As of now it doesn't consider cgroups when collecting shadow stats and
metrics so counter values from different cgroups will be saved in a
same slot. This resulted in an incorrect numbers when those cgroups
have different workloads.
For example, let's look at the below - the cgroup A and C runs same
workload which burns a cpu while cgroup B runs a light workload.
$ perf stat -a -e cycles,instructions --for-each-cgroup A,B,C sleep 1
Performance counter stats for 'system wide':
3,958,116,522 cyclesA
6,722,650,929 instructions A #2.53 insn per cycle
1,132,741 cyclesB
571,743 instructions B #0.00 insn per cycle
4,007,799,935 cyclesC
6,793,181,523 instructions C #2.56 insn per cycle
1.001050869 seconds time elapsed
When I run perf stat with single workload, it usually shows IPC around 1.7.
We can verify it (6,722,650,929.0 / 3,958,116,522 = 1.698) for cgroup A.
But in this case, since cgroups are ignored, cycles are averaged so it
used the lower value for IPC calculation and resulted in around 2.5.
avg cycle: (3958116522 + 1132741 + 4007799935) / 3 = 2655683066
IPC (A) : 6722650929 / 2655683066 = 2.531
IPC (B) : 571743 / 2655683066 = 0.0002
IPC (C) : 6793181523 / 2655683066 = 2.557
We can simply compare cgroup pointers in the evsel and it'll be NULL
when cgroups are not specified. With this patch, I can see correct
numbers like below:
$ perf stat -a -e cycles,instructions --for-each-cgroup A,B,C sleep 1
Performance counter stats for 'system wide':
4,171,051,687 cyclesA
7,219,793,922 instructions A #1.73 insn per cycle
1,051,189 cyclesB
583,102 instructions B #0.55 insn per cycle
4,171,124,710 cyclesC
7,192,944,580 instructions C #1.72 insn per cycle
1.007909814 seconds time elapsed
Signed-off-by: Namhyung Kim
---
tools/perf/util/stat-shadow.c | 243 ++
1 file changed, 132 insertions(+), 111 deletions(-)
diff --git a/tools/perf/util/stat-shadow.c b/tools/perf/util/stat-shadow.c
index 901265127e36..10d0f5a0fd4a 100644
--- a/tools/perf/util/stat-shadow.c
+++ b/tools/perf/util/stat-shadow.c
@@ -8,6 +8,7 @@
#include "evlist.h"
#include "expr.h"
#include "metricgroup.h"
+#include "cgroup.h"
#include
/*
@@ -28,6 +29,7 @@ struct saved_value {
enum stat_type type;
int ctx;
int cpu;
+ struct cgroup *cgrp;
struct runtime_stat *stat;
struct stats stats;
u64 metric_total;
@@ -57,6 +59,9 @@ static int saved_value_cmp(struct rb_node *rb_node, const
void *entry)
if (a->ctx != b->ctx)
return a->ctx - b->ctx;
+ if (a->cgrp != b->cgrp)
+ return (char *)a->cgrp < (char *)b->cgrp ? -1 : +1;
+
if (a->evsel == NULL && b->evsel == NULL) {
if (a->stat == b->stat)
return 0;
@@ -100,7 +105,8 @@ static struct saved_value *saved_value_lookup(struct evsel
*evsel,
bool create,
enum stat_type type,
int ctx,
- struct runtime_stat *st)
+ struct runtime_stat *st,
+ struct cgroup *cgrp)
{
struct rblist *rblist;
struct rb_node *nd;
@@ -110,6 +116,7 @@ static struct saved_value *saved_value_lookup(struct evsel
*evsel,
.type = type,
.ctx = ctx,
.stat = st,
+ .cgrp = cgrp,
};
rblist = >value_list;
@@ -193,10 +200,11 @@ void perf_stat__reset_shadow_per_stat(struct runtime_stat
*st)
static void update_runtime_stat(struct runtime_stat *st,
enum stat_type type,
- int ctx, int cpu, u64 count)
+ int ctx, int cpu, u64 count,
+ struct cgroup *cgrp)
{
struct saved_value *v = saved_value_lookup(NULL, cpu, true,
- type, ctx, st);
+ type, ctx, st, cgrp);
if (v)
update_stats(>stats, count);
@@ -212,80 +220,81 @@ void perf_stat__update_shadow_stats(struct evsel
*counter, u64 count,
{
int ctx = evsel_context(counter);
u64 count_ns = count;
+ struct cgroup *cgrp = counter->cgrp;
struct saved_value *v;
count *= counter->scale;
if (evsel__is_clock(counter))
- update_runtime_stat(st, STAT_NSECS, 0, cpu, count_ns);
+
