[SCSI] qla2xxx: Fix for a multiqueue bug in CPU affinity mode

Hold the hardware lock while do the response completion in work queue threads as
it involves sharing a common request queue among multiple threads.

Signed-off-by: Anirban Chakraborty <anirban.chakraborty@qlogic.com>
Signed-off-by: Giridhar Malavali <giridhar.malavali@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

diff --git a/drivers/scsi/qla2xxx/qla_mid.c b/drivers/scsi/qla2xxx/qla_mid.c
index 2a4c7f4e7b69b0a4fcaf13cf044044c60639a3f6..b901aa267e7d35e5fb1530056ceedab8d2c75b25 100644 (file)
--- a/drivers/scsi/qla2xxx/qla_mid.c
+++ b/drivers/scsi/qla2xxx/qla_mid.c
@@ -639,8 +639,10 @@ static void qla_do_work(struct work_struct *work)
        struct rsp_que *rsp = container_of(work, struct rsp_que, q_work);
        struct scsi_qla_host *vha;
 
+       spin_lock_irq(&rsp->hw->hardware_lock);
        vha = qla25xx_get_host(rsp);
        qla24xx_process_response_queue(vha, rsp);
+       spin_unlock_irq(&rsp->hw->hardware_lock);
 }
 
 /* create response queue */