Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <1545816338-1171-1-git-send-email-wei.w.wang@intel.com>
 <1545816338-1171-6-git-send-email-wei.w.wang@intel.com> <CALMp9eTOV3Twep9gL-9S+Pe_k-=v17CcJTLb5=+7_pjvWf9RfQ@mail.gmail.com>
 <5C2DC132.9050103@intel.com> <CALMp9eQ0nzKBP7az72G+eRe8C4KQQUBdj+7oD_WJHSFfYkj+Vw@mail.gmail.com>
 <5C3318A1.9090009@intel.com>
In-Reply-To: <5C3318A1.9090009@intel.com>
From:   Jim Mattson <jmattson@google.com>
Date:   Mon, 7 Jan 2019 10:05:27 -0800
Message-ID: <CALMp9eSyC5arXkrFdJKCx=cG7uF6tVgkbSJbUJArf4cFje13rQ@mail.gmail.com>
Subject: Re: [PATCH v4 05/10] KVM/x86: expose MSR_IA32_PERF_CAPABILITIES to
 the guest
To:     Wei Wang <wei.w.wang@intel.com>
Cc:     LKML <linux-kernel@vger.kernel.org>,
        kvm list <kvm@vger.kernel.org>,
        Paolo Bonzini <pbonzini@redhat.com>,
        Andi Kleen <ak@linux.intel.com>,
        Peter Zijlstra <peterz@infradead.org>,
        Kan Liang <kan.liang@intel.com>,
        Ingo Molnar <mingo@redhat.com>,
        =?UTF-8?B?UmFkaW0gS3LEjW3DocWZ?= <rkrcmar@redhat.com>,
        like.xu@intel.com, Jann Horn <jannh@google.com>,
        arei.gonglei@huawei.com
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Mon, Jan 7, 2019 at 1:09 AM Wei Wang <wei.w.wang@intel.com> wrote:
>
> On 01/03/2019 11:25 PM, Jim Mattson wrote:
> > On Wed, Jan 2, 2019 at 11:55 PM Wei Wang <wei.w.wang@intel.com> wrote:
> >
> >> Right, thanks. Probably better to change it to below:
> >>
> >> msr_info->data = 0;
> >> data = native_read_msr(MSR_IA32_PERF_CAPABILITIES);
> >> if (vcpu->kvm->arch.lbr_in_guest)
> >>       msr_info->data |= (data & X86_PERF_CAP_MASK_LBR_FMT);
> >>
> > This still breaks backwards compatibility. Returning 0 and raising #GP
> > are not the same.
>
> I'm not sure about raising GP# in this case.
>
> This PERF_CAP msr contains more things than the lbr format.
> For example, a guest with lbr=false option could read it to get PEBS_FMT,
> which is PERF_CAP[11:8]. We should offer those bits in this case.
>
> When lbr=false, the lbr feature is not usable by the guest,
> so I think whatever value (0 or other value) of the LBR_FMT bits that
> we give to the guest might not be important.

The issue is compatibility. Prior to your change, reading this MSR
from a VM would raise #GP. After your change, it won't. That means
that if you have a VM migrating between hosts with kernel versions
before and after this change, the results will be inconsistent. In the
forward migration path, RDMSR will first raise #GP, and later will
return 0. In the backward migration path, RDMSR will first return 0,
and later it will raise #GP.

To avoid these inconsistencies, the new MSR behavior should be
explicitly enabled by an opt-in ioctl from userspace (not necessarily
the same ioctl that enables LBR).