summaryrefslogtreecommitdiff
path: root/sys/dev/softraid_raid5.c
blob: 533b926f00bd17b28880394685314890bdd371d8 (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
/* $OpenBSD: softraid_raid5.c,v 1.4 2014/01/19 11:43:05 jsing Exp $ */
/*
 * Copyright (c) 2009 Marco Peereboom <marco@peereboom.us>
 * Copyright (c) 2009 Jordan Hargrave <jordan@openbsd.org>
 *
 * Permission to use, copy, modify, and distribute this software for any
 * purpose with or without fee is hereby granted, provided that the above
 * copyright notice and this permission notice appear in all copies.
 *
 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
 */

#include "bio.h"

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/buf.h>
#include <sys/device.h>
#include <sys/ioctl.h>
#include <sys/proc.h>
#include <sys/malloc.h>
#include <sys/kernel.h>
#include <sys/disk.h>
#include <sys/rwlock.h>
#include <sys/queue.h>
#include <sys/fcntl.h>
#include <sys/disklabel.h>
#include <sys/mount.h>
#include <sys/sensors.h>
#include <sys/stat.h>
#include <sys/pool.h>
#include <sys/conf.h>
#include <sys/uio.h>

#include <scsi/scsi_all.h>
#include <scsi/scsiconf.h>
#include <scsi/scsi_disk.h>

#include <dev/softraidvar.h>
#include <dev/rndvar.h>

/* RAID 5 functions. */
int	sr_raid5_create(struct sr_discipline *, struct bioc_createraid *,
	    int, int64_t);
int	sr_raid5_assemble(struct sr_discipline *, struct bioc_createraid *,
	    int, void *);
int	sr_raid5_init(struct sr_discipline *);
int	sr_raid5_rw(struct sr_workunit *);
int	sr_raid5_openings(struct sr_discipline *);
void	sr_raid5_intr(struct buf *);
int	sr_raid5_wu_done(struct sr_workunit *);
void	sr_raid5_set_chunk_state(struct sr_discipline *, int, int);
void	sr_raid5_set_vol_state(struct sr_discipline *);

int	sr_raid5_addio(struct sr_workunit *wu, int, daddr_t, daddr_t,
	    void *, int, int, void *);
int	sr_raid5_regenerate(struct sr_workunit *, int, daddr_t, daddr_t,
	    void *);
void	sr_raid5_xor(void *, void *, int);

void	sr_dump(void *, int);
void	sr_raid5_scrub(struct sr_discipline *);

/* discipline initialisation. */
void
sr_raid5_discipline_init(struct sr_discipline *sd)
{
	/* Fill out discipline members. */
	sd->sd_type = SR_MD_RAID5;
	strlcpy(sd->sd_name, "RAID 5", sizeof(sd->sd_name));
	sd->sd_capabilities = SR_CAP_SYSTEM_DISK | SR_CAP_AUTO_ASSEMBLE |
	    SR_CAP_REDUNDANT;
	sd->sd_max_ccb_per_wu = 4; /* only if stripsize <= MAXPHYS */
	sd->sd_max_wu = SR_RAID5_NOWU;

	/* Setup discipline specific function pointers. */
	sd->sd_assemble = sr_raid5_assemble;
	sd->sd_create = sr_raid5_create;
	sd->sd_openings = sr_raid5_openings;
	sd->sd_scsi_rw = sr_raid5_rw;
	sd->sd_scsi_intr = sr_raid5_intr;
	sd->sd_scsi_wu_done = sr_raid5_wu_done;
	sd->sd_set_chunk_state = sr_raid5_set_chunk_state;
	sd->sd_set_vol_state = sr_raid5_set_vol_state;
}

int
sr_raid5_create(struct sr_discipline *sd, struct bioc_createraid *bc,
    int no_chunk, int64_t coerced_size)
{
	if (no_chunk < 3) {
		sr_error(sd->sd_sc, "%s requires three or more chunks",
		    sd->sd_name);
		return EINVAL;
	}

	/*
	 * XXX add variable strip size later even though MAXPHYS is really
	 * the clever value, users like to tinker with that type of stuff.
	 */
	sd->sd_meta->ssdi.ssd_strip_size = MAXPHYS;
	sd->sd_meta->ssdi.ssd_size = (coerced_size &
	    ~(((u_int64_t)sd->sd_meta->ssdi.ssd_strip_size >>
	    DEV_BSHIFT) - 1)) * (no_chunk - 1);

	return sr_raid5_init(sd);
}

int
sr_raid5_assemble(struct sr_discipline *sd, struct bioc_createraid *bc,
    int no_chunk, void *data)
{
	return sr_raid5_init(sd);
}

int
sr_raid5_init(struct sr_discipline *sd)
{
	/* Initialise runtime values. */
	sd->mds.mdd_raid5.sr5_strip_bits =
	    sr_validate_stripsize(sd->sd_meta->ssdi.ssd_strip_size);
	if (sd->mds.mdd_raid5.sr5_strip_bits == -1) {
		sr_error(sd->sd_sc, "invalid strip size");
		return EINVAL;
	}

	return 0;
}

int
sr_raid5_openings(struct sr_discipline *sd)
{
	return (sd->sd_max_wu >> 1); /* 2 wu's per IO */
}

void
sr_raid5_set_chunk_state(struct sr_discipline *sd, int c, int new_state)
{
	int			old_state, s;

	DNPRINTF(SR_D_STATE, "%s: %s: %s: sr_raid_set_chunk_state %d -> %d\n",
	    DEVNAME(sd->sd_sc), sd->sd_meta->ssd_devname,
	    sd->sd_vol.sv_chunks[c]->src_meta.scmi.scm_devname, c, new_state);

	/* ok to go to splbio since this only happens in error path */
	s = splbio();
	old_state = sd->sd_vol.sv_chunks[c]->src_meta.scm_status;

	/* multiple IOs to the same chunk that fail will come through here */
	if (old_state == new_state)
		goto done;

	switch (old_state) {
	case BIOC_SDONLINE:
		switch (new_state) {
		case BIOC_SDOFFLINE:
		case BIOC_SDSCRUB:
			break;
		default:
			goto die;
		}
		break;

	case BIOC_SDOFFLINE:
		if (new_state == BIOC_SDREBUILD) {
			;
		} else
			goto die;
		break;

	case BIOC_SDSCRUB:
		switch (new_state) {
		case BIOC_SDONLINE:
		case BIOC_SDOFFLINE:
			break;
		default:
			goto die;
		}
		break;

	case BIOC_SDREBUILD:
		switch (new_state) {
		case BIOC_SDONLINE:
		case BIOC_SDOFFLINE:
			break;
		default:
			goto die;
		}
		break;

	default:
die:
		splx(s); /* XXX */
		panic("%s: %s: %s: invalid chunk state transition "
		    "%d -> %d", DEVNAME(sd->sd_sc),
		    sd->sd_meta->ssd_devname,
		    sd->sd_vol.sv_chunks[c]->src_meta.scmi.scm_devname,
		    old_state, new_state);
		/* NOTREACHED */
	}

	sd->sd_vol.sv_chunks[c]->src_meta.scm_status = new_state;
	sd->sd_set_vol_state(sd);

	sd->sd_must_flush = 1;
	workq_add_task(NULL, 0, sr_meta_save_callback, sd, NULL);
done:
	splx(s);
}

void
sr_raid5_set_vol_state(struct sr_discipline *sd)
{
	int			states[SR_MAX_STATES];
	int			new_state, i, s, nd;
	int			old_state = sd->sd_vol_status;

	DNPRINTF(SR_D_STATE, "%s: %s: sr_raid_set_vol_state\n",
	    DEVNAME(sd->sd_sc), sd->sd_meta->ssd_devname);

	nd = sd->sd_meta->ssdi.ssd_chunk_no;

	for (i = 0; i < SR_MAX_STATES; i++)
		states[i] = 0;

	for (i = 0; i < nd; i++) {
		s = sd->sd_vol.sv_chunks[i]->src_meta.scm_status;
		if (s >= SR_MAX_STATES)
			panic("%s: %s: %s: invalid chunk state",
			    DEVNAME(sd->sd_sc),
			    sd->sd_meta->ssd_devname,
			    sd->sd_vol.sv_chunks[i]->src_meta.scmi.scm_devname);
		states[s]++;
	}

	if (states[BIOC_SDONLINE] == nd)
		new_state = BIOC_SVONLINE;
	else if (states[BIOC_SDONLINE] < nd - 1)
		new_state = BIOC_SVOFFLINE;
	else if (states[BIOC_SDSCRUB] != 0)
		new_state = BIOC_SVSCRUB;
	else if (states[BIOC_SDREBUILD] != 0)
		new_state = BIOC_SVREBUILD;
	else if (states[BIOC_SDONLINE] == nd - 1)
		new_state = BIOC_SVDEGRADED;
	else {
#ifdef SR_DEBUG
		DNPRINTF(SR_D_STATE, "%s: invalid volume state, old state "
		    "was %d\n", DEVNAME(sd->sd_sc), old_state);
		for (i = 0; i < nd; i++)
			DNPRINTF(SR_D_STATE, "%s: chunk %d status = %d\n",
			    DEVNAME(sd->sd_sc), i,
			    sd->sd_vol.sv_chunks[i]->src_meta.scm_status);
#endif
		panic("invalid volume state");
	}

	DNPRINTF(SR_D_STATE, "%s: %s: sr_raid5_set_vol_state %d -> %d\n",
	    DEVNAME(sd->sd_sc), sd->sd_meta->ssd_devname,
	    old_state, new_state);

	switch (old_state) {
	case BIOC_SVONLINE:
		switch (new_state) {
		case BIOC_SVONLINE: /* can go to same state */
		case BIOC_SVDEGRADED:
		case BIOC_SVREBUILD: /* happens on boot */
			break;
		default:
			goto die;
		}
		break;

	case BIOC_SVOFFLINE:
		/* XXX this might be a little too much */
		goto die;

	case BIOC_SVSCRUB:
		switch (new_state) {
		case BIOC_SVONLINE:
		case BIOC_SVOFFLINE:
		case BIOC_SVDEGRADED:
		case BIOC_SVSCRUB: /* can go to same state */
			break;
		default:
			goto die;
		}
		break;

	case BIOC_SVREBUILD:
		switch (new_state) {
		case BIOC_SVONLINE:
		case BIOC_SVOFFLINE:
		case BIOC_SVDEGRADED:
		case BIOC_SVREBUILD: /* can go to the same state */
			break;
		default:
			goto die;
		}
		break;

	case BIOC_SVDEGRADED:
		switch (new_state) {
		case BIOC_SVOFFLINE:
		case BIOC_SVREBUILD:
		case BIOC_SVDEGRADED: /* can go to the same state */
			break;
		default:
			goto die;
		}
		break;

	default:
die:
		panic("%s: %s: invalid volume state transition %d -> %d",
		    DEVNAME(sd->sd_sc), sd->sd_meta->ssd_devname,
		    old_state, new_state);
		/* NOTREACHED */
	}

	sd->sd_vol_status = new_state;
}

static inline int
sr_raid5_chunk_online(struct sr_discipline *sd, int chunk)
{
	switch (sd->sd_vol.sv_chunks[chunk]->src_meta.scm_status) {
	case BIOC_SDONLINE:
	case BIOC_SDSCRUB:
		return 1;
	default:
		return 0;
	}
}

int
sr_raid5_rw(struct sr_workunit *wu)
{
	struct sr_workunit	*wu_r = NULL;
	struct sr_discipline	*sd = wu->swu_dis;
	struct scsi_xfer	*xs = wu->swu_xs;
	struct sr_chunk		*scp;
	daddr_t			blk, lba;
	int64_t			chunk_offs, lbaoffs, phys_offs, strip_offs;
	int64_t			strip_bits, strip_no, strip_size;
	int64_t			chunk, no_chunk;
	int64_t			length, parity, datalen, row_size;
	void			*xorbuf, *data;
	int			s;

	/* blk and scsi error will be handled by sr_validate_io */
	if (sr_validate_io(wu, &blk, "sr_raid5_rw"))
		goto bad;

	strip_size = sd->sd_meta->ssdi.ssd_strip_size;
	strip_bits = sd->mds.mdd_raid5.sr5_strip_bits;
	no_chunk = sd->sd_meta->ssdi.ssd_chunk_no - 1;
	row_size = (no_chunk << strip_bits) >> DEV_BSHIFT;

	data = xs->data;
	datalen = xs->datalen;
	lbaoffs	= blk << DEV_BSHIFT;

	if (xs->flags & SCSI_DATA_OUT) {
		if ((wu_r = sr_scsi_wu_get(sd, SCSI_NOSLEEP)) == NULL){
			printf("%s: can't get wu_r", DEVNAME(sd->sd_sc));
			goto bad;
		}
		wu_r->swu_state = SR_WU_INPROGRESS;
		wu_r->swu_flags |= SR_WUF_DISCIPLINE;
	}

	wu->swu_blk_start = 0;
	while (datalen != 0) {
		strip_no = lbaoffs >> strip_bits;
		strip_offs = lbaoffs & (strip_size - 1);
		chunk_offs = (strip_no / no_chunk) << strip_bits;
		phys_offs = chunk_offs + strip_offs +
		    (sd->sd_meta->ssd_data_offset << DEV_BSHIFT);

		/* get size remaining in this stripe */
		length = MIN(strip_size - strip_offs, datalen);

		/* map disk offset to parity/data drive */
		chunk = strip_no % no_chunk;

		/* RAID5 - left asymmetric algorithm */
		parity = no_chunk - ((strip_no / no_chunk) % (no_chunk + 1));
		if (chunk >= parity)
			chunk++;

		lba = phys_offs >> DEV_BSHIFT;

		/* XXX big hammer.. exclude I/O from entire stripe */
		if (wu->swu_blk_start == 0)
			wu->swu_blk_start = (strip_no / no_chunk) * row_size;
		wu->swu_blk_end = (strip_no / no_chunk) * row_size + (row_size - 1);

		scp = sd->sd_vol.sv_chunks[chunk];
		if (xs->flags & SCSI_DATA_IN) {
			switch (scp->src_meta.scm_status) {
			case BIOC_SDONLINE:
			case BIOC_SDSCRUB:
				/* drive is good. issue single read request */
				if (sr_raid5_addio(wu, chunk, lba, length,
				    data, xs->flags, 0, NULL))
					goto bad;
				break;
			case BIOC_SDOFFLINE:
			case BIOC_SDREBUILD:
			case BIOC_SDHOTSPARE:
				if (sr_raid5_regenerate(wu, chunk, lba,
				    length, data))
					goto bad;
				break;
			default:
				printf("%s: is offline, can't read\n",
				    DEVNAME(sd->sd_sc));
				goto bad;
			}
		} else {
			/* XXX handle writes to failed/offline disk? */
			if (scp->src_meta.scm_status == BIOC_SDOFFLINE)
				goto bad;

			/*
			 * initialize XORBUF with contents of new data to be
			 * written. This will be XORed with old data and old
			 * parity in the intr routine. The result in xorbuf
			 * is the new parity data.
			 */
			xorbuf = sr_block_get(sd, length);
			if (xorbuf == NULL)
				goto bad;
			memcpy(xorbuf, data, length);

			/* xor old data */
			if (sr_raid5_addio(wu_r, chunk, lba, length, NULL,
			    SCSI_DATA_IN, 0, xorbuf))
				goto bad;

			/* xor old parity */
			if (sr_raid5_addio(wu_r, parity, lba, length, NULL,
			    SCSI_DATA_IN, 0, xorbuf))
				goto bad;

			/* write new data */
			if (sr_raid5_addio(wu, chunk, lba, length, data,
			    xs->flags, 0, NULL))
				goto bad;

			/* write new parity */
			if (sr_raid5_addio(wu, parity, lba, length, xorbuf,
			    xs->flags, SR_CCBF_FREEBUF, NULL))
				goto bad;
		}

		/* advance to next block */
		lbaoffs += length;
		datalen -= length;
		data += length;
	}

	s = splbio();
	if (wu_r) {
		if (wu_r->swu_io_count > 0) {
			/* collide write request with reads */
			wu_r->swu_blk_start = wu->swu_blk_start;
			wu_r->swu_blk_end = wu->swu_blk_end;

			wu->swu_state = SR_WU_DEFERRED;
			wu_r->swu_collider = wu;
			TAILQ_INSERT_TAIL(&sd->sd_wu_defq, wu, swu_link);

			wu = wu_r;
		} else {
			sr_scsi_wu_put(sd, wu_r);
		}
	}
	splx(s);

	sr_schedule_wu(wu);

	return (0);

bad:
	/* XXX - can leak xorbuf on error. */
	/* wu is unwound by sr_wu_put */
	if (wu_r)
		sr_scsi_wu_put(sd, wu_r);
	return (1);
}

int
sr_raid5_regenerate(struct sr_workunit *wu, int chunk, daddr_t blkno,
    daddr_t len, void *data)
{
	struct sr_discipline	*sd = wu->swu_dis;
	int			i;

	/*
	 * Regenerate a block on a RAID 5 volume by xoring the data and parity
	 * from all of the remaining online chunks. This requires the parity
	 * to already be correct.
 	 */

	DNPRINTF(SR_D_DIS, "%s: %s sr_raid5_regenerate chunk %d offline, "
	    "regenerating block %llu\n",
	    DEVNAME(sd->sd_sc), sd->sd_meta->ssd_devname, chunk, blkno);

	memset(data, 0, len);
	for (i = 0; i < sd->sd_meta->ssdi.ssd_chunk_no; i++) {
		if (i == chunk)
			continue;
		if (!sr_raid5_chunk_online(sd, i))
			goto bad;
		if (sr_raid5_addio(wu, i, blkno, len, NULL, SCSI_DATA_IN,
		    0, data))
			goto bad;
	}
	return (0);

bad:
	return (1);
}

void
sr_raid5_intr(struct buf *bp)
{
	struct sr_ccb		*ccb = (struct sr_ccb *)bp;
	struct sr_workunit	*wu = ccb->ccb_wu;
	struct sr_discipline	*sd = wu->swu_dis;
	int			s;

	DNPRINTF(SR_D_INTR, "%s: sr_raid5_intr bp %p xs %p\n",
	    DEVNAME(sd->sd_sc), bp, wu->swu_xs);

	s = splbio();
	sr_ccb_done(ccb);

	/* XXX - Should this be done via the workq? */

	/* XOR data to result. */
	if (ccb->ccb_state == SR_CCB_OK && ccb->ccb_opaque)
		sr_raid5_xor(ccb->ccb_opaque, ccb->ccb_buf.b_data,
		    ccb->ccb_buf.b_bcount);

	/* Free allocated data buffer. */
	if (ccb->ccb_flags & SR_CCBF_FREEBUF) {
		sr_block_put(sd, ccb->ccb_buf.b_data, ccb->ccb_buf.b_bcount);
		ccb->ccb_buf.b_data = NULL;
	}

	sr_wu_done(wu);
	splx(s);
}

int
sr_raid5_wu_done(struct sr_workunit *wu)
{
	struct sr_discipline	*sd = wu->swu_dis;
	struct scsi_xfer	*xs = wu->swu_xs;

	/* XXX - we have no way of propagating errors... */
	if (wu->swu_flags & SR_WUF_DISCIPLINE)
		return SR_WU_OK;

	/* XXX - This is insufficient for RAID 5. */
	if (wu->swu_ios_succeeded > 0) {
		xs->error = XS_NOERROR;
		return SR_WU_OK;
	}

	if (xs->flags & SCSI_DATA_IN) {
		printf("%s: retrying read on block %lld\n",
		    sd->sd_meta->ssd_devname, (long long)wu->swu_blk_start);
		sr_wu_release_ccbs(wu);
		wu->swu_state = SR_WU_RESTART;
		if (sd->sd_scsi_rw(wu) == 0)
			return SR_WU_RESTART;
	} else {
		printf("%s: permanently fail write on block %lld\n",
		    sd->sd_meta->ssd_devname, (long long)wu->swu_blk_start);
	}

	wu->swu_state = SR_WU_FAILED;
	xs->error = XS_DRIVER_STUFFUP;

	return SR_WU_FAILED;
}

int
sr_raid5_addio(struct sr_workunit *wu, int chunk, daddr_t blkno,
    daddr_t len, void *data, int xsflags, int ccbflags, void *xorbuf)
{
	struct sr_discipline	*sd = wu->swu_dis;
	struct sr_ccb		*ccb;

	DNPRINTF(SR_D_DIS, "sr_raid5_addio: %s chunk %d block %lld "
	    "length %lld %s\n", (xsflags & SCSI_DATA_IN) ? "read" : "write",
	    chunk, (long long)blkno, (long long)len, xorbuf ? "X0R" : "-");

	/* Allocate temporary buffer. */
	if (data == NULL) {
		data = sr_block_get(sd, len);
		if (data == NULL)
			return (-1);
		ccbflags |= SR_CCBF_FREEBUF;
	}

	ccb = sr_ccb_rw(sd, chunk, blkno, len, data, xsflags, ccbflags);
	if (ccb == NULL) {
		if (ccbflags & SR_CCBF_FREEBUF)
			sr_block_put(sd, data, len);
		return (-1);
	}
	ccb->ccb_opaque = xorbuf;
	sr_wu_enqueue_ccb(wu, ccb);

	return (0);
}

void
sr_dump(void *blk, int len)
{
	uint8_t			*b = blk;
	int			i, j, c;

	for (i = 0; i < len; i += 16) {
		for (j = 0; j < 16; j++)
			printf("%.2x ", b[i + j]);
		printf("  ");
		for (j = 0; j < 16; j++) {
			c = b[i + j];
			if (c < ' ' || c > 'z' || i + j > len)
				c = '.';
			printf("%c", c);
		}
		printf("\n");
	}
}

void
sr_raid5_xor(void *a, void *b, int len)
{
	uint32_t		*xa = a, *xb = b;

	len >>= 2;
	while (len--)
		*xa++ ^= *xb++;
}

#if 0
void
sr_raid5_scrub(struct sr_discipline *sd)
{
	int64_t strip_no, strip_size, no_chunk, parity, max_strip, strip_bits;
	int64_t i;
	struct sr_workunit *wu_r, *wu_w;
	int s, slept;
	void *xorbuf;

	wu_w = sr_scsi_wu_get(sd, 0);
	wu_r = sr_scsi_wu_get(sd, 0);

	no_chunk = sd->sd_meta->ssdi.ssd_chunk_no - 1;
	strip_size = sd->sd_meta->ssdi.ssd_strip_size;
	strip_bits = sd->mds.mdd_raid5.sr5_strip_bits;
	max_strip = sd->sd_meta->ssdi.ssd_size >> strip_bits;

	for (strip_no = 0; strip_no < max_strip; strip_no++) {
		parity = no_chunk - ((strip_no / no_chunk) % (no_chunk + 1));

		xorbuf = sr_block_get(sd, strip_size);
		for (i = 0; i <= no_chunk; i++) {
			if (i != parity)
				sr_raid5_addio(wu_r, i, 0xBADCAFE, strip_size,
				    NULL, SCSI_DATA_IN, 0, xorbuf);
		}
		sr_raid5_addio(wu_w, parity, 0xBADCAFE, strip_size, xorbuf,
		    SCSI_DATA_OUT, SR_CCBF_FREEBUF, NULL);

		wu_r->swu_flags |= SR_WUF_REBUILD;

		/* Collide wu_w with wu_r */
		wu_w->swu_state = SR_WU_DEFERRED;
		wu_w->swu_flags |= SR_WUF_REBUILD | SR_WUF_WAKEUP;
		wu_r->swu_collider = wu_w;

		s = splbio();
		TAILQ_INSERT_TAIL(&sd->sd_wu_defq, wu_w, swu_link);
		splx(s);

		wu_r->swu_state = SR_WU_INPROGRESS;
		sr_schedule_wu(wu_r);

		slept = 0;
		while ((wu_w->swu_flags & SR_WUF_REBUILDIOCOMP) == 0) {
			tsleep(wu_w, PRIBIO, "sr_scrub", 0);
			slept = 1;
		}
		if (!slept)
			tsleep(sd->sd_sc, PWAIT, "sr_yield", 1);
	}
done:
	return;
}
#endif