[PATCH v2] tools/perf/metricgroup: Fix printing event names of metric group with multiple events incase of overlapping events

* [PATCH v2] tools/perf/metricgroup: Fix printing event names of metric group with multiple events incase of overlapping events
@ 2020-01-22  6:47 Kajol Jain
  2020-01-27 15:41 ` Jiri Olsa
  0 siblings, 1 reply; 4+ messages in thread
From: Kajol Jain @ 2020-01-22  6:47 UTC (permalink / raw)
  To: acme
  Cc: linux-kernel, linux-perf-users, kjain, Alexander Shishkin,
	Andi Kleen, Jiri Olsa, Kan Liang, Peter Zijlstra, Jin Yao,
	Madhavan Srinivasan, Anju T Sudhakar, Ravi Bangoria

Commit f01642e4912b ("perf metricgroup: Support multiple
events for metricgroup") introduced support for multiple events
in a metric group. But with the current upstream, metric events
names are not printed properly incase we try to run multiple
metric groups with overlapping event.

With current upstream version, incase of overlapping metric events
issue is, we always start our comparision logic from start.
So, the events which already matched with some metric group also
take part in comparision logic. Because of that when we have overlapping
events, we end up matching current metric group event with already matched
one.

For example, in skylake machine we have metric event CoreIPC and
Instructions. Both of them need 'inst_retired.any' event value.
As events in Instructions is subset of events in CoreIPC, they
endup in pointing to same 'inst_retired.any' value.

In skylake platform:

command:# ./perf stat -M CoreIPC,Instructions  -C 0 sleep 1

 Performance counter stats for 'CPU(s) 0':

     1,254,992,790      inst_retired.any          # 1254992790.0
                                                    Instructions
                                                  #      1.3 CoreIPC
       977,172,805      cycles
     1,254,992,756      inst_retired.any

       1.000802596 seconds time elapsed

command:# sudo ./perf stat -M UPI,IPC sleep 1

   Performance counter stats for 'sleep 1':

           948,650      uops_retired.retire_slots
           866,182      inst_retired.any          #      0.7 IPC
           866,182      inst_retired.any
         1,175,671      cpu_clk_unhalted.thread

Patch fixes the issue by adding a new variable 'evlist_iter' in evlist
itself to keep track of events which already matched with some group.
It points to event in perf_evlist from where next match should start.
Because we need to make sure, we match correct set of events belongs to
corresponding metric group.

With this patch:
In skylake platform:

command:# ./perf stat -M CoreIPC,Instructions  -C 0 sleep 1

 Performance counter stats for 'CPU(s) 0':

       149,481,533      inst_retired.any          #      0.8 CoreIPC
       186,244,218      cycles
       149,479,362      inst_retired.any          # 149479362.0
                                                        Instructions

       1.001655885 seconds time elapsed

command:# ./perf stat -M UPI,IPC sleep 1
 Performance counter stats for 'CPU(s) 0':

        16,858,849      uops_retired.retire_slots #      1.3 UPI
        12,529,178      inst_retired.any
        12,529,558      inst_retired.any          #      0.3 IPC
        39,936,071      cpu_clk_unhalted.thread

       1.001413978 seconds time elapsed

Signed-off-by: Kajol Jain <kjain@linux.ibm.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Jin Yao <yao.jin@linux.intel.com>
Cc: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Cc: Anju T Sudhakar <anju@linux.vnet.ibm.com>
Cc: Ravi Bangoria <ravi.bangoria@linux.ibm.com>
---
 tools/perf/util/evlist.h      |  1 +
 tools/perf/util/metricgroup.c | 14 +++++++++++++-
 2 files changed, 14 insertions(+), 1 deletion(-)

---
Changelog:
v1 -> v2
- Rather then adding static variable in metricgroup.c,
  add a new variable in evlist itself with name 'evlist_iter'
---

diff --git a/tools/perf/util/evlist.h b/tools/perf/util/evlist.h
index f5bd5c386df1..255f872aee92 100644
--- a/tools/perf/util/evlist.h
+++ b/tools/perf/util/evlist.h
@@ -54,6 +54,7 @@ struct evlist {
 	bool		 enabled;
 	int		 id_pos;
 	int		 is_pos;
+	int		 evlist_iter;
 	u64		 combined_sample_type;
 	enum bkw_mmap_state bkw_mmap_state;
 	struct {
diff --git a/tools/perf/util/metricgroup.c b/tools/perf/util/metricgroup.c
index 02aee946b6c1..911fab4ac04b 100644
--- a/tools/perf/util/metricgroup.c
+++ b/tools/perf/util/metricgroup.c
@@ -96,10 +96,13 @@ static struct evsel *find_evsel_group(struct evlist *perf_evlist,
 				      struct evsel **metric_events)
 {
 	struct evsel *ev;
-	int i = 0;
+	int i = 0, j = 0;
 	bool leader_found;
 
 	evlist__for_each_entry (perf_evlist, ev) {
+		j++;
+		if (j <= perf_evlist->evlist_iter)
+			continue;
 		if (!strcmp(ev->name, ids[i])) {
 			if (!metric_events[i])
 				metric_events[i] = ev;
@@ -147,6 +150,8 @@ static struct evsel *find_evsel_group(struct evlist *perf_evlist,
 		}
 	}
 
+	perf_evlist->evlist_iter = j;
+
 	return metric_events[0];
 }
 
@@ -161,6 +166,13 @@ static int metricgroup__setup_events(struct list_head *groups,
 	struct egroup *eg;
 	struct evsel *evsel;
 
+	/*
+	 * Reset the evlist_iter. This is needed because
+	 * when we start comparision logic in find_evsel_group(),
+	 * need to start it from very first event present in evlist.
+	 */
+	perf_evlist->evlist_iter = 0;
+
 	list_for_each_entry (eg, groups, nd) {
 		struct evsel **metric_events;
 
-- 
2.21.0


^ permalink raw reply related	[flat|nested] 4+ messages in thread