linux-kernel - Re: [PATCH] kvm/x86/vmx: switch MSR_MISC_FEATURES

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] [day] [month] [year] [list]

Date:   Thu, 14 Mar 2019 15:44:39 +0100
From:   Paolo Bonzini <pbonzini@...hat.com>
To:     Xiaoyao Li <xiaoyao.li@...ux.intel.com>
Cc:     Kyle Huey <khuey@...ehuey.com>, Chao Gao <chao.gao@...el.com>,
        Radim Krčmář <rkrcmar@...hat.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        "H. Peter Anvin" <hpa@...or.com>, x86@...nel.org,
        kvm@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] kvm/x86/vmx: switch MSR_MISC_FEATURES_ENABLES between
 host and guest

On 14/03/19 13:37, Xiaoyao Li wrote:
>> Adding a RDMSR for this to each vmentry is too heavy.  Since we emulate
>> MSR_MISC_FEATURES_ENABLES, you can just clear the MSR on vcpu_load and
>> restore it on vcpu_put.
> One question here. Just clear the MSR on vcpu_load instead of writing the
> emulated value to MSR? 
> 
> I think writing the emulated value to MSR is better. As I mentioned in case 3,
> if hardware has cpuid faulting feature. Using hardware capability is more
> efficient than emulation that the emulation solution needs VM exit to inject
> #GP.   

You can do that too, yes.  You can add it to vmx_msr_index and it will
be handled like that.

However, that wouldn't work on AMD (which doesn't use the
kvm_define_shared_msr infrastructure) and also on hosts that don't have
MSR_MISC_FEATURES_ENABLES, so my suggestion is to add this optimization
as a separate patch.

Paolo