From owner-freebsd-stable@FreeBSD.ORG Tue Mar 28 12:33:41 2006 Return-Path: X-Original-To: freebsd-stable@freebsd.org Delivered-To: freebsd-stable@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id A260F16A424 for ; Tue, 28 Mar 2006 12:33:41 +0000 (UTC) (envelope-from andrew@areilly.bpc-users.org) Received: from omta02sl.mx.bigpond.com (omta02sl.mx.bigpond.com [144.140.93.154]) by mx1.FreeBSD.org (Postfix) with ESMTP id 7AE5243D6A for ; Tue, 28 Mar 2006 12:33:33 +0000 (GMT) (envelope-from andrew@areilly.bpc-users.org) Received: from areilly.bpc-users.org ([141.168.4.160]) by omta02sl.mx.bigpond.com with ESMTP id <20060328123332.JPOV24931.omta02sl.mx.bigpond.com@areilly.bpc-users.org> for ; Tue, 28 Mar 2006 12:33:32 +0000 Received: (qmail 90347 invoked by uid 501); 28 Mar 2006 12:34:40 -0000 Date: Tue, 28 Mar 2006 23:34:40 +1100 From: Andrew Reilly To: JoaoBR Message-ID: <20060328123440.GA90303@gurney.reilly.home> References: <20060327093011.GA21070@math.jussieu.fr> <4427E3B1.3020704@kernel32.de> <20060328104007.GD87799@gurney.reilly.home> <200603280747.55047.joao@matik.com.br> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <200603280747.55047.joao@matik.com.br> User-Agent: Mutt/1.4.2.1i Cc: freebsd-stable@freebsd.org, shih@math.jussieu.fr, Marian Hettwer Subject: Re: watchdog network card 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, 28 Mar 2006 12:33:41 -0000 On Tue, Mar 28, 2006 at 07:47:54AM -0300, JoaoBR wrote: > nve does not run polling mode but dc does Hmm. Neither it does. I could have sworn that I saw it listed in the polling(4) man page. Oh, well. It wasn't working anyway, and I haven't tried to use it yet: the dc is working with polling on, though. > I guess you have an IRQ conflict, nve and dc on the same hw interrupt, and > that setting dc in polling mode worked around this problem then nve isn't sharing an interrupt with other things. Dc might have been, but I used the BIOS to pin the DC down to an unused IRQ (can't do that with the nve since it's on the motherboard). It's possible that a BIOS upgrade might help, but I haven't had time to try that yet, either. FWIW, the isr allocation, according to dmesg.boot is: ohci0: 21 ehci0: 22 atapci1: 21 atapci2: 22 dc0: 19 fwohci0: 18 nve0: 23 sio0: 4 sio1: 3 ppc0: 7 atkbdc0: 1 atkbd0: 1 psm0: 12 It's a bit alarming that the disk controllers are sharing interrupts with the USB controller, but they seem to be working OK. I'm not using the USB much. Hmm vmstat -i thinks differently. Why don't atapci[12] show up here? interrupt total rate irq1: atkbd0 881 0 irq3: sio1 1 0 irq4: sio0 1 0 irq12: psm0 98112 0 irq15: ata1 48 0 irq16: oss 129834685 147 irq18: fwohci0 519767 0 irq19: dc0 5 0 irq21: ohci0+ 1177487 1 irq22: ehci0+ 51 0 irq23: nve0 22 0 cpu0: timer 1762687264 1999 Total 1894318324 2149 > you could check vmstat -i with and without polling enabled to see it Yeah, but turning polling off kills the network connection, and I need this machine to be working. Maybe I'll try the comparison the next time I take it down for it's regular upgrade to _STABLE. Cheers, -- Andrew