cgroup: prepare for the default unified hierarchy

index 306ad0e..ca19735 100644
@@ -3535,7 +3535,7 @@ static void offline_css(struct cgroup_subsys_state *css)
 
 	css->flags &= ~CSS_ONLINE;
 	css->cgroup->nr_css--;
-	RCU_INIT_POINTER(css->cgroup->subsys[ss->id], css);
+	RCU_INIT_POINTER(css->cgroup->subsys[ss->id], NULL);
 }
 
 /**

index ca19735..7bf1459 100644
@@ -3868,6 +3868,15 @@ static int cgroup_destroy_locked(struct cgroup *cgrp)
 		return -EBUSY;
 
 	/*
+	 * Mark @cgrp dead.  This prevents further task migration and child
+	 * creation by disabling cgroup_lock_live_group().  Note that
+	 * CGRP_DEAD assertion is depended upon by css_next_child() to
+	 * resume iteration after dropping RCU read lock.  See
+	 * css_next_child() for details.
+	 */
+	set_bit(CGRP_DEAD, &cgrp->flags);
+
+	/*
 	 * Initiate massacre of all css's.  cgroup_destroy_css_killed()
 	 * will be invoked to perform the rest of destruction once the
 	 * percpu refs of all css's are confirmed to be killed.  This

@@ -3878,15 +3887,6 @@ static int cgroup_destroy_locked(struct cgroup *cgrp)
 		kill_css(css);
 	mutex_lock(&cgroup_mutex);
 
-	/*
-	 * Mark @cgrp dead.  This prevents further task migration and child
-	 * creation by disabling cgroup_lock_live_group().  Note that
-	 * CGRP_DEAD assertion is depended upon by css_next_child() to
-	 * resume iteration after dropping RCU read lock.  See
-	 * css_next_child() for details.
-	 */
-	set_bit(CGRP_DEAD, &cgrp->flags);
-
 	/* CGRP_DEAD is set, remove from ->release_list for the last time */
 	raw_spin_lock(&release_list_lock);
 	if (!list_empty(&cgrp->release_list))

index 7bf1459..12a0bda 100644
@@ -338,16 +338,24 @@ struct cgrp_cset_link {
 	struct list_head	cgrp_link;
 };
 
-/* The default css_set - used by init and its children prior to any
+/*
+ * The default css_set - used by init and its children prior to any
  * hierarchies being mounted. It contains a pointer to the root state
  * for each subsystem. Also used to anchor the list of css_sets. Not
  * reference-counted, to improve performance when child cgroups
  * haven't been created.
  */
+static struct css_set init_css_set = {
+	.refcount		= ATOMIC_INIT(1),
+	.cgrp_links		= LIST_HEAD_INIT(init_css_set.cgrp_links),
+	.tasks			= LIST_HEAD_INIT(init_css_set.tasks),
+	.mg_tasks		= LIST_HEAD_INIT(init_css_set.mg_tasks),
+	.mg_preload_node	= LIST_HEAD_INIT(init_css_set.mg_preload_node),
+	.mg_node		= LIST_HEAD_INIT(init_css_set.mg_node),
+};
 
-static struct css_set init_css_set;
 static struct cgrp_cset_link init_cgrp_cset_link;
-static int css_set_count;
+static int css_set_count	= 1;	/* 1 for init_css_set */
 
 /*
  * hash table for cgroup groups. This improves the performance to find

@@ -1352,7 +1360,8 @@ static void init_cgroup_housekeeping(struct cgroup *cgrp)
 	cgrp->dummy_css.cgroup = cgrp;
 }
 
-static void init_cgroup_root(struct cgroupfs_root *root)
+static void init_cgroup_root(struct cgroupfs_root *root,
+			     struct cgroup_sb_opts *opts)
 {
 	struct cgroup *cgrp = &root->top_cgroup;
 

@@ -1361,20 +1370,6 @@ static void init_cgroup_root(struct cgroupfs_root *root)
 	cgrp->root = root;
 	init_cgroup_housekeeping(cgrp);
 	idr_init(&root->cgroup_idr);
-}
-
-static struct cgroupfs_root *cgroup_root_from_opts(struct cgroup_sb_opts *opts)
-{
-	struct cgroupfs_root *root;
-
-	if (!opts->subsys_mask && !opts->none)
-		return ERR_PTR(-EINVAL);
-
-	root = kzalloc(sizeof(*root), GFP_KERNEL);
-	if (!root)
-		return ERR_PTR(-ENOMEM);
-
-	init_cgroup_root(root);
 
 	root->flags = opts->flags;
 	if (opts->release_agent)

@@ -1383,7 +1378,6 @@ static struct cgroupfs_root *cgroup_root_from_opts(struct cgroup_sb_opts *opts)
 		strcpy(root->name, opts->name);
 	if (opts->cpuset_clone_children)
 		set_bit(CGRP_CPUSET_CLONE_CHILDREN, &root->top_cgroup.flags);
-	return root;
 }
 
 static int cgroup_setup_root(struct cgroupfs_root *root, unsigned long ss_mask)

@@ -1548,13 +1542,24 @@ retry:
 		goto out_unlock;
 	}
 
-	/* no such thing, create a new one */
-	root = cgroup_root_from_opts(&opts);
-	if (IS_ERR(root)) {
-		ret = PTR_ERR(root);
+	/*
+	 * No such thing, create a new one.  name= matching without subsys
+	 * specification is allowed for already existing hierarchies but we
+	 * can't create new one without subsys specification.
+	 */
+	if (!opts.subsys_mask && !opts.none) {
+		ret = -EINVAL;
+		goto out_unlock;
+	}
+
+	root = kzalloc(sizeof(*root), GFP_KERNEL);
+	if (!root) {
+		ret = -ENOMEM;
 		goto out_unlock;
 	}
 
+	init_cgroup_root(root, &opts);
+
 	ret = cgroup_setup_root(root, opts.subsys_mask);
 	if (ret)
 		cgroup_free_root(root);

@@ -4030,26 +4035,13 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss)
  */
 int __init cgroup_init_early(void)
 {
+	static struct cgroup_sb_opts __initdata opts = { };
 	struct cgroup_subsys *ss;
 	int i;
 
-	atomic_set(&init_css_set.refcount, 1);
-	INIT_LIST_HEAD(&init_css_set.cgrp_links);
-	INIT_LIST_HEAD(&init_css_set.tasks);
-	INIT_LIST_HEAD(&init_css_set.mg_tasks);
-	INIT_LIST_HEAD(&init_css_set.mg_preload_node);
-	INIT_LIST_HEAD(&init_css_set.mg_node);
-	INIT_HLIST_NODE(&init_css_set.hlist);
-	css_set_count = 1;
-	init_cgroup_root(&cgroup_dummy_root);
-	cgroup_root_count = 1;
+	init_cgroup_root(&cgroup_dummy_root, &opts);
 	RCU_INIT_POINTER(init_task.cgroups, &init_css_set);
 
-	init_cgrp_cset_link.cset = &init_css_set;
-	init_cgrp_cset_link.cgrp = cgroup_dummy_top;
-	list_add(&init_cgrp_cset_link.cset_link, &cgroup_dummy_top->cset_links);
-	list_add(&init_cgrp_cset_link.cgrp_link, &init_css_set.cgrp_links);
-
 	for_each_subsys(ss, i) {
 		WARN(!ss->css_alloc || !ss->css_free || ss->name || ss->id,
 		     "invalid cgroup_subsys %d:%s css_alloc=%p css_free=%p name:id=%d:%s\n",

@@ -4077,22 +4069,10 @@ int __init cgroup_init(void)
 {
 	struct cgroup_subsys *ss;
 	unsigned long key;
-	int i, err;
+	int ssid, err;
 
 	BUG_ON(cgroup_init_cftypes(NULL, cgroup_base_files));
 
-	for_each_subsys(ss, i) {
-		if (!ss->early_init)
-			cgroup_init_subsys(ss);
-
-		/*
-		 * cftype registration needs kmalloc and can't be done
-		 * during early_init.  Register base cftypes separately.
-		 */
-		if (ss->base_cftypes)
-			WARN_ON(cgroup_add_cftypes(ss, ss->base_cftypes));
-	}
-
 	/* allocate id for the dummy hierarchy */
 	mutex_lock(&cgroup_mutex);
 

@@ -4106,8 +4086,26 @@ int __init cgroup_init(void)
 			0, 1, GFP_KERNEL);
 	BUG_ON(err < 0);
 
+	cgroup_root_count = 1;
+	init_cgrp_cset_link.cset = &init_css_set;
+	init_cgrp_cset_link.cgrp = cgroup_dummy_top;
+	list_add(&init_cgrp_cset_link.cset_link, &cgroup_dummy_top->cset_links);
+	list_add(&init_cgrp_cset_link.cgrp_link, &init_css_set.cgrp_links);
+
 	mutex_unlock(&cgroup_mutex);
 
+	for_each_subsys(ss, ssid) {
+		if (!ss->early_init)
+			cgroup_init_subsys(ss);
+
+		/*
+		 * cftype registration needs kmalloc and can't be done
+		 * during early_init.  Register base cftypes separately.
+		 */
+		if (ss->base_cftypes)
+			WARN_ON(cgroup_add_cftypes(ss, ss->base_cftypes));
+	}
+
 	cgroup_kobj = kobject_create_and_add("cgroup", fs_kobj);
 	if (!cgroup_kobj)
 		return -ENOMEM;

index 7bf1459..12a0bda 100644
@@ -338,16 +338,24 @@ struct cgrp_cset_link {
 	struct list_head	cgrp_link;
 };
 
-/* The default css_set - used by init and its children prior to any
+/*
+ * The default css_set - used by init and its children prior to any
  * hierarchies being mounted. It contains a pointer to the root state
  * for each subsystem. Also used to anchor the list of css_sets. Not
  * reference-counted, to improve performance when child cgroups
  * haven't been created.
  */
+static struct css_set init_css_set = {
+	.refcount		= ATOMIC_INIT(1),
+	.cgrp_links		= LIST_HEAD_INIT(init_css_set.cgrp_links),
+	.tasks			= LIST_HEAD_INIT(init_css_set.tasks),
+	.mg_tasks		= LIST_HEAD_INIT(init_css_set.mg_tasks),
+	.mg_preload_node	= LIST_HEAD_INIT(init_css_set.mg_preload_node),
+	.mg_node		= LIST_HEAD_INIT(init_css_set.mg_node),
+};
 
-static struct css_set init_css_set;
 static struct cgrp_cset_link init_cgrp_cset_link;
-static int css_set_count;
+static int css_set_count	= 1;	/* 1 for init_css_set */
 
 /*
  * hash table for cgroup groups. This improves the performance to find

@@ -1352,7 +1360,8 @@ static void init_cgroup_housekeeping(struct cgroup *cgrp)
 	cgrp->dummy_css.cgroup = cgrp;
 }
 
-static void init_cgroup_root(struct cgroupfs_root *root)
+static void init_cgroup_root(struct cgroupfs_root *root,
+			     struct cgroup_sb_opts *opts)
 {
 	struct cgroup *cgrp = &root->top_cgroup;
 

@@ -1361,20 +1370,6 @@ static void init_cgroup_root(struct cgroupfs_root *root)
 	cgrp->root = root;
 	init_cgroup_housekeeping(cgrp);
 	idr_init(&root->cgroup_idr);
-}
-
-static struct cgroupfs_root *cgroup_root_from_opts(struct cgroup_sb_opts *opts)
-{
-	struct cgroupfs_root *root;
-
-	if (!opts->subsys_mask && !opts->none)
-		return ERR_PTR(-EINVAL);
-
-	root = kzalloc(sizeof(*root), GFP_KERNEL);
-	if (!root)
-		return ERR_PTR(-ENOMEM);
-
-	init_cgroup_root(root);
 
 	root->flags = opts->flags;
 	if (opts->release_agent)

@@ -1383,7 +1378,6 @@ static struct cgroupfs_root *cgroup_root_from_opts(struct cgroup_sb_opts *opts)
 		strcpy(root->name, opts->name);
 	if (opts->cpuset_clone_children)
 		set_bit(CGRP_CPUSET_CLONE_CHILDREN, &root->top_cgroup.flags);
-	return root;
 }
 
 static int cgroup_setup_root(struct cgroupfs_root *root, unsigned long ss_mask)

@@ -1548,13 +1542,24 @@ retry:
 		goto out_unlock;
 	}
 
-	/* no such thing, create a new one */
-	root = cgroup_root_from_opts(&opts);
-	if (IS_ERR(root)) {
-		ret = PTR_ERR(root);
+	/*
+	 * No such thing, create a new one.  name= matching without subsys
+	 * specification is allowed for already existing hierarchies but we
+	 * can't create new one without subsys specification.
+	 */
+	if (!opts.subsys_mask && !opts.none) {
+		ret = -EINVAL;
+		goto out_unlock;
+	}
+
+	root = kzalloc(sizeof(*root), GFP_KERNEL);
+	if (!root) {
+		ret = -ENOMEM;
 		goto out_unlock;
 	}
 
+	init_cgroup_root(root, &opts);
+
 	ret = cgroup_setup_root(root, opts.subsys_mask);
 	if (ret)
 		cgroup_free_root(root);

@@ -4030,26 +4035,13 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss)
  */
 int __init cgroup_init_early(void)
 {
+	static struct cgroup_sb_opts __initdata opts = { };
 	struct cgroup_subsys *ss;
 	int i;
 
-	atomic_set(&init_css_set.refcount, 1);
-	INIT_LIST_HEAD(&init_css_set.cgrp_links);
-	INIT_LIST_HEAD(&init_css_set.tasks);
-	INIT_LIST_HEAD(&init_css_set.mg_tasks);
-	INIT_LIST_HEAD(&init_css_set.mg_preload_node);
-	INIT_LIST_HEAD(&init_css_set.mg_node);
-	INIT_HLIST_NODE(&init_css_set.hlist);
-	css_set_count = 1;
-	init_cgroup_root(&cgroup_dummy_root);
-	cgroup_root_count = 1;
+	init_cgroup_root(&cgroup_dummy_root, &opts);
 	RCU_INIT_POINTER(init_task.cgroups, &init_css_set);
 
-	init_cgrp_cset_link.cset = &init_css_set;
-	init_cgrp_cset_link.cgrp = cgroup_dummy_top;
-	list_add(&init_cgrp_cset_link.cset_link, &cgroup_dummy_top->cset_links);
-	list_add(&init_cgrp_cset_link.cgrp_link, &init_css_set.cgrp_links);
-
 	for_each_subsys(ss, i) {
 		WARN(!ss->css_alloc || !ss->css_free || ss->name || ss->id,
 		     "invalid cgroup_subsys %d:%s css_alloc=%p css_free=%p name:id=%d:%s\n",

@@ -4077,22 +4069,10 @@ int __init cgroup_init(void)
 {
 	struct cgroup_subsys *ss;
 	unsigned long key;
-	int i, err;
+	int ssid, err;
 
 	BUG_ON(cgroup_init_cftypes(NULL, cgroup_base_files));
 
-	for_each_subsys(ss, i) {
-		if (!ss->early_init)
-			cgroup_init_subsys(ss);
-
-		/*
-		 * cftype registration needs kmalloc and can't be done
-		 * during early_init.  Register base cftypes separately.
-		 */
-		if (ss->base_cftypes)
-			WARN_ON(cgroup_add_cftypes(ss, ss->base_cftypes));
-	}
-
 	/* allocate id for the dummy hierarchy */
 	mutex_lock(&cgroup_mutex);
 

@@ -4106,8 +4086,26 @@ int __init cgroup_init(void)
 			0, 1, GFP_KERNEL);
 	BUG_ON(err < 0);
 
+	cgroup_root_count = 1;
+	init_cgrp_cset_link.cset = &init_css_set;
+	init_cgrp_cset_link.cgrp = cgroup_dummy_top;
+	list_add(&init_cgrp_cset_link.cset_link, &cgroup_dummy_top->cset_links);
+	list_add(&init_cgrp_cset_link.cgrp_link, &init_css_set.cgrp_links);
+
 	mutex_unlock(&cgroup_mutex);
 
+	for_each_subsys(ss, ssid) {
+		if (!ss->early_init)
+			cgroup_init_subsys(ss);
+
+		/*
+		 * cftype registration needs kmalloc and can't be done
+		 * during early_init.  Register base cftypes separately.
+		 */
+		if (ss->base_cftypes)
+			WARN_ON(cgroup_add_cftypes(ss, ss->base_cftypes));
+	}
+
 	cgroup_kobj = kobject_create_and_add("cgroup", fs_kobj);
 	if (!cgroup_kobj)
 		return -ENOMEM;

index acbb9a4..9f4f253 100644
@@ -508,39 +508,23 @@ struct cgroup_subsys_state *seq_css(struct seq_file *seq);
 
 static inline int cgroup_name(struct cgroup *cgrp, char *buf, size_t buflen)
 {
-	/* dummy_top doesn't have a kn associated */
-	if (cgrp->kn)
-		return kernfs_name(cgrp->kn, buf, buflen);
-	else
-		return strlcpy(buf, "/", buflen);
+	return kernfs_name(cgrp->kn, buf, buflen);
 }
 
 static inline char * __must_check cgroup_path(struct cgroup *cgrp, char *buf,
 					      size_t buflen)
 {
-	/* dummy_top doesn't have a kn associated */
-	if (cgrp->kn)
-		return kernfs_path(cgrp->kn, buf, buflen);
-	strlcpy(buf, "/", buflen);
-	return (buflen <= 2) ? NULL : buf;
+	return kernfs_path(cgrp->kn, buf, buflen);
 }
 
 static inline void pr_cont_cgroup_name(struct cgroup *cgrp)
 {
-	/* dummy_top doesn't have a kn associated */
-	if (cgrp->kn)
-		pr_cont_kernfs_name(cgrp->kn);
-	else
-		pr_cont("/");
+	pr_cont_kernfs_name(cgrp->kn);
 }
 
 static inline void pr_cont_cgroup_path(struct cgroup *cgrp)
 {
-	/* dummy_top doesn't have a kn associated */
-	if (cgrp->kn)
-		pr_cont_kernfs_path(cgrp->kn);
-	else
-		pr_cont("/");
+	pr_cont_kernfs_path(cgrp->kn);
 }
 
 char *task_cgroup_path(struct task_struct *task, char *buf, size_t buflen);

index 3752a01..77294fc 100644
@@ -24,7 +24,7 @@
 
 #ifdef CONFIG_CGROUPS
 
-struct cgroupfs_root;
+struct cgroup_root;
 struct cgroup_subsys;
 struct inode;
 struct cgroup;

@@ -179,7 +179,7 @@ struct cgroup {
 	/* Private pointers for each registered subsystem */
 	struct cgroup_subsys_state __rcu *subsys[CGROUP_SUBSYS_COUNT];
 
-	struct cgroupfs_root *root;
+	struct cgroup_root *root;
 
 	/*
 	 * List of cgrp_cset_links pointing at css_sets with tasks in this

@@ -211,7 +211,7 @@ struct cgroup {
 
 #define MAX_CGROUP_ROOT_NAMELEN 64
 
-/* cgroupfs_root->flags */
+/* cgroup_root->flags */
 enum {
 	/*
 	 * Unfortunately, cgroup core and various controllers are riddled

@@ -272,18 +272,18 @@ enum {
 };
 
 /*
- * A cgroupfs_root represents the root of a cgroup hierarchy, and may be
+ * A cgroup_root represents the root of a cgroup hierarchy, and may be
  * associated with a kernfs_root to form an active hierarchy.  This is
  * internal to cgroup core.  Don't access directly from controllers.
  */
-struct cgroupfs_root {
+struct cgroup_root {
 	struct kernfs_root *kf_root;
 
 	/* Unique id for this hierarchy. */
 	int hierarchy_id;
 
 	/* The root cgroup.  Root is destroyed on its release. */
-	struct cgroup top_cgroup;
+	struct cgroup cgrp;
 
 	/* Number of cgroups in the hierarchy, used only for /proc/cgroups */
 	atomic_t nr_cgrps;

@@ -598,7 +598,7 @@ struct cgroup_subsys {
 	const char *name;
 
 	/* link to parent, protected by cgroup_lock() */
-	struct cgroupfs_root *root;
+	struct cgroup_root *root;
 
 	/*
 	 * List of cftypes.  Each entry is the first entry of an array

517 lines not on this page · raw message

index 79993ac..7e9fa50 100644
@@ -250,6 +250,9 @@ enum {
 	 *
 	 * - "cgroup.clone_children" is removed.
 	 *
+	 * - If mount is requested with sane_behavior but without any
+	 *   subsystem, the default unified hierarchy is mounted.
+	 *
 	 * - cpuset: tasks will be kept in empty cpusets when hotplug happens
 	 *   and take masks of ancestors with non-empty cpus/mems, instead of
 	 *   being moved to an ancestor.

@@ -468,6 +471,13 @@ struct cftype {
 #endif
 };
 
+extern struct cgroup_root cgrp_dfl_root;
+
+static inline bool cgroup_on_dfl(const struct cgroup *cgrp)
+{
+	return cgrp->root == &cgrp_dfl_root;
+}
+
 /*
  * See the comment above CGRP_ROOT_SANE_BEHAVIOR for details.  This
  * function can be called as long as @cgrp is accessible.

169 lines not on this page · raw message

index 7e9fa50..43d1ed3 100644
@@ -384,6 +384,7 @@ enum {
 	CFTYPE_NOT_ON_ROOT	= (1 << 1),	/* don't create on root cgrp */
 	CFTYPE_INSANE		= (1 << 2),	/* don't create if sane_behavior */
 	CFTYPE_NO_PREFIX	= (1 << 3),	/* (DON'T USE FOR NEW FILES) no subsys prefix */
+	CFTYPE_ONLY_ON_DFL	= (1 << 4),	/* only on default hierarchy */
 };
 
 #define MAX_CFTYPE_NAME		64

index c42230c..79e3573 100644
@@ -2356,6 +2356,8 @@ static int cgroup_addrm_files(struct cgroup *cgrp, struct cftype cfts[],
 
 	for (cft = cfts; cft->name[0] != '\0'; cft++) {
 		/* does cft->flags tell us to skip this file on @cgrp? */
+		if ((cft->flags & CFTYPE_ONLY_ON_DFL) && !cgroup_on_dfl(cgrp))
+			continue;
 		if ((cft->flags & CFTYPE_INSANE) && cgroup_sane_behavior(cgrp))
 			continue;
 		if ((cft->flags & CFTYPE_NOT_ON_ROOT) && !cgrp->parent)

index 12a0bda..2483f4e 100644
@@ -289,8 +289,8 @@ static int notify_on_release(const struct cgroup *cgrp)
 	for ((ssid) = 0; (ssid) < CGROUP_SUBSYS_COUNT &&		\
 	     (((ss) = cgroup_subsys[ssid]) || true); (ssid)++)
 
-/* iterate across the active hierarchies */
-#define for_each_active_root(root)					\
+/* iterate across the hierarchies */
+#define for_each_root(root)						\
 	list_for_each_entry((root), &cgroup_roots, root_list)
 
 /**

@@ -354,7 +354,6 @@ static struct css_set init_css_set = {
 	.mg_node		= LIST_HEAD_INIT(init_css_set.mg_node),
 };
 
-static struct cgrp_cset_link init_cgrp_cset_link;
 static int css_set_count	= 1;	/* 1 for init_css_set */
 
 /*

@@ -693,14 +692,13 @@ static struct cgroupfs_root *cgroup_root_from_kf(struct kernfs_root *kf_root)
 	return top_cgrp->root;
 }
 
-static int cgroup_init_root_id(struct cgroupfs_root *root, int start, int end)
+static int cgroup_init_root_id(struct cgroupfs_root *root)
 {
 	int id;
 
 	lockdep_assert_held(&cgroup_mutex);
 
-	id = idr_alloc_cyclic(&cgroup_hierarchy_idr, root, start, end,
-			      GFP_KERNEL);
+	id = idr_alloc_cyclic(&cgroup_hierarchy_idr, root, 0, 0, GFP_KERNEL);
 	if (id < 0)
 		return id;
 

@@ -1405,8 +1403,7 @@ static int cgroup_setup_root(struct cgroupfs_root *root, unsigned long ss_mask)
 	if (ret)
 		goto out;
 
-	/* ID 0 is reserved for dummy root, 1 for unified hierarchy */
-	ret = cgroup_init_root_id(root, 2, 0);
+	ret = cgroup_init_root_id(root);
 	if (ret)
 		goto out;
 

@@ -1486,9 +1483,12 @@ retry:
 		goto out_unlock;
 
 	/* look for a matching existing root */
-	for_each_active_root(root) {
+	for_each_root(root) {
 		bool name_match = false;
 
+		if (root == &cgroup_dummy_root)
+			continue;
+
 		/*
 		 * If we asked for a name then it must match.  Also, if
 		 * name matches but sybsys_mask doesn't, we should fail.

@@ -2106,9 +2106,12 @@ int cgroup_attach_task_all(struct task_struct *from, struct task_struct *tsk)
 	int retval = 0;
 
 	mutex_lock(&cgroup_mutex);
-	for_each_active_root(root) {
+	for_each_root(root) {
 		struct cgroup *from_cgrp;
 
+		if (root == &cgroup_dummy_root)
+			continue;
+
 		down_read(&css_set_rwsem);
 		from_cgrp = task_cgroup_from_root(from, root);
 		up_read(&css_set_rwsem);

@@ -4073,26 +4076,17 @@ int __init cgroup_init(void)
 
 	BUG_ON(cgroup_init_cftypes(NULL, cgroup_base_files));
 
-	/* allocate id for the dummy hierarchy */
+	mutex_lock(&cgroup_tree_mutex);
 	mutex_lock(&cgroup_mutex);
 
 	/* Add init_css_set to the hash table */
 	key = css_set_hash(init_css_set.subsys);
 	hash_add(css_set_table, &init_css_set.hlist, key);
 
-	BUG_ON(cgroup_init_root_id(&cgroup_dummy_root, 0, 1));
-
-	err = idr_alloc(&cgroup_dummy_root.cgroup_idr, cgroup_dummy_top,
-			0, 1, GFP_KERNEL);
-	BUG_ON(err < 0);
-
-	cgroup_root_count = 1;
-	init_cgrp_cset_link.cset = &init_css_set;
-	init_cgrp_cset_link.cgrp = cgroup_dummy_top;
-	list_add(&init_cgrp_cset_link.cset_link, &cgroup_dummy_top->cset_links);
-	list_add(&init_cgrp_cset_link.cgrp_link, &init_css_set.cgrp_links);
+	BUG_ON(cgroup_setup_root(&cgroup_dummy_root, 0));
 
 	mutex_unlock(&cgroup_mutex);
+	mutex_unlock(&cgroup_tree_mutex);
 
 	for_each_subsys(ss, ssid) {
 		if (!ss->early_init)

@@ -4176,11 +4170,14 @@ int proc_cgroup_show(struct seq_file *m, void *v)
 	mutex_lock(&cgroup_mutex);
 	down_read(&css_set_rwsem);
 
-	for_each_active_root(root) {
+	for_each_root(root) {
 		struct cgroup_subsys *ss;
 		struct cgroup *cgrp;
 		int ssid, count = 0;
 
+		if (root == &cgroup_dummy_root)
+			continue;
+
 		seq_printf(m, "%d:", root->hierarchy_id);
 		for_each_subsys(ss, ssid)
 			if (root->subsys_mask & (1 << ssid))

index acbb9a4..9f4f253 100644
@@ -508,39 +508,23 @@ struct cgroup_subsys_state *seq_css(struct seq_file *seq);
 
 static inline int cgroup_name(struct cgroup *cgrp, char *buf, size_t buflen)
 {
-	/* dummy_top doesn't have a kn associated */
-	if (cgrp->kn)
-		return kernfs_name(cgrp->kn, buf, buflen);
-	else
-		return strlcpy(buf, "/", buflen);
+	return kernfs_name(cgrp->kn, buf, buflen);
 }
 
 static inline char * __must_check cgroup_path(struct cgroup *cgrp, char *buf,
 					      size_t buflen)
 {
-	/* dummy_top doesn't have a kn associated */
-	if (cgrp->kn)
-		return kernfs_path(cgrp->kn, buf, buflen);
-	strlcpy(buf, "/", buflen);
-	return (buflen <= 2) ? NULL : buf;
+	return kernfs_path(cgrp->kn, buf, buflen);
 }
 
 static inline void pr_cont_cgroup_name(struct cgroup *cgrp)
 {
-	/* dummy_top doesn't have a kn associated */
-	if (cgrp->kn)
-		pr_cont_kernfs_name(cgrp->kn);
-	else
-		pr_cont("/");
+	pr_cont_kernfs_name(cgrp->kn);
 }
 
 static inline void pr_cont_cgroup_path(struct cgroup *cgrp)
 {
-	/* dummy_top doesn't have a kn associated */
-	if (cgrp->kn)
-		pr_cont_kernfs_path(cgrp->kn);
-	else
-		pr_cont("/");
+	pr_cont_kernfs_path(cgrp->kn);
 }
 
 char *task_cgroup_path(struct task_struct *task, char *buf, size_t buflen);

index 2483f4e..9b9a294 100644
@@ -175,8 +175,8 @@ static int need_forkexit_callback __read_mostly;
 static struct cftype cgroup_base_files[];
 
 static void cgroup_put(struct cgroup *cgrp);
-static int rebind_subsystems(struct cgroupfs_root *root,
-			     unsigned long added_mask, unsigned removed_mask);
+static int rebind_subsystems(struct cgroupfs_root *dst_root,
+			     unsigned long ss_mask);
 static void cgroup_destroy_css_killed(struct cgroup *cgrp);
 static int cgroup_destroy_locked(struct cgroup *cgrp);
 static int cgroup_addrm_files(struct cgroup *cgrp, struct cftype cfts[],

@@ -739,7 +739,7 @@ static void cgroup_destroy_root(struct cgroupfs_root *root)
 	BUG_ON(!list_empty(&cgrp->children));
 
 	/* Rebind all subsystems back to the default hierarchy */
-	WARN_ON(rebind_subsystems(root, 0, root->subsys_mask));
+	rebind_subsystems(&cgroup_dummy_root, root->subsys_mask);
 
 	/*
 	 * Release all the links from cset_links to this hierarchy's

@@ -976,69 +976,77 @@ static void cgroup_clear_dir(struct cgroup *cgrp, unsigned long subsys_mask)
 	}
 }
 
-static int rebind_subsystems(struct cgroupfs_root *root,
-			     unsigned long added_mask, unsigned removed_mask)
+static int rebind_subsystems(struct cgroupfs_root *dst_root,
+			     unsigned long ss_mask)
 {
-	struct cgroup *cgrp = &root->top_cgroup;
+	struct cgroup *dst_top = &dst_root->top_cgroup;
 	struct cgroup_subsys *ss;
-	int i, ret;
+	int ssid, ret;
 
 	lockdep_assert_held(&cgroup_tree_mutex);
 	lockdep_assert_held(&cgroup_mutex);
 
-	/* Check that any added subsystems are currently free */
-	for_each_subsys(ss, i)
-		if ((added_mask & (1 << i)) && ss->root != &cgroup_dummy_root)
+	for_each_subsys(ss, ssid) {
+		if (!(ss_mask & (1 << ssid)))
+			continue;
+
+		/* if @ss is on the dummy_root, we can always move it */
+		if (ss->root == &cgroup_dummy_root)
+			continue;
+
+		/* if @ss has non-root cgroups attached to it, can't move */
+		if (!list_empty(&ss->root->top_cgroup.children))
 			return -EBUSY;
 
-	ret = cgroup_populate_dir(cgrp, added_mask);
-	if (ret)
-		return ret;
+		/* can't move between two non-dummy roots either */
+		if (dst_root != &cgroup_dummy_root)
+			return -EBUSY;
+	}
+
+	if (dst_root != &cgroup_dummy_root) {
+		ret = cgroup_populate_dir(dst_top, ss_mask);
+		if (ret)
+			return ret;
+	}
 
 	/*
 	 * Nothing can fail from this point on.  Remove files for the
 	 * removed subsystems and rebind each subsystem.
 	 */
 	mutex_unlock(&cgroup_mutex);
-	cgroup_clear_dir(cgrp, removed_mask);
+	for_each_subsys(ss, ssid)
+		if ((ss_mask & (1 << ssid)) && ss->root != &cgroup_dummy_root)
+			cgroup_clear_dir(&ss->root->top_cgroup, 1 << ssid);
 	mutex_lock(&cgroup_mutex);
 
-	for_each_subsys(ss, i) {
-		unsigned long bit = 1UL << i;
-
-		if (bit & added_mask) {
-			/* We're binding this subsystem to this hierarchy */
-			BUG_ON(cgroup_css(cgrp, ss));
-			BUG_ON(!cgroup_css(cgroup_dummy_top, ss));
-			BUG_ON(cgroup_css(cgroup_dummy_top, ss)->cgroup != cgroup_dummy_top);
+	for_each_subsys(ss, ssid) {
+		struct cgroupfs_root *src_root;
+		struct cgroup *src_top;
+		struct cgroup_subsys_state *css;
 
-			rcu_assign_pointer(cgrp->subsys[i],
-					   cgroup_css(cgroup_dummy_top, ss));
-			cgroup_css(cgrp, ss)->cgroup = cgrp;
+		if (!(ss_mask & (1 << ssid)))
+			continue;
 
-			ss->root = root;
-			if (ss->bind)
-				ss->bind(cgroup_css(cgrp, ss));
+		src_root = ss->root;
+		src_top = &src_root->top_cgroup;
+		css = cgroup_css(src_top, ss);
 
-			/* refcount was already taken, and we're keeping it */
-			root->subsys_mask |= bit;
-		} else if (bit & removed_mask) {
-			/* We're removing this subsystem */
-			BUG_ON(cgroup_css(cgrp, ss) != cgroup_css(cgroup_dummy_top, ss));
-			BUG_ON(cgroup_css(cgrp, ss)->cgroup != cgrp);
+		WARN_ON(!css || cgroup_css(dst_top, ss));
 
-			if (ss->bind)
-				ss->bind(cgroup_css(cgroup_dummy_top, ss));
+		RCU_INIT_POINTER(src_top->subsys[ssid], NULL);
+		rcu_assign_pointer(dst_top->subsys[ssid], css);
+		ss->root = dst_root;
+		css->cgroup = dst_top;
 
-			cgroup_css(cgroup_dummy_top, ss)->cgroup = cgroup_dummy_top;
-			RCU_INIT_POINTER(cgrp->subsys[i], NULL);
+		src_root->subsys_mask &= ~(1 << ssid);
+		dst_root->subsys_mask |= 1 << ssid;
 
-			cgroup_subsys[i]->root = &cgroup_dummy_root;
-			root->subsys_mask &= ~bit;
-		}
+		if (ss->bind)
+			ss->bind(css);
 	}
 
-	kernfs_activate(cgrp->kn);
+	if (dst_root != &cgroup_dummy_root)
+		kernfs_activate(dst_top->kn);
 	return 0;
 }
 

@@ -1277,10 +1285,12 @@ static int cgroup_remount(struct kernfs_root *kf_root, int *flags, char *data)
 		goto out_unlock;
 	}
 
-	ret = rebind_subsystems(root, added_mask, removed_mask);
+	ret = rebind_subsystems(root, added_mask);
 	if (ret)
 		goto out_unlock;
 
+	rebind_subsystems(&cgroup_dummy_root, removed_mask);
+
 	if (opts.release_agent) {
 		spin_lock(&release_agent_path_lock);
 		strcpy(root->release_agent_path, opts.release_agent);

@@ -1420,7 +1430,7 @@ static int cgroup_setup_root(struct cgroupfs_root *root, unsigned long ss_mask)
 	if (ret)
 		goto destroy_root;
 
-	ret = rebind_subsystems(root, ss_mask, 0);
+	ret = rebind_subsystems(root, ss_mask);
 	if (ret)
 		goto destroy_root;
 

@@ -4026,6 +4036,8 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss)
 
 	BUG_ON(online_css(css));
 
+	cgroup_dummy_root.subsys_mask |= 1 << ss->id;
+
 	mutex_unlock(&cgroup_mutex);
 	mutex_unlock(&cgroup_tree_mutex);
 }

index 9f4f253..3752a01 100644
@@ -173,6 +173,9 @@ struct cgroup {
 	 */
 	u64 serial_nr;
 
+	/* The bitmask of subsystems attached to this cgroup */
+	unsigned long subsys_mask;
+
 	/* Private pointers for each registered subsystem */
 	struct cgroup_subsys_state __rcu *subsys[CGROUP_SUBSYS_COUNT];
 

@@ -276,9 +279,6 @@ enum {
 struct cgroupfs_root {
 	struct kernfs_root *kf_root;
 
-	/* The bitmask of subsystems attached to this hierarchy */
-	unsigned long subsys_mask;
-
 	/* Unique id for this hierarchy. */
 	int hierarchy_id;
 

137 lines not on this page · raw message

index 3752a01..77294fc 100644
@@ -24,7 +24,7 @@
 
 #ifdef CONFIG_CGROUPS
 
-struct cgroupfs_root;
+struct cgroup_root;
 struct cgroup_subsys;
 struct inode;
 struct cgroup;

@@ -179,7 +179,7 @@ struct cgroup {
 	/* Private pointers for each registered subsystem */
 	struct cgroup_subsys_state __rcu *subsys[CGROUP_SUBSYS_COUNT];
 
-	struct cgroupfs_root *root;
+	struct cgroup_root *root;
 
 	/*
 	 * List of cgrp_cset_links pointing at css_sets with tasks in this

@@ -211,7 +211,7 @@ struct cgroup {
 
 #define MAX_CGROUP_ROOT_NAMELEN 64
 
-/* cgroupfs_root->flags */
+/* cgroup_root->flags */
 enum {
 	/*
 	 * Unfortunately, cgroup core and various controllers are riddled

@@ -272,18 +272,18 @@ enum {
 };
 
 /*
- * A cgroupfs_root represents the root of a cgroup hierarchy, and may be
+ * A cgroup_root represents the root of a cgroup hierarchy, and may be
  * associated with a kernfs_root to form an active hierarchy.  This is
  * internal to cgroup core.  Don't access directly from controllers.
  */
-struct cgroupfs_root {
+struct cgroup_root {
 	struct kernfs_root *kf_root;
 
 	/* Unique id for this hierarchy. */
 	int hierarchy_id;
 
 	/* The root cgroup.  Root is destroyed on its release. */
-	struct cgroup top_cgroup;
+	struct cgroup cgrp;
 
 	/* Number of cgroups in the hierarchy, used only for /proc/cgroups */
 	atomic_t nr_cgrps;

@@ -598,7 +598,7 @@ struct cgroup_subsys {
 	const char *name;
 
 	/* link to parent, protected by cgroup_lock() */
-	struct cgroupfs_root *root;
+	struct cgroup_root *root;
 
 	/*
 	 * List of cftypes.  Each entry is the first entry of an array

517 lines not on this page · raw message

index 861c363..033745c 100644
@@ -1408,13 +1408,13 @@ static int tg_set_conf(struct cgroup_subsys_state *css, struct cftype *cft,
 }
 
 static int tg_set_conf_u64(struct cgroup_subsys_state *css, struct cftype *cft,
-			   const char *buf)
+			   char *buf)
 {
 	return tg_set_conf(css, cft, buf, true);
 }
 
 static int tg_set_conf_uint(struct cgroup_subsys_state *css, struct cftype *cft,
-			    const char *buf)
+			    char *buf)
 {
 	return tg_set_conf(css, cft, buf, false);
 }

index 4611879..f5de45b 100644
@@ -1701,13 +1701,13 @@ static int __cfqg_set_weight_device(struct cgroup_subsys_state *css,
 }
 
 static int cfqg_set_weight_device(struct cgroup_subsys_state *css,
-				  struct cftype *cft, const char *buf)
+				  struct cftype *cft, char *buf)
 {
 	return __cfqg_set_weight_device(css, cft, buf, false);
 }
 
 static int cfqg_set_leaf_weight_device(struct cgroup_subsys_state *css,
-				       struct cftype *cft, const char *buf)
+				       struct cftype *cft, char *buf)
 {
 	return __cfqg_set_weight_device(css, cft, buf, true);
 }

index 77294fc..79993ac 100644
@@ -454,7 +454,7 @@ struct cftype {
 	 * Returns 0 or -ve error code.
 	 */
 	int (*write_string)(struct cgroup_subsys_state *css, struct cftype *cft,
-			    const char *buffer);
+			    char *buffer);
 	/*
 	 * trigger() callback can be used to get some kick from the
 	 * userspace, when the actual string written is not important

index 04f6262..36d1a53 100644
@@ -2143,7 +2143,7 @@ static int cgroup_procs_write(struct cgroup_subsys_state *css,
 }
 
 static int cgroup_release_agent_write(struct cgroup_subsys_state *css,
-				      struct cftype *cft, const char *buffer)
+				      struct cftype *cft, char *buffer)
 {
 	struct cgroup_root *root = css->cgroup->root;
 

index 2ea98b2..2bc4a22 100644
@@ -442,7 +442,7 @@ static void freezer_change_state(struct freezer *freezer, bool freeze)
 }
 
 static int freezer_write(struct cgroup_subsys_state *css, struct cftype *cft,
-			 const char *buffer)
+			 char *buffer)
 {
 	bool freeze;
 

index 8d53245..efbf9ba 100644
@@ -1610,7 +1610,7 @@ out_unlock:
  * Common handling for a write to a "cpus" or "mems" file.
  */
 static int cpuset_write_resmask(struct cgroup_subsys_state *css,
-				struct cftype *cft, const char *buf)
+				struct cftype *cft, char *buf)
 {
 	struct cpuset *cs = css_cs(css);
 	struct cpuset *trialcs;

index b135853..595d7fd 100644
@@ -254,7 +254,7 @@ static u64 hugetlb_cgroup_read_u64(struct cgroup_subsys_state *css,
 }
 
 static int hugetlb_cgroup_write(struct cgroup_subsys_state *css,
-				struct cftype *cft, const char *buffer)
+				struct cftype *cft, char *buffer)
 {
 	int idx, name, ret;
 	unsigned long long val;

index d9c6ac1..96f94a9 100644
@@ -5242,7 +5242,7 @@ static int memcg_update_kmem_limit(struct mem_cgroup *memcg,
  * RES_LIMIT.
  */
 static int mem_cgroup_write(struct cgroup_subsys_state *css, struct cftype *cft,
-			    const char *buffer)
+			    char *buffer)
 {
 	struct mem_cgroup *memcg = mem_cgroup_from_css(css);
 	enum res_type type;

@@ -6063,7 +6063,7 @@ static void memcg_event_ptable_queue_proc(struct file *file,
  * Interpretation of args is defined by control file implementation.
  */
 static int memcg_write_event_control(struct cgroup_subsys_state *css,
-				     struct cftype *cft, const char *buffer)
+				     struct cftype *cft, char *buffer)
 {
 	struct mem_cgroup *memcg = mem_cgroup_from_css(css);
 	struct mem_cgroup_event *event;

index f9f3a40..3825f66 100644
@@ -186,7 +186,7 @@ static int read_priomap(struct seq_file *sf, void *v)
 }
 
 static int write_priomap(struct cgroup_subsys_state *css, struct cftype *cft,
-			 const char *buffer)
+			 char *buffer)
 {
 	char devname[IFNAMSIZ + 1];
 	struct net_device *dev;

index 20a0aca..d4f015a 100644
@@ -103,7 +103,7 @@ static int tcp_update_limit(struct mem_cgroup *memcg, u64 val)
 }
 
 static int tcp_cgroup_write(struct cgroup_subsys_state *css, struct cftype *cft,
-			    const char *buffer)
+			    char *buffer)
 {
 	struct mem_cgroup *memcg = mem_cgroup_from_css(css);
 	unsigned long long val;

index 7f88bcd..8365909 100644
@@ -496,7 +496,7 @@ static inline bool has_children(struct dev_cgroup *devcgroup)
  * parent cgroup has the access you're asking for.
  */
 static int devcgroup_update_access(struct dev_cgroup *devcgroup,
-				   int filetype, const char *buffer)
+				   int filetype, char *buffer)
 {
 	const char *b;
 	char temp[12];		/* 11 + 1 characters needed for a u32 */

@@ -652,7 +652,7 @@ static int devcgroup_update_access(struct dev_cgroup *devcgroup,
 }
 
 static int devcgroup_access_write(struct cgroup_subsys_state *css,
-				  struct cftype *cft, const char *buffer)
+				  struct cftype *cft, char *buffer)
 {
 	int retval;
 

index 79993ac..7e9fa50 100644
@@ -250,6 +250,9 @@ enum {
 	 *
 	 * - "cgroup.clone_children" is removed.
 	 *
+	 * - If mount is requested with sane_behavior but without any
+	 *   subsystem, the default unified hierarchy is mounted.
+	 *
 	 * - cpuset: tasks will be kept in empty cpusets when hotplug happens
 	 *   and take masks of ancestors with non-empty cpus/mems, instead of
 	 *   being moved to an ancestor.

@@ -468,6 +471,13 @@ struct cftype {
 #endif
 };
 
+extern struct cgroup_root cgrp_dfl_root;
+
+static inline bool cgroup_on_dfl(const struct cgroup *cgrp)
+{
+	return cgrp->root == &cgrp_dfl_root;
+}
+
 /*
  * See the comment above CGRP_ROOT_SANE_BEHAVIOR for details.  This
  * function can be called as long as @cgrp is accessible.

169 lines not on this page · raw message

index 7e9fa50..43d1ed3 100644
@@ -384,6 +384,7 @@ enum {
 	CFTYPE_NOT_ON_ROOT	= (1 << 1),	/* don't create on root cgrp */
 	CFTYPE_INSANE		= (1 << 2),	/* don't create if sane_behavior */
 	CFTYPE_NO_PREFIX	= (1 << 3),	/* (DON'T USE FOR NEW FILES) no subsys prefix */
+	CFTYPE_ONLY_ON_DFL	= (1 << 4),	/* only on default hierarchy */
 };
 
 #define MAX_CFTYPE_NAME		64

index c42230c..79e3573 100644
@@ -2356,6 +2356,8 @@ static int cgroup_addrm_files(struct cgroup *cgrp, struct cftype cfts[],
 
 	for (cft = cfts; cft->name[0] != '\0'; cft++) {
 		/* does cft->flags tell us to skip this file on @cgrp? */
+		if ((cft->flags & CFTYPE_ONLY_ON_DFL) && !cgroup_on_dfl(cgrp))
+			continue;
 		if ((cft->flags & CFTYPE_INSANE) && cgroup_sane_behavior(cgrp))
 			continue;
 		if ((cft->flags & CFTYPE_NOT_ON_ROOT) && !cgrp->parent)