From owner-freebsd-current@FreeBSD.ORG Sun Jan 20 18:15:36 2008 Return-Path: Delivered-To: freebsd-current@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id E275916A41B for ; Sun, 20 Jan 2008 18:15:36 +0000 (UTC) (envelope-from mailinglist@nesluop.dk) Received: from unixdunk.nesluop.dk (cpe.atm4-0-51450.0x535f05ca.hrnxx12.customer.tele.dk [83.95.5.202]) by mx1.freebsd.org (Postfix) with SMTP id 3CA1813C448 for ; Sun, 20 Jan 2008 18:15:35 +0000 (UTC) (envelope-from mailinglist@nesluop.dk) Received: (qmail 79512 invoked by uid 89); 20 Jan 2008 18:08:50 -0000 Received: from unknown (HELO ?192.168.1.100?) (192.168.1.100) by unixdunk.nesluop.dk with SMTP; 20 Jan 2008 18:08:50 -0000 Message-ID: <47938E33.6000601@nesluop.dk> Date: Sun, 20 Jan 2008 19:08:51 +0100 From: Chris Poulsen User-Agent: Thunderbird 2.0.0.9 (Windows/20071031) MIME-Version: 1.0 To: pyunyh@gmail.com References: <20071225234723.GA1018@cdnetworks.co.kr> <4772D649.3010001@nesluop.dk> <20071227002252.GE1018@cdnetworks.co.kr> <20080116012154.GB84758@cdnetworks.co.kr> <478E7DF3.4080908@nesluop.dk> <20080117014013.GA89210@cdnetworks.co.kr> <478F98D7.5040007@nesluop.dk> <20080118010100.GC92718@cdnetworks.co.kr> <20080118082609.GA93423@cdnetworks.co.kr> <4790EBA8.9090500@nesluop.dk> <20080119060354.GA98043@cdnetworks.co.kr> In-Reply-To: <20080119060354.GA98043@cdnetworks.co.kr> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: kevlo@freebsd.org, FreeBSD Current Subject: Re: Problem with nfe stability and throughput X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 20 Jan 2008 18:15:37 -0000 Hi, Pyun YongHyeon wrote: > Did you have to bring nfe(4) down and up manually due to network lockups? > Yes, I did it manually. > Maybe this would come from atphy(4) bug. atphy(4) does not seem to > reliabily detect an established link. > I saw several nfe0 down/up messages in my log after i did the manual down/up, but they stopped appearing as I stopped stressing the link. > > ifconfig nfe0 yields: > > > > nfe0: flags=8843 metric 0 mtu 1500 > > options=48 > > ether 00:1d:60:6d:73:ec > > inet 192.168.1.11 netmask 0xffffff00 broadcast 192.168.1.255 > > media: Ethernet autoselect (100baseTX ) > > status: active > > > > How about the following change? > From /usr/src/sys/dev/mii/athpy.c: > 288 ssr = PHY_READ(sc, ATPHY_SSR); > 289 if ((((bmcr & BMCR_AUTOEN) != 0) && ((bmsr & BMSR_ACOMP) == 0)) || > 290 (ssr & ATPHY_SSR_SPD_DPLX_RESOLVED) == 0) { > 291 /* Erg, still trying, I guess... */ > 292 mii->mii_media_active |= IFM_NONE; > 293 return; > 294 } > > To: > 288 ssr = PHY_READ(sc, ATPHY_SSR); > 289 if ((ssr & ATPHY_SSR_SPD_DPLX_RESOLVED) == 0) { > 290 /* Erg, still trying, I guess... */ > 291 mii->mii_media_active |= IFM_NONE; > 292 return; > 293 } > I applied the patch and rebooted the server remotely earlier this day (was feeling a little adventurous). The driver seems to run somewhat stable (not fast and not 100% reliable, but good enough for my current usage). When I get home, I'll try stressing the network some, to see how badly it misbehaves. I tried pinging the machine and the result of that is that 10-20 pings are OK and then usually a single timeout follows and then another 10-20 OK pings etc. SSH transfer from the machine does not look to be especially fast, it seemed to settle around 150 Kilo byte / sec (I'm able to pull around 1,3-2,0 MB/sec off this one using rl0 (there is some wlan involved so transfer speeds are not as reliable as the ftp uploads from my other bsd machine. > Thanks for your patience and testing. > I'd like to get my HW up and running :) - and I don't mind going through a little trouble, if it can help fbsd get even better. So I also appreciate you taking your time to give this a shot. I'll drop a mail when I get home tomorrow and have had a chance to beat up the network a little ;) -- Regards Chris