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

Groups > linux.kernel > #1716498 > unrolled thread

[PATCH 2/2] blk-mq: don't use WORK_CPU_UNBOUND

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  [PATCH 2/2] blk-mq: don't use WORK_CPU_UNBOUND Laurent Vivier <lvivier@redhat.com> - 2017-08-21 16:00 +0200
    Re: [PATCH 2/2] blk-mq: don't use WORK_CPU_UNBOUND Tejun Heo <tj@kernel.org> - 2017-08-21 16:50 +0200

#1716498 — [PATCH 2/2] blk-mq: don't use WORK_CPU_UNBOUND

cpumask is the list of CPUs present when the queue is built.
If a new CPU is hotplugged, this list is not updated,
and when the scheduler asks for a CPU id, blk_mq_hctx_next_cpu()
can return WORK_CPU_UNBOUND.
And in this case _blk_mq_run_hw_queue() can be executed by the new CPU
(that is not present in cpumask) and raises the following warning:

    WARN_ON(!cpumask_test_cpu(raw_smp_processor_id(), hctx->cpumask) &&
            cpu_online(hctx->next_cpu));

To fix this problem, this patch modifies blk_mq_hctx_next_cpu()
to only return a CPU id present in cpumask.

Signed-off-by: Laurent Vivier <lvivier@redhat.com>
---
 block/blk-mq.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/block/blk-mq.c b/block/blk-mq.c
index 4603b115e234..bdac1e654814 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -1126,9 +1126,6 @@ static void __blk_mq_run_hw_queue(struct blk_mq_hw_ctx *hctx)
  */
 static int blk_mq_hctx_next_cpu(struct blk_mq_hw_ctx *hctx)
 {
-	if (hctx->queue->nr_hw_queues == 1)
-		return WORK_CPU_UNBOUND;
-
 	if (--hctx->next_cpu_batch <= 0) {
 		int next_cpu;
 
-- 
2.13.5

[toc] | [next] | [standalone]

#1716568

On Mon, Aug 21, 2017 at 03:49:51PM +0200, Laurent Vivier wrote:
> cpumask is the list of CPUs present when the queue is built.
> If a new CPU is hotplugged, this list is not updated,
> and when the scheduler asks for a CPU id, blk_mq_hctx_next_cpu()
> can return WORK_CPU_UNBOUND.
> And in this case _blk_mq_run_hw_queue() can be executed by the new CPU
> (that is not present in cpumask) and raises the following warning:
> 
>     WARN_ON(!cpumask_test_cpu(raw_smp_processor_id(), hctx->cpumask) &&
>             cpu_online(hctx->next_cpu));
> 
> To fix this problem, this patch modifies blk_mq_hctx_next_cpu()
> to only return a CPU id present in cpumask.

So, this one isn't needed either.

Thanks.

-- 
tejun

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web