[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <aa182eec-3d0b-1452-19cc-20654190a2ae@redhat.com>
Date: Mon, 4 May 2020 18:57:00 +0200
From: Paolo Bonzini <pbonzini@...hat.com>
To: Like Xu <like.xu@...ux.intel.com>,
Jim Mattson <jmattson@...gle.com>,
Sean Christopherson <sean.j.christopherson@...el.com>
Cc: Vitaly Kuznetsov <vkuznets@...hat.com>,
Wanpeng Li <wanpengli@...cent.com>,
Joerg Roedel <joro@...tes.org>, kvm@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [RESEND PATCH] KVM: x86/pmu: Support full width counting
On 27/04/20 09:19, Like Xu wrote:
> + if (vmx_supported_perf_capabilities())
> + kvm_cpu_cap_check_and_set(X86_FEATURE_PDCM);
I think we can always set it, worst case it will be zero.
However, blocking intel_pmu_set_msr altogether is incorrect. Instead,
you need to:
- list the MSR in msr_based_features_all so that it appears in
KVM_GET_MSR_FEATURE_INDEX_LIST
- return the supported bits in vmx_get_msr_feature
- allow host-initiated writes (as long as they only set supported bits)
of the MSR in intel_pmu_set_msr.
Thanks,
Paolo
Powered by blists - more mailing lists