linux-kernel - Re: [BUG] general protection fault when connecting an old mp3/usb device

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: <aKYqvyfSO-jSTZAr@equinox>
Date: Wed, 20 Aug 2025 21:06:23 +0100
From: Phillip Potter <phil@...lpotter.co.uk>
To: David Wang <00107082@....com>
Cc: Bart Van Assche <bvanassche@....org>, phil@...lpotter.co.uk,
	linux-scsi@...r.kernel.org, linux-kernel@...r.kernel.org,
	James.Bottomley@...senpartnership.com, martin.petersen@...cle.com
Subject: Re: [BUG] general protection fault when connecting an old mp3/usb
 device

On Wed, Aug 20, 2025 at 06:02:05PM +0800, David Wang wrote:
> 
> >Phillip, is this behavior perhaps introduced by commit 5ec9d26b78c4
> >("cdrom: Call cdrom_mrw_exit from cdrom_release function")? Please do
> 
> I manage to reproduce this, but It turns out this is not about my old mp3 device, but about my phone:
> Just connect-umount-discnnect, and repeat, after several rounds, an error log would show up.
> (I should pay attention to the USB Product  name in log......
> It is just that I check the log only when I have trouble  connecting my mp3 device, and assuming the log is about my mp3 device.)
> 

Interesting. It seems your phone is emulating a CD-ROM/optical device,
which is why the driver is being loaded and attached to it.

> 
> 
> And after I upgrade to 6.17-rc1, it could not be reproduced ( I managed to test 10+ rounds of connect/umount/disconnect cycle)
> So I think commit 5ec9d26b78c4 does fix my problem.
> 
> Thanks
> David
> 
> >Bart.

Yes, as you point out, this commit became part of 6.17-rc1 - it was not
present in 6.16.0. It actually removes the cdrom_mrw_exit call from the
module unregistration function, and puts it in cdrom_release instead.

This seems to have helped with the original issue - that of crashes when
removing USB optical drives on Chromebooks, and has likely helped here
for the same reason. Good to hear, and thanks for testing/confirming
that 6.17-rc1 has helped.

Regards,
Phil