Re: [PATCH v5 1/5] perf metric: Event "Compat" value supports matching multiple identifiers

From: Jing Zhang
Date: Wed Aug 02 2023 - 05:38:43 EST




在 2023/7/31 下午6:59, Jing Zhang 写道:
>
> 在 2023/7/28 下午4:11, John Garry 写道:
>> On 28/07/2023 07:17, Jing Zhang wrote:
>>> The jevent "Compat" is used for uncore PMU alias or metric definitions.
>>>
>>> The same PMU driver has different PMU identifiers due to different hardware
>>> versions and types, but they may have some common PMU event/metric. Since a
>>> Compat value can only match one identifier, when adding the same event
>>> alias and metric to PMUs with different identifiers, each identifier needs
>>> to be defined once, which is not streamlined enough.
>>>
>>> So let "Compat" value supports matching multiple identifiers. For example,
>>> the Compat value {abcde;123*}
>> why not use a comma-separated list? that is more common
>>
>
> Hi John,
>
> I use a semicolon instead of a comma because I want to distinguish it from the function
> of the comma in "Unit" and avoid confusion between the use of commas in "Unit" and "Compat".
> Because in “Compat”, the semicolon means "or". So I think semicolons are more appropriate,
> what do you think?
>
>>> can match the PMU identifier "abcde" and the
>>> the PMU identifier with the prefix "123",
>>
>> I have to admit that this is not a great example as it does not match an expected real-life scenario. I mean, I would not expect a PMU identifier for the same PMU to be in either format "abcde" or "123*". I would expect to be in only ever one format.
>>
>
> Get, I'll pick a more appropriate example {43401;436*}(CMN600 r0p0 and all CMN650).
>
>>> where "*" is a wildcard.
>>> Tokens in Unit field are delimited by ';' with no spaces.
>>>
>>> Signed-off-by: Jing Zhang <renyu.zj@xxxxxxxxxxxxxxxxx>
>>> ---
>>>   tools/perf/util/metricgroup.c |  2 +-
>>>   tools/perf/util/pmu.c         | 27 ++++++++++++++++++++++++++-
>>>   tools/perf/util/pmu.h         |  1 +
>>>   3 files changed, 28 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/tools/perf/util/metricgroup.c b/tools/perf/util/metricgroup.c
>>> index 5e9c657..ff81bc5 100644
>>> --- a/tools/perf/util/metricgroup.c
>>> +++ b/tools/perf/util/metricgroup.c
>>> @@ -477,7 +477,7 @@ static int metricgroup__sys_event_iter(const struct pmu_metric *pm,
>>>         while ((pmu = perf_pmu__scan(pmu))) {
>>>   -        if (!pmu->id || strcmp(pmu->id, pm->compat))
>>> +        if (!pmu->id || !pmu_uncore_identifier_match(pmu->id, pm->compat))
>>>               continue;
>>>             return d->fn(pm, table, d->data);
>>> diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c
>>> index ad209c8..3ae249b 100644
>>> --- a/tools/perf/util/pmu.c
>>> +++ b/tools/perf/util/pmu.c
>>> @@ -776,6 +776,31 @@ static bool pmu_uncore_alias_match(const char *pmu_name, const char *name)
>>>       return res;
>>>   }
>>>   +bool pmu_uncore_identifier_match(const char *id, const char *compat)
>>> +{
>>> +    char *tmp = NULL, *tok, *str;
>>> +    bool res;
>>> +    int n;
>>> +
>>> +    str = strdup(compat);
>>
>> why duplicate this? are you modifying something?
>>
>
> This is really a redundant step, I will remove it.
>

Hi John,

I reviewed this code again and found that it still needs to duplicate "compat" because "compat" is a
const str* type and cannot be used as a parameter for the strtok_r function. If it is cast to char*,
using "compat" as a parameter for strtok_r is also unsafe and can cause a "Segmentation fault" error.
Therefore, let's keep the step of duplicating "compat".

Thanks,
Jing


>>> +    if (!str)
>>> +        return false;
>>> +
>>> +    tok = strtok_r(str, ";", &tmp);
>>> +    for (; tok; tok = strtok_r(NULL, ";", &tmp)) {
>>> +        n = strlen(tok);
>>> +        if ((tok[n - 1] == '*' && !strncmp(id, tok, n - 1)) ||
>>> +            !strcmp(id, tok)) {
>>> +            res = true;
>>> +            goto out;
>>> +        }
>>> +    }
>>> +    res = false;
>>> +out:
>>> +    free(str);
>>> +    return res;
>>> +}
>>> +
>>>   struct pmu_add_cpu_aliases_map_data {
>>>       struct list_head *head;
>>>       const char *name;
>>> @@ -847,7 +872,7 @@ static int pmu_add_sys_aliases_iter_fn(const struct pmu_event *pe,
>>
>> This is not for metrics specifically. You are really doing 2x things here. I suggest that you split the patch out into 1st pmu.c change and 2nd metricgroup.c change
>>
>
> Ok, will do.
>
>>>       if (!pe->compat || !pe->pmu)
>>>           return 0;
>>>   -    if (!strcmp(pmu->id, pe->compat) &&
>>> +    if (pmu_uncore_identifier_match(pmu->id, pe->compat) &&
>>>           pmu_uncore_alias_match(pe->pmu, pmu->name)) {
>>
>> nit: let's change order to check alias and then identifier
>>
>
> Will do.
>
>
> Thanks,
> Jing
>
>>>           __perf_pmu__new_alias(idata->head, -1,
>>>                         (char *)pe->name,
>>> diff --git a/tools/perf/util/pmu.h b/tools/perf/util/pmu.h
>>> index b9a02de..9d4385d 100644
>>> --- a/tools/perf/util/pmu.h
>>> +++ b/tools/perf/util/pmu.h
>>> @@ -241,6 +241,7 @@ void pmu_add_cpu_aliases_table(struct list_head *head, struct perf_pmu *pmu,
>>>   char *perf_pmu__getcpuid(struct perf_pmu *pmu);
>>>   const struct pmu_events_table *pmu_events_table__find(void);
>>>   const struct pmu_metrics_table *pmu_metrics_table__find(void);
>>> +bool pmu_uncore_identifier_match(const char *id, const char *compat);
>>>   void perf_pmu_free_alias(struct perf_pmu_alias *alias);
>>>     int perf_pmu__convert_scale(const char *scale, char **end, double *sval);
>>
>> Thanks,
>> John