Re: [PATCH 5/6] blk-mq: enable checking two part inflight counts at the same time

[Omar Sandoval <osandov@osandov.com>]

On Tue, Aug 08, 2017 at 05:47:15PM -0600, Jens Axboe wrote:
> On 08/08/2017 04:48 PM, Omar Sandoval wrote:
> > On Fri, Aug 04, 2017 at 09:04:21AM -0600, Jens Axboe wrote:
> >> Modify blk_mq_in_flight() to count both a partition and root at
> >> the same time. Then we only have to call it once, instead of
> >> potentially looping the tags twice.
> > 
> > Reviewed-by: Omar Sandoval <osandov@fb.com>
> > 
> > One comment below.
> > 
> >> Signed-off-by: Jens Axboe <axboe@kernel.dk>
> > 
> > [snip]
> > 
> >> diff --git a/block/blk-mq.c b/block/blk-mq.c
> >> index fe1aa1f5f069..410ed246bc9b 100644
> >> --- a/block/blk-mq.c
> >> +++ b/block/blk-mq.c
> >> @@ -98,11 +98,13 @@ static void blk_mq_check_inflight(struct blk_mq_hw_ctx *hctx,
> >>  		return;
> >>  
> >>  	/*
> >> -	 * Count as inflight if it either matches the partition we asked
> >> -	 * for, or if it's the root
> >> +	 * Count as inflight if it matches the partition, count separately
> >> +	 * (but all) if we got asked for the root
> >>  	 */
> >> -	if (rq->part == mi->part || mi->part->partno)
> >> +	if (rq->part == mi->part)
> >>  		mi->inflight[0]++;
> > 
> > Similar concern as with patch 3, why special case the part0 case below?
> 
> Not sure I follow, both are initialized for this case. Or do you mean
> the increment?

Yeah I mean the increment. If I'm reading this right, for the !part0
case, inflight[0] is the count of in-flight requests for the partition
and inflight[1] is the count of in-flight requests for the whole device.
For the part0 case, inflight[0] is the count of in-flight requests for
the root device and inflight[1] is always 0. Can we make inflight[1] the
same in the part0 and !part0 cases?

> The comment isn't great, I should update that. Basically
> we want to increment [1] if this is a partition, and increment [0] if it
> matches the one that was asked for.
> 
> -- 
> Jens Axboe
>