2009-04-20 15:00:56 +02:00
|
|
|
|
/*
|
2009-06-02 23:37:05 +02:00
|
|
|
|
* builtin-top.c
|
|
|
|
|
*
|
|
|
|
|
* Builtin top command: Display a continuously updated profile of
|
|
|
|
|
* any workload, CPU or specific PID.
|
|
|
|
|
*
|
|
|
|
|
* Copyright (C) 2008, Red Hat Inc, Ingo Molnar <mingo@redhat.com>
|
|
|
|
|
*
|
|
|
|
|
* Improvements and fixes by:
|
|
|
|
|
*
|
|
|
|
|
* Arjan van de Ven <arjan@linux.intel.com>
|
|
|
|
|
* Yanmin Zhang <yanmin.zhang@intel.com>
|
|
|
|
|
* Wu Fengguang <fengguang.wu@intel.com>
|
|
|
|
|
* Mike Galbraith <efault@gmx.de>
|
|
|
|
|
* Paul Mackerras <paulus@samba.org>
|
|
|
|
|
*
|
|
|
|
|
* Released under the GPL v2. (and only v2, not any later version)
|
2009-04-20 15:00:56 +02:00
|
|
|
|
*/
|
2009-06-02 23:37:05 +02:00
|
|
|
|
#include "builtin.h"
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-05-23 18:28:58 +02:00
|
|
|
|
#include "perf.h"
|
2009-06-02 23:37:05 +02:00
|
|
|
|
|
2009-05-28 19:55:41 +02:00
|
|
|
|
#include "util/symbol.h"
|
2009-06-04 15:19:47 +02:00
|
|
|
|
#include "util/color.h"
|
2009-04-27 08:02:14 +02:00
|
|
|
|
#include "util/util.h"
|
2009-05-28 19:55:41 +02:00
|
|
|
|
#include "util/rbtree.h"
|
2009-05-26 09:17:18 +02:00
|
|
|
|
#include "util/parse-options.h"
|
|
|
|
|
#include "util/parse-events.h"
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
|
|
|
|
#include <assert.h>
|
|
|
|
|
#include <fcntl.h>
|
2009-05-26 09:17:18 +02:00
|
|
|
|
|
2009-04-20 15:00:56 +02:00
|
|
|
|
#include <stdio.h>
|
2009-05-26 09:17:18 +02:00
|
|
|
|
|
2009-04-20 15:00:56 +02:00
|
|
|
|
#include <errno.h>
|
|
|
|
|
#include <time.h>
|
|
|
|
|
#include <sched.h>
|
|
|
|
|
#include <pthread.h>
|
|
|
|
|
|
|
|
|
|
#include <sys/syscall.h>
|
|
|
|
|
#include <sys/ioctl.h>
|
|
|
|
|
#include <sys/poll.h>
|
|
|
|
|
#include <sys/prctl.h>
|
|
|
|
|
#include <sys/wait.h>
|
|
|
|
|
#include <sys/uio.h>
|
|
|
|
|
#include <sys/mman.h>
|
|
|
|
|
|
|
|
|
|
#include <linux/unistd.h>
|
|
|
|
|
#include <linux/types.h>
|
|
|
|
|
|
2009-06-06 09:58:57 +02:00
|
|
|
|
static int fd[MAX_NR_CPUS][MAX_COUNTERS];
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-06-06 09:58:57 +02:00
|
|
|
|
static int system_wide = 0;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-06-06 09:58:57 +02:00
|
|
|
|
static int default_interval = 100000;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-06-04 08:53:05 +02:00
|
|
|
|
static __u64 count_filter = 5;
|
|
|
|
|
static int print_entries = 15;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-06-04 08:53:05 +02:00
|
|
|
|
static int target_pid = -1;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
static int profile_cpu = -1;
|
|
|
|
|
static int nr_cpus = 0;
|
|
|
|
|
static unsigned int realtime_prio = 0;
|
|
|
|
|
static int group = 0;
|
|
|
|
|
static unsigned int page_size;
|
2009-06-05 13:27:02 +02:00
|
|
|
|
static unsigned int mmap_pages = 16;
|
|
|
|
|
static int freq = 0;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
|
|
|
|
static char *sym_filter;
|
|
|
|
|
static unsigned long filter_start;
|
|
|
|
|
static unsigned long filter_end;
|
|
|
|
|
|
|
|
|
|
static int delay_secs = 2;
|
|
|
|
|
static int zero;
|
|
|
|
|
static int dump_symtab;
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Symbols
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
static uint64_t min_ip;
|
|
|
|
|
static uint64_t max_ip = -1ll;
|
|
|
|
|
|
|
|
|
|
struct sym_entry {
|
2009-05-28 19:55:41 +02:00
|
|
|
|
struct rb_node rb_node;
|
|
|
|
|
struct list_head node;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
unsigned long count[MAX_COUNTERS];
|
2009-05-29 22:03:07 +02:00
|
|
|
|
unsigned long snap_count;
|
|
|
|
|
double weight;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
int skip;
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
struct sym_entry *sym_filter_entry;
|
|
|
|
|
|
2009-06-06 09:58:57 +02:00
|
|
|
|
struct dso *kernel_dso;
|
2009-05-28 19:55:41 +02:00
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Symbols will be added here in record_ip and will get out
|
|
|
|
|
* after decayed.
|
|
|
|
|
*/
|
|
|
|
|
static LIST_HEAD(active_symbols);
|
2009-05-29 22:03:07 +02:00
|
|
|
|
static pthread_mutex_t active_symbols_lock = PTHREAD_MUTEX_INITIALIZER;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Ordering weight: count-1 * count-2 * ... / count-n
|
|
|
|
|
*/
|
|
|
|
|
static double sym_weight(const struct sym_entry *sym)
|
|
|
|
|
{
|
2009-05-29 22:03:07 +02:00
|
|
|
|
double weight = sym->snap_count;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
int counter;
|
|
|
|
|
|
|
|
|
|
for (counter = 1; counter < nr_counters-1; counter++)
|
|
|
|
|
weight *= sym->count[counter];
|
|
|
|
|
|
|
|
|
|
weight /= (sym->count[counter] + 1);
|
|
|
|
|
|
|
|
|
|
return weight;
|
|
|
|
|
}
|
|
|
|
|
|
2009-06-05 14:29:10 +02:00
|
|
|
|
static long samples;
|
|
|
|
|
static long userspace_samples;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
static const char CONSOLE_CLEAR[] = "[H[2J";
|
|
|
|
|
|
2009-05-29 22:03:07 +02:00
|
|
|
|
static void __list_insert_active_sym(struct sym_entry *syme)
|
2009-05-28 19:55:41 +02:00
|
|
|
|
{
|
|
|
|
|
list_add(&syme->node, &active_symbols);
|
|
|
|
|
}
|
|
|
|
|
|
2009-05-29 22:03:07 +02:00
|
|
|
|
static void list_remove_active_sym(struct sym_entry *syme)
|
|
|
|
|
{
|
|
|
|
|
pthread_mutex_lock(&active_symbols_lock);
|
|
|
|
|
list_del_init(&syme->node);
|
|
|
|
|
pthread_mutex_unlock(&active_symbols_lock);
|
|
|
|
|
}
|
|
|
|
|
|
2009-05-28 19:55:41 +02:00
|
|
|
|
static void rb_insert_active_sym(struct rb_root *tree, struct sym_entry *se)
|
|
|
|
|
{
|
|
|
|
|
struct rb_node **p = &tree->rb_node;
|
|
|
|
|
struct rb_node *parent = NULL;
|
|
|
|
|
struct sym_entry *iter;
|
|
|
|
|
|
|
|
|
|
while (*p != NULL) {
|
|
|
|
|
parent = *p;
|
|
|
|
|
iter = rb_entry(parent, struct sym_entry, rb_node);
|
|
|
|
|
|
2009-05-29 22:03:07 +02:00
|
|
|
|
if (se->weight > iter->weight)
|
2009-05-28 19:55:41 +02:00
|
|
|
|
p = &(*p)->rb_left;
|
|
|
|
|
else
|
|
|
|
|
p = &(*p)->rb_right;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
rb_link_node(&se->rb_node, parent, p);
|
|
|
|
|
rb_insert_color(&se->rb_node, tree);
|
|
|
|
|
}
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
|
|
|
|
static void print_sym_table(void)
|
|
|
|
|
{
|
2009-06-03 21:48:40 +02:00
|
|
|
|
int printed = 0, j;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
int counter;
|
2009-06-05 14:29:10 +02:00
|
|
|
|
float samples_per_sec = samples/delay_secs;
|
|
|
|
|
float ksamples_per_sec = (samples-userspace_samples)/delay_secs;
|
|
|
|
|
float sum_ksamples = 0.0;
|
2009-05-28 19:55:41 +02:00
|
|
|
|
struct sym_entry *syme, *n;
|
|
|
|
|
struct rb_root tmp = RB_ROOT;
|
|
|
|
|
struct rb_node *nd;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-06-05 14:29:10 +02:00
|
|
|
|
samples = userspace_samples = 0;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-05-28 19:55:41 +02:00
|
|
|
|
/* Sort the active symbols */
|
2009-05-29 22:03:07 +02:00
|
|
|
|
pthread_mutex_lock(&active_symbols_lock);
|
|
|
|
|
syme = list_entry(active_symbols.next, struct sym_entry, node);
|
|
|
|
|
pthread_mutex_unlock(&active_symbols_lock);
|
|
|
|
|
|
|
|
|
|
list_for_each_entry_safe_from(syme, n, &active_symbols, node) {
|
|
|
|
|
syme->snap_count = syme->count[0];
|
|
|
|
|
if (syme->snap_count != 0) {
|
|
|
|
|
syme->weight = sym_weight(syme);
|
2009-05-28 19:55:41 +02:00
|
|
|
|
rb_insert_active_sym(&tmp, syme);
|
2009-06-05 14:29:10 +02:00
|
|
|
|
sum_ksamples += syme->snap_count;
|
perf top: Reduce display overhead
Iterate over the symbol table once per display interval, and
copy/sort/tally/decay only those symbols which are active.
Before:
top - 10:14:53 up 4:08, 17 users, load average: 1.17, 1.53, 1.49
Tasks: 273 total, 5 running, 268 sleeping, 0 stopped, 0 zombie
Cpu(s): 6.9%us, 38.2%sy, 0.0%ni, 19.9%id, 0.0%wa, 0.0%hi, 35.0%si, 0.0%st
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ P COMMAND
28504 root 20 0 1044 260 164 S 58 0.0 0:04.19 2 netserver
28499 root 20 0 1040 412 316 R 51 0.0 0:04.15 0 netperf
28500 root 20 0 1040 408 316 R 50 0.0 0:04.14 1 netperf
28503 root 20 0 1044 260 164 S 50 0.0 0:04.01 1 netserver
28501 root 20 0 1044 260 164 S 49 0.0 0:03.99 0 netserver
28502 root 20 0 1040 412 316 S 43 0.0 0:03.96 2 netperf
28468 root 20 0 1892m 325m 972 S 16 10.8 0:10.50 3 perf
28467 root 20 0 1892m 325m 972 R 2 10.8 0:00.72 3 perf
After:
top - 10:16:30 up 4:10, 17 users, load average: 2.27, 1.88, 1.62
Tasks: 273 total, 6 running, 267 sleeping, 0 stopped, 0 zombie
Cpu(s): 2.5%us, 39.7%sy, 0.0%ni, 24.6%id, 0.0%wa, 0.0%hi, 33.3%si, 0.0%st
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ P COMMAND
28590 root 20 0 1040 412 316 S 54 0.0 0:07.85 2 netperf
28589 root 20 0 1044 260 164 R 54 0.0 0:07.84 0 netserver
28588 root 20 0 1040 412 316 R 50 0.0 0:07.89 1 netperf
28591 root 20 0 1044 256 164 S 50 0.0 0:07.82 1 netserver
28587 root 20 0 1040 408 316 R 47 0.0 0:07.61 0 netperf
28592 root 20 0 1044 260 164 R 47 0.0 0:07.85 2 netserver
28378 root 20 0 8732 1300 860 R 2 0.0 0:01.81 3 top
28577 root 20 0 1892m 165m 972 R 2 5.5 0:00.48 3 perf
28578 root 20 0 1892m 165m 972 S 2 5.5 0:00.04 3 perf
[ Impact: optimization ]
Signed-off-by: Mike Galbraith <efault@gmx.de>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-05-25 09:57:56 +02:00
|
|
|
|
|
|
|
|
|
for (j = 0; j < nr_counters; j++)
|
2009-05-28 19:55:41 +02:00
|
|
|
|
syme->count[j] = zero ? 0 : syme->count[j] * 7 / 8;
|
|
|
|
|
} else
|
2009-05-29 22:03:07 +02:00
|
|
|
|
list_remove_active_sym(syme);
|
perf top: Reduce display overhead
Iterate over the symbol table once per display interval, and
copy/sort/tally/decay only those symbols which are active.
Before:
top - 10:14:53 up 4:08, 17 users, load average: 1.17, 1.53, 1.49
Tasks: 273 total, 5 running, 268 sleeping, 0 stopped, 0 zombie
Cpu(s): 6.9%us, 38.2%sy, 0.0%ni, 19.9%id, 0.0%wa, 0.0%hi, 35.0%si, 0.0%st
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ P COMMAND
28504 root 20 0 1044 260 164 S 58 0.0 0:04.19 2 netserver
28499 root 20 0 1040 412 316 R 51 0.0 0:04.15 0 netperf
28500 root 20 0 1040 408 316 R 50 0.0 0:04.14 1 netperf
28503 root 20 0 1044 260 164 S 50 0.0 0:04.01 1 netserver
28501 root 20 0 1044 260 164 S 49 0.0 0:03.99 0 netserver
28502 root 20 0 1040 412 316 S 43 0.0 0:03.96 2 netperf
28468 root 20 0 1892m 325m 972 S 16 10.8 0:10.50 3 perf
28467 root 20 0 1892m 325m 972 R 2 10.8 0:00.72 3 perf
After:
top - 10:16:30 up 4:10, 17 users, load average: 2.27, 1.88, 1.62
Tasks: 273 total, 6 running, 267 sleeping, 0 stopped, 0 zombie
Cpu(s): 2.5%us, 39.7%sy, 0.0%ni, 24.6%id, 0.0%wa, 0.0%hi, 33.3%si, 0.0%st
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ P COMMAND
28590 root 20 0 1040 412 316 S 54 0.0 0:07.85 2 netperf
28589 root 20 0 1044 260 164 R 54 0.0 0:07.84 0 netserver
28588 root 20 0 1040 412 316 R 50 0.0 0:07.89 1 netperf
28591 root 20 0 1044 256 164 S 50 0.0 0:07.82 1 netserver
28587 root 20 0 1040 408 316 R 47 0.0 0:07.61 0 netperf
28592 root 20 0 1044 260 164 R 47 0.0 0:07.85 2 netserver
28378 root 20 0 8732 1300 860 R 2 0.0 0:01.81 3 top
28577 root 20 0 1892m 165m 972 R 2 5.5 0:00.48 3 perf
28578 root 20 0 1892m 165m 972 S 2 5.5 0:00.04 3 perf
[ Impact: optimization ]
Signed-off-by: Mike Galbraith <efault@gmx.de>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-05-25 09:57:56 +02:00
|
|
|
|
}
|
|
|
|
|
|
2009-06-04 20:48:04 +02:00
|
|
|
|
puts(CONSOLE_CLEAR);
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
|
|
|
|
printf(
|
|
|
|
|
"------------------------------------------------------------------------------\n");
|
2009-06-03 19:17:25 +02:00
|
|
|
|
printf( " PerfTop:%8.0f irqs/sec kernel:%4.1f%% [",
|
2009-06-05 14:29:10 +02:00
|
|
|
|
samples_per_sec,
|
|
|
|
|
100.0 - (100.0*((samples_per_sec-ksamples_per_sec)/samples_per_sec)));
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-06-05 13:27:02 +02:00
|
|
|
|
if (nr_counters == 1) {
|
2009-06-06 09:58:57 +02:00
|
|
|
|
printf("%Ld", attrs[0].sample_period);
|
2009-06-05 13:27:02 +02:00
|
|
|
|
if (freq)
|
|
|
|
|
printf("Hz ");
|
|
|
|
|
else
|
|
|
|
|
printf(" ");
|
|
|
|
|
}
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
|
|
|
|
for (counter = 0; counter < nr_counters; counter++) {
|
|
|
|
|
if (counter)
|
|
|
|
|
printf("/");
|
|
|
|
|
|
|
|
|
|
printf("%s", event_name(counter));
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
printf( "], ");
|
|
|
|
|
|
2009-05-26 09:17:18 +02:00
|
|
|
|
if (target_pid != -1)
|
|
|
|
|
printf(" (target_pid: %d", target_pid);
|
2009-04-20 15:00:56 +02:00
|
|
|
|
else
|
|
|
|
|
printf(" (all");
|
|
|
|
|
|
|
|
|
|
if (profile_cpu != -1)
|
|
|
|
|
printf(", cpu: %d)\n", profile_cpu);
|
|
|
|
|
else {
|
2009-05-26 09:17:18 +02:00
|
|
|
|
if (target_pid != -1)
|
2009-04-20 15:00:56 +02:00
|
|
|
|
printf(")\n");
|
|
|
|
|
else
|
|
|
|
|
printf(", %d CPUs)\n", nr_cpus);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
printf("------------------------------------------------------------------------------\n\n");
|
|
|
|
|
|
|
|
|
|
if (nr_counters == 1)
|
2009-06-05 14:29:10 +02:00
|
|
|
|
printf(" samples pcnt");
|
2009-04-20 15:00:56 +02:00
|
|
|
|
else
|
2009-06-05 14:29:10 +02:00
|
|
|
|
printf(" weight samples pcnt");
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
|
|
|
|
printf(" RIP kernel function\n"
|
2009-06-05 14:29:10 +02:00
|
|
|
|
" ______ _______ _____ ________________ _______________\n\n"
|
2009-04-20 15:00:56 +02:00
|
|
|
|
);
|
|
|
|
|
|
2009-05-28 19:55:41 +02:00
|
|
|
|
for (nd = rb_first(&tmp); nd; nd = rb_next(nd)) {
|
|
|
|
|
struct sym_entry *syme = rb_entry(nd, struct sym_entry, rb_node);
|
|
|
|
|
struct symbol *sym = (struct symbol *)(syme + 1);
|
2009-06-04 15:19:47 +02:00
|
|
|
|
char *color = PERF_COLOR_NORMAL;
|
|
|
|
|
double pcnt;
|
perf top: Reduce display overhead
Iterate over the symbol table once per display interval, and
copy/sort/tally/decay only those symbols which are active.
Before:
top - 10:14:53 up 4:08, 17 users, load average: 1.17, 1.53, 1.49
Tasks: 273 total, 5 running, 268 sleeping, 0 stopped, 0 zombie
Cpu(s): 6.9%us, 38.2%sy, 0.0%ni, 19.9%id, 0.0%wa, 0.0%hi, 35.0%si, 0.0%st
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ P COMMAND
28504 root 20 0 1044 260 164 S 58 0.0 0:04.19 2 netserver
28499 root 20 0 1040 412 316 R 51 0.0 0:04.15 0 netperf
28500 root 20 0 1040 408 316 R 50 0.0 0:04.14 1 netperf
28503 root 20 0 1044 260 164 S 50 0.0 0:04.01 1 netserver
28501 root 20 0 1044 260 164 S 49 0.0 0:03.99 0 netserver
28502 root 20 0 1040 412 316 S 43 0.0 0:03.96 2 netperf
28468 root 20 0 1892m 325m 972 S 16 10.8 0:10.50 3 perf
28467 root 20 0 1892m 325m 972 R 2 10.8 0:00.72 3 perf
After:
top - 10:16:30 up 4:10, 17 users, load average: 2.27, 1.88, 1.62
Tasks: 273 total, 6 running, 267 sleeping, 0 stopped, 0 zombie
Cpu(s): 2.5%us, 39.7%sy, 0.0%ni, 24.6%id, 0.0%wa, 0.0%hi, 33.3%si, 0.0%st
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ P COMMAND
28590 root 20 0 1040 412 316 S 54 0.0 0:07.85 2 netperf
28589 root 20 0 1044 260 164 R 54 0.0 0:07.84 0 netserver
28588 root 20 0 1040 412 316 R 50 0.0 0:07.89 1 netperf
28591 root 20 0 1044 256 164 S 50 0.0 0:07.82 1 netserver
28587 root 20 0 1040 408 316 R 47 0.0 0:07.61 0 netperf
28592 root 20 0 1044 260 164 R 47 0.0 0:07.85 2 netserver
28378 root 20 0 8732 1300 860 R 2 0.0 0:01.81 3 top
28577 root 20 0 1892m 165m 972 R 2 5.5 0:00.48 3 perf
28578 root 20 0 1892m 165m 972 S 2 5.5 0:00.04 3 perf
[ Impact: optimization ]
Signed-off-by: Mike Galbraith <efault@gmx.de>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-05-25 09:57:56 +02:00
|
|
|
|
|
2009-06-04 08:53:05 +02:00
|
|
|
|
if (++printed > print_entries || syme->snap_count < count_filter)
|
2009-05-29 22:03:07 +02:00
|
|
|
|
continue;
|
perf top: Reduce display overhead
Iterate over the symbol table once per display interval, and
copy/sort/tally/decay only those symbols which are active.
Before:
top - 10:14:53 up 4:08, 17 users, load average: 1.17, 1.53, 1.49
Tasks: 273 total, 5 running, 268 sleeping, 0 stopped, 0 zombie
Cpu(s): 6.9%us, 38.2%sy, 0.0%ni, 19.9%id, 0.0%wa, 0.0%hi, 35.0%si, 0.0%st
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ P COMMAND
28504 root 20 0 1044 260 164 S 58 0.0 0:04.19 2 netserver
28499 root 20 0 1040 412 316 R 51 0.0 0:04.15 0 netperf
28500 root 20 0 1040 408 316 R 50 0.0 0:04.14 1 netperf
28503 root 20 0 1044 260 164 S 50 0.0 0:04.01 1 netserver
28501 root 20 0 1044 260 164 S 49 0.0 0:03.99 0 netserver
28502 root 20 0 1040 412 316 S 43 0.0 0:03.96 2 netperf
28468 root 20 0 1892m 325m 972 S 16 10.8 0:10.50 3 perf
28467 root 20 0 1892m 325m 972 R 2 10.8 0:00.72 3 perf
After:
top - 10:16:30 up 4:10, 17 users, load average: 2.27, 1.88, 1.62
Tasks: 273 total, 6 running, 267 sleeping, 0 stopped, 0 zombie
Cpu(s): 2.5%us, 39.7%sy, 0.0%ni, 24.6%id, 0.0%wa, 0.0%hi, 33.3%si, 0.0%st
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ P COMMAND
28590 root 20 0 1040 412 316 S 54 0.0 0:07.85 2 netperf
28589 root 20 0 1044 260 164 R 54 0.0 0:07.84 0 netserver
28588 root 20 0 1040 412 316 R 50 0.0 0:07.89 1 netperf
28591 root 20 0 1044 256 164 S 50 0.0 0:07.82 1 netserver
28587 root 20 0 1040 408 316 R 47 0.0 0:07.61 0 netperf
28592 root 20 0 1044 260 164 R 47 0.0 0:07.85 2 netserver
28378 root 20 0 8732 1300 860 R 2 0.0 0:01.81 3 top
28577 root 20 0 1892m 165m 972 R 2 5.5 0:00.48 3 perf
28578 root 20 0 1892m 165m 972 S 2 5.5 0:00.04 3 perf
[ Impact: optimization ]
Signed-off-by: Mike Galbraith <efault@gmx.de>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-05-25 09:57:56 +02:00
|
|
|
|
|
2009-06-05 14:29:10 +02:00
|
|
|
|
pcnt = 100.0 - (100.0 * ((sum_ksamples - syme->snap_count) /
|
|
|
|
|
sum_ksamples));
|
perf top: Reduce display overhead
Iterate over the symbol table once per display interval, and
copy/sort/tally/decay only those symbols which are active.
Before:
top - 10:14:53 up 4:08, 17 users, load average: 1.17, 1.53, 1.49
Tasks: 273 total, 5 running, 268 sleeping, 0 stopped, 0 zombie
Cpu(s): 6.9%us, 38.2%sy, 0.0%ni, 19.9%id, 0.0%wa, 0.0%hi, 35.0%si, 0.0%st
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ P COMMAND
28504 root 20 0 1044 260 164 S 58 0.0 0:04.19 2 netserver
28499 root 20 0 1040 412 316 R 51 0.0 0:04.15 0 netperf
28500 root 20 0 1040 408 316 R 50 0.0 0:04.14 1 netperf
28503 root 20 0 1044 260 164 S 50 0.0 0:04.01 1 netserver
28501 root 20 0 1044 260 164 S 49 0.0 0:03.99 0 netserver
28502 root 20 0 1040 412 316 S 43 0.0 0:03.96 2 netperf
28468 root 20 0 1892m 325m 972 S 16 10.8 0:10.50 3 perf
28467 root 20 0 1892m 325m 972 R 2 10.8 0:00.72 3 perf
After:
top - 10:16:30 up 4:10, 17 users, load average: 2.27, 1.88, 1.62
Tasks: 273 total, 6 running, 267 sleeping, 0 stopped, 0 zombie
Cpu(s): 2.5%us, 39.7%sy, 0.0%ni, 24.6%id, 0.0%wa, 0.0%hi, 33.3%si, 0.0%st
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ P COMMAND
28590 root 20 0 1040 412 316 S 54 0.0 0:07.85 2 netperf
28589 root 20 0 1044 260 164 R 54 0.0 0:07.84 0 netserver
28588 root 20 0 1040 412 316 R 50 0.0 0:07.89 1 netperf
28591 root 20 0 1044 256 164 S 50 0.0 0:07.82 1 netserver
28587 root 20 0 1040 408 316 R 47 0.0 0:07.61 0 netperf
28592 root 20 0 1044 260 164 R 47 0.0 0:07.85 2 netserver
28378 root 20 0 8732 1300 860 R 2 0.0 0:01.81 3 top
28577 root 20 0 1892m 165m 972 R 2 5.5 0:00.48 3 perf
28578 root 20 0 1892m 165m 972 S 2 5.5 0:00.04 3 perf
[ Impact: optimization ]
Signed-off-by: Mike Galbraith <efault@gmx.de>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-05-25 09:57:56 +02:00
|
|
|
|
|
2009-06-04 15:19:47 +02:00
|
|
|
|
/*
|
|
|
|
|
* We color high-overhead entries in red, low-overhead
|
|
|
|
|
* entries in green - and keep the middle ground normal:
|
|
|
|
|
*/
|
|
|
|
|
if (pcnt >= 5.0)
|
|
|
|
|
color = PERF_COLOR_RED;
|
|
|
|
|
if (pcnt < 0.5)
|
|
|
|
|
color = PERF_COLOR_GREEN;
|
|
|
|
|
|
perf top: Reduce display overhead
Iterate over the symbol table once per display interval, and
copy/sort/tally/decay only those symbols which are active.
Before:
top - 10:14:53 up 4:08, 17 users, load average: 1.17, 1.53, 1.49
Tasks: 273 total, 5 running, 268 sleeping, 0 stopped, 0 zombie
Cpu(s): 6.9%us, 38.2%sy, 0.0%ni, 19.9%id, 0.0%wa, 0.0%hi, 35.0%si, 0.0%st
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ P COMMAND
28504 root 20 0 1044 260 164 S 58 0.0 0:04.19 2 netserver
28499 root 20 0 1040 412 316 R 51 0.0 0:04.15 0 netperf
28500 root 20 0 1040 408 316 R 50 0.0 0:04.14 1 netperf
28503 root 20 0 1044 260 164 S 50 0.0 0:04.01 1 netserver
28501 root 20 0 1044 260 164 S 49 0.0 0:03.99 0 netserver
28502 root 20 0 1040 412 316 S 43 0.0 0:03.96 2 netperf
28468 root 20 0 1892m 325m 972 S 16 10.8 0:10.50 3 perf
28467 root 20 0 1892m 325m 972 R 2 10.8 0:00.72 3 perf
After:
top - 10:16:30 up 4:10, 17 users, load average: 2.27, 1.88, 1.62
Tasks: 273 total, 6 running, 267 sleeping, 0 stopped, 0 zombie
Cpu(s): 2.5%us, 39.7%sy, 0.0%ni, 24.6%id, 0.0%wa, 0.0%hi, 33.3%si, 0.0%st
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ P COMMAND
28590 root 20 0 1040 412 316 S 54 0.0 0:07.85 2 netperf
28589 root 20 0 1044 260 164 R 54 0.0 0:07.84 0 netserver
28588 root 20 0 1040 412 316 R 50 0.0 0:07.89 1 netperf
28591 root 20 0 1044 256 164 S 50 0.0 0:07.82 1 netserver
28587 root 20 0 1040 408 316 R 47 0.0 0:07.61 0 netperf
28592 root 20 0 1044 260 164 R 47 0.0 0:07.85 2 netserver
28378 root 20 0 8732 1300 860 R 2 0.0 0:01.81 3 top
28577 root 20 0 1892m 165m 972 R 2 5.5 0:00.48 3 perf
28578 root 20 0 1892m 165m 972 S 2 5.5 0:00.04 3 perf
[ Impact: optimization ]
Signed-off-by: Mike Galbraith <efault@gmx.de>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-05-25 09:57:56 +02:00
|
|
|
|
if (nr_counters == 1)
|
2009-06-05 14:29:10 +02:00
|
|
|
|
printf("%20.2f - ", syme->weight);
|
perf top: Reduce display overhead
Iterate over the symbol table once per display interval, and
copy/sort/tally/decay only those symbols which are active.
Before:
top - 10:14:53 up 4:08, 17 users, load average: 1.17, 1.53, 1.49
Tasks: 273 total, 5 running, 268 sleeping, 0 stopped, 0 zombie
Cpu(s): 6.9%us, 38.2%sy, 0.0%ni, 19.9%id, 0.0%wa, 0.0%hi, 35.0%si, 0.0%st
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ P COMMAND
28504 root 20 0 1044 260 164 S 58 0.0 0:04.19 2 netserver
28499 root 20 0 1040 412 316 R 51 0.0 0:04.15 0 netperf
28500 root 20 0 1040 408 316 R 50 0.0 0:04.14 1 netperf
28503 root 20 0 1044 260 164 S 50 0.0 0:04.01 1 netserver
28501 root 20 0 1044 260 164 S 49 0.0 0:03.99 0 netserver
28502 root 20 0 1040 412 316 S 43 0.0 0:03.96 2 netperf
28468 root 20 0 1892m 325m 972 S 16 10.8 0:10.50 3 perf
28467 root 20 0 1892m 325m 972 R 2 10.8 0:00.72 3 perf
After:
top - 10:16:30 up 4:10, 17 users, load average: 2.27, 1.88, 1.62
Tasks: 273 total, 6 running, 267 sleeping, 0 stopped, 0 zombie
Cpu(s): 2.5%us, 39.7%sy, 0.0%ni, 24.6%id, 0.0%wa, 0.0%hi, 33.3%si, 0.0%st
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ P COMMAND
28590 root 20 0 1040 412 316 S 54 0.0 0:07.85 2 netperf
28589 root 20 0 1044 260 164 R 54 0.0 0:07.84 0 netserver
28588 root 20 0 1040 412 316 R 50 0.0 0:07.89 1 netperf
28591 root 20 0 1044 256 164 S 50 0.0 0:07.82 1 netserver
28587 root 20 0 1040 408 316 R 47 0.0 0:07.61 0 netperf
28592 root 20 0 1044 260 164 R 47 0.0 0:07.85 2 netserver
28378 root 20 0 8732 1300 860 R 2 0.0 0:01.81 3 top
28577 root 20 0 1892m 165m 972 R 2 5.5 0:00.48 3 perf
28578 root 20 0 1892m 165m 972 S 2 5.5 0:00.04 3 perf
[ Impact: optimization ]
Signed-off-by: Mike Galbraith <efault@gmx.de>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-05-25 09:57:56 +02:00
|
|
|
|
else
|
2009-06-05 14:29:10 +02:00
|
|
|
|
printf("%9.1f %10ld - ", syme->weight, syme->snap_count);
|
2009-06-04 15:19:47 +02:00
|
|
|
|
|
|
|
|
|
color_fprintf(stdout, color, "%4.1f%%", pcnt);
|
|
|
|
|
printf(" - %016llx : %s\n", sym->start, sym->name);
|
2009-04-20 15:00:56 +02:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void *display_thread(void *arg)
|
|
|
|
|
{
|
2009-06-04 20:48:04 +02:00
|
|
|
|
struct pollfd stdin_poll = { .fd = 0, .events = POLLIN };
|
|
|
|
|
int delay_msecs = delay_secs * 1000;
|
|
|
|
|
|
2009-06-03 19:17:25 +02:00
|
|
|
|
printf("PerfTop refresh period: %d seconds\n", delay_secs);
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-06-04 20:48:04 +02:00
|
|
|
|
do {
|
2009-04-20 15:00:56 +02:00
|
|
|
|
print_sym_table();
|
2009-06-04 20:48:04 +02:00
|
|
|
|
} while (!poll(&stdin_poll, 1, delay_msecs) == 1);
|
|
|
|
|
|
|
|
|
|
printf("key pressed - exiting.\n");
|
|
|
|
|
exit(0);
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
|
2009-05-28 19:55:41 +02:00
|
|
|
|
static int symbol_filter(struct dso *self, struct symbol *sym)
|
2009-04-20 15:00:56 +02:00
|
|
|
|
{
|
2009-05-28 19:55:41 +02:00
|
|
|
|
static int filter_match;
|
|
|
|
|
struct sym_entry *syme;
|
|
|
|
|
const char *name = sym->name;
|
|
|
|
|
|
|
|
|
|
if (!strcmp(name, "_text") ||
|
|
|
|
|
!strcmp(name, "_etext") ||
|
|
|
|
|
!strcmp(name, "_sinittext") ||
|
|
|
|
|
!strncmp("init_module", name, 11) ||
|
|
|
|
|
!strncmp("cleanup_module", name, 14) ||
|
|
|
|
|
strstr(name, "_text_start") ||
|
|
|
|
|
strstr(name, "_text_end"))
|
2009-04-20 15:00:56 +02:00
|
|
|
|
return 1;
|
|
|
|
|
|
2009-05-28 19:55:41 +02:00
|
|
|
|
syme = dso__sym_priv(self, sym);
|
2009-06-05 14:29:10 +02:00
|
|
|
|
/* Tag samples to be skipped. */
|
2009-05-28 19:55:41 +02:00
|
|
|
|
if (!strcmp("default_idle", name) ||
|
|
|
|
|
!strcmp("cpu_idle", name) ||
|
|
|
|
|
!strcmp("enter_idle", name) ||
|
|
|
|
|
!strcmp("exit_idle", name) ||
|
|
|
|
|
!strcmp("mwait_idle", name))
|
|
|
|
|
syme->skip = 1;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
|
|
|
|
if (filter_match == 1) {
|
2009-05-28 19:55:41 +02:00
|
|
|
|
filter_end = sym->start;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
filter_match = -1;
|
|
|
|
|
if (filter_end - filter_start > 10000) {
|
2009-05-28 19:55:41 +02:00
|
|
|
|
fprintf(stderr,
|
|
|
|
|
"hm, too large filter symbol <%s> - skipping.\n",
|
2009-04-20 15:00:56 +02:00
|
|
|
|
sym_filter);
|
2009-05-28 19:55:41 +02:00
|
|
|
|
fprintf(stderr, "symbol filter start: %016lx\n",
|
|
|
|
|
filter_start);
|
|
|
|
|
fprintf(stderr, " end: %016lx\n",
|
|
|
|
|
filter_end);
|
2009-04-20 15:00:56 +02:00
|
|
|
|
filter_end = filter_start = 0;
|
|
|
|
|
sym_filter = NULL;
|
|
|
|
|
sleep(1);
|
|
|
|
|
}
|
|
|
|
|
}
|
2009-05-28 19:55:41 +02:00
|
|
|
|
|
|
|
|
|
if (filter_match == 0 && sym_filter && !strcmp(name, sym_filter)) {
|
2009-04-20 15:00:56 +02:00
|
|
|
|
filter_match = 1;
|
2009-05-28 19:55:41 +02:00
|
|
|
|
filter_start = sym->start;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
}
|
|
|
|
|
|
2009-05-28 19:55:41 +02:00
|
|
|
|
|
2009-04-20 15:00:56 +02:00
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2009-05-28 19:55:41 +02:00
|
|
|
|
static int parse_symbols(void)
|
2009-04-20 15:00:56 +02:00
|
|
|
|
{
|
2009-05-28 19:55:41 +02:00
|
|
|
|
struct rb_node *node;
|
|
|
|
|
struct symbol *sym;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-05-28 19:55:41 +02:00
|
|
|
|
kernel_dso = dso__new("[kernel]", sizeof(struct sym_entry));
|
|
|
|
|
if (kernel_dso == NULL)
|
|
|
|
|
return -1;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-06-04 14:13:04 +02:00
|
|
|
|
if (dso__load_kernel(kernel_dso, NULL, symbol_filter, 1) != 0)
|
2009-05-28 19:55:41 +02:00
|
|
|
|
goto out_delete_dso;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-05-28 19:55:41 +02:00
|
|
|
|
node = rb_first(&kernel_dso->syms);
|
|
|
|
|
sym = rb_entry(node, struct symbol, rb_node);
|
|
|
|
|
min_ip = sym->start;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-05-28 19:55:41 +02:00
|
|
|
|
node = rb_last(&kernel_dso->syms);
|
|
|
|
|
sym = rb_entry(node, struct symbol, rb_node);
|
2009-05-29 08:23:16 +02:00
|
|
|
|
max_ip = sym->end;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-05-28 19:55:41 +02:00
|
|
|
|
if (dump_symtab)
|
2009-05-29 06:46:46 +02:00
|
|
|
|
dso__fprintf(kernel_dso, stderr);
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-05-28 19:55:41 +02:00
|
|
|
|
return 0;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-05-28 19:55:41 +02:00
|
|
|
|
out_delete_dso:
|
|
|
|
|
dso__delete(kernel_dso);
|
|
|
|
|
kernel_dso = NULL;
|
|
|
|
|
return -1;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#define TRACE_COUNT 3
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Binary search in the histogram table and record the hit:
|
|
|
|
|
*/
|
|
|
|
|
static void record_ip(uint64_t ip, int counter)
|
|
|
|
|
{
|
2009-05-28 19:55:41 +02:00
|
|
|
|
struct symbol *sym = dso__find_symbol(kernel_dso, ip);
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-05-28 19:55:41 +02:00
|
|
|
|
if (sym != NULL) {
|
|
|
|
|
struct sym_entry *syme = dso__sym_priv(kernel_dso, sym);
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-05-28 19:55:41 +02:00
|
|
|
|
if (!syme->skip) {
|
|
|
|
|
syme->count[counter]++;
|
2009-05-29 22:03:07 +02:00
|
|
|
|
pthread_mutex_lock(&active_symbols_lock);
|
2009-05-28 19:55:41 +02:00
|
|
|
|
if (list_empty(&syme->node) || !syme->node.next)
|
2009-05-29 22:03:07 +02:00
|
|
|
|
__list_insert_active_sym(syme);
|
|
|
|
|
pthread_mutex_unlock(&active_symbols_lock);
|
2009-05-28 19:55:41 +02:00
|
|
|
|
return;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2009-06-05 14:29:10 +02:00
|
|
|
|
samples--;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void process_event(uint64_t ip, int counter)
|
|
|
|
|
{
|
2009-06-05 14:29:10 +02:00
|
|
|
|
samples++;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
|
|
|
|
if (ip < min_ip || ip > max_ip) {
|
2009-06-05 14:29:10 +02:00
|
|
|
|
userspace_samples++;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
record_ip(ip, counter);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
struct mmap_data {
|
2009-06-06 09:58:57 +02:00
|
|
|
|
int counter;
|
|
|
|
|
void *base;
|
|
|
|
|
unsigned int mask;
|
|
|
|
|
unsigned int prev;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static unsigned int mmap_read_head(struct mmap_data *md)
|
|
|
|
|
{
|
|
|
|
|
struct perf_counter_mmap_page *pc = md->base;
|
|
|
|
|
int head;
|
|
|
|
|
|
|
|
|
|
head = pc->data_head;
|
|
|
|
|
rmb();
|
|
|
|
|
|
|
|
|
|
return head;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
struct timeval last_read, this_read;
|
|
|
|
|
|
|
|
|
|
static void mmap_read(struct mmap_data *md)
|
|
|
|
|
{
|
|
|
|
|
unsigned int head = mmap_read_head(md);
|
|
|
|
|
unsigned int old = md->prev;
|
|
|
|
|
unsigned char *data = md->base + page_size;
|
|
|
|
|
int diff;
|
|
|
|
|
|
|
|
|
|
gettimeofday(&this_read, NULL);
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* If we're further behind than half the buffer, there's a chance
|
2009-06-05 14:29:10 +02:00
|
|
|
|
* the writer will bite our tail and mess up the samples under us.
|
2009-04-20 15:00:56 +02:00
|
|
|
|
*
|
|
|
|
|
* If we somehow ended up ahead of the head, we got messed up.
|
|
|
|
|
*
|
|
|
|
|
* In either case, truncate and restart at head.
|
|
|
|
|
*/
|
|
|
|
|
diff = head - old;
|
|
|
|
|
if (diff > md->mask / 2 || diff < 0) {
|
|
|
|
|
struct timeval iv;
|
|
|
|
|
unsigned long msecs;
|
|
|
|
|
|
|
|
|
|
timersub(&this_read, &last_read, &iv);
|
|
|
|
|
msecs = iv.tv_sec*1000 + iv.tv_usec/1000;
|
|
|
|
|
|
|
|
|
|
fprintf(stderr, "WARNING: failed to keep up with mmap data."
|
|
|
|
|
" Last read %lu msecs ago.\n", msecs);
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* head points to a known good entry, start there.
|
|
|
|
|
*/
|
|
|
|
|
old = head;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
last_read = this_read;
|
|
|
|
|
|
|
|
|
|
for (; old != head;) {
|
|
|
|
|
struct ip_event {
|
|
|
|
|
struct perf_event_header header;
|
|
|
|
|
__u64 ip;
|
2009-05-26 09:17:18 +02:00
|
|
|
|
__u32 pid, target_pid;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
};
|
|
|
|
|
struct mmap_event {
|
|
|
|
|
struct perf_event_header header;
|
2009-05-26 09:17:18 +02:00
|
|
|
|
__u32 pid, target_pid;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
__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;
|
|
|
|
|
|
2009-04-20 15:22:22 +02:00
|
|
|
|
size_t size = event->header.size;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Event straddles the mmap boundary -- header should always
|
|
|
|
|
* be inside due to u64 alignment of output.
|
|
|
|
|
*/
|
|
|
|
|
if ((old & md->mask) + size != ((old + size) & md->mask)) {
|
|
|
|
|
unsigned int offset = old;
|
|
|
|
|
unsigned int len = min(sizeof(*event), size), cpy;
|
|
|
|
|
void *dst = &event_copy;
|
|
|
|
|
|
|
|
|
|
do {
|
|
|
|
|
cpy = min(md->mask + 1 - (offset & md->mask), len);
|
|
|
|
|
memcpy(dst, &data[offset & md->mask], cpy);
|
|
|
|
|
offset += cpy;
|
|
|
|
|
dst += cpy;
|
|
|
|
|
len -= cpy;
|
|
|
|
|
} while (len);
|
|
|
|
|
|
|
|
|
|
event = &event_copy;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
old += size;
|
|
|
|
|
|
|
|
|
|
if (event->header.misc & PERF_EVENT_MISC_OVERFLOW) {
|
2009-06-02 17:38:21 +02:00
|
|
|
|
if (event->header.type & PERF_SAMPLE_IP)
|
2009-04-20 15:00:56 +02:00
|
|
|
|
process_event(event->ip.ip, md->counter);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
md->prev = old;
|
|
|
|
|
}
|
|
|
|
|
|
2009-05-24 08:35:49 +02:00
|
|
|
|
static struct pollfd event_array[MAX_NR_CPUS * MAX_COUNTERS];
|
|
|
|
|
static struct mmap_data mmap_array[MAX_NR_CPUS][MAX_COUNTERS];
|
|
|
|
|
|
2009-05-26 09:17:18 +02:00
|
|
|
|
static int __cmd_top(void)
|
2009-04-20 15:00:56 +02:00
|
|
|
|
{
|
2009-06-06 09:58:57 +02:00
|
|
|
|
struct perf_counter_attr *attr;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
pthread_t thread;
|
|
|
|
|
int i, counter, group_fd, nr_poll = 0;
|
|
|
|
|
unsigned int cpu;
|
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < nr_cpus; i++) {
|
|
|
|
|
group_fd = -1;
|
|
|
|
|
for (counter = 0; counter < nr_counters; counter++) {
|
|
|
|
|
|
|
|
|
|
cpu = profile_cpu;
|
2009-05-26 09:17:18 +02:00
|
|
|
|
if (target_pid == -1 && profile_cpu == -1)
|
2009-04-20 15:00:56 +02:00
|
|
|
|
cpu = i;
|
|
|
|
|
|
2009-06-06 09:58:57 +02:00
|
|
|
|
attr = attrs + counter;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
2009-06-06 09:58:57 +02:00
|
|
|
|
attr->sample_type = PERF_SAMPLE_IP | PERF_SAMPLE_TID;
|
|
|
|
|
attr->freq = freq;
|
|
|
|
|
|
|
|
|
|
fd[i][counter] = sys_perf_counter_open(attr, target_pid, cpu, group_fd, 0);
|
2009-04-20 15:00:56 +02:00
|
|
|
|
if (fd[i][counter] < 0) {
|
|
|
|
|
int err = errno;
|
2009-06-03 19:17:25 +02:00
|
|
|
|
|
|
|
|
|
error("syscall returned with %d (%s)\n",
|
2009-04-20 15:00:56 +02:00
|
|
|
|
fd[i][counter], strerror(err));
|
|
|
|
|
if (err == EPERM)
|
|
|
|
|
printf("Are you root?\n");
|
|
|
|
|
exit(-1);
|
|
|
|
|
}
|
|
|
|
|
assert(fd[i][counter] >= 0);
|
|
|
|
|
fcntl(fd[i][counter], F_SETFL, O_NONBLOCK);
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* First counter acts as the group leader:
|
|
|
|
|
*/
|
|
|
|
|
if (group && group_fd == -1)
|
|
|
|
|
group_fd = fd[i][counter];
|
|
|
|
|
|
|
|
|
|
event_array[nr_poll].fd = fd[i][counter];
|
|
|
|
|
event_array[nr_poll].events = POLLIN;
|
|
|
|
|
nr_poll++;
|
|
|
|
|
|
|
|
|
|
mmap_array[i][counter].counter = counter;
|
|
|
|
|
mmap_array[i][counter].prev = 0;
|
|
|
|
|
mmap_array[i][counter].mask = mmap_pages*page_size - 1;
|
|
|
|
|
mmap_array[i][counter].base = mmap(NULL, (mmap_pages+1)*page_size,
|
|
|
|
|
PROT_READ, MAP_SHARED, fd[i][counter], 0);
|
2009-06-03 19:17:25 +02:00
|
|
|
|
if (mmap_array[i][counter].base == MAP_FAILED)
|
|
|
|
|
die("failed to mmap with %d (%s)\n", errno, strerror(errno));
|
2009-04-20 15:00:56 +02:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (pthread_create(&thread, NULL, display_thread, NULL)) {
|
|
|
|
|
printf("Could not create display thread.\n");
|
|
|
|
|
exit(-1);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (realtime_prio) {
|
|
|
|
|
struct sched_param param;
|
|
|
|
|
|
|
|
|
|
param.sched_priority = realtime_prio;
|
|
|
|
|
if (sched_setscheduler(0, SCHED_FIFO, ¶m)) {
|
|
|
|
|
printf("Could not set realtime priority.\n");
|
|
|
|
|
exit(-1);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
while (1) {
|
2009-06-05 14:29:10 +02:00
|
|
|
|
int hits = samples;
|
2009-04-20 15:00:56 +02:00
|
|
|
|
|
|
|
|
|
for (i = 0; i < nr_cpus; i++) {
|
|
|
|
|
for (counter = 0; counter < nr_counters; counter++)
|
|
|
|
|
mmap_read(&mmap_array[i][counter]);
|
|
|
|
|
}
|
|
|
|
|
|
2009-06-05 14:29:10 +02:00
|
|
|
|
if (hits == samples)
|
2009-04-20 15:00:56 +02:00
|
|
|
|
ret = poll(event_array, nr_poll, 100);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
2009-05-26 09:17:18 +02:00
|
|
|
|
|
|
|
|
|
static const char * const top_usage[] = {
|
|
|
|
|
"perf top [<options>]",
|
|
|
|
|
NULL
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static const struct option options[] = {
|
|
|
|
|
OPT_CALLBACK('e', "event", NULL, "event",
|
2009-06-06 12:24:17 +02:00
|
|
|
|
"event selector. use 'perf list' to list available events",
|
|
|
|
|
parse_events),
|
2009-05-26 09:17:18 +02:00
|
|
|
|
OPT_INTEGER('c', "count", &default_interval,
|
|
|
|
|
"event period to sample"),
|
|
|
|
|
OPT_INTEGER('p', "pid", &target_pid,
|
|
|
|
|
"profile events on existing pid"),
|
|
|
|
|
OPT_BOOLEAN('a', "all-cpus", &system_wide,
|
|
|
|
|
"system-wide collection from all CPUs"),
|
|
|
|
|
OPT_INTEGER('C', "CPU", &profile_cpu,
|
|
|
|
|
"CPU to profile on"),
|
|
|
|
|
OPT_INTEGER('m', "mmap-pages", &mmap_pages,
|
|
|
|
|
"number of mmap data pages"),
|
|
|
|
|
OPT_INTEGER('r', "realtime", &realtime_prio,
|
|
|
|
|
"collect data with this RT SCHED_FIFO priority"),
|
2009-05-26 15:25:34 +02:00
|
|
|
|
OPT_INTEGER('d', "delay", &delay_secs,
|
2009-05-26 09:17:18 +02:00
|
|
|
|
"number of seconds to delay between refreshes"),
|
|
|
|
|
OPT_BOOLEAN('D', "dump-symtab", &dump_symtab,
|
|
|
|
|
"dump the symbol table used for profiling"),
|
2009-06-04 08:53:05 +02:00
|
|
|
|
OPT_INTEGER('f', "count-filter", &count_filter,
|
2009-05-26 09:17:18 +02:00
|
|
|
|
"only display functions with more events than this"),
|
|
|
|
|
OPT_BOOLEAN('g', "group", &group,
|
|
|
|
|
"put the counters into a counter group"),
|
|
|
|
|
OPT_STRING('s', "sym-filter", &sym_filter, "pattern",
|
|
|
|
|
"only display symbols matchig this pattern"),
|
|
|
|
|
OPT_BOOLEAN('z', "zero", &group,
|
|
|
|
|
"zero history across updates"),
|
2009-06-04 08:53:05 +02:00
|
|
|
|
OPT_INTEGER('F', "freq", &freq,
|
2009-05-26 09:17:18 +02:00
|
|
|
|
"profile at this frequency"),
|
2009-06-04 08:53:05 +02:00
|
|
|
|
OPT_INTEGER('E', "entries", &print_entries,
|
|
|
|
|
"display this many functions"),
|
2009-05-26 09:17:18 +02:00
|
|
|
|
OPT_END()
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
int cmd_top(int argc, const char **argv, const char *prefix)
|
|
|
|
|
{
|
|
|
|
|
int counter;
|
|
|
|
|
|
|
|
|
|
page_size = sysconf(_SC_PAGE_SIZE);
|
|
|
|
|
|
|
|
|
|
argc = parse_options(argc, argv, options, top_usage, 0);
|
|
|
|
|
if (argc)
|
|
|
|
|
usage_with_options(top_usage, options);
|
|
|
|
|
|
|
|
|
|
if (freq) {
|
|
|
|
|
default_interval = freq;
|
|
|
|
|
freq = 1;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* CPU and PID are mutually exclusive */
|
|
|
|
|
if (target_pid != -1 && profile_cpu != -1) {
|
|
|
|
|
printf("WARNING: PID switch overriding CPU\n");
|
|
|
|
|
sleep(1);
|
|
|
|
|
profile_cpu = -1;
|
|
|
|
|
}
|
|
|
|
|
|
2009-06-06 09:58:57 +02:00
|
|
|
|
if (!nr_counters)
|
2009-05-26 09:17:18 +02:00
|
|
|
|
nr_counters = 1;
|
|
|
|
|
|
2009-06-05 19:31:01 +02:00
|
|
|
|
if (delay_secs < 1)
|
|
|
|
|
delay_secs = 1;
|
|
|
|
|
|
2009-06-06 09:58:57 +02:00
|
|
|
|
parse_symbols();
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Fill in the ones not specifically initialized via -c:
|
|
|
|
|
*/
|
2009-05-26 09:17:18 +02:00
|
|
|
|
for (counter = 0; counter < nr_counters; counter++) {
|
2009-06-06 09:58:57 +02:00
|
|
|
|
if (attrs[counter].sample_period)
|
2009-05-26 09:17:18 +02:00
|
|
|
|
continue;
|
|
|
|
|
|
2009-06-06 09:58:57 +02:00
|
|
|
|
attrs[counter].sample_period = default_interval;
|
2009-05-26 09:17:18 +02:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
nr_cpus = sysconf(_SC_NPROCESSORS_ONLN);
|
|
|
|
|
assert(nr_cpus <= MAX_NR_CPUS);
|
|
|
|
|
assert(nr_cpus >= 0);
|
|
|
|
|
|
|
|
|
|
if (target_pid != -1 || profile_cpu != -1)
|
|
|
|
|
nr_cpus = 1;
|
|
|
|
|
|
|
|
|
|
return __cmd_top();
|
|
|
|
|
}
|