_t		q_id;
+	int64_t			delta;
+};
+
 extern void	xfs_trans_mod_dquot(struct xfs_trans *tp, struct xfs_dquot *dqp,
 				    uint field, int64_t delta);
 extern void	xfs_trans_dqjoin(struct xfs_trans *, struct xfs_dquot *);
@@ -120,7 +136,7 @@ enum {
 	XFS_QM_TRANS_PRJ,
 	XFS_QM_TRANS_DQTYPES
 };
-#define XFS_QM_TRANS_MAXDQS		2
+#define XFS_QM_TRANS_MAXDQS		5
 struct xfs_dquot_acct {
 	struct xfs_dqtrx	dqs[XFS_QM_TRANS_DQTYPES][XFS_QM_TRANS_MAXDQS];
 };
diff --git a/fs/xfs/xfs_qm_bhv.c b/fs/xfs/xfs_qm_bhv.c
index b77673dd05..271c1021c7 100644
--- a/fs/xfs/xfs_qm_bhv.c
+++ b/fs/xfs/xfs_qm_bhv.c
@@ -9,6 +9,7 @@
 #include "xfs_format.h"
 #include "xfs_log_format.h"
 #include "xfs_trans_resv.h"
+#include "xfs_mount.h"
 #include "xfs_quota.h"
 #include "xfs_mount.h"
 #include "xfs_inode.h"
diff --git a/fs/xfs/xfs_quota.h b/fs/xfs/xfs_quota.h
index e0d56489f3..23d71a55bb 100644
--- a/fs/xfs/xfs_quota.h
+++ b/fs/xfs/xfs_quota.h
@@ -74,6 +74,22 @@ struct xfs_dqtrx {
 	int64_t		qt_icount_delta;  /* dquot inode count changes */
 };
 
+enum xfs_apply_dqtrx_type {
+	XFS_APPLY_DQTRX_COMMIT = 0,
+	XFS_APPLY_DQTRX_UNRESERVE,
+};
+
+/*
+ * Parameters for applying dqtrx changes to a dquot.  The hook function arg
+ * parameter is enum xfs_apply_dqtrx_type.
+ */
+struct xfs_apply_dqtrx_params {
+	uintptr_t		tx_id;
+	xfs_ino_t		ino;
+	xfs_dqtype_t		q_type;
+	xfs_dqid_t		q_id;
+};
+
 #ifdef CONFIG_XFS_QUOTA
 extern void xfs_trans_dup_dqinfo(struct xfs_trans *, struct xfs_trans *);
 extern void xfs_trans_free_dqinfo(struct xfs_trans *);
@@ -107,13 +123,31 @@ extern int xfs_qm_newmount(struct xfs_mount *, uint *, uint *);
 extern void xfs_qm_mount_quotas(struct xfs_mount *);
 extern void xfs_qm_unmount(struct xfs_mount *);
 extern void xfs_qm_unmount_quotas(struct xfs_mount *);
-
-static inline int
-xfs_quota_reserve_blkres(struct xfs_inode *ip, int64_t blocks)
-{
-	return xfs_trans_reserve_quota_nblks(NULL, ip, blocks, 0, false);
-}
 bool xfs_inode_near_dquot_enforcement(struct xfs_inode *ip, xfs_dqtype_t type);
+
+# ifdef CONFIG_XFS_LIVE_HOOKS
+void xfs_trans_mod_ino_dquot(struct xfs_trans *tp, struct xfs_inode *ip,
+		struct xfs_dquot *dqp, unsigned int field, int64_t delta);
+
+struct xfs_quotainfo;
+
+struct xfs_dqtrx_hook {
+	struct xfs_hook		mod_hook;
+	struct xfs_hook		apply_hook;
+};
+
+void xfs_dqtrx_hook_disable(void);
+void xfs_dqtrx_hook_enable(void);
+
+int xfs_dqtrx_hook_add(struct xfs_quotainfo *qi, struct xfs_dqtrx_hook *hook);
+void xfs_dqtrx_hook_del(struct xfs_quotainfo *qi, struct xfs_dqtrx_hook *hook);
+void xfs_dqtrx_hook_setup(struct xfs_dqtrx_hook *hook, notifier_fn_t mod_fn,
+		notifier_fn_t apply_fn);
+# else
+#  define xfs_trans_mod_ino_dquot(tp, ip, dqp, field, delta) \
+		xfs_trans_mod_dquot((tp), (dqp), (field), (delta))
+# endif /* CONFIG_XFS_LIVE_HOOKS */
+
 #else
 static inline int
 xfs_qm_vop_dqalloc(struct xfs_inode *ip, kuid_t kuid, kgid_t kgid,
@@ -148,12 +182,6 @@ static inline int xfs_trans_reserve_quota_bydquots(struct xfs_trans *tp,
 }
 
 static inline int
-xfs_quota_reserve_blkres(struct xfs_inode *ip, int64_t blocks)
-{
-	return 0;
-}
-
-static inline int
 xfs_trans_reserve_quota_icreate(struct xfs_trans *tp, struct xfs_dquot *udqp,
 		struct xfs_dquot *gdqp, struct xfs_dquot *pdqp, int64_t dblocks)
 {
@@ -173,12 +201,25 @@ xfs_trans_reserve_quota_icreate(struct xfs_trans *tp, struct xfs_dquot *udqp,
 #define xfs_qm_unmount(mp)
 #define xfs_qm_unmount_quotas(mp)
 #define xfs_inode_near_dquot_enforcement(ip, type)			(false)
+
+# ifdef CONFIG_XFS_LIVE_HOOKS
+#  define xfs_dqtrx_hook_enable()		((void)0)
+#  define xfs_dqtrx_hook_disable()		((void)0)
+# endif /* CONFIG_XFS_LIVE_HOOKS */
+
 #endif /* CONFIG_XFS_QUOTA */
 
 static inline int
-xfs_quota_unreserve_blkres(struct xfs_inode *ip, int64_t blocks)
+xfs_quota_reserve_blkres(struct xfs_inode *ip, int64_t blocks)
+{
+	return xfs_trans_reserve_quota_nblks(NULL, ip, blocks, 0, false);
+}
+
+static inline void
+xfs_quota_unreserve_blkres(struct xfs_inode *ip, uint64_t blocks)
 {
-	return xfs_quota_reserve_blkres(ip, -blocks);
+	/* don't return an error as unreserving quotas can't fail */
+	xfs_quota_reserve_blkres(ip, -(int64_t)blocks);
 }
 
 extern int xfs_mount_reset_sbqflags(struct xfs_mount *);
diff --git a/fs/xfs/xfs_refcount_item.c b/fs/xfs/xfs_refcount_item.c
index 20ad8086da..14919b33e4 100644
--- a/fs/xfs/xfs_refcount_item.c
+++ b/fs/xfs/xfs_refcount_item.c
@@ -36,9 +36,9 @@ STATIC void
 xfs_cui_item_free(
 	struct xfs_cui_log_item	*cuip)
 {
-	kmem_free(cuip->cui_item.li_lv_shadow);
+	kvfree(cuip->cui_item.li_lv_shadow);
 	if (cuip->cui_format.cui_nextents > XFS_CUI_MAX_FAST_EXTENTS)
-		kmem_free(cuip);
+		kfree(cuip);
 	else
 		kmem_cache_free(xfs_cui_cache, cuip);
 }
@@ -143,8 +143,8 @@ xfs_cui_init(
 
 	ASSERT(nextents > 0);
 	if (nextents > XFS_CUI_MAX_FAST_EXTENTS)
-		cuip = kmem_zalloc(xfs_cui_log_item_sizeof(nextents),
-				0);
+		cuip = kzalloc(xfs_cui_log_item_sizeof(nextents),
+				GFP_KERNEL | __GFP_NOFAIL);
 	else
 		cuip = kmem_cache_zalloc(xfs_cui_cache,
 					 GFP_KERNEL | __GFP_NOFAIL);
@@ -207,7 +207,7 @@ xfs_cud_item_release(
 	struct xfs_cud_log_item	*cudp = CUD_ITEM(lip);
 
 	xfs_cui_release(cudp->cud_cuip);
-	kmem_free(cudp->cud_item.li_lv_shadow);
+	kvfree(cudp->cud_item.li_lv_shadow);
 	kmem_cache_free(xfs_cud_cache, cudp);
 }
 
@@ -425,7 +425,7 @@ xfs_cui_recover_work(
 	struct xfs_refcount_intent	*ri;
 
 	ri = kmem_cache_alloc(xfs_refcount_intent_cache,
-			GFP_NOFS | __GFP_NOFAIL);
+			GFP_KERNEL | __GFP_NOFAIL);
 	ri->ri_type = pmap->pe_flags & XFS_REFCOUNT_EXTENT_TYPE_MASK;
 	ri->ri_startblock = pmap->pe_startblock;
 	ri->ri_blockcount = pmap->pe_len;
diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c
index d5ca8bcae6..265a2a418b 100644
--- a/fs/xfs/xfs_reflink.c
+++ b/fs/xfs/xfs_reflink.c
@@ -29,6 +29,7 @@
 #include "xfs_iomap.h"
 #include "xfs_ag.h"
 #include "xfs_ag_resv.h"
+#include "xfs_health.h"
 
 /*
  * Copy on Write of Shared Blocks
@@ -429,13 +430,6 @@ xfs_reflink_fill_cow_hole(
 	if (error)
 		return error;
 
-	/*
-	 * Allocation succeeded but the requested range was not even partially
-	 * satisfied?  Bail out!
-	 */
-	if (nimaps == 0)
-		return -ENOSPC;
-
 convert:
 	return xfs_reflink_convert_unwritten(ip, imap, cmap, convert_now);
 
@@ -498,13 +492,6 @@ xfs_reflink_fill_delalloc(
 		error = xfs_trans_commit(tp);
 		if (error)
 			return error;
-
-		/*
-		 * Allocation succeeded but the requested range was not even
-		 * partially satisfied?  Bail out!
-		 */
-		if (nimaps == 0)
-			return -ENOSPC;
 	} while (cmap->br_startoff + cmap->br_blockcount <= imap->br_startoff);
 
 	return xfs_reflink_convert_unwritten(ip, imap, cmap, convert_now);
@@ -527,7 +514,7 @@ xfs_reflink_allocate_cow(
 	int			error;
 	bool			found;
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 	if (!ip->i_cowfp) {
 		ASSERT(!xfs_is_reflink_inode(ip));
 		xfs_ifork_init_cow(ip);
@@ -605,10 +592,8 @@ xfs_reflink_cancel_cow_blocks(
 		trace_xfs_reflink_cancel_cow(ip, &del);
 
 		if (isnullstartblock(del.br_startblock)) {
-			error = xfs_bmap_del_extent_delay(ip, XFS_COW_FORK,
-					&icur, &got, &del);
-			if (error)
-				break;
+			xfs_bmap_del_extent_delay(ip, XFS_COW_FORK, &icur, &got,
+					&del);
 		} else if (del.br_state == XFS_EXT_UNWRITTEN || cancel_real) {
 			ASSERT((*tpp)->t_highest_agno == NULLAGNUMBER);
 
@@ -631,10 +616,7 @@ xfs_reflink_cancel_cow_blocks(
 			xfs_bmap_del_extent_cow(ip, &icur, &got, &del);
 
 			/* Remove the quota reservation */
-			error = xfs_quota_unreserve_blkres(ip,
-					del.br_blockcount);
-			if (error)
-				break;
+			xfs_quota_unreserve_blkres(ip, del.br_blockcount);
 		} else {
 			/* Didn't do anything, push cursor back. */
 			xfs_iext_prev(ifp, &icur);
@@ -730,12 +712,6 @@ xfs_reflink_end_cow_extent(
 	int			nmaps;
 	int			error;
 
-	/* No COW extents?  That's easy! */
-	if (ifp->if_bytes == 0) {
-		*offset_fsb = end_fsb;
-		return 0;
-	}
-
 	resblks = XFS_EXTENTADD_SPACE_RES(mp, XFS_DATA_FORK);
 	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, resblks, 0,
 			XFS_TRANS_RESERVE, &tp);
@@ -750,14 +726,6 @@ xfs_reflink_end_cow_extent(
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
 	xfs_trans_ijoin(tp, ip, 0);
 
-	error = xfs_iext_count_may_overflow(ip, XFS_DATA_FORK,
-			XFS_IEXT_REFLINK_END_COW_CNT);
-	if (error == -EFBIG)
-		error = xfs_iext_count_upgrade(tp, ip,
-				XFS_IEXT_REFLINK_END_COW_CNT);
-	if (error)
-		goto out_cancel;
-
 	/*
 	 * In case of racing, overlapping AIO writes no COW extents might be
 	 * left by the time I/O completes for the loser of the race.  In that
@@ -786,6 +754,11 @@ xfs_reflink_end_cow_extent(
 	del = got;
 	xfs_trim_extent(&del, *offset_fsb, end_fsb - *offset_fsb);
 
+	error = xfs_iext_count_extend(tp, ip, XFS_DATA_FORK,
+			XFS_IEXT_REFLINK_END_COW_CNT);
+	if (error)
+		goto out_cancel;
+
 	/* Grab the corresponding mapping in the data fork. */
 	nmaps = 1;
 	error = xfs_bmapi_read(ip, del.br_startoff, del.br_blockcount, &data,
@@ -805,7 +778,7 @@ xfs_reflink_end_cow_extent(
 		 * If the extent we're remapping is backed by storage (written
 		 * or not), unmap the extent and drop its refcount.
 		 */
-		xfs_bmap_unmap_extent(tp, ip, &data);
+		xfs_bmap_unmap_extent(tp, ip, XFS_DATA_FORK, &data);
 		xfs_refcount_decrease_extent(tp, &data);
 		xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT,
 				-data.br_blockcount);
@@ -829,7 +802,7 @@ xfs_reflink_end_cow_extent(
 	xfs_refcount_free_cow_extent(tp, del.br_startblock, del.br_blockcount);
 
 	/* Map the new blocks into the data fork. */
-	xfs_bmap_map_extent(tp, ip, &del);
+	xfs_bmap_map_extent(tp, ip, XFS_DATA_FORK, &del);
 
 	/* Charge this new data fork mapping to the on-disk quota. */
 	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_DELBCOUNT,
@@ -1227,8 +1200,10 @@ xfs_reflink_remap_extent(
 	 * extent if they're both holes or both the same physical extent.
 	 */
 	if (dmap->br_startblock == smap.br_startblock) {
-		if (dmap->br_state != smap.br_state)
+		if (dmap->br_state != smap.br_state) {
+			xfs_bmap_mark_sick(ip, XFS_DATA_FORK);
 			error = -EFSCORRUPTED;
+		}
 		goto out_cancel;
 	}
 
@@ -1280,9 +1255,7 @@ xfs_reflink_remap_extent(
 	if (dmap_written)
 		++iext_delta;
 
-	error = xfs_iext_count_may_overflow(ip, XFS_DATA_FORK, iext_delta);
-	if (error == -EFBIG)
-		error = xfs_iext_count_upgrade(tp, ip, iext_delta);
+	error = xfs_iext_count_extend(tp, ip, XFS_DATA_FORK, iext_delta);
 	if (error)
 		goto out_cancel;
 
@@ -1291,7 +1264,7 @@ xfs_reflink_remap_extent(
 		 * If the extent we're unmapping is backed by storage (written
 		 * or not), unmap the extent and drop its refcount.
 		 */
-		xfs_bmap_unmap_extent(tp, ip, &smap);
+		xfs_bmap_unmap_extent(tp, ip, XFS_DATA_FORK, &smap);
 		xfs_refcount_decrease_extent(tp, &smap);
 		qdelta -= smap.br_blockcount;
 	} else if (smap.br_startblock == DELAYSTARTBLOCK) {
@@ -1316,7 +1289,7 @@ xfs_reflink_remap_extent(
 	 */
 	if (dmap_written) {
 		xfs_refcount_increase_extent(tp, dmap);
-		xfs_bmap_map_extent(tp, ip, dmap);
+		xfs_bmap_map_extent(tp, ip, XFS_DATA_FORK, dmap);
 		qdelta += dmap->br_blockcount;
 	}
 
@@ -1391,6 +1364,7 @@ xfs_reflink_remap_blocks(
 		ASSERT(nimaps == 1 && imap.br_startoff == srcoff);
 		if (imap.br_startblock == DELAYSTARTBLOCK) {
 			ASSERT(imap.br_startblock != DELAYSTARTBLOCK);
+			xfs_bmap_mark_sick(src, XFS_DATA_FORK);
 			error = -EFSCORRUPTED;
 			break;
 		}
@@ -1413,6 +1387,7 @@ xfs_reflink_remap_blocks(
 		destoff += imap.br_blockcount;
 		len -= imap.br_blockcount;
 		remapped_len += imap.br_blockcount;
+		cond_resched();
 	}
 
 	if (error)
diff --git a/fs/xfs/xfs_rmap_item.c b/fs/xfs/xfs_rmap_item.c
index 79ad0087ae..e473124e29 100644
--- a/fs/xfs/xfs_rmap_item.c
+++ b/fs/xfs/xfs_rmap_item.c
@@ -36,9 +36,9 @@ STATIC void
 xfs_rui_item_free(
 	struct xfs_rui_log_item	*ruip)
 {
-	kmem_free(ruip->rui_item.li_lv_shadow);
+	kvfree(ruip->rui_item.li_lv_shadow);
 	if (ruip->rui_format.rui_nextents > XFS_RUI_MAX_FAST_EXTENTS)
-		kmem_free(ruip);
+		kfree(ruip);
 	else
 		kmem_cache_free(xfs_rui_cache, ruip);
 }
@@ -142,7 +142,8 @@ xfs_rui_init(
 
 	ASSERT(nextents > 0);
 	if (nextents > XFS_RUI_MAX_FAST_EXTENTS)
-		ruip = kmem_zalloc(xfs_rui_log_item_sizeof(nextents), 0);
+		ruip = kzalloc(xfs_rui_log_item_sizeof(nextents),
+				GFP_KERNEL | __GFP_NOFAIL);
 	else
 		ruip = kmem_cache_zalloc(xfs_rui_cache,
 					 GFP_KERNEL | __GFP_NOFAIL);
@@ -205,7 +206,7 @@ xfs_rud_item_release(
 	struct xfs_rud_log_item	*rudp = RUD_ITEM(lip);
 
 	xfs_rui_release(rudp->rud_ruip);
-	kmem_free(rudp->rud_item.li_lv_shadow);
+	kvfree(rudp->rud_item.li_lv_shadow);
 	kmem_cache_free(xfs_rud_cache, rudp);
 }
 
@@ -454,7 +455,7 @@ xfs_rui_recover_work(
 {
 	struct xfs_rmap_intent		*ri;
 
-	ri = kmem_cache_alloc(xfs_rmap_intent_cache, GFP_NOFS | __GFP_NOFAIL);
+	ri = kmem_cache_alloc(xfs_rmap_intent_cache, GFP_KERNEL | __GFP_NOFAIL);
 
 	switch (map->me_flags & XFS_RMAP_EXTENT_TYPE_MASK) {
 	case XFS_RMAP_EXTENT_MAP:
diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c
index 8649d981a0..5a7ddfed1b 100644
--- a/fs/xfs/xfs_rtalloc.c
+++ b/fs/xfs/xfs_rtalloc.c
@@ -22,6 +22,8 @@
 #include "xfs_sb.h"
 #include "xfs_rtbitmap.h"
 #include "xfs_quota.h"
+#include "xfs_log_priv.h"
+#include "xfs_health.h"
 
 /*
  * Return whether there are any free extents in the size range given
@@ -693,11 +695,8 @@ xfs_growfs_rt_alloc(
 		xfs_ilock(ip, XFS_ILOCK_EXCL);
 		xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
 
-		error = xfs_iext_count_may_overflow(ip, XFS_DATA_FORK,
+		error = xfs_iext_count_extend(tp, ip, XFS_DATA_FORK,
 				XFS_IEXT_ADD_NOSPLIT_CNT);
-		if (error == -EFBIG)
-			error = xfs_iext_count_upgrade(tp, ip,
-					XFS_IEXT_ADD_NOSPLIT_CNT);
 		if (error)
 			goto out_trans_cancel;
 
@@ -707,8 +706,6 @@ xfs_growfs_rt_alloc(
 		nmap = 1;
 		error = xfs_bmapi_write(tp, ip, oblocks, nblocks - oblocks,
 					XFS_BMAPI_METADATA, 0, &map, &nmap);
-		if (!error && nmap < 1)
-			error = -ENOSPC;
 		if (error)
 			goto out_trans_cancel;
 		/*
@@ -903,7 +900,7 @@ xfs_growfs_rt(
 	/*
 	 * Allocate a new (fake) mount/sb.
 	 */
-	nmp = kmem_alloc(sizeof(*nmp), 0);
+	nmp = kmalloc(sizeof(*nmp), GFP_KERNEL | __GFP_NOFAIL);
 	/*
 	 * Loop over the bitmap blocks.
 	 * We will do everything one bitmap block at a time.
@@ -955,10 +952,10 @@ xfs_growfs_rt(
 		nargs.tp = tp;
 
 		/*
-		 * Lock out other callers by grabbing the bitmap inode lock.
+		 * Lock out other callers by grabbing the bitmap and summary
+		 * inode locks and joining them to the transaction.
 		 */
-		xfs_ilock(mp->m_rbmip, XFS_ILOCK_EXCL | XFS_ILOCK_RTBITMAP);
-		xfs_trans_ijoin(tp, mp->m_rbmip, XFS_ILOCK_EXCL);
+		xfs_rtbitmap_lock(tp, mp);
 		/*
 		 * Update the bitmap inode's size ondisk and incore.  We need
 		 * to update the incore size so that inode inactivation won't
@@ -969,11 +966,6 @@ xfs_growfs_rt(
 		i_size_write(VFS_I(mp->m_rbmip), mp->m_rbmip->i_disk_size);
 		xfs_trans_log_inode(tp, mp->m_rbmip, XFS_ILOG_CORE);
 		/*
-		 * Get the summary inode into the transaction.
-		 */
-		xfs_ilock(mp->m_rsumip, XFS_ILOCK_EXCL | XFS_ILOCK_RTSUM);
-		xfs_trans_ijoin(tp, mp->m_rsumip, XFS_ILOCK_EXCL);
-		/*
 		 * Update the summary inode's size.  We need to update the
 		 * incore size so that inode inactivation won't punch what it
 		 * thinks are "posteof" blocks.
@@ -1050,7 +1042,7 @@ out_free:
 	/*
 	 * Free the fake mp structure.
 	 */
-	kmem_free(nmp);
+	kfree(nmp);
 
 	/*
 	 * If we had to allocate a new rsum_cache, we either need to free the
@@ -1059,10 +1051,10 @@ out_free:
 	 */
 	if (rsum_cache != mp->m_rsum_cache) {
 		if (error) {
-			kmem_free(mp->m_rsum_cache);
+			kvfree(mp->m_rsum_cache);
 			mp->m_rsum_cache = rsum_cache;
 		} else {
-			kmem_free(rsum_cache);
+			kvfree(rsum_cache);
 		}
 	}
 
@@ -1140,10 +1132,10 @@ xfs_rtalloc_reinit_frextents(
 	uint64_t		val = 0;
 	int			error;
 
-	xfs_ilock(mp->m_rbmip, XFS_ILOCK_SHARED | XFS_ILOCK_RTBITMAP);
+	xfs_rtbitmap_lock_shared(mp, XFS_RBMLOCK_BITMAP);
 	error = xfs_rtalloc_query_all(mp, NULL, xfs_rtalloc_count_frextent,
 			&val);
-	xfs_iunlock(mp->m_rbmip, XFS_ILOCK_SHARED | XFS_ILOCK_RTBITMAP);
+	xfs_rtbitmap_unlock_shared(mp, XFS_RBMLOCK_BITMAP);
 	if (error)
 		return error;
 
@@ -1202,6 +1194,8 @@ xfs_rtmount_inodes(
 
 	sbp = &mp->m_sb;
 	error = xfs_iget(mp, NULL, sbp->sb_rbmino, 0, 0, &mp->m_rbmip);
+	if (xfs_metadata_is_sick(error))
+		xfs_rt_mark_sick(mp, XFS_SICK_RT_BITMAP);
 	if (error)
 		return error;
 	ASSERT(mp->m_rbmip != NULL);
@@ -1211,6 +1205,8 @@ xfs_rtmount_inodes(
 		goto out_rele_bitmap;
 
 	error = xfs_iget(mp, NULL, sbp->sb_rsumino, 0, 0, &mp->m_rsumip);
+	if (xfs_metadata_is_sick(error))
+		xfs_rt_mark_sick(mp, XFS_SICK_RT_SUMMARY);
 	if (error)
 		goto out_rele_bitmap;
 	ASSERT(mp->m_rsumip != NULL);
@@ -1233,7 +1229,7 @@ void
 xfs_rtunmount_inodes(
 	struct xfs_mount	*mp)
 {
-	kmem_free(mp->m_rsum_cache);
+	kvfree(mp->m_rsum_cache);
 	if (mp->m_rbmip)
 		xfs_irele(mp->m_rbmip);
 	if (mp->m_rsumip)
@@ -1260,7 +1256,7 @@ xfs_rtpick_extent(
 	uint64_t		seq;		/* sequence number of file creation */
 	struct timespec64	ts;		/* timespec in inode */
 
-	ASSERT(xfs_isilocked(mp->m_rbmip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(mp->m_rbmip, XFS_ILOCK_EXCL);
 
 	ts = inode_get_atime(VFS_I(mp->m_rbmip));
 	if (!(mp->m_rbmip->i_diflags & XFS_DIFLAG_NEWRTBM)) {
@@ -1340,6 +1336,8 @@ xfs_bmap_rtalloc(
 	int			error;
 
 	align = xfs_get_extsz_hint(ap->ip);
+	if (!align)
+		align = 1;
 retry:
 	error = xfs_bmap_extsize_align(mp, &ap->got, &ap->prev,
 					align, 1, ap->eof, 0,
@@ -1376,10 +1374,7 @@ retry:
 	 * Lock out modifications to both the RT bitmap and summary inodes
 	 */
 	if (!rtlocked) {
-		xfs_ilock(mp->m_rbmip, XFS_ILOCK_EXCL|XFS_ILOCK_RTBITMAP);
-		xfs_trans_ijoin(ap->tp, mp->m_rbmip, XFS_ILOCK_EXCL);
-		xfs_ilock(mp->m_rsumip, XFS_ILOCK_EXCL|XFS_ILOCK_RTSUM);
-		xfs_trans_ijoin(ap->tp, mp->m_rsumip, XFS_ILOCK_EXCL);
+		xfs_rtbitmap_lock(ap->tp, mp);
 		rtlocked = true;
 	}
 
diff --git a/fs/xfs/xfs_stats.c b/fs/xfs/xfs_stats.c
index 90a77cd3eb..ed97d72caa 100644
--- a/fs/xfs/xfs_stats.c
+++ b/fs/xfs/xfs_stats.c
@@ -50,7 +50,9 @@ int xfs_stats_format(struct xfsstats __percpu *stats, char *buf)
 		{ "ibt2",		xfsstats_offset(xs_fibt_2)	},
 		{ "fibt2",		xfsstats_offset(xs_rmap_2)	},
 		{ "rmapbt",		xfsstats_offset(xs_refcbt_2)	},
-		{ "refcntbt",		xfsstats_offset(xs_qm_dqreclaims)},
+		{ "refcntbt",		xfsstats_offset(xs_rmap_mem_2)	},
+		{ "rmapbt_mem",		xfsstats_offset(xs_rcbag_2)	},
+		{ "rcbagbt",		xfsstats_offset(xs_qm_dqreclaims)},
 		/* we print both series of quota information together */
 		{ "qm",			xfsstats_offset(xs_xstrat_bytes)},
 	};
diff --git a/fs/xfs/xfs_stats.h b/fs/xfs/xfs_stats.h
index 43ffba74f0..a61fb56ed2 100644
--- a/fs/xfs/xfs_stats.h
+++ b/fs/xfs/xfs_stats.h
@@ -125,6 +125,8 @@ struct __xfsstats {
 	uint32_t		xs_fibt_2[__XBTS_MAX];
 	uint32_t		xs_rmap_2[__XBTS_MAX];
 	uint32_t		xs_refcbt_2[__XBTS_MAX];
+	uint32_t		xs_rmap_mem_2[__XBTS_MAX];
+	uint32_t		xs_rcbag_2[__XBTS_MAX];
 	uint32_t		xs_qm_dqreclaims;
 	uint32_t		xs_qm_dqreclaim_misses;
 	uint32_t		xs_qm_dquot_dups;
diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c
index 98401de832..27e9f749c4 100644
--- a/fs/xfs/xfs_super.c
+++ b/fs/xfs/xfs_super.c
@@ -43,7 +43,10 @@
 #include "xfs_iunlink_item.h"
 #include "xfs_dahash_test.h"
 #include "xfs_rtbitmap.h"
+#include "xfs_exchmaps_item.h"
+#include "xfs_parent.h"
 #include "scrub/stats.h"
+#include "scrub/rcbag_btree.h"
 
 #include <linux/magic.h>
 #include <linux/fs_context.h>
@@ -361,16 +364,16 @@ STATIC int
 xfs_blkdev_get(
 	xfs_mount_t		*mp,
 	const char		*name,
-	struct bdev_handle	**handlep)
+	struct file		**bdev_filep)
 {
 	int			error = 0;
 
-	*handlep = bdev_open_by_path(name,
+	*bdev_filep = bdev_file_open_by_path(name,
 		BLK_OPEN_READ | BLK_OPEN_WRITE | BLK_OPEN_RESTRICT_WRITES,
 		mp->m_super, &fs_holder_ops);
-	if (IS_ERR(*handlep)) {
-		error = PTR_ERR(*handlep);
-		*handlep = NULL;
+	if (IS_ERR(*bdev_filep)) {
+		error = PTR_ERR(*bdev_filep);
+		*bdev_filep = NULL;
 		xfs_warn(mp, "Invalid device [%s], error=%d", name, error);
 	}
 
@@ -435,26 +438,26 @@ xfs_open_devices(
 {
 	struct super_block	*sb = mp->m_super;
 	struct block_device	*ddev = sb->s_bdev;
-	struct bdev_handle	*logdev_handle = NULL, *rtdev_handle = NULL;
+	struct file		*logdev_file = NULL, *rtdev_file = NULL;
 	int			error;
 
 	/*
 	 * Open real time and log devices - order is important.
 	 */
 	if (mp->m_logname) {
-		error = xfs_blkdev_get(mp, mp->m_logname, &logdev_handle);
+		error = xfs_blkdev_get(mp, mp->m_logname, &logdev_file);
 		if (error)
 			return error;
 	}
 
 	if (mp->m_rtname) {
-		error = xfs_blkdev_get(mp, mp->m_rtname, &rtdev_handle);
+		error = xfs_blkdev_get(mp, mp->m_rtname, &rtdev_file);
 		if (error)
 			goto out_close_logdev;
 
-		if (rtdev_handle->bdev == ddev ||
-		    (logdev_handle &&
-		     rtdev_handle->bdev == logdev_handle->bdev)) {
+		if (file_bdev(rtdev_file) == ddev ||
+		    (logdev_file &&
+		     file_bdev(rtdev_file) == file_bdev(logdev_file))) {
 			xfs_warn(mp,
 	"Cannot mount filesystem with identical rtdev and ddev/logdev.");
 			error = -EINVAL;
@@ -466,25 +469,25 @@ xfs_open_devices(
 	 * Setup xfs_mount buffer target pointers
 	 */
 	error = -ENOMEM;
-	mp->m_ddev_targp = xfs_alloc_buftarg(mp, sb->s_bdev_handle);
+	mp->m_ddev_targp = xfs_alloc_buftarg(mp, sb->s_bdev_file);
 	if (!mp->m_ddev_targp)
 		goto out_close_rtdev;
 
-	if (rtdev_handle) {
-		mp->m_rtdev_targp = xfs_alloc_buftarg(mp, rtdev_handle);
+	if (rtdev_file) {
+		mp->m_rtdev_targp = xfs_alloc_buftarg(mp, rtdev_file);
 		if (!mp->m_rtdev_targp)
 			goto out_free_ddev_targ;
 	}
 
-	if (logdev_handle && logdev_handle->bdev != ddev) {
-		mp->m_logdev_targp = xfs_alloc_buftarg(mp, logdev_handle);
+	if (logdev_file && file_bdev(logdev_file) != ddev) {
+		mp->m_logdev_targp = xfs_alloc_buftarg(mp, logdev_file);
 		if (!mp->m_logdev_targp)
 			goto out_free_rtdev_targ;
 	} else {
 		mp->m_logdev_targp = mp->m_ddev_targp;
 		/* Handle won't be used, drop it */
-		if (logdev_handle)
-			bdev_release(logdev_handle);
+		if (logdev_file)
+			bdev_fput(logdev_file);
 	}
 
 	return 0;
@@ -495,11 +498,11 @@ xfs_open_devices(
  out_free_ddev_targ:
 	xfs_free_buftarg(mp->m_ddev_targp);
  out_close_rtdev:
-	 if (rtdev_handle)
-		bdev_release(rtdev_handle);
+	 if (rtdev_file)
+		bdev_fput(rtdev_file);
  out_close_logdev:
-	if (logdev_handle)
-		bdev_release(logdev_handle);
+	if (logdev_file)
+		bdev_fput(logdev_file);
 	return error;
 }
 
@@ -715,9 +718,7 @@ xfs_fs_inode_init_once(
 	/* xfs inode */
 	atomic_set(&ip->i_pincount, 0);
 	spin_lock_init(&ip->i_flags_lock);
-
-	mrlock_init(&ip->i_lock, MRLOCK_ALLOW_EQUAL_PRI|MRLOCK_BARRIER,
-		     "xfsino", ip->i_ino);
+	init_rwsem(&ip->i_lock);
 }
 
 /*
@@ -760,7 +761,7 @@ xfs_mount_free(
 	debugfs_remove(mp->m_debugfs);
 	kfree(mp->m_rtname);
 	kfree(mp->m_logname);
-	kmem_free(mp);
+	kfree(mp);
 }
 
 STATIC int
@@ -1052,12 +1053,18 @@ xfs_init_percpu_counters(
 	if (error)
 		goto free_fdblocks;
 
-	error = percpu_counter_init(&mp->m_frextents, 0, GFP_KERNEL);
+	error = percpu_counter_init(&mp->m_delalloc_rtextents, 0, GFP_KERNEL);
 	if (error)
 		goto free_delalloc;
 
+	error = percpu_counter_init(&mp->m_frextents, 0, GFP_KERNEL);
+	if (error)
+		goto free_delalloc_rt;
+
 	return 0;
 
+free_delalloc_rt:
+	percpu_counter_destroy(&mp->m_delalloc_rtextents);
 free_delalloc:
 	percpu_counter_destroy(&mp->m_delalloc_blks);
 free_fdblocks:
@@ -1087,6 +1094,9 @@ xfs_destroy_percpu_counters(
 	percpu_counter_destroy(&mp->m_ifree);
 	percpu_counter_destroy(&mp->m_fdblocks);
 	ASSERT(xfs_is_shutdown(mp) ||
+	       percpu_counter_sum(&mp->m_delalloc_rtextents) == 0);
+	percpu_counter_destroy(&mp->m_delalloc_rtextents);
+	ASSERT(xfs_is_shutdown(mp) ||
 	       percpu_counter_sum(&mp->m_delalloc_blks) == 0);
 	percpu_counter_destroy(&mp->m_delalloc_blks);
 	percpu_counter_destroy(&mp->m_frextents);
@@ -1580,17 +1590,21 @@ xfs_fs_fill_super(
 	if (error)
 		goto out_free_sb;
 
-	/* V4 support is undergoing deprecation. */
-	if (!xfs_has_crc(mp)) {
-#ifdef CONFIG_XFS_SUPPORT_V4
+	/*
+	 * V4 support is undergoing deprecation.
+	 *
+	 * Note: this has to use an open coded m_features check as xfs_has_crc
+	 * always returns false for !CONFIG_XFS_SUPPORT_V4.
+	 */
+	if (!(mp->m_features & XFS_FEAT_CRC)) {
+		if (!IS_ENABLED(CONFIG_XFS_SUPPORT_V4)) {
+			xfs_warn(mp,
+	"Deprecated V4 format (crc=0) not supported by kernel.");
+			error = -EINVAL;
+			goto out_free_sb;
+		}
 		xfs_warn_once(mp,
 	"Deprecated V4 format (crc=0) will not be supported after September 2030.");
-#else
-		xfs_warn(mp,
-	"Deprecated V4 format (crc=0) not supported by kernel.");
-		error = -EINVAL;
-		goto out_free_sb;
-#endif
 	}
 
 	/* ASCII case insensitivity is undergoing deprecation. */
@@ -1728,6 +1742,14 @@ xfs_fs_fill_super(
 		goto out_filestream_unmount;
 	}
 
+	if (xfs_has_exchange_range(mp))
+		xfs_warn(mp,
+	"EXPERIMENTAL exchange-range feature enabled. Use at your own risk!");
+
+	if (xfs_has_parent(mp))
+		xfs_warn(mp,
+	"EXPERIMENTAL parent pointer feature enabled. Use at your own risk!");
+
 	error = xfs_mountfs(mp);
 	if (error)
 		goto out_filestream_unmount;
@@ -1874,11 +1896,7 @@ xfs_remount_ro(
 	xfs_inodegc_stop(mp);
 
 	/* Free the per-AG metadata reservation pool. */
-	error = xfs_fs_unreserve_ag_blocks(mp);
-	if (error) {
-		xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
-		return error;
-	}
+	xfs_fs_unreserve_ag_blocks(mp);
 
 	/*
 	 * Before we sync the metadata, we need to free up the reserve block
@@ -1986,7 +2004,7 @@ static int xfs_init_fs_context(
 {
 	struct xfs_mount	*mp;
 
-	mp = kmem_alloc(sizeof(struct xfs_mount), KM_ZERO);
+	mp = kzalloc(sizeof(struct xfs_mount), GFP_KERNEL | __GFP_NOFAIL);
 	if (!mp)
 		return -ENOMEM;
 
@@ -2012,6 +2030,8 @@ static int xfs_init_fs_context(
 	mp->m_logbsize = -1;
 	mp->m_allocsize_log = 16; /* 64k */
 
+	xfs_hooks_init(&mp->m_dir_update_hooks);
+
 	fc->s_fs_info = mp;
 	fc->ops = &xfs_context_ops;
 
@@ -2043,8 +2063,7 @@ xfs_init_caches(void)
 
 	xfs_buf_cache = kmem_cache_create("xfs_buf", sizeof(struct xfs_buf), 0,
 					 SLAB_HWCACHE_ALIGN |
-					 SLAB_RECLAIM_ACCOUNT |
-					 SLAB_MEM_SPREAD,
+					 SLAB_RECLAIM_ACCOUNT,
 					 NULL);
 	if (!xfs_buf_cache)
 		goto out;
@@ -2059,10 +2078,14 @@ xfs_init_caches(void)
 	if (error)
 		goto out_destroy_log_ticket_cache;
 
-	error = xfs_defer_init_item_caches();
+	error = rcbagbt_init_cur_cache();
 	if (error)
 		goto out_destroy_btree_cur_cache;
 
+	error = xfs_defer_init_item_caches();
+	if (error)
+		goto out_destroy_rcbagbt_cur_cache;
+
 	xfs_da_state_cache = kmem_cache_create("xfs_da_state",
 					      sizeof(struct xfs_da_state),
 					      0, 0, NULL);
@@ -2109,14 +2132,14 @@ xfs_init_caches(void)
 					   sizeof(struct xfs_inode), 0,
 					   (SLAB_HWCACHE_ALIGN |
 					    SLAB_RECLAIM_ACCOUNT |
-					    SLAB_MEM_SPREAD | SLAB_ACCOUNT),
+					    SLAB_ACCOUNT),
 					   xfs_fs_inode_init_once);
 	if (!xfs_inode_cache)
 		goto out_destroy_efi_cache;
 
 	xfs_ili_cache = kmem_cache_create("xfs_ili",
 					 sizeof(struct xfs_inode_log_item), 0,
-					 SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD,
+					 SLAB_RECLAIM_ACCOUNT,
 					 NULL);
 	if (!xfs_ili_cache)
 		goto out_destroy_inode_cache;
@@ -2181,8 +2204,32 @@ xfs_init_caches(void)
 	if (!xfs_iunlink_cache)
 		goto out_destroy_attri_cache;
 
+	xfs_xmd_cache = kmem_cache_create("xfs_xmd_item",
+					 sizeof(struct xfs_xmd_log_item),
+					 0, 0, NULL);
+	if (!xfs_xmd_cache)
+		goto out_destroy_iul_cache;
+
+	xfs_xmi_cache = kmem_cache_create("xfs_xmi_item",
+					 sizeof(struct xfs_xmi_log_item),
+					 0, 0, NULL);
+	if (!xfs_xmi_cache)
+		goto out_destroy_xmd_cache;
+
+	xfs_parent_args_cache = kmem_cache_create("xfs_parent_args",
+					     sizeof(struct xfs_parent_args),
+					     0, 0, NULL);
+	if (!xfs_parent_args_cache)
+		goto out_destroy_xmi_cache;
+
 	return 0;
 
+ out_destroy_xmi_cache:
+	kmem_cache_destroy(xfs_xmi_cache);
+ out_destroy_xmd_cache:
+	kmem_cache_destroy(xfs_xmd_cache);
+ out_destroy_iul_cache:
+	kmem_cache_destroy(xfs_iunlink_cache);
  out_destroy_attri_cache:
 	kmem_cache_destroy(xfs_attri_cache);
  out_destroy_attrd_cache:
@@ -2219,6 +2266,8 @@ xfs_init_caches(void)
 	kmem_cache_destroy(xfs_da_state_cache);
  out_destroy_defer_item_cache:
 	xfs_defer_destroy_item_caches();
+ out_destroy_rcbagbt_cur_cache:
+	rcbagbt_destroy_cur_cache();
  out_destroy_btree_cur_cache:
 	xfs_btree_destroy_cur_caches();
  out_destroy_log_ticket_cache:
@@ -2237,6 +2286,9 @@ xfs_destroy_caches(void)
 	 * destroy caches.
 	 */
 	rcu_barrier();
+	kmem_cache_destroy(xfs_parent_args_cache);
+	kmem_cache_destroy(xfs_xmd_cache);
+	kmem_cache_destroy(xfs_xmi_cache);
 	kmem_cache_destroy(xfs_iunlink_cache);
 	kmem_cache_destroy(xfs_attri_cache);
 	kmem_cache_destroy(xfs_attrd_cache);
@@ -2256,6 +2308,7 @@ xfs_destroy_caches(void)
 	kmem_cache_destroy(xfs_ifork_cache);
 	kmem_cache_destroy(xfs_da_state_cache);
 	xfs_defer_destroy_item_caches();
+	rcbagbt_destroy_cur_cache();
 	xfs_btree_destroy_cur_caches();
 	kmem_cache_destroy(xfs_log_ticket_cache);
 	kmem_cache_destroy(xfs_buf_cache);
diff --git a/fs/xfs/xfs_symlink.c b/fs/xfs/xfs_symlink.c
index 92974a4414..17aee806ec 100644
--- a/fs/xfs/xfs_symlink.c
+++ b/fs/xfs/xfs_symlink.c
@@ -24,77 +24,9 @@
 #include "xfs_ialloc.h"
 #include "xfs_error.h"
 #include "xfs_health.h"
-
-/* ----- Kernel only functions below ----- */
-int
-xfs_readlink_bmap_ilocked(
-	struct xfs_inode	*ip,
-	char			*link)
-{
-	struct xfs_mount	*mp = ip->i_mount;
-	struct xfs_bmbt_irec	mval[XFS_SYMLINK_MAPS];
-	struct xfs_buf		*bp;
-	xfs_daddr_t		d;
-	char			*cur_chunk;
-	int			pathlen = ip->i_disk_size;
-	int			nmaps = XFS_SYMLINK_MAPS;
-	int			byte_cnt;
-	int			n;
-	int			error = 0;
-	int			fsblocks = 0;
-	int			offset;
-
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL));
-
-	fsblocks = xfs_symlink_blocks(mp, pathlen);
-	error = xfs_bmapi_read(ip, 0, fsblocks, mval, &nmaps, 0);
-	if (error)
-		goto out;
-
-	offset = 0;
-	for (n = 0; n < nmaps; n++) {
-		d = XFS_FSB_TO_DADDR(mp, mval[n].br_startblock);
-		byte_cnt = XFS_FSB_TO_B(mp, mval[n].br_blockcount);
-
-		error = xfs_buf_read(mp->m_ddev_targp, d, BTOBB(byte_cnt), 0,
-				&bp, &xfs_symlink_buf_ops);
-		if (error)
-			return error;
-		byte_cnt = XFS_SYMLINK_BUF_SPACE(mp, byte_cnt);
-		if (pathlen < byte_cnt)
-			byte_cnt = pathlen;
-
-		cur_chunk = bp->b_addr;
-		if (xfs_has_crc(mp)) {
-			if (!xfs_symlink_hdr_ok(ip->i_ino, offset,
-							byte_cnt, bp)) {
-				error = -EFSCORRUPTED;
-				xfs_alert(mp,
-"symlink header does not match required off/len/owner (0x%x/Ox%x,0x%llx)",
-					offset, byte_cnt, ip->i_ino);
-				xfs_buf_relse(bp);
-				goto out;
-
-			}
-
-			cur_chunk += sizeof(struct xfs_dsymlink_hdr);
-		}
-
-		memcpy(link + offset, cur_chunk, byte_cnt);
-
-		pathlen -= byte_cnt;
-		offset += byte_cnt;
-
-		xfs_buf_relse(bp);
-	}
-	ASSERT(pathlen == 0);
-
-	link[ip->i_disk_size] = '\0';
-	error = 0;
-
- out:
-	return error;
-}
+#include "xfs_symlink_remote.h"
+#include "xfs_parent.h"
+#include "xfs_defer.h"
 
 int
 xfs_readlink(
@@ -103,7 +35,7 @@ xfs_readlink(
 {
 	struct xfs_mount	*mp = ip->i_mount;
 	xfs_fsize_t		pathlen;
-	int			error = -EFSCORRUPTED;
+	int			error;
 
 	trace_xfs_readlink(ip);
 
@@ -116,14 +48,14 @@ xfs_readlink(
 
 	pathlen = ip->i_disk_size;
 	if (!pathlen)
-		goto out;
+		goto out_corrupt;
 
 	if (pathlen < 0 || pathlen > XFS_SYMLINK_MAXLEN) {
 		xfs_alert(mp, "%s: inode (%llu) bad symlink length (%lld)",
 			 __func__, (unsigned long long) ip->i_ino,
 			 (long long) pathlen);
 		ASSERT(0);
-		goto out;
+		goto out_corrupt;
 	}
 
 	if (ip->i_df.if_format == XFS_DINODE_FMT_LOCAL) {
@@ -132,17 +64,20 @@ xfs_readlink(
 		 * if if_data is junk.
 		 */
 		if (XFS_IS_CORRUPT(ip->i_mount, !ip->i_df.if_data))
-			goto out;
+			goto out_corrupt;
 
 		memcpy(link, ip->i_df.if_data, pathlen + 1);
 		error = 0;
 	} else {
-		error = xfs_readlink_bmap_ilocked(ip, link);
+		error = xfs_symlink_remote_read(ip, link);
 	}
 
- out:
 	xfs_iunlock(ip, XFS_ILOCK_SHARED);
 	return error;
+ out_corrupt:
+	xfs_iunlock(ip, XFS_ILOCK_SHARED);
+	xfs_inode_mark_sick(ip, XFS_SICK_INO_SYMLINK);
+	return -EFSCORRUPTED;
 }
 
 int
@@ -160,21 +95,14 @@ xfs_symlink(
 	int			error = 0;
 	int			pathlen;
 	bool                    unlock_dp_on_error = false;
-	xfs_fileoff_t		first_fsb;
 	xfs_filblks_t		fs_blocks;
-	int			nmaps;
-	struct xfs_bmbt_irec	mval[XFS_SYMLINK_MAPS];
-	xfs_daddr_t		d;
-	const char		*cur_chunk;
-	int			byte_cnt;
-	int			n;
-	struct xfs_buf		*bp;
 	prid_t			prid;
 	struct xfs_dquot	*udqp = NULL;
 	struct xfs_dquot	*gdqp = NULL;
 	struct xfs_dquot	*pdqp = NULL;
 	uint			resblks;
 	xfs_ino_t		ino;
+	struct xfs_parent_args	*ppargs;
 
 	*ipp = NULL;
 
@@ -205,18 +133,24 @@ xfs_symlink(
 
 	/*
 	 * The symlink will fit into the inode data fork?
-	 * There can't be any attributes so we get the whole variable part.
+	 * If there are no parent pointers, then there wont't be any attributes.
+	 * So we get the whole variable part, and do not need to reserve extra
+	 * blocks.  Otherwise, we need to reserve the blocks.
 	 */
-	if (pathlen <= XFS_LITINO(mp))
+	if (pathlen <= XFS_LITINO(mp) && !xfs_has_parent(mp))
 		fs_blocks = 0;
 	else
 		fs_blocks = xfs_symlink_blocks(mp, pathlen);
-	resblks = XFS_SYMLINK_SPACE_RES(mp, link_name->len, fs_blocks);
+	resblks = xfs_symlink_space_res(mp, link_name->len, fs_blocks);
+
+	error = xfs_parent_start(mp, &ppargs);
+	if (error)
+		goto out_release_dquots;
 
 	error = xfs_trans_alloc_icreate(mp, &M_RES(mp)->tr_symlink, udqp, gdqp,
 			pdqp, resblks, &tp);
 	if (error)
-		goto out_release_dquots;
+		goto out_parent;
 
 	xfs_ilock(dp, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT);
 	unlock_dp_on_error = true;
@@ -236,7 +170,7 @@ xfs_symlink(
 	if (!error)
 		error = xfs_init_new_inode(idmap, tp, dp, ino,
 				S_IFLNK | (mode & ~S_IFMT), 1, 0, prid,
-				false, &ip);
+				xfs_has_parent(mp), &ip);
 	if (error)
 		goto out_trans_cancel;
 
@@ -247,8 +181,7 @@ xfs_symlink(
 	 * the transaction cancel unlocking dp so don't do it explicitly in the
 	 * error path.
 	 */
-	xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
-	unlock_dp_on_error = false;
+	xfs_trans_ijoin(tp, dp, 0);
 
 	/*
 	 * Also attach the dquot(s) to it, if applicable.
@@ -256,62 +189,11 @@ xfs_symlink(
 	xfs_qm_vop_create_dqattach(tp, ip, udqp, gdqp, pdqp);
 
 	resblks -= XFS_IALLOC_SPACE_RES(mp);
-	/*
-	 * If the symlink will fit into the inode, write it inline.
-	 */
-	if (pathlen <= xfs_inode_data_fork_size(ip)) {
-		xfs_init_local_fork(ip, XFS_DATA_FORK, target_path, pathlen);
-
-		ip->i_disk_size = pathlen;
-		ip->i_df.if_format = XFS_DINODE_FMT_LOCAL;
-		xfs_trans_log_inode(tp, ip, XFS_ILOG_DDATA | XFS_ILOG_CORE);
-	} else {
-		int	offset;
-
-		first_fsb = 0;
-		nmaps = XFS_SYMLINK_MAPS;
-
-		error = xfs_bmapi_write(tp, ip, first_fsb, fs_blocks,
-				  XFS_BMAPI_METADATA, resblks, mval, &nmaps);
-		if (error)
-			goto out_trans_cancel;
-
-		resblks -= fs_blocks;
-		ip->i_disk_size = pathlen;
-		xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
-
-		cur_chunk = target_path;
-		offset = 0;
-		for (n = 0; n < nmaps; n++) {
-			char	*buf;
-
-			d = XFS_FSB_TO_DADDR(mp, mval[n].br_startblock);
-			byte_cnt = XFS_FSB_TO_B(mp, mval[n].br_blockcount);
-			error = xfs_trans_get_buf(tp, mp->m_ddev_targp, d,
-					       BTOBB(byte_cnt), 0, &bp);
-			if (error)
-				goto out_trans_cancel;
-			bp->b_ops = &xfs_symlink_buf_ops;
-
-			byte_cnt = XFS_SYMLINK_BUF_SPACE(mp, byte_cnt);
-			byte_cnt = min(byte_cnt, pathlen);
-
-			buf = bp->b_addr;
-			buf += xfs_symlink_hdr_set(mp, ip->i_ino, offset,
-						   byte_cnt, bp);
-
-			memcpy(buf, cur_chunk, byte_cnt);
-
-			cur_chunk += byte_cnt;
-			pathlen -= byte_cnt;
-			offset += byte_cnt;
-
-			xfs_trans_buf_set_type(tp, bp, XFS_BLFT_SYMLINK_BUF);
-			xfs_trans_log_buf(tp, bp, 0, (buf + byte_cnt - 1) -
-							(char *)bp->b_addr);
-		}
-		ASSERT(pathlen == 0);
-	}
+	error = xfs_symlink_write_target(tp, ip, ip->i_ino, target_path,
+			pathlen, fs_blocks, resblks);
+	if (error)
+		goto out_trans_cancel;
+	resblks -= fs_blocks;
 	i_size_write(VFS_I(ip), ip->i_disk_size);
 
 	/*
@@ -323,6 +205,15 @@ xfs_symlink(
 	xfs_trans_ichgtime(tp, dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
 	xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
 
+	/* Add parent pointer for the new symlink. */
+	if (ppargs) {
+		error = xfs_parent_addname(tp, ppargs, dp, link_name, ip);
+		if (error)
+			goto out_trans_cancel;
+	}
+
+	xfs_dir_update_hook(dp, ip, 1, link_name);
+
 	/*
 	 * If this is a synchronous mount, make sure that the
 	 * symlink transaction goes to disk before returning to
@@ -340,6 +231,9 @@ xfs_symlink(
 	xfs_qm_dqrele(pdqp);
 
 	*ipp = ip;
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	xfs_iunlock(dp, XFS_ILOCK_EXCL);
+	xfs_parent_finish(mp, ppargs);
 	return 0;
 
 out_trans_cancel:
@@ -351,9 +245,12 @@ out_release_inode:
 	 * transactions and deadlocks from xfs_inactive.
 	 */
 	if (ip) {
+		xfs_iunlock(ip, XFS_ILOCK_EXCL);
 		xfs_finish_inode_setup(ip);
 		xfs_irele(ip);
 	}
+out_parent:
+	xfs_parent_finish(mp, ppargs);
 out_release_dquots:
 	xfs_qm_dqrele(udqp);
 	xfs_qm_dqrele(gdqp);
@@ -375,19 +272,12 @@ out_release_dquots:
  */
 STATIC int
 xfs_inactive_symlink_rmt(
-	struct xfs_inode *ip)
+	struct xfs_inode	*ip)
 {
-	struct xfs_buf	*bp;
-	int		done;
-	int		error;
-	int		i;
-	xfs_mount_t	*mp;
-	xfs_bmbt_irec_t	mval[XFS_SYMLINK_MAPS];
-	int		nmaps;
-	int		size;
-	xfs_trans_t	*tp;
-
-	mp = ip->i_mount;
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_trans	*tp;
+	int			error;
+
 	ASSERT(!xfs_need_iread_extents(&ip->i_df));
 	/*
 	 * We're freeing a symlink that has some
@@ -411,44 +301,14 @@ xfs_inactive_symlink_rmt(
 	 * locked for the second transaction.  In the error paths we need it
 	 * held so the cancel won't rele it, see below.
 	 */
-	size = (int)ip->i_disk_size;
 	ip->i_disk_size = 0;
 	VFS_I(ip)->i_mode = (VFS_I(ip)->i_mode & ~S_IFMT) | S_IFREG;
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
-	/*
-	 * Find the block(s) so we can inval and unmap them.
-	 */
-	done = 0;
-	nmaps = ARRAY_SIZE(mval);
-	error = xfs_bmapi_read(ip, 0, xfs_symlink_blocks(mp, size),
-				mval, &nmaps, 0);
-	if (error)
-		goto error_trans_cancel;
-	/*
-	 * Invalidate the block(s). No validation is done.
-	 */
-	for (i = 0; i < nmaps; i++) {
-		error = xfs_trans_get_buf(tp, mp->m_ddev_targp,
-				XFS_FSB_TO_DADDR(mp, mval[i].br_startblock),
-				XFS_FSB_TO_BB(mp, mval[i].br_blockcount), 0,
-				&bp);
-		if (error)
-			goto error_trans_cancel;
-		xfs_trans_binval(tp, bp);
-	}
-	/*
-	 * Unmap the dead block(s) to the dfops.
-	 */
-	error = xfs_bunmapi(tp, ip, 0, size, 0, nmaps, &done);
+
+	error = xfs_symlink_remote_truncate(tp, ip);
 	if (error)
 		goto error_trans_cancel;
-	ASSERT(done);
 
-	/*
-	 * Commit the transaction. This first logs the EFI and the inode, then
-	 * rolls and commits the transaction that frees the extents.
-	 */
-	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 	error = xfs_trans_commit(tp);
 	if (error) {
 		ASSERT(xfs_is_shutdown(mp));
@@ -496,6 +356,7 @@ xfs_inactive_symlink(
 			 __func__, (unsigned long long)ip->i_ino, pathlen);
 		xfs_iunlock(ip, XFS_ILOCK_EXCL);
 		ASSERT(0);
+		xfs_inode_mark_sick(ip, XFS_SICK_INO_SYMLINK);
 		return -EFSCORRUPTED;
 	}
 
diff --git a/fs/xfs/xfs_symlink.h b/fs/xfs/xfs_symlink.h
index d1ca1ce62a..0d29a50e66 100644
--- a/fs/xfs/xfs_symlink.h
+++ b/fs/xfs/xfs_symlink.h
@@ -10,7 +10,6 @@
 int xfs_symlink(struct mnt_idmap *idmap, struct xfs_inode *dp,
 		struct xfs_name *link_name, const char *target_path,
 		umode_t mode, struct xfs_inode **ipp);
-int xfs_readlink_bmap_ilocked(struct xfs_inode *ip, char *link);
 int xfs_readlink(struct xfs_inode *ip, char *link);
 int xfs_inactive_symlink(struct xfs_inode *ip);
 
diff --git a/fs/xfs/xfs_sysfs.c b/fs/xfs/xfs_sysfs.c
index 17485666b6..d2391eec37 100644
--- a/fs/xfs/xfs_sysfs.c
+++ b/fs/xfs/xfs_sysfs.c
@@ -193,7 +193,6 @@ always_cow_show(
 }
 XFS_SYSFS_ATTR_RW(always_cow);
 
-#ifdef DEBUG
 /*
  * Override how many threads the parallel work queue is allowed to create.
  * This has to be a debug-only global (instead of an errortag) because one of
@@ -260,7 +259,6 @@ larp_show(
 	return snprintf(buf, PAGE_SIZE, "%d\n", xfs_globals.larp);
 }
 XFS_SYSFS_ATTR_RW(larp);
-#endif /* DEBUG */
 
 STATIC ssize_t
 bload_leaf_slack_store(
@@ -319,10 +317,8 @@ static struct attribute *xfs_dbg_attrs[] = {
 	ATTR_LIST(log_recovery_delay),
 	ATTR_LIST(mount_delay),
 	ATTR_LIST(always_cow),
-#ifdef DEBUG
 	ATTR_LIST(pwork_threads),
 	ATTR_LIST(larp),
-#endif
 	ATTR_LIST(bload_leaf_slack),
 	ATTR_LIST(bload_node_slack),
 	NULL,
diff --git a/fs/xfs/xfs_trace.c b/fs/xfs/xfs_trace.c
index 8a5dc1538a..9c7fbaae27 100644
--- a/fs/xfs/xfs_trace.c
+++ b/fs/xfs/xfs_trace.c
@@ -36,6 +36,12 @@
 #include "xfs_error.h"
 #include <linux/iomap.h>
 #include "xfs_iomap.h"
+#include "xfs_buf_mem.h"
+#include "xfs_btree_mem.h"
+#include "xfs_bmap.h"
+#include "xfs_exchmaps.h"
+#include "xfs_exchrange.h"
+#include "xfs_parent.h"
 
 /*
  * We include this last to have the helpers above available for the trace
diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
index 0984a1c884..25ff6fe1eb 100644
--- a/fs/xfs/xfs_trace.h
+++ b/fs/xfs/xfs_trace.h
@@ -31,6 +31,8 @@
  * pos: file offset, in bytes
  * bytecount: number of bytes
  *
+ * dablk: directory or xattr block offset, in filesystem blocks
+ *
  * disize: ondisk file size, in bytes
  * isize: incore file size, in bytes
  *
@@ -79,11 +81,21 @@ union xfs_btree_ptr;
 struct xfs_dqtrx;
 struct xfs_icwalk;
 struct xfs_perag;
+struct xfbtree;
+struct xfs_btree_ops;
+struct xfs_bmap_intent;
+struct xfs_exchmaps_intent;
+struct xfs_exchmaps_req;
+struct xfs_exchrange;
+struct xfs_getparents;
+struct xfs_parent_irec;
+struct xfs_attrlist_cursor_kern;
 
 #define XFS_ATTR_FILTER_FLAGS \
 	{ XFS_ATTR_ROOT,	"ROOT" }, \
 	{ XFS_ATTR_SECURE,	"SECURE" }, \
-	{ XFS_ATTR_INCOMPLETE,	"INCOMPLETE" }
+	{ XFS_ATTR_INCOMPLETE,	"INCOMPLETE" }, \
+	{ XFS_ATTR_PARENT,	"PARENT" }
 
 DECLARE_EVENT_CLASS(xfs_attr_list_class,
 	TP_PROTO(struct xfs_attr_list_context *ctx),
@@ -156,7 +168,7 @@ TRACE_EVENT(xlog_intent_recovery_failed,
 	),
 	TP_fast_assign(
 		__entry->dev = mp->m_super->s_dev;
-		__assign_str(name, ops->name);
+		__assign_str(name);
 		__entry->error = error;
 	),
 	TP_printk("dev %d:%d optype %s error %d",
@@ -640,6 +652,7 @@ DEFINE_BUF_ITEM_EVENT(xfs_trans_read_buf);
 DEFINE_BUF_ITEM_EVENT(xfs_trans_read_buf_recur);
 DEFINE_BUF_ITEM_EVENT(xfs_trans_log_buf);
 DEFINE_BUF_ITEM_EVENT(xfs_trans_brelse);
+DEFINE_BUF_ITEM_EVENT(xfs_trans_bdetach);
 DEFINE_BUF_ITEM_EVENT(xfs_trans_bjoin);
 DEFINE_BUF_ITEM_EVENT(xfs_trans_bhold);
 DEFINE_BUF_ITEM_EVENT(xfs_trans_bhold_release);
@@ -1650,7 +1663,6 @@ DEFINE_EVENT(xfs_extent_busy_class, name, \
 		 xfs_agblock_t agbno, xfs_extlen_t len), \
 	TP_ARGS(mp, agno, agbno, len))
 DEFINE_BUSY_EVENT(xfs_extent_busy);
-DEFINE_BUSY_EVENT(xfs_extent_busy_enomem);
 DEFINE_BUSY_EVENT(xfs_extent_busy_force);
 DEFINE_BUSY_EVENT(xfs_extent_busy_reuse);
 DEFINE_BUSY_EVENT(xfs_extent_busy_clear);
@@ -1710,12 +1722,10 @@ DECLARE_EVENT_CLASS(xfs_agf_class,
 		__entry->agno = be32_to_cpu(agf->agf_seqno),
 		__entry->flags = flags;
 		__entry->length = be32_to_cpu(agf->agf_length),
-		__entry->bno_root = be32_to_cpu(agf->agf_roots[XFS_BTNUM_BNO]),
-		__entry->cnt_root = be32_to_cpu(agf->agf_roots[XFS_BTNUM_CNT]),
-		__entry->bno_level =
-				be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]),
-		__entry->cnt_level =
-				be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]),
+		__entry->bno_root = be32_to_cpu(agf->agf_bno_root),
+		__entry->cnt_root = be32_to_cpu(agf->agf_cnt_root),
+		__entry->bno_level = be32_to_cpu(agf->agf_bno_level),
+		__entry->cnt_level = be32_to_cpu(agf->agf_cnt_level),
 		__entry->flfirst = be32_to_cpu(agf->agf_flfirst),
 		__entry->fllast = be32_to_cpu(agf->agf_fllast),
 		__entry->flcount = be32_to_cpu(agf->agf_flcount),
@@ -1890,28 +1900,28 @@ DEFINE_ALLOC_EVENT(xfs_alloc_vextent_near_bno);
 DEFINE_ALLOC_EVENT(xfs_alloc_vextent_finish);
 
 TRACE_EVENT(xfs_alloc_cur_check,
-	TP_PROTO(struct xfs_mount *mp, xfs_btnum_t btnum, xfs_agblock_t bno,
+	TP_PROTO(struct xfs_btree_cur *cur, xfs_agblock_t bno,
 		 xfs_extlen_t len, xfs_extlen_t diff, bool new),
-	TP_ARGS(mp, btnum, bno, len, diff, new),
+	TP_ARGS(cur, bno, len, diff, new),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
-		__field(xfs_btnum_t, btnum)
+		__string(name, cur->bc_ops->name)
 		__field(xfs_agblock_t, bno)
 		__field(xfs_extlen_t, len)
 		__field(xfs_extlen_t, diff)
 		__field(bool, new)
 	),
 	TP_fast_assign(
-		__entry->dev = mp->m_super->s_dev;
-		__entry->btnum = btnum;
+		__entry->dev = cur->bc_mp->m_super->s_dev;
+		__assign_str(name);
 		__entry->bno = bno;
 		__entry->len = len;
 		__entry->diff = diff;
 		__entry->new = new;
 	),
-	TP_printk("dev %d:%d btree %s agbno 0x%x fsbcount 0x%x diff 0x%x new %d",
+	TP_printk("dev %d:%d %sbt agbno 0x%x fsbcount 0x%x diff 0x%x new %d",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
-		  __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS),
+		  __get_str(name),
 		  __entry->bno, __entry->len, __entry->diff, __entry->new)
 )
 
@@ -1926,6 +1936,7 @@ DECLARE_EVENT_CLASS(xfs_da_class,
 		__field(xfs_dahash_t, hashval)
 		__field(xfs_ino_t, inumber)
 		__field(uint32_t, op_flags)
+		__field(xfs_ino_t, owner)
 	),
 	TP_fast_assign(
 		__entry->dev = VFS_I(args->dp)->i_sb->s_dev;
@@ -1936,9 +1947,10 @@ DECLARE_EVENT_CLASS(xfs_da_class,
 		__entry->hashval = args->hashval;
 		__entry->inumber = args->inumber;
 		__entry->op_flags = args->op_flags;
+		__entry->owner = args->owner;
 	),
 	TP_printk("dev %d:%d ino 0x%llx name %.*s namelen %d hashval 0x%x "
-		  "inumber 0x%llx op_flags %s",
+		  "inumber 0x%llx op_flags %s owner 0x%llx",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __entry->ino,
 		  __entry->namelen,
@@ -1946,7 +1958,8 @@ DECLARE_EVENT_CLASS(xfs_da_class,
 		  __entry->namelen,
 		  __entry->hashval,
 		  __entry->inumber,
-		  __print_flags(__entry->op_flags, "|", XFS_DA_OP_FLAGS))
+		  __print_flags(__entry->op_flags, "|", XFS_DA_OP_FLAGS),
+		  __entry->owner)
 )
 
 #define DEFINE_DIR2_EVENT(name) \
@@ -1990,7 +2003,6 @@ DECLARE_EVENT_CLASS(xfs_attr_class,
 		__field(int, valuelen)
 		__field(xfs_dahash_t, hashval)
 		__field(unsigned int, attr_filter)
-		__field(unsigned int, attr_flags)
 		__field(uint32_t, op_flags)
 	),
 	TP_fast_assign(
@@ -2002,11 +2014,10 @@ DECLARE_EVENT_CLASS(xfs_attr_class,
 		__entry->valuelen = args->valuelen;
 		__entry->hashval = args->hashval;
 		__entry->attr_filter = args->attr_filter;
-		__entry->attr_flags = args->attr_flags;
 		__entry->op_flags = args->op_flags;
 	),
 	TP_printk("dev %d:%d ino 0x%llx name %.*s namelen %d valuelen %d "
-		  "hashval 0x%x filter %s flags %s op_flags %s",
+		  "hashval 0x%x filter %s op_flags %s",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __entry->ino,
 		  __entry->namelen,
@@ -2016,9 +2027,6 @@ DECLARE_EVENT_CLASS(xfs_attr_class,
 		  __entry->hashval,
 		  __print_flags(__entry->attr_filter, "|",
 				XFS_ATTR_FILTER_FLAGS),
-		   __print_flags(__entry->attr_flags, "|",
-				{ XATTR_CREATE,		"CREATE" },
-				{ XATTR_REPLACE,	"REPLACE" }),
 		  __print_flags(__entry->op_flags, "|", XFS_DA_OP_FLAGS))
 )
 
@@ -2452,21 +2460,12 @@ DEFINE_DISCARD_EVENT(xfs_discard_toosmall);
 DEFINE_DISCARD_EVENT(xfs_discard_exclude);
 DEFINE_DISCARD_EVENT(xfs_discard_busy);
 
-/* btree cursor events */
-TRACE_DEFINE_ENUM(XFS_BTNUM_BNOi);
-TRACE_DEFINE_ENUM(XFS_BTNUM_CNTi);
-TRACE_DEFINE_ENUM(XFS_BTNUM_BMAPi);
-TRACE_DEFINE_ENUM(XFS_BTNUM_INOi);
-TRACE_DEFINE_ENUM(XFS_BTNUM_FINOi);
-TRACE_DEFINE_ENUM(XFS_BTNUM_RMAPi);
-TRACE_DEFINE_ENUM(XFS_BTNUM_REFCi);
-
 DECLARE_EVENT_CLASS(xfs_btree_cur_class,
 	TP_PROTO(struct xfs_btree_cur *cur, int level, struct xfs_buf *bp),
 	TP_ARGS(cur, level, bp),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
-		__field(xfs_btnum_t, btnum)
+		__string(name, cur->bc_ops->name)
 		__field(int, level)
 		__field(int, nlevels)
 		__field(int, ptr)
@@ -2474,15 +2473,15 @@ DECLARE_EVENT_CLASS(xfs_btree_cur_class,
 	),
 	TP_fast_assign(
 		__entry->dev = cur->bc_mp->m_super->s_dev;
-		__entry->btnum = cur->bc_btnum;
+		__assign_str(name);
 		__entry->level = level;
 		__entry->nlevels = cur->bc_nlevels;
 		__entry->ptr = cur->bc_levels[level].ptr;
 		__entry->daddr = bp ? xfs_buf_daddr(bp) : -1;
 	),
-	TP_printk("dev %d:%d btree %s level %d/%d ptr %d daddr 0x%llx",
+	TP_printk("dev %d:%d %sbt level %d/%d ptr %d daddr 0x%llx",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
-		  __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS),
+		  __get_str(name),
 		  __entry->level,
 		  __entry->nlevels,
 		  __entry->ptr,
@@ -2496,6 +2495,90 @@ DEFINE_EVENT(xfs_btree_cur_class, name, \
 DEFINE_BTREE_CUR_EVENT(xfs_btree_updkeys);
 DEFINE_BTREE_CUR_EVENT(xfs_btree_overlapped_query_range);
 
+TRACE_EVENT(xfs_btree_alloc_block,
+	TP_PROTO(struct xfs_btree_cur *cur, union xfs_btree_ptr *ptr, int stat,
+		 int error),
+	TP_ARGS(cur, ptr, stat, error),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_ino_t, ino)
+		__string(name, cur->bc_ops->name)
+		__field(int, error)
+		__field(xfs_agblock_t, agbno)
+	),
+	TP_fast_assign(
+		__entry->dev = cur->bc_mp->m_super->s_dev;
+		switch (cur->bc_ops->type) {
+		case XFS_BTREE_TYPE_INODE:
+			__entry->agno = 0;
+			__entry->ino = cur->bc_ino.ip->i_ino;
+			break;
+		case XFS_BTREE_TYPE_AG:
+			__entry->agno = cur->bc_ag.pag->pag_agno;
+			__entry->ino = 0;
+			break;
+		case XFS_BTREE_TYPE_MEM:
+			__entry->agno = 0;
+			__entry->ino = 0;
+			break;
+		}
+		__assign_str(name);
+		__entry->error = error;
+		if (!error && stat) {
+			if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) {
+				xfs_fsblock_t	fsb = be64_to_cpu(ptr->l);
+
+				__entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp,
+								fsb);
+				__entry->agbno = XFS_FSB_TO_AGBNO(cur->bc_mp,
+								fsb);
+			} else {
+				__entry->agbno = be32_to_cpu(ptr->s);
+			}
+		} else {
+			__entry->agbno = NULLAGBLOCK;
+		}
+	),
+	TP_printk("dev %d:%d %sbt agno 0x%x ino 0x%llx agbno 0x%x error %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __get_str(name),
+		  __entry->agno,
+		  __entry->ino,
+		  __entry->agbno,
+		  __entry->error)
+);
+
+TRACE_EVENT(xfs_btree_free_block,
+	TP_PROTO(struct xfs_btree_cur *cur, struct xfs_buf *bp),
+	TP_ARGS(cur, bp),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_ino_t, ino)
+		__string(name, cur->bc_ops->name)
+		__field(xfs_agblock_t, agbno)
+	),
+	TP_fast_assign(
+		__entry->dev = cur->bc_mp->m_super->s_dev;
+		__entry->agno = xfs_daddr_to_agno(cur->bc_mp,
+							xfs_buf_daddr(bp));
+		if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE)
+			__entry->ino = cur->bc_ino.ip->i_ino;
+		else
+			__entry->ino = 0;
+		__assign_str(name);
+		__entry->agbno = xfs_daddr_to_agbno(cur->bc_mp,
+							xfs_buf_daddr(bp));
+	),
+	TP_printk("dev %d:%d %sbt agno 0x%x ino 0x%llx agbno 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __get_str(name),
+		  __entry->agno,
+		  __entry->ino,
+		  __entry->agbno)
+);
+
 /* deferred ops */
 struct xfs_defer_pending;
 
@@ -2560,7 +2643,7 @@ DECLARE_EVENT_CLASS(xfs_defer_pending_class,
 	),
 	TP_fast_assign(
 		__entry->dev = mp ? mp->m_super->s_dev : 0;
-		__assign_str(name, dfp->dfp_ops->name);
+		__assign_str(name);
 		__entry->intent = dfp->dfp_intent;
 		__entry->flags = dfp->dfp_flags;
 		__entry->committed = dfp->dfp_done != NULL;
@@ -2579,7 +2662,25 @@ DEFINE_EVENT(xfs_defer_pending_class, name, \
 	TP_PROTO(struct xfs_mount *mp, struct xfs_defer_pending *dfp), \
 	TP_ARGS(mp, dfp))
 
-DECLARE_EVENT_CLASS(xfs_phys_extent_deferred_class,
+DEFINE_DEFER_EVENT(xfs_defer_cancel);
+DEFINE_DEFER_EVENT(xfs_defer_trans_roll);
+DEFINE_DEFER_EVENT(xfs_defer_trans_abort);
+DEFINE_DEFER_EVENT(xfs_defer_finish);
+DEFINE_DEFER_EVENT(xfs_defer_finish_done);
+
+DEFINE_DEFER_ERROR_EVENT(xfs_defer_trans_roll_error);
+DEFINE_DEFER_ERROR_EVENT(xfs_defer_finish_error);
+
+DEFINE_DEFER_PENDING_EVENT(xfs_defer_create_intent);
+DEFINE_DEFER_PENDING_EVENT(xfs_defer_cancel_list);
+DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_finish);
+DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_abort);
+DEFINE_DEFER_PENDING_EVENT(xfs_defer_relog_intent);
+DEFINE_DEFER_PENDING_EVENT(xfs_defer_isolate_paused);
+DEFINE_DEFER_PENDING_EVENT(xfs_defer_item_pause);
+DEFINE_DEFER_PENDING_EVENT(xfs_defer_item_unpause);
+
+DECLARE_EVENT_CLASS(xfs_free_extent_deferred_class,
 	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
 		 int type, xfs_agblock_t agbno, xfs_extlen_t len),
 	TP_ARGS(mp, agno, type, agbno, len),
@@ -2604,92 +2705,17 @@ DECLARE_EVENT_CLASS(xfs_phys_extent_deferred_class,
 		  __entry->agbno,
 		  __entry->len)
 );
-#define DEFINE_PHYS_EXTENT_DEFERRED_EVENT(name) \
-DEFINE_EVENT(xfs_phys_extent_deferred_class, name, \
+#define DEFINE_FREE_EXTENT_DEFERRED_EVENT(name) \
+DEFINE_EVENT(xfs_free_extent_deferred_class, name, \
 	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \
 		 int type, \
 		 xfs_agblock_t bno, \
 		 xfs_extlen_t len), \
 	TP_ARGS(mp, agno, type, bno, len))
-
-DECLARE_EVENT_CLASS(xfs_map_extent_deferred_class,
-	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
-		 int op,
-		 xfs_agblock_t agbno,
-		 xfs_ino_t ino,
-		 int whichfork,
-		 xfs_fileoff_t offset,
-		 xfs_filblks_t len,
-		 xfs_exntst_t state),
-	TP_ARGS(mp, agno, op, agbno, ino, whichfork, offset, len, state),
-	TP_STRUCT__entry(
-		__field(dev_t, dev)
-		__field(xfs_agnumber_t, agno)
-		__field(xfs_ino_t, ino)
-		__field(xfs_agblock_t, agbno)
-		__field(int, whichfork)
-		__field(xfs_fileoff_t, l_loff)
-		__field(xfs_filblks_t, l_len)
-		__field(xfs_exntst_t, l_state)
-		__field(int, op)
-	),
-	TP_fast_assign(
-		__entry->dev = mp->m_super->s_dev;
-		__entry->agno = agno;
-		__entry->ino = ino;
-		__entry->agbno = agbno;
-		__entry->whichfork = whichfork;
-		__entry->l_loff = offset;
-		__entry->l_len = len;
-		__entry->l_state = state;
-		__entry->op = op;
-	),
-	TP_printk("dev %d:%d op %d agno 0x%x agbno 0x%x owner 0x%llx %s fileoff 0x%llx fsbcount 0x%llx state %d",
-		  MAJOR(__entry->dev), MINOR(__entry->dev),
-		  __entry->op,
-		  __entry->agno,
-		  __entry->agbno,
-		  __entry->ino,
-		  __print_symbolic(__entry->whichfork, XFS_WHICHFORK_STRINGS),
-		  __entry->l_loff,
-		  __entry->l_len,
-		  __entry->l_state)
-);
-#define DEFINE_MAP_EXTENT_DEFERRED_EVENT(name) \
-DEFINE_EVENT(xfs_map_extent_deferred_class, name, \
-	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \
-		 int op, \
-		 xfs_agblock_t agbno, \
-		 xfs_ino_t ino, \
-		 int whichfork, \
-		 xfs_fileoff_t offset, \
-		 xfs_filblks_t len, \
-		 xfs_exntst_t state), \
-	TP_ARGS(mp, agno, op, agbno, ino, whichfork, offset, len, state))
-
-DEFINE_DEFER_EVENT(xfs_defer_cancel);
-DEFINE_DEFER_EVENT(xfs_defer_trans_roll);
-DEFINE_DEFER_EVENT(xfs_defer_trans_abort);
-DEFINE_DEFER_EVENT(xfs_defer_finish);
-DEFINE_DEFER_EVENT(xfs_defer_finish_done);
-
-DEFINE_DEFER_ERROR_EVENT(xfs_defer_trans_roll_error);
-DEFINE_DEFER_ERROR_EVENT(xfs_defer_finish_error);
-
-DEFINE_DEFER_PENDING_EVENT(xfs_defer_create_intent);
-DEFINE_DEFER_PENDING_EVENT(xfs_defer_cancel_list);
-DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_finish);
-DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_abort);
-DEFINE_DEFER_PENDING_EVENT(xfs_defer_relog_intent);
-DEFINE_DEFER_PENDING_EVENT(xfs_defer_isolate_paused);
-DEFINE_DEFER_PENDING_EVENT(xfs_defer_item_pause);
-DEFINE_DEFER_PENDING_EVENT(xfs_defer_item_unpause);
-
-#define DEFINE_BMAP_FREE_DEFERRED_EVENT DEFINE_PHYS_EXTENT_DEFERRED_EVENT
-DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_bmap_free_defer);
-DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_bmap_free_deferred);
-DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_agfl_free_defer);
-DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_agfl_free_deferred);
+DEFINE_FREE_EXTENT_DEFERRED_EVENT(xfs_bmap_free_defer);
+DEFINE_FREE_EXTENT_DEFERRED_EVENT(xfs_bmap_free_deferred);
+DEFINE_FREE_EXTENT_DEFERRED_EVENT(xfs_agfl_free_defer);
+DEFINE_FREE_EXTENT_DEFERRED_EVENT(xfs_agfl_free_deferred);
 
 DECLARE_EVENT_CLASS(xfs_defer_pending_item_class,
 	TP_PROTO(struct xfs_mount *mp, struct xfs_defer_pending *dfp,
@@ -2706,7 +2732,7 @@ DECLARE_EVENT_CLASS(xfs_defer_pending_item_class,
 	),
 	TP_fast_assign(
 		__entry->dev = mp ? mp->m_super->s_dev : 0;
-		__assign_str(name, dfp->dfp_ops->name);
+		__assign_str(name);
 		__entry->intent = dfp->dfp_intent;
 		__entry->item = item;
 		__entry->committed = dfp->dfp_done != NULL;
@@ -2854,12 +2880,63 @@ DEFINE_EVENT(xfs_rmapbt_class, name, \
 		 uint64_t owner, uint64_t offset, unsigned int flags), \
 	TP_ARGS(mp, agno, agbno, len, owner, offset, flags))
 
-#define DEFINE_RMAP_DEFERRED_EVENT DEFINE_MAP_EXTENT_DEFERRED_EVENT
+DECLARE_EVENT_CLASS(xfs_rmap_deferred_class,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 int op,
+		 xfs_agblock_t agbno,
+		 xfs_ino_t ino,
+		 int whichfork,
+		 xfs_fileoff_t offset,
+		 xfs_filblks_t len,
+		 xfs_exntst_t state),
+	TP_ARGS(mp, agno, op, agbno, ino, whichfork, offset, len, state),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_ino_t, ino)
+		__field(xfs_agblock_t, agbno)
+		__field(int, whichfork)
+		__field(xfs_fileoff_t, l_loff)
+		__field(xfs_filblks_t, l_len)
+		__field(xfs_exntst_t, l_state)
+		__field(int, op)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->ino = ino;
+		__entry->agbno = agbno;
+		__entry->whichfork = whichfork;
+		__entry->l_loff = offset;
+		__entry->l_len = len;
+		__entry->l_state = state;
+		__entry->op = op;
+	),
+	TP_printk("dev %d:%d op %d agno 0x%x agbno 0x%x owner 0x%llx %s fileoff 0x%llx fsbcount 0x%llx state %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->op,
+		  __entry->agno,
+		  __entry->agbno,
+		  __entry->ino,
+		  __print_symbolic(__entry->whichfork, XFS_WHICHFORK_STRINGS),
+		  __entry->l_loff,
+		  __entry->l_len,
+		  __entry->l_state)
+);
+#define DEFINE_RMAP_DEFERRED_EVENT(name) \
+DEFINE_EVENT(xfs_rmap_deferred_class, name, \
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \
+		 int op, \
+		 xfs_agblock_t agbno, \
+		 xfs_ino_t ino, \
+		 int whichfork, \
+		 xfs_fileoff_t offset, \
+		 xfs_filblks_t len, \
+		 xfs_exntst_t state), \
+	TP_ARGS(mp, agno, op, agbno, ino, whichfork, offset, len, state))
 DEFINE_RMAP_DEFERRED_EVENT(xfs_rmap_defer);
 DEFINE_RMAP_DEFERRED_EVENT(xfs_rmap_deferred);
 
-DEFINE_BUSY_EVENT(xfs_rmapbt_alloc_block);
-DEFINE_BUSY_EVENT(xfs_rmapbt_free_block);
 DEFINE_RMAPBT_EVENT(xfs_rmap_update);
 DEFINE_RMAPBT_EVENT(xfs_rmap_insert);
 DEFINE_RMAPBT_EVENT(xfs_rmap_delete);
@@ -2876,7 +2953,66 @@ DEFINE_RMAPBT_EVENT(xfs_rmap_find_right_neighbor_result);
 DEFINE_RMAPBT_EVENT(xfs_rmap_find_left_neighbor_result);
 
 /* deferred bmbt updates */
-#define DEFINE_BMAP_DEFERRED_EVENT	DEFINE_RMAP_DEFERRED_EVENT
+TRACE_DEFINE_ENUM(XFS_BMAP_MAP);
+TRACE_DEFINE_ENUM(XFS_BMAP_UNMAP);
+
+DECLARE_EVENT_CLASS(xfs_bmap_deferred_class,
+	TP_PROTO(struct xfs_bmap_intent *bi),
+	TP_ARGS(bi),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(dev_t, opdev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_ino_t, ino)
+		__field(xfs_agblock_t, agbno)
+		__field(xfs_fsblock_t, rtbno)
+		__field(int, whichfork)
+		__field(xfs_fileoff_t, l_loff)
+		__field(xfs_filblks_t, l_len)
+		__field(xfs_exntst_t, l_state)
+		__field(int, op)
+	),
+	TP_fast_assign(
+		struct xfs_inode	*ip = bi->bi_owner;
+
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		if (xfs_ifork_is_realtime(ip, bi->bi_whichfork)) {
+			__entry->agno = 0;
+			__entry->agbno = 0;
+			__entry->rtbno = bi->bi_bmap.br_startblock;
+			__entry->opdev = ip->i_mount->m_rtdev_targp->bt_dev;
+		} else {
+			__entry->agno = XFS_FSB_TO_AGNO(ip->i_mount,
+						bi->bi_bmap.br_startblock);
+			__entry->agbno = XFS_FSB_TO_AGBNO(ip->i_mount,
+						bi->bi_bmap.br_startblock);
+			__entry->rtbno = 0;
+			__entry->opdev = __entry->dev;
+		}
+		__entry->ino = ip->i_ino;
+		__entry->whichfork = bi->bi_whichfork;
+		__entry->l_loff = bi->bi_bmap.br_startoff;
+		__entry->l_len = bi->bi_bmap.br_blockcount;
+		__entry->l_state = bi->bi_bmap.br_state;
+		__entry->op = bi->bi_type;
+	),
+	TP_printk("dev %d:%d op %s opdev %d:%d ino 0x%llx agno 0x%x agbno 0x%x rtbno 0x%llx %s fileoff 0x%llx fsbcount 0x%llx state %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->op, XFS_BMAP_INTENT_STRINGS),
+		  MAJOR(__entry->opdev), MINOR(__entry->opdev),
+		  __entry->ino,
+		  __entry->agno,
+		  __entry->agbno,
+		  __entry->rtbno,
+		  __print_symbolic(__entry->whichfork, XFS_WHICHFORK_STRINGS),
+		  __entry->l_loff,
+		  __entry->l_len,
+		  __entry->l_state)
+);
+#define DEFINE_BMAP_DEFERRED_EVENT(name) \
+DEFINE_EVENT(xfs_bmap_deferred_class, name, \
+	TP_PROTO(struct xfs_bmap_intent *bi), \
+	TP_ARGS(bi))
 DEFINE_BMAP_DEFERRED_EVENT(xfs_bmap_defer);
 DEFINE_BMAP_DEFERRED_EVENT(xfs_bmap_deferred);
 
@@ -2932,7 +3068,6 @@ DEFINE_AG_RESV_EVENT(xfs_ag_resv_free_extent);
 DEFINE_AG_RESV_EVENT(xfs_ag_resv_critical);
 DEFINE_AG_RESV_EVENT(xfs_ag_resv_needed);
 
-DEFINE_AG_ERROR_EVENT(xfs_ag_resv_free_error);
 DEFINE_AG_ERROR_EVENT(xfs_ag_resv_init_error);
 
 /* refcount tracepoint classes */
@@ -3217,8 +3352,6 @@ DEFINE_EVENT(xfs_refcount_triple_extent_class, name, \
 	TP_ARGS(mp, agno, i1, i2, i3))
 
 /* refcount btree tracepoints */
-DEFINE_BUSY_EVENT(xfs_refcountbt_alloc_block);
-DEFINE_BUSY_EVENT(xfs_refcountbt_free_block);
 DEFINE_AG_BTREE_LOOKUP_EVENT(xfs_refcount_lookup);
 DEFINE_REFCOUNT_EXTENT_EVENT(xfs_refcount_get);
 DEFINE_REFCOUNT_EXTENT_EVENT(xfs_refcount_update);
@@ -3255,7 +3388,39 @@ DEFINE_AG_ERROR_EVENT(xfs_refcount_find_right_extent_error);
 DEFINE_AG_EXTENT_EVENT(xfs_refcount_find_shared);
 DEFINE_AG_EXTENT_EVENT(xfs_refcount_find_shared_result);
 DEFINE_AG_ERROR_EVENT(xfs_refcount_find_shared_error);
-#define DEFINE_REFCOUNT_DEFERRED_EVENT DEFINE_PHYS_EXTENT_DEFERRED_EVENT
+
+DECLARE_EVENT_CLASS(xfs_refcount_deferred_class,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 int type, xfs_agblock_t agbno, xfs_extlen_t len),
+	TP_ARGS(mp, agno, type, agbno, len),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(int, type)
+		__field(xfs_agblock_t, agbno)
+		__field(xfs_extlen_t, len)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->type = type;
+		__entry->agbno = agbno;
+		__entry->len = len;
+	),
+	TP_printk("dev %d:%d op %d agno 0x%x agbno 0x%x fsbcount 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->type,
+		  __entry->agno,
+		  __entry->agbno,
+		  __entry->len)
+);
+#define DEFINE_REFCOUNT_DEFERRED_EVENT(name) \
+DEFINE_EVENT(xfs_refcount_deferred_class, name, \
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \
+		 int type, \
+		 xfs_agblock_t bno, \
+		 xfs_extlen_t len), \
+	TP_ARGS(mp, agno, type, bno, len))
 DEFINE_REFCOUNT_DEFERRED_EVENT(xfs_refcount_defer);
 DEFINE_REFCOUNT_DEFERRED_EVENT(xfs_refcount_deferred);
 
@@ -3926,9 +4091,11 @@ DEFINE_EVENT(xfs_fs_corrupt_class, name,	\
 	TP_PROTO(struct xfs_mount *mp, unsigned int flags), \
 	TP_ARGS(mp, flags))
 DEFINE_FS_CORRUPT_EVENT(xfs_fs_mark_sick);
+DEFINE_FS_CORRUPT_EVENT(xfs_fs_mark_corrupt);
 DEFINE_FS_CORRUPT_EVENT(xfs_fs_mark_healthy);
 DEFINE_FS_CORRUPT_EVENT(xfs_fs_unfixed_corruption);
 DEFINE_FS_CORRUPT_EVENT(xfs_rt_mark_sick);
+DEFINE_FS_CORRUPT_EVENT(xfs_rt_mark_corrupt);
 DEFINE_FS_CORRUPT_EVENT(xfs_rt_mark_healthy);
 DEFINE_FS_CORRUPT_EVENT(xfs_rt_unfixed_corruption);
 
@@ -3955,6 +4122,7 @@ DEFINE_EVENT(xfs_ag_corrupt_class, name,	\
 		 unsigned int flags), \
 	TP_ARGS(mp, agno, flags))
 DEFINE_AG_CORRUPT_EVENT(xfs_ag_mark_sick);
+DEFINE_AG_CORRUPT_EVENT(xfs_ag_mark_corrupt);
 DEFINE_AG_CORRUPT_EVENT(xfs_ag_mark_healthy);
 DEFINE_AG_CORRUPT_EVENT(xfs_ag_unfixed_corruption);
 
@@ -3980,7 +4148,9 @@ DEFINE_EVENT(xfs_inode_corrupt_class, name,	\
 	TP_PROTO(struct xfs_inode *ip, unsigned int flags), \
 	TP_ARGS(ip, flags))
 DEFINE_INODE_CORRUPT_EVENT(xfs_inode_mark_sick);
+DEFINE_INODE_CORRUPT_EVENT(xfs_inode_mark_corrupt);
 DEFINE_INODE_CORRUPT_EVENT(xfs_inode_mark_healthy);
+DEFINE_INODE_CORRUPT_EVENT(xfs_inode_unfixed_corruption);
 
 TRACE_EVENT(xfs_iwalk_ag,
 	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
@@ -4040,31 +4210,6 @@ TRACE_EVENT(xfs_pwork_init,
 		  __entry->nr_threads, __entry->pid)
 )
 
-DECLARE_EVENT_CLASS(xfs_kmem_class,
-	TP_PROTO(ssize_t size, int flags, unsigned long caller_ip),
-	TP_ARGS(size, flags, caller_ip),
-	TP_STRUCT__entry(
-		__field(ssize_t, size)
-		__field(int, flags)
-		__field(unsigned long, caller_ip)
-	),
-	TP_fast_assign(
-		__entry->size = size;
-		__entry->flags = flags;
-		__entry->caller_ip = caller_ip;
-	),
-	TP_printk("size %zd flags 0x%x caller %pS",
-		  __entry->size,
-		  __entry->flags,
-		  (char *)__entry->caller_ip)
-)
-
-#define DEFINE_KMEM_EVENT(name) \
-DEFINE_EVENT(xfs_kmem_class, name, \
-	TP_PROTO(ssize_t size, int flags, unsigned long caller_ip), \
-	TP_ARGS(size, flags, caller_ip))
-DEFINE_KMEM_EVENT(kmem_alloc);
-
 TRACE_EVENT(xfs_check_new_dalign,
 	TP_PROTO(struct xfs_mount *mp, int new_dalign, xfs_ino_t calc_rootino),
 	TP_ARGS(mp, new_dalign, calc_rootino),
@@ -4091,7 +4236,7 @@ TRACE_EVENT(xfs_btree_commit_afakeroot,
 	TP_ARGS(cur),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
-		__field(xfs_btnum_t, btnum)
+		__string(name, cur->bc_ops->name)
 		__field(xfs_agnumber_t, agno)
 		__field(xfs_agblock_t, agbno)
 		__field(unsigned int, levels)
@@ -4099,15 +4244,15 @@ TRACE_EVENT(xfs_btree_commit_afakeroot,
 	),
 	TP_fast_assign(
 		__entry->dev = cur->bc_mp->m_super->s_dev;
-		__entry->btnum = cur->bc_btnum;
+		__assign_str(name);
 		__entry->agno = cur->bc_ag.pag->pag_agno;
 		__entry->agbno = cur->bc_ag.afake->af_root;
 		__entry->levels = cur->bc_ag.afake->af_levels;
 		__entry->blocks = cur->bc_ag.afake->af_blocks;
 	),
-	TP_printk("dev %d:%d btree %s agno 0x%x levels %u blocks %u root %u",
+	TP_printk("dev %d:%d %sbt agno 0x%x levels %u blocks %u root %u",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
-		  __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS),
+		  __get_str(name),
 		  __entry->agno,
 		  __entry->levels,
 		  __entry->blocks,
@@ -4119,7 +4264,7 @@ TRACE_EVENT(xfs_btree_commit_ifakeroot,
 	TP_ARGS(cur),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
-		__field(xfs_btnum_t, btnum)
+		__string(name, cur->bc_ops->name)
 		__field(xfs_agnumber_t, agno)
 		__field(xfs_agino_t, agino)
 		__field(unsigned int, levels)
@@ -4128,7 +4273,7 @@ TRACE_EVENT(xfs_btree_commit_ifakeroot,
 	),
 	TP_fast_assign(
 		__entry->dev = cur->bc_mp->m_super->s_dev;
-		__entry->btnum = cur->bc_btnum;
+		__assign_str(name);
 		__entry->agno = XFS_INO_TO_AGNO(cur->bc_mp,
 					cur->bc_ino.ip->i_ino);
 		__entry->agino = XFS_INO_TO_AGINO(cur->bc_mp,
@@ -4137,9 +4282,9 @@ TRACE_EVENT(xfs_btree_commit_ifakeroot,
 		__entry->blocks = cur->bc_ino.ifake->if_blocks;
 		__entry->whichfork = cur->bc_ino.whichfork;
 	),
-	TP_printk("dev %d:%d btree %s agno 0x%x agino 0x%x whichfork %s levels %u blocks %u",
+	TP_printk("dev %d:%d %sbt agno 0x%x agino 0x%x whichfork %s levels %u blocks %u",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
-		  __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS),
+		  __get_str(name),
 		  __entry->agno,
 		  __entry->agino,
 		  __print_symbolic(__entry->whichfork, XFS_WHICHFORK_STRINGS),
@@ -4156,7 +4301,7 @@ TRACE_EVENT(xfs_btree_bload_level_geometry,
 		blocks_with_extra),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
-		__field(xfs_btnum_t, btnum)
+		__string(name, cur->bc_ops->name)
 		__field(unsigned int, level)
 		__field(unsigned int, nlevels)
 		__field(uint64_t, nr_this_level)
@@ -4167,7 +4312,7 @@ TRACE_EVENT(xfs_btree_bload_level_geometry,
 	),
 	TP_fast_assign(
 		__entry->dev = cur->bc_mp->m_super->s_dev;
-		__entry->btnum = cur->bc_btnum;
+		__assign_str(name);
 		__entry->level = level;
 		__entry->nlevels = cur->bc_nlevels;
 		__entry->nr_this_level = nr_this_level;
@@ -4176,9 +4321,9 @@ TRACE_EVENT(xfs_btree_bload_level_geometry,
 		__entry->blocks = blocks;
 		__entry->blocks_with_extra = blocks_with_extra;
 	),
-	TP_printk("dev %d:%d btree %s level %u/%u nr_this_level %llu nr_per_block %u desired_npb %u blocks %llu blocks_with_extra %llu",
+	TP_printk("dev %d:%d %sbt level %u/%u nr_this_level %llu nr_per_block %u desired_npb %u blocks %llu blocks_with_extra %llu",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
-		  __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS),
+		  __get_str(name),
 		  __entry->level,
 		  __entry->nlevels,
 		  __entry->nr_this_level,
@@ -4195,7 +4340,7 @@ TRACE_EVENT(xfs_btree_bload_block,
 	TP_ARGS(cur, level, block_idx, nr_blocks, ptr, nr_records),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
-		__field(xfs_btnum_t, btnum)
+		__string(name, cur->bc_ops->name)
 		__field(unsigned int, level)
 		__field(unsigned long long, block_idx)
 		__field(unsigned long long, nr_blocks)
@@ -4205,11 +4350,11 @@ TRACE_EVENT(xfs_btree_bload_block,
 	),
 	TP_fast_assign(
 		__entry->dev = cur->bc_mp->m_super->s_dev;
-		__entry->btnum = cur->bc_btnum;
+		__assign_str(name);
 		__entry->level = level;
 		__entry->block_idx = block_idx;
 		__entry->nr_blocks = nr_blocks;
-		if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+		if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) {
 			xfs_fsblock_t	fsb = be64_to_cpu(ptr->l);
 
 			__entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, fsb);
@@ -4220,9 +4365,9 @@ TRACE_EVENT(xfs_btree_bload_block,
 		}
 		__entry->nr_records = nr_records;
 	),
-	TP_printk("dev %d:%d btree %s level %u block %llu/%llu agno 0x%x agbno 0x%x recs %u",
+	TP_printk("dev %d:%d %sbt level %u block %llu/%llu agno 0x%x agbno 0x%x recs %u",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
-		  __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS),
+		  __get_str(name),
 		  __entry->level,
 		  __entry->block_idx,
 		  __entry->nr_blocks,
@@ -4428,7 +4573,7 @@ TRACE_EVENT(xfs_force_shutdown,
 		__entry->dev = mp->m_super->s_dev;
 		__entry->ptag = ptag;
 		__entry->flags = flags;
-		__assign_str(fname, fname);
+		__assign_str(fname);
 		__entry->line_num = line_num;
 	),
 	TP_printk("dev %d:%d tag %s flags %s file %s line_num %d",
@@ -4472,6 +4617,577 @@ DEFINE_PERAG_INTENTS_EVENT(xfs_perag_wait_intents);
 
 #endif /* CONFIG_XFS_DRAIN_INTENTS */
 
+#ifdef CONFIG_XFS_MEMORY_BUFS
+TRACE_EVENT(xmbuf_create,
+	TP_PROTO(struct xfs_buftarg *btp),
+	TP_ARGS(btp),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(unsigned long, ino)
+		__array(char, pathname, 256)
+	),
+	TP_fast_assign(
+		char		pathname[257];
+		char		*path;
+		struct file	*file = btp->bt_file;
+
+		__entry->dev = btp->bt_mount->m_super->s_dev;
+		__entry->ino = file_inode(file)->i_ino;
+		memset(pathname, 0, sizeof(pathname));
+		path = file_path(file, pathname, sizeof(pathname) - 1);
+		if (IS_ERR(path))
+			path = "(unknown)";
+		strncpy(__entry->pathname, path, sizeof(__entry->pathname));
+	),
+	TP_printk("dev %d:%d xmino 0x%lx path '%s'",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->pathname)
+);
+
+TRACE_EVENT(xmbuf_free,
+	TP_PROTO(struct xfs_buftarg *btp),
+	TP_ARGS(btp),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(unsigned long, ino)
+		__field(unsigned long long, bytes)
+		__field(loff_t, size)
+	),
+	TP_fast_assign(
+		struct file	*file = btp->bt_file;
+		struct inode	*inode = file_inode(file);
+
+		__entry->dev = btp->bt_mount->m_super->s_dev;
+		__entry->size = i_size_read(inode);
+		__entry->bytes = (inode->i_blocks << SECTOR_SHIFT) + inode->i_bytes;
+		__entry->ino = inode->i_ino;
+	),
+	TP_printk("dev %d:%d xmino 0x%lx mem_bytes 0x%llx isize 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->bytes,
+		  __entry->size)
+);
+#endif /* CONFIG_XFS_MEMORY_BUFS */
+
+#ifdef CONFIG_XFS_BTREE_IN_MEM
+TRACE_EVENT(xfbtree_init,
+	TP_PROTO(struct xfs_mount *mp, struct xfbtree *xfbt,
+		 const struct xfs_btree_ops *ops),
+	TP_ARGS(mp, xfbt, ops),
+	TP_STRUCT__entry(
+		__field(const void *, btree_ops)
+		__field(unsigned long, xfino)
+		__field(unsigned int, leaf_mxr)
+		__field(unsigned int, leaf_mnr)
+		__field(unsigned int, node_mxr)
+		__field(unsigned int, node_mnr)
+		__field(unsigned long long, owner)
+	),
+	TP_fast_assign(
+		__entry->btree_ops = ops;
+		__entry->xfino = file_inode(xfbt->target->bt_file)->i_ino;
+		__entry->leaf_mxr = xfbt->maxrecs[0];
+		__entry->node_mxr = xfbt->maxrecs[1];
+		__entry->leaf_mnr = xfbt->minrecs[0];
+		__entry->node_mnr = xfbt->minrecs[1];
+		__entry->owner = xfbt->owner;
+	),
+	TP_printk("xfino 0x%lx btree_ops %pS owner 0x%llx leaf_mxr %u leaf_mnr %u node_mxr %u node_mnr %u",
+		  __entry->xfino,
+		  __entry->btree_ops,
+		  __entry->owner,
+		  __entry->leaf_mxr,
+		  __entry->leaf_mnr,
+		  __entry->node_mxr,
+		  __entry->node_mnr)
+);
+
+DECLARE_EVENT_CLASS(xfbtree_buf_class,
+	TP_PROTO(struct xfbtree *xfbt, struct xfs_buf *bp),
+	TP_ARGS(xfbt, bp),
+	TP_STRUCT__entry(
+		__field(unsigned long, xfino)
+		__field(xfs_daddr_t, bno)
+		__field(int, nblks)
+		__field(int, hold)
+		__field(int, pincount)
+		__field(unsigned int, lockval)
+		__field(unsigned int, flags)
+	),
+	TP_fast_assign(
+		__entry->xfino = file_inode(xfbt->target->bt_file)->i_ino;
+		__entry->bno = xfs_buf_daddr(bp);
+		__entry->nblks = bp->b_length;
+		__entry->hold = atomic_read(&bp->b_hold);
+		__entry->pincount = atomic_read(&bp->b_pin_count);
+		__entry->lockval = bp->b_sema.count;
+		__entry->flags = bp->b_flags;
+	),
+	TP_printk("xfino 0x%lx daddr 0x%llx bbcount 0x%x hold %d pincount %d lock %d flags %s",
+		  __entry->xfino,
+		  (unsigned long long)__entry->bno,
+		  __entry->nblks,
+		  __entry->hold,
+		  __entry->pincount,
+		  __entry->lockval,
+		  __print_flags(__entry->flags, "|", XFS_BUF_FLAGS))
+)
+
+#define DEFINE_XFBTREE_BUF_EVENT(name) \
+DEFINE_EVENT(xfbtree_buf_class, name, \
+	TP_PROTO(struct xfbtree *xfbt, struct xfs_buf *bp), \
+	TP_ARGS(xfbt, bp))
+DEFINE_XFBTREE_BUF_EVENT(xfbtree_create_root_buf);
+DEFINE_XFBTREE_BUF_EVENT(xfbtree_trans_commit_buf);
+DEFINE_XFBTREE_BUF_EVENT(xfbtree_trans_cancel_buf);
+
+DECLARE_EVENT_CLASS(xfbtree_freesp_class,
+	TP_PROTO(struct xfbtree *xfbt, struct xfs_btree_cur *cur,
+		 xfs_fileoff_t fileoff),
+	TP_ARGS(xfbt, cur, fileoff),
+	TP_STRUCT__entry(
+		__field(unsigned long, xfino)
+		__string(btname, cur->bc_ops->name)
+		__field(int, nlevels)
+		__field(xfs_fileoff_t, fileoff)
+	),
+	TP_fast_assign(
+		__entry->xfino = file_inode(xfbt->target->bt_file)->i_ino;
+		__assign_str(btname);
+		__entry->nlevels = cur->bc_nlevels;
+		__entry->fileoff = fileoff;
+	),
+	TP_printk("xfino 0x%lx %sbt nlevels %d fileoff 0x%llx",
+		  __entry->xfino,
+		  __get_str(btname),
+		  __entry->nlevels,
+		  (unsigned long long)__entry->fileoff)
+)
+
+#define DEFINE_XFBTREE_FREESP_EVENT(name) \
+DEFINE_EVENT(xfbtree_freesp_class, name, \
+	TP_PROTO(struct xfbtree *xfbt, struct xfs_btree_cur *cur, \
+		 xfs_fileoff_t fileoff), \
+	TP_ARGS(xfbt, cur, fileoff))
+DEFINE_XFBTREE_FREESP_EVENT(xfbtree_alloc_block);
+DEFINE_XFBTREE_FREESP_EVENT(xfbtree_free_block);
+#endif /* CONFIG_XFS_BTREE_IN_MEM */
+
+/* exchmaps tracepoints */
+#define XFS_EXCHMAPS_STRINGS \
+	{ XFS_EXCHMAPS_ATTR_FORK,		"ATTRFORK" }, \
+	{ XFS_EXCHMAPS_SET_SIZES,		"SETSIZES" }, \
+	{ XFS_EXCHMAPS_INO1_WRITTEN,		"INO1_WRITTEN" }, \
+	{ XFS_EXCHMAPS_CLEAR_INO1_REFLINK,	"CLEAR_INO1_REFLINK" }, \
+	{ XFS_EXCHMAPS_CLEAR_INO2_REFLINK,	"CLEAR_INO2_REFLINK" }, \
+	{ __XFS_EXCHMAPS_INO2_SHORTFORM,	"INO2_SF" }
+
+DEFINE_INODE_IREC_EVENT(xfs_exchmaps_mapping1_skip);
+DEFINE_INODE_IREC_EVENT(xfs_exchmaps_mapping1);
+DEFINE_INODE_IREC_EVENT(xfs_exchmaps_mapping2);
+DEFINE_ITRUNC_EVENT(xfs_exchmaps_update_inode_size);
+
+#define XFS_EXCHRANGE_INODES \
+	{ 1,	"file1" }, \
+	{ 2,	"file2" }
+
+DECLARE_EVENT_CLASS(xfs_exchrange_inode_class,
+	TP_PROTO(struct xfs_inode *ip, int whichfile),
+	TP_ARGS(ip, whichfile),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(int, whichfile)
+		__field(xfs_ino_t, ino)
+		__field(int, format)
+		__field(xfs_extnum_t, nex)
+		__field(int, broot_size)
+		__field(int, fork_off)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ip)->i_sb->s_dev;
+		__entry->whichfile = whichfile;
+		__entry->ino = ip->i_ino;
+		__entry->format = ip->i_df.if_format;
+		__entry->nex = ip->i_df.if_nextents;
+		__entry->fork_off = xfs_inode_fork_boff(ip);
+	),
+	TP_printk("dev %d:%d ino 0x%llx whichfile %s format %s num_extents %llu forkoff 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __print_symbolic(__entry->whichfile, XFS_EXCHRANGE_INODES),
+		  __print_symbolic(__entry->format, XFS_INODE_FORMAT_STR),
+		  __entry->nex,
+		  __entry->fork_off)
+)
+
+#define DEFINE_EXCHRANGE_INODE_EVENT(name) \
+DEFINE_EVENT(xfs_exchrange_inode_class, name, \
+	TP_PROTO(struct xfs_inode *ip, int whichfile), \
+	TP_ARGS(ip, whichfile))
+
+DEFINE_EXCHRANGE_INODE_EVENT(xfs_exchrange_before);
+DEFINE_EXCHRANGE_INODE_EVENT(xfs_exchrange_after);
+DEFINE_INODE_ERROR_EVENT(xfs_exchrange_error);
+
+#define XFS_EXCHANGE_RANGE_FLAGS_STRS \
+	{ XFS_EXCHANGE_RANGE_TO_EOF,		"TO_EOF" }, \
+	{ XFS_EXCHANGE_RANGE_DSYNC	,	"DSYNC" }, \
+	{ XFS_EXCHANGE_RANGE_DRY_RUN,		"DRY_RUN" }, \
+	{ XFS_EXCHANGE_RANGE_FILE1_WRITTEN,	"F1_WRITTEN" }, \
+	{ __XFS_EXCHANGE_RANGE_UPD_CMTIME1,	"CMTIME1" }, \
+	{ __XFS_EXCHANGE_RANGE_UPD_CMTIME2,	"CMTIME2" }
+
+/* file exchange-range tracepoint class */
+DECLARE_EVENT_CLASS(xfs_exchrange_class,
+	TP_PROTO(const struct xfs_exchrange *fxr, struct xfs_inode *ip1,
+		 struct xfs_inode *ip2),
+	TP_ARGS(fxr, ip1, ip2),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ip1_ino)
+		__field(loff_t, ip1_isize)
+		__field(loff_t, ip1_disize)
+		__field(xfs_ino_t, ip2_ino)
+		__field(loff_t, ip2_isize)
+		__field(loff_t, ip2_disize)
+
+		__field(loff_t, file1_offset)
+		__field(loff_t, file2_offset)
+		__field(unsigned long long, length)
+		__field(unsigned long long, flags)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ip1)->i_sb->s_dev;
+		__entry->ip1_ino = ip1->i_ino;
+		__entry->ip1_isize = VFS_I(ip1)->i_size;
+		__entry->ip1_disize = ip1->i_disk_size;
+		__entry->ip2_ino = ip2->i_ino;
+		__entry->ip2_isize = VFS_I(ip2)->i_size;
+		__entry->ip2_disize = ip2->i_disk_size;
+
+		__entry->file1_offset = fxr->file1_offset;
+		__entry->file2_offset = fxr->file2_offset;
+		__entry->length = fxr->length;
+		__entry->flags = fxr->flags;
+	),
+	TP_printk("dev %d:%d flags %s bytecount 0x%llx "
+		  "ino1 0x%llx isize 0x%llx disize 0x%llx pos 0x%llx -> "
+		  "ino2 0x%llx isize 0x%llx disize 0x%llx pos 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		   __print_flags_u64(__entry->flags, "|", XFS_EXCHANGE_RANGE_FLAGS_STRS),
+		  __entry->length,
+		  __entry->ip1_ino,
+		  __entry->ip1_isize,
+		  __entry->ip1_disize,
+		  __entry->file1_offset,
+		  __entry->ip2_ino,
+		  __entry->ip2_isize,
+		  __entry->ip2_disize,
+		  __entry->file2_offset)
+)
+
+#define DEFINE_EXCHRANGE_EVENT(name)	\
+DEFINE_EVENT(xfs_exchrange_class, name,	\
+	TP_PROTO(const struct xfs_exchrange *fxr, struct xfs_inode *ip1, \
+		 struct xfs_inode *ip2), \
+	TP_ARGS(fxr, ip1, ip2))
+DEFINE_EXCHRANGE_EVENT(xfs_exchrange_prep);
+DEFINE_EXCHRANGE_EVENT(xfs_exchrange_flush);
+DEFINE_EXCHRANGE_EVENT(xfs_exchrange_mappings);
+
+TRACE_EVENT(xfs_exchmaps_overhead,
+	TP_PROTO(struct xfs_mount *mp, unsigned long long bmbt_blocks,
+		 unsigned long long rmapbt_blocks),
+	TP_ARGS(mp, bmbt_blocks, rmapbt_blocks),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(unsigned long long, bmbt_blocks)
+		__field(unsigned long long, rmapbt_blocks)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->bmbt_blocks = bmbt_blocks;
+		__entry->rmapbt_blocks = rmapbt_blocks;
+	),
+	TP_printk("dev %d:%d bmbt_blocks 0x%llx rmapbt_blocks 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->bmbt_blocks,
+		  __entry->rmapbt_blocks)
+);
+
+DECLARE_EVENT_CLASS(xfs_exchmaps_estimate_class,
+	TP_PROTO(const struct xfs_exchmaps_req *req),
+	TP_ARGS(req),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino1)
+		__field(xfs_ino_t, ino2)
+		__field(xfs_fileoff_t, startoff1)
+		__field(xfs_fileoff_t, startoff2)
+		__field(xfs_filblks_t, blockcount)
+		__field(uint64_t, flags)
+		__field(xfs_filblks_t, ip1_bcount)
+		__field(xfs_filblks_t, ip2_bcount)
+		__field(xfs_filblks_t, ip1_rtbcount)
+		__field(xfs_filblks_t, ip2_rtbcount)
+		__field(unsigned long long, resblks)
+		__field(unsigned long long, nr_exchanges)
+	),
+	TP_fast_assign(
+		__entry->dev = req->ip1->i_mount->m_super->s_dev;
+		__entry->ino1 = req->ip1->i_ino;
+		__entry->ino2 = req->ip2->i_ino;
+		__entry->startoff1 = req->startoff1;
+		__entry->startoff2 = req->startoff2;
+		__entry->blockcount = req->blockcount;
+		__entry->flags = req->flags;
+		__entry->ip1_bcount = req->ip1_bcount;
+		__entry->ip2_bcount = req->ip2_bcount;
+		__entry->ip1_rtbcount = req->ip1_rtbcount;
+		__entry->ip2_rtbcount = req->ip2_rtbcount;
+		__entry->resblks = req->resblks;
+		__entry->nr_exchanges = req->nr_exchanges;
+	),
+	TP_printk("dev %d:%d ino1 0x%llx fileoff1 0x%llx ino2 0x%llx fileoff2 0x%llx fsbcount 0x%llx flags (%s) bcount1 0x%llx rtbcount1 0x%llx bcount2 0x%llx rtbcount2 0x%llx resblks 0x%llx nr_exchanges %llu",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino1, __entry->startoff1,
+		  __entry->ino2, __entry->startoff2,
+		  __entry->blockcount,
+		  __print_flags_u64(__entry->flags, "|", XFS_EXCHMAPS_STRINGS),
+		  __entry->ip1_bcount,
+		  __entry->ip1_rtbcount,
+		  __entry->ip2_bcount,
+		  __entry->ip2_rtbcount,
+		  __entry->resblks,
+		  __entry->nr_exchanges)
+);
+
+#define DEFINE_EXCHMAPS_ESTIMATE_EVENT(name)	\
+DEFINE_EVENT(xfs_exchmaps_estimate_class, name,	\
+	TP_PROTO(const struct xfs_exchmaps_req *req), \
+	TP_ARGS(req))
+DEFINE_EXCHMAPS_ESTIMATE_EVENT(xfs_exchmaps_initial_estimate);
+DEFINE_EXCHMAPS_ESTIMATE_EVENT(xfs_exchmaps_final_estimate);
+
+DECLARE_EVENT_CLASS(xfs_exchmaps_intent_class,
+	TP_PROTO(struct xfs_mount *mp, const struct xfs_exchmaps_intent *xmi),
+	TP_ARGS(mp, xmi),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino1)
+		__field(xfs_ino_t, ino2)
+		__field(uint64_t, flags)
+		__field(xfs_fileoff_t, startoff1)
+		__field(xfs_fileoff_t, startoff2)
+		__field(xfs_filblks_t, blockcount)
+		__field(xfs_fsize_t, isize1)
+		__field(xfs_fsize_t, isize2)
+		__field(xfs_fsize_t, new_isize1)
+		__field(xfs_fsize_t, new_isize2)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->ino1 = xmi->xmi_ip1->i_ino;
+		__entry->ino2 = xmi->xmi_ip2->i_ino;
+		__entry->flags = xmi->xmi_flags;
+		__entry->startoff1 = xmi->xmi_startoff1;
+		__entry->startoff2 = xmi->xmi_startoff2;
+		__entry->blockcount = xmi->xmi_blockcount;
+		__entry->isize1 = xmi->xmi_ip1->i_disk_size;
+		__entry->isize2 = xmi->xmi_ip2->i_disk_size;
+		__entry->new_isize1 = xmi->xmi_isize1;
+		__entry->new_isize2 = xmi->xmi_isize2;
+	),
+	TP_printk("dev %d:%d ino1 0x%llx fileoff1 0x%llx ino2 0x%llx fileoff2 0x%llx fsbcount 0x%llx flags (%s) isize1 0x%llx newisize1 0x%llx isize2 0x%llx newisize2 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino1, __entry->startoff1,
+		  __entry->ino2, __entry->startoff2,
+		  __entry->blockcount,
+		  __print_flags_u64(__entry->flags, "|", XFS_EXCHMAPS_STRINGS),
+		  __entry->isize1, __entry->new_isize1,
+		  __entry->isize2, __entry->new_isize2)
+);
+
+#define DEFINE_EXCHMAPS_INTENT_EVENT(name)	\
+DEFINE_EVENT(xfs_exchmaps_intent_class, name,	\
+	TP_PROTO(struct xfs_mount *mp, const struct xfs_exchmaps_intent *xmi), \
+	TP_ARGS(mp, xmi))
+DEFINE_EXCHMAPS_INTENT_EVENT(xfs_exchmaps_defer);
+DEFINE_EXCHMAPS_INTENT_EVENT(xfs_exchmaps_recover);
+
+TRACE_EVENT(xfs_exchmaps_delta_nextents_step,
+	TP_PROTO(struct xfs_mount *mp,
+		 const struct xfs_bmbt_irec *left,
+		 const struct xfs_bmbt_irec *curr,
+		 const struct xfs_bmbt_irec *new,
+		 const struct xfs_bmbt_irec *right,
+		 int delta, unsigned int state),
+	TP_ARGS(mp, left, curr, new, right, delta, state),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_fileoff_t, loff)
+		__field(xfs_fsblock_t, lstart)
+		__field(xfs_filblks_t, lcount)
+		__field(xfs_fileoff_t, coff)
+		__field(xfs_fsblock_t, cstart)
+		__field(xfs_filblks_t, ccount)
+		__field(xfs_fileoff_t, noff)
+		__field(xfs_fsblock_t, nstart)
+		__field(xfs_filblks_t, ncount)
+		__field(xfs_fileoff_t, roff)
+		__field(xfs_fsblock_t, rstart)
+		__field(xfs_filblks_t, rcount)
+		__field(int, delta)
+		__field(unsigned int, state)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->loff = left->br_startoff;
+		__entry->lstart = left->br_startblock;
+		__entry->lcount = left->br_blockcount;
+		__entry->coff = curr->br_startoff;
+		__entry->cstart = curr->br_startblock;
+		__entry->ccount = curr->br_blockcount;
+		__entry->noff = new->br_startoff;
+		__entry->nstart = new->br_startblock;
+		__entry->ncount = new->br_blockcount;
+		__entry->roff = right->br_startoff;
+		__entry->rstart = right->br_startblock;
+		__entry->rcount = right->br_blockcount;
+		__entry->delta = delta;
+		__entry->state = state;
+	),
+	TP_printk("dev %d:%d left 0x%llx:0x%llx:0x%llx; curr 0x%llx:0x%llx:0x%llx <- new 0x%llx:0x%llx:0x%llx; right 0x%llx:0x%llx:0x%llx delta %d state 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		__entry->loff, __entry->lstart, __entry->lcount,
+		__entry->coff, __entry->cstart, __entry->ccount,
+		__entry->noff, __entry->nstart, __entry->ncount,
+		__entry->roff, __entry->rstart, __entry->rcount,
+		__entry->delta, __entry->state)
+);
+
+TRACE_EVENT(xfs_exchmaps_delta_nextents,
+	TP_PROTO(const struct xfs_exchmaps_req *req, int64_t d_nexts1,
+		 int64_t d_nexts2),
+	TP_ARGS(req, d_nexts1, d_nexts2),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino1)
+		__field(xfs_ino_t, ino2)
+		__field(xfs_extnum_t, nexts1)
+		__field(xfs_extnum_t, nexts2)
+		__field(int64_t, d_nexts1)
+		__field(int64_t, d_nexts2)
+	),
+	TP_fast_assign(
+		int whichfork = xfs_exchmaps_reqfork(req);
+
+		__entry->dev = req->ip1->i_mount->m_super->s_dev;
+		__entry->ino1 = req->ip1->i_ino;
+		__entry->ino2 = req->ip2->i_ino;
+		__entry->nexts1 = xfs_ifork_ptr(req->ip1, whichfork)->if_nextents;
+		__entry->nexts2 = xfs_ifork_ptr(req->ip2, whichfork)->if_nextents;
+		__entry->d_nexts1 = d_nexts1;
+		__entry->d_nexts2 = d_nexts2;
+	),
+	TP_printk("dev %d:%d ino1 0x%llx nexts %llu ino2 0x%llx nexts %llu delta1 %lld delta2 %lld",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino1, __entry->nexts1,
+		  __entry->ino2, __entry->nexts2,
+		  __entry->d_nexts1, __entry->d_nexts2)
+);
+
+DECLARE_EVENT_CLASS(xfs_getparents_rec_class,
+	TP_PROTO(struct xfs_inode *ip, const struct xfs_getparents *ppi,
+		 const struct xfs_attr_list_context *context,
+	         const struct xfs_getparents_rec *pptr),
+	TP_ARGS(ip, ppi, context, pptr),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(unsigned int, firstu)
+		__field(unsigned short, reclen)
+		__field(unsigned int, bufsize)
+		__field(xfs_ino_t, parent_ino)
+		__field(unsigned int, parent_gen)
+		__string(name, pptr->gpr_name)
+	),
+	TP_fast_assign(
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->firstu = context->firstu;
+		__entry->reclen = pptr->gpr_reclen;
+		__entry->bufsize = ppi->gp_bufsize;
+		__entry->parent_ino = pptr->gpr_parent.ha_fid.fid_ino;
+		__entry->parent_gen = pptr->gpr_parent.ha_fid.fid_gen;
+		__assign_str(name);
+	),
+	TP_printk("dev %d:%d ino 0x%llx firstu %u reclen %u bufsize %u parent_ino 0x%llx parent_gen 0x%x name '%s'",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->firstu,
+		  __entry->reclen,
+		  __entry->bufsize,
+		  __entry->parent_ino,
+		  __entry->parent_gen,
+		  __get_str(name))
+)
+#define DEFINE_XFS_GETPARENTS_REC_EVENT(name) \
+DEFINE_EVENT(xfs_getparents_rec_class, name, \
+	TP_PROTO(struct xfs_inode *ip, const struct xfs_getparents *ppi, \
+		 const struct xfs_attr_list_context *context, \
+	         const struct xfs_getparents_rec *pptr), \
+	TP_ARGS(ip, ppi, context, pptr))
+DEFINE_XFS_GETPARENTS_REC_EVENT(xfs_getparents_put_listent);
+DEFINE_XFS_GETPARENTS_REC_EVENT(xfs_getparents_expand_lastrec);
+
+DECLARE_EVENT_CLASS(xfs_getparents_class,
+	TP_PROTO(struct xfs_inode *ip, const struct xfs_getparents *ppi,
+		 const struct xfs_attrlist_cursor_kern *cur),
+	TP_ARGS(ip, ppi, cur),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(unsigned short, iflags)
+		__field(unsigned short, oflags)
+		__field(unsigned int, bufsize)
+		__field(unsigned int, hashval)
+		__field(unsigned int, blkno)
+		__field(unsigned int, offset)
+		__field(int, initted)
+	),
+	TP_fast_assign(
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->iflags = ppi->gp_iflags;
+		__entry->oflags = ppi->gp_oflags;
+		__entry->bufsize = ppi->gp_bufsize;
+		__entry->hashval = cur->hashval;
+		__entry->blkno = cur->blkno;
+		__entry->offset = cur->offset;
+		__entry->initted = cur->initted;
+	),
+	TP_printk("dev %d:%d ino 0x%llx iflags 0x%x oflags 0x%x bufsize %u cur_init? %d hashval 0x%x blkno %u offset %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->iflags,
+		  __entry->oflags,
+		  __entry->bufsize,
+		  __entry->initted,
+		  __entry->hashval,
+		  __entry->blkno,
+		  __entry->offset)
+)
+#define DEFINE_XFS_GETPARENTS_EVENT(name) \
+DEFINE_EVENT(xfs_getparents_class, name, \
+	TP_PROTO(struct xfs_inode *ip, const struct xfs_getparents *ppi, \
+		 const struct xfs_attrlist_cursor_kern *cur), \
+	TP_ARGS(ip, ppi, cur))
+DEFINE_XFS_GETPARENTS_EVENT(xfs_getparents_begin);
+DEFINE_XFS_GETPARENTS_EVENT(xfs_getparents_end);
+
 #endif /* _TRACE_XFS_H */
 
 #undef TRACE_INCLUDE_PATH
diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
index 12d45e93f0..828da4ac43 100644
--- a/fs/xfs/xfs_trans.c
+++ b/fs/xfs/xfs_trans.c
@@ -163,7 +163,7 @@ xfs_trans_reserve(
 	 * fail if the count would go below zero.
 	 */
 	if (blocks > 0) {
-		error = xfs_mod_fdblocks(mp, -((int64_t)blocks), rsvd);
+		error = xfs_dec_fdblocks(mp, blocks, rsvd);
 		if (error != 0)
 			return -ENOSPC;
 		tp->t_blk_res += blocks;
@@ -210,7 +210,7 @@ xfs_trans_reserve(
 	 * fail if the count would go below zero.
 	 */
 	if (rtextents > 0) {
-		error = xfs_mod_frextents(mp, -((int64_t)rtextents));
+		error = xfs_dec_frextents(mp, rtextents);
 		if (error) {
 			error = -ENOSPC;
 			goto undo_log;
@@ -234,7 +234,7 @@ undo_log:
 
 undo_blocks:
 	if (blocks > 0) {
-		xfs_mod_fdblocks(mp, (int64_t)blocks, rsvd);
+		xfs_add_fdblocks(mp, blocks);
 		tp->t_blk_res = 0;
 	}
 	return error;
@@ -593,38 +593,44 @@ xfs_trans_unreserve_and_mod_sb(
 	struct xfs_trans	*tp)
 {
 	struct xfs_mount	*mp = tp->t_mountp;
-	bool			rsvd = (tp->t_flags & XFS_TRANS_RESERVE) != 0;
-	int64_t			blkdelta = 0;
-	int64_t			rtxdelta = 0;
+	int64_t			blkdelta = tp->t_blk_res;
+	int64_t			rtxdelta = tp->t_rtx_res;
 	int64_t			idelta = 0;
 	int64_t			ifreedelta = 0;
-	int			error;
 
-	/* calculate deltas */
-	if (tp->t_blk_res > 0)
-		blkdelta = tp->t_blk_res;
-	if ((tp->t_fdblocks_delta != 0) &&
-	    (xfs_has_lazysbcount(mp) ||
-	     (tp->t_flags & XFS_TRANS_SB_DIRTY)))
+	/*
+	 * Calculate the deltas.
+	 *
+	 * t_fdblocks_delta and t_frextents_delta can be positive or negative:
+	 *
+	 *  - positive values indicate blocks freed in the transaction.
+	 *  - negative values indicate blocks allocated in the transaction
+	 *
+	 * Negative values can only happen if the transaction has a block
+	 * reservation that covers the allocated block.  The end result is
+	 * that the calculated delta values must always be positive and we
+	 * can only put back previous allocated or reserved blocks here.
+	 */
+	ASSERT(tp->t_blk_res || tp->t_fdblocks_delta >= 0);
+	if (xfs_has_lazysbcount(mp) || (tp->t_flags & XFS_TRANS_SB_DIRTY)) {
 	        blkdelta += tp->t_fdblocks_delta;
+		ASSERT(blkdelta >= 0);
+	}
 
-	if (tp->t_rtx_res > 0)
-		rtxdelta = tp->t_rtx_res;
-	if ((tp->t_frextents_delta != 0) &&
-	    (tp->t_flags & XFS_TRANS_SB_DIRTY))
+	ASSERT(tp->t_rtx_res || tp->t_frextents_delta >= 0);
+	if (tp->t_flags & XFS_TRANS_SB_DIRTY) {
 		rtxdelta += tp->t_frextents_delta;
+		ASSERT(rtxdelta >= 0);
+	}
 
-	if (xfs_has_lazysbcount(mp) ||
-	     (tp->t_flags & XFS_TRANS_SB_DIRTY)) {
+	if (xfs_has_lazysbcount(mp) || (tp->t_flags & XFS_TRANS_SB_DIRTY)) {
 		idelta = tp->t_icount_delta;
 		ifreedelta = tp->t_ifree_delta;
 	}
 
 	/* apply the per-cpu counters */
-	if (blkdelta) {
-		error = xfs_mod_fdblocks(mp, blkdelta, rsvd);
-		ASSERT(!error);
-	}
+	if (blkdelta)
+		xfs_add_fdblocks(mp, blkdelta);
 
 	if (idelta)
 		percpu_counter_add_batch(&mp->m_icount, idelta,
@@ -633,10 +639,8 @@ xfs_trans_unreserve_and_mod_sb(
 	if (ifreedelta)
 		percpu_counter_add(&mp->m_ifree, ifreedelta);
 
-	if (rtxdelta) {
-		error = xfs_mod_frextents(mp, rtxdelta);
-		ASSERT(!error);
-	}
+	if (rtxdelta)
+		xfs_add_frextents(mp, rtxdelta);
 
 	if (!(tp->t_flags & XFS_TRANS_SB_DIRTY))
 		return;
@@ -672,7 +676,6 @@ xfs_trans_unreserve_and_mod_sb(
 	 */
 	ASSERT(mp->m_sb.sb_imax_pct >= 0);
 	ASSERT(mp->m_sb.sb_rextslog >= 0);
-	return;
 }
 
 /* Add the given log item to the transaction's list of log items. */
@@ -1273,7 +1276,7 @@ xfs_trans_reserve_more_inode(
 	unsigned int		rtx = xfs_extlen_to_rtxlen(mp, rblocks);
 	int			error;
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 
 	error = xfs_trans_reserve(tp, &resv, dblocks, rtx);
 	if (error)
@@ -1291,9 +1294,9 @@ xfs_trans_reserve_more_inode(
 		return 0;
 
 	/* Quota failed, give back the new reservation. */
-	xfs_mod_fdblocks(mp, dblocks, tp->t_flags & XFS_TRANS_RESERVE);
+	xfs_add_fdblocks(mp, dblocks);
 	tp->t_blk_res -= dblocks;
-	xfs_mod_frextents(mp, rtx);
+	xfs_add_frextents(mp, rtx);
 	tp->t_rtx_res -= rtx;
 	return error;
 }
@@ -1430,6 +1433,8 @@ out_cancel:
  * The caller must ensure that the on-disk dquots attached to this inode have
  * already been allocated and initialized.  The ILOCKs will be dropped when the
  * transaction is committed or cancelled.
+ *
+ * Caller is responsible for unlocking the inodes manually upon return
  */
 int
 xfs_trans_alloc_dir(
@@ -1460,8 +1465,8 @@ retry:
 
 	xfs_lock_two_inodes(dp, XFS_ILOCK_EXCL, ip, XFS_ILOCK_EXCL);
 
-	xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
-	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
+	xfs_trans_ijoin(tp, dp, 0);
+	xfs_trans_ijoin(tp, ip, 0);
 
 	error = xfs_qm_dqattach_locked(dp, false);
 	if (error) {
@@ -1484,6 +1489,9 @@ retry:
 	if (error == -EDQUOT || error == -ENOSPC) {
 		if (!retried) {
 			xfs_trans_cancel(tp);
+			xfs_iunlock(dp, XFS_ILOCK_EXCL);
+			if (dp != ip)
+				xfs_iunlock(ip, XFS_ILOCK_EXCL);
 			xfs_blockgc_free_quota(dp, 0);
 			retried = true;
 			goto retry;
diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h
index 08ce757c74..1636663707 100644
--- a/fs/xfs/xfs_trans.h
+++ b/fs/xfs/xfs_trans.h
@@ -215,6 +215,7 @@ struct xfs_buf	*xfs_trans_getsb(struct xfs_trans *);
 
 void		xfs_trans_brelse(xfs_trans_t *, struct xfs_buf *);
 void		xfs_trans_bjoin(xfs_trans_t *, struct xfs_buf *);
+void		xfs_trans_bdetach(struct xfs_trans *tp, struct xfs_buf *bp);
 void		xfs_trans_bhold(xfs_trans_t *, struct xfs_buf *);
 void		xfs_trans_bhold_release(xfs_trans_t *, struct xfs_buf *);
 void		xfs_trans_binval(xfs_trans_t *, struct xfs_buf *);
@@ -267,19 +268,14 @@ static inline void
 xfs_trans_set_context(
 	struct xfs_trans	*tp)
 {
-	ASSERT(current->journal_info == NULL);
 	tp->t_pflags = memalloc_nofs_save();
-	current->journal_info = tp;
 }
 
 static inline void
 xfs_trans_clear_context(
 	struct xfs_trans	*tp)
 {
-	if (current->journal_info == tp) {
-		memalloc_nofs_restore(tp->t_pflags);
-		current->journal_info = NULL;
-	}
+	memalloc_nofs_restore(tp->t_pflags);
 }
 
 static inline void
@@ -287,10 +283,8 @@ xfs_trans_switch_context(
 	struct xfs_trans	*old_tp,
 	struct xfs_trans	*new_tp)
 {
-	ASSERT(current->journal_info == old_tp);
 	new_tp->t_pflags = old_tp->t_pflags;
 	old_tp->t_pflags = 0;
-	current->journal_info = new_tp;
 }
 
 #endif	/* __XFS_TRANS_H__ */
diff --git a/fs/xfs/xfs_trans_ail.c b/fs/xfs/xfs_trans_ail.c
index 1098452e7f..e4c343096f 100644
--- a/fs/xfs/xfs_trans_ail.c
+++ b/fs/xfs/xfs_trans_ail.c
@@ -901,7 +901,8 @@ xfs_trans_ail_init(
 {
 	struct xfs_ail	*ailp;
 
-	ailp = kmem_zalloc(sizeof(struct xfs_ail), KM_MAYFAIL);
+	ailp = kzalloc(sizeof(struct xfs_ail),
+			GFP_KERNEL | __GFP_RETRY_MAYFAIL);
 	if (!ailp)
 		return -ENOMEM;
 
@@ -921,7 +922,7 @@ xfs_trans_ail_init(
 	return 0;
 
 out_free_ailp:
-	kmem_free(ailp);
+	kfree(ailp);
 	return -ENOMEM;
 }
 
@@ -932,5 +933,5 @@ xfs_trans_ail_destroy(
 	struct xfs_ail	*ailp = mp->m_ail;
 
 	kthread_stop(ailp->ail_task);
-	kmem_free(ailp);
+	kfree(ailp);
 }
diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c
index 6549e50d85..e28ab74af4 100644
--- a/fs/xfs/xfs_trans_buf.c
+++ b/fs/xfs/xfs_trans_buf.c
@@ -393,6 +393,48 @@ xfs_trans_brelse(
 }
 
 /*
+ * Forcibly detach a buffer previously joined to the transaction.  The caller
+ * will retain its locked reference to the buffer after this function returns.
+ * The buffer must be completely clean and must not be held to the transaction.
+ */
+void
+xfs_trans_bdetach(
+	struct xfs_trans	*tp,
+	struct xfs_buf		*bp)
+{
+	struct xfs_buf_log_item	*bip = bp->b_log_item;
+
+	ASSERT(tp != NULL);
+	ASSERT(bp->b_transp == tp);
+	ASSERT(bip->bli_item.li_type == XFS_LI_BUF);
+	ASSERT(atomic_read(&bip->bli_refcount) > 0);
+
+	trace_xfs_trans_bdetach(bip);
+
+	/*
+	 * Erase all recursion count, since we're removing this buffer from the
+	 * transaction.
+	 */
+	bip->bli_recur = 0;
+
+	/*
+	 * The buffer must be completely clean.  Specifically, it had better
+	 * not be dirty, stale, logged, ordered, or held to the transaction.
+	 */
+	ASSERT(!test_bit(XFS_LI_DIRTY, &bip->bli_item.li_flags));
+	ASSERT(!(bip->bli_flags & XFS_BLI_DIRTY));
+	ASSERT(!(bip->bli_flags & XFS_BLI_HOLD));
+	ASSERT(!(bip->bli_flags & XFS_BLI_LOGGED));
+	ASSERT(!(bip->bli_flags & XFS_BLI_ORDERED));
+	ASSERT(!(bip->bli_flags & XFS_BLI_STALE));
+
+	/* Unlink the log item from the transaction and drop the log item. */
+	xfs_trans_del_item(&bip->bli_item);
+	xfs_buf_item_put(bip);
+	bp->b_transp = NULL;
+}
+
+/*
  * Mark the buffer as not needing to be unlocked when the buf item's
  * iop_committing() routine is called.  The buffer must already be locked
  * and associated with the given transaction.
diff --git a/fs/xfs/xfs_trans_dquot.c b/fs/xfs/xfs_trans_dquot.c
index aa00cf67ad..b368e13424 100644
--- a/fs/xfs/xfs_trans_dquot.c
+++ b/fs/xfs/xfs_trans_dquot.c
@@ -17,6 +17,7 @@
 #include "xfs_qm.h"
 #include "xfs_trace.h"
 #include "xfs_error.h"
+#include "xfs_health.h"
 
 STATIC void	xfs_trans_alloc_dqinfo(xfs_trans_t *);
 
@@ -120,6 +121,116 @@ xfs_trans_dup_dqinfo(
 	}
 }
 
+#ifdef CONFIG_XFS_LIVE_HOOKS
+/*
+ * Use a static key here to reduce the overhead of quota live updates.  If the
+ * compiler supports jump labels, the static branch will be replaced by a nop
+ * sled when there are no hook users.  Online fsck is currently the only
+ * caller, so this is a reasonable tradeoff.
+ *
+ * Note: Patching the kernel code requires taking the cpu hotplug lock.  Other
+ * parts of the kernel allocate memory with that lock held, which means that
+ * XFS callers cannot hold any locks that might be used by memory reclaim or
+ * writeback when calling the static_branch_{inc,dec} functions.
+ */
+DEFINE_STATIC_XFS_HOOK_SWITCH(xfs_dqtrx_hooks_switch);
+
+void
+xfs_dqtrx_hook_disable(void)
+{
+	xfs_hooks_switch_off(&xfs_dqtrx_hooks_switch);
+}
+
+void
+xfs_dqtrx_hook_enable(void)
+{
+	xfs_hooks_switch_on(&xfs_dqtrx_hooks_switch);
+}
+
+/* Schedule a transactional dquot update on behalf of an inode. */
+void
+xfs_trans_mod_ino_dquot(
+	struct xfs_trans		*tp,
+	struct xfs_inode		*ip,
+	struct xfs_dquot		*dqp,
+	unsigned int			field,
+	int64_t				delta)
+{
+	xfs_trans_mod_dquot(tp, dqp, field, delta);
+
+	if (xfs_hooks_switched_on(&xfs_dqtrx_hooks_switch)) {
+		struct xfs_mod_ino_dqtrx_params	p = {
+			.tx_id		= (uintptr_t)tp,
+			.ino		= ip->i_ino,
+			.q_type		= xfs_dquot_type(dqp),
+			.q_id		= dqp->q_id,
+			.delta		= delta
+		};
+		struct xfs_quotainfo	*qi = tp->t_mountp->m_quotainfo;
+
+		xfs_hooks_call(&qi->qi_mod_ino_dqtrx_hooks, field, &p);
+	}
+}
+
+/* Call the specified functions during a dquot counter update. */
+int
+xfs_dqtrx_hook_add(
+	struct xfs_quotainfo	*qi,
+	struct xfs_dqtrx_hook	*hook)
+{
+	int			error;
+
+	/*
+	 * Transactional dquot updates first call the mod hook when changes
+	 * are attached to the transaction and then call the apply hook when
+	 * those changes are committed (or canceled).
+	 *
+	 * The apply hook must be installed before the mod hook so that we
+	 * never fail to catch the end of a quota update sequence.
+	 */
+	error = xfs_hooks_add(&qi->qi_apply_dqtrx_hooks, &hook->apply_hook);
+	if (error)
+		goto out;
+
+	error = xfs_hooks_add(&qi->qi_mod_ino_dqtrx_hooks, &hook->mod_hook);
+	if (error)
+		goto out_apply;
+
+	return 0;
+
+out_apply:
+	xfs_hooks_del(&qi->qi_apply_dqtrx_hooks, &hook->apply_hook);
+out:
+	return error;
+}
+
+/* Stop calling the specified function during a dquot counter update. */
+void
+xfs_dqtrx_hook_del(
+	struct xfs_quotainfo	*qi,
+	struct xfs_dqtrx_hook	*hook)
+{
+	/*
+	 * The mod hook must be removed before apply hook to avoid giving the
+	 * hook consumer with an incomplete update.  No hooks should be running
+	 * after these functions return.
+	 */
+	xfs_hooks_del(&qi->qi_mod_ino_dqtrx_hooks, &hook->mod_hook);
+	xfs_hooks_del(&qi->qi_apply_dqtrx_hooks, &hook->apply_hook);
+}
+
+/* Configure dquot update hook functions. */
+void
+xfs_dqtrx_hook_setup(
+	struct xfs_dqtrx_hook	*hook,
+	notifier_fn_t		mod_fn,
+	notifier_fn_t		apply_fn)
+{
+	xfs_hook_setup(&hook->mod_hook, mod_fn);
+	xfs_hook_setup(&hook->apply_hook, apply_fn);
+}
+#endif /* CONFIG_XFS_LIVE_HOOKS */
+
 /*
  * Wrap around mod_dquot to account for both user and group quotas.
  */
@@ -137,11 +248,11 @@ xfs_trans_mod_dquot_byino(
 		return;
 
 	if (XFS_IS_UQUOTA_ON(mp) && ip->i_udquot)
-		(void) xfs_trans_mod_dquot(tp, ip->i_udquot, field, delta);
+		xfs_trans_mod_ino_dquot(tp, ip, ip->i_udquot, field, delta);
 	if (XFS_IS_GQUOTA_ON(mp) && ip->i_gdquot)
-		(void) xfs_trans_mod_dquot(tp, ip->i_gdquot, field, delta);
+		xfs_trans_mod_ino_dquot(tp, ip, ip->i_gdquot, field, delta);
 	if (XFS_IS_PQUOTA_ON(mp) && ip->i_pdquot)
-		(void) xfs_trans_mod_dquot(tp, ip->i_pdquot, field, delta);
+		xfs_trans_mod_ino_dquot(tp, ip, ip->i_pdquot, field, delta);
 }
 
 STATIC struct xfs_dqtrx *
@@ -268,24 +379,29 @@ xfs_trans_mod_dquot(
 
 /*
  * Given an array of dqtrx structures, lock all the dquots associated and join
- * them to the transaction, provided they have been modified.  We know that the
- * highest number of dquots of one type - usr, grp and prj - involved in a
- * transaction is 3 so we don't need to make this very generic.
+ * them to the transaction, provided they have been modified.
  */
 STATIC void
 xfs_trans_dqlockedjoin(
 	struct xfs_trans	*tp,
 	struct xfs_dqtrx	*q)
 {
+	unsigned int		i;
 	ASSERT(q[0].qt_dquot != NULL);
 	if (q[1].qt_dquot == NULL) {
 		xfs_dqlock(q[0].qt_dquot);
 		xfs_trans_dqjoin(tp, q[0].qt_dquot);
-	} else {
-		ASSERT(XFS_QM_TRANS_MAXDQS == 2);
+	} else if (q[2].qt_dquot == NULL) {
 		xfs_dqlock2(q[0].qt_dquot, q[1].qt_dquot);
 		xfs_trans_dqjoin(tp, q[0].qt_dquot);
 		xfs_trans_dqjoin(tp, q[1].qt_dquot);
+	} else {
+		xfs_dqlockn(q);
+		for (i = 0; i < XFS_QM_TRANS_MAXDQS; i++) {
+			if (q[i].qt_dquot == NULL)
+				break;
+			xfs_trans_dqjoin(tp, q[i].qt_dquot);
+		}
 	}
 }
 
@@ -321,6 +437,29 @@ xfs_apply_quota_reservation_deltas(
 	}
 }
 
+#ifdef CONFIG_XFS_LIVE_HOOKS
+/* Call downstream hooks now that it's time to apply dquot deltas. */
+static inline void
+xfs_trans_apply_dquot_deltas_hook(
+	struct xfs_trans		*tp,
+	struct xfs_dquot		*dqp)
+{
+	if (xfs_hooks_switched_on(&xfs_dqtrx_hooks_switch)) {
+		struct xfs_apply_dqtrx_params	p = {
+			.tx_id		= (uintptr_t)tp,
+			.q_type		= xfs_dquot_type(dqp),
+			.q_id		= dqp->q_id,
+		};
+		struct xfs_quotainfo	*qi = tp->t_mountp->m_quotainfo;
+
+		xfs_hooks_call(&qi->qi_apply_dqtrx_hooks,
+				XFS_APPLY_DQTRX_COMMIT, &p);
+	}
+}
+#else
+# define xfs_trans_apply_dquot_deltas_hook(tp, dqp)	((void)0)
+#endif /* CONFIG_XFS_LIVE_HOOKS */
+
 /*
  * Called by xfs_trans_commit() and similar in spirit to
  * xfs_trans_apply_sb_deltas().
@@ -366,6 +505,8 @@ xfs_trans_apply_dquot_deltas(
 
 			ASSERT(XFS_DQ_IS_LOCKED(dqp));
 
+			xfs_trans_apply_dquot_deltas_hook(tp, dqp);
+
 			/*
 			 * adjust the actual number of blocks used
 			 */
@@ -465,6 +606,29 @@ xfs_trans_apply_dquot_deltas(
 	}
 }
 
+#ifdef CONFIG_XFS_LIVE_HOOKS
+/* Call downstream hooks now that it's time to cancel dquot deltas. */
+static inline void
+xfs_trans_unreserve_and_mod_dquots_hook(
+	struct xfs_trans		*tp,
+	struct xfs_dquot		*dqp)
+{
+	if (xfs_hooks_switched_on(&xfs_dqtrx_hooks_switch)) {
+		struct xfs_apply_dqtrx_params	p = {
+			.tx_id		= (uintptr_t)tp,
+			.q_type		= xfs_dquot_type(dqp),
+			.q_id		= dqp->q_id,
+		};
+		struct xfs_quotainfo	*qi = tp->t_mountp->m_quotainfo;
+
+		xfs_hooks_call(&qi->qi_apply_dqtrx_hooks,
+				XFS_APPLY_DQTRX_UNRESERVE, &p);
+	}
+}
+#else
+# define xfs_trans_unreserve_and_mod_dquots_hook(tp, dqp)	((void)0)
+#endif /* CONFIG_XFS_LIVE_HOOKS */
+
 /*
  * Release the reservations, and adjust the dquots accordingly.
  * This is called only when the transaction is being aborted. If by
@@ -495,6 +659,9 @@ xfs_trans_unreserve_and_mod_dquots(
 			 */
 			if ((dqp = qtrx->qt_dquot) == NULL)
 				break;
+
+			xfs_trans_unreserve_and_mod_dquots_hook(tp, dqp);
+
 			/*
 			 * Unreserve the original reservation. We don't care
 			 * about the number of blocks used field, or deltas.
@@ -706,6 +873,7 @@ error_return:
 error_corrupt:
 	xfs_dqunlock(dqp);
 	xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
+	xfs_fs_mark_sick(mp, XFS_SICK_FS_QUOTACHECK);
 	return -EFSCORRUPTED;
 }
 
@@ -796,7 +964,7 @@ xfs_trans_reserve_quota_nblks(
 		return 0;
 
 	ASSERT(!xfs_is_quota_inode(&mp->m_sb, ip->i_ino));
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
 
 	if (force)
 		qflags |= XFS_QMOPT_FORCE_RES;
diff --git a/fs/xfs/xfs_xattr.c b/fs/xfs/xfs_xattr.c
index 364104e1b3..ab3d22f662 100644
--- a/fs/xfs/xfs_xattr.c
+++ b/fs/xfs/xfs_xattr.c
@@ -17,15 +17,13 @@
 #include "xfs_acl.h"
 #include "xfs_log.h"
 #include "xfs_xattr.h"
+#include "xfs_quota.h"
 
 #include <linux/posix_acl_xattr.h>
 
 /*
  * Get permission to use log-assisted atomic exchange of file extents.
- *
- * Callers must not be running any transactions or hold any inode locks, and
- * they must release the permission by calling xlog_drop_incompat_feat
- * when they're done.
+ * Callers must not be running any transactions or hold any ILOCKs.
  */
 static inline int
 xfs_attr_grab_log_assist(
@@ -33,17 +31,8 @@ xfs_attr_grab_log_assist(
 {
 	int			error = 0;
 
-	/*
-	 * Protect ourselves from an idle log clearing the logged xattrs log
-	 * incompat feature bit.
-	 */
-	xlog_use_incompat_feat(mp->m_log);
-
-	/*
-	 * If log-assisted xattrs are already enabled, the caller can use the
-	 * log assisted swap functions with the log-incompat reference we got.
-	 */
-	if (xfs_sb_version_haslogxattrs(&mp->m_sb))
+	/* xattr update log intent items are already enabled */
+	if (xfs_is_using_logged_xattrs(mp))
 		return 0;
 
 	/*
@@ -52,31 +41,20 @@ xfs_attr_grab_log_assist(
 	 * a V5 filesystem for the superblock field, but we'll require rmap
 	 * or reflink to avoid having to deal with really old kernels.
 	 */
-	if (!xfs_has_reflink(mp) && !xfs_has_rmapbt(mp)) {
-		error = -EOPNOTSUPP;
-		goto drop_incompat;
-	}
+	if (!xfs_has_reflink(mp) && !xfs_has_rmapbt(mp))
+		return -EOPNOTSUPP;
 
 	/* Enable log-assisted xattrs. */
 	error = xfs_add_incompat_log_feature(mp,
 			XFS_SB_FEAT_INCOMPAT_LOG_XATTRS);
 	if (error)
-		goto drop_incompat;
+		return error;
+	xfs_set_using_logged_xattrs(mp);
 
 	xfs_warn_mount(mp, XFS_OPSTATE_WARNED_LARP,
  "EXPERIMENTAL logged extended attributes feature in use. Use at your own risk!");
 
 	return 0;
-drop_incompat:
-	xlog_drop_incompat_feat(mp->m_log);
-	return error;
-}
-
-static inline void
-xfs_attr_rele_log_assist(
-	struct xfs_mount	*mp)
-{
-	xlog_drop_incompat_feat(mp->m_log);
 }
 
 static inline bool
@@ -93,17 +71,31 @@ xfs_attr_want_log_assist(
 
 /*
  * Set or remove an xattr, having grabbed the appropriate logging resources
- * prior to calling libxfs.
+ * prior to calling libxfs.  Callers of this function are only required to
+ * initialize the inode, attr_filter, name, namelen, value, and valuelen fields
+ * of @args.
  */
 int
 xfs_attr_change(
-	struct xfs_da_args	*args)
+	struct xfs_da_args	*args,
+	enum xfs_attr_update	op)
 {
 	struct xfs_mount	*mp = args->dp->i_mount;
-	bool			use_logging = false;
 	int			error;
 
-	ASSERT(!(args->op_flags & XFS_DA_OP_LOGGED));
+	if (xfs_is_shutdown(mp))
+		return -EIO;
+
+	error = xfs_qm_dqattach(args->dp);
+	if (error)
+		return error;
+
+	/*
+	 * We have no control over the attribute names that userspace passes us
+	 * to remove, so we have to allow the name lookup prior to attribute
+	 * removal to fail as well.
+	 */
+	args->op_flags = XFS_DA_OP_OKNOENT;
 
 	if (xfs_attr_want_log_assist(mp)) {
 		error = xfs_attr_grab_log_assist(mp);
@@ -111,14 +103,14 @@ xfs_attr_change(
 			return error;
 
 		args->op_flags |= XFS_DA_OP_LOGGED;
-		use_logging = true;
 	}
 
-	error = xfs_attr_set(args);
+	args->owner = args->dp->i_ino;
+	args->geo = mp->m_attr_geo;
+	args->whichfork = XFS_ATTR_FORK;
+	xfs_attr_sethash(args);
 
-	if (use_logging)
-		xfs_attr_rele_log_assist(mp);
-	return error;
+	return xfs_attr_set(args, op, args->attr_filter & XFS_ATTR_ROOT);
 }
 
 
@@ -145,6 +137,20 @@ xfs_xattr_get(const struct xattr_handler *handler, struct dentry *unused,
 	return args.valuelen;
 }
 
+static inline enum xfs_attr_update
+xfs_xattr_flags_to_op(
+	int		flags,
+	const void	*value)
+{
+	if (!value)
+		return XFS_ATTRUPDATE_REMOVE;
+	if (flags & XATTR_CREATE)
+		return XFS_ATTRUPDATE_CREATE;
+	if (flags & XATTR_REPLACE)
+		return XFS_ATTRUPDATE_REPLACE;
+	return XFS_ATTRUPDATE_UPSERT;
+}
+
 static int
 xfs_xattr_set(const struct xattr_handler *handler,
 	      struct mnt_idmap *idmap, struct dentry *unused,
@@ -154,7 +160,6 @@ xfs_xattr_set(const struct xattr_handler *handler,
 	struct xfs_da_args	args = {
 		.dp		= XFS_I(inode),
 		.attr_filter	= handler->flags,
-		.attr_flags	= flags,
 		.name		= name,
 		.namelen	= strlen(name),
 		.value		= (void *)value,
@@ -162,7 +167,7 @@ xfs_xattr_set(const struct xattr_handler *handler,
 	};
 	int			error;
 
-	error = xfs_attr_change(&args);
+	error = xfs_attr_change(&args, xfs_xattr_flags_to_op(flags, value));
 	if (!error && (handler->flags & XFS_ATTR_ROOT))
 		xfs_forget_acl(inode, name);
 	return error;
@@ -237,6 +242,7 @@ xfs_xattr_put_listent(
 	int		flags,
 	unsigned char	*name,
 	int		namelen,
+	void		*value,
 	int		valuelen)
 {
 	char *prefix;
@@ -244,6 +250,10 @@ xfs_xattr_put_listent(
 
 	ASSERT(context->count >= 0);
 
+	/* Don't expose private xattr namespaces. */
+	if (flags & XFS_ATTR_PRIVATE_NSP_MASK)
+		return;
+
 	if (flags & XFS_ATTR_ROOT) {
 #ifdef CONFIG_XFS_POSIX_ACL
 		if (namelen == SGI_ACL_FILE_SIZE &&
diff --git a/fs/xfs/xfs_xattr.h b/fs/xfs/xfs_xattr.h
index cec766cad2..c3eb858fb5 100644
--- a/fs/xfs/xfs_xattr.h
+++ b/fs/xfs/xfs_xattr.h
@@ -6,7 +6,8 @@
 #ifndef __XFS_XATTR_H__
 #define __XFS_XATTR_H__
 
-int xfs_attr_change(struct xfs_da_args *args);
+enum xfs_attr_update;
+int xfs_attr_change(struct xfs_da_args *args, enum xfs_attr_update op);
 
 extern const struct xattr_handler * const xfs_xattr_handlers[];
 
diff --git a/fs/zonefs/file.c b/fs/zonefs/file.c
index dba5dcb62b..3b103715ac 100644
--- a/fs/zonefs/file.c
+++ b/fs/zonefs/file.c
@@ -125,7 +125,8 @@ static void zonefs_readahead(struct readahead_control *rac)
  * which implies that the page range can only be within the fixed inode size.
  */
 static int zonefs_write_map_blocks(struct iomap_writepage_ctx *wpc,
-				   struct inode *inode, loff_t offset)
+				   struct inode *inode, loff_t offset,
+				   unsigned int len)
 {
 	struct zonefs_zone *z = zonefs_inode_zone(inode);
 
diff --git a/fs/zonefs/super.c b/fs/zonefs/super.c
index b6e8e7c962..964fa7f240 100644
--- a/fs/zonefs/super.c
+++ b/fs/zonefs/super.c
@@ -15,12 +15,13 @@
 #include <linux/writeback.h>
 #include <linux/quotaops.h>
 #include <linux/seq_file.h>
-#include <linux/parser.h>
 #include <linux/uio.h>
 #include <linux/mman.h>
 #include <linux/sched/mm.h>
 #include <linux/crc32.h>
 #include <linux/task_io_accounting_ops.h>
+#include <linux/fs_parser.h>
+#include <linux/fs_context.h>
 
 #include "zonefs.h"
 
@@ -113,7 +114,7 @@ static int zonefs_zone_mgmt(struct super_block *sb,
 
 	trace_zonefs_zone_mgmt(sb, z, op);
 	ret = blkdev_zone_mgmt(sb->s_bdev, op, z->z_sector,
-			       z->z_size >> SECTOR_SHIFT, GFP_NOFS);
+			       z->z_size >> SECTOR_SHIFT);
 	if (ret) {
 		zonefs_err(sb,
 			   "Zone management operation %s at %llu failed %d\n",
@@ -470,58 +471,47 @@ static int zonefs_statfs(struct dentry *dentry, struct kstatfs *buf)
 }
 
 enum {
-	Opt_errors_ro, Opt_errors_zro, Opt_errors_zol, Opt_errors_repair,
-	Opt_explicit_open, Opt_err,
+	Opt_errors, Opt_explicit_open,
 };
 
-static const match_table_t tokens = {
-	{ Opt_errors_ro,	"errors=remount-ro"},
-	{ Opt_errors_zro,	"errors=zone-ro"},
-	{ Opt_errors_zol,	"errors=zone-offline"},
-	{ Opt_errors_repair,	"errors=repair"},
-	{ Opt_explicit_open,	"explicit-open" },
-	{ Opt_err,		NULL}
+struct zonefs_context {
+	unsigned long s_mount_opts;
 };
 
-static int zonefs_parse_options(struct super_block *sb, char *options)
-{
-	struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
-	substring_t args[MAX_OPT_ARGS];
-	char *p;
-
-	if (!options)
-		return 0;
-
-	while ((p = strsep(&options, ",")) != NULL) {
-		int token;
+static const struct constant_table zonefs_param_errors[] = {
+	{"remount-ro",		ZONEFS_MNTOPT_ERRORS_RO},
+	{"zone-ro",		ZONEFS_MNTOPT_ERRORS_ZRO},
+	{"zone-offline",	ZONEFS_MNTOPT_ERRORS_ZOL},
+	{"repair", 		ZONEFS_MNTOPT_ERRORS_REPAIR},
+	{}
+};
 
-		if (!*p)
-			continue;
+static const struct fs_parameter_spec zonefs_param_spec[] = {
+	fsparam_enum	("errors",		Opt_errors, zonefs_param_errors),
+	fsparam_flag	("explicit-open",	Opt_explicit_open),
+	{}
+};
 
-		token = match_token(p, tokens, args);
-		switch (token) {
-		case Opt_errors_ro:
-			sbi->s_mount_opts &= ~ZONEFS_MNTOPT_ERRORS_MASK;
-			sbi->s_mount_opts |= ZONEFS_MNTOPT_ERRORS_RO;
-			break;
-		case Opt_errors_zro:
-			sbi->s_mount_opts &= ~ZONEFS_MNTOPT_ERRORS_MASK;
-			sbi->s_mount_opts |= ZONEFS_MNTOPT_ERRORS_ZRO;
-			break;
-		case Opt_errors_zol:
-			sbi->s_mount_opts &= ~ZONEFS_MNTOPT_ERRORS_MASK;
-			sbi->s_mount_opts |= ZONEFS_MNTOPT_ERRORS_ZOL;
-			break;
-		case Opt_errors_repair:
-			sbi->s_mount_opts &= ~ZONEFS_MNTOPT_ERRORS_MASK;
-			sbi->s_mount_opts |= ZONEFS_MNTOPT_ERRORS_REPAIR;
-			break;
-		case Opt_explicit_open:
-			sbi->s_mount_opts |= ZONEFS_MNTOPT_EXPLICIT_OPEN;
-			break;
-		default:
-			return -EINVAL;
-		}
+static int zonefs_parse_param(struct fs_context *fc, struct fs_parameter *param)
+{
+	struct zonefs_context *ctx = fc->fs_private;
+	struct fs_parse_result result;
+	int opt;
+
+	opt = fs_parse(fc, zonefs_param_spec, param, &result);
+	if (opt < 0)
+		return opt;
+
+	switch (opt) {
+	case Opt_errors:
+		ctx->s_mount_opts &= ~ZONEFS_MNTOPT_ERRORS_MASK;
+		ctx->s_mount_opts |= result.uint_32;
+		break;
+	case Opt_explicit_open:
+		ctx->s_mount_opts |= ZONEFS_MNTOPT_EXPLICIT_OPEN;
+		break;
+	default:
+		return -EINVAL;
 	}
 
 	return 0;
@@ -543,13 +533,6 @@ static int zonefs_show_options(struct seq_file *seq, struct dentry *root)
 	return 0;
 }
 
-static int zonefs_remount(struct super_block *sb, int *flags, char *data)
-{
-	sync_filesystem(sb);
-
-	return zonefs_parse_options(sb, data);
-}
-
 static int zonefs_inode_setattr(struct mnt_idmap *idmap,
 				struct dentry *dentry, struct iattr *iattr)
 {
@@ -1065,7 +1048,7 @@ static int zonefs_init_zgroup(struct super_block *sb,
 	zonefs_info(sb, "Zone group \"%s\" has %u file%s\n",
 		    zonefs_zgroup_name(ztype),
 		    zgroup->g_nr_zones,
-		    zgroup->g_nr_zones > 1 ? "s" : "");
+		    str_plural(zgroup->g_nr_zones));
 
 	return 0;
 }
@@ -1205,7 +1188,6 @@ static const struct super_operations zonefs_sops = {
 	.alloc_inode	= zonefs_alloc_inode,
 	.free_inode	= zonefs_free_inode,
 	.statfs		= zonefs_statfs,
-	.remount_fs	= zonefs_remount,
 	.show_options	= zonefs_show_options,
 };
 
@@ -1250,9 +1232,10 @@ static void zonefs_release_zgroup_inodes(struct super_block *sb)
  * sub-directories and files according to the device zone configuration and
  * format options.
  */
-static int zonefs_fill_super(struct super_block *sb, void *data, int silent)
+static int zonefs_fill_super(struct super_block *sb, struct fs_context *fc)
 {
 	struct zonefs_sb_info *sbi;
+	struct zonefs_context *ctx = fc->fs_private;
 	struct inode *inode;
 	enum zonefs_ztype ztype;
 	int ret;
@@ -1289,7 +1272,7 @@ static int zonefs_fill_super(struct super_block *sb, void *data, int silent)
 	sbi->s_uid = GLOBAL_ROOT_UID;
 	sbi->s_gid = GLOBAL_ROOT_GID;
 	sbi->s_perm = 0640;
-	sbi->s_mount_opts = ZONEFS_MNTOPT_ERRORS_RO;
+	sbi->s_mount_opts = ctx->s_mount_opts;
 
 	atomic_set(&sbi->s_wro_seq_files, 0);
 	sbi->s_max_wro_seq_files = bdev_max_open_zones(sb->s_bdev);
@@ -1300,10 +1283,6 @@ static int zonefs_fill_super(struct super_block *sb, void *data, int silent)
 	if (ret)
 		return ret;
 
-	ret = zonefs_parse_options(sb, data);
-	if (ret)
-		return ret;
-
 	zonefs_info(sb, "Mounting %u zones", bdev_nr_zones(sb->s_bdev));
 
 	if (!sbi->s_max_wro_seq_files &&
@@ -1364,12 +1343,6 @@ cleanup:
 	return ret;
 }
 
-static struct dentry *zonefs_mount(struct file_system_type *fs_type,
-				   int flags, const char *dev_name, void *data)
-{
-	return mount_bdev(fs_type, flags, dev_name, data, zonefs_fill_super);
-}
-
 static void zonefs_kill_super(struct super_block *sb)
 {
 	struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
@@ -1384,22 +1357,72 @@ static void zonefs_kill_super(struct super_block *sb)
 	kfree(sbi);
 }
 
+static void zonefs_free_fc(struct fs_context *fc)
+{
+	struct zonefs_context *ctx = fc->fs_private;
+
+	kfree(ctx);
+}
+
+static int zonefs_get_tree(struct fs_context *fc)
+{
+	return get_tree_bdev(fc, zonefs_fill_super);
+}
+
+static int zonefs_reconfigure(struct fs_context *fc)
+{
+	struct zonefs_context *ctx = fc->fs_private;
+	struct super_block *sb = fc->root->d_sb;
+	struct zonefs_sb_info *sbi = sb->s_fs_info;
+
+	sync_filesystem(fc->root->d_sb);
+	/* Copy new options from ctx into sbi. */
+	sbi->s_mount_opts = ctx->s_mount_opts;
+
+	return 0;
+}
+
+static const struct fs_context_operations zonefs_context_ops = {
+	.parse_param    = zonefs_parse_param,
+	.get_tree       = zonefs_get_tree,
+	.reconfigure	= zonefs_reconfigure,
+	.free           = zonefs_free_fc,
+};
+
+/*
+ * Set up the filesystem mount context.
+ */
+static int zonefs_init_fs_context(struct fs_context *fc)
+{
+	struct zonefs_context *ctx;
+
+	ctx = kzalloc(sizeof(struct zonefs_context), GFP_KERNEL);
+	if (!ctx)
+		return -ENOMEM;
+	ctx->s_mount_opts = ZONEFS_MNTOPT_ERRORS_RO;
+	fc->ops = &zonefs_context_ops;
+	fc->fs_private = ctx;
+
+	return 0;
+}
+
 /*
  * File system definition and registration.
  */
 static struct file_system_type zonefs_type = {
-	.owner		= THIS_MODULE,
-	.name		= "zonefs",
-	.mount		= zonefs_mount,
-	.kill_sb	= zonefs_kill_super,
-	.fs_flags	= FS_REQUIRES_DEV,
+	.owner			= THIS_MODULE,
+	.name			= "zonefs",
+	.kill_sb		= zonefs_kill_super,
+	.fs_flags		= FS_REQUIRES_DEV,
+	.init_fs_context	= zonefs_init_fs_context,
+	.parameters		= zonefs_param_spec,
 };
 
 static int __init zonefs_init_inodecache(void)
 {
 	zonefs_inode_cachep = kmem_cache_create("zonefs_inode_cache",
 			sizeof(struct zonefs_inode_info), 0,
-			(SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD | SLAB_ACCOUNT),
+			SLAB_RECLAIM_ACCOUNT | SLAB_ACCOUNT,
 			NULL);
 	if (zonefs_inode_cachep == NULL)
 		return -ENOMEM;