diff options
| author | Trond Myklebust <trond.myklebust@primarydata.com> | 2017-05-05 13:02:42 -0400 |
|---|---|---|
| committer | Trond Myklebust <trond.myklebust@primarydata.com> | 2017-05-08 09:27:59 -0400 |
| commit | 2e84611b3f4fa50e1f4c12f2966fcc7fb955d944 (patch) | |
| tree | 07d5defa8924f2fa10e2fa0b32c1cfff877cb931 /fs | |
| parent | 0795bf8357c1887e2a95e6e4f5b89d0896a0d929 (diff) | |
NFSv4: Fix an rcu lock leak
The intention in the original patch was to release the lock when
we put the inode, however something got screwed up.
Reported-by: Jason Yan <yanaijie@huawei.com>
Fixes: 7b410d9ce460f ("pNFS: Delay getting the layout header in..")
Cc: stable@vger.kernel.org # v4.10+
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Diffstat (limited to 'fs')
| -rw-r--r-- | fs/nfs/callback_proc.c | 6 |
1 files changed, 4 insertions, 2 deletions
diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c index e7f041447afd..52479f180ea1 100644 --- a/fs/nfs/callback_proc.c +++ b/fs/nfs/callback_proc.c @@ -131,10 +131,11 @@ restart: if (!inode) continue; if (!nfs_sb_active(inode->i_sb)) { - rcu_read_lock(); + rcu_read_unlock(); spin_unlock(&clp->cl_lock); iput(inode); spin_lock(&clp->cl_lock); + rcu_read_lock(); goto restart; } return inode; @@ -170,10 +171,11 @@ restart: if (!inode) continue; if (!nfs_sb_active(inode->i_sb)) { - rcu_read_lock(); + rcu_read_unlock(); spin_unlock(&clp->cl_lock); iput(inode); spin_lock(&clp->cl_lock); + rcu_read_lock(); goto restart; } return inode; |