linux-kernel - Re: KMSAN: uninit-value in get_term

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

Message-ID: <s5himo9as9j.wl-tiwai@suse.de>
Date:   Mon, 28 Oct 2019 14:55:52 +0100
From:   Takashi Iwai <tiwai@...e.de>
To:     Greg KH <gregkh@...uxfoundation.org>
Cc:     syzbot <syzbot+8f2612936028bfd28f28@...kaller.appspotmail.com>,
        allison@...utok.net, alsa-devel@...a-project.org,
        benquike@...il.com, dan.carpenter@...cle.com, glider@...gle.com,
        linux-kernel@...r.kernel.org, perex@...ex.cz,
        syzkaller-bugs@...glegroups.com, tglx@...utronix.de,
        tiwai@...e.com, wang6495@....edu, yuehaibing@...wei.com
Subject: Re: KMSAN: uninit-value in get_term_name

On Mon, 28 Oct 2019 14:30:50 +0100,
Greg KH wrote:
> 
> On Mon, Oct 28, 2019 at 02:13:20PM +0100, Takashi Iwai wrote:
> > On Mon, 28 Oct 2019 11:32:07 +0100,
> > syzbot wrote:
> > > 
> > > Uninit was stored to memory at:
> > >  kmsan_save_stack_with_flags mm/kmsan/kmsan.c:151 [inline]
> > >  kmsan_internal_chain_origin+0xbd/0x180 mm/kmsan/kmsan.c:319
> > >  __msan_chain_origin+0x6b/0xd0 mm/kmsan/kmsan_instr.c:179
> > >  parse_term_proc_unit+0x73d/0x7e0 sound/usb/mixer.c:896
> > >  __check_input_term+0x13ef/0x2360 sound/usb/mixer.c:989
> > 
> > So this comes from the invalid descriptor for a processing unit, and
> > it's very likely the same issue as already spotted -- the validator up
> > to 5.3-rc4 had a bug that passed the invalid descriptor falsely.
> > This should have been covered by 5.3-rc5, commit ba8bf0967a15 ("ALSA:
> > usb-audio: Fix copy&paste error in the validator").
> 
> SHould we be backporting the validator to any older kernels as well?

Yes, that would be nice.  I didn't mark them for stable just because
they are a bit largish and wanted to let them tested for 5.4 for a
while.

The following commits are relevant (from top/old to bottom/new).

57f8770620e9b51c61089751f0b5ad3dbe376ff2
    ALSA: usb-audio: More validations of descriptor units
68e9fde245591d18200f8a9054cac22339437adb
    ALSA: usb-audio: Simplify parse_audio_unit()
52c3e317a857091fd746e15179a637f32be4d337
    ALSA: usb-audio: Unify the release of usb_mixer_elem_info objects
b8e4f1fdfa422398c2d6c47bfb7d1feb3046d70a
    ALSA: usb-audio: Remove superfluous bLength checks
e0ccdef92653f8867e2d1667facfd3c23699f540
    ALSA: usb-audio: Clean up check_input_term()
60849562a5db4a1eee2160167e4dce4590d3eafe
    ALSA: usb-audio: Fix possible NULL dereference at create_yamaha_midi_quirk()
b39e077fcb283dd96dd251a3abeba585402c61fe
    ALSA: usb-audio: remove some dead code
ba8bf0967a154796be15c4983603aad0b05c3138
    ALSA: usb-audio: Fix copy&paste error in the validator


thanks,

Takashi