ceph: fix dentry rehashing on virtual .snap dir

If a lookup fails on the magic .snap directory, we bind it to a magic
snap directory inode in ceph_lookup_finish().  That code assumes the dentry
is unhashed, but a recent server-side change started returning NULL leases
on lookup failure, causing the .snap dentry to be hashed and NULL by
ceph_fill_trace().

This causes dentry hash chain corruption, or a dies when d_rehash()
includes
	BUG_ON(!d_unhashed(entry));

So, avoid processing the NULL dentry lease if it the dentry matches the
snapdir name in ceph_fill_trace().  That allows the lookup completion to
properly bind it to the snapdir inode.  BUG there if dentry is hashed to
be sure.

Signed-off-by: Sage Weil <sage@newdream.net>

2 files changed, 10 insertions, 1 deletions

diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c
index 8a9116e15b7..aed8fda3302 100644
--- a/fs/ceph/dir.c
+++ b/fs/ceph/dir.c
@@ -488,6 +488,7 @@ struct dentry *ceph_finish_lookup(struct ceph_mds_request *req,
                 struct inode *inode = ceph_get_snapdir(parent);
                 dout("ENOENT on snapdir %p '%.*s', linking to snapdir %p\n",
                      dentry, dentry->d_name.len, dentry->d_name.name, inode);
+                BUG_ON(!d_unhashed(dentry));
                 d_add(dentry, inode);
                 err = 0;
         }
diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index aca82d55cc5..26f883c275e 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -886,6 +886,7 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req,
         struct inode *in = NULL;
         struct ceph_mds_reply_inode *ininfo;
         struct ceph_vino vino;
+        struct ceph_client *client = ceph_sb_to_client(sb);
         int i = 0;
         int err = 0;
 
@@ -949,7 +950,14 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req,
                         return err;
         }
 
-        if (rinfo->head->is_dentry && !req->r_aborted) {
+        /*
+         * ignore null lease/binding on snapdir ENOENT, or else we
+         * will have trouble splicing in the virtual snapdir later
+         */
+        if (rinfo->head->is_dentry && !req->r_aborted &&
+            (rinfo->head->is_target || strncmp(req->r_dentry->d_name.name,
+                                               client->mount_args->snapdir_name,
+                                               req->r_dentry->d_name.len))) {
                 /*
                  * lookup link rename   : null -> possibly existing inode
                  * mknod symlink mkdir  : null -> new inode