net/sched/sch_tbf.c

   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/pkt_sched.h>
  23
  24
  25 /*      Simple Token Bucket Filter.
  26         =======================================
  27
  28         SOURCE.
  29         -------
  30
  31         None.
  32
  33         Description.
  34         ------------
  35
  36         A data flow obeys TBF with rate R and depth B, if for any
  37         time interval t_i...t_f the number of transmitted bits
  38         does not exceed B + R*(t_f-t_i).
  39
  40         Packetized version of this definition:
  41         The sequence of packets of sizes s_i served at moments t_i
  42         obeys TBF, if for any i<=k:
  43
  44         s_i+....+s_k <= B + R*(t_k - t_i)
  45
  46         Algorithm.
  47         ----------
  48
  49         Let N(t_i) be B/R initially and N(t) grow continuously with time as:
  50
  51         N(t+delta) = min{B/R, N(t) + delta}
  52
  53         If the first packet in queue has length S, it may be
  54         transmitted only at the time t_* when S/R <= N(t_*),
  55         and in this case N(t) jumps:
  56
  57         N(t_* + 0) = N(t_* - 0) - S/R.
  58
  59
  60
  61         Actually, QoS requires two TBF to be applied to a data stream.
  62         One of them controls steady state burst size, another
  63         one with rate P (peak rate) and depth M (equal to link MTU)
  64         limits bursts at a smaller time scale.
  65
  66         It is easy to see that P>R, and B>M. If P is infinity, this double
  67         TBF is equivalent to a single one.
  68
  69         When TBF works in reshaping mode, latency is estimated as:
  70
  71         lat = max ((L-B)/R, (L-M)/P)
  72
  73
  74         NOTES.
  75         ------
  76
  77         If TBF throttles, it starts a watchdog timer, which will wake it up
  78         when it is ready to transmit.
  79         Note that the minimal timer resolution is 1/HZ.
  80         If no new packets arrive during this period,
  81         or if the device is not awaken by EOI for some previous packet,
  82         TBF can stop its activity for 1/HZ.
  83
  84
  85         This means, that with depth B, the maximal rate is
  86
  87         R_crit = B*HZ
  88
  89         F.e. for 10Mbit ethernet and HZ=100 the minimal allowed B is ~10Kbytes.
  90
  91         Note that the peak rate TBF is much more tough: with MTU 1500
  92         P_crit = 150Kbytes/sec. So, if you need greater peak
  93         rates, use alpha with HZ=1000 :-)
  94
  95         With classful TBF, limit is just kept for backwards compatibility.
  96         It is passed to the default bfifo qdisc - if the inner qdisc is
  97         changed the limit is not effective anymore.
  98 */
  99
 100 struct tbf_sched_data {
 101 /* Parameters */
 102         u32             limit;          /* Maximal length of backlog: bytes */
 103         u32             buffer;         /* Token bucket depth/rate: MUST BE >= MTU/B */
 104         u32             mtu;
 105         u32             max_size;
 106         struct qdisc_rate_table *R_tab;
 107         struct qdisc_rate_table *P_tab;
 108
 109 /* Variables */
 110         long    tokens;                 /* Current number of B tokens */
 111         long    ptokens;                /* Current number of P tokens */
 112         psched_time_t   t_c;            /* Time check-point */
 113         struct Qdisc    *qdisc;         /* Inner qdisc, default - bfifo queue */
 114         struct qdisc_watchdog watchdog; /* Watchdog timer */
 115 };
 116
 117 #define L2T(q, L)   qdisc_l2t((q)->R_tab, L)
 118 #define L2T_P(q, L) qdisc_l2t((q)->P_tab, L)
 119
 120 static int tbf_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 121 {
 122         struct tbf_sched_data *q = qdisc_priv(sch);
 123         int ret;
 124
 125         if (qdisc_pkt_len(skb) > q->max_size)
 126                 return qdisc_reshape_fail(skb, sch);
 127
 128         ret = qdisc_enqueue(skb, q->qdisc);
 129         if (ret != NET_XMIT_SUCCESS) {
 130                 if (net_xmit_drop_count(ret))
 131                         sch->qstats.drops++;
 132                 return ret;
 133         }
 134
 135         sch->q.qlen++;
 136         return NET_XMIT_SUCCESS;
 137 }
 138
 139 static unsigned int tbf_drop(struct Qdisc *sch)
 140 {
 141         struct tbf_sched_data *q = qdisc_priv(sch);
 142         unsigned int len = 0;
 143
 144         if (q->qdisc->ops->drop && (len = q->qdisc->ops->drop(q->qdisc)) != 0) {
 145                 sch->q.qlen--;
 146                 sch->qstats.drops++;
 147         }
 148         return len;
 149 }
 150
 151 static struct sk_buff *tbf_dequeue(struct Qdisc *sch)
 152 {
 153         struct tbf_sched_data *q = qdisc_priv(sch);
 154         struct sk_buff *skb;
 155
 156         skb = q->qdisc->ops->peek(q->qdisc);
 157
 158         if (skb) {
 159                 psched_time_t now;
 160                 long toks;
 161                 long ptoks = 0;
 162                 unsigned int len = qdisc_pkt_len(skb);
 163
 164                 now = psched_get_time();
 165                 toks = psched_tdiff_bounded(now, q->t_c, q->buffer);
 166
 167                 if (q->P_tab) {
 168                         ptoks = toks + q->ptokens;
 169                         if (ptoks > (long)q->mtu)
 170                                 ptoks = q->mtu;
 171                         ptoks -= L2T_P(q, len);
 172                 }
 173                 toks += q->tokens;
 174                 if (toks > (long)q->buffer)
 175                         toks = q->buffer;
 176                 toks -= L2T(q, len);
 177
 178                 if ((toks|ptoks) >= 0) {
 179                         skb = qdisc_dequeue_peeked(q->qdisc);
 180                         if (unlikely(!skb))
 181                                 return NULL;
 182
 183                         q->t_c = now;
 184                         q->tokens = toks;
 185                         q->ptokens = ptoks;
 186                         sch->q.qlen--;
 187                         qdisc_unthrottled(sch);
 188                         qdisc_bstats_update(sch, skb);
 189                         return skb;
 190                 }
 191
 192                 qdisc_watchdog_schedule(&q->watchdog,
 193                                         now + max_t(long, -toks, -ptoks));
 194
 195                 /* Maybe we have a shorter packet in the queue,
 196                    which can be sent now. It sounds cool,
 197                    but, however, this is wrong in principle.
 198                    We MUST NOT reorder packets under these circumstances.
 199
 200                    Really, if we split the flow into independent
 201                    subflows, it would be a very good solution.
 202                    This is the main idea of all FQ algorithms
 203                    (cf. CSZ, HPFQ, HFSC)
 204                  */
 205
 206                 sch->qstats.overlimits++;
 207         }
 208         return NULL;
 209 }
 210
 211 static void tbf_reset(struct Qdisc *sch)
 212 {
 213         struct tbf_sched_data *q = qdisc_priv(sch);
 214
 215         qdisc_reset(q->qdisc);
 216         sch->q.qlen = 0;
 217         q->t_c = psched_get_time();
 218         q->tokens = q->buffer;
 219         q->ptokens = q->mtu;
 220         qdisc_watchdog_cancel(&q->watchdog);
 221 }
 222
 223 static const struct nla_policy tbf_policy[TCA_TBF_MAX + 1] = {
 224         [TCA_TBF_PARMS] = { .len = sizeof(struct tc_tbf_qopt) },
 225         [TCA_TBF_RTAB]  = { .type = NLA_BINARY, .len = TC_RTAB_SIZE },
 226         [TCA_TBF_PTAB]  = { .type = NLA_BINARY, .len = TC_RTAB_SIZE },
 227 };
 228
 229 static int tbf_change(struct Qdisc *sch, struct nlattr *opt)
 230 {
 231         int err;
 232         struct tbf_sched_data *q = qdisc_priv(sch);
 233         struct nlattr *tb[TCA_TBF_PTAB + 1];
 234         struct tc_tbf_qopt *qopt;
 235         struct qdisc_rate_table *rtab = NULL;
 236         struct qdisc_rate_table *ptab = NULL;
 237         struct Qdisc *child = NULL;
 238         int max_size, n;
 239
 240         err = nla_parse_nested(tb, TCA_TBF_PTAB, opt, tbf_policy);
 241         if (err < 0)
 242                 return err;
 243
 244         err = -EINVAL;
 245         if (tb[TCA_TBF_PARMS] == NULL)
 246                 goto done;
 247
 248         qopt = nla_data(tb[TCA_TBF_PARMS]);
 249         rtab = qdisc_get_rtab(&qopt->rate, tb[TCA_TBF_RTAB]);
 250         if (rtab == NULL)
 251                 goto done;
 252
 253         if (qopt->peakrate.rate) {
 254                 if (qopt->peakrate.rate > qopt->rate.rate)
 255                         ptab = qdisc_get_rtab(&qopt->peakrate, tb[TCA_TBF_PTAB]);
 256                 if (ptab == NULL)
 257                         goto done;
 258         }
 259
 260         for (n = 0; n < 256; n++)
 261                 if (rtab->data[n] > qopt->buffer)
 262                         break;
 263         max_size = (n << qopt->rate.cell_log) - 1;
 264         if (ptab) {
 265                 int size;
 266
 267                 for (n = 0; n < 256; n++)
 268                         if (ptab->data[n] > qopt->mtu)
 269                                 break;
 270                 size = (n << qopt->peakrate.cell_log) - 1;
 271                 if (size < max_size)
 272                         max_size = size;
 273         }
 274         if (max_size < 0)
 275                 goto done;
 276
 277         if (q->qdisc != &noop_qdisc) {
 278                 err = fifo_set_limit(q->qdisc, qopt->limit);
 279                 if (err)
 280                         goto done;
 281         } else if (qopt->limit > 0) {
 282                 child = fifo_create_dflt(sch, &bfifo_qdisc_ops, qopt->limit);
 283                 if (IS_ERR(child)) {
 284                         err = PTR_ERR(child);
 285                         goto done;
 286                 }
 287         }
 288
 289         sch_tree_lock(sch);
 290         if (child) {
 291                 qdisc_tree_decrease_qlen(q->qdisc, q->qdisc->q.qlen);
 292                 qdisc_destroy(q->qdisc);
 293                 q->qdisc = child;
 294         }
 295         q->limit = qopt->limit;
 296         q->mtu = qopt->mtu;
 297         q->max_size = max_size;
 298         q->buffer = qopt->buffer;
 299         q->tokens = q->buffer;
 300         q->ptokens = q->mtu;
 301
 302         swap(q->R_tab, rtab);
 303         swap(q->P_tab, ptab);
 304
 305         sch_tree_unlock(sch);
 306         err = 0;
 307 done:
 308         if (rtab)
 309                 qdisc_put_rtab(rtab);
 310         if (ptab)
 311                 qdisc_put_rtab(ptab);
 312         return err;
 313 }
 314
 315 static int tbf_init(struct Qdisc *sch, struct nlattr *opt)
 316 {
 317         struct tbf_sched_data *q = qdisc_priv(sch);
 318
 319         if (opt == NULL)
 320                 return -EINVAL;
 321
 322         q->t_c = psched_get_time();
 323         qdisc_watchdog_init(&q->watchdog, sch);
 324         q->qdisc = &noop_qdisc;
 325
 326         return tbf_change(sch, opt);
 327 }
 328
 329 static void tbf_destroy(struct Qdisc *sch)
 330 {
 331         struct tbf_sched_data *q = qdisc_priv(sch);
 332
 333         qdisc_watchdog_cancel(&q->watchdog);
 334
 335         if (q->P_tab)
 336                 qdisc_put_rtab(q->P_tab);
 337         if (q->R_tab)
 338                 qdisc_put_rtab(q->R_tab);
 339
 340         qdisc_destroy(q->qdisc);
 341 }
 342
 343 static int tbf_dump(struct Qdisc *sch, struct sk_buff *skb)
 344 {
 345         struct tbf_sched_data *q = qdisc_priv(sch);
 346         struct nlattr *nest;
 347         struct tc_tbf_qopt opt;
 348
 349         nest = nla_nest_start(skb, TCA_OPTIONS);
 350         if (nest == NULL)
 351                 goto nla_put_failure;
 352
 353         opt.limit = q->limit;
 354         opt.rate = q->R_tab->rate;
 355         if (q->P_tab)
 356                 opt.peakrate = q->P_tab->rate;
 357         else
 358                 memset(&opt.peakrate, 0, sizeof(opt.peakrate));
 359         opt.mtu = q->mtu;
 360         opt.buffer = q->buffer;
 361         NLA_PUT(skb, TCA_TBF_PARMS, sizeof(opt), &opt);
 362
 363         nla_nest_end(skb, nest);
 364         return skb->len;
 365
 366 nla_put_failure:
 367         nla_nest_cancel(skb, nest);
 368         return -1;
 369 }
 370
 371 static int tbf_dump_class(struct Qdisc *sch, unsigned long cl,
 372                           struct sk_buff *skb, struct tcmsg *tcm)
 373 {
 374         struct tbf_sched_data *q = qdisc_priv(sch);
 375
 376         tcm->tcm_handle |= TC_H_MIN(1);
 377         tcm->tcm_info = q->qdisc->handle;
 378
 379         return 0;
 380 }
 381
 382 static int tbf_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new,
 383                      struct Qdisc **old)
 384 {
 385         struct tbf_sched_data *q = qdisc_priv(sch);
 386
 387         if (new == NULL)
 388                 new = &noop_qdisc;
 389
 390         sch_tree_lock(sch);
 391         *old = q->qdisc;
 392         q->qdisc = new;
 393         qdisc_tree_decrease_qlen(*old, (*old)->q.qlen);
 394         qdisc_reset(*old);
 395         sch_tree_unlock(sch);
 396
 397         return 0;
 398 }
 399
 400 static struct Qdisc *tbf_leaf(struct Qdisc *sch, unsigned long arg)
 401 {
 402         struct tbf_sched_data *q = qdisc_priv(sch);
 403         return q->qdisc;
 404 }
 405
 406 static unsigned long tbf_get(struct Qdisc *sch, u32 classid)
 407 {
 408         return 1;
 409 }
 410
 411 static void tbf_put(struct Qdisc *sch, unsigned long arg)
 412 {
 413 }
 414
 415 static void tbf_walk(struct Qdisc *sch, struct qdisc_walker *walker)
 416 {
 417         if (!walker->stop) {
 418                 if (walker->count >= walker->skip)
 419                         if (walker->fn(sch, 1, walker) < 0) {
 420                                 walker->stop = 1;
 421                                 return;
 422                         }
 423                 walker->count++;
 424         }
 425 }
 426
 427 static const struct Qdisc_class_ops tbf_class_ops = {
 428         .graft          =       tbf_graft,
 429         .leaf           =       tbf_leaf,
 430         .get            =       tbf_get,
 431         .put            =       tbf_put,
 432         .walk           =       tbf_walk,
 433         .dump           =       tbf_dump_class,
 434 };
 435
 436 static struct Qdisc_ops tbf_qdisc_ops __read_mostly = {
 437         .next           =       NULL,
 438         .cl_ops         =       &tbf_class_ops,
 439         .id             =       "tbf",
 440         .priv_size      =       sizeof(struct tbf_sched_data),
 441         .enqueue        =       tbf_enqueue,
 442         .dequeue        =       tbf_dequeue,
 443         .peek           =       qdisc_peek_dequeued,
 444         .drop           =       tbf_drop,
 445         .init           =       tbf_init,
 446         .reset          =       tbf_reset,
 447         .destroy        =       tbf_destroy,
 448         .change         =       tbf_change,
 449         .dump           =       tbf_dump,
 450         .owner          =       THIS_MODULE,
 451 };
 452
 453 static int __init tbf_module_init(void)
 454 {
 455         return register_qdisc(&tbf_qdisc_ops);
 456 }
 457
 458 static void __exit tbf_module_exit(void)
 459 {
 460         unregister_qdisc(&tbf_qdisc_ops);
 461 }
 462 module_init(tbf_module_init)
 463 module_exit(tbf_module_exit)
 464 MODULE_LICENSE("GPL");