ccid2.c 21.5 KB
Newer Older
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
/*
 *  net/dccp/ccids/ccid2.c
 *
 *  Copyright (c) 2005, 2006 Andrea Bittau <a.bittau@cs.ucl.ac.uk>
 *
 *  Changes to meet Linux coding standards, and DCCP infrastructure fixes.
 *
 *  Copyright (c) 2006 Arnaldo Carvalho de Melo <acme@conectiva.com.br>
 *
 *  This program is free software; you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License as published by
 *  the Free Software Foundation; either version 2 of the License, or
 *  (at your option) any later version.
 *
 *  This program is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *  GNU General Public License for more details.
 *
 *  You should have received a copy of the GNU General Public License
 *  along with this program; if not, write to the Free Software
 *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 */

/*
26
 * This implementation should follow RFC 4341
27 28 29 30 31 32 33
 */

#include "../ccid.h"
#include "../dccp.h"
#include "ccid2.h"


34
#ifdef CONFIG_IP_DCCP_CCID2_DEBUG
35 36
static int ccid2_debug;
#define ccid2_pr_debug(format, a...)	DCCP_PR_DEBUG(ccid2_debug, format, ##a)
37 38 39 40 41

static void ccid2_hc_tx_check_sanity(const struct ccid2_hc_tx_sock *hctx)
{
	int len = 0;
	int pipe = 0;
42
	struct ccid2_seq *seqp = hctx->ccid2hctx_seqh;
43 44 45 46 47 48 49 50 51

	/* there is data in the chain */
	if (seqp != hctx->ccid2hctx_seqt) {
		seqp = seqp->ccid2s_prev;
		len++;
		if (!seqp->ccid2s_acked)
			pipe++;

		while (seqp != hctx->ccid2hctx_seqt) {
52
			struct ccid2_seq *prev = seqp->ccid2s_prev;
53 54 55 56 57 58

			len++;
			if (!prev->ccid2s_acked)
				pipe++;

			/* packets are sent sequentially */
59 60
			BUG_ON(dccp_delta_seqno(seqp->ccid2s_seq,
						prev->ccid2s_seq ) >= 0);
61 62
			BUG_ON(time_before(seqp->ccid2s_sent,
					   prev->ccid2s_sent));
63 64 65 66 67 68 69 70 71 72 73

			seqp = prev;
		}
	}

	BUG_ON(pipe != hctx->ccid2hctx_pipe);
	ccid2_pr_debug("len of chain=%d\n", len);

	do {
		seqp = seqp->ccid2s_prev;
		len++;
74
	} while (seqp != hctx->ccid2hctx_seqh);
75 76

	ccid2_pr_debug("total len=%d\n", len);
77
	BUG_ON(len != hctx->ccid2hctx_seqbufc * CCID2_SEQBUF_LEN);
78 79
}
#else
80 81
#define ccid2_pr_debug(format, a...)
#define ccid2_hc_tx_check_sanity(hctx)
82 83
#endif

G
Gerrit Renker 已提交
84
static int ccid2_hc_tx_alloc_seq(struct ccid2_hc_tx_sock *hctx)
85 86 87 88 89 90 91 92 93 94
{
	struct ccid2_seq *seqp;
	int i;

	/* check if we have space to preserve the pointer to the buffer */
	if (hctx->ccid2hctx_seqbufc >= (sizeof(hctx->ccid2hctx_seqbuf) /
					sizeof(struct ccid2_seq*)))
		return -ENOMEM;

	/* allocate buffer and initialize linked list */
G
Gerrit Renker 已提交
95
	seqp = kmalloc(CCID2_SEQBUF_LEN * sizeof(struct ccid2_seq), gfp_any());
96 97 98
	if (seqp == NULL)
		return -ENOMEM;

G
Gerrit Renker 已提交
99
	for (i = 0; i < (CCID2_SEQBUF_LEN - 1); i++) {
100 101 102
		seqp[i].ccid2s_next = &seqp[i + 1];
		seqp[i + 1].ccid2s_prev = &seqp[i];
	}
G
Gerrit Renker 已提交
103 104
	seqp[CCID2_SEQBUF_LEN - 1].ccid2s_next = seqp;
	seqp->ccid2s_prev = &seqp[CCID2_SEQBUF_LEN - 1];
105 106 107 108 109 110 111 112 113

	/* This is the first allocation.  Initiate the head and tail.  */
	if (hctx->ccid2hctx_seqbufc == 0)
		hctx->ccid2hctx_seqh = hctx->ccid2hctx_seqt = seqp;
	else {
		/* link the existing list with the one we just created */
		hctx->ccid2hctx_seqh->ccid2s_next = seqp;
		seqp->ccid2s_prev = hctx->ccid2hctx_seqh;

G
Gerrit Renker 已提交
114 115
		hctx->ccid2hctx_seqt->ccid2s_prev = &seqp[CCID2_SEQBUF_LEN - 1];
		seqp[CCID2_SEQBUF_LEN - 1].ccid2s_next = hctx->ccid2hctx_seqt;
116 117 118 119 120 121 122 123 124
	}

	/* store the original pointer to the buffer so we can free it */
	hctx->ccid2hctx_seqbuf[hctx->ccid2hctx_seqbufc] = seqp;
	hctx->ccid2hctx_seqbufc++;

	return 0;
}

125
static int ccid2_hc_tx_send_packet(struct sock *sk, struct sk_buff *skb)
126
{
127
	struct ccid2_hc_tx_sock *hctx = ccid2_hc_tx_sk(sk);
128 129 130 131 132 133 134 135 136 137 138 139

	ccid2_pr_debug("pipe=%d cwnd=%d\n", hctx->ccid2hctx_pipe,
		       hctx->ccid2hctx_cwnd);

	if (hctx->ccid2hctx_pipe < hctx->ccid2hctx_cwnd) {
		/* OK we can send... make sure previous packet was sent off */
		if (!hctx->ccid2hctx_sendwait) {
			hctx->ccid2hctx_sendwait = 1;
			return 0;
		}
	}

140
	return 1; /* XXX CCID should dequeue when ready instead of polling */
141 142
}

143
static void ccid2_change_l_ack_ratio(struct sock *sk, u32 val)
144 145
{
	struct dccp_sock *dp = dccp_sk(sk);
146 147
	u32 max_ratio = DIV_ROUND_UP(ccid2_hc_tx_sk(sk)->ccid2hctx_cwnd, 2);

148
	/*
149 150 151 152
	 * Ensure that Ack Ratio does not exceed ceil(cwnd/2), which is (2) from
	 * RFC 4341, 6.1.2. We ignore the statement that Ack Ratio 2 is always
	 * acceptable since this causes starvation/deadlock whenever cwnd < 2.
	 * The same problem arises when Ack Ratio is 0 (ie. Ack Ratio disabled).
153
	 */
154 155 156
	if (val == 0 || val > max_ratio) {
		DCCP_WARN("Limiting Ack Ratio (%u) to %u\n", val, max_ratio);
		val = max_ratio;
157
	}
158 159
	if (val > 0xFFFF)		/* RFC 4340, 11.3 */
		val = 0xFFFF;
160

161 162 163
	if (val == dp->dccps_l_ack_ratio)
		return;

164
	ccid2_pr_debug("changing local ack ratio to %u\n", val);
165 166 167
	dp->dccps_l_ack_ratio = val;
}

G
Gerrit Renker 已提交
168
static void ccid2_change_cwnd(struct ccid2_hc_tx_sock *hctx, u32 val)
169
{
G
Gerrit Renker 已提交
170 171
	hctx->ccid2hctx_cwnd = val? : 1;
	ccid2_pr_debug("changed cwnd to %u\n", hctx->ccid2hctx_cwnd);
172 173
}

174 175 176 177 178 179 180 181 182 183 184
static void ccid2_change_srtt(struct ccid2_hc_tx_sock *hctx, long val)
{
	ccid2_pr_debug("change SRTT to %ld\n", val);
	hctx->ccid2hctx_srtt = val;
}

static void ccid2_change_pipe(struct ccid2_hc_tx_sock *hctx, long val)
{
	hctx->ccid2hctx_pipe = val;
}

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
static void ccid2_start_rto_timer(struct sock *sk);

static void ccid2_hc_tx_rto_expire(unsigned long data)
{
	struct sock *sk = (struct sock *)data;
	struct ccid2_hc_tx_sock *hctx = ccid2_hc_tx_sk(sk);
	long s;

	bh_lock_sock(sk);
	if (sock_owned_by_user(sk)) {
		sk_reset_timer(sk, &hctx->ccid2hctx_rtotimer,
			       jiffies + HZ / 5);
		goto out;
	}

	ccid2_pr_debug("RTO_EXPIRE\n");

	ccid2_hc_tx_check_sanity(hctx);

	/* back-off timer */
	hctx->ccid2hctx_rto <<= 1;

	s = hctx->ccid2hctx_rto / HZ;
	if (s > 60)
		hctx->ccid2hctx_rto = 60 * HZ;

	ccid2_start_rto_timer(sk);

	/* adjust pipe, cwnd etc */
214
	ccid2_change_pipe(hctx, 0);
215 216 217
	hctx->ccid2hctx_ssthresh = hctx->ccid2hctx_cwnd >> 1;
	if (hctx->ccid2hctx_ssthresh < 2)
		hctx->ccid2hctx_ssthresh = 2;
218
	ccid2_change_cwnd(hctx, 1);
219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234

	/* clear state about stuff we sent */
	hctx->ccid2hctx_seqt	= hctx->ccid2hctx_seqh;
	hctx->ccid2hctx_ssacks	= 0;
	hctx->ccid2hctx_acks	= 0;
	hctx->ccid2hctx_sent	= 0;

	/* clear ack ratio state. */
	hctx->ccid2hctx_arsent	 = 0;
	hctx->ccid2hctx_ackloss  = 0;
	hctx->ccid2hctx_rpseq	 = 0;
	hctx->ccid2hctx_rpdupack = -1;
	ccid2_change_l_ack_ratio(sk, 1);
	ccid2_hc_tx_check_sanity(hctx);
out:
	bh_unlock_sock(sk);
235
	sock_put(sk);
236 237 238 239 240 241 242 243 244 245 246 247 248
}

static void ccid2_start_rto_timer(struct sock *sk)
{
	struct ccid2_hc_tx_sock *hctx = ccid2_hc_tx_sk(sk);

	ccid2_pr_debug("setting RTO timeout=%ld\n", hctx->ccid2hctx_rto);

	BUG_ON(timer_pending(&hctx->ccid2hctx_rtotimer));
	sk_reset_timer(sk, &hctx->ccid2hctx_rtotimer,
		       jiffies + hctx->ccid2hctx_rto);
}

249
static void ccid2_hc_tx_packet_sent(struct sock *sk, int more, unsigned int len)
250 251 252
{
	struct dccp_sock *dp = dccp_sk(sk);
	struct ccid2_hc_tx_sock *hctx = ccid2_hc_tx_sk(sk);
253
	struct ccid2_seq *next;
254 255 256 257 258 259
	u64 seq;

	ccid2_hc_tx_check_sanity(hctx);

	BUG_ON(!hctx->ccid2hctx_sendwait);
	hctx->ccid2hctx_sendwait = 0;
260
	ccid2_change_pipe(hctx, hctx->ccid2hctx_pipe + 1);
261 262 263 264 265 266 267 268 269 270 271 272 273
	BUG_ON(hctx->ccid2hctx_pipe < 0);

	/* There is an issue.  What if another packet is sent between
	 * packet_send() and packet_sent().  Then the sequence number would be
	 * wrong.
	 * -sorbo.
	 */
	seq = dp->dccps_gss;

	hctx->ccid2hctx_seqh->ccid2s_seq   = seq;
	hctx->ccid2hctx_seqh->ccid2s_acked = 0;
	hctx->ccid2hctx_seqh->ccid2s_sent  = jiffies;

274 275 276
	next = hctx->ccid2hctx_seqh->ccid2s_next;
	/* check if we need to alloc more space */
	if (next == hctx->ccid2hctx_seqt) {
G
Gerrit Renker 已提交
277 278 279 280 281
		if (ccid2_hc_tx_alloc_seq(hctx)) {
			DCCP_CRIT("packet history - out of memory!");
			/* FIXME: find a more graceful way to bail out */
			return;
		}
282 283
		next = hctx->ccid2hctx_seqh->ccid2s_next;
		BUG_ON(next == hctx->ccid2hctx_seqt);
284
	}
285 286 287 288
	hctx->ccid2hctx_seqh = next;

	ccid2_pr_debug("cwnd=%d pipe=%d\n", hctx->ccid2hctx_cwnd,
		       hctx->ccid2hctx_pipe);
289 290 291 292 293 294 295 296

	hctx->ccid2hctx_sent++;

	/* Ack Ratio.  Need to maintain a concept of how many windows we sent */
	hctx->ccid2hctx_arsent++;
	/* We had an ack loss in this window... */
	if (hctx->ccid2hctx_ackloss) {
		if (hctx->ccid2hctx_arsent >= hctx->ccid2hctx_cwnd) {
297 298
			hctx->ccid2hctx_arsent	= 0;
			hctx->ccid2hctx_ackloss	= 0;
299
		}
300 301
	} else {
		/* No acks lost up to now... */
302 303 304
		/* decrease ack ratio if enough packets were sent */
		if (dp->dccps_l_ack_ratio > 1) {
			/* XXX don't calculate denominator each time */
305 306
			int denom = dp->dccps_l_ack_ratio * dp->dccps_l_ack_ratio -
				    dp->dccps_l_ack_ratio;
307 308 309 310 311 312 313

			denom = hctx->ccid2hctx_cwnd * hctx->ccid2hctx_cwnd / denom;

			if (hctx->ccid2hctx_arsent >= denom) {
				ccid2_change_l_ack_ratio(sk, dp->dccps_l_ack_ratio - 1);
				hctx->ccid2hctx_arsent = 0;
			}
314 315
		} else {
			/* we can't increase ack ratio further [1] */
316 317 318 319 320
			hctx->ccid2hctx_arsent = 0; /* or maybe set it to cwnd*/
		}
	}

	/* setup RTO timer */
321
	if (!timer_pending(&hctx->ccid2hctx_rtotimer))
322
		ccid2_start_rto_timer(sk);
323

324
#ifdef CONFIG_IP_DCCP_CCID2_DEBUG
325
	ccid2_pr_debug("pipe=%d\n", hctx->ccid2hctx_pipe);
R
Randy Dunlap 已提交
326
	ccid2_pr_debug("Sent: seq=%llu\n", (unsigned long long)seq);
327 328 329 330 331
	do {
		struct ccid2_seq *seqp = hctx->ccid2hctx_seqt;

		while (seqp != hctx->ccid2hctx_seqh) {
			ccid2_pr_debug("out seq=%llu acked=%d time=%lu\n",
332
				       (unsigned long long)seqp->ccid2s_seq,
R
Randy Dunlap 已提交
333
				       seqp->ccid2s_acked, seqp->ccid2s_sent);
334 335
			seqp = seqp->ccid2s_next;
		}
336
	} while (0);
337 338 339 340 341 342 343 344 345 346 347 348
	ccid2_pr_debug("=========\n");
	ccid2_hc_tx_check_sanity(hctx);
#endif
}

/* XXX Lame code duplication!
 * returns -1 if none was found.
 * else returns the next offset to use in the function call.
 */
static int ccid2_ackvector(struct sock *sk, struct sk_buff *skb, int offset,
			   unsigned char **vec, unsigned char *veclen)
{
349 350 351 352 353 354 355
	const struct dccp_hdr *dh = dccp_hdr(skb);
	unsigned char *options = (unsigned char *)dh + dccp_hdr_len(skb);
	unsigned char *opt_ptr;
	const unsigned char *opt_end = (unsigned char *)dh +
					(dh->dccph_doff * 4);
	unsigned char opt, len;
	unsigned char *value;
356 357 358 359 360 361 362 363

	BUG_ON(offset < 0);
	options += offset;
	opt_ptr = options;
	if (opt_ptr >= opt_end)
		return -1;

	while (opt_ptr != opt_end) {
364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386
		opt   = *opt_ptr++;
		len   = 0;
		value = NULL;

		/* Check if this isn't a single byte option */
		if (opt > DCCPO_MAX_RESERVED) {
			if (opt_ptr == opt_end)
				goto out_invalid_option;

			len = *opt_ptr++;
			if (len < 3)
				goto out_invalid_option;
			/*
			 * Remove the type and len fields, leaving
			 * just the value size
			 */
			len     -= 2;
			value   = opt_ptr;
			opt_ptr += len;

			if (opt_ptr > opt_end)
				goto out_invalid_option;
		}
387 388 389 390 391 392 393 394 395 396 397 398 399

		switch (opt) {
		case DCCPO_ACK_VECTOR_0:
		case DCCPO_ACK_VECTOR_1:
			*vec	= value;
			*veclen = len;
			return offset + (opt_ptr - options);
		}
	}

	return -1;

out_invalid_option:
400
	DCCP_BUG("Invalid option - this should not happen (previous parsing)!");
401 402 403
	return -1;
}

404
static void ccid2_hc_tx_kill_rto_timer(struct sock *sk)
405
{
406 407 408 409
	struct ccid2_hc_tx_sock *hctx = ccid2_hc_tx_sk(sk);

	sk_stop_timer(sk, &hctx->ccid2hctx_rtotimer);
	ccid2_pr_debug("deleted RTO timer\n");
410 411 412
}

static inline void ccid2_new_ack(struct sock *sk,
413
				 struct ccid2_seq *seqp,
414 415 416 417 418 419 420 421 422 423 424 425 426
				 unsigned int *maxincr)
{
	struct ccid2_hc_tx_sock *hctx = ccid2_hc_tx_sk(sk);

	/* slow start */
	if (hctx->ccid2hctx_cwnd < hctx->ccid2hctx_ssthresh) {
		hctx->ccid2hctx_acks = 0;

		/* We can increase cwnd at most maxincr [ack_ratio/2] */
		if (*maxincr) {
			/* increase every 2 acks */
			hctx->ccid2hctx_ssacks++;
			if (hctx->ccid2hctx_ssacks == 2) {
427
				ccid2_change_cwnd(hctx, hctx->ccid2hctx_cwnd+1);
428 429 430
				hctx->ccid2hctx_ssacks = 0;
				*maxincr = *maxincr - 1;
			}
431 432
		} else {
			/* increased cwnd enough for this single ack */
433 434
			hctx->ccid2hctx_ssacks = 0;
		}
435
	} else {
436 437 438 439
		hctx->ccid2hctx_ssacks = 0;
		hctx->ccid2hctx_acks++;

		if (hctx->ccid2hctx_acks >= hctx->ccid2hctx_cwnd) {
440
			ccid2_change_cwnd(hctx, hctx->ccid2hctx_cwnd + 1);
441 442 443 444 445 446
			hctx->ccid2hctx_acks = 0;
		}
	}

	/* update RTO */
	if (hctx->ccid2hctx_srtt == -1 ||
447 448
	    time_after(jiffies, hctx->ccid2hctx_lastrtt + hctx->ccid2hctx_srtt)) {
		unsigned long r = (long)jiffies - (long)seqp->ccid2s_sent;
449 450 451 452 453
		int s;

		/* first measurement */
		if (hctx->ccid2hctx_srtt == -1) {
			ccid2_pr_debug("R: %lu Time=%lu seq=%llu\n",
454
				       r, jiffies,
R
Randy Dunlap 已提交
455
				       (unsigned long long)seqp->ccid2s_seq);
456
			ccid2_change_srtt(hctx, r);
457
			hctx->ccid2hctx_rttvar = r >> 1;
458
		} else {
459 460
			/* RTTVAR */
			long tmp = hctx->ccid2hctx_srtt - r;
461 462
			long srtt;

463 464 465 466 467 468 469 470 471
			if (tmp < 0)
				tmp *= -1;

			tmp >>= 2;
			hctx->ccid2hctx_rttvar *= 3;
			hctx->ccid2hctx_rttvar >>= 2;
			hctx->ccid2hctx_rttvar += tmp;

			/* SRTT */
472 473 474
			srtt = hctx->ccid2hctx_srtt;
			srtt *= 7;
			srtt >>= 3;
475
			tmp = r >> 3;
476 477
			srtt += tmp;
			ccid2_change_srtt(hctx, srtt);
478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498
		}
		s = hctx->ccid2hctx_rttvar << 2;
		/* clock granularity is 1 when based on jiffies */
		if (!s)
			s = 1;
		hctx->ccid2hctx_rto = hctx->ccid2hctx_srtt + s;

		/* must be at least a second */
		s = hctx->ccid2hctx_rto / HZ;
		/* DCCP doesn't require this [but I like it cuz my code sux] */
#if 1
		if (s < 1)
			hctx->ccid2hctx_rto = HZ;
#endif
		/* max 60 seconds */
		if (s > 60)
			hctx->ccid2hctx_rto = HZ * 60;

		hctx->ccid2hctx_lastrtt = jiffies;

		ccid2_pr_debug("srtt: %ld rttvar: %ld rto: %ld (HZ=%d) R=%lu\n",
499 500
			       hctx->ccid2hctx_srtt, hctx->ccid2hctx_rttvar,
			       hctx->ccid2hctx_rto, HZ, r);
501 502 503 504
		hctx->ccid2hctx_sent = 0;
	}

	/* we got a new ack, so re-start RTO timer */
505
	ccid2_hc_tx_kill_rto_timer(sk);
506 507 508
	ccid2_start_rto_timer(sk);
}

509
static void ccid2_hc_tx_dec_pipe(struct sock *sk)
510
{
511 512
	struct ccid2_hc_tx_sock *hctx = ccid2_hc_tx_sk(sk);

513
	ccid2_change_pipe(hctx, hctx->ccid2hctx_pipe-1);
514 515 516
	BUG_ON(hctx->ccid2hctx_pipe < 0);

	if (hctx->ccid2hctx_pipe == 0)
517
		ccid2_hc_tx_kill_rto_timer(sk);
518 519
}

520
static void ccid2_congestion_event(struct sock *sk, struct ccid2_seq *seqp)
521
{
522 523
	struct ccid2_hc_tx_sock *hctx = ccid2_hc_tx_sk(sk);

524 525 526 527 528 529 530 531 532 533 534
	if (time_before(seqp->ccid2s_sent, hctx->ccid2hctx_last_cong)) {
		ccid2_pr_debug("Multiple losses in an RTT---treating as one\n");
		return;
	}

	hctx->ccid2hctx_last_cong = jiffies;

	ccid2_change_cwnd(hctx, hctx->ccid2hctx_cwnd >> 1);
	hctx->ccid2hctx_ssthresh = hctx->ccid2hctx_cwnd;
	if (hctx->ccid2hctx_ssthresh < 2)
		hctx->ccid2hctx_ssthresh = 2;
535 536 537 538

	/* Avoid spurious timeouts resulting from Ack Ratio > cwnd */
	if (dccp_sk(sk)->dccps_l_ack_ratio > hctx->ccid2hctx_cwnd)
		ccid2_change_l_ack_ratio(sk, hctx->ccid2hctx_cwnd);
539 540
}

541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564
static void ccid2_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb)
{
	struct dccp_sock *dp = dccp_sk(sk);
	struct ccid2_hc_tx_sock *hctx = ccid2_hc_tx_sk(sk);
	u64 ackno, seqno;
	struct ccid2_seq *seqp;
	unsigned char *vector;
	unsigned char veclen;
	int offset = 0;
	int done = 0;
	unsigned int maxincr = 0;

	ccid2_hc_tx_check_sanity(hctx);
	/* check reverse path congestion */
	seqno = DCCP_SKB_CB(skb)->dccpd_seq;

	/* XXX this whole "algorithm" is broken.  Need to fix it to keep track
	 * of the seqnos of the dupacks so that rpseq and rpdupack are correct
	 * -sorbo.
	 */
	/* need to bootstrap */
	if (hctx->ccid2hctx_rpdupack == -1) {
		hctx->ccid2hctx_rpdupack = 0;
		hctx->ccid2hctx_rpseq = seqno;
565
	} else {
566
		/* check if packet is consecutive */
567 568
		if (dccp_delta_seqno(hctx->ccid2hctx_rpseq, seqno) == 1)
			hctx->ccid2hctx_rpseq = seqno;
569 570 571 572 573 574 575 576 577 578
		/* it's a later packet */
		else if (after48(seqno, hctx->ccid2hctx_rpseq)) {
			hctx->ccid2hctx_rpdupack++;

			/* check if we got enough dupacks */
			if (hctx->ccid2hctx_rpdupack >=
			    hctx->ccid2hctx_numdupack) {
				hctx->ccid2hctx_rpdupack = -1; /* XXX lame */
				hctx->ccid2hctx_rpseq = 0;

579
				ccid2_change_l_ack_ratio(sk, 2 * dp->dccps_l_ack_ratio);
580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597
			}
		}
	}

	/* check forward path congestion */
	/* still didn't send out new data packets */
	if (hctx->ccid2hctx_seqh == hctx->ccid2hctx_seqt)
		return;

	switch (DCCP_SKB_CB(skb)->dccpd_type) {
	case DCCP_PKT_ACK:
	case DCCP_PKT_DATAACK:
		break;
	default:
		return;
	}

	ackno = DCCP_SKB_CB(skb)->dccpd_ack_seq;
A
Andrea Bittau 已提交
598 599 600 601 602 603 604 605 606 607 608
	if (after48(ackno, hctx->ccid2hctx_high_ack))
		hctx->ccid2hctx_high_ack = ackno;

	seqp = hctx->ccid2hctx_seqt;
	while (before48(seqp->ccid2s_seq, ackno)) {
		seqp = seqp->ccid2s_next;
		if (seqp == hctx->ccid2hctx_seqh) {
			seqp = hctx->ccid2hctx_seqh->ccid2s_prev;
			break;
		}
	}
609 610 611 612 613 614 615 616 617 618 619 620 621 622

	/* If in slow-start, cwnd can increase at most Ack Ratio / 2 packets for
	 * this single ack.  I round up.
	 * -sorbo.
	 */
	maxincr = dp->dccps_l_ack_ratio >> 1;
	maxincr++;

	/* go through all ack vectors */
	while ((offset = ccid2_ackvector(sk, skb, offset,
					 &vector, &veclen)) != -1) {
		/* go through this ack vector */
		while (veclen--) {
			const u8 rl = *vector & DCCP_ACKVEC_LEN_MASK;
623
			u64 ackno_end_rl = SUB48(ackno, rl);
624

R
Randy Dunlap 已提交
625 626 627
			ccid2_pr_debug("ackvec start:%llu end:%llu\n",
				       (unsigned long long)ackno,
				       (unsigned long long)ackno_end_rl);
628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645
			/* if the seqno we are analyzing is larger than the
			 * current ackno, then move towards the tail of our
			 * seqnos.
			 */
			while (after48(seqp->ccid2s_seq, ackno)) {
				if (seqp == hctx->ccid2hctx_seqt) {
					done = 1;
					break;
				}
				seqp = seqp->ccid2s_prev;
			}
			if (done)
				break;

			/* check all seqnos in the range of the vector
			 * run length
			 */
			while (between48(seqp->ccid2s_seq,ackno_end_rl,ackno)) {
646 647
				const u8 state = *vector &
						 DCCP_ACKVEC_STATE_MASK;
648 649 650 651

				/* new packet received or marked */
				if (state != DCCP_ACKVEC_STATE_NOT_RECEIVED &&
				    !seqp->ccid2s_acked) {
652
					if (state ==
653
					    DCCP_ACKVEC_STATE_ECN_MARKED) {
654
						ccid2_congestion_event(sk,
655
								       seqp);
656
					} else
657 658 659 660 661
						ccid2_new_ack(sk, seqp,
							      &maxincr);

					seqp->ccid2s_acked = 1;
					ccid2_pr_debug("Got ack for %llu\n",
R
Randy Dunlap 已提交
662
						       (unsigned long long)seqp->ccid2s_seq);
663
					ccid2_hc_tx_dec_pipe(sk);
664 665 666 667 668
				}
				if (seqp == hctx->ccid2hctx_seqt) {
					done = 1;
					break;
				}
669
				seqp = seqp->ccid2s_prev;
670 671 672 673
			}
			if (done)
				break;

674
			ackno = SUB48(ackno_end_rl, 1);
675 676 677 678 679 680 681 682 683
			vector++;
		}
		if (done)
			break;
	}

	/* The state about what is acked should be correct now
	 * Check for NUMDUPACK
	 */
A
Andrea Bittau 已提交
684 685 686 687 688 689 690 691
	seqp = hctx->ccid2hctx_seqt;
	while (before48(seqp->ccid2s_seq, hctx->ccid2hctx_high_ack)) {
		seqp = seqp->ccid2s_next;
		if (seqp == hctx->ccid2hctx_seqh) {
			seqp = hctx->ccid2hctx_seqh->ccid2s_prev;
			break;
		}
	}
692 693 694 695
	done = 0;
	while (1) {
		if (seqp->ccid2s_acked) {
			done++;
696
			if (done == hctx->ccid2hctx_numdupack)
697 698
				break;
		}
699
		if (seqp == hctx->ccid2hctx_seqt)
700 701 702 703 704 705 706 707 708 709 710 711 712
			break;
		seqp = seqp->ccid2s_prev;
	}

	/* If there are at least 3 acknowledgements, anything unacknowledged
	 * below the last sequence number is considered lost
	 */
	if (done == hctx->ccid2hctx_numdupack) {
		struct ccid2_seq *last_acked = seqp;

		/* check for lost packets */
		while (1) {
			if (!seqp->ccid2s_acked) {
713
				ccid2_pr_debug("Packet lost: %llu\n",
R
Randy Dunlap 已提交
714
					       (unsigned long long)seqp->ccid2s_seq);
715 716 717 718
				/* XXX need to traverse from tail -> head in
				 * order to detect multiple congestion events in
				 * one ack vector.
				 */
719
				ccid2_congestion_event(sk, seqp);
720
				ccid2_hc_tx_dec_pipe(sk);
721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740
			}
			if (seqp == hctx->ccid2hctx_seqt)
				break;
			seqp = seqp->ccid2s_prev;
		}

		hctx->ccid2hctx_seqt = last_acked;
	}

	/* trim acked packets in tail */
	while (hctx->ccid2hctx_seqt != hctx->ccid2hctx_seqh) {
		if (!hctx->ccid2hctx_seqt->ccid2s_acked)
			break;

		hctx->ccid2hctx_seqt = hctx->ccid2hctx_seqt->ccid2s_next;
	}

	ccid2_hc_tx_check_sanity(hctx);
}

741
static int ccid2_hc_tx_init(struct ccid *ccid, struct sock *sk)
742
{
743
	struct ccid2_hc_tx_sock *hctx = ccid_priv(ccid);
744

745
	ccid2_change_cwnd(hctx, 1);
746 747 748 749 750
	/* Initialize ssthresh to infinity.  This means that we will exit the
	 * initial slow-start after the first packet loss.  This is what we
	 * want.
	 */
	hctx->ccid2hctx_ssthresh  = ~0;
751 752 753
	hctx->ccid2hctx_numdupack = 3;

	/* XXX init ~ to window size... */
G
Gerrit Renker 已提交
754
	if (ccid2_hc_tx_alloc_seq(hctx))
755
		return -ENOMEM;
756

757
	hctx->ccid2hctx_rto	 = 3 * HZ;
758
	ccid2_change_srtt(hctx, -1);
759 760
	hctx->ccid2hctx_rttvar	 = -1;
	hctx->ccid2hctx_rpdupack = -1;
761
	hctx->ccid2hctx_last_cong = jiffies;
762 763
	setup_timer(&hctx->ccid2hctx_rtotimer, ccid2_hc_tx_rto_expire,
			(unsigned long)sk);
764 765 766 767 768 769 770

	ccid2_hc_tx_check_sanity(hctx);
	return 0;
}

static void ccid2_hc_tx_exit(struct sock *sk)
{
771
	struct ccid2_hc_tx_sock *hctx = ccid2_hc_tx_sk(sk);
772
	int i;
773

774
	ccid2_hc_tx_kill_rto_timer(sk);
775 776 777 778

	for (i = 0; i < hctx->ccid2hctx_seqbufc; i++)
		kfree(hctx->ccid2hctx_seqbuf[i]);
	hctx->ccid2hctx_seqbufc = 0;
779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797
}

static void ccid2_hc_rx_packet_recv(struct sock *sk, struct sk_buff *skb)
{
	const struct dccp_sock *dp = dccp_sk(sk);
	struct ccid2_hc_rx_sock *hcrx = ccid2_hc_rx_sk(sk);

	switch (DCCP_SKB_CB(skb)->dccpd_type) {
	case DCCP_PKT_DATA:
	case DCCP_PKT_DATAACK:
		hcrx->ccid2hcrx_data++;
		if (hcrx->ccid2hcrx_data >= dp->dccps_r_ack_ratio) {
			dccp_send_ack(sk);
			hcrx->ccid2hcrx_data = 0;
		}
		break;
	}
}

798
static struct ccid_operations ccid2 = {
I
Ian McDonald 已提交
799
	.ccid_id		= DCCPC_CCID2,
800 801
	.ccid_name		= "ccid2",
	.ccid_owner		= THIS_MODULE,
802
	.ccid_hc_tx_obj_size	= sizeof(struct ccid2_hc_tx_sock),
803 804 805 806 807
	.ccid_hc_tx_init	= ccid2_hc_tx_init,
	.ccid_hc_tx_exit	= ccid2_hc_tx_exit,
	.ccid_hc_tx_send_packet	= ccid2_hc_tx_send_packet,
	.ccid_hc_tx_packet_sent	= ccid2_hc_tx_packet_sent,
	.ccid_hc_tx_packet_recv	= ccid2_hc_tx_packet_recv,
808
	.ccid_hc_rx_obj_size	= sizeof(struct ccid2_hc_rx_sock),
809 810 811
	.ccid_hc_rx_packet_recv	= ccid2_hc_rx_packet_recv,
};

812
#ifdef CONFIG_IP_DCCP_CCID2_DEBUG
813
module_param(ccid2_debug, bool, 0444);
814
MODULE_PARM_DESC(ccid2_debug, "Enable debug messages");
815
#endif
816 817 818 819 820 821 822 823 824 825 826 827 828 829

static __init int ccid2_module_init(void)
{
	return ccid_register(&ccid2);
}
module_init(ccid2_module_init);

static __exit void ccid2_module_exit(void)
{
	ccid_unregister(&ccid2);
}
module_exit(ccid2_module_exit);

MODULE_AUTHOR("Andrea Bittau <a.bittau@cs.ucl.ac.uk>");
830
MODULE_DESCRIPTION("DCCP TCP-Like (CCID2) CCID");
831 832
MODULE_LICENSE("GPL");
MODULE_ALIAS("net-dccp-ccid-2");