perf record: Enable counters only when kernel is execing subcommand
authorZhang, Yanmin <yanmin_zhang@linux.intel.com>
Thu, 18 Mar 2010 14:36:04 +0000 (11:36 -0300)
committerIngo Molnar <mingo@elte.hu>
Thu, 18 Mar 2010 15:21:11 +0000 (16:21 +0100)
'perf record' starts counters before subcommand is execed, so
the statistics is not precise because it includes data of some
preparation steps. I fix it with the patch.

In addition, change the condition to fork/exec subcommand. If
there is a subcommand parameter, perf always fork/exec it. The
usage example is:

 # perf record -f -a sleep 10

So this command could collect statistics for 10 seconds
precisely. User still could stop it by CTRL+C. Without the new
capability, user could only input CTRL+C to stop it without
precise time clock.

Signed-off-by: Zhang Yanmin <yanmin_zhang@linux.intel.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Avi Kivity <avi@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Sheng Yang <sheng@linux.intel.com>
Cc: oerg Roedel <joro@8bytes.org>
Cc: Jes Sorensen <Jes.Sorensen@redhat.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Gleb Natapov <gleb@redhat.com>
Cc: <zhiteng.huang@intel.com>
Cc: Zachary Amsden <zamsden@redhat.com>
LKML-Reference: <1268922965-14774-2-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
tools/perf/builtin-record.c

index 962cdbf..e2b35ad 100644 (file)
@@ -225,7 +225,7 @@ static struct perf_header_attr *get_header_attr(struct perf_event_attr *a, int n
        return h_attr;
 }
 
-static void create_counter(int counter, int cpu, pid_t pid, bool forks)
+static void create_counter(int counter, int cpu, pid_t pid)
 {
        char *filter = filters[counter];
        struct perf_event_attr *attr = attrs + counter;
@@ -275,10 +275,10 @@ static void create_counter(int counter, int cpu, pid_t pid, bool forks)
        attr->mmap              = track;
        attr->comm              = track;
        attr->inherit           = inherit;
-       attr->disabled          = 1;
-
-       if (forks)
+       if (target_pid == -1 && !system_wide) {
+               attr->disabled = 1;
                attr->enable_on_exec = 1;
+       }
 
 try_again:
        fd[nr_cpu][counter] = sys_perf_event_open(attr, pid, cpu, group_fd, 0);
@@ -380,17 +380,15 @@ try_again:
                        exit(-1);
                }
        }
-
-       ioctl(fd[nr_cpu][counter], PERF_EVENT_IOC_ENABLE);
 }
 
-static void open_counters(int cpu, pid_t pid, bool forks)
+static void open_counters(int cpu, pid_t pid)
 {
        int counter;
 
        group_fd = -1;
        for (counter = 0; counter < nr_counters; counter++)
-               create_counter(counter, cpu, pid, forks);
+               create_counter(counter, cpu, pid);
 
        nr_cpu++;
 }
@@ -425,7 +423,7 @@ static int __cmd_record(int argc, const char **argv)
        int err;
        unsigned long waking = 0;
        int child_ready_pipe[2], go_pipe[2];
-       const bool forks = target_pid == -1 && argc > 0;
+       const bool forks = argc > 0;
        char buf;
 
        page_size = sysconf(_SC_PAGE_SIZE);
@@ -496,13 +494,13 @@ static int __cmd_record(int argc, const char **argv)
        atexit(atexit_header);
 
        if (forks) {
-               pid = fork();
+               child_pid = fork();
                if (pid < 0) {
                        perror("failed to fork");
                        exit(-1);
                }
 
-               if (!pid) {
+               if (!child_pid) {
                        close(child_ready_pipe[0]);
                        close(go_pipe[1]);
                        fcntl(go_pipe[0], F_SETFD, FD_CLOEXEC);
@@ -531,11 +529,6 @@ static int __cmd_record(int argc, const char **argv)
                        exit(-1);
                }
 
-               child_pid = pid;
-
-               if (!system_wide)
-                       target_pid = pid;
-
                close(child_ready_pipe[1]);
                close(go_pipe[0]);
                /*
@@ -548,13 +541,17 @@ static int __cmd_record(int argc, const char **argv)
                close(child_ready_pipe[0]);
        }
 
+       if (forks && target_pid == -1 && !system_wide)
+               pid = child_pid;
+       else
+               pid = target_pid;
 
        if ((!system_wide && !inherit) || profile_cpu != -1) {
-               open_counters(profile_cpu, target_pid, forks);
+               open_counters(profile_cpu, pid);
        } else {
                nr_cpus = read_cpu_map();
                for (i = 0; i < nr_cpus; i++)
-                       open_counters(cpumap[i], target_pid, forks);
+                       open_counters(cpumap[i], pid);
        }
 
        if (file_new) {