1 files changed, 157 insertions, 75 deletions

diff --git a/fs/quota/dquot.c b/fs/quota/dquot.c
index 770a2b143..868936076 100644
--- a/fs/quota/dquot.c
+++ b/fs/quota/dquot.c
@@ -223,18 +223,31 @@ static void put_quota_format(struct quota_format_type *fmt)
 
 /*
  * Dquot List Management:
- * The quota code uses three lists for dquot management: the inuse_list,
- * free_dquots, and dquot_hash[] array. A single dquot structure may be
- * on all three lists, depending on its current state.
+ * The quota code uses five lists for dquot management: the inuse_list,
+ * releasing_dquots, free_dquots, dqi_dirty_list, and dquot_hash[] array.
+ * A single dquot structure may be on some of those lists, depending on
+ * its current state.
  *
  * All dquots are placed to the end of inuse_list when first created, and this
  * list is used for invalidate operation, which must look at every dquot.
  *
- * Unused dquots (dq_count == 0) are added to the free_dquots list when freed,
- * and this list is searched whenever we need an available dquot.  Dquots are
- * removed from the list as soon as they are used again, and
- * dqstats.free_dquots gives the number of dquots on the list. When
- * dquot is invalidated it's completely released from memory.
+ * When the last reference of a dquot is dropped, the dquot is added to
+ * releasing_dquots. We'll then queue work item which will call
+ * synchronize_srcu() and after that perform the final cleanup of all the
+ * dquots on the list. Each cleaned up dquot is moved to free_dquots list.
+ * Both releasing_dquots and free_dquots use the dq_free list_head in the dquot
+ * struct.
+ *
+ * Unused and cleaned up dquots are in the free_dquots list and this list is
+ * searched whenever we need an available dquot. Dquots are removed from the
+ * list as soon as they are used again and dqstats.free_dquots gives the number
+ * of dquots on the list. When dquot is invalidated it's completely released
+ * from memory.
+ *
+ * Dirty dquots are added to the dqi_dirty_list of quota_info when mark
+ * dirtied, and this list is searched when writing dirty dquots back to
+ * quota file. Note that some filesystems do dirty dquot tracking on their
+ * own (e.g. in a journal) and thus don't use dqi_dirty_list.
  *
  * Dquots with a specific identity (device, type and id) are placed on
  * one of the dquot_hash[] hash chains. The provides an efficient search
@@ -243,6 +256,7 @@ static void put_quota_format(struct quota_format_type *fmt)
 
 static LIST_HEAD(inuse_list);
 static LIST_HEAD(free_dquots);
+static LIST_HEAD(releasing_dquots);
 static unsigned int dq_hash_bits, dq_hash_mask;
 static struct hlist_head *dquot_hash;
 
@@ -253,6 +267,9 @@ static qsize_t inode_get_rsv_space(struct inode *inode);
 static qsize_t __inode_get_rsv_space(struct inode *inode);
 static int __dquot_initialize(struct inode *inode, int type);
 
+static void quota_release_workfn(struct work_struct *work);
+static DECLARE_DELAYED_WORK(quota_release_work, quota_release_workfn);
+
 static inline unsigned int
 hashfn(const struct super_block *sb, struct kqid qid)
 {
@@ -300,12 +317,21 @@ static inline void put_dquot_last(struct dquot *dquot)
 	dqstats_inc(DQST_FREE_DQUOTS);
 }
 
+static inline void put_releasing_dquots(struct dquot *dquot)
+{
+	list_add_tail(&dquot->dq_free, &releasing_dquots);
+	set_bit(DQ_RELEASING_B, &dquot->dq_flags);
+}
+
 static inline void remove_free_dquot(struct dquot *dquot)
 {
 	if (list_empty(&dquot->dq_free))
 		return;
 	list_del_init(&dquot->dq_free);
-	dqstats_dec(DQST_FREE_DQUOTS);
+	if (!test_bit(DQ_RELEASING_B, &dquot->dq_flags))
+		dqstats_dec(DQST_FREE_DQUOTS);
+	else
+		clear_bit(DQ_RELEASING_B, &dquot->dq_flags);
 }
 
 static inline void put_inuse(struct dquot *dquot)
@@ -331,6 +357,11 @@ static void wait_on_dquot(struct dquot *dquot)
 	mutex_unlock(&dquot->dq_lock);
 }
 
+static inline int dquot_active(struct dquot *dquot)
+{
+	return test_bit(DQ_ACTIVE_B, &dquot->dq_flags);
+}
+
 static inline int dquot_dirty(struct dquot *dquot)
 {
 	return test_bit(DQ_MOD_B, &dquot->dq_flags);
@@ -346,14 +377,14 @@ int dquot_mark_dquot_dirty(struct dquot *dquot)
 {
 	int ret = 1;
 
-	if (!test_bit(DQ_ACTIVE_B, &dquot->dq_flags))
+	if (!dquot_active(dquot))
 		return 0;
 
 	if (sb_dqopt(dquot->dq_sb)->flags & DQUOT_NOLIST_DIRTY)
 		return test_and_set_bit(DQ_MOD_B, &dquot->dq_flags);
 
 	/* If quota is dirty already, we don't have to acquire dq_list_lock */
-	if (test_bit(DQ_MOD_B, &dquot->dq_flags))
+	if (dquot_dirty(dquot))
 		return 1;
 
 	spin_lock(&dq_list_lock);
@@ -432,7 +463,7 @@ int dquot_acquire(struct dquot *dquot)
 	smp_mb__before_atomic();
 	set_bit(DQ_READ_B, &dquot->dq_flags);
 	/* Instantiate dquot if needed */
-	if (!test_bit(DQ_ACTIVE_B, &dquot->dq_flags) && !dquot->dq_off) {
+	if (!dquot_active(dquot) && !dquot->dq_off) {
 		ret = dqopt->ops[dquot->dq_id.type]->commit_dqblk(dquot);
 		/* Write the info if needed */
 		if (info_dirty(&dqopt->info[dquot->dq_id.type])) {
@@ -471,7 +502,7 @@ int dquot_commit(struct dquot *dquot)
 		goto out_lock;
 	/* Inactive dquot can be only if there was error during read/init
 	 * => we have better not writing it */
-	if (test_bit(DQ_ACTIVE_B, &dquot->dq_flags))
+	if (dquot_active(dquot))
 		ret = dqopt->ops[dquot->dq_id.type]->commit_dqblk(dquot);
 	else
 		ret = -EIO;
@@ -532,6 +563,8 @@ static void invalidate_dquots(struct super_block *sb, int type)
 	struct dquot *dquot, *tmp;
 
 restart:
+	flush_delayed_work(&quota_release_work);
+
 	spin_lock(&dq_list_lock);
 	list_for_each_entry_safe(dquot, tmp, &inuse_list, dq_inuse) {
 		if (dquot->dq_sb != sb)
@@ -540,7 +573,7 @@ restart:
 			continue;
 		/* Wait for dquot users */
 		if (atomic_read(&dquot->dq_count)) {
-			dqgrab(dquot);
+			atomic_inc(&dquot->dq_count);
 			spin_unlock(&dq_list_lock);
 			/*
 			 * Once dqput() wakes us up, we know it's time to free
@@ -559,6 +592,15 @@ restart:
 			goto restart;
 		}
 		/*
+		 * The last user already dropped its reference but dquot didn't
+		 * get fully cleaned up yet. Restart the scan which flushes the
+		 * work cleaning up released dquots.
+		 */
+		if (test_bit(DQ_RELEASING_B, &dquot->dq_flags)) {
+			spin_unlock(&dq_list_lock);
+			goto restart;
+		}
+		/*
 		 * Quota now has no users and it has been written on last
 		 * dqput()
 		 */
@@ -582,14 +624,13 @@ int dquot_scan_active(struct super_block *sb,
 
 	spin_lock(&dq_list_lock);
 	list_for_each_entry(dquot, &inuse_list, dq_inuse) {
-		if (!test_bit(DQ_ACTIVE_B, &dquot->dq_flags))
+		if (!dquot_active(dquot))
 			continue;
 		if (dquot->dq_sb != sb)
 			continue;
 		/* Now we have active dquot so we can just increase use count */
 		atomic_inc(&dquot->dq_count);
 		spin_unlock(&dq_list_lock);
-		dqstats_inc(DQST_LOOKUPS);
 		dqput(old_dquot);
 		old_dquot = dquot;
 		/*
@@ -598,7 +639,7 @@ int dquot_scan_active(struct super_block *sb,
 		 * outstanding call and recheck the DQ_ACTIVE_B after that.
 		 */
 		wait_on_dquot(dquot);
-		if (test_bit(DQ_ACTIVE_B, &dquot->dq_flags)) {
+		if (dquot_active(dquot)) {
 			ret = fn(dquot, priv);
 			if (ret < 0)
 				goto out;
@@ -614,6 +655,18 @@ out:
 }
 EXPORT_SYMBOL(dquot_scan_active);
 
+static inline int dquot_write_dquot(struct dquot *dquot)
+{
+	int ret = dquot->dq_sb->dq_op->write_dquot(dquot);
+	if (ret < 0) {
+		quota_error(dquot->dq_sb, "Can't write quota structure "
+			    "(error %d). Quota may get out of sync!", ret);
+		/* Clear dirty bit anyway to avoid infinite loop. */
+		clear_dquot_dirty(dquot);
+	}
+	return ret;
+}
+
 /* Write all dquot structures to quota files */
 int dquot_writeback_dquots(struct super_block *sb, int type)
 {
@@ -637,24 +690,23 @@ int dquot_writeback_dquots(struct super_block *sb, int type)
 			dquot = list_first_entry(&dirty, struct dquot,
 						 dq_dirty);
 
-			WARN_ON(!test_bit(DQ_ACTIVE_B, &dquot->dq_flags));
+			WARN_ON(!dquot_active(dquot));
+			/* If the dquot is releasing we should not touch it */
+			if (test_bit(DQ_RELEASING_B, &dquot->dq_flags)) {
+				spin_unlock(&dq_list_lock);
+				flush_delayed_work(&quota_release_work);
+				spin_lock(&dq_list_lock);
+				continue;
+			}
 
 			/* Now we have active dquot from which someone is
  			 * holding reference so we can safely just increase
 			 * use count */
 			dqgrab(dquot);
 			spin_unlock(&dq_list_lock);
-			dqstats_inc(DQST_LOOKUPS);
-			err = sb->dq_op->write_dquot(dquot);
-			if (err) {
-				/*
-				 * Clear dirty bit anyway to avoid infinite
-				 * loop here.
-				 */
-				clear_dquot_dirty(dquot);
-				if (!ret)
-					ret = err;
-			}
+			err = dquot_write_dquot(dquot);
+			if (err && !ret)
+				ret = err;
 			dqput(dquot);
 			spin_lock(&dq_list_lock);
 		}
@@ -748,12 +800,52 @@ static struct shrinker dqcache_shrinker = {
 };
 
 /*
+ * Safely release dquot and put reference to dquot.
+ */
+static void quota_release_workfn(struct work_struct *work)
+{
+	struct dquot *dquot;
+	struct list_head rls_head;
+
+	spin_lock(&dq_list_lock);
+	/* Exchange the list head to avoid livelock. */
+	list_replace_init(&releasing_dquots, &rls_head);
+	spin_unlock(&dq_list_lock);
+	synchronize_srcu(&dquot_srcu);
+
+restart:
+	spin_lock(&dq_list_lock);
+	while (!list_empty(&rls_head)) {
+		dquot = list_first_entry(&rls_head, struct dquot, dq_free);
+		WARN_ON_ONCE(atomic_read(&dquot->dq_count));
+		/*
+		 * Note that DQ_RELEASING_B protects us from racing with
+		 * invalidate_dquots() calls so we are safe to work with the
+		 * dquot even after we drop dq_list_lock.
+		 */
+		if (dquot_dirty(dquot)) {
+			spin_unlock(&dq_list_lock);
+			/* Commit dquot before releasing */
+			dquot_write_dquot(dquot);
+			goto restart;
+		}
+		if (dquot_active(dquot)) {
+			spin_unlock(&dq_list_lock);
+			dquot->dq_sb->dq_op->release_dquot(dquot);
+			goto restart;
+		}
+		/* Dquot is inactive and clean, now move it to free list */
+		remove_free_dquot(dquot);
+		put_dquot_last(dquot);
+	}
+	spin_unlock(&dq_list_lock);
+}
+
+/*
  * Put reference to dquot
  */
 void dqput(struct dquot *dquot)
 {
-	int ret;
-
 	if (!dquot)
 		return;
 #ifdef CONFIG_QUOTA_DEBUG
@@ -765,7 +857,7 @@ void dqput(struct dquot *dquot)
 	}
 #endif
 	dqstats_inc(DQST_DROPS);
-we_slept:
+
 	spin_lock(&dq_list_lock);
 	if (atomic_read(&dquot->dq_count) > 1) {
 		/* We have more than one user... nothing to do */
@@ -777,35 +869,16 @@ we_slept:
 		spin_unlock(&dq_list_lock);
 		return;
 	}
+
 	/* Need to release dquot? */
-	if (dquot_dirty(dquot)) {
-		spin_unlock(&dq_list_lock);
-		/* Commit dquot before releasing */
-		ret = dquot->dq_sb->dq_op->write_dquot(dquot);
-		if (ret < 0) {
-			quota_error(dquot->dq_sb, "Can't write quota structure"
-				    " (error %d). Quota may get out of sync!",
-				    ret);
-			/*
-			 * We clear dirty bit anyway, so that we avoid
-			 * infinite loop here
-			 */
-			clear_dquot_dirty(dquot);
-		}
-		goto we_slept;
-	}
-	if (test_bit(DQ_ACTIVE_B, &dquot->dq_flags)) {
-		spin_unlock(&dq_list_lock);
-		dquot->dq_sb->dq_op->release_dquot(dquot);
-		goto we_slept;
-	}
-	atomic_dec(&dquot->dq_count);
 #ifdef CONFIG_QUOTA_DEBUG
 	/* sanity check */
 	BUG_ON(!list_empty(&dquot->dq_free));
 #endif
-	put_dquot_last(dquot);
+	put_releasing_dquots(dquot);
+	atomic_dec(&dquot->dq_count);
 	spin_unlock(&dq_list_lock);
+	queue_delayed_work(system_unbound_wq, &quota_release_work, 1);
 }
 EXPORT_SYMBOL(dqput);
 
@@ -892,10 +965,10 @@ we_slept:
 		dqstats_inc(DQST_LOOKUPS);
 	}
 	/* Wait for dq_lock - after this we know that either dquot_release() is
-	 * already finished or it will be canceled due to dq_count > 1 test */
+	 * already finished or it will be canceled due to dq_count > 0 test */
 	wait_on_dquot(dquot);
 	/* Read the dquot / allocate space in quota file */
-	if (!test_bit(DQ_ACTIVE_B, &dquot->dq_flags)) {
+	if (!dquot_active(dquot)) {
 		int err;
 
 		err = sb->dq_op->acquire_dquot(dquot);
@@ -1408,7 +1481,7 @@ static int info_bdq_free(struct dquot *dquot, qsize_t space)
 	return QUOTA_NL_NOWARN;
 }
 
-static int dquot_active(const struct inode *inode)
+static int inode_quota_active(const struct inode *inode)
 {
 	struct super_block *sb = inode->i_sb;
 
@@ -1431,7 +1504,7 @@ static int __dquot_initialize(struct inode *inode, int type)
 	qsize_t rsv;
 	int ret = 0;
 
-	if (!dquot_active(inode))
+	if (!inode_quota_active(inode))
 		return 0;
 
 	dquots = i_dquot(inode);
@@ -1539,7 +1612,7 @@ bool dquot_initialize_needed(struct inode *inode)
 	struct dquot **dquots;
 	int i;
 
-	if (!dquot_active(inode))
+	if (!inode_quota_active(inode))
 		return false;
 
 	dquots = i_dquot(inode);
@@ -1650,7 +1723,7 @@ int __dquot_alloc_space(struct inode *inode, qsize_t number, int flags)
 	int reserve = flags & DQUOT_SPACE_RESERVE;
 	struct dquot **dquots;
 
-	if (!dquot_active(inode)) {
+	if (!inode_quota_active(inode)) {
 		if (reserve) {
 			spin_lock(&inode->i_lock);
 			*inode_reserved_space(inode) += number;
@@ -1722,7 +1795,7 @@ int dquot_alloc_inode(struct inode *inode)
 	struct dquot_warn warn[MAXQUOTAS];
 	struct dquot * const *dquots;
 
-	if (!dquot_active(inode))
+	if (!inode_quota_active(inode))
 		return 0;
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
 		warn[cnt].w_type = QUOTA_NL_NOWARN;
@@ -1765,7 +1838,7 @@ int dquot_claim_space_nodirty(struct inode *inode, qsize_t number)
 	struct dquot **dquots;
 	int cnt, index;
 
-	if (!dquot_active(inode)) {
+	if (!inode_quota_active(inode)) {
 		spin_lock(&inode->i_lock);
 		*inode_reserved_space(inode) -= number;
 		__inode_add_bytes(inode, number);
@@ -1807,7 +1880,7 @@ void dquot_reclaim_space_nodirty(struct inode *inode, qsize_t number)
 	struct dquot **dquots;
 	int cnt, index;
 
-	if (!dquot_active(inode)) {
+	if (!inode_quota_active(inode)) {
 		spin_lock(&inode->i_lock);
 		*inode_reserved_space(inode) += number;
 		__inode_sub_bytes(inode, number);
@@ -1851,7 +1924,7 @@ void __dquot_free_space(struct inode *inode, qsize_t number, int flags)
 	struct dquot **dquots;
 	int reserve = flags & DQUOT_SPACE_RESERVE, index;
 
-	if (!dquot_active(inode)) {
+	if (!inode_quota_active(inode)) {
 		if (reserve) {
 			spin_lock(&inode->i_lock);
 			*inode_reserved_space(inode) -= number;
@@ -1906,7 +1979,7 @@ void dquot_free_inode(struct inode *inode)
 	struct dquot * const *dquots;
 	int index;
 
-	if (!dquot_active(inode))
+	if (!inode_quota_active(inode))
 		return;
 
 	dquots = i_dquot(inode);
@@ -2077,7 +2150,7 @@ int dquot_transfer(struct inode *inode, struct iattr *iattr)
 	struct super_block *sb = inode->i_sb;
 	int ret;
 
-	if (!dquot_active(inode))
+	if (!inode_quota_active(inode))
 		return 0;
 
 	if (iattr->ia_valid & ATTR_UID && !uid_eq(iattr->ia_uid, inode->i_uid)){
@@ -2312,6 +2385,20 @@ static int vfs_setup_quota_inode(struct inode *inode, int type)
 	if (sb_has_quota_loaded(sb, type))
 		return -EBUSY;
 
+	/*
+	 * Quota files should never be encrypted.  They should be thought of as
+	 * filesystem metadata, not user data.  New-style internal quota files
+	 * cannot be encrypted by users anyway, but old-style external quota
+	 * files could potentially be incorrectly created in an encrypted
+	 * directory, hence this explicit check.  Some reasons why encrypted
+	 * quota files don't work include: (1) some filesystems that support
+	 * encryption don't handle it in their quota_read and quota_write, and
+	 * (2) cleaning up encrypted quota files at unmount would need special
+	 * consideration, as quota files are cleaned up later than user files.
+	 */
+	if (IS_ENCRYPTED(inode))
+		return -EINVAL;
+
 	dqopt->files[type] = igrab(inode);
 	if (!dqopt->files[type])
 		return -EIO;
@@ -2407,7 +2494,8 @@ int dquot_load_quota_sb(struct super_block *sb, int type, int format_id,
 
 	error = add_dquot_ref(sb, type);
 	if (error)
-		dquot_disable(sb, type, flags);
+		dquot_disable(sb, type,
+			      DQUOT_USAGE_ENABLED | DQUOT_LIMITS_ENABLED);
 
 	return error;
 out_fmt:
@@ -2534,21 +2622,15 @@ int dquot_quota_on_mount(struct super_block *sb, char *qf_name,
 	struct dentry *dentry;
 	int error;
 
-	dentry = lookup_one_len_unlocked(qf_name, sb->s_root, strlen(qf_name));
+	dentry = lookup_positive_unlocked(qf_name, sb->s_root, strlen(qf_name));
 	if (IS_ERR(dentry))
 		return PTR_ERR(dentry);
 
-	if (d_really_is_negative(dentry)) {
-		error = -ENOENT;
-		goto out;
-	}
-
 	error = security_quota_on(dentry);
 	if (!error)
 		error = vfs_load_quota_inode(d_inode(dentry), type, format_id,
 				DQUOT_USAGE_ENABLED | DQUOT_LIMITS_ENABLED);
 
-out:
 	dput(dentry);
 	return error;
 }