From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263284AbVFYCdE (ORCPT ); Fri, 24 Jun 2005 22:33:04 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263288AbVFYCdD (ORCPT ); Fri, 24 Jun 2005 22:33:03 -0400 Received: from rwcrmhc13.comcast.net ([204.127.198.39]:25317 "EHLO rwcrmhc13.comcast.net") by vger.kernel.org with ESMTP id S263284AbVFYCc5 (ORCPT ); Fri, 24 Jun 2005 22:32:57 -0400 Message-ID: <42BCC257.5060900@comcast.net> Date: Fri, 24 Jun 2005 22:32:55 -0400 From: Ed Sweetman User-Agent: Debian Thunderbird 1.0.2 (X11/20050602) X-Accept-Language: en-us, en MIME-Version: 1.0 To: Ed Sweetman CC: Pierre Ossman , Bjorn Helgaas , LKML , jgarzik@pobox.com Subject: Re: 2.6.12 breaks 8139cp References: <42B9D21F.7040908@drzeus.cx> <200506221534.03716.bjorn.helgaas@hp.com> <42BA69AC.5090202@drzeus.cx> <200506231143.34769.bjorn.helgaas@hp.com> <42BB3428.6030708@drzeus.cx> <42BCBB60.7000508@comcast.net> <42BCBF01.206@comcast.net> In-Reply-To: <42BCBF01.206@comcast.net> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Sorry, i'm retarded, i pasted the wrong changelog line. [8139TOO]: Use rtnl_lock_interruptible() The 8139too thread needs to use rtnl_lock_interruptible so it can avoid doing the actual work once it's been kill_proc()ed on module removal time. Based on debugging and an earlier patch that adds a driver-private semaphore from Herbert Xu. Signed-off-by: David S. Miller This seems to be the only patch that contains a change to the 8139too code between working and non-working code. nothing in mm's patchset seems to have anything to cause these problems. Ed Sweetman wrote: > this is in the changefile to 2.6.11. This seems to be the real > culprit. I guarantee you if this patch is reverted, there will be no > problems. > > > [PATCH] r8169: hint for Tx flow control > > return 1 in start_xmit() when the required descriptors are not > available > and wait for more room. > > Signed-off-by: Francois Romieu > > > > > Ed Sweetman wrote: > >> Pierre Ossman wrote: >> >>> Bjorn Helgaas wrote: >>> >>> >>> >>>> Your 2.6.11 dmesg mentions the VIA IRQ fixup, but the 2.6.12 one >>>> doesn't. I bet something's broken there. >>>> >>>> Can you try the attached debugging patch? And please collect the >>>> output of lspci, too. >>>> >>>> >>>> >>> >>> >>> >>> I tried the attached patch and it had no effect. I also tried porting >>> the 2.6.11 way of handling the VIA quirk but it didn't have any effect. >>> I'll try a more complete port tomorrow (it was a bit of a hack this >>> time). >>> >>> >>> >> 2.6.11-mm4 doesn't work. So i'm guessing 2.6.11 wont work either >> which may be why backporting it's via fixes didn't do anything. I'm >> gonna try vanilla and if that by some crazy chance works, then it'll >> be fairly easy to see what change did it since mm has a nice Changelog. >> - >> To unsubscribe from this list: send the line "unsubscribe >> linux-kernel" in >> the body of a message to majordomo@vger.kernel.org >> More majordomo info at http://vger.kernel.org/majordomo-info.html >> Please read the FAQ at http://www.tux.org/lkml/ >>