[PATCH -mm] x86_64: Adjust the timing of initializing cyc2ns_scale.

[PATCH -mm] x86_64: Adjust the timing of initializing cyc2ns_scale.

[Toyo Abe]

The x86_64-mm-monotonic-clock.patch in 2.6.18-rc4-mm2 made a change to
the updating of monotonic_base. It now uses cycles_2_ns().

I suggest that a set_cyc2ns_scale() should be done prior to the setup_irq().
Because cycles_2_ns() can be called from the timer ISR right after the irq0
is enabled.


Signed-off-by: Toyo Abe <toyoa@mvista.com>

---

 arch/x86_64/kernel/time.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/x86_64/kernel/time.c b/arch/x86_64/kernel/time.c
index 5bbd05d..03ddc1a 100644
--- a/arch/x86_64/kernel/time.c
+++ b/arch/x86_64/kernel/time.c
@@ -928,10 +928,10 @@ #endif
 	vxtime.quot = (USEC_PER_SEC << US_SCALE) / vxtime_hz;
 	vxtime.tsc_quot = (USEC_PER_MSEC << US_SCALE) / cpu_khz;
 	vxtime.last_tsc = get_cycles_sync();
-	setup_irq(0, &irq0);
-
 	set_cyc2ns_scale(cpu_khz);
 
+	setup_irq(0, &irq0);
+
 	hotcpu_notifier(time_cpu_notifier, 0);
 	time_cpu_notifier(NULL, CPU_ONLINE, (void *)(long)smp_processor_id());
 

Re: [PATCH -mm] x86_64: Adjust the timing of initializing cyc2ns_scale.

[Andi Kleen]

On Friday 25 August 2006 18:45, Toyo Abe wrote:
> The x86_64-mm-monotonic-clock.patch in 2.6.18-rc4-mm2 made a change to
> the updating of monotonic_base. It now uses cycles_2_ns().
> 
> I suggest that a set_cyc2ns_scale() should be done prior to the setup_irq().
> Because cycles_2_ns() can be called from the timer ISR right after the irq0
> is enabled.

Added thanks. I folded it into the original patch.

Did you actually see a failure or was this just from code review?
-Andi

Re: [PATCH -mm] x86_64: Adjust the timing of initializing cyc2ns_scale.

[Toyo]

From: Andi Kleen <ak@suse.de>
Subject: Re: [PATCH -mm] x86_64: Adjust the timing of initializing cyc2ns_scale.
Date: Fri, 25 Aug 2006 18:55:57 +0200

> On Friday 25 August 2006 18:45, Toyo Abe wrote:
> > The x86_64-mm-monotonic-clock.patch in 2.6.18-rc4-mm2 made a change to
> > the updating of monotonic_base. It now uses cycles_2_ns().
> > 
> > I suggest that a set_cyc2ns_scale() should be done prior to the setup_irq().
> > Because cycles_2_ns() can be called from the timer ISR right after the irq0
> > is enabled.
> 
> Added thanks. I folded it into the original patch.
> 
> Did you actually see a failure or was this just from code review?
> -Andi
> 
I didn't see any failure on it. It was just a speculation. But I think
it's better thing here.

Best Regards,
toyo