1 // SPDX-License-Identifier: GPL-2.0-only
2 /*
3 * net/sched/sch_choke.c CHOKE scheduler
4 *
5 * Copyright (c) 2011 Stephen Hemminger <shemminger@vyatta.com>
6 * Copyright (c) 2011 Eric Dumazet <eric.dumazet@gmail.com>
7 */
8
9 #include <linux/module.h>
10 #include <linux/types.h>
11 #include <linux/kernel.h>
12 #include <linux/skbuff.h>
13 #include <linux/vmalloc.h>
14 #include <net/pkt_sched.h>
15 #include <net/pkt_cls.h>
16 #include <net/inet_ecn.h>
17 #include <net/red.h>
18 #include <net/flow_dissector.h>
19
20 /*
21 CHOKe stateless AQM for fair bandwidth allocation
22 =================================================
23
24 CHOKe (CHOose and Keep for responsive flows, CHOose and Kill for
25 unresponsive flows) is a variant of RED that penalizes misbehaving flows but
26 maintains no flow state. The difference from RED is an additional step
27 during the enqueuing process. If average queue size is over the
28 low threshold (qmin), a packet is chosen at random from the queue.
29 If both the new and chosen packet are from the same flow, both
30 are dropped. Unlike RED, CHOKe is not really a "classful" qdisc because it
31 needs to access packets in queue randomly. It has a minimal class
32 interface to allow overriding the builtin flow classifier with
33 filters.
34
35 Source:
36 R. Pan, B. Prabhakar, and K. Psounis, "CHOKe, A Stateless
37 Active Queue Management Scheme for Approximating Fair Bandwidth Allocation",
38 IEEE INFOCOM, 2000.
39
40 A. Tang, J. Wang, S. Low, "Understanding CHOKe: Throughput and Spatial
41 Characteristics", IEEE/ACM Transactions on Networking, 2004
42
43 */
44
45 /* Upper bound on size of sk_buff table (packets) */
46 #define CHOKE_MAX_QUEUE (128*1024 - 1)
47
48 struct choke_sched_data {
49 /* Parameters */
50 u32 limit;
51 unsigned char flags;
52
53 struct red_parms parms;
54
55 /* Variables */
56 struct red_vars vars;
57 struct {
58 u32 prob_drop; /* Early probability drops */
59 u32 prob_mark; /* Early probability marks */
60 u32 forced_drop; /* Forced drops, qavg > max_thresh */
61 u32 forced_mark; /* Forced marks, qavg > max_thresh */
62 u32 pdrop; /* Drops due to queue limits */
63 u32 matched; /* Drops to flow match */
64 } stats;
65
66 unsigned int head;
67 unsigned int tail;
68
69 unsigned int tab_mask; /* size - 1 */
70
71 struct sk_buff **tab;
72 };
73
74 /* number of elements in queue including holes */
choke_len(const struct choke_sched_data * q)75 static unsigned int choke_len(const struct choke_sched_data *q)
76 {
77 return (q->tail - q->head) & q->tab_mask;
78 }
79
80 /* Is ECN parameter configured */
use_ecn(const struct choke_sched_data * q)81 static int use_ecn(const struct choke_sched_data *q)
82 {
83 return q->flags & TC_RED_ECN;
84 }
85
86 /* Should packets over max just be dropped (versus marked) */
use_harddrop(const struct choke_sched_data * q)87 static int use_harddrop(const struct choke_sched_data *q)
88 {
89 return q->flags & TC_RED_HARDDROP;
90 }
91
92 /* Move head pointer forward to skip over holes */
choke_zap_head_holes(struct choke_sched_data * q)93 static void choke_zap_head_holes(struct choke_sched_data *q)
94 {
95 do {
96 q->head = (q->head + 1) & q->tab_mask;
97 if (q->head == q->tail)
98 break;
99 } while (q->tab[q->head] == NULL);
100 }
101
102 /* Move tail pointer backwards to reuse holes */
choke_zap_tail_holes(struct choke_sched_data * q)103 static void choke_zap_tail_holes(struct choke_sched_data *q)
104 {
105 do {
106 q->tail = (q->tail - 1) & q->tab_mask;
107 if (q->head == q->tail)
108 break;
109 } while (q->tab[q->tail] == NULL);
110 }
111
112 /* Drop packet from queue array by creating a "hole" */
choke_drop_by_idx(struct Qdisc * sch,unsigned int idx,struct sk_buff ** to_free)113 static void choke_drop_by_idx(struct Qdisc *sch, unsigned int idx,
114 struct sk_buff **to_free)
115 {
116 struct choke_sched_data *q = qdisc_priv(sch);
117 struct sk_buff *skb = q->tab[idx];
118
119 q->tab[idx] = NULL;
120
121 if (idx == q->head)
122 choke_zap_head_holes(q);
123 if (idx == q->tail)
124 choke_zap_tail_holes(q);
125
126 --sch->q.qlen;
127 qdisc_qstats_backlog_dec(sch, skb);
128 qdisc_tree_reduce_backlog(sch, 1, qdisc_pkt_len(skb));
129 qdisc_drop(skb, sch, to_free);
130 }
131
132 struct choke_skb_cb {
133 u8 keys_valid;
134 struct flow_keys_digest keys;
135 };
136
choke_skb_cb(const struct sk_buff * skb)137 static inline struct choke_skb_cb *choke_skb_cb(const struct sk_buff *skb)
138 {
139 qdisc_cb_private_validate(skb, sizeof(struct choke_skb_cb));
140 return (struct choke_skb_cb *)qdisc_skb_cb(skb)->data;
141 }
142
143 /*
144 * Compare flow of two packets
145 * Returns true only if source and destination address and port match.
146 * false for special cases
147 */
choke_match_flow(struct sk_buff * skb1,struct sk_buff * skb2)148 static bool choke_match_flow(struct sk_buff *skb1,
149 struct sk_buff *skb2)
150 {
151 struct flow_keys temp;
152
153 if (skb1->protocol != skb2->protocol)
154 return false;
155
156 if (!choke_skb_cb(skb1)->keys_valid) {
157 choke_skb_cb(skb1)->keys_valid = 1;
158 skb_flow_dissect_flow_keys(skb1, &temp, 0);
159 make_flow_keys_digest(&choke_skb_cb(skb1)->keys, &temp);
160 }
161
162 if (!choke_skb_cb(skb2)->keys_valid) {
163 choke_skb_cb(skb2)->keys_valid = 1;
164 skb_flow_dissect_flow_keys(skb2, &temp, 0);
165 make_flow_keys_digest(&choke_skb_cb(skb2)->keys, &temp);
166 }
167
168 return !memcmp(&choke_skb_cb(skb1)->keys,
169 &choke_skb_cb(skb2)->keys,
170 sizeof(choke_skb_cb(skb1)->keys));
171 }
172
173 /*
174 * Select a packet at random from queue
175 * HACK: since queue can have holes from previous deletion; retry several
176 * times to find a random skb but then just give up and return the head
177 * Will return NULL if queue is empty (q->head == q->tail)
178 */
choke_peek_random(const struct choke_sched_data * q,unsigned int * pidx)179 static struct sk_buff *choke_peek_random(const struct choke_sched_data *q,
180 unsigned int *pidx)
181 {
182 struct sk_buff *skb;
183 int retrys = 3;
184
185 do {
186 *pidx = (q->head + get_random_u32_below(choke_len(q))) & q->tab_mask;
187 skb = q->tab[*pidx];
188 if (skb)
189 return skb;
190 } while (--retrys > 0);
191
192 return q->tab[*pidx = q->head];
193 }
194
195 /*
196 * Compare new packet with random packet in queue
197 * returns true if matched and sets *pidx
198 */
choke_match_random(const struct choke_sched_data * q,struct sk_buff * nskb,unsigned int * pidx)199 static bool choke_match_random(const struct choke_sched_data *q,
200 struct sk_buff *nskb,
201 unsigned int *pidx)
202 {
203 struct sk_buff *oskb;
204
205 if (q->head == q->tail)
206 return false;
207
208 oskb = choke_peek_random(q, pidx);
209 return choke_match_flow(oskb, nskb);
210 }
211
choke_enqueue(struct sk_buff * skb,struct Qdisc * sch,struct sk_buff ** to_free)212 static int choke_enqueue(struct sk_buff *skb, struct Qdisc *sch,
213 struct sk_buff **to_free)
214 {
215 struct choke_sched_data *q = qdisc_priv(sch);
216 const struct red_parms *p = &q->parms;
217
218 choke_skb_cb(skb)->keys_valid = 0;
219 /* Compute average queue usage (see RED) */
220 q->vars.qavg = red_calc_qavg(p, &q->vars, sch->q.qlen);
221 if (red_is_idling(&q->vars))
222 red_end_of_idle_period(&q->vars);
223
224 /* Is queue small? */
225 if (q->vars.qavg <= p->qth_min)
226 q->vars.qcount = -1;
227 else {
228 unsigned int idx;
229
230 /* Draw a packet at random from queue and compare flow */
231 if (choke_match_random(q, skb, &idx)) {
232 WRITE_ONCE(q->stats.matched, q->stats.matched + 1);
233 choke_drop_by_idx(sch, idx, to_free);
234 goto congestion_drop;
235 }
236
237 /* Queue is large, always mark/drop */
238 if (q->vars.qavg > p->qth_max) {
239 q->vars.qcount = -1;
240
241 qdisc_qstats_overlimit(sch);
242 if (use_harddrop(q) || !use_ecn(q) ||
243 !INET_ECN_set_ce(skb)) {
244 WRITE_ONCE(q->stats.forced_drop,
245 q->stats.forced_drop + 1);
246 goto congestion_drop;
247 }
248
249 WRITE_ONCE(q->stats.forced_mark,
250 q->stats.forced_mark + 1);
251 } else if (++q->vars.qcount) {
252 if (red_mark_probability(p, &q->vars, q->vars.qavg)) {
253 q->vars.qcount = 0;
254 q->vars.qR = red_random(p);
255
256 qdisc_qstats_overlimit(sch);
257 if (!use_ecn(q) || !INET_ECN_set_ce(skb)) {
258 WRITE_ONCE(q->stats.prob_drop,
259 q->stats.prob_drop + 1);
260 goto congestion_drop;
261 }
262
263 WRITE_ONCE(q->stats.prob_mark,
264 q->stats.prob_mark + 1);
265 }
266 } else
267 q->vars.qR = red_random(p);
268 }
269
270 /* Admit new packet */
271 if (sch->q.qlen < q->limit) {
272 q->tab[q->tail] = skb;
273 q->tail = (q->tail + 1) & q->tab_mask;
274 ++sch->q.qlen;
275 qdisc_qstats_backlog_inc(sch, skb);
276 return NET_XMIT_SUCCESS;
277 }
278
279 WRITE_ONCE(q->stats.pdrop, q->stats.pdrop + 1);
280 return qdisc_drop(skb, sch, to_free);
281
282 congestion_drop:
283 qdisc_drop(skb, sch, to_free);
284 return NET_XMIT_CN;
285 }
286
choke_dequeue(struct Qdisc * sch)287 static struct sk_buff *choke_dequeue(struct Qdisc *sch)
288 {
289 struct choke_sched_data *q = qdisc_priv(sch);
290 struct sk_buff *skb;
291
292 if (q->head == q->tail) {
293 if (!red_is_idling(&q->vars))
294 red_start_of_idle_period(&q->vars);
295 return NULL;
296 }
297
298 skb = q->tab[q->head];
299 q->tab[q->head] = NULL;
300 choke_zap_head_holes(q);
301 --sch->q.qlen;
302 qdisc_qstats_backlog_dec(sch, skb);
303 qdisc_bstats_update(sch, skb);
304
305 return skb;
306 }
307
choke_reset(struct Qdisc * sch)308 static void choke_reset(struct Qdisc *sch)
309 {
310 struct choke_sched_data *q = qdisc_priv(sch);
311
312 while (q->head != q->tail) {
313 struct sk_buff *skb = q->tab[q->head];
314
315 q->head = (q->head + 1) & q->tab_mask;
316 if (!skb)
317 continue;
318 rtnl_qdisc_drop(skb, sch);
319 }
320
321 if (q->tab)
322 memset(q->tab, 0, (q->tab_mask + 1) * sizeof(struct sk_buff *));
323 q->head = q->tail = 0;
324 red_restart(&q->vars);
325 }
326
327 static const struct nla_policy choke_policy[TCA_CHOKE_MAX + 1] = {
328 [TCA_CHOKE_PARMS] = { .len = sizeof(struct tc_red_qopt) },
329 [TCA_CHOKE_STAB] = { .len = RED_STAB_SIZE },
330 [TCA_CHOKE_MAX_P] = { .type = NLA_U32 },
331 };
332
333
choke_free(void * addr)334 static void choke_free(void *addr)
335 {
336 kvfree(addr);
337 }
338
choke_change(struct Qdisc * sch,struct nlattr * opt,struct netlink_ext_ack * extack)339 static int choke_change(struct Qdisc *sch, struct nlattr *opt,
340 struct netlink_ext_ack *extack)
341 {
342 struct choke_sched_data *q = qdisc_priv(sch);
343 struct nlattr *tb[TCA_CHOKE_MAX + 1];
344 const struct tc_red_qopt *ctl;
345 int err;
346 struct sk_buff **old = NULL;
347 unsigned int mask;
348 u32 max_P;
349 u8 *stab;
350
351 if (opt == NULL)
352 return -EINVAL;
353
354 err = nla_parse_nested_deprecated(tb, TCA_CHOKE_MAX, opt,
355 choke_policy, NULL);
356 if (err < 0)
357 return err;
358
359 if (tb[TCA_CHOKE_PARMS] == NULL ||
360 tb[TCA_CHOKE_STAB] == NULL)
361 return -EINVAL;
362
363 max_P = nla_get_u32_default(tb[TCA_CHOKE_MAX_P], 0);
364
365 ctl = nla_data(tb[TCA_CHOKE_PARMS]);
366 stab = nla_data(tb[TCA_CHOKE_STAB]);
367 if (!red_check_params(ctl->qth_min, ctl->qth_max, ctl->Wlog, ctl->Scell_log, stab))
368 return -EINVAL;
369
370 if (ctl->limit > CHOKE_MAX_QUEUE)
371 return -EINVAL;
372
373 mask = roundup_pow_of_two(ctl->limit + 1) - 1;
374 if (mask != q->tab_mask) {
375 struct sk_buff **ntab;
376
377 ntab = kvzalloc_objs(struct sk_buff *, mask + 1);
378 if (!ntab)
379 return -ENOMEM;
380
381 sch_tree_lock(sch);
382 old = q->tab;
383 if (old) {
384 unsigned int oqlen = sch->q.qlen, tail = 0;
385 unsigned dropped = 0;
386
387 while (q->head != q->tail) {
388 struct sk_buff *skb = q->tab[q->head];
389
390 q->head = (q->head + 1) & q->tab_mask;
391 if (!skb)
392 continue;
393 if (tail < mask) {
394 ntab[tail++] = skb;
395 continue;
396 }
397 dropped += qdisc_pkt_len(skb);
398 qdisc_qstats_backlog_dec(sch, skb);
399 --sch->q.qlen;
400 rtnl_qdisc_drop(skb, sch);
401 }
402 qdisc_tree_reduce_backlog(sch, oqlen - sch->q.qlen, dropped);
403 q->head = 0;
404 q->tail = tail;
405 }
406
407 q->tab_mask = mask;
408 q->tab = ntab;
409 } else
410 sch_tree_lock(sch);
411
412 WRITE_ONCE(q->flags, ctl->flags);
413 WRITE_ONCE(q->limit, ctl->limit);
414
415 red_set_parms(&q->parms, ctl->qth_min, ctl->qth_max, ctl->Wlog,
416 ctl->Plog, ctl->Scell_log,
417 stab,
418 max_P);
419 red_set_vars(&q->vars);
420
421 if (q->head == q->tail)
422 red_end_of_idle_period(&q->vars);
423
424 sch_tree_unlock(sch);
425 choke_free(old);
426 return 0;
427 }
428
choke_init(struct Qdisc * sch,struct nlattr * opt,struct netlink_ext_ack * extack)429 static int choke_init(struct Qdisc *sch, struct nlattr *opt,
430 struct netlink_ext_ack *extack)
431 {
432 return choke_change(sch, opt, extack);
433 }
434
choke_dump(struct Qdisc * sch,struct sk_buff * skb)435 static int choke_dump(struct Qdisc *sch, struct sk_buff *skb)
436 {
437 struct choke_sched_data *q = qdisc_priv(sch);
438 u8 Wlog = READ_ONCE(q->parms.Wlog);
439 struct nlattr *opts = NULL;
440 struct tc_red_qopt opt = {
441 .limit = READ_ONCE(q->limit),
442 .flags = READ_ONCE(q->flags),
443 .qth_min = READ_ONCE(q->parms.qth_min) >> Wlog,
444 .qth_max = READ_ONCE(q->parms.qth_max) >> Wlog,
445 .Wlog = Wlog,
446 .Plog = READ_ONCE(q->parms.Plog),
447 .Scell_log = READ_ONCE(q->parms.Scell_log),
448 };
449
450 opts = nla_nest_start_noflag(skb, TCA_OPTIONS);
451 if (opts == NULL)
452 goto nla_put_failure;
453
454 if (nla_put(skb, TCA_CHOKE_PARMS, sizeof(opt), &opt) ||
455 nla_put_u32(skb, TCA_CHOKE_MAX_P, READ_ONCE(q->parms.max_P)))
456 goto nla_put_failure;
457 return nla_nest_end(skb, opts);
458
459 nla_put_failure:
460 nla_nest_cancel(skb, opts);
461 return -EMSGSIZE;
462 }
463
choke_dump_stats(struct Qdisc * sch,struct gnet_dump * d)464 static int choke_dump_stats(struct Qdisc *sch, struct gnet_dump *d)
465 {
466 struct choke_sched_data *q = qdisc_priv(sch);
467 struct tc_choke_xstats st = {
468 .early = READ_ONCE(q->stats.prob_drop) +
469 READ_ONCE(q->stats.forced_drop),
470 .marked = READ_ONCE(q->stats.prob_mark) +
471 READ_ONCE(q->stats.forced_mark),
472 .pdrop = READ_ONCE(q->stats.pdrop),
473 .matched = READ_ONCE(q->stats.matched),
474 };
475
476 return gnet_stats_copy_app(d, &st, sizeof(st));
477 }
478
choke_destroy(struct Qdisc * sch)479 static void choke_destroy(struct Qdisc *sch)
480 {
481 struct choke_sched_data *q = qdisc_priv(sch);
482
483 choke_free(q->tab);
484 }
485
choke_peek_head(struct Qdisc * sch)486 static struct sk_buff *choke_peek_head(struct Qdisc *sch)
487 {
488 struct choke_sched_data *q = qdisc_priv(sch);
489
490 return (q->head != q->tail) ? q->tab[q->head] : NULL;
491 }
492
493 static struct Qdisc_ops choke_qdisc_ops __read_mostly = {
494 .id = "choke",
495 .priv_size = sizeof(struct choke_sched_data),
496
497 .enqueue = choke_enqueue,
498 .dequeue = choke_dequeue,
499 .peek = choke_peek_head,
500 .init = choke_init,
501 .destroy = choke_destroy,
502 .reset = choke_reset,
503 .change = choke_change,
504 .dump = choke_dump,
505 .dump_stats = choke_dump_stats,
506 .owner = THIS_MODULE,
507 };
508 MODULE_ALIAS_NET_SCH("choke");
509
choke_module_init(void)510 static int __init choke_module_init(void)
511 {
512 return register_qdisc(&choke_qdisc_ops);
513 }
514
choke_module_exit(void)515 static void __exit choke_module_exit(void)
516 {
517 unregister_qdisc(&choke_qdisc_ops);
518 }
519
520 module_init(choke_module_init)
521 module_exit(choke_module_exit)
522
523 MODULE_LICENSE("GPL");
524 MODULE_DESCRIPTION("Choose and keep responsive flows scheduler");
525