[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20200518120205.GF277222@hirez.programming.kicks-ass.net>
Date: Mon, 18 May 2020 14:02:05 +0200
From: Peter Zijlstra <peterz@...radead.org>
To: Like Xu <like.xu@...ux.intel.com>
Cc: Paolo Bonzini <pbonzini@...hat.com>, linux-kernel@...r.kernel.org,
kvm@...r.kernel.org,
Sean Christopherson <sean.j.christopherson@...el.com>,
Vitaly Kuznetsov <vkuznets@...hat.com>,
Wanpeng Li <wanpengli@...cent.com>,
Jim Mattson <jmattson@...gle.com>,
Joerg Roedel <joro@...tes.org>,
Thomas Gleixner <tglx@...utronix.de>, ak@...ux.intel.com,
wei.w.wang@...el.com
Subject: Re: [PATCH v11 05/11] perf/x86: Keep LBR stack unchanged in host
context for guest LBR event
On Thu, May 14, 2020 at 04:30:48PM +0800, Like Xu wrote:
> @@ -544,7 +562,12 @@ void intel_pmu_lbr_enable_all(bool pmi)
> {
> struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events);
>
> - if (cpuc->lbr_users)
> + /*
> + * When the LBR hardware is scheduled for a guest LBR event,
> + * the guest will dis/enables LBR itself at the appropriate time,
> + * including configuring MSR_LBR_SELECT.
> + */
> + if (cpuc->lbr_users && !cpuc->guest_lbr_enabled)
> __intel_pmu_lbr_enable(pmi);
> }
No!, that should be done through perf_event_attr::exclude_host, as I
believe all the other KVM event do it.
Powered by blists - more mailing lists