1 files changed, 135 insertions, 174 deletions

diff --git a/block/blk-mq.c b/block/blk-mq.c
index c2697db59109..4806b867e37d 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -89,7 +89,7 @@ struct mq_inflight {
 	unsigned int inflight[2];
 };
 
-static bool blk_mq_check_inflight(struct request *rq, void *priv)
+static bool blk_mq_check_in_driver(struct request *rq, void *priv)
 {
 	struct mq_inflight *mi = priv;
 
@@ -101,24 +101,14 @@ static bool blk_mq_check_inflight(struct request *rq, void *priv)
 	return true;
 }
 
-unsigned int blk_mq_in_flight(struct request_queue *q,
-		struct block_device *part)
+void blk_mq_in_driver_rw(struct block_device *part, unsigned int inflight[2])
 {
 	struct mq_inflight mi = { .part = part };
 
-	blk_mq_queue_tag_busy_iter(q, blk_mq_check_inflight, &mi);
-
-	return mi.inflight[0] + mi.inflight[1];
-}
-
-void blk_mq_in_flight_rw(struct request_queue *q, struct block_device *part,
-		unsigned int inflight[2])
-{
-	struct mq_inflight mi = { .part = part };
-
-	blk_mq_queue_tag_busy_iter(q, blk_mq_check_inflight, &mi);
-	inflight[0] = mi.inflight[0];
-	inflight[1] = mi.inflight[1];
+	blk_mq_queue_tag_busy_iter(bdev_get_queue(part), blk_mq_check_in_driver,
+				   &mi);
+	inflight[READ] = mi.inflight[READ];
+	inflight[WRITE] = mi.inflight[WRITE];
 }
 
 #ifdef CONFIG_LOCKDEP
@@ -584,9 +574,13 @@ static struct request *blk_mq_rq_cache_fill(struct request_queue *q,
 	struct blk_mq_alloc_data data = {
 		.q		= q,
 		.flags		= flags,
+		.shallow_depth	= 0,
 		.cmd_flags	= opf,
+		.rq_flags	= 0,
 		.nr_tags	= plug->nr_ios,
 		.cached_rqs	= &plug->cached_rqs,
+		.ctx		= NULL,
+		.hctx		= NULL
 	};
 	struct request *rq;
 
@@ -646,8 +640,13 @@ struct request *blk_mq_alloc_request(struct request_queue *q, blk_opf_t opf,
 		struct blk_mq_alloc_data data = {
 			.q		= q,
 			.flags		= flags,
+			.shallow_depth	= 0,
 			.cmd_flags	= opf,
+			.rq_flags	= 0,
 			.nr_tags	= 1,
+			.cached_rqs	= NULL,
+			.ctx		= NULL,
+			.hctx		= NULL
 		};
 		int ret;
 
@@ -675,8 +674,13 @@ struct request *blk_mq_alloc_request_hctx(struct request_queue *q,
 	struct blk_mq_alloc_data data = {
 		.q		= q,
 		.flags		= flags,
+		.shallow_depth	= 0,
 		.cmd_flags	= opf,
+		.rq_flags	= 0,
 		.nr_tags	= 1,
+		.cached_rqs	= NULL,
+		.ctx		= NULL,
+		.hctx		= NULL
 	};
 	u64 alloc_time_ns = 0;
 	struct request *rq;
@@ -2080,7 +2084,7 @@ static void blk_mq_commit_rqs(struct blk_mq_hw_ctx *hctx, int queued,
  * Returns true if we did some work AND can potentially do more.
  */
 bool blk_mq_dispatch_rq_list(struct blk_mq_hw_ctx *hctx, struct list_head *list,
-			     unsigned int nr_budgets)
+			     bool get_budget)
 {
 	enum prep_dispatch prep;
 	struct request_queue *q = hctx->queue;
@@ -2102,7 +2106,7 @@ bool blk_mq_dispatch_rq_list(struct blk_mq_hw_ctx *hctx, struct list_head *list,
 		rq = list_first_entry(list, struct request, queuelist);
 
 		WARN_ON_ONCE(hctx != rq->mq_hctx);
-		prep = blk_mq_prep_dispatch_rq(rq, !nr_budgets);
+		prep = blk_mq_prep_dispatch_rq(rq, get_budget);
 		if (prep != PREP_DISPATCH_OK)
 			break;
 
@@ -2111,12 +2115,6 @@ bool blk_mq_dispatch_rq_list(struct blk_mq_hw_ctx *hctx, struct list_head *list,
 		bd.rq = rq;
 		bd.last = list_empty(list);
 
-		/*
-		 * once the request is queued to lld, no need to cover the
-		 * budget any more
-		 */
-		if (nr_budgets)
-			nr_budgets--;
 		ret = q->mq_ops->queue_rq(hctx, &bd);
 		switch (ret) {
 		case BLK_STS_OK:
@@ -2150,7 +2148,11 @@ out:
 			((hctx->flags & BLK_MQ_F_TAG_QUEUE_SHARED) ||
 			blk_mq_is_shared_tags(hctx->flags));
 
-		if (nr_budgets)
+		/*
+		 * If the caller allocated budgets, free the budgets of the
+		 * requests that have not yet been passed to the block driver.
+		 */
+		if (!get_budget)
 			blk_mq_release_budgets(q, list);
 
 		spin_lock(&hctx->lock);
@@ -2778,15 +2780,15 @@ static blk_status_t blk_mq_request_issue_directly(struct request *rq, bool last)
 	return __blk_mq_issue_directly(hctx, rq, last);
 }
 
-static void blk_mq_plug_issue_direct(struct blk_plug *plug)
+static void blk_mq_issue_direct(struct rq_list *rqs)
 {
 	struct blk_mq_hw_ctx *hctx = NULL;
 	struct request *rq;
 	int queued = 0;
 	blk_status_t ret = BLK_STS_OK;
 
-	while ((rq = rq_list_pop(&plug->mq_list))) {
-		bool last = rq_list_empty(&plug->mq_list);
+	while ((rq = rq_list_pop(rqs))) {
+		bool last = rq_list_empty(rqs);
 
 		if (hctx != rq->mq_hctx) {
 			if (hctx) {
@@ -2817,15 +2819,64 @@ out:
 		blk_mq_commit_rqs(hctx, queued, false);
 }
 
-static void __blk_mq_flush_plug_list(struct request_queue *q,
-				     struct blk_plug *plug)
+static void __blk_mq_flush_list(struct request_queue *q, struct rq_list *rqs)
 {
 	if (blk_queue_quiesced(q))
 		return;
-	q->mq_ops->queue_rqs(&plug->mq_list);
+	q->mq_ops->queue_rqs(rqs);
+}
+
+static unsigned blk_mq_extract_queue_requests(struct rq_list *rqs,
+					      struct rq_list *queue_rqs)
+{
+	struct request *rq = rq_list_pop(rqs);
+	struct request_queue *this_q = rq->q;
+	struct request **prev = &rqs->head;
+	struct rq_list matched_rqs = {};
+	struct request *last = NULL;
+	unsigned depth = 1;
+
+	rq_list_add_tail(&matched_rqs, rq);
+	while ((rq = *prev)) {
+		if (rq->q == this_q) {
+			/* move rq from rqs to matched_rqs */
+			*prev = rq->rq_next;
+			rq_list_add_tail(&matched_rqs, rq);
+			depth++;
+		} else {
+			/* leave rq in rqs */
+			prev = &rq->rq_next;
+			last = rq;
+		}
+	}
+
+	rqs->tail = last;
+	*queue_rqs = matched_rqs;
+	return depth;
+}
+
+static void blk_mq_dispatch_queue_requests(struct rq_list *rqs, unsigned depth)
+{
+	struct request_queue *q = rq_list_peek(rqs)->q;
+
+	trace_block_unplug(q, depth, true);
+
+	/*
+	 * Peek first request and see if we have a ->queue_rqs() hook.
+	 * If we do, we can dispatch the whole list in one go.
+	 * We already know at this point that all requests belong to the
+	 * same queue, caller must ensure that's the case.
+	 */
+	if (q->mq_ops->queue_rqs) {
+		blk_mq_run_dispatch_ops(q, __blk_mq_flush_list(q, rqs));
+		if (rq_list_empty(rqs))
+			return;
+	}
+
+	blk_mq_run_dispatch_ops(q, blk_mq_issue_direct(rqs));
 }
 
-static void blk_mq_dispatch_plug_list(struct blk_plug *plug, bool from_sched)
+static void blk_mq_dispatch_list(struct rq_list *rqs, bool from_sched)
 {
 	struct blk_mq_hw_ctx *this_hctx = NULL;
 	struct blk_mq_ctx *this_ctx = NULL;
@@ -2835,7 +2886,7 @@ static void blk_mq_dispatch_plug_list(struct blk_plug *plug, bool from_sched)
 	LIST_HEAD(list);
 
 	do {
-		struct request *rq = rq_list_pop(&plug->mq_list);
+		struct request *rq = rq_list_pop(rqs);
 
 		if (!this_hctx) {
 			this_hctx = rq->mq_hctx;
@@ -2848,9 +2899,9 @@ static void blk_mq_dispatch_plug_list(struct blk_plug *plug, bool from_sched)
 		}
 		list_add_tail(&rq->queuelist, &list);
 		depth++;
-	} while (!rq_list_empty(&plug->mq_list));
+	} while (!rq_list_empty(rqs));
 
-	plug->mq_list = requeue_list;
+	*rqs = requeue_list;
 	trace_block_unplug(this_hctx->queue, depth, !from_sched);
 
 	percpu_ref_get(&this_hctx->queue->q_usage_counter);
@@ -2870,9 +2921,21 @@ static void blk_mq_dispatch_plug_list(struct blk_plug *plug, bool from_sched)
 	percpu_ref_put(&this_hctx->queue->q_usage_counter);
 }
 
+static void blk_mq_dispatch_multiple_queue_requests(struct rq_list *rqs)
+{
+	do {
+		struct rq_list queue_rqs;
+		unsigned depth;
+
+		depth = blk_mq_extract_queue_requests(rqs, &queue_rqs);
+		blk_mq_dispatch_queue_requests(&queue_rqs, depth);
+		while (!rq_list_empty(&queue_rqs))
+			blk_mq_dispatch_list(&queue_rqs, false);
+	} while (!rq_list_empty(rqs));
+}
+
 void blk_mq_flush_plug_list(struct blk_plug *plug, bool from_schedule)
 {
-	struct request *rq;
 	unsigned int depth;
 
 	/*
@@ -2887,34 +2950,19 @@ void blk_mq_flush_plug_list(struct blk_plug *plug, bool from_schedule)
 	depth = plug->rq_count;
 	plug->rq_count = 0;
 
-	if (!plug->multiple_queues && !plug->has_elevator && !from_schedule) {
-		struct request_queue *q;
-
-		rq = rq_list_peek(&plug->mq_list);
-		q = rq->q;
-		trace_block_unplug(q, depth, true);
-
-		/*
-		 * Peek first request and see if we have a ->queue_rqs() hook.
-		 * If we do, we can dispatch the whole plug list in one go. We
-		 * already know at this point that all requests belong to the
-		 * same queue, caller must ensure that's the case.
-		 */
-		if (q->mq_ops->queue_rqs) {
-			blk_mq_run_dispatch_ops(q,
-				__blk_mq_flush_plug_list(q, plug));
-			if (rq_list_empty(&plug->mq_list))
-				return;
+	if (!plug->has_elevator && !from_schedule) {
+		if (plug->multiple_queues) {
+			blk_mq_dispatch_multiple_queue_requests(&plug->mq_list);
+			return;
 		}
 
-		blk_mq_run_dispatch_ops(q,
-				blk_mq_plug_issue_direct(plug));
+		blk_mq_dispatch_queue_requests(&plug->mq_list, depth);
 		if (rq_list_empty(&plug->mq_list))
 			return;
 	}
 
 	do {
-		blk_mq_dispatch_plug_list(plug, from_schedule);
+		blk_mq_dispatch_list(&plug->mq_list, from_schedule);
 	} while (!rq_list_empty(&plug->mq_list));
 }
 
@@ -2969,8 +3017,14 @@ static struct request *blk_mq_get_new_requests(struct request_queue *q,
 {
 	struct blk_mq_alloc_data data = {
 		.q		= q,
-		.nr_tags	= 1,
+		.flags		= 0,
+		.shallow_depth	= 0,
 		.cmd_flags	= bio->bi_opf,
+		.rq_flags	= 0,
+		.nr_tags	= 1,
+		.cached_rqs	= NULL,
+		.ctx		= NULL,
+		.hctx		= NULL
 	};
 	struct request *rq;
 
@@ -3080,8 +3134,6 @@ void blk_mq_submit_bio(struct bio *bio)
 		goto new_request;
 	}
 
-	bio = blk_queue_bounce(bio, q);
-
 	/*
 	 * The cached request already holds a q_usage_counter reference and we
 	 * don't have to acquire a new one if we use it.
@@ -4094,8 +4146,6 @@ static void blk_mq_map_swqueue(struct request_queue *q)
 	struct blk_mq_ctx *ctx;
 	struct blk_mq_tag_set *set = q->tag_set;
 
-	mutex_lock(&q->elevator_lock);
-
 	queue_for_each_hw_ctx(q, hctx, i) {
 		cpumask_clear(hctx->cpumask);
 		hctx->nr_ctx = 0;
@@ -4200,8 +4250,6 @@ static void blk_mq_map_swqueue(struct request_queue *q)
 		hctx->next_cpu = blk_mq_first_mapped_cpu(hctx);
 		hctx->next_cpu_batch = BLK_MQ_CPU_WORK_BATCH;
 	}
-
-	mutex_unlock(&q->elevator_lock);
 }
 
 /*
@@ -4505,16 +4553,9 @@ static void __blk_mq_realloc_hw_ctxs(struct blk_mq_tag_set *set,
 }
 
 static void blk_mq_realloc_hw_ctxs(struct blk_mq_tag_set *set,
-				   struct request_queue *q, bool lock)
+				   struct request_queue *q)
 {
-	if (lock) {
-		/* protect against switching io scheduler  */
-		mutex_lock(&q->elevator_lock);
-		__blk_mq_realloc_hw_ctxs(set, q);
-		mutex_unlock(&q->elevator_lock);
-	} else {
-		__blk_mq_realloc_hw_ctxs(set, q);
-	}
+	__blk_mq_realloc_hw_ctxs(set, q);
 
 	/* unregister cpuhp callbacks for exited hctxs */
 	blk_mq_remove_hw_queues_cpuhp(q);
@@ -4546,7 +4587,7 @@ int blk_mq_init_allocated_queue(struct blk_mq_tag_set *set,
 
 	xa_init(&q->hctx_table);
 
-	blk_mq_realloc_hw_ctxs(set, q, false);
+	blk_mq_realloc_hw_ctxs(set, q);
 	if (!q->nr_hw_queues)
 		goto err_hctxs;
 
@@ -4563,8 +4604,8 @@ int blk_mq_init_allocated_queue(struct blk_mq_tag_set *set,
 	q->nr_requests = set->queue_depth;
 
 	blk_mq_init_cpu_queues(q, set->nr_hw_queues);
-	blk_mq_add_queue_tag_set(set, q);
 	blk_mq_map_swqueue(q);
+	blk_mq_add_queue_tag_set(set, q);
 	return 0;
 
 err_hctxs:
@@ -4784,6 +4825,8 @@ int blk_mq_alloc_tag_set(struct blk_mq_tag_set *set)
 			goto out_free_srcu;
 	}
 
+	init_rwsem(&set->update_nr_hwq_lock);
+
 	ret = -ENOMEM;
 	set->tags = kcalloc_node(set->nr_hw_queues,
 				 sizeof(struct blk_mq_tags *), GFP_KERNEL,
@@ -4923,88 +4966,10 @@ int blk_mq_update_nr_requests(struct request_queue *q, unsigned int nr)
 	return ret;
 }
 
-/*
- * request_queue and elevator_type pair.
- * It is just used by __blk_mq_update_nr_hw_queues to cache
- * the elevator_type associated with a request_queue.
- */
-struct blk_mq_qe_pair {
-	struct list_head node;
-	struct request_queue *q;
-	struct elevator_type *type;
-};
-
-/*
- * Cache the elevator_type in qe pair list and switch the
- * io scheduler to 'none'
- */
-static bool blk_mq_elv_switch_none(struct list_head *head,
-		struct request_queue *q)
-{
-	struct blk_mq_qe_pair *qe;
-
-	qe = kmalloc(sizeof(*qe), GFP_NOIO | __GFP_NOWARN | __GFP_NORETRY);
-	if (!qe)
-		return false;
-
-	/* Accessing q->elevator needs protection from ->elevator_lock. */
-	mutex_lock(&q->elevator_lock);
-
-	if (!q->elevator) {
-		kfree(qe);
-		goto unlock;
-	}
-
-	INIT_LIST_HEAD(&qe->node);
-	qe->q = q;
-	qe->type = q->elevator->type;
-	/* keep a reference to the elevator module as we'll switch back */
-	__elevator_get(qe->type);
-	list_add(&qe->node, head);
-	elevator_disable(q);
-unlock:
-	mutex_unlock(&q->elevator_lock);
-
-	return true;
-}
-
-static struct blk_mq_qe_pair *blk_lookup_qe_pair(struct list_head *head,
-						struct request_queue *q)
-{
-	struct blk_mq_qe_pair *qe;
-
-	list_for_each_entry(qe, head, node)
-		if (qe->q == q)
-			return qe;
-
-	return NULL;
-}
-
-static void blk_mq_elv_switch_back(struct list_head *head,
-				  struct request_queue *q)
-{
-	struct blk_mq_qe_pair *qe;
-	struct elevator_type *t;
-
-	qe = blk_lookup_qe_pair(head, q);
-	if (!qe)
-		return;
-	t = qe->type;
-	list_del(&qe->node);
-	kfree(qe);
-
-	mutex_lock(&q->elevator_lock);
-	elevator_switch(q, t);
-	/* drop the reference acquired in blk_mq_elv_switch_none */
-	elevator_put(t);
-	mutex_unlock(&q->elevator_lock);
-}
-
 static void __blk_mq_update_nr_hw_queues(struct blk_mq_tag_set *set,
 							int nr_hw_queues)
 {
 	struct request_queue *q;
-	LIST_HEAD(head);
 	int prev_nr_hw_queues = set->nr_hw_queues;
 	unsigned int memflags;
 	int i;
@@ -5019,30 +4984,24 @@ static void __blk_mq_update_nr_hw_queues(struct blk_mq_tag_set *set,
 		return;
 
 	memflags = memalloc_noio_save();
-	list_for_each_entry(q, &set->tag_list, tag_set_list)
-		blk_mq_freeze_queue_nomemsave(q);
-
-	/*
-	 * Switch IO scheduler to 'none', cleaning up the data associated
-	 * with the previous scheduler. We will switch back once we are done
-	 * updating the new sw to hw queue mappings.
-	 */
-	list_for_each_entry(q, &set->tag_list, tag_set_list)
-		if (!blk_mq_elv_switch_none(&head, q))
-			goto switch_back;
-
 	list_for_each_entry(q, &set->tag_list, tag_set_list) {
 		blk_mq_debugfs_unregister_hctxs(q);
 		blk_mq_sysfs_unregister_hctxs(q);
 	}
 
-	if (blk_mq_realloc_tag_set_tags(set, nr_hw_queues) < 0)
+	list_for_each_entry(q, &set->tag_list, tag_set_list)
+		blk_mq_freeze_queue_nomemsave(q);
+
+	if (blk_mq_realloc_tag_set_tags(set, nr_hw_queues) < 0) {
+		list_for_each_entry(q, &set->tag_list, tag_set_list)
+			blk_mq_unfreeze_queue_nomemrestore(q);
 		goto reregister;
+	}
 
 fallback:
 	blk_mq_update_queue_map(set);
 	list_for_each_entry(q, &set->tag_list, tag_set_list) {
-		blk_mq_realloc_hw_ctxs(set, q, true);
+		__blk_mq_realloc_hw_ctxs(set, q);
 
 		if (q->nr_hw_queues != set->nr_hw_queues) {
 			int i = prev_nr_hw_queues;
@@ -5058,18 +5017,18 @@ fallback:
 		blk_mq_map_swqueue(q);
 	}
 
+	/* elv_update_nr_hw_queues() unfreeze queue for us */
+	list_for_each_entry(q, &set->tag_list, tag_set_list)
+		elv_update_nr_hw_queues(q);
+
 reregister:
 	list_for_each_entry(q, &set->tag_list, tag_set_list) {
 		blk_mq_sysfs_register_hctxs(q);
 		blk_mq_debugfs_register_hctxs(q);
-	}
-
-switch_back:
-	list_for_each_entry(q, &set->tag_list, tag_set_list)
-		blk_mq_elv_switch_back(&head, q);
 
-	list_for_each_entry(q, &set->tag_list, tag_set_list)
-		blk_mq_unfreeze_queue_nomemrestore(q);
+		blk_mq_remove_hw_queues_cpuhp(q);
+		blk_mq_add_hw_queues_cpuhp(q);
+	}
 	memalloc_noio_restore(memflags);
 
 	/* Free the excess tags when nr_hw_queues shrink. */
@@ -5079,9 +5038,11 @@ switch_back:
 
 void blk_mq_update_nr_hw_queues(struct blk_mq_tag_set *set, int nr_hw_queues)
 {
+	down_write(&set->update_nr_hwq_lock);
 	mutex_lock(&set->tag_list_lock);
 	__blk_mq_update_nr_hw_queues(set, nr_hw_queues);
 	mutex_unlock(&set->tag_list_lock);
+	up_write(&set->update_nr_hwq_lock);
 }
 EXPORT_SYMBOL_GPL(blk_mq_update_nr_hw_queues);