xref: /openbmc/linux/net/ipv4/tcp_illinois.c (revision afb46f79) 
1 /*
2  * TCP Illinois congestion control.
3  * Home page:
4  *	http://www.ews.uiuc.edu/~shaoliu/tcpillinois/index.html
5  *
6  * The algorithm is described in:
7  * "TCP-Illinois: A Loss and Delay-Based Congestion Control Algorithm
8  *  for High-Speed Networks"
9  * http://www.ifp.illinois.edu/~srikant/Papers/liubassri06perf.pdf
10  *
11  * Implemented from description in paper and ns-2 simulation.
12  * Copyright (C) 2007 Stephen Hemminger <shemminger@linux-foundation.org>
13  */
14 
15 #include <linux/module.h>
16 #include <linux/skbuff.h>
17 #include <linux/inet_diag.h>
18 #include <asm/div64.h>
19 #include <net/tcp.h>
20 
21 #define ALPHA_SHIFT	7
22 #define ALPHA_SCALE	(1u<<ALPHA_SHIFT)
23 #define ALPHA_MIN	((3*ALPHA_SCALE)/10)	/* ~0.3 */
24 #define ALPHA_MAX	(10*ALPHA_SCALE)	/* 10.0 */
25 #define ALPHA_BASE	ALPHA_SCALE		/* 1.0 */
26 #define RTT_MAX		(U32_MAX / ALPHA_MAX)	/* 3.3 secs */
27 
28 #define BETA_SHIFT	6
29 #define BETA_SCALE	(1u<<BETA_SHIFT)
30 #define BETA_MIN	(BETA_SCALE/8)		/* 0.125 */
31 #define BETA_MAX	(BETA_SCALE/2)		/* 0.5 */
32 #define BETA_BASE	BETA_MAX
33 
34 static int win_thresh __read_mostly = 15;
35 module_param(win_thresh, int, 0);
36 MODULE_PARM_DESC(win_thresh, "Window threshold for starting adaptive sizing");
37 
38 static int theta __read_mostly = 5;
39 module_param(theta, int, 0);
40 MODULE_PARM_DESC(theta, "# of fast RTT's before full growth");
41 
42 /* TCP Illinois Parameters */
43 struct illinois {
44 	u64	sum_rtt;	/* sum of rtt's measured within last rtt */
45 	u16	cnt_rtt;	/* # of rtts measured within last rtt */
46 	u32	base_rtt;	/* min of all rtt in usec */
47 	u32	max_rtt;	/* max of all rtt in usec */
48 	u32	end_seq;	/* right edge of current RTT */
49 	u32	alpha;		/* Additive increase */
50 	u32	beta;		/* Muliplicative decrease */
51 	u16	acked;		/* # packets acked by current ACK */
52 	u8	rtt_above;	/* average rtt has gone above threshold */
53 	u8	rtt_low;	/* # of rtts measurements below threshold */
54 };
55 
56 static void rtt_reset(struct sock *sk)
57 {
58 	struct tcp_sock *tp = tcp_sk(sk);
59 	struct illinois *ca = inet_csk_ca(sk);
60 
61 	ca->end_seq = tp->snd_nxt;
62 	ca->cnt_rtt = 0;
63 	ca->sum_rtt = 0;
64 
65 	/* TODO: age max_rtt? */
66 }
67 
68 static void tcp_illinois_init(struct sock *sk)
69 {
70 	struct illinois *ca = inet_csk_ca(sk);
71 
72 	ca->alpha = ALPHA_MAX;
73 	ca->beta = BETA_BASE;
74 	ca->base_rtt = 0x7fffffff;
75 	ca->max_rtt = 0;
76 
77 	ca->acked = 0;
78 	ca->rtt_low = 0;
79 	ca->rtt_above = 0;
80 
81 	rtt_reset(sk);
82 }
83 
84 /* Measure RTT for each ack. */
85 static void tcp_illinois_acked(struct sock *sk, u32 pkts_acked, s32 rtt)
86 {
87 	struct illinois *ca = inet_csk_ca(sk);
88 
89 	ca->acked = pkts_acked;
90 
91 	/* dup ack, no rtt sample */
92 	if (rtt < 0)
93 		return;
94 
95 	/* ignore bogus values, this prevents wraparound in alpha math */
96 	if (rtt > RTT_MAX)
97 		rtt = RTT_MAX;
98 
99 	/* keep track of minimum RTT seen so far */
100 	if (ca->base_rtt > rtt)
101 		ca->base_rtt = rtt;
102 
103 	/* and max */
104 	if (ca->max_rtt < rtt)
105 		ca->max_rtt = rtt;
106 
107 	++ca->cnt_rtt;
108 	ca->sum_rtt += rtt;
109 }
110 
111 /* Maximum queuing delay */
112 static inline u32 max_delay(const struct illinois *ca)
113 {
114 	return ca->max_rtt - ca->base_rtt;
115 }
116 
117 /* Average queuing delay */
118 static inline u32 avg_delay(const struct illinois *ca)
119 {
120 	u64 t = ca->sum_rtt;
121 
122 	do_div(t, ca->cnt_rtt);
123 	return t - ca->base_rtt;
124 }
125 
126 /*
127  * Compute value of alpha used for additive increase.
128  * If small window then use 1.0, equivalent to Reno.
129  *
130  * For larger windows, adjust based on average delay.
131  * A. If average delay is at minimum (we are uncongested),
132  *    then use large alpha (10.0) to increase faster.
133  * B. If average delay is at maximum (getting congested)
134  *    then use small alpha (0.3)
135  *
136  * The result is a convex window growth curve.
137  */
138 static u32 alpha(struct illinois *ca, u32 da, u32 dm)
139 {
140 	u32 d1 = dm / 100;	/* Low threshold */
141 
142 	if (da <= d1) {
143 		/* If never got out of low delay zone, then use max */
144 		if (!ca->rtt_above)
145 			return ALPHA_MAX;
146 
147 		/* Wait for 5 good RTT's before allowing alpha to go alpha max.
148 		 * This prevents one good RTT from causing sudden window increase.
149 		 */
150 		if (++ca->rtt_low < theta)
151 			return ca->alpha;
152 
153 		ca->rtt_low = 0;
154 		ca->rtt_above = 0;
155 		return ALPHA_MAX;
156 	}
157 
158 	ca->rtt_above = 1;
159 
160 	/*
161 	 * Based on:
162 	 *
163 	 *      (dm - d1) amin amax
164 	 * k1 = -------------------
165 	 *         amax - amin
166 	 *
167 	 *       (dm - d1) amin
168 	 * k2 = ----------------  - d1
169 	 *        amax - amin
170 	 *
171 	 *             k1
172 	 * alpha = ----------
173 	 *          k2 + da
174 	 */
175 
176 	dm -= d1;
177 	da -= d1;
178 	return (dm * ALPHA_MAX) /
179 		(dm + (da  * (ALPHA_MAX - ALPHA_MIN)) / ALPHA_MIN);
180 }
181 
182 /*
183  * Beta used for multiplicative decrease.
184  * For small window sizes returns same value as Reno (0.5)
185  *
186  * If delay is small (10% of max) then beta = 1/8
187  * If delay is up to 80% of max then beta = 1/2
188  * In between is a linear function
189  */
190 static u32 beta(u32 da, u32 dm)
191 {
192 	u32 d2, d3;
193 
194 	d2 = dm / 10;
195 	if (da <= d2)
196 		return BETA_MIN;
197 
198 	d3 = (8 * dm) / 10;
199 	if (da >= d3 || d3 <= d2)
200 		return BETA_MAX;
201 
202 	/*
203 	 * Based on:
204 	 *
205 	 *       bmin d3 - bmax d2
206 	 * k3 = -------------------
207 	 *         d3 - d2
208 	 *
209 	 *       bmax - bmin
210 	 * k4 = -------------
211 	 *         d3 - d2
212 	 *
213 	 * b = k3 + k4 da
214 	 */
215 	return (BETA_MIN * d3 - BETA_MAX * d2 + (BETA_MAX - BETA_MIN) * da)
216 		/ (d3 - d2);
217 }
218 
219 /* Update alpha and beta values once per RTT */
220 static void update_params(struct sock *sk)
221 {
222 	struct tcp_sock *tp = tcp_sk(sk);
223 	struct illinois *ca = inet_csk_ca(sk);
224 
225 	if (tp->snd_cwnd < win_thresh) {
226 		ca->alpha = ALPHA_BASE;
227 		ca->beta = BETA_BASE;
228 	} else if (ca->cnt_rtt > 0) {
229 		u32 dm = max_delay(ca);
230 		u32 da = avg_delay(ca);
231 
232 		ca->alpha = alpha(ca, da, dm);
233 		ca->beta = beta(da, dm);
234 	}
235 
236 	rtt_reset(sk);
237 }
238 
239 /*
240  * In case of loss, reset to default values
241  */
242 static void tcp_illinois_state(struct sock *sk, u8 new_state)
243 {
244 	struct illinois *ca = inet_csk_ca(sk);
245 
246 	if (new_state == TCP_CA_Loss) {
247 		ca->alpha = ALPHA_BASE;
248 		ca->beta = BETA_BASE;
249 		ca->rtt_low = 0;
250 		ca->rtt_above = 0;
251 		rtt_reset(sk);
252 	}
253 }
254 
255 /*
256  * Increase window in response to successful acknowledgment.
257  */
258 static void tcp_illinois_cong_avoid(struct sock *sk, u32 ack, u32 acked,
259 				    u32 in_flight)
260 {
261 	struct tcp_sock *tp = tcp_sk(sk);
262 	struct illinois *ca = inet_csk_ca(sk);
263 
264 	if (after(ack, ca->end_seq))
265 		update_params(sk);
266 
267 	/* RFC2861 only increase cwnd if fully utilized */
268 	if (!tcp_is_cwnd_limited(sk, in_flight))
269 		return;
270 
271 	/* In slow start */
272 	if (tp->snd_cwnd <= tp->snd_ssthresh)
273 		tcp_slow_start(tp, acked);
274 
275 	else {
276 		u32 delta;
277 
278 		/* snd_cwnd_cnt is # of packets since last cwnd increment */
279 		tp->snd_cwnd_cnt += ca->acked;
280 		ca->acked = 1;
281 
282 		/* This is close approximation of:
283 		 * tp->snd_cwnd += alpha/tp->snd_cwnd
284 		*/
285 		delta = (tp->snd_cwnd_cnt * ca->alpha) >> ALPHA_SHIFT;
286 		if (delta >= tp->snd_cwnd) {
287 			tp->snd_cwnd = min(tp->snd_cwnd + delta / tp->snd_cwnd,
288 					   (u32) tp->snd_cwnd_clamp);
289 			tp->snd_cwnd_cnt = 0;
290 		}
291 	}
292 }
293 
294 static u32 tcp_illinois_ssthresh(struct sock *sk)
295 {
296 	struct tcp_sock *tp = tcp_sk(sk);
297 	struct illinois *ca = inet_csk_ca(sk);
298 
299 	/* Multiplicative decrease */
300 	return max(tp->snd_cwnd - ((tp->snd_cwnd * ca->beta) >> BETA_SHIFT), 2U);
301 }
302 
303 
304 /* Extract info for Tcp socket info provided via netlink. */
305 static void tcp_illinois_info(struct sock *sk, u32 ext,
306 			      struct sk_buff *skb)
307 {
308 	const struct illinois *ca = inet_csk_ca(sk);
309 
310 	if (ext & (1 << (INET_DIAG_VEGASINFO - 1))) {
311 		struct tcpvegas_info info = {
312 			.tcpv_enabled = 1,
313 			.tcpv_rttcnt = ca->cnt_rtt,
314 			.tcpv_minrtt = ca->base_rtt,
315 		};
316 
317 		if (info.tcpv_rttcnt > 0) {
318 			u64 t = ca->sum_rtt;
319 
320 			do_div(t, info.tcpv_rttcnt);
321 			info.tcpv_rtt = t;
322 		}
323 		nla_put(skb, INET_DIAG_VEGASINFO, sizeof(info), &info);
324 	}
325 }
326 
327 static struct tcp_congestion_ops tcp_illinois __read_mostly = {
328 	.init		= tcp_illinois_init,
329 	.ssthresh	= tcp_illinois_ssthresh,
330 	.cong_avoid	= tcp_illinois_cong_avoid,
331 	.set_state	= tcp_illinois_state,
332 	.get_info	= tcp_illinois_info,
333 	.pkts_acked	= tcp_illinois_acked,
334 
335 	.owner		= THIS_MODULE,
336 	.name		= "illinois",
337 };
338 
339 static int __init tcp_illinois_register(void)
340 {
341 	BUILD_BUG_ON(sizeof(struct illinois) > ICSK_CA_PRIV_SIZE);
342 	return tcp_register_congestion_control(&tcp_illinois);
343 }
344 
345 static void __exit tcp_illinois_unregister(void)
346 {
347 	tcp_unregister_congestion_control(&tcp_illinois);
348 }
349 
350 module_init(tcp_illinois_register);
351 module_exit(tcp_illinois_unregister);
352 
353 MODULE_AUTHOR("Stephen Hemminger, Shao Liu");
354 MODULE_LICENSE("GPL");
355 MODULE_DESCRIPTION("TCP Illinois");
356 MODULE_VERSION("1.0");
357