Re: [PATCH 1/1] blk-mq: fix blk_mq_hw_ctx active request accounting

[Jens Axboe <axboe@kernel.dk>]

On 5/13/23 2:54?PM, Simon Horman wrote:
> On Sat, May 13, 2023 at 03:05:34PM -0400, Tian Lan wrote:
>> From: Tian Lan <tian.lan@twosigma.com>
>>
>> The nr_active counter continues to increase over time which causes the
>> blk_mq_get_tag to hang until the thread is rescheduled to a different
>> core despite there are still tags available.
>>
>> kernel-stack
>>
>>   INFO: task inboundIOReacto:3014879 blocked for more than 2 seconds
>>   Not tainted 6.1.15-amd64 #1 Debian 6.1.15~debian11
>>   "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
>>   task:inboundIOReacto state:D stack:0  pid:3014879 ppid:4557 flags:0x00000000
>>     Call Trace:
>>     <TASK>
>>     __schedule+0x351/0xa20
>>     scheduler+0x5d/0xe0
>>     io_schedule+0x42/0x70
>>     blk_mq_get_tag+0x11a/0x2a0
>>     ? dequeue_task_stop+0x70/0x70
>>     __blk_mq_alloc_requests+0x191/0x2e0
>>
>> kprobe output showing RQF_MQ_INFLIGHT bit is not cleared before
>> __blk_mq_free_request being called.
>>
>>   320    320  kworker/29:1H __blk_mq_free_request rq_flags 0x220c0 in-flight 1
>>          b'__blk_mq_free_request+0x1 [kernel]'
>>          b'bt_iter+0x50 [kernel]'
>>          b'blk_mq_queue_tag_busy_iter+0x318 [kernel]'
>>          b'blk_mq_timeout_work+0x7c [kernel]'
>>          b'process_one_work+0x1c4 [kernel]'
>>          b'worker_thread+0x4d [kernel]'
>>          b'kthread+0xe6 [kernel]'
>>          b'ret_from_fork+0x1f [kernel]'
>>
>> Signed-off-by: Tian Lan <tian.lan@twosigma.com>
>> ---
>>  block/blk-mq.c | 7 ++++---
>>  1 file changed, 4 insertions(+), 3 deletions(-)
>>
>> diff --git a/block/blk-mq.c b/block/blk-mq.c
>> index 9c8dc70020bc..732a39d88cd6 100644
>> --- a/block/blk-mq.c
>> +++ b/block/blk-mq.c
>> @@ -716,6 +716,10 @@ static void __blk_mq_free_request(struct request *rq)
>>  	blk_crypto_free_request(rq);
>>  	blk_pm_mark_last_busy(rq);
>>  	rq->mq_hctx = NULL;
>> +
>> +	if (rq->rq_flags & RQF_MQ_INFLIGHT)
>> +		__blk_mq_dec_active_requests(hctx);
>> +
>>  	if (rq->tag != BLK_MQ_NO_TAG)
>>  		blk_mq_put_tag(hctx->tags, ctx, rq->tag);
>>  	if (sched_tag != BLK_MQ_NO_TAG)
>> @@ -733,9 +737,6 @@ void blk_mq_free_request(struct request *rq)
>>  	    q->elevator->type->ops.finish_request)
>>  		q->elevator->type->ops.finish_request(rq);
>>  
>> -	if (rq->rq_flags & RQF_MQ_INFLIGHT)
>> -		__blk_mq_dec_active_requests(hctx);
>> -
> 
> Unless I am mistaken, hctx is now unused in this function.

Indeed. Tian, first one didn't compile and this one will most certainly
spew a warning. That's 2 for 2 so far. I realize that you tested this
one a different kernel in prod, but don't seen stuff you haven't even
compiled.

-- 
Jens Axboe