2 * net/sched/sch_tbf.c Token Bucket Filter queue.
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.
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
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>
25 /* Simple Token Bucket Filter.
26 =======================================
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).
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:
44 s_i+....+s_k <= B + R*(t_k - t_i)
49 Let N(t_i) be B/R initially and N(t) grow continuously with time as:
51 N(t+delta) = min{B/R, N(t) + delta}
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:
57 N(t_* + 0) = N(t_* - 0) - S/R.
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.
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.
69 When TBF works in reshaping mode, latency is estimated as:
71 lat = max ((L-B)/R, (L-M)/P)
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.
85 This means, that with depth B, the maximal rate is
89 F.e. for 10Mbit ethernet and HZ=100 the minimal allowed B is ~10Kbytes.
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 :-)
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.
100 struct tbf_sched_data
103 u32 limit; /* Maximal length of backlog: bytes */
104 u32 buffer; /* Token bucket depth/rate: MUST BE >= MTU/B */
107 struct qdisc_rate_table *R_tab;
108 struct qdisc_rate_table *P_tab;
111 long tokens; /* Current number of B tokens */
112 long ptokens; /* Current number of P tokens */
113 psched_time_t t_c; /* Time check-point */
114 struct Qdisc *qdisc; /* Inner qdisc, default - bfifo queue */
115 struct qdisc_watchdog watchdog; /* Watchdog timer */
118 #define L2T(q,L) qdisc_l2t((q)->R_tab,L)
119 #define L2T_P(q,L) qdisc_l2t((q)->P_tab,L)
121 static int tbf_enqueue(struct sk_buff *skb, struct Qdisc* sch)
123 struct tbf_sched_data *q = qdisc_priv(sch);
126 if (qdisc_pkt_len(skb) > q->max_size) {
128 #ifdef CONFIG_NET_CLS_ACT
129 if (sch->reshape_fail == NULL || sch->reshape_fail(skb, sch))
133 return NET_XMIT_DROP;
136 ret = qdisc_enqueue(skb, q->qdisc);
143 sch->bstats.bytes += qdisc_pkt_len(skb);
144 sch->bstats.packets++;
148 static int tbf_requeue(struct sk_buff *skb, struct Qdisc* sch)
150 struct tbf_sched_data *q = qdisc_priv(sch);
153 if ((ret = q->qdisc->ops->requeue(skb, q->qdisc)) == 0) {
155 sch->qstats.requeues++;
161 static unsigned int tbf_drop(struct Qdisc* sch)
163 struct tbf_sched_data *q = qdisc_priv(sch);
164 unsigned int len = 0;
166 if (q->qdisc->ops->drop && (len = q->qdisc->ops->drop(q->qdisc)) != 0) {
173 static struct sk_buff *tbf_dequeue(struct Qdisc* sch)
175 struct tbf_sched_data *q = qdisc_priv(sch);
178 skb = q->qdisc->dequeue(q->qdisc);
184 unsigned int len = qdisc_pkt_len(skb);
186 now = psched_get_time();
187 toks = psched_tdiff_bounded(now, q->t_c, q->buffer);
190 ptoks = toks + q->ptokens;
191 if (ptoks > (long)q->mtu)
193 ptoks -= L2T_P(q, len);
196 if (toks > (long)q->buffer)
200 if ((toks|ptoks) >= 0) {
205 sch->flags &= ~TCQ_F_THROTTLED;
209 qdisc_watchdog_schedule(&q->watchdog,
210 now + max_t(long, -toks, -ptoks));
212 /* Maybe we have a shorter packet in the queue,
213 which can be sent now. It sounds cool,
214 but, however, this is wrong in principle.
215 We MUST NOT reorder packets under these circumstances.
217 Really, if we split the flow into independent
218 subflows, it would be a very good solution.
219 This is the main idea of all FQ algorithms
220 (cf. CSZ, HPFQ, HFSC)
223 if (q->qdisc->ops->requeue(skb, q->qdisc) != NET_XMIT_SUCCESS) {
224 /* When requeue fails skb is dropped */
225 qdisc_tree_decrease_qlen(q->qdisc, 1);
229 sch->qstats.overlimits++;
234 static void tbf_reset(struct Qdisc* sch)
236 struct tbf_sched_data *q = qdisc_priv(sch);
238 qdisc_reset(q->qdisc);
240 q->t_c = psched_get_time();
241 q->tokens = q->buffer;
243 qdisc_watchdog_cancel(&q->watchdog);
246 static const struct nla_policy tbf_policy[TCA_TBF_MAX + 1] = {
247 [TCA_TBF_PARMS] = { .len = sizeof(struct tc_tbf_qopt) },
248 [TCA_TBF_RTAB] = { .type = NLA_BINARY, .len = TC_RTAB_SIZE },
249 [TCA_TBF_PTAB] = { .type = NLA_BINARY, .len = TC_RTAB_SIZE },
252 static int tbf_change(struct Qdisc* sch, struct nlattr *opt)
255 struct tbf_sched_data *q = qdisc_priv(sch);
256 struct nlattr *tb[TCA_TBF_PTAB + 1];
257 struct tc_tbf_qopt *qopt;
258 struct qdisc_rate_table *rtab = NULL;
259 struct qdisc_rate_table *ptab = NULL;
260 struct Qdisc *child = NULL;
263 err = nla_parse_nested(tb, TCA_TBF_PTAB, opt, tbf_policy);
268 if (tb[TCA_TBF_PARMS] == NULL)
271 qopt = nla_data(tb[TCA_TBF_PARMS]);
272 rtab = qdisc_get_rtab(&qopt->rate, tb[TCA_TBF_RTAB]);
276 if (qopt->peakrate.rate) {
277 if (qopt->peakrate.rate > qopt->rate.rate)
278 ptab = qdisc_get_rtab(&qopt->peakrate, tb[TCA_TBF_PTAB]);
283 for (n = 0; n < 256; n++)
284 if (rtab->data[n] > qopt->buffer) break;
285 max_size = (n << qopt->rate.cell_log)-1;
289 for (n = 0; n < 256; n++)
290 if (ptab->data[n] > qopt->mtu) break;
291 size = (n << qopt->peakrate.cell_log)-1;
292 if (size < max_size) max_size = size;
297 if (qopt->limit > 0) {
298 child = fifo_create_dflt(sch, &bfifo_qdisc_ops, qopt->limit);
300 err = PTR_ERR(child);
307 qdisc_tree_decrease_qlen(q->qdisc, q->qdisc->q.qlen);
308 qdisc_destroy(xchg(&q->qdisc, child));
310 q->limit = qopt->limit;
312 q->max_size = max_size;
313 q->buffer = qopt->buffer;
314 q->tokens = q->buffer;
316 rtab = xchg(&q->R_tab, rtab);
317 ptab = xchg(&q->P_tab, ptab);
318 sch_tree_unlock(sch);
322 qdisc_put_rtab(rtab);
324 qdisc_put_rtab(ptab);
328 static int tbf_init(struct Qdisc* sch, struct nlattr *opt)
330 struct tbf_sched_data *q = qdisc_priv(sch);
335 q->t_c = psched_get_time();
336 qdisc_watchdog_init(&q->watchdog, sch);
337 q->qdisc = &noop_qdisc;
339 return tbf_change(sch, opt);
342 static void tbf_destroy(struct Qdisc *sch)
344 struct tbf_sched_data *q = qdisc_priv(sch);
346 qdisc_watchdog_cancel(&q->watchdog);
349 qdisc_put_rtab(q->P_tab);
351 qdisc_put_rtab(q->R_tab);
353 qdisc_destroy(q->qdisc);
356 static int tbf_dump(struct Qdisc *sch, struct sk_buff *skb)
358 struct tbf_sched_data *q = qdisc_priv(sch);
360 struct tc_tbf_qopt opt;
362 nest = nla_nest_start(skb, TCA_OPTIONS);
364 goto nla_put_failure;
366 opt.limit = q->limit;
367 opt.rate = q->R_tab->rate;
369 opt.peakrate = q->P_tab->rate;
371 memset(&opt.peakrate, 0, sizeof(opt.peakrate));
373 opt.buffer = q->buffer;
374 NLA_PUT(skb, TCA_TBF_PARMS, sizeof(opt), &opt);
376 nla_nest_end(skb, nest);
380 nla_nest_cancel(skb, nest);
384 static int tbf_dump_class(struct Qdisc *sch, unsigned long cl,
385 struct sk_buff *skb, struct tcmsg *tcm)
387 struct tbf_sched_data *q = qdisc_priv(sch);
389 if (cl != 1) /* only one class */
392 tcm->tcm_handle |= TC_H_MIN(1);
393 tcm->tcm_info = q->qdisc->handle;
398 static int tbf_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new,
401 struct tbf_sched_data *q = qdisc_priv(sch);
407 *old = xchg(&q->qdisc, new);
408 qdisc_tree_decrease_qlen(*old, (*old)->q.qlen);
410 sch_tree_unlock(sch);
415 static struct Qdisc *tbf_leaf(struct Qdisc *sch, unsigned long arg)
417 struct tbf_sched_data *q = qdisc_priv(sch);
421 static unsigned long tbf_get(struct Qdisc *sch, u32 classid)
426 static void tbf_put(struct Qdisc *sch, unsigned long arg)
430 static int tbf_change_class(struct Qdisc *sch, u32 classid, u32 parentid,
431 struct nlattr **tca, unsigned long *arg)
436 static int tbf_delete(struct Qdisc *sch, unsigned long arg)
441 static void tbf_walk(struct Qdisc *sch, struct qdisc_walker *walker)
444 if (walker->count >= walker->skip)
445 if (walker->fn(sch, 1, walker) < 0) {
453 static struct tcf_proto **tbf_find_tcf(struct Qdisc *sch, unsigned long cl)
458 static const struct Qdisc_class_ops tbf_class_ops =
464 .change = tbf_change_class,
465 .delete = tbf_delete,
467 .tcf_chain = tbf_find_tcf,
468 .dump = tbf_dump_class,
471 static struct Qdisc_ops tbf_qdisc_ops __read_mostly = {
473 .cl_ops = &tbf_class_ops,
475 .priv_size = sizeof(struct tbf_sched_data),
476 .enqueue = tbf_enqueue,
477 .dequeue = tbf_dequeue,
478 .requeue = tbf_requeue,
482 .destroy = tbf_destroy,
483 .change = tbf_change,
485 .owner = THIS_MODULE,
488 static int __init tbf_module_init(void)
490 return register_qdisc(&tbf_qdisc_ops);
493 static void __exit tbf_module_exit(void)
495 unregister_qdisc(&tbf_qdisc_ops);
497 module_init(tbf_module_init)
498 module_exit(tbf_module_exit)
499 MODULE_LICENSE("GPL");