From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([140.186.70.92]:40314)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <mulyadi.santosa@gmail.com>) id 1QWcnG-0007rN-5k
	for qemu-devel@nongnu.org; Tue, 14 Jun 2011 19:13:11 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <mulyadi.santosa@gmail.com>) id 1QWcnE-00044p-VN
	for qemu-devel@nongnu.org; Tue, 14 Jun 2011 19:13:10 -0400
Received: from mail-gx0-f173.google.com ([209.85.161.173]:41250)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <mulyadi.santosa@gmail.com>) id 1QWcnE-00044k-Sm
	for qemu-devel@nongnu.org; Tue, 14 Jun 2011 19:13:08 -0400
Received: by gxk26 with SMTP id 26so4869357gxk.4
	for <qemu-devel@nongnu.org>; Tue, 14 Jun 2011 16:13:07 -0700 (PDT)
MIME-Version: 1.0
In-Reply-To: <4DF7E18F.6050708@cray.com>
References: <4DF7E18F.6050708@cray.com>
From: Mulyadi Santosa <mulyadi.santosa@gmail.com>
Date: Wed, 15 Jun 2011 06:12:26 +0700
Message-ID: <BANLkTikq=OYKfO16p4OmR25y_DyhqDDG7w@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Subject: Re: [Qemu-devel] High speed polling
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Clay Andreasen <cda@cray.com>
Cc: qemu-devel@nongnu.org

Hi Clay.....

On Wed, Jun 15, 2011 at 05:32, Clay Andreasen <cda@cray.com> wrote:
> I have a network device simulation that I am connecting to multiple
> instances of Qemu (nodes) via a shared memory queue. =A0It works pretty w=
ell
> as
> long as all of the nodes are initiating communication but when one node i=
s
> passive, it must poll to get packets. =A0So far the fastest I have been a=
ble
> to
> get it to poll is about every 2M emulated clocks.

that means, once every ... 1 mili second? (I doubt it, but
theoritically possible)

> This is with CONFIG_HIGH_RES_TIMERS and CONFIG_NO_HZ on the host.
> I also set MIN_TIMER_REARM_NS in qemu-timer.c to 100000.
> Is there some way to increase the polling rate by about an order of
> magnitude?

with the way current kernel did fair scheduling, your problem is
timing handling latency actually i.e timer is fast, but the handler
needs some time to catch up.

what I can suggest is experimenting with rt- kernel and put yourself
into such high priority process with realtime scheduling, keep
reducing MIN_TIMER_REARM_NS..

--=20
regards,

Mulyadi Santosa
Freelance Linux trainer and consultant

blog: the-hydra.blogspot.com
training: mulyaditraining.blogspot.com