Date: Wed, 22 Jan 2003 22:58:02 -0600 (CST) From: Mike Silbersack <silby@silby.com> To: Stephen Cravey <cravey@gotbrains.org> Cc: freebsd-bugs@FreeBSD.org Subject: Re: kern/47251: 5.0 RELEASE, panic in mii_physubr.c (xl0 Driver) Message-ID: <20030122225549.G9454-100000@patrocles.silby.com> In-Reply-To: <20030122150625.L20398@www.gotbrains.org> References: <20030121171746.F37185@www.gotbrains.org> <20030121220726.C4494-100000@patrocles.silby.com> <20030122150625.L20398@www.gotbrains.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 22 Jan 2003, Stephen Cravey wrote: > Actually, I meant as something like -Release-p1 5.1? :) > In any case, the DELAY(1) patch seemed to ALMOST completely solve the > problem for me. It manifested once after about 100 megs had been pushed > through the interface. > > The patch you forgot to MFC makes my problem infinitely worse. After > applying it and rebuilding ther kernel (after cleaning the old object > files) the system is now almost totally inaccessible. Rather than being > getting errors now, the stack seems to totally halt on that session. I can > open new sessions, but the old sessions are totally locked. After about 20 > minutes, I've had one become useable again. Seems strange to me, but you > know more about the driver than I do. > > The problem is now manifesting almost as soon as I log in. I can get the > MOTD to show up usually, but most of the time I can't get the output from > 'ls' to show up before it dies. Putty sometimes responds with an "invalid > CRC on packet" error. (2-10 minutes later it seems) > > I'm moving back to the DELAY(1) patch until someone comes up with > something else to try. > > -Stephen Well, damn. Only one person had encountered the problem you're experiencing, and the extra MII_SETs fixed it for him. Looks like I'm going to have to add the DELAY calls back; I'll do that (or come up with something better!) this weekend. For future reference, could you include the output of pciconf -l and your system's dmesg? Perhaps the systems which require DELAYs have something in common. Thanks, Mike "Silby" Silbersack To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-bugs" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030122225549.G9454-100000>