ceph: fix security xattr deadlock

When security is enabled, security module can call filesystem's
getxattr/setxattr callbacks during d_instantiate(). For cephfs,
d_instantiate() is usually called by MDS' dispatch thread, while
handling MDS reply. If the MDS reply does not include xattrs and
corresponding caps, getxattr/setxattr need to send a new request
to MDS and waits for the reply. This makes MDS' dispatch sleep,
nobody handles later MDS replies.

The fix is make sure lookup/atomic_open reply include xattrs and
corresponding caps. So getxattr can be handled by cached xattrs.
This requires some modification to both MDS and request message.
(Client tells MDS what caps it wants; MDS encodes proper caps in
the reply)

Smack security module may call setxattr during d_instantiate().
Unlike getxattr, we can't force MDS to issue CEPH_CAP_XATTR_EXCL
to us. So just make setxattr return error when called by MDS'
dispatch thread.

Signed-off-by: Yan, Zheng <zyan@redhat.com>

1 files changed, 14 insertions, 4 deletions

diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index 66edef1..8b136dc 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -1389,7 +1389,7 @@ int ceph_readdir_prepopulate(struct ceph_mds_request *req,
 	struct qstr dname;
 	struct dentry *dn;
 	struct inode *in;
-	int err = 0, ret, i;
+	int err = 0, skipped = 0, ret, i;
 	struct inode *snapdir = NULL;
 	struct ceph_mds_request_head *rhead = req->r_request->front.iov_base;
 	struct ceph_dentry_info *di;
@@ -1501,7 +1501,17 @@ retry_lookup:
 		}
 
 		if (d_really_is_negative(dn)) {
-			struct dentry *realdn = splice_dentry(dn, in);
+			struct dentry *realdn;
+
+			if (ceph_security_xattr_deadlock(in)) {
+				dout(" skip splicing dn %p to inode %p"
+				     " (security xattr deadlock)\n", dn, in);
+				iput(in);
+				skipped++;
+				goto next_item;
+			}
+
+			realdn = splice_dentry(dn, in);
 			if (IS_ERR(realdn)) {
 				err = PTR_ERR(realdn);
 				d_drop(dn);
@@ -1518,7 +1528,7 @@ retry_lookup:
 				    req->r_session,
 				    req->r_request_started);
 
-		if (err == 0 && cache_ctl.index >= 0) {
+		if (err == 0 && skipped == 0 && cache_ctl.index >= 0) {
 			ret = fill_readdir_cache(d_inode(parent), dn,
 						 &cache_ctl, req);
 			if (ret < 0)
@@ -1529,7 +1539,7 @@ next_item:
 			dput(dn);
 	}
 out:
-	if (err == 0) {
+	if (err == 0 && skipped == 0) {
 		req->r_did_prepopulate = true;
 		req->r_readdir_cache_idx = cache_ctl.index;
 	}