Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20211020120726.4022086-1-pizhenwei@bytedance.com>
 <CANRm+CxAVA-L0wjm72eohXXWvh9fS7wVFzfKHuEjrsiRFuk9fg@mail.gmail.com>
 <YXB4FHfzh99707EH@google.com> <08757159-1673-5c7b-3efc-e5b54e82d6c3@bytedance.com>
 <CANRm+CzcTUWYJeaj3eWKH84YZYgeMZz3kbpn13c8i97iYGGHFQ@mail.gmail.com>
 <5b718b32-cd92-920e-c474-27b9cafeec60@bytedance.com> <ec014e8d-eb5f-03cc-3ed1-da58039ef034@bytedance.com>
 <YXgnDBOXh2v3gzU4@google.com>
In-Reply-To: <YXgnDBOXh2v3gzU4@google.com>
From:   Wanpeng Li <kernellwp@gmail.com>
Date:   Wed, 27 Oct 2021 08:46:49 +0800
Message-ID: <CANRm+CygFFnvjrJc7U0zNrSVzH9nC=0Bh=SjhG2pjvRXo_zXgQ@mail.gmail.com>
Subject: Re: [PATCH] x86/kvm: Introduce boot parameter no-kvm-pvipi
To:     Sean Christopherson <seanjc@google.com>
Cc:     zhenwei pi <pizhenwei@bytedance.com>,
        Paolo Bonzini <pbonzini@redhat.com>,
        Jonathan Corbet <corbet@lwn.net>,
        Wanpeng Li <wanpengli@tencent.com>,
        LKML <linux-kernel@vger.kernel.org>, linux-doc@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Wed, 27 Oct 2021 at 00:04, Sean Christopherson <seanjc@google.com> wrote:
>
> On Mon, Oct 25, 2021, zhenwei pi wrote:
> > Hi, Wanpeng & Sean
> >
> > Also benchmark redis(by 127.0.0.1) in a guest(2vCPU), 'no-kvm-pvipi' gets
> > better performance.
> >
> > Test env:
> > Host side: pin 2vCPU on 2core in a die.
> > Guest side: run command:
> >   taskset -c 1 ./redis-server --appendonly no
> >   taskset -c 0 ./redis-benchmark -h 127.0.0.1 -d 1024 -n 10000000 -t get
> >
> > 1> without no-kvm-pvipi:
> > redis QPS: 193203.12 requests per second
> > kvm_pv_send_ipi exit: ~18K/s
> >
> > 2> with no-kvm-pvipi:
> > redis QPS: 196028.47 requests per second
> > avic_incomplete_ipi_interception exit: ~5K/s
>
> Numbers look sane, but I don't think that adding a guest-side kernel param is
> the correct "fix".  As evidenced by Wanpeng's tests, PV IPI can outperform AVIC
> in overcommit scenarios, and there's also no guarantee that AVIC/APICv is even

Our evaluation is a dedicated scenario w/ big VM. The testing from
above is a one-sided view.

    Wanpeng