sched: x86, track TSC-unstable events

track TSC-unstable events and propagate it to the scheduler code.
Also allow sched_clock() to be used when the TSC is unstable,
the rq_clock() wrapper creates a reliable clock out of it.

Signed-off-by: Ingo Molnar <mingo@elte.hu>

3 files changed, 17 insertions, 1 deletions

diff --git a/arch/i386/kernel/tsc.c b/arch/i386/kernel/tsc.c
index f64b81f3033b..ea63a30ca3e8 100644
--- a/arch/i386/kernel/tsc.c
+++ b/arch/i386/kernel/tsc.c
@@ -4,6 +4,7 @@
  * See comments there for proper credits.
  */
 
+#include <linux/sched.h>
 #include <linux/clocksource.h>
 #include <linux/workqueue.h>
 #include <linux/cpufreq.h>
@@ -106,8 +107,13 @@ unsigned long long sched_clock(void)
 
 	/*
 	 * Fall back to jiffies if there's no TSC available:
+	 * ( But note that we still use it if the TSC is marked
+	 *   unstable. We do this because unlike Time Of Day,
+	 *   the scheduler clock tolerates small errors and it's
+	 *   very important for it to be as fast as the platform
+	 *   can achive it. )
 	 */
-	if (unlikely(!tsc_enabled))
+	if (unlikely(!tsc_enabled && !tsc_unstable))
 		/* No locking but a rare wrong value is not a big deal: */
 		return (jiffies_64 - INITIAL_JIFFIES) * (1000000000 / HZ);
 
@@ -277,6 +283,7 @@ static struct clocksource clocksource_tsc = {
 
 void mark_tsc_unstable(char *reason)
 {
+	sched_clock_unstable_event();
 	if (!tsc_unstable) {
 		tsc_unstable = 1;
 		tsc_enabled = 0;
diff --git a/include/linux/sched.h b/include/linux/sched.h
index be2460e6f55b..fa895b309da0 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -1321,6 +1321,8 @@ extern void sched_exec(void);
 #define sched_exec()   {}
 #endif
 
+extern void sched_clock_unstable_event(void);
+
 #ifdef CONFIG_HOTPLUG_CPU
 extern void idle_task_exit(void);
 #else
diff --git a/kernel/sched.c b/kernel/sched.c
index 01ba4b1848a0..6150cd70f448 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -68,6 +68,13 @@ unsigned long long __attribute__((weak)) sched_clock(void)
 }
 
 /*
+ * CPU frequency is/was unstable - start new by setting prev_clock_raw:
+ */
+void sched_clock_unstable_event(void)
+{
+}
+
+/*
  * Convert user-nice values [ -20 ... 0 ... 19 ]
  * to static priority [ MAX_RT_PRIO..MAX_PRIO-1 ],
  * and back.