sysfs, kernfs: implement kernfs_create/destroy_root()

There currently is single kernfs hierarchy in the whole system which
is used for sysfs.  kernfs needs to support multiple hierarchies to
allow other users.  This patch introduces struct kernfs_root which
serves as the root of each kernfs hierarchy and implements
kernfs_create/destroy_root().

* Each kernfs_root is associated with a root sd (sysfs_dentry).  The
  root is freed when the root sd is released and kernfs_destory_root()
  simply invokes kernfs_remove() on the root sd.  sysfs_remove_one()
  is updated to handle release of the root sd.  Note that ps_iattr
  update in sysfs_remove_one() is trivially updated for readability.

* Root sd's are now dynamically allocated using sysfs_new_dirent().
  Update sysfs_alloc_ino() so that it gives out ino from 1 so that the
  root sd still gets ino 1.

* While kernfs currently only points to the root sd, it'll soon grow
  fields which are specific to each hierarchy.  As determining a given
  sd's root will be necessary, sd->s_dir.root is added.  This backlink
  fits better as a separate field in sd; however, sd->s_dir is inside
  union with space to spare, so use it to save space and provide
  kernfs_root() accessor to determine the root sd.

* As hierarchies may be destroyed now, each mount needs to hold onto
  the hierarchy it's attached to.  Update sysfs_fill_super() and
  sysfs_kill_sb() so that they get and put the kernfs_root
  respectively.

* sysfs_root is replaced with kernfs_root which is dynamically created
  by invoking kernfs_create_root() from sysfs_init().

This patch doesn't introduce any visible behavior changes.

v2: kernfs_create_root() forgot to set @sd->priv.  Fixed.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

3 files changed, 100 insertions, 20 deletions

diff --git a/fs/kernfs/dir.c b/fs/kernfs/dir.c
index a4ca4de3cb2..246740a741e 100644
--- a/fs/kernfs/dir.c
+++ b/fs/kernfs/dir.c
@@ -211,7 +211,7 @@ static int sysfs_alloc_ino(unsigned int *pino)
 
  retry:
 	spin_lock(&sysfs_ino_lock);
-	rc = ida_get_new_above(&sysfs_ino_ida, 2, &ino);
+	rc = ida_get_new_above(&sysfs_ino_ida, 1, &ino);
 	spin_unlock(&sysfs_ino_lock);
 
 	if (rc == -EAGAIN) {
@@ -253,9 +253,11 @@ EXPORT_SYMBOL_GPL(kernfs_get);
 void kernfs_put(struct sysfs_dirent *sd)
 {
 	struct sysfs_dirent *parent_sd;
+	struct kernfs_root *root;
 
 	if (!sd || !atomic_dec_and_test(&sd->s_count))
 		return;
+	root = kernfs_root(sd);
  repeat:
 	/* Moving/renaming is always done while holding reference.
 	 * sd->s_parent won't change beneath us.
@@ -278,8 +280,13 @@ void kernfs_put(struct sysfs_dirent *sd)
 	kmem_cache_free(sysfs_dir_cachep, sd);
 
 	sd = parent_sd;
-	if (sd && atomic_dec_and_test(&sd->s_count))
-		goto repeat;
+	if (sd) {
+		if (atomic_dec_and_test(&sd->s_count))
+			goto repeat;
+	} else {
+		/* just released the root sd, free @root too */
+		kfree(root);
+	}
 }
 EXPORT_SYMBOL_GPL(kernfs_put);
 
@@ -493,13 +500,15 @@ static void sysfs_remove_one(struct sysfs_addrm_cxt *acxt,
 	if (sd->s_flags & SYSFS_FLAG_REMOVED)
 		return;
 
-	sysfs_unlink_sibling(sd);
+	if (sd->s_parent) {
+		sysfs_unlink_sibling(sd);
 
-	/* Update timestamps on the parent */
-	ps_iattr = sd->s_parent->s_iattr;
-	if (ps_iattr) {
-		struct iattr *ps_iattrs = &ps_iattr->ia_iattr;
-		ps_iattrs->ia_ctime = ps_iattrs->ia_mtime = CURRENT_TIME;
+		/* Update timestamps on the parent */
+		ps_iattr = sd->s_parent->s_iattr;
+		if (ps_iattr) {
+			ps_iattr->ia_iattr.ia_ctime = CURRENT_TIME;
+			ps_iattr->ia_iattr.ia_mtime = CURRENT_TIME;
+		}
 	}
 
 	sd->s_flags |= SYSFS_FLAG_REMOVED;
@@ -604,6 +613,49 @@ struct sysfs_dirent *kernfs_find_and_get_ns(struct sysfs_dirent *parent,
 EXPORT_SYMBOL_GPL(kernfs_find_and_get_ns);
 
 /**
+ * kernfs_create_root - create a new kernfs hierarchy
+ * @priv: opaque data associated with the new directory
+ *
+ * Returns the root of the new hierarchy on success, ERR_PTR() value on
+ * failure.
+ */
+struct kernfs_root *kernfs_create_root(void *priv)
+{
+	struct kernfs_root *root;
+	struct sysfs_dirent *sd;
+
+	root = kzalloc(sizeof(*root), GFP_KERNEL);
+	if (!root)
+		return ERR_PTR(-ENOMEM);
+
+	sd = sysfs_new_dirent("", S_IFDIR | S_IRUGO | S_IXUGO, SYSFS_DIR);
+	if (!sd) {
+		kfree(root);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	sd->s_flags &= ~SYSFS_FLAG_REMOVED;
+	sd->priv = priv;
+	sd->s_dir.root = root;
+
+	root->sd = sd;
+
+	return root;
+}
+
+/**
+ * kernfs_destroy_root - destroy a kernfs hierarchy
+ * @root: root of the hierarchy to destroy
+ *
+ * Destroy the hierarchy anchored at @root by removing all existing
+ * directories and destroying @root.
+ */
+void kernfs_destroy_root(struct kernfs_root *root)
+{
+	kernfs_remove(root->sd);	/* will also free @root */
+}
+
+/**
  * kernfs_create_dir_ns - create a directory
  * @parent: parent in which to create a new directory
  * @name: name of the new directory
@@ -626,6 +678,7 @@ struct sysfs_dirent *kernfs_create_dir_ns(struct sysfs_dirent *parent,
 	if (!sd)
 		return ERR_PTR(-ENOMEM);
 
+	sd->s_dir.root = parent->s_dir.root;
 	sd->s_ns = ns;
 	sd->priv = priv;
 
diff --git a/fs/kernfs/kernfs-internal.h b/fs/kernfs/kernfs-internal.h
index 62ae35f997f..7dfe0627835 100644
--- a/fs/kernfs/kernfs-internal.h
+++ b/fs/kernfs/kernfs-internal.h
@@ -25,6 +25,12 @@ struct sysfs_elem_dir {
 	unsigned long		subdirs;
 	/* children rbtree starts here and goes through sd->s_rb */
 	struct rb_root		children;
+
+	/*
+	 * The kernfs hierarchy this directory belongs to.  This fits
+	 * better directly in sysfs_dirent but is here to save space.
+	 */
+	struct kernfs_root	*root;
 };
 
 struct sysfs_elem_symlink {
@@ -104,6 +110,20 @@ static inline unsigned int sysfs_type(struct sysfs_dirent *sd)
 	return sd->s_flags & SYSFS_TYPE_MASK;
 }
 
+/**
+ * kernfs_root - find out the kernfs_root a sysfs_dirent belongs to
+ * @sd: sysfs_dirent of interest
+ *
+ * Return the kernfs_root @sd belongs to.
+ */
+static inline struct kernfs_root *kernfs_root(struct sysfs_dirent *sd)
+{
+	/* if parent exists, it's always a dir; otherwise, @sd is a dir */
+	if (sd->s_parent)
+		sd = sd->s_parent;
+	return sd->s_dir.root;
+}
+
 /*
  * Context structure to be used while adding/removing nodes.
  */
diff --git a/fs/sysfs/mount.c b/fs/sysfs/mount.c
index 7cbd1fce282..0b5661b462f 100644
--- a/fs/sysfs/mount.c
+++ b/fs/sysfs/mount.c
@@ -32,15 +32,8 @@ static const struct super_operations sysfs_ops = {
 	.evict_inode	= sysfs_evict_inode,
 };
 
-static struct sysfs_dirent sysfs_root = {
-	.s_name		= "",
-	.s_count	= ATOMIC_INIT(1),
-	.s_flags	= SYSFS_DIR,
-	.s_mode		= S_IFDIR | S_IRUGO | S_IXUGO,
-	.s_ino		= 1,
-};
-
-struct sysfs_dirent *sysfs_root_sd = &sysfs_root;
+static struct kernfs_root *sysfs_root;
+struct sysfs_dirent *sysfs_root_sd;
 
 static int sysfs_fill_super(struct super_block *sb)
 {
@@ -68,6 +61,7 @@ static int sysfs_fill_super(struct super_block *sb)
 		pr_debug("%s: could not get root dentry!\n", __func__);
 		return -ENOMEM;
 	}
+	kernfs_get(sysfs_root_sd);
 	root->d_fsdata = sysfs_root_sd;
 	sb->s_root = root;
 	sb->s_d_op = &sysfs_dentry_ops;
@@ -138,11 +132,15 @@ static struct dentry *sysfs_mount(struct file_system_type *fs_type,
 static void sysfs_kill_sb(struct super_block *sb)
 {
 	struct sysfs_super_info *info = sysfs_info(sb);
-	/* Remove the superblock from fs_supers/s_instances
+	struct sysfs_dirent *root_sd = sb->s_root->d_fsdata;
+
+	/*
+	 * Remove the superblock from fs_supers/s_instances
 	 * so we can't find it, before freeing sysfs_super_info.
 	 */
 	kill_anon_super(sb);
 	free_sysfs_super_info(info);
+	kernfs_put(root_sd);
 }
 
 static struct file_system_type sysfs_fs_type = {
@@ -166,12 +164,21 @@ int __init sysfs_init(void)
 	if (err)
 		goto out_err;
 
+	sysfs_root = kernfs_create_root(NULL);
+	if (IS_ERR(sysfs_root)) {
+		err = PTR_ERR(sysfs_root);
+		goto out_err;
+	}
+	sysfs_root_sd = sysfs_root->sd;
+
 	err = register_filesystem(&sysfs_fs_type);
 	if (err)
-		goto out_err;
+		goto out_destroy_root;
 
 	return 0;
 
+out_destroy_root:
+	kernfs_destroy_root(sysfs_root);
 out_err:
 	kmem_cache_destroy(sysfs_dir_cachep);
 	sysfs_dir_cachep = NULL;