Re: [PATCH RFC v2 12/24] hpsa: use reserved commands

[John Garry]

On 17/03/2020 09:48, Hannes Reinecke wrote:
> On 3/17/20 10:38 AM, John Garry wrote:
> > On 11/03/2020 08:10, Ming Lei wrote:
> > > > ands(struct ctlr_info *h)
> > > > @@ -5803,6 +5803,7 @@ static int hpsa_scsi_host_alloc(struct 
> > > > ctlr_info *h)
> > > > ÂÂÂÂÂ sh->max_lun = HPSA_MAX_LUN;
> > > > ÂÂÂÂÂ sh->max_id = HPSA_MAX_LUN;
> > > > ÂÂÂÂÂ sh->can_queue = h->nr_cmds - HPSA_NRESERVED_CMDS;
> > > > +ÂÂÂ sh->nr_reserved_cmds = HPSA_NRESERVED_CMDS;
> > > Now .nr_reserved_cmds has been passed to blk-mq, you need to increase
> > > sh->can_queue to h->nr_cmds, because .can_queue is the whole queue depth
> > > (include the part of reserved tags), otherwise, IO tags will be
> > > decreased.
> >
> > Sounds correct.
> I will have having a look at the patchset; I thought I did a patch to 
> modify .can_queue so that it would cover only the usable tags, not the 
> reserved ones.

To me, it makes sense to leave .can_queue unmodified, carry it down to 
blk-mq and allow blk_mq_init_bitmap_tags() find the queue depth:

static struct blk_mq_tags *blk_mq_init_bitmap_tags(struct blk_mq_tags *tags,
  int node, int alloc_policy)
{
	unsigned int depth = tags->nr_tags - tags->nr_reserved_tags; *
	bool round_robin = alloc_policy == BLK_TAG_ALLOC_RR;

	if (bt_alloc(&tags->bitmap_tags, depth, round_robin, node))
		goto free_tags;

Cheers,
John