nvme: use blk_mq_start_hw_queues() in nvme_kill_queues()
authorMing Lei <ming.lei@redhat.com>
Mon, 22 May 2017 15:05:03 +0000 (23:05 +0800)
committerChristoph Hellwig <hch@lst.de>
Mon, 22 May 2017 18:50:09 +0000 (20:50 +0200)
Inside nvme_kill_queues(), we have to start hw queues for
draining requests in sw queues, .dispatch list and requeue list,
so use blk_mq_start_hw_queues() instead of blk_mq_start_stopped_hw_queues()
which only run queues if queues are stopped, but the queues may have
been started already, for example nvme_start_queues() is called in reset work
function.

blk_mq_start_hw_queues() run hw queues in current context, instead
of running asynchronously like before. Given nvme_kill_queues() is
run from either remove context or reset worker context, both are fine
to run hw queue directly. And the mutex of namespaces_mutex isn't a
problem too becasue nvme_start_freeze() runs hw queue in this way
already.

Cc: stable@vger.kernel.org
Reported-by: Zhang Yi <yizhan@redhat.com>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
drivers/nvme/host/core.c

index d5e0906262ead5dd9f202385d7ace81444c25a0f..40d5e4a9e8d7b537967dcc4fb745a3959dd2b8d7 100644 (file)
@@ -2437,7 +2437,13 @@ void nvme_kill_queues(struct nvme_ctrl *ctrl)
                revalidate_disk(ns->disk);
                blk_set_queue_dying(ns->queue);
                blk_mq_abort_requeue_list(ns->queue);
-               blk_mq_start_stopped_hw_queues(ns->queue, true);
+
+               /*
+                * Forcibly start all queues to avoid having stuck requests.
+                * Note that we must ensure the queues are not stopped
+                * when the final removal happens.
+                */
+               blk_mq_start_hw_queues(ns->queue);
        }
        mutex_unlock(&ctrl->namespaces_mutex);
 }