summaryrefslogtreecommitdiffstats
path: root/net/ipv4/tcp_dctcp.c
blob: bb23bb5b387a0ca9f41b3b2bb008325831f2fe6b (plain)
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
// SPDX-License-Identifier: GPL-2.0-or-later
/* DataCenter TCP (DCTCP) congestion control.
 *
 * http://simula.stanford.edu/~alizade/Site/DCTCP.html
 *
 * This is an implementation of DCTCP over Reno, an enhancement to the
 * TCP congestion control algorithm designed for data centers. DCTCP
 * leverages Explicit Congestion Notification (ECN) in the network to
 * provide multi-bit feedback to the end hosts. DCTCP's goal is to meet
 * the following three data center transport requirements:
 *
 *  - High burst tolerance (incast due to partition/aggregate)
 *  - Low latency (short flows, queries)
 *  - High throughput (continuous data updates, large file transfers)
 *    with commodity shallow buffered switches
 *
 * The algorithm is described in detail in the following two papers:
 *
 * 1) Mohammad Alizadeh, Albert Greenberg, David A. Maltz, Jitendra Padhye,
 *    Parveen Patel, Balaji Prabhakar, Sudipta Sengupta, and Murari Sridharan:
 *      "Data Center TCP (DCTCP)", Data Center Networks session
 *      Proc. ACM SIGCOMM, New Delhi, 2010.
 *   http://simula.stanford.edu/~alizade/Site/DCTCP_files/dctcp-final.pdf
 *
 * 2) Mohammad Alizadeh, Adel Javanmard, and Balaji Prabhakar:
 *      "Analysis of DCTCP: Stability, Convergence, and Fairness"
 *      Proc. ACM SIGMETRICS, San Jose, 2011.
 *   http://simula.stanford.edu/~alizade/Site/DCTCP_files/dctcp_analysis-full.pdf
 *
 * Initial prototype from Abdul Kabbani, Masato Yasuda and Mohammad Alizadeh.
 *
 * Authors:
 *
 *	Daniel Borkmann <dborkman@redhat.com>
 *	Florian Westphal <fw@strlen.de>
 *	Glenn Judd <glenn.judd@morganstanley.com>
 */

#include <linux/btf.h>
#include <linux/btf_ids.h>
#include <linux/module.h>
#include <linux/mm.h>
#include <net/tcp.h>
#include <linux/inet_diag.h>
#include "tcp_dctcp.h"

#define DCTCP_MAX_ALPHA	1024U

struct dctcp {
	u32 old_delivered;
	u32 old_delivered_ce;
	u32 prior_rcv_nxt;
	u32 dctcp_alpha;
	u32 next_seq;
	u32 ce_state;
	u32 loss_cwnd;
	struct tcp_plb_state plb;
};

static unsigned int dctcp_shift_g __read_mostly = 4; /* g = 1/2^4 */
module_param(dctcp_shift_g, uint, 0644);
MODULE_PARM_DESC(dctcp_shift_g, "parameter g for updating dctcp_alpha");

static unsigned int dctcp_alpha_on_init __read_mostly = DCTCP_MAX_ALPHA;
module_param(dctcp_alpha_on_init, uint, 0644);
MODULE_PARM_DESC(dctcp_alpha_on_init, "parameter for initial alpha value");

static struct tcp_congestion_ops dctcp_reno;

static void dctcp_reset(const struct tcp_sock *tp, struct dctcp *ca)
{
	ca->next_seq = tp->snd_nxt;

	ca->old_delivered = tp->delivered;
	ca->old_delivered_ce = tp->delivered_ce;
}

__bpf_kfunc static void dctcp_init(struct sock *sk)
{
	const struct tcp_sock *tp = tcp_sk(sk);

	if ((tp->ecn_flags & TCP_ECN_OK) ||
	    (sk->sk_state == TCP_LISTEN ||
	     sk->sk_state == TCP_CLOSE)) {
		struct dctcp *ca = inet_csk_ca(sk);

		ca->prior_rcv_nxt = tp->rcv_nxt;

		ca->dctcp_alpha = min(dctcp_alpha_on_init, DCTCP_MAX_ALPHA);

		ca->loss_cwnd = 0;
		ca->ce_state = 0;

		dctcp_reset(tp, ca);
		tcp_plb_init(sk, &ca->plb);

		return;
	}

	/* No ECN support? Fall back to Reno. Also need to clear
	 * ECT from sk since it is set during 3WHS for DCTCP.
	 */
	inet_csk(sk)->icsk_ca_ops = &dctcp_reno;
	INET_ECN_dontxmit(sk);
}

__bpf_kfunc static u32 dctcp_ssthresh(struct sock *sk)
{
	struct dctcp *ca = inet_csk_ca(sk);
	struct tcp_sock *tp = tcp_sk(sk);

	ca->loss_cwnd = tcp_snd_cwnd(tp);
	return max(tcp_snd_cwnd(tp) - ((tcp_snd_cwnd(tp) * ca->dctcp_alpha) >> 11U), 2U);
}

__bpf_kfunc static void dctcp_update_alpha(struct sock *sk, u32 flags)
{
	const struct tcp_sock *tp = tcp_sk(sk);
	struct dctcp *ca = inet_csk_ca(sk);

	/* Expired RTT */
	if (!before(tp->snd_una, ca->next_seq)) {
		u32 delivered = tp->delivered - ca->old_delivered;
		u32 delivered_ce = tp->delivered_ce - ca->old_delivered_ce;
		u32 alpha = ca->dctcp_alpha;
		u32 ce_ratio = 0;

		if (delivered > 0) {
			/* dctcp_alpha keeps EWMA of fraction of ECN marked
			 * packets. Because of EWMA smoothing, PLB reaction can
			 * be slow so we use ce_ratio which is an instantaneous
			 * measure of congestion. ce_ratio is the fraction of
			 * ECN marked packets in the previous RTT.
			 */
			if (delivered_ce > 0)
				ce_ratio = (delivered_ce << TCP_PLB_SCALE) / delivered;
			tcp_plb_update_state(sk, &ca->plb, (int)ce_ratio);
			tcp_plb_check_rehash(sk, &ca->plb);
		}

		/* alpha = (1 - g) * alpha + g * F */

		alpha -= min_not_zero(alpha, alpha >> dctcp_shift_g);
		if (delivered_ce) {

			/* If dctcp_shift_g == 1, a 32bit value would overflow
			 * after 8 M packets.
			 */
			delivered_ce <<= (10 - dctcp_shift_g);
			delivered_ce /= max(1U, delivered);

			alpha = min(alpha + delivered_ce, DCTCP_MAX_ALPHA);
		}
		/* dctcp_alpha can be read from dctcp_get_info() without
		 * synchro, so we ask compiler to not use dctcp_alpha
		 * as a temporary variable in prior operations.
		 */
		WRITE_ONCE(ca->dctcp_alpha, alpha);
		dctcp_reset(tp, ca);
	}
}

static void dctcp_react_to_loss(struct sock *sk)
{
	struct dctcp *ca = inet_csk_ca(sk);
	struct tcp_sock *tp = tcp_sk(sk);

	ca->loss_cwnd = tcp_snd_cwnd(tp);
	tp->snd_ssthresh = max(tcp_snd_cwnd(tp) >> 1U, 2U);
}

__bpf_kfunc static void dctcp_state(struct sock *sk, u8 new_state)
{
	if (new_state == TCP_CA_Recovery &&
	    new_state != inet_csk(sk)->icsk_ca_state)
		dctcp_react_to_loss(sk);
	/* We handle RTO in dctcp_cwnd_event to ensure that we perform only
	 * one loss-adjustment per RTT.
	 */
}

__bpf_kfunc static void dctcp_cwnd_event(struct sock *sk, enum tcp_ca_event ev)
{
	struct dctcp *ca = inet_csk_ca(sk);

	switch (ev) {
	case CA_EVENT_ECN_IS_CE:
	case CA_EVENT_ECN_NO_CE:
		dctcp_ece_ack_update(sk, ev, &ca->prior_rcv_nxt, &ca->ce_state);
		break;
	case CA_EVENT_LOSS:
		tcp_plb_update_state_upon_rto(sk, &ca->plb);
		dctcp_react_to_loss(sk);
		break;
	case CA_EVENT_TX_START:
		tcp_plb_check_rehash(sk, &ca->plb); /* Maybe rehash when inflight is 0 */
		break;
	default:
		/* Don't care for the rest. */
		break;
	}
}

static size_t dctcp_get_info(struct sock *sk, u32 ext, int *attr,
			     union tcp_cc_info *info)
{
	const struct dctcp *ca = inet_csk_ca(sk);
	const struct tcp_sock *tp = tcp_sk(sk);

	/* Fill it also in case of VEGASINFO due to req struct limits.
	 * We can still correctly retrieve it later.
	 */
	if (ext & (1 << (INET_DIAG_DCTCPINFO - 1)) ||
	    ext & (1 << (INET_DIAG_VEGASINFO - 1))) {
		memset(&info->dctcp, 0, sizeof(info->dctcp));
		if (inet_csk(sk)->icsk_ca_ops != &dctcp_reno) {
			info->dctcp.dctcp_enabled = 1;
			info->dctcp.dctcp_ce_state = (u16) ca->ce_state;
			info->dctcp.dctcp_alpha = ca->dctcp_alpha;
			info->dctcp.dctcp_ab_ecn = tp->mss_cache *
						   (tp->delivered_ce - ca->old_delivered_ce);
			info->dctcp.dctcp_ab_tot = tp->mss_cache *
						   (tp->delivered - ca->old_delivered);
		}

		*attr = INET_DIAG_DCTCPINFO;
		return sizeof(info->dctcp);
	}
	return 0;
}

__bpf_kfunc static u32 dctcp_cwnd_undo(struct sock *sk)
{
	const struct dctcp *ca = inet_csk_ca(sk);
	struct tcp_sock *tp = tcp_sk(sk);

	return max(tcp_snd_cwnd(tp), ca->loss_cwnd);
}

static struct tcp_congestion_ops dctcp __read_mostly = {
	.init		= dctcp_init,
	.in_ack_event   = dctcp_update_alpha,
	.cwnd_event	= dctcp_cwnd_event,
	.ssthresh	= dctcp_ssthresh,
	.cong_avoid	= tcp_reno_cong_avoid,
	.undo_cwnd	= dctcp_cwnd_undo,
	.set_state	= dctcp_state,
	.get_info	= dctcp_get_info,
	.flags		= TCP_CONG_NEEDS_ECN,
	.owner		= THIS_MODULE,
	.name		= "dctcp",
};

static struct tcp_congestion_ops dctcp_reno __read_mostly = {
	.ssthresh	= tcp_reno_ssthresh,
	.cong_avoid	= tcp_reno_cong_avoid,
	.undo_cwnd	= tcp_reno_undo_cwnd,
	.get_info	= dctcp_get_info,
	.owner		= THIS_MODULE,
	.name		= "dctcp-reno",
};

BTF_SET8_START(tcp_dctcp_check_kfunc_ids)
#ifdef CONFIG_X86
#ifdef CONFIG_DYNAMIC_FTRACE
BTF_ID_FLAGS(func, dctcp_init)
BTF_ID_FLAGS(func, dctcp_update_alpha)
BTF_ID_FLAGS(func, dctcp_cwnd_event)
BTF_ID_FLAGS(func, dctcp_ssthresh)
BTF_ID_FLAGS(func, dctcp_cwnd_undo)
BTF_ID_FLAGS(func, dctcp_state)
#endif
#endif
BTF_SET8_END(tcp_dctcp_check_kfunc_ids)

static const struct btf_kfunc_id_set tcp_dctcp_kfunc_set = {
	.owner = THIS_MODULE,
	.set   = &tcp_dctcp_check_kfunc_ids,
};

static int __init dctcp_register(void)
{
	int ret;

	BUILD_BUG_ON(sizeof(struct dctcp) > ICSK_CA_PRIV_SIZE);

	ret = register_btf_kfunc_id_set(BPF_PROG_TYPE_STRUCT_OPS, &tcp_dctcp_kfunc_set);
	if (ret < 0)
		return ret;
	return tcp_register_congestion_control(&dctcp);
}

static void __exit dctcp_unregister(void)
{
	tcp_unregister_congestion_control(&dctcp);
}

module_init(dctcp_register);
module_exit(dctcp_unregister);

MODULE_AUTHOR("Daniel Borkmann <dborkman@redhat.com>");
MODULE_AUTHOR("Florian Westphal <fw@strlen.de>");
MODULE_AUTHOR("Glenn Judd <glenn.judd@morganstanley.com>");

MODULE_LICENSE("GPL v2");
MODULE_DESCRIPTION("DataCenter TCP (DCTCP)");