On Friday 30 October 2009 14:36:31 Jason Wessel wrote:
> Thomas Renninger wrote:
> > Hi,
> >
> > we currently have a problem with latest Intel CPUs/BIOSes:
> > If the machine gets booted it hangs in:
> > init/main.c:
> > kdb_init();
> >
> > As soon as an irq via keyboard is triggered, the machine continues
> > booting and everything is fine.
> >
> > Is this known already? Can someone point me to a patch?
> >
> > If not, is there a repo having the latest code, I could try
> > out whether the latest version works.
> >
> > IMO something like the attached patch makes sense?
> > This fixes the problem with kdb=off (our default).
> > But I am anxious that kdb code could still get activitated somehow
> > and the machine might oops when trying to access not initialized
> > kdb structures? Could someone give an estimation how risky that
> > patch really is?
> >
> > Still the real bug should be found and if this is not known, can someone
> > give me a pointer to what this problem could be related and where I
> > should look closer at, please.
> >
>
> In the version of kdb which attached to kgdb, I wonder if you see the
> same sort of behavior?
> I am not sure which kernel base you need, but kdb prototype attached to
> kgdb exists for the 2.6.31 and 2.6.32 kernel.
This is about SLES11 (2.6.27 based)
The header in the main kdb patch is: kdb-v4.4-2.6.27-rc8-common-1.1
The same happens with our current factory (2.6.32) and therefore probably
also with our openSUSE 11.2 (gets released soon, we are at RC2) kernel.
The kdb patch seem to got an upgrade, but is still somewhat old (for
factory and 11.2)?:
The header there is: kdb-v4.4-2.6.30-common-4
Hmm, I wonder whether we should just rip out kdb for openSUSE if we are
on an old codebase there or at least do an upgrade (which might be a bit
late...).
> The code you patched does not exist in the kdb+kgdb code base.
If this is solved in latest kdb versions, I wonder whether my patch is
a *really* safe option for our SLE11 maintenance kernel and the root cause
does not need to be find for older verions. I can hardly
judge whether later code might use an uninitialized (kdb_init() not called)
structure even with kdb_on=0, theoretically kdb code should never be accessed
with kdb=off?
I'll browse the git repo, thanks a lot for your quick help.
Any other hint is very much appreciated of course.
Thomas
_______________________________________________
kdb mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/kdb
