GNU Linux-libre 4.19.286-gnu1
[releases.git] / net / dccp / ccids / lib / packet_history.c
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 }