tjh.dev
Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
perf jevents: Add set of common metrics based on default ones
Add support to getting a common set of metrics from a default
table. It simplifies the generation to add json metrics at the same
time. The metrics added are CPUs_utilized, cs_per_second,
migrations_per_second, page_faults_per_second, insn_per_cycle,
stalled_cycles_per_instruction, frontend_cycles_idle,
backend_cycles_idle, cycles_frequency, branch_frequency and
branch_miss_rate based on the shadow metric definitions.
Following this change the default perf stat output on an alderlake
looks like:
```
$ perf stat -a -- sleep 2
Performance counter stats for 'system wide':
0.00 msec cpu-clock # 0.000 CPUs utilized
77,739 context-switches
15,033 cpu-migrations
321,313 page-faults
14,355,634,225 cpu_atom/instructions/ # 1.40 insn per cycle (35.37%)
134,561,560,583 cpu_core/instructions/ # 3.44 insn per cycle (57.85%)
10,263,836,145 cpu_atom/cycles/ (35.42%)
39,138,632,894 cpu_core/cycles/ (57.60%)
2,989,658,777 cpu_atom/branches/ (42.60%)
32,170,570,388 cpu_core/branches/ (57.39%)
29,789,870 cpu_atom/branch-misses/ # 1.00% of all branches (42.69%)
165,991,152 cpu_core/branch-misses/ # 0.52% of all branches (57.19%)
(software) # nan cs/sec cs_per_second
TopdownL1 (cpu_core) # 11.9 % tma_bad_speculation
# 19.6 % tma_frontend_bound (63.97%)
TopdownL1 (cpu_core) # 18.8 % tma_backend_bound
# 49.7 % tma_retiring (63.97%)
(software) # nan faults/sec page_faults_per_second
# nan GHz cycles_frequency (42.88%)
# nan GHz cycles_frequency (69.88%)
TopdownL1 (cpu_atom) # 11.7 % tma_bad_speculation
# 29.9 % tma_retiring (50.07%)
TopdownL1 (cpu_atom) # 31.3 % tma_frontend_bound (43.09%)
(cpu_atom) # nan M/sec branch_frequency (43.09%)
# nan M/sec branch_frequency (70.07%)
# nan migrations/sec migrations_per_second
TopdownL1 (cpu_atom) # 27.1 % tma_backend_bound (43.08%)
(software) # 0.0 CPUs CPUs_utilized
# 1.4 instructions insn_per_cycle (43.04%)
# 3.5 instructions insn_per_cycle (69.99%)
# 1.0 % branch_miss_rate (35.46%)
# 0.5 % branch_miss_rate (65.02%)
2.005626564 seconds time elapsed
```
Signed-off-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
authored by
Ian Rogers
and committed by
Namhyung Kim
c7adeb09
2e514084
+212
-42
+86
tools/perf/pmu-events/arch/common/common/metrics.json
+86
tools/perf/pmu-events/arch/common/common/metrics.json
···
1
+
[
2
+
{
3
+
"BriefDescription": "Average CPU utilization",
4
+
"MetricExpr": "(software@cpu\\-clock\\,name\\=cpu\\-clock@ if #target_cpu else software@task\\-clock\\,name\\=task\\-clock@) / (duration_time * 1e9)",
5
+
"MetricGroup": "Default",
6
+
"MetricName": "CPUs_utilized",
7
+
"ScaleUnit": "1CPUs",
8
+
"MetricConstraint": "NO_GROUP_EVENTS"
9
+
},
10
+
{
11
+
"BriefDescription": "Context switches per CPU second",
12
+
"MetricExpr": "(software@context\\-switches\\,name\\=context\\-switches@ * 1e9) / (software@cpu\\-clock\\,name\\=cpu\\-clock@ if #target_cpu else software@task\\-clock\\,name\\=task\\-clock@)",
13
+
"MetricGroup": "Default",
14
+
"MetricName": "cs_per_second",
15
+
"ScaleUnit": "1cs/sec",
16
+
"MetricConstraint": "NO_GROUP_EVENTS"
17
+
},
18
+
{
19
+
"BriefDescription": "Process migrations to a new CPU per CPU second",
20
+
"MetricExpr": "(software@cpu\\-migrations\\,name\\=cpu\\-migrations@ * 1e9) / (software@cpu\\-clock\\,name\\=cpu\\-clock@ if #target_cpu else software@task\\-clock\\,name\\=task\\-clock@)",
21
+
"MetricGroup": "Default",
22
+
"MetricName": "migrations_per_second",
23
+
"ScaleUnit": "1migrations/sec",
24
+
"MetricConstraint": "NO_GROUP_EVENTS"
25
+
},
26
+
{
27
+
"BriefDescription": "Page faults per CPU second",
28
+
"MetricExpr": "(software@page\\-faults\\,name\\=page\\-faults@ * 1e9) / (software@cpu\\-clock\\,name\\=cpu\\-clock@ if #target_cpu else software@task\\-clock\\,name\\=task\\-clock@)",
29
+
"MetricGroup": "Default",
30
+
"MetricName": "page_faults_per_second",
31
+
"ScaleUnit": "1faults/sec",
32
+
"MetricConstraint": "NO_GROUP_EVENTS"
33
+
},
34
+
{
35
+
"BriefDescription": "Instructions Per Cycle",
36
+
"MetricExpr": "instructions / cpu\\-cycles",
37
+
"MetricGroup": "Default",
38
+
"MetricName": "insn_per_cycle",
39
+
"MetricThreshold": "insn_per_cycle < 1",
40
+
"ScaleUnit": "1instructions"
41
+
},
42
+
{
43
+
"BriefDescription": "Max front or backend stalls per instruction",
44
+
"MetricExpr": "max(stalled\\-cycles\\-frontend, stalled\\-cycles\\-backend) / instructions",
45
+
"MetricGroup": "Default",
46
+
"MetricName": "stalled_cycles_per_instruction"
47
+
},
48
+
{
49
+
"BriefDescription": "Frontend stalls per cycle",
50
+
"MetricExpr": "stalled\\-cycles\\-frontend / cpu\\-cycles",
51
+
"MetricGroup": "Default",
52
+
"MetricName": "frontend_cycles_idle",
53
+
"MetricThreshold": "frontend_cycles_idle > 0.1"
54
+
},
55
+
{
56
+
"BriefDescription": "Backend stalls per cycle",
57
+
"MetricExpr": "stalled\\-cycles\\-backend / cpu\\-cycles",
58
+
"MetricGroup": "Default",
59
+
"MetricName": "backend_cycles_idle",
60
+
"MetricThreshold": "backend_cycles_idle > 0.2"
61
+
},
62
+
{
63
+
"BriefDescription": "Cycles per CPU second",
64
+
"MetricExpr": "cpu\\-cycles / (software@cpu\\-clock\\,name\\=cpu\\-clock@ if #target_cpu else software@task\\-clock\\,name\\=task\\-clock@)",
65
+
"MetricGroup": "Default",
66
+
"MetricName": "cycles_frequency",
67
+
"ScaleUnit": "1GHz",
68
+
"MetricConstraint": "NO_GROUP_EVENTS"
69
+
},
70
+
{
71
+
"BriefDescription": "Branches per CPU second",
72
+
"MetricExpr": "branches / (software@cpu\\-clock\\,name\\=cpu\\-clock@ if #target_cpu else software@task\\-clock\\,name\\=task\\-clock@)",
73
+
"MetricGroup": "Default",
74
+
"MetricName": "branch_frequency",
75
+
"ScaleUnit": "1000M/sec",
76
+
"MetricConstraint": "NO_GROUP_EVENTS"
77
+
},
78
+
{
79
+
"BriefDescription": "Branch miss rate",
80
+
"MetricExpr": "branch\\-misses / branches",
81
+
"MetricGroup": "Default",
82
+
"MetricName": "branch_miss_rate",
83
+
"MetricThreshold": "branch_miss_rate > 0.05",
84
+
"ScaleUnit": "100%"
85
+
}
86
+
]
+84
-31
tools/perf/pmu-events/empty-pmu-events.c
+84
-31
tools/perf/pmu-events/empty-pmu-events.c
···
1303
1303
/* offset=127519 */ "sys_ccn_pmu.read_cycles\000uncore\000ccn read-cycles event\000config=0x2c\0000x01\00000\000\000\000\000\000"
1304
1304
/* offset=127596 */ "uncore_sys_cmn_pmu\000"
1305
1305
/* offset=127615 */ "sys_cmn_pmu.hnf_cache_miss\000uncore\000Counts total cache misses in first lookup result (high priority)\000eventid=1,type=5\000(434|436|43c|43a).*\00000\000\000\000\000\000"
1306
-
/* offset=127758 */ "CPI\000\0001 / IPC\000\000\000\000\000\000\000\00000"
1307
-
/* offset=127780 */ "IPC\000group1\000inst_retired.any / cpu_clk_unhalted.thread\000\000\000\000\000\000\000\00000"
1308
-
/* offset=127843 */ "Frontend_Bound_SMT\000\000idq_uops_not_delivered.core / (4 * (cpu_clk_unhalted.thread / 2 * (1 + cpu_clk_unhalted.one_thread_active / cpu_clk_unhalted.ref_xclk)))\000\000\000\000\000\000\000\00000"
1309
-
/* offset=128009 */ "dcache_miss_cpi\000\000l1d\\-loads\\-misses / inst_retired.any\000\000\000\000\000\000\000\00000"
1310
-
/* offset=128073 */ "icache_miss_cycles\000\000l1i\\-loads\\-misses / inst_retired.any\000\000\000\000\000\000\000\00000"
1311
-
/* offset=128140 */ "cache_miss_cycles\000group1\000dcache_miss_cpi + icache_miss_cycles\000\000\000\000\000\000\000\00000"
1312
-
/* offset=128211 */ "DCache_L2_All_Hits\000\000l2_rqsts.demand_data_rd_hit + l2_rqsts.pf_hit + l2_rqsts.rfo_hit\000\000\000\000\000\000\000\00000"
1313
-
/* offset=128305 */ "DCache_L2_All_Miss\000\000max(l2_rqsts.all_demand_data_rd - l2_rqsts.demand_data_rd_hit, 0) + l2_rqsts.pf_miss + l2_rqsts.rfo_miss\000\000\000\000\000\000\000\00000"
1314
-
/* offset=128439 */ "DCache_L2_All\000\000DCache_L2_All_Hits + DCache_L2_All_Miss\000\000\000\000\000\000\000\00000"
1315
-
/* offset=128503 */ "DCache_L2_Hits\000\000d_ratio(DCache_L2_All_Hits, DCache_L2_All)\000\000\000\000\000\000\000\00000"
1316
-
/* offset=128571 */ "DCache_L2_Misses\000\000d_ratio(DCache_L2_All_Miss, DCache_L2_All)\000\000\000\000\000\000\000\00000"
1317
-
/* offset=128641 */ "M1\000\000ipc + M2\000\000\000\000\000\000\000\00000"
1318
-
/* offset=128663 */ "M2\000\000ipc + M1\000\000\000\000\000\000\000\00000"
1319
-
/* offset=128685 */ "M3\000\0001 / M3\000\000\000\000\000\000\000\00000"
1320
-
/* offset=128705 */ "L1D_Cache_Fill_BW\000\00064 * l1d.replacement / 1e9 / duration_time\000\000\000\000\000\000\000\00000"
1306
+
/* offset=127758 */ "CPUs_utilized\000Default\000(software@cpu\\-clock\\,name\\=cpu\\-clock@ if #target_cpu else software@task\\-clock\\,name\\=task\\-clock@) / (duration_time * 1e9)\000\000Average CPU utilization\000\0001CPUs\000\000\000\00001"
1307
+
/* offset=127943 */ "cs_per_second\000Default\000software@context\\-switches\\,name\\=context\\-switches@ * 1e9 / (software@cpu\\-clock\\,name\\=cpu\\-clock@ if #target_cpu else software@task\\-clock\\,name\\=task\\-clock@)\000\000Context switches per CPU second\000\0001cs/sec\000\000\000\00001"
1308
+
/* offset=128175 */ "migrations_per_second\000Default\000software@cpu\\-migrations\\,name\\=cpu\\-migrations@ * 1e9 / (software@cpu\\-clock\\,name\\=cpu\\-clock@ if #target_cpu else software@task\\-clock\\,name\\=task\\-clock@)\000\000Process migrations to a new CPU per CPU second\000\0001migrations/sec\000\000\000\00001"
1309
+
/* offset=128434 */ "page_faults_per_second\000Default\000software@page\\-faults\\,name\\=page\\-faults@ * 1e9 / (software@cpu\\-clock\\,name\\=cpu\\-clock@ if #target_cpu else software@task\\-clock\\,name\\=task\\-clock@)\000\000Page faults per CPU second\000\0001faults/sec\000\000\000\00001"
1310
+
/* offset=128664 */ "insn_per_cycle\000Default\000instructions / cpu\\-cycles\000insn_per_cycle < 1\000Instructions Per Cycle\000\0001instructions\000\000\000\00000"
1311
+
/* offset=128776 */ "stalled_cycles_per_instruction\000Default\000max(stalled\\-cycles\\-frontend, stalled\\-cycles\\-backend) / instructions\000\000Max front or backend stalls per instruction\000\000\000\000\000\00000"
1312
+
/* offset=128939 */ "frontend_cycles_idle\000Default\000stalled\\-cycles\\-frontend / cpu\\-cycles\000frontend_cycles_idle > 0.1\000Frontend stalls per cycle\000\000\000\000\000\00000"
1313
+
/* offset=129068 */ "backend_cycles_idle\000Default\000stalled\\-cycles\\-backend / cpu\\-cycles\000backend_cycles_idle > 0.2\000Backend stalls per cycle\000\000\000\000\000\00000"
1314
+
/* offset=129193 */ "cycles_frequency\000Default\000cpu\\-cycles / (software@cpu\\-clock\\,name\\=cpu\\-clock@ if #target_cpu else software@task\\-clock\\,name\\=task\\-clock@)\000\000Cycles per CPU second\000\0001GHz\000\000\000\00001"
1315
+
/* offset=129368 */ "branch_frequency\000Default\000branches / (software@cpu\\-clock\\,name\\=cpu\\-clock@ if #target_cpu else software@task\\-clock\\,name\\=task\\-clock@)\000\000Branches per CPU second\000\0001000M/sec\000\000\000\00001"
1316
+
/* offset=129547 */ "branch_miss_rate\000Default\000branch\\-misses / branches\000branch_miss_rate > 0.05\000Branch miss rate\000\000100%\000\000\000\00000"
1317
+
/* offset=129650 */ "CPI\000\0001 / IPC\000\000\000\000\000\000\000\00000"
1318
+
/* offset=129672 */ "IPC\000group1\000inst_retired.any / cpu_clk_unhalted.thread\000\000\000\000\000\000\000\00000"
1319
+
/* offset=129735 */ "Frontend_Bound_SMT\000\000idq_uops_not_delivered.core / (4 * (cpu_clk_unhalted.thread / 2 * (1 + cpu_clk_unhalted.one_thread_active / cpu_clk_unhalted.ref_xclk)))\000\000\000\000\000\000\000\00000"
1320
+
/* offset=129901 */ "dcache_miss_cpi\000\000l1d\\-loads\\-misses / inst_retired.any\000\000\000\000\000\000\000\00000"
1321
+
/* offset=129965 */ "icache_miss_cycles\000\000l1i\\-loads\\-misses / inst_retired.any\000\000\000\000\000\000\000\00000"
1322
+
/* offset=130032 */ "cache_miss_cycles\000group1\000dcache_miss_cpi + icache_miss_cycles\000\000\000\000\000\000\000\00000"
1323
+
/* offset=130103 */ "DCache_L2_All_Hits\000\000l2_rqsts.demand_data_rd_hit + l2_rqsts.pf_hit + l2_rqsts.rfo_hit\000\000\000\000\000\000\000\00000"
1324
+
/* offset=130197 */ "DCache_L2_All_Miss\000\000max(l2_rqsts.all_demand_data_rd - l2_rqsts.demand_data_rd_hit, 0) + l2_rqsts.pf_miss + l2_rqsts.rfo_miss\000\000\000\000\000\000\000\00000"
1325
+
/* offset=130331 */ "DCache_L2_All\000\000DCache_L2_All_Hits + DCache_L2_All_Miss\000\000\000\000\000\000\000\00000"
1326
+
/* offset=130395 */ "DCache_L2_Hits\000\000d_ratio(DCache_L2_All_Hits, DCache_L2_All)\000\000\000\000\000\000\000\00000"
1327
+
/* offset=130463 */ "DCache_L2_Misses\000\000d_ratio(DCache_L2_All_Miss, DCache_L2_All)\000\000\000\000\000\000\000\00000"
1328
+
/* offset=130533 */ "M1\000\000ipc + M2\000\000\000\000\000\000\000\00000"
1329
+
/* offset=130555 */ "M2\000\000ipc + M1\000\000\000\000\000\000\000\00000"
1330
+
/* offset=130577 */ "M3\000\0001 / M3\000\000\000\000\000\000\000\00000"
1331
+
/* offset=130597 */ "L1D_Cache_Fill_BW\000\00064 * l1d.replacement / 1e9 / duration_time\000\000\000\000\000\000\000\00000"
1321
1332
;
1322
1333
1323
1334
static const struct compact_pmu_event pmu_events__common_default_core[] = {
···
2614
2603
},
2615
2604
};
2616
2605
2606
+
static const struct compact_pmu_event pmu_metrics__common_default_core[] = {
2607
+
{ 127758 }, /* CPUs_utilized\000Default\000(software@cpu\\-clock\\,name\\=cpu\\-clock@ if #target_cpu else software@task\\-clock\\,name\\=task\\-clock@) / (duration_time * 1e9)\000\000Average CPU utilization\000\0001CPUs\000\000\000\00001 */
2608
+
{ 129068 }, /* backend_cycles_idle\000Default\000stalled\\-cycles\\-backend / cpu\\-cycles\000backend_cycles_idle > 0.2\000Backend stalls per cycle\000\000\000\000\000\00000 */
2609
+
{ 129368 }, /* branch_frequency\000Default\000branches / (software@cpu\\-clock\\,name\\=cpu\\-clock@ if #target_cpu else software@task\\-clock\\,name\\=task\\-clock@)\000\000Branches per CPU second\000\0001000M/sec\000\000\000\00001 */
2610
+
{ 129547 }, /* branch_miss_rate\000Default\000branch\\-misses / branches\000branch_miss_rate > 0.05\000Branch miss rate\000\000100%\000\000\000\00000 */
2611
+
{ 127943 }, /* cs_per_second\000Default\000software@context\\-switches\\,name\\=context\\-switches@ * 1e9 / (software@cpu\\-clock\\,name\\=cpu\\-clock@ if #target_cpu else software@task\\-clock\\,name\\=task\\-clock@)\000\000Context switches per CPU second\000\0001cs/sec\000\000\000\00001 */
2612
+
{ 129193 }, /* cycles_frequency\000Default\000cpu\\-cycles / (software@cpu\\-clock\\,name\\=cpu\\-clock@ if #target_cpu else software@task\\-clock\\,name\\=task\\-clock@)\000\000Cycles per CPU second\000\0001GHz\000\000\000\00001 */
2613
+
{ 128939 }, /* frontend_cycles_idle\000Default\000stalled\\-cycles\\-frontend / cpu\\-cycles\000frontend_cycles_idle > 0.1\000Frontend stalls per cycle\000\000\000\000\000\00000 */
2614
+
{ 128664 }, /* insn_per_cycle\000Default\000instructions / cpu\\-cycles\000insn_per_cycle < 1\000Instructions Per Cycle\000\0001instructions\000\000\000\00000 */
2615
+
{ 128175 }, /* migrations_per_second\000Default\000software@cpu\\-migrations\\,name\\=cpu\\-migrations@ * 1e9 / (software@cpu\\-clock\\,name\\=cpu\\-clock@ if #target_cpu else software@task\\-clock\\,name\\=task\\-clock@)\000\000Process migrations to a new CPU per CPU second\000\0001migrations/sec\000\000\000\00001 */
2616
+
{ 128434 }, /* page_faults_per_second\000Default\000software@page\\-faults\\,name\\=page\\-faults@ * 1e9 / (software@cpu\\-clock\\,name\\=cpu\\-clock@ if #target_cpu else software@task\\-clock\\,name\\=task\\-clock@)\000\000Page faults per CPU second\000\0001faults/sec\000\000\000\00001 */
2617
+
{ 128776 }, /* stalled_cycles_per_instruction\000Default\000max(stalled\\-cycles\\-frontend, stalled\\-cycles\\-backend) / instructions\000\000Max front or backend stalls per instruction\000\000\000\000\000\00000 */
2618
+
2619
+
};
2620
+
2621
+
static const struct pmu_table_entry pmu_metrics__common[] = {
2622
+
{
2623
+
.entries = pmu_metrics__common_default_core,
2624
+
.num_entries = ARRAY_SIZE(pmu_metrics__common_default_core),
2625
+
.pmu_name = { 0 /* default_core\000 */ },
2626
+
},
2627
+
};
2628
+
2617
2629
static const struct compact_pmu_event pmu_events__test_soc_cpu_default_core[] = {
2618
2630
{ 126205 }, /* bp_l1_btb_correct\000branch\000L1 BTB Correction\000event=0x8a\000\00000\000\000\000\000\000 */
2619
2631
{ 126267 }, /* bp_l2_btb_correct\000branch\000L2 BTB Correction\000event=0x8b\000\00000\000\000\000\000\000 */
···
2698
2664
};
2699
2665
2700
2666
static const struct compact_pmu_event pmu_metrics__test_soc_cpu_default_core[] = {
2701
-
{ 127758 }, /* CPI\000\0001 / IPC\000\000\000\000\000\000\000\00000 */
2702
-
{ 128439 }, /* DCache_L2_All\000\000DCache_L2_All_Hits + DCache_L2_All_Miss\000\000\000\000\000\000\000\00000 */
2703
-
{ 128211 }, /* DCache_L2_All_Hits\000\000l2_rqsts.demand_data_rd_hit + l2_rqsts.pf_hit + l2_rqsts.rfo_hit\000\000\000\000\000\000\000\00000 */
2704
-
{ 128305 }, /* DCache_L2_All_Miss\000\000max(l2_rqsts.all_demand_data_rd - l2_rqsts.demand_data_rd_hit, 0) + l2_rqsts.pf_miss + l2_rqsts.rfo_miss\000\000\000\000\000\000\000\00000 */
2705
-
{ 128503 }, /* DCache_L2_Hits\000\000d_ratio(DCache_L2_All_Hits, DCache_L2_All)\000\000\000\000\000\000\000\00000 */
2706
-
{ 128571 }, /* DCache_L2_Misses\000\000d_ratio(DCache_L2_All_Miss, DCache_L2_All)\000\000\000\000\000\000\000\00000 */
2707
-
{ 127843 }, /* Frontend_Bound_SMT\000\000idq_uops_not_delivered.core / (4 * (cpu_clk_unhalted.thread / 2 * (1 + cpu_clk_unhalted.one_thread_active / cpu_clk_unhalted.ref_xclk)))\000\000\000\000\000\000\000\00000 */
2708
-
{ 127780 }, /* IPC\000group1\000inst_retired.any / cpu_clk_unhalted.thread\000\000\000\000\000\000\000\00000 */
2709
-
{ 128705 }, /* L1D_Cache_Fill_BW\000\00064 * l1d.replacement / 1e9 / duration_time\000\000\000\000\000\000\000\00000 */
2710
-
{ 128641 }, /* M1\000\000ipc + M2\000\000\000\000\000\000\000\00000 */
2711
-
{ 128663 }, /* M2\000\000ipc + M1\000\000\000\000\000\000\000\00000 */
2712
-
{ 128685 }, /* M3\000\0001 / M3\000\000\000\000\000\000\000\00000 */
2713
-
{ 128140 }, /* cache_miss_cycles\000group1\000dcache_miss_cpi + icache_miss_cycles\000\000\000\000\000\000\000\00000 */
2714
-
{ 128009 }, /* dcache_miss_cpi\000\000l1d\\-loads\\-misses / inst_retired.any\000\000\000\000\000\000\000\00000 */
2715
-
{ 128073 }, /* icache_miss_cycles\000\000l1i\\-loads\\-misses / inst_retired.any\000\000\000\000\000\000\000\00000 */
2667
+
{ 129650 }, /* CPI\000\0001 / IPC\000\000\000\000\000\000\000\00000 */
2668
+
{ 130331 }, /* DCache_L2_All\000\000DCache_L2_All_Hits + DCache_L2_All_Miss\000\000\000\000\000\000\000\00000 */
2669
+
{ 130103 }, /* DCache_L2_All_Hits\000\000l2_rqsts.demand_data_rd_hit + l2_rqsts.pf_hit + l2_rqsts.rfo_hit\000\000\000\000\000\000\000\00000 */
2670
+
{ 130197 }, /* DCache_L2_All_Miss\000\000max(l2_rqsts.all_demand_data_rd - l2_rqsts.demand_data_rd_hit, 0) + l2_rqsts.pf_miss + l2_rqsts.rfo_miss\000\000\000\000\000\000\000\00000 */
2671
+
{ 130395 }, /* DCache_L2_Hits\000\000d_ratio(DCache_L2_All_Hits, DCache_L2_All)\000\000\000\000\000\000\000\00000 */
2672
+
{ 130463 }, /* DCache_L2_Misses\000\000d_ratio(DCache_L2_All_Miss, DCache_L2_All)\000\000\000\000\000\000\000\00000 */
2673
+
{ 129735 }, /* Frontend_Bound_SMT\000\000idq_uops_not_delivered.core / (4 * (cpu_clk_unhalted.thread / 2 * (1 + cpu_clk_unhalted.one_thread_active / cpu_clk_unhalted.ref_xclk)))\000\000\000\000\000\000\000\00000 */
2674
+
{ 129672 }, /* IPC\000group1\000inst_retired.any / cpu_clk_unhalted.thread\000\000\000\000\000\000\000\00000 */
2675
+
{ 130597 }, /* L1D_Cache_Fill_BW\000\00064 * l1d.replacement / 1e9 / duration_time\000\000\000\000\000\000\000\00000 */
2676
+
{ 130533 }, /* M1\000\000ipc + M2\000\000\000\000\000\000\000\00000 */
2677
+
{ 130555 }, /* M2\000\000ipc + M1\000\000\000\000\000\000\000\00000 */
2678
+
{ 130577 }, /* M3\000\0001 / M3\000\000\000\000\000\000\000\00000 */
2679
+
{ 130032 }, /* cache_miss_cycles\000group1\000dcache_miss_cpi + icache_miss_cycles\000\000\000\000\000\000\000\00000 */
2680
+
{ 129901 }, /* dcache_miss_cpi\000\000l1d\\-loads\\-misses / inst_retired.any\000\000\000\000\000\000\000\00000 */
2681
+
{ 129965 }, /* icache_miss_cycles\000\000l1i\\-loads\\-misses / inst_retired.any\000\000\000\000\000\000\000\00000 */
2716
2682
2717
2683
};
2718
2684
···
2793
2759
.pmus = pmu_events__common,
2794
2760
.num_pmus = ARRAY_SIZE(pmu_events__common),
2795
2761
},
2796
-
.metric_table = {},
2762
+
.metric_table = {
2763
+
.pmus = pmu_metrics__common,
2764
+
.num_pmus = ARRAY_SIZE(pmu_metrics__common),
2765
+
},
2797
2766
},
2798
2767
{
2799
2768
.arch = "testarch",
···
3243
3206
const struct pmu_events_map *map = map_for_cpu(cpu);
3244
3207
3245
3208
return map ? &map->metric_table : NULL;
3209
+
}
3210
+
3211
+
const struct pmu_metrics_table *pmu_metrics_table__default(void)
3212
+
{
3213
+
int i = 0;
3214
+
3215
+
for (;;) {
3216
+
const struct pmu_events_map *map = &pmu_events_map[i++];
3217
+
3218
+
if (!map->arch)
3219
+
break;
3220
+
3221
+
if (!strcmp(map->cpuid, "common"))
3222
+
return &map->metric_table;
3223
+
}
3224
+
return NULL;
3246
3225
}
3247
3226
3248
3227
const struct pmu_events_table *find_core_events_table(const char *arch, const char *cpuid)
+20
-1
tools/perf/pmu-events/jevents.py
+20
-1
tools/perf/pmu-events/jevents.py
···
755
755
\t\t.pmus = pmu_events__common,
756
756
\t\t.num_pmus = ARRAY_SIZE(pmu_events__common),
757
757
\t},
758
-
\t.metric_table = {},
758
+
\t.metric_table = {
759
+
\t\t.pmus = pmu_metrics__common,
760
+
\t\t.num_pmus = ARRAY_SIZE(pmu_metrics__common),
761
+
\t},
759
762
},
760
763
""")
761
764
else:
···
1238
1235
const struct pmu_events_map *map = map_for_cpu(cpu);
1239
1236
1240
1237
return map ? &map->metric_table : NULL;
1238
+
}
1239
+
1240
+
const struct pmu_metrics_table *pmu_metrics_table__default(void)
1241
+
{
1242
+
int i = 0;
1243
+
1244
+
for (;;) {
1245
+
const struct pmu_events_map *map = &pmu_events_map[i++];
1246
+
1247
+
if (!map->arch)
1248
+
break;
1249
+
1250
+
if (!strcmp(map->cpuid, "common"))
1251
+
return &map->metric_table;
1252
+
}
1253
+
return NULL;
1241
1254
}
1242
1255
1243
1256
const struct pmu_events_table *find_core_events_table(const char *arch, const char *cpuid)
+1
tools/perf/pmu-events/pmu-events.h
+1
tools/perf/pmu-events/pmu-events.h
···
127
127
const struct pmu_events_table *perf_pmu__find_events_table(struct perf_pmu *pmu);
128
128
const struct pmu_events_table *perf_pmu__default_core_events_table(void);
129
129
const struct pmu_metrics_table *pmu_metrics_table__find(void);
130
+
const struct pmu_metrics_table *pmu_metrics_table__default(void);
130
131
const struct pmu_events_table *find_core_events_table(const char *arch, const char *cpuid);
131
132
const struct pmu_metrics_table *find_core_metrics_table(const char *arch, const char *cpuid);
132
133
int pmu_for_each_core_event(pmu_event_iter_fn fn, void *data);
+21
-10
tools/perf/util/metricgroup.c
+21
-10
tools/perf/util/metricgroup.c
···
424
424
.fn = fn,
425
425
.data = data,
426
426
};
427
+
const struct pmu_metrics_table *tables[2] = {
428
+
table,
429
+
pmu_metrics_table__default(),
430
+
};
427
431
428
-
if (table) {
429
-
int ret = pmu_metrics_table__for_each_metric(table, fn, data);
432
+
for (size_t i = 0; i < ARRAY_SIZE(tables); i++) {
433
+
int ret;
430
434
435
+
if (!tables[i])
436
+
continue;
437
+
438
+
ret = pmu_metrics_table__for_each_metric(tables[i], fn, data);
431
439
if (ret)
432
440
return ret;
433
441
}
···
1589
1581
1590
1582
bool metricgroup__has_metric_or_groups(const char *pmu, const char *metric_or_groups)
1591
1583
{
1592
-
const struct pmu_metrics_table *table = pmu_metrics_table__find();
1584
+
const struct pmu_metrics_table *tables[2] = {
1585
+
pmu_metrics_table__find(),
1586
+
pmu_metrics_table__default(),
1587
+
};
1593
1588
struct metricgroup__has_metric_data data = {
1594
1589
.pmu = pmu,
1595
1590
.metric_or_groups = metric_or_groups,
1596
1591
};
1597
1592
1598
-
if (!table)
1599
-
return false;
1600
-
1601
-
return pmu_metrics_table__for_each_metric(table,
1602
-
metricgroup__has_metric_or_groups_callback,
1603
-
&data)
1604
-
? true : false;
1593
+
for (size_t i = 0; i < ARRAY_SIZE(tables); i++) {
1594
+
if (pmu_metrics_table__for_each_metric(tables[i],
1595
+
metricgroup__has_metric_or_groups_callback,
1596
+
&data))
1597
+
return true;
1598
+
}
1599
+
return false;
1605
1600
}
1606
1601
1607
1602
static int metricgroup__topdown_max_level_callback(const struct pmu_metric *pm,