From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:47737) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1V6HG5-0000Ay-NI for qemu-devel@nongnu.org; Mon, 05 Aug 2013 05:39:26 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1V6HG0-000413-F3 for qemu-devel@nongnu.org; Mon, 05 Aug 2013 05:39:21 -0400 Received: from cantor2.suse.de ([195.135.220.15]:45911 helo=mx2.suse.de) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1V6HG0-0003yf-5n for qemu-devel@nongnu.org; Mon, 05 Aug 2013 05:39:16 -0400 Message-ID: <51FF72BD.6060202@suse.de> Date: Mon, 05 Aug 2013 11:39:09 +0200 From: =?ISO-8859-1?Q?Andreas_F=E4rber?= MIME-Version: 1.0 References: <1375461052-2476-1-git-send-email-v.maffione@gmail.com> <20130805092725.GD13743@stefanha-thinkpad.muc.redhat.com> In-Reply-To: <20130805092725.GD13743@stefanha-thinkpad.muc.redhat.com> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH v6] e1000: add interrupt mitigation support List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Stefan Hajnoczi Cc: aliguori@us.ibm.com, mst@redhat.com, rizzo@iet.unipi.it, jasowang@redhat.com, qemu-devel@nongnu.org, Vincenzo Maffione , pbonzini@redhat.com, g.lettieri@iet.unipi.it, lersek@redhat.com Am 05.08.2013 11:27, schrieb Stefan Hajnoczi: > On Fri, Aug 02, 2013 at 06:30:52PM +0200, Vincenzo Maffione wrote: >> This patch partially implements the e1000 interrupt mitigation mechani= sms. >> Using a single QEMUTimer, it emulates the ITR register (which is the n= ewer >> mitigation register, recommended by Intel) and approximately emulates >> RADV and TADV registers. TIDV and RDTR register functionalities are no= t >> emulated (RDTR is only used to validate RADV, according to the e1000 s= pecs). >> >> RADV, TADV, TIDV and RDTR registers make up the older e1000 mitigation >> mechanism and would need a timer each to be completely emulated. Howev= er, >> a single timer has been used in order to reach a good compromise betwe= en >> emulation accuracy and simplicity/efficiency. >> >> The implemented mechanism can be enabled/disabled specifying the comma= nd >> line e1000-specific boolean parameter "mitigation", e.g. >> >> qemu-system-x86_64 -device e1000,mitigation=3Don,... ... >> >> For more information, see the Software developer's manual at >> http://download.intel.com/design/network/manuals/8254x_GBe_SDM.pdf. >> >> Interrupt mitigation boosts performance when the guest suffers from >> an high interrupt rate (i.e. receiving short UDP packets at high packe= t >> rate). For some numerical results see the following link >> http://info.iet.unipi.it/~luigi/papers/20130520-rizzo-vm.pdf >> >> Signed-off-by: Vincenzo Maffione >> --- >> Added pc-*-1.7 machines (default machine moved to pc-i440fx-1.7). >> >> hw/i386/pc_piix.c | 18 ++++++- >> hw/i386/pc_q35.c | 16 ++++++- >> hw/net/e1000.c | 131 ++++++++++++++++++++++++++++++++++++++++++= +++++++-- >> include/hw/i386/pc.h | 8 ++++ >> 4 files changed, 167 insertions(+), 6 deletions(-) >=20 > I will merge this into my net tree soon but I'd like to wait for > Andreas' review to make sure he's comfortable with the machine type > changes too. I'm okay with the machine changes, just not comfortable reviewing the net code. Feel free to add a "Reviewed-by: Andreas F=E4rber (for pc-* machines)" if you wish. Andreas --=20 SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 N=FCrnberg, Germany GF: Jeff Hawn, Jennifer Guild, Felix Imend=F6rffer; HRB 16746 AG N=FCrnbe= rg