linux-kernel - Re: [PATCH -next RFC v2 2/8] blk-mq: call 'bt_wait_ptr()' later in blk_mq_get

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <f6bcc53d-1419-7190-fd9a-8c5fa7178fe1@acm.org>
Date:   Fri, 8 Apr 2022 07:20:41 -0700
From:   Bart Van Assche <bvanassche@....org>
To:     Yu Kuai <yukuai3@...wei.com>, axboe@...nel.dk,
        andriy.shevchenko@...ux.intel.com, john.garry@...wei.com,
        ming.lei@...hat.com
Cc:     linux-block@...r.kernel.org, linux-kernel@...r.kernel.org,
        yi.zhang@...wei.com
Subject: Re: [PATCH -next RFC v2 2/8] blk-mq: call 'bt_wait_ptr()' later in
 blk_mq_get_tag()

On 4/8/22 00:39, Yu Kuai wrote:
> bt_wait_ptr() will increase 'wait_index', however, if blk_mq_get_tag()
> get a tag successfully after bt_wait_ptr() is called and before
> sbitmap_prepare_to_wait() is called, then the 'ws' is skipped. This
> behavior might cause 8 waitqueues to be unbalanced.
> 
> Move bt_wait_ptr() later should reduce the problem when the disk is
> under high io preesure.
> 
> Signed-off-by: Yu Kuai <yukuai3@...wei.com>
> ---
>   block/blk-mq-tag.c | 4 +---
>   1 file changed, 1 insertion(+), 3 deletions(-)
> 
> diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c
> index 68ac23d0b640..228a0001694f 100644
> --- a/block/blk-mq-tag.c
> +++ b/block/blk-mq-tag.c
> @@ -155,7 +155,6 @@ unsigned int blk_mq_get_tag(struct blk_mq_alloc_data *data)
>   	if (data->flags & BLK_MQ_REQ_NOWAIT)
>   		return BLK_MQ_NO_TAG;
>   
> -	ws = bt_wait_ptr(bt, data->hctx);
>   	do {
>   		struct sbitmap_queue *bt_prev;
>   
> @@ -174,6 +173,7 @@ unsigned int blk_mq_get_tag(struct blk_mq_alloc_data *data)
>   		if (tag != BLK_MQ_NO_TAG)
>   			break;
>   
> +		ws = bt_wait_ptr(bt, data->hctx);
>   		sbitmap_prepare_to_wait(bt, ws, &wait, TASK_UNINTERRUPTIBLE);
>   
>   		tag = __blk_mq_get_tag(data, bt);
> @@ -201,8 +201,6 @@ unsigned int blk_mq_get_tag(struct blk_mq_alloc_data *data)
>   		 */
>   		if (bt != bt_prev)
>   			sbitmap_queue_wake_up(bt_prev);
> -
> -		ws = bt_wait_ptr(bt, data->hctx);
>   	} while (1);

Is it necessary to call bt_wait_ptr() during every loop iteration or 
only if bt != bt_prev? Would calling bt_wait_ptr() only if bt != bt_prev 
help to reduce unfairness further?

Thanks,

Bart.