Re: [RFC v2 PATCH 2/4] block: add queue runtime pm callbacks

[Alan Stern <stern@xxxxxxxxxxxxxxxxxxx>]

On Wed, 23 May 2012, Lin Ming wrote:

> Let's consider below code.
> 
> @@ -587,6 +591,11 @@ void __elv_add_request(struct request_queue *q,
> struct request *rq, int where)
>  {
>        trace_block_rq_insert(q, rq);
> 
> +       if (!(rq->cmd_flags & REQ_PM))
> +               if (q->nr_pending++ == 0 && (q->rpm_status == RPM_SUSPENDED ||
> +                               q->rpm_status == RPM_SUSPENDING) && q->dev)
> +                       pm_request_resume(q->dev);
> +
>        rq->q = q;
> 
>        if (rq->cmd_flags & REQ_SOFTBARRIER) {
> 
> Block layer reads runtime status and pm core writes this status.
> PM core uses dev->power.lock to protect this status.
> 
> I was thinking will it have problem if block layer does not acquire
> dev->power.lock?
> From your explanation below, it seems does not have problem.

I don't think it's a problem, because all you're doing is reading 
dev->power.rpm_status -- you're not writing it.

On the other hand, there's nothing really wrong with keeping your own
local copy of rpm_status.  You could think of it as being the queue's
status as opposed to the device's status.  (Also, some people might
argue that dev->power.rpm_status is supposed to be private to the
runtime PM core and shouldn't be used by other code.)

Alan Stern

--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html