Misc perf stat improvements

index 8065ce8..62efec9 100644
@@ -783,13 +783,6 @@ int perf_evlist__prepare_workload(struct perf_evlist *evlist,
 		fcntl(go_pipe[0], F_SETFD, FD_CLOEXEC);
 
 		/*
-		 * Do a dummy execvp to get the PLT entry resolved,
-		 * so we avoid the resolver overhead on the real
-		 * execvp call.
-		 */
-		execvp("", (char **)argv);
-
-		/*
 		 * Tell the parent we're ready to go
 		 */
 		close(child_ready_pipe[1]);

index c9c7494..60e0d84 100644
@@ -605,16 +605,16 @@ int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads)
 	return evsel->fd != NULL ? 0 : -ENOMEM;
 }
 
-int perf_evsel__set_filter(struct perf_evsel *evsel, int ncpus, int nthreads,
-			   const char *filter)
+static int perf_evsel__run_ioctl(struct perf_evsel *evsel, int ncpus, int nthreads,
+			  int ioc,  void *arg)
 {
 	int cpu, thread;
 
 	for (cpu = 0; cpu < ncpus; cpu++) {
 		for (thread = 0; thread < nthreads; thread++) {
 			int fd = FD(evsel, cpu, thread),
-			    err = ioctl(fd, PERF_EVENT_IOC_SET_FILTER, filter);
 
+			err = ioctl(fd, ioc, arg);
 			if (err)
 				return err;
 		}

@@ -623,6 +623,21 @@ int perf_evsel__set_filter(struct perf_evsel *evsel, int ncpus, int nthreads,
 	return 0;
 }
 
+int perf_evsel__set_filter(struct perf_evsel *evsel, int ncpus, int nthreads,
+			   const char *filter)
+{
+	return perf_evsel__run_ioctl(evsel, ncpus, nthreads,
+				     PERF_EVENT_IOC_SET_FILTER,
+				     (void *)filter);
+}
+
+int perf_evsel__enable(struct perf_evsel *evsel, int ncpus, int nthreads)
+{
+	return perf_evsel__run_ioctl(evsel, ncpus, nthreads,
+				     PERF_EVENT_IOC_ENABLE,
+				     0);
+}
+
 int perf_evsel__alloc_id(struct perf_evsel *evsel, int ncpus, int nthreads)
 {
 	evsel->sample_id = xyarray__new(ncpus, nthreads, sizeof(struct perf_sample_id));

index 3f156cc..b057e9c 100644
@@ -142,6 +142,7 @@ void perf_evsel__set_sample_id(struct perf_evsel *evsel);
 
 int perf_evsel__set_filter(struct perf_evsel *evsel, int ncpus, int nthreads,
 			   const char *filter);
+int perf_evsel__enable(struct perf_evsel *evsel, int ncpus, int nthreads);
 
 int perf_evsel__open_per_cpu(struct perf_evsel *evsel,
 			     struct cpu_map *cpus);

index 352fbd7..2e637e4 100644
@@ -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[] = {

index 2e637e4..f686d5f 100644
@@ -416,6 +416,8 @@ static void print_interval(void)
 		list_for_each_entry(counter, &evsel_list->entries, node)
 			print_counter_aggr(counter, prefix);
 	}
+
+	fflush(output);
 }
 
 static void handle_initial_delay(void)

index f686d5f..940fcfd 100644
@@ -982,6 +982,13 @@ static void print_aggr(char *prefix)
 					fprintf(output, "%s%s",
 						csv_sep, counter->cgrp->name);
 
+				if (csv_output)
+					fprintf(output, "%s%" PRIu64 "%s%.2f",
+						csv_sep,
+						run,
+						csv_sep,
+						100.0 * run / ena);
+
 				fputc('\n', output);
 				continue;
 			}

@@ -997,6 +1004,12 @@ static void print_aggr(char *prefix)
 				if (run != ena)
 					fprintf(output, "  (%.2f%%)",
 						100.0 * run / ena);
+			} else {
+				fprintf(output, "%s%" PRIu64 "%s%.2f",
+						csv_sep,
+						run,
+						csv_sep,
+						100.0 * run / ena);
 			}
 			fputc('\n', output);
 		}

@@ -1012,6 +1025,10 @@ static void print_counter_aggr(struct perf_evsel *counter, char *prefix)
 	struct perf_stat *ps = counter->priv;
 	double avg = avg_stats(&ps->res_stats[0]);
 	int scaled = counter->counts->scaled;
+	double avg_enabled, avg_running;
+
+	avg_enabled = avg_stats(&ps->res_stats[1]);
+	avg_running = avg_stats(&ps->res_stats[2]);
 
 	if (prefix)
 		fprintf(output, "%s", prefix);

@@ -1027,6 +1044,13 @@ static void print_counter_aggr(struct perf_evsel *counter, char *prefix)
 		if (counter->cgrp)
 			fprintf(output, "%s%s", csv_sep, counter->cgrp->name);
 
+		if (csv_output)
+			fprintf(output, "%s%.0f%s%.2f",
+				csv_sep,
+				avg_running,
+				csv_sep,
+				100.0 * avg_running / avg_enabled);
+
 		fputc('\n', output);
 		return;
 	}

@@ -1038,19 +1062,14 @@ static void print_counter_aggr(struct perf_evsel *counter, char *prefix)
 
 	print_noise(counter, avg);
 
-	if (csv_output) {
-		fputc('\n', output);
-		return;
-	}
-
-	if (scaled) {
-		double avg_enabled, avg_running;
-
-		avg_enabled = avg_stats(&ps->res_stats[1]);
-		avg_running = avg_stats(&ps->res_stats[2]);
-
+	if (csv_output)
+		fprintf(output, "%s%.0f%s%.2f",
+			csv_sep,
+			avg_running,
+			csv_sep,
+			100.0 * avg_running / avg_enabled);
+	else
 		fprintf(output, " [%5.2f%%]", 100 * avg_running / avg_enabled);
-	}
 	fprintf(output, "\n");
 }
 

@@ -1085,6 +1104,13 @@ static void print_counter(struct perf_evsel *counter, char *prefix)
 				fprintf(output, "%s%s",
 					csv_sep, counter->cgrp->name);
 
+			if (csv_output)
+				fprintf(output, "%s%" PRIu64 "%s%.2f",
+					csv_sep,
+					run,
+					csv_sep,
+					100.0 * run / ena);
+
 			fputc('\n', output);
 			continue;
 		}

@@ -1100,7 +1126,14 @@ static void print_counter(struct perf_evsel *counter, char *prefix)
 			if (run != ena)
 				fprintf(output, "  (%.2f%%)",
 					100.0 * run / ena);
+		} else {
+			fprintf(output, "%s%" PRIu64 "%s%.2f",
+					csv_sep,
+					run,
+					csv_sep,
+					100.0 * run / ena);
 		}
+
 		fputc('\n', output);
 	}
 }