From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([140.186.70.92]:59569) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1QAMrM-000183-Hn for qemu-devel@nongnu.org; Thu, 14 Apr 2011 09:45:25 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1QAMrL-0003wi-KH for qemu-devel@nongnu.org; Thu, 14 Apr 2011 09:45:24 -0400 Received: from mx1.redhat.com ([209.132.183.28]:34485) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1QAMrL-0003wa-7Q for qemu-devel@nongnu.org; Thu, 14 Apr 2011 09:45:23 -0400 Message-ID: <4DA6FA65.4010503@redhat.com> Date: Thu, 14 Apr 2011 16:45:09 +0300 From: Avi Kivity MIME-Version: 1.0 References: <20110413201615.GA15989@volta.aurel32.net> <4DA69EDB.8080600@web.de> In-Reply-To: <4DA69EDB.8080600@web.de> Content-Type: text/plain; charset=ISO-8859-15; format=flowed Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH v2 0/3] io-thread optimizations List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Jan Kiszka Cc: Paolo Bonzini , Anthony Liguori , Marcelo Tosatti , qemu-devel@nongnu.org, Aurelien Jarno On 04/14/2011 10:14 AM, Jan Kiszka wrote: > On 2011-04-13 22:16, Aurelien Jarno wrote: > > On Mon, Apr 11, 2011 at 10:27:41PM +0200, Jan Kiszka wrote: > >> These patches were posted before. They bring down the overhead of the > >> io-thread mode for TCG here, specifically when emulating SMP. > >> > >> The major change in this version, besides rebasing, is the exclusion of > >> KVM from the main loop polling optimization. > >> > >> > >> > >> Jan Kiszka (3): > >> Do not drop global mutex for polled main loop runs > >> Poll main loop after I/O events were received > >> Do not kick vcpus in TCG mode > >> > >> cpus.c | 2 +- > >> sysemu.h | 2 +- > >> vl.c | 22 ++++++++++++++++------ > >> 3 files changed, 18 insertions(+), 8 deletions(-) > >> > > > > Thanks for working on improving the io-thread with TCG. Your patches > > make sense, but they don't seems to fix the slowdown observed when > > enabling the io-thread. Well maybe they were not supposed to. This is > > for example the results of netperf between guest and host using virtio: > > > > no io-thread 122 MB/s > > io-thread 97 MB/s > > io-thread + patches 98 MB/s > > > > Can you capture ftraces of io-thread enabled& disabled runs? They just > need to cover a hand full of frames. > Also interesting would be the context switch rates on the host. If they're large, perhaps using user-space threading instead of native threads would help. -- error compiling committee.c: too many arguments to function