[openib-general] OOPS: ib_mad crashery on bootup

[Sean Hefty]

Hal Rosenstock wrote:
> On Fri, 2005-05-20 at 13:27, Tom Duffy wrote:
> 
>>On Fri, 2005-05-20 at 09:54 -0700, Shirley Ma wrote:
>>
>>>>Code: 4c 8b 7d 20 48 89 04 24 48 89 ef 31 db e8 bf 3c 25 f8 49 8b
>>>>RIP <ffffffff880f633f>{:ib_mad:ib_mad_send_done_handler+31}
> 
> 
> ib_mad_send_done_handler+31 looks like the spin_lock_irqsave call

Possibly the MAD or work completion or stack has been corrupted?

I've tried to reproduce this running MAD stress tests over just the core, 
mad, and mt modules, but haven't been successful yet.

- Sean