[pvrusb2] Kernel oops in pvrusb2 driver

[Mike Isely]

Handling hot unplug in any driver state has always been a difficult 
business.  I'm actually kind of surprised that we've gone on this long 
(years) without another race like this being reported.  I need to 
re-evaluate how this is being handled.  It's possible that changes in 
the surrounding kernel may now be invalidating the strategy I had been 
using in the driver to cleanly self-tear-down when a hot-unplug happens.

Actually it's been quite a while since I've dug into the driver.  One of 
the key issues has been just not having enough time.  However with the 
new year, one of the big time sinks that had been distracting me is now 
officially ended, so I will see about digging back into all this again!

  -Mike


On Tue, 10 Jan 2012, Gary Buhrmaster wrote:

> On Tue, Jan 10, 2012 at 19:06, Mike Isely <isely at isely.net> wrote:
> >
> > Well even if it's unlikely, a kernel oops should still be chased.  Was
> > this on an SMP system?
> 
> Yes, i5-2500, 4 CPUs.
> 
> > And I assume you were running the pvrusb2 driver
> > that is part of the 3.1.7 kernel source tree (as opposed to compiling a
> > separate driver)?
> 
> Correct.  I can test with additional debugging options, or
> compile driver code if that is helpful.  I figured a reliable
> reproducer was a good first step in the process...
> 
> Gary
> _______________________________________________
> pvrusb2 mailing list
> pvrusb2 at isely.net
> http://www.isely.net/cgi-bin/mailman/listinfo/pvrusb2
> 

-- 

Mike Isely
isely @ isely (dot) net
PGP: 03 54 43 4D 75 E5 CC 92 71 16 01 E2 B5 F5 C1 E8