Received-SPF: pass (google.com: domain of linux-kernel+bounces-22103-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
Date: Wed, 10 Jan 2024 12:40:06 +0100
From: Jan Kara <jack@suse.cz>
To: Hillf Danton <hdanton@sina.com>
Cc: syzbot <syzbot+a43d4f48b8397d0e41a9@syzkaller.appspotmail.com>,
	Theodore Tso <tytso@mit.edu>, Jan Kara <jack@suse.cz>,
	Matthew Wilcox <willy@infradead.org>,
	Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>,
	Linus Torvalds <torvalds@linux-foundation.org>,
	linux-kernel@vger.kernel.org, syzkaller-bugs@googlegroups.com
Subject: Re: [syzbot] [ext4?] INFO: task hung in ext4_quota_write
Message-ID: <20240110114006.thu3mpuvyb7pdhzz@quack3>
References: <20240106024807.3011-1-hdanton@sina.com>
 <20240110111259.3285-1-hdanton@sina.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20240110111259.3285-1-hdanton@sina.com>

On Wed 10-01-24 19:12:59, Hillf Danton wrote:
> On Tue, 09 Jan 2024 10:17:07 -0800
> > Hello,
> > 
> > syzbot has tested the proposed patch but the reproducer is still triggering an issue:
> > kernel BUG in ext4_quota_write
> > 
> > EXT4-fs error (device loop0) in ext4_process_orphan:347: Corrupt filesystem
> > EXT4-fs (loop0): 1 truncate cleaned up
> > EXT4-fs (loop0): mounted filesystem 00000000-0000-0000-0000-000000000000 r/w without journal. Quota mode: writeback.
> > ext4 filesystem being mounted at /root/syzkaller-testdir1916097639/syzkaller.TbSsym/0/file1 supports timestamps until 2038-01-19 (0x7fffffff)
> > ------------[ cut here ]------------
> > kernel BUG at fs/ext4/super.c:7251!
> 
> Given the BUG_ON in the debug patch tested, could deadlock be the reason
> behind the trigger instead of IO in flight? Or is it due to corrupted
> filesystem at the first place?

Thanks for the investigation! Based on your test results as well as on
results by Edward Adam Davis <eadavis@qq.com> I'd say syzbot has created a
cycle in the quota tree or something like that. Sadly the fs image provided
by syzbot is corrupted to the extent that e2fsprogs refuse to touch it so
I'll have to check manually why the kernel is mounting this image or what's
going on with the reproducer...

								Honza
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR