sch_generic.h 21.0 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
#define TCQ_F_INVISIBLE		0x80 /* invisible by default in dump */
70
	u32			limit;
71
	const struct Qdisc_ops	*ops;
E
Eric Dumazet 已提交
72
	struct qdisc_size_table	__rcu *stab;
73
	struct hlist_node       hash;
L
Linus Torvalds 已提交
74 75
	u32			handle;
	u32			parent;
76 77
	void			*u32_node;

78 79
	struct netdev_queue	*dev_queue;

80
	struct net_rate_estimator __rcu *rate_est;
81 82 83
	struct gnet_stats_basic_cpu __percpu *cpu_bstats;
	struct gnet_stats_queue	__percpu *cpu_qstats;

84 85 86
	/*
	 * For performance sake on SMP, we put highly modified fields at the end
	 */
87
	struct sk_buff		*gso_skb ____cacheline_aligned_in_smp;
88
	struct qdisc_skb_head	q;
89
	struct gnet_stats_basic_packed bstats;
90
	seqcount_t		running;
91
	struct gnet_stats_queue	qstats;
92 93 94
	unsigned long		state;
	struct Qdisc            *next_sched;
	struct sk_buff		*skb_bad_txq;
95
	struct rcu_head		rcu_head;
96 97 98 99
	int			padded;
	atomic_t		refcnt;

	spinlock_t		busylock ____cacheline_aligned_in_smp;
L
Linus Torvalds 已提交
100 101
};

102
static inline bool qdisc_is_running(const struct Qdisc *qdisc)
103
{
104
	return (raw_read_seqcount(&qdisc->running) & 1) ? true : false;
105 106 107 108
}

static inline bool qdisc_run_begin(struct Qdisc *qdisc)
{
109 110
	if (qdisc_is_running(qdisc))
		return false;
111 112 113 114 115
	/* 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_);
116
	return true;
117 118 119 120
}

static inline void qdisc_run_end(struct Qdisc *qdisc)
{
121
	write_seqcount_end(&qdisc->running);
122 123
}

124 125 126 127 128 129 130 131 132 133 134 135 136 137 138
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
139
struct Qdisc_class_ops {
L
Linus Torvalds 已提交
140
	/* Child qdisc manipulation */
141
	struct netdev_queue *	(*select_queue)(struct Qdisc *, struct tcmsg *);
L
Linus Torvalds 已提交
142 143 144
	int			(*graft)(struct Qdisc *, unsigned long cl,
					struct Qdisc *, struct Qdisc **);
	struct Qdisc *		(*leaf)(struct Qdisc *, unsigned long cl);
145
	void			(*qlen_notify)(struct Qdisc *, unsigned long);
L
Linus Torvalds 已提交
146 147 148 149 150

	/* Class manipulation routines */
	unsigned long		(*get)(struct Qdisc *, u32 classid);
	void			(*put)(struct Qdisc *, unsigned long);
	int			(*change)(struct Qdisc *, u32, u32,
151
					struct nlattr **, unsigned long *);
L
Linus Torvalds 已提交
152 153 154 155
	int			(*delete)(struct Qdisc *, unsigned long);
	void			(*walk)(struct Qdisc *, struct qdisc_walker * arg);

	/* Filter manipulation */
156
	struct tcf_block *	(*tcf_block)(struct Qdisc *, unsigned long);
157
	bool			(*tcf_cl_offload)(u32 classid);
L
Linus Torvalds 已提交
158 159 160 161 162 163 164 165 166 167 168
	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
169
struct Qdisc_ops {
L
Linus Torvalds 已提交
170
	struct Qdisc_ops	*next;
171
	const struct Qdisc_class_ops	*cl_ops;
L
Linus Torvalds 已提交
172 173 174
	char			id[IFNAMSIZ];
	int			priv_size;

175 176 177
	int 			(*enqueue)(struct sk_buff *skb,
					   struct Qdisc *sch,
					   struct sk_buff **to_free);
L
Linus Torvalds 已提交
178
	struct sk_buff *	(*dequeue)(struct Qdisc *);
179
	struct sk_buff *	(*peek)(struct Qdisc *);
L
Linus Torvalds 已提交
180

181
	int			(*init)(struct Qdisc *, struct nlattr *arg);
L
Linus Torvalds 已提交
182 183
	void			(*reset)(struct Qdisc *);
	void			(*destroy)(struct Qdisc *);
184
	int			(*change)(struct Qdisc *, struct nlattr *arg);
185
	void			(*attach)(struct Qdisc *);
L
Linus Torvalds 已提交
186 187 188 189 190 191 192 193

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

	struct module		*owner;
};


E
Eric Dumazet 已提交
194
struct tcf_result {
L
Linus Torvalds 已提交
195 196 197 198
	unsigned long	class;
	u32		classid;
};

E
Eric Dumazet 已提交
199
struct tcf_proto_ops {
200
	struct list_head	head;
L
Linus Torvalds 已提交
201 202
	char			kind[IFNAMSIZ];

203 204 205
	int			(*classify)(struct sk_buff *,
					    const struct tcf_proto *,
					    struct tcf_result *);
L
Linus Torvalds 已提交
206
	int			(*init)(struct tcf_proto*);
207
	void			(*destroy)(struct tcf_proto*);
L
Linus Torvalds 已提交
208 209

	unsigned long		(*get)(struct tcf_proto*, u32 handle);
210
	int			(*change)(struct net *net, struct sk_buff *,
211
					struct tcf_proto*, unsigned long,
212
					u32 handle, struct nlattr **,
213
					unsigned long *, bool);
214
	int			(*delete)(struct tcf_proto*, unsigned long, bool*);
L
Linus Torvalds 已提交
215 216 217
	void			(*walk)(struct tcf_proto*, struct tcf_walker *arg);

	/* rtnetlink specific */
218
	int			(*dump)(struct net*, struct tcf_proto*, unsigned long,
L
Linus Torvalds 已提交
219 220 221 222 223
					struct sk_buff *skb, struct tcmsg*);

	struct module		*owner;
};

E
Eric Dumazet 已提交
224
struct tcf_proto {
L
Linus Torvalds 已提交
225
	/* Fast access part */
J
John Fastabend 已提交
226 227
	struct tcf_proto __rcu	*next;
	void __rcu		*root;
228 229 230
	int			(*classify)(struct sk_buff *,
					    const struct tcf_proto *,
					    struct tcf_result *);
A
Al Viro 已提交
231
	__be16			protocol;
L
Linus Torvalds 已提交
232 233 234 235 236 237

	/* All the rest */
	u32			prio;
	u32			classid;
	struct Qdisc		*q;
	void			*data;
238
	const struct tcf_proto_ops	*ops;
239
	struct tcf_block	*block;
J
John Fastabend 已提交
240
	struct rcu_head		rcu;
L
Linus Torvalds 已提交
241 242
};

243 244
struct qdisc_skb_cb {
	unsigned int		pkt_len;
245
	u16			slave_dev_queue_mapping;
246
	u16			tc_classid;
247 248
#define QDISC_CB_PRIV_LEN 20
	unsigned char		data[QDISC_CB_PRIV_LEN];
249 250
};

251 252
struct tcf_chain {
	struct tcf_proto __rcu *filter_chain;
253 254 255
	struct tcf_proto __rcu **p_filter_chain;
};

256 257 258 259
struct tcf_block {
	struct tcf_chain *chain;
};

260 261 262
static inline void qdisc_cb_private_validate(const struct sk_buff *skb, int sz)
{
	struct qdisc_skb_cb *qcb;
263 264

	BUILD_BUG_ON(sizeof(skb->cb) < offsetof(struct qdisc_skb_cb, data) + sz);
265 266 267
	BUILD_BUG_ON(sizeof(qcb->data) < sz);
}

268
static inline int qdisc_qlen(const struct Qdisc *q)
269 270 271 272
{
	return q->q.qlen;
}

273
static inline struct qdisc_skb_cb *qdisc_skb_cb(const struct sk_buff *skb)
274 275 276 277
{
	return (struct qdisc_skb_cb *)skb->cb;
}

278 279 280 281 282
static inline spinlock_t *qdisc_lock(struct Qdisc *qdisc)
{
	return &qdisc->q.lock;
}

283
static inline struct Qdisc *qdisc_root(const struct Qdisc *qdisc)
284
{
285 286 287
	struct Qdisc *q = rcu_dereference_rtnl(qdisc->dev_queue->qdisc);

	return q;
288 289
}

290
static inline struct Qdisc *qdisc_root_sleeping(const struct Qdisc *qdisc)
291 292 293 294
{
	return qdisc->dev_queue->qdisc_sleeping;
}

295 296 297 298 299 300 301 302 303 304 305
/* 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.
 */
306
static inline spinlock_t *qdisc_root_lock(const struct Qdisc *qdisc)
307 308 309
{
	struct Qdisc *root = qdisc_root(qdisc);

310
	ASSERT_RTNL();
311
	return qdisc_lock(root);
312 313
}

314
static inline spinlock_t *qdisc_root_sleeping_lock(const struct Qdisc *qdisc)
315 316 317 318 319 320 321
{
	struct Qdisc *root = qdisc_root_sleeping(qdisc);

	ASSERT_RTNL();
	return qdisc_lock(root);
}

322 323 324 325 326 327 328 329
static inline seqcount_t *qdisc_root_sleeping_running(const struct Qdisc *qdisc)
{
	struct Qdisc *root = qdisc_root_sleeping(qdisc);

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

330
static inline struct net_device *qdisc_dev(const struct Qdisc *qdisc)
331 332 333
{
	return qdisc->dev_queue->dev;
}
L
Linus Torvalds 已提交
334

335
static inline void sch_tree_lock(const struct Qdisc *q)
336
{
J
Jarek Poplawski 已提交
337
	spin_lock_bh(qdisc_root_sleeping_lock(q));
338 339
}

340
static inline void sch_tree_unlock(const struct Qdisc *q)
341
{
J
Jarek Poplawski 已提交
342
	spin_unlock_bh(qdisc_root_sleeping_lock(q));
343 344 345 346
}

#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
347

348 349
extern struct Qdisc noop_qdisc;
extern struct Qdisc_ops noop_qdisc_ops;
350 351
extern struct Qdisc_ops pfifo_fast_ops;
extern struct Qdisc_ops mq_qdisc_ops;
P
Phil Sutter 已提交
352
extern struct Qdisc_ops noqueue_qdisc_ops;
353
extern const struct Qdisc_ops *default_qdisc_ops;
354 355 356 357 358 359
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;
}
360

E
Eric Dumazet 已提交
361
struct Qdisc_class_common {
362 363 364 365
	u32			classid;
	struct hlist_node	hnode;
};

E
Eric Dumazet 已提交
366
struct Qdisc_class_hash {
367 368 369 370 371 372 373 374 375 376 377 378 379 380
	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 *
381
qdisc_class_find(const struct Qdisc_class_hash *hash, u32 id)
382 383 384 385 386
{
	struct Qdisc_class_common *cl;
	unsigned int h;

	h = qdisc_class_hash(id, hash->hashmask);
387
	hlist_for_each_entry(cl, &hash->hash[h], hnode) {
388 389 390 391 392 393
		if (cl->classid == id)
			return cl;
	}
	return NULL;
}

394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410
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);
411 412
void qdisc_tree_reduce_backlog(struct Qdisc *qdisc, unsigned int n,
			       unsigned int len);
413
struct Qdisc *qdisc_alloc(struct netdev_queue *dev_queue,
414
			  const struct Qdisc_ops *ops);
415
struct Qdisc *qdisc_create_dflt(struct netdev_queue *dev_queue,
416
				const struct Qdisc_ops *ops, u32 parentid);
417 418
void __qdisc_calculate_pkt_len(struct sk_buff *skb,
			       const struct qdisc_size_table *stab);
419
int skb_do_redirect(struct sk_buff *);
L
Linus Torvalds 已提交
420

421 422 423
static inline void skb_reset_tc(struct sk_buff *skb)
{
#ifdef CONFIG_NET_CLS_ACT
424
	skb->tc_redirected = 0;
425 426 427
#endif
}

428 429 430
static inline bool skb_at_tc_ingress(const struct sk_buff *skb)
{
#ifdef CONFIG_NET_CLS_ACT
431
	return skb->tc_at_ingress;
432 433 434 435 436
#else
	return false;
#endif
}

437 438 439 440 441 442 443 444 445 446 447
static inline bool skb_skip_tc_classify(struct sk_buff *skb)
{
#ifdef CONFIG_NET_CLS_ACT
	if (skb->tc_skip_classify) {
		skb->tc_skip_classify = 0;
		return true;
	}
#endif
	return false;
}

448 449
/* 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)
450
{
451 452
	struct Qdisc *qdisc;

453
	for (; i < dev->num_tx_queues; i++) {
454
		qdisc = rtnl_dereference(netdev_get_tx_queue(dev, i)->qdisc);
455 456 457 458 459 460
		if (qdisc) {
			spin_lock_bh(qdisc_lock(qdisc));
			qdisc_reset(qdisc);
			spin_unlock_bh(qdisc_lock(qdisc));
		}
	}
461 462 463 464
}

static inline void qdisc_reset_all_tx(struct net_device *dev)
{
465
	qdisc_reset_all_tx_gt(dev, 0);
466 467
}

468 469 470
/* Are all TX queues of the device empty?  */
static inline bool qdisc_all_tx_empty(const struct net_device *dev)
{
471
	unsigned int i;
472 473

	rcu_read_lock();
474 475
	for (i = 0; i < dev->num_tx_queues; i++) {
		struct netdev_queue *txq = netdev_get_tx_queue(dev, i);
476
		const struct Qdisc *q = rcu_dereference(txq->qdisc);
477

478 479
		if (q->q.qlen) {
			rcu_read_unlock();
480
			return false;
481
		}
482
	}
483
	rcu_read_unlock();
484
	return true;
485 486
}

487
/* Are any of the TX qdiscs changing?  */
488
static inline bool qdisc_tx_changing(const struct net_device *dev)
489
{
490
	unsigned int i;
491

492 493
	for (i = 0; i < dev->num_tx_queues; i++) {
		struct netdev_queue *txq = netdev_get_tx_queue(dev, i);
494
		if (rcu_access_pointer(txq->qdisc) != txq->qdisc_sleeping)
495 496 497
			return true;
	}
	return false;
498 499
}

500
/* Is the device using the noop qdisc on all queues?  */
501 502
static inline bool qdisc_tx_is_noop(const struct net_device *dev)
{
503
	unsigned int i;
504

505 506
	for (i = 0; i < dev->num_tx_queues; i++) {
		struct netdev_queue *txq = netdev_get_tx_queue(dev, i);
507
		if (rcu_access_pointer(txq->qdisc) != &noop_qdisc)
508 509 510
			return false;
	}
	return true;
511 512
}

513
static inline unsigned int qdisc_pkt_len(const struct sk_buff *skb)
514
{
515
	return qdisc_skb_cb(skb)->pkt_len;
516 517
}

518
/* additional qdisc xmit flags (NET_XMIT_MASK in linux/netdevice.h) */
519 520
enum net_xmit_qdisc_t {
	__NET_XMIT_STOLEN = 0x00010000,
521
	__NET_XMIT_BYPASS = 0x00020000,
522 523
};

524
#ifdef CONFIG_NET_CLS_ACT
525 526 527 528 529
#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
530 531
static inline void qdisc_calculate_pkt_len(struct sk_buff *skb,
					   const struct Qdisc *sch)
532
{
533
#ifdef CONFIG_NET_SCHED
E
Eric Dumazet 已提交
534 535 536 537
	struct qdisc_size_table *stab = rcu_dereference_bh(sch->stab);

	if (stab)
		__qdisc_calculate_pkt_len(skb, stab);
538
#endif
E
Eric Dumazet 已提交
539 540
}

541 542
static inline int qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch,
				struct sk_buff **to_free)
E
Eric Dumazet 已提交
543 544
{
	qdisc_calculate_pkt_len(skb, sch);
545
	return sch->enqueue(skb, sch, to_free);
546 547
}

548 549 550 551
static inline bool qdisc_is_percpu_stats(const struct Qdisc *q)
{
	return q->flags & TCQ_F_CPUSTATS;
}
552

553 554 555 556 557 558 559
static inline void _bstats_update(struct gnet_stats_basic_packed *bstats,
				  __u64 bytes, __u32 packets)
{
	bstats->bytes += bytes;
	bstats->packets += packets;
}

560 561 562
static inline void bstats_update(struct gnet_stats_basic_packed *bstats,
				 const struct sk_buff *skb)
{
563 564 565 566 567 568 569 570 571 572 573
	_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);
574 575
}

576 577
static inline void bstats_cpu_update(struct gnet_stats_basic_cpu *bstats,
				     const struct sk_buff *skb)
578 579 580 581 582 583
{
	u64_stats_update_begin(&bstats->syncp);
	bstats_update(&bstats->bstats, skb);
	u64_stats_update_end(&bstats->syncp);
}

584 585 586 587 588 589
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);
}

590 591
static inline void qdisc_bstats_update(struct Qdisc *sch,
				       const struct sk_buff *skb)
592
{
593
	bstats_update(&sch->bstats, skb);
594 595
}

596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612
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;
}

613
static inline void qstats_drop_inc(struct gnet_stats_queue *qstats)
614
{
615
	qstats->drops++;
616 617
}

618
static inline void qstats_overlimit_inc(struct gnet_stats_queue *qstats)
619
{
620 621
	qstats->overlimits++;
}
622

623 624 625 626 627 628 629
static inline void qdisc_qstats_drop(struct Qdisc *sch)
{
	qstats_drop_inc(&sch->qstats);
}

static inline void qdisc_qstats_cpu_drop(struct Qdisc *sch)
{
630
	this_cpu_inc(sch->cpu_qstats->drops);
631 632
}

633 634 635 636 637
static inline void qdisc_qstats_overlimit(struct Qdisc *sch)
{
	sch->qstats.overlimits++;
}

638 639 640 641 642 643 644
static inline void qdisc_skb_head_init(struct qdisc_skb_head *qh)
{
	qh->head = NULL;
	qh->tail = NULL;
	qh->qlen = 0;
}

645
static inline int __qdisc_enqueue_tail(struct sk_buff *skb, struct Qdisc *sch,
646
				       struct qdisc_skb_head *qh)
647
{
648 649 650 651 652 653 654 655 656 657 658
	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++;
659
	qdisc_qstats_backlog_inc(sch, skb);
660 661 662 663 664 665 666 667 668

	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);
}

669
static inline struct sk_buff *__qdisc_dequeue_head(struct qdisc_skb_head *qh)
670
{
671 672 673 674 675 676 677 678 679
	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;
	}
680

681 682 683 684 685 686 687
	return skb;
}

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

688
	if (likely(skb != NULL)) {
689
		qdisc_qstats_backlog_dec(sch, skb);
690 691
		qdisc_bstats_update(sch, skb);
	}
692 693 694 695

	return skb;
}

696 697 698 699 700 701 702 703 704
/* 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;
}

705
static inline unsigned int __qdisc_queue_drop_head(struct Qdisc *sch,
706
						   struct qdisc_skb_head *qh,
707
						   struct sk_buff **to_free)
708
{
709
	struct sk_buff *skb = __qdisc_dequeue_head(qh);
710 711 712

	if (likely(skb != NULL)) {
		unsigned int len = qdisc_pkt_len(skb);
713

714
		qdisc_qstats_backlog_dec(sch, skb);
715
		__qdisc_drop(skb, to_free);
716 717 718 719 720 721
		return len;
	}

	return 0;
}

722 723
static inline unsigned int qdisc_queue_drop_head(struct Qdisc *sch,
						 struct sk_buff **to_free)
724
{
725
	return __qdisc_queue_drop_head(sch, &sch->q, to_free);
726 727
}

728 729
static inline struct sk_buff *qdisc_peek_head(struct Qdisc *sch)
{
730 731 732
	const struct qdisc_skb_head *qh = &sch->q;

	return qh->head;
733 734
}

735 736 737 738
/* 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 */
739
	if (!sch->gso_skb) {
740
		sch->gso_skb = sch->dequeue(sch);
741
		if (sch->gso_skb) {
742
			/* it's still part of the queue */
743
			qdisc_qstats_backlog_inc(sch, sch->gso_skb);
744
			sch->q.qlen++;
745
		}
746
	}
747 748 749 750 751 752 753 754 755

	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;

756
	if (skb) {
757
		sch->gso_skb = NULL;
758
		qdisc_qstats_backlog_dec(sch, skb);
759 760
		sch->q.qlen--;
	} else {
761
		skb = sch->dequeue(sch);
762
	}
763 764 765 766

	return skb;
}

767
static inline void __qdisc_reset_queue(struct qdisc_skb_head *qh)
768 769 770 771 772
{
	/*
	 * We do not know the backlog in bytes of this list, it
	 * is up to the caller to correct it
	 */
773 774 775 776 777 778 779
	ASSERT_RTNL();
	if (qh->qlen) {
		rtnl_kfree_skbs(qh->head, qh->tail);

		qh->head = NULL;
		qh->tail = NULL;
		qh->qlen = 0;
780
	}
781 782 783 784
}

static inline void qdisc_reset_queue(struct Qdisc *sch)
{
785
	__qdisc_reset_queue(&sch->q);
786 787 788
	sch->qstats.backlog = 0;
}

789 790 791 792 793 794 795 796 797
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) {
798
		qdisc_tree_reduce_backlog(old, old->q.qlen, old->qstats.backlog);
799 800 801 802 803 804 805
		qdisc_reset(old);
	}
	sch_tree_unlock(sch);

	return old;
}

806 807 808 809 810 811
static inline void rtnl_qdisc_drop(struct sk_buff *skb, struct Qdisc *sch)
{
	rtnl_kfree_skbs(skb, skb);
	qdisc_qstats_drop(sch);
}

812 813 814

static inline int qdisc_drop(struct sk_buff *skb, struct Qdisc *sch,
			     struct sk_buff **to_free)
815
{
816
	__qdisc_drop(skb, to_free);
817
	qdisc_qstats_drop(sch);
818 819 820 821

	return NET_XMIT_DROP;
}

822 823 824 825 826
/* 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)
{
827 828 829
	int slot = pktlen + rtab->rate.cell_align + rtab->rate.overhead;
	if (slot < 0)
		slot = 0;
830 831
	slot >>= rtab->rate.cell_log;
	if (slot > 255)
E
Eric Dumazet 已提交
832
		return rtab->data[255]*(slot >> 8) + rtab->data[slot & 0xFF];
833 834 835
	return rtab->data[slot];
}

836
struct psched_ratecfg {
837
	u64	rate_bytes_ps; /* bytes per second */
838 839
	u32	mult;
	u16	overhead;
840
	u8	linklayer;
841
	u8	shift;
842 843 844 845 846
};

static inline u64 psched_l2t_ns(const struct psched_ratecfg *r,
				unsigned int len)
{
847 848 849 850 851 852
	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;
853 854
}

855
void psched_ratecfg_precompute(struct psched_ratecfg *r,
856 857
			       const struct tc_ratespec *conf,
			       u64 rate64);
858

859 860
static inline void psched_ratecfg_getrate(struct tc_ratespec *res,
					  const struct psched_ratecfg *r)
861
{
862
	memset(res, 0, sizeof(*res));
863 864 865 866 867 868 869

	/* 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);

870
	res->overhead = r->overhead;
871
	res->linklayer = (r->linklayer & TC_LINKLAYER_MASK);
872 873
}

L
Linus Torvalds 已提交
874
#endif