struct x86_pmu {
const char *name;
int version;
- int (*handle_irq)(struct pt_regs *, int);
+ int (*handle_irq)(struct pt_regs *);
void (*disable_all)(void);
void (*enable_all)(void);
void (*enable)(struct hw_perf_counter *, int);
*/
static const u64 intel_perfmon_event_map[] =
{
- [PERF_COUNT_CPU_CYCLES] = 0x003c,
- [PERF_COUNT_INSTRUCTIONS] = 0x00c0,
- [PERF_COUNT_CACHE_REFERENCES] = 0x4f2e,
- [PERF_COUNT_CACHE_MISSES] = 0x412e,
- [PERF_COUNT_BRANCH_INSTRUCTIONS] = 0x00c4,
- [PERF_COUNT_BRANCH_MISSES] = 0x00c5,
- [PERF_COUNT_BUS_CYCLES] = 0x013c,
+ [PERF_COUNT_HW_CPU_CYCLES] = 0x003c,
+ [PERF_COUNT_HW_INSTRUCTIONS] = 0x00c0,
+ [PERF_COUNT_HW_CACHE_REFERENCES] = 0x4f2e,
+ [PERF_COUNT_HW_CACHE_MISSES] = 0x412e,
+ [PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = 0x00c4,
+ [PERF_COUNT_HW_BRANCH_MISSES] = 0x00c5,
+ [PERF_COUNT_HW_BUS_CYCLES] = 0x013c,
};
static u64 intel_pmu_event_map(int event)
return intel_perfmon_event_map[event];
}
+/*
+ * Generalized hw caching related event table, filled
+ * in on a per model basis. A value of 0 means
+ * 'not supported', -1 means 'event makes no sense on
+ * this CPU', any other value means the raw event
+ * ID.
+ */
+
+#define C(x) PERF_COUNT_HW_CACHE_##x
+
+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 const u64 nehalem_hw_cache_event_ids
+ [PERF_COUNT_HW_CACHE_MAX]
+ [PERF_COUNT_HW_CACHE_OP_MAX]
+ [PERF_COUNT_HW_CACHE_RESULT_MAX] =
+{
+ [ C(L1D) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x0f40, /* L1D_CACHE_LD.MESI */
+ [ C(RESULT_MISS) ] = 0x0140, /* L1D_CACHE_LD.I_STATE */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = 0x0f41, /* L1D_CACHE_ST.MESI */
+ [ C(RESULT_MISS) ] = 0x0141, /* L1D_CACHE_ST.I_STATE */
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = 0x014e, /* L1D_PREFETCH.REQUESTS */
+ [ C(RESULT_MISS) ] = 0x024e, /* L1D_PREFETCH.MISS */
+ },
+ },
+ [ C(L1I ) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x0380, /* L1I.READS */
+ [ C(RESULT_MISS) ] = 0x0280, /* L1I.MISSES */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = 0x0,
+ [ C(RESULT_MISS) ] = 0x0,
+ },
+ },
+ [ C(LL ) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x0324, /* L2_RQSTS.LOADS */
+ [ C(RESULT_MISS) ] = 0x0224, /* L2_RQSTS.LD_MISS */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = 0x0c24, /* L2_RQSTS.RFOS */
+ [ C(RESULT_MISS) ] = 0x0824, /* L2_RQSTS.RFO_MISS */
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = 0x4f2e, /* LLC Reference */
+ [ C(RESULT_MISS) ] = 0x412e, /* LLC Misses */
+ },
+ },
+ [ C(DTLB) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x0f40, /* L1D_CACHE_LD.MESI (alias) */
+ [ C(RESULT_MISS) ] = 0x0108, /* DTLB_LOAD_MISSES.ANY */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = 0x0f41, /* L1D_CACHE_ST.MESI (alias) */
+ [ C(RESULT_MISS) ] = 0x010c, /* MEM_STORE_RETIRED.DTLB_MISS */
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = 0x0,
+ [ C(RESULT_MISS) ] = 0x0,
+ },
+ },
+ [ C(ITLB) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x01c0, /* INST_RETIRED.ANY_P */
+ [ C(RESULT_MISS) ] = 0x20c8, /* ITLB_MISS_RETIRED */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ },
+ [ C(BPU ) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x00c4, /* BR_INST_RETIRED.ALL_BRANCHES */
+ [ C(RESULT_MISS) ] = 0x03e8, /* BPU_CLEARS.ANY */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ },
+};
+
+static const u64 core2_hw_cache_event_ids
+ [PERF_COUNT_HW_CACHE_MAX]
+ [PERF_COUNT_HW_CACHE_OP_MAX]
+ [PERF_COUNT_HW_CACHE_RESULT_MAX] =
+{
+ [ C(L1D) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x0f40, /* L1D_CACHE_LD.MESI */
+ [ C(RESULT_MISS) ] = 0x0140, /* L1D_CACHE_LD.I_STATE */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = 0x0f41, /* L1D_CACHE_ST.MESI */
+ [ C(RESULT_MISS) ] = 0x0141, /* L1D_CACHE_ST.I_STATE */
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = 0x104e, /* L1D_PREFETCH.REQUESTS */
+ [ C(RESULT_MISS) ] = 0,
+ },
+ },
+ [ C(L1I ) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x0080, /* L1I.READS */
+ [ C(RESULT_MISS) ] = 0x0081, /* L1I.MISSES */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = 0,
+ [ C(RESULT_MISS) ] = 0,
+ },
+ },
+ [ C(LL ) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x4f29, /* L2_LD.MESI */
+ [ C(RESULT_MISS) ] = 0x4129, /* L2_LD.ISTATE */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = 0x4f2A, /* L2_ST.MESI */
+ [ C(RESULT_MISS) ] = 0x412A, /* L2_ST.ISTATE */
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = 0,
+ [ C(RESULT_MISS) ] = 0,
+ },
+ },
+ [ C(DTLB) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x0f40, /* L1D_CACHE_LD.MESI (alias) */
+ [ C(RESULT_MISS) ] = 0x0208, /* DTLB_MISSES.MISS_LD */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = 0x0f41, /* L1D_CACHE_ST.MESI (alias) */
+ [ C(RESULT_MISS) ] = 0x0808, /* DTLB_MISSES.MISS_ST */
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = 0,
+ [ C(RESULT_MISS) ] = 0,
+ },
+ },
+ [ C(ITLB) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x00c0, /* INST_RETIRED.ANY_P */
+ [ C(RESULT_MISS) ] = 0x1282, /* ITLBMISSES */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ },
+ [ C(BPU ) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x00c4, /* BR_INST_RETIRED.ANY */
+ [ C(RESULT_MISS) ] = 0x00c5, /* BP_INST_RETIRED.MISPRED */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ },
+};
+
+static const u64 atom_hw_cache_event_ids
+ [PERF_COUNT_HW_CACHE_MAX]
+ [PERF_COUNT_HW_CACHE_OP_MAX]
+ [PERF_COUNT_HW_CACHE_RESULT_MAX] =
+{
+ [ C(L1D) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x2140, /* L1D_CACHE.LD */
+ [ C(RESULT_MISS) ] = 0,
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = 0x2240, /* L1D_CACHE.ST */
+ [ C(RESULT_MISS) ] = 0,
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = 0x0,
+ [ C(RESULT_MISS) ] = 0,
+ },
+ },
+ [ C(L1I ) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x0380, /* L1I.READS */
+ [ C(RESULT_MISS) ] = 0x0280, /* L1I.MISSES */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = 0,
+ [ C(RESULT_MISS) ] = 0,
+ },
+ },
+ [ C(LL ) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x4f29, /* L2_LD.MESI */
+ [ C(RESULT_MISS) ] = 0x4129, /* L2_LD.ISTATE */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = 0x4f2A, /* L2_ST.MESI */
+ [ C(RESULT_MISS) ] = 0x412A, /* L2_ST.ISTATE */
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = 0,
+ [ C(RESULT_MISS) ] = 0,
+ },
+ },
+ [ C(DTLB) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x2140, /* L1D_CACHE_LD.MESI (alias) */
+ [ C(RESULT_MISS) ] = 0x0508, /* DTLB_MISSES.MISS_LD */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = 0x2240, /* L1D_CACHE_ST.MESI (alias) */
+ [ C(RESULT_MISS) ] = 0x0608, /* DTLB_MISSES.MISS_ST */
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = 0,
+ [ C(RESULT_MISS) ] = 0,
+ },
+ },
+ [ C(ITLB) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x00c0, /* INST_RETIRED.ANY_P */
+ [ C(RESULT_MISS) ] = 0x0282, /* ITLB.MISSES */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ },
+ [ C(BPU ) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x00c4, /* BR_INST_RETIRED.ANY */
+ [ C(RESULT_MISS) ] = 0x00c5, /* BP_INST_RETIRED.MISPRED */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ },
+};
+
static u64 intel_pmu_raw_event(u64 event)
{
#define CORE_EVNTSEL_EVENT_MASK 0x000000FFULL
#define CORE_EVNTSEL_INV_MASK 0x00800000ULL
#define CORE_EVNTSEL_COUNTER_MASK 0xFF000000ULL
-#define CORE_EVNTSEL_MASK \
+#define CORE_EVNTSEL_MASK \
(CORE_EVNTSEL_EVENT_MASK | \
CORE_EVNTSEL_UNIT_MASK | \
CORE_EVNTSEL_EDGE_MASK | \
return event & CORE_EVNTSEL_MASK;
}
+static const u64 amd_0f_hw_cache_event_ids
+ [PERF_COUNT_HW_CACHE_MAX]
+ [PERF_COUNT_HW_CACHE_OP_MAX]
+ [PERF_COUNT_HW_CACHE_RESULT_MAX] =
+{
+ [ C(L1D) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0,
+ [ C(RESULT_MISS) ] = 0,
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = 0,
+ [ C(RESULT_MISS) ] = 0,
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = 0,
+ [ C(RESULT_MISS) ] = 0,
+ },
+ },
+ [ C(L1I ) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x0080, /* Instruction cache fetches */
+ [ C(RESULT_MISS) ] = 0x0081, /* Instruction cache misses */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = 0,
+ [ C(RESULT_MISS) ] = 0,
+ },
+ },
+ [ C(LL ) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0,
+ [ C(RESULT_MISS) ] = 0,
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = 0,
+ [ C(RESULT_MISS) ] = 0,
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = 0,
+ [ C(RESULT_MISS) ] = 0,
+ },
+ },
+ [ C(DTLB) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0,
+ [ C(RESULT_MISS) ] = 0,
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = 0,
+ [ C(RESULT_MISS) ] = 0,
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = 0,
+ [ C(RESULT_MISS) ] = 0,
+ },
+ },
+ [ C(ITLB) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x0080, /* Instruction fecthes */
+ [ C(RESULT_MISS) ] = 0x0085, /* Instr. fetch ITLB misses */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ },
+ [ C(BPU ) ] = {
+ [ C(OP_READ) ] = {
+ [ C(RESULT_ACCESS) ] = 0x00c2, /* Retired Branch Instr. */
+ [ C(RESULT_MISS) ] = 0x00c3, /* Retired Mispredicted BI */
+ },
+ [ C(OP_WRITE) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ [ C(OP_PREFETCH) ] = {
+ [ C(RESULT_ACCESS) ] = -1,
+ [ C(RESULT_MISS) ] = -1,
+ },
+ },
+};
+
/*
* AMD Performance Monitor K7 and later.
*/
static const u64 amd_perfmon_event_map[] =
{
- [PERF_COUNT_CPU_CYCLES] = 0x0076,
- [PERF_COUNT_INSTRUCTIONS] = 0x00c0,
- [PERF_COUNT_CACHE_REFERENCES] = 0x0080,
- [PERF_COUNT_CACHE_MISSES] = 0x0081,
- [PERF_COUNT_BRANCH_INSTRUCTIONS] = 0x00c4,
- [PERF_COUNT_BRANCH_MISSES] = 0x00c5,
+ [PERF_COUNT_HW_CPU_CYCLES] = 0x0076,
+ [PERF_COUNT_HW_INSTRUCTIONS] = 0x00c0,
+ [PERF_COUNT_HW_CACHE_REFERENCES] = 0x0080,
+ [PERF_COUNT_HW_CACHE_MISSES] = 0x0081,
+ [PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = 0x00c4,
+ [PERF_COUNT_HW_BRANCH_MISSES] = 0x00c5,
};
static u64 amd_pmu_event_map(int event)
return x86_pmu.handle_irq != NULL;
}
+static inline int
+set_ext_hw_attr(struct hw_perf_counter *hwc, struct perf_counter_attr *attr)
+{
+ unsigned int cache_type, cache_op, cache_result;
+ u64 config, val;
+
+ config = attr->config;
+
+ cache_type = (config >> 0) & 0xff;
+ if (cache_type >= PERF_COUNT_HW_CACHE_MAX)
+ return -EINVAL;
+
+ cache_op = (config >> 8) & 0xff;
+ if (cache_op >= PERF_COUNT_HW_CACHE_OP_MAX)
+ return -EINVAL;
+
+ cache_result = (config >> 16) & 0xff;
+ if (cache_result >= PERF_COUNT_HW_CACHE_RESULT_MAX)
+ return -EINVAL;
+
+ val = hw_cache_event_ids[cache_type][cache_op][cache_result];
+
+ if (val == 0)
+ return -ENOENT;
+
+ if (val == -1)
+ return -EINVAL;
+
+ hwc->config |= val;
+
+ return 0;
+}
+
/*
- * Setup the hardware configuration for a given hw_event_type
+ * Setup the hardware configuration for a given attr_type
*/
static int __hw_perf_counter_init(struct perf_counter *counter)
{
- struct perf_counter_hw_event *hw_event = &counter->hw_event;
+ struct perf_counter_attr *attr = &counter->attr;
struct hw_perf_counter *hwc = &counter->hw;
int err;
/*
* Count user and OS events unless requested not to.
*/
- if (!hw_event->exclude_user)
+ if (!attr->exclude_user)
hwc->config |= ARCH_PERFMON_EVENTSEL_USR;
- if (!hw_event->exclude_kernel)
+ if (!attr->exclude_kernel)
hwc->config |= ARCH_PERFMON_EVENTSEL_OS;
- /*
- * Use NMI events all the time:
- */
- hwc->nmi = 1;
- hw_event->nmi = 1;
-
- if (!hwc->sample_period)
+ if (!hwc->sample_period) {
hwc->sample_period = x86_pmu.max_period;
+ hwc->last_period = hwc->sample_period;
+ atomic64_set(&hwc->period_left, hwc->sample_period);
+ }
- atomic64_set(&hwc->period_left,
- min(x86_pmu.max_period, hwc->sample_period));
+ counter->destroy = hw_perf_counter_destroy;
/*
* Raw event type provide the config in the event structure
*/
- if (perf_event_raw(hw_event)) {
- hwc->config |= x86_pmu.raw_event(perf_event_config(hw_event));
- } else {
- if (perf_event_id(hw_event) >= x86_pmu.max_events)
- return -EINVAL;
- /*
- * The generic map:
- */
- hwc->config |= x86_pmu.event_map(perf_event_id(hw_event));
+ if (attr->type == PERF_TYPE_RAW) {
+ hwc->config |= x86_pmu.raw_event(attr->config);
+ return 0;
}
- counter->destroy = hw_perf_counter_destroy;
+ if (attr->type == PERF_TYPE_HW_CACHE)
+ return set_ext_hw_attr(hwc, attr);
+
+ if (attr->config >= x86_pmu.max_events)
+ return -EINVAL;
+ /*
+ * The generic map:
+ */
+ hwc->config |= x86_pmu.event_map(attr->config);
return 0;
}
* Set the next IRQ period, based on the hwc->period_left value.
* To be called with the counter disabled in hw:
*/
-static void
+static int
x86_perf_counter_set_period(struct perf_counter *counter,
struct hw_perf_counter *hwc, int idx)
{
s64 left = atomic64_read(&hwc->period_left);
- s64 period = min(x86_pmu.max_period, hwc->sample_period);
- int err;
+ s64 period = hwc->sample_period;
+ int err, ret = 0;
/*
* If we are way outside a reasoable range then just skip forward:
if (unlikely(left <= -period)) {
left = period;
atomic64_set(&hwc->period_left, left);
+ hwc->last_period = period;
+ ret = 1;
}
if (unlikely(left <= 0)) {
left += period;
atomic64_set(&hwc->period_left, left);
+ hwc->last_period = period;
+ ret = 1;
}
/*
* Quirk: certain CPUs dont like it if just 1 event is left:
if (unlikely(left < 2))
left = 2;
+ if (left > x86_pmu.max_period)
+ left = x86_pmu.max_period;
+
per_cpu(prev_left[idx], smp_processor_id()) = left;
/*
err = checking_wrmsrl(hwc->counter_base + idx,
(u64)(-left) & x86_pmu.counter_mask);
+
+ return ret;
}
static inline void
if (!x86_pmu.num_counters_fixed)
return -1;
+ /*
+ * Quirk, IA32_FIXED_CTRs do not work on current Atom processors:
+ */
+ if (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL &&
+ boot_cpu_data.x86_model == 28)
+ return -1;
+
event = hwc->config & ARCH_PERFMON_EVENT_MASK;
- if (unlikely(event == x86_pmu.event_map(PERF_COUNT_INSTRUCTIONS)))
+ if (unlikely(event == x86_pmu.event_map(PERF_COUNT_HW_INSTRUCTIONS)))
return X86_PMC_IDX_FIXED_INSTRUCTIONS;
- if (unlikely(event == x86_pmu.event_map(PERF_COUNT_CPU_CYCLES)))
+ if (unlikely(event == x86_pmu.event_map(PERF_COUNT_HW_CPU_CYCLES)))
return X86_PMC_IDX_FIXED_CPU_CYCLES;
- if (unlikely(event == x86_pmu.event_map(PERF_COUNT_BUS_CYCLES)))
+ if (unlikely(event == x86_pmu.event_map(PERF_COUNT_HW_BUS_CYCLES)))
return X86_PMC_IDX_FIXED_BUS_CYCLES;
return -1;
* Save and restart an expired counter. Called by NMI contexts,
* so it has to be careful about preempting normal counter ops:
*/
-static void intel_pmu_save_and_restart(struct perf_counter *counter)
+static int intel_pmu_save_and_restart(struct perf_counter *counter)
{
struct hw_perf_counter *hwc = &counter->hw;
int idx = hwc->idx;
+ int ret;
x86_perf_counter_update(counter, hwc, idx);
- x86_perf_counter_set_period(counter, hwc, idx);
+ ret = x86_perf_counter_set_period(counter, hwc, idx);
if (counter->state == PERF_COUNTER_STATE_ACTIVE)
intel_pmu_enable_counter(hwc, idx);
+
+ return ret;
}
static void intel_pmu_reset(void)
* This handler is triggered by the local APIC, so the APIC IRQ handling
* rules apply:
*/
-static int intel_pmu_handle_irq(struct pt_regs *regs, int nmi)
+static int intel_pmu_handle_irq(struct pt_regs *regs)
{
+ struct perf_sample_data data;
struct cpu_hw_counters *cpuc;
- struct cpu_hw_counters;
int bit, cpu, loops;
u64 ack, status;
+ data.regs = regs;
+ data.addr = 0;
+
cpu = smp_processor_id();
cpuc = &per_cpu(cpu_hw_counters, cpu);
if (!test_bit(bit, cpuc->active_mask))
continue;
- intel_pmu_save_and_restart(counter);
- if (perf_counter_overflow(counter, nmi, regs, 0))
+ if (!intel_pmu_save_and_restart(counter))
+ continue;
+
+ if (perf_counter_overflow(counter, 1, &data))
intel_pmu_disable_counter(&counter->hw, bit);
}
return 1;
}
-static int amd_pmu_handle_irq(struct pt_regs *regs, int nmi)
+static int amd_pmu_handle_irq(struct pt_regs *regs)
{
- int cpu, idx, handled = 0;
+ struct perf_sample_data data;
struct cpu_hw_counters *cpuc;
struct perf_counter *counter;
struct hw_perf_counter *hwc;
+ int cpu, idx, handled = 0;
u64 val;
+ data.regs = regs;
+ data.addr = 0;
+
cpu = smp_processor_id();
cpuc = &per_cpu(cpu_hw_counters, cpu);
if (val & (1ULL << (x86_pmu.counter_bits - 1)))
continue;
- /* counter overflow */
- x86_perf_counter_set_period(counter, hwc, idx);
- handled = 1;
- inc_irq_stat(apic_perf_irqs);
- if (perf_counter_overflow(counter, nmi, regs, 0))
+ /*
+ * counter overflow
+ */
+ handled = 1;
+ data.period = counter->hw.last_period;
+
+ if (!x86_perf_counter_set_period(counter, hwc, idx))
+ continue;
+
+ if (perf_counter_overflow(counter, 1, &data))
amd_pmu_disable_counter(hwc, idx);
}
- return handled;
-}
+ if (handled)
+ inc_irq_stat(apic_perf_irqs);
-void smp_perf_counter_interrupt(struct pt_regs *regs)
-{
- irq_enter();
- apic_write(APIC_LVTPC, LOCAL_PERF_VECTOR);
- ack_APIC_irq();
- x86_pmu.handle_irq(regs, 0);
- irq_exit();
+ return handled;
}
void smp_perf_pending_interrupt(struct pt_regs *regs)
* If the first NMI handles both, the latter will be empty and daze
* the CPU.
*/
- x86_pmu.handle_irq(regs, 1);
+ x86_pmu.handle_irq(regs);
return NOTIFY_STOP;
}
if (version < 2)
return -ENODEV;
- x86_pmu = intel_pmu;
- x86_pmu.version = version;
- x86_pmu.num_counters = eax.split.num_counters;
+ x86_pmu = intel_pmu;
+ x86_pmu.version = version;
+ x86_pmu.num_counters = eax.split.num_counters;
+ x86_pmu.counter_bits = eax.split.bit_width;
+ x86_pmu.counter_mask = (1ULL << eax.split.bit_width) - 1;
/*
* Quirk: v2 perfmon does not report fixed-purpose counters, so
* assume at least 3 counters:
*/
- x86_pmu.num_counters_fixed = max((int)edx.split.num_counters_fixed, 3);
-
- x86_pmu.counter_bits = eax.split.bit_width;
- x86_pmu.counter_mask = (1ULL << eax.split.bit_width) - 1;
+ x86_pmu.num_counters_fixed = max((int)edx.split.num_counters_fixed, 3);
rdmsrl(MSR_CORE_PERF_GLOBAL_CTRL, x86_pmu.intel_ctrl);
+ /*
+ * Install the hw-cache-events table:
+ */
+ switch (boot_cpu_data.x86_model) {
+ case 15: /* original 65 nm celeron/pentium/core2/xeon, "Merom"/"Conroe" */
+ case 22: /* single-core 65 nm celeron/core2solo "Merom-L"/"Conroe-L" */
+ case 23: /* current 45 nm celeron/core2/xeon "Penryn"/"Wolfdale" */
+ case 29: /* six-core 45 nm xeon "Dunnington" */
+ memcpy(hw_cache_event_ids, core2_hw_cache_event_ids,
+ sizeof(hw_cache_event_ids));
+
+ pr_cont("Core2 events, ");
+ break;
+ default:
+ case 26:
+ memcpy(hw_cache_event_ids, nehalem_hw_cache_event_ids,
+ sizeof(hw_cache_event_ids));
+
+ pr_cont("Nehalem/Corei7 events, ");
+ break;
+ case 28:
+ memcpy(hw_cache_event_ids, atom_hw_cache_event_ids,
+ sizeof(hw_cache_event_ids));
+
+ pr_cont("Atom events, ");
+ break;
+ }
return 0;
}
static int amd_pmu_init(void)
{
+ /* Performance-monitoring supported from K7 and later: */
+ if (boot_cpu_data.x86 < 6)
+ return -ENODEV;
+
x86_pmu = amd_pmu;
+
+ switch (boot_cpu_data.x86) {
+ case 0x0f:
+ case 0x10:
+ case 0x11:
+ memcpy(hw_cache_event_ids, amd_0f_hw_cache_event_ids,
+ sizeof(hw_cache_event_ids));
+
+ pr_cont("AMD Family 0f/10/11 events, ");
+ break;
+ }
return 0;
}
{
int err;
+ pr_info("Performance Counters: ");
+
switch (boot_cpu_data.x86_vendor) {
case X86_VENDOR_INTEL:
err = intel_pmu_init();
default:
return;
}
- if (err != 0)
+ if (err != 0) {
+ pr_cont("no PMU driver, software counters only.\n");
return;
+ }
- pr_info("%s Performance Monitoring support detected.\n", x86_pmu.name);
- pr_info("... version: %d\n", x86_pmu.version);
- pr_info("... bit width: %d\n", x86_pmu.counter_bits);
+ pr_cont("%s PMU driver.\n", x86_pmu.name);
- pr_info("... num counters: %d\n", x86_pmu.num_counters);
if (x86_pmu.num_counters > X86_PMC_MAX_GENERIC) {
x86_pmu.num_counters = X86_PMC_MAX_GENERIC;
WARN(1, KERN_ERR "hw perf counters %d > max(%d), clipping!",
perf_counter_mask = (1 << x86_pmu.num_counters) - 1;
perf_max_counters = x86_pmu.num_counters;
- pr_info("... value mask: %016Lx\n", x86_pmu.counter_mask);
- pr_info("... max period: %016Lx\n", x86_pmu.max_period);
-
if (x86_pmu.num_counters_fixed > X86_PMC_MAX_FIXED) {
x86_pmu.num_counters_fixed = X86_PMC_MAX_FIXED;
WARN(1, KERN_ERR "hw perf counters fixed %d > max(%d), clipping!",
x86_pmu.num_counters_fixed, X86_PMC_MAX_FIXED);
}
- pr_info("... fixed counters: %d\n", x86_pmu.num_counters_fixed);
perf_counter_mask |=
((1LL << x86_pmu.num_counters_fixed)-1) << X86_PMC_IDX_FIXED;
- pr_info("... counter mask: %016Lx\n", perf_counter_mask);
-
perf_counters_lapic_init();
register_die_notifier(&perf_counter_nmi_notifier);
+
+ pr_info("... version: %d\n", x86_pmu.version);
+ pr_info("... bit width: %d\n", x86_pmu.counter_bits);
+ pr_info("... generic counters: %d\n", x86_pmu.num_counters);
+ pr_info("... value mask: %016Lx\n", x86_pmu.counter_mask);
+ pr_info("... max period: %016Lx\n", x86_pmu.max_period);
+ pr_info("... fixed-purpose counters: %d\n", x86_pmu.num_counters_fixed);
+ pr_info("... counter mask: %016Lx\n", perf_counter_mask);
}
static inline void x86_pmu_read(struct perf_counter *counter)