| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Tue, 11 Jan 2022 17:01:10 -0300
From: Arnaldo Carvalho de Melo <acme@...nel.org>
To: Ian Rogers <irogers@...gle.com>
Cc: Andi Kleen <ak@...ux.intel.com>, Jiri Olsa <jolsa@...hat.com>,
Namhyung Kim <namhyung@...nel.org>,
John Garry <john.garry@...wei.com>,
Kajol Jain <kjain@...ux.ibm.com>,
"Paul A . Clarke" <pc@...ibm.com>,
Riccardo Mancini <rickyman7@...il.com>,
Kan Liang <kan.liang@...ux.intel.com>,
Peter Zijlstra <peterz@...radead.org>,
Ingo Molnar <mingo@...hat.com>,
Mark Rutland <mark.rutland@....com>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
linux-perf-users@...r.kernel.org, linux-kernel@...r.kernel.org,
Vineet Singh <vineet.singh@...el.com>,
James Clark <james.clark@....com>,
Mathieu Poirier <mathieu.poirier@...aro.org>,
Suzuki K Poulose <suzuki.poulose@....com>,
Mike Leach <mike.leach@...aro.org>,
Leo Yan <leo.yan@...aro.org>, coresight@...ts.linaro.org,
linux-arm-kernel@...ts.infradead.org, zhengjun.xing@...el.com,
eranian@...gle.com
Subject: Re: [PATCH v4 48/48] perf cpumap: Give CPUs their own type.
Em Tue, Jan 11, 2022 at 11:16:16AM -0800, Ian Rogers escreveu:
> On Mon, Jan 10, 2022 at 11:50 AM Arnaldo Carvalho de Melo
> <acme@...nel.org> wrote:
> >
> > Em Tue, Jan 04, 2022 at 10:13:51PM -0800, Ian Rogers escreveu:
> > > A common problem is confusing CPU map indices with the CPU, by wrapping
> > > the CPU with a struct then this is avoided. This approach is similar to
> > > atomic_t.
> >
> > This one needed this to build with BUILD_BPF_SKEL=1, please check, I'll
> > soon push this to tmp.perf/perf_cpu so that you can take a look and test
> > it.
>
>
> Thanks Arnaldo, I did eye-ball one issue where cpu_map__default_new's
> behavior would be impacted (nr would be 1 rather than nr_cpus). The
> fix is:
>
> --- a/tools/lib/perf/cpumap.c
> +++ b/tools/lib/perf/cpumap.c
> @@ -18,7 +18,7 @@ static struct perf_cpu_map *perf_cpu_map__alloc(int nr_cpus)
> if (!cpus)
> return NULL;
>
> - cpus->nr = 1;
> + cpus->nr = nr_cpus;
> refcount_set(&cpus->refcnt, 1);
> return cpus;
> }
> Could you add that into this patch as well?
Sure.
I´m now running perf-test after each of those csets, so far so good.
> Thanks,
> Ian
>
>
> > - Arnaldo
> >
> > diff --git a/tools/perf/util/bpf_counter.c b/tools/perf/util/bpf_counter.c
> > index 80d1a3a31052fe55..328479df5e16a638 100644
> > --- a/tools/perf/util/bpf_counter.c
> > +++ b/tools/perf/util/bpf_counter.c
> > @@ -540,7 +540,7 @@ static int bperf__load(struct evsel *evsel, struct target *target)
> > filter_type == BPERF_FILTER_TGID)
> > key = evsel->core.threads->map[i].pid;
> > else if (filter_type == BPERF_FILTER_CPU)
> > - key = evsel->core.cpus->map[i];
> > + key = evsel->core.cpus->map[i].cpu;
> > else
> > break;
> >
> > @@ -584,7 +584,7 @@ static int bperf_sync_counters(struct evsel *evsel)
> >
> > num_cpu = all_cpu_map->nr;
> > for (i = 0; i < num_cpu; i++) {
> > - cpu = all_cpu_map->map[i];
> > + cpu = all_cpu_map->map[i].cpu;
> > bperf_trigger_reading(evsel->bperf_leader_prog_fd, cpu);
> > }
> > return 0;
> > @@ -605,7 +605,7 @@ static int bperf__disable(struct evsel *evsel)
> > static int bperf__read(struct evsel *evsel)
> > {
> > struct bperf_follower_bpf *skel = evsel->follower_skel;
> > - __u32 num_cpu_bpf = cpu__max_cpu();
> > + __u32 num_cpu_bpf = cpu__max_cpu().cpu;
> > struct bpf_perf_event_value values[num_cpu_bpf];
> > int reading_map_fd, err = 0;
> > __u32 i;
> > @@ -615,6 +615,7 @@ static int bperf__read(struct evsel *evsel)
> > reading_map_fd = bpf_map__fd(skel->maps.accum_readings);
> >
> > for (i = 0; i < bpf_map__max_entries(skel->maps.accum_readings); i++) {
> > + struct perf_cpu entry;
> > __u32 cpu;
> >
> > err = bpf_map_lookup_elem(reading_map_fd, &i, values);
> > @@ -624,14 +625,15 @@ static int bperf__read(struct evsel *evsel)
> > case BPERF_FILTER_GLOBAL:
> > assert(i == 0);
> >
> > - perf_cpu_map__for_each_cpu(cpu, j, all_cpu_map) {
> > + perf_cpu_map__for_each_cpu(entry, j, all_cpu_map) {
> > + cpu = entry.cpu;
> > perf_counts(evsel->counts, cpu, 0)->val = values[cpu].counter;
> > perf_counts(evsel->counts, cpu, 0)->ena = values[cpu].enabled;
> > perf_counts(evsel->counts, cpu, 0)->run = values[cpu].running;
> > }
> > break;
> > case BPERF_FILTER_CPU:
> > - cpu = evsel->core.cpus->map[i];
> > + cpu = evsel->core.cpus->map[i].cpu;
> > perf_counts(evsel->counts, i, 0)->val = values[cpu].counter;
> > perf_counts(evsel->counts, i, 0)->ena = values[cpu].enabled;
> > perf_counts(evsel->counts, i, 0)->run = values[cpu].running;
> > diff --git a/tools/perf/util/bpf_counter_cgroup.c b/tools/perf/util/bpf_counter_cgroup.c
> > index cbc6c2bca488f6bf..631e34a0b66ff084 100644
> > --- a/tools/perf/util/bpf_counter_cgroup.c
> > +++ b/tools/perf/util/bpf_counter_cgroup.c
> > @@ -48,7 +48,7 @@ static int bperf_load_program(struct evlist *evlist)
> > struct cgroup *cgrp, *leader_cgrp;
> > __u32 i, cpu;
> > __u32 nr_cpus = evlist->core.all_cpus->nr;
> > - int total_cpus = cpu__max_cpu();
> > + int total_cpus = cpu__max_cpu().cpu;
> > int map_size, map_fd;
> > int prog_fd, err;
> >
> > @@ -125,7 +125,7 @@ static int bperf_load_program(struct evlist *evlist)
> > for (cpu = 0; cpu < nr_cpus; cpu++) {
> > int fd = FD(evsel, cpu);
> > __u32 idx = evsel->core.idx * total_cpus +
> > - evlist->core.all_cpus->map[cpu];
> > + evlist->core.all_cpus->map[cpu].cpu;
> >
> > err = bpf_map_update_elem(map_fd, &idx, &fd,
> > BPF_ANY);
> > @@ -212,7 +212,7 @@ static int bperf_cgrp__sync_counters(struct evlist *evlist)
> > int prog_fd = bpf_program__fd(skel->progs.trigger_read);
> >
> > for (i = 0; i < nr_cpus; i++) {
> > - cpu = evlist->core.all_cpus->map[i];
> > + cpu = evlist->core.all_cpus->map[i].cpu;
> > bperf_trigger_reading(prog_fd, cpu);
> > }
> >
> > @@ -245,7 +245,7 @@ static int bperf_cgrp__read(struct evsel *evsel)
> > {
> > struct evlist *evlist = evsel->evlist;
> > int i, cpu, nr_cpus = evlist->core.all_cpus->nr;
> > - int total_cpus = cpu__max_cpu();
> > + int total_cpus = cpu__max_cpu().cpu;
> > struct perf_counts_values *counts;
> > struct bpf_perf_event_value *values;
> > int reading_map_fd, err = 0;
> > @@ -272,7 +272,7 @@ static int bperf_cgrp__read(struct evsel *evsel)
> > }
> >
> > for (i = 0; i < nr_cpus; i++) {
> > - cpu = evlist->core.all_cpus->map[i];
> > + cpu = evlist->core.all_cpus->map[i].cpu;
> >
> > counts = perf_counts(evsel->counts, i, 0);
> > counts->val = values[cpu].counter;
> > diff --git a/tools/perf/util/bpf_ftrace.c b/tools/perf/util/bpf_ftrace.c
> > index 28dc4c60c7884818..d756cc66eef32ae8 100644
> > --- a/tools/perf/util/bpf_ftrace.c
> > +++ b/tools/perf/util/bpf_ftrace.c
> > @@ -63,7 +63,7 @@ int perf_ftrace__latency_prepare_bpf(struct perf_ftrace *ftrace)
> > fd = bpf_map__fd(skel->maps.cpu_filter);
> >
> > for (i = 0; i < ncpus; i++) {
> > - cpu = perf_cpu_map__cpu(ftrace->evlist->core.cpus, i);
> > + cpu = perf_cpu_map__cpu(ftrace->evlist->core.cpus, i).cpu;
> > bpf_map_update_elem(fd, &cpu, &val, BPF_ANY);
> > }
> > }
> > @@ -122,7 +122,7 @@ int perf_ftrace__latency_read_bpf(struct perf_ftrace *ftrace __maybe_unused,
> > int i, fd, err;
> > u32 idx;
> > u64 *hist;
> > - int ncpus = cpu__max_cpu();
> > + int ncpus = cpu__max_cpu().cpu;
> >
> > fd = bpf_map__fd(skel->maps.latency);
> >
--
- Arnaldo
Powered by blists - more mailing lists