Re: [PATCHv2 5/8] perf tools: Get precise_ip from the pmu config

[Jiri Olsa]

On Thu, Mar 07, 2019 at 08:51:23AM -0800, Andi Kleen wrote:
> On Thu, Mar 07, 2019 at 04:35:00PM +0100, Jiri Olsa wrote:
> > On Tue, Mar 05, 2019 at 08:40:17AM -0800, Andi Kleen wrote:
> > > On Tue, Mar 05, 2019 at 05:28:54PM +0100, Jiri Olsa wrote:
> > > > On Tue, Mar 05, 2019 at 08:13:19AM -0800, Andi Kleen wrote:
> > > > > On Tue, Mar 05, 2019 at 04:25:33PM +0100, Jiri Olsa wrote:
> > > > > > Getting precise_ip field from the perf_pmu::max_precise
> > > > > > config read from sysfs. If it's not available falling
> > > > > > back to current detection function.
> > > > > 
> > > > > max_precise depends on the event.  This won't work for all
> > > > > events. For example only instructions and cycles support
> > > > > ppp
> > > > 
> > > > I'm getting  precise_ip=3 on mem-* events as well, that's why I
> > > > was fixing this.. now it's not working for any event
> > > 
> > > I don't think it means anything for mem-*
> > > 
> > > There's some support for it on Goldmont plus for other events,
> > > but it doesn't support mem-*. On big core it's only
> > > for instructions and cycles, all implemented with the same
> > > event. All other PEBS events only have two levels
> > > switching between the two IPs.
> > 
> > ok, so how about this, it's the change I posted merged with the patch
> 
> Still seems like a hack. Even though I don't know of a case that
> would break now. But it would be better to have the precise probing
> in the open retry loop, instead of trying to reinvent it here.

how about something like below?

jirka


---
 tools/perf/util/evlist.c | 29 ----------------
 tools/perf/util/evlist.h |  2 --
 tools/perf/util/evsel.c  | 73 +++++++++++++++++++++++++++++++++-------
 3 files changed, 60 insertions(+), 44 deletions(-)

diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
index ed20f4379956..3a243c249c29 100644
--- a/tools/perf/util/evlist.c
+++ b/tools/perf/util/evlist.c
@@ -230,35 +230,6 @@ void perf_evlist__set_leader(struct perf_evlist *evlist)
 	}
 }
 
-void perf_event_attr__set_max_precise_ip(struct perf_event_attr *pattr)
-{
-	struct perf_event_attr attr = {
-		.type		= PERF_TYPE_HARDWARE,
-		.config		= PERF_COUNT_HW_CPU_CYCLES,
-		.exclude_kernel	= 1,
-		.precise_ip	= 3,
-	};
-
-	event_attr_init(&attr);
-
-	/*
-	 * Unnamed union member, not supported as struct member named
-	 * initializer in older compilers such as gcc 4.4.7
-	 */
-	attr.sample_period = 1;
-
-	while (attr.precise_ip != 0) {
-		int fd = sys_perf_event_open(&attr, 0, -1, -1, 0);
-		if (fd != -1) {
-			close(fd);
-			break;
-		}
-		--attr.precise_ip;
-	}
-
-	pattr->precise_ip = attr.precise_ip;
-}
-
 int __perf_evlist__add_default(struct perf_evlist *evlist, bool precise)
 {
 	struct perf_evsel *evsel = perf_evsel__new_cycles(precise);
diff --git a/tools/perf/util/evlist.h b/tools/perf/util/evlist.h
index 744906dd4887..cb15e3366eba 100644
--- a/tools/perf/util/evlist.h
+++ b/tools/perf/util/evlist.h
@@ -303,8 +303,6 @@ void perf_evlist__to_front(struct perf_evlist *evlist,
 void perf_evlist__set_tracking_event(struct perf_evlist *evlist,
 				     struct perf_evsel *tracking_evsel);
 
-void perf_event_attr__set_max_precise_ip(struct perf_event_attr *attr);
-
 struct perf_evsel *
 perf_evlist__find_evsel_by_str(struct perf_evlist *evlist, const char *str);
 
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index 3bbf73e979c0..7cd70f99201a 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -295,7 +295,6 @@ struct perf_evsel *perf_evsel__new_cycles(bool precise)
 	if (!precise)
 		goto new_event;
 
-	perf_event_attr__set_max_precise_ip(&attr);
 	/*
 	 * Now let the usual logic to set up the perf_event_attr defaults
 	 * to kick in when we return and before perf_evsel__open() is called.
@@ -305,6 +304,8 @@ struct perf_evsel *perf_evsel__new_cycles(bool precise)
 	if (evsel == NULL)
 		goto out;
 
+	evsel->precise_max = true;
+
 	/* use asprintf() because free(evsel) assumes name is allocated */
 	if (asprintf(&evsel->name, "cycles%s%s%.*s",
 		     (attr.precise_ip || attr.exclude_kernel) ? ":" : "",
@@ -1083,7 +1084,7 @@ void perf_evsel__config(struct perf_evsel *evsel, struct record_opts *opts,
 	}
 
 	if (evsel->precise_max)
-		perf_event_attr__set_max_precise_ip(attr);
+		attr->precise_ip = 3;
 
 	if (opts->all_user) {
 		attr->exclude_kernel = 1;
@@ -1749,6 +1750,60 @@ static bool ignore_missing_thread(struct perf_evsel *evsel,
 	return true;
 }
 
+static void display_attr(struct perf_event_attr *attr)
+{
+	if (verbose >= 2) {
+		fprintf(stderr, "%.60s\n", graph_dotted_line);
+		fprintf(stderr, "perf_event_attr:\n");
+		perf_event_attr__fprintf(stderr, attr, __open_attr__fprintf, NULL);
+		fprintf(stderr, "%.60s\n", graph_dotted_line);
+	}
+}
+
+static int perf_event_open(struct perf_evsel *evsel,
+			   pid_t pid, int cpu, int group_fd,
+			   unsigned long flags)
+{
+	int precise_ip = evsel->attr.precise_ip;
+	int fd;
+
+	while (1) {
+		pr_debug2("sys_perf_event_open: pid %d  cpu %d  group_fd %d  flags %#lx",
+			  pid, cpu, group_fd, flags);
+
+		fd = sys_perf_event_open(&evsel->attr, pid, cpu, group_fd, flags);
+		if (fd >= 0)
+			break;
+
+		/*
+		 * Do quick precise_ip fallback if requested:
+		 *  - there is some precise_ip
+		 *  - maximum precise is requested
+		 *  - we failed with ENOTSUP error, which is
+		 *    associated with wrong precise_ip config
+		 */
+		if (!precise_ip || !evsel->precise_max || (errno != ENOTSUP))
+			break;
+
+		/*
+		 * We tried all the precise_ip values, and it's
+		 * still failing, so leave it to standard fallabck.
+		 */
+		if (!evsel->attr.precise_ip) {
+			evsel->attr.precise_ip = precise_ip;
+			break;
+		}
+
+		pr_debug2("\nsys_perf_event_open failed, error %d\n", -ENOTSUP);
+
+		evsel->attr.precise_ip--;
+		pr_debug2("decreasing precise_ip by one (%d)\n", evsel->attr.precise_ip);
+		display_attr(&evsel->attr);
+	}
+
+	return fd;
+}
+
 int perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus,
 		     struct thread_map *threads)
 {
@@ -1824,12 +1879,7 @@ int perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus,
 	if (perf_missing_features.sample_id_all)
 		evsel->attr.sample_id_all = 0;
 
-	if (verbose >= 2) {
-		fprintf(stderr, "%.60s\n", graph_dotted_line);
-		fprintf(stderr, "perf_event_attr:\n");
-		perf_event_attr__fprintf(stderr, &evsel->attr, __open_attr__fprintf, NULL);
-		fprintf(stderr, "%.60s\n", graph_dotted_line);
-	}
+	display_attr(&evsel->attr);
 
 	for (cpu = 0; cpu < cpus->nr; cpu++) {
 
@@ -1841,13 +1891,10 @@ int perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus,
 
 			group_fd = get_group_fd(evsel, cpu, thread);
 retry_open:
-			pr_debug2("sys_perf_event_open: pid %d  cpu %d  group_fd %d  flags %#lx",
-				  pid, cpus->map[cpu], group_fd, flags);
-
 			test_attr__ready();
 
-			fd = sys_perf_event_open(&evsel->attr, pid, cpus->map[cpu],
-						 group_fd, flags);
+			fd = perf_event_open(evsel, pid, cpus->map[cpu],
+					     group_fd, flags);
 
 			FD(evsel, cpu, thread) = fd;
 
-- 
2.17.2