2 files changed, 26 insertions, 1 deletions

diff --git a/tools/perf/Documentation/perf-stat.txt b/tools/perf/Documentation/perf-stat.txt
index 2fe87fb558f0..73c9759005a3 100644
--- a/tools/perf/Documentation/perf-stat.txt
+++ b/tools/perf/Documentation/perf-stat.txt
@@ -132,6 +132,11 @@ is a useful mode to detect imbalance between physical cores.  To enable this mod
 use --per-core in addition to -a. (system-wide).  The output includes the
 core number and the number of online logical processors on that physical processor.
 
+-D msecs::
+--initial-delay msecs::
+After starting the program, wait msecs before measuring. This is useful to
+filter out the startup phase of the program, which is often very different.
+
 EXAMPLES
 --------
 
diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index 352fbd7ff4a1..2e637e4c951d 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -100,6 +100,7 @@ static const char		*pre_cmd			= NULL;
 static const char               *post_cmd                       = NULL;
 static bool                     sync_run                        = false;
 static unsigned int             interval                        = 0;
+static unsigned int             initial_delay                   = 0;
 static bool                     forever                         = false;
 static struct timespec          ref_time;
 static struct cpu_map           *aggr_map;
@@ -254,7 +255,8 @@ static int create_perf_stat_counter(struct perf_evsel *evsel)
         if (!perf_target__has_task(&target) &&
             perf_evsel__is_group_leader(evsel)) {
                 attr->disabled = 1;
-                attr->enable_on_exec = 1;
+                if (!initial_delay)
+                        attr->enable_on_exec = 1;
         }
 
         return perf_evsel__open_per_thread(evsel, evsel_list->threads);
@@ -416,6 +418,20 @@ static void print_interval(void)
         }
 }
 
+static void handle_initial_delay(void)
+{
+        struct perf_evsel *counter;
+
+        if (initial_delay) {
+                const int ncpus = cpu_map__nr(evsel_list->cpus),
+                        nthreads = thread_map__nr(evsel_list->threads);
+
+                usleep(initial_delay * 1000);
+                list_for_each_entry(counter, &evsel_list->entries, node)
+                        perf_evsel__enable(counter, ncpus, nthreads);
+        }
+}
+
 static int __run_perf_stat(int argc, const char **argv)
 {
         char msg[512];
@@ -486,6 +502,7 @@ static int __run_perf_stat(int argc, const char **argv)
 
         if (forks) {
                 perf_evlist__start_workload(evsel_list);
+                handle_initial_delay();
 
                 if (interval) {
                         while (!waitpid(child_pid, &status, WNOHANG)) {
@@ -497,6 +514,7 @@ static int __run_perf_stat(int argc, const char **argv)
                 if (WIFSIGNALED(status))
                         psignal(WTERMSIG(status), argv[0]);
         } else {
+                handle_initial_delay();
                 while (!done) {
                         nanosleep(&ts, NULL);
                         if (interval)
@@ -1419,6 +1437,8 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
                      "aggregate counts per processor socket", AGGR_SOCKET),
         OPT_SET_UINT(0, "per-core", &aggr_mode,
                      "aggregate counts per physical processor core", AGGR_CORE),
+        OPT_UINTEGER('D', "delay", &initial_delay,
+                     "ms to wait before starting measurement after program start"),
         OPT_END()
         };
         const char * const stat_usage[] = {