Hi,

I don't think that runqueues(cpu)->idle is set when you try to access it
from head.S. It is much easier to reintroduce init_tasks and leave head.S
as it is. Then no changes are needed in Ingo's patch for sched.h and
sched.c. Ingo, could you please comment?

Please find attached an ia64 patch which should be applied over
2.4.17 + ia64 + sched-O1-2.4.17-H7.patch


It doesn't boot yet, the system (2CPU BigSur) crashes in
  schedule() <- cpu_idle() <- rest_init() <- start_kernel()
Maybe somebody has an idea?

Thanks,

Erich

---
Erich Focht                                    <efocht@ess.nec.de>
NEC European Supercomputer Systems, European HPC Technology Center


On Fri, 11 Jan 2002, Nick Pollitt wrote:

> I'm trying to get Ingo's scheduler working on IA64 but I've hit a 
> dead-end with the head.S code.  Ingo's patch removes init_tasks,
> so I've modified the assembly in head.S to point at 
> runqueues(cpu)->idle, I think - it dies very early in the boot, 
> and I'm not familiar with ia64 assembly.
> 
> Other issues, I had to build offsets.h by hand, and I moved some
> stuff from sched.c to sched.h.  Other than that, it's H6 + ia64.
> 
> Anyone have any feedback on getting this booting?
> 
> Thanks
> Nick
> 
> 
> On Fri, Jan 11, 2002 at 06:49:28PM +0100, Ingo Molnar wrote:
> > 
> > the -H6 patch is available:
> > 
> >     http://redhat.com/~mingo/O(1)-scheduler/sched-O1-2.5.2-pre11-H6.patch
> >     http://redhat.com/~mingo/O(1)-scheduler/sched-O1-2.4.17-H6.patch
> > 
> 
> -- 
> Nick Pollitt                                   phone: 650.933.7406
> Scalable Linux Project                           fax: 650.932.0317
> Silicon Graphics, Inc.                       npollitt@engr.sgi.com
>