xref: /linux/net/sched/sch_tbf.c (revision 56d06fa29edd58c448766014afd833b7ff51247b)
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 
25 
26 /*	Simple Token Bucket Filter.
27 	=======================================
28 
29 	SOURCE.
30 	-------
31 
32 	None.
33 
34 	Description.
35 	------------
36 
37 	A data flow obeys TBF with rate R and depth B, if for any
38 	time interval t_i...t_f the number of transmitted bits
39 	does not exceed B + R*(t_f-t_i).
40 
41 	Packetized version of this definition:
42 	The sequence of packets of sizes s_i served at moments t_i
43 	obeys TBF, if for any i<=k:
44 
45 	s_i+....+s_k <= B + R*(t_k - t_i)
46 
47 	Algorithm.
48 	----------
49 
50 	Let N(t_i) be B/R initially and N(t) grow continuously with time as:
51 
52 	N(t+delta) = min{B/R, N(t) + delta}
53 
54 	If the first packet in queue has length S, it may be
55 	transmitted only at the time t_* when S/R <= N(t_*),
56 	and in this case N(t) jumps:
57 
58 	N(t_* + 0) = N(t_* - 0) - S/R.
59 
60 
61 
62 	Actually, QoS requires two TBF to be applied to a data stream.
63 	One of them controls steady state burst size, another
64 	one with rate P (peak rate) and depth M (equal to link MTU)
65 	limits bursts at a smaller time scale.
66 
67 	It is easy to see that P>R, and B>M. If P is infinity, this double
68 	TBF is equivalent to a single one.
69 
70 	When TBF works in reshaping mode, latency is estimated as:
71 
72 	lat = max ((L-B)/R, (L-M)/P)
73 
74 
75 	NOTES.
76 	------
77 
78 	If TBF throttles, it starts a watchdog timer, which will wake it up
79 	when it is ready to transmit.
80 	Note that the minimal timer resolution is 1/HZ.
81 	If no new packets arrive during this period,
82 	or if the device is not awaken by EOI for some previous packet,
83 	TBF can stop its activity for 1/HZ.
84 
85 
86 	This means, that with depth B, the maximal rate is
87 
88 	R_crit = B*HZ
89 
90 	F.e. for 10Mbit ethernet and HZ=100 the minimal allowed B is ~10Kbytes.
91 
92 	Note that the peak rate TBF is much more tough: with MTU 1500
93 	P_crit = 150Kbytes/sec. So, if you need greater peak
94 	rates, use alpha with HZ=1000 :-)
95 
96 	With classful TBF, limit is just kept for backwards compatibility.
97 	It is passed to the default bfifo qdisc - if the inner qdisc is
98 	changed the limit is not effective anymore.
99 */
100 
101 struct tbf_sched_data {
102 /* Parameters */
103 	u32		limit;		/* Maximal length of backlog: bytes */
104 	u32		max_size;
105 	s64		buffer;		/* Token bucket depth/rate: MUST BE >= MTU/B */
106 	s64		mtu;
107 	struct psched_ratecfg rate;
108 	struct psched_ratecfg peak;
109 
110 /* Variables */
111 	s64	tokens;			/* Current number of B tokens */
112 	s64	ptokens;		/* Current number of P tokens */
113 	s64	t_c;			/* Time check-point */
114 	struct Qdisc	*qdisc;		/* Inner qdisc, default - bfifo queue */
115 	struct qdisc_watchdog watchdog;	/* Watchdog timer */
116 };
117 
118 
119 /* Time to Length, convert time in ns to length in bytes
120  * to determinate how many bytes can be sent in given time.
121  */
122 static u64 psched_ns_t2l(const struct psched_ratecfg *r,
123 			 u64 time_in_ns)
124 {
125 	/* The formula is :
126 	 * len = (time_in_ns * r->rate_bytes_ps) / NSEC_PER_SEC
127 	 */
128 	u64 len = time_in_ns * r->rate_bytes_ps;
129 
130 	do_div(len, NSEC_PER_SEC);
131 
132 	if (unlikely(r->linklayer == TC_LINKLAYER_ATM)) {
133 		do_div(len, 53);
134 		len = len * 48;
135 	}
136 
137 	if (len > r->overhead)
138 		len -= r->overhead;
139 	else
140 		len = 0;
141 
142 	return len;
143 }
144 
145 /*
146  * Return length of individual segments of a gso packet,
147  * including all headers (MAC, IP, TCP/UDP)
148  */
149 static unsigned int skb_gso_mac_seglen(const struct sk_buff *skb)
150 {
151 	unsigned int hdr_len = skb_transport_header(skb) - skb_mac_header(skb);
152 	return hdr_len + skb_gso_transport_seglen(skb);
153 }
154 
155 /* GSO packet is too big, segment it so that tbf can transmit
156  * each segment in time
157  */
158 static int tbf_segment(struct sk_buff *skb, struct Qdisc *sch)
159 {
160 	struct tbf_sched_data *q = qdisc_priv(sch);
161 	struct sk_buff *segs, *nskb;
162 	netdev_features_t features = netif_skb_features(skb);
163 	unsigned int len = 0, prev_len = qdisc_pkt_len(skb);
164 	int ret, nb;
165 
166 	segs = skb_gso_segment(skb, features & ~NETIF_F_GSO_MASK);
167 
168 	if (IS_ERR_OR_NULL(segs))
169 		return qdisc_reshape_fail(skb, sch);
170 
171 	nb = 0;
172 	while (segs) {
173 		nskb = segs->next;
174 		segs->next = NULL;
175 		qdisc_skb_cb(segs)->pkt_len = segs->len;
176 		len += segs->len;
177 		ret = qdisc_enqueue(segs, q->qdisc);
178 		if (ret != NET_XMIT_SUCCESS) {
179 			if (net_xmit_drop_count(ret))
180 				qdisc_qstats_drop(sch);
181 		} else {
182 			nb++;
183 		}
184 		segs = nskb;
185 	}
186 	sch->q.qlen += nb;
187 	if (nb > 1)
188 		qdisc_tree_reduce_backlog(sch, 1 - nb, prev_len - len);
189 	consume_skb(skb);
190 	return nb > 0 ? NET_XMIT_SUCCESS : NET_XMIT_DROP;
191 }
192 
193 static int tbf_enqueue(struct sk_buff *skb, struct Qdisc *sch)
194 {
195 	struct tbf_sched_data *q = qdisc_priv(sch);
196 	int ret;
197 
198 	if (qdisc_pkt_len(skb) > q->max_size) {
199 		if (skb_is_gso(skb) && skb_gso_mac_seglen(skb) <= q->max_size)
200 			return tbf_segment(skb, sch);
201 		return qdisc_reshape_fail(skb, sch);
202 	}
203 	ret = qdisc_enqueue(skb, q->qdisc);
204 	if (ret != NET_XMIT_SUCCESS) {
205 		if (net_xmit_drop_count(ret))
206 			qdisc_qstats_drop(sch);
207 		return ret;
208 	}
209 
210 	sch->q.qlen++;
211 	return NET_XMIT_SUCCESS;
212 }
213 
214 static unsigned int tbf_drop(struct Qdisc *sch)
215 {
216 	struct tbf_sched_data *q = qdisc_priv(sch);
217 	unsigned int len = 0;
218 
219 	if (q->qdisc->ops->drop && (len = q->qdisc->ops->drop(q->qdisc)) != 0) {
220 		sch->q.qlen--;
221 		qdisc_qstats_drop(sch);
222 	}
223 	return len;
224 }
225 
226 static bool tbf_peak_present(const struct tbf_sched_data *q)
227 {
228 	return q->peak.rate_bytes_ps;
229 }
230 
231 static struct sk_buff *tbf_dequeue(struct Qdisc *sch)
232 {
233 	struct tbf_sched_data *q = qdisc_priv(sch);
234 	struct sk_buff *skb;
235 
236 	skb = q->qdisc->ops->peek(q->qdisc);
237 
238 	if (skb) {
239 		s64 now;
240 		s64 toks;
241 		s64 ptoks = 0;
242 		unsigned int len = qdisc_pkt_len(skb);
243 
244 		now = ktime_get_ns();
245 		toks = min_t(s64, now - q->t_c, q->buffer);
246 
247 		if (tbf_peak_present(q)) {
248 			ptoks = toks + q->ptokens;
249 			if (ptoks > q->mtu)
250 				ptoks = q->mtu;
251 			ptoks -= (s64) psched_l2t_ns(&q->peak, len);
252 		}
253 		toks += q->tokens;
254 		if (toks > q->buffer)
255 			toks = q->buffer;
256 		toks -= (s64) psched_l2t_ns(&q->rate, len);
257 
258 		if ((toks|ptoks) >= 0) {
259 			skb = qdisc_dequeue_peeked(q->qdisc);
260 			if (unlikely(!skb))
261 				return NULL;
262 
263 			q->t_c = now;
264 			q->tokens = toks;
265 			q->ptokens = ptoks;
266 			sch->q.qlen--;
267 			qdisc_unthrottled(sch);
268 			qdisc_bstats_update(sch, skb);
269 			return skb;
270 		}
271 
272 		qdisc_watchdog_schedule_ns(&q->watchdog,
273 					   now + max_t(long, -toks, -ptoks),
274 					   true);
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 		qdisc_qstats_overlimit(sch);
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_get_ns();
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 	buffer = min_t(u64, PSCHED_TICKS2NS(qopt->buffer), ~0U);
345 	mtu = min_t(u64, PSCHED_TICKS2NS(qopt->mtu), ~0U);
346 
347 	if (tb[TCA_TBF_RATE64])
348 		rate64 = nla_get_u64(tb[TCA_TBF_RATE64]);
349 	psched_ratecfg_precompute(&rate, &qopt->rate, rate64);
350 
351 	if (tb[TCA_TBF_BURST]) {
352 		max_size = nla_get_u32(tb[TCA_TBF_BURST]);
353 		buffer = psched_l2t_ns(&rate, max_size);
354 	} else {
355 		max_size = min_t(u64, psched_ns_t2l(&rate, buffer), ~0U);
356 	}
357 
358 	if (qopt->peakrate.rate) {
359 		if (tb[TCA_TBF_PRATE64])
360 			prate64 = nla_get_u64(tb[TCA_TBF_PRATE64]);
361 		psched_ratecfg_precompute(&peak, &qopt->peakrate, prate64);
362 		if (peak.rate_bytes_ps <= rate.rate_bytes_ps) {
363 			pr_warn_ratelimited("sch_tbf: peakrate %llu is lower than or equals to rate %llu !\n",
364 					peak.rate_bytes_ps, rate.rate_bytes_ps);
365 			err = -EINVAL;
366 			goto done;
367 		}
368 
369 		if (tb[TCA_TBF_PBURST]) {
370 			u32 pburst = nla_get_u32(tb[TCA_TBF_PBURST]);
371 			max_size = min_t(u32, max_size, pburst);
372 			mtu = psched_l2t_ns(&peak, pburst);
373 		} else {
374 			max_size = min_t(u64, max_size, psched_ns_t2l(&peak, mtu));
375 		}
376 	} else {
377 		memset(&peak, 0, sizeof(peak));
378 	}
379 
380 	if (max_size < psched_mtu(qdisc_dev(sch)))
381 		pr_warn_ratelimited("sch_tbf: burst %llu is lower than device %s mtu (%u) !\n",
382 				    max_size, qdisc_dev(sch)->name,
383 				    psched_mtu(qdisc_dev(sch)));
384 
385 	if (!max_size) {
386 		err = -EINVAL;
387 		goto done;
388 	}
389 
390 	if (q->qdisc != &noop_qdisc) {
391 		err = fifo_set_limit(q->qdisc, qopt->limit);
392 		if (err)
393 			goto done;
394 	} else if (qopt->limit > 0) {
395 		child = fifo_create_dflt(sch, &bfifo_qdisc_ops, qopt->limit);
396 		if (IS_ERR(child)) {
397 			err = PTR_ERR(child);
398 			goto done;
399 		}
400 	}
401 
402 	sch_tree_lock(sch);
403 	if (child) {
404 		qdisc_tree_reduce_backlog(q->qdisc, q->qdisc->q.qlen,
405 					  q->qdisc->qstats.backlog);
406 		qdisc_destroy(q->qdisc);
407 		q->qdisc = child;
408 	}
409 	q->limit = qopt->limit;
410 	if (tb[TCA_TBF_PBURST])
411 		q->mtu = mtu;
412 	else
413 		q->mtu = PSCHED_TICKS2NS(qopt->mtu);
414 	q->max_size = max_size;
415 	if (tb[TCA_TBF_BURST])
416 		q->buffer = buffer;
417 	else
418 		q->buffer = PSCHED_TICKS2NS(qopt->buffer);
419 	q->tokens = q->buffer;
420 	q->ptokens = q->mtu;
421 
422 	memcpy(&q->rate, &rate, sizeof(struct psched_ratecfg));
423 	memcpy(&q->peak, &peak, sizeof(struct psched_ratecfg));
424 
425 	sch_tree_unlock(sch);
426 	err = 0;
427 done:
428 	return err;
429 }
430 
431 static int tbf_init(struct Qdisc *sch, struct nlattr *opt)
432 {
433 	struct tbf_sched_data *q = qdisc_priv(sch);
434 
435 	if (opt == NULL)
436 		return -EINVAL;
437 
438 	q->t_c = ktime_get_ns();
439 	qdisc_watchdog_init(&q->watchdog, sch);
440 	q->qdisc = &noop_qdisc;
441 
442 	return tbf_change(sch, opt);
443 }
444 
445 static void tbf_destroy(struct Qdisc *sch)
446 {
447 	struct tbf_sched_data *q = qdisc_priv(sch);
448 
449 	qdisc_watchdog_cancel(&q->watchdog);
450 	qdisc_destroy(q->qdisc);
451 }
452 
453 static int tbf_dump(struct Qdisc *sch, struct sk_buff *skb)
454 {
455 	struct tbf_sched_data *q = qdisc_priv(sch);
456 	struct nlattr *nest;
457 	struct tc_tbf_qopt opt;
458 
459 	sch->qstats.backlog = q->qdisc->qstats.backlog;
460 	nest = nla_nest_start(skb, TCA_OPTIONS);
461 	if (nest == NULL)
462 		goto nla_put_failure;
463 
464 	opt.limit = q->limit;
465 	psched_ratecfg_getrate(&opt.rate, &q->rate);
466 	if (tbf_peak_present(q))
467 		psched_ratecfg_getrate(&opt.peakrate, &q->peak);
468 	else
469 		memset(&opt.peakrate, 0, sizeof(opt.peakrate));
470 	opt.mtu = PSCHED_NS2TICKS(q->mtu);
471 	opt.buffer = PSCHED_NS2TICKS(q->buffer);
472 	if (nla_put(skb, TCA_TBF_PARMS, sizeof(opt), &opt))
473 		goto nla_put_failure;
474 	if (q->rate.rate_bytes_ps >= (1ULL << 32) &&
475 	    nla_put_u64(skb, TCA_TBF_RATE64, q->rate.rate_bytes_ps))
476 		goto nla_put_failure;
477 	if (tbf_peak_present(q) &&
478 	    q->peak.rate_bytes_ps >= (1ULL << 32) &&
479 	    nla_put_u64(skb, TCA_TBF_PRATE64, q->peak.rate_bytes_ps))
480 		goto nla_put_failure;
481 
482 	return nla_nest_end(skb, nest);
483 
484 nla_put_failure:
485 	nla_nest_cancel(skb, nest);
486 	return -1;
487 }
488 
489 static int tbf_dump_class(struct Qdisc *sch, unsigned long cl,
490 			  struct sk_buff *skb, struct tcmsg *tcm)
491 {
492 	struct tbf_sched_data *q = qdisc_priv(sch);
493 
494 	tcm->tcm_handle |= TC_H_MIN(1);
495 	tcm->tcm_info = q->qdisc->handle;
496 
497 	return 0;
498 }
499 
500 static int tbf_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new,
501 		     struct Qdisc **old)
502 {
503 	struct tbf_sched_data *q = qdisc_priv(sch);
504 
505 	if (new == NULL)
506 		new = &noop_qdisc;
507 
508 	*old = qdisc_replace(sch, new, &q->qdisc);
509 	return 0;
510 }
511 
512 static struct Qdisc *tbf_leaf(struct Qdisc *sch, unsigned long arg)
513 {
514 	struct tbf_sched_data *q = qdisc_priv(sch);
515 	return q->qdisc;
516 }
517 
518 static unsigned long tbf_get(struct Qdisc *sch, u32 classid)
519 {
520 	return 1;
521 }
522 
523 static void tbf_put(struct Qdisc *sch, unsigned long arg)
524 {
525 }
526 
527 static void tbf_walk(struct Qdisc *sch, struct qdisc_walker *walker)
528 {
529 	if (!walker->stop) {
530 		if (walker->count >= walker->skip)
531 			if (walker->fn(sch, 1, walker) < 0) {
532 				walker->stop = 1;
533 				return;
534 			}
535 		walker->count++;
536 	}
537 }
538 
539 static const struct Qdisc_class_ops tbf_class_ops = {
540 	.graft		=	tbf_graft,
541 	.leaf		=	tbf_leaf,
542 	.get		=	tbf_get,
543 	.put		=	tbf_put,
544 	.walk		=	tbf_walk,
545 	.dump		=	tbf_dump_class,
546 };
547 
548 static struct Qdisc_ops tbf_qdisc_ops __read_mostly = {
549 	.next		=	NULL,
550 	.cl_ops		=	&tbf_class_ops,
551 	.id		=	"tbf",
552 	.priv_size	=	sizeof(struct tbf_sched_data),
553 	.enqueue	=	tbf_enqueue,
554 	.dequeue	=	tbf_dequeue,
555 	.peek		=	qdisc_peek_dequeued,
556 	.drop		=	tbf_drop,
557 	.init		=	tbf_init,
558 	.reset		=	tbf_reset,
559 	.destroy	=	tbf_destroy,
560 	.change		=	tbf_change,
561 	.dump		=	tbf_dump,
562 	.owner		=	THIS_MODULE,
563 };
564 
565 static int __init tbf_module_init(void)
566 {
567 	return register_qdisc(&tbf_qdisc_ops);
568 }
569 
570 static void __exit tbf_module_exit(void)
571 {
572 	unregister_qdisc(&tbf_qdisc_ops);
573 }
574 module_init(tbf_module_init)
575 module_exit(tbf_module_exit)
576 MODULE_LICENSE("GPL");
577