[PATCH 2/8] blk-mq: skip non-mq queues in blk_mq_quiesce_queue

Chao Leng lengchao at huawei.com
Thu Oct 20 20:16:05 PDT 2022



On 2022/10/21 10:47, Chao Leng wrote:
> 
> 
> On 2022/10/20 18:56, Christoph Hellwig wrote:
>> For submit_bio based queues there is no (S)RCU critical section during
>> I/O submission and thus nothing to wait for in blk_mq_wait_quiesce_done,
>> so skip doing any synchronization.
>>
>> Signed-off-by: Christoph Hellwig <hch at lst.de>
>> ---
>>   block/blk-mq.c | 4 +++-
>>   1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/block/blk-mq.c b/block/blk-mq.c
>> index 33292c01875d5..df967c8af9fee 100644
>> --- a/block/blk-mq.c
>> +++ b/block/blk-mq.c
>> @@ -280,7 +280,9 @@ EXPORT_SYMBOL_GPL(blk_mq_wait_quiesce_done);
>>   void blk_mq_quiesce_queue(struct request_queue *q)
>>   {
>>       blk_mq_quiesce_queue_nowait(q);
>> -    blk_mq_wait_quiesce_done(q);
>> +    /* nothing to wait for non-mq queues */
>> +    if (queue_is_mq(q))
>> +        blk_mq_wait_quiesce_done(q);
> For the non-mq queues, maybe we should add a synchronize_rcu.
>      if (queue_is_mq(q))
>          blk_mq_wait_quiesce_done(q);
>      else
>          synchronize_rcu();
Please ignore this, synchronize_rcu is not needed.
>>   }
>>   EXPORT_SYMBOL_GPL(blk_mq_quiesce_queue);
>>
> 
> .



More information about the Linux-nvme mailing list