sch_generic.h 20.6 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8
#ifndef __NET_SCHED_GENERIC_H
#define __NET_SCHED_GENERIC_H

#include <linux/netdevice.h>
#include <linux/types.h>
#include <linux/rcupdate.h>
#include <linux/pkt_sched.h>
#include <linux/pkt_cls.h>
9
#include <linux/percpu.h>
10
#include <linux/dynamic_queue_limits.h>
L
Linus Torvalds 已提交
11
#include <net/gen_stats.h>
12
#include <net/rtnetlink.h>
L
Linus Torvalds 已提交
13 14 15 16 17 18

struct Qdisc_ops;
struct qdisc_walker;
struct tcf_walker;
struct module;

E
Eric Dumazet 已提交
19
struct qdisc_rate_table {
L
Linus Torvalds 已提交
20 21 22 23 24 25
	struct tc_ratespec rate;
	u32		data[256];
	struct qdisc_rate_table *next;
	int		refcnt;
};

E
Eric Dumazet 已提交
26
enum qdisc_state_t {
27
	__QDISC_STATE_SCHED,
28
	__QDISC_STATE_DEACTIVATED,
29 30
};

31
struct qdisc_size_table {
E
Eric Dumazet 已提交
32
	struct rcu_head		rcu;
33 34 35 36 37 38
	struct list_head	list;
	struct tc_sizespec	szopts;
	int			refcnt;
	u16			data[];
};

39 40 41 42 43 44 45 46
/* similar to sk_buff_head, but skb->prev pointer is undefined. */
struct qdisc_skb_head {
	struct sk_buff	*head;
	struct sk_buff	*tail;
	__u32		qlen;
	spinlock_t	lock;
};

E
Eric Dumazet 已提交
47
struct Qdisc {
48 49 50 51
	int 			(*enqueue)(struct sk_buff *skb,
					   struct Qdisc *sch,
					   struct sk_buff **to_free);
	struct sk_buff *	(*dequeue)(struct Qdisc *sch);
52
	unsigned int		flags;
53
#define TCQ_F_BUILTIN		1
54 55 56
#define TCQ_F_INGRESS		2
#define TCQ_F_CAN_BYPASS	4
#define TCQ_F_MQROOT		8
57 58 59 60 61 62 63
#define TCQ_F_ONETXQUEUE	0x10 /* dequeue_skb() can assume all skbs are for
				      * q->dev_queue : It can test
				      * netif_xmit_frozen_or_stopped() before
				      * dequeueing next packet.
				      * Its true for MQ/MQPRIO slaves, or non
				      * multiqueue device.
				      */
64
#define TCQ_F_WARN_NONWC	(1 << 16)
65
#define TCQ_F_CPUSTATS		0x20 /* run using percpu statistics */
66 67 68
#define TCQ_F_NOPARENT		0x40 /* root of its hierarchy :
				      * qdisc_tree_decrease_qlen() should stop.
				      */
69
	u32			limit;
70
	const struct Qdisc_ops	*ops;
E
Eric Dumazet 已提交
71
	struct qdisc_size_table	__rcu *stab;
72
	struct hlist_node       hash;
L
Linus Torvalds 已提交
73 74
	u32			handle;
	u32			parent;
75 76
	void			*u32_node;

77 78
	struct netdev_queue	*dev_queue;

79
	struct net_rate_estimator __rcu *rate_est;
80 81 82
	struct gnet_stats_basic_cpu __percpu *cpu_bstats;
	struct gnet_stats_queue	__percpu *cpu_qstats;

83 84 85
	/*
	 * For performance sake on SMP, we put highly modified fields at the end
	 */
86
	struct sk_buff		*gso_skb ____cacheline_aligned_in_smp;
87
	struct qdisc_skb_head	q;
88
	struct gnet_stats_basic_packed bstats;
89
	seqcount_t		running;
90
	struct gnet_stats_queue	qstats;
91 92 93
	unsigned long		state;
	struct Qdisc            *next_sched;
	struct sk_buff		*skb_bad_txq;
94
	struct rcu_head		rcu_head;
95 96 97 98
	int			padded;
	atomic_t		refcnt;

	spinlock_t		busylock ____cacheline_aligned_in_smp;
L
Linus Torvalds 已提交
99 100
};

101
static inline bool qdisc_is_running(const struct Qdisc *qdisc)
102
{
103
	return (raw_read_seqcount(&qdisc->running) & 1) ? true : false;
104 105 106 107
}

static inline bool qdisc_run_begin(struct Qdisc *qdisc)
{
108 109
	if (qdisc_is_running(qdisc))
		return false;
110 111 112 113 114
	/* Variant of write_seqcount_begin() telling lockdep a trylock
	 * was attempted.
	 */
	raw_write_seqcount_begin(&qdisc->running);
	seqcount_acquire(&qdisc->running.dep_map, 0, 1, _RET_IP_);
115
	return true;
116 117 118 119
}

static inline void qdisc_run_end(struct Qdisc *qdisc)
{
120
	write_seqcount_end(&qdisc->running);
121 122
}

123 124 125 126 127 128 129 130 131 132 133 134 135 136 137
static inline bool qdisc_may_bulk(const struct Qdisc *qdisc)
{
	return qdisc->flags & TCQ_F_ONETXQUEUE;
}

static inline int qdisc_avail_bulklimit(const struct netdev_queue *txq)
{
#ifdef CONFIG_BQL
	/* Non-BQL migrated drivers will return 0, too. */
	return dql_avail(&txq->dql);
#else
	return 0;
#endif
}

E
Eric Dumazet 已提交
138
struct Qdisc_class_ops {
L
Linus Torvalds 已提交
139
	/* Child qdisc manipulation */
140
	struct netdev_queue *	(*select_queue)(struct Qdisc *, struct tcmsg *);
L
Linus Torvalds 已提交
141 142 143
	int			(*graft)(struct Qdisc *, unsigned long cl,
					struct Qdisc *, struct Qdisc **);
	struct Qdisc *		(*leaf)(struct Qdisc *, unsigned long cl);
144
	void			(*qlen_notify)(struct Qdisc *, unsigned long);
L
Linus Torvalds 已提交
145 146 147 148 149

	/* Class manipulation routines */
	unsigned long		(*get)(struct Qdisc *, u32 classid);
	void			(*put)(struct Qdisc *, unsigned long);
	int			(*change)(struct Qdisc *, u32, u32,
150
					struct nlattr **, unsigned long *);
L
Linus Torvalds 已提交
151 152 153 154
	int			(*delete)(struct Qdisc *, unsigned long);
	void			(*walk)(struct Qdisc *, struct qdisc_walker * arg);

	/* Filter manipulation */
J
John Fastabend 已提交
155
	struct tcf_proto __rcu ** (*tcf_chain)(struct Qdisc *, unsigned long);
156
	bool			(*tcf_cl_offload)(u32 classid);
L
Linus Torvalds 已提交
157 158 159 160 161 162 163 164 165 166 167
	unsigned long		(*bind_tcf)(struct Qdisc *, unsigned long,
					u32 classid);
	void			(*unbind_tcf)(struct Qdisc *, unsigned long);

	/* rtnetlink specific */
	int			(*dump)(struct Qdisc *, unsigned long,
					struct sk_buff *skb, struct tcmsg*);
	int			(*dump_stats)(struct Qdisc *, unsigned long,
					struct gnet_dump *);
};

E
Eric Dumazet 已提交
168
struct Qdisc_ops {
L
Linus Torvalds 已提交
169
	struct Qdisc_ops	*next;
170
	const struct Qdisc_class_ops	*cl_ops;
L
Linus Torvalds 已提交
171 172 173
	char			id[IFNAMSIZ];
	int			priv_size;

174 175 176
	int 			(*enqueue)(struct sk_buff *skb,
					   struct Qdisc *sch,
					   struct sk_buff **to_free);
L
Linus Torvalds 已提交
177
	struct sk_buff *	(*dequeue)(struct Qdisc *);
178
	struct sk_buff *	(*peek)(struct Qdisc *);
L
Linus Torvalds 已提交
179

180
	int			(*init)(struct Qdisc *, struct nlattr *arg);
L
Linus Torvalds 已提交
181 182
	void			(*reset)(struct Qdisc *);
	void			(*destroy)(struct Qdisc *);
183
	int			(*change)(struct Qdisc *, struct nlattr *arg);
184
	void			(*attach)(struct Qdisc *);
L
Linus Torvalds 已提交
185 186 187 188 189 190 191 192

	int			(*dump)(struct Qdisc *, struct sk_buff *);
	int			(*dump_stats)(struct Qdisc *, struct gnet_dump *);

	struct module		*owner;
};


E
Eric Dumazet 已提交
193
struct tcf_result {
L
Linus Torvalds 已提交
194 195 196 197
	unsigned long	class;
	u32		classid;
};

E
Eric Dumazet 已提交
198
struct tcf_proto_ops {
199
	struct list_head	head;
L
Linus Torvalds 已提交
200 201
	char			kind[IFNAMSIZ];

202 203 204
	int			(*classify)(struct sk_buff *,
					    const struct tcf_proto *,
					    struct tcf_result *);
L
Linus Torvalds 已提交
205
	int			(*init)(struct tcf_proto*);
206
	bool			(*destroy)(struct tcf_proto*, bool);
L
Linus Torvalds 已提交
207 208

	unsigned long		(*get)(struct tcf_proto*, u32 handle);
209
	int			(*change)(struct net *net, struct sk_buff *,
210
					struct tcf_proto*, unsigned long,
211
					u32 handle, struct nlattr **,
212
					unsigned long *, bool);
L
Linus Torvalds 已提交
213 214 215 216
	int			(*delete)(struct tcf_proto*, unsigned long);
	void			(*walk)(struct tcf_proto*, struct tcf_walker *arg);

	/* rtnetlink specific */
217
	int			(*dump)(struct net*, struct tcf_proto*, unsigned long,
L
Linus Torvalds 已提交
218 219 220 221 222
					struct sk_buff *skb, struct tcmsg*);

	struct module		*owner;
};

E
Eric Dumazet 已提交
223
struct tcf_proto {
L
Linus Torvalds 已提交
224
	/* Fast access part */
J
John Fastabend 已提交
225 226
	struct tcf_proto __rcu	*next;
	void __rcu		*root;
227 228 229
	int			(*classify)(struct sk_buff *,
					    const struct tcf_proto *,
					    struct tcf_result *);
A
Al Viro 已提交
230
	__be16			protocol;
L
Linus Torvalds 已提交
231 232 233 234 235 236

	/* All the rest */
	u32			prio;
	u32			classid;
	struct Qdisc		*q;
	void			*data;
237
	const struct tcf_proto_ops	*ops;
J
John Fastabend 已提交
238
	struct rcu_head		rcu;
L
Linus Torvalds 已提交
239 240
};

241 242
struct qdisc_skb_cb {
	unsigned int		pkt_len;
243
	u16			slave_dev_queue_mapping;
244
	u16			tc_classid;
245 246
#define QDISC_CB_PRIV_LEN 20
	unsigned char		data[QDISC_CB_PRIV_LEN];
247 248
};

249 250 251
static inline void qdisc_cb_private_validate(const struct sk_buff *skb, int sz)
{
	struct qdisc_skb_cb *qcb;
252 253

	BUILD_BUG_ON(sizeof(skb->cb) < offsetof(struct qdisc_skb_cb, data) + sz);
254 255 256
	BUILD_BUG_ON(sizeof(qcb->data) < sz);
}

257
static inline int qdisc_qlen(const struct Qdisc *q)
258 259 260 261
{
	return q->q.qlen;
}

262
static inline struct qdisc_skb_cb *qdisc_skb_cb(const struct sk_buff *skb)
263 264 265 266
{
	return (struct qdisc_skb_cb *)skb->cb;
}

267 268 269 270 271
static inline spinlock_t *qdisc_lock(struct Qdisc *qdisc)
{
	return &qdisc->q.lock;
}

272
static inline struct Qdisc *qdisc_root(const struct Qdisc *qdisc)
273
{
274 275 276
	struct Qdisc *q = rcu_dereference_rtnl(qdisc->dev_queue->qdisc);

	return q;
277 278
}

279
static inline struct Qdisc *qdisc_root_sleeping(const struct Qdisc *qdisc)
280 281 282 283
{
	return qdisc->dev_queue->qdisc_sleeping;
}

284 285 286 287 288 289 290 291 292 293 294
/* The qdisc root lock is a mechanism by which to top level
 * of a qdisc tree can be locked from any qdisc node in the
 * forest.  This allows changing the configuration of some
 * aspect of the qdisc tree while blocking out asynchronous
 * qdisc access in the packet processing paths.
 *
 * It is only legal to do this when the root will not change
 * on us.  Otherwise we'll potentially lock the wrong qdisc
 * root.  This is enforced by holding the RTNL semaphore, which
 * all users of this lock accessor must do.
 */
295
static inline spinlock_t *qdisc_root_lock(const struct Qdisc *qdisc)
296 297 298
{
	struct Qdisc *root = qdisc_root(qdisc);

299
	ASSERT_RTNL();
300
	return qdisc_lock(root);
301 302
}

303
static inline spinlock_t *qdisc_root_sleeping_lock(const struct Qdisc *qdisc)
304 305 306 307 308 309 310
{
	struct Qdisc *root = qdisc_root_sleeping(qdisc);

	ASSERT_RTNL();
	return qdisc_lock(root);
}

311 312 313 314 315 316 317 318
static inline seqcount_t *qdisc_root_sleeping_running(const struct Qdisc *qdisc)
{
	struct Qdisc *root = qdisc_root_sleeping(qdisc);

	ASSERT_RTNL();
	return &root->running;
}

319
static inline struct net_device *qdisc_dev(const struct Qdisc *qdisc)
320 321 322
{
	return qdisc->dev_queue->dev;
}
L
Linus Torvalds 已提交
323

324
static inline void sch_tree_lock(const struct Qdisc *q)
325
{
J
Jarek Poplawski 已提交
326
	spin_lock_bh(qdisc_root_sleeping_lock(q));
327 328
}

329
static inline void sch_tree_unlock(const struct Qdisc *q)
330
{
J
Jarek Poplawski 已提交
331
	spin_unlock_bh(qdisc_root_sleeping_lock(q));
332 333 334 335
}

#define tcf_tree_lock(tp)	sch_tree_lock((tp)->q)
#define tcf_tree_unlock(tp)	sch_tree_unlock((tp)->q)
L
Linus Torvalds 已提交
336

337 338
extern struct Qdisc noop_qdisc;
extern struct Qdisc_ops noop_qdisc_ops;
339 340
extern struct Qdisc_ops pfifo_fast_ops;
extern struct Qdisc_ops mq_qdisc_ops;
P
Phil Sutter 已提交
341
extern struct Qdisc_ops noqueue_qdisc_ops;
342
extern const struct Qdisc_ops *default_qdisc_ops;
343 344 345 346 347 348
static inline const struct Qdisc_ops *
get_default_qdisc_ops(const struct net_device *dev, int ntx)
{
	return ntx < dev->real_num_tx_queues ?
			default_qdisc_ops : &pfifo_fast_ops;
}
349

E
Eric Dumazet 已提交
350
struct Qdisc_class_common {
351 352 353 354
	u32			classid;
	struct hlist_node	hnode;
};

E
Eric Dumazet 已提交
355
struct Qdisc_class_hash {
356 357 358 359 360 361 362 363 364 365 366 367 368 369
	struct hlist_head	*hash;
	unsigned int		hashsize;
	unsigned int		hashmask;
	unsigned int		hashelems;
};

static inline unsigned int qdisc_class_hash(u32 id, u32 mask)
{
	id ^= id >> 8;
	id ^= id >> 4;
	return id & mask;
}

static inline struct Qdisc_class_common *
370
qdisc_class_find(const struct Qdisc_class_hash *hash, u32 id)
371 372 373 374 375
{
	struct Qdisc_class_common *cl;
	unsigned int h;

	h = qdisc_class_hash(id, hash->hashmask);
376
	hlist_for_each_entry(cl, &hash->hash[h], hnode) {
377 378 379 380 381 382
		if (cl->classid == id)
			return cl;
	}
	return NULL;
}

383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399
int qdisc_class_hash_init(struct Qdisc_class_hash *);
void qdisc_class_hash_insert(struct Qdisc_class_hash *,
			     struct Qdisc_class_common *);
void qdisc_class_hash_remove(struct Qdisc_class_hash *,
			     struct Qdisc_class_common *);
void qdisc_class_hash_grow(struct Qdisc *, struct Qdisc_class_hash *);
void qdisc_class_hash_destroy(struct Qdisc_class_hash *);

void dev_init_scheduler(struct net_device *dev);
void dev_shutdown(struct net_device *dev);
void dev_activate(struct net_device *dev);
void dev_deactivate(struct net_device *dev);
void dev_deactivate_many(struct list_head *head);
struct Qdisc *dev_graft_qdisc(struct netdev_queue *dev_queue,
			      struct Qdisc *qdisc);
void qdisc_reset(struct Qdisc *qdisc);
void qdisc_destroy(struct Qdisc *qdisc);
400 401
void qdisc_tree_reduce_backlog(struct Qdisc *qdisc, unsigned int n,
			       unsigned int len);
402
struct Qdisc *qdisc_alloc(struct netdev_queue *dev_queue,
403
			  const struct Qdisc_ops *ops);
404
struct Qdisc *qdisc_create_dflt(struct netdev_queue *dev_queue,
405
				const struct Qdisc_ops *ops, u32 parentid);
406 407
void __qdisc_calculate_pkt_len(struct sk_buff *skb,
			       const struct qdisc_size_table *stab);
408
bool tcf_destroy(struct tcf_proto *tp, bool force);
J
John Fastabend 已提交
409
void tcf_destroy_chain(struct tcf_proto __rcu **fl);
410
int skb_do_redirect(struct sk_buff *);
L
Linus Torvalds 已提交
411

412 413 414 415 416 417 418 419 420
static inline bool skb_at_tc_ingress(const struct sk_buff *skb)
{
#ifdef CONFIG_NET_CLS_ACT
	return G_TC_AT(skb->tc_verd) & AT_INGRESS;
#else
	return false;
#endif
}

421 422
/* Reset all TX qdiscs greater then index of a device.  */
static inline void qdisc_reset_all_tx_gt(struct net_device *dev, unsigned int i)
423
{
424 425
	struct Qdisc *qdisc;

426
	for (; i < dev->num_tx_queues; i++) {
427
		qdisc = rtnl_dereference(netdev_get_tx_queue(dev, i)->qdisc);
428 429 430 431 432 433
		if (qdisc) {
			spin_lock_bh(qdisc_lock(qdisc));
			qdisc_reset(qdisc);
			spin_unlock_bh(qdisc_lock(qdisc));
		}
	}
434 435 436 437
}

static inline void qdisc_reset_all_tx(struct net_device *dev)
{
438
	qdisc_reset_all_tx_gt(dev, 0);
439 440
}

441 442 443
/* Are all TX queues of the device empty?  */
static inline bool qdisc_all_tx_empty(const struct net_device *dev)
{
444
	unsigned int i;
445 446

	rcu_read_lock();
447 448
	for (i = 0; i < dev->num_tx_queues; i++) {
		struct netdev_queue *txq = netdev_get_tx_queue(dev, i);
449
		const struct Qdisc *q = rcu_dereference(txq->qdisc);
450

451 452
		if (q->q.qlen) {
			rcu_read_unlock();
453
			return false;
454
		}
455
	}
456
	rcu_read_unlock();
457
	return true;
458 459
}

460
/* Are any of the TX qdiscs changing?  */
461
static inline bool qdisc_tx_changing(const struct net_device *dev)
462
{
463
	unsigned int i;
464

465 466
	for (i = 0; i < dev->num_tx_queues; i++) {
		struct netdev_queue *txq = netdev_get_tx_queue(dev, i);
467
		if (rcu_access_pointer(txq->qdisc) != txq->qdisc_sleeping)
468 469 470
			return true;
	}
	return false;
471 472
}

473
/* Is the device using the noop qdisc on all queues?  */
474 475
static inline bool qdisc_tx_is_noop(const struct net_device *dev)
{
476
	unsigned int i;
477

478 479
	for (i = 0; i < dev->num_tx_queues; i++) {
		struct netdev_queue *txq = netdev_get_tx_queue(dev, i);
480
		if (rcu_access_pointer(txq->qdisc) != &noop_qdisc)
481 482 483
			return false;
	}
	return true;
484 485
}

486
static inline unsigned int qdisc_pkt_len(const struct sk_buff *skb)
487
{
488
	return qdisc_skb_cb(skb)->pkt_len;
489 490
}

491
/* additional qdisc xmit flags (NET_XMIT_MASK in linux/netdevice.h) */
492 493
enum net_xmit_qdisc_t {
	__NET_XMIT_STOLEN = 0x00010000,
494
	__NET_XMIT_BYPASS = 0x00020000,
495 496
};

497
#ifdef CONFIG_NET_CLS_ACT
498 499 500 501 502
#define net_xmit_drop_count(e)	((e) & __NET_XMIT_STOLEN ? 0 : 1)
#else
#define net_xmit_drop_count(e)	(1)
#endif

E
Eric Dumazet 已提交
503 504
static inline void qdisc_calculate_pkt_len(struct sk_buff *skb,
					   const struct Qdisc *sch)
505
{
506
#ifdef CONFIG_NET_SCHED
E
Eric Dumazet 已提交
507 508 509 510
	struct qdisc_size_table *stab = rcu_dereference_bh(sch->stab);

	if (stab)
		__qdisc_calculate_pkt_len(skb, stab);
511
#endif
E
Eric Dumazet 已提交
512 513
}

514 515
static inline int qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch,
				struct sk_buff **to_free)
E
Eric Dumazet 已提交
516 517
{
	qdisc_calculate_pkt_len(skb, sch);
518
	return sch->enqueue(skb, sch, to_free);
519 520
}

521 522 523 524
static inline bool qdisc_is_percpu_stats(const struct Qdisc *q)
{
	return q->flags & TCQ_F_CPUSTATS;
}
525

526 527 528 529 530 531 532
static inline void _bstats_update(struct gnet_stats_basic_packed *bstats,
				  __u64 bytes, __u32 packets)
{
	bstats->bytes += bytes;
	bstats->packets += packets;
}

533 534 535
static inline void bstats_update(struct gnet_stats_basic_packed *bstats,
				 const struct sk_buff *skb)
{
536 537 538 539 540 541 542 543 544 545 546
	_bstats_update(bstats,
		       qdisc_pkt_len(skb),
		       skb_is_gso(skb) ? skb_shinfo(skb)->gso_segs : 1);
}

static inline void _bstats_cpu_update(struct gnet_stats_basic_cpu *bstats,
				      __u64 bytes, __u32 packets)
{
	u64_stats_update_begin(&bstats->syncp);
	_bstats_update(&bstats->bstats, bytes, packets);
	u64_stats_update_end(&bstats->syncp);
547 548
}

549 550
static inline void bstats_cpu_update(struct gnet_stats_basic_cpu *bstats,
				     const struct sk_buff *skb)
551 552 553 554 555 556
{
	u64_stats_update_begin(&bstats->syncp);
	bstats_update(&bstats->bstats, skb);
	u64_stats_update_end(&bstats->syncp);
}

557 558 559 560 561 562
static inline void qdisc_bstats_cpu_update(struct Qdisc *sch,
					   const struct sk_buff *skb)
{
	bstats_cpu_update(this_cpu_ptr(sch->cpu_bstats), skb);
}

563 564
static inline void qdisc_bstats_update(struct Qdisc *sch,
				       const struct sk_buff *skb)
565
{
566
	bstats_update(&sch->bstats, skb);
567 568
}

569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585
static inline void qdisc_qstats_backlog_dec(struct Qdisc *sch,
					    const struct sk_buff *skb)
{
	sch->qstats.backlog -= qdisc_pkt_len(skb);
}

static inline void qdisc_qstats_backlog_inc(struct Qdisc *sch,
					    const struct sk_buff *skb)
{
	sch->qstats.backlog += qdisc_pkt_len(skb);
}

static inline void __qdisc_qstats_drop(struct Qdisc *sch, int count)
{
	sch->qstats.drops += count;
}

586
static inline void qstats_drop_inc(struct gnet_stats_queue *qstats)
587
{
588
	qstats->drops++;
589 590
}

591
static inline void qstats_overlimit_inc(struct gnet_stats_queue *qstats)
592
{
593 594
	qstats->overlimits++;
}
595

596 597 598 599 600 601 602
static inline void qdisc_qstats_drop(struct Qdisc *sch)
{
	qstats_drop_inc(&sch->qstats);
}

static inline void qdisc_qstats_cpu_drop(struct Qdisc *sch)
{
603
	this_cpu_inc(sch->cpu_qstats->drops);
604 605
}

606 607 608 609 610
static inline void qdisc_qstats_overlimit(struct Qdisc *sch)
{
	sch->qstats.overlimits++;
}

611 612 613 614 615 616 617
static inline void qdisc_skb_head_init(struct qdisc_skb_head *qh)
{
	qh->head = NULL;
	qh->tail = NULL;
	qh->qlen = 0;
}

618
static inline int __qdisc_enqueue_tail(struct sk_buff *skb, struct Qdisc *sch,
619
				       struct qdisc_skb_head *qh)
620
{
621 622 623 624 625 626 627 628 629 630 631
	struct sk_buff *last = qh->tail;

	if (last) {
		skb->next = NULL;
		last->next = skb;
		qh->tail = skb;
	} else {
		qh->tail = skb;
		qh->head = skb;
	}
	qh->qlen++;
632
	qdisc_qstats_backlog_inc(sch, skb);
633 634 635 636 637 638 639 640 641

	return NET_XMIT_SUCCESS;
}

static inline int qdisc_enqueue_tail(struct sk_buff *skb, struct Qdisc *sch)
{
	return __qdisc_enqueue_tail(skb, sch, &sch->q);
}

642
static inline struct sk_buff *__qdisc_dequeue_head(struct qdisc_skb_head *qh)
643
{
644 645 646 647 648 649 650 651 652
	struct sk_buff *skb = qh->head;

	if (likely(skb != NULL)) {
		qh->head = skb->next;
		qh->qlen--;
		if (qh->head == NULL)
			qh->tail = NULL;
		skb->next = NULL;
	}
653

654 655 656 657 658 659 660
	return skb;
}

static inline struct sk_buff *qdisc_dequeue_head(struct Qdisc *sch)
{
	struct sk_buff *skb = __qdisc_dequeue_head(&sch->q);

661
	if (likely(skb != NULL)) {
662
		qdisc_qstats_backlog_dec(sch, skb);
663 664
		qdisc_bstats_update(sch, skb);
	}
665 666 667 668

	return skb;
}

669 670 671 672 673 674 675 676 677
/* Instead of calling kfree_skb() while root qdisc lock is held,
 * queue the skb for future freeing at end of __dev_xmit_skb()
 */
static inline void __qdisc_drop(struct sk_buff *skb, struct sk_buff **to_free)
{
	skb->next = *to_free;
	*to_free = skb;
}

678
static inline unsigned int __qdisc_queue_drop_head(struct Qdisc *sch,
679
						   struct qdisc_skb_head *qh,
680
						   struct sk_buff **to_free)
681
{
682
	struct sk_buff *skb = __qdisc_dequeue_head(qh);
683 684 685

	if (likely(skb != NULL)) {
		unsigned int len = qdisc_pkt_len(skb);
686

687
		qdisc_qstats_backlog_dec(sch, skb);
688
		__qdisc_drop(skb, to_free);
689 690 691 692 693 694
		return len;
	}

	return 0;
}

695 696
static inline unsigned int qdisc_queue_drop_head(struct Qdisc *sch,
						 struct sk_buff **to_free)
697
{
698
	return __qdisc_queue_drop_head(sch, &sch->q, to_free);
699 700
}

701 702
static inline struct sk_buff *qdisc_peek_head(struct Qdisc *sch)
{
703 704 705
	const struct qdisc_skb_head *qh = &sch->q;

	return qh->head;
706 707
}

708 709 710 711
/* generic pseudo peek method for non-work-conserving qdisc */
static inline struct sk_buff *qdisc_peek_dequeued(struct Qdisc *sch)
{
	/* we can reuse ->gso_skb because peek isn't called for root qdiscs */
712
	if (!sch->gso_skb) {
713
		sch->gso_skb = sch->dequeue(sch);
714
		if (sch->gso_skb) {
715
			/* it's still part of the queue */
716
			qdisc_qstats_backlog_inc(sch, sch->gso_skb);
717
			sch->q.qlen++;
718
		}
719
	}
720 721 722 723 724 725 726 727 728

	return sch->gso_skb;
}

/* use instead of qdisc->dequeue() for all qdiscs queried with ->peek() */
static inline struct sk_buff *qdisc_dequeue_peeked(struct Qdisc *sch)
{
	struct sk_buff *skb = sch->gso_skb;

729
	if (skb) {
730
		sch->gso_skb = NULL;
731
		qdisc_qstats_backlog_dec(sch, skb);
732 733
		sch->q.qlen--;
	} else {
734
		skb = sch->dequeue(sch);
735
	}
736 737 738 739

	return skb;
}

740
static inline void __qdisc_reset_queue(struct qdisc_skb_head *qh)
741 742 743 744 745
{
	/*
	 * We do not know the backlog in bytes of this list, it
	 * is up to the caller to correct it
	 */
746 747 748 749 750 751 752
	ASSERT_RTNL();
	if (qh->qlen) {
		rtnl_kfree_skbs(qh->head, qh->tail);

		qh->head = NULL;
		qh->tail = NULL;
		qh->qlen = 0;
753
	}
754 755 756 757
}

static inline void qdisc_reset_queue(struct Qdisc *sch)
{
758
	__qdisc_reset_queue(&sch->q);
759 760 761
	sch->qstats.backlog = 0;
}

762 763 764 765 766 767 768 769 770
static inline struct Qdisc *qdisc_replace(struct Qdisc *sch, struct Qdisc *new,
					  struct Qdisc **pold)
{
	struct Qdisc *old;

	sch_tree_lock(sch);
	old = *pold;
	*pold = new;
	if (old != NULL) {
771
		qdisc_tree_reduce_backlog(old, old->q.qlen, old->qstats.backlog);
772 773 774 775 776 777 778
		qdisc_reset(old);
	}
	sch_tree_unlock(sch);

	return old;
}

779 780 781 782 783 784
static inline void rtnl_qdisc_drop(struct sk_buff *skb, struct Qdisc *sch)
{
	rtnl_kfree_skbs(skb, skb);
	qdisc_qstats_drop(sch);
}

785 786 787

static inline int qdisc_drop(struct sk_buff *skb, struct Qdisc *sch,
			     struct sk_buff **to_free)
788
{
789
	__qdisc_drop(skb, to_free);
790
	qdisc_qstats_drop(sch);
791 792 793 794

	return NET_XMIT_DROP;
}

795 796 797 798 799
/* Length to Time (L2T) lookup in a qdisc_rate_table, to determine how
   long it will take to send a packet given its size.
 */
static inline u32 qdisc_l2t(struct qdisc_rate_table* rtab, unsigned int pktlen)
{
800 801 802
	int slot = pktlen + rtab->rate.cell_align + rtab->rate.overhead;
	if (slot < 0)
		slot = 0;
803 804
	slot >>= rtab->rate.cell_log;
	if (slot > 255)
E
Eric Dumazet 已提交
805
		return rtab->data[255]*(slot >> 8) + rtab->data[slot & 0xFF];
806 807 808
	return rtab->data[slot];
}

809
struct psched_ratecfg {
810
	u64	rate_bytes_ps; /* bytes per second */
811 812
	u32	mult;
	u16	overhead;
813
	u8	linklayer;
814
	u8	shift;
815 816 817 818 819
};

static inline u64 psched_l2t_ns(const struct psched_ratecfg *r,
				unsigned int len)
{
820 821 822 823 824 825
	len += r->overhead;

	if (unlikely(r->linklayer == TC_LINKLAYER_ATM))
		return ((u64)(DIV_ROUND_UP(len,48)*53) * r->mult) >> r->shift;

	return ((u64)len * r->mult) >> r->shift;
826 827
}

828
void psched_ratecfg_precompute(struct psched_ratecfg *r,
829 830
			       const struct tc_ratespec *conf,
			       u64 rate64);
831

832 833
static inline void psched_ratecfg_getrate(struct tc_ratespec *res,
					  const struct psched_ratecfg *r)
834
{
835
	memset(res, 0, sizeof(*res));
836 837 838 839 840 841 842

	/* legacy struct tc_ratespec has a 32bit @rate field
	 * Qdisc using 64bit rate should add new attributes
	 * in order to maintain compatibility.
	 */
	res->rate = min_t(u64, r->rate_bytes_ps, ~0U);

843
	res->overhead = r->overhead;
844
	res->linklayer = (r->linklayer & TC_LINKLAYER_MASK);
845 846
}

L
Linus Torvalds 已提交
847
#endif