Re: [PATCH RFC V2 3/5] blk-mq: Facilitate a shared sbitmap per tagset

From: Ming Lei
Date: Thu Nov 21 2019 - 03:55:56 EST


On Tue, Nov 19, 2019 at 10:27:36PM +0800, 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 sbitmap per blk_mq_tag_set,
> which may be requested at init time.
>
> New flag BLK_MQ_F_TAG_HCTX_SHARED should be set when requesting the
> tagset to indicate whether the shared sbitmap should be used.
>
> Even when BLK_MQ_F_TAG_HCTX_SHARED is set, we still allocate a full set of
> tags and requests per hctx; the reason for this is that if we only allocate
> tags and requests for a single hctx - like hctx0 - we may break block
> drivers which expect a request be associated with a specific hctx, i.e.
> not hctx0.
>
> This is based on work originally from Ming Lei in [3] and from Bart's
> suggestion in [4].
>
> [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/
> [4] https://lore.kernel.org/linux-block/ff77beff-5fd9-9f05-12b6-826922bace1f@xxxxxxxxxx/T/#m3db0a602f095cbcbff27e9c884d6b4ae826144be
>
> Signed-off-by: John Garry <john.garry@xxxxxxxxxx>
> ---
> block/bfq-iosched.c | 4 +--
> block/blk-mq-debugfs.c | 4 +--
> block/blk-mq-sched.c | 14 ++++++++
> block/blk-mq-tag.c | 74 +++++++++++++++++++++++++++++++++---------
> block/blk-mq-tag.h | 15 +++++++--
> block/blk-mq.c | 57 +++++++++++++++++++++++++++++---
> block/blk-mq.h | 5 +++
> block/kyber-iosched.c | 4 +--
> include/linux/blk-mq.h | 7 ++++
> 9 files changed, 157 insertions(+), 27 deletions(-)
>
> diff --git a/block/bfq-iosched.c b/block/bfq-iosched.c
> index 0319d6339822..9633c864af07 100644
> --- a/block/bfq-iosched.c
> +++ b/block/bfq-iosched.c
> @@ -6327,8 +6327,8 @@ static void bfq_depth_updated(struct blk_mq_hw_ctx *hctx)
> struct blk_mq_tags *tags = hctx->sched_tags;
> unsigned int min_shallow;
>
> - min_shallow = bfq_update_depths(bfqd, &tags->bitmap_tags);
> - sbitmap_queue_min_shallow_depth(&tags->bitmap_tags, min_shallow);
> + min_shallow = bfq_update_depths(bfqd, tags->pbitmap_tags);
> + sbitmap_queue_min_shallow_depth(tags->pbitmap_tags, min_shallow);
> }
>
> static int bfq_init_hctx(struct blk_mq_hw_ctx *hctx, unsigned int index)
> diff --git a/block/blk-mq-debugfs.c b/block/blk-mq-debugfs.c
> index 33a40ae1d60f..9cf2f09c08e4 100644
> --- a/block/blk-mq-debugfs.c
> +++ b/block/blk-mq-debugfs.c
> @@ -483,7 +483,7 @@ static int hctx_tags_bitmap_show(void *data, struct seq_file *m)
> res = mutex_lock_interruptible(&q->sysfs_lock);
> if (res)
> goto out;
> - if (hctx->tags)
> + if (hctx->tags) /* We should just iterate the relevant bits for this hctx FIXME */
> sbitmap_bitmap_show(&hctx->tags->bitmap_tags.sb, m);
> mutex_unlock(&q->sysfs_lock);
>
> @@ -517,7 +517,7 @@ static int hctx_sched_tags_bitmap_show(void *data, struct seq_file *m)
> res = mutex_lock_interruptible(&q->sysfs_lock);
> if (res)
> goto out;
> - if (hctx->sched_tags)
> + if (hctx->sched_tags) /* We should just iterate the relevant bits for this hctx FIXME */
> sbitmap_bitmap_show(&hctx->sched_tags->bitmap_tags.sb, m);
> mutex_unlock(&q->sysfs_lock);
>
> diff --git a/block/blk-mq-sched.c b/block/blk-mq-sched.c
> index ca22afd47b3d..b23547f10949 100644
> --- a/block/blk-mq-sched.c
> +++ b/block/blk-mq-sched.c
> @@ -492,6 +492,7 @@ static void blk_mq_sched_tags_teardown(struct request_queue *q)
>
> int blk_mq_init_sched(struct request_queue *q, struct elevator_type *e)
> {
> + struct blk_mq_tag_set *tag_set = q->tag_set;
> struct blk_mq_hw_ctx *hctx;
> struct elevator_queue *eq;
> unsigned int i;
> @@ -537,6 +538,19 @@ int blk_mq_init_sched(struct request_queue *q, struct elevator_type *e)
> blk_mq_debugfs_register_sched_hctx(q, hctx);
> }
>
> + if (blk_mq_is_sbitmap_shared(tag_set)) {
> + if (!blk_mq_init_sched_shared_sbitmap(tag_set, q->nr_requests)) {
> + ret = -ENOMEM;
> + goto err;
> + }
> + queue_for_each_hw_ctx(q, hctx, i) {
> + struct blk_mq_tags *tags = hctx->sched_tags;
> +
> + tags->pbitmap_tags = &tag_set->sched_shared_bitmap_tags;
> + tags->pbreserved_tags = &tag_set->sched_shared_breserved_tags;

This kind of sharing is wrong, sched tags should be request queue wide
instead of tagset wide, and each request queue has its own & independent
scheduler queue.

> + }
> + }
> +
> return 0;
>
> err:
> diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c
> index 42792942b428..6625bebb46c3 100644
> --- a/block/blk-mq-tag.c
> +++ b/block/blk-mq-tag.c
> @@ -35,9 +35,9 @@ bool __blk_mq_tag_busy(struct blk_mq_hw_ctx *hctx)
> */
> void blk_mq_tag_wakeup_all(struct blk_mq_tags *tags, bool include_reserve)
> {
> - sbitmap_queue_wake_all(&tags->bitmap_tags);
> + sbitmap_queue_wake_all(tags->pbitmap_tags);
> if (include_reserve)
> - sbitmap_queue_wake_all(&tags->breserved_tags);
> + sbitmap_queue_wake_all(tags->pbreserved_tags);
> }
>
> /*
> @@ -113,10 +113,10 @@ unsigned int blk_mq_get_tag(struct blk_mq_alloc_data *data)
> WARN_ON_ONCE(1);
> return BLK_MQ_TAG_FAIL;
> }
> - bt = &tags->breserved_tags;
> + bt = tags->pbreserved_tags;
> tag_offset = 0;
> } else {
> - bt = &tags->bitmap_tags;
> + bt = tags->pbitmap_tags;
> tag_offset = tags->nr_reserved_tags;
> }
>
> @@ -162,9 +162,9 @@ unsigned int blk_mq_get_tag(struct blk_mq_alloc_data *data)
> data->ctx);
> tags = blk_mq_tags_from_data(data);
> if (data->flags & BLK_MQ_REQ_RESERVED)
> - bt = &tags->breserved_tags;
> + bt = tags->pbreserved_tags;
> else
> - bt = &tags->bitmap_tags;
> + bt = tags->pbitmap_tags;
>
> /*
> * If destination hw queue is changed, fake wake up on
> @@ -190,10 +190,10 @@ void blk_mq_put_tag(struct blk_mq_tags *tags, struct blk_mq_ctx *ctx,
> const int real_tag = tag - tags->nr_reserved_tags;
>
> BUG_ON(real_tag >= tags->nr_tags);
> - sbitmap_queue_clear(&tags->bitmap_tags, real_tag, ctx->cpu);
> + sbitmap_queue_clear(tags->pbitmap_tags, real_tag, ctx->cpu);
> } else {
> BUG_ON(tag >= tags->nr_reserved_tags);
> - sbitmap_queue_clear(&tags->breserved_tags, tag, ctx->cpu);
> + sbitmap_queue_clear(tags->pbreserved_tags, tag, ctx->cpu);
> }
> }
>
> @@ -220,7 +220,7 @@ static bool bt_iter(struct sbitmap *bitmap, unsigned int bitnr, void *data)
> * We can hit rq == NULL here, because the tagging functions
> * test and set the bit before assigning ->rqs[].
> */
> - if (rq && rq->q == hctx->queue)
> + if (rq && rq->q == hctx->queue && rq->mq_hctx == hctx)
> return iter_data->fn(hctx, rq, iter_data->data, reserved);
> return true;
> }
> @@ -321,8 +321,8 @@ static void blk_mq_all_tag_busy_iter(struct blk_mq_tags *tags,
> busy_tag_iter_fn *fn, void *priv)
> {
> if (tags->nr_reserved_tags)
> - bt_tags_for_each(tags, &tags->breserved_tags, fn, priv, true);
> - bt_tags_for_each(tags, &tags->bitmap_tags, fn, priv, false);
> + bt_tags_for_each(tags, tags->pbreserved_tags, fn, priv, true);
> + bt_tags_for_each(tags, tags->pbitmap_tags, fn, priv, false);
> }
>
> /**
> @@ -419,8 +419,8 @@ void blk_mq_queue_tag_busy_iter(struct request_queue *q, busy_iter_fn *fn,
> continue;
>
> if (tags->nr_reserved_tags)
> - bt_for_each(hctx, &tags->breserved_tags, fn, priv, true);
> - bt_for_each(hctx, &tags->bitmap_tags, fn, priv, false);
> + bt_for_each(hctx, tags->pbreserved_tags, fn, priv, true);
> + bt_for_each(hctx, tags->pbitmap_tags, fn, priv, false);
> }
> blk_queue_exit(q);
> }
> @@ -444,6 +444,9 @@ static struct blk_mq_tags *blk_mq_init_bitmap_tags(struct blk_mq_tags *tags,
> node))
> goto free_bitmap_tags;
>
> + tags->pbitmap_tags = &tags->bitmap_tags;
> + tags->pbreserved_tags = &tags->breserved_tags;
> +
> return tags;
> free_bitmap_tags:
> sbitmap_queue_free(&tags->bitmap_tags);
> @@ -452,7 +455,46 @@ static struct blk_mq_tags *blk_mq_init_bitmap_tags(struct blk_mq_tags *tags,
> return NULL;
> }
>
> -struct blk_mq_tags *blk_mq_init_tags(unsigned int total_tags,
> +bool blk_mq_init_shared_sbitmap(struct blk_mq_tag_set *tag_set)
> +{
> + unsigned int depth = tag_set->queue_depth -tag_set->reserved_tags;
> + int alloc_policy = BLK_MQ_FLAG_TO_ALLOC_POLICY(tag_set->flags);
> + bool round_robin = alloc_policy == BLK_TAG_ALLOC_RR;
> + int node = tag_set->numa_node;
> +
> + if (bt_alloc(&tag_set->shared_bitmap_tags, depth, round_robin, node))
> + return false;
> + if (bt_alloc(&tag_set->shared_breserved_tags, tag_set->reserved_tags, round_robin,
> + node))
> + goto free_bitmap_tags;
> +
> + return true;
> +free_bitmap_tags:
> + sbitmap_queue_free(&tag_set->shared_bitmap_tags);
> + return false;
> +}
> +
> +bool blk_mq_init_sched_shared_sbitmap(struct blk_mq_tag_set *tag_set, unsigned long nr_requests)
> +{
> + unsigned int depth = nr_requests -tag_set->reserved_tags;
> + int alloc_policy = BLK_MQ_FLAG_TO_ALLOC_POLICY(tag_set->flags);
> + bool round_robin = alloc_policy == BLK_TAG_ALLOC_RR;
> + int node = tag_set->numa_node;
> +
> + if (bt_alloc(&tag_set->sched_shared_bitmap_tags, depth, round_robin, node))
> + return false;
> + if (bt_alloc(&tag_set->sched_shared_breserved_tags, tag_set->reserved_tags, round_robin,
> + node))
> + goto free_bitmap_tags;
> +
> + return true;
> +free_bitmap_tags:
> + sbitmap_queue_free(&tag_set->sched_shared_bitmap_tags);
> + return false;
> +}
> +
> +struct blk_mq_tags *blk_mq_init_tags(struct blk_mq_tag_set *set,
> + unsigned int total_tags,
> unsigned int reserved_tags,
> int node, int alloc_policy)
> {
> @@ -470,6 +512,8 @@ struct blk_mq_tags *blk_mq_init_tags(unsigned int total_tags,
> tags->nr_tags = total_tags;
> tags->nr_reserved_tags = reserved_tags;
>
> + if (blk_mq_is_sbitmap_shared(set))
> + return tags;
> return blk_mq_init_bitmap_tags(tags, node, alloc_policy);
> }
>
> @@ -526,7 +570,7 @@ int blk_mq_tag_update_depth(struct blk_mq_hw_ctx *hctx,
> * Don't need (or can't) update reserved tags here, they
> * remain static and should never need resizing.
> */
> - sbitmap_queue_resize(&tags->bitmap_tags,
> + sbitmap_queue_resize(tags->pbitmap_tags,
> tdepth - tags->nr_reserved_tags);
> }
>
> diff --git a/block/blk-mq-tag.h b/block/blk-mq-tag.h
> index e36bec8e0970..198b6fbe2c22 100644
> --- a/block/blk-mq-tag.h
> +++ b/block/blk-mq-tag.h
> @@ -13,20 +13,31 @@ struct blk_mq_tags {
>
> atomic_t active_queues;
>
> + /* We should consider deleting these so they are not referenced by mistake */
> struct sbitmap_queue bitmap_tags;
> struct sbitmap_queue breserved_tags;
>
> + struct sbitmap_queue *pbitmap_tags;
> + struct sbitmap_queue *pbreserved_tags;
> +
> struct request **rqs;
> struct request **static_rqs;
> struct list_head page_list;
> };
>
>
> -extern struct blk_mq_tags *blk_mq_init_tags(unsigned int nr_tags, unsigned int reserved_tags, int node, int alloc_policy);
> +extern bool blk_mq_init_shared_sbitmap(struct blk_mq_tag_set *tag_set);
> +extern bool blk_mq_init_sched_shared_sbitmap(struct blk_mq_tag_set *tag_set,
> + unsigned long nr_requests);
> +extern struct blk_mq_tags *blk_mq_init_tags(struct blk_mq_tag_set *tag_set,
> + unsigned int nr_tags,
> + unsigned int reserved_tags,
> + int node, int alloc_policy);
> extern void blk_mq_free_tags(struct blk_mq_tags *tags);
>
> extern unsigned int blk_mq_get_tag(struct blk_mq_alloc_data *data);
> -extern void blk_mq_put_tag(struct blk_mq_tags *tags, struct blk_mq_ctx *ctx, unsigned int tag);
> +extern void blk_mq_put_tag(struct blk_mq_tags *tags,
> + struct blk_mq_ctx *ctx, unsigned int tag);
> extern int blk_mq_tag_update_depth(struct blk_mq_hw_ctx *hctx,
> struct blk_mq_tags **tags,
> unsigned int depth, bool can_grow);
> diff --git a/block/blk-mq.c b/block/blk-mq.c
> index 2d2956249ae9..8d5b21919c9a 100644
> --- a/block/blk-mq.c
> +++ b/block/blk-mq.c
> @@ -1089,7 +1089,7 @@ static int blk_mq_dispatch_wake(wait_queue_entry_t *wait, unsigned mode,
> struct sbitmap_queue *sbq;
>
> list_del_init(&wait->entry);
> - sbq = &hctx->tags->bitmap_tags;
> + sbq = hctx->tags->pbitmap_tags;
> atomic_dec(&sbq->ws_active);
> }
> spin_unlock(&hctx->dispatch_wait_lock);
> @@ -1107,7 +1107,7 @@ static int blk_mq_dispatch_wake(wait_queue_entry_t *wait, unsigned mode,
> static bool blk_mq_mark_tag_wait(struct blk_mq_hw_ctx *hctx,
> struct request *rq)
> {
> - struct sbitmap_queue *sbq = &hctx->tags->bitmap_tags;
> + struct sbitmap_queue *sbq = hctx->tags->pbitmap_tags;
> struct wait_queue_head *wq;
> wait_queue_entry_t *wait;
> bool ret;
> @@ -2097,7 +2097,7 @@ struct blk_mq_tags *blk_mq_alloc_rq_map(struct blk_mq_tag_set *set,
> if (node == NUMA_NO_NODE)
> node = set->numa_node;
>
> - tags = blk_mq_init_tags(nr_tags, reserved_tags, node,
> + tags = blk_mq_init_tags(set, nr_tags, reserved_tags, node,
> BLK_MQ_FLAG_TO_ALLOC_POLICY(set->flags));
> if (!tags)
> return NULL;
> @@ -3100,6 +3100,20 @@ int blk_mq_alloc_tag_set(struct blk_mq_tag_set *set)
> if (ret)
> goto out_free_mq_map;
>
> + if (blk_mq_is_sbitmap_shared(set)) {
> + if (!blk_mq_init_shared_sbitmap(set)) {
> + ret = -ENOMEM;
> + goto out_free_mq_map;
> + }
> +
> + for (i = 0; i < set->nr_hw_queues; i++) {
> + struct blk_mq_tags *tags = set->tags[i];
> +
> + tags->pbitmap_tags = &set->shared_bitmap_tags;
> + tags->pbreserved_tags = &set->shared_breserved_tags;
> + }
> + }
> +
> mutex_init(&set->tag_list_lock);
> INIT_LIST_HEAD(&set->tag_list);
>
> @@ -3137,6 +3151,7 @@ int blk_mq_update_nr_requests(struct request_queue *q, unsigned int nr)
> {
> struct blk_mq_tag_set *set = q->tag_set;
> struct blk_mq_hw_ctx *hctx;
> + bool sched_tags = false;
> int i, ret;
>
> if (!set)
> @@ -3160,6 +3175,7 @@ int blk_mq_update_nr_requests(struct request_queue *q, unsigned int nr)
> ret = blk_mq_tag_update_depth(hctx, &hctx->tags, nr,
> false);
> } else {
> + sched_tags = true;
> ret = blk_mq_tag_update_depth(hctx, &hctx->sched_tags,
> nr, true);
> }
> @@ -3169,8 +3185,41 @@ int blk_mq_update_nr_requests(struct request_queue *q, unsigned int nr)
> q->elevator->type->ops.depth_updated(hctx);
> }
>
> - if (!ret)
> + /*
> + * if ret is 0, all queues should have been updated to the same depth
> + * if not, then maybe some have been updated - yuk, need to handle this for shared sbitmap...
> + * if some are updated, we should probably roll back the change altogether. FIXME
> + */
> + if (!ret) {
> + if (blk_mq_is_sbitmap_shared(set)) {
> + if (sched_tags) {
> + sbitmap_queue_free(&set->sched_shared_bitmap_tags);
> + sbitmap_queue_free(&set->sched_shared_breserved_tags);
> + if (!blk_mq_init_sched_shared_sbitmap(set, nr))
> + return -ENOMEM; /* fixup error handling */
> +
> + queue_for_each_hw_ctx(q, hctx, i) {
> + hctx->sched_tags->pbitmap_tags = &set->sched_shared_bitmap_tags;
> + hctx->sched_tags->pbreserved_tags = &set->sched_shared_breserved_tags;
> + }
> + } else {
> + sbitmap_queue_free(&set->shared_bitmap_tags);
> + sbitmap_queue_free(&set->shared_breserved_tags);
> + if (!blk_mq_init_shared_sbitmap(set))
> + return -ENOMEM; /* fixup error handling */

No, we can't re-allocate driver tags here which are shared by all LUNs.
And you should see that 'can_grow' is set as false for driver tags
in blk_mq_update_nr_requests(), which can only touch per-request-queue
data, not tagset wide data.


Thanks,
Ming