linux-kernel - Re: [PATCH] x86/mce: Check for hypervisor before enabling additional error logging

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:   Tue, 10 Nov 2020 09:50:43 +0100
From:   Paolo Bonzini <pbonzini@...hat.com>
To:     Borislav Petkov <bp@...en8.de>, "Luck, Tony" <tony.luck@...el.com>
Cc:     Jim Mattson <jmattson@...gle.com>, 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>, x86 <x86@...nel.org>,
        "kvm@...r.kernel.org" <kvm@...r.kernel.org>
Subject: Re: [PATCH] x86/mce: Check for hypervisor before enabling additional
 error logging

On 10/11/20 07:31, Borislav Petkov wrote:
>>   
>> +	if (boot_cpu_has(X86_FEATURE_HYPERVISOR))
>> +		return;
>> +
> Frankly, I'm tired of wagging the dog because the tail can't. If
> qemu/kvm can't emulate a CPU model fully then it should ignore those
> unknown MSR accesses by default, i.e., that "ignore_msrs" functionality
> should be on by default I'd say...
> 
> We certainly can't be sprinkling this check everytime the kernel tries
> to do something as basic as read an MSR.

You don't have to, also because it's wrong.  Fortunately it's much 
simpler than that:

1) ignore_msrs _cannot_ be on by default.  You cannot know in advance 
that for all non-architectural MSRs it's okay for them to read as zero 
and eat writes.  For some non-architectural MSR which never reads as 
zero on real hardware, who knows that there isn't some code using the 
contents of the MSR as a divisor, and causing a division by zero 
exception with ignore_msrs=1?

2) it's not just KVM.  _Any_ hypervisor is bound to have this issue for 
some non-architectural MSRs.  KVM just gets the flak because Linux CI 
environments (for obvious reasons) use it more than they use Hyper-V or 
ESXi or VirtualBox.

3) because of (1) and (2), the solution is very simple.  If the MSR is 
architectural, its absence is a KVM bug and we'll fix it in all stable 
versions.  If the MSR is not architectural (and 17Fh isn't; not only 
it's not mentioned in the SDM, even Google is failing me), never ever 
assume that the CPUID family/model/stepping implies a given MSR is 
there, and just use rdmsr_safe/wrmsr_safe.

So, for this patch,

Nacked-by: Paolo Bonzini <pbonzini@...hat.com>

Paolo