[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: kernel panic with ext3 and cbq



Hi,

On Thu, Jul 12, 2001 at 05:36:42PM -0300, Renato Araújo wrote:
 
> I'm usign Kernel 2.2.19 with ext3 0.0.7b patches. I'm also
> using cbq to control bandwidth. The cbq is correctly creating
> the classes and queueing disciplines but,  whenever I call cbq stop
> I get a kernel panic:
> 
>  > Aiee, killing interrupt handler
>  > Kernel panic: Attempted to kill idle task!
>  > In interrupt handler - not syncing

Sounds like you have found a cbq bug.  ext3, by default, still
enables slab poisoning as a debugging aid.  It has, as a result,
uncovered bugs in several drivers so far.  If you look at the
mm/slab.c chunk in the ext3 patch, you'll see the bit which enables
slab poisoning.  I'll bet that if you apply just that chunk on its own
to your kernel, you'll see the problem appear --- that's cast iron
evidence that the driver is buggy; it's probably referencing memory
after it has been freed.

Cheers,
 Stephen





[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]