kdb patch for 2.6.11 and x86_64

[Ray Bryant]

On Thursday 25 August 2005 12:04, Jack F Vogel wrote:
> On Thu, Aug 25, 2005 at 10:51:14AM -0500, Ray Bryant wrote:
> > On Wednesday 24 August 2005 19:21, Jack F Vogel wrote:
> >
> > <snip>
> >
> > Hi Jack,
> >
> > > Hey Ray :)
> > >
> > > Is it possible for you to install SLES9 SP2 and check if
> > > the modular kdb there works?
> >
> > I tested it far enough to find out that I could get into KDB and type go
> > and have the system continue to run.   Didn't do anything more than a
> > "sniff" test though.
>
> I am pretty sure the problem is the changes that Jim
> posted, can you apply that and test. I am trying to
> reproduce and test that as well, but have next to no
> access to hardware right now.
>
> Jack

I applied Jim's patch.  The good news is that it applies cleanly.  The bad 
news is that it doesn't fix my problem.

I does appear that there is a bogus value in the notifier die_chain.   I put a 
dump_stack() in notifier_call_chain() and that is getting invoked from 
do_page_fault() so I am working under the assumption at the present time that 
I took an unhandled kernel page fault and the bogus value in the die_chain 
was causing the triple fault.   Why the first of these happened is still a 
mystery.  (2.6.11 without the kdb patches boots fine.)

At the moment I'm trying to track down where the bogus value in the die_chain 
is coming from.
-- 
Ray Bryant
AMD Performance Labs                   Austin, Tx
512-602-0038 (o)                 512-507-7807 (c)

---------------------------
Use http://oss.sgi.com/ecartis to modify your settings or to unsubscribe.