Re: [PATCH RFC 3/5] blk-mq: Facilitate a shared tags per tagset

[Hannes Reinecke]

On 11/13/19 3:57 PM, John Garry wrote:
> On 13/11/2019 14:06, Hannes Reinecke wrote:
>> On 11/13/19 2:36 PM, John Garry wrote:
>>> Some SCSI HBAs (such as HPSA, megaraid, mpt3sas, hisi_sas_v3 ..) support
>>> multiple reply queues with single hostwide tags.
>>>
>>> In addition, these drivers want to use interrupt assignment in
>>> pci_alloc_irq_vectors(PCI_IRQ_AFFINITY). However, as discussed in [0],
>>> CPU hotplug may cause in-flight IO completion to not be serviced when an
>>> interrupt is shutdown.
>>>
>>> To solve that problem, Ming's patchset to drain hctx's should ensure no
>>> IOs are missed in-flight [1].
>>>
>>> However, to take advantage of that patchset, we need to map the HBA HW
>>> queues to blk mq hctx's; to do that, we need to expose the HBA HW
>>> queues.
>>>
>>> In making that transition, the per-SCSI command request tags are no
>>> longer unique per Scsi host - they are just unique per hctx. As such,
>>> the
>>> HBA LLDD would have to generate this tag internally, which has a certain
>>> performance overhead.
>>>
>>> However another problem is that blk mq assumes the host may accept
>>> (Scsi_host.can_queue * #hw queue) commands. In [2], we removed the Scsi
>>> host busy counter, which would stop the LLDD being sent more than
>>> .can_queue commands; however, we should still ensure that the block
>>> layer
>>> does not issue more than .can_queue commands to the Scsi host.
>>>
>>> To solve this problem, introduce a shared tags per blk_mq_tag_set, which
>>> may be requested when allocating the tagset.
>>>
>>> New flag BLK_MQ_F_TAG_HCTX_SHARED should be set when requesting the
>>> tagset.
>>>
>>> This is based on work originally from Ming Lei in [3].
>>>
>>> [0]
>>> https://lore.kernel.org/linux-block/alpine.DEB.2.21.1904051331270.1802@xxxxxxxxxxxxxxxxxxxxxxx/
>>>
>>> [1]
>>> https://lore.kernel.org/linux-block/20191014015043.25029-1-ming.lei@xxxxxxxxxx/
>>>
>>> [2]
>>> https://lore.kernel.org/linux-scsi/20191025065855.6309-1-ming.lei@xxxxxxxxxx/
>>>
>>> [3]
>>> https://lore.kernel.org/linux-block/20190531022801.10003-1-ming.lei@xxxxxxxxxx/
>>>
>>>
>>> Signed-off-by: John Garry <john.garry@xxxxxxxxxx>
>>> ---
>>> Â block/blk-core.cÂÂÂÂÂÂ |Â 1 +
>>> Â block/blk-flush.cÂÂÂÂÂ |Â 2 +
>>> Â block/blk-mq-debugfs.c |Â 2 +-
>>> Â block/blk-mq-tag.cÂÂÂÂ | 85 ++++++++++++++++++++++++++++++++++++++++++
>>> Â block/blk-mq-tag.hÂÂÂÂ |Â 1 +
>>> Â block/blk-mq.cÂÂÂÂÂÂÂÂ | 61 +++++++++++++++++++++++++-----
>>> Â block/blk-mq.hÂÂÂÂÂÂÂÂ |Â 9 +++++
>>> Â include/linux/blk-mq.h |Â 3 ++
>>> Â include/linux/blkdev.h |Â 1 +
>>> Â 9 files changed, 155 insertions(+), 10 deletions(-)
>>>
>> [ .. ]
>>> @@ -396,15 +398,17 @@ static struct request
>>> *blk_mq_get_request(struct request_queue *q,
>>> ÂÂÂÂÂÂÂÂÂ blk_mq_tag_busy(data->hctx);
>>> ÂÂÂÂÂ }
>>> Â -ÂÂÂ tag = blk_mq_get_tag(data);
>>> -ÂÂÂ if (tag == BLK_MQ_TAG_FAIL) {
>>> -ÂÂÂÂÂÂÂ if (clear_ctx_on_error)
>>> -ÂÂÂÂÂÂÂÂÂÂÂ data->ctx = NULL;
>>> -ÂÂÂÂÂÂÂ blk_queue_exit(q);
>>> -ÂÂÂÂÂÂÂ return NULL;
>>> +ÂÂÂ if (data->hctx->shared_tags) {
>>> +ÂÂÂÂÂÂÂ shared_tag = blk_mq_get_shared_tag(data);
>>> +ÂÂÂÂÂÂÂ if (shared_tag == BLK_MQ_TAG_FAIL)
>>> +ÂÂÂÂÂÂÂÂÂÂÂ goto err_shared_tag;
>>> ÂÂÂÂÂ }
>>> Â -ÂÂÂ rq = blk_mq_rq_ctx_init(data, tag, data->cmd_flags,
>>> alloc_time_ns);
>>> +ÂÂÂ tag = blk_mq_get_tag(data);
>>> +ÂÂÂ if (tag == BLK_MQ_TAG_FAIL)
>>> +ÂÂÂÂÂÂÂ goto err_tag;
>>> +
>>> +ÂÂÂ rq = blk_mq_rq_ctx_init(data, tag, shared_tag, data->cmd_flags,
>>> alloc_time_ns);
>>> ÂÂÂÂÂ if (!op_is_flush(data->cmd_flags)) {
>>> ÂÂÂÂÂÂÂÂÂ rq->elv.icq = NULL;
>>> ÂÂÂÂÂÂÂÂÂ if (e && e->type->ops.prepare_request) {
> 
> Hi Hannes,
> 
>> Why do you need to keep a parallel tag accounting between 'normal' and
>> 'shared' tags here?
>> Isn't is sufficient to get a shared tag only, and us that in lieo of the
>> 'real' one?
> 
> In theory, yes. Just the 'shared' tag should be adequate.
> 
> A problem I see with this approach is that we lose the identity of which
> tags are allocated for each hctx. As an example for this, consider
> blk_mq_queue_tag_busy_iter(), which iterates the bits for each hctx.
> Now, if you're just using shared tags only, that wouldn't work.
> 
> Consider blk_mq_can_queue() as another example - this tells us if a hctx
> has any bits unset, while with only using shared tags it would tell if
> any bits unset over all queues, and this change in semantics could break
> things. At a glance, function __blk_mq_tag_idle() looks problematic also.
> 
> And this is where it becomes messy to implement.
> 
Oh, my. Indeed, that's correct.

But then we don't really care _which_ shared tag is assigned; so
wouldn't be we better off by just having an atomic counter here?
Cache locality will be blown anyway ...

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      Teamlead Storage & Networking
hare@xxxxxxx			                  +49 911 74053 688
SUSE Software Solutions Germany GmbH, Maxfeldstr. 5, 90409 NÃrnberg
HRB 247165 (AG MÃnchen), GF: Felix ImendÃrffer