Re: RFC/T: Possible fix for bcm43xx periodic work bug

[Michael Buesch <mb@bu3sch.de>]

On Friday 08 September 2006 11:42, Erik Mouw wrote:
> On Thu, Sep 07, 2006 at 01:17:05PM -0500, Larry Finger wrote:
> > I think I have a fix for the bcm43xx bug that leads to NETDEV WATCHDOG tx timeouts and would like it
> > to get as much testing as possible as this bug affects V2.6.18-rcX. If the problem is truly
> > fixed, I hope to get the fix into mainline before release of the bug into the stable series.
> 
> FWIW, I finally tracked down the bug that hangs my laptop to the
> bcm43xx driver. At first I got the impression it was the cpufreq code
> (which has been flaky in early 2.6.18-rc kernels), but after disabling
> that my laptop still crashed. After that I disabled preempt cause I got
> a couple of lockdep warnings when I had it enabled. That didn't make a
> difference, laptop still hangs after some time (runs a couple of hours
> at most). Right now I'm on wired network and my laptop finally doesn't
> hang anymore (up for two days).
> 
> The hang is very hard to trigger (i.e.: it happens at random, I see no
> pattern) and locks up the machine completely. I've tried to capture
> kernel messages through serial console, but that doesn't work (lock up
> before any messages are printed).
> 
> This is with any 2.6.18-rc kernel without additional patches or
> proprietary modules, 2.6.17 works ok.

The crash is fixed in wireless-2.6.
The actual cause of the controller restart not. So at least it
does not crash anymore.

-- 
Greetings Michael.