summaryrefslogtreecommitdiffstats
path: root/drivers/usb/mon/mon_bin.c
blob: 81aab527024cbe637e120649c341d0106e6ffb6d (plain)
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
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
/*
 * The USB Monitor, inspired by Dave Harding's USBMon.
 *
 * This is a binary format reader.
 *
 * Copyright (C) 2006 Paolo Abeni (paolo.abeni@email.it)
 * Copyright (C) 2006,2007 Pete Zaitcev (zaitcev@redhat.com)
 */

#include <linux/kernel.h>
#include <linux/types.h>
#include <linux/fs.h>
#include <linux/cdev.h>
#include <linux/usb.h>
#include <linux/poll.h>
#include <linux/compat.h>
#include <linux/mm.h>

#include <asm/uaccess.h>

#include "usb_mon.h"

/*
 * Defined by USB 2.0 clause 9.3, table 9.2.
 */
#define SETUP_LEN  8

/* ioctl macros */
#define MON_IOC_MAGIC 0x92

#define MON_IOCQ_URB_LEN _IO(MON_IOC_MAGIC, 1)
/* #2 used to be MON_IOCX_URB, removed before it got into Linus tree */
#define MON_IOCG_STATS _IOR(MON_IOC_MAGIC, 3, struct mon_bin_stats)
#define MON_IOCT_RING_SIZE _IO(MON_IOC_MAGIC, 4)
#define MON_IOCQ_RING_SIZE _IO(MON_IOC_MAGIC, 5)
#define MON_IOCX_GET   _IOW(MON_IOC_MAGIC, 6, struct mon_bin_get)
#define MON_IOCX_MFETCH _IOWR(MON_IOC_MAGIC, 7, struct mon_bin_mfetch)
#define MON_IOCH_MFLUSH _IO(MON_IOC_MAGIC, 8)
#ifdef CONFIG_COMPAT
#define MON_IOCX_GET32 _IOW(MON_IOC_MAGIC, 6, struct mon_bin_get32)
#define MON_IOCX_MFETCH32 _IOWR(MON_IOC_MAGIC, 7, struct mon_bin_mfetch32)
#endif

/*
 * Some architectures have enormous basic pages (16KB for ia64, 64KB for ppc).
 * But it's all right. Just use a simple way to make sure the chunk is never
 * smaller than a page.
 *
 * N.B. An application does not know our chunk size.
 *
 * Woops, get_zeroed_page() returns a single page. I guess we're stuck with
 * page-sized chunks for the time being.
 */
#define CHUNK_SIZE   PAGE_SIZE
#define CHUNK_ALIGN(x)   (((x)+CHUNK_SIZE-1) & ~(CHUNK_SIZE-1))

/*
 * The magic limit was calculated so that it allows the monitoring
 * application to pick data once in two ticks. This way, another application,
 * which presumably drives the bus, gets to hog CPU, yet we collect our data.
 * If HZ is 100, a 480 mbit/s bus drives 614 KB every jiffy. USB has an
 * enormous overhead built into the bus protocol, so we need about 1000 KB.
 *
 * This is still too much for most cases, where we just snoop a few
 * descriptor fetches for enumeration. So, the default is a "reasonable"
 * amount for systems with HZ=250 and incomplete bus saturation.
 *
 * XXX What about multi-megabyte URBs which take minutes to transfer?
 */
#define BUFF_MAX  CHUNK_ALIGN(1200*1024)
#define BUFF_DFL   CHUNK_ALIGN(300*1024)
#define BUFF_MIN     CHUNK_ALIGN(8*1024)

/*
 * The per-event API header (2 per URB).
 *
 * This structure is seen in userland as defined by the documentation.
 */
struct mon_bin_hdr {
	u64 id;			/* URB ID - from submission to callback */
	unsigned char type;	/* Same as in text API; extensible. */
	unsigned char xfer_type;	/* ISO, Intr, Control, Bulk */
	unsigned char epnum;	/* Endpoint number and transfer direction */
	unsigned char devnum;	/* Device address */
	unsigned short busnum;	/* Bus number */
	char flag_setup;
	char flag_data;
	s64 ts_sec;		/* gettimeofday */
	s32 ts_usec;		/* gettimeofday */
	int status;
	unsigned int len_urb;	/* Length of data (submitted or actual) */
	unsigned int len_cap;	/* Delivered length */
	unsigned char setup[SETUP_LEN];	/* Only for Control S-type */
};

/* per file statistic */
struct mon_bin_stats {
	u32 queued;
	u32 dropped;
};

struct mon_bin_get {
	struct mon_bin_hdr __user *hdr;	/* Only 48 bytes, not 64. */
	void __user *data;
	size_t alloc;		/* Length of data (can be zero) */
};

struct mon_bin_mfetch {
	u32 __user *offvec;	/* Vector of events fetched */
	u32 nfetch;		/* Number of events to fetch (out: fetched) */
	u32 nflush;		/* Number of events to flush */
};

#ifdef CONFIG_COMPAT
struct mon_bin_get32 {
	u32 hdr32;
	u32 data32;
	u32 alloc32;
};

struct mon_bin_mfetch32 {
        u32 offvec32;
        u32 nfetch32;
        u32 nflush32;
};
#endif

/* Having these two values same prevents wrapping of the mon_bin_hdr */
#define PKT_ALIGN   64
#define PKT_SIZE    64

/* max number of USB bus supported */
#define MON_BIN_MAX_MINOR 128

/*
 * The buffer: map of used pages.
 */
struct mon_pgmap {
	struct page *pg;
	unsigned char *ptr;	/* XXX just use page_to_virt everywhere? */
};

/*
 * This gets associated with an open file struct.
 */
struct mon_reader_bin {
	/* The buffer: one per open. */
	spinlock_t b_lock;		/* Protect b_cnt, b_in */
	unsigned int b_size;		/* Current size of the buffer - bytes */
	unsigned int b_cnt;		/* Bytes used */
	unsigned int b_in, b_out;	/* Offsets into buffer - bytes */
	unsigned int b_read;		/* Amount of read data in curr. pkt. */
	struct mon_pgmap *b_vec;	/* The map array */
	wait_queue_head_t b_wait;	/* Wait for data here */

	struct mutex fetch_lock;	/* Protect b_read, b_out */
	int mmap_active;

	/* A list of these is needed for "bus 0". Some time later. */
	struct mon_reader r;

	/* Stats */
	unsigned int cnt_lost;
};

static inline struct mon_bin_hdr *MON_OFF2HDR(const struct mon_reader_bin *rp,
    unsigned int offset)
{
	return (struct mon_bin_hdr *)
	    (rp->b_vec[offset / CHUNK_SIZE].ptr + offset % CHUNK_SIZE);
}

#define MON_RING_EMPTY(rp)	((rp)->b_cnt == 0)

static unsigned char xfer_to_pipe[4] = {
	PIPE_CONTROL, PIPE_ISOCHRONOUS, PIPE_BULK, PIPE_INTERRUPT
};

static struct class *mon_bin_class;
static dev_t mon_bin_dev0;
static struct cdev mon_bin_cdev;

static void mon_buff_area_fill(const struct mon_reader_bin *rp,
    unsigned int offset, unsigned int size);
static int mon_bin_wait_event(struct file *file, struct mon_reader_bin *rp);
static int mon_alloc_buff(struct mon_pgmap *map, int npages);
static void mon_free_buff(struct mon_pgmap *map, int npages);

/*
 * This is a "chunked memcpy". It does not manipulate any counters.
 * But it returns the new offset for repeated application.
 */
unsigned int mon_copy_to_buff(const struct mon_reader_bin *this,
    unsigned int off, const unsigned char *from, unsigned int length)
{
	unsigned int step_len;
	unsigned char *buf;
	unsigned int in_page;

	while (length) {
		/*
		 * Determine step_len.
		 */
		step_len = length;
		in_page = CHUNK_SIZE - (off & (CHUNK_SIZE-1));
		if (in_page < step_len)
			step_len = in_page;

		/*
		 * Copy data and advance pointers.
		 */
		buf = this->b_vec[off / CHUNK_SIZE].ptr + off % CHUNK_SIZE;
		memcpy(buf, from, step_len);
		if ((off += step_len) >= this->b_size) off = 0;
		from += step_len;
		length -= step_len;
	}
	return off;
}

/*
 * This is a little worse than the above because it's "chunked copy_to_user".
 * The return value is an error code, not an offset.
 */
static int copy_from_buf(const struct mon_reader_bin *this, unsigned int off,
    char __user *to, int length)
{
	unsigned int step_len;
	unsigned char *buf;
	unsigned int in_page;

	while (length) {
		/*
		 * Determine step_len.
		 */
		step_len = length;
		in_page = CHUNK_SIZE - (off & (CHUNK_SIZE-1));
		if (in_page < step_len)
			step_len = in_page;

		/*
		 * Copy data and advance pointers.
		 */
		buf = this->b_vec[off / CHUNK_SIZE].ptr + off % CHUNK_SIZE;
		if (copy_to_user(to, buf, step_len))
			return -EINVAL;
		if ((off += step_len) >= this->b_size) off = 0;
		to += step_len;
		length -= step_len;
	}
	return 0;
}

/*
 * Allocate an (aligned) area in the buffer.
 * This is called under b_lock.
 * Returns ~0 on failure.
 */
static unsigned int mon_buff_area_alloc(struct mon_reader_bin *rp,
    unsigned int size)
{
	unsigned int offset;

	size = (size + PKT_ALIGN-1) & ~(PKT_ALIGN-1);
	if (rp->b_cnt + size > rp->b_size)
		return ~0;
	offset = rp->b_in;
	rp->b_cnt += size;
	if ((rp->b_in += size) >= rp->b_size)
		rp->b_in -= rp->b_size;
	return offset;
}

/*
 * This is the same thing as mon_buff_area_alloc, only it does not allow
 * buffers to wrap. This is needed by applications which pass references
 * into mmap-ed buffers up their stacks (libpcap can do that).
 *
 * Currently, we always have the header stuck with the data, although
 * it is not strictly speaking necessary.
 *
 * When a buffer would wrap, we place a filler packet to mark the space.
 */
static unsigned int mon_buff_area_alloc_contiguous(struct mon_reader_bin *rp,
    unsigned int size)
{
	unsigned int offset;
	unsigned int fill_size;

	size = (size + PKT_ALIGN-1) & ~(PKT_ALIGN-1);
	if (rp->b_cnt + size > rp->b_size)
		return ~0;
	if (rp->b_in + size > rp->b_size) {
		/*
		 * This would wrap. Find if we still have space after
		 * skipping to the end of the buffer. If we do, place
		 * a filler packet and allocate a new packet.
		 */
		fill_size = rp->b_size - rp->b_in;
		if (rp->b_cnt + size + fill_size > rp->b_size)
			return ~0;
		mon_buff_area_fill(rp, rp->b_in, fill_size);

		offset = 0;
		rp->b_in = size;
		rp->b_cnt += size + fill_size;
	} else if (rp->b_in + size == rp->b_size) {
		offset = rp->b_in;
		rp->b_in = 0;
		rp->b_cnt += size;
	} else {
		offset = rp->b_in;
		rp->b_in += size;
		rp->b_cnt += size;
	}
	return offset;
}

/*
 * Return a few (kilo-)bytes to the head of the buffer.
 * This is used if a DMA fetch fails.
 */
static void mon_buff_area_shrink(struct mon_reader_bin *rp, unsigned int size)
{

	size = (size + PKT_ALIGN-1) & ~(PKT_ALIGN-1);
	rp->b_cnt -= size;
	if (rp->b_in < size)
		rp->b_in += rp->b_size;
	rp->b_in -= size;
}

/*
 * This has to be called under both b_lock and fetch_lock, because
 * it accesses both b_cnt and b_out.
 */
static void mon_buff_area_free(struct mon_reader_bin *rp, unsigned int size)
{

	size = (size + PKT_ALIGN-1) & ~(PKT_ALIGN-1);
	rp->b_cnt -= size;
	if ((rp->b_out += size) >= rp->b_size)
		rp->b_out -= rp->b_size;
}

static void mon_buff_area_fill(const struct mon_reader_bin *rp,
    unsigned int offset, unsigned int size)
{
	struct mon_bin_hdr *ep;

	ep = MON_OFF2HDR(rp, offset);
	memset(ep, 0, PKT_SIZE);
	ep->type = '@';
	ep->len_cap = size - PKT_SIZE;
}

static inline char mon_bin_get_setup(unsigned char *setupb,
    const struct urb *urb, char ev_type)
{

	if (!usb_endpoint_xfer_control(&urb->ep->desc) || ev_type != 'S')
		return '-';

	if (urb->setup_packet == NULL)
		return 'Z';

	memcpy(setupb, urb->setup_packet, SETUP_LEN);
	return 0;
}

static char mon_bin_get_data(const struct mon_reader_bin *rp,
    unsigned int offset, struct urb *urb, unsigned int length)
{

	if (urb->dev->bus->uses_dma &&
	    (urb->transfer_flags & URB_NO_TRANSFER_DMA_MAP)) {
		mon_dmapeek_vec(rp, offset, urb->transfer_dma, length);
		return 0;
	}

	if (urb->transfer_buffer == NULL)
		return 'Z';

	mon_copy_to_buff(rp, offset, urb->transfer_buffer, length);
	return 0;
}

static void mon_bin_event(struct mon_reader_bin *rp, struct urb *urb,
    char ev_type, int status)
{
	const struct usb_endpoint_descriptor *epd = &urb->ep->desc;
	unsigned long flags;
	struct timeval ts;
	unsigned int urb_length;
	unsigned int offset;
	unsigned int length;
	unsigned char dir;
	struct mon_bin_hdr *ep;
	char data_tag = 0;

	do_gettimeofday(&ts);

	spin_lock_irqsave(&rp->b_lock, flags);

	/*
	 * Find the maximum allowable length, then allocate space.
	 */
	urb_length = (ev_type == 'S') ?
	    urb->transfer_buffer_length : urb->actual_length;
	length = urb_length;

	if (length >= rp->b_size/5)
		length = rp->b_size/5;

	if (usb_urb_dir_in(urb)) {
		if (ev_type == 'S') {
			length = 0;
			data_tag = '<';
		}
		/* Cannot rely on endpoint number in case of control ep.0 */
		dir = USB_DIR_IN;
	} else {
		if (ev_type == 'C') {
			length = 0;
			data_tag = '>';
		}
		dir = 0;
	}

	if (rp->mmap_active)
		offset = mon_buff_area_alloc_contiguous(rp, length + PKT_SIZE);
	else
		offset = mon_buff_area_alloc(rp, length + PKT_SIZE);
	if (offset == ~0) {
		rp->cnt_lost++;
		spin_unlock_irqrestore(&rp->b_lock, flags);
		return;
	}

	ep = MON_OFF2HDR(rp, offset);
	if ((offset += PKT_SIZE) >= rp->b_size) offset = 0;

	/*
	 * Fill the allocated area.
	 */
	memset(ep, 0, PKT_SIZE);
	ep->type = ev_type;
	ep->xfer_type = xfer_to_pipe[usb_endpoint_type(epd)];
	ep->epnum = dir | usb_endpoint_num(epd);
	ep->devnum = urb->dev->devnum;
	ep->busnum = urb->dev->bus->busnum;
	ep->id = (unsigned long) urb;
	ep->ts_sec = ts.tv_sec;
	ep->ts_usec = ts.tv_usec;
	ep->status = status;
	ep->len_urb = urb_length;
	ep->len_cap = length;

	ep->flag_setup = mon_bin_get_setup(ep->setup, urb, ev_type);
	if (length != 0) {
		ep->flag_data = mon_bin_get_data(rp, offset, urb, length);
		if (ep->flag_data != 0) {	/* Yes, it's 0x00, not '0' */
			ep->len_cap = 0;
			mon_buff_area_shrink(rp, length);
		}
	} else {
		ep->flag_data = data_tag;
	}

	spin_unlock_irqrestore(&rp->b_lock, flags);

	wake_up(&rp->b_wait);
}

static void mon_bin_submit(void *data, struct urb *urb)
{
	struct mon_reader_bin *rp = data;
	mon_bin_event(rp, urb, 'S', -EINPROGRESS);
}

static void mon_bin_complete(void *data, struct urb *urb, int status)
{
	struct mon_reader_bin *rp = data;
	mon_bin_event(rp, urb, 'C', status);
}

static void mon_bin_error(void *data, struct urb *urb, int error)
{
	struct mon_reader_bin *rp = data;
	unsigned long flags;
	unsigned int offset;
	struct mon_bin_hdr *ep;

	spin_lock_irqsave(&rp->b_lock, flags);

	offset = mon_buff_area_alloc(rp, PKT_SIZE);
	if (offset == ~0) {
		/* Not incrementing cnt_lost. Just because. */
		spin_unlock_irqrestore(&rp->b_lock, flags);
		return;
	}

	ep = MON_OFF2HDR(rp, offset);

	memset(ep, 0, PKT_SIZE);
	ep->type = 'E';
	ep->xfer_type = xfer_to_pipe[usb_endpoint_type(&urb->ep->desc)];
	ep->epnum = usb_urb_dir_in(urb) ? USB_DIR_IN : 0;
	ep->epnum |= usb_endpoint_num(&urb->ep->desc);
	ep->devnum = urb->dev->devnum;
	ep->busnum = urb->dev->bus->busnum;
	ep->id = (unsigned long) urb;
	ep->status = error;

	ep->flag_setup = '-';
	ep->flag_data = 'E';

	spin_unlock_irqrestore(&rp->b_lock, flags);

	wake_up(&rp->b_wait);
}

static int mon_bin_open(struct inode *inode, struct file *file)
{
	struct mon_bus *mbus;
	struct mon_reader_bin *rp;
	size_t size;
	int rc;

	mutex_lock(&mon_lock);
	if ((mbus = mon_bus_lookup(iminor(inode))) == NULL) {
		mutex_unlock(&mon_lock);
		return -ENODEV;
	}
	if (mbus != &mon_bus0 && mbus->u_bus == NULL) {
		printk(KERN_ERR TAG ": consistency error on open\n");
		mutex_unlock(&mon_lock);
		return -ENODEV;
	}

	rp = kzalloc(sizeof(struct mon_reader_bin), GFP_KERNEL);
	if (rp == NULL) {
		rc = -ENOMEM;
		goto err_alloc;
	}
	spin_lock_init(&rp->b_lock);
	init_waitqueue_head(&rp->b_wait);
	mutex_init(&rp->fetch_lock);

	rp->b_size = BUFF_DFL;

	size = sizeof(struct mon_pgmap) * (rp->b_size/CHUNK_SIZE);
	if ((rp->b_vec = kzalloc(size, GFP_KERNEL)) == NULL) {
		rc = -ENOMEM;
		goto err_allocvec;
	}

	if ((rc = mon_alloc_buff(rp->b_vec, rp->b_size/CHUNK_SIZE)) < 0)
		goto err_allocbuff;

	rp->r.m_bus = mbus;
	rp->r.r_data = rp;
	rp->r.rnf_submit = mon_bin_submit;
	rp->r.rnf_error = mon_bin_error;
	rp->r.rnf_complete = mon_bin_complete;

	mon_reader_add(mbus, &rp->r);

	file->private_data = rp;
	mutex_unlock(&mon_lock);
	return 0;

err_allocbuff:
	kfree(rp->b_vec);
err_allocvec:
	kfree(rp);
err_alloc:
	mutex_unlock(&mon_lock);
	return rc;
}

/*
 * Extract an event from buffer and copy it to user space.
 * Wait if there is no event ready.
 * Returns zero or error.
 */
static int mon_bin_get_event(struct file *file, struct mon_reader_bin *rp,
    struct mon_bin_hdr __user *hdr, void __user *data, unsigned int nbytes)
{
	unsigned long flags;
	struct mon_bin_hdr *ep;
	size_t step_len;
	unsigned int offset;
	int rc;

	mutex_lock(&rp->fetch_lock);

	if ((rc = mon_bin_wait_event(file, rp)) < 0) {
		mutex_unlock(&rp->fetch_lock);
		return rc;
	}

	ep = MON_OFF2HDR(rp, rp->b_out);

	if (copy_to_user(hdr, ep, sizeof(struct mon_bin_hdr))) {
		mutex_unlock(&rp->fetch_lock);
		return -EFAULT;
	}

	step_len = min(ep->len_cap, nbytes);
	if ((offset = rp->b_out + PKT_SIZE) >= rp->b_size) offset = 0;

	if (copy_from_buf(rp, offset, data, step_len)) {
		mutex_unlock(&rp->fetch_lock);
		return -EFAULT;
	}

	spin_lock_irqsave(&rp->b_lock, flags);
	mon_buff_area_free(rp, PKT_SIZE + ep->len_cap);
	spin_unlock_irqrestore(&rp->b_lock, flags);
	rp->b_read = 0;

	mutex_unlock(&rp->fetch_lock);
	return 0;
}

static int mon_bin_release(struct inode *inode, struct file *file)
{
	struct mon_reader_bin *rp = file->private_data;
	struct mon_bus* mbus = rp->r.m_bus;

	mutex_lock(&mon_lock);

	if (mbus->nreaders <= 0) {
		printk(KERN_ERR TAG ": consistency error on close\n");
		mutex_unlock(&mon_lock);
		return 0;
	}
	mon_reader_del(mbus, &rp->r);

	mon_free_buff(rp->b_vec, rp->b_size/CHUNK_SIZE);
	kfree(rp->b_vec);
	kfree(rp);

	mutex_unlock(&mon_lock);
	return 0;
}

static ssize_t mon_bin_read(struct file *file, char __user *buf,
    size_t nbytes, loff_t *ppos)
{
	struct mon_reader_bin *rp = file->private_data;
	unsigned long flags;
	struct mon_bin_hdr *ep;
	unsigned int offset;
	size_t step_len;
	char *ptr;
	ssize_t done = 0;
	int rc;

	mutex_lock(&rp->fetch_lock);

	if ((rc = mon_bin_wait_event(file, rp)) < 0) {
		mutex_unlock(&rp->fetch_lock);
		return rc;
	}

	ep = MON_OFF2HDR(rp, rp->b_out);

	if (rp->b_read < sizeof(struct mon_bin_hdr)) {
		step_len = min(nbytes, sizeof(struct mon_bin_hdr) - rp->b_read);
		ptr = ((char *)ep) + rp->b_read;
		if (step_len && copy_to_user(buf, ptr, step_len)) {
			mutex_unlock(&rp->fetch_lock);
			return -EFAULT;
		}
		nbytes -= step_len;
		buf += step_len;
		rp->b_read += step_len;
		done += step_len;
	}

	if (rp->b_read >= sizeof(struct mon_bin_hdr)) {
		step_len = min(nbytes, (size_t)ep->len_cap);
		offset = rp->b_out + PKT_SIZE;
		offset += rp->b_read - sizeof(struct mon_bin_hdr);
		if (offset >= rp->b_size)
			offset -= rp->b_size;
		if (copy_from_buf(rp, offset, buf, step_len)) {
			mutex_unlock(&rp->fetch_lock);
			return -EFAULT;
		}
		nbytes -= step_len;
		buf += step_len;
		rp->b_read += step_len;
		done += step_len;
	}

	/*
	 * Check if whole packet was read, and if so, jump to the next one.
	 */
	if (rp->b_read >= sizeof(struct mon_bin_hdr) + ep->len_cap) {
		spin_lock_irqsave(&rp->b_lock, flags);
		mon_buff_area_free(rp, PKT_SIZE + ep->len_cap);
		spin_unlock_irqrestore(&rp->b_lock, flags);
		rp->b_read = 0;
	}

	mutex_unlock(&rp->fetch_lock);
	return done;
}

/*
 * Remove at most nevents from chunked buffer.
 * Returns the number of removed events.
 */
static int mon_bin_flush(struct mon_reader_bin *rp, unsigned nevents)
{
	unsigned long flags;
	struct mon_bin_hdr *ep;
	int i;

	mutex_lock(&rp->fetch_lock);
	spin_lock_irqsave(&rp->b_lock, flags);
	for (i = 0; i < nevents; ++i) {
		if (MON_RING_EMPTY(rp))
			break;

		ep = MON_OFF2HDR(rp, rp->b_out);
		mon_buff_area_free(rp, PKT_SIZE + ep->len_cap);
	}
	spin_unlock_irqrestore(&rp->b_lock, flags);
	rp->b_read = 0;
	mutex_unlock(&rp->fetch_lock);
	return i;
}

/*
 * Fetch at most max event offsets into the buffer and put them into vec.
 * The events are usually freed later with mon_bin_flush.
 * Return the effective number of events fetched.
 */
static int mon_bin_fetch(struct file *file, struct mon_reader_bin *rp,
    u32 __user *vec, unsigned int max)
{
	unsigned int cur_out;
	unsigned int bytes, avail;
	unsigned int size;
	unsigned int nevents;
	struct mon_bin_hdr *ep;
	unsigned long flags;
	int rc;

	mutex_lock(&rp->fetch_lock);

	if ((rc = mon_bin_wait_event(file, rp)) < 0) {
		mutex_unlock(&rp->fetch_lock);
		return rc;
	}

	spin_lock_irqsave(&rp->b_lock, flags);
	avail = rp->b_cnt;
	spin_unlock_irqrestore(&rp->b_lock, flags);

	cur_out = rp->b_out;
	nevents = 0;
	bytes = 0;
	while (bytes < avail) {
		if (nevents >= max)
			break;

		ep = MON_OFF2HDR(rp, cur_out);
		if (put_user(cur_out, &vec[nevents])) {
			mutex_unlock(&rp->fetch_lock);
			return -EFAULT;
		}

		nevents++;
		size = ep->len_cap + PKT_SIZE;
		size = (size + PKT_ALIGN-1) & ~(PKT_ALIGN-1);
		if ((cur_out += size) >= rp->b_size)
			cur_out -= rp->b_size;
		bytes += size;
	}

	mutex_unlock(&rp->fetch_lock);
	return nevents;
}

/*
 * Count events. This is almost the same as the above mon_bin_fetch,
 * only we do not store offsets into user vector, and we have no limit.
 */
static int mon_bin_queued(struct mon_reader_bin *rp)
{
	unsigned int cur_out;
	unsigned int bytes, avail;
	unsigned int size;
	unsigned int nevents;
	struct mon_bin_hdr *ep;
	unsigned long flags;

	mutex_lock(&rp->fetch_lock);

	spin_lock_irqsave(&rp->b_lock, flags);
	avail = rp->b_cnt;
	spin_unlock_irqrestore(&rp->b_lock, flags);

	cur_out = rp->b_out;
	nevents = 0;
	bytes = 0;
	while (bytes < avail) {
		ep = MON_OFF2HDR(rp, cur_out);

		nevents++;
		size = ep->len_cap + PKT_SIZE;
		size = (size + PKT_ALIGN-1) & ~(PKT_ALIGN-1);
		if ((cur_out += size) >= rp->b_size)
			cur_out -= rp->b_size;
		bytes += size;
	}

	mutex_unlock(&rp->fetch_lock);
	return nevents;
}

/*
 */
static int mon_bin_ioctl(struct inode *inode, struct file *file,
    unsigned int cmd, unsigned long arg)
{
	struct mon_reader_bin *rp = file->private_data;
	// struct mon_bus* mbus = rp->r.m_bus;
	int ret = 0;
	struct mon_bin_hdr *ep;
	unsigned long flags;

	switch (cmd) {

	case MON_IOCQ_URB_LEN:
		/*
		 * N.B. This only returns the size of data, without the header.
		 */
		spin_lock_irqsave(&rp->b_lock, flags);
		if (!MON_RING_EMPTY(rp)) {
			ep = MON_OFF2HDR(rp, rp->b_out);
			ret = ep->len_cap;
		}
		spin_unlock_irqrestore(&rp->b_lock, flags);
		break;

	case MON_IOCQ_RING_SIZE:
		ret = rp->b_size;
		break;

	case MON_IOCT_RING_SIZE:
		/*
		 * Changing the buffer size will flush it's contents; the new
		 * buffer is allocated before releasing the old one to be sure
		 * the device will stay functional also in case of memory
		 * pressure.
		 */
		{
		int size;
		struct mon_pgmap *vec;

		if (arg < BUFF_MIN || arg > BUFF_MAX)
			return -EINVAL;

		size = CHUNK_ALIGN(arg);
		if ((vec = kzalloc(sizeof(struct mon_pgmap) * (size/CHUNK_SIZE),
		    GFP_KERNEL)) == NULL) {
			ret = -ENOMEM;
			break;
		}

		ret = mon_alloc_buff(vec, size/CHUNK_SIZE);
		if (ret < 0) {
			kfree(vec);
			break;
		}

		mutex_lock(&rp->fetch_lock);
		spin_lock_irqsave(&rp->b_lock, flags);
		mon_free_buff(rp->b_vec, size/CHUNK_SIZE);
		kfree(rp->b_vec);
		rp->b_vec  = vec;
		rp->b_size = size;
		rp->b_read = rp->b_in = rp->b_out = rp->b_cnt = 0;
		rp->cnt_lost = 0;
		spin_unlock_irqrestore(&rp->b_lock, flags);
		mutex_unlock(&rp->fetch_lock);
		}
		break;

	case MON_IOCH_MFLUSH:
		ret = mon_bin_flush(rp, arg);
		break;

	case MON_IOCX_GET:
		{
		struct mon_bin_get getb;

		if (copy_from_user(&getb, (void __user *)arg,
					    sizeof(struct mon_bin_get)))
			return -EFAULT;

		if (getb.alloc > 0x10000000)	/* Want to cast to u32 */
			return -EINVAL;
		ret = mon_bin_get_event(file, rp,
			  getb.hdr, getb.data, (unsigned int)getb.alloc);
		}
		break;

#ifdef CONFIG_COMPAT
	case MON_IOCX_GET32: {
		struct mon_bin_get32 getb;

		if (copy_from_user(&getb, (void __user *)arg,
					    sizeof(struct mon_bin_get32)))
			return -EFAULT;

		ret = mon_bin_get_event(file, rp,
		    compat_ptr(getb.hdr32), compat_ptr(getb.data32),
		    getb.alloc32);
		}
		break;
#endif

	case MON_IOCX_MFETCH:
		{
		struct mon_bin_mfetch mfetch;
		struct mon_bin_mfetch __user *uptr;

		uptr = (struct mon_bin_mfetch __user *)arg;

		if (copy_from_user(&mfetch, uptr, sizeof(mfetch)))
			return -EFAULT;

		if (mfetch.nflush) {
			ret = mon_bin_flush(rp, mfetch.nflush);
			if (ret < 0)
				return ret;
			if (put_user(ret, &uptr->nflush))
				return -EFAULT;
		}
		ret = mon_bin_fetch(file, rp, mfetch.offvec, mfetch.nfetch);
		if (ret < 0)
			return ret;
		if (put_user(ret, &uptr->nfetch))
			return -EFAULT;
		ret = 0;
		}
		break;

#ifdef CONFIG_COMPAT
	case MON_IOCX_MFETCH32:
		{
		struct mon_bin_mfetch32 mfetch;
		struct mon_bin_mfetch32 __user *uptr;

		uptr = (struct mon_bin_mfetch32 __user *) compat_ptr(arg);

		if (copy_from_user(&mfetch, uptr, sizeof(mfetch)))
			return -EFAULT;

		if (mfetch.nflush32) {
			ret = mon_bin_flush(rp, mfetch.nflush32);
			if (ret < 0)
				return ret;
			if (put_user(ret, &uptr->nflush32))
				return -EFAULT;
		}
		ret = mon_bin_fetch(file, rp, compat_ptr(mfetch.offvec32),
		    mfetch.nfetch32);
		if (ret < 0)
			return ret;
		if (put_user(ret, &uptr->nfetch32))
			return -EFAULT;
		ret = 0;
		}
		break;
#endif

	case MON_IOCG_STATS: {
		struct mon_bin_stats __user *sp;
		unsigned int nevents;
		unsigned int ndropped;

		spin_lock_irqsave(&rp->b_lock, flags);
		ndropped = rp->cnt_lost;
		rp->cnt_lost = 0;
		spin_unlock_irqrestore(&rp->b_lock, flags);
		nevents = mon_bin_queued(rp);

		sp = (struct mon_bin_stats __user *)arg;
		if (put_user(rp->cnt_lost, &sp->dropped))
			return -EFAULT;
		if (put_user(nevents, &sp->queued))
			return -EFAULT;

		}
		break;

	default:
		return -ENOTTY;
	}

	return ret;
}

static unsigned int
mon_bin_poll(struct file *file, struct poll_table_struct *wait)
{
	struct mon_reader_bin *rp = file->private_data;
	unsigned int mask = 0;
	unsigned long flags;

	if (file->f_mode & FMODE_READ)
		poll_wait(file, &rp->b_wait, wait);

	spin_lock_irqsave(&rp->b_lock, flags);
	if (!MON_RING_EMPTY(rp))
		mask |= POLLIN | POLLRDNORM;    /* readable */
	spin_unlock_irqrestore(&rp->b_lock, flags);
	return mask;
}

#if 0

/*
 * open and close: just keep track of how many times the device is
 * mapped, to use the proper memory allocation function.
 */
static void mon_bin_vma_open(struct vm_area_struct *vma)
{
	struct mon_reader_bin *rp = vma->vm_private_data;
	rp->mmap_active++;
}

static void mon_bin_vma_close(struct vm_area_struct *vma)
{
	struct mon_reader_bin *rp = vma->vm_private_data;
	rp->mmap_active--;
}

/*
 * Map ring pages to user space.
 */
struct page *mon_bin_vma_nopage(struct vm_area_struct *vma,
                                unsigned long address, int *type)
{
	struct mon_reader_bin *rp = vma->vm_private_data;
	unsigned long offset, chunk_idx;
	struct page *pageptr;

	offset = (address - vma->vm_start) + (vma->vm_pgoff << PAGE_SHIFT);
	if (offset >= rp->b_size)
		return NOPAGE_SIGBUS;
	chunk_idx = offset / CHUNK_SIZE;
	pageptr = rp->b_vec[chunk_idx].pg;
	get_page(pageptr);
	if (type)
		*type = VM_FAULT_MINOR;
	return pageptr;
}

struct vm_operations_struct mon_bin_vm_ops = {
	.open =     mon_bin_vma_open,
	.close =    mon_bin_vma_close,
	.nopage =   mon_bin_vma_nopage,
};

int mon_bin_mmap(struct file *filp, struct vm_area_struct *vma)
{
	/* don't do anything here: "nopage" will set up page table entries */
	vma->vm_ops = &mon_bin_vm_ops;
	vma->vm_flags |= VM_RESERVED;
	vma->vm_private_data = filp->private_data;
	mon_bin_vma_open(vma);
	return 0;
}

#endif  /*  0  */

static struct file_operations mon_fops_binary = {
	.owner =	THIS_MODULE,
	.open =		mon_bin_open,
	.llseek =	no_llseek,
	.read =		mon_bin_read,
	/* .write =	mon_text_write, */
	.poll =		mon_bin_poll,
	.ioctl =	mon_bin_ioctl,
	.release =	mon_bin_release,
};

static int mon_bin_wait_event(struct file *file, struct mon_reader_bin *rp)
{
	DECLARE_WAITQUEUE(waita, current);
	unsigned long flags;

	add_wait_queue(&rp->b_wait, &waita);
	set_current_state(TASK_INTERRUPTIBLE);

	spin_lock_irqsave(&rp->b_lock, flags);
	while (MON_RING_EMPTY(rp)) {
		spin_unlock_irqrestore(&rp->b_lock, flags);

		if (file->f_flags & O_NONBLOCK) {
			set_current_state(TASK_RUNNING);
			remove_wait_queue(&rp->b_wait, &waita);
			return -EWOULDBLOCK; /* Same as EAGAIN in Linux */
		}
		schedule();
		if (signal_pending(current)) {
			remove_wait_queue(&rp->b_wait, &waita);
			return -EINTR;
		}
		set_current_state(TASK_INTERRUPTIBLE);

		spin_lock_irqsave(&rp->b_lock, flags);
	}
	spin_unlock_irqrestore(&rp->b_lock, flags);

	set_current_state(TASK_RUNNING);
	remove_wait_queue(&rp->b_wait, &waita);
	return 0;
}

static int mon_alloc_buff(struct mon_pgmap *map, int npages)
{
	int n;
	unsigned long vaddr;

	for (n = 0; n < npages; n++) {
		vaddr = get_zeroed_page(GFP_KERNEL);
		if (vaddr == 0) {
			while (n-- != 0)
				free_page((unsigned long) map[n].ptr);
			return -ENOMEM;
		}
		map[n].ptr = (unsigned char *) vaddr;
		map[n].pg = virt_to_page(vaddr);
	}
	return 0;
}

static void mon_free_buff(struct mon_pgmap *map, int npages)
{
	int n;

	for (n = 0; n < npages; n++)
		free_page((unsigned long) map[n].ptr);
}

int mon_bin_add(struct mon_bus *mbus, const struct usb_bus *ubus)
{
	struct device *dev;
	unsigned minor = ubus? ubus->busnum: 0;

	if (minor >= MON_BIN_MAX_MINOR)
		return 0;

	dev = device_create(mon_bin_class, ubus? ubus->controller: NULL,
			MKDEV(MAJOR(mon_bin_dev0), minor), "usbmon%d", minor);
	if (IS_ERR(dev))
		return 0;

	mbus->classdev = dev;
	return 1;
}

void mon_bin_del(struct mon_bus *mbus)
{
	device_destroy(mon_bin_class, mbus->classdev->devt);
}

int __init mon_bin_init(void)
{
	int rc;

	mon_bin_class = class_create(THIS_MODULE, "usbmon");
	if (IS_ERR(mon_bin_class)) {
		rc = PTR_ERR(mon_bin_class);
		goto err_class;
	}

	rc = alloc_chrdev_region(&mon_bin_dev0, 0, MON_BIN_MAX_MINOR, "usbmon");
	if (rc < 0)
		goto err_dev;

	cdev_init(&mon_bin_cdev, &mon_fops_binary);
	mon_bin_cdev.owner = THIS_MODULE;

	rc = cdev_add(&mon_bin_cdev, mon_bin_dev0, MON_BIN_MAX_MINOR);
	if (rc < 0)
		goto err_add;

	return 0;

err_add:
	unregister_chrdev_region(mon_bin_dev0, MON_BIN_MAX_MINOR);
err_dev:
	class_destroy(mon_bin_class);
err_class:
	return rc;
}

void mon_bin_exit(void)
{
	cdev_del(&mon_bin_cdev);
	unregister_chrdev_region(mon_bin_dev0, MON_BIN_MAX_MINOR);
	class_destroy(mon_bin_class);
}