1 /*
2  *  Copyright (c) 2007   The University of Aberdeen, Scotland, UK
3  *  Copyright (c) 2005-7 The University of Waikato, Hamilton, New Zealand.
4  *
5  *  An implementation of the DCCP protocol
6  *
7  *  This code has been developed by the University of Waikato WAND
8  *  research group. For further information please see http://www.wand.net.nz/
9  *  or e-mail Ian McDonald - ian.mcdonald@jandi.co.nz
10  *
11  *  This code also uses code from Lulea University, rereleased as GPL by its
12  *  authors:
13  *  Copyright (c) 2003 Nils-Erik Mattsson, Joacim Haggmark, Magnus Erixzon
14  *
15  *  Changes to meet Linux coding standards, to make it meet latest ccid3 draft
16  *  and to make it work as a loadable module in the DCCP stack written by
17  *  Arnaldo Carvalho de Melo <acme@conectiva.com.br>.
18  *
19  *  Copyright (c) 2005 Arnaldo Carvalho de Melo <acme@conectiva.com.br>
20  *
21  *  This program is free software; you can redistribute it and/or modify
22  *  it under the terms of the GNU General Public License as published by
23  *  the Free Software Foundation; either version 2 of the License, or
24  *  (at your option) any later version.
25  *
26  *  This program is distributed in the hope that it will be useful,
27  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
28  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
29  *  GNU General Public License for more details.
30  *
31  *  You should have received a copy of the GNU General Public License
32  *  along with this program; if not, write to the Free Software
33  *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
34  */
35 
36 #include <linux/string.h>
37 #include <linux/slab.h>
38 #include "packet_history.h"
39 #include "../../dccp.h"
40 
41 /*
42  * Transmitter History Routines
43  */
44 static struct kmem_cache *tfrc_tx_hist_slab;
45 
46 int __init tfrc_tx_packet_history_init(void)
47 {
48 	tfrc_tx_hist_slab = kmem_cache_create("tfrc_tx_hist",
49 					      sizeof(struct tfrc_tx_hist_entry),
50 					      0, SLAB_HWCACHE_ALIGN, NULL);
51 	return tfrc_tx_hist_slab == NULL ? -ENOBUFS : 0;
52 }
53 
54 void tfrc_tx_packet_history_exit(void)
55 {
56 	if (tfrc_tx_hist_slab != NULL) {
57 		kmem_cache_destroy(tfrc_tx_hist_slab);
58 		tfrc_tx_hist_slab = NULL;
59 	}
60 }
61 
62 int tfrc_tx_hist_add(struct tfrc_tx_hist_entry **headp, u64 seqno)
63 {
64 	struct tfrc_tx_hist_entry *entry = kmem_cache_alloc(tfrc_tx_hist_slab, gfp_any());
65 
66 	if (entry == NULL)
67 		return -ENOBUFS;
68 	entry->seqno = seqno;
69 	entry->stamp = ktime_get_real();
70 	entry->next  = *headp;
71 	*headp	     = entry;
72 	return 0;
73 }
74 
75 void tfrc_tx_hist_purge(struct tfrc_tx_hist_entry **headp)
76 {
77 	struct tfrc_tx_hist_entry *head = *headp;
78 
79 	while (head != NULL) {
80 		struct tfrc_tx_hist_entry *next = head->next;
81 
82 		kmem_cache_free(tfrc_tx_hist_slab, head);
83 		head = next;
84 	}
85 
86 	*headp = NULL;
87 }
88 
89 /*
90  *	Receiver History Routines
91  */
92 static struct kmem_cache *tfrc_rx_hist_slab;
93 
94 int __init tfrc_rx_packet_history_init(void)
95 {
96 	tfrc_rx_hist_slab = kmem_cache_create("tfrc_rxh_cache",
97 					      sizeof(struct tfrc_rx_hist_entry),
98 					      0, SLAB_HWCACHE_ALIGN, NULL);
99 	return tfrc_rx_hist_slab == NULL ? -ENOBUFS : 0;
100 }
101 
102 void tfrc_rx_packet_history_exit(void)
103 {
104 	if (tfrc_rx_hist_slab != NULL) {
105 		kmem_cache_destroy(tfrc_rx_hist_slab);
106 		tfrc_rx_hist_slab = NULL;
107 	}
108 }
109 
110 static inline void tfrc_rx_hist_entry_from_skb(struct tfrc_rx_hist_entry *entry,
111 					       const struct sk_buff *skb,
112 					       const u64 ndp)
113 {
114 	const struct dccp_hdr *dh = dccp_hdr(skb);
115 
116 	entry->tfrchrx_seqno = DCCP_SKB_CB(skb)->dccpd_seq;
117 	entry->tfrchrx_ccval = dh->dccph_ccval;
118 	entry->tfrchrx_type  = dh->dccph_type;
119 	entry->tfrchrx_ndp   = ndp;
120 	entry->tfrchrx_tstamp = ktime_get_real();
121 }
122 
123 void tfrc_rx_hist_add_packet(struct tfrc_rx_hist *h,
124 			     const struct sk_buff *skb,
125 			     const u64 ndp)
126 {
127 	struct tfrc_rx_hist_entry *entry = tfrc_rx_hist_last_rcv(h);
128 
129 	tfrc_rx_hist_entry_from_skb(entry, skb, ndp);
130 }
131 
132 /* has the packet contained in skb been seen before? */
133 int tfrc_rx_hist_duplicate(struct tfrc_rx_hist *h, struct sk_buff *skb)
134 {
135 	const u64 seq = DCCP_SKB_CB(skb)->dccpd_seq;
136 	int i;
137 
138 	if (dccp_delta_seqno(tfrc_rx_hist_loss_prev(h)->tfrchrx_seqno, seq) <= 0)
139 		return 1;
140 
141 	for (i = 1; i <= h->loss_count; i++)
142 		if (tfrc_rx_hist_entry(h, i)->tfrchrx_seqno == seq)
143 			return 1;
144 
145 	return 0;
146 }
147 
148 static void tfrc_rx_hist_swap(struct tfrc_rx_hist *h, const u8 a, const u8 b)
149 {
150 	const u8 idx_a = tfrc_rx_hist_index(h, a),
151 		 idx_b = tfrc_rx_hist_index(h, b);
152 
153 	swap(h->ring[idx_a], h->ring[idx_b]);
154 }
155 
156 /*
157  * Private helper functions for loss detection.
158  *
159  * In the descriptions, `Si' refers to the sequence number of entry number i,
160  * whose NDP count is `Ni' (lower case is used for variables).
161  * Note: All __xxx_loss functions expect that a test against duplicates has been
162  *       performed already: the seqno of the skb must not be less than the seqno
163  *       of loss_prev; and it must not equal that of any valid history entry.
164  */
165 static void __do_track_loss(struct tfrc_rx_hist *h, struct sk_buff *skb, u64 n1)
166 {
167 	u64 s0 = tfrc_rx_hist_loss_prev(h)->tfrchrx_seqno,
168 	    s1 = DCCP_SKB_CB(skb)->dccpd_seq;
169 
170 	if (!dccp_loss_free(s0, s1, n1)) {	/* gap between S0 and S1 */
171 		h->loss_count = 1;
172 		tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 1), skb, n1);
173 	}
174 }
175 
176 static void __one_after_loss(struct tfrc_rx_hist *h, struct sk_buff *skb, u32 n2)
177 {
178 	u64 s0 = tfrc_rx_hist_loss_prev(h)->tfrchrx_seqno,
179 	    s1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_seqno,
180 	    s2 = DCCP_SKB_CB(skb)->dccpd_seq;
181 
182 	if (likely(dccp_delta_seqno(s1, s2) > 0)) {	/* S1  <  S2 */
183 		h->loss_count = 2;
184 		tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 2), skb, n2);
185 		return;
186 	}
187 
188 	/* S0  <  S2  <  S1 */
189 
190 	if (dccp_loss_free(s0, s2, n2)) {
191 		u64 n1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_ndp;
192 
193 		if (dccp_loss_free(s2, s1, n1)) {
194 			/* hole is filled: S0, S2, and S1 are consecutive */
195 			h->loss_count = 0;
196 			h->loss_start = tfrc_rx_hist_index(h, 1);
197 		} else
198 			/* gap between S2 and S1: just update loss_prev */
199 			tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_loss_prev(h), skb, n2);
200 
201 	} else {	/* gap between S0 and S2 */
202 		/*
203 		 * Reorder history to insert S2 between S0 and S1
204 		 */
205 		tfrc_rx_hist_swap(h, 0, 3);
206 		h->loss_start = tfrc_rx_hist_index(h, 3);
207 		tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 1), skb, n2);
208 		h->loss_count = 2;
209 	}
210 }
211 
212 /* return 1 if a new loss event has been identified */
213 static int __two_after_loss(struct tfrc_rx_hist *h, struct sk_buff *skb, u32 n3)
214 {
215 	u64 s0 = tfrc_rx_hist_loss_prev(h)->tfrchrx_seqno,
216 	    s1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_seqno,
217 	    s2 = tfrc_rx_hist_entry(h, 2)->tfrchrx_seqno,
218 	    s3 = DCCP_SKB_CB(skb)->dccpd_seq;
219 
220 	if (likely(dccp_delta_seqno(s2, s3) > 0)) {	/* S2  <  S3 */
221 		h->loss_count = 3;
222 		tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 3), skb, n3);
223 		return 1;
224 	}
225 
226 	/* S3  <  S2 */
227 
228 	if (dccp_delta_seqno(s1, s3) > 0) {		/* S1  <  S3  <  S2 */
229 		/*
230 		 * Reorder history to insert S3 between S1 and S2
231 		 */
232 		tfrc_rx_hist_swap(h, 2, 3);
233 		tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 2), skb, n3);
234 		h->loss_count = 3;
235 		return 1;
236 	}
237 
238 	/* S0  <  S3  <  S1 */
239 
240 	if (dccp_loss_free(s0, s3, n3)) {
241 		u64 n1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_ndp;
242 
243 		if (dccp_loss_free(s3, s1, n1)) {
244 			/* hole between S0 and S1 filled by S3 */
245 			u64 n2 = tfrc_rx_hist_entry(h, 2)->tfrchrx_ndp;
246 
247 			if (dccp_loss_free(s1, s2, n2)) {
248 				/* entire hole filled by S0, S3, S1, S2 */
249 				h->loss_start = tfrc_rx_hist_index(h, 2);
250 				h->loss_count = 0;
251 			} else {
252 				/* gap remains between S1 and S2 */
253 				h->loss_start = tfrc_rx_hist_index(h, 1);
254 				h->loss_count = 1;
255 			}
256 
257 		} else /* gap exists between S3 and S1, loss_count stays at 2 */
258 			tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_loss_prev(h), skb, n3);
259 
260 		return 0;
261 	}
262 
263 	/*
264 	 * The remaining case:  S0  <  S3  <  S1  <  S2;  gap between S0 and S3
265 	 * Reorder history to insert S3 between S0 and S1.
266 	 */
267 	tfrc_rx_hist_swap(h, 0, 3);
268 	h->loss_start = tfrc_rx_hist_index(h, 3);
269 	tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 1), skb, n3);
270 	h->loss_count = 3;
271 
272 	return 1;
273 }
274 
275 /* recycle RX history records to continue loss detection if necessary */
276 static void __three_after_loss(struct tfrc_rx_hist *h)
277 {
278 	/*
279 	 * At this stage we know already that there is a gap between S0 and S1
280 	 * (since S0 was the highest sequence number received before detecting
281 	 * the loss). To recycle the loss record, it is	thus only necessary to
282 	 * check for other possible gaps between S1/S2 and between S2/S3.
283 	 */
284 	u64 s1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_seqno,
285 	    s2 = tfrc_rx_hist_entry(h, 2)->tfrchrx_seqno,
286 	    s3 = tfrc_rx_hist_entry(h, 3)->tfrchrx_seqno;
287 	u64 n2 = tfrc_rx_hist_entry(h, 2)->tfrchrx_ndp,
288 	    n3 = tfrc_rx_hist_entry(h, 3)->tfrchrx_ndp;
289 
290 	if (dccp_loss_free(s1, s2, n2)) {
291 
292 		if (dccp_loss_free(s2, s3, n3)) {
293 			/* no gap between S2 and S3: entire hole is filled */
294 			h->loss_start = tfrc_rx_hist_index(h, 3);
295 			h->loss_count = 0;
296 		} else {
297 			/* gap between S2 and S3 */
298 			h->loss_start = tfrc_rx_hist_index(h, 2);
299 			h->loss_count = 1;
300 		}
301 
302 	} else {	/* gap between S1 and S2 */
303 		h->loss_start = tfrc_rx_hist_index(h, 1);
304 		h->loss_count = 2;
305 	}
306 }
307 
308 /**
309  *  tfrc_rx_handle_loss  -  Loss detection and further processing
310  *  @h:		    The non-empty RX history object
311  *  @lh:	    Loss Intervals database to update
312  *  @skb:	    Currently received packet
313  *  @ndp:	    The NDP count belonging to @skb
314  *  @calc_first_li: Caller-dependent computation of first loss interval in @lh
315  *  @sk:	    Used by @calc_first_li (see tfrc_lh_interval_add)
316  *
317  *  Chooses action according to pending loss, updates LI database when a new
318  *  loss was detected, and does required post-processing. Returns 1 when caller
319  *  should send feedback, 0 otherwise.
320  *  Since it also takes care of reordering during loss detection and updates the
321  *  records accordingly, the caller should not perform any more RX history
322  *  operations when loss_count is greater than 0 after calling this function.
323  */
324 int tfrc_rx_handle_loss(struct tfrc_rx_hist *h,
325 			struct tfrc_loss_hist *lh,
326 			struct sk_buff *skb, const u64 ndp,
327 			u32 (*calc_first_li)(struct sock *), struct sock *sk)
328 {
329 	int is_new_loss = 0;
330 
331 	if (h->loss_count == 0) {
332 		__do_track_loss(h, skb, ndp);
333 	} else if (h->loss_count == 1) {
334 		__one_after_loss(h, skb, ndp);
335 	} else if (h->loss_count != 2) {
336 		DCCP_BUG("invalid loss_count %d", h->loss_count);
337 	} else if (__two_after_loss(h, skb, ndp)) {
338 		/*
339 		 * Update Loss Interval database and recycle RX records
340 		 */
341 		is_new_loss = tfrc_lh_interval_add(lh, h, calc_first_li, sk);
342 		__three_after_loss(h);
343 	}
344 	return is_new_loss;
345 }
346 
347 int tfrc_rx_hist_alloc(struct tfrc_rx_hist *h)
348 {
349 	int i;
350 
351 	for (i = 0; i <= TFRC_NDUPACK; i++) {
352 		h->ring[i] = kmem_cache_alloc(tfrc_rx_hist_slab, GFP_ATOMIC);
353 		if (h->ring[i] == NULL)
354 			goto out_free;
355 	}
356 
357 	h->loss_count = h->loss_start = 0;
358 	return 0;
359 
360 out_free:
361 	while (i-- != 0) {
362 		kmem_cache_free(tfrc_rx_hist_slab, h->ring[i]);
363 		h->ring[i] = NULL;
364 	}
365 	return -ENOBUFS;
366 }
367 
368 void tfrc_rx_hist_purge(struct tfrc_rx_hist *h)
369 {
370 	int i;
371 
372 	for (i = 0; i <= TFRC_NDUPACK; ++i)
373 		if (h->ring[i] != NULL) {
374 			kmem_cache_free(tfrc_rx_hist_slab, h->ring[i]);
375 			h->ring[i] = NULL;
376 		}
377 }
378 
379 /**
380  * tfrc_rx_hist_rtt_last_s - reference entry to compute RTT samples against
381  */
382 static inline struct tfrc_rx_hist_entry *
383 			tfrc_rx_hist_rtt_last_s(const struct tfrc_rx_hist *h)
384 {
385 	return h->ring[0];
386 }
387 
388 /**
389  * tfrc_rx_hist_rtt_prev_s - previously suitable (wrt rtt_last_s) RTT-sampling entry
390  */
391 static inline struct tfrc_rx_hist_entry *
392 			tfrc_rx_hist_rtt_prev_s(const struct tfrc_rx_hist *h)
393 {
394 	return h->ring[h->rtt_sample_prev];
395 }
396 
397 /**
398  * tfrc_rx_hist_sample_rtt  -  Sample RTT from timestamp / CCVal
399  * Based on ideas presented in RFC 4342, 8.1. Returns 0 if it was not able
400  * to compute a sample with given data - calling function should check this.
401  */
402 u32 tfrc_rx_hist_sample_rtt(struct tfrc_rx_hist *h, const struct sk_buff *skb)
403 {
404 	u32 sample = 0,
405 	    delta_v = SUB16(dccp_hdr(skb)->dccph_ccval,
406 			    tfrc_rx_hist_rtt_last_s(h)->tfrchrx_ccval);
407 
408 	if (delta_v < 1 || delta_v > 4) {	/* unsuitable CCVal delta */
409 		if (h->rtt_sample_prev == 2) {	/* previous candidate stored */
410 			sample = SUB16(tfrc_rx_hist_rtt_prev_s(h)->tfrchrx_ccval,
411 				       tfrc_rx_hist_rtt_last_s(h)->tfrchrx_ccval);
412 			if (sample)
413 				sample = 4 / sample *
414 				         ktime_us_delta(tfrc_rx_hist_rtt_prev_s(h)->tfrchrx_tstamp,
415 							tfrc_rx_hist_rtt_last_s(h)->tfrchrx_tstamp);
416 			else    /*
417 				 * FIXME: This condition is in principle not
418 				 * possible but occurs when CCID is used for
419 				 * two-way data traffic. I have tried to trace
420 				 * it, but the cause does not seem to be here.
421 				 */
422 				DCCP_BUG("please report to dccp@vger.kernel.org"
423 					 " => prev = %u, last = %u",
424 					 tfrc_rx_hist_rtt_prev_s(h)->tfrchrx_ccval,
425 					 tfrc_rx_hist_rtt_last_s(h)->tfrchrx_ccval);
426 		} else if (delta_v < 1) {
427 			h->rtt_sample_prev = 1;
428 			goto keep_ref_for_next_time;
429 		}
430 
431 	} else if (delta_v == 4) /* optimal match */
432 		sample = ktime_to_us(net_timedelta(tfrc_rx_hist_rtt_last_s(h)->tfrchrx_tstamp));
433 	else {			 /* suboptimal match */
434 		h->rtt_sample_prev = 2;
435 		goto keep_ref_for_next_time;
436 	}
437 
438 	if (unlikely(sample > DCCP_SANE_RTT_MAX)) {
439 		DCCP_WARN("RTT sample %u too large, using max\n", sample);
440 		sample = DCCP_SANE_RTT_MAX;
441 	}
442 
443 	h->rtt_sample_prev = 0;	       /* use current entry as next reference */
444 keep_ref_for_next_time:
445 
446 	return sample;
447 }
448