Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.37 as permitted sender) client-ip=23.128.96.37;
MIME-Version: 1.0
References: <20231004002038.907778-1-jmattson@google.com> <01009a2a-929e-ce16-6f44-1d314e6bcba5@intel.com>
 <CALMp9eR+Qudg++J_dmY_SGbM_kr=GQcRRcjuUxtm9rfaC_qeXQ@mail.gmail.com>
 <20231004075836.GBZR0bLC/Y09sSSYWw@fat_crate.local> <8c810f89-43f3-3742-60b8-1ba622321be8@redhat.com>
 <CALMp9eR=URBsz1qmTcDU5ixncUTkNgxJahLbfyZXYr-2RkBPng@mail.gmail.com> <b9a454c9-e8aa-02f0-29d5-57753d797cfb@redhat.com>
In-Reply-To: <b9a454c9-e8aa-02f0-29d5-57753d797cfb@redhat.com>
From:   Jim Mattson <jmattson@google.com>
Date:   Thu, 5 Oct 2023 10:27:42 -0700
Message-ID: <CALMp9eQ70wepTpmw9O3D+YnN1kaQ1m7etE4sa24bnya-vNai9Q@mail.gmail.com>
Subject: Re: [PATCH] x86: KVM: Add feature flag for AMD's FsGsKernelGsBaseNonSerializing
To:     Paolo Bonzini <pbonzini@redhat.com>
Cc:     Borislav Petkov <bp@alien8.de>,
        Dave Hansen <dave.hansen@intel.com>, kvm@vger.kernel.org,
        linux-kernel@vger.kernel.org,
        Pawan Gupta <pawan.kumar.gupta@linux.intel.com>,
        Jiaxi Chen <jiaxi.chen@linux.intel.com>,
        Kim Phillips <kim.phillips@amd.com>,
        Sean Christopherson <seanjc@google.com>,
        "H. Peter Anvin" <hpa@zytor.com>, x86@kernel.org,
        Dave Hansen <dave.hansen@linux.intel.com>,
        Ingo Molnar <mingo@redhat.com>,
        Thomas Gleixner <tglx@linutronix.de>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Thu, Oct 5, 2023 at 10:14=E2=80=AFAM Paolo Bonzini <pbonzini@redhat.com>=
 wrote:
>
> On 10/5/23 19:06, Jim Mattson wrote:
> > On Thu, Oct 5, 2023 at 9:39=E2=80=AFAM Paolo Bonzini<pbonzini@redhat.co=
m>  wrote:
> >
> >> I agree with Jim that it would be nice to have some bits from Intel, a=
nd
> >> some bits from AMD, that current processors always return as 1.  Futur=
e
> >> processors can change those to 0 as desired.
> > That's not quite what I meant.
> >
> > I'm suggesting a leaf devoted to single bit negative features. If a
> > bit is set in hardware, it means that something has been taken away.
> > Hypervisors don't need to know exactly what was taken away. For this
> > leaf only, hypervisors will always pass through a non-zero bit, even
> > if they have know idea what it means.
>
> Understood, but I'm suggesting that these might even have the right
> polarity: if a bit is set it means that something is there and might not
> in the future, even if we don't know exactly what.  We can pass through
> the bit, we can AND bits across the migration pool to define what to
> pass to the guest, we can force-set the leaves to zero (feature
> removed).  Either way, the point is to group future defeatures together.

Oh, yeah. Your suggestion is better. :)