Merge tag 'perf-tools-for-v5.19-2022-05-23' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux

Pull perf tool updates from Arnaldo Carvalho de Melo:
 "Intel PT:

   - Allow hardware tracing on KVM test programs. In this case, the VM
     is not running an OS, but only the functions loaded into it by the
     hypervisor test program, and conveniently, loaded at the same
     virtual addresses.

   - Improve documentation:
      - Add link to perf wiki's page

   - Cleanups:
      - Delete now unused perf-with-kcore.sh script
      - Remove unused machines__find_host()

  ARM SPE (Statistical Profile Extensions):

   - Add man page entry.

  Vendor Events:

   - Update various Intel event topics

   - Update various microarch events

   - Fix various cstate metrics

   - Fix Alderlake metric groups

   - Add sapphirerapids events

   - Add JSON files for ARM Cortex A34, A35, A55, A510, A65, A73, A75,
     A77, A78, A710, X1, X2 and Neoverse E1

   - Update Cortex A57/A72

  perf stat:

   - Introduce stats for the user and system rusage times

  perf c2c:

   - Prep work to support ARM systems

  perf annotate:

   - Add --percent-limit option

  perf lock:

   - Add -t/--thread option for report

   - Do not discard broken lock stats

  perf bench:

   - Add breakpoint benchmarks

  perf test:

   - Limit to only run executable scripts in tests

   - Add basic perf record tests

   - Add stat record+report test

   - Add basic stat and topdown group test

   - Skip several tests when the user hasn't permission to perform them

   - Fix test case 81 ("perf record tests") on s390x

  perf version:

   - debuginfod support improvements

  perf scripting python:

   - Expose symbol offset and source information

  perf build:

   - Error for BPF skeletons without LIBBPF

   - Use Python devtools for version autodetection rather than runtime

  Miscellaneous:

   - Add riscv64 support to 'perf jitdump'

   - Various fixes/tidy ups related to cpu_map

   - Fixes for handling Intel hybrid systems"

* tag 'perf-tools-for-v5.19-2022-05-23' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux: (122 commits)
  perf intel-pt: Add guest_code support
  perf kvm report: Add guest_code support
  perf script: Add guest_code support
  perf tools: Add guest_code support
  perf tools: Factor out thread__set_guest_comm()
  perf tools: Add machine to machines back pointer
  perf vendors events arm64: Update Cortex A57/A72
  perf vendors events arm64: Arm Neoverse E1
  perf vendors events arm64: Arm Cortex-X2
  perf vendors events arm64: Arm Cortex-X1
  perf vendors events arm64: Arm Cortex-A710
  perf vendors events arm64: Arm Cortex-A78
  perf vendors events arm64: Arm Cortex-A77
  perf vendors events arm64: Arm Cortex-A75
  perf vendors events arm64: Arm Cortex-A73
  perf vendors events arm64: Arm Cortex-A65
  perf vendors events arm64: Arm Cortex-A510
  perf vendors events arm64: Arm Cortex-A55
  perf vendors events arm64: Arm Cortex-A35
  perf vendors events arm64: Arm Cortex-A34
  ...

1 files changed, 54 insertions, 17 deletions

diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index a96f106dc93a..7e6cc8bdf061 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -216,6 +216,7 @@ static struct perf_stat_config stat_config = {
 	.run_count		= 1,
 	.metric_only_len	= METRIC_ONLY_LEN,
 	.walltime_nsecs_stats	= &walltime_nsecs_stats,
+	.ru_stats		= &ru_stats,
 	.big_num		= true,
 	.ctl_fd			= -1,
 	.ctl_fd_ack		= -1,
@@ -271,11 +272,8 @@ static void evlist__check_cpu_maps(struct evlist *evlist)
 			pr_warning("     %s: %s\n", evsel->name, buf);
 		}
 
-		for_each_group_evsel(pos, leader) {
-			evsel__set_leader(pos, pos);
-			pos->core.nr_members = 0;
-		}
-		evsel->core.leader->nr_members = 0;
+		for_each_group_evsel(pos, leader)
+			evsel__remove_from_group(pos, leader);
 	}
 }
 
@@ -341,15 +339,35 @@ static int evsel__write_stat_event(struct evsel *counter, int cpu_map_idx, u32 t
 static int read_single_counter(struct evsel *counter, int cpu_map_idx,
 			       int thread, struct timespec *rs)
 {
-	if (counter->tool_event == PERF_TOOL_DURATION_TIME) {
-		u64 val = rs->tv_nsec + rs->tv_sec*1000000000ULL;
-		struct perf_counts_values *count =
-			perf_counts(counter->counts, cpu_map_idx, thread);
-		count->ena = count->run = val;
-		count->val = val;
-		return 0;
+	switch(counter->tool_event) {
+		case PERF_TOOL_DURATION_TIME: {
+			u64 val = rs->tv_nsec + rs->tv_sec*1000000000ULL;
+			struct perf_counts_values *count =
+				perf_counts(counter->counts, cpu_map_idx, thread);
+			count->ena = count->run = val;
+			count->val = val;
+			return 0;
+		}
+		case PERF_TOOL_USER_TIME:
+		case PERF_TOOL_SYSTEM_TIME: {
+			u64 val;
+			struct perf_counts_values *count =
+				perf_counts(counter->counts, cpu_map_idx, thread);
+			if (counter->tool_event == PERF_TOOL_USER_TIME)
+				val = ru_stats.ru_utime_usec_stat.mean;
+			else
+				val = ru_stats.ru_stime_usec_stat.mean;
+			count->ena = count->run = val;
+			count->val = val;
+			return 0;
+		}
+		default:
+		case PERF_TOOL_NONE:
+			return evsel__read_counter(counter, cpu_map_idx, thread);
+		case PERF_TOOL_MAX:
+			/* This should never be reached */
+			return 0;
 	}
-	return evsel__read_counter(counter, cpu_map_idx, thread);
 }
 
 /*
@@ -1010,8 +1028,10 @@ try_again_reset:
 		evlist__reset_prev_raw_counts(evsel_list);
 		runtime_stat_reset(&stat_config);
 		perf_stat__reset_shadow_per_stat(&rt_stat);
-	} else
+	} else {
 		update_stats(&walltime_nsecs_stats, t1 - t0);
+		update_rusage_stats(&ru_stats, &stat_config.ru_data);
+	}
 
 	/*
 	 * Closing a group leader splits the group, and as we only disable
@@ -1235,6 +1255,8 @@ static struct option stat_options[] = {
 	OPT_SET_UINT('A', "no-aggr", &stat_config.aggr_mode,
 		    "disable CPU count aggregation", AGGR_NONE),
 	OPT_BOOLEAN(0, "no-merge", &stat_config.no_merge, "Do not merge identical named events"),
+	OPT_BOOLEAN(0, "hybrid-merge", &stat_config.hybrid_merge,
+		    "Merge identical named hybrid events"),
 	OPT_STRING('x', "field-separator", &stat_config.csv_sep, "separator",
 		   "print counts with custom separator"),
 	OPT_CALLBACK('G', "cgroup", &evsel_list, "name",
@@ -1842,11 +1864,23 @@ static int add_default_attributes(void)
 		unsigned int max_level = 1;
 		char *str = NULL;
 		bool warn = false;
+		const char *pmu_name = "cpu";
 
 		if (!force_metric_only)
 			stat_config.metric_only = true;
 
-		if (pmu_have_event("cpu", topdown_metric_L2_attrs[5])) {
+		if (perf_pmu__has_hybrid()) {
+			if (!evsel_list->hybrid_pmu_name) {
+				pr_warning("WARNING: default to use cpu_core topdown events\n");
+				evsel_list->hybrid_pmu_name = perf_pmu__hybrid_type_to_pmu("core");
+			}
+
+			pmu_name = evsel_list->hybrid_pmu_name;
+			if (!pmu_name)
+				return -1;
+		}
+
+		if (pmu_have_event(pmu_name, topdown_metric_L2_attrs[5])) {
 			metric_attrs = topdown_metric_L2_attrs;
 			max_level = 2;
 		}
@@ -1857,10 +1891,11 @@ static int add_default_attributes(void)
 		} else if (!stat_config.topdown_level)
 			stat_config.topdown_level = max_level;
 
-		if (topdown_filter_events(metric_attrs, &str, 1) < 0) {
+		if (topdown_filter_events(metric_attrs, &str, 1, pmu_name) < 0) {
 			pr_err("Out of memory\n");
 			return -1;
 		}
+
 		if (metric_attrs[0] && str) {
 			if (!stat_config.interval && !stat_config.metric_only) {
 				fprintf(stat_config.output,
@@ -1884,10 +1919,12 @@ static int add_default_attributes(void)
 		}
 
 		if (topdown_filter_events(topdown_attrs, &str,
-				arch_topdown_check_group(&warn)) < 0) {
+				arch_topdown_check_group(&warn),
+				pmu_name) < 0) {
 			pr_err("Out of memory\n");
 			return -1;
 		}
+
 		if (topdown_attrs[0] && str) {
 			struct parse_events_error errinfo;
 			if (warn)