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;
Subject: Re: [PATCH 1/2] KVM: X86: Single target IPI fastpath
To:     Wanpeng Li <kernellwp@gmail.com>, linux-kernel@vger.kernel.org,
        kvm@vger.kernel.org
Cc:     =?UTF-8?B?UmFkaW0gS3LEjW3DocWZ?= <rkrcmar@redhat.com>,
        Sean Christopherson <sean.j.christopherson@intel.com>,
        Vitaly Kuznetsov <vkuznets@redhat.com>,
        Wanpeng Li <wanpengli@tencent.com>,
        Jim Mattson <jmattson@google.com>,
        Joerg Roedel <joro@8bytes.org>
References: <1573283135-5502-1-git-send-email-wanpengli@tencent.com>
From:   Paolo Bonzini <pbonzini@redhat.com>
Openpgp: preference=signencrypt
Message-ID: <4418c734-68e1-edaf-c939-f24d041acf2e@redhat.com>
Date:   Mon, 11 Nov 2019 22:59:14 +0100
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101
 Thunderbird/60.8.0
MIME-Version: 1.0
In-Reply-To: <1573283135-5502-1-git-send-email-wanpengli@tencent.com>
Content-Language: en-US
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 09/11/19 08:05, Wanpeng Li wrote:
> From: Wanpeng Li <wanpengli@tencent.com>
>=20
> This patch tries to optimize x2apic physical destination mode, fixed deli=
very
> mode single target IPI by delivering IPI to receiver immediately after se=
nder
> writes ICR vmexit to avoid various checks when possible.
>=20
> Testing on Xeon Skylake server:
>=20
> The virtual IPI latency from sender send to receiver receive reduces more=
 than
> 330+ cpu cycles.
>=20
> Running hackbench(reschedule ipi) in the guest, the avg handle time of MS=
R_WRITE
> caused vmexit reduces more than 1000+ cpu cycles:
>=20
> Before patch:
>=20
>   VM-EXIT    Samples  Samples%     Time%    Min Time    Max Time   Avg ti=
me
> MSR_WRITE    5417390    90.01%    16.31%      0.69us    159.60us    1.08u=
s
>=20
> After patch:
>=20
>   VM-EXIT    Samples  Samples%     Time%    Min Time    Max Time   Avg ti=
me
> MSR_WRITE    6726109    90.73%    62.18%      0.48us    191.27us    0.58u=
s

Do you have retpolines enabled?  The bulk of the speedup might come just
from the indirect jump.

Paolo