summaryrefslogtreecommitdiffstats
path: root/fs/ceph
diff options
context:
space:
mode:
Diffstat (limited to 'fs/ceph')
-rw-r--r--fs/ceph/addr.c24
-rw-r--r--fs/ceph/dir.c28
-rw-r--r--fs/ceph/file.c66
-rw-r--r--fs/ceph/inode.c48
-rw-r--r--fs/ceph/mds_client.c270
-rw-r--r--fs/ceph/mds_client.h28
-rw-r--r--fs/ceph/super.c3
7 files changed, 433 insertions, 34 deletions
diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c
index ee9caf7916..8c16bc5250 100644
--- a/fs/ceph/addr.c
+++ b/fs/ceph/addr.c
@@ -193,7 +193,7 @@ static void ceph_netfs_expand_readahead(struct netfs_io_request *rreq)
* block, but do not exceed the file size, unless the original
* request already exceeds it.
*/
- new_end = min(round_up(end, lo->stripe_unit), rreq->i_size);
+ new_end = umin(round_up(end, lo->stripe_unit), rreq->i_size);
if (new_end > end && new_end <= rreq->start + max_len)
rreq->len = new_end - rreq->start;
@@ -498,11 +498,6 @@ const struct netfs_request_ops ceph_netfs_ops = {
};
#ifdef CONFIG_CEPH_FSCACHE
-static void ceph_set_page_fscache(struct page *page)
-{
- set_page_fscache(page);
-}
-
static void ceph_fscache_write_terminated(void *priv, ssize_t error, bool was_async)
{
struct inode *inode = priv;
@@ -517,13 +512,9 @@ static void ceph_fscache_write_to_cache(struct inode *inode, u64 off, u64 len, b
struct fscache_cookie *cookie = ceph_fscache_cookie(ci);
fscache_write_to_cache(cookie, inode->i_mapping, off, len, i_size_read(inode),
- ceph_fscache_write_terminated, inode, caching);
+ ceph_fscache_write_terminated, inode, true, caching);
}
#else
-static inline void ceph_set_page_fscache(struct page *page)
-{
-}
-
static inline void ceph_fscache_write_to_cache(struct inode *inode, u64 off, u64 len, bool caching)
{
}
@@ -715,8 +706,6 @@ static int writepage_nounlock(struct page *page, struct writeback_control *wbc)
len = wlen;
set_page_writeback(page);
- if (caching)
- ceph_set_page_fscache(page);
ceph_fscache_write_to_cache(inode, page_off, len, caching);
if (IS_ENCRYPTED(inode)) {
@@ -800,8 +789,6 @@ static int ceph_writepage(struct page *page, struct writeback_control *wbc)
return AOP_WRITEPAGE_ACTIVATE;
}
- wait_on_page_fscache(page);
-
err = writepage_nounlock(page, wbc);
if (err == -ERESTARTSYS) {
/* direct memory reclaimer was killed by SIGKILL. return 0
@@ -1075,7 +1062,7 @@ get_more_pages:
unlock_page(page);
break;
}
- if (PageWriteback(page) || PageFsCache(page)) {
+ if (PageWriteback(page)) {
if (wbc->sync_mode == WB_SYNC_NONE) {
doutc(cl, "%p under writeback\n", page);
unlock_page(page);
@@ -1083,7 +1070,6 @@ get_more_pages:
}
doutc(cl, "waiting on writeback %p\n", page);
wait_on_page_writeback(page);
- wait_on_page_fscache(page);
}
if (!clear_page_dirty_for_io(page)) {
@@ -1268,8 +1254,6 @@ new_request:
}
set_page_writeback(page);
- if (caching)
- ceph_set_page_fscache(page);
len += thp_size(page);
}
ceph_fscache_write_to_cache(inode, offset, len, caching);
@@ -1513,7 +1497,7 @@ static int ceph_write_begin(struct file *file, struct address_space *mapping,
if (r < 0)
return r;
- folio_wait_fscache(folio);
+ folio_wait_private_2(folio); /* [DEPRECATED] */
WARN_ON_ONCE(!folio_test_locked(folio));
*pagep = &folio->page;
return 0;
diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c
index 0e9f56eaba..82a2e2a06a 100644
--- a/fs/ceph/dir.c
+++ b/fs/ceph/dir.c
@@ -1336,8 +1336,12 @@ static int ceph_unlink(struct inode *dir, struct dentry *dentry)
struct inode *inode = d_inode(dentry);
struct ceph_mds_request *req;
bool try_async = ceph_test_mount_opt(fsc, ASYNC_DIROPS);
+ struct dentry *dn;
int err = -EROFS;
int op;
+ char *path;
+ int pathlen;
+ u64 pathbase;
if (ceph_snap(dir) == CEPH_SNAPDIR) {
/* rmdir .snap/foo is RMSNAP */
@@ -1351,6 +1355,30 @@ static int ceph_unlink(struct inode *dir, struct dentry *dentry)
CEPH_MDS_OP_RMDIR : CEPH_MDS_OP_UNLINK;
} else
goto out;
+
+ dn = d_find_alias(dir);
+ if (!dn) {
+ try_async = false;
+ } else {
+ path = ceph_mdsc_build_path(mdsc, dn, &pathlen, &pathbase, 0);
+ if (IS_ERR(path)) {
+ try_async = false;
+ err = 0;
+ } else {
+ err = ceph_mds_check_access(mdsc, path, MAY_WRITE);
+ }
+ ceph_mdsc_free_path(path, pathlen);
+ dput(dn);
+
+ /* For none EACCES cases will let the MDS do the mds auth check */
+ if (err == -EACCES) {
+ return err;
+ } else if (err < 0) {
+ try_async = false;
+ err = 0;
+ }
+ }
+
retry:
req = ceph_mdsc_create_request(mdsc, op, USE_AUTH_MDS);
if (IS_ERR(req)) {
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index 16873d0769..4b8d59ebda 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -366,6 +366,12 @@ int ceph_open(struct inode *inode, struct file *file)
struct ceph_file_info *fi = file->private_data;
int err;
int flags, fmode, wanted;
+ struct dentry *dentry;
+ char *path;
+ int pathlen;
+ u64 pathbase;
+ bool do_sync = false;
+ int mask = MAY_READ;
if (fi) {
doutc(cl, "file %p is already opened\n", file);
@@ -387,6 +393,31 @@ int ceph_open(struct inode *inode, struct file *file)
fmode = ceph_flags_to_mode(flags);
wanted = ceph_caps_for_mode(fmode);
+ if (fmode & CEPH_FILE_MODE_WR)
+ mask |= MAY_WRITE;
+ dentry = d_find_alias(inode);
+ if (!dentry) {
+ do_sync = true;
+ } else {
+ path = ceph_mdsc_build_path(mdsc, dentry, &pathlen, &pathbase, 0);
+ if (IS_ERR(path)) {
+ do_sync = true;
+ err = 0;
+ } else {
+ err = ceph_mds_check_access(mdsc, path, mask);
+ }
+ ceph_mdsc_free_path(path, pathlen);
+ dput(dentry);
+
+ /* For none EACCES cases will let the MDS do the mds auth check */
+ if (err == -EACCES) {
+ return err;
+ } else if (err < 0) {
+ do_sync = true;
+ err = 0;
+ }
+ }
+
/* snapped files are read-only */
if (ceph_snap(inode) != CEPH_NOSNAP && (file->f_mode & FMODE_WRITE))
return -EROFS;
@@ -402,7 +433,7 @@ int ceph_open(struct inode *inode, struct file *file)
* asynchronously.
*/
spin_lock(&ci->i_ceph_lock);
- if (__ceph_is_any_real_caps(ci) &&
+ if (!do_sync && __ceph_is_any_real_caps(ci) &&
(((fmode & CEPH_FILE_MODE_WR) == 0) || ci->i_auth_cap)) {
int mds_wanted = __ceph_caps_mds_wanted(ci, true);
int issued = __ceph_caps_issued(ci, NULL);
@@ -420,7 +451,7 @@ int ceph_open(struct inode *inode, struct file *file)
ceph_check_caps(ci, 0);
return ceph_init_file(inode, file, fmode);
- } else if (ceph_snap(inode) != CEPH_NOSNAP &&
+ } else if (!do_sync && ceph_snap(inode) != CEPH_NOSNAP &&
(ci->i_snap_caps & wanted) == wanted) {
__ceph_touch_fmode(ci, mdsc, fmode);
spin_unlock(&ci->i_ceph_lock);
@@ -759,6 +790,9 @@ int ceph_atomic_open(struct inode *dir, struct dentry *dentry,
bool try_async = ceph_test_mount_opt(fsc, ASYNC_DIROPS);
int mask;
int err;
+ char *path;
+ int pathlen;
+ u64 pathbase;
doutc(cl, "%p %llx.%llx dentry %p '%pd' %s flags %d mode 0%o\n",
dir, ceph_vinop(dir), dentry, dentry,
@@ -776,6 +810,34 @@ int ceph_atomic_open(struct inode *dir, struct dentry *dentry,
*/
flags &= ~O_TRUNC;
+ dn = d_find_alias(dir);
+ if (!dn) {
+ try_async = false;
+ } else {
+ path = ceph_mdsc_build_path(mdsc, dn, &pathlen, &pathbase, 0);
+ if (IS_ERR(path)) {
+ try_async = false;
+ err = 0;
+ } else {
+ int fmode = ceph_flags_to_mode(flags);
+
+ mask = MAY_READ;
+ if (fmode & CEPH_FILE_MODE_WR)
+ mask |= MAY_WRITE;
+ err = ceph_mds_check_access(mdsc, path, mask);
+ }
+ ceph_mdsc_free_path(path, pathlen);
+ dput(dn);
+
+ /* For none EACCES cases will let the MDS do the mds auth check */
+ if (err == -EACCES) {
+ return err;
+ } else if (err < 0) {
+ try_async = false;
+ err = 0;
+ }
+ }
+
retry:
if (flags & O_CREAT) {
if (ceph_quota_is_max_files_exceeded(dir))
diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index 7b2e77517f..249ddfbb1b 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -577,6 +577,8 @@ struct inode *ceph_alloc_inode(struct super_block *sb)
/* Set parameters for the netfs library */
netfs_inode_init(&ci->netfs, &ceph_netfs_ops, false);
+ /* [DEPRECATED] Use PG_private_2 to mark folio being written to the cache. */
+ __set_bit(NETFS_ICTX_USE_PGPRIV2, &ci->netfs.flags);
spin_lock_init(&ci->i_ceph_lock);
@@ -2480,6 +2482,34 @@ int __ceph_setattr(struct mnt_idmap *idmap, struct inode *inode,
bool lock_snap_rwsem = false;
bool fill_fscrypt;
int truncate_retry = 20; /* The RMW will take around 50ms */
+ struct dentry *dentry;
+ char *path;
+ int pathlen;
+ u64 pathbase;
+ bool do_sync = false;
+
+ dentry = d_find_alias(inode);
+ if (!dentry) {
+ do_sync = true;
+ } else {
+ path = ceph_mdsc_build_path(mdsc, dentry, &pathlen, &pathbase, 0);
+ if (IS_ERR(path)) {
+ do_sync = true;
+ err = 0;
+ } else {
+ err = ceph_mds_check_access(mdsc, path, MAY_WRITE);
+ }
+ ceph_mdsc_free_path(path, pathlen);
+ dput(dentry);
+
+ /* For none EACCES cases will let the MDS do the mds auth check */
+ if (err == -EACCES) {
+ return err;
+ } else if (err < 0) {
+ do_sync = true;
+ err = 0;
+ }
+ }
retry:
prealloc_cf = ceph_alloc_cap_flush();
@@ -2526,7 +2556,7 @@ retry:
/* It should never be re-set once set */
WARN_ON_ONCE(ci->fscrypt_auth);
- if (issued & CEPH_CAP_AUTH_EXCL) {
+ if (!do_sync && (issued & CEPH_CAP_AUTH_EXCL)) {
dirtied |= CEPH_CAP_AUTH_EXCL;
kfree(ci->fscrypt_auth);
ci->fscrypt_auth = (u8 *)cia->fscrypt_auth;
@@ -2555,7 +2585,7 @@ retry:
ceph_vinop(inode),
from_kuid(&init_user_ns, inode->i_uid),
from_kuid(&init_user_ns, attr->ia_uid));
- if (issued & CEPH_CAP_AUTH_EXCL) {
+ if (!do_sync && (issued & CEPH_CAP_AUTH_EXCL)) {
inode->i_uid = fsuid;
dirtied |= CEPH_CAP_AUTH_EXCL;
} else if ((issued & CEPH_CAP_AUTH_SHARED) == 0 ||
@@ -2573,7 +2603,7 @@ retry:
ceph_vinop(inode),
from_kgid(&init_user_ns, inode->i_gid),
from_kgid(&init_user_ns, attr->ia_gid));
- if (issued & CEPH_CAP_AUTH_EXCL) {
+ if (!do_sync && (issued & CEPH_CAP_AUTH_EXCL)) {
inode->i_gid = fsgid;
dirtied |= CEPH_CAP_AUTH_EXCL;
} else if ((issued & CEPH_CAP_AUTH_SHARED) == 0 ||
@@ -2587,7 +2617,7 @@ retry:
if (ia_valid & ATTR_MODE) {
doutc(cl, "%p %llx.%llx mode 0%o -> 0%o\n", inode,
ceph_vinop(inode), inode->i_mode, attr->ia_mode);
- if (issued & CEPH_CAP_AUTH_EXCL) {
+ if (!do_sync && (issued & CEPH_CAP_AUTH_EXCL)) {
inode->i_mode = attr->ia_mode;
dirtied |= CEPH_CAP_AUTH_EXCL;
} else if ((issued & CEPH_CAP_AUTH_SHARED) == 0 ||
@@ -2606,11 +2636,11 @@ retry:
inode, ceph_vinop(inode),
atime.tv_sec, atime.tv_nsec,
attr->ia_atime.tv_sec, attr->ia_atime.tv_nsec);
- if (issued & CEPH_CAP_FILE_EXCL) {
+ if (!do_sync && (issued & CEPH_CAP_FILE_EXCL)) {
ci->i_time_warp_seq++;
inode_set_atime_to_ts(inode, attr->ia_atime);
dirtied |= CEPH_CAP_FILE_EXCL;
- } else if ((issued & CEPH_CAP_FILE_WR) &&
+ } else if (!do_sync && (issued & CEPH_CAP_FILE_WR) &&
timespec64_compare(&atime,
&attr->ia_atime) < 0) {
inode_set_atime_to_ts(inode, attr->ia_atime);
@@ -2646,7 +2676,7 @@ retry:
CEPH_FSCRYPT_BLOCK_SIZE));
req->r_fscrypt_file = attr->ia_size;
fill_fscrypt = true;
- } else if ((issued & CEPH_CAP_FILE_EXCL) && attr->ia_size >= isize) {
+ } else if (!do_sync && (issued & CEPH_CAP_FILE_EXCL) && attr->ia_size >= isize) {
if (attr->ia_size > isize) {
i_size_write(inode, attr->ia_size);
inode->i_blocks = calc_inode_blocks(attr->ia_size);
@@ -2683,11 +2713,11 @@ retry:
inode, ceph_vinop(inode),
mtime.tv_sec, mtime.tv_nsec,
attr->ia_mtime.tv_sec, attr->ia_mtime.tv_nsec);
- if (issued & CEPH_CAP_FILE_EXCL) {
+ if (!do_sync && (issued & CEPH_CAP_FILE_EXCL)) {
ci->i_time_warp_seq++;
inode_set_mtime_to_ts(inode, attr->ia_mtime);
dirtied |= CEPH_CAP_FILE_EXCL;
- } else if ((issued & CEPH_CAP_FILE_WR) &&
+ } else if (!do_sync && (issued & CEPH_CAP_FILE_WR) &&
timespec64_compare(&mtime, &attr->ia_mtime) < 0) {
inode_set_mtime_to_ts(inode, attr->ia_mtime);
dirtied |= CEPH_CAP_FILE_WR;
diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
index 360b686c3c..c2157f6e0c 100644
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@ -4112,10 +4112,13 @@ static void handle_session(struct ceph_mds_session *session,
void *p = msg->front.iov_base;
void *end = p + msg->front.iov_len;
struct ceph_mds_session_head *h;
- u32 op;
+ struct ceph_mds_cap_auth *cap_auths = NULL;
+ u32 op, cap_auths_num = 0;
u64 seq, features = 0;
int wake = 0;
bool blocklisted = false;
+ u32 i;
+
/* decode */
ceph_decode_need(&p, end, sizeof(*h), bad);
@@ -4160,7 +4163,101 @@ static void handle_session(struct ceph_mds_session *session,
}
}
+ if (msg_version >= 6) {
+ ceph_decode_32_safe(&p, end, cap_auths_num, bad);
+ doutc(cl, "cap_auths_num %d\n", cap_auths_num);
+
+ if (cap_auths_num && op != CEPH_SESSION_OPEN) {
+ WARN_ON_ONCE(op != CEPH_SESSION_OPEN);
+ goto skip_cap_auths;
+ }
+
+ cap_auths = kcalloc(cap_auths_num,
+ sizeof(struct ceph_mds_cap_auth),
+ GFP_KERNEL);
+ if (!cap_auths) {
+ pr_err_client(cl, "No memory for cap_auths\n");
+ return;
+ }
+
+ for (i = 0; i < cap_auths_num; i++) {
+ u32 _len, j;
+
+ /* struct_v, struct_compat, and struct_len in MDSCapAuth */
+ ceph_decode_skip_n(&p, end, 2 + sizeof(u32), bad);
+
+ /* struct_v, struct_compat, and struct_len in MDSCapMatch */
+ ceph_decode_skip_n(&p, end, 2 + sizeof(u32), bad);
+ ceph_decode_64_safe(&p, end, cap_auths[i].match.uid, bad);
+ ceph_decode_32_safe(&p, end, _len, bad);
+ if (_len) {
+ cap_auths[i].match.gids = kcalloc(_len, sizeof(u32),
+ GFP_KERNEL);
+ if (!cap_auths[i].match.gids) {
+ pr_err_client(cl, "No memory for gids\n");
+ goto fail;
+ }
+
+ cap_auths[i].match.num_gids = _len;
+ for (j = 0; j < _len; j++)
+ ceph_decode_32_safe(&p, end,
+ cap_auths[i].match.gids[j],
+ bad);
+ }
+
+ ceph_decode_32_safe(&p, end, _len, bad);
+ if (_len) {
+ cap_auths[i].match.path = kcalloc(_len + 1, sizeof(char),
+ GFP_KERNEL);
+ if (!cap_auths[i].match.path) {
+ pr_err_client(cl, "No memory for path\n");
+ goto fail;
+ }
+ ceph_decode_copy(&p, cap_auths[i].match.path, _len);
+
+ /* Remove the tailing '/' */
+ while (_len && cap_auths[i].match.path[_len - 1] == '/') {
+ cap_auths[i].match.path[_len - 1] = '\0';
+ _len -= 1;
+ }
+ }
+
+ ceph_decode_32_safe(&p, end, _len, bad);
+ if (_len) {
+ cap_auths[i].match.fs_name = kcalloc(_len + 1, sizeof(char),
+ GFP_KERNEL);
+ if (!cap_auths[i].match.fs_name) {
+ pr_err_client(cl, "No memory for fs_name\n");
+ goto fail;
+ }
+ ceph_decode_copy(&p, cap_auths[i].match.fs_name, _len);
+ }
+
+ ceph_decode_8_safe(&p, end, cap_auths[i].match.root_squash, bad);
+ ceph_decode_8_safe(&p, end, cap_auths[i].readable, bad);
+ ceph_decode_8_safe(&p, end, cap_auths[i].writeable, bad);
+ doutc(cl, "uid %lld, num_gids %u, path %s, fs_name %s, root_squash %d, readable %d, writeable %d\n",
+ cap_auths[i].match.uid, cap_auths[i].match.num_gids,
+ cap_auths[i].match.path, cap_auths[i].match.fs_name,
+ cap_auths[i].match.root_squash,
+ cap_auths[i].readable, cap_auths[i].writeable);
+ }
+ }
+
+skip_cap_auths:
mutex_lock(&mdsc->mutex);
+ if (op == CEPH_SESSION_OPEN) {
+ if (mdsc->s_cap_auths) {
+ for (i = 0; i < mdsc->s_cap_auths_num; i++) {
+ kfree(mdsc->s_cap_auths[i].match.gids);
+ kfree(mdsc->s_cap_auths[i].match.path);
+ kfree(mdsc->s_cap_auths[i].match.fs_name);
+ }
+ kfree(mdsc->s_cap_auths);
+ }
+ mdsc->s_cap_auths_num = cap_auths_num;
+ mdsc->s_cap_auths = cap_auths;
+ }
if (op == CEPH_SESSION_CLOSE) {
ceph_get_mds_session(session);
__unregister_session(mdsc, session);
@@ -4290,6 +4387,13 @@ bad:
pr_err_client(cl, "corrupt message mds%d len %d\n", mds,
(int)msg->front.iov_len);
ceph_msg_dump(msg);
+fail:
+ for (i = 0; i < cap_auths_num; i++) {
+ kfree(cap_auths[i].match.gids);
+ kfree(cap_auths[i].match.path);
+ kfree(cap_auths[i].match.fs_name);
+ }
+ kfree(cap_auths);
return;
}
@@ -5499,6 +5603,170 @@ void send_flush_mdlog(struct ceph_mds_session *s)
mutex_unlock(&s->s_mutex);
}
+static int ceph_mds_auth_match(struct ceph_mds_client *mdsc,
+ struct ceph_mds_cap_auth *auth,
+ char *tpath)
+{
+ const struct cred *cred = get_current_cred();
+ u32 caller_uid = from_kuid(&init_user_ns, cred->fsuid);
+ u32 caller_gid = from_kgid(&init_user_ns, cred->fsgid);
+ struct ceph_client *cl = mdsc->fsc->client;
+ const char *spath = mdsc->fsc->mount_options->server_path;
+ bool gid_matched = false;
+ u32 gid, tlen, len;
+ int i, j;
+
+ doutc(cl, "match.uid %lld\n", auth->match.uid);
+ if (auth->match.uid != MDS_AUTH_UID_ANY) {
+ if (auth->match.uid != caller_uid)
+ return 0;
+ if (auth->match.num_gids) {
+ for (i = 0; i < auth->match.num_gids; i++) {
+ if (caller_gid == auth->match.gids[i])
+ gid_matched = true;
+ }
+ if (!gid_matched && cred->group_info->ngroups) {
+ for (i = 0; i < cred->group_info->ngroups; i++) {
+ gid = from_kgid(&init_user_ns,
+ cred->group_info->gid[i]);
+ for (j = 0; j < auth->match.num_gids; j++) {
+ if (gid == auth->match.gids[j]) {
+ gid_matched = true;
+ break;
+ }
+ }
+ if (gid_matched)
+ break;
+ }
+ }
+ if (!gid_matched)
+ return 0;
+ }
+ }
+
+ /* path match */
+ if (auth->match.path) {
+ if (!tpath)
+ return 0;
+
+ tlen = strlen(tpath);
+ len = strlen(auth->match.path);
+ if (len) {
+ char *_tpath = tpath;
+ bool free_tpath = false;
+ int m, n;
+
+ doutc(cl, "server path %s, tpath %s, match.path %s\n",
+ spath, tpath, auth->match.path);
+ if (spath && (m = strlen(spath)) != 1) {
+ /* mount path + '/' + tpath + an extra space */
+ n = m + 1 + tlen + 1;
+ _tpath = kmalloc(n, GFP_NOFS);
+ if (!_tpath)
+ return -ENOMEM;
+ /* remove the leading '/' */
+ snprintf(_tpath, n, "%s/%s", spath + 1, tpath);
+ free_tpath = true;
+ tlen = strlen(_tpath);
+ }
+
+ /*
+ * Please note the tailing '/' for match.path has already
+ * been removed when parsing.
+ *
+ * Remove the tailing '/' for the target path.
+ */
+ while (tlen && _tpath[tlen - 1] == '/') {
+ _tpath[tlen - 1] = '\0';
+ tlen -= 1;
+ }
+ doutc(cl, "_tpath %s\n", _tpath);
+
+ /*
+ * In case first == _tpath && tlen == len:
+ * match.path=/foo --> /foo _path=/foo --> match
+ * match.path=/foo/ --> /foo _path=/foo --> match
+ *
+ * In case first == _tmatch.path && tlen > len:
+ * match.path=/foo/ --> /foo _path=/foo/ --> match
+ * match.path=/foo --> /foo _path=/foo/ --> match
+ * match.path=/foo/ --> /foo _path=/foo/d --> match
+ * match.path=/foo --> /foo _path=/food --> mismatch
+ *
+ * All the other cases --> mismatch
+ */
+ char *first = strstr(_tpath, auth->match.path);
+ if (first != _tpath) {
+ if (free_tpath)
+ kfree(_tpath);
+ return 0;
+ }
+
+ if (tlen > len && _tpath[len] != '/') {
+ if (free_tpath)
+ kfree(_tpath);
+ return 0;
+ }
+ }
+ }
+
+ doutc(cl, "matched\n");
+ return 1;
+}
+
+int ceph_mds_check_access(struct ceph_mds_client *mdsc, char *tpath, int mask)
+{
+ const struct cred *cred = get_current_cred();
+ u32 caller_uid = from_kuid(&init_user_ns, cred->fsuid);
+ u32 caller_gid = from_kgid(&init_user_ns, cred->fsgid);
+ struct ceph_mds_cap_auth *rw_perms_s = NULL;
+ struct ceph_client *cl = mdsc->fsc->client;
+ bool root_squash_perms = true;
+ int i, err;
+
+ doutc(cl, "tpath '%s', mask %d, caller_uid %d, caller_gid %d\n",
+ tpath, mask, caller_uid, caller_gid);
+
+ for (i = 0; i < mdsc->s_cap_auths_num; i++) {
+ struct ceph_mds_cap_auth *s = &mdsc->s_cap_auths[i];
+
+ err = ceph_mds_auth_match(mdsc, s, tpath);
+ if (err < 0) {
+ return err;
+ } else if (err > 0) {
+ /* always follow the last auth caps' permision */
+ root_squash_perms = true;
+ rw_perms_s = NULL;
+ if ((mask & MAY_WRITE) && s->writeable &&
+ s->match.root_squash && (!caller_uid || !caller_gid))
+ root_squash_perms = false;
+
+ if (((mask & MAY_WRITE) && !s->writeable) ||
+ ((mask & MAY_READ) && !s->readable))
+ rw_perms_s = s;
+ }
+ }
+
+ doutc(cl, "root_squash_perms %d, rw_perms_s %p\n", root_squash_perms,
+ rw_perms_s);
+ if (root_squash_perms && rw_perms_s == NULL) {
+ doutc(cl, "access allowed\n");
+ return 0;
+ }
+
+ if (!root_squash_perms) {
+ doutc(cl, "root_squash is enabled and user(%d %d) isn't allowed to write",
+ caller_uid, caller_gid);
+ }
+ if (rw_perms_s) {
+ doutc(cl, "mds auth caps readable/writeable %d/%d while request r/w %d/%d",
+ rw_perms_s->readable, rw_perms_s->writeable,
+ !!(mask & MAY_READ), !!(mask & MAY_WRITE));
+ }
+ doutc(cl, "access denied\n");
+ return -EACCES;
+}
+
/*
* called before mount is ro, and before dentries are torn down.
* (hmm, does this still race with new lookups?)
diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h
index b88e804152..cfa18cf915 100644
--- a/fs/ceph/mds_client.h
+++ b/fs/ceph/mds_client.h
@@ -35,8 +35,9 @@ enum ceph_feature_type {
CEPHFS_FEATURE_32BITS_RETRY_FWD,
CEPHFS_FEATURE_NEW_SNAPREALM_INFO,
CEPHFS_FEATURE_HAS_OWNER_UIDGID,
+ CEPHFS_FEATURE_MDS_AUTH_CAPS_CHECK,
- CEPHFS_FEATURE_MAX = CEPHFS_FEATURE_HAS_OWNER_UIDGID,
+ CEPHFS_FEATURE_MAX = CEPHFS_FEATURE_MDS_AUTH_CAPS_CHECK,
};
#define CEPHFS_FEATURES_CLIENT_SUPPORTED { \
@@ -52,6 +53,7 @@ enum ceph_feature_type {
CEPHFS_FEATURE_OP_GETVXATTR, \
CEPHFS_FEATURE_32BITS_RETRY_FWD, \
CEPHFS_FEATURE_HAS_OWNER_UIDGID, \
+ CEPHFS_FEATURE_MDS_AUTH_CAPS_CHECK, \
}
/*
@@ -71,6 +73,24 @@ enum ceph_feature_type {
struct ceph_fs_client;
struct ceph_cap;
+#define MDS_AUTH_UID_ANY -1
+
+struct ceph_mds_cap_match {
+ s64 uid; /* default to MDS_AUTH_UID_ANY */
+ u32 num_gids;
+ u32 *gids; /* use these GIDs */
+ char *path; /* require path to be child of this
+ (may be "" or "/" for any) */
+ char *fs_name;
+ bool root_squash; /* default to false */
+};
+
+struct ceph_mds_cap_auth {
+ struct ceph_mds_cap_match match;
+ bool readable;
+ bool writeable;
+};
+
/*
* parsed info about a single inode. pointers are into the encoded
* on-wire structures within the mds reply message payload.
@@ -513,6 +533,9 @@ struct ceph_mds_client {
struct rw_semaphore pool_perm_rwsem;
struct rb_root pool_perm_tree;
+ u32 s_cap_auths_num;
+ struct ceph_mds_cap_auth *s_cap_auths;
+
char nodename[__NEW_UTS_LEN + 1];
};
@@ -581,6 +604,9 @@ extern void ceph_queue_cap_unlink_work(struct ceph_mds_client *mdsc);
extern int ceph_iterate_session_caps(struct ceph_mds_session *session,
int (*cb)(struct inode *, int mds, void *),
void *arg);
+extern int ceph_mds_check_access(struct ceph_mds_client *mdsc, char *tpath,
+ int mask);
+
extern void ceph_mdsc_pre_umount(struct ceph_mds_client *mdsc);
static inline void ceph_mdsc_free_path(char *path, int len)
diff --git a/fs/ceph/super.c b/fs/ceph/super.c
index 885cb5d4e7..0cdf84cd17 100644
--- a/fs/ceph/super.c
+++ b/fs/ceph/super.c
@@ -961,7 +961,8 @@ static int __init init_caches(void)
if (!ceph_mds_request_cachep)
goto bad_mds_req;
- ceph_wb_pagevec_pool = mempool_create_kmalloc_pool(10, CEPH_MAX_WRITE_SIZE >> PAGE_SHIFT);
+ ceph_wb_pagevec_pool = mempool_create_kmalloc_pool(10,
+ (CEPH_MAX_WRITE_SIZE >> PAGE_SHIFT) * sizeof(struct page *));
if (!ceph_wb_pagevec_pool)
goto bad_pagevec_pool;