Xen devel - Jun 2010 - [PATCH4/6] cpuidle: remove hpet access in hpet_broadcast_exit

cpuidle: remove hpet access in hpet_broadcast_exit

hpet_broadcast_exit calls reprogram_hpet to stop possible hpet intr if the last
deep-cstate waken up cpu is waken by unexpected intrs instead of hpet broadcast
handler. This can be removed without brings much useless intrs, but bring chance
for further optimization. It is a tradeoff between grace & optimization.

BTW, move the cpumask set out of critial section in hpet_broadcast_enter to
shorten it.

Signed-off-by: Wei Gang <gang.wei@intel.com>

diff -r 451a46c667c9 xen/arch/x86/hpet.c
--- a/xen/arch/x86/hpet.c	Sun Jun 13 07:23:26 2010 +0800
+++ b/xen/arch/x86/hpet.c	Sun Jun 13 07:27:14 2010 +0800
@@ -664,14 +664,12 @@ void hpet_broadcast_enter(void)
 
     /* Disable LAPIC timer interrupts. */
     disable_APIC_timer();
+    cpu_set(cpu, ch->cpumask);
 
     spin_lock(&ch->lock);
-
-    cpu_set(cpu, ch->cpumask);
     /* reprogram if current cpu expire time is nearer */
     if ( this_cpu(timer_deadline_end) < ch->next_event )
         reprogram_hpet_evt_channel(ch, this_cpu(timer_deadline_end), NOW(), 1);
-
     spin_unlock(&ch->lock);
 }
 
@@ -692,11 +690,7 @@ void hpet_broadcast_exit(void)
         raise_softirq(TIMER_SOFTIRQ);
 
     spin_lock_irq(&ch->lock);
-
     cpu_clear(cpu, ch->cpumask);
-    if ( cpus_empty(ch->cpumask) && ch->next_event != STIME_MAX )
-        reprogram_hpet_evt_channel(ch, STIME_MAX, 0, 0);
-
     spin_unlock_irq(&ch->lock);
 
     if ( ch != &legacy_hpet_event )

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

>>> On 17.06.10 at 09:38, "Wei, Gang"
<gang.wei@intel.com> wrote:
> BTW, move the cpumask set out of critial section in hpet_broadcast_enter to
> shorten it.
Could you explain why you can move it out of the locked section for
_enter(), but can''t for _exit() - this isn''t obvious to me
(and would
perhaps deserve a comment in the sources).

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

On Thursday, 2010-6-17 4:26 PM, Jan Beulich wrote:
>>>> On 17.06.10 at 09:38, "Wei, Gang"
<gang.wei@intel.com> wrote:
>> BTW, move the cpumask set out of critial section in
>> hpet_broadcast_enter to shorten it.
> 
> Could you explain why you can move it out of the locked section for
> _enter(), but can''t for _exit() - this isn''t obvious to
me (and would
> perhaps deserve a comment in the sources).
Keeping cpumask clear within the locked section in _exit() to avoid
handle_hpet_broadcast() to access other cpu''s timer_deadline_start/end
after the other cpu''s mask was cleared. For set, there are no such
issue. And cpumask set/clear it self should be atom type.

Jimmy
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel