[SCSI] libfcoe: fcoe_ctlr_destroy use cancel_work_sync instead of flush_work
authorChris Leech <christopher.leech@intel.com>
Tue, 25 Aug 2009 20:59:08 +0000 (13:59 -0700)
committerJames Bottomley <James.Bottomley@suse.de>
Thu, 10 Sep 2009 17:07:28 +0000 (12:07 -0500)
Use cancel_work_sync() in place of flush_work(), so that
fcoe_ctlr_destroy() can be called from a workqueue.

Also, purge the receive queue after the recv_work has been cancled because
if recv_work isn't run it's not guaranteed to be empty now.

Signed-off-by: Chris Leech <christopher.leech@intel.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
drivers/scsi/fcoe/libfcoe.c

index d0b5208e1a45b43d5c41f20fa83ec99c858732f7..62a4c202607202fa81f9229b215d4f0ed5fad3b6 100644 (file)
@@ -148,13 +148,17 @@ static void fcoe_ctlr_reset_fcfs(struct fcoe_ctlr *fip)
  */
 void fcoe_ctlr_destroy(struct fcoe_ctlr *fip)
 {
-       flush_work(&fip->recv_work);
+       cancel_work_sync(&fip->recv_work);
+       spin_lock_bh(&fip->fip_recv_list.lock);
+       __skb_queue_purge(&fip->fip_recv_list);
+       spin_unlock_bh(&fip->fip_recv_list.lock);
+
        spin_lock_bh(&fip->lock);
        fip->state = FIP_ST_DISABLED;
        fcoe_ctlr_reset_fcfs(fip);
        spin_unlock_bh(&fip->lock);
        del_timer_sync(&fip->timer);
-       flush_work(&fip->link_work);
+       cancel_work_sync(&fip->link_work);
 }
 EXPORT_SYMBOL(fcoe_ctlr_destroy);