linux-kernel - Re: [tip: ras/core] x86/mce: Enable additional error logging on certain Intel CPUs

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]

Date:   Mon, 9 Nov 2020 14:36:29 -0800
From:   Jim Mattson <jmattson@...gle.com>
To:     "Luck, Tony" <tony.luck@...el.com>
Cc:     Qian Cai <cai@...hat.com>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux-tip-commits@...r.kernel.org" 
        <linux-tip-commits@...r.kernel.org>, Boris Petkov <bp@...en8.de>,
        Borislav Petkov <bp@...e.de>, x86 <x86@...nel.org>,
        Paolo Bonzini <pbonzini@...hat.com>,
        "kvm@...r.kernel.org" <kvm@...r.kernel.org>
Subject: Re: [tip: ras/core] x86/mce: Enable additional error logging on
 certain Intel CPUs

On Mon, Nov 9, 2020 at 2:09 PM Luck, Tony <tony.luck@...el.com> wrote:
>
> What does KVM do with model specific MSRs?

"Model specific model-specific registers?" :-)

KVM only implements a small subset of MSRs. By default, any access to
the rest raises #GP.

> Looks like you let the guest believe it was running on one of Sandy Bridge, Ivy Bridge, Haswell (Xeon).
>
> So, the core MCE code tried to enable extended error reporting.
>
> If there is a mode to have KVM let the guest think that it read/wrote MSR 0x17F,
> but actually, doesn't do it ... that would seem to be a reasonable thing to do here.

There is an 'ignore_msrs' module parameter, to sink writes and return
zero on reads for unknown MSRs, but I don't think it's commonly used.

I thought Linux had long ago gone the route of turning rdmsr/wrmsr
into rdmsr_safe/wrmsr_safe, so that the guest would ignore the #GPs on
writes and return zero to the caller for #GPs on reads.