perf vendor events intel: update
Subsystems (2)
cover perf vendor events intel: update
Sync with the latest perfmon events from:
https://github.com/intel/perfmon
by running the script:
https://github.com/intel/perfmon/blob/main/scripts/create_perf_json.py
and copying the resulting json and mapfile.csv changes into the perf tree.
Chun-Tse Shao (6):
perf vendor events intel: Update arrowlake events from 1.17 to 1.19 perf vendor events intel: Update emeraldrapids events from 1.23 to
1.24
perf vendor events intel: Update graniterapids events from 1.18 to
1.19perf vendor events intel: Update lunarlake events from 1.22 to 1.25 perf vendor events intel: Update pantherlake events from 1.05 to 1.06 perf vendor events intel: Update tigerlake events from 1.18 to 1.19
.../pmu-events/arch/x86/arrowlake/cache.json | 30 ++- .../arch/x86/arrowlake/floating-point.json | 45 ++++ .../pmu-events/arch/x86/arrowlake/memory.json | 18 ++ .../arch/x86/arrowlake/pipeline.json | 129 +++++++++- .../arch/x86/emeraldrapids/cache.json | 9 + .../graniterapids/uncore-interconnect.json | 10 + .../arch/x86/graniterapids/uncore-memory.json | 2 +- .../pmu-events/arch/x86/lunarlake/cache.json | 2 +- .../arch/x86/lunarlake/pipeline.json | 27 ++- .../arch/x86/lunarlake/uncore-memory.json | 208 ++++++++++++++++- tools/perf/pmu-events/arch/x86/mapfile.csv | 12 +- .../arch/x86/pantherlake/counter.json | 5 + .../arch/x86/pantherlake/pipeline.json | 29 ++- .../x86/pantherlake/uncore-interconnect.json | 10 + .../arch/x86/pantherlake/uncore-memory.json | 221 +++++++++++++++++- 15 files changed, 728 insertions(+), 29 deletions(-) create mode 100644 tools/perf/pmu-events/arch/x86/pantherlake/uncore-interconnect.json
Signature · 1 line
2.54.0.1099.g489fc7bff1-goog
PATCH 1/6 perf vendor events intel: Update arrowlake events from 1.17 to 1.19
The updated events were published in:
https://github.com/intel/perfmon/commit/b84e75626ae78558b8f526a276e4597c5ca6c429
--- .../pmu-events/arch/x86/arrowlake/cache.json | 30 +++- .../arch/x86/arrowlake/floating-point.json | 45 ++++++ .../pmu-events/arch/x86/arrowlake/memory.json | 18 +++ .../arch/x86/arrowlake/pipeline.json | 129 +++++++++++++++++- tools/perf/pmu-events/arch/x86/mapfile.csv | 2 +- 5 files changed, 217 insertions(+), 7 deletions(-)
tools/perf/pmu-events/arch/x86/arrowlake/cache.json +29 -1
@@ -1,6 +1,6 @@ [ { - "BriefDescription": "Counts the number of request that were not accepted into the L2Q because the L2Q is FULL.", + "BriefDescription": "Counts the number of requests that were not accepted into the L2Q because the L2Q is FULL.", "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x31", "EventName": "CORE_REJECT_L2Q.ANY",
@@ -8,6 +8,15 @@ "SampleAfterValue": "1000003", "Unit": "cpu_atom" }, + { + "BriefDescription": "Counts the number of requests that were not accepted into the L2Q because the L2Q is FULL.", + "Counter": "0,1,2,3,4,5,6,7", + "EventCode": "0x31", + "EventName": "CORE_REJECT_L2Q.ANY", + "PublicDescription": "Counts the number of (demand and L1 prefetchers) core requests rejected by the L2Q due to a full or nearly full w condition which likely indicates back pressure from L2Q. It also counts requests that would have gone directly to the XQ, but are rejected due to a full or nearly full condition, indicating back pressure from the IDI link. The L2Q may also reject transactions from a core to insure fairness between cores, or to delay a cores dirty eviction when the address conflicts incoming external snoops. (Note that L2 prefetcher requests that are dropped are not counted by this event.) Counts on a per core basis.", + "SampleAfterValue": "200003", + "Unit": "cpu_lowpower" + }, { "BriefDescription": "Counts the number of L1D cacheline (dirty) evictions caused by load misses, stores, and prefetches.", "Counter": "0,1,2,3,4,5,6,7",
@@ -310,6 +319,15 @@ "SampleAfterValue": "1000003", "Unit": "cpu_atom" }, + { + "BriefDescription": "Counts the number of demand and prefetch transactions that the External Queue (XQ) rejects due to a full or near full condition.", + "Counter": "0,1,2,3,4,5,6,7", + "EventCode": "0x30", + "EventName": "L2_REJECT_XQ.ANY", + "PublicDescription": "Counts the number of demand and prefetch transactions that the External Queue (XQ) rejects due to a full or near full condition which likely indicates back pressure from the IDI link. The XQ may reject transactions from the L2Q (non-cacheable requests), BBL (L2 misses) and WOB (L2 write-back victims).", + "SampleAfterValue": "200003", + "Unit": "cpu_lowpower" + }, { "BriefDescription": "Counts the number of L2 Cache Accesses Counts the total number of L2 Cache Accesses - sum of hits, misses, rejects front door requests for CRd/DRd/RFO/ItoM/L2 Prefetches only, per core event", "Counter": "0,1,2,3,4,5,6,7",
@@ -1382,6 +1400,16 @@ "UMask": "0x83", "Unit": "cpu_atom" }, + { + "BriefDescription": "Counts the number of memory uops retired. A single uop that performs both a load AND a store will be counted as 1, not 2 (e.g. ADD [mem], CONST)", + "Counter": "0,1,2,3,4,5,6,7", + "Data_LA": "1", + "EventCode": "0xd0", + "EventName": "MEM_UOPS_RETIRED.ALL", + "SampleAfterValue": "200003", + "UMask": "0x83", + "Unit": "cpu_lowpower" + }, { "BriefDescription": "Counts the number of load uops retired.", "Counter": "0,1,2,3,4,5,6,7",
tools/perf/pmu-events/arch/x86/arrowlake/floating-point.json +45 -0
75 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/arrowlake/memory.json +18 -0
@@ -173,6 +173,15 @@ "UMask": "0x2", "Unit": "cpu_atom" }, + { + "BriefDescription": "Counts the number of cycles that the head (oldest load) of the load buffer is stalled due to request buffers full or lock in progress.", + "Counter": "0,1,2,3,4,5,6,7", + "EventCode": "0x05", + "EventName": "LD_HEAD.WCB_FULL", + "SampleAfterValue": "1000003", + "UMask": "0x2", + "Unit": "cpu_lowpower" + }, { "BriefDescription": "Counts the number of cycles that the head (oldest load) of the load buffer and retirement are both stalled due to request buffers full or lock in progress.", "Counter": "0,1,2,3,4,5,6,7",
@@ -182,6 +191,15 @@ "UMask": "0x82", "Unit": "cpu_atom" }, + { + "BriefDescription": "Counts the number of cycles that the head (oldest load) of the load buffer and retirement are both stalled due to request buffers full or lock in progress.", + "Counter": "0,1,2,3,4,5,6,7", + "EventCode": "0x05", + "EventName": "LD_HEAD.WCB_FULL_AT_RET", + "SampleAfterValue": "1000003", + "UMask": "0x82", + "Unit": "cpu_lowpower" + }, { "BriefDescription": "Counts the number of memory ordering machine clears triggered due to a snoop from an external agent. Does not count internally generated machine clears such as those due to disambiguations.", "Counter": "0,1,2,3,4,5,6,7",
tools/perf/pmu-events/arch/x86/arrowlake/pipeline.json +124 -5
234 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/mapfile.csv +1 -1
@@ -1,7 +1,7 @@ Family-model,Version,Filename,EventType GenuineIntel-6-(97|9A|B7|BA|BF),v1.39,alderlake,core GenuineIntel-6-BE,v1.39,alderlaken,core -GenuineIntel-6-C[56],v1.17,arrowlake,core +GenuineIntel-6-C[56],v1.19,arrowlake,core GenuineIntel-6-(1C|26|27|35|36),v5,bonnell,core GenuineIntel-6-(3D|47),v30,broadwell,core GenuineIntel-6-56,v12,broadwellde,core
Signature · 1 line
2.54.0.1099.g489fc7bff1-goog
[email protected]
AI reviewRe: [PATCH v1 1/6] perf vendor events intel: Update arrowlake events from 1.17 to 1.19
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider: - [Medium] The `UMask` field was erroneously deleted from the `BR_INST_RETIRED.COND_TAKEN_FWD` event…
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] The UMask field was erroneously deleted from the BR_INST_RETIRED.COND_TAKEN_FWD event definition, causing it to silently alias to BR_INST_RETIRED.ALL_BRANCHES. --
commit a0347536085c6d8e17e90fb1536ff53c3e33e1ee
Author: Chun-Tse Shao <[email protected]>
perf vendor events intel: Update arrowlake events from 1.17 to 1.19
This commit updates the Intel Arrowlake performance monitoring events from version 1.17 to 1.19. The updated events were published in the upstream perfmon repository.
Quoted 4 lines
tools/perf/pmu-events/arch/x86/arrowlake/pipeline.json +0
-0
[ ... ]
tools/perf/pmu-events/arch/x86/arrowlake/pipeline.json line 212
Quoted 5 lines
tools/perf/pmu-events/arch/x86/arrowlake/pipeline.json +0
-1@@ -209,7 +209,6 @@ "EventName": "BR_INST_RETIRED.COND_TAKEN_FWD", "PublicDescription": "Counts taken forward conditional branch instructions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", - "UMask": "0x102",
[Severity: Medium]
Was it intentional to completely remove the UMask field from BR_INST_RETIRED.COND_TAKEN_FWD instead of updating it?
When a PMU event has no UMask specified, the perf event parser defaults the umask to 0x00. Since BR_INST_RETIRED.ALL_BRANCHES uses EventCode 0xc4 without a UMask, the BR_INST_RETIRED.COND_TAKEN_FWD event (also 0xc4) now perfectly aliases to ALL_BRANCHES.
This breaks top-down microarchitecture analysis metrics that rely on it, such as tma_info_bad_spec_ipmisp_cond_taken_fwd, as the metric expression cpu_core@BR_INST_RETIRED.COND_TAKEN_FWD@ / cpu_core@BR_INST_RETIRED.ALL_BRANCHES@ will erroneously evaluate to exactly 1.0.
Quoted 3 lines
tools/perf/pmu-events/arch/x86/arrowlake/pipeline.json +0
-0"Unit": "cpu_core" }, {
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1
PATCH 2/6 perf vendor events intel: Update emeraldrapids events from 1.23 to 1.24
In reply to Chun-Tse ShaoThe updated events were published in:
https://github.com/intel/perfmon/commit/3f1d40d1953193e75c6b5a559638cf1f67bacaed
--- tools/perf/pmu-events/arch/x86/emeraldrapids/cache.json | 9 +++++++++ tools/perf/pmu-events/arch/x86/mapfile.csv | 2 +- 2 files changed, 10 insertions(+), 1 deletion(-)
tools/perf/pmu-events/arch/x86/emeraldrapids/cache.json +9 -0
@@ -368,6 +368,15 @@ "SampleAfterValue": "200003", "UMask": "0x40" }, + { + "BriefDescription": "Cycles when L1D is locked", + "Counter": "0,1,2,3", + "EventCode": "0x42", + "EventName": "LOCK_CYCLES.CACHE_LOCK_DURATION", + "PublicDescription": "This event counts the number of cycles when the L1D is locked. It is a superset of the 0x1 mask (BUS_LOCK_CLOCKS.BUS_LOCK_DURATION).", + "SampleAfterValue": "2000003", + "UMask": "0x2" + }, { "BriefDescription": "Core-originated cacheable requests that missed L3 (Except hardware prefetches to the L3)", "Counter": "0,1,2,3,4,5,6,7",
tools/perf/pmu-events/arch/x86/mapfile.csv +1 -1
@@ -9,7 +9,7 @@ GenuineIntel-6-4F,v23,broadwellx,core GenuineIntel-6-55-[56789ABCDEF],v1.25,cascadelakex,core GenuineIntel-6-DD,v1.02,clearwaterforest,core GenuineIntel-6-9[6C],v1.05,elkhartlake,core -GenuineIntel-6-CF,v1.23,emeraldrapids,core +GenuineIntel-6-CF,v1.24,emeraldrapids,core GenuineIntel-6-5[CF],v13,goldmont,core GenuineIntel-6-7A,v1.01,goldmontplus,core GenuineIntel-6-B6,v1.12,grandridge,core
Signature · 1 line
2.54.0.1099.g489fc7bff1-goog
PATCH 3/6 perf vendor events intel: Update graniterapids events from 1.18 to 1.19
In reply to Chun-Tse ShaoThe updated events were published in:
https://github.com/intel/perfmon/commit/875354c88686ef50387d9601f52354a6da8f24cc
--- .../arch/x86/graniterapids/uncore-interconnect.json | 10 ++++++++++ .../arch/x86/graniterapids/uncore-memory.json | 2 +- tools/perf/pmu-events/arch/x86/mapfile.csv | 2 +- 3 files changed, 12 insertions(+), 2 deletions(-)
tools/perf/pmu-events/arch/x86/graniterapids/uncore-interconnect.json +10 -0
@@ -808,6 +808,16 @@ "PerPkg": "1", "Unit": "IRP" }, + { + "BriefDescription": "Counts Timeouts - Set 0 : Cache Inserts of Write Transactions as Secondary", + "Counter": "0,1,2,3", + "EventCode": "0x1E", + "EventName": "UNC_I_MISC0.2ND_WR_INSERT", + "Experimental": "1", + "PerPkg": "1", + "UMask": "0x8", + "Unit": "IRP" + }, { "BriefDescription": "Counts Timeouts - Set 0 : Fastpath Rejects", "Counter": "0,1,2,3",
tools/perf/pmu-events/arch/x86/graniterapids/uncore-memory.json +1 -1
@@ -539,7 +539,7 @@ "Unit": "IMC" }, { - "BriefDescription": "DRAM Precharge commands. : Precharge due to (?) : Counts the number of DRAM Precharge commands sent on this channel.", + "BriefDescription": "DRAM Precharge commands. : Precharge due to page table : Counts the number of DRAM Precharge commands sent on this channel.", "Counter": "0,1,2,3", "EventCode": "0x03", "EventName": "UNC_M_PRE_COUNT.PGT",
tools/perf/pmu-events/arch/x86/mapfile.csv +1 -1
@@ -13,7 +13,7 @@ GenuineIntel-6-CF,v1.24,emeraldrapids,core GenuineIntel-6-5[CF],v13,goldmont,core GenuineIntel-6-7A,v1.01,goldmontplus,core GenuineIntel-6-B6,v1.12,grandridge,core -GenuineIntel-6-A[DE],v1.18,graniterapids,core +GenuineIntel-6-A[DE],v1.19,graniterapids,core GenuineIntel-6-(3C|45|46),v36,haswell,core GenuineIntel-6-3F,v29,haswellx,core GenuineIntel-6-7[DE],v1.24,icelake,core
Signature · 1 line
2.54.0.1099.g489fc7bff1-goog
PATCH 4/6 perf vendor events intel: Update lunarlake events from 1.22 to 1.25
In reply to Chun-Tse ShaoThe updated events were published in:
https://github.com/intel/perfmon/commit/5535a3e8cc14ae8ef58013cf3d8e9480018b911a
--- .../pmu-events/arch/x86/lunarlake/cache.json | 2 +- .../arch/x86/lunarlake/pipeline.json | 27 ++- .../arch/x86/lunarlake/uncore-memory.json | 208 +++++++++++++++++- tools/perf/pmu-events/arch/x86/mapfile.csv | 2 +- 4 files changed, 228 insertions(+), 11 deletions(-)
tools/perf/pmu-events/arch/x86/lunarlake/cache.json +1 -1
@@ -1,6 +1,6 @@ [ { - "BriefDescription": "Counts the number of request that were not accepted into the L2Q because the L2Q is FULL.", + "BriefDescription": "Counts the number of requests that were not accepted into the L2Q because the L2Q is FULL.", "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x31", "EventName": "CORE_REJECT_L2Q.ANY",
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json +22 -5
@@ -190,7 +190,6 @@ "EventName": "BR_INST_RETIRED.COND_TAKEN_FWD", "PublicDescription": "Counts taken forward conditional branch instructions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", - "UMask": "0x102", "Unit": "cpu_core" }, {
@@ -324,6 +323,15 @@ "UMask": "0xdf", "Unit": "cpu_atom" }, + { + "BriefDescription": "Counts the number of taken branch instructions retired", + "Counter": "0,1,2,3,4,5,6,7", + "EventCode": "0xc4", + "EventName": "BR_INST_RETIRED.TAKEN", + "SampleAfterValue": "200003", + "UMask": "0x80", + "Unit": "cpu_atom" + }, { "BriefDescription": "Counts the total number of mispredicted branch instructions retired for all branch types.", "Counter": "0,1,2,3,4,5,6,7",
@@ -446,7 +454,7 @@ "EventName": "BR_MISP_RETIRED.COND_TAKEN_BWD_COST", "PublicDescription": "number of branch instructions retired that were mispredicted and taken backward. This precise event may be used to get the misprediction cost via the Retire_Latency field of PEBS. It fires on the instruction that immediately follows the mispredicted branch. Available PDIST counters: 0,1", "SampleAfterValue": "400009", - "UMask": "0x8001", + "UMask": "0x41", "Unit": "cpu_core" }, {
@@ -475,7 +483,7 @@ "EventName": "BR_MISP_RETIRED.COND_TAKEN_FWD_COST", "PublicDescription": "number of branch instructions retired that were mispredicted and taken forward. This precise event may be used to get the misprediction cost via the Retire_Latency field of PEBS. It fires on the instruction that immediately follows the mispredicted branch. Available PDIST counters: 0,1", "SampleAfterValue": "400009", - "UMask": "0x8002", + "UMask": "0x140", "Unit": "cpu_core" }, {
@@ -575,11 +583,11 @@ "Unit": "cpu_core" }, { - "BriefDescription": "This event counts the number of mispredicted ret instructions retired. Non PEBS", + "BriefDescription": "This event counts the number of mispredicted ret instructions retired.", "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.RET", - "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts mispredicted return instructions retired. Available PDIST counters: 0,1", + "PublicDescription": "This event counts the number of mispredicted ret instructions retired. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x8", "Unit": "cpu_core"
@@ -1373,6 +1381,15 @@ "UMask": "0x88", "Unit": "cpu_atom" }, + { + "BriefDescription": "Counts number of virtual trap actually taken (e.g. highest priority event during retirement). It can count virtual trap from FPC port 0 or port 1 (x87/SSE) equally in a single counter.", + "Counter": "0,1,2,3,4,5,6,7", + "EventCode": "0xc3", + "EventName": "MACHINE_CLEARS.FPC_VIRTUAL_TRAP", + "SampleAfterValue": "20003", + "UMask": "0x40", + "Unit": "cpu_atom" + }, { "BriefDescription": "Counts the number of nukes due to memory renaming", "Counter": "0,1,2,3,4,5,6,7",
tools/perf/pmu-events/arch/x86/lunarlake/uncore-memory.json +204 -4
242 lines not on this page · raw message
tools/perf/pmu-events/arch/x86/mapfile.csv +1 -1
@@ -22,7 +22,7 @@ GenuineIntel-6-3A,v24,ivybridge,core GenuineIntel-6-3E,v24,ivytown,core GenuineIntel-6-2D,v24,jaketown,core GenuineIntel-6-(57|85),v16,knightslanding,core -GenuineIntel-6-BD,v1.22,lunarlake,core +GenuineIntel-6-BD,v1.25,lunarlake,core GenuineIntel-6-(AA|AC|B5),v1.21,meteorlake,core GenuineIntel-6-1[AEF],v4,nehalemep,core GenuineIntel-6-2E,v4,nehalemex,core
Signature · 1 line
2.54.0.1099.g489fc7bff1-goog
[email protected]
AI reviewRe: [PATCH v1 4/6] perf vendor events intel: Update lunarlake events from 1.22 to 1.25
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider: - [Medium] Accidental deletion of the UMask field for the BR_INST_RETIRED.COND_TAKEN_FWD PMU event, making…
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Accidental deletion of the UMask field for the BR_INST_RETIRED.COND_TAKEN_FWD PMU event, making it an alias of ALL_BRANCHES. --
commit 5c9c0d0a9a04ac6fce773f8ccf00f8df75a3f68d
Author: Chun-Tse Shao <[email protected]>
perf vendor events intel: Update lunarlake events from 1.22 to 1.25
This commit updates the Lunar Lake PMU events from version 1.22 to 1.25 based on the upstream intel/perfmon repository.
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json line 193
Quoted 9 lines
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json +0
-1@@ -190,7 +190,6 @@ "EventName": "BR_INST_RETIRED.COND_TAKEN_FWD", "PublicDescription": "Counts taken forward conditional branch instructions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", - "UMask": "0x102",
[Severity: Medium]
Was the UMask field intentionally removed here?
When UMask is missing, it is omitted from the generated event string. The kernel PMU driver interprets omitted mask bits as zero, which makes EventCode 0xc4 with UMask 0x00 an exact match for the ALL_BRANCHES event.
Does this cause BR_INST_RETIRED.COND_TAKEN_FWD to incorrectly count all branches instead of only forward-taken conditional branches, potentially corrupting Topdown metrics that rely on this branching ratio?
(This same UMask removal also appears to happen in the parallel Arrow Lake update patch within this series).
Quoted 2 lines
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json +0
-0"Unit": "cpu_core" },
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=4
Re: [PATCH v1 4/6] perf vendor events intel: Update lunarlake events from 1.22 to 1.25
+ Dapeng and Edward
Quoted 30 lines · On Tue, Jun 9, 2026 at 3:02 PM <[email protected]> wrote:
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Accidental deletion of the UMask field for the BR_INST_RETIRED.COND_TAKEN_FWD PMU event, making it an alias ofALL_BRANCHES. --commit
5c9c0d0a9a04ac6fce773f8ccf00f8df75a3f68d
Author: Chun-Tse Shao <[email protected]>perf vendor events intel: Update lunarlake events from 1.22 to 1.25
This commit updates the Lunar Lake PMU events from version 1.22 to 1.25 based on the upstream intel/perfmon repository.
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json line 193
Quoted 9 lines
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json +0
-1@@ -190,7 +190,6 @@ "EventName": "BR_INST_RETIRED.COND_TAKEN_FWD", "PublicDescription": "Counts taken forward conditional branch instructions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", - "UMask": "0x102",[Severity: Medium]
Was the UMask field intentionally removed here?When UMask is missing, it is omitted from the generated event string. The kernel PMU driver interprets omitted mask bits as zero, which makes EventCode
0xc4with UMask0x00an exact match for theALL_BRANCHESevent.
Quoted 6 lines
Does this cause BR_INST_RETIRED.COND_TAKEN_FWD to incorrectly count all branches instead of only forward-taken conditional branches, potentially corrupting Topdown metrics that rely on this branching ratio?
(This same UMask removal also appears to happen in the parallel Arrow Lake update patch within this series).
From perfmon: https://github.com/intel/perfmon/commit/5535a3e8cc14ae8ef58013cf3d8e9480018b911a
It seems like the umask of BR_INST_RETIRED.COND_TAKEN_FWD is changed
from 0x02 to 0x00.
Can you confirm if that is correct?
Thanks,
CT
Quoted 6 lines
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json line 194
Quoted 3 lines
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json +0
-0"Unit": "cpu_core" },--
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=4
Re: [PATCH v1 4/6] perf vendor events intel: Update lunarlake events from 1.22 to 1.25
Quoted 35 lines · Chun-Tse Shao
+ Dapeng and Edward
Quoted 32 lines · On Tue, Jun 9, 2026 at 3:02 PM <[email protected]> wrote:
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Accidental deletion of the UMask field for the BR_INST_RETIRED.COND_TAKEN_FWD PMU event, making it an alias ofALL_BRANCHES. --commit
5c9c0d0a9a04ac6fce773f8ccf00f8df75a3f68d
Author: Chun-Tse Shao <[email protected]>perf vendor events intel: Update lunarlake events from 1.22 to 1.25
This commit updates the Lunar Lake PMU events from version 1.22 to 1.25 based on the upstream intel/perfmon repository.
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json line 193
Quoted 9 lines
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json +0
-1@@ -190,7 +190,6 @@ "EventName": "BR_INST_RETIRED.COND_TAKEN_FWD", "PublicDescription": "Counts taken forward conditional branch instructions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", - "UMask": "0x102",[Severity: Medium]
Was the UMask field intentionally removed here?When UMask is missing, it is omitted from the generated event string. The kernel PMU driver interprets omitted mask bits as zero, which makes EventCode
0xc4with UMask0x00an exact match for theALL_BRANCHESevent.Does this cause BR_INST_RETIRED.COND_TAKEN_FWD to incorrectly count all branches instead of only forward-taken conditional branches, potentially corrupting Topdown metrics that rely on this branching ratio?
Quoted 6 lines
Quoted 2 lines
(This same UMask removal also appears to happen in the parallel Arrow Lake update patch within this series).
From perfmon: https://github.com/intel/perfmon/commit/5535a3e8cc14ae8ef58013cf3d8e9480018b911a
It seems like the umask ofBR_INST_RETIRED.COND_TAKEN_FWDis changed
from0x02to0x00.
Can you confirm if that is correct?
I suppose it's correct, but need double check. Thanks.
Quoted 8 lines
Thanks,
CTQuoted 4 lines
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json line 194
Quoted 2 lines
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json +0
-0"Unit": "cpu_core" },--
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=4
Re: [PATCH v1 4/6] perf vendor events intel: Update lunarlake events from 1.22 to 1.25
Quoted 36 lines · Dapeng
Quoted 35 lines · Chun-Tse Shao
+ Dapeng and Edward
Quoted 32 lines · On Tue, Jun 9, 2026 at 3:02 PM <[email protected]> wrote:
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Accidental deletion of the UMask field for the BR_INST_RETIRED.COND_TAKEN_FWD PMU event, making it an alias ofALL_BRANCHES. --commit
5c9c0d0a9a04ac6fce773f8ccf00f8df75a3f68d
Author: Chun-Tse Shao <[email protected]>perf vendor events intel: Update lunarlake events from 1.22 to 1.25
This commit updates the Lunar Lake PMU events from version 1.22 to 1.25 based on the upstream intel/perfmon repository.
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json line 193
Quoted 9 lines
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json +0
-1@@ -190,7 +190,6 @@ "EventName": "BR_INST_RETIRED.COND_TAKEN_FWD", "PublicDescription": "Counts taken forward conditional branch instructions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", - "UMask": "0x102",[Severity: Medium]
Was the UMask field intentionally removed here?When UMask is missing, it is omitted from the generated event string. The kernel PMU driver interprets omitted mask bits as zero, which makes EventCode
0xc4with UMask0x00an exact match for theALL_BRANCHESevent.Does this cause BR_INST_RETIRED.COND_TAKEN_FWD to incorrectly count all branches instead of only forward-taken conditional branches, potentially corrupting Topdown metrics that rely on this branching ratio?
Quoted 7 lines
Quoted 6 lines
Quoted 2 lines
(This same UMask removal also appears to happen in the parallel Arrow Lake update patch within this series).
From perfmon: https://github.com/intel/perfmon/commit/5535a3e8cc14ae8ef58013cf3d8e9480018b911a
It seems like the umask ofBR_INST_RETIRED.COND_TAKEN_FWDis changed
from0x02to0x00.
Can you confirm if that is correct?I suppose it's correct, but need double check. Thanks.
Ed confirmed the change is correct. The umask should be 0x00 instead of 0x00. Thanks.
Quoted 9 lines
Quoted 9 lines
Thanks,
CTQuoted 4 lines
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json line 194
Quoted 2 lines
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json +0
-0"Unit": "cpu_core" },--
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=4
Re: [PATCH v1 4/6] perf vendor events intel: Update lunarlake events from 1.22 to 1.25
Quoted 39 lines · Dapeng
Quoted 38 lines · Dapeng
Quoted 37 lines · Chun-Tse Shao
+ Dapeng and Edward
Quoted 34 lines · On Tue, Jun 9, 2026 at 3:02 PM <[email protected]> wrote:
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Accidental deletion of the UMask field for the BR_INST_RETIRED.COND_TAKEN_FWD PMU event, making it an alias ofALL_BRANCHES. --commit
5c9c0d0a9a04ac6fce773f8ccf00f8df75a3f68d
Author: Chun-Tse Shao <[email protected]>perf vendor events intel: Update lunarlake events from 1.22 to 1.25
This commit updates the Lunar Lake PMU events from version 1.22 to 1.25 based on the upstream intel/perfmon repository.
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json line 193
Quoted 9 lines
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json +0
-1@@ -190,7 +190,6 @@ "EventName": "BR_INST_RETIRED.COND_TAKEN_FWD", "PublicDescription": "Counts taken forward conditional branch instructions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", - "UMask": "0x102",[Severity: Medium]
Was the UMask field intentionally removed here?When UMask is missing, it is omitted from the generated event string. The kernel PMU driver interprets omitted mask bits as zero, which makes EventCode
0xc4with UMask0x00an exact match for theALL_BRANCHESevent.Does this cause BR_INST_RETIRED.COND_TAKEN_FWD to incorrectly count all branches instead of only forward-taken conditional branches, potentially corrupting Topdown metrics that rely on this branching ratio?
(This same UMask removal also appears to happen in the parallel Arrow Lake update patch within this series).
Quoted 7 lines
Quoted 5 lines
Quoted 4 lines
From perfmon: https://github.com/intel/perfmon/commit/5535a3e8cc14ae8ef58013cf3d8e9480018b911a
It seems like the umask ofBR_INST_RETIRED.COND_TAKEN_FWDis changed
from0x02to0x00.
Can you confirm if that is correct?I suppose it's correct, but need double check. Thanks.
Ed confirmed the change is correct. The umask should be
0x00instead of0x00. Thanks.
Correct the typo. "instead of 0x02". Thanks.
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Thanks,
CTQuoted 4 lines
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json line 194
Quoted 2 lines
tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json +0
-0"Unit": "cpu_core" },--
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=4
PATCH 5/6 perf vendor events intel: Update pantherlake events from 1.05 to 1.06
In reply to Chun-Tse ShaoThe updated events were published in:
https://github.com/intel/perfmon/commit/ffc03fc3b414127c5a36bbb648e500c4afeff134
--- tools/perf/pmu-events/arch/x86/mapfile.csv | 2 +- .../arch/x86/pantherlake/counter.json | 5 + .../arch/x86/pantherlake/pipeline.json | 29 ++- .../x86/pantherlake/uncore-interconnect.json | 10 + .../arch/x86/pantherlake/uncore-memory.json | 221 +++++++++++++++++- 5 files changed, 260 insertions(+), 7 deletions(-) create mode 100644 tools/perf/pmu-events/arch/x86/pantherlake/uncore-interconnect.json
tools/perf/pmu-events/arch/x86/mapfile.csv +1 -1
@@ -26,7 +26,7 @@ GenuineIntel-6-BD,v1.25,lunarlake,core GenuineIntel-6-(AA|AC|B5),v1.21,meteorlake,core GenuineIntel-6-1[AEF],v4,nehalemep,core GenuineIntel-6-2E,v4,nehalemex,core -GenuineIntel-6-(CC|D5),v1.05,pantherlake,core +GenuineIntel-6-(CC|D5),v1.06,pantherlake,core GenuineIntel-6-A7,v1.04,rocketlake,core GenuineIntel-6-2A,v19,sandybridge,core GenuineIntel-6-8F,v1.39,sapphirerapids,core
tools/perf/pmu-events/arch/x86/pantherlake/counter.json +5 -0
@@ -13,5 +13,10 @@ "Unit": "iMC", "CountersNumFixed": "0", "CountersNumGeneric": "5" + }, + { + "Unit": "SANTA", + "CountersNumFixed": 1, + "CountersNumGeneric": "0" } ]
tools/perf/pmu-events/arch/x86/pantherlake/pipeline.json +25 -4
@@ -887,11 +887,32 @@ "Unit": "cpu_core" }, { - "BriefDescription": "This event counts the number of mispredicted ret instructions retired. Non PEBS [This event is alias to BR_MISP_RETIRED.RET]", + "BriefDescription": "This event is deprecated. [This event is alias to BR_MISP_RETIRED.NEAR_RETURN]", + "Counter": "0,1,2,3,4,5,6,7", + "Deprecated": "1", + "EventCode": "0xc5", + "EventName": "BR_MISP_RETIRED.NEAR_RET", + "PublicDescription": "This event is deprecated. [This event is alias to BR_MISP_RETIRED.NEAR_RETURN] Available PDIST counters: 0,1", + "SampleAfterValue": "1000003", + "UMask": "0x8", + "Unit": "cpu_atom" + }, + { + "BriefDescription": "Counts the number of mispredicted near RET branch instructions retired. [This event is alias to BR_MISP_RETIRED.NEAR_RET]", + "Counter": "0,1,2,3,4,5,6,7", + "EventCode": "0xc5", + "EventName": "BR_MISP_RETIRED.NEAR_RETURN", + "PublicDescription": "Counts the number of mispredicted near RET branch instructions retired. [This event is alias to BR_MISP_RETIRED.NEAR_RET] Available PDIST counters: 0,1", + "SampleAfterValue": "1000003", + "UMask": "0x8", + "Unit": "cpu_atom" + }, + { + "BriefDescription": "This event counts the number of mispredicted ret instructions retired [This event is alias to BR_MISP_RETIRED.RET]", "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.NEAR_RETURN", - "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts mispredicted return instructions retired. [This event is alias to BR_MISP_RETIRED.RET] Available PDIST counters: 0,1", + "PublicDescription": "This event counts the number of mispredicted ret instructions retired [This event is alias to BR_MISP_RETIRED.RET] Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x8", "Unit": "cpu_core"
@@ -1726,7 +1747,7 @@ "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number of machines clears due to memory renaming.", + "BriefDescription": "Counts the number of machine clears due to memory renaming.", "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.MRN_NUKE",
@@ -1930,7 +1951,7 @@ "Unit": "cpu_atom" }, { - "BriefDescription": "Counts the number issue slots not consumed due to a color request for an FCW or MXCSR control register when all 4 colors (copies) are already in use.", + "BriefDescription": "Counts the number of issue slots not consumed due to a color request for an FCW or MXCSR control register when all 4 colors (copies) are already in use.", "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x75", "EventName": "SERIALIZATION.COLOR_STALLS",
tools/perf/pmu-events/arch/x86/pantherlake/uncore-interconnect.json new file +10 -0
@@ -0,0 +1,10 @@ +[ + { + "BriefDescription": "This 48-bit fixed counter counts the UCLK cycles.", + "Counter": "FIXED", + "EventCode": "0xff", + "EventName": "UNC_CLOCK.SOCKET", + "PerPkg": "1", + "Unit": "SANTA" + } +]
tools/perf/pmu-events/arch/x86/pantherlake/uncore-memory.json +219 -2
243 lines not on this page · raw message
Signature · 1 line
2.54.0.1099.g489fc7bff1-goog
PATCH 6/6 perf vendor events intel: Update tigerlake events from 1.18 to 1.19
In reply to Chun-Tse ShaoThe updated events were published in:
https://github.com/intel/perfmon/commit/8353ffb63efcad6b6fac1a8c05d76e2d6317ae23
--- tools/perf/pmu-events/arch/x86/mapfile.csv | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
tools/perf/pmu-events/arch/x86/mapfile.csv +1 -1
@@ -35,7 +35,7 @@ GenuineIntel-6-(37|4A|4C|4D|5A),v15,silvermont,core GenuineIntel-6-(4E|5E|8E|9E|A5|A6),v59,skylake,core GenuineIntel-6-55-[01234],v1.37,skylakex,core GenuineIntel-6-86,v1.23,snowridgex,core -GenuineIntel-6-8[CD],v1.18,tigerlake,core +GenuineIntel-6-8[CD],v1.19,tigerlake,core GenuineIntel-6-2C,v5,westmereep-dp,core GenuineIntel-6-25,v4,westmereep-sp,core GenuineIntel-6-2F,v4,westmereex,core
Signature · 1 line
2.54.0.1099.g489fc7bff1-goog
[email protected]
AI reviewRe: [PATCH v1 6/6] perf vendor events intel: Update tigerlake events from 1.18 to 1.19
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider: - [Medium] Incomplete patch: The commit updates the Tigerlake event version in mapfile.csv but fails to…
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Incomplete patch: The commit updates the Tigerlake event version in mapfile.csv but fails to include the actual updated JSON event files. --
commit 827bd8b5ad34fdfc9cb54712c5bfc0e4b556fa4f
Author: Chun-Tse Shao <[email protected]>
perf vendor events intel: Update tigerlake events from 1.18 to 1.19
This commit updates the tigerlake event version string in mapfile.csv from 1.18 to 1.19 to track the latest Intel perfmon events.
Quoted 4 lines
tools/perf/pmu-events/arch/x86/mapfile.csv +0
-0
[ ... ]
tools/perf/pmu-events/arch/x86/mapfile.csv GenuineIntel-6-(37|4A|4C|4D|5A),v15,silvermont,core · line 38
Quoted 6 lines
tools/perf/pmu-events/arch/x86/mapfile.csv +1
-1@@ -35,7 +35,7 @@ GenuineIntel-6-(37|4A|4C|4D|5A),v15,silvermont,core GenuineIntel-6-(4E|5E|8E|9E|A5|A6),v59,skylake,core GenuineIntel-6-55-[01234],v1.37,skylakex,core GenuineIntel-6-86,v1.23,snowridgex,core -GenuineIntel-6-8[CD],v1.18,tigerlake,core +GenuineIntel-6-8[CD],v1.19,tigerlake,core
[Severity: Medium]
Are the corresponding JSON files for the tigerlake events missing from this commit?
The patch updates the version string to v1.19 here in mapfile.csv, but it doesn't include the actual updated JSON event files in the tools/perf/pmu-events/arch/x86/tigerlake/ directory.
Because jevents.py parses the local JSON files to generate the C event tables, won't perf continue to be compiled with the outdated v1.18 events while falsely reporting v1.19?
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=6
LGTM. Thanks.
Quoted 35 lines · Chun-Tse Shao
Sync with the latest perfmon events from:
https://github.com/intel/perfmon
by running the script:
https://github.com/intel/perfmon/blob/main/scripts/create_perf_json.py
and copying the resulting json and mapfile.csv changes into the perf tree.Chun-Tse Shao (6):
perf vendor events intel: Update arrowlake events from 1.17 to 1.19 perf vendor events intel: Update emeraldrapids events from 1.23 to1.24 perf vendor events intel: Update graniterapids events from 1.18 to 1.19perf vendor events intel: Update lunarlake events from 1.22 to 1.25 perf vendor events intel: Update pantherlake events from 1.05 to 1.06 perf vendor events intel: Update tigerlake events from 1.18 to 1.19
.../pmu-events/arch/x86/arrowlake/cache.json | 30 ++- .../arch/x86/arrowlake/floating-point.json | 45 ++++ .../pmu-events/arch/x86/arrowlake/memory.json | 18 ++ .../arch/x86/arrowlake/pipeline.json | 129 +++++++++- .../arch/x86/emeraldrapids/cache.json | 9 + .../graniterapids/uncore-interconnect.json | 10 + .../arch/x86/graniterapids/uncore-memory.json | 2 +- .../pmu-events/arch/x86/lunarlake/cache.json | 2 +- .../arch/x86/lunarlake/pipeline.json | 27 ++- .../arch/x86/lunarlake/uncore-memory.json | 208 ++++++++++++++++- tools/perf/pmu-events/arch/x86/mapfile.csv | 12 +- .../arch/x86/pantherlake/counter.json | 5 + .../arch/x86/pantherlake/pipeline.json | 29 ++- .../x86/pantherlake/uncore-interconnect.json | 10 + .../arch/x86/pantherlake/uncore-memory.json | 221 +++++++++++++++++- 15 files changed, 728 insertions(+), 29 deletions(-) create mode 100644 tools/perf/pmu-events/arch/x86/pantherlake/uncore-interconnect.json
Quoted 4 lines · Dapeng
LGTM. Thanks.
Reviewed-by Dapeng Mi <[email protected]>
The tigerlake update is just the version number. On the perfmon git the change was removing zero fields:
https://github.com/intel/perfmon/commit/8353ffb63efcad6b6fac1a8c05d76e2d6317ae23
but zero fields are dropped from the perf JSON, resulting in no differences.
Thanks,
Ian
Quoted 36 lines · Chun-Tse Shao
Quoted 35 lines
Sync with the latest perfmon events from:
https://github.com/intel/perfmon
by running the script:
https://github.com/intel/perfmon/blob/main/scripts/create_perf_json.py
and copying the resulting json and mapfile.csv changes into the perf tree.Chun-Tse Shao (6):
perf vendor events intel: Update arrowlake events from 1.17 to 1.19 perf vendor events intel: Update emeraldrapids events from 1.23 to1.24 perf vendor events intel: Update graniterapids events from 1.18 to 1.19perf vendor events intel: Update lunarlake events from 1.22 to 1.25 perf vendor events intel: Update pantherlake events from 1.05 to 1.06 perf vendor events intel: Update tigerlake events from 1.18 to 1.19
.../pmu-events/arch/x86/arrowlake/cache.json | 30 ++- .../arch/x86/arrowlake/floating-point.json | 45 ++++ .../pmu-events/arch/x86/arrowlake/memory.json | 18 ++ .../arch/x86/arrowlake/pipeline.json | 129 +++++++++- .../arch/x86/emeraldrapids/cache.json | 9 + .../graniterapids/uncore-interconnect.json | 10 + .../arch/x86/graniterapids/uncore-memory.json | 2 +- .../pmu-events/arch/x86/lunarlake/cache.json | 2 +- .../arch/x86/lunarlake/pipeline.json | 27 ++- .../arch/x86/lunarlake/uncore-memory.json | 208 ++++++++++++++++- tools/perf/pmu-events/arch/x86/mapfile.csv | 12 +- .../arch/x86/pantherlake/counter.json | 5 + .../arch/x86/pantherlake/pipeline.json | 29 ++- .../x86/pantherlake/uncore-interconnect.json | 10 + .../arch/x86/pantherlake/uncore-memory.json | 221 +++++++++++++++++- 15 files changed, 728 insertions(+), 29 deletions(-) create mode 100644 tools/perf/pmu-events/arch/x86/pantherlake/uncore-interconnect.json
Quoted 8 lines · Chun-Tse Shao
Sync with the latest perfmon events from:
https://github.com/intel/perfmon
by running the script:
https://github.com/intel/perfmon/blob/main/scripts/create_perf_json.py
and copying the resulting json and mapfile.csv changes into the perf tree.[...]
Applied to perf-tools-next, thanks!
Best regards,
Namhyung