Re: Oops in 2.6.10-rc1

[Linus Torvalds <torvalds@osdl.org>]

On Sun, 7 Nov 2004, Christian Kujau wrote:
> 
> if someone could give me a hint here what to do next or perhaps tell me
> that the whole things was totally pointless - please say so.
> i am somehow lost as to which is the right person to bug here.

Since you seem to be a BK user, try doing a

	bk revtool sound/pci/ens1370.c 

and see if you can find the change that caused your problem. Of course, 
the real change might be somewhere else in the sound driver initialization 
path, so it's not like just that one file might be the cause. Regardöess, 
the more you can pinpoint when the problem started, the better.

Also, if you enable frame pointers (under kernel debugging), the traceback
will look a bit better. As it is, your oops looks looks like something has
jumped off into la-la-land by jumping through a bad pointer (the value is
still in %ecx), but it's definitely not clear _where_ that happened.  
Your trace points to pci_enable_device_bars(), but that may well be just
stale stack contents.

		Linus