Marcelo,
This one bit me when I actually started using the (relatively) new
services disable_APIC_timer() and enable_APIC_timer(). Enable_APIC_timer()
is coded correctly; this patch fixes the bug in the disable service.
Please apply. Patch is against 2.4.19-pre10.
Joe
--- linux/arch/i386/kernel/apic.c.orig Thu Jun 6 15:21:26 2002
+++ linux/arch/i386/kernel/apic.c Thu Jun 6 15:24:05 2002
@@ -941,7 +941,7 @@
smp_call_function(setup_APIC_timer, (void *)calibration_result, 1, 1);
}
-void __init disable_APIC_timer(void)
+void disable_APIC_timer(void)
{
if (using_apic_timer) {
unsigned long v;
On Thu, 2002-06-06 at 21:59, Joe Korty wrote:
> Marcelo,
> This one bit me when I actually started using the (relatively) new
> services disable_APIC_timer() and enable_APIC_timer(). Enable_APIC_timer()
> is coded correctly; this patch fixes the bug in the disable service.
When is this getting called from a non boot up situation ?
> On Thu, 2002-06-06 at 21:59, Joe Korty wrote:
>> Marcelo,
>> This one bit me when I actually started using the (relatively) new
>> services disable_APIC_timer() and enable_APIC_timer(). Enable_APIC_timer()
>> is coded correctly; this patch fixes the bug in the disable service.
>
> When is this getting called from a non boot up situation ?
I am calling it from some cpu shielding code I've written and am
debugging.
Joe
On Fri, 2002-06-07 at 15:49, Joe Korty wrote:
>
> I am calling it from some cpu shielding code I've written and am
> debugging.
So the __init isnt actually a bug. It might be an appropriate change if
your code ever becomes part of the main tree thats all
> On Fri, 2002-06-07 at 15:49, Joe Korty wrote:
> >
> > I am calling it from some cpu shielding code I've written and am
> > debugging.
>
> So the __init isnt actually a bug. It might be an appropriate change if
> your code ever becomes part of the main tree thats all
True. In that case, an __init should be added to enable_APIC_timer.