diff mbox

[11/12] nvme: Use BLK_MQ_S_STOPPED instead of QUEUE_FLAG_STOPPED in blk-mq code

Message ID 805e1911-cd10-0563-c76b-256d76054b08@sandisk.com (mailing list archive)
State Not Applicable
Headers show

Commit Message

Bart Van Assche Oct. 26, 2016, 10:56 p.m. UTC 
Make nvme_requeue_req() check BLK_MQ_S_STOPPED instead of
QUEUE_FLAG_STOPPED. Remove the QUEUE_FLAG_STOPPED manipulations
that became superfluous because of this change. Change
blk_queue_stopped() tests into blk_mq_queue_stopped().

This patch fixes a race condition: using queue_flag_clear_unlocked()
is not safe if any other function that manipulates the queue flags
can be called concurrently, e.g. blk_cleanup_queue().

Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Cc: Keith Busch <keith.busch@intel.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Sagi Grimberg <sagi@grimberg.me>
---
 drivers/nvme/host/core.c | 7 +------
 1 file changed, 1 insertion(+), 6 deletions(-)

Comments

Sagi Grimberg Oct. 27, 2016, 12:19 p.m. UTC | #1 
Looks good,

Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Keith Busch Oct. 28, 2016, 4:01 p.m. UTC | #2 
On Wed, Oct 26, 2016 at 03:56:04PM -0700, Bart Van Assche wrote:
> diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
> index 7bb73ba..b662416 100644
> --- a/drivers/nvme/host/core.c
> +++ b/drivers/nvme/host/core.c
> @@ -205,7 +205,7 @@ void nvme_requeue_req(struct request *req)
>  
>  	blk_mq_requeue_request(req, false);
>  	spin_lock_irqsave(req->q->queue_lock, flags);
> -	if (!blk_queue_stopped(req->q))
> +	if (!blk_mq_queue_stopped(req->q))
>  		blk_mq_kick_requeue_list(req->q);
>  	spin_unlock_irqrestore(req->q->queue_lock, flags);
>  }
> @@ -2079,10 +2079,6 @@ void nvme_stop_queues(struct nvme_ctrl *ctrl)
>  
>  	mutex_lock(&ctrl->namespaces_mutex);
>  	list_for_each_entry(ns, &ctrl->namespaces, list) {
> -		spin_lock_irq(ns->queue->queue_lock);
> -		queue_flag_set(QUEUE_FLAG_STOPPED, ns->queue);
> -		spin_unlock_irq(ns->queue->queue_lock);
> -
>  		blk_mq_cancel_requeue_work(ns->queue);
>  		blk_mq_stop_hw_queues(ns->queue);

There's actually a reason the queue stoppage is using a different flag
than blk_mq_queue_stopped: kicking the queue starts stopped queues.
The driver has to ensure the requeue work can't be kicked prior to
cancelling the current requeue work. Once we know requeue work isn't
running and can't restart again, then we're safe to stop the hw queues.

It's a pretty obscure condition, requiring the controller post an
error completion at the same time the driver decides to reset the
controller. Here's the sequence with the wrong outcome:

 CPU A                   CPU B
 -----                   -----
nvme_stop_queues         nvme_requeue_req
 blk_mq_stop_hw_queues    if (blk_mq_queue_stopped) <- returns false
  blk_mq_stop_hw_queue     blk_mq_kick_requeue_list
                         blk_mq_requeue_work
                          blk_mq_start_hw_queues
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index 7bb73ba..b662416 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -205,7 +205,7 @@  void nvme_requeue_req(struct request *req)
 
 	blk_mq_requeue_request(req, false);
 	spin_lock_irqsave(req->q->queue_lock, flags);
-	if (!blk_queue_stopped(req->q))
+	if (!blk_mq_queue_stopped(req->q))
 		blk_mq_kick_requeue_list(req->q);
 	spin_unlock_irqrestore(req->q->queue_lock, flags);
 }
@@ -2079,10 +2079,6 @@  void nvme_stop_queues(struct nvme_ctrl *ctrl)
 
 	mutex_lock(&ctrl->namespaces_mutex);
 	list_for_each_entry(ns, &ctrl->namespaces, list) {
-		spin_lock_irq(ns->queue->queue_lock);
-		queue_flag_set(QUEUE_FLAG_STOPPED, ns->queue);
-		spin_unlock_irq(ns->queue->queue_lock);
-
 		blk_mq_cancel_requeue_work(ns->queue);
 		blk_mq_stop_hw_queues(ns->queue);
 	}
@@ -2096,7 +2092,6 @@  void nvme_start_queues(struct nvme_ctrl *ctrl)
 
 	mutex_lock(&ctrl->namespaces_mutex);
 	list_for_each_entry(ns, &ctrl->namespaces, list) {
-		queue_flag_clear_unlocked(QUEUE_FLAG_STOPPED, ns->queue);
 		blk_mq_start_stopped_hw_queues(ns->queue, true);
 		blk_mq_kick_requeue_list(ns->queue);
 	}