linux-kernel - Re: [PATCH v2] blk-mq: release scheduler resource when request complete

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]

Date:   Thu, 17 Aug 2023 11:24:50 -0600
From:   Jens Axboe <axboe@...nel.dk>
To:     Chengming Zhou <chengming.zhou@...ux.dev>,
        Bart Van Assche <bvanassche@....org>, hch@....de
Cc:     oe-lkp@...ts.linux.dev, lkp@...el.com,
        Chuck Lever <chuck.lever@...cle.com>,
        linux-block@...r.kernel.org, cel@...nel.org,
        linux-kernel@...r.kernel.org,
        kernel test robot <oliver.sang@...el.com>
Subject: Re: [PATCH v2] blk-mq: release scheduler resource when request
 complete

On 8/17/23 11:20 AM, Jens Axboe wrote:
> On 8/17/23 9:29 AM, Chengming Zhou wrote:
>> On 2023/8/17 22:50, Bart Van Assche wrote:
>>> On 8/17/23 07:41, kernel test robot wrote:
>>>> [  222.622837][ T2216] statistics for priority 1: i 276 m 0 d 276 c 278
>>>> [ 222.629307][ T2216] WARNING: CPU: 0 PID: 2216 at block/mq-deadline.c:680 dd_exit_sched (block/mq-deadline.c:680 (discriminator 3))
>>>
>>> The above information shows that dd_inserted_request() has been called
>>> 276 times and also that dd_finish_request() has been called 278 times.
>>
>> Thanks much for your help.
>>
>> This patch indeed introduced a regression, postflush requests will be completed
>> twice, so here dd_finish_request() is more than dd_inserted_request().
>>
>> diff --git a/block/blk-mq.c b/block/blk-mq.c
>> index a8c63bef8ff1..7cd47ffc04ce 100644
>> --- a/block/blk-mq.c
>> +++ b/block/blk-mq.c
>> @@ -686,8 +686,10 @@ static void blk_mq_finish_request(struct request *rq)
>>  {
>>         struct request_queue *q = rq->q;
>>
>> -       if (rq->rq_flags & RQF_USE_SCHED)
>> +       if (rq->rq_flags & RQF_USE_SCHED) {
>>                 q->elevator->type->ops.finish_request(rq);
>> +               rq->rq_flags &= ~RQF_USE_SCHED;
>> +       }
>>  }
>>
>>
>> Clear RQF_USE_SCHED flag here should fix this problem, which should be ok
>> since finish_request() is the last callback, this flag isn't needed anymore.
>>
>> Jens, should I send this diff as another patch or resend updated v3?
> 
> I don't think this is the right solution, it makes all kinds of
> assumptions on what that flag is and when it's safe to clear it. It's a
> very fragile fix, I think we need to do better than that.

Well maybe this is actually fine, since we're freeing the request now
anyway. I can fold it in the fix, I'll add a comment as well. If this is
subtle enough that it caused this issue, we definitely should have a
comment on why we're clearing this flag.

-- 
Jens Axboe