From mboxrd@z Thu Jan 1 00:00:00 1970 From: Beth Kon Subject: Re: [Qemu-devel] [RFC][PATCH] Add HPET emulation to qemu (v2) Date: Tue, 12 Aug 2008 12:33:52 -0400 Message-ID: <1218558832.7401.13.camel@beth-ubuntu> References: <1217675114-17670-1-git-send-email-eak@us.ibm.com> <20080802113812.GD4535@implementation> <48947346.5080605@codemonkey.ws> <20080802172103.GH4535@implementation> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Anthony Liguori , qemu-devel@nongnu.org, kvm@vger.kernel.org To: Samuel Thibault Return-path: Received: from e4.ny.us.ibm.com ([32.97.182.144]:34148 "EHLO e4.ny.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750989AbYHLQeW (ORCPT ); Tue, 12 Aug 2008 12:34:22 -0400 Received: from d01relay02.pok.ibm.com (d01relay02.pok.ibm.com [9.56.227.234]) by e4.ny.us.ibm.com (8.13.8/8.13.8) with ESMTP id m7CGYLLw000668 for ; Tue, 12 Aug 2008 12:34:21 -0400 Received: from d01av03.pok.ibm.com (d01av03.pok.ibm.com [9.56.224.217]) by d01relay02.pok.ibm.com (8.13.8/8.13.8/NCO v9.0) with ESMTP id m7CGYL2x230344 for ; Tue, 12 Aug 2008 12:34:21 -0400 Received: from d01av03.pok.ibm.com (loopback [127.0.0.1]) by d01av03.pok.ibm.com (8.12.11.20060308/8.13.3) with ESMTP id m7CGYLUR016031 for ; Tue, 12 Aug 2008 12:34:21 -0400 In-Reply-To: <20080802172103.GH4535@implementation> Sender: kvm-owner@vger.kernel.org List-ID: On Sat, 2008-08-02 at 18:21 +0100, Samuel Thibault wrote: > Anthony Liguori, le Sat 02 Aug 2008 09:46:30 -0500, a =C3=A9crit : > > Samuel Thibault wrote: > > >Beth Kon, le Sat 02 Aug 2008 06:05:14 -0500, a =C3=A9crit : > > > =20 > > >>I was trying to reproduce the wakeup every 10ms that=20 > > >>Samuel Thibault mentioned, thinking the HPET would improve it.=20 > > >>But for an idle guest in both cases (with and without HPET), the=20 > > >>number of wakeups per second was relatively low (28). > > >> =20 > > > > > >I was referring to vl.c's timeout =3D 10; which makes the select c= all > > >use a timeout of 10ms. That said, "/* If all cpus are halted then = wait > > >until the next IRQ */", so maybe that's why you get slower wakeups= per > > >second. I'm still surprised because of the call to qemu_mod_timer= in > > >pit_irq_timer_update which should setup at least a 100Hz timer wit= h > > >linux guests (when they don't have HPET available). > > > =20 > >=20 > > The patch disables that when hpet is active. >=20 > That's why I would expect, indeed, but he is reporting that _without_ > HPET he gets low wakeups per second already. >=20 > Samuel Yes, 28 is incorrect. I was misinterpreting the output of powertop, shown here: Wakeups-from-idle per second : 27.7 interval: 10.0s no ACPI power usage estimate available Top causes for wakeups: 46.1% ( 63.9) qemu-system-x86 : schedule_timeout (process_timeout) 36.5% ( 50.6) qemu-system-x86 : sys_timer_settime (posix_timer_fn) ... The "Wakeups-from-idle per second" reports 27.7, but the powertop sourc= e code shows that this value is the total wakeups-per-second divided by the number of online processors. So the proper number of wakeups-per-second caused by the select is 63.9, which makes more sense.=20 Looking at the main_loop code, there is no way to get a timeout of greater than 10 without setting icount.=20 > -- > To unsubscribe from this list: send the line "unsubscribe kvm" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html --=20 Elizabeth Kon (Beth) IBM Linux Technology Center Open Hypervisor Team email: eak@us.ibm.com