linux-kernel - Re: [PATCH 3/3] KVM:SVM: Enable INVPCID feature on AMD

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, 15 Jun 2020 15:46:53 +0200
From:   Paolo Bonzini <pbonzini@...hat.com>
To:     Jim Mattson <jmattson@...gle.com>, Babu Moger <babu.moger@....com>
Cc:     Wanpeng Li <wanpengli@...cent.com>, Joerg Roedel <joro@...tes.org>,
        the arch/x86 maintainers <x86@...nel.org>,
        Sean Christopherson <sean.j.christopherson@...el.com>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        "H . Peter Anvin" <hpa@...or.com>,
        Vitaly Kuznetsov <vkuznets@...hat.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        LKML <linux-kernel@...r.kernel.org>,
        kvm list <kvm@...r.kernel.org>
Subject: Re: [PATCH 3/3] KVM:SVM: Enable INVPCID feature on AMD

On 13/06/20 02:04, Jim Mattson wrote:
>> I think I have misunderstood this part. I was not inteding to change the
>> #GP behaviour. I will remove this part. My intension of these series is to
>> handle invpcid in shadow page mode. I have verified that part. Hope I did
>> not miss anything else.
> You don't really have to intercept INVPCID when tdp is in use, right?
> There are certainly plenty of operations for which kvm does not
> properly raise #UD when they aren't enumerated in the guest CPUID.
> 

Indeed; for RDRAND and RDSEED it makes sense to do so because the
instructions may introduce undesirable nondeterminism (or block all the
packages in your core as they have been doing for a few weeks).
Therefore on Intel we trap them and raise #UD; on AMD this is not
possible because RDRAND has no intercept.

In general however we do not try to hard to raise #UD and that is
usually not even possible.

Paolo