perf stat: Enable BPF counter with --for-each-cgroup

Recently bperf was added to use BPF to count perf events for various
purposes. This is an extension for the approach and targetting to
cgroup usages.

Unlike the other bperf, it doesn't share the events with other
processes but it'd reduce unnecessary events (and the overhead of
multiplexing) for each monitored cgroup within the perf session.

When --for-each-cgroup is used with --bpf-counters, it will open
cgroup-switches event per cpu internally and attach the new BPF
program to read given perf_events and to aggregate the results for
cgroups. It's only called when task is switched to a task in a
different cgroup.

Signed-off-by: Namhyung Kim <[e-mail]>
Acked-by: Song Liu <[e-mail]>
Cc: Andi Kleen <[e-mail]>
Cc: Ian Rogers <[e-mail]>
Cc: Jiri Olsa <[e-mail]>
Cc: Peter Zijlstra <[e-mail]>
Cc: Stephane Eranian <[e-mail]>
Link: http://lore.kernel.org/lkml/[e-mail]
Signed-off-by: Arnaldo Carvalho de Melo <[e-mail]>

tools/perf/Makefile.perf

···
 SKEL_TMP_OUT := $(abspath $(SKEL_OUT)/.tmp)
 SKELETONS := $(SKEL_OUT)/bpf_prog_profiler.skel.h
 SKELETONS += $(SKEL_OUT)/bperf_leader.skel.h $(SKEL_OUT)/bperf_follower.skel.h
+SKELETONS += $(SKEL_OUT)/bperf_cgroup.skel.h
 
 ifdef BUILD_BPF_SKEL
 BPFTOOL := $(SKEL_TMP_OUT)/bootstrap/bpftool
···
 	CFLAGS= $(MAKE) -C ../bpf/bpftool \
 		OUTPUT=$(SKEL_TMP_OUT)/ bootstrap
 
-$(SKEL_TMP_OUT)/%.bpf.o: util/bpf_skel/%.bpf.c $(LIBBPF) | $(SKEL_TMP_OUT)
+VMLINUX_BTF_PATHS ?= $(if $(O),$(O)/vmlinux)				\
+		     $(if $(KBUILD_OUTPUT),$(KBUILD_OUTPUT)/vmlinux)	\
+		     ../../vmlinux					\
+		     /sys/kernel/btf/vmlinux				\
+		     /boot/vmlinux-$(shell uname -r)
+VMLINUX_BTF ?= $(abspath $(firstword $(wildcard $(VMLINUX_BTF_PATHS))))
+
+$(SKEL_OUT)/vmlinux.h: $(VMLINUX_BTF) $(BPFTOOL)
+ifeq ($(VMLINUX_H),)
+	$(QUIET_GEN)$(BPFTOOL) btf dump file $< format c > $@
+else
+	$(Q)cp "$(VMLINUX_H)" $@
+endif
+
+$(SKEL_TMP_OUT)/%.bpf.o: util/bpf_skel/%.bpf.c $(LIBBPF) $(SKEL_OUT)/vmlinux.h | $(SKEL_TMP_OUT)
 	$(QUIET_CLANG)$(CLANG) -g -O2 -target bpf -Wall -Werror $(BPF_INCLUDE) \
 	  -c $(filter util/bpf_skel/%.bpf.c,$^) -o $@ && $(LLVM_STRIP) -g $@
 

tools/perf/util/Build

···
 perf-$(CONFIG_LIBBPF) += bpf-loader.o
 perf-$(CONFIG_LIBBPF) += bpf_map.o
 perf-$(CONFIG_PERF_BPF_SKEL) += bpf_counter.o
+perf-$(CONFIG_PERF_BPF_SKEL) += bpf_counter_cgroup.o
 perf-$(CONFIG_BPF_PROLOGUE) += bpf-prologue.o
 perf-$(CONFIG_LIBELF) += symbol-elf.o
 perf-$(CONFIG_LIBELF) += probe-file.o

tools/perf/util/bpf_counter.c

···
 #include "evsel.h"
 #include "evlist.h"
 #include "target.h"
+#include "cgroup.h"
 #include "cpumap.h"
 #include "thread_map.h"
 
···
 	.destroy    = bperf__destroy,
 };
 
+extern struct bpf_counter_ops bperf_cgrp_ops;
+
 static inline bool bpf_counter_skip(struct evsel *evsel)
 {
 	return list_empty(&evsel->bpf_counter_list) &&
···
 {
 	if (target->bpf_str)
 		evsel->bpf_counter_ops = &bpf_program_profiler_ops;
+	else if (cgrp_event_expanded && target->use_bpf)
+		evsel->bpf_counter_ops = &bperf_cgrp_ops;
 	else if (target->use_bpf || evsel->bpf_counter ||
 		 evsel__match_bpf_counter_events(evsel->name))
 		evsel->bpf_counter_ops = &bperf_ops;

tools/perf/util/bpf_counter_cgroup.c

···
+// SPDX-License-Identifier: GPL-2.0
+
+/* Copyright (c) 2021 Facebook */
+/* Copyright (c) 2021 Google */
+
+#include <assert.h>
+#include <limits.h>
+#include <unistd.h>
+#include <sys/file.h>
+#include <sys/time.h>
+#include <sys/resource.h>
+#include <linux/err.h>
+#include <linux/zalloc.h>
+#include <linux/perf_event.h>
+#include <api/fs/fs.h>
+#include <perf/bpf_perf.h>
+
+#include "affinity.h"
+#include "bpf_counter.h"
+#include "cgroup.h"
+#include "counts.h"
+#include "debug.h"
+#include "evsel.h"
+#include "evlist.h"
+#include "target.h"
+#include "cpumap.h"
+#include "thread_map.h"
+
+#include "bpf_skel/bperf_cgroup.skel.h"
+
+static struct perf_event_attr cgrp_switch_attr = {
+	.type = PERF_TYPE_SOFTWARE,
+	.config = PERF_COUNT_SW_CGROUP_SWITCHES,
+	.size = sizeof(cgrp_switch_attr),
+	.sample_period = 1,
+	.disabled = 1,
+};
+
+static struct evsel *cgrp_switch;
+static struct bperf_cgroup_bpf *skel;
+
+#define FD(evt, cpu) (*(int *)xyarray__entry(evt->core.fd, cpu, 0))
+
+static int bperf_load_program(struct evlist *evlist)
+{
+	struct bpf_link *link;
+	struct evsel *evsel;
+	struct cgroup *cgrp, *leader_cgrp;
+	__u32 i, cpu;
+	__u32 nr_cpus = evlist->core.all_cpus->nr;
+	int total_cpus = cpu__max_cpu();
+	int map_size, map_fd;
+	int prog_fd, err;
+
+	skel = bperf_cgroup_bpf__open();
+	if (!skel) {
+		pr_err("Failed to open cgroup skeleton\n");
+		return -1;
+	}
+
+	skel->rodata->num_cpus = total_cpus;
+	skel->rodata->num_events = evlist->core.nr_entries / nr_cgroups;
+
+	BUG_ON(evlist->core.nr_entries % nr_cgroups != 0);
+
+	/* we need one copy of events per cpu for reading */
+	map_size = total_cpus * evlist->core.nr_entries / nr_cgroups;
+	bpf_map__resize(skel->maps.events, map_size);
+	bpf_map__resize(skel->maps.cgrp_idx, nr_cgroups);
+	/* previous result is saved in a per-cpu array */
+	map_size = evlist->core.nr_entries / nr_cgroups;
+	bpf_map__resize(skel->maps.prev_readings, map_size);
+	/* cgroup result needs all events (per-cpu) */
+	map_size = evlist->core.nr_entries;
+	bpf_map__resize(skel->maps.cgrp_readings, map_size);
+
+	set_max_rlimit();
+
+	err = bperf_cgroup_bpf__load(skel);
+	if (err) {
+		pr_err("Failed to load cgroup skeleton\n");
+		goto out;
+	}
+
+	if (cgroup_is_v2("perf_event") > 0)
+		skel->bss->use_cgroup_v2 = 1;
+
+	err = -1;
+
+	cgrp_switch = evsel__new(&cgrp_switch_attr);
+	if (evsel__open_per_cpu(cgrp_switch, evlist->core.all_cpus, -1) < 0) {
+		pr_err("Failed to open cgroup switches event\n");
+		goto out;
+	}
+
+	for (i = 0; i < nr_cpus; i++) {
+		link = bpf_program__attach_perf_event(skel->progs.on_cgrp_switch,
+						      FD(cgrp_switch, i));
+		if (IS_ERR(link)) {
+			pr_err("Failed to attach cgroup program\n");
+			err = PTR_ERR(link);
+			goto out;
+		}
+	}
+
+	/*
+	 * Update cgrp_idx map from cgroup-id to event index.
+	 */
+	cgrp = NULL;
+	i = 0;
+
+	evlist__for_each_entry(evlist, evsel) {
+		if (cgrp == NULL || evsel->cgrp == leader_cgrp) {
+			leader_cgrp = evsel->cgrp;
+			evsel->cgrp = NULL;
+
+			/* open single copy of the events w/o cgroup */
+			err = evsel__open_per_cpu(evsel, evlist->core.all_cpus, -1);
+			if (err) {
+				pr_err("Failed to open first cgroup events\n");
+				goto out;
+			}
+
+			map_fd = bpf_map__fd(skel->maps.events);
+			for (cpu = 0; cpu < nr_cpus; cpu++) {
+				int fd = FD(evsel, cpu);
+				__u32 idx = evsel->idx * total_cpus +
+					evlist->core.all_cpus->map[cpu];
+
+				err = bpf_map_update_elem(map_fd, &idx, &fd,
+							  BPF_ANY);
+				if (err < 0) {
+					pr_err("Failed to update perf_event fd\n");
+					goto out;
+				}
+			}
+
+			evsel->cgrp = leader_cgrp;
+		}
+		evsel->supported = true;
+
+		if (evsel->cgrp == cgrp)
+			continue;
+
+		cgrp = evsel->cgrp;
+
+		if (read_cgroup_id(cgrp) < 0) {
+			pr_err("Failed to get cgroup id\n");
+			err = -1;
+			goto out;
+		}
+
+		map_fd = bpf_map__fd(skel->maps.cgrp_idx);
+		err = bpf_map_update_elem(map_fd, &cgrp->id, &i, BPF_ANY);
+		if (err < 0) {
+			pr_err("Failed to update cgroup index map\n");
+			goto out;
+		}
+
+		i++;
+	}
+
+	/*
+	 * bperf uses BPF_PROG_TEST_RUN to get accurate reading. Check
+	 * whether the kernel support it
+	 */
+	prog_fd = bpf_program__fd(skel->progs.trigger_read);
+	err = bperf_trigger_reading(prog_fd, 0);
+	if (err) {
+		pr_warning("The kernel does not support test_run for raw_tp BPF programs.\n"
+			   "Therefore, --for-each-cgroup might show inaccurate readings\n");
+		err = 0;
+	}
+
+out:
+	return err;
+}
+
+static int bperf_cgrp__load(struct evsel *evsel,
+			    struct target *target __maybe_unused)
+{
+	static bool bperf_loaded = false;
+
+	evsel->bperf_leader_prog_fd = -1;
+	evsel->bperf_leader_link_fd = -1;
+
+	if (!bperf_loaded && bperf_load_program(evsel->evlist))
+		return -1;
+
+	bperf_loaded = true;
+	/* just to bypass bpf_counter_skip() */
+	evsel->follower_skel = (struct bperf_follower_bpf *)skel;
+
+	return 0;
+}
+
+static int bperf_cgrp__install_pe(struct evsel *evsel __maybe_unused,
+				  int cpu __maybe_unused, int fd __maybe_unused)
+{
+	/* nothing to do */
+	return 0;
+}
+
+/*
+ * trigger the leader prog on each cpu, so the cgrp_reading map could get
+ * the latest results.
+ */
+static int bperf_cgrp__sync_counters(struct evlist *evlist)
+{
+	int i, cpu;
+	int nr_cpus = evlist->core.all_cpus->nr;
+	int prog_fd = bpf_program__fd(skel->progs.trigger_read);
+
+	for (i = 0; i < nr_cpus; i++) {
+		cpu = evlist->core.all_cpus->map[i];
+		bperf_trigger_reading(prog_fd, cpu);
+	}
+
+	return 0;
+}
+
+static int bperf_cgrp__enable(struct evsel *evsel)
+{
+	if (evsel->idx)
+		return 0;
+
+	bperf_cgrp__sync_counters(evsel->evlist);
+
+	skel->bss->enabled = 1;
+	return 0;
+}
+
+static int bperf_cgrp__disable(struct evsel *evsel)
+{
+	if (evsel->idx)
+		return 0;
+
+	bperf_cgrp__sync_counters(evsel->evlist);
+
+	skel->bss->enabled = 0;
+	return 0;
+}
+
+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();
+	struct perf_counts_values *counts;
+	struct bpf_perf_event_value *values;
+	int reading_map_fd, err = 0;
+	__u32 idx;
+
+	if (evsel->idx)
+		return 0;
+
+	bperf_cgrp__sync_counters(evsel->evlist);
+
+	values = calloc(total_cpus, sizeof(*values));
+	if (values == NULL)
+		return -ENOMEM;
+
+	reading_map_fd = bpf_map__fd(skel->maps.cgrp_readings);
+
+	evlist__for_each_entry(evlist, evsel) {
+		idx = evsel->idx;
+		err = bpf_map_lookup_elem(reading_map_fd, &idx, values);
+		if (err) {
+			pr_err("bpf map lookup falied: idx=%u, event=%s, cgrp=%s\n",
+			       idx, evsel__name(evsel), evsel->cgrp->name);
+			goto out;
+		}
+
+		for (i = 0; i < nr_cpus; i++) {
+			cpu = evlist->core.all_cpus->map[i];
+
+			counts = perf_counts(evsel->counts, i, 0);
+			counts->val = values[cpu].counter;
+			counts->ena = values[cpu].enabled;
+			counts->run = values[cpu].running;
+		}
+	}
+
+out:
+	free(values);
+	return err;
+}
+
+static int bperf_cgrp__destroy(struct evsel *evsel)
+{
+	if (evsel->idx)
+		return 0;
+
+	bperf_cgroup_bpf__destroy(skel);
+	evsel__delete(cgrp_switch);  // it'll destroy on_switch progs too
+
+	return 0;
+}
+
+struct bpf_counter_ops bperf_cgrp_ops = {
+	.load       = bperf_cgrp__load,
+	.enable     = bperf_cgrp__enable,
+	.disable    = bperf_cgrp__disable,
+	.read       = bperf_cgrp__read,
+	.install_pe = bperf_cgrp__install_pe,
+	.destroy    = bperf_cgrp__destroy,
+};

tools/perf/util/bpf_skel/bperf_cgroup.bpf.c

···
+// SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+// Copyright (c) 2021 Facebook
+// Copyright (c) 2021 Google
+#include "vmlinux.h"
+#include <bpf/bpf_helpers.h>
+#include <bpf/bpf_tracing.h>
+#include <bpf/bpf_core_read.h>
+
+#define MAX_LEVELS  10  // max cgroup hierarchy level: arbitrary
+#define MAX_EVENTS  32  // max events per cgroup: arbitrary
+
+// NOTE: many of map and global data will be modified before loading
+//       from the userspace (perf tool) using the skeleton helpers.
+
+// single set of global perf events to measure
+struct {
+	__uint(type, BPF_MAP_TYPE_PERF_EVENT_ARRAY);
+	__uint(key_size, sizeof(__u32));
+	__uint(value_size, sizeof(int));
+	__uint(max_entries, 1);
+} events SEC(".maps");
+
+// from cgroup id to event index
+struct {
+	__uint(type, BPF_MAP_TYPE_HASH);
+	__uint(key_size, sizeof(__u64));
+	__uint(value_size, sizeof(__u32));
+	__uint(max_entries, 1);
+} cgrp_idx SEC(".maps");
+
+// per-cpu event snapshots to calculate delta
+struct {
+	__uint(type, BPF_MAP_TYPE_PERCPU_ARRAY);
+	__uint(key_size, sizeof(__u32));
+	__uint(value_size, sizeof(struct bpf_perf_event_value));
+} prev_readings SEC(".maps");
+
+// aggregated event values for each cgroup (per-cpu)
+// will be read from the user-space
+struct {
+	__uint(type, BPF_MAP_TYPE_PERCPU_ARRAY);
+	__uint(key_size, sizeof(__u32));
+	__uint(value_size, sizeof(struct bpf_perf_event_value));
+} cgrp_readings SEC(".maps");
+
+const volatile __u32 num_events = 1;
+const volatile __u32 num_cpus = 1;
+
+int enabled = 0;
+int use_cgroup_v2 = 0;
+
+static inline int get_cgroup_v1_idx(__u32 *cgrps, int size)
+{
+	struct task_struct *p = (void *)bpf_get_current_task();
+	struct cgroup *cgrp;
+	register int i = 0;
+	__u32 *elem;
+	int level;
+	int cnt;
+
+	cgrp = BPF_CORE_READ(p, cgroups, subsys[perf_event_cgrp_id], cgroup);
+	level = BPF_CORE_READ(cgrp, level);
+
+	for (cnt = 0; i < MAX_LEVELS; i++) {
+		__u64 cgrp_id;
+
+		if (i > level)
+			break;
+
+		// convert cgroup-id to a map index
+		cgrp_id = BPF_CORE_READ(cgrp, ancestor_ids[i]);
+		elem = bpf_map_lookup_elem(&cgrp_idx, &cgrp_id);
+		if (!elem)
+			continue;
+
+		cgrps[cnt++] = *elem;
+		if (cnt == size)
+			break;
+	}
+
+	return cnt;
+}
+
+static inline int get_cgroup_v2_idx(__u32 *cgrps, int size)
+{
+	register int i = 0;
+	__u32 *elem;
+	int cnt;
+
+	for (cnt = 0; i < MAX_LEVELS; i++) {
+		__u64 cgrp_id = bpf_get_current_ancestor_cgroup_id(i);
+
+		if (cgrp_id == 0)
+			break;
+
+		// convert cgroup-id to a map index
+		elem = bpf_map_lookup_elem(&cgrp_idx, &cgrp_id);
+		if (!elem)
+			continue;
+
+		cgrps[cnt++] = *elem;
+		if (cnt == size)
+			break;
+	}
+
+	return cnt;
+}
+
+static int bperf_cgroup_count(void)
+{
+	register __u32 idx = 0;  // to have it in a register to pass BPF verifier
+	register int c = 0;
+	struct bpf_perf_event_value val, delta, *prev_val, *cgrp_val;
+	__u32 cpu = bpf_get_smp_processor_id();
+	__u32 cgrp_idx[MAX_LEVELS];
+	int cgrp_cnt;
+	__u32 key, cgrp;
+	long err;
+
+	if (use_cgroup_v2)
+		cgrp_cnt = get_cgroup_v2_idx(cgrp_idx, MAX_LEVELS);
+	else
+		cgrp_cnt = get_cgroup_v1_idx(cgrp_idx, MAX_LEVELS);
+
+	for ( ; idx < MAX_EVENTS; idx++) {
+		if (idx == num_events)
+			break;
+
+		// XXX: do not pass idx directly (for verifier)
+		key = idx;
+		// this is per-cpu array for diff
+		prev_val = bpf_map_lookup_elem(&prev_readings, &key);
+		if (!prev_val) {
+			val.counter = val.enabled = val.running = 0;
+			bpf_map_update_elem(&prev_readings, &key, &val, BPF_ANY);
+
+			prev_val = bpf_map_lookup_elem(&prev_readings, &key);
+			if (!prev_val)
+				continue;
+		}
+
+		// read from global perf_event array
+		key = idx * num_cpus + cpu;
+		err = bpf_perf_event_read_value(&events, key, &val, sizeof(val));
+		if (err)
+			continue;
+
+		if (enabled) {
+			delta.counter = val.counter - prev_val->counter;
+			delta.enabled = val.enabled - prev_val->enabled;
+			delta.running = val.running - prev_val->running;
+
+			for (c = 0; c < MAX_LEVELS; c++) {
+				if (c == cgrp_cnt)
+					break;
+
+				cgrp = cgrp_idx[c];
+
+				// aggregate the result by cgroup
+				key = cgrp * num_events + idx;
+				cgrp_val = bpf_map_lookup_elem(&cgrp_readings, &key);
+				if (cgrp_val) {
+					cgrp_val->counter += delta.counter;
+					cgrp_val->enabled += delta.enabled;
+					cgrp_val->running += delta.running;
+				} else {
+					bpf_map_update_elem(&cgrp_readings, &key,
+							    &delta, BPF_ANY);
+				}
+			}
+		}
+
+		*prev_val = val;
+	}
+	return 0;
+}
+
+// This will be attached to cgroup-switches event for each cpu
+SEC("perf_events")
+int BPF_PROG(on_cgrp_switch)
+{
+	return bperf_cgroup_count();
+}
+
+SEC("raw_tp/sched_switch")
+int BPF_PROG(trigger_read)
+{
+	return bperf_cgroup_count();
+}
+
+char LICENSE[] SEC("license") = "Dual BSD/GPL";

tools/perf/util/cgroup.c

···
 #include <regex.h>
 
 int nr_cgroups;
+bool cgrp_event_expanded;
 
 /* used to match cgroup name with patterns */
 struct cgroup_name {
···
 	}
 
 	ret = 0;
+	cgrp_event_expanded = true;
 
 out_err:
 	evlist__delete(orig_list);

tools/perf/util/cgroup.h

···
 };
 
 extern int nr_cgroups; /* number of explicit cgroups defined */
+extern bool cgrp_event_expanded;
 
 struct cgroup *cgroup__get(struct cgroup *cgroup);
 void cgroup__put(struct cgroup *cgroup);