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;
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
From:   Andy Lutomirski <luto@amacapital.net>
Mime-Version: 1.0 (1.0)
Subject: Re: [PATCH v2] x86/kvm: Disable KVM_ASYNC_PF_SEND_ALWAYS
Date:   Tue, 7 Apr 2020 15:29:23 -0700
Message-Id: <0255CF03-D45D-45E0-BC61-79159B94ED44@amacapital.net>
References: <2776fced-54c2-40eb-7921-1c68236c7f70@redhat.com>
Cc:     Thomas Gleixner <tglx@linutronix.de>,
        Vivek Goyal <vgoyal@redhat.com>,
        Peter Zijlstra <peterz@infradead.org>,
        Andy Lutomirski <luto@kernel.org>,
        LKML <linux-kernel@vger.kernel.org>, X86 ML <x86@kernel.org>,
        kvm list <kvm@vger.kernel.org>, stable <stable@vger.kernel.org>
In-Reply-To: <2776fced-54c2-40eb-7921-1c68236c7f70@redhat.com>
To:     Paolo Bonzini <pbonzini@redhat.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


> On Apr 7, 2020, at 3:07 PM, Paolo Bonzini <pbonzini@redhat.com> wrote:
>=20
> =EF=BB=BFOn 07/04/20 23:41, Andy Lutomirski wrote:
>> 2. Access to bad memory results in #MC.  Sure, #MC is a turd, but
>> it=E2=80=99s an *architectural* turd. By all means, have a nice simple PV=

>> mechanism to tell the #MC code exactly what went wrong, but keep the
>> overall flow the same as in the native case.
>>=20
>> I think I like #2 much better. It has another nice effect: a good
>> implementation will serve as a way to exercise the #MC code without
>> needing to muck with EINJ or with whatever magic Tony uses. The
>> average kernel developer does not have access to a box with testable
>> memory failure reporting.
>=20
> I prefer #VE, but I can see how #MC has some appeal.  However, #VE has a
> mechanism to avoid reentrancy, unlike #MC.  How would that be better
> than the current mess with an NMI happening in the first few
> instructions of the #PF handler?
>=20
>=20

It has to be an IST vector due to the possibility of hitting a memory failur=
e right after SYSCALL.  I imagine that making #VE use IST would be unfortuna=
te.

I think #MC has a mechanism to prevent reentrancy to a limited extent. How d=
oes #VE avoid reentrancy?=