1/* eBPF example program:
2 *
3 * - Creates arraymap in kernel with 4 bytes keys and 8 byte values
4 *
5 * - Loads eBPF program
6 *
7 *   The eBPF program accesses the map passed in to store two pieces of
8 *   information. The number of invocations of the program, which maps
9 *   to the number of packets received, is stored to key 0. Key 1 is
10 *   incremented on each iteration by the number of bytes stored in
11 *   the skb.
12 *
13 * - Attaches the new program to a cgroup using BPF_PROG_ATTACH
14 *
15 * - Every second, reads map[0] and map[1] to see how many bytes and
16 *   packets were seen on any socket of tasks in the given cgroup.
17 */
18
19#define _GNU_SOURCE
20
21#include <stdio.h>
22#include <stdlib.h>
23#include <stddef.h>
24#include <string.h>
25#include <unistd.h>
26#include <assert.h>
27#include <errno.h>
28#include <fcntl.h>
29
30#include <linux/bpf.h>
31#include <bpf/bpf.h>
32
33#include "bpf_insn.h"
34
35enum {
36	MAP_KEY_PACKETS,
37	MAP_KEY_BYTES,
38};
39
40char bpf_log_buf[BPF_LOG_BUF_SIZE];
41
42static int prog_load(int map_fd, int verdict)
43{
44	struct bpf_insn prog[] = {
45		BPF_MOV64_REG(BPF_REG_6, BPF_REG_1), /* save r6 so it's not clobbered by BPF_CALL */
46
47		/* Count packets */
48		BPF_MOV64_IMM(BPF_REG_0, MAP_KEY_PACKETS), /* r0 = 0 */
49		BPF_STX_MEM(BPF_W, BPF_REG_10, BPF_REG_0, -4), /* *(u32 *)(fp - 4) = r0 */
50		BPF_MOV64_REG(BPF_REG_2, BPF_REG_10),
51		BPF_ALU64_IMM(BPF_ADD, BPF_REG_2, -4), /* r2 = fp - 4 */
52		BPF_LD_MAP_FD(BPF_REG_1, map_fd), /* load map fd to r1 */
53		BPF_RAW_INSN(BPF_JMP | BPF_CALL, 0, 0, 0, BPF_FUNC_map_lookup_elem),
54		BPF_JMP_IMM(BPF_JEQ, BPF_REG_0, 0, 2),
55		BPF_MOV64_IMM(BPF_REG_1, 1), /* r1 = 1 */
56		BPF_RAW_INSN(BPF_STX | BPF_XADD | BPF_DW, BPF_REG_0, BPF_REG_1, 0, 0), /* xadd r0 += r1 */
57
58		/* Count bytes */
59		BPF_MOV64_IMM(BPF_REG_0, MAP_KEY_BYTES), /* r0 = 1 */
60		BPF_STX_MEM(BPF_W, BPF_REG_10, BPF_REG_0, -4), /* *(u32 *)(fp - 4) = r0 */
61		BPF_MOV64_REG(BPF_REG_2, BPF_REG_10),
62		BPF_ALU64_IMM(BPF_ADD, BPF_REG_2, -4), /* r2 = fp - 4 */
63		BPF_LD_MAP_FD(BPF_REG_1, map_fd),
64		BPF_RAW_INSN(BPF_JMP | BPF_CALL, 0, 0, 0, BPF_FUNC_map_lookup_elem),
65		BPF_JMP_IMM(BPF_JEQ, BPF_REG_0, 0, 2),
66		BPF_LDX_MEM(BPF_W, BPF_REG_1, BPF_REG_6, offsetof(struct __sk_buff, len)), /* r1 = skb->len */
67		BPF_RAW_INSN(BPF_STX | BPF_XADD | BPF_DW, BPF_REG_0, BPF_REG_1, 0, 0), /* xadd r0 += r1 */
68
69		BPF_MOV64_IMM(BPF_REG_0, verdict), /* r0 = verdict */
70		BPF_EXIT_INSN(),
71	};
72	size_t insns_cnt = sizeof(prog) / sizeof(struct bpf_insn);
73
74	return bpf_load_program(BPF_PROG_TYPE_CGROUP_SKB,
75				prog, insns_cnt, "GPL", 0,
76				bpf_log_buf, BPF_LOG_BUF_SIZE);
77}
78
79static int usage(const char *argv0)
80{
81	printf("Usage: %s [-d] [-D] <cg-path> <egress|ingress>\n", argv0);
82	printf("	-d	Drop Traffic\n");
83	printf("	-D	Detach filter, and exit\n");
84	return EXIT_FAILURE;
85}
86
87static int attach_filter(int cg_fd, int type, int verdict)
88{
89	int prog_fd, map_fd, ret, key;
90	long long pkt_cnt, byte_cnt;
91
92	map_fd = bpf_create_map(BPF_MAP_TYPE_ARRAY,
93				sizeof(key), sizeof(byte_cnt),
94				256, 0);
95	if (map_fd < 0) {
96		printf("Failed to create map: '%s'\n", strerror(errno));
97		return EXIT_FAILURE;
98	}
99
100	prog_fd = prog_load(map_fd, verdict);
101	printf("Output from kernel verifier:\n%s\n-------\n", bpf_log_buf);
102
103	if (prog_fd < 0) {
104		printf("Failed to load prog: '%s'\n", strerror(errno));
105		return EXIT_FAILURE;
106	}
107
108	ret = bpf_prog_attach(prog_fd, cg_fd, type, 0);
109	if (ret < 0) {
110		printf("Failed to attach prog to cgroup: '%s'\n",
111		       strerror(errno));
112		return EXIT_FAILURE;
113	}
114	while (1) {
115		key = MAP_KEY_PACKETS;
116		assert(bpf_map_lookup_elem(map_fd, &key, &pkt_cnt) == 0);
117
118		key = MAP_KEY_BYTES;
119		assert(bpf_map_lookup_elem(map_fd, &key, &byte_cnt) == 0);
120
121		printf("cgroup received %lld packets, %lld bytes\n",
122		       pkt_cnt, byte_cnt);
123		sleep(1);
124	}
125
126	return EXIT_SUCCESS;
127}
128
129int main(int argc, char **argv)
130{
131	int detach_only = 0, verdict = 1;
132	enum bpf_attach_type type;
133	int opt, cg_fd, ret;
134
135	while ((opt = getopt(argc, argv, "Dd")) != -1) {
136		switch (opt) {
137		case 'd':
138			verdict = 0;
139			break;
140		case 'D':
141			detach_only = 1;
142			break;
143		default:
144			return usage(argv[0]);
145		}
146	}
147
148	if (argc - optind < 2)
149		return usage(argv[0]);
150
151	if (strcmp(argv[optind + 1], "ingress") == 0)
152		type = BPF_CGROUP_INET_INGRESS;
153	else if (strcmp(argv[optind + 1], "egress") == 0)
154		type = BPF_CGROUP_INET_EGRESS;
155	else
156		return usage(argv[0]);
157
158	cg_fd = open(argv[optind], O_DIRECTORY | O_RDONLY);
159	if (cg_fd < 0) {
160		printf("Failed to open cgroup path: '%s'\n", strerror(errno));
161		return EXIT_FAILURE;
162	}
163
164	if (detach_only) {
165		ret = bpf_prog_detach(cg_fd, type);
166		printf("bpf_prog_detach() returned '%s' (%d)\n",
167		       strerror(errno), errno);
168	} else
169		ret = attach_filter(cg_fd, type, verdict);
170
171	return ret;
172}
173