perf events: Add Intel x86 mapping for PERF_COUNT_HW_REF_CPU_CYCLES
authorStephane Eranian <eranian@google.com>
Sat, 10 Dec 2011 23:28:53 +0000 (00:28 +0100)
committerIngo Molnar <mingo@elte.hu>
Wed, 21 Dec 2011 09:26:39 +0000 (10:26 +0100)
Add event maps for Intel x86 processors (with architected PMU v2 or later).

On AMD, there is frequency scaling but no Turbo. There is no core
cycle event not subject to frequency scaling, therefore we do not
provide a mapping.

Signed-off-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1323559734-3488-4-git-send-email-eranian@google.com
Signed-off-by: Ingo Molnar <mingo@elte.hu>
arch/x86/kernel/cpu/perf_event_intel.c

index 61f865f947b3ec0002dced7ef777bf9473e97427..cbfaaa2475eabde2c56936d301297178bea34dd1 100644 (file)
@@ -28,6 +28,7 @@ static u64 intel_perfmon_event_map[PERF_COUNT_HW_MAX] __read_mostly =
   [PERF_COUNT_HW_BRANCH_INSTRUCTIONS]  = 0x00c4,
   [PERF_COUNT_HW_BRANCH_MISSES]                = 0x00c5,
   [PERF_COUNT_HW_BUS_CYCLES]           = 0x013c,
+  [PERF_COUNT_HW_REF_CPU_CYCLES]       = 0x0300, /* pseudo-encoding */
 };
 
 static struct event_constraint intel_core_event_constraints[] __read_mostly =