From: "Darrick J. Wong" <djwong@kernel.org>
To: Christoph Hellwig <hch@lst.de>
Cc: Andrey Albershteyn <aalbersh@kernel.org>, linux-xfs@vger.kernel.org
Subject: Re: [PATCH 07/10] libxfs: fix buftarg initialization for the RT device
Date: Fri, 25 Sep 2026 15:42:38 -0700 [thread overview]
Message-ID: <20260925224238.GV2705364@frogsfrogsfrogs> (raw)
In-Reply-To: <20260925051336.2997014-8-hch@lst.de>
On Fri, Sep 25, 2026 at 07:12:53AM +0200, Christoph Hellwig wrote:
> libxfs_buftarg_init currently sets the rtdev targp to the ddev one when
> no explicit RT device is provided, which allows tools like repair to
> proceed without a specified RT device when they should not, and
> potentially can cause data corruption.
>
> Fix this by delaying the RT buftarg initialization to libxfs_mount,
> where we can check if the superblock indicates an internal RT device.
> This works because all external callers of libxfs_buftarg_init never
> need the RT device as they just examine the sb and/or log,
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>
😬😬😬
Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>
--D
> ---
> libxfs/init.c | 11 +++++------
> 1 file changed, 5 insertions(+), 6 deletions(-)
>
> diff --git a/libxfs/init.c b/libxfs/init.c
> index 842a2e1e09a9..97c5286c00a9 100644
> --- a/libxfs/init.c
> +++ b/libxfs/init.c
> @@ -554,8 +554,7 @@ libxfs_buftarg_reinit(
> }
>
> if (mp->m_rtdev_targp && mp->m_rtdev_targp != mp->m_ddev_targp) {
> - if ((xi->rt.dev && xi->rt.dev == xi->data.dev) ||
> - mp->m_rtdev_targp->bt_bdev != xi->rt.dev ||
> + if (mp->m_rtdev_targp->bt_bdev != xi->rt.dev ||
> mp->m_rtdev_targp->bt_mount != mp) {
> fprintf(stderr,
> _("%s: bad buftarg reinit, rtdev\n"),
> @@ -587,10 +586,6 @@ libxfs_buftarg_init(
> mp->m_logdev_targp = mp->m_ddev_targp;
> else
> mp->m_logdev_targp = libxfs_buftarg_alloc(mp, xi, &xi->log);
> - if (!xi->rt.dev || xi->rt.dev == xi->data.dev)
> - mp->m_rtdev_targp = mp->m_ddev_targp;
> - else
> - mp->m_rtdev_targp = libxfs_buftarg_alloc(mp, xi, &xi->rt);
> }
>
> /* Compute maximum possible height for per-AG btree types for this fs. */
> @@ -739,6 +734,10 @@ libxfs_mount(
> if (flags & LIBXFS_MOUNT_REPORT_CORRUPTION)
> xfs_set_reporting_corruption(mp);
> libxfs_buftarg_init(mp, xi);
> + if (sb->sb_rtstart)
> + mp->m_rtdev_targp = mp->m_ddev_targp;
> + else if (xi->rt.dev)
> + mp->m_rtdev_targp = libxfs_buftarg_alloc(mp, xi, &xi->rt);
> libxfs_setup_write_error_injection(mp);
>
> if (xi->data.name)
> --
> 2.53.0
>
>
next prev parent reply other threads:[~2026-09-25 22:42 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-25 5:12 fix missing RT detection in libxfs v2 Christoph Hellwig
2026-09-25 5:12 ` [PATCH 01/10] libxfs: refactor the reinit case in libxfs_buftarg_init Christoph Hellwig
2026-09-25 22:32 ` Darrick J. Wong
2026-09-25 5:12 ` [PATCH 02/10] libxfs: split error injection option parsing out of libxfs_buftarg_init Christoph Hellwig
2026-09-25 22:33 ` Darrick J. Wong
2026-09-25 5:12 ` [PATCH 03/10] libxfs: remove buftarg member aliases Christoph Hellwig
2026-09-25 5:12 ` [PATCH 04/10] libxfs: unify buftarg handling with the kernel Christoph Hellwig
2026-09-25 22:37 ` Darrick J. Wong
2026-09-26 5:59 ` Christoph Hellwig
2026-09-27 15:18 ` Darrick J. Wong
2026-09-25 5:12 ` [PATCH 05/10] libxfs: make the size check in rtmount_init internal RT device aware Christoph Hellwig
2026-09-25 22:38 ` Darrick J. Wong
2026-09-25 5:12 ` [PATCH 06/10] libxfs: don't clear ->dev in libxfs_device_close Christoph Hellwig
2026-09-25 22:39 ` Darrick J. Wong
2026-09-25 5:12 ` [PATCH 07/10] libxfs: fix buftarg initialization for the RT device Christoph Hellwig
2026-09-25 22:42 ` Darrick J. Wong [this message]
2026-09-25 5:12 ` [PATCH 08/10] libxfs: rename bt_bdev* Christoph Hellwig
2026-09-25 22:43 ` Darrick J. Wong
2026-09-25 5:12 ` [PATCH 09/10] libfrog: improve ramdisk handling in platform_flush_device Christoph Hellwig
2026-09-25 22:47 ` Darrick J. Wong
2026-09-26 6:02 ` Christoph Hellwig
2026-09-27 15:22 ` Darrick J. Wong
2026-09-25 5:12 ` [PATCH 10/10] xfs_db: remove a double-whitespace in rdump_regfile_data Christoph Hellwig
2026-09-25 22:47 ` Darrick J. Wong
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=20260925224238.GV2705364@frogsfrogsfrogs \
--to=djwong@kernel.org \
--cc=aalbersh@kernel.org \
--cc=hch@lst.de \
--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