Linux USB
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "William Chuang" <founder@logarcheon.com>
Cc: linux-usb@vger.kernel.org
Subject: Re: [PATCH] configfs: give the root directory inode its own lockdep class
Date: Tue, 17 Nov 2026 07:13:47 +0000	[thread overview]
Message-ID: <sashiko-outbox-218297@kernel.org> (raw)
In-Reply-To: <20261117065314.4309-1-founder@logarcheon.com>

Thank you for your contribution! Sashiko AI review found 1 potential
issue to consider:
- [High] configfs: root inode lockdep class masks real rename() vs
  mkdir() deadlock
--

commit 058eebb48180eef94995fcd78f689f610830d617
Author: William Chuang <founder@logarcheon.com>

configfs: give the root directory inode its own lockdep class

This commit assigns a dedicated lockdep class to the configfs root
directory inode's i_rwsem to avoid a lockdep circular dependency report
between USB gadget and target subsystems.

> diff --git a/fs/configfs/mount.c b/fs/configfs/mount.c
> index d8cac1cbf3bd5..d20d2fd1e1809 100644
> --- a/fs/configfs/mount.c
> +++ b/fs/configfs/mount.c
> @@ -58,6 +58,15 @@ static struct configfs_dirent configfs_root = {
>  	.s_iattr	= NULL,
>  };
>  
> +/*
> + * configfs_depend_item_unlocked() takes the root inode's lock while a
> + * subsystem's own locks may be held, but nothing takes a subsystem lock
> + * while holding the root inode's lock.  Directories created by mkdir(2)
> + * share the default i_rwsem class and are locked before their subsystem's
> + * su_mutex, so give the root inode a class of its own.
> + */
> +static struct lock_class_key configfs_root_inode_class;
> +
>  static int configfs_fill_super(struct super_block *sb, struct fs_context *fc)
>  {
>  	struct inode *inode;
> @@ -74,6 +83,7 @@ static int configfs_fill_super(struct super_block *sb, struct fs_context *fc)
>  	if (inode) {
>  		inode->i_op = &configfs_root_inode_operations;
>  		inode->i_fop = &configfs_dir_operations;
> +		lockdep_set_class(&inode->i_rwsem, &configfs_root_inode_class);

[Severity: High]
Does assigning a separate lockdep class to the configfs root inode's
i_rwsem hide a real ABBA deadlock between renameat2() and mkdir()?

In filename_renameat2(), __start_renaming() -> lock_rename() ->
lock_two_directories() runs before vfs_rename() checks
old_dir->i_op->rename.  When one parent directory is the configfs root
and the other is a descendant directory (such as target/usb_gadget/naa.*),
lock_two_directories() sees that the root is an ancestor and locks
root_inode->i_rwsem before the descendant directory's i_rwsem:

fs/namei.c:lock_two_directories() {
	...
	r = d_ancestor(p2, p1);
	if (r) {
		inode_lock_nested(p2->d_inode, I_MUTEX_PARENT);
		inode_lock_nested(p1->d_inode, I_MUTEX_PARENT2);
		return r;
	}
}

Meanwhile, mkdir() on target/usb_gadget/naa.*/tpgt_1 locks those same two
inodes in the opposite order:

CPU0                                  CPU1
----                                  ----
lock(&naa_inode->i_rwsem)
                                      lock(&root_inode->i_rwsem)
                                      lock(&naa_inode->i_rwsem)
lock(&root_inode->i_rwsem)

 *** DEADLOCK ***

CPU0: do_mkdirat() -> configfs_mkdir() -> target_fabric_make_tpg() ->
      usbg_make_tpg() -> configfs_depend_item_unlocked()
CPU1: do_renameat2() -> filename_renameat2() -> __start_renaming() ->
      lock_rename() -> lock_two_directories()

Splitting root_inode->i_rwsem into configfs_root_inode_class prevents
lockdep from detecting this lock ordering inversion until a rename()
actually races with mkdir() or symlink().

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261117065314.4309-1-founder@logarcheon.com?part=1

      reply	other threads:[~2026-11-17  7:13 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-17  6:53 [PATCH] configfs: give the root directory inode its own lockdep class William Chuang
2026-11-17  7:13 ` sashiko-bot [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=sashiko-outbox-218297@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=founder@logarcheon.com \
    --cc=linux-usb@vger.kernel.org \
    --cc=sashiko-reviews@lists.linux.dev \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox