BPF List
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Christian Brauner" <brauner@kernel.org>
Cc: bpf@vger.kernel.org
Subject: Re: [PATCH 1/2] fs: Add bpf_sock_read_xattr() kfunc to read socket xattrs
Date: Wed, 17 Jun 2026 11:32:08 +0000	[thread overview]
Message-ID: <20260617113209.8D4BA1F000E9@smtp.kernel.org> (raw)
In-Reply-To: <20260617-work-bpf-sock-xattr-v1-1-a1276f7c9da3@kernel.org>

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Missing READ_ONCE() on lockless access to sock->file can lead to a TOCTOU NULL pointer dereference.
--

commit 9bf8c99c6cda345fa08537f5df5f36bc0aa367b9
Author: Christian Brauner <brauner@kernel.org>

fs: Add bpf_sock_read_xattr() kfunc to read socket xattrs

This commit introduces the bpf_sock_read_xattr() kfunc to allow BPF LSM
programs to locklessly read user.* extended attributes from a socket's
sockfs inode. The kfunc is registered as KF_RCU, and retrieves the
attributes using an RCU-protected rhashtable lookup.

Link: https://github.com/systemd/systemd/pull/40559 [1]

> diff --git a/net/socket.c b/net/socket.c
> index 9e8dc769ff7af..3566f8c8ea3fb 100644
> --- a/net/socket.c
> +++ b/net/socket.c
[ ... ]
> @@ -465,6 +465,31 @@ static const struct xattr_handler sockfs_user_xattr_handler = {
>  	.set = sockfs_user_xattr_set,
>  };
>  
> +/**
> + * sock_read_xattr - read a user.* xattr from a socket's sockfs inode
[ ... ]
> +int sock_read_xattr(struct socket *sock, const char *name, void *value, size_t size)
> +{
> +	struct file *file = sock->file;
> +	struct sockfs_inode *si;
> +
> +	if (!file || sock_from_file(file) != sock)

[Severity: High]
Can a concurrent close cause a NULL pointer dereference here?

Since bpf_sock_read_xattr() is registered with KF_RCU, BPF programs can
call it under RCU protection where the socket might not have a held file
reference. If the last file descriptor is closed concurrently,
__sock_release() can set sock->file = NULL without waiting for an RCU grace
period.

Because sock->file is read without READ_ONCE(), the compiler is permitted
to reload the pointer from memory. If the initial read for the !file check
sees a valid pointer, but a subsequent read for the inlined
sock_from_file(file) call sees NULL due to the concurrent modification,
sock_from_file() will attempt to dereference file->f_op on a NULL pointer.

Does this need struct file *file = READ_ONCE(sock->file) to prevent a
TOCTOU issue?

> +		return -EOPNOTSUPP;
> +
> +	si = SOCKFS_I(SOCK_INODE(sock));
> +	return simple_xattr_get(&sockfs_xa_cache, &si->xattrs, name, value, size);
> +}

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260617-work-bpf-sock-xattr-v1-0-a1276f7c9da3@kernel.org?part=1

  reply	other threads:[~2026-06-17 11:32 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-06-17 11:18 [PATCH 0/2] Add bpf_sock_read_xattr() kfunc to read socket xattrs Christian Brauner
2026-06-17 11:18 ` [PATCH 1/2] fs: " Christian Brauner
2026-06-17 11:32   ` sashiko-bot [this message]
2026-06-17 14:03     ` Christian Brauner
2026-06-17 11:18 ` [PATCH 2/2] selftests/bpf: Add test for bpf_sock_read_xattr() kfunc Christian Brauner

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=20260617113209.8D4BA1F000E9@smtp.kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=bpf@vger.kernel.org \
    --cc=brauner@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