Re: Crashes in arm qemu emulations due to 'cpufreq: governor: Replace timers with utilization ...'

From: Russell King - ARM Linux
Date: Mon Feb 15 2016 - 14:07:23 EST


On Mon, Feb 15, 2016 at 07:54:26PM +0100, Rafael J. Wysocki wrote:
> On Mon, Feb 15, 2016 at 7:49 PM, Marc Zyngier <marc.zyngier@xxxxxxx> wrote:
> > Given that OMAP3 is a UP system, there is zero chance that it has
> > registered the magic hook that delivers IPIs (its interrupt controller
> > is not even capable of doing so).
> >
> > I don't really know the context, but IPIs on a UP system seem at best odd.
>
> That would explain it, thanks.
>
> So it looks like we should always use irq_work_queue() on UP even if
> CONFIG_SMP is set, shouldn't we?

irq_work_queue_on() doesn't check whether 'cpu' is the CPU that we're
running on. This is a problem where we want to be able to run a kernel
built for SMP on a UP system.

I guess the question is whether irq_work_queue_on() is buggy, or whether
our implementation of arch_send_call_function_single_ipi() is buggy.
Should arch_send_call_function_single_ipi() do something on UP systems,
if so what?

We don't have IPIs on UP systems, so we can't raise any interrupts.
So, should we call generic_smp_call_function_interrupt() directly
from it?

Some clues would be good...

--
RMK's Patch system: http://www.arm.linux.org.uk/developer/patches/
FTTC broadband for 0.8mile line: currently at 9.6Mbps down 400kbps up
according to speedtest.net.