1 /* 2 * net/sched/sch_tbf.c Token Bucket Filter queue. 3 * 4 * This program is free software; you can redistribute it and/or 5 * modify it under the terms of the GNU General Public License 6 * as published by the Free Software Foundation; either version 7 * 2 of the License, or (at your option) any later version. 8 * 9 * Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru> 10 * Dmitry Torokhov <dtor@mail.ru> - allow attaching inner qdiscs - 11 * original idea by Martin Devera 12 * 13 */ 14 15 #include <linux/module.h> 16 #include <linux/types.h> 17 #include <linux/kernel.h> 18 #include <linux/string.h> 19 #include <linux/errno.h> 20 #include <linux/skbuff.h> 21 #include <net/netlink.h> 22 #include <net/sch_generic.h> 23 #include <net/pkt_sched.h> 24 #include <net/tcp.h> 25 26 27 /* Simple Token Bucket Filter. 28 ======================================= 29 30 SOURCE. 31 ------- 32 33 None. 34 35 Description. 36 ------------ 37 38 A data flow obeys TBF with rate R and depth B, if for any 39 time interval t_i...t_f the number of transmitted bits 40 does not exceed B + R*(t_f-t_i). 41 42 Packetized version of this definition: 43 The sequence of packets of sizes s_i served at moments t_i 44 obeys TBF, if for any i<=k: 45 46 s_i+....+s_k <= B + R*(t_k - t_i) 47 48 Algorithm. 49 ---------- 50 51 Let N(t_i) be B/R initially and N(t) grow continuously with time as: 52 53 N(t+delta) = min{B/R, N(t) + delta} 54 55 If the first packet in queue has length S, it may be 56 transmitted only at the time t_* when S/R <= N(t_*), 57 and in this case N(t) jumps: 58 59 N(t_* + 0) = N(t_* - 0) - S/R. 60 61 62 63 Actually, QoS requires two TBF to be applied to a data stream. 64 One of them controls steady state burst size, another 65 one with rate P (peak rate) and depth M (equal to link MTU) 66 limits bursts at a smaller time scale. 67 68 It is easy to see that P>R, and B>M. If P is infinity, this double 69 TBF is equivalent to a single one. 70 71 When TBF works in reshaping mode, latency is estimated as: 72 73 lat = max ((L-B)/R, (L-M)/P) 74 75 76 NOTES. 77 ------ 78 79 If TBF throttles, it starts a watchdog timer, which will wake it up 80 when it is ready to transmit. 81 Note that the minimal timer resolution is 1/HZ. 82 If no new packets arrive during this period, 83 or if the device is not awaken by EOI for some previous packet, 84 TBF can stop its activity for 1/HZ. 85 86 87 This means, that with depth B, the maximal rate is 88 89 R_crit = B*HZ 90 91 F.e. for 10Mbit ethernet and HZ=100 the minimal allowed B is ~10Kbytes. 92 93 Note that the peak rate TBF is much more tough: with MTU 1500 94 P_crit = 150Kbytes/sec. So, if you need greater peak 95 rates, use alpha with HZ=1000 :-) 96 97 With classful TBF, limit is just kept for backwards compatibility. 98 It is passed to the default bfifo qdisc - if the inner qdisc is 99 changed the limit is not effective anymore. 100 */ 101 102 struct tbf_sched_data { 103 /* Parameters */ 104 u32 limit; /* Maximal length of backlog: bytes */ 105 s64 buffer; /* Token bucket depth/rate: MUST BE >= MTU/B */ 106 s64 mtu; 107 u32 max_size; 108 struct psched_ratecfg rate; 109 struct psched_ratecfg peak; 110 bool peak_present; 111 112 /* Variables */ 113 s64 tokens; /* Current number of B tokens */ 114 s64 ptokens; /* Current number of P tokens */ 115 s64 t_c; /* Time check-point */ 116 struct Qdisc *qdisc; /* Inner qdisc, default - bfifo queue */ 117 struct qdisc_watchdog watchdog; /* Watchdog timer */ 118 }; 119 120 121 /* Time to Length, convert time in ns to length in bytes 122 * to determinate how many bytes can be sent in given time. 123 */ 124 static u64 psched_ns_t2l(const struct psched_ratecfg *r, 125 u64 time_in_ns) 126 { 127 /* The formula is : 128 * len = (time_in_ns * r->rate_bytes_ps) / NSEC_PER_SEC 129 */ 130 u64 len = time_in_ns * r->rate_bytes_ps; 131 132 do_div(len, NSEC_PER_SEC); 133 134 if (unlikely(r->linklayer == TC_LINKLAYER_ATM)) { 135 do_div(len, 53); 136 len = len * 48; 137 } 138 139 if (len > r->overhead) 140 len -= r->overhead; 141 else 142 len = 0; 143 144 return len; 145 } 146 147 /* 148 * Return length of individual segments of a gso packet, 149 * including all headers (MAC, IP, TCP/UDP) 150 */ 151 static unsigned int skb_gso_seglen(const struct sk_buff *skb) 152 { 153 unsigned int hdr_len = skb_transport_header(skb) - skb_mac_header(skb); 154 const struct skb_shared_info *shinfo = skb_shinfo(skb); 155 156 if (likely(shinfo->gso_type & (SKB_GSO_TCPV4 | SKB_GSO_TCPV6))) 157 hdr_len += tcp_hdrlen(skb); 158 else 159 hdr_len += sizeof(struct udphdr); 160 return hdr_len + shinfo->gso_size; 161 } 162 163 /* GSO packet is too big, segment it so that tbf can transmit 164 * each segment in time 165 */ 166 static int tbf_segment(struct sk_buff *skb, struct Qdisc *sch) 167 { 168 struct tbf_sched_data *q = qdisc_priv(sch); 169 struct sk_buff *segs, *nskb; 170 netdev_features_t features = netif_skb_features(skb); 171 int ret, nb; 172 173 segs = skb_gso_segment(skb, features & ~NETIF_F_GSO_MASK); 174 175 if (IS_ERR_OR_NULL(segs)) 176 return qdisc_reshape_fail(skb, sch); 177 178 nb = 0; 179 while (segs) { 180 nskb = segs->next; 181 segs->next = NULL; 182 qdisc_skb_cb(segs)->pkt_len = segs->len; 183 ret = qdisc_enqueue(segs, q->qdisc); 184 if (ret != NET_XMIT_SUCCESS) { 185 if (net_xmit_drop_count(ret)) 186 sch->qstats.drops++; 187 } else { 188 nb++; 189 } 190 segs = nskb; 191 } 192 sch->q.qlen += nb; 193 if (nb > 1) 194 qdisc_tree_decrease_qlen(sch, 1 - nb); 195 consume_skb(skb); 196 return nb > 0 ? NET_XMIT_SUCCESS : NET_XMIT_DROP; 197 } 198 199 static int tbf_enqueue(struct sk_buff *skb, struct Qdisc *sch) 200 { 201 struct tbf_sched_data *q = qdisc_priv(sch); 202 int ret; 203 204 if (qdisc_pkt_len(skb) > q->max_size) { 205 if (skb_is_gso(skb) && skb_gso_seglen(skb) <= q->max_size) 206 return tbf_segment(skb, sch); 207 return qdisc_reshape_fail(skb, sch); 208 } 209 ret = qdisc_enqueue(skb, q->qdisc); 210 if (ret != NET_XMIT_SUCCESS) { 211 if (net_xmit_drop_count(ret)) 212 sch->qstats.drops++; 213 return ret; 214 } 215 216 sch->q.qlen++; 217 return NET_XMIT_SUCCESS; 218 } 219 220 static unsigned int tbf_drop(struct Qdisc *sch) 221 { 222 struct tbf_sched_data *q = qdisc_priv(sch); 223 unsigned int len = 0; 224 225 if (q->qdisc->ops->drop && (len = q->qdisc->ops->drop(q->qdisc)) != 0) { 226 sch->q.qlen--; 227 sch->qstats.drops++; 228 } 229 return len; 230 } 231 232 static struct sk_buff *tbf_dequeue(struct Qdisc *sch) 233 { 234 struct tbf_sched_data *q = qdisc_priv(sch); 235 struct sk_buff *skb; 236 237 skb = q->qdisc->ops->peek(q->qdisc); 238 239 if (skb) { 240 s64 now; 241 s64 toks; 242 s64 ptoks = 0; 243 unsigned int len = qdisc_pkt_len(skb); 244 245 now = ktime_to_ns(ktime_get()); 246 toks = min_t(s64, now - q->t_c, q->buffer); 247 248 if (q->peak_present) { 249 ptoks = toks + q->ptokens; 250 if (ptoks > q->mtu) 251 ptoks = q->mtu; 252 ptoks -= (s64) psched_l2t_ns(&q->peak, len); 253 } 254 toks += q->tokens; 255 if (toks > q->buffer) 256 toks = q->buffer; 257 toks -= (s64) psched_l2t_ns(&q->rate, len); 258 259 if ((toks|ptoks) >= 0) { 260 skb = qdisc_dequeue_peeked(q->qdisc); 261 if (unlikely(!skb)) 262 return NULL; 263 264 q->t_c = now; 265 q->tokens = toks; 266 q->ptokens = ptoks; 267 sch->q.qlen--; 268 qdisc_unthrottled(sch); 269 qdisc_bstats_update(sch, skb); 270 return skb; 271 } 272 273 qdisc_watchdog_schedule_ns(&q->watchdog, 274 now + max_t(long, -toks, -ptoks)); 275 276 /* Maybe we have a shorter packet in the queue, 277 which can be sent now. It sounds cool, 278 but, however, this is wrong in principle. 279 We MUST NOT reorder packets under these circumstances. 280 281 Really, if we split the flow into independent 282 subflows, it would be a very good solution. 283 This is the main idea of all FQ algorithms 284 (cf. CSZ, HPFQ, HFSC) 285 */ 286 287 sch->qstats.overlimits++; 288 } 289 return NULL; 290 } 291 292 static void tbf_reset(struct Qdisc *sch) 293 { 294 struct tbf_sched_data *q = qdisc_priv(sch); 295 296 qdisc_reset(q->qdisc); 297 sch->q.qlen = 0; 298 q->t_c = ktime_to_ns(ktime_get()); 299 q->tokens = q->buffer; 300 q->ptokens = q->mtu; 301 qdisc_watchdog_cancel(&q->watchdog); 302 } 303 304 static const struct nla_policy tbf_policy[TCA_TBF_MAX + 1] = { 305 [TCA_TBF_PARMS] = { .len = sizeof(struct tc_tbf_qopt) }, 306 [TCA_TBF_RTAB] = { .type = NLA_BINARY, .len = TC_RTAB_SIZE }, 307 [TCA_TBF_PTAB] = { .type = NLA_BINARY, .len = TC_RTAB_SIZE }, 308 [TCA_TBF_RATE64] = { .type = NLA_U64 }, 309 [TCA_TBF_PRATE64] = { .type = NLA_U64 }, 310 [TCA_TBF_BURST] = { .type = NLA_U32 }, 311 [TCA_TBF_PBURST] = { .type = NLA_U32 }, 312 }; 313 314 static int tbf_change(struct Qdisc *sch, struct nlattr *opt) 315 { 316 int err; 317 struct tbf_sched_data *q = qdisc_priv(sch); 318 struct nlattr *tb[TCA_TBF_MAX + 1]; 319 struct tc_tbf_qopt *qopt; 320 struct Qdisc *child = NULL; 321 struct psched_ratecfg rate; 322 struct psched_ratecfg peak; 323 u64 max_size; 324 s64 buffer, mtu; 325 u64 rate64 = 0, prate64 = 0; 326 327 err = nla_parse_nested(tb, TCA_TBF_MAX, opt, tbf_policy); 328 if (err < 0) 329 return err; 330 331 err = -EINVAL; 332 if (tb[TCA_TBF_PARMS] == NULL) 333 goto done; 334 335 qopt = nla_data(tb[TCA_TBF_PARMS]); 336 if (qopt->rate.linklayer == TC_LINKLAYER_UNAWARE) 337 qdisc_put_rtab(qdisc_get_rtab(&qopt->rate, 338 tb[TCA_TBF_RTAB])); 339 340 if (qopt->peakrate.linklayer == TC_LINKLAYER_UNAWARE) 341 qdisc_put_rtab(qdisc_get_rtab(&qopt->peakrate, 342 tb[TCA_TBF_PTAB])); 343 344 if (q->qdisc != &noop_qdisc) { 345 err = fifo_set_limit(q->qdisc, qopt->limit); 346 if (err) 347 goto done; 348 } else if (qopt->limit > 0) { 349 child = fifo_create_dflt(sch, &bfifo_qdisc_ops, qopt->limit); 350 if (IS_ERR(child)) { 351 err = PTR_ERR(child); 352 goto done; 353 } 354 } 355 356 buffer = min_t(u64, PSCHED_TICKS2NS(qopt->buffer), ~0U); 357 mtu = min_t(u64, PSCHED_TICKS2NS(qopt->mtu), ~0U); 358 359 if (tb[TCA_TBF_RATE64]) 360 rate64 = nla_get_u64(tb[TCA_TBF_RATE64]); 361 psched_ratecfg_precompute(&rate, &qopt->rate, rate64); 362 363 if (tb[TCA_TBF_BURST]) { 364 max_size = nla_get_u32(tb[TCA_TBF_BURST]); 365 buffer = psched_l2t_ns(&rate, max_size); 366 } else { 367 max_size = min_t(u64, psched_ns_t2l(&rate, buffer), ~0U); 368 } 369 370 if (qopt->peakrate.rate) { 371 if (tb[TCA_TBF_PRATE64]) 372 prate64 = nla_get_u64(tb[TCA_TBF_PRATE64]); 373 psched_ratecfg_precompute(&peak, &qopt->peakrate, prate64); 374 if (peak.rate_bytes_ps <= rate.rate_bytes_ps) { 375 pr_warn_ratelimited("sch_tbf: peakrate %llu is lower than or equals to rate %llu !\n", 376 peak.rate_bytes_ps, rate.rate_bytes_ps); 377 err = -EINVAL; 378 goto done; 379 } 380 381 if (tb[TCA_TBF_PBURST]) { 382 u32 pburst = nla_get_u32(tb[TCA_TBF_PBURST]); 383 max_size = min_t(u32, max_size, pburst); 384 mtu = psched_l2t_ns(&peak, pburst); 385 } else { 386 max_size = min_t(u64, max_size, psched_ns_t2l(&peak, mtu)); 387 } 388 } 389 390 if (max_size < psched_mtu(qdisc_dev(sch))) 391 pr_warn_ratelimited("sch_tbf: burst %llu is lower than device %s mtu (%u) !\n", 392 max_size, qdisc_dev(sch)->name, 393 psched_mtu(qdisc_dev(sch))); 394 395 if (!max_size) { 396 err = -EINVAL; 397 goto done; 398 } 399 400 sch_tree_lock(sch); 401 if (child) { 402 qdisc_tree_decrease_qlen(q->qdisc, q->qdisc->q.qlen); 403 qdisc_destroy(q->qdisc); 404 q->qdisc = child; 405 } 406 q->limit = qopt->limit; 407 if (tb[TCA_TBF_PBURST]) 408 q->mtu = mtu; 409 else 410 q->mtu = PSCHED_TICKS2NS(qopt->mtu); 411 q->max_size = max_size; 412 if (tb[TCA_TBF_BURST]) 413 q->buffer = buffer; 414 else 415 q->buffer = PSCHED_TICKS2NS(qopt->buffer); 416 q->tokens = q->buffer; 417 q->ptokens = q->mtu; 418 419 memcpy(&q->rate, &rate, sizeof(struct psched_ratecfg)); 420 if (qopt->peakrate.rate) { 421 memcpy(&q->peak, &peak, sizeof(struct psched_ratecfg)); 422 q->peak_present = true; 423 } else { 424 q->peak_present = false; 425 } 426 427 sch_tree_unlock(sch); 428 err = 0; 429 done: 430 return err; 431 } 432 433 static int tbf_init(struct Qdisc *sch, struct nlattr *opt) 434 { 435 struct tbf_sched_data *q = qdisc_priv(sch); 436 437 if (opt == NULL) 438 return -EINVAL; 439 440 q->t_c = ktime_to_ns(ktime_get()); 441 qdisc_watchdog_init(&q->watchdog, sch); 442 q->qdisc = &noop_qdisc; 443 444 return tbf_change(sch, opt); 445 } 446 447 static void tbf_destroy(struct Qdisc *sch) 448 { 449 struct tbf_sched_data *q = qdisc_priv(sch); 450 451 qdisc_watchdog_cancel(&q->watchdog); 452 qdisc_destroy(q->qdisc); 453 } 454 455 static int tbf_dump(struct Qdisc *sch, struct sk_buff *skb) 456 { 457 struct tbf_sched_data *q = qdisc_priv(sch); 458 struct nlattr *nest; 459 struct tc_tbf_qopt opt; 460 461 sch->qstats.backlog = q->qdisc->qstats.backlog; 462 nest = nla_nest_start(skb, TCA_OPTIONS); 463 if (nest == NULL) 464 goto nla_put_failure; 465 466 opt.limit = q->limit; 467 psched_ratecfg_getrate(&opt.rate, &q->rate); 468 if (q->peak_present) 469 psched_ratecfg_getrate(&opt.peakrate, &q->peak); 470 else 471 memset(&opt.peakrate, 0, sizeof(opt.peakrate)); 472 opt.mtu = PSCHED_NS2TICKS(q->mtu); 473 opt.buffer = PSCHED_NS2TICKS(q->buffer); 474 if (nla_put(skb, TCA_TBF_PARMS, sizeof(opt), &opt)) 475 goto nla_put_failure; 476 if (q->rate.rate_bytes_ps >= (1ULL << 32) && 477 nla_put_u64(skb, TCA_TBF_RATE64, q->rate.rate_bytes_ps)) 478 goto nla_put_failure; 479 if (q->peak_present && 480 q->peak.rate_bytes_ps >= (1ULL << 32) && 481 nla_put_u64(skb, TCA_TBF_PRATE64, q->peak.rate_bytes_ps)) 482 goto nla_put_failure; 483 484 nla_nest_end(skb, nest); 485 return skb->len; 486 487 nla_put_failure: 488 nla_nest_cancel(skb, nest); 489 return -1; 490 } 491 492 static int tbf_dump_class(struct Qdisc *sch, unsigned long cl, 493 struct sk_buff *skb, struct tcmsg *tcm) 494 { 495 struct tbf_sched_data *q = qdisc_priv(sch); 496 497 tcm->tcm_handle |= TC_H_MIN(1); 498 tcm->tcm_info = q->qdisc->handle; 499 500 return 0; 501 } 502 503 static int tbf_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new, 504 struct Qdisc **old) 505 { 506 struct tbf_sched_data *q = qdisc_priv(sch); 507 508 if (new == NULL) 509 new = &noop_qdisc; 510 511 sch_tree_lock(sch); 512 *old = q->qdisc; 513 q->qdisc = new; 514 qdisc_tree_decrease_qlen(*old, (*old)->q.qlen); 515 qdisc_reset(*old); 516 sch_tree_unlock(sch); 517 518 return 0; 519 } 520 521 static struct Qdisc *tbf_leaf(struct Qdisc *sch, unsigned long arg) 522 { 523 struct tbf_sched_data *q = qdisc_priv(sch); 524 return q->qdisc; 525 } 526 527 static unsigned long tbf_get(struct Qdisc *sch, u32 classid) 528 { 529 return 1; 530 } 531 532 static void tbf_put(struct Qdisc *sch, unsigned long arg) 533 { 534 } 535 536 static void tbf_walk(struct Qdisc *sch, struct qdisc_walker *walker) 537 { 538 if (!walker->stop) { 539 if (walker->count >= walker->skip) 540 if (walker->fn(sch, 1, walker) < 0) { 541 walker->stop = 1; 542 return; 543 } 544 walker->count++; 545 } 546 } 547 548 static const struct Qdisc_class_ops tbf_class_ops = { 549 .graft = tbf_graft, 550 .leaf = tbf_leaf, 551 .get = tbf_get, 552 .put = tbf_put, 553 .walk = tbf_walk, 554 .dump = tbf_dump_class, 555 }; 556 557 static struct Qdisc_ops tbf_qdisc_ops __read_mostly = { 558 .next = NULL, 559 .cl_ops = &tbf_class_ops, 560 .id = "tbf", 561 .priv_size = sizeof(struct tbf_sched_data), 562 .enqueue = tbf_enqueue, 563 .dequeue = tbf_dequeue, 564 .peek = qdisc_peek_dequeued, 565 .drop = tbf_drop, 566 .init = tbf_init, 567 .reset = tbf_reset, 568 .destroy = tbf_destroy, 569 .change = tbf_change, 570 .dump = tbf_dump, 571 .owner = THIS_MODULE, 572 }; 573 574 static int __init tbf_module_init(void) 575 { 576 return register_qdisc(&tbf_qdisc_ops); 577 } 578 579 static void __exit tbf_module_exit(void) 580 { 581 unregister_qdisc(&tbf_qdisc_ops); 582 } 583 module_init(tbf_module_init) 584 module_exit(tbf_module_exit) 585 MODULE_LICENSE("GPL"); 586