perf sched: Use perf_evsel__{int,str}val
authorArnaldo Carvalho de Melo <acme@redhat.com>
Tue, 11 Sep 2012 22:29:17 +0000 (19:29 -0300)
committerArnaldo Carvalho de Melo <acme@redhat.com>
Tue, 11 Sep 2012 22:33:51 +0000 (19:33 -0300)
This patch also stops reading the common fields, as they were not being used except
for one ->common_pid case that was replaced by sample->tid, i.e. the info is already
in the perf_sample struct.

Also it only fills the _event structures when there is a handler.

  [root@sandy ~]# perf sched record sleep 30s
  [ perf record: Woken up 1 times to write data ]
  [ perf record: Captured and wrote 8.585 MB perf.data (~375063 samples) ]

Before:

  [root@sandy ~]# perf stat -r 10 perf sched lat > /dev/null

   Performance counter stats for 'perf sched lat' (10 runs):

          129.117838 task-clock                #    0.994 CPUs utilized            ( +-  0.28% )
                  14 context-switches          #    0.111 K/sec                    ( +-  2.10% )
                   0 cpu-migrations            #    0.002 K/sec                    ( +- 66.67% )
               7,654 page-faults               #    0.059 M/sec                    ( +-  0.67% )
         438,121,661 cycles                    #    3.393 GHz                      ( +-  0.06% ) [83.06%]
         150,808,605 stalled-cycles-frontend   #   34.42% frontend cycles idle     ( +-  0.14% ) [83.10%]
          80,748,941 stalled-cycles-backend    #   18.43% backend  cycles idle     ( +-  0.64% ) [66.73%]
         758,605,879 instructions              #    1.73  insns per cycle
                                               #    0.20  stalled cycles per insn  ( +-  0.08% ) [83.54%]
         162,164,321 branches                  # 1255.940 M/sec                    ( +-  0.10% ) [83.70%]
           1,609,903 branch-misses             #    0.99% of all branches          ( +-  0.08% ) [83.62%]

         0.129949153 seconds time elapsed                                          ( +-  0.28% )

After:

  [root@sandy ~]# perf stat -r 10 perf sched lat > /dev/null

   Performance counter stats for 'perf sched lat' (10 runs):

          103.592215 task-clock                #    0.993 CPUs utilized            ( +-  0.33% )
                  12 context-switches          #    0.114 K/sec                    ( +-  3.29% )
                   0 cpu-migrations            #    0.000 K/sec
               7,605 page-faults               #    0.073 M/sec                    ( +-  0.00% )
         345,796,112 cycles                    #    3.338 GHz                      ( +-  0.07% ) [82.90%]
         106,876,796 stalled-cycles-frontend   #   30.91% frontend cycles idle     ( +-  0.38% ) [83.23%]
          62,060,877 stalled-cycles-backend    #   17.95% backend  cycles idle     ( +-  0.80% ) [67.14%]
         628,246,586 instructions              #    1.82  insns per cycle
                                               #    0.17  stalled cycles per insn  ( +-  0.04% ) [83.64%]
         134,962,057 branches                  # 1302.820 M/sec                    ( +-  0.10% ) [83.64%]
           1,233,037 branch-misses             #    0.91% of all branches          ( +-  0.29% ) [83.41%]

         0.104333272 seconds time elapsed                                          ( +-  0.33% )

  [root@sandy ~]#

Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-weu9t63zkrfrazkn0gxj48xy@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
tools/perf/builtin-sched.c

index 79f88fa..0df5e7a 100644 (file)
@@ -98,82 +98,40 @@ struct work_atoms {
 typedef int (*sort_fn_t)(struct work_atoms *, struct work_atoms *);
 
 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];
+       char *prev_comm;
        u32  prev_pid;
        u32  prev_prio;
        u64  prev_state;
-       char next_comm[16];
+       char *next_comm;
        u32  next_pid;
        u32  next_prio;
 };
 
 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];
+       char *comm;
        u32  pid;
        u64  runtime;
        u64  vruntime;
 };
 
 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];
+       char *comm;
        u32  pid;
-
        u32  prio;
        u32  success;
        u32  cpu;
 };
 
 struct trace_fork_event {
-       u32  size;
-
-       u16  common_type;
-       u8   common_flags;
-       u8   common_preempt_count;
-       u32  common_pid;
-       u32  common_tgid;
-
-       char parent_comm[16];
+       char *parent_comm;
        u32  parent_pid;
-       char child_comm[16];
+       char *child_comm;
        u32   child_pid;
 };
 
 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];
+       char *comm;
        u32  pid;
-
        u32  prio;
        u32  cpu;
 };
@@ -184,7 +142,7 @@ struct trace_sched_handler {
        int (*switch_event)(struct perf_sched *sched,
                            struct trace_switch_event *event,
                            struct machine *machine,
-                           struct event_format *tp_format,
+                           struct perf_evsel *evsel,
                            struct perf_sample *sample);
 
        int (*runtime_event)(struct perf_sched *sched,
@@ -195,12 +153,12 @@ struct trace_sched_handler {
        int (*wakeup_event)(struct perf_sched *sched,
                            struct trace_wakeup_event *event,
                            struct machine *machine,
-                           struct event_format *tp_format,
+                           struct perf_evsel *evsel,
                            struct perf_sample *sample);
 
        int (*fork_event)(struct perf_sched *sched,
                          struct trace_fork_event *event,
-                         struct event_format *tp_format);
+                         struct perf_evsel *evsel);
 
        int (*migrate_task_event)(struct perf_sched *sched,
                                  struct trace_migrate_task_event *event,
@@ -740,40 +698,22 @@ static void test_calibrations(struct perf_sched *sched)
        printf("the sleep test took %" PRIu64 " nsecs\n", T1 - T0);
 }
 
-#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)
-
 static int
 replay_wakeup_event(struct perf_sched *sched,
                    struct trace_wakeup_event *wakeup_event,
                    struct machine *machine __maybe_unused,
-                   struct event_format *event, struct perf_sample *sample)
+                   struct perf_evsel *evsel, struct perf_sample *sample)
 {
        struct task_desc *waker, *wakee;
 
        if (verbose) {
-               printf("sched_wakeup event %p\n", event);
+               printf("sched_wakeup event %p\n", evsel);
 
                printf(" ... pid %d woke up %s/%d\n",
-                      wakeup_event->common_pid, wakeup_event->comm, wakeup_event->pid);
+                      sample->tid, wakeup_event->comm, wakeup_event->pid);
        }
 
-       waker = register_pid(sched, wakeup_event->common_pid, "<unknown>");
+       waker = register_pid(sched, sample->tid, "<unknown>");
        wakee = register_pid(sched, wakeup_event->pid, wakeup_event->comm);
 
        add_sched_event_wakeup(sched, waker, sample->time, wakee);
@@ -784,7 +724,7 @@ static int
 replay_switch_event(struct perf_sched *sched,
                    struct trace_switch_event *switch_event,
                    struct machine *machine __maybe_unused,
-                   struct event_format *event,
+                   struct perf_evsel *evsel,
                    struct perf_sample *sample)
 {
        struct task_desc *prev, __maybe_unused *next;
@@ -793,7 +733,7 @@ replay_switch_event(struct perf_sched *sched,
        s64 delta;
 
        if (verbose)
-               printf("sched_switch event %p\n", event);
+               printf("sched_switch event %p\n", evsel);
 
        if (cpu >= MAX_CPUS || cpu < 0)
                return 0;
@@ -829,10 +769,10 @@ replay_switch_event(struct perf_sched *sched,
 
 static int
 replay_fork_event(struct perf_sched *sched, struct trace_fork_event *fork_event,
-                 struct event_format *event)
+                 struct perf_evsel *evsel)
 {
        if (verbose) {
-               printf("sched_fork event %p\n", event);
+               printf("sched_fork event %p\n", evsel);
                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);
        }
@@ -931,7 +871,7 @@ static int thread_atoms_insert(struct perf_sched *sched, struct thread *thread)
 
 static int latency_fork_event(struct perf_sched *sched __maybe_unused,
                              struct trace_fork_event *fork_event __maybe_unused,
-                             struct event_format *event __maybe_unused)
+                             struct perf_evsel *evsel __maybe_unused)
 {
        /* should insert the newcomer */
        return 0;
@@ -1015,7 +955,7 @@ static int
 latency_switch_event(struct perf_sched *sched,
                     struct trace_switch_event *switch_event,
                     struct machine *machine,
-                    struct event_format *event __maybe_unused,
+                    struct perf_evsel *evsel __maybe_unused,
                     struct perf_sample *sample)
 {
        struct work_atoms *out_events, *in_events;
@@ -1106,7 +1046,7 @@ static int
 latency_wakeup_event(struct perf_sched *sched,
                     struct trace_wakeup_event *wakeup_event,
                     struct machine *machine,
-                    struct event_format *event __maybe_unused,
+                    struct perf_evsel *evsel __maybe_unused,
                     struct perf_sample *sample)
 {
        struct work_atoms *atoms;
@@ -1350,34 +1290,32 @@ static void perf_sched__sort_lat(struct perf_sched *sched)
 }
 
 static int process_sched_wakeup_event(struct perf_tool *tool,
-                                     struct event_format *event,
+                                     struct perf_evsel *evsel,
                                      struct perf_sample *sample,
                                      struct machine *machine)
 {
        struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
-       void *data = sample->raw_data;
-       struct trace_wakeup_event wakeup_event;
-       int err = 0;
-
-       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);
+       if (sched->tp_handler->wakeup_event) {
+               struct trace_wakeup_event event = {
+                       .comm    = perf_evsel__strval(evsel, sample, "comm"),
+                       .pid     = perf_evsel__intval(evsel, sample, "pid"),
+                       .prio    = perf_evsel__intval(evsel, sample, "prio"),
+                       .success = perf_evsel__intval(evsel, sample, "success"),
+                       .cpu     = perf_evsel__intval(evsel, sample, "cpu"),
+               };
 
-       if (sched->tp_handler->wakeup_event)
-               err = sched->tp_handler->wakeup_event(sched, &wakeup_event, machine, event, sample);
+               return sched->tp_handler->wakeup_event(sched, &event, machine, evsel, sample);
+       }
 
-       return err;
+       return 0;
 }
 
 static int
 map_switch_event(struct perf_sched *sched,
                 struct trace_switch_event *switch_event,
                 struct machine *machine,
-                struct event_format *event __maybe_unused,
+                struct perf_evsel *evsel __maybe_unused,
                 struct perf_sample *sample)
 {
        struct thread *sched_out __maybe_unused, *sched_in;
@@ -1455,120 +1393,113 @@ map_switch_event(struct perf_sched *sched,
 }
 
 static int process_sched_switch_event(struct perf_tool *tool,
-                                     struct event_format *event,
+                                     struct perf_evsel *evsel,
                                      struct perf_sample *sample,
                                      struct machine *machine)
 {
        struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
        int this_cpu = sample->cpu, err = 0;
-       void *data = sample->raw_data;
-       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);
+       u32 prev_pid = perf_evsel__intval(evsel, sample, "prev_pid"),
+           next_pid = perf_evsel__intval(evsel, sample, "next_pid");
 
        if (sched->curr_pid[this_cpu] != (u32)-1) {
                /*
                 * Are we trying to switch away a PID that is
                 * not current?
                 */
-               if (sched->curr_pid[this_cpu] != switch_event.prev_pid)
+               if (sched->curr_pid[this_cpu] != prev_pid)
                        sched->nr_context_switch_bugs++;
        }
-       if (sched->tp_handler->switch_event)
-               err = sched->tp_handler->switch_event(sched, &switch_event, machine, event, sample);
 
-       sched->curr_pid[this_cpu] = switch_event.next_pid;
+       if (sched->tp_handler->switch_event) {
+               struct trace_switch_event event = {
+                       .prev_comm  = perf_evsel__strval(evsel, sample, "prev_comm"),
+                       .prev_pid   = prev_pid,
+                       .prev_prio  = perf_evsel__intval(evsel, sample, "prev_prio"),
+                       .prev_state = perf_evsel__intval(evsel, sample, "prev_state"),
+                       .next_comm  = perf_evsel__strval(evsel, sample, "next_comm"),
+                       .next_pid   = next_pid,
+                       .next_prio  = perf_evsel__intval(evsel, sample, "next_prio"),
+               };
+
+               err = sched->tp_handler->switch_event(sched, &event, machine, evsel, sample);
+       }
+
+       sched->curr_pid[this_cpu] = next_pid;
        return err;
 }
 
 static int process_sched_runtime_event(struct perf_tool *tool,
-                                      struct event_format *event,
+                                      struct perf_evsel *evsel,
                                       struct perf_sample *sample,
                                       struct machine *machine)
 {
        struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
-       void *data = sample->raw_data;
-       struct trace_runtime_event runtime_event;
-       int err = 0;
 
-       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);
-
-       if (sched->tp_handler->runtime_event)
-               err = sched->tp_handler->runtime_event(sched, &runtime_event, machine, sample);
+       if (sched->tp_handler->runtime_event) {
+               struct trace_runtime_event event = {
+                       .comm     = perf_evsel__strval(evsel, sample, "comm"),
+                       .pid      = perf_evsel__intval(evsel, sample, "pid"),
+                       .runtime  = perf_evsel__intval(evsel, sample, "runtime"),
+                       .vruntime = perf_evsel__intval(evsel, sample, "vruntime"),
+               };
+               return sched->tp_handler->runtime_event(sched, &event, machine, sample);
+       }
 
-       return err;
+       return 0;
 }
 
 static int process_sched_fork_event(struct perf_tool *tool,
-                                   struct event_format *event,
+                                   struct perf_evsel *evsel,
                                    struct perf_sample *sample,
                                    struct machine *machine __maybe_unused)
 {
        struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
-       void *data = sample->raw_data;
-       struct trace_fork_event fork_event;
-       int err = 0;
-
-       FILL_COMMON_FIELDS(fork_event, event, data);
 
-       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);
-
-       if (sched->tp_handler->fork_event)
-               err = sched->tp_handler->fork_event(sched, &fork_event, event);
+       if (sched->tp_handler->fork_event) {
+               struct trace_fork_event event = {
+                       .parent_comm = perf_evsel__strval(evsel, sample, "parent_comm"),
+                       .child_comm  = perf_evsel__strval(evsel, sample, "child_comm"),
+                       .parent_pid  = perf_evsel__intval(evsel, sample, "parent_pid"),
+                       .child_pid  = perf_evsel__intval(evsel, sample, "child_pid"),
+               };
+               return sched->tp_handler->fork_event(sched, &event, evsel);
+       }
 
-       return err;
+       return 0;
 }
 
 static int process_sched_exit_event(struct perf_tool *tool __maybe_unused,
-                                   struct event_format *event,
+                                   struct perf_evsel *evsel,
                                    struct perf_sample *sample __maybe_unused,
                                    struct machine *machine __maybe_unused)
 {
-       if (verbose)
-               printf("sched_exit event %p\n", event);
-
+       pr_debug("sched_exit event %p\n", evsel);
        return 0;
 }
 
 static int process_sched_migrate_task_event(struct perf_tool *tool,
-                                           struct event_format *event,
+                                           struct perf_evsel *evsel,
                                            struct perf_sample *sample,
                                            struct machine *machine)
 {
        struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
-       void *data = sample->raw_data;
-       struct trace_migrate_task_event migrate_task_event;
-       int err = 0;
 
-       FILL_COMMON_FIELDS(migrate_task_event, event, data);
-
-       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);
-
-       if (sched->tp_handler->migrate_task_event)
-               err = sched->tp_handler->migrate_task_event(sched, &migrate_task_event, machine, sample);
+       if (sched->tp_handler->migrate_task_event) {
+               struct trace_migrate_task_event event = {
+                       .comm = perf_evsel__strval(evsel, sample, "comm"),
+                       .pid  = perf_evsel__intval(evsel, sample, "pid"),
+                       .prio = perf_evsel__intval(evsel, sample, "prio"),
+                       .cpu  = perf_evsel__intval(evsel, sample, "cpu"),
+               };
+               return sched->tp_handler->migrate_task_event(sched, &event, machine, sample);
+       }
 
-       return err;
+       return 0;
 }
 
 typedef int (*tracepoint_handler)(struct perf_tool *tool,
-                                 struct event_format *tp_format,
+                                 struct perf_evsel *evsel,
                                  struct perf_sample *sample,
                                  struct machine *machine);
 
@@ -1592,7 +1523,7 @@ static int perf_sched__process_tracepoint_sample(struct perf_tool *tool __maybe_
 
        if (evsel->handler.func != NULL) {
                tracepoint_handler f = evsel->handler.func;
-               err = f(tool, evsel->tp_format, sample, machine);
+               err = f(tool, evsel, sample, machine);
        }
 
        return err;