cpufreq.c

// SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause)
// Copyright (c) 2020 Wenbo Zhang
//
// Based on cpufreq(8) from BPF-Perf-Tools-Book by Brendan Gregg.
// 10-OCT-2020   Wenbo Zhang   Created this.
#include <argp.h>
#include <signal.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <linux/perf_event.h>
#include <asm/unistd.h>
#include <fcntl.h>
#include <bpf/libbpf.h>
#include <bpf/bpf.h>
#include "cpufreq.h"
#include "cpufreq.skel.h"
#include "trace_helpers.h"

static struct env {
	int duration;
	int freq;
	bool verbose;
	char *cgroupspath;
	bool cg;
} env = {
	.duration = -1,
	.freq = 99,
};

const char *argp_program_version = "cpufreq 0.1";
const char *argp_program_bug_address =
	"https://github.com/iovisor/bcc/tree/master/libbpf-tools";
const char argp_program_doc[] =
"Sampling CPU freq system-wide & by process. Ctrl-C to end.\n"
"\n"
"USAGE: cpufreq [--help] [-d DURATION] [-f FREQUENCY] [-c CG]\n"
"\n"
"EXAMPLES:\n"
"    cpufreq         # sample CPU freq at 99HZ (default)\n"
"    cpufreq -d 5    # sample for 5 seconds only\n"
"    cpufreq -c CG   # Trace process under cgroupsPath CG\n"
"    cpufreq -f 199  # sample CPU freq at 199HZ\n";

static const struct argp_option opts[] = {
	{ "duration", 'd', "DURATION", 0, "Duration to sample in seconds", 0 },
	{ "frequency", 'f', "FREQUENCY", 0, "Sample with a certain frequency", 0 },
	{ "cgroup", 'c', "/sys/fs/cgroup/unified", 0, "Trace process in cgroup path", 0 },
	{ "verbose", 'v', NULL, 0, "Verbose debug output", 0 },
	{ NULL, 'h', NULL, OPTION_HIDDEN, "Show the full help", 0 },
	{},
};

static error_t parse_arg(int key, char *arg, struct argp_state *state)
{
	switch (key) {
	case 'h':
		argp_state_help(state, stderr, ARGP_HELP_STD_HELP);
		break;
	case 'v':
		env.verbose = true;
		break;
	case 'd':
		errno = 0;
		env.duration = strtol(arg, NULL, 10);
		if (errno || env.duration <= 0) {
			fprintf(stderr, "Invalid duration: %s\n", arg);
			argp_usage(state);
		}
		break;
	case 'c':
		env.cgroupspath = arg;
		env.cg = true;
		break;
	case 'f':
		errno = 0;
		env.freq = strtol(arg, NULL, 10);
		if (errno || env.freq <= 0) {
			fprintf(stderr, "Invalid freq (in HZ): %s\n", arg);
			argp_usage(state);
		}
		break;
	default:
		return ARGP_ERR_UNKNOWN;
	}
	return 0;
}

static int nr_cpus;

static int open_and_attach_perf_event(int freq, struct bpf_program *prog,
				struct bpf_link *links[])
{
	struct perf_event_attr attr = {
		.type = PERF_TYPE_SOFTWARE,
		.freq = 1,
		.sample_period = freq,
		.config = PERF_COUNT_SW_CPU_CLOCK,
	};
	int i, fd;

	for (i = 0; i < nr_cpus; i++) {
		fd = syscall(__NR_perf_event_open, &attr, -1, i, -1, 0);
		if (fd < 0) {
			/* Ignore CPU that is offline */
			if (errno == ENODEV)
				continue;
			fprintf(stderr, "failed to init perf sampling: %s\n",
				strerror(errno));
			return -1;
		}
		links[i] = bpf_program__attach_perf_event(prog, fd);
		if (!links[i]) {
			fprintf(stderr, "failed to attach perf event on cpu: %d\n", i);
			close(fd);
			return -1;
		}
	}

	return 0;
}

static int libbpf_print_fn(enum libbpf_print_level level, const char *format, va_list args)
{
	if (level == LIBBPF_DEBUG && !env.verbose)
		return 0;
	return vfprintf(stderr, format, args);
}

static void sig_handler(int sig)
{
}

static int init_freqs_mhz(__u32 *freqs_mhz, struct bpf_link *links[])
{
	char path[64];
	FILE *f;
	int i;

	for (i = 0; i < nr_cpus; i++) {
		if (!links[i]) {
			continue;
		}
		snprintf(path, sizeof(path),
			"/sys/devices/system/cpu/cpu%d/cpufreq/scaling_cur_freq",
			i);

		f = fopen(path, "r");
		if (!f) {
			fprintf(stderr, "failed to open '%s': %s\n", path,
				strerror(errno));
			return -1;
		}
		if (fscanf(f, "%u\n", &freqs_mhz[i]) != 1) {
			fprintf(stderr, "failed to parse '%s': %s\n", path,
				strerror(errno));
			fclose(f);
			return -1;
		}
		/*
		 * scaling_cur_freq is in kHz. To be handled with
		 * a small data size, it's converted in mHz.
		 */
		freqs_mhz[i] /= 1000;
		fclose(f);
	}

	return 0;
}

static void print_linear_hists(struct bpf_map *hists,
			struct cpufreq_bpf__bss *bss)
{
	struct hkey lookup_key = {}, next_key;
	int err, fd = bpf_map__fd(hists);
	struct hist hist;

	while (!bpf_map_get_next_key(fd, &lookup_key, &next_key)) {
		err = bpf_map_lookup_elem(fd, &next_key, &hist);
		if (err < 0) {
			fprintf(stderr, "failed to lookup hist: %d\n", err);
			return;
		}
		print_linear_hist(hist.slots, MAX_SLOTS, 0, HIST_STEP_SIZE,
				next_key.comm);
		printf("\n");
		lookup_key = next_key;
	}

	printf("\n");
	print_linear_hist(bss->syswide.slots, MAX_SLOTS, 0, HIST_STEP_SIZE,
			  "syswide");
}

int main(int argc, char **argv)
{
	static const struct argp argp = {
		.options = opts,
		.parser = parse_arg,
		.doc = argp_program_doc,
	};
	struct bpf_link *links[MAX_CPU_NR] = {};
	struct cpufreq_bpf *obj;
	int err, i;
	int idx, cg_map_fd;
	int cgfd = -1;

	err = argp_parse(&argp, argc, argv, 0, NULL, NULL);
	if (err)
		return err;

	libbpf_set_print(libbpf_print_fn);

	nr_cpus = libbpf_num_possible_cpus();
	if (nr_cpus < 0) {
		fprintf(stderr, "failed to get # of possible cpus: '%s'!\n",
			strerror(-nr_cpus));
		return 1;
	}
	if (nr_cpus > MAX_CPU_NR) {
		fprintf(stderr, "the number of cpu cores is too big, please "
			"increase MAX_CPU_NR's value and recompile");
		return 1;
	}

	obj = cpufreq_bpf__open_and_load();
	if (!obj) {
		fprintf(stderr, "failed to open and/or load BPF object\n");
		return 1;
	}

	if (!obj->bss) {
		fprintf(stderr, "Memory-mapping BPF maps is supported starting from Linux 5.7, please upgrade.\n");
		goto cleanup;
	}

	obj->bss->filter_cg = env.cg;

	/* update cgroup path fd to map */
	if (env.cg) {
		idx = 0;
		cg_map_fd = bpf_map__fd(obj->maps.cgroup_map);
		cgfd = open(env.cgroupspath, O_RDONLY);
		if (cgfd < 0) {
			fprintf(stderr, "Failed opening Cgroup path: %s", env.cgroupspath);
			goto cleanup;
		}
		if (bpf_map_update_elem(cg_map_fd, &idx, &cgfd, BPF_ANY)) {
			fprintf(stderr, "Failed adding target cgroup to map");
			goto cleanup;
		}
	}

	err = open_and_attach_perf_event(env.freq, obj->progs.do_sample, links);
	if (err)
		goto cleanup;
	err = init_freqs_mhz(obj->bss->freqs_mhz, links);
	if (err) {
		fprintf(stderr, "failed to init freqs\n");
		goto cleanup;
	}

	err = cpufreq_bpf__attach(obj);
	if (err) {
		fprintf(stderr, "failed to attach BPF programs\n");
		goto cleanup;
	}

	printf("Sampling CPU freq system-wide & by process. Ctrl-C to end.\n");

	signal(SIGINT, sig_handler);

	/*
	 * We'll get sleep interrupted when someone presses Ctrl-C (which will
	 * be "handled" with noop by sig_handler).
	 */
	sleep(env.duration);
	printf("\n");

	print_linear_hists(obj->maps.hists, obj->bss);

cleanup:
	for (i = 0; i < nr_cpus; i++)
		bpf_link__destroy(links[i]);
	cpufreq_bpf__destroy(obj);
	if (cgfd > 0)
		close(cgfd);

	return err != 0;
}