Re: [PATCH 2/4] block: Avoid that request_fn is invoked on a dead queue

[Bart Van Assche <bvanassche@acm.org>]

On 10/17/12 01:38, Tejun Heo wrote:
>>   /**
>> + * __blk_run_queue_uncond - run a queue whether or not it has been stopped
>> + * @q:	The queue to run
>> + *
>> + * Description:
>> + *    Invoke request handling on a queue if there are any pending requests.
>> + *    May be used to restart request handling after a request has completed.
>> + *    This variant runs the queue whether or not the queue has been
>> + *    stopped. Must be called with the queue lock held and interrupts
>> + *    disabled. See also @blk_run_queue.
>> + */
>> +void __blk_run_queue_uncond(struct request_queue *q)
>> +{
>> +	if (unlikely(blk_queue_dead(q)))
>> +		return;
>> +
>> +	q->request_fn(q);
>> +}
>> +
>> +/**
>>    * __blk_run_queue - run a single device queue
>>    * @q:	The queue to run
>>    *
>> @@ -305,7 +324,7 @@ void __blk_run_queue(struct request_queue *q)
>>   	if (unlikely(blk_queue_stopped(q)))
>>   		return;
>>
>> -	q->request_fn(q);
>> +	__blk_run_queue_uncond(q);
>
> __blk_run_queue_uncond() is a cold path and I don't think adding a
> test there matters but I think it would be better if we avoid an extra
> branch if possible for __blk_run_queue().  Can't we merge
> blk_queue_stopped/dead() testing?

How about declaring the function __blk_run_queue_uncond() inline ? That 
should allow the compiler to combine the two tests into a single test.

Bart.