* Re: 098bcea71b8d security relevant
[not found] ` <2026111606-uneaten-passport-9d89@gregkh>
@ 2026-11-16 10:49 ` Petr Tesarik
2026-11-16 10:55 ` Greg KH
0 siblings, 1 reply; 2+ messages in thread
From: Petr Tesarik @ 2026-11-16 10:49 UTC (permalink / raw)
To: Greg KH; +Cc: Borislav Petkov, Andrea Mattiazzo, cve, stable
On Mon, 16 Nov 2026 10:54:47 +0100
Greg KH <gregkh@linuxfoundation.org> wrote:
> On Mon, Nov 16, 2026 at 09:41:33AM +0100, Petr Tesarik wrote:
> > On Fri, 13 Nov 2026 22:24:19 -0800
> > Borislav Petkov <bp@alien8.de> wrote:
> >
> > > On Fri, Nov 13, 2026 at 02:30:48PM -0500, Greg KH wrote:
> > > > System configurations do not seem to meet the cve.org definition of
> > > > vulnerability, right?
> > >
> > > Yah, a system without any microcode loaded by the BIOS is not the common
> > > situation but rather an exception.
> >
> > One of the systems used by SUSE Performance Team boots the kernel
> > without applying any microcode patch first. We assumed that the
> > microcode from early initrd was applied in that case, but it was
> > applied only to BSP, leaving all other CPUs unpatched.
> >
> > Thank you for explaining to me that failing to apply a security fix
> > (i.e. failing to load the available microcode update) is not security
> > relevant, because the admin should know better and update their BIOS
> > instead.
>
> I'm not saying it is not security relevant, only that it does not match
> the definition of what cve.org considers a vulnerability and so is
> allowed to have a CVE id assigned to. Many such security things like
> this fall into this category, as you know.
>
> > I still do not like the fact that the Linux kernel then lies about the
> > vulnerability status under /sys/devices/system/cpu/vulnerabilities/,
> > because those files are based solely on data from BSP;
> > cpu_show_common() contains:
> >
> > if (!boot_cpu_has_bug(bug))
> > return sysfs_emit(buf, "Not affected\n");
> >
> > This gives the admin a false sense of security, but I agree it can be
> > perceived as a configuration issue: The admin should not rely on that
> > data, because they can simply check /proc/cpuinfo to see for themselves
> > that all APs actually do not have any microcode applied to them, and
> > then go read AMD errata to understand which CPU vulnerabilities are
> > mitigated with a microcode update.
>
> So is this bug not fixed? If so, it probably should be.
It is fixed with 098bcea71b8d (by correctly updating the AP microcode).
I guess the commit should be included in stable branches even if it is
not a security fix.
Petr T
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: 098bcea71b8d security relevant
2026-11-16 10:49 ` 098bcea71b8d security relevant Petr Tesarik
@ 2026-11-16 10:55 ` Greg KH
0 siblings, 0 replies; 2+ messages in thread
From: Greg KH @ 2026-11-16 10:55 UTC (permalink / raw)
To: Petr Tesarik; +Cc: Borislav Petkov, Andrea Mattiazzo, cve, stable
On Mon, Nov 16, 2026 at 11:49:02AM +0100, Petr Tesarik wrote:
> On Mon, 16 Nov 2026 10:54:47 +0100
> Greg KH <gregkh@linuxfoundation.org> wrote:
>
> > On Mon, Nov 16, 2026 at 09:41:33AM +0100, Petr Tesarik wrote:
> > > On Fri, 13 Nov 2026 22:24:19 -0800
> > > Borislav Petkov <bp@alien8.de> wrote:
> > >
> > > > On Fri, Nov 13, 2026 at 02:30:48PM -0500, Greg KH wrote:
> > > > > System configurations do not seem to meet the cve.org definition of
> > > > > vulnerability, right?
> > > >
> > > > Yah, a system without any microcode loaded by the BIOS is not the common
> > > > situation but rather an exception.
> > >
> > > One of the systems used by SUSE Performance Team boots the kernel
> > > without applying any microcode patch first. We assumed that the
> > > microcode from early initrd was applied in that case, but it was
> > > applied only to BSP, leaving all other CPUs unpatched.
> > >
> > > Thank you for explaining to me that failing to apply a security fix
> > > (i.e. failing to load the available microcode update) is not security
> > > relevant, because the admin should know better and update their BIOS
> > > instead.
> >
> > I'm not saying it is not security relevant, only that it does not match
> > the definition of what cve.org considers a vulnerability and so is
> > allowed to have a CVE id assigned to. Many such security things like
> > this fall into this category, as you know.
> >
> > > I still do not like the fact that the Linux kernel then lies about the
> > > vulnerability status under /sys/devices/system/cpu/vulnerabilities/,
> > > because those files are based solely on data from BSP;
> > > cpu_show_common() contains:
> > >
> > > if (!boot_cpu_has_bug(bug))
> > > return sysfs_emit(buf, "Not affected\n");
> > >
> > > This gives the admin a false sense of security, but I agree it can be
> > > perceived as a configuration issue: The admin should not rely on that
> > > data, because they can simply check /proc/cpuinfo to see for themselves
> > > that all APs actually do not have any microcode applied to them, and
> > > then go read AMD errata to understand which CPU vulnerabilities are
> > > mitigated with a microcode update.
> >
> > So is this bug not fixed? If so, it probably should be.
>
> It is fixed with 098bcea71b8d (by correctly updating the AP microcode).
> I guess the commit should be included in stable branches even if it is
> not a security fix.
It is in the 6.18.y and 7.2 and newer kernels already. If it needs to
go to older kernel releases, let stable@vger.kernel.org know with a
working backport and they will be glad to queue it up.
thanks,
greg k-h
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-11-16 10:55 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
[not found] <c78fbc1b-b190-4111-a6a8-f456b5c85a51@suse.de>
[not found] ` <2026111326-acre-backhand-a1d5@gregkh>
[not found] ` <20261113202510.2107c905@mordecai>
[not found] ` <2026111303-colonial-wrongly-7d9c@gregkh>
[not found] ` <20261114062419.GBavf-kxerE_YDukB7@fat_crate.local>
[not found] ` <20261116094133.4aa496f0@mordecai>
[not found] ` <2026111606-uneaten-passport-9d89@gregkh>
2026-11-16 10:49 ` 098bcea71b8d security relevant Petr Tesarik
2026-11-16 10:55 ` Greg KH
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox