1 files changed, 107 insertions, 124 deletions

diff --git a/tools/perf/arch/x86/util/intel-pt.c b/tools/perf/arch/x86/util/intel-pt.c
index 082e5f2a415a..6de7e2d21075 100644
--- a/tools/perf/arch/x86/util/intel-pt.c
+++ b/tools/perf/arch/x86/util/intel-pt.c
@@ -11,6 +11,7 @@
 #include <linux/bitops.h>
 #include <linux/log2.h>
 #include <linux/zalloc.h>
+#include <linux/err.h>
 #include <cpuid.h>
 
 #include "../../../util/session.h"
@@ -22,7 +23,7 @@
 #include "../../../util/mmap.h"
 #include <subcmd/parse-options.h>
 #include "../../../util/parse-events.h"
-#include "../../../util/pmu.h"
+#include "../../../util/pmus.h"
 #include "../../../util/debug.h"
 #include "../../../util/auxtrace.h"
 #include "../../../util/perf_api_probe.h"
@@ -59,43 +60,34 @@ struct intel_pt_recording {
 	size_t				priv_size;
 };
 
-static int intel_pt_parse_terms_with_default(const char *pmu_name,
-					     struct list_head *formats,
+static int intel_pt_parse_terms_with_default(const struct perf_pmu *pmu,
 					     const char *str,
 					     u64 *config)
 {
-	struct list_head *terms;
+	struct parse_events_terms terms;
 	struct perf_event_attr attr = { .size = 0, };
 	int err;
 
-	terms = malloc(sizeof(struct list_head));
-	if (!terms)
-		return -ENOMEM;
-
-	INIT_LIST_HEAD(terms);
-
-	err = parse_events_terms(terms, str);
+	parse_events_terms__init(&terms);
+	err = parse_events_terms(&terms, str, /*input=*/ NULL);
 	if (err)
 		goto out_free;
 
 	attr.config = *config;
-	err = perf_pmu__config_terms(pmu_name, formats, &attr, terms, true,
-				     NULL);
+	err = perf_pmu__config_terms(pmu, &attr, &terms, /*zero=*/true, /*err=*/NULL);
 	if (err)
 		goto out_free;
 
 	*config = attr.config;
 out_free:
-	parse_events_terms__delete(terms);
+	parse_events_terms__exit(&terms);
 	return err;
 }
 
-static int intel_pt_parse_terms(const char *pmu_name, struct list_head *formats,
-				const char *str, u64 *config)
+static int intel_pt_parse_terms(const struct perf_pmu *pmu, const char *str, u64 *config)
 {
 	*config = 0;
-	return intel_pt_parse_terms_with_default(pmu_name, formats, str,
-						 config);
+	return intel_pt_parse_terms_with_default(pmu, str, config);
 }
 
 static u64 intel_pt_masked_bits(u64 mask, u64 bits)
@@ -125,7 +117,7 @@ static int intel_pt_read_config(struct perf_pmu *intel_pt_pmu, const char *str,
 
 	*res = 0;
 
-	mask = perf_pmu__format_bits(&intel_pt_pmu->format, str);
+	mask = perf_pmu__format_bits(intel_pt_pmu, str);
 	if (!mask)
 		return -EINVAL;
 
@@ -185,7 +177,7 @@ static int intel_pt_pick_bit(int bits, int target)
 	return pick;
 }
 
-static u64 intel_pt_default_config(struct perf_pmu *intel_pt_pmu)
+static u64 intel_pt_default_config(const struct perf_pmu *intel_pt_pmu)
 {
 	char buf[256];
 	int mtc, mtc_periods = 0, mtc_period;
@@ -193,16 +185,19 @@ static u64 intel_pt_default_config(struct perf_pmu *intel_pt_pmu)
 	int pos = 0;
 	u64 config;
 	char c;
+	int dirfd;
+
+	dirfd = perf_pmu__event_source_devices_fd();
 
 	pos += scnprintf(buf + pos, sizeof(buf) - pos, "tsc");
 
-	if (perf_pmu__scan_file(intel_pt_pmu, "caps/mtc", "%d",
-				&mtc) != 1)
+	if (perf_pmu__scan_file_at(intel_pt_pmu, dirfd, "caps/mtc", "%d",
+				   &mtc) != 1)
 		mtc = 1;
 
 	if (mtc) {
-		if (perf_pmu__scan_file(intel_pt_pmu, "caps/mtc_periods", "%x",
-					&mtc_periods) != 1)
+		if (perf_pmu__scan_file_at(intel_pt_pmu, dirfd, "caps/mtc_periods", "%x",
+					   &mtc_periods) != 1)
 			mtc_periods = 0;
 		if (mtc_periods) {
 			mtc_period = intel_pt_pick_bit(mtc_periods, 3);
@@ -211,13 +206,13 @@ static u64 intel_pt_default_config(struct perf_pmu *intel_pt_pmu)
 		}
 	}
 
-	if (perf_pmu__scan_file(intel_pt_pmu, "caps/psb_cyc", "%d",
-				&psb_cyc) != 1)
+	if (perf_pmu__scan_file_at(intel_pt_pmu, dirfd, "caps/psb_cyc", "%d",
+				   &psb_cyc) != 1)
 		psb_cyc = 1;
 
 	if (psb_cyc && mtc_periods) {
-		if (perf_pmu__scan_file(intel_pt_pmu, "caps/psb_periods", "%x",
-					&psb_periods) != 1)
+		if (perf_pmu__scan_file_at(intel_pt_pmu, dirfd, "caps/psb_periods", "%x",
+					   &psb_periods) != 1)
 			psb_periods = 0;
 		if (psb_periods) {
 			psb_period = intel_pt_pick_bit(psb_periods, 3);
@@ -226,15 +221,15 @@ static u64 intel_pt_default_config(struct perf_pmu *intel_pt_pmu)
 		}
 	}
 
-	if (perf_pmu__scan_file(intel_pt_pmu, "format/pt", "%c", &c) == 1 &&
-	    perf_pmu__scan_file(intel_pt_pmu, "format/branch", "%c", &c) == 1)
+	if (perf_pmu__scan_file_at(intel_pt_pmu, dirfd, "format/pt", "%c", &c) == 1 &&
+	    perf_pmu__scan_file_at(intel_pt_pmu, dirfd, "format/branch", "%c", &c) == 1)
 		pos += scnprintf(buf + pos, sizeof(buf) - pos, ",pt,branch");
 
 	pr_debug2("%s default config: %s\n", intel_pt_pmu->name, buf);
 
-	intel_pt_parse_terms(intel_pt_pmu->name, &intel_pt_pmu->format, buf,
-			     &config);
+	intel_pt_parse_terms(intel_pt_pmu, buf, &config);
 
+	close(dirfd);
 	return config;
 }
 
@@ -261,20 +256,17 @@ static int intel_pt_parse_snapshot_options(struct auxtrace_record *itr,
 	return 0;
 }
 
-struct perf_event_attr *
-intel_pt_pmu_default_config(struct perf_pmu *intel_pt_pmu)
+void intel_pt_pmu_default_config(const struct perf_pmu *intel_pt_pmu,
+				 struct perf_event_attr *attr)
 {
-	struct perf_event_attr *attr;
+	static u64 config;
+	static bool initialized;
 
-	attr = zalloc(sizeof(struct perf_event_attr));
-	if (!attr)
-		return NULL;
-
-	attr->config = intel_pt_default_config(intel_pt_pmu);
-
-	intel_pt_pmu->selectable = true;
-
-	return attr;
+	if (!initialized) {
+		config = intel_pt_default_config(intel_pt_pmu);
+		initialized = true;
+	}
+	attr->config = config;
 }
 
 static const char *intel_pt_find_filter(struct evlist *evlist,
@@ -306,6 +298,7 @@ intel_pt_info_priv_size(struct auxtrace_record *itr, struct evlist *evlist)
 
 	ptr->priv_size = (INTEL_PT_AUXTRACE_PRIV_MAX * sizeof(u64)) +
 			 intel_pt_filter_bytes(filter);
+	ptr->priv_size += sizeof(u64); /* Cap Event Trace */
 
 	return ptr->priv_size;
 }
@@ -335,28 +328,27 @@ static int intel_pt_info_fill(struct auxtrace_record *itr,
 	unsigned long max_non_turbo_ratio;
 	size_t filter_str_len;
 	const char *filter;
+	int event_trace;
 	__u64 *info;
 	int err;
 
 	if (priv_size != ptr->priv_size)
 		return -EINVAL;
 
-	intel_pt_parse_terms(intel_pt_pmu->name, &intel_pt_pmu->format,
-			     "tsc", &tsc_bit);
-	intel_pt_parse_terms(intel_pt_pmu->name, &intel_pt_pmu->format,
-			     "noretcomp", &noretcomp_bit);
-	intel_pt_parse_terms(intel_pt_pmu->name, &intel_pt_pmu->format,
-			     "mtc", &mtc_bit);
-	mtc_freq_bits = perf_pmu__format_bits(&intel_pt_pmu->format,
-					      "mtc_period");
-	intel_pt_parse_terms(intel_pt_pmu->name, &intel_pt_pmu->format,
-			     "cyc", &cyc_bit);
+	intel_pt_parse_terms(intel_pt_pmu, "tsc", &tsc_bit);
+	intel_pt_parse_terms(intel_pt_pmu, "noretcomp", &noretcomp_bit);
+	intel_pt_parse_terms(intel_pt_pmu, "mtc", &mtc_bit);
+	mtc_freq_bits = perf_pmu__format_bits(intel_pt_pmu, "mtc_period");
+	intel_pt_parse_terms(intel_pt_pmu, "cyc", &cyc_bit);
 
 	intel_pt_tsc_ctc_ratio(&tsc_ctc_ratio_n, &tsc_ctc_ratio_d);
 
 	if (perf_pmu__scan_file(intel_pt_pmu, "max_nonturbo_ratio",
 				"%lu", &max_non_turbo_ratio) != 1)
 		max_non_turbo_ratio = 0;
+	if (perf_pmu__scan_file(intel_pt_pmu, "caps/event_trace",
+				"%d", &event_trace) != 1)
+		event_trace = 0;
 
 	filter = intel_pt_find_filter(session->evlist, ptr->intel_pt_pmu);
 	filter_str_len = filter ? strlen(filter) : 0;
@@ -377,7 +369,7 @@ static int intel_pt_info_fill(struct auxtrace_record *itr,
 			ui__warning("Intel Processor Trace: TSC not available\n");
 	}
 
-	per_cpu_mmaps = !perf_cpu_map__empty(session->evlist->core.cpus);
+	per_cpu_mmaps = !perf_cpu_map__is_any_cpu_or_is_empty(session->evlist->core.user_requested_cpus);
 
 	auxtrace_info->type = PERF_AUXTRACE_INTEL_PT;
 	auxtrace_info->priv[INTEL_PT_PMU_TYPE] = intel_pt_pmu->type;
@@ -407,36 +399,34 @@ static int intel_pt_info_fill(struct auxtrace_record *itr,
 		info += len >> 3;
 	}
 
+	*info++ = event_trace;
+
 	return 0;
 }
 
+#ifdef HAVE_LIBTRACEEVENT
 static int intel_pt_track_switches(struct evlist *evlist)
 {
 	const char *sched_switch = "sched:sched_switch";
 	struct evsel *evsel;
 	int err;
 
-	if (!perf_evlist__can_select_event(evlist, sched_switch))
+	if (!evlist__can_select_event(evlist, sched_switch))
 		return -EPERM;
 
-	err = parse_events(evlist, sched_switch, NULL);
-	if (err) {
-		pr_debug2("%s: failed to parse %s, error %d\n",
+	evsel = evlist__add_sched_switch(evlist, true);
+	if (IS_ERR(evsel)) {
+		err = PTR_ERR(evsel);
+		pr_debug2("%s: failed to create %s, error = %d\n",
 			  __func__, sched_switch, err);
 		return err;
 	}
 
-	evsel = evlist__last(evlist);
-
-	evsel__set_sample_bit(evsel, CPU);
-	evsel__set_sample_bit(evsel, TIME);
-
-	evsel->core.system_wide = true;
-	evsel->no_aux_samples = true;
 	evsel->immediate = true;
 
 	return 0;
 }
+#endif
 
 static void intel_pt_valid_str(char *str, size_t len, u64 valid)
 {
@@ -484,7 +474,7 @@ static void intel_pt_valid_str(char *str, size_t len, u64 valid)
 	}
 }
 
-static int intel_pt_val_config_term(struct perf_pmu *intel_pt_pmu,
+static int intel_pt_val_config_term(struct perf_pmu *intel_pt_pmu, int dirfd,
 				    const char *caps, const char *name,
 				    const char *supported, u64 config)
 {
@@ -494,16 +484,16 @@ static int intel_pt_val_config_term(struct perf_pmu *intel_pt_pmu,
 	u64 bits;
 	int ok;
 
-	if (perf_pmu__scan_file(intel_pt_pmu, caps, "%llx", &valid) != 1)
+	if (perf_pmu__scan_file_at(intel_pt_pmu, dirfd, caps, "%llx", &valid) != 1)
 		valid = 0;
 
 	if (supported &&
-	    perf_pmu__scan_file(intel_pt_pmu, supported, "%d", &ok) == 1 && !ok)
+	    perf_pmu__scan_file_at(intel_pt_pmu, dirfd, supported, "%d", &ok) == 1 && !ok)
 		valid = 0;
 
 	valid |= 1;
 
-	bits = perf_pmu__format_bits(&intel_pt_pmu->format, name);
+	bits = perf_pmu__format_bits(intel_pt_pmu, name);
 
 	config &= bits;
 
@@ -527,56 +517,45 @@ out_err:
 static int intel_pt_validate_config(struct perf_pmu *intel_pt_pmu,
 				    struct evsel *evsel)
 {
-	int err;
+	int err, dirfd;
 	char c;
 
 	if (!evsel)
 		return 0;
 
+	dirfd = perf_pmu__event_source_devices_fd();
+	if (dirfd < 0)
+		return dirfd;
+
 	/*
 	 * If supported, force pass-through config term (pt=1) even if user
 	 * sets pt=0, which avoids senseless kernel errors.
 	 */
-	if (perf_pmu__scan_file(intel_pt_pmu, "format/pt", "%c", &c) == 1 &&
+	if (perf_pmu__scan_file_at(intel_pt_pmu, dirfd, "format/pt", "%c", &c) == 1 &&
 	    !(evsel->core.attr.config & 1)) {
 		pr_warning("pt=0 doesn't make sense, forcing pt=1\n");
 		evsel->core.attr.config |= 1;
 	}
 
-	err = intel_pt_val_config_term(intel_pt_pmu, "caps/cycle_thresholds",
+	err = intel_pt_val_config_term(intel_pt_pmu, dirfd, "caps/cycle_thresholds",
 				       "cyc_thresh", "caps/psb_cyc",
 				       evsel->core.attr.config);
 	if (err)
-		return err;
+		goto out;
 
-	err = intel_pt_val_config_term(intel_pt_pmu, "caps/mtc_periods",
+	err = intel_pt_val_config_term(intel_pt_pmu, dirfd, "caps/mtc_periods",
 				       "mtc_period", "caps/mtc",
 				       evsel->core.attr.config);
 	if (err)
-		return err;
+		goto out;
 
-	return intel_pt_val_config_term(intel_pt_pmu, "caps/psb_periods",
+	err = intel_pt_val_config_term(intel_pt_pmu, dirfd, "caps/psb_periods",
 					"psb_period", "caps/psb_cyc",
 					evsel->core.attr.config);
-}
-
-static void intel_pt_config_sample_mode(struct perf_pmu *intel_pt_pmu,
-					struct evsel *evsel)
-{
-	u64 user_bits = 0, bits;
-	struct evsel_config_term *term = evsel__get_config_term(evsel, CFG_CHG);
-
-	if (term)
-		user_bits = term->val.cfg_chg;
-
-	bits = perf_pmu__format_bits(&intel_pt_pmu->format, "psb_period");
 
-	/* Did user change psb_period */
-	if (bits & user_bits)
-		return;
-
-	/* Set psb_period to 0 */
-	evsel->core.attr.config &= ~bits;
+out:
+	close(dirfd);
+	return err;
 }
 
 static void intel_pt_min_max_sample_sz(struct evlist *evlist,
@@ -625,7 +604,7 @@ static int intel_pt_recording_options(struct auxtrace_record *itr,
 	struct perf_pmu *intel_pt_pmu = ptr->intel_pt_pmu;
 	bool have_timing_info, need_immediate = false;
 	struct evsel *evsel, *intel_pt_evsel = NULL;
-	const struct perf_cpu_map *cpus = evlist->core.cpus;
+	const struct perf_cpu_map *cpus = evlist->core.user_requested_cpus;
 	bool privileged = perf_event_paranoid_check(-1);
 	u64 tsc_bit;
 	int err;
@@ -642,6 +621,7 @@ static int intel_pt_recording_options(struct auxtrace_record *itr,
 			evsel->core.attr.freq = 0;
 			evsel->core.attr.sample_period = 1;
 			evsel->no_aux_samples = true;
+			evsel->needs_auxtrace_mmap = true;
 			intel_pt_evsel = evsel;
 			opts->full_auxtrace = true;
 		}
@@ -669,7 +649,8 @@ static int intel_pt_recording_options(struct auxtrace_record *itr,
 		return 0;
 
 	if (opts->auxtrace_sample_mode)
-		intel_pt_config_sample_mode(intel_pt_pmu, intel_pt_evsel);
+		evsel__set_config_if_unset(intel_pt_pmu, intel_pt_evsel,
+					   "psb_period", 0);
 
 	err = intel_pt_validate_config(intel_pt_pmu, intel_pt_evsel);
 	if (err)
@@ -776,8 +757,13 @@ static int intel_pt_recording_options(struct auxtrace_record *itr,
 		}
 	}
 
-	intel_pt_parse_terms(intel_pt_pmu->name, &intel_pt_pmu->format,
-			     "tsc", &tsc_bit);
+	if (!opts->auxtrace_snapshot_mode && !opts->auxtrace_sample_mode) {
+		u32 aux_watermark = opts->auxtrace_mmap_pages * page_size / 4;
+
+		intel_pt_evsel->core.attr.aux_watermark = aux_watermark;
+	}
+
+	intel_pt_parse_terms(intel_pt_pmu, "tsc", &tsc_bit);
 
 	if (opts->full_auxtrace && (intel_pt_evsel->core.attr.config & tsc_bit))
 		have_timing_info = true;
@@ -788,7 +774,7 @@ static int intel_pt_recording_options(struct auxtrace_record *itr,
 	 * Per-cpu recording needs sched_switch events to distinguish different
 	 * threads.
 	 */
-	if (have_timing_info && !perf_cpu_map__empty(cpus) &&
+	if (have_timing_info && !perf_cpu_map__is_any_cpu_or_is_empty(cpus) &&
 	    !record_opts__no_switch_events(opts)) {
 		if (perf_can_record_switch_events()) {
 			bool cpu_wide = !target__none(&opts->target) &&
@@ -797,18 +783,11 @@ static int intel_pt_recording_options(struct auxtrace_record *itr,
 			if (!cpu_wide && perf_can_record_cpu_wide()) {
 				struct evsel *switch_evsel;
 
-				err = parse_events(evlist, "dummy:u", NULL);
-				if (err)
-					return err;
+				switch_evsel = evlist__add_dummy_on_all_cpus(evlist);
+				if (!switch_evsel)
+					return -ENOMEM;
 
-				switch_evsel = evlist__last(evlist);
-
-				switch_evsel->core.attr.freq = 0;
-				switch_evsel->core.attr.sample_period = 1;
 				switch_evsel->core.attr.context_switch = 1;
-
-				switch_evsel->core.system_wide = true;
-				switch_evsel->no_aux_samples = true;
 				switch_evsel->immediate = true;
 
 				evsel__set_sample_bit(switch_evsel, TID);
@@ -827,6 +806,7 @@ static int intel_pt_recording_options(struct auxtrace_record *itr,
 					ptr->have_sched_switch = 2;
 			}
 		} else {
+#ifdef HAVE_LIBTRACEEVENT
 			err = intel_pt_track_switches(evlist);
 			if (err == -EPERM)
 				pr_debug2("Unable to select sched:sched_switch\n");
@@ -834,6 +814,7 @@ static int intel_pt_recording_options(struct auxtrace_record *itr,
 				return err;
 			else
 				ptr->have_sched_switch = 1;
+#endif
 		}
 	}
 
@@ -846,36 +827,38 @@ static int intel_pt_recording_options(struct auxtrace_record *itr,
 		 * To obtain the auxtrace buffer file descriptor, the auxtrace
 		 * event must come first.
 		 */
-		perf_evlist__to_front(evlist, intel_pt_evsel);
+		evlist__to_front(evlist, intel_pt_evsel);
 		/*
 		 * In the case of per-cpu mmaps, we need the CPU on the
 		 * AUX event.
 		 */
-		if (!perf_cpu_map__empty(cpus))
+		if (!perf_cpu_map__is_any_cpu_or_is_empty(cpus))
 			evsel__set_sample_bit(intel_pt_evsel, CPU);
 	}
 
 	/* Add dummy event to keep tracking */
 	if (opts->full_auxtrace) {
+		bool need_system_wide_tracking;
 		struct evsel *tracking_evsel;
 
-		err = parse_events(evlist, "dummy:u", NULL);
-		if (err)
-			return err;
-
-		tracking_evsel = evlist__last(evlist);
+		/*
+		 * User space tasks can migrate between CPUs, so when tracing
+		 * selected CPUs, sideband for all CPUs is still needed.
+		 */
+		need_system_wide_tracking = opts->target.cpu_list &&
+					    !intel_pt_evsel->core.attr.exclude_user;
 
-		perf_evlist__set_tracking_event(evlist, tracking_evsel);
+		tracking_evsel = evlist__add_aux_dummy(evlist, need_system_wide_tracking);
+		if (!tracking_evsel)
+			return -ENOMEM;
 
-		tracking_evsel->core.attr.freq = 0;
-		tracking_evsel->core.attr.sample_period = 1;
+		evlist__set_tracking_event(evlist, tracking_evsel);
 
-		tracking_evsel->no_aux_samples = true;
 		if (need_immediate)
 			tracking_evsel->immediate = true;
 
 		/* In per-cpu case, always need the time of mmap events etc */
-		if (!perf_cpu_map__empty(cpus)) {
+		if (!perf_cpu_map__is_any_cpu_or_is_empty(cpus)) {
 			evsel__set_sample_bit(tracking_evsel, TIME);
 			/* And the CPU for switch events */
 			evsel__set_sample_bit(tracking_evsel, CPU);
@@ -887,7 +870,7 @@ static int intel_pt_recording_options(struct auxtrace_record *itr,
 	 * Warn the user when we do not have enough information to decode i.e.
 	 * per-cpu with no sched_switch (except workload-only).
 	 */
-	if (!ptr->have_sched_switch && !perf_cpu_map__empty(cpus) &&
+	if (!ptr->have_sched_switch && !perf_cpu_map__is_any_cpu_or_is_empty(cpus) &&
 	    !target__none(&opts->target) &&
 	    !intel_pt_evsel->core.attr.exclude_user)
 		ui__warning("Intel Processor Trace decoding will not be possible except for kernel tracing!\n");
@@ -1183,7 +1166,7 @@ static u64 intel_pt_reference(struct auxtrace_record *itr __maybe_unused)
 
 struct auxtrace_record *intel_pt_recording_init(int *err)
 {
-	struct perf_pmu *intel_pt_pmu = perf_pmu__find(INTEL_PT_PMU_NAME);
+	struct perf_pmu *intel_pt_pmu = perf_pmus__find(INTEL_PT_PMU_NAME);
 	struct intel_pt_recording *ptr;
 
 	if (!intel_pt_pmu)