[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <2025100643-tarot-gender-4430@gregkh>
Date: Mon, 6 Oct 2025 11:53:59 +0200
From: "gregkh@...uxfoundation.org" <gregkh@...uxfoundation.org>
To: Siddh Raman Pant <siddh.raman.pant@...cle.com>
Cc: "cve@...nel.org" <cve@...nel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: CVE-2025-39751: ALSA: hda/ca0132: Fix buffer overflow in
add_tuning_control
On Mon, Oct 06, 2025 at 09:19:42AM +0000, Siddh Raman Pant wrote:
> On Mon, Oct 06 2025 at 13:44:23 +0530, gregkh@...uxfoundation.org
> wrote:
> > On Mon, Oct 06, 2025 at 07:07:00AM +0000, Siddh Raman Pant wrote:
> > > On Thu, 11 Sep 2025 18:52:52 +0200, Greg Kroah-Hartman wrote:
> > > > ALSA: hda/ca0132: Fix buffer overflow in add_tuning_control
> > > >
> > > > The 'sprintf' call in 'add_tuning_control' may exceed the 44-byte
> > > > buffer if either string argument is too long. This triggers a compiler
> > > > warning.
> > > > Replaced 'sprintf' with 'snprintf' to limit string lengths to prevent
> > > > overflow.
> > > >
> > > > The Linux kernel CVE team has assigned CVE-2025-39751 to this issue.
> > >
> > > While the change is good for defensive reasons, there isn't actually
> > > any buffer overflow as it is to "fix".
> > >
> > > The largest string possible is "Wedge Angle Playback Volume", whose
> > > length is less than 44.
> >
> > Thanks for the info. What was the compiler warning about then if it
> > could detect just how big the string would always be as these are static
> > values?
>
> Probably a false positive.
>
> GCC docs does say:
>
> -Wformat-overflow
> -Wformat-overflow=level
>
> Warn about calls to formatted input/output functions such
> as sprintf and vsprintf that might overflow the
> destination
> buffer. When the exact number of bytes written by a format
> directive cannot be determined at compile-time it is
> estimated based on heuristics that depend on the level
> argument and on optimization. While enabling optimization
> will in most cases improve the accuracy of the warning, it
> may also result in false positives.
I can't seem to duplicate this warning on a newer version of gcc than
the original test used:
https://lore.kernel.org/oe-kbuild-all/202506100642.95jpuMY1-lkp@intel.com/
But that value of "767" is very specific, which feels odd to me.
> > Should this CVE be rejected?
>
> Yes.
Ok, will do, but this still seems odd, you should patch your kernel just
to be safe :)
thanks,
greg k-h
Powered by blists - more mailing lists