Re: [PATCH v4 2/2] perf vendor events hygon: Add Hygon Data Fabric PMU events and metrics
From: Qi Liu
Date: Tue Aug 18 2026 - 02:14:03 EST
On 8/18/2026 12:28 PM, Ian Rogers wrote:
On Mon, Aug 17, 2026 at 8:12 PM Qi Liu <liuqi@xxxxxxxx> wrote:Thank you for the suggestion. I agree that 64B is the common case, will add "ScaleUnit": "64Bytes" for the I/O request events and document the possible 32B-request caveat.
Hi Ian，
Thanks for your reply:)
On 8/18/2026 6:12 AM, Ian Rogers wrote:
On Tue, Aug 4, 2026 at 12:00 AM Qi Liu <liuqi@xxxxxxxx> wrote:Thank you for pointing this out. Yes, for these CPU memory-request events, Data
Add perf JSON event definitions and metrics for the Hygon Data Fabric (DF)
uncore PMU.
Signed-off-by: Qi Liu <liuqi@xxxxxxxx>
---
tools/perf/pmu-events/Build | 2 +-
.../arch/x86/hygonm1/data-fabric.json | 87 ++++++++++
.../arch/x86/hygonm4/data-fabric.json | 122 ++++++++++++++
.../arch/x86/hygonm6/data-fabric.json | 154 ++++++++++++++++++
tools/perf/pmu-events/arch/x86/mapfile.csv | 3 +
tools/perf/pmu-events/jevents.py | 2 +
6 files changed, 369 insertions(+), 1 deletion(-)
create mode 100644 tools/perf/pmu-events/arch/x86/hygonm1/data-fabric.json
create mode 100644 tools/perf/pmu-events/arch/x86/hygonm4/data-fabric.json
create mode 100644 tools/perf/pmu-events/arch/x86/hygonm6/data-fabric.json
diff --git a/tools/perf/pmu-events/Build b/tools/perf/pmu-events/Build
index 372773b998e4..79a9b4cad202 100644
--- a/tools/perf/pmu-events/Build
+++ b/tools/perf/pmu-events/Build
@@ -95,7 +95,7 @@ endif
ifeq ($(JEVENTS_ARCH),$(filter $(JEVENTS_ARCH),x86 all))
# Generate Intel Json
-INTELS := $(shell ls -d pmu-events/arch/x86/*|grep -v amdzen|grep -v mapfile.csv)
+INTELS := $(shell ls -d pmu-events/arch/x86/*|grep -v amdzen|grep -v hygon|grep -v mapfile.csv)
INTEL_METRICS = $(foreach x,$(INTELS),$(OUTPUT)$(x)/extra-metrics.json)
INTEL_METRICGROUPS = $(foreach x,$(INTELS),$(OUTPUT)$(x)/extra-metricgroups.json)
GEN_JSON += $(INTEL_METRICS) $(INTEL_METRICGROUPS)
diff --git a/tools/perf/pmu-events/arch/x86/hygonm1/data-fabric.json b/tools/perf/pmu-events/arch/x86/hygonm1/data-fabric.json
new file mode 100644
index 000000000000..99c243bd9ba0
--- /dev/null
+++ b/tools/perf/pmu-events/arch/x86/hygonm1/data-fabric.json
@@ -0,0 +1,87 @@
+[
+ {
+ "EventName": "cpu_read_requests_cs_0",
+ "BriefDescription": "The number of CPU read requests to memory: Data (64B): CS0",
Does "Data (64B)" mean that each count should be scaled by 64 and the
units are bytes? In which case adding:
"ScaleUnit": "64Bytes",
Would cause the count to be multiplied by 64 in the perf stat output
and the unit Bytes to be added.
(64B) means that each count represents a 64-byte data unit. I will add
"ScaleUnit": "64Bytes" so that perf stat reports the scaled value in Bytes.
The I/O request events are different: each request may carry either 32B or 64B,
so I will keep those unscaled and retain the separate minimum/maximum bandwidth
metrics.
Sounds good to me. Fwiw, in other drivers I've scaled by 64 and then
when the bandwidth is below expectations the explanation are 32B
requests.
Thank you for pointing this out. I will use “Coherent Subordinate” in the event descriptions while retaining the existing CS abbreviation and event names.Would it be possible to define terms like CS0?Thank you. I will replace uses such as CS0 with Coherent Slave 0 (CS0) in the
event descriptions.
Thanks for adding the definition! The Linux coding style advocates
against adding new usage of the term slave:
https://www.kernel.org/doc/Documentation/process/coding-style.rst
Maybe subordinate to keep the initials the same.
Thanks,
Qi
Thank you for the suggestion. ConstId stands for Component Select ID. It+ "EventCode": "0x2",
+ "UMask": "0x07",
+ "ConstId": "0x0",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "cpu_read_requests_cs_1",
+ "BriefDescription": "The number of CPU read requests to memory: Data (64B): CS1",
+ "EventCode": "0x2",
+ "UMask": "0x07",
+ "ConstId": "0x1",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "cpu_write_requests_cs_0",
+ "BriefDescription": "The number of CPU write requests to memory: Data (64B): CS0",
+ "EventCode": "0x2",
+ "UMask": "0x08",
+ "ConstId": "0x0",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "cpu_write_requests_cs_1",
+ "BriefDescription": "The number of CPU write requests to memory: Data (64B): CS1",
+ "EventCode": "0x2",
+ "UMask": "0x08",
+ "ConstId": "0x1",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "io_read_requests_cs_0",
+ "BriefDescription": "The number of inbound I/O read requests to memory: Data (32B or 64B): CS0",
+ "EventCode": "0x3",
+ "UMask": "0x0F",
+ "ConstId": "0x0",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "io_read_requests_cs_1",
+ "BriefDescription": "The number of inbound I/O read requests to memory: Data (32B or 64B): CS1",
+ "EventCode": "0x3",
+ "UMask": "0x0F",
+ "ConstId": "0x1",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "io_write_requests_cs_0",
+ "BriefDescription": "The number of inbound I/O write requests to memory: Data (32B or 64B): CS0",
+ "EventCode": "0x3",
+ "UMask": "0x70",
+ "ConstId": "0x0",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "io_write_requests_cs_1",
+ "BriefDescription": "The number of inbound I/O write requests to memory: Data (32B or 64B): CS1",
+ "EventCode": "0x3",
+ "UMask": "0x70",
+ "ConstId": "0x1",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "MetricName": "cpu_memory_bandwidth",
+ "MetricExpr": "(cpu_read_requests_cs_0 + cpu_read_requests_cs_1 + cpu_write_requests_cs_0 + cpu_write_requests_cs_1) * 64 / (duration_time * 1e6)",
+ "BriefDescription": "Total CPU memory access bandwidth",
+ "MetricGroup": "DF_Memory_BW",
+ "ScaleUnit": "1MB/s"
+ },
+ {
+ "MetricName": "io_memory_max_bandwidth",
+ "MetricExpr": "(io_read_requests_cs_0 + io_read_requests_cs_1 + io_write_requests_cs_0 + io_write_requests_cs_1) * 64 / (duration_time * 1e6)",
+ "BriefDescription": "Maximum inbound I/O memory access bandwidth (upper bound, assuming 64B per request. Actual inbound I/O requests may be 32B or 64B)",
+ "MetricGroup": "DF_Memory_BW",
+ "ScaleUnit": "1MB/s"
+ },
+ {
+ "MetricName": "io_memory_min_bandwidth",
+ "MetricExpr": "(io_read_requests_cs_0 + io_read_requests_cs_1 + io_write_requests_cs_0 + io_write_requests_cs_1) * 32 / (duration_time * 1e6)",
+ "BriefDescription": "Minimum inbound I/O memory access bandwidth (lower bound, assuming 32B per request. Actual inbound I/O requests may be 32B or 64B)",
+ "MetricGroup": "DF_Memory_BW",
+ "ScaleUnit": "1MB/s"
+ }
+]
diff --git a/tools/perf/pmu-events/arch/x86/hygonm4/data-fabric.json b/tools/perf/pmu-events/arch/x86/hygonm4/data-fabric.json
new file mode 100644
index 000000000000..f32dcea7061c
--- /dev/null
+++ b/tools/perf/pmu-events/arch/x86/hygonm4/data-fabric.json
@@ -0,0 +1,122 @@
+[
+ {
+ "EventName": "cpu_read_requests_cs_0",
+ "BriefDescription": "The number of CPU read requests to memory: Data (64B): CS0",
+ "EventCode": "0x2",
+ "UMask": "0x07",
+ "ConstId": "0x0",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "cpu_read_requests_cs_1",
+ "BriefDescription": "The number of CPU read requests to memory: Data (64B): CS1",
+ "EventCode": "0x2",
+ "UMask": "0x07",
+ "ConstId": "0x1",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "cpu_read_requests_cs_2",
+ "BriefDescription": "The number of CPU read requests to memory: Data (64B): CS2",
+ "EventCode": "0x2",
+ "UMask": "0x07",
+ "ConstId": "0x2",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "cpu_write_requests_cs_0",
+ "BriefDescription": "The number of CPU write requests to memory: Data (64B): CS0",
+ "EventCode": "0x2",
+ "UMask": "0x08",
+ "ConstId": "0x0",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "cpu_write_requests_cs_1",
+ "BriefDescription": "The number of CPU write requests to memory: Data (64B): CS1",
+ "EventCode": "0x2",
+ "UMask": "0x08",
+ "ConstId": "0x1",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "cpu_write_requests_cs_2",
+ "BriefDescription": "The number of CPU write requests to memory: Data (64B): CS2",
+ "EventCode": "0x2",
+ "UMask": "0x08",
+ "ConstId": "0x2",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "io_read_requests_cs_0",
+ "BriefDescription": "The number of inbound I/O read requests to memory: Data (32B or 64B): CS0",
+ "EventCode": "0x3",
+ "UMask": "0x0F",
+ "ConstId": "0x0",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "io_read_requests_cs_1",
+ "BriefDescription": "The number of inbound I/O read requests to memory: Data (32B or 64B): CS1",
+ "EventCode": "0x3",
+ "UMask": "0x0F",
+ "ConstId": "0x1",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "io_read_requests_cs_2",
+ "BriefDescription": "The number of inbound I/O read requests to memory: Data (32B or 64B): CS2",
+ "EventCode": "0x3",
+ "UMask": "0x0F",
+ "ConstId": "0x2",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "io_write_requests_cs_0",
+ "BriefDescription": "The number of inbound I/O write requests to memory: Data (32B or 64B): CS0",
+ "EventCode": "0x3",
+ "UMask": "0x70",
+ "ConstId": "0x0",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "io_write_requests_cs_1",
+ "BriefDescription": "The number of inbound I/O write requests to memory: Data (32B or 64B): CS1",
+ "EventCode": "0x3",
+ "UMask": "0x70",
+ "ConstId": "0x1",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "io_write_requests_cs_2",
+ "BriefDescription": "The number of inbound I/O write requests to memory: Data (32B or 64B): CS2",
+ "EventCode": "0x3",
+ "UMask": "0x70",
+ "ConstId": "0x2",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "MetricName": "cpu_memory_bandwidth",
+ "MetricExpr": "(cpu_read_requests_cs_0 + cpu_read_requests_cs_1 + cpu_read_requests_cs_2 + cpu_write_requests_cs_0 + cpu_write_requests_cs_1 + cpu_write_requests_cs_2) * 64 / (duration_time * 1e6)",
+ "MetricConstraint": "NO_GROUP_EVENTS",
+ "BriefDescription": "Total CPU memory access bandwidth",
+ "MetricGroup": "DF_Memory_BW",
+ "ScaleUnit": "1MB/s"
+ },
+ {
+ "MetricName": "io_memory_max_bandwidth",
+ "MetricExpr": "(io_read_requests_cs_0 + io_read_requests_cs_1 + io_read_requests_cs_2 + io_write_requests_cs_0 + io_write_requests_cs_1 + io_write_requests_cs_2) * 64 / (duration_time * 1e6)",
+ "MetricConstraint": "NO_GROUP_EVENTS",
+ "BriefDescription": "Maximum inbound I/O memory access bandwidth (upper bound, assuming 64B per request. Actual inbound I/O requests may be 32B or 64B)",
+ "MetricGroup": "DF_Memory_BW",
+ "ScaleUnit": "1MB/s"
+ },
+ {
+ "MetricName": "io_memory_min_bandwidth",
+ "MetricExpr": "(io_read_requests_cs_0 + io_read_requests_cs_1 + io_read_requests_cs_2 + io_write_requests_cs_0 + io_write_requests_cs_1 + io_write_requests_cs_2) * 32 / (duration_time * 1e6)",
+ "MetricConstraint": "NO_GROUP_EVENTS",
+ "BriefDescription": "Minimum inbound I/O memory access bandwidth (lower bound, assuming 32B per request. Actual inbound I/O requests may be 32B or 64B)",
+ "MetricGroup": "DF_Memory_BW",
+ "ScaleUnit": "1MB/s"
+ }
+]
diff --git a/tools/perf/pmu-events/arch/x86/hygonm6/data-fabric.json b/tools/perf/pmu-events/arch/x86/hygonm6/data-fabric.json
new file mode 100644
index 000000000000..bff5c0959d82
--- /dev/null
+++ b/tools/perf/pmu-events/arch/x86/hygonm6/data-fabric.json
@@ -0,0 +1,154 @@
+[
+ {
+ "EventName": "cpu_read_requests_cs_0",
+ "BriefDescription": "The number of CPU read requests to memory: Data (64B): CS0",
+ "EventCode": "0x2",
+ "UMask": "0x07",
+ "ConstId": "0x0",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "cpu_read_requests_cs_1",
+ "BriefDescription": "The number of CPU read requests to memory: Data (64B): CS1",
+ "EventCode": "0x2",
+ "UMask": "0x07",
+ "ConstId": "0x1",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "cpu_read_requests_cs_2",
+ "BriefDescription": "The number of CPU read requests to memory: Data (64B): CS2",
+ "EventCode": "0x2",
+ "UMask": "0x07",
+ "ConstId": "0x2",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "cpu_read_requests_cs_3",
+ "BriefDescription": "The number of CPU read requests to memory: Data (64B): CS3",
+ "EventCode": "0x2",
+ "UMask": "0x07",
+ "ConstId": "0x3",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "cpu_write_requests_cs_0",
+ "BriefDescription": "The number of CPU write requests to memory: Data (64B): CS0",
+ "EventCode": "0x2",
+ "UMask": "0x08",
+ "ConstId": "0x0",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "cpu_write_requests_cs_1",
+ "BriefDescription": "The number of CPU write requests to memory: Data (64B): CS1",
+ "EventCode": "0x2",
+ "UMask": "0x08",
+ "ConstId": "0x1",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "cpu_write_requests_cs_2",
+ "BriefDescription": "The number of CPU write requests to memory: Data (64B): CS2",
+ "EventCode": "0x2",
+ "UMask": "0x08",
+ "ConstId": "0x2",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "cpu_write_requests_cs_3",
+ "BriefDescription": "The number of CPU write requests to memory: Data (64B): CS3",
+ "EventCode": "0x2",
+ "UMask": "0x08",
+ "ConstId": "0x3",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "io_read_requests_cs_0",
+ "BriefDescription": "The number of inbound I/O read requests to memory: Data (32B or 64B): CS0",
+ "EventCode": "0x3",
+ "UMask": "0x0F",
+ "ConstId": "0x0",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "io_read_requests_cs_1",
+ "BriefDescription": "The number of inbound I/O read requests to memory: Data (32B or 64B): CS1",
+ "EventCode": "0x3",
+ "UMask": "0x0F",
+ "ConstId": "0x1",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "io_read_requests_cs_2",
+ "BriefDescription": "The number of inbound I/O read requests to memory: Data (32B or 64B): CS2",
+ "EventCode": "0x3",
+ "UMask": "0x0F",
+ "ConstId": "0x2",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "io_read_requests_cs_3",
+ "BriefDescription": "The number of inbound I/O read requests to memory: Data (32B or 64B): CS3",
+ "EventCode": "0x3",
+ "UMask": "0x0F",
+ "ConstId": "0x3",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "io_write_requests_cs_0",
+ "BriefDescription": "The number of inbound I/O write requests to memory: Data (32B or 64B): CS0",
+ "EventCode": "0x3",
+ "UMask": "0x70",
+ "ConstId": "0x0",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "io_write_requests_cs_1",
+ "BriefDescription": "The number of inbound I/O write requests to memory: Data (32B or 64B): CS1",
+ "EventCode": "0x3",
+ "UMask": "0x70",
+ "ConstId": "0x1",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "io_write_requests_cs_2",
+ "BriefDescription": "The number of inbound I/O write requests to memory: Data (32B or 64B): CS2",
+ "EventCode": "0x3",
+ "UMask": "0x70",
+ "ConstId": "0x2",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "EventName": "io_write_requests_cs_3",
+ "BriefDescription": "The number of inbound I/O write requests to memory: Data (32B or 64B): CS3",
+ "EventCode": "0x3",
+ "UMask": "0x70",
+ "ConstId": "0x3",
+ "Unit": "HYGON_DF"
+ },
+ {
+ "MetricName": "cpu_memory_bandwidth",
+ "MetricExpr": "(cpu_read_requests_cs_0 + cpu_read_requests_cs_1 + cpu_read_requests_cs_2 + cpu_read_requests_cs_3 + cpu_write_requests_cs_0 + cpu_write_requests_cs_1 + cpu_write_requests_cs_2 + cpu_write_requests_cs_3) * 64 / (duration_time * 1e6)",
+ "MetricConstraint": "NO_GROUP_EVENTS",
+ "BriefDescription": "Total CPU memory access bandwidth",
+ "MetricGroup": "DF_Memory_BW",
+ "ScaleUnit": "1MB/s"
+ },
+ {
+ "MetricName": "io_memory_max_bandwidth",
+ "MetricExpr": "(io_read_requests_cs_0 + io_read_requests_cs_1 + io_read_requests_cs_2 + io_read_requests_cs_3 + io_write_requests_cs_0 + io_write_requests_cs_1 + io_write_requests_cs_2 + io_write_requests_cs_3) * 64 / (duration_time * 1e6)",
+ "MetricConstraint": "NO_GROUP_EVENTS",
+ "BriefDescription": "Maximum inbound I/O memory access bandwidth (upper bound, assuming 64B per request. Actual inbound I/O requests may be 32B or 64B)",
+ "MetricGroup": "DF_Memory_BW",
+ "ScaleUnit": "1MB/s"
+ },
+ {
+ "MetricName": "io_memory_min_bandwidth",
+ "MetricExpr": "(io_read_requests_cs_0 + io_read_requests_cs_1 + io_read_requests_cs_2 + io_read_requests_cs_3 + io_write_requests_cs_0 + io_write_requests_cs_1 + io_write_requests_cs_2 + io_write_requests_cs_3) * 32 / (duration_time * 1e6)",
+ "MetricConstraint": "NO_GROUP_EVENTS",
+ "BriefDescription": "Minimum inbound I/O memory access bandwidth (lower bound, assuming 32B per request. Actual inbound I/O requests may be 32B or 64B)",
+ "MetricGroup": "DF_Memory_BW",
+ "ScaleUnit": "1MB/s"
+ }
+]
diff --git a/tools/perf/pmu-events/arch/x86/mapfile.csv b/tools/perf/pmu-events/arch/x86/mapfile.csv
index 50a403b429b1..ba17117245d1 100644
--- a/tools/perf/pmu-events/arch/x86/mapfile.csv
+++ b/tools/perf/pmu-events/arch/x86/mapfile.csv
@@ -45,3 +45,6 @@ AuthenticAMD-25-([245][[:xdigit:]]|[[:xdigit:]]),v1,amdzen3,core
AuthenticAMD-25-[[:xdigit:]]+,v1,amdzen4,core
AuthenticAMD-26-([12467][[:xdigit:]]|[[:xdigit:]]),v1,amdzen5,core
AuthenticAMD-26-[[:xdigit:]]+,v1,amdzen6,core
+HygonGenuine-24-[0123],v1,hygonm1,core
+HygonGenuine-24-[45],v1,hygonm4,core
+HygonGenuine-24-(1[012345678]|[6789ABCDEF]),v1,hygonm6,core
diff --git a/tools/perf/pmu-events/jevents.py b/tools/perf/pmu-events/jevents.py
index 376dc2d24162..63ae3a763a4f 100755
--- a/tools/perf/pmu-events/jevents.py
+++ b/tools/perf/pmu-events/jevents.py
@@ -294,6 +294,7 @@ class JsonEvent:
'L3PMC': 'amd_l3',
'DFPMC': 'amd_df',
'UMCPMC': 'amd_umc',
+ 'HYGON_DF': 'hygon_df',
'cpu_core': 'cpu_core',
'cpu_atom': 'cpu_atom',
'ali_drw': 'ali_drw',
@@ -390,6 +391,7 @@ class JsonEvent:
('Invert', 'inv='),
('SampleAfterValue', 'period='),
('UMask', 'umask='),
+ ('ConstId', 'constid='),
Is Const a shortening of Constant? It seems a little redundant in the
format name. Would CsId be more intention revealing? I don't know what
CS means, but this at least looks a little more intuitive to me than
Constant :-)
identifies the selected Data Fabric component instance, and I agree that the
current name is not particularly intuitive. I will rename it to CompId
(Component ID) in the next revision.
Many thanks!
Ian
Thanks,
Qi> Thanks,
Ian
('NodeType', 'type='),
('RdWrMask', 'rdwrmask='),
('EnAllCores', 'enallcores='),
--
2.34.1