Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1328488

Re: [PATCH] NVMe: do not touch sq door bell if nvmeq has been suspended

From Sagi Grimberg <sagig@dev.mellanox.co.il>
Newsgroups linux.kernel
Subject Re: [PATCH] NVMe: do not touch sq door bell if nvmeq has been suspended
Date 2016-02-07 14:50 +0100
Message-ID <qZxSW-6QG-5@gated-at.bofh.it> (permalink)
References (2 earlier) <qXIza-2Za-31@gated-at.bofh.it> <qY6UO-3c5-9@gated-at.bofh.it> <qY8Df-4ol-3@gated-at.bofh.it> <qY8Df-4ol-1@gated-at.bofh.it> <qZxSW-6QG-7@gated-at.bofh.it>
Organization linux.* mail to news gateway

Show all headers | View raw


> Keith,
>
> Is the following solution OK?
> synchronize_rcu guarantee that no queue_rq is running concurrently with device disable code.
> Together with your another patch (adding blk_sync_queue), both sync/async path shall be handled correctly.

This can be acceptable I think.

> Do you think synchronize_rcu shall be added to blk_sync_queue?

Or, we'll add it to blk_mq_stop_hw_queues() and then scsi
will enjoy it as well.

> diff --git a/block/blk-mq.c b/block/blk-mq.c
> index 4c0622f..bfe9132 100644
> --- a/block/blk-mq.c
> +++ b/block/blk-mq.c
> @@ -865,7 +865,9 @@ void blk_mq_run_hw_queue(struct blk_mq_hw_ctx *hctx, bool async)
>          if (!async) {
>                  int cpu = get_cpu();
>                  if (cpumask_test_cpu(cpu, hctx->cpumask)) {
> +                       rcu_read_lock();
>                          __blk_mq_run_hw_queue(hctx);
> +                       rcu_read_unlock();

I think the rcu is better folded into __blk_mq_run_hw_queue
to cover all the call sites.

Back to linux.kernel | Previous | NextPrevious in thread | Next in thread | Find similar | Unroll thread


Thread

[PATCH] NVMe: do not touch sq door bell if nvmeq has been suspended Wenbo Wang <mail_weber_wang@163.com> - 2016-02-01 16:50 +0100
  RE: [PATCH] NVMe: do not touch sq door bell if nvmeq has been  suspended "Busch, Keith" <keith.busch@intel.com> - 2016-02-01 17:00 +0100
    RE: [PATCH] NVMe: do not touch sq door bell if nvmeq has been  suspended Wenbo Wang <wenbo.wang@memblaze.com> - 2016-02-01 17:20 +0100
  Re: [PATCH] NVMe: do not touch sq door bell if nvmeq has been  suspended Jens Axboe <axboe@fb.com> - 2016-02-01 18:00 +0100
    Re: [PATCH] NVMe: do not touch sq door bell if nvmeq has been  suspended Sagi Grimberg <sagig@dev.mellanox.co.il> - 2016-02-02 13:50 +0100
      Re: [PATCH] NVMe: do not touch sq door bell if nvmeq has been  suspended Keith Busch <keith.busch@intel.com> - 2016-02-02 15:30 +0100
        Re: [PATCH] NVMe: do not touch sq door bell if nvmeq has been  suspended Sagi Grimberg <sagig@dev.mellanox.co.il> - 2016-02-02 15:40 +0100
          Re: [PATCH] NVMe: do not touch sq door bell if nvmeq has been  suspended Keith Busch <keith.busch@intel.com> - 2016-02-02 15:50 +0100
            Re: [PATCH] NVMe: do not touch sq door bell if nvmeq has been  suspended Sagi Grimberg <sagig@dev.mellanox.co.il> - 2016-02-02 18:30 +0100
    Re: [PATCH] NVMe: do not touch sq door bell if nvmeq has been  suspended Keith Busch <keith.busch@intel.com> - 2016-02-02 18:30 +0100
    Re: [PATCH] NVMe: do not touch sq door bell if nvmeq has been  suspended Keith Busch <keith.busch@intel.com> - 2016-02-03 15:50 +0100
      Re: [PATCH] NVMe: do not touch sq door bell if nvmeq has been  suspended Keith Busch <keith.busch@intel.com> - 2016-02-03 17:40 +0100
        Re: [PATCH] NVMe: do not touch sq door bell if nvmeq has been  suspended Sagi Grimberg <sagig@dev.mellanox.co.il> - 2016-02-07 14:50 +0100
        Re: [PATCH] NVMe: do not touch sq door bell if nvmeq has been  suspended Keith Busch <keith.busch@intel.com> - 2016-02-08 16:10 +0100

csiph-web