[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Zhhx5kMSpksInMq9@google.com>
Date: Thu, 11 Apr 2024 16:27:34 -0700
From: Sean Christopherson <seanjc@...gle.com>
To: Jim Mattson <jmattson@...gle.com>
Cc: Xiong Zhang <xiong.y.zhang@...ux.intel.com>, pbonzini@...hat.com,
peterz@...radead.org, mizhang@...gle.com, kan.liang@...el.com,
zhenyuw@...ux.intel.com, dapeng1.mi@...ux.intel.com, kvm@...r.kernel.org,
linux-perf-users@...r.kernel.org, linux-kernel@...r.kernel.org,
zhiyuan.lv@...el.com, eranian@...gle.com, irogers@...gle.com,
samantha.alt@...el.com, like.xu.linux@...il.com, chao.gao@...el.com,
Xiong Zhang <xiong.y.zhang@...el.com>
Subject: Re: [RFC PATCH 15/41] KVM: x86/pmu: Manage MSR interception for IA32_PERF_GLOBAL_CTRL
On Thu, Apr 11, 2024, Jim Mattson wrote:
> On Thu, Apr 11, 2024 at 2:21 PM Sean Christopherson <seanjc@...gle.com> wrote:
> >
> > On Fri, Jan 26, 2024, Xiong Zhang wrote:
> > > + if (is_passthrough_pmu_enabled(&vmx->vcpu)) {
> > > + /*
> > > + * Setup auto restore guest PERF_GLOBAL_CTRL MSR at vm entry.
> > > + */
> > > + if (vmentry_ctrl & VM_ENTRY_LOAD_IA32_PERF_GLOBAL_CTRL)
> > > + vmcs_write64(GUEST_IA32_PERF_GLOBAL_CTRL, 0);
> > > + else {
> > > + i = vmx_find_loadstore_msr_slot(&vmx->msr_autoload.guest,
> > > + MSR_CORE_PERF_GLOBAL_CTRL);
> > > + if (i < 0) {
> > > + i = vmx->msr_autoload.guest.nr++;
> > > + vmcs_write32(VM_ENTRY_MSR_LOAD_COUNT,
> > > + vmx->msr_autoload.guest.nr);
> > > + }
> > > + vmx->msr_autoload.guest.val[i].index = MSR_CORE_PERF_GLOBAL_CTRL;
> > > + vmx->msr_autoload.guest.val[i].value = 0;
> >
> > Eww, no. Just make cpu_has_load_perf_global_ctrl() and VM_EXIT_SAVE_IA32_PERF_GLOBAL_CTRL
> > hard requirements for enabling passthrough mode. And then have clear_atomic_switch_msr()
> > yell if KVM tries to disable loading MSR_CORE_PERF_GLOBAL_CTRL.
>
> Weren't you just complaining about the PMU version 4 constraint in
> another patch? And here, you are saying, "Don't support anything older
> than Sapphire Rapids."
Heh, I didn't realize VM_EXIT_SAVE_IA32_PERF_GLOBAL_CTRL was SPR+ when I wrote
this, I thought it existed alongside the "load" controls.
> Sapphire Rapids has PMU version 4, so if we require
> VM_EXIT_SAVE_IA32_PERF_GLOBAL_CTRL, PMU version 4 is irrelevant.
Powered by blists - more mailing lists