Merge tag 'perf-core-for-mingo' of git://git.kernel.org/pub/scm/linux…

…/kernel/git/acme/linux into perf/core

Pull perf/core improvements and fixes from Arnaldo Carvalho de Melo:

. Check for flex and bison before continuing building, from Borislav Petkov.

. Make event_copy local to mmaps, fixing buffer wrap around problems, from
  David Ahern.

. Add option for runtime switching perf data file in perf report, just press
  's' and a menu with the valid files found in the current directory will be
  presented, from Feng Tang.

. Add support to display whole group data for raw columns, from Jiri Olsa.

. Fix SIGALRM and pipe read race for the rwtop perl script. from Jiri Olsa.

. Fix perf_evsel::exclude_GH handling and add a test to catch regressions, from
  Jiri Olsa.

. Error checking fixes, from Namhyung Kim.

. Fix calloc argument ordering, from Paul Gortmaker.

. Fix set event list leader, from Stephane Eranian.

. Add per processor socket count aggregation in perf stat, from Stephane Eranian.

. Fix perf python binding breakage.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>

tools/perf/Documentation/perf-stat.txt

@@ -116,9 +116,16 @@ perf stat --repeat 10 --null --sync --pre 'make -s O=defconfig-build/clean' -- m
 
 -I msecs::
 --interval-print msecs::
-	print count deltas every N milliseconds (minimum: 100ms)
+	Print count deltas every N milliseconds (minimum: 100ms)
 	example: perf stat -I 1000 -e cycles -a sleep 5
 
+--aggr-socket::
+Aggregate counts per processor socket for system-wide mode measurements.  This
+is a useful mode to detect imbalance between sockets.  To enable this mode,
+use --aggr-socket in addition to -a. (system-wide).  The output includes the
+socket number and the number of online processors on that socket. This is
+useful to gauge the amount of aggregation.
+
 EXAMPLES
 --------
 

tools/perf/Makefile

@@ -149,6 +149,8 @@ RM = rm -f
 MKDIR = mkdir
 FIND = find
 INSTALL = install
+FLEX = flex
+BISON= bison
 
 # sparse is architecture-neutral, which means that we need to tell it
 # explicitly what architecture to check for. Fix this up for yours..
@@ -158,6 +160,14 @@ ifneq ($(MAKECMDGOALS),clean)
 ifneq ($(MAKECMDGOALS),tags)
 -include config/feature-tests.mak
 
+ifeq ($(call get-executable,$(FLEX)),)
+	dummy := $(error Error: $(FLEX) is missing on this system, please install it)
+endif
+
+ifeq ($(call get-executable,$(BISON)),)
+	dummy := $(error Error: $(BISON) is missing on this system, please install it)
+endif
+
 ifeq ($(call try-cc,$(SOURCE_HELLO),$(CFLAGS) -Werror -fstack-protector-all,-fstack-protector-all),y)
 	CFLAGS := $(CFLAGS) -fstack-protector-all
 endif
@@ -282,9 +292,6 @@ endif
 
 export PERL_PATH
 
-FLEX = flex
-BISON= bison
-
 $(OUTPUT)util/parse-events-flex.c: util/parse-events.l $(OUTPUT)util/parse-events-bison.c
 	$(QUIET_FLEX)$(FLEX) --header-file=$(OUTPUT)util/parse-events-flex.h $(PARSER_DEBUG_FLEX) -t util/parse-events.l > $(OUTPUT)util/parse-events-flex.c
 

tools/perf/builtin-annotate.c

@@ -309,7 +309,8 @@ int cmd_annotate(int argc, const char **argv, const char *prefix __maybe_unused)
 	if (symbol__init() < 0)
 		return -1;
 
-	setup_sorting(annotate_usage, options);
+	if (setup_sorting() < 0)
+		usage_with_options(annotate_usage, options);
 
 	if (argc) {
 		/*

tools/perf/builtin-diff.c

@@ -605,7 +605,9 @@ int cmd_diff(int argc, const char **argv, const char *prefix __maybe_unused)
 
 	ui_init();
 
-	setup_sorting(diff_usage, options);
+	if (setup_sorting() < 0)
+		usage_with_options(diff_usage, options);
+
 	setup_pager();
 
 	sort_entry__setup_elide(&sort_dso, symbol_conf.dso_list, "dso", NULL);

tools/perf/builtin-evlist.c

@@ -39,7 +39,7 @@ int cmd_evlist(int argc, const char **argv, const char *prefix __maybe_unused)
 	OPT_BOOLEAN('F', "freq", &details.freq, "Show the sample frequency"),
 	OPT_BOOLEAN('v', "verbose", &details.verbose,
 		    "Show all event attr details"),
-	OPT_BOOLEAN('g', "group", &symbol_conf.event_group,
+	OPT_BOOLEAN('g', "group", &details.event_group,
 		    "Show event group information"),
 	OPT_END()
 	};
@@ -52,7 +52,7 @@ int cmd_evlist(int argc, const char **argv, const char *prefix __maybe_unused)
 	if (argc)
 		usage_with_options(evlist_usage, options);
 
-	if (symbol_conf.event_group && (details.verbose || details.freq)) {
+	if (details.event_group && (details.verbose || details.freq)) {
 		pr_err("--group option is not compatible with other options\n");
 		usage_with_options(evlist_usage, options);
 	}

tools/perf/builtin-report.c

@@ -468,9 +468,17 @@ static int __cmd_report(struct perf_report *rep)
 
 	if (use_browser > 0) {
 		if (use_browser == 1) {
-			perf_evlist__tui_browse_hists(session->evlist, help,
-						      NULL,
-						      &session->header.env);
+			ret = perf_evlist__tui_browse_hists(session->evlist,
+							help,
+							NULL,
+							&session->header.env);
+			/*
+			 * Usually "ret" is the last pressed key, and we only
+			 * care if the key notifies us to switch data file.
+			 */
+			if (ret != K_SWITCH_INPUT_DATA)
+				ret = 0;
+
 		} else if (use_browser == 2) {
 			perf_evlist__gtk_browse_hists(session->evlist, help,
 						      NULL);
@@ -708,6 +716,16 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
 		else
 			input_name = "perf.data";
 	}
+
+	if (strcmp(input_name, "-") != 0)
+		setup_browser(true);
+	else {
+		use_browser = 0;
+		perf_hpp__column_enable(PERF_HPP__OVERHEAD);
+		perf_hpp__init();
+	}
+
+repeat:
 	session = perf_session__new(input_name, O_RDONLY,
 				    report.force, false, &report.tool);
 	if (session == NULL)
@@ -733,15 +751,8 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
 
 	}
 
-	if (strcmp(input_name, "-") != 0)
-		setup_browser(true);
-	else {
-		use_browser = 0;
-		perf_hpp__column_enable(PERF_HPP__OVERHEAD);
-		perf_hpp__init();
-	}
-
-	setup_sorting(report_usage, options);
+	if (setup_sorting() < 0)
+		usage_with_options(report_usage, options);
 
 	/*
 	 * Only in the newt browser we are doing integrated annotation,
@@ -809,6 +820,12 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
 	}
 
 	ret = __cmd_report(&report);
+	if (ret == K_SWITCH_INPUT_DATA) {
+		perf_session__delete(session);
+		goto repeat;
+	} else
+		ret = 0;
+
 error:
 	perf_session__delete(session);
 	return ret;

tools/perf/builtin-stat.c

@@ -68,6 +68,7 @@
 static void print_stat(int argc, const char **argv);
 static void print_counter_aggr(struct perf_evsel *counter, char *prefix);
 static void print_counter(struct perf_evsel *counter, char *prefix);
+static void print_aggr_socket(char *prefix);
 
 static struct perf_evlist	*evsel_list;
 
@@ -79,6 +80,7 @@ static int			run_count			=  1;
 static bool			no_inherit			= false;
 static bool			scale				=  true;
 static bool			no_aggr				= false;
+static bool			aggr_socket			= false;
 static pid_t			child_pid			= -1;
 static bool			null_run			=  false;
 static int			detailed_run			=  0;
@@ -93,6 +95,7 @@ static const char		*post_cmd			= NULL;
 static bool			sync_run			= false;
 static unsigned int		interval			= 0;
 static struct timespec		ref_time;
+static struct cpu_map		*sock_map;
 
 static volatile int done = 0;
 
@@ -312,7 +315,9 @@ static void print_interval(void)
 	sprintf(prefix, "%6lu.%09lu%s", rs.tv_sec, rs.tv_nsec, csv_sep);
 
 	if (num_print_interval == 0 && !csv_output) {
-		if (no_aggr)
+		if (aggr_socket)
+			fprintf(output, "#           time socket cpus             counts events\n");
+		else if (no_aggr)
 			fprintf(output, "#           time CPU                 counts events\n");
 		else
 			fprintf(output, "#           time             counts events\n");
@@ -321,7 +326,9 @@ static void print_interval(void)
 	if (++num_print_interval == 25)
 		num_print_interval = 0;
 
-	if (no_aggr) {
+	if (aggr_socket)
+		print_aggr_socket(prefix);
+	else if (no_aggr) {
 		list_for_each_entry(counter, &evsel_list->entries, node)
 			print_counter(counter, prefix);
 	} else {
@@ -349,6 +356,12 @@ static int __run_perf_stat(int argc __maybe_unused, const char **argv)
 		ts.tv_nsec = 0;
 	}
 
+	if (aggr_socket
+	    && cpu_map__build_socket_map(evsel_list->cpus, &sock_map)) {
+		perror("cannot build socket map");
+		return -1;
+	}
+
 	if (forks && (pipe(child_ready_pipe) < 0 || pipe(go_pipe) < 0)) {
 		perror("failed to create pipes");
 		return -1;
@@ -529,13 +542,21 @@ static void print_noise(struct perf_evsel *evsel, double avg)
 	print_noise_pct(stddev_stats(&ps->res_stats[0]), avg);
 }
 
-static void nsec_printout(int cpu, struct perf_evsel *evsel, double avg)
+static void nsec_printout(int cpu, int nr, struct perf_evsel *evsel, double avg)
 {
 	double msecs = avg / 1e6;
 	char cpustr[16] = { '\0', };
 	const char *fmt = csv_output ? "%s%.6f%s%s" : "%s%18.6f%s%-25s";
 
-	if (no_aggr)
+	if (aggr_socket)
+		sprintf(cpustr, "S%*d%s%*d%s",
+			csv_output ? 0 : -5,
+			cpu,
+			csv_sep,
+			csv_output ? 0 : 4,
+			nr,
+			csv_sep);
+	else if (no_aggr)
 		sprintf(cpustr, "CPU%*d%s",
 			csv_output ? 0 : -4,
 			perf_evsel__cpus(evsel)->map[cpu], csv_sep);
@@ -734,7 +755,7 @@ static void print_ll_cache_misses(int cpu,
 	fprintf(output, " of all LL-cache hits   ");
 }
 
-static void abs_printout(int cpu, struct perf_evsel *evsel, double avg)
+static void abs_printout(int cpu, int nr, struct perf_evsel *evsel, double avg)
 {
 	double total, ratio = 0.0;
 	char cpustr[16] = { '\0', };
@@ -747,7 +768,15 @@ static void abs_printout(int cpu, struct perf_evsel *evsel, double avg)
 	else
 		fmt = "%s%18.0f%s%-25s";
 
-	if (no_aggr)
+	if (aggr_socket)
+		sprintf(cpustr, "S%*d%s%*d%s",
+			csv_output ? 0 : -5,
+			cpu,
+			csv_sep,
+			csv_output ? 0 : 4,
+			nr,
+			csv_sep);
+	else if (no_aggr)
 		sprintf(cpustr, "CPU%*d%s",
 			csv_output ? 0 : -4,
 			perf_evsel__cpus(evsel)->map[cpu], csv_sep);
@@ -853,6 +882,70 @@ static void abs_printout(int cpu, struct perf_evsel *evsel, double avg)
 	}
 }
 
+static void print_aggr_socket(char *prefix)
+{
+	struct perf_evsel *counter;
+	u64 ena, run, val;
+	int cpu, s, s2, sock, nr;
+
+	if (!sock_map)
+		return;
+
+	for (s = 0; s < sock_map->nr; s++) {
+		sock = cpu_map__socket(sock_map, s);
+		list_for_each_entry(counter, &evsel_list->entries, node) {
+			val = ena = run = 0;
+			nr = 0;
+			for (cpu = 0; cpu < perf_evsel__nr_cpus(counter); cpu++) {
+				s2 = cpu_map__get_socket(evsel_list->cpus, cpu);
+				if (s2 != sock)
+					continue;
+				val += counter->counts->cpu[cpu].val;
+				ena += counter->counts->cpu[cpu].ena;
+				run += counter->counts->cpu[cpu].run;
+				nr++;
+			}
+			if (prefix)
+				fprintf(output, "%s", prefix);
+
+			if (run == 0 || ena == 0) {
+				fprintf(output, "S%*d%s%*d%s%*s%s%*s",
+					csv_output ? 0 : -5,
+					s,
+					csv_sep,
+					csv_output ? 0 : 4,
+					nr,
+					csv_sep,
+					csv_output ? 0 : 18,
+					counter->supported ? CNTR_NOT_COUNTED : CNTR_NOT_SUPPORTED,
+					csv_sep,
+					csv_output ? 0 : -24,
+					perf_evsel__name(counter));
+				if (counter->cgrp)
+					fprintf(output, "%s%s",
+						csv_sep, counter->cgrp->name);
+
+				fputc('\n', output);
+				continue;
+			}
+
+			if (nsec_counter(counter))
+				nsec_printout(sock, nr, counter, val);
+			else
+				abs_printout(sock, nr, counter, val);
+
+			if (!csv_output) {
+				print_noise(counter, 1.0);
+
+				if (run != ena)
+					fprintf(output, "  (%.2f%%)",
+						100.0 * run / ena);
+			}
+			fputc('\n', output);
+		}
+	}
+}
+
 /*
  * Print out the results of a single counter:
  * aggregated counts in system-wide mode
@@ -882,9 +975,9 @@ static void print_counter_aggr(struct perf_evsel *counter, char *prefix)
 	}
 
 	if (nsec_counter(counter))
-		nsec_printout(-1, counter, avg);
+		nsec_printout(-1, 0, counter, avg);
 	else
-		abs_printout(-1, counter, avg);
+		abs_printout(-1, 0, counter, avg);
 
 	print_noise(counter, avg);
 
@@ -940,9 +1033,9 @@ static void print_counter(struct perf_evsel *counter, char *prefix)
 		}
 
 		if (nsec_counter(counter))
-			nsec_printout(cpu, counter, val);
+			nsec_printout(cpu, 0, counter, val);
 		else
-			abs_printout(cpu, counter, val);
+			abs_printout(cpu, 0, counter, val);
 
 		if (!csv_output) {
 			print_noise(counter, 1.0);
@@ -980,7 +1073,9 @@ static void print_stat(int argc, const char **argv)
 		fprintf(output, ":\n\n");
 	}
 
-	if (no_aggr) {
+	if (aggr_socket)
+		print_aggr_socket(NULL);
+	else if (no_aggr) {
 		list_for_each_entry(counter, &evsel_list->entries, node)
 			print_counter(counter, NULL);
 	} else {
@@ -1228,6 +1323,7 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
 			"command to run after to the measured command"),
 	OPT_UINTEGER('I', "interval-print", &interval,
 		    "print counts at regular interval in ms (>= 100)"),
+	OPT_BOOLEAN(0, "aggr-socket", &aggr_socket, "aggregate counts per processor socket"),
 	OPT_END()
 	};
 	const char * const stat_usage[] = {
@@ -1314,6 +1410,14 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
 		usage_with_options(stat_usage, options);
 	}
 
+	if (aggr_socket) {
+		if (!perf_target__has_cpu(&target)) {
+			fprintf(stderr, "--aggr-socket only available in system-wide mode (-a)\n");
+			usage_with_options(stat_usage, options);
+		}
+		no_aggr = true;
+	}
+
 	if (add_default_attributes())
 		goto out;