Re: [PATCH v1] perf vendor events intel: Add umasks to PCU events.

From: Liang, Kan
Date: Wed Feb 28 2024 - 09:55:19 EST

Next message: Daniel Vetter: "Re: [PATCH] drm/scheduler: Simplify the allocation of slab caches in drm_sched_fence_slab_init"
Previous message: Arnd Bergmann: "Re: [PATCH] net: ethernet: ti: am65-cpsw-qos: fix non-bql configs"
In reply to: Ian Rogers: "Re: [PATCH v1] perf vendor events intel: Add umasks to PCU events."
Next in thread: Ian Rogers: "Re: [PATCH v1] perf vendor events intel: Add umasks to PCU events."
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On 2024-02-27 9:12 p.m., Ian Rogers wrote:
> On Tue, Feb 27, 2024 at 6:12 AM Liang, Kan <kan.liang@xxxxxxxxxxxxxxx> wrote:
>>
>>
>>
>> On 2024-02-26 3:15 p.m., Ian Rogers wrote:
>>> UMasks were being dropped leading to all PCU
>>> UNC_P_POWER_STATE_OCCUPANCY events having the same encoding. Don't
>>> drop the umask trying to be consistent with other sources of events
>>> like libpfm4 [1]. This applies the change from [2].
>>>
>>> [1] https://sourceforge.net/p/perfmon2/libpfm4/ci/master/tree/lib/events/intel_skx_unc_pcu_events.h#l30
>>> [2] https://github.com/captain5050/perfmon/commit/661a229996493448095fb55415ed568ceabc631b
>>>
>>> Signed-off-by: Ian Rogers <irogers@xxxxxxxxxx>
>>> ---
>>
>> Thanks Ian.
>>
>> Reviewed-by: Kan Liang <kan.liang@xxxxxxxxxxxxxxx>
>>
>> Thanks,
>> Kan
>>
>>> tools/perf/pmu-events/arch/x86/broadwellde/uncore-power.json | 3 +++
>>> tools/perf/pmu-events/arch/x86/broadwellx/uncore-power.json | 3 +++
>>> tools/perf/pmu-events/arch/x86/cascadelakex/uncore-power.json | 3 +++
>>> tools/perf/pmu-events/arch/x86/haswellx/uncore-power.json | 3 +++
>>> tools/perf/pmu-events/arch/x86/icelakex/uncore-power.json | 3 +++
>>> tools/perf/pmu-events/arch/x86/ivytown/uncore-power.json | 3 +++
>>> tools/perf/pmu-events/arch/x86/jaketown/uncore-power.json | 3 +++
>>> tools/perf/pmu-events/arch/x86/skylakex/uncore-power.json | 3 +++
>>> tools/perf/pmu-events/arch/x86/snowridgex/uncore-power.json | 3 +++
>>> 9 files changed, 27 insertions(+)
>>>
>>> diff --git a/tools/perf/pmu-events/arch/x86/broadwellde/uncore-power.json b/tools/perf/pmu-events/arch/x86/broadwellde/uncore-power.json
>>> index 83d20130c217..9a0bc34c08e1 100644
>>> --- a/tools/perf/pmu-events/arch/x86/broadwellde/uncore-power.json
>>> +++ b/tools/perf/pmu-events/arch/x86/broadwellde/uncore-power.json
>>> @@ -396,6 +396,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C0",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0x40",
>>> "Unit": "PCU"
>>> },
>>> {
>>> @@ -404,6 +405,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C3",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0x80",
>>> "Unit": "PCU"
>>> },
>>> {
>>> @@ -412,6 +414,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C6",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0xc0",
>>> "Unit": "PCU"
>>> },
>>> {
>>> diff --git a/tools/perf/pmu-events/arch/x86/broadwellx/uncore-power.json b/tools/perf/pmu-events/arch/x86/broadwellx/uncore-power.json
>>> index 83d20130c217..9a0bc34c08e1 100644
>>> --- a/tools/perf/pmu-events/arch/x86/broadwellx/uncore-power.json
>>> +++ b/tools/perf/pmu-events/arch/x86/broadwellx/uncore-power.json
>>> @@ -396,6 +396,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C0",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0x40",
>>> "Unit": "PCU"
>>> },
>>> {
>>> @@ -404,6 +405,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C3",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0x80",
>>> "Unit": "PCU"
>>> },
>>> {
>>> @@ -412,6 +414,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C6",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0xc0",
>>> "Unit": "PCU"
>>> },
>>> {
>>> diff --git a/tools/perf/pmu-events/arch/x86/cascadelakex/uncore-power.json b/tools/perf/pmu-events/arch/x86/cascadelakex/uncore-power.json
>>> index c6254af7a468..ceef46046488 100644
>>> --- a/tools/perf/pmu-events/arch/x86/cascadelakex/uncore-power.json
>>> +++ b/tools/perf/pmu-events/arch/x86/cascadelakex/uncore-power.json
>>> @@ -144,6 +144,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C0",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0x40",
>>> "Unit": "PCU"
>>> },
>>> {
>>> @@ -152,6 +153,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C3",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0x80",
>>> "Unit": "PCU"
>>> },
>>> {
>>> @@ -160,6 +162,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C6",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0xc0",
>>> "Unit": "PCU"
>>> },
>>> {
>>> diff --git a/tools/perf/pmu-events/arch/x86/haswellx/uncore-power.json b/tools/perf/pmu-events/arch/x86/haswellx/uncore-power.json
>>> index daebf1050acb..15097511cbd8 100644
>>> --- a/tools/perf/pmu-events/arch/x86/haswellx/uncore-power.json
>>> +++ b/tools/perf/pmu-events/arch/x86/haswellx/uncore-power.json
>>> @@ -428,6 +428,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C0",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0x40",
>>> "Unit": "PCU"
>>> },
>>> {
>>> @@ -436,6 +437,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C3",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0x80",
>>> "Unit": "PCU"
>>> },
>>> {
>>> @@ -444,6 +446,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C6",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0xc0",
>>> "Unit": "PCU"
>>> },
>>> {
>>> diff --git a/tools/perf/pmu-events/arch/x86/icelakex/uncore-power.json b/tools/perf/pmu-events/arch/x86/icelakex/uncore-power.json
>>> index ee4dac6fc797..920cab6ffe37 100644
>>> --- a/tools/perf/pmu-events/arch/x86/icelakex/uncore-power.json
>>> +++ b/tools/perf/pmu-events/arch/x86/icelakex/uncore-power.json
>>> @@ -151,6 +151,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C0",
>>> "PerPkg": "1",
>>> "PublicDescription": "Number of cores in C-State : C0 and C1 : This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0x40",
>>> "Unit": "PCU"
>>> },
>>> {
>>> @@ -159,6 +160,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C3",
>>> "PerPkg": "1",
>>> "PublicDescription": "Number of cores in C-State : C3 : This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0x80",
>>> "Unit": "PCU"
>>> },
>>> {
>>> @@ -167,6 +169,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C6",
>>> "PerPkg": "1",
>>> "PublicDescription": "Number of cores in C-State : C6 and C7 : This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0xc0",
>>> "Unit": "PCU"
>>> },
>>> {
>>> diff --git a/tools/perf/pmu-events/arch/x86/ivytown/uncore-power.json b/tools/perf/pmu-events/arch/x86/ivytown/uncore-power.json
>>> index 5df1ebfb89ea..4dac5810324b 100644
>>> --- a/tools/perf/pmu-events/arch/x86/ivytown/uncore-power.json
>>> +++ b/tools/perf/pmu-events/arch/x86/ivytown/uncore-power.json
>>> @@ -516,6 +516,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C0",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0x40",
>
> There's something wrong on Ivytown, I don't see a umask format in uncore_pmu:
> ```
> $ ls /sys/devices/uncore_pcu/format/
> edge filter_band0 filter_band2 occ_edge occ_sel
> event filter_band1 filter_band3 occ_invert thresh
> ```
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/x86/events/intel/uncore_snbep.c#n1606
>
> It does exist for skylake though:
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/x86/events/intel/uncore_snbep.c#n4145
>
> Kan, what's the right way to deal with this?

There is no umask in the uncore spec. Please see P120 of the spec.
https://www.intel.com/content/www/us/en/develop/download/intel-xeon-processor-e5-2600-v2-product-family-uncore-performance-monitoring-reference.html

It should be occ_sel.
The occ_sel = (the umask from the event list >> 6) & 0x3.

I don't think we want to change either the kernel or the perf tool.
There is nothing to help either if we add the occ_sel in the original
event list. So it seems the only choice should be handling it in the
convertor tool.

Is it possible to check the UNC_P_POWER_STATE_OCCUPANCY.CORES_C0 in IVT
and use "Filter": "occ_sel=0x1" to replace "UMask": "0x40"?

It seems everything in the filter will directly be appended. Is my
understanding correct?

Thanks,
Kan

>
> Thanks,
> Ian
>
>>> "Unit": "PCU"
>>> },
>>> {
>>> @@ -524,6 +525,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C3",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0x80",
>>> "Unit": "PCU"
>>> },
>>> {
>>> @@ -532,6 +534,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C6",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0xc0",
>>> "Unit": "PCU"
>>> },
>>> {
>>> diff --git a/tools/perf/pmu-events/arch/x86/jaketown/uncore-power.json b/tools/perf/pmu-events/arch/x86/jaketown/uncore-power.json
>>> index b3ee5d741015..a83e07dce947 100644
>>> --- a/tools/perf/pmu-events/arch/x86/jaketown/uncore-power.json
>>> +++ b/tools/perf/pmu-events/arch/x86/jaketown/uncore-power.json
>>> @@ -235,6 +235,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C0",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in C0. It can be used by itself to get the average number of cores in C0, with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0x40",
>>> "Unit": "PCU"
>>> },
>>> {
>>> @@ -243,6 +244,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C3",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in C0. It can be used by itself to get the average number of cores in C0, with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0x80",
>>> "Unit": "PCU"
>>> },
>>> {
>>> @@ -251,6 +253,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C6",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in C0. It can be used by itself to get the average number of cores in C0, with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0xc0",
>>> "Unit": "PCU"
>>> },
>>> {
>>> diff --git a/tools/perf/pmu-events/arch/x86/skylakex/uncore-power.json b/tools/perf/pmu-events/arch/x86/skylakex/uncore-power.json
>>> index c6254af7a468..ceef46046488 100644
>>> --- a/tools/perf/pmu-events/arch/x86/skylakex/uncore-power.json
>>> +++ b/tools/perf/pmu-events/arch/x86/skylakex/uncore-power.json
>>> @@ -144,6 +144,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C0",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0x40",
>>> "Unit": "PCU"
>>> },
>>> {
>>> @@ -152,6 +153,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C3",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0x80",
>>> "Unit": "PCU"
>>> },
>>> {
>>> @@ -160,6 +162,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C6",
>>> "PerPkg": "1",
>>> "PublicDescription": "This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0xc0",
>>> "Unit": "PCU"
>>> },
>>> {
>>> diff --git a/tools/perf/pmu-events/arch/x86/snowridgex/uncore-power.json b/tools/perf/pmu-events/arch/x86/snowridgex/uncore-power.json
>>> index a61ffca2dfea..dcf268467db9 100644
>>> --- a/tools/perf/pmu-events/arch/x86/snowridgex/uncore-power.json
>>> +++ b/tools/perf/pmu-events/arch/x86/snowridgex/uncore-power.json
>>> @@ -150,6 +150,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C0",
>>> "PerPkg": "1",
>>> "PublicDescription": "Number of cores in C-State : C0 and C1 : This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0x40",
>>> "Unit": "PCU"
>>> },
>>> {
>>> @@ -158,6 +159,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C3",
>>> "PerPkg": "1",
>>> "PublicDescription": "Number of cores in C-State : C3 : This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0x80",
>>> "Unit": "PCU"
>>> },
>>> {
>>> @@ -166,6 +168,7 @@
>>> "EventName": "UNC_P_POWER_STATE_OCCUPANCY.CORES_C6",
>>> "PerPkg": "1",
>>> "PublicDescription": "Number of cores in C-State : C6 and C7 : This is an occupancy event that tracks the number of cores that are in the chosen C-State. It can be used by itself to get the average number of cores in that C-state with thresholding to generate histograms, or with other PCU events and occupancy triggering to capture other details.",
>>> + "UMask": "0xc0",
>>> "Unit": "PCU"
>>> },
>>> {
>

Next message: Daniel Vetter: "Re: [PATCH] drm/scheduler: Simplify the allocation of slab caches in drm_sched_fence_slab_init"
Previous message: Arnd Bergmann: "Re: [PATCH] net: ethernet: ti: am65-cpsw-qos: fix non-bql configs"
In reply to: Ian Rogers: "Re: [PATCH v1] perf vendor events intel: Add umasks to PCU events."
Next in thread: Ian Rogers: "Re: [PATCH v1] perf vendor events intel: Add umasks to PCU events."
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]