Re: [PATCH v3 2/7] perf metric: Event "Compat" value supports matching multiple identifiers

[Jing Zhang]

在 2023/6/16 下午7:41, John Garry 写道:
> On 15/06/2023 03:18, Jing Zhang wrote:
>>>>>> Unit is the format of the event_source device name. We should match based on that as well as compat. I need to check the code again to understand how that is done... it has changed a good bit in 3 years.
>>>>>>
>>>>> This situation only happens on uncore metric. I happened to write wrong Unit, but the metric still matches.
>>>>>
>>>> I'm just double checking this now. I think any possible fix should be easy enough for current code but may be tricky for backport with lots of metric code changes.
>>> I also have code to re-work sys event metric support such that we don't require "compat" or "Unit" values for a metric when the metric is described in terms of event aliases. That code is 2 years old, so may take a bit of time to rebase. I'll look to do that now.
>>>
>> Sounds good!
> 
> BTW, I am just looking at your cmn JSONs in this series, and we have something like this:
> 
> index 0000000..e70ac1a
> --- /dev/null
> +++ b/tools/perf/pmu-events/arch/arm64/arm/cmn/sys/metrics.json
> @@ -0,0 +1,74 @@
> +[
> +    {
> +        "MetricName": "slc_miss_rate",
> +        "BriefDescription": "The system level cache miss rate include.",
> +        "MetricGroup": "arm_cmn",
> +        "MetricExpr": "hnf_cache_miss / hnf_slc_sf_cache_access",
> 
> So this expression uses event aliases hnf_cache_miss and hnf_slc_sf_cache_access - where are they defined in a JSON?
> 

Hi John,

I defined the aliases for these events in the JSON file during the RFC version. However, I later removed the alias
definitions for these events in subsequent versions due to the possibility of non-uniqueness and difficulty in defining
their EventCode. But this does not affect their usage in metrics. In other words, metrics can use the aliases without
defining event aliases in the JSON file.

In the past, I always thought that the function of the alias was to explain the meaning of these events in the perf list.
Or maybe I'm missing something?

Thanks,
Jing

> I could not see them. If they are not needed, then I may be missing something...
> 
> Thanks,
> John
> 
>