Re: [PATCH 1/6] blk-mq: make blk_mq_stop_hw_queue() reliably block queue processing

[Ming Lei]

On Wed, Jun 18, 2014 at 11:21 PM, Tejun Heo <tj@xxxxxxxxxx> wrote:
> blk_mq_stop_hw_queue() has the following two issues.
>
> * BLK_MQ_S_STOPPED should be set before canceling the work items;
>   otherwise, a new instance may proceed beyond STOPPED checking
>   inbetween.
>
> * cancel_delayed_work() doesn't do anything to work instance already
>   executing.  Use cancel_delayed_work_sync() to ensure that the
>   currently in-flight one is finished.
>
> Signed-off-by: Tejun Heo <tj@xxxxxxxxxx>
> Cc: Jens Axboe <axboe@xxxxxxxxx>
> Cc: Nicholas A. Bellinger <nab@xxxxxxxxxxxxxxx>
> ---
>  block/blk-mq.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/block/blk-mq.c b/block/blk-mq.c
> index e11f5f8..4787c3d 100644
> --- a/block/blk-mq.c
> +++ b/block/blk-mq.c
> @@ -862,9 +862,9 @@ EXPORT_SYMBOL(blk_mq_run_queues);
>
>  void blk_mq_stop_hw_queue(struct blk_mq_hw_ctx *hctx)
>  {
> -       cancel_delayed_work(&hctx->run_work);
> -       cancel_delayed_work(&hctx->delay_work);
>         set_bit(BLK_MQ_S_STOPPED, &hctx->state);
> +       cancel_delayed_work_sync(&hctx->run_work);
> +       cancel_delayed_work_sync(&hctx->delay_work);

The function can be called from atomic context, like virtio-blk.

Currently blk_mq_stop_hw_queue() is called in case of queue
being busy, so looks it is ok with cancel_delayed_work().


Thanks,
-- 
Ming Lei
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/