On Thu, 7 Mar 2002, Anton Blanchard wrote:
> > we encountered problems with the initial distribution of the
> > migration_tasks across the CPUs. Machines with 16 and more CPUs
> > sometimes won't boot.
>
> We found this on a 31 way and have sent a fix to Ingo already, we needed
> to do a set_current_state(TASK_UNINTERRUPTIBLE) before schedule_timeout()
> or we just busy loop.
OK, good to know that it wasn't a tipical IA64 problem.
As far as I understand, your solution still relies on the specific
behavior of the scheduler for distributing the tasks uniformly and on
synchronizing among migration_tasks. I'd really prefer replacing this
shaky method by the usage of migration_task on CPU#0. Thich would always
work, no matter how many tasks are around. This saves some debugging work
to those who play around with the scheduler or want to start some kernel
threads earlier...
Regards,
Erich
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Mar 07 2002 - 21:01:03 EST