Linux XFS filesystem development
 help / color / mirror / Atom feed
From: Christoph Hellwig <hch@infradead.org>
To: Qing Ming <a0yami@mailbox.org>
Cc: Carlos Maiolino <cem@kernel.org>,
	linux-xfs@vger.kernel.org, linux-kernel@vger.kernel.org,
	"Darrick J. Wong" <djwong@kernel.org>
Subject: Re: [PATCH] xfs: remove file privileges after XFS_IOC_SWAPEXT
Date: Sun, 24 May 2026 22:53:28 -0700	[thread overview]
Message-ID: <ahPj2PLdjNC9QWIF@infradead.org> (raw)
In-Reply-To: <20260524063820.45459-1-a0yami@mailbox.org>

On Sun, May 24, 2026 at 02:38:20PM +0800, Qing Ming wrote:
> XFS_IOC_SWAPEXT exchanges the data forks of two regular files. This
> changes file contents and therefore needs the same privilege stripping
> that ordinary write paths apply.
> 
> The legacy ioctl currently completes the exchange without removing
> SUID/SGID bits or file capabilities. As a result, a privileged inode can
> retain those attributes after its data fork has been replaced.
> 
> Pass the file objects into xfs_swap_extents() and call file_remove_privs()
> for both files after the exchange commits, before dropping the outer
> inode and mapping locks. This matches the XFS_IOC_EXCHANGE_RANGE finish
> path.

Not sure this makes much sense, as xfs_swap_extents is used for
defragmentation, including system-wide one, and this would drop the
suid bit from existing suid bit files and break the system.

I don't think there is a security issue as the target file needs to
be writable to the user performing the operation, and the owning uid/gid
has to match as well.

I think the issue is more that XFS_IOC_COMMIT_RANGE drops the suid
and could thus cause problems when used by fsr.


      reply	other threads:[~2026-05-25  5:53 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-05-24  6:38 [PATCH] xfs: remove file privileges after XFS_IOC_SWAPEXT Qing Ming
2026-05-25  5:53 ` Christoph Hellwig [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=ahPj2PLdjNC9QWIF@infradead.org \
    --to=hch@infradead.org \
    --cc=a0yami@mailbox.org \
    --cc=cem@kernel.org \
    --cc=djwong@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-xfs@vger.kernel.org \
    /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