summaryrefslogtreecommitdiffstats
path: root/sys/geom/geom_ccd.c
blob: 672ffe2fbef166c382fdfe5cc682567afa01249f (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
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
/* $FreeBSD$ */

/*	$NetBSD: ccd.c,v 1.22 1995/12/08 19:13:26 thorpej Exp $	*/

/*
 * Copyright (c) 1995 Jason R. Thorpe.
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *	This product includes software developed for the NetBSD Project
 *	by Jason R. Thorpe.
 * 4. The name of the author may not be used to endorse or promote products
 *    derived from this software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
 * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 */

/*
 * Copyright (c) 1988 University of Utah.
 * Copyright (c) 1990, 1993
 *	The Regents of the University of California.  All rights reserved.
 *
 * This code is derived from software contributed to Berkeley by
 * the Systems Programming Group of the University of Utah Computer
 * Science Department.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *	This product includes software developed by the University of
 *	California, Berkeley and its contributors.
 * 4. Neither the name of the University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 *
 * from: Utah $Hdr: cd.c 1.6 90/11/28$
 *
 *	@(#)cd.c	8.2 (Berkeley) 11/16/93
 */

/*
 * "Concatenated" disk driver.
 *
 * Dynamic configuration and disklabel support by:
 *	Jason R. Thorpe <thorpej@nas.nasa.gov>
 *	Numerical Aerodynamic Simulation Facility
 *	Mail Stop 258-6
 *	NASA Ames Research Center
 *	Moffett Field, CA 94035
 */

#include "ccd.h"

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/module.h>
#include <sys/proc.h>
#include <sys/buf.h>
#include <sys/malloc.h>
#include <sys/namei.h>
#include <sys/conf.h>
#include <sys/stat.h>
#include <sys/sysctl.h>
#include <sys/disklabel.h>
#include <ufs/ffs/fs.h> 
#include <sys/devicestat.h>
#include <sys/fcntl.h>
#include <sys/vnode.h>

#include <sys/ccdvar.h>

#include <vm/vm_zone.h>

#if defined(CCDDEBUG) && !defined(DEBUG)
#define DEBUG
#endif

#ifdef DEBUG
#define CCDB_FOLLOW	0x01
#define CCDB_INIT	0x02
#define CCDB_IO		0x04
#define CCDB_LABEL	0x08
#define CCDB_VNODE	0x10
static int ccddebug = CCDB_FOLLOW | CCDB_INIT | CCDB_IO | CCDB_LABEL |
    CCDB_VNODE;
SYSCTL_INT(_debug, OID_AUTO, ccddebug, CTLFLAG_RW, &ccddebug, 0, "");
#undef DEBUG
#endif

#define	ccdunit(x)	dkunit(x)
#define ccdpart(x)	dkpart(x)

/*
   This is how mirroring works (only writes are special):

   When initiating a write, ccdbuffer() returns two "struct ccdbuf *"s
   linked together by the cb_mirror field.  "cb_pflags &
   CCDPF_MIRROR_DONE" is set to 0 on both of them.

   When a component returns to ccdiodone(), it checks if "cb_pflags &
   CCDPF_MIRROR_DONE" is set or not.  If not, it sets the partner's
   flag and returns.  If it is, it means its partner has already
   returned, so it will go to the regular cleanup.

 */

struct ccdbuf {
	struct buf	cb_buf;		/* new I/O buf */
	struct buf	*cb_obp;	/* ptr. to original I/O buf */
	struct ccdbuf	*cb_freenext;	/* free list link */
	int		cb_unit;	/* target unit */
	int		cb_comp;	/* target component */
	int		cb_pflags;	/* mirror/parity status flag */
	struct ccdbuf	*cb_mirror;	/* mirror counterpart */
};

/* bits in cb_pflags */
#define CCDPF_MIRROR_DONE 1	/* if set, mirror counterpart is done */

#define CCDLABELDEV(dev)	\
	(makedev(major((dev)), dkmakeminor(ccdunit((dev)), 0, RAW_PART)))

static d_open_t ccdopen;
static d_close_t ccdclose;
static d_strategy_t ccdstrategy;
static d_ioctl_t ccdioctl;
static d_dump_t ccddump;
static d_psize_t ccdsize;

#define NCCDFREEHIWAT	16

#define CDEV_MAJOR 74
#define BDEV_MAJOR 21

static struct cdevsw ccd_cdevsw = {
	/* open */	ccdopen,
	/* close */	ccdclose,
	/* read */	physread,
	/* write */	physwrite,
	/* ioctl */	ccdioctl,
	/* poll */	nopoll,
	/* mmap */	nommap,
	/* strategy */	ccdstrategy,
	/* name */	"ccd",
	/* maj */	CDEV_MAJOR,
	/* dump */	ccddump,
	/* psize */	ccdsize,
	/* flags */	D_DISK,
	/* bmaj */	BDEV_MAJOR
};

/* called during module initialization */
static	void ccdattach __P((void));
static	int ccd_modevent __P((module_t, int, void *));

/* called by biodone() at interrupt time */
static	void ccdiodone __P((struct ccdbuf *cbp));

static	void ccdstart __P((struct ccd_softc *, struct buf *));
static	void ccdinterleave __P((struct ccd_softc *, int));
static	void ccdintr __P((struct ccd_softc *, struct buf *));
static	int ccdinit __P((struct ccddevice *, char **, struct proc *));
static	int ccdlookup __P((char *, struct proc *p, struct vnode **));
static	void ccdbuffer __P((struct ccdbuf **ret, struct ccd_softc *,
		struct buf *, daddr_t, caddr_t, long));
static	void ccdgetdisklabel __P((dev_t));
static	void ccdmakedisklabel __P((struct ccd_softc *));
static	int ccdlock __P((struct ccd_softc *));
static	void ccdunlock __P((struct ccd_softc *));

#ifdef DEBUG
static	void printiinfo __P((struct ccdiinfo *));
#endif

/* Non-private for the benefit of libkvm. */
struct	ccd_softc *ccd_softc;
struct	ccddevice *ccddevs;
struct	ccdbuf *ccdfreebufs;
static	int numccdfreebufs;
static	int numccd = 0;

/*
 * getccdbuf() -	Allocate and zero a ccd buffer.
 *
 *	This routine is called at splbio().
 */

static __inline
struct ccdbuf *
getccdbuf(struct ccdbuf *cpy)
{
	struct ccdbuf *cbp;

	/*
	 * Allocate from freelist or malloc as necessary
	 */
	if ((cbp = ccdfreebufs) != NULL) {
		ccdfreebufs = cbp->cb_freenext;
		--numccdfreebufs;
	} else {
		cbp = malloc(sizeof(struct ccdbuf), M_DEVBUF, M_WAITOK);
	}

	/*
	 * Used by mirroring code
	 */
	if (cpy)
		bcopy(cpy, cbp, sizeof(struct ccdbuf));
	else
		bzero(cbp, sizeof(struct ccdbuf));

	/*
	 * independant struct buf initialization
	 */
	LIST_INIT(&cbp->cb_buf.b_dep);
	BUF_LOCKINIT(&cbp->cb_buf);
	BUF_LOCK(&cbp->cb_buf, LK_EXCLUSIVE);
	BUF_KERNPROC(&cbp->cb_buf);

	return(cbp);
}

/*
 * putccdbuf() -	Free a ccd buffer.
 *
 *	This routine is called at splbio().
 */

static __inline
void
putccdbuf(struct ccdbuf *cbp)
{
	BUF_UNLOCK(&cbp->cb_buf);
	BUF_LOCKFREE(&cbp->cb_buf);

	if (numccdfreebufs < NCCDFREEHIWAT) {
		cbp->cb_freenext = ccdfreebufs;
		ccdfreebufs = cbp;
		++numccdfreebufs;
	} else {
		free((caddr_t)cbp, M_DEVBUF);
	}
}


/*
 * Number of blocks to untouched in front of a component partition.
 * This is to avoid violating its disklabel area when it starts at the
 * beginning of the slice.
 */
#if !defined(CCD_OFFSET)
#define CCD_OFFSET 16
#endif

/*
 * Called by main() during pseudo-device attachment.  All we need
 * to do is allocate enough space for devices to be configured later, and
 * add devsw entries.
 */
static void
ccdattach()
{
	int i;
	int num = NCCD;

	if (num > 1)
		printf("ccd0-%d: Concatenated disk drivers\n", num-1);
	else
		printf("ccd0: Concatenated disk driver\n");

	ccd_softc = (struct ccd_softc *)malloc(num * sizeof(struct ccd_softc),
	    M_DEVBUF, M_NOWAIT);
	ccddevs = (struct ccddevice *)malloc(num * sizeof(struct ccddevice),
	    M_DEVBUF, M_NOWAIT);
	if ((ccd_softc == NULL) || (ccddevs == NULL)) {
		printf("WARNING: no memory for concatenated disks\n");
		if (ccd_softc != NULL)
			free(ccd_softc, M_DEVBUF);
		if (ccddevs != NULL)
			free(ccddevs, M_DEVBUF);
		return;
	}
	numccd = num;
	bzero(ccd_softc, num * sizeof(struct ccd_softc));
	bzero(ccddevs, num * sizeof(struct ccddevice));

	cdevsw_add(&ccd_cdevsw);
	/* XXX: is this necessary? */
	for (i = 0; i < numccd; ++i)
		ccddevs[i].ccd_dk = -1;
}

static int
ccd_modevent(mod, type, data)
	module_t mod;
	int type;
	void *data;
{
	int error = 0;

	switch (type) {
	case MOD_LOAD:
		ccdattach();
		break;

	case MOD_UNLOAD:
		printf("ccd0: Unload not supported!\n");
		error = EOPNOTSUPP;
		break;

	default:	/* MOD_SHUTDOWN etc */
		break;
	}
	return (error);
}

DEV_MODULE(ccd, ccd_modevent, NULL);

static int
ccdinit(ccd, cpaths, p)
	struct ccddevice *ccd;
	char **cpaths;
	struct proc *p;
{
	struct ccd_softc *cs = &ccd_softc[ccd->ccd_unit];
	struct ccdcinfo *ci = NULL;	/* XXX */
	size_t size;
	int ix;
	struct vnode *vp;
	size_t minsize;
	int maxsecsize;
	struct partinfo dpart;
	struct ccdgeom *ccg = &cs->sc_geom;
	char tmppath[MAXPATHLEN];
	int error = 0;

#ifdef DEBUG
	if (ccddebug & (CCDB_FOLLOW|CCDB_INIT))
		printf("ccdinit: unit %d\n", ccd->ccd_unit);
#endif

	cs->sc_size = 0;
	cs->sc_ileave = ccd->ccd_interleave;
	cs->sc_nccdisks = ccd->ccd_ndev;

	/* Allocate space for the component info. */
	cs->sc_cinfo = malloc(cs->sc_nccdisks * sizeof(struct ccdcinfo),
	    M_DEVBUF, M_WAITOK);

	/*
	 * Verify that each component piece exists and record
	 * relevant information about it.
	 */
	maxsecsize = 0;
	minsize = 0;
	for (ix = 0; ix < cs->sc_nccdisks; ix++) {
		vp = ccd->ccd_vpp[ix];
		ci = &cs->sc_cinfo[ix];
		ci->ci_vp = vp;

		/*
		 * Copy in the pathname of the component.
		 */
		bzero(tmppath, sizeof(tmppath));	/* sanity */
		if ((error = copyinstr(cpaths[ix], tmppath,
		    MAXPATHLEN, &ci->ci_pathlen)) != 0) {
#ifdef DEBUG
			if (ccddebug & (CCDB_FOLLOW|CCDB_INIT))
				printf("ccd%d: can't copy path, error = %d\n",
				    ccd->ccd_unit, error);
#endif
			goto fail;
		}
		ci->ci_path = malloc(ci->ci_pathlen, M_DEVBUF, M_WAITOK);
		bcopy(tmppath, ci->ci_path, ci->ci_pathlen);

		ci->ci_dev = vn_todev(vp);

		/*
		 * Get partition information for the component.
		 */
		if ((error = VOP_IOCTL(vp, DIOCGPART, (caddr_t)&dpart,
		    FREAD, p->p_ucred, p)) != 0) {
#ifdef DEBUG
			if (ccddebug & (CCDB_FOLLOW|CCDB_INIT))
				 printf("ccd%d: %s: ioctl failed, error = %d\n",
				     ccd->ccd_unit, ci->ci_path, error);
#endif
			goto fail;
		}
		if (dpart.part->p_fstype == FS_BSDFFS) {
			maxsecsize =
			    ((dpart.disklab->d_secsize > maxsecsize) ?
			    dpart.disklab->d_secsize : maxsecsize);
			size = dpart.part->p_size - CCD_OFFSET;
		} else {
#ifdef DEBUG
			if (ccddebug & (CCDB_FOLLOW|CCDB_INIT))
				printf("ccd%d: %s: incorrect partition type\n",
				    ccd->ccd_unit, ci->ci_path);
#endif
			error = EFTYPE;
			goto fail;
		}

		/*
		 * Calculate the size, truncating to an interleave
		 * boundary if necessary.
		 */

		if (cs->sc_ileave > 1)
			size -= size % cs->sc_ileave;

		if (size == 0) {
#ifdef DEBUG
			if (ccddebug & (CCDB_FOLLOW|CCDB_INIT))
				printf("ccd%d: %s: size == 0\n",
				    ccd->ccd_unit, ci->ci_path);
#endif
			error = ENODEV;
			goto fail;
		}

		if (minsize == 0 || size < minsize)
			minsize = size;
		ci->ci_size = size;
		cs->sc_size += size;
	}

	/*
	 * Don't allow the interleave to be smaller than
	 * the biggest component sector.
	 */
	if ((cs->sc_ileave > 0) &&
	    (cs->sc_ileave < (maxsecsize / DEV_BSIZE))) {
#ifdef DEBUG
		if (ccddebug & (CCDB_FOLLOW|CCDB_INIT))
			printf("ccd%d: interleave must be at least %d\n",
			    ccd->ccd_unit, (maxsecsize / DEV_BSIZE));
#endif
		error = EINVAL;
		goto fail;
	}

	/*
	 * If uniform interleave is desired set all sizes to that of
	 * the smallest component.  This will guarentee that a single
	 * interleave table is generated.
	 *
	 * Lost space must be taken into account when calculating the
	 * overall size.  Half the space is lost when CCDF_MIRROR is
	 * specified.  One disk is lost when CCDF_PARITY is specified.
	 */
	if (ccd->ccd_flags & CCDF_UNIFORM) {
		for (ci = cs->sc_cinfo;
		     ci < &cs->sc_cinfo[cs->sc_nccdisks]; ci++) {
			ci->ci_size = minsize;
		}
		if (ccd->ccd_flags & CCDF_MIRROR) {
			/*
			 * Check to see if an even number of components
			 * have been specified.  The interleave must also
			 * be non-zero in order for us to be able to 
			 * guarentee the topology.
			 */
			if (cs->sc_nccdisks % 2) {
				printf("ccd%d: mirroring requires an even number of disks\n", ccd->ccd_unit );
				error = EINVAL;
				goto fail;
			}
			if (cs->sc_ileave == 0) {
				printf("ccd%d: an interleave must be specified when mirroring\n", ccd->ccd_unit);
				error = EINVAL;
				goto fail;
			}
			cs->sc_size = (cs->sc_nccdisks/2) * minsize;
		} else if (ccd->ccd_flags & CCDF_PARITY) {
			cs->sc_size = (cs->sc_nccdisks-1) * minsize;
		} else {
			if (cs->sc_ileave == 0) {
				printf("ccd%d: an interleave must be specified when using parity\n", ccd->ccd_unit);
				error = EINVAL;
				goto fail;
			}
			cs->sc_size = cs->sc_nccdisks * minsize;
		}
	}

	/*
	 * Construct the interleave table.
	 */
	ccdinterleave(cs, ccd->ccd_unit);

	/*
	 * Create pseudo-geometry based on 1MB cylinders.  It's
	 * pretty close.
	 */
	ccg->ccg_secsize = maxsecsize;
	ccg->ccg_ntracks = 1;
	ccg->ccg_nsectors = 1024 * 1024 / ccg->ccg_secsize;
	ccg->ccg_ncylinders = cs->sc_size / ccg->ccg_nsectors;

	/*
	 * Add an devstat entry for this device.
	 */
	devstat_add_entry(&cs->device_stats, "ccd", ccd->ccd_unit,
			  ccg->ccg_secsize, DEVSTAT_ALL_SUPPORTED,
			  DEVSTAT_TYPE_STORARRAY |DEVSTAT_TYPE_IF_OTHER,
			  DEVSTAT_PRIORITY_ARRAY);

	cs->sc_flags |= CCDF_INITED;
	cs->sc_cflags = ccd->ccd_flags;	/* So we can find out later... */
	cs->sc_unit = ccd->ccd_unit;
	return (0);
fail:
	while (ci > cs->sc_cinfo) {
		ci--;
		free(ci->ci_path, M_DEVBUF);
	}
	free(cs->sc_cinfo, M_DEVBUF);
	return (error);
}

static void
ccdinterleave(cs, unit)
	struct ccd_softc *cs;
	int unit;
{
	struct ccdcinfo *ci, *smallci;
	struct ccdiinfo *ii;
	daddr_t bn, lbn;
	int ix;
	u_long size;

#ifdef DEBUG
	if (ccddebug & CCDB_INIT)
		printf("ccdinterleave(%x): ileave %d\n", cs, cs->sc_ileave);
#endif

	/*
	 * Allocate an interleave table.  The worst case occurs when each
	 * of N disks is of a different size, resulting in N interleave
	 * tables.
	 *
	 * Chances are this is too big, but we don't care.
	 */
	size = (cs->sc_nccdisks + 1) * sizeof(struct ccdiinfo);
	cs->sc_itable = (struct ccdiinfo *)malloc(size, M_DEVBUF, M_WAITOK);
	bzero((caddr_t)cs->sc_itable, size);

	/*
	 * Trivial case: no interleave (actually interleave of disk size).
	 * Each table entry represents a single component in its entirety.
	 *
	 * An interleave of 0 may not be used with a mirror or parity setup.
	 */
	if (cs->sc_ileave == 0) {
		bn = 0;
		ii = cs->sc_itable;

		for (ix = 0; ix < cs->sc_nccdisks; ix++) {
			/* Allocate space for ii_index. */
			ii->ii_index = malloc(sizeof(int), M_DEVBUF, M_WAITOK);
			ii->ii_ndisk = 1;
			ii->ii_startblk = bn;
			ii->ii_startoff = 0;
			ii->ii_index[0] = ix;
			bn += cs->sc_cinfo[ix].ci_size;
			ii++;
		}
		ii->ii_ndisk = 0;
#ifdef DEBUG
		if (ccddebug & CCDB_INIT)
			printiinfo(cs->sc_itable);
#endif
		return;
	}

	/*
	 * The following isn't fast or pretty; it doesn't have to be.
	 */
	size = 0;
	bn = lbn = 0;
	for (ii = cs->sc_itable; ; ii++) {
		/*
		 * Allocate space for ii_index.  We might allocate more then
		 * we use.
		 */
		ii->ii_index = malloc((sizeof(int) * cs->sc_nccdisks),
		    M_DEVBUF, M_WAITOK);

		/*
		 * Locate the smallest of the remaining components
		 */
		smallci = NULL;
		for (ci = cs->sc_cinfo; ci < &cs->sc_cinfo[cs->sc_nccdisks]; 
		    ci++) {
			if (ci->ci_size > size &&
			    (smallci == NULL ||
			     ci->ci_size < smallci->ci_size)) {
				smallci = ci;
			}
		}

		/*
		 * Nobody left, all done
		 */
		if (smallci == NULL) {
			ii->ii_ndisk = 0;
			break;
		}

		/*
		 * Record starting logical block using an sc_ileave blocksize.
		 */
		ii->ii_startblk = bn / cs->sc_ileave;

		/*
		 * Record starting comopnent block using an sc_ileave 
		 * blocksize.  This value is relative to the beginning of
		 * a component disk.
		 */
		ii->ii_startoff = lbn;

		/*
		 * Determine how many disks take part in this interleave
		 * and record their indices.
		 */
		ix = 0;
		for (ci = cs->sc_cinfo; 
		    ci < &cs->sc_cinfo[cs->sc_nccdisks]; ci++) {
			if (ci->ci_size >= smallci->ci_size) {
				ii->ii_index[ix++] = ci - cs->sc_cinfo;
			}
		}
		ii->ii_ndisk = ix;
		bn += ix * (smallci->ci_size - size);
		lbn = smallci->ci_size / cs->sc_ileave;
		size = smallci->ci_size;
	}
#ifdef DEBUG
	if (ccddebug & CCDB_INIT)
		printiinfo(cs->sc_itable);
#endif
}

/* ARGSUSED */
static int
ccdopen(dev, flags, fmt, p)
	dev_t dev;
	int flags, fmt;
	struct proc *p;
{
	int unit = ccdunit(dev);
	struct ccd_softc *cs;
	struct disklabel *lp;
	int error = 0, part, pmask;

#ifdef DEBUG
	if (ccddebug & CCDB_FOLLOW)
		printf("ccdopen(%x, %x)\n", dev, flags);
#endif
	if (unit >= numccd)
		return (ENXIO);
	cs = &ccd_softc[unit];

	if ((error = ccdlock(cs)) != 0)
		return (error);

	lp = &cs->sc_label;

	part = ccdpart(dev);
	pmask = (1 << part);

	/*
	 * If we're initialized, check to see if there are any other
	 * open partitions.  If not, then it's safe to update
	 * the in-core disklabel.
	 */
	if ((cs->sc_flags & CCDF_INITED) && (cs->sc_openmask == 0))
		ccdgetdisklabel(dev);

	/* Check that the partition exists. */
	if (part != RAW_PART && ((part >= lp->d_npartitions) ||
	    (lp->d_partitions[part].p_fstype == FS_UNUSED))) {
		error = ENXIO;
		goto done;
	}

	cs->sc_openmask |= pmask;
 done:
	ccdunlock(cs);
	return (0);
}

/* ARGSUSED */
static int
ccdclose(dev, flags, fmt, p)
	dev_t dev;
	int flags, fmt;
	struct proc *p;
{
	int unit = ccdunit(dev);
	struct ccd_softc *cs;
	int error = 0, part;

#ifdef DEBUG
	if (ccddebug & CCDB_FOLLOW)
		printf("ccdclose(%x, %x)\n", dev, flags);
#endif

	if (unit >= numccd)
		return (ENXIO);
	cs = &ccd_softc[unit];

	if ((error = ccdlock(cs)) != 0)
		return (error);

	part = ccdpart(dev);

	/* ...that much closer to allowing unconfiguration... */
	cs->sc_openmask &= ~(1 << part);
	ccdunlock(cs);
	return (0);
}

static void
ccdstrategy(bp)
	struct buf *bp;
{
	int unit = ccdunit(bp->b_dev);
	struct ccd_softc *cs = &ccd_softc[unit];
	int s;
	int wlabel;
	struct disklabel *lp;

#ifdef DEBUG
	if (ccddebug & CCDB_FOLLOW)
		printf("ccdstrategy(%x): unit %d\n", bp, unit);
#endif
	if ((cs->sc_flags & CCDF_INITED) == 0) {
		bp->b_error = ENXIO;
		bp->b_flags |= B_ERROR;
		goto done;
	}

	/* If it's a nil transfer, wake up the top half now. */
	if (bp->b_bcount == 0)
		goto done;

	lp = &cs->sc_label;

	/*
	 * Do bounds checking and adjust transfer.  If there's an
	 * error, the bounds check will flag that for us.
	 */
	wlabel = cs->sc_flags & (CCDF_WLABEL|CCDF_LABELLING);
	if (ccdpart(bp->b_dev) != RAW_PART) {
		if (bounds_check_with_label(bp, lp, wlabel) <= 0)
			goto done;
	} else {
		int pbn;        /* in sc_secsize chunks */
		long sz;        /* in sc_secsize chunks */

		pbn = bp->b_blkno / (cs->sc_geom.ccg_secsize / DEV_BSIZE);
		sz = howmany(bp->b_bcount, cs->sc_geom.ccg_secsize);

		/*
		 * If out of bounds return an error. If at the EOF point,
		 * simply read or write less.
		 */

		if (pbn < 0 || pbn >= cs->sc_size) {
			bp->b_resid = bp->b_bcount;
			if (pbn != cs->sc_size) {
				bp->b_error = EINVAL;
				bp->b_flags |= B_ERROR | B_INVAL;
			}
			goto done;
		}

		/*
		 * If the request crosses EOF, truncate the request.
		 */
		if (pbn + sz > cs->sc_size) {
			bp->b_bcount = (cs->sc_size - pbn) * 
			    cs->sc_geom.ccg_secsize;
		}
	}

	bp->b_resid = bp->b_bcount;

	/*
	 * "Start" the unit.
	 */
	s = splbio();
	ccdstart(cs, bp);
	splx(s);
	return;
done:
	biodone(bp);
}

static void
ccdstart(cs, bp)
	struct ccd_softc *cs;
	struct buf *bp;
{
	long bcount, rcount;
	struct ccdbuf *cbp[4];
	/* XXX! : 2 reads and 2 writes for RAID 4/5 */
	caddr_t addr;
	daddr_t bn;
	struct partition *pp;

#ifdef DEBUG
	if (ccddebug & CCDB_FOLLOW)
		printf("ccdstart(%x, %x)\n", cs, bp);
#endif

	/* Record the transaction start  */
	devstat_start_transaction(&cs->device_stats);

	/*
	 * Translate the partition-relative block number to an absolute.
	 */
	bn = bp->b_blkno;
	if (ccdpart(bp->b_dev) != RAW_PART) {
		pp = &cs->sc_label.d_partitions[ccdpart(bp->b_dev)];
		bn += pp->p_offset;
	}

	/*
	 * Allocate component buffers and fire off the requests
	 */
	addr = bp->b_data;
	for (bcount = bp->b_bcount; bcount > 0; bcount -= rcount) {
		ccdbuffer(cbp, cs, bp, bn, addr, bcount);
		rcount = cbp[0]->cb_buf.b_bcount;

		if (cs->sc_cflags & CCDF_MIRROR) {
			/*
			 * Mirroring.  Writes go to both disks, reads are
			 * taken from whichever disk seems most appropriate.
			 *
			 * We attempt to localize reads to the disk whos arm
			 * is nearest the read request.  We ignore seeks due
			 * to writes when making this determination and we
			 * also try to avoid hogging.
			 */
			if ((cbp[0]->cb_buf.b_flags & B_READ) == 0) {
				cbp[0]->cb_buf.b_vp->v_numoutput++;
				cbp[1]->cb_buf.b_vp->v_numoutput++;
				VOP_STRATEGY(cbp[0]->cb_buf.b_vp, 
				    &cbp[0]->cb_buf);
				VOP_STRATEGY(cbp[1]->cb_buf.b_vp, 
				    &cbp[1]->cb_buf);
			} else {
				int pick = cs->sc_pick;
				daddr_t range = cs->sc_size / 16;

				if (bn < cs->sc_blk[pick] - range ||
				    bn > cs->sc_blk[pick] + range
				) {
					cs->sc_pick = pick = 1 - pick;
				}
				cs->sc_blk[pick] = bn + btodb(rcount);
				VOP_STRATEGY(cbp[pick]->cb_buf.b_vp, 
				    &cbp[pick]->cb_buf);
			}
		} else {
			/*
			 * Not mirroring
			 */
			if ((cbp[0]->cb_buf.b_flags & B_READ) == 0)
				cbp[0]->cb_buf.b_vp->v_numoutput++;
			VOP_STRATEGY(cbp[0]->cb_buf.b_vp, &cbp[0]->cb_buf);
		}
		bn += btodb(rcount);
		addr += rcount;
	}
}

/*
 * Build a component buffer header.
 */
static void
ccdbuffer(cb, cs, bp, bn, addr, bcount)
	struct ccdbuf **cb;
	struct ccd_softc *cs;
	struct buf *bp;
	daddr_t bn;
	caddr_t addr;
	long bcount;
{
	struct ccdcinfo *ci, *ci2 = NULL;	/* XXX */
	struct ccdbuf *cbp;
	daddr_t cbn, cboff;
	off_t cbc;

#ifdef DEBUG
	if (ccddebug & CCDB_IO)
		printf("ccdbuffer(%x, %x, %d, %x, %d)\n",
		       cs, bp, bn, addr, bcount);
#endif
	/*
	 * Determine which component bn falls in.
	 */
	cbn = bn;
	cboff = 0;

	if (cs->sc_ileave == 0) {
		/*
		 * Serially concatenated and neither a mirror nor a parity
		 * config.  This is a special case.
		 */
		daddr_t sblk;

		sblk = 0;
		for (ci = cs->sc_cinfo; cbn >= sblk + ci->ci_size; ci++)
			sblk += ci->ci_size;
		cbn -= sblk;
	} else {
		struct ccdiinfo *ii;
		int ccdisk, off;

		/*
		 * Calculate cbn, the logical superblock (sc_ileave chunks),
		 * and cboff, a normal block offset (DEV_BSIZE chunks) relative
		 * to cbn.
		 */
		cboff = cbn % cs->sc_ileave;	/* DEV_BSIZE gran */
		cbn = cbn / cs->sc_ileave;	/* DEV_BSIZE * ileave gran */

		/*
		 * Figure out which interleave table to use.
		 */
		for (ii = cs->sc_itable; ii->ii_ndisk; ii++) {
			if (ii->ii_startblk > cbn)
				break;
		}
		ii--;

		/*
		 * off is the logical superblock relative to the beginning 
		 * of this interleave block.  
		 */
		off = cbn - ii->ii_startblk;

		/*
		 * We must calculate which disk component to use (ccdisk),
		 * and recalculate cbn to be the superblock relative to
		 * the beginning of the component.  This is typically done by
		 * adding 'off' and ii->ii_startoff together.  However, 'off'
		 * must typically be divided by the number of components in
		 * this interleave array to be properly convert it from a
		 * CCD-relative logical superblock number to a 
		 * component-relative superblock number.
		 */
		if (ii->ii_ndisk == 1) {
			/*
			 * When we have just one disk, it can't be a mirror
			 * or a parity config.
			 */
			ccdisk = ii->ii_index[0];
			cbn = ii->ii_startoff + off;
		} else {
			if (cs->sc_cflags & CCDF_MIRROR) {
				/*
				 * We have forced a uniform mapping, resulting
				 * in a single interleave array.  We double
				 * up on the first half of the available
				 * components and our mirror is in the second
				 * half.  This only works with a single 
				 * interleave array because doubling up
				 * doubles the number of sectors, so there
				 * cannot be another interleave array because
				 * the next interleave array's calculations
				 * would be off.
				 */
				int ndisk2 = ii->ii_ndisk / 2;
				ccdisk = ii->ii_index[off % ndisk2];
				cbn = ii->ii_startoff + off / ndisk2;
				ci2 = &cs->sc_cinfo[ccdisk + ndisk2];
			} else if (cs->sc_cflags & CCDF_PARITY) {
				/* 
				 * XXX not implemented yet
				 */
				int ndisk2 = ii->ii_ndisk - 1;
				ccdisk = ii->ii_index[off % ndisk2];
				cbn = ii->ii_startoff + off / ndisk2;
				if (cbn % ii->ii_ndisk <= ccdisk)
					ccdisk++;
			} else {
				ccdisk = ii->ii_index[off % ii->ii_ndisk];
				cbn = ii->ii_startoff + off / ii->ii_ndisk;
			}
		}

		ci = &cs->sc_cinfo[ccdisk];

		/*
		 * Convert cbn from a superblock to a normal block so it
		 * can be used to calculate (along with cboff) the normal
		 * block index into this particular disk.
		 */
		cbn *= cs->sc_ileave;
	}

	/*
	 * Fill in the component buf structure.
	 */
	cbp = getccdbuf(NULL);
	cbp->cb_buf.b_flags = bp->b_flags | B_CALL;
	cbp->cb_buf.b_iodone = (void (*)(struct buf *))ccdiodone;
	cbp->cb_buf.b_dev = ci->ci_dev;		/* XXX */
	cbp->cb_buf.b_blkno = cbn + cboff + CCD_OFFSET;
	cbp->cb_buf.b_offset = dbtob(cbn + cboff + CCD_OFFSET);
	cbp->cb_buf.b_data = addr;
	cbp->cb_buf.b_vp = ci->ci_vp;
	if (cs->sc_ileave == 0)
              cbc = dbtob((off_t)(ci->ci_size - cbn));
	else
              cbc = dbtob((off_t)(cs->sc_ileave - cboff));
	cbp->cb_buf.b_bcount = (cbc < bcount) ? cbc : bcount;
 	cbp->cb_buf.b_bufsize = cbp->cb_buf.b_bcount;

	/*
	 * context for ccdiodone
	 */
	cbp->cb_obp = bp;
	cbp->cb_unit = cs - ccd_softc;
	cbp->cb_comp = ci - cs->sc_cinfo;

#ifdef DEBUG
	if (ccddebug & CCDB_IO)
		printf(" dev %x(u%d): cbp %x bn %d addr %x bcnt %d\n",
		       ci->ci_dev, ci-cs->sc_cinfo, cbp, cbp->cb_buf.b_blkno,
		       cbp->cb_buf.b_data, cbp->cb_buf.b_bcount);
#endif
	cb[0] = cbp;

	/*
	 * Note: both I/O's setup when reading from mirror, but only one
	 * will be executed.
	 */
	if (cs->sc_cflags & CCDF_MIRROR) {
		/* mirror, setup second I/O */
		cbp = getccdbuf(cb[0]);
		cbp->cb_buf.b_dev = ci2->ci_dev;
		cbp->cb_buf.b_vp = ci2->ci_vp;
		cbp->cb_comp = ci2 - cs->sc_cinfo;
		cb[1] = cbp;
		/* link together the ccdbuf's and clear "mirror done" flag */
		cb[0]->cb_mirror = cb[1];
		cb[1]->cb_mirror = cb[0];
		cb[0]->cb_pflags &= ~CCDPF_MIRROR_DONE;
		cb[1]->cb_pflags &= ~CCDPF_MIRROR_DONE;
	}
}

static void
ccdintr(cs, bp)
	struct ccd_softc *cs;
	struct buf *bp;
{
#ifdef DEBUG
	if (ccddebug & CCDB_FOLLOW)
		printf("ccdintr(%x, %x)\n", cs, bp);
#endif
	/*
	 * Request is done for better or worse, wakeup the top half.
	 */
	if (bp->b_flags & B_ERROR)
		bp->b_resid = bp->b_bcount;
	devstat_end_transaction_buf(&cs->device_stats, bp);
	biodone(bp);
}

/*
 * Called at interrupt time.
 * Mark the component as done and if all components are done,
 * take a ccd interrupt.
 */
static void
ccdiodone(cbp)
	struct ccdbuf *cbp;
{
	struct buf *bp = cbp->cb_obp;
	int unit = cbp->cb_unit;
	int count, s;

	s = splbio();
#ifdef DEBUG
	if (ccddebug & CCDB_FOLLOW)
		printf("ccdiodone(%x)\n", cbp);
	if (ccddebug & CCDB_IO) {
		printf("ccdiodone: bp %x bcount %d resid %d\n",
		       bp, bp->b_bcount, bp->b_resid);
		printf(" dev %x(u%d), cbp %x bn %d addr %x bcnt %d\n",
		       cbp->cb_buf.b_dev, cbp->cb_comp, cbp,
		       cbp->cb_buf.b_blkno, cbp->cb_buf.b_data,
		       cbp->cb_buf.b_bcount);
	}
#endif
	/*
	 * If an error occured, report it.  If this is a mirrored 
	 * configuration and the first of two possible reads, do not
	 * set the error in the bp yet because the second read may
	 * succeed.
	 */

	if (cbp->cb_buf.b_flags & B_ERROR) {
		const char *msg = "";

		if ((ccd_softc[unit].sc_cflags & CCDF_MIRROR) &&
		    (cbp->cb_buf.b_flags & B_READ) &&
		    (cbp->cb_pflags & CCDPF_MIRROR_DONE) == 0) {
			/*
			 * We will try our read on the other disk down
			 * below, also reverse the default pick so if we 
			 * are doing a scan we do not keep hitting the
			 * bad disk first.
			 */
			struct ccd_softc *cs = &ccd_softc[unit];

			msg = ", trying other disk";
			cs->sc_pick = 1 - cs->sc_pick;
			cs->sc_blk[cs->sc_pick] = bp->b_blkno;
		} else {
			bp->b_flags |= B_ERROR;
			bp->b_error = cbp->cb_buf.b_error ? 
			    cbp->cb_buf.b_error : EIO;
		}
		printf("ccd%d: error %d on component %d block %d (ccd block %d)%s\n",
		       unit, bp->b_error, cbp->cb_comp, 
		       (int)cbp->cb_buf.b_blkno, bp->b_blkno, msg);
	}

	/*
	 * Process mirror.  If we are writing, I/O has been initiated on both
	 * buffers and we fall through only after both are finished.
	 *
	 * If we are reading only one I/O is initiated at a time.  If an
	 * error occurs we initiate the second I/O and return, otherwise 
	 * we free the second I/O without initiating it.
	 */

	if (ccd_softc[unit].sc_cflags & CCDF_MIRROR) {
		if ((cbp->cb_buf.b_flags & B_READ) == 0) {
			/*
			 * When writing, handshake with the second buffer
			 * to determine when both are done.  If both are not
			 * done, return here.
			 */
			if ((cbp->cb_pflags & CCDPF_MIRROR_DONE) == 0) {
				cbp->cb_mirror->cb_pflags |= CCDPF_MIRROR_DONE;
				putccdbuf(cbp);
				splx(s);
				return;
			}
		} else {
			/*
			 * When reading, either dispose of the second buffer
			 * or initiate I/O on the second buffer if an error 
			 * occured with this one.
			 */
			if ((cbp->cb_pflags & CCDPF_MIRROR_DONE) == 0) {
				if (cbp->cb_buf.b_flags & B_ERROR) {
					cbp->cb_mirror->cb_pflags |= 
					    CCDPF_MIRROR_DONE;
					VOP_STRATEGY(
					    cbp->cb_mirror->cb_buf.b_vp, 
					    &cbp->cb_mirror->cb_buf
					);
					putccdbuf(cbp);
					splx(s);
					return;
				} else {
					putccdbuf(cbp->cb_mirror);
					/* fall through */
				}
			}
		}
	}

	/*
	 * use b_bufsize to determine how big the original request was rather
	 * then b_bcount, because b_bcount may have been truncated for EOF.
	 *
	 * XXX We check for an error, but we do not test the resid for an
	 * aligned EOF condition.  This may result in character & block
	 * device access not recognizing EOF properly when read or written 
	 * sequentially, but will not effect filesystems.
	 */
	count = cbp->cb_buf.b_bufsize;
	putccdbuf(cbp);

	/*
	 * If all done, "interrupt".
	 */
	bp->b_resid -= count;
	if (bp->b_resid < 0)
		panic("ccdiodone: count");
	if (bp->b_resid == 0)
		ccdintr(&ccd_softc[unit], bp);
	splx(s);
}

static int
ccdioctl(dev, cmd, data, flag, p)
	dev_t dev;
	u_long cmd;
	caddr_t data;
	int flag;
	struct proc *p;
{
	int unit = ccdunit(dev);
	int i, j, lookedup = 0, error = 0;
	int part, pmask, s;
	struct ccd_softc *cs;
	struct ccd_ioctl *ccio = (struct ccd_ioctl *)data;
	struct ccddevice ccd;
	char **cpp;
	struct vnode **vpp;

	if (unit >= numccd)
		return (ENXIO);
	cs = &ccd_softc[unit];

	bzero(&ccd, sizeof(ccd));

	switch (cmd) {
	case CCDIOCSET:
		if (cs->sc_flags & CCDF_INITED)
			return (EBUSY);

		if ((flag & FWRITE) == 0)
			return (EBADF);

		if ((error = ccdlock(cs)) != 0)
			return (error);

		/* Fill in some important bits. */
		ccd.ccd_unit = unit;
		ccd.ccd_interleave = ccio->ccio_ileave;
		if (ccd.ccd_interleave == 0 &&
		    ((ccio->ccio_flags & CCDF_MIRROR) ||
		     (ccio->ccio_flags & CCDF_PARITY))) {
			printf("ccd%d: disabling mirror/parity, interleave is 0\n", unit);
			ccio->ccio_flags &= ~(CCDF_MIRROR | CCDF_PARITY);
		}
		if ((ccio->ccio_flags & CCDF_MIRROR) &&
		    (ccio->ccio_flags & CCDF_PARITY)) {
			printf("ccd%d: can't specify both mirror and parity, using mirror\n", unit);
			ccio->ccio_flags &= ~CCDF_PARITY;
		}
		if ((ccio->ccio_flags & (CCDF_MIRROR | CCDF_PARITY)) &&
		    !(ccio->ccio_flags & CCDF_UNIFORM)) {
			printf("ccd%d: mirror/parity forces uniform flag\n",
			       unit);
			ccio->ccio_flags |= CCDF_UNIFORM;
		}
		ccd.ccd_flags = ccio->ccio_flags & CCDF_USERMASK;

		/*
		 * Allocate space for and copy in the array of
		 * componet pathnames and device numbers.
		 */
		cpp = malloc(ccio->ccio_ndisks * sizeof(char *),
		    M_DEVBUF, M_WAITOK);
		vpp = malloc(ccio->ccio_ndisks * sizeof(struct vnode *),
		    M_DEVBUF, M_WAITOK);

		error = copyin((caddr_t)ccio->ccio_disks, (caddr_t)cpp,
		    ccio->ccio_ndisks * sizeof(char **));
		if (error) {
			free(vpp, M_DEVBUF);
			free(cpp, M_DEVBUF);
			ccdunlock(cs);
			return (error);
		}

#ifdef DEBUG
		if (ccddebug & CCDB_INIT)
			for (i = 0; i < ccio->ccio_ndisks; ++i)
				printf("ccdioctl: component %d: 0x%x\n",
				    i, cpp[i]);
#endif

		for (i = 0; i < ccio->ccio_ndisks; ++i) {
#ifdef DEBUG
			if (ccddebug & CCDB_INIT)
				printf("ccdioctl: lookedup = %d\n", lookedup);
#endif
			if ((error = ccdlookup(cpp[i], p, &vpp[i])) != 0) {
				for (j = 0; j < lookedup; ++j)
					(void)vn_close(vpp[j], FREAD|FWRITE,
					    p->p_ucred, p);
				free(vpp, M_DEVBUF);
				free(cpp, M_DEVBUF);
				ccdunlock(cs);
				return (error);
			}
			++lookedup;
		}
		ccd.ccd_cpp = cpp;
		ccd.ccd_vpp = vpp;
		ccd.ccd_ndev = ccio->ccio_ndisks;

		/*
		 * Initialize the ccd.  Fills in the softc for us.
		 */
		if ((error = ccdinit(&ccd, cpp, p)) != 0) {
			for (j = 0; j < lookedup; ++j)
				(void)vn_close(vpp[j], FREAD|FWRITE,
				    p->p_ucred, p);
			bzero(&ccd_softc[unit], sizeof(struct ccd_softc));
			free(vpp, M_DEVBUF);
			free(cpp, M_DEVBUF);
			ccdunlock(cs);
			return (error);
		}

		/*
		 * The ccd has been successfully initialized, so
		 * we can place it into the array and read the disklabel.
		 */
		bcopy(&ccd, &ccddevs[unit], sizeof(ccd));
		ccio->ccio_unit = unit;
		ccio->ccio_size = cs->sc_size;
		ccdgetdisklabel(dev);

		ccdunlock(cs);

		break;

	case CCDIOCCLR:
		if ((cs->sc_flags & CCDF_INITED) == 0)
			return (ENXIO);

		if ((flag & FWRITE) == 0)
			return (EBADF);

		if ((error = ccdlock(cs)) != 0)
			return (error);

		/* Don't unconfigure if any other partitions are open */
		part = ccdpart(dev);
		pmask = (1 << part);
		if ((cs->sc_openmask & ~pmask)) {
			ccdunlock(cs);
			return (EBUSY);
		}

		/*
		 * Free ccd_softc information and clear entry.
		 */

		/* Close the components and free their pathnames. */
		for (i = 0; i < cs->sc_nccdisks; ++i) {
			/*
			 * XXX: this close could potentially fail and
			 * cause Bad Things.  Maybe we need to force
			 * the close to happen?
			 */
#ifdef DEBUG
			if (ccddebug & CCDB_VNODE)
				vprint("CCDIOCCLR: vnode info",
				    cs->sc_cinfo[i].ci_vp);
#endif
			(void)vn_close(cs->sc_cinfo[i].ci_vp, FREAD|FWRITE,
			    p->p_ucred, p);
			free(cs->sc_cinfo[i].ci_path, M_DEVBUF);
		}

		/* Free interleave index. */
		for (i = 0; cs->sc_itable[i].ii_ndisk; ++i)
			free(cs->sc_itable[i].ii_index, M_DEVBUF);

		/* Free component info and interleave table. */
		free(cs->sc_cinfo, M_DEVBUF);
		free(cs->sc_itable, M_DEVBUF);
		cs->sc_flags &= ~CCDF_INITED;

		/*
		 * Free ccddevice information and clear entry.
		 */
		free(ccddevs[unit].ccd_cpp, M_DEVBUF);
		free(ccddevs[unit].ccd_vpp, M_DEVBUF);
		ccd.ccd_dk = -1;
		bcopy(&ccd, &ccddevs[unit], sizeof(ccd));

		/*
		 * And remove the devstat entry.
		 */
		devstat_remove_entry(&cs->device_stats);

		/* This must be atomic. */
		s = splhigh();
		ccdunlock(cs);
		bzero(cs, sizeof(struct ccd_softc));
		splx(s);

		break;

	case DIOCGDINFO:
		if ((cs->sc_flags & CCDF_INITED) == 0)
			return (ENXIO);

		*(struct disklabel *)data = cs->sc_label;
		break;

	case DIOCGPART:
		if ((cs->sc_flags & CCDF_INITED) == 0)
			return (ENXIO);

		((struct partinfo *)data)->disklab = &cs->sc_label;
		((struct partinfo *)data)->part =
		    &cs->sc_label.d_partitions[ccdpart(dev)];
		break;

	case DIOCWDINFO:
	case DIOCSDINFO:
		if ((cs->sc_flags & CCDF_INITED) == 0)
			return (ENXIO);

		if ((flag & FWRITE) == 0)
			return (EBADF);

		if ((error = ccdlock(cs)) != 0)
			return (error);

		cs->sc_flags |= CCDF_LABELLING;

		error = setdisklabel(&cs->sc_label,
		    (struct disklabel *)data, 0);
		if (error == 0) {
			if (cmd == DIOCWDINFO)
				error = writedisklabel(CCDLABELDEV(dev),
				    &cs->sc_label);
		}

		cs->sc_flags &= ~CCDF_LABELLING;

		ccdunlock(cs);

		if (error)
			return (error);
		break;

	case DIOCWLABEL:
		if ((cs->sc_flags & CCDF_INITED) == 0)
			return (ENXIO);

		if ((flag & FWRITE) == 0)
			return (EBADF);
		if (*(int *)data != 0)
			cs->sc_flags |= CCDF_WLABEL;
		else
			cs->sc_flags &= ~CCDF_WLABEL;
		break;

	default:
		return (ENOTTY);
	}

	return (0);
}

static int
ccdsize(dev)
	dev_t dev;
{
	struct ccd_softc *cs;
	int part, size;

	if (ccdopen(dev, 0, S_IFCHR, curproc))
		return (-1);

	cs = &ccd_softc[ccdunit(dev)];
	part = ccdpart(dev);

	if ((cs->sc_flags & CCDF_INITED) == 0)
		return (-1);

	if (cs->sc_label.d_partitions[part].p_fstype != FS_SWAP)
		size = -1;
	else
		size = cs->sc_label.d_partitions[part].p_size;

	if (ccdclose(dev, 0, S_IFCHR, curproc))
		return (-1);

	return (size);
}

static int
ccddump(dev)
	dev_t dev;
{

	/* Not implemented. */
	return ENXIO;
}

/*
 * Lookup the provided name in the filesystem.  If the file exists,
 * is a valid block device, and isn't being used by anyone else,
 * set *vpp to the file's vnode.
 */
static int
ccdlookup(path, p, vpp)
	char *path;
	struct proc *p;
	struct vnode **vpp;	/* result */
{
	struct nameidata nd;
	struct vnode *vp;
	int error;

	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, path, p);
	if ((error = vn_open(&nd, FREAD|FWRITE, 0)) != 0) {
#ifdef DEBUG
		if (ccddebug & CCDB_FOLLOW|CCDB_INIT)
			printf("ccdlookup: vn_open error = %d\n", error);
#endif
		return (error);
	}
	vp = nd.ni_vp;

	if (vp->v_usecount > 1) {
		error = EBUSY;
		goto bad;
	}

	if (!vn_isdisk(vp, &error)) 
		goto bad;

#ifdef DEBUG
	if (ccddebug & CCDB_VNODE)
		vprint("ccdlookup: vnode info", vp);
#endif

	VOP_UNLOCK(vp, 0, p);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	*vpp = vp;
	return (0);
bad:
	VOP_UNLOCK(vp, 0, p);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	/* vn_close does vrele() for vp */
	(void)vn_close(vp, FREAD|FWRITE, p->p_ucred, p);
	return (error);
}

/*
 * Read the disklabel from the ccd.  If one is not present, fake one
 * up.
 */
static void
ccdgetdisklabel(dev)
	dev_t dev;
{
	int unit = ccdunit(dev);
	struct ccd_softc *cs = &ccd_softc[unit];
	char *errstring;
	struct disklabel *lp = &cs->sc_label;
	struct ccdgeom *ccg = &cs->sc_geom;

	bzero(lp, sizeof(*lp));

	lp->d_secperunit = cs->sc_size;
	lp->d_secsize = ccg->ccg_secsize;
	lp->d_nsectors = ccg->ccg_nsectors;
	lp->d_ntracks = ccg->ccg_ntracks;
	lp->d_ncylinders = ccg->ccg_ncylinders;
	lp->d_secpercyl = lp->d_ntracks * lp->d_nsectors;

	strncpy(lp->d_typename, "ccd", sizeof(lp->d_typename));
	lp->d_type = DTYPE_CCD;
	strncpy(lp->d_packname, "fictitious", sizeof(lp->d_packname));
	lp->d_rpm = 3600;
	lp->d_interleave = 1;
	lp->d_flags = 0;

	lp->d_partitions[RAW_PART].p_offset = 0;
	lp->d_partitions[RAW_PART].p_size = cs->sc_size;
	lp->d_partitions[RAW_PART].p_fstype = FS_UNUSED;
	lp->d_npartitions = RAW_PART + 1;

	lp->d_bbsize = BBSIZE;				/* XXX */
	lp->d_sbsize = SBSIZE;				/* XXX */

	lp->d_magic = DISKMAGIC;
	lp->d_magic2 = DISKMAGIC;
	lp->d_checksum = dkcksum(&cs->sc_label);

	/*
	 * Call the generic disklabel extraction routine.
	 */
	errstring = readdisklabel(CCDLABELDEV(dev), &cs->sc_label);
	if (errstring != NULL)
		ccdmakedisklabel(cs);

#ifdef DEBUG
	/* It's actually extremely common to have unlabeled ccds. */
	if (ccddebug & CCDB_LABEL)
		if (errstring != NULL)
			printf("ccd%d: %s\n", unit, errstring);
#endif
}

/*
 * Take care of things one might want to take care of in the event
 * that a disklabel isn't present.
 */
static void
ccdmakedisklabel(cs)
	struct ccd_softc *cs;
{
	struct disklabel *lp = &cs->sc_label;

	/*
	 * For historical reasons, if there's no disklabel present
	 * the raw partition must be marked FS_BSDFFS.
	 */
	lp->d_partitions[RAW_PART].p_fstype = FS_BSDFFS;

	strncpy(lp->d_packname, "default label", sizeof(lp->d_packname));
}

/*
 * Wait interruptibly for an exclusive lock.
 *
 * XXX
 * Several drivers do this; it should be abstracted and made MP-safe.
 */
static int
ccdlock(cs)
	struct ccd_softc *cs;
{
	int error;

	while ((cs->sc_flags & CCDF_LOCKED) != 0) {
		cs->sc_flags |= CCDF_WANTED;
		if ((error = tsleep(cs, PRIBIO | PCATCH, "ccdlck", 0)) != 0)
			return (error);
	}
	cs->sc_flags |= CCDF_LOCKED;
	return (0);
}

/*
 * Unlock and wake up any waiters.
 */
static void
ccdunlock(cs)
	struct ccd_softc *cs;
{

	cs->sc_flags &= ~CCDF_LOCKED;
	if ((cs->sc_flags & CCDF_WANTED) != 0) {
		cs->sc_flags &= ~CCDF_WANTED;
		wakeup(cs);
	}
}

#ifdef DEBUG
static void
printiinfo(ii)
	struct ccdiinfo *ii;
{
	int ix, i;

	for (ix = 0; ii->ii_ndisk; ix++, ii++) {
		printf(" itab[%d]: #dk %d sblk %d soff %d",
		       ix, ii->ii_ndisk, ii->ii_startblk, ii->ii_startoff);
		for (i = 0; i < ii->ii_ndisk; i++)
			printf(" %d", ii->ii_index[i]);
		printf("\n");
	}
}
#endif


/* Local Variables: */
/* c-argdecl-indent: 8 */
/* c-continued-statement-offset: 8 */
/* c-indent-level: 8 */
/* End: */
OpenPOWER on IntegriCloud