Re: Crash with SMP on post 2.6.15 -git kernel

[Stephen Hemminger <shemminger@osdl.org>]

On Thu, 12 Jan 2006 00:36:25 +0100
Andi Kleen <ak@suse.de> wrote:

> On Wednesday 11 January 2006 23:57, Stephen Hemminger wrote:
> > On Wed, 11 Jan 2006 02:27:30 +0100
> > Andi Kleen <ak@suse.de> wrote:
> > 
> > > On Wednesday 11 January 2006 01:54, Stephen Hemminger wrote:
> > > 6
> > > > [   37.047264] CPU: L1 I Cache: 64K (64 bytes/line), D cache 64K (64 bytes/line)[   37.070722] CPU: L2 Cache: 1024K (64 bytes/line)
> > > > [   37.085894] mtrr: v2.0 (20020519)
> > > > [   37.350186] Using local APIC timer interrupts.
> > > > [   37.414873] Detected 12.464 MHz APIC timer.
> > > > [   37.428717] Booting processor 1/2 APIC 0x1
> > > > 
> > > > Machine then goes blank and reboots...
> > > 
> > > Don't know what it could be - I didn't merge anything. Maybe revert the kexec patches? 
> > > Does the -git6 snapshot still work?  Possibly do a binary search to narrow
> > > it down.
> > > 
> > > -Andi
> > That was not the bad config. It turns out the problem is the new code
> > for kdump. If CONFIG_KDUMP is turned on it crashes.
> 
> Ok thanks. I took a quick look and didn't saw anything obvious
> below arch/x86_64 that could cause it. Maybe will try later again in a simulator 
> or Vivek will probably fix it when he wakes up again.
> 

Could the PHYSICAL_START default be wrong for SMP on x86_64?
Never looked inside the boot startup code to check.
-- 
Stephen Hemminger <shemminger@osdl.org>
OSDL http://developer.osdl.org/~shemminger