Re: [PATCH] block: Make blk_drain_queue() work for stopped queues

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

On 03/19/12 17:04, Tejun Heo wrote:
> Thanks for the pointer. It would be great if you can describe / link
> the actual case the patch is trying to solve.

Another example is the sd scanning code (sd_probe_async()) which can
queue SCSI commands concurrently with the blk_cleanup_queue() call
called (indirectly) from scsi_remove_host().

>
>> there is still a tiny race window left by that patch - new
>> requests can get queued after the SCSI request function has been invoked
>> by scsi_free_queue() and before blk_cleanup_queue() gets invoked. Hence
>> the proposal to change the block layer to make sure that all queued
>> requests get processed eventually.
> I don't think it's a good idea to push requests out to stopped queue.
> Wouldn't aborting all pending requests be better?

Sure, but is that possible from inside the block layer ? With patch
http://marc.info/?l=linux-scsi&m=133207725114386 applied, the following
code is present at the start of scsi_request_fn():

	if (unlikely(blk_queue_dead(q))) {
 		while ((req = blk_peek_request(q)) != NULL)
 			scsi_kill_request(req, q);

Bart.