From: sashiko-bot@kernel.org
To: "Shuai Liu" <847807867@qq.com>, linux-raid@vger.kernel.org
Cc: yukuai@fygo.io
Subject: Re: [PATCH] md: reject out-of-bounds desc_nr in super_90_validate()
Date: Mon, 16 Nov 2026 08:54:17 +0000 [thread overview]
Message-ID: <sashiko-outbox-216570@kernel.org> (raw)
In-Reply-To: <tencent_3583131EEE6BD2102C15F21A4DEEB86A6208@qq.com>
Thank you for your contribution! Sashiko AI review found 1 potential
issue to consider:
- [Medium] md: ignored super_90_validate() error in md_add_new_disk()
and add_bound_rdev()
It also noticed 1 related pre-existing issue this patch does not
introduce:
- [Medium] md: missing raid_disk and delta_disks validation in
super_90_validate() [1]
--
commit 8c2da3468482d3b2e5d7496a1240f196823bd4e4
Author: Shuai Liu <847807867@qq.com>
md: reject out-of-bounds desc_nr in super_90_validate()
This commit adds a bounds check on rdev->desc_nr in super_90_validate() to
reject values outside [0, MD_SB_DISKS) with -EINVAL before indexing
sb->disks[].
> diff --git a/drivers/md/md.c b/drivers/md/md.c
> index 03e35386a8028..83009b0d84092 100644
> --- a/drivers/md/md.c
> +++ b/drivers/md/md.c
> @@ -1495,6 +1495,9 @@ static int super_90_validate(struct mddev *mddev, struct md_rdev *freshest, stru
> mdp_super_t *sb = page_address(rdev->sb_page);
> __u64 ev1 = md_event(sb);
>
> + if (rdev->desc_nr < 0 || rdev->desc_nr >= MD_SB_DISKS)
> + return -EINVAL;
> +
> rdev->raid_disk = -1;
> clear_bit(Faulty, &rdev->flags);
> clear_bit(In_sync, &rdev->flags);
[Severity: Medium]
Can a hot-added disk with an out-of-bounds sb->this_disk.number still be
accepted onto a running persistent v0.90 array?
Both md_add_new_disk() and add_bound_rdev() ignore the return value of
validate_super(), and unlike super_1_validate(), super_90_validate() does
not reset rdev->desc_nr = -1 on failure. When sb->this_disk.number >=
MD_SB_DISKS and MD_DISK_JOURNAL is set, bind_rdev_to_array() skips its
max_disks check and add_bound_rdev() proceeds to hot_add_disk() with the
out-of-bounds rdev->desc_nr intact. When sb->this_disk.number < 0,
bind_rdev_to_array() reassigns rdev->desc_nr to an in-range slot and the
second validate_super() call in add_bound_rdev() reads a different
descriptor slot than sb->this_disk.number.
[Severity: Medium]
This problem wasn't introduced by this patch, but can unvalidated
sb->delta_disks and desc->raid_disk values in super_90_validate() cause
out-of-bounds indexing into personality disk arrays [1]?
super_90_validate() copies sb->delta_disks into mddev->delta_disks and
desc->raid_disk into rdev->raid_disk without bounds checks:
drivers/md/md.c:super_90_validate() {
...
if (mddev->minor_version >= 91) {
mddev->reshape_position = sb->reshape_position;
mddev->delta_disks = sb->delta_disks;
...
} else if (desc->state & (1<<MD_DISK_SYNC)) {
set_bit(In_sync, &rdev->flags);
rdev->raid_disk = desc->raid_disk;
rdev->saved_raid_disk = desc->raid_disk;
...
}
When hot-adding via add_bound_rdev(), mddev->pers->hot_add_disk() is
called directly after super_90_validate() without analyze_sbs(). Even
during assembly, a desc->raid_disk >= 0x80000000 becomes a negative
rdev->raid_disk (< -1) with In_sync set, or a negative sb->delta_disks
inflates the upper bound in analyze_sbs() past mddev->raid_disks.
Personality handlers such as raid1_add_disk() then index
conf->mirrors[rdev->raid_disk] out of bounds.
[1] https://sashiko.dev/#/bug/linux-9ad89505-1b9e-435b-941d-6a126f8f4adb
--
Sashiko AI review · https://sashiko.dev/#/patchset/tencent_3583131EEE6BD2102C15F21A4DEEB86A6208@qq.com?part=1
next prev parent reply other threads:[~2026-11-16 8:54 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-15 22:06 [syzbot] [raid?] KASAN: slab-out-of-bounds Read in super_90_validate syzbot
2026-11-16 8:32 ` [PATCH] md: reject out-of-bounds desc_nr in super_90_validate() Shuai Liu
2026-11-16 8:54 ` sashiko-bot [this message]
2026-11-16 8:59 ` [PATCH v2] " Shuai Liu
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-216570@kernel.org \
--to=sashiko-bot@kernel.org \
--cc=847807867@qq.com \
--cc=linux-raid@vger.kernel.org \
--cc=sashiko-reviews@lists.linux.dev \
--cc=yukuai@fygo.io \
/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