1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
|
// SPDX-License-Identifier: GPL-2.0
/* Copyright (c) 2019 Facebook */
#include <linux/bpf.h>
#include <netinet/in.h>
#include <stdbool.h>
#include <bpf/bpf_helpers.h>
#include <bpf/bpf_endian.h>
enum bpf_linum_array_idx {
EGRESS_LINUM_IDX,
INGRESS_LINUM_IDX,
READ_SK_DST_PORT_LINUM_IDX,
__NR_BPF_LINUM_ARRAY_IDX,
};
struct {
__uint(type, BPF_MAP_TYPE_ARRAY);
__uint(max_entries, __NR_BPF_LINUM_ARRAY_IDX);
__type(key, __u32);
__type(value, __u32);
} linum_map SEC(".maps");
struct bpf_spinlock_cnt {
struct bpf_spin_lock lock;
__u32 cnt;
};
struct {
__uint(type, BPF_MAP_TYPE_SK_STORAGE);
__uint(map_flags, BPF_F_NO_PREALLOC);
__type(key, int);
__type(value, struct bpf_spinlock_cnt);
} sk_pkt_out_cnt SEC(".maps");
struct {
__uint(type, BPF_MAP_TYPE_SK_STORAGE);
__uint(map_flags, BPF_F_NO_PREALLOC);
__type(key, int);
__type(value, struct bpf_spinlock_cnt);
} sk_pkt_out_cnt10 SEC(".maps");
struct tcp_sock {
__u32 lsndtime;
} __attribute__((preserve_access_index));
struct bpf_tcp_sock listen_tp = {};
struct sockaddr_in6 srv_sa6 = {};
struct bpf_tcp_sock cli_tp = {};
struct bpf_tcp_sock srv_tp = {};
struct bpf_sock listen_sk = {};
struct bpf_sock srv_sk = {};
struct bpf_sock cli_sk = {};
__u64 parent_cg_id = 0;
__u64 child_cg_id = 0;
__u64 lsndtime = 0;
static bool is_loopback6(__u32 *a6)
{
return !a6[0] && !a6[1] && !a6[2] && a6[3] == bpf_htonl(1);
}
static void skcpy(struct bpf_sock *dst,
const struct bpf_sock *src)
{
dst->bound_dev_if = src->bound_dev_if;
dst->family = src->family;
dst->type = src->type;
dst->protocol = src->protocol;
dst->mark = src->mark;
dst->priority = src->priority;
dst->src_ip4 = src->src_ip4;
dst->src_ip6[0] = src->src_ip6[0];
dst->src_ip6[1] = src->src_ip6[1];
dst->src_ip6[2] = src->src_ip6[2];
dst->src_ip6[3] = src->src_ip6[3];
dst->src_port = src->src_port;
dst->dst_ip4 = src->dst_ip4;
dst->dst_ip6[0] = src->dst_ip6[0];
dst->dst_ip6[1] = src->dst_ip6[1];
dst->dst_ip6[2] = src->dst_ip6[2];
dst->dst_ip6[3] = src->dst_ip6[3];
dst->dst_port = src->dst_port;
dst->state = src->state;
}
static void tpcpy(struct bpf_tcp_sock *dst,
const struct bpf_tcp_sock *src)
{
dst->snd_cwnd = src->snd_cwnd;
dst->srtt_us = src->srtt_us;
dst->rtt_min = src->rtt_min;
dst->snd_ssthresh = src->snd_ssthresh;
dst->rcv_nxt = src->rcv_nxt;
dst->snd_nxt = src->snd_nxt;
dst->snd_una = src->snd_una;
dst->mss_cache = src->mss_cache;
dst->ecn_flags = src->ecn_flags;
dst->rate_delivered = src->rate_delivered;
dst->rate_interval_us = src->rate_interval_us;
dst->packets_out = src->packets_out;
dst->retrans_out = src->retrans_out;
dst->total_retrans = src->total_retrans;
dst->segs_in = src->segs_in;
dst->data_segs_in = src->data_segs_in;
dst->segs_out = src->segs_out;
dst->data_segs_out = src->data_segs_out;
dst->lost_out = src->lost_out;
dst->sacked_out = src->sacked_out;
dst->bytes_received = src->bytes_received;
dst->bytes_acked = src->bytes_acked;
}
/* Always return CG_OK so that no pkt will be filtered out */
#define CG_OK 1
#define RET_LOG() ({ \
linum = __LINE__; \
bpf_map_update_elem(&linum_map, &linum_idx, &linum, BPF_ANY); \
return CG_OK; \
})
SEC("cgroup_skb/egress")
int egress_read_sock_fields(struct __sk_buff *skb)
{
struct bpf_spinlock_cnt cli_cnt_init = { .lock = {}, .cnt = 0xeB9F };
struct bpf_spinlock_cnt *pkt_out_cnt, *pkt_out_cnt10;
struct bpf_tcp_sock *tp, *tp_ret;
struct bpf_sock *sk, *sk_ret;
__u32 linum, linum_idx;
struct tcp_sock *ktp;
linum_idx = EGRESS_LINUM_IDX;
sk = skb->sk;
if (!sk)
RET_LOG();
/* Not testing the egress traffic or the listening socket,
* which are covered by the cgroup_skb/ingress test program.
*/
if (sk->family != AF_INET6 || !is_loopback6(sk->src_ip6) ||
sk->state == BPF_TCP_LISTEN)
return CG_OK;
if (sk->src_port == bpf_ntohs(srv_sa6.sin6_port)) {
/* Server socket */
sk_ret = &srv_sk;
tp_ret = &srv_tp;
} else if (sk->dst_port == srv_sa6.sin6_port) {
/* Client socket */
sk_ret = &cli_sk;
tp_ret = &cli_tp;
} else {
/* Not the testing egress traffic */
return CG_OK;
}
/* It must be a fullsock for cgroup_skb/egress prog */
sk = bpf_sk_fullsock(sk);
if (!sk)
RET_LOG();
/* Not the testing egress traffic */
if (sk->protocol != IPPROTO_TCP)
return CG_OK;
tp = bpf_tcp_sock(sk);
if (!tp)
RET_LOG();
skcpy(sk_ret, sk);
tpcpy(tp_ret, tp);
if (sk_ret == &srv_sk) {
ktp = bpf_skc_to_tcp_sock(sk);
if (!ktp)
RET_LOG();
lsndtime = ktp->lsndtime;
child_cg_id = bpf_sk_cgroup_id(ktp);
if (!child_cg_id)
RET_LOG();
parent_cg_id = bpf_sk_ancestor_cgroup_id(ktp, 2);
if (!parent_cg_id)
RET_LOG();
/* The userspace has created it for srv sk */
pkt_out_cnt = bpf_sk_storage_get(&sk_pkt_out_cnt, ktp, 0, 0);
pkt_out_cnt10 = bpf_sk_storage_get(&sk_pkt_out_cnt10, ktp,
0, 0);
} else {
pkt_out_cnt = bpf_sk_storage_get(&sk_pkt_out_cnt, sk,
&cli_cnt_init,
BPF_SK_STORAGE_GET_F_CREATE);
pkt_out_cnt10 = bpf_sk_storage_get(&sk_pkt_out_cnt10,
sk, &cli_cnt_init,
BPF_SK_STORAGE_GET_F_CREATE);
}
if (!pkt_out_cnt || !pkt_out_cnt10)
RET_LOG();
/* Even both cnt and cnt10 have lock defined in their BTF,
* intentionally one cnt takes lock while one does not
* as a test for the spinlock support in BPF_MAP_TYPE_SK_STORAGE.
*/
pkt_out_cnt->cnt += 1;
bpf_spin_lock(&pkt_out_cnt10->lock);
pkt_out_cnt10->cnt += 10;
bpf_spin_unlock(&pkt_out_cnt10->lock);
return CG_OK;
}
SEC("cgroup_skb/ingress")
int ingress_read_sock_fields(struct __sk_buff *skb)
{
struct bpf_tcp_sock *tp;
__u32 linum, linum_idx;
struct bpf_sock *sk;
linum_idx = INGRESS_LINUM_IDX;
sk = skb->sk;
if (!sk)
RET_LOG();
/* Not the testing ingress traffic to the server */
if (sk->family != AF_INET6 || !is_loopback6(sk->src_ip6) ||
sk->src_port != bpf_ntohs(srv_sa6.sin6_port))
return CG_OK;
/* Only interested in the listening socket */
if (sk->state != BPF_TCP_LISTEN)
return CG_OK;
/* It must be a fullsock for cgroup_skb/ingress prog */
sk = bpf_sk_fullsock(sk);
if (!sk)
RET_LOG();
tp = bpf_tcp_sock(sk);
if (!tp)
RET_LOG();
skcpy(&listen_sk, sk);
tpcpy(&listen_tp, tp);
return CG_OK;
}
/*
* NOTE: 4-byte load from bpf_sock at dst_port offset is quirky. It
* gets rewritten by the access converter to a 2-byte load for
* backward compatibility. Treating the load result as a be16 value
* makes the code portable across little- and big-endian platforms.
*/
static __noinline bool sk_dst_port__load_word(struct bpf_sock *sk)
{
__u32 *word = (__u32 *)&sk->dst_port;
return word[0] == bpf_htons(0xcafe);
}
static __noinline bool sk_dst_port__load_half(struct bpf_sock *sk)
{
__u16 *half;
asm volatile ("");
half = (__u16 *)&sk->dst_port;
return half[0] == bpf_htons(0xcafe);
}
static __noinline bool sk_dst_port__load_byte(struct bpf_sock *sk)
{
__u8 *byte = (__u8 *)&sk->dst_port;
return byte[0] == 0xca && byte[1] == 0xfe;
}
SEC("cgroup_skb/egress")
int read_sk_dst_port(struct __sk_buff *skb)
{
__u32 linum, linum_idx;
struct bpf_sock *sk;
linum_idx = READ_SK_DST_PORT_LINUM_IDX;
sk = skb->sk;
if (!sk)
RET_LOG();
/* Ignore everything but the SYN from the client socket */
if (sk->state != BPF_TCP_SYN_SENT)
return CG_OK;
if (!sk_dst_port__load_word(sk))
RET_LOG();
if (!sk_dst_port__load_half(sk))
RET_LOG();
if (!sk_dst_port__load_byte(sk))
RET_LOG();
return CG_OK;
}
char _license[] SEC("license") = "GPL";
|
