Prev: [PATCH 0/4, v2] x86: enlightenment for ticket spin locks
Next: perf: Default PMU ops
From: Ronny Tschüter on 29 Jun 2010 10:50
I use Linux performance counters to trace some performance related
events. Unfortunately power:power_start events are not reported on my
machine. In /drivers/acpi/processor_idle.c I found three different ways
to call a C-state, but only one way (via architectural FFH based C-state
and using MONITOR/MWAIT) is currently instrumented. Here is my patch to
instrument the remaining two possibilities:

diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c
index b1b3856..08e7c42 100644
--- a/drivers/acpi/processor_idle.c
+++ b/drivers/acpi/processor_idle.c
@@ -43,6 +43,7 @@
#include <linux/clockchips.h>
#include <linux/cpuidle.h>
#include <linux/irqflags.h>
+#include <trace/events/power.h>

/*
* Include the apic definitions for x86 to have the APIC timer related
defines
@@ -136,6 +137,7 @@ static void acpi_safe_halt(void)
*/
smp_mb();
if (!need_resched()) {
+ trace_power_start(POWER_CSTATE,1);
safe_halt();
local_irq_disable();
}
@@ -797,6 +799,10 @@ static inline void acpi_idle_do_entry(struct
acpi_processor_cx *cx)
} else {
int unused;
/* IO port based C-state */
+ trace_power_start(POWER_CSTATE,
+ cx->type == ACPI_STATE_C1 ? 1 :
+ cx->type == ACPI_STATE_C2 ? 2 :
+ 3);
inb(cx->address);
/* Dummy wait op - must do something useless after P_LVL2 read
because chipsets cannot guarantee that STPCLK# signal
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
 | 
Pages: 1
Prev: [PATCH 0/4, v2] x86: enlightenment for ticket spin locks
Next: perf: Default PMU ops