linux/tools/perf/builtin-sched.c

1918 lines
43 KiB
C
Raw Normal View History

#include "builtin.h"
#include "perf.h"
#include "util/util.h"
#include "util/cache.h"
#include "util/symbol.h"
#include "util/thread.h"
#include "util/header.h"
#include "util/session.h"
#include "util/parse-options.h"
#include "util/trace-event.h"
#include "util/debug.h"
#include <sys/prctl.h>
#include <semaphore.h>
#include <pthread.h>
#include <math.h>
static char const *input_name = "perf.data";
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
static char default_sort_order[] = "avg, max, switch, runtime";
static const char *sort_order = default_sort_order;
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
static int profile_cpu = -1;
#define PR_SET_NAME 15 /* Set process name */
#define MAX_CPUS 4096
static u64 run_measurement_overhead;
static u64 sleep_measurement_overhead;
#define COMM_LEN 20
#define SYM_LEN 129
#define MAX_PID 65536
static unsigned long nr_tasks;
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
struct sched_atom;
struct task_desc {
unsigned long nr;
unsigned long pid;
char comm[COMM_LEN];
unsigned long nr_events;
unsigned long curr_event;
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
struct sched_atom **atoms;
pthread_t thread;
sem_t sleep_sem;
sem_t ready_for_work;
sem_t work_done_sem;
u64 cpu_usage;
};
enum sched_event_type {
SCHED_EVENT_RUN,
SCHED_EVENT_SLEEP,
SCHED_EVENT_WAKEUP,
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
SCHED_EVENT_MIGRATION,
};
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
struct sched_atom {
enum sched_event_type type;
int specific_wait;
u64 timestamp;
u64 duration;
unsigned long nr;
sem_t *wait_sem;
struct task_desc *wakee;
};
static struct task_desc *pid_to_task[MAX_PID];
static struct task_desc **tasks;
static pthread_mutex_t start_work_mutex = PTHREAD_MUTEX_INITIALIZER;
static u64 start_time;
static pthread_mutex_t work_done_wait_mutex = PTHREAD_MUTEX_INITIALIZER;
static unsigned long nr_run_events;
static unsigned long nr_sleep_events;
static unsigned long nr_wakeup_events;
static unsigned long nr_sleep_corrections;
static unsigned long nr_run_events_optimized;
static unsigned long targetless_wakeups;
static unsigned long multitarget_wakeups;
static u64 cpu_usage;
static u64 runavg_cpu_usage;
static u64 parent_cpu_usage;
static u64 runavg_parent_cpu_usage;
static unsigned long nr_runs;
static u64 sum_runtime;
static u64 sum_fluct;
static u64 run_avg;
static unsigned int replay_repeat = 10;
perf tools: Implement counter output multiplexing Finish the -M/--multiplex option implementation: - separate it out from group_fd - correctly set it via the ioctl and dont mmap counters that are multiplexed - modify the perf record event loop to deal with buffer-less counters. - remove the -g option from perf sched record - account for unordered events in perf sched latency - (add -f to perf sched record to ease measurements) - skip idle threads (pid==0) in latency output The result is better latency output by 'perf sched latency': ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- ksoftirqd/8 | 0.071 ms | 2 | avg: 0.458 ms | max: 0.913 ms | at-spi-registry | 0.609 ms | 19 | avg: 0.013 ms | max: 0.023 ms | perf | 3.316 ms | 16 | avg: 0.013 ms | max: 0.054 ms | Xorg | 0.392 ms | 19 | avg: 0.011 ms | max: 0.018 ms | sleep | 0.537 ms | 2 | avg: 0.009 ms | max: 0.009 ms | ----------------------------------------------------------------------------------- TOTAL: | 4.925 ms | 58 | --------------------------------------------- Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 16:15:54 +00:00
static unsigned long nr_timestamps;
static unsigned long nr_unordered_timestamps;
static unsigned long nr_state_machine_bugs;
static unsigned long nr_context_switch_bugs;
static unsigned long nr_events;
static unsigned long nr_lost_chunks;
static unsigned long nr_lost_events;
#define TASK_STATE_TO_CHAR_STR "RSDTtZX"
enum thread_state {
THREAD_SLEEPING = 0,
THREAD_WAIT_CPU,
THREAD_SCHED_IN,
THREAD_IGNORE
};
struct work_atom {
struct list_head list;
enum thread_state state;
perf tools: Fix processing of randomly serialized sched traces Currently it's possible to meet such too high latency results with 'perf sched latency'. ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- xfce4-panel | 0.222 ms | 2 | avg: 4718.345 ms | max: 9436.493 ms | scsi_eh_3 | 3.962 ms | 36 | avg: 55.957 ms | max: 1977.829 ms | The origin is on traces that are sometimes badly serialized across cpus. For example the raw traces that raised such results for xfce4-panel: (1) [init]-0 [000] 1494.663899990: sched_switch: task swapper:0 [140] (R) ==> xfce4-panel:4569 [120] (2) xfce4-panel-4569 [000] 1494.663928373: sched_switch: task xfce4-panel:4569 [120] (S) ==> swapper:0 [140] (3) Xorg-4276 [001] 1494.663860125: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (4) Xorg-4276 [001] 1504.098252756: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (5) perf-5219 [000] 1504.100353302: sched_switch: task perf:5219 [120] (S) ==> xfce4-panel:4569 [120] The traces are processed in the order they arrive. Then in (2), xfce4-panel sleeps, it is first waken up in (3) and eventually scheduled in (5). The latency reported is then 1504 - 1495 = 9 secs, as reported by perf sched. But this is wrong, we are confident in the fact the traces are nicely serialized while we should actually more trust the timestamps. If we reorder by timestamps we get: (1) Xorg-4276 [001] 1494.663860125: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (2) [init]-0 [000] 1494.663899990: sched_switch: task swapper:0 [140] (R) ==> xfce4-panel:4569 [120] (3) xfce4-panel-4569 [000] 1494.663928373: sched_switch: task xfce4-panel:4569 [120] (S) ==> swapper:0 [140] (4) Xorg-4276 [001] 1504.098252756: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (5) perf-5219 [000] 1504.100353302: sched_switch: task perf:5219 [120] (S) ==> xfce4-panel:4569 [120] Now the trace make more sense, xfce4-panel is sleeping. Then it is woken up in (1), scheduled in (2) It goes to sleep in (3), woken up in (4) and scheduled in (5). Now, latency captured between (1) and (2) is of 39 us. And between (4) and (5) it is 2.1 ms. Such pattern of bad serializing is the origin of the high latencies reported by perf sched. Basically, we need to check whether wake up time is higher than schedule out time. If it's not the case, we need to tag the current work atom as invalid. Beside that, we may need to work later on a better ordering of the traces given by the kernel. After this patch: xfce4-session | 0.221 ms | 1 | avg: 0.538 ms | max: 0.538 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 01:01:12 +00:00
u64 sched_out_time;
u64 wake_up_time;
u64 sched_in_time;
u64 runtime;
};
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
struct work_atoms {
struct list_head work_list;
struct thread *thread;
struct rb_node node;
u64 max_lat;
u64 max_lat_at;
u64 total_lat;
u64 nb_atoms;
u64 total_runtime;
};
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
typedef int (*sort_fn_t)(struct work_atoms *, struct work_atoms *);
static struct rb_root atom_root, sorted_atom_root;
static u64 all_runtime;
static u64 all_count;
static u64 get_nsecs(void)
{
struct timespec ts;
clock_gettime(CLOCK_MONOTONIC, &ts);
return ts.tv_sec * 1000000000ULL + ts.tv_nsec;
}
static void burn_nsecs(u64 nsecs)
{
u64 T0 = get_nsecs(), T1;
do {
T1 = get_nsecs();
} while (T1 + run_measurement_overhead < T0 + nsecs);
}
static void sleep_nsecs(u64 nsecs)
{
struct timespec ts;
ts.tv_nsec = nsecs % 999999999;
ts.tv_sec = nsecs / 999999999;
nanosleep(&ts, NULL);
}
static void calibrate_run_measurement_overhead(void)
{
u64 T0, T1, delta, min_delta = 1000000000ULL;
int i;
for (i = 0; i < 10; i++) {
T0 = get_nsecs();
burn_nsecs(0);
T1 = get_nsecs();
delta = T1-T0;
min_delta = min(min_delta, delta);
}
run_measurement_overhead = min_delta;
printf("run measurement overhead: %" PRIu64 " nsecs\n", min_delta);
}
static void calibrate_sleep_measurement_overhead(void)
{
u64 T0, T1, delta, min_delta = 1000000000ULL;
int i;
for (i = 0; i < 10; i++) {
T0 = get_nsecs();
sleep_nsecs(10000);
T1 = get_nsecs();
delta = T1-T0;
min_delta = min(min_delta, delta);
}
min_delta -= 10000;
sleep_measurement_overhead = min_delta;
printf("sleep measurement overhead: %" PRIu64 " nsecs\n", min_delta);
}
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
static struct sched_atom *
get_new_event(struct task_desc *task, u64 timestamp)
{
struct sched_atom *event = zalloc(sizeof(*event));
unsigned long idx = task->nr_events;
size_t size;
event->timestamp = timestamp;
event->nr = idx;
task->nr_events++;
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
size = sizeof(struct sched_atom *) * task->nr_events;
task->atoms = realloc(task->atoms, size);
BUG_ON(!task->atoms);
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
task->atoms[idx] = event;
return event;
}
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
static struct sched_atom *last_event(struct task_desc *task)
{
if (!task->nr_events)
return NULL;
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
return task->atoms[task->nr_events - 1];
}
static void
add_sched_event_run(struct task_desc *task, u64 timestamp, u64 duration)
{
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
struct sched_atom *event, *curr_event = last_event(task);
/*
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
* optimize an existing RUN event by merging this one
* to it:
*/
if (curr_event && curr_event->type == SCHED_EVENT_RUN) {
nr_run_events_optimized++;
curr_event->duration += duration;
return;
}
event = get_new_event(task, timestamp);
event->type = SCHED_EVENT_RUN;
event->duration = duration;
nr_run_events++;
}
static void
add_sched_event_wakeup(struct task_desc *task, u64 timestamp,
struct task_desc *wakee)
{
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
struct sched_atom *event, *wakee_event;
event = get_new_event(task, timestamp);
event->type = SCHED_EVENT_WAKEUP;
event->wakee = wakee;
wakee_event = last_event(wakee);
if (!wakee_event || wakee_event->type != SCHED_EVENT_SLEEP) {
targetless_wakeups++;
return;
}
if (wakee_event->wait_sem) {
multitarget_wakeups++;
return;
}
wakee_event->wait_sem = zalloc(sizeof(*wakee_event->wait_sem));
sem_init(wakee_event->wait_sem, 0, 0);
wakee_event->specific_wait = 1;
event->wait_sem = wakee_event->wait_sem;
nr_wakeup_events++;
}
static void
add_sched_event_sleep(struct task_desc *task, u64 timestamp,
u64 task_state __used)
{
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
struct sched_atom *event = get_new_event(task, timestamp);
event->type = SCHED_EVENT_SLEEP;
nr_sleep_events++;
}
static struct task_desc *register_pid(unsigned long pid, const char *comm)
{
struct task_desc *task;
BUG_ON(pid >= MAX_PID);
task = pid_to_task[pid];
if (task)
return task;
task = zalloc(sizeof(*task));
task->pid = pid;
task->nr = nr_tasks;
strcpy(task->comm, comm);
/*
* every task starts in sleeping state - this gets ignored
* if there's no wakeup pointing to this sleep state:
*/
add_sched_event_sleep(task, 0, 0);
pid_to_task[pid] = task;
nr_tasks++;
tasks = realloc(tasks, nr_tasks*sizeof(struct task_task *));
BUG_ON(!tasks);
tasks[task->nr] = task;
if (verbose)
printf("registered task #%ld, PID %ld (%s)\n", nr_tasks, pid, comm);
return task;
}
static void print_task_traces(void)
{
struct task_desc *task;
unsigned long i;
for (i = 0; i < nr_tasks; i++) {
task = tasks[i];
printf("task %6ld (%20s:%10ld), nr_events: %ld\n",
task->nr, task->comm, task->pid, task->nr_events);
}
}
static void add_cross_task_wakeups(void)
{
struct task_desc *task1, *task2;
unsigned long i, j;
for (i = 0; i < nr_tasks; i++) {
task1 = tasks[i];
j = i + 1;
if (j == nr_tasks)
j = 0;
task2 = tasks[j];
add_sched_event_wakeup(task1, 0, task2);
}
}
static void
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
process_sched_event(struct task_desc *this_task __used, struct sched_atom *atom)
{
int ret = 0;
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
switch (atom->type) {
case SCHED_EVENT_RUN:
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
burn_nsecs(atom->duration);
break;
case SCHED_EVENT_SLEEP:
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
if (atom->wait_sem)
ret = sem_wait(atom->wait_sem);
BUG_ON(ret);
break;
case SCHED_EVENT_WAKEUP:
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
if (atom->wait_sem)
ret = sem_post(atom->wait_sem);
BUG_ON(ret);
break;
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
case SCHED_EVENT_MIGRATION:
break;
default:
BUG_ON(1);
}
}
static u64 get_cpu_usage_nsec_parent(void)
{
struct rusage ru;
u64 sum;
int err;
err = getrusage(RUSAGE_SELF, &ru);
BUG_ON(err);
sum = ru.ru_utime.tv_sec*1e9 + ru.ru_utime.tv_usec*1e3;
sum += ru.ru_stime.tv_sec*1e9 + ru.ru_stime.tv_usec*1e3;
return sum;
}
static int self_open_counters(void)
{
struct perf_event_attr attr;
int fd;
memset(&attr, 0, sizeof(attr));
attr.type = PERF_TYPE_SOFTWARE;
attr.config = PERF_COUNT_SW_TASK_CLOCK;
fd = sys_perf_event_open(&attr, 0, -1, -1, 0);
if (fd < 0)
die("Error: sys_perf_event_open() syscall returned"
"with %d (%s)\n", fd, strerror(errno));
return fd;
}
static u64 get_cpu_usage_nsec_self(int fd)
{
u64 runtime;
int ret;
ret = read(fd, &runtime, sizeof(runtime));
BUG_ON(ret != sizeof(runtime));
return runtime;
}
static void *thread_func(void *ctx)
{
struct task_desc *this_task = ctx;
u64 cpu_usage_0, cpu_usage_1;
unsigned long i, ret;
char comm2[22];
int fd;
sprintf(comm2, ":%s", this_task->comm);
prctl(PR_SET_NAME, comm2);
fd = self_open_counters();
again:
ret = sem_post(&this_task->ready_for_work);
BUG_ON(ret);
ret = pthread_mutex_lock(&start_work_mutex);
BUG_ON(ret);
ret = pthread_mutex_unlock(&start_work_mutex);
BUG_ON(ret);
cpu_usage_0 = get_cpu_usage_nsec_self(fd);
for (i = 0; i < this_task->nr_events; i++) {
this_task->curr_event = i;
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
process_sched_event(this_task, this_task->atoms[i]);
}
cpu_usage_1 = get_cpu_usage_nsec_self(fd);
this_task->cpu_usage = cpu_usage_1 - cpu_usage_0;
ret = sem_post(&this_task->work_done_sem);
BUG_ON(ret);
ret = pthread_mutex_lock(&work_done_wait_mutex);
BUG_ON(ret);
ret = pthread_mutex_unlock(&work_done_wait_mutex);
BUG_ON(ret);
goto again;
}
static void create_tasks(void)
{
struct task_desc *task;
pthread_attr_t attr;
unsigned long i;
int err;
err = pthread_attr_init(&attr);
BUG_ON(err);
err = pthread_attr_setstacksize(&attr,
(size_t) max(16 * 1024, PTHREAD_STACK_MIN));
BUG_ON(err);
err = pthread_mutex_lock(&start_work_mutex);
BUG_ON(err);
err = pthread_mutex_lock(&work_done_wait_mutex);
BUG_ON(err);
for (i = 0; i < nr_tasks; i++) {
task = tasks[i];
sem_init(&task->sleep_sem, 0, 0);
sem_init(&task->ready_for_work, 0, 0);
sem_init(&task->work_done_sem, 0, 0);
task->curr_event = 0;
err = pthread_create(&task->thread, &attr, thread_func, task);
BUG_ON(err);
}
}
static void wait_for_tasks(void)
{
u64 cpu_usage_0, cpu_usage_1;
struct task_desc *task;
unsigned long i, ret;
start_time = get_nsecs();
cpu_usage = 0;
pthread_mutex_unlock(&work_done_wait_mutex);
for (i = 0; i < nr_tasks; i++) {
task = tasks[i];
ret = sem_wait(&task->ready_for_work);
BUG_ON(ret);
sem_init(&task->ready_for_work, 0, 0);
}
ret = pthread_mutex_lock(&work_done_wait_mutex);
BUG_ON(ret);
cpu_usage_0 = get_cpu_usage_nsec_parent();
pthread_mutex_unlock(&start_work_mutex);
for (i = 0; i < nr_tasks; i++) {
task = tasks[i];
ret = sem_wait(&task->work_done_sem);
BUG_ON(ret);
sem_init(&task->work_done_sem, 0, 0);
cpu_usage += task->cpu_usage;
task->cpu_usage = 0;
}
cpu_usage_1 = get_cpu_usage_nsec_parent();
if (!runavg_cpu_usage)
runavg_cpu_usage = cpu_usage;
runavg_cpu_usage = (runavg_cpu_usage*9 + cpu_usage)/10;
parent_cpu_usage = cpu_usage_1 - cpu_usage_0;
if (!runavg_parent_cpu_usage)
runavg_parent_cpu_usage = parent_cpu_usage;
runavg_parent_cpu_usage = (runavg_parent_cpu_usage*9 +
parent_cpu_usage)/10;
ret = pthread_mutex_lock(&start_work_mutex);
BUG_ON(ret);
for (i = 0; i < nr_tasks; i++) {
task = tasks[i];
sem_init(&task->sleep_sem, 0, 0);
task->curr_event = 0;
}
}
static void run_one_test(void)
{
u64 T0, T1, delta, avg_delta, fluct;
T0 = get_nsecs();
wait_for_tasks();
T1 = get_nsecs();
delta = T1 - T0;
sum_runtime += delta;
nr_runs++;
avg_delta = sum_runtime / nr_runs;
if (delta < avg_delta)
fluct = avg_delta - delta;
else
fluct = delta - avg_delta;
sum_fluct += fluct;
if (!run_avg)
run_avg = delta;
run_avg = (run_avg*9 + delta)/10;
printf("#%-3ld: %0.3f, ",
nr_runs, (double)delta/1000000.0);
printf("ravg: %0.2f, ",
(double)run_avg/1e6);
printf("cpu: %0.2f / %0.2f",
(double)cpu_usage/1e6, (double)runavg_cpu_usage/1e6);
#if 0
/*
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
* rusage statistics done by the parent, these are less
* accurate than the sum_exec_runtime based statistics:
*/
printf(" [%0.2f / %0.2f]",
(double)parent_cpu_usage/1e6,
(double)runavg_parent_cpu_usage/1e6);
#endif
printf("\n");
if (nr_sleep_corrections)
printf(" (%ld sleep corrections)\n", nr_sleep_corrections);
nr_sleep_corrections = 0;
}
static void test_calibrations(void)
{
u64 T0, T1;
T0 = get_nsecs();
burn_nsecs(1e6);
T1 = get_nsecs();
printf("the run test took %" PRIu64 " nsecs\n", T1 - T0);
T0 = get_nsecs();
sleep_nsecs(1e6);
T1 = get_nsecs();
printf("the sleep test took %" PRIu64 " nsecs\n", T1 - T0);
}
perf sched: Fix bad event alignment perf sched raises the following error when it meets a sched switch event: perf: builtin-sched.c:286: register_pid: Assertion `!(pid >= 65536)' failed. Abandon Currently in x86-64, the sched switch events have a hole in the middle of the structure: u16 common_type; u8 common_flags; u8 common_preempt_count; u32 common_pid; u32 common_tgid; char prev_comm[16]; u32 prev_pid; u32 prev_prio; <--- there u64 prev_state; char next_comm[16]; u32 next_pid; u32 next_prio; Gcc inserts a 4 bytes hole there for prev_state to be u64 aligned. And the events are exported to userspace with this hole. But in userspace, from perf sched, we fetch it using a structure that has a new field in the beginning: u32 size. This is because our trace is exported with its size as a field. But now that we have this new field, the hole in the middle disappears because it makes prev_state becoming well aligned. And since we are using a pointer to the raw trace using this struct, instead of reading prev_state, we are reading the hole. We could fix it by keeping the size seperate from the struct but actually there a lot of other potential problems: some fields may be saved as long in a 64 bits system and later read as long in a 32 bits system. Also this direct cast doesn't care about the endianness differences between the host traced machine and the machine in which we do the post processing. So instead of using such dangerous direct casts, fetch the values using the trace parsing API that already takes care of all these problems. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-12 00:43:45 +00:00
#define FILL_FIELD(ptr, field, event, data) \
ptr.field = (typeof(ptr.field)) raw_field_value(event, #field, data)
#define FILL_ARRAY(ptr, array, event, data) \
do { \
void *__array = raw_field_ptr(event, #array, data); \
memcpy(ptr.array, __array, sizeof(ptr.array)); \
} while(0)
#define FILL_COMMON_FIELDS(ptr, event, data) \
do { \
FILL_FIELD(ptr, common_type, event, data); \
FILL_FIELD(ptr, common_flags, event, data); \
FILL_FIELD(ptr, common_preempt_count, event, data); \
FILL_FIELD(ptr, common_pid, event, data); \
FILL_FIELD(ptr, common_tgid, event, data); \
} while (0)
struct trace_switch_event {
u32 size;
u16 common_type;
u8 common_flags;
u8 common_preempt_count;
u32 common_pid;
u32 common_tgid;
char prev_comm[16];
u32 prev_pid;
u32 prev_prio;
u64 prev_state;
char next_comm[16];
u32 next_pid;
u32 next_prio;
};
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
struct trace_runtime_event {
u32 size;
u16 common_type;
u8 common_flags;
u8 common_preempt_count;
u32 common_pid;
u32 common_tgid;
char comm[16];
u32 pid;
u64 runtime;
u64 vruntime;
};
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
struct trace_wakeup_event {
u32 size;
u16 common_type;
u8 common_flags;
u8 common_preempt_count;
u32 common_pid;
u32 common_tgid;
char comm[16];
u32 pid;
u32 prio;
u32 success;
u32 cpu;
};
struct trace_fork_event {
u32 size;
perf sched: Fix bad event alignment perf sched raises the following error when it meets a sched switch event: perf: builtin-sched.c:286: register_pid: Assertion `!(pid >= 65536)' failed. Abandon Currently in x86-64, the sched switch events have a hole in the middle of the structure: u16 common_type; u8 common_flags; u8 common_preempt_count; u32 common_pid; u32 common_tgid; char prev_comm[16]; u32 prev_pid; u32 prev_prio; <--- there u64 prev_state; char next_comm[16]; u32 next_pid; u32 next_prio; Gcc inserts a 4 bytes hole there for prev_state to be u64 aligned. And the events are exported to userspace with this hole. But in userspace, from perf sched, we fetch it using a structure that has a new field in the beginning: u32 size. This is because our trace is exported with its size as a field. But now that we have this new field, the hole in the middle disappears because it makes prev_state becoming well aligned. And since we are using a pointer to the raw trace using this struct, instead of reading prev_state, we are reading the hole. We could fix it by keeping the size seperate from the struct but actually there a lot of other potential problems: some fields may be saved as long in a 64 bits system and later read as long in a 32 bits system. Also this direct cast doesn't care about the endianness differences between the host traced machine and the machine in which we do the post processing. So instead of using such dangerous direct casts, fetch the values using the trace parsing API that already takes care of all these problems. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-12 00:43:45 +00:00
u16 common_type;
u8 common_flags;
u8 common_preempt_count;
u32 common_pid;
u32 common_tgid;
char parent_comm[16];
u32 parent_pid;
char child_comm[16];
u32 child_pid;
};
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
struct trace_migrate_task_event {
u32 size;
u16 common_type;
u8 common_flags;
u8 common_preempt_count;
u32 common_pid;
u32 common_tgid;
char comm[16];
u32 pid;
u32 prio;
u32 cpu;
};
struct trace_sched_handler {
void (*switch_event)(struct trace_switch_event *,
struct perf_session *,
struct event *,
int cpu,
u64 timestamp,
struct thread *thread);
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
void (*runtime_event)(struct trace_runtime_event *,
struct perf_session *,
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
struct event *,
int cpu,
u64 timestamp,
struct thread *thread);
void (*wakeup_event)(struct trace_wakeup_event *,
struct perf_session *,
struct event *,
int cpu,
u64 timestamp,
struct thread *thread);
void (*fork_event)(struct trace_fork_event *,
struct event *,
int cpu,
u64 timestamp,
struct thread *thread);
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
void (*migrate_task_event)(struct trace_migrate_task_event *,
struct perf_session *session,
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
struct event *,
int cpu,
u64 timestamp,
struct thread *thread);
};
perf sched: Fix bad event alignment perf sched raises the following error when it meets a sched switch event: perf: builtin-sched.c:286: register_pid: Assertion `!(pid >= 65536)' failed. Abandon Currently in x86-64, the sched switch events have a hole in the middle of the structure: u16 common_type; u8 common_flags; u8 common_preempt_count; u32 common_pid; u32 common_tgid; char prev_comm[16]; u32 prev_pid; u32 prev_prio; <--- there u64 prev_state; char next_comm[16]; u32 next_pid; u32 next_prio; Gcc inserts a 4 bytes hole there for prev_state to be u64 aligned. And the events are exported to userspace with this hole. But in userspace, from perf sched, we fetch it using a structure that has a new field in the beginning: u32 size. This is because our trace is exported with its size as a field. But now that we have this new field, the hole in the middle disappears because it makes prev_state becoming well aligned. And since we are using a pointer to the raw trace using this struct, instead of reading prev_state, we are reading the hole. We could fix it by keeping the size seperate from the struct but actually there a lot of other potential problems: some fields may be saved as long in a 64 bits system and later read as long in a 32 bits system. Also this direct cast doesn't care about the endianness differences between the host traced machine and the machine in which we do the post processing. So instead of using such dangerous direct casts, fetch the values using the trace parsing API that already takes care of all these problems. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-12 00:43:45 +00:00
static void
replay_wakeup_event(struct trace_wakeup_event *wakeup_event,
struct perf_session *session __used,
struct event *event,
int cpu __used,
u64 timestamp __used,
struct thread *thread __used)
{
struct task_desc *waker, *wakee;
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
if (verbose) {
printf("sched_wakeup event %p\n", event);
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
printf(" ... pid %d woke up %s/%d\n",
wakeup_event->common_pid,
wakeup_event->comm,
wakeup_event->pid);
}
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
waker = register_pid(wakeup_event->common_pid, "<unknown>");
wakee = register_pid(wakeup_event->pid, wakeup_event->comm);
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
add_sched_event_wakeup(waker, timestamp, wakee);
}
perf sched: Fix 'perf sched latency' output on 32-bit systems Before: ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- perf |4853313.251 ms | 10 | avg: 0.046 ms | max: 0.337 ms | flush-8:0 |2426659.202 ms | 5 | avg: 0.015 ms | max: 0.016 ms | sleep |485331.966 ms | 1 | avg: 0.012 ms | max: 0.012 ms | ksoftirqd/1 |485331.320 ms | 1 | avg: 0.005 ms | max: 0.005 ms | ----------------------------------------------------------------------------------- TOTAL: |8250635.739 ms | 17 | --------------------------------------------- After: ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- perf | 0.206 ms | 10 | avg: 0.046 ms | max: 0.337 ms | flush-8:0 | 2.680 ms | 5 | avg: 0.015 ms | max: 0.016 ms | sleep | 0.662 ms | 1 | avg: 0.012 ms | max: 0.012 ms | ksoftirqd/1 | 0.015 ms | 1 | avg: 0.005 ms | max: 0.005 ms | ----------------------------------------------------------------------------------- TOTAL: | 3.563 ms | 17 | --------------------------------------------- Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 16:22:53 +00:00
static u64 cpu_last_switched[MAX_CPUS];
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
static void
replay_switch_event(struct trace_switch_event *switch_event,
struct perf_session *session __used,
struct event *event,
int cpu,
u64 timestamp,
struct thread *thread __used)
{
struct task_desc *prev, __used *next;
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
u64 timestamp0;
s64 delta;
if (verbose)
printf("sched_switch event %p\n", event);
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
if (cpu >= MAX_CPUS || cpu < 0)
return;
timestamp0 = cpu_last_switched[cpu];
if (timestamp0)
delta = timestamp - timestamp0;
else
delta = 0;
if (delta < 0)
die("hm, delta: %" PRIu64 " < 0 ?\n", delta);
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
if (verbose) {
printf(" ... switch from %s/%d to %s/%d [ran %" PRIu64 " nsecs]\n",
switch_event->prev_comm, switch_event->prev_pid,
switch_event->next_comm, switch_event->next_pid,
delta);
}
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
prev = register_pid(switch_event->prev_pid, switch_event->prev_comm);
next = register_pid(switch_event->next_pid, switch_event->next_comm);
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
cpu_last_switched[cpu] = timestamp;
add_sched_event_run(prev, timestamp, delta);
add_sched_event_sleep(prev, timestamp, switch_event->prev_state);
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
}
static void
replay_fork_event(struct trace_fork_event *fork_event,
struct event *event,
int cpu __used,
u64 timestamp __used,
struct thread *thread __used)
{
if (verbose) {
printf("sched_fork event %p\n", event);
printf("... parent: %s/%d\n", fork_event->parent_comm, fork_event->parent_pid);
printf("... child: %s/%d\n", fork_event->child_comm, fork_event->child_pid);
}
register_pid(fork_event->parent_pid, fork_event->parent_comm);
register_pid(fork_event->child_pid, fork_event->child_comm);
}
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
static struct trace_sched_handler replay_ops = {
.wakeup_event = replay_wakeup_event,
.switch_event = replay_switch_event,
.fork_event = replay_fork_event,
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
};
struct sort_dimension {
const char *name;
sort_fn_t cmp;
struct list_head list;
};
static LIST_HEAD(cmp_pid);
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
static int
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
thread_lat_cmp(struct list_head *list, struct work_atoms *l, struct work_atoms *r)
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
{
struct sort_dimension *sort;
int ret = 0;
BUG_ON(list_empty(list));
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
list_for_each_entry(sort, list, list) {
ret = sort->cmp(l, r);
if (ret)
return ret;
}
return ret;
}
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
static struct work_atoms *
thread_atoms_search(struct rb_root *root, struct thread *thread,
struct list_head *sort_list)
{
struct rb_node *node = root->rb_node;
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
struct work_atoms key = { .thread = thread };
while (node) {
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
struct work_atoms *atoms;
int cmp;
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
atoms = container_of(node, struct work_atoms, node);
cmp = thread_lat_cmp(sort_list, &key, atoms);
if (cmp > 0)
node = node->rb_left;
else if (cmp < 0)
node = node->rb_right;
else {
BUG_ON(thread != atoms->thread);
return atoms;
}
}
return NULL;
}
static void
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
__thread_latency_insert(struct rb_root *root, struct work_atoms *data,
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
struct list_head *sort_list)
{
struct rb_node **new = &(root->rb_node), *parent = NULL;
while (*new) {
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
struct work_atoms *this;
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
int cmp;
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
this = container_of(*new, struct work_atoms, node);
parent = *new;
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
cmp = thread_lat_cmp(sort_list, data, this);
if (cmp > 0)
new = &((*new)->rb_left);
else
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
new = &((*new)->rb_right);
}
rb_link_node(&data->node, parent, new);
rb_insert_color(&data->node, root);
}
static void thread_atoms_insert(struct thread *thread)
{
struct work_atoms *atoms = zalloc(sizeof(*atoms));
if (!atoms)
die("No memory");
atoms->thread = thread;
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
INIT_LIST_HEAD(&atoms->work_list);
__thread_latency_insert(&atom_root, atoms, &cmp_pid);
}
static void
latency_fork_event(struct trace_fork_event *fork_event __used,
struct event *event __used,
int cpu __used,
u64 timestamp __used,
struct thread *thread __used)
{
/* should insert the newcomer */
}
__used
static char sched_out_state(struct trace_switch_event *switch_event)
{
const char *str = TASK_STATE_TO_CHAR_STR;
return str[switch_event->prev_state];
}
static void
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
add_sched_out_event(struct work_atoms *atoms,
char run_state,
u64 timestamp)
{
struct work_atom *atom = zalloc(sizeof(*atom));
if (!atom)
die("Non memory");
perf tools: Fix processing of randomly serialized sched traces Currently it's possible to meet such too high latency results with 'perf sched latency'. ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- xfce4-panel | 0.222 ms | 2 | avg: 4718.345 ms | max: 9436.493 ms | scsi_eh_3 | 3.962 ms | 36 | avg: 55.957 ms | max: 1977.829 ms | The origin is on traces that are sometimes badly serialized across cpus. For example the raw traces that raised such results for xfce4-panel: (1) [init]-0 [000] 1494.663899990: sched_switch: task swapper:0 [140] (R) ==> xfce4-panel:4569 [120] (2) xfce4-panel-4569 [000] 1494.663928373: sched_switch: task xfce4-panel:4569 [120] (S) ==> swapper:0 [140] (3) Xorg-4276 [001] 1494.663860125: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (4) Xorg-4276 [001] 1504.098252756: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (5) perf-5219 [000] 1504.100353302: sched_switch: task perf:5219 [120] (S) ==> xfce4-panel:4569 [120] The traces are processed in the order they arrive. Then in (2), xfce4-panel sleeps, it is first waken up in (3) and eventually scheduled in (5). The latency reported is then 1504 - 1495 = 9 secs, as reported by perf sched. But this is wrong, we are confident in the fact the traces are nicely serialized while we should actually more trust the timestamps. If we reorder by timestamps we get: (1) Xorg-4276 [001] 1494.663860125: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (2) [init]-0 [000] 1494.663899990: sched_switch: task swapper:0 [140] (R) ==> xfce4-panel:4569 [120] (3) xfce4-panel-4569 [000] 1494.663928373: sched_switch: task xfce4-panel:4569 [120] (S) ==> swapper:0 [140] (4) Xorg-4276 [001] 1504.098252756: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (5) perf-5219 [000] 1504.100353302: sched_switch: task perf:5219 [120] (S) ==> xfce4-panel:4569 [120] Now the trace make more sense, xfce4-panel is sleeping. Then it is woken up in (1), scheduled in (2) It goes to sleep in (3), woken up in (4) and scheduled in (5). Now, latency captured between (1) and (2) is of 39 us. And between (4) and (5) it is 2.1 ms. Such pattern of bad serializing is the origin of the high latencies reported by perf sched. Basically, we need to check whether wake up time is higher than schedule out time. If it's not the case, we need to tag the current work atom as invalid. Beside that, we may need to work later on a better ordering of the traces given by the kernel. After this patch: xfce4-session | 0.221 ms | 1 | avg: 0.538 ms | max: 0.538 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 01:01:12 +00:00
atom->sched_out_time = timestamp;
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
if (run_state == 'R') {
atom->state = THREAD_WAIT_CPU;
perf tools: Fix processing of randomly serialized sched traces Currently it's possible to meet such too high latency results with 'perf sched latency'. ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- xfce4-panel | 0.222 ms | 2 | avg: 4718.345 ms | max: 9436.493 ms | scsi_eh_3 | 3.962 ms | 36 | avg: 55.957 ms | max: 1977.829 ms | The origin is on traces that are sometimes badly serialized across cpus. For example the raw traces that raised such results for xfce4-panel: (1) [init]-0 [000] 1494.663899990: sched_switch: task swapper:0 [140] (R) ==> xfce4-panel:4569 [120] (2) xfce4-panel-4569 [000] 1494.663928373: sched_switch: task xfce4-panel:4569 [120] (S) ==> swapper:0 [140] (3) Xorg-4276 [001] 1494.663860125: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (4) Xorg-4276 [001] 1504.098252756: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (5) perf-5219 [000] 1504.100353302: sched_switch: task perf:5219 [120] (S) ==> xfce4-panel:4569 [120] The traces are processed in the order they arrive. Then in (2), xfce4-panel sleeps, it is first waken up in (3) and eventually scheduled in (5). The latency reported is then 1504 - 1495 = 9 secs, as reported by perf sched. But this is wrong, we are confident in the fact the traces are nicely serialized while we should actually more trust the timestamps. If we reorder by timestamps we get: (1) Xorg-4276 [001] 1494.663860125: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (2) [init]-0 [000] 1494.663899990: sched_switch: task swapper:0 [140] (R) ==> xfce4-panel:4569 [120] (3) xfce4-panel-4569 [000] 1494.663928373: sched_switch: task xfce4-panel:4569 [120] (S) ==> swapper:0 [140] (4) Xorg-4276 [001] 1504.098252756: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (5) perf-5219 [000] 1504.100353302: sched_switch: task perf:5219 [120] (S) ==> xfce4-panel:4569 [120] Now the trace make more sense, xfce4-panel is sleeping. Then it is woken up in (1), scheduled in (2) It goes to sleep in (3), woken up in (4) and scheduled in (5). Now, latency captured between (1) and (2) is of 39 us. And between (4) and (5) it is 2.1 ms. Such pattern of bad serializing is the origin of the high latencies reported by perf sched. Basically, we need to check whether wake up time is higher than schedule out time. If it's not the case, we need to tag the current work atom as invalid. Beside that, we may need to work later on a better ordering of the traces given by the kernel. After this patch: xfce4-session | 0.221 ms | 1 | avg: 0.538 ms | max: 0.538 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 01:01:12 +00:00
atom->wake_up_time = atom->sched_out_time;
}
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
list_add_tail(&atom->list, &atoms->work_list);
}
static void
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
add_runtime_event(struct work_atoms *atoms, u64 delta, u64 timestamp __used)
{
struct work_atom *atom;
BUG_ON(list_empty(&atoms->work_list));
atom = list_entry(atoms->work_list.prev, struct work_atom, list);
atom->runtime += delta;
atoms->total_runtime += delta;
}
static void
add_sched_in_event(struct work_atoms *atoms, u64 timestamp)
{
struct work_atom *atom;
u64 delta;
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
if (list_empty(&atoms->work_list))
return;
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
atom = list_entry(atoms->work_list.prev, struct work_atom, list);
if (atom->state != THREAD_WAIT_CPU)
return;
if (timestamp < atom->wake_up_time) {
atom->state = THREAD_IGNORE;
return;
}
atom->state = THREAD_SCHED_IN;
atom->sched_in_time = timestamp;
delta = atom->sched_in_time - atom->wake_up_time;
atoms->total_lat += delta;
if (delta > atoms->max_lat) {
atoms->max_lat = delta;
atoms->max_lat_at = timestamp;
}
atoms->nb_atoms++;
}
static void
latency_switch_event(struct trace_switch_event *switch_event,
struct perf_session *session,
struct event *event __used,
int cpu,
u64 timestamp,
struct thread *thread __used)
{
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
struct work_atoms *out_events, *in_events;
struct thread *sched_out, *sched_in;
u64 timestamp0;
s64 delta;
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
BUG_ON(cpu >= MAX_CPUS || cpu < 0);
timestamp0 = cpu_last_switched[cpu];
cpu_last_switched[cpu] = timestamp;
if (timestamp0)
delta = timestamp - timestamp0;
else
delta = 0;
if (delta < 0)
die("hm, delta: %" PRIu64 " < 0 ?\n", delta);
sched_out = perf_session__findnew(session, switch_event->prev_pid);
sched_in = perf_session__findnew(session, switch_event->next_pid);
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
out_events = thread_atoms_search(&atom_root, sched_out, &cmp_pid);
if (!out_events) {
thread_atoms_insert(sched_out);
out_events = thread_atoms_search(&atom_root, sched_out, &cmp_pid);
if (!out_events)
die("out-event: Internal tree error");
}
add_sched_out_event(out_events, sched_out_state(switch_event), timestamp);
in_events = thread_atoms_search(&atom_root, sched_in, &cmp_pid);
if (!in_events) {
thread_atoms_insert(sched_in);
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
in_events = thread_atoms_search(&atom_root, sched_in, &cmp_pid);
if (!in_events)
die("in-event: Internal tree error");
/*
* Take came in we have not heard about yet,
* add in an initial atom in runnable state:
*/
add_sched_out_event(in_events, 'R', timestamp);
}
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
add_sched_in_event(in_events, timestamp);
}
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
static void
latency_runtime_event(struct trace_runtime_event *runtime_event,
struct perf_session *session,
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
struct event *event __used,
int cpu,
u64 timestamp,
struct thread *this_thread __used)
{
struct thread *thread = perf_session__findnew(session, runtime_event->pid);
struct work_atoms *atoms = thread_atoms_search(&atom_root, thread, &cmp_pid);
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
BUG_ON(cpu >= MAX_CPUS || cpu < 0);
if (!atoms) {
thread_atoms_insert(thread);
atoms = thread_atoms_search(&atom_root, thread, &cmp_pid);
if (!atoms)
die("in-event: Internal tree error");
add_sched_out_event(atoms, 'R', timestamp);
}
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
add_runtime_event(atoms, runtime_event->runtime, timestamp);
}
static void
latency_wakeup_event(struct trace_wakeup_event *wakeup_event,
struct perf_session *session,
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
struct event *__event __used,
int cpu __used,
u64 timestamp,
struct thread *thread __used)
{
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
struct work_atoms *atoms;
struct work_atom *atom;
struct thread *wakee;
/* Note for later, it may be interesting to observe the failing cases */
if (!wakeup_event->success)
return;
wakee = perf_session__findnew(session, wakeup_event->pid);
atoms = thread_atoms_search(&atom_root, wakee, &cmp_pid);
if (!atoms) {
thread_atoms_insert(wakee);
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
atoms = thread_atoms_search(&atom_root, wakee, &cmp_pid);
if (!atoms)
die("wakeup-event: Internal tree error");
add_sched_out_event(atoms, 'S', timestamp);
}
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
BUG_ON(list_empty(&atoms->work_list));
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
atom = list_entry(atoms->work_list.prev, struct work_atom, list);
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
/*
* You WILL be missing events if you've recorded only
* one CPU, or are only looking at only one, so don't
* make useless noise.
*/
if (profile_cpu == -1 && atom->state != THREAD_SLEEPING)
nr_state_machine_bugs++;
perf tools: Implement counter output multiplexing Finish the -M/--multiplex option implementation: - separate it out from group_fd - correctly set it via the ioctl and dont mmap counters that are multiplexed - modify the perf record event loop to deal with buffer-less counters. - remove the -g option from perf sched record - account for unordered events in perf sched latency - (add -f to perf sched record to ease measurements) - skip idle threads (pid==0) in latency output The result is better latency output by 'perf sched latency': ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- ksoftirqd/8 | 0.071 ms | 2 | avg: 0.458 ms | max: 0.913 ms | at-spi-registry | 0.609 ms | 19 | avg: 0.013 ms | max: 0.023 ms | perf | 3.316 ms | 16 | avg: 0.013 ms | max: 0.054 ms | Xorg | 0.392 ms | 19 | avg: 0.011 ms | max: 0.018 ms | sleep | 0.537 ms | 2 | avg: 0.009 ms | max: 0.009 ms | ----------------------------------------------------------------------------------- TOTAL: | 4.925 ms | 58 | --------------------------------------------- Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 16:15:54 +00:00
nr_timestamps++;
if (atom->sched_out_time > timestamp) {
nr_unordered_timestamps++;
perf tools: Fix processing of randomly serialized sched traces Currently it's possible to meet such too high latency results with 'perf sched latency'. ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- xfce4-panel | 0.222 ms | 2 | avg: 4718.345 ms | max: 9436.493 ms | scsi_eh_3 | 3.962 ms | 36 | avg: 55.957 ms | max: 1977.829 ms | The origin is on traces that are sometimes badly serialized across cpus. For example the raw traces that raised such results for xfce4-panel: (1) [init]-0 [000] 1494.663899990: sched_switch: task swapper:0 [140] (R) ==> xfce4-panel:4569 [120] (2) xfce4-panel-4569 [000] 1494.663928373: sched_switch: task xfce4-panel:4569 [120] (S) ==> swapper:0 [140] (3) Xorg-4276 [001] 1494.663860125: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (4) Xorg-4276 [001] 1504.098252756: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (5) perf-5219 [000] 1504.100353302: sched_switch: task perf:5219 [120] (S) ==> xfce4-panel:4569 [120] The traces are processed in the order they arrive. Then in (2), xfce4-panel sleeps, it is first waken up in (3) and eventually scheduled in (5). The latency reported is then 1504 - 1495 = 9 secs, as reported by perf sched. But this is wrong, we are confident in the fact the traces are nicely serialized while we should actually more trust the timestamps. If we reorder by timestamps we get: (1) Xorg-4276 [001] 1494.663860125: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (2) [init]-0 [000] 1494.663899990: sched_switch: task swapper:0 [140] (R) ==> xfce4-panel:4569 [120] (3) xfce4-panel-4569 [000] 1494.663928373: sched_switch: task xfce4-panel:4569 [120] (S) ==> swapper:0 [140] (4) Xorg-4276 [001] 1504.098252756: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (5) perf-5219 [000] 1504.100353302: sched_switch: task perf:5219 [120] (S) ==> xfce4-panel:4569 [120] Now the trace make more sense, xfce4-panel is sleeping. Then it is woken up in (1), scheduled in (2) It goes to sleep in (3), woken up in (4) and scheduled in (5). Now, latency captured between (1) and (2) is of 39 us. And between (4) and (5) it is 2.1 ms. Such pattern of bad serializing is the origin of the high latencies reported by perf sched. Basically, we need to check whether wake up time is higher than schedule out time. If it's not the case, we need to tag the current work atom as invalid. Beside that, we may need to work later on a better ordering of the traces given by the kernel. After this patch: xfce4-session | 0.221 ms | 1 | avg: 0.538 ms | max: 0.538 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 01:01:12 +00:00
return;
perf tools: Implement counter output multiplexing Finish the -M/--multiplex option implementation: - separate it out from group_fd - correctly set it via the ioctl and dont mmap counters that are multiplexed - modify the perf record event loop to deal with buffer-less counters. - remove the -g option from perf sched record - account for unordered events in perf sched latency - (add -f to perf sched record to ease measurements) - skip idle threads (pid==0) in latency output The result is better latency output by 'perf sched latency': ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- ksoftirqd/8 | 0.071 ms | 2 | avg: 0.458 ms | max: 0.913 ms | at-spi-registry | 0.609 ms | 19 | avg: 0.013 ms | max: 0.023 ms | perf | 3.316 ms | 16 | avg: 0.013 ms | max: 0.054 ms | Xorg | 0.392 ms | 19 | avg: 0.011 ms | max: 0.018 ms | sleep | 0.537 ms | 2 | avg: 0.009 ms | max: 0.009 ms | ----------------------------------------------------------------------------------- TOTAL: | 4.925 ms | 58 | --------------------------------------------- Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 16:15:54 +00:00
}
perf tools: Fix processing of randomly serialized sched traces Currently it's possible to meet such too high latency results with 'perf sched latency'. ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- xfce4-panel | 0.222 ms | 2 | avg: 4718.345 ms | max: 9436.493 ms | scsi_eh_3 | 3.962 ms | 36 | avg: 55.957 ms | max: 1977.829 ms | The origin is on traces that are sometimes badly serialized across cpus. For example the raw traces that raised such results for xfce4-panel: (1) [init]-0 [000] 1494.663899990: sched_switch: task swapper:0 [140] (R) ==> xfce4-panel:4569 [120] (2) xfce4-panel-4569 [000] 1494.663928373: sched_switch: task xfce4-panel:4569 [120] (S) ==> swapper:0 [140] (3) Xorg-4276 [001] 1494.663860125: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (4) Xorg-4276 [001] 1504.098252756: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (5) perf-5219 [000] 1504.100353302: sched_switch: task perf:5219 [120] (S) ==> xfce4-panel:4569 [120] The traces are processed in the order they arrive. Then in (2), xfce4-panel sleeps, it is first waken up in (3) and eventually scheduled in (5). The latency reported is then 1504 - 1495 = 9 secs, as reported by perf sched. But this is wrong, we are confident in the fact the traces are nicely serialized while we should actually more trust the timestamps. If we reorder by timestamps we get: (1) Xorg-4276 [001] 1494.663860125: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (2) [init]-0 [000] 1494.663899990: sched_switch: task swapper:0 [140] (R) ==> xfce4-panel:4569 [120] (3) xfce4-panel-4569 [000] 1494.663928373: sched_switch: task xfce4-panel:4569 [120] (S) ==> swapper:0 [140] (4) Xorg-4276 [001] 1504.098252756: sched_wakeup: task xfce4-panel:4569 [120] success=1 [000] (5) perf-5219 [000] 1504.100353302: sched_switch: task perf:5219 [120] (S) ==> xfce4-panel:4569 [120] Now the trace make more sense, xfce4-panel is sleeping. Then it is woken up in (1), scheduled in (2) It goes to sleep in (3), woken up in (4) and scheduled in (5). Now, latency captured between (1) and (2) is of 39 us. And between (4) and (5) it is 2.1 ms. Such pattern of bad serializing is the origin of the high latencies reported by perf sched. Basically, we need to check whether wake up time is higher than schedule out time. If it's not the case, we need to tag the current work atom as invalid. Beside that, we may need to work later on a better ordering of the traces given by the kernel. After this patch: xfce4-session | 0.221 ms | 1 | avg: 0.538 ms | max: 0.538 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 01:01:12 +00:00
atom->state = THREAD_WAIT_CPU;
atom->wake_up_time = timestamp;
}
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
static void
latency_migrate_task_event(struct trace_migrate_task_event *migrate_task_event,
struct perf_session *session,
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
struct event *__event __used,
int cpu __used,
u64 timestamp,
struct thread *thread __used)
{
struct work_atoms *atoms;
struct work_atom *atom;
struct thread *migrant;
/*
* Only need to worry about migration when profiling one CPU.
*/
if (profile_cpu == -1)
return;
migrant = perf_session__findnew(session, migrate_task_event->pid);
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
atoms = thread_atoms_search(&atom_root, migrant, &cmp_pid);
if (!atoms) {
thread_atoms_insert(migrant);
register_pid(migrant->pid, migrant->comm);
atoms = thread_atoms_search(&atom_root, migrant, &cmp_pid);
if (!atoms)
die("migration-event: Internal tree error");
add_sched_out_event(atoms, 'R', timestamp);
}
BUG_ON(list_empty(&atoms->work_list));
atom = list_entry(atoms->work_list.prev, struct work_atom, list);
atom->sched_in_time = atom->sched_out_time = atom->wake_up_time = timestamp;
nr_timestamps++;
if (atom->sched_out_time > timestamp)
nr_unordered_timestamps++;
}
static struct trace_sched_handler lat_ops = {
.wakeup_event = latency_wakeup_event,
.switch_event = latency_switch_event,
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
.runtime_event = latency_runtime_event,
.fork_event = latency_fork_event,
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
.migrate_task_event = latency_migrate_task_event,
};
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
static void output_lat_thread(struct work_atoms *work_list)
{
int i;
int ret;
u64 avg;
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
if (!work_list->nb_atoms)
return;
perf tools: Implement counter output multiplexing Finish the -M/--multiplex option implementation: - separate it out from group_fd - correctly set it via the ioctl and dont mmap counters that are multiplexed - modify the perf record event loop to deal with buffer-less counters. - remove the -g option from perf sched record - account for unordered events in perf sched latency - (add -f to perf sched record to ease measurements) - skip idle threads (pid==0) in latency output The result is better latency output by 'perf sched latency': ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- ksoftirqd/8 | 0.071 ms | 2 | avg: 0.458 ms | max: 0.913 ms | at-spi-registry | 0.609 ms | 19 | avg: 0.013 ms | max: 0.023 ms | perf | 3.316 ms | 16 | avg: 0.013 ms | max: 0.054 ms | Xorg | 0.392 ms | 19 | avg: 0.011 ms | max: 0.018 ms | sleep | 0.537 ms | 2 | avg: 0.009 ms | max: 0.009 ms | ----------------------------------------------------------------------------------- TOTAL: | 4.925 ms | 58 | --------------------------------------------- Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 16:15:54 +00:00
/*
* Ignore idle threads:
*/
if (!strcmp(work_list->thread->comm, "swapper"))
perf tools: Implement counter output multiplexing Finish the -M/--multiplex option implementation: - separate it out from group_fd - correctly set it via the ioctl and dont mmap counters that are multiplexed - modify the perf record event loop to deal with buffer-less counters. - remove the -g option from perf sched record - account for unordered events in perf sched latency - (add -f to perf sched record to ease measurements) - skip idle threads (pid==0) in latency output The result is better latency output by 'perf sched latency': ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- ksoftirqd/8 | 0.071 ms | 2 | avg: 0.458 ms | max: 0.913 ms | at-spi-registry | 0.609 ms | 19 | avg: 0.013 ms | max: 0.023 ms | perf | 3.316 ms | 16 | avg: 0.013 ms | max: 0.054 ms | Xorg | 0.392 ms | 19 | avg: 0.011 ms | max: 0.018 ms | sleep | 0.537 ms | 2 | avg: 0.009 ms | max: 0.009 ms | ----------------------------------------------------------------------------------- TOTAL: | 4.925 ms | 58 | --------------------------------------------- Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 16:15:54 +00:00
return;
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
all_runtime += work_list->total_runtime;
all_count += work_list->nb_atoms;
ret = printf(" %s:%d ", work_list->thread->comm, work_list->thread->pid);
for (i = 0; i < 24 - ret; i++)
printf(" ");
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
avg = work_list->total_lat / work_list->nb_atoms;
printf("|%11.3f ms |%9" PRIu64 " | avg:%9.3f ms | max:%9.3f ms | max at: %9.6f s\n",
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
(double)work_list->total_runtime / 1e6,
work_list->nb_atoms, (double)avg / 1e6,
(double)work_list->max_lat / 1e6,
(double)work_list->max_lat_at / 1e9);
}
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
static int pid_cmp(struct work_atoms *l, struct work_atoms *r)
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
{
if (l->thread->pid < r->thread->pid)
return -1;
if (l->thread->pid > r->thread->pid)
return 1;
return 0;
}
static struct sort_dimension pid_sort_dimension = {
.name = "pid",
.cmp = pid_cmp,
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
};
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
static int avg_cmp(struct work_atoms *l, struct work_atoms *r)
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
{
u64 avgl, avgr;
if (!l->nb_atoms)
return -1;
if (!r->nb_atoms)
return 1;
avgl = l->total_lat / l->nb_atoms;
avgr = r->total_lat / r->nb_atoms;
if (avgl < avgr)
return -1;
if (avgl > avgr)
return 1;
return 0;
}
static struct sort_dimension avg_sort_dimension = {
.name = "avg",
.cmp = avg_cmp,
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
};
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
static int max_cmp(struct work_atoms *l, struct work_atoms *r)
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
{
if (l->max_lat < r->max_lat)
return -1;
if (l->max_lat > r->max_lat)
return 1;
return 0;
}
static struct sort_dimension max_sort_dimension = {
.name = "max",
.cmp = max_cmp,
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
};
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
static int switch_cmp(struct work_atoms *l, struct work_atoms *r)
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
{
if (l->nb_atoms < r->nb_atoms)
return -1;
if (l->nb_atoms > r->nb_atoms)
return 1;
return 0;
}
static struct sort_dimension switch_sort_dimension = {
.name = "switch",
.cmp = switch_cmp,
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
};
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
static int runtime_cmp(struct work_atoms *l, struct work_atoms *r)
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
{
if (l->total_runtime < r->total_runtime)
return -1;
if (l->total_runtime > r->total_runtime)
return 1;
return 0;
}
static struct sort_dimension runtime_sort_dimension = {
.name = "runtime",
.cmp = runtime_cmp,
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
};
static struct sort_dimension *available_sorts[] = {
&pid_sort_dimension,
&avg_sort_dimension,
&max_sort_dimension,
&switch_sort_dimension,
&runtime_sort_dimension,
};
#define NB_AVAILABLE_SORTS (int)(sizeof(available_sorts) / sizeof(struct sort_dimension *))
static LIST_HEAD(sort_list);
static int sort_dimension__add(const char *tok, struct list_head *list)
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
{
int i;
for (i = 0; i < NB_AVAILABLE_SORTS; i++) {
if (!strcmp(available_sorts[i]->name, tok)) {
list_add_tail(&available_sorts[i]->list, list);
return 0;
}
}
return -1;
}
static void setup_sorting(void);
static void sort_lat(void)
{
struct rb_node *node;
for (;;) {
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
struct work_atoms *data;
node = rb_first(&atom_root);
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
if (!node)
break;
rb_erase(node, &atom_root);
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
data = rb_entry(node, struct work_atoms, node);
__thread_latency_insert(&sorted_atom_root, data, &sort_list);
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
}
}
static struct trace_sched_handler *trace_handler;
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
static void
process_sched_wakeup_event(void *data, struct perf_session *session,
struct event *event,
int cpu __used,
u64 timestamp __used,
struct thread *thread __used)
{
struct trace_wakeup_event wakeup_event;
FILL_COMMON_FIELDS(wakeup_event, event, data);
FILL_ARRAY(wakeup_event, comm, event, data);
FILL_FIELD(wakeup_event, pid, event, data);
FILL_FIELD(wakeup_event, prio, event, data);
FILL_FIELD(wakeup_event, success, event, data);
FILL_FIELD(wakeup_event, cpu, event, data);
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
if (trace_handler->wakeup_event)
trace_handler->wakeup_event(&wakeup_event, session, event,
cpu, timestamp, thread);
}
/*
* Track the current task - that way we can know whether there's any
* weird events, such as a task being switched away that is not current.
*/
static int max_cpu;
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
static u32 curr_pid[MAX_CPUS] = { [0 ... MAX_CPUS-1] = -1 };
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
static struct thread *curr_thread[MAX_CPUS];
static char next_shortname1 = 'A';
static char next_shortname2 = '0';
static void
map_switch_event(struct trace_switch_event *switch_event,
struct perf_session *session,
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
struct event *event __used,
int this_cpu,
u64 timestamp,
struct thread *thread __used)
{
struct thread *sched_out __used, *sched_in;
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
int new_shortname;
u64 timestamp0;
s64 delta;
int cpu;
BUG_ON(this_cpu >= MAX_CPUS || this_cpu < 0);
if (this_cpu > max_cpu)
max_cpu = this_cpu;
timestamp0 = cpu_last_switched[this_cpu];
cpu_last_switched[this_cpu] = timestamp;
if (timestamp0)
delta = timestamp - timestamp0;
else
delta = 0;
if (delta < 0)
die("hm, delta: %" PRIu64 " < 0 ?\n", delta);
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
sched_out = perf_session__findnew(session, switch_event->prev_pid);
sched_in = perf_session__findnew(session, switch_event->next_pid);
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
curr_thread[this_cpu] = sched_in;
printf(" ");
new_shortname = 0;
if (!sched_in->shortname[0]) {
sched_in->shortname[0] = next_shortname1;
sched_in->shortname[1] = next_shortname2;
if (next_shortname1 < 'Z') {
next_shortname1++;
} else {
next_shortname1='A';
if (next_shortname2 < '9') {
next_shortname2++;
} else {
next_shortname2='0';
}
}
new_shortname = 1;
}
for (cpu = 0; cpu <= max_cpu; cpu++) {
if (cpu != this_cpu)
printf(" ");
else
printf("*");
if (curr_thread[cpu]) {
if (curr_thread[cpu]->pid)
printf("%2s ", curr_thread[cpu]->shortname);
else
printf(". ");
} else
printf(" ");
}
printf(" %12.6f secs ", (double)timestamp/1e9);
if (new_shortname) {
printf("%s => %s:%d\n",
sched_in->shortname, sched_in->comm, sched_in->pid);
} else {
printf("\n");
}
}
static void
process_sched_switch_event(void *data, struct perf_session *session,
struct event *event,
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
int this_cpu,
u64 timestamp __used,
struct thread *thread __used)
{
struct trace_switch_event switch_event;
FILL_COMMON_FIELDS(switch_event, event, data);
FILL_ARRAY(switch_event, prev_comm, event, data);
FILL_FIELD(switch_event, prev_pid, event, data);
FILL_FIELD(switch_event, prev_prio, event, data);
FILL_FIELD(switch_event, prev_state, event, data);
FILL_ARRAY(switch_event, next_comm, event, data);
FILL_FIELD(switch_event, next_pid, event, data);
FILL_FIELD(switch_event, next_prio, event, data);
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
if (curr_pid[this_cpu] != (u32)-1) {
/*
* Are we trying to switch away a PID that is
* not current?
*/
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
if (curr_pid[this_cpu] != switch_event.prev_pid)
nr_context_switch_bugs++;
}
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
if (trace_handler->switch_event)
trace_handler->switch_event(&switch_event, session, event,
this_cpu, timestamp, thread);
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
curr_pid[this_cpu] = switch_event.next_pid;
}
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
static void
process_sched_runtime_event(void *data, struct perf_session *session,
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
struct event *event,
int cpu __used,
u64 timestamp __used,
struct thread *thread __used)
{
struct trace_runtime_event runtime_event;
FILL_ARRAY(runtime_event, comm, event, data);
FILL_FIELD(runtime_event, pid, event, data);
FILL_FIELD(runtime_event, runtime, event, data);
FILL_FIELD(runtime_event, vruntime, event, data);
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
if (trace_handler->runtime_event)
trace_handler->runtime_event(&runtime_event, session, event, cpu, timestamp, thread);
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
}
static void
process_sched_fork_event(void *data,
struct event *event,
int cpu __used,
u64 timestamp __used,
struct thread *thread __used)
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
{
perf sched: Fix bad event alignment perf sched raises the following error when it meets a sched switch event: perf: builtin-sched.c:286: register_pid: Assertion `!(pid >= 65536)' failed. Abandon Currently in x86-64, the sched switch events have a hole in the middle of the structure: u16 common_type; u8 common_flags; u8 common_preempt_count; u32 common_pid; u32 common_tgid; char prev_comm[16]; u32 prev_pid; u32 prev_prio; <--- there u64 prev_state; char next_comm[16]; u32 next_pid; u32 next_prio; Gcc inserts a 4 bytes hole there for prev_state to be u64 aligned. And the events are exported to userspace with this hole. But in userspace, from perf sched, we fetch it using a structure that has a new field in the beginning: u32 size. This is because our trace is exported with its size as a field. But now that we have this new field, the hole in the middle disappears because it makes prev_state becoming well aligned. And since we are using a pointer to the raw trace using this struct, instead of reading prev_state, we are reading the hole. We could fix it by keeping the size seperate from the struct but actually there a lot of other potential problems: some fields may be saved as long in a 64 bits system and later read as long in a 32 bits system. Also this direct cast doesn't care about the endianness differences between the host traced machine and the machine in which we do the post processing. So instead of using such dangerous direct casts, fetch the values using the trace parsing API that already takes care of all these problems. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-12 00:43:45 +00:00
struct trace_fork_event fork_event;
FILL_COMMON_FIELDS(fork_event, event, data);
perf sched: Fix bad event alignment perf sched raises the following error when it meets a sched switch event: perf: builtin-sched.c:286: register_pid: Assertion `!(pid >= 65536)' failed. Abandon Currently in x86-64, the sched switch events have a hole in the middle of the structure: u16 common_type; u8 common_flags; u8 common_preempt_count; u32 common_pid; u32 common_tgid; char prev_comm[16]; u32 prev_pid; u32 prev_prio; <--- there u64 prev_state; char next_comm[16]; u32 next_pid; u32 next_prio; Gcc inserts a 4 bytes hole there for prev_state to be u64 aligned. And the events are exported to userspace with this hole. But in userspace, from perf sched, we fetch it using a structure that has a new field in the beginning: u32 size. This is because our trace is exported with its size as a field. But now that we have this new field, the hole in the middle disappears because it makes prev_state becoming well aligned. And since we are using a pointer to the raw trace using this struct, instead of reading prev_state, we are reading the hole. We could fix it by keeping the size seperate from the struct but actually there a lot of other potential problems: some fields may be saved as long in a 64 bits system and later read as long in a 32 bits system. Also this direct cast doesn't care about the endianness differences between the host traced machine and the machine in which we do the post processing. So instead of using such dangerous direct casts, fetch the values using the trace parsing API that already takes care of all these problems. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-12 00:43:45 +00:00
FILL_ARRAY(fork_event, parent_comm, event, data);
FILL_FIELD(fork_event, parent_pid, event, data);
FILL_ARRAY(fork_event, child_comm, event, data);
FILL_FIELD(fork_event, child_pid, event, data);
perf sched: Fix bad event alignment perf sched raises the following error when it meets a sched switch event: perf: builtin-sched.c:286: register_pid: Assertion `!(pid >= 65536)' failed. Abandon Currently in x86-64, the sched switch events have a hole in the middle of the structure: u16 common_type; u8 common_flags; u8 common_preempt_count; u32 common_pid; u32 common_tgid; char prev_comm[16]; u32 prev_pid; u32 prev_prio; <--- there u64 prev_state; char next_comm[16]; u32 next_pid; u32 next_prio; Gcc inserts a 4 bytes hole there for prev_state to be u64 aligned. And the events are exported to userspace with this hole. But in userspace, from perf sched, we fetch it using a structure that has a new field in the beginning: u32 size. This is because our trace is exported with its size as a field. But now that we have this new field, the hole in the middle disappears because it makes prev_state becoming well aligned. And since we are using a pointer to the raw trace using this struct, instead of reading prev_state, we are reading the hole. We could fix it by keeping the size seperate from the struct but actually there a lot of other potential problems: some fields may be saved as long in a 64 bits system and later read as long in a 32 bits system. Also this direct cast doesn't care about the endianness differences between the host traced machine and the machine in which we do the post processing. So instead of using such dangerous direct casts, fetch the values using the trace parsing API that already takes care of all these problems. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-12 00:43:45 +00:00
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
if (trace_handler->fork_event)
trace_handler->fork_event(&fork_event, event,
cpu, timestamp, thread);
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
}
static void
process_sched_exit_event(struct event *event,
int cpu __used,
u64 timestamp __used,
struct thread *thread __used)
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
{
if (verbose)
printf("sched_exit event %p\n", event);
}
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
static void
process_sched_migrate_task_event(void *data, struct perf_session *session,
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
struct event *event,
int cpu __used,
u64 timestamp __used,
struct thread *thread __used)
{
struct trace_migrate_task_event migrate_task_event;
FILL_COMMON_FIELDS(migrate_task_event, event, data);
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
FILL_ARRAY(migrate_task_event, comm, event, data);
FILL_FIELD(migrate_task_event, pid, event, data);
FILL_FIELD(migrate_task_event, prio, event, data);
FILL_FIELD(migrate_task_event, cpu, event, data);
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
if (trace_handler->migrate_task_event)
trace_handler->migrate_task_event(&migrate_task_event, session,
event, cpu, timestamp, thread);
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
}
static void process_raw_event(union perf_event *raw_event __used,
struct perf_session *session, void *data, int cpu,
u64 timestamp, struct thread *thread)
{
struct event *event;
int type;
type = trace_parse_common_type(data);
event = trace_find_event(type);
if (!strcmp(event->name, "sched_switch"))
process_sched_switch_event(data, session, event, cpu, timestamp, thread);
perf sched: Add support for sched:sched_stat_runtime events This allows more precise 'perf sched latency' output: --------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | --------------------------------------------------------------------------------------- ksoftirqd/0-4 | 0.010 ms | 2 | avg: 2.476 ms | max: 2.977 ms | perf-12328 | 15.844 ms | 66 | avg: 1.118 ms | max: 9.979 ms | bdi-default-235 | 0.009 ms | 1 | avg: 0.998 ms | max: 0.998 ms | events/1-8 | 0.020 ms | 2 | avg: 0.998 ms | max: 0.998 ms | events/0-7 | 0.018 ms | 2 | avg: 0.992 ms | max: 0.996 ms | sleep-12329 | 0.742 ms | 3 | avg: 0.906 ms | max: 2.289 ms | sshd-12122 | 0.163 ms | 2 | avg: 0.283 ms | max: 0.562 ms | loop-getpid-lon-12322 | 1023.636 ms | 69 | avg: 0.208 ms | max: 5.996 ms | loop-getpid-lon-12321 | 1038.638 ms | 5 | avg: 0.073 ms | max: 0.171 ms | migration/1-5 | 0.000 ms | 1 | avg: 0.006 ms | max: 0.006 ms | --------------------------------------------------------------------------------------- TOTAL: | 2079.078 ms | 153 | ------------------------------------------------- Also, streamline the code a bit more, add asserts for various state machine failures (they should be debugged if they occur) and fix a few odd ends. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 18:04:48 +00:00
if (!strcmp(event->name, "sched_stat_runtime"))
process_sched_runtime_event(data, session, event, cpu, timestamp, thread);
if (!strcmp(event->name, "sched_wakeup"))
process_sched_wakeup_event(data, session, event, cpu, timestamp, thread);
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
if (!strcmp(event->name, "sched_wakeup_new"))
process_sched_wakeup_event(data, session, event, cpu, timestamp, thread);
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
if (!strcmp(event->name, "sched_process_fork"))
process_sched_fork_event(data, event, cpu, timestamp, thread);
perf sched: Implement the scheduling workload replay engine Integrate the schedbench.c bits with the raw trace events that we get from the perf machinery, and activate the workload replayer/simulator. Example of a captured 'make -j' workload: $ perf sched run measurement overhead: 90 nsecs sleep measurement overhead: 2724743 nsecs the run test took 1000081 nsecs the sleep test took 2981111 nsecs version = 0.5 ... nr_run_events: 70 nr_sleep_events: 66 nr_wakeup_events: 9 target-less wakeups: 71 multi-target wakeups: 47 run events optimized: 139 task 0 ( perf: 6607), nr_events: 2 task 1 ( perf: 6608), nr_events: 6 task 2 ( : 0), nr_events: 1 task 3 ( make: 6609), nr_events: 5 task 4 ( sh: 6610), nr_events: 4 task 5 ( make: 6611), nr_events: 6 task 6 ( sh: 6612), nr_events: 4 task 7 ( make: 6613), nr_events: 5 task 8 ( migration/11: 25), nr_events: 1 task 9 ( migration/13: 29), nr_events: 1 task 10 ( migration/15: 33), nr_events: 1 task 11 ( migration/9: 21), nr_events: 1 task 12 ( sh: 6614), nr_events: 4 task 13 ( make: 6615), nr_events: 5 task 14 ( sh: 6616), nr_events: 4 task 15 ( make: 6617), nr_events: 7 task 16 ( migration/3: 9), nr_events: 1 task 17 ( migration/5: 13), nr_events: 1 task 18 ( migration/7: 17), nr_events: 1 task 19 ( migration/1: 5), nr_events: 1 task 20 ( sh: 6618), nr_events: 4 task 21 ( make: 6619), nr_events: 5 task 22 ( sh: 6620), nr_events: 4 task 23 ( make: 6621), nr_events: 10 task 24 ( sh: 6623), nr_events: 3 task 25 ( gcc: 6624), nr_events: 4 task 26 ( gcc: 6625), nr_events: 4 task 27 ( gcc: 6626), nr_events: 5 task 28 ( collect2: 6627), nr_events: 5 task 29 ( sh: 6622), nr_events: 1 task 30 ( make: 6628), nr_events: 7 task 31 ( sh: 6630), nr_events: 4 task 32 ( gcc: 6631), nr_events: 4 task 33 ( sh: 6629), nr_events: 1 task 34 ( gcc: 6632), nr_events: 4 task 35 ( gcc: 6633), nr_events: 4 task 36 ( collect2: 6634), nr_events: 4 task 37 ( make: 6635), nr_events: 8 task 38 ( sh: 6637), nr_events: 4 task 39 ( sh: 6636), nr_events: 1 task 40 ( gcc: 6638), nr_events: 4 task 41 ( gcc: 6639), nr_events: 4 task 42 ( gcc: 6640), nr_events: 4 task 43 ( collect2: 6641), nr_events: 4 task 44 ( make: 6642), nr_events: 6 task 45 ( sh: 6643), nr_events: 5 task 46 ( sh: 6644), nr_events: 3 task 47 ( sh: 6645), nr_events: 4 task 48 ( make: 6646), nr_events: 6 task 49 ( sh: 6647), nr_events: 3 task 50 ( make: 6648), nr_events: 5 task 51 ( sh: 6649), nr_events: 5 task 52 ( sh: 6650), nr_events: 6 task 53 ( make: 6651), nr_events: 4 task 54 ( make: 6652), nr_events: 5 task 55 ( make: 6653), nr_events: 4 task 56 ( make: 6654), nr_events: 4 task 57 ( make: 6655), nr_events: 5 task 58 ( sh: 6656), nr_events: 4 task 59 ( gcc: 6657), nr_events: 9 task 60 ( ksoftirqd/3: 10), nr_events: 1 task 61 ( gcc: 6658), nr_events: 4 task 62 ( make: 6659), nr_events: 5 task 63 ( sh: 6660), nr_events: 3 task 64 ( gcc: 6661), nr_events: 5 task 65 ( collect2: 6662), nr_events: 4 ------------------------------------------------------------ #1 : 256.745, ravg: 256.74, cpu: 0.00 / 0.00 #2 : 439.372, ravg: 275.01, cpu: 0.00 / 0.00 #3 : 411.971, ravg: 288.70, cpu: 0.00 / 0.00 #4 : 385.500, ravg: 298.38, cpu: 0.00 / 0.00 #5 : 366.526, ravg: 305.20, cpu: 0.00 / 0.00 #6 : 381.281, ravg: 312.81, cpu: 0.00 / 0.00 #7 : 410.756, ravg: 322.60, cpu: 0.00 / 0.00 #8 : 368.009, ravg: 327.14, cpu: 0.00 / 0.00 #9 : 408.098, ravg: 335.24, cpu: 0.00 / 0.00 #10 : 368.582, ravg: 338.57, cpu: 0.00 / 0.00 I.e. we successfully analyzed the trace, replayed it via real threads and measured the replayed workload's scheduling properties. This is how it looked like in 'top' output: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7164 mingo 20 0 1434m 8080 888 R 57.0 0.1 0:02.04 :perf 7165 mingo 20 0 1434m 8080 888 R 41.8 0.1 0:01.52 :perf 7228 mingo 20 0 1434m 8080 888 R 39.8 0.1 0:01.44 :gcc 7225 mingo 20 0 1434m 8080 888 R 33.8 0.1 0:01.26 :gcc 7202 mingo 20 0 1434m 8080 888 R 31.2 0.1 0:01.16 :sh 7222 mingo 20 0 1434m 8080 888 R 25.2 0.1 0:00.96 :sh 7211 mingo 20 0 1434m 8080 888 R 21.9 0.1 0:00.82 :sh 7213 mingo 20 0 1434m 8080 888 D 19.2 0.1 0:00.74 :sh 7194 mingo 20 0 1434m 8080 888 D 18.6 0.1 0:00.72 :make There's still various kinks in it - more patches to come. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-11 10:12:54 +00:00
if (!strcmp(event->name, "sched_process_exit"))
process_sched_exit_event(event, cpu, timestamp, thread);
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
if (!strcmp(event->name, "sched_migrate_task"))
process_sched_migrate_task_event(data, session, event, cpu, timestamp, thread);
}
static int process_sample_event(union perf_event *event,
struct perf_sample *sample,
struct perf_evsel *evsel __used,
struct perf_session *session)
{
struct thread *thread;
if (!(session->sample_type & PERF_SAMPLE_RAW))
return 0;
thread = perf_session__findnew(session, sample->pid);
if (thread == NULL) {
pr_debug("problem processing %d event, skipping it.\n",
event->header.type);
return -1;
}
dump_printf(" ... thread: %s:%d\n", thread->comm, thread->pid);
if (profile_cpu != -1 && profile_cpu != (int)sample->cpu)
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
return 0;
process_raw_event(event, session, sample->raw_data, sample->cpu,
sample->time, thread);
return 0;
}
static struct perf_event_ops event_ops = {
.sample = process_sample_event,
.comm = perf_event__process_comm,
.lost = perf_event__process_lost,
.fork = perf_event__process_task,
.ordered_samples = true,
};
static int read_events(void)
{
int err = -EINVAL;
perf session: Fallback to unordered processing if no sample_id_all If we are running the new perf on an old kernel without support for sample_id_all, we should fall back to the old unordered processing of events. If we didn't than we would *always* process events without timestamps out of order, whether or not we hit a reordering race. In other words, instead of there being a chance of not attributing samples correctly, we would guarantee that samples would not be attributed. While processing all events without timestamps before events with timestamps may seem like an intuitive solution, it falls down as PERF_RECORD_EXIT events would also be processed before any samples. Even with a workaround for that case, samples before/after an exec would not be attributed correctly. This patch allows commands to indicate whether they need to fall back to unordered processing, so that commands that do not care about timestamps on every event will not be affected. If we do fallback, this will print out a warning if report -D was invoked. This patch adds the test in perf_session__new so that we only need to test once per session. Commands that do not use an event_ops (such as record and top) can simply pass NULL in it's place. Acked-by: Thomas Gleixner <tglx@linutronix.de> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> LKML-Reference: <1291951882-sup-6069@au1.ibm.com> Signed-off-by: Ian Munsie <imunsie@au1.ibm.com> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2010-12-10 03:09:16 +00:00
struct perf_session *session = perf_session__new(input_name, O_RDONLY,
0, false, &event_ops);
if (session == NULL)
return -ENOMEM;
if (perf_session__has_traces(session, "record -R")) {
err = perf_session__process_events(session, &event_ops);
nr_events = session->hists.stats.nr_events[0];
nr_lost_events = session->hists.stats.total_lost;
nr_lost_chunks = session->hists.stats.nr_events[PERF_RECORD_LOST];
}
perf_session__delete(session);
return err;
}
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
static void print_bad_events(void)
{
if (nr_unordered_timestamps && nr_timestamps) {
printf(" INFO: %.3f%% unordered timestamps (%ld out of %ld)\n",
(double)nr_unordered_timestamps/(double)nr_timestamps*100.0,
nr_unordered_timestamps, nr_timestamps);
}
if (nr_lost_events && nr_events) {
printf(" INFO: %.3f%% lost events (%ld out of %ld, in %ld chunks)\n",
(double)nr_lost_events/(double)nr_events*100.0,
nr_lost_events, nr_events, nr_lost_chunks);
}
if (nr_state_machine_bugs && nr_timestamps) {
printf(" INFO: %.3f%% state machine bugs (%ld out of %ld)",
(double)nr_state_machine_bugs/(double)nr_timestamps*100.0,
nr_state_machine_bugs, nr_timestamps);
if (nr_lost_events)
printf(" (due to lost events?)");
printf("\n");
}
if (nr_context_switch_bugs && nr_timestamps) {
printf(" INFO: %.3f%% context switch bugs (%ld out of %ld)",
(double)nr_context_switch_bugs/(double)nr_timestamps*100.0,
nr_context_switch_bugs, nr_timestamps);
if (nr_lost_events)
printf(" (due to lost events?)");
printf("\n");
}
}
static void __cmd_lat(void)
{
struct rb_node *next;
setup_pager();
read_events();
sort_lat();
printf("\n ---------------------------------------------------------------------------------------------------------------\n");
printf(" Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | Maximum delay at |\n");
printf(" ---------------------------------------------------------------------------------------------------------------\n");
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
next = rb_first(&sorted_atom_root);
while (next) {
struct work_atoms *work_list;
work_list = rb_entry(next, struct work_atoms, node);
output_lat_thread(work_list);
next = rb_next(next);
}
printf(" -----------------------------------------------------------------------------------------\n");
printf(" TOTAL: |%11.3f ms |%9" PRIu64 " |\n",
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
(double)all_runtime/1e6, all_count);
printf(" ---------------------------------------------------\n");
print_bad_events();
printf("\n");
}
static struct trace_sched_handler map_ops = {
.wakeup_event = NULL,
.switch_event = map_switch_event,
.runtime_event = NULL,
.fork_event = NULL,
};
static void __cmd_map(void)
{
max_cpu = sysconf(_SC_NPROCESSORS_CONF);
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
setup_pager();
read_events();
print_bad_events();
}
static void __cmd_replay(void)
{
unsigned long i;
calibrate_run_measurement_overhead();
calibrate_sleep_measurement_overhead();
test_calibrations();
read_events();
printf("nr_run_events: %ld\n", nr_run_events);
printf("nr_sleep_events: %ld\n", nr_sleep_events);
printf("nr_wakeup_events: %ld\n", nr_wakeup_events);
if (targetless_wakeups)
printf("target-less wakeups: %ld\n", targetless_wakeups);
if (multitarget_wakeups)
printf("multi-target wakeups: %ld\n", multitarget_wakeups);
if (nr_run_events_optimized)
printf("run atoms optimized: %ld\n",
nr_run_events_optimized);
print_task_traces();
add_cross_task_wakeups();
create_tasks();
printf("------------------------------------------------------------\n");
for (i = 0; i < replay_repeat; i++)
run_one_test();
}
static const char * const sched_usage[] = {
"perf sched [<options>] {record|latency|map|replay|trace}",
NULL
};
static const struct option sched_options[] = {
OPT_STRING('i', "input", &input_name, "file",
"input file name"),
perf: Fix endianness argument compatibility with OPT_BOOLEAN() and introduce OPT_INCR() Parsing an option from the command line with OPT_BOOLEAN on a bool data type would not work on a big-endian machine due to the manner in which the boolean was being cast into an int and incremented. For example, running 'perf probe --list' on a PowerPC machine would fail to properly set the list_events bool and would therefore print out the usage information and terminate. This patch makes OPT_BOOLEAN work as expected with a bool datatype. For cases where the original OPT_BOOLEAN was intentionally being used to increment an int each time it was passed in on the command line, this patch introduces OPT_INCR with the old behaviour of OPT_BOOLEAN (the verbose variable is currently the only such example of this). I have reviewed every use of OPT_BOOLEAN to verify that a true C99 bool was passed. Where integers were used, I verified that they were only being used for boolean logic and changed them to bools to ensure that they would not be mistakenly used as ints. The major exception was the verbose variable which now uses OPT_INCR instead of OPT_BOOLEAN. Signed-off-by: Ian Munsie <imunsie@au.ibm.com> Acked-by: David S. Miller <davem@davemloft.net> Cc: <stable@kernel.org> # NOTE: wont apply to .3[34].x cleanly, please backport Cc: Git development list <git@vger.kernel.org> Cc: Ian Munsie <imunsie@au1.ibm.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> Cc: Rusty Russell <rusty@rustcorp.com.au> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Eric B Munson <ebmunson@us.ibm.com> Cc: Valdis.Kletnieks@vt.edu Cc: WANG Cong <amwang@redhat.com> Cc: Thiago Farina <tfransosi@gmail.com> Cc: Masami Hiramatsu <mhiramat@redhat.com> Cc: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com> Cc: Jaswinder Singh Rajput <jaswinderrajput@gmail.com> Cc: Arjan van de Ven <arjan@linux.intel.com> Cc: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp> Cc: Mike Galbraith <efault@gmx.de> Cc: Tom Zanussi <tzanussi@gmail.com> Cc: Anton Blanchard <anton@samba.org> Cc: John Kacur <jkacur@redhat.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Steven Rostedt <rostedt@goodmis.org> LKML-Reference: <1271147857-11604-1-git-send-email-imunsie@au.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-04-13 08:37:33 +00:00
OPT_INCR('v', "verbose", &verbose,
"be more verbose (show symbol address, etc)"),
OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace,
"dump raw trace in ASCII"),
OPT_END()
};
static const char * const latency_usage[] = {
"perf sched latency [<options>]",
NULL
};
static const struct option latency_options[] = {
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
OPT_STRING('s', "sort", &sort_order, "key[,key2...]",
"sort by key(s): runtime, switch, avg, max"),
perf: Fix endianness argument compatibility with OPT_BOOLEAN() and introduce OPT_INCR() Parsing an option from the command line with OPT_BOOLEAN on a bool data type would not work on a big-endian machine due to the manner in which the boolean was being cast into an int and incremented. For example, running 'perf probe --list' on a PowerPC machine would fail to properly set the list_events bool and would therefore print out the usage information and terminate. This patch makes OPT_BOOLEAN work as expected with a bool datatype. For cases where the original OPT_BOOLEAN was intentionally being used to increment an int each time it was passed in on the command line, this patch introduces OPT_INCR with the old behaviour of OPT_BOOLEAN (the verbose variable is currently the only such example of this). I have reviewed every use of OPT_BOOLEAN to verify that a true C99 bool was passed. Where integers were used, I verified that they were only being used for boolean logic and changed them to bools to ensure that they would not be mistakenly used as ints. The major exception was the verbose variable which now uses OPT_INCR instead of OPT_BOOLEAN. Signed-off-by: Ian Munsie <imunsie@au.ibm.com> Acked-by: David S. Miller <davem@davemloft.net> Cc: <stable@kernel.org> # NOTE: wont apply to .3[34].x cleanly, please backport Cc: Git development list <git@vger.kernel.org> Cc: Ian Munsie <imunsie@au1.ibm.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> Cc: Rusty Russell <rusty@rustcorp.com.au> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Eric B Munson <ebmunson@us.ibm.com> Cc: Valdis.Kletnieks@vt.edu Cc: WANG Cong <amwang@redhat.com> Cc: Thiago Farina <tfransosi@gmail.com> Cc: Masami Hiramatsu <mhiramat@redhat.com> Cc: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com> Cc: Jaswinder Singh Rajput <jaswinderrajput@gmail.com> Cc: Arjan van de Ven <arjan@linux.intel.com> Cc: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp> Cc: Mike Galbraith <efault@gmx.de> Cc: Tom Zanussi <tzanussi@gmail.com> Cc: Anton Blanchard <anton@samba.org> Cc: John Kacur <jkacur@redhat.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Steven Rostedt <rostedt@goodmis.org> LKML-Reference: <1271147857-11604-1-git-send-email-imunsie@au.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-04-13 08:37:33 +00:00
OPT_INCR('v', "verbose", &verbose,
"be more verbose (show symbol address, etc)"),
perf sched: Add -C option to measure on a specific CPU To refresh, trying to sched record only one CPU results in bogus latencies as below. I fixed^Wmade it stop doing the bad thing today, by following task migration events properly. Before: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- Xorg:4943 | 1.290 ms | 1 | avg: 1670.132 ms | max: 1670.132 ms | hald-addon-stor:3569 | 0.091 ms | 3 | avg: 658.609 ms | max: 1975.797 ms | hald-addon-stor:3573 | 0.209 ms | 4 | avg: 499.138 ms | max: 1990.565 ms | audispd:4270 | 0.012 ms | 1 | avg: 0.015 ms | max: 0.015 ms | .... marge:/root/tmp # perf sched trace|grep 'Xorg:4943' swapper-0 [000] 401.184013288: sched_stat_runtime: task: Xorg:4943 runtime: 1233188 [ns], vruntime: 19105169779 [ns] rt2870TimerQHan-4947 [000] 402.854140127: sched_stat_wait: task: Xorg:4943 wait: 580073 [ns] rt2870TimerQHan-4947 [000] 402.854141770: sched_migrate_task: task Xorg:4943 [140] from: 1 to: 0 rt2870TimerQHan-4947 [000] 402.854143854: sched_stat_wait: task: Xorg:4943 wait: 0 [ns] rt2870TimerQHan-4947 [000] 402.854145397: sched_switch: task rt2870TimerQHan:4947 [140] (D) ==> Xorg:4943 [140] Xorg-4943 [000] 402.854193133: sched_stat_runtime: task: Xorg:4943 runtime: 56546 [ns], vruntime: 11766332500 [ns] Xorg-4943 [000] 402.854196842: sched_switch: task Xorg:4943 [140] (S) ==> swapper:0 [140] After: marge:/root/tmp # taskset -c 1 perf sched record -C 0 -- sleep 10 marge:/root/tmp # perf sched lat ----------------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------------- amarokapp:11150 | 271.297 ms | 878 | avg: 0.130 ms | max: 1.057 ms | konsole:5965 | 1.370 ms | 12 | avg: 0.092 ms | max: 0.855 ms | Xorg:4943 | 179.980 ms | 1109 | avg: 0.087 ms | max: 1.206 ms | hald-addon-stor:3574 | 0.212 ms | 9 | avg: 0.040 ms | max: 0.169 ms | hald-addon-stor:3570 | 0.223 ms | 9 | avg: 0.037 ms | max: 0.223 ms | klauncher:5864 | 0.550 ms | 8 | avg: 0.032 ms | max: 0.048 ms | The 'Maximum delay ms' results are now sane. Signed-off-by: Mike Galbraith <efault@gmx.de> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-10-10 12:46:04 +00:00
OPT_INTEGER('C', "CPU", &profile_cpu,
"CPU to profile on"),
OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace,
"dump raw trace in ASCII"),
OPT_END()
};
static const char * const replay_usage[] = {
"perf sched replay [<options>]",
NULL
};
static const struct option replay_options[] = {
OPT_UINTEGER('r', "repeat", &replay_repeat,
"repeat the workload replay N times (-1: infinite)"),
perf: Fix endianness argument compatibility with OPT_BOOLEAN() and introduce OPT_INCR() Parsing an option from the command line with OPT_BOOLEAN on a bool data type would not work on a big-endian machine due to the manner in which the boolean was being cast into an int and incremented. For example, running 'perf probe --list' on a PowerPC machine would fail to properly set the list_events bool and would therefore print out the usage information and terminate. This patch makes OPT_BOOLEAN work as expected with a bool datatype. For cases where the original OPT_BOOLEAN was intentionally being used to increment an int each time it was passed in on the command line, this patch introduces OPT_INCR with the old behaviour of OPT_BOOLEAN (the verbose variable is currently the only such example of this). I have reviewed every use of OPT_BOOLEAN to verify that a true C99 bool was passed. Where integers were used, I verified that they were only being used for boolean logic and changed them to bools to ensure that they would not be mistakenly used as ints. The major exception was the verbose variable which now uses OPT_INCR instead of OPT_BOOLEAN. Signed-off-by: Ian Munsie <imunsie@au.ibm.com> Acked-by: David S. Miller <davem@davemloft.net> Cc: <stable@kernel.org> # NOTE: wont apply to .3[34].x cleanly, please backport Cc: Git development list <git@vger.kernel.org> Cc: Ian Munsie <imunsie@au1.ibm.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> Cc: Rusty Russell <rusty@rustcorp.com.au> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Eric B Munson <ebmunson@us.ibm.com> Cc: Valdis.Kletnieks@vt.edu Cc: WANG Cong <amwang@redhat.com> Cc: Thiago Farina <tfransosi@gmail.com> Cc: Masami Hiramatsu <mhiramat@redhat.com> Cc: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com> Cc: Jaswinder Singh Rajput <jaswinderrajput@gmail.com> Cc: Arjan van de Ven <arjan@linux.intel.com> Cc: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp> Cc: Mike Galbraith <efault@gmx.de> Cc: Tom Zanussi <tzanussi@gmail.com> Cc: Anton Blanchard <anton@samba.org> Cc: John Kacur <jkacur@redhat.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Steven Rostedt <rostedt@goodmis.org> LKML-Reference: <1271147857-11604-1-git-send-email-imunsie@au.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-04-13 08:37:33 +00:00
OPT_INCR('v', "verbose", &verbose,
"be more verbose (show symbol address, etc)"),
OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace,
"dump raw trace in ASCII"),
OPT_END()
};
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
static void setup_sorting(void)
{
char *tmp, *tok, *str = strdup(sort_order);
for (tok = strtok_r(str, ", ", &tmp);
tok; tok = strtok_r(NULL, ", ", &tmp)) {
if (sort_dimension__add(tok, &sort_list) < 0) {
error("Unknown --sort key: `%s'", tok);
usage_with_options(latency_usage, latency_options);
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
}
}
free(str);
sort_dimension__add("pid", &cmp_pid);
perf sched: Implement multidimensional sorting Implement multidimensional sorting on perf sched so that you can sort either by number of switches, latency average, latency maximum, runtime. perf sched -l -s avg,max (this is the default) ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- gnome-power-man | 0.113 ms | 1 | avg: 4998.531 ms | max: 4998.531 ms | xfdesktop | 1.190 ms | 7 | avg: 136.475 ms | max: 940.933 ms | xfce-mcs-manage | 2.194 ms | 22 | avg: 38.534 ms | max: 735.174 ms | notification-da | 2.749 ms | 31 | avg: 27.436 ms | max: 731.791 ms | xfce4-session | 3.343 ms | 28 | avg: 26.796 ms | max: 734.891 ms | xfwm4 | 3.159 ms | 22 | avg: 12.406 ms | max: 241.333 ms | xchat | 42.789 ms | 214 | avg: 11.886 ms | max: 100.349 ms | xfce4-terminal | 5.386 ms | 22 | avg: 11.414 ms | max: 241.611 ms | firefox | 151.992 ms | 123 | avg: 9.543 ms | max: 153.717 ms | xfce4-panel | 24.324 ms | 47 | avg: 8.189 ms | max: 242.352 ms | :5090 | 6.932 ms | 111 | avg: 8.131 ms | max: 102.665 ms | events/0 | 0.758 ms | 12 | avg: 1.964 ms | max: 21.879 ms | Xorg | 280.558 ms | 340 | avg: 1.864 ms | max: 99.526 ms | geany | 63.391 ms | 295 | avg: 1.099 ms | max: 9.334 ms | reiserfs/0 | 0.039 ms | 2 | avg: 0.854 ms | max: 1.487 ms | kondemand/0 | 8.251 ms | 245 | avg: 0.691 ms | max: 34.372 ms | Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 01:36:29 +00:00
}
static const char *record_args[] = {
"record",
"-a",
"-R",
perf tools: Implement counter output multiplexing Finish the -M/--multiplex option implementation: - separate it out from group_fd - correctly set it via the ioctl and dont mmap counters that are multiplexed - modify the perf record event loop to deal with buffer-less counters. - remove the -g option from perf sched record - account for unordered events in perf sched latency - (add -f to perf sched record to ease measurements) - skip idle threads (pid==0) in latency output The result is better latency output by 'perf sched latency': ----------------------------------------------------------------------------------- Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | ----------------------------------------------------------------------------------- ksoftirqd/8 | 0.071 ms | 2 | avg: 0.458 ms | max: 0.913 ms | at-spi-registry | 0.609 ms | 19 | avg: 0.013 ms | max: 0.023 ms | perf | 3.316 ms | 16 | avg: 0.013 ms | max: 0.054 ms | Xorg | 0.392 ms | 19 | avg: 0.011 ms | max: 0.018 ms | sleep | 0.537 ms | 2 | avg: 0.009 ms | max: 0.009 ms | ----------------------------------------------------------------------------------- TOTAL: | 4.925 ms | 58 | --------------------------------------------- Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-13 16:15:54 +00:00
"-f",
"-m", "1024",
"-c", "1",
"-e", "sched:sched_switch",
"-e", "sched:sched_stat_wait",
"-e", "sched:sched_stat_sleep",
"-e", "sched:sched_stat_iowait",
"-e", "sched:sched_stat_runtime",
"-e", "sched:sched_process_exit",
"-e", "sched:sched_process_fork",
"-e", "sched:sched_wakeup",
"-e", "sched:sched_migrate_task",
};
static int __cmd_record(int argc, const char **argv)
{
unsigned int rec_argc, i, j;
const char **rec_argv;
rec_argc = ARRAY_SIZE(record_args) + argc - 1;
rec_argv = calloc(rec_argc + 1, sizeof(char *));
if (rec_argv == NULL)
return -ENOMEM;
for (i = 0; i < ARRAY_SIZE(record_args); i++)
rec_argv[i] = strdup(record_args[i]);
for (j = 1; j < (unsigned int)argc; j++, i++)
rec_argv[i] = argv[j];
BUG_ON(i != rec_argc);
return cmd_record(i, rec_argv, NULL);
}
int cmd_sched(int argc, const char **argv, const char *prefix __used)
{
argc = parse_options(argc, argv, sched_options, sched_usage,
PARSE_OPT_STOP_AT_NON_OPTION);
if (!argc)
usage_with_options(sched_usage, sched_options);
/*
* Aliased to 'perf script' for now:
*/
if (!strcmp(argv[0], "script"))
return cmd_script(argc, argv, prefix);
symbol__init();
if (!strncmp(argv[0], "rec", 3)) {
return __cmd_record(argc, argv);
} else if (!strncmp(argv[0], "lat", 3)) {
trace_handler = &lat_ops;
if (argc > 1) {
argc = parse_options(argc, argv, latency_options, latency_usage, 0);
if (argc)
usage_with_options(latency_usage, latency_options);
}
setup_sorting();
__cmd_lat();
perf sched: Add 'perf sched map' scheduling event map printout This prints a textual context-switching outline of workload captured via perf sched record. For example, on a 16 CPU box it outputs: N1 O1 . . . S1 . . . B0 . *I0 C1 . M1 . 23002.773423 secs N1 O1 . *Q0 . S1 . . . B0 . I0 C1 . M1 . 23002.773423 secs N1 O1 . Q0 . S1 . . . B0 . *R1 C1 . M1 . 23002.773485 secs N1 O1 . Q0 . S1 . *S0 . B0 . R1 C1 . M1 . 23002.773478 secs *L0 O1 . Q0 . S1 . S0 . B0 . R1 C1 . M1 . 23002.773523 secs L0 O1 . *. . S1 . S0 . B0 . R1 C1 . M1 . 23002.773531 secs L0 O1 . . . S1 . S0 . B0 . R1 C1 *T1 M1 . 23002.773547 secs T1 => irqbalance:2089 L0 O1 . . . S1 . S0 . *P0 . R1 C1 T1 M1 . 23002.773549 secs *N1 O1 . . . S1 . S0 . P0 . R1 C1 T1 M1 . 23002.773566 secs N1 O1 . . . *J0 . S0 . P0 . R1 C1 T1 M1 . 23002.773571 secs N1 O1 . . . J0 . S0 *B0 P0 . R1 C1 T1 M1 . 23002.773592 secs N1 O1 . . . J0 . *U0 B0 P0 . R1 C1 T1 M1 . 23002.773582 secs N1 O1 . . . *S1 . U0 B0 P0 . R1 C1 T1 M1 . 23002.773604 secs N1 O1 . . . S1 . U0 B0 *. . R1 C1 T1 M1 . 23002.773615 secs N1 O1 . . . S1 . U0 B0 . . *K0 C1 T1 M1 . 23002.773631 secs N1 O1 . *M0 . S1 . U0 B0 . . K0 C1 T1 M1 . 23002.773624 secs N1 O1 . M0 . S1 . U0 *. . . K0 C1 T1 M1 . 23002.773644 secs N1 O1 . M0 . S1 . U0 . . . *R1 C1 T1 M1 . 23002.773662 secs N1 O1 . M0 . S1 . *. . . . R1 C1 T1 M1 . 23002.773648 secs N1 O1 . *. . S1 . . . . . R1 C1 T1 M1 . 23002.773680 secs N1 O1 . . . *L0 . . . . . R1 C1 T1 M1 . 23002.773717 secs *N0 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773709 secs *N1 O1 . . . L0 . . . . . R1 C1 T1 M1 . 23002.773747 secs Columns stand for individual CPUs, from CPU0 to CPU15, and the two-letter shortcuts stand for tasks that are running on a CPU. '*' denotes the CPU that had the event. A dot signals an idle CPU. New tasks are assigned new two-letter shortcuts - when they occur first they are printed. In the above example 'T1' stood for irqbalance: T1 => irqbalance:2089 Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-16 15:40:48 +00:00
} else if (!strcmp(argv[0], "map")) {
trace_handler = &map_ops;
setup_sorting();
__cmd_map();
} else if (!strncmp(argv[0], "rep", 3)) {
trace_handler = &replay_ops;
if (argc) {
argc = parse_options(argc, argv, replay_options, replay_usage, 0);
if (argc)
usage_with_options(replay_usage, replay_options);
}
__cmd_replay();
} else {
usage_with_options(sched_usage, sched_options);
}
return 0;
}