cgroup: replace __DEVEL__sane_behavior with cgroup2 fs type

With major controllers - cpu, memory and io - shaping up for the
unified hierarchy, cgroup2 is about ready to be, gradually, released
into the wild.  Replace __DEVEL__sane_behavior flag which was used to
select the unified hierarchy with a separate filesystem type "cgroup2"
so that unified hierarchy can be mounted as follows.

  mount -t cgroup2 none $MOUNT_POINT

The cgroup2 fs has its own magic number - 0x63677270 ("cgrp").

v2: Assign a different magic number to cgroup2 fs.

Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
(cherry picked from commit 67e9c74b8a873408c27ac9a8e4c1d1c8d72c93ff)
Signed-off-by: Alex Shi <alex.shi@linaro.org>

Author: htejun

Documentation/cgroups/unified-hierarchy.txt

@@ -94,11 +94,9 @@ the process.
 
 2-1. Mounting
 
-Currently, unified hierarchy can be mounted with the following mount
-command.  Note that this is still under development and scheduled to
-change soon.
+Unified hierarchy can be mounted with the following mount command.
 
- mount -t cgroup -o __DEVEL__sane_behavior cgroup $MOUNT_POINT
+ mount -t cgroup2 none $MOUNT_POINT
 
 All controllers which support the unified hierarchy and are not bound
 to other hierarchies are automatically bound to unified hierarchy and

include/linux/cgroup-defs.h

@@ -66,7 +66,6 @@ enum {
 
 /* cgroup_root->flags */
 enum {
-	CGRP_ROOT_SANE_BEHAVIOR	= (1 << 0), /* __DEVEL__sane_behavior specified */
 	CGRP_ROOT_NOPREFIX	= (1 << 1), /* mounted subsystems have no named prefix */
 	CGRP_ROOT_XATTR		= (1 << 2), /* supports extended attributes */
 };

include/uapi/linux/magic.h

@@ -54,6 +54,7 @@
 
 #define SMB_SUPER_MAGIC		0x517B
 #define CGROUP_SUPER_MAGIC	0x27e0eb
+#define CGROUP2_SUPER_MAGIC	0x63677270
 
 
 #define STACK_END_MAGIC		0x57AC6E9D

kernel/cgroup.c

@@ -211,6 +211,7 @@ static unsigned long have_free_callback __read_mostly;
 /* Ditto for the can_fork callback. */
 static unsigned long have_canfork_callback __read_mostly;
 
+static struct file_system_type cgroup2_fs_type;
 static struct cftype cgroup_dfl_base_files[];
 static struct cftype cgroup_legacy_base_files[];
 
@@ -1650,10 +1651,6 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts)
 			all_ss = true;
 			continue;
 		}
-		if (!strcmp(token, "__DEVEL__sane_behavior")) {
-			opts->flags |= CGRP_ROOT_SANE_BEHAVIOR;
-			continue;
-		}
 		if (!strcmp(token, "noprefix")) {
 			opts->flags |= CGRP_ROOT_NOPREFIX;
 			continue;
@@ -1720,15 +1717,6 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts)
 			return -ENOENT;
 	}
 
-	if (opts->flags & CGRP_ROOT_SANE_BEHAVIOR) {
-		pr_warn("sane_behavior: this is still under development and its behaviors will change, proceed at your own risk\n");
-		if (nr_opts != 1) {
-			pr_err("sane_behavior: no other mount options allowed\n");
-			return -EINVAL;
-		}
-		return 0;
-	}
-
 	/*
 	 * If the 'all' option was specified select all the subsystems,
 	 * otherwise if 'none', 'name=' and a subsystem name options were
@@ -2007,6 +1995,7 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type,
 			 int flags, const char *unused_dev_name,
 			 void *data)
 {
+	bool is_v2 = fs_type == &cgroup2_fs_type;
 	struct super_block *pinned_sb = NULL;
 	struct cgroup_subsys *ss;
 	struct cgroup_root *root;
@@ -2023,22 +2012,24 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type,
 	if (!use_task_css_set_links)
 		cgroup_enable_task_cg_lists();
 
+	if (is_v2) {
+		if (data) {
+			pr_err("cgroup2: unknown option \"%s\"\n", (char *)data);
+			return ERR_PTR(-EINVAL);
+		}
+		cgrp_dfl_root_visible = true;
+		root = &cgrp_dfl_root;
+		cgroup_get(&root->cgrp);
+		goto out_mount;
+	}
+
 	mutex_lock(&cgroup_mutex);
 
 	/* First find the desired set of subsystems */
 	ret = parse_cgroupfs_options(data, &opts);
 	if (ret)
 		goto out_unlock;
 
-	/* look for a matching existing root */
-	if (opts.flags & CGRP_ROOT_SANE_BEHAVIOR) {
-		cgrp_dfl_root_visible = true;
-		root = &cgrp_dfl_root;
-		cgroup_get(&root->cgrp);
-		ret = 0;
-		goto out_unlock;
-	}
-
 	/*
 	 * Destruction of cgroup root is asynchronous, so subsystems may
 	 * still be dying after the previous unmount.  Let's drain the
@@ -2149,9 +2140,10 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type,
 
 	if (ret)
 		return ERR_PTR(ret);
-
+out_mount:
 	dentry = kernfs_mount(fs_type, flags, root->kf_root,
-				CGROUP_SUPER_MAGIC, &new_sb);
+			      is_v2 ? CGROUP2_SUPER_MAGIC : CGROUP_SUPER_MAGIC,
+			      &new_sb);
 	if (IS_ERR(dentry) || !new_sb)
 		cgroup_put(&root->cgrp);
 
@@ -2194,6 +2186,12 @@ static struct file_system_type cgroup_fs_type = {
 	.kill_sb = cgroup_kill_sb,
 };
 
+static struct file_system_type cgroup2_fs_type = {
+	.name = "cgroup2",
+	.mount = cgroup_mount,
+	.kill_sb = cgroup_kill_sb,
+};
+
 /**
  * task_cgroup_path - cgroup path of a task in the first cgroup hierarchy
  * @task: target task
@@ -5383,6 +5381,7 @@ int __init cgroup_init(void)
 
 	WARN_ON(sysfs_create_mount_point(fs_kobj, "cgroup"));
 	WARN_ON(register_filesystem(&cgroup_fs_type));
+	WARN_ON(register_filesystem(&cgroup2_fs_type));
 	WARN_ON(!proc_create("cgroups", 0, NULL, &proc_cgroupstats_operations));
 
 	return 0;