[SCSI] bnx2fc: fix panic in bnx2fc_post_io_req
authorBhanu Prakash Gollapudi <bprakash@broadcom.com>
Tue, 24 Jan 2012 02:00:46 +0000 (18:00 -0800)
committerJames Bottomley <JBottomley@Parallels.com>
Sun, 19 Feb 2012 14:08:58 +0000 (08:08 -0600)
System panics while accessing stale pointer - timer_work_queue - in the IO path
before bnx2fc_stop is called. Fix is to destroy the workqueue after the destroy
operation is complete.

Signed-off-by: Bhanu Prakash Gollapudi <bprakash@broadcom.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
drivers/scsi/bnx2fc/bnx2fc_fcoe.c

index 8c6156a10d90233a45bee5f27f343a33c7311929..10cf5d96ce296a9eed6f86565972784da4aba15a 100644 (file)
@@ -1537,6 +1537,7 @@ static void __bnx2fc_destroy(struct bnx2fc_interface *interface)
 static int bnx2fc_destroy(struct net_device *netdev)
 {
        struct bnx2fc_interface *interface = NULL;
+       struct workqueue_struct *timer_work_queue;
        int rc = 0;
 
        rtnl_lock();
@@ -1549,9 +1550,9 @@ static int bnx2fc_destroy(struct net_device *netdev)
                goto netdev_err;
        }
 
-
-       destroy_workqueue(interface->timer_work_queue);
+       timer_work_queue = interface->timer_work_queue;
        __bnx2fc_destroy(interface);
+       destroy_workqueue(timer_work_queue);
 
 netdev_err:
        mutex_unlock(&bnx2fc_dev_lock);