NLM: Ensure that we resend all pending blocking locks after a reclaim
authorTrond Myklebust <Trond.Myklebust@netapp.com>
Tue, 19 Feb 2013 17:04:42 +0000 (12:04 -0500)
committerTrond Myklebust <Trond.Myklebust@netapp.com>
Tue, 19 Feb 2013 17:18:27 +0000 (12:18 -0500)
Currently, nlmclnt_lock will break out of the for(;;) loop when
the reclaimer wakes up the blocking lock thread by setting
nlm_lck_denied_grace_period. This causes the lock request to fail
with an ENOLCK error.
The intention was always to ensure that we resend the lock request
after the grace period has expired.

Reported-by: Wangyuan Zhang <Wangyuan.Zhang@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: stable@vger.kernel.org
fs/lockd/clntproc.c

index 54f9e6ce0430ae88d742709bc47499c0d1b71731..52e5120bb159f14cd286f91ed2f72848f8aeb5d9 100644 (file)
@@ -550,6 +550,9 @@ again:
                status = nlmclnt_block(block, req, NLMCLNT_POLL_TIMEOUT);
                if (status < 0)
                        break;
                status = nlmclnt_block(block, req, NLMCLNT_POLL_TIMEOUT);
                if (status < 0)
                        break;
+               /* Resend the blocking lock request after a server reboot */
+               if (resp->status ==  nlm_lck_denied_grace_period)
+                       continue;
                if (resp->status != nlm_lck_blocked)
                        break;
        }
                if (resp->status != nlm_lck_blocked)
                        break;
        }