From mboxrd@z Thu Jan  1 00:00:00 1970
From: Cyrill Gorcunov <gorcunov@gmail.com>
Subject: Re: [PATCH 1/3] kvm tools: Make virtio-net kvm__irq_line thread safe
Date: Fri, 29 Apr 2011 14:47:57 +0400
Message-ID: <BANLkTimAYMs9aMwMCbaO8yCuC=7HUnn_vA@mail.gmail.com>
References: <1304058985-13833-1-git-send-email-asias.hejun@gmail.com>
	<BANLkTikn8JM33gRCsODgKWPyq7UZPjdL6g@mail.gmail.com>
	<4DBA6519.5070509@gmail.com>
	<1304063606.10069.7.camel@lappy>
	<BANLkTi=bAKpEgR95v=MnwEAWfAEVdzOh_g@mail.gmail.com>
	<BANLkTikHjqC_KjibLPozVE+9OGXiefGYkw@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: QUOTED-PRINTABLE
Cc: Sasha Levin <levinsasha928@gmail.com>,
	Asias He <asias.hejun@gmail.com>, Ingo Molnar <mingo@elte.hu>,
	Prasad Joshi <prasadjoshi124@gmail.com>,
	"kvm@vger.kernel.org" <kvm@vger.kernel.org>
To: Pekka Enberg <penberg@kernel.org>
Return-path: <kvm-owner@vger.kernel.org>
Received: from mail-pv0-f174.google.com ([74.125.83.174]:38374 "EHLO
	mail-pv0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1753121Ab1D2Kr5 convert rfc822-to-8bit (ORCPT
	<rfc822;kvm@vger.kernel.org>); Fri, 29 Apr 2011 06:47:57 -0400
Received: by pvg12 with SMTP id 12so2237968pvg.19
        for <kvm@vger.kernel.org>; Fri, 29 Apr 2011 03:47:57 -0700 (PDT)
In-Reply-To: <BANLkTikHjqC_KjibLPozVE+9OGXiefGYkw@mail.gmail.com>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

i suspect so, threadpool makes all things to be more async as they
were before and proper lockings does matter, i could test any patches
at evening, and i must admit i never get well uderstanding on virtio
device operations yet (in terms of virtio ring :)

On Friday, April 29, 2011, Pekka Enberg <penberg@kernel.org> wrote:
> On Fri, Apr 29, 2011 at 1:02 PM, Cyrill Gorcunov <gorcunov@gmail.com>=
 wrote:
>> So, if I understand all the things correct -- making virtio devices =
to
>> belong separated irqs
>> issued some race conditions on read\write operations between host an=
d
>> guest and adding
>> thread pool revealed it, right? (because previously we were doing al=
l
>> the work inside i/o
>> path on guest site).
>
> So does reverting commit a37089da817ce7aad9789aeb9fc09b68e088ad9a
> ("kvm tools: Use threadpool for virtio-net") fix things? I think the
> problem here is that now RX path relies on VIRTIO_PCI_QUEUE_NOTIFY to
> happen in order for it to trigger KVM_IRQ_LINE which is wrong. Using
> shared IRQs obviously masks the problem which is why reverting
> Cyrill's commit makes the problem go away.
>
>  =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0Pekka
>