packet_history.c (12880B)
1// SPDX-License-Identifier: GPL-2.0-or-later 2/* 3 * Copyright (c) 2007 The University of Aberdeen, Scotland, UK 4 * Copyright (c) 2005-7 The University of Waikato, Hamilton, New Zealand. 5 * 6 * An implementation of the DCCP protocol 7 * 8 * This code has been developed by the University of Waikato WAND 9 * research group. For further information please see https://www.wand.net.nz/ 10 * or e-mail Ian McDonald - ian.mcdonald@jandi.co.nz 11 * 12 * This code also uses code from Lulea University, rereleased as GPL by its 13 * authors: 14 * Copyright (c) 2003 Nils-Erik Mattsson, Joacim Haggmark, Magnus Erixzon 15 * 16 * Changes to meet Linux coding standards, to make it meet latest ccid3 draft 17 * and to make it work as a loadable module in the DCCP stack written by 18 * Arnaldo Carvalho de Melo <acme@conectiva.com.br>. 19 * 20 * Copyright (c) 2005 Arnaldo Carvalho de Melo <acme@conectiva.com.br> 21 */ 22 23#include <linux/string.h> 24#include <linux/slab.h> 25#include "packet_history.h" 26#include "../../dccp.h" 27 28/* 29 * Transmitter History Routines 30 */ 31static struct kmem_cache *tfrc_tx_hist_slab; 32 33int __init tfrc_tx_packet_history_init(void) 34{ 35 tfrc_tx_hist_slab = kmem_cache_create("tfrc_tx_hist", 36 sizeof(struct tfrc_tx_hist_entry), 37 0, SLAB_HWCACHE_ALIGN, NULL); 38 return tfrc_tx_hist_slab == NULL ? -ENOBUFS : 0; 39} 40 41void tfrc_tx_packet_history_exit(void) 42{ 43 if (tfrc_tx_hist_slab != NULL) { 44 kmem_cache_destroy(tfrc_tx_hist_slab); 45 tfrc_tx_hist_slab = NULL; 46 } 47} 48 49int tfrc_tx_hist_add(struct tfrc_tx_hist_entry **headp, u64 seqno) 50{ 51 struct tfrc_tx_hist_entry *entry = kmem_cache_alloc(tfrc_tx_hist_slab, gfp_any()); 52 53 if (entry == NULL) 54 return -ENOBUFS; 55 entry->seqno = seqno; 56 entry->stamp = ktime_get_real(); 57 entry->next = *headp; 58 *headp = entry; 59 return 0; 60} 61 62void tfrc_tx_hist_purge(struct tfrc_tx_hist_entry **headp) 63{ 64 struct tfrc_tx_hist_entry *head = *headp; 65 66 while (head != NULL) { 67 struct tfrc_tx_hist_entry *next = head->next; 68 69 kmem_cache_free(tfrc_tx_hist_slab, head); 70 head = next; 71 } 72 73 *headp = NULL; 74} 75 76/* 77 * Receiver History Routines 78 */ 79static struct kmem_cache *tfrc_rx_hist_slab; 80 81int __init tfrc_rx_packet_history_init(void) 82{ 83 tfrc_rx_hist_slab = kmem_cache_create("tfrc_rxh_cache", 84 sizeof(struct tfrc_rx_hist_entry), 85 0, SLAB_HWCACHE_ALIGN, NULL); 86 return tfrc_rx_hist_slab == NULL ? -ENOBUFS : 0; 87} 88 89void tfrc_rx_packet_history_exit(void) 90{ 91 if (tfrc_rx_hist_slab != NULL) { 92 kmem_cache_destroy(tfrc_rx_hist_slab); 93 tfrc_rx_hist_slab = NULL; 94 } 95} 96 97static inline void tfrc_rx_hist_entry_from_skb(struct tfrc_rx_hist_entry *entry, 98 const struct sk_buff *skb, 99 const u64 ndp) 100{ 101 const struct dccp_hdr *dh = dccp_hdr(skb); 102 103 entry->tfrchrx_seqno = DCCP_SKB_CB(skb)->dccpd_seq; 104 entry->tfrchrx_ccval = dh->dccph_ccval; 105 entry->tfrchrx_type = dh->dccph_type; 106 entry->tfrchrx_ndp = ndp; 107 entry->tfrchrx_tstamp = ktime_get_real(); 108} 109 110void tfrc_rx_hist_add_packet(struct tfrc_rx_hist *h, 111 const struct sk_buff *skb, 112 const u64 ndp) 113{ 114 struct tfrc_rx_hist_entry *entry = tfrc_rx_hist_last_rcv(h); 115 116 tfrc_rx_hist_entry_from_skb(entry, skb, ndp); 117} 118 119/* has the packet contained in skb been seen before? */ 120int tfrc_rx_hist_duplicate(struct tfrc_rx_hist *h, struct sk_buff *skb) 121{ 122 const u64 seq = DCCP_SKB_CB(skb)->dccpd_seq; 123 int i; 124 125 if (dccp_delta_seqno(tfrc_rx_hist_loss_prev(h)->tfrchrx_seqno, seq) <= 0) 126 return 1; 127 128 for (i = 1; i <= h->loss_count; i++) 129 if (tfrc_rx_hist_entry(h, i)->tfrchrx_seqno == seq) 130 return 1; 131 132 return 0; 133} 134 135static void tfrc_rx_hist_swap(struct tfrc_rx_hist *h, const u8 a, const u8 b) 136{ 137 const u8 idx_a = tfrc_rx_hist_index(h, a), 138 idx_b = tfrc_rx_hist_index(h, b); 139 140 swap(h->ring[idx_a], h->ring[idx_b]); 141} 142 143/* 144 * Private helper functions for loss detection. 145 * 146 * In the descriptions, `Si' refers to the sequence number of entry number i, 147 * whose NDP count is `Ni' (lower case is used for variables). 148 * Note: All __xxx_loss functions expect that a test against duplicates has been 149 * performed already: the seqno of the skb must not be less than the seqno 150 * of loss_prev; and it must not equal that of any valid history entry. 151 */ 152static void __do_track_loss(struct tfrc_rx_hist *h, struct sk_buff *skb, u64 n1) 153{ 154 u64 s0 = tfrc_rx_hist_loss_prev(h)->tfrchrx_seqno, 155 s1 = DCCP_SKB_CB(skb)->dccpd_seq; 156 157 if (!dccp_loss_free(s0, s1, n1)) { /* gap between S0 and S1 */ 158 h->loss_count = 1; 159 tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 1), skb, n1); 160 } 161} 162 163static void __one_after_loss(struct tfrc_rx_hist *h, struct sk_buff *skb, u32 n2) 164{ 165 u64 s0 = tfrc_rx_hist_loss_prev(h)->tfrchrx_seqno, 166 s1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_seqno, 167 s2 = DCCP_SKB_CB(skb)->dccpd_seq; 168 169 if (likely(dccp_delta_seqno(s1, s2) > 0)) { /* S1 < S2 */ 170 h->loss_count = 2; 171 tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 2), skb, n2); 172 return; 173 } 174 175 /* S0 < S2 < S1 */ 176 177 if (dccp_loss_free(s0, s2, n2)) { 178 u64 n1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_ndp; 179 180 if (dccp_loss_free(s2, s1, n1)) { 181 /* hole is filled: S0, S2, and S1 are consecutive */ 182 h->loss_count = 0; 183 h->loss_start = tfrc_rx_hist_index(h, 1); 184 } else 185 /* gap between S2 and S1: just update loss_prev */ 186 tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_loss_prev(h), skb, n2); 187 188 } else { /* gap between S0 and S2 */ 189 /* 190 * Reorder history to insert S2 between S0 and S1 191 */ 192 tfrc_rx_hist_swap(h, 0, 3); 193 h->loss_start = tfrc_rx_hist_index(h, 3); 194 tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 1), skb, n2); 195 h->loss_count = 2; 196 } 197} 198 199/* return 1 if a new loss event has been identified */ 200static int __two_after_loss(struct tfrc_rx_hist *h, struct sk_buff *skb, u32 n3) 201{ 202 u64 s0 = tfrc_rx_hist_loss_prev(h)->tfrchrx_seqno, 203 s1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_seqno, 204 s2 = tfrc_rx_hist_entry(h, 2)->tfrchrx_seqno, 205 s3 = DCCP_SKB_CB(skb)->dccpd_seq; 206 207 if (likely(dccp_delta_seqno(s2, s3) > 0)) { /* S2 < S3 */ 208 h->loss_count = 3; 209 tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 3), skb, n3); 210 return 1; 211 } 212 213 /* S3 < S2 */ 214 215 if (dccp_delta_seqno(s1, s3) > 0) { /* S1 < S3 < S2 */ 216 /* 217 * Reorder history to insert S3 between S1 and S2 218 */ 219 tfrc_rx_hist_swap(h, 2, 3); 220 tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 2), skb, n3); 221 h->loss_count = 3; 222 return 1; 223 } 224 225 /* S0 < S3 < S1 */ 226 227 if (dccp_loss_free(s0, s3, n3)) { 228 u64 n1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_ndp; 229 230 if (dccp_loss_free(s3, s1, n1)) { 231 /* hole between S0 and S1 filled by S3 */ 232 u64 n2 = tfrc_rx_hist_entry(h, 2)->tfrchrx_ndp; 233 234 if (dccp_loss_free(s1, s2, n2)) { 235 /* entire hole filled by S0, S3, S1, S2 */ 236 h->loss_start = tfrc_rx_hist_index(h, 2); 237 h->loss_count = 0; 238 } else { 239 /* gap remains between S1 and S2 */ 240 h->loss_start = tfrc_rx_hist_index(h, 1); 241 h->loss_count = 1; 242 } 243 244 } else /* gap exists between S3 and S1, loss_count stays at 2 */ 245 tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_loss_prev(h), skb, n3); 246 247 return 0; 248 } 249 250 /* 251 * The remaining case: S0 < S3 < S1 < S2; gap between S0 and S3 252 * Reorder history to insert S3 between S0 and S1. 253 */ 254 tfrc_rx_hist_swap(h, 0, 3); 255 h->loss_start = tfrc_rx_hist_index(h, 3); 256 tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 1), skb, n3); 257 h->loss_count = 3; 258 259 return 1; 260} 261 262/* recycle RX history records to continue loss detection if necessary */ 263static void __three_after_loss(struct tfrc_rx_hist *h) 264{ 265 /* 266 * At this stage we know already that there is a gap between S0 and S1 267 * (since S0 was the highest sequence number received before detecting 268 * the loss). To recycle the loss record, it is thus only necessary to 269 * check for other possible gaps between S1/S2 and between S2/S3. 270 */ 271 u64 s1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_seqno, 272 s2 = tfrc_rx_hist_entry(h, 2)->tfrchrx_seqno, 273 s3 = tfrc_rx_hist_entry(h, 3)->tfrchrx_seqno; 274 u64 n2 = tfrc_rx_hist_entry(h, 2)->tfrchrx_ndp, 275 n3 = tfrc_rx_hist_entry(h, 3)->tfrchrx_ndp; 276 277 if (dccp_loss_free(s1, s2, n2)) { 278 279 if (dccp_loss_free(s2, s3, n3)) { 280 /* no gap between S2 and S3: entire hole is filled */ 281 h->loss_start = tfrc_rx_hist_index(h, 3); 282 h->loss_count = 0; 283 } else { 284 /* gap between S2 and S3 */ 285 h->loss_start = tfrc_rx_hist_index(h, 2); 286 h->loss_count = 1; 287 } 288 289 } else { /* gap between S1 and S2 */ 290 h->loss_start = tfrc_rx_hist_index(h, 1); 291 h->loss_count = 2; 292 } 293} 294 295/** 296 * tfrc_rx_handle_loss - Loss detection and further processing 297 * @h: The non-empty RX history object 298 * @lh: Loss Intervals database to update 299 * @skb: Currently received packet 300 * @ndp: The NDP count belonging to @skb 301 * @calc_first_li: Caller-dependent computation of first loss interval in @lh 302 * @sk: Used by @calc_first_li (see tfrc_lh_interval_add) 303 * 304 * Chooses action according to pending loss, updates LI database when a new 305 * loss was detected, and does required post-processing. Returns 1 when caller 306 * should send feedback, 0 otherwise. 307 * Since it also takes care of reordering during loss detection and updates the 308 * records accordingly, the caller should not perform any more RX history 309 * operations when loss_count is greater than 0 after calling this function. 310 */ 311int tfrc_rx_handle_loss(struct tfrc_rx_hist *h, 312 struct tfrc_loss_hist *lh, 313 struct sk_buff *skb, const u64 ndp, 314 u32 (*calc_first_li)(struct sock *), struct sock *sk) 315{ 316 int is_new_loss = 0; 317 318 if (h->loss_count == 0) { 319 __do_track_loss(h, skb, ndp); 320 } else if (h->loss_count == 1) { 321 __one_after_loss(h, skb, ndp); 322 } else if (h->loss_count != 2) { 323 DCCP_BUG("invalid loss_count %d", h->loss_count); 324 } else if (__two_after_loss(h, skb, ndp)) { 325 /* 326 * Update Loss Interval database and recycle RX records 327 */ 328 is_new_loss = tfrc_lh_interval_add(lh, h, calc_first_li, sk); 329 __three_after_loss(h); 330 } 331 return is_new_loss; 332} 333 334int tfrc_rx_hist_alloc(struct tfrc_rx_hist *h) 335{ 336 int i; 337 338 for (i = 0; i <= TFRC_NDUPACK; i++) { 339 h->ring[i] = kmem_cache_alloc(tfrc_rx_hist_slab, GFP_ATOMIC); 340 if (h->ring[i] == NULL) 341 goto out_free; 342 } 343 344 h->loss_count = h->loss_start = 0; 345 return 0; 346 347out_free: 348 while (i-- != 0) { 349 kmem_cache_free(tfrc_rx_hist_slab, h->ring[i]); 350 h->ring[i] = NULL; 351 } 352 return -ENOBUFS; 353} 354 355void tfrc_rx_hist_purge(struct tfrc_rx_hist *h) 356{ 357 int i; 358 359 for (i = 0; i <= TFRC_NDUPACK; ++i) 360 if (h->ring[i] != NULL) { 361 kmem_cache_free(tfrc_rx_hist_slab, h->ring[i]); 362 h->ring[i] = NULL; 363 } 364} 365 366/** 367 * tfrc_rx_hist_rtt_last_s - reference entry to compute RTT samples against 368 * @h: The non-empty RX history object 369 */ 370static inline struct tfrc_rx_hist_entry * 371 tfrc_rx_hist_rtt_last_s(const struct tfrc_rx_hist *h) 372{ 373 return h->ring[0]; 374} 375 376/** 377 * tfrc_rx_hist_rtt_prev_s - previously suitable (wrt rtt_last_s) RTT-sampling entry 378 * @h: The non-empty RX history object 379 */ 380static inline struct tfrc_rx_hist_entry * 381 tfrc_rx_hist_rtt_prev_s(const struct tfrc_rx_hist *h) 382{ 383 return h->ring[h->rtt_sample_prev]; 384} 385 386/** 387 * tfrc_rx_hist_sample_rtt - Sample RTT from timestamp / CCVal 388 * @h: receive histogram 389 * @skb: packet containing timestamp. 390 * 391 * Based on ideas presented in RFC 4342, 8.1. Returns 0 if it was not able 392 * to compute a sample with given data - calling function should check this. 393 */ 394u32 tfrc_rx_hist_sample_rtt(struct tfrc_rx_hist *h, const struct sk_buff *skb) 395{ 396 u32 sample = 0, 397 delta_v = SUB16(dccp_hdr(skb)->dccph_ccval, 398 tfrc_rx_hist_rtt_last_s(h)->tfrchrx_ccval); 399 400 if (delta_v < 1 || delta_v > 4) { /* unsuitable CCVal delta */ 401 if (h->rtt_sample_prev == 2) { /* previous candidate stored */ 402 sample = SUB16(tfrc_rx_hist_rtt_prev_s(h)->tfrchrx_ccval, 403 tfrc_rx_hist_rtt_last_s(h)->tfrchrx_ccval); 404 if (sample) 405 sample = 4 / sample * 406 ktime_us_delta(tfrc_rx_hist_rtt_prev_s(h)->tfrchrx_tstamp, 407 tfrc_rx_hist_rtt_last_s(h)->tfrchrx_tstamp); 408 else /* 409 * FIXME: This condition is in principle not 410 * possible but occurs when CCID is used for 411 * two-way data traffic. I have tried to trace 412 * it, but the cause does not seem to be here. 413 */ 414 DCCP_BUG("please report to dccp@vger.kernel.org" 415 " => prev = %u, last = %u", 416 tfrc_rx_hist_rtt_prev_s(h)->tfrchrx_ccval, 417 tfrc_rx_hist_rtt_last_s(h)->tfrchrx_ccval); 418 } else if (delta_v < 1) { 419 h->rtt_sample_prev = 1; 420 goto keep_ref_for_next_time; 421 } 422 423 } else if (delta_v == 4) /* optimal match */ 424 sample = ktime_to_us(net_timedelta(tfrc_rx_hist_rtt_last_s(h)->tfrchrx_tstamp)); 425 else { /* suboptimal match */ 426 h->rtt_sample_prev = 2; 427 goto keep_ref_for_next_time; 428 } 429 430 if (unlikely(sample > DCCP_SANE_RTT_MAX)) { 431 DCCP_WARN("RTT sample %u too large, using max\n", sample); 432 sample = DCCP_SANE_RTT_MAX; 433 } 434 435 h->rtt_sample_prev = 0; /* use current entry as next reference */ 436keep_ref_for_next_time: 437 438 return sample; 439}