[PATCH 1/2] cgroup: s/CGRP_CLONE_CHILDREN/CGRP_CPUSET_CLONE_CHILDREN/

[PATCH 1/2] cgroup: s/CGRP_CLONE_CHILDREN/CGRP_CPUSET_CLONE_CHILDREN/

[Tejun Heo]

>From 6e405c1ae4018d813e8ed9e0bd463d6976aebfa8 Mon Sep 17 00:00:00 2001
From: Tejun Heo <tj@kernel.org>
Date: Tue, 13 Nov 2012 12:21:50 -0800

clone_children is only meaningful for cpuset and will stay that way.
Rename the flag to reflect that and update documentation.  Also, drop
clone_children() wrapper in cgroup.c.  The thin wrapper is used only a
few times and one of them will go away soon.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Glauber Costa <glommer@parallels.com>
---
These two patches are based on top of "cgroup: allow->post_create() to
fail" patchset.

 http://thread.gmane.org/gmane.linux.kernel.cgroups/5047
 git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git review-cgroup-online-may-fail

and availalbe in the following branch.

 git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git review-cgroup-remove-post_clone

Thanks.

 Documentation/cgroups/cgroups.txt |  8 +++-----
 include/linux/cgroup.h            |  6 ++++--
 kernel/cgroup.c                   | 28 ++++++++++++----------------
 3 files changed, 19 insertions(+), 23 deletions(-)

diff --git a/Documentation/cgroups/cgroups.txt b/Documentation/cgroups/cgroups.txt
index b06eea2..24cdf76 100644
--- a/Documentation/cgroups/cgroups.txt
+++ b/Documentation/cgroups/cgroups.txt
@@ -299,11 +299,9 @@ a cgroup hierarchy's release_agent path is empty.
 1.5 What does clone_children do ?
 ---------------------------------
 
-If the clone_children flag is enabled (1) in a cgroup, then all
-cgroups created beneath will call the post_clone callbacks for each
-subsystem of the newly created cgroup. Usually when this callback is
-implemented for a subsystem, it copies the values of the parent
-subsystem, this is the case for the cpuset.
+This flag only affects the cpuset controller. If the clone_children
+flag is enabled (1) in a cgroup, a new cpuset cgroup will copy its
+configuration from the parent during initialization.
 
 1.6 How do I use cgroups ?
 --------------------------
diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h
index 89c631d..926d8d1 100644
--- a/include/linux/cgroup.h
+++ b/include/linux/cgroup.h
@@ -143,9 +143,11 @@ enum {
 	/* Control Group requires release notifications to userspace */
 	CGRP_NOTIFY_ON_RELEASE,
 	/*
-	 * Clone cgroup values when creating a new child cgroup
+	 * Clone the parent's configuration when creating a new child
+	 * cpuset cgroup.  For historical reasons, this option can be
+	 * specified at mount time and thus is implemented here.
 	 */
-	CGRP_CLONE_CHILDREN,
+	CGRP_CPUSET_CLONE_CHILDREN,
 };
 
 struct cgroup {
diff --git a/kernel/cgroup.c b/kernel/cgroup.c
index 46c5119..a49572e 100644
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -296,11 +296,6 @@ static int notify_on_release(const struct cgroup *cgrp)
 	return test_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags);
 }
 
-static int clone_children(const struct cgroup *cgrp)
-{
-	return test_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
-}
-
 /*
  * for_each_subsys() allows you to iterate on each subsystem attached to
  * an active hierarchy
@@ -1101,7 +1096,7 @@ static int cgroup_show_options(struct seq_file *seq, struct dentry *dentry)
 		seq_puts(seq, ",xattr");
 	if (strlen(root->release_agent_path))
 		seq_printf(seq, ",release_agent=%s", root->release_agent_path);
-	if (clone_children(&root->top_cgroup))
+	if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &root->top_cgroup.flags))
 		seq_puts(seq, ",clone_children");
 	if (strlen(root->name))
 		seq_printf(seq, ",name=%s", root->name);
@@ -1113,7 +1108,7 @@ struct cgroup_sb_opts {
 	unsigned long subsys_mask;
 	unsigned long flags;
 	char *release_agent;
-	bool clone_children;
+	bool cpuset_clone_children;
 	char *name;
 	/* User explicitly requested empty subsystem */
 	bool none;
@@ -1164,7 +1159,7 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts)
 			continue;
 		}
 		if (!strcmp(token, "clone_children")) {
-			opts->clone_children = true;
+			opts->cpuset_clone_children = true;
 			continue;
 		}
 		if (!strcmp(token, "xattr")) {
@@ -1474,8 +1469,8 @@ static struct cgroupfs_root *cgroup_root_from_opts(struct cgroup_sb_opts *opts)
 		strcpy(root->release_agent_path, opts->release_agent);
 	if (opts->name)
 		strcpy(root->name, opts->name);
-	if (opts->clone_children)
-		set_bit(CGRP_CLONE_CHILDREN, &root->top_cgroup.flags);
+	if (opts->cpuset_clone_children)
+		set_bit(CGRP_CPUSET_CLONE_CHILDREN, &root->top_cgroup.flags);
 	return root;
 }
 
@@ -3905,7 +3900,7 @@ fail:
 static u64 cgroup_clone_children_read(struct cgroup *cgrp,
 				    struct cftype *cft)
 {
-	return clone_children(cgrp);
+	return test_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
 }
 
 static int cgroup_clone_children_write(struct cgroup *cgrp,
@@ -3913,9 +3908,9 @@ static int cgroup_clone_children_write(struct cgroup *cgrp,
 				     u64 val)
 {
 	if (val)
-		set_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
+		set_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
 	else
-		clear_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
+		clear_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
 	return 0;
 }
 
@@ -4130,8 +4125,8 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry,
 	if (notify_on_release(parent))
 		set_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags);
 
-	if (clone_children(parent))
-		set_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
+	if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &parent->flags))
+		set_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
 
 	for_each_subsys(root, ss) {
 		struct cgroup_subsys_state *css;
@@ -4148,7 +4143,8 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry,
 				goto err_free_all;
 		}
 		/* At error, ->css_free() callback has to free assigned ID. */
-		if (clone_children(parent) && ss->post_clone)
+		if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &parent->flags) &&
+		    ss->post_clone)
 			ss->post_clone(cgrp);
 
 		if (ss->broken_hierarchy && !ss->warned_broken_hierarchy &&
-- 
1.7.11.7

[PATCH 1/2] cgroup: s/CGRP_CLONE_CHILDREN/CGRP_CPUSET_CLONE_CHILDREN/

[Tejun Heo]

From 6e405c1ae4018d813e8ed9e0bd463d6976aebfa8 Mon Sep 17 00:00:00 2001
From: Tejun Heo <tj@kernel.org>
Date: Tue, 13 Nov 2012 12:21:50 -0800

clone_children is only meaningful for cpuset and will stay that way.
Rename the flag to reflect that and update documentation.  Also, drop
clone_children() wrapper in cgroup.c.  The thin wrapper is used only a
few times and one of them will go away soon.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Glauber Costa <glommer@parallels.com>
---
These two patches are based on top of "cgroup: allow->post_create() to
fail" patchset.

 http://thread.gmane.org/gmane.linux.kernel.cgroups/5047
 git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git review-cgroup-online-may-fail

and availalbe in the following branch.

 git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git review-cgroup-remove-post_clone

Thanks.

 Documentation/cgroups/cgroups.txt |  8 +++-----
 include/linux/cgroup.h            |  6 ++++--
 kernel/cgroup.c                   | 28 ++++++++++++----------------
 3 files changed, 19 insertions(+), 23 deletions(-)

diff --git a/Documentation/cgroups/cgroups.txt b/Documentation/cgroups/cgroups.txt
index b06eea2..24cdf76 100644
--- a/Documentation/cgroups/cgroups.txt
+++ b/Documentation/cgroups/cgroups.txt
@@ -299,11 +299,9 @@ a cgroup hierarchy's release_agent path is empty.
 1.5 What does clone_children do ?
 ---------------------------------
 
-If the clone_children flag is enabled (1) in a cgroup, then all
-cgroups created beneath will call the post_clone callbacks for each
-subsystem of the newly created cgroup. Usually when this callback is
-implemented for a subsystem, it copies the values of the parent
-subsystem, this is the case for the cpuset.
+This flag only affects the cpuset controller. If the clone_children
+flag is enabled (1) in a cgroup, a new cpuset cgroup will copy its
+configuration from the parent during initialization.
 
 1.6 How do I use cgroups ?
 --------------------------
diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h
index 89c631d..926d8d1 100644
--- a/include/linux/cgroup.h
+++ b/include/linux/cgroup.h
@@ -143,9 +143,11 @@ enum {
 	/* Control Group requires release notifications to userspace */
 	CGRP_NOTIFY_ON_RELEASE,
 	/*
-	 * Clone cgroup values when creating a new child cgroup
+	 * Clone the parent's configuration when creating a new child
+	 * cpuset cgroup.  For historical reasons, this option can be
+	 * specified at mount time and thus is implemented here.
 	 */
-	CGRP_CLONE_CHILDREN,
+	CGRP_CPUSET_CLONE_CHILDREN,
 };
 
 struct cgroup {
diff --git a/kernel/cgroup.c b/kernel/cgroup.c
index 46c5119..a49572e 100644
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -296,11 +296,6 @@ static int notify_on_release(const struct cgroup *cgrp)
 	return test_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags);
 }
 
-static int clone_children(const struct cgroup *cgrp)
-{
-	return test_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
-}
-
 /*
  * for_each_subsys() allows you to iterate on each subsystem attached to
  * an active hierarchy
@@ -1101,7 +1096,7 @@ static int cgroup_show_options(struct seq_file *seq, struct dentry *dentry)
 		seq_puts(seq, ",xattr");
 	if (strlen(root->release_agent_path))
 		seq_printf(seq, ",release_agent=%s", root->release_agent_path);
-	if (clone_children(&root->top_cgroup))
+	if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &root->top_cgroup.flags))
 		seq_puts(seq, ",clone_children");
 	if (strlen(root->name))
 		seq_printf(seq, ",name=%s", root->name);
@@ -1113,7 +1108,7 @@ struct cgroup_sb_opts {
 	unsigned long subsys_mask;
 	unsigned long flags;
 	char *release_agent;
-	bool clone_children;
+	bool cpuset_clone_children;
 	char *name;
 	/* User explicitly requested empty subsystem */
 	bool none;
@@ -1164,7 +1159,7 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts)
 			continue;
 		}
 		if (!strcmp(token, "clone_children")) {
-			opts->clone_children = true;
+			opts->cpuset_clone_children = true;
 			continue;
 		}
 		if (!strcmp(token, "xattr")) {
@@ -1474,8 +1469,8 @@ static struct cgroupfs_root *cgroup_root_from_opts(struct cgroup_sb_opts *opts)
 		strcpy(root->release_agent_path, opts->release_agent);
 	if (opts->name)
 		strcpy(root->name, opts->name);
-	if (opts->clone_children)
-		set_bit(CGRP_CLONE_CHILDREN, &root->top_cgroup.flags);
+	if (opts->cpuset_clone_children)
+		set_bit(CGRP_CPUSET_CLONE_CHILDREN, &root->top_cgroup.flags);
 	return root;
 }
 
@@ -3905,7 +3900,7 @@ fail:
 static u64 cgroup_clone_children_read(struct cgroup *cgrp,
 				    struct cftype *cft)
 {
-	return clone_children(cgrp);
+	return test_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
 }
 
 static int cgroup_clone_children_write(struct cgroup *cgrp,
@@ -3913,9 +3908,9 @@ static int cgroup_clone_children_write(struct cgroup *cgrp,
 				     u64 val)
 {
 	if (val)
-		set_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
+		set_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
 	else
-		clear_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
+		clear_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
 	return 0;
 }
 
@@ -4130,8 +4125,8 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry,
 	if (notify_on_release(parent))
 		set_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags);
 
-	if (clone_children(parent))
-		set_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
+	if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &parent->flags))
+		set_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
 
 	for_each_subsys(root, ss) {
 		struct cgroup_subsys_state *css;
@@ -4148,7 +4143,8 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry,
 				goto err_free_all;
 		}
 		/* At error, ->css_free() callback has to free assigned ID. */
-		if (clone_children(parent) && ss->post_clone)
+		if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &parent->flags) &&
+		    ss->post_clone)
 			ss->post_clone(cgrp);
 
 		if (ss->broken_hierarchy && !ss->warned_broken_hierarchy &&
-- 
1.7.11.7

[PATCH 2/2] cgroup, cpuset: remove cgroup_subsys->post_clone()

[Tejun Heo]

From bbf2566f9f4fc79ff3320e83cafb69533efc9ea0 Mon Sep 17 00:00:00 2001
From: Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
Date: Tue, 13 Nov 2012 12:21:50 -0800

Currently CGRP_CPUSET_CLONE_CHILDREN triggers ->post_clone().  Now
that clone_children is cpuset specific, there's no reason to have this
rather odd option activation mechanism in cgroup core.  cpuset can
check the flag from its ->css_allocate() and take the necessary
action.

Move cpuset_post_clone() logic to the end of cpuset_css_alloc() and
remove cgroup_subsys->post_clone().

Loosely based on Glauber's "generalize post_clone into post_create"
patch.

Signed-off-by: Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
Original-patch-by: Glauber Costa <glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
Original-patch: <1351686554-22592-2-git-send-email-glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
Cc: Glauber Costa <glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
---
 Documentation/cgroups/cgroups.txt |  8 ----
 include/linux/cgroup.h            |  1 -
 kernel/cgroup.c                   |  4 --
 kernel/cpuset.c                   | 80 ++++++++++++++++++---------------------
 4 files changed, 36 insertions(+), 57 deletions(-)

diff --git a/Documentation/cgroups/cgroups.txt b/Documentation/cgroups/cgroups.txt
index 24cdf76..bcf1a00 100644
--- a/Documentation/cgroups/cgroups.txt
+++ b/Documentation/cgroups/cgroups.txt
@@ -642,14 +642,6 @@ void exit(struct task_struct *task)
 
 Called during task exit.
 
-void post_clone(struct cgroup *cgrp)
-(cgroup_mutex held by caller)
-
-Called during cgroup_create() to do any parameter
-initialization which might be required before a task could attach.  For
-example, in cpusets, no task may attach before 'cpus' and 'mems' are set
-up.
-
 void bind(struct cgroup *root)
 (cgroup_mutex held by caller)
 
diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h
index 926d8d1..d5fc8a7 100644
--- a/include/linux/cgroup.h
+++ b/include/linux/cgroup.h
@@ -449,7 +449,6 @@ struct cgroup_subsys {
 	void (*fork)(struct task_struct *task);
 	void (*exit)(struct cgroup *cgrp, struct cgroup *old_cgrp,
 		     struct task_struct *task);
-	void (*post_clone)(struct cgroup *cgrp);
 	void (*bind)(struct cgroup *root);
 
 	int subsys_id;
diff --git a/kernel/cgroup.c b/kernel/cgroup.c
index a49572e..35aa18b 100644
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -4142,10 +4142,6 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry,
 			if (err)
 				goto err_free_all;
 		}
-		/* At error, ->css_free() callback has to free assigned ID. */
-		if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &parent->flags) &&
-		    ss->post_clone)
-			ss->post_clone(cgrp);
 
 		if (ss->broken_hierarchy && !ss->warned_broken_hierarchy &&
 		    parent->parent) {
diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index 0693133..b017887 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -1784,56 +1784,20 @@ static struct cftype files[] = {
 };
 
 /*
- * post_clone() is called during cgroup_create() when the
- * clone_children mount argument was specified.  The cgroup
- * can not yet have any tasks.
- *
- * Currently we refuse to set up the cgroup - thereby
- * refusing the task to be entered, and as a result refusing
- * the sys_unshare() or clone() which initiated it - if any
- * sibling cpusets have exclusive cpus or mem.
- *
- * If this becomes a problem for some users who wish to
- * allow that scenario, then cpuset_post_clone() could be
- * changed to grant parent->cpus_allowed-sibling_cpus_exclusive
- * (and likewise for mems) to the new cgroup. Called with cgroup_mutex
- * held.
- */
-static void cpuset_post_clone(struct cgroup *cgroup)
-{
-	struct cgroup *parent, *child;
-	struct cpuset *cs, *parent_cs;
-
-	parent = cgroup->parent;
-	list_for_each_entry(child, &parent->children, sibling) {
-		cs = cgroup_cs(child);
-		if (is_mem_exclusive(cs) || is_cpu_exclusive(cs))
-			return;
-	}
-	cs = cgroup_cs(cgroup);
-	parent_cs = cgroup_cs(parent);
-
-	mutex_lock(&callback_mutex);
-	cs->mems_allowed = parent_cs->mems_allowed;
-	cpumask_copy(cs->cpus_allowed, parent_cs->cpus_allowed);
-	mutex_unlock(&callback_mutex);
-	return;
-}
-
-/*
  *	cpuset_css_alloc - allocate a cpuset css
  *	cont:	control group that the new cpuset will be part of
  */
 
 static struct cgroup_subsys_state *cpuset_css_alloc(struct cgroup *cont)
 {
-	struct cpuset *cs;
-	struct cpuset *parent;
+	struct cgroup *parent_cg = cont->parent;
+	struct cgroup *tmp_cg;
+	struct cpuset *parent, *cs;
 
-	if (!cont->parent) {
+	if (!parent_cg)
 		return &top_cpuset.css;
-	}
-	parent = cgroup_cs(cont->parent);
+	parent = cgroup_cs(parent_cg);
+
 	cs = kmalloc(sizeof(*cs), GFP_KERNEL);
 	if (!cs)
 		return ERR_PTR(-ENOMEM);
@@ -1855,7 +1819,36 @@ static struct cgroup_subsys_state *cpuset_css_alloc(struct cgroup *cont)
 
 	cs->parent = parent;
 	number_of_cpusets++;
-	return &cs->css ;
+
+	if (!test_bit(CGRP_CPUSET_CLONE_CHILDREN, &cont->flags))
+		goto skip_clone;
+
+	/*
+	 * Clone @parent's configuration if CGRP_CPUSET_CLONE_CHILDREN is
+	 * set.  This flag handling is implemented in cgroup core for
+	 * histrical reasons - the flag may be specified during mount.
+	 *
+	 * Currently, if any sibling cpusets have exclusive cpus or mem, we
+	 * refuse to clone the configuration - thereby refusing the task to
+	 * be entered, and as a result refusing the sys_unshare() or
+	 * clone() which initiated it.  If this becomes a problem for some
+	 * users who wish to allow that scenario, then this could be
+	 * changed to grant parent->cpus_allowed-sibling_cpus_exclusive
+	 * (and likewise for mems) to the new cgroup.
+	 */
+	list_for_each_entry(tmp_cg, &parent_cg->children, sibling) {
+		struct cpuset *tmp_cs = cgroup_cs(tmp_cg);
+
+		if (is_mem_exclusive(tmp_cs) || is_cpu_exclusive(tmp_cs))
+			goto skip_clone;
+	}
+
+	mutex_lock(&callback_mutex);
+	cs->mems_allowed = parent->mems_allowed;
+	cpumask_copy(cs->cpus_allowed, parent->cpus_allowed);
+	mutex_unlock(&callback_mutex);
+skip_clone:
+	return &cs->css;
 }
 
 /*
@@ -1882,7 +1875,6 @@ struct cgroup_subsys cpuset_subsys = {
 	.css_free = cpuset_css_free,
 	.can_attach = cpuset_can_attach,
 	.attach = cpuset_attach,
-	.post_clone = cpuset_post_clone,
 	.subsys_id = cpuset_subsys_id,
 	.base_cftypes = files,
 	.early_init = 1,
-- 
1.7.11.7

[PATCH 2/2] cgroup, cpuset: remove cgroup_subsys->post_clone()

[Tejun Heo]

>From bbf2566f9f4fc79ff3320e83cafb69533efc9ea0 Mon Sep 17 00:00:00 2001
From: Tejun Heo <tj@kernel.org>
Date: Tue, 13 Nov 2012 12:21:50 -0800

Currently CGRP_CPUSET_CLONE_CHILDREN triggers ->post_clone().  Now
that clone_children is cpuset specific, there's no reason to have this
rather odd option activation mechanism in cgroup core.  cpuset can
check the flag from its ->css_allocate() and take the necessary
action.

Move cpuset_post_clone() logic to the end of cpuset_css_alloc() and
remove cgroup_subsys->post_clone().

Loosely based on Glauber's "generalize post_clone into post_create"
patch.

Signed-off-by: Tejun Heo <tj@kernel.org>
Original-patch-by: Glauber Costa <glommer@parallels.com>
Original-patch: <1351686554-22592-2-git-send-email-glommer@parallels.com>
Cc: Glauber Costa <glommer@parallels.com>
---
 Documentation/cgroups/cgroups.txt |  8 ----
 include/linux/cgroup.h            |  1 -
 kernel/cgroup.c                   |  4 --
 kernel/cpuset.c                   | 80 ++++++++++++++++++---------------------
 4 files changed, 36 insertions(+), 57 deletions(-)

diff --git a/Documentation/cgroups/cgroups.txt b/Documentation/cgroups/cgroups.txt
index 24cdf76..bcf1a00 100644
--- a/Documentation/cgroups/cgroups.txt
+++ b/Documentation/cgroups/cgroups.txt
@@ -642,14 +642,6 @@ void exit(struct task_struct *task)
 
 Called during task exit.
 
-void post_clone(struct cgroup *cgrp)
-(cgroup_mutex held by caller)
-
-Called during cgroup_create() to do any parameter
-initialization which might be required before a task could attach.  For
-example, in cpusets, no task may attach before 'cpus' and 'mems' are set
-up.
-
 void bind(struct cgroup *root)
 (cgroup_mutex held by caller)
 
diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h
index 926d8d1..d5fc8a7 100644
--- a/include/linux/cgroup.h
+++ b/include/linux/cgroup.h
@@ -449,7 +449,6 @@ struct cgroup_subsys {
 	void (*fork)(struct task_struct *task);
 	void (*exit)(struct cgroup *cgrp, struct cgroup *old_cgrp,
 		     struct task_struct *task);
-	void (*post_clone)(struct cgroup *cgrp);
 	void (*bind)(struct cgroup *root);
 
 	int subsys_id;
diff --git a/kernel/cgroup.c b/kernel/cgroup.c
index a49572e..35aa18b 100644
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -4142,10 +4142,6 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry,
 			if (err)
 				goto err_free_all;
 		}
-		/* At error, ->css_free() callback has to free assigned ID. */
-		if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &parent->flags) &&
-		    ss->post_clone)
-			ss->post_clone(cgrp);
 
 		if (ss->broken_hierarchy && !ss->warned_broken_hierarchy &&
 		    parent->parent) {
diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index 0693133..b017887 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -1784,56 +1784,20 @@ static struct cftype files[] = {
 };
 
 /*
- * post_clone() is called during cgroup_create() when the
- * clone_children mount argument was specified.  The cgroup
- * can not yet have any tasks.
- *
- * Currently we refuse to set up the cgroup - thereby
- * refusing the task to be entered, and as a result refusing
- * the sys_unshare() or clone() which initiated it - if any
- * sibling cpusets have exclusive cpus or mem.
- *
- * If this becomes a problem for some users who wish to
- * allow that scenario, then cpuset_post_clone() could be
- * changed to grant parent->cpus_allowed-sibling_cpus_exclusive
- * (and likewise for mems) to the new cgroup. Called with cgroup_mutex
- * held.
- */
-static void cpuset_post_clone(struct cgroup *cgroup)
-{
-	struct cgroup *parent, *child;
-	struct cpuset *cs, *parent_cs;
-
-	parent = cgroup->parent;
-	list_for_each_entry(child, &parent->children, sibling) {
-		cs = cgroup_cs(child);
-		if (is_mem_exclusive(cs) || is_cpu_exclusive(cs))
-			return;
-	}
-	cs = cgroup_cs(cgroup);
-	parent_cs = cgroup_cs(parent);
-
-	mutex_lock(&callback_mutex);
-	cs->mems_allowed = parent_cs->mems_allowed;
-	cpumask_copy(cs->cpus_allowed, parent_cs->cpus_allowed);
-	mutex_unlock(&callback_mutex);
-	return;
-}
-
-/*
  *	cpuset_css_alloc - allocate a cpuset css
  *	cont:	control group that the new cpuset will be part of
  */
 
 static struct cgroup_subsys_state *cpuset_css_alloc(struct cgroup *cont)
 {
-	struct cpuset *cs;
-	struct cpuset *parent;
+	struct cgroup *parent_cg = cont->parent;
+	struct cgroup *tmp_cg;
+	struct cpuset *parent, *cs;
 
-	if (!cont->parent) {
+	if (!parent_cg)
 		return &top_cpuset.css;
-	}
-	parent = cgroup_cs(cont->parent);
+	parent = cgroup_cs(parent_cg);
+
 	cs = kmalloc(sizeof(*cs), GFP_KERNEL);
 	if (!cs)
 		return ERR_PTR(-ENOMEM);
@@ -1855,7 +1819,36 @@ static struct cgroup_subsys_state *cpuset_css_alloc(struct cgroup *cont)
 
 	cs->parent = parent;
 	number_of_cpusets++;
-	return &cs->css ;
+
+	if (!test_bit(CGRP_CPUSET_CLONE_CHILDREN, &cont->flags))
+		goto skip_clone;
+
+	/*
+	 * Clone @parent's configuration if CGRP_CPUSET_CLONE_CHILDREN is
+	 * set.  This flag handling is implemented in cgroup core for
+	 * histrical reasons - the flag may be specified during mount.
+	 *
+	 * Currently, if any sibling cpusets have exclusive cpus or mem, we
+	 * refuse to clone the configuration - thereby refusing the task to
+	 * be entered, and as a result refusing the sys_unshare() or
+	 * clone() which initiated it.  If this becomes a problem for some
+	 * users who wish to allow that scenario, then this could be
+	 * changed to grant parent->cpus_allowed-sibling_cpus_exclusive
+	 * (and likewise for mems) to the new cgroup.
+	 */
+	list_for_each_entry(tmp_cg, &parent_cg->children, sibling) {
+		struct cpuset *tmp_cs = cgroup_cs(tmp_cg);
+
+		if (is_mem_exclusive(tmp_cs) || is_cpu_exclusive(tmp_cs))
+			goto skip_clone;
+	}
+
+	mutex_lock(&callback_mutex);
+	cs->mems_allowed = parent->mems_allowed;
+	cpumask_copy(cs->cpus_allowed, parent->cpus_allowed);
+	mutex_unlock(&callback_mutex);
+skip_clone:
+	return &cs->css;
 }
 
 /*
@@ -1882,7 +1875,6 @@ struct cgroup_subsys cpuset_subsys = {
 	.css_free = cpuset_css_free,
 	.can_attach = cpuset_can_attach,
 	.attach = cpuset_attach,
-	.post_clone = cpuset_post_clone,
 	.subsys_id = cpuset_subsys_id,
 	.base_cftypes = files,
 	.early_init = 1,
-- 
1.7.11.7

Re: [PATCH 1/2] cgroup: s/CGRP_CLONE_CHILDREN/CGRP_CPUSET_CLONE_CHILDREN/

[Serge Hallyn]

Quoting Tejun Heo (tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org):
> From 6e405c1ae4018d813e8ed9e0bd463d6976aebfa8 Mon Sep 17 00:00:00 2001
> From: Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
> Date: Tue, 13 Nov 2012 12:21:50 -0800
> 
> clone_children is only meaningful for cpuset and will stay that way.
> Rename the flag to reflect that and update documentation.  Also, drop
> clone_children() wrapper in cgroup.c.  The thin wrapper is used only a
> few times and one of them will go away soon.
> 
> Signed-off-by: Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>

Thanks.

(both)
Acked-by: Serge E. Hallyn <serge.hallyn-GeWIH/nMZzLQT0dZR+AlfA@public.gmane.org>

> Cc: Glauber Costa <glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
> ---
> These two patches are based on top of "cgroup: allow->post_create() to
> fail" patchset.
> 
>  http://thread.gmane.org/gmane.linux.kernel.cgroups/5047
>  git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git review-cgroup-online-may-fail
> 
> and availalbe in the following branch.
> 
>  git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git review-cgroup-remove-post_clone
> 
> Thanks.
> 
>  Documentation/cgroups/cgroups.txt |  8 +++-----
>  include/linux/cgroup.h            |  6 ++++--
>  kernel/cgroup.c                   | 28 ++++++++++++----------------
>  3 files changed, 19 insertions(+), 23 deletions(-)
> 
> diff --git a/Documentation/cgroups/cgroups.txt b/Documentation/cgroups/cgroups.txt
> index b06eea2..24cdf76 100644
> --- a/Documentation/cgroups/cgroups.txt
> +++ b/Documentation/cgroups/cgroups.txt
> @@ -299,11 +299,9 @@ a cgroup hierarchy's release_agent path is empty.
>  1.5 What does clone_children do ?
>  ---------------------------------
>  
> -If the clone_children flag is enabled (1) in a cgroup, then all
> -cgroups created beneath will call the post_clone callbacks for each
> -subsystem of the newly created cgroup. Usually when this callback is
> -implemented for a subsystem, it copies the values of the parent
> -subsystem, this is the case for the cpuset.
> +This flag only affects the cpuset controller. If the clone_children
> +flag is enabled (1) in a cgroup, a new cpuset cgroup will copy its
> +configuration from the parent during initialization.
>  
>  1.6 How do I use cgroups ?
>  --------------------------
> diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h
> index 89c631d..926d8d1 100644
> --- a/include/linux/cgroup.h
> +++ b/include/linux/cgroup.h
> @@ -143,9 +143,11 @@ enum {
>  	/* Control Group requires release notifications to userspace */
>  	CGRP_NOTIFY_ON_RELEASE,
>  	/*
> -	 * Clone cgroup values when creating a new child cgroup
> +	 * Clone the parent's configuration when creating a new child
> +	 * cpuset cgroup.  For historical reasons, this option can be
> +	 * specified at mount time and thus is implemented here.
>  	 */
> -	CGRP_CLONE_CHILDREN,
> +	CGRP_CPUSET_CLONE_CHILDREN,
>  };
>  
>  struct cgroup {
> diff --git a/kernel/cgroup.c b/kernel/cgroup.c
> index 46c5119..a49572e 100644
> --- a/kernel/cgroup.c
> +++ b/kernel/cgroup.c
> @@ -296,11 +296,6 @@ static int notify_on_release(const struct cgroup *cgrp)
>  	return test_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags);
>  }
>  
> -static int clone_children(const struct cgroup *cgrp)
> -{
> -	return test_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
> -}
> -
>  /*
>   * for_each_subsys() allows you to iterate on each subsystem attached to
>   * an active hierarchy
> @@ -1101,7 +1096,7 @@ static int cgroup_show_options(struct seq_file *seq, struct dentry *dentry)
>  		seq_puts(seq, ",xattr");
>  	if (strlen(root->release_agent_path))
>  		seq_printf(seq, ",release_agent=%s", root->release_agent_path);
> -	if (clone_children(&root->top_cgroup))
> +	if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &root->top_cgroup.flags))
>  		seq_puts(seq, ",clone_children");
>  	if (strlen(root->name))
>  		seq_printf(seq, ",name=%s", root->name);
> @@ -1113,7 +1108,7 @@ struct cgroup_sb_opts {
>  	unsigned long subsys_mask;
>  	unsigned long flags;
>  	char *release_agent;
> -	bool clone_children;
> +	bool cpuset_clone_children;
>  	char *name;
>  	/* User explicitly requested empty subsystem */
>  	bool none;
> @@ -1164,7 +1159,7 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts)
>  			continue;
>  		}
>  		if (!strcmp(token, "clone_children")) {
> -			opts->clone_children = true;
> +			opts->cpuset_clone_children = true;
>  			continue;
>  		}
>  		if (!strcmp(token, "xattr")) {
> @@ -1474,8 +1469,8 @@ static struct cgroupfs_root *cgroup_root_from_opts(struct cgroup_sb_opts *opts)
>  		strcpy(root->release_agent_path, opts->release_agent);
>  	if (opts->name)
>  		strcpy(root->name, opts->name);
> -	if (opts->clone_children)
> -		set_bit(CGRP_CLONE_CHILDREN, &root->top_cgroup.flags);
> +	if (opts->cpuset_clone_children)
> +		set_bit(CGRP_CPUSET_CLONE_CHILDREN, &root->top_cgroup.flags);
>  	return root;
>  }
>  
> @@ -3905,7 +3900,7 @@ fail:
>  static u64 cgroup_clone_children_read(struct cgroup *cgrp,
>  				    struct cftype *cft)
>  {
> -	return clone_children(cgrp);
> +	return test_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
>  }
>  
>  static int cgroup_clone_children_write(struct cgroup *cgrp,
> @@ -3913,9 +3908,9 @@ static int cgroup_clone_children_write(struct cgroup *cgrp,
>  				     u64 val)
>  {
>  	if (val)
> -		set_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
> +		set_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
>  	else
> -		clear_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
> +		clear_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
>  	return 0;
>  }
>  
> @@ -4130,8 +4125,8 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry,
>  	if (notify_on_release(parent))
>  		set_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags);
>  
> -	if (clone_children(parent))
> -		set_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
> +	if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &parent->flags))
> +		set_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
>  
>  	for_each_subsys(root, ss) {
>  		struct cgroup_subsys_state *css;
> @@ -4148,7 +4143,8 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry,
>  				goto err_free_all;
>  		}
>  		/* At error, ->css_free() callback has to free assigned ID. */
> -		if (clone_children(parent) && ss->post_clone)
> +		if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &parent->flags) &&
> +		    ss->post_clone)
>  			ss->post_clone(cgrp);
>  
>  		if (ss->broken_hierarchy && !ss->warned_broken_hierarchy &&
> -- 
> 1.7.11.7
> 
> _______________________________________________
> Containers mailing list
> Containers-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org
> https://lists.linuxfoundation.org/mailman/listinfo/containers

Re: [PATCH 1/2] cgroup: s/CGRP_CLONE_CHILDREN/CGRP_CPUSET_CLONE_CHILDREN/

[Serge Hallyn]

Quoting Tejun Heo (tj@kernel.org):
> From 6e405c1ae4018d813e8ed9e0bd463d6976aebfa8 Mon Sep 17 00:00:00 2001
> From: Tejun Heo <tj@kernel.org>
> Date: Tue, 13 Nov 2012 12:21:50 -0800
> 
> clone_children is only meaningful for cpuset and will stay that way.
> Rename the flag to reflect that and update documentation.  Also, drop
> clone_children() wrapper in cgroup.c.  The thin wrapper is used only a
> few times and one of them will go away soon.
> 
> Signed-off-by: Tejun Heo <tj@kernel.org>

Thanks.

(both)
Acked-by: Serge E. Hallyn <serge.hallyn@ubuntu.com>

> Cc: Glauber Costa <glommer@parallels.com>
> ---
> These two patches are based on top of "cgroup: allow->post_create() to
> fail" patchset.
> 
>  http://thread.gmane.org/gmane.linux.kernel.cgroups/5047
>  git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git review-cgroup-online-may-fail
> 
> and availalbe in the following branch.
> 
>  git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git review-cgroup-remove-post_clone
> 
> Thanks.
> 
>  Documentation/cgroups/cgroups.txt |  8 +++-----
>  include/linux/cgroup.h            |  6 ++++--
>  kernel/cgroup.c                   | 28 ++++++++++++----------------
>  3 files changed, 19 insertions(+), 23 deletions(-)
> 
> diff --git a/Documentation/cgroups/cgroups.txt b/Documentation/cgroups/cgroups.txt
> index b06eea2..24cdf76 100644
> --- a/Documentation/cgroups/cgroups.txt
> +++ b/Documentation/cgroups/cgroups.txt
> @@ -299,11 +299,9 @@ a cgroup hierarchy's release_agent path is empty.
>  1.5 What does clone_children do ?
>  ---------------------------------
>  
> -If the clone_children flag is enabled (1) in a cgroup, then all
> -cgroups created beneath will call the post_clone callbacks for each
> -subsystem of the newly created cgroup. Usually when this callback is
> -implemented for a subsystem, it copies the values of the parent
> -subsystem, this is the case for the cpuset.
> +This flag only affects the cpuset controller. If the clone_children
> +flag is enabled (1) in a cgroup, a new cpuset cgroup will copy its
> +configuration from the parent during initialization.
>  
>  1.6 How do I use cgroups ?
>  --------------------------
> diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h
> index 89c631d..926d8d1 100644
> --- a/include/linux/cgroup.h
> +++ b/include/linux/cgroup.h
> @@ -143,9 +143,11 @@ enum {
>  	/* Control Group requires release notifications to userspace */
>  	CGRP_NOTIFY_ON_RELEASE,
>  	/*
> -	 * Clone cgroup values when creating a new child cgroup
> +	 * Clone the parent's configuration when creating a new child
> +	 * cpuset cgroup.  For historical reasons, this option can be
> +	 * specified at mount time and thus is implemented here.
>  	 */
> -	CGRP_CLONE_CHILDREN,
> +	CGRP_CPUSET_CLONE_CHILDREN,
>  };
>  
>  struct cgroup {
> diff --git a/kernel/cgroup.c b/kernel/cgroup.c
> index 46c5119..a49572e 100644
> --- a/kernel/cgroup.c
> +++ b/kernel/cgroup.c
> @@ -296,11 +296,6 @@ static int notify_on_release(const struct cgroup *cgrp)
>  	return test_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags);
>  }
>  
> -static int clone_children(const struct cgroup *cgrp)
> -{
> -	return test_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
> -}
> -
>  /*
>   * for_each_subsys() allows you to iterate on each subsystem attached to
>   * an active hierarchy
> @@ -1101,7 +1096,7 @@ static int cgroup_show_options(struct seq_file *seq, struct dentry *dentry)
>  		seq_puts(seq, ",xattr");
>  	if (strlen(root->release_agent_path))
>  		seq_printf(seq, ",release_agent=%s", root->release_agent_path);
> -	if (clone_children(&root->top_cgroup))
> +	if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &root->top_cgroup.flags))
>  		seq_puts(seq, ",clone_children");
>  	if (strlen(root->name))
>  		seq_printf(seq, ",name=%s", root->name);
> @@ -1113,7 +1108,7 @@ struct cgroup_sb_opts {
>  	unsigned long subsys_mask;
>  	unsigned long flags;
>  	char *release_agent;
> -	bool clone_children;
> +	bool cpuset_clone_children;
>  	char *name;
>  	/* User explicitly requested empty subsystem */
>  	bool none;
> @@ -1164,7 +1159,7 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts)
>  			continue;
>  		}
>  		if (!strcmp(token, "clone_children")) {
> -			opts->clone_children = true;
> +			opts->cpuset_clone_children = true;
>  			continue;
>  		}
>  		if (!strcmp(token, "xattr")) {
> @@ -1474,8 +1469,8 @@ static struct cgroupfs_root *cgroup_root_from_opts(struct cgroup_sb_opts *opts)
>  		strcpy(root->release_agent_path, opts->release_agent);
>  	if (opts->name)
>  		strcpy(root->name, opts->name);
> -	if (opts->clone_children)
> -		set_bit(CGRP_CLONE_CHILDREN, &root->top_cgroup.flags);
> +	if (opts->cpuset_clone_children)
> +		set_bit(CGRP_CPUSET_CLONE_CHILDREN, &root->top_cgroup.flags);
>  	return root;
>  }
>  
> @@ -3905,7 +3900,7 @@ fail:
>  static u64 cgroup_clone_children_read(struct cgroup *cgrp,
>  				    struct cftype *cft)
>  {
> -	return clone_children(cgrp);
> +	return test_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
>  }
>  
>  static int cgroup_clone_children_write(struct cgroup *cgrp,
> @@ -3913,9 +3908,9 @@ static int cgroup_clone_children_write(struct cgroup *cgrp,
>  				     u64 val)
>  {
>  	if (val)
> -		set_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
> +		set_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
>  	else
> -		clear_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
> +		clear_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
>  	return 0;
>  }
>  
> @@ -4130,8 +4125,8 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry,
>  	if (notify_on_release(parent))
>  		set_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags);
>  
> -	if (clone_children(parent))
> -		set_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
> +	if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &parent->flags))
> +		set_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
>  
>  	for_each_subsys(root, ss) {
>  		struct cgroup_subsys_state *css;
> @@ -4148,7 +4143,8 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry,
>  				goto err_free_all;
>  		}
>  		/* At error, ->css_free() callback has to free assigned ID. */
> -		if (clone_children(parent) && ss->post_clone)
> +		if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &parent->flags) &&
> +		    ss->post_clone)
>  			ss->post_clone(cgrp);
>  
>  		if (ss->broken_hierarchy && !ss->warned_broken_hierarchy &&
> -- 
> 1.7.11.7
> 
> _______________________________________________
> Containers mailing list
> Containers@lists.linux-foundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/containers

Re: [PATCH 1/2] cgroup: s/CGRP_CLONE_CHILDREN/CGRP_CPUSET_CLONE_CHILDREN/

[Serge Hallyn]

Quoting Tejun Heo (tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org):
> From 6e405c1ae4018d813e8ed9e0bd463d6976aebfa8 Mon Sep 17 00:00:00 2001
> From: Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
> Date: Tue, 13 Nov 2012 12:21:50 -0800
> 
> clone_children is only meaningful for cpuset and will stay that way.
> Rename the flag to reflect that and update documentation.  Also, drop
> clone_children() wrapper in cgroup.c.  The thin wrapper is used only a
> few times and one of them will go away soon.
> 
> Signed-off-by: Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>

Thanks.

(both)
Acked-by: Serge E. Hallyn <serge.hallyn-GeWIH/nMZzLQT0dZR+AlfA@public.gmane.org>

> Cc: Glauber Costa <glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
> ---
> These two patches are based on top of "cgroup: allow->post_create() to
> fail" patchset.
> 
>  http://thread.gmane.org/gmane.linux.kernel.cgroups/5047
>  git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git review-cgroup-online-may-fail
> 
> and availalbe in the following branch.
> 
>  git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git review-cgroup-remove-post_clone
> 
> Thanks.
> 
>  Documentation/cgroups/cgroups.txt |  8 +++-----
>  include/linux/cgroup.h            |  6 ++++--
>  kernel/cgroup.c                   | 28 ++++++++++++----------------
>  3 files changed, 19 insertions(+), 23 deletions(-)
> 
> diff --git a/Documentation/cgroups/cgroups.txt b/Documentation/cgroups/cgroups.txt
> index b06eea2..24cdf76 100644
> --- a/Documentation/cgroups/cgroups.txt
> +++ b/Documentation/cgroups/cgroups.txt
> @@ -299,11 +299,9 @@ a cgroup hierarchy's release_agent path is empty.
>  1.5 What does clone_children do ?
>  ---------------------------------
>  
> -If the clone_children flag is enabled (1) in a cgroup, then all
> -cgroups created beneath will call the post_clone callbacks for each
> -subsystem of the newly created cgroup. Usually when this callback is
> -implemented for a subsystem, it copies the values of the parent
> -subsystem, this is the case for the cpuset.
> +This flag only affects the cpuset controller. If the clone_children
> +flag is enabled (1) in a cgroup, a new cpuset cgroup will copy its
> +configuration from the parent during initialization.
>  
>  1.6 How do I use cgroups ?
>  --------------------------
> diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h
> index 89c631d..926d8d1 100644
> --- a/include/linux/cgroup.h
> +++ b/include/linux/cgroup.h
> @@ -143,9 +143,11 @@ enum {
>  	/* Control Group requires release notifications to userspace */
>  	CGRP_NOTIFY_ON_RELEASE,
>  	/*
> -	 * Clone cgroup values when creating a new child cgroup
> +	 * Clone the parent's configuration when creating a new child
> +	 * cpuset cgroup.  For historical reasons, this option can be
> +	 * specified at mount time and thus is implemented here.
>  	 */
> -	CGRP_CLONE_CHILDREN,
> +	CGRP_CPUSET_CLONE_CHILDREN,
>  };
>  
>  struct cgroup {
> diff --git a/kernel/cgroup.c b/kernel/cgroup.c
> index 46c5119..a49572e 100644
> --- a/kernel/cgroup.c
> +++ b/kernel/cgroup.c
> @@ -296,11 +296,6 @@ static int notify_on_release(const struct cgroup *cgrp)
>  	return test_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags);
>  }
>  
> -static int clone_children(const struct cgroup *cgrp)
> -{
> -	return test_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
> -}
> -
>  /*
>   * for_each_subsys() allows you to iterate on each subsystem attached to
>   * an active hierarchy
> @@ -1101,7 +1096,7 @@ static int cgroup_show_options(struct seq_file *seq, struct dentry *dentry)
>  		seq_puts(seq, ",xattr");
>  	if (strlen(root->release_agent_path))
>  		seq_printf(seq, ",release_agent=%s", root->release_agent_path);
> -	if (clone_children(&root->top_cgroup))
> +	if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &root->top_cgroup.flags))
>  		seq_puts(seq, ",clone_children");
>  	if (strlen(root->name))
>  		seq_printf(seq, ",name=%s", root->name);
> @@ -1113,7 +1108,7 @@ struct cgroup_sb_opts {
>  	unsigned long subsys_mask;
>  	unsigned long flags;
>  	char *release_agent;
> -	bool clone_children;
> +	bool cpuset_clone_children;
>  	char *name;
>  	/* User explicitly requested empty subsystem */
>  	bool none;
> @@ -1164,7 +1159,7 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts)
>  			continue;
>  		}
>  		if (!strcmp(token, "clone_children")) {
> -			opts->clone_children = true;
> +			opts->cpuset_clone_children = true;
>  			continue;
>  		}
>  		if (!strcmp(token, "xattr")) {
> @@ -1474,8 +1469,8 @@ static struct cgroupfs_root *cgroup_root_from_opts(struct cgroup_sb_opts *opts)
>  		strcpy(root->release_agent_path, opts->release_agent);
>  	if (opts->name)
>  		strcpy(root->name, opts->name);
> -	if (opts->clone_children)
> -		set_bit(CGRP_CLONE_CHILDREN, &root->top_cgroup.flags);
> +	if (opts->cpuset_clone_children)
> +		set_bit(CGRP_CPUSET_CLONE_CHILDREN, &root->top_cgroup.flags);
>  	return root;
>  }
>  
> @@ -3905,7 +3900,7 @@ fail:
>  static u64 cgroup_clone_children_read(struct cgroup *cgrp,
>  				    struct cftype *cft)
>  {
> -	return clone_children(cgrp);
> +	return test_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
>  }
>  
>  static int cgroup_clone_children_write(struct cgroup *cgrp,
> @@ -3913,9 +3908,9 @@ static int cgroup_clone_children_write(struct cgroup *cgrp,
>  				     u64 val)
>  {
>  	if (val)
> -		set_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
> +		set_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
>  	else
> -		clear_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
> +		clear_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
>  	return 0;
>  }
>  
> @@ -4130,8 +4125,8 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry,
>  	if (notify_on_release(parent))
>  		set_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags);
>  
> -	if (clone_children(parent))
> -		set_bit(CGRP_CLONE_CHILDREN, &cgrp->flags);
> +	if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &parent->flags))
> +		set_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags);
>  
>  	for_each_subsys(root, ss) {
>  		struct cgroup_subsys_state *css;
> @@ -4148,7 +4143,8 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry,
>  				goto err_free_all;
>  		}
>  		/* At error, ->css_free() callback has to free assigned ID. */
> -		if (clone_children(parent) && ss->post_clone)
> +		if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &parent->flags) &&
> +		    ss->post_clone)
>  			ss->post_clone(cgrp);
>  
>  		if (ss->broken_hierarchy && !ss->warned_broken_hierarchy &&
> -- 
> 1.7.11.7
> 
> _______________________________________________
> Containers mailing list
> Containers-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org
> https://lists.linuxfoundation.org/mailman/listinfo/containers

Re: [PATCH 2/2] cgroup, cpuset: remove cgroup_subsys->post_clone()

[Glauber Costa]

On 11/13/2012 09:27 PM, Tejun Heo wrote:
> From bbf2566f9f4fc79ff3320e83cafb69533efc9ea0 Mon Sep 17 00:00:00 2001
> From: Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
> Date: Tue, 13 Nov 2012 12:21:50 -0800
> 
> Currently CGRP_CPUSET_CLONE_CHILDREN triggers ->post_clone().  Now
> that clone_children is cpuset specific, there's no reason to have this
> rather odd option activation mechanism in cgroup core.  cpuset can
> check the flag from its ->css_allocate() and take the necessary
> action.
> 
> Move cpuset_post_clone() logic to the end of cpuset_css_alloc() and
> remove cgroup_subsys->post_clone().
> 
> Loosely based on Glauber's "generalize post_clone into post_create"
> patch.
> 
> Signed-off-by: Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
> Original-patch-by: Glauber Costa <glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
> Original-patch: <1351686554-22592-2-git-send-email-glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
> Cc: Glauber Costa <glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>

I don't have any preference one way or the other. This patch seems to do
the job.

Re: [PATCH 2/2] cgroup, cpuset: remove cgroup_subsys->post_clone()

[Glauber Costa]

On 11/13/2012 09:27 PM, Tejun Heo wrote:
> From bbf2566f9f4fc79ff3320e83cafb69533efc9ea0 Mon Sep 17 00:00:00 2001
> From: Tejun Heo <tj@kernel.org>
> Date: Tue, 13 Nov 2012 12:21:50 -0800
> 
> Currently CGRP_CPUSET_CLONE_CHILDREN triggers ->post_clone().  Now
> that clone_children is cpuset specific, there's no reason to have this
> rather odd option activation mechanism in cgroup core.  cpuset can
> check the flag from its ->css_allocate() and take the necessary
> action.
> 
> Move cpuset_post_clone() logic to the end of cpuset_css_alloc() and
> remove cgroup_subsys->post_clone().
> 
> Loosely based on Glauber's "generalize post_clone into post_create"
> patch.
> 
> Signed-off-by: Tejun Heo <tj@kernel.org>
> Original-patch-by: Glauber Costa <glommer@parallels.com>
> Original-patch: <1351686554-22592-2-git-send-email-glommer@parallels.com>
> Cc: Glauber Costa <glommer@parallels.com>

I don't have any preference one way or the other. This patch seems to do
the job.

Re: [PATCH 2/2] cgroup, cpuset: remove cgroup_subsys->post_clone()

[Glauber Costa]

On 11/13/2012 09:27 PM, Tejun Heo wrote:
> From bbf2566f9f4fc79ff3320e83cafb69533efc9ea0 Mon Sep 17 00:00:00 2001
> From: Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
> Date: Tue, 13 Nov 2012 12:21:50 -0800
> 
> Currently CGRP_CPUSET_CLONE_CHILDREN triggers ->post_clone().  Now
> that clone_children is cpuset specific, there's no reason to have this
> rather odd option activation mechanism in cgroup core.  cpuset can
> check the flag from its ->css_allocate() and take the necessary
> action.
> 
> Move cpuset_post_clone() logic to the end of cpuset_css_alloc() and
> remove cgroup_subsys->post_clone().
> 
> Loosely based on Glauber's "generalize post_clone into post_create"
> patch.
> 
> Signed-off-by: Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
> Original-patch-by: Glauber Costa <glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
> Original-patch: <1351686554-22592-2-git-send-email-glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
> Cc: Glauber Costa <glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>

I don't have any preference one way or the other. This patch seems to do
the job.

Re: [PATCH 1/2] cgroup: s/CGRP_CLONE_CHILDREN/CGRP_CPUSET_CLONE_CHILDREN/

[Li Zefan]

On 2012/11/14 4:27, Tejun Heo wrote:
>>From 6e405c1ae4018d813e8ed9e0bd463d6976aebfa8 Mon Sep 17 00:00:00 2001
> From: Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
> Date: Tue, 13 Nov 2012 12:21:50 -0800
> 
> clone_children is only meaningful for cpuset and will stay that way.
> Rename the flag to reflect that and update documentation.  Also, drop
> clone_children() wrapper in cgroup.c.  The thin wrapper is used only a
> few times and one of them will go away soon.
> 
> Signed-off-by: Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
> Cc: Glauber Costa <glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>

Acked-by: Li Zefan <lizefan-hv44wF8Li93QT0dZR+AlfA@public.gmane.org>

Re: [PATCH 1/2] cgroup: s/CGRP_CLONE_CHILDREN/CGRP_CPUSET_CLONE_CHILDREN/

[Li Zefan]

On 2012/11/14 4:27, Tejun Heo wrote:
>>From 6e405c1ae4018d813e8ed9e0bd463d6976aebfa8 Mon Sep 17 00:00:00 2001
> From: Tejun Heo <tj@kernel.org>
> Date: Tue, 13 Nov 2012 12:21:50 -0800
> 
> clone_children is only meaningful for cpuset and will stay that way.
> Rename the flag to reflect that and update documentation.  Also, drop
> clone_children() wrapper in cgroup.c.  The thin wrapper is used only a
> few times and one of them will go away soon.
> 
> Signed-off-by: Tejun Heo <tj@kernel.org>
> Cc: Glauber Costa <glommer@parallels.com>

Acked-by: Li Zefan <lizefan@huawei.com>

Re: [PATCH 2/2] cgroup, cpuset: remove cgroup_subsys->post_clone()

[Li Zefan]

On 2012/11/14 4:27, Tejun Heo wrote:
>>From bbf2566f9f4fc79ff3320e83cafb69533efc9ea0 Mon Sep 17 00:00:00 2001
> From: Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
> Date: Tue, 13 Nov 2012 12:21:50 -0800
> 
> Currently CGRP_CPUSET_CLONE_CHILDREN triggers ->post_clone().  Now
> that clone_children is cpuset specific, there's no reason to have this
> rather odd option activation mechanism in cgroup core.  cpuset can
> check the flag from its ->css_allocate() and take the necessary
> action.
> 
> Move cpuset_post_clone() logic to the end of cpuset_css_alloc() and
> remove cgroup_subsys->post_clone().
> 
> Loosely based on Glauber's "generalize post_clone into post_create"
> patch.
> 
> Signed-off-by: Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
> Original-patch-by: Glauber Costa <glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
> Original-patch: <1351686554-22592-2-git-send-email-glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
> Cc: Glauber Costa <glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>

Acked-by: Li Zefan <lizefan-hv44wF8Li93QT0dZR+AlfA@public.gmane.org>

Re: [PATCH 2/2] cgroup, cpuset: remove cgroup_subsys->post_clone()

[Li Zefan]

On 2012/11/14 4:27, Tejun Heo wrote:
>>From bbf2566f9f4fc79ff3320e83cafb69533efc9ea0 Mon Sep 17 00:00:00 2001
> From: Tejun Heo <tj@kernel.org>
> Date: Tue, 13 Nov 2012 12:21:50 -0800
> 
> Currently CGRP_CPUSET_CLONE_CHILDREN triggers ->post_clone().  Now
> that clone_children is cpuset specific, there's no reason to have this
> rather odd option activation mechanism in cgroup core.  cpuset can
> check the flag from its ->css_allocate() and take the necessary
> action.
> 
> Move cpuset_post_clone() logic to the end of cpuset_css_alloc() and
> remove cgroup_subsys->post_clone().
> 
> Loosely based on Glauber's "generalize post_clone into post_create"
> patch.
> 
> Signed-off-by: Tejun Heo <tj@kernel.org>
> Original-patch-by: Glauber Costa <glommer@parallels.com>
> Original-patch: <1351686554-22592-2-git-send-email-glommer@parallels.com>
> Cc: Glauber Costa <glommer@parallels.com>

Acked-by: Li Zefan <lizefan@huawei.com>

Re: [PATCH 2/2] cgroup, cpuset: remove cgroup_subsys->post_clone()

[Tejun Heo]

> > Signed-off-by: Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
> > Original-patch-by: Glauber Costa <glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
> > Original-patch: <1351686554-22592-2-git-send-email-glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
> > Cc: Glauber Costa <glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
> 
> Acked-by: Li Zefan <lizefan-hv44wF8Li93QT0dZR+AlfA@public.gmane.org>

Both patches applied to cgroup/for-3.8.  Thanks.

-- 
tejun

Re: [PATCH 2/2] cgroup, cpuset: remove cgroup_subsys->post_clone()

[Tejun Heo]

> > Signed-off-by: Tejun Heo <tj@kernel.org>
> > Original-patch-by: Glauber Costa <glommer@parallels.com>
> > Original-patch: <1351686554-22592-2-git-send-email-glommer@parallels.com>
> > Cc: Glauber Costa <glommer@parallels.com>
> 
> Acked-by: Li Zefan <lizefan@huawei.com>

Both patches applied to cgroup/for-3.8.  Thanks.

-- 
tejun

Re: [PATCH 2/2] cgroup, cpuset: remove cgroup_subsys->post_clone()

[Tejun Heo]

> > Signed-off-by: Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
> > Original-patch-by: Glauber Costa <glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
> > Original-patch: <1351686554-22592-2-git-send-email-glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
> > Cc: Glauber Costa <glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
> 
> Acked-by: Li Zefan <lizefan-hv44wF8Li93QT0dZR+AlfA@public.gmane.org>

Both patches applied to cgroup/for-3.8.  Thanks.

-- 
tejun