1 /* 2 * net/sched/sch_htb.c Hierarchical token bucket, feed tree version 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: Martin Devera, <devik@cdi.cz> 10 * 11 * Credits (in time order) for older HTB versions: 12 * Stef Coene <stef.coene@docum.org> 13 * HTB support at LARTC mailing list 14 * Ondrej Kraus, <krauso@barr.cz> 15 * found missing INIT_QDISC(htb) 16 * Vladimir Smelhaus, Aamer Akhter, Bert Hubert 17 * helped a lot to locate nasty class stall bug 18 * Andi Kleen, Jamal Hadi, Bert Hubert 19 * code review and helpful comments on shaping 20 * Tomasz Wrona, <tw@eter.tym.pl> 21 * created test case so that I was able to fix nasty bug 22 * Wilfried Weissmann 23 * spotted bug in dequeue code and helped with fix 24 * Jiri Fojtasek 25 * fixed requeue routine 26 * and many others. thanks. 27 */ 28 #include <linux/module.h> 29 #include <linux/moduleparam.h> 30 #include <linux/types.h> 31 #include <linux/kernel.h> 32 #include <linux/string.h> 33 #include <linux/errno.h> 34 #include <linux/skbuff.h> 35 #include <linux/list.h> 36 #include <linux/compiler.h> 37 #include <linux/rbtree.h> 38 #include <linux/workqueue.h> 39 #include <linux/slab.h> 40 #include <net/netlink.h> 41 #include <net/sch_generic.h> 42 #include <net/pkt_sched.h> 43 #include <net/pkt_cls.h> 44 45 /* HTB algorithm. 46 Author: devik@cdi.cz 47 ======================================================================== 48 HTB is like TBF with multiple classes. It is also similar to CBQ because 49 it allows to assign priority to each class in hierarchy. 50 In fact it is another implementation of Floyd's formal sharing. 51 52 Levels: 53 Each class is assigned level. Leaf has ALWAYS level 0 and root 54 classes have level TC_HTB_MAXDEPTH-1. Interior nodes has level 55 one less than their parent. 56 */ 57 58 static int htb_hysteresis __read_mostly = 0; /* whether to use mode hysteresis for speedup */ 59 #define HTB_VER 0x30011 /* major must be matched with number suplied by TC as version */ 60 61 #if HTB_VER >> 16 != TC_HTB_PROTOVER 62 #error "Mismatched sch_htb.c and pkt_sch.h" 63 #endif 64 65 /* Module parameter and sysfs export */ 66 module_param (htb_hysteresis, int, 0640); 67 MODULE_PARM_DESC(htb_hysteresis, "Hysteresis mode, less CPU load, less accurate"); 68 69 static int htb_rate_est = 0; /* htb classes have a default rate estimator */ 70 module_param(htb_rate_est, int, 0640); 71 MODULE_PARM_DESC(htb_rate_est, "setup a default rate estimator (4sec 16sec) for htb classes"); 72 73 /* used internaly to keep status of single class */ 74 enum htb_cmode { 75 HTB_CANT_SEND, /* class can't send and can't borrow */ 76 HTB_MAY_BORROW, /* class can't send but may borrow */ 77 HTB_CAN_SEND /* class can send */ 78 }; 79 80 struct htb_prio { 81 union { 82 struct rb_root row; 83 struct rb_root feed; 84 }; 85 struct rb_node *ptr; 86 /* When class changes from state 1->2 and disconnects from 87 * parent's feed then we lost ptr value and start from the 88 * first child again. Here we store classid of the 89 * last valid ptr (used when ptr is NULL). 90 */ 91 u32 last_ptr_id; 92 }; 93 94 /* interior & leaf nodes; props specific to leaves are marked L: 95 * To reduce false sharing, place mostly read fields at beginning, 96 * and mostly written ones at the end. 97 */ 98 struct htb_class { 99 struct Qdisc_class_common common; 100 struct psched_ratecfg rate; 101 struct psched_ratecfg ceil; 102 s64 buffer, cbuffer;/* token bucket depth/rate */ 103 s64 mbuffer; /* max wait time */ 104 u32 prio; /* these two are used only by leaves... */ 105 int quantum; /* but stored for parent-to-leaf return */ 106 107 struct tcf_proto __rcu *filter_list; /* class attached filters */ 108 struct tcf_block *block; 109 int filter_cnt; 110 111 int level; /* our level (see above) */ 112 unsigned int children; 113 struct htb_class *parent; /* parent class */ 114 115 struct net_rate_estimator __rcu *rate_est; 116 117 /* 118 * Written often fields 119 */ 120 struct gnet_stats_basic_packed bstats; 121 struct tc_htb_xstats xstats; /* our special stats */ 122 123 /* token bucket parameters */ 124 s64 tokens, ctokens;/* current number of tokens */ 125 s64 t_c; /* checkpoint time */ 126 127 union { 128 struct htb_class_leaf { 129 int deficit[TC_HTB_MAXDEPTH]; 130 struct Qdisc *q; 131 } leaf; 132 struct htb_class_inner { 133 struct htb_prio clprio[TC_HTB_NUMPRIO]; 134 } inner; 135 } un; 136 s64 pq_key; 137 138 int prio_activity; /* for which prios are we active */ 139 enum htb_cmode cmode; /* current mode of the class */ 140 struct rb_node pq_node; /* node for event queue */ 141 struct rb_node node[TC_HTB_NUMPRIO]; /* node for self or feed tree */ 142 143 unsigned int drops ____cacheline_aligned_in_smp; 144 unsigned int overlimits; 145 }; 146 147 struct htb_level { 148 struct rb_root wait_pq; 149 struct htb_prio hprio[TC_HTB_NUMPRIO]; 150 }; 151 152 struct htb_sched { 153 struct Qdisc_class_hash clhash; 154 int defcls; /* class where unclassified flows go to */ 155 int rate2quantum; /* quant = rate / rate2quantum */ 156 157 /* filters for qdisc itself */ 158 struct tcf_proto __rcu *filter_list; 159 struct tcf_block *block; 160 161 #define HTB_WARN_TOOMANYEVENTS 0x1 162 unsigned int warned; /* only one warning */ 163 int direct_qlen; 164 struct work_struct work; 165 166 /* non shaped skbs; let them go directly thru */ 167 struct qdisc_skb_head direct_queue; 168 long direct_pkts; 169 170 struct qdisc_watchdog watchdog; 171 172 s64 now; /* cached dequeue time */ 173 174 /* time of nearest event per level (row) */ 175 s64 near_ev_cache[TC_HTB_MAXDEPTH]; 176 177 int row_mask[TC_HTB_MAXDEPTH]; 178 179 struct htb_level hlevel[TC_HTB_MAXDEPTH]; 180 }; 181 182 /* find class in global hash table using given handle */ 183 static inline struct htb_class *htb_find(u32 handle, struct Qdisc *sch) 184 { 185 struct htb_sched *q = qdisc_priv(sch); 186 struct Qdisc_class_common *clc; 187 188 clc = qdisc_class_find(&q->clhash, handle); 189 if (clc == NULL) 190 return NULL; 191 return container_of(clc, struct htb_class, common); 192 } 193 194 static unsigned long htb_search(struct Qdisc *sch, u32 handle) 195 { 196 return (unsigned long)htb_find(handle, sch); 197 } 198 /** 199 * htb_classify - classify a packet into class 200 * 201 * It returns NULL if the packet should be dropped or -1 if the packet 202 * should be passed directly thru. In all other cases leaf class is returned. 203 * We allow direct class selection by classid in priority. The we examine 204 * filters in qdisc and in inner nodes (if higher filter points to the inner 205 * node). If we end up with classid MAJOR:0 we enqueue the skb into special 206 * internal fifo (direct). These packets then go directly thru. If we still 207 * have no valid leaf we try to use MAJOR:default leaf. It still unsuccessful 208 * then finish and return direct queue. 209 */ 210 #define HTB_DIRECT ((struct htb_class *)-1L) 211 212 static struct htb_class *htb_classify(struct sk_buff *skb, struct Qdisc *sch, 213 int *qerr) 214 { 215 struct htb_sched *q = qdisc_priv(sch); 216 struct htb_class *cl; 217 struct tcf_result res; 218 struct tcf_proto *tcf; 219 int result; 220 221 /* allow to select class by setting skb->priority to valid classid; 222 * note that nfmark can be used too by attaching filter fw with no 223 * rules in it 224 */ 225 if (skb->priority == sch->handle) 226 return HTB_DIRECT; /* X:0 (direct flow) selected */ 227 cl = htb_find(skb->priority, sch); 228 if (cl) { 229 if (cl->level == 0) 230 return cl; 231 /* Start with inner filter chain if a non-leaf class is selected */ 232 tcf = rcu_dereference_bh(cl->filter_list); 233 } else { 234 tcf = rcu_dereference_bh(q->filter_list); 235 } 236 237 *qerr = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; 238 while (tcf && (result = tcf_classify(skb, tcf, &res, false)) >= 0) { 239 #ifdef CONFIG_NET_CLS_ACT 240 switch (result) { 241 case TC_ACT_QUEUED: 242 case TC_ACT_STOLEN: 243 case TC_ACT_TRAP: 244 *qerr = NET_XMIT_SUCCESS | __NET_XMIT_STOLEN; 245 /* fall through */ 246 case TC_ACT_SHOT: 247 return NULL; 248 } 249 #endif 250 cl = (void *)res.class; 251 if (!cl) { 252 if (res.classid == sch->handle) 253 return HTB_DIRECT; /* X:0 (direct flow) */ 254 cl = htb_find(res.classid, sch); 255 if (!cl) 256 break; /* filter selected invalid classid */ 257 } 258 if (!cl->level) 259 return cl; /* we hit leaf; return it */ 260 261 /* we have got inner class; apply inner filter chain */ 262 tcf = rcu_dereference_bh(cl->filter_list); 263 } 264 /* classification failed; try to use default class */ 265 cl = htb_find(TC_H_MAKE(TC_H_MAJ(sch->handle), q->defcls), sch); 266 if (!cl || cl->level) 267 return HTB_DIRECT; /* bad default .. this is safe bet */ 268 return cl; 269 } 270 271 /** 272 * htb_add_to_id_tree - adds class to the round robin list 273 * 274 * Routine adds class to the list (actually tree) sorted by classid. 275 * Make sure that class is not already on such list for given prio. 276 */ 277 static void htb_add_to_id_tree(struct rb_root *root, 278 struct htb_class *cl, int prio) 279 { 280 struct rb_node **p = &root->rb_node, *parent = NULL; 281 282 while (*p) { 283 struct htb_class *c; 284 parent = *p; 285 c = rb_entry(parent, struct htb_class, node[prio]); 286 287 if (cl->common.classid > c->common.classid) 288 p = &parent->rb_right; 289 else 290 p = &parent->rb_left; 291 } 292 rb_link_node(&cl->node[prio], parent, p); 293 rb_insert_color(&cl->node[prio], root); 294 } 295 296 /** 297 * htb_add_to_wait_tree - adds class to the event queue with delay 298 * 299 * The class is added to priority event queue to indicate that class will 300 * change its mode in cl->pq_key microseconds. Make sure that class is not 301 * already in the queue. 302 */ 303 static void htb_add_to_wait_tree(struct htb_sched *q, 304 struct htb_class *cl, s64 delay) 305 { 306 struct rb_node **p = &q->hlevel[cl->level].wait_pq.rb_node, *parent = NULL; 307 308 cl->pq_key = q->now + delay; 309 if (cl->pq_key == q->now) 310 cl->pq_key++; 311 312 /* update the nearest event cache */ 313 if (q->near_ev_cache[cl->level] > cl->pq_key) 314 q->near_ev_cache[cl->level] = cl->pq_key; 315 316 while (*p) { 317 struct htb_class *c; 318 parent = *p; 319 c = rb_entry(parent, struct htb_class, pq_node); 320 if (cl->pq_key >= c->pq_key) 321 p = &parent->rb_right; 322 else 323 p = &parent->rb_left; 324 } 325 rb_link_node(&cl->pq_node, parent, p); 326 rb_insert_color(&cl->pq_node, &q->hlevel[cl->level].wait_pq); 327 } 328 329 /** 330 * htb_next_rb_node - finds next node in binary tree 331 * 332 * When we are past last key we return NULL. 333 * Average complexity is 2 steps per call. 334 */ 335 static inline void htb_next_rb_node(struct rb_node **n) 336 { 337 *n = rb_next(*n); 338 } 339 340 /** 341 * htb_add_class_to_row - add class to its row 342 * 343 * The class is added to row at priorities marked in mask. 344 * It does nothing if mask == 0. 345 */ 346 static inline void htb_add_class_to_row(struct htb_sched *q, 347 struct htb_class *cl, int mask) 348 { 349 q->row_mask[cl->level] |= mask; 350 while (mask) { 351 int prio = ffz(~mask); 352 mask &= ~(1 << prio); 353 htb_add_to_id_tree(&q->hlevel[cl->level].hprio[prio].row, cl, prio); 354 } 355 } 356 357 /* If this triggers, it is a bug in this code, but it need not be fatal */ 358 static void htb_safe_rb_erase(struct rb_node *rb, struct rb_root *root) 359 { 360 if (RB_EMPTY_NODE(rb)) { 361 WARN_ON(1); 362 } else { 363 rb_erase(rb, root); 364 RB_CLEAR_NODE(rb); 365 } 366 } 367 368 369 /** 370 * htb_remove_class_from_row - removes class from its row 371 * 372 * The class is removed from row at priorities marked in mask. 373 * It does nothing if mask == 0. 374 */ 375 static inline void htb_remove_class_from_row(struct htb_sched *q, 376 struct htb_class *cl, int mask) 377 { 378 int m = 0; 379 struct htb_level *hlevel = &q->hlevel[cl->level]; 380 381 while (mask) { 382 int prio = ffz(~mask); 383 struct htb_prio *hprio = &hlevel->hprio[prio]; 384 385 mask &= ~(1 << prio); 386 if (hprio->ptr == cl->node + prio) 387 htb_next_rb_node(&hprio->ptr); 388 389 htb_safe_rb_erase(cl->node + prio, &hprio->row); 390 if (!hprio->row.rb_node) 391 m |= 1 << prio; 392 } 393 q->row_mask[cl->level] &= ~m; 394 } 395 396 /** 397 * htb_activate_prios - creates active classe's feed chain 398 * 399 * The class is connected to ancestors and/or appropriate rows 400 * for priorities it is participating on. cl->cmode must be new 401 * (activated) mode. It does nothing if cl->prio_activity == 0. 402 */ 403 static void htb_activate_prios(struct htb_sched *q, struct htb_class *cl) 404 { 405 struct htb_class *p = cl->parent; 406 long m, mask = cl->prio_activity; 407 408 while (cl->cmode == HTB_MAY_BORROW && p && mask) { 409 m = mask; 410 while (m) { 411 int prio = ffz(~m); 412 m &= ~(1 << prio); 413 414 if (p->un.inner.clprio[prio].feed.rb_node) 415 /* parent already has its feed in use so that 416 * reset bit in mask as parent is already ok 417 */ 418 mask &= ~(1 << prio); 419 420 htb_add_to_id_tree(&p->un.inner.clprio[prio].feed, cl, prio); 421 } 422 p->prio_activity |= mask; 423 cl = p; 424 p = cl->parent; 425 426 } 427 if (cl->cmode == HTB_CAN_SEND && mask) 428 htb_add_class_to_row(q, cl, mask); 429 } 430 431 /** 432 * htb_deactivate_prios - remove class from feed chain 433 * 434 * cl->cmode must represent old mode (before deactivation). It does 435 * nothing if cl->prio_activity == 0. Class is removed from all feed 436 * chains and rows. 437 */ 438 static void htb_deactivate_prios(struct htb_sched *q, struct htb_class *cl) 439 { 440 struct htb_class *p = cl->parent; 441 long m, mask = cl->prio_activity; 442 443 while (cl->cmode == HTB_MAY_BORROW && p && mask) { 444 m = mask; 445 mask = 0; 446 while (m) { 447 int prio = ffz(~m); 448 m &= ~(1 << prio); 449 450 if (p->un.inner.clprio[prio].ptr == cl->node + prio) { 451 /* we are removing child which is pointed to from 452 * parent feed - forget the pointer but remember 453 * classid 454 */ 455 p->un.inner.clprio[prio].last_ptr_id = cl->common.classid; 456 p->un.inner.clprio[prio].ptr = NULL; 457 } 458 459 htb_safe_rb_erase(cl->node + prio, 460 &p->un.inner.clprio[prio].feed); 461 462 if (!p->un.inner.clprio[prio].feed.rb_node) 463 mask |= 1 << prio; 464 } 465 466 p->prio_activity &= ~mask; 467 cl = p; 468 p = cl->parent; 469 470 } 471 if (cl->cmode == HTB_CAN_SEND && mask) 472 htb_remove_class_from_row(q, cl, mask); 473 } 474 475 static inline s64 htb_lowater(const struct htb_class *cl) 476 { 477 if (htb_hysteresis) 478 return cl->cmode != HTB_CANT_SEND ? -cl->cbuffer : 0; 479 else 480 return 0; 481 } 482 static inline s64 htb_hiwater(const struct htb_class *cl) 483 { 484 if (htb_hysteresis) 485 return cl->cmode == HTB_CAN_SEND ? -cl->buffer : 0; 486 else 487 return 0; 488 } 489 490 491 /** 492 * htb_class_mode - computes and returns current class mode 493 * 494 * It computes cl's mode at time cl->t_c+diff and returns it. If mode 495 * is not HTB_CAN_SEND then cl->pq_key is updated to time difference 496 * from now to time when cl will change its state. 497 * Also it is worth to note that class mode doesn't change simply 498 * at cl->{c,}tokens == 0 but there can rather be hysteresis of 499 * 0 .. -cl->{c,}buffer range. It is meant to limit number of 500 * mode transitions per time unit. The speed gain is about 1/6. 501 */ 502 static inline enum htb_cmode 503 htb_class_mode(struct htb_class *cl, s64 *diff) 504 { 505 s64 toks; 506 507 if ((toks = (cl->ctokens + *diff)) < htb_lowater(cl)) { 508 *diff = -toks; 509 return HTB_CANT_SEND; 510 } 511 512 if ((toks = (cl->tokens + *diff)) >= htb_hiwater(cl)) 513 return HTB_CAN_SEND; 514 515 *diff = -toks; 516 return HTB_MAY_BORROW; 517 } 518 519 /** 520 * htb_change_class_mode - changes classe's mode 521 * 522 * This should be the only way how to change classe's mode under normal 523 * cirsumstances. Routine will update feed lists linkage, change mode 524 * and add class to the wait event queue if appropriate. New mode should 525 * be different from old one and cl->pq_key has to be valid if changing 526 * to mode other than HTB_CAN_SEND (see htb_add_to_wait_tree). 527 */ 528 static void 529 htb_change_class_mode(struct htb_sched *q, struct htb_class *cl, s64 *diff) 530 { 531 enum htb_cmode new_mode = htb_class_mode(cl, diff); 532 533 if (new_mode == cl->cmode) 534 return; 535 536 if (new_mode == HTB_CANT_SEND) 537 cl->overlimits++; 538 539 if (cl->prio_activity) { /* not necessary: speed optimization */ 540 if (cl->cmode != HTB_CANT_SEND) 541 htb_deactivate_prios(q, cl); 542 cl->cmode = new_mode; 543 if (new_mode != HTB_CANT_SEND) 544 htb_activate_prios(q, cl); 545 } else 546 cl->cmode = new_mode; 547 } 548 549 /** 550 * htb_activate - inserts leaf cl into appropriate active feeds 551 * 552 * Routine learns (new) priority of leaf and activates feed chain 553 * for the prio. It can be called on already active leaf safely. 554 * It also adds leaf into droplist. 555 */ 556 static inline void htb_activate(struct htb_sched *q, struct htb_class *cl) 557 { 558 WARN_ON(cl->level || !cl->un.leaf.q || !cl->un.leaf.q->q.qlen); 559 560 if (!cl->prio_activity) { 561 cl->prio_activity = 1 << cl->prio; 562 htb_activate_prios(q, cl); 563 } 564 } 565 566 /** 567 * htb_deactivate - remove leaf cl from active feeds 568 * 569 * Make sure that leaf is active. In the other words it can't be called 570 * with non-active leaf. It also removes class from the drop list. 571 */ 572 static inline void htb_deactivate(struct htb_sched *q, struct htb_class *cl) 573 { 574 WARN_ON(!cl->prio_activity); 575 576 htb_deactivate_prios(q, cl); 577 cl->prio_activity = 0; 578 } 579 580 static void htb_enqueue_tail(struct sk_buff *skb, struct Qdisc *sch, 581 struct qdisc_skb_head *qh) 582 { 583 struct sk_buff *last = qh->tail; 584 585 if (last) { 586 skb->next = NULL; 587 last->next = skb; 588 qh->tail = skb; 589 } else { 590 qh->tail = skb; 591 qh->head = skb; 592 } 593 qh->qlen++; 594 } 595 596 static int htb_enqueue(struct sk_buff *skb, struct Qdisc *sch, 597 struct sk_buff **to_free) 598 { 599 int uninitialized_var(ret); 600 struct htb_sched *q = qdisc_priv(sch); 601 struct htb_class *cl = htb_classify(skb, sch, &ret); 602 603 if (cl == HTB_DIRECT) { 604 /* enqueue to helper queue */ 605 if (q->direct_queue.qlen < q->direct_qlen) { 606 htb_enqueue_tail(skb, sch, &q->direct_queue); 607 q->direct_pkts++; 608 } else { 609 return qdisc_drop(skb, sch, to_free); 610 } 611 #ifdef CONFIG_NET_CLS_ACT 612 } else if (!cl) { 613 if (ret & __NET_XMIT_BYPASS) 614 qdisc_qstats_drop(sch); 615 __qdisc_drop(skb, to_free); 616 return ret; 617 #endif 618 } else if ((ret = qdisc_enqueue(skb, cl->un.leaf.q, 619 to_free)) != NET_XMIT_SUCCESS) { 620 if (net_xmit_drop_count(ret)) { 621 qdisc_qstats_drop(sch); 622 cl->drops++; 623 } 624 return ret; 625 } else { 626 htb_activate(q, cl); 627 } 628 629 qdisc_qstats_backlog_inc(sch, skb); 630 sch->q.qlen++; 631 return NET_XMIT_SUCCESS; 632 } 633 634 static inline void htb_accnt_tokens(struct htb_class *cl, int bytes, s64 diff) 635 { 636 s64 toks = diff + cl->tokens; 637 638 if (toks > cl->buffer) 639 toks = cl->buffer; 640 toks -= (s64) psched_l2t_ns(&cl->rate, bytes); 641 if (toks <= -cl->mbuffer) 642 toks = 1 - cl->mbuffer; 643 644 cl->tokens = toks; 645 } 646 647 static inline void htb_accnt_ctokens(struct htb_class *cl, int bytes, s64 diff) 648 { 649 s64 toks = diff + cl->ctokens; 650 651 if (toks > cl->cbuffer) 652 toks = cl->cbuffer; 653 toks -= (s64) psched_l2t_ns(&cl->ceil, bytes); 654 if (toks <= -cl->mbuffer) 655 toks = 1 - cl->mbuffer; 656 657 cl->ctokens = toks; 658 } 659 660 /** 661 * htb_charge_class - charges amount "bytes" to leaf and ancestors 662 * 663 * Routine assumes that packet "bytes" long was dequeued from leaf cl 664 * borrowing from "level". It accounts bytes to ceil leaky bucket for 665 * leaf and all ancestors and to rate bucket for ancestors at levels 666 * "level" and higher. It also handles possible change of mode resulting 667 * from the update. Note that mode can also increase here (MAY_BORROW to 668 * CAN_SEND) because we can use more precise clock that event queue here. 669 * In such case we remove class from event queue first. 670 */ 671 static void htb_charge_class(struct htb_sched *q, struct htb_class *cl, 672 int level, struct sk_buff *skb) 673 { 674 int bytes = qdisc_pkt_len(skb); 675 enum htb_cmode old_mode; 676 s64 diff; 677 678 while (cl) { 679 diff = min_t(s64, q->now - cl->t_c, cl->mbuffer); 680 if (cl->level >= level) { 681 if (cl->level == level) 682 cl->xstats.lends++; 683 htb_accnt_tokens(cl, bytes, diff); 684 } else { 685 cl->xstats.borrows++; 686 cl->tokens += diff; /* we moved t_c; update tokens */ 687 } 688 htb_accnt_ctokens(cl, bytes, diff); 689 cl->t_c = q->now; 690 691 old_mode = cl->cmode; 692 diff = 0; 693 htb_change_class_mode(q, cl, &diff); 694 if (old_mode != cl->cmode) { 695 if (old_mode != HTB_CAN_SEND) 696 htb_safe_rb_erase(&cl->pq_node, &q->hlevel[cl->level].wait_pq); 697 if (cl->cmode != HTB_CAN_SEND) 698 htb_add_to_wait_tree(q, cl, diff); 699 } 700 701 /* update basic stats except for leaves which are already updated */ 702 if (cl->level) 703 bstats_update(&cl->bstats, skb); 704 705 cl = cl->parent; 706 } 707 } 708 709 /** 710 * htb_do_events - make mode changes to classes at the level 711 * 712 * Scans event queue for pending events and applies them. Returns time of 713 * next pending event (0 for no event in pq, q->now for too many events). 714 * Note: Applied are events whose have cl->pq_key <= q->now. 715 */ 716 static s64 htb_do_events(struct htb_sched *q, const int level, 717 unsigned long start) 718 { 719 /* don't run for longer than 2 jiffies; 2 is used instead of 720 * 1 to simplify things when jiffy is going to be incremented 721 * too soon 722 */ 723 unsigned long stop_at = start + 2; 724 struct rb_root *wait_pq = &q->hlevel[level].wait_pq; 725 726 while (time_before(jiffies, stop_at)) { 727 struct htb_class *cl; 728 s64 diff; 729 struct rb_node *p = rb_first(wait_pq); 730 731 if (!p) 732 return 0; 733 734 cl = rb_entry(p, struct htb_class, pq_node); 735 if (cl->pq_key > q->now) 736 return cl->pq_key; 737 738 htb_safe_rb_erase(p, wait_pq); 739 diff = min_t(s64, q->now - cl->t_c, cl->mbuffer); 740 htb_change_class_mode(q, cl, &diff); 741 if (cl->cmode != HTB_CAN_SEND) 742 htb_add_to_wait_tree(q, cl, diff); 743 } 744 745 /* too much load - let's continue after a break for scheduling */ 746 if (!(q->warned & HTB_WARN_TOOMANYEVENTS)) { 747 pr_warn("htb: too many events!\n"); 748 q->warned |= HTB_WARN_TOOMANYEVENTS; 749 } 750 751 return q->now; 752 } 753 754 /* Returns class->node+prio from id-tree where classe's id is >= id. NULL 755 * is no such one exists. 756 */ 757 static struct rb_node *htb_id_find_next_upper(int prio, struct rb_node *n, 758 u32 id) 759 { 760 struct rb_node *r = NULL; 761 while (n) { 762 struct htb_class *cl = 763 rb_entry(n, struct htb_class, node[prio]); 764 765 if (id > cl->common.classid) { 766 n = n->rb_right; 767 } else if (id < cl->common.classid) { 768 r = n; 769 n = n->rb_left; 770 } else { 771 return n; 772 } 773 } 774 return r; 775 } 776 777 /** 778 * htb_lookup_leaf - returns next leaf class in DRR order 779 * 780 * Find leaf where current feed pointers points to. 781 */ 782 static struct htb_class *htb_lookup_leaf(struct htb_prio *hprio, const int prio) 783 { 784 int i; 785 struct { 786 struct rb_node *root; 787 struct rb_node **pptr; 788 u32 *pid; 789 } stk[TC_HTB_MAXDEPTH], *sp = stk; 790 791 BUG_ON(!hprio->row.rb_node); 792 sp->root = hprio->row.rb_node; 793 sp->pptr = &hprio->ptr; 794 sp->pid = &hprio->last_ptr_id; 795 796 for (i = 0; i < 65535; i++) { 797 if (!*sp->pptr && *sp->pid) { 798 /* ptr was invalidated but id is valid - try to recover 799 * the original or next ptr 800 */ 801 *sp->pptr = 802 htb_id_find_next_upper(prio, sp->root, *sp->pid); 803 } 804 *sp->pid = 0; /* ptr is valid now so that remove this hint as it 805 * can become out of date quickly 806 */ 807 if (!*sp->pptr) { /* we are at right end; rewind & go up */ 808 *sp->pptr = sp->root; 809 while ((*sp->pptr)->rb_left) 810 *sp->pptr = (*sp->pptr)->rb_left; 811 if (sp > stk) { 812 sp--; 813 if (!*sp->pptr) { 814 WARN_ON(1); 815 return NULL; 816 } 817 htb_next_rb_node(sp->pptr); 818 } 819 } else { 820 struct htb_class *cl; 821 struct htb_prio *clp; 822 823 cl = rb_entry(*sp->pptr, struct htb_class, node[prio]); 824 if (!cl->level) 825 return cl; 826 clp = &cl->un.inner.clprio[prio]; 827 (++sp)->root = clp->feed.rb_node; 828 sp->pptr = &clp->ptr; 829 sp->pid = &clp->last_ptr_id; 830 } 831 } 832 WARN_ON(1); 833 return NULL; 834 } 835 836 /* dequeues packet at given priority and level; call only if 837 * you are sure that there is active class at prio/level 838 */ 839 static struct sk_buff *htb_dequeue_tree(struct htb_sched *q, const int prio, 840 const int level) 841 { 842 struct sk_buff *skb = NULL; 843 struct htb_class *cl, *start; 844 struct htb_level *hlevel = &q->hlevel[level]; 845 struct htb_prio *hprio = &hlevel->hprio[prio]; 846 847 /* look initial class up in the row */ 848 start = cl = htb_lookup_leaf(hprio, prio); 849 850 do { 851 next: 852 if (unlikely(!cl)) 853 return NULL; 854 855 /* class can be empty - it is unlikely but can be true if leaf 856 * qdisc drops packets in enqueue routine or if someone used 857 * graft operation on the leaf since last dequeue; 858 * simply deactivate and skip such class 859 */ 860 if (unlikely(cl->un.leaf.q->q.qlen == 0)) { 861 struct htb_class *next; 862 htb_deactivate(q, cl); 863 864 /* row/level might become empty */ 865 if ((q->row_mask[level] & (1 << prio)) == 0) 866 return NULL; 867 868 next = htb_lookup_leaf(hprio, prio); 869 870 if (cl == start) /* fix start if we just deleted it */ 871 start = next; 872 cl = next; 873 goto next; 874 } 875 876 skb = cl->un.leaf.q->dequeue(cl->un.leaf.q); 877 if (likely(skb != NULL)) 878 break; 879 880 qdisc_warn_nonwc("htb", cl->un.leaf.q); 881 htb_next_rb_node(level ? &cl->parent->un.inner.clprio[prio].ptr: 882 &q->hlevel[0].hprio[prio].ptr); 883 cl = htb_lookup_leaf(hprio, prio); 884 885 } while (cl != start); 886 887 if (likely(skb != NULL)) { 888 bstats_update(&cl->bstats, skb); 889 cl->un.leaf.deficit[level] -= qdisc_pkt_len(skb); 890 if (cl->un.leaf.deficit[level] < 0) { 891 cl->un.leaf.deficit[level] += cl->quantum; 892 htb_next_rb_node(level ? &cl->parent->un.inner.clprio[prio].ptr : 893 &q->hlevel[0].hprio[prio].ptr); 894 } 895 /* this used to be after charge_class but this constelation 896 * gives us slightly better performance 897 */ 898 if (!cl->un.leaf.q->q.qlen) 899 htb_deactivate(q, cl); 900 htb_charge_class(q, cl, level, skb); 901 } 902 return skb; 903 } 904 905 static struct sk_buff *htb_dequeue(struct Qdisc *sch) 906 { 907 struct sk_buff *skb; 908 struct htb_sched *q = qdisc_priv(sch); 909 int level; 910 s64 next_event; 911 unsigned long start_at; 912 913 /* try to dequeue direct packets as high prio (!) to minimize cpu work */ 914 skb = __qdisc_dequeue_head(&q->direct_queue); 915 if (skb != NULL) { 916 ok: 917 qdisc_bstats_update(sch, skb); 918 qdisc_qstats_backlog_dec(sch, skb); 919 sch->q.qlen--; 920 return skb; 921 } 922 923 if (!sch->q.qlen) 924 goto fin; 925 q->now = ktime_get_ns(); 926 start_at = jiffies; 927 928 next_event = q->now + 5LLU * NSEC_PER_SEC; 929 930 for (level = 0; level < TC_HTB_MAXDEPTH; level++) { 931 /* common case optimization - skip event handler quickly */ 932 int m; 933 s64 event = q->near_ev_cache[level]; 934 935 if (q->now >= event) { 936 event = htb_do_events(q, level, start_at); 937 if (!event) 938 event = q->now + NSEC_PER_SEC; 939 q->near_ev_cache[level] = event; 940 } 941 942 if (next_event > event) 943 next_event = event; 944 945 m = ~q->row_mask[level]; 946 while (m != (int)(-1)) { 947 int prio = ffz(m); 948 949 m |= 1 << prio; 950 skb = htb_dequeue_tree(q, prio, level); 951 if (likely(skb != NULL)) 952 goto ok; 953 } 954 } 955 qdisc_qstats_overlimit(sch); 956 if (likely(next_event > q->now)) 957 qdisc_watchdog_schedule_ns(&q->watchdog, next_event); 958 else 959 schedule_work(&q->work); 960 fin: 961 return skb; 962 } 963 964 /* reset all classes */ 965 /* always caled under BH & queue lock */ 966 static void htb_reset(struct Qdisc *sch) 967 { 968 struct htb_sched *q = qdisc_priv(sch); 969 struct htb_class *cl; 970 unsigned int i; 971 972 for (i = 0; i < q->clhash.hashsize; i++) { 973 hlist_for_each_entry(cl, &q->clhash.hash[i], common.hnode) { 974 if (cl->level) 975 memset(&cl->un.inner, 0, sizeof(cl->un.inner)); 976 else { 977 if (cl->un.leaf.q) 978 qdisc_reset(cl->un.leaf.q); 979 } 980 cl->prio_activity = 0; 981 cl->cmode = HTB_CAN_SEND; 982 } 983 } 984 qdisc_watchdog_cancel(&q->watchdog); 985 __qdisc_reset_queue(&q->direct_queue); 986 sch->q.qlen = 0; 987 sch->qstats.backlog = 0; 988 memset(q->hlevel, 0, sizeof(q->hlevel)); 989 memset(q->row_mask, 0, sizeof(q->row_mask)); 990 } 991 992 static const struct nla_policy htb_policy[TCA_HTB_MAX + 1] = { 993 [TCA_HTB_PARMS] = { .len = sizeof(struct tc_htb_opt) }, 994 [TCA_HTB_INIT] = { .len = sizeof(struct tc_htb_glob) }, 995 [TCA_HTB_CTAB] = { .type = NLA_BINARY, .len = TC_RTAB_SIZE }, 996 [TCA_HTB_RTAB] = { .type = NLA_BINARY, .len = TC_RTAB_SIZE }, 997 [TCA_HTB_DIRECT_QLEN] = { .type = NLA_U32 }, 998 [TCA_HTB_RATE64] = { .type = NLA_U64 }, 999 [TCA_HTB_CEIL64] = { .type = NLA_U64 }, 1000 }; 1001 1002 static void htb_work_func(struct work_struct *work) 1003 { 1004 struct htb_sched *q = container_of(work, struct htb_sched, work); 1005 struct Qdisc *sch = q->watchdog.qdisc; 1006 1007 rcu_read_lock(); 1008 __netif_schedule(qdisc_root(sch)); 1009 rcu_read_unlock(); 1010 } 1011 1012 static int htb_init(struct Qdisc *sch, struct nlattr *opt, 1013 struct netlink_ext_ack *extack) 1014 { 1015 struct htb_sched *q = qdisc_priv(sch); 1016 struct nlattr *tb[TCA_HTB_MAX + 1]; 1017 struct tc_htb_glob *gopt; 1018 int err; 1019 1020 qdisc_watchdog_init(&q->watchdog, sch); 1021 INIT_WORK(&q->work, htb_work_func); 1022 1023 if (!opt) 1024 return -EINVAL; 1025 1026 err = tcf_block_get(&q->block, &q->filter_list, sch, extack); 1027 if (err) 1028 return err; 1029 1030 err = nla_parse_nested(tb, TCA_HTB_MAX, opt, htb_policy, NULL); 1031 if (err < 0) 1032 return err; 1033 1034 if (!tb[TCA_HTB_INIT]) 1035 return -EINVAL; 1036 1037 gopt = nla_data(tb[TCA_HTB_INIT]); 1038 if (gopt->version != HTB_VER >> 16) 1039 return -EINVAL; 1040 1041 err = qdisc_class_hash_init(&q->clhash); 1042 if (err < 0) 1043 return err; 1044 1045 qdisc_skb_head_init(&q->direct_queue); 1046 1047 if (tb[TCA_HTB_DIRECT_QLEN]) 1048 q->direct_qlen = nla_get_u32(tb[TCA_HTB_DIRECT_QLEN]); 1049 else 1050 q->direct_qlen = qdisc_dev(sch)->tx_queue_len; 1051 1052 if ((q->rate2quantum = gopt->rate2quantum) < 1) 1053 q->rate2quantum = 1; 1054 q->defcls = gopt->defcls; 1055 1056 return 0; 1057 } 1058 1059 static int htb_dump(struct Qdisc *sch, struct sk_buff *skb) 1060 { 1061 struct htb_sched *q = qdisc_priv(sch); 1062 struct nlattr *nest; 1063 struct tc_htb_glob gopt; 1064 1065 /* Its safe to not acquire qdisc lock. As we hold RTNL, 1066 * no change can happen on the qdisc parameters. 1067 */ 1068 1069 gopt.direct_pkts = q->direct_pkts; 1070 gopt.version = HTB_VER; 1071 gopt.rate2quantum = q->rate2quantum; 1072 gopt.defcls = q->defcls; 1073 gopt.debug = 0; 1074 1075 nest = nla_nest_start(skb, TCA_OPTIONS); 1076 if (nest == NULL) 1077 goto nla_put_failure; 1078 if (nla_put(skb, TCA_HTB_INIT, sizeof(gopt), &gopt) || 1079 nla_put_u32(skb, TCA_HTB_DIRECT_QLEN, q->direct_qlen)) 1080 goto nla_put_failure; 1081 1082 return nla_nest_end(skb, nest); 1083 1084 nla_put_failure: 1085 nla_nest_cancel(skb, nest); 1086 return -1; 1087 } 1088 1089 static int htb_dump_class(struct Qdisc *sch, unsigned long arg, 1090 struct sk_buff *skb, struct tcmsg *tcm) 1091 { 1092 struct htb_class *cl = (struct htb_class *)arg; 1093 struct nlattr *nest; 1094 struct tc_htb_opt opt; 1095 1096 /* Its safe to not acquire qdisc lock. As we hold RTNL, 1097 * no change can happen on the class parameters. 1098 */ 1099 tcm->tcm_parent = cl->parent ? cl->parent->common.classid : TC_H_ROOT; 1100 tcm->tcm_handle = cl->common.classid; 1101 if (!cl->level && cl->un.leaf.q) 1102 tcm->tcm_info = cl->un.leaf.q->handle; 1103 1104 nest = nla_nest_start(skb, TCA_OPTIONS); 1105 if (nest == NULL) 1106 goto nla_put_failure; 1107 1108 memset(&opt, 0, sizeof(opt)); 1109 1110 psched_ratecfg_getrate(&opt.rate, &cl->rate); 1111 opt.buffer = PSCHED_NS2TICKS(cl->buffer); 1112 psched_ratecfg_getrate(&opt.ceil, &cl->ceil); 1113 opt.cbuffer = PSCHED_NS2TICKS(cl->cbuffer); 1114 opt.quantum = cl->quantum; 1115 opt.prio = cl->prio; 1116 opt.level = cl->level; 1117 if (nla_put(skb, TCA_HTB_PARMS, sizeof(opt), &opt)) 1118 goto nla_put_failure; 1119 if ((cl->rate.rate_bytes_ps >= (1ULL << 32)) && 1120 nla_put_u64_64bit(skb, TCA_HTB_RATE64, cl->rate.rate_bytes_ps, 1121 TCA_HTB_PAD)) 1122 goto nla_put_failure; 1123 if ((cl->ceil.rate_bytes_ps >= (1ULL << 32)) && 1124 nla_put_u64_64bit(skb, TCA_HTB_CEIL64, cl->ceil.rate_bytes_ps, 1125 TCA_HTB_PAD)) 1126 goto nla_put_failure; 1127 1128 return nla_nest_end(skb, nest); 1129 1130 nla_put_failure: 1131 nla_nest_cancel(skb, nest); 1132 return -1; 1133 } 1134 1135 static int 1136 htb_dump_class_stats(struct Qdisc *sch, unsigned long arg, struct gnet_dump *d) 1137 { 1138 struct htb_class *cl = (struct htb_class *)arg; 1139 struct gnet_stats_queue qs = { 1140 .drops = cl->drops, 1141 .overlimits = cl->overlimits, 1142 }; 1143 __u32 qlen = 0; 1144 1145 if (!cl->level && cl->un.leaf.q) { 1146 qlen = cl->un.leaf.q->q.qlen; 1147 qs.backlog = cl->un.leaf.q->qstats.backlog; 1148 } 1149 cl->xstats.tokens = clamp_t(s64, PSCHED_NS2TICKS(cl->tokens), 1150 INT_MIN, INT_MAX); 1151 cl->xstats.ctokens = clamp_t(s64, PSCHED_NS2TICKS(cl->ctokens), 1152 INT_MIN, INT_MAX); 1153 1154 if (gnet_stats_copy_basic(qdisc_root_sleeping_running(sch), 1155 d, NULL, &cl->bstats) < 0 || 1156 gnet_stats_copy_rate_est(d, &cl->rate_est) < 0 || 1157 gnet_stats_copy_queue(d, NULL, &qs, qlen) < 0) 1158 return -1; 1159 1160 return gnet_stats_copy_app(d, &cl->xstats, sizeof(cl->xstats)); 1161 } 1162 1163 static int htb_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new, 1164 struct Qdisc **old, struct netlink_ext_ack *extack) 1165 { 1166 struct htb_class *cl = (struct htb_class *)arg; 1167 1168 if (cl->level) 1169 return -EINVAL; 1170 if (new == NULL && 1171 (new = qdisc_create_dflt(sch->dev_queue, &pfifo_qdisc_ops, 1172 cl->common.classid, extack)) == NULL) 1173 return -ENOBUFS; 1174 1175 *old = qdisc_replace(sch, new, &cl->un.leaf.q); 1176 return 0; 1177 } 1178 1179 static struct Qdisc *htb_leaf(struct Qdisc *sch, unsigned long arg) 1180 { 1181 struct htb_class *cl = (struct htb_class *)arg; 1182 return !cl->level ? cl->un.leaf.q : NULL; 1183 } 1184 1185 static void htb_qlen_notify(struct Qdisc *sch, unsigned long arg) 1186 { 1187 struct htb_class *cl = (struct htb_class *)arg; 1188 1189 htb_deactivate(qdisc_priv(sch), cl); 1190 } 1191 1192 static inline int htb_parent_last_child(struct htb_class *cl) 1193 { 1194 if (!cl->parent) 1195 /* the root class */ 1196 return 0; 1197 if (cl->parent->children > 1) 1198 /* not the last child */ 1199 return 0; 1200 return 1; 1201 } 1202 1203 static void htb_parent_to_leaf(struct htb_sched *q, struct htb_class *cl, 1204 struct Qdisc *new_q) 1205 { 1206 struct htb_class *parent = cl->parent; 1207 1208 WARN_ON(cl->level || !cl->un.leaf.q || cl->prio_activity); 1209 1210 if (parent->cmode != HTB_CAN_SEND) 1211 htb_safe_rb_erase(&parent->pq_node, 1212 &q->hlevel[parent->level].wait_pq); 1213 1214 parent->level = 0; 1215 memset(&parent->un.inner, 0, sizeof(parent->un.inner)); 1216 parent->un.leaf.q = new_q ? new_q : &noop_qdisc; 1217 parent->tokens = parent->buffer; 1218 parent->ctokens = parent->cbuffer; 1219 parent->t_c = ktime_get_ns(); 1220 parent->cmode = HTB_CAN_SEND; 1221 } 1222 1223 static void htb_destroy_class(struct Qdisc *sch, struct htb_class *cl) 1224 { 1225 if (!cl->level) { 1226 WARN_ON(!cl->un.leaf.q); 1227 qdisc_destroy(cl->un.leaf.q); 1228 } 1229 gen_kill_estimator(&cl->rate_est); 1230 tcf_block_put(cl->block); 1231 kfree(cl); 1232 } 1233 1234 static void htb_destroy(struct Qdisc *sch) 1235 { 1236 struct htb_sched *q = qdisc_priv(sch); 1237 struct hlist_node *next; 1238 struct htb_class *cl; 1239 unsigned int i; 1240 1241 cancel_work_sync(&q->work); 1242 qdisc_watchdog_cancel(&q->watchdog); 1243 /* This line used to be after htb_destroy_class call below 1244 * and surprisingly it worked in 2.4. But it must precede it 1245 * because filter need its target class alive to be able to call 1246 * unbind_filter on it (without Oops). 1247 */ 1248 tcf_block_put(q->block); 1249 1250 for (i = 0; i < q->clhash.hashsize; i++) { 1251 hlist_for_each_entry(cl, &q->clhash.hash[i], common.hnode) { 1252 tcf_block_put(cl->block); 1253 cl->block = NULL; 1254 } 1255 } 1256 for (i = 0; i < q->clhash.hashsize; i++) { 1257 hlist_for_each_entry_safe(cl, next, &q->clhash.hash[i], 1258 common.hnode) 1259 htb_destroy_class(sch, cl); 1260 } 1261 qdisc_class_hash_destroy(&q->clhash); 1262 __qdisc_reset_queue(&q->direct_queue); 1263 } 1264 1265 static int htb_delete(struct Qdisc *sch, unsigned long arg) 1266 { 1267 struct htb_sched *q = qdisc_priv(sch); 1268 struct htb_class *cl = (struct htb_class *)arg; 1269 struct Qdisc *new_q = NULL; 1270 int last_child = 0; 1271 1272 /* TODO: why don't allow to delete subtree ? references ? does 1273 * tc subsys guarantee us that in htb_destroy it holds no class 1274 * refs so that we can remove children safely there ? 1275 */ 1276 if (cl->children || cl->filter_cnt) 1277 return -EBUSY; 1278 1279 if (!cl->level && htb_parent_last_child(cl)) { 1280 new_q = qdisc_create_dflt(sch->dev_queue, &pfifo_qdisc_ops, 1281 cl->parent->common.classid, 1282 NULL); 1283 last_child = 1; 1284 } 1285 1286 sch_tree_lock(sch); 1287 1288 if (!cl->level) { 1289 unsigned int qlen = cl->un.leaf.q->q.qlen; 1290 unsigned int backlog = cl->un.leaf.q->qstats.backlog; 1291 1292 qdisc_reset(cl->un.leaf.q); 1293 qdisc_tree_reduce_backlog(cl->un.leaf.q, qlen, backlog); 1294 } 1295 1296 /* delete from hash and active; remainder in destroy_class */ 1297 qdisc_class_hash_remove(&q->clhash, &cl->common); 1298 if (cl->parent) 1299 cl->parent->children--; 1300 1301 if (cl->prio_activity) 1302 htb_deactivate(q, cl); 1303 1304 if (cl->cmode != HTB_CAN_SEND) 1305 htb_safe_rb_erase(&cl->pq_node, 1306 &q->hlevel[cl->level].wait_pq); 1307 1308 if (last_child) 1309 htb_parent_to_leaf(q, cl, new_q); 1310 1311 sch_tree_unlock(sch); 1312 1313 htb_destroy_class(sch, cl); 1314 return 0; 1315 } 1316 1317 static int htb_change_class(struct Qdisc *sch, u32 classid, 1318 u32 parentid, struct nlattr **tca, 1319 unsigned long *arg, struct netlink_ext_ack *extack) 1320 { 1321 int err = -EINVAL; 1322 struct htb_sched *q = qdisc_priv(sch); 1323 struct htb_class *cl = (struct htb_class *)*arg, *parent; 1324 struct nlattr *opt = tca[TCA_OPTIONS]; 1325 struct nlattr *tb[TCA_HTB_MAX + 1]; 1326 struct tc_htb_opt *hopt; 1327 u64 rate64, ceil64; 1328 int warn = 0; 1329 1330 /* extract all subattrs from opt attr */ 1331 if (!opt) 1332 goto failure; 1333 1334 err = nla_parse_nested(tb, TCA_HTB_MAX, opt, htb_policy, NULL); 1335 if (err < 0) 1336 goto failure; 1337 1338 err = -EINVAL; 1339 if (tb[TCA_HTB_PARMS] == NULL) 1340 goto failure; 1341 1342 parent = parentid == TC_H_ROOT ? NULL : htb_find(parentid, sch); 1343 1344 hopt = nla_data(tb[TCA_HTB_PARMS]); 1345 if (!hopt->rate.rate || !hopt->ceil.rate) 1346 goto failure; 1347 1348 /* Keeping backward compatible with rate_table based iproute2 tc */ 1349 if (hopt->rate.linklayer == TC_LINKLAYER_UNAWARE) 1350 qdisc_put_rtab(qdisc_get_rtab(&hopt->rate, tb[TCA_HTB_RTAB], 1351 NULL)); 1352 1353 if (hopt->ceil.linklayer == TC_LINKLAYER_UNAWARE) 1354 qdisc_put_rtab(qdisc_get_rtab(&hopt->ceil, tb[TCA_HTB_CTAB], 1355 NULL)); 1356 1357 if (!cl) { /* new class */ 1358 struct Qdisc *new_q; 1359 int prio; 1360 struct { 1361 struct nlattr nla; 1362 struct gnet_estimator opt; 1363 } est = { 1364 .nla = { 1365 .nla_len = nla_attr_size(sizeof(est.opt)), 1366 .nla_type = TCA_RATE, 1367 }, 1368 .opt = { 1369 /* 4s interval, 16s averaging constant */ 1370 .interval = 2, 1371 .ewma_log = 2, 1372 }, 1373 }; 1374 1375 /* check for valid classid */ 1376 if (!classid || TC_H_MAJ(classid ^ sch->handle) || 1377 htb_find(classid, sch)) 1378 goto failure; 1379 1380 /* check maximal depth */ 1381 if (parent && parent->parent && parent->parent->level < 2) { 1382 pr_err("htb: tree is too deep\n"); 1383 goto failure; 1384 } 1385 err = -ENOBUFS; 1386 cl = kzalloc(sizeof(*cl), GFP_KERNEL); 1387 if (!cl) 1388 goto failure; 1389 1390 err = tcf_block_get(&cl->block, &cl->filter_list, sch, extack); 1391 if (err) { 1392 kfree(cl); 1393 goto failure; 1394 } 1395 if (htb_rate_est || tca[TCA_RATE]) { 1396 err = gen_new_estimator(&cl->bstats, NULL, 1397 &cl->rate_est, 1398 NULL, 1399 qdisc_root_sleeping_running(sch), 1400 tca[TCA_RATE] ? : &est.nla); 1401 if (err) { 1402 tcf_block_put(cl->block); 1403 kfree(cl); 1404 goto failure; 1405 } 1406 } 1407 1408 cl->children = 0; 1409 RB_CLEAR_NODE(&cl->pq_node); 1410 1411 for (prio = 0; prio < TC_HTB_NUMPRIO; prio++) 1412 RB_CLEAR_NODE(&cl->node[prio]); 1413 1414 /* create leaf qdisc early because it uses kmalloc(GFP_KERNEL) 1415 * so that can't be used inside of sch_tree_lock 1416 * -- thanks to Karlis Peisenieks 1417 */ 1418 new_q = qdisc_create_dflt(sch->dev_queue, &pfifo_qdisc_ops, 1419 classid, NULL); 1420 sch_tree_lock(sch); 1421 if (parent && !parent->level) { 1422 unsigned int qlen = parent->un.leaf.q->q.qlen; 1423 unsigned int backlog = parent->un.leaf.q->qstats.backlog; 1424 1425 /* turn parent into inner node */ 1426 qdisc_reset(parent->un.leaf.q); 1427 qdisc_tree_reduce_backlog(parent->un.leaf.q, qlen, backlog); 1428 qdisc_destroy(parent->un.leaf.q); 1429 if (parent->prio_activity) 1430 htb_deactivate(q, parent); 1431 1432 /* remove from evt list because of level change */ 1433 if (parent->cmode != HTB_CAN_SEND) { 1434 htb_safe_rb_erase(&parent->pq_node, &q->hlevel[0].wait_pq); 1435 parent->cmode = HTB_CAN_SEND; 1436 } 1437 parent->level = (parent->parent ? parent->parent->level 1438 : TC_HTB_MAXDEPTH) - 1; 1439 memset(&parent->un.inner, 0, sizeof(parent->un.inner)); 1440 } 1441 /* leaf (we) needs elementary qdisc */ 1442 cl->un.leaf.q = new_q ? new_q : &noop_qdisc; 1443 1444 cl->common.classid = classid; 1445 cl->parent = parent; 1446 1447 /* set class to be in HTB_CAN_SEND state */ 1448 cl->tokens = PSCHED_TICKS2NS(hopt->buffer); 1449 cl->ctokens = PSCHED_TICKS2NS(hopt->cbuffer); 1450 cl->mbuffer = 60ULL * NSEC_PER_SEC; /* 1min */ 1451 cl->t_c = ktime_get_ns(); 1452 cl->cmode = HTB_CAN_SEND; 1453 1454 /* attach to the hash list and parent's family */ 1455 qdisc_class_hash_insert(&q->clhash, &cl->common); 1456 if (parent) 1457 parent->children++; 1458 if (cl->un.leaf.q != &noop_qdisc) 1459 qdisc_hash_add(cl->un.leaf.q, true); 1460 } else { 1461 if (tca[TCA_RATE]) { 1462 err = gen_replace_estimator(&cl->bstats, NULL, 1463 &cl->rate_est, 1464 NULL, 1465 qdisc_root_sleeping_running(sch), 1466 tca[TCA_RATE]); 1467 if (err) 1468 return err; 1469 } 1470 sch_tree_lock(sch); 1471 } 1472 1473 rate64 = tb[TCA_HTB_RATE64] ? nla_get_u64(tb[TCA_HTB_RATE64]) : 0; 1474 1475 ceil64 = tb[TCA_HTB_CEIL64] ? nla_get_u64(tb[TCA_HTB_CEIL64]) : 0; 1476 1477 psched_ratecfg_precompute(&cl->rate, &hopt->rate, rate64); 1478 psched_ratecfg_precompute(&cl->ceil, &hopt->ceil, ceil64); 1479 1480 /* it used to be a nasty bug here, we have to check that node 1481 * is really leaf before changing cl->un.leaf ! 1482 */ 1483 if (!cl->level) { 1484 u64 quantum = cl->rate.rate_bytes_ps; 1485 1486 do_div(quantum, q->rate2quantum); 1487 cl->quantum = min_t(u64, quantum, INT_MAX); 1488 1489 if (!hopt->quantum && cl->quantum < 1000) { 1490 warn = -1; 1491 cl->quantum = 1000; 1492 } 1493 if (!hopt->quantum && cl->quantum > 200000) { 1494 warn = 1; 1495 cl->quantum = 200000; 1496 } 1497 if (hopt->quantum) 1498 cl->quantum = hopt->quantum; 1499 if ((cl->prio = hopt->prio) >= TC_HTB_NUMPRIO) 1500 cl->prio = TC_HTB_NUMPRIO - 1; 1501 } 1502 1503 cl->buffer = PSCHED_TICKS2NS(hopt->buffer); 1504 cl->cbuffer = PSCHED_TICKS2NS(hopt->cbuffer); 1505 1506 sch_tree_unlock(sch); 1507 1508 if (warn) 1509 pr_warn("HTB: quantum of class %X is %s. Consider r2q change.\n", 1510 cl->common.classid, (warn == -1 ? "small" : "big")); 1511 1512 qdisc_class_hash_grow(sch, &q->clhash); 1513 1514 *arg = (unsigned long)cl; 1515 return 0; 1516 1517 failure: 1518 return err; 1519 } 1520 1521 static struct tcf_block *htb_tcf_block(struct Qdisc *sch, unsigned long arg, 1522 struct netlink_ext_ack *extack) 1523 { 1524 struct htb_sched *q = qdisc_priv(sch); 1525 struct htb_class *cl = (struct htb_class *)arg; 1526 1527 return cl ? cl->block : q->block; 1528 } 1529 1530 static unsigned long htb_bind_filter(struct Qdisc *sch, unsigned long parent, 1531 u32 classid) 1532 { 1533 struct htb_class *cl = htb_find(classid, sch); 1534 1535 /*if (cl && !cl->level) return 0; 1536 * The line above used to be there to prevent attaching filters to 1537 * leaves. But at least tc_index filter uses this just to get class 1538 * for other reasons so that we have to allow for it. 1539 * ---- 1540 * 19.6.2002 As Werner explained it is ok - bind filter is just 1541 * another way to "lock" the class - unlike "get" this lock can 1542 * be broken by class during destroy IIUC. 1543 */ 1544 if (cl) 1545 cl->filter_cnt++; 1546 return (unsigned long)cl; 1547 } 1548 1549 static void htb_unbind_filter(struct Qdisc *sch, unsigned long arg) 1550 { 1551 struct htb_class *cl = (struct htb_class *)arg; 1552 1553 if (cl) 1554 cl->filter_cnt--; 1555 } 1556 1557 static void htb_walk(struct Qdisc *sch, struct qdisc_walker *arg) 1558 { 1559 struct htb_sched *q = qdisc_priv(sch); 1560 struct htb_class *cl; 1561 unsigned int i; 1562 1563 if (arg->stop) 1564 return; 1565 1566 for (i = 0; i < q->clhash.hashsize; i++) { 1567 hlist_for_each_entry(cl, &q->clhash.hash[i], common.hnode) { 1568 if (arg->count < arg->skip) { 1569 arg->count++; 1570 continue; 1571 } 1572 if (arg->fn(sch, (unsigned long)cl, arg) < 0) { 1573 arg->stop = 1; 1574 return; 1575 } 1576 arg->count++; 1577 } 1578 } 1579 } 1580 1581 static const struct Qdisc_class_ops htb_class_ops = { 1582 .graft = htb_graft, 1583 .leaf = htb_leaf, 1584 .qlen_notify = htb_qlen_notify, 1585 .find = htb_search, 1586 .change = htb_change_class, 1587 .delete = htb_delete, 1588 .walk = htb_walk, 1589 .tcf_block = htb_tcf_block, 1590 .bind_tcf = htb_bind_filter, 1591 .unbind_tcf = htb_unbind_filter, 1592 .dump = htb_dump_class, 1593 .dump_stats = htb_dump_class_stats, 1594 }; 1595 1596 static struct Qdisc_ops htb_qdisc_ops __read_mostly = { 1597 .cl_ops = &htb_class_ops, 1598 .id = "htb", 1599 .priv_size = sizeof(struct htb_sched), 1600 .enqueue = htb_enqueue, 1601 .dequeue = htb_dequeue, 1602 .peek = qdisc_peek_dequeued, 1603 .init = htb_init, 1604 .reset = htb_reset, 1605 .destroy = htb_destroy, 1606 .dump = htb_dump, 1607 .owner = THIS_MODULE, 1608 }; 1609 1610 static int __init htb_module_init(void) 1611 { 1612 return register_qdisc(&htb_qdisc_ops); 1613 } 1614 static void __exit htb_module_exit(void) 1615 { 1616 unregister_qdisc(&htb_qdisc_ops); 1617 } 1618 1619 module_init(htb_module_init) 1620 module_exit(htb_module_exit) 1621 MODULE_LICENSE("GPL"); 1622