linux-ext4 - Re: [syzbot] [ext4?] general protection fault in hrtimer

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]

Date: Fri, 10 Nov 2023 12:08:35 -0500
From: "Theodore Ts'o" <tytso@....edu>
To: Aleksandr Nogikh <nogikh@...gle.com>
Cc: Thomas Gleixner <tglx@...utronix.de>,
        syzbot <syzbot+b408cd9b40ec25380ee1@...kaller.appspotmail.com>,
        adilger.kernel@...ger.ca, linux-ext4@...r.kernel.org,
        linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org,
        syzkaller-bugs@...glegroups.com
Subject: Re: [syzbot] [ext4?] general protection fault in hrtimer_nanosleep

On Thu, Nov 09, 2023 at 09:00:18PM -0800, Aleksandr Nogikh wrote:
> 
> The reproducer does work on the attached disk image, but definitely
> not very often. I've just run it 10 times or so and got interleaved
> BUG/KFENCE bug reports like this (twice):
> https://pastebin.com/W0TkRsnw
> 
> These seem to be related to ext4 rather than hrtimers though.

So what would be nice is if there was a way to ask the syzkaller
tester to use a different config or to change the reproducer somehow
--- for example, is it *really* necessary to twiddle the bluetooth
subsystem, as demonstrated by the spew in the console?

I've certainly spent hours cutting down the reproducer to a simple C
program which is readable by humans, which makes it *clear* the syzbot
minimizer doesn't do a good job.  Why should a time-limited maintainer
spend hours trying to cut down the reproducer, when a robot should be
able to do that for us?  And when often it doesn't reproduce on
anything via syzbot test, but not when run using KVM, this is why we
need to have a simple way of trigger a test where things are as close
as possible to whatever syzbot is using.

Cheers,

						- Ted