linux-kernel - Re: [PATCH (REPOST)] hfs: don't use BUG() when we can continue

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <b992789a-84f5-4f57-88f6-76efedd7d00e@I-love.SAKURA.ne.jp>
Date: Thu, 5 Dec 2024 23:14:36 +0900
From: Tetsuo Handa <penguin-kernel@...ove.SAKURA.ne.jp>
To: Matthew Wilcox <willy@...radead.org>
Cc: Andrew Morton <akpm@...ux-foundation.org>,
        linux-fsdevel <linux-fsdevel@...r.kernel.org>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH (REPOST)] hfs: don't use BUG() when we can continue

On 2024/12/05 22:59, Matthew Wilcox wrote:
> On Thu, Dec 05, 2024 at 10:45:11PM +0900, Tetsuo Handa wrote:
>> syzkaller can mount crafted filesystem images.
>> Don't crash the kernel when we can continue.
> 
> The one in hfs_test_inode() makes sense, but we should never get to
> hfs_release_folio() or hfs_write_inode() with a bad inode, even with a
> corrupted fs.  Right?

Unfortunately, https://syzkaller.appspot.com/bug?extid=97e301b4b82ae803d21b
reports that we can reach hfs_write_inode() with a corrupted filesystem.
If we should never get to hfs_write_inode(), how do we want to handle
this problem? Is someone familiar with hfs enough to perform an in-kernel
fsck upon mounting?