1 files changed, 95 insertions, 69 deletions

diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index ff7e1d6cfcd2..ba20bf7c011d 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -161,6 +161,7 @@ struct record {
 	struct evlist		*sb_evlist;
 	pthread_t		thread_id;
 	int			realtime_prio;
+	bool			latency;
 	bool			switch_output_event_set;
 	bool			no_buildid;
 	bool			no_buildid_set;
@@ -171,6 +172,7 @@ struct record {
 	bool			timestamp_filename;
 	bool			timestamp_boundary;
 	bool			off_cpu;
+	const char		*filter_action;
 	struct switch_output	switch_output;
 	unsigned long long	samples;
 	unsigned long		output_max_size;	/* = 0: unlimited */
@@ -193,6 +195,15 @@ static const char *affinity_tags[PERF_AFFINITY_MAX] = {
 	"SYS", "NODE", "CPU"
 };
 
+static int build_id__process_mmap(const struct perf_tool *tool, union perf_event *event,
+				  struct perf_sample *sample, struct machine *machine);
+static int build_id__process_mmap2(const struct perf_tool *tool, union perf_event *event,
+				   struct perf_sample *sample, struct machine *machine);
+static int process_timestamp_boundary(const struct perf_tool *tool,
+				      union perf_event *event,
+				      struct perf_sample *sample,
+				      struct machine *machine);
+
 #ifndef HAVE_GETTID
 static inline pid_t gettid(void)
 {
@@ -332,7 +343,7 @@ static int record__aio_complete(struct mmap *md, struct aiocb *cblock)
 	} else {
 		/*
 		 * aio write request may require restart with the
-		 * reminder if the kernel didn't write whole
+		 * remainder if the kernel didn't write whole
 		 * chunk at once.
 		 */
 		rem_off = cblock->aio_offset + written;
@@ -400,7 +411,7 @@ static int record__aio_pushfn(struct mmap *map, void *to, void *buf, size_t size
 	 *
 	 * Coping can be done in two steps in case the chunk of profiling data
 	 * crosses the upper bound of the kernel buffer. In this case we first move
-	 * part of data from map->start till the upper bound and then the reminder
+	 * part of data from map->start till the upper bound and then the remainder
 	 * from the beginning of the kernel buffer till the end of the data chunk.
 	 */
 
@@ -608,7 +619,7 @@ static int record__comp_enabled(struct record *rec)
 	return rec->opts.comp_level > 0;
 }
 
-static int process_synthesized_event(struct perf_tool *tool,
+static int process_synthesized_event(const struct perf_tool *tool,
 				     union perf_event *event,
 				     struct perf_sample *sample __maybe_unused,
 				     struct machine *machine __maybe_unused)
@@ -619,7 +630,7 @@ static int process_synthesized_event(struct perf_tool *tool,
 
 static struct mutex synth_lock;
 
-static int process_locked_synthesized_event(struct perf_tool *tool,
+static int process_locked_synthesized_event(const struct perf_tool *tool,
 				     union perf_event *event,
 				     struct perf_sample *sample __maybe_unused,
 				     struct machine *machine __maybe_unused)
@@ -704,7 +715,7 @@ static void record__sig_exit(void)
 
 #ifdef HAVE_AUXTRACE_SUPPORT
 
-static int record__process_auxtrace(struct perf_tool *tool,
+static int record__process_auxtrace(const struct perf_tool *tool,
 				    struct mmap *map,
 				    union perf_event *event, void *data1,
 				    size_t len1, void *data2, size_t len2)
@@ -850,7 +861,9 @@ static int record__auxtrace_init(struct record *rec)
 	if (err)
 		return err;
 
-	auxtrace_regroup_aux_output(rec->evlist);
+	err = auxtrace_parse_aux_action(rec->evlist);
+	if (err)
+		return err;
 
 	return auxtrace_parse_filters(rec->evlist);
 }
@@ -1355,8 +1368,6 @@ static int record__open(struct record *rec)
 	struct record_opts *opts = &rec->opts;
 	int rc = 0;
 
-	evlist__config(evlist, opts, &callchain_param);
-
 	evlist__for_each_entry(evlist, pos) {
 try_again:
 		if (evsel__open(pos, pos->core.cpus, pos->core.threads) < 0) {
@@ -1391,7 +1402,7 @@ try_again:
 "even with a suitable vmlinux or kallsyms file.\n\n");
 	}
 
-	if (evlist__apply_filters(evlist, &pos)) {
+	if (evlist__apply_filters(evlist, &pos, &opts->target)) {
 		pr_err("failed to set filter \"%s\" on event %s with %d (%s)\n",
 			pos->filter ?: "BPF", evsel__name(pos), errno,
 			str_error_r(errno, msg, sizeof(msg)));
@@ -1418,7 +1429,7 @@ static void set_timestamp_boundary(struct record *rec, u64 sample_time)
 		rec->evlist->last_sample_time = sample_time;
 }
 
-static int process_sample_event(struct perf_tool *tool,
+static int process_sample_event(const struct perf_tool *tool,
 				union perf_event *event,
 				struct perf_sample *sample,
 				struct evsel *evsel,
@@ -1460,7 +1471,7 @@ static int process_buildids(struct record *rec)
 	 * first/last samples.
 	 */
 	if (rec->buildid_all && !rec->timestamp_boundary)
-		rec->tool.sample = NULL;
+		rec->tool.sample = process_event_sample_stub;
 
 	return perf_session__process_events(session);
 }
@@ -1740,10 +1751,8 @@ static void record__init_features(struct record *rec)
 	if (rec->no_buildid)
 		perf_header__clear_feat(&session->header, HEADER_BUILD_ID);
 
-#ifdef HAVE_LIBTRACEEVENT
 	if (!have_tracepoints(&rec->evlist->core.entries))
 		perf_header__clear_feat(&session->header, HEADER_TRACING_DATA);
-#endif
 
 	if (!rec->opts.branch_stack)
 		perf_header__clear_feat(&session->header, HEADER_BRANCH_STACK);
@@ -1790,7 +1799,7 @@ record__finish_output(struct record *rec)
 		process_buildids(rec);
 
 		if (rec->buildid_all)
-			dsos__hit_all(rec->session);
+			perf_session__dsos_hit_all(rec->session);
 	}
 	perf_session__write_header(rec->session, rec->evlist, fd, true);
 
@@ -1909,9 +1918,10 @@ static void __record__save_lost_samples(struct record *rec, struct evsel *evsel,
 					u16 misc_flag)
 {
 	struct perf_sample_id *sid;
-	struct perf_sample sample = {};
+	struct perf_sample sample;
 	int id_hdr_size;
 
+	perf_sample__init(&sample, /*all=*/true);
 	lost->lost = lost_count;
 	if (evsel->core.ids) {
 		sid = xyarray__entry(evsel->core.sample_id, cpu_idx, thread_idx);
@@ -1923,12 +1933,13 @@ static void __record__save_lost_samples(struct record *rec, struct evsel *evsel,
 	lost->header.size = sizeof(*lost) + id_hdr_size;
 	lost->header.misc = misc_flag;
 	record__write(rec, NULL, lost, lost->header.size);
+	perf_sample__exit(&sample);
 }
 
 static void record__read_lost_samples(struct record *rec)
 {
 	struct perf_session *session = rec->session;
-	struct perf_record_lost_samples *lost = NULL;
+	struct perf_record_lost_samples_and_ids lost;
 	struct evsel *evsel;
 
 	/* there was an error during record__open */
@@ -1953,20 +1964,13 @@ static void record__read_lost_samples(struct record *rec)
 
 				if (perf_evsel__read(&evsel->core, x, y, &count) < 0) {
 					pr_debug("read LOST count failed\n");
-					goto out;
+					return;
 				}
 
 				if (count.lost) {
-					if (!lost) {
-						lost = zalloc(sizeof(*lost) +
-							      session->machines.host.id_hdr_size);
-						if (!lost) {
-							pr_debug("Memory allocation failed\n");
-							return;
-						}
-						lost->header.type = PERF_RECORD_LOST_SAMPLES;
-					}
-					__record__save_lost_samples(rec, evsel, lost,
+					memset(&lost, 0, sizeof(lost));
+					lost.lost.header.type = PERF_RECORD_LOST_SAMPLES;
+					__record__save_lost_samples(rec, evsel, &lost.lost,
 								    x, y, count.lost, 0);
 				}
 			}
@@ -1974,21 +1978,12 @@ static void record__read_lost_samples(struct record *rec)
 
 		lost_count = perf_bpf_filter__lost_count(evsel);
 		if (lost_count) {
-			if (!lost) {
-				lost = zalloc(sizeof(*lost) +
-					      session->machines.host.id_hdr_size);
-				if (!lost) {
-					pr_debug("Memory allocation failed\n");
-					return;
-				}
-				lost->header.type = PERF_RECORD_LOST_SAMPLES;
-			}
-			__record__save_lost_samples(rec, evsel, lost, 0, 0, lost_count,
+			memset(&lost, 0, sizeof(lost));
+			lost.lost.header.type = PERF_RECORD_LOST_SAMPLES;
+			__record__save_lost_samples(rec, evsel, &lost.lost, 0, 0, lost_count,
 						    PERF_RECORD_MISC_LOST_SAMPLES_BPF);
 		}
 	}
-out:
-	free(lost);
 }
 
 static volatile sig_atomic_t workload_exec_errno;
@@ -2382,13 +2377,8 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 	signal(SIGTERM, sig_handler);
 	signal(SIGSEGV, sigsegv_handler);
 
-	if (rec->opts.record_namespaces)
-		tool->namespace_events = true;
-
 	if (rec->opts.record_cgroup) {
-#ifdef HAVE_FILE_HANDLE
-		tool->cgroup_events = true;
-#else
+#ifndef HAVE_FILE_HANDLE
 		pr_err("cgroup tracking is not supported\n");
 		return -1;
 #endif
@@ -2404,6 +2394,18 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 		signal(SIGUSR2, SIG_IGN);
 	}
 
+	perf_tool__init(tool, /*ordered_events=*/true);
+	tool->sample		= process_sample_event;
+	tool->fork		= perf_event__process_fork;
+	tool->exit		= perf_event__process_exit;
+	tool->comm		= perf_event__process_comm;
+	tool->namespaces	= perf_event__process_namespaces;
+	tool->mmap		= build_id__process_mmap;
+	tool->mmap2		= build_id__process_mmap2;
+	tool->itrace_start	= process_timestamp_boundary;
+	tool->aux		= process_timestamp_boundary;
+	tool->namespace_events	= rec->opts.record_namespaces;
+	tool->cgroup_events	= rec->opts.record_cgroup;
 	session = perf_session__new(data, tool);
 	if (IS_ERR(session)) {
 		pr_err("Perf session creation failed.\n");
@@ -2483,6 +2485,8 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 
 	evlist__uniquify_name(rec->evlist);
 
+	evlist__config(rec->evlist, opts, &callchain_param);
+
 	/* Debug message used by test scripts */
 	pr_debug3("perf record opening and mmapping events\n");
 	if (record__open(rec) != 0) {
@@ -2531,6 +2535,9 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 		goto out_free_threads;
 	}
 
+	if (!evlist__needs_bpf_sb_event(rec->evlist))
+		opts->no_bpf_event = true;
+
 	err = record__setup_sb_evlist(rec);
 	if (err)
 		goto out_free_threads;
@@ -2881,10 +2888,10 @@ out_delete_session:
 	}
 #endif
 	zstd_fini(&session->zstd_data);
-	perf_session__delete(session);
-
 	if (!opts->no_bpf_event)
 		evlist__stop_sb_thread(rec->sb_evlist);
+
+	perf_session__delete(session);
 	return status;
 }
 
@@ -3198,7 +3205,7 @@ static int switch_output_setup(struct record *rec)
 	unsigned long val;
 
 	/*
-	 * If we're using --switch-output-events, then we imply its 
+	 * If we're using --switch-output-events, then we imply its
 	 * --switch-output=signal, as we'll send a SIGUSR2 from the side band
 	 *  thread to its parent.
 	 */
@@ -3259,7 +3266,7 @@ static const char * const __record_usage[] = {
 };
 const char * const *record_usage = __record_usage;
 
-static int build_id__process_mmap(struct perf_tool *tool, union perf_event *event,
+static int build_id__process_mmap(const struct perf_tool *tool, union perf_event *event,
 				  struct perf_sample *sample, struct machine *machine)
 {
 	/*
@@ -3271,7 +3278,7 @@ static int build_id__process_mmap(struct perf_tool *tool, union perf_event *even
 	return perf_event__process_mmap(tool, event, sample, machine);
 }
 
-static int build_id__process_mmap2(struct perf_tool *tool, union perf_event *event,
+static int build_id__process_mmap2(const struct perf_tool *tool, union perf_event *event,
 				   struct perf_sample *sample, struct machine *machine)
 {
 	/*
@@ -3284,7 +3291,7 @@ static int build_id__process_mmap2(struct perf_tool *tool, union perf_event *eve
 	return perf_event__process_mmap2(tool, event, sample, machine);
 }
 
-static int process_timestamp_boundary(struct perf_tool *tool,
+static int process_timestamp_boundary(const struct perf_tool *tool,
 				      union perf_event *event __maybe_unused,
 				      struct perf_sample *sample,
 				      struct machine *machine __maybe_unused)
@@ -3342,18 +3349,6 @@ static struct record record = {
 		.ctl_fd_ack          = -1,
 		.synth               = PERF_SYNTH_ALL,
 	},
-	.tool = {
-		.sample		= process_sample_event,
-		.fork		= perf_event__process_fork,
-		.exit		= perf_event__process_exit,
-		.comm		= perf_event__process_comm,
-		.namespaces	= perf_event__process_namespaces,
-		.mmap		= build_id__process_mmap,
-		.mmap2		= build_id__process_mmap2,
-		.itrace_start	= process_timestamp_boundary,
-		.aux		= process_timestamp_boundary,
-		.ordered_events	= true,
-	},
 };
 
 const char record_callchain_help[] = CALLCHAIN_RECORD_HELP
@@ -3382,6 +3377,9 @@ static struct option __record_options[] = {
 		     parse_events_option),
 	OPT_CALLBACK(0, "filter", &record.evlist, "filter",
 		     "event filter", parse_filter),
+	OPT_BOOLEAN(0, "latency", &record.latency,
+		    "Enable data collection for latency profiling.\n"
+		    "\t\t\t  Use perf report --latency for latency-centric profile."),
 	OPT_CALLBACK_NOOPT(0, "exclude-perf", &record.evlist,
 			   NULL, "don't record events from perf itself",
 			   exclude_perf),
@@ -3573,6 +3571,8 @@ static struct option __record_options[] = {
 			    "write collected trace data into several data files using parallel threads",
 			    record__parse_threads),
 	OPT_BOOLEAN(0, "off-cpu", &record.off_cpu, "Enable off-cpu analysis"),
+	OPT_STRING(0, "setup-filter", &record.filter_action, "pin|unpin",
+		   "BPF filter action"),
 	OPT_END()
 };
 
@@ -4026,6 +4026,22 @@ int cmd_record(int argc, const char **argv)
 
 	}
 
+	if (record.latency) {
+		/*
+		 * There is no fundamental reason why latency profiling
+		 * can't work for system-wide mode, but exact semantics
+		 * and details are to be defined.
+		 * See the following thread for details:
+		 * https://lore.kernel.org/all/Z4XDJyvjiie3howF@google.com/
+		 */
+		if (record.opts.target.system_wide) {
+			pr_err("Failed: latency profiling is not supported with system-wide collection.\n");
+			err = -EINVAL;
+			goto out_opts;
+		}
+		record.opts.record_switch_events = true;
+	}
+
 	if (rec->buildid_mmap) {
 		if (!perf_can_record_build_id()) {
 			pr_err("Failed: no support to record build id in mmap events, update your kernel.\n");
@@ -4102,6 +4118,18 @@ int cmd_record(int argc, const char **argv)
 		pr_warning("WARNING: --timestamp-filename option is not available in parallel streaming mode.\n");
 	}
 
+	if (rec->filter_action) {
+		if (!strcmp(rec->filter_action, "pin"))
+			err = perf_bpf_filter__pin();
+		else if (!strcmp(rec->filter_action, "unpin"))
+			err = perf_bpf_filter__unpin();
+		else {
+			pr_warning("Unknown BPF filter action: %s\n", rec->filter_action);
+			err = -EINVAL;
+		}
+		goto out_opts;
+	}
+
 	/*
 	 * Allow aliases to facilitate the lookup of symbols for address
 	 * filters. Refer to auxtrace_parse_filters().
@@ -4154,9 +4182,7 @@ int cmd_record(int argc, const char **argv)
 		record.opts.tail_synthesize = true;
 
 	if (rec->evlist->core.nr_entries == 0) {
-		bool can_profile_kernel = perf_event_paranoid_check(1);
-
-		err = parse_event(rec->evlist, can_profile_kernel ? "cycles:P" : "cycles:Pu");
+		err = parse_event(rec->evlist, "cycles:P");
 		if (err)
 			goto out;
 	}
@@ -4258,13 +4284,13 @@ int cmd_record(int argc, const char **argv)
 
 	err = __cmd_record(&record, argc, argv);
 out:
-	evlist__delete(rec->evlist);
+	record__free_thread_masks(rec, rec->nr_threads);
+	rec->nr_threads = 0;
 	symbol__exit();
 	auxtrace_record__free(rec->itr);
 out_opts:
-	record__free_thread_masks(rec, rec->nr_threads);
-	rec->nr_threads = 0;
 	evlist__close_control(rec->opts.ctl_fd, rec->opts.ctl_fd_ack, &rec->opts.ctl_fd_close);
+	evlist__delete(rec->evlist);
 	return err;
 }