lkml.org 
[lkml]   [2011]   [Mar]   [1]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Subject[PATCH v2 2/4] perf-events: Fix LLC-* events on Intel Nehalem/Westmere
From
Date
The generic perf LLC-* events do count the L2 caches, not the real
L3 LLC on Intel Nehalem and Westmere. This lead to quite some confusion.

Fixing this properly requires use of the special OFFCORE_RESPONSE
events which need a separate mask register. This has been implemented
in a earlier patch.

Now use this infrastructure to set correct events for the LLC-*
on Nehalem and Westmere

Signed-off-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Lin Ming <ming.m.lin@intel.com>
---
arch/x86/kernel/cpu/perf_event.c | 15 ++++--
arch/x86/kernel/cpu/perf_event_intel.c | 81 +++++++++++++++++++++++++++-----
2 files changed, 79 insertions(+), 17 deletions(-)

diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index ec1e5bf..28b475c 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -308,6 +308,10 @@ static u64 __read_mostly hw_cache_event_ids
[PERF_COUNT_HW_CACHE_MAX]
[PERF_COUNT_HW_CACHE_OP_MAX]
[PERF_COUNT_HW_CACHE_RESULT_MAX];
+static u64 __read_mostly hw_cache_extra_regs
+ [PERF_COUNT_HW_CACHE_MAX]
+ [PERF_COUNT_HW_CACHE_OP_MAX]
+ [PERF_COUNT_HW_CACHE_RESULT_MAX];

/*
* Propagate event elapsed time into the generic event.
@@ -522,8 +526,9 @@ static inline int x86_pmu_initialized(void)
}

static inline int
-set_ext_hw_attr(struct hw_perf_event *hwc, struct perf_event_attr *attr)
+set_ext_hw_attr(struct hw_perf_event *hwc, struct perf_event *event)
{
+ struct perf_event_attr *attr = &event->attr;
unsigned int cache_type, cache_op, cache_result;
u64 config, val;

@@ -550,8 +555,8 @@ set_ext_hw_attr(struct hw_perf_event *hwc, struct perf_event_attr *attr)
return -EINVAL;

hwc->config |= val;
-
- return 0;
+ attr->config1 = hw_cache_extra_regs[cache_type][cache_op][cache_result];
+ return x86_pmu_extra_regs(val, event);
}

static int x86_setup_perfctr(struct perf_event *event)
@@ -576,10 +581,10 @@ static int x86_setup_perfctr(struct perf_event *event)
}

if (attr->type == PERF_TYPE_RAW)
- return 0;
+ return x86_pmu_extra_regs(event->attr.config, event);

if (attr->type == PERF_TYPE_HW_CACHE)
- return set_ext_hw_attr(hwc, attr);
+ return set_ext_hw_attr(hwc, event);

if (attr->config >= x86_pmu.max_events)
return -EINVAL;
diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c
index 81db295..e82a7f9 100644
--- a/arch/x86/kernel/cpu/perf_event_intel.c
+++ b/arch/x86/kernel/cpu/perf_event_intel.c
@@ -172,16 +172,26 @@ static __initconst const u64 westmere_hw_cache_event_ids
},
[ C(LL ) ] = {
[ C(OP_READ) ] = {
- [ C(RESULT_ACCESS) ] = 0x0324, /* L2_RQSTS.LOADS */
- [ C(RESULT_MISS) ] = 0x0224, /* L2_RQSTS.LD_MISS */
+ /* OFFCORE_RESPONSE_0.ANY_DATA.LOCAL_CACHE */
+ [ C(RESULT_ACCESS) ] = 0x01b7,
+ /* OFFCORE_RESPONSE_1.ANY_DATA.ANY_LLC_MISS */
+ [ C(RESULT_MISS) ] = 0x01bb,
},
+ /*
+ * Use RFO, not WRITEBACK, because a write miss would typically occur
+ * on RFO.
+ */
[ C(OP_WRITE) ] = {
- [ C(RESULT_ACCESS) ] = 0x0c24, /* L2_RQSTS.RFOS */
- [ C(RESULT_MISS) ] = 0x0824, /* L2_RQSTS.RFO_MISS */
+ /* OFFCORE_RESPONSE_1.ANY_RFO.LOCAL_CACHE */
+ [ C(RESULT_ACCESS) ] = 0x01bb,
+ /* OFFCORE_RESPONSE_0.ANY_RFO.ANY_LLC_MISS */
+ [ C(RESULT_MISS) ] = 0x01b7,
},
[ C(OP_PREFETCH) ] = {
- [ C(RESULT_ACCESS) ] = 0x4f2e, /* LLC Reference */
- [ C(RESULT_MISS) ] = 0x412e, /* LLC Misses */
+ /* OFFCORE_RESPONSE_0.PREFETCH.LOCAL_CACHE */
+ [ C(RESULT_ACCESS) ] = 0x01b7,
+ /* OFFCORE_RESPONSE_1.PREFETCH.ANY_LLC_MISS */
+ [ C(RESULT_MISS) ] = 0x01bb,
},
},
[ C(DTLB) ] = {
@@ -228,6 +238,39 @@ static __initconst const u64 westmere_hw_cache_event_ids
},
};

+/*
+ * OFFCORE_RESPONSE MSR bits (subset), See IA32 SDM Vol 3 30.6.1.3
+ */
+
+#define DMND_DATA_RD (1 << 0)
+#define DMND_RFO (1 << 1)
+#define DMND_WB (1 << 3)
+#define PF_DATA_RD (1 << 4)
+#define PF_DATA_RFO (1 << 5)
+#define RESP_UNCORE_HIT (1 << 8)
+#define RESP_MISS (0xf600) /* non uncore hit */
+
+static __initconst const u64 nehalem_hw_cache_extra_regs
+ [PERF_COUNT_HW_CACHE_MAX]
+ [PERF_COUNT_HW_CACHE_OP_MAX]
+ [PERF_COUNT_HW_CACHE_RESULT_MAX] =
+{
+ [ C(LL ) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = DMND_DATA_RD|RESP_UNCORE_HIT,
+ [ C(RESULT_MISS) ] = DMND_DATA_RD|RESP_MISS,
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = DMND_RFO|DMND_WB|RESP_UNCORE_HIT,
+ [ C(RESULT_MISS) ] = DMND_RFO|DMND_WB|RESP_MISS,
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = PF_DATA_RD|PF_DATA_RFO|RESP_UNCORE_HIT,
+ [ C(RESULT_MISS) ] = PF_DATA_RD|PF_DATA_RFO|RESP_MISS,
+ },
+ }
+};
+
static __initconst const u64 nehalem_hw_cache_event_ids
[PERF_COUNT_HW_CACHE_MAX]
[PERF_COUNT_HW_CACHE_OP_MAX]
@@ -263,16 +306,26 @@ static __initconst const u64 nehalem_hw_cache_event_ids
},
[ C(LL ) ] = {
[ C(OP_READ) ] = {
- [ C(RESULT_ACCESS) ] = 0x0324, /* L2_RQSTS.LOADS */
- [ C(RESULT_MISS) ] = 0x0224, /* L2_RQSTS.LD_MISS */
+ /* OFFCORE_RESPONSE.ANY_DATA.LOCAL_CACHE */
+ [ C(RESULT_ACCESS) ] = 0x01b7,
+ /* OFFCORE_RESPONSE.ANY_DATA.ANY_LLC_MISS */
+ [ C(RESULT_MISS) ] = 0x01b7,
},
+ /*
+ * Use RFO, not WRITEBACK, because a write miss would typically occur
+ * on RFO.
+ */
[ C(OP_WRITE) ] = {
- [ C(RESULT_ACCESS) ] = 0x0c24, /* L2_RQSTS.RFOS */
- [ C(RESULT_MISS) ] = 0x0824, /* L2_RQSTS.RFO_MISS */
+ /* OFFCORE_RESPONSE.ANY_RFO.LOCAL_CACHE */
+ [ C(RESULT_ACCESS) ] = 0x01b7,
+ /* OFFCORE_RESPONSE.ANY_RFO.ANY_LLC_MISS */
+ [ C(RESULT_MISS) ] = 0x01b7,
},
[ C(OP_PREFETCH) ] = {
- [ C(RESULT_ACCESS) ] = 0x4f2e, /* LLC Reference */
- [ C(RESULT_MISS) ] = 0x412e, /* LLC Misses */
+ /* OFFCORE_RESPONSE.PREFETCH.LOCAL_CACHE */
+ [ C(RESULT_ACCESS) ] = 0x01b7,
+ /* OFFCORE_RESPONSE.PREFETCH.ANY_LLC_MISS */
+ [ C(RESULT_MISS) ] = 0x01b7,
},
},
[ C(DTLB) ] = {
@@ -1227,6 +1280,8 @@ static __init int intel_pmu_init(void)
case 46: /* 45 nm nehalem-ex, "Beckton" */
memcpy(hw_cache_event_ids, nehalem_hw_cache_event_ids,
sizeof(hw_cache_event_ids));
+ memcpy(hw_cache_extra_regs, nehalem_hw_cache_extra_regs,
+ sizeof(hw_cache_extra_regs));

intel_pmu_lbr_init_nhm();

@@ -1252,6 +1307,8 @@ static __init int intel_pmu_init(void)
case 44: /* 32 nm nehalem, "Gulftown" */
memcpy(hw_cache_event_ids, westmere_hw_cache_event_ids,
sizeof(hw_cache_event_ids));
+ memcpy(hw_cache_extra_regs, nehalem_hw_cache_extra_regs,
+ sizeof(hw_cache_extra_regs));

intel_pmu_lbr_init_nhm();

--
1.7.3







\
 
 \ /
  Last update: 2011-03-01 17:51    [W:0.037 / U:1.808 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site