blkcg: factor out blkio_group creation

Currently both blk-throttle and cfq-iosched implement their own
blkio_group creation code in throtl_get_tg() and cfq_get_cfqg().  This
patch factors out the common code into blkg_lookup_create(), which
returns ERR_PTR value so that transitional failures due to queue
bypass can be distinguished from other failures.

* New plkio_policy_ops methods blkio_alloc_group_fn() and
  blkio_link_group_fn added.  Both are transitional and will be
  removed once the blkg management code is fully moved into
  blk-cgroup.c.

* blkio_alloc_group_fn() allocates policy-specific blkg which is
  usually a larger data structure with blkg as the first entry and
  intiailizes it.  Note that initialization of blkg proper, including
  percpu stats, is responsibility of blk-cgroup proper.

  Note that default config (weight, bps...) initialization is done
  from this method; otherwise, we end up violating locking order
  between blkcg and q locks via blkcg_get_CONF() functions.

* blkio_link_group_fn() is called under queue_lock and responsible for
  linking the blkg to the queue.  blkcg side is handled by blk-cgroup
  proper.

* The common blkg creation function is named blkg_lookup_create() and
  blkiocg_lookup_group() is renamed to blkg_lookup() for consistency.
  Also, throtl / cfq related functions are similarly [re]named for
  consistency.

This simplifies blkcg policy implementations and enables further
cleanup.

-v2: Vivek noticed that blkg_lookup_create() incorrectly tested
     blk_queue_dead() instead of blk_queue_bypass() leading a user of
     the function ending up creating a new blkg on bypassing queue.
     This is a bug introduced while relocating bypass patches before
     this one.  Fixed.

-v3: ERR_PTR patch folded into this one.  @for_root added to
     blkg_lookup_create() to allow creating root group on a bypassed
     queue during elevator switch.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>

block/blk-cgroup.c

@@ -465,38 +465,93 @@ void blkiocg_update_io_merged_stats(struct blkio_group *blkg, bool direction,
 }
 EXPORT_SYMBOL_GPL(blkiocg_update_io_merged_stats);
 
-/*
- * This function allocates the per cpu stats for blkio_group. Should be called
- * from sleepable context as alloc_per_cpu() requires that.
- */
-int blkio_alloc_blkg_stats(struct blkio_group *blkg)
+struct blkio_group *blkg_lookup_create(struct blkio_cgroup *blkcg,
+				       struct request_queue *q,
+				       enum blkio_policy_id plid,
+				       bool for_root)
+	__releases(q->queue_lock) __acquires(q->queue_lock)
 {
-	/* Allocate memory for per cpu stats */
-	blkg->stats_cpu = alloc_percpu(struct blkio_group_stats_cpu);
-	if (!blkg->stats_cpu)
-		return -ENOMEM;
-	return 0;
-}
-EXPORT_SYMBOL_GPL(blkio_alloc_blkg_stats);
+	struct blkio_policy_type *pol = blkio_policy[plid];
+	struct blkio_group *blkg, *new_blkg;
 
-void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg,
-		struct blkio_group *blkg, struct request_queue *q, dev_t dev,
-		enum blkio_policy_id plid)
-{
-	unsigned long flags;
+	WARN_ON_ONCE(!rcu_read_lock_held());
+	lockdep_assert_held(q->queue_lock);
 
-	spin_lock_irqsave(&blkcg->lock, flags);
-	spin_lock_init(&blkg->stats_lock);
-	rcu_assign_pointer(blkg->q, q);
-	blkg->blkcg_id = css_id(&blkcg->css);
+	/*
+	 * This could be the first entry point of blkcg implementation and
+	 * we shouldn't allow anything to go through for a bypassing queue.
+	 * The following can be removed if blkg lookup is guaranteed to
+	 * fail on a bypassing queue.
+	 */
+	if (unlikely(blk_queue_bypass(q)) && !for_root)
+		return ERR_PTR(blk_queue_dead(q) ? -EINVAL : -EBUSY);
+
+	blkg = blkg_lookup(blkcg, q, plid);
+	if (blkg)
+		return blkg;
+
+	if (!css_tryget(&blkcg->css))
+		return ERR_PTR(-EINVAL);
+
+	/*
+	 * Allocate and initialize.
+	 *
+	 * FIXME: The following is broken.  Percpu memory allocation
+	 * requires %GFP_KERNEL context and can't be performed from IO
+	 * path.  Allocation here should inherently be atomic and the
+	 * following lock dancing can be removed once the broken percpu
+	 * allocation is fixed.
+	 */
+	spin_unlock_irq(q->queue_lock);
+	rcu_read_unlock();
+
+	new_blkg = pol->ops.blkio_alloc_group_fn(q, blkcg);
+	if (new_blkg) {
+		new_blkg->stats_cpu = alloc_percpu(struct blkio_group_stats_cpu);
+
+		spin_lock_init(&new_blkg->stats_lock);
+		rcu_assign_pointer(new_blkg->q, q);
+		new_blkg->blkcg_id = css_id(&blkcg->css);
+		new_blkg->plid = plid;
+		cgroup_path(blkcg->css.cgroup, new_blkg->path,
+			    sizeof(new_blkg->path));
+	}
+
+	rcu_read_lock();
+	spin_lock_irq(q->queue_lock);
+	css_put(&blkcg->css);
+
+	/* did bypass get turned on inbetween? */
+	if (unlikely(blk_queue_bypass(q)) && !for_root) {
+		blkg = ERR_PTR(blk_queue_dead(q) ? -EINVAL : -EBUSY);
+		goto out;
+	}
+
+	/* did someone beat us to it? */
+	blkg = blkg_lookup(blkcg, q, plid);
+	if (unlikely(blkg))
+		goto out;
+
+	/* did alloc fail? */
+	if (unlikely(!new_blkg || !new_blkg->stats_cpu)) {
+		blkg = ERR_PTR(-ENOMEM);
+		goto out;
+	}
+
+	/* insert */
+	spin_lock(&blkcg->lock);
+	swap(blkg, new_blkg);
 	hlist_add_head_rcu(&blkg->blkcg_node, &blkcg->blkg_list);
-	blkg->plid = plid;
-	spin_unlock_irqrestore(&blkcg->lock, flags);
-	/* Need to take css reference ? */
-	cgroup_path(blkcg->css.cgroup, blkg->path, sizeof(blkg->path));
-	blkg->dev = dev;
+	pol->ops.blkio_link_group_fn(q, blkg);
+	spin_unlock(&blkcg->lock);
+out:
+	if (new_blkg) {
+		free_percpu(new_blkg->stats_cpu);
+		kfree(new_blkg);
+	}
+	return blkg;
 }
-EXPORT_SYMBOL_GPL(blkiocg_add_blkio_group);
+EXPORT_SYMBOL_GPL(blkg_lookup_create);
 
 static void __blkiocg_del_blkio_group(struct blkio_group *blkg)
 {
@@ -533,9 +588,9 @@ int blkiocg_del_blkio_group(struct blkio_group *blkg)
 EXPORT_SYMBOL_GPL(blkiocg_del_blkio_group);
 
 /* called under rcu_read_lock(). */
-struct blkio_group *blkiocg_lookup_group(struct blkio_cgroup *blkcg,
-					 struct request_queue *q,
-					 enum blkio_policy_id plid)
+struct blkio_group *blkg_lookup(struct blkio_cgroup *blkcg,
+				struct request_queue *q,
+				enum blkio_policy_id plid)
 {
 	struct blkio_group *blkg;
 	struct hlist_node *n;
@@ -545,7 +600,7 @@ struct blkio_group *blkiocg_lookup_group(struct blkio_cgroup *blkcg,
 			return blkg;
 	return NULL;
 }
-EXPORT_SYMBOL_GPL(blkiocg_lookup_group);
+EXPORT_SYMBOL_GPL(blkg_lookup);
 
 void blkg_destroy_all(struct request_queue *q)
 {

block/blk-cgroup.h

@@ -204,6 +204,10 @@ extern unsigned int blkcg_get_read_iops(struct blkio_cgroup *blkcg,
 extern unsigned int blkcg_get_write_iops(struct blkio_cgroup *blkcg,
 				     dev_t dev);
 
+typedef struct blkio_group *(blkio_alloc_group_fn)(struct request_queue *q,
+						   struct blkio_cgroup *blkcg);
+typedef void (blkio_link_group_fn)(struct request_queue *q,
+			struct blkio_group *blkg);
 typedef void (blkio_unlink_group_fn)(struct request_queue *q,
 			struct blkio_group *blkg);
 typedef bool (blkio_clear_queue_fn)(struct request_queue *q);
@@ -219,6 +223,8 @@ typedef void (blkio_update_group_write_iops_fn)(struct request_queue *q,
 			struct blkio_group *blkg, unsigned int write_iops);
 
 struct blkio_policy_ops {
+	blkio_alloc_group_fn *blkio_alloc_group_fn;
+	blkio_link_group_fn *blkio_link_group_fn;
 	blkio_unlink_group_fn *blkio_unlink_group_fn;
 	blkio_clear_queue_fn *blkio_clear_queue_fn;
 	blkio_update_group_weight_fn *blkio_update_group_weight_fn;
@@ -307,14 +313,14 @@ static inline void blkiocg_set_start_empty_time(struct blkio_group *blkg) {}
 extern struct blkio_cgroup blkio_root_cgroup;
 extern struct blkio_cgroup *cgroup_to_blkio_cgroup(struct cgroup *cgroup);
 extern struct blkio_cgroup *task_blkio_cgroup(struct task_struct *tsk);
-extern void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg,
-	struct blkio_group *blkg, struct request_queue *q, dev_t dev,
-	enum blkio_policy_id plid);
-extern int blkio_alloc_blkg_stats(struct blkio_group *blkg);
 extern int blkiocg_del_blkio_group(struct blkio_group *blkg);
-extern struct blkio_group *blkiocg_lookup_group(struct blkio_cgroup *blkcg,
-						struct request_queue *q,
-						enum blkio_policy_id plid);
+extern struct blkio_group *blkg_lookup(struct blkio_cgroup *blkcg,
+				       struct request_queue *q,
+				       enum blkio_policy_id plid);
+struct blkio_group *blkg_lookup_create(struct blkio_cgroup *blkcg,
+				       struct request_queue *q,
+				       enum blkio_policy_id plid,
+				       bool for_root);
 void blkiocg_update_timeslice_used(struct blkio_group *blkg,
 					unsigned long time,
 					unsigned long unaccounted_time);
@@ -335,17 +341,11 @@ cgroup_to_blkio_cgroup(struct cgroup *cgroup) { return NULL; }
 static inline struct blkio_cgroup *
 task_blkio_cgroup(struct task_struct *tsk) { return NULL; }
 
-static inline void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg,
-		struct blkio_group *blkg, void *key, dev_t dev,
-		enum blkio_policy_id plid) {}
-
-static inline int blkio_alloc_blkg_stats(struct blkio_group *blkg) { return 0; }
-
 static inline int
 blkiocg_del_blkio_group(struct blkio_group *blkg) { return 0; }
 
-static inline struct blkio_group *
-blkiocg_lookup_group(struct blkio_cgroup *blkcg, void *key) { return NULL; }
+static inline struct blkio_group *blkg_lookup(struct blkio_cgroup *blkcg,
+					      void *key) { return NULL; }
 static inline void blkiocg_update_timeslice_used(struct blkio_group *blkg,
 						unsigned long time,
 						unsigned long unaccounted_time)