Blame view
net/sched/sch_api.c
44.5 KB
1da177e4c
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 |
/* * net/sched/sch_api.c Packet scheduler API. * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version * 2 of the License, or (at your option) any later version. * * Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru> * * Fixes: * * Rani Assaf <rani@magic.metawire.com> :980802: JIFFIES and CPU clock sources are repaired. * Eduardo J. Blanco <ejbs@netlabs.com.uy> :990222: kmod support * Jamal Hadi Salim <hadi@nortelnetworks.com>: 990601: ingress support */ |
1da177e4c
|
17 18 19 |
#include <linux/module.h> #include <linux/types.h> #include <linux/kernel.h> |
1da177e4c
|
20 |
#include <linux/string.h> |
1da177e4c
|
21 |
#include <linux/errno.h> |
1da177e4c
|
22 |
#include <linux/skbuff.h> |
1da177e4c
|
23 24 25 26 27 |
#include <linux/init.h> #include <linux/proc_fs.h> #include <linux/seq_file.h> #include <linux/kmod.h> #include <linux/list.h> |
4179477f6
|
28 |
#include <linux/hrtimer.h> |
25bfcd5a7
|
29 |
#include <linux/lockdep.h> |
5a0e3ad6a
|
30 |
#include <linux/slab.h> |
1da177e4c
|
31 |
|
457c4cbc5
|
32 |
#include <net/net_namespace.h> |
b854272b3
|
33 |
#include <net/sock.h> |
dc5fc579b
|
34 |
#include <net/netlink.h> |
1da177e4c
|
35 |
#include <net/pkt_sched.h> |
7316ae88c
|
36 37 |
static int qdisc_notify(struct net *net, struct sk_buff *oskb, struct nlmsghdr *n, u32 clid, |
1da177e4c
|
38 |
struct Qdisc *old, struct Qdisc *new); |
7316ae88c
|
39 40 41 |
static int tclass_notify(struct net *net, struct sk_buff *oskb, struct nlmsghdr *n, struct Qdisc *q, unsigned long cl, int event); |
1da177e4c
|
42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 |
/* Short review. ------------- This file consists of two interrelated parts: 1. queueing disciplines manager frontend. 2. traffic classes manager frontend. Generally, queueing discipline ("qdisc") is a black box, which is able to enqueue packets and to dequeue them (when device is ready to send something) in order and at times determined by algorithm hidden in it. qdisc's are divided to two categories: - "queues", which have no internal structure visible from outside. - "schedulers", which split all the packets to "traffic classes", using "packet classifiers" (look at cls_api.c) In turn, classes may have child qdiscs (as rule, queues) attached to them etc. etc. etc. The goal of the routines in this file is to translate information supplied by user in the form of handles to more intelligible for kernel form, to make some sanity checks and part of work, which is common to all qdiscs and to provide rtnetlink notifications. All real intelligent work is done inside qdisc modules. Every discipline has two major routines: enqueue and dequeue. ---dequeue dequeue usually returns a skb to send. It is allowed to return NULL, but it does not mean that queue is empty, it just means that discipline does not want to send anything this time. Queue is really empty if q->q.qlen == 0. For complicated disciplines with multiple queues q->q is not real packet queue, but however q->q.qlen must be valid. ---enqueue enqueue returns 0, if packet was enqueued successfully. If packet (this one or another one) was dropped, it returns not zero error code. NET_XMIT_DROP - this packet dropped Expected action: do not backoff, but wait until queue will clear. NET_XMIT_CN - probably this packet enqueued, but another one dropped. Expected action: backoff or ignore NET_XMIT_POLICED - dropped by police. Expected action: backoff or error to real-time apps. Auxiliary routines: |
99c0db267
|
100 101 102 |
---peek like dequeue but without removing a packet from the queue |
1da177e4c
|
103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 |
---reset returns qdisc to initial state: purge all buffers, clear all timers, counters (except for statistics) etc. ---init initializes newly created qdisc. ---destroy destroys resources allocated by init and during lifetime of qdisc. ---change changes qdisc parameters. */ /* Protects list of registered TC modules. It is pure SMP lock. */ static DEFINE_RWLOCK(qdisc_mod_lock); /************************************************ * Queueing disciplines manipulation. * ************************************************/ /* The list of all installed queueing disciplines. */ static struct Qdisc_ops *qdisc_base; |
21eb21898
|
133 |
/* Register/unregister queueing discipline */ |
1da177e4c
|
134 135 136 137 138 139 140 141 142 143 144 145 146 |
int register_qdisc(struct Qdisc_ops *qops) { struct Qdisc_ops *q, **qp; int rc = -EEXIST; write_lock(&qdisc_mod_lock); for (qp = &qdisc_base; (q = *qp) != NULL; qp = &q->next) if (!strcmp(qops->id, q->id)) goto out; if (qops->enqueue == NULL) qops->enqueue = noop_qdisc_ops.enqueue; |
99c0db267
|
147 |
if (qops->peek == NULL) { |
68fd26b59
|
148 |
if (qops->dequeue == NULL) |
99c0db267
|
149 |
qops->peek = noop_qdisc_ops.peek; |
68fd26b59
|
150 151 |
else goto out_einval; |
99c0db267
|
152 |
} |
1da177e4c
|
153 154 |
if (qops->dequeue == NULL) qops->dequeue = noop_qdisc_ops.dequeue; |
68fd26b59
|
155 156 |
if (qops->cl_ops) { const struct Qdisc_class_ops *cops = qops->cl_ops; |
3e9e5a592
|
157 |
if (!(cops->get && cops->put && cops->walk && cops->leaf)) |
68fd26b59
|
158 159 160 161 162 |
goto out_einval; if (cops->tcf_chain && !(cops->bind_tcf && cops->unbind_tcf)) goto out_einval; } |
1da177e4c
|
163 164 165 166 167 168 |
qops->next = NULL; *qp = qops; rc = 0; out: write_unlock(&qdisc_mod_lock); return rc; |
68fd26b59
|
169 170 171 172 |
out_einval: rc = -EINVAL; goto out; |
1da177e4c
|
173 |
} |
62e3ba1b5
|
174 |
EXPORT_SYMBOL(register_qdisc); |
1da177e4c
|
175 176 177 178 179 180 181 |
int unregister_qdisc(struct Qdisc_ops *qops) { struct Qdisc_ops *q, **qp; int err = -ENOENT; write_lock(&qdisc_mod_lock); |
cc7ec456f
|
182 |
for (qp = &qdisc_base; (q = *qp) != NULL; qp = &q->next) |
1da177e4c
|
183 184 185 186 187 188 189 190 191 192 |
if (q == qops) break; if (q) { *qp = q->next; q->next = NULL; err = 0; } write_unlock(&qdisc_mod_lock); return err; } |
62e3ba1b5
|
193 |
EXPORT_SYMBOL(unregister_qdisc); |
1da177e4c
|
194 |
|
6da7c8fcb
|
195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 |
/* Get default qdisc if not otherwise specified */ void qdisc_get_default(char *name, size_t len) { read_lock(&qdisc_mod_lock); strlcpy(name, default_qdisc_ops->id, len); read_unlock(&qdisc_mod_lock); } static struct Qdisc_ops *qdisc_lookup_default(const char *name) { struct Qdisc_ops *q = NULL; for (q = qdisc_base; q; q = q->next) { if (!strcmp(name, q->id)) { if (!try_module_get(q->owner)) q = NULL; break; } } return q; } /* Set new default qdisc to use */ int qdisc_set_default(const char *name) { const struct Qdisc_ops *ops; if (!capable(CAP_NET_ADMIN)) return -EPERM; write_lock(&qdisc_mod_lock); ops = qdisc_lookup_default(name); if (!ops) { /* Not found, drop lock and try to load module */ write_unlock(&qdisc_mod_lock); request_module("sch_%s", name); write_lock(&qdisc_mod_lock); ops = qdisc_lookup_default(name); } if (ops) { /* Set new default */ module_put(default_qdisc_ops->owner); default_qdisc_ops = ops; } write_unlock(&qdisc_mod_lock); return ops ? 0 : -ENOENT; } |
1da177e4c
|
246 247 248 |
/* We know handle. Find qdisc among all qdisc's attached to device (root qdisc, all its children, children of children etc.) */ |
6113b748f
|
249 |
static struct Qdisc *qdisc_match_from_root(struct Qdisc *root, u32 handle) |
8123b421e
|
250 251 252 253 254 255 256 257 258 259 260 261 262 |
{ struct Qdisc *q; if (!(root->flags & TCQ_F_BUILTIN) && root->handle == handle) return root; list_for_each_entry(q, &root->list, list) { if (q->handle == handle) return q; } return NULL; } |
95dc19299
|
263 |
void qdisc_list_add(struct Qdisc *q) |
f6e0b239a
|
264 |
{ |
37314363c
|
265 266 |
if ((q->parent != TC_H_ROOT) && !(q->flags & TCQ_F_INGRESS)) { struct Qdisc *root = qdisc_dev(q)->qdisc; |
e57a784d8
|
267 |
|
37314363c
|
268 |
WARN_ON_ONCE(root == &noop_qdisc); |
e57a784d8
|
269 |
list_add_tail(&q->list, &root->list); |
37314363c
|
270 |
} |
f6e0b239a
|
271 |
} |
95dc19299
|
272 |
EXPORT_SYMBOL(qdisc_list_add); |
f6e0b239a
|
273 274 275 |
void qdisc_list_del(struct Qdisc *q) { |
f6486d40b
|
276 |
if ((q->parent != TC_H_ROOT) && !(q->flags & TCQ_F_INGRESS)) |
f6e0b239a
|
277 |
list_del(&q->list); |
f6e0b239a
|
278 279 |
} EXPORT_SYMBOL(qdisc_list_del); |
ead81cc5f
|
280 |
struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle) |
1da177e4c
|
281 |
{ |
f6e0b239a
|
282 |
struct Qdisc *q; |
af356afa0
|
283 284 285 |
q = qdisc_match_from_root(dev->qdisc, handle); if (q) goto out; |
f6e0b239a
|
286 |
|
24824a09e
|
287 288 289 290 |
if (dev_ingress_queue(dev)) q = qdisc_match_from_root( dev_ingress_queue(dev)->qdisc_sleeping, handle); |
f6486d40b
|
291 |
out: |
f6e0b239a
|
292 |
return q; |
1da177e4c
|
293 294 295 296 297 298 |
} static struct Qdisc *qdisc_leaf(struct Qdisc *p, u32 classid) { unsigned long cl; struct Qdisc *leaf; |
20fea08b5
|
299 |
const struct Qdisc_class_ops *cops = p->ops->cl_ops; |
1da177e4c
|
300 301 302 303 304 305 306 307 308 309 310 311 312 |
if (cops == NULL) return NULL; cl = cops->get(p, classid); if (cl == 0) return NULL; leaf = cops->leaf(p, cl); cops->put(p, cl); return leaf; } /* Find queueing discipline by name */ |
1e90474c3
|
313 |
static struct Qdisc_ops *qdisc_lookup_ops(struct nlattr *kind) |
1da177e4c
|
314 315 316 317 318 319 |
{ struct Qdisc_ops *q = NULL; if (kind) { read_lock(&qdisc_mod_lock); for (q = qdisc_base; q; q = q->next) { |
1e90474c3
|
320 |
if (nla_strcmp(kind, q->id) == 0) { |
1da177e4c
|
321 322 323 324 325 326 327 328 329 |
if (!try_module_get(q->owner)) q = NULL; break; } } read_unlock(&qdisc_mod_lock); } return q; } |
8a8e3d84b
|
330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 |
/* The linklayer setting were not transferred from iproute2, in older * versions, and the rate tables lookup systems have been dropped in * the kernel. To keep backward compatible with older iproute2 tc * utils, we detect the linklayer setting by detecting if the rate * table were modified. * * For linklayer ATM table entries, the rate table will be aligned to * 48 bytes, thus some table entries will contain the same value. The * mpu (min packet unit) is also encoded into the old rate table, thus * starting from the mpu, we find low and high table entries for * mapping this cell. If these entries contain the same value, when * the rate tables have been modified for linklayer ATM. * * This is done by rounding mpu to the nearest 48 bytes cell/entry, * and then roundup to the next cell, calc the table entry one below, * and compare. */ static __u8 __detect_linklayer(struct tc_ratespec *r, __u32 *rtab) { int low = roundup(r->mpu, 48); int high = roundup(low+1, 48); int cell_low = low >> r->cell_log; int cell_high = (high >> r->cell_log) - 1; /* rtab is too inaccurate at rates > 100Mbit/s */ if ((r->rate > (100000000/8)) || (rtab[0] == 0)) { pr_debug("TC linklayer: Giving up ATM detection "); return TC_LINKLAYER_ETHERNET; } if ((cell_high > cell_low) && (cell_high < 256) && (rtab[cell_low] == rtab[cell_high])) { pr_debug("TC linklayer: Detected ATM, low(%d)=high(%d)=%u ", cell_low, cell_high, rtab[cell_high]); return TC_LINKLAYER_ATM; } return TC_LINKLAYER_ETHERNET; } |
1da177e4c
|
370 |
static struct qdisc_rate_table *qdisc_rtab_list; |
1e90474c3
|
371 |
struct qdisc_rate_table *qdisc_get_rtab(struct tc_ratespec *r, struct nlattr *tab) |
1da177e4c
|
372 373 |
{ struct qdisc_rate_table *rtab; |
40edeff6e
|
374 375 376 |
if (tab == NULL || r->rate == 0 || r->cell_log == 0 || nla_len(tab) != TC_RTAB_SIZE) return NULL; |
1da177e4c
|
377 |
for (rtab = qdisc_rtab_list; rtab; rtab = rtab->next) { |
40edeff6e
|
378 379 |
if (!memcmp(&rtab->rate, r, sizeof(struct tc_ratespec)) && !memcmp(&rtab->data, nla_data(tab), 1024)) { |
1da177e4c
|
380 381 382 383 |
rtab->refcnt++; return rtab; } } |
1da177e4c
|
384 385 386 387 |
rtab = kmalloc(sizeof(*rtab), GFP_KERNEL); if (rtab) { rtab->rate = *r; rtab->refcnt = 1; |
1e90474c3
|
388 |
memcpy(rtab->data, nla_data(tab), 1024); |
8a8e3d84b
|
389 390 |
if (r->linklayer == TC_LINKLAYER_UNAWARE) r->linklayer = __detect_linklayer(r, rtab->data); |
1da177e4c
|
391 392 393 394 395 |
rtab->next = qdisc_rtab_list; qdisc_rtab_list = rtab; } return rtab; } |
62e3ba1b5
|
396 |
EXPORT_SYMBOL(qdisc_get_rtab); |
1da177e4c
|
397 398 399 400 401 402 403 |
void qdisc_put_rtab(struct qdisc_rate_table *tab) { struct qdisc_rate_table *rtab, **rtabp; if (!tab || --tab->refcnt) return; |
cc7ec456f
|
404 405 406 |
for (rtabp = &qdisc_rtab_list; (rtab = *rtabp) != NULL; rtabp = &rtab->next) { |
1da177e4c
|
407 408 409 410 411 412 413 |
if (rtab == tab) { *rtabp = rtab->next; kfree(rtab); return; } } } |
62e3ba1b5
|
414 |
EXPORT_SYMBOL(qdisc_put_rtab); |
1da177e4c
|
415 |
|
175f9c1bb
|
416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 |
static LIST_HEAD(qdisc_stab_list); static DEFINE_SPINLOCK(qdisc_stab_lock); static const struct nla_policy stab_policy[TCA_STAB_MAX + 1] = { [TCA_STAB_BASE] = { .len = sizeof(struct tc_sizespec) }, [TCA_STAB_DATA] = { .type = NLA_BINARY }, }; static struct qdisc_size_table *qdisc_get_stab(struct nlattr *opt) { struct nlattr *tb[TCA_STAB_MAX + 1]; struct qdisc_size_table *stab; struct tc_sizespec *s; unsigned int tsize = 0; u16 *tab = NULL; int err; err = nla_parse_nested(tb, TCA_STAB_MAX, opt, stab_policy); if (err < 0) return ERR_PTR(err); if (!tb[TCA_STAB_BASE]) return ERR_PTR(-EINVAL); s = nla_data(tb[TCA_STAB_BASE]); if (s->tsize > 0) { if (!tb[TCA_STAB_DATA]) return ERR_PTR(-EINVAL); tab = nla_data(tb[TCA_STAB_DATA]); tsize = nla_len(tb[TCA_STAB_DATA]) / sizeof(u16); } |
00093fab9
|
447 |
if (tsize != s->tsize || (!tab && tsize > 0)) |
175f9c1bb
|
448 |
return ERR_PTR(-EINVAL); |
f3b9605d7
|
449 |
spin_lock(&qdisc_stab_lock); |
175f9c1bb
|
450 451 452 453 454 455 456 |
list_for_each_entry(stab, &qdisc_stab_list, list) { if (memcmp(&stab->szopts, s, sizeof(*s))) continue; if (tsize > 0 && memcmp(stab->data, tab, tsize * sizeof(u16))) continue; stab->refcnt++; |
f3b9605d7
|
457 |
spin_unlock(&qdisc_stab_lock); |
175f9c1bb
|
458 459 |
return stab; } |
f3b9605d7
|
460 |
spin_unlock(&qdisc_stab_lock); |
175f9c1bb
|
461 462 463 464 465 466 467 468 469 |
stab = kmalloc(sizeof(*stab) + tsize * sizeof(u16), GFP_KERNEL); if (!stab) return ERR_PTR(-ENOMEM); stab->refcnt = 1; stab->szopts = *s; if (tsize > 0) memcpy(stab->data, tab, tsize * sizeof(u16)); |
f3b9605d7
|
470 |
spin_lock(&qdisc_stab_lock); |
175f9c1bb
|
471 |
list_add_tail(&stab->list, &qdisc_stab_list); |
f3b9605d7
|
472 |
spin_unlock(&qdisc_stab_lock); |
175f9c1bb
|
473 474 475 |
return stab; } |
a2da570d6
|
476 477 478 479 |
static void stab_kfree_rcu(struct rcu_head *head) { kfree(container_of(head, struct qdisc_size_table, rcu)); } |
175f9c1bb
|
480 481 482 483 |
void qdisc_put_stab(struct qdisc_size_table *tab) { if (!tab) return; |
f3b9605d7
|
484 |
spin_lock(&qdisc_stab_lock); |
175f9c1bb
|
485 486 487 |
if (--tab->refcnt == 0) { list_del(&tab->list); |
a2da570d6
|
488 |
call_rcu_bh(&tab->rcu, stab_kfree_rcu); |
175f9c1bb
|
489 |
} |
f3b9605d7
|
490 |
spin_unlock(&qdisc_stab_lock); |
175f9c1bb
|
491 492 493 494 495 496 497 498 |
} EXPORT_SYMBOL(qdisc_put_stab); static int qdisc_dump_stab(struct sk_buff *skb, struct qdisc_size_table *stab) { struct nlattr *nest; nest = nla_nest_start(skb, TCA_STAB); |
3aa4614da
|
499 500 |
if (nest == NULL) goto nla_put_failure; |
1b34ec43c
|
501 502 |
if (nla_put(skb, TCA_STAB_BASE, sizeof(stab->szopts), &stab->szopts)) goto nla_put_failure; |
175f9c1bb
|
503 504 505 506 507 508 509 |
nla_nest_end(skb, nest); return skb->len; nla_put_failure: return -1; } |
a2da570d6
|
510 |
void __qdisc_calculate_pkt_len(struct sk_buff *skb, const struct qdisc_size_table *stab) |
175f9c1bb
|
511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 |
{ int pkt_len, slot; pkt_len = skb->len + stab->szopts.overhead; if (unlikely(!stab->szopts.tsize)) goto out; slot = pkt_len + stab->szopts.cell_align; if (unlikely(slot < 0)) slot = 0; slot >>= stab->szopts.cell_log; if (likely(slot < stab->szopts.tsize)) pkt_len = stab->data[slot]; else pkt_len = stab->data[stab->szopts.tsize - 1] * (slot / stab->szopts.tsize) + stab->data[slot % stab->szopts.tsize]; pkt_len <<= stab->szopts.size_log; out: if (unlikely(pkt_len < 1)) pkt_len = 1; qdisc_skb_cb(skb)->pkt_len = pkt_len; } |
a2da570d6
|
536 |
EXPORT_SYMBOL(__qdisc_calculate_pkt_len); |
175f9c1bb
|
537 |
|
6e765a009
|
538 |
void qdisc_warn_nonwc(const char *txt, struct Qdisc *qdisc) |
b00355db3
|
539 540 |
{ if (!(qdisc->flags & TCQ_F_WARN_NONWC)) { |
cc7ec456f
|
541 542 543 |
pr_warn("%s: %s qdisc %X: is non-work-conserving? ", txt, qdisc->ops->id, qdisc->handle >> 16); |
b00355db3
|
544 545 546 547 |
qdisc->flags |= TCQ_F_WARN_NONWC; } } EXPORT_SYMBOL(qdisc_warn_nonwc); |
4179477f6
|
548 549 550 |
static enum hrtimer_restart qdisc_watchdog(struct hrtimer *timer) { struct qdisc_watchdog *wd = container_of(timer, struct qdisc_watchdog, |
2fbd3da38
|
551 |
timer); |
4179477f6
|
552 |
|
fd245a4ad
|
553 |
qdisc_unthrottled(wd->qdisc); |
8608db031
|
554 |
__netif_schedule(qdisc_root(wd->qdisc)); |
1936502d0
|
555 |
|
4179477f6
|
556 557 558 559 560 |
return HRTIMER_NORESTART; } void qdisc_watchdog_init(struct qdisc_watchdog *wd, struct Qdisc *qdisc) { |
2fbd3da38
|
561 562 |
hrtimer_init(&wd->timer, CLOCK_MONOTONIC, HRTIMER_MODE_ABS); wd->timer.function = qdisc_watchdog; |
4179477f6
|
563 564 565 |
wd->qdisc = qdisc; } EXPORT_SYMBOL(qdisc_watchdog_init); |
34c5d292c
|
566 |
void qdisc_watchdog_schedule_ns(struct qdisc_watchdog *wd, u64 expires) |
4179477f6
|
567 |
{ |
2540e0511
|
568 569 570 |
if (test_bit(__QDISC_STATE_DEACTIVATED, &qdisc_root_sleeping(wd->qdisc)->state)) return; |
fd245a4ad
|
571 |
qdisc_throttled(wd->qdisc); |
46baac38e
|
572 573 |
hrtimer_start(&wd->timer, |
34c5d292c
|
574 |
ns_to_ktime(expires), |
46baac38e
|
575 |
HRTIMER_MODE_ABS); |
4179477f6
|
576 |
} |
34c5d292c
|
577 |
EXPORT_SYMBOL(qdisc_watchdog_schedule_ns); |
4179477f6
|
578 579 580 |
void qdisc_watchdog_cancel(struct qdisc_watchdog *wd) { |
2fbd3da38
|
581 |
hrtimer_cancel(&wd->timer); |
fd245a4ad
|
582 |
qdisc_unthrottled(wd->qdisc); |
4179477f6
|
583 584 |
} EXPORT_SYMBOL(qdisc_watchdog_cancel); |
1da177e4c
|
585 |
|
a94f779f9
|
586 |
static struct hlist_head *qdisc_class_hash_alloc(unsigned int n) |
6fe1c7a55
|
587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 |
{ unsigned int size = n * sizeof(struct hlist_head), i; struct hlist_head *h; if (size <= PAGE_SIZE) h = kmalloc(size, GFP_KERNEL); else h = (struct hlist_head *) __get_free_pages(GFP_KERNEL, get_order(size)); if (h != NULL) { for (i = 0; i < n; i++) INIT_HLIST_HEAD(&h[i]); } return h; } static void qdisc_class_hash_free(struct hlist_head *h, unsigned int n) { unsigned int size = n * sizeof(struct hlist_head); if (size <= PAGE_SIZE) kfree(h); else free_pages((unsigned long)h, get_order(size)); } void qdisc_class_hash_grow(struct Qdisc *sch, struct Qdisc_class_hash *clhash) { struct Qdisc_class_common *cl; |
b67bfe0d4
|
617 |
struct hlist_node *next; |
6fe1c7a55
|
618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 |
struct hlist_head *nhash, *ohash; unsigned int nsize, nmask, osize; unsigned int i, h; /* Rehash when load factor exceeds 0.75 */ if (clhash->hashelems * 4 <= clhash->hashsize * 3) return; nsize = clhash->hashsize * 2; nmask = nsize - 1; nhash = qdisc_class_hash_alloc(nsize); if (nhash == NULL) return; ohash = clhash->hash; osize = clhash->hashsize; sch_tree_lock(sch); for (i = 0; i < osize; i++) { |
b67bfe0d4
|
636 |
hlist_for_each_entry_safe(cl, next, &ohash[i], hnode) { |
6fe1c7a55
|
637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 |
h = qdisc_class_hash(cl->classid, nmask); hlist_add_head(&cl->hnode, &nhash[h]); } } clhash->hash = nhash; clhash->hashsize = nsize; clhash->hashmask = nmask; sch_tree_unlock(sch); qdisc_class_hash_free(ohash, osize); } EXPORT_SYMBOL(qdisc_class_hash_grow); int qdisc_class_hash_init(struct Qdisc_class_hash *clhash) { unsigned int size = 4; clhash->hash = qdisc_class_hash_alloc(size); if (clhash->hash == NULL) return -ENOMEM; clhash->hashsize = size; clhash->hashmask = size - 1; clhash->hashelems = 0; return 0; } EXPORT_SYMBOL(qdisc_class_hash_init); void qdisc_class_hash_destroy(struct Qdisc_class_hash *clhash) { qdisc_class_hash_free(clhash->hash, clhash->hashsize); } EXPORT_SYMBOL(qdisc_class_hash_destroy); void qdisc_class_hash_insert(struct Qdisc_class_hash *clhash, struct Qdisc_class_common *cl) { unsigned int h; INIT_HLIST_NODE(&cl->hnode); h = qdisc_class_hash(cl->classid, clhash->hashmask); hlist_add_head(&cl->hnode, &clhash->hash[h]); clhash->hashelems++; } EXPORT_SYMBOL(qdisc_class_hash_insert); void qdisc_class_hash_remove(struct Qdisc_class_hash *clhash, struct Qdisc_class_common *cl) { hlist_del(&cl->hnode); clhash->hashelems--; } EXPORT_SYMBOL(qdisc_class_hash_remove); |
fa0f5aa74
|
689 690 691 |
/* Allocate an unique handle from space managed by kernel * Possible range is [8000-FFFF]:0000 (0x8000 values) */ |
1da177e4c
|
692 693 |
static u32 qdisc_alloc_handle(struct net_device *dev) { |
fa0f5aa74
|
694 |
int i = 0x8000; |
1da177e4c
|
695 696 697 698 699 700 |
static u32 autohandle = TC_H_MAKE(0x80000000U, 0); do { autohandle += TC_H_MAKE(0x10000U, 0); if (autohandle == TC_H_MAKE(TC_H_ROOT, 0)) autohandle = TC_H_MAKE(0x80000000U, 0); |
fa0f5aa74
|
701 702 703 704 |
if (!qdisc_lookup(dev, autohandle)) return autohandle; cond_resched(); } while (--i > 0); |
1da177e4c
|
705 |
|
fa0f5aa74
|
706 |
return 0; |
1da177e4c
|
707 |
} |
43effa1e5
|
708 709 |
void qdisc_tree_decrease_qlen(struct Qdisc *sch, unsigned int n) { |
20fea08b5
|
710 |
const struct Qdisc_class_ops *cops; |
43effa1e5
|
711 712 |
unsigned long cl; u32 parentid; |
2c8c8e6f9
|
713 |
int drops; |
43effa1e5
|
714 715 716 |
if (n == 0) return; |
2c8c8e6f9
|
717 |
drops = max_t(int, n, 0); |
43effa1e5
|
718 |
while ((parentid = sch->parent)) { |
066a3b5b2
|
719 720 |
if (TC_H_MAJ(parentid) == TC_H_MAJ(TC_H_INGRESS)) return; |
5ce2d488f
|
721 |
sch = qdisc_lookup(qdisc_dev(sch), TC_H_MAJ(parentid)); |
ffc8fefaf
|
722 723 724 725 |
if (sch == NULL) { WARN_ON(parentid != TC_H_ROOT); return; } |
43effa1e5
|
726 727 728 729 730 731 732 |
cops = sch->ops->cl_ops; if (cops->qlen_notify) { cl = cops->get(sch, parentid); cops->qlen_notify(sch, cl); cops->put(sch, cl); } sch->q.qlen -= n; |
2c8c8e6f9
|
733 |
sch->qstats.drops += drops; |
43effa1e5
|
734 735 736 |
} } EXPORT_SYMBOL(qdisc_tree_decrease_qlen); |
1da177e4c
|
737 |
|
7316ae88c
|
738 739 |
static void notify_and_destroy(struct net *net, struct sk_buff *skb, struct nlmsghdr *n, u32 clid, |
99194cff3
|
740 741 742 |
struct Qdisc *old, struct Qdisc *new) { if (new || old) |
7316ae88c
|
743 |
qdisc_notify(net, skb, n, clid, old, new); |
1da177e4c
|
744 |
|
4d8863a29
|
745 |
if (old) |
99194cff3
|
746 |
qdisc_destroy(old); |
99194cff3
|
747 748 749 750 751 752 753 754 755 |
} /* Graft qdisc "new" to class "classid" of qdisc "parent" or * to device "dev". * * When appropriate send a netlink notification using 'skb' * and "n". * * On success, destroy old qdisc. |
1da177e4c
|
756 757 758 |
*/ static int qdisc_graft(struct net_device *dev, struct Qdisc *parent, |
99194cff3
|
759 760 |
struct sk_buff *skb, struct nlmsghdr *n, u32 classid, struct Qdisc *new, struct Qdisc *old) |
1da177e4c
|
761 |
{ |
99194cff3
|
762 |
struct Qdisc *q = old; |
7316ae88c
|
763 |
struct net *net = dev_net(dev); |
1da177e4c
|
764 |
int err = 0; |
1da177e4c
|
765 |
|
10297b993
|
766 |
if (parent == NULL) { |
99194cff3
|
767 768 769 770 |
unsigned int i, num_q, ingress; ingress = 0; num_q = dev->num_tx_queues; |
8d50b53d6
|
771 772 |
if ((q && q->flags & TCQ_F_INGRESS) || (new && new->flags & TCQ_F_INGRESS)) { |
99194cff3
|
773 774 |
num_q = 1; ingress = 1; |
24824a09e
|
775 776 |
if (!dev_ingress_queue(dev)) return -ENOENT; |
99194cff3
|
777 778 779 780 |
} if (dev->flags & IFF_UP) dev_deactivate(dev); |
6ec1c69a8
|
781 782 783 784 |
if (new && new->ops->attach) { new->ops->attach(new); num_q = 0; } |
99194cff3
|
785 |
for (i = 0; i < num_q; i++) { |
24824a09e
|
786 |
struct netdev_queue *dev_queue = dev_ingress_queue(dev); |
99194cff3
|
787 788 789 |
if (!ingress) dev_queue = netdev_get_tx_queue(dev, i); |
8d50b53d6
|
790 791 792 |
old = dev_graft_qdisc(dev_queue, new); if (new && i > 0) atomic_inc(&new->refcnt); |
036d6a673
|
793 794 |
if (!ingress) qdisc_destroy(old); |
1da177e4c
|
795 |
} |
99194cff3
|
796 |
|
036d6a673
|
797 |
if (!ingress) { |
7316ae88c
|
798 799 |
notify_and_destroy(net, skb, n, classid, dev->qdisc, new); |
036d6a673
|
800 801 802 803 |
if (new && !new->ops->attach) atomic_inc(&new->refcnt); dev->qdisc = new ? : &noop_qdisc; } else { |
7316ae88c
|
804 |
notify_and_destroy(net, skb, n, classid, old, new); |
036d6a673
|
805 |
} |
af356afa0
|
806 |
|
99194cff3
|
807 808 |
if (dev->flags & IFF_UP) dev_activate(dev); |
1da177e4c
|
809 |
} else { |
20fea08b5
|
810 |
const struct Qdisc_class_ops *cops = parent->ops->cl_ops; |
1da177e4c
|
811 |
|
c9f1d0389
|
812 813 |
err = -EOPNOTSUPP; if (cops && cops->graft) { |
1da177e4c
|
814 815 |
unsigned long cl = cops->get(parent, classid); if (cl) { |
99194cff3
|
816 |
err = cops->graft(parent, cl, new, &old); |
1da177e4c
|
817 |
cops->put(parent, cl); |
c9f1d0389
|
818 819 |
} else err = -ENOENT; |
1da177e4c
|
820 |
} |
99194cff3
|
821 |
if (!err) |
7316ae88c
|
822 |
notify_and_destroy(net, skb, n, classid, old, new); |
1da177e4c
|
823 824 825 |
} return err; } |
25bfcd5a7
|
826 827 828 |
/* lockdep annotation is needed for ingress; egress gets it only for name */ static struct lock_class_key qdisc_tx_lock; static struct lock_class_key qdisc_rx_lock; |
1da177e4c
|
829 830 831 832 833 834 835 |
/* Allocate and initialize new qdisc. Parameters are passed via opt. */ static struct Qdisc * |
bb949fbd1
|
836 |
qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, |
23bcf634c
|
837 838 |
struct Qdisc *p, u32 parent, u32 handle, struct nlattr **tca, int *errp) |
1da177e4c
|
839 840 |
{ int err; |
1e90474c3
|
841 |
struct nlattr *kind = tca[TCA_KIND]; |
1da177e4c
|
842 843 |
struct Qdisc *sch; struct Qdisc_ops *ops; |
175f9c1bb
|
844 |
struct qdisc_size_table *stab; |
1da177e4c
|
845 846 |
ops = qdisc_lookup_ops(kind); |
95a5afca4
|
847 |
#ifdef CONFIG_MODULES |
1da177e4c
|
848 849 |
if (ops == NULL && kind != NULL) { char name[IFNAMSIZ]; |
1e90474c3
|
850 |
if (nla_strlcpy(name, kind, IFNAMSIZ) < IFNAMSIZ) { |
1da177e4c
|
851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 |
/* We dropped the RTNL semaphore in order to * perform the module load. So, even if we * succeeded in loading the module we have to * tell the caller to replay the request. We * indicate this using -EAGAIN. * We replay the request because the device may * go away in the mean time. */ rtnl_unlock(); request_module("sch_%s", name); rtnl_lock(); ops = qdisc_lookup_ops(kind); if (ops != NULL) { /* We will try again qdisc_lookup_ops, * so don't keep a reference. */ module_put(ops->owner); err = -EAGAIN; goto err_out; } } } #endif |
b9e2cc0f0
|
874 |
err = -ENOENT; |
1da177e4c
|
875 876 |
if (ops == NULL) goto err_out; |
5ce2d488f
|
877 |
sch = qdisc_alloc(dev_queue, ops); |
3d54b82fd
|
878 879 |
if (IS_ERR(sch)) { err = PTR_ERR(sch); |
1da177e4c
|
880 |
goto err_out2; |
3d54b82fd
|
881 |
} |
1da177e4c
|
882 |
|
ffc8fefaf
|
883 |
sch->parent = parent; |
3d54b82fd
|
884 |
if (handle == TC_H_INGRESS) { |
1da177e4c
|
885 |
sch->flags |= TCQ_F_INGRESS; |
3d54b82fd
|
886 |
handle = TC_H_MAKE(TC_H_INGRESS, 0); |
25bfcd5a7
|
887 |
lockdep_set_class(qdisc_lock(sch), &qdisc_rx_lock); |
fd44de7cc
|
888 |
} else { |
fd44de7cc
|
889 890 891 892 893 894 |
if (handle == 0) { handle = qdisc_alloc_handle(dev); err = -ENOMEM; if (handle == 0) goto err_out3; } |
25bfcd5a7
|
895 |
lockdep_set_class(qdisc_lock(sch), &qdisc_tx_lock); |
1abbe1394
|
896 897 |
if (!netif_is_multiqueue(dev)) sch->flags |= TCQ_F_ONETXQUEUE; |
1da177e4c
|
898 |
} |
3d54b82fd
|
899 |
sch->handle = handle; |
1da177e4c
|
900 |
|
1e90474c3
|
901 |
if (!ops->init || (err = ops->init(sch, tca[TCA_OPTIONS])) == 0) { |
175f9c1bb
|
902 903 904 905 |
if (tca[TCA_STAB]) { stab = qdisc_get_stab(tca[TCA_STAB]); if (IS_ERR(stab)) { err = PTR_ERR(stab); |
7c64b9f3f
|
906 |
goto err_out4; |
175f9c1bb
|
907 |
} |
a2da570d6
|
908 |
rcu_assign_pointer(sch->stab, stab); |
175f9c1bb
|
909 |
} |
1e90474c3
|
910 |
if (tca[TCA_RATE]) { |
f6f9b93f1
|
911 |
spinlock_t *root_lock; |
23bcf634c
|
912 913 914 |
err = -EOPNOTSUPP; if (sch->flags & TCQ_F_MQROOT) goto err_out4; |
f6f9b93f1
|
915 |
if ((sch->parent != TC_H_ROOT) && |
23bcf634c
|
916 917 |
!(sch->flags & TCQ_F_INGRESS) && (!p || !(p->flags & TCQ_F_MQROOT))) |
f6f9b93f1
|
918 919 920 |
root_lock = qdisc_root_sleeping_lock(sch); else root_lock = qdisc_lock(sch); |
023e09a76
|
921 |
err = gen_new_estimator(&sch->bstats, &sch->rate_est, |
f6f9b93f1
|
922 |
root_lock, tca[TCA_RATE]); |
23bcf634c
|
923 924 |
if (err) goto err_out4; |
023e09a76
|
925 |
} |
f6e0b239a
|
926 927 |
qdisc_list_add(sch); |
1da177e4c
|
928 |
|
1da177e4c
|
929 930 931 932 |
return sch; } err_out3: dev_put(dev); |
3d54b82fd
|
933 |
kfree((char *) sch - sch->padded); |
1da177e4c
|
934 935 936 937 |
err_out2: module_put(ops->owner); err_out: *errp = err; |
1da177e4c
|
938 |
return NULL; |
23bcf634c
|
939 940 941 942 943 944 |
err_out4: /* * Any broken qdiscs that would require a ops->reset() here? * The qdisc was never in action so it shouldn't be necessary. */ |
a2da570d6
|
945 |
qdisc_put_stab(rtnl_dereference(sch->stab)); |
23bcf634c
|
946 947 948 |
if (ops->destroy) ops->destroy(sch); goto err_out3; |
1da177e4c
|
949 |
} |
1e90474c3
|
950 |
static int qdisc_change(struct Qdisc *sch, struct nlattr **tca) |
1da177e4c
|
951 |
{ |
a2da570d6
|
952 |
struct qdisc_size_table *ostab, *stab = NULL; |
175f9c1bb
|
953 |
int err = 0; |
1da177e4c
|
954 |
|
175f9c1bb
|
955 |
if (tca[TCA_OPTIONS]) { |
1da177e4c
|
956 957 |
if (sch->ops->change == NULL) return -EINVAL; |
1e90474c3
|
958 |
err = sch->ops->change(sch, tca[TCA_OPTIONS]); |
1da177e4c
|
959 960 961 |
if (err) return err; } |
175f9c1bb
|
962 963 964 965 966 967 |
if (tca[TCA_STAB]) { stab = qdisc_get_stab(tca[TCA_STAB]); if (IS_ERR(stab)) return PTR_ERR(stab); } |
a2da570d6
|
968 969 970 |
ostab = rtnl_dereference(sch->stab); rcu_assign_pointer(sch->stab, stab); qdisc_put_stab(ostab); |
175f9c1bb
|
971 |
|
23bcf634c
|
972 |
if (tca[TCA_RATE]) { |
71bcb09a5
|
973 974 |
/* NB: ignores errors from replace_estimator because change can't be undone. */ |
23bcf634c
|
975 976 |
if (sch->flags & TCQ_F_MQROOT) goto out; |
1da177e4c
|
977 |
gen_replace_estimator(&sch->bstats, &sch->rate_est, |
71bcb09a5
|
978 979 |
qdisc_root_sleeping_lock(sch), tca[TCA_RATE]); |
23bcf634c
|
980 981 |
} out: |
1da177e4c
|
982 983 |
return 0; } |
cc7ec456f
|
984 985 |
struct check_loop_arg { struct qdisc_walker w; |
1da177e4c
|
986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 |
struct Qdisc *p; int depth; }; static int check_loop_fn(struct Qdisc *q, unsigned long cl, struct qdisc_walker *w); static int check_loop(struct Qdisc *q, struct Qdisc *p, int depth) { struct check_loop_arg arg; if (q->ops->cl_ops == NULL) return 0; arg.w.stop = arg.w.skip = arg.w.count = 0; arg.w.fn = check_loop_fn; arg.depth = depth; arg.p = p; q->ops->cl_ops->walk(q, &arg.w); return arg.w.stop ? -ELOOP : 0; } static int check_loop_fn(struct Qdisc *q, unsigned long cl, struct qdisc_walker *w) { struct Qdisc *leaf; |
20fea08b5
|
1011 |
const struct Qdisc_class_ops *cops = q->ops->cl_ops; |
1da177e4c
|
1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 |
struct check_loop_arg *arg = (struct check_loop_arg *)w; leaf = cops->leaf(q, cl); if (leaf) { if (leaf == arg->p || arg->depth > 7) return -ELOOP; return check_loop(leaf, arg->p, arg->depth + 1); } return 0; } /* * Delete/get qdisc. */ |
661d2967b
|
1026 |
static int tc_get_qdisc(struct sk_buff *skb, struct nlmsghdr *n) |
1da177e4c
|
1027 |
{ |
3b1e0a655
|
1028 |
struct net *net = sock_net(skb->sk); |
02ef22ca4
|
1029 |
struct tcmsg *tcm = nlmsg_data(n); |
1e90474c3
|
1030 |
struct nlattr *tca[TCA_MAX + 1]; |
1da177e4c
|
1031 |
struct net_device *dev; |
de179c8c1
|
1032 |
u32 clid; |
1da177e4c
|
1033 1034 1035 |
struct Qdisc *q = NULL; struct Qdisc *p = NULL; int err; |
4e8bbb819
|
1036 |
if ((n->nlmsg_type != RTM_GETQDISC) && |
5f013c9bc
|
1037 |
!netlink_ns_capable(skb, net->user_ns, CAP_NET_ADMIN)) |
dfc47ef86
|
1038 |
return -EPERM; |
1e90474c3
|
1039 1040 1041 |
err = nlmsg_parse(n, sizeof(*tcm), tca, TCA_MAX, NULL); if (err < 0) return err; |
de179c8c1
|
1042 1043 1044 1045 1046 |
dev = __dev_get_by_index(net, tcm->tcm_ifindex); if (!dev) return -ENODEV; clid = tcm->tcm_parent; |
1da177e4c
|
1047 1048 1049 |
if (clid) { if (clid != TC_H_ROOT) { if (TC_H_MAJ(clid) != TC_H_MAJ(TC_H_INGRESS)) { |
cc7ec456f
|
1050 1051 |
p = qdisc_lookup(dev, TC_H_MAJ(clid)); if (!p) |
1da177e4c
|
1052 1053 |
return -ENOENT; q = qdisc_leaf(p, clid); |
cc7ec456f
|
1054 1055 |
} else if (dev_ingress_queue(dev)) { q = dev_ingress_queue(dev)->qdisc_sleeping; |
10297b993
|
1056 |
} |
1da177e4c
|
1057 |
} else { |
af356afa0
|
1058 |
q = dev->qdisc; |
1da177e4c
|
1059 1060 1061 1062 1063 1064 1065 |
} if (!q) return -ENOENT; if (tcm->tcm_handle && q->handle != tcm->tcm_handle) return -EINVAL; } else { |
cc7ec456f
|
1066 1067 |
q = qdisc_lookup(dev, tcm->tcm_handle); if (!q) |
1da177e4c
|
1068 1069 |
return -ENOENT; } |
1e90474c3
|
1070 |
if (tca[TCA_KIND] && nla_strcmp(tca[TCA_KIND], q->ops->id)) |
1da177e4c
|
1071 1072 1073 1074 1075 1076 1077 |
return -EINVAL; if (n->nlmsg_type == RTM_DELQDISC) { if (!clid) return -EINVAL; if (q->handle == 0) return -ENOENT; |
cc7ec456f
|
1078 1079 |
err = qdisc_graft(dev, p, skb, n, clid, NULL, q); if (err != 0) |
1da177e4c
|
1080 |
return err; |
1da177e4c
|
1081 |
} else { |
7316ae88c
|
1082 |
qdisc_notify(net, skb, n, clid, NULL, q); |
1da177e4c
|
1083 1084 1085 1086 1087 |
} return 0; } /* |
cc7ec456f
|
1088 |
* Create/change qdisc. |
1da177e4c
|
1089 |
*/ |
661d2967b
|
1090 |
static int tc_modify_qdisc(struct sk_buff *skb, struct nlmsghdr *n) |
1da177e4c
|
1091 |
{ |
3b1e0a655
|
1092 |
struct net *net = sock_net(skb->sk); |
1da177e4c
|
1093 |
struct tcmsg *tcm; |
1e90474c3
|
1094 |
struct nlattr *tca[TCA_MAX + 1]; |
1da177e4c
|
1095 1096 1097 1098 |
struct net_device *dev; u32 clid; struct Qdisc *q, *p; int err; |
5f013c9bc
|
1099 |
if (!netlink_ns_capable(skb, net->user_ns, CAP_NET_ADMIN)) |
dfc47ef86
|
1100 |
return -EPERM; |
1da177e4c
|
1101 1102 |
replay: /* Reinit, just in case something touches this. */ |
de179c8c1
|
1103 1104 1105 |
err = nlmsg_parse(n, sizeof(*tcm), tca, TCA_MAX, NULL); if (err < 0) return err; |
02ef22ca4
|
1106 |
tcm = nlmsg_data(n); |
1da177e4c
|
1107 1108 |
clid = tcm->tcm_parent; q = p = NULL; |
cc7ec456f
|
1109 1110 |
dev = __dev_get_by_index(net, tcm->tcm_ifindex); if (!dev) |
1da177e4c
|
1111 |
return -ENODEV; |
1e90474c3
|
1112 |
|
1da177e4c
|
1113 1114 1115 |
if (clid) { if (clid != TC_H_ROOT) { if (clid != TC_H_INGRESS) { |
cc7ec456f
|
1116 1117 |
p = qdisc_lookup(dev, TC_H_MAJ(clid)); if (!p) |
1da177e4c
|
1118 1119 |
return -ENOENT; q = qdisc_leaf(p, clid); |
cc7ec456f
|
1120 1121 |
} else if (dev_ingress_queue_create(dev)) { q = dev_ingress_queue(dev)->qdisc_sleeping; |
1da177e4c
|
1122 1123 |
} } else { |
af356afa0
|
1124 |
q = dev->qdisc; |
1da177e4c
|
1125 1126 1127 1128 1129 1130 1131 1132 |
} /* It may be default qdisc, ignore it */ if (q && q->handle == 0) q = NULL; if (!q || !tcm->tcm_handle || q->handle != tcm->tcm_handle) { if (tcm->tcm_handle) { |
cc7ec456f
|
1133 |
if (q && !(n->nlmsg_flags & NLM_F_REPLACE)) |
1da177e4c
|
1134 1135 1136 |
return -EEXIST; if (TC_H_MIN(tcm->tcm_handle)) return -EINVAL; |
cc7ec456f
|
1137 1138 |
q = qdisc_lookup(dev, tcm->tcm_handle); if (!q) |
1da177e4c
|
1139 |
goto create_n_graft; |
cc7ec456f
|
1140 |
if (n->nlmsg_flags & NLM_F_EXCL) |
1da177e4c
|
1141 |
return -EEXIST; |
1e90474c3
|
1142 |
if (tca[TCA_KIND] && nla_strcmp(tca[TCA_KIND], q->ops->id)) |
1da177e4c
|
1143 1144 1145 1146 1147 1148 1149 |
return -EINVAL; if (q == p || (p && check_loop(q, p, 0))) return -ELOOP; atomic_inc(&q->refcnt); goto graft; } else { |
cc7ec456f
|
1150 |
if (!q) |
1da177e4c
|
1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 |
goto create_n_graft; /* This magic test requires explanation. * * We know, that some child q is already * attached to this parent and have choice: * either to change it or to create/graft new one. * * 1. We are allowed to create/graft only * if CREATE and REPLACE flags are set. * * 2. If EXCL is set, requestor wanted to say, * that qdisc tcm_handle is not expected * to exist, so that we choose create/graft too. * * 3. The last case is when no flags are set. * Alas, it is sort of hole in API, we * cannot decide what to do unambiguously. * For now we select create/graft, if * user gave KIND, which does not match existing. */ |
cc7ec456f
|
1172 1173 1174 |
if ((n->nlmsg_flags & NLM_F_CREATE) && (n->nlmsg_flags & NLM_F_REPLACE) && ((n->nlmsg_flags & NLM_F_EXCL) || |
1e90474c3
|
1175 1176 |
(tca[TCA_KIND] && nla_strcmp(tca[TCA_KIND], q->ops->id)))) |
1da177e4c
|
1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 |
goto create_n_graft; } } } else { if (!tcm->tcm_handle) return -EINVAL; q = qdisc_lookup(dev, tcm->tcm_handle); } /* Change qdisc parameters */ if (q == NULL) return -ENOENT; |
cc7ec456f
|
1189 |
if (n->nlmsg_flags & NLM_F_EXCL) |
1da177e4c
|
1190 |
return -EEXIST; |
1e90474c3
|
1191 |
if (tca[TCA_KIND] && nla_strcmp(tca[TCA_KIND], q->ops->id)) |
1da177e4c
|
1192 1193 1194 |
return -EINVAL; err = qdisc_change(q, tca); if (err == 0) |
7316ae88c
|
1195 |
qdisc_notify(net, skb, n, clid, NULL, q); |
1da177e4c
|
1196 1197 1198 |
return err; create_n_graft: |
cc7ec456f
|
1199 |
if (!(n->nlmsg_flags & NLM_F_CREATE)) |
1da177e4c
|
1200 |
return -ENOENT; |
24824a09e
|
1201 1202 1203 1204 1205 1206 1207 1208 |
if (clid == TC_H_INGRESS) { if (dev_ingress_queue(dev)) q = qdisc_create(dev, dev_ingress_queue(dev), p, tcm->tcm_parent, tcm->tcm_parent, tca, &err); else err = -ENOENT; } else { |
926e61b7c
|
1209 |
struct netdev_queue *dev_queue; |
6ec1c69a8
|
1210 1211 |
if (p && p->ops->cl_ops && p->ops->cl_ops->select_queue) |
926e61b7c
|
1212 1213 1214 1215 1216 |
dev_queue = p->ops->cl_ops->select_queue(p, tcm); else if (p) dev_queue = p->dev_queue; else dev_queue = netdev_get_tx_queue(dev, 0); |
6ec1c69a8
|
1217 |
|
926e61b7c
|
1218 |
q = qdisc_create(dev, dev_queue, p, |
bb949fbd1
|
1219 |
tcm->tcm_parent, tcm->tcm_handle, |
ffc8fefaf
|
1220 |
tca, &err); |
6ec1c69a8
|
1221 |
} |
1da177e4c
|
1222 1223 1224 1225 1226 1227 1228 |
if (q == NULL) { if (err == -EAGAIN) goto replay; return err; } graft: |
e5befbd95
|
1229 1230 1231 1232 1233 |
err = qdisc_graft(dev, p, skb, n, clid, q, NULL); if (err) { if (q) qdisc_destroy(q); return err; |
1da177e4c
|
1234 |
} |
e5befbd95
|
1235 |
|
1da177e4c
|
1236 1237 1238 1239 |
return 0; } static int tc_fill_qdisc(struct sk_buff *skb, struct Qdisc *q, u32 clid, |
15e473046
|
1240 |
u32 portid, u32 seq, u16 flags, int event) |
1da177e4c
|
1241 1242 1243 |
{ struct tcmsg *tcm; struct nlmsghdr *nlh; |
27a884dc3
|
1244 |
unsigned char *b = skb_tail_pointer(skb); |
1da177e4c
|
1245 |
struct gnet_dump d; |
a2da570d6
|
1246 |
struct qdisc_size_table *stab; |
1da177e4c
|
1247 |
|
fba373d2b
|
1248 |
cond_resched(); |
15e473046
|
1249 |
nlh = nlmsg_put(skb, portid, seq, event, sizeof(*tcm), flags); |
02ef22ca4
|
1250 1251 1252 |
if (!nlh) goto out_nlmsg_trim; tcm = nlmsg_data(nlh); |
1da177e4c
|
1253 |
tcm->tcm_family = AF_UNSPEC; |
9ef1d4c7c
|
1254 1255 |
tcm->tcm__pad1 = 0; tcm->tcm__pad2 = 0; |
5ce2d488f
|
1256 |
tcm->tcm_ifindex = qdisc_dev(q)->ifindex; |
1da177e4c
|
1257 1258 1259 |
tcm->tcm_parent = clid; tcm->tcm_handle = q->handle; tcm->tcm_info = atomic_read(&q->refcnt); |
1b34ec43c
|
1260 1261 |
if (nla_put_string(skb, TCA_KIND, q->ops->id)) goto nla_put_failure; |
1da177e4c
|
1262 |
if (q->ops->dump && q->ops->dump(q, skb) < 0) |
1e90474c3
|
1263 |
goto nla_put_failure; |
1da177e4c
|
1264 |
q->qstats.qlen = q->q.qlen; |
a2da570d6
|
1265 1266 |
stab = rtnl_dereference(q->stab); if (stab && qdisc_dump_stab(skb, stab) < 0) |
175f9c1bb
|
1267 |
goto nla_put_failure; |
102396ae6
|
1268 1269 |
if (gnet_stats_start_copy_compat(skb, TCA_STATS2, TCA_STATS, TCA_XSTATS, qdisc_root_sleeping_lock(q), &d) < 0) |
1e90474c3
|
1270 |
goto nla_put_failure; |
1da177e4c
|
1271 1272 |
if (q->ops->dump_stats && q->ops->dump_stats(q, &d) < 0) |
1e90474c3
|
1273 |
goto nla_put_failure; |
1da177e4c
|
1274 1275 |
if (gnet_stats_copy_basic(&d, &q->bstats) < 0 || |
d250a5f90
|
1276 |
gnet_stats_copy_rate_est(&d, &q->bstats, &q->rate_est) < 0 || |
1da177e4c
|
1277 |
gnet_stats_copy_queue(&d, &q->qstats) < 0) |
1e90474c3
|
1278 |
goto nla_put_failure; |
10297b993
|
1279 |
|
1da177e4c
|
1280 |
if (gnet_stats_finish_copy(&d) < 0) |
1e90474c3
|
1281 |
goto nla_put_failure; |
10297b993
|
1282 |
|
27a884dc3
|
1283 |
nlh->nlmsg_len = skb_tail_pointer(skb) - b; |
1da177e4c
|
1284 |
return skb->len; |
02ef22ca4
|
1285 |
out_nlmsg_trim: |
1e90474c3
|
1286 |
nla_put_failure: |
dc5fc579b
|
1287 |
nlmsg_trim(skb, b); |
1da177e4c
|
1288 1289 |
return -1; } |
53b0f0804
|
1290 1291 1292 1293 |
static bool tc_qdisc_dump_ignore(struct Qdisc *q) { return (q->flags & TCQ_F_BUILTIN) ? true : false; } |
7316ae88c
|
1294 1295 1296 |
static int qdisc_notify(struct net *net, struct sk_buff *oskb, struct nlmsghdr *n, u32 clid, struct Qdisc *old, struct Qdisc *new) |
1da177e4c
|
1297 1298 |
{ struct sk_buff *skb; |
15e473046
|
1299 |
u32 portid = oskb ? NETLINK_CB(oskb).portid : 0; |
1da177e4c
|
1300 1301 1302 1303 |
skb = alloc_skb(NLMSG_GOODSIZE, GFP_KERNEL); if (!skb) return -ENOBUFS; |
53b0f0804
|
1304 |
if (old && !tc_qdisc_dump_ignore(old)) { |
15e473046
|
1305 |
if (tc_fill_qdisc(skb, old, clid, portid, n->nlmsg_seq, |
cc7ec456f
|
1306 |
0, RTM_DELQDISC) < 0) |
1da177e4c
|
1307 1308 |
goto err_out; } |
53b0f0804
|
1309 |
if (new && !tc_qdisc_dump_ignore(new)) { |
15e473046
|
1310 |
if (tc_fill_qdisc(skb, new, clid, portid, n->nlmsg_seq, |
cc7ec456f
|
1311 |
old ? NLM_F_REPLACE : 0, RTM_NEWQDISC) < 0) |
1da177e4c
|
1312 1313 1314 1315 |
goto err_out; } if (skb->len) |
15e473046
|
1316 |
return rtnetlink_send(skb, net, portid, RTNLGRP_TC, |
cc7ec456f
|
1317 |
n->nlmsg_flags & NLM_F_ECHO); |
1da177e4c
|
1318 1319 1320 1321 1322 |
err_out: kfree_skb(skb); return -EINVAL; } |
307236730
|
1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 |
static int tc_dump_qdisc_root(struct Qdisc *root, struct sk_buff *skb, struct netlink_callback *cb, int *q_idx_p, int s_q_idx) { int ret = 0, q_idx = *q_idx_p; struct Qdisc *q; if (!root) return 0; q = root; if (q_idx < s_q_idx) { q_idx++; } else { if (!tc_qdisc_dump_ignore(q) && |
15e473046
|
1338 |
tc_fill_qdisc(skb, q, q->parent, NETLINK_CB(cb->skb).portid, |
307236730
|
1339 1340 1341 1342 1343 1344 1345 1346 1347 |
cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWQDISC) <= 0) goto done; q_idx++; } list_for_each_entry(q, &root->list, list) { if (q_idx < s_q_idx) { q_idx++; continue; } |
cc7ec456f
|
1348 |
if (!tc_qdisc_dump_ignore(q) && |
15e473046
|
1349 |
tc_fill_qdisc(skb, q, q->parent, NETLINK_CB(cb->skb).portid, |
307236730
|
1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 |
cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWQDISC) <= 0) goto done; q_idx++; } out: *q_idx_p = q_idx; return ret; done: ret = -1; goto out; } |
1da177e4c
|
1362 1363 |
static int tc_dump_qdisc(struct sk_buff *skb, struct netlink_callback *cb) { |
3b1e0a655
|
1364 |
struct net *net = sock_net(skb->sk); |
1da177e4c
|
1365 1366 1367 |
int idx, q_idx; int s_idx, s_q_idx; struct net_device *dev; |
1da177e4c
|
1368 1369 1370 |
s_idx = cb->args[0]; s_q_idx = q_idx = cb->args[1]; |
f1e9016da
|
1371 |
|
7562f876c
|
1372 |
idx = 0; |
15dc36ebb
|
1373 1374 |
ASSERT_RTNL(); for_each_netdev(net, dev) { |
307236730
|
1375 |
struct netdev_queue *dev_queue; |
1da177e4c
|
1376 |
if (idx < s_idx) |
7562f876c
|
1377 |
goto cont; |
1da177e4c
|
1378 1379 |
if (idx > s_idx) s_q_idx = 0; |
1da177e4c
|
1380 |
q_idx = 0; |
307236730
|
1381 |
|
af356afa0
|
1382 |
if (tc_dump_qdisc_root(dev->qdisc, skb, cb, &q_idx, s_q_idx) < 0) |
307236730
|
1383 |
goto done; |
24824a09e
|
1384 1385 1386 1387 |
dev_queue = dev_ingress_queue(dev); if (dev_queue && tc_dump_qdisc_root(dev_queue->qdisc_sleeping, skb, cb, &q_idx, s_q_idx) < 0) |
307236730
|
1388 |
goto done; |
7562f876c
|
1389 1390 |
cont: idx++; |
1da177e4c
|
1391 1392 1393 |
} done: |
1da177e4c
|
1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 |
cb->args[0] = idx; cb->args[1] = q_idx; return skb->len; } /************************************************ * Traffic classes manipulation. * ************************************************/ |
661d2967b
|
1405 |
static int tc_ctl_tclass(struct sk_buff *skb, struct nlmsghdr *n) |
1da177e4c
|
1406 |
{ |
3b1e0a655
|
1407 |
struct net *net = sock_net(skb->sk); |
02ef22ca4
|
1408 |
struct tcmsg *tcm = nlmsg_data(n); |
1e90474c3
|
1409 |
struct nlattr *tca[TCA_MAX + 1]; |
1da177e4c
|
1410 1411 |
struct net_device *dev; struct Qdisc *q = NULL; |
20fea08b5
|
1412 |
const struct Qdisc_class_ops *cops; |
1da177e4c
|
1413 1414 |
unsigned long cl = 0; unsigned long new_cl; |
de179c8c1
|
1415 1416 1417 |
u32 portid; u32 clid; u32 qid; |
1da177e4c
|
1418 |
int err; |
4e8bbb819
|
1419 |
if ((n->nlmsg_type != RTM_GETTCLASS) && |
5f013c9bc
|
1420 |
!netlink_ns_capable(skb, net->user_ns, CAP_NET_ADMIN)) |
dfc47ef86
|
1421 |
return -EPERM; |
1e90474c3
|
1422 1423 1424 |
err = nlmsg_parse(n, sizeof(*tcm), tca, TCA_MAX, NULL); if (err < 0) return err; |
de179c8c1
|
1425 1426 1427 |
dev = __dev_get_by_index(net, tcm->tcm_ifindex); if (!dev) return -ENODEV; |
1da177e4c
|
1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 |
/* parent == TC_H_UNSPEC - unspecified parent. parent == TC_H_ROOT - class is root, which has no parent. parent == X:0 - parent is root class. parent == X:Y - parent is a node in hierarchy. parent == 0:Y - parent is X:Y, where X:0 is qdisc. handle == 0:0 - generate handle from kernel pool. handle == 0:Y - class is X:Y, where X:0 is qdisc. handle == X:Y - clear. handle == X:0 - root class. */ /* Step 1. Determine qdisc handle X:0 */ |
de179c8c1
|
1442 1443 1444 |
portid = tcm->tcm_parent; clid = tcm->tcm_handle; qid = TC_H_MAJ(clid); |
15e473046
|
1445 1446 |
if (portid != TC_H_ROOT) { u32 qid1 = TC_H_MAJ(portid); |
1da177e4c
|
1447 1448 1449 1450 1451 1452 1453 1454 |
if (qid && qid1) { /* If both majors are known, they must be identical. */ if (qid != qid1) return -EINVAL; } else if (qid1) { qid = qid1; } else if (qid == 0) |
af356afa0
|
1455 |
qid = dev->qdisc->handle; |
1da177e4c
|
1456 1457 |
/* Now qid is genuine qdisc handle consistent |
cc7ec456f
|
1458 1459 |
* both with parent and child. * |
15e473046
|
1460 |
* TC_H_MAJ(portid) still may be unspecified, complete it now. |
1da177e4c
|
1461 |
*/ |
15e473046
|
1462 1463 |
if (portid) portid = TC_H_MAKE(qid, portid); |
1da177e4c
|
1464 1465 |
} else { if (qid == 0) |
af356afa0
|
1466 |
qid = dev->qdisc->handle; |
1da177e4c
|
1467 1468 1469 |
} /* OK. Locate qdisc */ |
cc7ec456f
|
1470 1471 |
q = qdisc_lookup(dev, qid); if (!q) |
1da177e4c
|
1472 1473 1474 1475 1476 1477 1478 1479 1480 |
return -ENOENT; /* An check that it supports classes */ cops = q->ops->cl_ops; if (cops == NULL) return -EINVAL; /* Now try to get class */ if (clid == 0) { |
15e473046
|
1481 |
if (portid == TC_H_ROOT) |
1da177e4c
|
1482 1483 1484 1485 1486 1487 1488 1489 1490 |
clid = qid; } else clid = TC_H_MAKE(qid, clid); if (clid) cl = cops->get(q, clid); if (cl == 0) { err = -ENOENT; |
cc7ec456f
|
1491 1492 |
if (n->nlmsg_type != RTM_NEWTCLASS || !(n->nlmsg_flags & NLM_F_CREATE)) |
1da177e4c
|
1493 1494 1495 |
goto out; } else { switch (n->nlmsg_type) { |
10297b993
|
1496 |
case RTM_NEWTCLASS: |
1da177e4c
|
1497 |
err = -EEXIST; |
cc7ec456f
|
1498 |
if (n->nlmsg_flags & NLM_F_EXCL) |
1da177e4c
|
1499 1500 1501 |
goto out; break; case RTM_DELTCLASS: |
de6d5cdf8
|
1502 1503 1504 |
err = -EOPNOTSUPP; if (cops->delete) err = cops->delete(q, cl); |
1da177e4c
|
1505 |
if (err == 0) |
7316ae88c
|
1506 |
tclass_notify(net, skb, n, q, cl, RTM_DELTCLASS); |
1da177e4c
|
1507 1508 |
goto out; case RTM_GETTCLASS: |
7316ae88c
|
1509 |
err = tclass_notify(net, skb, n, q, cl, RTM_NEWTCLASS); |
1da177e4c
|
1510 1511 1512 1513 1514 1515 1516 1517 |
goto out; default: err = -EINVAL; goto out; } } new_cl = cl; |
de6d5cdf8
|
1518 1519 |
err = -EOPNOTSUPP; if (cops->change) |
15e473046
|
1520 |
err = cops->change(q, clid, portid, tca, &new_cl); |
1da177e4c
|
1521 |
if (err == 0) |
7316ae88c
|
1522 |
tclass_notify(net, skb, n, q, new_cl, RTM_NEWTCLASS); |
1da177e4c
|
1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 |
out: if (cl) cops->put(q, cl); return err; } static int tc_fill_tclass(struct sk_buff *skb, struct Qdisc *q, unsigned long cl, |
15e473046
|
1534 |
u32 portid, u32 seq, u16 flags, int event) |
1da177e4c
|
1535 1536 1537 |
{ struct tcmsg *tcm; struct nlmsghdr *nlh; |
27a884dc3
|
1538 |
unsigned char *b = skb_tail_pointer(skb); |
1da177e4c
|
1539 |
struct gnet_dump d; |
20fea08b5
|
1540 |
const struct Qdisc_class_ops *cl_ops = q->ops->cl_ops; |
1da177e4c
|
1541 |
|
fba373d2b
|
1542 |
cond_resched(); |
15e473046
|
1543 |
nlh = nlmsg_put(skb, portid, seq, event, sizeof(*tcm), flags); |
02ef22ca4
|
1544 1545 1546 |
if (!nlh) goto out_nlmsg_trim; tcm = nlmsg_data(nlh); |
1da177e4c
|
1547 |
tcm->tcm_family = AF_UNSPEC; |
16ebb5e0b
|
1548 1549 |
tcm->tcm__pad1 = 0; tcm->tcm__pad2 = 0; |
5ce2d488f
|
1550 |
tcm->tcm_ifindex = qdisc_dev(q)->ifindex; |
1da177e4c
|
1551 1552 1553 |
tcm->tcm_parent = q->handle; tcm->tcm_handle = q->handle; tcm->tcm_info = 0; |
1b34ec43c
|
1554 1555 |
if (nla_put_string(skb, TCA_KIND, q->ops->id)) goto nla_put_failure; |
1da177e4c
|
1556 |
if (cl_ops->dump && cl_ops->dump(q, cl, skb, tcm) < 0) |
1e90474c3
|
1557 |
goto nla_put_failure; |
1da177e4c
|
1558 |
|
102396ae6
|
1559 1560 |
if (gnet_stats_start_copy_compat(skb, TCA_STATS2, TCA_STATS, TCA_XSTATS, qdisc_root_sleeping_lock(q), &d) < 0) |
1e90474c3
|
1561 |
goto nla_put_failure; |
1da177e4c
|
1562 1563 |
if (cl_ops->dump_stats && cl_ops->dump_stats(q, cl, &d) < 0) |
1e90474c3
|
1564 |
goto nla_put_failure; |
1da177e4c
|
1565 1566 |
if (gnet_stats_finish_copy(&d) < 0) |
1e90474c3
|
1567 |
goto nla_put_failure; |
1da177e4c
|
1568 |
|
27a884dc3
|
1569 |
nlh->nlmsg_len = skb_tail_pointer(skb) - b; |
1da177e4c
|
1570 |
return skb->len; |
02ef22ca4
|
1571 |
out_nlmsg_trim: |
1e90474c3
|
1572 |
nla_put_failure: |
dc5fc579b
|
1573 |
nlmsg_trim(skb, b); |
1da177e4c
|
1574 1575 |
return -1; } |
7316ae88c
|
1576 1577 1578 |
static int tclass_notify(struct net *net, struct sk_buff *oskb, struct nlmsghdr *n, struct Qdisc *q, unsigned long cl, int event) |
1da177e4c
|
1579 1580 |
{ struct sk_buff *skb; |
15e473046
|
1581 |
u32 portid = oskb ? NETLINK_CB(oskb).portid : 0; |
1da177e4c
|
1582 1583 1584 1585 |
skb = alloc_skb(NLMSG_GOODSIZE, GFP_KERNEL); if (!skb) return -ENOBUFS; |
15e473046
|
1586 |
if (tc_fill_tclass(skb, q, cl, portid, n->nlmsg_seq, 0, event) < 0) { |
1da177e4c
|
1587 1588 1589 |
kfree_skb(skb); return -EINVAL; } |
15e473046
|
1590 |
return rtnetlink_send(skb, net, portid, RTNLGRP_TC, |
cc7ec456f
|
1591 |
n->nlmsg_flags & NLM_F_ECHO); |
1da177e4c
|
1592 |
} |
cc7ec456f
|
1593 1594 1595 1596 |
struct qdisc_dump_args { struct qdisc_walker w; struct sk_buff *skb; struct netlink_callback *cb; |
1da177e4c
|
1597 1598 1599 1600 1601 |
}; static int qdisc_class_dump(struct Qdisc *q, unsigned long cl, struct qdisc_walker *arg) { struct qdisc_dump_args *a = (struct qdisc_dump_args *)arg; |
15e473046
|
1602 |
return tc_fill_tclass(a->skb, q, cl, NETLINK_CB(a->cb->skb).portid, |
1da177e4c
|
1603 1604 |
a->cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWTCLASS); } |
307236730
|
1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 |
static int tc_dump_tclass_qdisc(struct Qdisc *q, struct sk_buff *skb, struct tcmsg *tcm, struct netlink_callback *cb, int *t_p, int s_t) { struct qdisc_dump_args arg; if (tc_qdisc_dump_ignore(q) || *t_p < s_t || !q->ops->cl_ops || (tcm->tcm_parent && TC_H_MAJ(tcm->tcm_parent) != q->handle)) { (*t_p)++; return 0; } if (*t_p > s_t) memset(&cb->args[1], 0, sizeof(cb->args)-sizeof(cb->args[0])); arg.w.fn = qdisc_class_dump; arg.skb = skb; arg.cb = cb; arg.w.stop = 0; arg.w.skip = cb->args[1]; arg.w.count = 0; q->ops->cl_ops->walk(q, &arg.w); cb->args[1] = arg.w.count; if (arg.w.stop) return -1; (*t_p)++; return 0; } static int tc_dump_tclass_root(struct Qdisc *root, struct sk_buff *skb, struct tcmsg *tcm, struct netlink_callback *cb, int *t_p, int s_t) { struct Qdisc *q; if (!root) return 0; if (tc_dump_tclass_qdisc(root, skb, tcm, cb, t_p, s_t) < 0) return -1; list_for_each_entry(q, &root->list, list) { if (tc_dump_tclass_qdisc(q, skb, tcm, cb, t_p, s_t) < 0) return -1; } return 0; } |
1da177e4c
|
1653 1654 |
static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb) { |
02ef22ca4
|
1655 |
struct tcmsg *tcm = nlmsg_data(cb->nlh); |
3b1e0a655
|
1656 |
struct net *net = sock_net(skb->sk); |
307236730
|
1657 |
struct netdev_queue *dev_queue; |
1da177e4c
|
1658 |
struct net_device *dev; |
307236730
|
1659 |
int t, s_t; |
1da177e4c
|
1660 |
|
573ce260b
|
1661 |
if (nlmsg_len(cb->nlh) < sizeof(*tcm)) |
1da177e4c
|
1662 |
return 0; |
cc7ec456f
|
1663 1664 |
dev = dev_get_by_index(net, tcm->tcm_ifindex); if (!dev) |
1da177e4c
|
1665 1666 1667 1668 |
return 0; s_t = cb->args[0]; t = 0; |
af356afa0
|
1669 |
if (tc_dump_tclass_root(dev->qdisc, skb, tcm, cb, &t, s_t) < 0) |
307236730
|
1670 |
goto done; |
24824a09e
|
1671 1672 1673 1674 |
dev_queue = dev_ingress_queue(dev); if (dev_queue && tc_dump_tclass_root(dev_queue->qdisc_sleeping, skb, tcm, cb, &t, s_t) < 0) |
307236730
|
1675 |
goto done; |
1da177e4c
|
1676 |
|
307236730
|
1677 |
done: |
1da177e4c
|
1678 1679 1680 1681 1682 1683 1684 |
cb->args[0] = t; dev_put(dev); return skb->len; } /* Main classifier routine: scans classifier chain attached |
cc7ec456f
|
1685 1686 |
* to this qdisc, (optionally) tests for protocol and asks * specific classifiers. |
1da177e4c
|
1687 |
*/ |
dc7f9f6e8
|
1688 |
int tc_classify_compat(struct sk_buff *skb, const struct tcf_proto *tp, |
73ca4918f
|
1689 1690 1691 |
struct tcf_result *res) { __be16 protocol = skb->protocol; |
cc7ec456f
|
1692 |
int err; |
73ca4918f
|
1693 1694 |
for (; tp; tp = tp->next) { |
cc7ec456f
|
1695 1696 1697 1698 1699 1700 |
if (tp->protocol != protocol && tp->protocol != htons(ETH_P_ALL)) continue; err = tp->classify(skb, tp, res); if (err >= 0) { |
73ca4918f
|
1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 |
#ifdef CONFIG_NET_CLS_ACT if (err != TC_ACT_RECLASSIFY && skb->tc_verd) skb->tc_verd = SET_TC_VERD(skb->tc_verd, 0); #endif return err; } } return -1; } EXPORT_SYMBOL(tc_classify_compat); |
dc7f9f6e8
|
1711 |
int tc_classify(struct sk_buff *skb, const struct tcf_proto *tp, |
73ca4918f
|
1712 |
struct tcf_result *res) |
1da177e4c
|
1713 1714 |
{ int err = 0; |
1da177e4c
|
1715 |
#ifdef CONFIG_NET_CLS_ACT |
dc7f9f6e8
|
1716 |
const struct tcf_proto *otp = tp; |
1da177e4c
|
1717 |
reclassify: |
52bc97470
|
1718 |
#endif |
1da177e4c
|
1719 |
|
73ca4918f
|
1720 |
err = tc_classify_compat(skb, tp, res); |
1da177e4c
|
1721 |
#ifdef CONFIG_NET_CLS_ACT |
73ca4918f
|
1722 1723 1724 1725 1726 |
if (err == TC_ACT_RECLASSIFY) { u32 verd = G_TC_VERD(skb->tc_verd); tp = otp; if (verd++ >= MAX_REC_LOOP) { |
e87cc4728
|
1727 1728 1729 1730 1731 |
net_notice_ratelimited("%s: packet reclassify loop rule prio %u protocol %02x ", tp->q->ops->id, tp->prio & 0xffff, ntohs(tp->protocol)); |
73ca4918f
|
1732 |
return TC_ACT_SHOT; |
1da177e4c
|
1733 |
} |
73ca4918f
|
1734 1735 |
skb->tc_verd = SET_TC_VERD(skb->tc_verd, verd); goto reclassify; |
1da177e4c
|
1736 |
} |
73ca4918f
|
1737 1738 |
#endif return err; |
1da177e4c
|
1739 |
} |
73ca4918f
|
1740 |
EXPORT_SYMBOL(tc_classify); |
1da177e4c
|
1741 |
|
a48b5a614
|
1742 1743 1744 1745 1746 1747 |
void tcf_destroy(struct tcf_proto *tp) { tp->ops->destroy(tp); module_put(tp->ops->owner); kfree(tp); } |
ff31ab56c
|
1748 |
void tcf_destroy_chain(struct tcf_proto **fl) |
a48b5a614
|
1749 1750 |
{ struct tcf_proto *tp; |
ff31ab56c
|
1751 1752 |
while ((tp = *fl) != NULL) { *fl = tp->next; |
a48b5a614
|
1753 1754 1755 1756 |
tcf_destroy(tp); } } EXPORT_SYMBOL(tcf_destroy_chain); |
1da177e4c
|
1757 1758 1759 |
#ifdef CONFIG_PROC_FS static int psched_show(struct seq_file *seq, void *v) { |
3c0cfc135
|
1760 1761 1762 |
struct timespec ts; hrtimer_get_res(CLOCK_MONOTONIC, &ts); |
1da177e4c
|
1763 1764 |
seq_printf(seq, "%08x %08x %08x %08x ", |
ca44d6e60
|
1765 |
(u32)NSEC_PER_USEC, (u32)PSCHED_TICKS2NS(1), |
514bca322
|
1766 |
1000000, |
3c0cfc135
|
1767 |
(u32)NSEC_PER_SEC/(u32)ktime_to_ns(timespec_to_ktime(ts))); |
1da177e4c
|
1768 1769 1770 1771 1772 1773 |
return 0; } static int psched_open(struct inode *inode, struct file *file) { |
7e5ab1578
|
1774 |
return single_open(file, psched_show, NULL); |
1da177e4c
|
1775 |
} |
da7071d7e
|
1776 |
static const struct file_operations psched_fops = { |
1da177e4c
|
1777 1778 1779 1780 1781 |
.owner = THIS_MODULE, .open = psched_open, .read = seq_read, .llseek = seq_lseek, .release = single_release, |
10297b993
|
1782 |
}; |
7316ae88c
|
1783 1784 1785 1786 |
static int __net_init psched_net_init(struct net *net) { struct proc_dir_entry *e; |
d4beaa66a
|
1787 |
e = proc_create("psched", 0, net->proc_net, &psched_fops); |
7316ae88c
|
1788 1789 1790 1791 1792 1793 1794 1795 |
if (e == NULL) return -ENOMEM; return 0; } static void __net_exit psched_net_exit(struct net *net) { |
ece31ffd5
|
1796 |
remove_proc_entry("psched", net->proc_net); |
7316ae88c
|
1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 |
} #else static int __net_init psched_net_init(struct net *net) { return 0; } static void __net_exit psched_net_exit(struct net *net) { } |
1da177e4c
|
1807 |
#endif |
7316ae88c
|
1808 1809 1810 1811 |
static struct pernet_operations psched_net_ops = { .init = psched_net_init, .exit = psched_net_exit, }; |
1da177e4c
|
1812 1813 |
static int __init pktsched_init(void) { |
7316ae88c
|
1814 1815 1816 1817 |
int err; err = register_pernet_subsys(&psched_net_ops); if (err) { |
cc7ec456f
|
1818 |
pr_err("pktsched_init: " |
7316ae88c
|
1819 1820 1821 1822 |
"cannot initialize per netns operations "); return err; } |
6da7c8fcb
|
1823 |
register_qdisc(&pfifo_fast_ops); |
1da177e4c
|
1824 1825 |
register_qdisc(&pfifo_qdisc_ops); register_qdisc(&bfifo_qdisc_ops); |
57dbb2d83
|
1826 |
register_qdisc(&pfifo_head_drop_qdisc_ops); |
6ec1c69a8
|
1827 |
register_qdisc(&mq_qdisc_ops); |
1da177e4c
|
1828 |
|
c7ac8679b
|
1829 1830 1831 1832 1833 1834 |
rtnl_register(PF_UNSPEC, RTM_NEWQDISC, tc_modify_qdisc, NULL, NULL); rtnl_register(PF_UNSPEC, RTM_DELQDISC, tc_get_qdisc, NULL, NULL); rtnl_register(PF_UNSPEC, RTM_GETQDISC, tc_get_qdisc, tc_dump_qdisc, NULL); rtnl_register(PF_UNSPEC, RTM_NEWTCLASS, tc_ctl_tclass, NULL, NULL); rtnl_register(PF_UNSPEC, RTM_DELTCLASS, tc_ctl_tclass, NULL, NULL); rtnl_register(PF_UNSPEC, RTM_GETTCLASS, tc_ctl_tclass, tc_dump_tclass, NULL); |
be577ddc2
|
1835 |
|
1da177e4c
|
1836 1837 1838 1839 |
return 0; } subsys_initcall(pktsched_init); |