1 files changed, 91 insertions, 138 deletions

diff --git a/block/bfq-cgroup.c b/block/bfq-cgroup.c
index 144bca006463..9fb9f3533150 100644
--- a/block/bfq-cgroup.c
+++ b/block/bfq-cgroup.c
@@ -127,7 +127,7 @@ static void bfqg_stats_update_group_wait_time(struct bfqg_stats *stats)
 	if (!bfqg_stats_waiting(stats))
 		return;
 
-	now = ktime_get_ns();
+	now = blk_time_get_ns();
 	if (now > stats->start_group_wait_time)
 		bfq_stat_add(&stats->group_wait_time,
 			      now - stats->start_group_wait_time);
@@ -144,7 +144,7 @@ static void bfqg_stats_set_start_group_wait_time(struct bfq_group *bfqg,
 		return;
 	if (bfqg == curr_bfqg)
 		return;
-	stats->start_group_wait_time = ktime_get_ns();
+	stats->start_group_wait_time = blk_time_get_ns();
 	bfqg_stats_mark_waiting(stats);
 }
 
@@ -156,7 +156,7 @@ static void bfqg_stats_end_empty_time(struct bfqg_stats *stats)
 	if (!bfqg_stats_empty(stats))
 		return;
 
-	now = ktime_get_ns();
+	now = blk_time_get_ns();
 	if (now > stats->start_empty_time)
 		bfq_stat_add(&stats->empty_time,
 			      now - stats->start_empty_time);
@@ -183,7 +183,7 @@ void bfqg_stats_set_start_empty_time(struct bfq_group *bfqg)
 	if (bfqg_stats_empty(stats))
 		return;
 
-	stats->start_empty_time = ktime_get_ns();
+	stats->start_empty_time = blk_time_get_ns();
 	bfqg_stats_mark_empty(stats);
 }
 
@@ -192,7 +192,7 @@ void bfqg_stats_update_idle_time(struct bfq_group *bfqg)
 	struct bfqg_stats *stats = &bfqg->stats;
 
 	if (bfqg_stats_idling(stats)) {
-		u64 now = ktime_get_ns();
+		u64 now = blk_time_get_ns();
 
 		if (now > stats->start_idle_time)
 			bfq_stat_add(&stats->idle_time,
@@ -205,7 +205,7 @@ void bfqg_stats_set_start_idle_time(struct bfq_group *bfqg)
 {
 	struct bfqg_stats *stats = &bfqg->stats;
 
-	stats->start_idle_time = ktime_get_ns();
+	stats->start_idle_time = blk_time_get_ns();
 	bfqg_stats_mark_idling(stats);
 }
 
@@ -224,7 +224,7 @@ void bfqg_stats_update_io_add(struct bfq_group *bfqg, struct bfq_queue *bfqq,
 {
 	blkg_rwstat_add(&bfqg->stats.queued, opf, 1);
 	bfqg_stats_end_empty_time(&bfqg->stats);
-	if (!(bfqq == ((struct bfq_data *)bfqg->bfqd)->in_service_queue))
+	if (!(bfqq == bfqg->bfqd->in_service_queue))
 		bfqg_stats_set_start_group_wait_time(bfqg, bfqq_group(bfqq));
 }
 
@@ -242,7 +242,7 @@ void bfqg_stats_update_completion(struct bfq_group *bfqg, u64 start_time_ns,
 				  u64 io_start_time_ns, blk_opf_t opf)
 {
 	struct bfqg_stats *stats = &bfqg->stats;
-	u64 now = ktime_get_ns();
+	u64 now = blk_time_get_ns();
 
 	if (now > io_start_time_ns)
 		blkg_rwstat_add(&stats->service_time, opf,
@@ -316,14 +316,12 @@ struct bfq_group *bfqq_group(struct bfq_queue *bfqq)
 
 static void bfqg_get(struct bfq_group *bfqg)
 {
-	bfqg->ref++;
+	refcount_inc(&bfqg->ref);
 }
 
 static void bfqg_put(struct bfq_group *bfqg)
 {
-	bfqg->ref--;
-
-	if (bfqg->ref == 0)
+	if (refcount_dec_and_test(&bfqg->ref))
 		kfree(bfqg);
 }
 
@@ -499,15 +497,9 @@ static struct blkcg_policy_data *bfq_cpd_alloc(gfp_t gfp)
 	bgd = kzalloc(sizeof(*bgd), gfp);
 	if (!bgd)
 		return NULL;
-	return &bgd->pd;
-}
-
-static void bfq_cpd_init(struct blkcg_policy_data *cpd)
-{
-	struct bfq_group_data *d = cpd_to_bfqgd(cpd);
 
-	d->weight = cgroup_subsys_on_dfl(io_cgrp_subsys) ?
-		CGROUP_WEIGHT_DFL : BFQ_WEIGHT_LEGACY_DFL;
+	bgd->weight = CGROUP_WEIGHT_DFL;
+	return &bgd->pd;
 }
 
 static void bfq_cpd_free(struct blkcg_policy_data *cpd)
@@ -515,12 +507,12 @@ static void bfq_cpd_free(struct blkcg_policy_data *cpd)
 	kfree(cpd_to_bfqgd(cpd));
 }
 
-static struct blkg_policy_data *bfq_pd_alloc(gfp_t gfp, struct request_queue *q,
-					     struct blkcg *blkcg)
+static struct blkg_policy_data *bfq_pd_alloc(struct gendisk *disk,
+		struct blkcg *blkcg, gfp_t gfp)
 {
 	struct bfq_group *bfqg;
 
-	bfqg = kzalloc_node(sizeof(*bfqg), gfp, q->node);
+	bfqg = kzalloc_node(sizeof(*bfqg), gfp, disk->node_id);
 	if (!bfqg)
 		return NULL;
 
@@ -530,7 +522,7 @@ static struct blkg_policy_data *bfq_pd_alloc(gfp_t gfp, struct request_queue *q,
 	}
 
 	/* see comments in bfq_bic_update_cgroup for why refcounting */
-	bfqg_get(bfqg);
+	refcount_set(&bfqg->ref, 1);
 	return &bfqg->pd;
 }
 
@@ -552,7 +544,7 @@ static void bfq_pd_init(struct blkg_policy_data *pd)
 				   */
 	bfqg->bfqd = bfqd;
 	bfqg->active_entities = 0;
-	bfqg->online = true;
+	bfqg->num_queues_with_pending_reqs = 0;
 	bfqg->rq_pos_tree = RB_ROOT;
 }
 
@@ -610,8 +602,12 @@ struct bfq_group *bfq_bio_bfqg(struct bfq_data *bfqd, struct bio *bio)
 	struct bfq_group *bfqg;
 
 	while (blkg) {
+		if (!blkg->online) {
+			blkg = blkg->parent;
+			continue;
+		}
 		bfqg = blkg_to_bfqg(blkg);
-		if (bfqg->online) {
+		if (bfqg->pd.online) {
 			bio_associate_blkg_from_css(bio, &blkg->blkcg->css);
 			return bfqg;
 		}
@@ -641,6 +637,7 @@ void bfq_bfqq_move(struct bfq_data *bfqd, struct bfq_queue *bfqq,
 {
 	struct bfq_entity *entity = &bfqq->entity;
 	struct bfq_group *old_parent = bfqq_group(bfqq);
+	bool has_pending_reqs = false;
 
 	/*
 	 * No point to move bfqq to the same group, which can happen when
@@ -661,6 +658,11 @@ void bfq_bfqq_move(struct bfq_data *bfqd, struct bfq_queue *bfqq,
 	 */
 	bfqq->ref++;
 
+	if (entity->in_groups_with_pending_reqs) {
+		has_pending_reqs = true;
+		bfq_del_bfqq_in_groups_with_pending_reqs(bfqq);
+	}
+
 	/* If bfqq is empty, then bfq_bfqq_expire also invokes
 	 * bfq_del_bfqq_busy, thereby removing bfqq and its entity
 	 * from data structures related to current group. Otherwise we
@@ -677,29 +679,67 @@ void bfq_bfqq_move(struct bfq_data *bfqd, struct bfq_queue *bfqq,
 		bfq_put_idle_entity(bfq_entity_service_tree(entity), entity);
 	bfqg_and_blkg_put(old_parent);
 
-	if (entity->parent &&
-	    entity->parent->last_bfqq_created == bfqq)
-		entity->parent->last_bfqq_created = NULL;
-	else if (bfqd->last_bfqq_created == bfqq)
-		bfqd->last_bfqq_created = NULL;
-
+	bfq_reassign_last_bfqq(bfqq, NULL);
 	entity->parent = bfqg->my_entity;
 	entity->sched_data = &bfqg->sched_data;
 	/* pin down bfqg and its associated blkg  */
 	bfqg_and_blkg_get(bfqg);
 
+	if (has_pending_reqs)
+		bfq_add_bfqq_in_groups_with_pending_reqs(bfqq);
+
 	if (bfq_bfqq_busy(bfqq)) {
 		if (unlikely(!bfqd->nonrot_with_queueing))
 			bfq_pos_tree_add_move(bfqd, bfqq);
 		bfq_activate_bfqq(bfqd, bfqq);
 	}
 
-	if (!bfqd->in_service_queue && !bfqd->rq_in_driver)
+	if (!bfqd->in_service_queue && !bfqd->tot_rq_in_driver)
 		bfq_schedule_dispatch(bfqd);
 	/* release extra ref taken above, bfqq may happen to be freed now */
 	bfq_put_queue(bfqq);
 }
 
+static void bfq_sync_bfqq_move(struct bfq_data *bfqd,
+			       struct bfq_queue *sync_bfqq,
+			       struct bfq_io_cq *bic,
+			       struct bfq_group *bfqg,
+			       unsigned int act_idx)
+{
+	struct bfq_queue *bfqq;
+
+	if (!sync_bfqq->new_bfqq && !bfq_bfqq_coop(sync_bfqq)) {
+		/* We are the only user of this bfqq, just move it */
+		if (sync_bfqq->entity.sched_data != &bfqg->sched_data)
+			bfq_bfqq_move(bfqd, sync_bfqq, bfqg);
+		return;
+	}
+
+	/*
+	 * The queue was merged to a different queue. Check
+	 * that the merge chain still belongs to the same
+	 * cgroup.
+	 */
+	for (bfqq = sync_bfqq; bfqq; bfqq = bfqq->new_bfqq)
+		if (bfqq->entity.sched_data != &bfqg->sched_data)
+			break;
+	if (bfqq) {
+		/*
+		 * Some queue changed cgroup so the merge is not valid
+		 * anymore. We cannot easily just cancel the merge (by
+		 * clearing new_bfqq) as there may be other processes
+		 * using this queue and holding refs to all queues
+		 * below sync_bfqq->new_bfqq. Similarly if the merge
+		 * already happened, we need to detach from bfqq now
+		 * so that we cannot merge bio to a request from the
+		 * old cgroup.
+		 */
+		bfq_put_cooperator(sync_bfqq);
+		bic_set_bfqq(bic, NULL, true, act_idx);
+		bfq_release_process_ref(bfqd, sync_bfqq);
+	}
+}
+
 /**
  * __bfq_bic_change_cgroup - move @bic to @bfqg.
  * @bfqd: the queue descriptor.
@@ -710,60 +750,25 @@ void bfq_bfqq_move(struct bfq_data *bfqd, struct bfq_queue *bfqq,
  * sure that the reference to cgroup is valid across the call (see
  * comments in bfq_bic_update_cgroup on this issue)
  */
-static void *__bfq_bic_change_cgroup(struct bfq_data *bfqd,
-				     struct bfq_io_cq *bic,
-				     struct bfq_group *bfqg)
+static void __bfq_bic_change_cgroup(struct bfq_data *bfqd,
+				    struct bfq_io_cq *bic,
+				    struct bfq_group *bfqg)
 {
-	struct bfq_queue *async_bfqq = bic_to_bfqq(bic, 0);
-	struct bfq_queue *sync_bfqq = bic_to_bfqq(bic, 1);
-	struct bfq_entity *entity;
+	unsigned int act_idx;
 
-	if (async_bfqq) {
-		entity = &async_bfqq->entity;
+	for (act_idx = 0; act_idx < bfqd->num_actuators; act_idx++) {
+		struct bfq_queue *async_bfqq = bic_to_bfqq(bic, false, act_idx);
+		struct bfq_queue *sync_bfqq = bic_to_bfqq(bic, true, act_idx);
 
-		if (entity->sched_data != &bfqg->sched_data) {
-			bic_set_bfqq(bic, NULL, 0);
+		if (async_bfqq &&
+		    async_bfqq->entity.sched_data != &bfqg->sched_data) {
+			bic_set_bfqq(bic, NULL, false, act_idx);
 			bfq_release_process_ref(bfqd, async_bfqq);
 		}
-	}
 
-	if (sync_bfqq) {
-		if (!sync_bfqq->new_bfqq && !bfq_bfqq_coop(sync_bfqq)) {
-			/* We are the only user of this bfqq, just move it */
-			if (sync_bfqq->entity.sched_data != &bfqg->sched_data)
-				bfq_bfqq_move(bfqd, sync_bfqq, bfqg);
-		} else {
-			struct bfq_queue *bfqq;
-
-			/*
-			 * The queue was merged to a different queue. Check
-			 * that the merge chain still belongs to the same
-			 * cgroup.
-			 */
-			for (bfqq = sync_bfqq; bfqq; bfqq = bfqq->new_bfqq)
-				if (bfqq->entity.sched_data !=
-				    &bfqg->sched_data)
-					break;
-			if (bfqq) {
-				/*
-				 * Some queue changed cgroup so the merge is
-				 * not valid anymore. We cannot easily just
-				 * cancel the merge (by clearing new_bfqq) as
-				 * there may be other processes using this
-				 * queue and holding refs to all queues below
-				 * sync_bfqq->new_bfqq. Similarly if the merge
-				 * already happened, we need to detach from
-				 * bfqq now so that we cannot merge bio to a
-				 * request from the old cgroup.
-				 */
-				bfq_put_cooperator(sync_bfqq);
-				bfq_release_process_ref(bfqd, sync_bfqq);
-				bic_set_bfqq(bic, NULL, 1);
-			}
-		}
+		if (sync_bfqq)
+			bfq_sync_bfqq_move(bfqd, sync_bfqq, bic, bfqg, act_idx);
 	}
-
-	return bfqg;
 }
 
 void bfq_bic_update_cgroup(struct bfq_io_cq *bic, struct bio *bio)
@@ -787,57 +792,6 @@ void bfq_bic_update_cgroup(struct bfq_io_cq *bic, struct bio *bio)
 	 */
 	bfq_link_bfqg(bfqd, bfqg);
 	__bfq_bic_change_cgroup(bfqd, bic, bfqg);
-	/*
-	 * Update blkg_path for bfq_log_* functions. We cache this
-	 * path, and update it here, for the following
-	 * reasons. Operations on blkg objects in blk-cgroup are
-	 * protected with the request_queue lock, and not with the
-	 * lock that protects the instances of this scheduler
-	 * (bfqd->lock). This exposes BFQ to the following sort of
-	 * race.
-	 *
-	 * The blkg_lookup performed in bfq_get_queue, protected
-	 * through rcu, may happen to return the address of a copy of
-	 * the original blkg. If this is the case, then the
-	 * bfqg_and_blkg_get performed in bfq_get_queue, to pin down
-	 * the blkg, is useless: it does not prevent blk-cgroup code
-	 * from destroying both the original blkg and all objects
-	 * directly or indirectly referred by the copy of the
-	 * blkg.
-	 *
-	 * On the bright side, destroy operations on a blkg invoke, as
-	 * a first step, hooks of the scheduler associated with the
-	 * blkg. And these hooks are executed with bfqd->lock held for
-	 * BFQ. As a consequence, for any blkg associated with the
-	 * request queue this instance of the scheduler is attached
-	 * to, we are guaranteed that such a blkg is not destroyed, and
-	 * that all the pointers it contains are consistent, while we
-	 * are holding bfqd->lock. A blkg_lookup performed with
-	 * bfqd->lock held then returns a fully consistent blkg, which
-	 * remains consistent until this lock is held.
-	 *
-	 * Thanks to the last fact, and to the fact that: (1) bfqg has
-	 * been obtained through a blkg_lookup in the above
-	 * assignment, and (2) bfqd->lock is being held, here we can
-	 * safely use the policy data for the involved blkg (i.e., the
-	 * field bfqg->pd) to get to the blkg associated with bfqg,
-	 * and then we can safely use any field of blkg. After we
-	 * release bfqd->lock, even just getting blkg through this
-	 * bfqg may cause dangling references to be traversed, as
-	 * bfqg->pd may not exist any more.
-	 *
-	 * In view of the above facts, here we cache, in the bfqg, any
-	 * blkg data we may need for this bic, and for its associated
-	 * bfq_queue. As of now, we need to cache only the path of the
-	 * blkg, which is used in the bfq_log_* functions.
-	 *
-	 * Finally, note that bfqg itself needs to be protected from
-	 * destruction on the blkg_free of the original blkg (which
-	 * invokes bfq_pd_free). We use an additional private
-	 * refcounter for bfqg, to let it disappear only after no
-	 * bfq_queue refers to it any longer.
-	 */
-	blkg_path(bfqg_to_blkg(bfqg), bfqg->blkg_path, sizeof(bfqg->blkg_path));
 	bic->blkcg_serial_nr = serial_nr;
 }
 
@@ -968,7 +922,6 @@ static void bfq_pd_offline(struct blkg_policy_data *pd)
 
 put_async_queues:
 	bfq_put_async_queues(bfqd, bfqg);
-	bfqg->online = false;
 
 	spin_unlock_irqrestore(&bfqd->lock, flags);
 	/*
@@ -1096,9 +1049,11 @@ static ssize_t bfq_io_set_device_weight(struct kernfs_open_file *of,
 	struct bfq_group *bfqg;
 	u64 v;
 
-	ret = blkg_conf_prep(blkcg, &blkcg_policy_bfq, buf, &ctx);
+	blkg_conf_init(&ctx, buf);
+
+	ret = blkg_conf_prep(blkcg, &blkcg_policy_bfq, &ctx);
 	if (ret)
-		return ret;
+		goto out;
 
 	if (sscanf(ctx.body, "%llu", &v) == 1) {
 		/* require "default" on dfl */
@@ -1120,7 +1075,7 @@ static ssize_t bfq_io_set_device_weight(struct kernfs_open_file *of,
 		ret = 0;
 	}
 out:
-	blkg_conf_finish(&ctx);
+	blkg_conf_exit(&ctx);
 	return ret ?: nbytes;
 }
 
@@ -1274,7 +1229,7 @@ struct bfq_group *bfq_create_group_hierarchy(struct bfq_data *bfqd, int node)
 {
 	int ret;
 
-	ret = blkcg_activate_policy(bfqd->queue, &blkcg_policy_bfq);
+	ret = blkcg_activate_policy(bfqd->queue->disk, &blkcg_policy_bfq);
 	if (ret)
 		return NULL;
 
@@ -1286,8 +1241,6 @@ struct blkcg_policy blkcg_policy_bfq = {
 	.legacy_cftypes		= bfq_blkcg_legacy_files,
 
 	.cpd_alloc_fn		= bfq_cpd_alloc,
-	.cpd_init_fn		= bfq_cpd_init,
-	.cpd_bind_fn	        = bfq_cpd_init,
 	.cpd_free_fn		= bfq_cpd_free,
 
 	.pd_alloc_fn		= bfq_pd_alloc,