memcg: Fix possible use-after-free in memcg_write_event_control()

memcg_write_event_control() accesses the dentry->d_name of the specified
control fd to route the write call.  As a cgroup interface file can't be
renamed, it's safe to access d_name as long as the specified file is a
regular cgroup file.  Also, as these cgroup interface files can't be
removed before the directory, it's safe to access the parent too.

Prior to 347c4a8 ("memcg: remove cgroup_event->cft"), there was a
call to __file_cft() which verified that the specified file is a regular
cgroupfs file before further accesses.  The cftype pointer returned from
__file_cft() was no longer necessary and the commit inadvertently
dropped the file type check with it allowing any file to slip through.
With the invarients broken, the d_name and parent accesses can now race
against renames and removals of arbitrary files and cause
use-after-free's.

Fix the bug by resurrecting the file type check in __file_cft().  Now
that cgroupfs is implemented through kernfs, checking the file
operations needs to go through a layer of indirection.  Instead, let's
check the superblock and dentry type.

Signed-off-by: Tejun Heo <[email protected]>
Fixes: 347c4a8 ("memcg: remove cgroup_event->cft")
Cc: [email protected] # v3.14+
Reported-by: Jann Horn <[email protected]>
Acked-by: Johannes Weiner <[email protected]>
Acked-by: Roman Gushchin <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>

include/linux/cgroup.h

@@ -68,6 +68,7 @@ struct css_task_iter {
 	struct list_head		iters_node;	/* css_set->task_iters */
 };
 
+extern struct file_system_type cgroup_fs_type;
 extern struct cgroup_root cgrp_dfl_root;
 extern struct css_set init_css_set;
 

kernel/cgroup/cgroup-internal.h

@@ -167,7 +167,6 @@ struct cgroup_mgctx {
 extern spinlock_t css_set_lock;
 extern struct cgroup_subsys *cgroup_subsys[];
 extern struct list_head cgroup_roots;
-extern struct file_system_type cgroup_fs_type;
 
 /* iterate across the hierarchies */
 #define for_each_root(root)						\

mm/memcontrol.c

@@ -4832,6 +4832,7 @@ static ssize_t memcg_write_event_control(struct kernfs_open_file *of,
 	unsigned int efd, cfd;
 	struct fd efile;
 	struct fd cfile;
+	struct dentry *cdentry;
 	const char *name;
 	char *endp;
 	int ret;
@@ -4885,6 +4886,16 @@ static ssize_t memcg_write_event_control(struct kernfs_open_file *of,
 	if (ret < 0)
 		goto out_put_cfile;
 
+	/*
+	 * The control file must be a regular cgroup1 file. As a regular cgroup
+	 * file can't be renamed, it's safe to access its name afterwards.
+	 */
+	cdentry = cfile.file->f_path.dentry;
+	if (cdentry->d_sb->s_type != &cgroup_fs_type || !d_is_reg(cdentry)) {
+		ret = -EINVAL;
+		goto out_put_cfile;
+	}
+
 	/*
 	 * Determine the event callbacks and set them in @event.  This used
 	 * to be done via struct cftype but cgroup core no longer knows
@@ -4893,7 +4904,7 @@ static ssize_t memcg_write_event_control(struct kernfs_open_file *of,
 	 *
 	 * DO NOT ADD NEW FILES.
 	 */
-	name = cfile.file->f_path.dentry->d_name.name;
+	name = cdentry->d_name.name;
 
 	if (!strcmp(name, "memory.usage_in_bytes")) {
 		event->register_event = mem_cgroup_usage_register_event;
@@ -4917,7 +4928,7 @@ static ssize_t memcg_write_event_control(struct kernfs_open_file *of,
 	 * automatically removed on cgroup destruction but the removal is
 	 * asynchronous, so take an extra ref on @css.
 	 */
-	cfile_css = css_tryget_online_from_dir(cfile.file->f_path.dentry->d_parent,
+	cfile_css = css_tryget_online_from_dir(cdentry->d_parent,
 					       &memory_cgrp_subsys);
 	ret = -EINVAL;
 	if (IS_ERR(cfile_css))