Re: [PATCH v3 1/3] perf: Use monotonic clock as a source for timestamps

On Tue, 2014-11-04 at 07:23 +0000, Peter Zijlstra wrote:

On Tue, Nov 04, 2014 at 12:28:36AM +0000, Pawel Moll wrote:

quoted

+int sysctl_perf_sample_time_clk_id = CLOCK_MONOTONIC;

const ?

Sure (unless we have to change it as mentioned below)

quoted

 /*
  * perf samples are done in some very critical code paths (NMIs).
  * If they take too much CPU time, the system can lock up and not

@@ -324,7 +326,7 @@ extern __weak const char *perf_pmu_name(void)
 
 static inline u64 perf_clock(void)
 {
-	return local_clock();
+	return ktime_get_mono_fast_ns();
 }

Do we maybe want to make it boot-time switchable back to local_clock for
people with bad systems and or backwards compat issues?

Very good idea, should have came up with it myself :-)

Does __setup("perf_use_local_clock") sound reasonable? Then we have to
decide whether to hide the sysctl "perf_sample_time_clk_id" (my
preferred option, will see how difficult it is) or just provide an
invalid clock_id (eg. -1) in it.

Cheers!

Pawel