perf/x86/intel: Add 1G DTLB load/store miss support for SKL
authorKan Liang <Kan.liang@intel.com>
Mon, 19 Jun 2017 14:26:09 +0000 (07:26 -0700)
committerIngo Molnar <mingo@kernel.org>
Thu, 22 Jun 2017 09:07:08 +0000 (11:07 +0200)
Current DTLB load/store miss events (0x608/0x649) only counts 4K,2M and
4M page size.
Need to extend the events to support any page size (4K/2M/4M/1G).

The complete DTLB load/store miss events are:

  DTLB_LOAD_MISSES.WALK_COMPLETED 0xe08
  DTLB_STORE_MISSES.WALK_COMPLETED 0xe49

Signed-off-by: Kan Liang <Kan.liang@intel.com>
Cc: <stable@vger.kernel.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: eranian@google.com
Link: http://lkml.kernel.org/r/20170619142609.11058-1-kan.liang@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
arch/x86/events/intel/core.c

index a6d91d4..110ce82 100644 (file)
@@ -431,11 +431,11 @@ static __initconst const u64 skl_hw_cache_event_ids
  [ C(DTLB) ] = {
        [ C(OP_READ) ] = {
                [ C(RESULT_ACCESS) ] = 0x81d0,  /* MEM_INST_RETIRED.ALL_LOADS */
-               [ C(RESULT_MISS)   ] = 0x608,   /* DTLB_LOAD_MISSES.WALK_COMPLETED */
+               [ C(RESULT_MISS)   ] = 0xe08,   /* DTLB_LOAD_MISSES.WALK_COMPLETED */
        },
        [ C(OP_WRITE) ] = {
                [ C(RESULT_ACCESS) ] = 0x82d0,  /* MEM_INST_RETIRED.ALL_STORES */
-               [ C(RESULT_MISS)   ] = 0x649,   /* DTLB_STORE_MISSES.WALK_COMPLETED */
+               [ C(RESULT_MISS)   ] = 0xe49,   /* DTLB_STORE_MISSES.WALK_COMPLETED */
        },
        [ C(OP_PREFETCH) ] = {
                [ C(RESULT_ACCESS) ] = 0x0,