Merge branch 'for-3.5/core' of git://git.kernel.dk/linux-block

Merge block/IO core bits from Jens Axboe:
 "This is a bit bigger on the core side than usual, but that is purely
  because we decided to hold off on parts of Tejun's submission on 3.4
  to give it a bit more time to simmer.  As a consequence, it's seen a
  long cycle in for-next.

  It contains:

   - Bug fix from Dan, wrong locking type.
   - Relax splice gifting restriction from Eric.
   - A ton of updates from Tejun, primarily for blkcg.  This improves
     the code a lot, making the API nicer and cleaner, and also includes
     fixes for how we handle and tie policies and re-activate on
     switches.  The changes also include generic bug fixes.
   - A simple fix from Vivek, along with a fix for doing proper delayed
     allocation of the blkcg stats."

Fix up annoying conflict just due to different merge resolution in
Documentation/feature-removal-schedule.txt

* 'for-3.5/core' of git://git.kernel.dk/linux-block: (92 commits)
  blkcg: tg_stats_alloc_lock is an irq lock
  vmsplice: relax alignement requirements for SPLICE_F_GIFT
  blkcg: use radix tree to index blkgs from blkcg
  blkcg: fix blkcg->css ref leak in __blkg_lookup_create()
  block: fix elvpriv allocation failure handling
  block: collapse blk_alloc_request() into get_request()
  blkcg: collapse blkcg_policy_ops into blkcg_policy
  blkcg: embed struct blkg_policy_data in policy specific data
  blkcg: mass rename of blkcg API
  blkcg: style cleanups for blk-cgroup.h
  blkcg: remove blkio_group->path[]
  blkcg: blkg_rwstat_read() was missing inline
  blkcg: shoot down blkgs if all policies are deactivated
  blkcg: drop stuff unused after per-queue policy activation update
  blkcg: implement per-queue policy activation
  blkcg: add request_queue->root_blkg
  blkcg: make request_queue bypassing on allocation
  blkcg: make sure blkg_lookup() returns %NULL if @q is bypassing
  blkcg: make blkg_conf_prep() take @pol and return with queue lock held
  blkcg: remove static policy ID enums
  ...

13 files changed, 2310 insertions, 2907 deletions

diff --git a/block/Kconfig.iosched b/block/Kconfig.iosched
index 3199b76f795d..421bef9c4c48 100644
--- a/block/Kconfig.iosched
+++ b/block/Kconfig.iosched
@@ -23,8 +23,6 @@ config IOSCHED_DEADLINE
 
 config IOSCHED_CFQ
 	tristate "CFQ I/O scheduler"
-	# If BLK_CGROUP is a module, CFQ has to be built as module.
-	depends on (BLK_CGROUP=m && m) || !BLK_CGROUP || BLK_CGROUP=y
 	default y
 	---help---
 	  The CFQ I/O scheduler tries to distribute bandwidth equally
@@ -34,8 +32,6 @@ config IOSCHED_CFQ
 
 	  This is the default I/O scheduler.
 
-	  Note: If BLK_CGROUP=m, then CFQ can be built only as module.
-
 config CFQ_GROUP_IOSCHED
 	bool "CFQ Group Scheduling support"
 	depends on IOSCHED_CFQ && BLK_CGROUP
diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index 126c341955de..02cf6335e9bd 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -11,1570 +11,612 @@
  * 	              Nauman Rafique <nauman@google.com>
  */
 #include <linux/ioprio.h>
-#include <linux/seq_file.h>
 #include <linux/kdev_t.h>
 #include <linux/module.h>
 #include <linux/err.h>
 #include <linux/blkdev.h>
 #include <linux/slab.h>
-#include "blk-cgroup.h"
 #include <linux/genhd.h>
+#include <linux/delay.h>
+#include <linux/atomic.h>
+#include "blk-cgroup.h"
+#include "blk.h"
 
 #define MAX_KEY_LEN 100
 
-static DEFINE_SPINLOCK(blkio_list_lock);
-static LIST_HEAD(blkio_list);
+static DEFINE_MUTEX(blkcg_pol_mutex);
 
-struct blkio_cgroup blkio_root_cgroup = { .weight = 2*BLKIO_WEIGHT_DEFAULT };
-EXPORT_SYMBOL_GPL(blkio_root_cgroup);
+struct blkcg blkcg_root = { .cfq_weight = 2 * CFQ_WEIGHT_DEFAULT };
+EXPORT_SYMBOL_GPL(blkcg_root);
 
-/* for encoding cft->private value on file */
-#define BLKIOFILE_PRIVATE(x, val)	(((x) << 16) | (val))
-/* What policy owns the file, proportional or throttle */
-#define BLKIOFILE_POLICY(val)		(((val) >> 16) & 0xffff)
-#define BLKIOFILE_ATTR(val)		((val) & 0xffff)
+static struct blkcg_policy *blkcg_policy[BLKCG_MAX_POLS];
 
-static inline void blkio_policy_insert_node(struct blkio_cgroup *blkcg,
-					    struct blkio_policy_node *pn)
+struct blkcg *cgroup_to_blkcg(struct cgroup *cgroup)
 {
-	list_add(&pn->node, &blkcg->policy_list);
+	return container_of(cgroup_subsys_state(cgroup, blkio_subsys_id),
+			    struct blkcg, css);
 }
+EXPORT_SYMBOL_GPL(cgroup_to_blkcg);
 
-static inline bool cftype_blkg_same_policy(struct cftype *cft,
-			struct blkio_group *blkg)
+static struct blkcg *task_blkcg(struct task_struct *tsk)
 {
-	enum blkio_policy_id plid = BLKIOFILE_POLICY(cft->private);
-
-	if (blkg->plid == plid)
-		return 1;
-
-	return 0;
+	return container_of(task_subsys_state(tsk, blkio_subsys_id),
+			    struct blkcg, css);
 }
 
-/* Determines if policy node matches cgroup file being accessed */
-static inline bool pn_matches_cftype(struct cftype *cft,
-			struct blkio_policy_node *pn)
+struct blkcg *bio_blkcg(struct bio *bio)
 {
-	enum blkio_policy_id plid = BLKIOFILE_POLICY(cft->private);
-	int fileid = BLKIOFILE_ATTR(cft->private);
-
-	return (plid == pn->plid && fileid == pn->fileid);
+	if (bio && bio->bi_css)
+		return container_of(bio->bi_css, struct blkcg, css);
+	return task_blkcg(current);
 }
+EXPORT_SYMBOL_GPL(bio_blkcg);
 
-/* Must be called with blkcg->lock held */
-static inline void blkio_policy_delete_node(struct blkio_policy_node *pn)
+static bool blkcg_policy_enabled(struct request_queue *q,
+				 const struct blkcg_policy *pol)
 {
-	list_del(&pn->node);
+	return pol && test_bit(pol->plid, q->blkcg_pols);
 }
 
-/* Must be called with blkcg->lock held */
-static struct blkio_policy_node *
-blkio_policy_search_node(const struct blkio_cgroup *blkcg, dev_t dev,
-		enum blkio_policy_id plid, int fileid)
+/**
+ * blkg_free - free a blkg
+ * @blkg: blkg to free
+ *
+ * Free @blkg which may be partially allocated.
+ */
+static void blkg_free(struct blkcg_gq *blkg)
 {
-	struct blkio_policy_node *pn;
-
-	list_for_each_entry(pn, &blkcg->policy_list, node) {
-		if (pn->dev == dev && pn->plid == plid && pn->fileid == fileid)
-			return pn;
-	}
+	int i;
 
-	return NULL;
-}
+	if (!blkg)
+		return;
 
-struct blkio_cgroup *cgroup_to_blkio_cgroup(struct cgroup *cgroup)
-{
-	return container_of(cgroup_subsys_state(cgroup, blkio_subsys_id),
-			    struct blkio_cgroup, css);
-}
-EXPORT_SYMBOL_GPL(cgroup_to_blkio_cgroup);
+	for (i = 0; i < BLKCG_MAX_POLS; i++) {
+		struct blkcg_policy *pol = blkcg_policy[i];
+		struct blkg_policy_data *pd = blkg->pd[i];
 
-struct blkio_cgroup *task_blkio_cgroup(struct task_struct *tsk)
-{
-	return container_of(task_subsys_state(tsk, blkio_subsys_id),
-			    struct blkio_cgroup, css);
-}
-EXPORT_SYMBOL_GPL(task_blkio_cgroup);
+		if (!pd)
+			continue;
 
-static inline void
-blkio_update_group_weight(struct blkio_group *blkg, unsigned int weight)
-{
-	struct blkio_policy_type *blkiop;
+		if (pol && pol->pd_exit_fn)
+			pol->pd_exit_fn(blkg);
 
-	list_for_each_entry(blkiop, &blkio_list, list) {
-		/* If this policy does not own the blkg, do not send updates */
-		if (blkiop->plid != blkg->plid)
-			continue;
-		if (blkiop->ops.blkio_update_group_weight_fn)
-			blkiop->ops.blkio_update_group_weight_fn(blkg->key,
-							blkg, weight);
+		kfree(pd);
 	}
+
+	kfree(blkg);
 }
 
-static inline void blkio_update_group_bps(struct blkio_group *blkg, u64 bps,
-				int fileid)
+/**
+ * blkg_alloc - allocate a blkg
+ * @blkcg: block cgroup the new blkg is associated with
+ * @q: request_queue the new blkg is associated with
+ *
+ * Allocate a new blkg assocating @blkcg and @q.
+ */
+static struct blkcg_gq *blkg_alloc(struct blkcg *blkcg, struct request_queue *q)
 {
-	struct blkio_policy_type *blkiop;
-
-	list_for_each_entry(blkiop, &blkio_list, list) {
-
-		/* If this policy does not own the blkg, do not send updates */
-		if (blkiop->plid != blkg->plid)
-			continue;
-
-		if (fileid == BLKIO_THROTL_read_bps_device
-		    && blkiop->ops.blkio_update_group_read_bps_fn)
-			blkiop->ops.blkio_update_group_read_bps_fn(blkg->key,
-								blkg, bps);
+	struct blkcg_gq *blkg;
+	int i;
 
-		if (fileid == BLKIO_THROTL_write_bps_device
-		    && blkiop->ops.blkio_update_group_write_bps_fn)
-			blkiop->ops.blkio_update_group_write_bps_fn(blkg->key,
-								blkg, bps);
-	}
-}
+	/* alloc and init base part */
+	blkg = kzalloc_node(sizeof(*blkg), GFP_ATOMIC, q->node);
+	if (!blkg)
+		return NULL;
 
-static inline void blkio_update_group_iops(struct blkio_group *blkg,
-			unsigned int iops, int fileid)
-{
-	struct blkio_policy_type *blkiop;
+	blkg->q = q;
+	INIT_LIST_HEAD(&blkg->q_node);
+	blkg->blkcg = blkcg;
+	blkg->refcnt = 1;
 
-	list_for_each_entry(blkiop, &blkio_list, list) {
+	for (i = 0; i < BLKCG_MAX_POLS; i++) {
+		struct blkcg_policy *pol = blkcg_policy[i];
+		struct blkg_policy_data *pd;
 
-		/* If this policy does not own the blkg, do not send updates */
-		if (blkiop->plid != blkg->plid)
+		if (!blkcg_policy_enabled(q, pol))
 			continue;
 
-		if (fileid == BLKIO_THROTL_read_iops_device
-		    && blkiop->ops.blkio_update_group_read_iops_fn)
-			blkiop->ops.blkio_update_group_read_iops_fn(blkg->key,
-								blkg, iops);
+		/* alloc per-policy data and attach it to blkg */
+		pd = kzalloc_node(pol->pd_size, GFP_ATOMIC, q->node);
+		if (!pd) {
+			blkg_free(blkg);
+			return NULL;
+		}
 
-		if (fileid == BLKIO_THROTL_write_iops_device
-		    && blkiop->ops.blkio_update_group_write_iops_fn)
-			blkiop->ops.blkio_update_group_write_iops_fn(blkg->key,
-								blkg,iops);
+		blkg->pd[i] = pd;
+		pd->blkg = blkg;
 	}
-}
 
-/*
- * Add to the appropriate stat variable depending on the request type.
- * This should be called with the blkg->stats_lock held.
- */
-static void blkio_add_stat(uint64_t *stat, uint64_t add, bool direction,
-				bool sync)
-{
-	if (direction)
-		stat[BLKIO_STAT_WRITE] += add;
-	else
-		stat[BLKIO_STAT_READ] += add;
-	if (sync)
-		stat[BLKIO_STAT_SYNC] += add;
-	else
-		stat[BLKIO_STAT_ASYNC] += add;
-}
+	/* invoke per-policy init */
+	for (i = 0; i < BLKCG_MAX_POLS; i++) {
+		struct blkcg_policy *pol = blkcg_policy[i];
 
-/*
- * Decrements the appropriate stat variable if non-zero depending on the
- * request type. Panics on value being zero.
- * This should be called with the blkg->stats_lock held.
- */
-static void blkio_check_and_dec_stat(uint64_t *stat, bool direction, bool sync)
-{
-	if (direction) {
-		BUG_ON(stat[BLKIO_STAT_WRITE] == 0);
-		stat[BLKIO_STAT_WRITE]--;
-	} else {
-		BUG_ON(stat[BLKIO_STAT_READ] == 0);
-		stat[BLKIO_STAT_READ]--;
-	}
-	if (sync) {
-		BUG_ON(stat[BLKIO_STAT_SYNC] == 0);
-		stat[BLKIO_STAT_SYNC]--;
-	} else {
-		BUG_ON(stat[BLKIO_STAT_ASYNC] == 0);
-		stat[BLKIO_STAT_ASYNC]--;
+		if (blkcg_policy_enabled(blkg->q, pol))
+			pol->pd_init_fn(blkg);
 	}
-}
 
-#ifdef CONFIG_DEBUG_BLK_CGROUP
-/* This should be called with the blkg->stats_lock held. */
-static void blkio_set_start_group_wait_time(struct blkio_group *blkg,
-						struct blkio_group *curr_blkg)
-{
-	if (blkio_blkg_waiting(&blkg->stats))
-		return;
-	if (blkg == curr_blkg)
-		return;
-	blkg->stats.start_group_wait_time = sched_clock();
-	blkio_mark_blkg_waiting(&blkg->stats);
+	return blkg;
 }
 
-/* This should be called with the blkg->stats_lock held. */
-static void blkio_update_group_wait_time(struct blkio_group_stats *stats)
+static struct blkcg_gq *__blkg_lookup(struct blkcg *blkcg,
+				      struct request_queue *q)
 {
-	unsigned long long now;
+	struct blkcg_gq *blkg;
 
-	if (!blkio_blkg_waiting(stats))
-		return;
+	blkg = rcu_dereference(blkcg->blkg_hint);
+	if (blkg && blkg->q == q)
+		return blkg;
 
-	now = sched_clock();
-	if (time_after64(now, stats->start_group_wait_time))
-		stats->group_wait_time += now - stats->start_group_wait_time;
-	blkio_clear_blkg_waiting(stats);
+	/*
+	 * Hint didn't match.  Look up from the radix tree.  Note that we
+	 * may not be holding queue_lock and thus are not sure whether
+	 * @blkg from blkg_tree has already been removed or not, so we
+	 * can't update hint to the lookup result.  Leave it to the caller.
+	 */
+	blkg = radix_tree_lookup(&blkcg->blkg_tree, q->id);
+	if (blkg && blkg->q == q)
+		return blkg;
+
+	return NULL;
 }
 
-/* This should be called with the blkg->stats_lock held. */
-static void blkio_end_empty_time(struct blkio_group_stats *stats)
+/**
+ * blkg_lookup - lookup blkg for the specified blkcg - q pair
+ * @blkcg: blkcg of interest
+ * @q: request_queue of interest
+ *
+ * Lookup blkg for the @blkcg - @q pair.  This function should be called
+ * under RCU read lock and is guaranteed to return %NULL if @q is bypassing
+ * - see blk_queue_bypass_start() for details.
+ */
+struct blkcg_gq *blkg_lookup(struct blkcg *blkcg, struct request_queue *q)
 {
-	unsigned long long now;
-
-	if (!blkio_blkg_empty(stats))
-		return;
+	WARN_ON_ONCE(!rcu_read_lock_held());
 
-	now = sched_clock();
-	if (time_after64(now, stats->start_empty_time))
-		stats->empty_time += now - stats->start_empty_time;
-	blkio_clear_blkg_empty(stats);
+	if (unlikely(blk_queue_bypass(q)))
+		return NULL;
+	return __blkg_lookup(blkcg, q);
 }
+EXPORT_SYMBOL_GPL(blkg_lookup);
 
-void blkiocg_update_set_idle_time_stats(struct blkio_group *blkg)
+static struct blkcg_gq *__blkg_lookup_create(struct blkcg *blkcg,
+					     struct request_queue *q)
+	__releases(q->queue_lock) __acquires(q->queue_lock)
 {
-	unsigned long flags;
+	struct blkcg_gq *blkg;
+	int ret;
 
-	spin_lock_irqsave(&blkg->stats_lock, flags);
-	BUG_ON(blkio_blkg_idling(&blkg->stats));
-	blkg->stats.start_idle_time = sched_clock();
-	blkio_mark_blkg_idling(&blkg->stats);
-	spin_unlock_irqrestore(&blkg->stats_lock, flags);
-}
-EXPORT_SYMBOL_GPL(blkiocg_update_set_idle_time_stats);
+	WARN_ON_ONCE(!rcu_read_lock_held());
+	lockdep_assert_held(q->queue_lock);
 
-void blkiocg_update_idle_time_stats(struct blkio_group *blkg)
-{
-	unsigned long flags;
-	unsigned long long now;
-	struct blkio_group_stats *stats;
-
-	spin_lock_irqsave(&blkg->stats_lock, flags);
-	stats = &blkg->stats;
-	if (blkio_blkg_idling(stats)) {
-		now = sched_clock();
-		if (time_after64(now, stats->start_idle_time))
-			stats->idle_time += now - stats->start_idle_time;
-		blkio_clear_blkg_idling(stats);
+	/* lookup and update hint on success, see __blkg_lookup() for details */
+	blkg = __blkg_lookup(blkcg, q);
+	if (blkg) {
+		rcu_assign_pointer(blkcg->blkg_hint, blkg);
+		return blkg;
 	}
-	spin_unlock_irqrestore(&blkg->stats_lock, flags);
-}
-EXPORT_SYMBOL_GPL(blkiocg_update_idle_time_stats);
 
-void blkiocg_update_avg_queue_size_stats(struct blkio_group *blkg)
-{
-	unsigned long flags;
-	struct blkio_group_stats *stats;
-
-	spin_lock_irqsave(&blkg->stats_lock, flags);
-	stats = &blkg->stats;
-	stats->avg_queue_size_sum +=
-			stats->stat_arr[BLKIO_STAT_QUEUED][BLKIO_STAT_READ] +
-			stats->stat_arr[BLKIO_STAT_QUEUED][BLKIO_STAT_WRITE];
-	stats->avg_queue_size_samples++;
-	blkio_update_group_wait_time(stats);
-	spin_unlock_irqrestore(&blkg->stats_lock, flags);
-}
-EXPORT_SYMBOL_GPL(blkiocg_update_avg_queue_size_stats);
+	/* blkg holds a reference to blkcg */
+	if (!css_tryget(&blkcg->css))
+		return ERR_PTR(-EINVAL);
 
-void blkiocg_set_start_empty_time(struct blkio_group *blkg)
-{
-	unsigned long flags;
-	struct blkio_group_stats *stats;
+	/* allocate */
+	ret = -ENOMEM;
+	blkg = blkg_alloc(blkcg, q);
+	if (unlikely(!blkg))
+		goto err_put;
 
-	spin_lock_irqsave(&blkg->stats_lock, flags);
-	stats = &blkg->stats;
+	/* insert */
+	ret = radix_tree_preload(GFP_ATOMIC);
+	if (ret)
+		goto err_free;
 
-	if (stats->stat_arr[BLKIO_STAT_QUEUED][BLKIO_STAT_READ] ||
-			stats->stat_arr[BLKIO_STAT_QUEUED][BLKIO_STAT_WRITE]) {
-		spin_unlock_irqrestore(&blkg->stats_lock, flags);
-		return;
+	spin_lock(&blkcg->lock);
+	ret = radix_tree_insert(&blkcg->blkg_tree, q->id, blkg);
+	if (likely(!ret)) {
+		hlist_add_head_rcu(&blkg->blkcg_node, &blkcg->blkg_list);
+		list_add(&blkg->q_node, &q->blkg_list);
 	}
+	spin_unlock(&blkcg->lock);
 
-	/*
-	 * group is already marked empty. This can happen if cfqq got new
-	 * request in parent group and moved to this group while being added
-	 * to service tree. Just ignore the event and move on.
-	 */
-	if(blkio_blkg_empty(stats)) {
-		spin_unlock_irqrestore(&blkg->stats_lock, flags);
-		return;
-	}
+	radix_tree_preload_end();
 
-	stats->start_empty_time = sched_clock();
-	blkio_mark_blkg_empty(stats);
-	spin_unlock_irqrestore(&blkg->stats_lock, flags);
+	if (!ret)
+		return blkg;
+err_free:
+	blkg_free(blkg);
+err_put:
+	css_put(&blkcg->css);
+	return ERR_PTR(ret);
 }
-EXPORT_SYMBOL_GPL(blkiocg_set_start_empty_time);
 
-void blkiocg_update_dequeue_stats(struct blkio_group *blkg,
-			unsigned long dequeue)
+struct blkcg_gq *blkg_lookup_create(struct blkcg *blkcg,
+				    struct request_queue *q)
 {
-	blkg->stats.dequeue += dequeue;
-}
-EXPORT_SYMBOL_GPL(blkiocg_update_dequeue_stats);
-#else
-static inline void blkio_set_start_group_wait_time(struct blkio_group *blkg,
-					struct blkio_group *curr_blkg) {}
-static inline void blkio_end_empty_time(struct blkio_group_stats *stats) {}
-#endif
-
-void blkiocg_update_io_add_stats(struct blkio_group *blkg,
-			struct blkio_group *curr_blkg, bool direction,
-			bool sync)
-{
-	unsigned long flags;
-
-	spin_lock_irqsave(&blkg->stats_lock, flags);
-	blkio_add_stat(blkg->stats.stat_arr[BLKIO_STAT_QUEUED], 1, direction,
-			sync);
-	blkio_end_empty_time(&blkg->stats);
-	blkio_set_start_group_wait_time(blkg, curr_blkg);
-	spin_unlock_irqrestore(&blkg->stats_lock, flags);
+	/*
+	 * This could be the first entry point of blkcg implementation and
+	 * we shouldn't allow anything to go through for a bypassing queue.
+	 */
+	if (unlikely(blk_queue_bypass(q)))
+		return ERR_PTR(blk_queue_dead(q) ? -EINVAL : -EBUSY);
+	return __blkg_lookup_create(blkcg, q);
 }
-EXPORT_SYMBOL_GPL(blkiocg_update_io_add_stats);
+EXPORT_SYMBOL_GPL(blkg_lookup_create);
 
-void blkiocg_update_io_remove_stats(struct blkio_group *blkg,
-						bool direction, bool sync)
+static void blkg_destroy(struct blkcg_gq *blkg)
 {
-	unsigned long flags;
+	struct request_queue *q = blkg->q;
+	struct blkcg *blkcg = blkg->blkcg;
 
-	spin_lock_irqsave(&blkg->stats_lock, flags);
-	blkio_check_and_dec_stat(blkg->stats.stat_arr[BLKIO_STAT_QUEUED],
-					direction, sync);
-	spin_unlock_irqrestore(&blkg->stats_lock, flags);
-}
-EXPORT_SYMBOL_GPL(blkiocg_update_io_remove_stats);
+	lockdep_assert_held(q->queue_lock);
+	lockdep_assert_held(&blkcg->lock);
 
-void blkiocg_update_timeslice_used(struct blkio_group *blkg, unsigned long time,
-				unsigned long unaccounted_time)
-{
-	unsigned long flags;
-
-	spin_lock_irqsave(&blkg->stats_lock, flags);
-	blkg->stats.time += time;
-#ifdef CONFIG_DEBUG_BLK_CGROUP
-	blkg->stats.unaccounted_time += unaccounted_time;
-#endif
-	spin_unlock_irqrestore(&blkg->stats_lock, flags);
-}
-EXPORT_SYMBOL_GPL(blkiocg_update_timeslice_used);
+	/* Something wrong if we are trying to remove same group twice */
+	WARN_ON_ONCE(list_empty(&blkg->q_node));
+	WARN_ON_ONCE(hlist_unhashed(&blkg->blkcg_node));
 
-/*
- * should be called under rcu read lock or queue lock to make sure blkg pointer
- * is valid.
- */
-void blkiocg_update_dispatch_stats(struct blkio_group *blkg,
-				uint64_t bytes, bool direction, bool sync)
-{
-	struct blkio_group_stats_cpu *stats_cpu;
-	unsigned long flags;
+	radix_tree_delete(&blkcg->blkg_tree, blkg->q->id);
+	list_del_init(&blkg->q_node);
+	hlist_del_init_rcu(&blkg->blkcg_node);
 
 	/*
-	 * Disabling interrupts to provide mutual exclusion between two
-	 * writes on same cpu. It probably is not needed for 64bit. Not
-	 * optimizing that case yet.
+	 * Both setting lookup hint to and clearing it from @blkg are done
+	 * under queue_lock.  If it's not pointing to @blkg now, it never
+	 * will.  Hint assignment itself can race safely.
 	 */
-	local_irq_save(flags);
-
-	stats_cpu = this_cpu_ptr(blkg->stats_cpu);
-
-	u64_stats_update_begin(&stats_cpu->syncp);
-	stats_cpu->sectors += bytes >> 9;
-	blkio_add_stat(stats_cpu->stat_arr_cpu[BLKIO_STAT_CPU_SERVICED],
-			1, direction, sync);
-	blkio_add_stat(stats_cpu->stat_arr_cpu[BLKIO_STAT_CPU_SERVICE_BYTES],
-			bytes, direction, sync);
-	u64_stats_update_end(&stats_cpu->syncp);
-	local_irq_restore(flags);
-}
-EXPORT_SYMBOL_GPL(blkiocg_update_dispatch_stats);
-
-void blkiocg_update_completion_stats(struct blkio_group *blkg,
-	uint64_t start_time, uint64_t io_start_time, bool direction, bool sync)
-{
-	struct blkio_group_stats *stats;
-	unsigned long flags;
-	unsigned long long now = sched_clock();
-
-	spin_lock_irqsave(&blkg->stats_lock, flags);
-	stats = &blkg->stats;
-	if (time_after64(now, io_start_time))
-		blkio_add_stat(stats->stat_arr[BLKIO_STAT_SERVICE_TIME],
-				now - io_start_time, direction, sync);
-	if (time_after64(io_start_time, start_time))
-		blkio_add_stat(stats->stat_arr[BLKIO_STAT_WAIT_TIME],
-				io_start_time - start_time, direction, sync);
-	spin_unlock_irqrestore(&blkg->stats_lock, flags);
-}
-EXPORT_SYMBOL_GPL(blkiocg_update_completion_stats);
-
-/*  Merged stats are per cpu.  */
-void blkiocg_update_io_merged_stats(struct blkio_group *blkg, bool direction,
-					bool sync)
-{
-	struct blkio_group_stats_cpu *stats_cpu;
-	unsigned long flags;
+	if (rcu_dereference_raw(blkcg->blkg_hint) == blkg)
+		rcu_assign_pointer(blkcg->blkg_hint, NULL);
 
 	/*
-	 * Disabling interrupts to provide mutual exclusion between two
-	 * writes on same cpu. It probably is not needed for 64bit. Not
-	 * optimizing that case yet.
+	 * Put the reference taken at the time of creation so that when all
+	 * queues are gone, group can be destroyed.
 	 */
-	local_irq_save(flags);
-
-	stats_cpu = this_cpu_ptr(blkg->stats_cpu);
-
-	u64_stats_update_begin(&stats_cpu->syncp);
-	blkio_add_stat(stats_cpu->stat_arr_cpu[BLKIO_STAT_CPU_MERGED], 1,
-				direction, sync);
-	u64_stats_update_end(&stats_cpu->syncp);
-	local_irq_restore(flags);
+	blkg_put(blkg);
 }
-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.
+/**
+ * blkg_destroy_all - destroy all blkgs associated with a request_queue
+ * @q: request_queue of interest
+ *
+ * Destroy all blkgs associated with @q.
  */
-int blkio_alloc_blkg_stats(struct blkio_group *blkg)
+static void blkg_destroy_all(struct request_queue *q)
 {
-	/* 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 blkcg_gq *blkg, *n;
 
-void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg,
-		struct blkio_group *blkg, void *key, dev_t dev,
-		enum blkio_policy_id plid)
-{
-	unsigned long flags;
-
-	spin_lock_irqsave(&blkcg->lock, flags);
-	spin_lock_init(&blkg->stats_lock);
-	rcu_assign_pointer(blkg->key, key);
-	blkg->blkcg_id = css_id(&blkcg->css);
-	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;
-}
-EXPORT_SYMBOL_GPL(blkiocg_add_blkio_group);
+	lockdep_assert_held(q->queue_lock);
 
-static void __blkiocg_del_blkio_group(struct blkio_group *blkg)
-{
-	hlist_del_init_rcu(&blkg->blkcg_node);
-	blkg->blkcg_id = 0;
-}
+	list_for_each_entry_safe(blkg, n, &q->blkg_list, q_node) {
+		struct blkcg *blkcg = blkg->blkcg;
 
-/*
- * returns 0 if blkio_group was still on cgroup list. Otherwise returns 1
- * indicating that blk_group was unhashed by the time we got to it.
- */
-int blkiocg_del_blkio_group(struct blkio_group *blkg)
-{
-	struct blkio_cgroup *blkcg;
-	unsigned long flags;
-	struct cgroup_subsys_state *css;
-	int ret = 1;
-
-	rcu_read_lock();
-	css = css_lookup(&blkio_subsys, blkg->blkcg_id);
-	if (css) {
-		blkcg = container_of(css, struct blkio_cgroup, css);
-		spin_lock_irqsave(&blkcg->lock, flags);
-		if (!hlist_unhashed(&blkg->blkcg_node)) {
-			__blkiocg_del_blkio_group(blkg);
-			ret = 0;
-		}
-		spin_unlock_irqrestore(&blkcg->lock, flags);
+		spin_lock(&blkcg->lock);
+		blkg_destroy(blkg);
+		spin_unlock(&blkcg->lock);
 	}
-
-	rcu_read_unlock();
-	return ret;
 }
-EXPORT_SYMBOL_GPL(blkiocg_del_blkio_group);
 
-/* called under rcu_read_lock(). */
-struct blkio_group *blkiocg_lookup_group(struct blkio_cgroup *blkcg, void *key)
+static void blkg_rcu_free(struct rcu_head *rcu_head)
 {
-	struct blkio_group *blkg;
-	struct hlist_node *n;
-	void *__key;
-
-	hlist_for_each_entry_rcu(blkg, n, &blkcg->blkg_list, blkcg_node) {
-		__key = blkg->key;
-		if (__key == key)
-			return blkg;
-	}
-
-	return NULL;
+	blkg_free(container_of(rcu_head, struct blkcg_gq, rcu_head));
 }
-EXPORT_SYMBOL_GPL(blkiocg_lookup_group);
 
-static void blkio_reset_stats_cpu(struct blkio_group *blkg)
+void __blkg_release(struct blkcg_gq *blkg)
 {
-	struct blkio_group_stats_cpu *stats_cpu;
-	int i, j, k;
+	/* release the extra blkcg reference this blkg has been holding */
+	css_put(&blkg->blkcg->css);
+
 	/*
-	 * Note: On 64 bit arch this should not be an issue. This has the
-	 * possibility of returning some inconsistent value on 32bit arch
-	 * as 64bit update on 32bit is non atomic. Taking care of this
-	 * corner case makes code very complicated, like sending IPIs to
-	 * cpus, taking care of stats of offline cpus etc.
+	 * A group is freed in rcu manner. But having an rcu lock does not
+	 * mean that one can access all the fields of blkg and assume these
+	 * are valid. For example, don't try to follow throtl_data and
+	 * request queue links.
 	 *
-	 * reset stats is anyway more of a debug feature and this sounds a
-	 * corner case. So I am not complicating the code yet until and
-	 * unless this becomes a real issue.
+	 * Having a reference to blkg under an rcu allows acess to only
+	 * values local to groups like group stats and group rate limits
 	 */
-	for_each_possible_cpu(i) {
-		stats_cpu = per_cpu_ptr(blkg->stats_cpu, i);
-		stats_cpu->sectors = 0;
-		for(j = 0; j < BLKIO_STAT_CPU_NR; j++)
-			for (k = 0; k < BLKIO_STAT_TOTAL; k++)
-				stats_cpu->stat_arr_cpu[j][k] = 0;
-	}
+	call_rcu(&blkg->rcu_head, blkg_rcu_free);
 }
+EXPORT_SYMBOL_GPL(__blkg_release);
 
-static int
-blkiocg_reset_stats(struct cgroup *cgroup, struct cftype *cftype, u64 val)
+static int blkcg_reset_stats(struct cgroup *cgroup, struct cftype *cftype,
+			     u64 val)
 {
-	struct blkio_cgroup *blkcg;
-	struct blkio_group *blkg;
-	struct blkio_group_stats *stats;
+	struct blkcg *blkcg = cgroup_to_blkcg(cgroup);
+	struct blkcg_gq *blkg;
 	struct hlist_node *n;
-	uint64_t queued[BLKIO_STAT_TOTAL];
 	int i;
-#ifdef CONFIG_DEBUG_BLK_CGROUP
-	bool idling, waiting, empty;
-	unsigned long long now = sched_clock();
-#endif
 
-	blkcg = cgroup_to_blkio_cgroup(cgroup);
+	mutex_lock(&blkcg_pol_mutex);
 	spin_lock_irq(&blkcg->lock);
-	hlist_for_each_entry(blkg, n, &blkcg->blkg_list, blkcg_node) {
-		spin_lock(&blkg->stats_lock);
-		stats = &blkg->stats;
-#ifdef CONFIG_DEBUG_BLK_CGROUP
-		idling = blkio_blkg_idling(stats);
-		waiting = blkio_blkg_waiting(stats);
-		empty = blkio_blkg_empty(stats);
-#endif
-		for (i = 0; i < BLKIO_STAT_TOTAL; i++)
-			queued[i] = stats->stat_arr[BLKIO_STAT_QUEUED][i];
-		memset(stats, 0, sizeof(struct blkio_group_stats));
-		for (i = 0; i < BLKIO_STAT_TOTAL; i++)
-			stats->stat_arr[BLKIO_STAT_QUEUED][i] = queued[i];
-#ifdef CONFIG_DEBUG_BLK_CGROUP
-		if (idling) {
-			blkio_mark_blkg_idling(stats);
-			stats->start_idle_time = now;
-		}
-		if (waiting) {
-			blkio_mark_blkg_waiting(stats);
-			stats->start_group_wait_time = now;
-		}
-		if (empty) {
-			blkio_mark_blkg_empty(stats);
-			stats->start_empty_time = now;
-		}
-#endif
-		spin_unlock(&blkg->stats_lock);
-
-		/* Reset Per cpu stats which don't take blkg->stats_lock */
-		blkio_reset_stats_cpu(blkg);
-	}
-
-	spin_unlock_irq(&blkcg->lock);
-	return 0;
-}
-
-static void blkio_get_key_name(enum stat_sub_type type, dev_t dev, char *str,
-				int chars_left, bool diskname_only)
-{
-	snprintf(str, chars_left, "%d:%d", MAJOR(dev), MINOR(dev));
-	chars_left -= strlen(str);
-	if (chars_left <= 0) {
-		printk(KERN_WARNING
-			"Possibly incorrect cgroup stat display format");
-		return;
-	}
-	if (diskname_only)
-		return;
-	switch (type) {
-	case BLKIO_STAT_READ:
-		strlcat(str, " Read", chars_left);
-		break;
-	case BLKIO_STAT_WRITE:
-		strlcat(str, " Write", chars_left);
-		break;
-	case BLKIO_STAT_SYNC:
-		strlcat(str, " Sync", chars_left);
-		break;
-	case BLKIO_STAT_ASYNC:
-		strlcat(str, " Async", chars_left);
-		break;
-	case BLKIO_STAT_TOTAL:
-		strlcat(str, " Total", chars_left);
-		break;
-	default:
-		strlcat(str, " Invalid", chars_left);
-	}
-}
-
-static uint64_t blkio_fill_stat(char *str, int chars_left, uint64_t val,
-				struct cgroup_map_cb *cb, dev_t dev)
-{
-	blkio_get_key_name(0, dev, str, chars_left, true);
-	cb->fill(cb, str, val);
-	return val;
-}
-
-
-static uint64_t blkio_read_stat_cpu(struct blkio_group *blkg,
-			enum stat_type_cpu type, enum stat_sub_type sub_type)
-{
-	int cpu;
-	struct blkio_group_stats_cpu *stats_cpu;
-	u64 val = 0, tval;
-
-	for_each_possible_cpu(cpu) {
-		unsigned int start;
-		stats_cpu  = per_cpu_ptr(blkg->stats_cpu, cpu);
-
-		do {
-			start = u64_stats_fetch_begin(&stats_cpu->syncp);
-			if (type == BLKIO_STAT_CPU_SECTORS)
-				tval = stats_cpu->sectors;
-			else
-				tval = stats_cpu->stat_arr_cpu[type][sub_type];
-		} while(u64_stats_fetch_retry(&stats_cpu->syncp, start));
-
-		val += tval;
-	}
-
-	return val;
-}
-
-static uint64_t blkio_get_stat_cpu(struct blkio_group *blkg,
-		struct cgroup_map_cb *cb, dev_t dev, enum stat_type_cpu type)
-{
-	uint64_t disk_total, val;
-	char key_str[MAX_KEY_LEN];
-	enum stat_sub_type sub_type;
 
-	if (type == BLKIO_STAT_CPU_SECTORS) {
-		val = blkio_read_stat_cpu(blkg, type, 0);
-		return blkio_fill_stat(key_str, MAX_KEY_LEN - 1, val, cb, dev);
-	}
-
-	for (sub_type = BLKIO_STAT_READ; sub_type < BLKIO_STAT_TOTAL;
-			sub_type++) {
-		blkio_get_key_name(sub_type, dev, key_str, MAX_KEY_LEN, false);
-		val = blkio_read_stat_cpu(blkg, type, sub_type);
-		cb->fill(cb, key_str, val);
-	}
-
-	disk_total = blkio_read_stat_cpu(blkg, type, BLKIO_STAT_READ) +
-			blkio_read_stat_cpu(blkg, type, BLKIO_STAT_WRITE);
-
-	blkio_get_key_name(BLKIO_STAT_TOTAL, dev, key_str, MAX_KEY_LEN, false);
-	cb->fill(cb, key_str, disk_total);
-	return disk_total;
-}
-
-/* This should be called with blkg->stats_lock held */
-static uint64_t blkio_get_stat(struct blkio_group *blkg,
-		struct cgroup_map_cb *cb, dev_t dev, enum stat_type type)
-{
-	uint64_t disk_total;
-	char key_str[MAX_KEY_LEN];
-	enum stat_sub_type sub_type;
-
-	if (type == BLKIO_STAT_TIME)
-		return blkio_fill_stat(key_str, MAX_KEY_LEN - 1,
-					blkg->stats.time, cb, dev);
-#ifdef CONFIG_DEBUG_BLK_CGROUP
-	if (type == BLKIO_STAT_UNACCOUNTED_TIME)
-		return blkio_fill_stat(key_str, MAX_KEY_LEN - 1,
-					blkg->stats.unaccounted_time, cb, dev);
-	if (type == BLKIO_STAT_AVG_QUEUE_SIZE) {
-		uint64_t sum = blkg->stats.avg_queue_size_sum;
-		uint64_t samples = blkg->stats.avg_queue_size_samples;
-		if (samples)
-			do_div(sum, samples);
-		else
-			sum = 0;
-		return blkio_fill_stat(key_str, MAX_KEY_LEN - 1, sum, cb, dev);
-	}
-	if (type == BLKIO_STAT_GROUP_WAIT_TIME)
-		return blkio_fill_stat(key_str, MAX_KEY_LEN - 1,
-					blkg->stats.group_wait_time, cb, dev);
-	if (type == BLKIO_STAT_IDLE_TIME)
-		return blkio_fill_stat(key_str, MAX_KEY_LEN - 1,
-					blkg->stats.idle_time, cb, dev);
-	if (type == BLKIO_STAT_EMPTY_TIME)
-		return blkio_fill_stat(key_str, MAX_KEY_LEN - 1,
-					blkg->stats.empty_time, cb, dev);
-	if (type == BLKIO_STAT_DEQUEUE)
-		return blkio_fill_stat(key_str, MAX_KEY_LEN - 1,
-					blkg->stats.dequeue, cb, dev);
-#endif
-
-	for (sub_type = BLKIO_STAT_READ; sub_type < BLKIO_STAT_TOTAL;
-			sub_type++) {
-		blkio_get_key_name(sub_type, dev, key_str, MAX_KEY_LEN, false);
-		cb->fill(cb, key_str, blkg->stats.stat_arr[type][sub_type]);
-	}
-	disk_total = blkg->stats.stat_arr[type][BLKIO_STAT_READ] +
-			blkg->stats.stat_arr[type][BLKIO_STAT_WRITE];
-	blkio_get_key_name(BLKIO_STAT_TOTAL, dev, key_str, MAX_KEY_LEN, false);
-	cb->fill(cb, key_str, disk_total);
-	return disk_total;
-}
-
-static int blkio_policy_parse_and_set(char *buf,
-	struct blkio_policy_node *newpn, enum blkio_policy_id plid, int fileid)
-{
-	struct gendisk *disk = NULL;
-	char *s[4], *p, *major_s = NULL, *minor_s = NULL;
-	unsigned long major, minor;
-	int i = 0, ret = -EINVAL;
-	int part;
-	dev_t dev;
-	u64 temp;
-
-	memset(s, 0, sizeof(s));
-
-	while ((p = strsep(&buf, " ")) != NULL) {
-		if (!*p)
-			continue;
-
-		s[i++] = p;
-
-		/* Prevent from inputing too many things */
-		if (i == 3)
-			break;
-	}
-
-	if (i != 2)
-		goto out;
-
-	p = strsep(&s[0], ":");
-	if (p != NULL)
-		major_s = p;
-	else
-		goto out;
-
-	minor_s = s[0];
-	if (!minor_s)
-		goto out;
-
-	if (strict_strtoul(major_s, 10, &major))
-		goto out;
-
-	if (strict_strtoul(minor_s, 10, &minor))
-		goto out;
-
-	dev = MKDEV(major, minor);
-
-	if (strict_strtoull(s[1], 10, &temp))
-		goto out;
-
-	/* For rule removal, do not check for device presence. */
-	if (temp) {
-		disk = get_gendisk(dev, &part);
-		if (!disk || part) {
-			ret = -ENODEV;
-			goto out;
-		}
-	}
-
-	newpn->dev = dev;
-
-	switch (plid) {
-	case BLKIO_POLICY_PROP:
-		if ((temp < BLKIO_WEIGHT_MIN && temp > 0) ||
-		     temp > BLKIO_WEIGHT_MAX)
-			goto out;
-
-		newpn->plid = plid;
-		newpn->fileid = fileid;
-		newpn->val.weight = temp;
-		break;
-	case BLKIO_POLICY_THROTL:
-		switch(fileid) {
-		case BLKIO_THROTL_read_bps_device:
-		case BLKIO_THROTL_write_bps_device:
-			newpn->plid = plid;
-			newpn->fileid = fileid;
-			newpn->val.bps = temp;
-			break;
-		case BLKIO_THROTL_read_iops_device:
-		case BLKIO_THROTL_write_iops_device:
-			if (temp > THROTL_IOPS_MAX)
-				goto out;
-
-			newpn->plid = plid;
-			newpn->fileid = fileid;
-			newpn->val.iops = (unsigned int)temp;
-			break;
-		}
-		break;
-	default:
-		BUG();
-	}
-	ret = 0;
-out:
-	put_disk(disk);
-	return ret;
-}
-
-unsigned int blkcg_get_weight(struct blkio_cgroup *blkcg,
-			      dev_t dev)
-{
-	struct blkio_policy_node *pn;
-	unsigned long flags;
-	unsigned int weight;
-
-	spin_lock_irqsave(&blkcg->lock, flags);
-
-	pn = blkio_policy_search_node(blkcg, dev, BLKIO_POLICY_PROP,
-				BLKIO_PROP_weight_device);
-	if (pn)
-		weight = pn->val.weight;
-	else
-		weight = blkcg->weight;
-
-	spin_unlock_irqrestore(&blkcg->lock, flags);
-
-	return weight;
-}
-EXPORT_SYMBOL_GPL(blkcg_get_weight);
-
-uint64_t blkcg_get_read_bps(struct blkio_cgroup *blkcg, dev_t dev)
-{
-	struct blkio_policy_node *pn;
-	unsigned long flags;
-	uint64_t bps = -1;
-
-	spin_lock_irqsave(&blkcg->lock, flags);
-	pn = blkio_policy_search_node(blkcg, dev, BLKIO_POLICY_THROTL,
-				BLKIO_THROTL_read_bps_device);
-	if (pn)
-		bps = pn->val.bps;
-	spin_unlock_irqrestore(&blkcg->lock, flags);
-
-	return bps;
-}
-
-uint64_t blkcg_get_write_bps(struct blkio_cgroup *blkcg, dev_t dev)
-{
-	struct blkio_policy_node *pn;
-	unsigned long flags;
-	uint64_t bps = -1;
-
-	spin_lock_irqsave(&blkcg->lock, flags);
-	pn = blkio_policy_search_node(blkcg, dev, BLKIO_POLICY_THROTL,
-				BLKIO_THROTL_write_bps_device);
-	if (pn)
-		bps = pn->val.bps;
-	spin_unlock_irqrestore(&blkcg->lock, flags);
-
-	return bps;
-}
-
-unsigned int blkcg_get_read_iops(struct blkio_cgroup *blkcg, dev_t dev)
-{
-	struct blkio_policy_node *pn;
-	unsigned long flags;
-	unsigned int iops = -1;
-
-	spin_lock_irqsave(&blkcg->lock, flags);
-	pn = blkio_policy_search_node(blkcg, dev, BLKIO_POLICY_THROTL,
-				BLKIO_THROTL_read_iops_device);
-	if (pn)
-		iops = pn->val.iops;
-	spin_unlock_irqrestore(&blkcg->lock, flags);
-
-	return iops;
-}
-
-unsigned int blkcg_get_write_iops(struct blkio_cgroup *blkcg, dev_t dev)
-{
-	struct blkio_policy_node *pn;
-	unsigned long flags;
-	unsigned int iops = -1;
-
-	spin_lock_irqsave(&blkcg->lock, flags);
-	pn = blkio_policy_search_node(blkcg, dev, BLKIO_POLICY_THROTL,
-				BLKIO_THROTL_write_iops_device);
-	if (pn)
-		iops = pn->val.iops;
-	spin_unlock_irqrestore(&blkcg->lock, flags);
-
-	return iops;
-}
+	/*
+	 * Note that stat reset is racy - it doesn't synchronize against
+	 * stat updates.  This is a debug feature which shouldn't exist
+	 * anyway.  If you get hit by a race, retry.
+	 */
+	hlist_for_each_entry(blkg, n, &blkcg->blkg_list, blkcg_node) {
+		for (i = 0; i < BLKCG_MAX_POLS; i++) {
+			struct blkcg_policy *pol = blkcg_policy[i];
 
-/* Checks whether user asked for deleting a policy rule */
-static bool blkio_delete_rule_command(struct blkio_policy_node *pn)
-{
-	switch(pn->plid) {
-	case BLKIO_POLICY_PROP:
-		if (pn->val.weight == 0)
-			return 1;
-		break;
-	case BLKIO_POLICY_THROTL:
-		switch(pn->fileid) {
-		case BLKIO_THROTL_read_bps_device:
-		case BLKIO_THROTL_write_bps_device:
-			if (pn->val.bps == 0)
-				return 1;
-			break;
-		case BLKIO_THROTL_read_iops_device:
-		case BLKIO_THROTL_write_iops_device:
-			if (pn->val.iops == 0)
-				return 1;
+			if (blkcg_policy_enabled(blkg->q, pol) &&
+			    pol->pd_reset_stats_fn)
+				pol->pd_reset_stats_fn(blkg);
 		}
-		break;
-	default:
-		BUG();
 	}
 
+	spin_unlock_irq(&blkcg->lock);
+	mutex_unlock(&blkcg_pol_mutex);
 	return 0;
 }
 
-static void blkio_update_policy_rule(struct blkio_policy_node *oldpn,
-					struct blkio_policy_node *newpn)
-{
-	switch(oldpn->plid) {
-	case BLKIO_POLICY_PROP:
-		oldpn->val.weight = newpn->val.weight;
-		break;
-	case BLKIO_POLICY_THROTL:
-		switch(newpn->fileid) {
-		case BLKIO_THROTL_read_bps_device:
-		case BLKIO_THROTL_write_bps_device:
-			oldpn->val.bps = newpn->val.bps;
-			break;
-		case BLKIO_THROTL_read_iops_device:
-		case BLKIO_THROTL_write_iops_device:
-			oldpn->val.iops = newpn->val.iops;
-		}
-		break;
-	default:
-		BUG();
-	}
-}
-
-/*
- * Some rules/values in blkg have changed. Propagate those to respective
- * policies.
- */
-static void blkio_update_blkg_policy(struct blkio_cgroup *blkcg,
-		struct blkio_group *blkg, struct blkio_policy_node *pn)
+static const char *blkg_dev_name(struct blkcg_gq *blkg)
 {
-	unsigned int weight, iops;
-	u64 bps;
-
-	switch(pn->plid) {
-	case BLKIO_POLICY_PROP:
-		weight = pn->val.weight ? pn->val.weight :
-				blkcg->weight;
-		blkio_update_group_weight(blkg, weight);
-		break;
-	case BLKIO_POLICY_THROTL:
-		switch(pn->fileid) {
-		case BLKIO_THROTL_read_bps_device:
-		case BLKIO_THROTL_write_bps_device:
-			bps = pn->val.bps ? pn->val.bps : (-1);
-			blkio_update_group_bps(blkg, bps, pn->fileid);
-			break;
-		case BLKIO_THROTL_read_iops_device:
-		case BLKIO_THROTL_write_iops_device:
-			iops = pn->val.iops ? pn->val.iops : (-1);
-			blkio_update_group_iops(blkg, iops, pn->fileid);
-			break;
-		}
-		break;
-	default:
-		BUG();
-	}
+	/* some drivers (floppy) instantiate a queue w/o disk registered */
+	if (blkg->q->backing_dev_info.dev)
+		return dev_name(blkg->q->backing_dev_info.dev);
+	return NULL;
 }
 
-/*
- * A policy node rule has been updated. Propagate this update to all the
- * block groups which might be affected by this update.
+/**
+ * blkcg_print_blkgs - helper for printing per-blkg data
+ * @sf: seq_file to print to
+ * @blkcg: blkcg of interest
+ * @prfill: fill function to print out a blkg
+ * @pol: policy in question
+ * @data: data to be passed to @prfill
+ * @show_total: to print out sum of prfill return values or not
+ *
+ * This function invokes @prfill on each blkg of @blkcg if pd for the
+ * policy specified by @pol exists.  @prfill is invoked with @sf, the
+ * policy data and @data.  If @show_total is %true, the sum of the return
+ * values from @prfill is printed with "Total" label at the end.
+ *
+ * This is to be used to construct print functions for
+ * cftype->read_seq_string method.
  */
-static void blkio_update_policy_node_blkg(struct blkio_cgroup *blkcg,
-				struct blkio_policy_node *pn)
+void blkcg_print_blkgs(struct seq_file *sf, struct blkcg *blkcg,
+		       u64 (*prfill)(struct seq_file *,
+				     struct blkg_policy_data *, int),
+		       const struct blkcg_policy *pol, int data,
+		       bool show_total)
 {
-	struct blkio_group *blkg;
+	struct blkcg_gq *blkg;
 	struct hlist_node *n;
+	u64 total = 0;
 
-	spin_lock(&blkio_list_lock);
 	spin_lock_irq(&blkcg->lock);
-
-	hlist_for_each_entry(blkg, n, &blkcg->blkg_list, blkcg_node) {
-		if (pn->dev != blkg->dev || pn->plid != blkg->plid)
-			continue;
-		blkio_update_blkg_policy(blkcg, blkg, pn);
-	}
-
+	hlist_for_each_entry(blkg, n, &blkcg->blkg_list, blkcg_node)
+		if (blkcg_policy_enabled(blkg->q, pol))
+			total += prfill(sf, blkg->pd[pol->plid], data);
 	spin_unlock_irq(&blkcg->lock);
-	spin_unlock(&blkio_list_lock);
+
+	if (show_total)
+		seq_printf(sf, "Total %llu\n", (unsigned long long)total);
 }
+EXPORT_SYMBOL_GPL(blkcg_print_blkgs);
 
-static int blkiocg_file_write(struct cgroup *cgrp, struct cftype *cft,
- 				       const char *buffer)
+/**
+ * __blkg_prfill_u64 - prfill helper for a single u64 value
+ * @sf: seq_file to print to
+ * @pd: policy private data of interest
+ * @v: value to print
+ *
+ * Print @v to @sf for the device assocaited with @pd.
+ */
+u64 __blkg_prfill_u64(struct seq_file *sf, struct blkg_policy_data *pd, u64 v)
 {
-	int ret = 0;
-	char *buf;
-	struct blkio_policy_node *newpn, *pn;
-	struct blkio_cgroup *blkcg;
-	int keep_newpn = 0;
-	enum blkio_policy_id plid = BLKIOFILE_POLICY(cft->private);
-	int fileid = BLKIOFILE_ATTR(cft->private);
-
-	buf = kstrdup(buffer, GFP_KERNEL);
-	if (!buf)
-		return -ENOMEM;
-
-	newpn = kzalloc(sizeof(*newpn), GFP_KERNEL);
-	if (!newpn) {
-		ret = -ENOMEM;
-		goto free_buf;
-	}
-
-	ret = blkio_policy_parse_and_set(buf, newpn, plid, fileid);
-	if (ret)
-		goto free_newpn;
-
-	blkcg = cgroup_to_blkio_cgroup(cgrp);
-
-	spin_lock_irq(&blkcg->lock);
-
-	pn = blkio_policy_search_node(blkcg, newpn->dev, plid, fileid);
-	if (!pn) {
-		if (!blkio_delete_rule_command(newpn)) {
-			blkio_policy_insert_node(blkcg, newpn);
-			keep_newpn = 1;
-		}
-		spin_unlock_irq(&blkcg->lock);
-		goto update_io_group;
-	}
-
-	if (blkio_delete_rule_command(newpn)) {
-		blkio_policy_delete_node(pn);
-		kfree(pn);
-		spin_unlock_irq(&blkcg->lock);
-		goto update_io_group;
-	}
-	spin_unlock_irq(&blkcg->lock);
+	const char *dname = blkg_dev_name(pd->blkg);
 
-	blkio_update_policy_rule(pn, newpn);
+	if (!dname)
+		return 0;
 
-update_io_group:
-	blkio_update_policy_node_blkg(blkcg, newpn);
-
-free_newpn:
-	if (!keep_newpn)
-		kfree(newpn);
-free_buf:
-	kfree(buf);
-	return ret;
+	seq_printf(sf, "%s %llu\n", dname, (unsigned long long)v);
+	return v;
 }
+EXPORT_SYMBOL_GPL(__blkg_prfill_u64);
 
-static void
-blkio_print_policy_node(struct seq_file *m, struct blkio_policy_node *pn)
-{
-	switch(pn->plid) {
-		case BLKIO_POLICY_PROP:
-			if (pn->fileid == BLKIO_PROP_weight_device)
-				seq_printf(m, "%u:%u\t%u\n", MAJOR(pn->dev),
-					MINOR(pn->dev), pn->val.weight);
-			break;
-		case BLKIO_POLICY_THROTL:
-			switch(pn->fileid) {
-			case BLKIO_THROTL_read_bps_device:
-			case BLKIO_THROTL_write_bps_device:
-				seq_printf(m, "%u:%u\t%llu\n", MAJOR(pn->dev),
-					MINOR(pn->dev), pn->val.bps);
-				break;
-			case BLKIO_THROTL_read_iops_device:
-			case BLKIO_THROTL_write_iops_device:
-				seq_printf(m, "%u:%u\t%u\n", MAJOR(pn->dev),
-					MINOR(pn->dev), pn->val.iops);
-				break;
-			}
-			break;
-		default:
-			BUG();
-	}
-}
+/**
+ * __blkg_prfill_rwstat - prfill helper for a blkg_rwstat
+ * @sf: seq_file to print to
+ * @pd: policy private data of interest
+ * @rwstat: rwstat to print
+ *
+ * Print @rwstat to @sf for the device assocaited with @pd.
+ */
+u64 __blkg_prfill_rwstat(struct seq_file *sf, struct blkg_policy_data *pd,
+			 const struct blkg_rwstat *rwstat)
+{
+	static const char *rwstr[] = {
+		[BLKG_RWSTAT_READ]	= "Read",
+		[BLKG_RWSTAT_WRITE]	= "Write",
+		[BLKG_RWSTAT_SYNC]	= "Sync",
+		[BLKG_RWSTAT_ASYNC]	= "Async",
+	};
+	const char *dname = blkg_dev_name(pd->blkg);
+	u64 v;
+	int i;
 
-/* cgroup files which read their data from policy nodes end up here */
-static void blkio_read_policy_node_files(struct cftype *cft,
-			struct blkio_cgroup *blkcg, struct seq_file *m)
-{
-	struct blkio_policy_node *pn;
-
-	if (!list_empty(&blkcg->policy_list)) {
-		spin_lock_irq(&blkcg->lock);
-		list_for_each_entry(pn, &blkcg->policy_list, node) {
-			if (!pn_matches_cftype(cft, pn))
-				continue;
-			blkio_print_policy_node(m, pn);
-		}
-		spin_unlock_irq(&blkcg->lock);
-	}
-}
+	if (!dname)
+		return 0;
 
-static int blkiocg_file_read(struct cgroup *cgrp, struct cftype *cft,
-				struct seq_file *m)
-{
-	struct blkio_cgroup *blkcg;
-	enum blkio_policy_id plid = BLKIOFILE_POLICY(cft->private);
-	int name = BLKIOFILE_ATTR(cft->private);
-
-	blkcg = cgroup_to_blkio_cgroup(cgrp);
-
-	switch(plid) {
-	case BLKIO_POLICY_PROP:
-		switch(name) {
-		case BLKIO_PROP_weight_device:
-			blkio_read_policy_node_files(cft, blkcg, m);
-			return 0;
-		default:
-			BUG();
-		}
-		break;
-	case BLKIO_POLICY_THROTL:
-		switch(name){
-		case BLKIO_THROTL_read_bps_device:
-		case BLKIO_THROTL_write_bps_device:
-		case BLKIO_THROTL_read_iops_device:
-		case BLKIO_THROTL_write_iops_device:
-			blkio_read_policy_node_files(cft, blkcg, m);
-			return 0;
-		default:
-			BUG();
-		}
-		break;
-	default:
-		BUG();
-	}
+	for (i = 0; i < BLKG_RWSTAT_NR; i++)
+		seq_printf(sf, "%s %s %llu\n", dname, rwstr[i],
+			   (unsigned long long)rwstat->cnt[i]);
 
-	return 0;
+	v = rwstat->cnt[BLKG_RWSTAT_READ] + rwstat->cnt[BLKG_RWSTAT_WRITE];
+	seq_printf(sf, "%s Total %llu\n", dname, (unsigned long long)v);
+	return v;
 }
 
-static int blkio_read_blkg_stats(struct blkio_cgroup *blkcg,
-		struct cftype *cft, struct cgroup_map_cb *cb,
-		enum stat_type type, bool show_total, bool pcpu)
+/**
+ * blkg_prfill_stat - prfill callback for blkg_stat
+ * @sf: seq_file to print to
+ * @pd: policy private data of interest
+ * @off: offset to the blkg_stat in @pd
+ *
+ * prfill callback for printing a blkg_stat.
+ */
+u64 blkg_prfill_stat(struct seq_file *sf, struct blkg_policy_data *pd, int off)
 {
-	struct blkio_group *blkg;
-	struct hlist_node *n;
-	uint64_t cgroup_total = 0;
-
-	rcu_read_lock();
-	hlist_for_each_entry_rcu(blkg, n, &blkcg->blkg_list, blkcg_node) {
-		if (blkg->dev) {
-			if (!cftype_blkg_same_policy(cft, blkg))
-				continue;
-			if (pcpu)
-				cgroup_total += blkio_get_stat_cpu(blkg, cb,
-						blkg->dev, type);
-			else {
-				spin_lock_irq(&blkg->stats_lock);
-				cgroup_total += blkio_get_stat(blkg, cb,
-						blkg->dev, type);
-				spin_unlock_irq(&blkg->stats_lock);
-			}
-		}
-	}
-	if (show_total)
-		cb->fill(cb, "Total", cgroup_total);
-	rcu_read_unlock();
-	return 0;
+	return __blkg_prfill_u64(sf, pd, blkg_stat_read((void *)pd + off));
 }
+EXPORT_SYMBOL_GPL(blkg_prfill_stat);
 
-/* All map kind of cgroup file get serviced by this function */
-static int blkiocg_file_read_map(struct cgroup *cgrp, struct cftype *cft,
-				struct cgroup_map_cb *cb)
+/**
+ * blkg_prfill_rwstat - prfill callback for blkg_rwstat
+ * @sf: seq_file to print to
+ * @pd: policy private data of interest
+ * @off: offset to the blkg_rwstat in @pd
+ *
+ * prfill callback for printing a blkg_rwstat.
+ */
+u64 blkg_prfill_rwstat(struct seq_file *sf, struct blkg_policy_data *pd,
+		       int off)
 {
-	struct blkio_cgroup *blkcg;
-	enum blkio_policy_id plid = BLKIOFILE_POLICY(cft->private);
-	int name = BLKIOFILE_ATTR(cft->private);
-
-	blkcg = cgroup_to_blkio_cgroup(cgrp);
-
-	switch(plid) {
-	case BLKIO_POLICY_PROP:
-		switch(name) {
-		case BLKIO_PROP_time:
-			return blkio_read_blkg_stats(blkcg, cft, cb,
-						BLKIO_STAT_TIME, 0, 0);
-		case BLKIO_PROP_sectors:
-			return blkio_read_blkg_stats(blkcg, cft, cb,
-						BLKIO_STAT_CPU_SECTORS, 0, 1);
-		case BLKIO_PROP_io_service_bytes:
-			return blkio_read_blkg_stats(blkcg, cft, cb,
-					BLKIO_STAT_CPU_SERVICE_BYTES, 1, 1);
-		case BLKIO_PROP_io_serviced:
-			return blkio_read_blkg_stats(blkcg, cft, cb,
-						BLKIO_STAT_CPU_SERVICED, 1, 1);
-		case BLKIO_PROP_io_service_time:
-			return blkio_read_blkg_stats(blkcg, cft, cb,
-						BLKIO_STAT_SERVICE_TIME, 1, 0);
-		case BLKIO_PROP_io_wait_time:
-			return blkio_read_blkg_stats(blkcg, cft, cb,
-						BLKIO_STAT_WAIT_TIME, 1, 0);
-		case BLKIO_PROP_io_merged:
-			return blkio_read_blkg_stats(blkcg, cft, cb,
-						BLKIO_STAT_CPU_MERGED, 1, 1);
-		case BLKIO_PROP_io_queued:
-			return blkio_read_blkg_stats(blkcg, cft, cb,
-						BLKIO_STAT_QUEUED, 1, 0);
-#ifdef CONFIG_DEBUG_BLK_CGROUP
-		case BLKIO_PROP_unaccounted_time:
-			return blkio_read_blkg_stats(blkcg, cft, cb,
-					BLKIO_STAT_UNACCOUNTED_TIME, 0, 0);
-		case BLKIO_PROP_dequeue:
-			return blkio_read_blkg_stats(blkcg, cft, cb,
-						BLKIO_STAT_DEQUEUE, 0, 0);
-		case BLKIO_PROP_avg_queue_size:
-			return blkio_read_blkg_stats(blkcg, cft, cb,
-					BLKIO_STAT_AVG_QUEUE_SIZE, 0, 0);
-		case BLKIO_PROP_group_wait_time:
-			return blkio_read_blkg_stats(blkcg, cft, cb,
-					BLKIO_STAT_GROUP_WAIT_TIME, 0, 0);
-		case BLKIO_PROP_idle_time:
-			return blkio_read_blkg_stats(blkcg, cft, cb,
-						BLKIO_STAT_IDLE_TIME, 0, 0);
-		case BLKIO_PROP_empty_time:
-			return blkio_read_blkg_stats(blkcg, cft, cb,
-						BLKIO_STAT_EMPTY_TIME, 0, 0);
-#endif
-		default:
-			BUG();
-		}
-		break;
-	case BLKIO_POLICY_THROTL:
-		switch(name){
-		case BLKIO_THROTL_io_service_bytes:
-			return blkio_read_blkg_stats(blkcg, cft, cb,
-						BLKIO_STAT_CPU_SERVICE_BYTES, 1, 1);
-		case BLKIO_THROTL_io_serviced:
-			return blkio_read_blkg_stats(blkcg, cft, cb,
-						BLKIO_STAT_CPU_SERVICED, 1, 1);
-		default:
-			BUG();
-		}
-		break;
-	default:
-		BUG();
-	}
+	struct blkg_rwstat rwstat = blkg_rwstat_read((void *)pd + off);
 
-	return 0;
+	return __blkg_prfill_rwstat(sf, pd, &rwstat);
 }
+EXPORT_SYMBOL_GPL(blkg_prfill_rwstat);
 
-static int blkio_weight_write(struct blkio_cgroup *blkcg, u64 val)
+/**
+ * blkg_conf_prep - parse and prepare for per-blkg config update
+ * @blkcg: target block cgroup
+ * @pol: target policy
+ * @input: input string
+ * @ctx: blkg_conf_ctx to be filled
+ *
+ * Parse per-blkg config update from @input and initialize @ctx with the
+ * result.  @ctx->blkg points to the blkg to be updated and @ctx->v the new
+ * value.  This function returns with RCU read lock and queue lock held and
+ * must be paired with blkg_conf_finish().
+ */
+int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
+		   const char *input, struct blkg_conf_ctx *ctx)
+	__acquires(rcu) __acquires(disk->queue->queue_lock)
 {
-	struct blkio_group *blkg;
-	struct hlist_node *n;
-	struct blkio_policy_node *pn;
+	struct gendisk *disk;
+	struct blkcg_gq *blkg;
+	unsigned int major, minor;
+	unsigned long long v;
+	int part, ret;
 
-	if (val < BLKIO_WEIGHT_MIN || val > BLKIO_WEIGHT_MAX)
+	if (sscanf(input, "%u:%u %llu", &major, &minor, &v) != 3)
 		return -EINVAL;
 
-	spin_lock(&blkio_list_lock);
-	spin_lock_irq(&blkcg->lock);
-	blkcg->weight = (unsigned int)val;
-
-	hlist_for_each_entry(blkg, n, &blkcg->blkg_list, blkcg_node) {
-		pn = blkio_policy_search_node(blkcg, blkg->dev,
-				BLKIO_POLICY_PROP, BLKIO_PROP_weight_device);
-		if (pn)
-			continue;
-
-		blkio_update_group_weight(blkg, blkcg->weight);
-	}
-	spin_unlock_irq(&blkcg->lock);
-	spin_unlock(&blkio_list_lock);
-	return 0;
-}
+	disk = get_gendisk(MKDEV(major, minor), &part);
+	if (!disk || part)
+		return -EINVAL;
 
-static u64 blkiocg_file_read_u64 (struct cgroup *cgrp, struct cftype *cft) {
-	struct blkio_cgroup *blkcg;
-	enum blkio_policy_id plid = BLKIOFILE_POLICY(cft->private);
-	int name = BLKIOFILE_ATTR(cft->private);
+	rcu_read_lock();
+	spin_lock_irq(disk->queue->queue_lock);
 
-	blkcg = cgroup_to_blkio_cgroup(cgrp);
+	if (blkcg_policy_enabled(disk->queue, pol))
+		blkg = blkg_lookup_create(blkcg, disk->queue);
+	else
+		blkg = ERR_PTR(-EINVAL);
 
-	switch(plid) {
-	case BLKIO_POLICY_PROP:
-		switch(name) {
-		case BLKIO_PROP_weight:
-			return (u64)blkcg->weight;
+	if (IS_ERR(blkg)) {
+		ret = PTR_ERR(blkg);
+		rcu_read_unlock();
+		spin_unlock_irq(disk->queue->queue_lock);
+		put_disk(disk);
+		/*
+		 * If queue was bypassing, we should retry.  Do so after a
+		 * short msleep().  It isn't strictly necessary but queue
+		 * can be bypassing for some time and it's always nice to
+		 * avoid busy looping.
+		 */
+		if (ret == -EBUSY) {
+			msleep(10);
+			ret = restart_syscall();
 		}
-		break;
-	default:
-		BUG();
+		return ret;
 	}
+
+	ctx->disk = disk;
+	ctx->blkg = blkg;
+	ctx->v = v;
 	return 0;
 }
+EXPORT_SYMBOL_GPL(blkg_conf_prep);
 
-static int
-blkiocg_file_write_u64(struct cgroup *cgrp, struct cftype *cft, u64 val)
+/**
+ * blkg_conf_finish - finish up per-blkg config update
+ * @ctx: blkg_conf_ctx intiailized by blkg_conf_prep()
+ *
+ * Finish up after per-blkg config update.  This function must be paired
+ * with blkg_conf_prep().
+ */
+void blkg_conf_finish(struct blkg_conf_ctx *ctx)
+	__releases(ctx->disk->queue->queue_lock) __releases(rcu)
 {
-	struct blkio_cgroup *blkcg;
-	enum blkio_policy_id plid = BLKIOFILE_POLICY(cft->private);
-	int name = BLKIOFILE_ATTR(cft->private);
-
-	blkcg = cgroup_to_blkio_cgroup(cgrp);
-
-	switch(plid) {
-	case BLKIO_POLICY_PROP:
-		switch(name) {
-		case BLKIO_PROP_weight:
-			return blkio_weight_write(blkcg, val);
-		}
-		break;
-	default:
-		BUG();
-	}
-
-	return 0;
+	spin_unlock_irq(ctx->disk->queue->queue_lock);
+	rcu_read_unlock();
+	put_disk(ctx->disk);
 }
+EXPORT_SYMBOL_GPL(blkg_conf_finish);
 
-struct cftype blkio_files[] = {
-	{
-		.name = "weight_device",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,
-				BLKIO_PROP_weight_device),
-		.read_seq_string = blkiocg_file_read,
-		.write_string = blkiocg_file_write,
-		.max_write_len = 256,
-	},
-	{
-		.name = "weight",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,
-				BLKIO_PROP_weight),
-		.read_u64 = blkiocg_file_read_u64,
-		.write_u64 = blkiocg_file_write_u64,
-	},
-	{
-		.name = "time",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,
-				BLKIO_PROP_time),
-		.read_map = blkiocg_file_read_map,
-	},
-	{
-		.name = "sectors",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,
-				BLKIO_PROP_sectors),
-		.read_map = blkiocg_file_read_map,
-	},
-	{
-		.name = "io_service_bytes",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,
-				BLKIO_PROP_io_service_bytes),
-		.read_map = blkiocg_file_read_map,
-	},
-	{
-		.name = "io_serviced",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,
-				BLKIO_PROP_io_serviced),
-		.read_map = blkiocg_file_read_map,
-	},
-	{
-		.name = "io_service_time",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,
-				BLKIO_PROP_io_service_time),
-		.read_map = blkiocg_file_read_map,
-	},
-	{
-		.name = "io_wait_time",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,
-				BLKIO_PROP_io_wait_time),
-		.read_map = blkiocg_file_read_map,
-	},
-	{
-		.name = "io_merged",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,
-				BLKIO_PROP_io_merged),
-		.read_map = blkiocg_file_read_map,
-	},
-	{
-		.name = "io_queued",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,
-				BLKIO_PROP_io_queued),
-		.read_map = blkiocg_file_read_map,
-	},
+struct cftype blkcg_files[] = {
 	{
 		.name = "reset_stats",
-		.write_u64 = blkiocg_reset_stats,
-	},
-#ifdef CONFIG_BLK_DEV_THROTTLING
-	{
-		.name = "throttle.read_bps_device",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_THROTL,
-				BLKIO_THROTL_read_bps_device),
-		.read_seq_string = blkiocg_file_read,
-		.write_string = blkiocg_file_write,
-		.max_write_len = 256,
-	},
-
-	{
-		.name = "throttle.write_bps_device",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_THROTL,
-				BLKIO_THROTL_write_bps_device),
-		.read_seq_string = blkiocg_file_read,
-		.write_string = blkiocg_file_write,
-		.max_write_len = 256,
-	},
-
-	{
-		.name = "throttle.read_iops_device",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_THROTL,
-				BLKIO_THROTL_read_iops_device),
-		.read_seq_string = blkiocg_file_read,
-		.write_string = blkiocg_file_write,
-		.max_write_len = 256,
-	},
-
-	{
-		.name = "throttle.write_iops_device",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_THROTL,
-				BLKIO_THROTL_write_iops_device),
-		.read_seq_string = blkiocg_file_read,
-		.write_string = blkiocg_file_write,
-		.max_write_len = 256,
-	},
-	{
-		.name = "throttle.io_service_bytes",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_THROTL,
-				BLKIO_THROTL_io_service_bytes),
-		.read_map = blkiocg_file_read_map,
-	},
-	{
-		.name = "throttle.io_serviced",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_THROTL,
-				BLKIO_THROTL_io_serviced),
-		.read_map = blkiocg_file_read_map,
-	},
-#endif /* CONFIG_BLK_DEV_THROTTLING */
-
-#ifdef CONFIG_DEBUG_BLK_CGROUP
-	{
-		.name = "avg_queue_size",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,
-				BLKIO_PROP_avg_queue_size),
-		.read_map = blkiocg_file_read_map,
+		.write_u64 = blkcg_reset_stats,
 	},
-	{
-		.name = "group_wait_time",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,
-				BLKIO_PROP_group_wait_time),
-		.read_map = blkiocg_file_read_map,
-	},
-	{
-		.name = "idle_time",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,
-				BLKIO_PROP_idle_time),
-		.read_map = blkiocg_file_read_map,
-	},
-	{
-		.name = "empty_time",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,
-				BLKIO_PROP_empty_time),
-		.read_map = blkiocg_file_read_map,
-	},
-	{
-		.name = "dequeue",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,
-				BLKIO_PROP_dequeue),
-		.read_map = blkiocg_file_read_map,
-	},
-	{
-		.name = "unaccounted_time",
-		.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,
-				BLKIO_PROP_unaccounted_time),
-		.read_map = blkiocg_file_read_map,
-	},
-#endif
 	{ }	/* terminate */
 };
 
-static void blkiocg_destroy(struct cgroup *cgroup)
+/**
+ * blkcg_pre_destroy - cgroup pre_destroy callback
+ * @cgroup: cgroup of interest
+ *
+ * This function is called when @cgroup is about to go away and responsible
+ * for shooting down all blkgs associated with @cgroup.  blkgs should be
+ * removed while holding both q and blkcg locks.  As blkcg lock is nested
+ * inside q lock, this function performs reverse double lock dancing.
+ *
+ * This is the blkcg counterpart of ioc_release_fn().
+ */
+static int blkcg_pre_destroy(struct cgroup *cgroup)
 {
-	struct blkio_cgroup *blkcg = cgroup_to_blkio_cgroup(cgroup);
-	unsigned long flags;
-	struct blkio_group *blkg;
-	void *key;
-	struct blkio_policy_type *blkiop;
-	struct blkio_policy_node *pn, *pntmp;
+	struct blkcg *blkcg = cgroup_to_blkcg(cgroup);
 
-	rcu_read_lock();
-	do {
-		spin_lock_irqsave(&blkcg->lock, flags);
+	spin_lock_irq(&blkcg->lock);
 
-		if (hlist_empty(&blkcg->blkg_list)) {
-			spin_unlock_irqrestore(&blkcg->lock, flags);
-			break;
+	while (!hlist_empty(&blkcg->blkg_list)) {
+		struct blkcg_gq *blkg = hlist_entry(blkcg->blkg_list.first,
+						struct blkcg_gq, blkcg_node);
+		struct request_queue *q = blkg->q;
+
+		if (spin_trylock(q->queue_lock)) {
+			blkg_destroy(blkg);
+			spin_unlock(q->queue_lock);
+		} else {
+			spin_unlock_irq(&blkcg->lock);
+			cpu_relax();
+			spin_lock_irq(&blkcg->lock);
 		}
+	}
 
-		blkg = hlist_entry(blkcg->blkg_list.first, struct blkio_group,
-					blkcg_node);
-		key = rcu_dereference(blkg->key);
-		__blkiocg_del_blkio_group(blkg);
-
-		spin_unlock_irqrestore(&blkcg->lock, flags);
-
-		/*
-		 * This blkio_group is being unlinked as associated cgroup is
-		 * going away. Let all the IO controlling policies know about
-		 * this event.
-		 */
-		spin_lock(&blkio_list_lock);
-		list_for_each_entry(blkiop, &blkio_list, list) {
-			if (blkiop->plid != blkg->plid)
-				continue;
-			blkiop->ops.blkio_unlink_group_fn(key, blkg);
-		}
-		spin_unlock(&blkio_list_lock);
-	} while (1);
+	spin_unlock_irq(&blkcg->lock);
+	return 0;
+}
 
-	list_for_each_entry_safe(pn, pntmp, &blkcg->policy_list, node) {
-		blkio_policy_delete_node(pn);
-		kfree(pn);
-	}
+static void blkcg_destroy(struct cgroup *cgroup)
+{
+	struct blkcg *blkcg = cgroup_to_blkcg(cgroup);
 
-	free_css_id(&blkio_subsys, &blkcg->css);
-	rcu_read_unlock();
-	if (blkcg != &blkio_root_cgroup)
+	if (blkcg != &blkcg_root)
 		kfree(blkcg);
 }
 
-static struct cgroup_subsys_state *blkiocg_create(struct cgroup *cgroup)
+static struct cgroup_subsys_state *blkcg_create(struct cgroup *cgroup)
 {
-	struct blkio_cgroup *blkcg;
+	static atomic64_t id_seq = ATOMIC64_INIT(0);
+	struct blkcg *blkcg;
 	struct cgroup *parent = cgroup->parent;
 
 	if (!parent) {
-		blkcg = &blkio_root_cgroup;
+		blkcg = &blkcg_root;
 		goto done;
 	}
 
@@ -1582,22 +624,68 @@ static struct cgroup_subsys_state *blkiocg_create(struct cgroup *cgroup)
 	if (!blkcg)
 		return ERR_PTR(-ENOMEM);
 
-	blkcg->weight = BLKIO_WEIGHT_DEFAULT;
+	blkcg->cfq_weight = CFQ_WEIGHT_DEFAULT;
+	blkcg->id = atomic64_inc_return(&id_seq); /* root is 0, start from 1 */
 done:
 	spin_lock_init(&blkcg->lock);
+	INIT_RADIX_TREE(&blkcg->blkg_tree, GFP_ATOMIC);
 	INIT_HLIST_HEAD(&blkcg->blkg_list);
 
-	INIT_LIST_HEAD(&blkcg->policy_list);
 	return &blkcg->css;
 }
 
+/**
+ * blkcg_init_queue - initialize blkcg part of request queue
+ * @q: request_queue to initialize
+ *
+ * Called from blk_alloc_queue_node(). Responsible for initializing blkcg
+ * part of new request_queue @q.
+ *
+ * RETURNS:
+ * 0 on success, -errno on failure.
+ */
+int blkcg_init_queue(struct request_queue *q)
+{
+	might_sleep();
+
+	return blk_throtl_init(q);
+}
+
+/**
+ * blkcg_drain_queue - drain blkcg part of request_queue
+ * @q: request_queue to drain
+ *
+ * Called from blk_drain_queue().  Responsible for draining blkcg part.
+ */
+void blkcg_drain_queue(struct request_queue *q)
+{
+	lockdep_assert_held(q->queue_lock);
+
+	blk_throtl_drain(q);
+}
+
+/**
+ * blkcg_exit_queue - exit and release blkcg part of request_queue
+ * @q: request_queue being released
+ *
+ * Called from blk_release_queue().  Responsible for exiting blkcg part.
+ */
+void blkcg_exit_queue(struct request_queue *q)
+{
+	spin_lock_irq(q->queue_lock);
+	blkg_destroy_all(q);
+	spin_unlock_irq(q->queue_lock);
+
+	blk_throtl_exit(q);
+}
+
 /*
  * We cannot support shared io contexts, as we have no mean to support
  * two tasks with the same ioc in two different groups without major rework
  * of the main cic data structures.  For now we allow a task to change
  * its cgroup only if it's the only owner of its ioc.
  */
-static int blkiocg_can_attach(struct cgroup *cgrp, struct cgroup_taskset *tset)
+static int blkcg_can_attach(struct cgroup *cgrp, struct cgroup_taskset *tset)
 {
 	struct task_struct *task;
 	struct io_context *ioc;
@@ -1616,63 +704,213 @@ static int blkiocg_can_attach(struct cgroup *cgrp, struct cgroup_taskset *tset)
 	return ret;
 }
 
-static void blkiocg_attach(struct cgroup *cgrp, struct cgroup_taskset *tset)
-{
-	struct task_struct *task;
-	struct io_context *ioc;
-
-	cgroup_taskset_for_each(task, cgrp, tset) {
-		/* we don't lose anything even if ioc allocation fails */
-		ioc = get_task_io_context(task, GFP_ATOMIC, NUMA_NO_NODE);
-		if (ioc) {
-			ioc_cgroup_changed(ioc);
-			put_io_context(ioc);
-		}
-	}
-}
-
 struct cgroup_subsys blkio_subsys = {
 	.name = "blkio",
-	.create = blkiocg_create,
-	.can_attach = blkiocg_can_attach,
-	.attach = blkiocg_attach,
-	.destroy = blkiocg_destroy,
-#ifdef CONFIG_BLK_CGROUP
-	/* note: blkio_subsys_id is otherwise defined in blk-cgroup.h */
+	.create = blkcg_create,
+	.can_attach = blkcg_can_attach,
+	.pre_destroy = blkcg_pre_destroy,
+	.destroy = blkcg_destroy,
 	.subsys_id = blkio_subsys_id,
-#endif
-	.base_cftypes = blkio_files,
-	.use_id = 1,
+	.base_cftypes = blkcg_files,
 	.module = THIS_MODULE,
 };
 EXPORT_SYMBOL_GPL(blkio_subsys);
 
-void blkio_policy_register(struct blkio_policy_type *blkiop)
+/**
+ * blkcg_activate_policy - activate a blkcg policy on a request_queue
+ * @q: request_queue of interest
+ * @pol: blkcg policy to activate
+ *
+ * Activate @pol on @q.  Requires %GFP_KERNEL context.  @q goes through
+ * bypass mode to populate its blkgs with policy_data for @pol.
+ *
+ * Activation happens with @q bypassed, so nobody would be accessing blkgs
+ * from IO path.  Update of each blkg is protected by both queue and blkcg
+ * locks so that holding either lock and testing blkcg_policy_enabled() is
+ * always enough for dereferencing policy data.
+ *
+ * The caller is responsible for synchronizing [de]activations and policy
+ * [un]registerations.  Returns 0 on success, -errno on failure.
+ */
+int blkcg_activate_policy(struct request_queue *q,
+			  const struct blkcg_policy *pol)
 {
-	spin_lock(&blkio_list_lock);
-	list_add_tail(&blkiop->list, &blkio_list);
-	spin_unlock(&blkio_list_lock);
+	LIST_HEAD(pds);
+	struct blkcg_gq *blkg;
+	struct blkg_policy_data *pd, *n;
+	int cnt = 0, ret;
+
+	if (blkcg_policy_enabled(q, pol))
+		return 0;
+
+	blk_queue_bypass_start(q);
+
+	/* make sure the root blkg exists and count the existing blkgs */
+	spin_lock_irq(q->queue_lock);
+
+	rcu_read_lock();
+	blkg = __blkg_lookup_create(&blkcg_root, q);
+	rcu_read_unlock();
+
+	if (IS_ERR(blkg)) {
+		ret = PTR_ERR(blkg);
+		goto out_unlock;
+	}
+	q->root_blkg = blkg;
+
+	list_for_each_entry(blkg, &q->blkg_list, q_node)
+		cnt++;
+
+	spin_unlock_irq(q->queue_lock);
+
+	/* allocate policy_data for all existing blkgs */
+	while (cnt--) {
+		pd = kzalloc_node(pol->pd_size, GFP_KERNEL, q->node);
+		if (!pd) {
+			ret = -ENOMEM;
+			goto out_free;
+		}
+		list_add_tail(&pd->alloc_node, &pds);
+	}
+
+	/*
+	 * Install the allocated pds.  With @q bypassing, no new blkg
+	 * should have been created while the queue lock was dropped.
+	 */
+	spin_lock_irq(q->queue_lock);
+
+	list_for_each_entry(blkg, &q->blkg_list, q_node) {
+		if (WARN_ON(list_empty(&pds))) {
+			/* umm... this shouldn't happen, just abort */
+			ret = -ENOMEM;
+			goto out_unlock;
+		}
+		pd = list_first_entry(&pds, struct blkg_policy_data, alloc_node);
+		list_del_init(&pd->alloc_node);
+
+		/* grab blkcg lock too while installing @pd on @blkg */
+		spin_lock(&blkg->blkcg->lock);
+
+		blkg->pd[pol->plid] = pd;
+		pd->blkg = blkg;
+		pol->pd_init_fn(blkg);
+
+		spin_unlock(&blkg->blkcg->lock);
+	}
+
+	__set_bit(pol->plid, q->blkcg_pols);
+	ret = 0;
+out_unlock:
+	spin_unlock_irq(q->queue_lock);
+out_free:
+	blk_queue_bypass_end(q);
+	list_for_each_entry_safe(pd, n, &pds, alloc_node)
+		kfree(pd);
+	return ret;
 }
-EXPORT_SYMBOL_GPL(blkio_policy_register);
+EXPORT_SYMBOL_GPL(blkcg_activate_policy);
 
-void blkio_policy_unregister(struct blkio_policy_type *blkiop)
+/**
+ * blkcg_deactivate_policy - deactivate a blkcg policy on a request_queue
+ * @q: request_queue of interest
+ * @pol: blkcg policy to deactivate
+ *
+ * Deactivate @pol on @q.  Follows the same synchronization rules as
+ * blkcg_activate_policy().
+ */
+void blkcg_deactivate_policy(struct request_queue *q,
+			     const struct blkcg_policy *pol)
 {
-	spin_lock(&blkio_list_lock);
-	list_del_init(&blkiop->list);
-	spin_unlock(&blkio_list_lock);
+	struct blkcg_gq *blkg;
+
+	if (!blkcg_policy_enabled(q, pol))
+		return;
+
+	blk_queue_bypass_start(q);
+	spin_lock_irq(q->queue_lock);
+
+	__clear_bit(pol->plid, q->blkcg_pols);
+
+	/* if no policy is left, no need for blkgs - shoot them down */
+	if (bitmap_empty(q->blkcg_pols, BLKCG_MAX_POLS))
+		blkg_destroy_all(q);
+
+	list_for_each_entry(blkg, &q->blkg_list, q_node) {
+		/* grab blkcg lock too while removing @pd from @blkg */
+		spin_lock(&blkg->blkcg->lock);
+
+		if (pol->pd_exit_fn)
+			pol->pd_exit_fn(blkg);
+
+		kfree(blkg->pd[pol->plid]);
+		blkg->pd[pol->plid] = NULL;
+
+		spin_unlock(&blkg->blkcg->lock);
+	}
+
+	spin_unlock_irq(q->queue_lock);
+	blk_queue_bypass_end(q);
 }
-EXPORT_SYMBOL_GPL(blkio_policy_unregister);
+EXPORT_SYMBOL_GPL(blkcg_deactivate_policy);
 
-static int __init init_cgroup_blkio(void)
+/**
+ * blkcg_policy_register - register a blkcg policy
+ * @pol: blkcg policy to register
+ *
+ * Register @pol with blkcg core.  Might sleep and @pol may be modified on
+ * successful registration.  Returns 0 on success and -errno on failure.
+ */
+int blkcg_policy_register(struct blkcg_policy *pol)
 {
-	return cgroup_load_subsys(&blkio_subsys);
+	int i, ret;
+
+	if (WARN_ON(pol->pd_size < sizeof(struct blkg_policy_data)))
+		return -EINVAL;
+
+	mutex_lock(&blkcg_pol_mutex);
+
+	/* find an empty slot */
+	ret = -ENOSPC;
+	for (i = 0; i < BLKCG_MAX_POLS; i++)
+		if (!blkcg_policy[i])
+			break;
+	if (i >= BLKCG_MAX_POLS)
+		goto out_unlock;
+
+	/* register and update blkgs */
+	pol->plid = i;
+	blkcg_policy[i] = pol;
+
+	/* everything is in place, add intf files for the new policy */
+	if (pol->cftypes)
+		WARN_ON(cgroup_add_cftypes(&blkio_subsys, pol->cftypes));
+	ret = 0;
+out_unlock:
+	mutex_unlock(&blkcg_pol_mutex);
+	return ret;
 }
+EXPORT_SYMBOL_GPL(blkcg_policy_register);
 
-static void __exit exit_cgroup_blkio(void)
+/**
+ * blkcg_policy_unregister - unregister a blkcg policy
+ * @pol: blkcg policy to unregister
+ *
+ * Undo blkcg_policy_register(@pol).  Might sleep.
+ */
+void blkcg_policy_unregister(struct blkcg_policy *pol)
 {
-	cgroup_unload_subsys(&blkio_subsys);
-}
+	mutex_lock(&blkcg_pol_mutex);
 
-module_init(init_cgroup_blkio);
-module_exit(exit_cgroup_blkio);
-MODULE_LICENSE("GPL");
+	if (WARN_ON(blkcg_policy[pol->plid] != pol))
+		goto out_unlock;
+
+	/* kill the intf files first */
+	if (pol->cftypes)
+		cgroup_rm_cftypes(&blkio_subsys, pol->cftypes);
+
+	/* unregister and update blkgs */
+	blkcg_policy[pol->plid] = NULL;
+out_unlock:
+	mutex_unlock(&blkcg_pol_mutex);
+}
+EXPORT_SYMBOL_GPL(blkcg_policy_unregister);
diff --git a/block/blk-cgroup.h b/block/blk-cgroup.h
index 6f3ace7e792f..8ac457ce7783 100644
--- a/block/blk-cgroup.h
+++ b/block/blk-cgroup.h
@@ -15,350 +15,371 @@
 
 #include <linux/cgroup.h>
 #include <linux/u64_stats_sync.h>
-
-enum blkio_policy_id {
-	BLKIO_POLICY_PROP = 0,		/* Proportional Bandwidth division */
-	BLKIO_POLICY_THROTL,		/* Throttling */
-};
+#include <linux/seq_file.h>
+#include <linux/radix-tree.h>
 
 /* Max limits for throttle policy */
 #define THROTL_IOPS_MAX		UINT_MAX
 
-#if defined(CONFIG_BLK_CGROUP) || defined(CONFIG_BLK_CGROUP_MODULE)
-
-#ifndef CONFIG_BLK_CGROUP
-/* When blk-cgroup is a module, its subsys_id isn't a compile-time constant */
-extern struct cgroup_subsys blkio_subsys;
-#define blkio_subsys_id blkio_subsys.subsys_id
-#endif
-
-enum stat_type {
-	/* Total time spent (in ns) between request dispatch to the driver and
-	 * request completion for IOs doen by this cgroup. This may not be
-	 * accurate when NCQ is turned on. */
-	BLKIO_STAT_SERVICE_TIME = 0,
-	/* Total time spent waiting in scheduler queue in ns */
-	BLKIO_STAT_WAIT_TIME,
-	/* Number of IOs queued up */
-	BLKIO_STAT_QUEUED,
-	/* All the single valued stats go below this */
-	BLKIO_STAT_TIME,
-#ifdef CONFIG_DEBUG_BLK_CGROUP
-	/* Time not charged to this cgroup */
-	BLKIO_STAT_UNACCOUNTED_TIME,
-	BLKIO_STAT_AVG_QUEUE_SIZE,
-	BLKIO_STAT_IDLE_TIME,
-	BLKIO_STAT_EMPTY_TIME,
-	BLKIO_STAT_GROUP_WAIT_TIME,
-	BLKIO_STAT_DEQUEUE
-#endif
-};
+/* CFQ specific, out here for blkcg->cfq_weight */
+#define CFQ_WEIGHT_MIN		10
+#define CFQ_WEIGHT_MAX		1000
+#define CFQ_WEIGHT_DEFAULT	500
 
-/* Per cpu stats */
-enum stat_type_cpu {
-	BLKIO_STAT_CPU_SECTORS,
-	/* Total bytes transferred */
-	BLKIO_STAT_CPU_SERVICE_BYTES,
-	/* Total IOs serviced, post merge */
-	BLKIO_STAT_CPU_SERVICED,
-	/* Number of IOs merged */
-	BLKIO_STAT_CPU_MERGED,
-	BLKIO_STAT_CPU_NR
-};
+#ifdef CONFIG_BLK_CGROUP
 
-enum stat_sub_type {
-	BLKIO_STAT_READ = 0,
-	BLKIO_STAT_WRITE,
-	BLKIO_STAT_SYNC,
-	BLKIO_STAT_ASYNC,
-	BLKIO_STAT_TOTAL
-};
+enum blkg_rwstat_type {
+	BLKG_RWSTAT_READ,
+	BLKG_RWSTAT_WRITE,
+	BLKG_RWSTAT_SYNC,
+	BLKG_RWSTAT_ASYNC,
 
-/* blkg state flags */
-enum blkg_state_flags {
-	BLKG_waiting = 0,
-	BLKG_idling,
-	BLKG_empty,
+	BLKG_RWSTAT_NR,
+	BLKG_RWSTAT_TOTAL = BLKG_RWSTAT_NR,
 };
 
-/* cgroup files owned by proportional weight policy */
-enum blkcg_file_name_prop {
-	BLKIO_PROP_weight = 1,
-	BLKIO_PROP_weight_device,
-	BLKIO_PROP_io_service_bytes,
-	BLKIO_PROP_io_serviced,
-	BLKIO_PROP_time,
-	BLKIO_PROP_sectors,
-	BLKIO_PROP_unaccounted_time,
-	BLKIO_PROP_io_service_time,
-	BLKIO_PROP_io_wait_time,
-	BLKIO_PROP_io_merged,
-	BLKIO_PROP_io_queued,
-	BLKIO_PROP_avg_queue_size,
-	BLKIO_PROP_group_wait_time,
-	BLKIO_PROP_idle_time,
-	BLKIO_PROP_empty_time,
-	BLKIO_PROP_dequeue,
-};
+struct blkcg_gq;
 
-/* cgroup files owned by throttle policy */
-enum blkcg_file_name_throtl {
-	BLKIO_THROTL_read_bps_device,
-	BLKIO_THROTL_write_bps_device,
-	BLKIO_THROTL_read_iops_device,
-	BLKIO_THROTL_write_iops_device,
-	BLKIO_THROTL_io_service_bytes,
-	BLKIO_THROTL_io_serviced,
-};
+struct blkcg {
+	struct cgroup_subsys_state	css;
+	spinlock_t			lock;
 
-struct blkio_cgroup {
-	struct cgroup_subsys_state css;
-	unsigned int weight;
-	spinlock_t lock;
-	struct hlist_head blkg_list;
-	struct list_head policy_list; /* list of blkio_policy_node */
-};
+	struct radix_tree_root		blkg_tree;
+	struct blkcg_gq			*blkg_hint;
+	struct hlist_head		blkg_list;
+
+	/* for policies to test whether associated blkcg has changed */
+	uint64_t			id;
 
-struct blkio_group_stats {
-	/* total disk time and nr sectors dispatched by this group */
-	uint64_t time;
-	uint64_t stat_arr[BLKIO_STAT_QUEUED + 1][BLKIO_STAT_TOTAL];
-#ifdef CONFIG_DEBUG_BLK_CGROUP
-	/* Time not charged to this cgroup */
-	uint64_t unaccounted_time;
-
-	/* Sum of number of IOs queued across all samples */
-	uint64_t avg_queue_size_sum;
-	/* Count of samples taken for average */
-	uint64_t avg_queue_size_samples;
-	/* How many times this group has been removed from service tree */
-	unsigned long dequeue;
-
-	/* Total time spent waiting for it to be assigned a timeslice. */
-	uint64_t group_wait_time;
-	uint64_t start_group_wait_time;
-
-	/* Time spent idling for this blkio_group */
-	uint64_t idle_time;
-	uint64_t start_idle_time;
-	/*
-	 * Total time when we have requests queued and do not contain the
-	 * current active queue.
-	 */
-	uint64_t empty_time;
-	uint64_t start_empty_time;
-	uint16_t flags;
-#endif
+	/* TODO: per-policy storage in blkcg */
+	unsigned int			cfq_weight;	/* belongs to cfq */
 };
 
-/* Per cpu blkio group stats */
-struct blkio_group_stats_cpu {
-	uint64_t sectors;
-	uint64_t stat_arr_cpu[BLKIO_STAT_CPU_NR][BLKIO_STAT_TOTAL];
-	struct u64_stats_sync syncp;
+struct blkg_stat {
+	struct u64_stats_sync		syncp;
+	uint64_t			cnt;
 };
 
-struct blkio_group {
-	/* An rcu protected unique identifier for the group */
-	void *key;
-	struct hlist_node blkcg_node;
-	unsigned short blkcg_id;
-	/* Store cgroup path */
-	char path[128];
-	/* The device MKDEV(major, minor), this group has been created for */
-	dev_t dev;
-	/* policy which owns this blk group */
-	enum blkio_policy_id plid;
-
-	/* Need to serialize the stats in the case of reset/update */
-	spinlock_t stats_lock;
-	struct blkio_group_stats stats;
-	/* Per cpu stats pointer */
-	struct blkio_group_stats_cpu __percpu *stats_cpu;
+struct blkg_rwstat {
+	struct u64_stats_sync		syncp;
+	uint64_t			cnt[BLKG_RWSTAT_NR];
 };
 
-struct blkio_policy_node {
-	struct list_head node;
-	dev_t dev;
-	/* This node belongs to max bw policy or porportional weight policy */
-	enum blkio_policy_id plid;
-	/* cgroup file to which this rule belongs to */
-	int fileid;
-
-	union {
-		unsigned int weight;
-		/*
-		 * Rate read/write in terms of bytes per second
-		 * Whether this rate represents read or write is determined
-		 * by file type "fileid".
-		 */
-		u64 bps;
-		unsigned int iops;
-	} val;
+/*
+ * A blkcg_gq (blkg) is association between a block cgroup (blkcg) and a
+ * request_queue (q).  This is used by blkcg policies which need to track
+ * information per blkcg - q pair.
+ *
+ * There can be multiple active blkcg policies and each has its private
+ * data on each blkg, the size of which is determined by
+ * blkcg_policy->pd_size.  blkcg core allocates and frees such areas
+ * together with blkg and invokes pd_init/exit_fn() methods.
+ *
+ * Such private data must embed struct blkg_policy_data (pd) at the
+ * beginning and pd_size can't be smaller than pd.
+ */
+struct blkg_policy_data {
+	/* the blkg this per-policy data belongs to */
+	struct blkcg_gq			*blkg;
+
+	/* used during policy activation */
+	struct list_head		alloc_node;
 };
 
-extern unsigned int blkcg_get_weight(struct blkio_cgroup *blkcg,
-				     dev_t dev);
-extern uint64_t blkcg_get_read_bps(struct blkio_cgroup *blkcg,
-				     dev_t dev);
-extern uint64_t blkcg_get_write_bps(struct blkio_cgroup *blkcg,
-				     dev_t dev);
-extern unsigned int blkcg_get_read_iops(struct blkio_cgroup *blkcg,
-				     dev_t dev);
-extern unsigned int blkcg_get_write_iops(struct blkio_cgroup *blkcg,
-				     dev_t dev);
-
-typedef void (blkio_unlink_group_fn) (void *key, struct blkio_group *blkg);
-
-typedef void (blkio_update_group_weight_fn) (void *key,
-			struct blkio_group *blkg, unsigned int weight);
-typedef void (blkio_update_group_read_bps_fn) (void * key,
-			struct blkio_group *blkg, u64 read_bps);
-typedef void (blkio_update_group_write_bps_fn) (void *key,
-			struct blkio_group *blkg, u64 write_bps);
-typedef void (blkio_update_group_read_iops_fn) (void *key,
-			struct blkio_group *blkg, unsigned int read_iops);
-typedef void (blkio_update_group_write_iops_fn) (void *key,
-			struct blkio_group *blkg, unsigned int write_iops);
-
-struct blkio_policy_ops {
-	blkio_unlink_group_fn *blkio_unlink_group_fn;
-	blkio_update_group_weight_fn *blkio_update_group_weight_fn;
-	blkio_update_group_read_bps_fn *blkio_update_group_read_bps_fn;
-	blkio_update_group_write_bps_fn *blkio_update_group_write_bps_fn;
-	blkio_update_group_read_iops_fn *blkio_update_group_read_iops_fn;
-	blkio_update_group_write_iops_fn *blkio_update_group_write_iops_fn;
+/* association between a blk cgroup and a request queue */
+struct blkcg_gq {
+	/* Pointer to the associated request_queue */
+	struct request_queue		*q;
+	struct list_head		q_node;
+	struct hlist_node		blkcg_node;
+	struct blkcg			*blkcg;
+	/* reference count */
+	int				refcnt;
+
+	struct blkg_policy_data		*pd[BLKCG_MAX_POLS];
+
+	struct rcu_head			rcu_head;
 };
 
-struct blkio_policy_type {
-	struct list_head list;
-	struct blkio_policy_ops ops;
-	enum blkio_policy_id plid;
+typedef void (blkcg_pol_init_pd_fn)(struct blkcg_gq *blkg);
+typedef void (blkcg_pol_exit_pd_fn)(struct blkcg_gq *blkg);
+typedef void (blkcg_pol_reset_pd_stats_fn)(struct blkcg_gq *blkg);
+
+struct blkcg_policy {
+	int				plid;
+	/* policy specific private data size */
+	size_t				pd_size;
+	/* cgroup files for the policy */
+	struct cftype			*cftypes;
+
+	/* operations */
+	blkcg_pol_init_pd_fn		*pd_init_fn;
+	blkcg_pol_exit_pd_fn		*pd_exit_fn;
+	blkcg_pol_reset_pd_stats_fn	*pd_reset_stats_fn;
 };
 
+extern struct blkcg blkcg_root;
+
+struct blkcg *cgroup_to_blkcg(struct cgroup *cgroup);
+struct blkcg *bio_blkcg(struct bio *bio);
+struct blkcg_gq *blkg_lookup(struct blkcg *blkcg, struct request_queue *q);
+struct blkcg_gq *blkg_lookup_create(struct blkcg *blkcg,
+				    struct request_queue *q);
+int blkcg_init_queue(struct request_queue *q);
+void blkcg_drain_queue(struct request_queue *q);
+void blkcg_exit_queue(struct request_queue *q);
+
 /* Blkio controller policy registration */
-extern void blkio_policy_register(struct blkio_policy_type *);
-extern void blkio_policy_unregister(struct blkio_policy_type *);
+int blkcg_policy_register(struct blkcg_policy *pol);
+void blkcg_policy_unregister(struct blkcg_policy *pol);
+int blkcg_activate_policy(struct request_queue *q,
+			  const struct blkcg_policy *pol);
+void blkcg_deactivate_policy(struct request_queue *q,
+			     const struct blkcg_policy *pol);
+
+void blkcg_print_blkgs(struct seq_file *sf, struct blkcg *blkcg,
+		       u64 (*prfill)(struct seq_file *,
+				     struct blkg_policy_data *, int),
+		       const struct blkcg_policy *pol, int data,
+		       bool show_total);
+u64 __blkg_prfill_u64(struct seq_file *sf, struct blkg_policy_data *pd, u64 v);
+u64 __blkg_prfill_rwstat(struct seq_file *sf, struct blkg_policy_data *pd,
+			 const struct blkg_rwstat *rwstat);
+u64 blkg_prfill_stat(struct seq_file *sf, struct blkg_policy_data *pd, int off);
+u64 blkg_prfill_rwstat(struct seq_file *sf, struct blkg_policy_data *pd,
+		       int off);
+
+struct blkg_conf_ctx {
+	struct gendisk			*disk;
+	struct blkcg_gq			*blkg;
+	u64				v;
+};
+
+int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
+		   const char *input, struct blkg_conf_ctx *ctx);
+void blkg_conf_finish(struct blkg_conf_ctx *ctx);
+
+
+/**
+ * blkg_to_pdata - get policy private data
+ * @blkg: blkg of interest
+ * @pol: policy of interest
+ *
+ * Return pointer to private data associated with the @blkg-@pol pair.
+ */
+static inline struct blkg_policy_data *blkg_to_pd(struct blkcg_gq *blkg,
+						  struct blkcg_policy *pol)
+{
+	return blkg ? blkg->pd[pol->plid] : NULL;
+}
+
+/**
+ * pdata_to_blkg - get blkg associated with policy private data
+ * @pd: policy private data of interest
+ *
+ * @pd is policy private data.  Determine the blkg it's associated with.
+ */
+static inline struct blkcg_gq *pd_to_blkg(struct blkg_policy_data *pd)
+{
+	return pd ? pd->blkg : NULL;
+}
+
+/**
+ * blkg_path - format cgroup path of blkg
+ * @blkg: blkg of interest
+ * @buf: target buffer
+ * @buflen: target buffer length
+ *
+ * Format the path of the cgroup of @blkg into @buf.
+ */
+static inline int blkg_path(struct blkcg_gq *blkg, char *buf, int buflen)
+{
+	int ret;
+
+	rcu_read_lock();
+	ret = cgroup_path(blkg->blkcg->css.cgroup, buf, buflen);
+	rcu_read_unlock();
+	if (ret)
+		strncpy(buf, "<unavailable>", buflen);
+	return ret;
+}
 
-static inline char *blkg_path(struct blkio_group *blkg)
+/**
+ * blkg_get - get a blkg reference
+ * @blkg: blkg to get
+ *
+ * The caller should be holding queue_lock and an existing reference.
+ */
+static inline void blkg_get(struct blkcg_gq *blkg)
 {
-	return blkg->path;
+	lockdep_assert_held(blkg->q->queue_lock);
+	WARN_ON_ONCE(!blkg->refcnt);
+	blkg->refcnt++;
 }
 
-#else
+void __blkg_release(struct blkcg_gq *blkg);
 
-struct blkio_group {
+/**
+ * blkg_put - put a blkg reference
+ * @blkg: blkg to put
+ *
+ * The caller should be holding queue_lock.
+ */
+static inline void blkg_put(struct blkcg_gq *blkg)
+{
+	lockdep_assert_held(blkg->q->queue_lock);
+	WARN_ON_ONCE(blkg->refcnt <= 0);
+	if (!--blkg->refcnt)
+		__blkg_release(blkg);
+}
+
+/**
+ * blkg_stat_add - add a value to a blkg_stat
+ * @stat: target blkg_stat
+ * @val: value to add
+ *
+ * Add @val to @stat.  The caller is responsible for synchronizing calls to
+ * this function.
+ */
+static inline void blkg_stat_add(struct blkg_stat *stat, uint64_t val)
+{
+	u64_stats_update_begin(&stat->syncp);
+	stat->cnt += val;
+	u64_stats_update_end(&stat->syncp);
+}
+
+/**
+ * blkg_stat_read - read the current value of a blkg_stat
+ * @stat: blkg_stat to read
+ *
+ * Read the current value of @stat.  This function can be called without
+ * synchroniztion and takes care of u64 atomicity.
+ */
+static inline uint64_t blkg_stat_read(struct blkg_stat *stat)
+{
+	unsigned int start;
+	uint64_t v;
+
+	do {
+		start = u64_stats_fetch_begin(&stat->syncp);
+		v = stat->cnt;
+	} while (u64_stats_fetch_retry(&stat->syncp, start));
+
+	return v;
+}
+
+/**
+ * blkg_stat_reset - reset a blkg_stat
+ * @stat: blkg_stat to reset
+ */
+static inline void blkg_stat_reset(struct blkg_stat *stat)
+{
+	stat->cnt = 0;
+}
+
+/**
+ * blkg_rwstat_add - add a value to a blkg_rwstat
+ * @rwstat: target blkg_rwstat
+ * @rw: mask of REQ_{WRITE|SYNC}
+ * @val: value to add
+ *
+ * Add @val to @rwstat.  The counters are chosen according to @rw.  The
+ * caller is responsible for synchronizing calls to this function.
+ */
+static inline void blkg_rwstat_add(struct blkg_rwstat *rwstat,
+				   int rw, uint64_t val)
+{
+	u64_stats_update_begin(&rwstat->syncp);
+
+	if (rw & REQ_WRITE)
+		rwstat->cnt[BLKG_RWSTAT_WRITE] += val;
+	else
+		rwstat->cnt[BLKG_RWSTAT_READ] += val;
+	if (rw & REQ_SYNC)
+		rwstat->cnt[BLKG_RWSTAT_SYNC] += val;
+	else
+		rwstat->cnt[BLKG_RWSTAT_ASYNC] += val;
+
+	u64_stats_update_end(&rwstat->syncp);
+}
+
+/**
+ * blkg_rwstat_read - read the current values of a blkg_rwstat
+ * @rwstat: blkg_rwstat to read
+ *
+ * Read the current snapshot of @rwstat and return it as the return value.
+ * This function can be called without synchronization and takes care of
+ * u64 atomicity.
+ */
+static inline struct blkg_rwstat blkg_rwstat_read(struct blkg_rwstat *rwstat)
+{
+	unsigned int start;
+	struct blkg_rwstat tmp;
+
+	do {
+		start = u64_stats_fetch_begin(&rwstat->syncp);
+		tmp = *rwstat;
+	} while (u64_stats_fetch_retry(&rwstat->syncp, start));
+
+	return tmp;
+}
+
+/**
+ * blkg_rwstat_sum - read the total count of a blkg_rwstat
+ * @rwstat: blkg_rwstat to read
+ *
+ * Return the total count of @rwstat regardless of the IO direction.  This
+ * function can be called without synchronization and takes care of u64
+ * atomicity.
+ */
+static inline uint64_t blkg_rwstat_sum(struct blkg_rwstat *rwstat)
+{
+	struct blkg_rwstat tmp = blkg_rwstat_read(rwstat);
+
+	return tmp.cnt[BLKG_RWSTAT_READ] + tmp.cnt[BLKG_RWSTAT_WRITE];
+}
+
+/**
+ * blkg_rwstat_reset - reset a blkg_rwstat
+ * @rwstat: blkg_rwstat to reset
+ */
+static inline void blkg_rwstat_reset(struct blkg_rwstat *rwstat)
+{
+	memset(rwstat->cnt, 0, sizeof(rwstat->cnt));
+}
+
+#else	/* CONFIG_BLK_CGROUP */
+
+struct cgroup;
+
+struct blkg_policy_data {
 };
 
-struct blkio_policy_type {
+struct blkcg_gq {
 };
 
-static inline void blkio_policy_register(struct blkio_policy_type *blkiop) { }
-static inline void blkio_policy_unregister(struct blkio_policy_type *blkiop) { }
-
-static inline char *blkg_path(struct blkio_group *blkg) { return NULL; }
-
-#endif
-
-#define BLKIO_WEIGHT_MIN	10
-#define BLKIO_WEIGHT_MAX	1000
-#define BLKIO_WEIGHT_DEFAULT	500
-
-#ifdef CONFIG_DEBUG_BLK_CGROUP
-void blkiocg_update_avg_queue_size_stats(struct blkio_group *blkg);
-void blkiocg_update_dequeue_stats(struct blkio_group *blkg,
-				unsigned long dequeue);
-void blkiocg_update_set_idle_time_stats(struct blkio_group *blkg);
-void blkiocg_update_idle_time_stats(struct blkio_group *blkg);
-void blkiocg_set_start_empty_time(struct blkio_group *blkg);
-
-#define BLKG_FLAG_FNS(name)						\
-static inline void blkio_mark_blkg_##name(				\
-		struct blkio_group_stats *stats)			\
-{									\
-	stats->flags |= (1 << BLKG_##name);				\
-}									\
-static inline void blkio_clear_blkg_##name(				\
-		struct blkio_group_stats *stats)			\
-{									\
-	stats->flags &= ~(1 << BLKG_##name);				\
-}									\
-static inline int blkio_blkg_##name(struct blkio_group_stats *stats)	\
-{									\
-	return (stats->flags & (1 << BLKG_##name)) != 0;		\
-}									\
-
-BLKG_FLAG_FNS(waiting)
-BLKG_FLAG_FNS(idling)
-BLKG_FLAG_FNS(empty)
-#undef BLKG_FLAG_FNS
-#else
-static inline void blkiocg_update_avg_queue_size_stats(
-						struct blkio_group *blkg) {}
-static inline void blkiocg_update_dequeue_stats(struct blkio_group *blkg,
-						unsigned long dequeue) {}
-static inline void blkiocg_update_set_idle_time_stats(struct blkio_group *blkg)
-{}
-static inline void blkiocg_update_idle_time_stats(struct blkio_group *blkg) {}
-static inline void blkiocg_set_start_empty_time(struct blkio_group *blkg) {}
-#endif
-
-#if defined(CONFIG_BLK_CGROUP) || defined(CONFIG_BLK_CGROUP_MODULE)
-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, void *key, 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,
-						void *key);
-void blkiocg_update_timeslice_used(struct blkio_group *blkg,
-					unsigned long time,
-					unsigned long unaccounted_time);
-void blkiocg_update_dispatch_stats(struct blkio_group *blkg, uint64_t bytes,
-						bool direction, bool sync);
-void blkiocg_update_completion_stats(struct blkio_group *blkg,
-	uint64_t start_time, uint64_t io_start_time, bool direction, bool sync);
-void blkiocg_update_io_merged_stats(struct blkio_group *blkg, bool direction,
-					bool sync);
-void blkiocg_update_io_add_stats(struct blkio_group *blkg,
-		struct blkio_group *curr_blkg, bool direction, bool sync);
-void blkiocg_update_io_remove_stats(struct blkio_group *blkg,
-					bool direction, bool sync);
-#else
-struct cgroup;
-static inline struct blkio_cgroup *
-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 void blkiocg_update_timeslice_used(struct blkio_group *blkg,
-						unsigned long time,
-						unsigned long unaccounted_time)
-{}
-static inline void blkiocg_update_dispatch_stats(struct blkio_group *blkg,
-				uint64_t bytes, bool direction, bool sync) {}
-static inline void blkiocg_update_completion_stats(struct blkio_group *blkg,
-		uint64_t start_time, uint64_t io_start_time, bool direction,
-		bool sync) {}
-static inline void blkiocg_update_io_merged_stats(struct blkio_group *blkg,
-						bool direction, bool sync) {}
-static inline void blkiocg_update_io_add_stats(struct blkio_group *blkg,
-		struct blkio_group *curr_blkg, bool direction, bool sync) {}
-static inline void blkiocg_update_io_remove_stats(struct blkio_group *blkg,
-						bool direction, bool sync) {}
-#endif
-#endif /* _BLK_CGROUP_H */
+struct blkcg_policy {
+};
+
+static inline struct blkcg *cgroup_to_blkcg(struct cgroup *cgroup) { return NULL; }
+static inline struct blkcg *bio_blkcg(struct bio *bio) { return NULL; }
+static inline struct blkcg_gq *blkg_lookup(struct blkcg *blkcg, void *key) { return NULL; }
+static inline int blkcg_init_queue(struct request_queue *q) { return 0; }
+static inline void blkcg_drain_queue(struct request_queue *q) { }
+static inline void blkcg_exit_queue(struct request_queue *q) { }
+static inline int blkcg_policy_register(struct blkcg_policy *pol) { return 0; }
+static inline void blkcg_policy_unregister(struct blkcg_policy *pol) { }
+static inline int blkcg_activate_policy(struct request_queue *q,
+					const struct blkcg_policy *pol) { return 0; }
+static inline void blkcg_deactivate_policy(struct request_queue *q,
+					   const struct blkcg_policy *pol) { }
+
+static inline struct blkg_policy_data *blkg_to_pd(struct blkcg_gq *blkg,
+						  struct blkcg_policy *pol) { return NULL; }
+static inline struct blkcg_gq *pd_to_blkg(struct blkg_policy_data *pd) { return NULL; }
+static inline char *blkg_path(struct blkcg_gq *blkg) { return NULL; }
+static inline void blkg_get(struct blkcg_gq *blkg) { }
+static inline void blkg_put(struct blkcg_gq *blkg) { }
+
+#endif	/* CONFIG_BLK_CGROUP */
+#endif	/* _BLK_CGROUP_H */
diff --git a/block/blk-core.c b/block/blk-core.c
index 1f61b74867e4..3c923a7aeb56 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -29,11 +29,13 @@
 #include <linux/fault-inject.h>
 #include <linux/list_sort.h>
 #include <linux/delay.h>
+#include <linux/ratelimit.h>
 
 #define CREATE_TRACE_POINTS
 #include <trace/events/block.h>
 
 #include "blk.h"
+#include "blk-cgroup.h"
 
 EXPORT_TRACEPOINT_SYMBOL_GPL(block_bio_remap);
 EXPORT_TRACEPOINT_SYMBOL_GPL(block_rq_remap);
@@ -280,7 +282,7 @@ EXPORT_SYMBOL(blk_stop_queue);
  *
  *     This function does not cancel any asynchronous activity arising
  *     out of elevator or throttling code. That would require elevaotor_exit()
- *     and blk_throtl_exit() to be called with queue lock initialized.
+ *     and blkcg_exit_queue() to be called with queue lock initialized.
  *
  */
 void blk_sync_queue(struct request_queue *q)
@@ -365,17 +367,23 @@ void blk_drain_queue(struct request_queue *q, bool drain_all)
 
 		spin_lock_irq(q->queue_lock);
 
-		elv_drain_elevator(q);
-		if (drain_all)
-			blk_throtl_drain(q);
+		/*
+		 * The caller might be trying to drain @q before its
+		 * elevator is initialized.
+		 */
+		if (q->elevator)
+			elv_drain_elevator(q);
+
+		blkcg_drain_queue(q);
 
 		/*
 		 * This function might be called on a queue which failed
-		 * driver init after queue creation.  Some drivers
-		 * (e.g. fd) get unhappy in such cases.  Kick queue iff
-		 * dispatch queue has something on it.
+		 * driver init after queue creation or is not yet fully
+		 * active yet.  Some drivers (e.g. fd and loop) get unhappy
+		 * in such cases.  Kick queue iff dispatch queue has
+		 * something on it and @q has request_fn set.
 		 */
-		if (!list_empty(&q->queue_head))
+		if (!list_empty(&q->queue_head) && q->request_fn)
 			__blk_run_queue(q);
 
 		drain |= q->rq.elvpriv;
@@ -403,6 +411,49 @@ void blk_drain_queue(struct request_queue *q, bool drain_all)
 }
 
 /**
+ * blk_queue_bypass_start - enter queue bypass mode
+ * @q: queue of interest
+ *
+ * In bypass mode, only the dispatch FIFO queue of @q is used.  This
+ * function makes @q enter bypass mode and drains all requests which were
+ * throttled or issued before.  On return, it's guaranteed that no request
+ * is being throttled or has ELVPRIV set and blk_queue_bypass() %true
+ * inside queue or RCU read lock.
+ */
+void blk_queue_bypass_start(struct request_queue *q)
+{
+	bool drain;
+
+	spin_lock_irq(q->queue_lock);
+	drain = !q->bypass_depth++;
+	queue_flag_set(QUEUE_FLAG_BYPASS, q);
+	spin_unlock_irq(q->queue_lock);
+
+	if (drain) {
+		blk_drain_queue(q, false);
+		/* ensure blk_queue_bypass() is %true inside RCU read lock */
+		synchronize_rcu();
+	}
+}
+EXPORT_SYMBOL_GPL(blk_queue_bypass_start);
+
+/**
+ * blk_queue_bypass_end - leave queue bypass mode
+ * @q: queue of interest
+ *
+ * Leave bypass mode and restore the normal queueing behavior.
+ */
+void blk_queue_bypass_end(struct request_queue *q)
+{
+	spin_lock_irq(q->queue_lock);
+	if (!--q->bypass_depth)
+		queue_flag_clear(QUEUE_FLAG_BYPASS, q);
+	WARN_ON_ONCE(q->bypass_depth < 0);
+	spin_unlock_irq(q->queue_lock);
+}
+EXPORT_SYMBOL_GPL(blk_queue_bypass_end);
+
+/**
  * blk_cleanup_queue - shutdown a request queue
  * @q: request queue to shutdown
  *
@@ -418,6 +469,19 @@ void blk_cleanup_queue(struct request_queue *q)
 	queue_flag_set_unlocked(QUEUE_FLAG_DEAD, q);
 
 	spin_lock_irq(lock);
+
+	/*
+	 * Dead queue is permanently in bypass mode till released.  Note
+	 * that, unlike blk_queue_bypass_start(), we aren't performing
+	 * synchronize_rcu() after entering bypass mode to avoid the delay
+	 * as some drivers create and destroy a lot of queues while
+	 * probing.  This is still safe because blk_release_queue() will be
+	 * called only after the queue refcnt drops to zero and nothing,
+	 * RCU or not, would be traversing the queue by then.
+	 */
+	q->bypass_depth++;
+	queue_flag_set(QUEUE_FLAG_BYPASS, q);
+
 	queue_flag_set(QUEUE_FLAG_NOMERGES, q);
 	queue_flag_set(QUEUE_FLAG_NOXMERGES, q);
 	queue_flag_set(QUEUE_FLAG_DEAD, q);
@@ -428,13 +492,8 @@ void blk_cleanup_queue(struct request_queue *q)
 	spin_unlock_irq(lock);
 	mutex_unlock(&q->sysfs_lock);
 
-	/*
-	 * Drain all requests queued before DEAD marking.  The caller might
-	 * be trying to tear down @q before its elevator is initialized, in
-	 * which case we don't want to call into draining.
-	 */
-	if (q->elevator)
-		blk_drain_queue(q, true);
+	/* drain all requests queued before DEAD marking */
+	blk_drain_queue(q, true);
 
 	/* @q won't process any more request, flush async actions */
 	del_timer_sync(&q->backing_dev_info.laptop_mode_wb_timer);
@@ -498,14 +557,15 @@ struct request_queue *blk_alloc_queue_node(gfp_t gfp_mask, int node_id)
 	if (err)
 		goto fail_id;
 
-	if (blk_throtl_init(q))
-		goto fail_id;
-
 	setup_timer(&q->backing_dev_info.laptop_mode_wb_timer,
 		    laptop_mode_timer_fn, (unsigned long) q);
 	setup_timer(&q->timeout, blk_rq_timed_out_timer, (unsigned long) q);
+	INIT_LIST_HEAD(&q->queue_head);
 	INIT_LIST_HEAD(&q->timeout_list);
 	INIT_LIST_HEAD(&q->icq_list);
+#ifdef CONFIG_BLK_CGROUP
+	INIT_LIST_HEAD(&q->blkg_list);
+#endif
 	INIT_LIST_HEAD(&q->flush_queue[0]);
 	INIT_LIST_HEAD(&q->flush_queue[1]);
 	INIT_LIST_HEAD(&q->flush_data_in_flight);
@@ -522,6 +582,18 @@ struct request_queue *blk_alloc_queue_node(gfp_t gfp_mask, int node_id)
 	 */
 	q->queue_lock = &q->__queue_lock;
 
+	/*
+	 * A queue starts its life with bypass turned on to avoid
+	 * unnecessary bypass on/off overhead and nasty surprises during
+	 * init.  The initial bypass will be finished at the end of
+	 * blk_init_allocated_queue().
+	 */
+	q->bypass_depth = 1;
+	__set_bit(QUEUE_FLAG_BYPASS, &q->queue_flags);
+
+	if (blkcg_init_queue(q))
+		goto fail_id;
+
 	return q;
 
 fail_id:
@@ -614,15 +686,15 @@ blk_init_allocated_queue(struct request_queue *q, request_fn_proc *rfn,
 
 	q->sg_reserved_size = INT_MAX;
 
-	/*
-	 * all done
-	 */
-	if (!elevator_init(q, NULL)) {
-		blk_queue_congestion_threshold(q);
-		return q;
-	}
+	/* init elevator */
+	if (elevator_init(q, NULL))
+		return NULL;
 
-	return NULL;
+	blk_queue_congestion_threshold(q);
+
+	/* all done, end the initial bypass */
+	blk_queue_bypass_end(q);
+	return q;
 }
 EXPORT_SYMBOL(blk_init_allocated_queue);
 
@@ -648,33 +720,6 @@ static inline void blk_free_request(struct request_queue *q, struct request *rq)
 	mempool_free(rq, q->rq.rq_pool);
 }
 
-static struct request *
-blk_alloc_request(struct request_queue *q, struct io_cq *icq,
-		  unsigned int flags, gfp_t gfp_mask)
-{
-	struct request *rq = mempool_alloc(q->rq.rq_pool, gfp_mask);
-
-	if (!rq)
-		return NULL;
-
-	blk_rq_init(q, rq);
-
-	rq->cmd_flags = flags | REQ_ALLOCED;
-
-	if (flags & REQ_ELVPRIV) {
-		rq->elv.icq = icq;
-		if (unlikely(elv_set_request(q, rq, gfp_mask))) {
-			mempool_free(rq, q->rq.rq_pool);
-			return NULL;
-		}
-		/* @rq->elv.icq holds on to io_context until @rq is freed */
-		if (icq)
-			get_io_context(icq->ioc);
-	}
-
-	return rq;
-}
-
 /*
  * ioc_batching returns true if the ioc is a valid batching request and
  * should be given priority access to a request.
@@ -763,6 +808,22 @@ static bool blk_rq_should_init_elevator(struct bio *bio)
 }
 
 /**
+ * rq_ioc - determine io_context for request allocation
+ * @bio: request being allocated is for this bio (can be %NULL)
+ *
+ * Determine io_context to use for request allocation for @bio.  May return
+ * %NULL if %current->io_context doesn't exist.
+ */
+static struct io_context *rq_ioc(struct bio *bio)
+{
+#ifdef CONFIG_BLK_CGROUP
+	if (bio && bio->bi_ioc)
+		return bio->bi_ioc;
+#endif
+	return current->io_context;
+}
+
+/**
  * get_request - get a free request
  * @q: request_queue to allocate request from
  * @rw_flags: RW and SYNC flags
@@ -779,7 +840,7 @@ static bool blk_rq_should_init_elevator(struct bio *bio)
 static struct request *get_request(struct request_queue *q, int rw_flags,
 				   struct bio *bio, gfp_t gfp_mask)
 {
-	struct request *rq = NULL;
+	struct request *rq;
 	struct request_list *rl = &q->rq;
 	struct elevator_type *et;
 	struct io_context *ioc;
@@ -789,7 +850,7 @@ static struct request *get_request(struct request_queue *q, int rw_flags,
 	int may_queue;
 retry:
 	et = q->elevator->type;
-	ioc = current->io_context;
+	ioc = rq_ioc(bio);
 
 	if (unlikely(blk_queue_dead(q)))
 		return NULL;
@@ -808,7 +869,7 @@ retry:
 			 */
 			if (!ioc && !retried) {
 				spin_unlock_irq(q->queue_lock);
-				create_io_context(current, gfp_mask, q->node);
+				create_io_context(gfp_mask, q->node);
 				spin_lock_irq(q->queue_lock);
 				retried = true;
 				goto retry;
@@ -831,7 +892,7 @@ retry:
 					 * process is not a "batcher", and not
 					 * exempted by the IO scheduler
 					 */
-					goto out;
+					return NULL;
 				}
 			}
 		}
@@ -844,7 +905,7 @@ retry:
 	 * allocated with any setting of ->nr_requests
 	 */
 	if (rl->count[is_sync] >= (3 * q->nr_requests / 2))
-		goto out;
+		return NULL;
 
 	rl->count[is_sync]++;
 	rl->starved[is_sync] = 0;
@@ -859,8 +920,7 @@ retry:
 	 * Also, lookup icq while holding queue_lock.  If it doesn't exist,
 	 * it will be created after releasing queue_lock.
 	 */
-	if (blk_rq_should_init_elevator(bio) &&
-	    !test_bit(QUEUE_FLAG_ELVSWITCH, &q->queue_flags)) {
+	if (blk_rq_should_init_elevator(bio) && !blk_queue_bypass(q)) {
 		rw_flags |= REQ_ELVPRIV;
 		rl->elvpriv++;
 		if (et->icq_cache && ioc)
@@ -871,41 +931,36 @@ retry:
 		rw_flags |= REQ_IO_STAT;
 	spin_unlock_irq(q->queue_lock);
 
-	/* create icq if missing */
-	if ((rw_flags & REQ_ELVPRIV) && unlikely(et->icq_cache && !icq)) {
-		icq = ioc_create_icq(q, gfp_mask);
-		if (!icq)
-			goto fail_icq;
-	}
-
-	rq = blk_alloc_request(q, icq, rw_flags, gfp_mask);
+	/* allocate and init request */
+	rq = mempool_alloc(q->rq.rq_pool, gfp_mask);
+	if (!rq)
+		goto fail_alloc;
 
-fail_icq:
-	if (unlikely(!rq)) {
-		/*
-		 * Allocation failed presumably due to memory. Undo anything
-		 * we might have messed up.
-		 *
-		 * Allocating task should really be put onto the front of the
-		 * wait queue, but this is pretty rare.
-		 */
-		spin_lock_irq(q->queue_lock);
-		freed_request(q, rw_flags);
+	blk_rq_init(q, rq);
+	rq->cmd_flags = rw_flags | REQ_ALLOCED;
+
+	/* init elvpriv */
+	if (rw_flags & REQ_ELVPRIV) {
+		if (unlikely(et->icq_cache && !icq)) {
+			create_io_context(gfp_mask, q->node);
+			ioc = rq_ioc(bio);
+			if (!ioc)
+				goto fail_elvpriv;
+
+			icq = ioc_create_icq(ioc, q, gfp_mask);
+			if (!icq)
+				goto fail_elvpriv;
+		}
 
-		/*
-		 * in the very unlikely event that allocation failed and no
-		 * requests for this direction was pending, mark us starved
-		 * so that freeing of a request in the other direction will
-		 * notice us. another possible fix would be to split the
-		 * rq mempool into READ and WRITE
-		 */
-rq_starved:
-		if (unlikely(rl->count[is_sync] == 0))
-			rl->starved[is_sync] = 1;
+		rq->elv.icq = icq;
+		if (unlikely(elv_set_request(q, rq, bio, gfp_mask)))
+			goto fail_elvpriv;
 
-		goto out;
+		/* @rq->elv.icq holds io_context until @rq is freed */
+		if (icq)
+			get_io_context(icq->ioc);
 	}
-
+out:
 	/*
 	 * ioc may be NULL here, and ioc_batching will be false. That's
 	 * OK, if the queue is under the request limit then requests need
@@ -916,8 +971,48 @@ rq_starved:
 		ioc->nr_batch_requests--;
 
 	trace_block_getrq(q, bio, rw_flags & 1);
-out:
 	return rq;
+
+fail_elvpriv:
+	/*
+	 * elvpriv init failed.  ioc, icq and elvpriv aren't mempool backed
+	 * and may fail indefinitely under memory pressure and thus
+	 * shouldn't stall IO.  Treat this request as !elvpriv.  This will
+	 * disturb iosched and blkcg but weird is bettern than dead.
+	 */
+	printk_ratelimited(KERN_WARNING "%s: request aux data allocation failed, iosched may be disturbed\n",
+			   dev_name(q->backing_dev_info.dev));
+
+	rq->cmd_flags &= ~REQ_ELVPRIV;
+	rq->elv.icq = NULL;
+
+	spin_lock_irq(q->queue_lock);
+	rl->elvpriv--;
+	spin_unlock_irq(q->queue_lock);
+	goto out;
+
+fail_alloc:
+	/*
+	 * Allocation failed presumably due to memory. Undo anything we
+	 * might have messed up.
+	 *
+	 * Allocating task should really be put onto the front of the wait
+	 * queue, but this is pretty rare.
+	 */
+	spin_lock_irq(q->queue_lock);
+	freed_request(q, rw_flags);
+
+	/*
+	 * in the very unlikely event that allocation failed and no
+	 * requests for this direction was pending, mark us starved so that
+	 * freeing of a request in the other direction will notice
+	 * us. another possible fix would be to split the rq mempool into
+	 * READ and WRITE
+	 */
+rq_starved:
+	if (unlikely(rl->count[is_sync] == 0))
+		rl->starved[is_sync] = 1;
+	return NULL;
 }
 
 /**
@@ -961,7 +1056,7 @@ static struct request *get_request_wait(struct request_queue *q, int rw_flags,
 		 * up to a big batch of them for a small period time.
 		 * See ioc_batching, ioc_set_batching
 		 */
-		create_io_context(current, GFP_NOIO, q->node);
+		create_io_context(GFP_NOIO, q->node);
 		ioc_set_batching(q, current->io_context);
 
 		spin_lock_irq(q->queue_lock);
diff --git a/block/blk-ioc.c b/block/blk-ioc.c
index fb95dd2f889a..1e2d53b04858 100644
--- a/block/blk-ioc.c
+++ b/block/blk-ioc.c
@@ -155,20 +155,20 @@ void put_io_context(struct io_context *ioc)
 }
 EXPORT_SYMBOL(put_io_context);
 
-/* Called by the exiting task */
-void exit_io_context(struct task_struct *task)
+/**
+ * put_io_context_active - put active reference on ioc
+ * @ioc: ioc of interest
+ *
+ * Undo get_io_context_active().  If active reference reaches zero after
+ * put, @ioc can never issue further IOs and ioscheds are notified.
+ */
+void put_io_context_active(struct io_context *ioc)
 {
-	struct io_context *ioc;
-	struct io_cq *icq;
 	struct hlist_node *n;
 	unsigned long flags;
+	struct io_cq *icq;
 
-	task_lock(task);
-	ioc = task->io_context;
-	task->io_context = NULL;
-	task_unlock(task);
-
-	if (!atomic_dec_and_test(&ioc->nr_tasks)) {
+	if (!atomic_dec_and_test(&ioc->active_ref)) {
 		put_io_context(ioc);
 		return;
 	}
@@ -197,6 +197,20 @@ retry:
 	put_io_context(ioc);
 }
 
+/* Called by the exiting task */
+void exit_io_context(struct task_struct *task)
+{
+	struct io_context *ioc;
+
+	task_lock(task);
+	ioc = task->io_context;
+	task->io_context = NULL;
+	task_unlock(task);
+
+	atomic_dec(&ioc->nr_tasks);
+	put_io_context_active(ioc);
+}
+
 /**
  * ioc_clear_queue - break any ioc association with the specified queue
  * @q: request_queue being cleared
@@ -218,19 +232,18 @@ void ioc_clear_queue(struct request_queue *q)
 	}
 }
 
-void create_io_context_slowpath(struct task_struct *task, gfp_t gfp_flags,
-				int node)
+int create_task_io_context(struct task_struct *task, gfp_t gfp_flags, int node)
 {
 	struct io_context *ioc;
 
 	ioc = kmem_cache_alloc_node(iocontext_cachep, gfp_flags | __GFP_ZERO,
 				    node);
 	if (unlikely(!ioc))
-		return;
+		return -ENOMEM;
 
 	/* initialize */
 	atomic_long_set(&ioc->refcount, 1);
-	atomic_set(&ioc->nr_tasks, 1);
+	atomic_set(&ioc->active_ref, 1);
 	spin_lock_init(&ioc->lock);
 	INIT_RADIX_TREE(&ioc->icq_tree, GFP_ATOMIC | __GFP_HIGH);
 	INIT_HLIST_HEAD(&ioc->icq_list);
@@ -250,6 +263,8 @@ void create_io_context_slowpath(struct task_struct *task, gfp_t gfp_flags,
 	else
 		kmem_cache_free(iocontext_cachep, ioc);
 	task_unlock(task);
+
+	return 0;
 }
 
 /**
@@ -281,7 +296,7 @@ struct io_context *get_task_io_context(struct task_struct *task,
 			return ioc;
 		}
 		task_unlock(task);
-	} while (create_io_context(task, gfp_flags, node));
+	} while (!create_task_io_context(task, gfp_flags, node));
 
 	return NULL;
 }
@@ -325,26 +340,23 @@ EXPORT_SYMBOL(ioc_lookup_icq);
 
 /**
  * ioc_create_icq - create and link io_cq
+ * @ioc: io_context of interest
  * @q: request_queue of interest
  * @gfp_mask: allocation mask
  *
- * Make sure io_cq linking %current->io_context and @q exists.  If either
- * io_context and/or icq don't exist, they will be created using @gfp_mask.
+ * Make sure io_cq linking @ioc and @q exists.  If icq doesn't exist, they
+ * will be created using @gfp_mask.
  *
  * The caller is responsible for ensuring @ioc won't go away and @q is
  * alive and will stay alive until this function returns.
  */
-struct io_cq *ioc_create_icq(struct request_queue *q, gfp_t gfp_mask)
+struct io_cq *ioc_create_icq(struct io_context *ioc, struct request_queue *q,
+			     gfp_t gfp_mask)
 {
 	struct elevator_type *et = q->elevator->type;
-	struct io_context *ioc;
 	struct io_cq *icq;
 
 	/* allocate stuff */
-	ioc = create_io_context(current, gfp_mask, q->node);
-	if (!ioc)
-		return NULL;
-
 	icq = kmem_cache_alloc_node(et->icq_cache, gfp_mask | __GFP_ZERO,
 				    q->node);
 	if (!icq)
@@ -382,74 +394,6 @@ struct io_cq *ioc_create_icq(struct request_queue *q, gfp_t gfp_mask)
 	return icq;
 }
 
-void ioc_set_icq_flags(struct io_context *ioc, unsigned int flags)
-{
-	struct io_cq *icq;
-	struct hlist_node *n;
-
-	hlist_for_each_entry(icq, n, &ioc->icq_list, ioc_node)
-		icq->flags |= flags;
-}
-
-/**
- * ioc_ioprio_changed - notify ioprio change
- * @ioc: io_context of interest
- * @ioprio: new ioprio
- *
- * @ioc's ioprio has changed to @ioprio.  Set %ICQ_IOPRIO_CHANGED for all
- * icq's.  iosched is responsible for checking the bit and applying it on
- * request issue path.
- */
-void ioc_ioprio_changed(struct io_context *ioc, int ioprio)
-{
-	unsigned long flags;
-
-	spin_lock_irqsave(&ioc->lock, flags);
-	ioc->ioprio = ioprio;
-	ioc_set_icq_flags(ioc, ICQ_IOPRIO_CHANGED);
-	spin_unlock_irqrestore(&ioc->lock, flags);
-}
-
-/**
- * ioc_cgroup_changed - notify cgroup change
- * @ioc: io_context of interest
- *
- * @ioc's cgroup has changed.  Set %ICQ_CGROUP_CHANGED for all icq's.
- * iosched is responsible for checking the bit and applying it on request
- * issue path.
- */
-void ioc_cgroup_changed(struct io_context *ioc)
-{
-	unsigned long flags;
-
-	spin_lock_irqsave(&ioc->lock, flags);
-	ioc_set_icq_flags(ioc, ICQ_CGROUP_CHANGED);
-	spin_unlock_irqrestore(&ioc->lock, flags);
-}
-EXPORT_SYMBOL(ioc_cgroup_changed);
-
-/**
- * icq_get_changed - fetch and clear icq changed mask
- * @icq: icq of interest
- *
- * Fetch and clear ICQ_*_CHANGED bits from @icq.  Grabs and releases
- * @icq->ioc->lock.
- */
-unsigned icq_get_changed(struct io_cq *icq)
-{
-	unsigned int changed = 0;
-	unsigned long flags;
-
-	if (unlikely(icq->flags & ICQ_CHANGED_MASK)) {
-		spin_lock_irqsave(&icq->ioc->lock, flags);
-		changed = icq->flags & ICQ_CHANGED_MASK;
-		icq->flags &= ~ICQ_CHANGED_MASK;
-		spin_unlock_irqrestore(&icq->ioc->lock, flags);
-	}
-	return changed;
-}
-EXPORT_SYMBOL(icq_get_changed);
-
 static int __init blk_ioc_init(void)
 {
 	iocontext_cachep = kmem_cache_create("blkdev_ioc",
diff --git a/block/blk-sysfs.c b/block/blk-sysfs.c
index cf150011d808..aa41b47c22d2 100644
--- a/block/blk-sysfs.c
+++ b/block/blk-sysfs.c
@@ -9,6 +9,7 @@
 #include <linux/blktrace_api.h>
 
 #include "blk.h"
+#include "blk-cgroup.h"
 
 struct queue_sysfs_entry {
 	struct attribute attr;
@@ -479,6 +480,8 @@ static void blk_release_queue(struct kobject *kobj)
 
 	blk_sync_queue(q);
 
+	blkcg_exit_queue(q);
+
 	if (q->elevator) {
 		spin_lock_irq(q->queue_lock);
 		ioc_clear_queue(q);
@@ -486,15 +489,12 @@ static void blk_release_queue(struct kobject *kobj)
 		elevator_exit(q->elevator);
 	}
 
-	blk_throtl_exit(q);
-
 	if (rl->rq_pool)
 		mempool_destroy(rl->rq_pool);
 
 	if (q->queue_tags)
 		__blk_queue_free_tags(q);
 
-	blk_throtl_release(q);
 	blk_trace_shutdown(q);
 
 	bdi_destroy(&q->backing_dev_info);
diff --git a/block/blk-throttle.c b/block/blk-throttle.c
index f2ddb94626bd..5b0659512047 100644
--- a/block/blk-throttle.c
+++ b/block/blk-throttle.c
@@ -21,6 +21,8 @@ static int throtl_quantum = 32;
 /* Throttling is performed over 100ms slice and after that slice is renewed */
 static unsigned long throtl_slice = HZ/10;	/* 100 ms */
 
+static struct blkcg_policy blkcg_policy_throtl;
+
 /* A workqueue to queue throttle related work */
 static struct workqueue_struct *kthrotld_workqueue;
 static void throtl_schedule_delayed_work(struct throtl_data *td,
@@ -38,9 +40,17 @@ struct throtl_rb_root {
 
 #define rb_entry_tg(node)	rb_entry((node), struct throtl_grp, rb_node)
 
+/* Per-cpu group stats */
+struct tg_stats_cpu {
+	/* total bytes transferred */
+	struct blkg_rwstat		service_bytes;
+	/* total IOs serviced, post merge */
+	struct blkg_rwstat		serviced;
+};
+
 struct throtl_grp {
-	/* List of throtl groups on the request queue*/
-	struct hlist_node tg_node;
+	/* must be the first member */
+	struct blkg_policy_data pd;
 
 	/* active throtl group service_tree member */
 	struct rb_node rb_node;
@@ -52,8 +62,6 @@ struct throtl_grp {
 	 */
 	unsigned long disptime;
 
-	struct blkio_group blkg;
-	atomic_t ref;
 	unsigned int flags;
 
 	/* Two lists for READ and WRITE */
@@ -80,18 +88,18 @@ struct throtl_grp {
 	/* Some throttle limits got updated for the group */
 	int limits_changed;
 
-	struct rcu_head rcu_head;
+	/* Per cpu stats pointer */
+	struct tg_stats_cpu __percpu *stats_cpu;
+
+	/* List of tgs waiting for per cpu stats memory to be allocated */
+	struct list_head stats_alloc_node;
 };
 
 struct throtl_data
 {
-	/* List of throtl groups */
-	struct hlist_head tg_list;
-
 	/* service tree for active throtl groups */
 	struct throtl_rb_root tg_service_tree;
 
-	struct throtl_grp *root_tg;
 	struct request_queue *queue;
 
 	/* Total Number of queued bios on READ and WRITE lists */
@@ -108,6 +116,33 @@ struct throtl_data
 	int limits_changed;
 };
 
+/* list and work item to allocate percpu group stats */
+static DEFINE_SPINLOCK(tg_stats_alloc_lock);
+static LIST_HEAD(tg_stats_alloc_list);
+
+static void tg_stats_alloc_fn(struct work_struct *);
+static DECLARE_DELAYED_WORK(tg_stats_alloc_work, tg_stats_alloc_fn);
+
+static inline struct throtl_grp *pd_to_tg(struct blkg_policy_data *pd)
+{
+	return pd ? container_of(pd, struct throtl_grp, pd) : NULL;
+}
+
+static inline struct throtl_grp *blkg_to_tg(struct blkcg_gq *blkg)
+{
+	return pd_to_tg(blkg_to_pd(blkg, &blkcg_policy_throtl));
+}
+
+static inline struct blkcg_gq *tg_to_blkg(struct throtl_grp *tg)
+{
+	return pd_to_blkg(&tg->pd);
+}
+
+static inline struct throtl_grp *td_root_tg(struct throtl_data *td)
+{
+	return blkg_to_tg(td->queue->root_blkg);
+}
+
 enum tg_state_flags {
 	THROTL_TG_FLAG_on_rr = 0,	/* on round-robin busy list */
 };
@@ -128,244 +163,150 @@ static inline int throtl_tg_##name(const struct throtl_grp *tg)		\
 
 THROTL_TG_FNS(on_rr);
 
-#define throtl_log_tg(td, tg, fmt, args...)				\
-	blk_add_trace_msg((td)->queue, "throtl %s " fmt,		\
-				blkg_path(&(tg)->blkg), ##args);      	\
+#define throtl_log_tg(td, tg, fmt, args...)	do {			\
+	char __pbuf[128];						\
+									\
+	blkg_path(tg_to_blkg(tg), __pbuf, sizeof(__pbuf));		\
+	blk_add_trace_msg((td)->queue, "throtl %s " fmt, __pbuf, ##args); \
+} while (0)
 
 #define throtl_log(td, fmt, args...)	\
 	blk_add_trace_msg((td)->queue, "throtl " fmt, ##args)
 
-static inline struct throtl_grp *tg_of_blkg(struct blkio_group *blkg)
-{
-	if (blkg)
-		return container_of(blkg, struct throtl_grp, blkg);
-
-	return NULL;
-}
-
 static inline unsigned int total_nr_queued(struct throtl_data *td)
 {
 	return td->nr_queued[0] + td->nr_queued[1];
 }
 
-static inline struct throtl_grp *throtl_ref_get_tg(struct throtl_grp *tg)
-{
-	atomic_inc(&tg->ref);
-	return tg;
-}
-
-static void throtl_free_tg(struct rcu_head *head)
+/*
+ * Worker for allocating per cpu stat for tgs. This is scheduled on the
+ * system_nrt_wq once there are some groups on the alloc_list waiting for
+ * allocation.
+ */
+static void tg_stats_alloc_fn(struct work_struct *work)
 {
-	struct throtl_grp *tg;
+	static struct tg_stats_cpu *stats_cpu;	/* this fn is non-reentrant */
+	struct delayed_work *dwork = to_delayed_work(work);
+	bool empty = false;
+
+alloc_stats:
+	if (!stats_cpu) {
+		stats_cpu = alloc_percpu(struct tg_stats_cpu);
+		if (!stats_cpu) {
+			/* allocation failed, try again after some time */
+			queue_delayed_work(system_nrt_wq, dwork,
+					   msecs_to_jiffies(10));
+			return;
+		}
+	}
 
-	tg = container_of(head, struct throtl_grp, rcu_head);
-	free_percpu(tg->blkg.stats_cpu);
-	kfree(tg);
-}
+	spin_lock_irq(&tg_stats_alloc_lock);
 
-static void throtl_put_tg(struct throtl_grp *tg)
-{
-	BUG_ON(atomic_read(&tg->ref) <= 0);
-	if (!atomic_dec_and_test(&tg->ref))
-		return;
+	if (!list_empty(&tg_stats_alloc_list)) {
+		struct throtl_grp *tg = list_first_entry(&tg_stats_alloc_list,
+							 struct throtl_grp,
+							 stats_alloc_node);
+		swap(tg->stats_cpu, stats_cpu);
+		list_del_init(&tg->stats_alloc_node);
+	}
 
-	/*
-	 * A group is freed in rcu manner. But having an rcu lock does not
-	 * mean that one can access all the fields of blkg and assume these
-	 * are valid. For example, don't try to follow throtl_data and
-	 * request queue links.
-	 *
-	 * Having a reference to blkg under an rcu allows acess to only
-	 * values local to groups like group stats and group rate limits
-	 */
-	call_rcu(&tg->rcu_head, throtl_free_tg);
+	empty = list_empty(&tg_stats_alloc_list);
+	spin_unlock_irq(&tg_stats_alloc_lock);
+	if (!empty)
+		goto alloc_stats;
 }
 
-static void throtl_init_group(struct throtl_grp *tg)
+static void throtl_pd_init(struct blkcg_gq *blkg)
 {
-	INIT_HLIST_NODE(&tg->tg_node);
+	struct throtl_grp *tg = blkg_to_tg(blkg);
+	unsigned long flags;
+
 	RB_CLEAR_NODE(&tg->rb_node);
 	bio_list_init(&tg->bio_lists[0]);
 	bio_list_init(&tg->bio_lists[1]);
 	tg->limits_changed = false;
 
-	/* Practically unlimited BW */
-	tg->bps[0] = tg->bps[1] = -1;
-	tg->iops[0] = tg->iops[1] = -1;
+	tg->bps[READ] = -1;
+	tg->bps[WRITE] = -1;
+	tg->iops[READ] = -1;
+	tg->iops[WRITE] = -1;
 
 	/*
-	 * Take the initial reference that will be released on destroy
-	 * This can be thought of a joint reference by cgroup and
-	 * request queue which will be dropped by either request queue
-	 * exit or cgroup deletion path depending on who is exiting first.
+	 * Ugh... We need to perform per-cpu allocation for tg->stats_cpu
+	 * but percpu allocator can't be called from IO path.  Queue tg on
+	 * tg_stats_alloc_list and allocate from work item.
 	 */
-	atomic_set(&tg->ref, 1);
+	spin_lock_irqsave(&tg_stats_alloc_lock, flags);
+	list_add(&tg->stats_alloc_node, &tg_stats_alloc_list);
+	queue_delayed_work(system_nrt_wq, &tg_stats_alloc_work, 0);
+	spin_unlock_irqrestore(&tg_stats_alloc_lock, flags);
 }
 
-/* Should be called with rcu read lock held (needed for blkcg) */
-static void
-throtl_add_group_to_td_list(struct throtl_data *td, struct throtl_grp *tg)
+static void throtl_pd_exit(struct blkcg_gq *blkg)
 {
-	hlist_add_head(&tg->tg_node, &td->tg_list);
-	td->nr_undestroyed_grps++;
-}
-
-static void
-__throtl_tg_fill_dev_details(struct throtl_data *td, struct throtl_grp *tg)
-{
-	struct backing_dev_info *bdi = &td->queue->backing_dev_info;
-	unsigned int major, minor;
-
-	if (!tg || tg->blkg.dev)
-		return;
-
-	/*
-	 * Fill in device details for a group which might not have been
-	 * filled at group creation time as queue was being instantiated
-	 * and driver had not attached a device yet
-	 */
-	if (bdi->dev && dev_name(bdi->dev)) {
-		sscanf(dev_name(bdi->dev), "%u:%u", &major, &minor);
-		tg->blkg.dev = MKDEV(major, minor);
-	}
-}
-
-/*
- * Should be called with without queue lock held. Here queue lock will be
- * taken rarely. It will be taken only once during life time of a group
- * if need be
- */
-static void
-throtl_tg_fill_dev_details(struct throtl_data *td, struct throtl_grp *tg)
-{
-	if (!tg || tg->blkg.dev)
-		return;
-
-	spin_lock_irq(td->queue->queue_lock);
-	__throtl_tg_fill_dev_details(td, tg);
-	spin_unlock_irq(td->queue->queue_lock);
-}
-
-static void throtl_init_add_tg_lists(struct throtl_data *td,
-			struct throtl_grp *tg, struct blkio_cgroup *blkcg)
-{
-	__throtl_tg_fill_dev_details(td, tg);
-
-	/* Add group onto cgroup list */
-	blkiocg_add_blkio_group(blkcg, &tg->blkg, (void *)td,
-				tg->blkg.dev, BLKIO_POLICY_THROTL);
+	struct throtl_grp *tg = blkg_to_tg(blkg);
+	unsigned long flags;
 
-	tg->bps[READ] = blkcg_get_read_bps(blkcg, tg->blkg.dev);
-	tg->bps[WRITE] = blkcg_get_write_bps(blkcg, tg->blkg.dev);
-	tg->iops[READ] = blkcg_get_read_iops(blkcg, tg->blkg.dev);
-	tg->iops[WRITE] = blkcg_get_write_iops(blkcg, tg->blkg.dev);
+	spin_lock_irqsave(&tg_stats_alloc_lock, flags);
+	list_del_init(&tg->stats_alloc_node);
+	spin_unlock_irqrestore(&tg_stats_alloc_lock, flags);
 
-	throtl_add_group_to_td_list(td, tg);
+	free_percpu(tg->stats_cpu);
 }
 
-/* Should be called without queue lock and outside of rcu period */
-static struct throtl_grp *throtl_alloc_tg(struct throtl_data *td)
+static void throtl_pd_reset_stats(struct blkcg_gq *blkg)
 {
-	struct throtl_grp *tg = NULL;
-	int ret;
+	struct throtl_grp *tg = blkg_to_tg(blkg);
+	int cpu;
 
-	tg = kzalloc_node(sizeof(*tg), GFP_ATOMIC, td->queue->node);
-	if (!tg)
-		return NULL;
+	if (tg->stats_cpu == NULL)
+		return;
 
-	ret = blkio_alloc_blkg_stats(&tg->blkg);
+	for_each_possible_cpu(cpu) {
+		struct tg_stats_cpu *sc = per_cpu_ptr(tg->stats_cpu, cpu);
 
-	if (ret) {
-		kfree(tg);
-		return NULL;
+		blkg_rwstat_reset(&sc->service_bytes);
+		blkg_rwstat_reset(&sc->serviced);
 	}
-
-	throtl_init_group(tg);
-	return tg;
 }
 
-static struct
-throtl_grp *throtl_find_tg(struct throtl_data *td, struct blkio_cgroup *blkcg)
+static struct throtl_grp *throtl_lookup_tg(struct throtl_data *td,
+					   struct blkcg *blkcg)
 {
-	struct throtl_grp *tg = NULL;
-	void *key = td;
-
 	/*
-	 * This is the common case when there are no blkio cgroups.
- 	 * Avoid lookup in this case
- 	 */
-	if (blkcg == &blkio_root_cgroup)
-		tg = td->root_tg;
-	else
-		tg = tg_of_blkg(blkiocg_lookup_group(blkcg, key));
+	 * This is the common case when there are no blkcgs.  Avoid lookup
+	 * in this case
+	 */
+	if (blkcg == &blkcg_root)
+		return td_root_tg(td);
 
-	__throtl_tg_fill_dev_details(td, tg);
-	return tg;
+	return blkg_to_tg(blkg_lookup(blkcg, td->queue));
 }
 
-static struct throtl_grp * throtl_get_tg(struct throtl_data *td)
+static struct throtl_grp *throtl_lookup_create_tg(struct throtl_data *td,
+						  struct blkcg *blkcg)
 {
-	struct throtl_grp *tg = NULL, *__tg = NULL;
-	struct blkio_cgroup *blkcg;
 	struct request_queue *q = td->queue;
-
-	/* no throttling for dead queue */
-	if (unlikely(blk_queue_dead(q)))
-		return NULL;
-
-	rcu_read_lock();
-	blkcg = task_blkio_cgroup(current);
-	tg = throtl_find_tg(td, blkcg);
-	if (tg) {
-		rcu_read_unlock();
-		return tg;
-	}
-
-	/*
-	 * Need to allocate a group. Allocation of group also needs allocation
-	 * of per cpu stats which in-turn takes a mutex() and can block. Hence
-	 * we need to drop rcu lock and queue_lock before we call alloc.
-	 */
-	rcu_read_unlock();
-	spin_unlock_irq(q->queue_lock);
-
-	tg = throtl_alloc_tg(td);
-
-	/* Group allocated and queue is still alive. take the lock */
-	spin_lock_irq(q->queue_lock);
-
-	/* Make sure @q is still alive */
-	if (unlikely(blk_queue_dead(q))) {
-		kfree(tg);
-		return NULL;
-	}
-
-	/*
-	 * Initialize the new group. After sleeping, read the blkcg again.
-	 */
-	rcu_read_lock();
-	blkcg = task_blkio_cgroup(current);
+	struct throtl_grp *tg = NULL;
 
 	/*
-	 * If some other thread already allocated the group while we were
-	 * not holding queue lock, free up the group
+	 * This is the common case when there are no blkcgs.  Avoid lookup
+	 * in this case
 	 */
-	__tg = throtl_find_tg(td, blkcg);
-
-	if (__tg) {
-		kfree(tg);
-		rcu_read_unlock();
-		return __tg;
-	}
-
-	/* Group allocation failed. Account the IO to root group */
-	if (!tg) {
-		tg = td->root_tg;
-		return tg;
+	if (blkcg == &blkcg_root) {
+		tg = td_root_tg(td);
+	} else {
+		struct blkcg_gq *blkg;
+
+		blkg = blkg_lookup_create(blkcg, q);
+
+		/* if %NULL and @q is alive, fall back to root_tg */
+		if (!IS_ERR(blkg))
+			tg = blkg_to_tg(blkg);
+		else if (!blk_queue_dead(q))
+			tg = td_root_tg(td);
 	}
 
-	throtl_init_add_tg_lists(td, tg, blkcg);
-	rcu_read_unlock();
 	return tg;
 }
 
@@ -734,16 +675,41 @@ static bool tg_may_dispatch(struct throtl_data *td, struct throtl_grp *tg,
 	return 0;
 }
 
+static void throtl_update_dispatch_stats(struct blkcg_gq *blkg, u64 bytes,
+					 int rw)
+{
+	struct throtl_grp *tg = blkg_to_tg(blkg);
+	struct tg_stats_cpu *stats_cpu;
+	unsigned long flags;
+
+	/* If per cpu stats are not allocated yet, don't do any accounting. */
+	if (tg->stats_cpu == NULL)
+		return;
+
+	/*
+	 * Disabling interrupts to provide mutual exclusion between two
+	 * writes on same cpu. It probably is not needed for 64bit. Not
+	 * optimizing that case yet.
+	 */
+	local_irq_save(flags);
+
+	stats_cpu = this_cpu_ptr(tg->stats_cpu);
+
+	blkg_rwstat_add(&stats_cpu->serviced, rw, 1);
+	blkg_rwstat_add(&stats_cpu->service_bytes, rw, bytes);
+
+	local_irq_restore(flags);
+}
+
 static void throtl_charge_bio(struct throtl_grp *tg, struct bio *bio)
 {
 	bool rw = bio_data_dir(bio);
-	bool sync = rw_is_sync(bio->bi_rw);
 
 	/* Charge the bio to the group */
 	tg->bytes_disp[rw] += bio->bi_size;
 	tg->io_disp[rw]++;
 
-	blkiocg_update_dispatch_stats(&tg->blkg, bio->bi_size, rw, sync);
+	throtl_update_dispatch_stats(tg_to_blkg(tg), bio->bi_size, bio->bi_rw);
 }
 
 static void throtl_add_bio_tg(struct throtl_data *td, struct throtl_grp *tg,
@@ -753,7 +719,7 @@ static void throtl_add_bio_tg(struct throtl_data *td, struct throtl_grp *tg,
 
 	bio_list_add(&tg->bio_lists[rw], bio);
 	/* Take a bio reference on tg */
-	throtl_ref_get_tg(tg);
+	blkg_get(tg_to_blkg(tg));
 	tg->nr_queued[rw]++;
 	td->nr_queued[rw]++;
 	throtl_enqueue_tg(td, tg);
@@ -786,8 +752,8 @@ static void tg_dispatch_one_bio(struct throtl_data *td, struct throtl_grp *tg,
 
 	bio = bio_list_pop(&tg->bio_lists[rw]);
 	tg->nr_queued[rw]--;
-	/* Drop bio reference on tg */
-	throtl_put_tg(tg);
+	/* Drop bio reference on blkg */
+	blkg_put(tg_to_blkg(tg));
 
 	BUG_ON(td->nr_queued[rw] <= 0);
 	td->nr_queued[rw]--;
@@ -865,8 +831,8 @@ static int throtl_select_dispatch(struct throtl_data *td, struct bio_list *bl)
 
 static void throtl_process_limit_change(struct throtl_data *td)
 {
-	struct throtl_grp *tg;
-	struct hlist_node *pos, *n;
+	struct request_queue *q = td->queue;
+	struct blkcg_gq *blkg, *n;
 
 	if (!td->limits_changed)
 		return;
@@ -875,7 +841,9 @@ static void throtl_process_limit_change(struct throtl_data *td)
 
 	throtl_log(td, "limits changed");
 
-	hlist_for_each_entry_safe(tg, pos, n, &td->tg_list, tg_node) {
+	list_for_each_entry_safe(blkg, n, &q->blkg_list, q_node) {
+		struct throtl_grp *tg = blkg_to_tg(blkg);
+
 		if (!tg->limits_changed)
 			continue;
 
@@ -973,120 +941,159 @@ throtl_schedule_delayed_work(struct throtl_data *td, unsigned long delay)
 	}
 }
 
-static void
-throtl_destroy_tg(struct throtl_data *td, struct throtl_grp *tg)
+static u64 tg_prfill_cpu_rwstat(struct seq_file *sf,
+				struct blkg_policy_data *pd, int off)
 {
-	/* Something wrong if we are trying to remove same group twice */
-	BUG_ON(hlist_unhashed(&tg->tg_node));
+	struct throtl_grp *tg = pd_to_tg(pd);
+	struct blkg_rwstat rwstat = { }, tmp;
+	int i, cpu;
 
-	hlist_del_init(&tg->tg_node);
+	for_each_possible_cpu(cpu) {
+		struct tg_stats_cpu *sc = per_cpu_ptr(tg->stats_cpu, cpu);
 
-	/*
-	 * Put the reference taken at the time of creation so that when all
-	 * queues are gone, group can be destroyed.
-	 */
-	throtl_put_tg(tg);
-	td->nr_undestroyed_grps--;
+		tmp = blkg_rwstat_read((void *)sc + off);
+		for (i = 0; i < BLKG_RWSTAT_NR; i++)
+			rwstat.cnt[i] += tmp.cnt[i];
+	}
+
+	return __blkg_prfill_rwstat(sf, pd, &rwstat);
 }
 
-static void throtl_release_tgs(struct throtl_data *td)
+static int tg_print_cpu_rwstat(struct cgroup *cgrp, struct cftype *cft,
+			       struct seq_file *sf)
 {
-	struct hlist_node *pos, *n;
-	struct throtl_grp *tg;
+	struct blkcg *blkcg = cgroup_to_blkcg(cgrp);
 
-	hlist_for_each_entry_safe(tg, pos, n, &td->tg_list, tg_node) {
-		/*
-		 * If cgroup removal path got to blk_group first and removed
-		 * it from cgroup list, then it will take care of destroying
-		 * cfqg also.
-		 */
-		if (!blkiocg_del_blkio_group(&tg->blkg))
-			throtl_destroy_tg(td, tg);
-	}
+	blkcg_print_blkgs(sf, blkcg, tg_prfill_cpu_rwstat, &blkcg_policy_throtl,
+			  cft->private, true);
+	return 0;
 }
 
-/*
- * Blk cgroup controller notification saying that blkio_group object is being
- * delinked as associated cgroup object is going away. That also means that
- * no new IO will come in this group. So get rid of this group as soon as
- * any pending IO in the group is finished.
- *
- * This function is called under rcu_read_lock(). key is the rcu protected
- * pointer. That means "key" is a valid throtl_data pointer as long as we are
- * rcu read lock.
- *
- * "key" was fetched from blkio_group under blkio_cgroup->lock. That means
- * it should not be NULL as even if queue was going away, cgroup deltion
- * path got to it first.
- */
-void throtl_unlink_blkio_group(void *key, struct blkio_group *blkg)
+static u64 tg_prfill_conf_u64(struct seq_file *sf, struct blkg_policy_data *pd,
+			      int off)
 {
-	unsigned long flags;
-	struct throtl_data *td = key;
+	struct throtl_grp *tg = pd_to_tg(pd);
+	u64 v = *(u64 *)((void *)tg + off);
 
-	spin_lock_irqsave(td->queue->queue_lock, flags);
-	throtl_destroy_tg(td, tg_of_blkg(blkg));
-	spin_unlock_irqrestore(td->queue->queue_lock, flags);
+	if (v == -1)
+		return 0;
+	return __blkg_prfill_u64(sf, pd, v);
 }
 
-static void throtl_update_blkio_group_common(struct throtl_data *td,
-				struct throtl_grp *tg)
+static u64 tg_prfill_conf_uint(struct seq_file *sf, struct blkg_policy_data *pd,
+			       int off)
 {
-	xchg(&tg->limits_changed, true);
-	xchg(&td->limits_changed, true);
-	/* Schedule a work now to process the limit change */
-	throtl_schedule_delayed_work(td, 0);
+	struct throtl_grp *tg = pd_to_tg(pd);
+	unsigned int v = *(unsigned int *)((void *)tg + off);
+
+	if (v == -1)
+		return 0;
+	return __blkg_prfill_u64(sf, pd, v);
 }
 
-/*
- * For all update functions, key should be a valid pointer because these
- * update functions are called under blkcg_lock, that means, blkg is
- * valid and in turn key is valid. queue exit path can not race because
- * of blkcg_lock
- *
- * Can not take queue lock in update functions as queue lock under blkcg_lock
- * is not allowed. Under other paths we take blkcg_lock under queue_lock.
- */
-static void throtl_update_blkio_group_read_bps(void *key,
-				struct blkio_group *blkg, u64 read_bps)
+static int tg_print_conf_u64(struct cgroup *cgrp, struct cftype *cft,
+			     struct seq_file *sf)
 {
-	struct throtl_data *td = key;
-	struct throtl_grp *tg = tg_of_blkg(blkg);
-
-	tg->bps[READ] = read_bps;
-	throtl_update_blkio_group_common(td, tg);
+	blkcg_print_blkgs(sf, cgroup_to_blkcg(cgrp), tg_prfill_conf_u64,
+			  &blkcg_policy_throtl, cft->private, false);
+	return 0;
 }
 
-static void throtl_update_blkio_group_write_bps(void *key,
-				struct blkio_group *blkg, u64 write_bps)
+static int tg_print_conf_uint(struct cgroup *cgrp, struct cftype *cft,
+			      struct seq_file *sf)
 {
-	struct throtl_data *td = key;
-	struct throtl_grp *tg = tg_of_blkg(blkg);
-
-	tg->bps[WRITE] = write_bps;
-	throtl_update_blkio_group_common(td, tg);
+	blkcg_print_blkgs(sf, cgroup_to_blkcg(cgrp), tg_prfill_conf_uint,
+			  &blkcg_policy_throtl, cft->private, false);
+	return 0;
 }
 
-static void throtl_update_blkio_group_read_iops(void *key,
-			struct blkio_group *blkg, unsigned int read_iops)
+static int tg_set_conf(struct cgroup *cgrp, struct cftype *cft, const char *buf,
+		       bool is_u64)
 {
-	struct throtl_data *td = key;
-	struct throtl_grp *tg = tg_of_blkg(blkg);
+	struct blkcg *blkcg = cgroup_to_blkcg(cgrp);
+	struct blkg_conf_ctx ctx;
+	struct throtl_grp *tg;
+	struct throtl_data *td;
+	int ret;
+
+	ret = blkg_conf_prep(blkcg, &blkcg_policy_throtl, buf, &ctx);
+	if (ret)
+		return ret;
+
+	tg = blkg_to_tg(ctx.blkg);
+	td = ctx.blkg->q->td;
+
+	if (!ctx.v)
+		ctx.v = -1;
+
+	if (is_u64)
+		*(u64 *)((void *)tg + cft->private) = ctx.v;
+	else
+		*(unsigned int *)((void *)tg + cft->private) = ctx.v;
+
+	/* XXX: we don't need the following deferred processing */
+	xchg(&tg->limits_changed, true);
+	xchg(&td->limits_changed, true);
+	throtl_schedule_delayed_work(td, 0);
 
-	tg->iops[READ] = read_iops;
-	throtl_update_blkio_group_common(td, tg);
+	blkg_conf_finish(&ctx);
+	return 0;
 }
 
-static void throtl_update_blkio_group_write_iops(void *key,
-			struct blkio_group *blkg, unsigned int write_iops)
+static int tg_set_conf_u64(struct cgroup *cgrp, struct cftype *cft,
+			   const char *buf)
 {
-	struct throtl_data *td = key;
-	struct throtl_grp *tg = tg_of_blkg(blkg);
+	return tg_set_conf(cgrp, cft, buf, true);
+}
 
-	tg->iops[WRITE] = write_iops;
-	throtl_update_blkio_group_common(td, tg);
+static int tg_set_conf_uint(struct cgroup *cgrp, struct cftype *cft,
+			    const char *buf)
+{
+	return tg_set_conf(cgrp, cft, buf, false);
 }
 
+static struct cftype throtl_files[] = {
+	{
+		.name = "throttle.read_bps_device",
+		.private = offsetof(struct throtl_grp, bps[READ]),
+		.read_seq_string = tg_print_conf_u64,
+		.write_string = tg_set_conf_u64,
+		.max_write_len = 256,
+	},
+	{
+		.name = "throttle.write_bps_device",
+		.private = offsetof(struct throtl_grp, bps[WRITE]),
+		.read_seq_string = tg_print_conf_u64,
+		.write_string = tg_set_conf_u64,
+		.max_write_len = 256,
+	},
+	{
+		.name = "throttle.read_iops_device",
+		.private = offsetof(struct throtl_grp, iops[READ]),
+		.read_seq_string = tg_print_conf_uint,
+		.write_string = tg_set_conf_uint,
+		.max_write_len = 256,
+	},
+	{
+		.name = "throttle.write_iops_device",
+		.private = offsetof(struct throtl_grp, iops[WRITE]),
+		.read_seq_string = tg_print_conf_uint,
+		.write_string = tg_set_conf_uint,
+		.max_write_len = 256,
+	},
+	{
+		.name = "throttle.io_service_bytes",
+		.private = offsetof(struct tg_stats_cpu, service_bytes),
+		.read_seq_string = tg_print_cpu_rwstat,
+	},
+	{
+		.name = "throttle.io_serviced",
+		.private = offsetof(struct tg_stats_cpu, serviced),
+		.read_seq_string = tg_print_cpu_rwstat,
+	},
+	{ }	/* terminate */
+};
+
 static void throtl_shutdown_wq(struct request_queue *q)
 {
 	struct throtl_data *td = q->td;
@@ -1094,19 +1101,13 @@ static void throtl_shutdown_wq(struct request_queue *q)
 	cancel_delayed_work_sync(&td->throtl_work);
 }
 
-static struct blkio_policy_type blkio_policy_throtl = {
-	.ops = {
-		.blkio_unlink_group_fn = throtl_unlink_blkio_group,
-		.blkio_update_group_read_bps_fn =
-					throtl_update_blkio_group_read_bps,
-		.blkio_update_group_write_bps_fn =
-					throtl_update_blkio_group_write_bps,
-		.blkio_update_group_read_iops_fn =
-					throtl_update_blkio_group_read_iops,
-		.blkio_update_group_write_iops_fn =
-					throtl_update_blkio_group_write_iops,
-	},
-	.plid = BLKIO_POLICY_THROTL,
+static struct blkcg_policy blkcg_policy_throtl = {
+	.pd_size		= sizeof(struct throtl_grp),
+	.cftypes		= throtl_files,
+
+	.pd_init_fn		= throtl_pd_init,
+	.pd_exit_fn		= throtl_pd_exit,
+	.pd_reset_stats_fn	= throtl_pd_reset_stats,
 };
 
 bool blk_throtl_bio(struct request_queue *q, struct bio *bio)
@@ -1114,7 +1115,7 @@ bool blk_throtl_bio(struct request_queue *q, struct bio *bio)
 	struct throtl_data *td = q->td;
 	struct throtl_grp *tg;
 	bool rw = bio_data_dir(bio), update_disptime = true;
-	struct blkio_cgroup *blkcg;
+	struct blkcg *blkcg;
 	bool throttled = false;
 
 	if (bio->bi_rw & REQ_THROTTLED) {
@@ -1122,33 +1123,31 @@ bool blk_throtl_bio(struct request_queue *q, struct bio *bio)
 		goto out;
 	}
 
+	/* bio_associate_current() needs ioc, try creating */
+	create_io_context(GFP_ATOMIC, q->node);
+
 	/*
 	 * A throtl_grp pointer retrieved under rcu can be used to access
 	 * basic fields like stats and io rates. If a group has no rules,
 	 * just update the dispatch stats in lockless manner and return.
 	 */
-
 	rcu_read_lock();
-	blkcg = task_blkio_cgroup(current);
-	tg = throtl_find_tg(td, blkcg);
+	blkcg = bio_blkcg(bio);
+	tg = throtl_lookup_tg(td, blkcg);
 	if (tg) {
-		throtl_tg_fill_dev_details(td, tg);
-
 		if (tg_no_rule_group(tg, rw)) {
-			blkiocg_update_dispatch_stats(&tg->blkg, bio->bi_size,
-					rw, rw_is_sync(bio->bi_rw));
-			rcu_read_unlock();
-			goto out;
+			throtl_update_dispatch_stats(tg_to_blkg(tg),
+						     bio->bi_size, bio->bi_rw);
+			goto out_unlock_rcu;
 		}
 	}
-	rcu_read_unlock();
 
 	/*
 	 * Either group has not been allocated yet or it is not an unlimited
 	 * IO group
 	 */
 	spin_lock_irq(q->queue_lock);
-	tg = throtl_get_tg(td);
+	tg = throtl_lookup_create_tg(td, blkcg);
 	if (unlikely(!tg))
 		goto out_unlock;
 
@@ -1189,6 +1188,7 @@ queue_bio:
 			tg->io_disp[rw], tg->iops[rw],
 			tg->nr_queued[READ], tg->nr_queued[WRITE]);
 
+	bio_associate_current(bio);
 	throtl_add_bio_tg(q->td, tg, bio);
 	throttled = true;
 
@@ -1199,6 +1199,8 @@ queue_bio:
 
 out_unlock:
 	spin_unlock_irq(q->queue_lock);
+out_unlock_rcu:
+	rcu_read_unlock();
 out:
 	return throttled;
 }
@@ -1241,79 +1243,31 @@ void blk_throtl_drain(struct request_queue *q)
 int blk_throtl_init(struct request_queue *q)
 {
 	struct throtl_data *td;
-	struct throtl_grp *tg;
+	int ret;
 
 	td = kzalloc_node(sizeof(*td), GFP_KERNEL, q->node);
 	if (!td)
 		return -ENOMEM;
 
-	INIT_HLIST_HEAD(&td->tg_list);
 	td->tg_service_tree = THROTL_RB_ROOT;
 	td->limits_changed = false;
 	INIT_DELAYED_WORK(&td->throtl_work, blk_throtl_work);
 
-	/* alloc and Init root group. */
+	q->td = td;
 	td->queue = q;
-	tg = throtl_alloc_tg(td);
 
-	if (!tg) {
+	/* activate policy */
+	ret = blkcg_activate_policy(q, &blkcg_policy_throtl);
+	if (ret)
 		kfree(td);
-		return -ENOMEM;
-	}
-
-	td->root_tg = tg;
-
-	rcu_read_lock();
-	throtl_init_add_tg_lists(td, tg, &blkio_root_cgroup);
-	rcu_read_unlock();
-
-	/* Attach throtl data to request queue */
-	q->td = td;
-	return 0;
+	return ret;
 }
 
 void blk_throtl_exit(struct request_queue *q)
 {
-	struct throtl_data *td = q->td;
-	bool wait = false;
-
-	BUG_ON(!td);
-
-	throtl_shutdown_wq(q);
-
-	spin_lock_irq(q->queue_lock);
-	throtl_release_tgs(td);
-
-	/* If there are other groups */
-	if (td->nr_undestroyed_grps > 0)
-		wait = true;
-
-	spin_unlock_irq(q->queue_lock);
-
-	/*
-	 * Wait for tg->blkg->key accessors to exit their grace periods.
-	 * Do this wait only if there are other undestroyed groups out
-	 * there (other than root group). This can happen if cgroup deletion
-	 * path claimed the responsibility of cleaning up a group before
-	 * queue cleanup code get to the group.
-	 *
-	 * Do not call synchronize_rcu() unconditionally as there are drivers
-	 * which create/delete request queue hundreds of times during scan/boot
-	 * and synchronize_rcu() can take significant time and slow down boot.
-	 */
-	if (wait)
-		synchronize_rcu();
-
-	/*
-	 * Just being safe to make sure after previous flush if some body did
-	 * update limits through cgroup and another work got queued, cancel
-	 * it.
-	 */
+	BUG_ON(!q->td);
 	throtl_shutdown_wq(q);
-}
-
-void blk_throtl_release(struct request_queue *q)
-{
+	blkcg_deactivate_policy(q, &blkcg_policy_throtl);
 	kfree(q->td);
 }
 
@@ -1323,8 +1277,7 @@ static int __init throtl_init(void)
 	if (!kthrotld_workqueue)
 		panic("Failed to create kthrotld\n");
 
-	blkio_policy_register(&blkio_policy_throtl);
-	return 0;
+	return blkcg_policy_register(&blkcg_policy_throtl);
 }
 
 module_init(throtl_init);
diff --git a/block/blk.h b/block/blk.h
index d45be871329e..85f6ae42f7d3 100644
--- a/block/blk.h
+++ b/block/blk.h
@@ -23,7 +23,8 @@ void blk_rq_bio_prep(struct request_queue *q, struct request *rq,
 			struct bio *bio);
 int blk_rq_append_bio(struct request_queue *q, struct request *rq,
 		      struct bio *bio);
-void blk_drain_queue(struct request_queue *q, bool drain_all);
+void blk_queue_bypass_start(struct request_queue *q);
+void blk_queue_bypass_end(struct request_queue *q);
 void blk_dequeue_request(struct request *rq);
 void __blk_queue_free_tags(struct request_queue *q);
 bool __blk_end_bidi_request(struct request *rq, int error,
@@ -144,9 +145,6 @@ void blk_queue_congestion_threshold(struct request_queue *q);
 
 int blk_dev_init(void);
 
-void elv_quiesce_start(struct request_queue *q);
-void elv_quiesce_end(struct request_queue *q);
-
 
 /*
  * Return the threshold (number of used requests) at which the queue is
@@ -186,32 +184,30 @@ static inline int blk_do_io_stat(struct request *rq)
  */
 void get_io_context(struct io_context *ioc);
 struct io_cq *ioc_lookup_icq(struct io_context *ioc, struct request_queue *q);
-struct io_cq *ioc_create_icq(struct request_queue *q, gfp_t gfp_mask);
+struct io_cq *ioc_create_icq(struct io_context *ioc, struct request_queue *q,
+			     gfp_t gfp_mask);
 void ioc_clear_queue(struct request_queue *q);
 
-void create_io_context_slowpath(struct task_struct *task, gfp_t gfp_mask,
-				int node);
+int create_task_io_context(struct task_struct *task, gfp_t gfp_mask, int node);
 
 /**
  * create_io_context - try to create task->io_context
- * @task: target task
  * @gfp_mask: allocation mask
  * @node: allocation node
  *
- * If @task->io_context is %NULL, allocate a new io_context and install it.
- * Returns the current @task->io_context which may be %NULL if allocation
- * failed.
+ * If %current->io_context is %NULL, allocate a new io_context and install
+ * it.  Returns the current %current->io_context which may be %NULL if
+ * allocation failed.
  *
  * Note that this function can't be called with IRQ disabled because
- * task_lock which protects @task->io_context is IRQ-unsafe.
+ * task_lock which protects %current->io_context is IRQ-unsafe.
  */
-static inline struct io_context *create_io_context(struct task_struct *task,
-						   gfp_t gfp_mask, int node)
+static inline struct io_context *create_io_context(gfp_t gfp_mask, int node)
 {
 	WARN_ON_ONCE(irqs_disabled());
-	if (unlikely(!task->io_context))
-		create_io_context_slowpath(task, gfp_mask, node);
-	return task->io_context;
+	if (unlikely(!current->io_context))
+		create_task_io_context(current, gfp_mask, node);
+	return current->io_context;
 }
 
 /*
@@ -222,7 +218,6 @@ extern bool blk_throtl_bio(struct request_queue *q, struct bio *bio);
 extern void blk_throtl_drain(struct request_queue *q);
 extern int blk_throtl_init(struct request_queue *q);
 extern void blk_throtl_exit(struct request_queue *q);
-extern void blk_throtl_release(struct request_queue *q);
 #else /* CONFIG_BLK_DEV_THROTTLING */
 static inline bool blk_throtl_bio(struct request_queue *q, struct bio *bio)
 {
@@ -231,7 +226,6 @@ static inline bool blk_throtl_bio(struct request_queue *q, struct bio *bio)
 static inline void blk_throtl_drain(struct request_queue *q) { }
 static inline int blk_throtl_init(struct request_queue *q) { return 0; }
 static inline void blk_throtl_exit(struct request_queue *q) { }
-static inline void blk_throtl_release(struct request_queue *q) { }
 #endif /* CONFIG_BLK_DEV_THROTTLING */
 
 #endif /* BLK_INTERNAL_H */
diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
index 3c38536bd52c..673c977cc2bf 100644
--- a/block/cfq-iosched.c
+++ b/block/cfq-iosched.c
@@ -15,7 +15,9 @@
 #include <linux/ioprio.h>
 #include <linux/blktrace_api.h>
 #include "blk.h"
-#include "cfq.h"
+#include "blk-cgroup.h"
+
+static struct blkcg_policy blkcg_policy_cfq __maybe_unused;
 
 /*
  * tunables
@@ -171,8 +173,53 @@ enum wl_type_t {
 	SYNC_WORKLOAD = 2
 };
 
+struct cfqg_stats {
+#ifdef CONFIG_CFQ_GROUP_IOSCHED
+	/* total bytes transferred */
+	struct blkg_rwstat		service_bytes;
+	/* total IOs serviced, post merge */
+	struct blkg_rwstat		serviced;
+	/* number of ios merged */
+	struct blkg_rwstat		merged;
+	/* total time spent on device in ns, may not be accurate w/ queueing */
+	struct blkg_rwstat		service_time;
+	/* total time spent waiting in scheduler queue in ns */
+	struct blkg_rwstat		wait_time;
+	/* number of IOs queued up */
+	struct blkg_rwstat		queued;
+	/* total sectors transferred */
+	struct blkg_stat		sectors;
+	/* total disk time and nr sectors dispatched by this group */
+	struct blkg_stat		time;
+#ifdef CONFIG_DEBUG_BLK_CGROUP
+	/* time not charged to this cgroup */
+	struct blkg_stat		unaccounted_time;
+	/* sum of number of ios queued across all samples */
+	struct blkg_stat		avg_queue_size_sum;
+	/* count of samples taken for average */
+	struct blkg_stat		avg_queue_size_samples;
+	/* how many times this group has been removed from service tree */
+	struct blkg_stat		dequeue;
+	/* total time spent waiting for it to be assigned a timeslice. */
+	struct blkg_stat		group_wait_time;
+	/* time spent idling for this blkcg_gq */
+	struct blkg_stat		idle_time;
+	/* total time with empty current active q with other requests queued */
+	struct blkg_stat		empty_time;
+	/* fields after this shouldn't be cleared on stat reset */
+	uint64_t			start_group_wait_time;
+	uint64_t			start_idle_time;
+	uint64_t			start_empty_time;
+	uint16_t			flags;
+#endif	/* CONFIG_DEBUG_BLK_CGROUP */
+#endif	/* CONFIG_CFQ_GROUP_IOSCHED */
+};
+
 /* This is per cgroup per device grouping structure */
 struct cfq_group {
+	/* must be the first member */
+	struct blkg_policy_data pd;
+
 	/* group service_tree member */
 	struct rb_node rb_node;
 
@@ -180,7 +227,7 @@ struct cfq_group {
 	u64 vdisktime;
 	unsigned int weight;
 	unsigned int new_weight;
-	bool needs_update;
+	unsigned int dev_weight;
 
 	/* number of cfqq currently on this group */
 	int nr_cfqq;
@@ -206,20 +253,21 @@ struct cfq_group {
 	unsigned long saved_workload_slice;
 	enum wl_type_t saved_workload;
 	enum wl_prio_t saved_serving_prio;
-	struct blkio_group blkg;
-#ifdef CONFIG_CFQ_GROUP_IOSCHED
-	struct hlist_node cfqd_node;
-	int ref;
-#endif
+
 	/* number of requests that are on the dispatch list or inside driver */
 	int dispatched;
 	struct cfq_ttime ttime;
+	struct cfqg_stats stats;
 };
 
 struct cfq_io_cq {
 	struct io_cq		icq;		/* must be the first member */
 	struct cfq_queue	*cfqq[2];
 	struct cfq_ttime	ttime;
+	int			ioprio;		/* the current ioprio */
+#ifdef CONFIG_CFQ_GROUP_IOSCHED
+	uint64_t		blkcg_id;	/* the current blkcg ID */
+#endif
 };
 
 /*
@@ -229,7 +277,7 @@ struct cfq_data {
 	struct request_queue *queue;
 	/* Root service tree for cfq_groups */
 	struct cfq_rb_root grp_service_tree;
-	struct cfq_group root_group;
+	struct cfq_group *root_group;
 
 	/*
 	 * The priority currently being served
@@ -303,12 +351,6 @@ struct cfq_data {
 	struct cfq_queue oom_cfqq;
 
 	unsigned long last_delayed_sync;
-
-	/* List of cfq groups being managed on this device*/
-	struct hlist_head cfqg_list;
-
-	/* Number of groups which are on blkcg->blkg_list */
-	unsigned int nr_blkcg_linked_grps;
 };
 
 static struct cfq_group *cfq_get_next_cfqg(struct cfq_data *cfqd);
@@ -371,21 +413,284 @@ CFQ_CFQQ_FNS(deep);
 CFQ_CFQQ_FNS(wait_busy);
 #undef CFQ_CFQQ_FNS
 
+static inline struct cfq_group *pd_to_cfqg(struct blkg_policy_data *pd)
+{
+	return pd ? container_of(pd, struct cfq_group, pd) : NULL;
+}
+
+static inline struct cfq_group *blkg_to_cfqg(struct blkcg_gq *blkg)
+{
+	return pd_to_cfqg(blkg_to_pd(blkg, &blkcg_policy_cfq));
+}
+
+static inline struct blkcg_gq *cfqg_to_blkg(struct cfq_group *cfqg)
+{
+	return pd_to_blkg(&cfqg->pd);
+}
+
+#if defined(CONFIG_CFQ_GROUP_IOSCHED) && defined(CONFIG_DEBUG_BLK_CGROUP)
+
+/* cfqg stats flags */
+enum cfqg_stats_flags {
+	CFQG_stats_waiting = 0,
+	CFQG_stats_idling,
+	CFQG_stats_empty,
+};
+
+#define CFQG_FLAG_FNS(name)						\
+static inline void cfqg_stats_mark_##name(struct cfqg_stats *stats)	\
+{									\
+	stats->flags |= (1 << CFQG_stats_##name);			\
+}									\
+static inline void cfqg_stats_clear_##name(struct cfqg_stats *stats)	\
+{									\
+	stats->flags &= ~(1 << CFQG_stats_##name);			\
+}									\
+static inline int cfqg_stats_##name(struct cfqg_stats *stats)		\
+{									\
+	return (stats->flags & (1 << CFQG_stats_##name)) != 0;		\
+}									\
+
+CFQG_FLAG_FNS(waiting)
+CFQG_FLAG_FNS(idling)
+CFQG_FLAG_FNS(empty)
+#undef CFQG_FLAG_FNS
+
+/* This should be called with the queue_lock held. */
+static void cfqg_stats_update_group_wait_time(struct cfqg_stats *stats)
+{
+	unsigned long long now;
+
+	if (!cfqg_stats_waiting(stats))
+		return;
+
+	now = sched_clock();
+	if (time_after64(now, stats->start_group_wait_time))
+		blkg_stat_add(&stats->group_wait_time,
+			      now - stats->start_group_wait_time);
+	cfqg_stats_clear_waiting(stats);
+}
+
+/* This should be called with the queue_lock held. */
+static void cfqg_stats_set_start_group_wait_time(struct cfq_group *cfqg,
+						 struct cfq_group *curr_cfqg)
+{
+	struct cfqg_stats *stats = &cfqg->stats;
+
+	if (cfqg_stats_waiting(stats))
+		return;
+	if (cfqg == curr_cfqg)
+		return;
+	stats->start_group_wait_time = sched_clock();
+	cfqg_stats_mark_waiting(stats);
+}
+
+/* This should be called with the queue_lock held. */
+static void cfqg_stats_end_empty_time(struct cfqg_stats *stats)
+{
+	unsigned long long now;
+
+	if (!cfqg_stats_empty(stats))
+		return;
+
+	now = sched_clock();
+	if (time_after64(now, stats->start_empty_time))
+		blkg_stat_add(&stats->empty_time,
+			      now - stats->start_empty_time);
+	cfqg_stats_clear_empty(stats);
+}
+
+static void cfqg_stats_update_dequeue(struct cfq_group *cfqg)
+{
+	blkg_stat_add(&cfqg->stats.dequeue, 1);
+}
+
+static void cfqg_stats_set_start_empty_time(struct cfq_group *cfqg)
+{
+	struct cfqg_stats *stats = &cfqg->stats;
+
+	if (blkg_rwstat_sum(&stats->queued))
+		return;
+
+	/*
+	 * group is already marked empty. This can happen if cfqq got new
+	 * request in parent group and moved to this group while being added
+	 * to service tree. Just ignore the event and move on.
+	 */
+	if (cfqg_stats_empty(stats))
+		return;
+
+	stats->start_empty_time = sched_clock();
+	cfqg_stats_mark_empty(stats);
+}
+
+static void cfqg_stats_update_idle_time(struct cfq_group *cfqg)
+{
+	struct cfqg_stats *stats = &cfqg->stats;
+
+	if (cfqg_stats_idling(stats)) {
+		unsigned long long now = sched_clock();
+
+		if (time_after64(now, stats->start_idle_time))
+			blkg_stat_add(&stats->idle_time,
+				      now - stats->start_idle_time);
+		cfqg_stats_clear_idling(stats);
+	}
+}
+
+static void cfqg_stats_set_start_idle_time(struct cfq_group *cfqg)
+{
+	struct cfqg_stats *stats = &cfqg->stats;
+
+	BUG_ON(cfqg_stats_idling(stats));
+
+	stats->start_idle_time = sched_clock();
+	cfqg_stats_mark_idling(stats);
+}
+
+static void cfqg_stats_update_avg_queue_size(struct cfq_group *cfqg)
+{
+	struct cfqg_stats *stats = &cfqg->stats;
+
+	blkg_stat_add(&stats->avg_queue_size_sum,
+		      blkg_rwstat_sum(&stats->queued));
+	blkg_stat_add(&stats->avg_queue_size_samples, 1);
+	cfqg_stats_update_group_wait_time(stats);
+}
+
+#else	/* CONFIG_CFQ_GROUP_IOSCHED && CONFIG_DEBUG_BLK_CGROUP */
+
+static inline void cfqg_stats_set_start_group_wait_time(struct cfq_group *cfqg, struct cfq_group *curr_cfqg) { }
+static inline void cfqg_stats_end_empty_time(struct cfqg_stats *stats) { }
+static inline void cfqg_stats_update_dequeue(struct cfq_group *cfqg) { }
+static inline void cfqg_stats_set_start_empty_time(struct cfq_group *cfqg) { }
+static inline void cfqg_stats_update_idle_time(struct cfq_group *cfqg) { }
+static inline void cfqg_stats_set_start_idle_time(struct cfq_group *cfqg) { }
+static inline void cfqg_stats_update_avg_queue_size(struct cfq_group *cfqg) { }
+
+#endif	/* CONFIG_CFQ_GROUP_IOSCHED && CONFIG_DEBUG_BLK_CGROUP */
+
 #ifdef CONFIG_CFQ_GROUP_IOSCHED
-#define cfq_log_cfqq(cfqd, cfqq, fmt, args...)	\
+
+static inline void cfqg_get(struct cfq_group *cfqg)
+{
+	return blkg_get(cfqg_to_blkg(cfqg));
+}
+
+static inline void cfqg_put(struct cfq_group *cfqg)
+{
+	return blkg_put(cfqg_to_blkg(cfqg));
+}
+
+#define cfq_log_cfqq(cfqd, cfqq, fmt, args...)	do {			\
+	char __pbuf[128];						\
+									\
+	blkg_path(cfqg_to_blkg((cfqq)->cfqg), __pbuf, sizeof(__pbuf));	\
 	blk_add_trace_msg((cfqd)->queue, "cfq%d%c %s " fmt, (cfqq)->pid, \
-			cfq_cfqq_sync((cfqq)) ? 'S' : 'A', \
-			blkg_path(&(cfqq)->cfqg->blkg), ##args)
+			  cfq_cfqq_sync((cfqq)) ? 'S' : 'A',		\
+			  __pbuf, ##args);				\
+} while (0)
 
-#define cfq_log_cfqg(cfqd, cfqg, fmt, args...)				\
-	blk_add_trace_msg((cfqd)->queue, "%s " fmt,			\
-				blkg_path(&(cfqg)->blkg), ##args)       \
+#define cfq_log_cfqg(cfqd, cfqg, fmt, args...)	do {			\
+	char __pbuf[128];						\
+									\
+	blkg_path(cfqg_to_blkg(cfqg), __pbuf, sizeof(__pbuf));		\
+	blk_add_trace_msg((cfqd)->queue, "%s " fmt, __pbuf, ##args);	\
+} while (0)
+
+static inline void cfqg_stats_update_io_add(struct cfq_group *cfqg,
+					    struct cfq_group *curr_cfqg, int rw)
+{
+	blkg_rwstat_add(&cfqg->stats.queued, rw, 1);
+	cfqg_stats_end_empty_time(&cfqg->stats);
+	cfqg_stats_set_start_group_wait_time(cfqg, curr_cfqg);
+}
+
+static inline void cfqg_stats_update_timeslice_used(struct cfq_group *cfqg,
+			unsigned long time, unsigned long unaccounted_time)
+{
+	blkg_stat_add(&cfqg->stats.time, time);
+#ifdef CONFIG_DEBUG_BLK_CGROUP
+	blkg_stat_add(&cfqg->stats.unaccounted_time, unaccounted_time);
+#endif
+}
+
+static inline void cfqg_stats_update_io_remove(struct cfq_group *cfqg, int rw)
+{
+	blkg_rwstat_add(&cfqg->stats.queued, rw, -1);
+}
+
+static inline void cfqg_stats_update_io_merged(struct cfq_group *cfqg, int rw)
+{
+	blkg_rwstat_add(&cfqg->stats.merged, rw, 1);
+}
+
+static inline void cfqg_stats_update_dispatch(struct cfq_group *cfqg,
+					      uint64_t bytes, int rw)
+{
+	blkg_stat_add(&cfqg->stats.sectors, bytes >> 9);
+	blkg_rwstat_add(&cfqg->stats.serviced, rw, 1);
+	blkg_rwstat_add(&cfqg->stats.service_bytes, rw, bytes);
+}
+
+static inline void cfqg_stats_update_completion(struct cfq_group *cfqg,
+			uint64_t start_time, uint64_t io_start_time, int rw)
+{
+	struct cfqg_stats *stats = &cfqg->stats;
+	unsigned long long now = sched_clock();
+
+	if (time_after64(now, io_start_time))
+		blkg_rwstat_add(&stats->service_time, rw, now - io_start_time);
+	if (time_after64(io_start_time, start_time))
+		blkg_rwstat_add(&stats->wait_time, rw,
+				io_start_time - start_time);
+}
+
+static void cfq_pd_reset_stats(struct blkcg_gq *blkg)
+{
+	struct cfq_group *cfqg = blkg_to_cfqg(blkg);
+	struct cfqg_stats *stats = &cfqg->stats;
+
+	/* queued stats shouldn't be cleared */
+	blkg_rwstat_reset(&stats->service_bytes);
+	blkg_rwstat_reset(&stats->serviced);
+	blkg_rwstat_reset(&stats->merged);
+	blkg_rwstat_reset(&stats->service_time);
+	blkg_rwstat_reset(&stats->wait_time);
+	blkg_stat_reset(&stats->time);
+#ifdef CONFIG_DEBUG_BLK_CGROUP
+	blkg_stat_reset(&stats->unaccounted_time);
+	blkg_stat_reset(&stats->avg_queue_size_sum);
+	blkg_stat_reset(&stats->avg_queue_size_samples);
+	blkg_stat_reset(&stats->dequeue);
+	blkg_stat_reset(&stats->group_wait_time);
+	blkg_stat_reset(&stats->idle_time);
+	blkg_stat_reset(&stats->empty_time);
+#endif
+}
+
+#else	/* CONFIG_CFQ_GROUP_IOSCHED */
+
+static inline void cfqg_get(struct cfq_group *cfqg) { }
+static inline void cfqg_put(struct cfq_group *cfqg) { }
 
-#else
 #define cfq_log_cfqq(cfqd, cfqq, fmt, args...)	\
 	blk_add_trace_msg((cfqd)->queue, "cfq%d " fmt, (cfqq)->pid, ##args)
 #define cfq_log_cfqg(cfqd, cfqg, fmt, args...)		do {} while (0)
-#endif
+
+static inline void cfqg_stats_update_io_add(struct cfq_group *cfqg,
+			struct cfq_group *curr_cfqg, int rw) { }
+static inline void cfqg_stats_update_timeslice_used(struct cfq_group *cfqg,
+			unsigned long time, unsigned long unaccounted_time) { }
+static inline void cfqg_stats_update_io_remove(struct cfq_group *cfqg, int rw) { }
+static inline void cfqg_stats_update_io_merged(struct cfq_group *cfqg, int rw) { }
+static inline void cfqg_stats_update_dispatch(struct cfq_group *cfqg,
+					      uint64_t bytes, int rw) { }
+static inline void cfqg_stats_update_completion(struct cfq_group *cfqg,
+			uint64_t start_time, uint64_t io_start_time, int rw) { }
+
+#endif	/* CONFIG_CFQ_GROUP_IOSCHED */
+
 #define cfq_log(cfqd, fmt, args...)	\
 	blk_add_trace_msg((cfqd)->queue, "cfq " fmt, ##args)
 
@@ -466,8 +771,9 @@ static inline int cfqg_busy_async_queues(struct cfq_data *cfqd,
 }
 
 static void cfq_dispatch_insert(struct request_queue *, struct request *);
-static struct cfq_queue *cfq_get_queue(struct cfq_data *, bool,
-				       struct io_context *, gfp_t);
+static struct cfq_queue *cfq_get_queue(struct cfq_data *cfqd, bool is_sync,
+				       struct cfq_io_cq *cic, struct bio *bio,
+				       gfp_t gfp_mask);
 
 static inline struct cfq_io_cq *icq_to_cic(struct io_cq *icq)
 {
@@ -545,7 +851,7 @@ static inline u64 cfq_scale_slice(unsigned long delta, struct cfq_group *cfqg)
 {
 	u64 d = delta << CFQ_SERVICE_SHIFT;
 
-	d = d * BLKIO_WEIGHT_DEFAULT;
+	d = d * CFQ_WEIGHT_DEFAULT;
 	do_div(d, cfqg->weight);
 	return d;
 }
@@ -872,9 +1178,9 @@ static void
 cfq_update_group_weight(struct cfq_group *cfqg)
 {
 	BUG_ON(!RB_EMPTY_NODE(&cfqg->rb_node));
-	if (cfqg->needs_update) {
+	if (cfqg->new_weight) {
 		cfqg->weight = cfqg->new_weight;
-		cfqg->needs_update = false;
+		cfqg->new_weight = 0;
 	}
 }
 
@@ -936,7 +1242,7 @@ cfq_group_notify_queue_del(struct cfq_data *cfqd, struct cfq_group *cfqg)
 	cfq_log_cfqg(cfqd, cfqg, "del_from_rr group");
 	cfq_group_service_tree_del(st, cfqg);
 	cfqg->saved_workload_slice = 0;
-	cfq_blkiocg_update_dequeue_stats(&cfqg->blkg, 1);
+	cfqg_stats_update_dequeue(cfqg);
 }
 
 static inline unsigned int cfq_cfqq_slice_usage(struct cfq_queue *cfqq,
@@ -1008,178 +1314,59 @@ static void cfq_group_served(struct cfq_data *cfqd, struct cfq_group *cfqg,
 		     "sl_used=%u disp=%u charge=%u iops=%u sect=%lu",
 		     used_sl, cfqq->slice_dispatch, charge,
 		     iops_mode(cfqd), cfqq->nr_sectors);
-	cfq_blkiocg_update_timeslice_used(&cfqg->blkg, used_sl,
-					  unaccounted_sl);
-	cfq_blkiocg_set_start_empty_time(&cfqg->blkg);
+	cfqg_stats_update_timeslice_used(cfqg, used_sl, unaccounted_sl);
+	cfqg_stats_set_start_empty_time(cfqg);
 }
 
-#ifdef CONFIG_CFQ_GROUP_IOSCHED
-static inline struct cfq_group *cfqg_of_blkg(struct blkio_group *blkg)
-{
-	if (blkg)
-		return container_of(blkg, struct cfq_group, blkg);
-	return NULL;
-}
-
-static void cfq_update_blkio_group_weight(void *key, struct blkio_group *blkg,
-					  unsigned int weight)
-{
-	struct cfq_group *cfqg = cfqg_of_blkg(blkg);
-	cfqg->new_weight = weight;
-	cfqg->needs_update = true;
-}
-
-static void cfq_init_add_cfqg_lists(struct cfq_data *cfqd,
-			struct cfq_group *cfqg, struct blkio_cgroup *blkcg)
-{
-	struct backing_dev_info *bdi = &cfqd->queue->backing_dev_info;
-	unsigned int major, minor;
-
-	/*
-	 * Add group onto cgroup list. It might happen that bdi->dev is
-	 * not initialized yet. Initialize this new group without major
-	 * and minor info and this info will be filled in once a new thread
-	 * comes for IO.
-	 */
-	if (bdi->dev) {
-		sscanf(dev_name(bdi->dev), "%u:%u", &major, &minor);
-		cfq_blkiocg_add_blkio_group(blkcg, &cfqg->blkg,
-					(void *)cfqd, MKDEV(major, minor));
-	} else
-		cfq_blkiocg_add_blkio_group(blkcg, &cfqg->blkg,
-					(void *)cfqd, 0);
-
-	cfqd->nr_blkcg_linked_grps++;
-	cfqg->weight = blkcg_get_weight(blkcg, cfqg->blkg.dev);
-
-	/* Add group on cfqd list */
-	hlist_add_head(&cfqg->cfqd_node, &cfqd->cfqg_list);
-}
-
-/*
- * Should be called from sleepable context. No request queue lock as per
- * cpu stats are allocated dynamically and alloc_percpu needs to be called
- * from sleepable context.
+/**
+ * cfq_init_cfqg_base - initialize base part of a cfq_group
+ * @cfqg: cfq_group to initialize
+ *
+ * Initialize the base part which is used whether %CONFIG_CFQ_GROUP_IOSCHED
+ * is enabled or not.
  */
-static struct cfq_group * cfq_alloc_cfqg(struct cfq_data *cfqd)
+static void cfq_init_cfqg_base(struct cfq_group *cfqg)
 {
-	struct cfq_group *cfqg = NULL;
-	int i, j, ret;
 	struct cfq_rb_root *st;
-
-	cfqg = kzalloc_node(sizeof(*cfqg), GFP_ATOMIC, cfqd->queue->node);
-	if (!cfqg)
-		return NULL;
+	int i, j;
 
 	for_each_cfqg_st(cfqg, i, j, st)
 		*st = CFQ_RB_ROOT;
 	RB_CLEAR_NODE(&cfqg->rb_node);
 
 	cfqg->ttime.last_end_request = jiffies;
-
-	/*
-	 * Take the initial reference that will be released on destroy
-	 * This can be thought of a joint reference by cgroup and
-	 * elevator which will be dropped by either elevator exit
-	 * or cgroup deletion path depending on who is exiting first.
-	 */
-	cfqg->ref = 1;
-
-	ret = blkio_alloc_blkg_stats(&cfqg->blkg);
-	if (ret) {
-		kfree(cfqg);
-		return NULL;
-	}
-
-	return cfqg;
 }
 
-static struct cfq_group *
-cfq_find_cfqg(struct cfq_data *cfqd, struct blkio_cgroup *blkcg)
+#ifdef CONFIG_CFQ_GROUP_IOSCHED
+static void cfq_pd_init(struct blkcg_gq *blkg)
 {
-	struct cfq_group *cfqg = NULL;
-	void *key = cfqd;
-	struct backing_dev_info *bdi = &cfqd->queue->backing_dev_info;
-	unsigned int major, minor;
-
-	/*
-	 * This is the common case when there are no blkio cgroups.
-	 * Avoid lookup in this case
-	 */
-	if (blkcg == &blkio_root_cgroup)
-		cfqg = &cfqd->root_group;
-	else
-		cfqg = cfqg_of_blkg(blkiocg_lookup_group(blkcg, key));
-
-	if (cfqg && !cfqg->blkg.dev && bdi->dev && dev_name(bdi->dev)) {
-		sscanf(dev_name(bdi->dev), "%u:%u", &major, &minor);
-		cfqg->blkg.dev = MKDEV(major, minor);
-	}
+	struct cfq_group *cfqg = blkg_to_cfqg(blkg);
 
-	return cfqg;
+	cfq_init_cfqg_base(cfqg);
+	cfqg->weight = blkg->blkcg->cfq_weight;
 }
 
 /*
  * Search for the cfq group current task belongs to. request_queue lock must
  * be held.
  */
-static struct cfq_group *cfq_get_cfqg(struct cfq_data *cfqd)
+static struct cfq_group *cfq_lookup_create_cfqg(struct cfq_data *cfqd,
+						struct blkcg *blkcg)
 {
-	struct blkio_cgroup *blkcg;
-	struct cfq_group *cfqg = NULL, *__cfqg = NULL;
 	struct request_queue *q = cfqd->queue;
+	struct cfq_group *cfqg = NULL;
 
-	rcu_read_lock();
-	blkcg = task_blkio_cgroup(current);
-	cfqg = cfq_find_cfqg(cfqd, blkcg);
-	if (cfqg) {
-		rcu_read_unlock();
-		return cfqg;
-	}
-
-	/*
-	 * Need to allocate a group. Allocation of group also needs allocation
-	 * of per cpu stats which in-turn takes a mutex() and can block. Hence
-	 * we need to drop rcu lock and queue_lock before we call alloc.
-	 *
-	 * Not taking any queue reference here and assuming that queue is
-	 * around by the time we return. CFQ queue allocation code does
-	 * the same. It might be racy though.
-	 */
-
-	rcu_read_unlock();
-	spin_unlock_irq(q->queue_lock);
-
-	cfqg = cfq_alloc_cfqg(cfqd);
-
-	spin_lock_irq(q->queue_lock);
-
-	rcu_read_lock();
-	blkcg = task_blkio_cgroup(current);
-
-	/*
-	 * If some other thread already allocated the group while we were
-	 * not holding queue lock, free up the group
-	 */
-	__cfqg = cfq_find_cfqg(cfqd, blkcg);
+	/* avoid lookup for the common case where there's no blkcg */
+	if (blkcg == &blkcg_root) {
+		cfqg = cfqd->root_group;
+	} else {
+		struct blkcg_gq *blkg;
 
-	if (__cfqg) {
-		kfree(cfqg);
-		rcu_read_unlock();
-		return __cfqg;
+		blkg = blkg_lookup_create(blkcg, q);
+		if (!IS_ERR(blkg))
+			cfqg = blkg_to_cfqg(blkg);
 	}
 
-	if (!cfqg)
-		cfqg = &cfqd->root_group;
-
-	cfq_init_add_cfqg_lists(cfqd, cfqg, blkcg);
-	rcu_read_unlock();
-	return cfqg;
-}
-
-static inline struct cfq_group *cfq_ref_get_cfqg(struct cfq_group *cfqg)
-{
-	cfqg->ref++;
 	return cfqg;
 }
 
@@ -1187,94 +1374,224 @@ static void cfq_link_cfqq_cfqg(struct cfq_queue *cfqq, struct cfq_group *cfqg)
 {
 	/* Currently, all async queues are mapped to root group */
 	if (!cfq_cfqq_sync(cfqq))
-		cfqg = &cfqq->cfqd->root_group;
+		cfqg = cfqq->cfqd->root_group;
 
 	cfqq->cfqg = cfqg;
 	/* cfqq reference on cfqg */
-	cfqq->cfqg->ref++;
+	cfqg_get(cfqg);
 }
 
-static void cfq_put_cfqg(struct cfq_group *cfqg)
+static u64 cfqg_prfill_weight_device(struct seq_file *sf,
+				     struct blkg_policy_data *pd, int off)
 {
-	struct cfq_rb_root *st;
-	int i, j;
+	struct cfq_group *cfqg = pd_to_cfqg(pd);
 
-	BUG_ON(cfqg->ref <= 0);
-	cfqg->ref--;
-	if (cfqg->ref)
-		return;
-	for_each_cfqg_st(cfqg, i, j, st)
-		BUG_ON(!RB_EMPTY_ROOT(&st->rb));
-	free_percpu(cfqg->blkg.stats_cpu);
-	kfree(cfqg);
+	if (!cfqg->dev_weight)
+		return 0;
+	return __blkg_prfill_u64(sf, pd, cfqg->dev_weight);
 }
 
-static void cfq_destroy_cfqg(struct cfq_data *cfqd, struct cfq_group *cfqg)
+static int cfqg_print_weight_device(struct cgroup *cgrp, struct cftype *cft,
+				    struct seq_file *sf)
 {
-	/* Something wrong if we are trying to remove same group twice */
-	BUG_ON(hlist_unhashed(&cfqg->cfqd_node));
+	blkcg_print_blkgs(sf, cgroup_to_blkcg(cgrp),
+			  cfqg_prfill_weight_device, &blkcg_policy_cfq, 0,
+			  false);
+	return 0;
+}
 
-	hlist_del_init(&cfqg->cfqd_node);
+static int cfq_print_weight(struct cgroup *cgrp, struct cftype *cft,
+			    struct seq_file *sf)
+{
+	seq_printf(sf, "%u\n", cgroup_to_blkcg(cgrp)->cfq_weight);
+	return 0;
+}
 
-	BUG_ON(cfqd->nr_blkcg_linked_grps <= 0);
-	cfqd->nr_blkcg_linked_grps--;
+static int cfqg_set_weight_device(struct cgroup *cgrp, struct cftype *cft,
+				  const char *buf)
+{
+	struct blkcg *blkcg = cgroup_to_blkcg(cgrp);
+	struct blkg_conf_ctx ctx;
+	struct cfq_group *cfqg;
+	int ret;
 
-	/*
-	 * Put the reference taken at the time of creation so that when all
-	 * queues are gone, group can be destroyed.
-	 */
-	cfq_put_cfqg(cfqg);
+	ret = blkg_conf_prep(blkcg, &blkcg_policy_cfq, buf, &ctx);
+	if (ret)
+		return ret;
+
+	ret = -EINVAL;
+	cfqg = blkg_to_cfqg(ctx.blkg);
+	if (!ctx.v || (ctx.v >= CFQ_WEIGHT_MIN && ctx.v <= CFQ_WEIGHT_MAX)) {
+		cfqg->dev_weight = ctx.v;
+		cfqg->new_weight = cfqg->dev_weight ?: blkcg->cfq_weight;
+		ret = 0;
+	}
+
+	blkg_conf_finish(&ctx);
+	return ret;
 }
 
-static void cfq_release_cfq_groups(struct cfq_data *cfqd)
+static int cfq_set_weight(struct cgroup *cgrp, struct cftype *cft, u64 val)
 {
-	struct hlist_node *pos, *n;
-	struct cfq_group *cfqg;
+	struct blkcg *blkcg = cgroup_to_blkcg(cgrp);
+	struct blkcg_gq *blkg;
+	struct hlist_node *n;
 
-	hlist_for_each_entry_safe(cfqg, pos, n, &cfqd->cfqg_list, cfqd_node) {
-		/*
-		 * If cgroup removal path got to blk_group first and removed
-		 * it from cgroup list, then it will take care of destroying
-		 * cfqg also.
-		 */
-		if (!cfq_blkiocg_del_blkio_group(&cfqg->blkg))
-			cfq_destroy_cfqg(cfqd, cfqg);
+	if (val < CFQ_WEIGHT_MIN || val > CFQ_WEIGHT_MAX)
+		return -EINVAL;
+
+	spin_lock_irq(&blkcg->lock);
+	blkcg->cfq_weight = (unsigned int)val;
+
+	hlist_for_each_entry(blkg, n, &blkcg->blkg_list, blkcg_node) {
+		struct cfq_group *cfqg = blkg_to_cfqg(blkg);
+
+		if (cfqg && !cfqg->dev_weight)
+			cfqg->new_weight = blkcg->cfq_weight;
 	}
+
+	spin_unlock_irq(&blkcg->lock);
+	return 0;
 }
 
-/*
- * Blk cgroup controller notification saying that blkio_group object is being
- * delinked as associated cgroup object is going away. That also means that
- * no new IO will come in this group. So get rid of this group as soon as
- * any pending IO in the group is finished.
- *
- * This function is called under rcu_read_lock(). key is the rcu protected
- * pointer. That means "key" is a valid cfq_data pointer as long as we are rcu
- * read lock.
- *
- * "key" was fetched from blkio_group under blkio_cgroup->lock. That means
- * it should not be NULL as even if elevator was exiting, cgroup deltion
- * path got to it first.
- */
-static void cfq_unlink_blkio_group(void *key, struct blkio_group *blkg)
+static int cfqg_print_stat(struct cgroup *cgrp, struct cftype *cft,
+			   struct seq_file *sf)
 {
-	unsigned long  flags;
-	struct cfq_data *cfqd = key;
+	struct blkcg *blkcg = cgroup_to_blkcg(cgrp);
 
-	spin_lock_irqsave(cfqd->queue->queue_lock, flags);
-	cfq_destroy_cfqg(cfqd, cfqg_of_blkg(blkg));
-	spin_unlock_irqrestore(cfqd->queue->queue_lock, flags);
+	blkcg_print_blkgs(sf, blkcg, blkg_prfill_stat, &blkcg_policy_cfq,
+			  cft->private, false);
+	return 0;
 }
 
-#else /* GROUP_IOSCHED */
-static struct cfq_group *cfq_get_cfqg(struct cfq_data *cfqd)
+static int cfqg_print_rwstat(struct cgroup *cgrp, struct cftype *cft,
+			     struct seq_file *sf)
 {
-	return &cfqd->root_group;
+	struct blkcg *blkcg = cgroup_to_blkcg(cgrp);
+
+	blkcg_print_blkgs(sf, blkcg, blkg_prfill_rwstat, &blkcg_policy_cfq,
+			  cft->private, true);
+	return 0;
 }
 
-static inline struct cfq_group *cfq_ref_get_cfqg(struct cfq_group *cfqg)
+#ifdef CONFIG_DEBUG_BLK_CGROUP
+static u64 cfqg_prfill_avg_queue_size(struct seq_file *sf,
+				      struct blkg_policy_data *pd, int off)
 {
-	return cfqg;
+	struct cfq_group *cfqg = pd_to_cfqg(pd);
+	u64 samples = blkg_stat_read(&cfqg->stats.avg_queue_size_samples);
+	u64 v = 0;
+
+	if (samples) {
+		v = blkg_stat_read(&cfqg->stats.avg_queue_size_sum);
+		do_div(v, samples);
+	}
+	__blkg_prfill_u64(sf, pd, v);
+	return 0;
+}
+
+/* print avg_queue_size */
+static int cfqg_print_avg_queue_size(struct cgroup *cgrp, struct cftype *cft,
+				     struct seq_file *sf)
+{
+	struct blkcg *blkcg = cgroup_to_blkcg(cgrp);
+
+	blkcg_print_blkgs(sf, blkcg, cfqg_prfill_avg_queue_size,
+			  &blkcg_policy_cfq, 0, false);
+	return 0;
+}
+#endif	/* CONFIG_DEBUG_BLK_CGROUP */
+
+static struct cftype cfq_blkcg_files[] = {
+	{
+		.name = "weight_device",
+		.read_seq_string = cfqg_print_weight_device,
+		.write_string = cfqg_set_weight_device,
+		.max_write_len = 256,
+	},
+	{
+		.name = "weight",
+		.read_seq_string = cfq_print_weight,
+		.write_u64 = cfq_set_weight,
+	},
+	{
+		.name = "time",
+		.private = offsetof(struct cfq_group, stats.time),
+		.read_seq_string = cfqg_print_stat,
+	},
+	{
+		.name = "sectors",
+		.private = offsetof(struct cfq_group, stats.sectors),
+		.read_seq_string = cfqg_print_stat,
+	},
+	{
+		.name = "io_service_bytes",
+		.private = offsetof(struct cfq_group, stats.service_bytes),
+		.read_seq_string = cfqg_print_rwstat,
+	},
+	{
+		.name = "io_serviced",
+		.private = offsetof(struct cfq_group, stats.serviced),
+		.read_seq_string = cfqg_print_rwstat,
+	},
+	{
+		.name = "io_service_time",
+		.private = offsetof(struct cfq_group, stats.service_time),
+		.read_seq_string = cfqg_print_rwstat,
+	},
+	{
+		.name = "io_wait_time",
+		.private = offsetof(struct cfq_group, stats.wait_time),
+		.read_seq_string = cfqg_print_rwstat,
+	},
+	{
+		.name = "io_merged",
+		.private = offsetof(struct cfq_group, stats.merged),
+		.read_seq_string = cfqg_print_rwstat,
+	},
+	{
+		.name = "io_queued",
+		.private = offsetof(struct cfq_group, stats.queued),
+		.read_seq_string = cfqg_print_rwstat,
+	},
+#ifdef CONFIG_DEBUG_BLK_CGROUP
+	{
+		.name = "avg_queue_size",
+		.read_seq_string = cfqg_print_avg_queue_size,
+	},
+	{
+		.name = "group_wait_time",
+		.private = offsetof(struct cfq_group, stats.group_wait_time),
+		.read_seq_string = cfqg_print_stat,
+	},
+	{
+		.name = "idle_time",
+		.private = offsetof(struct cfq_group, stats.idle_time),
+		.read_seq_string = cfqg_print_stat,
+	},
+	{
+		.name = "empty_time",
+		.private = offsetof(struct cfq_group, stats.empty_time),
+		.read_seq_string = cfqg_print_stat,
+	},
+	{
+		.name = "dequeue",
+		.private = offsetof(struct cfq_group, stats.dequeue),
+		.read_seq_string = cfqg_print_stat,
+	},
+	{
+		.name = "unaccounted_time",
+		.private = offsetof(struct cfq_group, stats.unaccounted_time),
+		.read_seq_string = cfqg_print_stat,
+	},
+#endif	/* CONFIG_DEBUG_BLK_CGROUP */
+	{ }	/* terminate */
+};
+#else /* GROUP_IOSCHED */
+static struct cfq_group *cfq_lookup_create_cfqg(struct cfq_data *cfqd,
+						struct blkcg *blkcg)
+{
+	return cfqd->root_group;
 }
 
 static inline void
@@ -1282,9 +1599,6 @@ cfq_link_cfqq_cfqg(struct cfq_queue *cfqq, struct cfq_group *cfqg) {
 	cfqq->cfqg = cfqg;
 }
 
-static void cfq_release_cfq_groups(struct cfq_data *cfqd) {}
-static inline void cfq_put_cfqg(struct cfq_group *cfqg) {}
-
 #endif /* GROUP_IOSCHED */
 
 /*
@@ -1551,12 +1865,10 @@ static void cfq_reposition_rq_rb(struct cfq_queue *cfqq, struct request *rq)
 {
 	elv_rb_del(&cfqq->sort_list, rq);
 	cfqq->queued[rq_is_sync(rq)]--;
-	cfq_blkiocg_update_io_remove_stats(&(RQ_CFQG(rq))->blkg,
-					rq_data_dir(rq), rq_is_sync(rq));
+	cfqg_stats_update_io_remove(RQ_CFQG(rq), rq->cmd_flags);
 	cfq_add_rq_rb(rq);
-	cfq_blkiocg_update_io_add_stats(&(RQ_CFQG(rq))->blkg,
-			&cfqq->cfqd->serving_group->blkg, rq_data_dir(rq),
-			rq_is_sync(rq));
+	cfqg_stats_update_io_add(RQ_CFQG(rq), cfqq->cfqd->serving_group,
+				 rq->cmd_flags);
 }
 
 static struct request *
@@ -1612,8 +1924,7 @@ static void cfq_remove_request(struct request *rq)
 	cfq_del_rq_rb(rq);
 
 	cfqq->cfqd->rq_queued--;
-	cfq_blkiocg_update_io_remove_stats(&(RQ_CFQG(rq))->blkg,
-					rq_data_dir(rq), rq_is_sync(rq));
+	cfqg_stats_update_io_remove(RQ_CFQG(rq), rq->cmd_flags);
 	if (rq->cmd_flags & REQ_PRIO) {
 		WARN_ON(!cfqq->prio_pending);
 		cfqq->prio_pending--;
@@ -1648,8 +1959,7 @@ static void cfq_merged_request(struct request_queue *q, struct request *req,
 static void cfq_bio_merged(struct request_queue *q, struct request *req,
 				struct bio *bio)
 {
-	cfq_blkiocg_update_io_merged_stats(&(RQ_CFQG(req))->blkg,
-					bio_data_dir(bio), cfq_bio_sync(bio));
+	cfqg_stats_update_io_merged(RQ_CFQG(req), bio->bi_rw);
 }
 
 static void
@@ -1671,8 +1981,7 @@ cfq_merged_requests(struct request_queue *q, struct request *rq,
 	if (cfqq->next_rq == next)
 		cfqq->next_rq = rq;
 	cfq_remove_request(next);
-	cfq_blkiocg_update_io_merged_stats(&(RQ_CFQG(rq))->blkg,
-					rq_data_dir(next), rq_is_sync(next));
+	cfqg_stats_update_io_merged(RQ_CFQG(rq), next->cmd_flags);
 
 	cfqq = RQ_CFQQ(next);
 	/*
@@ -1713,7 +2022,7 @@ static int cfq_allow_merge(struct request_queue *q, struct request *rq,
 static inline void cfq_del_timer(struct cfq_data *cfqd, struct cfq_queue *cfqq)
 {
 	del_timer(&cfqd->idle_slice_timer);
-	cfq_blkiocg_update_idle_time_stats(&cfqq->cfqg->blkg);
+	cfqg_stats_update_idle_time(cfqq->cfqg);
 }
 
 static void __cfq_set_active_queue(struct cfq_data *cfqd,
@@ -1722,7 +2031,7 @@ static void __cfq_set_active_queue(struct cfq_data *cfqd,
 	if (cfqq) {
 		cfq_log_cfqq(cfqd, cfqq, "set_active wl_prio:%d wl_type:%d",
 				cfqd->serving_prio, cfqd->serving_type);
-		cfq_blkiocg_update_avg_queue_size_stats(&cfqq->cfqg->blkg);
+		cfqg_stats_update_avg_queue_size(cfqq->cfqg);
 		cfqq->slice_start = 0;
 		cfqq->dispatch_start = jiffies;
 		cfqq->allocated_slice = 0;
@@ -2043,7 +2352,7 @@ static void cfq_arm_slice_timer(struct cfq_data *cfqd)
 	 * task has exited, don't wait
 	 */
 	cic = cfqd->active_cic;
-	if (!cic || !atomic_read(&cic->icq.ioc->nr_tasks))
+	if (!cic || !atomic_read(&cic->icq.ioc->active_ref))
 		return;
 
 	/*
@@ -2070,7 +2379,7 @@ static void cfq_arm_slice_timer(struct cfq_data *cfqd)
 		sl = cfqd->cfq_slice_idle;
 
 	mod_timer(&cfqd->idle_slice_timer, jiffies + sl);
-	cfq_blkiocg_update_set_idle_time_stats(&cfqq->cfqg->blkg);
+	cfqg_stats_set_start_idle_time(cfqq->cfqg);
 	cfq_log_cfqq(cfqd, cfqq, "arm_idle: %lu group_idle: %d", sl,
 			group_idle ? 1 : 0);
 }
@@ -2093,8 +2402,7 @@ static void cfq_dispatch_insert(struct request_queue *q, struct request *rq)
 
 	cfqd->rq_in_flight[cfq_cfqq_sync(cfqq)]++;
 	cfqq->nr_sectors += blk_rq_sectors(rq);
-	cfq_blkiocg_update_dispatch_stats(&cfqq->cfqg->blkg, blk_rq_bytes(rq),
-					rq_data_dir(rq), rq_is_sync(rq));
+	cfqg_stats_update_dispatch(cfqq->cfqg, blk_rq_bytes(rq), rq->cmd_flags);
 }
 
 /*
@@ -2677,7 +2985,7 @@ static void cfq_put_queue(struct cfq_queue *cfqq)
 
 	BUG_ON(cfq_cfqq_on_rr(cfqq));
 	kmem_cache_free(cfq_pool, cfqq);
-	cfq_put_cfqg(cfqg);
+	cfqg_put(cfqg);
 }
 
 static void cfq_put_cooperator(struct cfq_queue *cfqq)
@@ -2736,7 +3044,7 @@ static void cfq_exit_icq(struct io_cq *icq)
 	}
 }
 
-static void cfq_init_prio_data(struct cfq_queue *cfqq, struct io_context *ioc)
+static void cfq_init_prio_data(struct cfq_queue *cfqq, struct cfq_io_cq *cic)
 {
 	struct task_struct *tsk = current;
 	int ioprio_class;
@@ -2744,7 +3052,7 @@ static void cfq_init_prio_data(struct cfq_queue *cfqq, struct io_context *ioc)
 	if (!cfq_cfqq_prio_changed(cfqq))
 		return;
 
-	ioprio_class = IOPRIO_PRIO_CLASS(ioc->ioprio);
+	ioprio_class = IOPRIO_PRIO_CLASS(cic->ioprio);
 	switch (ioprio_class) {
 	default:
 		printk(KERN_ERR "cfq: bad prio %x\n", ioprio_class);
@@ -2756,11 +3064,11 @@ static void cfq_init_prio_data(struct cfq_queue *cfqq, struct io_context *ioc)
 		cfqq->ioprio_class = task_nice_ioclass(tsk);
 		break;
 	case IOPRIO_CLASS_RT:
-		cfqq->ioprio = task_ioprio(ioc);
+		cfqq->ioprio = IOPRIO_PRIO_DATA(cic->ioprio);
 		cfqq->ioprio_class = IOPRIO_CLASS_RT;
 		break;
 	case IOPRIO_CLASS_BE:
-		cfqq->ioprio = task_ioprio(ioc);
+		cfqq->ioprio = IOPRIO_PRIO_DATA(cic->ioprio);
 		cfqq->ioprio_class = IOPRIO_CLASS_BE;
 		break;
 	case IOPRIO_CLASS_IDLE:
@@ -2778,19 +3086,24 @@ static void cfq_init_prio_data(struct cfq_queue *cfqq, struct io_context *ioc)
 	cfq_clear_cfqq_prio_changed(cfqq);
 }
 
-static void changed_ioprio(struct cfq_io_cq *cic)
+static void check_ioprio_changed(struct cfq_io_cq *cic, struct bio *bio)
 {
+	int ioprio = cic->icq.ioc->ioprio;
 	struct cfq_data *cfqd = cic_to_cfqd(cic);
 	struct cfq_queue *cfqq;
 
-	if (unlikely(!cfqd))
+	/*
+	 * Check whether ioprio has changed.  The condition may trigger
+	 * spuriously on a newly created cic but there's no harm.
+	 */
+	if (unlikely(!cfqd) || likely(cic->ioprio == ioprio))
 		return;
 
 	cfqq = cic->cfqq[BLK_RW_ASYNC];
 	if (cfqq) {
 		struct cfq_queue *new_cfqq;
-		new_cfqq = cfq_get_queue(cfqd, BLK_RW_ASYNC, cic->icq.ioc,
-						GFP_ATOMIC);
+		new_cfqq = cfq_get_queue(cfqd, BLK_RW_ASYNC, cic, bio,
+					 GFP_ATOMIC);
 		if (new_cfqq) {
 			cic->cfqq[BLK_RW_ASYNC] = new_cfqq;
 			cfq_put_queue(cfqq);
@@ -2800,6 +3113,8 @@ static void changed_ioprio(struct cfq_io_cq *cic)
 	cfqq = cic->cfqq[BLK_RW_SYNC];
 	if (cfqq)
 		cfq_mark_cfqq_prio_changed(cfqq);
+
+	cic->ioprio = ioprio;
 }
 
 static void cfq_init_cfqq(struct cfq_data *cfqd, struct cfq_queue *cfqq,
@@ -2823,17 +3138,24 @@ static void cfq_init_cfqq(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 }
 
 #ifdef CONFIG_CFQ_GROUP_IOSCHED
-static void changed_cgroup(struct cfq_io_cq *cic)
+static void check_blkcg_changed(struct cfq_io_cq *cic, struct bio *bio)
 {
-	struct cfq_queue *sync_cfqq = cic_to_cfqq(cic, 1);
 	struct cfq_data *cfqd = cic_to_cfqd(cic);
-	struct request_queue *q;
+	struct cfq_queue *sync_cfqq;
+	uint64_t id;
 
-	if (unlikely(!cfqd))
-		return;
+	rcu_read_lock();
+	id = bio_blkcg(bio)->id;
+	rcu_read_unlock();
 
-	q = cfqd->queue;
+	/*
+	 * Check whether blkcg has changed.  The condition may trigger
+	 * spuriously on a newly created cic but there's no harm.
+	 */
+	if (unlikely(!cfqd) || likely(cic->blkcg_id == id))
+		return;
 
+	sync_cfqq = cic_to_cfqq(cic, 1);
 	if (sync_cfqq) {
 		/*
 		 * Drop reference to sync queue. A new sync queue will be
@@ -2843,21 +3165,26 @@ static void changed_cgroup(struct cfq_io_cq *cic)
 		cic_set_cfqq(cic, NULL, 1);
 		cfq_put_queue(sync_cfqq);
 	}
+
+	cic->blkcg_id = id;
 }
+#else
+static inline void check_blkcg_changed(struct cfq_io_cq *cic, struct bio *bio) { }
 #endif  /* CONFIG_CFQ_GROUP_IOSCHED */
 
 static struct cfq_queue *
-cfq_find_alloc_queue(struct cfq_data *cfqd, bool is_sync,
-		     struct io_context *ioc, gfp_t gfp_mask)
+cfq_find_alloc_queue(struct cfq_data *cfqd, bool is_sync, struct cfq_io_cq *cic,
+		     struct bio *bio, gfp_t gfp_mask)
 {
+	struct blkcg *blkcg;
 	struct cfq_queue *cfqq, *new_cfqq = NULL;
-	struct cfq_io_cq *cic;
 	struct cfq_group *cfqg;
 
 retry:
-	cfqg = cfq_get_cfqg(cfqd);
-	cic = cfq_cic_lookup(cfqd, ioc);
-	/* cic always exists here */
+	rcu_read_lock();
+
+	blkcg = bio_blkcg(bio);
+	cfqg = cfq_lookup_create_cfqg(cfqd, blkcg);
 	cfqq = cic_to_cfqq(cic, is_sync);
 
 	/*
@@ -2870,6 +3197,7 @@ retry:
 			cfqq = new_cfqq;
 			new_cfqq = NULL;
 		} else if (gfp_mask & __GFP_WAIT) {
+			rcu_read_unlock();
 			spin_unlock_irq(cfqd->queue->queue_lock);
 			new_cfqq = kmem_cache_alloc_node(cfq_pool,
 					gfp_mask | __GFP_ZERO,
@@ -2885,7 +3213,7 @@ retry:
 
 		if (cfqq) {
 			cfq_init_cfqq(cfqd, cfqq, current->pid, is_sync);
-			cfq_init_prio_data(cfqq, ioc);
+			cfq_init_prio_data(cfqq, cic);
 			cfq_link_cfqq_cfqg(cfqq, cfqg);
 			cfq_log_cfqq(cfqd, cfqq, "alloced");
 		} else
@@ -2895,6 +3223,7 @@ retry:
 	if (new_cfqq)
 		kmem_cache_free(cfq_pool, new_cfqq);
 
+	rcu_read_unlock();
 	return cfqq;
 }
 
@@ -2904,6 +3233,9 @@ cfq_async_queue_prio(struct cfq_data *cfqd, int ioprio_class, int ioprio)
 	switch (ioprio_class) {
 	case IOPRIO_CLASS_RT:
 		return &cfqd->async_cfqq[0][ioprio];
+	case IOPRIO_CLASS_NONE:
+		ioprio = IOPRIO_NORM;
+		/* fall through */
 	case IOPRIO_CLASS_BE:
 		return &cfqd->async_cfqq[1][ioprio];
 	case IOPRIO_CLASS_IDLE:
@@ -2914,11 +3246,11 @@ cfq_async_queue_prio(struct cfq_data *cfqd, int ioprio_class, int ioprio)
 }
 
 static struct cfq_queue *
-cfq_get_queue(struct cfq_data *cfqd, bool is_sync, struct io_context *ioc,
-	      gfp_t gfp_mask)
+cfq_get_queue(struct cfq_data *cfqd, bool is_sync, struct cfq_io_cq *cic,
+	      struct bio *bio, gfp_t gfp_mask)
 {
-	const int ioprio = task_ioprio(ioc);
-	const int ioprio_class = task_ioprio_class(ioc);
+	const int ioprio_class = IOPRIO_PRIO_CLASS(cic->ioprio);
+	const int ioprio = IOPRIO_PRIO_DATA(cic->ioprio);
 	struct cfq_queue **async_cfqq = NULL;
 	struct cfq_queue *cfqq = NULL;
 
@@ -2928,7 +3260,7 @@ cfq_get_queue(struct cfq_data *cfqd, bool is_sync, struct io_context *ioc,
 	}
 
 	if (!cfqq)
-		cfqq = cfq_find_alloc_queue(cfqd, is_sync, ioc, gfp_mask);
+		cfqq = cfq_find_alloc_queue(cfqd, is_sync, cic, bio, gfp_mask);
 
 	/*
 	 * pin the queue now that it's allocated, scheduler exit will prune it
@@ -3010,7 +3342,7 @@ cfq_update_idle_window(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 
 	if (cfqq->next_rq && (cfqq->next_rq->cmd_flags & REQ_NOIDLE))
 		enable_idle = 0;
-	else if (!atomic_read(&cic->icq.ioc->nr_tasks) ||
+	else if (!atomic_read(&cic->icq.ioc->active_ref) ||
 		 !cfqd->cfq_slice_idle ||
 		 (!cfq_cfqq_deep(cfqq) && CFQQ_SEEKY(cfqq)))
 		enable_idle = 0;
@@ -3174,8 +3506,7 @@ cfq_rq_enqueued(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 				cfq_clear_cfqq_wait_request(cfqq);
 				__blk_run_queue(cfqd->queue);
 			} else {
-				cfq_blkiocg_update_idle_time_stats(
-						&cfqq->cfqg->blkg);
+				cfqg_stats_update_idle_time(cfqq->cfqg);
 				cfq_mark_cfqq_must_dispatch(cfqq);
 			}
 		}
@@ -3197,14 +3528,13 @@ static void cfq_insert_request(struct request_queue *q, struct request *rq)
 	struct cfq_queue *cfqq = RQ_CFQQ(rq);
 
 	cfq_log_cfqq(cfqd, cfqq, "insert_request");
-	cfq_init_prio_data(cfqq, RQ_CIC(rq)->icq.ioc);
+	cfq_init_prio_data(cfqq, RQ_CIC(rq));
 
 	rq_set_fifo_time(rq, jiffies + cfqd->cfq_fifo_expire[rq_is_sync(rq)]);
 	list_add_tail(&rq->queuelist, &cfqq->fifo);
 	cfq_add_rq_rb(rq);
-	cfq_blkiocg_update_io_add_stats(&(RQ_CFQG(rq))->blkg,
-			&cfqd->serving_group->blkg, rq_data_dir(rq),
-			rq_is_sync(rq));
+	cfqg_stats_update_io_add(RQ_CFQG(rq), cfqd->serving_group,
+				 rq->cmd_flags);
 	cfq_rq_enqueued(cfqd, cfqq, rq);
 }
 
@@ -3300,9 +3630,8 @@ static void cfq_completed_request(struct request_queue *q, struct request *rq)
 	cfqd->rq_in_driver--;
 	cfqq->dispatched--;
 	(RQ_CFQG(rq))->dispatched--;
-	cfq_blkiocg_update_completion_stats(&cfqq->cfqg->blkg,
-			rq_start_time_ns(rq), rq_io_start_time_ns(rq),
-			rq_data_dir(rq), rq_is_sync(rq));
+	cfqg_stats_update_completion(cfqq->cfqg, rq_start_time_ns(rq),
+				     rq_io_start_time_ns(rq), rq->cmd_flags);
 
 	cfqd->rq_in_flight[cfq_cfqq_sync(cfqq)]--;
 
@@ -3399,7 +3728,7 @@ static int cfq_may_queue(struct request_queue *q, int rw)
 
 	cfqq = cic_to_cfqq(cic, rw_is_sync(rw));
 	if (cfqq) {
-		cfq_init_prio_data(cfqq, cic->icq.ioc);
+		cfq_init_prio_data(cfqq, cic);
 
 		return __cfq_may_queue(cfqq);
 	}
@@ -3421,7 +3750,7 @@ static void cfq_put_request(struct request *rq)
 		cfqq->allocated[rw]--;
 
 		/* Put down rq reference on cfqg */
-		cfq_put_cfqg(RQ_CFQG(rq));
+		cfqg_put(RQ_CFQG(rq));
 		rq->elv.priv[0] = NULL;
 		rq->elv.priv[1] = NULL;
 
@@ -3465,32 +3794,25 @@ split_cfqq(struct cfq_io_cq *cic, struct cfq_queue *cfqq)
  * Allocate cfq data structures associated with this request.
  */
 static int
-cfq_set_request(struct request_queue *q, struct request *rq, gfp_t gfp_mask)
+cfq_set_request(struct request_queue *q, struct request *rq, struct bio *bio,
+		gfp_t gfp_mask)
 {
 	struct cfq_data *cfqd = q->elevator->elevator_data;
 	struct cfq_io_cq *cic = icq_to_cic(rq->elv.icq);
 	const int rw = rq_data_dir(rq);
 	const bool is_sync = rq_is_sync(rq);
 	struct cfq_queue *cfqq;
-	unsigned int changed;
 
 	might_sleep_if(gfp_mask & __GFP_WAIT);
 
 	spin_lock_irq(q->queue_lock);
 
-	/* handle changed notifications */
-	changed = icq_get_changed(&cic->icq);
-	if (unlikely(changed & ICQ_IOPRIO_CHANGED))
-		changed_ioprio(cic);
-#ifdef CONFIG_CFQ_GROUP_IOSCHED
-	if (unlikely(changed & ICQ_CGROUP_CHANGED))
-		changed_cgroup(cic);
-#endif
-
+	check_ioprio_changed(cic, bio);
+	check_blkcg_changed(cic, bio);
 new_queue:
 	cfqq = cic_to_cfqq(cic, is_sync);
 	if (!cfqq || cfqq == &cfqd->oom_cfqq) {
-		cfqq = cfq_get_queue(cfqd, is_sync, cic->icq.ioc, gfp_mask);
+		cfqq = cfq_get_queue(cfqd, is_sync, cic, bio, gfp_mask);
 		cic_set_cfqq(cic, cfqq, is_sync);
 	} else {
 		/*
@@ -3516,8 +3838,9 @@ new_queue:
 	cfqq->allocated[rw]++;
 
 	cfqq->ref++;
+	cfqg_get(cfqq->cfqg);
 	rq->elv.priv[0] = cfqq;
-	rq->elv.priv[1] = cfq_ref_get_cfqg(cfqq->cfqg);
+	rq->elv.priv[1] = cfqq->cfqg;
 	spin_unlock_irq(q->queue_lock);
 	return 0;
 }
@@ -3614,7 +3937,6 @@ static void cfq_exit_queue(struct elevator_queue *e)
 {
 	struct cfq_data *cfqd = e->elevator_data;
 	struct request_queue *q = cfqd->queue;
-	bool wait = false;
 
 	cfq_shutdown_timer_wq(cfqd);
 
@@ -3624,89 +3946,52 @@ static void cfq_exit_queue(struct elevator_queue *e)
 		__cfq_slice_expired(cfqd, cfqd->active_queue, 0);
 
 	cfq_put_async_queues(cfqd);
-	cfq_release_cfq_groups(cfqd);
-
-	/*
-	 * If there are groups which we could not unlink from blkcg list,
-	 * wait for a rcu period for them to be freed.
-	 */
-	if (cfqd->nr_blkcg_linked_grps)
-		wait = true;
 
 	spin_unlock_irq(q->queue_lock);
 
 	cfq_shutdown_timer_wq(cfqd);
 
-	/*
-	 * Wait for cfqg->blkg->key accessors to exit their grace periods.
-	 * Do this wait only if there are other unlinked groups out
-	 * there. This can happen if cgroup deletion path claimed the
-	 * responsibility of cleaning up a group before queue cleanup code
-	 * get to the group.
-	 *
-	 * Do not call synchronize_rcu() unconditionally as there are drivers
-	 * which create/delete request queue hundreds of times during scan/boot
-	 * and synchronize_rcu() can take significant time and slow down boot.
-	 */
-	if (wait)
-		synchronize_rcu();
-
-#ifdef CONFIG_CFQ_GROUP_IOSCHED
-	/* Free up per cpu stats for root group */
-	free_percpu(cfqd->root_group.blkg.stats_cpu);
+#ifndef CONFIG_CFQ_GROUP_IOSCHED
+	kfree(cfqd->root_group);
 #endif
+	blkcg_deactivate_policy(q, &blkcg_policy_cfq);
 	kfree(cfqd);
 }
 
-static void *cfq_init_queue(struct request_queue *q)
+static int cfq_init_queue(struct request_queue *q)
 {
 	struct cfq_data *cfqd;
-	int i, j;
-	struct cfq_group *cfqg;
-	struct cfq_rb_root *st;
+	struct blkcg_gq *blkg __maybe_unused;
+	int i, ret;
 
 	cfqd = kmalloc_node(sizeof(*cfqd), GFP_KERNEL | __GFP_ZERO, q->node);
 	if (!cfqd)
-		return NULL;
+		return -ENOMEM;
+
+	cfqd->queue = q;
+	q->elevator->elevator_data = cfqd;
 
 	/* Init root service tree */
 	cfqd->grp_service_tree = CFQ_RB_ROOT;
 
-	/* Init root group */
-	cfqg = &cfqd->root_group;
-	for_each_cfqg_st(cfqg, i, j, st)
-		*st = CFQ_RB_ROOT;
-	RB_CLEAR_NODE(&cfqg->rb_node);
-
-	/* Give preference to root group over other groups */
-	cfqg->weight = 2*BLKIO_WEIGHT_DEFAULT;
-
+	/* Init root group and prefer root group over other groups by default */
 #ifdef CONFIG_CFQ_GROUP_IOSCHED
-	/*
-	 * Set root group reference to 2. One reference will be dropped when
-	 * all groups on cfqd->cfqg_list are being deleted during queue exit.
-	 * Other reference will remain there as we don't want to delete this
-	 * group as it is statically allocated and gets destroyed when
-	 * throtl_data goes away.
-	 */
-	cfqg->ref = 2;
-
-	if (blkio_alloc_blkg_stats(&cfqg->blkg)) {
-		kfree(cfqg);
-		kfree(cfqd);
-		return NULL;
-	}
-
-	rcu_read_lock();
+	ret = blkcg_activate_policy(q, &blkcg_policy_cfq);
+	if (ret)
+		goto out_free;
 
-	cfq_blkiocg_add_blkio_group(&blkio_root_cgroup, &cfqg->blkg,
-					(void *)cfqd, 0);
-	rcu_read_unlock();
-	cfqd->nr_blkcg_linked_grps++;
+	cfqd->root_group = blkg_to_cfqg(q->root_blkg);
+#else
+	ret = -ENOMEM;
+	cfqd->root_group = kzalloc_node(sizeof(*cfqd->root_group),
+					GFP_KERNEL, cfqd->queue->node);
+	if (!cfqd->root_group)
+		goto out_free;
 
-	/* Add group on cfqd->cfqg_list */
-	hlist_add_head(&cfqg->cfqd_node, &cfqd->cfqg_list);
+	cfq_init_cfqg_base(cfqd->root_group);
 #endif
+	cfqd->root_group->weight = 2 * CFQ_WEIGHT_DEFAULT;
+
 	/*
 	 * Not strictly needed (since RB_ROOT just clears the node and we
 	 * zeroed cfqd on alloc), but better be safe in case someone decides
@@ -3718,13 +4003,17 @@ static void *cfq_init_queue(struct request_queue *q)
 	/*
 	 * Our fallback cfqq if cfq_find_alloc_queue() runs into OOM issues.
 	 * Grab a permanent reference to it, so that the normal code flow
-	 * will not attempt to free it.
+	 * will not attempt to free it.  oom_cfqq is linked to root_group
+	 * but shouldn't hold a reference as it'll never be unlinked.  Lose
+	 * the reference from linking right away.
 	 */
 	cfq_init_cfqq(cfqd, &cfqd->oom_cfqq, 1, 0);
 	cfqd->oom_cfqq.ref++;
-	cfq_link_cfqq_cfqg(&cfqd->oom_cfqq, &cfqd->root_group);
 
-	cfqd->queue = q;
+	spin_lock_irq(q->queue_lock);
+	cfq_link_cfqq_cfqg(&cfqd->oom_cfqq, cfqd->root_group);
+	cfqg_put(cfqd->root_group);
+	spin_unlock_irq(q->queue_lock);
 
 	init_timer(&cfqd->idle_slice_timer);
 	cfqd->idle_slice_timer.function = cfq_idle_slice_timer;
@@ -3750,7 +4039,11 @@ static void *cfq_init_queue(struct request_queue *q)
 	 * second, in order to have larger depth for async operations.
 	 */
 	cfqd->last_delayed_sync = jiffies - HZ;
-	return cfqd;
+	return 0;
+
+out_free:
+	kfree(cfqd);
+	return ret;
 }
 
 /*
@@ -3877,15 +4170,13 @@ static struct elevator_type iosched_cfq = {
 };
 
 #ifdef CONFIG_CFQ_GROUP_IOSCHED
-static struct blkio_policy_type blkio_policy_cfq = {
-	.ops = {
-		.blkio_unlink_group_fn =	cfq_unlink_blkio_group,
-		.blkio_update_group_weight_fn =	cfq_update_blkio_group_weight,
-	},
-	.plid = BLKIO_POLICY_PROP,
+static struct blkcg_policy blkcg_policy_cfq = {
+	.pd_size		= sizeof(struct cfq_group),
+	.cftypes		= cfq_blkcg_files,
+
+	.pd_init_fn		= cfq_pd_init,
+	.pd_reset_stats_fn	= cfq_pd_reset_stats,
 };
-#else
-static struct blkio_policy_type blkio_policy_cfq;
 #endif
 
 static int __init cfq_init(void)
@@ -3906,24 +4197,31 @@ static int __init cfq_init(void)
 #else
 		cfq_group_idle = 0;
 #endif
+
+	ret = blkcg_policy_register(&blkcg_policy_cfq);
+	if (ret)
+		return ret;
+
 	cfq_pool = KMEM_CACHE(cfq_queue, 0);
 	if (!cfq_pool)
-		return -ENOMEM;
+		goto err_pol_unreg;
 
 	ret = elv_register(&iosched_cfq);
-	if (ret) {
-		kmem_cache_destroy(cfq_pool);
-		return ret;
-	}
-
-	blkio_policy_register(&blkio_policy_cfq);
+	if (ret)
+		goto err_free_pool;
 
 	return 0;
+
+err_free_pool:
+	kmem_cache_destroy(cfq_pool);
+err_pol_unreg:
+	blkcg_policy_unregister(&blkcg_policy_cfq);
+	return ret;
 }
 
 static void __exit cfq_exit(void)
 {
-	blkio_policy_unregister(&blkio_policy_cfq);
+	blkcg_policy_unregister(&blkcg_policy_cfq);
 	elv_unregister(&iosched_cfq);
 	kmem_cache_destroy(cfq_pool);
 }
diff --git a/block/cfq.h b/block/cfq.h
deleted file mode 100644
index 2a155927e37c..000000000000
--- a/block/cfq.h
+++ /dev/null
@@ -1,115 +0,0 @@
-#ifndef _CFQ_H
-#define _CFQ_H
-#include "blk-cgroup.h"
-
-#ifdef CONFIG_CFQ_GROUP_IOSCHED
-static inline void cfq_blkiocg_update_io_add_stats(struct blkio_group *blkg,
-	struct blkio_group *curr_blkg, bool direction, bool sync)
-{
-	blkiocg_update_io_add_stats(blkg, curr_blkg, direction, sync);
-}
-
-static inline void cfq_blkiocg_update_dequeue_stats(struct blkio_group *blkg,
-			unsigned long dequeue)
-{
-	blkiocg_update_dequeue_stats(blkg, dequeue);
-}
-
-static inline void cfq_blkiocg_update_timeslice_used(struct blkio_group *blkg,
-			unsigned long time, unsigned long unaccounted_time)
-{
-	blkiocg_update_timeslice_used(blkg, time, unaccounted_time);
-}
-
-static inline void cfq_blkiocg_set_start_empty_time(struct blkio_group *blkg)
-{
-	blkiocg_set_start_empty_time(blkg);
-}
-
-static inline void cfq_blkiocg_update_io_remove_stats(struct blkio_group *blkg,
-				bool direction, bool sync)
-{
-	blkiocg_update_io_remove_stats(blkg, direction, sync);
-}
-
-static inline void cfq_blkiocg_update_io_merged_stats(struct blkio_group *blkg,
-		bool direction, bool sync)
-{
-	blkiocg_update_io_merged_stats(blkg, direction, sync);
-}
-
-static inline void cfq_blkiocg_update_idle_time_stats(struct blkio_group *blkg)
-{
-	blkiocg_update_idle_time_stats(blkg);
-}
-
-static inline void
-cfq_blkiocg_update_avg_queue_size_stats(struct blkio_group *blkg)
-{
-	blkiocg_update_avg_queue_size_stats(blkg);
-}
-
-static inline void
-cfq_blkiocg_update_set_idle_time_stats(struct blkio_group *blkg)
-{
-	blkiocg_update_set_idle_time_stats(blkg);
-}
-
-static inline void cfq_blkiocg_update_dispatch_stats(struct blkio_group *blkg,
-				uint64_t bytes, bool direction, bool sync)
-{
-	blkiocg_update_dispatch_stats(blkg, bytes, direction, sync);
-}
-
-static inline void cfq_blkiocg_update_completion_stats(struct blkio_group *blkg, uint64_t start_time, uint64_t io_start_time, bool direction, bool sync)
-{
-	blkiocg_update_completion_stats(blkg, start_time, io_start_time,
-				direction, sync);
-}
-
-static inline void cfq_blkiocg_add_blkio_group(struct blkio_cgroup *blkcg,
-			struct blkio_group *blkg, void *key, dev_t dev) {
-	blkiocg_add_blkio_group(blkcg, blkg, key, dev, BLKIO_POLICY_PROP);
-}
-
-static inline int cfq_blkiocg_del_blkio_group(struct blkio_group *blkg)
-{
-	return blkiocg_del_blkio_group(blkg);
-}
-
-#else /* CFQ_GROUP_IOSCHED */
-static inline void cfq_blkiocg_update_io_add_stats(struct blkio_group *blkg,
-	struct blkio_group *curr_blkg, bool direction, bool sync) {}
-
-static inline void cfq_blkiocg_update_dequeue_stats(struct blkio_group *blkg,
-			unsigned long dequeue) {}
-
-static inline void cfq_blkiocg_update_timeslice_used(struct blkio_group *blkg,
-			unsigned long time, unsigned long unaccounted_time) {}
-static inline void cfq_blkiocg_set_start_empty_time(struct blkio_group *blkg) {}
-static inline void cfq_blkiocg_update_io_remove_stats(struct blkio_group *blkg,
-				bool direction, bool sync) {}
-static inline void cfq_blkiocg_update_io_merged_stats(struct blkio_group *blkg,
-		bool direction, bool sync) {}
-static inline void cfq_blkiocg_update_idle_time_stats(struct blkio_group *blkg)
-{
-}
-static inline void
-cfq_blkiocg_update_avg_queue_size_stats(struct blkio_group *blkg) {}
-
-static inline void
-cfq_blkiocg_update_set_idle_time_stats(struct blkio_group *blkg) {}
-
-static inline void cfq_blkiocg_update_dispatch_stats(struct blkio_group *blkg,
-				uint64_t bytes, bool direction, bool sync) {}
-static inline void cfq_blkiocg_update_completion_stats(struct blkio_group *blkg, uint64_t start_time, uint64_t io_start_time, bool direction, bool sync) {}
-
-static inline void cfq_blkiocg_add_blkio_group(struct blkio_cgroup *blkcg,
-			struct blkio_group *blkg, void *key, dev_t dev) {}
-static inline int cfq_blkiocg_del_blkio_group(struct blkio_group *blkg)
-{
-	return 0;
-}
-
-#endif /* CFQ_GROUP_IOSCHED */
-#endif
diff --git a/block/deadline-iosched.c b/block/deadline-iosched.c
index 7bf12d793fcd..599b12e5380f 100644
--- a/block/deadline-iosched.c
+++ b/block/deadline-iosched.c
@@ -337,13 +337,13 @@ static void deadline_exit_queue(struct elevator_queue *e)
 /*
  * initialize elevator private data (deadline_data).
  */
-static void *deadline_init_queue(struct request_queue *q)
+static int deadline_init_queue(struct request_queue *q)
 {
 	struct deadline_data *dd;
 
 	dd = kmalloc_node(sizeof(*dd), GFP_KERNEL | __GFP_ZERO, q->node);
 	if (!dd)
-		return NULL;
+		return -ENOMEM;
 
 	INIT_LIST_HEAD(&dd->fifo_list[READ]);
 	INIT_LIST_HEAD(&dd->fifo_list[WRITE]);
@@ -354,7 +354,9 @@ static void *deadline_init_queue(struct request_queue *q)
 	dd->writes_starved = writes_starved;
 	dd->front_merges = 1;
 	dd->fifo_batch = fifo_batch;
-	return dd;
+
+	q->elevator->elevator_data = dd;
+	return 0;
 }
 
 /*
diff --git a/block/elevator.c b/block/elevator.c
index f016855a46b0..6a55d418896f 100644
--- a/block/elevator.c
+++ b/block/elevator.c
@@ -38,6 +38,7 @@
 #include <trace/events/block.h>
 
 #include "blk.h"
+#include "blk-cgroup.h"
 
 static DEFINE_SPINLOCK(elv_list_lock);
 static LIST_HEAD(elv_list);
@@ -121,15 +122,6 @@ static struct elevator_type *elevator_get(const char *name)
 	return e;
 }
 
-static int elevator_init_queue(struct request_queue *q,
-			       struct elevator_queue *eq)
-{
-	eq->elevator_data = eq->type->ops.elevator_init_fn(q);
-	if (eq->elevator_data)
-		return 0;
-	return -ENOMEM;
-}
-
 static char chosen_elevator[ELV_NAME_MAX];
 
 static int __init elevator_setup(char *str)
@@ -188,7 +180,6 @@ static void elevator_release(struct kobject *kobj)
 int elevator_init(struct request_queue *q, char *name)
 {
 	struct elevator_type *e = NULL;
-	struct elevator_queue *eq;
 	int err;
 
 	if (unlikely(q->elevator))
@@ -222,17 +213,16 @@ int elevator_init(struct request_queue *q, char *name)
 		}
 	}
 
-	eq = elevator_alloc(q, e);
-	if (!eq)
+	q->elevator = elevator_alloc(q, e);
+	if (!q->elevator)
 		return -ENOMEM;
 
-	err = elevator_init_queue(q, eq);
+	err = e->ops.elevator_init_fn(q);
 	if (err) {
-		kobject_put(&eq->kobj);
+		kobject_put(&q->elevator->kobj);
 		return err;
 	}
 
-	q->elevator = eq;
 	return 0;
 }
 EXPORT_SYMBOL(elevator_init);
@@ -564,25 +554,6 @@ void elv_drain_elevator(struct request_queue *q)
 	}
 }
 
-void elv_quiesce_start(struct request_queue *q)
-{
-	if (!q->elevator)
-		return;
-
-	spin_lock_irq(q->queue_lock);
-	queue_flag_set(QUEUE_FLAG_ELVSWITCH, q);
-	spin_unlock_irq(q->queue_lock);
-
-	blk_drain_queue(q, false);
-}
-
-void elv_quiesce_end(struct request_queue *q)
-{
-	spin_lock_irq(q->queue_lock);
-	queue_flag_clear(QUEUE_FLAG_ELVSWITCH, q);
-	spin_unlock_irq(q->queue_lock);
-}
-
 void __elv_add_request(struct request_queue *q, struct request *rq, int where)
 {
 	trace_block_rq_insert(q, rq);
@@ -692,12 +663,13 @@ struct request *elv_former_request(struct request_queue *q, struct request *rq)
 	return NULL;
 }
 
-int elv_set_request(struct request_queue *q, struct request *rq, gfp_t gfp_mask)
+int elv_set_request(struct request_queue *q, struct request *rq,
+		    struct bio *bio, gfp_t gfp_mask)
 {
 	struct elevator_queue *e = q->elevator;
 
 	if (e->type->ops.elevator_set_req_fn)
-		return e->type->ops.elevator_set_req_fn(q, rq, gfp_mask);
+		return e->type->ops.elevator_set_req_fn(q, rq, bio, gfp_mask);
 	return 0;
 }
 
@@ -801,8 +773,9 @@ static struct kobj_type elv_ktype = {
 	.release	= elevator_release,
 };
 
-int __elv_register_queue(struct request_queue *q, struct elevator_queue *e)
+int elv_register_queue(struct request_queue *q)
 {
+	struct elevator_queue *e = q->elevator;
 	int error;
 
 	error = kobject_add(&e->kobj, &q->kobj, "%s", "iosched");
@@ -820,11 +793,6 @@ int __elv_register_queue(struct request_queue *q, struct elevator_queue *e)
 	}
 	return error;
 }
-
-int elv_register_queue(struct request_queue *q)
-{
-	return __elv_register_queue(q, q->elevator);
-}
 EXPORT_SYMBOL(elv_register_queue);
 
 void elv_unregister_queue(struct request_queue *q)
@@ -907,53 +875,60 @@ EXPORT_SYMBOL_GPL(elv_unregister);
  */
 static int elevator_switch(struct request_queue *q, struct elevator_type *new_e)
 {
-	struct elevator_queue *old_elevator, *e;
+	struct elevator_queue *old = q->elevator;
+	bool registered = old->registered;
 	int err;
 
-	/* allocate new elevator */
-	e = elevator_alloc(q, new_e);
-	if (!e)
-		return -ENOMEM;
+	/*
+	 * Turn on BYPASS and drain all requests w/ elevator private data.
+	 * Block layer doesn't call into a quiesced elevator - all requests
+	 * are directly put on the dispatch list without elevator data
+	 * using INSERT_BACK.  All requests have SOFTBARRIER set and no
+	 * merge happens either.
+	 */
+	blk_queue_bypass_start(q);
+
+	/* unregister and clear all auxiliary data of the old elevator */
+	if (registered)
+		elv_unregister_queue(q);
+
+	spin_lock_irq(q->queue_lock);
+	ioc_clear_queue(q);
+	spin_unlock_irq(q->queue_lock);
 
-	err = elevator_init_queue(q, e);
+	/* allocate, init and register new elevator */
+	err = -ENOMEM;
+	q->elevator = elevator_alloc(q, new_e);
+	if (!q->elevator)
+		goto fail_init;
+
+	err = new_e->ops.elevator_init_fn(q);
 	if (err) {
-		kobject_put(&e->kobj);
-		return err;
+		kobject_put(&q->elevator->kobj);
+		goto fail_init;
 	}
 
-	/* turn on BYPASS and drain all requests w/ elevator private data */
-	elv_quiesce_start(q);
-
-	/* unregister old queue, register new one and kill old elevator */
-	if (q->elevator->registered) {
-		elv_unregister_queue(q);
-		err = __elv_register_queue(q, e);
+	if (registered) {
+		err = elv_register_queue(q);
 		if (err)
 			goto fail_register;
 	}
 
-	/* done, clear io_cq's, switch elevators and turn off BYPASS */
-	spin_lock_irq(q->queue_lock);
-	ioc_clear_queue(q);
-	old_elevator = q->elevator;
-	q->elevator = e;
-	spin_unlock_irq(q->queue_lock);
-
-	elevator_exit(old_elevator);
-	elv_quiesce_end(q);
+	/* done, kill the old one and finish */
+	elevator_exit(old);
+	blk_queue_bypass_end(q);
 
-	blk_add_trace_msg(q, "elv switch: %s", e->type->elevator_name);
+	blk_add_trace_msg(q, "elv switch: %s", new_e->elevator_name);
 
 	return 0;
 
 fail_register:
-	/*
-	 * switch failed, exit the new io scheduler and reattach the old
-	 * one again (along with re-adding the sysfs dir)
-	 */
-	elevator_exit(e);
+	elevator_exit(q->elevator);
+fail_init:
+	/* switch failed, restore and re-register old elevator */
+	q->elevator = old;
 	elv_register_queue(q);
-	elv_quiesce_end(q);
+	blk_queue_bypass_end(q);
 
 	return err;
 }
diff --git a/block/noop-iosched.c b/block/noop-iosched.c
index 413a0b1d788c..5d1bf70e33d5 100644
--- a/block/noop-iosched.c
+++ b/block/noop-iosched.c
@@ -59,15 +59,17 @@ noop_latter_request(struct request_queue *q, struct request *rq)
 	return list_entry(rq->queuelist.next, struct request, queuelist);
 }
 
-static void *noop_init_queue(struct request_queue *q)
+static int noop_init_queue(struct request_queue *q)
 {
 	struct noop_data *nd;
 
 	nd = kmalloc_node(sizeof(*nd), GFP_KERNEL, q->node);
 	if (!nd)
-		return NULL;
+		return -ENOMEM;
+
 	INIT_LIST_HEAD(&nd->queue);
-	return nd;
+	q->elevator->elevator_data = nd;
+	return 0;
 }
 
 static void noop_exit_queue(struct elevator_queue *e)