Re: [PATCH v4 14/35] perf evlist: Remove __evlist__add_default

From: Ian Rogers
Date: Sat May 27 2023 - 01:58:25 EST


On Fri, May 26, 2023 at 6:45 PM Arnaldo Carvalho de Melo
<acme@xxxxxxxxxx> wrote:
>
> Em Fri, May 26, 2023 at 02:53:49PM -0700, Ian Rogers escreveu:
> > __evlist__add_default adds a cycles event to a typically empty evlist
> > and was extended for hybrid with evlist__add_default_hybrid, as more
> > than 1 PMU was necessary. Rather than have dedicated logic for the
> > cycles event, this change switches to parsing 'cycles:P' which will
> > handle wildcarding the PMUs appropriately for hybrid.
>
> I think I reported this earlier, but at this point 'perf test python'
> breaks, I fixed it in the tmp.perf-tools-next:
>
> 19: 'import perf' in python : FAILED!
> ⬢[acme@toolbox perf-tools-next]$ git log --oneline -1
> fe4f622c4fc7a02a (HEAD) perf evlist: Remove __evlist__add_default
> ⬢[acme@toolbox perf-tools-next]$
> ⬢[acme@toolbox perf-tools-next]$ perf test -v python
> Couldn't bump rlimit(MEMLOCK), failures may take place when creating BPF maps, etc
> 19: 'import perf' in python :
> --- start ---
> test child forked, pid 2976621
> python usage test: "echo "import sys ; sys.path.append('/tmp/build/perf-tools-next/python'); import perf" | '/usr/bin/python3' "
> Traceback (most recent call last):
> File "<stdin>", line 1, in <module>
> ImportError: /tmp/build/perf-tools-next/python/perf.cpython-310-x86_64-linux-gnu.so: undefined symbol: parse_event
> test child finished with -1
> ---- end ----
> 'import perf' in python: FAILED!
> ⬢[acme@toolbox perf-tools-next]$
>
> Probably there will be a few more cases in the next patches, please
> check.

I'll rebase and resend. I needed to add:
https://lore.kernel.org/lkml/20230527055517.2711487-1-irogers@xxxxxxxxxx/
to repro the failure. The test was passing without it.

Thanks,
Ian

> - Arnaldo
>
> > Signed-off-by: Ian Rogers <irogers@xxxxxxxxxx>
> > Reviewed-by: Kan Liang <kan.liang@xxxxxxxxxxxxxxx>
> > ---
> > tools/perf/arch/x86/util/evsel.c | 20 --------------
> > tools/perf/builtin-record.c | 13 +++------
> > tools/perf/builtin-top.c | 10 ++++---
> > tools/perf/util/evlist-hybrid.c | 25 -----------------
> > tools/perf/util/evlist-hybrid.h | 1 -
> > tools/perf/util/evlist.c | 22 ++++++---------
> > tools/perf/util/evlist.h | 7 -----
> > tools/perf/util/evsel.c | 46 --------------------------------
> > tools/perf/util/evsel.h | 3 ---
> > 9 files changed, 17 insertions(+), 130 deletions(-)
> >
> > diff --git a/tools/perf/arch/x86/util/evsel.c b/tools/perf/arch/x86/util/evsel.c
> > index ea3972d785d1..153cdca94cd4 100644
> > --- a/tools/perf/arch/x86/util/evsel.c
> > +++ b/tools/perf/arch/x86/util/evsel.c
> > @@ -16,26 +16,6 @@ void arch_evsel__set_sample_weight(struct evsel *evsel)
> > evsel__set_sample_bit(evsel, WEIGHT_STRUCT);
> > }
> >
> > -void arch_evsel__fixup_new_cycles(struct perf_event_attr *attr)
> > -{
> > - struct perf_env env = { .total_mem = 0, } ;
> > -
> > - if (!perf_env__cpuid(&env))
> > - return;
> > -
> > - /*
> > - * On AMD, precise cycles event sampling internally uses IBS pmu.
> > - * But IBS does not have filtering capabilities and perf by default
> > - * sets exclude_guest = 1. This makes IBS pmu event init fail and
> > - * thus perf ends up doing non-precise sampling. Avoid it by clearing
> > - * exclude_guest.
> > - */
> > - if (env.cpuid && strstarts(env.cpuid, "AuthenticAMD"))
> > - attr->exclude_guest = 0;
> > -
> > - free(env.cpuid);
> > -}
> > -
> > /* Check whether the evsel's PMU supports the perf metrics */
> > bool evsel__sys_has_perf_metrics(const struct evsel *evsel)
> > {
> > diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
> > index 88f7b4241153..d80b54a6f450 100644
> > --- a/tools/perf/builtin-record.c
> > +++ b/tools/perf/builtin-record.c
> > @@ -4161,18 +4161,11 @@ int cmd_record(int argc, const char **argv)
> > record.opts.tail_synthesize = true;
> >
> > if (rec->evlist->core.nr_entries == 0) {
> > - if (perf_pmu__has_hybrid()) {
> > - err = evlist__add_default_hybrid(rec->evlist,
> > - !record.opts.no_samples);
> > - } else {
> > - err = __evlist__add_default(rec->evlist,
> > - !record.opts.no_samples);
> > - }
> > + bool can_profile_kernel = perf_event_paranoid_check(1);
> >
> > - if (err < 0) {
> > - pr_err("Not enough memory for event selector list\n");
> > + err = parse_event(rec->evlist, can_profile_kernel ? "cycles:P" : "cycles:Pu");
> > + if (err)
> > goto out;
> > - }
> > }
> >
> > if (rec->opts.target.tid && !rec->opts.no_inherit_set)
> > diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
> > index 48ee49e95c5e..27a7f068207d 100644
> > --- a/tools/perf/builtin-top.c
> > +++ b/tools/perf/builtin-top.c
> > @@ -1653,10 +1653,12 @@ int cmd_top(int argc, const char **argv)
> > if (annotate_check_args(&top.annotation_opts) < 0)
> > goto out_delete_evlist;
> >
> > - if (!top.evlist->core.nr_entries &&
> > - evlist__add_default(top.evlist) < 0) {
> > - pr_err("Not enough memory for event selector list\n");
> > - goto out_delete_evlist;
> > + if (!top.evlist->core.nr_entries) {
> > + bool can_profile_kernel = perf_event_paranoid_check(1);
> > + int err = parse_event(top.evlist, can_profile_kernel ? "cycles:P" : "cycles:Pu");
> > +
> > + if (err)
> > + goto out_delete_evlist;
> > }
> >
> > status = evswitch__init(&top.evswitch, top.evlist, stderr);
> > diff --git a/tools/perf/util/evlist-hybrid.c b/tools/perf/util/evlist-hybrid.c
> > index 0f59c80f27b2..64f78d06fe19 100644
> > --- a/tools/perf/util/evlist-hybrid.c
> > +++ b/tools/perf/util/evlist-hybrid.c
> > @@ -16,31 +16,6 @@
> > #include <perf/evsel.h>
> > #include <perf/cpumap.h>
> >
> > -int evlist__add_default_hybrid(struct evlist *evlist, bool precise)
> > -{
> > - struct evsel *evsel;
> > - struct perf_pmu *pmu;
> > - __u64 config;
> > - struct perf_cpu_map *cpus;
> > -
> > - perf_pmu__for_each_hybrid_pmu(pmu) {
> > - config = PERF_COUNT_HW_CPU_CYCLES |
> > - ((__u64)pmu->type << PERF_PMU_TYPE_SHIFT);
> > - evsel = evsel__new_cycles(precise, PERF_TYPE_HARDWARE,
> > - config);
> > - if (!evsel)
> > - return -ENOMEM;
> > -
> > - cpus = perf_cpu_map__get(pmu->cpus);
> > - evsel->core.cpus = cpus;
> > - evsel->core.own_cpus = perf_cpu_map__get(cpus);
> > - evsel->pmu_name = strdup(pmu->name);
> > - evlist__add(evlist, evsel);
> > - }
> > -
> > - return 0;
> > -}
> > -
> > bool evlist__has_hybrid(struct evlist *evlist)
> > {
> > struct evsel *evsel;
> > diff --git a/tools/perf/util/evlist-hybrid.h b/tools/perf/util/evlist-hybrid.h
> > index 4b000eda6626..0cded76eb344 100644
> > --- a/tools/perf/util/evlist-hybrid.h
> > +++ b/tools/perf/util/evlist-hybrid.h
> > @@ -7,7 +7,6 @@
> > #include "evlist.h"
> > #include <unistd.h>
> >
> > -int evlist__add_default_hybrid(struct evlist *evlist, bool precise);
> > bool evlist__has_hybrid(struct evlist *evlist);
> >
> > #endif /* __PERF_EVLIST_HYBRID_H */
> > diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
> > index 9dfa977193b3..63f8821a5395 100644
> > --- a/tools/perf/util/evlist.c
> > +++ b/tools/perf/util/evlist.c
> > @@ -93,8 +93,15 @@ struct evlist *evlist__new(void)
> > struct evlist *evlist__new_default(void)
> > {
> > struct evlist *evlist = evlist__new();
> > + bool can_profile_kernel;
> > + int err;
> > +
> > + if (!evlist)
> > + return NULL;
> >
> > - if (evlist && evlist__add_default(evlist)) {
> > + can_profile_kernel = perf_event_paranoid_check(1);
> > + err = parse_event(evlist, can_profile_kernel ? "cycles:P" : "cycles:Pu");
> > + if (err) {
> > evlist__delete(evlist);
> > evlist = NULL;
> > }
> > @@ -237,19 +244,6 @@ static void evlist__set_leader(struct evlist *evlist)
> > perf_evlist__set_leader(&evlist->core);
> > }
> >
> > -int __evlist__add_default(struct evlist *evlist, bool precise)
> > -{
> > - struct evsel *evsel;
> > -
> > - evsel = evsel__new_cycles(precise, PERF_TYPE_HARDWARE,
> > - PERF_COUNT_HW_CPU_CYCLES);
> > - if (evsel == NULL)
> > - return -ENOMEM;
> > -
> > - evlist__add(evlist, evsel);
> > - return 0;
> > -}
> > -
> > static struct evsel *evlist__dummy_event(struct evlist *evlist)
> > {
> > struct perf_event_attr attr = {
> > diff --git a/tools/perf/util/evlist.h b/tools/perf/util/evlist.h
> > index 5e7ff44f3043..664c6bf7b3e0 100644
> > --- a/tools/perf/util/evlist.h
> > +++ b/tools/perf/util/evlist.h
> > @@ -100,13 +100,6 @@ void evlist__delete(struct evlist *evlist);
> > void evlist__add(struct evlist *evlist, struct evsel *entry);
> > void evlist__remove(struct evlist *evlist, struct evsel *evsel);
> >
> > -int __evlist__add_default(struct evlist *evlist, bool precise);
> > -
> > -static inline int evlist__add_default(struct evlist *evlist)
> > -{
> > - return __evlist__add_default(evlist, true);
> > -}
> > -
> > int evlist__add_attrs(struct evlist *evlist, struct perf_event_attr *attrs, size_t nr_attrs);
> >
> > int __evlist__add_default_attrs(struct evlist *evlist,
> > diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
> > index 8c8f371ea2b5..1df8f967d2eb 100644
> > --- a/tools/perf/util/evsel.c
> > +++ b/tools/perf/util/evsel.c
> > @@ -316,48 +316,6 @@ struct evsel *evsel__new_idx(struct perf_event_attr *attr, int idx)
> > return evsel;
> > }
> >
> > -static bool perf_event_can_profile_kernel(void)
> > -{
> > - return perf_event_paranoid_check(1);
> > -}
> > -
> > -struct evsel *evsel__new_cycles(bool precise __maybe_unused, __u32 type, __u64 config)
> > -{
> > - struct perf_event_attr attr = {
> > - .type = type,
> > - .config = config,
> > - .exclude_kernel = !perf_event_can_profile_kernel(),
> > - };
> > - struct evsel *evsel;
> > -
> > - event_attr_init(&attr);
> > -
> > - /*
> > - * Now let the usual logic to set up the perf_event_attr defaults
> > - * to kick in when we return and before perf_evsel__open() is called.
> > - */
> > - evsel = evsel__new(&attr);
> > - if (evsel == NULL)
> > - goto out;
> > -
> > - arch_evsel__fixup_new_cycles(&evsel->core.attr);
> > -
> > - evsel->precise_max = true;
> > -
> > - /* use asprintf() because free(evsel) assumes name is allocated */
> > - if (asprintf(&evsel->name, "cycles%s%s%.*s",
> > - (attr.precise_ip || attr.exclude_kernel) ? ":" : "",
> > - attr.exclude_kernel ? "u" : "",
> > - attr.precise_ip ? attr.precise_ip + 1 : 0, "ppp") < 0)
> > - goto error_free;
> > -out:
> > - return evsel;
> > -error_free:
> > - evsel__delete(evsel);
> > - evsel = NULL;
> > - goto out;
> > -}
> > -
> > int copy_config_terms(struct list_head *dst, struct list_head *src)
> > {
> > struct evsel_config_term *pos, *tmp;
> > @@ -1131,10 +1089,6 @@ void __weak arch_evsel__set_sample_weight(struct evsel *evsel)
> > evsel__set_sample_bit(evsel, WEIGHT);
> > }
> >
> > -void __weak arch_evsel__fixup_new_cycles(struct perf_event_attr *attr __maybe_unused)
> > -{
> > -}
> > -
> > void __weak arch__post_evsel_config(struct evsel *evsel __maybe_unused,
> > struct perf_event_attr *attr __maybe_unused)
> > {
> > diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h
> > index df8928745fc6..429b172cc94d 100644
> > --- a/tools/perf/util/evsel.h
> > +++ b/tools/perf/util/evsel.h
> > @@ -243,8 +243,6 @@ static inline struct evsel *evsel__newtp(const char *sys, const char *name)
> > }
> > #endif
> >
> > -struct evsel *evsel__new_cycles(bool precise, __u32 type, __u64 config);
> > -
> > #ifdef HAVE_LIBTRACEEVENT
> > struct tep_event *event_format__new(const char *sys, const char *name);
> > #endif
> > @@ -312,7 +310,6 @@ void __evsel__reset_sample_bit(struct evsel *evsel, enum perf_event_sample_forma
> > void evsel__set_sample_id(struct evsel *evsel, bool use_sample_identifier);
> >
> > void arch_evsel__set_sample_weight(struct evsel *evsel);
> > -void arch_evsel__fixup_new_cycles(struct perf_event_attr *attr);
> > void arch__post_evsel_config(struct evsel *evsel, struct perf_event_attr *attr);
> >
> > int evsel__set_filter(struct evsel *evsel, const char *filter);
> > --
> > 2.41.0.rc0.172.g3f132b7071-goog
> >
>
> --
>
> - Arnaldo