Re: [PATCH] block: put the reference of the io scheduler module after switching back

[Yu Kuai <yukuai1@huaweicloud.com>]

Hi!

在 2022/10/11 20:54, Yu Kuai 写道:
> Hi!
> 
> 在 2022/10/11 18:21, Jinlong Chen 写道:
>> We got a reference of the io scheduler module in
>> blk_mq_elv_switch_none to prevent the module from
>> being removed. We need to put that reference back
>> once we are done.
> 
> However, if blk_mq_elv_switch_none() failed due to OOM, and module
> reference is not acquired, blk_mq_elv_switch_back() will still be
> called. This seems not to be the right fix.

Please ignore this.

I got it now if switch none failed, old queue is not added to the list,
and later blk_lookup_qe_pair() will fail to lookup the queue.

Thanks,
Kuai
> 
> Thanks,
> Kuai
>>
>> Signed-off-by: Jinlong Chen <chenjinlong2016@outlook.com>
>> ---
>>   block/blk-mq.c | 7 +++++++
>>   1 file changed, 7 insertions(+)
>>
>> diff --git a/block/blk-mq.c b/block/blk-mq.c
>> index 8070b6c10e8d..8dfe3bf3e599 100644
>> --- a/block/blk-mq.c
>> +++ b/block/blk-mq.c
>> @@ -4595,6 +4595,13 @@ static void blk_mq_elv_switch_back(struct 
>> list_head *head,
>>       mutex_lock(&q->sysfs_lock);
>>       elevator_switch(q, t);
>> +    /**
>> +     * We got a reference of the io scheduler module in
>> +     * blk_mq_elv_switch_none to prevent the module from
>> +     * being removed. We need to put that reference back
>> +     * once we are done.
>> +     */
>> +    module_put(t->elevator_owner);
>>       mutex_unlock(&q->sysfs_lock);
>>   }
>>
> 
> .
>