NLM: Don't hang forever on NLM unlock requests
If the NLM daemon is killed on the NFS server, we can currently end up
hanging forever on an 'unlock' request, instead of aborting. Basically,
if the rpcbind request fails, or the server keeps returning garbage, we
really want to quit instead of retrying.
Tested-by: Vasily Averin <vvs@sw.ru>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: stable@kernel.org
diff --git a/fs/lockd/clntproc.c b/fs/lockd/clntproc.c
index adb45ec..e374050 100644
--- a/fs/lockd/clntproc.c
+++ b/fs/lockd/clntproc.c
@@ -708,7 +708,13 @@
if (task->tk_status < 0) {
dprintk("lockd: unlock failed (err = %d)\n", -task->tk_status);
- goto retry_rebind;
+ switch (task->tk_status) {
+ case -EACCES:
+ case -EIO:
+ goto die;
+ default:
+ goto retry_rebind;
+ }
}
if (status == NLM_LCK_DENIED_GRACE_PERIOD) {
rpc_delay(task, NLMCLNT_GRACE_WAIT);
diff --git a/include/linux/sunrpc/sched.h b/include/linux/sunrpc/sched.h
index f73c482..fe2d8e6 100644
--- a/include/linux/sunrpc/sched.h
+++ b/include/linux/sunrpc/sched.h
@@ -84,7 +84,8 @@
#endif
unsigned char tk_priority : 2,/* Task priority */
tk_garb_retry : 2,
- tk_cred_retry : 2;
+ tk_cred_retry : 2,
+ tk_rebind_retry : 2;
};
#define tk_xprt tk_client->cl_xprt
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index b84d739..566bcfd 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -1175,6 +1175,9 @@
status = -EOPNOTSUPP;
break;
}
+ if (task->tk_rebind_retry == 0)
+ break;
+ task->tk_rebind_retry--;
rpc_delay(task, 3*HZ);
goto retry_timeout;
case -ETIMEDOUT:
diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c
index 6b43ee7..a27406b 100644
--- a/net/sunrpc/sched.c
+++ b/net/sunrpc/sched.c
@@ -792,6 +792,7 @@
/* Initialize retry counters */
task->tk_garb_retry = 2;
task->tk_cred_retry = 2;
+ task->tk_rebind_retry = 2;
task->tk_priority = task_setup_data->priority - RPC_PRIORITY_LOW;
task->tk_owner = current->tgid;