From owner-freebsd-current Mon Apr 15 2:11:42 2002 Delivered-To: freebsd-current@freebsd.org Received: from van-laarhoven.org (ap-z-5ab8.adsl.wanadoo.nl [212.129.218.184]) by hub.freebsd.org (Postfix) with SMTP id 0364437B416 for ; Mon, 15 Apr 2002 02:11:38 -0700 (PDT) Received: (qmail 57885 invoked from network); 15 Apr 2002 09:11:36 -0000 Received: from heather.van-laarhoven.org (10.66.0.2) by uitsmijter.van-laarhoven.org with SMTP; 15 Apr 2002 09:11:36 -0000 Date: Mon, 15 Apr 2002 11:11:36 +0200 (CEST) From: Nick Hibma To: Makoto Matsushita Cc: John Baldwin , Will Andrews , "current@FreeBSD.org" Subject: Re: plug aue ethernet goes to panic In-Reply-To: <20020413175429.GA65074@squall.waterspout.com> Message-ID: <20020415110325.J36693-100000@heather.van-laarhoven.org> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG device_set_ivars is always called (in usbd_probe_and_attach) with as an argument a stack variable. Also, the ivar is not stored or anything in the if_aue.c driver. So this problem sounds like a problem in revisions of various files. Please check that your kernel modules kernel are in sync. Do this by rebuilding the kernel and the modules from scratch. Also, after you've installed your kernel check that all your kernel files have been updated. Do you by any chance have a stale /modules or /boot/modules directory lying around? You should have only kernel modules in /boot/kernel*/ and NOT in /modules* or /boot/modules*. If the problem persists, please mail me the output of ident /sys/dev/usb/*.[ch] find /modules /boot -type f -ls Thanks. Nick On Sat, 13 Apr 2002, Will Andrews wrote: > On Sat, Apr 13, 2002 at 12:44:36PM -0400, John Baldwin wrote: > > Can you get a backtrace in ddb? It looks like a null pointer dereference, and > > knowing where it happened would help. Finding the file and line of the > > instruction pointer using addr2line on kernel.debug would be helpful as well. > > It *is* a null pointer deref. Joe and I looked at this problem, > and it seems a function called device_get_ivars() isn't doing > its job in sys/dev/usb/usb_port.h. So the variable is getting > filled with a NULL pointer for the iface element and is later > deref'd. That's the limit of my debugging, and Joe is looking > into the problem actively. > > Regards, > -- n_hibma@van-laarhoven.org http://www.van-laarhoven.org/ n_hibma@FreeBSD.org http://www.etla.net/~n_hibma/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message