Re: [PATCH] blk-mq: Fix blk_mq_tagset_busy_iter() for shared tags

[John Garry <john.garry@huawei.com>]

On 13/10/2021 16:13, John Garry wrote:
>> diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c
>> index 72a2724a4eee..2a2ad6dfcc33 100644
>> --- a/block/blk-mq-tag.c
>> +++ b/block/blk-mq-tag.c
>> @@ -232,8 +232,9 @@ static bool bt_iter(struct sbitmap *bitmap, 
>> unsigned int bitnr, void *data)
>>       if (!rq)
>>           return true;
>> -    if (rq->q == hctx->queue && rq->mq_hctx == hctx)
>> -        ret = iter_data->fn(hctx, rq, iter_data->data, reserved);
>> +    if (rq->q == hctx->queue && (rq->mq_hctx == hctx ||
>> +                blk_mq_is_shared_tags(hctx->flags)))
>> +        ret = iter_data->fn(rq->mq_hctx, rq, iter_data->data, reserved);
>>       blk_mq_put_rq_ref(rq);
>>       return ret;
>>   }
>> @@ -460,6 +461,9 @@ void blk_mq_queue_tag_busy_iter(struct 
>> request_queue *q, busy_iter_fn *fn,
>>           if (tags->nr_reserved_tags)
>>               bt_for_each(hctx, &tags->breserved_tags, fn, priv, true);
>>           bt_for_each(hctx, &tags->bitmap_tags, fn, priv, false);
>> +
>> +        if (blk_mq_is_shared_tags(hctx->flags))
>> +            break;
>>       }
>>       blk_queue_exit(q);
>>   }
>>
> 
> I suppose that is ok, and means that we iter once.
> 
> However, I have to ask, where is the big user of 
> blk_mq_queue_tag_busy_iter() coming from? I saw this from Kashyap's mail:
> 
>  > 1.31%     1.31%  kworker/57:1H-k  [kernel.vmlinux]
>  >       native_queued_spin_lock_slowpath
>  >       ret_from_fork
>  >       kthread
>  >       worker_thread
>  >       process_one_work
>  >       blk_mq_timeout_work
>  >       blk_mq_queue_tag_busy_iter
>  >       bt_iter
>  >       blk_mq_find_and_get_req
>  >       _raw_spin_lock_irqsave
>  >       native_queued_spin_lock_slowpath
> 
> How or why blk_mq_timeout_work()?

Just some update: I tried hisi_sas with 10x SAS SSDs, megaraid sas with 
1x SATA HDD (that's all I have), and null blk with lots of devices, and 
I still can't see high usage of blk_mq_queue_tag_busy_iter().

So how about we get this patch processed (to fix 
blk_mq_tagset_busy_iter()), as it is independent of 
blk_mq_queue_tag_busy_iter()? And then wait for some update or some more 
info from Kashyap regarding blk_mq_queue_tag_busy_iter()

Thanks,
John