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


Groups > linux.kernel > #1486723

Re: [PATCH 04/14] blk-mq: Do not limit number of queues to 'nr_cpu_ids' in allocations

From Omar Sandoval <osandov@osandov.com>
Newsgroups linux.kernel
Subject Re: [PATCH 04/14] blk-mq: Do not limit number of queues to 'nr_cpu_ids' in allocations
Date 2016-09-19 19:50 +0200
Message-ID <sjaRA-8rr-11@gated-at.bofh.it> (permalink)
References <siERH-5bE-3@gated-at.bofh.it> <siERH-5bE-11@gated-at.bofh.it>
Organization linux.* mail to news gateway

Show all headers | View raw


On Sun, Sep 18, 2016 at 09:37:14AM +0200, Alexander Gordeev wrote:
> Currently maximum number of used hardware queues is limited to
> number of CPUs in the system. However, using 'nr_cpu_ids' as
> the limit for (de-)allocations of data structures instead of
> existing data structures' counters (a) worsens readability and
> (b) leads to unused memory when number of hardware queues is
> less than number of CPUs.
> 
> CC: linux-block@vger.kernel.org
> Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
> ---
>  block/blk-mq.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/block/blk-mq.c b/block/blk-mq.c
> index 276ec7b..2c77b68 100644
> --- a/block/blk-mq.c
> +++ b/block/blk-mq.c
> @@ -2054,8 +2054,8 @@ struct request_queue *blk_mq_init_allocated_queue(struct blk_mq_tag_set *set,
>  	if (!q->queue_ctx)
>  		goto err_exit;
>  
> -	q->queue_hw_ctx = kzalloc_node(nr_cpu_ids * sizeof(*(q->queue_hw_ctx)),
> -						GFP_KERNEL, set->numa_node);
> +	q->queue_hw_ctx = kzalloc_node(set->nr_hw_queues *
> +			sizeof(*(q->queue_hw_ctx)), GFP_KERNEL, set->numa_node);
>  	if (!q->queue_hw_ctx)
>  		goto err_percpu;
>  
> @@ -2319,7 +2319,7 @@ int blk_mq_alloc_tag_set(struct blk_mq_tag_set *set)
>  	if (set->nr_hw_queues > nr_cpu_ids)
>  		set->nr_hw_queues = nr_cpu_ids;
>  
> -	set->tags = kzalloc_node(nr_cpu_ids * sizeof(struct blk_mq_tags *),
> +	set->tags = kzalloc_node(set->nr_hw_queues * sizeof(*set->tags),
>  				 GFP_KERNEL, set->numa_node);
>  	if (!set->tags)
>  		return -ENOMEM;
> @@ -2360,7 +2360,7 @@ void blk_mq_free_tag_set(struct blk_mq_tag_set *set)
>  {
>  	int i;
>  
> -	for (i = 0; i < nr_cpu_ids; i++) {
> +	for (i = 0; i < set->nr_hw_queues; i++) {
>  		if (set->tags[i])
>  			blk_mq_free_rq_map(set, set->tags[i], i);
>  	}

I don't think this is safe since we might increase the number of
hardware queues (blk_mq_update_nr_hw_queues()).

-- 
Omar

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


Thread

[PATCH 00/14] blk-mq: Minor fixes and cleanups Alexander Gordeev <agordeev@redhat.com> - 2016-09-18 09:40 +0200
  [PATCH 14/14] blk-mq: Set flush_start_tag to BLK_MQ_MAX_DEPTH Alexander Gordeev <agordeev@redhat.com> - 2016-09-18 09:40 +0200
  [PATCH 01/14] blk-mq: Fix memory leaks on queue cleanup Alexander Gordeev <agordeev@redhat.com> - 2016-09-18 09:40 +0200
    Re: [PATCH 01/14] blk-mq: Fix memory leaks on queue cleanup Christoph Hellwig <hch@infradead.org> - 2016-09-19 16:00 +0200
  [PATCH 04/14] blk-mq: Do not limit number of queues to 'nr_cpu_ids' in allocations Alexander Gordeev <agordeev@redhat.com> - 2016-09-18 09:40 +0200
    Re: [PATCH 04/14] blk-mq: Do not limit number of queues to  'nr_cpu_ids' in allocations Omar Sandoval <osandov@osandov.com> - 2016-09-19 19:50 +0200
      Re: [PATCH 04/14] blk-mq: Do not limit number of queues to  'nr_cpu_ids' in allocations Alexander Gordeev <agordeev@redhat.com> - 2016-09-20 13:40 +0200
        Re: [PATCH 04/14] blk-mq: Do not limit number of queues to  'nr_cpu_ids' in allocations Omar Sandoval <osandov@osandov.com> - 2016-09-20 19:30 +0200
  [PATCH 11/14] blk-mq: Move hardware context init code into single location Alexander Gordeev <agordeev@redhat.com> - 2016-09-18 09:40 +0200
  [PATCH 02/14] blk-mq: Fix a potential NULL pointer assignment to hctx tags Alexander Gordeev <agordeev@redhat.com> - 2016-09-18 09:40 +0200
    Re: [PATCH 02/14] blk-mq: Fix a potential NULL pointer assignment to  hctx tags Omar Sandoval <osandov@osandov.com> - 2016-09-19 20:40 +0200
  [PATCH 03/14] block: Get rid of unused request_queue::nr_queues member Alexander Gordeev <agordeev@redhat.com> - 2016-09-18 09:40 +0200
    Re: [PATCH 03/14] block: Get rid of unused request_queue::nr_queues  member Omar Sandoval <osandov@osandov.com> - 2016-09-19 20:40 +0200
  [PATCH 05/14] blk-mq: Remove a redundant assignment Alexander Gordeev <agordeev@redhat.com> - 2016-09-18 09:50 +0200
    Re: [PATCH 05/14] blk-mq: Remove a redundant assignment Omar Sandoval <osandov@osandov.com> - 2016-09-19 20:40 +0200
  [PATCH 07/14] blk-mq: Cleanup a loop exit condition Alexander Gordeev <agordeev@redhat.com> - 2016-09-18 09:50 +0200
    Re: [PATCH 07/14] blk-mq: Cleanup a loop exit condition Omar Sandoval <osandov@osandov.com> - 2016-09-19 21:10 +0200
      Re: [PATCH 07/14] blk-mq: Cleanup a loop exit condition Alexander Gordeev <agordeev@redhat.com> - 2016-09-20 13:30 +0200
  [PATCH 08/14] blk-mq: Get rid of unnecessary blk_mq_free_hw_queues() Alexander Gordeev <agordeev@redhat.com> - 2016-09-18 09:50 +0200
  [PATCH 09/14] blk-mq: Move duplicating code to blk_mq_exit_hctx() Alexander Gordeev <agordeev@redhat.com> - 2016-09-18 09:50 +0200
    Re: [PATCH 09/14] blk-mq: Move duplicating code to blk_mq_exit_hctx() Omar Sandoval <osandov@osandov.com> - 2016-09-19 20:00 +0200
  [PATCH 12/14] blk-mq: Rework blk_mq_init_hctx() function Alexander Gordeev <agordeev@redhat.com> - 2016-09-18 09:50 +0200
  [PATCH 06/14] blk-mq: Fix hardware context data node selection Alexander Gordeev <agordeev@redhat.com> - 2016-09-18 09:50 +0200
    Re: [PATCH 06/14] blk-mq: Fix hardware context data node selection Omar Sandoval <osandov@osandov.com> - 2016-09-19 21:00 +0200
  [PATCH 13/14] blk-mq: Pair blk_mq_hctx_kobj_init() with blk_mq_hctx_kobj_put() Alexander Gordeev <agordeev@redhat.com> - 2016-09-18 09:50 +0200

csiph-web