Merge branch 'perfcounters/urgent' into perfcounters/core
Conflicts:
kernel/perf_counter.c
Merge reason: update to latest upstream (-rc6) and resolve
the conflict with urgent fixes.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
diff --git a/arch/x86/include/asm/perf_counter.h b/arch/x86/include/asm/perf_counter.h
index fa64e40..e7b7c93 100644
--- a/arch/x86/include/asm/perf_counter.h
+++ b/arch/x86/include/asm/perf_counter.h
@@ -84,6 +84,16 @@
#define MSR_ARCH_PERFMON_FIXED_CTR2 0x30b
#define X86_PMC_IDX_FIXED_BUS_CYCLES (X86_PMC_IDX_FIXED + 2)
+/*
+ * We model BTS tracing as another fixed-mode PMC.
+ *
+ * We choose a value in the middle of the fixed counter range, since lower
+ * values are used by actual fixed counters and higher values are used
+ * to indicate other overflow conditions in the PERF_GLOBAL_STATUS msr.
+ */
+#define X86_PMC_IDX_FIXED_BTS (X86_PMC_IDX_FIXED + 16)
+
+
#ifdef CONFIG_PERF_COUNTERS
extern void init_hw_perf_counters(void);
extern void perf_counters_lapic_init(void);
diff --git a/arch/x86/kernel/cpu/perf_counter.c b/arch/x86/kernel/cpu/perf_counter.c
index 900332b..396e35d 100644
--- a/arch/x86/kernel/cpu/perf_counter.c
+++ b/arch/x86/kernel/cpu/perf_counter.c
@@ -6,6 +6,7 @@
* Copyright (C) 2009 Jaswinder Singh Rajput
* Copyright (C) 2009 Advanced Micro Devices, Inc., Robert Richter
* Copyright (C) 2008-2009 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com>
+ * Copyright (C) 2009 Intel Corporation, <markus.t.metzger@intel.com>
*
* For licencing details see kernel-base/COPYING
*/
@@ -20,6 +21,7 @@
#include <linux/sched.h>
#include <linux/uaccess.h>
#include <linux/highmem.h>
+#include <linux/cpu.h>
#include <asm/apic.h>
#include <asm/stacktrace.h>
@@ -27,12 +29,52 @@
static u64 perf_counter_mask __read_mostly;
+/* The maximal number of PEBS counters: */
+#define MAX_PEBS_COUNTERS 4
+
+/* The size of a BTS record in bytes: */
+#define BTS_RECORD_SIZE 24
+
+/* The size of a per-cpu BTS buffer in bytes: */
+#define BTS_BUFFER_SIZE (BTS_RECORD_SIZE * 1024)
+
+/* The BTS overflow threshold in bytes from the end of the buffer: */
+#define BTS_OVFL_TH (BTS_RECORD_SIZE * 64)
+
+
+/*
+ * Bits in the debugctlmsr controlling branch tracing.
+ */
+#define X86_DEBUGCTL_TR (1 << 6)
+#define X86_DEBUGCTL_BTS (1 << 7)
+#define X86_DEBUGCTL_BTINT (1 << 8)
+#define X86_DEBUGCTL_BTS_OFF_OS (1 << 9)
+#define X86_DEBUGCTL_BTS_OFF_USR (1 << 10)
+
+/*
+ * A debug store configuration.
+ *
+ * We only support architectures that use 64bit fields.
+ */
+struct debug_store {
+ u64 bts_buffer_base;
+ u64 bts_index;
+ u64 bts_absolute_maximum;
+ u64 bts_interrupt_threshold;
+ u64 pebs_buffer_base;
+ u64 pebs_index;
+ u64 pebs_absolute_maximum;
+ u64 pebs_interrupt_threshold;
+ u64 pebs_counter_reset[MAX_PEBS_COUNTERS];
+};
+
struct cpu_hw_counters {
struct perf_counter *counters[X86_PMC_IDX_MAX];
unsigned long used_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)];
unsigned long active_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)];
unsigned long interrupts;
int enabled;
+ struct debug_store *ds;
};
/*
@@ -58,6 +100,8 @@
int apic;
u64 max_period;
u64 intel_ctrl;
+ void (*enable_bts)(u64 config);
+ void (*disable_bts)(void);
};
static struct x86_pmu x86_pmu __read_mostly;
@@ -577,6 +621,9 @@
u64 prev_raw_count, new_raw_count;
s64 delta;
+ if (idx == X86_PMC_IDX_FIXED_BTS)
+ return 0;
+
/*
* Careful: an NMI might modify the previous counter value.
*
@@ -666,10 +713,109 @@
#endif
}
+static inline bool bts_available(void)
+{
+ return x86_pmu.enable_bts != NULL;
+}
+
+static inline void init_debug_store_on_cpu(int cpu)
+{
+ struct debug_store *ds = per_cpu(cpu_hw_counters, cpu).ds;
+
+ if (!ds)
+ return;
+
+ wrmsr_on_cpu(cpu, MSR_IA32_DS_AREA,
+ (u32)((u64)(long)ds), (u32)((u64)(long)ds >> 32));
+}
+
+static inline void fini_debug_store_on_cpu(int cpu)
+{
+ if (!per_cpu(cpu_hw_counters, cpu).ds)
+ return;
+
+ wrmsr_on_cpu(cpu, MSR_IA32_DS_AREA, 0, 0);
+}
+
+static void release_bts_hardware(void)
+{
+ int cpu;
+
+ if (!bts_available())
+ return;
+
+ get_online_cpus();
+
+ for_each_online_cpu(cpu)
+ fini_debug_store_on_cpu(cpu);
+
+ for_each_possible_cpu(cpu) {
+ struct debug_store *ds = per_cpu(cpu_hw_counters, cpu).ds;
+
+ if (!ds)
+ continue;
+
+ per_cpu(cpu_hw_counters, cpu).ds = NULL;
+
+ kfree((void *)(long)ds->bts_buffer_base);
+ kfree(ds);
+ }
+
+ put_online_cpus();
+}
+
+static int reserve_bts_hardware(void)
+{
+ int cpu, err = 0;
+
+ if (!bts_available())
+ return -EOPNOTSUPP;
+
+ get_online_cpus();
+
+ for_each_possible_cpu(cpu) {
+ struct debug_store *ds;
+ void *buffer;
+
+ err = -ENOMEM;
+ buffer = kzalloc(BTS_BUFFER_SIZE, GFP_KERNEL);
+ if (unlikely(!buffer))
+ break;
+
+ ds = kzalloc(sizeof(*ds), GFP_KERNEL);
+ if (unlikely(!ds)) {
+ kfree(buffer);
+ break;
+ }
+
+ ds->bts_buffer_base = (u64)(long)buffer;
+ ds->bts_index = ds->bts_buffer_base;
+ ds->bts_absolute_maximum =
+ ds->bts_buffer_base + BTS_BUFFER_SIZE;
+ ds->bts_interrupt_threshold =
+ ds->bts_absolute_maximum - BTS_OVFL_TH;
+
+ per_cpu(cpu_hw_counters, cpu).ds = ds;
+ err = 0;
+ }
+
+ if (err)
+ release_bts_hardware();
+ else {
+ for_each_online_cpu(cpu)
+ init_debug_store_on_cpu(cpu);
+ }
+
+ put_online_cpus();
+
+ return err;
+}
+
static void hw_perf_counter_destroy(struct perf_counter *counter)
{
if (atomic_dec_and_mutex_lock(&active_counters, &pmc_reserve_mutex)) {
release_pmc_hardware();
+ release_bts_hardware();
mutex_unlock(&pmc_reserve_mutex);
}
}
@@ -712,6 +858,42 @@
return 0;
}
+static void intel_pmu_enable_bts(u64 config)
+{
+ unsigned long debugctlmsr;
+
+ debugctlmsr = get_debugctlmsr();
+
+ debugctlmsr |= X86_DEBUGCTL_TR;
+ debugctlmsr |= X86_DEBUGCTL_BTS;
+ debugctlmsr |= X86_DEBUGCTL_BTINT;
+
+ if (!(config & ARCH_PERFMON_EVENTSEL_OS))
+ debugctlmsr |= X86_DEBUGCTL_BTS_OFF_OS;
+
+ if (!(config & ARCH_PERFMON_EVENTSEL_USR))
+ debugctlmsr |= X86_DEBUGCTL_BTS_OFF_USR;
+
+ update_debugctlmsr(debugctlmsr);
+}
+
+static void intel_pmu_disable_bts(void)
+{
+ struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters);
+ unsigned long debugctlmsr;
+
+ if (!cpuc->ds)
+ return;
+
+ debugctlmsr = get_debugctlmsr();
+
+ debugctlmsr &=
+ ~(X86_DEBUGCTL_TR | X86_DEBUGCTL_BTS | X86_DEBUGCTL_BTINT |
+ X86_DEBUGCTL_BTS_OFF_OS | X86_DEBUGCTL_BTS_OFF_USR);
+
+ update_debugctlmsr(debugctlmsr);
+}
+
/*
* Setup the hardware configuration for a given attr_type
*/
@@ -728,9 +910,13 @@
err = 0;
if (!atomic_inc_not_zero(&active_counters)) {
mutex_lock(&pmc_reserve_mutex);
- if (atomic_read(&active_counters) == 0 && !reserve_pmc_hardware())
- err = -EBUSY;
- else
+ if (atomic_read(&active_counters) == 0) {
+ if (!reserve_pmc_hardware())
+ err = -EBUSY;
+ else
+ reserve_bts_hardware();
+ }
+ if (!err)
atomic_inc(&active_counters);
mutex_unlock(&pmc_reserve_mutex);
}
@@ -817,7 +1003,18 @@
static void intel_pmu_disable_all(void)
{
+ struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters);
+
+ if (!cpuc->enabled)
+ return;
+
+ cpuc->enabled = 0;
+ barrier();
+
wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, 0);
+
+ if (test_bit(X86_PMC_IDX_FIXED_BTS, cpuc->active_mask))
+ intel_pmu_disable_bts();
}
static void amd_pmu_disable_all(void)
@@ -875,7 +1072,25 @@
static void intel_pmu_enable_all(void)
{
+ struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters);
+
+ if (cpuc->enabled)
+ return;
+
+ cpuc->enabled = 1;
+ barrier();
+
wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, x86_pmu.intel_ctrl);
+
+ if (test_bit(X86_PMC_IDX_FIXED_BTS, cpuc->active_mask)) {
+ struct perf_counter *counter =
+ cpuc->counters[X86_PMC_IDX_FIXED_BTS];
+
+ if (WARN_ON_ONCE(!counter))
+ return;
+
+ intel_pmu_enable_bts(counter->hw.config);
+ }
}
static void amd_pmu_enable_all(void)
@@ -962,6 +1177,11 @@
static inline void
intel_pmu_disable_counter(struct hw_perf_counter *hwc, int idx)
{
+ if (unlikely(idx == X86_PMC_IDX_FIXED_BTS)) {
+ intel_pmu_disable_bts();
+ return;
+ }
+
if (unlikely(hwc->config_base == MSR_ARCH_PERFMON_FIXED_CTR_CTRL)) {
intel_pmu_disable_fixed(hwc, idx);
return;
@@ -990,6 +1210,9 @@
s64 period = hwc->sample_period;
int err, ret = 0;
+ if (idx == X86_PMC_IDX_FIXED_BTS)
+ return 0;
+
/*
* If we are way outside a reasoable range then just skip forward:
*/
@@ -1072,6 +1295,14 @@
static void intel_pmu_enable_counter(struct hw_perf_counter *hwc, int idx)
{
+ if (unlikely(idx == X86_PMC_IDX_FIXED_BTS)) {
+ if (!__get_cpu_var(cpu_hw_counters).enabled)
+ return;
+
+ intel_pmu_enable_bts(hwc->config);
+ return;
+ }
+
if (unlikely(hwc->config_base == MSR_ARCH_PERFMON_FIXED_CTR_CTRL)) {
intel_pmu_enable_fixed(hwc, idx);
return;
@@ -1093,11 +1324,16 @@
{
unsigned int event;
+ event = hwc->config & ARCH_PERFMON_EVENT_MASK;
+
+ if (unlikely((event ==
+ x86_pmu.event_map(PERF_COUNT_HW_BRANCH_INSTRUCTIONS)) &&
+ (hwc->sample_period == 1)))
+ return X86_PMC_IDX_FIXED_BTS;
+
if (!x86_pmu.num_counters_fixed)
return -1;
- event = hwc->config & ARCH_PERFMON_EVENT_MASK;
-
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_HW_CPU_CYCLES)))
@@ -1118,7 +1354,25 @@
int idx;
idx = fixed_mode_idx(counter, hwc);
- if (idx >= 0) {
+ if (idx == X86_PMC_IDX_FIXED_BTS) {
+ /*
+ * Try to use BTS for branch tracing. If that is not
+ * available, try to get a generic counter.
+ */
+ if (unlikely(!cpuc->ds))
+ goto try_generic;
+
+ /*
+ * Try to get the fixed counter, if that is already taken
+ * then try to get a generic counter:
+ */
+ if (test_and_set_bit(idx, cpuc->used_mask))
+ goto try_generic;
+
+ hwc->config_base = 0;
+ hwc->counter_base = 0;
+ hwc->idx = idx;
+ } else if (idx >= 0) {
/*
* Try to get the fixed counter, if that is already taken
* then try to get a generic counter:
@@ -1229,6 +1483,45 @@
local_irq_restore(flags);
}
+static void intel_pmu_drain_bts_buffer(struct cpu_hw_counters *cpuc,
+ struct perf_sample_data *data)
+{
+ struct debug_store *ds = cpuc->ds;
+ struct bts_record {
+ u64 from;
+ u64 to;
+ u64 flags;
+ };
+ struct perf_counter *counter = cpuc->counters[X86_PMC_IDX_FIXED_BTS];
+ unsigned long orig_ip = data->regs->ip;
+ u64 at;
+
+ if (!counter)
+ return;
+
+ if (!ds)
+ return;
+
+ for (at = ds->bts_buffer_base;
+ at < ds->bts_index;
+ at += sizeof(struct bts_record)) {
+ struct bts_record *rec = (struct bts_record *)(long)at;
+
+ data->regs->ip = rec->from;
+ data->addr = rec->to;
+
+ perf_counter_output(counter, 1, data);
+ }
+
+ ds->bts_index = ds->bts_buffer_base;
+
+ data->regs->ip = orig_ip;
+ data->addr = 0;
+
+ /* There's new data available. */
+ counter->pending_kill = POLL_IN;
+}
+
static void x86_pmu_disable(struct perf_counter *counter)
{
struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters);
@@ -1253,6 +1546,15 @@
* that we are disabling:
*/
x86_perf_counter_update(counter, hwc, idx);
+
+ /* Drain the remaining BTS records. */
+ if (unlikely(idx == X86_PMC_IDX_FIXED_BTS)) {
+ struct perf_sample_data data;
+ struct pt_regs regs;
+
+ data.regs = ®s;
+ intel_pmu_drain_bts_buffer(cpuc, &data);
+ }
cpuc->counters[idx] = NULL;
clear_bit(idx, cpuc->used_mask);
@@ -1280,6 +1582,7 @@
static void intel_pmu_reset(void)
{
+ struct debug_store *ds = __get_cpu_var(cpu_hw_counters).ds;
unsigned long flags;
int idx;
@@ -1297,6 +1600,8 @@
for (idx = 0; idx < x86_pmu.num_counters_fixed; idx++) {
checking_wrmsrl(MSR_ARCH_PERFMON_FIXED_CTR0 + idx, 0ull);
}
+ if (ds)
+ ds->bts_index = ds->bts_buffer_base;
local_irq_restore(flags);
}
@@ -1362,6 +1667,7 @@
cpuc = &__get_cpu_var(cpu_hw_counters);
perf_disable();
+ intel_pmu_drain_bts_buffer(cpuc, &data);
status = intel_pmu_get_status();
if (!status) {
perf_enable();
@@ -1571,6 +1877,8 @@
* the generic counter period:
*/
.max_period = (1ULL << 31) - 1,
+ .enable_bts = intel_pmu_enable_bts,
+ .disable_bts = intel_pmu_disable_bts,
};
static struct x86_pmu amd_pmu = {
@@ -1962,3 +2270,8 @@
return entry;
}
+
+void hw_perf_counter_setup_online(int cpu)
+{
+ init_debug_store_on_cpu(cpu);
+}
diff --git a/tools/perf/Documentation/perf-report.txt b/tools/perf/Documentation/perf-report.txt
index e72e931..370344a 100644
--- a/tools/perf/Documentation/perf-report.txt
+++ b/tools/perf/Documentation/perf-report.txt
@@ -27,6 +27,9 @@
-n
--show-nr-samples
Show the number of samples for each symbol
+-T
+--threads
+ Show per-thread event counters
-C::
--comms=::
Only consider symbols in these comms. CSV that understands
diff --git a/tools/perf/Makefile b/tools/perf/Makefile
index c045b42..68218cf 100644
--- a/tools/perf/Makefile
+++ b/tools/perf/Makefile
@@ -310,6 +310,7 @@
LIB_H += util/symbol.h
LIB_H += util/module.h
LIB_H += util/color.h
+LIB_H += util/values.h
LIB_OBJS += util/abspath.o
LIB_OBJS += util/alias.o
@@ -337,6 +338,9 @@
LIB_OBJS += util/pager.o
LIB_OBJS += util/header.o
LIB_OBJS += util/callchain.o
+LIB_OBJS += util/values.o
+LIB_OBJS += util/debug.o
+LIB_OBJS += util/map.o
BUILTIN_OBJS += builtin-annotate.o
BUILTIN_OBJS += builtin-help.o
diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c
index 1dba568..543c452 100644
--- a/tools/perf/builtin-annotate.c
+++ b/tools/perf/builtin-annotate.c
@@ -26,7 +26,6 @@
#define SHOW_HV 4
static char const *input_name = "perf.data";
-static char *vmlinux = "vmlinux";
static char default_sort_order[] = "comm,symbol";
static char *sort_order = default_sort_order;
@@ -37,9 +36,6 @@
static int dump_trace = 0;
#define dprintf(x...) do { if (dump_trace) printf(x); } while (0)
-static int verbose;
-
-static int modules;
static int full_paths;
@@ -48,40 +44,6 @@
static unsigned long page_size;
static unsigned long mmap_window = 32;
-struct ip_event {
- struct perf_event_header header;
- u64 ip;
- u32 pid, tid;
-};
-
-struct mmap_event {
- struct perf_event_header header;
- u32 pid, tid;
- u64 start;
- u64 len;
- u64 pgoff;
- char filename[PATH_MAX];
-};
-
-struct comm_event {
- struct perf_event_header header;
- u32 pid, tid;
- char comm[16];
-};
-
-struct fork_event {
- struct perf_event_header header;
- u32 pid, ppid;
-};
-
-typedef union event_union {
- struct perf_event_header header;
- struct ip_event ip;
- struct mmap_event mmap;
- struct comm_event comm;
- struct fork_event fork;
-} event_t;
-
struct sym_ext {
struct rb_node node;
@@ -89,175 +51,6 @@
char *path;
};
-static LIST_HEAD(dsos);
-static struct dso *kernel_dso;
-static struct dso *vdso;
-
-
-static void dsos__add(struct dso *dso)
-{
- list_add_tail(&dso->node, &dsos);
-}
-
-static struct dso *dsos__find(const char *name)
-{
- struct dso *pos;
-
- list_for_each_entry(pos, &dsos, node)
- if (strcmp(pos->name, name) == 0)
- return pos;
- return NULL;
-}
-
-static struct dso *dsos__findnew(const char *name)
-{
- struct dso *dso = dsos__find(name);
- int nr;
-
- if (dso)
- return dso;
-
- dso = dso__new(name, 0);
- if (!dso)
- goto out_delete_dso;
-
- nr = dso__load(dso, NULL, verbose);
- if (nr < 0) {
- if (verbose)
- fprintf(stderr, "Failed to open: %s\n", name);
- goto out_delete_dso;
- }
- if (!nr && verbose) {
- fprintf(stderr,
- "No symbols found in: %s, maybe install a debug package?\n",
- name);
- }
-
- dsos__add(dso);
-
- return dso;
-
-out_delete_dso:
- dso__delete(dso);
- return NULL;
-}
-
-static void dsos__fprintf(FILE *fp)
-{
- struct dso *pos;
-
- list_for_each_entry(pos, &dsos, node)
- dso__fprintf(pos, fp);
-}
-
-static struct symbol *vdso__find_symbol(struct dso *dso, u64 ip)
-{
- return dso__find_symbol(dso, ip);
-}
-
-static int load_kernel(void)
-{
- int err;
-
- kernel_dso = dso__new("[kernel]", 0);
- if (!kernel_dso)
- return -1;
-
- err = dso__load_kernel(kernel_dso, vmlinux, NULL, verbose, modules);
- if (err <= 0) {
- dso__delete(kernel_dso);
- kernel_dso = NULL;
- } else
- dsos__add(kernel_dso);
-
- vdso = dso__new("[vdso]", 0);
- if (!vdso)
- return -1;
-
- vdso->find_symbol = vdso__find_symbol;
-
- dsos__add(vdso);
-
- return err;
-}
-
-struct map {
- struct list_head node;
- u64 start;
- u64 end;
- u64 pgoff;
- u64 (*map_ip)(struct map *, u64);
- struct dso *dso;
-};
-
-static u64 map__map_ip(struct map *map, u64 ip)
-{
- return ip - map->start + map->pgoff;
-}
-
-static u64 vdso__map_ip(struct map *map __used, u64 ip)
-{
- return ip;
-}
-
-static struct map *map__new(struct mmap_event *event)
-{
- struct map *self = malloc(sizeof(*self));
-
- if (self != NULL) {
- const char *filename = event->filename;
-
- self->start = event->start;
- self->end = event->start + event->len;
- self->pgoff = event->pgoff;
-
- self->dso = dsos__findnew(filename);
- if (self->dso == NULL)
- goto out_delete;
-
- if (self->dso == vdso)
- self->map_ip = vdso__map_ip;
- else
- self->map_ip = map__map_ip;
- }
- return self;
-out_delete:
- free(self);
- return NULL;
-}
-
-static struct map *map__clone(struct map *self)
-{
- struct map *map = malloc(sizeof(*self));
-
- if (!map)
- return NULL;
-
- memcpy(map, self, sizeof(*self));
-
- return map;
-}
-
-static int map__overlap(struct map *l, struct map *r)
-{
- if (l->start > r->start) {
- struct map *t = l;
- l = r;
- r = t;
- }
-
- if (l->end > r->start)
- return 1;
-
- return 0;
-}
-
-static size_t map__fprintf(struct map *self, FILE *fp)
-{
- return fprintf(fp, " %Lx-%Lx %Lx %s\n",
- self->start, self->end, self->pgoff, self->dso->name);
-}
-
struct thread {
struct rb_node rb_node;
@@ -927,7 +720,7 @@
process_mmap_event(event_t *event, unsigned long offset, unsigned long head)
{
struct thread *thread = threads__findnew(event->mmap.pid);
- struct map *map = map__new(&event->mmap);
+ struct map *map = map__new(&event->mmap, NULL, 0);
dprintf("%p [%p]: PERF_EVENT_MMAP %d: [%p(%p) @ %p]: %s\n",
(void *)(offset + head),
diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index 89a5ddc..65b4115 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -15,6 +15,7 @@
#include "util/string.h"
#include "util/header.h"
+#include "util/event.h"
#include <unistd.h>
#include <sched.h>
@@ -42,7 +43,6 @@
static int force = 0;
static int append_file = 0;
static int call_graph = 0;
-static int verbose = 0;
static int inherit_stat = 0;
static int no_samples = 0;
static int sample_address = 0;
@@ -62,24 +62,6 @@
struct perf_header *header;
-struct mmap_event {
- struct perf_event_header header;
- u32 pid;
- u32 tid;
- u64 start;
- u64 len;
- u64 pgoff;
- char filename[PATH_MAX];
-};
-
-struct comm_event {
- struct perf_event_header header;
- u32 pid;
- u32 tid;
- char comm[16];
-};
-
-
struct mmap_data {
int counter;
void *base;
diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index b53a60f..6321951 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -17,6 +17,7 @@
#include "util/string.h"
#include "util/callchain.h"
#include "util/strlist.h"
+#include "util/values.h"
#include "perf.h"
#include "util/header.h"
@@ -29,7 +30,6 @@
#define SHOW_HV 4
static char const *input_name = "perf.data";
-static char *vmlinux = NULL;
static char default_sort_order[] = "comm,dso,symbol";
static char *sort_order = default_sort_order;
@@ -45,14 +45,15 @@
#define dprintf(x...) do { if (dump_trace) printf(x); } while (0)
#define cdprintf(x...) do { if (dump_trace) color_fprintf(stdout, color, x); } while (0)
-static int verbose;
-#define eprintf(x...) do { if (verbose) fprintf(stderr, x); } while (0)
-
-static int modules;
-
static int full_paths;
static int show_nr_samples;
+static int show_threads;
+static struct perf_read_values show_threads_values;
+
+static char default_pretty_printing_style[] = "normal";
+static char *pretty_printing_style = default_pretty_printing_style;
+
static unsigned long page_size;
static unsigned long mmap_window = 32;
@@ -66,6 +67,10 @@
static int callchain;
+static char __cwd[PATH_MAX];
+static char *cwd = __cwd;
+static int cwdlen;
+
static
struct callchain_param callchain_param = {
.mode = CHAIN_GRAPH_REL,
@@ -74,59 +79,6 @@
static u64 sample_type;
-struct ip_event {
- struct perf_event_header header;
- u64 ip;
- u32 pid, tid;
- unsigned char __more_data[];
-};
-
-struct mmap_event {
- struct perf_event_header header;
- u32 pid, tid;
- u64 start;
- u64 len;
- u64 pgoff;
- char filename[PATH_MAX];
-};
-
-struct comm_event {
- struct perf_event_header header;
- u32 pid, tid;
- char comm[16];
-};
-
-struct fork_event {
- struct perf_event_header header;
- u32 pid, ppid;
- u32 tid, ptid;
-};
-
-struct lost_event {
- struct perf_event_header header;
- u64 id;
- u64 lost;
-};
-
-struct read_event {
- struct perf_event_header header;
- u32 pid,tid;
- u64 value;
- u64 time_enabled;
- u64 time_running;
- u64 id;
-};
-
-typedef union event_union {
- struct perf_event_header header;
- struct ip_event ip;
- struct mmap_event mmap;
- struct comm_event comm;
- struct fork_event fork;
- struct lost_event lost;
- struct read_event read;
-} event_t;
-
static int repsep_fprintf(FILE *fp, const char *fmt, ...)
{
int n;
@@ -154,215 +106,6 @@
return n;
}
-static LIST_HEAD(dsos);
-static struct dso *kernel_dso;
-static struct dso *vdso;
-static struct dso *hypervisor_dso;
-
-static void dsos__add(struct dso *dso)
-{
- list_add_tail(&dso->node, &dsos);
-}
-
-static struct dso *dsos__find(const char *name)
-{
- struct dso *pos;
-
- list_for_each_entry(pos, &dsos, node)
- if (strcmp(pos->name, name) == 0)
- return pos;
- return NULL;
-}
-
-static struct dso *dsos__findnew(const char *name)
-{
- struct dso *dso = dsos__find(name);
- int nr;
-
- if (dso)
- return dso;
-
- dso = dso__new(name, 0);
- if (!dso)
- goto out_delete_dso;
-
- nr = dso__load(dso, NULL, verbose);
- if (nr < 0) {
- eprintf("Failed to open: %s\n", name);
- goto out_delete_dso;
- }
- if (!nr)
- eprintf("No symbols found in: %s, maybe install a debug package?\n", name);
-
- dsos__add(dso);
-
- return dso;
-
-out_delete_dso:
- dso__delete(dso);
- return NULL;
-}
-
-static void dsos__fprintf(FILE *fp)
-{
- struct dso *pos;
-
- list_for_each_entry(pos, &dsos, node)
- dso__fprintf(pos, fp);
-}
-
-static struct symbol *vdso__find_symbol(struct dso *dso, u64 ip)
-{
- return dso__find_symbol(dso, ip);
-}
-
-static int load_kernel(void)
-{
- int err;
-
- kernel_dso = dso__new("[kernel]", 0);
- if (!kernel_dso)
- return -1;
-
- err = dso__load_kernel(kernel_dso, vmlinux, NULL, verbose, modules);
- if (err <= 0) {
- dso__delete(kernel_dso);
- kernel_dso = NULL;
- } else
- dsos__add(kernel_dso);
-
- vdso = dso__new("[vdso]", 0);
- if (!vdso)
- return -1;
-
- vdso->find_symbol = vdso__find_symbol;
-
- dsos__add(vdso);
-
- hypervisor_dso = dso__new("[hypervisor]", 0);
- if (!hypervisor_dso)
- return -1;
- dsos__add(hypervisor_dso);
-
- return err;
-}
-
-static char __cwd[PATH_MAX];
-static char *cwd = __cwd;
-static int cwdlen;
-
-static int strcommon(const char *pathname)
-{
- int n = 0;
-
- while (n < cwdlen && pathname[n] == cwd[n])
- ++n;
-
- return n;
-}
-
-struct map {
- struct list_head node;
- u64 start;
- u64 end;
- u64 pgoff;
- u64 (*map_ip)(struct map *, u64);
- struct dso *dso;
-};
-
-static u64 map__map_ip(struct map *map, u64 ip)
-{
- return ip - map->start + map->pgoff;
-}
-
-static u64 vdso__map_ip(struct map *map __used, u64 ip)
-{
- return ip;
-}
-
-static inline int is_anon_memory(const char *filename)
-{
- return strcmp(filename, "//anon") == 0;
-}
-
-static struct map *map__new(struct mmap_event *event)
-{
- struct map *self = malloc(sizeof(*self));
-
- if (self != NULL) {
- const char *filename = event->filename;
- char newfilename[PATH_MAX];
- int anon;
-
- if (cwd) {
- int n = strcommon(filename);
-
- if (n == cwdlen) {
- snprintf(newfilename, sizeof(newfilename),
- ".%s", filename + n);
- filename = newfilename;
- }
- }
-
- anon = is_anon_memory(filename);
-
- if (anon) {
- snprintf(newfilename, sizeof(newfilename), "/tmp/perf-%d.map", event->pid);
- filename = newfilename;
- }
-
- self->start = event->start;
- self->end = event->start + event->len;
- self->pgoff = event->pgoff;
-
- self->dso = dsos__findnew(filename);
- if (self->dso == NULL)
- goto out_delete;
-
- if (self->dso == vdso || anon)
- self->map_ip = vdso__map_ip;
- else
- self->map_ip = map__map_ip;
- }
- return self;
-out_delete:
- free(self);
- return NULL;
-}
-
-static struct map *map__clone(struct map *self)
-{
- struct map *map = malloc(sizeof(*self));
-
- if (!map)
- return NULL;
-
- memcpy(map, self, sizeof(*self));
-
- return map;
-}
-
-static int map__overlap(struct map *l, struct map *r)
-{
- if (l->start > r->start) {
- struct map *t = l;
- l = r;
- r = t;
- }
-
- if (l->end > r->start)
- return 1;
-
- return 0;
-}
-
-static size_t map__fprintf(struct map *self, FILE *fp)
-{
- return fprintf(fp, " %Lx-%Lx %Lx %s\n",
- self->start, self->end, self->pgoff, self->dso->name);
-}
-
-
struct thread {
struct rb_node rb_node;
struct list_head maps;
@@ -1397,6 +1140,9 @@
size_t ret = 0;
unsigned int width;
char *col_width = col_width_list_str;
+ int raw_printing_style;
+
+ raw_printing_style = !strcmp(pretty_printing_style, "raw");
init_rem_hits();
@@ -1473,6 +1219,10 @@
free(rem_sq_bracket);
+ if (show_threads)
+ perf_read_values_display(fp, &show_threads_values,
+ raw_printing_style);
+
return ret;
}
@@ -1611,7 +1361,7 @@
process_mmap_event(event_t *event, unsigned long offset, unsigned long head)
{
struct thread *thread = threads__findnew(event->mmap.pid);
- struct map *map = map__new(&event->mmap);
+ struct map *map = map__new(&event->mmap, cwd, cwdlen);
dprintf("%p [%p]: PERF_EVENT_MMAP %d/%d: [%p(%p) @ %p]: %s\n",
(void *)(offset + head),
@@ -1760,6 +1510,16 @@
{
struct perf_counter_attr *attr = perf_header__find_attr(event->read.id);
+ if (show_threads) {
+ char *name = attr ? __event_name(attr->type, attr->config)
+ : "unknown";
+ perf_read_values_add_value(&show_threads_values,
+ event->read.pid, event->read.tid,
+ event->read.id,
+ name,
+ event->read.value);
+ }
+
dprintf("%p [%p]: PERF_EVENT_READ: %d %d %s %Lu\n",
(void *)(offset + head),
(void *)(long)(event->header.size),
@@ -1841,6 +1601,9 @@
register_idle_thread();
+ if (show_threads)
+ perf_read_values_init(&show_threads_values);
+
input = open(input_name, O_RDONLY);
if (input < 0) {
fprintf(stderr, " failed to open file: %s", input_name);
@@ -1995,6 +1758,9 @@
output__resort(total);
output__fprintf(stdout, total);
+ if (show_threads)
+ perf_read_values_destroy(&show_threads_values);
+
return rc;
}
@@ -2068,6 +1834,10 @@
"load module symbols - WARNING: use only with -k and LIVE kernel"),
OPT_BOOLEAN('n', "show-nr-samples", &show_nr_samples,
"Show a column with the number of samples"),
+ OPT_BOOLEAN('T', "threads", &show_threads,
+ "Show per-thread event counters"),
+ OPT_STRING(0, "pretty", &pretty_printing_style, "key",
+ "pretty printing style key: normal raw"),
OPT_STRING('s', "sort", &sort_order, "key[,key2...]",
"sort by key(s): pid, comm, dso, symbol, parent"),
OPT_BOOLEAN('P', "full-paths", &full_paths,
diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index b4b06c7..4b9dd4a 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -63,7 +63,6 @@
#define MAX_RUN 100
static int system_wide = 0;
-static int verbose = 0;
static unsigned int nr_cpus = 0;
static int run_idx = 0;
diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index 7de28ce..9a6dbbf 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -68,8 +68,6 @@
static unsigned int page_size;
static unsigned int mmap_pages = 16;
static int freq = 0;
-static int verbose = 0;
-static char *vmlinux = NULL;
static int delay_secs = 2;
static int zero;
@@ -338,8 +336,6 @@
printf("%d lines not displayed, maybe increase display entries [e]\n", more);
}
-struct dso *kernel_dso;
-
/*
* Symbols will be added here in record_ip and will get out
* after decayed.
@@ -937,26 +933,6 @@
last_read = this_read;
for (; old != head;) {
- struct ip_event {
- struct perf_event_header header;
- u64 ip;
- u32 pid, target_pid;
- };
- struct mmap_event {
- struct perf_event_header header;
- u32 pid, target_pid;
- u64 start;
- u64 len;
- u64 pgoff;
- char filename[PATH_MAX];
- };
-
- typedef union event_union {
- struct perf_event_header header;
- struct ip_event ip;
- struct mmap_event mmap;
- } event_t;
-
event_t *event = (event_t *)&data[old & md->mask];
event_t event_copy;
diff --git a/tools/perf/builtin.h b/tools/perf/builtin.h
index 51d1682..3a63e41 100644
--- a/tools/perf/builtin.h
+++ b/tools/perf/builtin.h
@@ -22,5 +22,6 @@
extern int cmd_top(int argc, const char **argv, const char *prefix);
extern int cmd_version(int argc, const char **argv, const char *prefix);
extern int cmd_list(int argc, const char **argv, const char *prefix);
+extern int cmd_trace(int argc, const char **argv, const char *prefix);
#endif
diff --git a/tools/perf/perf.h b/tools/perf/perf.h
index e5148e2..f550921 100644
--- a/tools/perf/perf.h
+++ b/tools/perf/perf.h
@@ -48,6 +48,7 @@
#include "../../include/linux/perf_counter.h"
#include "util/types.h"
+#include "util/debug.h"
/*
* prctl(PR_TASK_PERF_COUNTERS_DISABLE) will (cheaply) disable all
diff --git a/tools/perf/util/callchain.h b/tools/perf/util/callchain.h
index a926ae4..43cf3ea 100644
--- a/tools/perf/util/callchain.h
+++ b/tools/perf/util/callchain.h
@@ -4,6 +4,7 @@
#include "../perf.h"
#include <linux/list.h>
#include <linux/rbtree.h>
+#include "util.h"
#include "symbol.h"
enum chain_mode {
diff --git a/tools/perf/util/debug.c b/tools/perf/util/debug.c
new file mode 100644
index 0000000..7cb8464
--- /dev/null
+++ b/tools/perf/util/debug.c
@@ -0,0 +1,22 @@
+/* For general debugging purposes */
+
+#include "../perf.h"
+#include <string.h>
+#include <stdarg.h>
+#include <stdio.h>
+
+int verbose = 0;
+
+int eprintf(const char *fmt, ...)
+{
+ va_list args;
+ int ret = 0;
+
+ if (verbose) {
+ va_start(args, fmt);
+ ret = vfprintf(stderr, fmt, args);
+ va_end(args);
+ }
+
+ return ret;
+}
diff --git a/tools/perf/util/debug.h b/tools/perf/util/debug.h
new file mode 100644
index 0000000..2ae9090
--- /dev/null
+++ b/tools/perf/util/debug.h
@@ -0,0 +1,5 @@
+/* For debugging general purposes */
+
+extern int verbose;
+
+int eprintf(const char *fmt, ...) __attribute__((format(printf, 1, 2)));
diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h
new file mode 100644
index 0000000..fa7c50b
--- /dev/null
+++ b/tools/perf/util/event.h
@@ -0,0 +1,90 @@
+#ifndef __PERF_EVENT_H
+#define __PERF_EVENT_H
+#include "../perf.h"
+#include "util.h"
+#include <linux/list.h>
+
+/*
+ * PERF_SAMPLE_IP | PERF_SAMPLE_TID | *
+ */
+struct ip_event {
+ struct perf_event_header header;
+ u64 ip;
+ u32 pid, tid;
+ unsigned char __more_data[];
+};
+
+struct mmap_event {
+ struct perf_event_header header;
+ u32 pid, tid;
+ u64 start;
+ u64 len;
+ u64 pgoff;
+ char filename[PATH_MAX];
+};
+
+struct comm_event {
+ struct perf_event_header header;
+ u32 pid, tid;
+ char comm[16];
+};
+
+struct fork_event {
+ struct perf_event_header header;
+ u32 pid, ppid;
+ u32 tid, ptid;
+};
+
+struct lost_event {
+ struct perf_event_header header;
+ u64 id;
+ u64 lost;
+};
+
+/*
+ * PERF_FORMAT_ENABLED | PERF_FORMAT_RUNNING | PERF_FORMAT_ID
+ */
+struct read_event {
+ struct perf_event_header header;
+ u32 pid,tid;
+ u64 value;
+ u64 time_enabled;
+ u64 time_running;
+ u64 id;
+};
+
+typedef union event_union {
+ struct perf_event_header header;
+ struct ip_event ip;
+ struct mmap_event mmap;
+ struct comm_event comm;
+ struct fork_event fork;
+ struct lost_event lost;
+ struct read_event read;
+} event_t;
+
+struct map {
+ struct list_head node;
+ u64 start;
+ u64 end;
+ u64 pgoff;
+ u64 (*map_ip)(struct map *, u64);
+ struct dso *dso;
+};
+
+static inline u64 map__map_ip(struct map *map, u64 ip)
+{
+ return ip - map->start + map->pgoff;
+}
+
+static inline u64 vdso__map_ip(struct map *map __used, u64 ip)
+{
+ return ip;
+}
+
+struct map *map__new(struct mmap_event *event, char *cwd, int cwdlen);
+struct map *map__clone(struct map *self);
+int map__overlap(struct map *l, struct map *r);
+size_t map__fprintf(struct map *self, FILE *fp);
+
+#endif
diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c
new file mode 100644
index 0000000..804e023
--- /dev/null
+++ b/tools/perf/util/map.c
@@ -0,0 +1,97 @@
+#include "event.h"
+#include "symbol.h"
+#include <stdlib.h>
+#include <string.h>
+#include <stdio.h>
+
+static inline int is_anon_memory(const char *filename)
+{
+ return strcmp(filename, "//anon") == 0;
+}
+
+static int strcommon(const char *pathname, char *cwd, int cwdlen)
+{
+ int n = 0;
+
+ while (n < cwdlen && pathname[n] == cwd[n])
+ ++n;
+
+ return n;
+}
+
+ struct map *map__new(struct mmap_event *event, char *cwd, int cwdlen)
+{
+ struct map *self = malloc(sizeof(*self));
+
+ if (self != NULL) {
+ const char *filename = event->filename;
+ char newfilename[PATH_MAX];
+ int anon;
+
+ if (cwd) {
+ int n = strcommon(filename, cwd, cwdlen);
+
+ if (n == cwdlen) {
+ snprintf(newfilename, sizeof(newfilename),
+ ".%s", filename + n);
+ filename = newfilename;
+ }
+ }
+
+ anon = is_anon_memory(filename);
+
+ if (anon) {
+ snprintf(newfilename, sizeof(newfilename), "/tmp/perf-%d.map", event->pid);
+ filename = newfilename;
+ }
+
+ self->start = event->start;
+ self->end = event->start + event->len;
+ self->pgoff = event->pgoff;
+
+ self->dso = dsos__findnew(filename);
+ if (self->dso == NULL)
+ goto out_delete;
+
+ if (self->dso == vdso || anon)
+ self->map_ip = vdso__map_ip;
+ else
+ self->map_ip = map__map_ip;
+ }
+ return self;
+out_delete:
+ free(self);
+ return NULL;
+}
+
+struct map *map__clone(struct map *self)
+{
+ struct map *map = malloc(sizeof(*self));
+
+ if (!map)
+ return NULL;
+
+ memcpy(map, self, sizeof(*self));
+
+ return map;
+}
+
+int map__overlap(struct map *l, struct map *r)
+{
+ if (l->start > r->start) {
+ struct map *t = l;
+ l = r;
+ r = t;
+ }
+
+ if (l->end > r->start)
+ return 1;
+
+ return 0;
+}
+
+size_t map__fprintf(struct map *self, FILE *fp)
+{
+ return fprintf(fp, " %Lx-%Lx %Lx %s\n",
+ self->start, self->end, self->pgoff, self->dso->name);
+}
diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
index 5c0f42e..0b98623 100644
--- a/tools/perf/util/symbol.c
+++ b/tools/perf/util/symbol.c
@@ -924,6 +924,103 @@
return err;
}
+LIST_HEAD(dsos);
+struct dso *kernel_dso;
+struct dso *vdso;
+struct dso *hypervisor_dso;
+
+char *vmlinux = "vmlinux";
+int modules;
+
+static void dsos__add(struct dso *dso)
+{
+ list_add_tail(&dso->node, &dsos);
+}
+
+static struct dso *dsos__find(const char *name)
+{
+ struct dso *pos;
+
+ list_for_each_entry(pos, &dsos, node)
+ if (strcmp(pos->name, name) == 0)
+ return pos;
+ return NULL;
+}
+
+struct dso *dsos__findnew(const char *name)
+{
+ struct dso *dso = dsos__find(name);
+ int nr;
+
+ if (dso)
+ return dso;
+
+ dso = dso__new(name, 0);
+ if (!dso)
+ goto out_delete_dso;
+
+ nr = dso__load(dso, NULL, verbose);
+ if (nr < 0) {
+ eprintf("Failed to open: %s\n", name);
+ goto out_delete_dso;
+ }
+ if (!nr)
+ eprintf("No symbols found in: %s, maybe install a debug package?\n", name);
+
+ dsos__add(dso);
+
+ return dso;
+
+out_delete_dso:
+ dso__delete(dso);
+ return NULL;
+}
+
+void dsos__fprintf(FILE *fp)
+{
+ struct dso *pos;
+
+ list_for_each_entry(pos, &dsos, node)
+ dso__fprintf(pos, fp);
+}
+
+static struct symbol *vdso__find_symbol(struct dso *dso, u64 ip)
+{
+ return dso__find_symbol(dso, ip);
+}
+
+int load_kernel(void)
+{
+ int err;
+
+ kernel_dso = dso__new("[kernel]", 0);
+ if (!kernel_dso)
+ return -1;
+
+ err = dso__load_kernel(kernel_dso, vmlinux, NULL, verbose, modules);
+ if (err <= 0) {
+ dso__delete(kernel_dso);
+ kernel_dso = NULL;
+ } else
+ dsos__add(kernel_dso);
+
+ vdso = dso__new("[vdso]", 0);
+ if (!vdso)
+ return -1;
+
+ vdso->find_symbol = vdso__find_symbol;
+
+ dsos__add(vdso);
+
+ hypervisor_dso = dso__new("[hypervisor]", 0);
+ if (!hypervisor_dso)
+ return -1;
+ dsos__add(hypervisor_dso);
+
+ return err;
+}
+
+
void symbol__init(void)
{
elf_version(EV_CURRENT);
diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h
index b53bf01..48b8e57 100644
--- a/tools/perf/util/symbol.h
+++ b/tools/perf/util/symbol.h
@@ -6,6 +6,7 @@
#include <linux/list.h>
#include <linux/rbtree.h>
#include "module.h"
+#include "event.h"
#ifdef HAVE_CPLUS_DEMANGLE
extern char *cplus_demangle(const char *, int);
@@ -72,9 +73,20 @@
symbol_filter_t filter, int verbose, int modules);
int dso__load_modules(struct dso *self, symbol_filter_t filter, int verbose);
int dso__load(struct dso *self, symbol_filter_t filter, int verbose);
+struct dso *dsos__findnew(const char *name);
+void dsos__fprintf(FILE *fp);
size_t dso__fprintf(struct dso *self, FILE *fp);
char dso__symtab_origin(const struct dso *self);
+int load_kernel(void);
+
void symbol__init(void);
+
+extern struct list_head dsos;
+extern struct dso *kernel_dso;
+extern struct dso *vdso;
+extern struct dso *hypervisor_dso;
+extern char *vmlinux;
+extern int modules;
#endif /* _PERF_SYMBOL_ */
diff --git a/tools/perf/util/util.h b/tools/perf/util/util.h
index 68fe157..d61a6f0 100644
--- a/tools/perf/util/util.h
+++ b/tools/perf/util/util.h
@@ -83,6 +83,7 @@
#include <inttypes.h>
#include "../../../include/linux/magic.h"
+
#ifndef NO_ICONV
#include <iconv.h>
#endif
diff --git a/tools/perf/util/values.c b/tools/perf/util/values.c
new file mode 100644
index 0000000..614cfaf
--- /dev/null
+++ b/tools/perf/util/values.c
@@ -0,0 +1,231 @@
+#include <stdlib.h>
+
+#include "util.h"
+#include "values.h"
+
+void perf_read_values_init(struct perf_read_values *values)
+{
+ values->threads_max = 16;
+ values->pid = malloc(values->threads_max * sizeof(*values->pid));
+ values->tid = malloc(values->threads_max * sizeof(*values->tid));
+ values->value = malloc(values->threads_max * sizeof(*values->value));
+ if (!values->pid || !values->tid || !values->value)
+ die("failed to allocate read_values threads arrays");
+ values->threads = 0;
+
+ values->counters_max = 16;
+ values->counterrawid = malloc(values->counters_max
+ * sizeof(*values->counterrawid));
+ values->countername = malloc(values->counters_max
+ * sizeof(*values->countername));
+ if (!values->counterrawid || !values->countername)
+ die("failed to allocate read_values counters arrays");
+ values->counters = 0;
+}
+
+void perf_read_values_destroy(struct perf_read_values *values)
+{
+ int i;
+
+ if (!values->threads_max || !values->counters_max)
+ return;
+
+ for (i = 0; i < values->threads; i++)
+ free(values->value[i]);
+ free(values->pid);
+ free(values->tid);
+ free(values->counterrawid);
+ for (i = 0; i < values->counters; i++)
+ free(values->countername[i]);
+ free(values->countername);
+}
+
+static void perf_read_values__enlarge_threads(struct perf_read_values *values)
+{
+ values->threads_max *= 2;
+ values->pid = realloc(values->pid,
+ values->threads_max * sizeof(*values->pid));
+ values->tid = realloc(values->tid,
+ values->threads_max * sizeof(*values->tid));
+ values->value = realloc(values->value,
+ values->threads_max * sizeof(*values->value));
+ if (!values->pid || !values->tid || !values->value)
+ die("failed to enlarge read_values threads arrays");
+}
+
+static int perf_read_values__findnew_thread(struct perf_read_values *values,
+ u32 pid, u32 tid)
+{
+ int i;
+
+ for (i = 0; i < values->threads; i++)
+ if (values->pid[i] == pid && values->tid[i] == tid)
+ return i;
+
+ if (values->threads == values->threads_max)
+ perf_read_values__enlarge_threads(values);
+
+ i = values->threads++;
+ values->pid[i] = pid;
+ values->tid[i] = tid;
+ values->value[i] = malloc(values->counters_max * sizeof(**values->value));
+ if (!values->value[i])
+ die("failed to allocate read_values counters array");
+
+ return i;
+}
+
+static void perf_read_values__enlarge_counters(struct perf_read_values *values)
+{
+ int i;
+
+ values->counters_max *= 2;
+ values->counterrawid = realloc(values->counterrawid,
+ values->counters_max * sizeof(*values->counterrawid));
+ values->countername = realloc(values->countername,
+ values->counters_max * sizeof(*values->countername));
+ if (!values->counterrawid || !values->countername)
+ die("failed to enlarge read_values counters arrays");
+
+ for (i = 0; i < values->threads; i++) {
+ values->value[i] = realloc(values->value[i],
+ values->counters_max * sizeof(**values->value));
+ if (!values->value[i])
+ die("failed to enlarge read_values counters arrays");
+ }
+}
+
+static int perf_read_values__findnew_counter(struct perf_read_values *values,
+ u64 rawid, char *name)
+{
+ int i;
+
+ for (i = 0; i < values->counters; i++)
+ if (values->counterrawid[i] == rawid)
+ return i;
+
+ if (values->counters == values->counters_max)
+ perf_read_values__enlarge_counters(values);
+
+ i = values->counters++;
+ values->counterrawid[i] = rawid;
+ values->countername[i] = strdup(name);
+
+ return i;
+}
+
+void perf_read_values_add_value(struct perf_read_values *values,
+ u32 pid, u32 tid,
+ u64 rawid, char *name, u64 value)
+{
+ int tindex, cindex;
+
+ tindex = perf_read_values__findnew_thread(values, pid, tid);
+ cindex = perf_read_values__findnew_counter(values, rawid, name);
+
+ values->value[tindex][cindex] = value;
+}
+
+static void perf_read_values__display_pretty(FILE *fp,
+ struct perf_read_values *values)
+{
+ int i, j;
+ int pidwidth, tidwidth;
+ int *counterwidth;
+
+ counterwidth = malloc(values->counters * sizeof(*counterwidth));
+ if (!counterwidth)
+ die("failed to allocate counterwidth array");
+ tidwidth = 3;
+ pidwidth = 3;
+ for (j = 0; j < values->counters; j++)
+ counterwidth[j] = strlen(values->countername[j]);
+ for (i = 0; i < values->threads; i++) {
+ int width;
+
+ width = snprintf(NULL, 0, "%d", values->pid[i]);
+ if (width > pidwidth)
+ pidwidth = width;
+ width = snprintf(NULL, 0, "%d", values->tid[i]);
+ if (width > tidwidth)
+ tidwidth = width;
+ for (j = 0; j < values->counters; j++) {
+ width = snprintf(NULL, 0, "%Lu", values->value[i][j]);
+ if (width > counterwidth[j])
+ counterwidth[j] = width;
+ }
+ }
+
+ fprintf(fp, "# %*s %*s", pidwidth, "PID", tidwidth, "TID");
+ for (j = 0; j < values->counters; j++)
+ fprintf(fp, " %*s", counterwidth[j], values->countername[j]);
+ fprintf(fp, "\n");
+
+ for (i = 0; i < values->threads; i++) {
+ fprintf(fp, " %*d %*d", pidwidth, values->pid[i],
+ tidwidth, values->tid[i]);
+ for (j = 0; j < values->counters; j++)
+ fprintf(fp, " %*Lu",
+ counterwidth[j], values->value[i][j]);
+ fprintf(fp, "\n");
+ }
+}
+
+static void perf_read_values__display_raw(FILE *fp,
+ struct perf_read_values *values)
+{
+ int width, pidwidth, tidwidth, namewidth, rawwidth, countwidth;
+ int i, j;
+
+ tidwidth = 3; /* TID */
+ pidwidth = 3; /* PID */
+ namewidth = 4; /* "Name" */
+ rawwidth = 3; /* "Raw" */
+ countwidth = 5; /* "Count" */
+
+ for (i = 0; i < values->threads; i++) {
+ width = snprintf(NULL, 0, "%d", values->pid[i]);
+ if (width > pidwidth)
+ pidwidth = width;
+ width = snprintf(NULL, 0, "%d", values->tid[i]);
+ if (width > tidwidth)
+ tidwidth = width;
+ }
+ for (j = 0; j < values->counters; j++) {
+ width = strlen(values->countername[j]);
+ if (width > namewidth)
+ namewidth = width;
+ width = snprintf(NULL, 0, "%llx", values->counterrawid[j]);
+ if (width > rawwidth)
+ rawwidth = width;
+ }
+ for (i = 0; i < values->threads; i++) {
+ for (j = 0; j < values->counters; j++) {
+ width = snprintf(NULL, 0, "%Lu", values->value[i][j]);
+ if (width > countwidth)
+ countwidth = width;
+ }
+ }
+
+ fprintf(fp, "# %*s %*s %*s %*s %*s\n",
+ pidwidth, "PID", tidwidth, "TID",
+ namewidth, "Name", rawwidth, "Raw",
+ countwidth, "Count");
+ for (i = 0; i < values->threads; i++)
+ for (j = 0; j < values->counters; j++)
+ fprintf(fp, " %*d %*d %*s %*llx %*Lu\n",
+ pidwidth, values->pid[i],
+ tidwidth, values->tid[i],
+ namewidth, values->countername[j],
+ rawwidth, values->counterrawid[j],
+ countwidth, values->value[i][j]);
+}
+
+void perf_read_values_display(FILE *fp, struct perf_read_values *values,
+ int raw)
+{
+ if (raw)
+ perf_read_values__display_raw(fp, values);
+ else
+ perf_read_values__display_pretty(fp, values);
+}
diff --git a/tools/perf/util/values.h b/tools/perf/util/values.h
new file mode 100644
index 0000000..f8960fd
--- /dev/null
+++ b/tools/perf/util/values.h
@@ -0,0 +1,27 @@
+#ifndef _PERF_VALUES_H
+#define _PERF_VALUES_H
+
+#include "types.h"
+
+struct perf_read_values {
+ int threads;
+ int threads_max;
+ u32 *pid, *tid;
+ int counters;
+ int counters_max;
+ u64 *counterrawid;
+ char **countername;
+ u64 **value;
+};
+
+void perf_read_values_init(struct perf_read_values *values);
+void perf_read_values_destroy(struct perf_read_values *values);
+
+void perf_read_values_add_value(struct perf_read_values *values,
+ u32 pid, u32 tid,
+ u64 rawid, char *name, u64 value);
+
+void perf_read_values_display(FILE *fp, struct perf_read_values *values,
+ int raw);
+
+#endif /* _PERF_VALUES_H */