be2net: call napi_disable() for all event queues
authorIvan Vecera <ivecera@redhat.com>
Wed, 27 Nov 2013 07:59:32 +0000 (08:59 +0100)
committerDavid S. Miller <davem@davemloft.net>
Thu, 28 Nov 2013 23:54:02 +0000 (18:54 -0500)
The recent be2net commit 6384a4d (adds a support for busy polling)
introduces a regression that results in kernel crash. It incorrectly
modified be_close() so napi_disable() is called only for the first queue.
This breaks a correct pairing of napi_enable/_disable for the rest
of event queues and causes a crash in subsequent be_open() call.

v2: Applied suggestions from Sathya

Fixes: 6384a4d ("be2net: add support for ndo_busy_poll")
Cc: Sathya Perla <sathya.perla@emulex.com>
Cc: Subbu Seetharaman <subbu.seetharaman@emulex.com>
Cc: Ajit Khaparde <ajit.khaparde@emulex.com>
Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Acked-by: Sathya Perla <sathya.perla@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/net/ethernet/emulex/benet/be_main.c

index 78a0e859760b8bc6c75e4b67ea919776a23bc674..fee64bf10446092fa718caa12819e1be65ba0d6a 100644 (file)
@@ -2658,8 +2658,8 @@ static int be_close(struct net_device *netdev)
 
        be_roce_dev_close(adapter);
 
-       for_all_evt_queues(adapter, eqo, i) {
-               if (adapter->flags & BE_FLAGS_NAPI_ENABLED) {
+       if (adapter->flags & BE_FLAGS_NAPI_ENABLED) {
+               for_all_evt_queues(adapter, eqo, i) {
                        napi_disable(&eqo->napi);
                        be_disable_busy_poll(eqo);
                }