From mboxrd@z Thu Jan 1 00:00:00 1970 From: Larry Finger Subject: Re: bcm43xx regression 2.6.19rc3 -> rc5, rtnl_lock trouble? Date: Thu, 16 Nov 2006 16:40:59 -0600 Message-ID: <455CE8FB.2000803@lwfinger.net> References: <455B63EC.8070704@madrabbit.org> <455BFC47.3020006@madrabbit.org> <455CAB2F.1060709@lwfinger.net> <200611162016.42095.mb@bu3sch.de> <455CBDD7.6000507@madrabbit.org> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Cc: netdev-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Bcm43xx-dev-0fE9KPoRgkgATYTw5x5z8w@public.gmane.org, Michael Buesch Return-path: To: Ray Lee In-Reply-To: <455CBDD7.6000507-0Cg02Ec9UG4BXFe83j6qeQ@public.gmane.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: bcm43xx-dev-bounces-0fE9KPoRgkgATYTw5x5z8w@public.gmane.org Errors-To: bcm43xx-dev-bounces-0fE9KPoRgkgATYTw5x5z8w@public.gmane.org List-Id: netdev.vger.kernel.org Ray Lee wrote: > > If I could figure out a way to make it repeatable, I'd happily do a blind > bisect. As it stands, I can't trigger it manually. I've got a while true; do > iwconfig eth1; done running to hit the ioctls (as the trace in my first > message showed one of them in use), but that's might be a red herring. After > this email, I'll be shutting down all my browser and email windows to try to > kill the large network traffic generating apps as well. Also, does the 0x812 > core only hit for certain access points? I have a b and 2 b/g's in range. > > So, barring me finding a way to reproduce it, there's the good news that > there's only three bcm43xx patches between what worked and what didn't: > > ray@phoenix:~/work/kernel/linux-2.6$ hg log -I drivers/net/wireless/bcm43xx -r > v2.6.19-rc3:tip > changeset: 40500:4ef6746b2f06 > user: Al Viro > date: Wed Oct 25 12:01:11 2006 +0700 > summary: [PATCH] missing include of dma-mapping.h > > changeset: 40964:ca97546422bd > user: Michael Buesch > date: Wed Nov 01 08:15:40 2006 +0500 > summary: [PATCH] bcm43xx: Fix low-traffic netdev watchdog TX timeouts > > changeset: 40965:f5021f3521c2 > user: Larry Finger > date: Wed Nov 01 08:15:41 2006 +0500 > summary: [PATCH] bcm43xx: fix unexpected LED control values in BCM4303 sprom > > I think we can safely rule out a #include addition. The LED control values one > should be harmless too, yes? So I'm at a loss as to how it could be anything > but the middle one. > > I'm open to suggestions on how to make the problem trigger more than once > every two days... I don't know what might be causing the lock problems. I'm more concerned with the NETDEV WATCHDOG timeouts. AFAIK, you are the only one still reporting this error. On my system, I get an occasional MAC suspend failure, sometimes followed by an BCM43xx_IRQ_XMIT_ERROR. From what I read in your post, the timeouts happen a lot more often than once every two days. Once we get those fixed, then we can concentrate on the locking. Larry