[OLK-5.10] AMD Genoa and Turin Perf patches backport for OLK-5.10 · Pull Request !12311 · openEuler/kernel

perf support:
sample type PERF_SAMPLE_DATA_PAGE_SIZE support
perf tools: Support PERF_SAMPLE_WEIGHT_STRUCT
perf record ibs: Warn about sampling period skew
HYBRID_TOPOLOGY feature
Record non-CPU PMU capabilities
perf script ibs: Support new IBS bits in raw trace dump
perf tool support for IBS DataSrc and L3 miss filtering
"ref-cycles" event support for Zen 4 and later microarchitectures

Install perf and perftest tool to verify patched kernel.
```javascript
#cd /kernel/tools/perf/
#sudo make perf
#cp perf /usr/bin/
```

*Run Test cases*
-----------------------------
* perf/x86/amd/core: Define a proper ref-cycles event for Zen 4 and later
* perf/x86/amd/core: Update and fix stalled-cycles-* events for Zen 2 and later
* x86: Fix comment for X86_FEATURE_ZEN

```javascript
perf list | grep ref-cycles

ref-cycles                                         [Hardware event]
ref-cycles OR cpu/ref-cycles/                      [Kernel PMU event]
```
```javascript
perf stat -e "{ref-cycles, cpu/ref-cycles/}" sleep 5

Performance counter stats for 'sleep 5':
  371,296      ref-cycles:u
  
  371,296       cpu/ref-cycles/u

5.002200577 seconds time elapsed
  
  0.000000000 seconds user
  
  0.001862000 seconds sys
```
-------------------------------------------------------------------
* perf script ibs: Support new IBS bits in raw trace dump
* perf tool ibs: Sync AMD IBS header file
* perf report: Add support to print a textual representation of IBS raw sample data
* perf evlist: Use the right prefix for 'struct evlist' sample id lookup methods
* perf evlist: Use the right prefix for 'struct evlist' sample parsing methods
* perf evlist: Use the right prefix for 'struct evlist' raw samples methods
* perf report: Add tools/arch/x86/include/asm/amd-ibs.h

```javascript
$sudo perf record -c 130 -a -e ibs_op/l3missonly=1/ --raw-samples
$sudo perf report -D
```
...

ibs_op_ctl:     0000004500070008 MaxCnt       128 L3MissOnly 1 En 1

Val 1 CntCtl 0=cycles CurCnt        69
ibs_op_data:    0000000000710002 CompToRetCtr     2 TagToRetCtr   113

BrnRet 0  RipInvalid 0 BrnFuse 0 Microcode 0
ibs_op_data2:   0000000000000002 CacheHitSt 0=M-state RmtNode 0

DataSrc 2=A peer cache in a near CCX
ibs_op_data3:   000000681d1700a1 LdOp 1 StOp 0 DcL1TlbMiss 0

DcL2TlbMiss 0 DcL1TlbHit2M 0 DcL1TlbHit1G 1 DcL2TlbHit2M 0

DcMiss 1 DcMisAcc 0 DcWcMemAcc 0 DcUcMemAcc 0 DcLockedOp 0

DcMissNoMabAlloc 1 DcLinAddrValid 1 DcPhyAddrValid 1

DcL2TlbHit1G 0 L2Miss 1 SwPf 0 OpMemWidth  8 bytes

OpDcMissOpenMemReqs  7 DcMissLat   104 TlbRefillLat     0
```

*IBS Fetch PMU ex:*
```javascript
$sudo perf record -c 130 -a -e ibs_fetch/l3missonly=1/ --raw-samples
sudo ./perf report -D

...
ibs_fetch_ctl:  3c1f00c700080008 MaxCnt     128 Cnt     128 Lat   199
  
En 1 Val 1 Comp 1 IcMiss 1 PhyAddrValid        1 L1TlbPgSz 4KB
  
L1TlbMiss 0 L2TlbMiss 0 RandEn 0 L2Miss 1 L3MissOnly 1
  
FetchOcMiss 1 FetchL3Miss 1
```
----------------------------------------------------------------------------------
* perf header: Record non-CPU PMU capabilities
* perf inject: Keep some features sections from input file
* libperf: Add preadn()
* perf header: Add ability to keep feature sections

```javascript
  $ perf inject -i perf.data-from-desktop -o injected-perf.data

$ perf script --header-only -i injected-perf.data | head -15
  # ========
  # captured on    : Fri May 20 15:08:54 2022
  # header version : 1
  # data offset    : 1208
  # data size      : 837480
  # feat offset    : 838688
  # hostname : Desktop
  # os release : 5.13.0-41-generic
  # perf version : 5.18.rc5.gac837f7ca7ed
  # arch : x86_64
  # nrcpus online : 28
  # nrcpus avail : 28
  # cpudesc : Intel(R) Core(TM) i9-9940X CPU @ 3.30GHz
  # cpuid : GenuineIntel,6,85,4
  # total memory : 65548656 kB
```
--------------------------------------------------------------------------------------
* perf tools: Support PERF_SAMPLE_WEIGHT_STRUCT
* perf script: Fix PERF_SAMPLE_WEIGHT_STRUCT support
* perf record: Add support for PERF_SAMPLE_CODE_PAGE_SIZE

```javascript
sudo perf record --code-page-size sleep 1

Error:
Asking for the code page size isn't supported by this kernel.
```
-----------------------------------------------------------------------------------------
* perf evsel: Emit warning about kernel not supporting the data page size sample_type bit
* perf record: Support new sample type for data page size

```javascript
sudo perf record --data-page-size sleep 1

Error:
Asking for the data page size isn't supported by this kernel.
```
-----------------------------------------------------------------------------------------
* perf/x86/amd: Do not WARN() on every IRQ
* perf/x86/amd/core: Avoid register reset when CPU is dead

```javascript
git clone https://github.com/deater/perf_event_tests.git

[sudo sysctl kernel.perf_event_paranoid=2]

./fuzzer/perf_fuzzer -t OC -s 100
....
....
Close:  5/5 Successful
Read:   0/0 Successful
Write:  0/0 Successful
Ioctl:  0/0 Successful:

```

--------------------------------------------------------------------------------
* perf header: Store PMU caps in an array of strings
* perf header: Pass "cpu" pmu name while printing caps
* perf header: Support HYBRID_CPU_PMU_CAPS feature
* perf header: Support HYBRID_TOPOLOGY feature
* perf pmu: Add hybrid helper functions
* perf pmu: Save detected hybrid pmus to a global pmu list
* perf env: Conditionally compile BPF support code on having HAVE_LIBBPF_SUPPORT
* perf pmu: Parse pmu caps sysfs only once

```javascript
sudo ./perf record -c 10000 -C 0 -e ibs_fetch/l3missonly=1/ sleep 5
 
WARNING: Hw internally resets sampling period when L3 Miss Filtering is enabled
and tagged operation does not cause L3 Miss. This causes sampling period skew.
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.369 MB perf.data ]
[amd@volcano9f8e-oshost malathi]$ sudo ./perf report --header-only -I

...
...
search HYBRID_TOPO
```

GVPopenEuler/kernel关闭

搜索帮助

GVP openEuler/kernel
关闭