If ceph_atomic_open is handed a !d_in_lookup dentry, then that means
that it already passed d_revalidate so we *know* what the state of the
thing was just recently. Just bail out at that point and let the caller
handle that case.
This also addresses a subtle bug in dentry handling. Non-O_CREAT opens
call atomic_open with the parent's i_rwsem shared, but calling
d_splice_alias on a hashed dentry requires the exclusive lock.
ceph_atomic_open could receive a hashed, negative dentry on a
non-O_CREAT open. If another client were to race in and create the file
before we issue our OPEN, ceph_fill_trace could end up calling
d_splice_alias on the dentry with the new inode.
Suggested-by: Al Viro <viro@xxxxxxxxxxxxxxxxxx>
Signed-off-by: Jeff Layton <jlayton@xxxxxxxxxx>
---
fs/ceph/file.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index d277f71abe0b..691b7b1a6075 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -462,6 +462,8 @@ int ceph_atomic_open(struct inode *dir, struct dentry *dentry,
err = ceph_security_init_secctx(dentry, mode, &as_ctx);
if (err < 0)
goto out_ctx;
+ } else if (!d_in_lookup(dentry)) {
+ return finish_no_open(file, dentry);
}
/* do the open */
--
2.21.0
