[PATCHSET cgroup,block/for-4.3] cgroup, blkcg: establish controller interface conventions

[PATCHSET cgroup,block/for-4.3] cgroup, blkcg: establish controller interface conventions

[Tejun Heo]

Hello,

On the legacy hierarchies, different controllers implemented interface
without much guidance and as a result there isn't much consistency
across controllers even when they're doing similar things.

Controllers which already implemented interface for the unified
hierarchy are already pretty consistent and this patchset explicitly
sets up the conventions so that it's easier to make future conversions
conformant.

For blkcg, the only thing which needs to be changed is the range and
default value of io.weight which is currently [10, 1000] and 500.  For
the unified hierarchy, cgroup now defines the standard range and
default value to be [1, 10000] and 100 so that there's enough room for
100x bias in both directions.

This patchset contains the following four patches.

 0001-cgroup-export-cgrp_dfl_root.patch
 0002-cgroup-define-controller-file-conventions.patch
 0003-blkcg-s-CFQ_WEIGHT_-CFQ_WEIGHT_LEGACY_.patch
 0004-blkcg-use-CGROUP_WEIGHT_-scale-for-io.weight-on-the-.patch

0001-0002 explicitly defines the controller interface file conventions
on the unified hierarchy.  0003-0004 update blkcg so that it conforms
to the standard weight range.

This patchset is also available in the following git branch.

 git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git review-blkcg-normalize-weight

and is on top of

  block/for-linus f3f5da624e0a ("block: Do a full clone when splitting discard bios")
+ [1] [PATCHSET block/for-4.3] writeback: cgroup writeback updates
+ [2] [PATCHSET v2 block/for-4.3] block, cgroup: make cfq charge async IOs to the appropriate blkcgs
+ [3] [PATCHSET v3 block/for-4.3] blkcg: blkcg policy methods and data handling cleanup
+ [4] [PATCHSET v2 block/for-4.3] blkcg: blkcg stats cleanup
+ [5] [PATCHSET block/for-4.3] blkcg: implement interface for the unified hierarchy

diffstat follows.  Thanks.

 Documentation/cgroups/unified-hierarchy.txt |   77 ++++++++++++++++++++++++----
 block/blk-cgroup.c                          |   21 +++++++
 block/cfq-iosched.c                         |   63 ++++++++++++++++------
 include/linux/blk-cgroup.h                  |    2 
 include/linux/cgroup.h                      |    9 +++
 kernel/cgroup.c                             |    1 
 6 files changed, 145 insertions(+), 28 deletions(-)

--
tejun

[L] http://lkml.kernel.org/g/1435268337-1738-1-git-send-email-tj@kernel.org
[1] http://lkml.kernel.org/g/1436281823-1947-1-git-send-email-tj@kernel.org
[2] http://lkml.kernel.org/g/1436283361-3889-1-git-send-email-tj@kernel.org
[3] http://lkml.kernel.org/g/1436637654-28110-1-git-send-email-tj@kernel.org
[4] http://lkml.kernel.org/g/1436724043-12986-1-git-send-email-tj@kernel.org
[5] http://lkml.kernel.org/g/1437763434-28699-1-git-send-email-tj@kernel.org

[PATCH 1/4] cgroup: export cgrp_dfl_root

[Tejun Heo]

While cgroup subsystems can't be modules, blkcg supports dynamically
loadable policies which interact with cgroup core.  Export
cgrp_dfl_root so that cgroup_on_dfl() can be used in those modules.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Li Zefan <lizefan@huawei.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
---
 kernel/cgroup.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/kernel/cgroup.c b/kernel/cgroup.c
index 1276569..9c8eb17 100644
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -145,6 +145,7 @@ static const char *cgroup_subsys_name[] = {
  * part of that cgroup.
  */
 struct cgroup_root cgrp_dfl_root;
+EXPORT_SYMBOL_GPL(cgrp_dfl_root);
 
 /*
  * The default hierarchy always exists but is hidden until mounted for the
-- 
2.4.3

Re: [PATCH 1/4] cgroup: export cgrp_dfl_root

[Zefan Li]

On 2015/8/3 4:23, Tejun Heo wrote:
> While cgroup subsystems can't be modules, blkcg supports dynamically
> loadable policies which interact with cgroup core.  Export
> cgrp_dfl_root so that cgroup_on_dfl() can be used in those modules.
> 
> Signed-off-by: Tejun Heo <tj@kernel.org>
> Cc: Li Zefan <lizefan@huawei.com>
> Cc: Johannes Weiner <hannes@cmpxchg.org>

acked

Re: [PATCH 1/4] cgroup: export cgrp_dfl_root

[Tejun Heo]

On Sun, Aug 02, 2015 at 04:23:41PM -0400, Tejun Heo wrote:
> While cgroup subsystems can't be modules, blkcg supports dynamically
> loadable policies which interact with cgroup core.  Export
> cgrp_dfl_root so that cgroup_on_dfl() can be used in those modules.
> 
> Signed-off-by: Tejun Heo <tj@kernel.org>
> Cc: Li Zefan <lizefan@huawei.com>
> Cc: Johannes Weiner <hannes@cmpxchg.org>

Applied to cgroup/for-4.3-unified-base

Thanks.

-- 
tejun

[PATCH 2/4] cgroup: define controller file conventions

[Tejun Heo]

Traditionally, each cgroup controller implemented whatever interface
it wanted leading to interfaces which are widely inconsistent.
Examining the requirements of the controllers readily yield that there
are only a few control schemes shared among all.

Two major controllers already had to implement new interface for the
unified hierarchy due to significant structural changes.  Let's take
the chance to establish common conventions throughout all controllers.

This patch defines CGROUP_WEIGHT_MIN/DFL/MAX to be used on all weight
based control knobs and documents the conventions that controllers
should follow on the unified hierarchy.  Except for io.weight knob,
all existing unified hierarchy knobs are already compliant.  A
follow-up patch will update io.weight.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Li Zefan <lizefan@huawei.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
---
 Documentation/cgroups/unified-hierarchy.txt | 75 ++++++++++++++++++++++++++---
 include/linux/cgroup.h                      |  9 ++++
 2 files changed, 76 insertions(+), 8 deletions(-)

diff --git a/Documentation/cgroups/unified-hierarchy.txt b/Documentation/cgroups/unified-hierarchy.txt
index 4e23d4c..ee6ce34 100644
--- a/Documentation/cgroups/unified-hierarchy.txt
+++ b/Documentation/cgroups/unified-hierarchy.txt
@@ -23,10 +23,13 @@ CONTENTS
 5. Other Changes
   5-1. [Un]populated Notification
   5-2. Other Core Changes
-  5-3. Per-Controller Changes
-    5-3-1. blkio
-    5-3-2. cpuset
-    5-3-3. memory
+  5-3. Controller file conventions
+    5-3-1. Format
+    5-3-2. Control knobs
+  5-4. Per-Controller Changes
+    5-4-1. blkio
+    5-4-2. cpuset
+    5-4-3. memory
 6. Planned Changes
   6-1. CAP for resource control
 
@@ -372,9 +375,65 @@ supported and the interface files "release_agent" and
 - The "cgroup.clone_children" file is removed.
 
 
-5-3. Per-Controller Changes
+5-3. Controller file conventions
 
-5-3-1. io
+5-3-1. Format
+
+In general, all controller files should be in one of the following
+formats whenever possible.
+
+- Values only files
+
+  VAL0 VAL1...\n
+
+- Flat keyed files
+
+  KEY0 VAL0\n
+  KEY1 VAL1\n
+  ...
+
+- Nested keyed files
+
+  KEY0 SUB_KEY0=VAL00 SUB_KEY1=VAL01...
+  KEY1 SUB_KEY0=VAL10 SUB_KEY1=VAL11...
+  ...
+
+For a writeable file, the format for writing should generally match
+reading; however, controllers may allow omitting later fields or
+implement restricted shortcuts for most common use cases.
+
+For both flat and nested keyed files, only the values for a single key
+can be written at a time.  For nested keyed files, the sub key pairs
+may be specified in any order and not all pairs have to be specified.
+
+
+5-3-2. Control knobs
+
+- Settings for a single feature should generally be implemented in a
+  single file.
+
+- In general, the root cgroup should be exempt from resource control
+  and thus shouldn't have resource control knobs.
+
+- If a controller implements ratio based resource distribution, the
+  control knob should be named "weight" and have the range [1, 10000]
+  and 100 should be the default value.  The values are chosen to allow
+  enough and symmetric bias in both directions while keeping it
+  intuitive (the default is 100%).
+
+- If a controller implements an absolute resource limit, the control
+  knob should be named "max".  The special token "max" should be used
+  to represent no limit for both reading and writing.
+
+- If a setting has configurable default value and specific overrides,
+  the default settings should be keyed with "default" and appear as
+  the first entry in the file.  Specific entries can use "default" as
+  its value to indicate inheritance of the default value.
+
+
+5-4. Per-Controller Changes
+
+5-4-1. io
 
 - blkio is renamed to io.  The interface is overhauled anyway.  The
   new name is more in line with the other two major controllers, cpu
@@ -432,7 +491,7 @@ supported and the interface files "release_agent" and
 	This file is available only on non-root cgroups.
 
 
-5-3-2. cpuset
+5-4-2. cpuset
 
 - Tasks are kept in empty cpusets after hotplug and take on the masks
   of the nearest non-empty ancestor, instead of being moved to it.
@@ -441,7 +500,7 @@ supported and the interface files "release_agent" and
   masks of the nearest non-empty ancestor.
 
 
-5-3-3. memory
+5-4-3. memory
 
 - use_hierarchy is on by default and the cgroup file for the flag is
   not created.
diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h
index a593e29..c6bf9d3 100644
--- a/include/linux/cgroup.h
+++ b/include/linux/cgroup.h
@@ -22,6 +22,15 @@
 
 #ifdef CONFIG_CGROUPS
 
+/*
+ * All weight knobs on the default hierarhcy should use the following min,
+ * default and max values.  The default value is the logarithmic center of
+ * MIN and MAX and allows 100x to be expressed in both directions.
+ */
+#define CGROUP_WEIGHT_MIN		1
+#define CGROUP_WEIGHT_DFL		100
+#define CGROUP_WEIGHT_MAX		10000
+
 /* a css_task_iter should be treated as an opaque object */
 struct css_task_iter {
 	struct cgroup_subsys		*ss;
-- 
2.4.3

Re: [PATCH 2/4] cgroup: define controller file conventions

[Johannes Weiner]

On Sun, Aug 02, 2015 at 04:23:42PM -0400, Tejun Heo wrote:
> Traditionally, each cgroup controller implemented whatever interface
> it wanted leading to interfaces which are widely inconsistent.
> Examining the requirements of the controllers readily yield that there
> are only a few control schemes shared among all.
> 
> Two major controllers already had to implement new interface for the
> unified hierarchy due to significant structural changes.  Let's take
> the chance to establish common conventions throughout all controllers.
> 
> This patch defines CGROUP_WEIGHT_MIN/DFL/MAX to be used on all weight
> based control knobs and documents the conventions that controllers
> should follow on the unified hierarchy.  Except for io.weight knob,
> all existing unified hierarchy knobs are already compliant.  A
> follow-up patch will update io.weight.
> 
> Signed-off-by: Tejun Heo <tj@kernel.org>
> Cc: Li Zefan <lizefan@huawei.com>
> Cc: Johannes Weiner <hannes@cmpxchg.org>

Acked-by: Johannes Weiner <hannes@cmpxchg.org>

Re: [PATCH 2/4] cgroup: define controller file conventions

[Zefan Li]

On 2015/8/3 4:23, Tejun Heo wrote:
> Traditionally, each cgroup controller implemented whatever interface
> it wanted leading to interfaces which are widely inconsistent.
> Examining the requirements of the controllers readily yield that there
> are only a few control schemes shared among all.
> 
> Two major controllers already had to implement new interface for the
> unified hierarchy due to significant structural changes.  Let's take
> the chance to establish common conventions throughout all controllers.
> 
> This patch defines CGROUP_WEIGHT_MIN/DFL/MAX to be used on all weight
> based control knobs and documents the conventions that controllers
> should follow on the unified hierarchy.  Except for io.weight knob,
> all existing unified hierarchy knobs are already compliant.  A
> follow-up patch will update io.weight.
> 
> Signed-off-by: Tejun Heo <tj@kernel.org>
> Cc: Li Zefan <lizefan@huawei.com>
> Cc: Johannes Weiner <hannes@cmpxchg.org>

acked

[PATCH 3/4] blkcg: s/CFQ_WEIGHT_*/CFQ_WEIGHT_LEGACY_*/

[Tejun Heo]

blkcg is gonna switch to cgroup common weight range as defined by
CGROUP_WEIGHT_* on the unified hierarchy.  In preparation, rename
CFQ_WEIGHT_* constants to CFQ_WEIGHT_LEGACY_*.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Arianna Avanzini <avanzini.arianna@gmail.com>
---
 block/cfq-iosched.c | 22 +++++++++++-----------
 1 file changed, 11 insertions(+), 11 deletions(-)

diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
index 97da571..0fe721e 100644
--- a/block/cfq-iosched.c
+++ b/block/cfq-iosched.c
@@ -68,9 +68,9 @@ static struct kmem_cache *cfq_pool;
 #define rb_entry_cfqg(node)	rb_entry((node), struct cfq_group, rb_node)
 
 /* blkio-related constants */
-#define CFQ_WEIGHT_MIN          10
-#define CFQ_WEIGHT_MAX          1000
-#define CFQ_WEIGHT_DEFAULT      500
+#define CFQ_WEIGHT_LEGACY_MIN	10
+#define CFQ_WEIGHT_LEGACY_DFL	500
+#define CFQ_WEIGHT_LEGACY_MAX	1000
 
 struct cfq_ttime {
 	unsigned long last_end_request;
@@ -1580,11 +1580,11 @@ static void cfq_cpd_init(struct blkcg_policy_data *cpd)
 	struct cfq_group_data *cgd = cpd_to_cfqgd(cpd);
 
 	if (cpd_to_blkcg(cpd) == &blkcg_root) {
-		cgd->weight = 2 * CFQ_WEIGHT_DEFAULT;
-		cgd->leaf_weight = 2 * CFQ_WEIGHT_DEFAULT;
+		cgd->weight = 2 * CFQ_WEIGHT_LEGACY_DFL;
+		cgd->leaf_weight = 2 * CFQ_WEIGHT_LEGACY_DFL;
 	} else {
-		cgd->weight = CFQ_WEIGHT_DEFAULT;
-		cgd->leaf_weight = CFQ_WEIGHT_DEFAULT;
+		cgd->weight = CFQ_WEIGHT_LEGACY_DFL;
+		cgd->leaf_weight = CFQ_WEIGHT_LEGACY_DFL;
 	}
 }
 
@@ -1769,7 +1769,7 @@ static ssize_t __cfqg_set_weight_device(struct kernfs_open_file *of,
 	cfqgd = blkcg_to_cfqgd(blkcg);
 
 	ret = -ERANGE;
-	if (!v || (v >= CFQ_WEIGHT_MIN && v <= CFQ_WEIGHT_MAX)) {
+	if (!v || (v >= CFQ_WEIGHT_LEGACY_MIN && v <= CFQ_WEIGHT_LEGACY_MAX)) {
 		if (!is_leaf_weight) {
 			cfqg->dev_weight = v;
 			cfqg->new_weight = v ?: cfqgd->weight;
@@ -1804,7 +1804,7 @@ static int __cfq_set_weight(struct cgroup_subsys_state *css, u64 val,
 	struct cfq_group_data *cfqgd;
 	int ret = 0;
 
-	if (val < CFQ_WEIGHT_MIN || val > CFQ_WEIGHT_MAX)
+	if (val < CFQ_WEIGHT_LEGACY_MIN || val > CFQ_WEIGHT_LEGACY_MAX)
 		return -EINVAL;
 
 	spin_lock_irq(&blkcg->lock);
@@ -4513,8 +4513,8 @@ static int cfq_init_queue(struct request_queue *q, struct elevator_type *e)
 
 	cfq_init_cfqg_base(cfqd->root_group);
 #endif
-	cfqd->root_group->weight = 2 * CFQ_WEIGHT_DEFAULT;
-	cfqd->root_group->leaf_weight = 2 * CFQ_WEIGHT_DEFAULT;
+	cfqd->root_group->weight = 2 * CFQ_WEIGHT_LEGACY_DFL;
+	cfqd->root_group->leaf_weight = 2 * CFQ_WEIGHT_LEGACY_DFL;
 
 	/*
 	 * Not strictly needed (since RB_ROOT just clears the node and we
-- 
2.4.3

[PATCH 4/4] blkcg: use CGROUP_WEIGHT_* scale for io.weight on the unified hierarchy

[Tejun Heo]

cgroup is trying to make interface consistent across different
controllers.  For weight based resource control, the knob should have
the range [1, 10000] and default to 100.  This patch updates
cfq-iosched so that the weight range conforms.  The internal
calculations have enough range and the widening of the weight range
shouldn't cause any problem.

* blkcg_policy->cpd_bind_fn() is added.  If present, this is invoked
  when blkcg is attached to a hierarchy.

* cfq_cpd_init() is updated to use the new default value on the
  unified hierarchy.

* cfq_cpd_bind() callback is implemented to clear per-blkg configs and
  apply the default config matching the hierarchy type.

* cfqd->root_group->[leaf_]weight initialization in cfq_init_queue()
  is moved into !CONFIG_CFQ_GROUP_IOSCHED block.  cfq_cpd_bind() is
  now responsible for initializing the initial weights when blkcg is
  enabled.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Arianna Avanzini <avanzini.arianna@gmail.com>
---
 Documentation/cgroups/unified-hierarchy.txt |  2 +-
 block/blk-cgroup.c                          | 21 +++++++++++
 block/cfq-iosched.c                         | 55 +++++++++++++++++++++--------
 include/linux/blk-cgroup.h                  |  2 ++
 4 files changed, 64 insertions(+), 16 deletions(-)

diff --git a/Documentation/cgroups/unified-hierarchy.txt b/Documentation/cgroups/unified-hierarchy.txt
index ee6ce34..bd319f0 100644
--- a/Documentation/cgroups/unified-hierarchy.txt
+++ b/Documentation/cgroups/unified-hierarchy.txt
@@ -459,7 +459,7 @@ may be specified in any order and not all pairs have to be specified.
 
 	The weight setting, currently only available and effective if
 	cfq-iosched is in use for the target device.  The weight is
-	between 10 and 1000 and defaults to 500.  The first line
+	between 1 and 10000 and defaults to 100.  The first line
 	always contains the default weight in the following format to
 	use when per-device setting is missing.
 
diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index 88bdb73..ac8370c 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -1143,11 +1143,32 @@ static int blkcg_can_attach(struct cgroup_subsys_state *css,
 	return ret;
 }
 
+static void blkcg_bind(struct cgroup_subsys_state *root_css)
+{
+	int i;
+
+	mutex_lock(&blkcg_pol_mutex);
+
+	for (i = 0; i < BLKCG_MAX_POLS; i++) {
+		struct blkcg_policy *pol = blkcg_policy[i];
+		struct blkcg *blkcg;
+
+		if (!pol || !pol->cpd_bind_fn)
+			continue;
+
+		list_for_each_entry(blkcg, &all_blkcgs, all_blkcgs_node)
+			if (blkcg->cpd[pol->plid])
+				pol->cpd_bind_fn(blkcg->cpd[pol->plid]);
+	}
+	mutex_unlock(&blkcg_pol_mutex);
+}
+
 struct cgroup_subsys io_cgrp_subsys = {
 	.css_alloc = blkcg_css_alloc,
 	.css_offline = blkcg_css_offline,
 	.css_free = blkcg_css_free,
 	.can_attach = blkcg_can_attach,
+	.bind = blkcg_bind,
 	.dfl_cftypes = blkcg_files,
 	.legacy_cftypes = blkcg_legacy_files,
 	.legacy_name = "blkio",
diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
index 0fe721e..04de884 100644
--- a/block/cfq-iosched.c
+++ b/block/cfq-iosched.c
@@ -1522,6 +1522,9 @@ static void cfq_init_cfqg_base(struct cfq_group *cfqg)
 }
 
 #ifdef CONFIG_CFQ_GROUP_IOSCHED
+static int __cfq_set_weight(struct cgroup_subsys_state *css, u64 val,
+			    bool on_dfl, bool reset_dev, bool is_leaf_weight);
+
 static void cfqg_stats_exit(struct cfqg_stats *stats)
 {
 	blkg_rwstat_exit(&stats->merged);
@@ -1578,14 +1581,14 @@ static struct blkcg_policy_data *cfq_cpd_alloc(gfp_t gfp)
 static void cfq_cpd_init(struct blkcg_policy_data *cpd)
 {
 	struct cfq_group_data *cgd = cpd_to_cfqgd(cpd);
+	unsigned int weight = cgroup_on_dfl(blkcg_root.css.cgroup) ?
+			      CGROUP_WEIGHT_DFL : CFQ_WEIGHT_LEGACY_DFL;
 
-	if (cpd_to_blkcg(cpd) == &blkcg_root) {
-		cgd->weight = 2 * CFQ_WEIGHT_LEGACY_DFL;
-		cgd->leaf_weight = 2 * CFQ_WEIGHT_LEGACY_DFL;
-	} else {
-		cgd->weight = CFQ_WEIGHT_LEGACY_DFL;
-		cgd->leaf_weight = CFQ_WEIGHT_LEGACY_DFL;
-	}
+	if (cpd_to_blkcg(cpd) == &blkcg_root)
+		weight *= 2;
+
+	cgd->weight = weight;
+	cgd->leaf_weight = weight;
 }
 
 static void cfq_cpd_free(struct blkcg_policy_data *cpd)
@@ -1593,6 +1596,19 @@ static void cfq_cpd_free(struct blkcg_policy_data *cpd)
 	kfree(cpd_to_cfqgd(cpd));
 }
 
+static void cfq_cpd_bind(struct blkcg_policy_data *cpd)
+{
+	struct blkcg *blkcg = cpd_to_blkcg(cpd);
+	bool on_dfl = cgroup_on_dfl(blkcg_root.css.cgroup);
+	unsigned int weight = on_dfl ? CGROUP_WEIGHT_DFL : CFQ_WEIGHT_LEGACY_DFL;
+
+	if (blkcg == &blkcg_root)
+		weight *= 2;
+
+	WARN_ON_ONCE(__cfq_set_weight(&blkcg->css, weight, on_dfl, true, false));
+	WARN_ON_ONCE(__cfq_set_weight(&blkcg->css, weight, on_dfl, true, true));
+}
+
 static struct blkg_policy_data *cfq_pd_alloc(gfp_t gfp, int node)
 {
 	struct cfq_group *cfqg;
@@ -1742,6 +1758,8 @@ static ssize_t __cfqg_set_weight_device(struct kernfs_open_file *of,
 					char *buf, size_t nbytes, loff_t off,
 					bool on_dfl, bool is_leaf_weight)
 {
+	unsigned int min = on_dfl ? CGROUP_WEIGHT_MIN : CFQ_WEIGHT_LEGACY_MIN;
+	unsigned int max = on_dfl ? CGROUP_WEIGHT_MAX : CFQ_WEIGHT_LEGACY_MAX;
 	struct blkcg *blkcg = css_to_blkcg(of_css(of));
 	struct blkg_conf_ctx ctx;
 	struct cfq_group *cfqg;
@@ -1769,7 +1787,7 @@ static ssize_t __cfqg_set_weight_device(struct kernfs_open_file *of,
 	cfqgd = blkcg_to_cfqgd(blkcg);
 
 	ret = -ERANGE;
-	if (!v || (v >= CFQ_WEIGHT_LEGACY_MIN && v <= CFQ_WEIGHT_LEGACY_MAX)) {
+	if (!v || (v >= min && v <= max)) {
 		if (!is_leaf_weight) {
 			cfqg->dev_weight = v;
 			cfqg->new_weight = v ?: cfqgd->weight;
@@ -1797,15 +1815,17 @@ static ssize_t cfqg_set_leaf_weight_device(struct kernfs_open_file *of,
 }
 
 static int __cfq_set_weight(struct cgroup_subsys_state *css, u64 val,
-			    bool is_leaf_weight)
+			    bool on_dfl, bool reset_dev, bool is_leaf_weight)
 {
+	unsigned int min = on_dfl ? CGROUP_WEIGHT_MIN : CFQ_WEIGHT_LEGACY_MIN;
+	unsigned int max = on_dfl ? CGROUP_WEIGHT_MAX : CFQ_WEIGHT_LEGACY_MAX;
 	struct blkcg *blkcg = css_to_blkcg(css);
 	struct blkcg_gq *blkg;
 	struct cfq_group_data *cfqgd;
 	int ret = 0;
 
-	if (val < CFQ_WEIGHT_LEGACY_MIN || val > CFQ_WEIGHT_LEGACY_MAX)
-		return -EINVAL;
+	if (val < min || val > max)
+		return -ERANGE;
 
 	spin_lock_irq(&blkcg->lock);
 	cfqgd = blkcg_to_cfqgd(blkcg);
@@ -1826,9 +1846,13 @@ static int __cfq_set_weight(struct cgroup_subsys_state *css, u64 val,
 			continue;
 
 		if (!is_leaf_weight) {
+			if (reset_dev)
+				cfqg->dev_weight = 0;
 			if (!cfqg->dev_weight)
 				cfqg->new_weight = cfqgd->weight;
 		} else {
+			if (reset_dev)
+				cfqg->dev_leaf_weight = 0;
 			if (!cfqg->dev_leaf_weight)
 				cfqg->new_leaf_weight = cfqgd->leaf_weight;
 		}
@@ -1842,13 +1866,13 @@ static int __cfq_set_weight(struct cgroup_subsys_state *css, u64 val,
 static int cfq_set_weight(struct cgroup_subsys_state *css, struct cftype *cft,
 			  u64 val)
 {
-	return __cfq_set_weight(css, val, false);
+	return __cfq_set_weight(css, val, false, false, false);
 }
 
 static int cfq_set_leaf_weight(struct cgroup_subsys_state *css,
 			       struct cftype *cft, u64 val)
 {
-	return __cfq_set_weight(css, val, true);
+	return __cfq_set_weight(css, val, false, false, true);
 }
 
 static int cfqg_print_stat(struct seq_file *sf, void *v)
@@ -2135,7 +2159,7 @@ static ssize_t cfq_set_weight_on_dfl(struct kernfs_open_file *of,
 	/* "WEIGHT" or "default WEIGHT" sets the default weight */
 	v = simple_strtoull(buf, &endp, 0);
 	if (*endp == '\0' || sscanf(buf, "default %llu", &v) == 1) {
-		ret = __cfq_set_weight(of_css(of), v, false);
+		ret = __cfq_set_weight(of_css(of), v, true, false, false);
 		return ret ?: nbytes;
 	}
 
@@ -4512,9 +4536,9 @@ static int cfq_init_queue(struct request_queue *q, struct elevator_type *e)
 		goto out_free;
 
 	cfq_init_cfqg_base(cfqd->root_group);
-#endif
 	cfqd->root_group->weight = 2 * CFQ_WEIGHT_LEGACY_DFL;
 	cfqd->root_group->leaf_weight = 2 * CFQ_WEIGHT_LEGACY_DFL;
+#endif
 
 	/*
 	 * Not strictly needed (since RB_ROOT just clears the node and we
@@ -4715,6 +4739,7 @@ static struct blkcg_policy blkcg_policy_cfq = {
 	.cpd_alloc_fn		= cfq_cpd_alloc,
 	.cpd_init_fn		= cfq_cpd_init,
 	.cpd_free_fn		= cfq_cpd_free,
+	.cpd_bind_fn		= cfq_cpd_bind,
 
 	.pd_alloc_fn		= cfq_pd_alloc,
 	.pd_init_fn		= cfq_pd_init,
diff --git a/include/linux/blk-cgroup.h b/include/linux/blk-cgroup.h
index 9a7c4bd..0a5cc7a 100644
--- a/include/linux/blk-cgroup.h
+++ b/include/linux/blk-cgroup.h
@@ -138,6 +138,7 @@ struct blkcg_gq {
 typedef struct blkcg_policy_data *(blkcg_pol_alloc_cpd_fn)(gfp_t gfp);
 typedef void (blkcg_pol_init_cpd_fn)(struct blkcg_policy_data *cpd);
 typedef void (blkcg_pol_free_cpd_fn)(struct blkcg_policy_data *cpd);
+typedef void (blkcg_pol_bind_cpd_fn)(struct blkcg_policy_data *cpd);
 typedef struct blkg_policy_data *(blkcg_pol_alloc_pd_fn)(gfp_t gfp, int node);
 typedef void (blkcg_pol_init_pd_fn)(struct blkg_policy_data *pd);
 typedef void (blkcg_pol_online_pd_fn)(struct blkg_policy_data *pd);
@@ -155,6 +156,7 @@ struct blkcg_policy {
 	blkcg_pol_alloc_cpd_fn		*cpd_alloc_fn;
 	blkcg_pol_init_cpd_fn		*cpd_init_fn;
 	blkcg_pol_free_cpd_fn		*cpd_free_fn;
+	blkcg_pol_bind_cpd_fn		*cpd_bind_fn;
 
 	blkcg_pol_alloc_pd_fn		*pd_alloc_fn;
 	blkcg_pol_init_pd_fn		*pd_init_fn;
-- 
2.4.3