From mboxrd@z Thu Jan  1 00:00:00 1970
From: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com>
Subject: Re: [PART1 RFC v2 10/10] svm: Manage vcpu load/unload when enable
 AVIC
Date: Mon, 14 Mar 2016 18:58:05 +0700
Message-ID: <56E6A74D.8090807@amd.com>
References: <1457124368-2025-1-git-send-email-Suravee.Suthikulpanit@amd.com>
 <1457124368-2025-11-git-send-email-Suravee.Suthikulpanit@amd.com>
 <20160309214629.GE19459@potion.brq.redhat.com>
 <20160310140141.GA23501@potion.brq.redhat.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8;
	format=flowed
Content-Transfer-Encoding: QUOTED-PRINTABLE
Cc: <pbonzini@redhat.com>, <joro@8bytes.org>, <bp@alien8.de>,
	<gleb@kernel.org>, <alex.williamson@redhat.com>,
	<kvm@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
	<wei@redhat.com>, <sherry.hurwitz@amd.com>
To: =?UTF-8?B?UmFkaW0gS3LEjW3DocWZ?= <rkrcmar@redhat.com>
Return-path: <kvm-owner@vger.kernel.org>
Received: from mail-bn1on0057.outbound.protection.outlook.com ([157.56.110.57]:39904
	"EHLO na01-bn1-obe.outbound.protection.outlook.com"
	rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP
	id S1755494AbcCNL62 (ORCPT <rfc822;kvm@vger.kernel.org>);
	Mon, 14 Mar 2016 07:58:28 -0400
In-Reply-To: <20160310140141.GA23501@potion.brq.redhat.com>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

Hi,

On 03/10/2016 09:01 PM, Radim Kr=C4=8Dm=C3=A1=C5=99 wrote:
> Well, we haven't reached an agreement on is_running yet.  The situati=
on:
> if we don't unset vcpu1.is_running when vcpu1 is scheduled out and vc=
pu2
> gets scheduled on vcpu1's physical core, then vcpu2 would receive a
> doorbell intended to vcpu1.

That's why, in V2, I added the logic to check if the is_running bit is=20
set for the current vcpu (e.g. vcpu1) when unloaded, then restore the=20
bit during loading later of if it was set during previous unloaded. Thi=
s=20
way, when we load the new vcpu (e.g. vcpu2), the is_running will be set=
=20
as it was before unloading.

> We'd like to keep is_running set when there is no reason to vmexit, b=
ut
> not if a guest can negatively affect other guests.

Not sure how this can affect other guests?

> How does receiving a stray doorbell affect the performance?

As far as I know, the doorbell only affects the CPU during vmrun. Once=20
received, it will check the IRR in vAPIC backing page.  So, I think if=20
IRR bit is not set, the affect should be rather minimal.

Thanks,
Suravee