From: CJ <firefly0158@163.com>
To: shaggy@kernel.org, zheng.yu@northwestern.edu
Cc: jfs-discussion@lists.sourceforge.net, linux-kernel@vger.kernel.org
Subject: [BUG] jfs: KASAN slab-use-after-free in dtSplitPage
Date: Mon, 14 Sep 2026 16:20:08 +0800 (CST) [thread overview]
Message-ID: <63f1d46b.7a22.1a09f008c52.Coremail.firefly0158@163.com> (raw)
Hi,
I am reporting a KASAN slab-use-after-free in the JFS directory B-tree code,
triggered by a syzkaller reproducer that mounts a crafted JFS image. The issue
is reproducible with HEAD commit cee9395acd8043be0644b25c34bfa86623f2b935
(v7.3-rc1, Linux 7.3.0-rc1).
The reproducer mounts a crafted JFS filesystem image on a loop device and
performs operations that force a directory page split.
KASAN reports a slab-use-after-free with a 1-byte write, with the report
attributing the access to dtMoveEntry inlined into dtSplitPage at
fs/jfs/jfs_dtree.c. The write lands in an object from a slab cache that had
already been freed when the split moved entries between pages.
One possible cause is that dtSplitPage keeps using a directory page buffer after
the split path has released it, so the entry move writes into freed memory. In
the original report the same access was reported at a smaller offset inside the
same function; the offsets differ because the function is large and the build
differs, but the title, the access type and the inlined caller are identical.
This looks like a page-buffer lifetime problem in the dtree split path rather
than a bounds error. I am reporting the sanitizer finding and the path as
observed.
This appears to be a recurrence of the syzbot issue whose external id is
63c40a3ef45fce944739, title "KASAN: slab-use-after-free in dtSplitPage". It
remains reproducible on v7.3-rc1.
Reproducer:
syz reproducer:https://pastebin.com/raw/Sdk1vY9w
console output: https://pastebin.com/raw/LmDQmmyF
kernel config: https://pastebin.com/raw/QSKSFD7u
Kernel:
HEAD commit: cee9395acd8043be0644b25c34bfa86623f2b935
git tree: upstream (linux.git), tested through the v7.3-rc1 annotated tag object
e5e04726cdd043e309677071ab1b65a4b18f422b
kernel version: 7.3.0-rc1 #1 PREEMPT(full)
tested tag: v7.3-rc1 (Linux 7.3-rc1, 2026-08-30)
Let me know if you need more details or testing.
Best regards,
Changjian
reply other threads:[~2026-09-14 8:20 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
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=63f1d46b.7a22.1a09f008c52.Coremail.firefly0158@163.com \
--to=firefly0158@163.com \
--cc=jfs-discussion@lists.sourceforge.net \
--cc=linux-kernel@vger.kernel.org \
--cc=shaggy@kernel.org \
--cc=zheng.yu@northwestern.edu \
/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
all inboxes | Powered by JetHome®