From owner-freebsd-stable@FreeBSD.ORG Tue Sep 25 09:14:05 2007 Return-Path: Delivered-To: freebsd-stable@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 224FD16A46B for ; Tue, 25 Sep 2007 09:14:05 +0000 (UTC) (envelope-from matrix@itlegion.ru) Received: from corpmail.itlegion.ru (corpmail.itlegion.ru [84.21.226.211]) by mx1.freebsd.org (Postfix) with SMTP id 557E513C458 for ; Tue, 25 Sep 2007 09:14:03 +0000 (UTC) (envelope-from matrix@itlegion.ru) Received: (qmail 95927 invoked from network); 25 Sep 2007 13:14:01 +0400 Received: from unknown (HELO Artem) (192.168.0.12) by 84.21.226.211 with SMTP; 25 Sep 2007 13:14:01 +0400 X-AntiVirus: Checked by Dr.Web [version: 4.33, engine: 4.33.5.10110, virus records: 254434, updated: 25.09.2007] Message-ID: <00b801c7ff54$6796f2b0$0c00a8c0@Artem> From: "Artem Kuchin" To: "Nikos Vassiliadis" , References: <01ae01c7fd3e$e6ff28f0$0c00a8c0@Artem> <001301c7fed4$7e5f38c0$0c00a8c0@Artem> <200709251048.20808.nvass@teledomenet.gr> Date: Tue, 25 Sep 2007 13:13:47 +0400 Organization: IT Legion MIME-Version: 1.0 Content-Type: text/plain; format=flowed; charset="koi8-r"; reply-type=original Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2900.3138 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3138 Cc: Subject: Re: device polling and weird timer interrupt count from vmstat X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 25 Sep 2007 09:14:05 -0000 Nikos Vassiliadis wrote: > On Monday 24 September 2007 20:58, Artem Kuchin wrote: >>> What i don't understand is why timer rate on each cpu is 1995? I >>> have set it to 1000, not 1995 or 2000. I have seen it showing 2000 >>> on another box. >>> So >>> 1) why not 1000? > > > I can only make assumptions about the doubling, and I don't want to. > FreeBSD is not a RTOS and some milliseconds variation is > understandable. > >>> 2) if it is supposed to be doubled (why?) when why not 2000? > > I can only make assumptions about the doubling, and I don't want to. > FreeBSD is not a RTOS and some milliseconds variation is > understandable. > >>> 3) Is timer int really generated on each cpu? > > Apparently, why do you doubt it? > >>> Am i really wasting cpu >>> time on ~4000 ints per second? > > You can lower it you know, if you feel that you are waisting > that much resources. Ofcourse you'll break your traffic flow > that way, since latency will increase. > > You seem very upset about it, are you sure you want to use > polling(4)? it uses much more resources than interrupts. I am not upset about all this. But i want to understand why is it doubled. Because when i turn on polling i think that timer freq supposed to be just like i set HZ. However, actually, i am now thinking about another issue. I have dual CPU with HT. If i turn on HT (and it does help in my case) it shoud 2000 int x4 = 8000 ints per second. SO, i have saved 200 int/second from NIC and got myself 8000 ints/second from timer. Did i really win anything? I wish there were some good explanation on this. >>> 4) does twe driver use polling? whay about twa? how to check it in >>> the sources? > > Polling is only used on some network interface drivers. Polling(4) > does not offer generic device-polling facilities. > > By the way, you know your post has an aggressive sense, don't you? > Please don't do this when asking questions and want replies. Hmm.. Really? I didn't mean it, i was just trying to me as short and as technical as possible. Alright, i'll give a though on how to be more.. polite, i guess. -- Regards, Artem