libbpf-tools: add dcstat for monitoring disk cache statistics.

libbpf-tools/.gitignore

@@ -14,6 +14,7 @@
 /capable
 /cpudist
 /cpufreq
+/dcstat
 /drsnoop
 /execsnoop
 /exitsnoop

libbpf-tools/Makefile

@@ -51,6 +51,7 @@ APPS = \
 	capable \
 	cpudist \
 	cpufreq \
+	dcstat \
 	drsnoop \
 	execsnoop \
 	exitsnoop \

libbpf-tools/dcstat.bpf.c

@@ -0,0 +1,36 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright 2024 Sony Group Corporation
+//
+// Based on dcstat(8) from BCC by Brendan Gregg
+
+#include <vmlinux.h>
+#include <bpf/bpf_helpers.h>
+#include <bpf/bpf_tracing.h>
+#include "dcstat.h"
+
+__u64 stats[S_MAXSTAT] = {};
+
+static __always_inline int inc_stats(int key)
+{
+	__atomic_add_fetch(&stats[key], 1, __ATOMIC_RELAXED);
+	return 0;
+}
+
+SEC("kprobe/lookup_fast")
+int BPF_KPROBE(lookup_fast)
+{
+	return inc_stats(S_REFS);
+}
+
+SEC("kretprobe/d_lookup")
+int BPF_KRETPROBE(d_lookup_ret)
+{
+	inc_stats(S_SLOW);
+	if (PT_REGS_RC(ctx) == 0) {
+		inc_stats(S_MISS);
+	}
+
+	return 0;
+}
+
+char LICENSE[] SEC("license") = "GPL";

libbpf-tools/dcstat.c

@@ -0,0 +1,216 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright 2024 Sony Group Corporation
+//
+// Based on dcstat(8) from BCC by Brendan Gregg
+
+#include <argp.h>
+#include <unistd.h>
+#include <time.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <bpf/bpf.h>
+#include "dcstat.h"
+#include "dcstat.skel.h"
+#include "btf_helpers.h"
+#include "trace_helpers.h"
+
+const char *argp_program_version = "dcstat 0.1";
+const char *argp_program_bug_address =
+	"https://github.com/iovisor/bcc/tree/master/libbpf-tools";
+static const char argp_program_doc[] =
+	"\ndcstat: Count various disk cache statistics\n"
+	"\n"
+	"EXAMPLES:\n"
+	"    dcstat      # interval one second\n"
+	"    dcstat 5 3  # interval five seconds, three output lines\n";
+static char args_doc[] = "[interval [count]]";
+
+static const struct argp_option opts[] =
+{
+	{ "verbose", 'v', NULL, 0, "Verbose debug output" },
+	{ NULL, 'h', NULL, OPTION_HIDDEN, "Show the full help" },
+	{},
+};
+
+static struct env
+{
+	bool verbose;
+	int count;
+	int interval;
+} env = {
+	.interval = 1,
+};
+
+static error_t parse_arg(int key, char *arg, struct argp_state *state)
+{
+	long interval;
+	long count;
+
+	switch (key) {
+	case 'h':
+		argp_state_help(state, stderr, ARGP_HELP_STD_HELP);
+		break;
+	case 'v':
+		env.verbose = true;
+		break;
+	case ARGP_KEY_ARG:
+		switch (state->arg_num) {
+		case 0:
+			errno = 0;
+			interval = strtol(arg, NULL, 10);
+			if (errno || interval <= 0 || interval > INT_MAX) {
+				fprintf(stderr, "invalid interval: %s\n", arg);
+				argp_usage(state);
+			}
+			env.interval = interval;
+			break;
+		case 1:
+			errno = 0;
+			count = strtol(arg, NULL, 10);
+			if (errno || count < 0 || count > INT_MAX) {
+				fprintf(stderr, "invalid count: %s\n", arg);
+				argp_usage(state);
+			}
+			env.count = count;
+			break;
+		default:
+			argp_usage(state);
+			break;
+		}
+		break;
+	default:
+		return ARGP_ERR_UNKNOWN;
+	}
+	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 const char *strftime_now(char *s, size_t max, const char *format)
+{
+	struct tm *tm;
+	time_t t;
+
+	t = time(NULL);
+	tm = localtime(&t);
+	if (tm == NULL) {
+		fprintf(stderr, "localtime: %s\n", strerror(errno));
+		return "<failed>";
+	}
+
+	if (strftime(s, max, format, tm) == 0) {
+		fprintf(stderr, "strftime error\n");
+		return "<failed>";
+	}
+
+	return s;
+}
+
+static const char *stat_types_names[] =
+{
+	[S_REFS] = "REFS",
+	[S_SLOW] = "SLOW",
+	[S_MISS] = "MISS",
+};
+
+void print_header(void)
+{
+	int i;
+
+	printf("%-8s  ", "TIME");
+	for (i = 0; i < S_MAXSTAT; i++) {
+		printf("%8s/s", stat_types_names[i]);
+	}
+	printf("  %-8s\n", "HITS");
+}
+
+static void print_and_reset_stats(__u64 stats[S_MAXSTAT])
+{
+	char s[16];
+	__u64 val, ref, miss, hit;
+	double pct = 0;
+	int i;
+
+	printf("%-8s: ", strftime_now(s, sizeof(s), "%H:%M:%S"));
+	for (i = 0; i < S_MAXSTAT; i++) {
+		val = __atomic_exchange_n(&stats[i], 0, __ATOMIC_RELAXED);
+		if (i == 0)
+			ref = val;
+		if (i == 2)
+			miss = val;
+		printf(" %8llu", val / env.interval);
+	}
+	hit = ref - miss;
+	pct = 100.0 * hit / ref;
+	printf(" %8.2f\n", pct);
+}
+
+int main(int argc, char **argv)
+{
+	LIBBPF_OPTS(bpf_object_open_opts, open_opts);
+	static const struct argp argp = {
+		.options = opts,
+		.parser = parse_arg,
+		.doc = argp_program_doc,
+		.args_doc = args_doc,
+	};
+	struct dcstat_bpf *skel;
+	int err;
+
+	err = argp_parse(&argp, argc, argv, 0, NULL, NULL);
+	if (err)
+		return err;
+
+	libbpf_set_print(libbpf_print_fn);
+
+	err = ensure_core_btf(&open_opts);
+	if (err) {
+		fprintf(stderr, "failed to fetch necessary BTF for CO-RE: %s\n",
+			strerror(-err));
+		return 1;
+	}
+
+	skel = dcstat_bpf__open();
+	if (!skel) {
+		fprintf(stderr, "failed to open BPF skelect\n");
+		return 1;
+	}
+
+	err = dcstat_bpf__load(skel);
+	if (err) {
+		fprintf(stderr, "failed to load BPF skelect: %d\n", err);
+		goto cleanup;
+	}
+
+	if (!skel->bss) {
+		fprintf(stderr, "Memory-mapping BPF maps is supported starting from Linux "
+			"5.7, please upgrade.\n");
+		goto cleanup;
+	}
+
+	err = dcstat_bpf__attach(skel);
+	if (err) {
+		fprintf(stderr, "failed to attach BPF programs: %s\n",
+		strerror(-err));
+		goto cleanup;
+	}
+
+	print_header();
+	do {
+		sleep(env.interval);
+		print_and_reset_stats(skel->bss->stats);
+	} while (!env.count || --env.count);
+
+cleanup:
+	dcstat_bpf__destroy(skel);
+	cleanup_core_btf(&open_opts);
+
+	return err != 0;
+}

libbpf-tools/dcstat.h

@@ -0,0 +1,14 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright 2024 Sony Group Corporation
+
+#ifndef __DCSTAT_H
+#define __DCSTAT_H
+
+enum stats {
+	S_REFS,
+	S_SLOW,
+	S_MISS,
+	S_MAXSTAT
+};
+
+#endif /* __DCSTAT_H */