linux-kernel - Re: [PATCH 2/2] KVM: x86: Forbid KVM_SET_CPUID{,2} after KVM

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: <877dbbq5om.fsf@redhat.com>
Date:   Fri, 07 Jan 2022 10:02:49 +0100
From:   Vitaly Kuznetsov <vkuznets@...hat.com>
To:     Paolo Bonzini <pbonzini@...hat.com>,
        Igor Mammedov <imammedo@...hat.com>
Cc:     kvm@...r.kernel.org, Sean Christopherson <seanjc@...gle.com>,
        Wanpeng Li <wanpengli@...cent.com>,
        Jim Mattson <jmattson@...gle.com>, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 2/2] KVM: x86: Forbid KVM_SET_CPUID{,2} after KVM_RUN

Vitaly Kuznetsov <vkuznets@...hat.com> writes:

> Paolo Bonzini <pbonzini@...hat.com> writes:
>
>> On 1/3/22 13:56, Vitaly Kuznetsov wrote:
>>>   'allowlist' of things which can change (and put
>>> *APICids there) and only fail KVM_SET_CPUID{,2} when we see something
>>> else changing.
>>
>> We could also go the other way and only deny changes that result in 
>> changed CPU caps.  That should be easier to implement since we have 
>> already a mapping from CPU capability words to CPUID leaves and registers.
>>
>
> Good idea, I'll look into it (if noone beats me to it).

(just getting to it)

On the other hand, e.g. MAXPHYADDR (CPUID 0x80000008.EAX) is not a CPU
cap but it's one of the main reasons why we want to forbid
KVM_SET_CPUID{,2} after KVM_RUN in the first place. I'm also not sure
about allowing PV feature bits changes (KVM, Hyper-V, Xen) and I don't
think there's a need for that.

I'm again leaning towards an allowlist and currently I see only two
candidates:

CPUID.01H.EBX bits 31:24 (initial LAPIC id)
CPUID.0BH.EDX (x2APIC id)

Anything else I'm missing?

-- 
Vitaly