[RFC][PATCH 7/7] x86, tsc: Employ static_branch_likely()

From: Peter Zijlstra
Date: Fri Jul 24 2015 - 14:03:40 EST


Because of the static_key restrictions we had to take an unconditional
jump for the most likely case, causing $I bloat.

Rewrite to use the new primitives.

Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
---
arch/x86/kernel/tsc.c | 22 ++++++++++------------
1 file changed, 10 insertions(+), 12 deletions(-)

--- a/arch/x86/kernel/tsc.c
+++ b/arch/x86/kernel/tsc.c
@@ -38,7 +38,7 @@ static int __read_mostly tsc_unstable;
erroneous rdtsc usage on !cpu_has_tsc processors */
static int __read_mostly tsc_disabled = -1;

-static struct static_key __use_tsc = STATIC_KEY_INIT;
+static DEFINE_STATIC_KEY_FALSE(__use_tsc);

int tsc_clocksource_reliable;

@@ -274,7 +274,12 @@ static void set_cyc2ns_scale(unsigned lo
*/
u64 native_sched_clock(void)
{
- u64 tsc_now;
+ if (static_branch_likely(&__use_tsc)) {
+ u64 tsc_now = rdtsc();
+
+ /* return the value in ns */
+ return cycles_2_ns(tsc_now);
+ }

/*
* Fall back to jiffies if there's no TSC available:
@@ -284,16 +289,9 @@ u64 native_sched_clock(void)
* very important for it to be as fast as the platform
* can achieve it. )
*/
- if (!static_key_false(&__use_tsc)) {
- /* No locking but a rare wrong value is not a big deal: */
- return (jiffies_64 - INITIAL_JIFFIES) * (1000000000 / HZ);
- }
-
- /* read the Time Stamp Counter: */
- tsc_now = rdtsc();

- /* return the value in ns */
- return cycles_2_ns(tsc_now);
+ /* No locking but a rare wrong value is not a big deal: */
+ return (jiffies_64 - INITIAL_JIFFIES) * (1000000000 / HZ);
}

/* We need to define a real function for sched_clock, to override the
@@ -1204,7 +1202,7 @@ void __init tsc_init(void)
/* now allow native_sched_clock() to use rdtsc */

tsc_disabled = 0;
- static_key_slow_inc(&__use_tsc);
+ static_branch_enable(&__use_tsc);

if (!no_sched_irq_time)
enable_sched_clock_irqtime();


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/