sch_generic.h 21.3 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>
11
#include <linux/list.h>
12
#include <linux/refcount.h>
L
Linus Torvalds 已提交
13
#include <net/gen_stats.h>
14
#include <net/rtnetlink.h>
L
Linus Torvalds 已提交
15 16 17 18 19 20

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

E
Eric Dumazet 已提交
21
struct qdisc_rate_table {
L
Linus Torvalds 已提交
22 23 24 25 26 27
	struct tc_ratespec rate;
	u32		data[256];
	struct qdisc_rate_table *next;
	int		refcnt;
};

E
Eric Dumazet 已提交
28
enum qdisc_state_t {
29
	__QDISC_STATE_SCHED,
30
	__QDISC_STATE_DEACTIVATED,
31 32
};

33
struct qdisc_size_table {
E
Eric Dumazet 已提交
34
	struct rcu_head		rcu;
35 36 37 38 39 40
	struct list_head	list;
	struct tc_sizespec	szopts;
	int			refcnt;
	u16			data[];
};

41 42 43 44 45 46 47 48
/* 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
49
struct Qdisc {
50 51 52 53
	int 			(*enqueue)(struct sk_buff *skb,
					   struct Qdisc *sch,
					   struct sk_buff **to_free);
	struct sk_buff *	(*dequeue)(struct Qdisc *sch);
54
	unsigned int		flags;
55
#define TCQ_F_BUILTIN		1
56 57 58
#define TCQ_F_INGRESS		2
#define TCQ_F_CAN_BYPASS	4
#define TCQ_F_MQROOT		8
59 60 61 62 63 64 65
#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.
				      */
66
#define TCQ_F_WARN_NONWC	(1 << 16)
67
#define TCQ_F_CPUSTATS		0x20 /* run using percpu statistics */
68 69 70
#define TCQ_F_NOPARENT		0x40 /* root of its hierarchy :
				      * qdisc_tree_decrease_qlen() should stop.
				      */
71
#define TCQ_F_INVISIBLE		0x80 /* invisible by default in dump */
72
	u32			limit;
73
	const struct Qdisc_ops	*ops;
E
Eric Dumazet 已提交
74
	struct qdisc_size_table	__rcu *stab;
75
	struct hlist_node       hash;
L
Linus Torvalds 已提交
76 77
	u32			handle;
	u32			parent;
78 79
	void			*u32_node;

80 81
	struct netdev_queue	*dev_queue;

82
	struct net_rate_estimator __rcu *rate_est;
83 84 85
	struct gnet_stats_basic_cpu __percpu *cpu_bstats;
	struct gnet_stats_queue	__percpu *cpu_qstats;

86 87 88
	/*
	 * For performance sake on SMP, we put highly modified fields at the end
	 */
89
	struct sk_buff		*gso_skb ____cacheline_aligned_in_smp;
90
	struct qdisc_skb_head	q;
91
	struct gnet_stats_basic_packed bstats;
92
	seqcount_t		running;
93
	struct gnet_stats_queue	qstats;
94 95 96
	unsigned long		state;
	struct Qdisc            *next_sched;
	struct sk_buff		*skb_bad_txq;
97
	struct rcu_head		rcu_head;
98
	int			padded;
99
	refcount_t		refcnt;
100 101

	spinlock_t		busylock ____cacheline_aligned_in_smp;
L
Linus Torvalds 已提交
102 103
};

104
static inline bool qdisc_is_running(const struct Qdisc *qdisc)
105
{
106
	return (raw_read_seqcount(&qdisc->running) & 1) ? true : false;
107 108 109 110
}

static inline bool qdisc_run_begin(struct Qdisc *qdisc)
{
111 112
	if (qdisc_is_running(qdisc))
		return false;
113 114 115 116 117
	/* 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_);
118
	return true;
119 120 121 122
}

static inline void qdisc_run_end(struct Qdisc *qdisc)
{
123
	write_seqcount_end(&qdisc->running);
124 125
}

126 127 128 129 130 131 132 133 134 135 136 137 138 139 140
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
141
struct Qdisc_class_ops {
L
Linus Torvalds 已提交
142
	/* Child qdisc manipulation */
143
	struct netdev_queue *	(*select_queue)(struct Qdisc *, struct tcmsg *);
L
Linus Torvalds 已提交
144 145 146
	int			(*graft)(struct Qdisc *, unsigned long cl,
					struct Qdisc *, struct Qdisc **);
	struct Qdisc *		(*leaf)(struct Qdisc *, unsigned long cl);
147
	void			(*qlen_notify)(struct Qdisc *, unsigned long);
L
Linus Torvalds 已提交
148 149 150 151 152

	/* Class manipulation routines */
	unsigned long		(*get)(struct Qdisc *, u32 classid);
	void			(*put)(struct Qdisc *, unsigned long);
	int			(*change)(struct Qdisc *, u32, u32,
153
					struct nlattr **, unsigned long *);
L
Linus Torvalds 已提交
154 155 156 157
	int			(*delete)(struct Qdisc *, unsigned long);
	void			(*walk)(struct Qdisc *, struct qdisc_walker * arg);

	/* Filter manipulation */
158
	struct tcf_block *	(*tcf_block)(struct Qdisc *, unsigned long);
159
	bool			(*tcf_cl_offload)(u32 classid);
L
Linus Torvalds 已提交
160 161 162 163 164 165 166 167 168 169 170
	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
171
struct Qdisc_ops {
L
Linus Torvalds 已提交
172
	struct Qdisc_ops	*next;
173
	const struct Qdisc_class_ops	*cl_ops;
L
Linus Torvalds 已提交
174 175 176
	char			id[IFNAMSIZ];
	int			priv_size;

177 178 179
	int 			(*enqueue)(struct sk_buff *skb,
					   struct Qdisc *sch,
					   struct sk_buff **to_free);
L
Linus Torvalds 已提交
180
	struct sk_buff *	(*dequeue)(struct Qdisc *);
181
	struct sk_buff *	(*peek)(struct Qdisc *);
L
Linus Torvalds 已提交
182

183
	int			(*init)(struct Qdisc *, struct nlattr *arg);
L
Linus Torvalds 已提交
184 185
	void			(*reset)(struct Qdisc *);
	void			(*destroy)(struct Qdisc *);
186
	int			(*change)(struct Qdisc *, struct nlattr *arg);
187
	void			(*attach)(struct Qdisc *);
L
Linus Torvalds 已提交
188 189 190 191 192 193 194 195

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

	struct module		*owner;
};


E
Eric Dumazet 已提交
196
struct tcf_result {
197 198 199 200 201 202 203
	union {
		struct {
			unsigned long	class;
			u32		classid;
		};
		const struct tcf_proto *goto_tp;
	};
L
Linus Torvalds 已提交
204 205
};

E
Eric Dumazet 已提交
206
struct tcf_proto_ops {
207
	struct list_head	head;
L
Linus Torvalds 已提交
208 209
	char			kind[IFNAMSIZ];

210 211 212
	int			(*classify)(struct sk_buff *,
					    const struct tcf_proto *,
					    struct tcf_result *);
L
Linus Torvalds 已提交
213
	int			(*init)(struct tcf_proto*);
214
	void			(*destroy)(struct tcf_proto*);
L
Linus Torvalds 已提交
215 216

	unsigned long		(*get)(struct tcf_proto*, u32 handle);
217
	int			(*change)(struct net *net, struct sk_buff *,
218
					struct tcf_proto*, unsigned long,
219
					u32 handle, struct nlattr **,
220
					unsigned long *, bool);
221
	int			(*delete)(struct tcf_proto*, unsigned long, bool*);
L
Linus Torvalds 已提交
222 223 224
	void			(*walk)(struct tcf_proto*, struct tcf_walker *arg);

	/* rtnetlink specific */
225
	int			(*dump)(struct net*, struct tcf_proto*, unsigned long,
L
Linus Torvalds 已提交
226 227 228 229 230
					struct sk_buff *skb, struct tcmsg*);

	struct module		*owner;
};

E
Eric Dumazet 已提交
231
struct tcf_proto {
L
Linus Torvalds 已提交
232
	/* Fast access part */
J
John Fastabend 已提交
233 234
	struct tcf_proto __rcu	*next;
	void __rcu		*root;
235 236 237
	int			(*classify)(struct sk_buff *,
					    const struct tcf_proto *,
					    struct tcf_result *);
A
Al Viro 已提交
238
	__be16			protocol;
L
Linus Torvalds 已提交
239 240 241 242 243 244

	/* All the rest */
	u32			prio;
	u32			classid;
	struct Qdisc		*q;
	void			*data;
245
	const struct tcf_proto_ops	*ops;
246
	struct tcf_chain	*chain;
J
John Fastabend 已提交
247
	struct rcu_head		rcu;
L
Linus Torvalds 已提交
248 249
};

250 251
struct qdisc_skb_cb {
	unsigned int		pkt_len;
252
	u16			slave_dev_queue_mapping;
253
	u16			tc_classid;
254 255
#define QDISC_CB_PRIV_LEN 20
	unsigned char		data[QDISC_CB_PRIV_LEN];
256 257
};

258 259
struct tcf_chain {
	struct tcf_proto __rcu *filter_chain;
260
	struct tcf_proto __rcu **p_filter_chain;
261 262 263 264
	struct list_head list;
	struct tcf_block *block;
	u32 index; /* chain index */
	unsigned int refcnt;
265 266
};

267
struct tcf_block {
268
	struct list_head chain_list;
269 270
};

271 272 273
static inline void qdisc_cb_private_validate(const struct sk_buff *skb, int sz)
{
	struct qdisc_skb_cb *qcb;
274 275

	BUILD_BUG_ON(sizeof(skb->cb) < offsetof(struct qdisc_skb_cb, data) + sz);
276 277 278
	BUILD_BUG_ON(sizeof(qcb->data) < sz);
}

279
static inline int qdisc_qlen(const struct Qdisc *q)
280 281 282 283
{
	return q->q.qlen;
}

284
static inline struct qdisc_skb_cb *qdisc_skb_cb(const struct sk_buff *skb)
285 286 287 288
{
	return (struct qdisc_skb_cb *)skb->cb;
}

289 290 291 292 293
static inline spinlock_t *qdisc_lock(struct Qdisc *qdisc)
{
	return &qdisc->q.lock;
}

294
static inline struct Qdisc *qdisc_root(const struct Qdisc *qdisc)
295
{
296 297 298
	struct Qdisc *q = rcu_dereference_rtnl(qdisc->dev_queue->qdisc);

	return q;
299 300
}

301
static inline struct Qdisc *qdisc_root_sleeping(const struct Qdisc *qdisc)
302 303 304 305
{
	return qdisc->dev_queue->qdisc_sleeping;
}

306 307 308 309 310 311 312 313 314 315 316
/* 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.
 */
317
static inline spinlock_t *qdisc_root_lock(const struct Qdisc *qdisc)
318 319 320
{
	struct Qdisc *root = qdisc_root(qdisc);

321
	ASSERT_RTNL();
322
	return qdisc_lock(root);
323 324
}

325
static inline spinlock_t *qdisc_root_sleeping_lock(const struct Qdisc *qdisc)
326 327 328 329 330 331 332
{
	struct Qdisc *root = qdisc_root_sleeping(qdisc);

	ASSERT_RTNL();
	return qdisc_lock(root);
}

333 334 335 336 337 338 339 340
static inline seqcount_t *qdisc_root_sleeping_running(const struct Qdisc *qdisc)
{
	struct Qdisc *root = qdisc_root_sleeping(qdisc);

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

341
static inline struct net_device *qdisc_dev(const struct Qdisc *qdisc)
342 343 344
{
	return qdisc->dev_queue->dev;
}
L
Linus Torvalds 已提交
345

346
static inline void sch_tree_lock(const struct Qdisc *q)
347
{
J
Jarek Poplawski 已提交
348
	spin_lock_bh(qdisc_root_sleeping_lock(q));
349 350
}

351
static inline void sch_tree_unlock(const struct Qdisc *q)
352
{
J
Jarek Poplawski 已提交
353
	spin_unlock_bh(qdisc_root_sleeping_lock(q));
354 355 356 357
}

#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
358

359 360
extern struct Qdisc noop_qdisc;
extern struct Qdisc_ops noop_qdisc_ops;
361 362
extern struct Qdisc_ops pfifo_fast_ops;
extern struct Qdisc_ops mq_qdisc_ops;
P
Phil Sutter 已提交
363
extern struct Qdisc_ops noqueue_qdisc_ops;
364
extern const struct Qdisc_ops *default_qdisc_ops;
365 366 367 368 369 370
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;
}
371

E
Eric Dumazet 已提交
372
struct Qdisc_class_common {
373 374 375 376
	u32			classid;
	struct hlist_node	hnode;
};

E
Eric Dumazet 已提交
377
struct Qdisc_class_hash {
378 379 380 381 382 383 384 385 386 387 388 389 390 391
	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 *
392
qdisc_class_find(const struct Qdisc_class_hash *hash, u32 id)
393 394 395 396 397
{
	struct Qdisc_class_common *cl;
	unsigned int h;

	h = qdisc_class_hash(id, hash->hashmask);
398
	hlist_for_each_entry(cl, &hash->hash[h], hnode) {
399 400 401 402 403 404
		if (cl->classid == id)
			return cl;
	}
	return NULL;
}

405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421
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);
422 423
void qdisc_tree_reduce_backlog(struct Qdisc *qdisc, unsigned int n,
			       unsigned int len);
424
struct Qdisc *qdisc_alloc(struct netdev_queue *dev_queue,
425
			  const struct Qdisc_ops *ops);
426
struct Qdisc *qdisc_create_dflt(struct netdev_queue *dev_queue,
427
				const struct Qdisc_ops *ops, u32 parentid);
428 429
void __qdisc_calculate_pkt_len(struct sk_buff *skb,
			       const struct qdisc_size_table *stab);
430
int skb_do_redirect(struct sk_buff *);
L
Linus Torvalds 已提交
431

432 433 434
static inline void skb_reset_tc(struct sk_buff *skb)
{
#ifdef CONFIG_NET_CLS_ACT
435
	skb->tc_redirected = 0;
436 437 438
#endif
}

439 440 441
static inline bool skb_at_tc_ingress(const struct sk_buff *skb)
{
#ifdef CONFIG_NET_CLS_ACT
442
	return skb->tc_at_ingress;
443 444 445 446 447
#else
	return false;
#endif
}

448 449 450 451 452 453 454 455 456 457 458
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;
}

459 460
/* 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)
461
{
462 463
	struct Qdisc *qdisc;

464
	for (; i < dev->num_tx_queues; i++) {
465
		qdisc = rtnl_dereference(netdev_get_tx_queue(dev, i)->qdisc);
466 467 468 469 470 471
		if (qdisc) {
			spin_lock_bh(qdisc_lock(qdisc));
			qdisc_reset(qdisc);
			spin_unlock_bh(qdisc_lock(qdisc));
		}
	}
472 473 474 475
}

static inline void qdisc_reset_all_tx(struct net_device *dev)
{
476
	qdisc_reset_all_tx_gt(dev, 0);
477 478
}

479 480 481
/* Are all TX queues of the device empty?  */
static inline bool qdisc_all_tx_empty(const struct net_device *dev)
{
482
	unsigned int i;
483 484

	rcu_read_lock();
485 486
	for (i = 0; i < dev->num_tx_queues; i++) {
		struct netdev_queue *txq = netdev_get_tx_queue(dev, i);
487
		const struct Qdisc *q = rcu_dereference(txq->qdisc);
488

489 490
		if (q->q.qlen) {
			rcu_read_unlock();
491
			return false;
492
		}
493
	}
494
	rcu_read_unlock();
495
	return true;
496 497
}

498
/* Are any of the TX qdiscs changing?  */
499
static inline bool qdisc_tx_changing(const struct net_device *dev)
500
{
501
	unsigned int i;
502

503 504
	for (i = 0; i < dev->num_tx_queues; i++) {
		struct netdev_queue *txq = netdev_get_tx_queue(dev, i);
505
		if (rcu_access_pointer(txq->qdisc) != txq->qdisc_sleeping)
506 507 508
			return true;
	}
	return false;
509 510
}

511
/* Is the device using the noop qdisc on all queues?  */
512 513
static inline bool qdisc_tx_is_noop(const struct net_device *dev)
{
514
	unsigned int i;
515

516 517
	for (i = 0; i < dev->num_tx_queues; i++) {
		struct netdev_queue *txq = netdev_get_tx_queue(dev, i);
518
		if (rcu_access_pointer(txq->qdisc) != &noop_qdisc)
519 520 521
			return false;
	}
	return true;
522 523
}

524
static inline unsigned int qdisc_pkt_len(const struct sk_buff *skb)
525
{
526
	return qdisc_skb_cb(skb)->pkt_len;
527 528
}

529
/* additional qdisc xmit flags (NET_XMIT_MASK in linux/netdevice.h) */
530 531
enum net_xmit_qdisc_t {
	__NET_XMIT_STOLEN = 0x00010000,
532
	__NET_XMIT_BYPASS = 0x00020000,
533 534
};

535
#ifdef CONFIG_NET_CLS_ACT
536 537 538 539 540
#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
541 542
static inline void qdisc_calculate_pkt_len(struct sk_buff *skb,
					   const struct Qdisc *sch)
543
{
544
#ifdef CONFIG_NET_SCHED
E
Eric Dumazet 已提交
545 546 547 548
	struct qdisc_size_table *stab = rcu_dereference_bh(sch->stab);

	if (stab)
		__qdisc_calculate_pkt_len(skb, stab);
549
#endif
E
Eric Dumazet 已提交
550 551
}

552 553
static inline int qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch,
				struct sk_buff **to_free)
E
Eric Dumazet 已提交
554 555
{
	qdisc_calculate_pkt_len(skb, sch);
556
	return sch->enqueue(skb, sch, to_free);
557 558
}

559 560 561 562
static inline bool qdisc_is_percpu_stats(const struct Qdisc *q)
{
	return q->flags & TCQ_F_CPUSTATS;
}
563

564 565 566 567 568 569 570
static inline void _bstats_update(struct gnet_stats_basic_packed *bstats,
				  __u64 bytes, __u32 packets)
{
	bstats->bytes += bytes;
	bstats->packets += packets;
}

571 572 573
static inline void bstats_update(struct gnet_stats_basic_packed *bstats,
				 const struct sk_buff *skb)
{
574 575 576 577 578 579 580 581 582 583 584
	_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);
585 586
}

587 588
static inline void bstats_cpu_update(struct gnet_stats_basic_cpu *bstats,
				     const struct sk_buff *skb)
589 590 591 592 593 594
{
	u64_stats_update_begin(&bstats->syncp);
	bstats_update(&bstats->bstats, skb);
	u64_stats_update_end(&bstats->syncp);
}

595 596 597 598 599 600
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);
}

601 602
static inline void qdisc_bstats_update(struct Qdisc *sch,
				       const struct sk_buff *skb)
603
{
604
	bstats_update(&sch->bstats, skb);
605 606
}

607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623
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;
}

624
static inline void qstats_drop_inc(struct gnet_stats_queue *qstats)
625
{
626
	qstats->drops++;
627 628
}

629
static inline void qstats_overlimit_inc(struct gnet_stats_queue *qstats)
630
{
631 632
	qstats->overlimits++;
}
633

634 635 636 637 638 639 640
static inline void qdisc_qstats_drop(struct Qdisc *sch)
{
	qstats_drop_inc(&sch->qstats);
}

static inline void qdisc_qstats_cpu_drop(struct Qdisc *sch)
{
641
	this_cpu_inc(sch->cpu_qstats->drops);
642 643
}

644 645 646 647 648
static inline void qdisc_qstats_overlimit(struct Qdisc *sch)
{
	sch->qstats.overlimits++;
}

649 650 651 652 653 654 655
static inline void qdisc_skb_head_init(struct qdisc_skb_head *qh)
{
	qh->head = NULL;
	qh->tail = NULL;
	qh->qlen = 0;
}

656
static inline int __qdisc_enqueue_tail(struct sk_buff *skb, struct Qdisc *sch,
657
				       struct qdisc_skb_head *qh)
658
{
659 660 661 662 663 664 665 666 667 668 669
	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++;
670
	qdisc_qstats_backlog_inc(sch, skb);
671 672 673 674 675 676 677 678 679

	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);
}

680
static inline struct sk_buff *__qdisc_dequeue_head(struct qdisc_skb_head *qh)
681
{
682 683 684 685 686 687 688 689 690
	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;
	}
691

692 693 694 695 696 697 698
	return skb;
}

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

699
	if (likely(skb != NULL)) {
700
		qdisc_qstats_backlog_dec(sch, skb);
701 702
		qdisc_bstats_update(sch, skb);
	}
703 704 705 706

	return skb;
}

707 708 709 710 711 712 713 714 715
/* 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;
}

716
static inline unsigned int __qdisc_queue_drop_head(struct Qdisc *sch,
717
						   struct qdisc_skb_head *qh,
718
						   struct sk_buff **to_free)
719
{
720
	struct sk_buff *skb = __qdisc_dequeue_head(qh);
721 722 723

	if (likely(skb != NULL)) {
		unsigned int len = qdisc_pkt_len(skb);
724

725
		qdisc_qstats_backlog_dec(sch, skb);
726
		__qdisc_drop(skb, to_free);
727 728 729 730 731 732
		return len;
	}

	return 0;
}

733 734
static inline unsigned int qdisc_queue_drop_head(struct Qdisc *sch,
						 struct sk_buff **to_free)
735
{
736
	return __qdisc_queue_drop_head(sch, &sch->q, to_free);
737 738
}

739 740
static inline struct sk_buff *qdisc_peek_head(struct Qdisc *sch)
{
741 742 743
	const struct qdisc_skb_head *qh = &sch->q;

	return qh->head;
744 745
}

746 747 748 749
/* 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 */
750
	if (!sch->gso_skb) {
751
		sch->gso_skb = sch->dequeue(sch);
752
		if (sch->gso_skb) {
753
			/* it's still part of the queue */
754
			qdisc_qstats_backlog_inc(sch, sch->gso_skb);
755
			sch->q.qlen++;
756
		}
757
	}
758 759 760 761 762 763 764 765 766

	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;

767
	if (skb) {
768
		sch->gso_skb = NULL;
769
		qdisc_qstats_backlog_dec(sch, skb);
770 771
		sch->q.qlen--;
	} else {
772
		skb = sch->dequeue(sch);
773
	}
774 775 776 777

	return skb;
}

778
static inline void __qdisc_reset_queue(struct qdisc_skb_head *qh)
779 780 781 782 783
{
	/*
	 * We do not know the backlog in bytes of this list, it
	 * is up to the caller to correct it
	 */
784 785 786 787 788 789 790
	ASSERT_RTNL();
	if (qh->qlen) {
		rtnl_kfree_skbs(qh->head, qh->tail);

		qh->head = NULL;
		qh->tail = NULL;
		qh->qlen = 0;
791
	}
792 793 794 795
}

static inline void qdisc_reset_queue(struct Qdisc *sch)
{
796
	__qdisc_reset_queue(&sch->q);
797 798 799
	sch->qstats.backlog = 0;
}

800 801 802 803 804 805 806 807 808
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) {
809 810 811
		unsigned int qlen = old->q.qlen;
		unsigned int backlog = old->qstats.backlog;

812
		qdisc_reset(old);
813
		qdisc_tree_reduce_backlog(old, qlen, backlog);
814 815 816 817 818 819
	}
	sch_tree_unlock(sch);

	return old;
}

820 821 822 823 824 825
static inline void rtnl_qdisc_drop(struct sk_buff *skb, struct Qdisc *sch)
{
	rtnl_kfree_skbs(skb, skb);
	qdisc_qstats_drop(sch);
}

826 827 828

static inline int qdisc_drop(struct sk_buff *skb, struct Qdisc *sch,
			     struct sk_buff **to_free)
829
{
830
	__qdisc_drop(skb, to_free);
831
	qdisc_qstats_drop(sch);
832 833 834 835

	return NET_XMIT_DROP;
}

836 837 838 839 840
/* 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)
{
841 842 843
	int slot = pktlen + rtab->rate.cell_align + rtab->rate.overhead;
	if (slot < 0)
		slot = 0;
844 845
	slot >>= rtab->rate.cell_log;
	if (slot > 255)
E
Eric Dumazet 已提交
846
		return rtab->data[255]*(slot >> 8) + rtab->data[slot & 0xFF];
847 848 849
	return rtab->data[slot];
}

850
struct psched_ratecfg {
851
	u64	rate_bytes_ps; /* bytes per second */
852 853
	u32	mult;
	u16	overhead;
854
	u8	linklayer;
855
	u8	shift;
856 857 858 859 860
};

static inline u64 psched_l2t_ns(const struct psched_ratecfg *r,
				unsigned int len)
{
861 862 863 864 865 866
	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;
867 868
}

869
void psched_ratecfg_precompute(struct psched_ratecfg *r,
870 871
			       const struct tc_ratespec *conf,
			       u64 rate64);
872

873 874
static inline void psched_ratecfg_getrate(struct tc_ratespec *res,
					  const struct psched_ratecfg *r)
875
{
876
	memset(res, 0, sizeof(*res));
877 878 879 880 881 882 883

	/* 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);

884
	res->overhead = r->overhead;
885
	res->linklayer = (r->linklayer & TC_LINKLAYER_MASK);
886 887
}

L
Linus Torvalds 已提交
888
#endif