aboutsummaryrefslogblamecommitdiffstats
path: root/net/sched/sch_api.c
blob: 05e6e0a799da994f74681d71203c2c7fcf51f817 (plain) (tree)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
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
100
101
102
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
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
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
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
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
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
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
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
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
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
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
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
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
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762

























































































































































































































































































































































































































































































































































































































































































































































































                                                                                                  
                                                                





                                        
                                                                   

                                    

                           

































































































































































































































                                                                                                                
                                                                 






                                                        
                                                                   

























































































































































































































































































                                                                                                                 
                            




                                
/*
 * 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
 */

#include <linux/config.h>
#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <linux/rtnetlink.h>
#include <linux/init.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
#include <linux/kmod.h>
#include <linux/list.h>
#include <linux/bitops.h>

#include <net/sock.h>
#include <net/pkt_sched.h>

#include <asm/processor.h>
#include <asm/uaccess.h>
#include <asm/system.h>

static int qdisc_notify(struct sk_buff *oskb, struct nlmsghdr *n, u32 clid,
			struct Qdisc *old, struct Qdisc *new);
static int tclass_notify(struct sk_buff *oskb, struct nlmsghdr *n,
			 struct Qdisc *q, unsigned long cl, int event);

/*

   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:

   ---requeue

   requeues once dequeued packet. It is used for non-standard or
   just buggy devices, which can defer output even if dev->tbusy=0.

   ---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;

/* Register/uregister queueing discipline */

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;
	if (qops->requeue == NULL)
		qops->requeue = noop_qdisc_ops.requeue;
	if (qops->dequeue == NULL)
		qops->dequeue = noop_qdisc_ops.dequeue;

	qops->next = NULL;
	*qp = qops;
	rc = 0;
out:
	write_unlock(&qdisc_mod_lock);
	return rc;
}

int unregister_qdisc(struct Qdisc_ops *qops)
{
	struct Qdisc_ops *q, **qp;
	int err = -ENOENT;

	write_lock(&qdisc_mod_lock);
	for (qp = &qdisc_base; (q=*qp)!=NULL; qp = &q->next)
		if (q == qops)
			break;
	if (q) {
		*qp = q->next;
		q->next = NULL;
		err = 0;
	}
	write_unlock(&qdisc_mod_lock);
	return err;
}

/* We know handle. Find qdisc among all qdisc's attached to device
   (root qdisc, all its children, children of children etc.)
 */

struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle)
{
	struct Qdisc *q;

	read_lock_bh(&qdisc_tree_lock);
	list_for_each_entry(q, &dev->qdisc_list, list) {
		if (q->handle == handle) {
			read_unlock_bh(&qdisc_tree_lock);
			return q;
		}
	}
	read_unlock_bh(&qdisc_tree_lock);
	return NULL;
}

static struct Qdisc *qdisc_leaf(struct Qdisc *p, u32 classid)
{
	unsigned long cl;
	struct Qdisc *leaf;
	struct Qdisc_class_ops *cops = p->ops->cl_ops;

	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 */

static struct Qdisc_ops *qdisc_lookup_ops(struct rtattr *kind)
{
	struct Qdisc_ops *q = NULL;

	if (kind) {
		read_lock(&qdisc_mod_lock);
		for (q = qdisc_base; q; q = q->next) {
			if (rtattr_strcmp(kind, q->id) == 0) {
				if (!try_module_get(q->owner))
					q = NULL;
				break;
			}
		}
		read_unlock(&qdisc_mod_lock);
	}
	return q;
}

static struct qdisc_rate_table *qdisc_rtab_list;

struct qdisc_rate_table *qdisc_get_rtab(struct tc_ratespec *r, struct rtattr *tab)
{
	struct qdisc_rate_table *rtab;

	for (rtab = qdisc_rtab_list; rtab; rtab = rtab->next) {
		if (memcmp(&rtab->rate, r, sizeof(struct tc_ratespec)) == 0) {
			rtab->refcnt++;
			return rtab;
		}
	}

	if (tab == NULL || r->rate == 0 || r->cell_log == 0 || RTA_PAYLOAD(tab) != 1024)
		return NULL;

	rtab = kmalloc(sizeof(*rtab), GFP_KERNEL);
	if (rtab) {
		rtab->rate = *r;
		rtab->refcnt = 1;
		memcpy(rtab->data, RTA_DATA(tab), 1024);
		rtab->next = qdisc_rtab_list;
		qdisc_rtab_list = rtab;
	}
	return rtab;
}

void qdisc_put_rtab(struct qdisc_rate_table *tab)
{
	struct qdisc_rate_table *rtab, **rtabp;

	if (!tab || --tab->refcnt)
		return;

	for (rtabp = &qdisc_rtab_list; (rtab=*rtabp) != NULL; rtabp = &rtab->next) {
		if (rtab == tab) {
			*rtabp = rtab->next;
			kfree(rtab);
			return;
		}
	}
}


/* Allocate an unique handle from space managed by kernel */

static u32 qdisc_alloc_handle(struct net_device *dev)
{
	int i = 0x10000;
	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);
	} while	(qdisc_lookup(dev, autohandle) && --i > 0);

	return i>0 ? autohandle : 0;
}

/* Attach toplevel qdisc to device dev */

static struct Qdisc *
dev_graft_qdisc(struct net_device *dev, struct Qdisc *qdisc)
{
	struct Qdisc *oqdisc;

	if (dev->flags & IFF_UP)
		dev_deactivate(dev);

	qdisc_lock_tree(dev);
	if (qdisc && qdisc->flags&TCQ_F_INGRESS) {
		oqdisc = dev->qdisc_ingress;
		/* Prune old scheduler */
		if (oqdisc && atomic_read(&oqdisc->refcnt) <= 1) {
			/* delete */
			qdisc_reset(oqdisc);
			dev->qdisc_ingress = NULL;
		} else {  /* new */
			dev->qdisc_ingress = qdisc;
		}

	} else {

		oqdisc = dev->qdisc_sleeping;

		/* Prune old scheduler */
		if (oqdisc && atomic_read(&oqdisc->refcnt) <= 1)
			qdisc_reset(oqdisc);

		/* ... and graft new one */
		if (qdisc == NULL)
			qdisc = &noop_qdisc;
		dev->qdisc_sleeping = qdisc;
		dev->qdisc = &noop_qdisc;
	}

	qdisc_unlock_tree(dev);

	if (dev->flags & IFF_UP)
		dev_activate(dev);

	return oqdisc;
}


/* Graft qdisc "new" to class "classid" of qdisc "parent" or
   to device "dev".

   Old qdisc is not destroyed but returned in *old.
 */

static int qdisc_graft(struct net_device *dev, struct Qdisc *parent,
		       u32 classid,
		       struct Qdisc *new, struct Qdisc **old)
{
	int err = 0;
	struct Qdisc *q = *old;


	if (parent == NULL) { 
		if (q && q->flags&TCQ_F_INGRESS) {
			*old = dev_graft_qdisc(dev, q);
		} else {
			*old = dev_graft_qdisc(dev, new);
		}
	} else {
		struct Qdisc_class_ops *cops = parent->ops->cl_ops;

		err = -EINVAL;

		if (cops) {
			unsigned long cl = cops->get(parent, classid);
			if (cl) {
				err = cops->graft(parent, cl, new, old);
				if (new)
					new->parent = classid;
				cops->put(parent, cl);
			}
		}
	}
	return err;
}

/*
   Allocate and initialize new qdisc.

   Parameters are passed via opt.
 */

static struct Qdisc *
qdisc_create(struct net_device *dev, u32 handle, struct rtattr **tca, int *errp)
{
	int err;
	struct rtattr *kind = tca[TCA_KIND-1];
	void *p = NULL;
	struct Qdisc *sch;
	struct Qdisc_ops *ops;
	int size;

	ops = qdisc_lookup_ops(kind);
#ifdef CONFIG_KMOD
	if (ops == NULL && kind != NULL) {
		char name[IFNAMSIZ];
		if (rtattr_strlcpy(name, kind, IFNAMSIZ) < IFNAMSIZ) {
			/* 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

	err = -EINVAL;
	if (ops == NULL)
		goto err_out;

	/* ensure that the Qdisc and the private data are 32-byte aligned */
	size = ((sizeof(*sch) + QDISC_ALIGN_CONST) & ~QDISC_ALIGN_CONST);
	size += ops->priv_size + QDISC_ALIGN_CONST;

	p = kmalloc(size, GFP_KERNEL);
	err = -ENOBUFS;
	if (!p)
		goto err_out2;
	memset(p, 0, size);
	sch = (struct Qdisc *)(((unsigned long)p + QDISC_ALIGN_CONST)
	                       & ~QDISC_ALIGN_CONST);
	sch->padded = (char *)sch - (char *)p;

	INIT_LIST_HEAD(&sch->list);
	skb_queue_head_init(&sch->q);

	if (handle == TC_H_INGRESS)
		sch->flags |= TCQ_F_INGRESS;

	sch->ops = ops;
	sch->enqueue = ops->enqueue;
	sch->dequeue = ops->dequeue;
	sch->dev = dev;
	dev_hold(dev);
	atomic_set(&sch->refcnt, 1);
	sch->stats_lock = &dev->queue_lock;
	if (handle == 0) {
		handle = qdisc_alloc_handle(dev);
		err = -ENOMEM;
		if (handle == 0)
			goto err_out3;
	}

	if (handle == TC_H_INGRESS)
                sch->handle =TC_H_MAKE(TC_H_INGRESS, 0);
        else
                sch->handle = handle;

	if (!ops->init || (err = ops->init(sch, tca[TCA_OPTIONS-1])) == 0) {
		qdisc_lock_tree(dev);
		list_add_tail(&sch->list, &dev->qdisc_list);
		qdisc_unlock_tree(dev);

#ifdef CONFIG_NET_ESTIMATOR
		if (tca[TCA_RATE-1])
			gen_new_estimator(&sch->bstats, &sch->rate_est,
				sch->stats_lock, tca[TCA_RATE-1]);
#endif
		return sch;
	}
err_out3:
	dev_put(dev);
err_out2:
	module_put(ops->owner);
err_out:
	*errp = err;
	if (p)
		kfree(p);
	return NULL;
}

static int qdisc_change(struct Qdisc *sch, struct rtattr **tca)
{
	if (tca[TCA_OPTIONS-1]) {
		int err;

		if (sch->ops->change == NULL)
			return -EINVAL;
		err = sch->ops->change(sch, tca[TCA_OPTIONS-1]);
		if (err)
			return err;
	}
#ifdef CONFIG_NET_ESTIMATOR
	if (tca[TCA_RATE-1])
		gen_replace_estimator(&sch->bstats, &sch->rate_est,
			sch->stats_lock, tca[TCA_RATE-1]);
#endif
	return 0;
}

struct check_loop_arg
{
	struct qdisc_walker 	w;
	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;
	struct Qdisc_class_ops *cops = q->ops->cl_ops;
	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.
 */

static int tc_get_qdisc(struct sk_buff *skb, struct nlmsghdr *n, void *arg)
{
	struct tcmsg *tcm = NLMSG_DATA(n);
	struct rtattr **tca = arg;
	struct net_device *dev;
	u32 clid = tcm->tcm_parent;
	struct Qdisc *q = NULL;
	struct Qdisc *p = NULL;
	int err;

	if ((dev = __dev_get_by_index(tcm->tcm_ifindex)) == NULL)
		return -ENODEV;

	if (clid) {
		if (clid != TC_H_ROOT) {
			if (TC_H_MAJ(clid) != TC_H_MAJ(TC_H_INGRESS)) {
				if ((p = qdisc_lookup(dev, TC_H_MAJ(clid))) == NULL)
					return -ENOENT;
				q = qdisc_leaf(p, clid);
			} else { /* ingress */
				q = dev->qdisc_ingress;
                        }
		} else {
			q = dev->qdisc_sleeping;
		}
		if (!q)
			return -ENOENT;

		if (tcm->tcm_handle && q->handle != tcm->tcm_handle)
			return -EINVAL;
	} else {
		if ((q = qdisc_lookup(dev, tcm->tcm_handle)) == NULL)
			return -ENOENT;
	}

	if (tca[TCA_KIND-1] && rtattr_strcmp(tca[TCA_KIND-1], q->ops->id))
		return -EINVAL;

	if (n->nlmsg_type == RTM_DELQDISC) {
		if (!clid)
			return -EINVAL;
		if (q->handle == 0)
			return -ENOENT;
		if ((err = qdisc_graft(dev, p, clid, NULL, &q)) != 0)
			return err;
		if (q) {
			qdisc_notify(skb, n, clid, q, NULL);
			spin_lock_bh(&dev->queue_lock);
			qdisc_destroy(q);
			spin_unlock_bh(&dev->queue_lock);
		}
	} else {
		qdisc_notify(skb, n, clid, NULL, q);
	}
	return 0;
}

/*
   Create/change qdisc.
 */

static int tc_modify_qdisc(struct sk_buff *skb, struct nlmsghdr *n, void *arg)
{
	struct tcmsg *tcm;
	struct rtattr **tca;
	struct net_device *dev;
	u32 clid;
	struct Qdisc *q, *p;
	int err;

replay:
	/* Reinit, just in case something touches this. */
	tcm = NLMSG_DATA(n);
	tca = arg;
	clid = tcm->tcm_parent;
	q = p = NULL;

	if ((dev = __dev_get_by_index(tcm->tcm_ifindex)) == NULL)
		return -ENODEV;

	if (clid) {
		if (clid != TC_H_ROOT) {
			if (clid != TC_H_INGRESS) {
				if ((p = qdisc_lookup(dev, TC_H_MAJ(clid))) == NULL)
					return -ENOENT;
				q = qdisc_leaf(p, clid);
			} else { /*ingress */
				q = dev->qdisc_ingress;
			}
		} else {
			q = dev->qdisc_sleeping;
		}

		/* 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) {
				if (q && !(n->nlmsg_flags&NLM_F_REPLACE))
					return -EEXIST;
				if (TC_H_MIN(tcm->tcm_handle))
					return -EINVAL;
				if ((q = qdisc_lookup(dev, tcm->tcm_handle)) == NULL)
					goto create_n_graft;
				if (n->nlmsg_flags&NLM_F_EXCL)
					return -EEXIST;
				if (tca[TCA_KIND-1] && rtattr_strcmp(tca[TCA_KIND-1], q->ops->id))
					return -EINVAL;
				if (q == p ||
				    (p && check_loop(q, p, 0)))
					return -ELOOP;
				atomic_inc(&q->refcnt);
				goto graft;
			} else {
				if (q == NULL)
					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.
				 */
				if ((n->nlmsg_flags&NLM_F_CREATE) &&
				    (n->nlmsg_flags&NLM_F_REPLACE) &&
				    ((n->nlmsg_flags&NLM_F_EXCL) ||
				     (tca[TCA_KIND-1] &&
				      rtattr_strcmp(tca[TCA_KIND-1], q->ops->id))))
					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;
	if (n->nlmsg_flags&NLM_F_EXCL)
		return -EEXIST;
	if (tca[TCA_KIND-1] && rtattr_strcmp(tca[TCA_KIND-1], q->ops->id))
		return -EINVAL;
	err = qdisc_change(q, tca);
	if (err == 0)
		qdisc_notify(skb, n, clid, NULL, q);
	return err;

create_n_graft:
	if (!(n->nlmsg_flags&NLM_F_CREATE))
		return -ENOENT;
	if (clid == TC_H_INGRESS)
		q = qdisc_create(dev, tcm->tcm_parent, tca, &err);
        else
		q = qdisc_create(dev, tcm->tcm_handle, tca, &err);
	if (q == NULL) {
		if (err == -EAGAIN)
			goto replay;
		return err;
	}

graft:
	if (1) {
		struct Qdisc *old_q = NULL;
		err = qdisc_graft(dev, p, clid, q, &old_q);
		if (err) {
			if (q) {
				spin_lock_bh(&dev->queue_lock);
				qdisc_destroy(q);
				spin_unlock_bh(&dev->queue_lock);
			}
			return err;
		}
		qdisc_notify(skb, n, clid, old_q, q);
		if (old_q) {
			spin_lock_bh(&dev->queue_lock);
			qdisc_destroy(old_q);
			spin_unlock_bh(&dev->queue_lock);
		}
	}
	return 0;
}

static int tc_fill_qdisc(struct sk_buff *skb, struct Qdisc *q, u32 clid,
			 u32 pid, u32 seq, u16 flags, int event)
{
	struct tcmsg *tcm;
	struct nlmsghdr  *nlh;
	unsigned char	 *b = skb->tail;
	struct gnet_dump d;

	nlh = NLMSG_NEW(skb, pid, seq, event, sizeof(*tcm), flags);
	tcm = NLMSG_DATA(nlh);
	tcm->tcm_family = AF_UNSPEC;
	tcm->tcm__pad1 = 0;
	tcm->tcm__pad2 = 0;
	tcm->tcm_ifindex = q->dev->ifindex;
	tcm->tcm_parent = clid;
	tcm->tcm_handle = q->handle;
	tcm->tcm_info = atomic_read(&q->refcnt);
	RTA_PUT(skb, TCA_KIND, IFNAMSIZ, q->ops->id);
	if (q->ops->dump && q->ops->dump(q, skb) < 0)
		goto rtattr_failure;
	q->qstats.qlen = q->q.qlen;

	if (gnet_stats_start_copy_compat(skb, TCA_STATS2, TCA_STATS,
			TCA_XSTATS, q->stats_lock, &d) < 0)
		goto rtattr_failure;

	if (q->ops->dump_stats && q->ops->dump_stats(q, &d) < 0)
		goto rtattr_failure;

	if (gnet_stats_copy_basic(&d, &q->bstats) < 0 ||
#ifdef CONFIG_NET_ESTIMATOR
	    gnet_stats_copy_rate_est(&d, &q->rate_est) < 0 ||
#endif
	    gnet_stats_copy_queue(&d, &q->qstats) < 0)
		goto rtattr_failure;
	
	if (gnet_stats_finish_copy(&d) < 0)
		goto rtattr_failure;
	
	nlh->nlmsg_len = skb->tail - b;
	return skb->len;

nlmsg_failure:
rtattr_failure:
	skb_trim(skb, b - skb->data);
	return -1;
}

static int qdisc_notify(struct sk_buff *oskb, struct nlmsghdr *n,
			u32 clid, struct Qdisc *old, struct Qdisc *new)
{
	struct sk_buff *skb;
	u32 pid = oskb ? NETLINK_CB(oskb).pid : 0;

	skb = alloc_skb(NLMSG_GOODSIZE, GFP_KERNEL);
	if (!skb)
		return -ENOBUFS;

	if (old && old->handle) {
		if (tc_fill_qdisc(skb, old, clid, pid, n->nlmsg_seq, 0, RTM_DELQDISC) < 0)
			goto err_out;
	}
	if (new) {
		if (tc_fill_qdisc(skb, new, clid, pid, n->nlmsg_seq, old ? NLM_F_REPLACE : 0, RTM_NEWQDISC) < 0)
			goto err_out;
	}

	if (skb->len)
		return rtnetlink_send(skb, pid, RTMGRP_TC, n->nlmsg_flags&NLM_F_ECHO);

err_out:
	kfree_skb(skb);
	return -EINVAL;
}

static int tc_dump_qdisc(struct sk_buff *skb, struct netlink_callback *cb)
{
	int idx, q_idx;
	int s_idx, s_q_idx;
	struct net_device *dev;
	struct Qdisc *q;

	s_idx = cb->args[0];
	s_q_idx = q_idx = cb->args[1];
	read_lock(&dev_base_lock);
	for (dev=dev_base, idx=0; dev; dev = dev->next, idx++) {
		if (idx < s_idx)
			continue;
		if (idx > s_idx)
			s_q_idx = 0;
		read_lock_bh(&qdisc_tree_lock);
		q_idx = 0;
		list_for_each_entry(q, &dev->qdisc_list, list) {
			if (q_idx < s_q_idx) {
				q_idx++;
				continue;
			}
			if (tc_fill_qdisc(skb, q, q->parent, NETLINK_CB(cb->skb).pid,
					  cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWQDISC) <= 0) {
				read_unlock_bh(&qdisc_tree_lock);
				goto done;
			}
			q_idx++;
		}
		read_unlock_bh(&qdisc_tree_lock);
	}

done:
	read_unlock(&dev_base_lock);

	cb->args[0] = idx;
	cb->args[1] = q_idx;

	return skb->len;
}



/************************************************
 *	Traffic classes manipulation.		*
 ************************************************/



static int tc_ctl_tclass(struct sk_buff *skb, struct nlmsghdr *n, void *arg)
{
	struct tcmsg *tcm = NLMSG_DATA(n);
	struct rtattr **tca = arg;
	struct net_device *dev;
	struct Qdisc *q = NULL;
	struct Qdisc_class_ops *cops;
	unsigned long cl = 0;
	unsigned long new_cl;
	u32 pid = tcm->tcm_parent;
	u32 clid = tcm->tcm_handle;
	u32 qid = TC_H_MAJ(clid);
	int err;

	if ((dev = __dev_get_by_index(tcm->tcm_ifindex)) == NULL)
		return -ENODEV;

	/*
	   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 */

	if (pid != TC_H_ROOT) {
		u32 qid1 = TC_H_MAJ(pid);

		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)
			qid = dev->qdisc_sleeping->handle;

		/* Now qid is genuine qdisc handle consistent
		   both with parent and child.

		   TC_H_MAJ(pid) still may be unspecified, complete it now.
		 */
		if (pid)
			pid = TC_H_MAKE(qid, pid);
	} else {
		if (qid == 0)
			qid = dev->qdisc_sleeping->handle;
	}

	/* OK. Locate qdisc */
	if ((q = qdisc_lookup(dev, qid)) == NULL) 
		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) {
		if (pid == TC_H_ROOT)
			clid = qid;
	} else
		clid = TC_H_MAKE(qid, clid);

	if (clid)
		cl = cops->get(q, clid);

	if (cl == 0) {
		err = -ENOENT;
		if (n->nlmsg_type != RTM_NEWTCLASS || !(n->nlmsg_flags&NLM_F_CREATE))
			goto out;
	} else {
		switch (n->nlmsg_type) {
		case RTM_NEWTCLASS:	
			err = -EEXIST;
			if (n->nlmsg_flags&NLM_F_EXCL)
				goto out;
			break;
		case RTM_DELTCLASS:
			err = cops->delete(q, cl);
			if (err == 0)
				tclass_notify(skb, n, q, cl, RTM_DELTCLASS);
			goto out;
		case RTM_GETTCLASS:
			err = tclass_notify(skb, n, q, cl, RTM_NEWTCLASS);
			goto out;
		default:
			err = -EINVAL;
			goto out;
		}
	}

	new_cl = cl;
	err = cops->change(q, clid, pid, tca, &new_cl);
	if (err == 0)
		tclass_notify(skb, n, q, new_cl, RTM_NEWTCLASS);

out:
	if (cl)
		cops->put(q, cl);

	return err;
}


static int tc_fill_tclass(struct sk_buff *skb, struct Qdisc *q,
			  unsigned long cl,
			  u32 pid, u32 seq, u16 flags, int event)
{
	struct tcmsg *tcm;
	struct nlmsghdr  *nlh;
	unsigned char	 *b = skb->tail;
	struct gnet_dump d;
	struct Qdisc_class_ops *cl_ops = q->ops->cl_ops;

	nlh = NLMSG_NEW(skb, pid, seq, event, sizeof(*tcm), flags);
	tcm = NLMSG_DATA(nlh);
	tcm->tcm_family = AF_UNSPEC;
	tcm->tcm_ifindex = q->dev->ifindex;
	tcm->tcm_parent = q->handle;
	tcm->tcm_handle = q->handle;
	tcm->tcm_info = 0;
	RTA_PUT(skb, TCA_KIND, IFNAMSIZ, q->ops->id);
	if (cl_ops->dump && cl_ops->dump(q, cl, skb, tcm) < 0)
		goto rtattr_failure;

	if (gnet_stats_start_copy_compat(skb, TCA_STATS2, TCA_STATS,
			TCA_XSTATS, q->stats_lock, &d) < 0)
		goto rtattr_failure;

	if (cl_ops->dump_stats && cl_ops->dump_stats(q, cl, &d) < 0)
		goto rtattr_failure;

	if (gnet_stats_finish_copy(&d) < 0)
		goto rtattr_failure;

	nlh->nlmsg_len = skb->tail - b;
	return skb->len;

nlmsg_failure:
rtattr_failure:
	skb_trim(skb, b - skb->data);
	return -1;
}

static int tclass_notify(struct sk_buff *oskb, struct nlmsghdr *n,
			  struct Qdisc *q, unsigned long cl, int event)
{
	struct sk_buff *skb;
	u32 pid = oskb ? NETLINK_CB(oskb).pid : 0;

	skb = alloc_skb(NLMSG_GOODSIZE, GFP_KERNEL);
	if (!skb)
		return -ENOBUFS;

	if (tc_fill_tclass(skb, q, cl, pid, n->nlmsg_seq, 0, event) < 0) {
		kfree_skb(skb);
		return -EINVAL;
	}

	return rtnetlink_send(skb, pid, RTMGRP_TC, n->nlmsg_flags&NLM_F_ECHO);
}

struct qdisc_dump_args
{
	struct qdisc_walker w;
	struct sk_buff *skb;
	struct netlink_callback *cb;
};

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;

	return tc_fill_tclass(a->skb, q, cl, NETLINK_CB(a->cb->skb).pid,
			      a->cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWTCLASS);
}

static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb)
{
	int t;
	int s_t;
	struct net_device *dev;
	struct Qdisc *q;
	struct tcmsg *tcm = (struct tcmsg*)NLMSG_DATA(cb->nlh);
	struct qdisc_dump_args arg;

	if (cb->nlh->nlmsg_len < NLMSG_LENGTH(sizeof(*tcm)))
		return 0;
	if ((dev = dev_get_by_index(tcm->tcm_ifindex)) == NULL)
		return 0;

	s_t = cb->args[0];
	t = 0;

	read_lock_bh(&qdisc_tree_lock);
	list_for_each_entry(q, &dev->qdisc_list, list) {
		if (t < s_t || !q->ops->cl_ops ||
		    (tcm->tcm_parent &&
		     TC_H_MAJ(tcm->tcm_parent) != q->handle)) {
			t++;
			continue;
		}
		if (t > 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)
			break;
		t++;
	}
	read_unlock_bh(&qdisc_tree_lock);

	cb->args[0] = t;

	dev_put(dev);
	return skb->len;
}

/* Main classifier routine: scans classifier chain attached
   to this qdisc, (optionally) tests for protocol and asks
   specific classifiers.
 */
int tc_classify(struct sk_buff *skb, struct tcf_proto *tp,
	struct tcf_result *res)
{
	int err = 0;
	u32 protocol = skb->protocol;
#ifdef CONFIG_NET_CLS_ACT
	struct tcf_proto *otp = tp;
reclassify:
#endif
	protocol = skb->protocol;

	for ( ; tp; tp = tp->next) {
		if ((tp->protocol == protocol ||
			tp->protocol == __constant_htons(ETH_P_ALL)) &&
			(err = tp->classify(skb, tp, res)) >= 0) {
#ifdef CONFIG_NET_CLS_ACT
			if ( TC_ACT_RECLASSIFY == err) {
				__u32 verd = (__u32) G_TC_VERD(skb->tc_verd);
				tp = otp;

				if (MAX_REC_LOOP < verd++) {
					printk("rule prio %d protocol %02x reclassify is buggy packet dropped\n",
						tp->prio&0xffff, ntohs(tp->protocol));
					return TC_ACT_SHOT;
				}
				skb->tc_verd = SET_TC_VERD(skb->tc_verd,verd);
				goto reclassify;
			} else {
				if (skb->tc_verd) 
					skb->tc_verd = SET_TC_VERD(skb->tc_verd,0);
				return err;
			}
#else

			return err;
#endif
		}

	}
	return -1;
}

static int psched_us_per_tick = 1;
static int psched_tick_per_us = 1;

#ifdef CONFIG_PROC_FS
static int psched_show(struct seq_file *seq, void *v)
{
	seq_printf(seq, "%08x %08x %08x %08x\n",
		      psched_tick_per_us, psched_us_per_tick,
		      1000000, HZ);

	return 0;
}

static int psched_open(struct inode *inode, struct file *file)
{
	return single_open(file, psched_show, PDE(inode)->data);
}

static struct file_operations psched_fops = {
	.owner = THIS_MODULE,
	.open = psched_open,
	.read  = seq_read,
	.llseek = seq_lseek,
	.release = single_release,
};	
#endif

#ifdef CONFIG_NET_SCH_CLK_CPU
psched_tdiff_t psched_clock_per_hz;
int psched_clock_scale;
EXPORT_SYMBOL(psched_clock_per_hz);
EXPORT_SYMBOL(psched_clock_scale);

psched_time_t psched_time_base;
cycles_t psched_time_mark;
EXPORT_SYMBOL(psched_time_mark);
EXPORT_SYMBOL(psched_time_base);

/*
 * Periodically adjust psched_time_base to avoid overflow
 * with 32-bit get_cycles(). Safe up to 4GHz CPU.
 */
static void psched_tick(unsigned long);
static struct timer_list psched_timer = TIMER_INITIALIZER(psched_tick, 0, 0);

static void psched_tick(unsigned long dummy)
{
	if (sizeof(cycles_t) == sizeof(u32)) {
		psched_time_t dummy_stamp;
		PSCHED_GET_TIME(dummy_stamp);
		psched_timer.expires = jiffies + 1*HZ;
		add_timer(&psched_timer);
	}
}

int __init psched_calibrate_clock(void)
{
	psched_time_t stamp, stamp1;
	struct timeval tv, tv1;
	psched_tdiff_t delay;
	long rdelay;
	unsigned long stop;

	psched_tick(0);
	stop = jiffies + HZ/10;
	PSCHED_GET_TIME(stamp);
	do_gettimeofday(&tv);
	while (time_before(jiffies, stop)) {
		barrier();
		cpu_relax();
	}
	PSCHED_GET_TIME(stamp1);
	do_gettimeofday(&tv1);

	delay = PSCHED_TDIFF(stamp1, stamp);
	rdelay = tv1.tv_usec - tv.tv_usec;
	rdelay += (tv1.tv_sec - tv.tv_sec)*1000000;
	if (rdelay > delay)
		return -1;
	delay /= rdelay;
	psched_tick_per_us = delay;
	while ((delay>>=1) != 0)
		psched_clock_scale++;
	psched_us_per_tick = 1<<psched_clock_scale;
	psched_clock_per_hz = (psched_tick_per_us*(1000000/HZ))>>psched_clock_scale;
	return 0;
}
#endif

static int __init pktsched_init(void)
{
	struct rtnetlink_link *link_p;

#ifdef CONFIG_NET_SCH_CLK_CPU
	if (psched_calibrate_clock() < 0)
		return -1;
#elif defined(CONFIG_NET_SCH_CLK_JIFFIES)
	psched_tick_per_us = HZ<<PSCHED_JSCALE;
	psched_us_per_tick = 1000000;
#endif

	link_p = rtnetlink_links[PF_UNSPEC];

	/* Setup rtnetlink links. It is made here to avoid
	   exporting large number of public symbols.
	 */

	if (link_p) {
		link_p[RTM_NEWQDISC-RTM_BASE].doit = tc_modify_qdisc;
		link_p[RTM_DELQDISC-RTM_BASE].doit = tc_get_qdisc;
		link_p[RTM_GETQDISC-RTM_BASE].doit = tc_get_qdisc;
		link_p[RTM_GETQDISC-RTM_BASE].dumpit = tc_dump_qdisc;
		link_p[RTM_NEWTCLASS-RTM_BASE].doit = tc_ctl_tclass;
		link_p[RTM_DELTCLASS-RTM_BASE].doit = tc_ctl_tclass;
		link_p[RTM_GETTCLASS-RTM_BASE].doit = tc_ctl_tclass;
		link_p[RTM_GETTCLASS-RTM_BASE].dumpit = tc_dump_tclass;
	}

	register_qdisc(&pfifo_qdisc_ops);
	register_qdisc(&bfifo_qdisc_ops);
	proc_net_fops_create("psched", 0, &psched_fops);

	return 0;
}

subsys_initcall(pktsched_init);

EXPORT_SYMBOL(qdisc_lookup);
EXPORT_SYMBOL(qdisc_get_rtab);
EXPORT_SYMBOL(qdisc_put_rtab);
EXPORT_SYMBOL(register_qdisc);
EXPORT_SYMBOL(unregister_qdisc);
EXPORT_SYMBOL(tc_classify);